Syndecan-2 expression enriches for hematopoietic stem cells and regulates stem cell repopulating capacity Tracking no : BLD-2020-010447 R 2

Abstract: The discovery of novel hematopoietic stem cell (HSC) surface markers can enhance understanding of HSC identity and function. We have discovered a population of primitive bone marrow (BM) HSCs distinguished by their expression of the heparan sulfate proteoglycan, Syndecan-2, which serves as both a marker and regulator of HSC function. Syndecan-2 expression was increased 10-fold in CD150+CD48-CD34-c-Kit+Sca1+Lineagecells (long-term – HSCs, LT-HSCs) compared to differentiated hematopoietic cells. Isolation of BM cells based solely on Syndecan-2 surface expression produced a 24-fold enrichment for LT-HSCs, 6fold enrichment for alpha-catulin+c-kit+ HSCs, and yielded HSCs with superior in vivo repopulating capacity compared to CD150+ cells. Competitive repopulation assays revealed the HSC frequency to be 17fold higher in Syndecan-2+CD34-KSL cells compared to Syndecan-2-CD34-KSL cells and indistinguishable from CD150+CD34-KSL cells. Syndecan-2 expression also identified nearly all repopulating HSCs within the CD150+CD34-KSL population. Mechanistically, Syndecan-2 regulates HSC repopulating capacity through control of expression of Cdkn1c (p57) and HSC quiescence. Loss of Syndecan-2 expression caused increased HSC cell cycle entry, downregulation of Cdkn1c and loss of HSC long-term repopulating capacity. Syndecan-2 is a novel marker of HSCs which regulates HSC repopulating capacity via control of HSC quiescence.


Introduction
Hematopoietic stem cells (HSCs) comprise less than 0.01% of the total nucleated cells in the adult bone marrow (BM). 1 Long-term -HSCs (LT-HSCs) possess self-renewal capacity and unrestricted hematopoietic differentiation potential. 2 Characterization of the immunophenotype of HSCs is important both to understand the fundamental biology of HSCs and to facilitate the development of HSC-based regenerative therapies.
Murine HSCs display unique surface protein expression patterns, which enabled the purification of HSCs using antibody staining and fluorescence activated cell sorting (FACS).
Here, we demonstrate that Syndecan-2 expression enriches for HSCs with enhanced self-renewal capacity and Syndecan-2 regulates HSC repopulating capacity via control of HSC quiescence.

Methods
For detailed methods, see Supplemental Methods.

Mice
All mouse procedures were performed using 8-12-week-old mixed gender mice in accordance with protocols approved by UCLA and Cedars Sinai Medical Center (PI, John Chute).

Flow cytometry
BM cells were isolated from murine long bones, lineage depleted, and stained using antibodies and/or 7-AAD/Annexin-V for cell death analysis. Stained cells were analyzed using a BD FACS Canto II or sorted using a BD FACS Aria.

Competitive transplants
HSCs were resuspended in 10% FBS/PBS supplemented with 2x10 5 competitor BM cells. Cells were transplanted into lethally irradiated (900 cGy) mice via tail vein injection. PB was analyzed every four weeks for donor chimerism using flow cytometry. For secondary transplants, BM was harvested from primary transplanted mice 16 weeks post-transplant and 3x10 6 BM cells were Downloaded from http://ashpublications.org/blood/article-pdf/doi/10.1182/blood.2020010447/1835121/blood.2020010447.pdf by guest on 28 November 2021 transplanted along with 2x10 5 competitor BM cells. PB was analyzed every four weeks for donor chimerism.

Homing assay
BM HSPCs were resuspended in 10% FBS/PBS and transplanted intravenously into lethally irradiated mice. At +24 hours, BM cells were harvested and analyzed for donor cells by flow cytometry.

HSC cultures, lentiviral transduction and colony assays
BM HSCs were sorted using FACS, plated in TSF media and cultured in humidified, 5% CO 2 incubator. Cultured cells were analyzed after 3 or 7 days of culture. Sorted 34 -KSL cells were pre-stimulated for 48 hours in StemSpan media supplemented with mouse SCF, IL-3, and IL-6.
Cells were spin-occulated with shControl, shSdc2, shCdkn1c, MSCV Control or MSCV-Sdc2 viral supernatant and incubated for 48 hours. For colony forming cell assays, BM cells were plated in Methocult and incubated for 10 days prior to colony counts.

Gene expression and RNA sequence analysis
RNA was isolated using the RNeasy Micro Kit. For qRT-PCR analysis, RNA was reverse transcribed into cDNA using the High-Capacity cDNA Reverse Transcription Kit and gene expression was analyzed using an Applied Biosystems QuantStudio 6 PCR Machine. For RNA sequencing, RNA was sequenced using a HiSeq3000 Sequencing System and IPA was performed. RNA sequencing data are available at GEO under accession number GSE151733.

Cell cycle analysis
BM cells were stained for surface markers using antibodies and then fixed, permeabilized and stained for Ki-67 and 7AAD for flow cytometric analysis.

Confocal microscopy
Cells were plated on fibronectin-coated slides, fixed, permeabilized, and stained using antibodies to p57, p16, p21 and p27. Imaging was performed using the Leica Stellaris system and analyzed with ImageJ.
We next evaluated whether Syndecan-2 as a sole marker could enrich for functional HSCs and HSPCs. Syndecan-2 + Lincells produced significantly more colony forming cells (CFCs) compared to CD150 + Lincells ( Figure 1J). 12,15 Furthermore, mice transplanted with Syndecan-2 + Lincells displayed increased total donor cell chimerism and increased donor B cell and myeloid cell chimerism in the BM at 16 weeks post-transplant compared to mice transplanted with CD150 + Lin -BM cells or Syndecan-2 -Lin -BM cells ( Figure 1K, L). Therefore, Syndecan-2 surface expression alone enriches for HSCs with multilineage repopulating capacity.
Syndecan-2 expression was minimally detected on CMP, MEP, GMP, and CLP cells and terminally differentiated hematopoietic cells (supplemental Figure 1H-I). Syndecan-2 + BM cells contained small percentages of hematopoietic progenitor cells and terminally differentiated cells compared to Syndecan-2 -BM cells (supplemental Figure 1J-L).
Mice transplanted with Sdc2 High HSCs demonstrated significantly higher total and multilineage donor chimerism following competitive transplantation compared to Sdc2 Med or Sdc2 -HSCs (supplemental Figure 2E). Secondary competitive repopulation assays to assess long-term HSC repopulating capacity revealed increased donor hematopoietic cell engraftment in the PB ( Figure 2G) and BM (supplemental Figure 2F) of recipients of Sdc2 + HSCs, without lineage skewing, compared to mice transplanted with Sdc2 -HSCs (supplemental Figure 2F, G). Secondary recipient mice also displayed increased donor BM CD45.1 + CD150 + CD48 -KSL HSCs compared to recipients of Sdc2 -HSCs ( Figure 2H).
While quiescent HSCs are protected from exhaustion, 33 HSCs must maintain the ability to enter the cell cycle to support hematopoietic demands. We isolated 34 -KSL HSCs, Sdc2 -HSCs and Sdc2 + HSCs and plated these populations in complete IMDM supplemented with thrombopoietin, stem cell factor and Flt3 ligand (TSF media) to promote cell cycling ( Figure 4A).
Total cell expansion was observed in all populations after 7 days, but Sdc2 + HSCs produced significantly more 34 -KSL HSCs in culture compared to Sdc2 -HSCs, which were nearly depleted of phenotypic HSCs ( Figure 4B-E).
In order to determine if Syndecan-2 surface expression is interchangeable on HSCs, Sdc2 + HSCs or Sdc2 -HSCs were cultured with TSF media. By day +3 of culture, more than 80% of the 34 -KSL cells in cultures initiated with Sdc2 + HSCs became Sdc2 -, whereas more than 30% of the 34 -KSL cells in cultures initiated with Sdc2 -HSCs became Sdc2 + ( Figure 3B). Cultures initiated with Sdc2 + HSCs produced significantly more Sdc2 + and Sdc2 -HSCs ( Figure 4H, I), consistent with increased total cell expansion in the cultures initiated with Sdc2 + HSCs ( Figure 4B, E). We also sorted Sdc2 + 34 -KSL cells and Sdc2 -34 -KSL cells from day +7 cultures and re-plated each population in TSF media and quantified the percentages of Sdc2 + 34 -KSL cells in each group after 14 total days of culture (supplemental Figure 3C). Both Sdc2 + 34 -KSL cells and Sdc2 -34 -KSL cells derived from originating Sdc2 -HSCs continued to produce a higher percentage of Sdc2 + 34 -KSL cells in culture compared to the progeny of Sdc2 + HSCs (supplemental Figure 3D). These results suggest that a subset of Sdc2 -HSCs can convert to Sdc2 + 34 -KSL cells in response to cytokine stimulation.
In cultures initiated with Sdc2 -HSCs, > 70% of 34 -KSL cells were in G 0 at day 7,  Figure 4N). Secondary competitive transplantation assays utilizing BM cells isolated at 16 weeks post-transplant from primary recipient mice confirmed that long-term engraftment potential was increased in the cultured progeny of Sdc2 + HSC cultures compared to the progeny of Sdc2 -HSCs or 34 -KSL HSCs (supplemental Figure   3G).

Syndecan-2 regulates HSC quiescence through control of Cdkn1c
Syndecans mediate context-specific effects on cell proliferation, whereas we have observed that Syndecan-2 expression is associated with HSC quiescence. 34,35 We next measured the cell cycle status of BM 34 -KSL HSCs following treatment with lentiviral particles containing shControl or shSdc2. HSCs treated with shSdc2 displayed decreased percentages of cells in G 0 compared to shControl-treated HSCs, indicating exit from quiescence in response to Syndecan-2 silencing ( Figure 6A-C). Since the Cip/Kip family of cyclin dependent kinase (Cdk) inhibitors regulates HSC cell cycling, 36,37 we measured the expression of Cdkn1c, which encodes the cell cycle inhibitor, p57, in shSdc2-treated HSCs and shControl-treated HSCs. At baseline, Sdc2 + HSCs expressed nearly 10-fold increased levels of Cdkn1c compared to Sdc2 -HSCs ( Figure   6D). Silencing of Sdc2 suppressed Cdkn1c expression in BM 34 -KSL cells ( Figure 6D). Sdc2 knockdown in BM 34 -KSL cells also caused a moderate increase in the expression of Cdkn1a and p16 compared to control HSCs, but did not impact Cdkn1b expression ( Figure 6E).
Microscopic protein expression analyses revealed decreased p57 and increased p16 protein expression in shSdc2-treated HSCs compared to shControl-treated HSCs (supplemental Figure   5E -H).
Given the role of Sdc2 in regulating Cdkn1c expression, we next knocked down both Sdc2 and Cdkn1c in Sdc2 + and Sdc2 -HSCs. Silencing of Cdkn1c, Sdc2, or both in Sdc2 -HSCs did not further decrease Cdkn1c levels, suggesting that Sdc2 -HSCs minimally express Cdkn1c ( Figure 6D). Silencing of Cdkn1c, Sdc2, or both in Sdc2 -HSCs also did not impact the expression of Cdkn1a, Cdkn1b or p16 ( Figure 6E). In Sdc2 + HSCs, silencing of Cdkn1c and Sdc2 did not further decrease Cdkn1c expression or increase Cdkn1a expression compared to knockdown of Cdkn1c alone ( Figure 6D, E). Double knockdown of Sdc2 and Cdkn1c in Sdc2 + HSCs decreased CFU-GEMM colony production compared to shControl-treated HSCs, but this reduction was comparable to knockdown of either Sdc2 or Cdkn1c alone ( Figure 6F). These results suggest that Sdc2 mediates molecular and hematopoietic effects through Cdkn1c.
Silencing of Sdc2 expression decreased HSC quiescence in control Sdc2 + HSC cultures, but Sdc2 silencing had no significant effect on HSC quiescence in the presence of SIS3 treatment ( Figure 6G). We next assessed the cell cycle status of CD150 + CD48 -34 -KSL Sdc2 + HSCs upon knockdown of Sdc2, Cdkn1c or both. Silencing of Sdc2 alone, Cdkn1c alone or both Sdc2 and Cdkn1c caused comparable reductions in the percentages of HSCs in G 0 and G 1 , with concordant increases in percentages of HSCs in G 2 /S/M phase ( Figure 6H). These data suggest that Syndecan-2 regulates HSC cell cycle status through Cdkn1c.
To understand why Sdc2 + HSCs are susceptible to inhibition of TGFβ-1-dependent SMAD3 phosphorylation while Sdc2 -HSCs are not, we quantified phospho-SMAD3 expression in Sdc2 + and Sdc2 -HSCs. Sdc2 + HSCs exhibited elevated SMAD3 activation at baseline compared to Sdc2 -HSCs, but Sdc2 + and Sdc2 -HSCs were equally sensitive to TGFβ stimulation ( Figure 6I), suggesting that Syndecan-2 expression marks HSCs with elevated basal SMAD3 signaling. We next treated HSCs with SIS3 to understand whether Sdc2 + or Sdc2 -HSCs are dependent on SMAD3 signaling for HSC maintenance. SIS3 treatment of Sdc2 + HSCs significantly decreased CFU-GEMM generation compared to control treated Sdc2 + HSCs ( Figure 6J) and significantly decreased percentages of 34 -KSL cells in culture ( Figure 6K). Since TGFβ/SMAD signaling regulates Cdkn1c expression 41 , these data are consistent with our genetic studies suggesting that Syndecan-2 regulates HSC quiescence and function through control of Cdkn1c.

Discussion
Syndecan-2 is a member of the family of HSPGs, which via their role as extracellular binding partners for secreted proteins, 18,19 regulate morphogen gradients during development. 26,[49][50][51][52] Syndecans also modulate the differentiation, proliferation and regeneration of adult neural stem cells. 24,25 In hematopoiesis, HSPGs produced by stromal cells contribute to cytokine-mediated regulation of HSPC growth [53][54][55] and HSPC retention in the BM. 26 We discovered that BM HSCs are highly enriched for Syndecan-2 expression and isolation of BM cells based on Syndecan-2 surface expression yields a 24-fold enrichment for CD150 + 48 -KSL LT-HSCs and approximately 6-fold enrichment for α-catulin + c-kit + HSCs compared to Syndecan-2 negative BM cells. LT-HSCs. 29 Here, via limiting dilution analysis of competitive repopulation assays, we determined that the HSC frequency within Syndecan-2 + 34 -KSL BM was indistinguishable from CD150 + 34 -KSL BM cells and more than 10fold increased compared to Syndecan-2 -34 -KSL BM cells. Therefore, Syndecan-2 represents a novel cell surface marker that can be utilized to enrich and purify HSC populations. Prior studies have also shown that CD150 + HSC populations are heterogeneous with regard to their myeloid-or lymphoid-potential 15,32 , their rate of entry into cell cycle 56 and their self-renewal potential in vivo 57 . Here, we demonstrated that CD150 + 34 -KSL Sdc2 + cells contain nearly all of the in vivo repopulating capacity of the CD150 + 34 -KSL HSCs. Therefore, Syndecan-2 surface expression may have utility toward resolving the heterogeneity in self-renewal capacity within the CD150 + 34 -KSL HSC population. In keeping with this conclusion, we found that Sdc2 + HSCs are highly enriched for expression of numerous genes associated with enhanced hematopoietic cell stemness. Finally, our studies also suggest that Syndecan-2 surface expression on phenotypic HSCs changes in response to cytokine stimulation and following competitive transplantation in mice; further studies will address whether loss or gain of Syndecan-2 expression by phenotypic HSCs also reflects a change in self-renewal capacity.
Beyond its importance as a marker of LT-HSCs, Syndecan-2 uniquely regulates HSC function via control of HSC quiescence. Silencing of Syndecan-2 expression caused an increase in HSC cycling that persisted for several months following transplantation in mice.
Commensurate with this, Syndecan-2 silencing depleted HSCs with long-term in vivo repopulating capacity. The functional role of other HSC markers such as CD150 and endothelial protein C receptor has not been defined. 8,12,15 Similarly, α-catulin GFP/GFP mice have been shown to have normal hematopoiesis and normal HSC content and function. 29 Surface expression of endothelial cell-specific adhesion molecule 1 (Esam1) and JAM-C have been shown to enrich for murine HSCs and deletion of Esam1 was associated with lineage skewing, whereas JAM-C deficient mice displayed increased myeloid cells. 10,58 In our studies, selection of BM lincells or CD34 -KSL HSCs for Sdc2 expression caused significant gain of HSC repopulating function in vivo, whereas silencing of Sdc2 expression in HSCs increased HSC cycling and depleted long-term repopulating HSCs. Syndecan-2 is uniquely modified by the addition of three heparan sulfate chains, whereas Syndecan-1 and Syndecan-3 are decorated by both chondroitin and heparan sulfate chains, 59 which have been shown to have opposing physiological roles in other cell types. 60,61 As such, our studies demonstrating the unique self-renewal capacity of Syndecan-2 + HSCs provides the basis for the development of strategies to increase heparan sulfate content on LT-HSCs.
Syndecan-2, expressed as a transmembrane HSPG, can serve as a co-receptor for TGF-β and either promote or inhibit TGF-β receptormediated signaling in a contextdependent manner. 62, 63 We have found that Syndecan-2 promotes HSC quiescence and enhanced repopulating capacity through induction of Cdkn1c expression, which can be suppressed by inhibition of TGF-β signaling. These data suggest that Syndecan-2 promotes HSC quiescence and enhanced repopulating capacity through activation of TGF-β receptor signaling and sustainment of Cdkn1c expression. TGF-β signaling and Smad pathway activation have also been shown to regulate p57 protein levels via control of proteolysis, so we will further explore this mechanism in regulating the quiescence of Syndecan-2 + HSCs. 64 Since Syndecans can regulate the activity of other growth factors, integrins and extracellular matrix proteins that may act on HSCs, 65 our findings also provide the basis to explore additional Syndecan-2mediated pathways that regulate HSC quiescence and repopulating capacity. In summary, Syndecan-2 expression enriches for HSCs and Syndecan-2 regulates HSC repopulating capacity through control of HSC quiescence.

Conflict-of-interest disclosure
The authors declare no competing financial interests.

Data Sharing Statement
For original data, please contact john.chute@cshs.org or Christina.Termini@cshs.org.