Introduction

Oxaliplatin, a platinum compound with a 1,2-diaminocyclohexane carrier ligand that differs from other platinum drugs such as cisplatin and carboplatin, has been in use for the treatment of various digestive tract cancers [1]. It is often used in combination regimen along with other drugs such as capecitabine, gemcitabine, 5fluorouracil/folinic acid, and epirubicin. Oxaliplatin-based combination regimens have shown superior activity in the form of prolonged disease progression-free survival and overall survival in both advanced and metastatic settings of digestive tract cancers [2,3,4,5,6,7].

Among non-haematological toxicity profile of oxaliplatin, peripheral neuropathy (PN) is one of the toxicities which affect nearly 75% of all the treated patients [8]. Initially, oxaliplatin-induced peripheral neuropathy (OXAIPN) was recognised as a prominent side effect, but now has emerged as the major dose limiting toxicity. OXAIPN is unique among other platinum drugs and it represents as two clinically distinct syndromes, i.e. acute OXAIPN and chronic OXAIPN. The acute syndrome is transient and occurs during or within hours of the drug infusion. It is usually characterised by distal or perioral paresthesias and dysesthesias. Existing studies have shown that 90% of the patients treated with oxaliplatin develop any grade of acute OXAIPN and 22% of them would require prolongation of drug infusion or cessation of treatment [9]. The chronic OXAIPN is dose dependent and develops gradually due to repeated doses of administration of oxaliplatin and persistent between cycles of chemotherapy treatment. It will be seen at around 750 mg of cumulative dose of oxaliplatin [10,11,12,13,14].

The development of oxaliplatin-induced peripheral neuropathy (OXAIPN) may lead to non-adherence and can hinder the overall success of the therapy. As the susceptibility to develop OXAIPN differs greatly from patient to patient, the prediction of the inter-individual variability in relation to the development of OXAIPN could improve the quality of life by allowing the individualization therapy. The exact mechanism of OXAIPN remains unknown and there have been studies attempting to define predictive markers for OXAIPN in addition to identify effective prophylactic or therapeutic agents.

Although, to date, various pharmacogenetics studies revealed significant association between the incidence of OXAIPN and genetic variants within genes involved in oxaliplatin metabolism (GSTP1) [15], oxalate metabolic pathway (AGXT) [16], cell cycle control (CCNH), excretion and transport pathway (ABCG2) [17], voltage-gated sodium and potassium channels [18, 19], DNA repair pathways [20], the current data are insufficient to identify the single-nucleotide polymorphisms (SNPs) that contribute to the variation in susceptibility to OXAIPN. Clinical validation of genetic variants can improve our understanding of OXAIPN pathogenesis and more importantly can directly enhance patient care by enabling identification of patients at high risk of oxaliplatin-induced neuropathy who should be treated with modified doses of oxaliplatin or non-oxaliplain containing chemotherapeutic regimens. In addition, South Indian population represents a genetically distinct group [21]. As the difference in the distribution of alleles could confound the association between genotype–phenotypes and are the primary reasons for the conflicting results, it is important to study the effect of genetic variants in each population. Hence In the present study, we aimed to investigate a panel of SNPs within candidate genes involved in the oxaliplatin metabolism, cell cycle control, detoxification or excretion pathways to predict the acute OXAIPN in South Indian Tamil patients with digestive tract cancers treated with oxaliplatin-based combination chemotherapy.

Methodology

Study design and patient selection

This prospective, single institutional cohort study was carried out between November 2014 and December 2016 at Regional Cancer Centre (RCC), in collaboration with the departments of pharmacology, Medical Oncology and Neurology from Jawaharlal Institute of Postgraduate Medical Education and Research (JIPMER), Puducherry, India. The sample size was calculated using the power and sample size calculation software (version 3.0, 2009, Vanderbilt University, Nashville, Tennessee, USA) by taking the following factors into the account: power 90%, 5% type-I error rate, acute OXAIPN occurrence rate of 20% and relative risk 2% and minor allele frequency 10%. Patients with histologically confirmed digestive tract cancer scheduled to receive oxaliplatin-based combination chemotherapy (GEMOX, Modified FOLFOX-4, CAPOX, and EOX) either as adjuvant and neoadjuvant treatments or as palliative treatment were prospectively screened and registered onto the study. Patients aged greater than 18 years, patients with eastern cooperative oncology group performance status (ECOG-PS) less than 2, patients with the normal hepatic and renal functions and patients with normal electrophysiological profile were included into the study. Pregnant and lactating women, patients receiving other drugs which cause peripheral neuropathy were excluded from our study.

Treatment and end points of the study

The patient distribution to oxaliplatin-based combination chemotherapy was made according to medical oncologist’s decision. For the FOLFOX and GEMOX regimens, oxaliplatin was administered at 85 and 100 mg/m2, every 2 weeks, for 12 cycles and with CAPOX and EOX regimens, patients were given oxaliplatin at 130 mg/m2, every three weeks, for 8 cycles. All the study patients received oxaliplatin intravenously for 2 h. In the neoadjuvant and adjuvant settings, patients have received either 12 cycles of FOLFOX/GEMOX or 8 cycles of CAPX/EOX according to the primary malignancy; however, a number of cycles in palliative setting were not fixed and were decided based on response and tolerance. The primary end point of the study was the development of grade 3/4 toxicity of acute OXAIPN.

Assessment of acute OXAIPN in the study cohort

All the patients who were registered into the study were clinically assessed at baseline (visit 1) and during the treatment period (at every cycle) to detect the incidence of OXAIPN. For the assessment of acute OXAIPN, a simple descriptive questionnaire (a yes/no response format) was used to quantify the frequency of the 11 most common acute OXAIPN syndrome symptoms (Table 1) which were previously described by Velasco et al. [22] and Argyriou et al. [23]. For this study, severity was graded based on the sum of number of symptoms manifested during the course of the treatment. If the patients exhibited 1–3 symptoms the severity of OXAIPN was graded as grade 1, if the symptoms were present between 4 and 6 then the severity was graded 2 and if 6–9 symptoms were present then it was graded 3 and more than 9 symptoms as grade 4.

Table 1 Acute OXAIPN symptoms scored in patients receiving oxaliplatin-based chemotherapy in the study cohort
Full size table

DNA extraction and genotyping

Before the initiation of oxaliplatin-based combination regimen therapy, approximately 5 ml of peripheral venous blood was collected from each study subject in tubes containing 10% ethylene diamine tetra acetic acid (EDTA). The genomic DNA (lymphocyte DNA) from patient blood sample was extracted using the standard phenol–chloroform extraction method [24] and was quantified by multianalyzer (TECAN Infinite M200, Switzerland).

The genotyping was done on Real-Time Thermocycler (ABI Prism 7300, Foster City, CA, USA) using validated Real-Time TaqMan single-nucleotide polymorphism (SNP) genotyping probes (Applied Biosystems, Foster City, CA, USA). All the samples were analysed in duplicates along with negative controls to ensure authenticity of the results.

Selection of SNPs

The following five SNPs, namely GSTP1 (rs1965), ABCG2 (rs3114018), CCNH (rs2230641, rs3093816) and AGXT (rs4426527) were selected for the study from the SNP database of the National Centre for Biotechnology Information. The details of the five SNPs are given in Table 2.

Table 2 Characteristic features, rs IDs and assay IDs of the five SNPs studied
Full size table

Statistical analyses

The genotypes for each SNP were analysed as a three-group categorical variable (reference model) and tested for deviation from Hardy–Weinberg equilibrium (HWE). The Fisher’s exact test was used to determine the influence of genotype with toxicity groups. Two-sided p values of < 0.05 were considered statistically significant. Bonferroni correction was also carried out for multiple testing. A p value < 0.01 (0.05 divided by 5, the total number of SNPs studied) was considered statistically significant after Bonferroni correction.

All statistical analyses were carried out using Graph Pad Instat version 3.0 (Graph Pad Software Inc., San Diego, CA, USA).

Results

Characteristics of the study patients

A total of 284 patients who were scheduled to receive oxaliplatin-based combination chemotherapy were screened initially and a total of 228 subjects were eventually included in the study between November 2014 and December 2016. During screening, 56 cancer patients were excluded due to various reasons such as presence of clinical neuropathy or for receiving oxaliplatin previously.

Out of 228 patients with digestive tract malignancies, 111 (49%) patients had gastric cancer, 108 (47%) patients had colorectal cancer and 9 (4%) patients had pancreatic and gall bladder cancers. The median age was 53 years (range 19–75). There were 142 males and 86 females in the study. The detailed demographic and clinical characteristics of the study subjects are given in Table 3. All the patients received a mean cumulative dose of 772.8 ± 267.8 mg/m2 of oxaliplatin.

Table 3 Baseline demographics and clinical characteristics of study patients (N = 228)
Full size table

Incidence and severity of acute OXAIPN

Among 228 patients, acute OXAIPN was observed in 111 (48.7%) patients. Gradation of OXAIPN revealed grade1 in 71 (64.0%) patients, grade 2 in 35 (31.5%) patients and grade 3 in 5 (4.5%) patients. The median time for the onset of acute OXAIPN was 21 (1–84) days. During the study period, 75 out of 228 patients did not receive the planned treatment. The treatment was withdrawn in patients exhibiting grade 3 toxicity of nausea, vomiting/diarrhoea and thrombocytopenia (n = 32), patients defaulted during treatment (n = 4) and death due to disease progression (n = 39). The details on the occurrence of acute OXAIPN at different cycles of oxaliplatin treatment in study subjects are given in Supplementary Table.1.

The genotype and allele frequencies of 5 SNPs and their association with acute OXAIPN in the study cohort

The genotype statuses of the five SNPs were determined for all 228 patients. The genotype and allele frequencies of the five variants observed in our population are depicted in Table 4. The genotype frequency and distribution of all the SNPs alleles were found to be within Hardy–Weinberg equilibrium (HWE) probability limits.

Table 4 Observed genotype and allele frequency of SNPs in the present study population (N = 228 patients)
Full size table

In the present study, genotypes of five SNPs were correlated with the occurrence of acute OXAIPN and its severity in different genetic models [25]. The CCNH genetic variants rs2230641 and rs3093816 were significantly associated with both the incidence and severity of acute OXAIPN. Patients harbouring a single copy of the mutant allele for the polymorphisms rs2230641, rs3093816 were observed to have higher risk for developing acute and severe OXAIPN. Dominant model genetic analysis also revealed that the mutant allele in both the loci conferred statistically significant risk to develop acute OXAIPN after applying Bonferroni correction (p = 0.001 for rs2230641; p = 0.001 for rs3093816) (p < 0.01 after Bonferroni correction) (Table 5). We found no significant association between carriers of rs1695 (GSTP1), rs3114018 (ABCG2) and rs4426527 (AGXT) variants with the incidence and severity of acute OXAIPN in our study cohort (Table 5).

Table 5 Association of various SNPs with the incidence and severity of acute OXAIPN
Full size table

Discussion

Oxaliplatin was one among the essential medicines listed by World Health Organisation (WHO) for the management of solid tumours [26]. However, the usage of this drug may produce untoward side effects such as acute and/or chronic neuropathies. Oxaliplatin-induced peripheral neuropathy may affect a large number of cancer patients and may warrant cessation of treatment. The recovery period from OXAIPN is variable; however, the nerve damage affects quality of life of cancer patients. Several studies were carried out to elucidate the molecular mechanism of OXAIPN; yet, a clear mechanism of peripheral neuropathy induced by oxaliplatin still remains obscure. A variation in response to treatment and development of toxicity among the patients undergoing chemotherapy suggests an underlying genetic cause [27, 28]. Hence, the identification of genetic biomarkers associated with OXAIPN might be helpful in predicting the risk for toxicity in patients treated with oxaliplatin.

In the present study, we evaluated the influence of five genetic polymorphisms in genes associated with oxaliplatin metabolism (GSTP1), oxalate metabolic pathway (AGXT), cell cycle control (CCNH), excretion and transport pathways (ABCG2) for the development of acute OXAIPN. In our study group, we observed that the variations in exon (rs 2230641) and intron (rs3093816) of CCNH gene involved in cell cycle control was associated with the development of acute OXAIPN. We observed that the presence of single mutant allele in each locus enhances the risk to develop OXAIPN, suggesting that even a single copy of the mutant allele is a risk factor for acute OXAIPN.

CCNH plays a major role in cell cycle control and it was proposed to be associated with the repair of dorsal root ganglia after oxaliplatin exposure [29,30,31]. Though the exact functional implications of these two genetic variants are not known, it was hypothesised that the non-synonymous variant may lead to the synthesis of proteins with compromised DNA repairing functions, which could in turn lead to OXAIPN. The role of the intronic variant (rs3093816) of CCNH with the development of OXAIPN remains obscure. Polymorphisms within the introns were reported to affect the splicing of transcripts and may lead to the reduced synthesis of functional proteins [32, 33]. Hence, the functional effect of the intronic variant of CCNH has to be validated by further molecular studies.

Our findings are concurrent with the reports of Custodio et al., study results. They reported that the missense variant of CCNH (rs 2230641) in addition with the variant in ABCG2 (rs3114018) was associated with chronic OXAIPN in Spanish Caucasoids [17]. However, in our study, we observed that the ABCG2 (rs3114018) variant failed to confer significant association with the development of OXAIPN. Though the frequency of the mutant allele C of ABCG2 rs3114018 was more than that of the reference allele A, it failed to confer statistical significance in our cohort. In contrast to the above, Kanai et al. screened 12 variants in genes associated with DNA repair, drug transport and metabolism in a Japanese cohort and reported that none of them were associated with oxaliplatin-induced peripheral neuropathy [34]. The differences observed might be due to variable mutant allele frequencies between the ethnic populations.

Argyriou et al., briefed that oxaliplatin therapy induces peripheral neuropathy by affecting mitochondrial function, axonal transport, cytoskeleton structure and nodal axonal voltage-gated Na+ channels [9]. Hence, we screened genetic variants in genes such as GSTP1 (rs1695), ABCG2 (rs3114018) and AGXT (rs4426527) and found that none of these polymorphisms influenced the development of OXAIPN in our population. Grothey et al. reported that Caucasians harbouring a single copy of the mutant allele G or with homozygous GG genotype for GSTP1 (rs1695) variant were at a higher risk to develop neurotoxicity [35]. In contrast to their report, the observed mutant allele frequency of GSTP1 (rs1695) variant was 43.7% and it did not confer significant risk for acute OXAIPN in our population. The differences can be attributed to population heterogeneity. A meta-analysis revealed no association between GSTP1 (rs1695) and oxaliplatin-induced neuropathies [36].

Our finding suggests that the genetic variants within CCNH gene might serve as a common biomarker to predict acute OXAIPN in south Indians. In an observational research, sample size cannot to be planned based on formal sample size calculation but, in general, a large number of patients are required to detect smaller associations [37]. Hence, our findings require validation in a larger cohort. For the first time we report that the intronic variant in CCNH (rs rs3093816) was also associated with the incidence of acute OXAIPN. The molecular mechanism of neurotoxicity induced by the CCNH intronic variant (rs3093816) and its interaction with the other genetic variants has to be elucidated by functional studies.

Conclusion

Multiple genetic factors are involved in oxaliplatin-induced neurotoxicity. The exonic variant of CCNH (rs2230641) might serve as a common biomarker for acute OXAIPN. The role of intronic variant of CCNH (rs3093816) in the development of oxaliplatin-induced acute peripheral neuropathy has to be elucidated.