Epigenetic differences in the tumor suppressor genes MLH1 and p16INK4a between Nepalese and Swedish patients with colorectal cancer

Introduction

Colorectal cancer (CRC) ranks as the third most prevalent cancer disease worldwide, yet its incidence rates exhibit a staggering tenfold variation globally. Notably, European societies report the highest rates, ranging between 40 and 90 cases per 100,000 individuals, while many African regions, like middle Africa, demonstrate significantly lower rates, at approximately 12 cases per 100,000 [1]. Similarly, South Asia presents a diverse landscape, with affluent societies such as South Korea, Japan, and Singapore reporting higher rates compared to lower-income countries in the region [2]. In India, the age-adjusted incidence rate, standardized against a global population standard, drops to as low as 4 cases per 100,000 individuals [2].

The primary risk factor attributed to the onset of CRC is advancing age [3]. European studies estimate a mean age of 72 years among patients with CRC, while in low-income Asian countries, this figure ranges from 43.7 to 51.2 years [4, 5]. Disparities in age between affluent and low-income nations may significantly contribute to the notable divergence in CRC incidence rates within these populations. Interestingly, while there has not been a general increase in CRC incidence over recent decades, there has been a distinct rise observed among individuals under 50 years of age [6]. Additionally, a decline in the age of CRC diagnosis is evident in Asian countries [7]. A retrospective analysis spanning a decade (1999–2008) at a tertiary hospital in Nepal uncovered a noteworthy trend: the mean age of patients with CRC decreased from 34±4.7 years in the initial 5 years to 31±5.1 years in the subsequent half, indicating a declining trend in the age at diagnosis [8].

Numerous studies have delved into the genetic and epigenetic changes underlying the onset of CRC. Of particular interest is DNA hypermethylation, a crucial process implicated in the suppression of tumor suppressor genes. This aberrant methylation of CpG-rich promoter regions of tumor suppressor genes leads to transcriptional inactivation, a hallmark of CRC development [9]. These alterations manifest early and have been discerned in macroscopically normal mucosa adjacent to both benign and malignant tumors. Primarily, they occur at cytosine bases within CpG dinucleotides outside CpG islands [10]. Additionally, DNA replication errors in microsatellite repeat sequences can result in microsatellite instability (MSI), observed in 10–15 % of sporadic CRC cases, signaling deficiencies in DNA mismatch repair (MMR) genes. Epigenetic modifications in the MutL homolog 1 gene (MLH1) frequently contribute to MSI, particularly through alterations in its promoter region, which is widely recognized as an early event in CRC tumorigenesis [11], [12], [13].

Frequent hypermethylation has also been documented in the INK4b-ARF-INK4a locus, which encodes two critical cyclin-dependent kinase inhibitors, p15INK4a and p16INK4a, both pivotal in regulating the G1 phase of the cell cycle. Within this locus, an additional protein, p14ARF (alternative reading frame), is also encoded [14]. A prior investigation revealed that hypermethylation of the p16INK4a gene promoter was detected in 36 % of mucosa samples obtained 10 cm from the tumor in patients with CRC [15]. Notably, hypermethylation of p16INK4a in mucosa correlated with poorer patient outcomes, suggesting that epigenetic changes in the surrounding mucosa might precede genetic alterations in the early stages of carcinogenesis [13].

While epigenetics has been extensively studied in Western populations, the association between epigenetics and CRC remains largely unexplored in resource-constrained countries such as Nepal. This study aimed to bridge this gap by comparing DNA methylation levels in two commonly hypermethylated tumor suppressor genes, MLH1 and p16INK4a, within the tumor and matched mucosa of patients with CRC from both Nepal and Sweden. Additionally, the study aimed to analyze the correlation between the methylation data and the demographic and clinicopathological profiles of the patients.

Patients and methods

Results

In the study, 56 % of participants were male, and 44 % were female. The median age was 53 years, ranging from 19 to 78 years. Stage III CRC was predominant in the Nepalese group (54 %) compared to the Swedish group (38 %). The majority of the Nepalese patients exhibited well-differentiated tumors (41 %), whereas moderately differentiated tumors were more common among Swedish patients (53 %). Colon cancer represented 74 % of cases among Nepalese patients and 62 % among Swedish patients, with the predominant site of lesions being the right colon (Table 1).

MLH1 methylation at individual CpG sites

MLH1 methylation levels could be determined in all samples from both the Nepalese and Swedish patient cohorts. Distinct variations in MLH1 methylation levels between the mucosa and tumors of cohorts were evident at individual CpG sites (Figure 1, Supplementary Table 2). In mucosa samples from Nepalese patients, MLH1 methylation levels ranged from 0 to 50 %, contrasting with 0 to 8% in Swedish patients. MLH1 methylation levels in tumor samples ranged from 0 to 46 % and 0 to 74 % in the two patient cohorts, respectively.

Figure 1:

MLH1 methylation at each CpG site in mucosa and tumor tissue of Nepalese patients (A and B) and Swedish patients (C and D) with colorectal cancer.

p16INK4a methylation at individual CpG sites

p16INK4a methylation was assessable in all samples from Swedish patients and in 32 out of 39 mucosa samples and 31 out of 39 tumor samples from Nepalese patients. In contrast to the generally uniform methylation pattern observed for MLH1 in mucosa, there was notable variability in methylation levels among different CpG sites for p16INK4a, particularly with elevated levels at CpG sites 4 and 6 in mucosa samples from Nepalese patients (Figure 2). Methylation levels of p16INK4a ranged from 0 to 46 % in mucosa samples from Nepalese patients, contrasting with 0 to 21 % in Swedish patients. Although p16INK4a hypermethylation was more prevalent in tumor samples from Swedish patients, the highest methylation level was detected in a tumor from one of the Nepalese patients (case 19, Figure 3B). Methylation levels of p16INK4a in tumors ranged from 1 to 100 % and from 1 to 75 % in the Nepalese and Swedish patient cohorts, respectively.

Figure 2:

p16INK4a methylation at each CpG site in mucosa and tumor tissue of Nepalese patients (A and B) and Swedish patients (C and D) with colorectal cancer. CpG site 5 was not included due to a single nucleotide polymorphism at this position.

Figure 3:

Scatter dot plots showing MLH1 and p16INK4a methylation levels in mucosa and tumor tissues of (A) Nepalese and (B) Swedish patients with colorectal cancer. Data are presented as mean values with SEM bars.

Methylation according to age

Patients were stratified into three age groups (<50, 50–60, and >60 years old) and assessed for MLH1 and p16INK4a methylation in both mucosa and tumor tissue. The results are presented in Figure 4. There was no significant difference in methylation levels between age groups within the Nepalese patient cohort. However, Swedish patients aged over 60 years exhibited a higher p16INK4a methylation level compared to those younger than 50 years (p=0.0046). Furthermore, Nepalese patients had a higher MLH1 methylation level in mucosa compared to Swedish patients in the age groups of 50–60 (p=0.011) and >60 (p=0.016). The two Swedish patients displaying either MLH1 or p16INK4a methylation in mucosa belonged to the youngest age group. Additional details regarding the percentages of Nepalese and Swedish patients with MLH1 hypermethylation at respective CpG sites in mucosa and tumor tissue, segmented by age group, can be found in Supplementary Table 3.

Figure 4:

Scatter dot plots showing MLH1 and p16INK4a methylation levels in mucosa and tumor tissues of (A) Nepalese and (B) Swedish patients with colorectal cancer by age group. Data are presented as mean values with SEM bars.

Among Swedish patients, the percentage of individuals lacking both MLH1 and p16INK4a methylation in tumor samples decreased with age, dropping from 82 % in the <50 years age group to 33 % in the >60 years age group (Figure 5A). Conversely, in the Nepalese patient cohort, as much as 75 % of patients in the >60 years age group lacked methylation of both genes.

Figure 5:

Bars and boxes comparing the MLH1 and p16INK4a methylation status in tumors deriving from Nepalese and Swedish patients with colorectal cancer by (A) age group and (B) sex. The methylation status is classified as follows: no methylation, exclusive MLH1 methylation, exclusive p16INK4a methylation, or methylation in both genes.

Methylation according to tumor location

No statistically significant difference was observed in the methylation levels of MLH1 or p16INK4a in mucosa or tumor tissues among Nepalese patients based on tumor location (Figure 6A). In Swedish patients, however, significantly higher methylation levels of both genes were detected in tumor samples obtained from the right side of the colon (Figure 6B).

Figure 6:

Scatter dot plots showing MLH1 and p16INK4a methylation levels in mucosa and tumor tissues of (A) Nepalese and (B) Swedish patients with colorectal cancer by tumor site. Data are presented as mean values with SEM bars.

Discussion

In this study, we compared the methylation status of the tumor suppressor genes MLH1 and p16INK4a in colorectal tumor tissue and matching mucosa samples from Nepalese and Swedish patients with CRC. Our findings revealed a significantly higher frequency of DNA methylation in the mucosa of Nepalese patients compared to Swedish patients. Additionally, the methylation levels observed in some mucosa samples from Nepalese patients were notably elevated compared to previous reports [17]. The underlying reason for this unexpectedly high methylation level in the mucosa of Nepalese patients remains unclear, but it may be linked to environmental and lifestyle factors known to influence gene regulation through epigenetic mechanisms, including DNA methylation. Such factors encompass age, sex, diet, lifestyle choices, environmental exposures, and medications [18, 19] as discussed below.

Age-related promoter hypermethylation has been documented for several tumor suppressor genes, including the mismatch repair gene MLH1, which contributes to 15–30 % of stomach and colon carcinomas [20]. Additionally, it is known that hypermethylation is more prevalent among females than males. Consequently, age and sex differences between patient groups may yield disparate results. However, in this study, the two patient cohorts were matched for these covariates; thus, any observed differences in methylation status between groups would likely be attributed to other factors [18]. Nevertheless, MLH1 hypermethylation at all five analyzed CpG sites in DNA derived from mucosa was more common in older Nepalese patients. Moreover, the percentage of patients with methylation of both MLH1 and p16INK4a in tumors was observed to increase with age in both Nepalese and Swedish patients. Consistent with previous reports, both MLH1 and p16INK4a methylation in tumor tissue was more prevalent in females compared to males among Swedish patients [21, 22]. MLH1 methylation in the mucosa of Nepalese patients was also more prevalent in females. However, the methylation status in tumors of Nepalese patients did not follow this pattern, as cases with MLH1 and p16INK4a hypermethylation in tumors were equally distributed among female and male patients.

In addition to age and sex, the occurrence of hypermethylation may also be influenced by tumor location. Typically, hypermethylation is more prevalent in tumors situated on the right side compared to those on the left side of the colon [23]. Consistent with previous findings, the results of this study revealed a higher level of methylation in right-sided colon tumors among both Nepalese and Swedish patients. However, methylation levels in mucosa samples did not exhibit variations based on tumor location.

There was no observed correlation between the methylation levels in mucosa and tumor samples within any of the patient cohorts. This finding aligns with similar results reported by Coppede et al., who analyzed five CRC-related genes in patients with CRC [19]. Among these genes, APC exhibited the highest frequency of methylation in both mucosa and tumor tissue, yet the methylation level in mucosa remained below 10 % and did not correlate with that in tumors. Conversely, Ramirez et al., in a methylation analysis of six genes including MLH1 and p16INK4a, found that up to 71 % of patients exhibited similar methylation patterns between mucosa and tumor tissues [13]. Additionally, 47 % of the mucosa samples displayed methylation in one or two loci, while 12 % showed methylation in three or more loci. The concept of “field cancerization” has been proposed in normal mucosa adjacent to tumor cells [24]. The presence of hypermethylation in mucosa cells of patients with colon cancer indicates that these cells harbor some of the epigenetic alterations associated with the tumor [25]. Such alterations in the mucosa surrounding CRC lesions may serve as predictors of outcome. As demonstrated in a previous study, hypermethylation of p16INK4a in CRC mucosa was associated with a worse disease-free survival [26].

The incidence of CRC is progressively rising, particularly in low- and middle-income countries, likely due to shifting demographics and dietary patterns [27]. Notably, Asians relocating to the United States eventually develop a similar CRC incidence as native-born Americans [28], suggesting a potential influence of dietary habits. It has been posited that the prevalent vegetarian diet in Asian and African societies may confer protection against CRC [29]. Obesity, strongly linked to CRC, is also recognized as a risk factor for DNA hypermethylation [30, 31]. In Sweden, the prevalence of obesity has been steadily increasing over recent decades, whereas Nepalese patients may experience undernourishment comparatively [32]. Obesity could thus contribute to a higher incidence of DNA hypermethylation and CRC development in Swedish patients compared to Nepalese patients. Moreover, dietary differences may play another significant role: the predominant Nepali vegetarian diet may shape a gut microbiota distinct from that resulting from a typical Western-style diet prevalent in Sweden. Additionally, environmental factors vary between the two populations. For instance, Kathmandu, the capital of Nepal, faces significant pollution, and its tap water is often contaminated with bacteria and parasites like Giardia Lamblia. These diverse factors, known to influence DNA methylation and subsequent gene regulation and expression, underscore the disparities between the Nepalese and Swedish populations [18].

Currently, we can only speculate on the reasons for the high methylation levels found in the mucosa of Nepalese patients. One hypothesis would be that aberrant methylation may be triggered in large areas of the gut in response to alterations in the microbiota. The interesting connection between the microbiota and the onset of CRC is an emerging area of research that is garnering increasing attention. For example, Fusobacterium nucleatum has been linked to promoter methylation of MLH1 and p16INK4a, while Streptococcus spp. seem to be associated with MLH1 hypermethylation [33]. In another recent study involving mice, even a nonpathogenic microbiota was found to significantly influence DNA methylation in the gut, contributing to the development of aberrant methylomes observed in aging and cancer [34].

With the increasing incidence of CRC among individuals under 50 years old, there is a recent recommendation to lower the age for CRC screening to 45 years [35]. However, routine colonoscopy screening may not be feasible in all countries, particularly those with low CRC incidence. Nevertheless, analyzing methylation in mucosa samples obtained during colonoscopy could serve as a screening tool for high-risk patients and potentially as a follow-up tool for predicting recurrent disease. Moreover, ongoing research is focused on developing and investigating novel techniques for detecting and quantifying methylated DNA using less invasive methods.

The carcinoembryonic antigen (CEA) marker is utilized to monitor treatment response in patients with CRC, while the Fecal Occult Blood Test (FOBT) has been employed for screening high-risk populations; however, its sensitivity and specificity, particularly for proximal CRC, are limited [36]. Given this context, detecting epigenetic alterations in body fluids could offer a safer and more effective screening method for CRC in high-risk populations [27]. Recently, the U.S. Food and Drug Administration (FDA) approved methylated Septin 9 (SEPT9) as a diagnostic marker for CRC screening. SEPT9 exhibits higher sensitivity than the fecal immunochemical test (FIT) and CEA in CRC detection [37]. Nevertheless, SEPT9 analysis is blood-based, and utilizing methylation tests based on fecal samples could provide an effective noninvasive alternative for diagnosing, screening, and monitoring CRC. However, CpG island methylation of certain tumor suppressor genes, including MLH1 and p16INK4a, can also be detected in both mucosal and fecal DNA of healthy individuals. Therefore, it is crucial to identify informative CRC-associated markers whose levels significantly differ from the normal baseline [17].

Conclusions

Nepalese and Swedish patients displayed equal levels of MLH1 and p16INK4a methylation in tumors, but Nepalese patients had a higher level of MLH1 and p16INK4a methylation in mucosa compared to Swedish patients. These discrepancies may be attributed to environmental and lifestyle factors known to influence gene regulation through epigenetic mechanisms, including DNA methylation. Whether the elevated hypermethylation level in the mucosa of Nepalese patients with CRC represents an early aberration during tumorigenesis that could be valuable for screening high-risk patients or predicting recurrence is currently unknown. Ongoing studies will address this question by assessing the prevalence of tumor suppressor gene hypermethylation in the colorectal mucosa of Nepalese control subjects.