Skip to content
Publicly Available Published by De Gruyter December 19, 2017

Thymidylate synthase gene variants as predictors of clinical response and toxicity to fluoropyrimidine-based chemotherapy for colorectal cancer

  • Carlos A. Castro-Rojas , Antonio R. Esparza-Mota , Francisco Hernandez-Cabrera , Viktor J. Romero-Diaz , Juan F. Gonzalez-Guerrero , Hector Maldonado-Garza , Irma S. Garcia-Gonzalez , Sergio Buenaventura-Cisneros , Josefina Y. Sanchez-Lopez , Rocio Ortiz-Lopez , Alberto Camacho-Morales , Oralia Barboza-Quintana EMAIL logo and Augusto Rojas-Martinez EMAIL logo



Fluoropyrimidines form the chemotherapy backbone of advanced and metastatic colorectal cancer (CRC). These drugs are frequently associated with toxicity events that result in dose adjustments and even suspension of the treatment. The thymidylate synthase (TYMS) gene is a potential marker of response and toxicity to fluoropyirimidines as this enzyme is the molecular target of these drugs. Our aim was to assess the association between variants of TYMS with response and toxicity to fluoropyrimidines in patients with CRC in independent retrospective and prospective studies.


Variants namely rs45445694, rs183205964, rs2853542 and rs151264360 of TYMS were genotyped in 105 CRC patients and were evaluated to define their association with clinical response and toxicity to fluoropyrimidines. Additionally, the relationship between genotypes and tumor gene expression was analyzed by quantitative polymerase chain reaction.


The 2R/2R (rs45445694) was associated with clinical response (p=0.05, odds ratio (OR)=3.45) and severe toxicity (p=0.0014, OR=5.21, from pooled data). Expression analysis in tumor tissues suggested a correlation between the 2R/2R genotype and low TYMS expression.


The allele 2R (rs45445694) predicts severe toxicity and objective response in advanced CRC patients. In addition, the alleles G(rs2853542) and 6bp-(rs151264360) are independent predictors of response failure to chemotherapy. This is the first study made on a Latin American population that points out TYMS gene variants have predictive values for response and toxicity in patients with CRC treated with fluoropyrimidine-based chemotherapy.


Tumors of the large bowel are a leading cause of morbidity and mortality due to cancer in Mexico [1]. 5-Fluorouracil (5-FU) and capecitabine are fluoropyrimidines constituting the corner stone for chemotherapy against gastrointestinal tumors. Its application in the treatment of metastatic colorectal cancer has represented a significant development in terms of overall survival, compared to the support therapy for patients [2]. Despite this, there is not one predictive biomarker of response for the current first-line scheme based on fluoropyrimidines that can be applied to patients with advanced colorectal cancer (CRC) and the search for useful predictive biomarkers of response to fluoropyrimidines to guide effective therapeutic decisions is still in progress [3], [4]. In the context of personalized medicine, these types of biomarkers are fundamental to ensure the maximum possible benefit for a specific treatment and to extend the spectrum of potential diseases that could benefit from a specific therapy.

Another key consideration of biomarkers for cancer chemotherapy is their prediction proficiency for adverse events. In the case of fluoropyrimidines, serious toxicity is observed in up to 14% of the cases with oral treatment and up to 42% with systemic therapy [5]. Concerning the predictive markers of toxicity to fluoropyrimidines, variants in the gene dihydropyrimidine dehydrogenase responsible for the limiting step in the metabolism of 5-FU were associated with severe toxic events, resulting in even fatal outcomes. However, the low allelic frequency of these variants in the general population, coupled with their low sensitivity to predict toxic events in the clinic, points out to the presence of other possible biomarkers associated with fluoropyrimidines toxicity [6], [7].

The TYMS gene codifies for the thymidylate synthase enzyme involved in the de novo synthesis of deoxythymidine monophosphate. As this enzyme is the pharmacological target for the fluoropyrimidines, TYMS gene variations are attractive candidates for pharmacogenetic studies. However, the coding region of the TYMS gene is highly stable, so only few variants with some functional effect were reported, even in tumor tissues; so attention has turned to the variants present in the noncoding regions of the gene [8]. The potential use of variants in the 5′UTR (untranslated region) and 3′UTR flanks has been investigated in several studies of the TYMS gene as genetic biomarkers to predict clinical response in gastrointestinal tumors. However, the results are conflicting and there is no consensus on the clinical value of these variants [9], [10], [11]. For instance, while some studies report association between a favorable clinical response and variants at the 5′UTR and 3′UTR regulatory regions of the TYMS gene, other reports have failed to replicate such results [12], [13], [14], [15], [16]. None of these studies was performed on the Latin-American populations and studies in this group may contribute to the understanding of the biological significance of TYMS variants and their pharmacological relevance for CRC patients undergoing chemotherapy.

The present study deals with the identification of germ-line variants in noncoding regions of the TYMS gene in Mexican patients with CRC from two independent analyses: retrospective and prospective studies, aimed at evaluating the predictive power of these variants in the development of serious toxicity events and lack of response to first-line chemotherapeutic scheme based on fluoropyrimidines for advanced colorectal tumors. Additionally, we analyzed the expression of the TYMS gene in fresh samples of tumor and nontumor tissues of colorectal origin and their association with these variants.

Materials and methods

Study design and patients

We reviewed clinical records and tested samples of 99 patients with advanced CRC (stages III–IV according to the TNM classification) receiving 5-FU or capecitabine in the High Specialties Unit N°25 of the Mexican Institute for Social Security (IMSS) during 2010–2012 for retrospective analyses. These study subjects participated in the genome wide association study entitled “Genetic study of Common Hereditary Bowel Cancers in Hispania and the Americas” (FP7 Program). In addition, 68 patients with similar clinic characteristics and chemotherapeutic regime were recruited for a prospective study to analyze treatment toxicity and clinical response during 2013–2015 in the same hospital. Clinical protocols were approved by the respective Ethics Committees of the following participating institutions: University Hospital of Universidad Autonoma de Nuevo Leon (approval code BI13-002), High Specialties Unit N°25, IMSS and West National Medical Center, IMSS (approval code R-2015-785-024). All patients signed an informed consent for their connection with the described studies. Categorization of objective clinical response was performed following the Guide to Response Evaluation Criteria in Solid Tumors (RECIST v1.1, 2009) and adverse events were evaluated by the Guide to Common Terminology Criteria for Adverse Events (CTCAE v4.03, 2010). Patients for whom they could not get complete clinical information or whose biological samples did not have the necessary quality for subsequent analysis were excluded. After examining the clinical data and genomic DNA quality, 105 patients were selected for these pharmacogenetic studies (42 and 63 from the two recruiting campaigns).


Patients received the FOLFOX scheme consisting of IV infusion of 400 mg/m2 leucovorin and 100 mg/m2 oxaliplatin in 2 h and a subsequent bolus of 400 mg/m2 5-FU, followed by continuous infusion of 2400–3000 mg/m2 5-FU in 48 h every 2 weeks. As an alternative, some patients received the XELOX scheme consisting of IV infusion of 130 mg/m2 oxaliplatin every 3 weeks and 1000 mg capecitabine twice daily after meals.

The general physical condition of each patient was determined prior to the therapy according to the Eastern Cooperative Oncology Group (ECOG) scale and laboratory tests were performed to assess the baseline liver and kidney functions [17].


Genomic DNA was prepared from blood samples anticoagulated with etheylenediaminetetraacetic acid and stored at −20 °C until further processing. The DNA was isolated with the DNeasy Blood & Tissue Kit (QIAGEN, Hilden, Germany) following manufacturer instructions. The variable number of tandem repeats (VNTR) variants of the 5′UTR flank of TYMS (rs45445694), and their additional single nucleotide variant (SNV) G>C in the first repeat of the 2R allele (rs183205964, named 2RG or 2RC), the SNV G>C in the second repeat of the 3R allele (rs2853542, named 3RG or 3RC), and the 6 bp insertion in the second repeat of the 3R allele (rs538469385) (all located into the rs45445694 variant) were performed by polymerase chain reaction-restriction fragment length polymorphism using the primers 5012F19:5′-CGGGAAAAGGCGCGCGGAA-3′ and 5147R18: 5′-GCAGCTCCGAGCCGGCCA-3 and HaeIII (New England Biolabs, Ipswich, MA, USA). A 6-bp deletion variant at the 3′UTR region of TYMS (rs151264360) was genotyped as previously reported [18]. Genotyping analyses were verified in duplicate for each specimen.

Quantitative TYMS expression analysis

All samples were blind evaluated by a pathologist to separate cancerous and normal fractions. Tumor and adjacent tissue samples were recovered from surgical specimens preserved in Tissue-Tek® Optimal Cutting Temperature (Sakura Finetek, Torrance, CA, USA) at −80 °C until processing. Some 10 μm slices cut at −20 °C were immediately processed for gDNA as previously reported [19] and for total RNA using the TRizol® reagent (Invitrogen, Carlsbad, CA, USA). The quality of the RNA was assessed by electrophoresis and spectrophotometry. The cDNA was synthesized with the SuperScript® III First-Strand Synthesis kit (Invitrogen). The quantitative polymerase chain reaction study was performed using the SYBR® Green Master Mix (Bio-Rad. Hercules, CA, USA) in the CFX96 Real Time System (Bio-Rad) using the primers T2211F: 5′-TACCTGGGGCAGATCCAACACATC-3′ and T2211R: 5′-ACAGCAACTCCTCCAAAACACCC-3. The B2M gene expression was selected for normalization. Expression studies were performed in triplicate for both tested genes. Quantification of the relative expression of the TYMS gene was performed by the 2[-Delta Delta C(T)] method previously reported [20].

Statistical analysis

Deviations from the Hardy-Weinberg law were evaluated by the Pearson’s χ2-test. Univariate analyses were performed with one-tailed Fisher’s exact test and the Cochran-Armitage trend test. The likelihood ratio test was employed to compare variant effects in different genetic models using the SNPstats program [21]. The selection of the best model was based on the joint analysis of the Akaike information criterion and the Bayesian information criterion. Multivariate logistic regression-based analysis was used for the adjustment of the genetic model selected with clinical or demographic covariates. Comparison of TYMS gene expression between tumor and adjacent-tissue samples was performed using the Mann-Whitney U test on linearized relative expression data. Two-tailed p<0.05 values were considered significant in the different performed tests. All statistical tests were performed using the IBM SPSS Statistics software v22 (IBM Corp. Armonk, NY, USA).


Association analysis for TYMS variants and toxicity in the retrospective study

A total of 42 patients from the retrospective study were included in these analysis. The variants analyzed in the present study are described in Table 1. Regarding the rs45445694 VNTR variant, only alleles with 2(2R) and 3(3R) repeats were identified in the study sample. In addition, the G>C SNV inside the second repeat of the 3R variant (rs2853542) and the 6-bp deletion variant located in the 3′UTR region (rs151264360) were also identified. Variants ascertained in this study are shown in Figure 1. All other reported variants were not detected in the study subjects.

Table 1:

List of analyzed variants.

dbSNP IDaChromosome coordinatesbNomenclaturecDescription
rs4544569418:657646LRG_783t1:c.-97_70CCGCGCCACTTGGCCTGCCTCCGTCCCG [1], [2], [3], [4], [7], [8], [9]VNTR of 28 nt in the 5′UTR region of the TYMS gene
rs18320596418:657657LRG_783t1:c.-86G>CG>C transversion in the 12th nt of the first repeat of the 2R allele of the VNTR
rs285354218:657685LRG_783t1:c.58G>CG>C transversion in the 12th nt of the second repeat of the 3R allele of the VNTR
rs53846938518:657697LRG_783t1:c.-46_45insCCCCCGInsertion of 6 nt in the second repeat of the 3R allele of the VNTR
rs15126436018:673444LRG_783t1:c.447_452delTTAAAGDeletion of 6 nt in the 3′UTR region of the TYMS gene. Also named 1494del6
  1. According with Human Genome Variation Society nomenclature recommendations. adbSNP build 147. b RefSeq entry: NC_000018.10 GRCh38.p7. cRefSeq entry: NM_001071.

Figure 1: Topological description of the 5′UTR and 3′UTR variants in the TYMS gene.Bold arrows indicate the direction of the coding sequence. Open squares show the variable number repeat variant (rs45445694) with either 2 or 3 repeat alleles. The G>C variant (rs2853542) is located inside the second repeat of the 3R allele. The dashed square shows the position of the del6bp variant (rs151264360) in the 3′UTR region of the gene.
Figure 1:

Topological description of the 5′UTR and 3′UTR variants in the TYMS gene.

Bold arrows indicate the direction of the coding sequence. Open squares show the variable number repeat variant (rs45445694) with either 2 or 3 repeat alleles. The G>C variant (rs2853542) is located inside the second repeat of the 3R allele. The dashed square shows the position of the del6bp variant (rs151264360) in the 3′UTR region of the gene.

The patients were classified for toxicity according to CTCAE v4.03 and categorized into mild to moderate toxicity (grades 0–2) and serious toxicity (grades 3–4) groups. Some 26% (11/42) of the patients experienced serious toxicity (Table 2). The main causes of toxicity were gastrointestinal toxicity (nausea, vomiting, diarrhoea), followed by neurological toxicity (dysesthesia and sensory neuropathy). Some 7.1% (3/42) of the patients required dose adjustment and treatment was suspended in 11.9% (5/42) of the patients because of severe toxicity. The most important serious toxicities were diarrhoea and sensory neuropathy (14% and 7%, respectively). Death events related to the therapy were not reported.

Table 2:

Distribution of toxicity events in the retrospective and prospective analysis of CRC patients.

Toxicityn (%)a
Retrospective (n=42)Prospective (n=63)
TotalGrades 1–2Grades 3–4TotalGrades 1–2Grades 3–4
Gastrointestinal20 (47.6)13 (30.9)7 (16.7)29 (46.0)20 (31.7)9 (14.5)
Neurological15 (35.7)12 (28.6)3 (7.1)24 (38.1)21 (33.3)3 (4.8)
Hematological4 (9.5)3 (7.1)1 (2.4)15 (23.8)12 (19.0)3 (4.8)
Asthenia10 (23.8)9 (21.4)1 (2.4)16 (25.4)14 (22.2)2 (3.2)
Hand-foot syndrome15 (35.7)14 (33.3)1 (2.4)7 (11.1)7 (11.1)0 (0)
Other9 (21.4)5 (11.9)4 (9.5)3 (4.8)2 (3.2)1 (1.6)
All toxicities42 (100)31 (74.0)11 (26.0)63 (100)45 (71.0)18 (29.0)
  1. Toxicities grading according with CTCAE v4.03. aPercentage with respect to the total population of patients in each study.

The univariate analysis did not show associations among serious toxicity and age, sex, tumor stage or type of administered fluoropyrimidine (Supplemental Table 1). The likelihood ratio test determined that the 2R allele of the rs45445694 variant was significantly associated with an increased risk of serious global toxicity in a recessive genetic model (p=0.023, odds ratio (OR)=8.29, 95%CI 1.25–54.71) (Table 3). The G>C inside the 3R variant (rs 2853542) was not associated with serious toxicity to the fluoropyrimidines (p=0.208, 6% vs. 25%).

Table 3:

Association of the VNTR with severe toxicity to fluoropyirimidine-based chemotherapy.

Genetic modelGenotypen (%)OR (95% CI)p-Valuea
Toxicity (0–2)Toxicity (3–4)
Retrospective study of CRC patients (n=42):
Recessive3R/3R–2R/3R29 (93.5)7 (63.6)1.000.023
2R/2R2 (6.5)4 (36.4)8.29 (1.25–54.71)
Prospective study of CRC patients (n=63):
Recessive3R/3R–2R/3R37 (86)12 (60)1.000.024
2R/2R6 (14)8 (40)4.11 (1.19–14.25)
Pooled data (n=105):
Recessive3R/3R–2R/3R66 (89.2)19 (61.3)1.000.0014
2R/2R8 (10.8)12 (38.7)5.21 (1.86–14.59)
  1. CI, confidence interval. Statistically significant associations are highlighted in bold. aLikelihood ratio test.

Prospective analysis of association between variants and objective response

A total of 63 patients met the necessary conditions for their clinical and genetic evaluations in the prospective study. Patients were categorized into responder (partial or complete response) or nonresponder (stable disease, disease progression, or death due to disease) groups, according to RECIST v1.1. The total objective response rate (ORR) was 29%. Univariate analysis for therapeutic response against age (<57 vs. ≥57 years old), gender (male vs. female), and therapy (chemotherapy vs. chemotherapy+radiotherapy) did not show associations (Supplemental Table 2).

The allele discrimination analyses for this cohort identified the 2R (rs45445694), the G>C SNV inside the 3R allele (rs2853542, called 3RC allele) and the 6 bp deletion in the 3′UTR region of TYMS (rs151264360) in frequencies of 0.48, 0.46 and 0.33, respectively. Again, the G>C SNV inside the first repetition of the 2R allele (rs183205964) and the 6 bp insertion in the TYMS promoter (rs538469385) were not identified. All genotypes for the described TYMS variants were in Hardy-Weinberg equilibrium (Supplemental Table 3).

The association analysis for the VNTR (rs45445694) variant and tumor response rate showed that the 2R allele was associated with a positive patient response to chemotherapy in a recessive genetic model (p=0.05; OR=3.45, 95% CI 1.00–11.99) (Table 4).

Table 4:

Association of the VNTR and del6bp variants with the objective response of CRC patients to fluoropyrimidine-based chemotherapy in the prospective study.

Genetic modelGenotypen (%)OR (95% CI)p-ValueaAICbBICc
VNTR (rs45445694)
Co-dominant3R/3R12 (26.7%)4 (22.2%)1.000.1477.583.9
2R/3R26 (57.8%)7 (38.9%)0.81 (0.20–3.30)
2R/2R7 (15.6%)7 (38.9%)3.00 (0.64–14.02)
Dominant3R/3R12 (26.7%)4 (22.2%)1.000.7179.283.5
2R/3R-2R/2R33 (73.3%)14 (77.8%)1.27 (0.35–4.64)
Recessive3R/3R-2R/3R38 (84.4%)11 (61.1%)1.000.0575.679.9
2R/2R7 (15.6%)7 (38.9%)3.45 (1.00–11.99)
Over-dominant3R/3R-2R/2R19 (42.2%)11 (61.1%)1.000.1777.581.8
2R/3R26 (57.8%)7 (38.9%)0.47 (0.15–1.42)
Log-additive1.82 (0.80–4.15)0.1577.381.6
del6bp (rs151264360)
Co-dominant6bp−/6bp−6 (13.3%)0 (0%)1.000.00771.377.8
6bp+/6bp−24 (53.3%)5 (27.8%)NA (0.00–NA)
6bp+/6bp+15 (33.3%)13 (72.2%)NA (0.00–NA)
Dominant6bp−/6bp−6 (13.3%)0 (0%)1.000.03875.179.4
6bp+/6bp− 6bp+/6bp+39 (86.7%)18 (100%)NA (0.00–NA)
Recessive6bp−/6bp− 6bp+/6bp−30 (66.7%)5 (27.8%)1.000.00571.475.7
6bp+/6bp+15 (33.3%)13 (72.2%)5.20 (1.56–17.32)
Over-dominant6bp−/6bp 6bp+/6bp+21 (46.7%)13 (72.2%)1.000.06275.980.2
6bp+/6bp−24 (53.3%)5 (27.8%)0.34 (0.10–1.10)
Log-additive4.81 (1.56–14.77)0.00269.874.1
  1. NA, not applicable; CI, confidence interval. Statistically significant associations are highlighted in bold. aLikelihood ratio test. bAkaike’s information criterion. cBayesian information criterion.

According to the VNTR and the SNV G>C (rs2853542) of the TYMS gene patients were categorized into genotypes of low (2R/2R, 2R/3RC, 3RC/3RC) and high (2R/3RG, 3RG/3RC, 3RG/3RG) enzyme expression, based on functional analyses previously reported [17]. This categorization showed a significant association between low gene expression variants and positive tumor response (p=0.005, OR=6.84, 95% CI 1.73–27.02). In contrast, the number of 3RG alleles in the genotype significantly correlated with therapeutic failure (p=0.0033) (Figure 2). The 6 bp-(rs151264360) variant at the 3′UTR was significantly associated in an additive genetic model with a positive response to chemotherapy (p=0.02, OR=4.81 for each deleted allele, 95% CI 1.56–14.77) (Table 4).

Figure 2: Lack of response to chemotherapy with the number of 3RG alleles in the cohort of colorectal cancer patients.p=0.0033, Cochran-Armitage trend test.
Figure 2:

Lack of response to chemotherapy with the number of 3RG alleles in the cohort of colorectal cancer patients.

p=0.0033, Cochran-Armitage trend test.

Associations among toxicity and clinical, genetic and demographic factors in the prospective analysis

Some 18 out of 63 of patients who reported toxicity developed serious adverse events due to chemotherapy, according to the CTCAv4.01 Guide. Ten of these patients required dose adjustment and two required definitive therapy suspension (Table 2). Gastrointestinal type toxicity was the most frequent, followed by neurological and hematological toxicities (46%, 38% and 24%, respectively). No death events were reported in this prospective study.

None of the analyzed clinical and demographic factors (gender, age, type of treatment, drug, and ECOG performance status) showed association with serious toxicity to fluoropyrimidines in the univariate analysis (Supplemental Table 4).

Regarding the rs45445694, rs2853542, and rs151264360 TYMS variants, the 2R allele of rs45445694 was associated with severe toxicity to the fluoropyrimidine-based chemotherapy in a recessive genetic model in the multivariate analysis (p=0.0024, OR=4.11, 95% CI 1.19–14.25) (Table 3). The 3RG and 3RC alleles of rs2853542 were not associated with serious toxicity (27% vs. 22%, p=0.566) suggesting that this variant lacks informational value for toxicity to fluoropyrimidines.

Combined multivariate analysis

The multivariate logistic regression analysis for combined toxicity data pooled from the retrospective study and the prospective study of CRC patients (n=105, α=0.05, statistical power=80%), showed a significant association between the 2R/2R genotype adjusted by age and the risk of severe global toxicity to fluoropyrimidines (p=0.0014, OR=5.21, 95% CI 1.86–14.59) (Table 3). The post-hoc analysis for prediction of chemotherapy toxicity establishes a recessive model for the association of the 2R allele with toxicity. This study ensures significant detection of the toxicity effect in the studied sample.

TYMS gene expression in colorectal tissues

Tumor and adjacent-tissue samples were available from 16 patients for TYMS expression analysis. A trend to lower TYMS expression was observed in tumor tissues, although the comparison did not reach significance (Figure 3A). As expected, no associations between TYMS expression and rate of response to chemotherapy were found. The 2R/2R genotype of the VNTR was significantly associated with a low relative expression of TYMS in tumor tissues (p<0.05) (Figure 3B). Genotypic analyses in paired tumor and adjacent-tissue samples for the selected TYMS variants showed 100% correspondence, suggesting that none of the tumors presented loss of heterozygosity. No other association was found.

Figure 3: Comparison between TYMS gene relative expression in tumor and nontumor colorectal tissues.(A) Fold change in TYMS relative expression between tumor and nontumor adjacent colorectal tissues. p=0.34, Wilcoxon signed-rank test. Results from three biological replicates. (B) TYMS relative expression by genotype of the VNTR variant in tumor (n=11) and nontumor (n=16) colorectal tissues. p<0.05, Mann-Whitney U-test.
Figure 3:

Comparison between TYMS gene relative expression in tumor and nontumor colorectal tissues.

(A) Fold change in TYMS relative expression between tumor and nontumor adjacent colorectal tissues. p=0.34, Wilcoxon signed-rank test. Results from three biological replicates. (B) TYMS relative expression by genotype of the VNTR variant in tumor (n=11) and nontumor (n=16) colorectal tissues. p<0.05, Mann-Whitney U-test.


The TYMS enzyme is the pharmacological target of the fluoropyrimidines and is mainly responsible for the cytotoxic effect. It is reasonable that variants in the regulatory regions of TYMS may affect the gene expression and are potential candidates for predicting clinical response and toxicity to fluoropyrimidines [22], [23], [24]. Our retrospective analysis of CRC patients shows a significant association between the number of 2R alleles of the VNTR variant in the region 5′UTR of the gene and serious global toxicity to fluoropyrimidines, a result confirmed by the prospective analysis of patients with colorectal tumors. Multivariate analysis of pooled data (prospective and retrospective) shows a recessive model for 2R allele and phenotype of severe toxicity to the fluoropyrimidines with a notorious OR value (5.21). Accordingly, the 2R/2R genotype has been linked to toxicity to fluoropyrimidines [25], [26], [27], [28].

Some studies have related the 2R allele with lower gene expression; however, the relationship between the functional effect of two or more repetitions of this variant and the 5-FU effect is unclear [29], [30], [31].

We also found a significant association between 2R/2R genotype and ORR of colorectal cancer patients to chemotherapy. This result was reported previously in patients with advanced gastrointestinal cancer treated with fluoropyrimidines [32], [33]. The dual relationship between greater toxicity and better response with the genotype 2R/2R is demonstrated in the clinical context [34], [35], [36]. The 2R/2R genotype may potentiate the drug cytotoxic effect by increasing the chemosensitivity of both, tumor and nontumor cells [37].

The variant G>C (rs2853542) showed a tendency towards poorer chemotherapeutic response with each copy of the G allele. Although the clinical effect of this variant has been less studied, our results confirm the report of Meulendijks et al. [33] regarding the negative effect of the G allele on the antineoplastic response. In vitro studies suggest that this allele modifies a putative sequence for the binding of the USF-1 transcription factor, decreasing the gene transcriptional activity. This could be associated with increased tumor sensitivity to the fluoropyrimidines and better clinical response [38], [39]. The absence of this allele in the genotypes classified in the low expression category (i.e. 2R/2R, 2R/3RC, 3RC/3RC) may explain the significant association between better clinical response and low expression genotypes in this study [40]. This result was previously reported for esophageal cancer patients but not for CRC patients [41].

The 6bp+/6bp+ genotype of rs151264360 was associated with good response to fluoropyrimidines, as described by Arrazubi et al. [42] but contrasts with the report by Dotor et al. [43]. There are methodological differences in the treatment administration schemes and sample origins that may explain the discrepancies. Our results also differ with two studies performed on Asian populations. An important difference is that 6bp−/6bp− genotype frequencies described in these studies are higher than those found in our cohort of patients [14], [44]. A comparison of VNTR genotype frequencies in similar studies from different populations shows similarities between Mexican and Caucasian populations from Italy, Hungary and Spain (2R/2R: 16%–24%, 2R/3R: 47%–49%, 3R/3R: 29%–35%), but noticeable differences with East Asian populations (2R/2R: 4%–5%, 2R/3R: 21%–29%, 3R/3R: 67%–73%) [14], [15], [25], [45], [46]. The low frequencies of the 2R allele in the last geographical group may explain the absence of significant association between the VNTR and clinical response in these populations [47]. Similarity between frequencies found in this study with those from Caucasian populations may be explained by the Spaniard contributions to the admixed Mestizo genome in current north-eastern and western Mexicans (38% and 30%, respectively), as was previously established [48], [49].

The TYMS expression studies in our colorectal samples showed decreased gene expression in the 2R/2R tumor tissue samples. This observation agrees with reports indicating better ORR to fluoropyrimidines in patients whose colorectal tumors have lower TYMS expression [50]. Our studies showed 100% genotypic correspondence between tumor and adjacent-tissue samples for the studied TYMS variants, suggesting that none of the tumors presented loss of heterozygosity as previously reported, particularly for the rs45445694 VNTR variant [51], [52]. Functional studies have reported in vitro and in vivo effects to the G>C (rs2853542) and VNTR variant in TYMS expression [38], [53], [54].

Regarding the 6 bp allele in the 3′UTR region of the gene, we did not find association between this allele and TYMS expression. This allele is related with decreased mRNA stability in vitro studies which results in lower enzyme expression but this finding requires in vivo confirmation [55].

Evidence presented here supports that rs45445694 and rs151264360 are predictive biomarkers of response. However, the role of additional variants in noncoding regions of TYMS to chemotherapeutic response and toxicity to fluoropyrimidines cannot be ruled out [56], [57].


The 2R/2R genotype of the VNTR (rs45445694) in the promoter region of the TYMS gene is a predictive marker of clinical response and severe toxicity to the fluoropyrimidines in CRC patients. It is possible that this genotype decreases the expression of the gene in the colorectal tumor tissue. The G variant in rs2853542 and the 6bp deletion in rs151264360 (3′UTR region) are independent biomarkers of therapeutic failure.

Corresponding authors: Oralia Barboza-Quintana, MD, DMed, Universidad Autonoma de Nuevo Leon, Service of Anatomic Pathology and Cytopathology, University Hospital, Ave. F.I. Madero S/N, Col. Mitras Centro, Monterrey, C.P. 64460, Mexico, Phone: +52-81–83294154, and Augusto Rojas-Martinez, MD, DSc, Tecnologico de Monterrey, School of Medicine and Health Sciences, CITES Building – 3rd Floor, Ave. Morones Prieto 3000 Pte, Col. Doctores, Monterrey, C.P. 64710, Mexico, Phone: +52-81-88882117

  1. Author contributions: All the authors have accepted responsibility for the entire content of this submitted manuscript and approved submission.

  2. Research funding: CCR is a PhD student who is supported with funds from the Colombian Department of Science, Technology and Innovation (COLCIENCIAS). Funding for this study was provided by the National Council of Science and Technology (CONACYT) with the registry code SALUD-2013-01-202661.

  3. Employment or leadership: None declared.

  4. Honorarium: None declared.

  5. Competing interests: The funding organization played no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the report for publication.


1. Ferlay J, Soerjomataram I, Ervik M, Dikshit R, Eser S, Mathers C, et al. GLOBOCAN 2012 v1.0, cancer incidence and mortality worldwide: IARC CancerBase No. 11[Internet] Lyon, France: International Agency for Research on Cancer; 2013 [Accessed on 06-01-2017]. Available from: in Google Scholar

2. Loree JM, Cheung WY. Optimizing adjuvant therapy and survivorship care of stage III colon cancer. Future Oncol 2016;12:2021–35.10.2217/fon-2016-0109Search in Google Scholar PubMed

3. Duffy MJ, Lamerz R, Haglund C, Nicolini A, Kalousova M, Holubec L, et al. Tumor markers in colorectal cancer, gastric cancer and gastrointestinal stromal cancers: European group on tumor markers 2014 guidelines update. Int J Cancer 2014;134:2513–22.10.1002/ijc.28384Search in Google Scholar PubMed PubMed Central

4. Lech G, Slotwinski R, Slodkowski M, Krasnodebski IW. Colorectal cancer tumour markers and biomarkers: Recent therapeutic advances. World J Gastroenterol 2016;22:1745–55.10.3748/wjg.v22.i5.1745Search in Google Scholar PubMed PubMed Central

5. Braun MS, Seymour MT. Balancing the efficacy and toxicity of chemotherapy in colorectal cancer. Ther Adv Med Oncol 2011;3:43–52.10.1177/1758834010388342Search in Google Scholar PubMed PubMed Central

6. Lee AM, Shi Q, Pavey E, Alberts SR, Sargent DJ, Sinicrope FA, et al. DPYD variants as predictors of 5-fluorouracil toxicity in adjuvant colon cancer treatment (NCCTG N0147). J Natl Cancer Inst 2014;106:dju298.10.1093/jnci/dju298Search in Google Scholar PubMed PubMed Central

7. Morel A, Boisdron-Celle M, Fey L, Soulie P, Craipeau MC, Traore S, et al. Clinical relevance of different dihydropyrimidine dehydrogenase gene single nucleotide polymorphisms on 5-fluorouracil tolerance. Mol Cancer Ther 2006;5:2895–904.10.1158/1535-7163.MCT-06-0327Search in Google Scholar PubMed

8. Calascibetta A, Contino F, Feo S, Gulotta G, Cajozzo M, Antona A, et al. Analysis of the thymidylate synthase gene structure in colorectal cancer patients and its possible relation with the 5-fluorouracil drug response. J Nucleic Acids 2010;2010:306754.10.4061/2010/306754Search in Google Scholar PubMed PubMed Central

9. Schwarzenbach H. Predictive diagnostics in colorectal cancer: impact of genetic polymorphisms on individual outcomes and treatment with fluoropyrimidine-based chemotherapy. EPMA J 2010;1:485–94.10.1007/s13167-010-0022-5Search in Google Scholar PubMed PubMed Central

10. Gusella M, Padrini R. G>C SNP of thymidylate synthase with respect to colorectal cancer. Pharmacogenomics 2007;8:985–96.10.2217/14622416.8.8.985Search in Google Scholar PubMed

11. Lima A, Azevedo R, Sousa H, Seabra V, Medeiros R. Current approaches for TYMS polymorphisms and their importance in molecular epidemiology and pharmacogenetics. Pharmacogenomics 2013;14:1337–51.10.2217/pgs.13.118Search in Google Scholar PubMed

12. Sulzyc-BV, Bielicki D, Binczak-Kuleta A, Kaczmarczyk M, Pioch W, Machoy-Mokrzynska A, et al. Thymidylate synthase gene polymorphism and survival of colorectal cancer patients receiving adjuvant 5-fluorouracil. Genet Mol Biomarker 2013;17:799–806.10.1089/gtmb.2013.0171Search in Google Scholar PubMed PubMed Central

13. Afzal S, Gusella M, Jensen SA, Vainer B, Vogel U, Andersen JT, et al. The association of polymorphisms in 5-fluorouracil metabolism genes with outcome in adjuvant treatment of colorectal cancer. Pharmacogenomics 2011;12:1257–67.10.2217/pgs.11.83Search in Google Scholar PubMed

14. Kumamoto K, Ishibashi K, Okada N, Tajima Y, Kuwabara K, Kumagai Y, et al. Polymorphisms of GSTP1, ERCC2 and TS-3′UTR are associated with the clinical outcome of mFOLFOX6 in colorectal cancer patients. Oncol Lett 2013;6:648–54.10.3892/ol.2013.1467Search in Google Scholar PubMed PubMed Central

15. Hitre E, Budai B, Adleff V, Czegledi F, Horvath Z, Gyergyay F, et al. Influence of thymidylate synthase gene polymorphisms on the survival of colorectal cancer patients receiving adjuvant 5-fluorouracil. Pharmacogenet Genomics 2005;15:723–30.10.1097/01.fpc.0000175598.42141.59Search in Google Scholar PubMed

16. Park CM, Lee WY, Chun HK, Cho YB, Yun HR, Heo JS, et al. Relationship of polymorphism of the tandem repeat sequence in the thymidylate synthase gene and the survival of stage III colorectal cancer patients receiving adjuvant 5-flurouracil-based chemotherapy. J Surg Oncol 2010;101:22–7.10.1002/jso.21412Search in Google Scholar PubMed

17. Oken MM, Creech RH, Tormey DC, Horton J, Davis TE, McFadden ET, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol 1982;5:649–55.10.1097/00000421-198212000-00014Search in Google Scholar PubMed

18. Ulrich CM, Bigler J, Velicer CM, Greene EA, Farin FM, Potter JD. Searching expressed sequence tag databases: discovery and confirmation of a common polymorphism in the thymidylate synthase gene. Cancer Epidemiol Biomarkers Prev 2000;9:1381–5.Search in Google Scholar PubMed

19. Aljanabi SM, Martinez I. Universal and rapid salt-extraction of high quality genomic DNA for PCR-based techniques. Nucleic Acids Res 1997;25:4692–3.10.1093/nar/25.22.4692Search in Google Scholar PubMed PubMed Central

20. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 2001;25:402–8.10.1006/meth.2001.1262Search in Google Scholar PubMed

21. Sole X, Guino E, Valls J, Iniesta R, Moreno V. SNPStats: a web tool for the analysis of association studies. Bioinformatics 2006;22:1928–9.10.1093/bioinformatics/btl268Search in Google Scholar PubMed

22. Longley DB, Harkin DP, Johnston PG. 5-fluorouracil: mechanisms of action and clinical strategies. Nat Rev Cancer 2003;3:330–8.10.1038/nrc1074Search in Google Scholar PubMed

23. Pullarkat ST, Stoehlmacher J, Ghaderi V, Xiong YP, Ingles SA, Sherrod A, et al. Thymidylate synthase gene polymorphism determines response and toxicity of 5-FU chemotherapy. Pharmacogenomics J 2001;1:65–70.10.1038/sj.tpj.6500012Search in Google Scholar PubMed

24. Koda K, Miyauchi H, Kosugi C, Kaiho T, Takiguchi N, Kobayashi S, et al. Tumor 5-FU-related mRNA expression and efficacy of oral fluoropyrimidines in adjuvant chemotherapy of colorectal cancer. Anticancer Res 2016;36:5325–31.10.21873/anticanres.11105Search in Google Scholar PubMed

25. Romiti A, Roberto M, D’Antonio C, Onesti CE, Barucca V, Milano A, et al. The TYMS-TSER polymorphism is associated with toxicity of low-dose capecitabine in patients with advanced gastrointestinal cancer. Anti-cancer Drugs 2016;27:1044–9.10.1097/CAD.0000000000000429Search in Google Scholar PubMed

26. Wilks AB, Saif MW. First case of foot drop associated with capecitabine in a patient with thymidylate synthase polymorphism. Cureus 2017;9:e995.10.7759/cureus.995Search in Google Scholar PubMed PubMed Central

27. Lecomte T, Ferraz JM, Zinzindohoue F, Loriot MA, Tregouet DA, Landi B, et al. Thymidylate synthase gene polymorphism predicts toxicity in colorectal cancer patients receiving 5-fluorouracil-based chemotherapy. Clin Cancer Res 2004;10:5880–8.10.1158/1078-0432.CCR-04-0169Search in Google Scholar PubMed

28. Wang B, Walsh SJ, Saif MW. Pancytopenia and severe gastrointestinal toxicities associated with 5-fluorouracil in a patient with thymidylate synthase (TYMS) polymorphism. Cureus 2016;8:e798.10.7759/cureus.798Search in Google Scholar PubMed PubMed Central

29. Kawakami K, Omura K, Kanehira E, Watanabe Y. Polymorphic tandem repeats in the thymidylate synthase gene is associated with its protein expression in human gastrointestinal cancers. Anticancer Res 1999;19:3249–52.Search in Google Scholar PubMed

30. Popat S, Wort R, Houlston RS. Relationship between thymidylate synthase (TS) genotype and TS expression: a tissue microarray analysis of colorectal cancers. Int J Surg Pathol 2005;13:127–33.10.1177/106689690501300201Search in Google Scholar PubMed

31. Morganti M, Ciantelli M, Giglioni B, Putignano AL, Nobili S, Papi L, et al. Relationships between promoter polymorphisms in the thymidylate synthase gene and mRNA levels in colorectal cancers. Eur J Cancer 2005;41:2176–83.10.1016/j.ejca.2005.06.016Search in Google Scholar PubMed

32. Park DJ, Stoehlmacher J, Zhang W, Tsao-Wei D, Groshen S, Lenz HJ. Thymidylate synthase gene polymorphism predicts response to capecitabine in advanced colorectal cancer. Int J Colorectal Dis 2002;17:46–9.10.1007/s003840100358Search in Google Scholar PubMed

33. Meulendijks D, Rozeman EA, Cats A, Sikorska K, Joerger M, Deenen MJ, et al. Pharmacogenetic variants associated with outcome in patients with advanced gastric cancer treated with fluoropyrimidine and platinum-based triplet combinations: a pooled analysis of three prospective studies. Pharmacogenomics J 2017;17:441–51.10.1038/tpj.2016.81Search in Google Scholar PubMed

34. Soveri LM, Hermunen K, de Gramont A, Poussa T, Quinaux E, Bono P, et al. Association of adverse events and survival in colorectal cancer patients treated with adjuvant 5-fluorouracil and leucovorin: Is efficacy an impact of toxicity? Eur J Cancer 2014;50:2966–74.10.1016/j.ejca.2014.08.017Search in Google Scholar PubMed

35. Jennings BA, Kwok CS, Willis G, Matthews V, Wawruch P, Loke YK. Functional polymorphisms of folate metabolism and response to chemotherapy for colorectal cancer, a systematic review and meta-analysis. Pharmacogenet Genomics 2012;22:290–304.10.1097/FPC.0b013e328351875dSearch in Google Scholar PubMed

36. Shahrokni A, Rajebi MR, Saif MW. Toxicity and efficacy of 5-fluorouracil and capecitabine in a patient with TYMS gene polymorphism: a challenge or a dilemma? Clin Colorectal Cancer 2009;8:231–4.10.3816/CCC.2009.n.039Search in Google Scholar PubMed

37. Yawata A, Kim SR, Miyajima A, Kubo T, Ishida S, Saito Y, et al. Polymorphic tandem repeat sequences of the thymidylate synthase gene correlates with cellular-based sensitivity to fluoropyrimidine antitumor agents. Cancer Chemother Pharmacol 2005;56:465–72.10.1007/s00280-005-1018-zSearch in Google Scholar PubMed

38. Mandola MV, Stoehlmacher J, Muller–Weeks S, Cesarone G, Yu MC, Lenz HJ, et al. A novel single nucleotide polymorphism within the 5′ tandem repeat polymorphism of the thymidylate synthase gene abolishes USF-1 binding and alters transcriptional activity. Cancer Res 2003;63:2898–904.Search in Google Scholar PubMed

39. Pare L, Marcuello E, Altes A, del Rio E, Sedano L, Barnadas A, et al. Transcription factor-binding sites in the thymidylate synthase gene: predictors of outcome in patients with metastatic colorectal cancer treated with 5-fluorouracil and oxaliplatin? Pharmacogenomics J 2008;8:315–20.10.1038/sj.tpj.6500469Search in Google Scholar PubMed

40. Kawakami K, Watanabe G. Identification and functional analysis of single nucleotide polymorphism in the tandem repeat sequence ofthymidylate synthase gene. Cancer Res 2003;63:6004–7.Search in Google Scholar PubMed

41. Kuramochi H, Tanaka K, Oh D, Lehman BJ, Dunst CM, Yang DY, et al. Thymidylate synthase polymorphisms and mRNA expression are independent chemotherapy predictive markers in esophageal adenocarcinoma patients. Int J Oncol 2008;32:201–8.10.3892/ijo.32.1.201Search in Google Scholar PubMed

42. Arrazubi V, Suarez J, Guerrero D, Gomez M, Viudez A, Arias F, et al. Prognostic significance of thymidylate synthase polymorphisms in rectal cancer patients treated with neoadjuvant chemoradiotherapy. Colorectal Dis 2013;15:428–35.10.1111/codi.12009Search in Google Scholar PubMed

43. Dotor E, Cuatrecases M, Martinez IM, Navarro M, Vilardell F, Guino E, et al. Tumor thymidylate synthase 1494del6 genotype as a prognostic factor in colorectal cancer patients receiving fluorouracil-based adjuvant treatment. J Clin Oncol 2006;24:1603–11.10.1200/JCO.2005.03.5253Search in Google Scholar PubMed

44. Gao J, He Q, Hua D, Mao Y, Li Y, Shen L. Polymorphism of TS 3′-UTR predicts survival of Chinese advanced gastric cancer patients receiving first-line capecitabine plus paclitaxel. Clin Transl Oncol 2013;15:619–25.10.1007/s12094-012-0979-8Search in Google Scholar PubMed

45. Huang MY, Fang WY, Lee SC, Cheng TL, Wang JY, Lin SR. ERCC2 2251A>C genetic polymorphism was highly correlated with early relapse in high-risk stage II and stage III colorectal cancer patients: a preliminary study. BMC cancer 2008;8:50.10.1186/1471-2407-8-50Search in Google Scholar PubMed PubMed Central

46. Paez D, Pare L, Altes A, Sancho-Poch FJ, Petriz L, Garriga J, et al. Thymidylate synthase germline polymorphisms in rectal cancer patients treated with neoadjuvant chemoradiotherapy based on 5-fluorouracil. J Cancer Res Clin Oncol 2010;136:1681–9.10.1007/s00432-010-0826-7Search in Google Scholar PubMed

47. Patel JN. Cancer pharmacogenomics: implications on ethnic diversity and drug response. Pharmacogenet Genomics 2015;25:223–30.10.1097/FPC.0000000000000134Search in Google Scholar PubMed

48. Martinez FML, Beuten J, Leach RJ, Parra EJ, Cruz-Lopez M, Rangel VH, et al. Ancestry informative markers and admixture proportions in northeastern Mexico. J Hum Genet 2009;54:504–9.10.1038/jhg.2009.65Search in Google Scholar PubMed

49. Martinez MV, Valladares A, Cameron E, Chan A, Perera A, Globus GR, et al. Admixture in Mexico City: implications for admixture mapping of type 2 diabetes genetic risk factors. Hum Genet 2007;120:807–19.10.1007/s00439-006-0273-3Search in Google Scholar PubMed

50. Qiu LX, Tang QY, Bai JL, Qian XP, Li RT, Liu BR, et al. Predictive value of thymidylate synthase expression in advanced colorectal cancer patients receiving fluoropyrimidine-based chemotherapy: evidence from 24 studies. Int J Cancer 2008;123:2384–9.10.1002/ijc.23822Search in Google Scholar PubMed

51. Uchida K, Hayashi K, Kawakami K, Schneider S, Yochim JM, Kuramochi H, et al. Loss of heterozygosity at the thymidylate synthase (TS) locus on chromosome 18 affects tumor response and survival in individuals heterozygous for a 28-bp polymorphism in the TS gene. Clin Cancer Res 2004;10:433–9.10.1158/1078-0432.CCR-0200-03Search in Google Scholar

52. Kawakami K, Ishida Y, Danenberg KD, Omura K, Watanabe G, Danenberg PV. Functional polymorphism of the thymidylate synthase gene in colorectal cancer accompanied by frequent loss of heterozygosity. Jpn J Cancer Res 2002;93:1221–9.10.1111/j.1349-7006.2002.tb01227.xSearch in Google Scholar PubMed PubMed Central

53. Horie N, Aiba H, Oguro K, Hojo H, Takeishi K. Functional analysis and DNA polymorphism of the tandemly repeated sequences in the 5′-terminal regulatory region of the human gene forthymidylate synthase. Cell Struct Funct 1995;20:191–7.10.1247/csf.20.191Search in Google Scholar PubMed

54. de Bock CE, Garg MB, Scott N, Sakoff JA, Scorgie FE, Ackland SP, et al. Association of thymidylate synthase enhancer region polymorphisms with thymidylate synthase activity in vivo. Pharmacogenomics J 2011;11:307–14.10.1038/tpj.2010.43Search in Google Scholar PubMed

55. Mandola MV, Stoehlmacher J, Zhang W, Groshen S, Yu MC, Iqbal S, et al. A 6 bp polymorphism in the thymidylate synthase gene causes message instability and is associated with decreased intratumoral TS mRNA levels. Pharmacogenetics 2004;14: 319–27.10.1097/00008571-200405000-00007Search in Google Scholar PubMed

56. Kim SR, Ozawa S, Saito Y, Kurose K, Kaniwa N, Kamatani N, et al. Fourteen novel genetic variations and haplotype structures of the TYMS gene encoding human thymidylate synthase (TS). Drug Metab Pharmacokinet 2006;21:509–16.10.2133/dmpk.21.509Search in Google Scholar PubMed

57. Ghosh S, Hossain MZ, Borges M, Goggins MG, Ingersoll RG, Eshleman JR, et al. Analysis of polymorphisms and haplotype structure of the human thymidylate synthase genetic region: a tool for pharmacogenetic studies. PLoS One 2012;7:e34426.10.1371/journal.pone.0034426Search in Google Scholar PubMed PubMed Central

Supplemental Material:

The online version of this article offers supplementary material (

Received: 2017-9-19
Accepted: 2017-11-15
Published Online: 2017-12-19
Published in Print: 2017-12-20

©2017 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 30.3.2023 from
Scroll Up Arrow