Previous studies have confirmed a significant correlation between the infiltration degree of M2 macrophages in TME and the responsiveness to neoadjuvant therapy [7, 11]. However, the mechanism of M2 macrophages participating in neoadjuvant therapy is still unclear. In this study, all the functional genes of M2 macrophages were systematically screened by the TCGA database for the first time. The functional genes, namely SOCS1 and ORAI3, which may participate in the preoperative neoadjuvant therapy of ESCC, were screened and verified by the GEO database.
In this study, there are two main reasons that Score was used to be the final evaluation criterion for predicting pCR in the GEO database. Firstly, Score was the sum of the two genes’ regression coefficients multiplying by ratios of the expression of the two genes to the abundance of M2 macrophage genes, reflecting the relationship between the two genes and M2 macrophage better. Secondly, the expression of genes as the numerator in Score may be the leading factor of predicting the response of nCRT, not denominator. So it further indicated that the expression of potential genes had better predictive value than the abundance of M2 macrophage.
27 paraffin-embedded specimens of ESCC were used to verify the results of GEO database. We found no significant correlation between CD163 H-score and remission, which was consistent with the results of database analysis. Therefore, the predictive value of Score1 for neoadjuvant therapy can reflect the correlation between SOCS1 expression in the local tumor microenvironment and the curative effect. Finally, we found that the high expression of SOCS1 is significantly correlated with the favorable responsiveness to preoperative neoadjuvant therapy for ESCC.
SOCS1, as a cell signal suppressor, is mainly involved in the regulation of two signaling pathways, namely Janus tyrosine kinase / signal transducer and activator of transcription (JAK/STAT) signaling pathway and toll-like receptors (TLR) signaling pathway [12, 13, 14, 15, 16]. In JAK / STAT signaling pathway, SOCS1 can interact with JAK kinase domains, such as Jak1, JAK2, Tyk2, to inhibit the activity of the kinase and affect the downstream STAT protein phosphorylation, leading to the inactivation of the signaling pathway and playing a negative feedback regulation role [17, 18, 19]. In the TLR signaling pathway, SOCS1 can interact with TLR adaptor proteins, such as MAL and IRAK, leading to their degradation, thereby inhibiting this signaling pathway [16, 17, 20, 21]. As a tumor suppressor, SOCS1 gene promoter methylation leads to the inactivation of SOCS1 gene expression, which occurs in 53–71% of patients with pancreatic cancer and occurs in patients with acute myeloid leukemia and liver cancer [22, 23, 24]. McCormick found that in the signal pathway of IL-4 activating tyrosine phosphorylation of insulin receptor substrate-2 (IRS-2) to induce M2 macrophage polarization, SOCS1 protein can inhibit M2 macrophage differentiation by mediating IRS-2 ubiquitination and promoting protease degradation of phosphorylated IRS-2[25]. Whyte and Liu also found that SOCS1 was closely related to the differentiation of M2 macrophages [26, 27]. This study shows that SOCS1 can be used as a biomarker for the favorable responsiveness to preoperative neoadjuvant therapy for ESCC. The reason may be that SOCS1 can negatively regulate the differentiation of M2 macrophages in the tumor microenvironment, reduce the infiltration of M2 macrophages in the microenvironment, and thus reduce the immunosuppressive effect of M2 macrophages [28, 29, 30].
There are several limitations in this study: (1) The number of samples in training set GSE45670 and validation set GSE104958 is small, and the accuracy of the screened genes needs to be further studied and verified; (2) The number of paraffin-embedded specimens used for validation is small, and the results need to be further confirmed by expanding the sample size; (3) The ESCC specimens were from preoperative endoscopic biopsy, and the samples were taken from the surface of the tumor. So they may not reflect the whole tumor microenvironment. However, compared with the prediction efficiency of tumor microenvironment of preoperative endoscopic biopsy specimens and postoperative resection specimens in previous studies, it is found that the tumor microenvironment of preoperative endoscopic biopsy specimens can also better predict the prognosis and curative effect of patients [31, 32]. (4) The function of SOCS1 in M2 macrophage had not been studied, and further research is needed to verify the function.
In conclusion, the high expression of SOCS1 in the local tumor microenvironment is significantly related to the favorable responsiveness to neoadjuvant therapy in patients with ESCC, which can be used as a biomarker to predict the responsiveness to neoadjuvant therapy and has certain clinical value.