Exploring macrophage polarization as a prognostic indicator for colorectal cancer: Unveiling the impact of metalloproteinase mutations

Abstract

BACKGROUND

Macrophages play a crucial role in the tumor microenvironment, displaying remarkable plasticity that allows them to either suppress or promote tumor progression. Their polarization into M1 or M2 phenotypes could have significant prognostic implications, and manipulating this polarization may offer a novel approach to controlling colorectal neoplasms.

AIM

To evaluate the infiltration rates of M1 and M2 macrophages in colorectal neoplasia, specifically comparing cases with and without metalloproteinase mutations. Additionally, it sought to explore potential prognostic factors associated with the disease.

METHODS

The study involved two cohorts of patients diagnosed with colorectal neoplasia: 33 patients with metalloproteinase mutations and 33 without. Macrophage quantity and polarization were assessed using markers indicative of M1 (iNOS) and M2 (CD163, CD206) macrophages. Prognostic factors and survival outcomes related to colorectal cancer (CRC) were also analyzed.

RESULTS

Among the 61 patients, 28 (45.9%) exhibited metalloproteinase mutations, while 33 (54.1%) did not. Tumor staging revealed that 16.9% were in stage I, 34.2% in stage II, 42.4% in stage III, and 8.5% in stage IV. The study recorded 12 patient deaths (19.7%), with 21.2% from the control group and 17.9% from the mutation group. M2 macrophages, identified by CD163 and CD206 markers, had mean counts of 23 and 17, respectively, with standard deviations of 21 and 17. In contrast, M1 macrophages, identified by iNOS, had a mean count of five per site, with a standard deviation of 11.

CONCLUSION

The study found no statistically significant differences in macrophage density between groups, irrespective of metalloproteinase mutation status, age, gender, tumor region, staging, TILS, tumor recurrence, or clinical outcomes. No association was observed between macrophage polarization and CRC prognosis or survival. However, patients with metalloproteinase mutations demonstrated a better survival rate, suggesting a potential protective role of this mutation in colorectal neoplasia.

Key Words: Colorectal neoplasia; Polarized macrophages; Colorectal adenocarcinoma; Microsatellite instability

Core Tip: This study investigates the prognostic significance of macrophage polarization in colorectal cancer (CRC), focusing on patients with and without metalloproteinase mutations. Using M1 (inducible nitric oxide synthase) and M2 (CD163, CD206) markers, macrophage densities were quantified and correlated with survival and clinical outcomes. While no significant differences in macrophage density or polarization were observed between groups, the study highlights a potential protective role of metalloproteinase mutations, as patients harboring these mutations demonstrated improved survival rates. This insight may guide future research into the tumor microenvironment and its implications for CRC prognosis and therapy.

INTRODUCTION

Colorectal cancer (CRC) is the third most common type of malignant tumor worldwide. In 2020, global statistics revealed 1.9 million new CRC cases and 940000 deaths, marking a notable rise from previous years[1]. Similarly, in Brazil, estimates predict 45600 new cases annually from 2023 to 2025, equating to 21.10 cases per 100000 inhabitants. Women are expected to have slightly higher case numbers than men[2].

Risk factors for CRC can be classified as intrinsic or extrinsic. Intrinsic factors include genetic and epigenetic changes, such as oncogene activation and signals from the cellular microenvironment, which are pivotal in cancer development. Mutations in key genes like APC, KRAS, and TP53 are central to CRC pathogenesis. Extrinsic factors include lifestyle choices and environmental influences, such as obesity, lack of physical activity, smoking, alcohol use, and the consumption of processed foods. These extrinsic factors induce metabolic dysregulation and chronic inflammation, promoting carcinogenesis through cellular damage and alterations in pathways that regulate cell proliferation and programmed cell death[3].

CRC is a complex disease characterized by various molecular processes leading to the establishment of pathological features, from initiation to metastasis. These features, termed the "hallmarks of cancer," were outlined by Hanahan and Weinberg[4] in 2011 and include ten key characteristics: Genomic instability, sustained proliferative signaling, evasion of tumor suppressors, resistance to cell death, cellular immortality, angiogenesis promotion, metastasis capability, altered energy metabolism, immune evasion, and tumor-induced inflammation.

Significantly, many of these hallmarks are not solely dependent on the cancer cells but are also influenced by the tumor microenvironment (TME), which consists of the surrounding cellular and molecular milieu. The TME plays a pivotal role in cancer progression by fostering angiogenesis, immune evasion, and metastasis. Thus, the interactions between cancer cells and their microenvironment are critical for the progression and dissemination of CRC[4].

Cancer progression involves multiple steps, including genetic and epigenetic changes traditionally thought to affect only cancer cells. Recent understanding emphasizes the dynamic interactions between cancer cells and the TME. Paget's "seed and soil" analogy highlights this relationship, where the TME, consisting of fibroblasts, endothelial cells, immune cells, extracellular matrix components, and soluble factors, plays a crucial role[5,6].

The tumor stroma, composed of blood vessels, fibroblasts, and immune cells, supports cancer growth, promoting angiogenesis and suppressing immune responses. Cancer-associated fibroblasts (CAFs) are pivotal in this environment, releasing cancer cells from growth suppression and driving local inflammation and invasion. CAFs also promote angiogenesis and inhibit cytotoxic T lymphocytes, contributing to tumor progression[4].

Due to the complexity of cancer processes, efforts focus on targeting TME interactions for therapeutic advancements. Cancer cells often develop resistance to current treatments, making other TME components, such as more stable stromal cells, promising targets for intervention. Macrophages, abundantly present in tumors and associated with poor prognosis in some cancers, represent another potential therapeutic target[7].

The role of macrophages in the prognosis of CRC remains under-researched and requires further elucidation. Macrophages are immune cells present in nearly all tissues, and they express specific markers like CD68, CD163, and iNOS[7-9]. Originally identified by Metchnikoff in the 19th century, macrophages maintain tissue integrity through patrolling and clearing pathogens and debris[9]. Monocyte-derived macrophages originate from bone marrow, while embryonic macrophages are derived from the yolk sac and fetal liver. Both types are essential for tissue homeostasis and inflammation recovery, particularly through apoptotic cell removal via efferocytosis[9,10].

Carcinogens trigger inflammation, typically mediated by innate immunity. Macrophages, along with neutrophils, respond to harmful signals, engulfing pathogens and secreting molecules like nitric oxide to degrade them[10]. These cells are classified into M1 (pro-inflammatory) and M2 (anti-inflammatory) types[11]. M1 macrophages produce tumoricidal molecules and sustain inflammation by recruiting other immune cells like T cells[7,12]. CSF-1 has been shown to affect macrophage infiltration and tumor progression, with inhibition increasing CD8+ T-cell recruitment and reducing tumors[13].

When inflammation fails to control tumors, M2 macrophages resolve the inflammation, remove apoptotic cells, regulate tissue integrity, and are exploited by cancer cells to promote tumor growth and invasion[9,14,15]. Tumor-associated macrophages (TAMs) originate from monocytes recruited into the tumor tissue, differentiating in response to signaling through cytokines, chemokines, and growth factors produced by stromal cells and the TME[16]. Macrophages are key components of the TME, and M2 macrophages are predominant in most tumors, promoting invasive phenotypes[7]. High densities of M2 macrophages are linked to poor prognosis in various cancers, though in CRC, M2 macrophages correlate with a better prognosis. This discrepancy may arise from differences in macrophage marker identification and their phenotypic plasticity[11].

The TME in CRC is shaped by somatic and genetic mutations, alongside epigenetic changes. The immune system exerts selective pressure on tumor cells through the recognition of non-self antigens, mediated by immune checkpoints that regulate cell repair or destruction. Failures in this system may allow tumor cells to evade detection, leading to their persistence[17]. Immunosuppressive mechanisms in CRC include the recruitment of regulatory immune cells, such as Tregs and M1 macrophages. Additionally, the Warburg effect, where glycolysis is favored even under normoxia, supports neoplastic proliferation and creates an inhospitable environment for T cells[17].

Microsatellite instability (MSI) tumors, which comprise 15% of CRC cases, exhibit a high mutation rate and a significantly increased presentation of neoantigens compared to microsatellite-stable tumors, making them more susceptible to cytotoxic T-cell infiltration and immunotherapy[18]. DNA repair proteins, such as MLH1 and PMS2, form heterodimers essential for correcting small DNA defects. Mutations in these genes lead to the degradation of their corresponding dimers and an accumulation of mutational events, contributing to CRC development[18].

Li et al's meta-analysis[19] revealed that patients with wild-type metalloproteinases had better survival compared to mutated forms, although differing methods for macrophage identification complicate conclusions. Recent advances show promising immunotherapy outcomes in MSI-positive CRC patients, but the absence of MSI in some tumors leaves a gap for alternative therapeutic exploration. Macrophage interactions within the TME remain a critical area for future research to improve CRC treatments[20].

This study aims to evaluate the infiltration rates of M1 and M2 macrophages in CRC tumors, both with and without mutations in metalloproteinases. Specifically, the study seeks to examine the correlation between macrophage infiltration rates and key prognostic factors of CRC, as well as to explore the impact of these infiltration rates on patient survival.

MATERIALS AND METHODS

Study design

This retrospective study assessed 33 cases of colon and rectal adenocarcinoma with mutations in the metalloproteinases MLH1, PMS2, MSH6, and MSH2. A control group of 33 cases was selected, consisting of the next consecutive patients diagnosed with CRC, who exhibited a wild-type profile for the same metalloproteinases. The analysis focused on clinicopathological variables, including age at diagnosis, clinical staging (stages I-IV), treatment modalities (surgery or chemoradiotherapy), disease-free survival, and overall survival. The study followed the REMARK guidelines established by the NCI-EORTC Working Group on Cancer Diagnostics, ensuring adherence to the gold standard for biomarker research.

Study setting

The medical records were obtained from the Pathology Unit database of the General Hospital of Caxias do Sul. The data collection encompassed records between January 1, 2018, and December 31, 2021. The General Hospital of Caxias do Sul is a reference center for CRC treatment, providing comprehensive data on CRC cases, including molecular and histopathological profiles relevant for biomarker analysis.

Patients

The patient cohort consisted of 33 individuals with colon and rectal adenocarcinoma exhibiting mutations in metalloproteinases, alongside 33 control patients with a wild-type metalloproteinase profile. Clinical data, including diagnosis date, staging, and survival outcomes, were collected from electronic medical records. All aspects of patient confidentiality and data protection were strictly maintained, with approvals from the hospital’s Scientific and Editorial Council and the Research Ethics Committee, as well as compliance with privacy protocols.

Immunohistochemistry

The variables under investigation included the presence of macrophage membrane markers CD68, iNOS, CD206, and CD163. CD68 served as a marker for total lymphocytes, while CD206 and CD163 were used to identify M2-polarized macrophages, with iNOS indicating M1-polarized macrophages. Original hematoxylin-eosin stained slides were reviewed for diagnostic confirmation. Areas with the highest concentration of intratumoral macrophages were identified and subsequently examined on the immunohistochemistry slides.

Immunohistochemistry slides were prepared using 3 μm thick cuts from tissue fixed in 10% buffered formalin and embedded in paraffin. The slides underwent deparaffinization and rehydration, followed by antigen retrieval using Tris-EDTA (Thermo Fisher, Waltham, MA) at 95 °C for 20 minutes. Primary antibodies against iNOS (dilution 1:100, polyclonal, Sigma-Aldrich^®), CD206 (dilution 1:300, 5C11, Sigma-Aldrich^®), CD163 (dilution 1:200, polyclonal, ArigoBio^®), and CD68 (dilution 1:1500, K-p1, Cell Marque^®) were incubated overnight and subsequently stained with DAB. Macrophages from the tissue itself served as internal controls. The immunohistochemistry slides were examined independently by two pathologists to evaluate the percentage of stained macrophages. The percentage was assessed visually at high magnification (400×). Discrepancies in interpretation were simultaneously reviewed by both pathologists, leading to a consensus.

Statistical analysis

The data analysis involved descriptive and inferential statistical methods for group comparisons. Quantitative variables were described using mean and standard deviation, while qualitative variables were presented as absolute and relative frequencies. Inferential statistical analyses employed both parametric and non-parametric methods, considering the independent design of the variables. For quantitative variables, one-way and two-way analysis of variance (ANOVA) tests were performed, with normality and homogeneity of variances checked to apply the appropriate statistical tests. Multivariate analyses, including MANOVA, were also conducted.

For qualitative variables, the χ² test was used to assess independence. Fisher's exact test was applied for cases with two-way contingency tables, while Pearson's χ² test of maximum likelihood ratio was used for larger tables. Additionally, cluster analysis was performed using the k-means algorithm, determining the optimal number of clusters via the elbow method. This involved plotting explained variance against the number of clusters to select the most appropriate grouping. To evaluate the internal quality of clusters, the stability of the ideal solution was calculated using Jaccard bootstrap values with 10000 iterations. A two-tailed P value of < 0.05 was deemed statistically significant. All analyses were conducted using R programming language (version 4.3.1, The R Foundation for Statistical Computing).

RESULTS

The sample consisted of 66 cases. Of this total, five patients did not meet all the required data criteria and were excluded, leaving 61 patients for statistical evaluation. Regarding the presence of mutations in metalloproteinases, 33 patients (54.1%) did not exhibit mutations, while 28 patients (45.9%) presented the mutation (Table 1).

Table 1 Variable analysis regarding microsatellite instability (n = 61).

	Total		No mutation, n = 33		With mutation, n = 28		P value
	n	%	n	%	n	%
Age (year)							0.1351
≤ 60	15	24.6	11	33.3	4	14.3
> 60	46	75.4	22	66.7	24	85.7
Sex							0.0931
Female	43	70.5	20	60.6	23	82.1
Male	18	29.5	13	39.4	5	17.9
Location							0.1541
Right colon	44	72.1	21	54.5	23	78.6
Left colon	17	27.9	12	45.5	5	21.4
Staging							0.051
I	10	16.9	3	9.7	7	25.0
II	19	32.2	7	22.6	12	42.9
III	25	42.4	18	58.0	7	25
IV	5	8.5	3	9.7	2	7.1
Colectomy							0.248
Right side	39	63.9	18	54.5	21	75.0
Left side	15	24.6	10	30.3	5	17.9
Total	7	11.5	5	15.2	2	7.1
Lymphocytes							0.24
Low	30	50.0	18	54.6	12	44.4
Moderate	18	30.0	7	21.2	11	40.7
Intense	12	20.0	8	24.2	4	14.9
Reccurence							0.4431
No	49	87.5	26	83.9	23	92.0
Yes	7	12.5	5	16.1	2	8.0
Clinical outcome							1.01
Disease-free	49	80.3	26	78.8	23	82.1
Death	12	19.7	7	21.2	5	17.9

¹P values obtained via Fisher's exact test.

The samples were categorized into two groups: Patients aged up to 60 years and patients older than 60 years. This division showed that 46 patients (75.4%) were over 60 years old, with no statistically significant difference between the groups with and without metalloproteinase mutations (P = 0.135). Concerning gender, 43 samples (70.5%) were from female patients, and again, no statistical difference was found between the groups with and without metalloproteinase mutations (P = 0.093).

When evaluating tumor location, 44 patients (65.6%) had lesions located in the right colon, while 17 patients (27.9%) had lesions located in the left colon. No statistically significant difference was found between the groups with and without metalloproteinase mutations (P = 0.062). Surgeries performed included right-sided colectomy in 39 cases (63.9%), left-sided colectomy in 15 cases (24.6%), and total colectomy in seven cases (11.5%). Among the total colectomies, four patients had neoplasms in the right colon and three in the left colon. There was no statistical difference regarding the presence or absence of mutations in these two groups (P = 0.248).

Using staging to group the samples, 10 patients were in stage I (16.9%), 19 were in stage II (34.2%), 25 were in stage III (42.4%), and five were in stage IV (8.5%). When taking metalloproteinase mutations into account, stages I and II combined had the highest number of samples with mutations (67.9%), while stages III and IV together had the highest number of samples without mutations (67.7%). A statistically significant difference was identified between the staging groups with and without metalloproteinase mutations (P = 0.05).

By the end of the data collection period, seven patients (12.5%) exhibited disease recurrence, of whom two had metalloproteinase mutations (8%) and five did not (16.1%), with no statistically significant difference between the groups (P = 0.443).

There were 12 deaths (19.7%) during the study: Seven (21.2%) in the control group and five in the mutation group (17.9%). No statistically significant difference was identified between clinical outcomes and the presence or absence of mutations (P = 1.0). When analyzing by stage, one death occurred in a patient in stage I, five in stage II, three in stage III, and three in stage IV. The highest mortality occurred in stage IV, accounting for 60% of the cases (Table 2). Of the total deaths, three were due to coronavirus disease 2019 (COVID-19) and were unrelated to tumor recurrence. Among the patients who died from COVID-19 complications, one was in stage I and two in stage II, all of whom had metalloproteinase mutations. Additionally, two patients in the mutation group died from adenocarcinoma, and both were in stage IV.

Table 2 Variable analysis regarding staging (n = 61).

	Total		Stage I, n = 10		Stage II, n = 19		Stage III, n = 25		Stage IV, n = 5		P value
	n	%	n	%	n	%	n	%	n	%
Instability											0.05a
No mutation	33	52.5	3	30.0	8	40.0	18	72.0	4	66.66
With mutation	28	47.5	7	70.0	12	60.0	7	28.0	2	33.33
Age (year)											0.61
≤ 60	15	24.6	1	10.0	6	30.0	7	28.0	1	16.7
> 60	46	75.4	9	90.0	14	70.0	18	72.0	5	83.3
Sex											0.055
Female	43	70.5	9	90.0	17	85.0	14	56.0	3	50.0
Male	18	29.5	1	10.0	3	15.0	11	44.0	3	50.0
Location											0.722
Right colon	44	72.1	8	80.0	16	80.0	17	68.0	3	50.0
Left colon	17	27.9	2	20.0	4	20.0	8	32.0	3	50.0
Colectomy											0.342
Right side	39	63.9	7	70.0	15	75.0	14	56.0	3	50.0
Left side	15	24.6	3	30.0	3	15.0	6	24.0	3	50.0
Total	7	11.5	0	0.0	2	10.0	5	20.0	0	0.0
Lymphocytes											0.874
Low	30	50.0	3	33.3	10	50.0	14	56.0	3	50.0
Moderate	18	30.0	4	44.4	6	30.0	7	28.0	1	16.6
Intense	12	20.0	2	22.2	4	20.0	4	16.0	2	33.3
Recurrence											0.59
No	49	87.5	9	100.0	16	89.0	19	82.6	5	83.3
Yes	7	12.5	0	0.0	2	11.0	4	17.4	1	16.7
Clinical outcome											0.022a
Disease-free	49	80.3	8	80.0	17	85.0	22	88.0	2	33.3
Death	12	19.7	1	10.0	3	15.0	3	12.0	4	66.6

^aStatistically significant.

Qualitative variables expressed in values of absolute frequency and relative frequency.

The percentual value shown as relative frequency is taking into account the percentage of valid cases, that is, excluding the missing values.

In terms of time to death among patients with and without metalloproteinase mutations (Figure 1), the survival curves showed a statistically significant difference (P = 0.016). The group of patients with mutations exhibited longer survival, while those without mutations tended to have an earlier death. The data also showed considerable heterogeneity regarding the presence of macrophages within the neoplasms. In general, the quantitative variables exhibited a very high standard deviation, making their evaluation difficult. The CD68 marker for total macrophages showed an average of 32 macrophages per field with a standard deviation of 26. The CD163 and CD206 markers, indicative of M2-polarized macrophages, had averages of 23 and 17 macrophages, with standard deviations of 21 and 17, respectively (Table 3). In contrast, the iNOS marker for M1-polarized macrophages showed an average of five macrophages per field with a standard deviation of 11.

Figure 1 Survival curves of patients with or without metalloproteinase mutation. The blue line represents the patients without mutation (instability 1), while the purple line represents the patients with mutation (instability 2).

Table 3 Factorial analysis of macrophages regarding tumor location (n = 61).

		n	%	mean ± SD	P value
CD68					0.757
Right side	No mutation	21	47.7	31.667 ± 25.997
	With mutation	23	52.3	30.0 ± 25.849
Left side	No mutation	12	70.6	33.75 ± 28.534
	With mutation	5	29.4	37.2 ± 32.042
CD163					0.256
Right side	No mutation	21	47.7	23.881 ± 20.987
	With mutation	23	52.3	20.109 ± 19.934
Left side	No mutation	12	70.6	21.750 ± 24.076
	With mutation	5	29.4	33.3 ± 28.504
CD206					0.117
Right side	No mutation	21	47.7	19.057 ± 19.358
	With mutation	23	52.3	14.226 ± 15.599
Left side	No mutation	12	70.6	15.458 ± 12.648
	With mutation	5	29.4	27.4 ± 23.954
iNOS					0.179
Right side	No mutation	21	47.7	6.767 ± 13.487
	With mutation	23	52.3	3.57 ± 4.831
Left side	No mutation	12	70.6	3.383 ± 4.522
	With mutation	5	29.4	8.21 ± 15.199

When examining the effect of patient variables on macrophage density as assessed by the CD163, iNOS, and CD206 markers collectively, a statistically significant interaction effect was observed between instability and clinical outcome (Table 4, P = 0.019). Individually, the macrophage density obtained through these markers did not demonstrate a statistically significant difference between the groups when considering the studied variables, such as metalloproteinase mutation, age, gender, tumor location, staging, TME, recurrence, and clinical outcomes (Table 5).

Table 4 MANOVA analysis of macrophages CD163, iNOS e CD206 combined (n = 61).

	Degrees of freedom	F	Hotelling-Lawley test	P value
Instability	1	0.314	0.005	0.958
Age	1	0.537	0.028	0.659
Location	1	0.058	0.003	0.936
Staging	3	1.697	0.296	0.094
Recurrence	1	0.650	0.038	0.586
Clinical outcome	1	1.929	0.102	0.305
Instability vs location	1	1.061	0.006	0.322
Instability vs clinical outcome	1	2.592	0.141	0.019

Table 5 Macrophage evaluation (n = 61).

	Total		CD68		CD163		iNOS		CD206
	n	%	mean ± SD	P value	mean ± SD	P value	mean ± SD	P value	mean ± SD	P value
Instability				0.753		0.868		0.524		0.690
No mutation	33	54.1	32.424 ± 26.52		23.106 ± 21.809		5.536 ± 11.11		17.748 ± 17.097
With mutation	28	45.9	31.286 ± 26.541		22.464 ± 21.694		4.398 ± 7.518		16.579 ± 17.598
Age (year)				0.483		0.706		0.993		0.251
≤ 60	15	24.6	33.0 ± 21.696		22.5 ± 17.687		5.377 ± 9.082		20.067 ± 15.374
> 60	46	75.4	31.543 ± 27.861		22.913 ± 22.877		4.896 ± 9.819		16.280 ± 17.804
Sex				0.295		0.496		0.701		0.275
Female	43	70.5	33.953 ± 26.403		23.919 ± 21.043		4.526 ± 6.575		19.191 ± 18.764
Male	18	29.5	27.0 ± 26.183		20.167 ± 23.208		6.181 ± 14.644		12.483 ± 11.825
Location				0.77		0.923		0.332		0.551
Right colon	44	72.1	30.795 ± 25.63		21.909 ± 20.293		5.095 ± 4.803		16.532 ± 17.457
Left colon	17	27.9	34.765 ± 28.619		25.147 ± 25.121		4.803 ± 8.772		18.971 ± 16.879
Staging				0.371		0.454		0.402		0.973
I	10	16.9	24.5 ± 21.141		15.65 ± 15.431		5.06 ± 6.184		16.1 ± 15.911
II	19	32.2	37.368 ± 25.732		24.079 ± 21.545		6.029 ± 8.282		18.684 ± 17.811
III	25	42.4	28.84 ± 26.226		22.42 ± 22.71		2.868 ± 3.554		17.808 ± 19.248
IV	5	8.5	43.0 ± 38.987		34.4 ± 28.369		12.8 ± 28.066		10.94 ± 8.029
Colectomy				0.93		0.994		0.59		0.634
Right side	39	63.9	31.026 ± 26.462		22.474 ± 20.703		4.954 ± 10.326		18.069 ± 17.915
Left side	15	24.6	33.4 ± 28.306		23.567 ± 24.243		3.777 ± 5.634		16.033 ± 14.502
Total	7	11.5	33.571 ± 24.616		23.071 ± 23.816		8.0 ± 12.342		14.957 ± 20.454
Lymphocytes				0.636		0.593		0.544		0.417
Low	30	50.0	33.867 ± 28.295		36.25 ± 27.89		6.555 ± 11.843		16.033 ± 16.511
Moderate	18	30.0	26.944 ± 22.37		22.567 ± 23.103		2.339 ± 3.0		15.261 ± 15.029
Intense	12	20.0	36.25 ± 27.89		21.306 ± 20.569		5.583 ± 9.758		24.1 ± 21.661
Recurrence				0.499		0.719		0.388		0.941
No	49	87.5	33.98 ± 26.279		24.337 ± 22.125		5.637 ± 10.421		17.518 ± 16.553
Yes	7	12.5	30.714 ± 31.282		22.929 ± 22.606		2.507 ± 3.446		22.286 ± 25.063
Clinical outcome				0.252		0.429		0.208		0.167
Disease-free	49	80.3	30.261 ± 25.51		23.378 ± 21.317		4.578 ± 6.32		10.392 ± 9.601
Death	12	19.7	27.917 ± 30.261		20.5 ± 23.434		6.796 ± 17.917		18.882 ± 18.273

Due to the high degree of heterogeneity in the data, machine learning methods were employed to assess possible patterns in macrophage density. From these data, two distinct groups were identified: (1) Group 1 with higher macrophage density; and (2) Group 2 with lower macrophage density per field (Table 6). Additionally, a statistically significant difference was found between these groups concerning macrophage density as identified by the CD68 marker (P < 0.01). Group 1 had an average of 64 macrophages per field, while Group 2 had an average of 17 macrophages per field. Using the CD206 marker, a statistically significant difference was also found between the groups (P < 0.01), with Group 1 showing an average of 38 macrophages per field, and Group 2 showing an average of 7 macrophages per field. Lastly, the CD163 marker showed an average of 50 macrophages per field in Group 1 and 15 macrophages per field in Group 2 (P < 0.01).

Table 6 Variable analysis regarding grouping results (n = 61).

	Group 1, n = 19	Group 2, n = 42	P value
CD 68	64.737 ± 17.048	17.048 ± 12.61	< 0.01
CD163	50.053 ± 10.488	15.29 ± 8.897	< 0.01
iNOS	9.474 ± 15.19	2.996 ± 4.389	0.064
CD 206	38.211 ± 14.635	7.712 ± 6.462	< 0.01
Age (complete years)	64.947 ± 11.038	66.405 ± 12.655	0.548
Time to death (days)	161.333 ± 238.123	421.875 ± 391.745	0.279

Quantitative variables expressed in values of mean ± SD.

The clustering analysis results are supported by the correlations between the measurements obtained using the different markers to identify macrophages (Table 7). When evaluating these two clusters, none of the studied variables, such as the presence of metalloproteinase mutations, staging, tumor location, or survival, showed a statistically significant relationship (Table 8).

Table 7 Analysis of correlation between variables of macrophage measurement (n = 61).

Comparison	Rho	P value
CD68c CD163	0.897	< 0.001a
CD68c iNOS	0.556	< 0.001a
CD68c CD206	0.783	< 0.001a
CD68c age	-0.155	0.233
CD163c iNOS	0.466	< 0.001a
CD163c CD206	0.796	< 0.001a
CD163c age	-0.156	0.229
iNOSc CD206	0.383	0.002b
iNOSc age	-0.206	0.112
CD206c age	-0.233	0.07

^aP < 0.001.

^bP < 0.01.

^cP < 0.05.

Table 8 Analysis of variables regarding grouping results (n = 61).

	Total		Group 1, n = 19		Group 2, n = 42		P value
	n	%	n	%	n	%
Instability							0.785a
No mutation	33	52.5	11	57.9	22	52.4
With mutation	28	47.5	8	42.1	20	47.6
Age (year)							1.0a
≤ 60	15	24.6	5	26.3	10	23.8
> 60	46	75.4	14	73.7	32	76.2
Sex							0.14a
Female	43	70.5	16	84.2	27	64.3
Male	18	29.5	3	15.8	15	35.71
Location							0.76a
Right colon	44	72.1	13	68.4	31	73.8
Left colon	17	27.9	6	31.6	11	26.2
Staging							0.96
I	10	16.9	3	16.7	7	17.1
II	19	32.2	6	33.3	13	31.7
III	25	42.4	7	38.9	18	43.9
IV	5	8.5	2	11.1	3	7.3
Colectomy							0.592
Right side	39	63.9	13	68.4	26	61.9
Left side	15	24.6	5	26.3	10	23.8
Total	7	11.5	1	5.3	6	14.3
Lymphocytes							0.91
Low	30	50.0	18	54.6	12	44.4
Moderate	18	30.0	7	21.2	11	40.7
Intense	12	20.0	8	24.2	4	14.9
Recurrence							0.679a
No	49	87.5	16	84.2	33	78.5
Yes	7	12.5	3	15.8	4	9.5
Clinical Outcome							0.737a
Disease-free	50	82.0	16	84.2	34	81.0
Death	11	18.0	3	15.8	8	19.0

^aP values obtained via Fisher's exact test.

DISCUSSION

Macrophages are essential for maintaining tissue integrity, hemostasis, and repair, performing specialized functions that vary by organ and gene expression[21]. They originate from fetal hematopoiesis, first in the yolk sac and then in the liver, and integrate into tissues. In the colon, embryonic macrophages are replaced by monocyte-derived macrophages formed in the bone marrow postnatally[22]. Macrophages play critical roles in hemostasis, inflammation resolution, angiogenesis, and the clearance of apoptotic cells, exhibiting diverse functions influenced by tissue type and pathological conditions. This diversity is regulated by epigenetic responses related to interferon and metabolism[10]. Macrophages secrete inflammatory mediators, such as TNF-α, which can either protect against pathogens (M1) or lead to tissue damage; thus, regulation towards an anti-inflammatory phenotype (M2) is vital for healing[23].

TME consists of various non-tumor cells, including fibroblasts, endothelial cells, and immune cells, along with extracellular components such as chemokines and cytokines. Investigating macrophages is crucial due to their polarized functions[24]. Generally categorized as M1 (pro-inflammatory) or M2 (anti-inflammatory), M2 macrophages include TAMs, which are often linked to poor prognoses[9]. TAMs promote inflammation, facilitate immune evasion, and enhance tumor progression[6]. Macrophage regulation is influenced by tumor cell properties and TME characteristics, prompting increased interest in targeting TAMs for immunotherapy[10].

In this study, CD68 was used to identify macrophages, with iNOS marking M1 and CD163 and CD206 marking M2 macrophages. M1 macrophages are identified by high TNF-α and iNOS expression, while M2 macrophages show elevated arginase 1, IL-10, CD163, CD204, and CD206[25]. The suitability of CD163 as an M2 marker is debated, leading to the decision to use both CD163 and CD206 in this evaluation[26-30].

Multiple methodologies exist for assessing macrophages in tumor tissues. This study utilized the PD-L1 IHC 22C3 pharmDx^© protocol, originally designed for lung cancer, to evaluate macrophages in CRC samples[31]. Two pathologists identified the area of highest macrophage concentration using CD68 staining and calculated the percentage of macrophages marked by CD163, CD206, and iNOS through the Tumor Proportion Score.

Forssell et al[32] linked elevated macrophage levels at tumor margins to improved prognosis, although our study did not consider macrophage localization due to sample size constraints. Furthermore, we classified tumors based on metalloproteinase mutations, noting that these mutations are often linked to earlier disease stages, a relationship corroborated by previous studies[10,33,34]. Tumors with metalloproteinase mutations also displayed increased survival rates[14], as observed in our findings (P = 0.016). It is important to note that a correlation between overall survival and disease-free time was demonstrated previously, which aimed to identify a group that would benefit from adjuvant chemotherapy in stage II colon cancer patients[35].

In our sample, only two patients with repair protein mutations succumbed to CRC, while the remaining three deaths in this group involved patients in stages I and II of cancer, resulting from COVID-19. Other study variables, including age, sex, tumor location, type of surgery performed, TILs, and macrophage count, did not show significant associations with survival rates. It is important to note that a correlation between overall survival and disease-free time was demonstrated by Feng et al[35], which aimed to identify a group that would benefit from adjuvant chemotherapy in stage II colon cancer patients[35].

No significant differences in macrophage density were found between patients with and without metalloproteinase mutations, aligning with Edin et al[36], while Bauer et al[37] reported greater macrophage infiltration in MSI cases without examining prognostic factors.

Regarding the number of macrophages obtained from immunohistochemical analysis, our results depict a positive correlation between the quantity of macrophages identified using the markers CD68, CD163, CD206, and iNOS. These findings partially align with the study conducted by Koelzer et al[34] in 2016, which found a correlation between CD68 and CD163 but not between CD68 and iNOS[34].

Although TAMs are frequently presented as factors of poor prognosis[38,39], this relationship remains controversial in the context of CRC, as demonstrated by the meta-analysis conducted by Zhang et al[40], which compared the presence of macrophages with disease prognosis[40].

Such uncertainty persists in light of conflicting findings within the literature. For instance, Khorana et al[41] demonstrated that TAM infiltration and vascular endothelial growth factor are associated with a better prognosis[41]. Conversely, Shabo et al[27] observed higher recurrence rates and lower survival for rectal tumors with increased CD163 expression. Additionally, TAMs have potential as predictors of hepatic metastases, with a decrease in M1 macrophages and an increase in M2 macrophages observed in stage IIIB patients, who are more likely to develop metastases[42].

Another meta-analysis exploring 24 studies indicated that a high rate of TAM infiltration was associated with favorable five-year survival rates for CRC cases (hazard ratio = 0.69, 95% confidence interval = 0.55–0.87), albeit with high heterogeneity[19]. Recent work by Elomaa et al[43] further demonstrated that PD-L1⁺ macrophages—predominantly M1-polarized and closely associated with tumor cells—correlate with improved clinical outcomes. Moreover, patients without deficiencies in repair proteins exhibited better prognoses, suggesting that the molecular context, including potential protective effects conferred by metalloproteinase mutations, may influence macrophage behavior[19]. In our study, although overall macrophage quantity did not show significant differences across prognostic variables, this observation mirrors findings by Li et al[19], where 16 of 27 articles reported similar inconsistencies. In contrast, emerging evidence from single-cell analyses has identified distinct macrophage subpopulations, such as EEF1G⁺ macrophages, that are closely linked to metastatic potential and immunometabolic remodeling[44]. Together, these studies highlight that variations in macrophage subpopulation characterization and underlying molecular mechanisms may account for discrepancies in prognostic associations.

The cluster analysis in our study identified two well-defined groups based on macrophage quantity; however, cross-referencing these clusters with prognostic outcomes did not reveal significant differences. This contrasts with findings from Koelzer et al[34], where specific macrophage phenotypes were associated with patient prognosis. In support of this, Donadon et al[45] showed that macrophage morphology—delineating small and large subsets—correlates with distinct functional states and survival outcomes, emphasizing that quantitative assessment alone may be insufficient. Additionally, Hua et al[44] demonstrated that certain macrophage subsets are integral to establishing an immunosuppressive microenvironment that drives metastasis. These studies suggest that a nuanced evaluation of macrophage polarization—including spatial distribution, morphological diversity, and molecular signatures—is essential for unraveling the mechanisms underlying their potential protective effects, particularly those possibly mediated by metalloproteinase mutations.

Overall, discrepancies have been observed in the literature regarding the quantity of macrophages. The polarization of macrophages appears to be a crucial factor in tumor progression and, thus, holds potential for prognostic prediction. Furthermore, it is important to consider the site of infiltration, whether in the stroma or at the tumor front. Although the results of this study did not present statistically significant inferences, the findings pave the way for more in-depth investigations. Identifying patterns in macrophage quantity from a prognostic perspective, as well as the possibility of manipulating these cells through immunotherapy, represents key research directions aimed at advancing cancer medicine.

This study is subject to several limitations. The small sample size of patients with the identified mutation, while including all available cases at our service, remains relatively small for studies investigating complex tumor biology. This may limit the representativeness and statistical power of the results, potentially affecting the ability to fully elucidate the relationship between macrophage polarization and various prognostic factors. Additionally, the evaluation of macrophage polarization was not performed in detail, which could obscure important distinctions between M1 and M2 macrophages and their impact on tumor behavior. Furthermore, the retrospective nature of the study may introduce biases related to data collection and interpretation. Given these inherent limitations, future multicenter studies with larger, well-defined cohorts and comprehensive analyses of macrophage polarization and localization are essential to strengthen the statistical analysis and enhance the generalizability of the findings.

CONCLUSION

In this study, no differences were found in the presence of polarized macrophages, whether M1 or M2, between groups of patients with or without mutations in DNA repair genes. Additionally, no statistically significant differences were identified when evaluating predictive factors related to CRC and survival. These results may be attributed to the small sample size of patients, as well as the limited observation period.

ACKNOWLEDGEMENTS

The authors would like to extend our sincere gratitude to the Centro de Patologia Médica for their invaluable assistance with the immunohistochemistry analyses. We are also grateful to the Postgraduate Program at the Universidade de Caxias do Sul for the support provided through the PhD program, and to Hospital Geral de Caxias do Sul for their ongoing support in facilitating this research. Their contributions were instrumental in the completion of this study, which is Brambilla B PhD thesis.