The Roles of Ninjurin1 and Estrogen in Modulating Azoxymethane/Dextran Sodium Sulfate–Induced Colitis-Associated Colorectal Cancer in Male Mice
Article information
Abstract
Purpose
Ninjury-induced protein 1 (Ninj1) is associated with inflammation and tumor progression and shows increased expression in various cancers. This study aimed to investigate the role of Ninj1 in colitis-associated colorectal cancer (CRC) by focusing on its interaction with 17β-estradiol (E2).
Materials and Methods
Using an azoxymethane (AOM)/dextran sodium sulfate (DSS) mouse model of colitis-associated CRC, wild-type (WT) and Ninj1 knockout (KO) male mice were treated with or without E2.
Results
At week 2, Ninj1 KO mice exhibited attenuated colitis symptoms than WT mice following AOM/DSS treatment. E2 administration significantly alleviated these symptoms in both WT and Ninj1 KO mice, with reductions in the disease activity index, colon length shortening, and histopathological damage. The levels of pro-inflammatory mediators were reduced by E2 treatment in both groups, with the Ninj1 KO group showing a more pronounced response. At week 13, tumor development in Ninj1 KO mice was significantly lower than that in WT mice, particularly in the distal colon. E2 treatment inhibited tumor formation in WT mice and had a stronger inhibitory effect on distal colon tumorigenesis in Ninj1 KO mice. Immune cell populations, including the populations of macrophages and T cells, were also modulated by E2 in WT mice; however, these effects were diminished in Ninj1 KO mice.
Conclusion
These findings suggest that Ninj1 plays a role in modulating colitis and CRC progression, with E2 exerting anti-inflammatory and anti-tumorigenic effects that are influenced by Ninj1 status.
Introduction
Colorectal cancer (CRC) ranks as the third most frequently diagnosed cancer and a leading cause of cancer-related deaths in the United States [1]. CRC shows notable sex-specific variations in its incidence and outcomes [1]; although it generally occurs more frequently in males, recent findings have also indicated improved responses to adjuvant chemotherapy in females with right-sided tumors than in males [2]. The risk of CRC is lower in women with hereditary non-polyposis cancer syndrome [3]. However, female patients with CRC aged > 65 years demonstrate increased mortality and decreased 5-year survival rates compared to male patients in the same age group [4]. Globally, CRC exhibits sex-specific differences [5], suggesting a potential link between sex hormones and CRC development. The 2023 Cancer Statistics Report reaffirmed the prevalence of CRC, emphasizing its higher incidence in males and the clinical distinctions related to tumor location, immune response, and microsatellite instability (MSI) status [5], with a notable concentration of tumors with high MSI in the proximal colon [6].
In terms of sex hormones and CRC, estrogen, a key female hormone, has been shown to play a protective role against colon tumorigenesis [7], neurodegenerative diseases [8], and cardiovascular diseases [9] in both sexes. Higher levels of female hormones, influenced by factors such as pregnancy and hormone replacement therapy, are associated with a lower risk of CRC in females [10]. For example, clinical studies like the Women’s Health Initiative indicated a 40% reduction in the likelihood of invasive CRC with combined estrogen and progestin therapy [7]. However, the relationship between testosterone levels and CRC outcomes remains rather unclear. Testosterone, a primary male hormone, may contribute to the development of colon adenomas, thereby explaining the greater vulnerability of males to CRC; in fact, this hypothesis is supported by the results of animal studies. Specifically, male hormones such as dihydrotestosterone and testosterone enanthate have been shown to accelerate adenoma formation in the colon [11]. Interestingly, a study involving Japanese postmenopausal women found that elevated testosterone levels were associated with a 2.1-fold increase in CRC risk [12].
Nerve injury-induced protein 1 (Ninjurin1, Ninj1) is a 17-kDa cell adhesion molecule involved in processes such as nerve regeneration, cell migration, and immune response regulation [13]. Its expression, which is triggered by stress within the tumor microenvironment [13], is associated with cancer progression [14] and inflammation [15], particularly through its role in macrophage-mediated responses. Ninj1 is overexpressed in various cancers, including lung cancer [16], hepatocellular carcinoma [17], and CRC, and its expression in CRC is correlated with tumor growth, metastasis, and immune cell infiltration. In contrast, conflicting studies have reported that Ninj1 knockout (KO) models show more severe colitis in male mice and that Ninj1 overexpression reduces tumor growth in female mice with CRC [18]. Because Ninj1 regulates immune responses by interacting with macro-phages and lymphocytes [15], its role may differ in various microenvironments. Nevertheless, the precise roles of Ninj1, especially those related to sex-based differences in colitis and CRC, remain unclear.
The azoxymethane (AOM) and dextran sodium sulfate (DSS) mouse model is frequently used to explore the molecular mechanisms of colitis-associated CRC [19]. This model, which is generated by combined administration of AOM and DSS, effectively mimics the multistage progression of tumors and has been instrumental in studying sex-based differences in CRC development [19]. Research has shown that male mice experience more aggressive CRC than female mice following AOM/DSS treatment [20]. Administration of 17β-estradiol (E2) postpones CRC development by influencing the nuclear factor erythropoietin 2-related factor 2 (Nrf2) signaling pathways [21], whereas the elimination of endogenous estrogen through ovariectomy results in an increased number of tumors in the proximal colon [22]. In human female patients with CRC, Nrf2 mutations and programmed death ligand 1 expression have been shown to be involved in right-sided CRC (unpublished data). In terms of male sex hormones, a testosterone-deficient orchiectomy model demonstrated a link between testosterone levels and the formation of larger tumors, particularly in the distal colon, with the development of invasive submucosal cancers [23]. Furthermore, Ninj1 has been shown to be involved in CRC development in a testosterone-dependent manner [24]. On the basis of this background, we hypothesized that Ninj1 may amplify the inflammatory and tumorigenic responses, which could be reduced by E2 in the cascade of CRC. Thus, this study aimed to examine the role of Ninj1 in colitis-associated colon tumorigenesis based on tumor location and its potential interaction with E2. Male mice with and without Ninj1 were subjected to AOM/DSS treatment to induce colitis and CRC, and E2 was administered to evaluate its effects within this experimental framework.
Materials and Methods
1. Chemicals
AOM (# A5486) and E2 (# E8875) were purchased from Sigma-Aldrich, and DSS (# 160110) was purchased from MP Biomedicals. Phosphate-buffered saline (PBS), olive oil, and water were used to dissolve AOM, E2, and DSS, respectively.
2. WT and Ninj1 KO mice
Ninjurin1 KO (Ninj1−/−) mice (B6.129P2-Ninj1tm1Gto, generated by G. T. Oh) were kindly provided by Prof. K-W. Kim from Seoul National University. Wild-type (WT) (Ninj1+/+) and Ninj1 KO (Ninj1−/−) mice were obtained by backcrossing C57BL/6 mice (Orient Bio) and Ninjurin1 KO (Ninj1−/−) mice. The breeding colony was established and kept in cages at 23°C with a 12-hour light/dark cycle under controlled, pathogen-free conditions. Genomic DNA was extracted using a DNeasy Blood & Tissue Kit (Qiagen GmbH) and served as the template for genotyping by polymerase chain reaction (PCR) (Table 1).
3. Development of mouse models for colitis and CRC
Male WT and Ninj1 KO mice were randomly allocated to three groups: control, AOM/DSS, and AOM/DSS+E2. The method for establishing a colitis and CRC mouse model using AOM and DSS treatment has been described previously. Briefly, 1 week after intraperitoneal injection of 10 mg/kg AOM in mice in the AOM/DSS and AOM/DSS+E2 groups, 1.5% DSS was supplied in drinking water for 7 days (Fig. 1). Mice in the AOM/DSS+E2 group received intraperitoneal injections of E2 at a dose of 10 mg/kg once daily for 7 days (Fig. 1). The dose of E2 was determined on the basis of previous dose-effect relationship studies conducted by our research group. Mice in all other groups received olive oil. The animals were euthanized by CO2 asphyxiation at 2 (9 weeks of age) and 13 (20 weeks of age) weeks after AOM injection.

Experimental scheme to evaluate the role of E2 on AOM/DSS-induced colitis (week 2) and CRC (week 13) in male WT and Ninj1 KO mice. AOM (10 mg/kg) is administered to the mice on day 0. One week later, DSS (1.5%) is provided in the drinking water for 1 week. E2 (10 mg/kg) was administered by daily intraperitoneal injection for 1 week during the DSS treatment. The mice are sacrificed at weeks 2 and 13, following the AOM injection. AOM, azoxymethane; CRC, colorectal cancer; DSS, dextran sodium sulfate; E2, 17β-estradiol; KO, knockout; Ninj1, ninjurin1; WT, wild-type.
4. Evaluation of disease activity index
Clinical symptoms were assessed by evaluating the disease activity index (DAI), which rates body weight loss, stool consistency, and blood in the stool. Briefly, two researchers scored the DAI in a blinded manner by averaging the scores of weight loss, diarrhea, and rectal bleeding.
5. Macroscopic tumor assessment
Briefly, the colon was longitudinally opened, and fecal matter was removed by rinsing with PBS. The colon length (from cecum to rectum) was measured using a ruler. Polyps were counted independently by investigators unaware of the experimental details.
6. Evaluation of colonic epithelial damage score
For the evaluation of colonic epithelial damage, whole colons from 2-week samples were longitudinally opened and divided into proximal and distal parts. The distal colons were fixed in 4% paraformaldehyde, embedded in paraffin, sectioned at 5 μm thickness, deparaffinized, and stained with hematoxylin and eosin (H&E). Two sections were prepared for each sample, and the H&E-stained slides were scanned using a Pannoramic 250 Flash III digital slide scanner (3DHISTECH Ltd.) for analysis. The scanned images for entire tissue area were reviewed using 3DHISTECH Case-Viewer ver. 2.3.
Colonic damage scores were calculated by summing the scores for crypt damage and inflammatory cell infiltration [25,26]. Crypt damage was scored on a scale of 0 to 6: 0, normal; 1, hyperproliferation, irregular crypts, and goblet cell loss; 2, mild to moderate crypt loss (10%-50%); 3, severe crypt loss (50%-90%); 4, complete crypt loss with intact surface epithelium; 5, small- to medium-sized ulcer (< 10 crypt widths); 6, large ulcer (≥ 10 crypt widths). Inflammatory cell infiltration was scored separately for the mucosa (0, normal; 1, mild; 2, modest; 3, severe), submucosa (0, normal; 1, mild to modest; 2, severe), and muscle/serosa (0, normal; 1, moderate to severe). The total colonic epithelial damage score ranged from 0 to 12. The evaluation was conducted by two researchers in a blind manner.
7. Evaluation of adenoma and carcinoma
Whole colons were separated into proximal and distal parts, with the upper section defined as the proximal part and the lower section as the distal part of the colon. Briefly, colonic tissues with abnormal lesions were fixed in a 4% paraformaldehyde solution. After embedding in paraffin, each section was stained with H&E. For the 13-week samples, adenomas and cancers were identified by a specialized histopathologist (E. Shin) who was blinded to the experimental protocol.
8. Quantitative real-time polymerase chain reaction
Total RNA was extracted from colon tissue samples collected at 2 weeks and 13 weeks using TRIzol reagent (#15596026, Invitrogen). In particular, RNA for the 13-week examination was prepared from the colon polyps of AOM/DSS-treated mice and from the normal colon tissue of AOM/DSS-untreated mice. For quantitative real-time polymerase chain reaction (qRT-PCR), 2 μg of total RNA was converted to cDNA using a High-Capacity cDNA Reverse Transcription Kit following the manufacturer’s protocol (#4368814, Applied Biosystems). qRT-PCR was conducted with specific primers (Table 1) and 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 Gapdh.
9. Evaluation of inflammatory mediators
Total protein lysates were prepared, and enzyme-linked immunosorbent assay (ELISA) was conducted to determine the levels of myeloperoxidase (MPO), interleukin (IL)-1β, and IL-6 in colonic tissues. A mouse MPO ELISA kit (# HK210) was obtained from Hycult Biotechnology, while mouse IL-1β (# MLB00C) and IL-6 (# M6000B) ELISA kits were sourced from R&D Systems Inc. Notably, 13-week samples were categorized into tumor and non-tumor groups before ELISA analysis. All assays were conducted in triplicate.
10. Quantification of immune cell populations in the lamina propria of colon tissue
Lamina propria was extracted from the whole colon using a mouse Lamina Propria Dissociation Kit (#130-097-410), gentleMACS C tube (#130-096-334), and gentleMACS Octo Dissociator with Heaters (#130-096-427), all purchased from Miltenyi Biotec., following the the manufacturer’s protocol. All the steps were conducted on ice. In brief, small pieces of the colon were placed into a gentleMACS C tube containing an enzyme mixture prepared according to the provided instructions. The samples were then incubated in the gentleMACS Octo Dissociator at 37°C for 30 minutes (Program: 37C_m_LPDK_1). After centrifugation, the digested cell suspension was passed through a 70-μm cell strainer (#352350, BD Biosciences).
Subsequently, the colonic lamina propria suspensions were incubated with an Fc receptor-blocking antibody (#156604, BioLegend Inc.) in FACS buffer (0.5% fetal bovine serum in PBS) for 30 minutes to prevent nonspecific antibody binding. This step was followed by staining with fluorescently labeled specific antibodies. Details of the antibodies used can be found in the flow cytometry section of Table 2. Data acquisition was performed using a BD FACSAria III flow cytometer (BD Biosciences), and the results were analyzed with BD FACSDiva Software ver. 9.0.1 (BD Biosciences). Cell subtypes were identified based on the presence (expressed as+) of specific surface markers. Gating strategies are outlined in Table 3.
11. Statistical analysis
Data are presented as mean±standard error of the mean using GraphPad Prism software ver. 8.01 (GraphPad Software). Data were analyzed using PASW Statistics for Windows ver. 18.0 (SPSS Inc.). Statistical significance was determined using the Mann-Whitney U test or Fisher’s exact test. Values of p < 0.05 were considered statistically significant.
Results
1. E2 supplementation and Ninj1 KO suppressed colon inflammation at week 2
To assess the effects of Ninj1 on the anti-inflammatory effects of E2, we compared the findings for WT and Ninj1 KO male mice that received E2 following AOM/DSS administration with those that did not receive E2 (Fig. 1). First, we analyzed the colitis-related symptoms, including the DAI, colon length shortening, and colonic epithelial damage. In the WT AOM/DSS group, the DAI peaked at week 2 and subsequently decreased, whereas the Ninj1 KO AOM/DSS group showed an elevated DAI from weeks 2 to 3 (Fig. 2A). In the assessment of the E2 supplementation effect, the AOM/DSS-induced DAI significantly decreased with E2 treatment at week 2 in WT males (p=0.003) (Fig. 2B). E2 supplementation in the Ninj1 KO male AOM/DSS group resulted in reduction rates (58%) similar to those observed in the WT males (58%), but without statistical significance (Fig. 2B, left panel). At week 3, the increase in the DAI score induced by AOM/DSS treatment was strongly suppressed by E2 treatment in both WT and Ninj1 KO males (p=0.002 for WT, p=0.002 for Ninj1 KO; Fig. 2B, right panel). However, the shortening of colon length induced by AOM/DSS was not relieved by E2 treatment in both WT and Ninj1 KO males (Fig. 2C). Representative histopathological images are shown in Fig. 2D. Assessment of the colonic epithelial damage scores revealed that AOM/DSS-induced crypt loss and inflammatory cells in the colon tissues were inhibited by E2 supplementation in the Ninj1 KO males (p=0.005), similar to the findings in the WT males (p=0.027) (Fig. 2E). In terms of the impact of Ninj1 KO, the increased colitis-related symptoms with AOM/DSS treatment were alleviated in the Ninj1 KO males in comparison with the WT males, but only DAI and colonic epithelial damage score showed significant differences (Fig. 2).

Effect of E2 supplementation in male WT and Ninj1 KO mice in colitis symptoms. (A, B) DAI score during the experimental period (A) and at weeks 2 (B, left panel) and 3 (B, right panel). (C) Effect of E2 on AOM/DSS-mediated colon length shortening in male WT and Ninj1 KO mice at week 2. (D) Representative H&E staining images of colon tissues at week 2 (×100). The crypt within the colon tissues is normal in the WT and Ninj1 KO control mice. However, crypt loss and strong inflammatory cell infiltration within the colon tissues (red arrows) are observed in the male AOM/DSS-treated WT and Ninj1 KO mice. In the WT and Ninj1 KO groups, AOM/DSS-induced histologic damage is inhibited by E2 supplementation with only mild erosion (green arrows) visible. (E) The inhibitory effect of E2 on AOM/DSS-induced colonic epithelial damage is seen in male Ninj1 KO mice at a rate similar to that of male WT mice. (F-H) Determination of levels of pro-inflammatory mediators such as MPO (F), IL-1β (G), and IL-6 (H) by enzyme-linked immunosorbent assay at week 2. Data are expressed as the mean±SEM. Mann-Whitney U test for comparison difference between independent two groups was performed. *p < 0.05 for intergroup comparison for AOM/DSS vs. CON or AOM/DSS+E2 group. #p < 0.05 for WT vs. Ninj1 KO mice in CON, AOM/DSS, and AOM/DSS+E2 group. AOM, azoxymethane; CON, control; DSS, dextran sodium sulfate; E2, 17β-estradiol; H&E, hematoxylin and eosin; IL-1β, interleukin-1β; IL-6, interleukin 6; KO, knockout; MPO, myeloperoxidase; Ninj1, Ninjurin1; ns, not significant; SEM, standard error of the mean; WT, wild-type.
Subsequently, we conducted an ELISA to assess the levels of MPO, IL-1β, and IL-6, which are mediators linked to colitis, in the colon tissues at week 2. In terms of E2 supplementation effect, the levels of MPO, IL-1β, and IL-6 increased by AOM/DSS treatment were diminished by E2 administration in both WT and Ninj1 KO mice (Fig. 2F-H). However, only MPO and IL-1β levels showed significant reduction with E2 supplementation in Ninj1 KO mice in comparison with the WT mice (p=0.049 for MPO and p=0.027 for IL-β in WT; p=0.014 for MPO and p=0.014 for IL-β in Ninj1 KO) (Fig. 2F and G). In terms of the influence of Ninj1 KO, the increased IL-β and IL-6 ELISA levels with AOM/DSS treatment were significantly alleviated in the Ninj1 KO mice in comparison with the WT mice (Fig. 2G and H). Collectively, these data indicate that E2 and Ninj1 KO independently mitigated DSS-induced colitis, as evidenced by improvements in the DAI, prevention of colon shortening, reduced histopathological damage, and modulation of inflammatory mediators.
2. Ninj1 modulates tumor development and E2-mediated inhibition at week 13
Next, to evaluate the role of Ninj1 in the antitumor effect of E2, the number of tumors was classified according to size (≤ 2 mm and > 2 mm) and location (proximal and distal) by macroscopic assessment at week 13. The experimental scheme is shown in Fig. 1. In the WT mice, tumors were well-developed at week 13 after AOM/DSS administration, primarily in the distal colon in comparison to the proximal colon (Fig. 3A and B), which is consistent with the findings of a previous report [27]. Ninj1 KO strongly inhibited the development of small tumors (≤ 2 mm) induced by AOM/DSS in the distal colon in comparison with the findings in WT mice (p=0.014) (Fig. 3A). E2 treatment did not affect the number of tumors in either group (Fig. 3A). In the assessment of large tumors (> 2 mm), Ninj1 KO mice exhibited significantly more tumors in the proximal colon (p=0.001) than WT mice. E2 showed a strong inhibitory effect on distal colon tumors in WT mice (54% reduction, p=0.015). However, Ninj1 KO mice did not show a significant reduction (47%) (Fig. 3B), indicating that Ninj1 may play a role in modulating tumor development. The data for the incidence of microscopic neoplasms and tumor multiplicity in the colon are shown in Table 4. Representative histopathological H&E-stained images are shown in Fig. 3C. There were no inflammatory polyps; all cases were classified as adenoma or carcinoma. In Ninj1 KO mice, E2 supplementation caused significant inhibition (p=0.015) of the occurrence of adenomas and cancers induced by AOM/DSS treatment in the distal colon; adenomas and cancers were observed in only two out of nine mice. In contrast, the inhibitory effect of E2 on tumors originating from the distal colon was insufficient in WT mice; adenomas and cancers were observed in seven out of nine mice (Table 4, Fig. 3D). On the other hand, Ninj1 KO mice showed a smaller proportion of cancers with mucosal and submucosal invasion induced by AOM/DSS treatment than WT mice (67% in Ninj1 KO AOM/DSS vs. 100% in WT AOM/DSS). Interestingly, the submucosal invasive cancers in the distal colon observed in WT mice (20%; one out of five mice) did not develop in any of the nine Ninj1 KO mice (Table 4, Fig. 3D). Taken together, these findings indicate that E2 exhibited a more pronounced inhibitory effect on distal colon carcinogenesis in Ninj1 KO mice than in WT mice.

Anti-tumorigenic effect of E2 in both male WT and Ninj1 KO mice at week 13. (A, B) Average number of small tumors (≤ 2 mm) (A) and large tumors (> 2 mm) (B) in the proximal, distal, and whole colon of each group sacrificed at week 13 after the AOM injection. (C) Representative H&E staining images (×40). Black and red arrow heads indicate tumors and submucosal invasive cancer, respectively. (D) Quantification of adenoma/carcinoma incidence and invasion in each group by microscopic evaluation of the colonic tissues. Data are expressed as the mean±SEM. Mann-Whitney U test for comparison difference between independent two groups was performed. *p < 0.05 for intergroup comparison for AOM/DSS vs. CON or AOM/DSS+E2 group. #p < 0.05 for WT vs. Ninj1 KO mice in CON, AOM/DSS, and AOM/DSS+E2 group. AOM, azoxymethane; CON, control; DSS, dextran sodium sulfate; E2, 17β-estradiol; H&E, hematoxylin and eosin; KO, knockout; Ninj1, Ninjurin1; ns, not significant; SEM, standard error of the mean; WT, wild-type.
3. Suppressive effect of E2 supplementation and Ninj1 KO on pro-inflammatory mediator expression at week 13
To analyze the expression of pro-inflammatory mediators in 13-week samples, samples from the AOM/DSS and AOM/DSS+E2 groups were prepared from colon tumor tissue, whereas the control group samples were obtained from normal colon tissue. In the WT mice, the mRNA expression levels of pro-inflammatory mediators such as inducible nitric oxide synthase (iNos), cyclooxygenase 2 (Cox-2), interleukin-1 beta (Il-1β), and tumor necrosis factor-alpha (Tnf-α) induced by AOM/DSS treatment was significantly suppressed by E2 supplementation (p=0.009 for iNos, p=0.044 for Cox-2, p=0.020 for Il-1β, and p=0.039 for Tnf-α) (Fig. 4A). However, in the Ninj1 KO group, AOM/DSS treatment did not induce the mRNA expression of these mediators, except for Tnf-α (p=0.006), indicating that E2 supplementation had no inhibitory effect in all targets (Fig. 4A).

Effect of E2 administration in male WT and Ninj1 KO mice in the levels of pro-inflammatory mediators in colonic tissues. (A) AOM/DSS-treated group samples are prepared from colon tumor tissue, and control samples are prepared from normal colon tissue to analyze mRNA expression of pro-inflammatory mediators such as iNos, Cox-2, Il-1β, and Tnf-α by quantitative real-time polymerase chain reaction analysis at week 13. (B-D) Determination of levels of pro-inflammatory mediators such as MPO (B), IL-1β, (C) and IL-6 (D) by enzyme-linked immunosorbent assay in the tumor and non-tumor tissues of the colon at week 13. Data are expressed as the mean±SEM. Mann-Whitney U test for comparison difference between independent two groups was performed. *p < 0.05 for intergroup comparison for AOM/DSS versus CON or AOM/DSS+E2 group. #p < 0.05 for WT versus Ninj1 KO mice in CON, AOM/DSS, and AOM/DSS+E2 group. AOM, azoxymethane; Con, control; Cox-2, cyclooxygenase 2; DSS, dextran sodium sulfate; E2, 17β-estradiol; IL-1β, interleukin-1β; IL-6, interleukin 6, iNos, inducible nitric oxide synthase; KO, knockout; MPO, myeloperoxidase; Ninj1, Ninjurin1; ns, not significant; SEM, standard error of the mean; Tnf-α, tumor necrosis factor-α; WT, wild-type.
We also conducted ELISA to assess the levels of MPO, Il-1β, and IL-6, which are mediators linked to colitis-associated tumorigenesis, in the colon tumors at week 13. The 13-week samples were divided into tumor and non-tumor tissues for ELISA analysis. In the mice that received AOM/DSS, the MPO level was significantly higher in tumor tissues than in non-tumor tissues (p=0.008) (Fig. 4B). The AOM/DSS-induced increase in the MPO levels in the tumor tissue was suppressed by E2, regardless of Ninj1 status, but the difference was not significant (Fig. 4B, upper panel). However, in the non-tumor tissues of WT mice, AOM/DSS-induced MPO levels were significantly suppressed by E2 treatment (p=0.014) (Fig. 4B, lower panel). In contrast, in Ninj1 KO mice, the initial MPO levels in non-tumor tissues were not increased by AOM/DSS treatment, and E2 showed no inhibitory effect (Fig. 4B, lower panel). The ELISA results for Il-1β and IL-6 showed similar responses to E2 supplementation and Ninj1 KO in both tumor and non-tumor tissues (Fig. 4C and D). Taken together, these findings indicate that E2 suppressed the expression of pro-inflammatory mediators induced by AOM/DSS in WT mice, whereas Ninj1 KO mice showed minimal induction with no significant E2 inhibitory effect.
4. Influence of E2 supplementation and Ninj1 KO on the immune cell populations at week 13
To assess the influence of E2 supplementation and Ninj1 KO on the immune cell population within the colon during the tumorigenesis stage, we further analyzed the frequency of various immune cell populations, including macrophages (CD11b+ F4/80+) such as M1 macrophages (CD11b+ F4/80+ CD86+), M2 macrophages (CD11b+ F4/80+ CD206+), T cells (CD3e+), CD4+ T cells (CD3e+ CD4+), and CD8+ T cells (CD3e+ CD8a+), using the gating strategy (Figs. 5A and 6A). In the WT group, E2 supplementation did not affect the overall macrophage populations (Fig. 5B). However, it reversed the decrease in M1 macrophages (p=0.049) (Fig. 5C) and the increase in M2 macrophages (p=0.049) (Fig. 5D) caused by AOM/DSS treatment. Furthermore, E2 treatment resulted in a slight increase in the frequency of the T-cell population in comparison with that in the AOM/DSS group (p=0.049) (Fig. 6B). Similarly, the CD4+ T-cell population showed similar regulation to that of the overall T-cell population in response to E2 treatment (p=0.049) (Fig. 6C). The increase in the CD8+ T-cell population induced by AOM/DSS treatment was counteracted by E2 treatment (p=0.049) (Fig. 6D). However, the observed differences caused by AOM/DSS or E2 treatment were not evident in the Ninj1 KO group (Figs. 5 and 6).

Cell population analysis through flow cytometry in the colonic lamina propria of male WT and Ninj1 KO mice at week 13. (A) Gating strategy. Single cells were gated into P1 using FSC and SSC. Afterwards, P1 was gated once again with FSC-A and FSC-H, and this gate was used as the total cell to be analyzed later. Macrophages in the total cell population were gated based on CD11b+ F4/80+ cells, and this macrophage population was further divided into CD86+ M1 macrophages (CD11b+ F4/80+ CD86+) and CD206+ M2 macrophages (CD11b+ F4/80+ CD206+). (B-D) Frequencies of macrophages (CD11b+ F4/80+) (B), M1 macrophage (CD11b+ F4/80+ CD86+) (C), and M2 macrophage populations (CD11b+ F4/80+ CD206+) (D). Data are expressed as the mean±SEM. Mann-Whitney U test for comparison difference between independent two groups was performed. *p < 0.05 for intergroup comparison for AOM/DSS versus CON or AOM/DSS+E2 group. #p < 0.05 for WT versus Ninj1 KO mice in CON, AOM/DSS, and AOM/DSS+E2 group. AOM, azoxymethane; CON, control; DSS, dextran sodium sulfate; E2, 17β-estradiol; FSC, forward scatter; KO, knockout; Ninj1, Ninjurin1; SEM, standard error of the mean; SSC, side scatter; WT, wild-type.

Cell population analysis through flow cytometry in the colonic lamina propria of male WT and Ninj1 KO mice at week 13. (A) Gating strategy. Single cells were gated into P1 using FSC and SSC. Afterwards, P1 was gated once again with FSC-A and FSC-H, and this gate was used as the total cell to be analyzed later. In the total cell population, cells were gated into T cells (CD3e+), CD4 T cells (CD3e+ CD4+), and CD8 T cells (CD3e+ CD8+). (B-D) Frequencies of T cells (CD3e+) (B), CD4 T cells (CD3e+ CD4+) (C), and CD8 T cells (CD3e+ CD8+) (D). Data are expressed as the mean±SEM. Mann-Whitney U test for comparison difference between independent two groups was performed. *p < 0.05 for intergroup comparison for AOM/DSS versus CON or AOM/DSS+E2 group. #p < 0.05 for WT versus Ninj1 KO mice in CON, AOM/DSS, and AOM/DSS+E2 group. AOM, azoxymethane; CON, control; DSS, dextran sodium sulfate; E2, 17β-estradiol; FSC, forward scatter; KO, knockout; Ninj1, Ninjurin1; SEM, standard error of the mean; SSC, side scatter; WT, wild-type.
Discussion
In this study, we explored the role of Ninj1 in modulating inflammation and tumorigenesis in colitis-associated CRC, with a specific focus on its interaction with the female sex hormone E2. Previous research primarily examined the role of Ninj1 in inflammation and tumor development, emphasizing its involvement in immune modulation and its interactions with sex-specific factors, such as testosterone [24]. In contrast, this study introduces E2 supplementation as a novel aspect, exploring its anti-inflammatory and anti-tumorigenic effects in Ninj1 KO mice. Our findings indicate that Ninj1 KO mice exhibit different inflammatory and tumorigenic responses than WT mice in the AOM/DSS model. Additionally, E2 supplementation demonstrated protective effects against both inflammation and tumor development, particularly in the Ninj1 KO group, indicating a potential interaction between Ninj1 and estrogen signaling in the regulation of colitis-associated CRC. It provides new insights into sex hormone-dependent regulation of inflammation and tumor development, underscoring the importance of considering estrogen signaling in colitis and CRC pathogenesis, particularly in the context of Ninj1 deficiency.
Combining findings from the previous [24] and current study, these studies underscore the distinct roles of E2 and Ninj1 KO in mitigating colitis, highlighting their anti-inflammatory effects and the influence of sex-specific factors. In AOM/DSS-treated WT mice, colitis severity, indicated by DAI scores and histopathological damage, was more pronounced in males than females [24]. Ninj1 KO significantly alleviated these symptoms in both sexes, with a stronger impact on males, reducing DAI scores, inflammatory mediator levels, and colonic damage [24]. Furthermore, these findings are in line with previous research demonstrating the protective effects of estrogen against CRC, likely due to its anti-inflammatory properties and its modulation of Nrf2-mediated pathways [7,8]. In the present study, E2 supplementation demonstrated potent anti-inflammatory effects in male mice, significantly reducing DAI scores, crypt loss, and inflammatory mediators such as MPO, IL-1β, and IL-6. E2 administration similarly reduced DAI scores and levels of key inflammatory markers during the early stages of colitis (week 2) in both WT and Ninj1 KO mice [7,8,10]. These results are consistent with earlier studies showing that estrogen can suppress colonic inflammation [7,8]. While Ninj1 KO and E2 independently alleviated colitis severity, the effects of E2 treatment were comparable in both WT and Ninj1 KO males, suggesting their additive or indirect conditional effects rather than a direct interaction between Ninj1 and E2. Notably, the absence of Ninj1 enhanced the reduction of specific inflammatory markers, such as MPO and IL-1β, with E2 treatment. This significant reduction in pro-inflammatory mediators observed in the Ninj1 KO group further suggests that Ninj1 contributes to the amplification of the inflammatory response, indicating that Ninj1 plays a role in regulating macrophage-mediated inflammation [9,10]. Together, these findings underscore Ninj1 as a critical mediator of inflammation, particularly in males, and highlight the therapeutic potential of E2 in colitis. They emphasize the importance of incorporating sex-specific strategies in developing treatments for colitis and related diseases.
The tumorigenic response observed in this study highlights the complex role of Ninj1 in CRC development. While Ninj1 KO mice exhibited fewer small tumors (≤ 2 mm) than WT mice, they developed significantly more large tumors (> 2 mm) in the proximal colon at week 13, suggesting that Ninj1 plays a more prominent role in the progression of larger tumors than on the formation of smaller tumors. This study further underscores the distinct effects of Ninj1 deficiency and E2 treatment on tumor development in the distal and proximal colon, particularly in relation to tumor size and incidence. In WT male mice treated with AOM and DSS, tumor formation was predominantly observed in the distal colon, with significantly fewer tumors detected in the proximal colon. However, large tumors (> 2 mm) were significantly more frequent in the proximal colon of Ninj1 KO mice compared to WT mice, highlighting the critical role of Ninj1 in suppressing tumor growth in this region. This observation echoes findings from earlier studies on other cancers, where Ninj1 was shown to promote tumor growth and metastasis by regulating cell adhesion and immune cell infiltration [13,18]. Moreover, E2 treatment exhibited a potent inhibitory effect on tumor development in the distal colon of WT male mice, reducing the tumor burden by 54% following AOM/DSS-induced tumorigenesis. Importantly, in Ninj1 KO mice, E2 supplementation also significantly suppressed the formation of adenomas and cancers in the distal colon, demonstrating the retained antitumor efficacy of E2 despite the absence of Ninj1. However, the effects of E2 in the proximal colon were not significant, and the present study did not provide evidence of substantial modulation by either Ninj1 deficiency or E2 treatment in this region. Notably, our findings revealed that E2 reduced tumor incidence and multiplicity and suppressed the transition to invasive submucosal cancer, highlighting the protective role of estrogen in CRC, especially CRC in the distal colon, which is more prevalent in males [28]. Collectively, these results highlight the regional specificity of Ninj1 and E2 effects, with E2 exerting pronounced tumor-suppressive effects in the distal colon. While Ninj1 deficiency exacerbates tumor development in the proximal colon, it does not completely abrogate the inhibitory effects of E2 in the distal colon. These findings warrant further investigation into the mechanisms underlying the regional differences observed and their implications for CRC progression and therapeutic strategies.
Interestingly, the influence of Ninj1 on immune cell populations appears to be significant. E2 supplementation reverses AOM/DSS-induced alterations in macrophage polarization and T-cell populations in WT mice; however, these effects are diminished in Ninj1 KO mice [19,29]. This finding suggests that Ninj1 may play a role in shaping the tumor immune microenvironment, potentially by influencing macrophage and lymphocyte activity, which are critical for inflammation and tumor progression [29]. Although our study supports the notion that Ninj1 exacerbates inflammation and tumor development, the interaction between Ninj1 and E2, particularly with respect to immune cell regulation, requires further investigation [19,29].
Colitis-associated CRC accounts for a minor proportion of overall CRCs in humans. While the AOM/DSS model used in this study primarily represents inflammation-driven carcinogenesis, resembling ulcerative colitis-associated cancer, our findings may provide insights relevant to other CRC pathogenesis, including non-IBD–based CRCs such as chromosomal instability (CIN)– or MSI-driven cancers. CRC exhibits distinct molecular and pathological characteristics depending on tumor location and sex [5]. In males, left-sided CRC is more common and is often associated with CIN, which accounts for 60%-70% of CRCs. This type of CRC frequently involves mutations in genes such as adenomatous polyposis coli, Kirsten-ras (KRAS), deleted in colorectal cancer, and p53 [5]. Conversely, right-sided CRC is more prevalent in females and is often characterized by MSI-high, CpG island methylator phenotype–high, and BRAF mutations [5]. These molecular differences underline the need for tailored approaches when studying and treating CRC. In our AOM/DSS model, CRC occurred more frequently in males than in females, which aligns with the sex differences observed in human CRC. Moreover, the model demonstrated estrogen-mediated anti-inflammatory and anti-tumorigenic effects, which are also consistent with the protective role of estrogen observed in human colorectal cancer. Although we did not directly measure CIN in this study, the unexpected parallels between the AOM/DSS model and human CRC, particularly regarding sex differences, suggest that inflammation-driven mechanisms explored in this model may intersect with pathways involved in CIN- and MSI-driven carcinogenesis.
Our study adds to a growing body of evidence regarding the involvement of sex hormones in the development and progression of CRC [7,10,30]. The differential responses of male mice to AOM/DSS-induced colitis and CRC highlight the importance of sex-based differences in experimental models [20,23]. The protective effect of E2 against inflammation and tumorigenesis, as well as the exacerbation of colitis symptoms following the removal of endogenous estrogen, supports previous findings linking estrogen levels to CRC risk in females [28,30]. Moreover, the link between testosterone and increased tumor size, as observed in orchiectomy models, may limit the potential role of androgens in CRC development [11,23].
This study had several limitations. First, although the AOM/DSS mouse model is effective for studying colitis-associated CRC, colitis-associated CRC accounts for a minor proportion of overall CRCs. That is, it does not fully replicate the complexity of human CRC, limiting the direct applicability of the findings to clinical cases. Nevertheless, the results obtained with this model provide evidence to support the role of Ninj1 in human CRC. Second, although we observed changes in immune cell populations, the underlying molecular mechanisms by which Ninj1 influences immune responses have not yet been thoroughly investigated. Finally, the relatively short observation period limited insights into long-term tumor progression and metastasis, which are critical for CRC prognosis. Future studies addressing these points will enhance our understanding of the role of Ninj1 in CRC and its potential as a therapeutic target. Previously, we adopted the Nancy criteria to investigate the molecular activity of inflammation and tissue remodeling markers in endoscopically inflamed and uninflamed regions of ulcerative colitis (UC) [31]. Our findings revealed that mRNA expressions of transforming growth factor β, IL-1β, OLFM4, ferroptosis suppressor protein 1, vimentin, and α-smooth muscle actin were significantly higher, while that of E-cadherin was significantly lower in both inflamed and uninflamed regions of patients with UC compared to controls [31]. The Nancy criteria, a widely recognized grading system for evaluating inflammation in human UC, provide a detailed and objective framework [32]. However, for the current study focusing on colitis-associated cancer development, the Nancy grading system’s complexity and its primary emphasis on inflammation were less suited to our specific research objectives. Instead, we adopted a simplified scoring approach tailored to the study’s goals. While this simplified approach was appropriate for our study, it may limit direct comparisons with research employing the Nancy criteria. This highlights an area for refinement in future studies, where incorporating established systems like the Nancy criteria could improve objectivity and enable better comparability with existing literature.
In conclusion, our findings suggest that Ninj1 plays a significant role in modulating inflammation and tumor progression in colitis-associated CRC, and that E2 exerts protective effects, particularly in the absence of Ninj1. The interaction between Ninj1 and estrogen signaling pathways offers a promising avenue for further research, particularly in the context of sex-specific differences in the incidence and out- comes of CRC. Further studies are needed to explore the molecular mechanisms underlying the role of Ninj1 in immune regulation and tumor development, as well as its potential as a therapeutic target in CRC.
Notes
Ethical Statement
All animal procedures were conducted following approval from the Institutional Animal Care and Use Committee of Seoul National University Bundang Hospital (BA-1980-277-066-01) and adhered to the ARRIVE (Animals Research: Reporting In Vivo Experiments) guidelines.
Author Contributions
Conceived and designed the analysis: Song CH, Kim N.
Collected the data: Song CH.
Contributed data or analysis tools: Song CH, Shin E.
Performed the analysis: Song CH, Nam RH, Jang JY, Kim EH, Ha S.
Wrote the paper: Song CH.
Supervised the study, and critically revised the article: Kim N, Lee HN.
Provided ninjurin1 transgenic mouse: Choi H, Kim KW, Jeon S, Oh GT.
Conflicts of Interest
Conflict of interest relevant to this article was not reported.
Funding
This work was supported by grant no 06-2021-0018 and 14-2024-0009 from the Seoul National University Bundang Hospital Research fund. In addition, this work was also supported by a grant from the National Research Foundation of Korea (NRF) funded by the government of the Republic of Korea (2019R1A2C2085149).