Fibrocytes represent a novel MDSC subset circulating in patients with metastatic cancer

Abstract

Fibrocytes are hematopoietic stem cell-derived fibroblast precursors that are implicated in chronic inflammation, fibrosis, and wound healing. Myeloid-derived suppressor cells (MDSCs) expand in cancer-bearing hosts and contribute to tumor immune evasion. They are typically described as CD11b⁺HLA-DR⁻ in humans. We report abnormal expansions of CD11b⁺HLA-DR⁺ myeloid cells in peripheral blood mononuclear fractions of subjects with metastatic pediatric sarcomas. Like classical fibrocytes, they display cell surface α smooth muscle actin, collagen I/V, and mediate angiogenesis. However, classical fibrocytes serve as antigen presenters and augment immune reactivity, whereas fibrocytes from cancer subjects suppressed anti-CD3-mediated T-cell proliferation, primarily via indoleamine oxidase (IDO). The degree of fibrocyte expansion observed in individual subjects directly correlated with the frequency of circulating GATA3⁺CD4⁺ cells (R = 0.80) and monocytes from healthy donors cultured with IL-4 differentiated into fibrocytes with the same phenotypic profile and immunosuppressive properties as those observed in patients with cancer. We thus describe a novel subset of cancer-induced myeloid-derived suppressor cells, which bear the phenotypic and functional hallmarks of fibrocytes but mediate immune suppression. These cells are likely expanded in response to Th2 immune deviation and may contribute to tumor progression via both immune evasion and angiogenesis.

Figures

Figure 1

Expansion of a novel myeloid cell subset in cancer patients. (A) In cancer patients, gating on CD14– cells in monocyte-rich fractions elutriated from apheresis products showed expansion of CD11chiCD123– cells (blue box) that are not present in healthy donors. The green box denotes myeloid dendritic cells and the orange box denotes plasmacytoid dendritic cells. The CD14+ fraction is comprised entirely of monocytes (red box) in patients and healthy donors. (B) Percent of CD11c+CD123– cells in the patient cohort (n = 53) vs healthy donors (n = 10). The red dotted line represents the mean, and the error bar represents the SEM. (C) Expression profiles of gated subsets (color coded) are shown. CD11chiCD123–CD14– cells express common MDSC markers, including CD11b, CD15, CD66b, and CD124, but they also express HLA-DR and CD127. Quadrants set based on the background fluorescence using fluorescence minus one (FMO) controls. Representative flow cytometry plots from a representative subject are shown. (D) Hematoxylin and eosin–stained cytospins of electronically sorted subsets (original magnification ×200) from a representative subject. CD11chiCD123–CD14– cells resemble immature monocytes, whereas other subsets show the expected morphology.

Figure 2

Expanded CD11chiCD123–CD14– cells are fibrocytes that mediate angiogenesis. (A) Using the same gating strategy as shown in Figure 1A, CD11chiCD123–CD14– cells from a representative subject sample were analyzed for cell surface phenotype. The shaded areas represent background fluorescence on the designated population as indicated by FMO controls. This is representative of more than 10 experiments. (B) Comparison of tube formation by both electronically sorted CD11chiCD123–CD14– cells (designated fibrocytes) and CD14+ monocytes from 1 representative cancer subject, which were then plated in an angiogenesis assay. HUVEC cultures in FGF-2 were used as a positive control. This is representative of 5 experiments. (C) Mean ± SEM joint counts formed by HUVEC cocultured with FGF-2, monocytes vs fibrocytes in 5 separate tube formation assays using fibrocytes, and monocytes from 5 separate subjects.

Figure 3

Fibrocytes mediate immune suppression via IDO. (A) Fibrocytes do not co-stimulate T cells treated with submitogenic concentrations of anti-CD3. Elutriated lymphocytes (5 × 104 cells/well) and designated electronically sorted cell subsets (2.5 × 104 cells/well) were added to a 96-well plate precoated with the designated concentration of anti-CD3, and thymidine incorporation was measured on day 3. This is representative of 5 experiments using cells from 5 separate subjects. (B) Suppression of anti–CD3-mediated T-cell proliferation by fibrocytes. T cells were plated as in (A), but the designated electronically sorted cell subsets (purity >90%) were added (2.5 × 104 cells/well) at a ratio of 2 lymphocytes:1 suppressor, and thymidine incorporation was measured on day 3. This is representative of 6 experiments using cells from 6 separate subjects. (C) Semiquantitative real-time PCR was used to measure IDO, Arginase I, and INOS2. Relative amounts compared with normal peripheral blood mononuclear cells are shown. The mean ± SEM was derived from 6 experiments. (D) Fibrocyte-mediated suppression can be reversed by the addition of an IDO inhibitor, 1MT. Cultures were set up as described in (A) and, where designated, 1MT was added at a concentration of 20 μM at the beginning on the co-culture. Results are representative of 5 experiments using fibrocytes from 5 separate subjects. All experiments in Figure 3 used electronically sorted fibrocytes with purity >90%.

Figure 4

Frequency of circulating fibrocytes in cancer subjects correlates with the expansion of circulating GATA3+CD4+ T cells. (A) Correlation between % fibrocytes in the monocyte-rich apheresis fraction and GATA3+CD4+ T cells in the lymphocyte-rich apheresis fraction acquired simultaneously. Each closed dot represents a value measured in an individual subject from whom adequate simultaneous monocyte-rich and lymphocyte-rich samples were available (n = 22 subjects). Gray dots represent results from normal donors (P < .05). (B) Representative flow cytometry dot plot illustrating high levels of GATA3+CD4+ T cells from 1 subject with simultaneous fibrocyte expansion. CD4+ cells were analyzed immediately after thawing and permeabilization without stimulation.

Figure 5

IL-4 induces monocytes to differentiate into CD14– fibrocytes that are readily distinguished from CD14+ macrophages in the same culture. (A) Culture of monocyte-rich apheresis fractions from healthy donors with rhIL-4 induces differentiation into an adherent population that contains cells with a typical spindlelike appearance, consistent with a fibrocyte lineage (original magnification ×20). (B) Cell surface phenotype of IL-4–differentiated adherent cells identifies 2 subsets based on CD14 expression, which further shows differential expression of collagen and TSLPR. FMO controls on gated CD14+ vs CD14– populations are shown by shaded gray histograms. This is representative of more than 5 experiments from 5 separate healthy donors. (C) Fibrocyte (defined as CD14–CD123–CD11c+) numbers in monocyte cultures incubated for 9 days with rhIL-4 alone (10 ng/mL) vs rhIL-4 plus TSLP (10 ng/mL). This is representative of 4 separate experiments. (D) CD14+ macrophages vs CD14– fibrocytes generated using rhIL-4 as described in (C) were analyzed by reverse transcriptase PCR for IDO, TDO, ARG1, and INOS2. The mean ± SEM of relative increases normalized to PBMC are shown from 5 separate experiments. (E) RhIL4-induced fibrocytes suppress autologous T-cell proliferation to various concentrations of anti-CD3. Comparisons between 2 groups (with T-cell-alone group) were done using a 2-way analysis of variance for multiple comparisons, with significance determined as P < .05. This is representative of 3 experiments. (F) Using RNA expression profiles of cells generated and electronically sorted as described in (C), gene set enrichment analysis of CD14+ macrophages vs CD14– fibrocytes for the Lindstedt DC maturation C gene set is shown. The most underexpressed of 49 genes in CD14– fibrocytes vs the CD14+ macrophages is listed below the figure. This is representative of 3 separate expression profile-based analyses.