Abstract
Background
Caudal-type homeobox 2 (CDX2) and special AT-rich sequence-binding protein 2 (SATB2) are transcription factors playing important roles in intestinal homeostasis and participating in the regulation of intestinal inflammation. In colorectal cancer (CRC), reduced expression levels of CDX2 and SATB2 have been associated with poor differentiation and worse survival. However, their prognostic significance still needs further clarification, and the associations between CDX2 and SATB2 and immune cell infiltration into the CRC microenvironment are largely unknown.
Methods
We analyzed CDX2 and SATB2 expression in two large cohorts of stages I–IV CRC patients (N = 2302) and analyzed their associations with clinicopathologic parameters, the density of local immune cells (determined with three multiplex immunohistochemistry panels and conventional immunohistochemistry), and survival.
Results
In mismatch repair-proficient tumors, reduced CDX2 and SATB2 expression were associated with higher densities of immature monocytic cells, macrophages, and M2-like macrophages. Low expression of CDX2 was associated with shorter cancer-specific survival independent of conventional prognostic parameters in both cohorts. In the larger cohort, adjusted hazard ratio (HR) for negative (vs. high) CDX2 expression was 3.62 (95% CI 2.08–6.31, ptrend < 0.0001), and adjusted HR for negative (vs. high) SATB2 level was 1.61 (95% CI 0.97–2.67, ptrend = 0.002).
Conclusion
This study indicates that reduced CDX2 and SATB2 expression levels are associated with myeloid cell infiltration in the CRC microenvironment and represent markers for poor outcome. These findings highlight the potential of CDX2 and SATB2 as biomarkers for classifying CRC patients and support their role in regulating the tumor microenvironment.
Supplementary Information
The online version contains supplementary material available at 10.1007/s00262-025-03964-x.
Keywords: Colorectal cancer, CDX2, SATB2, Immune cells, Prognosis
Introduction
Caudal-type homeobox 2 (CDX2) and special AT-rich sequence-binding protein 2 (SATB2) are transcription factors regulating gene expression in the intestine. CDX2 is normally expressed within the nuclei of intestinal epithelial cells and plays a crucial role in the development and maintenance of the intestinal epithelium [1]. SATB2 is highly expressed in the lower gastrointestinal epithelium and specific neurons in brain and is particularly important in the development of neural, craniofacial, and osteoblastic structures [2]. In the colon, SATB2 is required for maintaining colonic identity of stem cells and differentiated cells [3].
CDX2 and SATB2 have been linked to colorectal cancer (CRC) progression. They both are used as immunohistochemical markers for colorectal origin in tumors and are advantageous in distinguishing metastatic colorectal adenocarcinoma from other adenocarcinomas [4–6]. However, their expression is reduced or lost in a subset of CRCs, most often poorly differentiated CRCs [7, 8]. Reduced expression has been found to be associated with worse survival and several unfavorable prognostic factors, including advanced stage, higher tumor grade, and BRAF mutation [8–14]. Decreased CDX2 and SATB2 expression is also associated with mismatch repair (MMR) protein deficiency in CRC [15, 16].
High immune cell infiltration generally predicts better outcome in CRC [17]. The immune cell infiltrate can be quantified using the Immunoscore, a scoring system based on the density of CD3+ and CD8+ cell populations both at the tumor center and at the invasive margin [18]. A high Immunoscore predicts longer survival, and the prognostic power of the Immunoscore has been validated by a worldwide consortium [19]. In recent years, CDX2 and SATB2 have been increasingly linked to intestinal inflammation. CDX2 has been shown to regulate inflammasome activity through expression of the NLRP3 suppressor TRIM31 [20]. In the mouse colon epithelium, loss of Cdx function leads to elevated infiltration of macrophages and expression of proinflammatory cytokines, including tumor necrosis factor-alpha, interleukin-1 beta, and interleukin-6 [20–22]. Also, the expression of CDX2 and SATB2 has been shown to be decreased in the epithelium of patients with inflammatory bowel disease, and Satb2-deficient mice are more prone to the development of colitis and colitis-associated cancer compared with control mice [23–25]. However, the potential role of CDX2 and SATB2 in the inflammatory response in human CRC is largely unknown.
The primary aim of this study was to evaluate the associations between CDX2 and SATB2 expression and tumor-infiltrating immune cells in CRC. The secondary aim was to confirm the prognostic value of CDX2 and SATB2 in CRC.
Materials and methods
Patients
This study was based on two independent cohorts. Cohort 1 was retrospectively collected in Central Finland Central Hospital in Jyväskylä. It included 1343 patients who had undergone tumor resection during 2000–2015 [26]. Cohort 2 was prospectively collected of 1011 newly diagnosed CRC patients operated in the Oulu University Hospital between 2006 and 2020, who had signed an informed consent to participate and were eligible to the study [27]. Patients who had received preoperative radiotherapy or chemoradiotherapy (Cohort 1: N = 243; Cohort 2: N = 235) were excluded from immune cell analyses, and patients who died within 30 days of initial surgery (Cohort 1: N = 40; Cohort 2: N = 5) were excluded from survival analysis.
Immunohistochemistry and image analysis
Immunohistochemistry was conducted on tissue microarrays (TMAs) that were constructed from formalin-fixed paraffin-embedded tissue samples and were designed to contain four cores (diameter 1 mm) per case, two from the center of the tumor (CT) and two from the invasive margin (IM) [26, 27]. MMR enzyme screening status for MLH1, MSH2, MSH6, and PMS2, and BRAF V600E mutation status had been previously examined with immunohistochemistry [26, 27]. The immune cell analyses for Cohort 1 were based on multiplex immunohistochemistry and quantitative image analyses. The multiplex immunohistochemistry staining protocol, reagents, and the primary antibodies along with their dilutions and epitope retrieval conditions are demonstrated in Supplementary Fig. S1. Analyses of tumor-infiltrating immune cells were conducted with supervised machine learning approaches built-in QuPath [26, 28, 29]. The cell types investigated in this study included CD3+ T cells, CD20+CD79A+ B cells, CD20-CD79A+ plasma cells, M1-like and M2-like macrophages, CD14+HLA-DR+ mature monocytic cells, CD14+HLA-DR- immature monocytic cells, CD66B+ granulocytes, and tryptase+ mast cells. Macrophages were categorized according to their polarization by calculating a polarization index for each macrophage [28]. Immune cell score was calculated according to the principles of Immunoscore® [26].
An automated immunostainer (BOND-III, Leica Biosystems, Germany) and the BOND Refine detection kit (Leica Biosystems) were used for the immunohistochemical staining of CDX2 (Epredia RM-2116, clone EPR2764Y, 1:500) and SATB2 (Cell Marque 384R-16, clone EP281, 1:80). BOND Epitope Retrieval Solution 2 (Leica Biosystems, AR9640) was used in the pretreatment (100 °C, 30 min). The slides were scanned with NanoZoomer XR (Hamamatsu Photonics, Hamamatsu City, Japan) scanner using 20× objective magnification. Scanned images were analyzed with QuPath (version 0.4.4). In QuPath, cell detection function was used to detect cells, add intensity features function was used to calculate Haralick’s texture features, and add smoothed features function was used to calculate smoothed features (within 20 µm). An object classifier was then trained to recognize tumor and stromal cells. The intensity of DAB (3,3’-Diaminobenzidine) chromogen was measured for each tumor cell with QuPath, and the cell data were processed using RStudio and R statistical programming (version 4.3.1, R core Team). For each cell, DAB intensity was classified into four categories. For CDX2, DAB < 0.25 intensity units indicated negative, DAB 0.25–0.55 weakly positive, DAB 0.55–0.85 moderately positive, and DAB > 0.85 strongly positive staining. For SATB2, DAB < 0.2 intensity units indicated negative, DAB 0.2–0.4 weakly positive, DAB 0.4–0.6 moderately positive, and DAB > 0.6 strongly positive staining. The mean histoscore for each core was calculated using the formula: 1 × percentage of tumor cells showing weak staining + 2 × percentage of tumor cells showing moderate staining + 3 × percentage of tumors cells showing intensive staining. For each tumor, a single histoscore was calculated as the average of histoscores from all individual cores, and tumors were categorized into negative (histoscore < 5), low (histoscore ≥ 5 and < 100), and high expression (histoscore ≥ 100) [30]. Examples of TMA cores with cell classification and different CDX2 and SATB2 expression groups are presented in Fig. 1.
Fig. 1.
CDX2 and SATB2 immunohistochemistry. A Cell detection and classification of one TMA core using QuPath. In the image analysis result image, red color represents cancer cells and yellow other cells. B Examples of TMA cores with high, low, and negative CDX2 and SATB2 expression
Statistical analyses
The statistical analyses were conducted using IBM SPSS Statistics for Windows version 29.0.0.0 (IBM Corporation, Armonk, NY, USA). Chi-square test was used to characterize the relationship between categorical variables. Correlations between two continuous variables were determined using the Spearman’s rank correlation test. The 2D visualization was created with Cytoscape software platform [31], utilizing the Prefuse force directed algorithm weighted by the statistical significances of the correlations between individual variables. Survival rates were visualized by the Kaplan–Meier method, and the statistical significance was tested with the log-rank test. Cancer-specific survival (CSS; time from operation to CRC death) was the primary endpoint. Hazard ratios (HRs) with 95% confidence interval (CI) were calculated using Cox proportional hazard models. Multivariable Cox regression models included the following pre-determined covariates: age (< 65, 65–75, and > 75), sex (male and female), stage (I, II, III, and IV), MMR enzyme status (proficient and deficient), BRAF V600E mutation status (wild-type and mutant), tumor location (proximal colon, distal colon, and rectum), year of operation (Cohort 1: 2000–2005, 2006–2010, and 2011–2015; Cohort 2: 2006-2010, 2011–2015, and 2016-2020), lymphatic or venous invasion (no and yes), grade (low-grade and high-grade), and preoperative radiotherapy or chemoradiotherapy (no and yes). A two-tailed p < 0.01 was considered statistically significant.
Results
Distribution of CDX2 and SATB2 staining and association with clinicopathological characteristics
The expression of CDX2 and SATB2 was detected in variable proportions of tumor cell nuclei, while stromal cells were negative for both. The associations of CDX2 and SATB2 histoscore with tumor and patient characteristics are shown in Tables 1 and 2. CDX2 and SATB2 expression could be evaluated from 1329 patients in Cohort 1 and 973 patients in Cohort 2. Of them, 32 (2%) in Cohort 1 and 22 (2%) in Cohort 2 had loss of CDX2 expression, 95 (7%) and 50 (5%) low CDX2 expression and 1202 (90%) and 901 (93%) high CDX2 expression. SATB2 expression was negative in 55 (4%) tumors in Cohort 1 and 120 (12%) tumors in Cohort 2, low in 265 (20%) and 206 (21%) and high in 1009 (76%) and 647 (67%) tumors. In both cohorts, lower CDX2 and SATB2 expression levels were associated with proximal tumor location, high tumor grade, MMR deficiency, and BRAF V600E mutation (p < 0.001 for all). CDX2 and SATB2 expression groups were associated with one another (p < 0.001). In Cohort 1, low expression of CDX2 and SATB2 was also related to advanced stage (p = 0.004 and p < 0.001, respectively) and lymphovascular invasion (p = 0.008 and p = 0.002, respectively).
Table 1.
Associations between CDX2 expression and clinicopathological parameters in colorectal cancers
| Characteristic | Cohort 1 | Cohort 2 | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Total N | High | Low | Negative | p | Total N | High | Low | Negative | p | |
| All cases | 1329 (100%) | 1202 (90%) | 95 (7%) | 32 (2%) | 973 (100%) | 901 (93%) | 50 (5%) | 22 (2%) | ||
| Sex | ||||||||||
| Female | 618 (47%) | 548 (46%) | 49 (52%) | 21 (66%) | 0.048 | 425 (44%) | 382 (58%) | 29 (58%) | 14 (64%) | 0.016 |
| Male | 711 (54%) | 654 (54%) | 46 (48%) | 11 (34%) | 548 (56%) | 519 (42%) | 21 (42%) | 8 (36%) | ||
| Age (years) | ||||||||||
| <65 | 395 (30%) | 364 (30%) | 22 (23%) | 9 (28%) | 0.243 | 330 (34%) | 303 (34%) | 20 (40%) | 7 (32%) | 0.769 |
| 65–75 | 454 (34%) | 415 (35%) | 31 (33%) | 8 (25%) | 356 (37%) | 334 (37%) | 14 (28%) | 8 (36%) | ||
| >75 | 480 (36%) | 423 (35%) | 42 (44%) | 15 (47%) | 287 (29%) | 264 (29%) | 16 (32%) | 7 (32%) | ||
| Tumor location | ||||||||||
| Proximal colon | 532 (40%) | 445 (37%) | 63 (66%) | 24 (75%) | <0.001 | 321 (33%) | 279 (31%) | 26 (52%) | 16 (73%) | <0.001 |
| Distal colon | 399 (30%) | 387 (32%) | 8 (8%) | 4 (13%) | 205 (21%) | 200 (22%) | 3 (6%) | 2 (9%) | ||
| Rectum | 398 (30%) | 370 (31%) | 24 (25%) | 4 (13%) | 447 (46%) | 422 (47%) | 21 (42%) | 4 (18%) | ||
| Preoperative RT/CRT | ||||||||||
|
No Yes |
1091 (82%) 239 (18%) |
984 (82%) 218 (18%) |
77 (81%) 18 (19%) |
30 (91%) 3 (9%) |
0.397 |
759 (78%) 214 (22%) |
701 (78%) 200 (22%) |
39 (78%) 11 (22%) |
19 (86%) 3 (14%) |
0.632 |
| AJCC disease stage | ||||||||||
| I | 246 (19%) | 238 (20%) | 7 (7%) | 0 (0.0%) | 0.004 | 213 (22%) | 202 (22%) | 8 (16%) | 3 (14%) | 0.124 |
| II | 485 (37%) | 433 (36%) | 40 (42%) | 13 (41%) | 306 (31%) | 280 (31%) | 18 (36%) | 8 (36%) | ||
| III | 425 (32%) | 382 (32%) | 30 (32%) | 13 (41%) | 346 (36%) | 325 (36%) | 16 (32%) | 5 (23%) | ||
| IV | 173 (13%) | 149 (12%) | 18 (19%) | 6 (19%) | 108 (11%) | 94 (10%) | 8 (16%) | 6 (27%) | ||
| Tumor grade | ||||||||||
| Low-grade | 1106 (83%) | 1051 (87%) | 45 (47%) | 10 (31%) | <0.001 | 829 (85%) | 793 (88%) | 29 (58%) | 7 (32%) | <0.001 |
| High-grade | 223 (17%) | 151 (13%) | 50 (53%) | 22 (69%) | 144 (15%) | 108 (12%) | 21 (42%) | 15 (68%) | ||
| Lymphovascular invasion | ||||||||||
| No | 1043 (78%) | 957 (80%) | 64 (67%) | 22 (69%) | 0.008 | 509 (52%) | 478 (53%) | 24 (48%) | 7 (32%) | 0.118 |
| Yes | 286 (22%) | 245 (20%) | 31 (33%) | 10 (31%) | 464 (48%) | 423 (47%) | 26 (52%) | 15 (68%) | ||
| MMR status | ||||||||||
| pMMR | 1160 (87%) | 1083 (90%) | 63 (66%) | 14 (44%) | <0.001 | 847 (87%) | 809 (90%) | 26 (52%) | 12 (55%) | <0.001 |
| dMMR | 169 (13%) | 119 (10%) | 32 (34%) | 18 (56%) | 126 (13%) | 92 (10%) | 24 (48%) | 10 (45%) | ||
| BRAF status | ||||||||||
| Wild-type | 1140 (86%) | 1064 (89%) | 61 (64%) | 15 (47%) | <0.001 | 862 (89%) | 825 (92%) | 27 (54%) | 10 (45%) | <0.001 |
| Mutant | 187 (4%) | 136 (11%) | 34 (36%) | 17 (53%) | 111 (11%) | 76 (8%) | 23 (46%) | 12 (55%) | ||
| SATB2 subgroups | ||||||||||
| High | 1009 (76%) | 981 (82%) | 24 (25%) | 4 (13%) | <0.001 | 647 (67%) | 637 (71%) | 9 (18%) | 1 (5%) | <0.001 |
| Low | 265 (20%) | 194 (16%) | 53 (56%) | 18 (56%) | 206 (21%) | 184 (20%) | 20 (40%) | 2 (9%) | ||
| Negative | 55 (4%) | 27 (2%) | 18 (19%) | 10 (31%) | 120 (12%) | 80 (9%) | 21 (42%) | 19 (86%) | ||
Abbreviations: AJCC, The American Joint Committee on Cancer; RT/CRT, radiotherapy/chemoradiotherapy; MMR, mismatch repair; pMMR, mismatch repair proficient; and dMMR, mismatch repair deficient
Table 2.
Associations between SATB2 expression and clinicopathological parameters in colorectal cancers
| Characteristic | Cohort 1 | Cohort 2 | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Total N | High | Low | Negative | p | Total N | High | Low | Negative | p | |
| All cases | 1329 (100%) | 1009 (76%) | 265 (20%) | 55 (4%) | 973 (100%) | 647 (67%) | 206 (21%) | 120 (12%) | ||
| Sex | ||||||||||
| Female | 618 (47%) | 451 (%) | 131 (%) | 36 (%) | 0.006 | 425 (44%) | 272 (42%) | 99 (48%) | 54 (45%) | 0.302 |
| Male | 711 (54%) | 558 (%) | 134 (%) | 19 (%) | 548 (56%) | 375 (58%) | 107 (52%) | 66 (55%) | ||
| Age (years) | ||||||||||
| <65 | 395 (30%) | 299 (30%) | 79 (30%) | 17 (31%) | 0.497 | 330 (34%) | 216 (33%) | 73 (35%) | 41 (34%) | 0.827 |
| 65–75 | 454 (34%) | 357 (35%) | 81 (31%) | 16 (29%) | 356 (37%) | 240 (37%) | 69 (34%) | 47 (39%) | ||
| >75 | 480 (36%) | 353 (35%) | 105 (40%) | 22 (40%) | 287 (29%) | 191 (30%) | 64 (31%) | 32 (27%) | ||
| Tumor location | ||||||||||
| Proximal colon | 532 (40%) | 352 (35%) | 139 (53%) | 41 (75%) | <0.001 | 321 (33%) | 179 (28%) | 81 (39%) | 61 (51%) | <0.001 |
| Distal colon | 399 (30%) | 338 (34%) | 55 (21%) | 6 (11%) | 205 (21%) | 146 (23%) | 43 (21%) | 16 (13%) | ||
| Rectum | 398 (30%) | 319 (32%) | 71 (27%) | 8 (15%) | 447 (46%) | 322 (50%) | 82 (40%) | 43 (36%) | ||
| Preoperative RT/CRT | ||||||||||
|
No Yes |
1091 (82%) 239 (18%) |
822 (82%) 187 (18%) |
217 (82%) 49 (18%) |
52 (95%) 3 (5%) |
759 (78%) 214 (22%) |
496 (77%) 151 (23%) |
164 (80%) 42 (20%) |
99 (83%) 21 (17%) |
0.301 | |
| AJCC disease stage | ||||||||||
| I | 246 (19%) | 212 (21%) | 27 (10%) | 6 (11%) | <0.001 | 213 (22%) | 155 (24%) | 40 (19%) | 18 (15%) | 0.156 |
| II | 485 (37%) | 375 (37%) | 90 (34%) | 21 (38%) | 306 (31%) | 201 (31%) | 64 (31%) | 41 (34%) | ||
| III | 425 (32%) | 306 (30%) | 102 (39%) | 17 (31%) | 346 (36%) | 229 (35%) | 75 (36%) | 42 (35%) | ||
| IV | 173 (13%) | 116 (12%) | 46 (17%) | 11 (20%) | 108 (11%) | 62 (10%) | 27 (13%) | 19 (16%) | ||
| Tumor grade | ||||||||||
| Low-grade | 1106 (83%) | 896 (89%) | 181 (68%) | 29 (53%) | <0.001 | 829 (85%) | 594 (92%) | 153 (74%) | 82 (68%) | <0.001 |
| High-grade | 223 (17%) | 113 (11%) | 84 (32%) | 26 (47%) | 144 (15%) | 53 (8%) | 53 (26%) | 38 (32%) | ||
| Lymphovascular invasion | ||||||||||
| No | 1043 (79%) | 813 (81%) | 187 (71%) | 43 (78%) | 0.002 | 509 (52%) | 356 (55%) | 102 (50%) | 51 (43%) | 0.028 |
| Yes | 286 (22%) | 196 (19%) | 78 (29%) | 12 (22%) | 464 (48%) | 291 (45%) | 104 (51%) | 69 (58%) | ||
| MMR status | ||||||||||
| pMMR | 1160 (87%) | 913 (90%) | 209 (79%) | 38 (69%) | <0.001 | 847 (87%) | 597 (92%) | 164 (80%) | 86 (72%) | <0.001 |
| dMMR | 169 (13%) | 96 (10%) | 56 (21%) | 17 (31%) | 126 (13%) | 50 (8%) | 42 (20%) | 34 (28%) | ||
| BRAF status | ||||||||||
| Wild-type | 1140 (86%) | 911 (91%) | 189 (71%) | 40 (73%) | <0.001 | 862 (89%) | 610 (94%) | 163 (79%) | 89 (74%) | <0.001 |
| Mutant | 187 (4%) | 96 (10%) | 76 (29%) | 15 (27%) | 111 (11%) | 37 (6%) | 43 (21%) | 31 (26%) | ||
Abbreviations: AJCC, The American Joint Committee on Cancer; RT/CRT, radiotherapy/chemoradiotherapy; MMR, mismatch repair; pMMR, mismatch repair proficient; and dMMR, mismatch repair deficient
Association with immune cell densities
Next, we investigated whether the expression levels of CDX2 and SATB2 are associated with immune cell infiltration patterns of the tumors. We utilized three multiplex immunohistochemistry panels in Cohort 1 and conducted separate analyses for MMR-proficient (pMMR) and MMR-deficient (dMMR) tumors, considering that both immune cell densities and SATB2 and CDX2 expression are associated with MMR status. Example multiplex immunohistochemistry images of CDX2 high/SATB2 low tumor and their corresponding cell maps are shown in Fig. 2, while images of CDX2 high/SATB2 high tumor and CDX2 negative/SATB2 negative tumor are presented in Supplementary Fig. S2. We found several weak correlations between CDX2 and SATB2 histoscore and the overall densities of different immune cells (Fig. 2, Supplementary Table S2). In pMMR tumors, lower CDX2 histoscore was associated with higher densities of macrophages (r = −0.226, p < 0.001), M2-like macrophages (r = −0.197, p < 0.001), mature monocytic cells (r = −0.135, p < 0.001), and immature monocytic cells (r = −0.168, p < 0.001). Lower SATB2 histoscore was associated with decreased densities of M1-like macrophages (r = 0.113, p < 0.001) and granulocytes (r = 0.121, p < 0.001) and increased densities of macrophages (r = −0.096, p = 0.004), M2-like macrophages (r = −0.164, p < 0.001), and immature monocytic cells (r = −0.130, p < 0.001). In dMMR tumors, CDX2 expression showed no significant correlations with immune cells, while lower SATB2 expression was associated with decreased densities of M1-like macrophages (r = 0.231, p = 0.003) and mast cells (r = 0.223, p = 0.005). We further analyzed the correlations between CDX2/SATB2 expression and immune cell densities separately at the tumor center and at the invasive margin (Supplementary Table S1), and the findings were largely similar to those for overall immune cell densities. To clarify the effects of the interaction between CDX2 loss and SATB2 loss on immune cell infiltration, we combined CDX2 expression status (high vs. low/negative) with SATB2 expression status (high vs. low/negative). CRCs with both CDX2 and SATB2 low/negative status had significantly lower M1-like macrophage and mast cell densities (Supplementary Fig. S3).
Fig. 2.
Immune cell analyses. A–F Multiplex immunohistochemistry images for visualizing three staining panels (A, C, E) and their corresponding cell phenotyping maps (B, D, F) in one tumor core and a close-up view. Scale bars are 100 µm. G, H Two-dimensional visualization of the relationships between CDX2 and SATB2 expression levels and immune cell densities in mismatch repair-proficient (G) and -deficient (H) tumors in Cohort 1. The edges (connecting lines) depict the associations between variables (only those with p < 0.01 are shown), and the edge length illustrates the significance of the association. The correlations between CDX2 and SATB2 and other variables are represented by green (positive correlation) and red (negative correlation) edges, with the label showing corresponding Spearman correlation coefficient for the correlation. The other associations are represented by gray edges
Survival analyses
Finally, we analyzed the prognostic value of CDX2 and SATB2 immunohistochemistry. There were 354 cancer deaths in Cohort 1 and 196 in Cohort 2. Kaplan-Meier plots showed that low CDX2 and SATB2 levels predicted shorter CSS compared to high levels in both cohorts (p < 0.001 for all, Fig. 3). The prognostic significance of CDX2 remained after adjusting for other confounding factors in multivariable Cox regression models in both cohorts (Table 3, Supplementary Table S2). The multivariable HR for negative (vs. high) CDX2 expression was 3.62 (95% CI 2.08–6.31, ptrend < 0.0001) in Cohort 1 and 3.55 (95% CI 1.72–7.31, ptrend < 0.0001) in Cohort 2. For SATB2, a trend of an association between low expression and worse survival was observed in multivariable Cox regression model in Cohort 1 (HR for negative vs. high SATB2 level 1.61, 95% CI 0.97–2.67, ptrend = 0.002) and a significant association in Cohort 2 (HR for negative vs. high SATB2 level 1.75, 95% CI 1.15–2.67, ptrend = 0.007) (Table 3, Supplementary Table S2). When we analyzed the survival according to combined CDX2/SATB2 expression category variable, patients with low/negative expression of both CDX2 and SATB2 showed the worst survival outcomes (Supplementary Table S3 and Supplementary Fig. S4).
Fig. 3.
Survival analyses. Kaplan–Meier analysis and log-rank test of colorectal cancer-specific survival according to CDX2 expression in Cohort 1 (A) and Cohort 2 (B) and SATB2 expression in Cohort 1 (C) and Cohort 2 (D)
Table 3.
Cancer-specific survival according to CDX2 and SATB2 expression in Cohort 1 and 2
| Cohort 1 | Cohort 2 | |||||||
|---|---|---|---|---|---|---|---|---|
| No. of cases | No. of events | Univariable HR (95% CI) | Multivariable HR (95% CI) | No. of cases | No. of events | Univariable HR (95% CI) | Multivariable HR (95% CI) | |
| CDX2 | ||||||||
| High | 1168 | 305 | 1 (referent) | 1 (referent) | 896 | 171 | 1 (referent) | 1 (referent) |
| Low | 91 | 33 | 1.68 (1.18–2.41) | 1.88 (1.23–2.81) | 50 | 14 | 1.51 (0.87–2.60) | 1.25 (0.67–2.32) |
| Negative | 30 | 16 | 2.90 (1.75–4.80) | 3.62 (2.08–6.31) | 22 | 11 | 3.86 (2.10–7.10) | 3.55 (1.72–7.31) |
| ptrend | <0.0001 | <0.0001 | <0.0001 | 0.003 | ||||
| SATB2 | ||||||||
| High | 981 | 241 | 1 (referent) | 1 (referent) | 645 | 108 | 1 (referent) | 1 (referent) |
| Low | 255 | 95 | 1.82 (1.44–2.31) | 1.47 (1.15–1.89) | 204 | 52 | 1.45 (1.04–2.02) | 1.32 (0.93–1.89) |
| Negative | 53 | 18 | 1.72 (1.07–2.78) | 1.61 (0.97–2.67) | 119 | 36 | 1.86 (1.28–2.72) | 1.75 (1.15–2.67) |
| ptrend | <0.0001 | 0.002 | 0.0005 | 0.007 | ||||
Multivariable Cox proportional hazards regression models were adjusted for sex, age (< 65, 65–75, and > 75), year of operation (Cohort 1: 2000–2005, 2006–2010, and 2011–2015, Cohort 2: 2006–2010, 2011–2015, and 2016–2020), tumor location (proximal colon, distal colon, and rectum), disease stage (I, II, III, and IV), tumor grade (well/moderately differentiated and poorly differentiated), lymphovascular invasion (negative and positive), MMR status (proficient and deficient), BRAF status (wild-type and mutant), and preoperative radiotherapy or chemoradiotherapy (no and yes)
ptrend values were calculated by using the three ordinal categories of CDX2 histoscore and SATB2 histoscore as continuous variables in univariable and multivariable Cox proportional hazard regression models
CI confidence interval and HR hazard ratio
To compare the prognostic value of CDX2 and SATB2 histoscore and immune cell score, we included these variables in the same multivariable Cox regression models (Supplementary Table S4). In these models, CDX2 histoscore had prognostic value in both cohorts (ptrend < 0.0001 in Cohort 1 and ptrend = 0.0005 in Cohort 2) independently of immune cell score (ptrend = 0.001 in Cohort 1 and ptrend = 0.002 in Cohort 2). Additionally, CDX2/SATB2 combination variable had prognostic value independent of immune cell score. The prognostic value of SATB2 histoscore remained in the multivariable model including immune cell score in Cohort 1 (ptrend = 0.0004), but not in Cohort 2 (ptrend = 0.019). Subgroup analyses were performed to further assess the prognostic significance of CDX2 and SATB2 expression status according to various clinicopathologic features (Supplementary Fig. S6), but no consistent findings of a significant difference according to any clinicopathologic characteristic were discovered across both cohorts.
Discussion
We evaluated the associations of CDX2 and SATB2 expression with immune cell infiltration and prognosis in two large CRC cohorts. We found that reduced CDX2 and SATB2 expression were associated with increased infiltration of myeloid cells, most prominently immature monocytic cells and M2-like macrophages, in pMMR tumors, while reduced SATB2 expression in dMMR tumors was associated with lower infiltration of M1-like macrophages and mast cells. Lower CDX2 expression predicted worse CSS in both cohorts independent of conventional prognostic parameters and immune cell infiltrates. These findings underscore the potential of CDX2 and SATB2 as biomarkers for stratifying CRC patients and highlight their significance in regulating the tumor microenvironment, which could inform the development of targeted therapies.
As CDX2 and SATB2 have been linked to intestinal inflammation, we hypothesized that loss of CDX2 and SATB2 expression might be associated with alterations in immune cell infiltration patterns in CRC. We found that both lower CDX2 and SATB2 expression were associated with increased macrophage, M2-like macrophage, and immature monocytic cell infiltrates in pMMR tumors. In addition, lower CDX2 expression was associated with increased mature monocytic cell densities and lower SATB expression with decreased M1-like macrophage and granulocyte densities. In dMMR tumors, lower SATB2 expression was associated with decreased densities of M1-like macrophages and mast cells. In combined analyses that included both CDX2 and SATB2 status, CRCs with low/negative expression of both CDX2 and SATB2 had significantly lower densities of M1-like macrophages and mast cells. Although the correlations were weak, these findings provide new insights into the potential role of CDX2 and SATB2 in CRC-associated inflammatory reactions. Previously, only a few studies have examined the associations between tumor-infiltrating immune cells and the expression of CDX2 or SATB2 in CRC. Chewchuk et al. [32] demonstrated that loss of Cdx function leads to attenuation of histocompatibility complex protein H2-T3, followed by a decrease in iCD8α cell number and degranulation, and an increase in macrophage infiltration in the murine intestinal epithelium. Our results align with this finding, as CDX2 expression showed a negative correlation with macrophage densities. Wang et al. have also reported a negative association between the expression of CDX2 and tumor-infiltrating macrophages [22]. However, they also showed a negative association between CDX2 expression and T-cell infiltrate, whereas we found no correlation between T-cell density and CDX2 or SATB2 expression status. Ni et al. [23] studied colitis model in intestinal epithelial-specific Satb2 knockout mice and control mice, and observed higher numbers of M1 macrophages in Satb2 knockout mice. In contrast with this, we found that lower SATB2 expression was associated with decreased M1-like macrophage infiltrate. Our results suggest that CDX2 and SATB2 expression could affect myeloid cell recruitment and macrophage balance (M1 vs. M2). While the underlying mechanisms remain hypothetical, it is possible that loss of CDX2/SATB2 expression leads to increased expression of cytokines that drive macrophage polarization toward M2 phenotype [21]. Additionally, the associations with altered immune cell densities might be related to disruption of the intestinal epithelial barrier, as CDX2 is essential for maintaining its integrity and function [1]. Compared to the previous studies, our analyses benefited from the use of multiplex immunohistochemistry that enabled more precise detection of different immune cell types. For example, there are no single specific markers for macrophage polarization states [33], and our method was based on calculating a polarization index based on the expression of four polarization markers.
Reduced CDX2 expression was an independent factor for worse CSS in multivariable analysis in both cohorts, whereas reduced SATB2 expression was independent prognostic factor in one cohort and showed a trend of an association in the other cohort. Previously, CDX2 loss has been identified as an independent poor prognostic factor in both non-metastatic and metastatic CRC [9, 11, 13, 22, 34, 35] although not all studies have confirmed its prognostic value [36, 37]. Independent prognostic significance of SATB2 expression has also been reported [8, 10, 35]. Schmitt et al. [8] previously reported a higher prognostic value for SATB2 loss compared to CDX2 loss, whereas we found that the prognostic impact of CDX2 loss was higher compared to SATB2 loss. Lee et al. [38] analyzed concomitant expression patterns of CDX2 and SATB2 as prognostic factors and found that concomitant loss of CDX2 and SATB2, but not isolated loss of CDX2 or SATB2, is a prognostic marker for stage III CRCs. The prognostic significance of CDX2 and SATB2 may be different in specific subgroups such as pMMR and dMMR tumors. Most studies have reported a prognostic value for CDX2 in pMMR tumors [12, 14, 34, 39], while Ma et al. [35] reported an impact only in patients with dMMR but not pMMR tumors. In our subgroup analyses, we found no statistically significant difference in the prognostic significance of CDX2 or SATB2 according to MMR status. Considering that CDX2 and SATB2 expression were associated with immune cell infiltration patterns, we also compared their prognostic value with the immune cell score. In this analysis, the prognostic significance of CDX2 and CDX2/SATB2 combination variable remained independent of immune cell score in both cohorts.
The frequency of CDX2 loss was 2% and SATB2 loss 4–12% in the two cohorts. These findings are in line with the previous studies reporting CDX2 loss in 1.3–5.9% and SATB2 loss in 6.6–28.8% of CRCs [8–10, 40, 41]. These results support the applicability of both transcription factors as markers of tumors with colorectal origin. When applying these markers in the diagnostics of tumors of uncertain origin, it needs to be noted that loss of CDX2 and SATB2 expression was associated with higher tumor grade, proximal tumor location, BRAF mutation, and dMMR, which also corresponds with the findings of the previous studies [7, 15, 36, 41, 42]. While CDX2 appears to have somewhat higher sensitivity for CRC than SATB2, the previous studies have indicated that SATB2 has higher specificity for CRC, as CDX2 is frequently also expressed by adenocarcinomas of the pancreas or the upper gastrointestinal tract among others [5, 43].
This study has some limitations. First, protein expression is heterogeneous in tumor tissue, and we used TMAs, which only represent small areas of the tumor. However, the previous studies have shown the applicability of TMAs in the analysis of immune cell infiltrates in CRC [44, 45]. Moreover, CDX2 and SATB2 expression generally showed little variation between cores sampled from different parts of the tumors. Second, data on cancer adjuvant treatments were not available, and we could not assess whether CDX2 or SATB2 expression would predict response to specific treatments. The strengths of this study include the utilization of two large CRC cohorts that contained extensive clinicopathologic data. The immune cell analyses were stratified by MMR status which is an important confounding factor that is associated with both immune cell densities and CDX2/SATB2 expression. Supervised machine learning technique facilitated quantitative image analysis and provided more accurate estimates of protein expression levels in comparison with visual evaluation.
In conclusion, reduced CDX2 and SATB2 expression are associated with increased infiltration of myeloid cells in pMMR tumors and exhibit unfavorable prognostic effect in CRC. These findings support the role of CDX2 and SATB2 in regulating the tumor microenvironment, as well as their potential as prognostic biomarkers in CRC.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
We acknowledge Dr. Sara Väyrynen for her contribution to the immune cell analyses, particularly for providing the image co-registration script that enabled multiplex immunohistochemical analyses. The study benefited from samples/data from Central Finland Biobank (Jyväskylä, Finland) and Northern Finland Biobank Borealis (Oulu, Finland).
Abbreviations
- AJCC
The American Joint Committee on Cancer
- CDX2
Caudal-type homeobox 2
- CI
Confidence interval
- CRC
Colorectal cancer
- CSS
Cancer-specific survival
- CT
Center of the tumor
- dMMR
Mismatch repair deficient
- HR
Hazard ratio
- IM
Invasive margin
- MMR
Mismatch repair
- pMMR
Mismatch repair proficient
- RT/CRT
Radiotherapy/chemoradiotherapy
- SATB2
Special AT-rich sequence-binding protein 2
- TMA
Tissue microarray
Author contributions
P. Sirniö and J. P. Väyrynen conceptualized the study. P. Sirniö, H. Elomaa, A. Tuomisto, V. K. Äijälä, H. Karjalainen, M. Kastinen, V. V. Tapiainen, O. Sirkiä, and J. P. Väyrynen curated the data. P. Sirniö and J. P. Väyrynen performed formal analysis. M. J. Mäkinen and J. P. Väyrynen acquired funding. All authors participated in investigation. P. Sirniö and J. P. Väyrynen contributed to the methodology. J. P. Väyrynen supervised the study. P. Sirniö and J. P. Väyrynen performed visualization. P. Sirniö and J. P. Väyrynen drafted the manuscript. All authors contributed to manuscript review and editing.
Funding
Open Access funding provided by University of Oulu (including Oulu University Hospital). The study was funded by Cancer Foundation Finland (59-5619 to JPV), Emil Aaltonen Foundation (220257P to JPV), Finnish Medical Foundation (6021 to JPV), Sigrid Jusélius Foundation (230229 to JPV), and Finnish State Research Funding (to MJM and JPV). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Data availability
The data generated and/or analyzed during this study are not publicly available. The sharing of data will require approval from relevant ethics committees and/or biobanks. Further information including the procedures to obtain and access data of Finnish Biobanks are described at https://finbb.fi/en/fingenious-service.
Declarations
Conflict of interest
T.T. Seppälä reports consultation fees from Tillots Pharma, Mehiläinen and Nouscom, being a co-owner and CEO of Healthfund Finland Ltd., and a position in the Clinical Advisory Board and as a minor shareholder of Lynsight Ltd. The other authors declare that they have no conflicts of interest.
Compliance with ethical standards
The study was performed according to the guidelines of the Declaration of Helsinki. For Cohort 1, the study was approved by the Regional medical research ethics committee of the Well-being services county of Central Finland (Dnro13U/2011, 1/2016, 8/2020, and 2/2023), the Finnish Medicines Agency (Fimea), and the Central Finland Biobank (BB23-0172). The need to obtain informed consent from the study patients was waived (FIMEA/2023/001573, 4/2023). For Cohort 2, the study was approved by the Regional medical research ethics committee of the Well-being services county of North Ostrobothnia (25/2002, 42/2005, 122/2009, and 37/2020), Fimea (FIMEA/2022/001941), and Biobank Borealis (BB-2017_1012). Patients signed an informed consent to participate.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Beck F (2004) The role of Cdx genes in the mammalian gut. Gut 53:1394–6. 10.1136/gut.2003.038240 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Huang X, Chen Q, Luo W, Pakvasa M, Zhang Y, Zheng L et al (2020) SATB2: A versatile transcriptional regulator of craniofacial and skeleton development, neurogenesis and tumorigenesis, and its applications in regenerative medicine. Genes Dis 9:95–107. 10.1016/J.GENDIS.2020.10.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Gu W, Wang H, Huang X, Kraiczy J, Singh PNP, Ng C et al (2022) SATB2 preserves colon stem cell identity and mediates ileum-colon conversion via enhancer remodeling. Cell Stem Cell 29:101-115.e10. 10.1016/J.STEM.2021.09.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Werling RW, Yaziji H, Bacchi CE, Gown AM (2003) CDX2, a highly sensitive and specific marker of adenocarcinomas of intestinal origin: an immunohistochemical survey of 476 primary and metastatic carcinomas. Am J Surg Pathol 27:303–10. 10.1097/00000478-200303000-00003 [DOI] [PubMed] [Google Scholar]
- 5.Zhang Y-J, Chen J-W, He X-S, Zhang H-Z, Ling Y-H, Wen J-H et al (2018) SATB2 is a promising biomarker for identifying a colorectal origin for liver metastatic adenocarcinomas. EBioMedicine 28:62–9. 10.1016/j.ebiom.2018.01.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Dragomir A, de Wit M, Johansson C, Uhlen M, Pontén F (2014) The role of SATB2 as a diagnostic marker for tumors of colorectal origin: results of a pathology-based clinical prospective study. Am J Clin Pathol 141:630–8. 10.1309/AJCPWW2URZ9JKQJU [DOI] [PubMed] [Google Scholar]
- 7.Olsen J, Espersen MLM, Jess P, Kirkeby LT, Troelsen JT (2014) The clinical perspectives of CDX2 expression in colorectal cancer: a qualitative systematic review. Surg Oncol 23:167–76. 10.1016/j.suronc.2014.07.003 [DOI] [PubMed] [Google Scholar]
- 8.Schmitt M, Silva M, Konukiewitz B, Lang C, Steiger K, Halfter K et al (2021) Loss of SATB2 occurs more frequently than CDX2 loss in colorectal carcinoma and identifies particularly aggressive cancers in high-risk subgroups. Cancers (Basel) 13:6177. 10.3390/cancers13246177 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bae JM, Lee TH, Cho N-Y, Kim T-Y, Kang GH (2015) Loss of CDX2 expression is associated with poor prognosis in colorectal cancer patients. World J Gastroenterol 21:1457–67. 10.3748/wjg.v21.i5.1457 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Eberhard J, Gaber A, Wangefjord S, Nodin B, Uhlén M, Ericson Lindquist K et al (2012) A cohort study of the prognostic and treatment predictive value of SATB2 expression in colorectal cancer. Br J Cancer 106:931–8. 10.1038/bjc.2012.34 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Dalerba P, Sahoo D, Paik S, Guo X, Yothers G, Song N et al (2016) CDX2 as a prognostic biomarker in Stage II and Stage III colon cancer. N Engl J Med 374:211–22. 10.1056/NEJMoa1506597 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Neumann J, Heinemann V, Engel J, Kirchner T, Stintzing S (2018) The prognostic impact of CDX2 correlates with the underlying mismatch repair status and BRAF mutational status but not with distant metastasis in colorectal cancer. Virchows Arch 473:199–207. 10.1007/S00428-018-2360-Y [DOI] [PubMed] [Google Scholar]
- 13.Aasebø K, Dragomir A, Sundström M, Mezheyeuski A, Edqvist P-H, Eide GE et al (2020) CDX2: a prognostic marker in metastatic colorectal cancer defining a better BRAF mutated and a worse KRAS mutated subgroup. Front Oncol 10:8. 10.3389/fonc.2020.00008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Konukiewitz B, Schmitt M, Silva M, Pohl J, Lang C, Steiger K et al (2021) Loss of CDX2 in colorectal cancer is associated with histopathologic subtypes and microsatellite instability but is prognostically inferior to hematoxylin-eosin-based morphologic parameters from the WHO classification. Br J Cancer 125:1632–46. 10.1038/s41416-021-01553-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ma C, Olevian DC, Lowenthal BM, Jayachandran P, Kozak MM, Chang DT et al (2018) Loss of SATB2 expression in colorectal carcinoma is associated With DNA mismatch repair protein deficiency and BRAF mutation. Am J Surg Pathol 42:1409–17. 10.1097/PAS.0000000000001116 [DOI] [PubMed] [Google Scholar]
- 16.Vlahović I, Rajc J, Švagelj I, Šolić K, Švagelj D (2023) Potential predictors for CDX2 expression loss and mismatch repair deficiency in colorectal cancer. Pathol Oncol Res 29:1610908. 10.3389/PORE.2023.1610908 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Alexander PG, Mcmillan DC, Park JH (2020) The local inflammatory response in colorectal cancer - Type, location or density? A systematic review and meta-analysis. Cancer Treat Rev 83:101949. 10.1016/j.ctrv.2019.101949 [DOI] [PubMed] [Google Scholar]
- 18.Galon J, Mlecnik B, Bindea G, Angell HK, Berger A, Lagorce C et al (2014) Towards the introduction of the “Immunoscore” in the classification of malignant tumours. J Pathol 232:199–209. 10.1002/path.4287 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Galon J, Pagès F, Marincola FM, Thurin M, Trinchieri G, Fox BA et al (2012) The immune score as a new possible approach for the classification of cancer. J Transl Med 10:1. 10.1186/1479-5876-10-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Jahan S, Awaja N, Hess B, Hajjar S, Sad S, Lohnes D (2022) The transcription factor Cdx2 regulates inflammasome activity through expression of the NLRP3 suppressor TRIM31 to maintain intestinal homeostasis. J Biol Chem 298:102386. 10.1016/j.jbc.2022.102386 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Balbinot C, Armant O, Elarouci N, Marisa L, Martin E, De Clara E et al (2018) The Cdx2 homeobox gene suppresses intestinal tumorigenesis through non-cell-autonomous mechanisms. J Exp Med 215:911–26. 10.1084/jem.20170934 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Wang Y-S, Kou Y, Zhu R-T, Han B-W, Li C-H, Wang H-J et al (2022) CDX2 as a predictive biomarker involved in immunotherapy response suppresses metastasis through EMT in colorectal cancer. Dis Markers 2022:9025668. 10.1155/2022/9025668 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ni H, Chen Y, Xia W, Wang C, Hu C, Sun L et al (2021) SATB2 defect promotes colitis and colitis-associated colorectal cancer by impairing Cl-/HCO3- exchange and homeostasis of gut microbiota. J Crohns Colitis 15:2088–102. 10.1093/ECCO-JCC/JJAB094 [DOI] [PubMed] [Google Scholar]
- 24.Coskun M, Olsen AK, Holm TL, Kvist PH, Nielsen OH, Riis LB et al (2012) TNF-α-induced down-regulation of CDX2 suppresses MEP1A expression in colitis. Biochim Biophys Acta 1822:843–51. 10.1016/j.bbadis.2012.01.012 [DOI] [PubMed] [Google Scholar]
- 25.Ma C, Henn P, Miller C, Herbst C, Hartman DJ, Pai RK (2019) Loss of SATB2 expression is a biomarker of inflammatory bowel disease-associated colorectal dysplasia and adenocarcinoma. Am J Surg Pathol 43:1314–22. 10.1097/PAS.0000000000001330 [DOI] [PubMed] [Google Scholar]
- 26.Elomaa H, Ahtiainen M, Väyrynen SA, Ogino S, Nowak JA, Friman M et al (2022) Prognostic significance of spatial and density analysis of T lymphocytes in colorectal cancer. Br J Cancer 127:514–23. 10.1038/s41416-022-01822-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Kastinen M, Sirniö P, Elomaa H, Äijälä VK, Karjalainen H, Tapiainen VV et al (2024) Establishing criteria for tumor necrosis as prognostic indicator in colorectal cancer. Am J Surg Pathol 48:1284–92. 10.1097/PAS.0000000000002286 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Elomaa H, Ahtiainen M, Väyrynen SA, Ogino S, Nowak JA, Lau MC et al (2023) Spatially resolved multimarker evaluation of CD274 (PD-L1)/PDCD1 (PD-1) immune checkpoint expression and macrophage polarisation in colorectal cancer. Br J Cancer 128:2104–15. 10.1038/s41416-023-02238-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Elomaa H, Härkönen J, Väyrynen SA, Ahtiainen M, Ogino S, Nowak JA et al (2024) Quantitative multiplexed analysis of indoleamine 2,3-dioxygenase (IDO) and arginase-1 (ARG1) expression and myeloid cell infiltration in colorectal cancer. Mod Pathol 37:100450. 10.1016/J.MODPAT.2024.100450 [DOI] [PubMed] [Google Scholar]
- 30.Bankhead P, Loughrey MB, Fernández JA, Dombrowski Y, McArt DG, Dunne PD et al (2017) QuPath: open source software for digital pathology image analysis. Sci Rep 7:16878. 10.1038/S41598-017-17204-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D et al (2003) Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 13:2498–504. 10.1101/gr.1239303 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Chewchuk S, Jahan S, Lohnes D (2021) Cdx2 regulates immune cell infiltration in the intestine. Sci Rep 11:15841. 10.1038/s41598-021-95412-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Murray PJ, Allen JE, Biswas SK, Fisher EA, Gilroy DW, Goerdt S et al (2014) Macrophage activation and polarization: nomenclature and experimental guidelines. Immunity 41:14–20. 10.1016/J.IMMUNI.2014.06.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Slik K, Turkki R, Carpén O, Kurki S, Korkeila E, Sundström J et al (2019) CDX2 loss with microsatellite stable phenotype predicts poor clinical outcome in stage II colorectal carcinoma. Am J Surg Pathol 43:1473–82. 10.1097/PAS.0000000000001356 [DOI] [PubMed] [Google Scholar]
- 35.Ma C, Olevian D, Miller C, Herbst C, Jayachandran P, Kozak MM et al (2019) SATB2 and CDX2 are prognostic biomarkers in DNA mismatch repair protein deficient colon cancer. Mod Pathol 32:1217–31. 10.1038/s41379-019-0265-1 [DOI] [PubMed] [Google Scholar]
- 36.Baba Y, Nosho K, Shima K, Freed E, Irahara N, Philips J et al (2009) Relationship of CDX2 loss with molecular features and prognosis in colorectal cancer. Clin Cancer Res 15:1665−4673. 10.1158/1078-0432.CCR-09-0401 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Cecchini MJ, Walsh JC, Parfitt J, Chakrabarti S, Correa RJ, MacKenzie MJ et al (2019) CDX2 and Muc2 immunohistochemistry as prognostic markers in stage II colon cancer. Hum Pathol 90:70–9. 10.1016/J.HUMPATH.2019.05.005 [DOI] [PubMed] [Google Scholar]
- 38.Lee J-A, Park HE, Jin H-Y, Jin L, Cho N-Y, Bae JM et al (2024) Concomitant expression patterns of CDX2 and SATB2 as prognostic factors in stage III colorectal cancers. Ann Diagn Pathol 71:152289. 10.1016/j.anndiagpath.2024.152289 [DOI] [PubMed] [Google Scholar]
- 39.Hestetun KE, Aasebø K, Rosenlund NB, Müller Y, Dahl O, Myklebust MP (2020) Mismatch repair phenotype determines the implications of tumor grade and CDX2 expression in stage II–III colon cancer. Mod Pathol 34:161–70. 10.1038/s41379-020-0634-9 [DOI] [PubMed] [Google Scholar]
- 40.Cígerová V, Adamkov M, Drahošová S, Grendár M (2021) Immunohistochemical expression and significance of SATB2 protein in colorectal cancer. Ann Diagn Pathol 52:151731. 10.1016/j.anndiagpath.2021.151731 [DOI] [PubMed] [Google Scholar]
- 41.Graule J, Uth K, Fischer E, Centeno I, Galván JA, Eichmann M et al (2018) CDX2 in colorectal cancer is an independent prognostic factor and regulated by promoter methylation and histone deacetylation in tumors of the serrated pathway. Clin Epigenetics 10:120. 10.1186/S13148-018-0548-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Dawson H, Koelzer VH, Lukesch AC, Mallaev M, Inderbitzin D, Lugli A et al (2013) Loss of Cdx2 expression in primary tumors and lymph node metastases is specific for mismatch repair-deficiency in colorectal cancer. Front Oncol 3:63999. 10.3389/FONC.2013.00265/BIBTEX [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Elnady MS, Eltatawy FA, Nosseir AG, Zamzam YA, El-Guindy DM (2022) Diagnostic accuracy of SATB2 in identifying primary and metastatic colorectal carcinoma: a comparative immunohistochemical study. Ecancermedicalscience 16:1491. 10.3332/ECANCER.2022.1491 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Galon J, Costes A, Sanchez-Cabo F, Kirilovsky A, Mlecnik B, Lagorce-Pages C et al (2006) Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Science 313:1960–1964. 10.1126/science.1129139 [DOI] [PubMed] [Google Scholar]
- 45.Ye L, Zhang T, Kang Z, Guo G, Sun Y, Lin K et al (2019) Tumor-infiltrating immune cells act as a marker for prognosis in colorectal cancer. Front Immunol 10:2368. 10.3389/FIMMU.2019.02368 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data generated and/or analyzed during this study are not publicly available. The sharing of data will require approval from relevant ethics committees and/or biobanks. Further information including the procedures to obtain and access data of Finnish Biobanks are described at https://finbb.fi/en/fingenious-service.