Analysis of the correlation between combined multiple cytokine detection and colorectal cancer.

Introduction
Colorectal cancer is one of the most common malignancies of the digestive system. In 2022, it had the third-highest incidence and the second-highest mortality rate worldwide1. Currently, various factors, such as obesity, dietary pattern changes, lack of exercise, sedentary behavior, aging population, significant intake of red or processed meat, smoking, and heavy alcohol consumption, have led to continuous increases in the incidence and mortality of colorectal cancer in China2. Nearly one-quarter of colorectal cancer patients have distant metastases at diagnosis3, and roughly half of them die. Clinically, tumor markers for colorectal cancer screening lack sufficient sensitivity and specificity. Thus, most patients require a combination of imaging, endoscopy, and diagnostic pathology. Therefore, early screening and diagnostic staging of colorectal cancer are necessary to improve the quality of life for these patients. Numerous studies have indicated that tumors often occur at chronic inflammation sites, primarily characterized by immune cell infiltration, tissue damage, fibrosis, and angiogenesis4. The cancer immunoediting hypothesis presumes that three immune response phases precede colorectal cancer development: elimination, equilibrium, and escape5. During these stages, the expression of immune cells is crucial. Thus, cytokines participate in various immune regulation, inflammation, and tumor-promoting effects as essential mediators in information transmission between immune cells. This study detected the levels of 14 cytokines, including interleukin-1β (IL-1β), IL-10, IL-17 A, IL-4, IL-5, IL-6, IL-8, IL-22, IL-33, IL-18, IL-2RA, tumor necrosis factor-α (TNF-α), TNF-β, and interferon-γ (IFN-γ), in the serum of patients with colorectal cancer. We analyzed the diagnostic value of individual and combined cytokine levels in patients with colorectal cancer by constructing different models and developing new evidence to diagnose colorectal cancer.

Materials and methods

Source of clinical specimens and ethics
The study population included patients newly diagnosed with colorectal cancer from the Shanxi Cancer Hospital between April 1, 2020, and November 30, 2020, using healthy individuals as control samples. A total of 56 colorectal cancer patients were enrolled, including 29 males and 27 females. (All study participants were recruited during the same period, ensuring data consistency; moreover, as a cross-sectional observational study aimed at analyzing cytokine expression characteristics at a specific time point, the research fully complies with the requirements of the Ethics Committee). Additionally, 25 healthy individuals were included, with 12 males and 13 females. No statistically significant differences were observed in terms of sex (P = 1.00) or age (P = 0.584) between the colorectal cancer and healthy control groups. The inclusion criteria for the 56 patients with colorectal cancer were: (1) no history of radiotherapy, chemotherapy, or targeted therapy before enrollment; (2) no abnormal liver and kidney function, concurrent immune system diseases; and (3) no use of antibiotics, probiotics, immunomodulators, hormones, or other medications within one month of peripheral blood collection. All patients in the colorectal cancer group were initially diagnosed with colorectal adenocarcinoma by pathological examination in hospitals at or above the secondary level. The exclusion criteria were (1) concurrent malignant tumors other than colorectal cancer; (2) severe complications associated with colorectal cancer, including active bleeding, acute intestinal obstruction, and tumor perforation; (3) severe underlying diseases, including heart failure, liver failure, active pulmonary tuberculosis, systemic lupus erythematosus (SLE), and type 2 diabetes mellitus (T2DM); (4) the receipt of chemotherapy or radiotherapy within 4 weeks prior to study enrollment; and (5) the receipt of antiviral agents, anti-infective agents, immunosuppressants, or biological agents within 2–4 weeks prior to study enrollment. The inclusion criteria for the healthy participants included: (1) adults aged > 18 years; (2) no history of gastrointestinal or other digestive system disorders, immune system diseases, diabetes, or other underlying diseases; and (3) no history of infectious diseases in the past month. Informed consent was obtained from all subjects for experiments involving human blood. The procedures were carried out under the principles expressed in the Declaration of Helsinki. This research was approved by the Ethics Committee of Shanxi Province Cancer Hospital.

Main reagents and instruments
The Aimplex® Human Custom 14-Plex Kit was procured from Beijing Kuangbo Biotechnology Co., Ltd. The pre-mixed multi-factor kit has pre-mixed antibody-coupled microspheres with pre-mixed biotin-coupled detection antibodies. The cytokines included in the detection of colorectal cancer-related serum factors were IFN-gamma (catalog number: A111122), IL-1beta (catalog number: A111128), IL-10 (catalog number: A111131), IL-17 A (catalog number: A111146), IL-4 (catalog number: A111173), IL-5 (catalog number: A111176), IL-6 (catalog number: A111179), TNF-alpha (catalog number: A111209), TNF-beta (catalog number: A111212), IL-8 (catalog number: A112224), IL-22 (catalog numbers: A111161), IL-33 (catalog number: A111419), IL-18 (catalog number: A111426), CD25 (catalog number: A113269), and the NR Basic kit (catalog number: P100001). The flow cytometer (BD FACSCantoTM II) was procured from BD Biosciences, USA.

Serum cytokine level measurements in patients with colorectal cancer using aimplex high-throughput multi-factor flow cytometry technology
Patients in the colorectal cancer and healthy control groups provided peripheral venous blood (2 mL per person). The blood samples were centrifuged at 1,200×g for 12 min, and serum was retrieved and stored at -80 °C for further use. Serum(15µL) was taken for each sample. The serum concentrations of 14 cytokines, including IL-1β, IL-10, IL-17 A, IL-4, IL-5, IL-6, IL-8, IL-22, IL-33, IL-18, IL-2RA, TNF-α, TNF-β, and IFN-γ, were simultaneously measured following the Aimplex® Human Custom 14-Plex Kit instructions. The experimental procedure is summarized as follows. Pre-mixed antibody-coupled microspheres were incubated with 15 µL of serum for 60 min and then incubated with the biotin-coupled detection antibodies for 30 min. Subsequently, streptavidin labeled with phycoerythrin was added and incubated for 20 min. Then, the concentrations of the 14 cytokines were analyzed by flow cytometry, and data processing was performed using FCAPArray 3.0 software. Three parallel samples were prepared and analyzed for each subject.

Statistical analysis
The experimental data were analyzed using R software (version 4.4.0). The Kolmogorov-Smirnov test was used to evaluate whether the data were normally distributed. Medians (interquartile range) and numbers (percentages) are used to describe degree of dispersion without normal distribution and categorical variables. The Mann-Whitney U and chi-squared tests were used to examine differences between groups. Logistic regression models were used to describe the correlation between cytokines and cancer incidence, applying natural logarithm transformation to all cytokine concentrations. Restricted cubic spline (RCS) curves were used to explore nonlinear associations between cytokines and cancer incidence. We used the Bayesian Kernel Machine Regression (BKMR) model to determine the mixed effects of the 14 cytokines on colorectal cancer occurrence based on the potential nonlinear relationships and interactions between various cytokines. Statistical significance was defined as P < 0.05, and a more stringent threshold of P < 0.01 was used to denote stronger statistical significance for key outcomes.

Result

General clinical data of study subjects
The colorectal cancer group comprised 56 patients with a median age of 61 (52, 66) years, including 29 males and 27 females. The sample size of this study was validated using the two-independent-samples t-test. The results indicated that the statistical power was ≥ 80%, which is sufficient to detect the hypothesized effect while minimizing the risk of type II error (false negative results). Therefore, the current sample size is adequate to draw robust conclusions. The control group had 25 healthy individuals with a median age of 62 (53.75, 68) years, including 12 males and 13 females.

Expression differences of 14 serum cytokines between patients with colorectal cancer and healthy individuals
The AimPlex high-throughput multi-factor flow cytometry detection (Table 1) results indicated that the levels of two cytokines, IL-2RA and IL-6, were significantly increased in the patient group compared to the healthy control group (P < 0.01). IFN-γ, IL-8, and IL-5 levels were elevated in the colorectal cancer group (P < 0.05). No statistically significant differences in the expression of the other cytokines were observed between the two groups (P > 0.05). Table 1 shows a comparison of the distribution of cytokines in the two groups.

Analysis of 14 serum cytokines in the logistic regression model
Table 2 demonstrates the construction of a multivariate logistic regression equation incorporating factors like age and sex. The results showed that IL-6 had a statistically significant association with colorectal cancer occurrence (P < 0.01). Colorectal cancer risk was increased by 4.66 times for every one-unit Ln-IL-6 increase. Additionally, Ln-IL-2RA, Ln-IFN-γ, and Ln-IL-5 were statistically significantly associated with the incidence of colorectal cancer (P < 0.05). Moreover, for every one-unit increase in Ln-IFN-γ, the colorectal cancer risk increased by 4.01 times.

RCS nonlinear curve model analysis
In this study, cytokines with statistically significant differences or marginal effects (P < 0.1) were analyzed, and marginal effects in the logistic regression model to RCS nonlinear analysis (represented in Figs. 1 and 2 (post-LN transformation)). IL-6 showed statistically significant differences in the untransformed and LN-transformed nonlinear models (P = 0.0029, P = 0.0052). The results showed a positive correlation trend with colorectal cancer incidence. Additionally, IL-10 was statistically significant in the model post-LN transformation (P = 0.0497), suggesting that the risk of colorectal cancer also increases as IL-10 levels increase.

BKMR mixed exposure model analysis
The above-mentioned 14 cytokines in this study were considered as mixed exposures. The BKMR model was used to analyze the impact of these 14 mixed cytokine exposures on the occurrence of colorectal cancer. In this model, when the 14 cytokines were fixed simultaneously at different percentiles and at the median, the risk of colorectal cancer increased by 0.95 times under mixed exposure conditions when all the cytokines were at the 75th percentile rather than at the 50th percentile (Table 3; Fig. 3). Figure 4 A demonstrates the dose-response relationships between the 14 cytokines and colorectal cancer. When other cytokines were fixed at the 50th percentile (P50), Ln-IL-10, Ln-IL-6, and Ln-IFN-γ had positive and linear correlations with colorectal cancer occurrence. In contrast, Ln-IL-17 A, Ln-IL-8, and Ln-IL-1β had negative and approximately linear correlations with colorectal cancer occurrence. Figure 4B shows that the colorectal cancer risk increased by 0.29 times, 0.248 times, and 0.4919 times when the remaining cytokines were exposed at P50. Moreover, the risk increased when the exposure levels of Ln-IL-6, Ln-IFN-γ, and Ln-TNF-β were elevated from the 25th (P25) to the 75th (P75) percentile. In contrast, the risk of colorectal cancer decreased by 0.046 times when the Ln-IL-2RA (natural logarithm-transformed interleukin-1 receptor antagonist) exposure level increased from P25 to P75. Figure 5 shows a bivariate exposure-response function plot for the cytokines. The plot indicated no interactive effects of the 14 cytokines and colorectal cancer occurrence.

Discussion
Colorectal cancer, one of the most common malignancies of the digestive tract, presents with early hidden symptoms. The disease lacks effective, sensitive, and specific methods to ensure early diagnosis. The tumor microenvironment primarily comprises immune cells, fibroblasts, extracellular matrix molecules, different enzyme molecules, and several cytokines, chemokines, and protein factors6 in a dynamic evolutionary process. Therefore, tumor progression accumulates numerous immunosuppressive cells and inflammatory-related cytokines, jointly promoting tumor immune escape, growth, and metastasis5. Currently, the relationship between tumors and inflammation is garnering increased attention. These inflammatory cytokines develop a complex network associated with tumor modulation and development.
IL-2RA, also called CD25, is the α-chain of the IL-2 receptor and is expressed on activated T lymphocytes. IL-2RA participates in T-lymphocyte proliferation and differentiation. One study using immunohistochemistry and in situ hybridization techniques reported that IL-2 is expressed in the cytoplasm and nucleus of tumor cells7. Thus, IL-2RA-positive cells exhibit faster cell division and chromosomal instability, promoting tumor invasiveness and poor patient prognosis8. This study identified a close association between IL-2RA and the occurrence of colorectal cancer through a rank sum test and logistic regression model analysis. Additionally, strong IL-2RA expression is commonly observed in solid tumors, including lung, prostate, and esophageal cancer9. Clinical studies have indicated that gastric and liver cancer patients often have abnormally high serum sIL-2R levels. Serum sIL-2R levels are significantly higher among patients with metastasis, recurrence, and stage III/IV tumors than those without invasion, metastasis, recurrence, and stage I/II tumors. Therefore, sIL-2R levels could become an independent prognostic indicator for gastric and liver cancer10,11. Furthermore, serum sIL-2R level monitoring in thyroid, breast, and ovarian cancer patients has revealed that serum sIL-2R levels increase when tumors metastasize or recur. Therefore, these tumors have the characteristics of synthesizing and secreting sIL-2R. Currently, the clinical use of drugs like IL-2/IL-2R complexes or antagonists can result in partial or complete remission in adult patients with T-cell leukemia9. However, their effectiveness in treating solid tumors remains limited. Therefore, IL-2 delivers good clinical value for diagnosing, prognosing, and assessing advanced colorectal tumors. Hence, research on its combination with antagonists could be a future direction for colorectal cancer treatments.
IL-6 is a pleiotropic cytokine involved in malignant tumor growth, invasion, and metastasis12,13. This study used the rank sum test, the logistic regression model, and the RCS curve model. They all showed statistically significant positive correlations between IL-6 and colorectal cancer occurrence. Taniguchi et al. reported that serum IL-6 levels are linked to the prognosis, survival, and disease staging of malignant tumors, including lung, esophageal, breast, ovarian, and kidney cancer14. Sasaki et al. reported that elevated serum IL-6 levels were positively associated with the presence of colorectal adenomas15,16. IL-6 activates the IL-6/STAT3 pathway by reverse signaling. This activation is crucial in the initiation and progression of colorectal cancer17,18. Furthermore, studies have demonstrated that anti-IL-6 receptor antibody treatment decreased the incidence of colitis-associated cancer in mice with colitis-associated cancer by reducing the expression of critical genes associated with aerobic glycolysis19.
Activated T lymphocytes and NK cells primarily produce IFN-γ. Studies have demonstrated elevated IFN-γ + cellular levels in Crohn’s disease20. Research using mouse models suggests that IFN-γ can induce early inflammation21. However, because of its immunosuppressive effects, IFN-γ is not a specific diagnostic indicator of colorectal cancer, as it also inhibits the cytotoxic functions of CTL and NK cells. IL-5, primarily synthesized by T lymphocytes and mast cells, is responsible for the recruitment, proliferation, maturation, and activation of eosinophils22, promoting their adhesion and migration to bone marrow proteins23. Additionally, IL-5 stimulates the growth and differentiation of B lymphocytes while inducing mast cells by enhancing the production of pro-tumor and pro-fibrotic cytokines24. Studies have revealed that IL-5 can stimulate tumor cell migration and activation via the STAT5 signaling pathway. Moreover, the IL-5/IL-5Rα pathway has pro-tumor effects within the tumor microenvironment of mice and human pancreatic tumors. IL-5/IL-5Rα expression elevates during pancreatic inflammation and decreases in advanced lesions24. In this study, modeling showed that IL-5 levels in the colorectal cancer group were significantly and statistically higher than in the healthy control group, consistent with the research mentioned above. Cytokines like IFN-γ and IL-5 are closely associated with the occurrence and development of colorectal cancer.
The current study analyzed the role of 14 cytokines as mixed exposure factors in colorectal cancer using the BKMR model. The results indicated that the risk of colorectal cancer increased by 0.95 times when all the cytokines were at the 75th percentile compared to the 50th percentile. A positive association was observed between the mixed effects of the 14 cytokines and colorectal cancer occurrence. Among them, the cytokines with the greatest contributions included IL-6, IL-5, IFN-γ, IL-2RA, IL-8, and IL-10. The current model provides a novel diagnostic and therapeutic approach for the future occurrence and development of colorectal cancer while enhancing the diagnostic value.
The current study analyzed 14 serum cytokines using rank sum tests, logistic regression models, RCS curve models, and the BKMR model. These tests and models provide new insight into the future diagnosis of colorectal cancer. In summary, mixed exposure analysis of these 14 cytokines can improve the diagnosis of colorectal cancer. Unfortunately, the study sample size was relatively small. Dynamic monitoring during the colorectal cancer development stages while dynamically analyzing changes in multiple serum cytokines before and after surgery in a larger sample population can help better translate the combined detection of various cytokines into a clinical diagnostic setting. Finally, the current study revealed a unique diagnostic method for the combined detection of 14 cytokines in colorectal cancer while providing experimental evidence for immunotherapy against malignant tumors.
Recent studies have revealed significant alterations in cytokine profiles in patients with colorectal cancer (CRC). Research indicates that the lipopolysaccharide (LPS)-induced production of TNF-α and IFN-γ is reduced in CRC patients, and this reduction is associated with disease stage. Notably, the production of these cytokines can be restored following surgical tumor resection25. Additionally, the abnormal expression of various cytokines and chemokines in serum has been observed, which may be correlated with tumor metastasis and chemotherapy tolerance26. These findings suggest that cytokines not only participate in CRC progression but also serve as potential molecular indicators for predicting treatment efficacy.
In the development and progression of CRC, pro-inflammatory cytokines, such as TNF-α and IL-6, can promote tumor growth, invasion, and metastasis by activating the nuclear factor-κB (NF-κB) signaling pathway. In contrast, cytokines, including IL-10, IL-18, and IFN-γ, exhibit potential tumor-suppressive effects.
Regarding novel therapeutic approaches for CRC, studies have demonstrated that N, N-dibenzylasparagine (NNDAsp) can inhibit the proliferation of CRC cells and induce their apoptosis27. Another study found that selenium-enriched Pleurotus ostreatus extract could suppress CRC cell proliferation and regulate cytokine expression28. Both strategies provide new directions for CRC treatment. With the continuous exploration and innovation of researchers, more advanced therapeutic methods for CRC could be developed in the near future.
Original contributions of this study: Firstly, this study represents the first and simultaneous detection of 14 serum cytokines to compare their expression differences between colorectal cancer patients and healthy individuals. This high-throughput multiplex assay is relatively uncommon in colorectal cancer diagnostic research and provides a more comprehensive immune status profile. Secondly, the introduction of analytical models such as Restricted Cubic Spline (RCS) and Bayesian Kernel Machine Regression (BKMR) enhances the reliability and scientific rigor of our findings. Finally, this work underscores the clinical value of serum cytokines as a potential biomarker panel for colorectal cancer, offering experimental evidence and theoretical support for future research into non-invasive and highly sensitive blood-based diagnostics.