Association between miRNA expression profiles and polymorphisms of dihydropyrimidine dehydrogenase drug‑metabolizing gene in patients with colorectal cancer receiving 5‑fluorouracil

Introduction

5-Fluorouracil (5-FU) is a key chemotherapeutic agent in colorectal cancer (CRC) treatment, primarily acting by inhibiting thymidylate synthase, which disrupts DNA synthesis, and by incorporating toxic metabolites into RNA to impair cell functions (1). Its efficacy is enhanced in combination regimens such as leucovorin, 5-FU and oxaliplatin) and FOLFIRI (leucovorin, 5-FU, and Irinotecan), often alongside leucovorin, stabilizing its cytotoxic effects (2). Despite its effectiveness, 5-FU treatment is frequently limited by resistance mechanisms and adverse effects, such as myelosuppression and gastrointestinal toxicity, highlighting the importance of biomarkers and personalized strategies to predict treatment response and mitigate side effects (3).

Advances in pharmacogenomics and combinatorial approaches continue to refine the use of 5-FU, offering improved outcomes for patients with CRC (4). The efficacy and safety of 5-FU are notably influenced by interindividual variability in drug metabolism, largely attributed to genetic polymorphisms in the dihydropyrimidine dehydrogenase (DPYD) gene, which encodes dihydropyrimidine dehydrogenase (DPD), the key enzyme responsible for 5-FU catabolism (5). Deficient or decreased DPD activity, caused by DPYD polymorphisms such as DPYD 85T>C and DPYD 1896T>C, leads to the accumulation of toxic 5-FU metabolites, resulting in severe and sometimes fatal toxicity, including neutropenia, diarrhea and mucositis (6,7).

The three DPYD variants DPYD 85T>C, DPYD 1627A>G, and DPYD 1896T>C are relatively more prevalent in East and Southeast Asian populations compared with European risk alleles such as DPYD*2A (c.1905+1G>A) and DPYD 13 (c.1679T>G) (8). Previous pharmacogenetic studies, including analyses in Thai cohorts and population databases, have consistently demonstrated higher allele frequencies for these variants in Asian populations (9,10). Each variant has potential functional relevance: DPYD 85T>C is associated with decreased DPD enzymatic activity and variable risk of fluoropyrimidine-related toxicity; DPYD 1627A>G represents a missense substitution with conflicting reports on its functional consequences but occurs at a relatively high frequency in Asian populations and DPYD 1896T>C has been observed in patients with severe 5-FU-induced toxicity, suggesting its role as a putative risk allele (6,7,11). Finally, these variants are not currently incorporated into international dosing guidelines such as those issued by the Clinical Pharmacogenetics Implementation Consortium (4,12). However, evidence (9,10) indicates that they may hold clinical value in non-European populations, thereby justifying further investigation.

MicroRNAs (miRNAs or miRs) are small non-coding RNAs ~22 nucleotides in length. miRNAs regulate gene expression at the post-transcriptional level by mRNA degradation or translational inhibition (13). The key roles of miRNAs involve several biological processes such as cell proliferation, differentiation and apoptosis (14). The aberrant expression of miRNAs has been investigated in various cancer types and related to drug resistance (15-17). Circulating miRNAs may be a non-invasive biomarker for diagnosing and predicting disease progression (18). Ferracin et al (19) found that miR-21-5p expression is high in the plasma of patients with CRC patients. A previous study also showed that miR-145, miR-106a and miR-17-3p were significantly differentially expressed between patients with pre- and post-operative stage II/III CRC. High levels of miR-17-3p and miR-106a are associated with shorter disease-free survival, suggesting they may serve as serum-miRNA-based biomarkers for prognosis and predicting disease recurrence in patients with stage II/III CRC (20). In addition, recent studies have identified altered expression of other miRNAs in patients with CRC patients, including upregulation of miR-374a in both plasma and tumor tissues (21), as well as specific miRNA signatures in saliva and lymphatic samples, which may serve as non-invasive diagnostic biomarkers and help identify patients at high risk of lymph node metastasis (22,23). Offer et al (24) demonstrated that miR-27a and miR-27b may be pharmacological modulators of hepatic DPD enzyme function. DPD is an important enzyme in the uracil catabolic pathway converting the anti-cancer drug 5-FU to the inactive metabolite 5-dihydrofluorouracil. Deficiency of DPD resulting from inadequate expression or deleterious variants in DPYD is associated with severe toxic responses to 5-FU (24).

Moreover, patients who are heterozygous for miR-27a SNV (rs895819) have an increased risk of fluoropyrimidine toxicity, which was investigated in both DPYD wild-type and DPYD variant carriers (25). This suggests that miR-27a rs895819 may serve as a biomarker for fluoropyrimidine-associated toxicity prediction. Previous studies have predominantly focused on common DPYD variants observed in European populations, such as DPYD 2A (c.1905+1G>A), DPYD 13 (c.1679T>G), and HapB3 (c.1236G>A), which are well-established markers of fluoropyrimidine toxicity (8,26). To the best of our knowledge, few studies (24,25) have concurrently assessed germline DPYD polymorphisms alongside miRNA expression profiles to investigate their combined impact on 5-FU toxicity. The present study aimed to determine the miRNA expression profiles in the DPYD variant (DPYD 85T>C and DPYD 1896T>C) in patients with CRC to facilitate use of miRNAs as potential targets for predicting the toxicity of 5-FU and the development of personalized patient profiles and therapeutic interventions.

Materials and methods

Subjects

A total of 48 patients with CRC were recruited between October 2020 and October 2023 at the Division of Oncology, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand. The mean age was 64.7±11.9 years, ranging from 18-90 years. Among the 48 patients 28 (58.3%) were male and 20 (41.7% were female. Inclusion criteria were as follows: CRC confirmed histologically or cytologically; age ≥18 years; no previous treatment with 5-FU; Eastern Cooperative Oncology Group (ECOG) (27) performance status of 0-2; life expectancy >3 months; white blood cell, 4.5-11.0x109/l; hemoglobin, 13.2-16.6 g/dl for male patients and 11.6-15.0 g/dl for female patients; neutrophil count <1.5x109/l; platelet count <8x1010/l and serum creatinine <1.5 mg/dl. Exclusion criteria were pregnancy and any laboratory evidence of renal or hepatic abnormality.

The present study was carried out in compliance with the Declaration of Helsinki and approved by the Ethics Committee of Ramathibodi Hospital, Mahidol University, Thailand (approval no. MURA2020/1613 Ref.2419). The study procedure was explained to the patients before the study andall patients signed the consent form to participate in the study.

Molecular analysis

EDTA blood samples were collected to perform DNA extraction using MagNA Pure Compact System (Roche Diagnostics GmbH). TaqMan® real-time (RT)PCR ViiA7™ system (Applied Biosystems; Thermo Fisher Scientific, Inc.) was used to detect three DPYD variants: DPYD 85T>C (rs1801265, cat. no. C_9491497_10), DPYD 1627 A>G (rs1801159, cat. no. C_1823316_20) and DPYD 1896T>C (rs17376848, cat. no. C_25471727_20). All reagents were obtained from Applied Biosystems, Thermo Fisher Scientific, Inc., and used according to the manufacturer's instructions.

Toxicity assessment

The Common Terminology Criteria for Adverse Events (CTCAE) v5.0(28) was used to evaluate toxicity at first and second cycles of treatment. The hematological toxicity included leukopenia, neutropenia, thrombocytopenia and anemia. Grade 1-4 was regarded as toxicity.

miRNA extraction and cDNA synthesis

Plasma samples from patients with CRC were used for miRNA extraction. Prior to extraction, plasma was centrifuged at 1,107 x g for 15 min at room temperature. Following the manufacturer's protocol, 100 µl plasma was processed using the miRNeasy® Serum/Plasma kit (cat. no. 217184; Qiagen GmbH) for miRNA isolation. The RNA concentration was measured using a NanoDrop 2000/2000c Spectrophotometer (Thermo Fisher Scientific, Inc.). To synthesize cDNA, 200 ng RNA was used with the miRCURY LNA™ RT kit (cat. no. 339340; Qiagen GmbH). The reaction mixture included 5X miRCURY SYBR Green RT reaction buffer, 10X miRCURY RT enzyme mix, Uni Sp6 RNA spike-in and the RNA template. The reverse transcription reaction was performed in a thermal cycler at 42˚C for 60 min, followed by enzyme inactivation at 95˚C for 5 min and a final hold at 4˚C. cDNA was then stored at -20˚C until further use.

miRNA array

The miRNA expression profiles were analyzed using RT-PCR with a customized miRNA array panel (cat. no. 217184; Qiagen GmbH) coated with specific primers for 43 miRNAs (sequences not provided) involved in the 5-FU metabolic pathway. miR target sequences are listed in Table SI. The reactions were performed using the miRCURY LNA™ SYBR Green PCR kit (cat. no. 339345, Qiagen GmbH). cDNA was diluted to 1:80 before mixing in the reaction containing 2X miRCURY SYBRGreen mix and nuclease-free water. The 96-well plate of the miRNA array was subjected to RT-PCR (Bio-Rad Laboratories, Inc.; cat. no. CFX 96). Thermocycling conditions were as follows: Initial heat activation at 95˚C for 2 min, followed by 50 cycles of denaturation at 95˚C for 10 sec and annealing/extension at 56˚C for 60 sec. The relative miRNA expression was determined using the 2-ΔΔCq method (29). The analysis of miRNA profiles was performed using the Qiagen web portal at GeneGlobe (geneglobe.qiagen.com). Heatmaps were generated using the GeneGlobe platform (geneglobe.qiagen.com/th, which applies an unsupervised clustering algorithm. Unsupervised clustering groups both samples and genes based on the similarity of their expression patterns, rather than relying on predefined group labels. This approach minimizes bias introduced by prior assumptions. This method was selected because it is widely regarded as a standard for gene expression analysis and provides an unbiased visualization of expression profiles (30,31).

Statistical analysis

Hardy-Weinberg equilibrium of DPYD was assessed using the χ2 test. Data were assessed for normality of distribution. Descriptive statistics for patients were presented as mean ± standard deviation (SD) for variables with a normal distribution. The association between DPYD variants status and the hematological toxicity was evaluated by Fisher's exact test. All tests were performed using SPSS software version 21.0 (IBM Corp.). P-values were calculated using unpaired Student's t-test for each miRNA. The test was performed as a parametric, unpaired, two-sample t-test assuming equal variance with a two-tailed distribution. The present study was designed as an exploratory screen to identify candidate miRNAs for further validation. Therefore, no multiplicity adjustment was performed. P<0.05 was considered to indicate a statistically significant difference.

Results

Patient demographics and clinical data

A total of 48 patients were included in the present study. The mean age was 64.7±11.9 years, with a male predominance (58.3%, n=28 vs. 41.7%, n=20). Most patients had a performance status, assessed by ECOG score, of 0 or 1, with a wide distribution of primary tumor locations, including the rectum, sigmoid colon and other sites across the colon. Metastases were most commonly observed at sites other than the liver and lung (22.9%, n=11), with the liver being the second most frequent site (18.8%, n=9). Tumors were predominantly moderately differentiated (Table I).

Table I Patient characteristics (n=48).

Table I

Patient characteristics (n=48).

Characteristic	Value
Mean age, years	64.7±11.9
Sex (%)
Male	28 (58.3)
Female	20 (41.7)
ECOG score (%)
0	21 (43.8)
1	22 (45.8)
2	5 (10.4)
Site of cancer (%)
Ascending colon	5 (10.4)
Transverse colon	3 (6.3)
Descending colon	4 (8.3)
Sigmoid colon	14 (29.2)
Rectosigmoid	5 (10.4)
Rectum	16 (33.3)
Colon, unspecified	1 (2.1)
Site of metastasis (%)
Liver	9 (18.8)
Lung	3 (6.3)
Liver and lung	2 (4.2)
Other	11 (22.9)
No	23 (47.9)
Histopathology (%)
Well differentiated	11 (22.9)
Moderately differentiated	36 (75.0)
Poorly differentiated	1 (2.1)

[i] ECOG, Eastern Cooperative Oncology Group.

Genotypic and allelic frequencies of DPYD variants

A total of three single nucleotide polymorphisms (SNPs) in the DPYD gene were analyzed: 85T>C (rs1801265), 1627A>G (rs1801159), and 1896T>C (rs17376848). All three variants were present in the study cohort, with genotype distributions consistent with Hardy-Weinberg equilibrium.

For the DPYD 85T>C polymorphism, the majority of individuals were homozygous wild-type (72.9%), while 27.1% were heterozygous; no homozygous variant carriers were observed. The corresponding allele frequencies were 0.86 for the wild-type allele and 0.14 for the variant allele.

For the DPYD 1627A>G variant, 70.8% of subjects carried the homozygous wild-type genotype, 25.0% were heterozygous, and 4.2% were homozygous for the variant allele, yielding allele frequencies of 0.83 (wild-type) and 0.17 (variant).

Similarly, the DPYD 1896T>C polymorphism demonstrated 72.9% homozygous wild-type, 27.1% heterozygous, and no homozygous variant genotypes, corresponding to allele frequencies of 0.86 and 0.14 for the wild-type and variant alleles, respectively (Table II).

Table II Dihydropyrimidine dehydrogenase genotype and allele frequency.

Table II

Dihydropyrimidine dehydrogenase genotype and allele frequency.

		Genotype frequency (%)			Allele frequency
SNP	rs ID no.	wt/wt	wt/vt	vt/vt	wt	vt
85T>C	1801265	35 (72.9)	13 (27.1)	0 (0.0)	0.86	0.14
1627A>G	1801159	34 (70.8)	12 (25.0)	2 (4.2)	0.83	0.17
1896 T>C	17376848	35 (72.9)	13 (27.1)	0 (0.0)	0.86	0.14

[i] wt/wt, homozygous wild-type; wt/vt, heterozygous variant; vt/vt, homozygous variant; SNP, single nucleotide polymorphism; rs, Reference SNP cluster ID.

Association of DPYD variants with hematological toxicity

Hematological toxicity, including anemia, leucopenia, neutropenia and thrombocytopenia, were assessed for their association with DPYD variants across two chemotherapy cycles (Table III). DPYD 85T>C showed a trend toward increased anemia, particularly during the second chemotherapy cycle. Although the difference was not significant, this suggests a possible cumulative effect of the variant on drug metabolism over time. By contrast, DPYD 1627A>G and DPYD 1896T>C were not associated with increased anemia in either cycle.

Table III Dihydropyrimidine dehydrogenase variants and hematological toxicity.

Table III

Dihydropyrimidine dehydrogenase variants and hematological toxicity.

				First cycle			Second cycle
Toxicity	SNP	Genotype	n	Grade 0 (%)	Grade 1-4 (%)	P-value	Grade 0 (%)	Grade 1-4 (%)	P-value
Anemia	85T>C	T/T	35	19 (54.3)	16 (45.7)	0.371	21 (60.0)	14 (40.0)	0.070
		T/C	13	5 (38.5)	8 (61.5)		4 (30.8)	9 (69.2)
	1627A>G	A/A	34	18 (52.9)	16 (47.1)	0.492	17 (50.0)	17 (50.0)	0.626
		A/G or G/G	14	6 (42.9)	8 (57.1)		8 (57.1)	6 (42.9)
	1896T>C	T/T	35	17 (48.6)	18 (51.4)	0.727	18 (51.4)	17 (48.6)	0.873
		T/C	13	7 (53.8)	6 (46.2)		7 (53.8)	6 (46.2)
Leucopenia	85T>C	T/T	35	28 (80.0)	7 (20.0)	0.446	27 (77.1)	8 (22.9)	0.064
		T/C	13	12 (92.3)	1 (7.7)		13 (100.0)	0 (0.0)
	1627A>G	A/A	34	28 (82.4)	6 (17.6)	>0.999	28 (82.4)	6 (17.6)	>0.999
		A/G or G/G	14	12 (85.7)	2 (14.3)		12 (85.7)	2 (14.3)
	1896T>C	T/T	35	27 (77.1)	8 (22.9)	0.64	29 (82.9)	6 (17.1)	>0.999
		T/C	13	13 (100.0)	0 (0.0)		11 (84.6)	2 (15.4)
Neutropenia	85T>C	T/T	35	27 (77.1)	8 (22.9)	0.724	25 (71.4)	10 (28.6)	0.498
		T/C	13	11 (84.6)	2 (15.4)		11 (84.6)	2 (15.4)
	1627A>G	A/A	34	26 (76.5)	8 (23.5)	0.724	25 (73.5)	9 (26.5)	>0.999
		A/G or G/G	14	12 (85.7)	2 (14.3)		11 (78.6)	3 (21.4)
	1896T>C	T/T	35	26 (74.3)	9 (25.7)	0.280	26 (74.3)	9 (25.7)	>0.999
		T/C	13	12 (92.3)	1 (7.7)		10 (76.9)	3 (23.1)
Thrombocytopenia	85T>C	T/T	35	33 (94.3)	2 (5.7)	0.583	28 (80.0)	7 (20.0)	0.446
		T/C	13	12 (92.3)	1 (7.7)		12 (92.3)	1 (7.7)
	1627A>G	A/G	34	32 (94.1)	2 (5.9)	>0.999	30 (88.2)	4 (11.8)	0.205
		A/G or G/G	14	13 (92.9)	1 (7.1)		10 (71.4)	4 (28.6)
	1896T>C	A/A	35	33 (94.3)	2 (5.7)	0.583	28 (80.0)	7 (20.0)	0.446
		A/G	13	12 (92.3)	1 (7.7)		12 (92.3)	1 (7.7)

[i] SNP, single nucleotide polymorphism.

For leukopenia, a lower incidence was observed among DPYD 85T>C variant carriers during the second cycle. This was not statistically significant. No consistent associations were found between DPYD 1627A>G or DPYD 1896T>C variants and any of the evaluated toxicities, including neutropenia and thrombocytopenia.

No significant associations were found between DPYD variants and thrombocytopenia. The incidence of grade 1-4 thrombocytopenia was similar across DPYD 85T>C, DPYD 1627A>G, and DPYD 1896T>C variants for both cycles. A slightly higher prevalence of thrombocytopenia was observed among A/G or G/G carriers of 1627A>G in the second cycle (28.6%) compared with A/A carriers (11.8%), but this difference was not statistically significant.

miRNA expression profiles associated with DPYD 85T>C

miRNAs were extracted from plasma samples of nine patients with CRC (five wild-type and four with the variant DPYD 85T>C) using the miRNeasy® Serum/Plasma kit. miRNA array was used to analyze 43 miRNAs related to the DPYD drug-metabolizing gene. miRNAs were arranged by unsupervised clustering, which organizes data according to expression similarity (Fig. 1A). This approach revealed natural groupings of WT and VT samples, reflecting their underlying expression profiles. The results demonstrated differential expression of miRNAs when comparing the variant DPYD 85T>C with wild-type patients with CRC (Fig. 1A). The scatter plot demonstrated up- and downregulated miRNAs in variant vs. wild-type patients, indicating a potential link between miRNA expression and the DPYD gene polymorphism (Fig. 1B). There were nine up- and 11 downregulated miRNAs in the variant DPYD 85T>C compared with the wild-type group (Table IV).

Figure 1 miR expression profiles in patients with VT (DPYD 85T>C) and WT colorectal cancer. A total of 43 miRs were detected in the plasma of patients with VT (n=4) and WT colorectal cancer (n=5) using miR array. (A) Heat map of differential expression of miRs in VT compared with WT colorectal cancer. Red and green show up- and downregulation, respectively. (B) Scatter plot showing miR profiles between VT and WT colorectal cancer. Each dot represents the fold-change in expression of miRNA. Red, green, and black dots represent up- and downregulated and unchanged miRNAs, respectively. miR, microRNA; VT, Variant; DPYD, dihydropyrimidine dehydrogenase gene; WT, Wild type; min, minimum; avg, average; max, maximum; sn, small nuclear.

Table IV Up- and downregulated miRs in Dihydropyrimidine dehydrogenase 85T>C compared with patients with wild-type colorectal cancer.

Table IV

Up- and downregulated miRs in Dihydropyrimidine dehydrogenase 85T>C compared with patients with wild-type colorectal cancer.

A, Upregulated miRs
No.	miR	Fold-change	P-value
1	miR-218-5p	3.44	0.001
2	miR-203a-5p	3.35	0.003
3	miR-215-3p	3.35	0.003
4	miR-587	3.35	0.003
5	miR-1915-3p	3.35	0.003
6	miR-129-5p	3.35	0.003
7	miR-3190-5p	3.35	0.003
8	miR-365a-3p	3.35	0.003
9	miR-302a-3p	3.33	0.001
B, Downregulated miRs
No.	miR	Fold-change	P-value
1	miR-425-5p	-72.63	0.015
2	miR-20a-5p	-63.79	0.039
3	miR-21-5p	-38.30	0.042
4	miR-27b-3p	-30.85	0.020
5	miR-27a-3p	-30.45	0.015
6	miR-22-3p	-22.59	0.031
7	miR-197-3p	-14.96	0.017
8	miR-361-5p	-10.53	0.019
9	miR-766-3p	-10.01	0.045
10	miR-139-5p	-8.92	0.045
11	miR-10b-5p	-7.11	0.044

[i] miR, microRNA.

miRNA expression profiles associated with DPYD 1896T>C

To assess the miRNA expression profiles in CRC patients with the DPYD 1896T>C variant compared with wild-type, miRNA was extracted from plasma samples of six patients with CRC (three wild-type and three with the DPYD 1896T>C variant). The results revealed differential miRNA expression between the variant DPYD 1896T>C and wild-type patients (Fig. 2A). The scatter plot demonstrated both up- and downregulated miRNAs in the variant compared with the wild-type group, suggesting a potential association between these miRNAs and the DPYD 1896T>C variant (Fig. 2B). There were five up-and nine downregulated miRNAs in the variant group compared with wild-type patients (Table V).

Figure 2 miRNA expression profiles in patients with VT (DPYD 1896T>C) and WT colorectal cancer. A total of 43 miRs were detected in the plasma of patients with (n=3) and WT colorectal cancer (n=3). (A) Heat map of differential expression of miRs in VT compared with WT colorectal cancer. Red and green show up- and downregulation, respectively. (B) Scatter plot showing miR profiles between VT and WT colorectal cancer. Each dot represents the fold-change in expression of miRNA. Red, green, and black dots represent up- and downregulated and unchanged miRNAs, respectively. miR, microRNA; VT, Variant; DPYD, dihydropyrimidine dehydrogenase gene; WT, Wild type; min, minimum; avg, average; max, maximum; snRNA, small nuclear RNA.

Table V Up- and downregulated miRs in dihydropyrimidine dehydrogenase 1896T>C compared with patients with wild-type colorectal cancer.

Table V

Up- and downregulated miRs in dihydropyrimidine dehydrogenase 1896T>C compared with patients with wild-type colorectal cancer.

A, Upregulated miRs
No.	miR	Fold-change	P-value
1	miR-21-5p	122.32	0.028
2	miR-22-3p	49.31	0.049
3	miR-143-3p	20.91	0.043
4	miR-10b-5p	6.42	<0.001
5	miR-193b-3p	2.31	0.016
B, Downregulated miRs
No.	miR	Fold-change	P-value
1	miR-587	-3.56	0.032
2	miR-1915-3p	-3.56	0.032
3	miR-129-5p	-3.56	0.032
4	miR-3190-5p	-3.56	0.032
5	miR-149-5p	-3.56	0.032
6	miR-215-3p	-3.56	0.032
7	miR-365a-3p	-3.56	0.032
8	miR-203a-5p	-3.56	0.032
9	miR-193a-3p	-2.37	0.010

[i] miR, microRNA.

miRNA profiling was performed on a randomly selected subset of 16 patients. Among these, only two individuals had the DPYD 1627A>G variant. Due to the very limited number of variant carriers, statistical analysis of the association between this SNP and miRNA expression as not conducted.

Discussion

DPYD encodes the enzyme DPD, which is responsible for the catabolism of pyrimidine-based compounds, including the chemotherapy drug 5-FU. Deficiency or reduced activity of DPYD leads to the accumulation of 5-FU, resulting in severe hematological toxicity, including neutropenia, thrombocytopenia, anemia and leukopenia. Patients with partial DPD deficiency have a 3.4-fold increased risk of developing grade IV neutropenia compared with those with normal DPD activity (32,33). Analysis of the DPYD gene in patients with grade IV neutropenia revealed that 50% of the individuals tested were either heterozygous or homozygous for the IVS14+1G>A mutation (34). In addition, the DPYD*5 gene mutation leads to decreased DPD enzyme activity and impaired 5-FU metabolism, which is associated with the accumulation of 5-FU and increased chemotherapeutic toxicity in gastric and colon carcinoma (35).

The present study assessed the association between DPYD 85T>C, 1627A>G and 1896T>C and hematological toxicity across two cycles of 5-FU-based chemotherapy in patients with CRC. Recent analyses (8,36) have demonstrated the relevance of DPYD variants in Asian cohorts. For example, Chan et al (8) conducted a systematic review of DPYD genotypes in non-European individuals with severe fluoropyrimidine toxicity, identifying high-frequency variants [c.1627A>G (DPYD*5) and c.85T>C (DPYD*9A)] in East and Southeast Asian patients that are not commonly included in European-focused genotyping panels. Similar to present study, the frequency of DPYD 85T>C variant was 14% in Thai patients with CRC. Among these, DPYD 85T>C showed a trend toward increased anemia, especially during the second chemotherapy cycle, with 69.2% of T/C carriers developing anemia compared to 40.0% of T/T carriers (P=0.070). Although not statistically significant, this trend suggests a possible cumulative effect of the variant on drug metabolism and hematologic toxicity. Previous studies (11,34,36) have reported similar associations between DPYD variants and hematological adverse effects. Patients with DPYD 85T>C variant have a significantly increased risk of hematological toxicity (11,35,37).

Detailleur et al (38) reported a high prevalence of the DPYD 85T>C variant among patients who experienced severe toxicity following treatment with 5-FU-based chemotherapy in a retrospective study. Different DPYD polymorphisms confer varying levels of residual DPD enzyme activity. For example, the 85T>C (DPYD*9A) variant is associated with a modest reduction in DPD activity rather than complete loss, which may explain the milder and variable toxicity profiles seen in certain carriers (6,9).

The meta-analysis by Leung and Chan (10) revealed that the DPYD 1627A>G variant has a high allele frequency (>20%) in patients from China, Korea, Japan and Thailand, while the DPYD 1896T>C variant shows an allele frequency >14% in Korean and Thai cohorts. The statistical power to detect associations for both polymorphisms is >75%. Similar to present study, the allele frequency was 17% in DPYD 1627A>G and 14% in DPYD 1896T>C. The DPYD c.1627A>G variant is a SNP that results in a missense mutation, causing an isoleucine-to-valine substitution at codon 543 (p.I543V) of the DPD enzyme. Several studies have reported that it may influence DPD enzymatic activity, particularly in compound heterozygous individuals or in the presence of additional risk alleles (6,10).

The present study found that DPYD 1627A>G and 1896T>C variants were not significantly associated with anemia, neutropenia, leukopenia or thrombocytopenia in either treatment cycle. He et al (9) reported that DPYD enzyme activity does not differ significantly among carriers of the 85T>C (DPYD 9A), 1627A>G (DPYD 5) or 1896T>C variants.

miRNAs are small, non-coding RNAs that regulate gene expression by binding to the 3' untranslated region (UTR) of target mRNAs, leading to mRNA degradation or inhibition of translation (39). In drug metabolism, miRNAs have been found to regulate a variety of enzymes involved in drug processing, including cytochrome P450 and UDP-glucuronosyltransferases (40). Sun et al (41) reported that miR-21, miR-215, miR-218, miR-326 and miR-328 are involved in the regulation of 5-FU metabolic pathways, and their differential expression is significantly associated with clinical outcomes and survival in patients with CRC receiving fluoropyrimidine-based adjuvant chemotherapy.

The present study compared the miRNA profiles of wild-type and DPYD 85T>C and 1896T>C variants in patients with CRC and revealed distinct miRNA profiles between wild-type and variant patients, with different patterns of up- and downregulation observed for each variant. Several miRNAs were upregulated in the DPYD 85T>C variant, but downregulated in the 1896T>C variant, including miR-587, miR-1915-3p, miR-129-5p, miR-3190-5p, miR-215-3p, miR-365a-3p and miR-203a-5p. By contrast, some miRNAs were upregulated in the 1896T>C variant but downregulated in the 85T>C variant, such as miR-21-5p, miR-22-3p and miR-10b-5p. This suggested that the regulation of miRNAs varies across different DPYD variants, and the expression of miRNAs may be influenced by genetic variants, potentially modulating their target genes and cellular functions in CRC, which may affect disease progression, treatment response and toxicity. miR-21-5p was highly expressed in patients with the DPYD 1896T>C variant, showing a 122-fold increase compared with wild-type patients, suggesting its potential as a diagnostic biomarker for DPYD 1896T>C. This finding aligns with a previous study, which reported elevated miR-21-5p levels in the serum of patients with CRC, with fluctuations observed after surgery and recurrence (42). Moreover, miR-21-5p levels are associated with TNM staging and lymph node metastasis, suggesting that miR-21-5p may serve as an oncogene in CRC progression and a valuable diagnostic biomarker (42). However, these observations warrant further validation in a larger cohort to confirm their clinical relevance.

Downregulation of miR-22-3p was observed in the 85T>C variant, which is consistent with previous studies (43,44) showing that miR-22-3p is downregulated in CRC and exerts antitumor effects. miR-22-3p decreases the proliferative, invasive and migratory capacity of CRC cells (43). Another study found that low miR-22 in CRC tissue and metastatic cell lines correlated with metastasis, advanced stage, and relapse, whereas ectopic miR-22 inhibited CRC growth and metastasis (44). In the present study, miR-10b-5p was also downregulated in DPYD 85T>C compared with wild-type patients. The aforementioned study showed that miR-10b-5p is a target of circular RNAs, such as circ_0021977. Additionally, p21 and p53 are potential target genes of miR-10b-5p (45). The circ_0021977/miR-10b-5p/p21/p53 axis suppresses CRC cell proliferation, migration and invasion, suggesting its role in CRC progression (45). miR-587 contributes to drug resistance by downregulating Protein Phosphatase 2, Regulatory Subunit A, Beta (PPP2R1B), a subunit of the Protein Phosphatase 2 (PP2A) complex, which results in increased AKT activation and enhanced 5-FU resistance through elevated X-linked Inhibitor of Apoptosis Protein (XIAP) expression. Targeting the miR-587/PPP2R1B/phosphorylated AKT/XIAP axis could offer therapeutic strategies to overcome drug resistance in CRC (46). In the present study, miR-587 was upregulated in patients with the 85T>C variant, which may be associated with 5-FU resistance. This preliminary finding suggests a possible role of miR-587 in CRC treatment and its relevance in drug resistance, warranting further investigation in larger cohorts. The present data also suggested that miR-1915-3p may be upregulated in the 85T>C variant. This aligns with previous report that exosomal delivery of miR-1915-3p can enhance the chemotherapeutic efficacy of oxaliplatin in CRC cells by suppressing epithelial-mesenchymal transition-promoting oncogenes, such as 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase 3) and USP2 (Ubiquitin-Specific Protease 2) (46). In addition, miR-129-5p was downregulated in the 1896T>C variant. This is consistent with previous research indicating that miR-129-5p is downregulated in colon cancer tissue and miR-129-5p mimics suppress the proliferation of colon cancer cells (48). However, further studies with larger sample sizes are needed to confirm these observations.

The present study demonstrated an upregulation of miR-3190-5p in the 85T>C variant while it was downregulated in the 1896T>C variant compared with wild-type. To the best of our knowledge, only one previous study has reported this miRNA in CRC (48). miR-3190-5p regulates the expression of ABCC4 (ATP-binding cassette sub-family C member 4) by binding the 3' UTR of the ABCC4 gene, and this regulatory effect is disrupted by the rs3742106 polymorphism (49). Additionally, miR-3190-5p increases the intracellular concentration of 5-FU, thereby enhancing the sensitivity of CRC cells to 5-FU (49), suggesting the miR-3190-5p and DPYD variant may serve as biomarkers for the personalized use of 5-FU in CRC treatment.

The present study demonstrated upregulation of miR-215-3p in the 85T>C variant and a downregulation in the 1896T>C variant compared with the wild-type. Previous studies (50-52) have indicated that the levels of miR-215-3p are associated with the sensitivity of CRC cells to 5-FU, with alterations in miR-215-3p affecting 5-FU sensitivity. Specifically, miR-215-3p has been shown to enhance the apoptosis of CRC cells treated with 5-FU (50). Mechanistically, miR-215-3p regulates C-X-C chemokine receptor type 1 (CXCR1) expression in human CRC HCT116) cells, and alterations in CXCR1 affect 5-FU sensitivity, influencing CRC cell response to the drug (50,51). The genetic variant-dependent regulation of miR-215-3p may provide insight into personalized treatment strategies, where miR-215-3p may serve as a biomarker for predicting response to 5-FU chemotherapy in patients with CRC. Moreover, p53 is a key regulator of the DNA damage response and apoptosis, both of which are critical in determining cell sensitivity to 5-FU (52). Studies have shown that p53 status influences miRNA expression profiles and may modulate response to fluoropyrimidine-based chemotherapy (53,54). Thus, variability in p53 function may interact with DPYD variants and impact downstream miRNA expression and treatment outcomes.

miR-365a-3p and miR-203a-5p were upregulated in the 85T>C variant and downregulated in the 1896T>C variant compared with the wild-type. miR-365a-3p inhibits CRC progression, at least in part by suppressing ADAM10 expression and the associated JAK/STAT signaling pathway (55). miR-203a-5p regulates tumorigenesis by downregulating suppressor of cytokine signaling 3 expression (56), highlighting the signaling axis as a potential therapeutic target in CRC.

Increasing evidence (24,25) suggests that miRNAs may regulate DPYD expression and, by extension, impact the metabolism of fluoropyrimidines. Offer et al (24) identified expression of miR-27a and miR-27b as potential pharmacological modulators of hepatic DPD enzyme function (24). DPD serves a crucial role in the uracil catabolic pathway by converting the chemotherapy drug 5-FU into its inactive metabolite 5-dihydrofluorouracil. A deficiency in DPD, resulting from insufficient expression or harmful variants in the DPYD gene, is associated with severe toxic reactions to 5-FU (24). Additionally, patients who are heterozygous for the miR-27a SNP (rs895819) have an increased risk of fluoropyrimidine toxicity. This has been studied in both individuals with wild-type and variant DPYD variant (25). These findings suggest that miR-27a rs895819 could serve as a potential biomarker for predicting fluoropyrimidine-associated toxicity. Hirota et al (57) demonstrated that DPYD is a target of several miRNAs, including miR-27a, miR-27b, miR-134 and miR-582-5p (57). Overexpression of these miRNAs leads to a significant reduction in reporter activity in a plasmid containing the 3' UTR of DPYD mRNA in a luciferase assay (57). Additionally, the overexpression of these miRNAs also results in a notable decrease in DPD protein levels in pancreatic carcinoma MIAPaca-2 cells, suggesting these miRNAs regulate DPD protein expression at the post-transcriptional level (57). In the present study, miR-27a-3p and miR-27b-3p were significantly downregulated in patients with CRC with the variant compared with wild-type patients, indicating the absence of DPYD deficiency in the DPYD 85T>C variant.

Hou et al (58) demonstrated similar miRNA expression in human colon cancer cells (HT29) in response to 5-FU treatment and nutrient starvation using miRNA microarray analysis (58). Bioinformatic predictions, pathway and gene network analyses revealed four downregulated miRNAs, including hsa-miR-302a-3p, and 27 upregulated miRNAs, which may regulate autophagy in CRC cells during 5-FU-based chemotherapy (58). The present study showed that miR-302a-3p was upregulated in DPYD 85T>C compared with wild-type CRC. The present study had limitations. The sample size for miRNA analysis was relatively small, particularly in the variant groups (n=4 for 85T>C and n=3 for 1896T>C), which may limit statistical power and generalizability. The small sample size may contribute to potential bias and increased variability and limit the ability to detect subtle but biologically meaningful differences. The present study focused on only three SNPs of DPYD associated with 5-FU-related toxicities; other SNPs of DPYD variants should be considered for further study. Additionally, the lack of longitudinal data on toxicity progression and survival outcomes limits the ability to establish causal links. Future studies with larger, multi-center cohorts and functional validation assays are warranted. These should include measurements of DPD enzyme activity and 5-FU pharmacokinetics to clarify the mechanistic association between the identified SNP variants, miRNA expression profiles and clinical responses to 5-FU in CRC.

In conclusion, as miRNAs are regulators of drug metabolism and toxicity, future research should focus on their potential role in modulating the impact of DPYD polymorphisms. Understanding the association between miRNA expression and DPYD activity may facilitate personalized treatment strategies, where miRNA profiling could be used alongside DPYD genotyping to predict toxicity and optimize drug dosing. Additionally, miRNA-based therapy may be explored as a potential method to modulate DPYD activity in patients with low enzyme activity, decreasing the risk of toxicity while maintaining therapeutic efficacy.

Supplementary Material

miR target sequences used in the miR array.

Acknowledgements

The authors would like to thank Professor Sam Ormond (Clinical Research Center, Faculty of Medicine, Thammasat University, Pathum Thani, Thailand) for English editorial assistance.

Funding

Funding: The present study was supported by the Health Systems Research Institute under Genomics Thailand Strategic Fund (grant no. 65-084) and Thailand Science Research and Innovation Fundamental Fund, fiscal year 2025.

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

PP designed the study, performed experiments, analyzed data and wrote and revised the manuscript. PC, ES, TR, SA and SS designed the study. PJ performed experiments. CS and CA conceived and designed the study and analyzed data. CA wrote and edited the manuscript. PP and CA confirm the authenticity of all the raw data. All authors have read and approved the final manuscript.

Ethics approval and consent to participate

The present study was carried out in compliance with the Declaration of Helsinki and received approval from the Ethics Committee of Ramathibodi Hospital, Mahidol University, Bangkok, Thailand (approval no. MURA2020/1613 Ref.2419). Written informed consent was provided by all participants before the start of the study.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References