A SCA1+ tumor cell population is enriched during tumor progression and has higher in vivo metastatic capacity. To investigate tumor cell heterogeneity during tumor progression, we first examined the expression of reported breast CSC markers Cd24, Cd44, Cd61, Sca1, and Cd49f (38) in the RNA-Seq data set from Ross et al. (39) encompassing several murine breast cancer models. This data set includes data from cultured cancer cells (In_Culture), orthotopic primary tumors (OT_PT), spontaneous lung metastases (OT_LuM), and experimental lung metastases (i.e., upon tail-vein injection, TV_LuM) (Figure 1A and Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/JCI166847DS1). Sca1 expression was consistently elevated in spontaneous lung metastasis in 4T1, 6DT1, Mvt1, and Met1 models compared with the respective primary tumors (PTs). In 4T1, 6DT1, and Mvt1 models, Sca1 expression was also elevated in experimental lung metastases compared with cultured cells. The expression of Cd24, Cd44, Cd61, and Cd49f was either unaltered or exhibited inconsistent patterns during progression (Supplemental Figure 1A). Thus, increased expression of Sca1 is consistently observed during metastasis across different preclinical breast cancer models.

SCA1+ population is enriched during in vivo metastasis across multiple breast cancer models. (A) Sca1 mRNA expression in the metastatic murine breast cancer models 4T1, 6DT1, Mvt1, and Met1, extracted from the Ross data set (39). Analyzed samples consist of cultured cells (In_Culture), orthotopic injected PTs (OP_PT), spontaneous lung metastases (OP_LuM), and lung metastases induced by i.v. injection (TV_LuM). Data are represented as the mean of reads per kilobase of transcript per million mapped reads (RPKM) ± SD. (B) Experimental setup for in vivo experimental validation. 4T1 tumor cells were orthotopically injected into the fourth mammary fat pad. Thirty days later, cells from PTs and lungs were isolated to examine CSC marker expression by flow cytometry. (C) Frequency of CSC marker expression in PTs and lung metastases. Results are shown as percentages of CD24-, CD44-, SCA1-, CD61-, and CD49f-positive cells gated in lineage-negative cells (CD45–CD31–TER119–). n = 6/group. (D–F) Experimental setup (D) of the in vivo experiment to assess tumor growth (E) and lung metastatic ability (metastatic index) (F) of 4T1-SCA1+ and 4T1-SCA1– populations isolated from tumors induced by parental 4T1 cells orthotopically injected into the fourth mammary fat pads. Metastases are assessed 21 days after tumor cell injection. n = 8/group. (G–I) Experimental setup (G) of in vivo experiment to assess lung colonization capacity upon tail-vein injection of sorted parental 4T1, 4T1-SCA1+, and 4T1-SCA1– cells. Lung metastatic nodule numbers (H) and representative images (I) of lungs from mice 10 days after injection. n = 5–6/group. Scale bar: 1 mm. Data are represented as means ± SEM and are representative of 3 independent experiments for C, E, F, and H. P values were calculated using unpaired, 2-tailed Student’s t test with Holm’s correction (A); unpaired 2-tailed Student’s t test (C and F); 2-way ANOVA with Tukey’s multiple-comparison test (E); or 1-way ANOVA with Tukey’s multiple-comparison test (H). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

To investigate whether the increased Sca1 expression within the tumor mass was due to increased gene expression in all cancer cells or to the enrichment of a SCA1+ population, we orthotopically injected 4T1 cells and determined the frequency of different cell populations present in the PT and lung metastases 30 days later by flow cytometry (Figure 1B). The frequency of a SCA1+ population, and to a lesser extent of a CD61+ population, increased in lung metastases compared with PTs (Figure 1C). In contrast, the CD24+, CD44+, and CD49f+ populations were not significantly altered. Enrichment of SCA1+ and CD61+ populations in lung metastasis was confirmed in the D2A1 breast cancer model (Supplemental Figure 1B).

These observations prompted us to ask whether SCA1+ cells actively contribute to metastasis formation. To this end, we isolated 4T1-SCA1+ and 4T1-SCA1– cells by magnetic activated cell sorting (MACS) from parental 4T1 cells, which contain low frequencies of SCA1+ populations (10%–15%) (Supplemental Figure 1C), and examined their metastatic capacity in vivo. Orthotopically injected 4T1-SCA1+ cells formed significantly more lung metastases than 4T1-SCA1– cells, while no significant difference was observed in PT growth (Figure 1, D–F). Upon tail-vein injection, 4T1-SCA1+ cells displayed significantly greater lung colonization ability compared with 4T1-SCA1– cells (Figure 1, G–I). In addition, 4T1-SCA1+ cells showed significantly higher in vitro mammosphere-forming efficiency than 4T1-SCA1– cells (Supplemental Figure 2A), while in vitro cell growth and cell motility were comparable (Supplemental Figure 2, B and C). These results suggest that 4T1-SCA1+ and 4T1-SCA1– cells have similar tumorigenic potential, while 4T1-SCA1+ cells have higher mammosphere-forming and metastasis-initiating capacities.

The SCA1+ tumor cell population is plastic in vitro and in vivo. Growing evidence indicates that cancer cells possess plastic features modulated by both cell-intrinsic factors and microenvironmental cues (39, 40). To characterize the observed plasticity of 4T1-SCA1+ cells, we first investigated isolated 4T1-SCA1+ and 4T1-SCA1– cells in vitro. The frequency of the SCA1+ population in MACS-enriched 4T1-SCA1+ cells (>93%) gradually decreased to around 50% after 4 days of culture, while the SCA1– cells (>99% negatively enriched by MACS) regenerated a SCA1+ population (from <1% to >20%) (Supplemental Figure 2D). Consistently, after orthotopic injection of 4T1-SCA1+ cells, the abundance of the SCA1+ population from the derived tumors decreased from 85% after MACS enrichment at injection time to about 40% after 23 days of in vivo growth, while in tumors generated from 4T1-SCA1– cells, it increased from less than 1% to 15% (Supplemental Figure 2E), similar to the frequency of SCA1+ population in tumors derived from parental 4T1 cells (Figure 1C). When tumor cells derived from PTs and lung metastases of parental 4T1–injected mice were cultured ex vivo, the abundance of the SCA1+ population significantly decreased, from 20% to 4.5% and from 60% to 19%, respectively (Supplemental Figure 2F). From these observations, we conclude that both SCA1+ and SCA1– populations exhibit high plasticity, which can be modulated by the TME in vivo.

Tu-Gr1+CD11b+ cells demonstrate higher potency in expanding the metastatic SCA1+ population. Immune cells in the TME are critical determinants of tumor cell behavior, including metastatic capacity (41). To collect evidence for a potential correlation between immune cells and the SCA1+ population, we characterized the inflammatory cells infiltrating the PTs following orthotopic injection of 4T1-SCA1+ and 4T1-SCA1– cells. We observed a significant increase of the Gr1hiLy6G+Ly6CloCD11b+ population and a significant decrease of the Gr1loLy6G–Ly6Chi CD11b+ population in tumors from 4T1-SCA1+– compared with 4T1-SCA1––injected mice (Figure 2A). To examine the direct contribution of the Tu-Gr1+CD11b+ cells in promoting the enrichment of the SCA1+ population, we MACS isolated Gr1+ cells from PTs, bone marrow, and spleen of 4T1 tumor–bearing BALB/c mice and cocultured them in vitro with parental 4T1 or sorted 4T1-SCA1– cells (Figure 2B). Gr1+ cells isolated from tumors (Tu-Gr1+CD11b+) induced a significant expansion of the SCA1+ population from both 4T1 and 4T1-SCA1– cells within 48 hours (Figure 2C). In contrast, Gr1+ cells from the spleen (Spl-Gr1+CD11b+) or bone marrow (BM-Gr1+CD11b+) were less efficient and induced the SCA1+ population only from 4T1-SCA1– cells at lower frequency (between 10% and 20%), comparable to unsorted 4T1 cells (Figure 2C). Interestingly, the Tu-Gr1+CD11b+–induced SCA1+ population appeared more stable in time compared with the isolated 4T1-SCA1+ cells (Supplemental Figure 2G). Moreover, Tu-Gr1+CD11b+ cells, rather than BM-Gr1+CD11b+ or Spl-Gr1+CD11b+ cells, were more effective in expanding the SCA1+ population in D2A1 cells, which have low intrinsic frequency of SCA1+ cells (4%–5%) (Supplemental Figure 3, A and B).

SCA1 expression is modulated by TME. (A) Frequency of different immune cell populations in PTs of mice orthotopically injected with 4T1-SCA1+ and 4T1-SCA1– cells 21 days after injection. Populations are determined in CD45-positive, viable cells. n = 8 mice/group. (B and C) Schematic (B) showing experimental design for isolating Gr1+ cells from different sites of tumor-bearing mice. Twenty-one days after tumor implantation, Gr1+ cells were isolated from bone marrow (BM-Gr1+CD11b+), spleen (Spl- Gr1+CD11b+), or PT (Tu-Gr1+CD11b+) and cocultured for 48 hours with parental 4T1 or sorted 4T1-SCA1– cells in vitro. SCA1 expression in tumor cells was examined by flow cytometry (C). Coculture conditions are indicated in bar graph. n = 3/group for 4T1; n = 5–9/group for 4T1-SCA1–. (D and E) Schematic of experimental coculture setup (E). MACS-enriched Gr1+ cells were cocultured with 4T1 or sorted 4T1-SCA1– cells with or without Transwell inserts of 0.4 μm pore size. Cells were seeded in bottom well and Gr1+CD11b+ cells in upper part of insert. After 48 hours, tumor cells were examined for SCA1 expression by FACS (E). Coculture conditions are indicated in bar graph. Ratio of tumor cells and Tu-Gr1+CD11b+ varied from 1:1 to 1:3. n = 3/group for 4T1; n = 5–9/group for 4T1-SCA1–. (F–H) Schematic of experimental metastasis setup for evaluation of metastatic capacity of Gr1+CD11b+-educated 4T1 cells in vivo (F). 4T1 tumor cells were primed with Tu-Gr1+CD11b+ or Spl-Gr1+CD11b+ in vitro without cell-cell contact for 48 hours and injected into tail veins. Lung metastases were quantified 10 days after injection (G). Representative H&E staining images of lung sections are shown (H). Scale bar: 1 mm. n = 6 mice/group. Data are represented as means ± SEM and are representative of 3 independent experiments. P values were calculated using unpaired 2-tailed Student’s t test (A and H) or 1-way ANOVA with Dunnett’s multiple-comparison test (C and E). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

To determine whether the expansion of a SCA1+ population requires direct contact with Tu-Gr1+CD11b+ or is mediated by soluble factors, we compared the induction in standard wells (contact) versus Transwells (without contact) (Figure 2D). There was no significant difference in the expansion of the SCA1+ population between these 2 conditions in 4T1 cells. Additionally, increasing the ratio of 4T1 to Tu-Gr1+CD11b+ cells from 1:1 to 1:3 did not lead to further expansion of the SCA1+ population. Similar results were observed in 4T1-SCA1– cells, which, however, appeared to rely more on cell-cell contact conditions (Figure 2E). Tu-Gr1+CD11b+ cells were also able to expand the SCA1+ population in D2A1 cells contact free in a dose-dependent manner (Supplemental Figure 3C). Importantly, when Tu-Gr1+CD11b+ cell–primed 4T1 cells were injected into the tail vein, they formed more lung metastases compared with Spl-Gr1+CD11b+ cell–primed 4T1 cells (Figure 2, F–H).

Tu-Gr1+CD11b+ is a mixed cell population, including Gr1hiLy6G+Ly6CloCD11b+ tumor-associated neutrophils (TANs) and Gr1loLy6G–Ly6C+CD11b+ monocytic cells. Both of these cell populations originate from Gr1+CD11b+ immature myeloid progenitors mobilized from the bone marrow to the TME and promote tumor progression and metastasis (42, 43). Their surface markers broadly overlap, impairing phenotypical distinction (43). In tumor-bearing mice, we observed homogenous Gr1+CD11b+ cells in the circulation, which is consistent with a previous report (44), while in the TME, they differentiated into Gr1hi and Gr1lo subpopulations in both 4T1- and D2A1-injected mice (Supplemental Figure 3D). To elucidate differences in their capacity to induce the SCA1+ population, we isolated Gr1hiCD11b+ and Gr1loCD11b+ cells from PTs by FACS sorting and cocultured them with 4T1 or 4T1-SCA1– cells (Supplemental Figure 3E). Both Gr1hi and Gr1lo subpopulations demonstrated the ability to expand the SCA1+ population, with Gr1hiCD11b+ cells being more potent (Supplemental Figure 3F). Together, these results imply that Tu-Gr1+CD11b+ cells induce the emergence of a highly metastatic SCA1+ population through secreted factors.

Tu-Gr1+CD11b+–induced and tumor-inherent SCA1+ populations display distinct gene expression profiles. To unravel the molecular basis for the metastatic capacity of the inherent 4T1-SCA1+ population and the Tu-Gr1+CD11b+–induced SCA1+ population, we first performed transcriptomic profiling of 4T1-SCA1+ and 4T1-SCA1– cells isolated from the parental 4T1 line. Pathway enrichment analysis showed that 4T1-SCA1+ and 4T1-SCA1– cells expressed genes associated with distinct signaling pathways (Figure 3A). The top 200 significantly upregulated and downregulated genes were extracted as SCA1-positive and SCA1-negative signatures, respectively (Supplemental Table 1). Next, we performed transcriptomic profiling of Tu-Gr1+CD11b+–primed 4T1, Spl-Gr1+CD11b+-primed 4T1, and parental 4T1 cells. Pathway enrichment analysis revealed that Tu-Gr1+CD11b+ and Spl-Gr1+CD11b+ priming induced distinct transcriptomic alterations in 4T1 cells (Figure 3B). Comparative analysis revealed that Tu-Gr1+CD11b+–primed 4T1 cells expressed both SCA1-positive and SCA1-negative signatures (Figure 3C), suggesting that Tu-Gr1+CD11b+ may twist tumor plasticity by converting the 4T1-SCA1– cells into 4T1-SCA1+ cells, rather than expanding the preexisting 4T1-SCA1+ population.

Transcriptomic analysis of SCA1+ tumor cells. (A) Heatmap showing the signature score of the hallmark pathways analysis in 4T1-SCA1+ and 4T1-SCA1– populations sorted from parental 4T1 cells. The colors code the expression levels relative to average levels, as indicated at the bottom. (B) Heatmap showing the signature score of the hallmarks pathway analysis in parental 4T1, Spl-Gr1+CD11b+–primed 4T1, and Tu-Gr1+CD11b+–primed 4T1 cells. The colors code the expression levels relative to average levels, as indicated at the bottom. (C) GSEA comparing the Tu-Gr1+CD11b+– and Spl-Gr1+CD11b+–primed 4T1 cells. GSEA shows positive correlations of both SCA1-positive and SCA1-negative signatures. NES, normalized enrichment score. (D) Venn diagrams showing that 56 upregulated genes and 1 downregulated gene are shared between inherent and Tu-Gr1+CD11b+–induced SCA1+ population in 4T1 tumor cells. (E) UMAP plot showing clusters of cancer cells and myeloid cell populations in orthotopically grown 4T1-derived PTs extracted from the Sebastian data set (see Methods for details). (F) Circos diagram showing the predicted potential interactions between cancer cells and different myeloid cell populations determined by CellPhoneDB (see Supplemental Methods for details) based on the Sebastian data set. Only Osmr and P2ry6 are shared with the common 56 gene list shown in panel D.

To test for possible common molecular mechanisms underlying the induction and metastatic capacity of the different SCA1+ populations, we compared the significantly differentially expressed genes (DEG) between 4T1-SCA1+ versus 4T1-SCA1– and Tu-Gr1+CD11b+–primed 4T1 versus Spl-Gr1+CD11b+–primed 4T1 cells. Strikingly, among a total of 1118 upregulated and 423 downregulated DEGs in the 2 conditions, only 56 upregulated and 1 downregulated genes were shared (Figure 3D and Supplemental Table 1). This was consistent with the notion that the SCA1+ population in Tu-Gr1+CD11b+–primed 4T1 cells was different from the inherent 4T1-SCA1+ cells. Nonetheless, the fact that the 4T1-SCA1+ cells and Tu-Gr1+CD11b+–primed 4T1 cells possessed similar in vivo metastatic capacities suggested that, among these common pathways, some were relevant for 4T1 metastasis formation. To address this question, we analyzed the scRNA-Seq data set from 4T1 PTs of Sebastian et al. (45) encompassing several cell types, including cancer cells, epithelial cells, fibroblasts, and distinct subpopulations of myeloid cells. To determine relevant ligand-receptor interactions between epithelial/cancer cells and myeloid/monocytic cells, we performed cell-cell interaction analysis from the scRNA-Seq data with CellPhoneDB (46) (Figure 3, E and F). The analysis identified 160 ligand-receptor interaction pairs (Supplemental Table 2). Among those pairs, OSM receptor (OSMR) and pyrimidinergic receptor P2Y6 (P2RY6) were the only ones present among the 56 common genes shown in Figure 3D. However, P2RY6 interacts with coatomer complex subunit α (COPA), a membrane protein involved in membrane traffic between endoplasmic reticulum and Golgi (47) and thus unlikely to mediate cell-cell contact-independent induction of the SCA1+ population. As the IL-6/JAK/STAT3 signaling pathway was upregulated both in the 4T1-SCA1+ population and Tu-Gr1+CD11b+–primed 4T1 cells (Figure 3, A and B), we examined the expression of Osm, Osmr, Il6st, Il6, and Il6 receptor (Il6ra) in the Sebastian data set (Supplemental Figure 4A). The expression of Il6 and Osm was restricted to myeloid cells, with Osm expression being more prominent, consistent with a previous report (34). Osmr was predominantly expressed in tumor cells, while Il6st and Il6ra were homogenously expressed in all cell types.

In contrast, Osm and Il6 expression were very low in all 4T1-SCA1+ cells and Tu-Gr1+CD11b+–primed 4T1 cell–related samples (normalized count number less than 7 on average) (Supplemental Figure 4, B and C). Osmr and Il6ra, however, were highly expressed in 4T1-SCA1+ cells compared with 4T1-SCA1– cells, while the expression of Il6st was abundant in both populations (Supplemental Figure 4B). Meanwhile, Osmr was significantly upregulated in Tu-Gr1+CD11b+–primed 4T1 compared with Spl-Gr1+CD11b+-primed 4T1 (Supplemental Figure 4C). Taken together, these results suggest that OSM/OMSR and IL-6/IL-6R signaling pathways may be involved in expansion of the SCA1+ population mediated by Tu-Gr1+CD11b+ cells.

Tu-Gr1+CD11b+ cells promote SCA1– to SCA1+ population conversion. The above results strongly implied that Tu-Gr1+CD11b+ cells convert SCA1– cells into the SCA1+ population. To further test this hypothesis, we compared the cell population dynamics in cultured cells and the orthotopic PTs by analyzing publicly available scRNA-Seq data sets. By integrating scRNA-Seq data from 3D cultured 4T1 cells (NCBI’s Gene Expression Omnibus [GEO] GSM4812003) (48) and tumor cells isolated from orthotopically injected 4T1 PTs (GEO GSM3502134) (49) (Figure 4, A and B), we observed 5 clusters: clusters 0, 1, 2, and 4 were predominant in cultured tumor cells, while cluster 3 was predominant in PT. Single-cell trajectories analysis confirmed that cluster 3 was at the end of the transformation process (Figure 4, C and D). The population dynamics also showed that the fraction of cells in clusters 1, 2, and 4 decreased during the transformation and the one in cluster 0 only minimally increased, while the fraction in cluster 3 massively increased (Figure 4E). Importantly, very few cultured 4T1 cells expressed Sca1, while it was abundantly expressed in most cells in the PT (Supplemental Figure 5A). Consistently, the fraction of cells expressing Osmr was higher in the PT compared with cultured cells (Supplemental Figure 5A). Similar observations were obtained when analyzing scRNA-Seq data from the ER+ human breast cancer model MCF-7. After integrating data from cultured MCF-7 cells (GEO GSM4681765) and MCF-7 cells that were isolated from tumors generated by MCF-7 cells injected in the mammary gland intraductally (GEO GSM5904917) (50), 6 clusters were identified (Figure 4F), with clusters 1 and 3 predominant in cultured MCF-7 cells, while clusters 2 and 4 were prevalent in PTs (Figure 4G). Clusters 2 and 4 expanded during in vitro–to–in vivo tumor cell transformation and represented nearly 50% of the in vivo PT cells (Figure 4, H–J). Although there is no human homolog of the Sca1 gene for comparison, OSMR-expressing cells were increased upon tumor implantation, especially in clusters 2 and 1 (Supplemental Figure 5B).

Transformation dynamics of tumor cell populations induced by the TME. (A) UMAP plots showing 4T1 clusters based on integrated scRNA-Seq data from 4T1 cells in 3D culture or in PT. (B) Distribution of specific clusters in 4T1 cells in 3D culture or PT. (C and D) UMAP plot (C) and box plot (D) showing the clusters in pseudo-time course during the transformation of 4T1 cells from ex vivo culture to in vivo. (E) Sankey diagram showing the dynamic of each cluster during the transformation of 4T1 cells from ex vivo culture to in vivo. Cluster 3 was largely expanded in vivo. (F) UMAP plots showing MCF-7 clusters based on integrated scRNA-Seq data from MCF-7 cells in culture or in PT. (G) Distribution of the specific clusters in cultured MCF-7 cells or MCF-7 PTs. (H–I) UMAP plot (H) and box plot (I) showing the clusters in pseudo-time course during the transformation of MCF-7 cells from ex vivo culture to in vivo. (J) Sankey diagram showing the dynamic of each cluster during the transformation of MCF-7 cells from ex vivo culture to in vivo. Clusters 2 and 4 were largely expanded in vivo. (K and L) GSEA analysis of SCA1-positive signature, SCA1-negative signature, and Tu-Gr1+CD11b+–induced signature of cells in cluster 3 in 4T1 data (K) and in cluster 2 and cluster 4 in MCF-7 data (L). Analyses are based on publicly available data (4T1: GEO GSM4812003 and GSM3502134; MCF-7: GEO GSM4681765 and GSM5904917).

To investigate the signals involved in this transformation, we performed Gene Set Enrichment Analysis (GSEA) for cluster 3 in 4T1 cells and clusters 2 and 4 in MCF-7 cells, respectively (Figure 4, K and L, and Supplemental Figure 5, C and D). By comparing the Hallmark gene signatures, the IL-6/JAK/STAT3 signature was significantly upregulated in both cell populations (Supplemental Figure 5, C and D). Interestingly, SCA1-positive and SCA1-negative signatures were both upregulated (Figure 4, K and L), coherently with our ex vivo induction experiment (Figure 3C). To validate the involvement of Tu-Gr1+CD11b+ during the cell population transformation, we extracted the top 50 upregulated genes (Supplemental Table 3) identified by comparing the Tu-Gr1+CD11b+– with the Sp-Gr1+CD11b+–stimulated 4T1 cells as Tu-Gr1+CD11b+–induced signature. The clusters predominant in the PT in both 4T1 (cluster 3) and MCF-7 (clusters 2 and 4) models showed an upregulation of the Tu-Gr1+CD11b+–induced signature (Figure 4, K and L). These data, together with the ex vivo coculture experiments (Figure 2, B–G, and Figure 3C), demonstrate that Tu-Gr1+CD11b+ cells convert the SCA1– population into SCA1+ population, possibly via the OSM/IL-6 signaling pathway.

OSM/IL-6/JAK pathway mediates Tu-Gr1+CD11b+–induced SCA1+ population enrichment. To experimentally interrogate the role of OSM/IL-6 in modulating the SCA1+ population, we first measured Osm and Il6 mRNA expression by quantitative PCR (qPCR) in Spl-Gr1+CD11b+ and Tu-Gr1+CD11b+ cells and protein levels by ELISA in conditioned supernatants. Indeed, we found that the expression of Osm and Il6 in both mRNA and protein levels was significantly elevated in Tu-Gr1+CD11b+ cells (Figure 5, A and B). In addition, higher OSM and IL-6 protein secretion were observed in Tu-Gr1hiLy6G+CD11b+ cells compared with Tu-Gr1loLy6C+CD11b+ cells (Figure 5C), correlating with their ability to induce 4T1-SCA1+ cells (Supplemental Figure 3E).

SCA1+ population is modulated by OSM/IL-6/JAK pathway. (A) Relative Osm and Il6 mRNA expression in Tu-Gr1+CD11b+ and Spl-Gr1+CD11b+ cells. n = 4–5/group. (B and C) OSM and IL-6 protein quantification in supernatant of (B) Tu-Gr1+CD11b+ and Spl-Gr1+CD11b+ (C) and Tu-Gr1hiLy6G+CD11b+ and Gr1loLy6G–CD11b+ cells. n = 4/group. (D) Fraction of SCA1+ cells of parental 4T1 or sorted 4T1-SCA1– cells upon exposure to OSM or IL-6 (10 ng/ml 48 hours). n = 6/group. (E) Effect of inhibition of OSM and IL-6 in Tu-Gr1+CD11b+ conditioned medium on SCA1 expression in parental 4T1 or sorted 4T1-SCA1– cells after 48 hours treatment. n = 3–6/group. (F) Experimental design for examining lung metastatic capacity of IL-6/OSM. 4T1-SCA1– cells were primed by Tu-Gr1+CD11b+ conditioned medium with dual depletion of OSM/IL-6 depleted or control, in vitro for 48 hours and then injected into tail vein. Lungs were examined for metastasis 10 days later. (G and H) Quantification of lung metastases (G) and representative H&E staining images of lung sections (H). n = 8–10/group. Scale bar: 1 mm. (I and J) Quantification of lung metastases in mice injected with 4T1 control and 4T1 silenced for Sca1 (shSca1-120 and shSca1-597) (I). n = 8/group. Representative H&E stained lung sections (J). Scale bar: 1 mm. (K) SCA1+ population stimulation in cultured parental 4T1 or sorted 4T1-SCA1– cells by recombinant OSM or IL-6 (10 ng/ml, 48 hours) in vitro in presence of ruxolitinib (5 μM) or DMSO control. n = 3–6/group. Data are represented as means ± SEM from 3 independent experiments. P values were calculated using unpaired 2-tailed Student’s t test (A–C and G); unpaired 2-tailed Student’s t test with Holm’s correction (I); 1-way ANOVA with Dunnett’s multiple-comparison test (D and K); or 1-way ANOVA with Tukey’s multiple-comparison test (E). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

To functionally validate the role of OSM/IL-6 in promoting the SCA1+ population, we first treated 4T1, 4T1-SCA1–, or D2A1 cells for 2 days with recombinant OSM and IL-6 proteins in vitro. Both treatments significantly increased the frequency of the SCA1+ population (Figure 5D and Supplemental Figure 6A). To validate the functional role of OSM and IL-6 produced by Tu-Gr1+CD11b+ cells, we set up an inhibition experiment. First, we confirmed that the conditioned medium produced by Tu-Gr1+CD11b+ cocultured with 4T1 cells retained the ability to enhance the SCA1+ population in 4T1 or 4T1-SCA1– cells (Supplemental Figure 6, B and C). Next, we depleted OSM and IL-6 in the Tu-Gr1+CD11b+ conditioned medium using neutralizing antibodies against OSM or IL-6 and cultured with 4T1 cells, resulting in a significant reduction of the emergence of the SCA1+ population from 4T1 or 4T1-SCA1– cells (Figure 5E). The combination of anti-OSM and anti–IL-6 antibodies did not have additive effects, suggesting that OSM and IL-6 contribute to promote the SCA1+ population through the same pathway. Importantly, tail-vein injection of 4T1-SCA1– cells treated ex vivo with OSM/IL-6–depleted Tu-Gr1+CD11b+ conditioned medium blunted its prometastatic effect in vivo, thus demonstrating the role of Tu-Gr1+CD11b+–derived OSM/IL-6 in promoting metastasis (Figure 5, F–H).

Next, we asked whether SCA1 itself functionally contributes to the OSM/IL-6–induced metastatic population in 4T1 cells. To this end, we stably silenced SCA1 in 4T1 cells by lentiviral-mediated shRNA delivery. The Sca1 silencing efficiency was validated via OSM treatment in vitro (Supplemental Figure 6D). Sca1 silencing with 2 independent shRNAs did not affect PT growth upon orthotopic cell injection nor did it reduce lung metastasis formation, indicating that SCA1 itself is not the mediator of metastasis (Figure 5, I and J, and Supplemental Figure 6, E and F).

OSMR and IL-6R signal by activating the intracellular JAK (51). To explore the involvement of the JAK pathway in the emergence of the SCA1+ population, we treated tumor cells with a JAK inhibitor (ruxolitinib) during exposure to recombinant OSM and IL-6. Ruxolitinib treatment prevented the emergence of the SCA1+ population in response to recombinant OSM and IL-6 in 4T1, 4T1-SCA1–, and D2A1 cells (Figure 5K and Supplemental Figure 6G). From these data, we conclude that the OSM/IL-6/JAK pathway has a crucial role in mediating the enrichment of the highly metastatic SCA1+ population induced by Tu-Gr1+CD11b+.

Tu-Gr1+CD11b+–induced SCA1+ population and 4T1-inherent SCA1+ population have distinct CSC and EMT gene expression profiles. OSM/IL-6/JAK signaling has been reported to support tumor progression by promoting a CSC phenotype and epithelial-mesenchymal plasticity (26, 37, 52). To characterize CSC and EMT features in 4T1-SCA1+ cells and a Tu-Gr1+CD11b+–induced SCA1+ population, we interrogated our RNA-Seq data sets for the expression of 9 stem cell and 8 EMT marker genes (Supplemental Figure 7). The expressions of the stem cell markers Oct4 (Pou5f1), Sox2, and Nanog were undetectable or very low in all (or some) samples. 4T1-SCA1+ cells showed higher expression of Aldh1a1, Aldh3a1, and Podxl, but lower expression of Klf4 and Sox9 compared with 4T1-SCA1– cells. There was no difference in the expression of Abcg2 and Has2. Tu-Gr1+CD11b+–primed 4T1 cells had higher expression of Klf4 and Has2, lower expression of Aldh1a1, Aldh3a1, and Sox9, and similar expression of Podxl when compared with Spl-Gr1+CD11b+–primed 4T1 cells. Among them, only Has2 expression was specifically elevated in Tu-Gr1+CD11b+–primed 4T1 cells compared with control 4T1 and Spl-Gr1+CD11b+–primed 4T1 cells (Supplemental Figure 7A). On the other hand, 4T1-SCA1+ cells had lower expression of Cdh1 and higher expression of Snail1, Twist1, Vim, and Foxc1, which support an EMT status, although Zeb1 expression was reduced (Supplemental Figure 7B). Globally, the expression of most of the EMT genes was similar between Tu-Gr1+CD11b+– and Spl-Gr1+CD11b+–primed 4T1 cells, except for Snail2 and Vim, whose expressions were suppressed in Tu-Gr1+CD11b+–primed 4T1 (Supplemental Figure 7B). Collectively, these results indicate that the Tu-Gr1+CD11b+–induced SCA1+ population and inherent SCA1+ population display different CSC and EMT transcriptional profiles, reinforcing the notion that Tu-Gr1+CD11b+–induced SCA1+ populations are not just enriched tumor-inherent SCA1+ populations.

Chemotherapy enriches a SCA1+ population with CSC features. CSC and cancer cell plasticity contribute to drug resistance in various tumor types (53–57). The above results, including the significantly elevated expression of Aldh3a1 (a marker of drug resistance) in the 4T1-SCA1+ population, prompted us to investigate the resistance to chemotherapy of this population. To this end, we treated 4T1 cells for 48 hours in vitro with methotrexate (MTX) or doxorubicin (Dox), 2 widely used chemotherapy drugs. The 48-hour treatment with either drug increased the frequency of the SCA1+ population in 4T1 cells (Supplemental Figure 8A). To mimic the clinically relevant situation of cancer cells escaping chemotherapy treatment, we exposed 4T1 cells to a MTX concentration slightly higher than the IC50 of the drug (28 nM) for up to 3 weeks and recovered the surviving cells by switching to normal culture medium (Figure 6A). The selected MR13 cell line was highly enriched in SCA1+ cells (>60%) (Supplemental Figure 8B). Compared with parental 4T1 cells, MR13 cells exhibited a higher mammosphere-forming efficiency (Figure 6B), lower proliferative capacity (Figure 6C), and increased cell mobility (Figure 6D) in vitro, consistent with CSC-like properties. When tested in a 48-hour cytotoxicity assay, MR13 cells were more resistant to MTX compared with parental 4T1 (Supplemental Figure 8C).

Long-term chemotherapy treatment of 4T1 cells induces a stable SCA1+ population (MR13) with higher metastatic capacity and CSC features. (A) Schematic of the experimental design to obtain chemotherapy-resistant MR13 cells from 4T1. (B) Quantification of the mammosphere-forming efficiency of 4T1 and MR13 tumor cells. n = 5–6/group. (C) Cell proliferation curves of 4T1 and MR13 tumor cells in vitro determined by the CV assay. Results are represented as mean of OD. n = 8/group. (D) Cell motility of 4T1 and MR13 tumor cells determined by the scratch wound healing assay. n = 5–6/group. Results are represented as cell-free area relative to the initial wound area from 3 independent experiments. (E) Growth curves of PTs in BALB/c mice orthotopically injected with 4T1 and MR13 tumor cells. n = 10–11/group. (F) Tumor weight of 4T1 and MR13 tumors recovered from BALB/c mice at day 22 after injection. n = 8/group. (G and H) Lung metastasis index 23 days after injection. The number of metastatic nodules is determined by H&E staining and normalized based on the PT weight (G). Representative H&E staining images of lung sections are shown (H). Scale bar: 1 mm. n = 8/group. (I) Frequency of different CD11b+ myeloid cells subpopulations in PTs from MR13- and 4T1-injected mice determined by flow cytometry 21 days after injection. n = 6/group. Subpopulations are determined in CD45-positive, viable cell population. (J) Percentage of SCA1+ tumor cells at time of injections (D0) of 4T1 and MR13 cells and in PTs recovered at day 21 (D21). SCA1 expression is determined in CD45-negative, viable cell population. n = 6/group. Data are represented as means ± SEM and are representative of 3 independent experiments. P values were calculated using unpaired, 2-tailed Student’s t test (B, F, G, I and J) or 2-way ANOVA with Tukey’s multiple comparison test for (C–E). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

MR13-derived tumors are highly metastatic and enriched in Gr1hiLy6G+CD11b+ cells. To characterize the in vivo behavior of MR13 cells, we implanted them orthotopically into BALB/c mice. MR13 cells formed smaller PTs that were more metastatic to the lung (Figure 6, E–H) and enriched in Gr1hiLy6G+CD11b+ cells compared with 4T1-derived tumors (Figure 6I), similarly to 4T1-SCA1+–derived tumors (Figure 2A). Strikingly, we observed metastases in the heart (Supplemental Figure 8D), which we never observed with parental 4T1 cells. MR13 cells retained a large fraction of the SCA1+ population in vitro, even when cultured in the absence of MTX (Supplemental Figure 8B) and upon in vivo expansion (Figure 6J). Such stability of the SCA1+ population contrasted with SCA1+ cells enriched by MACS in vitro or induced by the TME in vivo, both of which reverted to a SCA1– phenotype shortly after in vitro culture (Supplemental Figure 2, D–F). This observation suggests that the MR13 line was capable of self-sustaining its own SCA1+ population. Taken together, these data show that chemotherapy-selected MR13 cells are stably enriched for the SCA1+ population both in vitro and in vivo and are highly metastatic in vivo.

IL-6/IL-6R/JAK autocrine signaling maintains SCA1 positivity and metastatic capacity in MR13 cells. To better understand the chemotherapy-induced alterations in MR13 cells, we performed transcriptomic analyses comparing MR13 and parental 4T1 cells. Pathway enrichment analysis showed an upregulation of the IL-6/JAK/STAT3 signature in MR13 cells (Figure 7A). Importantly, MR13 gene expression significantly positively correlated with the SCA1-positive signature and negatively correlated with the SCA1-negative signature (Figure 7B). This suggested that chemotherapy enriches for the inherent 4T1-SCA1+ population, rather than converting SCA1– cells into SCA1+ cells. In addition, Osmr, Il6, Il6ra, and Il6st genes were all overexpressed in MR13 cells compared with parental 4T1 cells, while Osm expression was not altered (Figure 7C). To functionally validate the IL-6/JAK signaling pathway in SCA1+ population maintenance, we treated MR13 cells with ruxolitinib in vitro for 48 hours. The treatment significantly decreased the fraction of SCA1+ cells (Figure 7D), while cell proliferation (Figure 7E and Supplemental Figure 8E) and viability (Supplemental Figure 8, F and G) were not affected. Interestingly, in vitro pretreatment of MR13 cells with ruxolitinib for 3 days completely abolished their lung metastatic capacity upon tail-vein injection (Figure 7, F–H). Together, these data indicate that MR13 cells sustain the metastatic SCA1+ population through cell-autonomous activation of the IL-6/JAK signaling pathway.

IL-6/JAK pathway promotes SCA1+ persistence and metastatic capacity in chemotherapy-resistant MR13 tumor cells. (A and B) Gene expression analysis of parental 4T1 and chemotherapy-resistant MR13 cells. Heatmap represents the signature score of the hallmark pathways analysis. Results from 3 biological replicates are shown (A). GSEA results show that MR13 cells are positively enriched for the SCA1-positive signature and negatively enriched for the SCA1-negative signature (B). (C) Volcano plot showing the differential expression of Osm, Osmr, Il6st, Il6, and Il6ra mRNA in MR13 versus 4T1 tumor cells. (D) Fraction of SCA1+ population in MR13 tumor cells treated for 72 hours with the JAK inhibitor ruxolitinib (5 μM) relative to vehicle control (DMSO) treatment. n = 9/group. (E) Percentage of EdU-positive in SCA1+ population of MR13 cells after treatment with JAK inhibitor ruxolitinib or DMSO control for 72 hours. n = 6/group. (F) Schematic of the experimental design for testing the effect of ruxolitinib on MR13 lung metastatic capacity shown in G–H. MR13 tumor cells were treated with ruxolitinib (5 μM) or DMSO in vitro for 72 hours and then injected into the mouse tail vein. Lungs were examined for metastasis 10 days after tumor cell injection. (G and H) Quantification of lung metastases in mice injected i.v. with MR13 treated in vitro with ruxolitinib (5 μM) or DMSO (G). Representative images of H&E staining are shown (H). n = 8/group. Scale bar: 1 mm. Data are represented as means ± SEM and are representative of 3 independent experiments for D, E, and G. P values were calculated using 1-way ANOVA with Dunnett’s multiple-comparison test (D and E) or unpaired 2-tailed Student’s t test (G). ***P < 0.001.

Tu-Gr1+CD11b+–invoked tumor cell signature predicts shorter OS and RFS in breast cancer patients. To evaluate the clinical relevance of the crosstalk between Tu-Gr1+CD11b+ and tumor cells, we tested to determine whether the Tu-Gr1+CD11b+–induced 4T1 signature (top 50 DEGs) could predict cancer progression in other murine models and patients. First, we examined the signature expression in the PT and lung metastasis from different mouse models using the Ross et al. data set (39). The signature was significantly elevated in lung metastasis in 4T1, 6DT1, and Mvt1 breast cancer models, while in the Met1 model, it was unaltered (Figure 8A). Next, we examined the NKI295 cohort (58) consisting of expression data of invasive breast carcinoma from 295 women curated with site-specific metastasis data. Higher expression of the human orthologue of the murine signature predicted metastatic relapse to the lung (Figure 8B). Further, we interrogated the METABRIC data set comprising expression data from over 2,000 breast cancer patients (59) with the Tu-Gr1+CD11b+–induced signature. Thirty-two human orthologue genes (Supplemental Table 3) of the murine 50 gene signature were present in the METABRIC data set. Patients with higher expression levels of the orthologue signature had shorter OS (P = 0.0056) and RFS (P = 0.032) for all patients. Median OS and RFS values of patients with high versus low signature were 142.4 versus 169.6 and 196.4 versus 296.7 months, respectively (Figure 8, C and D). When interrogating breast cancer subtypes, higher signature expression was associated with a significantly reduced OS in ER+, PR+, and HER2– breast cancer patients (Supplemental Figure 9, A–F). Notably, OSM expression levels positively correlated with the signature (Supplemental Figure 9G), suggesting that OSM does contribute to altered signature expression in breast cancer patients. Of the 32 genes, 5 (MX1, IRF7, OAS1, CMPK2, ISG15) were discriminant for a shorter OS and RFS (Figure 8, E and F), and when combined they showed enhanced predictive power for OS (P = 0.00055) and RFS (P = 0.00069) (Figure 8, G and H).

Tu-Gr1+CD11b+–induced tumor cell signature predicts worse outcome in breast cancer patients. (A) Box plot showing the expression of Tu-Gr1+CD11b+–induced 4T1 cell signature in PT and lung metastasis in the 4T1, 6DT1, Mvt1, and Met1 murine metastatic breast cancer models, extracted from the Ross data set (39). The box extended from 25th to 75th percentile, with the median indicated as a line within the box. The whiskers shown are 1.5 times interquartile ranges. P values were calculated using unpaired 2-tailed Student’s t test. (B) Expression of human orthologue of Tu-Gr1+CD11b+–induced signature in breast cancer patients with (yes) or without (no) metastatic relapse to the lung in the NKI295 cohort (58). (C and D) Kaplan-Meier curves showing OS (C) or RFS (D) for breast cancer patients according to high or low expression of an orthologue 32 gene signature, based on the Tu-Gr1+CD11b+–induced 4T1 cell signature in the METABRIC data sets (59). (E and F) Forest plots showing Cox’s proportional hazard regression (HR) for OS (E) and RFS (F) of the individual 32 orthologues of the Tu-Gr1+CD11b+–induced signature, based on gene expression in tumor samples of the METABRIC data set. (G and H) Kaplan-Meier curves showing OS (G) and RFS (H) according to the minimal 5-gene orthologue gene signature expression in the METABRIC data sets. P values were calculated using unpaired 2-tailed Student’s t test (A and B) or log-rank test (C, D, G and H). *P < 0.05; **P < 0.01.