Thread by @jimjaxn100

Happy to announce our latest publication is online!
rdcu.be/c1f58
Huge effort from @ashkanshahbandi and co authors @Raegan2222 @SelinaSMJ_here @AshlynA_99 @LabMachado @Lorax_Bodyspray @Thesrao
#cancer #breastcancer #Immunotherapy #SABCS2022
a quick 🧵 tweetorial...

@ashkanshahbandi was interested in examining why some BC have residual disease after chemo Rx that persists. These tumors tend to be p53 wild type and tend to undergo cell cycle arrest/senescence rather than cell death #chemotherapy #residualdisease

Looking at immune regulatory pathways he found PD-L1 and CD80 were both highly induced in tumors from chemotherapy treated mice. By RNA, many checkpoints were induced in senescent human BC and mouse mammary tumors after chemo (Fig.1)

Interestingly, PD-L1 (green) and CD80 (red) were expressed on different cell populations

sc seq analysis by our core lab @GotTcells gave us some hints about these populations. We found chemo could NOT induce PD-L1 in cell lines as we saw in vivo in tumors. Interferon gamma also failed to induce PD-L1 in proliferating cells, BUT, cells made senescent THEN IFNg=BOOM

by RNA, we observed some genes that could be induced by Doxo alone, and some (like PD-L1 aka Cd274) that require senescence THEN IFNg. Some genes like Stat1, only require IFNg and don't care if chemo is given first. Tumor transplants in IFNg KO mice did not fully induce PD-L1

So why do cells need chemo Rx to be responsive to IFNg?
We did ATAC seq on tumors and tissue culture cells and the loci that gain accessibility after chemo were overwhelmingly at Irf motifs

So, what does this chromatin accessibility mean? It means that IRF1 transcription factor can now access loci of genes like PD-L1 and Oasl2

Some genes, like Socs1 and Stat1, don't care about chemo. Their chromatin is accessible all the time, and all they need is IFNg

Here's the schematic of how Oasl2-like genes (those that require chemo induced senescence PLUS IFNg are induced

So what about that other population? CD80 expressing cells are characterized by high p53 signaling, CD80 requires only p53 to be induced, and p53 is at the CD80 promoter. p53 CRISPR ko cells made by s.meyers @SelinaSMJ_here could not induce CD80 after Rx

So we know that senescent tumor cells make a bunch of cytokines. These tumors must have lots going on with their immune cells right? Not so much! Very little change in immune cell populations post chemo

Let's see if we can target the PD-L1 + CD80 on tumor cells, and reinvigorate the immune system. We thus used anti PD1/CTLA4. First, an untreated mammary tumor had absolutely no response to immunotherapy (top), but the chemo Rx, senescesnt cells did respond, tho cures were rare

obvi this was disappointing to @ashkanshahbandi and he did become somewhat dysregulated

but further investigation showed we did get a massive accumulation of lymphocytes/Tcells in the tumor (almost 20 fold up) and a change in tumor architecture.

Interestingly, when we looked at gene expression ONLY in tumor cells of residual disease, in chemo +aPD1 Rx tumors, PD-L1 expression went down, consistent w/ these tumor cells now being eliminated, and in aCTLA4 Rx, PD-L1 expression went up. ICI Rx were hitting their target cells

Some take home messages: we need to examine the post chemo residual disease for expression of checkpoints and immune modulatory genes
Chemo and senescence rewires a tumor completely, gene expression is dramatically changed,& these cells acquire new properties, eg IFNg sensitivity

One other thing to note: if you are looking at sen genes in vitro, you are missing a LOT vs sen in vivo. IFNg produced by non-tumor cells drives expression of >100 genes including PD-L1, in sen TUMOR cells.

This took a while to complete and get published. @ashkanshahbandi presented a lot of this at @YicsaC @CellSenResearch seminar series on Zoom during covid back in Aug 2020!

I would be remiss if I did not note the support of the @NatureCancer editors and staff, including @alexia_zz @LisaHaas0107. Thanks for everything!