In the ERC starting project CANCER-DC, we took advantage of our pioneer work in genome editing aiming to develop a new methodology that improves immune cell function and supports resistance to immunosuppressive tumor microenvironment (TME). We applied the methodology that we developed to target and improve the function of dendritic cells (DCs), innate immune cells that mediate the activation of the adaptive immune response. In the absence of proper DC maturation, the adaptive immune response is not effective.
We investigated cancer types in which immune cells incorporate the tumor mass but immunotherapy treatments are not beneficial to the patients. We hypothesize that modifications of immune cells to bypass the effect of immunosuppression in these cancer types can lead to effective treatments. We therefore chose to focus on ovarian cancer and pancreatic cancer, two aggressive cancer types with low survival rates.
To characterize the immunosuppressive signals that alter immune cell function.
We performed scRNA-seq of early lesions and tumors from mice and humans. We have used five different mouse models including genetic models and orthotopic injection of cells to the pancreas or ovary. In total, we have sequenced more than 350,000 cells.
The extensive characterization of cells that are associated with tumors allowed us to model how immunosuppressive TME is formed from early lesions till tumor and characterize the amount, gene expression programs, cell states, and cell-cell interactions of all the immune cells along the malignant process. We found that innate immune cells start to express immunosuppressive genes such as IL10 and TGFB and that T cells express PD-l1 and LAG3 at a very early stage before tumor formation.
We explored and examined which interactions and signaling molecules contribute to immune dysfunction and alter proper response against cancer.
These results were published in Nature Communications 2020 Sep 9;11(1):4516 and we are preparing an additional manuscript to compare the immune cell states between different cancer types.
In the next part of the project, we screen for genes that can improve DC function in the presence of suppressive signals that we detected in the first project.
To find genes that regulate DC function.
To find key regulatory genes that control DC function we conducted a set of genome-wide (GW) CRISPR screens in bone marrow differentiated dendritic cells (BMDCs). CRISPR screen is a powerful tool that allows one to find a set of genes that regulate a phenotype of interest.
We found that CD86, which is expressed by DCs and binds to CD28 on T cells to allow their transition from naïve to effector cells, is sensitive to immunosuppressive signals. Specifically, IL10 or incubation of BMDCs with cancer cells reduces the level of CD86. We also found that CD86 is not only essential for T cell priming but it is a rate-limiting factor. To reveal negative regulators of CD86, we performed three GW CRISPR screens, without any suppressive signals, in the presence of IL10, and after co-incubation with cancer cell line. In each condition, we also performed a secondary screen that was based on the GW screen, to reduce the number of false positive and false negative results. The last stage includes extensive validations to test which of the targeted genes indeed enhances the level of CD86. We used 2-3 gRNAs for more than 70 genes and could validate a significant enhancement of CD86 in 30 genes. Half of the genes that we validated were not known as regulators of CD86 before. In many cases, enhanced DC maturation induces the expression of immunosuppressive receptors to balance the response. To exclude such genes we performed another GW screen this time searching for positive regulators of PD-L1. We followed with a secondary screen and single gRNA validations. We validated 7 positive regulators of PD-L1 that were not reported before. Together, we found negative regulators of CD86 and positive regulators of PD-L1. We chose these receptors because they play a major role in T cell priming but in addition, they represent the activation program (CD86) and the suppressive program (PD-L1) of DCs. Thus, targeting the regulatory genes that we found can optimize DC mode of action in hostile TME.
These novel regulators of DC activation have a high potential to produce an effective immune response but these results also pose two major challenges: (i) Each regulator that was validated can affect hundreds of genes, and therefore the overall effect of each perturbation on genomic circuits and cell function cannot be determined from the screen results. It brings up the question of which of all the validated genes it is most beneficial to target. (ii) In addition, it may be that targeting several genes in the same cell can lead to better results depending on the genetic interactions between these genes.
To meet these challenges, we performed a high-order perturb-seq experiment. We generated a library of gRNA that targets 11 genes that had the desired phenotype and infected BMDCs at high multiplicity of infection to include 1-2 gRNA in each cell. We then performed scRNA-seq of perturbed cells. This experiment allows us to reveal the effect of each perturbation and each combination of perturbations on the gene expression program and protein level of CD86 and PD-L1. Based on this experiment we could expose the genetic interactions between the genes that we found and we discovered that targeting Cebpb – an enhancer binding protein and Med12 a subunit of the mediator complex in the same cell is predicted to modify DC function optimally.
In summary, we set a novel methodology that supports the finding of combinations of genes to target to improve immune cell function starting from screening all the genes in the genome. We named this new methodology High Multiplicity of Perturbations and Cellular Indexing of Transcriptomes and Epitopes (HMPCITE-seq) and we published a related paper in Nature Communications. 2023 Oct 9;14(1):6295.
To explore the effect of the combination of KO in Cebpb and Med12 on DC function.
Next, we investigated how the perturbations that we discover affect the ability of DCs to induce T cell priming and activation and how it affects tumor growth. We conducted a large set of experiments in vitro and in vivo and found that CD8 T cells that co-incubate with Cebpb Med12 BMDCs proliferate faster and secrete more perferin and granzyme B. We also injected BMDCs with KO in Cebpb to tumor-bearing mice and could measure a reduction in tumor size. Thus, our methodology allows us to rewire immune cells and improve their function in the presence of immunosuppressive signals.
