The present invention provides a method of predicting the risk of a patient for developing adverse drug reactions, particularly SJS or TEN. It was discovered that an HLA-B allele, HLA-B* 1502, is associated with SJS/TEN that is induced by a variety of drugs. The correlation with HLA-B* 1502 is most significant for carbamazepine-induced SJS/TEN, wherein all the patients tested have the HLA-B* 1502 allele. In addition, another HLA-B allele, HLA-B*5801, is particularly associated with SJS/TEN induced by allopurinol. Milder cutaneous reactions, such as maculopapular rash, erythema multiforme (EM), urticaria, and fixed drug eruption, are particularly associated with a third allele, HLA-B *4601. For any of the alleles, genetic markers (e.g., HLA markers, microsatellite, or single nucleotide polymorphism markers) located between DRB1 and HLA-A region of the specific HLA-B haplotype can also be used for the test.

Description of the Invention

BACKGROUND OF THE INVENTION

Adverse drug reactions (ADRs) are a major clinical problem. According to a widely cited meta-analysis, ADRs was ranked between the fourth and sixth most common cause of death (Lazarou et al., 1998). In particular, potentially serious cutaneous ADRs account for about 2-3% of all hospital admissions (Bigby et al., 1986). Although drug eruptions may be mild to moderate, such as maculopapular rash, erythema multiforme (EM), urticaria, and fixed drug eruption, more severe cutaneous ADRs are life-threatening and frequently result in death, such as Stevens-Johnson syndrome (SJS) and toxic epidermal necrolysis (TEN; Lyell's syndrome).

SJS is characterized by high fever, malaise, and a rapidly developing blistering exanthema of macules and target-like lesions accompanied by mucosal involvement. TEN has similar presentations with an even more extensive skin detachment and a higher mortality rate (30 to 40%). Although the incidence of SJS/TEN is rare with an annual estimated incidence of 3-5 per million people, these conditions can kill or severely disable previously otherwise healthy people (Roujeau and Stem, 1994). The severity of the condition has prompted pharmaceutical companies to withdraw a few newly released drugs.

Almost all SJS/TEN cases are caused by drugs, most commonly sulfonamides, anticonvulsants, allopurinol, nonsteroidal anti-inflammatory drugs (NSAIDs), and antimalarials (Roujeau et al., 1995). In Taiwan, anticonvulsants (carbamazepine, phenyloin and phenobarbital), and allopurinol are the most common drugs causing SJS/TEN. Other medications such as NSAID and antibiotics are also noted to cause severe ADR.

Recent developments of pharmacogenomics have implied that the susceptibility to ADRs is associated with genetic variants. A successful example of application of pharmacogenomic study to prevent drug-induced side effect is genotyping thiopurine methyltransferase (TPMT) before prescribing azathioprine, a drug for rheumatologic or cancer diseases (Yates et al., 1997). An individual's genomic polymorphism(s) of TPMT can cause enzyme deficiency and slow metabolizing rate, resulting in leukocytopenia. This kind of molecular diagnostics certified by CLIA (Clinical Laboratory Improvement Amendments) is now offered by reference laboratories in the USA (Prometheus Laboratory Inc.; Genaissance Pharmaceutical) and Europe. Although susceptibility to SJS/TEN on certain drugs is thought to be genetically determined (Gennis M A, 1991; Edwards S G, 1999), the responsible genetic factors have yet to be identified and currently there is no method clinically useful that can be used to predict who will develop SJS/TEN or to which drugs.

SUMMARY OF THE INVENTION

The present invention provides a method of predicting the risk of a patient for developing adverse drug reactions, particularly SJS or TEN. It was discovered that an HLA-B allele, HLA-B* 1502, is associated with SJS/TEN that is induced by a variety of drugs. The correlation with HLA-B* 1502 is most significant for carbamazepine-induced SJS/TEN, wherein all the patients tested have the HLA-B*1502 allele. In addition, another HLA-B allele, HLA-B*5801, is particularly associated with SJS/TEN induced by allopurinol. Milder cutaneous reactions induced by carbamazepine, such as maculopapular rash, erythema multiforme (EM), urticaria, and fixed drug eruption, are particularly associated with a third allele, HLA-B*4601.

Accordingly, the present application provides a method of assessing the risk of a patient for developing a cutaneous adverse drug reaction in response to a drug, comprising performing HLA typing using a biological sample from the patient. The drug is preferably selected from the group consisting of carbamazepine, allopurinol, phenyloin, sulfasalazine, amoxicillin, ibuprofen and ketoprofen. Alternatively, the drug is preferably not a nonsteroidal anti-inflammatory drug.

Specifically, one aspect of the present invention provides a method of assessing the risk of a patient for developing an adverse drug reaction in response to a drug, comprising determining the presence of an HLA-B allele selected from the group consisting of HLA-B* 1502, HLA-B*5801 and HLA-B*4601, wherein the presence of the HLA-B allele is indicative of a risk for an adverse drug reaction. The drug is preferably selected from the group consisting of carbamazepine, allopurinol, phenyloin, sulfasalazine, amoxicillin, ibuprofen and ketoprofen. Most preferably, the drug is carbamazepine or allopurinol.

The adverse drug reaction is preferably a cutaneous adverse drug reaction, such as Stevens-Johnson syndrome or toxic epidermal necrolysis. In a preferred embodiment, the drug is carbamazepine, and the allele is HLA-B* 1502. In another preferred embodiment, the allele HLA-B*5801 is used to predict the risk for cutaneous ADR, such as Stevens-Johnson syndrome or toxic epidermal necrolysis in response to allopurinol. Other subtypes of HLA-B15, B58 or B46 can also be used to predict the risk for ADR instead of HLA-B*1502, HLA-B*5801 or HLA-B*4601, such as HLA-B*1503.

The allele can be detected by using any method known in the art. For example, the presence of the allele can be determined by using an oligonucleotide that specifically hybridizes with the nucleic acid coding for the allele. Preferably, the DNA prepared from the peripheral blood of the patient is employed in the determination. The allele can also be detected by, for example, serological or microcytotoxicity methods.

The presence of the allele of interest can also be determined by detecting an equivalent genetic marker of the allele, which is a genetic marker that is linked to the allele. For example, the HLA-markers of HLA-B B*1502 haplotype include, without being limited to, DRB1*1202, Cw*0801, Cw*0806, A*1101, and MICA*019. The HLA-markers of the HLA-B*5801 haplotype comprise, for example, Cw*0302. In other words, the presence of the HLA-B*1502, 5801 or 4601 haplotype, rather than the alleles per se, is indicative of a risk for adverse drug reactions.

Another aspect of the present invention provides a method of pharmacogenomics profiling comprising determining the presence of at least one HLA-B allele selected from the group consisting of HLA-B* 1502, HLA-B *5801, and HLA-B*4601. Preferably, the presence of at least two alleles selected from the group is determined, such as HLA-B*1502 and HLA-B*5801. More preferably, the presence of all three alleles is determined. The method can optionally comprise the determination of other genetic factors. Those other genetic factors may be associated with the predisposition for any disease or medical condition, including adverse drug reactions. For example, these other genetic factors may be selected from the group consisting of thiopurine methyltransferase and the genes for the long-QT syndrome.

Further provided is a method of screening and/or identifying medicines that can be used to treat drug-induced SJS/TEN by using HLA-B* 1502, 5801 or 4601 as a target in drug development. For example, cells expressing any of the alleles can be contacted with medicine candidates, and the candidates that bind to the allele are likely to inhibit the function of the allele. The efficacy of the allele-binding candidate in treating drug induced SJS/TEN can then be further tested.

DETAILED DESCRIPTION OF THE INVENTION

The present invention provides a method of predicting the risk of a patient for developing adverse drug reactions, particularly SJS or TEN. It was discovered that an HLA-B allele, HLA-B* 1502, is associated with SJS/TEN that is induced by a variety of drugs. The correlation with HLA-B* 1502 is most significant for carbamazepine-induced SJS/TEN, wherein all the patients tested have the HLA-B* 1502 allele. In addition, another HLA-B allele, HLA-B*5801, is particularly associated with SJS/TEN induced by allopurinol. Milder cutaneous reactions associated with carbamazepine, such as maculopapular rash, erythema multiforme (EM), urticaria, and fixed drug eruption, are particularly associated with a third allele, HLA-B*4601.

Methods

There is evidence that the pathogenesis of several similar multisystem drug hypersensitivity reactions involves MHC-restricted presentation of drug or drug metabolites, with direct binding of these non-peptide antigens to MHC molecules or haptenation to endogenous proteins before T cell activation (Svensson et al., 2000). Skin-infiltrating CD8+ cytotoxic T cells were found to be dominant in the bullous reactions such as SJS/TEN (Hari et al., 2001), whereas CD4+ helper T cells were characteristic of milder cutaneous adverse drug reactions, such as maculopapular rash (Pichler et al., 1997). Since the major histocompatibility complex (MHC) is known to be important in determining T-cell mediated immune responsiveness to the antigens, such as metabolites of drugs, we evaluated whether the alleles of the major histocompatibility complex are associated with drug-induced SJS/TEN.

We performed HLA typing on patients with adverse drug reactions (Examples 1-2). The results indicate that HLA-B*1502 was present in 42 of 42 (100%) SJS/TEN patients who received carbamazepine. The allele was also found in 17 of 53 (32%) SJS/TEN-patients who received other drugs (8 phenyloin, 2 allopurinol, 2 amoxicillin, 1 sulfasalazine, 1 ketoprofen, 1 Ibuprofen, and 2 unknown drugs). Particularly, eight of 17 patients (47.05%) who developed SJS/TEN after taking phenyloin also carry the HLA-B* 1502 allele. On the other hand, the allele was only found in 4.1% (3/73) of the carbamazepine-tolerant group, 0% (0/32) of the phenyloin-tolerant group, 6.3% (9/142) of the patients who had milder adverse drug reactions other than SJS, and 5.3% (5/94) of the general population. By using the tolerant group as control, the odds ratio, sensitivity, specificity, positive predictive value, and negative predictive value for B* 1502 associated carbamazepine-induced SJS/TEN, were 1712, 100%, 95.89%, 96.0%, and 100%, respectively. For B*1502 associated phenyloin-induced SJS/TEN, the odds ratio, sensitivity, specificity, positive predictive value, and negative predictive value were 58, 47%, 100%, 100%, and 65.35%, respectively. Accordingly, the presence of this HLA-B allele can be used in the identification of high-risk patients for drug-induced SJS/TEN, particularly carbamazepine- and phenyloin-induced SJS/TEN.

The mild adverse reactions induced by carbamazepine appear to be associated with another allele, HLA-B*4601. Thus, 10 out of 16 (62.5%) of the patients with these milder reactions to carbamazepine had HLA-B*4601. In contrast, the allele was only found in 26% (19/73) of the carbamazepine-tolerant group. The odds ratio for B*4601 associated carbamazepine-induced milder cutaneous ADRs was 4.73. Consequently, HLA-B*4601 can be used in the risk assessment for mild cutaneous ADR induced by carbamazepine.

A third HLA-B allele, HLA-B*5801, was found in 17 out of 17 (100%) patients with SJS/TEN or hypersensitivity patients who received allopurinol, but only 18% in the general population. The odds ratio, sensitivity, specificity, positive predictive value, and negative predictive value for B*5801 associated allopurinol-induced SJS/TEN or hypersensitivity were 155, 100%, 82%, 84.7%, and 100%, respectively. HLA-B*5801 can thus be used to predict the risk for adverse drug reactions in response to allopurinol.

Accordingly, the present invention provides a method of assessing the risk of a patient for developing an adverse drug reaction after taking a drug, comprising determining the presence of an HLA-B allele selected from the group consisting of HLA-B* 1502, HLA-B*5801 and HLA-B*4601, wherein the presence of the HLA-B allele is indicative of a risk for an adverse drug reaction. In a preferred embodiment, HLA-B* 1502 is used to predict the risk for SJS/TEN, particularly carbamazepine-induced SJS/TEN.

Carbamazepine, also known as Tegretol, Tegol, G-32883, Biston, Calepsin, Carbatrol, Epitol, Finlepsin, Sirtal, Stazepine, Telesmin, or Timonil, is an aromatic anticonvulsant. Other aromatic anticonvulsants, including phenyloin (Dilantin) and phenobarbital, cause similar adverse drug reactions as carbamazepine. Therefore, HLA-B* 1502 can be employed to assess the risk for adverse drug reactions to these other aromatic anticonvulsants as well. The aromatic anticonvulsants for which HLA-B* 1502 can be used as a risk factor also include metabolites and derivatives of carbamazepine, phenyloin or phenobarbital. Metabolites of these drugs are known in the art (see, e.g., Gennis et al., 1991; Leeder, 1998; Naisbitt et al., 2003), such as carbamazepine-10, 11 epoxide, carbamazepine-10, 11-diol, carbamazepine 2,3-diol, dihydro carbamazepine, carbamazepine catechol and carbamazepine o-quinone, p-hydroxy phenyloin, phenyloin dihydrodiol, phenyloin catechol, phenyloin methylcatechol, and phenyloin o-quinone.

In another preferred embodiment, HLA-B*5801 is used to predict the risk for allopurinol-induced SJS/TEN. Allopurinol is a drug for hyperuricemia and chronic gout. As is with the other drugs, HLA-B*5801 can be used to assess the risk of the metabolites and derivatives of allopurinol as well.

Other subtypes of the HLA-B15, B58 or B46 locus may also be predispositive for cutaneous adverse drug reactions, particularly when the patient is of a different ethnic origin. Such subtype variation has been observed in the art. For example, ankylosing spondylitis is strongly associated with HLA-B27. Many alleles, or subtypes, have been reported for HLA-B27, such as B*2701-B*2723. These subtypes are distributed in different areas in the world and many are associated with ankylosing spondylitis (Khan, 2000; Feltkamp et al., 2001). We contemplate that HLA-B15, B58 or B46 are associated with cutaneous ADR as described herein, and other subtypes of HLA-B15, B58 or B46 may also be used for risk assessment instead of HLA-B*1502, 5801 or 4601, e.g., HLA-B*1503.

Furthermore, it should be noted that in addition to the specific HLA alleles per se, genetic markers that are linked to each of the specific alleles can be used to predict the corresponding ADR risk as well. This is because genetic markers near the HLA allele of interest tend to co-segregate, or show a linkage disequilibrium, with the allele of interest. Consequently, the presence of these markers (equivalent genetic markers) is indicative of the presence of the allele of interest, which, in turn, is indicative of a risk for ADR. As shown in Example 3, the HLA-B*1502 haplotype includes HLA markers such as DRB1*1202, Cw*0801, Cw*0806, A*1101, and MICA*019. The HLA markers of the HLA-B*5801 haplotype include, for example, Cw*0302.

The equivalent genetic marker can be any marker, including HLA markers, microsatellites, and single nucleotide polymorphism (SNP) markers. Preferably, the useful genetic markers are about 200 kb from the HLA-B locus or less. More preferably, the markers are about 100 kb, 80 kb, 60 kb, 40 kb, or 20 kb from HLA-B locus or less. Of particular interest are the markers located between DRB1 and the HLA-A region of a specific HLA-B haplotype.

The HLA alleles can be detected by using any method known in the art. Preferably, genomic DNA is hybridized to a probe that is specific for the allele of interest. The probe may be labeled for direct detection, or contacted by a second, detectable molecule that specifically binds to the probe. Alternatively, cDNA, RNA, or protein product of the allele can be detected. For example, serotyping or microcytotoxity methods can be used to determine the protein product of the allele. Similarly, the equivalent genetic markers can be detected by any methods known in the art.

To further increase the accuracy of risk prediction, the allele of interest and/or its equivalent genetic marker may be determined along with the genetic markers of accessory molecules and co-stimulatory molecules which are involved in the interaction between antigen-presenting cell and T-cell interaction. These genetic markers include microsatellite, and single nucleotide polymorphism (SNP) markers. The accessory and co-stimulatory molecules include cell surface molecules (e.g., CD80, CD86, CD28, CD4, CD8, T cell receptor (TCR), ICAM-1, CD11a, CD58, CD2, etc.), and inflammatory or pro-inflammatory cytokines, chemokines (e.g., TNF-.alpha.), and mediators (e.g., complements, apoptosis proteins, enzymes, extracellular matrix components, etc.). Also of interest are genetic markers of drug metabolizing enzymes which are involved in the bioactivation and detoxification of drugs. These genetic markers also include microsatellite and SNP markers. The drug metabolizing enzymes include phase I enzymes (e.g., cytochrome P450 superfamily etc.), and phase II enzymes (e.g., microsomal epoxide hydrolase, arylamine N-acetyltransferase, UDP-glucuronosyl-transferase, etc.).

The present invention further provides a method for pharmacogenomic profiling. Thus, a panel of genetic factors is determined for a given individual, and each genetic factor is associated with the predisposition for a disease or medical condition, including adverse drug reactions. In the present method, the panel of genetic factors includes at least one allele selected from the group consisting of HLA-B*1502, 5801 and 4601. The panel preferably includes at least two alleles, and most preferably all three alleles, from the group. In addition to HLA-B* 1502, 5801 and/or 4601, the panel may include any other known genetic factors, such as thiopurine methyltransferase and the genes for the long-QT syndrome. The genetic markers for accessory molecules, co-stimulatory molecules and/or drug metabolizing enzymes described above can also be included.

Further provided is a method of screening and/or identifying medicines that can be used to treat drug-induced SJS/TEN by using HLA-B* 1502, 5801 or 4601 as a target in drug development. For example, cells expressing any of the alleles can be contacted with medicine candidates, and the candidates that bind to the allele are likely to inhibit the expression and/or function of the allele. The efficacy of the candidate in treating drug induced SJS/TEN can then be further tested.

Kits

Another aspect of the present invention provides a kit comprising the means for detecting at least one allele selected from the group consisting of HLA-B* 1502, 5801 and 4601. The means is preferably a probe that binds specifically to the allele, and the kit preferably also contains detection reagents for the probe. The probe is preferably an oligonucleotide. The kit may further comprise tools and/or reagents for collecting biological samples from patients, as well as those for preparing genomic DNA, cDNA, RNA or the allele protein from the samples. For example, PCR primers for amplifying the relevant regions of the genomic DNA may be included.

The kit preferably comprises means for detecting at least two alleles selected from the group consisting of HLA-B * 1502, 5801 and 4601. Optionally, the kit may comprise means for detecting other genetic factors as well, particularly those useful in pharmacogenomic profiling. A preferred example is thiopurine methyltransferase.

Thus, in a preferred embodiment, the kit may comprise probes for detecting all three alleles, HLA-B* 1502, 5801 and 4601. More preferably, the kit further comprises the PCR primers suitable for each and every allele as well.
 

Claim 1 of 8 Claims

1. A method of assessing a risk of a human patient for developing an adverse drug reaction in response to a drug, comprising detecting the presence of an HLA allele in a sample obtained from the patient, wherein the HLA allele is HLA-B*5801, and correlating the presence of the HLA allele in the sample with an increased risk for an adverse drug reaction in the patient in response to the drug, wherein the adverse drug reaction is Stevens-Johnson syndrome or toxic epidermal necrolysis, and the drug is an allopurinol.
 

____________________________________________
If you want to learn more about this patent, please go directly to the U.S. Patent and Trademark Office Web site to access the full patent.