Population-Specific Pharmacogenetic Variants Associated with Chemotherapy Response in the Bulgarian Population

1. Introduction

Pharmacogenetic differences contribute substantially to the variability observed in individual responses to chemotherapy, affecting both treatment effectiveness and the occurrence of adverse drug reactions. Germline genetic variation in pathways related to drug biotransformation, transport, and DNA damage response can modify systemic drug exposure, tumor sensitivity, and toxicity patterns across multiple classes of anticancer agents [1,2].

A growing amount of evidence supports the clinical relevance of pharmacogenetic markers for widely used chemotherapeutic regimens, including platinum compounds, fluoropyrimidines, taxanes, anthracyclines, alkylating agents, and targeted therapies such as epidermal growth factor receptor (EGFR) inhibitors. Polymorphisms in genes encoding drug-metabolizing enzymes and transporters (e.g. DPYD, MTHFR, ABCB1, ABCC2, ABCG2, SLC22A2, SLCO1B3), as well as genes involved in oxidative stress regulation and DNA repair processes (e.g. SOD2, XRCC1, ERCC2, XPC, ATM, TP53), have been linked to variability in treatment response and clinical outcomes in cancer patients [3,4].

Although pharmacogenetic testing is increasingly implemented in personalized oncology, its broader clinical application is constrained by the limited availability of population-specific reference data. Publicly accessible databases, such as the Genome Aggregation Database (gnomAD), offer extensive allele frequency information but predominantly reflect individuals of Western European descent. Consequently, genetic profiles characteristic of Eastern European populations, including Bulgarians, are underrepresented, which may restrict the direct application of pharmacogenetic data from global datasets [5].

In the present study, targeted next-generation sequencing was applied using a panel of 484 genes associated with drug metabolism and response to anticancer therapy. While multiple forms of genetic variation were detected, variants of clinical relevance for chemotherapy metabolism were limited to single nucleotide polymorphisms (SNPs). The gene panel was designed to capture pharmacogenetic markers relevant to treatment regimens commonly administered to patients with colorectal cancer, breast cancer, and non-small cell lung cancer — among the most prevalent malignancies in Bulgaria and worldwide. These regimens included platinum-based agents, pyrimidine analogues, folinic acid, EGFR inhibitors, taxanes, anthracyclines, and Cyclophosphamide [3].

The objective of this study was to define the distribution of germline pharmacogenetic variants relevant to these regimens in Bulgarian cancer patients and to assess their allele frequencies in comparison with global reference populations. Identifying population-specific patterns of pharmacogenetic variation may enhance the clinical relevance of genetic testing and support the development of more precise, locally optimized chemotherapy strategies.

2. Materials and Methods

Initially, 140 individuals were enrolled in the study. Of these, 90 participants (50 cancer patients and 40 healthy controls) completed all study procedures without withdrawal and were included in the final analysis. The oncological group consisted of 26 patients with colorectal cancer, 13 with non-small cell lung cancer, and 11 with breast cancer. These cancer types were selected based on their high prevalence in Bulgaria and worldwide, the frequent occurrence of chemotherapy-associated adverse drug reactions, and the overlap in therapeutic regimens used for their treatment.

The study protocol was reviewed and approved by the Institutional Review Board of the Scientific Ethics Committee at the Medical University of Plovdiv (approval number C–09-2/10.04.2020). Written informed consent was obtained from all participants or their legal representatives prior to enrollment. All procedures were performed in accordance with institutional ethical requirements and the principles of the Declaration of Helsinki.

Peripheral blood samples were collected in EDTA-containing Monovettes for the extraction of genomic DNA. For cell-free DNA (cfDNA) analysis, two Cell-Free DNA BCT (Streck) tubes were obtained from each patient. Genomic DNA was isolated using the QIAamp DNA Blood Mini Kit (QIAGEN), and DNA concentration was assessed with a NanoDrop™ One Microvolume UV–Vis spectrophotometer (Thermo Scientific™). Cell-free DNA extraction was performed using the cfPure® Cell-Free DNA Extraction Kit (BioChain), with DNA quantification carried out on a DeNovix fluorimeter.

Genomic DNA was obtained from 50 samples, including 10 cancer patients and 40 healthy individuals, while cfDNA was isolated from an additional 40 cancer patients. All DNA samples were subjected to next-generation sequencing at Novogene Corporation Inc., where sequencing and initial bioinformatic processing were conducted. Variants identified from genomic DNA and cfDNA were not analyzed comparatively.

3. Results

From the initially enrolled 26 patients diagnosed with colorectal cancer, three were excluded from further analysis due to treatment with radiotherapy in the absence of chemotherapy. Consequently, 23 patients with colorectal cancer (CRC) were included in the final evaluation. These individuals received combinations of platinum-based agents (Oxaliplatin), folinic acid (Leucovorin), pyrimidine analogues (5-fluorouracil, Capecitabine), topoisomerase I inhibitors (Irinotecan), EGFR inhibitors (Panitumumab, Cetuximab), and monoclonal antibodies (Bevacizumab, Ramucirumab) (Supplementary Table S1). Observed adverse drug reactions (ADRs) and their severity grades are summarized in Supplementary Table S1, with toxicity classified according to the Common Terminology Criteria for Adverse Events (CTCAE).

The non-small cell lung cancer (NSCLC) cohort consisted of 13 patients treated with taxanes (Docetaxel, Paclitaxel), platinum compounds (Carboplatin, Cisplatin), topoisomerase II inhibitors (Etoposide), nucleoside analogues (Gemcitabine), monoclonal antibodies (Pembrolizumab, Atezolizumab, Bevacizumab, Nivolumab, Ramucirumab), EGFR inhibitors (Erlotinib), and tyrosine kinase inhibitors (Alectinib) (Supplementary Table S2). Documented ADRs and corresponding toxicity grades are presented in Supplementary Table S2.

The third malignancy-based subgroup included 11 women with breast cancer. Treatment regimens comprised taxanes, anthracyclines (Farmorubicin), pyrimidine analogues (5-fluorouracil), platinum-based chemotherapeutics (Carboplatin), alkylating agents (Cyclophosphamide), and hormone therapy (Supplementary Table S3). ADRs observed in this group are detailed in Supplementary Table S3.

Given the overlap in chemotherapeutic agents administered across malignancies, patients who received chemotherapy (n = 47) were regrouped according to the type of medication used in order to evaluate the clinical relevance of pharmacogenetic variants. The resulting treatment-based groups included patients treated with platinum-based agents (n = 35), pyrimidine analogues (n = 26), folinic acid (n = 21), EGFR inhibitors (n = 11), taxanes (n = 15), anthracyclines (n = 8), and Cyclophosphamide (n = 8) (Supplementary Tables S4–S10). These tables summarize the genes associated with the metabolism of each drug class and the zygosity status (homozygous or heterozygous) of the detected variants.

Allele frequencies of germline pharmacogenetic variants were compared between cancer patients and the 40 healthy controls, as well as between cancer patients and non-cancer reference populations from the Genome Aggregation Database (gnomAD). Statistical significance was assessed using Fisher’s exact test and Z-score analysis, with a threshold of p < 0.05 (Supplementary Tables S11–S17).

Among the 35 patients treated with platinum-based chemotherapeutics, high frequencies were observed for several variants associated with drug metabolism, including ABCC2 c.1249G>A (95%), GSTP1 c.313A>G (94%), SLCO1B3 c.699G>A (94%), and XPC c.1496C>T (91%). Significant differences compared with healthy controls were identified for XPC c.2815C>A, XPC c.1496C>T, and SLC22A2 c.808T>G. Although the ERCC2 c.934G>A variant did not differ significantly from the local control group, its frequency was significantly lower than that reported in non-cancer individuals in gnomAD (p = 0.0014; 0.0008) (Supplementary Table S11).

In patients receiving pyrimidine analogues (5-fluorouracil/Capecitabine; n = 26), the DPYD c.2194G>A variant occurred at a significantly higher frequency compared with both the healthy Bulgarian cohort and the global reference population (p < 0.05). XRCC1 c.1196A>G also demonstrated statistically significant higher prevalence relative to local controls (p = 0.0023) and gnomAD (p < 0.00001). No significant differences were detected for the remaining analyzed variants, including MTHFR, DPYD, CYP1B1, ABC transporters, GSTP1, TP53, and ERCC2 polymorphisms (Supplementary Table S12).

Analysis of the subgroup treated with folinic acid (n = 21) focused on MTHFR c.1286A>C, MTHFR c.665C>T, and ERCC2 c.2251A>C. Allele frequencies of these variants did not differ significantly from either the healthy control group or gnomAD data (Supplementary Table S13).

Among patients treated with EGFR inhibitors (n = 11), only the EGFR c.1562G>A variant was detected, with an allele frequency of 0.9091. While no significant difference was observed in comparison with the local control group, the variant was significantly more frequent relative to the global reference population (p ≈ 0.036) (Supplementary Table S14).

For patients receiving taxane-based therapy (n = 15), variants in SOD2, ABCB1, SLCO1B3, and ERCC2 were evaluated. SOD2 c.47T>C and ABCB1 c.2677T>G were present in 67% of patients but showed no statistically significant deviation from control populations. Both SLCO1B3 variants were detected in all individuals. However, their frequencies were comparable to those reported in local and global datasets. No significant differences were observed for ERCC2 c.2251A>C (Supplementary Table S15).

Patients treated with anthracyclines (Farmorubicin) and Cyclophosphamide (n = 8) represented an identical subgroup and were therefore analyzed simultaneously. Detected variants included CYP1B1 c.1294G>C, ABCG2 c.421C>A, ABCC2 c.1249G>A, GSTP1 c.313A>G, ATM c.5557G>A, TP53 c.215C>G, and XRCC1 c.1196A>G. All patients carried ABCC2 c.1249G>A and GSTP1 c.313A>G, while CYP1B1 c.1294G>C and TP53 c.215C>G were observed in 87.5% of cases. No statistically significant differences were identified when compared with control cohorts, except for a borderline significant deviation of XRCC1 c.1196A>G relative to gnomAD (Z-score p = 0.0366) (Supplementary Tables S16 and S17).

Finally, comparison of allele frequencies for 23 pharmacogenetic variants between the combined Bulgarian cohort (patients and controls; n = 90) and gnomAD revealed statistically significant differences for 10 variants (p < 0.05) (Table 1). Lower allele frequencies were observed for MTHFR c.1286A>C, DPYD c.85T>C, ABCG2 c.421C>A, and ERCC2 c.934G>A, whereas increased frequencies were detected for MTHFR c.665C>T, DPYD c.2194G>A, XPC c.2815C>A, EGFR c.1562G>A, XRCC1 c.1196A>G, and ERCC2 c.2251A>C.

4. Discussion

The present study outlines germline pharmacogenetic variability in Bulgarian cancer patients and identifies multiple deviations from allele frequencies reported in global reference datasets. The differences between patients, local controls, and gnomAD indicate that pharmacogenetic risk estimates based on large aggregated populations may not accurately reflect regional populations, particularly in Eastern Europe [6,7].

Variants within genes involved in folate metabolism displayed contrasting distribution patterns. The reduced frequency of MTHFR c.1286A>C in the Bulgarian cohort suggests a comparatively minor contribution to fluoropyrimidine- or platinum-related toxicity at the population level. In contrast, the increased frequency of MTHFR c.665C>T suggests a possible effect on treatment tolerance and adverse event risk, consistent with its reported impact on folate-dependent pathways relevant to chemotherapy response [6,8].

For fluoropyrimidine-associated genes, DPYD c.2194G>A emerged as a notable finding due to its increased prevalence relative to both local controls and global datasets. This observation highlights its relevance for clinical dose individualization and strategies aimed at reducing toxicity. Other DPYD variants examined in this study did not exhibit population-specific differences, in line with previous reports describing their limited predictive value in European patient cohorts [9,10].

Genes implicated in xenobiotic metabolism showed variable results. The distribution of CYP1B1 c.1294G>C resembled that of reference populations, suggesting that this variant alone is unlikely to account for substantial interpopulation differences in anthracycline or Cyclophosphamide toxicity [11,12]. Similarly, the low prevalence of ABCG2 c.421C>A in the Bulgarian cohort reduces its utility as a specific for the population marker, despite prior associations with chemotherapy-induced adverse reactions [6,13]. In contrast, the higher prevalence of SLC22A2 c.808T>G supports earlier findings linking this variant to platinum-related hepatotoxicity and highlights its potential relevance for regional risk stratification [7,14].

Notable population-specific trends were also observed among DNA repair genes. Increased frequencies of XPC c.2815C>A and c.1496C>T in patients treated with platinum-based regimens are concordant with previously described associations with hematologic and gastrointestinal toxicity. Likewise, the high prevalence of XRCC1 c.1196A>G across treatment groups suggests it may affect chemotherapy tolerance across multiple drug classes [7,15,16].

Differences in ERCC2 variant distribution further emphasize regional specificity. The higher occurrence of ERCC2 c.2251A>C aligns with prior reports linking this variant to increased toxicity risk in platinum- and fluoropyrimidine-based therapies, whereas the comparatively low frequency of ERCC2 c.934G>A may reduce its clinical relevance within combined chemotherapy regimens in this population [17,18].

The almost universal presence of EGFR c.1562G>A in patients receiving EGFR-targeted therapy is another important finding. As this variant has been linked to skin and gastrointestinal side effects, it may be worth considering when planning treatment, especially in settings where EGFR inhibitors or anti-VEGF drugs are commonly used [6].

Overall, the pharmacogenetic profile observed in this cohort reflects a combination of overrepresented and underrepresented variants that collectively shape chemotherapy tolerance in the Bulgarian population. Variants with increased frequencies — including MTHFR c.665C>T, DPYD c.2194G>A, XPC c.2815C>A, EGFR c.1562G>A, XRCC1 c.1196A>G, and ERCC2 c.2251A>C — may suggest a heightened susceptibility to treatment-related toxicity and support the integration of pre-therapeutic genotyping into routine clinical practice. Conversely, variants observed at lower frequencies, such as MTHFR c.1286A>C, DPYD c.85T>C, ABCG2 c.421C>A, and ERCC2 c.934G>A, may partially explain interpopulation differences in chemotherapy tolerance and underscore the necessity of region-specific pharmacogenomic data.

5. Conclusions

The findings of this study illustrate the distinct pharmacogenetic landscape of Bulgarian cancer patients, characterized by both increased and reduced frequencies of clinically relevant germline variants in comparison with global reference populations. Such population-specific patterns highlight the importance of regional genetic data for improving the interpretation of chemotherapy-related risk and therapeutic outcomes. Incorporation of locally derived pharmacogenetic information into existing national and international resources may enhance the accuracy of pharmacogenomic models and support more individualized and safer cancer treatment strategies. Expansion of this research through larger, multicenter Bulgarian cohorts will be essential to confirm these observations and to facilitate the broader integration of pharmacogenetic testing into routine oncology practice.