IFN-γ induces hematopoietic stem cell myelopoiesis through Meis1 in tumor

Abstract

Background

The role of inflammation-induced myeloid-biased hematopoiesis in driving resistance to immune checkpoint blockade (ICB) is recognized, yet the intricate mechanisms through which tumors orchestrate it are not fully defined.

Methods

MC38 tumor and Lewis lung cancer models were performed to evaluate hematopoietic stem cells (HSCs) differentiation biased. Key pro-inflammatory cytokines implicated in this process were screened through ELISA assay and bioinformatic analysis. Subsequent mechanistic investigations identified the central transcription factor governing tumor-induced myeloid-biased differentiation of HSCs. To demonstrate the functional impact on antitumor immunity, we quantified HSC-derived myeloid-derived suppressor cells (MDSCs) and assessed their suppressive effects on T cell function. Furthermore, the therapeutic potential of targeting this axis was evaluated using Emapalumab, an anti-IFN-γ antibody, to determine whether suppressing myeloid-biased HSCs could enhance the antitumor effects of ICB.

Results

Here, we found HSCs exhibit a persistent myeloid-biased differentiation phenotype in MC38 tumor and Lewis lung cancer models, which was induced by the pro-inflammatory cytokines IFN-γ. Transcriptional profiling indicated Meis homeobox 1 (Meis1) was enriched in tumor primed HSCs, and ablation of Meis1 in HSCs prevented HSCs-associated myeloid cell differentiation. The resulting HSC-derived MDSCs were identified as key factors of tumor progression. Therapeutic targeting of the myeloid differentiation axis with a combination of anti-PD-1 antibody and Emapalumab, an anti-IFN-γ antibody inhibited HSC-derived MDSCs production and enhanced T cells-mediated adaptive immunity to suppress tumor progression.

Conclusions

Our results highlight HSC-directed therapy as a novel approach for cancer treatment. Combining anti-PD-1 with Emapalumab potently enhances the response to ICB, offering a promising strategy to achieve superior and durable anticancer efficacy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13287-026-04968-9.

Introduction

The hematopoietic stem cells (HSCs) represent a rare population of quiescent and multi-lineage producing stem cells that persist throughout an individual’s lifespan [1]. The presence of external stimuli can rapidly induce partial HSCs to enhance immune cell production, which consequently leads to function attenuation and eventual exhaustion of stem cell [2, 3]. Both acute insults—hemorrhage, infection—and chronic inflammatory states trigger this emergency myelopoiesis, with mature myeloid cells being the dominant product [4–7]. This emergency hematopoiesis is triggered by injuries through both direct and indirect mechanisms [8]. Further investigations have demonstrated that pro-inflammatory cytokines, including interferon (IFN)-α, IFN-γ, granulocyte colony-stimulating factor (G-CSF), thrombopoietin (TPO), interleukin-1 (IL-1) and tumor necrosis factor alpha (TNF-α), play important roles in inducing inflammatory response of HSCs [9–15].

Systematic inflammatory stimuli can train HSCs to a long-term myelopoiesis to initiate trained innate immunity [16–18]. In response to pro-inflammatory cytokines, HSCs become active to facilitate rapid myelopoiesis, leading to increased levels of lineage-biased multipotent progenitors (MPPs) and lineage-committed myeloid progenitor cells [19–22]. The MPP compartments indicate an overproduction of megakaryocytic/erythroid-biased MPP2s and granulocyte/macrophage-biased MPP3s, as well as limited presence of lymphoid-biased MPP4s [23, 24]. Although inflammation has been established as a critical inducer of cancer, its impact on HSC function and lineage commitment during tumorigenesis remains unclear.

The progression of tumors, such as colon cancer and lung cancer, are often associated with chronic inflammation and can be obviously retarded by immune checkpoint blockade (ICB) therapy [25–27]. A high myeloid-to-lymphoid ratio is frequently observed in cancer patients and is associated with a poorer response to immunotherapy [28, 29]. During tumor progression, cancer cells can directly reshape the HSC niche, thereby playing a pivotal role in reprogramming host immunity. In melanoma, tumor-derived IL-3 signaling drives myeloid-biased hematopoiesis in the bone marrow, resulting in increased myeloid-derived suppressor cells (MDSCs), macrophages, and dendritic cells, and a concurrent reduction in mature erythrocytes, platelets, and differentiated B cells [30]. Tumor cells further reshape the spatial architecture of the bone-marrow niche, relocating HSCs to positions in closer proximity to mesenchymal stem cells (MSCs), which in turn drive HSC commitment toward the myeloid lineage [31]. Furthermore, previous studies have showed that myeloid cells, particularly myeloid-derived suppressor cells (MDSCs), can suppress T cells activity to facilitate tumor development [32, 33]. This suggests that inflammation-driven alterations in HSCs function may contribute to both tumor and reduced sensitivity to immunotherapy. Our study uncovered a critical role of bone marrow IFN-γ in promoting myeloid-bias of HSCs in colorectal cancer and lung cancer via a Meis-1-dependent mechanism. Importantly, we demonstrated that combination therapy with anti-IFN-γ antibody enhance the anti-tumor efficacy of ICB theray in models of colorectal cancer and lung cancer, offering a promising new therapeutic strategy for these malignancies.

Materials and methods

Mice

C57BL/6Smoc-Meis1^{tm1(flox)Smoc} mice were purchased from the Shanghai Model Organisms Center. Tg(Vav1-icre)A2Kio/J mice were purchased from The Jackson Laboratory. Scl-creERT; tdTomato mice were intraperitoneally injected with 4 mg/mice Tamoxifen in 200 µl corn oil as indicated. Genotypes were confirmed by PCR on tail DNA [34]. All mouse strains used in this study were in C57BL/6J genetic background. All animal protocols were approved by the Institutional Animal Care and Use Committee of the Guangzhou First People’s Hospital. The study protocol was approved by the Experimental Animal Ethics Committee of the Hospital Clinico San Carlos of Madrid and was in line with ARRIVE guidelines 2.0. All experiments were initiated when mice were between 6 and 8 weeks of age, which were a mix of male and female. The number of animals used per experiment is shown in the figure legend. The mice were anaesthetized by isoflurane 2% inhalation during the experimental procedures, and euthanized through CO₂ exposure at the end of the study.

Tumor implantation

To establish a tumor model, 5 × 10⁵ MC38 or lewis cells were subcutaneously injected into mice, and tumor volume measurement began five days post-injection. Humane end points for tumour induction studies were inability to breathe, eat, drink or move normally, behavioural abnormalities, tumour size > 1,000mm³, tumour ulceration and loss of body weight greater than 20%. Maximal tumour size was not exceeded before the end point of experiments in this study [35].

Flow cytometry and cell sorting

Single-cell suspensions from mouse femurs, tibias, or spleens were harvested where indicated. Red blood cells were lysed before antibody staining. For FACS analysis the following antibodies were used including, lineage cocktail [anti-CD3ɛ (clone 145-2C11), anti-CD4 (clone RM4-5), anti-CD8a (clone 53 − 6.7), anti-B220 (clone RA3-6B2), anti-Mac-1 (clone M1/70), anti-Gr-1 (clone RB6-8C5), anti-IgM (clone RMM-1), anti-Ter119 (clone TER-119)]-Pe-Cy5, anti-Sca-1-Pe-Cy7 (clone D7), anti-C-kit-APC (clone 2B8), anti-CD34-FITC (clone RAM34), anti-Flk2-BV421 (clone A2F10), anti-CD3-Apc-Cy7 (clone 17A2), anti-CD4-eFluor450 (clone RM4-5), anti-CD8a-FITC (clone 53 − 6.7), anti-CD45.1-BV605 (clone A20), anti-CD45.1-Pe-Cy5 (clone A20), anti-CD45.2-AF700 (clone 104), anti-CD45.2-BV421 (clone 104), anti-Mac-1-Pe-Cy7 (clone M1/70), anti-Gr-1-Apc-Cy7 (clone RB6-8C5), anti-CD3-APC (clone 17A2), anti-B220-eFluro450 (clone RA3-6B2), anti-CD4-BV605 (clone RM4-5), anti-CD8a-AF700 (clone 53 − 6.7), anti-CD127-Apc-Cy7 (clone A7R34), anti-F4/80-APC (clone BM8), anti-CD11C-Pacific Blue (clone N418), anti-Ly6C-FITC (clone HK1.4), anti-Ly6G-PE (clone 1A8). FACS analysis was performed by Attune (Thermo Fisher Scientific) or the full spectral flow cytometry Aurora (Cytek) and analyzed by FlowJo (Tree Star) or SpectroFlo (Cytek). Cell sorting was performed on MoFlo Astrios EQs (Beckman).

HSC competitive transplantation

On the 7th day after PBS, MC38 or Lewis injection, 500 HSCs from wild type (WT), Scl-creERT2;tdTomato or Vav-cre; Meis1^flox/flox (CD45.2⁺) mice were sorted and transplanted to lethally irradiated (9 Gy) CD45.1⁺ recipients together with 2 × 10⁵ CD45.1⁺ bone marrow cells.

HSCs ex vivo culture

HSCs from PBS, MC38 or Lewis cells treated WT mice and Vav-cre; Meis1^flox/flox mice were sorted into 96-well U-bottom plates at 200 cells per well in 200 µl StemSpan SFEM medium (9650, StemCell Technologies). The medium was supplemented with 10 µg ml^–1 heparin (H3149, Sigma), 10 ng ml^–1 recombinant mouse SCF (C775, Novoprotein), 100 ng ml^–1 recombinant mouse TPO (CP40, Novoprotein), and 1% penicillin/streptomycin (SV30010, HyClone). Emapalumab (100 µg ml^{− 1}, HY-P99191, MedChemExpress) [36], IFN-γ (20 ng ml^{− 1}, HY-P7071, MedChemExpress) [12], and Meis1 inhibitor (1 µM, HY-132870, MedChemExpress) [37] was added during culture at indicated time.

T cell in vitro stimulation

T cells were collected from spleens in mice where indicated. For intracellular cytokine staining, 2 × 10⁶ cells per well in a 96-well U-bottom plate were cultured in 200 µl of RPMI 1640 medium (11875093, Gibco) and stimulated with PMA (25 ng ml^–1, P8139, Sigma) and ionomycin (500 ng ml^–1, 407952, Sigma) at 37 °C for 6 h, and subsequently treated with Brefeldin A (10 mg ml^–1, B5936, Sigma) for the last 2 h. Intracellular cytokine staining was performed with the FOXP3/Transcription Factor Fixation/Permeabilization Concentrate and Diluent kit (00–5523-00, eBioscience) according to the manufacturer’s instructions. For T cells analysis, anti-CD4-eFluor450 (clone RM4-5), anti-CD8a-FITC (clone 53 − 6.7), anti-IFNγ-AF488 (clone XMG1.2), anti-IL2-PE (clone JES6-5H4), and anti-IL4-Pe-Cy7 (clone 11B11) were stained.

Cell culture

The murine colon cancer cell line (MC38) was obtained from National Infrastructure of Cell Line Resource (Beijing, China). The cells were cultured in DMEM medium supplemented with 10% FBS. 5 × 10⁵ MC38 cells in 200 µl PBS were subcutaneously implanted into mice. For tumor growth, the tumor volume was monitored every 2 days between day 7 to day 21 after tumor implantation.

Western blotting

For western blotting, the anti-STAT1 (rabbit, 1:1000, A19563, ABclonal Technology); anti-GAPDH (rabbit, 1:1000, 5174 S, Cell Signaling Technology) antibodies were incubated. The membranes were incubated with primary antibody for overnight at 4 °C and further incubated with secondary antibodies (rabbit, 1:10,000, W401B, Promega) for 1 h at room temperature. The blots were imaged using X-ray film or a digital imaging system (Odyssey Fc).

Quantitative RT-PCR analysis

HSCs were directly sorted into 1 ml Trizol reagent (15596026, Thermo Fisher Scientific). Total RNA was extracted by using Trizol reagent followed by quantification with NanoDrop ND-100 spectrophotometer (Thermo Fisher Scientific). cDNA was synthesized by reverse transcription for qPCR with a HiScript II one-step qRT-PCR Kit (Q221-01, Vazyme) on CFX96 realtime PCR system (Bio-Rad). GAPDH was used as the internal control. All primers for qRT-PCR were listed in Table S1, which could see in the Supplementary Material.

ELISA

The femurs and tibias were flushed out with 100 µl PBS. After 10 min of centrifugation at 3000 g, the supernatant was collected and measured using mouse IFN alpha ELISA kit (BMS6027, Thermo Fisher Scientific), IFN gamma ELISA kit (BMS606-2TEN, Thermo Fisher Scientific), IFN beta ELISA kit (424001, Thermo Fisher Scientific), TNF alpha ELISA kit (BMS607-2HS, Thermo Fisher Scientific), IL1alpha ELISA kit (BMS611TEN, Thermo Fisher Scientific), IL1beta ELISA kit (BMS6002-2TEN, Thermo Fisher Scientific), and IL6 ELISA kit (BMS603-2, Thermo Fisher Scientific), according to the manufacturer’s instructions.

RNA-seq data processing

The RNA-seq was processed with six independent replicates per group (Table S2). RNA-seq reads were aligned to mouse genome mm10 and gene annotation from the Ensembl database. Low quality reads were removed and the adaptor sequences were trimmed by Trim Galore (v0.5.0). The remaining pair-end reads were mapped to the reference genome by STAR (v2.6.1) with ENCODE options. The uniquely mapped reads were counted by HTSeq (v0.10.0), and normalized by trimmed mean of M values. Transcripts per million reads (TPM) and read counts were quantified by Salmon (v0.13.1) against the prepared GRCm38 index using salmon quant command. A gene is considered expressed if the TPM > 0.2 in one of six the biological replicates in any of the cell types. Differential expression analysis was assessed using the DESeq2 package in R. Genes were considered differentially expressed when the overall p value < 0.05 and absolute log2 fold change is above 1.5. Gene set enrichment analysis (GSEA) was performed using gsea (v3.12.0, R package) and visualized using enrichplot (v1.6.1, R package). Heatmap was generated using the Pheatmap (v1.0.12, R package).

Statistical analysis

Data are expressed as means ± s.e.m. Differences were considered statistically significant if P < 0.05. Student’s t-test was used for two populations (* P < 0.05, ** P < 0.01, *** P < 0.001), and one-way ANOVA followed by Dunnett’s test was used for multiple comparisons (ǂ P < 0.05, ǂǂ P < 0.01, ǂǂǂ P < 0.001).

Results

HSCs differentiation exhibits myeloid-bias during tumorigenesis

To explore the role of HSCs played in tumorigenesis, we test the function of HSCs in mice suffered from colon cancer or Lewis lung cancer. Initially, we analyzed the bone marrow (BM) from mice that had been challenged with MC38 cells, a murine colon cancer cell line, at 7, 14, and 21 days post-implantation (Fig. 1A). The HSCs number increased on day 7 and returned to baseline level by day 14. The number of MPP2s increased by over 200% on day 14 and returned to baseline levels by day 21, exhibiting a pronounced hysteresis compared to HSCs. Meanwhile, the number of MPP3s remained stable during colon cancer development and the number of MPP4s abruptly increased by almost 2 times on day 21 (Fig. 1B). Upon thorough analysis, the number of myeloid progenitors (common-myeloid progenitors (CMPs), granulocyte-macrophage progenitors (GMPs) and megakaryocyte-erythrocyte progenitors (MEPs)) elevated and the number of lymphoid progenitors CLPs declined during colon cancer tumorigenesis (Fig. 1C). To further investigate the change of HSCs function during tumorigenesis in vivo, we conducted a long-term competitive transplantation experiment by using HSCs from normal and MC38 colon cancer mice (Fig. 1D). The engraftment capability of HSCs derived from both normal mice and MC38 colon cancer mice was found to be similar, as evidenced by the analysis of peripheral blood cells. Upon further detailed analysis, it was observed that HSCs from MC38 colon cancer mice produced a higher number of myeloid cells and a lower number of B cells (Fig. 1E). To confirm the myeloid-bias nature of HSCs in tumorigenesis, we conducted a similar experiment using a Lewis lung cancer mouse model (Fig. 1F). Same as in MC38 colon cancer mice, the HSCs number increased on day 7 and returned to baseline level on day 14 in Lewis lung cancer mice (Fig. 1G). Differently, the MPP3s number experienced a nearly 300% elevation on day 14, while the MPP4s number decreased by half from day 7 to day 14 (Fig. 1G), which also leading to increased number of myeloid progenitors and decreased number of lymphoid progenitors CLPs (Fig. 1H). We also performed a competitive transplantation experiment using the HSCs from Lewis lung cancer mice (Supplementary Fig. S1A) and observed the corresponding myeloid-bias phenomenon (Supplementary Fig. S1B and S1C). Collectively, these in vivo analyses demonstrate a pronounced myeloid-skewed differentiation of HSCs during the process of tumorigenesis.

Fig. 1.

HSCs differentiation exhibits myeloid-bias during tumorigenesis. A Schematic experiment design for cell analysis in MC38 tumor model. Mice were treated with MC38 tumor cells and sacrificed at day 0, 7, 14, and 21. B HSPC populations number in bone marrow after injected with MC38 tumor cells at day 0, 7, 14, and 21. n = 3–4 mice. C Progenitors number in bone marrow after injected MC38 tumor cells at day 0, 7, 14, 21. n = 3–4 mice. D Schematic experiment design for HSCs competitive transplantation from with or without MC38 treatment mice. 500 HSCs from normal or MC38 treated mice, together with 2 × 10⁵ CD45.1 cells were transplanted into lethally irradiated (9 Gy) CD45.1 recipient mice. Peripheral blood was analyzed in recipient mice every 4 weeks from 4 to 16 weeks after transplantation. E Percentage of donor derived the overall, myeloid cell, B cell, and T cell in peripheral blood every 4 weeks from 4 to 16 weeks after transplantation. n = 5 mice. F Schematic experiment design for cell analysis in Lewis lung cancer model. Mice were treated with Lewis lung cancer cells and sacrificed at day 0, 7, 14, and 21. G HSPC populations number in bone marrow after injected Lewis lung cancer cells at day 0, 7, 14, and 21. n = 4 mice. H Progenitor cells number in bone marrow after injected Lewis lung cancer cells at day 0, 7, 14, and 21. n = 4 mice. The data shown represent the mean ± s.e.m. The p values in (B, C, E, G) and (H) were determined using two-tailed Student’s t-test. * p < 0.05, ** p < 0.01, *** p < 0.001, ns, not significant

Tumorigenesis induces myeloid-biased differentiation of HSCs

We further conducted RNA-seq analysis of HSCs from both normal and MC38 colon cancer mice to investigate the mechanism underlying the myeloid-bias of HSCs induced by tumorigenesis. Principal Component Analysis (PCA) revealed differences in the transcriptional profiles of HSCs from WT and MC38 colon cancer mice (Fig. 2A). Compared to those in HSCs from normal mice, 497 genes showed significantly upregulation while 406 genes exhibited downregulation in HSCs from mice subjected to MC38 cells implantation (Fig. 2B). Gene Ontology (GO) analysis of these differentially expressed genes further revealed the obvious upregulation in pathways related to myeloid cell differentiation and inflammation response (Fig. 2C), which suggested a connection between HSCs myeloid-bias and inflammation during tumorigenesis. Accordingly, Gene Set Enrichment Analysis (GSEA) showed significantly enrichment of pathways associated with myeloid cell differentiation and inflammation response (Fig. 2D). Moreover, HSCs from MC38 colon cancer treated Scl-creERT2;tdTomato mice produced a higher number of myeloid cells and a lower number of B cells (Supplementary Fig. S2A and S2B), which was consistent with Fig. 1E and F. Taken together, these results demonstrate that tumorigenesis induces myeloid-biased differentiation of HSCs.

Fig. 2.

Tumorigenesis induces myeloid-biased differentiation of HSCs. A Principal component analysis of bulk RNA-seq of HSCs from WT and MC38 treated mice at day 7. n = 3–4 from pool of 5–6 mice. B Volcano plot of HSCs from mice with or without MC38 treatment. Numbers of differential expression genes (DEGs) (p < 0.05, log2-fold change) were indicated. C The major Gene Ontology (GO) terms of increased (up) and decreased (down) DEGs after MC38 treated. D Gene set enrichment analysis (GSEA) plots of myeloid cell differentiation and inflammation response in HSCs between normal and MC38 treated mice

Tumorigenesis induces myeloid-biased differentiation of HSCs through excess IFN-γ excretion in bone marrow

From the RNAseq analysis, we noticed the inflammatory pathways were notably affected in HSCs from mice with cancer. It is widely known that inflammation can induce myeloid-bias in HSCs. To explore whether tumorigenesis induces myeloid-bias in HSCs through inflammation, we first performed cytokine profiling of the cytokines that have been proven to be involved in HSCs activation in bone marrow and circulation blood from mice with or without MC38 tumors. We only discovered IFN-γ level was significantly elevated in the bone marrow of mice with MC38 tumors compared to that of control mice (Fig. 3A). The cytokine-profiles related to HSCs activation showed no difference in the peripheral blood (PB) of both groups (Fig. 3B). Increased membrane protein expression of IFNGR1 was also detected on HSCs from mice with MC38 tumors (Fig. 3C). Accordingly, we identified the upregulated gene expression of Ifngr1 and Ifngr2, which encode the main receptors for IFN-γ, in HSCs from mice with MC38 tumors (Fig. 3D). GSEA analysis further revealed the increased enrichment of IFN-γ pathway (Fig. 3E) and heatmap analysis demonstrated increased gene expression of common IFN-γ-related genes, such as Ifi213, Stat1 and Ifitm6 (Fig. 3F). The protein level of Stat1 was obviously increased in HSCs of MC38 mice (Fig. 3G). These results indicated the activation of the IFN signaling pathway in the HSCs of mice under the invasion of MC38 tumor.

Fig. 3.

Tumorigenesis induces myeloid-biased differentiation of HSCs through excess IFN-γ excretion in bone marrow. A-B Serum cytokine levels of IFN-γ, IFN-α, IFN-β, TNF-α, IL-1α, IL-1β, IL-6 in the mice bone marrow (A) and serum (B) of MC38 treated or non-treated mice. n = 3–4 mice. C FACS analysis of IFNGR1 expression in HSCs from normal and MC38 treated mice. n = 3 mice. D qPCR analysis of associated cytokine receptors gene expression in HSCs from normal and MC38 treated mice. n = 4 mice. E GSEA plots of interferon gamma response signal pathway in HSCs between normal and MC38 treated mice. F Heatmap plots of genes associated with interferon gamma response signal pathway in HSCs from normal and MC38 treated mice. Ifi213, Stat1, and Ifitm6 were highlighted. n = 3–4 from pool of 5–6 mice. G Western blots of Stat1 expression in HSCs after MC38 treated. 1#, 2#, 3# indicated 3 independent mice. H Schematic experiment design for strategy of Emapalumab treated HSCs from normal and MC38 treated mice. HSCs from normal or MC38-treated mice were sorted for culture up to 7 days, with or without Emapalumab treatment. I HSCs, MPP2s, MPP3s, and MPP4s, each population ratio within HSPCs compartment after Emapalumab ex vivo treatment from normal and MC38 treated mice. n = 4–5 mice. J Frequency of HSCs, MPP2s, MPP3s, and MPP4s after Emapalumab ex vivo treatment from WT and MC38 treated mice. Value relative to mean of untreated control mice and log2-transformed. n = 4–5 mice. K Schematic diagram of the treatment regimen of Emapalumab (100 mg kg^− 1) in mice treated with or without MC38 injection. Normal and MC38 treated mice were injected with Emapalumab at day 5 and analyzed at day 14. L HSPC populations number in bone marrow after Emapalumab treatment in normal and MC38 treated mice at day 14. n = 4 mice. M Progenitors number in bone marrow after Emapalumab treatment in normal and MC38 treated mice at day 14. n = 3–4 mice. The data shown represent the mean ± s.e.m. The p values in (A, B, C) and (D) were determined using two-tailed Student’s t-test. * p < 0.05, ** p < 0.01, *** p < 0.001, ns, not significant. The p values in (J, L) and (M) were determined using one-way ANOVA multiple comparison test, ǂǂ p < 0.01, ǂǂǂ p < 0.001, ns, not significant

Based on the previously proven factor that IFN-γ can promotes myeloid-bias differentiation of HSCs [38], we hypothesize that tumorigenesis-induced myeloid-bias differentiation of HSCs depends on IFN-γ pathway. To verify this hypothesis, we treated HSCs from mice with or without MC38 tumors using Emapalumab, an anti-IFN-γ antibody which has already been utilized clinically for the treatment of primary hemophagocytic lymphohistiocytosis (Fig. 3H). After Emapalumab treatment, the formerly elevated frequency of MMP2 derived from HSCs from mice with MC38 tumors decreased to the same level as that in the control group (Fig. 3I and J). In vivo treatment with Emapalumab further indicated that the interruption of the IFN-γ pathway restored the number of MPP2 and MPP3, which increased by more than 2-fold upon tumorigenesis stimulation, to the normal level in mice with MC38 tumors (Fig. 3K and L). Consequently, the number of CMPs and GMPs significantly decreased following Emapalumab treatment in mice with MC38 tumors (Fig. 3M).

Hence, tumorigenesis causes excessive IFN-γ excretion in the bone marrow, activating the IFN-γ pathway in HSCs and eventually inducing myeloid-biased differentiation.

IFN-γ induce myeloid-biased differentiation of HSCs by upregulating Meis1

Next, we conducted heatmap analysis using the previous RNAseq data to explore the gene profile of myeloid differentiation in HSCs during tumorigenesis. We found a general increase in the expression of genes crucial for myeloid differentiation, such as Meis1, Spi1, Runx1, Gata2, and Cebpα, in HSCs of mice with MC38 tumors (Fig. 4A). Among these, the expression of Meis1 was most significantly increased. We further verified this finding through qPCR experiment, which revealed that the gene expression of Meis1 was elevated by more than two-folds while the expression of Spi1 and Runx1 was slightly increased (Fig. 4B). To clarify the role of Meis1 in tumorigenesis-induced myeloid-bias differentiation of HSCs, we implanted MC38 cells in the Vav-cre; Meis1^flox/flox mice, which were generated by crossing Vav-cre mice and Meis1^flox/flox mice for the deletion of Meis1 in HSCs (Fig. 4C). Meis1 deletion significantly reduced the number and proportion of MPP3 cells in mice with MC38 tumors. However, it exerts no influence on HSCs and MPP2 cells and even tenderly increased the number of MPP4 cells (Fig. 4D and E). The cell cycle showed loss of quiescence in Meis1 deleted HSCs after MC38 treatment, despite of unchanged HSCs number because of the increased differentiation ability (Supplementary Fig. S3A). The effects of Meis1 knockout in HSCs were measured and showed that Meis1 knockout in HSCs caused an increased number of HSC and impaired the quiescence of HSC by cell cycle analysis, which was consistent with previous studies (Supplementary Fig. S3B and S3C) [34, 39]. Although the number of CMP cells was not affected by Meis1 deletion, the number of GMP cells was apparently reduced in mice with Meis1 deletion, which was in line with the decrease in the number of MPP3 cells (Fig. 4F). To confirm that Meis1 played a critical role on tumorigenesis-induced myeloid-bias differentiation of HSCs, we carried out competitive transplantation using the HSCs from the Vav-cre; Meis1^flox/flox mice with MC38 tumors (Fig. 4G). Peripheral blood test showed an obvious decrease of the proportion of myeloid cells derived from HSCs from mice with Meis1 deletion compared to that from control mice under the influence of MC38 tumors (Fig. 4H). Further analysis revealed that the deletion of Meis1 led to a number decrease in CMP and GMP rather than multipotent progenitors (Fig. 4I–K). These results indicated that the deletion of Meis1 in HSCs after MC38 tumor priming didn’t affect the maintenance of the HSCs stemness, but restored the lineage differentiation balance by inhibiting myeloid-biased differentiation of MPPs. We also implanted MC38 cells in recipient mice that were transplanted with HSCs with or without Meis1 (Fig. 4L). We observed a significantly lower level in both the ratio and the number of MPP2 and only a significant decrease in the ratio of MMP3 derived from Meis1 knockout HSCs (Fig. 4M and N). The number of substantial myeloid progenitors, namely CMPs and MEPs, was consistently decreased, which providing compelling evidence to our previous hypothesis that the myeloid-bias of HSCs differentiation induced by MC38 relied on the regulation of Meis1 (Fig. 4O).

Fig. 4.

IFN-γ induce myeloid-biased differentiation of HSCs by upregulating Meis1. A Heatmap plots of myeloid differentiation associated transcription factors, Meis1, Spi1, Runx1, Gata2, Cebpa, and Cebpb in HSCs from normal and MC38 treated mice. n = 3–4 from pool of 5–6 mice. B qPCR analysis of myeloid differentiation associated transcription factors, Meis1, Spi1, Runx1, Gata2, Cebpa, and Cebpb in HSCs from normal and MC38 treated mice. n = 4 mice. C Schematic diagram of the effect of Meis1 gene on HSC function in MC38 tumor-bearing mice. The WT and Meis1 conditional knockout mice were treated with MC38 tumor cells and analyzed at day 14. D HSPC populations number in bone marrow in MC38 treated WT and Meis1 conditional knockout mice at day 14. n = 4 mice. E HSCs, MPP2s, MPP3s, and MPP4s, each population ratio within HSPCs compartment in MC38 treated Meis1 conditional knockout mice at day 14. n = 4 mice. F Progenitors number in bone marrow in MC38 treated Meis1 conditional knockout mice at day 14. n = 4 mice. G Schematic experiment design for HSCs competitive transplantation from MC38 treated Meis1 conditional knockout mice. 500 HSCs from MC38 treated WT and Meis1 conditional knockout mice, together with 2 × 10⁵ CD45.1 cells were transplanted into lethally irradiated (9 Gy) CD45.1 recipient mice. Peripheral blood was analyzed in recipient mice every 4 weeks from 4 to 16 weeks after transplantation. H Percentage of donor derived the overall, myeloid cell, B cell, and T cell in peripheral blood every 4 weeks. n = 5 mice. I-J Donor derived HSPC populations (I) and progenitor cells (J) number in bone marrow in recipient mice transplanted HSCs from MC38 primed Meis1 conditional knockout mice at day 14. n = 4 mice. K Donor derived HSCs, MPP2s, MPP3s, and MPP4s, each population ratio within HSPCs compartment in recipient mice transplanted HSCs from MC38 treated Meis1 conditional knockout mice at day 14. n = 4 mice. L Schematic experiment design for HSCs competitive transplantation from Meis1 conditional knockout mice. 500 HSCs from WT and Meis1 conditional knockout mice, together with 2 × 10⁵ CD45.1 cells were transplanted into lethally irradiated CD45.1 recipient mice. The recipient mice were injected with MC38 tumor cells and analyzed at day 14 after treated MC38 tumor cells. M Donor derived HSCs, MPP2s, MPP3s, and MPP4s, each population ratio within HSPCs compartment in recipient mice after MC38 tumor cells treatment at day 14. n = 4 mice. N and O Donor derived HSPC populations (N) and progenitor cells O number in bone marrow in recipient mice after MC38 tumor cells treatment at day 14. n = 4 mice. P Schematic experiment design for IFN-γ treatment HSCs from Meis1 conditional knockout mice ex vivo culture. HSCs from WT and Meis1 conditional knockout mice were sorted and cultured for 7 days by adding IFN-γ. (Q-R) HSPC populations (Q) and progenitor cells (R) number after ex vivo culture from WT and Meis1 conditional knockout mice at day 7. n = 4 mice. S HSCs, MPP2s, MPP3s, and MPP4s, each population ratio within HSPCs compartment after ex vivo culture from WT and Meis1 conditional knockout mice at day 7. n = 4 mice. The data shown represent the mean ± s.e.m. The p values in (B, D, F, H, I, J, N) and (O) were determined using two-tailed Student’s t-test. * p < 0.05, ** p < 0.01, *** p < 0.001, ns, not significant. The p values in (Q) and (R) were determined using one-way ANOVA multiple comparison test, ǂǂ p < 0.01, ǂǂǂ p < 0.001, ns, not significant

To investigate the relationship between the IFN-γ signaling pathway and Meis1 during the myeloid-bias of HSCs differentiation induced by tumorigenesis, we treated HSCs from the Vav-cre; Meis1^flox/flox mice with or without IFN-γ in vitro. Subsequently, we analyzed the HSPCs compartment and lineage output at day 7 after IFN-γ treatment (Fig. 4P). The Meis1 deficiency exerted no obvious impact on the number of HSPCs during in vitro culture. Consistent with previous studies, IFN-γ treatment led to a significant decrease in both ratio and number of HSCs and MPP4s, the main lymphoid-associated multipotent progenitors (Fig. 4Q and R). Conversely, the ratio and number of the two main myeloid-associated multipotent progenitors, MPP2s and MPP3s, were increased by IFN-γ treatment (Fig. 4Q and R). Accordingly, we found an increase in the number of CMPs and MEPs under the IFN-γ exposure in vitro (Fig. 4S). Meis1 deficiency effectively inhibited the proliferation impairment and myeloid-biased differentiation of HSCs induced by IFN-γ. These findings indicated that transcription factor Meis1 participated in the myeloid-biased output of HSCs induced by IFN-γ.

IFN-γ inhibition enhanced the efficacy of anti-PD-1 treatment in colon cancer

To determine the myeloid compartment derived from tumor primed HSCs and their impact on tumor progression, we quantified mature myeloid cell populations involved in innate immunity in MC38 tumor model. No significant changes in monocytes, macrophages and dendritic cells (DCs), but a reduction in neutrophils in BM were observed in MC38 tumor-bearing mice compared to normal control mice (Fig. 5A). We assessed MDSCs subsets and observed an expansion of monocytic MDSCs (M-MDSCs), but no alterations in polymorphonuclear MDSCs (PMN-MDSCs), in the BM during MC38 tumor progression (Fig. 5B). These data implicated important roles of neutrophils and M-MDSCs in the tumor-associated myeloid-bias of HSCs. To identify the mature myeloid cell types involved, we employed a lineage tracing approach to monitor the HSC proliferation and differentiation during tumor priming. In tamoxifen (TMX) induced Scl-Cre; tdTomato mice, increased tdTomato⁺ labeling was observed in monocytes, macrophages, and M-MDSCs at day 14 post-MC38 tumor challenge (Fig. 5C and D). These results indicated that HSCs-derived M-MDSCs, comprising both granulocytic and monocytic subtypes, may promote tumor progression.

Fig. 5.

IFN-γ inhibition enhanced the efficacy of anti-PD-1 treatment in colon cancer. A Mature innate immune cells number in the bone marrow of mice with or without MC38 tumor cells treatment. n = 3–4 mice. B MDSC subtype cells (PMN-MDSC: CD11b⁺Ly6C^lowLy6G⁺F4/80^–; M-MDSC: CD11b⁺Ly6C^highLy6G^–F4/80⁺) in bone marrow of mice with or without MC38 tumor cells treatment. n = 3–4 mice. C Percentage of tdTomato fluorescence labeled monocytes, macrophages, DCs, and neutrophils in bone barrow of mice with or without MC38 tumor cells treatment. n = 3 mice. D Percentage of tdTomato fluorescence labeled MDSC subtype cells in bone marrow of mice with or without MC38 tumor cells treatment. n = 3–4 mice. E Schematic experiment design for anti-PD-1 (5 mg kg^− 1) and Emapalumab (100 mg kg^− 1) combination therapy treatment of MC38 tumor model. The WT mice were injected with MC38 tumor cells at day 0 and treated with PBS, Emapalumab, anti-PD-1, and anti-PD-1 and Emapalumab at days 7 and 14. The mice were sacrificed at day 21. F Tumor growth after PBS, Emapalumab, anti-PD-1, and anti-PD-1 and Emapalumab treatment. The tumor volume was measured every 2 days between day 7 and day 21 after tumor implantation. n = 5 mice. G MDSC subtype cells number in bone marrow after PBS, Emapalumab, anti-PD-1, and anti-PD-1 and Emapalumab treatment at day 21. n = 3–4 mice. H Percentage of MDSC subtype cells in tumor after PBS, Emapalumab, anti-PD-1, and anti-PD-1 and Emapalumab treatment at day 21. n = 4 mice. I Activated T cells in spleen after PBS, Emapalumab, anti-PD-1, and anti-PD-1 and Emapalumab treatment at day 21. n = 4 mice. J Schematic diagram of the effect of Meis1 gene on MDSC in MC38 tumor mice. The WT and Meis1 conditional knockout mice were treated with MC38 tumor cells. K Tumor growth after ablation of Meis1 in HSCs. The tumor volume was measured every 2 days between day 7 and day 21 after tumor implantation. n = 5 mice. L MDSC subtype cells number in bone marrow after ablation of Meis1 in HSCs at day 21. n = 5 mice. M Percentage of MDSC subtype cells in tumor after ablation of Meis1 in HSCs at day 21. n = 4 mice. N Activated T cells in spleen after ablation of Meis1 in HSCs at day 21. n = 4–5 mice. The data shown represent the mean ± s.e.m. The p values in (A, B, C, D, K, L, M) and (N) were determined using two-tailed Student’s t-test. * p < 0.05, ** p < 0.01, *** p < 0.001, ns, not significant. The p values in (F, G, H) and (I) were determined using one-way ANOVA multiple comparison test, ǂǂ p < 0.01, ǂǂǂ p < 0.001, ns, not significant

The selective targeting of tumor-infiltrating MDSCs enhanced the antitumor efficacy of anti-PD-1 treatment by promoting T cell activation in our mouse model. To evaluate this strategy, we treated MC38 tumor-bearing mice with PBS, Emapalumab alone, anti-PD-1 alone and anti-PD-1 combined with Emapalumab (Fig. 5E). This combination therapy specifically targeted HSC-derived MDSCs to assess its anti-tumor potential.This combination therapy significantly suppressed tumor growth compared to all other treatments(Fig. 5F; Tabl S3). While PMN-MDSCs number in BM and ratios in tumor tissues remained unchanged, the treatment specifically reduced M-MDSCs level in these compartments (Fig. 5G and H). Furthermore, this combination regimen enhanced cytokine production by T cells, with increased IFN-γ, IL-2, and IL-4 in CD4 + T cells and elevated IFN-γ in CD8⁺ T cells. (Fig. 5I and Supplementary Figure. S4).

Based on the established role of Meis1 in driving myeloid-biased HSC differentiation in tumor-bearing mice, we investigated whether it also orchestrates the expansion of MDSCs during tumor progression. Following Meis1 ablation in the MC38 tumor model, we assessed tumor growth, MDSCs accumulation, and T cell responses (Fig. 5J). Meis1 deficiency significantly decelerated tumor progression (Fig. 5K). PMN-MDSC number in BM and ratio in tumor tissues were not affected by Meis1 ablation, but M-MDSCs accumulation were significantly reduced (Fig. 5L and M). Consistently, Meis1-deficient mice showed enhanced cytokine production in T cells, with CD4 + T cells produceing more IFN-γ, IL-2, and IL-4, and CD8 + T cells elevating IFN-γ secretion(Fig. 5N). These results demonstrate that Meis1 is a critical regulator that governs the expansion of M-MDSCs originating from tumor-induced, myeloid-biased HSCs, positioning its inhibition as a promising strategy to enhance response to anti-PD-1 treatment.

IFN-γ inhibition enhanced the efficacy of anti-PD-1 treatment in lung cancer

To evaluate the broader implications of targeting HSCs-dependent myeloid-biased differentiation, we examined the effect of anti-PD-1 and Emapalumab combination therapy in a Lewis lung cancer model. We first assessed whether Emapalumab modulates HSC differentiation in this model (Fig. 6A). Emapalumab significantly inhibited the differentiation of HSCs, MPP2s, and MPP3s compared to untreated group, a trend consistent with changes in the overall HSPC compartment (Fig. 6B and C). Furthermore, Emapalumab treatment led to a reduction in CMPs, GMPs, and MEPs, accompanied by an increase in CLPs(Fig. 6D). Based on the findings that Emapalumab corrected the myeloid-biased HSC differentiation in MC38 tumor model, we evaluated its therapeutic effect on Lewis lung cancer. Mice received biweekly administration of anti-PD-1 alone or in combination with Emapalumab (Fig. 6E). The combination therapy significantly suppressed tumor growth compared to all other treatments (Fig. 6F; Table S3). This enhanced efficacy was associated with a specific reduction in M-MDSCs, but not PMN-MDSCs in both BM and tumor tissues(Fig. 6G and H), accompanied by increased cytokine production from CD4⁺ (IFN-γ, IL-2, IL-4) and CD8⁺ T cells (IFN-γ) (Fig. 6I and Supplementary Figure. S5). These results confirm that anti-PD-1/Emapalumab combination therapy is effective in Lewis lung cancer, operating through suppression of HSC-derived MDSCs and potentiation of T cell responses. To further verify Meis1’s role in regulating MDSCs output derived from tumor-induced myeloid-biased HSC differentiation in Lewis lung cancer model, we assessed tumor growth MDSCs accumulation, and T cell activity aollowing Meis1 ablation in Lewis lung cancer model (Fig. 6J). Meis1 ablation significantly decelerated the tumor progression (Fig. 6K). While PMN-MDSCs number in BM and ratio in tumor were not obviously affected, the Meis1 defficiency specifically reduced M-MDSCs accumulation in these compartment (Fig. 6L and M). Consistently, CD4⁺ T cells exhibited enhanced production of IFN-γ, IL-2, and IL-4, and CD8⁺ T cells showed increased IFN-γ secretion in Meis1-ablated mice.(Fig. 6N). These data demonstrate the therapeutic potential of targeting IFN-γ or Meis1 to enhance the antitumor efficacy of anti-PD-1 therapy across multiple cancer types.

Fig. 6.

IFN-γ inhibition enhanced the efficacy of anti-PD-1 treatment in lung cancer. A Schematic experiment design for Emapalumab treatment in normal and Lewis lung cancer cells treated mice. Normal and Lewis treated mice were injected with Emapalumab at day 5 and analyzed at day 14. B HSPC populations number in bone marrow after Emapalumab treatment in normal and Lewis treated mice at day 14. n = 4–5 mice. C HSCs, MPP2s, MPP3s, and MPP4s, each population ratio within HSPCs compartment after Emapalumab treatment in normal and Lewis treated mice at day 14. n = 4–5 mice. D Progenitors number in bone marrow after Emapalumab treatment in normal and Lewis treated mice at day 14. n = 3–4 mice. E Schematic experiment design for anti-PD-1 and Emapalumab combination therapy treatment of Lewis lung cancer model. The WT mice were injected with Lewis lung cancer cells at day 0 and treated with PBS, Emapalumab, anti-PD-1, and anti-PD-1 and Emapalumab at days 7 and 14. The mice were sacrificed at day 21. F Tumor growth after PBS, Emapalumab, anti-PD-1, and anti-PD-1 and Emapalumab treatment. The tumor volume was measured every 2 days between day 7 and day 21 after tumor implantation. n = 5 mice. G MDSC subtype cells number in bone marrow of of Lewis treated mice after PBS, Emapalumab, anti-PD-1, and anti-PD-1 and Emapalumab treatment at day 21. n = 3–4 mice. H Percentage of MDSC subtype cells in tumor after PBS, Emapalumab, anti-PD-1, and anti-PD-1 and Emapalumab treatment at day 21. n = 4 mice. I Activated T cells in spleen of Lewis treated mice after PBS, Emapalumab, anti-PD-1, and anti-PD-1 and Emapalumab treatment at day 21. n = 4 mice. J Schematic experiment design for Vav-cre; Meis1^flox/flox mice treated with Lewis lung cancer cells. The WT and Meis1 conditional knockout mice were treated with Lewis lung cancer cells. K Tumor growth after ablation of Meis1 in HSCs. The tumor volume was measured every 2 days between day 7 and day 21 after tumor implantation. n = 5 mice. L MDSC subtype cells number in bone marrow after ablation of Meis1 in HSCs at day 21. n = 4 mice. M Percentage of MDSC subtype cells in tumor after ablation of Meis1 in HSCs at day 21. n = 4 mice. N Activated T cells in spleen after ablation of Meis1 in HSCs at day 21. n = 5 mice. The data shown represent the mean ± s.e.m. The p values in (K, L, M) and (N) were determined using two-tailed Student’s t-test. Data represent mean ± s.e.m. * p < 0.05, ** p < 0.01, *** p < 0.001, ns, not significant. The p values in (B, D, F, G, H) and (I) were determined using one-way ANOVA multiple comparison test, ǂǂ p < 0.01, ǂǂǂ p < 0.001, ns, not significant

Discussion

The imbalance of immunity is one of the most important mechanisms of tumor development and progression [40, 41]. Myeloid cells are the most abundant element of the tumor immune microenvironment [42, 43] and play a complex and crucial role in tumorigenesis. Under normal physiology, myeloid cells act as immune defenders and immune surveillants. Meanwhile, tumor recruits and educates them to facilitate immune escape [44]. Specifically, tumor-associated macrophages (TAMs) and MDSCs are the two primary myeloid cell populations induced by tumorigenesis. These cells play a critical role in suppressing the function of T cells and natural killer (NK) cells, thereby supporting tumor growth and metastasis [32, 33]. MDSCs are also a key contributor to resistance against ICB, a primary modality in modern antitumor treatment [45, 46]. It is widely acknowledged that TAMs originate from mature circulating macrophages [47] and MDSCs are differentiated from granulocyte-macrophage progenitors (GMPs) [48]. Moreover, clinical observation revealed elevated neutrophil-to-lymphocyte ratio and circulating GMPs in the peripheral blood of cancer patients [49–52]. Myeloid-Biased extramedullary hematopoiesis was further discovered in cancer patients to account for the source of increasing myeloid cells both in circulation and tumor tissues [52, 53]. Consistent with these clinical observations, our current study demonstrated a significant increase in myeloid cells, particularly MDSCs, in the peripheral blood of mice that were challenged with tumors. Further, we observed elevated myeloid HSPCs in these mice. Through HSCs transplantation experiments with HSCs-Scl-creERT2;tdTomato mice as donors to trace HSCs, we subsequently demonstrated that after tumor stimulation, bone marrow HSCs displayed a myeloid-biased differentiation tendency even when they were transplanted into normal mice. We also found Meis1 was key regulator of myeloid-biased differentiaton in tumor primed HSCs, which was paly an vital role in manipulating HSCs quiescence in previous study [34, 39]. Moreover, tumor-educated HSCs are preferred to differentiate into GMP rather than MEP, which explains the increase in circulating MDSCs. These findings provide evidences that solid tumors also educate bone marrow HSCs to supply circulating myeloid cells to facilitate tumor development.

Tumorigenesis is one of the most common pathogeneses of inflammation [54]. Under tumorigenesis, pro-inflammation cytokines, such as IL-6, IL-1, TNF-α and IFNs are secreted to create an inflammatory tumor microenvironment and promote tumor progression [55]. Also, it has been proved that these pro-inflammation cytokines assist tumor to educate myeloid cells to transform into TAMs and MDSCs, thereby facilitating tumorigenesis [32, 33]. HSCs are proved to exhibit myeloid-bias under inflammation environment induced by infection and aging [56, 57]. We observed a normal profile of circulating pro-inflammation cytokines, including IL-6, IL-1, TNF-α and IFNs, in mice with tumors. Meanwhile, the level of IFN-γ was obviously elevated in the bone marrow of mice with tumors, and the expression of IFN-γ receptor was enhanced in tumor-educated HSCs. This finding aligns with prior reports of interferon-stimulated gene upregulation in HSPCs from breast cancer models [31]. RNAseq analysis of tumor-educated HSCs also revealed enrichment of pathways related to inflammation response. IFN-γ was fully studied as a critical cytokine affecting the activity and function of HSCs. IFN-γ was shown to inhibit hematopoiesis as early as 1980 s [58, 59]. Further investigations revealed that IFN-γ hinders the quiescence and impairs the self-renewal of HSCs, ultimately resulting in the malfunction and exhaustion of HSCs [60–62]. As well, IFN-γ was found to induce myelopoiesis and skew multipotent progenitors toward a monocytic lineage via multiple mechanisms, such as inhibition of G-CSF-induced activation of STAT3, induction of expression of SOCS3 in GMPs, and decreasing the expression of RUNX1 and CEBPα in HSPCs [60, 63, 64]. Consistent with these previous studies, our findings indicated tumors induced myelopoiesis of HSCs by the activation of IFNγ pathway, leading to an increase in M-MDSC rather than PMN-MDSC in tumor-bearing mice. Furthermore, we demonstrated Meis1, a critical differentiation regulator of HSCs, was the key regulator of IFN-γ-induced myeloid bias of HSCs. Consequently, it remains to be explored whether chronic IFN-γ exposure in the bone marrow environment will eventually lead to HSCs malfunction and exhaustion, thereby increasing the morbidity of hypocytosis and secondary MDS, two hematological complications of solid tumors [65].

Emapalumab, an IFN-γ antibody, has been successfully utilized in the clinical treatment of hemophagocytic syndrome (HLH) due to its ability to inhibit the unbridled activation of T cells and NK cells induced by the high level of IFN-γ [66]. Our results indicated Emapalumab effectively rectified the tumor-induced myeloid bias of bone marrow HSCs and partially reduced the production of M-MDSC. However, it failed to retard tumor development. It is probably because IFN-γ inhibition also prevents the tumor-killing activities that relied on the activation of T cells and NK cells. Recent years, ICB therapy has revolutionized cancer therapy. Notably, PD-1 and PD-L1 antibodies significantly improved the survival of patients subject to different solid tumors [67]. Targeting MDSCs can boost the efficacy of ICB in cancer treatment [68–70]. So, we combined PD-1 antibody with Emapalumab to treat mice suffering from colon or lung cancer and discovered an enhanced response of tumors to PD-1 antibody. These findings further substantiate that IFN-γ is one of the main cytokines that tumors employ to induce the production of immune suppressive myeloid cells.

Conclusions

In conclusion, we demonstrate that tumorigenesis creates an IFN-γ-rich inflammatory bone marrow environment that drives Meis1-dependent myeloid bias in HSCs, which in turn translates into immunosuppression and tumor progression. This insight provides a fundamental rationale for a novel combinatorial approach to potentiate ICB in solid tumors.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

**Xue Han** : Conceptualization, Writing - original draft, Writing - review & editing, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation. **Minyi Zhao** : Conceptualization, Writing - review & editing, Investigation, Formal analysis. **Kexin Wang** : Investigation, Visualization, Validation, Formal analysis. **Weiwei Ma** : Investigation, Methodology. **Songqi Zhu** : Investigation, Methodology. **Ruiqing Zhou** : Investigation, Methodology. **Uet Yu** : Investigation, Formal analysis. **Bangxue Jiang** : Investigation. **Xiaoqing Bai** : Investigation. **Peng Lei** : Conceptualization, Project administration, Data curation, Writing - review & editing, Supervision, Funding acquisition. **Shunqing Wang** : Conceptualization, Data curation, Investigation, Writing - review & editing, Project administration, Supervision, Formal analysis, Methodology, Funding acquisition.

Funding

This work was supported by the National Natural Science Foundation of China [grant numbers 82100166]; the National Key R&D Program of China [grant number 2024YFC2510500].

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This work, titled “IFN-γ induces hematopoietic stem cell myelopoiesis through Meis1 in tumor”, has been approved by Ethnics Committee of Guangzhou First People’s Hospital. Thr approval number was N2025-36004. Date of approval was February 28th, 2025.

Competing interests

The authors declare no competing interests.