Oral Presentation Melbourne Immunotherapy Network Winter Retreat 2018

Targeting the tumour fibroblastic microenvironment to improve T cell function (#3)

Anne Fletcher 1 , Sara Cruz Migoni 1 2 , Konstantin Knoblich 1
  1. Monash University, Clayton, VIC, Australia
  2. University of Birmingham, Edgbaston, UK

Checkpoint inhibitor therapies show enormous promise in the treatment of cancer. Response rates are high for blood-borne cancers but relatively poor for solid tumours. Cancer associated fibroblasts (CAFs) are a hallmark of solid tumours and strongly associate with poor prognosis for patients.

We recently described (PLOS Biology, in press) a T cell immunosuppressive role for activated human fibroblasts within lymphoid organs. We hypothesised that CAFs are likely very similar, if not largely identical to activated fibroblasts elsewhere in the body, and may therefore similarly affect tumor infiltrating T cells, reducing the efficacy of checkpoint inhibitor therapy for solid tumours.

A novel meta-analysis of trial data showed that CAF load within solid tumours significantly correlated with lack of objective response to PD-1/PDL1 and CTLA4 checkpoint blockade. We then performed the first direct comparison of freshly isolated CAFs from human pancreatic, breast, and colorectal tumours, using flow cytometry and RNA-Seq. Phenotypes were strongly conserved, with few differences between tissue activated fibroblasts and intratumoral CAFs, supporting a model where CAFs are activated as a normal physiological response which is not tissue or cancer-dependent. Using in vitro co-cultures, all CAFs from breast, colorectal and pancreatic tumours, as well as all activated fibroblasts from adjacent tissues, strongly suppressed T cell activation at physiologically relevant ratios, regardless of tumour or tissue type. Mechanisms included prostaglandin E2 and IDO, and effects were reversible using available inhibitors. A novel assay developed to validate the reversal of T cell suppression using live tumour slices has shown a response in 2/3 tumours.

Together, this work shows that CAF phenotypes are highly conserved between tumours, tissues and patients; that CAFs are not truly cancer-associated but more likely inflammation associated; and that CAFs alter T cell activation using reversible mechanisms. Fibroblasts likely create a conserved therapeutically targetable T cell activation checkpoint.