Read on Twitter
Happy to share our latest work on tumor-infiltrating dendritic cells (DC)! https://doi.org/10.1084/jem.20200264 We know these cells can play key antitumor functions, but they require more attention and could become important targets for cancer #immunotherapy. @MGHCSB @MGH_RI @harvardmed 1/n
Multiple tumor-infiltrating DC states have been identified by scRNAseq in human cancers. But are these DC states conserved across cancer indications, and between humans and mice? If so, do they have identifiable gene signatures? And what are their origins and functions? 2/n
To answer these questions, we did a meta-analysis of DC states across solid human cancer-types and reviewed their origins and functions. Out today in @JEM Led by - @gengerhard @RubenBill5 @MariusMessemak1 in collaboration with @KleinLabHMS 3/n
Comparison of tumor-infiltrating DC states across scRNAseq studies is complicated by the fact that different investigators may have used varying experimental protocols and bioinformatic approaches, and given distinctive names to observed DC states. 4/n
We defined a whole-transcriptome reciprocal similarity score for each DC state comparison pair that quickly vanishes to zero if states do not show mutual correspondence. This is analogous to âreciprocal best hitâ approaches used to identify orthologous genes between species. 5/n
We identified 5 conserved tumor-infiltrating DC states: pDC, cDC1, cDC2, cDC2/Monocyte-like DC (cDC2/MoDC; also called DC3 in https://doi.org/10.1016/j.immuni.2020.06.002) and DC3 (also called mregDC in https://doi.org/10.1038/s41586-020-2134-y), LAMP3+ DC, or CCR7+ DC). 6/n
The 5 tumor-infiltrating DC states show a conserved gene signature across cancer types. Top enriched genes include marker genes that have been used to define DC states but also new genes that warrant investigation! 7/n
The tumor-infiltrating cDC1, cDC2, DC3, and pDC states are distinct from their monocyte and macrophage counterparts, whereas the cDC2/MoDC state shows similarities with these cells. 8/n
Where do tumor-infiltrating DC states originate from? cDC1, pDC, cDC2, and cDC2/MoDC states have their counterparts in blood, suggesting these cells are defined phenotypically before they enter the tumor... 9/n
The DC3 state is specific to tissue without counterpart in blood â likely produced locally. RNA-velocity results suggest that DC3 may derive from either the cDC1, cDC2/MoDC or possibly cDC2. 10/n
What are the functions of tumor-infiltrating DC states? The roles of all five tumor-infiltrating DC states is being unraveled by many research groups. Below is a quick recap (but there is much more info in the paper!): 11/n
cDC1: Can recruit and locally (re-)activate antitumor CD8+ T cells. Required for successful immunotherapy in various murine mouse models. Accordingly, cDC1 gene-signatures correlate with improved patient survival and clinical response to immunotherapy. 12/n
cDC2 and cDC2/MoDC: Classical gating strategies may not distinguish these cell states. As such, these cells require further functional investigation. They likely can regulate various types of T cell responses. 13/n
DC3s: Required for successful cancer immunotherapy in mice ( https://doi.org/10.1016/j.immuni.2018.09.024 ). Can license T cells to eliminate tumors. Can also be immunosuppressed ( https://doi.org/10.1038/s41586-020-2134-y). 14/n
pDC: May promote antitumor immunity (e.g., via interferons) but may also promote cancer growth by activating Tregs. More needs to be done to understand these cells better . 15/n
The DC states discussed here likely all contribute to antitumor immunity, but remain underinvestigated. Also, tumor immune evasion likely involves suppressing DC functions. Canât wait to learn more about these cells and hopefully to harness them successfully in therapy! 16/n
Importantly, the overall conservation of tumor-infiltrating DC states across cancers should facilitate their investigation, and findings made in the context of a given cancer may be relatively generalizable. #Cancer #Immunotherapy 17/n
Thank you to @MiriamMerad, Zemin Zhang ( http://cancer-pku.cn/ ) and @LambrechtsDlab for contributing scRNAseq data from different human cancer types to our meta-analysis. And a big thank you also to scRNAseq master @RapolasZilionis ...end
...end
And here is a better link to access the paper! (the link listed above doesn't seem to be active yet--thanks for letting me know) https://rupress.org/jem/article/218/1/e20200264/211613/Tumor-infiltrating-dendritic-cell-states-are#.X90S7W9Z8Cs.twitter @JExpMed
