Bioinformatics analysis of proteins in the complement and coagulation cascades in colon cancer: Discovering the potential biomarker SERPINA1

Introduction

Colon cancer represents the second leading cause of global cancer-related mortality, imposing a notable economic and healthcare burden (1). In China, it ranks among the top five most prevalent types of cancer, alongside lung, liver, thyroid and gastric cancers (2,3). Worldwide, the incidence rate increased from 3.96 (95% UI: 3.69–4.21) per 100,000 individuals in 1990 to 5.37 (95% UI: 4.91–5.86) per 100,000 in 2021, while the mortality rate decreased from 2.19 (95% UI: 2.01–2.36) per 100,000 in 1990 to 2.01 (95% UI: 1.84–2.19) per 100,000 in 2021 (4), affecting not only the elderly but also adolescents and young adults (5,6). The absence of early symptoms and specific biomarkers often leads to diagnosis at advanced stages (7,8). Despite advances in treatment modalities such as surgery and targeted therapies, the prognosis for colorectal cancer (CRC) remains unfavorable, with a 5-year mortality rate of ~12.8% (9). Metastatic disease confers a worse prognosis, with 5-year survival rates falling <50% even with surgical intervention (10). The pathogenesis of colon cancer is multifactorial, involving genetic predisposition, environmental factors and chronic inflammatory conditions (11). Emerging evidence further indicates that immune dysregulation plays a key role in driving carcinogenesis (12,13).

Dysregulation of the complement and coagulation cascade (CCC) pathway has been implicated in various types of cancer (14–17). Activation of this pathway promotes a procoagulant state, which can compromise vascular integrity and enhance leukocyte infiltration (18). Furthermore, specific CCC genes contribute to complement-mediated endothelial damage (19), a process involved in tumorigenesis. Despite its established relevance in pre-cancers, the role of the CCC pathway in colon cancer remains insufficiently explored (17).

To systematically investigate the role of CCC pathway proteins in colon cancer, a set of 88 genes associated with this pathway was first retrieved from the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Their expression patterns were subsequently analyzed using the GEPIA2 cancer database through the ‘Expression DIY’ and ‘Survival Analysis’ modules. This approach led to the identification of four key genes: Complement C3 (C3), factor VIII (F8), serpin family A member 1 (SERPINA1) and SERPING1, which are implicated in colon cancer progression. Among these, SERPINA1 was chosen for subsequent validation. The expression and clinical relevance of SERPINA1 were further assessed through a meta-analysis and immunohistochemistry (IHC) performed on tissue microarrays. Additionally, the relationship between SERPINA1 expression levels and overall patient survival was evaluated using available microarray datasets.

Materials and methods

Ethics approval

The present study received approval from Shanghai Biotechnology Co., Ltd. (approval no. SHYJS-CP-1707004) and adhered to the ethical standards outlined in the 2013 Declaration of Helsinki. Written informed consent was obtained from every participant included in the present study. From July 2006 to March 2007, tumor samples, along with matched normal colon tissue from adjacent sites, were collected from residual clinical material remaining after routine pathological assessment. Relevant clinical and pathological parameters, including demographic details, were retrospectively gathered from the institution's electronic health record system.

Screening of differentially expressed genes (DEGs) within the CCC pathway

To identify DEGs related to the CCC pathway in colon cancer, a two-step screening approach was employed. First, the gene list defining the CCC pathway was obtained from the KEGG database (http://www.genome.jp/kegg/mapper.html). Subsequently, the gene expression data for colon adenocarcinoma (COAD) were acquired from the GEPIA2 portal (http://gepia2.cancer-pku.cn/) to analyze gene expression differences. DEGs were identified using thresholds of absolute log2 fold change (|log2FC|)>1 and an adjusted q-value<0.01. Finally, the overlap between these COAD-associated DEGs and the 88 CCC pathway genes was determined using R software (version 4.4.2; Posit Software, PBC).

Bioinformatics analysis for functional enrichment and hub gene identification

To investigate the biological relevance of the identified CCC-related DEGs, a protein-protein interaction (PPI) network was constructed using the Search Tool for the Retrieval of Interacting Genes and proteins (STRING) database (version 12.0; http://www.string-db.org/). Interaction confidence was set at a minimum score of 0.7 and statistical significance was defined by a false discovery rate <0.05. The resulting network was visualized and analyzed in Cytoscape (v3.10.0) (20). Key functional submodules within the PPI network were detected using the Molecular Complex Detection (MCODE) plugin (21), with parameters set as follows: Degree cutoff=2, node score cutoff=0.2, k-core=2, maximum depth=100 and a minimum module size of 4 genes. Hub genes were subsequently identified via the cytoHubba plugin (22), applying the maximal clique centrality algorithm to rank nodes by their topological importance within the network.

Analysis of the relationship between gene expression and patient survival

The association between gene expression and patient survival was analyzed using the GEPIA2 online database (http://gepia2.cancer-pku.cn/#survival). Overall survival data for all 32 DEGs in COAD were assessed using both median (50% cutoff) or quartile (75% cutoff-High) options, with a 95% confidence interval (CI) and the hazard ratio (HR) applied. Following plot generation, the log-rank P-value and HR (high) were extracted from the image. After comparing the median and quartile group cutoffs, the plots showing genes with lower log-rank P-values were selected for further study. Additionally, the relationship between four specific genes (SERPINA1, C3, F8 and SERPING1) and patient survival was examined across 31 other tumor types available in the GEPIA2 database.

Potential drug target analyses

The therapeutic potential of SERPINA1, C3, F8 and SERPING1 was evaluated for druggability using the Open Targets Platform (23), which provided data on associated diseases and known targeting compounds. Further investigation into SERPINA1 was conducted via the Cancer Therapeutics Response Portal (CTRP) (https://portals.broadinstitute.org), where candidate compounds were prioritized based on the correlation between their sensitivity profiles and SERPINA1 expression levels, applying an interquartile multiplier cutoff.

Meta-analysis

SERPINA1 expression data were obtained from articles indexed in PubMed (https://pubmed.ncbi.nlm.nih.gov/) using the keywords ‘SERPINA1’ and ‘cancer’. The inclusion criteria were as follows: i) Research conducted on adult human samples; ii) comparison between cancer patients and healthy controls; and iii) primary data clearly presenting differential groups, with DEGs that could be extracted or recalculated. Studies focusing solely on the effects of specific treatments in cancer patients or only including samples from treated patients were excluded. Additionally, preprints, reanalyses of data from databases, and reviews were excluded. Retrieved articles were managed using Endnote (version 21). Two reviewers (DL and DXH) independently extracted the data following PRISMA guidelines (24), and discrepancies were resolved through discussion. For the meta-analysis of SERPINA1 expression data, a random-effects model was implemented in Stata 14.0 (StataCorp LP). The key metrics extracted per study were the log2FC, reflecting the cancer vs. control expression difference and the negative log P-value (−logP) as a measure of statistical strength. The overall effect size was calculated as a weighted average of the log2FC values, with the weight of each study determined by the-logP-value to emphasize more reliable findings. The random-effects model was used for weight assessment. The 95% CIs for the pooled estimate were derived from this weighting scheme. Results were visualized as forest plots and the between-study heterogeneity was evaluated using the I2 statistic. Significance was established if the 95% CI of the combined estimate did not include zero.

IHC validation of SERPINA1 expression using tissue microarray

IHC analysis was performed using a colon cancer tissue microarray (cat. no. HCo1A180Su16; Shanghai Outdo Biotech Co., Ltd.), which included 180 samples [tumor (n=104) and adjacent normal tissues (n=76)] from 104 patients: 6 stage I–II, 1 stage I–III, 51 stage II, 33 stage II–III and 13 stage III (according to the 7th edition of the AJCC TNM staging system (25). The patient cohort consisted of 59 men and 45 women, with a mean age of 68.2±10.8 years (range, 24–90 years). Samples had been collected between July 2006 and May 2007. By July 2015, 41 patients remained alive, while 63 were deceased. Tissue sections were incubated with an undiluted rabbit anti-α-1 antitrypsin (AAT; SERPINA1) primary antibody (cat. no. ZA-0007; Beijing Zhongshan Jinqiao Biotechnology Co., Ltd.) followed by goat anti-rabbit IgG H&L (HRP) diluted 1:1,000 (cat. no. ab6721; Abcam). The sections were then stained by diaminobenzidine (DAB). Staining and slide imaging were automated using the BOND RX Research Stainer (Leica Biosystems) and the TissueFAXS 200 system (TissueFAXS Viewer software v6.0.6245.146; TissueGnostics GmbH), respectively. SERPINA1 expression was quantified via optical density measurement with HistoQuest software (v6.0.1.114; TissueGnostics GmbH). Prior to quantification, the analysis pipeline was optimized by refining three key parameters: Total tissue area segmentation, positive cell detection and background signal correction. The final expression level was calculated as the percentage of positive cells within a total analyzed tissue volume of 3.16 mm3.

Statistical analysis

Statistical analyses and graphical presentations were performed with GraphPad Prism (v9.0.0; Dotmatics) or R Studio (v4.4.2). Values are expressed as the mean ± standard deviation. Group comparisons were made by Student's unpaired t-test. P<0.05 was considered to indicate a statistically significant difference.

Results

Identification and network analysis of complement and coagulation cascade (CCC)-related DEGs in COAD

A total of 32 genes from the 88-gene CCC pathway were identified as DEGs (|log2FC|>1 and q-value cutoff of 0.01) in COAD based on the GEPIA2 database. Among these, 13 genes were upregulated (P<0.05) and 19 were downregulated (P<0.05) (Fig. 1A and Table I), with broad representation across the intrinsic, classical, lectin and alternative complement pathways (Fig. S1). Enrichment analysis performed in the STRING database indicated that the ‘CCC pathway’ was the most significantly overrepresented KEGG term, followed by pathways linked to Staphylococcus aureus infection, pertussis and systemic lupus erythematosus (Fig. 1B). PPI interaction analysis revealed a densely interconnected network. Further clustering of this PPI network via the MCODE plugin in Cytoscape delineated two functional modules. Module 1, comprising 17 nodes and 121 edges, included complement C1s, SERPING1 and C3. Module 2, consisting of 6 nodes and 12 edges, encompassed the hub genes thrombomodulin and SERPINA1 (Fig. 1C and D).

Figure 1. PPI and functional analysis of differentially expressed genes in the complement and coagulation cascades pathway. (A) PPI network visualized using Cytoscape, with up- and down-regulated genes indicated in red and blue, respectively. (B) KEGG pathway enrichment results from the STRING database. Top hub genes identified within (C) Module 1 and (D) Module 2, respectively, by applying the cytoHubba algorithm to the MCODE-derived network clusters. KEGG, Kyoto Encyclopedia of Genes and Genomes; FDR, false discovery rate; PPI, protein-protein interaction.

Table I. Differentially expressed genes in colon adenocarcinoma from GEPIA2 database.

Table I.

Differentially expressed genes in colon adenocarcinoma from GEPIA2 database.

Genes	Median (tumor)	Median (normal)	log2FC	P-value
C7	1.67	111.39	−5.40	6.80×10-112
MASP1	0.30	23.52	−4.24	7.16×10-73
CFD	9.65	162.01	−3.94	6.45×10-97
CLU	56.28	744.12	−3.70	3.50×10-65
F13A1	1.65	17.67	−2.82	2.72×10-69
VSIG4	4.15	22.13	−2.17	2.24×10-41
SERPING1	67.62	261.89	−1.94	2.52×10-36
C1S	70.60	241.45	−1.76	5.43×10-35
C3	68.83	234.82	−1.76	4.95×10-21
C4A	18.13	59.54	−1.66	2.46×10-26
C4B	18.15	59.09	−1.65	1.28×10-25
CFH	6.62	19.90	−1.46	1.91×10-35
SERPINA5	0.34	2.64	−1.44	8.93×10-60
MASP2	0.24	1.92	−1.24	3.64×10-97
A2M	58.60	139.21	−1.23	1.56×10-28
C1R	77.29	176.88	−1.18	5.10×10-23
THBD	3.53	9.01	−1.14	1.64×10-27
F8	1.72	4.88	−1.11	3.70×10-25
VWF	16.77	36.48	−1.08	2.30×10-23
C4BPA	1.78	0.18	1.24	5.55×10-27
CD46	174.82	71.39	1.28	2.42×10-78
F2R	12.27	4.27	1.33	1.45×10-47
PROCR	35.05	12.79	1.39	2.04×10-35
CD55	65.26	22.58	1.49	9.23×10-42
CFB	69.48	21.30	1.66	3.87×10-45
C4BPB	7.37	1.57	1.70	7.87×10-27
PLAU	29.05	6.56	1.99	5.20×10-68
SERPINE2	47.67	10.86	2.04	5.04×10-41
PLAUR	59.00	11.74	2.24	1.58×10-66
C2	57.80	9.82	2.44	1.14×10-86
SERPINA1	120.77	17.31	2.73	6.53×10-33
F12	18.78	1.94	2.75	3.72×10-96

[i] log₂FC, log2 fold change.

Four genes related to the survival of patients with pan-cancers

Overall survival analysis using a median group cutoff demonstrated an association between four genes from the CCC pathway (SERPINA1, C3, F8 and SERPING1) and patient survival in COAD (Fig. 2). This association remained significant for SERPINA1 alone when a more stringent quartile cutoff with Cutoff-high of 75% was applied. Extending the analysis to other types of cancer revealed broader prognostic relevance. SERPINA1 expression was associated with patient survival in brain lower grade glioma (LGG), skin cutaneous melanoma (SKCM) and breast invasive carcinoma. The other three genes also showed significant associations across multiple types of cancer, with C3 linked to adrenocortical carcinoma, COAD, LGG and SKCM; F8 to COAD and kidney renal clear cell carcinoma; and SERPING1 to COAD, LGG and SKCM (Table II).

Figure 2. Association of four genes with overall survival in colon cancer. (A) Survival analysis for SERPINA1 using quartile cutoff. Survival analyses for (B) C3, (C) F8 and (D) SERPING1, respectively, using median cutoff. SERPINA1, serpin family A member 1; C3, complement C3; F8, factor VIII; HR, hazard ratio.

Table II. Survival analysis of four genes in pan-cancer panel from GEPIA2 database.

Table II.

Survival analysis of four genes in pan-cancer panel from GEPIA2 database.

	Type of cancer
Genes	COAD	LGG	SKCM	BRCA	ACC	CHOD	KICH	KRIC	MESO	PCPG	CESC	THYM	LUSC	SARC
SERPINA1
Log rank P-value	0.043	2.7×10-6	0.00026	9.3×10-12	-	-	-	-	-	-	-	-	-	-
HR (high)	0.51	2.4	0.61	0.61	-	-	-	-	-	-	-	-	-	-
C3
Log rank P-value	0.033	0.0031	0.017	-	9.5×10-5	0.033	0.032	4.1×10-5	0.0025	0.013	-	-	-	-
HR (high)	1.7	1.7	0.72	-	0.19	1.7	0.14	1.9	0.47	1.6×10-9	-	-	-	-
F8
Log rank P-value	0.015	-	-	-	0.019	-	-	0.0003	-	-	0.0026	0.036	-	-
HR (high)	1.8	-	-	-	0.39	-	-	0.57	-	-	0.49	6.9	-	-
SERPING1
Log rank P-value	0.032	3.3×10-7	0.00052	-	-	-	-	-	0.0003	-	-	-	0.037	3.9×10-5
HR (high)	1.7	2.6	0.62	-	-	-	-	-	0.42	-	-	-	1.3	0.43

[i] COAD, colon adenocarcinoma; LGG, brain lower grade glioma; SKCM, skin cutaneous melanoma; BRCA, breast invasive carcinoma; ACC, adrenocortical carcinoma; CHOL, cholangiocarcinoma; KICH, kidney chromophobe; KIRC, kidney renal clear cell carcinoma; MESO, mesothelioma; PCPG, pheochromocytoma and paraganglioma; CESC, cervical squamous cell carcinoma and endocervical adenocarcinoma; THYM, thymoma; LUSC, lung squamous cell carcinoma; SARC, sarcoma; HR, hazard ratio; SERPINA1, serpin family A member 1; C3, complement C3; F8, factor VIII.

Evaluating the druggability profile of SERPINA1, C3, F8 and SERPING1

To investigate the potential druggability of the four identified genes (SERPINA1, C3, F8 and SERPING1), established disease associations were analyzed using the Open Targets Platform (Table II). All genes displayed associations with various types of cancer, with strength of evidence varying. Gene-cancer association scores were obtained from the GeneCards Suite, which integrates multi-omics and clinical evidence to generate a normalized ALIscore (0–1) for each gene-disease pair, with higher scores indicating stronger association. For instance, SERPINA1 showed strong associations with gastric and breast cancer, whereas the link between SERPING1 and types of cancer, such as breast cancer and prostate cancer, was comparatively weak. Subsequent review of approved targeted therapies revealed that F8, C3 and SERPING1 are already targeted by existing drugs (Table III) (26–28). Notably, however, no clinically approved drugs currently target SERPINA1, suggesting its potential as a novel therapeutic candidate (https://www.genecards.org/card/SERPINA1).

Table III. Summary of disease associations and approved or investigational drugs targeting the four genes.

Table III.

Summary of disease associations and approved or investigational drugs targeting the four genes.

Genes	Type of cancer (scorea)	Known drugs	Main treatment	Earliest approved	(Refs.)
SERPINA1	Gastric cancer	No	/	/	N/A
	(0.82);
	Pancancer (0.81)
	Breast cancer
	(0.75);
	Colorectal cancer
	(0.59)
C3	Ovarian cancer	Pegcetacoplan	Paroxysmal	2021	(26)
	(0.83);		nocturnal
	Pancancer (0.81);		hemoglobinuria
	Gastric cancer		and immune
	(0.72); Colorectal		system disease
	cancer (0.36)
F8	Pancancer (0.67);	Moroctocog alfa	Hemophilia a	1999	(27)
	Esophageal
	cancer (0.39);
	Breast cancer
	(0.20); Lung
	cancer (0.14)
SERPING1	Breast cancer	Conestat alfa	Hereditary	2010	(28)
	(0.24); Pancancer		angioedema
	(0.18); Prostate
	cancer (0.15);
	Pituitary cancer
	(0.13)

a Association scores. A score of 1.0 corresponds to the highest level of relevance. N/A, not applicable; SERPINA1, serpin family A member 1; F8, factor VIII; C3, complement C3.

In the absence of approved drugs targeting SERPINA1, the CTRP database was used to identify potential compounds interacting with this gene. As presented in Fig. 3, five compounds showed a correlation with SERPINA1 expression, listed in order of decreasing correlation score: Omacetaxine mepesuccinate (0.2330), ML030 (0.2140), vincristine (0.2090), SB-743921 (0.2070) and KX2-391 (0.2020). These agents may represent potential therapeutic candidates for colon cancer via modulation of SERPINA1 activity.

Figure 3. Correlation analysis of small molecules with SERPINA1 expression using Cancer Therapeutics Response Portal data. (A) Distribution of Pearson correlation coefficients presented as a box plot (interquartile multiplier method). (B) Bar chart representation of the correlation scores visualized using GraphPad Prism 9 software. SERPINA1, serpin family A member 1.

Meta-analysis of SERPINA1 expression

Based on multiple criteria, including prognostic relevance, hub gene status and the absence of approved targeted therapies, SERPINA1 was prioritized for further investigation. Following a literature review in PubMed (https://pubmed.ncbi.nlm.nih.gov/) using the keywords ‘cancer’ and ‘SERPINA1’, four eligible studies (29–32) were included (Table SI). As shown in Fig. 4, the meta-analysis demonstrated that SERPINA1 was consistently upregulated in cancer tissues compared with healthy controls, with a pooled log2FC of 1.58 (95% CI: 0.56, 2.60). However, high heterogeneity was observed across the included studies (I2=85.6%; P<0.001). This may be attributed to the limited reference data (derived from only four articles) and inherent data heterogeneity. In the present study, the data came from two types of cancer: Colon cancer (n=3) and pancreatic ductal adenocarcinoma (n=1). In addition, different sample types were involved, including tissue (n=2), plasma exosomes (n=1) and serum (n=1).

Figure 4. Forest plot of SERPINA1 based on the Random Effects Model by meta-analysis. Individual study estimates are shown as black dots (log2FC) with horizontal error bars representing 95% ICs. The pooled effect size across studies is indicated by a diamond marker. A dashed vertical line at log2FC=0 corresponds to no differential expression. The plot also reports key summary statistics, including the ES (test of ES=0; z=3.05; P=0.002), the heterogeneity index (I2=85.6%) and the P-value for heterogeneity (P<0.001), supporting the robustness of this finding while indicating high heterogeneity. log2FC, log2 fold change; ES, effect size; CI, confidence interval; SERPINA1, serpin family A member 1.

SERPINA1 validation by IHC analysis

To validate SERPINA1 expression at the protein level, IHC analysis was performed. Although previous data from the GEPIA2 database and meta-analysis had indicated upregulation of SERPINA1 in colon cancer, direct protein validation using established methods such as IHC, western blot or ELISA remained necessary. To address this, IHC was conducted on a tissue microarray containing 180 spots of colon tumor and matched adjacent normal tissues. Anti-AAT staining demonstrated markedly stronger AAT accumulation in tumor tissues (Fig. 5A) compared with adjacent normal tissues (Fig. 5B). Quantitative optical density analysis further demonstrated that AAT was upregulated by 1.84-fold in tumors (P<0.0001; Fig. 5C). When comparing AAT expression between surviving and deceased patient groups, higher AAT levels (P<0.05) were observed in the deceased group (Fig. 5D). Consistently, survival analysis revealed significantly shorter overall survival in patients with higher SERPINA1 expression compared with those with lower expression (Fig. 5E).

Figure 5. IHC detection of SERPINA1 in colon tissues. (A) Representative IHC staining of SERPINA1 in colon tumor tissue. Lower left image depicts the whole tissue, while the boxed area is enlarged and displayed on the right. Red arrows depict positively stained cells. Scale bar, 20 µm. (B) Representative IHC staining of SERPINA1 in adjacent normal colon tissue. Lower left image shows the whole tissue, while the boxed area is enlarged and displayed on the right. Red arrows depict positively stained cells. Scale bar, 20 µm. (C) Quantitative comparison of SERPINA1 expression between tumor and adjacent normal tissues using optical density analysis. (D) Comparison of SERPINA1 expression between deceased and surviving patient groups based on optical density measurements. (E) Survival analysis comparing patients with high vs. low SERPINA1 expression. **P<0.01 and ****P<0.0001. IHC, immunohistochemistry; SERPINA1, serpin family A member 1.

Discussion

CRC is a highly prevalent malignancy of the digestive system, whose development is influenced by a combination of genetic predisposition, environmental factors, chronic inflammation and gut microbiota dysbiosis. Prior research has highlighted the involvement of the complement system, a central element of innate immunity, in CRC pathogenesis (33,34), as it contributes to host defense against pathogens and modulates intestinal inflammatory responses during cancer progression. Furthermore, complement signaling has been implicated across multiple stages of tumorigenesis, including initiation, proliferation, metastasis and response to therapy (34–36). Dysregulation of the CCC pathway has been documented in several malignancies, such as metastatic urothelial carcinoma (14), acute lymphoblastic leukemia (37), lower-grade glioma (15) and bladder cancer (38). Nevertheless, the role of the CCC pathway in CRC remains largely unexplored. In the present study, the expression profiles of 88 CCC genes in COAD were evaluated using the GEPIA2 database and 32 differentially expressed CCC genes (13 upregulated and 19 downregulated) were identified, suggesting that widespread dysregulation may be implicated in clinical outcomes for patients with CRC.

Following PPI and survival analyses, four hub genes associated with patient survival were identified: SERPINA1 (upregulated), C3 (downregulated), F8 (downregulated) and SERPING1 (downregulated). SERPINA1 is a serine protease inhibitor primarily targeting elastase, and is capable of irreversibly inhibiting trypsin, chymotrypsin and plasminogen activator. The aberrant isoform of SERPINA1 inhibits insulin-induced nitric oxide synthesis in platelets, shortens coagulation time and exhibits proteolytic activity against insulin and plasmin (39). SERPINA1 has been implicated in colon cancer (29–32) and in combination with fibrinogen demonstrates superior diagnostic efficacy compared with conventional markers such as carcinoembryonic antigen and carbohydrate antigen 19-9 (30). In CRC, SERPINA1 is highly expressed, is associated with unfavorable clinical outcomes and promotes CRC cell proliferation and migration via activation of the STAT3 pathway (31). C3 serves as a precursor for non-enzymatic constituents of the classical, alternative, lectin and granzyme K complement pathways. These pathways comprise a proteolytic cascade that drives pathogen phagocytosis and degradation while enhancing adaptive immune signaling. C3 deficiency exacerbates inflammatory responses in the colon (40). Although downregulated in colon cancer, elevated C3 expression is associated with worse overall survival in gastric cancer (34). F8, in the presence of calcium and phospholipids, functions as a cofactor for factor IXa during the conversion of factor X to its active form, factor Xa. Deficiency in coagulation factor VIII underlies hemophilia A, for which recombinant or plasma-derived factor VIII remains first-line therapy (41). SERPING1, another serine protease inhibitor, regulates the classical complement pathway. SERPING1 is downregulated in prostate cancer and reduced SERPING1 expression is associated with higher Gleason scores, advanced pathological grade and more progressive tumor stages (42). In summary, SERPINA1, C3, F8 and SERPING1 are four cancer-related genes likely to play notable roles in colon cancer, despite the limited number of studies specifically focused on their functions in this malignancy.

The druggability of SERPINA1, C3, F8 and SERPING1 was further evaluated based on reports from the Open Targets Platform (23). Among these, only SERPINA1 has not yet been established as a known drug target, to the best of our knowledge. However, based on predictive screening using the CTRP database, omacetaxine mepesuccinate was identified as a potential compound targeting SERPINA1. This agent is an approved anticancer drug currently used in the treatment of chronic myeloid leukemia (43,44). Taken together, these findings suggest that SERPINA1, C3, F8 and SERPING1 may represent promising candidate targets for therapeutic intervention.

SERPINA1 was selected for meta-analysis and IHC validation using tissue microarrays. While prior studies have examined the expression and function of SERPINA1 in colon cancer (30–32), the present approach distinguishes itself from these earlier works by employing different specimens and different validation methods. For instance, Li et al (30) and Peltier et al (32) employed plasma samples from patients with CRC and validated SERPINA1 via ELISA, whereas Ma et al (31) utilized a mouse CRC model. The present study, based on tissue microarray analysis, thus serves as a methodological complement to existing research. Furthermore, the expression pattern of SERPINA1 in colon cancer remains contentious. The cProSite database, which incorporates proteomic and phosphoproteomic data from the National Cancer Institute's Clinical Proteomic Tumor Analysis Consortium and International Cancer Proteogenome Consortium (45), indicates that SERPINA1 is downregulated in colon cancer (Fig. S2). By contrast, the GEPIA2 database, which performs gene expression analysis based on tumor and normal samples from the TCGA and GTEx databases (46), reports its upregulation. In the present study, SERPINA1 expression was elevated in tumor tissues compared with adjacent normal colon tissues, and was higher in deceased patients compared with surviving patients. Furthermore, in the GEPIA2 database cohort, patients with higher SERPINA1 expression exhibited longer survival. However, in the present study, increased SERPINA1 expression was associated with shortened survival in patients with colon cancer, a finding consistent with observations in pancreatic ductal adenocarcinoma (30). This discrepancy in survival outcomes between the GEPIA2 database and the present study may be attributed to the difference between RNA-sequencing data (from GEPIA2) and protein expression data (from the present study). The post-transcriptional and post-translational modifications of SERPINA1 may lead to inconsistencies between its mRNA and protein expression levels, which also highlights the necessity of protein-level experimental validation for bioinformatics and omics findings. Additionally, a review of published literature indicated that SERPINA1 was commonly upregulated across multiple types of cancer (30–32). Thus, the present study provides experimental evidence that helps to clarify the discrepant findings between the cProSite and GEPIA2 database cohorts.

The present study has several limitations that need to be addressed. First, the protein expression levels of the other three candidate genes (C3, F8 and SERPING1) were not experimentally validated. Since mRNA expression levels do not always correlate with protein abundance due to post-transcriptional regulation, future studies should therefore employ immunohistochemistry or western blot analysis to confirm their protein expression in colon cancer tissues. Second, the meta-analysis was performed solely for SERPINA1 expression. Consequently, the diagnostic or prognostic value of the other three genes (C3, F8, SERPING1) across different cohorts remains elusive. Further meta-analyses integrating multiple independent datasets are needed to evaluate their clinical relevance. Third, no experiments were conducted to validate potential compounds targeting SERPINA1. Thus, the druggability of SERPINA1 suggested by the Open Targets Platform is purely computational and lacks experimental support. Future research should perform in vitro or in vivo assays (e.g., viability, apoptosis or targeted inhibition assays) to test candidate compounds. Finally, the precise functional role of SERPINA1 in patients with colon cancer has yet to be experimentally elucidated. This is a critical limitation because without functional validation (e.g., via knockdown or overexpression models), the observed expression differences cannot be causally linked to tumor progression or patient outcomes. Future investigations using gain- or loss-of-function approaches in colon cancer cell lines or animal models are required to define its mechanistic role.

In summary, the present study analyzed the expression of 88 genes in the CCC pathway and identified 32 DEGs. Among these, four hub genes were associated with the survival of patients with colon cancer. SERPINA1, an upregulated gene, was further validated in colon tissue microarrays via IHC analysis, showing that its upregulation was associated with worse survival outcomes. These findings suggest that SERPINA1 may serve as a potential diagnostic biomarker for colon cancer and represents a promising candidate for therapeutic targeting.

Supplementary Material

Supporting Data

Supporting Data

Acknowledgements

Not applicable.

Funding

This research received financial support from University Student Innovation and Entrepreneurship Training Plan of Heilongjiang Province (grant no. S202510222084).

Availability of data and materials

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

Authors' contributions

PX conceived and designed the study and revised the manuscript. DL collected, analyzed and interpreted the data, and wrote the manuscript. DH designed the differential expression analysis pipeline and interpreted the transcriptome data. KL performed the functional enrichment analysis and curated all public datasets used in this study. HJ conducted statistical analysis of bioinformatics data and generated all related figures. DH, KL and HJ critically revised the bioinformatics sections of the manuscript for important intellectual content. PX and DL confirm the authenticity of all the raw data. All authors read and approved the final manuscript.

Ethics approval and consent to participate

The research complied with the principles outlined in the Declaration of Helsinki and received formal approval from Shanghai Biotechnology Co., Ltd. (approval no. SHYJS-CP-1707004). Written informed consent was secured from all participants prior to their involvement.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Use of artificial intelligence tools

During the preparation of this work, DeepSeek-V3.1 (https://chat.deepseek.com/) was used to improve the readability and language of the manuscript or to generate images, and subsequently, the authors revised and edited the content produced by the artificial intelligence tools as necessary, taking full responsibility for the ultimate content of the present manuscript.

Glossary

Abbreviations

Abbreviations:

References