Introduction
The incidence and mortality of colorectal cancer (CRC) rank third and second in the world, respectively [
1]. A recent epidemiological report in China pointed out that the detection rate of stage IV colorectal cancer has increased, and the mortality rate has also increased [
2]. Tumor metastasis is the main cause of death in these patients, and the most common site of metastasis is the liver. Surgery is currently the best treatment for liver metastases, and its 5-year survival rate can reach 20–45% [
3]. However, resectability depends on many factors, including the extent of liver metastases, and whether there are other unresectable extrahepatic diseases and patient complications [
4]. Therefore, for unresectable patients, effective treatments are urgently needed, and immunotherapy has become a hot spot of research. Studies have shown that immune checkpoint inhibitors have sustained clinical responses in colorectal cancer patients with high microsatellite instability (MSI-H) or mismatch repair deficiency (MMR-D), with significant clinical improvements reported [
5,
6]. However, immunotherapy has shown a weakened therapeutic effect on liver metastases [
7].The immunosuppressive microenvironment of the liver may be the reason for the weakened immunotherapy effect. In particular, studying the difference in the tumor microenvironment between primary colorectal cancer and liver metastases is necessary to understand the differential responsiveness to immunotherapy.
The tumor microenvironment is a complex and dynamic system that contains a variety of immune cells. Tumor-infiltrating T lymphocytes are regarded as the main effectors of anti-tumor immune response [
8,
9], and CD8 is recognized as the important marker of T-cell infiltration. Foxp3 is an important marker of regulatory T cells. In different types of tumors, a large amount of Treg cell infiltration is associated with poor clinical prognosis [
10], but its role in colon cancer is controversial [
11]. CD68- and CD163-labeled tumor-associated macrophages (TAMs) have both tumor-promoting and antitumor effects on tumors [
12,
13], and they both have significant correlation with prognosis. For example, a high CD163
+/CD68
+ ratio in the infiltrative margin of the tumor suggests a poor prognosis for the colorectal cancer patients. [
14]. We also include CD20-labeled B lymphocytes [
15], CD66b-labeled neutrophils [
16], CD11c-labeled dendritic cells (DCs) [
17] and CD56-labeled natural killer (NK) cells [
18], which all play an important roles in the occurrence and development of tumors. In addition, PD-L1 is an important marker for the evaluation of immunotherapy [
19], VEGFR-2 is one of the key markers of tumor angiogenesis [
20], INF-γ is an important proinflammatory factor that inhibits tumor growth [
21], and Ki67 evaluates cell proliferation and is also related to tumor prognosis [
22]. In this study, we used these biomarkers to evaluate the overall tumor microenvironment of primary colorectal cancer and liver metastases from multiple perspectives to gain a deeper understanding of the differences in the tumor microenvironment.
Discussion
In recent years, multiplex immunofluorescence has been demonstrated in several studies as a very valuable and meaningful tool for the immunoassay of tumor tissues [
31,
32]. Compared to previous studies, we utilized the multiplex immunofluorescence analysis to visualize the expression of multiple immune cells on a single slide, which would give us a comprehensive overview of the tumor microenvironment and avoid the spatial heterogeneity compared with conventional IHC on consecutive slides. In this study, we used PerkinElmer Inform software to compare and analyze various aspects of primary colorectal cancer and liver metastases (T cells, B cells, macrophages, neutrophils, DCs, NK cells, lymphokines, PD-L1, tumor angiogenesis and cell proliferation). Previous research has mainly focused on T lymphocyte [
8,
33,
34], but T cells are insufficient for representing the complex microenvironment of the immune system. Macrophages, B cells, neutrophils, etc., also play an important role in the tumor microenvironment [
35‐
38]. Therefore, the multicellular perspective allows us to better understand the tumor microenvironment between primary tumors and liver metastases. In addition, according to recent reports, the immune microenvironment of tumors mainly depends on two regions (TC and TF), but the definition of the TF region is not clear. In our research and observation, the spatial distribution of immune cells at the edge of tumor invasion was different. At a distance of less than 150 µm from the edge of the tumor, immune cells obviously clump together, similar to the formation of an “immune band”. At the same time, in the area more than or equal to 150 µm from the tumor, some immune cells also showed an increase, illustrating that the tumor microenvironment is dynamic. Therefore, it is necessary for us to conduct a more detailed spatial analysis of the tumor infiltration margin to better analyze the microenvironment and strengthen our understanding of the spatial distribution of the microenvironment.
In our research, we found that some immune cells have a certain correlation with the number and size of liver metastases. One possible explanation for this might be that the immunosuppressive effect of the tumor is stronger when the number of liver metastases is > 1 or the diameter is ≥ 3 cm, or it may only be due to individual differences in patients. In the analysis of the expression of a variety of immune cells between the primary tumor and liver metastases, it was found that, overall, the liver metastases and primary tumors have similar immune statuses in regard to the presence of T cells, B cells, dendritic cells, PD-L1, tumor angiogenesis, and cell proliferation. However, hepatic immunosuppression cannot be clearly seen from the overall differences in neutrophils, macrophages, regulatory T cells, and INF-γ in primary and liver metastases. From the perspective of spatial distribution, the immune cells of the primary tumor and liver metastases are mainly concentrated in the invasive front of the tumor, less than 150 µm from the tumor, which is much higher than the tumor center, which confirms previous research [
39]. Compared with the primary tumor, liver metastases express more immune cells in the peritumor area more than or equal to 150 µm away from the tumor, especially CD163-labeled macrophages, which have the highest expression in the PT region. It is well known that CD163 can enhance the migration and invasion of colorectal cancer cells [
40] and is related to the poor prognosis of patients [
41]. This may be an important reason for the immunosuppression of liver metastases.
Furthermore, we conducted a colocalization analysis of CD68- and CD163-labeled macrophages, and found that the spatial distribution of macrophages changed significantly. In the PT region, the distribution of macrophages in primary tumors and liver metastases is significantly different. The primary tumors were mainly CD68
+CD163
− cells, while the liver metastases were mainly CD68
+CD163
+ and CD68
−CD163
+ cells. Interestingly, CD68
−CD163
+ cells showed a significant increase in liver metastases in the spatial distribution among TC, TF, and PT regions compared with primary tumors. Research reports that macrophages in liver metastases can cause CD8 T-cell loss and reduce the effectiveness of immunotherapy [
26] which also suggests that the high expression of CD68
−CD163
+ macrophages may be related to the immunosuppression of liver metastases. According to the immune status in different spatial distribution areas of the liver metastasis and the primary tumor, we speculate that CD68
−CD163
+ macrophages may not only inhibit or eliminate the function of CD8 T cells, but also inhibit or eliminate B cells, dendritic cells and PD-L1. Furthermore, we also found significant differences in the expression of CD163 in different cell types by immunophenotyping. This represents a difficult problem in the administration of immunotherapy for patients with liver metastases. Notably, a recent phase I clinical trial study found that the combination of regorafenib and nivolumab in MSS colorectal cancer patients reached an objective response rate of 33%. However, none of the patients with liver metastases are responsive [
42], suggesting that the immunosuppressive TME in liver metastases may hamper the efficacy of PD-1 blockade. In our study, we found that CD68
+CD163
− macrophages were closely related to the immunosuppression of liver metastases. In future, it is promising that specific targeting and modulation of this group of macrophages may have potential to stimulate tumor immunity and enhance anti-tumor activity of PD-1 blockade. Addressing the problem of immunosuppression of macrophages in liver metastases may lead to breakthroughs in immunotherapy. To our knowledge, there is few studies to analyze the similarities and differences in the microenvironment between liver metastases and the primary tumor by using multiple types of cells from a pathological perspective.
However, our research has some limitations. First, this was a retrospective study. In addition, the sample size was small, and the selection of various immune cell markers was not sufficient. It does not fully represent all types of immune cells. The spatial extent of a PT region larger than 150 µm is not precisely defined, and not all tumors have a PT region. When the tumor is observed to be more poorly differentiated, we can only divide it into TF and TC regions, and cannot clearly distinguish the PT region. Therefore, more studies are needed in the future to confirm our results.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.