PDF
Citation
Print
Introduction
Colorectal cancer (CRC) is the third most commonly diagnosed cancer and the third leading cause of cancer-related mortality in the United States [1]. There is a well-established sexual dimorphism in CRC worldwide [2], with males demonstrating higher incidence rates than females [3]. CRC also exhibits different molecular and pathological characteristics depending on the tumor location, that right colon cancer, showing high microsatellite instability (MSI) status, and CpG island methylator phenotype (CIMP) [3], accounts for approximately 15% of all CRC cases and is frequently associated with females [4]. However, the rationale behind the increased frequency of CRC in the right side among females remains unknown.
Nuclear factor erythroid 2-related factor 2 (NRF2), a transcription factor, maintains redox homeostasis, suppressing inflammation and protecting cells from inflammation-induced carcinogenesis [5]. However, NRF2 activation also shields cancer cells from oxidative stress, promoting cancer progression and suggesting a dual role [6]. Recently, the programmed cell death receptor-1 ligand (PD-L1) locus was reported to contain an NRF2 binding site in human melanoma tissues, which means PD-L1 is regulated by NRF2 [7]. There was a report of the role of the NRF2–PD-L1 pathway in resistance to oxaliplatin in CRC patients [8]. PD-L1 overexpression in tumor cells leads to immune evasion, ultimately contributing to the formation of an immunosuppressive tumor microenvironment [9]. Furthermore PD-L1 also contributes to immune-independent tumorigenicity and interacts with key molecules in tumor progression, such as epidermal growth factor receptor (EGFR) [10]. Thus immunotherapy, specifically immune checkpoint inhibitors (ICIs) targeting PD-L1 or programmed cell death receptor-1 (PD-1), has emerged as a powerful tool in cancer treatment [9,10]. Several studies suggested sex differences in the effects of ICI treatments [11,12]. However, the role of immune environment is not well known in colorectal adenomas (AD) and colon carcinogenesis [13,14]. Furthermore, the expression of proinflammatory mediators and immune checkpoints and their potential correlation with sex and tumor location is lacking.
Previously we reported a lower PD-L1 expression in NRF2 knockout mice treated with 17β-estradiol (E2) compared to the wild-type CRC group induced by azoxymethane/dextran sulfate sodium treatment [15]. In addition, estrogen suppressed colon tumor development through the downregulation of PD-L1 expression in the MC38 colon tumor mice, thereby enhancing the therapeutic effect of anti–PD-L1 [16]. Additionally, PD-L1 expression, mismatch repair/MSI status, and EGFR expression were different in the right-sided colon cancer in females [17]. Thus, we hypothesized that the expression levels of NRF2 and immune checkpoints (PD-L1 and PD-1) might change sequentially from healthy control (HC) to AD and ultimately to CRC, especially in right colon. Hence, this study aimed to elucidate the molecular characteristics of CRC and AD depending on sex.
Materials and Methods
1. Study participants
Study participants who visited Seoul National University Bundang Hospital between April 2013 and July 2023 for regular checkups for CRC surveillance or gastrointestinal symptoms such as abdominal discomfort, diarrhea, and constipation were prospectively enrolled. Patients were recruited according to the following inclusion criteria: (1) histologically confirmed colorectal adenocarcinoma or (2) ADs ≥ 10 mm in diameter according to the endoscopic presentation. Controls were defined as participants without any evidence of CRC or adenoma, who agreed to participate in the study. The exclusion criteria were as follows: (1) history of CRC, polyps, or colectomy before the first surveillance colonoscopy; (2) hereditary CRC syndromes, such as familial adenomatous polyposis or Lynch syndrome; (3) a family history of CRC in at least one first-degree relative; (4) inflammatory bowel disease; and (5) incomplete colonoscopy or incomplete clinical information. Finally, 378 participants, including 88 HC, 139 patients with AD, and 151 patients with CRC were selected and further categorized into two groups according to sex. Data were collected from questionnaires on sex, age, body mass index, and social history such as alcohol consumption and smoking.
2. Endoscopic and histologic analysis
All patients were examined by an experienced gastroenterologist (N.K.; over 30 years old) using video colonoscopes (Olympus CF-240I, Olympus). The locations of the biopsy lesions were classified as follows: proximal colon (cecum, ascending colon, hepatic flexure, and transverse colon) and distal colon (splenic flexure of the colon, descending colon, sigmoid, and rectum). The histological diagnosis of the specimens was evaluated by a gastroenterology pathologist in accordance with the 2019 World Health Organization guidelines for digestive system tumors in a blinded manner [18].
3. Sample preparation
The tissue samples were collected during colonoscopy. Molecular experiments were conducted using tumor samples from patients with AD and CRC and normal mucosa located 20 cm above the anal verge in the HC group.
4. Quantitative real-time polymerase chain reaction
Colon tumor samples from patients with AD and CRC and normal colonic mucosa located 20 cm above the anal verge (from sigmoid colon) in the HC group were used for quantitative real-time polymerase chain reaction (qRT-PCR). After colonoscopy, colonic samples were immediately frozen in liquid nitrogen and stored at −80°C until RNA isolation. Total RNA was extracted from colon tissue samples using a TRIzol reagent (#15596026, Invitrogen). For qRT-PCR, 2 μg of total RNA was reverse-transcribed using a High-Capacity cDNA Reverse Transcription Kit according to the manufacturer’s instructions (#4368814, Applied Biosystems). qRT-PCR was performed utilizing specific primers (S1 Table) and the SYBR Green PCR Master Mix (#4367659, Applied Biosystems) on a QuantStudio 7 Flex Real-Time PCR instrument (#4484643, Applied Biosystems). Expression levels were normalized to those of glyceraldehyde 3-phosphate dehydrogenase.
5. DNA preparation and sodium bisulfite modification of DNA and methylation-specific PCR
Colon tumor samples from patients with AD and CRC and normal colonic mucosa located 20 cm above the anal verge (from sigmoid colon) in the HC group were used for qRT-PCR. After colonoscopy, colonic samples were immediately frozen in liquid nitrogen and stored at −80°C until DNA isolation. Genomic DNA was manually extracted from the colon tissue samples. Briefly, specimens were homogenized in proteinase K solution (20 mmol/L Tris-HCl [pH 8.0], 10 mmol/L ethylenediaminetetraacetic acid, 0.5% sodium dodecyl sulfate, and 10 mg/mL proteinase K) using a sterile micropestle and then incubated for 3 hours at 52°C. DNA was isolated from the homogenates through phenol/chloroform extraction and ethanol precipitation.
Genomic DNA (1 μg) was bisulfite-modified using the EZ DNA Methylation Kit (Zymo Research), following the manufacturer’s instructions. For methylation-specific PCR (MSP), aliquots of 2 μL was used with a primer set specific to methylated or unmethylated sequences, with a specific annealing temperature of 56-57°C. Primer sets for the NRF2 and KEAP1 promoter regions and for the unmethylated promoter region of ACTB as a reference gene have been previously described [19]. The methylated and unmethylated primers for NRF2 and KEAP1 are listed in S1 Table.
MSP was performed using Solg Taq DNA Polymerase on a Verti 96 well Thermal Cycler (Applied Biosystems). The PCR results were analyzed visually by running it on an agarose gel (3.5%) with Dyne LoadingSTAR. Data were quantified using the ImageJ software (National Institutes of Health). Finally, the ratio between methylated and unmethylated PCR band intensities was calculated using Microsoft Excel (Microsoft Corporation) [20]. First, the relative levels of PCR products for methylated or unmethylated molecules were normalized to the ACTB as follows: Methylation level=(Band density of methylated molecules)/(Band density of ACTB). And then presented as the ratio of methylated or unmethylated PCR products to the total PCR products as follows: Methylation ratio=(Methylation level)/[Total levels of DNA molecules (Methylated+Unmethylated levels)]×100.
6. Immunohistochemistry
Colon tumor samples from patients with AD and CRC and normal colonic mucosa located 20 cm above the anal verge (from sigmoid colon) in the HC group were used for immunohistochemistry (IHC). After colonoscopy, colonic tissues were fixed with 4% paraformaldehyde solution and embedded in paraffin. Paraffin-embedded sections (4 μm) were treated with 3% hydrogen peroxide, and nonspecific binding sites were blocked. The sections were incubated with anti-NRF2 (#ab31163, rabbit polyclonal, Abcam) or anti–PD-L1 (#22C3 pharmDx, mouse monoclonal, Dako) antibodies. An automatic immunostainer (BenchMark XT, Ventana Medical Systems) and UltraView Universal DAB detection kit (Ventana Medical Systems) were used for immunostaining. Scoring for NRF2 protein expression was measured based on staining intensity: 0 (no staining), 1+ (weak), 2+ (moderate), and 3+ (strong). S2 Fig. shows representative IHC images based on the NRF2 protein expression score. IHC findings for NRF2, ranging from 0 to 300, were calculated as the product of intensity score and stained area (percentage of the stained portion of the entire slide) [15].
For PD-L1, expression in the tumor cell membrane and in the membrane and/or cytoplasm of tumor-associated immune cells, macrophages, and lymphocytes was scored [17]. The combined positive score (CPS) was calculated as the total number of PD-L1–positive viable cells (tumor and mononuclear inflammatory cells) divided by the number of all viable tumor cells, multiplied by 100 [17]. The CPS was defined as the total number of PD-L1–positive mucosal epithelial and mononuclear immune cells divided by the number of mucosal epithelial cells.
7. Statistical analysis
Data are presented as mean±standard error of the mean using GraphPad Prism software ver. 8.01 (GraphPad Software Inc.). PASW Statistics for Windows ver. 18.0 (SPSS Inc.) was used for the data analysis. Statistical significance was determined using the Mann-Whitney U test. Statistical significance was set at a p < 0.05.
Results
1. Baseline characteristics of the study participants
A total of 378 participants, including 88 HCs, 139 patients with AD, and 151 patients with CRC, were recruited. The baseline clinical characteristics showed statistical significance for age (p < 0.001) and current or ex-smoker status (p=0.022) (Table 1). Among the groups, the frequency of individuals aged 65 years or older was the highest in the CRC group (60.9%).
Each group was further categorized according to sex, and a univariate analysis was conducted to identify the associated risk factors (Table 1). CRCs were more frequently found in the right colon of females, whereas males had a higher incidence of CRCs in the left colon. Furthermore, males aged < 65 years had a higher prevalence of ADs (62.2%) compared to older males (37.8%), whereas older females had a higher prevalence of ADs than younger females (58.5% and 41.5%, respectively).
2. NRF2 expression and its methylation in colorectal carcinogenesis
The analysis proceeded as illustrated in Fig. 1, focusing on CRC progression in total and for each sex. Second, we compared sex within each group. Finally, molecular marker changes were analyzed according to the location of CRC depending on sex to figure out sex difference of CRC. For this aim we focused on NRF2–PD-L1 pathway [8], which might provide underlying mechanism of sex difference of CRC.
To identify the role of NRF2 in the progression of CRC, first, we measured the NRF2 mRNA expression. It was significantly higher in the CRC group than in the HC and AD groups (p < 0.001 for CRC vs. HC and CRC vs. AD, respectively) (Fig. 2A). Similar results were noted in males and females (Fig. 2A). NRF2 is regulated by KEAP1, and the expression of these genes could be regulated by DNA methylation. To investigate this possibility, next, NRF2 and KEAP1 methylation levels were measured. The NRF2 methylation levels became significantly lower during colon carcinogenesis (Fig. 2B). In terms of sex, NRF2 methylation was significantly lower in the male AD group than in the HC group and significantly lower in the female CRC group than in the AD group (Fig. 2B). The KEAP1 methylation level was also significantly lower in the AD and CRC groups than in the HC group, with no differences between the AD and CRC groups in both males or females (Fig. 2C). The NRF2 expression in the colon tissue was confirmed using IHC. Representative IHC images of NRF2 are illustrated in Fig. 2D. NRF2 expression was significantly higher in the AD and CRC groups than that in the HC group (HC vs. AD, p < 0.001; HC vs. CRC, p < 0.001; respectively) (Fig. 2E). These differences were similar between males and females (Fig. 2E). However, NRF2 expression was significantly higher in the AD group than that in the CRC group (Fig. 2E). These patterns were observed in males (Fig. 2E).
3. Differential PD-L1 and PD-1 expression in colorectal carcinogenesis
PD-L1, an immune checkpoint inhibitor target, has been reported as a downstream target of NRF2 [7]. Our research showed that both NRF2 and PD-L1 were involved in CRC development [15]. Therefore, we investigated the expression of PD-1 along with PD-L1. PD-L1 mRNA expression exhibited no significant differences in total, males, and females (Fig. 3A). However, PD-L1 expression by IHC in the colon tissue was remarkably different. Representative images are shown in Fig. 3B. IHC for PD-L1 was performed in 174 patients (20 in the HC, 61 in the AD, and 93 in the CRC group). PD-L1 expression was significantly higher in the CRC group than in the HC and AD groups (HC vs. CRC, p=0.002; AD vs. CRC, p=0.007) (Fig. 3C) without difference between the HC and AD groups regardless of sex (Fig. 3C). In case of PD-1, the mRNA expression was significantly higher in the HC group compared to AD in total, but not in males and females (Fig. 3D).
4. PD-L1 and NRF2 expression levels in relation to MSI status
MSI-high (MSI-H) is known as a predictive biomarker for a positive response to immunotherapy. Since we observed high expression of PD-L1 and NRF2 in CRC patients, we tried to further investigate their association with MSI status. PD-L1 expression (average CPS and percentage above CPS 1 and 5) and NRF2 (expression score of IHC) was higher in MSI-H patients than in microsatellite stability/MSI-low patients, although statistical significance was not reached mainly because of small size of CRC (n=29) (S3 Table).
5. Altered mRNA expression of proinflammatory mediators and antioxidant enzyme gene in colorectal carcinogenesis
Next, we explored alterations in mRNA expression levels of proinflammatory mediators, including cyclooxygenase 2 (COX-2) and IL-1β. COX-2 mRNA expression was remarkably increased in the CRC group compared to that in both the HC and AD groups regardless of sex (Fig. 4A). Similarly, IL-1β mRNA expression demonstrated a significant increase in the CRC group compared to both the HC and AD groups, without significant differences between the HC and AD groups regardless of sex (Fig. 4B).
6. Sex-based analyses of molecular markers in colorectal carcinogenesis
Next, sex-based analyses of molecular markers within each group were conducted. NRF2 methylation levels were significantly lower in females than in males in the CRC group, without sex differences in the HC and AD groups (Fig. 5A). NRF2 IHC results showed a higher expression in females than in males in the CRC group, without sex differences in the HC and AD groups (Fig. 5B). The mRNA expression of PD-L1 was significantly higher in females than in males in the AD group (Fig. 5C). Similarly, the mRNA expression of PD-1 was significantly higher in female than in males with CRC (Fig. 5D). However, no sex differences were observed in the NRF2 mRNA expression, KEAP1 methylation, PD-L1 IHC expression, and the mRNA expression of COX-2 and IL-1β (S4 Fig.).
7. Sex-dependent expression of NRF2 and PD-L1 in CRC by tumor location
As CRC exhibits different molecular characteristics depending on the tumor location the expression of NRF2 and PD-L1 were further analyzed according to sex and location of CRC. NRF2 expression via IHC was significantly higher in female patients with right-sided CRC than in male patients with left-sided CRC and female patients with left-sided CRC (Fig. 5E). Conversely NRF2 methylation levels were significantly lower in female patients with left-sided CRC than in both male patients with left- and right-sided CRC (S5B Fig.). PD-L1 mRNA expression was also significantly higher in female right-sided CRC than in male left-sided CRC and female left-sided CRC (Fig. 5F). Except for NRF2 methylation levels, no significant differences were noted in NRF2 mRNA expression, KEAP1 methylation, PD-L1 expression, and mRNA expression levels of PD-1, COX-2, and IL-1β according to the location of CRC occurrence (S5 Fig.). These results support the association between increased NRF2 and PD-L1 expression and the occurrence of right-sided CRC in females.
Discussion
Our findings revealed distinct patterns of sex and location pathways in CRC in terms of NRF2 and PD-L1 pathway. Female patients with right colon cancer showed a higher NRF2 and PD-L1 expression than other patients, while females with left colon cancer showed lower NRF2 methylation levels (Fig. 6). These results indicate that different NRF2 and PD-L1 expressions might determine the differences in CRC in terms of sex and tumor location. This supports that PD-L1 locus contains an NRF2 binding site and a positive correlation between NRF2 and PD-L1 expression in human melanoma tissues [7].
Recently, the pathogenesis of left and right CRC was known to be different [21]. That is, chromosomal instability and damage to the tumor suppressor genes such as adenomatous polyposis coli, Kirstenras, and p53 are prevalent in left-sided CRC, accounting for approximately 70% of CRC [21,22]. Conversely, right-sided CRC frequently exhibits MSI-H, CIMP-high, and BRAF mutations [21,23]. Among the pathogenic mechanisms of CRC, KEAP1/NRF2 axis has recently been considered important, and that inflammation and oxidative stress play crucial roles in the occurrence and progression of CRC [24]. Actually, the dual role of NRF2 in CRC became to be well known. That is, NRF2 is known to translocate to the nucleus, forms dimers with small Maf proteins, and binds to ARE to activate downstream gene expression, exerting antioxidant effects under oxidative or electrophilic stress in general [24]. However, in the cancer state the overexpression of NRF2 promotes the progression and metastasis of tumor which was proved in CRC [25]. The presence of NRF2 inhibited the anti-tumor immune responses of estrogen, facilitating the development of CRC in mice model. In humans, the activation of TLR4/KEAP1/NRF2 axis or NRF2/HO-1 pathway was shown to be involved in the increased invasiveness, metastasis, and treatment resistance of CRC [25]. This situation has been explained that cancer cells might hijack NRF2 signaling to survive through reactive oxygen species reduction and reprogramming metabolism in the tumor environment [26,27], causing chemotherapy resistance by NRF2 [28].
In our study, NRF2 expression increased with NRF2 and KEAP1 hypomethylation in the CRC group, which led to an increase in PD-L1 expression and a decrease in HO-1 expression in the CRC group. As there was a correlation between high PD-L1 expression and mismatch repair-deficient (dMMR) in right-sided CRC [29], our findings suggest that NRF2, PD-L1, and dMMR could play a key role in the pathogenesis of right-sided CRC in females. In the previous studies, supplementation of estrogen decreased PD-L1 expression in tumor cells as well as the expression of proinflammatory enzymes, contributing to an immunopatent tumor microenvironment in the NRF2-knockout state. However, the overexpression of NRF2, including hypomethylation of NRF2, is deemed to inhibit this protective effect of estrogen [15]. Thus, we supposed that the increased expression of NRF2 in right-sided CRC might contribute to the increased expression of PD-L1 resulting in the immunosuppressive tumor microenvironment and development of CRC. However, as we did not provide the causal relationship, further studies are necessary to support this suggestion.
Nevertheless, our study is the first to analyze the correlation between the NRF2/KEAP1 pathway and PD-L1 expression in HC, AD, and CRC patients according to sex and tumor location. NRF2 can promote proliferation and migration of CRC as previously mentioned, and expression of NRF2 even can lead to drug resistance to chemotherapeutic drugs [28,30]. Additionally, silencing NRF2–PD-L1 signaling pathways has been suggested as an effective method improving oxaliplatin efficacy in colon cancer cells [8]. In the present study, we reconfirmed these results and further analyzed the differences between sexes and tumor location in human tissue samples. Collectively, we suggest that there are sex-related differences in the role of NRF2–PD-L1 pathway in the progression of CRC, and the involvement of estrogen would be one mechanism to explain sex difference in CRC. In conclusion, the expression and methylation patterns of NRF2 and PD-L1 in CRC may explain the CRC-specific sex and tumor location. These insights may contribute to a deeper understanding of the pathogenesis of CRC and highlight the need for personalized treatment strategies.
XML Download