Return to the LLNA Report  

Return to the NICEATM-ICCVAM Document Search Page  

Allergic contact dermatitis is a frequent occupational health problem and, in common with other forms of allergic disease, develops in two phases. The first or induction phase is initiated when a susceptible individual encounters on the skin sufficient amounts of an inducing allergen to stimulate a primary cutaneous immune response. This results in allergic sensitization. If the now sensitized individual is subsequently exposed, at the same or a different skin site, to the same allergen then an accelerated and more aggressive secondary immune response will be provoked at the site of contact. Allergen-responsive T lymphocytes are activated in the skin at the site of contact and release cytokines and other inflammatory mediators which cause the accumulation of mononuclear cells and the inflammatory reaction that is recognized clinically as allergic contact dermatitis.

For many years the species of choice for the identification of contact allergens was the guinea pig . A variety of guinea pig test methods has been described and while these vary in detail, the principles of the assays are in each case the same, sensitizing activity being measured as a function of challenge-induced erythematous and edematous reactions in previously sensitized animals. There is no doubt that some at least of these guinea pig methods have served toxicologists well. Nevertheless, it is clear that such assays are subject to some important limitations, including the fact that the endpoint is subjective and may be difficult to measure and interpret if colored or irritant chemicals are evaluated. Moreover, some of the more sensitive guinea pig methods demand the use of adjuvant. These limitations encouraged consideration of alternative approaches.

Some ten years ago the local lymph node assay was described (Kimber et al, 1986; Kimber et al, 1989; Kimber and Basketter, 1992; Kimber et al, 1994; Kimber, 1996). This method was founded on the belief that an increasingly sophisticated appreciation of the immune system would facilitate the design of alternative methods for the identification of chemical allergens that cause adverse effects through the stimulation of specific immune responses. The local lymph node assay employs mice, the experimental species where there is the most detailed information available about the induction and regulation of immunological responses. In contrast to guinea pig test methods, the local lymph node assay identifies potential skin sensitizing chemicals as a function of events associated with the induction, rather than elicitation, phase of skin sensitization.

The induction phase of skin sensitization is characterized by the stimulation of an allergen-specific immune response in lymph nodes draining the site of exposure. Epidermal Langerhans cells (LC) recognize, internalize and process the chemical hapten associated with protein. LC are induced to migrate to draining lymph nodes. While in transit they develop into immunostimulatory dendritic cells which in the lymph nodes are able to interact with, and present antigen to, responsive T lymphocytes (Kimber and Cumberbatch, 1992; Kimber and Dearman, 1996). Immune activation in draining lymph nodes is characterized by T lymphocyte division and differentiation, the production by activated cells of cytokines and other mediators and an increase in the size, weight and cellularity of the lymph nodes. The division of activated T cells results in an increase in the number of allergen-reactive lymphocytes; this clonal expansion being the cellular basis of immunological memory and allergic sensitization. The importance of clonal expansion is reflected by the fact that the vigor of proliferative responses induced by chemicals in draining lymph nodes correlates closely with the extent to which sensitization develops (Kimber and Dearman, 1991; Kimber and Dearman, 1996).

In initial investigations several parameters of draining lymph node activation were measured following topical exposure of mice to contact allergens and to non-sensitizing chemicals. These comprised changes in lymph node weight and cellularity and lymphocyte proliferation measured as a function of radiolabelled thymidine incorporation during culture of lymph node cells (Kimber et al, 1986; Kimber and Weisenberger, 1989a; Kimber, 1989). The marker that proved to be the most sensitive and selective correlate of skin sensitizing activity was the induction of lymph node cell proliferation and subsequent investigations focused upon this. Another change introduced following these preliminary experiments was to measure the proliferative activity in situ, by intravenous injection of tritiated thymidine, rather than following culture of isolated lymph node cells ( Kimber and Weisenberger 1989b; Kimber et al, 1989). It is this version of the method that has been evaluated extensively in the context of national and international collaborative trials and which has been the subject of detailed comparisons with guinea pig tests and with human data. The results of these evaluations and comparisons will be discussed later.

A criterion of positivity was required to facilitate decisions regarding the sensitizing potential of chemicals based on activity in the local lymph node assay. The decision was made, based on extensive experience gained with the method, that a chemical should be classified as a skin sensitizer if, at one or more test concentrations, proliferative activity three-fold or greater than that measured in concurrent vehicle treated controls was induced. The validity of the use of a stimulation index of 3 for the identification of contact allergens is discussed later in this submission.

In summary, the local lymph node assay provides a novel approach to the identification of skin allergens where immunobiological events stimulated during the induction phase of skin sensitization are measured. Decisions are based upon assessment of draining lymph node cell proliferative responses - responses that are known to be essential for, and to correlate with, the induction of skin sensitization.

For practical purposes the following recommendations are made for use of the local lymph node assay:

• A chemical which, at one or more test concentrations, elicits a three-fold or greater increase in proliferative activity compared with concurrent vehicle treated controls should be classified as being a contact allergen and handled and labeled accordingly.
• Chemicals that fail at all test concentrations to elicit a positive response in the local lymph node should be classified as lacking significant skin sensitizing potential and should be handled and labeled accordingly. No further confirmation of negative results is required.

There is currently some interest in comparing and contrasting the nature of immune responses induced in mice by different types of chemical allergens. It is very important to emphasize here, however, that the proposal is that the local lymph node assay can be used to identify those chemicals that are able to cause skin sensitization. A case is not being made here for use of the local lymph node assay in the identification of any other classes of chemical allergen. Moreover, this submission is focused on the standard LLNA. Consequently, papers describing modified versions of the assay are not reviewed in this document.

The proposal is that the local lymph node assay provides an alternative method for use in the identification of skin sensitizing chemicals and for confirming that chemicals lack a significant potential to cause skin sensitization. This does not necessarily imply that in all instances the local lymph node assay should be used in place of guinea pig tests, but rather that the assay is of equal merit and may be employed as a full alternative in which positive and negative results require no further confirmation.

The local lymph node assay is not an in vitro method and as a consequence will not eliminate the use of animals in the assessment of contact sensitizing activity. It will, however, permit a reduction in the number of animals required for this purpose. It has been estimated that, in practice, on average half the number of animals required for a standard guinea pig test is needed for conduct of a local lymph node assay. Moreover, the local lymph node assay does offer a substantial refinement of the way in which animals are used for contact sensitization testing. One important point is that, unlike some of the guinea pig methods, such as the guinea pig maximization test, the local lymph node assay does not require the use of adjuvant. Furthermore, the local lymph node assay is based upon consideration of immunobiological events stimulated by chemicals during the induction phase of sensitization. Unlike guinea pig tests the local lymph node assay does not require that challenged-induced dermal hypersensitivity reactions are elicited.

Due to the fact that the local lymph node assay requires far fewer animals than needed for standard guinea pig tests, it can be conducted for approximately half the cost. The time taken for conduct of a local lymph node assay is some eight times less than that needed for a standard guinea pig method.

It is estimated currently that in excess of 25 separate laboratories world-wide are conducting the local lymph node assay.

Protocol The standard protocol described previously (Kimber and Basketter, 1992) utilizes young adult (6-16 week old) female CBA/Ca stain mice. In strain comparisons, CBA/Ca mice were found to exhibit a more marked response to contact allergens than did the other strains examined (Kimber and Weisenberger, 1989a). However, female CBA/J and CBA/JHsd strain mice are also acceptable for use in the assay as, in several interlaboratory validation studies, they display responses comparable with those of CBA/Ca strain mice (Kimber et al, 1995; Loveless et al, 1996). Mice are housed under standard conditions, individually or by treatment group, in plastic shoe box type cages for the duration of the study. Food and tap water are provided ad libitum. Control of bias is addressed by randomization of mice prior to initiation of the study.

A sample protocol is provided in Appendix D.

Dose selection No additional animals are used for dose range finding. The current practice is to select at least three consecutive concentrations from the following range: 100, 50, 25, 10, 5, 2.5, 1, 0.5, 0.25 and 0.1% (w/v). The selection is made to provide the highest possible test concentration, limited by compatibility with the vehicle chosen (and the suitability of the resultant preparation for unoccluded dermal application ), while avoiding dermal trauma or systemic toxicity. The test chemical is dissolved in an appropriate vehicle. Vehicle selection is important and a variety of organic solvents is suitable. The following are recommended, in order of preference: acetone-olive oil (4:1) (AOO), acetone, dimethylformamide, methyl ethyl ketone, propylene glycol and dimethylsulfoxide (Kimber and Basketter, 1992). While aqueous vehicles are not recommended, aqueous and aqueous-organic mixtures such as 3:1 acetone:water have been used successfully.

Control Materials The current OECD positive control sensitizers hexyl cinnamic aldehyde, 2-mercaptobenzothiazole and benzocaine have each been evaluated in the local lymph node assay. Results with these positive controls in the local lymph node assay met or exceeded the minimum acceptable standard set forth by the OECD (Basketter et al, 1993). The strong sensitizer 2,4-dinitrochlorobenzene (DNCB) may be used as a positive control as it has produced consistent responses in the LLNA, including when tested in two recent international interlaboratory trials (Kimber et al, 1995; Loveless et al, 1996). Currently, there are no recommended negative controls for the LLNA as is the case with the reference guinea pig methods. However, methyl salicylate, tested at 1, 2.5, 5, 10 and 20% (w/v) in acetone:olive oil (4:1) (Kimber et al, 1995; Kimber et al, 1998) and para-aminobenzoic acid tested at 0.5, 1, 2.5, 5 and 10% (w/v) in acetone:olive oil (Loveless et al, 1996) have been used successfully as negative control chemicals in interlaboratory validation studies. In common with other skin sensitization tests, a control substance for irritation has not been defined for the LLNA.

In cases where individual mice are being used for determining the mean dpm value for an experimental group, statistical analysis may be performed. The value of statistical analyses, either alone or in conjunction with the three-fold stimulation index, has not yet been established and is still the subject of investigations. Where isotope incorporation is determined for individual mice, a mean dpm value ± standard error of the mean (SEM) is calculated for each experimental group. A stimulation index is derived for each experimental group by dividing the mean dpm of that group by the mean dpm of the vehicle-control group.

One approach to the development of statistical methods that may prove of value in the local lymph node assay is as follows. For statistical analyses, the mean dpm values for each treatment group and the vehicle control group are initially normalized by obtaining their log value. Bartlett's test (Bartlett, 1937) is then used to examine the data for homogeneity of the within-chemical treatment variance. When analysis of variance reveals significant differences in parametric data, experimental groups are compared with vehicle-treated controls using Dunnett's t test (Dunnett, 1955). For non-parametric data, a Kurskal-Wallis test (Kruskal and Wallis, 1952) followed by Dunn's multiple comparison procedure (Dunn, 1964) is used. Groups differing from vehicle-treated controls at the level of P<0.05 are considered significantly different. Alternately, if Bartlett's test for homogeneity of variance is not significant, comparisons with the control group (and other specific, pairwise comparisons of groups) are based on the least significant difference criterion. If Bartlett's test is significant, these comparisons are based on Wilcoxon's rank sum test.

In addition, an estimate of the test material concentration required to produce a stimulation index of 3 (EC₃) can be calculated using fitted quadratic regression analyses. An advantage of the EC₃ calculation is that data from the entire dose response curve are used to produce a single value of intrinsic potency (Loveless et al, 1996). The EC₃ value can then be used to rank order the skin sensitizing potential of chemicals. Stronger sensitizers such as DNCB and oxazolone have lower EC₃ values than more moderate sensitizers such as hexyl cinnamic aldehyde and eugenol (Loveless et al, 1996). Dose response analyses in the local lymph node assay, combined with the mathematical derivation of the lowest test concentration of a chemical required for a defined stimulation index, such as the EC₃, provides a convenient, reliable and realistic approach to evaluation of relative potency (Kimber and Basketer, 1997).

An examination of the application of statistical analyses to the local lymph node assay is continuing. At present, it is not clear whether, or in what way, an evaluation of statistical significance would add value to the interpretation of the local lymph node assay. This, together with consideration of EC₃ values for measurement of relative potency are areas of investigation that may pay dividends in the future, but which are not currently part of the standard protocol.

Summary of control data The recommended positive control material, hexyl cinnamic aldehyde (HCA), was tested independently by five laboratories over a dose range of 2.5, 5.0, 10.0, 25.0, and 50% (w/v) in AOO (Loveless et al., 1996). All five correctly identified HCA as a contact allergen. Four of the five laboratories found the lowest concentration to produce an SI of 3 or greater was 10%. The fifth laboratory reported an SI of 2.5 for this concentration. Calculations of the EC₃ for HCA ranged from 7.0 to 8.4%. DNCB was tested in two separate trials by the same five laboratories at concentrations of 0.01, 0.025, 0.05,0.1, and 0.25% (w/v) in AOO. EC₃ calculations for DNCB from both trials ranged from 0.03 to 0.09%.

Recently the stability with time of responses induced in the local lymph node assay by HCA has been evaluated in a single laboratory. Over a ten month period HCA elicited very similar EC₃ values in the local lymph node assay (Dearman et al, 1998). These issues are discussed further in Section D below.

Two of the interlaboratory evaluations of the LLNA were carried out under conditions where all details of the test materials and test conditions were not known to the participating laboratories. In the first of these studies, 20 substances were coded and supplied to each of 4 laboratories (Basketter et al, 1991). In a subsequent study, the chemical names were given, but no advice on dose/vehicle selection was provided (Scholes et al, 1992). The results from both of these investigations demonstrated a high degree of interlaboratory agreement. It is interesting to compare these results with those from unblinded interlaboratory studies of the GPMT and the Buehler test (Robinson et al, 1990; Andersen et al, 1985). In these instances, relatively poor interlaboratory reproducibility was achieved, which is in sharp contrast to experience with LLNA.

There are considerable data on intralaboratory reproducibility of the LLNA, some of which has been published (Basketter et al, 1996; Kimber et al, 1998) and some of which is based on unpublished individual laboratory experience. Table 1 summarizes the information on this topic.

Although it is not the aim within the current validation to examine assessment of relative skin sensitizing potency, it is possible to derive such information from the LLNA (Basketter et al, 1996; Kimber and Basketter, 1997). For this, the estimated concentration of the test chemical which is sufficient to cause a 3-fold stimulation (EC₃) is determined by interpolation of the dose response data. What precise value this may have for risk assessment is currently the subject of various pieces of work (eg Basketter et al, 1996; Kimber and Basketter, 1997; Basketter, 1998). However, the approach taken also allows better comparison of individual LLNA results. Examples of this type of data are contained in Table 2.

ND = No data

The first collaborative LLNA validation trial involved four independent laboratories in the UK which evaluated the same batch of eight chemicals, using the same protocol, vehicles and test concentrations. Each laboratory identified 2,4-dinitrochlorobenzene (DNCB), formalin, eugenol, isoeugenol, paraphenylenediamine (p-PDA), and potassium dichromate as positive with benzocaine and methyl salicylate as negatives. With the exception of isoeugenol, no significant differences between the laboratories were found with respect to the characteristics of dose-response curves (Kimber et al, 1991).

The same four laboratories participated in a more extensive evaluation involving 25 chemicals (Basketter et al, 1991). Of the 25 chemicals, equivalent predictions of sensitizing potential were made for 18 chemicals by all laboratories. An additional five chemicals were identified as potential sensitizers in the LLNA by two or three laboratories. Three of these subsequently gave a positive response in laboratories which initially failed to detect them when retested under identical or altered conditions (e.g. higher concentration, different vehicle). It should be noted that these investigations were conducted prior to publication of the definitive LLNA protocol.

For the final phase of this national collaboration, nine chemicals were evaluated and each laboratory independently selected the test concentrations and vehicles (Scholes et al, 1992). One modification that all laboratories employed was applying chemicals topically for three consecutive days and then terminating the experiment five days after the initiation of exposure, rather than four days. Chemicals were evaluated at three concentrations which were chosen independently by each laboratory with regard to potential toxicity. The choice of vehicle was based upon solubility and viscosity. For eight chemicals, equivalent predictions were made by all laboratories and by three of the four laboratories for the remaining chemical. Identical vehicles and concentrations were selected independently by all laboratories for two chemicals and by three laboratories for six chemicals. In those cases where different concentrations or vehicles were chosen, equivalent predictions (positive or negative LLNA results) were still made.

To determine what effect minor protocol modifications would have on the predictive value of the test, the LLNA was evaluated in an international study by five independent laboratories, two of which had participated in the UK national validation exercise. Modifications to the standard protocol included exposure of mice for four, rather than three, consecutive days, removal of auricular lymph nodes four rather than five days after study initiation, the use of an alternative isotope and analysis of lymph nodes from individual mice to allow for statistical evaluation proposed (reviewed in Gerberick et al, 1992; Ladics et al, 1995).

In the first phase of this international validation, two skin sensitizers, DNCB and potassium dichromate, and one non-sensitizer, methyl salicylate, were evaluated (Kimber et al, 1995). In the LLNA, the criteria for a positive result is a three-fold or greater stimulation of proliferative activity relative to vehicle controls. In the laboratories analyzing nodes from individual mice, a positive result was also defined, for the purpose of this investigation, as treatment groups differing from vehicle treated controls at a predetermined level of statistical significance (p<0.05 or p<0.01 depending upon the statistical method employed). By either criterion, and regardless of the protocol utilized, all five laboratories identified the two known sensitizers as being positive in the LLNA. Estimates of the test concentration required to yield a stimulation index of three (EC₃) were very similar for all laboratories for both chemicals. Using the stimulation index criteria, all laboratories reported a negative finding for methyl salicylate at all concentrations tested. Two of the three laboratories evaluating nodes from individual mice did detect a statistically significant increase in radioisotope incorporation at the highest of the five concentrations tested (20%).

In the second phase of the international collaborative trial, the sensitivity and selectivity of the assay were examined further by analysis of six additional chemicals: hexylcinnamic aldehyde (HCA), oxazolone, isoeugenol, eugenol, sodium lauryl sulphate (SLS), and para-aminobenzoic acid (pABA) (Loveless et al, 1996). The last two are considered to be non-sensitizing chemicals, while the remainder exhibit skin sensitizing potential to varying extents, with HCA being one of three chemicals recommended by the OECD for use as positive controls in skin sensitization studies (OECD, 1993). All laboratories retested DNCB under the conditions employed in phase I of the trial (Kimber et al, 1995) to provide information on the temporal stability of assay data. All five laboratories identified as positive the five moderate to strong sensitizers (DNCB, HCA, oxazolone, isoeugenol and eugenol). SLS, considered to be a non-sensitizing skin irritant, also induced a positive response in the assay. pABA, a non-sensitizing chemical, was negative in each laboratory.

Oxazolone was clearly the most potent sensitizer evaluated in Phase II, with predicted EC₃ values ranging from 0.0007 to 0.0026%. This chemical highlights the benefit of utilizing the entire dose response curve for predicting the concentration required for a SI of 3, since four of the five laboratories recorded stimulation indices of above three at the lowest concentration tested. It also demonstrates that determination of an EC₃ maybe useful in assessing the relative sensitizing potency of a class of chemicals. Results with HCA, eugenol, isoeugenol and pABA were similar to published LLNA results (Basketter et al, 1993; Basketter and Scholes, 1992; Basketter et al, 1994 ).

The results of Phase I and II provide strong support that the incorporation of minor procedural modifications did not affect the performance of the LLNA. In that regard applying a test chemical for either three or four consecutive days, with removal of lymph nodes five or four days, respectively, after the initiation of treatment, did not change the ability of the assay to detect skin allergens. Three consecutive daily exposures to a chemical is therefore considered sufficient for the purpose of the identification of potential skin sensitization hazard.

Concerning the choice of isotope utilized for detection of proliferation, there was no difference in the ability of ³HTdR or ¹²⁵IUdR to identify correctly the chemicals evaluated in this study. Either isotope can be used in the LLNA (Ladics et al, 1995; Kimber et al, 1995; Loveless et al, 1996).

An important modification assessed during Phase I and II of this international validation study was the analysis of proliferation within lymph nodes of individual mice as opposed to lymph nodes pooled for each experimental group. In the majority of cases, the lowest concentration yielding a positive response was identical by either method of analysis.

One objective of Phase II was to examine inter-experimental variability by evaluating DNCB twice. Three of the five laboratories obtained identical results to the first study (Kimber et al, 1995). Depending upon which of the criteria were used, the other two participating laboratories had either identical inter-experimental results or were within one adjacent concentration level. Therefore, the intralaboratory inter-experimental variability was very low.

The overall conclusion from this and the previous phase of the validation study (Kimber et al, 1995) is that five independent laboratories, despite the use of procedural modifications and different methods for data analysis, successfully and consistently employed the LLNA to reach identical conclusions on the sensitizing potential of nine chemicals.

The most recent interlaboratory validation study involved the same five laboratories working in collaboration with the US FDA. In this study (Kimber et al, 1998), a small series of chemicals used in topical drug products was examined. Again there was very close agreement between laboratories, with all five identifying correctly benzoyl peroxide, hydroquinone, penicillin G and methyl salicylate. Streptomycin sulfate induced equivocal responses, insofar as this material provoked a positive LLNA response in only one of the five laboratories, and then only at the highest concentration tested. Ethylenediamine dihydrochloride was uniformly negative. Collectively these data serve to confirm that the LLNA is sufficiently robust to yield equivalent results when performed independently in separate laboratories. The data indicate also that the LLNA is of value in assessing the skin sensitization potential of topical medicaments.

A total of 7 laboratories have been involved in interlaboratory validations of the LLNA. The results of the work have appeared in the several associated publications (Kimber et al, 1991; Basketter et al, 1991; Scholes et al, 1992; Kimber et al, 1995; Loveless et al, 1996; Kimber et al, 1998). This work has involved investigation of more than 40 different chemicals.

An overview of the time frame for the development and validation of the LLNA is displayed in Figure 1 (adapted from Chamberlain and Basketter, 1996). Information on consistency/performance over time has been given earlier in this section.

A variety of guinea pig tests has been developed for evaluation of the skin sensitizing potential of chemicals. Among those most widely applied are the guinea pig maximization test (GPMT) (Magnusson and Kligman, 1969,1970) and the occluded patch test of Buehler (Buehler, 1965, 1985; Robinson et al, 1990). These two assays are the preferred guinea pig sensitization tests outlined in the OECD 406 guideline for skin sensitization.

Figure 1. LLNA Timeline

LLNA Timeline Graphic

The GPMT used for comparisons with LLNA results is based on and similar to that described by Magnusson and Kligman (1970) which uses Freund’s adjuvant. Albino Dunkin-Hartley guinea pigs, weighing approximately 350g at the start of each study, are used. Preliminary irritation tests are carried out to determine the concentrations of the test substances suitable for induction of sensitization and for challenge. Guinea pigs are then treated by a series of six intradermal injections in the shoulder region to induce sensitization. After 6-8 days, sensitization is boosted by a 48 hr occluded patch placed over the injection site. Twelve to fourteen days later, the animals are challenged on one flank by a 24 hr occluded patch at the maximum non-irritant concentration. Challenge sites are scored for erythema (scale 0-3) and edema 24 and 48 hr after removal of the patches. The EC guidelines state that a material is positive if the incidence is ³ 30% (European Communities, 1993).

The standard Buehler test (BT) protocol uses an occluded topical patch technique for the induction and elicitation of contact sensitization (Buehler, 1965, 1985; Robinson et al, 1990). The procedure calls for 20 animals in the test (sensitized) group, 10 naive (control) animals for challenge, and 10 separate naive control animals for rechallenge. For induction, a single dorsal site is used for three 6 hour induction patches (applied occluded once per week to the same pre-shaven induction site on the dorsal surface of the test animals). Following a two week rest period, the test and non-induced control animals receive 6 hour challenge patches at a naive skin site for the primary challenge. The same test animals and additional new control animals can be rechallenged by this procedure 7-15 days after primary challenge at any remaining naive skin sites. Reactions are graded for erythema 24 and 48 hours after patch removal, according to a 5 point grading scale. The grades "1", "2" and "3" denote increasing severity of erythema with grades ³ "1" considered positive. The EC guidelines state that a material is positive if the incidence is ³ 15% (European Communities, 1993).

In addition to comparison of the LLNA with guinea pig sensitization test data, the LLNA has also been compared with human data (Basketter et al, 1994; Basketter et al, 1996). Specifically, the LLNA has been compared with the human maximization test (HMT) (Kligman, 1966a,b,c). This method was specifically designed to provide a rigorous assessment of the skin sensitization potential of chemicals in humans. In principle, a group of 25 subjects is subjected to 48 hour occlusive patch treatments with as high a concentration of test chemical as possible. This treatment is repeated five times over a two week period. If the substance is not sufficiently irritating, the irritancy is enhanced by prior treatment of the site for 24 hours with sodium lauryl sulfate prior to each 48 hour patch. The extent of sensitization in the panel is assessed by 48 hour treatments on a slightly irritated skin site using the maximum non-irritant concentration of the test substance. The challenge sites are scored at 48 hours and 96 hours post-application. In essence, this procedure can provide a stringent assessment of intrinsic sensitization hazard and its relative potency.

To define the role of the LLNA in predictive testing, results from the assay have been compared with predictions from guinea pig and human tests. In some instances, the LLNA results and the reference results (guinea pig or human) are presented together. In other cases, LLNA studies have been conducted with chemicals whose sensitization potential, or lack thereof, are well known. Basketter and Scholes (1992) investigated the correlation between results in the LLNA and those derived from the GPMT for materials that covered a range of chemical types and levels of skin sensitization potency. Kimber et al (1990) reported comparative analyses in which 24 chemicals, of previously unknown contact sensitizing potential, were evaluated in both the local lymph node assay and the occluded patch test of Buehler. The data reported demonstrate that the local lymph node assay identified successfully those chemicals that were classified as moderate or strong skin sensitizers in the Buehler test. Basketter et al (1991) evaluated the performance of the LLNA with 25 chemicals for which guinea pig maximization test or Buehler occluded patch test data were available. The 25 chemicals included preservatives, perfume ingredients, surfactants, plastics/resin chemicals and oil additives. A high level of agreement between the results of local lymph node assays and guinea pig test data was found.

As stated above, an essential point of comparison for the LLNA is with human data. Basketter et al (1994 and 1996) compared human maximization tests results with those obtained with the LLNA for the same 38 chemicals. The former being a rigorous assessment of the sensitization potential of chemicals in humans. The authors reported that the LLNA identifies those chemicals that are significant human contact allergens and that the specificity of the assay is good. A comprehensive review of published and unpublished LLNA data is given in Appendix A.

The predictive power of the LLNA in comparison to standard guinea pig methods is given in Appendix B. This type of information has been reviewed in detail in a recent paper (Basketter et al, 1996). While it is clear that the LLNA is not quite as sensitive as the GPMT, it is of similar or greater sensitivity than the Buehler test. It is important to note that this comparison is only true where the guinea pig tests have been conducted to the very highest standards. In terms of predictive identification of important skin sensitizers, the LLNA is at least as sensitive as, and much more reliable than, current guinea pig tests. Of the 130 chemicals tested in one of the reference guinea pig tests, approximately 88% gave the same result in the LLNA and the guinea pig tests. An overview of this information is contained in the 2 X 2 contingency table (Table 3).

Return to the LLNA Report  

Return to the NICEATM-ICCVAM Document Search Page