A wide variety of immune cells with lymphoid and myeloid cell lineage are distributed in the primary site of human cancer and are involved in the formation of the TME and ultimately influence the behaviour of the disease. For breast cancer, it is recognized that marked lymphocytic infiltration in the cancer tissue is associated with higher pCR rates after NAC [12, 19] and favourable prognosis for non-luminal type cancers [19, 20]. Breast cancers with significant lymphocytic infiltration have been defined as LPBC [12, 21], and the International TILs Working Group 2014 has recommended histological criteria based on the occupancy of stromal lymphocytes in cancer tissue to evaluate the degree of lymphocyte infiltration [22]. These criteria have been carefully established including the presentation of practical examples, and are helpful for histopathological examination, and their usefulness has been widely verified [20, 23, 24], though there has been some level of discrepancy in the results in some specific situations [25]. On the other hand, marked heterogeneity in the local immune response in breast cancer is recognized [26, 27], and the histological criteria for LPBC still require optimisation to account for this heterogeneity, such as incorporating other examination procedures.
The International TILs Working Group criteria do not consider the composition of immune cell lineage, as they presumably prioritise the reproducibility of practical histopathological evaluation, and the density of immune cells are evaluated on hematoxylin and eosin-stained specimens. The fundamental purpose of this current study was to evaluate the composition of immune cell lineage in the TME, which would provide a marked improvement of the LPBC evaluation criteria. As several previous studies adopted the multiplexed fluorescent labelling method to evaluate immune reactions in breast cancer [28, 29], the method is now an important and useful tool in the histological analysis of human cancer specimens. Previous studies mainly focused on the expression of markers related to cytotoxic immune reactions or immune check-point-related molecules. However, in this study we performed multiplexed labelling for key markers of lymphoid and myeloid lineage, including CD8 (typical cytotoxic T cells), CD4 (helper T cells), CD19 (pan B cells), CD11c (conventional DCs), and CD11b (pan myeloid cells), to analyse the composition of the fundamental immune cell lineages. Multiplexed labelling enabled us to count the immune cells expressing these markers simultaneously on identical fields of view.
The level of infiltration by these immune cells was associated with each other in many instances. Such tendency of association among various immune cell lineages is likely due to a certain harmonized balance among these lineages which can be maintained regardless of the degree of inflammation. This may provide validity to the histological criteria for LPBC by the overall inflammatory response. In this study, because we focused on multiplexed immunohistochemistry in a limited number of cases, the degree of TIL infiltration was not evaluated using the International TILs Working Group criteria, although, comparative examination between the overall degree of inflammation and the balance between individual cell lineages remains an important issue that should continue to be investigated.
The main issue of this study was to evaluate the relationship between the degree of infiltration of each immune cell lineage and therapeutic effect, and we revealed that only the degree of infiltration by CD8-positive TILs was significantly associated with therapeutic effect among the cell lineages examined. Those tumours with a high level of CD8-positive TIL infiltration can be considered to have a T-cell inflamed phenotype [30] which is functionally suppressed by inhibitory effects, including programmed death ligand 1 expressed on cancer cells or regulatory T cell infiltration [31]. However, some of the CD8-positive TILs may be activated as antigen-specific cytotoxic T cells, and our findings are consistent with this viewpoint of cancer immunity. We also have previously reported the significance of CD8-positive TILs in medullary carcinoma of the breast, which is regarded as a subtype of LPBC with marked lymphocytic infiltration. A high level of intra-tumoural CD8-positive TIL infiltration is a specific feature that distinguishes medullary carcinoma, which has a favourable prognosis despite its high grade atypical histology, from other lymphocyte-rich cancers [32]. Medullary carcinoma can be considered as a special subtype of LPBC with a T-cell inflamed phenotype.
Today, immune check-point inhibitors are commonly used for TNBC. CD8-positive cytotoxic T cells are the target cells to be re-activated by check-point inhibition. Therefore, it is highly probable that CD8 TILs would be a useful biomarker for LPBC to predict the effect of immune check-point inhibitors. A more detailed investigation and examination of the usefulness of CD8-positive TILs in biopsy specimens of TNBC, including a histological evaluation procedure for cut-off values, is urgently required.
The levels of infiltration by other cell lines, including CD4-positive TILs, CD19-positive TILs, CD11c-positive cells, and CD11b-positive cells, were not directly associated with therapeutic effect. However, assuming a possible influence by these lineages on the correlation between CD8-positive TIL infiltration and therapeutic effect, we performed a stratified analysis. Although no statistical significance was confirmed, the level of CD8-positive TIL infiltration tended to be associated with pCR in the subgroup that also had higher levels of CD4-positive TILs, but not with pCR in the subgroup with lower levels of CD4-positive TIL infiltration. These results indicate that CD4-positive TILs may have some positive influence on the correlation between the infiltration level of CD8-positive TILs and therapeutic effect. CD4-positive T-lymphocytes are composed of various subsets of T helper cells, including Th1, Th2, Th9, Th17, Th22, and Tfh, and regulator T cells, which are responsible for specific functions and interact in the TME [33]. Multiplexed immunohistochemistry can identify each T helper subset by detecting the master transcription factor of the lineage. Specific detection of T helper subsets and detailed evaluation of their level of infiltration and the balance between each subset are important issues that require further examination.
Stratified analysis of the level of myeloid cell infiltration also showed that higher levels of CD8-positive TIL infiltration were significantly associated with a higher pCR rate in the subgroup with lower levels of CD11c-positive cell infiltration, while such association was not observed in the subgroup with higher levels of CD11c-positive cells. CD11c-positive cells are members of various immune cell lineages. Both conventional DC1, which promotes effective cytotoxic immune reaction by cytotoxic T cells, and conventional DC2, which is responsible for priming CD4-positive T cells, are CD11c positive. In addition, monocyte-derived DCs and some macrophages express CD11c [34]. Considering that a high level of CD8-positive TIL infiltration was associated with higher pCR rates in the subgroup with a lower level of CD11c-positive cell infiltration, it is possible that some CD11c-positive cells have a suppressive function against CD8-positive TILs. The CD11c-positive cell lineage should be further investigated using a more detailed multiplexed labelling panel to sort independent cell lineages.
No particularly clear impact on the level of CD11b-positive cell infiltration and the correlation between CD8-positive TILs and pCR rate was identified. However, it is possible that CD11b-positive cells may have some effect on immune cell lineages other than CD8-positive TILs. Also considering the possible infiltration of myeloid-derived suppressor cells which are positive for CD11b, CD11b-positive cell lineages should also be examined more deeply with specific labelling panels.
Multivariate analysis revealed that remnant lymph node metastasis was associated with the DFS rate, rather than the level of CD8-positive TIL infiltration or therapeutic effect. The residual rate of lymph node metastasis was 51.7% in the non-pCR subgroup versus 48.3% in the pCR subgroup, which is similar and is consistent as an independent variable. Stratified analysis by residual lymph node metastasis revealed that CD8-positive TILs did not correlate with DFS in the group without remnant lymph node metastases which may include patients without lymph node metastasis before NAC, whereas higher levels of CD8-positive TIL infiltration were associated with predominantly longer DFS in those patients with lymph node metastases. This indicates that the level of CD8-positive TIL infiltration is a prognostic factor that lowers the risk of recurrence in the group with residual lymph node metastasis which has a significantly higher incidence of recurrence.
The case-specific TME, especially the level of CD8-positive TIL infiltration, which can be evaluated using biopsy specimens before adopting NAC, may predict the recurrence risk of patients with lymph node metastasis and would be helpful for considering more aggressive post-operative therapeutic strategies. Furthermore, it is highly probable that the level of CD8-positive TIL infiltration may serve as a useful biomarker for the therapeutic efficacy of immune check-point inhibitors which re-activate cytotoxic T cell function, and may also be a determinant for selected therapy.
In conclusion, we found that the level of CD8-positive TIL infiltration in pre-therapeutic biopsy specimens would be a helpful biomarker for predicting pCR, and high levels are also associated with favourable DFS in patients with residual lymph node metastasis after NAC. CD8-positive TILs are also a presumptive biomarker for predicting the therapeutic effect of immune check-point inhibitors which induce a cytotoxic immune response via CD8-positive lymphocytes. Exploration of a practical cut-off value for CD8-positive TIL infiltration for histological diagnosis is an important future issue, and it will also be necessary to examine the status of CD8-positive TILs in the TME in more detail, including their activated status, e.g., the expression of programmed death ligand 1, to examine their validity as a histological biomarker.