Activin A and B share the same downstream signalling pathway (activation of SMAD2/3) as TGF-β and consequently elicit many of the same functional responses as TGF-β, including immune suppression, activation of cancer-associated fibroblasts (CAFs) and extracellular matrix production and remodelling. However, activin’s role in tumourigenesis has been relatively overlooked compared to TGF-β’s. We generated and characterized a dual specificity human antibody that recognizes both activin A and B and compared its activity in syngeneic mouse models of breast cancer and pancreatic ductal adenocarcinoma (PDAC) with an activin A-specific antibody. We demonstrate that activin A and B are central to the function of CAFs and therapeutic inhibition of activin results in a reduction of collagen rich desmoplastic barriers, enabling the infiltration of cytotoxic T cells. This is correlated with an upregulation of the T cell chemoattractant CXCL10, which is normally repressed by activin signalling. Interestingly, despite greater T cell infiltration, activin A inhibition resulted in poorer survival in the KPC mouse model of PDAC and slightly larger tumours in the breast cancer model, indicating a tumour suppressive role of activin A-rich CAFs. Strikingly, however, treatment with the same anti-activin A antibody of PDAC tumours where SMAD4 is deleted in the tumour cells, resulted in increased survival, which was potentiated with additional treatment with immune checkpoint inhibitors. These results suggest that anti-activin therapy has potential for the cohort of PDAC patients exhibiting inactivation of SMAD4.

SMAD4 is a tumour suppressor gene that is mutated in up to 60% of pancreatic ductal adenocarcinoma (PDAC) tumours. Activins are members of the TGF-b signalling superfamily which signal through SMAD4 and are generally associated with an oncogenic process, with differing effects on tumour and stromal cells, such as cancer associated fibroblasts (CAFs). We sought to further elucidate the role of activin within PDAC and perturb activin signalling in vitro and in vivo. Analysis was performed of publicly available human datasets alongside human and murine single-cell RNA-sequencing datasets to study the effect of activin expression on survival and tumourigenesis. We developed anti-activin monoclonal antibodies to perturb activin signalling in PDAC mice, with deep analysis of tumour composition including immune cells and CAFs via a multitude of basic science techniques. We demonstrated that circulating activin levels are higher in PDAC patients than in healthy volunteers, whilst activin overexpression is associated with reduced overall survival in PDAC patients. Activin was shown to be predominantly expressed by contractile (collagen-depositing) CAFs, as opposed to inflammatory CAFs. Activin expression co-localises with fibroblasts, and anti-activin antibodies can reduce the collagen content of dense PDAC tumours. Additionally, anti-activin antibodies led to a statistically significant increase in CD4 and CD8 T-cell infiltration into murine PDAC tumours. We have shown that activin is overexpressed in PDAC, and high expression is associated with a more aggressive disease process. Anti-activin antibodies can alter the composition of desmoplastic PDAC tumours and increase the immune infiltrate of these typically “immune cold” tumours.

Deregulation of TGF‐β family signalling underlies some serious human diseases, including cancer, the Marfan syndromes, and other connective tissue disorders. Recent work in my lab has focused on working out how mutations in key components of TGF‐β family signalling pathways cause these diseases.

Fibrodysplasia ossificans progressiva (FOP) and diffuse intrinsic pontine glioma (DIPG) are debilitating diseases that share causal mutations in ACVR1, a TGF‐β family type I receptor. ACVR1R206H is a frequent mutation in both diseases. Pathogenic signaling via the SMAD1/5 pathway is mediated by Activin A, but how the mutation triggers aberrant signaling is not known. We show that ACVR1 is essential for Activin A‐mediated SMAD1/5 phosphorylation and is activated by two distinct mechanisms. Wild‐type ACVR1 is activated by the Activin type I receptors, ACVR1B/C. In contrast, ACVR1R206H activation does not require upstream kinases, but is predominantly activated via Activin A‐dependent receptor clustering, which induces its auto‐activation. We use optogenetics and live‐imaging approaches to demonstrate Activin A‐induced receptor clustering and show it requires the type II receptors ACVR2A/B. Our data provide molecular mechanistic insight into the pathogenesis of FOP and DIPG by linking the causal activating genetic mutation to disrupted signaling.

Introduction: Lung cancer is the result of a multistep accumulation of genetic and/or epigenetic alterations; therefore, a better understanding of the molecular mechanism by which these alterations ...