Overcoming resistance to Immune therapy

Last ASCO session this year and the session is jam-packed!

work in progress

Rational design of combination immune therapies
Michael Postow

Different ideas how to strengthen the immune response to tumours:

1. Increasing the 'visibility' of the tumour to the immune system

• Intratumoral oncolytic virus injection:
• T-Vec = Imlygic (Chesney et al.)
• Coxsackievirus A21 (Curti et al.)
• Entinostat (johnson et all)
• Cobimetinib + anti-PDL1

2. More T-Cells into the tumour

CD19 Chimeric antigenic receptor T-cells + Pembro (Maude et al)

CEA T-Cell bispecific antibody plust Atezo (Tabernero et al)- colon 

3. 'Force field'- make the environment more favourable for T-Cells

IDO- deletes tryptohpan, produces toxic kynurenine: inhibits T-Cells


4. Enhance T cell attack:

- block negative checkpoints
- activate positive checkpoints



How to test all possible combinations in a rational way?

- is there a reason to believe 2 agents work together?
- does the new treatment slow down tumour progression?





Resistance to Immunotherapy 
Tom Gajewski


Working model

T-cell inflamed versus non-flamed tumour environment

Vaccine-responders had a T-cell inflamed phenotype and favours response to checkpoint inhibitors (find old BMS study on Ipi)

source: Nature Immunology 2013 


CD8, FoxP3, PDL1, IDO present in T cell-inflamed, but absent in T cell non-inflamed


Primary resistance might by caused by absence of a T cell-inflamed environment. How can one create such a favourable environment?
The injections with viruses- like T-Vec- is thought to 'wake up' the environment


3 major hypothesis for primary resistance

- somatic differences (tumours are different)
- host differences (patients are genetically different- polymorphism in immune regulatory genes, so their immune systems react differently)
- environmental differences (like microbiota, pathogen exposure)

Most likely, all 3 factors are important



Biomarkers
Jeffery Weber

What can biomarkers tell you?

- risk prediction
- diagnostic
- monitor response to treatment
- predict response to treatment
- toxicity


Prognostic versus predictive?


Profiles associated with response to PD1 blockade

- high mutational load: better response to PD1
- Hugo Cell 2016- profiles of people who do not respond to PD1


AND AGAIN

PDL1 expression is associated with better response to PD1 blockade but it is not predictive. And in 30% of times, not even consistent in the SAME tumour. 

CD8 T cell infiltrate predicts response to PD1


JAK 1/2 mutations associated with acquired resistance to PD1-blockade

Literature
Daniel Sanghoon Shin et al. Cancer Discovery 2017
Hugo Cell 2016
Gao Cell 2016
Yusko, 2017 (unpublished)- mutational load associated with PD1 response but overlap huge, so clinical value limited right now
Huang et al. Nature 2017