M1 hot tumor-associated macrophages boost tissue-resident memory T cells infiltration and survival in human lung cancer

Background The role of tumor-associated macrophages (TAMs) in determining the outcome between the antitumor effects of the adaptive immune system and the tumor's anti-immunity stratagems, is controversial. Macrophages modulate their activities and phenotypes by integration of signals in the tumor microenvironment. Depending on how macrophages are activated, they may adopt so-called M1-like, antitumor or M2-like, protumor profiles. In many solid tumors, a dominance of M2-like macrophages is associated with poor outcomes but in some tumor types, strong M1-like profiles are linked to better outcomes. We aimed to investigate the interrelationship of these TAM populations to establish how they modulate the efficacy of the adaptive immune system in early lung cancer. Methods Macrophages from matched lung (non-tumor-associated macrophages (NTAMs)) and tumor samples (TAMs) from resected lung cancers were assessed by bulk and single-cell transcriptomic analysis. Protein expression of genes characteristic of M1-like (chemokine (C-X-C motif) ligand 9) or M2-like (matrix metallopeptidase 12) functions was confirmed by confocal microscopy. Immunohistochemistry related the distribution of TAM transcriptomic signatures to density of CD8 + tissue-resident memory T cells (T RM) in tumors and survival data from an independent cohort of 393 patients with lung cancer. Results TAMs have significantly different transcriptomic profiles from NTAMs with >1000 differentially expressed genes. TAMs displayed a strong M2-like signature with no significant variation between patients. However, single-cell RNA-sequencing supported by immuno-stained cells revealed that additionally, in 25% of patients the M2-like TAMs also co-expressed a strong/hot M1-like signature (M1 hot). Importantly, there was a strong association between the density of M1 hot TAMs and T RM cells in tumors that was in turn linked to better survival. Our data suggest a mechanism by which M1 hot TAMs may recruit T RM cells via CXCL9 expression and sustain them by making available more of the essential fatty acids on which T RM depend. Conclusions We showed that in early lung cancer, expression of M1-like and M2-like gene signatures are not mutually exclusive since the same TAMs can simultaneously display both gene-expression profiles. The presence of M1 hot TAMs was associated with a strong T RM tumor-infiltrate and better outcomes. Thus, therapeutic approaches to re-program TAMs to an M1 hot phenotype are likely to augment the adaptive antitumor responses.