Spatial Immunology in Liver Metastases from Colorectal Carcinoma according to the Histologic Growth Pattern

Abstract

Colorectal cancer liver metastases (CRC-LM) present differential histologic growth patterns (HGP) that determine the interaction between immune and tumor cells. We explored the spatial distribution of lymphocytic infiltrates in CRC-LM in the context of the HGP using multispectral digital pathology. We did not find statistically significant differences of immune cell densities in the central regions of desmoplastic (_dHGP) and non-desmoplastic (_ndHGP) metastases. The spatial evaluation reported that _dHGP-metastases displayed higher infiltration by CD8⁺ and CD20⁺ cells in peripheral regions as well as CD4⁺ and CD45RO⁺ cells in _ndHGP-metastases. However, the reactive stroma regions at the invasive margin (IM) of _ndHGP-metastases displayed higher density of CD4⁺, CD20⁺, and CD45RO⁺ cells. The antitumor status of the TIL infiltrates measured as CD8/CD4 reported higher values in the IM of encapsulated metastases up to 400 μm towards the tumor center (p < 0.05). Remarkably, the IM of _dHGP-metastases was characterized by higher infiltration of CD8⁺ cells in the epithelial compartment parameter assessed with the ratio CD8^epithelial/CD8^stromal, suggesting anti-tumoral activity in the encapsulating lesions. Taking together, the amount of CD8⁺ cells is comparable in the IM of both HGP metastases types. However, in _dHGP-metastases some cytotoxic cells reach the tumor nests while remaining retained in the stromal areas in _ndHGP-metastases.

Keywords: desmoplasia; growth pattern; immunology; liver metastases; lymphocyte; multiplex.

Figure 1

(A) Marker combination to identify each of the lymphocyte subsets assessed in the study. (B) Boxplots of cell densities (cells/mm²) for each of the subsets assessed and stratified by HGP. We differentiated cell counts according to different topographic locations in the tumor, either in the tumor periphery or in central tumor areas as well as differentiating whether infiltrates were located in the stroma or in tumor glands. Tumor periphery was defined as the tissue fraction from 100 μm of the proximal liver parenchyma until the first row of tumor cells, thus, including the capsule in dHGP CRC-LM and any rare portion of peritumoral stroma in ndHGP CRC-LM. Asterisks illustrate p value < 0.05, U-Mann Whitney test comparing dHGP vs. ndHGP.

Figure 2

(A) Distances considered in relation to the tumor-host interface. Measure A, using as a reference the tumor-liver interface, including as part of the tumor, the fibrous capsule in case of _dHGP or any small stromal area in the case of _ndHGP; Measure B, using the tumor margin as a reference and then, not considering capsule or any peritumoral stromal. We use this measure throughout the manuscript to refer to the invasive margin. (B) Detailed magnification of the tumor/liver interface and central tumoral areas for both main histologic growth patterns.

Figure 3

(A) Spatial assessment of lymphocyte subsets in _dHGP and _ndHGP CRC-LM according to two different measures, Measure A, from liver margin, Measure B, from the outer tumoral region (invasive margin). Lines show mean levels of cell densities. Dots show cell count/mm². X-axis is square-root transformed (distant segments, μm). Mann–Whitney U test was applied for statistical analysis. (B) Boxplots of cell densities (cells/mm²) for each of the subsets were assessed and stratified by HGP according to the invasive margin. Asterisks illustrate p value < 0.05, Mann–Whitney U test comparing _dHGP vs. _ndHGP.

Figure 4

Top panel, boxplot for CD8/CD4 ratio comparing _dHGP and _ndHGP in central tumoral areas and at the invasive margin. Bottom panel, boxplot displaying the fraction of CD8⁺ cells within the epithelial compartment, both at the invasive margin and central tumor areas. Asterisks illustrate p value < 0.05, Mann–Whitney U test comparing _dHGP vs. _ndHGP.