MEDICAL PROGRESS FAT TRANSPORT IN LIPOPROTEINS - AN INTEGRATED APPROACH TO MECHANISMS AND DISORDERS* DONALD S. FREDRICKSON, M.D.,t ROBERT I. LEVY, M.D.,* AND ROBERT S. LEES, M.D.9 BETHESDA, MARYLAND T HE subjects .of this review are the plasma lipo- proteins, their structure and functions and the ways in which they are disordered in certain dis- eases. The intent is not to discuss lipoproteins for their own sake, however, but to exploit their poten- tial for illuminating the common and often frustrat- ing clinical problem of hyperlipidemia. The finding of an abnormal concentration in plasma of choles- terol, glycerides or a given class of the lipoproteins often raises questions of cause and relief that have no certain answer. These will not necessarily be forthcoming in this report. What will be attempted is the reduction of current information about fat transport and metabolism to the minimum terms needed by a physician to obtain a rational approach to the patient with hyperlipidemia and to keep abreast of new developments in this rapidly ex- panding field. *From the Laboratory of Molecular Diseases, National Heart Institute. tDirector and chief, Laboratory of Molecular Diseases, National Heart Institute. tHead, Section on Lipoproteins. Laboratory of Molecular Diseases, National Heart Institute. IAssistant professor and associate physician, Rockefeller University. The integration of information and concepts about normal mechanisms and clinical disorders will pro- ceed from more theoretical to more practical grounds. The first part of the review will outline the normal tasks of fat transport and describe how the several plasma lipids and certain proteins interact in their performances. The proteins that have evolved mainly to participate in transport of esterified lipids and the lipoproteins that they form will be closely examined. This will include analysis of several in- heritable diseases in which one of these proteins is deficient to gain perspective on the functions that they apparently serve. A detailed discussion of hyperlipidemia will follow. This will be based on an approach developed pri- marily for the study of genetically determined ab- normalities, but acquired or nonfamilial disorders, including changes in lipid concentrations secondary to other known disease, will be dealt with as well. All these disorders are translated into hyperlipopro- teinemia on the premise - for which supporting evidence will be presented - that lipoprotein pat- terns offer necessary information not provided by analyses of plasma lipids alone. Some simple new nomenclature is offered since the older terminology 276: 3244, Reprinted from the New England Journd of Medicine 94-103, M&156, 215-226, 273-261 (January 5, 12, 19, 26, February 2), 1967 obscures the heterogeneity that has recently been discovered. Attention will also be paid to practical steps to diagnosis and conservative therapy suitable for the great majority of patients with hyperlipopro- teinemia. It will be necessary to cover the enormous litera- ture in the area of this report in a selective rather than a comprehensive fashion. Whenever possible general references will be introduced with major topics and the number of specific citations held to a minimum. THE NORMAL FAT-TRANSPORT TASKS'.~ It seems wisest to begin by considering the kinds and amounts of lipids that move through the extra- cellular waterways. Most of the lipid in plasma at any one time is usually not in rapid transit from one tissue to another. Cholesterol and phospholipids, which represent about two thirds of the plasma lipid, have a much slower turhover than fatty acids. In quantitative terms the major fat-transport tasks are movement of free fatty acids and fatty acid es- ters of glycerol (glycerides). The lipoprotein con- centrations in plasma are directly and indirectly influenced by this traffic, which in turn depends on many factors, an important one of which is the state of carbohydrate metabolism. FREE FATTY ACIDS By far the greatest amount of fat transported through the plasma compartment is in the form of free fatty acids. This is despite their relatively in- significant contribution to the total plasma lipid concentration. In the fasting subject, there are from 0.3 to 0.7 mEq. per liter of plasma, about 8 to 20 mg. of the total lipid concentration of 400 to 800 mg. per 100 ml. The content of individual free fatty acids is similar but not identical to that of adipose- tissue triglycerides, from which they mainly origi- nate. These fatty acids do not actually circulate in the "free" state but bound to albumin, The free- fatty-acid-albumin complex is formed immediately when fatty acids are released into the bloodstream and the complex releases its fatty acid at sites of utilization, which include liver, muscle, heart and many other tissues. This process is under sensitive metabolic control that is rapidly adjusted to meet the body's ever-changing need for metabolic fuel. Release of free fatty acids depends acutely on the availability of insulin, but is further adjusted by the sympathetic nervous system and by circulating cate- cholamines; adrenocortical hormones, thyroid hor- mone, glucagon and several anterior pituitary hor- mones also have permissive or secondary roles in mobilization of free fatty acids. In the postabsorptive state from 50 to 90 percent of the body's total energy needs are met by free fatty acids delivered from the adipose tissue. More than 25 gm. per hour may be transported through plasma during the day although much of this is re-esterified rather than oxidized immediately. The major direct contribution of the free fatty acids to plasma lipid concentrations is their conversion in the liver to glycerides. When flux of free fatty acids to the liver. is unusually high, in considerable ex- cess of the ability of that organ to burn them, the outpouring of such glyceride can be great, leading to a form of "endogenous hyperlipemia." Although the importance of free fatty acid transport cannot be overemphasized it would be tangential to our pur- poses to attempt to cover many details available in a number of reviews." Free fatty acids will be men- tioned further only when directly pertinent to lipo- protein metabolism and disorders thereof. Glyceride Transport Exogenous. From the first days of life a second major task must be performed, the disposal of glycerides that are ingested in the amount of I to 2 gm. per kilogram of body weight daily. These are hydrolyzed in the intestinal lumen and taken up along with smaller amounts of other lipids and lipid-soluble substances. In the mucosal cells glyc- erides are reformed and collected into particles. (The term particle is conventionally applied to cir- culating lipid-protein complexes large enough co be seen in the light microscope - that is, about 0.1 /.L or larger.5) The chylomicrons or "exogenous parti- cles" formed in the intestinal cells during fat ab- sorption are released into intestinal lymphatics and enter the bloodstream through the thoracic duct. Either at the capillary surface or immediately on entering cells in adipose tissue, liver, heart and other organs, the chylomicron glycerides are hydro- lyzed, and the constituent fatty acids reformed into other esters within the cell. Endogenous. The plasma of normal fasting sub- jects contains glyceride in concentrations of 10 to I90 mg. per 100 ml. This glyceride, which appears for the most part to be synthesized in the liver, is in the form of very-low-density lipoproteins that will be called pre-P in this review. If the concentration of such glyceride is abnormally increased the pre-P lipoproteins become larger or particulate in size. Such "endogenous particles" differ from chylomi- crons not only in their origin but also in their phys- ical properties and their content of cholesterol, phospholipid and glyceride.6.7 The turnover of glycerides in pre-fi lipoproteins seems to be. slower than that in chylomicrons. This is more probably due to their smaller average size687 than to any chemical differences, since it has been suggested that the clearing rate of particles is pro- portional to size.' At ordinary plasma concentrations the de Of removal has been calculated to be about 2 gm. per hour.9 This estimate is subject to a number of qualifications and possibly much variation, but its relatiVe*Y low order of magnitude implies that face brs that amelemte synthesis and secretion of these liPoProteins may rapidly produce endogenous Lye perlipemia. MINOR TRANSPORT TASKS Cholesterol TranspoW I4 In contrast to the large amounts of free fatty acids and glyceride that must be transported through the plasma compartment each day, the net movement of cholesterol between tissues appears to be small indeed. From 100 to 500 mg. of cholesterol per day is usually absorbed from the diet, to which is added roughly another gram of sterol resorbed from the intestinal lumen upon its secretion from bile or intestinal mucosa. Most tissues are capable of cho- lesterol synthesis, and none has been known to require that cholesterol be transported to it via the plasma to meet its demands, although this may be occurring, for example, in organs using cholesterol as a precursor for hormone synthesis. An unknown amount of cholesterol must be re- transported to the liver to take advantage of this organ's unique capability to degrade sterol to bile acids, which may then be excreted. The circulating red blood cells, for example, represent a total pool of about 4 gm. of cholesterol. With the replacement of 1 per cent of this mass per day about 40 mg. of cholesterol must be transferred from sites where red cells are broken down to where they are made, or the cholesterol must be catabolized. The total of all such anabolic and catabolic processes is a daily body turnover of cholesterol approximating 1 gm.*"s*i Cholesterol coming in from the intestine does so in association with chylomicrons, and much of this seems to disembark in the liver. From this point the exogenous cholesterol, like that which is made en- dogenously, is usually carried from tissue to tissue in the form of (Y and j3 lipoproteins. There is rapid exchange of free cholesterol between the various transport forms and many tissues, most swiftly be- tween plasma, liver and red blood cells; the re- sultant randomization of molecules makes any cal- culation of net transport from tracer studies most difficult. For present purposes, the important con- clusion to be drawn is that most of the cholesterol in plasma is not earmarked as cargo but is there for another purpose, as a structural component of lipo- proteins, vehicles for transport of other lipid, Phospholipid Transport The phospholipids, which in plasma are mostly phosphatidyl choline and sphingomyelin,i5 exceed all other lipid classes in their contribution to the total mass of lipids. If these molecules have a role in carrying specific fatty acids between tissues it has thus far been a silent one. The total plasma phos- pholipid pool is estimated to turn over about every three days in man.16 It is possible to reach erro- neous conclusions about the turnover of phospho- lipids if they are considered as a single pool, for many molecular species, differing in fatty acid con- tent, are rep-resented. One cannot escape the intui- tive conclusion, however, that the phospholipids, always predictable in their composition and varying only sluggishly in concentration, are mainly in plasma to function as "biologic detergents." Their high surface activity promotes stability at the oil- water interfaces represented by the lipoproteins and their interactions with plasma. Carotenoids and Fat-Soluble Vitamins*7-21 The diet contains several milligrams of various carotenoids per day. These pigments are of impor- tance to man because some of them are convertible in the intestine into vitamin A. This essential fat- soluble vitamin is present in the diet in only micro- gram quantities. During digestion, vitamin A alcohol (retinol) and ester, formed in the gut from B caro- tene, are transported along with unchanged carotene into the bloodstream in association with lymph chy- lomicrons. In the postabsorptive state the carotenes are found mainly in the p lipoproteins whereas vi- tamin A (retinal) is associated with still unidentified proteins of density greater than the lipoproteins (more than 1.21). Vitamin E is also transported from the intestine in the chylomicron lipids, and carried in the blood predominantly with the p lipoproteins. Little is known about the transport of vitamins D and K, but these too are probably adsorbed to the chylomicrons and carried on the fi lipoproteins in the postprandial state. To summarize the known demands for lipid transport in man and quite likely many other spe- cies, the movement of fatty acids for maintenance of "caloric homeostasis" far overshadows all other re- quirements. Food intake gives rise to several tidal waves of exogenous glyceride. As the last of these ebbs, and the delivery of dietary glucose declines as well, an increase in movement of free fatty acids out from the adipose tissues occurs. Some of these fatty acids and carbohydrates are converted to endo- genous glycerides. The ones that the liver cannot store are sent back to the adipose tissue. The inter- actions of these several transport circuits and their responses to changing demands will crop up fre- quently as this discussion proceeds, for they are the basis for several forms of hyperlipoproteinemia. In considering the manner in which lipids com- bine with proteins to serve the major transport tasks, we shall, again, only mention the important transport pathway for free fatty acids. Albumin is uniquely involved, and the protein-lipid complex formed is neither considered a lipoprotein nor measured in the usual quantification of lipids or lipoproteins. When albumin is deficient or free fatty acid concentrations are unusually high free fatty acids may travel with lipoproteins and affect the lipoprotein patterns, but these are special cases rarely encountered. The lipoproteins, on the other hand, are the keys to a more rational approach to hyperlipidemia, and the proteins involved and their various combinations with lipids will now be exam- ined in some detail. 3 THE PLASMA LIPOPROTEINS*,~.*~-~~ None of the plasma lipids are sufficiently polar to circulate free in solution. They depend upon inter- actions with protein, and the resulting "macromole- cules" or "micromicelles" are referred to by the generic term of lipoproteins. There is a tendency to restrict the term "lipoprotein" to the soluble com- plexes and to use the name "particle" for those that scatter light and approach a Z-phase distribution of lipid and plasma as they get large. For purposes of simplification the concept will be adopted here that there are 2 basic kinds of lipoproteins, the a and the /3, and that these act to solubilize varying amounts of glyceride. Followed to its logical con- clusion, this means that particles therefore contain lipoproteins. The advantages of this point of view, which has not been strictly proved to be true, will emerge presently. There is much uncertainty about the structural relations of the lipid to the proteins in these macro- molecules. It remains to be proved for any of the lipoprotein forms whether the protein serves as a film over the surface of all the lipid, exists as a cen- tral core or is sandwiched between alternate lipid moieties. The nature of the bonds between lipid and protein in the lipoproteins is also speculative. Few are covalent.*5 They are strong enough to resist dissociation during the physical processes used to isolate the lipoproteins; yet they allow for the ex- change of lipid between plasma lipoproteins them- selves and between tissue and plasma lipoproteins. There are many methods for isolating and charac- terizing lipoproteins. These take advantage of the fact that the lipoproteins behave as euglobulins but have physical properties that are determined by their content of both protein and lipids and permit sepa- ration by methods as diverse as salting out, ethanol- salt fractionation,w precipitation by antibodies and nonspecific polyanions;B electrophoresis,"-32 ultra- centrifugation33J4 and chromatography.35-m Lipoproteins isolated by other technics are in- variably equated with lipoproteins prepared by the 2 most widely used methods, ultracentrifugation and electrophoresis, which have the greatest range and adaptability. The analytical ultracentrifuge, which can quantitatively determine an almost limitless number of subgroups (S, classes) of lipoproteins varying by small increments of differences in their densities, has capabilities beyond the current re- quirements of many experimental or clinical studies.m The groups of lipoproteins important in clinical work are 4: high-density or a lipoproteins; low-density or p lipoproteins; very-low-density or pre-B- (also called az-) lipoproteins; and chylomicrons. Although everyone agrees with this subdivision as far as it goes there is not complete accord over how these groups of lipoproteins are related to each other or whether they contain subgroups that also have in- dependent metabolic behavior. This uncertainty is moving toward resolution through wider employment of immunochemical standards for comparing the content and measuring the purity of lipoproteins isolated by different pro- cedures. The lipoproteins are good antigens. Their specificity devolves from the Protein moiety, the lipids contributing in only a minor way as haptenes. The search for immunochemical determinants has exposed the presence of several different proteins and greatly enhanced the understanding of lipopro- teins in general. The Lipoprotein Apoproteins TWO different proteins are consistently isolated from plasma lipoproteins. A third has been found in I group of lipoproteins. These proteins, which some favor calling apoproteins in the lipid-free state, are generally designated as A (or (Y), B (or p) and C pro- teins. They differ in their terminal residues, total amino acid content and immunochemical behavior. The presence of 1 or both of the A and B proteins, combined with lipid, accounts for the known func- tions and chemical properties of all the plasma lipo- proteins with one exception. This is the frequent demonstration of aminoterminal serine and threo- nine in proteins obtained from very-low-density lipo- proteins, attributed by some to the C protein. Normally, the A protein is the only protein found in the cr,-migrating (high-density) lipoproteins and the B protein in the p-migrating (low-density) lipo- proteins. In some diseases the normal distribution of either protein is distorted. For example, A pro- tein appears in the low-density lipoproteins in ob structive liver disease and in the rare disease, abeta- lipoproteinemia. At times the B protein appears in the high-density lipoproteins; most commonly this implies that the manipulations involved in lipopro- tein isolation have been too severe. Both proteins can be isolated in the very-low-density lipoproteins and chylomicrons. The appearance of A and B proteins individually in the 2 soluble lipoproteins and together in the larger glyceride-rich particles is the basis for the ma- jor simplification of lipoprotein metabolism adopted for this review (Fig. 1). This concept40 is simplis- tic and vulnerable to a number of possible contra- dictions as knowledge of lipoproteins unfolds. In brief, it assumes that A and B proteins are the pri- mary components of the lipoproteins. In plasma they usually occur with predictable complements of lipid that feature differing proportions of cholesterol and phospholipids. When glycerides appear in quantity these 2 lipoproteins become involved with its trans- port. Glyceride (Fig. 1) thus becomes the third and most dynamic factor in determining the nature of the lipoprotein distribution in plasma. ALPHA LIPOPROTEIN A crude a lipoprotein was isolated from horse serum in 1929 by Macheboeuf," who used salt pre- cipitation in a way somewhat similar to the later 4 r TG FIGURE 1. Symbol for the 3 Most Important Factors in the Stmplest Concept of Faf Traruport by Lipoproteins - a = Al- phn Ltpoprotezn, /3 = Beta Lipoprotem; TC = Tn@ycende; /be Phospholipid (P), Cholesterol (C) cnz$ f'ro/41r (Stnppled /br A Apoprotezn and S&d for B Apoprotein) - All&ted Areas in fhr Sy.m- holy Comparable to Their Contribution by We& to the Iipoproterns. In all subsequent jigures the protein moiety only will be .shown in Cohn fractionation method.*" The Cohn fraction IV-I contains lipoproteins that have a, mobility on either free, starch or paper electrophoresis and correspond to the "high-density" lipoproteins isolated in the ultracentrifuge between the densities of 1.063 and 1.21. Alpha lipoproteins are normally not precipi- tated by polyanions like heparin that aggregate all the low-density lipoproteins and are thus the only lipoproteins remaining in the supematant after such treatment of plasma. They have the highest content of phospholipids and protein of all the lipoproteins. They are relatively stable and can be delipidated with solvents to yield a water-soluble A protein containing only a little phospholipid. The physical and chemical properties of the cr lipoproteins are better known than are those of the lower-density lipoproteins. The metabolism of the latter has been far better studied, however, owing in part to the interest generated by their frequent indictment for complicity in causing atherosclerosis. Inferences from what little is known of the com- parative biochemistry of lipoproteins suggests that on the evolutionary scale the a lipoproteins may be the older of the 2 major soluble lipoproteins. In many mammals (when not hibernating) the bulk of the lipid content in the postabsorptive plasma is present in a lipoproteins.34 In man and other pri- mates the total weight and lipid content of the /3 lipoproteins is greater. Nevertheless, in human plasma there is still much more A protein than B (approximately 140 mg. of the former as compared to 80 mg. of the latter per 100 ml.). Neither Cohn fractionation, electrophoresis nor polyanion precipitation alone provides (Y lipopro- teins uncontaminated by other serum proteins. Pre- parative ultracentrifugation must be employed when it is desired to isolate a large quantity of material as nearly pure as possible. In our experience, the highest yields are obtained by ultracentrifugation of plasma that has been brought to a density of 1.21 by the addition of salt. The supematant is washed free of nonlipoprotein material by ultracentrifugation repeated once or, at most, twice. All the low-density lipoproteins are then precipitated with heparin and manganese.42 Alpha lipoproteins isolated in this way ideally should contain only A protein. As deter- mined immunochemically, however, traces of B pro- tein are not uncommonly present. The A Protein22~27 Pure a lipoproteins can be extracted to obtain the A protein. Several stages of delipidation are possible. Exposure of the lipoproteins to cold ether alone extracts only some of the cholesterol and fails to remove any of the phospholipids. Pretreatment of the lipoproteins with trypsin, chymotrypsin, or phospholipases increases the amount of lipids ex- tracted by ether. Delipidation of lyophilized a lipo- proteins in the presence of starch granules43 re- moves all the neutral lipids and leaves a phos- pholipid protein that is quite stable and may rep- resent an important structural unit or stage in lipoprotein formation. The best method to date for obtaining A protein that is water soluble is pro- longed treatment of OL lipoproteins with ethanol and ether in the cold. The product still contains a small amount of tightly bound phospholipids. The essentially lipid-free A protein obtained by ethanol-ether extraction still contains over 3 per cent carbohydrate by weight.44345 The B protein iso- lated from p lipoproteins similarly contains carbo- hydrates.46 The total amount of carbohydrates present in the lipoproteins before delipidation, their exact chemical form and whatever function they are serving remain to be determined. The A protein obtained in this fashion seems to be a polymer or aggregate whose components read- ily separate and recombine in response to changes in pH, ionic strength and the presence or absence of urea or detergents like sodium dodecyl sul- fate.2794794H The protein may consist of identical sub- units whose molecular weight has been estimated to be between 23,000 and 36.000."~4~ Upon exposure of the A protein to depolymerizing agents, the polymer of S,,,4.5 yields several units of S,,,2.3. These units reaggregate rapidly in solution. It is believed that the A protein in native cr lipoprotein represents a polymer of 2 to 6 units. These subunits contain aminoterminal aspartic acid and carboxyterminal threonine. In the amino acid analyzer their composition is indistinguishable. They have large amounts of glutamic acid and leu- tine and small amounts (1 or 2 moles per mole of protein) of isoleucine, methionine and cystine.4!`n"" Upon electrophoresis in either agar or agarose the A protein migrates more slowly than the native a lipoprotein. On starch or Pevikon this difference is reversed. The A protein appears to have a high degree of helical structural organization whether lipid is present or not, suggesting that the tertiary structure of the protein is independent of its inter- actions with lipid.51 The A protein reacts with antiserums against a lipoproteins. When rarI-labeled A protein is intro- duced into plasma, it disappears at the same rate as a lipoprotein.*rss2 The A protein has great avidity for lipid. Whenever it is exposed to lipid suspensions in vitro or to other lipoproteins nearly all of it is recovered with the lipid or lipoproteins. Alpha LipoproteinP 55 There has not yet been convincing proof that A pro- tein is present in plasma except as LY lipoprotein. When plasma is ultracentrifuged at density 1.21 the lipolrrotein-poor infranate usually yields a small amount of immunoprecipitate with anti-a-lipoprotein serum. There is no certainty that this represents a lipoprotein or A protein that was present in the native state because the process of ultracentrifuga- tion delipidates some a lipoproteins and the prod- ucts of this transformation also sediment at density 1.21."" Recently, an "apoprotein" capable of recom- bining with lipid has been described in similar preparations (density greater than 1.21 infranatant fractions) from rat plasma.56 This protein has not yet been shown to be A (or B) protein. The lipid and protein sedimenting at density greater than 1.21, sometimes collectively called "very-high-density lipoproteins," still remain uncer- tainly related to the rest of the plasma lipoproteins. This fraction accounts for about 8 per cent of the plasma phospholipids and includes trace amounts of neutral lipid. Most of the phospholipids are lysolec- ithins. It has been proposed these may be the product of a plasma enzyme transferring a fatty acid molecule from lecithin to esterify cholesterol.5H The lysolecithin may be complexed with albumin,s!' but further experimental confirmation is needed. Normally most of the a lipoproteins can be iso- lated from plasma between the densities of 1.063 and 1.21. This is a rather broad density band, which can be further subdivided in the ultracentrifuge.3" The heavier a-lipoprotein fractions contain relative- ly more A protein and phospholipid and less cho- lesterol than the lighter ones. The "average" a lipo- proteins are composed (in dry weight) of 45 to 55 per cent protein, about 30 per cent phospholipid and about 18 per cent cholesterol. Five sixths of the latter is esterified. Small amounts of glyceride are also found in most preparations. The hydrated lipo- proteins contain about 15 per cent water. Estimates of molecular weight vary from 165,000 to 4OO,- 000 ?7,4!n,Rll AS detertrrined by light-scattering and viscometry the lipoprotein seems to be a prolate ellipsoid about 300 X 50 A.wl.~r Compared to the j3 lipoproteins, the a lipoproteins contain relatively more esterified cholesterol and phospholipids. The fatty acid patterns of the differ- ent lipid moieties are similar in the 2 lipoproteinsR2 The ratio (by weight) of sphingomyelin to lecithin in a lipoproteins is about 0.2."3 Immunochemistry?' The immunochemistry of the plasma lipoproteins is summarized in Figure 2. The a lipoproteins and their A protein are less antigenic than the larger and more lipid-rich /? lipoproteins. When the latter are present even in trace amounts they stimulate the production of a potent anti-&lipoprotein serum. This is a persistent problem in preparation of anti- serums to a lipoproteins. As previously indicated, antiserum prepared to either the A protein or the a lipoprotein usually crossreact. Only 1 immunologic form of a lipoprotein is usually detectable in fresh plasma. At least 1 other antigenic form becomes readily detectable after ultracentrifugation, freeze thawing or brief storage at room temperature. These 2 (Y lipoprotein antigens are only partially crnssre,ic- tive with most antiserums. We have designated the native a lipoprotein as "aLP,." This is represented by the forward migrating a-precipitation line in Figure 2. The slower migrating antigen (aLP,) is partially delipidated a lipoprotein believed to con- tain a smaller polymer of the basic subunit of A protein. Only aLP, is present in the high-density lipoproteins isolated between density 1.063 and 1.12 (called the HDL, subclass when it is quantified in the analytical ultracentrifuge)." Both aLP, and aLP, are present in the subclass separated in the ultracentrifuge between densities 1.12 and 1.21 (called HDL, in the analytical ultracentrifuge). Since aLP, is formed during ultracentrifugation it has ken suggested that the relative concentmtions of HDL, and HDL, may represent the degree to which native a lipoproteins are dissociated in the ultracentrifuge?r The plasma concentrations of these 2 "forms" of a lipoproteins vary greatly with cer- tain metabolic disorders as they are measured in the analytical ultracentrifuge." It is possible that this is a function not only of the degree of dissociability of the lipoproteins but also of other factors still unap- preciated that may control the amount of lipid asso- ciated with the A protein or its state of polymeriza- tion. Other small peaks have been seen in the analytical ultracentrifuge besides the HDL, and HDL, fractions.fis The possibility that these are technical artifacts has been raised.66 BETA LIPOPROTEINS The Lipoprotein The p lipoproteins represent a second homogene- 6 PAPER ELECTROPHORESIS 0 C"y--;Em fl St ULTRA CENTRIFUGE GM/ML GMfMl GM/ML I i I I I I I I I I I I I AGAROSE I I I I I I IMMUNO-ELECTROPHORESIS I I I I I I I I I I I NATIVE DELIPIDATED NATIVE DELIPIDATED NAT. DEL. NAT. DEL. FIGURE 2. Schematic Representation of the Major Portion o/ the Lipoprotein Spectrum as Defined b Paper Efertmphoreszr, the UltrcLcentnJGge and by lmmunoelectrophore.~i Using Anlu~rumt Reacting wtth Both a and /3 Lzpoproteinx The protein content is also depicted us in F~gwre I. ous component of the plasma lipoproteins. In the newborn infant practically all the plasma lipid not present in a lipoproteins is found in p lipopro- teins.67 This perhaps "ideal" condition also holds in most healthy, actively growing young human beings in the postabsorptive state. As the members of most populations grow older, however, there is a pro- gressive departure from this simple picture of plasma lipoproteins. Beta lipoproteins migrate with sharp boundaries in the B zone on most types of electrophoresis and can be isolated in bulk in the Cohn fraction III-O. In the ultracentrifuge p lipoproteins are isolated between the densities 1.006 and 1.063 and have a mean density of about 1.03.33 Like the o lipopro- teins, p lipoproteins are found at densities less than 1.006 when plasma glyceride concentrations'rise. In these cases they do not usually have their p mobil- ity but are part of the pre-/3 lipoproteins and chylo- microns. In the analytical ultracentrifuge, accord- ing to the technic of Gofman et al.,33164 the pure /3 lipoproteins normally are in the SJ subclass of O-20; they have a mean S, of 6 and are mostly in the sub class S, O-12. (S, refers to Svedberg units of flotation 7 expressed in lo-l3 cm. per second per dyne per gram). In man p lipoproteins are the major cholesterol- bearing lipoproteins. In the dry state a "typical" /3 lipoprotein consists (by weight) of 20 to 25 per cent protein, 8 per cent free and 35 per cent esterified cholesterol, 22 per cent phospholipid and 10 per cent triglyceride.68 In its native state it is extensive- ly hydrated. The fatty acid components of its con- stituent lipids are similar to those found in the total plasma lipids and in a lipoproteins.53 There is rela- tively more sphingomyelin in the /3 lipoproteins than in a, ratio of the sphingomyelin to lecithin being about 0.4.63 `The molecular weight of representative /3 lipopro- teins has been estimated at anywhere from I.3 to 3.2 X 106.6i~6~ Light-scattering studies indicate that the usual complex is dyssymetric and a prolate ellipsoid of about I50 x 350 A.61 Mild treatment with cold ether or n-heptane read- ily removes much of the cholesterol and glycerides from the p lipoproteins. **0*7*43 Attempts to remove the neutral lipids more completely or to strip away phospholipids usually yield a gel-like product that is irreversibly aggregated. Both the state of hydra- tion and the presence of lipid appear to be very important in maintaining the integrity of fl lipopro- teins. Recently, delipidation in the presence of de- polymerizing agents like urea and sodium dodecyl sulfate has shown promise of yielding a lipid-free and soluble protein from p lipoprotein."' The B Proteln71~76 The difficulties encountered in obtaining lipid- free B protein leave it less well characterized than the A protein. The low-density lipoproteins in egg yolk (lipovitellin) have been suggested as q model for p lipoprotein.77 When these are delipidated step- wise formation of smaller protein components, each approximately half the size of its immediate precur- sor, occurs. The total number of polypeptide chains is partially dependent on the total lipid content of the complex. Analogous but less complete experi- ments with human /3 lipoproteins suggest that the B protein in the native lipoprotein consists of several identical or at least similar peptide chains. Two identical protein units of molecular weight 380,000 have been reported in lipoproteins having an S, 7.9 flotation value. Other predictions of protein molecu- lar weight have been as low as 250,000. Recent work suggests a protein on the order of 100,000 molecular weight as a possible repeating subunit.76 The B protein contains aminoterminal glutamic acid, carboxyterminal serine and a total amino acid pat- tern that differs from the A protein particularly in the relative contents of isoleucine, leucine, glutamic acid and alanine.27,%`*74 Immunochemical studies of fi lipoproteins are subject to difficulties not encountered with Q owing to interactions with media commonly used for pre- cipitation reactions. Precipitation lines may appear that do not reflect true immunoprecipitation. Puri- fied agar and especially sulfate-free agar (agarose) have made it easier to work with p lipoproteins. In agarose they migrate in the p region and, although they still do not diffuse freely, usually give a sharp immunoprecipitation line with anti+-lipoprotein serums (Fig. 2). Most antiserums reacting with p lipoproteins also react with the soluble phospholipid-protein complex obtained by mild ether delipidation. Some anti- bodies fail to react with the products of more vigor- ous lipid extraction, suggesting that the neutral lip- ids of the lipoproteins are important as haptenes. It has not yet been shown that an anti-/?-lipoprotein serum will react with a completely lipid-free B pro- tein. This has made it impossible to learn whether B protein circulates in plasma or exists in tissues without appreciable amounts of associated lipid. Thus far it appears that al] /3 lipoproteins obtained from a given subject are antigenically homogeneous, but it is likely that as methods are improved anti- genie forms of p lipoprotein sharing only partial identity will be detected in the same individual as is the case with a lipoproteins. Several variations (polymorphism) in plasma p lipoproteins have al- ready been reported in man. These fall into 0 cate- gories. In the first ~I:IcY~. minor antigenic differences occur between the /3 lipoproteins in different sub jects; these appear to be genetically determined.*"" They are usually demonstrable by immunochemical technics using antiserums obtained from patients who have had multiple blood transfusions. Second- ly, there are cases in which p lipoproteins differ from the normal in their electrophoretic and ultra- centrifugal behavior. A peculiar "broad" p lipopro- tein found in Type III hyperlipoproteinemia will later be described in detail. There is also a recently de- scribed "double p-lipoprotein" anomaly first detected by starch-gel electrophoresis and restricted thus far to 1 family."5 Here there appears to be an altera- tion in the physical state, perhaps in degree of poly- merization, of some of the /3 lipoproteins, giving them greater density and a higher molecular weight than the normal. This intriguing abnormality may provide a means for learning more about what con- trols the normal structure of the lipoprotein. Finally, the p lipoproteins also contain enzyme activity, . especially esterases.nG It is not yet known whether this is located in the B protein or represents other proteins adsorbed to the fi lipoprotein molecules. STRUCTUREAND~NTRACELLULAR~~ETABOLISM OFTHE LIPOPROTEINS No further attempt will be made here to review the fine structure of lipoproteins. The subject must lean heavily upon analogies drawn from colloidal and surface chemistry or study of bimolecular leaf- lets such as myelin, and little progress beyond the theoretical has been made. Since the plasma lipo- proteins provide uniquely accessible models for approaching the structural features of other lipid- protein complexes in cells, knowledge in this area may be expected to increase rapidly. When certain difficulties encountered in the study of tissue lipoproteins are overcome it will also be possible to learn more about how and where the plasma lipoproteins are synthesized and broken down. It is not easy to apply to tissues the technics presently used with plasma lipoproteins. Frequent contamination of tissues with plasma, lability of the relatively weak lipid-protein bonds and frequent exchange or adherence of labeled precursors to lipo- proteins all complicate experiments dealing with cellular lipoproteins. There is fairly good evidence that the liver is capable of synthesizing lipoproteins identical to those found in plasma. Incorporation of 14 C labeled amino acids, and apparently net synthesis of fi lipo- proteins (density less than 1.063). occurs in rat-liver slices and perfused rat livers.H7"!# The ability of the rat liver to make lipoproteins that correspond in peptide pattern to plasma a lipoproteins has been 8 similarly demonstrated,!" and synthesis of lipopro- teins has been reported in rat-liver microsomes.!" Less conclusive experiments suggest that intesti- nal-mucosa cells can incorporate labeled amino acids into both LX and /3 lipoproteins.+`z-"4 This is in keeping with a recent report that the hepatecto- mized dog can still synthesize plasma lipoprotein,"5 but there yet is no basis for assessing the contribu- tion of the intestine to the plasma lipoprotein pools. Inhibitors of protein synthesis such as puromycin cause accumulation of lipid in both the intestinal and liver cells and therefore presumably inhibit either the synthesis or the release of plasma lipo- proteins.Yfi*"r In the liver the same effect has been shown for hepatotoxins such as carbon tetrachloride. Orotid acid not only produces a fatty liver but elim- inates nearly all the plasma fl lipoprotein in rats, a nearly specific effect that is quickly reversed by adenine."*a"" Turnover of plasma CY lipoproteins has been stud- ied by labeling of the protein with 1311. The biologic half-life of the protein, either in lipoprotein or as apoprotein is about four days.*~*s2~i'1" A similar half- life has been obtained with /3 lipoprotein tagged with either 1311 or 35S.`~,*"i The plasma lipoprotein proteins thus turn over faster than most other major plasma proteins with the exception of fibrinogen."" Recapitulation Those who feel no great concern with the minu- tiae concerning the plasma lipoproteins may find it easiest to consider the essential elements as being 3, all represented symbolically in Figure 1. Two of these are lipoproteins. Each consists of a different protein, the A (or CK) and the B (or p), that has un- usual affinity for lipids and prefers to circulate in extracellular fluid accompanied by a complement of cholesterol and phospholipid. The lipoproteins that result have densities and electrostatic charges that differ sufficiently to permit their operational defini- tion by several technics. By the most commonly employed methods, electrophoresis and the ultra- centrifuge, they are identified as a, or high-density, lipoproteins and p, or low-density, lipoproteins. The a and /3 lipoproteins normally account for about 90 per cent of the cholesterol and phospho- lipid in plasma. The concentrations of those lipids undergo very little tidal change in comparison to that of the other major group of neutral lipids found in plasma, the glycerides. Glycerides are the third element in any discussion of lipoproteins and fat transport (Fig, 1). They are the dynamic factor to which the a and /3 lipoproteins seem related as vehicles are to cargo. When appreciable glyceride is present in plasma it associates with both (Y and /3 lipoproteins in such a way as to decrease their den- sity and alter the net effect of the charges on the lipoproteins. The resultant combinations differ, de- pending on whether the glycerides are coming in from the diet or are made in the body. This distinc- tion is of great importance when one is in dealing with hyperlipoproteinemia and merits detailed con- sideration, which will be given in the next section of this review. GLYCERIDE TRANSPORT Two Forms of Plasma Glyceride Transport In subjects on fat-free diets the plasma lipid con- centrations vary little throughout the day and tend to be highest just before breakfast"`2 When the diet contains fat the levels of cholesterol and phospholip- ids in plasma are still relatively constant, but the glyceride concentration, as already intimated, is quite variable. It is lowest during the early morning hours (when blood samples for analysis are usually taken) and rises sharply, reaching a peak about three hours after breakfast. The concentration is boosted again by the midday meal, and after some decline in the afternoon, again rises after dinner, ebbing during the night, The form in which glycerides are transported in the plasma depends on whether their immediate source is dietary fat or endogenous synthesis (Fig. 3). In normal subjects before breakfast, and in those on fat-free diets at all times, the plasma glycerides are exclusively of endogenous origin and are carried predominantly on pre-fi lipoproteins and particles. Glycerides from dietary fat, on the other hand, are carried on the chylomicrons. METHODS Analytical methods for the detection and quanti- tation of (Y and fi lipoproteins are relatively simple, and the technics employed are generally compara- ble from one laboratory to another. In the analysis of the glyceride-rich lipoproteins, however, there is no such uniformity. Many different procedures are used, none of them ideal. The ones that are of greatest value at present have one common feature, the ability to provide a qualitative separation of lipoproteins and particles carrying mainly endoge- nous glycerides from those containing exogenous glycerides. These technics, collated in Table 1, will be considered one by one. Ultracentrifugation The preparative ultracentrifuge separates lipopro- teins according to their flotation characteristics, which depend upon size and density. It can be used to separate the main chylomicron (exogenous) glyceride mass from the endogenous particles. Such separations are not complete, however, for the flota- tion rates of chylomicrons and endogenous particles overlap.6*`~`"3 Thus, the "chylomicron" fractions pre- pared by exposure of plasma to centrifugal fields of about 105 g. per minute may contain a variable mix- ture of endogenous as well as exogenous particles. After this fraction has been removed a species of very-low-density lipoproteins can next be isolated that is generally considered endogenous in origin. It is of S, 20-400 and identical to the pre-/3 lipopro- teins separated by paper electrophoresis.ia'4 By ma- nipulation of the diet of subjects serving as lipopro- tein sources - for example, fat-free diets, high in carbohydrate, which induce high concentrations of endogenous glycerides, or high-fat diets, which pro- duce transient chylomicronemia - the preparative ultracentrifuge may be used to obtain quantities of pure or nearly pure particles of the two types.`,`" Precipitation Precipitation methods are among the oldest of the technics for preparing lipoproteins and are of great usefulness in the isolation of pure lipoproteins."' No reliable way has yet been found to use precipitation alone for precise subfractionation of very-low-den- sity lipoproteins and particles. Dextrdn sulfate, amy- lopectin sulfate, heparin or polyvinylpyrrolidone precipitate endogenous and exogenous particles together."' The combination of gradient flotation and precipitation by polyvinylpyrrolidone is used to sclia- rate chylomicrons into 2 classes, "primary" and "secondary" particles, and these exogenous particles can also be separated from endogenous particles." "`5 This technic can be employed to distinguish endog- enous and exogenous hypcrlipcmia and is particu- larly useful for preparative work since the floccu- lated particles can be recovered for lipid and other analyses. Electrophoresis It was noted quite early in the study of plasma by free electrophoresis that the turbidity in lipemic % l"l<`/ "rrl&l IK lOlr8/ 7 lVl,d % wd Izpd /,/ml r,pld >4on I) 5-2.5 7%95 2.12 3-15 >401t > 4oll 2. IO till-80 IO-20 10.20 20.400 IO-13 50-70 IO-25 15.25 samples obscured underlying protein peaks.`*`" Exog- enous particles in lymph were shown to migrate with albumin whereas those in plasma moved with p globulin. When the simpler technic of electropho- resis in an inert medium of starch granules was devised, similar turbid peaks were found and the separateness of an a, lipoprotein (pre-/3 lipoproteins on paper electrophoresis) estab1ished.i'" More re- cently it has been demonstrated that exogenous particles in plasma may accumulate in 2 regions on starch blocks.G~i"H Exogenous particles in the a, zone are called "primary" because they appear in plasma early after fat ingestion. The other group of exoge- nous particles appear somewhat later in the p zone and are called "secondary." The primary particles tend to be superimposed upon endogenous lipopro- teins and particles when the latter are present, Similar separations can be achieved with the sim- pler technic of paper electrophoresis. It was early noted that with barbital buffer, chylomicrons remain at the origin, as do the larger endogenous particles, whereas the smaller particles and SJ 20-400 lipopro- teins form a "pre-P" band.i')~~rl'!`ri*II When albumin is added to buffer to decrease adsorption to the paper sharper separations are obtained.iir*lr2 The chylomi- crons (both "primary" and "secondary" exogenous particles) remain at the point of application; the endogenous glyceride-rich lipoproteins are concen- trated in the pre-/3 position, with trailing of the larger endogenous particles between the origin and pre-P region.7s1i2 Nature of Pre-P Lipoproteins'* The pre-P lipoproteins and endogenous particles cover a wide density spectrum varying from 1.006 to 0.93. Their composition depends on their density. As the density decreases, the relative proportion of glyceride rises, and that of protein falls. The fatty acid pattern of a given subclass is not the same in all subjects; grosser differences may obtain between normal and abnormal subjects. Nevertheless, certain generalizations can be made about the composition of the usual pre-P group of lipoproteins (Table 1). Lipids make up over 85 per cent of the total weight of the complex, protein being 2 to 15 per cent, ~Iyc- eride is the major lipid class. When prep lipopro- teins and particles are isolated under conditions that minimize lipid exchange with other lipoproteins, the glyceride fatty acid composition is that of endog- enously synthesized fiit, low in linoleic acid and high in palmitic and oleic acids.",' 10 Proteins in Pre-P Complexes7H**13 The protein moiety of pre-fi lipoproteins was for a long time considered to be mainly or exclusively the B protein, identical to that in the p lipopro- teins.1'4 Native pre-P lipoproteins react with anti- p-lipoprotein serums and not with anti-a serums. Certain chemical findings, however, were never explained by the assumption of a single protein. The increased electrophoretic mobility of pre-fi lipoproteins over that of p lipoproteins, the higher ratio of phospholipid to cholesterol and a significanl content of aminoterminal aspartic acid (the same as in the A protein) in the protein residues, spurred a search for other companions of the B protein in pre+ lipoproteins. Recently, it has been shown that after careful delipidation of pre-P lipoproteins sig- nificant amounts of both (Y and p lipoproteins are immunologically identifiable (Fig. 2). Hydrolysis of pre-P lipoproteins catalyzed by postheparin enzymes also immediately increases the plasma content of a and fi lipoproteins. In subjects fed various experi- mental diets the concentrations of pre-P lipoproteins vary inversely with those of LY and p lipoproteins. The evidence is good that both the A and B pro- teins, probably as their lipoproteins, are normally present in the triglyceride-rich complexes grouped as pre-P (very-low-density) lipoproteins. C protein."" Analyses of the proteins obtained in pre-P lipoproteins and particles usually reveal some aminoterminal residues (threonine and serine) dif- fering horn those in the A or B proteins. The amounts have been variable, and the source conjec- tural.' In recent years one series of studies have dealt with the isolation of a third protein, called apoprotein C. This apoprotein has been isolated as a phospholipid-protein complex along with the A and B apoproteins after lyophilization and partial delipidation of very-low-density lipoproteins with heptane. Whether the particle sources of this pro- tein contain glycerides of endogenous or exogenous origin has not been established. The C apoprotein is characterized as a 7S protein with hydrated den- sity of 1.09 (pm. per milliliter) and a molecular weight of about 834,000. Although it appears homo- geneous in the analytical ultracentrifuge it migrates as a double band with a prealbumin position on starch-gel electrophoresis and contains both amino- terminal serine and threonine. The C apoprotein fails to react with antiserums to (Y and p lipoprotein. Peptide fingerprints obtained by trypsin or pepsin digestion have been considered different from those of the A and B apoproteins. The C apoprotein ap- parently binds phospholipid more avidly than A or 8. The bid of C protein for entry into the family of fat-transporting proteins is greeted with great inter- est. Although its presence in glyceride-rich particles is as indubitable as its separateness from apoprote- ins A and B, it remains for further work to demon- strate its specificity and its physiologic role, There is not yet sufficient knowledge about it to fit C pro- tein into any of the concepts being developed in this review, and it does not appear in any of the accompanying illustrations. Metabolism of Endogenous GlycerideL-3*4*115 II7 Considerable evidence points to the liver as the major'site of synthesis of the endogenous triglycer- ide and presumably elaboration of many if not all of the circulating pre-/3 lipoproteins. An important source of fatty acid substrate for glyceride synthesis is the plasma free fatty acids. The flux of free fatty acids into liver, heart, skeletal muscle and other tissues is governed by their rate of release from adipose tissue. Any factor that increases lipolysis or decreases glycerol esterification in the adipose tis- sue causes outpouring of free fatty acids."" Much of this free fatty acid is removed by the liver, and the excess beyond what it can use or store is resynthe- sized into glycerides and resecreted as pre-P lipo- proteins since the acids apparently are not returned to adipose tissue as such. A significant "overshoot" in release of free fatty acids, therefore, may tempo- rarily elevate the plasma glyceride concentration. Potential causes of hyper-pre-fi lipoproteinemia include a number of abnormal physiologic states that are accompanied by increased release of free fatty acids. Carbohydrate induction."s-124 Endogenous hyper- lipemia perhaps first became a distinctly recognized phenomenon in 1950, when Watkin and his co-work- ers,,!' noted that fat-free, high-carbohydrate diets increased the glyceride concentration in a group of hypertensive men. Many others have confirmed and extended these observations. The concept of carbo- hydrate-induced hyperlipemia is an important out- growth of such studies. Not all carbohydrate induc- tion is abnormal, however, for it is now clear that the response of practically all subjects to very high-carbohydrate feeding is conversion of much of the carbohydrate to fat and its eventual release into the plasma as pre-fi lipoproteins. In healthy young subjects the plasma glyceride rises to a peak three to ten days after the beginning of carbohydrate feeding and usually falls slowly thereafter despite continuation of the diet. The response is widely variable; the average rise in glyceride concentration is about 200 mg. per 100 ml. above the initial level, but in a "normal" subject may be as much as 400 mg. per 100 rn1.1z4 The ease of induction of endoge- nous hyperlipemia in normal subjects is probably due to the relatively limited rate of removal of such lipid from the plasma. The previously mentioned estimates of a turnover rate of glycerides in pre-P lipoproteins of about 2 gm. per hour in the adult have been obtained in subjects with normal and somewhat elevated glyceride concentrations.!' The capacity of removal mechanisms to adapt to higher loads has not been established. The question of whether endogenous and exogenous glycerides are removed by an identical mechanism also has not been resolved and is important to the understanding 11 of certain kinds of hyperlipemia.1'7 In some experi- ments tagged glycerides in fatty acids administered in a form comparable to the pre-P lipoproteins have been found predominantly in the liver in the fasting state and in the adipose tissue in the fed state."" In others, the role of the liver in removing endogenous glycerides has seemed unimportant.iZS Particles of Exogenous Glyceride (Chylomicrons)1-5 lwX~lli Dejinition. The term chylomicron was coined in 1920 for the visible particles that appear in lymph and blood in response to fat feeding and contain the fed triglycerides.12" For a time all particulate lipid was given this name, but it has become useful to return to the original meaning to provide some easy way of distinguishing dietary particles." It is easier to define chylomicrons in physiologic than in operational terms (Table 1). They float too rapidly in the ultracentrifuge to be quantified by optical means, and arbitrary centrifugal forces have lieen adopted for their definition and isolation. Sometimes, all particles having an S, greater than 400 hVC been considered chylomicrons. Endogenorls particles may also have flotation rates of this order, however, and the ultracentrifuge sometimes cannot be used to separate chylomicrons from significant quantities of endogenous particles. In free or starch-block electrophoresis chylomi- crons have a wide range of mobility. Those isolated from thoracic-duct lymph migrate with albumin.5 As already noted, in plasma they may move with 01~ globulin (primary particles) or with p globulin (sec- ondary particles).i"H Endogenous particles likewise have a2 mobility and cannot always be separated clearly from chylomicrons by this technic. One sep- aration of the 3 types of particles by polyvinylpyrrol- idone gradient tubes has already been described.6 `,For clinical purposes the presence of chylomi- crons in plasma is most conveniently indicated by paper electrophoresis. In albumin-containing buffer chylomicrons remain as a distinct band at the origin on electrophoresis whereas all other lipoprotein species migrate to some extent (Fig. 3).7 Composition (Table 1). In the light or electron microscope the chylomicrons appear to be spheres that vary in diameter from about 0.1 to 5.0 p. The upper limit may reflect aggregation during isolation, and it is possible that circulating chylomicrons are not larger than about 1 /.L. Chylomicrons, like pre-fi lipoproteins, are made up mainly of glyceride, with lesser amounts of phospholipid and cholesterol. The proportions depend on the average size of the par- ticles under study. Usually less than half the cho- lesterol is esterified, as compared to 65 to 75 per cent of esters in other lipoproteins. Patients with severe chylomicronemia may have a low percentage of esterified cholesterol in the plasma without the usual implication of abnormal liver function. The protein content of chylomicrons is usually between 0.5 and 2.5 per cent of their total weight. Reports of higher protein content are probably due to contamination of the isolates. There is no general agreement about the nature of this protein or whether it is an intrinsic part of the chylomi- cron.S*2is1g71Z' It therefore does not appear in Figure 2. Even in chylomicrons washed many times by ultracentrifugation several different serum proteins, including albumin and gamma globulin, can often be identified immunologically. Amino acid or pep- tide analyses on such small quantities of protein are subject to errors compounded by the probability that a mixture of proteins is usually present. Amino- terminal aspartic and glutamic acid are often ob tained, and the aminoterminal serine and threonine typical of apoprotein C are inconsistently present, Considering all the available evidence, the 2 pro- teins obtained most consistently after delipidation are the A and B proteins. These can be seen best by immunologic means, but peptide analyses have also demonstrated the A protein on one occasion.iw It is not known how the chylomicrons are actually stabilized, and some or all of the proteins present could be adsorbed to the surface or dissolved in these complexes without serving a functional pur- pose. It seems clear that upon entering plasma,, chylomicrons take on additional protein, for the protein content is higher than that of similar parti- cles isolated from lymph and, as noted earlier, the electrophoretic mobility also changesi"" We shall have occasion shortly to examine evidence that the 2 major plasma lipoproteins have more than a casual relation to the chylomicron. Origin of chylonaicrons."-`3' In the intestinal lu- men dietary glycerides are dispersed by the action of bile salts and rendered easily vulnerable to hy- drolysis by lipases. The resulting products are mi- celles containing glycerol, free fatty acids, monoglyc- erides and diglycerides that are taken up by in- testinal epithelial cells and there resynthesized to triglycerides. Cholesterol, phospholipid and proba- bly protein are added, and the resulting chylomi- crons are extruded from the mucosal cell into the lymphatic spaces. The precise details of chylomi- cron formation and movement from the base of the jejunal epithelial cell into the lymphatics are not known. Chylomicrons travel in the intestinal lymph through the regional lymph nodes and the thoracic duct into the venous blood. ChyEotnicron remor;al.iir Chylomicrons are rapidly removed from the plasma. When chylomicrons con- taining tagged glycerides are infused into human subjects they disappear with a half-time of five to fifteen minutes.iRi From the distribution of tagged glyceride fatty acids it appears that most organs of the body rapidly receive chylomicron fatty acids. The possibility that some of the glycerides are hy- drolyzed very rapidly and the resulting fatty acids distributed as free fatty acids to some tissues has to be considered in the interpretation of experimental studies of glyceride removal. it appears from such studies in several species that after an overnight fast 12 the major site of chylomicron removal may be the liver. In animals allowed access to food most of the chylomicrons may be cleared by adipose tissue. The concept of net hepatic removal of chylomicron glyc- erides has recently been challenged by experi- ments in which rat livers were perfused with chy- lomicrons in the presence and absence of heparin.i3* Without heparin, which may release lipolytic en- zymes into the medium, there was no net uptake or metabolism of the particles. When heparin was added to the perfusate, uptake and oxidation began immediately. Other perfusion experimentsi3" have not been in agreement, however, and the question of direct participation of the liver in glyceride re- moval remains unsettled. The role of lipolysis. 11r,~r~ It is fairly certain that a hydrolytic step must initiate or accompany the re- moval of chylomicrons from the circulation, The evidence is inferential and of the following nature. Although the endothelium of the hepatic sinusoids may contain spaces that could admit chylomicrons, the capillary walls in other tissues appear continu- ous. Passage of chylomicrons through the endothe- lial wall by pinocytosis has not been convincingly demonstrated by the electron microscope. Further- more, during removal of glycerides, the fatty acids appear quickly in the free fatty acids, perhaps 10 per cent of which represent fatty acids coming in with the glycerides.13" Tagged fatty acids appearing in liver do not remain in the glycerides in which they were delivered but are rapidly reshuffled, many appearing in phospholipids. The initial hydrolysis of the glycerides entering extrahepatic tissues is be- lieved by some to be catalyzed by lipoprotein lipase concentrated in the capillary wall.117 Lipoprotein lipase catalyzes the hydrolysis to fatty acid and glycerol of glycerides in chylomicrons and other lipoproteins, and will also split artificial fat emul- sions if they are first activated by added serum or plasma. Parenteral administration of heparin and other polyanions causes lipoprotein lipase activity to appear rapidly in plasma. The activity rapidly dis- appears if liver function is normal. There is evi- dence that postheparin lipolytic activity is due to more than 1 enzyme, including a phospho1ipase.r"" The intravascular lipolysis induced by heparin or similar agents is a gross exaggeration of normal chy- lomicron removal to the extent that the fatty acids cannot all be removed by the immediately adjacent tissues. The free fatty acid content does rise slightly during normal chylomicron clearing. How much lipoprotein lipase is present in the adipose tissue cells themselves as opposed to their capillary epi- thelium is unclear.*37 It has been detected in he- patic-vein bloodr3"*`3!' but never convincingly dem- onstrated in liver parenchymal cells. The level of lipoprotein lipase in adipose tissue bears a relation to the insulin activity. In rats made diabetic with alloxan the enzyme activity is low and restored to normal by insulin treatment.14" Recupitulation. Plasma glycerides are transported by 2 different systems. Dietary fat is carried from the intestine into the bloodstream as large particles (chylomicrons), consisting mainly of glycerides. Phospholipids, cholesterol and a little protein are also present and probably stabilize the particle. The presence of chylomicrons can be established, and the quantity estimated by several technics. One of the simplest and most useful is based on the lack of mobility of chylomicrons on paper electrophoresis. The plasma concentration of chylomicrons is varia- ble and dependent upon the timing of fat ingestion. Only quantities below the limits of detection by paper electrophoresis are normally present after an overnight fast. The disposal of these particles is relatively rapid and depends upon 2 major steps, The first is hydrolysis, which possibly takes place at the capillary wall and is catalyzed by lipoprotein lipase. The second is the ability of tissues beyond the capillary wall to take up the fatty acids released. In the adipose tissue, one of the major sites of dis- posal, re-esterification depends up011 .ui atlc,cl\i.ite supply of cr glycerolphosphate, which in turil is derived from glucose breakdown. The fate of the other constituents of the chylomicron, which seem to include small amounts of a and /3 lipoproteins, is not known. It is possible that chylomicrons after removal of some of their glycerides become pre-@ lipoproteins. Alpha and /!I lipoproteins, freed of their mantle of glycerides, may join others of their spe- cies in the plasma. Glycerides synthesized in the liver either from carbohydrate or from re-esterification of circulating free fatty acids are transported from that organ in pre-/i lipoproteins and larger endogenous particles. These represent a broad spectrum of density and size from particles large enough to be seen in the light microscope to others merging with the density limits of the soluble p lipoproteins. The endogenous particles have higher protein and cholesterol con- tent than the chylomicrons and different properties. On paper electrophoresis, they migrate in the pre-/3 range with increasing trailing as particle size in- creases. Their concentration in plasma changes slowly but can vary widely over the period of a few days, particularly in association with marked changes in the carbohydrate content of the diet. The removal rate of glyceride in the pre-b lipoproteins seems to be slower than that of glyceride in chylo- microns. The large endogenous particles may be cleared at about the same rate as chylomicrons, however, and it has not been shown that glyceride molecules of different origin have different disposal mechanisms. INHERITED LIPOPROTEIN DEFICIENCY STATES Several mutations in man have been discovered that have laid to rest all uncertainty that the A and B proteins in lipoproteins have separately evolved to serve different functions. They have also consid- erably illuminated some of the transport tasks that 13 each lipoprotein seems to serve and illustrated something about the genetic control of their plasma concentrations. In 2 genetically determined disorders the normal /3 or OL lipoproteins respectively are completely missing from the plasma. These are abeta-lipopro- teinemia and familial u-lipoprotein deficiency (Tan- gier disease). The latter is termed "deficiency" be- cause small amounts of a lipoproteins are present, but they appear to be abnormal judging from avail- able immunochemical evidence. The clinical details of these 2 diseases have been extensively reviewed, and attention can be focused here on the disabilities in fat transport that accompany absence of one or the other class of lipoproteins. In 1950, 2 patients with neurologic abnormalities and "crenated" red blood cells (acanthocytes)i44 became the first of about 30 patients with a singular disorder subsequently reported from America, Britain and Europe. At first the acanthocytosis seemed the most striking finding; a very low con- centration of plasma cholesterol was then noted and absence of /!I lipoproteins was demonstrated in 1960.145 Other investigators confirmed that several patients had severe deficiency or absence of all the plasma lipoproteins of density less than 1.063, in- cluding the absence of any chylomicronemia after the feeding of fat. The name abeta-lipoproteinemia was then suggested'45 and has gained general ac- ceptance on the assumption that this describes the primary inherited defect. The disease is usually expressed in infancy by retarded growth associated with steatorrhea and abdominal distention. Malabsorption becomes less marked in late childhood and is succeeded by pro- gressively more severe neurologic deficits, including loss of muscle strength and nystagmus and signs of degeneration of the posterolateral columns and cer- ebellar tracts. Pigmentary retinal degeneration and visual difficulties appear. Life expectancy is limited; at least 1 death has been associated with persistent cardiac arrythmias. Deposition of ceroid pigment in the tissues is a prominent finding. The plasma and tissue phospholipids are unusually low in essential fatty acids. A deficiency of the latter or some other dietary constituent whose intake is dependent upon fat absorption may be the basis of the changes in erythrocyte membranes and nerve-cell structures that seem to underlie the many features of the dis- ease. Sibs can be affected, and sometimes there is parental consanguinity, but vertical transmission has never been observed. The disease appears to be the expression of a double dose of a mutant autosomal allele. In 1 set of parents the fi lipoproteins have been low,rts but at. present there is no way to dis- tinguish most of the presumed heterozygotes. The plasma lipoprotein pattern in abeta-lipopro- teinemia is illustrated in Figure 4. The evidence indicates that there is no p lipoprotein in plasma. DENSITY Sf 1 PAPER STRIP 1 PROTEIN a PROTEIN Immunochemical methods capable of detecting less than l/10,000 of the /3 lipoprotein in normal plasma have revealed none in 6 patients from 4 kindreds."' The average plasma concentrations of cholesterol (20 to 90 mg. per 100 ml.) and of phospholipids (35 to 95 mg. per 100 ml.) are the lowest seen in any human disease. Glyceride concentrations approach the vanishing point - less than 10 to 30 mg. per 100 ml. AS shown in Figure 4 abeta-lipoproteinemia pro- vides one of the rare instances in which a lipopro- teins appear in the density region of I.019 to 1.063 that is usually the exclusive preserve of the p lipo- proteins. These cy lipoproteins have a mean density that is lower than usual, probably as the result of a large lipid complement carried by the A protein. The A protein seems otherwise to be perfectly nor- mal. The (II lipoproteins are immunochemically identical to those in normal subjects, and the A protein has an amino acid composition indistin- guishable from the normal.;!1 The omission of all P-lipoprotein-containing com- plexes (compare Figure 4 with Figure 5) is asso- ciated with the abolition of practically all glyceride transport in abeta-lipoproteinemin. The most ob vious defect is a total inability to form chylomi- crons. Patients with abeta-lipoproteinemia digest glycer- ides, absorb the resulting monoglycerides and free kitty acids into the intestinal mucosa cells and re- form them into triglycerides, but fail to release chy- lomicrons. Of great interest is the ability to produce in rats a similar defect in the intestine by abolishing protein synthesis with puromycin.":r,!" As patients with abeta-lipoproteinemia get older, they "learn" to absorb some fat, but must do so through another pathway such as transport of long-chain fatty acids via the portal circulation; they never have chylomi- crons in peripheral blood. When the patients are fed diets high in carbohy- drates designed to induce production of endogenous 14 DENSITY Sf "El* z.10 LOW PAPER STRIP 1006-10 1.019-10 1.063-o 1.2, - PROTEIN a PROTEIN glycerides and release of pre-fl lipoprotein, no in- crease in plasma glycerides occurs and no trace of pre-/3 lipoproteins appears.r+' Diets that contain more than 20 gm. of carbohydrate per kilogram of body weight have been used. A liver biopsy has been obtained on 1 patient who was regularly on a rela- tively high-carbohydrate diet and revealed tissue con- taining excessive amounts of glyceride.14' Thus, it appears that, like the intestinal mucosa, the liver in a patient with abeta-lipoproteinemia is able to pro- duce and store glycerides but cannot transfer the lipid into the plasma. The analogue of abeta-lipoproteinemia was dis- covered in 1960 and called Tangier disease after the first 2 examples, in siblings five and six-years-old, from Tangier Island in Chesapeake Bay. Since that time the authors have had the opportunity to study 6 more examples, in 3 pairs of siblings from differ- ent families unrelated to the Tangier population. The abnormalities found in the intestine and liver of patients with this disease, along with the nearly complete absence of glyceride in plasma, strongly invite the conclusion that the B protein has a specific function in the transport of both exogenous and endogenous glyceride from cells. This function cannot be adequately assumed by A protein. Hypobeta-lipoproteinemia A few patients with familial deficiency of /3 lipo- proteins (as opposed to their absence) have been reported.r4"-14# Beta lipoprotein concentrations have been 10 to 50 per cent of normal, with comparable decreases in plasma cholesterol, phospholipids, gly- cerides and probably in essential fatty acids. Some patients have had abnormally fragile red cells or acanthocytes and progressive neuromuscular diffi- culties developing in adulthood. In at least 1 of the reported families, the data indicate that the de- fect could be transmitted as an autosomal dominant; the mutations are probably different from that pro- ducing abeta-lipoproteinemia. In Tangier disease the plasma concentrations of cholesterol average 70 mg. per 100 ml. (with a range of 50 to 130 mg.), and that of phospholipids about 100 mg. (range of 70 to 140 mg.), both similar to those found in abeta-lipoproteinemia. The glyc- eride concentrations tend to be modestly elevated in the postabsorptive state (range of 120 to 280 mg.). The dramatic abnormality that uniquely character- izes the disease is the great size and peculiar or- ange color of the tonsils. Even small tags remaining after tonsillectomy have the telltale appearance. These changes are due to a gross deposition of cho- lesterol esters in reticuloendothelial tissues, some- times also associated with enlargement of liver, spleen and lymph nodes. All the known clinical manifestations of the disorder appear to be second- ary to the deposition of lipid, which is widespread and can also be found in cutaneous lesions and the cornea, blood vessels and reticuloendothelial cells in the rectal mucosa. One adult patient had pancy- topenia corrected by removal of an enlarged spleen. His affected sibling died at the age of forty-eight years with a probable myocardial infarction. Fat absorption and chylomicron formation are not abolished in hypobeta-lipoproteinemia. Presumably, sufficient p lipoprotein is formed to meet these re- quirements, but it is likely that these patients are operating at minimal levels of B protein and its lipoprotein. For example, in 1 patient with severe hypobeta-lipoproteinemia and acanthocytosis neuro- logic symptoms developed only after the fourth Alpha-lipoprotein deficiency seems certain to be the primary inheritable defect, and extensive study of the pedigrees has revealed the expression in a clear genetic mode. The presence of a single ab- normal autosomal gene results in abnormally low plasma concentrations of (Y lipoprotein (in terms of lipoprotein cholesterol, below 32 mg. per 100 ml. in males and 35 mg. per 100 ml. in females) but no significant tissue lipid deposition.r+' The homozy- gous abnormal genotype is expressed by near ab- sence of all plasma high-density or (I! lipoproteins and tissue accumulation of cholesterol esters. The plasma'of the homozygote usually appears to be completely devoid of (Y lipoproteins after paper electrophoresis or ordinary immunophoresis. With appropriate antiserums and concentration of the plasma, however, a small amount of (Y lipoprotein 15 pregnancy. Presumably, the hyperglyceridemia of pregnancy creates greater demand for glyceride transport and therefore for B protein. Familial hypobeta-lipoproteinemia must be dif- ferentiated from that secondary to acute infections, other severe debilitating illnesses or malabsorption due to different gastrointestinal lesions. Undoubted- ly as population surveys increase more patients will be found with similar decreases in P-lipoprotein production or increased catabolism. A spectrum of inherited states with different degrees of P-lipopro- tein deficiency may eventually be uncovered. Familial Alpha-Lipoprotein Deficiency (Tangier Disease)": can he detected. This lipoprotein can be isolated between the densities of D 1.063 and 1.21 in the ultracentrifuge and concentrated in il manner similar to that of normal a lipoprotein, It contains choles- terol and pllospholipid ,lntl has an a, mobility like that of' a lipoprotein. It also reacta with anti-a- lipoprotein sorum, .rntl when used ;LS an antigen. pro- vokes antibodies to itself as well as normal Iy lipo- protein\. Despite these similarities to native a lipoprotein, however, flIrther immunochemicnl stud- ies reveal that the Tangier a-lipoprotein, in both its lipidrich and its delipidated form, is not antigeni- tally identical to normal a, lipoprotein.`;" Thus, it has been gjverl, the separate designation, Tangier a lipoprotein; it is present in the plasma of patients with Tangier disease in about one twelfth of the concentration of a lipoprotein in normal subjects (about 30 mg. vs. 360 mg. per 100 ml.) and is the only form of 01 lipoprotein that can be identified by immunochemical technics. The heterozygous rela- ti\es of' the patients have both Tangier a lipopl-stein a, and a Ijpoprotein. The P-lipoproteins in Tangier disease appear to be normal on immunophoresis. Their phospholipid content is high, however, and some float at the ab- normally low density of 1.006. The lipoprotein pattern in postabsorptive plasma from patients with Tangier disease (Fig. 6) is unique. The paper electrophoretogram alone imme- diately permits a presumptive diagnosis that needs only to be confirmed by immunochemical analyses. In contrast to abeta-lipoproteinemia the patients with Tangier disease absorb fat normally from the intestine and make chylomicrons that behave phys- ically like the normal ones. Their ability to release endogenous glyceride into the circulation in re- sponse to a dietary carbohydrate load is also unim- DENSITY Sf - PAPER STRIP ,fj PROTEIN & PROTEIN paired.;" Thih endogenous hyperlipemia is as\oci- atcd, however, with peculiar triglyceride-laden lipoproteins of density less than 1.006 that appear as a broadened p band on electrophoresis (Fig. 6). This phenomenon provides a convincing demon- stration that the normal pre-P mobility of the very- low-density lipoproteins depends upon the presence of a lipoproteins in the complex.`" The ability of patients with severe a-lipoprotein deficiency to move both endogenous and exogenous glyceride into the plasma implies that a lipoproteins are not essential in this phase of glyceride transport. That they may have another role in facilitating the passage of glycerides and other lipids from plasma to tissues is suggested by 2 consistent observations in Tangier disease. Plasma glyceride levels in the fasting state usually are abnormally high on regular diets. Whenever either endogenous or exogenous hyperglyceridemia is induced by appropriate diets the maximum glyceride concentrations and the clu- ration of the hyperlipemia are greater than normal. HOW a lipoproteins assist in the exodus of glyceride from plasma is still open to speculation. As best as can be determined from immunochem- ical studies the reticuloendothelial tissues in pa- tients with Tangier disease do not seem to contain a lipoproteins in unusual quantities, and there is no evidence that the cholesterol-ester infiltration arises because cells have engorged a lipoproteins. A sin- gle study has revealed no increase in `cholesterol synthesis in the Tangier tonsil. It seems more likely that a lipoprotein enhances the stability of the cir- cuIating particles containing glyceride and other lipids. When the lipoprotein is missing, the unstable particles may be more susceptible to removal by phagocytic cells. The cholesterol esters may repre- sent the steadily increasing residue of such an up- take. Recapitulation Information derived from the observation of pa- tients with lipoprotein deficiency combined with other information permits several hypotheses con- cerning the participation of 2 particular proteins in fat transport. These are briefly summarized as fol- lows. The liver and intestine, and perhaps other tissues, have the capacity to synthesize proteins with a special affinity for lipids. In extracellular fluid these proteins appear in association with char- acteristic and different amounts of phospholipids and sterols to form 2 independent kinds of lipopro- teins. These a and p lipoproteins not only provide a means for keeping their constituent lipids in solu- tion but also assist in the stabilization of more lipids when they must be transported in plasma. The glyc- erides, one of the major forms in which fats des- tined for ultimate caloric use are carried, seem to depend upon lipoproteins for their movement out of cells and delivery to sites of removal. The B pro- tein, backbone of p lipoprotein, appears to have a special function in making it possible for intestinal 16 and liver cells to deliver glyceride into the extra- cellular fluids. How and at which cellular sites or surface5 it acts in this capacity is not known. In man, at least, the A protein, which forms the (Y lipo- proteins, cannot perform this function. Its absence is reflected in somewhat poorer removal of glyceride from plasma and tremendous and progressive cho- lesterol accumulation in tissues. Many control mechanisms must operate to main- tain plasma lipoprotein concentrations. Ultimately, these affect the rates of synthesis and catabolism of the lipoprotein proteins, but they probably are more directly responsive to changing requirements for lipid transport. The constantly shifting demand for glyceride transport requires a quick response. Other requirements, such as the need for stepped up movement of cholesterol when this lipid is ingested or synthesized in increased amounts, perliaps allow more time for adaptation. The states of metabolism of carbohydrate and protein also influence lipopro- tein concentrations, and the total number of genetic and environmental factors at play is obviouslv large. If even the least subtle of all the possible controls on lipoprotein metabolism were completely clarified it would be much easier to discuss the clinical problem of hyperlipoproteinemia. As it is we must take up that topic in the next section, employing terms more descriptive and less mechanistic than one would desire. HYPERLIPOPROTEINEMIA Definitions Up to this point we have concentrated on laying the support for 2 generalizations. The first is that, with the exception of free fatty acid concentrations, which have no lipoprotein equivalents, all abnor- malities in plasma lipid concentrations or dysli- Iridemia can be translated into dyslil)ol)roteinemiu. The second is that the shift of emphasis to lipopro- teins offers distinct advantages in the recognition and management of such disorders. We have already discussed the relatively few cases in which hype- lipoproteinemia is a clinical problem, and the re- mainder of the review will be devoted to hype&o- proteinemia. Hyperlipoproteinemia (hyperlipidemia) falls into 2 major subdivisions from the standpoint of differen- tial diagnosis on the basis of etiology. The second- ary kind is an expression of altered metabolism due to some other recognizable disease, such as the nephrotic syndrome or hypothyroidism. One deals with the disease and ignores the symptom; hyperli- poproteinemia will go away if the underlying dis- ease is successfully managed. Primary hyperlipopro- teinemia includes all that is left. It is either familial or sporadic. The term heritable is less commonly used than familial, but it is more accurate if there is good evidence of genotypic variation. Familial hy- perlipoproteinemia obviously need not be inheritable if it is due, for example, to patterns of excess in diet or alcohol intake that have been acquired by close relatives. In a desire to use more specific terminology cer- tain durable nouns from the "lipid era" should not be discarded. Hyl)eTcholesterolenEiu and ht/l)erglyc- eridemirr convey exact meaning, as does h!/)rerli- I,emio (hyperglyceridemia severe enough to cause lactescence in plasma). Exogenous hyl~erlil~ernicl is synonymous with fut-induced hyl)erlipemiu. Endog- enozls Izy~,erZilJemiu is sometimes carboh!ydrute in- dtcced. but these terms are not necessarily synony- mous. Two venerable modifiers, essential and idio- nathic, must be laid to rest. Their use no longer defines specific diseases and merely conceals heter- ogeneity. A diagnosis of "essential hyperlipemia," for example, should no longer be acceptable to ei- ther physician or patient. Sorting out the Abnormals There is no single test or maneuver that infallibly separates all those who have hyperlipoproteinemia from those who do not. The majority of laboratories still employ a combination of chemical measure- ments of plasma lipid concentrations for this pur- pose. Lipid Determinations The simplest screening method is a measurement of the total lipid content of plasma. Useful methods are available.`51 Some are relative, such as those'de- pending on turbiditylst; others, like the measure- ment of total fatty acids,`j"*`54 are more specific but also more complicated. All need careful standard- ization. Sometimes, laboratories report "total lipid" concentrations that are less than the sum of the concentrations of several lipid classes reported con- comitantly. Measurement of total lipids alone never provides a specific diagnosis, and occasionally ab- normal but reciprocal changes in cholesterol and glyceride concentrations can occur without throwing total lipid concentrations into a clearly abnormal range. Combination of lipid anulyses. Methods for de- termining cholesterol concentrations are widely availablelss and have even been successfully auto- mated.156 One should be included in any simple scheme for screening for hyperlipoproteinemia. It should be used in combination with 1 of 3 other tests or maneuvers: inspection of the serum for tur- bidity (a gross test for hyperglyceridemia); or the determination of either total lipids or glycerides. The best combination is measurement of cholesterol and of glyceride concentrationsijrr'jn If both are clearly within normal limits, hyperlipoproteinemia is ruled out with a degree of precision that is quite adequate for current use. Phospholipids. The addition of a determination of the total phospholipid concentration to plasma lipid analyses is not difficult. When the ratio of plasma cholesterol to phospholipids is high it indicates a relative preponderance of p lipoproteins. It is rela- tively low in the great excesses of a lipoprotein that may accompany obstructive liver disease. Phospho- 17 lipid determinations do not offer unique information to obtain quantitative lipoprotein pattems.%Jm*16' To about most other types of hyperlipoproteinemia, do so in a single plasma sample requires 3 or 4 however, and they do not supplement cholesterol serial runs and subsequent lipid determinations. and glyceride determinations in such a way as to The premium for adaptation of either the analytical eliminate the added value of lipoprotein determina- or the preparative ultracentrifuge to the study of tions. Plasma phospholipid concentrationsl5Y have many patients is thus high in terms of both instru- therefore not been included in this review. ment cost and operational time. Lipoprotein Patterns The 4 major groups of lipoproteins offer more variables than lipid determinations do. The patterns that these form can be used to diagnose several specific lipoprotein deficiency states already dis- cussed and to segregate at least 5 different syn- dromes or groups of diseases associated with hy- perlipoproteinemia (Table 2). Of all the methods for obtaining lipoprotein patterns only 2 kinds have the necessary range to achieve this segregation. The first are those based on flotation and use either the preparative ultracentrifuge alone or in combination with the analytical model. The latter comes closest to the ideal of defining hyperlipoproteinemia by a single operation or, more accurately, a series of such operations.64 As now adapted the analytical ultra- centrihge has the ability to draw a continuous plot of the concentrations of lipoproteins in flotation (S,) classes differing by very-small-density increments.% Such instrumentation is not generally available. Electrophoresis is less quantitative but much more convenient and economical. It is adaptable to the screening of large numbers of subjects at rela- tively low cost. Visual inspection of properly stained strips permits the immediate recognition of most normal patterns and certain abnormal ones of specific types. The definition of borderline abnor- malities and the resolution of different types is achieved by ancillary determinations. Four addi- tional steps have been adopted by us for sequential employment as they may be required to interpret the paper electrophoretogram. These include a de- termination of plasma cholesterol and glycerides, the precipitation of all lower-density lipoproteins and, occasionally, a single run in the preparative ultracentrifuge. The last is required to determine the quantity of fi lipoproteins and whether the lipo- proteins of fi mobility have normal density. The preparative ultracentrifuge can also be used The features of the electrophoretogram along with the further steps to its interpretation are sum- marized in Figure 7. The successful definition of lipoprotein patterns by such a systematic approach `T`ABLE 2. Types of Hyperlipopoteinemia as Defined by Vanour Inciexrs to Plasm Lipoprotein Co~mwtrcrtimr~." - 1 Milky II Clear 111 Turbid IV Turbid tt t t tt Chylomicrons s, 100-400 t present: all other lipo- proteins L Bttt s,n-12 t t t Pre-6 t 2 20.100 7' "Broad 8" s,o-I:, 1; pR.Se"t: 1%IO0 t t pre-j3 1 -t (? typical sub- (requires Ch pattwn). ultracenu-ifuge LO show beta lipoproteins of density I.006 t 2. V.11 iable HedV) Low-density Not known I' rpoproteins I .OO6- I.063 1 ; re-8 lipoproteins and chylomicrons; the fraction of density greater than 1.006 lesh the (I lipoproteins gives the concentrxion of p lipoproteins; the a lipoproteins represent the supernatant of the fxecifji- tation step. The entire set of procedures can he f'erformed on all hut very lipemic serums, in which the hepal-in and manganese may not precipitate the largest particles. Electrophores~r. The fractions of density greater and less than 1.006 separated in the ultracentrifuge are also subject- ed to paper efectrophoresis, and the mobility of their fipo- protein hands compared with each orher and that of. the lipoproteins in the whole plasma. The cholesterol concentrations shown in Table 3 were measured by horh the Abell modification'"5 and rhe Autoan- afyzer technic.`5" There was no significant difference between results obtained with the 2 methods. Triglycerides were determined by 3 technics.`57*`5BS'6' There were no Ggnihcant differences among results obtained by these 3 methods. @!tta,zt$ficn~zon /mm poprr \tr+. Many workers have shown that the a-lipoprotein and P-lipoprotein hands can he quantified from paper efectrophoretic strips, either direct+ or after elution.:"":" This has not been demonstrated Ior the pre-p and chylomicron bands. We have fjreferred to ure the strips only for qualitative patterns. With experience one can estimate from the appearance of rhe strips the plasma cho- lesterol and glyceride conrentrarions to within IO to I.5 fxl- cent of the chemical detern.inations. For purposes of com- parison each day'% electrophoresis run always includes a sample from a control subject whose lipoprotein concentla- tions are well known. Setting Normal Limits The interpretation of the lipoprotein pattern de- pends upon standards for normal concentrations of the lipoproteins. The setting of "limits of normal" for biologic quantities is often arbitrary. What is "usual" for one population may not be for another and is not necessarily healthy for either. Multiple modes of distribution for the concentration of plasma lipids or lipoproteins have rarely been dem- onstrated, and one is usually forced to rely on fidu- cial limits that assume a normal distribution.lfi" The standards differ with age and sometimes with sex. A set of cut-off points that help to define the usual limits of plasma concentration of the variables needed to interpret lipoprotein patterns is presented in Table 3. These are approximations, and for prac- tical purposes, some of the limits have been com- bined for the 2 sexes even though modest but significant differences actually exist. The sex differ- 19 ence in o-lipoprotein concentrations has been main- tained (Table 3) since the lower limits do have value in detecting heterozygotes for Tangier dis- ease."" The upper (5 per cent) limits for all the quantities are relatively high because the samples from which they are calculated are rather small. This should bias interpretation of patterns in the direction of mislabeling as "normal" subjects some who have marginal hyperlipoproteinemia. Age-related changes. Lipid and lipoprotein con- centrations do not progress stepwise as suggested by Table 3 but as continuous and nonlinear function of age that is not necessarily identical for males and fema]es,"`,l"4-l"" The lowest lipoprotein concentrations are those in cord blood. The mean cholesterol concentration is about 70 mg. per 100 rnl.,i~r~rRH and the a lipopro- teins are about half and the p lipoproteins about a third of the concentrations shown in Table 3 for the youngest age decrement.fir At this time there are practically no pre+ lipoproteins. Within the first few hours after birth the infant is forced to call upon the fat reserves that he has accumulated main- ly in the last trimester of pregnancy. The initially low concentrations of free fatty acids are doubled, and the respiratory quotient begins to fall.rfiv The transport of endogenous glyceride, perhaps required mainly to take care of overshoot in release of free fatty acids, begins in this early period. The mecha- nisms for transporting exogenous glyceride are also activated with the first feedings. Therefore, the de- mands upon p and possibly CY lipoproteins should increase very early. Indeed, the concentration of p lipoproteins doubles or trebles within the first week of life, and lesser but definite increases in LX lipo- proteins also occur.irlfi,rri A very slow ascent in lipo- protein and cholesterol concentrations continues until well into the third decade.172 For practical purposes pediatricians may use the limits in Table 3 for the first two decades without any correction except for the immediate postnatal period. In the third decade there begins a "third phase" in which concentrations of p and pre-P lipoproteins rise at a new and more perceptible rate.rfi~~r~s These increases are probably expressions of the change in fuel economy that is taking place at this time, Physical growth is ending, and the subject becomes more sedentary; caloric excess is easier to achieve, and perhaps other environmental and humoral fac- tors come into play. There is no general agreement about lipoprotein concentrations after about the age of sixty. From the available data it appears that the rise is over, at least for men. One must be very careful not to overinterpret lipoprotein determinations in very old subjects, and this sometimes poses difficulties in kindreds in which a younger propositus has hyper- lipoproteinemia. Naming the Patterns In Figure 7 the terms "Type I," "Type II" and SO forth appeared without comment. These are short- hand designations that were originally used to de- fine different phenotypes of hyperlipoproteinemia because the existing nomenclature for the familial syndromes was inadequate and frequently mislead- ing.r5g~`r"~`r4 They have proved to be of such value for ready communication both in the laboratory and in the clinic that they may be used to denote specific lipoprotein patterns whether they are asso- ciated with primary or secondary hyperlipopro- teinemia. The advantage is one of convenience. The nomenclature for lipoproteins is unhandy and may be confusing when different technics are being con- sidered. For example, the synonym "Type IV" is more convenient than either "hyperprebetalipo- 20 FIGURE 7. Step,\ in Interpreting the Paper Electrophoret(~,~~cm. proteinemia" or "increased very-low-density (less than 1.006) lipoproteins." Since the clinical features associated with the different patterns tend to be specific, the type designation is frequently used here for either lipoprotein pattern or syndrome. All the nomenclature for these diseases, especially the genetically determined ones, must someday be based on a description of the responsible metabolic defect. The type system is only a temporary solu- tion. The numbering of the lipoprotein patterns accord- ing to type has been arranged in a mnemonically convenient way. This can be seen by comparison of Figure 5, the normal lipoprotein pattern, with simi- lar figures that follow `and show the abnormal pat- terns. The numbering begins at the origin of the paper electrophoretic strip, with Type I referring to the presence of chylomicrons, Type II to hyperbe- talipoproteinemia and so on as the bands sequen- tially occur on the strip. We may now begin to examine each of 5 abnor- mal types of lipoprotein patterns associated with hyperlipoproteinemia. This classification is an evolving one based on continuing studies at the Clinical Center that have emphasized genetically determined disorders.159*17"175 It must be kept in mind that the lipoprotein patterns do not necessarily reflect genetic abnormalities, nor do they imply abnormal metabolism of the lipoproteins themselves as opposed to changing demands for fat transport. A single abnormal pattern may be the expression of one of several very different diseases. Finally, it is stressed once again that the index lipoprotein pat- tern for purposes of classification is that obtained on a normal diet. TYPE I HYPERLIPOPROTEINEMIA'~~ General Definition. The Type I lipoprotein pattern, in the system underlying this review, is characterized by the presence of chylomicrons in high concentration in plasma fourteen hours or more after the last meal of a normal diet. The chylomicrons carry dietary glycerides and are particles so large that they scatter light and cause hyperlipemia that is properly called exogenous or fat induced. Of several theoretically possible reasons for this type of hyperlipopro- teinemia there is at present good evidence concern- ing the existence of only one, a decrease in the activity of the enzyme lipoprotein lipase. Retarded removal of chylomicrons from plasma is associated with other disorders that may not involve decreased activity of this enzyme. Some of these will be taken up later under the mixed form of hyperlipemia, Type V. Lil)olJrotein puttern. By the Type 1 pattern is meant hyperchylomicronemia in fairly pure form. Pre-/3 lipoproteins may be slightly increased. A de- creuse in (Y and p lipoproteins is the rule. Paper electrophoresis using the albumin-barbital buffer is at its most valuable in the quick and usually clear- cut demonstration of a chylbmicron band (Fig. 8). Other ways to define the Type I lipoprotein pattern are listed in Table 2. The representation of chylomicrons in so trun- cated a density spectrum as available in Figure 8 is schematic in the extreme. The cut-off points used in the ultracentrifuge to separate chylomicrons are arbitrary, and a precise definition by flotation is not available. A crude and eminently practical definition is that of allowing lactescent plasma to stand over- night in the icebox. A discrete cream layer on the top usually represents chylomicrons. Plasma lipids are helpful in detecting hyperchylo- micronemia, but are not definitive. If the patient is on a regular diet the glycerides will exceed the cholesterol by a ratio (milligram per milligram) of about 8:l or higher. The proportion of free choles- terol will be high, about 50 per cent instead of the usual 30. Under the right conditions chylomicrons are precipitable by dextran sulfate or heparin along with other low-density lipoproteins. The chylomi- cron precipitate often does not sediment, however, and routine precipitation technics (Table 2) yield quite variable results in severe chylomicronemia. They also cannot discriminate between exogenous and endogenous hyperlipemia. DENSITY PROTEIN WQ! PROTEIN FIGURE 8. Lij+oprotein Pattern in Type I Hyperltjnprc/lelnemrcl (See Also Table ?). The key clinical features oj the familial syndrome ore early exprec- .rion, bouls of abdominal pain and other accompanrmtnt~ a/ www hyperlipemia, low posthepnn'n lipol~tic nctiuitv (PHLA) and autoso- mnl recessive tmwm0sion. 21 It will be noted in Figure 8 how scanty are the amounts of (Y and B lipoproteins in severe hyper- chylomicronemia. This is also seen with the excess- es of pre-/3 lipoproteins or chylomicrons in Types IV and V. From evidence developed earlier it appears that these smaller lipoproteins are "bound" to or otherwise associated with the triglyceride particles so that they are no longer detectable in their usual density range or electrophoretic position. Although concentrations of free fatty acids have not yet been proved to be helpful in diagnosis they are low in Type I and do not show the normal rise on feeding of fat. This is in accord with an assumed defect in hydrolysis of glyceride at the capillary wall. Secondary Type I There are cases in which abnormal chylomi- cronemia occurs in association with other forms of hyperlipoproteinemia. These are discussed later under Types IV and V and include difficulty in re- moving exogenous glyceride that appears to be sec- ondary to such problems as uncontrolled diabetes,`76 pancreatitis177-`7" and acute alcoholism.1"" Whether hyperchylomicronemia secondary to other diseases exactly mimics the Type I pattern seen in the fa- milial lipase deficiency is uncertain. Primary (Familial) Type I The most severe degree of hyperchylomicronemia is seen in patients who apparently are homozygous for a rare mutant gene regulating the activity of the clearing enzyme or enzymes. All familial Type I hyperlipoproteinemia ultimately may not turn out to represent lipoprotein lipase deficiency, but this is the only biochemical defect currently recognized. History. The history of this syndrome has been reviewed elsewhere.`"!) It is somewhat more straightforward than that of the other types of fa- milial hyperlipoproteinemia. The manifestations are so dramatic that few if any of the earliest case re- ports have been ignored. The example generally accepted as the first representative of the primary familial Type I syndrome was the case of a twelve- year-old boy reported in 1932 under the title "hepatosplenomegalic lipidosis" by Burger and Griitz.lb* The first patient with a relative who also had hyperlipemia was described in 1939 as having "idiopathic familial hyperlipemia."`S' After this de- scription appeared many other cases were reported under the title "idiopathic" or "essential hyperli- pemia," but close scrutiny of the case reports sug- gested that only a few of these qualify as the Type I syndrome. If fairly rigid criteria are used to define the syndrome, about 35 acceptable case reports can be found.`"!' Examples of primary Type I have been described by the following synonyms: idiopathic familial hyperlipemia; essential familial hyperli- pemia; retention hyperlipemia; alimentary "hepato- genlc " fat retention; familial lipemia; fat-induced hyperlipemia; and familial hyperchylomicronemia: Of these definitions only the last 2 seem acceptable. One defines the characteristic lipoprotein pattern, and the other the origin of the glycerides accumu- lating in plasma. All synonyms; including the short- hand term, Type I, can be replaced by a more specific one of lil,o))rotein-lil,ase deficient familial hyperchylomicronemia when it has been ascertained that the patient qualifies for such a diagnosis. For this one needs to have certain information beyond the lipoprotein pattern, including the patient's en- zyme response to heparin. Clinical munifestations. The important clinical manifestations summarized in the legend of Figure 8 are sufficiently reproducible to permit the synthe- sis of a "typical case history." It may begin at a pediatrician's office with the appearance of a mother whose one-month-old child looks healthy but has "bouts of colic" and an unusually prominent abdo- men. Recently, some yellow papules with a reddish base may have broken out over the skin and oral mucosa. An enlarged liver and spleen are felt, and hospitalization for observation is recommended. Here, the intern sees nearly white retinal vessels (lipemia retinalis). As the blood sample emerges in the syringe it looks like "cream-of-tomato soup." The lipoprotein pattern is established, and the baby's formula is changed to one containing only skim milk or other fat-free sources of calories. Within three days the hyperlipemia has cleared dramatically. The xanthomas will shortly resolve, the liver and spleen will decrease in size, and the apparent attacks of abdominal discomfort will disap- pear. Such a child is more fortunate than the patient with the same genotype whose abnormality is not detected until he is old enough to describe his ab- dominal discomfort in details that suggest one of a variety of acute surgical emergencies. He may un- dergo laparotomy before the nature of his syndrome has been appreciated. Although the majority of pa- tients are discovered before the age of ten, some may be adults when the diagnosis is first made.159 Diabetes. Glucose intolerance is not a feature of Type I and the usual oral and intravenous tests are normal even when the patient has severe hyperli- pemia. This is in contradistinction to other types of hyperlipemia, including some that are "fat-induced" (see Types IV and V). Diugnosis. The diagnosis of this syndrome is usu- ally not difficult and entails three steps: identihca- tion of the Type I lipoprotein pattern; ascertain- ment that the glyceride accumulation is immediately related to dietary fat intake; and measurement of plasma postheparin lipolytic activity (PHLA). Within a few days of switching a patient with a Type I pattern from a regular fat intake to less than 5 gm. of fat per day, one will note a rapid decline in plasma glycerides or hyperlipemia. The lipopro- tein pattern will evolve in a predictable fashion. Chylomicrons will disappear, the p-lipoprotein band will increase, as will the OL, but the most striking 22 change will be abnormal accumulation of pre-/3 lip- oproteins.7~`5g The glyceride concentrations rarely become completely normal on the fat-free diet be- cause the patient is now moderately "carbohydrate induced." Presumably, the means of removing en- dogenous glyceride are similarly affected by the inherited abnormality, and even the normal increase in endogenous glyceride that follows the shift to a high-carbohydrate diet causes a slightly greater hy- perlipemia than normal. The paper electrophoretic technic is most helpful for observation of the fasci- nating and instructive changes in lipoprotein pat- terns in Type I with different diets. Lipoprotein lipase activity should be assessed in all such patients.i*%1"s PHLA is best assayed under in vitro conditions capable of determining maximum reaction velocity. In 12 patients from 11 kindreds we have found values from 0.04 to 0.20 (in PEq. of free fatty acids per minute per milliliter of plasma). This is below the range of values (0.24 to 0.60, mean. about 0.40) found in over 100 subjects with either normal or Type II-IV lipoprotein patterns. One other patient with the familial Type I syndrome has persistently had normal PHLA, although his sibling, with an identical lipoprotein pattern, had abnor- mally low values.lj!l Phenocopies. Low PHLA, as defined by the assay used for the Type I patients described above, has been reported in patients with untreated diabetes"" and hypothyroidism.`"" This activity is also low in other syndromes, and the enzyme assay is not a specific determinant of genotype. One other prob- lem plagues the definition of the familial Type I syndrome. This is the possibility that in some of the patients the disorder has "converted" to another type of hyperlipoproteinemia because of pancreati- tis, diabetes or prolonged intake of an abnormal diet. One way to be fairly certain of the diagnosis is the detection of another typical example of Type I in the patient's family. Inheritance. In 5 kindreds with primary Type I, multiple sibs had had equally severe hyperli- pemia.m' Occasionally, very mild hyperlipemia has been present in 1 parent or 1 or more siblings. The study of kindreds has been inadequate, particularly concerning the need to demonstrate that a relative with hyperlipemia does indeed have the Type I pattern and not, for example, the much more com- mon Type IV pattern. Both parents of an affected child may have normal lipoprotein patterns although in several kindreds at least 1 has had mild hyperli- pemia. The relatively small number of involved sibs supports the assumption that a double dose of an abnormal gene accounts for the severe Type I phe- notype and that a single dose of the gene may pro- duce little or no detectable abnormality. An unusually higli number of parents or sibs have PHLA in or sliglltly below the lowest quartile of the distribution of the enzyme in normal subjects.`5!' The slightly low activity may be associated with normal postabsorptive glyceride concentrations. As illustrated by the presence of the 1 homozygous abnormal patient with normal PHLA, a more specific assay for enzyme activity is needed. At- tempts have been made to measure the enzyme in human adipose tissue.1~7 The activities are very low. In 1 report, levels significantly lower than normal were found in a family with fat-induced hyperli- pemia.iXX Mechanism. The few well studied examples of Type I suggest that the inheritable defect in all has been a deficiency of lipoprotein lipase activ- ity. IWW~IW Hydrolysis of glyceride at sites of re- moval is decreased. This imposes severe limitation on the rate of clearing of chylomicrons and probably that of glycerides in other particles or lipoproteins. The patients with familial Type I syndrome have almost no threshold for fat removal, as though the entire normal clearing mechanism were inoperative. This presents something of a paradox, for the evi- dence is far from convincing that lipoprotein lipase is present in the liver. It seems either that the liver does not have an essential role in clearing chylomi- crons or that lipoprotein lipase deficiency is not the basic defect in Type I. The generalized explanation of this disorder therefore hangs upon better under- standing of the physiologic mechanisms of fat clear- ing. There is a need to examine carefully all new ex- amples of Type I to be certain that lipoprotein lip- ase deficiency is indeed present. Among other possible mechanisms, the production of abnormal chylomicrons or presence of circulating inhibitors of lipoprotein lipase has been eliminated in some pa- tients.159 Defects in the re-esterification or further utilization of glyceride fatty acids that theoretically could limit the rate of clearing have not been ex- cluded, but the persistently low levels of free fatty acids in Type I argue against such possibilities. Management. The outlook for patients with the "pure" Type I familial syndrome is not yet pre- dictable. The well documented cases still include a relatively young group. Most have learned they must limit their daily fat intake if they are to avoid bouts of abdominal pain. The origins of the pain are still obscure, and it can occur as well in patients with other kinds of severe hyperlipemia. Some- times, it is accompanied by the usual chemical signs of pancreatitis. It has been speculated that this or- gan is compromised by fat embolization or perhaps local lipolysis, giving rise to irritatingly high free fatty acid concentrations. Proof is lacking for any mechanism. On other occasions, similar pain is not associated with any rise in serum lipase or amylase content. The pain can be restricted to a single organ such as the spleen, which may be exquisitely ten- der. Because pregnancy may exacerbate hyperli- pemia special care must be taken with these pa- tients during this time. Several patients have delivered normal babies. In addition to the painful attacks, the retarded removal of dietary fat has a number of side effects. 23 Chylomicrons awaiting access to their usual sites of disposal seem to be prime targets for uptake by reticuloendothelial cells. Large foam cells appear in the bone marrow; eruptive xanthomas arise in the skin, and the liver and spleen enlarge. These changes are less dramatic but also urge limitation of fat intake in the hope of avoiding possible compro- mise of the functions of these and other tissues. Atherosclerosis. In the 35 or so patients with the familial Type I syndrome there has so far been a lack of evidence of accelerated coronary-artery dis- ease. This is not proof, of course, that a high con- centration of glycerides in plasma in the form of chylomicrons represents no particular hazard for the vessel wall. The inference is tempting, however, and may be used at least in the argument against extreme limitation in fat intake. For in these pa- tients, this will lead to accumulation of pre-P lipo- proteins. The latter seem to have a less benign as- sociation with vascular disease. Diet. The treatment of Type I, then, is moderate restriction of dietary fat. The best motivated of these patients will usually select about 20 to 25 gm. of fat per day. So far as the degree of saturation of the fatty acids in the dietary glycerides is con- cerned, the source of the fat makes no difference to the degree of hyperlipemia. It has been popular in recent years to feed tri- glycerides containing medium-chain-length fatty acids (MCT) to patients with abnormal absorption or fat-induced hyperlipemia. This permits fat intake without chylomicron excess, since these fatty acids are taken into the body by a different mechanism, apparently through the portal system. This type of fat may induce higher concentrations of pre-fi lipo- proteins, and its long-term safety is uncertain. Parenterally infused fluids and no oral intake are the best treatment of the acute abdominal symptoms that may accompany Type I. Intubation may be necessary to relieve distention and ileus. Other "antihyperlipemic" medications thus far available have not been shown to have a place in treatment of this disease. All the patients have some rise in free fatty acids in response to heparin injec- tions, and this drug will promote some intravascular lipolysis. The small amount of lipolysis is not enough to clear the plasma, however, and the use of heparin in Type I is without good rationale. "Intermittent" Primary Forms of Type I Examples of severe intermittent hyperlipemia resembling Type I may occur. We have examined the plasma of a twelve-year-old girl (a patient of Dr. Allen Cracker) whose severe, `*fat-induced" hyper- lipemia seemed to be intermittent even on a regular diet. She had a typical Type I pattern on one occa- sion and a completely normal pattern some weeks later. There was no familial hyperlipoproteinemia. Such a case is puzzling in the extreme. It can be easily calculated that if one absorbed all of a IOO-gm. daily intake of fat, and the removal of this were suddenly and completely blocked, a severe hy- perlipemia of the order of 3000 mg. per 100 ml. could be attained in one day. Presumably, some "toxic" factors could so interrupt the clearing process at one step or another in a transient or sporadic fashion, but there is little or no knowledge of what such factors might be. Good examples of the inheritable Type I or their equally interesting phenocopies are hard to come by. They deserve uncommon attention. Possible lesser degrees of Type I. In our experi- ence with paper electrophoresis even the faintest of chylomicrons is very rarely detected in apparently healthy Americans of any age after an overnight fast. In young adults the chylomicron tide has disap- peared by midnight, six hours after the last meal. When the daily fat intake, spread over 3 meals, is increased to a total load that is two or three times greater than the usual American intake, chylomi- crons are still briskly removed, and the lowest glyc- eride measurements in the daily cycle are regis- tered just before breakfast. Our experience that hyperchylomicronemia is extremely unusual when electrophoresis is used for screening is not necessarily at odds with other suggestions that fat tolerance may decay with age or other conditions190 or that minor degrees of intoler- ance might be genetically determined.191 The paper electrophoretic technic is relatively insensitive and may not detect very small amounts of chylomicrons. The fast for twelve to fourteen hours is also long, and some shorter sampling time after the fat feeding might provide a better discriminant. The ideal fat tolerance test has not yet been devised. At present it is not known whether there are mild degrees of inability to remove dietary fat that are genetically determined or sporadic. Only the grosser abnormal- ities can be reliably detected. TYPE II HYPERLIPOPROTEINEMIA~~~,~~~~~ General De$nitions. By the Type II lipoprotein pattern we mean an increase in the concentration of lipopro- teins that have discrete /3 mobility by electrophore- sis and the normal density and chemical composi- tion of P-migrating lipoproteins. This kind of hyperbeta-lipoproteinemia must be distinguished from another in which the density of the fi-migrat- ing lipoproteins is abnormally low, giving rise to a pattern designated as Type III in this system for defining types of hyperlipoproteinemia. The bases for distinguishing the hyperbeta-lipoproteinemia of Types II and III arise from considerable clinical and genetic evidence that they are expressions of different inheritable metabolic disorders. Type II is a common pattern; it can be a resultant of diet or secondary to hypothyroidism and other diseases, and in many patients it proves to be an expression of a mutant gene or very similar genes, appearing in relatively high frequency in many populations, in- cluding the North American. 24 Lirvrotein pattern. The illustration of the Type II pbrrern in Figure 9 depicts two important fea- tures. Foremost is the discrete increase in fi lipo- proteins, which on the paper electrophoretic strip is visible as a sharp band taking increased amounts of lipid stain with almost malevolent intensity when one considers the clinical implications of the ab- normality. This sharp band is indicative of an in- crease in the low-density classes of both S, O-12 and 12-20. Normally, the concentration of the S, O-12 class is about five times that of the 12-20 class, and this differential is usually maintained in the Type II pattern. 64,1Y4,1Y6,lY, The second feature of the Type II pattern, and one that we are anxious to clarify, is that it must not exclude a small increase in glyceride concentrations. In analyses of 100 consecutive patients with estab- lished familial Type II patterns that are in prepara- tion for publication, the mean glyceride concentra- tion is about 150 mg. per 100 ml., and the upper limit about 500. This is associated with a modest increase in pre-P lipoproteins or lipoproteins of S, 2@~()(),lY4-`"7 After comparing the glyceride and P-lipoprotein concentrations associated with several hundred lipo- protein patterns, we conclude that in Type II a given increase in glycerides will be represented by about the same increase in pre-fi lipoproteins. This is noteworthy, since the overwhelming concentra- tions of p lipoproteins in Type II might be expected to "accommodate" abnormal amounts of glyceride without the presence of a pre-/3 band on the elec- trophoretic strip. Some Type II patients have no pre-P band (or increase in S, 20-100 lipoproteins); others may have distinct and sometimes abnormally increased amounts of pre-/I lipoproteins. The cur- rent definition of Type 11'74 differs from that de- scribed by us in our first publications related to phenotyping by lipoprotein pattems.i57,*73 Formerly, the combination of increased p and pre-/3 lipopro- teins was considered one kind of Type III pattern. Type III has since been more explicitly defined (as discussed below). Determination of hyperbeta-lipoproteinemia. The age trends of plasma cholesterol concentrations in 110th sexes in the general population tend to parallel those of the @lipoprotein (S, O-20) concentra- tions."4,`98 (See also Table 3.) The concentrations of S, 20-400 (pre-p) lipoproteins also rise with age.64'188*19Y Thus, an intense P-lipoprotein band on the electrophoretogram can definitely be interpreted as abnormal when the plasma cholesterol concen- tration is clearly beyond the upper limit of normal (Table 2) only if the triglyceride concentration is not in excess of 150 mg. per 100 ml. Instead of a triglyceride determination the plasma must there- fore be observed to be perfectly clear. When there is uncertainty the quantity of /3 lipoproteins must be specifically determined, either as described earlier or by some comparable technic. When the native plasma or serum is perfectly clear the various chem- DENSITY Sf I 1.006 -20 1.063- 0 1.21- PAPER STRIP PROTEIN Or PROTEIN FIGURE 9. Type II Hyperlipoproteinemia (See Also Table 3). The familial form is transmitted as an autosomal dominant and ic umalty accompanied tq palpetrral, tendon and tuberour wanthow, cornea1 OTCUJ and acceleratid a&rosc&~osis. Clyceri&s (p-e-/3 lipoproteins) may be modestly elevated; PHLA ir normal, and glu- come tolerance is usually normal. ical precipitation technics or those depending on antibodies to /3 lipoproteins yield their least equivo- cal results and provide a fairly reliable index to Type II hyperlipoproteinemia. The establishment of any dividing line between normal and abnormal P-lipoprotein concentrations is somewhat arbitrary, as already discussed. The limits provided in Table 3 are more likely to be too high than too low, at least in the context of defining the most healthy concentrations. The suggested upper limit for any age in Table 3 is 210 mg. per 100 ml. (as lipoprotein cholesterol), and in the younger age group, it is lower. The mean P-lipoprotein concen- trations for the normal levels in Table 3 are very close to those in a smaller series previously re- ported.i6i Sets of normal values for p lipoproteins have been obtained in larger samples by means of the analytical ultracentrifuge.64*iss The average p-lipo- protein concentrations (S, O-12 plus S, 12-20) in the latter analyses, when estimated arbitrarily in terms of cholesterol content, are slightly higher than those shown in Table 3. The mean increased with age into the sixth decade in these analyses as it does in Table 3. Major assistance in judging what are "normal" limits has been provided by kindreds with the fa- milial Type II syndrome. Affected adult members, on a regular diet, rarely have P-lipoprotein concen- trations below 250 mg. per 100 ml. Some of the young children affected and some of the adults on strict diets may fall below 200 mg. From inspection of the data being gathered in the Clinical Center studies it appears that bimodality in the distribution of /?-lipoprotein concentrations within Type II fam- ilies will be demonstrable. From a few Lebanese families in which several probable homozygous abnormal members appear it seems likely that 3 modes will be demonstrable when a population sample containing a sufficient number of marriages 25 between heterozygotes has been examined.*M) Pre- vious attempts to show several modes of distribution of plasma cholesterol concentrations in affected American kindreds have, at the most, been only partially successful.1*4*2111 Enoironment and the Type II pattern. The inter- pretation of possibly abnormal P-lipoprotein concen- trations in the "gray zone" of about I75 to 225 mg. per 100 ml. (as cholesterol) is difficult. Depending on the age of the patient, the case may have to be considered primary "Type II" even though the fam- ily history is negative. Such "probable abnormal patients" are not infrequent in groups with acceler- ated atherosclerosis. The interpretation of their lipo- protein patterns requires special attention to their living habits. It is generally accepted from epide- miologic studies and easily demonstrable experimen- tally that otherwise normal subjects who eat a great deal of cholesteroW"" or an increased amount of saturated fats.12 generally have higher plasma cho- lesterol concentrations than those who eat much less of these foodstuffs. There is no consensus of whether sterol or the type of dietary glyceride is the prime determinant. 204 A P-lipoprotein concentration cannot be properly interpreted, therefore, without knowledge of the patient's diet. Although the lipo- protein level can be driven down by certain diets to much lower levels in normal subjects than in pa- tients with the Type II pattern adequate study has not yet been made of the possibility that response to diet alone may segregate primary hyperbeta-lipo- proteinemia into sporadic and inheritable forms or further subdivide patients within these main groups. Dietary restrictions should not be imposed on a patient with a Type II pattern until the etiology has been carefully sought. The effort expended in changing a patient's food habits, and often those of his family requires that metabolic determinants other than diet be excluded and that the degree of involvement shared by the members of the family be ascertained. Secondary Type II Type II hyperlipoproteinemia accounts for a great deal of the "hypercholesterolemia" that the physi- cian sees. As with other abnormal lipoprotein pat- terns, the most systematic approach to diagnosis and treatment first requires the elimination of certain diseases of which the hyperlipoproteinemia is a secondary manifestation. Probably the most common of these is hypothy- roidism. Next most common are some types or stages of obstructive hepatic disease and hypopro- teinemias like the nephrotic syndrome; quite rarely one may encounter this pattern with myeloma, mac- roglobulinemia, idiopathic hypercalcemia and other unusual disorders. Obstructive jaundice usually provides a unique pattern that will be described separately. Others of these disorders will be dis- cussed in the concluding section of this review un- der Type IV, the abnormal lipoprotein pattern with which they are more commonly associated. Hypothyroidism. Myxedema may be accompanied by great increases in @lipoprotein concentrations and a Type II pattern that mimics that due to muta- tion~26tPl~~60~101 According to studies with `alI-tagged lipoproteins there is a decreased rate of removal of p lipoproteins from circulation in hypothyroid- isrn.mS*M6 Carotene concentrations are high,ms as they are in other forms of hypercholesterolemia.*or In hypothyroidism this has been attributed to de- creased catabolism of the p lipoprotein, the major carrier of carotene in plasma. Vitamin A levels can also be 1ow.m~ It has not been proved that de- creased conversion in the intestine of carotene to vitamin A does not occur in hypothyroidism, in- creasing the amount of carotene absorbed and transported. It is still not known how thyroid hor- mone regulates plasma levels of p lipoprotein. The assumption that such a regulatory defect may be independent of calorigenic activity of this hormone is the basis for frequent use of the dextroisomer of thyroxine in the treatment of hyperlipopro- teinemia.208 Hypothyroidism is sometimes associated with hyperlipemias as well. This will be discussed again in the differential diagnosis of Type IV. The proof that a Type II pattern is caused by hypothyroidism rests on the demonstration of hor- mone deficiency and improvement in the pattern as replacement therapy is undertaken. Sometimes pri- mary Type II is not discovered until the lipoprotein pattern fails to return to normal after a patient with myxedema has been made euthyroid. Familial Type ll~Jy~~~2-~Y~ One of the most common inheritable forms of hyperlipoproteinemia encountered in the world is Type II hyperbeta-lipoproteinemia. Perhaps it shouId be called a syndrome since it may be .a group of similar diseases. For convenience it will be colloquially referred to here as "familial Type II." In addition to hundreds of affected American families there are in the literature easily identifiable examples in Caucasians and Negroes from many geographic areas. There is at least 1 report among Asiatics.mY There are no clear-cut data on the gene frequency in any population, including the Ameri- can, and information about its possible occurrence in some areas of the world is completely lacking. One of the major purposes of the studies that are the basis for this review has been the establishment of better criteria for ascertainment of familial Type II and the promotion of their utilization in obtaining this much needed information. Our experience with familial Type II consists of the study, including complete lipoprotein analyses, of over 100 definitely familial examples from about 50 families and a similar number of less well stud- ied familial cases or single patients whose lipopro- tein pattern and clinical features are compatible with the syndrome. There are many other reports of single cases or whole pedigrees in the literature. A number of comprehensive studies indicate close 26 agreement with the present description of this dis- order. ~YZ-M~~O,ZIO The consideration that all familial Type II may be due to a mutation at the same genetic locus is sup- ported by the predictability of clinical expression and genetic mode in all the patients with this lipo- protein pattern. The heterozygote seems to be nearly always recognizable by lipoprotein pattern. Depending upon unknown factors that govern the degree and speed of expression, the accompanying xanthomatosis and atheromatosis frequently seen in the syndrome will develop in many heterozygotes. They usually survive the childbearing period and may have a normal life-span. The homozygous abnormal phenotype has in- creases in p lipoprotein that are nearly double those seen in the usual heterozygote, quite severe xantho- matosis and often death in childhood from cardio- vascular involvement. This includes endocardial deposits of lipid, sometimes causing aortic stenosis, coronary-artery disease and occlusive disease of the peripheral vessels. . c History. Familial Type II seems to have entered medical history surreptitiously, possibly as long as one hundred and thirty years ago. The date will remain uncertain because xanthomas were recog- nized sometime before it was realized that they were due to hyperlipidemia and long before differ- ent metabolic bases for the latter were appreciated. Palpebral xanthomas (xanthelasma) are shown with- out comment in Rayer's*ii atlas, published in 1835. The first descriptions of tendon and tuberous xan- thomas appear under the term "vitiligoidea" in Guy's Hospital Reports in 1851.2i2 The authors, Addison and Cull, were describing lesions associ- ated with the hyperlipoproteinemia due to obstructive biliary disease, as were the other authors of reports of xanthomas in the English and French literature for the next fifteen to twenty years. The appearance of cutaneous xanthomas in a patient free of jaundice with a relative who had the same lesions was brief- ly described between 1868 and 188Z.r13*2*4 There was no mention of plasma lipids, and the descrip- tions of the lesions are not convincing evidence for Type II. Synonyms for familial Type II appeared in medi- cal writing in roughly the following order (with the omission of certain terms for similar kinds of xan- thomas): hereditary xanthomatosis; familial xan- thoma; hereditary xanthoma tuberosum multiplex; general xanthelasma; xanthelasma multiplex; xan- thoma tendinosum; essential familial hypercholes- terolemic xanthomatosis; and, most recently, es- sential familial hypercholesterolemia.i!`* Neither "familial hypercholesterolemia," "xanthomas tendi- nosum" nor "familial hyperbeta-lipoproteinemia" is entirely satisfactory, however; for the quite different pattern syndromes, Types II and III, can both be included in these categorical listings. An ideal name for Type II must await elucidation of the primary abnormality. Diagnosis. The major features of familial Type II hyperlipoproteinemia form a triad. The first compo- nent is the lipoprotein pattern in the absence of other primary disease. The second is distribution of the Type II pattern in the family according to men- delian concepts of an autosomal dominant trait. Both these components should be present for ascertain- ment of what, in present-day genetic studies, can be considered the familial Type II phenotype. For practical clinical purposes the first component, when coupled with florid appearance of the third feature, - xanthomatosis, - is usually adequate for diagnosis. The 3 components of this triad deserve some further, more detailed exposition. Hyperbeta-lipoproteinemia is the first requisite. It has already been defined in detail, and stress laid on the fact that it may be accompanied by moderate hyperglyceridemia. The small pre+-lipoprotein band appearing in many Type II patterns does not give rise to significant hyperlipemia. Therefore, milky serum is incompatible with uncomplicated Type II. The observations of Type II families have not yet established whether the plasma concentration of glyceride or pre-/3 lipoproteins may provide an in- dependent variable that will segregate some fami- lies as representing a different "kind" of familial Type II syndrome. In 1 large kindred glyceride concentrations in affected and nonaffected subjects were not significantly different.136 The "sporadic" occurrence of modest hyperprebeta-lipoproteinemia has the earmarks of representing a "way of life" in many Americans, and one must be very cautious in considering it a discriminant for subgrouping other- wise similar diseases. Familiul appearance of the Type II pattern is the second diagnostic feature. This includes the nearly 100 per cent expectation of a similar lipoprotein abnormality in at least 1 parent and distribution of the abnormality in siblings and other blood relatives consonant with inheritance of a "highly penetrant" trait that is recognizably expressed whenever a sin- gle abnormal allele is present. One abnormal parent will theoretically affect half his progeny. With rare exceptions, the single dose of the abnormal gene is usually expressed by abnormally high /3-lipoprotein concentrations in the young child. In a single case umbilical-cord-blood cholesterol has been abnor- mally high,215 but from evidence in 4 other Type II children, cord-blood cholesterol levels have been normal and have increased into the abnormal range during the first year.216 Occasionally, expression may be delayed until the second decade.159 Tissue lipid deposition. Xanthomas and atheromas are very common features of Type II. The most characteristic manifestation is tendinous xanthomas, located particularly in the Achilles tendons and the extensor tendons of the hands and feet. Often, there are also tuberous xanthomas, particularly over ex- tensor surfaces, including the elbows, knees, hands and buttocks, or periorbital xanthomas (xanthelas- 27 mas). In a severely affected subject soft, tuberous xanthomas may appear at many sites over the body. A comeal arcus is common,*r7 assuming greater di- agnostic significance the younger the age at appear- ance. Accelerated arterial disease, especially involv- ing the coronary arteries, is a life-threatening feature. Some families may be relatively free of this manifestationr"R; others are seriously affected. Oc- clusive peripheral or cerebrovascular disease does not seem to be as common as coronary-artery dis- ease in Type 11.1~4~196 Not typical and tending strongly to militate against the diagnosis of Type II are eruptive xan- thomas, planar xanthomas in the creases of palms of the hands or tuberoeruptive xanthomas. The last term refers to raised, sometimes pedunculated and often confluent intracutaneous lesions that have the reddish, inflammatory appearance of "eruptive" xanthomas, as opposed to the saccular lesions covered by normal-looking skin that are "tuberous" xanthomas in the more accepted sense. We are una- ware of any patient with a Type II pattern and ten- don xanthomas both of whose parents have been shown clearly to have no hyperlipoproteinemia. A family history consonant with such tissue lipid deposition is therefore helpful in diagnosis if the parents cannot be sampled. It has been shown convincingly that tissue lipid deposits increase with time and are related to the degree of abnormality in the lipoprotein pattern's3 Homozygous abnormal subjects can have xanthomas very early in childhood and may even be born with them. Other clinical manifestations. The legend of Fig- ure 9 summarizes the important clinical features of Type II, most of which have already been discussed. The overall incidence of "chemical diabetes" in Type II is not obviously greater than that thought to be present in the general population. A much larger study with suitable glucose tolerance tests is needed, however, to establish the true incidence and to determine whether the hyperprebeta-lipopro- teinemia found in some Type II subjects is related to abnormal glucose tolerance. It also is not known how the presence of other abnormal genes, such as that determining diabetes, might affect the expres- 20 represents t+ supernalant fraction, and S, < 20. the infmnate. Only in Type III do lipoprotkm of /I mobzlzty appear in the I .006-dendy supemte. have been present in many reports of heterogenous patients with severe hyperlipoproteinemia and xan- thomatosis.23"2" It is inevitable that some Type III patients, with their tendency toward tendon xantho- mas, xanthelasmas and crippling atherosclerosis, should have been classified as having "essential hypercholesterolemia" (Type II). Others have been singled out by their milky plasma and yellow-red eruptions on the extremities, and these cases have been consigned to "essential hyperlipemia" - a sanctuary of the undifferentiated that one hopes will soon be abandoned. One study stands out during the period from 1950 to the present, a comparison -. of xanthoma type and lipoprotein patterns at the ,- ," \ Donner Laboratory.6**19 ! \ f Two groups of patients were examined, defined as having "xanthoma ten- dinosum" and "xanthoma tuberosum." The first group almost certainly contained examples of famil- ial Type II; all had marked increases in S, O-12 lipoproteins (hyperbeta-lipoproteinemia). The other group had skin lesions on "extensor surfaces of the elbows, buttocks, extensor aspects of the knees, the hands, expecially the volar surfaces, and over the ankle malleoli, especially laterally."igr In contrast to the patients with xanthoma "tendinosum," some of whom had lesions in childhood, all the 23 with xanthoma "tuberosum" had first noted their lesions in adulthood. Two of the latter were sibs, the only familial examples presented. In these patients the S, O-12 lipoproteins were lower than normal, and the S, 12-400 classes greatly elevated. When the classification of familial hyperlipidemia by the current system was begun at the Clinical Center, one of the objectives was the search for proof of the existence of separate phenotypes com- parable to those suggested by these early studies. The present concept of Type III evolved through trial and error. In the earliest published classifica- tion by electrophoretogramsi5g**r3 a "Type III" pat- tern was defined as an increase in both /3 and pre-P lipoproteins. It was noted that there seemed to be 2 different variations on this pattern. In 1 the 2 lipo- protein bands were intense and distinct In the other they merged in a broad, intensely staining /3 band. As described above, the former pattern was eventually found to be a variant of the Type II pat- tern. The "broad-beta" pattern has come to define a group of patients whose familial expression and clinical features were clearly different.`r4 The ab- normal flotation of the p lipoproteins was noted after the quantification steps were added to the typing system. It has been observed that the /3-lipo- protein anomaly is consistent and reproducible and not a technical artifact. This flotation anomaly is not a function of the absolute amounts of fl lipoprotein, cholesterol or glycerides in plasma. It has never been observed in patients with Types II or IV pat- terns even though their plasma lipid concentrations are sometimes very similar to those seen in Type III. The important difference between the inheri- tance of Types II and III will be described below. Clinical Manifestations Plasma lipids. The plasma cholesterol concentra- tions in Type III vary from 200 to more than 1400 mg., and the glycerides from 175 to more than 1500 mg. per 100 ml. Extreme variability of cholesterol and glyceride concentrations is a diagnostic feature and a distinct contrast to the steady elevations in Type II. The lipoprotein levels are quite sensitive to changes in the content of the diet and the amount of total calories. If the patient is switched from a regular diet to one rich in polyunsaturated fats and containing only 100 to 200 mg. of choles- terol per day both cholesterol and glyceride concen- trations may be lowered severalfold within a few weeks. The effect of negative caloric balance may be even more dramatic. Glyceride concentrations tend to rise moderately on high carbohydrate diets, but Type III is not always accompanied by carbo- hydrate inducibility exceeding the response of nor- mal subjects. Carbohydrate intolerance is the rule in most patients (over 90 per cent), and the ability to handle either an oral or intravenous glucose load seems to decrease with age faster than it does in the normal state. There have been no measurements of plasma insulin levels in Type III; only 1 of 24 patients thus far has ever taken insulin or hypogly- cemic agents for his mild diabetes. The 10 patients with Type III who have been available for testing have had a small accumulation of `chylomicrons after an overnight fast when they were fed more than 150 gm. of fat per day. Their hyperglyceridemia on. regular diets is not exoge- nous, however, as witnessed by their good thera- peutic response to low-cholesterol diets containing 100 gm. or more of fat. The "palmar striae" of Type III are not typical of any other type of ptimary hyper- lipoproteinemia. Xanthomas. The terms "xanthoma tendinosum" and "xanthoma tuberosum" do not accurately distin- guish Types II and III. Tendon lesions occur in both, as do thickened tibia1 tuberosities, cornea1 arcus and periorbital xanthomas. The distinctive 31 lesions in Type III are found on the palmar surfaces of the hands. The creases and sometimes the tips of the fingers or other areas, especially where rings are worn, contain yellow-white deposits that rise a little above the surface; they are often subtle changes and are sometimes overlooked by both patient and phy- sician. So-called planar xanthomas of similar ap- pearance but usually on the back of the hands have long been associated with obstructive liver disease; they can be seen on models constructed from jaun- diced patients over a hundred years ago."' Such lesions are not typical of any other type of primary hyperlipoproteinemia. On the elbows several kinds of lesions have been observed in Type III. Soft, pedunculated growths may be present, but they tend to be confluent and smaller than tuberous xanthomas usually seen in Type II and do not have the smooth normal skin cover of the latter. The Type III lesions bear the reddish, inflammatory appearance that is the trade- mark of skin lesions associated with hyperglycer- idemia. Sometimes, they are confluent, yellow-red papules raised just above the surface. Occasionally, when they seem to be involuting they closely re- semble dabs of peanut butter spread on the skin; these last two forms may commonly be amalgamated into brownish-yellow papules having a reddish base. Similar lesions in severe diabetes have been called "xanthoma diabeticorum." For all these the name "tuberoeruptive" may be helpful in suggesting their appearance of transition between tuberous and eruptive xanthomas. The lesions wax and wane, and may even disappear without obvious change in the plasma lipoprotein pattern. Similar-appearing xan- thomas may appear on the buttocks or knees. In distinct contrast to Type II, thus far only 1 patient in 24 with Type III at the Clinical Center has noted any xanthomas before the age of twentv-five. Vascular disease. Occlusive vascular disease is a serious and frequent accompaniment of Type III. In the samples observed by us occlusive peripheral vascular disease, especially of femoral and popliteal vessels with severe claudication, has been at least as frequent as coronary atherosclerosis. Genetics The Type III lipoprotein pattern and other clini- cal features occur in families, but in contrast to Type II, the distribution of affected members is not that of a trait expressed in single dosage of an ab- normal gene. The 24 patients have come from 19 kindreds. In 3 of these, 2 siblings have been in- volved, and in 2 others, a parent and one offspring. Fourteen have been males, and 10 females. The 104 blood relatives sampled do not include complete sets of parents and siblings from all kindreds. In 3 kindreds both parents of the propositus have been sampled. One of each set was normal. One spouse had Type III, and the other 2 Type IV patterns. It is most significant that in none of the 128 relatives and patients sampled was a Type II pattern present. Conversely, complete lipoprotein analyses of 105 patients with familial Type II and 90 of their rela- tives (representing 50 kindreds) failed to reveal a single instance of Type III. These include three presumably homozygous abnormal examples of Type II. From these data it is apparent that Type III and Type II are not simply different phenotypic expres- sions of the same genotype. They are at least 2 different diseases. A few Type IV patterns (endoge- nous hyperlipemia) have been noted in the Type III relatives. As will be discussed shortly, this is a frequent pattern in Americans and may be due to factors that are not genetic. Type IV patterns appear as well in some relatives of Type II patients, and any significant differences in distribution have not developed. It is conceivable, of course, that Type III and Type IV could prove to be related, in- cluding the possible expression of the "Type XII heterozygote" sometimes as a Type IV pattern. The possibility that several mutant alleles determine this pattern also remains open. Speculation on the Mechanism Type III, with its apparent accumulation of a subclass of very-low-density lipoproteins, requires consideration of a theory of "interconversion" of plasma lipoproteins. This is inherent in the concept developed earlier in this review (Fig. l), which in turn evolved from even more theoretical considera- tions.*6**3 Briefly stated, it can be imagined that large particles containing glyceride, p `and (I lipo- proteins are degraded in stepwise fashion. As glyc- eride is sequentially removed a rise and fall of lipoproteins of progressively lower S, occurs. Meta- bolic block at any point in this transformation would theoretically cause accumulation -of very-low-density lipoproteins of a given subclass.26 Actually, such a sequence has never been convincingly demon- strated during chylomicron clearance' and a disorder of metabolism of specific subclass of pre-P lipopro- teins has likewise not been proved. The tendency today is to assume that the concentration and aver- age density of pre-/I lipoproteins is a function of the amount of circulating glyceride, which in turn is determined by the state of energy metabolism at the cellular level. It is therefore unlikely that the peculiar Type III lipoprotein anomaly is the result of a metabolic block at a discrete point in the metabolism of other- wise normal pre-/3 lipoproteins. The anomaly has not been observed in other types of hyperlipopro- teinemia, even when the plasma concentrations of pre-P lipoproteins are being driven up or down by extreme changes in diet. For example, the glyceride levels in Type IV patients may be changed from 100 to 5000 mg. per 100 ml., and there is never lipoprotein of discrete p mobility floating at density 1.006. The Type III anomaly therefore appears more likely to be secondary to the presence of an abnormal lipoprotein. 32 The Type III pattern is mimicked in the "broad beta" band seen in Tangier disease (Fig. 6); in this instance p lipoproteins and glyceride form a com- plex that lacks the added negative charge con- tributed by cr lipoproteins, normally givin:: such complexes pre-/3 mobility. Although the "floating beta" of Type III is suggestive of a deficient partic- ipation of (Y lipoprotein in formation of the glycer- ide-rich complexes, dietary studies have shown that these patients make pre-/3 lipoproteins. Because no way has yet been found to delipidate these com- plexes and still obtain accurate quantification of their content of a and p lipoprotein, some obvious questions cannot be answered. The regular a and p lipoproteins (in their usual density class) react with antiserums as though normal, and the content of a (as measured after heparin precipitation) is normal. One is left at the moment with the suggestion of a peculiar affinity of Type III /3 lipoproteins for glyc- eride, for "floating beta" is present in Type III patients even when diets have brought their glycer- ides to within normal limits. More details of these studies will be forthcoming shortly from this labora- tory; they still leave Type III as a most provocative biochemical enigma. Management The prognosis in Type III is uncertain. Four of our patients have reached the seventh decade. The majority have evidence of vascular disease, how- ever, and it seems reasonable to utilize such con- servative means as are available to lower plasma lipid concentrations. The latter are characteristically labile and responsive to diet alone. Diet. Weight reduction is one key to reduction of plasma lipids in Type III and is always effective if the patient is overweight. After ideal weight is reached it may be maintained by diets containing 40 to 50 per cent of calories from fats that are as high in polyunsaturated fats and as low in choles- terol as is practicable. If the diet is carefully fol- lowed, normal or nearly normal lipid levels can be maintained for many months and the skin lesions may completely disappear. It is hoped that the vas- cular changes may be undergoing similar beneficial involution. The "floating" p lipoprotein will persist, however, in the lipoprotein pattern. Drugs. It is premature to comment on drug ther- apy for Type III since none of the agents now being investigated have been tested adequately in this syndrome. It is our opinion that drugs tending to be more effective against hyperglyceridemia than isolated hypercholesterolemia will be more effective in Type III. Agents in this category have been dis- cussed under Type II. Perhaps the most effective and generally useful drug will prove to be chloro- phenoxyisobutyric acid. Some evidence that it may lower lipid levels in Type III has been reported.228 TYPE IV HYPERLIPOPROTEINEMIA General Dejinition. The Type IV lipoprotein pattern is the 33 hallmark of endogenous hyperlipemia. It implies that glycerides synthesized in the body, usually in the liver, have been excreted into plasma at rates exceeding the capacity for removal. The appearance of a Type IV pattern often suggests that something has gone wrong with carbohydrate metabolism or caloric balance. There is evidence that it can some- times mean inordinate emotional stress, excessive alcoholic intake or some other conflict between the patient and his environment. The development of this type of hyperlipoproteinemia is often condi- tioned by genotype. The Type IV pattern (hereafter called simply "Type IV") is a valuable indicator of metabolic imbalance; it does not describe a specific disease. It is sometimes considered synonymous with "carbohydrate-induced hyperlipemia." This is probably too narrow a concept since most patients who have this pattern on a regu- lar diet do not lose it entirely when their diet is changed so that 10 per cent of their calories come from carbohydrate, and an occasional patient does not have an abnormal increase in plasma glycer- ides when fed 80 per cent of his calories as carbo- hydrate. If freed of a generic connotation, the term "carbohydrate induced"`*3 has great virtue, how- ever, for it focuses immediate attention on abnormal glucose tolerance and a family history of diabetes, both associated with Type IV with extraordinary frequency, 15!~XW241 At the Clinical Center, where patient selection has recently been biased toward other lipoprotein anomalies, 15 of approximately 100 kindreds with proved familial hyperlipopro- teinemia have been families in which affected members had only Type IV. The pattern has been observed in several hundred plasma samples from different individuals on a regular diet. Quantitative data from a large sample representative of the entire population are not available, but Type IV may be the most common of all types of hyperlipoproteinemia. Lipoprotein pattern. As shown in Figure 12, endog- enous hyperlipemia is manifested as hyperprebeta- lipoproteinemia. This is associated with an increase in glycerides above normal limits as defined in Table 2 and is commonly accompanied by a rise in cholesterol, on the order of about 1 mg. for each 5-mg. increase in glyceride concentration. With the use of other analytical methods Type IV is de- fined by increased concentrations of S, 20-400 and sometimes S, 400 lipoproteins in the analytical ul- tracentrifuge, very-low-density lipoproteins (density less than 1.006) as separated by the preparative ultracentrifuge, lipoproteins and "endogenous" par- ticles having a2 mobility on starch-block electropho- resis or intermediate density by polyvinylpyrrol- idone fractionation, or low-density "beta" lipoproteins as measured by most precipitation technics. Con- centrations of a and p lipoproteins are usually below normal in the Type IV pattern; chylomicrons are not present. The Type IV pattern is easy to spot by the elec- trophoretogram when lipemia is not extreme (glyc- I 006-20 - 1063-O - Origin b PI.2 /j a? PROTEIN 01 PROTEIN FIGURE 12. Type IV Hyperlipoprokwwmm (See Also Table 31, - Endogenous Hyperliprmia That Is Typzally Carbohydrtie Inducible and Accompanied b Glucose Intolerance. PHLA IS noml, and hyperuncemia is common. There are many cauws, and the patlpm 1.~ often familial. erides less than 2000 mg. per 100 ml.). In severe hyperlipemia the trail of endogenous particles from the origin is marked, and interpretation difficult. Diluting plasma onefold or twofold with saline so- lution may help to distinguish any boundary be- tween pre-P lipoproteins and chylomicrons (the two together being defined as the Type V pattern). The Type I pattern is not likely to be confused with either Type IV or Type V. The ability of paper electrophoresis to separate endogenous from exogenous glyceride appears to depend on the comparaitively greater ratio of mass to charge of the dietary particles. The distinction is not absolute, and it is likely that some fed Klycer- ides may quickly appear in pre-/3 lipoproteins. Pa- tients with even the most severe endogenous hy- perlipemia, however, do not have chylomicron bands in the absence of fat in the diet. Conversely, patients with Type IV who have accumulated heavy prebeta bands on high-carbohydrate diets usually have transient hyperchylomicronemia for a few days after fat is returned to the diet. As with all other types of hyperlipoproteinemia, the pattern obtained after a week or two of a normal diet must be the index for purposes of classification. In Type IV the concentrations of /3 and a lipopro- teins vary with the degree of hyperlipemia. As heavy concentrations of glycerides are reduced, the 2 lipoprotein bands become progressively greater. The rise and fall in P-lipoprotein concentrations is an especially prominent feature of the transition in the Type IV pattern induced by dietary changes. History. Observations of patients with ketosis and uncontrolled diabetes who had hyperlipemia despite malnutrition probably provided some of the earliest stimulus to a concept of endogenous hyperlipemia. Other patients with lactescent serum that did not improve markedly upon withdrawal of fat from the diet have often been observed. With the develop- ment of the concept of carbohydrate induction all doubt has been removed that excess glycerides in the plasma are frequently not of dietary origin. In- terest in postabsorptive increases in plasma glycer- ides, or what are here called pre-/3 lipoproteins, has also been maintained by many studies, not always in agreement, of an uneven distribution of this phe- nomenon in different populations. There has been special interest in whether this abnormality is un- usually common in young and middle-aged adults who have ischemic heart disease.24'-245 What are lacking almost completely in the Iitera- ture are family data in clear-cut examples of Type IV. We have summarized elsewhere a few reports of what seem likely to have been Type IV patterns occurring in siblings or in succeeding generations.`Sg Exploration of the pedigrees of affected patients has been distressingly meager and must be improved if the meaning of the abnormal pattern is to become clearer. Clinical Manifestations It is perhaps not pertinent to discuss the clinical features of what is only a chemical indicator, but patients with Type IV tend to have certain general characteristics that can be summarized as a prelude to discussion of primary and secondary forms of this hyperlipoproteinemia. Common features of hyperlipemia. Severe Type IV from any cause may show some abnormalities already described in exogenous hyperlipemia, such as eruptive xanthomas, lipemia retinalis and foam cells in bone marrow and other reticuloendothelial tissues. The patients are also subject to the same bouts of abdominal pain - with or without chemi- cal signs of pancreatitis. Hyperuricemia occurs in hyperlipemia,2'"#2m but its frequency in Type IV has not been established. Obesity and glucose tolerance. Secondary Type IV can occur in thin, nondiabetic subjects. So can the primary form, but in patients with the latter, excessive body weight and abnormal glucose toler- ance are extremely common. In our experience the incidence of abnormal glucose tolerance in those with familial Type IV patterns exceeds 90 per cent. Age at detection. Plasma triglyceride concentra- tions rise with age (Table 2). Apparently so does the incidence of Type IV. Familial occurrence has been seen in young children, as has the secondary form, but the overall frequency in children seems much less than in adults. Afherosclerosis. All the evidence concerning the association of coronary-artery disease and pre-P lip- oproteins is retrospective; some of it is derived from plasma glyceride concentrations, and some from lipoprotein analyses. Agreement is lacking, but the consensus indicates a positive correlation at least in relatively young rnen.*42-2% Noteworthy is the finding of a much higher concentration of pre-P lipoproteins in males who had had a myocardial infarct than in controls.z43 The possibility that many pa- tients with infarcts are placed on very low-fat, and 34 hence high-carbohydrate diets may be a factor in enhancing the lipoprotein abnormalities. Our col- lection of about 200 kindreds with all types of fa- milial hyperlipoproteinemia cannot yet be properly matched for age and sex for estimation of the inci- dence of atherosclerosis. The available experience leaves the impression that Type IV is not associated with atherosclerosis to the same extent as Types II and III. Diabetes and obesity complicate such com- parisons, and only more data of the proper kind will eventually resolve this important question. Mechanisms in Type IV Two abnormalities may be operative in produc- tion of Type IV. A brief summary of them can be confined for simplicity to the interplay between adipose tissue and liver. The roles of both organs in normal endogenous glyceride metabolism have been discussed earlier and need not be repeated. Excess glyceride release. The first possible abnor- mality is release of glycerides by the liver faster than the adipose tissue can remove them. In the steady state a new equilibrium will be established at some higher plasma level of pre-P lipoproteins. Examples in which this undoubtedly occurs include excessive flow of free fatty acids to the liver or ex- tremely high-carbohydrate feeding. Both substrates in great excess will be turned out by the liver as glycerides that usually seek storage in the expansi- ble reservoirs of the adipose tissue but may linger in the plasma. To demonstrate steady or cyclical increases in flux of free fatty acids in patients is not simple; it de- pends in part on the determination of a plasma sub- stance whose concentrations are notoriously labile. Increased mobilization of fatty acids has been in- voked to explain clinical examples of endogenous hyperlipemia. The evidence is convincing only in diabetic ketosis but highly provocative in some ex- amples of prolonged emotional stress. Carbohydrate induction, on the other hand, is easy to demonstrate.`23,`24.`5" It may not, however, be assumed to be the defect in any patient with Type IV unless the patient's response is clearly outside the normal degree of carbohydrate inducibility - a rise in glycerides to levels as high as 300 to 400 mg. per 100 ml. after four to seven days on a diet containing 6 or 7 gm. of carbohydrate per kilogram of body weight.124 The issue of exaggerated carbo- hydrate inducibility may not be solved by the use of carbohydrate loads that far exceed those the pa- tient is usually eating when he has hyperlipemia. Isotopic technics recently refined to measure he- patic glyceride synthesis!l*`4r-t4!' are more likely to reveal the nature of the abnormalities in the pres- ence of conventional diets. A combination of such technics with appropriate measurements of glucose turnover, plasma and tissue insulin activities, he- patic arteriovenous differences of glyceride and perhaps opportunity to measure hepatic glyceride and lipoprotein synthesis directly may be required to understand the abnormal "diversion" of substrate to glycerides that seems certain to underlie Type IV in some patients. Decreased glyceride removal. The second possible abnormality in Type IV is a decrease in the ability of the adipose tissue to clear glyceride that is en- tering plasma at the normal rate. Two major stages at which abnormality might occur were discussed earlier. A deficiency of lipase activity at the capil- lary wall or in the cell is one, and defective glycer- ide resynthesis in the cell is the other. Although PHLA levels tend to go down on high-carbohydrate diets they are usually normal in patients with florid endogenous hyperlipemia on regular diets.184 Granting the oblique nature of available assays, a deficiency in lipoprotein lipase activity is probably not the explanation for most of Type IV. Deficien- cies of PHLA in diabetes*~~*r6 and pancreatitis"" are perhaps more usually accompanied by mixed hy- perlipemia, discussed below under Type V. Direct and correlative measurements of the capacity of adipose tissue to convert fatty acids to glyceride are not simple in man and have not been applied to the problem of hyperlipemia. Normal glucose metabo- lism is essential in this process, and the abnormali- ties of glucose tolerance in Type IV call attention to this site. Of particular interest is the question of whether insulin activity at the level of the adipose- tissue cell may not be abnormal. Briefly noted as a third possible abnormality in Type IV is the formation of anomalous lipoproteins, resistant to normal processes of removal. Possible examples of this in Type IV are presently limited to certain dysglobulinemias described below. Secondary Type IV Type IV hyperlipoproteinemia is often secondary to other known diseases. In the differential diag- nosis 2 disorders outstrip all others in frequency and in confusion, for it is sometimes very difficult to determine whether the hyperlipoproteinemia or the associated disease is the underlying abnormality. These conditions are diabetes mellitus and pancrea- titis. Diabetes meElitus.64 As a rule, diabetes that is insulin-sensitive and generally occurring in younger nonobese subjects inclined to have ketosis is con- sidered the cause, and the Type IV hyperlipopro- teinemia developing when control is poor the tran- sient resultant. In severe uncontrolled diabetes hyperlipemia is almost universal and often accom- panied by eruptive xanthomas (xanthoma diabet- icorum) and the other clinical features of any hyperlipemia. Adipose-tissue stores are briskly mobilized, plasma free fatty acid concentrations are high, and hyperlipemia develops soon after ketosis. Patients with insulin-dependent diabetes may have Type IV ,patterns or, when exogenous hyperlipemia is superimposed on the endogenous, Type V. Hy- perlipemia clears rapidly when the diabetes is brought under control. 35 By contrast exogenous insulin has little effect on Type IV patterns or carbohydrate intolerance in the patients with mild diabetes in whom ketosis never develops. Commonest in the obese and middle-aged, the hyperlipoproteinemia responds to caloric re- striction and sometimes to orally administered hy- poglycemic agents.zms25' The abnormal mechanism is unlikely to be related to excessive transport of free fatty acids. Pancreatitis and alcoholisrn.`77~1n0,252-254 The com- bination of hyperlipemia and abdominal pain has been known for nearly half a century.*55 Probably the commonest cause is pancreatitis, not infre- quently in alcoholism. In a few subjects like the patients with the Type I pattern already discussed, hyperlipemia apparently antedated the pancreatitis; in the majority pancreatitis apparently preceded the hyperlipemia. In acute pancreatitis lipoprotein pat- terns have been described that resemble Types I, IV and V.ir"v254 For the chronic hyperlipemia per- sisting after an attack, Type IV is perhaps most com- mon, but the existing data are too skimpy to support a generalization. Lipoprotein patterns are therefore of little help to the physician in the immediate di- lemma posed by a patient with abdominal pain and hyperlipemia that are both of uncommon severity. Several suggestions are offered for management: to stop all fat intake until the pain is gone; to check for the chemical signs of pancreatitis; regardless of the outcome of these tests, to be cautious about undertaking any exploratory surgery; and when the patient has recovered, not to neglect to seek evi- dence of familial hyperlipemia in his relatives. No mechanism has been proved for the hyperlipemia in pancreatitis. AI~ohol.iH~~J5fi Oral and intravenous administration of ethanol has been shown to cause endogenous hyperglyceridemia, but the plasma glyceride con- centrations are less than those frequently seen in pancreatitis. The mechanisms whereby alcohol pro- duces hyperlipemia are still being debated.`"" In most patients with glyceride levels in excess of 1000 mg. per 100 ml. there is chemical evidence of hepatocellular damage as well as pancreatitis. Low PHLA has been reported in alcoholism.*57 Zieve"" has described the occurrence of hemolytic anemia and hyperlipemia in alcoholism. This association could be due to 2 distinct effects of alcohol since anemia a d ,hypC'rlipemia may occur singly in chronid a coholism. There is a tendency to be em- r pirical about intake of alcohol in patients with Type IV. Hyperlipemia observed after a "binge" usually cannot be reproduced under experimental condi- tions, and the association is difficult to prove. Glycogen-storage disease. Mild to severe hyperli- pemia is frequently found in several types of glyco- gen-storage diseases. fi4*2=+1 We have examined plasma from 4 patients with glycogenosis (Cori Type I); the Type IV pattern was present in each. Idiopathic hypercakemia. Hyperlipidemia"`!' or hypercholesterolemia'"!' has been reported to be a frequent feature of idiopathic hypercalcemia of in- fancy; changes in lipoproteins have been reported in only a few cases.`"!' The Type IV pattern has been observed in 2 children (studied in collabora- tion with Dr. Sidney Levin). In 1, it was possible to demonstrate that hyperlipemia was associated with exaggerated carbohydrate induction. It is not known whether the children who have recovered from this disorder, which may be associated with premature atherosclerosis, have persistent lipoprotein abnor- malities. Hypothyroidism. Thyroid deficiency should al- ways be considered in a patient with the Type IV pattern.260 It can occur, but is less common than the Type II anomaly, as previously described. The rela- tion can only be proved by therapeutic trial. Nephrotic syndrome.27s64.261 The nephrotic syn- drome can provide a panoply of lipoprotein pat- terns, ranging from a discrete increase in p lipopro- teins to tremendous increases in pre-/3 lipopro- teins.i0"**"`**6* The severity is related to the degree of hypoalbuminemia. The mechanism by which lipo- protein concentrations are increased in nephrosis has been much studied but not established. Suggestions have included a block in the "conversion" of pre-P lipoproteins to p lipoproteins"" (a theoretical mechanism, also discussed under Type III) possibly associated with a decreased uptake of triglyceride by the adipose tissue.27 Others have suggested that the hyperglyceridemia is secondary to an increased out- flow of free fatty acids from adipose tissue27 or that the proteinuria provokes increased hepatic synthesis of proteins (and lipoproteins).261 Dysglobulinemia.261-270 One of the diagnoses often suggested by the sudden appearance of severe hy- perlipoproteinemia is the presence of an abnormal plasma gamma globulin as in myeloma, cryoglobu- linemia or macroglobulinemia. There have been an increasing number of reports of multiple myeloma associated with Type II or Type IV pattern. Xan- thomas, tendinous, tuberous or eruptive in form, have sometimes been present. Xanthomatosis has been reported in multiple myeloma without definite hyperlipoproteinemia. The myeloma protein has usually been a P,,~ globulin. Usually, there is a di- rect relation between the amount of abnormal pro- tein and the degree of hyperlipoproteinemia. It has been variously considered that the abnormal protein itself serves as a lipid carrier, is complexed with /3 or pre-p lipoproteins in such a way as to retard their catab- olism or serves as a cryoprecipitin for the lipoproteins. Gestational hormones. Women who are taking progesterone analogues by mouth, alone or in com- bination with estrogens, for contraception or other purposes frequently have Type IV patterns.271 The same thing is true (if pregnancy.~r~ Other disorder.9. We have also observed Type IV patterns in association with gout, Niemann - Pick disease, Gaucher's disease, total lipoatrophy, Wer- ner's syndrome and other diseases. 36 Primary Type IV Doubtless, all the diseases of which Type IV may be only an accompaniment have not been distin- guished. Neither have all the more subtle environ- mental facts that may cause sporadic appearance of the pattern in the population. Just as the incidence of primary forms of Type IV is not known, so is information meager about the number of such cases that are genetically determined. The limited expe- rience at the Clinical Center appears to be the ma- jor source of data bearing on this last question. In the two or three years in which phenotyping has become a more than casual activity, approxi- mately 90 patients have been seen at the Clinical Center whose index patterns on regular diets could be classified as primary Type IV. This excludes many more in whom such a pattern has been de- tected in connection with collaborative studies, in analysis of pedigrees of propositi of other types or in other patients in whom obvious disorders causing secondary lipoprotein abnormalities could not be excluded. For 22 of the patients with Type IV pat- terns it has been possible to sample some of their parents, siblings and children. In only 5 kindreds have all parents and all sibs been available. The choice of the relatives sampled in all 22 kindreds was based solely on availability and is therefore pre- sumably subject to only such bias as this might introduce. Familial occurrence has been detected in 15 of the kindreds. The data from this study offer some interesting insight into the familial distribu- tion of Type IV. Eighty-six parents, children or sibs of the 22 in- dex patients with primary Type IV were sampled. Thirty-four (40 per cent) had Type IV patterns. If the "propositi" are added to the sample, the per- centage of abnormal family members rises to 52 (56 in 108). The Type IV pattern was the only abnormal one observed in these kindreds. Both the mother and the father of 5 "propositi" were sampled. In 2 sets both were abnormal; 1 in each of 3 sets was normal, and the other had Type IV. Ten of a total of 21 parents sampled were abnormal. Of the sibs of the index patients 52 were sampled, and 21 (40 per cent) had Type IV. The remaining 13 relatives ex- amined were some of the children of the propositi. Three of 13 had Type IV, and all the abnormal rel- atives were over twenty years of age. Of the 10 who were normal, 7 were still less than twenty years of age. These data are too incomplete to warrant any statistical evaluation or serious genetic interpreta- tion. Proper controls are lacking, and doubtless more than one biochemical defect is represented. It does appear, however, that the incidence of the same lipoprotein abnormality is surprisingly high in relatives of patients with primary Type IV. "Vertical transmission" does occur, and only 1 parent may show the defect. The data expose a broad opportu- nity and great need for further population study. Chicd features. The lipoprotein abnormalities in most of the familial examples of Type IV described above have been mild. A few cases have been quite severe. Examples have been described elsewhere in detail.~~!' The clinical manifestations include those described above for Type IV in general. Glucose intolerance has been nearly universal, being present in 22 of 23 subjects tested so far in the overall fa- milial sample. Some, but not all, of the familial exampI& have been associated with obesity. Suh- jects with the severest hyperlipemia have usually exceeded normal weight limits. Exaggerated carbo- hydrate inducibility has been present in 6 of 7 who were adequately tested. Management. The treatment of primary Type IV presently includes weight control, avoidance of ex- cessive dietary carbohydrate and hypolipemic agents, in that order. Obese patients with Type IV who are brought to ideal weight will have a dramatic improvement in lipoprotein patterns and, usually, in glucose tolerance. If excessive carbohydrate induci- bility can actually be tested this is ideal; if not, it is reasonable to assume that most patients will have less hyperlipemia if they maintain ideal weight on a diet in which 45 to 50 per cent of calories come from fats, particularly of fairly unsaturated, low-cho- lesterol sources. This helps the patient keep down his craving for sugars and itarches. If the program outlined above does not lower plasma glycerides to below 400 mg. per 100 ml., drug therapy may be considered in addition. The orally administered antidiabetic agents may be of value*s'**5' although they do not reduce lipoprotein patterns to normal levels in some patients. Hyperli- poproteinemia is no contraindication to their use in the management of accompanying diabetes. Certain agents promoting hypolipidemi:: were described in the discussion of Type II. The 2 most effective in decreasing hyperlipemia are chlorphenoxyisobutyric acid and nicotinic acid. The precautions mentioned earlier for use of these agents apply to Type IV as well. Heparin injections may temporarily reduce Type IV hyperlipoproteinemia.`22 Its chronic use in the treatment of hyperlipemia ,is not practical. TYPEV HYPERLIPOPROTEINEMIA"~ Definition There are some patients on a regular diet who have chylomicrons and increased amounts of pre-P lipoproteins in the usual postabsorptive sample. The Type V pattern has been reserved to designate this combination of exogenous and endogenous hyperli- pemia. The legitimacy of Type V as an indicator of a specific disease, syndrome or group of disorders may be properly questioned. It may be, for exam- ple, merely a stage of Type IV, produced by sudden imposition of a heavy load of dietary fat on clearing mechanisms intolerably burdened by endogenous glycerides. This sequence has been experimentally demonstrated in more than 1 of the patients with Type IV referred to in the last section. A Type V 37 pattern may also appear when the severe exogenous hyperlipemia of Type I is clearing on the low-fat diets that cause modest carbohydrate induction in such patients. Type V could he a genotypic variant of Type IV, possibly in terms of homozygous expres- sion of the same mutant (or mutants) or because of the interaction of other genes. Type IV patterns frequently occur in the few "Type V kindreds" sam- pled thus far. The significance of Type V, therefore, is not clear, but its recognition is important. It pro- vides the therapist with a guide to dietary manage- ment and the experimentalist with a tool to deter- mine the identity or separateness of clearing mechanisms for glycerides of different origin and the factors determining the rate of removal. The incidence of patients with the Type V pattern on normal diets is not known. Twenty-one such patients have been screened at the Clinical Center. A partial sampling of the relatives of these has brought the total number of examples of Type V to 27. Our sampling of patients with acute diseases associated with hyperlipemia has been extremely limited, and the frequency of this pattern might be much greater in general hospitals. There is no body of useful information about Type V in the medical literature to warrant a draw- ing of historical perspectives. Many examples of this abnormality have undoubtedly been reported, but it is only with the increasing use of polyvinylpyrro- lidone gradients and starch-block and paper electro- phoresis that the identification of mixed hyperli- pemia has been sufficiently good to permit the study of patients with Type V patterns. Lipoprotein Pattern As shown in Figure 13, the Type V pattern is complex. Cholesterol and glyceride concentrations are elevated but in proportions too variable to be helpful in diagnosis. Simple observation of the plasma after it has sat overnight in the cold can be useful. As they do in Type I, chylomicrons appear as a cream layer at the top. However, the infrana- tant layer in Type V remains lactescent. In the polyvinylpyrrolidone gradient tube, top, intermediate and bottom layers of turbid particles are usually seen. On starch-block electrophoresis primary and secondary particles are accompanied by endogenous particles of (Y mobility.6 On the paper electrophore- togram the pre-/3 lipoproteins usually leave their trail of endogenous glyceride from the origin, but the chylomicron band is identifiable as a discrete band at the point of application.7 When the chylo- microns are few, or when both bands are very in- tense, the distinction can be hard and requires ex- perience. Concentrations of (Y and p lipoprotein tend to be low. If precipitation technics are used, there is an increase in "low-density lipoproteins," but reliable distinction of chylomicrons from endog- enous particles requires ancillary analyses. Clinical Forms and Features The patients with Type V patterns studied so far -5 PROTElN a- PROTEIN FIGURE 13. Type V Hyperlzpopoteinemia (See Also Table 3), Whwh Is a Mixture o/ Endogenovr and Exogenous HyperIipemia Uncertainly Related to Type IV. PHL.A is normal OT low; glucose tolerarue IS usually abnowd. Abdomnal pawn, eruptive xanthomar and he~atosplenomegal? are seen as in Types I and IV. The pattern 1~ often famdral. bui there are "phenocopies" sicomiq to many disorderx have many clinical features in common. Both males and females are affected. Symptoms have frequently appeared first in the late teens or the third decade. The commonest complaint is recurrent abdominal pain; this has often led to exploratory surgery, which has provided no specific diagnosis in most cases. Occasionally, the surgeons have ,found milky fluid in the peritoneal cavity. Sometimes, the ap- pearance of eruptive xanthomas has been the first abnormality noted by the patient. As in Type I, these are commonly on the knees, buttocks, shoul- ders and back, but they may be anywhere. Lipemia retinalis, foam cells in the bone marrow and hepa-. tosplenomegaly may be present. The patients are often obese and frequently have a family history of both diabetes and obesity. Evidence of accelerated atherosclerosis in either patients or their families has not been striking. Hyperuricemia may or may not be present. Pancreatic lipase and amylase may not be elevated even during bouts of abdominal pain. Glucose tolerance is almost invariably ab- normal. The contrast with Type I is important. Patients with Type I usually have symptoms beginning in infancy. There is no relation of their hyperlipemia to obesity or diabetes. The levels of PHLA in Type I are distinctly reduced. In Type V they may be somewhat low but are usually in the range of nor- mal. The fall in hyperlipemia with a fat-free diet is usually rapid in both Types I and V, but the glyc- eride level is much more dependent upon the die- tary fat intake in Type I. Type V hyperlipemia is much more affected by negative caloric balance. For hoth Types I and V one must be very cautious in refeeding fat after the patient has been on a fat-free, high-carbohydrate diet. Marked chylomicronemia and severe abdominal pain may develop in one day. AS with other types of hyperlipoproteinemia, Type V can be separated into primary and second- 38 ary forms, a distinction that for the present means simply the presence or absence of other diseases or conditions known to be associated with hyperli- pemia. For example, the 21 index patients with Type V at the National Institutes of Health included 9 with what was considered to be primary and I2 with secondary. Of the latter, 3 had uncontrolled, insulin-dependent diabetes, and 5 chronic pancrea- titis, and 4 were patients with history of recent heavy alcohol intake.176 Full lipoprotein analyses have been done infrequently in cases of hyperli- pemia due to alcoholic excess or pancreatitis, but many reports suggest that Type V patterns were present. Genetic informatiotk Of 9 "primary" examples of Type V in the Clinical Center group studied thus far (5 males and 4 females), all have proved to have at least 1 relative with hyperlipemia. It has been possible to sample all the siblings of the propositi in 4 kindreds, and both parents of the propositi in 3. The parents of the fourth propositus are dead. All the 6 living parents had abnormal lipoprotein patterns. In 2 sets both had Type IV, and in the third, the father was Type V and the mother Type IV. The 4 propositi had 13 siblings. Four were normal; 3 had Type V, and 6 Type IV. Ten children of the propositi or their affected siblings were all normal, but none were over thirteen years of age. It is possible that the expression of the disorder is delayed, as was suggested in the discussion of Type IV, and young children should be treated separately in any population analyses until this point is clarified. To sum up the lipoprotein patterns seen in 23 adult members of 4 kindreds of patients presenting Type V patterns, including the propositi, their sib- lings and parents, 8 had Type V, and 11 Type IV, and only 4 were normal. Not included in these ana- lyses is another kindred only partially studiedny in which at least 2 of 6 siblings have severe Type V patterns. The incidence of abnormal glucose toler- ance tests in members of families with Type V and hyperlipoproteinemia (either Type IV or V patterns) is 16 of 17 subjects tested at the Clinical Center. The sampling is so small as to make any genetic interpretations perilous. Nevertheless, it is worth comparing the data from the 4 completely sampled families with Type V just described with those ob- tained in relatives of propositi with Type IV de- scribed in the preceding sections. The 4 propositi with Type V had 19 adult rela- tives (their sibs or parents) who were sampled. Four (20 per cent) were normal; 4 had Type V patterns, and 11 Type IV patterns. Fourteen patients who presented Type IV patterns and were then demon- strated to have primary familial hyperlipopro- teinemia provided 54 adult sibs or parents for com- parison. Twenty-four of these (44 per cent) were normal. All the 30 abnormal relatives were Type IV. The number of affected males and females was about equal. It is obvious that if 1 of the relatives with Type IV in the "Type V kindreds" had been seen first and had become the index patient, the distribution of patterns in the relatives of the 2 sets of propositi would have been different. This does not alter a strong impression from the available data that if the Type IV and Type V patterns represent the same genotype, such influences as determine 1 phenotype or the other seem not to be randomly distributed among all the affected families. With observation of the presence, so far, of abnormalities in both parents of the kindreds with Type V, one possibility is that Type IV and Type V are the het- erozygous and homozygous genotypes, respectively, for identical. mutant alleles. It is also possible that single genes for Type I and Type IV defects inter- act to produce Type V. The only tests of these and other possible explanations must come from the careful study of more kindreds. Very probably, the ultimate solution will require the development of better tests for genotype than these 2 lipoprotein patterns can provide. Biochemical Mechanisms Type V is a combination of abnormalities whose possible bases have already been discussed under Types I and IV. Speculation that the underlying lesion is comparable to that in Type IV, with the exogenous hyperlipemia a secondary feature attribut- able to diet, has already been mentioned. Accept- ance of this theory will in part depend upon more rigorous proof than is available that chylomicrons and pre-/3 lipoproteins compete at identical sites for removal of their glyceride. Although this is probably so, more information will be needed to explain some of the family data just discussed. If Type V represents the homozygous expression of Type IV one might expect that patients with V would be much more susceptible to carbohydrate induction. Adequate comparisons have not been made. If the mechanism in Type V is related to that in Type I, one would expect lipoprotein lipase de- ficiency to be present as detected by assay of PHLA. In 7 patients with Type V so tested the ac- tivity has varied from a low value of 0.17 in 1 pa- tient to others clearly in the normal range during intake of regular diets. The test lacks the ability to discriminate clearly partial defects in enzyme activ- ity as is evident in subjects assumed to be heterozy- gous for Type I. Tests are in progress to determine whether patients with Type V might have exagger- ated de-induction of PHLA during low-fat feeding or retarded adaptation of the enzyme levels to the challenge of higher loads of dietary fat. It has been reported that patients who have insu- lin-deficient diabetes, and Type V-like patterns when they are out of control, also have low PHLA activity that returns to normal when insulin is read- ministered and clearing of hyperlipemia occurs.lrfi Patients with hyperlipemia and pancreatitis have also been reported to have decreased PHLA.179 Dia- betic animals clear ingested or infused glycerides 39 better when treated with insulin,273 and a similar phenomenon has been shown in man.J76.274 Management If one may generalize about so elusive a syn- drome, Type V appears to combine the hazards of Type I and Type IV. The treatment of Type V, beyond dietary manipulation, has not been system- atically developed. The secondary and primary forms may be approached similarly from the stand- point of diet. Weight reduction in the obese patient is an excellent way to decrease hyperlipemia. As in Type IV, Type V patterns that return completely to normal levels on severe caloric restriction again become abnormal when isocaloric feeding is re- sumed at ideal weight. The hyperlipoproteinemia is usually much improved, however. Diets that are not unusually high in fats or carbohydrates should be used in maintenance. The patients must be followed carefully, both to forestall gain in weight and to discourage intemperance in alcohol, as well as to observe the lipoprotein pattern for unanticipated changes. Whenever severe endogenous hyperlipemia develops, care must be taken to reduce both calo- ries and carbohydrate until improvement results, for a sudden switch to isocaloric high-fat diets may cause an abdominal crisis. The use of drugs has not been carefully described for Type V. The same considerations as lnentioned under Type IV obtain. OTHER ABNORMAL LIPOPROTEIN PATTERNS With the conclusion of the discussion of 5 types of lipoprotein patterns seen in patients with hyper- lipoproteinemia, it will be clear to observers of other patients that some significant and perhaps diagnostic patterns have been left out of the num- bering scheme. One may draw upon the experience with blood-clotting factors for encouragement to limit the entries in any numbered system. New patterns will inevitably be discovered as represent- ative of other syndromes, and better technics will fragment the existing Types I through V. The no- menclature will evolve as it must. Nothing has been said of chemical abnormalities in the constituent lipids of the lipoproteins, such as the quantities of phytanic acid that appear in Refsum's syndrome*75 or small increases in cerebrosides in Gaucher's dis- ease.276 Some other specific lipoprotein patterns such as the absence or deficiency of (Y and p lipoproteins have already been discussed in this review. One other set of patterns encountered quite often re- mains to be described. These are the patterns in certain liver diseases. Obstructive Liver Disease The lipoprotein patterns and serum lipids in pa- tients with either intrahepatic or extrahepatic hiliary obstruction are unique. They are characterized by marked hypercholesterolemia and hyperphospholi- pidemia. The increase in serum cholesterol is main- ly confined to free sterol whereas the esterified cholesterol content is normal or low. The increase of phospholipid is relatively greater than that of the cholesterol. More often than not the serum triglyc- eride and free fatty acid concentrations are normal, and the plasma is icteric but perfectly clear.277,27" The content of high-density lipoproteins is severely reduced or even absent in analyses made by pre- parative or analytical ultracentrifugation.26,64 There is, on the other hand, a marked increase in the li- poproteins having density between I.019 and 1.063. Comparison of the ultracentrifugal data to early work with Cohn fractionation pointed up a seeming paradox.`r* Usually, large amounts of lipid were present in the fraction IV that contains a lipoprotein in normal serum. The increased low-density lipo- proteins were also shown to be abnormal, having lower than normal ratios of protein to lipid and cholesterol to phospholipid. They also did not react to antiserums specific for /3 lipoproteins.27K We have recently found that patients with biliary cirrhosis may have tremendous increases in OL lipo- proteins on the electrophoretogram and by immuno- chemical analyses. The (Y lipoproteins behave as though antigenically identical to the normal lipo- proteins but migrate much more slowly during pa- per electrophoresis. These (Y lipoproteins, which contain more phospholipid than normal, float in the density region 1.006 to 1.063 and swell the concen- tration of low-density lipoprotein. Other patients with biliary-tract obstruction may have an elevation in true p lipoprotein that also contains more phos- pholipid than usual. The relation of bile stasis to these abnormal lipoproteins is unclear. It has been suggested that the overall lipoprotein increase may be related to the surfactant properties of bile salts favoring disruption of lipid-protein complexes and promoting a different type of stabilization of the lipids in plasma. 277 The liver probably also releases a good deal more lipid into the plasma because the normal flow of phospholipids and cholesterol in the bile is obstructed and because interruption of the enterohepatic cycle of cholesterol has altered the synthesis of lipoproteins. The hyperlipoproteinemia does vary with the extent and degree of biliary ob- struction The drug cholestyramine has been varia- bly successful in lowering the hyperlipoproteinemia of biliary obstruction.279 When severe hepatocellular damage occurs in the liver there is a lowering of all lipoprotein fractions, and patients with parenchymal liver disease may have severe total deficiency of OL lipoproteins.7x CONCLUSIONS The plasma is often the only window from which one can see the state of intracellular metabolism. The view is limited, and all ingenuity is demanded to gain the sharpest perspectives. The plasma lipo- proteins have been discussed, and the vast informa- tion about them distilled from the point of view of the clinician who seeks the most rational approach I. 2. 3. 4 5 6 7 H Y IO II 12. 13. 14. I .i. 16. 17. 18. REFERENCES 29. to patients with certain abnormalities in fat metabo- lism. The lipoproteins serve this purpose better than chemical determination of plasma lipid con- centrations alone. It is hoped that the systematic way of using some of the simpler tools for lipopro- tein analyses described will encourage wider ap- plication of lipoprotein patterns in clinical practice so that more specific diagnoses will replace certain of the time-worn cliches and more attention will be focused on the many basic problems of etiology that are still to be solved. Major emphasis has been placed on the use of lipoprotein patterns to seek genetic factors that produce altered plasma lipid levels. It has been one of the areas most neglected. If a pedigree chart appeared on the clinical record of every patient with hyperlipidemia, the number of unanswered questions might be reduced by half; certainly, the satisfaction that the physician gains from the study of patients with hyperlipidemia would be doubled. Fredrickson, D. S., and Gordon, R. S., Jr. Transport of fatty acids. Phymiol. Rev. 38:585-630, 1958. Olson, R. E., and Vester, J. W. Nutrition-endoct ine interrela- tionships in control of fat transport in man. Physiol. RPU. 40 677-733, lY60. Fritz, I. B. Factors influencing rates of long-chain fatty acid oxidation and svnthesis in mammalian systems. Phpiol. Rev. 41:52-129, 1961.' Steinberg, D. Fatty acid mobilization - mechanisms of regula- tion and metabolic consequences. In The Control of Lipid Mel&- Item. Edited by J. K. (:rant. 191 pp. New York: University Press, l!lti3. (B&hem. Society Symposium 24.) Pp. 11 l-143. Dole, V P., and Hamlin, J. T., 111. Particulate fat in lymph xnd blood. Phy.riol. Reu. 42:674-X1 I, 1962. Bierman, E. L., Porte. D., Jr., O'Hara. D. D., Schwartz. M.. and Wood, F. C., Jr. Characterization of fat particles in plasma of hyperlipemic subjects maintained on fat-free high-carbohydrate diets. I. c:Bn. Inue.~frgution 44:261-270, 1965. l.ees, R. S., and Fredrickson. D. S. Differentiation of exogenous and endogenous hyperlipemia by paper electrophoresis. I, Clin. Invr\t~gdm 44: 1968-1977, 1965. Quarfordt, S. H., and Goodman, D. S. Heterogeneity in rate of plasma I-le.trance of chylomicrons of different size. Bmchim. el bw/d/v\. or,,, 11 6:382-385, 1966 Reaven. C. M.. Hill. D. B.. Gross, R. C., and Farquhar. ,J. E. Kinetics of triglyceride turnover of very low density lipoproteins of human plasma. ,I. Ckn. Irux@t$wn 44: IR%- 1833, lY65. BeI-KStrinn. S. Metrlwllanr III hile acids. Federabon f'roc. 21 (Sup]`. I I ):.5X-32. 1962. lhtnielsson, H. Present status of research on catah)liw~ and excretion of cholesterol. ,4dvrrnce.r tn I.lptd Research 1 :335-385, 1963. Spritr, N., Ahrens, E. H., Jr., and Grundy. S. Sterol balance in man as plasma cholesterol concentrations are altered by ex- changes of dietary fats. .I. Clin. Invc.\6+ion 44: 148". 1493. l!l65. Wilson. J. D.. and Lindsey, C. A., Jr. Studies on influence of dietary cholesterol on cholesterol metabolism in isotopic steady state in man. ,I. (:lm. Inur~fi~ulirm 44: 1805. I8 14. I9fi5. Goodman. D. S. Cholesterol ester metabolism. I'lryvio!. Keu. 45:747-83!l. l!Jl%. Philltps, G. B. Phospholipid composition of human serum ,, lipoprotein fractions separated by ultracentrifugation. ./. Clin. IsuP,fl,gnlMm 38:489-493. 1959. Moser. H. W., and Emerson. K., Jr. Estimation of phospholipid phosphorus turnover time in man: studies in normal individu- als. in patients with nephrotir- syndrome and in other types of hype! lipemta. .I. f:lru. /r~~w&v~lzon 34: 1286-l YYti, 1955. Deuel, H. J., Jr, Metabolism and nutritional value of carote- noids and vitamins A. In T/w /.+idr. Vol 3. 1065 pp. New York: Interscience, 1957. P 421. Krinsky. N. I., Cornwell, D. G., and Oncley. J. L. Transport of vitamin A and carotenoids in human plasma. Arch. Htochem. 73:2X%24fi. 1958. 19 20 21 22 23 24 25 26 27. 28. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. . McCormick, E. C.. Cornwell. D. G., and Br&rr. 1. B Studies on distribution of tocopherol in human serum lipl,proteins. J. I+d Re.wnrch 1 :22 1.228. I !J6(). Olson. J. A. Biosynthesis and metabolism of caroten&ids and retinal (vitamin A). J. Lqnd Rexnrrh 5:2Hl-299. 1964. Hung, H. S., and Goodman, D. S. Vitamin A and carotcnoids. I. Intestinal ahsorption and met&&cm of %-labeled vitamin A alcohol and P-carotene in rat. /. 8w[. Chem. 240.2X39-2844, 1965. Curd, F. R. N. Some naturally occurring lipoprotein systems. In Ltptdde Chemi.r/ry. Edited by D. J Hanahan. 325 pp. New York: Wiley, 1960. P. 260. Vaodenheuwl. F. A. Origin, metabohsm, and strttcture of normal human serum lipoproteins. Cnwd. J. Biochem. U Phynol. 40:1299-1326, 1962. Salem, L. Role of long-range forces in cohesion of lipoproteins. Crud. J. Btochem. U Physrol. 40: 1287.3297, 1962. Fisher. W., and Gurin, S. Structure of lipoproteins: ccwalently bound fatty actds. Sczence 143362. 1964. Lindgren, F. T., and Nichols, A. V. Structure and function of human serum lipoproteins. In The P1a.w~ Protknr. Vol. II. Edited by F. W. Putnam. 518 pp. New York: Academic Press, 1960. P. I. Scanu, A. M. Factors affecting lipoprotein metabolism. /Idzrn,,~cs in Lipid Research 3:63-138. 1965. Cohn, E. J., et al. Preparation and properties of serum and plasma proteins. IV. System for separation into fractions of protein and lipoprotein components of biological tissues and fluids. J. Am. Chem. Sot. 68:459-475. 1946. Cornwell, D. G., and Kruger, F. A. Molecular complexes in isolation and characterization of plasma lipoproteins. /. Lrfnd Research 2:110-134, 1961. Jencks, W. P., Hyatt, M. R., Jetton, M. R., Mattingly. T W.. and Durrum. E. L. Study of serum lipoproteins in normal and atherosclerotic patients by paper electrophoreuc techniques. /. C&n. Inuesfigation 35.980-990, 1956. Kunkel, H. G., and Trautman, R. ~2~ Lipoproteins of human serum: correlatton of ultracentrifugal and electrophoretic prop- erties. /. C/m. Invcstigalron 35:641-648, 1956. Nikkil% E. Studies on lipid-protein relationships in normal and pathological sera and effect of heparin on serum lipoproteins, Scandvuu. J. Cltn. kf Lab. fnuerfzgation 5 (Supp. X):1-101, 1958. de Lalla, 0. F., and Gofman, J, W. Ultracentrifugal analysis of serum lipoproteins. In Me&d.s u/ Biochnnacal Analysu. Edited by D. Click. 528 pp. New York: Interscience, 1954. P. 459. Havel, R. J., Eder, H. A., and Bragdon, J. H. Distribution and chemical composition of ultracentrifugally separated lipopro- teins in human serum. /. Clm. Inuestig&ron 34: 1345.1353, 1955. Hjerttn, S. Calcium phosphate chromatography of normal human serum and of electrophoretically isolated serum pro- teins. Biochim. PI biophys acln 31:216-235, 1959. Carlson, L. A. Chromatographic separation of serum lipopro- teins on glass powder columns: description of method and some applications. C/in. chum. ,cctu 5:528-538. 1960. Cram&r, K. Serum &lipoprotein lipids and protein in normal subjects of different sex and age. ,4,-f/r med. Scnrrdinnu. 171:413- 421, 1962. Peeters. H., and Lrga, E. Electrochromatography of serum lipoproteins. In Profidcr o//he BiologicalFluuf~. Edwd by H. Peeters. 476 pp. Amsterdam C: Elsevier. I963. P. 134. Ewing, A. M., Freeman, N. K., and Lindgren, F. T. Analysis of human serum lipoprotein distributions. Advonrcr VI Lt~td Ke- wzrch 3:25-6 I, 1965. Fredrickson, D. S., and Levy, R. 1. Functional view of plasma lipoproteins. In Procrpdinp OJ TL Mosbuch Colloyun~m ,I/ 1Jw .Sorwty /iw Phv.nolqiccrl Chemnlry. Heidelberg: Springer (in press). Mach&oeuf, M. M. A. Recherches sur les phosphoaminollpides et les &rides du s&rum et du plasma sanguins. I. Entrainrment des phospholipides, des st&ols et des St&ides par les diverses fractions au tours du fractionnement des puGides du serum. Bull. Sot. rhim. bud. 11 :268-293. 192% But-stein, M., and Samaille, J. Sur on dorage rdpide dll chrrles- terol IiC aux u- et aux fi-IipoprotPines du r&um (:/it,. ,/urn. ,,r,n 5:609, IOfiO. Gustalson. A. New method for partial delipidir~tion 01 rerum lipoproteins. /, I.+d ,!?Pwu& 6:~ 12-5 17, 1965. van Schultz, H. E. Uber (;lykoproteine. I)u~rfwhu mrrl. U'(hnwh~. 63.1742-1752. l!l5X. Epstein, F. H.. and Hlwk, W. D. Gly co rotein content of serum 1 lipproteins. I'roc. Ser. Ex/wr. Hzol. &? hlvrf. 101 ,740.742. I%!). M;Ii.shall. W. E., and Kurnmenw. I;. A. (:.lrbohydr;tte u,,,st,t,,- ents of human serum &It]wpro~ein: gxl.~c tobe, m;mm)se. glu- cosamine and sialic acid. Arch. llrrr,bvm. 96:?71-273, lWi2. 41 47 4x. 49. 50. 51. 52. 53. 54. .i .`1 56. 57 .iR 59 fro 61 62 63 64 6.5 ti6 67 cin 6Y 71) 71 72 73 74 75 Sanbar, S. S.. and Aldupovic. P. Effect of urea on behavior of prote,n moiety of human-serum a-lipoproteins in solution. Mrorhlm. et biophys. urtrr 71:235. 1963. Scanu, A Forms of human serum high density lipoprote,n protein. /. f.lptd Rnmrrh 7:295-306, 1966. Shore, V.. and Shore, B. Protein subunit of human serum l,poprotei,l\ of density 1.125-I ,200 gram/ml. Hwcbr~s &i H,o/~kr t. Re\eorrh. Commun. 9:455-460, 1962. Levy, R. I., and Fredrickson, D. S. Heterogeneity of plasma h,gh density lipoproteins. /. Clan. Investrgafion 44:426-44 I, 1965. Scanu. A Studies on wrrformation of human seru,,, high- density lipoproteins HDL, and HDL,. I'roc Nnl. Acud. SC. 54: 16YY- 1705. lY65. hurman. R. H., Sanbar. S. S.. Alaupovic. P., Bradlord, R. H.. and Howard, R. P. Stud,ec of metabolism of radioiodinated human serum alpha lipoprotein in normal and hyperlipidemlc subjects. /, Lob. W C/rw Md. 63: lY3-204, 1964. Freeman, N. K., Lindgren. F. T., and Nichols. A. V. Chemistry of serum lipoproteins. In I'ru~~e?.~ in the Chrmtrtry of Fats and O//w /,+rd,. Vol. 6. Edited by R. T. Holman. W. 0. Lundberg .md `I`. \l,tlkir,. 364 pp. New Yrwk: Ma milln,,, 1!%3. P. 21.5. Omle!. J. 1.. Lipid protem interactions. In Bmm LipA clnd I.I~+IIIPI~I\. und thu Lrlr~~~~fy~trophrv\. Edited by J. F&h,-Pi and H. B.nwr. 2 I3 pp. Amsterdam C: Elsrvier, 1463. P. 1. Fred,-ickson, D. S. Familial high-density lipoprotein deficiency: `Lanqier drsease. In T/I< ,~letnbolrr Ham 01 lahentrd Unelra. Set- ,md'ed,t,on. Edited by J. B. Stanbury. J. B. Wyngaarden and D. 5. Fredrickson. 1434 pp. New York: McGraw-Hill, I!)66 P. 486. Rohe,m. P. S., Miller, L., and Eder. H. A. Formatm,, of- plasma ~~~~~~lr;~~,,,rr;~6~,apoprotei,, in plama. /. Biol. Chem. Phillips. G. B. Lipid composition ut human serum lipoprotein fraction with density greater than 1.210. Proc. Ser. Exprr. B,o[. @