CD49a+ NK cells promote esophageal cancer development by inducing MDCSs infiltration via GM-CSF.

Abstract

Background: Although extensive evidence indicates that tissue-resident natural killer (trNK) cells are closely related to the development and clinical outcomes of various tumours, their roles in esophageal cancer remain obscure.

Methods: Tumour tissues collected from 54 esophageal squamous cell carcinoma (ESCC) patients during surgery and human ESCC tissue arrays including samples from 258 patients were subjected to analyse the phenotype and functions of immune cells.

Results: We observed a population of CD49a⁺ NK cells with a CD103^hiCD69^hi resident phenotype predominantly within the CD56^bright NK cell subset in the intratumoral tissue of ESCC patients. These CD49a⁺ trNK cells, characterised by high expression of inhibitory receptors (TIM-3, CD244, TIGIT, PD-1), reduced cytotoxic potential (lower CD16, granzyme B, and perforin), and elevated secretion of IFN-γ and TGF-β, exhibited an exhausted and regulatory phenotype. Their presence was associated with poor prognosis in early-stage ESCC, but not in advanced stages. Mechanistically, these cells were found to enhance the immunosuppressive tumour environment by increasing MDSCs through GM-CSF secretion, thereby facilitating tumour progression.

Conclusions: Our research reveals that tumour-infiltrated CD49a⁺ NK cells exhibit exhausted and regulatory profile, capable of inducing the accumulation of MDSCs by secreting GM-CSF, and predict poor clinical outcomes in early-stage ESCC patients. trNK cells in esophageal tissues exhibit antitumor potential via degranulation-mediated elimination of cancer cells, thereby performing immune surveillance. However, during tumour progression, upregulation of immune checkpoint molecules (e.g., TIM3 and CD244) on trNK cells results in reduced cytotoxic activity and an exhausted phenotype. Exhausted trNK cells further enhance the immunosuppressive tumour microenvironment by promoting MDSCs accumulation through GM-CSF secretion, ultimately facilitating tumour progression. The graph was created with BioRender.com.