Personalized Medicine for HLA-associated Drug-hypersensitivity Reactions

Mandvi Bharadwaj; Patricia Illing; Lyudmila Kostenko


Personalized Medicine. 2010;7(5):495-516. 

In This Article

Abstract and Introduction


Multiple genetic and nongenetic factors can modify the action of a drug, resulting in varied responses to a particular drug across different individuals. Personalized medicine incorporates the comprehensive knowledge of these factors to facilitate the selection of optimal therapy, reduce adverse drug reactions, increase patient compliance and increase the efficiency of therapy. Pharmacogenomics, which integrates the knowledge of an individual's genetic make-up for diagnostic decisions or therapeutic interventions is closely linked to personalized medicine, and is being increasingly used to prevent adverse drug reactions. There are various reports on genetic associations between particular HLA allotypes and drug hypersensitivities and the strongest associations reported thus far, are with the reverse transcriptase inhibitor, abacavir and HLA-B*5701, the gout prophylactic allopurinol and HLA-B*5801 and the antiepileptic carbamazepine and B*1502, providing a defined disease trigger and suggesting a general mechanism for these associations. Recognizing the strong HLA association, the US FDA has recommended genetic testing before starting abacavir and carbamazepine therapies. To incorporate HLA testing for other drug hypersensitivities and life-threatening reactions it is essential first to establish clear HLA associations, and second, to understand the immune-mechanism by which these drugs induce HLA-linked hypersensitivity. The latter will provide insight into the pathologic mechanisms of drug allergy allowing rational immunotherapy for these life-threatening reactions and the development of alternative drug therapies for hypersensitive patients.


Adverse drug reactions (ADRs) are one of the leading causes of morbidity and mortality in healthcare. The cost of drug-related morbidity and mortality in the USA alone has been estimated to be US$136 billion annually,[1] which is more than the total cost of cardiovascular or diabetic care. In addition to increasing the burden on the healthcare system, ADRs limit the use of key drug therapies. There are two major categories of ADRs, the predictable (Type A) ADR, associated with the pharmacological activity of the drug, and the idiosyncratic (Type B) ADR, also known as drug hypersensitivities, caused by an immune response to the drug.[2] Drug hypersensitivities constitute approximately 13% of ADRs[3] and are typically more severe in nature than the predictable drug reactions.[4] Delayed type hypersensitivity (DTH), a T-cell-mediated drug-hypersensitivity reaction, is one of the commonly reported idiosyncratic ADRs. Understanding the mechanisms behind these reactions is highly desirable for both prognosis and prevention.

Over the last few decades, attempts to understand the etiology of idiosyncratic drug reactions has led to the discovery of associations between various reactions and particular alleles of the human leukocyte antigen (HLA; summarized in Table 1). The MHC, situated on the short arm of chromosome 6, contains the genes for the HLA class I and class II (also known as MHC-I and MHC-II) and is an area of high linkage disequilibrium (LD). This region includes genes for the three classical HLA class I genes, HLA-A, -B and -C; and the three classical HLA class II genes, HLA-DR, -DP and -DQ.[5] The classical HLA genes are some of the most highly polymorphic genes within the human genome, and of these HLA-B; is the most polymorphic gene of the human genome possessing over 1000 functional alleles as listed in the IMGT/HLA database.[6,7,201]

The HLA molecules play a key role in the immune system, presenting endogenous and exogenous peptide antigens to T cells to initiate antigen specific immune responses. HLA class I and class II, generally recruit CD8+ and CD4+ T cells, respectively. Central to this function is their ability to bind peptides within their peptide-binding groove. Differences between the HLA alleles predominantly map to this location, generating differences in the peptide binding landscape that alter not only their peptide binding preferences, or peptide binding motifs, but in some cases the way in which similar alleles present the same peptides.[8] Differences in the peptides presented, the structure of the HLA molecules themselves and the orientation of peptides within the groove, collectively generate diversity in the T-cell repertoire recruited by different individuals.[8,9] The evolution of HLA diversity is thought to hinge on this phenomenon, as a heterozygous individual, possessing two different peptide binding specificities, will be able to display a greater range of peptides on their HLA. Against a background of changing pathogenic challenge this will increase the chance that they can display peptides from an invading pathogen and mount an effective immune response. Thus, alleles that occur too frequently within the population (generating more frequent homozygotes) experience negative selection, resulting in a diversification of the alleles in the population.[10] The duplication of the HLA genes, or polygeny, similarly increases the array of peptides on the cell surface.

Therefore, it seems plausible that we would observe differences in the ability of individuals to clear disease based on the array of HLA alleles they possess, hence, the antigens they can present and the T cells they recruit. However, very few HLA alleles have been associated with disease protection or susceptibility. Examples include slow progression of HIV-1 infection in individuals possessing HLA-B*5701[11] and protection against severe malaria in a West African population by the HLA-Bw53 and the HLA-DRB1*1302–HLA-DQB1*0502 haplotype.[12]

By contrast, an increasing number of HLA associations have been observed with drug-hypersensitivity reactions (summarized in Table 1) and with autoimmune diseases (summarized in Table 2). The strongest associations between specific alleles of the HLA and drug hypersensitivity discovered thus far are, abacavir hypersensitivity syndrome (AHS) and HLA-B*5701 in the Caucasian population;[13–15] HLA-B*5801 and allopurinol induced Stevens–Johnson syndrome (SJS), toxic epidermal necrolysis (TEN) and hypersensitivity syndrome (HSS) in several Asian populations;[16–18] and carbamazepine induced SJS/TEN and HLA-B*1502 in the Han Chinese and Thai populations.[19–21] All of these reactions follow the clinical pattern for T-cell mediated drug hypersensitivity with symptoms tending to occur at least 3–4 days after drug administration, improve on cessation of treatment, and occur more rapidly on re-exposure.[22–24] This progression is consistent with the proliferation of drug-reactive T cells from naive and memory populations on initial and repeat exposure, respectively. Furthermore T-cell involvement is observed within these reactions,[15,25–29] favoring the hypothesis that the hypersensitivity-associated alleles are mechanistically involved in disease progression.

It is the strength of these associations, and potential involvement in the immune reaction, that has seen these alleles suggested as candidates for pharmacogenetic testing either as predictive or diagnostic markers of true hypersensitivity reactions to the associated drugs. However, there are also several weaker HLA associations that have become apparent where the role of T cells is less well indicated in disease progression and it is clear even in the stronger associations, that other factors may be necessary to trigger hypersensitivity.

This article will explore various examples of HLA-associated drug hypersensitivity for drugs currently used in clinical practice, focusing on the strength of the associations and potential factors involved in disease manifestation. The utility of HLA typing as a diagnostic and/or predictive test for these potentially life-threatening reactions will also be discussed.