Multiparameter Analysis of Human Bone Marrow Stromal Cells Identifies Distinct Immunomodulatory and Differentiation-Competent Subtypes

Abstract

Bone marrow stromal cells (BMSCs, also called bone-marrow-derived mesenchymal stromal cells) provide hematopoietic support and immunoregulation and contain a stem cell fraction capable of skeletogenic differentiation. We used immortalized human BMSC clonal lines for multi-level analysis of functional markers for BMSC subsets. All clones expressed typical BMSC cell-surface antigens; however, clones with trilineage differentiation capacity exhibited enhanced vascular interaction gene sets, whereas non-differentiating clones were uniquely CD317 positive with significantly enriched immunomodulatory transcriptional networks and high IL-7 production. IL-7 lineage tracing and CD317 immunolocalization confirmed the existence of a rare non-differentiating BMSC subtype, distinct from Cxcl12-DsRed(+) perivascular stromal cells in vivo. Colony-forming CD317(+) IL-7(hi) cells, identified at ∼ 1%-3% frequency in heterogeneous human BMSC fractions, were found to have the same biomolecular profile as non-differentiating BMSC clones using Raman spectroscopy. Distinct functional identities can be assigned to BMSC subpopulations, which are likely to have specific roles in immune control, lymphopoiesis, and bone homeostasis.

Figure 1

Generation and Analysis of hTERT-BMSC Clones (A) Analysis of hTERT-BMSC clones (Y101, Y102, Y201, and Y202) by flow cytometry for surface markers typically assigned to BMSCs. Representative histograms are shown. (B) Clonal growth of hTERT-BMSCs (crystal violet stain). Scale bars, 200 μm. (C–E) Assessment of the OAC potential of hTERT-BMSC clones; representative images and quantification of cells differentiated with adipogenic (C), osteogenic (D), and chondrogenic supplements (E). Scale bar represents 200 μm (C and D) or 2 mm (E). Data represent average quantified values ± SD for two or three independent experiments performed using two to six replicates. Differentiation was considered to be statistically significant after the data from n > 3 experiments were analyzed by one-way ANOVA with Holm Sidak’s multiple comparisons if ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, or ^∗∗∗∗p < 0.0001 between both day 0 and the time-matched basal control. Additionally, one-way ANOVA with Holm Sidak’s comparisons was performed between the different cell lines at day 21, and significant differences compared with Y101 (∼) or Y102 ($) and Y201 (#) are indicated. See also Figure S1.

Figure 2

Expression Profiling Identifies BMSC Clones with Distinct Immunoregulatory Features (A) Hierarchical clustering heatmap of global gene expression profiles from each hTERT-BMSC clonal line and the parental BMSCs. (B) Pathway analysis showing most significantly enriched pathways. (C) Self-organizing heatmaps of the four hTERT-BMSC clones and five primary BMSC samples, including the parent population (FH181). Arrows indicate presence (closed) and absence (open) of overrepresented (red) and underrepresented (blue) metagene spots. (D) Heatmap showing expression of key genes involved in inflammatory-induced responses. (E) Immunocytochemical detection of IL-7 expression (red) by the four hTERT-BMSC clones (after 24 hr in culture) (DAPI nuclear counterstain in blue). Scale bars, 20 μm. (F) IL-7 expression by ELISA in the four hTERT-BMSC clones, primary BMSCs (K98, K102, and K110), and human dermal fibroblasts (HDF). Data represent mean values ± SD from three individual experiments performed in duplicate. See also Figure S2.

Figure 3

Lineage Tracing of IL-7-Expressing Cells Using IL-7cre Rosa26-EYFP Mice (A–F) Sporadic EYFP expression throughout bone marrow, associated occasionally with cells lining the vasculature (A, arrows) and prominent on endosteal surfaces (B). Infrequent osteocytes were EYFP positive (D–F). (G–J) Immunostaining of perilipin-positive adipocytes (red) in bone marrow of IL-7cre Rosa26-EYFP mice. (K and L) Absence of EYFP expression in white adipose tissue extracted from IL-7cre Rosa26-EYFP mice. (M–O) EYFP expression in chondrocytes of articular cartilage. Scale bars, 50 μm (unless otherwise stated). See also Figure S3.

Figure 4

Identification of CD317⁺ Bone Marrow Cells (A) Flow cytometric analysis of hTERT-BMSC; CD317 expression selectively discriminates Y102/Y202 from Y101/Y201. (B) Heatmap showing expression of key genes involved in anti-viral responses. (C) CFU-F assay and morphology of CD317⁻ and CD317⁺ cells sorted from heterogeneous BMSCs. Scale bars, 200 μm. Histogram represents mean ± SEM (n = 4 independent experiments performed with five replicates). (D) Analysis of cell perimeter (left) and cell area (right) of CD317⁻ and CD317⁺ cells sorted from heterogeneous BMSCs. Mean ± SEM, n = 4 independent experiments in which, on average, 113 cells were examined. ^∗p < 0.05 by unpaired non-parametric t test. (E) IL-7 mRNA (left) and protein expression (right, by ELISA) in CD317⁻ and CD317⁺ cells. Data represent mean IL-7 levels ± SD from FACS-sorted cells in triplicate from four or two donors, respectively. ^∗p < 0.05 by paired t test. (F) Immunohistochemistry of mouse femur bone marrow sections stained with antibodies to CD317 (red), CD31 (green) with Cxcl12-DsRed (white) and nuclear (DAPI, blue) staining. Representative merged and single-panel images shown. (G) Immunolocalization of LepR/CD295 (green) and CD317 (red) in mouse bone marrow showing LepR⁺CD317⁻ (upper panel) and occasional dual LepR⁺CD317⁺ cells (lower panel). Blue indicates nuclear DAPI stain. (H and I) Analysis of hTERT-BMSC clones and CD317⁺ cells by Raman spectroscopy, Raman shifts (H) and peak ratios (I) are shown. Raman peak assignments are provided in Table S2. See also Figure S4.