CD1d expression demarcates CDX4+ hemogenic mesoderm with definitive hematopoietic potential

Highlights • scRNAseq of early hPSC differentiation reveals a CDX1/2/4+ CD1d + mesodermal population.• KDR + CD1d + mesoderm efficiently gives rise to hemogenic endothelium with erythroid, myeloid, and lymphoid potential.• CD1d-derived CD34 + cells robustly express HOXA7/9.


Introduction
A long-held goal of regenerative medicine has been the in vitro production of hematopoietic stem cells (HSCs) that can be used therapeutically and as a platform for the study of hematological disease. Despite recent advances in the field of directed differentiation of human pluripotent stem cells (hPSCs), this goal remains unrealized in the absence of transgene expression (Sugimura et al., 2017). This is due, in part, to our limited understanding of the overall complexity and heterogeneity of embryonic hematopoietic development, where multiple, spatiotemporally separated programs have been identified . Broadly, these can be separated into extra-embryonic, erythromyeloid programs and the intra-embryonic (definitive) multilineage programs, the latter of which gives rise to the HSC from hemogenic endothelium (HE) within the dorsal aorta (Gritz and Hirschi, 2016).
However, the mesodermal origin(s) of HE remains poorly defined. The hPSC differentiation system has identified several markers of early mesoderm with hemogenic potential, such as KDR (Kennedy et al., 2007), PDGFRA (Davis et al., 2008); and APLNR (Yu et al., 2012). However, these markers do not discriminate between other mesodermal lineages, nor do they define progenitors of the extra-embryonic-like and intra-embryonic-like hematopoietic programs. We have previously identified that mesodermal expression of GYPA/GYPB (CD235a/b) demarcates a population within hPSC differentiation cultures that harbors exclusively extra-embryonic-like hematopoietic potential, and that the expression of CD235a/b on this nascent mesoderm is regulated by stagespecific WNT and ACTIVIN/NODAL signaling (Sturgeon et al., 2014). However, the identification of a cell surface antigen that positively identifies a mesodermal population with exclusively definitive hematopoietic potential, but not extra-embryonic-like potential, has remained elusive. WNT-mediated expression of CDX1/2/4 has been correlated with specification of KDR+CD235a-mesoderm harboring potential for CD34+HOXA+ intra-embryonic-like cells with hematopoietic potential (Ng et al., 2016;Creamer et al., 2017;Jung et al., 2021). Thus, we aimed to use these CDX genes to identify definitive hematopoietic precursors within mesodermal hPSC cultures.

Single cell RNAseq analyses
Day 3 differentiation cultures were dissociated and immediately fixed with methanol as previously described (Alles et al., 2017). Briefly, EB's were treated with trypsin-EDTA for 5 min, stopped with 5% FBS + IMDM, and spun at 300 × g for 5 min. After resuspending with PBS + 5% FBS and counting, cells were pelleted at 300 × g for 5 min at 4 • C, the supernatant was removed manually, and the cell pellet resuspended in 2 volumes (200 μl) of ice-cold PBS. To avoid cell clumping, 8 volumes (800 μl) of methanol (grade p.a.; pre-chilled to -20 • C) were added dropwise, while gently vortexing the cell suspension (final concentration: 80% methanol in PBS). The methanol-fixed cells were kept on ice for a minimum of 15 min and then stored at -80 • C. For rehydration, cells were kept on ice, pelleted at 1000 × g, washed and resuspended in PBS + 0.01% BSA, passed through a 40-μm cell strainer, counted and diluted for library prep (1000 cells per µL). Libraries were prepared following the manufacturer's instruction using the 10X Genomics Chromium Single Cell 3 ′ Library and Gel Bead Kit v2 (PN120237), Chromium Single Cell 3 ′ Chip kit v2 (PN-120236), and Chromium i7 Multiplex Kit (PN-120262). 17,000 cells were loaded into a well of the chip, capturing > 6000 cells. cDNA libraries were sequenced on an Illumina HiSeq 3000. Sequencing reads were processed using the Cell Ranger software pipeline (version 2.1.0). Using Seurat (Stuart et al., 2019) (version 3.9.9) implemented in R (version 4.0.3) for all steps described below, the dataset was filtered by removing genes expressed in <3 cells, and retaining cells with unique gene counts between 200 and 6000. The remaining UMI counts were normalized and transformed, and regression was performed to account for the percent of mitochondrial UMI counts. Principal component analysis was used to generate uniform manifold approximation and project (UMAP) plots and unsupervised clustering was performed using a resolution of 0.9, resulting in 12 cell clusters. One cluster was removed for excess mitochondrial gene contribution (median > 10%). Differential gene expression analysis was performed using the 'FindAllMarkers' function with a minimum expression threshold of 25%, 0.25 log2 fold change, and p adj (Benjamini-Hochberg) < 0.01 to determine the cell identity of each cluster. The dataset is publicly available at the Gene Expression Omibus (GEO) under the accession number: GSE139850.

RNA expression analysis
For qRT-PCR, total RNA was isolated with the RNAqueous RNA Isolation Kit (Ambion), followed immediately by reverse transcription into cDNA using random hexamers and Oligo (dT) with Superscript III Reverse Transcriptase (Invitrogen). Real-time quantitative PCR was performed on a StepOnePlus thermocycle (Applied Biosystems), using Power Green SYBR mix (Invitrogen). Gene expression values were calculated using the ΔC T method against the endogenous control (ACTB). Primer sequences are ACTB F: AAACTGGAACGGTGAAGGTGACAG R: CAATGTGCAAT-CAAAGTCCTCGGC.
We and others (Ng et al., 2016;Creamer et al., 2017;Jung et al., 2021;Davidson et al., 2003;Wang et al., 2008) have previously demonstrated that early expression of CDX1/2/4 correlates with definitive hematopoietic specification, and that CDX4 acts as critical regulator of HE development from its mesodermal precursor (Creamer et al., 2017). We therefore hypothesized that CDX(4)+ clusters may contain these precursors to definitive HE, and could be used to identify unique cell surface antigens for their identification in vitro. Differential gene expression testing performed on each cluster compared to all remaining clusters revealed that CDX4 was positively enriched only within the "Early Mesoderm" (cluster 5) and "Mesoderm A" (cluster 8), but not within "Mesoderm B" (Supplementary Table 1). This led us to designate "Early Mesoderm" and "Mesoderm A" as the "CDX4 hi " clusters within the dataset. While "Mesoderm B" did not have significantly enriched CDX4 in comparison to the rest of the dataset, low level CDX4 expression was detectable, leading us to designate this cluster as "CDX4 lo ", and the remaining clusters, comprised of all other cell types, as "CDX4-" (Fig. 1D/E). Interestingly, when differential gene analysis was performed across these newly defined groups, CD1D, a noncanonical MHC receptor found on antigen presenting cells (Beckman, 1994), was found to be enriched in the CDX4 hi clusters and correlated with high CDX1/2/4 expression (Supplementary Table 2, Fig. 1F).
Flow cytometric analysis of these differentiation cultures on day 3 following CHIR99021 and SB431542 treatment, 61.47 +/-3.121% (SEM, n = 6) of KDR + CD34-CD235a-mesoderm was CD1d+ ( Fig. 2A). While FACS-isolated CD1d + cells were 8-fold enriched and 20-fold enriched respectively in CDX1 and CDX4 expression (Fig. 2B), there was a non-statistically significant ~2-fold enrichment of CDX2 (p = 0.0545). This agrees with the observation in the scRNAseq dataset that the CDX4-and CDX4 lo populations also expressed CDX2 but lacked any CD1D expression (Fig. 1E/F). To determine the hematopoietic potential of each mesodermal population, KDR+CD235a-CD1d neg /+ cells were isolated by FACS and cultured an additional 5 days under hematoendothelial promoting conditions (Sturgeon et al., 2014). Interestingly, each population gave rise to a CD34+CD43-population at similar frequencies (Fig. 2Aii). Gene expression analysis of isolated CD34+ endothelium from each culture revealed that the CD1d+ derived endothelium was significantly enriched for HOXA7/9 and RUNX1 expression in comparison to those derived from CD1d neg mesoderm (Fig. 2C), suggesting that hemogenic potential may be restricted to the CD1d-derived CD34+ cells. We therefore isolated each by FACS to assess for the presence of functional HE (Ditadi et al., 2015). After culturing as a monolayer for an additional 9 days under endothelial-tohematopoietic transition (EHT) promoting conditions, the CD1d negderived CD34+ cells exhibited no signs of multilineage hematopoietic potential. However, the CD1d+ -derived CD34+ cells exhibited robust hematopoietic potential, with many round non-adherent hematopoietic cells observed in the culture (Fig. 2D). These cells were confirmed to be hematopoietic progenitors, as when we transferred this population to methylcellulose, many erythro-myeloid CFU were detected (Fig. 2E).
In summary, single cell transcriptomics of early stage hPSC differentiation cultures has revealed a subpopulation of mesoderm, preceding hemato-endothelial specification, that is highly enriched for both CDX4 and CD1D. Subsequent functional in vitro studies demonstrate that CD1d is a marker for CDX4 hi hemogenic mesoderm with multilineage definitive erythroid, myeloid, and lymphoid potential. While this does not demonstrate a cell-autonomous role for CDX4 in the development of the definitive hematopoietic program, the strong correlation between CD1d and CDX4 expression, coupled with our previous studies demonstrating that CDX4 regulates definitive hematopoietic development within Fig. 1. scRNAseq allows for characterization of CDX + mesoderm and reveals CD1D as a potential surface marker: A. hPSC's were differentiated following exposure to stage-specific BMP4, bFGF, CHIR99012, and SB43152 treatment. On day 3, the differentiation culture was harvested for scRNAseq. B. Seurat v3 R package was used to organize the cells into high dimensional space via UMAP and algorithmically distinct clusters denoted by different colors labeled with a putative cell type. C. Expression of differential lineage markers in violin plots broken down by cluster, scale = log10 (transcripts). D. UMAP of combined clusters with CDX4-, CDX4 lo , CDX4 hi groups as determined by global DEG analysis. E. Violin plots of the expression of CDX genes and CD1D within different groups, scale = log2 (transcripts). F. Expression of CDX1/2/4 and CD1D genes using the module function in Seurat v3 shown on the UMAP where red indicates higher expression and blue lower than expected by chance by random sampling of 5 genes. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
It also remains to be determined which mesodermal lineage(s) arise from CD1d neg mesoderm. It will be of great interest to determine if there are functional differences in the CD34+ endothelial progeny from CD1d+ and CD1d neg mesoderm. For example, our scRNAseq analyses show that the CDX4 lo cluster is enriched in cardiogenic genes such as TBX2/4 and HAND1/2 (Supplementary Table 2). One possibility is that CD1d neg mesoderm harbors endocardial potential, which is consistent with observations that early CDX(1/4) expression is negatively correlated with cardiogenic potential in vivo and in vitro (Lengerke et al., 2011).
Finally, the CDX4 hi mesodermal population we identified exhibits enrichment for HOXA1 and HOXA3, of which the latter is a regulator of HE development (Iacovino et al., 2011). Curiously, the CDX4 lo population was enriched for HOXA9, HOXA10, HOXA11, and HOXA13, suggesting there are differences in distal/proximal HOX gene expression within each mesodermal subset (Supplementary Table 2). HOXA cluster expression has also been proposed as an important gene element of definitive HE in vitro and in vivo (Ng et al., 2016;Jung et al., 2021), but it is unclear which elements are necessary and/or sufficient for the emergence of HSCs. 'Medial' HOXA genes (HOXA5/7/9) have been demonstrated to be important for nascent HSCs (Dou et al., 2016) and, interestingly, both CDX4 hi and CDX4 lo populations are enriched for HOXA7. Distinct states of HOXA gene patterning were also observed when CD1d neg /+ mesoderm was isolated and differentiated into CD34+ endothelium. This evidence reinforces the critical role of gene specification early in differentiation and that appropriate patterning of medial HOXA expression within hemogenic mesoderm via additional signal pathway manipulation may ultimately prove critical in the derivation of HSCs from hPSCs.
We previously demonstrated that expression of both KDR and CD235a, prior to CD34 acquisition, identifies a mesodermal population that harbors exclusively extra-embryonic-like hematopoietic potential (Sturgeon et al., 2014). Here, we show that the combined markers of KDR, CD235a, and CD1d can identify a hemogenic mesodermal population with exclusively definitive hematopoietic potential. With this insight, mid-and high-throughput screening techniques can be applied to identify novel signal pathways that regulate definitive hematopoietic development, which in turn, should enable the development of technologies for the in vitro specification of HSCs from hPSCs.

Declaration of Competing Interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Christopher Sturgeon reports a relationship with Clade Therapeutics, Inc. that includes: board membership.

Data availability
Data will be made available on request.