PHARMACOGENOMICS OF OPIOID RESPONSE IN EUROPEAN ADVANCED CANCER PATIENTS

Background: Strong opioids, which are drugs in the third step of the World Health Organization’s analgesic ladder, are the standard of care for treating pain in advanced cancer patients. Unfortunately, a significant minority of them do not benefit from analgesic therapy, or experience several side effects, such as nausea and vomiting. Genetics might play a role in predisposing patients to a good or poor response to opioids. To investigate this issue, two different genome-wide association studies (GWAS) were conducted to identify genetic variants modulating opioid efficacy and toxicity. Methods: We genotyped over two thousand European patients with advanced cancer, treated with morphine, buprenorphine, fentanyl, and oxycodone. Data about toxicity (nausea-vomiting score, NVS) and efficacy (pain intensity, PI) and other relevant clinical information were collected. We carried out two whole genome regression models (using REGENIE software) to test the association between genotypes and the two opioid response phenotypes, both defined as numerical scores, one measuring patient pain intensity and the other nausea/vomiting intensity. To understand the functional role of the variants significantly associated with opioid response, several “in silico” post-GWAS analyses were conducted. Colocalization between pain intensity phenotype and gene expression was explored, as well as the tissue-type enrichment for our significant variants, using stratified linkage disequilibrium score regression (sLDSC). Finally, Mendelian randomization, using our pain intensity GWAS datasets, was performed to investigate causal relationships between putative pain risk factors, such as inflammation, sleep alteration, and psychiatric conditions, and cancer pain intensity. Results: For the opioid-induced toxicity phenotype, 65 variants associated with NVS (at Pvalue < 1.0 × 10−5) were found. Among these, 14 variants on chromosome 2 mapped in intronic regions of NPAS2 gene, which encodes a circadian transcription factor. This gene is reported to influence the sleep-awake cycles; several studies suggested a reciprocal link between the circadian and pain systems, and the modulation of opioid drugs. Also, some of these variants were previously identified as splicing quantitative trait loci of the NPAS2 (HGNC:7895) gene. The second GWAS for opioid efficacy identified five variants (rs6062363, rs6062365, rs13043326, rs6089804, and rs1806952) whose minor alleles negatively correlated with pain intensity, at genome-wide statistically significance level (P-value < 5.0 × 10-8 ). This negative correlation indicates that subjects homozygous, for the minor alleles or heterozygous, experienced lower pain intensity than patients homozygous for the major allele. These variants mapped to a non-coding region of chromosome 20 downstream PCMTD2 (HGNC:15882) gene, and less than 200kbp far from OPRL1 gene. While information on the role of PCMTD2 is scarce, OPRL1 encodes the opioid related nociceptin receptor 1, belonging to the opioid receptor family. The identified polymorphisms were reported to be modulators of the expression of both PCMTD2 and OPRL1 genes, in eQTLGen database. Also, variants in the same chromosomal region were recently reported to be significantly associated with pain intensity in a GWAS conducted in subjects with different chronic pain conditions. The colocalization analyses showed that there was low evidence for shared variants regulating both pain intensity and PCMTD2 gene expression. sLDSC indicated that variants associated with pain intensity were enriched in liver and central nervous system. Finally, the Mendelian Randomization did not support causal relationships for the analyzed traits. Conclusions: The results obtained support the hypothesis of a genetic role in modulating the opioid response of advanced cancer patients. However, in vivo functional analyses are needed to understand the biological mechanism underlying the observed association. Also, validation in an independent but homogeneous patient series would be advisable. Those herein reported are the first two GWAS for opioid response, both in terms of efficacy and toxicity, so far performed in cancer patients. They represent the starting point for further pharmacogenomics studies of opioid therapy in larger sample sizes, which might be more representative of the whole population. This study provides new insights into the genetic factors influencing pain intensity suggesting new potential markers of opioid response, which could lead to personalized cancer pain management. The development of individualized pain treatment plans, ultimately, will pave the way to an improvement of advanced cancer patients’ quality of life.