The tumor immune microenvironment exerts a pivotal influence in tumor initiation and progression.
The etiopathogenesis of most colorectal carcinomas (CRC) follows the conventional adenoma–carcinoma sequence (ACS), in which colorectal adenomas with low-grade dysplasia (LGD) and/or high-grade dysplasia (HGD) evolve to invasive adenocarcinoma (ADC), through a stepwise accumulation of genetic and epigenetic alterations. The earliest events in colorectal tumorigenesis involve alterations of the WNT signaling pathway, most frequently through somatic mutation of the adenomatous polyposis coli
) gene 
. Germline mutations in the APC
gene are the genetic cause of familial adenomatous polyposis (FAP), an autosomal dominant syndrome characterized by numerous adenomatous lesions in the colorectum, and early onset CRC.
The tumor immune microenvironment has a crucial role in the development and progression of colorectal precursor and invasive lesions 
. Moreover, the amount, type, and location of tumor-infiltrating lymphocytes have been shown to have prognostic value in CRC patients 
. With the widespread use of targeted immunotherapy in solid tumors, including CRC 
, there is a growing research interest in exploring the role of the tumor immune microenvironment also as a predictive biomarker. Currently, monoclonal antibodies against immune checkpoints, e.g., CTLA4, PD-1, and PD-L1, have been approved in microsatellite instability-high (MSI-H) CRCs and microsatellite-stable (MSS) tumors with high tumor mutation burden (TMB) 
, reflecting the immunogenicity of these tumor subtypes 
In the current body of literature, great attention has been given to the immune landscape of invasive ADC, whereas few studies have compared the tumor immune microenvironment of pre-invasive and malignant lesions, with contradictory findings 
. Some studies have demonstrated a lower expression of Th1 cytokines, a shift to a Th2 cytokine profile 
, and a decrease in cytotoxic CD8+ T cell counts along ACS 
, whereas others have suggested an increase in the density of T lymphocytes, regulatory T cells, and macrophages during tumor progression 
. Likewise, little information is available regarding the comparison of human colorectal neoplastic lesions arising from a sporadic or hereditary background, since most studies on hereditary cancer have used mouse models, especially APCmin
. Furthermore, most of the studies in human colorectal cancer focused on MSI tumors. A study by Liu, G. et al. showed that the number of CD3+, CD8+, CD4+ and FoxP3+ cells in the invasive margin or tumor stroma is higher in Lynch syndrome cases than in sporadic cases 
. Therefore, an in-depth characterization of the human tumor immune landscape and its evolution along the colorectal ACS is needed, especially for MSS colorectal cancer.
2. Tumor Infiltration with Immune Cells Decreases throughout the Colorectal ACS
To investigate the evolution of the tumor immune microenvironment along the colorectal ACS, we analyzed the density and quality of the immune infiltration in adenomatous lesions with LGD, with HGD, and in invasive ADCs. By immunohistochemistry and counting representative areas, we analyzed and normalized the density of the overall population of CD3+ T lymphocytes per mm2, as well as the specific populations of CD4+ helper T cells, cytotoxic CD8+ T lymphocytes, Foxp3+ regulatory T cells, and CD57+ T lymphocytes/natural killer (NK) cells.
We observed an overall decrease in tumor-infiltrating immune cells along the colorectal ACS in both sporadic and hereditary FAP-related cohorts (Figure 1). Adenomatous lesions with LGD harbored higher counts of CD3+, CD4+, CD8+, and CD57+ immune cells when compared with adenomatous lesions with HGD and invasive ADCs, in both sporadic and hereditary FAP-related lesions. These findings were more pronounced in sporadic lesions, with p-values < 0.001 when adenomatous lesions with LGD were compared with ADCs. Although the density of Foxp3+ regulatory T cells followed the same decrease pattern in the sporadic context, this was not observed in lesions from FAP patients. As the spatial distribution of immune cell types within the tumors could be relevant in the interpretation of our results, we studied the immune cell density separately at the center and at the front of invasive ADCs. No differences were found between the two locations (data not shown) and the results on immune cell counts were considered as a whole.
Figure 1. Counts of immune cells along the colorectal adenoma-carcinoma sequence in sporadic (A) and hereditary Familial Adenomatous Polyposis (FAP)-related (B) lesions. LGD, low-grade dysplasia; HGD, high-grade dysplasia; ADC, invasive adenocarcinoma; * = p < 0.05; ** = p < 0.01; *** = p < 0.001; **** = p < 0.0001.
To assess whether the tumor immune microenvironment could vary depending on the hereditary or sporadic background of the neoplastic lesions, we investigated the immune landscape between the two cohorts for LGD, HGD, and invasive ADC lesions (Figure 2).
Figure 2. Comparison of immune cell counts between sporadic and hereditary Familial Adenomatous Polyposis (FAP)-related lesions in adenomatous lesions with low-grade dysplasia (A), adenomatous lesions with high-grade dysplasia (B), and invasive adenocarcinoma (C). * = p < 0.05; **** = p < 0.0001.
We found that adenomas with LGD from FAP patients displayed a marginal increase in CD8+ T cells (p < 0.05) and a significantly lower density of Foxp3+ (p < 0.0001) T cells when compared with sporadic adenomas with LGD. No differences were observed for the remaining adaptive immune cell populations, or subgroups of sporadic and hereditary lesions.
In conclusion, our findings suggest that colorectal neoplastic lesions trigger a host immune response, but become immunologically colder throughout the colorectal ACS. This was observed for both sporadic and hereditary lesions, although it was more pronounced in the former. In that regard, we would like to stress the higher density of Foxp3+ regulatory T cells in sporadic versus hereditary LGD lesions, suggesting that the host has a higher baseline immunological tolerance towards hereditary lesions.
3. TMB and MHC-I Expression Increase along the Colorectal ACS
We hypothesized that the decrease in immune cells counts along the ACS could be partially explained by the progressive loss of immunogenicity 
. To explore this hypothesis, we evaluated TMB and the expression of the major histocompatibility complex class I (MHC-I) protein in tumor cells as surrogates to tumor antigenicity, and to the ability of tumor cells to present neoantigens to the immune system, respectively.
We found that TMB increased throughout the ACS in sporadic cases (Figure 3A), from the median value of 2.76 mutation/mega base pair (mut/Mbp) in LGD lesions to 15.76 mut/Mbp in ADC cases (p < 0.05). In contrast, this trend was not observed in hereditary FAP-related lesions (Figure 3B), where the average number of mut/Mbp remained constant along the ACS, despite the presence of some outlier samples. Likewise, we found that MHC-I expression in neoplastic cells progressively increased throughout the ACS in both FAP-related (p = 0.033) and sporadic (p = 0.025) lesions (Figure 4). MHC-I was expressed at the cell membrane, as well as in the cytoplasm of neoplastic cells (Figure 5).
Figure 3. TMB analysis along the colorectal adenoma-carcinoma in sporadic (A) and hereditary Familial Adenomatous Polyposis (FAP)-related (B) lesions. * = p < 0.05. LGD, low-grade dysplasia (circles); HGD, high-grade dysplasia (squares); ADC, invasive adenocarcinoma (triangles); * = p < 0.05.
Figure 4. Percentage of cases in which MHC-I in tumor cells is >25% (score 1) along the colorectal adenoma-carcinoma sequence, in sporadic (A) and hereditary Familial Adenomatous Polyposis (FAP)-related (B) lesions. LGD, low-grade dysplasia; HGD, high-grade dysplasia; ADC, invasive adenocarcinoma.
Figure 5. MHC-I and PD-L1 expression along the colorectal adenoma-carcinoma squence in representative examples of sporadic adenomatous lesion with low-grade dysplasia (LGD) (upper panel), sporadic adenomatous lesion with high-grade dysplasia (HGD) (middle panel), and sporadic invasive adenocarcinoma (ADC) (lower panel). Note the membranous and cytoplasmic immunoreactivity of MHC-I in neoplastic cells of HGD and invasive ADC lesions. Hematoxylin and eosin (left panel) and immunohistochemistry for MHC-I (middle panel) and PD-L1 (right panel), 40× amplification. The insets in the right panel represent higher magnification of PD-L1 immunohistochemistry, 400× amplification.
Overall, the increase in TMB in sporadic lesions, and the increase in MHC-I expression in both sporadic and FAP lesions along the ACS suggest that the observed decrease in immune cell counts cannot be explained by a loss of immunogenicity.