@MeWantData @JswLab @xochitl__luna @CorradinLab @WhiteheadInst @KarlHerrup @martinowk @tuuliel @ophirshalem @UPenn_SongMing @alicechenp @SashaGusevPosts @panos_roussos @doctorcorces @johnomix @sj_marzi @TNRLab @WernigLab @ckgucsd @nevillesanjana @towfiqueRaj @FChrisBennett @CreminsLab @kristenbrennand.

@MeWantData @JswLab @xochitl__luna @CorradinLab @WhiteheadInst Would love to get feedback from the community #neuroinflammation #multiplesclerosis #Alzheimer #microglia #iPSC #CRISPR #Perturbseq #GWAS #epigenome #neurodegeneration.

@MeWantData @JswLab Importantly, this project was co-led by @xochitl__luna and was a huge group effort between the Jaenisch and @CorradinLab and all the core facilities @WhiteheadInst.

Please find slide 16 somewhere around here, sorry for the mess up!.

@MeWantData @JswLab 37/ we're currently looking more closely at the remaining 4 loci, and are performing functional experiments to see if we can figure out how these genes might be altering microglial phenotypes in MS and related neuroimmune conditions. Paper link:

@MeWantData @JswLab 35/ we identified the likely causal risk genes at 5/9 MS risk loci, and found that when perturbing these genes, downstream affected pathways were strongly enriched for MS heritability, suggesting regulatory cross-talk between MS risk mechanisms in microglia.

@MeWantData @JswLab 34/ in summary, we took a broad look at iPSC-neurons and glia to assess their ability to study GWAS risk SNP mechanisms, used outside variant analysis to predict pathogenic cell types at specific loci, and developed a new Perturb-seq platform for microglia to test our predictions.

@MeWantData @JswLab 33/ pathways when perturbing the causal effector genes at individual MS risk loci.

@MeWantData @JswLab 32/ Finally, we compared these DEGs with those seen in MS patient brain tissue, including microglia from gray matter, white matter, active/inactive lesions, etc. We found a significant overlap, suggesting that iPSC-microglia monoculture can actually recapitulate some MS-relevant.

@MeWantData @JswLab 31/ pathways were upregulated.

@MeWantData @JswLab 30/ we also find striking concordance between DEGs across these 5 loci - many genes were shared and directionally concordant. Homeostatic microglia pathways such as GTPase activity were downregulated in cells with MS risk locus perturbation, whereas oxphos and other metabolic.

@MeWantData @JswLab 29/ genes displayed trans DEGs (vast majority on other chromosomes) that were enriched for MS heritability! This wasn't just due to altering microglia genes, as AD h2g was not as enriched in these genomic regions. The effect was not dependent on the cis target gene or HLA locus

@MeWantData @JswLab 28/ as we had scRNA-seq data, we further looked to see if downstream (trans) DEGs could shed light onto MS pathophysiology. What we found was quite surprising - among the 5 loci where we identified the SNP target genes, we found that cells containing sgRNA perturbations of these.

@MeWantData @JswLab 27/ and three other loci: CD86, IRF5, and SLC25A19, the latter of which we show GRB2 as the likely SNP target gene

@MeWantData @JswLab 26/ complement C3, which is regulated by upstream enhancers with MS SNPs, but no effect on TNFSF14 which is actually closer to the GWAS signal.

@MeWantData @JswLab 25/ we identified the causal MS risk gene at 5/9 loci, including MERTK, a well know regulatory of phagocytosis.

@MeWantData @JswLab 24/ when we performed the screen, the microglia were very homogeneous, and even lacked markers of highly similar border-associated macrophages. We used MOI of 1.5, sequenced ~120K cells with ~10B read pairs, got ~400 cells/guide and ~2K genes/cell.

@MeWantData @JswLab 23/ this was the hardest part - we applied the Vpx lentiviral system to a Perturb-seq vector to allow for efficient, non-toxic, minimally-activating CRISPRi in microglia. We used dCas9-KRAB-expressing iPSCs with the ZIM3 KRAB domain for maximal knockdown.

@MeWantData 22/ we selected 9 loci we predicted would act through microglia for CRISPRi single cell screening (Perturb-seq), with help from @JswLab. The goal here was to perturb all candidate functional SNPs and see if they converged on a causal risk gene (or genes)

@MeWantData 21/ a really interesting locus if WWOX/MAF, which has both AD and MS GWAS SNPs, as well as MS outside variants, overlapping microglial enhancers spanning 500kb. PLAC-seq in ex vivo microglia and PCHi-C in iPSC-microglia really suggest MAF as the target gene here