MPS1i-induced chromosomal instability generates microenvironment conditions that favor anti-cancer M1-like macrophages while minimizing skewing to a pro-cancer M2-like phenotype.

(A) Timeline and schematic for treatment of B16F10 mouse melanoma cells with MPS1 inhibitors (MPS1i). B16F10 cells were treated with MPS1i (reversine) or the equivalent volume of DMSO vehicle control. Cells were treated for 24 h, after which they were washed twice with PBS and allowed to recover for an additional 48 h. After the recovery period elapsed, numerous follow-up experiments were conducted to characterize chromosomal instability, aneuploidy, and effects on BMDMs (BMDMs).

(B) (i) Representative DNA images of B16F10 cells 72 h after initial treatment, as outlined in (A). Scale bars = 10 μm. Yellow arrowheads point at micronuclei, which are used as a signature readout of chromosomal instability. (ii) Quantification of the percentage of B16F10 cells with micronuclei after MPS1i treatment, across a range of different MPS1i concentrations and DMSO vehicle control (mean ± SEM shown, n = 3 replicates per condition). Statistical significance was calculated by ordinary one-way ANOVA and Tukey’s multiple comparison test (* p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001).

(C) Inferred copy number in DMSO and MPS1i-treated B16F10 cells from single-cell RNA-sequencing and the InferCNV pipeline. Cells that are considered aneuploid (labeled as outliers in the InferCNV algorithm) show full-level chromosome gains and/or losses compared to the consensus copy number profile of the DMSO-treated B16F10 population. Approximately 34% of MPS1i-treated B16F10 show aneuploidy, as determined by InferCNV, compared to 10% in the DMSO-treated B16F10 population. Single-cell RNA-sequencing confirms that MPS1i induces chromosomal instability and aneuploidy in B16F10 cells.

(D) Timeline and schematic for whole-tumor bulk RNA-sequencing. Prior to tumor inoculation, B16F10 CD47 KO cells were treated with 2.5 μM MPS1i (reversine) or the equivalent volume of DMSO vehicle control. Cells were treated for 24 h, after which they were washed twice with PBS and allowed to recover for an additional 48 h. After the recovery period elapsed, all mice were subcutaneously injected with 5x105 B16F10 cells. Five or ten days after tumor inoculation, mice were euthanized, and tumors were excised and disaggregated for bulk RNA-sequencing.

(E) Heatmap of selected RNA transcripts related to M2 macrophage polarization and pro-tumor function that were differentially expressed in tumors comprised of MPS1i-treated B16F10 compared to tumors comprised of their DMSO-treated counterpart. Heatmap shows log2-transformed transcript reads which were then z-score normalized. Generally, M2 and pro-tumor macrophage markers are downregulated at day 5 and even further at day 10 compared to DMSO controls. Tumors analyzed were two comprised of DMSO-treated B16F10 cells five days post-inoculation, two comprised of MPS1i-treated B16F10 cells five days post-inoculation, and two comprised of MPS1i-treated B16F10 cells ten days post-inoculation.

(F) Top downregulated hallmark gene sets in tumors comprised of MPS1i-treated B16F10 cells, showing downregulated DNA repair, cell cycle, and growth pathways, consistent with observations of slowed growth in culture and in vivo – as subsequently quantified. Gene set enrichment analysis pathways were obtained from MSigDB. Abbreviations: CL homeostasis = cholesterol homeostasis; OXPHOS = oxidative phosphorylation.

(G) Scheme for conditioned media (secretome) treatment of BMDMs and subsequent characterization by single-cell RNA-sequencing. Media from DMSO and MPS1i-treated B16F10 cells was collected and added 1:1 with fresh media to 7-day differentiated BMDMs (with 20 ng/mL M-CSF). After 24 h, BMDMs were processed for single-cell RNA-sequencing.

(H) (Left) UMAP plots of expression profiles for all analyzed BMDMs, treated with the secretome from DMSO-treated (gray) or MPS1i-treated (red) B16F10 cells. Each circle represents an individual cell. (Right) Same UMAP plots but colors now represent cells clustered together based on similarity of global gene expression. (Inset) Composition of each cluster.

(I) Dot plot showing proportion of cells in each cluster expressing anti-inflammatory and M2-like polarization-associated genes (pro-cancer gene set overall). Indicated to the right are a heatmap scale for average gene expression and a dot size reference for what proportion of cells in the cluster express a gene. Clusters 1 and 3, both of which consist of ∼75% of BMDMs treated with conditioned media from DMSO-treated B16F10, generally show downregulated transcript levels for genes associated with skewing macrophages to a pro-inflammatory, anti-cancer, M1-like phenotype. Cluster 3, in particular, shows significant downregulated. On the contrary, clusters 0 and 2, which consist of 75% of BMDMs treated with conditioned media from MPS1i-treated B16F10, show upregulated expression of many of these anti-cancer genes. Cluster 2, in particular, shows signatures of high upregulation.

(J) Dot plot showing proportion of cells in each cluster expressing pro-inflammatory and M1-like polarization-associated genes (anti-cancer gene set overall). Indicated to the right are a heatmap scale for average gene expression and a dot size reference for what proportion of cells in the cluster express a gene. Clusters 1 and 3, both of which consist of ∼75% of BMDMs treated with conditioned media from DMSO-treated B16F10, generally show upregulated transcript levels for genes associated with skewing macrophages to an anti-inflammatory, pro-cancer, M2-like phenotype. On the contrary, clusters 0 and 2, which consist of 75% of BMDMs treated with conditioned media from MPS1i-treated B16F10, show either little-to-no expression or downregulation of many of these genes.

Chromosomally unstable tumoroids and tumors in early stages show increased anti-cancer macrophage polarization and activity.

(A) Timeline and schematic for generating engineered “immuno-tumoroids” for time-lapsed studies of macrophage-mediated phagocytosis of cancer cells. Tumoroids are formed by plating and culturing B16F10 cells on non-adhesive surfaces in U-bottom shaped wells. Prior to plating for tumoroid formation, B16F10 were treated with either DMSO or MPS1i (reversine) as depicted in Fig. 1A. ∼24 h after plating, bone marrow-derived macrophages (BMDMs) with or without opsonizing anti-Tyrp1 are added to the cohesive B16F10 tumoroids. Immuno-tumoroids are imaged at the listed timepoints. At day 5, a separate subset of prepared immuno-tumoroids were collected, dissociated, and stained for measurement of macrophage polarization markers by flow cytometry.

(B) Representative flow cytometry histograms for flow cytometry measurements of macrophage polarization markers. For M1-like (anti-tumor) markers, we chose MHCII and CD86. For M2-like (pro-tumor) markers, we chose CD163 and CD206.

(C) Geometric mean fluorescence intensity quantification by flow cytometry of anti-tumor macrophage markers (MHCII and CD86) and pro-tumor macrophage markers (CD163 and CD206) in immuno-tumoroid cultures, with either DMSO or MPS1i-treated B16F10 CD47 KO cells. Generally, immuno-tumoroid cultures with MPS1i-treated B16F10 CD47 KO cells induce surface expression of M1-like, anti-tumor markers in macrophage while also reducing expression of M2-like, pro-tumoral markers. This suggests that, in addition to reduced cancer cell proliferation, chromosomal instability in cancer cells in early stages can prime a microenvironment conducive to anti-tumor macrophages. All data were collected from five independent immune-tumoroid culture experiments (one 96-well per replicate). Statistical significance was calculated by an unpaired two-tailed t-test with Welch’s correction (** p < 0.01).

(D) Representative fluorescence images depicting either growth or repression of B16F10 CD47 KO cells (green) in immuno-tumoroids from days 1-5. BMDMs, shown in magenta, were added at a 3:1 ratio to initial B16F10 numbers after the day 1 images were acquired. Scale bar = 0.5 mm.

(E) Tumoroid growth was measured by calculating the GFP+ area at the indicated timepoints (mean ± SD, n = 16 total tumoroids from two independent experiments for each condition). All data were then normalized to average GFP+ area on day 1 of each drug treatment’s respective mouse IgG2a isotype control condition. Overall, macrophages can clear MPS1i-treated B16F10 cells regardless of either MPS1i treatment concentration or IgG opsonization.

(F) Experimental timeline for immune cell infiltration analyses in tumors comprised of either DMSO or MPS1i-treated B16F10 CD47 KO cells. Prior to tumor inoculation, B16F10 CD47 KO cells were treated with 2.5 μM MPS1i (reversine) or the equivalent volume of DMSO vehicle control. Cells were treated for 24 h, after which they were washed twice with PBS and allowed to recover for an additional 48 h. After the recovery period elapsed, all C57BL/6 mice were subcutaneously injected with 2x105 B16F10 cells. Four days (96 h) post-challenge, mice were treated with 250 μg of anti-Tyrp1 or mouse IgG2a isotype. Mice were then euthanized 24 h after antibody treatment, and their tumors were excised and disaggregated for immune infiltrate analysis by flow cytometry.

(G) Immune cell infiltrate measurements of B16F10 CD47 KO tumors comprised of DMSO or MPS1i-treated cells. (i) Quantification of the percentage of CD45+ (immune) cells in the excised tumors, showing that tumors comprised of MPS1i-treated cells show ∼3-fold increased tumor immune cell infiltrate compared to their DMSO counterparts. (ii) Quantification of the percentage of tumor infiltrating myeloid cells in the excised tumors. B16F10 CD47 KO tumors comprised of MPS1i-treated cells show ∼2.5-fold more myeloid cells compared to their DMSO counterparts. (iii) Quantification of tumor infiltrating F4/80+ macrophages relative to the total number of tumor cells, with the tumors comprised of MPS1i-treated cells showing a ∼6-fold increase compared to their DMSO counterparts. (iv) Quantification of the number of M1-like macrophages (CD86hiCD206lo) and M2-like macrophages (CD86loCD206hi) relative to the total number of macrophages in each tumor condition. Tumors comprised of MPS1i-treated cells have ∼6-fold more CD86hiCD206lo compared to their DMSO counterparts. They also have a similar increase in CD86loCD206hi macrophages, but the number of CD86hiCD206lo macrophages was still overall twice as high. Mean ± SEM shown (n = 8 mice challenged with DMSO-treated B16F10 cells, n = 5 mice challenged with MPS1i-treated B16F10 cells). Statistical significance was calculated by an unpaired two-tailed t-test with Welch’s correction (ns, not significant; ** p< 0.01; **** p < 0.0001).

MPS1i-induced chromosomally unstable cancer cells are maximally cleared when IgG-opsonized and depleted of the CD47 macrophage checkpoint.

(A) Timeline for in vitro treatment B16F10 cells prior to injection in mice and then subsequent therapeutic treatment for tumor-challenged mice. Prior to tumor inoculation, B16F10 CD47 KO cells were treated with 2.5 μM MPS1i (reversine) or the equivalent volume of DMSO vehicle control. Cells were treated for 24 h, after which they were washed twice with PBS and allowed to recover for an additional 48 h. After the recovery period elapsed, all mice were subcutaneously injected with 2x105 B16F10 cells. For anti-Tyrp1 and mouse IgG2a isotype control treatments, mice were treated intravenously or intraperitoneally with 250 μg with antibody on days 4, 5, 7, 9, 11, 13, and 15 after tumor challenge.

(B) Tumor growth curve of projected tumor area versus days after tumor challenge, with B16F10 sgCtrl cells (expressing WT levels of CD47). Each line represents a separate tumor and is fit with an exponential growth equation: A = A0ekt. Experimental conditions are as follows: n = 9 mice that were challenged with DMSO-treated B16F10 sgCtrl and subsequently treated with mouse IgG2a control, n = 11 mice that were challenged with DMSO-treated B16F10 sgCtrl and subsequently treated with anti-Tyrp1, n = 6 mice that were challenge with MPS1i-treated B16F10 sgCtrl and subsequently treated with mouse IgG2a control, and n = 7 mice that were challenge with MPS1i-treated B16F10 sgCtrl and subsequently treated with anti-Tyrp1. All data were collected across three independent experiments. Inset bar graphs depict response type for each indicated tumor challenge. A partial response was defined as a mouse that survived at least one week (20+ days) beyond the median survival of the B16F10 sgCtrl cohort treated with mouse IgG2a isotype control (13 days).

(C) Tumor growth curve of projected tumor area versus days after tumor challenge, with B16F10 CD47 KO cells. Each line represents a separate tumor and is fit with an exponential growth equation: A = A0ekt. Complete anti-tumor responses in which a tumor never grew are depicted with the same symbol as their growing counterparts and with solid lines at A = 0. Experimental conditions are as follows: n = 7 mice that were challenged with DMSO-treated B16F10 CD47 KO and subsequently treated with mouse IgG2a control, n = 9 mice that were challenged with DMSO-treated B16F10 CD47 KO and subsequently treated with anti-Tyrp1, n = 14 mice that were challenge with MPS1i-treated B16F10 CD47 KO and subsequently treated with mouse IgG2a control, and n = 35 mice that were challenge with MPS1i-treated B16F10 CD47 KO and subsequently treated with anti-Tyrp1. All data in which mice were challenged with DMSO-treated cells were collected from three independent experiments. Data for the condition in which mice were challenged with MPS1i-challenged cells and then given mouse IgG2a control were collected from four independent experiments. Data for the final condition in which mice were injected with MPS1i-treated cells and then treated with anti-Tyrp1 was collected from seven independent experiments. Inset bar graphs depict response type for each indicated tumor challenge. A partial response was defined as a mouse that survived at least one week (20+ days) beyond the median survival of the B16F10 sgCtrl cohort treated with mouse IgG2a isotype control (13 days).

(D) Survival curves of mice up to 100 days after the tumor challenges in (B) and (C). (i) Survival curves for mice challenged with B16F10 sgCtrl. (ii) Survival curves for mice challenged with B16F10 CD47 KO. (iii) Triangular matrix depicting p-values between the different tested in vivo conditions. Statistical significance was determined by the Log-rank (Mantel-Cox) test.

(E) (i) Schematic depicting the different engineering anti-tumor macrophage strategies that can be used for validating macrophages’ role in clearing CIN-afflicted B16F10 cells. Fresh bone marrow was isolated from the tibia of donor C57BL/6 mice and was then incubated with anti-SIRPα clone P84 (18 μg/mL) or anti-Tyrp1 (1 μg/mL) only, both together added subsequently, or neither. Donor marrow (2x107 cells) is then injected intravenously into C57BL/6 harboring B16F10 WT tumors on day 4 post-challenge. Additional anti-Tyrp1 injections, when necessary, are done on days 5, 7, 9, 11, 13, and 15 post-tumor challenge. (ii) Survival curves of mice up to 100 days after the B16F10 WT tumor challenge. All mice were initially challenged with 2x105 B16F10 WT cells. (iii) Triangular matrix depicting p-values between the different tested in vivo conditions. Statistical significance was determined by the Log-rank (Mantel-Cox) test.

MPS1i-induced chromosomal instability favors induction of pro-phagocytic de novo IgG and can lead to durable acquired immunity.

(A) Schematic illustrating protocol for sera collection from surviving mice from Fig. 3C-D and follow-up experiments to characterize potential de novo anti-cancer IgG antibodies and their functionality both in vitro and in vivo. Serum from all mice was collected at least 100 days after initial tumor challenge.

(B) (i) Representative flow cytometry histograms showing that convalescent sera from survivors in Fig. 3C-D contain IgG2a/c (top) and IgG2b (bottom) that bind to both WT and Tyrp1 KO B16F10 cells.

(ii) Median fluorescence intensity quantification of IgG2a/c and IgG2b binding from sera from surviving mice. Convalescent sera show statistically significant increase in both IgG2a/c and IgG2b titers. Binding to even Tyrp1 KO cells suggests broader recognition of antigens unique to B16F10. Statistical significance was calculated by an unpaired two-sample Kolmogorov-Smirnov test (**** p < 0.0001). For IgG2a/c quantification: for binding against Tyrp1+ cells, n = 23 distinct naïve serum samples and n = 28 distinct convalescent serum samples from surviving mice; for binding against Tyrp1 KO cells, n = 23 distinct naïve serum samples and n = 27 distinct convalescent serum samples. For IgG2b quantification: for binding against Tyrp1+ cells, n = 14 distinct naïve serum samples and n = 16 distinct convalescent serum samples from surviving mice; for binding against Tyrp1 KO cells, n = 13 distinct naïve serum samples and n = 17 distinct convalescent serum samples.

(iii) Western blotting of B16F10 lysate with either naïve sera or first challenge survivor sera as primary probe followed by anti-mouse IgG [H+L] secondary staining. Numerous bands appear when immunoblotting with convalescent survivor sera (and more than when immunoblotting with naïve sera), qualitatively confirming binding to numerous antigens and suggesting acquired immunity beyond Tyrp1.

(C) (i) Phagocytosis of serum-opsonized CD47 KO or CD47/Tyrp1 double KO B16F10 cells by BMDMs on 2D tissue culture plastic. Additionally, B16F10 cells opsonized with either anti-Tyrp1 or mouse IgG2a were included as controls for comparisons. Serum IgG derived from survivors has both opsonization and pro-phagocytic functional ability against B16F10. Furthermore, convalescent sera IgG from survivors is still able to drive engulfment of Tyrp1 KO cells, further suggesting targeting of antigens beyond Tyrp1. Statistical significance was calculated by two-way ANOVA and Tukey’s multiple comparison test (mean ± SD, n = 21-23 distinct sera samples collected from survivors for B16F10 CD47 KO phagocytosis per condition and n = 14-23 for B16F10 CD47/Tyrp1 double KO phagocytosis per condition).

(ii) Convalescent sera from first challenge survivors can repress growth of B16F10 CD47 KO immuno-tumoroids (with macrophages). Tumoroid growth was measured by calculating the GFP+ area at the indicated timepoints (mean ± SD, n = 16 total tumoroids from two independent experiments for each condition, except n = 8 for opsonization with sera from naïve mice). Statistical significance was calculated by Brown-Forsythe and Welch ANOVA tests with Dunnett T3 corrections for multiple comparisons (ns, not significant; ** p < 0.01; **** p < 0.0001). Scale bars = 0.5 mm.

(D) B16F10 CD47 KO cells were pre-opsonized with either convalescent sera from first challenge survivors, anti-Tyrp1 or mouse IgG2a isotype control. All mice were subcutaneously injected with 2x105 pre-opsonized B16F10 CD47 KO cells.

(i) Tumor growth curves shown are at early timepoints where growth is still in the linear regime. Linear fits highlight growth suppression of tumors comprised of B16F10 CD47 KO cells pre-opsonized with convalescent sera and anti-Tyrp1, compared to the mouse IgG2a isotype counterparts. Mean ± SEM for all timepoints, with n = 16 mice with tumors pre-opsonized with convalescent sera (each from a distinct survivor), n = 9 mice with tumors pre-opsonized with anti-Tyrp1, and n = 14 mice with tumors pre-opsonized with mouse IgG2a isotype control. Statistical significance was calculated by ordinary one-way ANOVA and Tukey’s multiple comparison test at days 9, 11, and 13 (* p < 0.05; **** p < 0.0001). Significance represented in plot legend is representative of all three timepoints.

(ii) Survival curves up to 100 days of mice from (D-i) with pre-opsonized tumors. Both convalescent sera and anti-Tyrp1 provide similar survival benefits, suggesting potent de novo IgG opsonization and anti-cancer function. Statistical significance was determined by the Log-rank (Mantel-Cox) test (ns, not significant; ** p < 0.01; **** p < 0.0001).

(E) Schematic illustrating the series of experimental tumor challenges to assess acquired immunity. Survivors from the first challenge (Fig. 3C-D) were again challenged with either DMSO or MPS1i-treated B16F10 CD47 KO cells. Survivors from this second tumor challenge were once again challenged, this time with untreated B16F10 CD47 KO.

(F) Survival curves of survivors from Fig. 3C-D for a second tumor challenge experiment. Prior to tumor inoculation, B16F10 CD47 KO cells were treated with 2.5 μM MPS1i (reversine) or the equivalent volume of DMSO vehicle control. Cells were treated for 24 h, after which they were washed twice with PBS and allowed to recover for an additional 48 h. After the recovery period elapsed, all mice were subcutaneously injected with 2x105 B16F10 CD47 KO cells. Experimental conditions are as follows: n = 7 age-matched naïve mice (never tumor-challenged) injected with DMSO-treated B16F10 CD47 KO cells, n = 11 surviving mice (from Fig. 3C-D) injected DMSO-treated B16F10 CD47 KO cells, n = 6 age-matched naïve mice injected with MPS1i-treated B16F10 CD47 KO cells, and n = 8 surviving mice (from Fig. 3C-D) injected with MPS1i-treated B16F10 CD47 KO cells. Previous survivors challenged with DMSO-treated B16F10 CD47 KO cells show increased median survival (21 days) compared to their naïve counterpart (14 days). All previous survivors that were again challenged with MPS1i-treated B16F10 CD47 KO cells survive. All mice challenged were from three independent experiments. Statistical significance was determined by the Log-rank (Mantel-Cox) test (** p < 0.01; *** p < 0.001; **** p < 0.0001).

(G) Non-survivors from the second tumor challenge in (F) were euthanized after tumor size was >150 mm2, and their tumors were excised and disaggregated for immune infiltrate analysis by flow cytometry. (i) Quantification of CD45+ (immune) cells in the excised tumors, showing that first challenge survivors still show ∼2.5-fold increased tumor immune cell infiltrate despite reaching terminal burden. n = 4 mice for age-matched naïve control, n = 5 mice that survived the first tumor challenge. (ii) Quantification of tumor infiltrating CD8a+ cytotoxic T cell relative to the total number of tumor cells. First challenge survivors show ∼2-fold increase in CD8a+ T cells. n = 4 mice for age-matched naïve control, n = 3 mice that survived the first tumor challenge. (iii) Quantification of tumor infiltrating F4/80+ macrophages relative to the total number of tumor cells. First challenge survivors show ∼3-fold increase in macrophages. n = 4 mice for age-matched naïve control, n = 5 mice that survived the first tumor challenge. (iv) Quantification of MHCII+ tumor infiltrating F4/80+ macrophages relative to the total number of F4/80 macrophages. First challenge survivors show ∼3-fold increase in MHCII+ macrophages. n = 4 mice for age-matched naïve control, n = 5 mice that survived the first tumor challenge. For all experiments, mean ± SEM shown, and statistical significance was calculated by an unpaired two-tailed t-test with Welch’s correction (* p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001).

(H) (i) Third tumor challenge survival curves of long-term survivors from (G). All mice were challenged with 2x105 B16F10 CD47 KO cells (n = 9 mice from three independent experiments), delivered subcutaneously. For benchmarking and statistical comparison, survival curves from Fig. 3F for naïve mice (n = 7 mice from three independent experiments) and second challenge (n = 11 mice from three independent experiments) are included. Long-term survivors challenged a third time show ∼70% survival without any additional therapeutic modality, suggesting significantly improved acquired immune response. Statistical significance was determined by the Log-rank (Mantel-Cox) test (** p < 0.01; *** p < 0.001; **** p < 0.0001).

(ii) Individual tumor growth curves for third challenge in long-term survivors (n = 9 mice from three independent experiments) shown in (H-i). In total, 4 mice developed tumors, two of which had to be euthanized prematurely due to tumor rupture despite not reach a terminal burden of 125 mm2. These mice are still considered and included as casualties in the survival curve analysis. The two remaining mice show significantly slower tumor growth than naïve mice challenged with regular B16F10 CD47 KO (median survival of 14 days) and can be considered durable partial responders.

Characterization of MPS1i-induced genome and chromosomal instability in B16F10 mouse melanoma.

(A) Quantification of micronuclei induced by different MPS1 inhibitors (AZ3146, BAY 12-17389, and reversine) at different concentrations. Statistical significance was calculated by ordinary one-way ANOVA and Tukey’s multiple comparison test (* p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001).

(B) Quantification of micronuclei induced by MPS1i up to 12 days after initial exposure, showing MPS1i yields micronuclei-positive persistent populations while untreated cells fail to produce micronuclei.

(C) Consensus clustering of B16F10 analyzed in single-cell RNA-sequencing. B16F10 cells were treated with 2.5 μM MPS1i (reversine) or the equivalent volume of DMSO vehicle control. Cells were treated for 24 h, after which they were washed twice with PBS and allowed to recover for an additional 48 h. Cells were then collected and processes using 10X Genomics Chromium Single Cell Gene Expression kit for RNA isolation and library preparation.

(D) (i) UMAP plots of expression profiles for all B16F10 analyzed in single-cell RNA-sequencing. Top: UMAP plots depict clusters in which B16F10 fall, as determined by consensus clustering. Bottom: UMAP plots highlight cells with detectable copy number variations (CNVs), as determined by InferCNV. Red circles indicate individual cells with at least one CNV. Gray circles represent individual cells with no detectable CNV, suggesting these are chromosomally stable.

(ii) Quantification of the composition (MPS1i-treated or DMSO-treated B16F10) of each cluster.

Characterization of MPS1i-induced genome and chromosomal instability in B16F10 mouse melanoma.

(A) Flow cytometry analysis of CD47 and Tyrp1 expression on B16F10 mouse melanoma cells. Representative histograms for anti-CD47 (left) and anti-Tyrp1 (right) binding to B16F10 CD47 knockout (KO) and sgRNA ctrl (wild-type CD47 expression). Binding was detected by using secondary antibodies conjugated with Alexa Fluor 647.

(B) Quantification of Tyrp1 expression by flow cytometry in both MPS1i- and DMSO-treated cells. Statistical significance was calculated by three-way ANOVA and Tukey’s multiple comparison test. Mean ± SD shown, n = 3 replicates per condition (**** p < 0.0001). Quantification of CD47 expression by flow cytometry in both MPS1i- and DMSO-treated B16F10 sgCtrl cells. Statistical significance was calculated by three-way ANOVA and Tukey’s multiple comparison test. Mean ± SD shown, n = 3 replicates per condition (**** p < 0.0001).

(C) Quantification of cell size measurements by flow cytometry in both MPS1i- and DMSO-treated cells. Statistical significance was calculated by three-way ANOVA and Tukey’s multiple comparison test. Mean ± SD shown, n = 3 replicates per condition (* p < 0.05; ** p < 0.01; *** p < 0.001).

(D) Phagocytosis of B16F10 cells in in a standard phagocytosis assay on two-dimensional tissue culture plastic, with or without the following: IgG opsonization by anti-Tyrp1, CD47 KO, and MPS1i-treatment of B16F10. B16F10 cells were treated with MPS1i or DMSO prior to this assay, following the same protocol described in Fig. 1A. (i) Schematic of the 2D phagocytosis assay.

(ii) Representative images of phagocytosis. B16F10 cells (green) were incubated with opsonizing anti-Tyrp1 or mouse IgG2a control and added to adherent bone marrow-derived macrophages (magenta). Random fields were imaged and then quantified to determine the percentage of phagocytic macrophages. Yellow arrowheads denote phagocytic events (complete engulfment of a target B16F10 cells). (iii) Quantification of phagocytic macrophages. Statistical significance was calculated by three-way ANOVA and Tukey’s multiple comparison test. Mean ± SD shown, n = 3 replicates per condition (** p < 0.01; **** p < 0.0001).

Macrophages readily clear CIN-afflicted tumoroids but only if CIN is accompanied by proliferation deficits.

(A) Timeline and schematic for generating engineered “immuno-tumoroids” for time-lapsed studies of macrophage-mediated phagocytosis of cancer cells. Tumoroids are formed by plating and culturing B16F10 cells on non-adhesive surfaces in U-bottom shaped wells. Prior to plating for tumoroid formation, B16F10 were treated with either MPS1i or DMSO (AZ3146, BAY 12-17389 or reversine). Tumoroid growth was measured by calculating the GFP+ area at the indicated timepoints (mean ± SD, n = 16 total tumoroids from two independent experiments for each condition). All data were then normalized to average GFP+ area on day 1 of each drug treatment’s respective mouse IgG2a isotype control condition. Immuno-tumoroids are imaged at the listed timepoints. (B) To assess macrophage-mediated clearance, ∼24 h after plating, bone marrow-derived macrophages (BMDMs) with or without opsonizing anti-Tyrp1 are added to the cohesive B16F10 tumoroids. Tumoroid growth curves for DMSO-treated B16F10 CD47 knockout (KO) cells. (C) Tumoroid growth curves (with BMDMs) for MPS1i-treated B16F10 CD47 KO cells (treated with varying concentrations of MPS1i): (i) AZ3146, (ii) BAY 12-17389. (D) Tumoroid growth curves (without BMDMs) for MPS1i-treated B16F10 CD47 KO cells (treated with varying concentrations of MPS1i): (i) BAY 12-17389, (ii) reversine. Both BAY 12-17389 and reversine induce proliferation deficits, which may play an essential role in mediating clearance by macrophages that could not be observed in AZ3146-treated cells.

MPS1i-induced chromosomal instability upregulates MHC-1 class molecules on B16F10, suggesting increased antigen presentation.

Flow cytometry analysis and quantification H-2Kb expression on B16F10 mouse melanoma cells. B16F10 cells were treated with MPS1i or the equivalent volume of DMSO vehicle control. Concentrations used: 2.5 μM AZ3146, 1.0 μM BAY 12-17380, and 2.5 μM reversine. Cells were treated for 24 h, after which they were washed twice with PBS and allowed to recover for an additional 48 h. Binding was detected by using primary antibodies conjugated with Alexa Fluor 647. Statistical significance was calculated by Brown-Forsythe and Welch ANOVA and Dunnett’s T3 multiple comparison test. Mean ± SEM shown, n = 3 independent replicates (* p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001).

Flow cytometry gating strategy for identification & quantification of macrophage infiltrate and characterization in CIN-afflicted and chromosomally stable B16F10 tumors.

Representative flow cytometry gating strategy for in vivo B16F10 CD47 KO tumor immune infiltrate five days after initial challenge. Tumors were comprised of cells treated with either (i) DMSO or (ii) MPS1i (2.5 μM), following the timeline schema in Fig. 2F. Singlets were separated from debris, doublets, and aggregates by FSC-A vs FSC-H and SSC-A vs SSC-H gates. Debris and dead cells were further removed by FSC-A vs SSC-A gating. Overall immune cell infiltrate was determined by CD45+ expression. Macrophages were then further identified based on Ly6G- and F4/80+ expression. Upon identification of macrophages, cells were further characterized for polarization based on CD86 (M1-like marker) and CD206 (M2-like marker) expression.

MPS1i-induced CIN favors clearance when paired with CD47 KO and IgG opsonization, regardless of the degree of CIN.

Tumor growth curve of projected tumor area versus days after tumor challenge, with B16F10 CD47 KO cells. Each line represents a separate tumor and is fit with an exponential growth equation: A = A0ekt. Complete anti-tumor responses in which a tumor never grew are depicted with the same symbol as their growing counterparts and with solid lines at A = 0.

(A) Tumor growth curve data in which there was no complete response, from Fig. 3C. n = 7 mice that were challenged with DMSO-treated B16F10 CD47 KO and subsequently treated with mouse IgG2a control, n = 9 mice that were challenged with DMSO-treated B16F10 CD47 KO and subsequently treated with anti-Tyrp1.

(B) Tumor growth curve data from mice challenges with CIN-afflicted B16F10, but with varying concentration of MPS1i used to cause different degrees of instability. All mice here were challenged with B16F10 CD47 KO and treated with anti-Tyrp1 for conditions of maximal phagocytosis and to better compare to the ∼97% cure rate in Fig. 3C, 4D-ii. n = 6 mice challenged with cells with each tested MPS1i concentration: 0.1, 0.5, 1.0 or 2.5 μM.

(C) Survival curves of mice up to 100 days after the tumor challenges in (A) and (B).

Survivors challenged with CIN-afflicted tumors generate anti-cancer IgG, regardless of the degree of CIN.

Median fluorescence intensity quantification of IgG2a/c and IgG2b binding from sera from surviving mice with different degrees of chromosomal instability by titrating reversine concentration used to treat B16F10 CD47 KO cells before injection. Convalescent sera are collected from survivors from Fig. S6. Binding of IgG2a/c and IgG2b against (A) B16F10 cells expressing Tyrp1 and (B) B16F10 Tyrp1 knockout (KO). n = 4-5 distinct sera samples collected from survivors.

Flow cytometry gating strategy for identification & quantification of immune infiltrate and characterization in re-challenge experiments.

Representative flow cytometry gating strategy for in vivo B16F10 CD47 KO tumor immune infiltrate from 2nd challenge non-survivors and age-matched naïve controls. Singlets were separated from debris, doublets, and aggregates by FSC-A vs FSC-H and SSC-A vs SSC-H gates. Debris and dead cells were further removed by FSC-A vs SSC-A gating. For T cell quantification: CD45+ cells were gated on and then analyzed for CD3e and CD8a surface protein expression. For macrophage quantification: myeloid cells were isolated by CD45+ and CD11b+ expression, after which macrophages were isolated based on F4/80 expression.

Growth of CIN-afflicted wild-type (WT) tumors in T- and B-cell deficient mice and T- and B-cell replete mice.

Similar growth delays are found for MPS1i-pretreated B16F10 cells in T- and B-cell deficient NSG mice and immunocompetent C57BL/6 mice. Both types of mice have functional macrophages. Parallel studies in vivo were done with WT B16F10 ctrl cells cultured 24 h in 2.5 μM MPS1i (reversine or DMSO, then washed 3x in growth media for 5 min each and allowed to recover in growth media for 48 h. 200,000 cells in 100 uL PBS were injected subcutaneously into right flanks, and the standard size limit was used to determine survival curves. The C57BL/6 experiments were done independently here (by co-author L.J.D.) from the similar results (by B.H.H.) shown in Fig.3D-i, which provides evidence of reproducibility.