Pluripotent stem cell-derived CAR-macrophage cells with antigen-dependent anti-cancer cell functions

Recently, CAR-iPSC-differentiated CAR-expressing T cells and NK cells have been reported to have potent cytotoxic activity against cancer cells, and they represent a new family of engineered stem cell-derived immune cells for CAR therapies [1, 2]. Myeloid cells such as macrophages have been utilized as a type of effector cells to combat cancer cells by means of their phagocytosis function [3, 4]. However, immortalized macrophage cell lines are not applicable to clinical settings, and bone marrow or PBMC-derived primary macrophages are not efficiently engineered, thus leaving iPSC-derived macrophage cells as a great source for myeloid cell-based immunotherapy. Upon challenge with antigen-expressing cancer cells, CAR-expressing iPSC-induced macrophage (CAR-iMac) cells showed antigen-dependent macrophage functions. Expression of a CAR targeting tumor-associated antigen conferred CAR-iMac cells in vitro and in vivo anti-tumor effects.

First, we started from deriving iPSCs from peripheral blood mononuclear cells (PBMC) of a healthy donor with non-integrating episomal vectors encoding reprogramming factors (Fig. 1a), and isolated single iPSC clones (Additional file 1: Fig. S1). The materials and methods are shown in detail in the Additional file 2. Then, we compared different CD19 CARs and chose the conventional CAR to introduce into the iPSCs by lentiviral transduction (Fig. 1a and Additional file 1: Fig. S2a–g). Then, we established a protocol of myeloid/macrophage differentiation to induce CAR-iPSCs toward myeloid cell lineages (Additional file 1: Fig. S2h). Differentiated cells showed typical macrophage marker gene expression (Fig. 1b, Additional file 1: Fig. S4b–i). The products can be collected from 20 to 30 day for multiple times to allow further expansion to have a final yield of above 50-fold of the starting iPSCs (Additional file 1: Fig. S3a), with high purity indicated by ~ 100% of CD11b and CD14 expression at later days (Fig. 1b). Key macrophage marker genes were induced, whereas pluripotent marker genes disappeared (Fig. 1c; Additional file 3: Table S1). RNA-sequencing using differentiated cells showed that iPSCs clustered with precursor cells, and late-day differentiated cells clustered with primary macrophage cells, or untransduced iPSC-differentiated macrophage cells from previous studies [5, 6](Fig. 1d, e). GO analyses showed strong enrichment of innate immunity-related functions in 28-day differentiated cells (Fig. 1f; Additional file 4: Table S2).

Next, we further dissected their subpopulations by performing single-cell RNA-sequencing analysis. These cells clustered away from undifferentiated CAR-iPSCs (Fig. 1g), and they appeared to be largely homogenous with only a small number of cells not clustered with the main population (Fig. 1h). Blasting the differentiated single cells in a database of human cell atlas containing single-cell RNA-sequencing data revealed that these iMac cells mainly clustered with macrophages (Fig. 1i and Additional file 1: Fig. S6a). Trajectory analysis revealed that CAR-iMac cells went through a path from HSC to macrophage and DC cells without major branches (Fig. 1j, Additional file 1: Fig. S5c).

Moreover, all 10 clusters of differentiated cells showed strong signatures of M2 state [7‐10] (Fig. 1k, Additional file 1: Fig. S6b). We further compared the single-cell data in the 10 clusters with bulk RNA-seq data from the LPS/IFN-γ-polarized M1 cells or IL-4/IL-10-polarized M2 cells, by examining M1/M2-associated genes (Fig. 1l, Additional file 1: Fig. S5d), and found that most clusters were more similar to the M2 state particularly when using metabolism genes as markers [5, 6, 11‐15].

Next, we incubated the CAR (CD19)-iMac cells, CAR (meso)-iMac cells, or control iMac cells with CD19-expressing K562 leukemia cells or mesothelin-expressing OVCAR3/ASPC1 ovarian/pancreatic cancer cells. Compared with K562 alone, K562-CD19 cells were more likely to be phagocytosed by CAR (CD19)-iMacs (Fig. 2a, b), and compared with control cells, CAR (meso)-iMac showed increased phagocytosis activity against OVCAR3 and ASPC1 cells (Fig. 2g, h and Additional file 1: Fig. S7g). Intracellular signaling such as phosphorylation of ERK and NF-κB(P65) proteins were increased in CAR-iMacs co-cultured with CD19-expressing K562 cells compared to K562 cells, or to CAR-iMac cells cultured alone (Fig. 2c). We also examined cytokine gene expression in CAR(CD19)-iMac and CAR (meso)-iMac cells when they were incubated with tumor cells and found antigen-dependent increase in M1 pro-inflammatory cytokine expression(Fig. 2d, j and Additional file 1: Fig. S7h). Moreover, transcriptional analysis showed that CAR(CD19)-iMac cells and CAR(meso)-iMac cells showed strong enrichment of up-regulated genes in GO or KEGG terms of “positive regulation of cytokine secretion,” “antigen processing and presentation,” and “Toll-like receptor signaling pathway,” indicating these cells are more wired toward the pro-inflammatory state, when they encounter the antigen (Fig. 2e, f, i and Additional file 1: Fig. S7i).

When injected into NSG mice, these CAR-iMac cells expanded in vivo till around day 3 for about two fold, and persisted till more than 20 days and gradually disappeared after around 30 days (Fig. 2k). To test their anti-tumor cell activity, we first intraperitoneally injected ovarian cancer cells HO8910 expressing a luciferase gene into NSG mice. In order to further polarize CAR-iMac cells toward M1, we treated them with IFN-γ and washed IFN-γ away before injection (Additional file 1: Fig. S8). CAR(meso)-iMac-treated mice showed reduced tumor burden compared to the control group on day 4, 11, and 14 (Fig. 2l, m). These data demonstrate that the CAR confers anti-cancer cell activities in iMacs in vivo. The efficacy and persistency need to be further improved by designing more effective CARs or further modifying the CAR-iMac cells to stay constitutively active.