Pharmacogenomic Biomarkers in Allergy and Immunology Practice

Pharmacogenomics can be a useful tool to help personalize approaches to treatment and identify patients who may be a responder or non-responder to a drug, identify patients at risk for both on-target and off-target adverse drug reactions, including drug hypersensitivity, and determine appropriate dosing of a drug. The US Food and Drug Administration (FDA) has developed a summary table listing therapeutic products from Drugs@FDA with pharmacogenomic information found in the drug labeling.(1) This listing is extensive covering over 400 drugs and associated biomarkers for therapeutic areas of hematology-oncology, infectious disease, rheumatology, pulmonary, psychiatry, neurology, endocrinology dermatology, and others. In this Editorial, we have focused on drugs either relevant to the direct use by Allergy Immunology (AI) specialists or their inpatient or outpatient consultative practice. We have summarized the key biomarkers, their FDA specific label notation, and their potential implications, particularly in relation to actionable pharmacogenetic variants where there has been published and/or evidence-based guidance.(Table 1). One major area of pharmacogenetics is in oncology where genetic variations of the cancer tissue may also affect drug responses. This is beyond the scope of this editorial but a few relevant examples as relates to AI specialists are discussed later. Pharmacogenetics and effects of inhaled corticosteroids in asthma patients is an area of interest to AI specialists but was not listed in the FDA summary table. Polymorphisms in a number of genes including TBX21, FCER2, and CRHR1 have been shown to effect responses to inhaled corticosteroids but with variable results and have been recently reviewed elsewhere.(2)

Table 1.

FDA Recommendations and Comments for Pharmacogenomic Biomarkers in Allergy & Immunology Therapeutics(1)^*

Medication Type	Medication(s)	Biomarker Variant	Recommendation/Comment
Anti-inflammatory Immunosuppressant	azathioprine	TPMT NUDT15	Risk for bone marrow toxicity in TPMT and NUDT15 poor and intermediate metabolizers. Dose reduction of 30–80% or up to 10 fold reduction is recommended in poor or intermediate metabolizers (combination of TMPT and NUDT15 to be considered) (7)
	dapsone	G6PD	G6PD deficient patients at higher risk for hemolytic anemia and methemoglobinemia. Screen prior to prescription Avoid in patients with congenital (CYB5R3 variants) or idiopathic methemoglobinemia
	hydroxychloroquine	G6PD	G6PD deficient patients evidence suggests they are not at risk for hemolysis and that pre-screening not indicated in standard risk patients
	mycophenolic acid	HPRT1	Should be avoided in patients with rare hereditary deficiency of HGPRT such as Lesch-Nyhan and Kelley-Seegmiller syndromes
	sulfasalazine	NAT	Slow acetylators may have increased adverse effects including allergic reactions to sulfasalazine. Pre-prescription testing not currently recommended.
		G6PD	Considered low risk and screening not recommended
NSAIDS	Celecoxib Meloxicam Piroxicam Flurbiprofen Ibuprofen	CYP2C9	If CYP2C9 status known, start 50% of dose in poor metabolizers (and consider alternatives not metabolized by CYP2C9 if relevant (naproxen, aspirin, ketorolac, sulindac)(7)
Proton pump inhibitors	esomeprazole, lansoprazole, omeprazole, pantoprazole	CYP2C19	Recent study suggests that children with normal or poor CYP2C19 metabolism may be at higher risk for Clostridioides difficile infection when PPI dosed with antibiotics (7)
			Consider increased starting dose for CYP2C19 ultrarapid metabolizer (*17/*17) or for rapid metabolizer (*1/*17) initiating treatment for Helicobacter pylori (7)
Long-acting inhalational beta-agonists	formoterol,	CYP2C19,CYP2D6	Unknown whether reduction in metabolism through CYP2C19 of CYP2D6 leads to increased systemic exposure or adverse effects
	arformoterol	CYP2D6,CYP2C19, UGT1A1	Reduced CYP2D6 activity did not impact systemic exposure
Anti-depressant	Doxepin Amitriptyline	CYP2D6	Consider measurement of plasma levels for patient on higher sustained doses Poor metabolizers of CYP2D6 and CYP2C19 may have higher levels When genetic testing available consider 50% reduction in dose for poor or intermediate metabolizers. Consider anti-depressants not metabolized by CYP2D6 for ultrarapid metabolizers (7)
Anti-emetic	Ondansetron	CYP2D6	No difference in starting dose for poor, intermediate and normal metabolizers of CYP2D6 For Ultrarapid metabolizers due to increased metabolism and reduced clinical response in some reports recommended to choose alternate drug not metabolized by CYP2D6 (e.g. granisetron).
	Dolasetron, palonosetron,	CYP2D6	No difference in PK in those with reduced CYP2D6 metabolism.(7)
Tyrosine kinase inhibitor	midostaurin	KIT	Responses seen in advanced systemic mastocytosis with or without KIT D816V mutation
	imatinib	KIT	Indicated for aggressive systemic mastocytosis without KIT D816V mutation
Antiretrovirals	Abacavir	HLA-B*57:01	Screen all patients prior to initiation of abacavir (3,7) Discontinue abacavir if any symptoms of hypersensitivity. HLA allele carriage rate: European (5–8%);Sub-saharan African (<1%);African American (2–3%); Southeast Asian (<1%) (5)
	Efavirenz	CYP2B6	Poor metabolizers at risk for QTc prolongation and central nervous system toxicity. Consider dose adjustment to 400 mg od for intermediate metabolizers (*1/*6,*1/*18,*4/*6,*4/*18,*6/*22,*18/*22) and 200 mg od for poor metabolizers(*6/*6,*18/*18,*6/*18). Consider therapeutic drug monitoring.(7)
	Dolutegravir	UGTI1A1	UGT1A1 poor metabolizers had 46% increased in AUC. Unclear if associated with toxicity outcomes. No dose adjustment currently recommended.
	Raltegravir	UGT1A1	No evidence for clinically significant alteration in PK. No dose adjustment recommended.
Antimicrobials	Trimethoprim-sulfamethoxazole	G6PD	Low risk of hemolysis at standard dosing. Routine screening not recommended.
	Voriconazole	CYP2C19	Poor metabolizers (*2/*2,*2/*3,*3/*3) have 4 x higher AUC than extensive metabolizers. For ultrarapid metabolizers (*17/*17 and rapid metabolizers (*1/*17) consider alternative antifungal. not dependent on CYP2C19 metabolism. For poor metabolizers consider alternative antifungal or initiating at a lower dose with therapeutic drug monitoring (TDM). TDM should also be considered when there is altered hepatic or renal function or drug interaction concerns.(7)
Opioids	Codeine and Tramadol	CYP2D6	Ultrarapid metabolizers (active metabolite morphine) at risk for oversedation and toxicity and caution that can transfer from mother to infant in breast milk (7)
			For ultrarapid metabolizers consider alternative not metabolized by CYP2D6 (hydromorphone, fentanyl) and avoid codeine, tramadol and oxycodone. For poor metabolizers consider alternatives due to potential lack of efficacy (avoid codeine, tramadol and oxycodone).(7)
Anti-coagulants	Warfarin	CYP2C9 VKORC1 CYP4F2	Use algorithm that incorporates patient characteristics (age, sex, race, interacting drugs etc.) with genetic variables as described in product monograph and consensus guidelines.(7)
Drugs associated with hypersensitivity reactions	Carbamazepine	HLA-B*15:02	Associated with carbamazepine Stevens-Johnson syndrome and toxic epidermal necrolysis (SJS/TEN) in high risk populations (South and Southeast Asians) should be screened for HLA-B*15:02 prior to carbamazepine prescription. Monitoring for hypersensitivity should occur as usual due to lack of 100% negative predictive value. (5,7). HLA allele carriage rate:Southeast Asian (10–15%);European (<0.1%); African (<0.1%)
	Carbamazepine	HLA-A*31:01	Associated with carbamazepine associated benign exanthem (maculopapular eruption (MPE)), drug reaction with eosinophilia and systemic symptoms, SJS/TEN and drug-induced liver injury in populations where it is prevalent (European, Japanese, Korean). Limitation of use is low positive predictive value and lack of 100% negative predictive value. In HLA-A*31:01 positive patient consider use of alternative agent and also avoidance of other structurally aromatic amine anticonvulsants (eg oxcarbazepine) (5,7) HLA allele carriage rate: European (≤6%); Japanese/South Korean (10–15%); South Central Asians (4%); African (≤2 %)
	Oxcarbazepine	HLA-B*15:02	There is a greater risk of oxcarbamazepine SJS/TEN and use is not recommended (7)
	Phenytoin	HLA-B*15:02	Associated with phenytoin SJS/TEN in high risk groups (South and Southeast Asian). Not recommended to use phenytoin or fosphenytoin. Caution should be applied using other structurally related anticonvulstants such as eslicarbamazepine, phenobarbital and lamotrigine) (7)
		CYP2C9	CYP2C9*3 is associated with phenytoin MPE, DRESS and SJS/TEN. For HLA-B*15:02 negative intermediate and poor metabolizers consideration for TDM guided dosage and dose reduction of 25–50% for intermediate (*1/*3, *2/*2) and poor metabolizers (*2/*3,*3/*3) (7)

^*in FDA product monograph, italicized text is from guideline documents or literature.

TPMT: thiopurine S-methyltransferase

NUDT15: nucleoside diphosphate-linked moiety X motif 15

G6PD: glucose-6 phosphate dehydrogenase

HGPRT: hypoxanthine-guanine phosphoribosyl-transferase

NAT: N-acetyl transferase

CYP: cytochrome P450

HLA: human leukocyte antigen

Every drug has on-target and off-target pharmacological effects. Variance in genes related to the on-target effect of the drug can affect efficacy or treatment outcome and can also influence adverse drug reactions predictable based on the pharmacological action of the drug. These genetic variants often involve either the cytochrome P450 (CYP P450) family of isoenzymes that are involved in the metabolism of many drugs, or drug transporters but can also involve variation in other enzymes or the on-target receptor where a drug is known to act. The off-target pharmacological effects of drugs can also vary significantly between individuals based on genetic variation and typically lead to variance in adverse drug reactions. The specific off-target effects are not known for all drugs and therefore are often not predictable based on the pharmacological action of the drug. They include cellular toxicities, pseudoallergic reactions, and both antibody and T-cell mediated hypersensitivity reactions. The recently described MRGPRX2 receptor on mast cells is a receptor ligand for many small molecules such as fluoroquinolones, vancomycin, neuromuscular blocking agents and opioids with the MRGPRX2 receptor can lead to non-IgE mediated mast cell activation resulting in pseudoallergic drug reactions.

A paradigm shift in personalized medicine has been the association between severe T-cell mediated drug hypersensitivity reactions and variation in class I human leukocyte antigens (HLA). The strong association between HLA-B*57:01 and abacavir hypersensitivity led to a clinical trial that provided strong evidence for the guideline-based routine use of HLA-B*57:01 screening prior to abacavir use, and another study in African Americans that supported the generalizability of the 100% negative predictive value of HLA-B*57:01 for abacavir hypersensitivity across race.(3) HLA-B*57:01 screening has now drastically reduced reporting of abacavir hypersensitivity. Over 10 years of post-marketing surveillance of HLA-B*57:01 testing supports that when the test is ordered and acted upon prior to abacavir prescription, true immunologically mediated abacavir hypersensitivity is prevented.(4) Strong associations between HLA-B*15:02 and the most life-threatening of T-cell mediated drug reactions, Stevens-Johnson syndrome and toxic epidermal necrolysis (SJS/TEN), in Southeast Asia has led to screening programs in Southeast Asia that have reduced the prevalence of carbamazepine SJS/TEN.(5) In 2007 the FDA issued a label change to recommend HLA-B*15:02 screening prior to carbamazepine prescription in these high risk populations. Associations between HLA-B*15:02 and oxcarbazepine SJS/TEN are also well described and included in the drug label. Weaker associations between HLA-B*15:02 and SJS/TEN associated with anticonvulsants such as phenytoin provide a mechanism to risk-stratify patients prior to anticonvulsant prescription. HLA-A*31:01 has been associated with several phenotypes of carbamazepine associated T-cell mediated reactions including maculopapular exanthem (MPE), drug reaction with eosinophilia and systemic symptoms (DRESS), and SJS/TEN in Europeans, Japanese and Koreans and provides helpful information in the decision of avoidance of carbamazepine in favor of other drugs. For phenytoin the poor metabolizing genotype CYP2C9*3 has been associated with a wide spectrum of drug hypersensitivity reactions including MPE, DRESS and SJS/TEN and provides additional information for recommended dose reduction in those not carrying a high-risk HLA class I allele. There are several other HLA associations of note that although not in the FDA label are of potential clinical relevance and utility. This includes: HLA-B*13:01 and dapsone hypersensitivity currently being implemented in China as a pre-prescription strategy for leprosy treatment; HLA-A*32:01 as a pre-emptive screening strategy and diagnostic aid for vancomycin DRESS; and HLA-B*58:01 which explains almost 100% of allopurinol-induced SJS/TEN and DRESS in Southeast Asians and approximately 60% in Europeans and other populations.(5, 6)

Several anti-inflammatory or immunosuppressant medications that may be used in refractory urticaria, urticarial vasculitis or autoimmune complications of common variable immunodeficiency have pharmacogenomic factors that may lead to increased adverse effects. Azathioprine is a thiopurine compound that is metabolized by the enzyme thiopurine S-methyltransferase (TPMT). Polymorphisms in the TPMT gene can lead to a marked decrease in enzymatic activity and increased risk for drug-induced leukopenia. In the product label the FDA recommends that clinicians consider testing for TPMT and nucleoside diphosphate-linked moiety X motif 15 (NUDT15) in patients who develop severe bone marrow toxicities. NUDT15 inactivates thiopurine metabolites and patients with loss of activity variants in NUDT15 are more susceptible to thiopurine toxicity. When TPMT and NUDT15 genetic information is available, guidance for dose reduction has been provided by several guidelines including the clinical pharmacogenetics implementation consortium (CPIC). Dapsone may cause a dose-related hemolytic anemia and methemoglobinemia. Patients with glucose-6 phosphate dehydrogenase (G6PD) deficiency are at increased risk for both of these complications. Although noted in the label for other drugs including hydroxychloroquine, sulfasalazine and trimethoprim-sulfamethoxazole, G6PD screening prior to prescription of these drugs is not recommended based on current evidence due to lower risk for these drugs at routine dosing. Mycophenolic acid should be avoided in patients with rare hereditary deficiency of hypoxanthine-guanine phosphoribosyl-transferase (HGPRT) such as Lesch-Nyhan and Kelley-Seegmiller syndromes because it may cause an exacerbation of disease symptoms characterized by the overproduction and accumulation of uric acid. Acetylation of sulfasalazine is mediated by N-acetyl transferase (NAT) and polymorphisms in the gene encoding for NAT can result in slow or fast acetylator phenotypes. Although some reports have associated toxicity with slow acetylator phenotypes, genotyping prior to administration is not recommended by any guideline. NSAIDS such as celecoxib, meloxicam, flurbiprofen and ibuprofen are metabolized by CYP2C9 and dosage reduction or alternatives can be considered in poor metabolizers. Other drugs such as warfarin although infrequently prescribed by AI specialists are associated with pharmacogenetic variation (Table 1) in CYP2C9, VKORC1 and CYP4F2 that can affect time to therapeutic anticoagulation and bleeding risk. Consensus guidelines and product monograph information exist to guide dosing.

Proton pump inhibitors (PPIs) such as esomeprazole, lansoprazole, omeprazole, and pantoprazole are metabolized by CYP2C19. PPIs may be used by AI specialists to treat gastroesophageal reflux, laryngopharyngeal reflux, or eosinophilic esophagitis. Patients who are rapid or ultrarapid metabolizers of PPIs are at risk for treatment failure particularly in the context of combination therapy for Helicobacter pylori infection and drugs known to induce CYP2C19 (e.g. St. John’s Wort) can also reduce the effectiveness of these PPIs. Concomitant administration of these PPIs may also lead to increased blood levels of tacrolimus. In addition, polymorphisms in CYP2C19 can lead to a poor metabolizer phenotype leading to increased levels of PPIs and potentially heightened risk of Clostridioides difficile as reported in children.(7) The long-acting beta-agonist formoterol and its enantiomer arformoterol are also metabolized by CYP2C19, CYP2D6 and UGT1A1 however in patients with reduction in these enzyme activities there is no impact on systemic exposure. Doxepin and amitriptyline may be used in chronic urticaria or itch due to their potent antihistamine activity. Poor metabolizers of CYP2D6 and CYP2C19 may have higher doxepin and amitriptyline levels. Although some data suggests that the pharmacokinetics of ondansetron (which may be used to treat reactions from food protein induced enterocolitis syndrome), do not differ between poor and extensive metabolizers of CYP2D6 there are reports of decreased antiemetic effect when used for chemotherapy associated or postoperative nausea and vomiting in CYP2D6 ultrarapid metabolizers.

Voriconazole metabolized by CYP2C19 has been used in allergic bronchopulmonary aspergillosis and invasive fungal infections and poor metabolizers have a four-fold increase in AUC compared with extensive metabolizers. Guidance exists for use of alternative agents, therapeutic drug monitoring and dose alteration in the setting of ultrarapid, rapid or poor metabolizing status.(7)

Midostaurin is a small molecule that inhibits multiple kinases including nonmutant and mutant KIT D816V and has been shown effective in advanced systemic mastocytosis. Patients with the KIT D816V mutation, and those who were negative or unknown for this mutation had responses to midostaurin (63% and 44% respectively).(8) In contrast, imatinib is approved for aggressive systemic mastocytosis in patients who are negative for KIT D816V mutation. Imatinib is also approved for treatment of hypereosinophilic syndrome (HES) and/or chronic eosinophilic leukemia. It is typically very effective in patients with the FIP1L1-PDGFRα fusion kinase and can be started at a lower dose (100 mg). It is also indicated for HES patients where the fusion protein is negative or unknown with a higher initial dose, typically 400 mg daily.

Patients often present to AI specialists with complaints of cutaneous reactions to opioids such as morphine that are consistent with non-IgE mediated mast cell activation. Currently it is not known whether ultarapid metabolizers of CYP2D6 which more efficiently metabolizes codeine and other opioids such as tramadol to active morphine metabolites are at higher risk of non-IgE mediated mast cell activation. However, risk of sedation and respiratory depression is increased in ultrarapid metabolizers and alternatives such as fentanyl or hydrocodone should be considered. Poor metabolizers may have decreased efficacy associated with codeine, tramadol and oxycodone.

The efficacy, safety and effectiveness of AI practice will increasingly benefit from personalized medicine approaches. The intent of this editorial was to draw attention to genetic variation in drugs relevant to AI practice and information in the FDA label which may help guide clinical practice through awareness without providing clear recommendations for the use of specific testing. There have been major success stories in pharmacogenetics such as routine guideline-based use of HLA-B*57:01 screening to prevent abacavir hypersensitivity, however, currently there are still many limitations to the routine application of pharmacogenetic information to clinical practice. Although pharmacogenetic testing, particularly done as panels or single allele HLA typing is not costly, in many cases there is a disconnect and lack of consensus between the FDA label, evidence, availability, and reimbursement of pharmacogenetic testing which represents a hurdle in its widespread implementation. Ongoing initiatives are now driving effective ways to implement pharmacogenetic testing in a real-world setting that includes electronic health record clinical decision support and integration and these efforts will be vital to clinical translation.(9, 10)