SFRP1 and SFRP2 Jointly Support Normal and Malignant Hematopoietic Precursor Cells in an in Vitro Hematopoietic Niche

The pleiotropic proteins, Secreted frizzled-related proteins (SFRPs), were first identified as WNT antagonists and recently recognized as extracellular WNT carriers. Extracellular SFRP2 was also reported to maintain stemness of mesenchymal stromal cells (MSCs) by suppressing BMP signals. We have uncovered that the MSCs that lack MED1 subunit of the Mediator transcriptional coregulatory complex do not sufficiently support hematopoietic stem/progenitor cells (HSPCs). As Sfrp1 and Sfrp2 expressions were among those that were substantially attenuated in Med1^-/- mouse embryonic fibroblasts (MEFs), and in view also that the Sfrp1 or Sfrp2 knockout mice reportedly showed attenuated HSPCs, we herewith studied the role of SFRP1 and SFRP2 in the hematopoietic niche. SFRP1/2 were variously expressed in several MSCs including OP-9 and MS-5 mouse MSCs and profoundly reduced in Med1^-/- MEFs, whereas the bone marrow (BM) hematopoietic cells did not express SFRP1 but did SFRP2 weakly. Although in malignant situations nuclear SFRPs were reported to bind and suppress β-catenin, immunofluorescent studies of the OP-9 cells showed SFRP1/2 were localized to the cell membrane. To know the role of SFRP1/2 in the hematopoietic microenvironment, a simplified in vitro hematopoietic niche model was made use of that was composed of a coculture of mitomycin C (MMC)-treated MS-5 or OP-9 MSCs and mouse BM hematopoietic cells. When the blocking antibody for either SFRP1 or SFRP2 was added, the number and the DNA synthesis of cocultured hematopoietic cells were attenuated, and these were further reduced when both antibodies were added. The long-term culture-initiating cells (LTC-ICs) were slightly attenuated by either of the antibodies, while in the presence of both antibodies the LTC-ICs were profoundly (10-fold) decreased, in the methylcellulose colony-forming assays. Therefore, the MSC-derived SFRP1/2 appear to function jointly as niche factors for maintenance of normal HSPCs.

We next asked if SFRP1/2 also maintain leukemic stem cells and tested the MSC-dependent human myeloblastoma cell line, MB-1, originally established from a patient with myeloid crisis chronic myeloid leukemia. When either of blocking antibodies for SFRP1or SFRP2 was added to the coculture of MB-1 cells and MMC-treated OP-9 cells, the mitogenicity and the number of MB-1 cells decreased slightly, and these were further (1/3 and 2/3 of the controls) diminished in the presence of both antibodies. The number of cobblestone areas, the in vitro characteristics of leukemic stem cells, were also attenuated with either antibody and further reduced (70% of the control) in the presence of both antibodies. These phenomena were confirmed by using another pair of different antibodies against SFRP1 and SFRP2. Therefore, SFRP1/2 appear to be niche factors also for leukemic stem cells.

We then analyzed the source of SFRP1/2 in the MB-1/OP-9 coculture by human- and mouse-specific quantitative RT-PCR. The expressions of Sfrp1 and Sfrp2 in OP-9 cells were oscillated; they decreased 1 day after initiating coculture and reverted after several days. The expressions of SFRP1 and SFRP2 in MB-1 cells were reciprocal; they were ectopically highly expressed 1 day after the coculture and then decreased. Thus in a malignant situation these expressions in niche and blood cells were oscillated and mutually reciprocal.

We next asked how SFRP1/2 support MB-1 cells, and analyzed the expression of the WNT and BMP target genes in MB-1 cells. Indeed, the expressions of both WNT target genes, AXIN2 and CCND1, and BMP target genes, ID1 and ID2, were all unchanged after the antibodies additions. Hence, SFRP1/2 may support MB-1 cells through acting on the MSCs. As SFRP2 reportedly maintains stemness of MSCs by inhibiting BMP/WNT signals, SFRP1/2 might act on OP-9 cells to maintain the ability to support hematopoietic cells. To address if SFRP1/2 have biological effects on OP-9 cells, we added anti SFRP1/2 blocking antibodies to MMC-untreated OP-9 cells, and found that the cell numbers were 20% decreased in the presence of either antibody, and further reduced (40%) with both antibodies, in agreement with the known role of BMP/WNT in growth inhibition of mesenchymal cells.

These findings collectively suggest that SFRP1 and SFRP2 are bona fide niche factors that are jointly necessary for normal and malignant HSPCs support, at least partly, by suppressing BMP/WNT signals of MSCs.

No relevant conflicts of interest to declare.