Background

To date, the malignant tumor has become responsible for the majority of global deaths, and its incidence has been increasing over recent decades [1]. For both sexes combined, CRC is the third most prevalent diagnosed gastrointestinal malignancies globally and remains the fifth common cause of mortality by cancer in China, which is responsible for more than 191,000 deaths annually [1, 2].

5-FU and capecitabine, an oral prodrug of 5-FU, are the backbone of the therapeutic scheme for CRC and many other solid tumors including monotherapy and combined with other chemotherapy drugs. Common combination regimens include FOLFOX [3]: (oxaliplatin combined with bolus/infusional 5-FU and leucovorin) and XELOX [4]: (capecitabine plus oxaliplatin). Unfortunately, approximately 10 to 30% of patients display varying degrees of adverse effects [5], most frequent manifestations include anemia, neutropenia, nausea, vomiting, diarrhea, and neurological toxicities. Various clinical studies revealed that dihydropyrimidine dehydrogenase (DPD), encoded by the DPYD gene, is the limiting velocity enzyme in the catabolism of fluoropyrimidines. DPD is responsible for catabolism, more than 80% of given fluoropyrimidines. Deficiency of the DPYD gene results in excess fluoropyrimidine accumulation in the blood and increases the risk of toxicity. DPD enzyme activity detection currently commonly used methods are western blotting and HPLC [6, 7]. Western blotting requires preliminary experiments to determine the best conditions, and there are many interference factors. HPLC is expensive and not widely available in hospitals. Numerous studies have revealed that the polymorphism of the DPYD gene mutation results in a significantly decreased enzyme activity, such as well-known IVS14+1G>A (DPYD*2A). The DPYD IVS14+1G>A alternate splicing in intron 14 resulting in creating a fragmentary protein and leading to the lack of normal DPD enzyme function. DPYD c. 1627A>G (DPYD*5A) and DPYD c. 85T>C (DPYD*9A) are linked with decreased enzyme activity. In codon 29 of DPYD at the 85 position, the point mutation of T to C gives rise to an amino acid alteration from Cys to Arg, and DPYD 1627A>G substitution results in the DPYD*5A allele due to the change in I543V. In different ethnic groups, DPYD genetic polymorphisms vary widely [8].

Glutathione-S-transferase P1 (GSTP1) is a rate-limiting enzyme that catalyzes the detoxification pathway of platinum drugs, for cell protection. The GSTP1 Ile105Val (rs1695, A>G) is a missense mutation and affects the activity of the GSTP1 enzyme. Oxaliplatin-related cumulative peripheral neuropathy was reported to be more frequent and severe in patients with heterozygous (AG) genotype when compared to patients with wild-type alleles (AA) [9]. Wiese et al. [10] conclude that GSTP1 c.313A>G polymorphism was associated with the overall survival of patients treated with oxaliplatin-based therapy. From the above, we hypothesize that DPYD and GSTP1 polymorphisms could be linked to acute toxicity in patients undergoing 5-FU and oxaliplatin-based chemotherapy.

Our study aims to board the possible relationship between DPYD and GSTP1 genomic variations and adverse reactions with 5-FU and oxaliplatin-based treatment in the Meizhou Hakka population, southern China, a unique Han Chinese ethnic group having its civilization [11]. We performed a systematic retrospective pharmacokinetics study in a cohort of 104 Meizhou Hakka CRC patients for the first time to explore the associations.

Patients and methods

Population study and data collection

One hundred and four registered patients with histologically confirmed CRC who received adjuvant chemotherapy regimen containing 5-FU, capecitabine, or oxaliplatin between September 2016 and December 2017 at the Meizhou People’s Hospital were investigated in this retrospective study. Patients had to follow the lists for inclusion criteria such as (1) histopathologically proven colorectal cancers in any stage; (2) aged over 18 years; (3) patients received adjuvant chemotherapy containing 5-FU, capecitabine, or oxaliplatin in any line of treatment and received at least 2 courses for treatment; (4) any previous chemotherapy completed half a year ago, without any other adjuvant chemotherapy or radiotherapy; (5) life expectancy > 4 months; (6) adequate hematological, liver, or kidney functions. Patients were excluded from enrollment as follows: (1) patients with concomitant neoplastic disease other than colorectal cancer; (2) any non-compliance of the inclusion criteria of above; (3) severe renal or hepatic disorder, severe hematological disease before treatment; (4) less than 6 months of previous treatment unless appear greater than grade 3 adverse events. Patients who disagree to sign an informed consent form were excluded from the research. Written informed consent was provided for each patient before any procedure of this study. All participants were then followed up 18 months to obtain their detailed clinical response information and recorded in a medical history sheet. The study protocol was approved by the institution’s review and Human Ethics Committees of the Meizhou Peoples’ Hospital (No. MPH-HEC 2016-A-40). Clinical characteristics distribution of 55 men and 49 women newly diagnosed with CRC from 2016 to 2017 were enrolled in the final study. The demographic and clinical-pathological features of our study cohort were evaluated as summarized in Table 1. The median age at diagnosis was 56 (range 25–78). All adverse drug reactions and toxicity underwent in the front 6 cycles of chemotherapy in each patient were monitored and graded in line with the National Cancer Institute Common Terminology Criteria for Adverse Events v3.0 (CTCAE) [12].

Table 1 Demographic and clinical characteristics of treated patients
Full size table

Genotyping

All collected blood samples were anticoagulants with EDTA. Genomic DNA from each enrolled patient was extracted using the DNA Isolation Kit for Blood/Bone Marrow/Tissue (Tiangen Biotech (Beijing) Co., Ltd.). DNA concentration and purity were evaluated using a NanoDrop2000 (Thermo Scientific). The primer sequences and the restriction enzymes used were provided by SINOMD Gene (Beijing) Co., Ltd. Genomic DNA was used for the cyclic condition consisted of one predenaturation step at 95 °C for 3 min initial cycle, then 45 cycles of 94 °C for 15 s, 63 °C for 1 min and extension at 72 °C for 1 min, and a final elongation step of 10 min at 72 °C. PCR products were purified with ExoSap-It (ABI PCR Product Cleanup Reagent). Single-nucleotide polymorphisms were analyzed using ABI 3500 Dx Genetic Analyzer (ABI Terminator v3.1 Cycle Sequencing kit) and Sequencing Analysis v5.4 (Life Technologies, CA, USA). We looked for the following SNPs: DPYD*2A, DPYD*5A, DPYD*9A, and GSTP1.

Toxicity grading and statistical analysis

All quantitative variables such as age, gender, smoking, alcohol intake, tumor location, neoplasms histologic type, grade, and stage of the tumor were analyzed. Toxicity including hematopoietic toxicity (anemia, leucopenia, neutropenia, and thrombocytopenia), gastrointestinal toxicity (mucositis, vomiting, and diarrhea), and dermatologic toxicity (ulceration and hand-foot syndrome) were grouped by four classes. Association between DPYD and GSTP1 variant status with a response and different toxicity was evaluated by Fisher’s exact test. The Hardy-Weinberg equilibrium (HWE) of DPYD and GSTP1 genotypes were assessed using the χ2 test. A p value of less than 0.05 was deemed to be statistically significant. All tests were carried out by the SPSS software version 21.0 (SPSS Inc., Chicago, USA)

Results

Histopathological type and clinical stage

As shown in Table 1, colon cancer was more predominant than rectum cancer (62.54 and 37.5%, respectively) about tumor location. All histopathological types were adenocarcinoma. Seventy-four (46.9%) patients were affected by stage III–IV tumors. Following the American Joint Committee on Cancer’s (AJCC) Cancer Staging 6th edition 2002 TNM grading systems, there were 28.8% of stage II, 44.2% of stage III, and 27% of stage IV about the grade of the tumor, but none of stage I.

Chemotherapy toxicity

The toxicity of chemotherapy in the study participants is shown in Table 2. Treatment of fluoropyrimidine and platinum combination chemotherapy was made at the time of the first diagnosis of CRC. All patients received chemotherapy regimens for at least 2 cycles and were evaluated for adverse event outcomes. Of the total 104 cases, 72 patients (69.2%) develop severe chemotherapy-related toxicities, including 18.3% suffered from anemia, 9.6% for leukopenia, 15.4% for neutropenia, 11.5% for thrombocytopenia, 8.7% for mucositis, 8.7% for diarrhea, 22.1% for vomiting, and 8.6% for dermatological toxicities.

Table 2 Toxicity of chemotherapy in the study participants
Full size table

Dihydropyrimidine dehydrogenase and glutathione S-transferase pi-1 genotypes

In our cohort, all patients were genotyped for DPYD*2A, DPYD*5A, DPYD*9A, and GSTP1 variants. The DPYD [IVS] 14+1G>A mutation was not identified in any of this study patients. In the case of DPYD*5A polymorphism, DPYD c. 1627A>G was present in 33 (31.7%) patients with heterozygous and 7 (6.7%) in homozygous, and there were 76% wild types and 24% heterozygote mutants to DPYD c. 85T>C. The results also show that the distribution of GSTP1 A/A, GSTP1 A/G, and GSTP1 G/G was 67.3%, 29.8%, and 2.9%, respectively (Table 3). Determined SNPs were found to be in Hardy-Weinberg Equilibrium.

Table 3 DPYD and GSTP1 genes frequency in the study patients
Full size table

Adverse events

The distribution of toxicity classification of patients with DPYD and GSTP1 wild type and mutant was shown in Tables 4 and 5, respectively. Among patients screened for DPYD mutant (n = 56), 40 patients had DPYD c. 1627A>G (DPYD*5A) variant, while DPYD c. 85T>C (DPYD*9A) variant exists in 25 patients. Out of them in 9 patients with both DPYD*5A and DPYD*9A heterozygous. The most commonly experienced grade 3 or grade 4 toxicity in both DPYD wild-type and mutant patients was anemia in hematological and vomiting in gastrointestinal toxicity (Table 4). The same results were found in patients with GSTP1 wild type and mutants (Table 5).

Table 4 The frequency of all grade toxicities in patients with DPYD wild type and mutants
Full size table
Table 5 The frequency of all grade toxicities in patients with GSTP1 wild type and mutants
Full size table

Association between SNPs and toxicities

Association between polymorphisms of DPYD and GSTP1 genes and toxicities of chemotherapy outline is shown in Tables 6, 7, and 8. Adverse events were experienced in most patients who received chemotherapy after the first or second cycle. We further examined the association between these SNPs and adverse events in all individuals. The existence of the GSTP1 c. 313A>G polymorphism was associated with platinum related to several toxicities such as vomiting and skin ulceration (p = 0.042 and p = 0.018, respectively). We also found that patients with the AA genotype for GSTP1 presented lower rates of serious vomiting (35.3%) than patients with the AG and GG genotype (66.7% and 100%, respectively, p = 0.027). Hand-foot syndrome and hematopoietic toxicity could not be demonstrated a statistically significant association with GSTP1 genotype in our study. However, 5-FU and capecitabine-related toxicities were not correlated to DPYD polymorphism, although there was a statistical trend for more diarrhea in the patients when carrying the mutant DPYD*5A A/G genotype.

Table 6 Association between polymorphisms of DPYD and GSTP1 genes and hematological toxicity of chemotherapy
Full size table
Table 7 Association between polymorphisms of DPYD and GSTP1 genes and gastrointestinal toxicity of chemotherapy
Full size table
Table 8 Association between polymorphisms of DPYD and GSTP1 genes and dermatology toxicity of chemotherapy
Full size table

Discussion

Fluorouracil (FU) is the most common component of chemotherapy agents for malignant tumors like leukemia, lung, breast, gastric, and colorectal cancers. Falzacappa et al. demonstrated combined fluorouracil, and rucaparib was able to enhance the efficacy and specificity of acute myeloid leukemia [13]. Adrian BogdanTigu et al. highlighted that the combination of allicin with 5-FU has an obvious against lung and colorectal cancer effect while reducing the dosage of fluorouracil [14]. For breast cancer, fluorouracil (5-FU) is widely used and a growing number of physicians are using capecitabine as first-line or second-line chemotherapy [15]. Adjuvant chemoradiotherapy based on 5-fluorouracil is the first choice after gastric cancer surgery. Li et al. suggested that accurate classification of GC patients who may benefit from 5-FU adjuvant chemoradiotherapy will help clinicians formulate more effective GC treatment regimens [16].

Fluorouracil-based treatment is generally well tolerated, except about 5 to 10% patient who suffers grade > 3 toxicity, which even impacts on the quality of life [17]. Intolerance of 5-FU is usually because of decreased activity of the fluoropyrimidine detoxifying enzyme dihydropyrimidine dehydrogenase (DPD) [18]. DPD is known to be the initial and rate-limiting enzyme catalyze degradation of more than 85% of the given 5-FU. Activity levels vary caused by variable phenotypes and genetic alterations [19]. Single-nucleotide polymorphisms (SNPs) in its encoding gene DPYD is the most common cause of deficiency in DPD enzyme activity [20]. Analyzing the impact of DPYD genotype in serum levels of 5-FU and correlative toxicity is vital to achieving good curative effect and prevent adverse events. In this research, we looked for adverse events related to fluoropyrimidines and platinum drugs-based treatment and their association with DPYD and GSTP1 gene polymorphisms. Vomiting and skin ulceration was the most chemotherapy-related toxicity observed in our study.

The allelic variant (A>G) at position 1627 amino acid in exon 13 of the DPYD gene results in an exchange of Isoleucine to Valine at codon 543 of DPYD*5A. There is a wide interethnic and inter geographical difference about gene polymorphism and frequency of DPYD*5A. DPYD*5A G/G + G/A genotype was identified in 38.4% participants, obviously more than that in African-Americans (22.7%) [21] and Caucasian subjects (28%) [22], and similar to that in Japanese (35%) [23]. DPYD*9A polymorphism is regarded as one prevalent DPYD variants. The TT and TC genotype proportion of DPYD*9A in our enrolled patients was 76% and 24%, respectively, whereas the homozygous DPYD*9A genotype was not found. The effect of DPYD*9A polymorphism on fluoropyrimidine treatment-related toxicity has been reported in previous researches. Khushman et al. [24] showed that DPYD c. T85C mutant variant was associated with diarrhea (p = 0.0055). Joerger [25] reported that the subjects carrying the DPYD*9A C allele significantly increase the risk for severe diarrhea compared to patients with the T/T genotype (p = 0.033). Nevertheless, several other studies suggest different points of view. Maarten [26], Amirfallah [27], and McLeod et al. [28] revealed that DPYD heterozygous for 85T>C was not associated with a decreased risk for grade 3 to 4 toxicity. In our study cohort, we haven't found a statistically significant relationship between DPYD*2A, *5A, *9A polymorphism and severe chemotherapy-related toxicity, owing to the limited size of the cohort in our study, and different allele frequency of these gene variants among different populations in different regions.

As a part of the phase II detoxification system, glutathione S-transferase (GST) superfamily members detoxifies cisplatin by the formation of glutathione conjugates, conversion to less toxic and increases its excretion from the body, resulting in the detoxification [29]. It is important to note that polymorphisms in GSTs are linked to the occurrence and development of various tumors by way of altering biological pathways and affecting protein expression [30]. Genes encoding GSTs, including GSTP1, different genotypes have different effects on the efficacy to detoxify exogenous and endogenous toxic species. Extensive research has been carried out to reveal the 313A>G (GSTP1) mutation is an A to G replacement at codon 105 leading to a transition in the amino acid from Isoleucine to Valine, which has been found to decrease enzymatic activity, and highly linked for causing platinum treatment-related toxicity in different cancers.

In previous studies, Santric [31] reported a significant association between GSTP1 Ile105Val polymorphism and oxaliplatin-related toxicity. Kudhair [32] analyzed Arab populations and revealed that carrying GSTP1 Ile105Val substitution is positively correlated with a higher risk for lung cancer in WP tobacco smokers. Conversely, Sophonnithiprasert [33] reported an insignificant difference between GSTP1 Ile105Val polymorphism and oxaliplatin-induced toxicity. Our research showed that GSTP1 c. 313A>G polymorphism was linked to acute toxicity in colorectal cancer patients undergoing skin ulceration, which was conflicting with research from Terrazzino [34] and Osti [35]. We also found that those participants having heterozygotes and homozygotes genotype showed a higher risk of high-grade severe toxicity compared to the wild type of GSTP1, in line with the reported study of Mir [36]. According to the research by Watson, it was suggested that GST enzyme activities were significantly reduced among individuals with Val-containing [37]. Besides, there was lower activity toward chemotherapy drug tolerance [38]. As mentioned above, the GSTP1 c.313 A > G mutation leads to a decrease in enzyme activity, thereby reducing the ability to detoxify platinum and other substances. Skin ulceration may be a direct result of the transport of the residual oxaliplatin active substance to the skin surface via sweat. Oxaliplatin results in the formation of intra and inter strands crosslinks, causing potential errors in DNA repair and damaged DNA. Finally, it diminishes numbers of endothelial cells and dermal matrix regeneration, leading to the occurrence and development of skin ulceration.

In Table 7, we found an interesting association between GSTP1 A/G genotype and severe vomiting in patients (p = 0.042), which consistent with Carron et al. [39]. As the detoxification ability of GSTP1 A/G genotype decreases, platinum accumulates in the body causes vomiting by inducing DNA damage in epithelial enterochromaffin cells of the intestine, thereby leading to serotonin release and stimulation of the vomiting center.

There are some limitations to our study. First, the number of subjects enrolling in the research is relatively small. Second, we did not investigate other SNPs of GST associated with oxaliplatin-related toxicities. The third limitation of this study is that we focus on only three out of more than 30,000 variants reported DPYD variants. Moreover, there is a difference in chemotherapy regimens. Future analysis will enroll a larger cohort and reveal the potential associations between more mutation combinations of DPYD and GST variants and severe toxicities.

Conclusions

In brief, skin ulceration and vomiting were the major dose-limiting toxicity of fluoropyrimidine and platinum-based regimens in this cohort. A significant correlation was confirmed between GSTP1 polymorphism and toxicity like vomiting and skin ulceration strengthening GSTP1 polymorphism as a predictor of severe events. Screening of genetic polymorphism of GSTP1 in colorectal cancer patients before chemotherapy could contribute to effectively decrease the occurrence of serious adverse toxicity, improve the predictive specificity for a severe event. We indicate that the detection of GSTP1 phenotype may become a routine laboratory test item. This may enable clinicians to implement the optimal therapeutic schedule to achieve satisfactory efficacy and minimize toxic and side effects, and then, lead to personalized therapy. Further studies should focus on validating results in a larger sample size, ascertain possible interactions between other DPYD and GST polymorphisms, and clinical outcomes.