Transcriptome analysis of myeloid cells exposed to leukemic cells in a CLL-xenograft system

Immune cells of the myeloid lineage and CLL cells support each other during leukemia progression and dissemination [19]. We have investigated the molecular interactions supporting this cell-cell interdependence with a special focus on ncRNAs, whose role in non-malignant immune cells is largely unknown.

We used a mouse Affymetrix Clariom D assay and Transcriptome Analysis Console (TAC) 3.0 software to perform a broad transcriptome-gene- and exon-level analysis of coding and ncRNA isoforms in different stages of leukemia (day 21, early stage; day 31, late stage leukemia) in Rag2−/−γc−/− mice xeno-transplanted with MEC1 cells (Fig. S1A). As shown in Fig. S1B, we found a significant enrichment of both upregulated and downregulated ncRNAs. MiRNAs miR-511-3p, miR-181c, miR-9-2 and the long ncRNAs (lncRNAs) HOX transcript antisense RNA 4 (HOTAIR 4) and HOX antisense intergenic RNA myeloid 1 (HOTAIRM1) were found upregulated in murine myeloid cells during leukemia progression (Table S8). ROCK2, a direct target of miR-511-3p [20], and DLEU2/miR-16-1, which are on the same gene cluster that maps the human chromosome 13q14 region, were found downmodulated in murine myeloid cells during leukemia progression (Table S9). Although the deletion of 13q14 is the most frequent genetic lesion in CLL cells, miR-15a/miR-16-1-mediated regulation of non-malignant myeloid cells has not been characterized. The MDR includes the first exon of the DLEU1 ncRNA, the deleted in leukemia (DLEU) 2 gene, encoding for a primary transcript, and the miR-15a/miR-16-1 cluster, which is located in an intron of DLEU2. We previously reported that DLEU2 transcript is downregulated in myeloid cells isolated from the BM of MEC1-xenotransplanted mice at early stage of leukemia development compared with myeloid cells from age-matched, wild-type (WT), untransplanted mice [19]. Other groups have found that miR-16-1 regulates macrophage protumor activation and polarization through the critical immune suppressor molecule PD-L1 [18, 21]. Recent findings have also demonstrated that myeloid cells are required for PD1/PD-L1 checkpoint activation [22] and that PD-L1 expression can be induced in protumor and immunosuppressive cell types of the myeloid lineage [23].

Our findings corroborate the hypothesis that DLEU2/miR-15a/miR-16-1 has a role in the protumor function of myeloid cells during leukemia progression.

The TCL1+/−MDR−/− mouse model of CLL

We used the mice with the germline deletion of the MDR (MDR−/−) [6] to further explore the hypothesis that DLEU2/miR-15a/miR-16-1 affects the protumor function of non-malignant myeloid and lymphoid cells during leukemia progression. MDR−/− mice were crossed with Eμ-TCL1 transgenic (tg) mice [24]. TCL1+/−, TCL1+/−MDR−/− (TM), MDR−/− and WT mice were developed, and we first analyzed their survival. The MDR deletion significantly shortened the life spans of the mice (Fig. 1A), thus offering a more rapid and reliable model of CLL development. We then characterized 4- and 9-month-old TCL1+/−, TCL1+/−MDR−/−, MDR−/− and WT mice (n = 6–10 mice/group). By using multi-color flow cytometry, we have analyzed the accumulation of CD19+CD5+ leukemic cells and key molecules (e.g., BCL2) potentially involved in the miR-15a/miR-16-1-mediated control of cell proliferation and differentiation. An increase of CD19+CD5+ leukemic cells in the spleen (SP) of 4-month-old TCL1+/−MDR−/− mice correlated with the upregulation of BCL2 (Fig. 1B, C). The splenic CD19+CD5+ leukemic expansion and BCL2 upregulation were exacerbated in 9-month-old TCL1+/−MDR−/− mice (Fig. 1D, E). The same trend was observed in the peripheral blood (PB) and bone marrow (BM), but it did not reach statistical significance due to the heterogeneity of the mice (Fig. S2A–H).

Fig. 1: Premature CLL-like expansion and shorter life span in TCL1 transgenic mice with MDR deletion.

A Kaplan–Meier survival curves for TCL1+/− (n = 51), TCL1+/−MDR−/− (TM; n = 20), MDR−/−(n = 18) and wild-type (WT; n = 17) mice. A statistical analysis of the groups was performed using the log-rank test (median survival: 10 months for TM, 13 months for TCL1+/−, 17 months for MDR−/−, 24 months for WT mice). Mice were included in the analysis after spontaneous death or after they had been killed because of symptoms of illness. B, C Using flow cytometry, 4-month-old TM mice (n = 7), age-matched TCL1+/− (n = 6), MDR−/− (n = 6) and WT control (n = 6) mice were analyzed for the accumulation of CD19+ CD5+ B cells in the spleen. B The mean values ± the standard deviations (SDs) of the absolute numbers of CD19+ CD5+ cells gated on CD19+ cells and (C) the mean values ± SDs of the absolute numbers of CD19+ BCL2+ cells gated on CD19+ CD5+ cells are shown in the graphs. A statistical analysis was performed using the Student t test *P < 0.05, **P < 0.01. D, E Using flow cytometry, 9-month-old TM (n = 9), age-matched TCL1+/− (n = 10), MDR−/− (n = 10), WT control (n = 7) mice and, 27-month-old MDR−/− (n = 3) were analyzed for the accumulation of CD19+ CD5+ B cells in the spleen. D The mean values ± SDs of the absolute numbers of CD19+ CD5+ cells gated on CD19+ cells and (E) the mean values ± SDs of the absolute numbers of CD19+ BCL2+ cells gated on CD19+ CD5+ cells are shown in the graphs. A statistical analysis was performed using the Student’s t test *P < 0.05, **P < 0.01. F–H B cells were purified from the spleen of 9-month-old TM, TCL1+/− and MDR−/− mice using magnetic negative selection. Lysates were processed for the reverse phase protein Array (RPPA). The heatmaps of unsupervised hierarchical clustering display differentially expressed proteins between (F) leukemic cells from TCL1+/− (n = 3) and MDR−/− (n = 6) mice, (G) leukemic cells from TM (n = 3) and MDR−/− (n = 6) mice, and (H) leukemic cells from TM (n = 3) and TCL1+/− (n = 3) mice.

To examine the proteomic profile of the leukemic cells purified from the spleen of leukemic TCL1+/−MDR−/− mice compared to TCL1+/− and to MDR−/− mice, we used the RPPA high-throughput technology (Fig. 1F–H). We found that several components of the BCR and PI3K/AKT signaling pathways (e.g., PKCa, Lyn, ERK1/2, JNK2) [25] were upregulated in the TCL1+/− compared with the MDR−/− mice (Fig. 1F). The same signaling pathways were found activated in the TCL1+/−MDR−/− mice compared to MDR−/− mice including the downstream upregulation of the anti-apoptotic protein MCL1, a well-known target of miR-15a/miR-16-1 [9] (Fig. 1G). Of note, serum- and glucocorticoid-inducible protein kinase 3 (SGK3) was upregulated in the splenic CLL cells of TCL1+/−MDR−/− mice compared with TCL1+/− and MDR−/− mice (Fig. 1G, H). SGK3, also known as cytokine-independent survival kinase, is a well-known downstream mediator of PI3K oncogenic signaling in various cancers, including breast cancer and ovarian cancer [26, 27]. SGK3 is involved in cell proliferation, growth, survival, and migration and is considered an intriguing target for anticancer drug development [28, 29].

Overall, we have generated a faster CLL mouse model with two known pro-tumorigenic genes, the hTCL1 oncogene and the MDR tumor-suppressor, in the context of a murine immunocompetent microenvironment.

Characterization of the immune cells in the TCL1+/−MDR−/− mouse model

To investigate the involvement of miR-15a/miR-16-1 in the protumor function of myeloid and lymphoid non-malignant cells in the leukemic microenvironment during leukemia progression, we examined the lymphoid and myeloid cell populations including lymphocytes, monocytes, macrophages, and monocytic-myeloid derived suppressor cells (M-MDSCs) in the circulation and lymphoid tissues of TCL1+/−MDR−/− and age-matched control mice. We analyzed key molecules potentially involved in the miR-16-1-mediated control of myeloid cell proliferation and differentiation (e.g., BCL2 and checkpoint molecules PD-1 and PD-L1) in 4- and 9-month-old mice. In the spleen of 4-month-old mice, we observed an altered composition of the T lymphocyte population and of the CD4:CD8 T cell ratio (Fig. 2A) and found a significant increase of the whole pool of CD8+ T cells and of CD8+ and CD4+ memory T cells expressing BCL2 in TCL1+/−MDR−/− and MDR−/− mice compared with TCL1+/− and WT mice (Fig. 2B–E). Furthermore, we found high numbers of PD1+ CD4+ effector memory (TEM) T cells (Fig. 2F) and an increase of potentially regulatory CD4+ CD25+ T cells expressing BCL2 in TCL1+/−MDR−/− compared with TCL1+/− mice (Fig. 2G, H). This aberrant T cell composition was associated with the significant increase of Ly6Clow monocytes in the circulation, BM, and SP of the 4-month old TCL1+/−MDR−/− mice (Fig. 3A–G). Ly6Clow monocytes differentiate into protumor-associated macrophages (TAMs) [30]. Of note, in the SP, we found high frequency of protumor TAMs expressing MRC1, BCL2 and PD-L1 (Fig. 3H–J). These findings demonstrate that the deletion of the DLEU2/miR-15a/miR-16-1 cluster impacts the myeloid and lymphoid cell compartment at early stage of leukemia development.

Fig. 2: Characterization of the T-cell compartment of young TCL1 transgenic mice with MDR deletion.

A–H The splenic T-cell compartment of 4-month-old TCL1+/−MDR−/− (TM; n = 7), age-matched TCL1+/−(n = 6), MDR−/− (n = 6), and wild-type (WT) control (n = 6) mice was analyzed using flow cytometry. A The mean values ± the standard deviations (SD) of the absolute numbers of CD4+ and CD8+ T cells, (B) the mean values ± SD of the absolute numbers of CD44−CD62L+ naïve, CD44+CD62Llow/neg effector and CD44+CD62L+ central memory CD8+ T cells and, (C) the mean value ± SDs of the absolute numbers of CD44-CD62L+ naïve, CD44+CD62Llow/neg effector and CD44+CD62L+ central memory CD8+ T cells expressing BCL2 are shown in graphs. D The mean value ± SDs of the absolute numbers of CD44−CD62L+ naïve, CD44+CD62Llow/neg effector and CD44+CD62L+ central memory CD4+ T cells, (E) the mean value ± SD of the absolute number of CD44−CD62L+ naïve, CD44+CD62Llow/neg effector and CD44+CD62L+ central memory CD4+ T cells expressing BCL2 or (F) PD1 are shown in graphs. G The mean value ± SD of the absolute number of CD4+CD25+ T cells and (H) CD4+CD25+BCL2+ T cells are shown in graphs. A statistical analysis was performed using the Student’s t test *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 3: Characterization of the monocytes and macrophages of young TCL1 transgenic mice with MDR deletion.

A–H Monocytes and macrophages of 4-month-old TCL1+/−MDR−/− (TM; n = 7), age-matched TCL1+/−(n = 6), MDR−/− (n = 6) and WT control (n = 6) mice were analyzed using flow cytometry. A The mean value of the relative contributions of CD11b+ CSF1R+ cells gated on CD45 in peripheral blood (PB) is shown in the graph. B The mean values of the absolute numbers of CD11b+ CSF1R+ cells gated on CD45 in BM and (C) SP are shown in the graphs. D The mean values of the absolute numbers of CD11b+ CSF1R+ BCL2+ monocytes is shown in graph. E The mean value of the relative contributions of CD11b+ Ly6Clow and CD11b+ Ly6Chigh cells to the whole monocyte subset (CD11b+ CSF1R+) gated on CD45+ in the PB is shown in the graph. F The mean values of the absolute numbers of CD11b+ Ly6Clow and CD11b+ Ly6Chigh cells to the whole monocyte subset (CD11b+ CSF1R+) gated on CD45+ in the BM and (G) SP are shown in the graphs. H The mean values of the absolute numbers of CD11b+ F4/80+ cells gated on CD45+ in the SP is shown in the graph. I The mean values of the absolute number of CD11b+ F4/80+ MRC1+ cells, (J) CD11b+ F4/80+ BCL2+ and CD11b+ F4/80+ PD-L1+ gated on CD45+ in the SP are shown in the graphs. A statistical analysis was performed using the Student’s t test *P < 0.05.

In 9-month-old mice, we confirmed an altered CD4:CD8 T-cell ratio in the circulation of TCL1+/−MDR−/− and MDR−/− mice with a significant increase of CD8+ effector memory TEM cells and of CD8+ central memory TCM expressing PD-1 in the SP (Fig. 4A–D). A small cohort of 27-month-old MDR−/− was included in the study to rule out the possibility of age-related effects in the mice with the germline deletion.

Fig. 4: Characterization of the T cell compartment of leukemic TCL1 transgenic mice with MDR deletion.

A–D The splenic T-cell compartment of 9-month-old TCL1+/−MDR−/− (TM; n = 9), age-matched TCL1+/− (n = 10), MDR−/− (n = 10), WT control (n = 7) mice and 27-month-old MDR−/− (n = 3) was analyzed using flow cytometry. A The relative contributions ± SD of the of the CD4+ and CD8+ T cells in the PB, (B) the mean values ± SDs of the absolute numbers of CD4+ and CD8+ T cells in the SP, (C) the mean values ± SDs of the absolute numbers of CD44-CD62L+ naïve, CD44+CD62Llow/neg effector and CD44+CD62L+ central memory CD8+ T cells and, (D) the mean values ± SDs of the absolute numbers of CD44−CD62L+ naïve, CD44+CD62Llow/neg effector and CD44+CD62L+ central memory CD8+ T cells expressing PD1 are shown in the graphs. A statistical analysis was performed using the Student’s t test *P < 0.05, **P < 0.01, ***P < 0.001.

In addition, an increase in CD19+CD5+ leukemic cells was associated with an increased number of BCL2-expressing monocytes in the (PB), BM and SP (Fig. 5A–F); increased numbers of PD-L1 expressing monocytes in the PB and BM (Fig. 5G, H); and an increased percentage of protumor F4/80+ MRC1+ macrophages in the PB (Fig. 5I, J) and SP (Fig. S3).

Fig. 5: Characterization of monocytes and macrophages of leukemic TCL1 transgenic mice with MDR deletion.

A–H The monocytes and macrophages of 9-month-old TCL1+/−MDR−/− (TM; n = 9), age-matched TCL1+/− (n = 10), MDR−/− (n = 10), WT control (n = 7) mice and 27-month-old MDR−/− (n = 3) were analyzed using flow cytometry. A The mean values ± SDs of the relative contributions of CD11b+ CSF1R+ cells gated on CD45 in PB is shown in the graph. B The mean values ± SDs of the absolute numbers of CD11b+ CSF1R+ cells gated on CD45 in BM and (C) SP are shown in the graphs. D The mean values ± SDs of relative contributions of CD11b+ CSF1R+ BCL2+ monocytes in PB is shown in the graph. E The mean values ± SDs of the absolute numbers of CD11b+ CSF1R+ BCL2+ monocytes in BM and (F) SP are shown in the graphs. G The mean values ± SDs of relative contributions of CD11b+ CSF1R+ PDL1+ monocytes in PB is shown in the graph. H The mean values ± SDs of the absolute numbers of CD11b+ CSF1R+ PDL1+ monocytes in BM is shown in the graph. I The mean values ± SDs of the relative contributions of CD11b+ F4/80+ cells gated on CD45+ in the PB is shown in the graph. J The mean values ± SDs of the relative contributions of CD11b+ F4/80+ MRC1+ cells gated on CD45+ in the PB is shown in the graph. A statistical analysis was performed using the Student’s t test *P < 0.05, **P < 0.01, ***P < 0.001.

To finally demonstrate the cell-autonomous, pro-leukemic activity of monocytes from 13q14 MDR−/− mice, we used a TCL1 tg transplantation system where we transplanted leukemic cells obtained from the SP of an Eμ-TCL1 tg mouse into syngeneic, immunocompetent recipients. At day 84 post-transplantation, mice were injected intravenously (i.v.) with monocytes purified from the BM of MDR−/− mice. Seven days later, they were killed. As shown in Fig. 6A–F, a significantly higher frequency of CD19+ CD5+ leukemic cells was observed in the spleen of mice adoptively transferred with MDR−/−-derived monocytes that was accompanied by an increased spleen weight.

Fig. 6: Pro-tumor function of monocytes and macrophages from MDR−/− mice in the TCL1 transgenic transplantation system.

A C57BL/6 mice intraperitoneally (i.p.) transplanted with leukemic B cells from a Eµ-TCL1 transgenic mouse donor, were left untreated (n = 3, red squares) or adoptively transferred (AT, day +84) with monocytes/macrophages purified from the BM of MDR−/− mice (n = 3, empty red squares). Mice were killed on day 91 and were analyzed using flow cytometry. B The mean values ± SD of the relative contributions of CD19+ CD5+ cells to the whole B cell pool at day 84 and (C) day 91 in the PB of mice are shown in the graphs. D The mean values ± SDs of the absolute number of CD19+ CD5+ cells gated on CD19+ in BM and (E) SP are shown in the graphs. F The spleen weights of untreated and adoptively transferred mice are shown in the graph. A statistical analysis was performed using the Student’s t test *P < 0.05.

miR-15a/ miR-16-1 in CLL patient-derived immune cells

To validate the molecular and functional information gathered from mouse models in human samples, we evaluated miR-15a/miR-16-1 expression in human immune cells, including CD19+ B cells, CD8+ effector and central memory T cells (TEM and TCM), CD4+ TREG cells, CD14+CD16++ non-classical (NC), CD14++CD16+ intermediate (I) and CD14++CD16- classical (C) monocytes, and CD14+ HLA-DRlow/neg monocytic-myeloid derived suppressor cells (M-MDSCs). Figs. S4A and 7A show the expression level of miR-15a and miR-16-1 relative to U6 control on CD19+ B cells, T cells and monocyte subsets separated from fresh PBMCs of a cohort of 7 patient samples from MDACC (cohort 1, patients 1-7; Table S1).

Fig. 7: Characterization of miR-16-1 and related target proteins in human immune cells.

A, B The relative quantification values were calculated as miR-16-1 expression relative to U6 control. (A) miR-16-1 expression in human CD19+ B cells, effector memory hCD8+ CD45RA−CD45RO+CD62L− TEM, central memory hCD8+ CD45RA-CD45RO+CD62L+ TCM, CD4+CD25+CD127low/− Tregs, CD14+ CD16++ non-classical, CD14++ CD16+ intermediate and CD14++ CD16− classical monocytes and CD14+HLADRlow/− M-MDSCs, separated from fresh PBMCs of CLL patients (n = 7, patients 1–7, Table S1) using fluorescence-activated cell sorting is shown in the graph. B Frozen PBMCs obtained from healthy donors (n = 6) and patients with CLL (n = 6, patients 1, 2, 3, 51, 52, 68) were plated for 2 h in a 6-well plate at 6 × 106/1.8 mL after thawing and were then stained for fluorescence activated cell sorting (FACS). Patients 51, 52, 68 were enrolled into trial with ibrutinib (IBT) and venetoclax (Ven) after this study, and are indicated as pre-treatment. miR-16-1 expression in human CD19+ B cells, effector memory hCD8+ CD45RA−CD45RO+CD62L− TEM, central memory hCD8+ CD45RA-CD45RO+CD62L+ TCM, CD4+CD25+CD127low/− TREG, CD14+ CD16++ non-classical (NC), CD14++ CD16+ intermediate (I) and CD14++ CD16− classical (C) monocytes and CD14+HLADRlow/− M-MDSCs, separated from PBMCs of healthy donors and patients with CLL is shown in graph. C–F Frozen PBMCs obtained from healthy donors (n = 6) and patients with CLL (n = 6, patients 1, 2, 3, 51, 52, 68) were plated for 2 h in a 6-well plate at 6 × 106/1.8 mL after thawing and were then stained and analyzed for the expression of BCL2, PD1 and PD-L1 on CD19+ B cells, effector memory hCD8+ CD45RA-CD45RO+CD62L− TEM, central memory hCD8+ CD45RA−CD45RO+CD62L+ TCM, CD4+CD25+CD127low/− Tregs, CD14+ CD16++ non-classical, CD14++ CD16+ intermediate and CD14++ CD16− classical monocytes and CD14+HLADRlow/− M-MDSCs using flow cytometry. C The mean values ± SDs of the relative contributions of BCL2 expressing immune cells from patients with CLL and healthy donor controls and are shown in the graph. D The mean values ± SDs of the absolute numbers of BCL2 expressing immune cells from patients with CLL and healthy donor controls are shown in the graph. E The mean values ± SDs of the relative contributions of PD-L1+ CD14+ CD16++ non-classical, PD-L1+ CD14++ CD16+ intermediate, PD-L1+ CD14++ CD16− classical monocytes and PD-L1+ CD14+HLADRlow/− M-MDSCs from patients with CLL and healthy donor controls are shown in graph. F The mean values ± SDs of the relative contributions of PD1+ CD8+ CD45RA−CD45RO+CD62L− TEM and PD1+ CD8+ CD45RA-CD45RO+CD62L+ TCM from patients with CLL and healthy donor controls are shown in graph. G MISTRG mice were adoptively transferred on day 0 with monocytes and M-MDSCs (i.v.: 50,000-546,000 cells depending on the patient, including NC, I, C monocyte subsets and M-MDSCs) separated using fluorescence-activated cell sorting from the PBMCs of patients 413, 680, 915 (Table S2) and NS-transfected for 15 min with 50pM of either miR-16-1 miRNA mimic or miRNA mimic control. Mice were euthanized 7 days later for the analysis of patient-derived macrophages. H miR-16-1 expression relative to U6 control in myeloid cell pools (including NC, I, C monocytes and M-MDSCs) separated from PBMCs of CLL patients (n = 3, patients 413, 680, 915, Table S2) and then NS-transfected with miR-16-1 miRNA mimic (miR-16-1 mimic), with mimic control (mimic control) or left untreated (Unt) is shown in the graph. I The percentages of BCL2+, PD-L1+, CD206+, CD163+ to the whole CD68+ macrophage pool gated on live cells detected in the PB of MISTRG mice are shown in the graphs. J Flow cytometry detection of BCL2, PD-L1, CD206 and CD163 by human CD68+ macrophages in the PB of MISTRG mice transplanted with cells from representative patient 915. The percentage of BCL2, PD-L1, CD206 and CD163-expressing macrophages is indicated. A statistical analysis was performed using the Student’s t test *P < 0.05, **P < 0.01, ***P < 0.001.

Finally, as frozen samples became available, we retrospectively analyzed, together with miRNAs, the levels of human BCL2/PD1/PD-L1 target proteins on immune cells from patients with CLL and compared them with the levels in age-matched, healthy donor controls (n = 6; Fig. 7B–F; Fig. S4B). This set of patient samples includes sub-cohort 1A within cohort 1 (Table S1, patients 1,2,3) and 3 patients enrolled into the investigator-initiated phase 2 trial NCT02756897 [15] with BTK inhibitor ibrutinib (IBT) and the BCL2 inhibitor venetoclax (Ven) (Table S1, patients 51, 52, 68). The immune cell composition of patients treated with IBT/Ven was longitudinally investigated in a larger cohort of patients, including the 3 patients mentioned in regard to Fig. 7, and is described in Fig. S5A–I. Low levels of miR-15a/miR-16-1 and high levels of BCL2 protein have been observed in B cells and immune cells from patients with CLL and from healthy donor controls (Fig. 7B–D, Fig. S4B), while increasing levels of PD1/PD-L1 checkpoint molecules were observed in immune cells from patients with CLL compared with healthy donor controls (Fig. 7E, F). As expected, in patients with CLL, the majority of cells expressing BCL2 were B cells (Fig. 7D).

To mechanistically evaluate the involvement of miR-15a/miR-16-1 in the differentiation of monocytes into macrophages and ultimately their protumor immunophenotype, we established a patient-derived xenograft (PDX) system in the humanized MISTRG mice. This strain particularly facilitates the engraftment of human innate immune cells including monocytes [31]. When adoptively transferred into MISTRG mice, patient-derived monocyte subsets and M-MDSCs differentiate in vivo into tumor-associated macrophages (Fig. S6). We exploited the NanoStraw technology [32, 33] to transfect patient derived monocytes and M-MDSCs (Fig. S7) with either miR-16-1 or miR-15a mimics. When we forced the expression of miR-16-1 on a pool of monocytes and M-MDSCs from 3 patients with CLL (Table S2, patients 413, 680, 915) and we analyzed macrophages in MISTRG mice (Figs. S8–S9, 7G–J), we found a downregulation of BCL2/PD-L1 related target proteins together with a downmodulation of TAM protumor markers including CD163 and CD206 (Fig. 7I, J). This regulation on macrophages was not induced by the forced expression of miR-15a (Fig. S10).

Overall, our findings confirmed the previous evidence on miR-15a/miR-16-1 on B and T cells, and our study was the first to investigate the role of miR-15a/miR-16-1 cluster on human monocyte subsets, M-MDSCs and macrophages. We demonstrated that human monocyte subsets and M-MDSCs have low expression levels of miR-15a/miR-16-1 and express BCL2 and PD-L1. We ultimately demonstrated that forced expression of miR-16-1 mitigates the protumor immunophenotype of monocyte-derived macrophages.