Elevated MS4A12 expression is indicative of resistance to concurrent chemoradiotherapy and inferior survival in patients with rectal cancer

Abstract

Introduction

In individuals presenting with locally advanced rectal cancer, the therapeutic strategy of neoadjuvant concurrent chemoradiotherapy (CCRT) aims to enhance tumor downstaging; however, only a subset of patients exhibit a favorable response. Molecular stratification, combined with the traditional tumor staging system (TNM), is a promising approach for predicting treatment efficacy and patient outcomes. Therefore, we intend to better grasp the molecular basis of CCRT resistance and guide therapeutic strategies with greater precision.

Methods

We utilized a public rectal cancer transcriptomic dataset (n = 46) to predict responsiveness to neoadjuvant CCRT by analyzing signal transduction-related genes. In our well-characterized rectal cancer cohort (n = 343), we assessed correlations between membrane-spanning 4-domains A12 (MS4A12) immunostaining and clinicopathological characteristics using Pearson’s chi-squared test. To calculate survival rates, we employed the Kaplan–Meier method with a log-rank test. Additionally, we conducted multivariate analyses with the Cox proportional hazards model to identify independent prognostic biomarkers.

Results

We identified that the MS4A12 gene is highly expressed in rectal cancer resistant to CCRT. Elevated MS4A12 expression, confirmed by immunohistochemical staining, is significantly associated with advanced tumor status after CCRT (p < 0.001), positive node status both before and after CCRT (p = 0.01 and p = 0.004), the presence of perineural and vascular invasion (p = 0.006 and p = 0.001), and low or no response to CCRT (p < 0.001). Notably, high MS4A12 immunoexpression is strongly correlated with reduced patient survival in rectal cancer. Mechanically, high MS4A12 expression is significantly associated with aberrant glycosylation and B-cell infiltration.

Conclusion

MS4A12 expression may offer a helpful predictive and prognostic indicator for identifying patients who could gain advantages from neoadjuvant CCRT.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13014-025-02709-5.

Introduction

Colorectal cancer, or CRC, is a disease that primarily affects the colon or rectum. According to the World Health Organization’s GLOBOCAN database [1], CRC ranks as the third most frequently diagnosed cancer in both males and females globally. Unlike many other types of cancer, CRC can frequently be prevented through regular screening and is highly manageable when identified in its early stages. Nonetheless, the reduction in CRC incidence rates has been gradual over the past decade for all anatomic subsites (proximal and distal colon and rectum) among which rectal cancer was the most notable due to a rise in incidence in individuals under the age of 65 [2]. Specifically, there has been an annual increase of approximately 2% among those younger than 50 years and 1% among those aged 50 to 64. Unfortunately, patients who experience early onset are frequently diagnosed with advanced-stage disease [2]. At present, the conventional approach for managing locally advanced rectal cancer is concurrent chemoradiotherapy (CCRT) prior to surgery, followed by adjuvant chemotherapy [3]. While many patients experience improved local control with CCRT, there remains ongoing debate regarding its impact on overall survival [4].

The regimens of neoadjuvant CCRT comprise intravenous 5-fluorouracil (5-FU) or its oral analog capecitabine and radiation (45–50 Gy, 25–28 fractions). Although neoadjuvant CCRT could improve tumor downstaging to increase the complete resection rate of surgery and enhance the quality of life by preserving the sphincter [3], only 15–27% of patients are complete responders to this treatment approach, which is linked to favorable survival [5]. Accordingly, there is a strong requirement for reliable biomarkers that can predict treatment outcomes and distinguish patient subgroups with resistant phenotypes. While the traditional tumor staging system (TNM) is widely acknowledged as a robust predictor of clinical outcomes, it exhibits shortcomings in categorizing patient subgroups with intermediate tumor response following neoadjuvant therapy into more informative prognostic categories [6]. This limitation can potentially be addressed by evaluating tumor regression grading, which involves quantifying the extent of cellular tumor response to neoadjuvant therapy [7]. However, the TNM classification and tumor regression grading are simply accessible in post-CCRT specimens, making them unavailable in pre-CCRT samples for aiding in the planning of adjusted neoadjuvant therapy. There’s a growing recognition that cancer drug resistance involves diverse molecular mechanisms, with the dysregulation of signal transduction emerging as a pivotal contributor. Therefore, gaining a deeper understanding of the molecular mechanisms underlying CCRT resistance may offer insights into predictive biomarker identification and how to target these signaling pathways for therapeutic purposes.

The human membrane-spanning 4-domains A12 (MS4A12) gene, mapped to chromosome 11q12.2, encodes a cell surface protein observed primarily in the apical side (lumen) of the colon [8] and colorectal cancer [9]. The level of MS4A12 is strictly regulated by transcriptional repression in normal colonic epithelial cells. In colon cancer cells, MS4A12 activation has been suggested to be a component of store-operated Ca²⁺ (SOC) entry and enhance epidermal growth factor (EGF)-mediated tumor growth-promoting effects [10]. The promoter of MS4A12 in colon cancer is specifically transactivated by the transcription factor caudal type homeobox 2 (CDX2), which is originally identified to be involved in the development and differentiation of the intestinal epithelium [11]. Moreover, high expression of mucin 2 (MUC2), one of the other CDX2 target genes showing colon-confined expression, has been indicated to function as an adverse prognostic factor for rectal cancer patients undergoing CCRT in our previous study [12]. Nevertheless, low MS4A12 expression has been associated with worse overall survival in primary colorectal cancer patients not receiving tumor treatment [9]. Additionally, it has also been indicated that low expression of MS4A12 inhibits differentiation and predicts poor survival in early-stage colon cancer patients [13]. Notably, in rectal cancer cells resistant to radiation, a majority of the differentially expressed genes are linked to the control of the differentiation of epithelial cells [14]. Additionally, various studies have yielded varied outcomes, influenced by factors, such as the study population, the stages of cancer, and the modality of treatment. Accordingly, whether high expression of MS4A12 could provide a protective barrier against CCRT and predict inferior survival in locally advanced rectal cancer patients is a question that remains unanswered.

Patients and methods

Analysis of the distinctive expression profiles of signal transduction-related genes

To predict the responsiveness of rectal cancer to neoadjuvant CCRT, we employed a dataset (GSE35452) containing transcriptome profiles. In this dataset, tumor biopsies were collected from 46 rectal cancer patients prior to receiving CCRT during screening colonoscopy. RNA extraction was performed on these biopsies, and gene expression profiling was carried out using the Affymetrix Human Genome U133 Plus 2.0 Array (Affymetrix, CA) platform. Subsequently, the gene expression levels were measured using the statistical software Nexus Expression 3 (BioDiscovery, CA) [15], and all probe sets were examined without any filtering. The biopsy samples were categorized into “resistance” and “sensitivity” groups based on their response to CCRT, and a supervised comparative analysis was carried out. We conducted a differential gene expression analysis with a focus on signal transduction (GO: 0007165). A positive log2 ratio indicates higher gene expression in the resistance group. Transcripts with log2 ratios ≥ 1 and p-values < 0.01 were chosen for further investigation.

Patient recruitment and eligibility

This study (IRB10302014) received approval from the Ethics Committee and Institutional Review Board (IRB) at Chi Mei Medical Center. The formalin-fixed paraffin-embedded (FFPE) cancer tissue samples, encompassing 343 patients diagnosed with rectal cancer (from 1998 to 2017), were sourced from our biobank. All these patients were initially diagnosed with rectal cancer through colonoscopy, and no distant metastasis was detected based on chest X‐ray and/or abdominopelvic computed tomography reports. Also, patients with missing data were excluded. Through retrospective examination of their medical records, we collected clinical and pathological data as well as their clinical outcomes. Patients were regularly followed up from the time of diagnosis until their death or the last tracking available. Before undergoing surgical resection, all patients received neoadjuvant concurrent chemoradiotherapy (CCRT), consisting of a total dose of 45–50 Gy of radiotherapy (25 fractions) administered concurrently with sustained delivery of 5-fluorouracil (5-FU)-based chemotherapy. For patients with tumor status beyond T3 or nodal status beyond N1, (neo) adjuvant chemotherapy was administered prior to or following CCRT. Our rectal cancer patients typically underwent regular medical consultations and follow-up screenings at intervals of 3–6 months over a period of 5 years. Almost all patients developed events within 5 years.

Histopathological evaluation and the scoring of immunohistochemistry

To ensure an impartial evaluation, patient information was concealed from two independent pathologists, Chien-Feng Li and Wan-Shan Li, who examined all rectal cancer tissue samples. Following the guidelines of the 8th edition of the AJCC tumor-node-metastasis (TNM) staging system, the tumor and node statuses were defined. The Dworak system for grading tumor regression was employed to forecast the effectiveness of CCRT and is defined as follows: 0–1, no or little response; 2–3, modest response; 4, complete response. Immunohistochemical (IHC) staining was performed on FFPE human rectal cancer tissue slides using MS4A12 primary antibody (13,293–1-AP, Proteintech, Rosemont, IL, USA) at dilution of 1:100 and peroxidase-labeled secondary antibody (Dako EnVision Detection System, Agilent Technologies, Santa Clara, CA, USA). To quantify MS4A12 immunostaining from the IHC images, the H-score was employed and calculated according to the following formula: H-score = ΣPi (i + 1), where Pi represents the proportion (ranging from 0 to 100%) of stained tumor cells for every intensity, and i signifies the intensity (ranging from 0 to 3 +) of staining. The median H-score of all assessed tumor specimens was employed to categorize MS4A12 immunoreactivity into low level (n = 171) and high level (n = 172).

Prediction and classification of gene functions

To provide hypothesis-generating insights into MS4A12’s potential involvement in rectal cancer, we conducted an analysis of the relationships between MS4A12 mRNA expression and its coexpressed genes by Spearman’s correlation using a rectal cancer dataset comprising 788 samples from the Memorial Sloan Kettering (MSK) database in the cBioPortal online platform. Subsequently, we retrieved the top 193 positively and 73negatively correlated genes, ranked by correlation coefficient and filtered using a q-value < 0.01. These selected genes were then subjected to functional annotation using the Gene Ontology (GO) classification system covering three domains: molecular function, biological process, and cellular component, and evaluated based on fold enrichment. In essence, we performed the PANTHER overrepresentation test using Fisher’s exact test to identify GO terms that were disproportionately represented among the genes identified as differentially expressed. The p-value in this test reflects the probability of observing the given number of genes associated with a particular GO term, under the null hypothesis. A p-value below 0.05 indicates that the distribution of the identified GO term is not random and deviates from the overall distribution. To reduce the potential of false positives (or type I errors), the initial p-value underwent adjustment for multiple hypothesis testing, resulting in a statistical measure known as the false discovery rate (FDR). Maintaining the FDR below a threshold of 0.05 enables the identification of a significant number of GO terms while keeping the rate of false positives relatively low. We used an R script along with the ggplot2 package to create visual representations of the meaningful GO terms.

Statistical analysis

For the statistical analysis, we utilized the IBM SPSS statistical software version 20.0. To visualize and present the significant GO terms, we employed an R script along with the ggplot2 package. Pearson’s chi-squared test was employed to investigate the associations between MS4A12 expression and clinicopathological features. We assessed three time-to-event endpoints: disease-specific survival (DSS), local recurrence-free survival (LRFS), and metastasis-free survival (MeFS). These endpoints are defined as the duration from the date of surgery to the occurrence of cancer-related death, first local recurrence, or first metastasis. In the time-to-event analyses, individuals who were lost to follow-up were censored. For univariate analysis, we calculated survival rates using the Kaplan–Meier method and used the log-rank test to determine whether the survival distributions of the two groups differed. The clinicopathological features that exhibited statistical significance at the univariate level were considered for inclusion in the multivariate analysis using the Cox proportional hazards model. Differences with a p-value or q-value less than 0.05 were regarded as statistically significant.

Results

MS4A12 stands out as the gene with the highest increase in expression among rectal cancer patients resistant to CCRT

In an effort to predict the responsiveness of rectal cancer to neoadjuvant CCRT, we employed a dataset (GSE35452) containing transcriptome profiles. Within this dataset, tumor biopsies were collected from 46 rectal cancer patients prior to receiving CCRT during routine screening colonoscopy. Among these biopsies, twenty-two (47.8%) were identified as displaying resistance to CCRT, while twenty-four (52.2%) exhibited sensitivity to CCRT. We conducted a comparative analysis between these two groups (resistance vs. sensitivity) to assess the differentially expressed genes associated with CCRT effectiveness. Afterward, we correlated these genes with those linked to signal transduction (GO: 0007165) and identified 5 genes: MS4A12, ITLN1, GCG, WISP3, and EPHA4, as presented in Table 1 and Fig. 1. Among these genes, MS4A12 displayed the most significant positive fold change in expression (log2 ratio = 1.3507, p = 0.0045), suggesting that it had the highest increase in expression in rectal cancer patients resistant to CCRT. Notably, the expression of MS4A12 is confined to the colon [8] and colorectal cancer [9]. This observation prompted us to proceed with clinical validation of MS4A12 expression in our thoroughly characterized rectal cancer cohort.

Table 1.

Summary of differentially expressed genes associated with signal transduction (GO: 0007165) in relation to response to CCRT in rectal carcinoma

Probe	Comparison Log Ratio	Comparison p-Value	Gene Symbol	Gene name	Biological process	Molecular function
220834_at	1.3507	0.0045	MS4A12	Membrane-spanning 4-domains; subfamily A; member 12	Signal transduction	Receptor activity
223597_at	1.3196	0.0094	ITLN1	Intelectin 1 (galactofuranose binding)	Signal transduction	GPI anchor binding, receptor binding, sugar binding
206422_at	1.2432	0.0065	GCG	Glucagon	G-protein coupled receptor protein signaling pathway, cell proliferation, feeding behavior, signal transduction	Hormone activity, receptor binding
210861_s_at	1.1635	< 0.0001	WISP3	WNT1 inducible signaling pathway protein 3	Cell–cell signaling, regulation of cell growth, signal transduction	Growth factor activity, insulin-like growth factor binding
227449_at	1.1273	< 0.0001	EPHA4	EPH receptor A4	Adult walking behavior, axon guidance, protein amino acid phosphorylation, signal transduction, transmembrane receptor protein tyrosine kinase signaling pathway	ATP binding, ephrin receptor activity, kinase activity, nucleotide binding, protein kinase activity, protein-tyrosine kinase activity, receptor activity, transferase activity, transmembrane receptor protein tyrosine kinase activity

Fig. 1.

Assessment of the differential expression profile of genes related to signal transduction and their association with concurrent chemoradiotherapy (CCRT) efficacy in rectal cancer. In this heatmap, Cluster 1 and Cluster 2 represent two groups of samples identified through unsupervised hierarchical clustering based on the expression profiles of the genes listed at the right. Each column corresponds to a sample identifier (GEO Sample, GSM), annotated by treatment response status: yellow indicates responders, and blue indicates non-responders. Cluster 1 (red) contains a higher proportion of non-responders (blue), while Cluster 2 (purple) includes more responders (yellow). In this presentation, genes with an average expression level are shown in black, those with higher expression levels are displayed in red, and those with lower expression levels are displayed in green. Notably, we identified membrane-spanning 4-domains A12 (MS4A12) as the gene with the highest increase in expression, particularly concerning signal transduction (GO: 0007165), within the group of CCRT non-responders

High immunoexpression of MS4A12 is correlated with an aggressive phenotype in rectal cancer

We utilized formalin-fixed paraffin-embedded (FFPE) human rectal cancer tissues obtained from our biobank to conduct immunohistochemical (IHC) staining and establish a correlation between MS4A12 immunostaining and clinicopathological features. The results of the IHC assay demonstrated that MS4A12 was more abundantly expressed in rectal cancer tissues resistant to CCRT compared to those sensitive to CCRT (Fig. 2A–C). To illustrate the heterogeneity of MS4A12 expression across histological subtypes, additional representative images, including mucinous adenocarcinoma, adenocarcinoma with lymphoepithelioma-like carcinoma features, and conventional adenocarcinoma, are provided in Supplementary Fig. 1. These images demonstrate variable staining intensities and patterns in different tumor architectures, supporting the robustness and distribution of MS4A12 expression in the tumor microenvironment. Moreover, the results of the IHC assay indicated that elevated MS4A12 immunoexpression was substantially associated with advanced tumor status after CCRT (p < 0.001; q = 0.002) as well as positive node status both before and after CCRT (p = 0.01; q = 0.015 and p = 0.004; q = 0.009) (Table 2). Additionally, high MS4A12 immunoexpression correlated significantly with the presence of perineural and vascular invasion (p = 0.006; q = 0.011 and p = 0.001; q = 0.003) (Table 2). Furthermore, high MS4A12 immunoexpression was notably linked with a lack of response or a low response to CCRT (p < 0.001; q = 0.002) (Table 2). Among the 73 surgical specimens exhibiting low-grade tumor regression (grade 0–1), 52 (71.2%) displayed high MS4A12 immunoexpression.

Fig. 2.

MS4A12 immunostaining patterns. IHC staining was conducted using an anti-MS4A12 antibody. A The non-tumor mucosa displayed no MS4A12 immunoexpression. In rectal cancer surgery samples, B CCRT responders exhibited low MS4A12 immunoexpression, while C CCRT non-responders showed high MS4A12 immunoexpression. Scale bar = 100 µm

Table 2.

Associations and comparisons between MS4A12 expression and clinicopathological factors in 343 rectal cancer patients receiving neoadjuvant CCRT

Parameter		No. of case	MS4A12 Expression		p-Value	q-Value
			Low Exp	High Exp
Gender	Male	223(65%)	114(51.1%)	109(48.9%)	0.522	0.528
	Female	120(35%)	57(47.5%)	63(52.5%)
Age	< 70	217(63.3%)	111(51.2%)	106(48.8%)	0.528	0.528
	≧70	126(36.7%)	60(47.6%)	66(52.4%)
Pre-Tx tumor status (Pre-T)	cT1–T2	133(38.8%)	73(54.9%)	60(45.1%)	0.138	0.177
	cT3–T4	210(61.2%)	98(46.7%)	112(53.3%)
Pre-Tx nodal status (Pre-N)	cN0	199(58%)	111(55.8%)	88(44.2%)	0.010*	0.015*
	cN1–N2	144(42%)	60(41.7%)	84(58.3%)
Post-Tx tumor status (Post-T)	ypT1–T2	155(45.2%)	96(61.9%)	59(38.1%)	< 0.001*	0.002*
	ypT3–T4	188(54.8%)	75(39.9%)	113(60.1%)
Post-Tx nodal status (Post-N)	ypN0	225(65.6%)	125(55.6%)	100(44.4%)	0.004*	0.009*
	ypN1–N2	118(34.4%)	46(39%)	72(61%)
Vascular invasion	Absent	305(88.9%)	162(53.1%)	143(46.9%)	0.001*	0.003*
	Present	38(11.1%)	9(23.7%)	29(76.3%)
Perineural invasion	Absent	305(88.9%)	160(52.5%)	145(47.5%)	0.006*	0.011*
	Present	38(11.1%)	11(28.9%)	27(71.1%)
Tumor regression grade	Grade 0–1	73(21.3%)	21(28.8%)	52(71.2%)	< 0.001*	0.002*
	Grade 2–3	248(72.3%)	135(54.4%)	113(45.6%)
	Grade 4	22(6.4%)	15(68.2%)	7(31.8%)

Tx, treatment; Exp., expression; c, clinical; p, pathological; y, postneoadjuvant chemoradiation therapy; ^*, statistically significant

High MS4A12 immunoexpression has a negative prognostic impact on rectal cancer

We analyzed three time-to-event endpoints: disease-specific survival (DSS), local recurrence-free survival (LRFS), and metastasis-free survival (MeFS). These endpoints are defined as the period from the surgical date to the happening of cancer-related death, first local recurrence, or first metastasis. As detailed in Table 3, rectal cancer was the cause of death in 63 patients (18.4%), and recurrence in the local area and distant metastasis were observed in 35 (10.2%) and 62 (18.1%) patients, correspondingly. We conducted both univariate and multivariate analyses to evaluate the prognostic significance of MS4A12 immunoexpression and clinicopathological features. The results of the univariate analysis demonstrated that advanced tumor status and positive node status after CCRT, vascular invasion, low-grade tumor regression, and high MS4A12 immunoexpression (Fig. 3A–C) were substantially associated with an unfavorable prognosis for all three endpoints (all p ≤ 0.0434) (Table 3). In addition, advanced tumor status before CCRT and perineural invasion were remarkably correlated only with poor DSS (p = 0.0074 and p = 0.018) (Table 3). Furthermore, in the multivariate analysis, the results revealed that low-grade tumor regression and high MS4A12 immunoexpression remained independent adverse prognicators significantly associated with all three endpoints (all p ≤ 0.004) (Table 4). Nearly all patients experienced events within a 60-month timeframe. Among the patients who did not experience events, the average duration of follow-up was 66.4 months (with a median of 59.2 months, ranging from 10.3 to 131.3 months).

Table 3.

Univariate log-rank analysis for notable clinicopathological variables and MS4A12 expression

Parameter		No. of case	DSS		LRFS		MeFS
			No. of event	p-Value	No. of event	p-Value	No. of event	p-Value
Gender	Male	223	42	0.9755	9	0.2176	25	0.3460
	Female	120	21		26		37
Age	< 70	217	34	0.0728	23	0.9398	38	0.5771
	≧70	126	29		12		24
Pre-Tx tumor status (Pre-T)	cT1–T2	133	15	0.0074*	13	0.6695	18	0.0663
	cT3–T4	210	48		22		44
Pre-Tx nodal status (Pre-N)	cN0	199	31	0.0932	19	0.4755	30	0.0697
	cN1–N2	144	32		16		32
Post-Tx tumor status (Post-T)	ypT1–T2	155	13	< 0.0001*	8	0.0020*	15	< 0.0001*
	ypT3–T4	188	50		27		47
Post-Tx nodal status (Post-N)	ypN0	225	31	0.0022*	18	0.0434*	32	0.0077*
	ypN1–N2	118	32		17		30
Vascular invasion	Absent	305	50	0.0023*	28	0.0379*	50	0.0136*
	Present	38	13		7		12
Perineural invasion	Absent	305	51	0.0180*	33	0.4263	53	0.1782
	Present	38	12		2		9
Tumor regression grade	Grade 0–1	73	28	< 0.0001*	14	0.0017*	25	< 0.0001*
	Grade 2–3	248	34		21		36
	Grade 4	22	1		0		1
MS4A12 expression	Low Exp	171	14	< 0.0001*	8	0.0004*	15	< 0.0001*
	High Exp	172	49		27		47

c, clinical; p, pathological; y, postneoadjuvant chemoradiation therapy; Exp., expression; DSS, disease-specific survival; LRFS, local recurrence-free survival; MeFS, metastasis-free survival; ^*, statistically significant.

Fig. 3.

Kaplan–Meier survival analysis. We applied the Kaplan–Meier method along with a log-rank test, and the outcomes revealed a significant correlation between elevated MS4A12 immunoexpression and poorer A disease-specific survival, B local recurrence-free survival, and C metastasis-free survival. Exp., expression

Table 4.

Multivariate analysis

Parameter	DSS			LRFS			MeFS
	H.R	95% CI	p-Value	H.R	95% CI	p-Value	H.R	95% CI	p-Value
MS4A12 high expression	2.507	2.978–7.978	0.003*	2.577	1.394–4.761	0.003*	2.556	4.250–21.929	0.002*
Tumor regression grade	2.611	1.577–4.310	< 0.001*	2.463	1.506–4.032	< 0.001*	2.079	1.261–3.445	0.004*
Post-Tx tumor status (Post-T)	1.937	1.076–2.777	0.050*	2.173	1.125–4.198	0.021*	1.812	0.953–3.445	0.070
Pre-Tx tumor status (Pre-T)	1.775	0.887–2.104	0.057	–	–	–	–	–	–
Post-Tx nodal status (Post-N)	1.108	0.669–1.589	0.710	1.100	0.637–1.901	0.732	1.144	0.656–1.996	0.635
Vascular invasion	1.596	0.805–2.3300	0.158	1.585	0.825–3.046	0.166	1.432	0.727–2.820	0.299
Perineural invasion	1.321	0.611–1.807	0.380	1.283	0.666–2.470	0.457	0.940	0.454–1.949	0.868

DSS, disease-specific survival; LRFS, local recurrence-free survival; MeFS, metastasis-free survival; ^*, statistically significant.

Functional prediction of MS4A12 from gene coexpression and the TME correlations

To gain initial insights into the unexplored roles of MS4A12 in rectal cancer, we analyzed its mRNA expression alongside coexpressed genes from a dataset of 788 rectal cancer samples in the Memorial Sloan Kettering (MSK) database. We identified the top 193 positively and 73 negatively correlated genes (Supplementary Tables 1–2), ranked by correlation coefficient and filtered using a q-value < 0.01, and conducted functional annotation using the Gene Ontology (GO) resource. In the context of biological processes, we observed that oligosaccharide biosynthetic process (GO: 0009312) was the most noteworthy GO term representing MS4A12 upregulation, featuring a fold enrichment of 17.23 (false discovery rate, FDR: 4.44 × 10⁻²) (Fig. 4A). Notably, we identified that the beta-1,3-galactosyltransferase 1 (B3GALT1), B3GALT5, beta-1,4-N-acetyl-galactosaminyltransferase 2 (B4GALNT2), ST6 N-acetylgalactosaminide alpha-2,6-sialyltransferase 1 (ST6GALNAC1), and ST6GALNAC6 genes were involved in the function mentioned above. In terms of molecular functions, no statistically significant results were found. Furthermore, in relation to cellular components, brush border membrane (GO: 0031526) was identified as the most predominant GO term presenting MS4A12 upregulation, with a fold enrichment of 13.33 (FDR: 5.47 × 10⁻⁵) (Fig. 4B). In addition to the tumor cell-autonomous signaling, the non-cell-autonomous signaling such as the tumor microenvironment (TME) is significantly implicated in CRC development and therapeutic response, either by promoting or inhibiting the process at different stages [16]. To explore the crosstalk between MS4A12 expression and the TME in rectal cancer, the Tumor IMmune Estimation Resource version 2.0 (TIMER2.0) database was used. Notably, unlike other cancer types, rectal cancer (READ, n = 166) uniquely shows consistent results across most deconvolution algorithms, where higher MS4A12 expression is significantly associated with greater B-cell infiltration (Fig. 5).

Fig. 4.

Enrichment of GO terms related to the upregulation of MS4A12. To identify the genes coexpressed with MS4A12 in rectal cancer, we utilized the Memorial Sloan Kettering (MSK) database (n = 788) and selected the top 193 transcripts whose expressions exhibited a positive correlation with MS4A12 level to conduct functional profiling using the Gene Ontology (GO) database. This profiling was based on A biological processes and B cellular components, and the results were ranked according to fold enrichment. To visualize and present the significant GO terms, we employed an R script along with the ggplot2 package. All the terms shown in the plots meet the significance threshold of FDR < 0.05

Fig. 5.

High MS4A12 expression is significantly correlated with abundant infiltration of B cells in rectal cancer. These data were obtained from the Tumor IMmune Estimation Resource version 2.0 (TIMER2.0) database. Partial_Cor refers to partial correlation, which quantifies the association between gene expression and immune cell infiltration while controlling for tumor purity. READ: rectum adenocarcinoma. (n = 166)

Discussion

In normal colonic epithelial cells, MS4A12 expression is tightly controlled through transcriptional repression. In colon cancer cells, the MS4A12 promoter is selectively activated by the transcription factor CDX2, initially known for its role in intestinal epithelium development and differentiation [10, 11]. The finding that MS4A12 expression is exclusively limited to colonic epithelial cells aligns with the pattern observed in other CDX2 target genes, which also exhibit restricted expression within the colon. Moreover, CDX2 was characterized as a lineage-survival oncogene deregulated in CRC [17], and MS4A12 was also found to be implicated in colon tumorigenesis [8]. These findings suggest that CDX2 may contribute to rectal cancer progression through MS4A12-mediated pathways. Nevertheless, these findings challenge the observations that low MS4A12 expression predicted poor survival in patients with early-stage colon cancer [13] and that lack of CDX2 expression was correlated with more advanced tumor stage and unfavorable survival in colon cancer [18]. Collectively, these observations suggest the complex multigene signatures and multistep processes of CRC development. Notably, it has been indicated that patients with CDX2-negative colon cancer might benefit from adjuvant chemotherapy, particularly in a stage II setting [18]. In the present study, we are the first to demonstrate that high MS4A12 expression is significantly correlated with advanced tumor stage, poor response to neoadjuvant CCRT, and inferior patient survival in rectal cancer, providing a helpful predictive and prognostic marker for identifying patients who could gain advantages from neoadjuvant CCRT.

The intestinal tract is crucial for digestion, and disruptions to the intestinal barrier can result in a range of illnesses. Functional oligosaccharides, characterized by their low degree of polymerization, have been suggested to repair the intestinal tract and maintain intestinal homeostasis by regulating the barrier, immune response, and microbiome [19]. Thus, they held promise as a therapeutic option for treating intestinal conditions, such as inflammatory bowel disease (IBD) [20] and CRC [21]. In contrast, it has been reported that aberrant glycosylation was correlated with drug resistance and poor prognosis in CRC [22]. Glycosylation refers to a set of post-translational modifications (PTMs) in which carbohydrate moieties are enzymatically (by glycosyltransferases) attached to proteins, lipids, or other molecules. The processes of glycosylation include the initiation and core extension steps (pathway-specific) and the elongation and capping steps (pathway-nonspecific) [23]. Intriguingly, several glycosyltransferase-related genes co-upregulated with MS4A12 were implicated in the processes of glycosylation (Fig. 4A), and these genes comprise beta-1,3-galactosyltransferase 1 (B3GALT1), B3GALT5, beta-1,4-N-acetyl-galactosaminyltransferase 2 (B4GALNT2), ST6 N-acetylgalactosaminide alpha-2,6-sialyltransferase 1 (ST6GALNAC1), and ST6GALNAC6.

Cancer frequently presents oncofetal features that manifest in the composition of their glycoconjugates. These alterations in glycosylation promote metastatic characteristics, hinder apoptosis, and result in resistance to chemotherapy [24]. The B3GALT family mediates the transfer of galactose (Gal) to N-acetylglucosamine (GlcNAc) residues and is responsible for type I chain synthesis. High B3GALT1 expression has been associated with high metastatic potential in hepatocarcinoma cells [25] and doxorubicin resistance in breast cancer cells [26]. Otherwise, high expression of B3GALT5 has been correlated with poorer relapse-free survival and overall survival in hepatocellular carcinoma [27] and breast cancer [28]. B4GALNT2 adds N-acetylgalactosamine (GalNAc) sugars and is responsible for the last step of the Sda/Cad blood antigen synthesis. B4GALNT2 is highly expressed in triple-negative breast cancer cells and its interaction with major histocompatibility complex, class I, B (HLA-B) might enhance cell proliferation and metastasis [29], whereas high B4GALNT2 expression has been linked to good prognosis in colon cancer through database analysis [30]. The ST6GALNAC family catalyzes the addition of sialic acids to GalNAc residues in O-glycosylated proteins or glycolipids. ST6GALNAC1 facilitates the development of the tumor-associated sialyl-Tn (sTn) antigen which is overexpressed in many tumor types and is linked to poor prognosis [31]. ST6GALNAC1 may enhance cancer cell stemness via the Akt pathway in colorectal [32] and ovarian [33] cancer cells. ST6GALNAC1-positive CRC has also been associated with 5-FU resistance and poor prognosis [32]. Moreover, ST6GALNAC1 may confer cisplatin and 5-FU resistance in gastric cancer cells [34]. On the other hand, ST6GALNAC6 exhibits a preference for adding sialic acids to glycolipids rather than glycoproteins, and its higher expression has been observed in stromal cells that suppress T-cell activation in CRC [35]. Additionally, ST6GALNAC6 may also enhance the migration of renal cell carcinoma cells and inhibit NK-cell cytotoxicity [36]. Nevertheless, the correlations between the expressions of MS4A12 and these glycosyltransferases and their involvement in CCRT resistance and rectal carcinogenesis deserve further investigation.

Accumulating evidence has indicated that the tumor microenvironment (TME), comprising tumor cells and their accompanying immune and stromal cells, significantly influences tumor behavior and the response to immunotherapy [37]. Since we found that many of the above-mentioned genes were involved in the TME favoring tumor progression, we were interested in further exploration and discussion. MS4A is a family of structurally related cell surface proteins, and one of the prominent sequence homologs of MS4A12 is MS4A1 (also known as CD20) [38]. CD20 is a B cell-specific membrane protein (except early pro-B cells and plasma cells) and represents an attractive target for B-cell lymphoma, while MS4A12 is hardly expressed on B cells. Interestingly, we found that MS4A12 expression was significantly positively associated with B-cell infiltration level in rectal cancer (n = 166) through TIMER2.0 analysis (Fig. 5). However, whether and how the involvement of MS4A12 in B-cell recruitment needs further examination. Moreover, a recent study has revealed that CD20⁺ B cells within tertiary lymphoid structures (TLSs), featured by the accumulation of IgG plasma cells, strong antigen presentation, high tumor mutation burden (TMB-H), high microsatellite instability (MSI-H), and immunosuppressive characteristics, were enriched in CRC [39]. These features may present a hot tumor immune environment and become a promising predictive biomarker for immunotherapy. Accordingly, an investigation is warranted to check whether rectal cancer patients with high MS4A12 expression could obtain significant therapeutic advantages in an immunotherapy cohort. As for CCRT efficacy, it has been indicated that B cells may contribute to CCRT resistance in CRC by promoting an immunosuppressive tumor microenvironment [39]. Regulatory B cells (Bregs) can secrete interleukin-10 (IL-10) and transforming growth factor beta (TGF-β) [40], suppressing cytotoxic T cell activity, while B cells expressing programmed death-ligand 1 (PD-L1) [39] may induce T cell exhaustion, impairing effective tumor clearance. Consequently, the observed association among high MS4A12 expression, increased B-cell infiltration, and poor response to CCRT may reflect a shift toward a suppressive immune landscape. However, the underlying mechanism remains unclear and requires further validation.

CDX2 and MUC2 are differentiation-associated markers, and their low expression has been linked to poor prognosis in treatment-naïve colorectal cancer [18, 41]. Low MS4A12 expression has also been associated with differentiation inhibition [13] and worse survival in colorectal cancer patients naïve to treatment [9]. However, the absence of CDX2 may benefit from adjuvant chemotherapy in colon cancer [18], and our previous work and current study also suggest that low MUC2 [12] and MS4A12 expressions, respectively, are associated with better response to CCRT in rectal cancer. This indicates that reduced expression of differentiation-related genes may reflect dedifferentiation and aggressiveness, but may also lead to increased treatment sensitivity, possibly due to reduced epithelial integrity. CDX2 is a transcription factor and not easily targetable, with unclear mechanisms underlying CCRT resistance. While MUC2 may contribute to therapy resistance by forming a physical barrier to drug penetration or immune cell access [12], its broad expression profile poses a risk of off-target toxicity. In contrast, high MS4A12 expression in this study was associated with CCRT resistance and poor prognosis, potentially through modulation of B cell-mediated immune suppression beyond differentiation status. Although both MUC2 and MS4A12 are CDX2 downstream targets, MS4A12 may provide additional prognostic value and mechanistic insight, warranting further investigation as a specific therapeutic target.

Our ongoing research is subject to certain limitations. Firstly, given that protein expression levels vary continuously with cancer progression, it becomes imperative to categorize protein expression into multiple levels to assess their respective impacts on survival. This process may delineate the optimal threshold of high MS4A12 expression for clinical assessment. Secondly, our analysis of tumor samples was retrospective and confined to a single institution; therefore, the clinical utility of MS4A12 expression warrants validation through prospective, large-scale, and multicenter investigations.

Conclusion

Unlike other MS4A proteins, typically limited to hematopoietic or lymphatic cell lineages, MS4A12 expression is specifically localized to the apical area of normal colonic epithelial cells [8]. In this study, we found that high MS4A12 expression was considerably linked to an aggressive phenotype, low or no response to CCRT, and inferior clinical outcomes in patients with rectal cancer. Clinically, MS4A12 expression may function as an early indicator of eventual clinical outcomes and assist in making therapeutic choices by helping identify individuals who are probable responders to particular treatments.

Author contributions

Conceptualization: H.-Y. Lai; methodology: C.-I. Chen, Y.-K. Kao, P.-W. Yang, P.-C. Chen, C.-C. Yang, W.-S. Li, H.-H. Tsai, and Y.-J. Wang; investigation: C.-I. Chen, Y.-K. Kao, P.-W. Yang, P.-C. Chen, C.-C. Yang, W.-S. Li, H.-H. Tsai, and Y.-J. Wang; formal analysis: C.-I. Chen, Y.-K. Kao, P.-W. Yang, P.-C. Chen, C.-C. Yang, W.-S. Li, H.-H. Tsai, and Y.-J. Wang; resources: H.-H. Tsai and Y.-J. Wang; validation: C.-I. Chen, Y.-K. Kao, P.-W. Yang, P.-C. Chen, C.-C. Yang, and W.-S. Li; visualization: C.-I. Chen, Y.-K. Kao, P.-W. Yang, P.-C. Chen, C.-C. Yang, and W.-S. Li; writing—original draft: H.-Y. Lai; writing—review & editing: H.-Y. Lai; funding acquisition: H.-Y. Lai; supervision: H.-Y. Lai. All the authors contributed to the article and approved the submitted version.

Funding

This work was financially supported by the China Medical University grants CMU113-MF-85 and CMU113-N-11, the China Medical University Hospital grant DMR-113–007, and the National Science and Technology Council grant NSTC113-2314-B-039–028-.

Availability of data and materials

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

As a rule, every participant signed informed consent before samples were deposited in the biobank. This study and its use of deidentified tumor samples from the biobank were approved by the Ethics Committee and Institutional Review Board of Chi Mei Medical Center (10302014). This study was conducted following the ethical guidelines of the Declaration of Helsinki and the regulations of our government.

Consent for publication

Not applicable.

Conflict of interest

The authors declare no competing interests.