Improved hematopoietic differentiation of mouse embryonic stem cells through manipulation of the RNA binding protein ARS2

Abstract

The RNA binding protein ARS2 is highly expressed in hematopoietic progenitor populations and is required for adult hematopoiesis. Recent molecular studies found that ARS2 coordinates interactions between nascent RNA polymerase II transcripts and downstream RNA processing machineries, yet how such interactions influence hematopoiesis remains largely unknown. Techniques to differentiate embryonic stem cells (ESC) to hematopoietic progenitor cells (HPC) and mature blood cells have increased molecular understanding of hematopoiesis. Taking such an in vitro approach to examine the influence of ARS2 on hematopoiesis, we found that ARS2 suppresses expression of some HSC signature genes and differentiation of ESC to a HPC population (CSMD-HPC) identified by markers expressed on bone marrow resident hematopoietic stem cells. In line with ARS2's ability to promote proliferation of cultured cells, ARS2 knockout ESC showed limited expansion and yielded less CSMD-HPC than wild-type ESC. In contrast, transient ARS2 knockdown led to doubling the number of CSMD-HPC generated per ESC without affecting further differentiation into mature T-cells. Overall, data indicate that ARS2 negatively regulates early hematopoietic differentiation of ESC, in stark contrast to its supportive role in adult hematopoiesis. Consequently, manipulation of ARS2 expression and/or function has potential utility in hematopoietic cell engineering and regenerative medicine.

1. Introduction

Hematopoietic stem cells (HSC) residing in adult bone marrow have the capacity to differentiate into all blood lineages while maintaining a pool of self-renewing HSC. As such, transplantation of HSC has been a highly effective therapeutic strategy for many hematological malignancies and autoimmune disorders [1–3]. Consequently, methods to efficiently expand or generate hematopoietic stem cells in vitro have been the focus of significant research efforts for many years. Despite clinical application of ex vivo methods to expand HSC, the inability to produce large quantities of HSC from a limited number of input cells (e.g. cord blood cells) imposes a major limitation on such efforts [4, 5]. Differentiation of HSC in vitro from an unlimited supply of pluripotent stem cells has the potential to overcome this limitation. Several systems have been established to form HSC from pluripotent embryonic stem cells (ESC), such as overexpression of Hoxb4 gene in embryoid bodies [6–8], formation of teratomas in mice [9–11] or introduction of several transcription factors to directly reprogram ESC [12]. While current protocols are successful in creating hematopoietic progenitor cells (HPC) capable of establishing chimerism [13] or functional lymphoid/myeloid progenitors [14], resulting hematopoietic cell populations have limited similarity to HSC found in vivo at the gene expression level and the tricky, time-consuming differentiation protocols employed have hampered bench-to-bedside translation.

Over the past few years there has been increasing appreciation of the importance of RNA binding proteins in regulating gene expression for precise orchestration of hematopoiesis and prevention of leukemic transformation [15, 16]. Arsenic resistance protein 2 (ARS2) is a highly conserved RNA binding protein that functions to co-transcriptionally coordinate many aspects of RNA maturation [17–24]. ARS2 is highly expressed in hematopoietic stem/progenitor cells, and its depletion is lethal during pre-implantation embryogenesis around E5.5 [25]. Our recent work found that ARS2 is essential for hematopoiesis in adult mice, with rapid loss of long-term repopulating bone marrow resident HSC and thymocytes observed in ARS2^KO animals [26]. While we were able to demonstrate increased apoptosis of ARS2^KO thymocytes, how ARS2 contributes to the formation and/or maintenance of HSC is unknown.

In this study, we set out to investigate the molecular role of ARS2 in primitive hematopoietic cells using in vitro differentiation of ARS2 depleted mouse ESC as a model. Data demonstrate that ARS2 knockout or knockdown drastically skewed hematopoietic differentiation of ESC toward cells (termed CSMD-HPC for cell surface and molecularly defined-HPC) that express HSC signature genes and surface markers characteristic of HSC found in bone marrow of adult mice [27]. Interestingly, we found that knockout of ARS2 limited expansion of ESC in ESC-HPC cultures. As a result, ARS2^KO ESC yielded significantly less CSMD-HPC than control ESC. In contrast, transient siRNA knockdown of ARS2 had limited effect on expansion of ESC in ESC-HPC cultures and resulted in double the number of CSMD-HPC generated over 8 days of culture. Importantly, terminal differentiation of in vitro derived CSMD-HPC to mature T-cells was unaffected by ARS2 knockdown. Rather, the efficiency of in vitro differentiation of mouse ESC to CSMD-HPC and terminally differentiated T-cells was doubled by transient ARS2 depletion. Findings indicate a role for ARS2 in limiting early hematopoietic differentiation, a role that may be co-opted by limitation of ARS2 expression, or potentially function, to improve in vitro differentiation of ESC to mature hematopoietic cells.

2. Materials and Methods

2.1. Embryonic stem cells (ESC)

Inducible ARS2^KO ESC were derived by crossing super-ovulated female Srrt^flox/flox (ARS2^f/f, C57BL/6 background) mice with male Cre-ERT2-expressing ARS2^f/f mice [26] and harvesting blastocysts on day 3.5 following vaginal plug identification [28]. Resulting ESC clones were genotyped by end-point PCR (Fig. S1) to confirm both Srrt alleles contained loxP sites and that Cre-ERT2 was expressed. Spectral karyotyping (SKY) was performed on clones with the correct genotype by the Pathology Resource Network at Roswell Park Comprehensive Cancer Center (Fig. S1) and a male clone with a normal karyotype was identified. JM8A3.N1 ESCs (male, C57BL/6N) were used as a wild type control. ESCs were cultured in 2i media (Table S1) on 0.1% gelatin coated plates. Maintenance medium was supplemented with 1μM 4-hydroxytamoxifen (4-OHT, Abcam) for 4 days to knock out ARS2 prior to ESC-HPC culture.

2.2. ESC-HPC culture

To generate hematopoietic progenitor cells (HPC), ESC were cultured on OP9-DL1 stroma cells in OP9 medium (Table S1) following the protocol of Kučerová-Levisohn et al. [29] with minor modification. Cells were re-plated in 6 well plates containing OP9-DL1 cells on day 5 at 1 x 10⁵ cells/well in medium containing 5 ng mL⁻¹ FLT3 ligand (FLT3L) and harvested at day 8 for analysis or further differentiation.

To generate T-cells, CD41⁺, c-Kit⁺, CD150⁺, Sca-1⁺, EPCR⁺ CSMD-HPC were FACS sorted into 24 well plates (100 cells/well) containing OP9-DL1 cells in OP9 media supplemented with 1 ng mL⁻¹ interleukin-7 (IL-7) and 5 ng mL⁻¹ FLT3L. Cells were fed every 3 days and plated onto fresh OP9-DL1 every 6 days. Cells were collected 12 and 20 days following seeding into 24 well plates for analysis.

2.3. Additional materials and methods

Detailed methods can be found in Supplemental Methods and materials can be found in the Key Resources Table.

KEY RESOURCES TABLE.

Reagent or resource	Source	Identifier
Antibodies
Rabbit polyclonal anti-ARS2 – N-terminal	Abeam	Cat#:ab192999
Mouse monoclonal anti-β-Actin (clone AC-15)	Millipore Sigma	Cat#:A1978
Mouse monoclonal anti-ARS2 (clone 2G10)	Gruber et al, 2012
BV421 anti-mouse Ly-6A/E (Sca-1) (clone D7)	Biolegend	Cat#: 108127
APC anti-mouse CD201 (EPCR) (clone RCR-16)	Biolegend	Cat#: 141505
APC/Cy7 anti-mouse CD117 (c-Kit) (clone 2B8)	Biolegend	Cat#: 105825
BV510 anti-mouse CD48 (clone HM48–1)	Biolegend	Cat#: 103443
PerCP/Cy5.5 anti-mouse CD150 (SLAM) (clone TC15- 12F12.2)	Biolegend	Cat#:115921
PE anti-mouse CD41 (clone MWReg30)	Biolegend	Cat#: 133905
PE/Cy7 anti-mouse CD34 (clone HM34)	Biolegend	Cat#: 128609
APC anti-mouse CD34 (clone HM34)	Biolegend	Cat#: 128611
FITC anti-mouse CD11b (clone M1/70)	Biolegend	Cat#: 101205
FITC anti-mouse/human CD45R/B220 (clone RA3–6B2)	Biolegend	Cat#: 103205
FITC anti-mouse Ly-6G (clone 1A8)	Biolegend	Cat#: 127605
FITC anti-mouse TER-119 (clone TER-119)	Biolegend	Cat#116205
FITC anti-mouse CD3 (clone 17A2)	Biolegend	Cat#: 100203
BV421 anti-mouse CD3 (clone (17A2)	Biolegend	Cat#100227
PE/Cy7 anti-mouse CD4 (clone GK1.5)	Biolegend	Cat#: 100421
APC anti-mouse CD8a (clone 53–6.7)	Biolegend	Cat#: 100712
Bacterial and Virus Strains
Biological Samples
Chemicals, Peptides, and Recombinant Proteins
4-hydroxytamoxifen (4-OHT)	Millipore Sigma	Cat#:H6278
Premium USDA Origin Fetal Bovine Serum (lot 221C16)	Biowest	Cat#:S1620
US Origin Fetal Bovine Serum	Sigma-Aldrich	Cat#: F6178
MethoCult GF M3434	StemCell Technologies	Cat#:03434
PD0325901	StemCell Technologies	Cat#: 72184
CHIR99021	StemCell Technologies	Cat#:72054
Mouse recombinant LIF	StemCell Technologies	Cat#:78056
Mouse recombinant FLT3L	StemCell Technologies	Cat#: 78011
Mouse recombinant IL-7	StemCell Technologies	Cat#:78054
Mouse recombinant SCF	StemCell Technologies	Cat#: 78064
Mouse recombinant IL-11	StemCell Technologies	Cat#: 78026
Critical Commercial Assays
Superscript IV Reverse Transcriptase	Invitrogen	Cat#: 18090050
Power SYBR Green PCR Master Mix	Invitrogen	Cat#:4368706
P3 Primary Cell 4D-NucleofectorX Kit	Lonza	Cat#:V4XP-3024
Deposited Data
Experimental Models: Cell Lines
Srrtfl,fl, CreERT2+ mouse ESC	This paper
JM8A3.N1 mouse ESC	KOMP	RRID:MMRRC 0509 61-UCD
Experimental Models: Organisms/Strains
C57BL/6J mice (bone marrow)	Jackson Laboratory	Stock#:000664
Srrtfl,fl, CreERT2+/− mice (C57BL/6 background)	Elahi et al. 2018
Oligonucleotides
Allstars Negative Control siRNA	Qiagen	Cat#:SI03650318
ARS2–1 siRNA	Gruber et al, 2009
ARS2–4 siRNA	Gruber et al, 2012
Primers forqRT-PCR, see Table S3.	IDT
Recombinant DNA
Software and Algorithms
Other

3. Results and Discussion

3.1. Hematopoietic differentiation of ARS2^KO mESC

To begin to assess the molecular role of ARS2 in hematopoiesis we derived karyotypically normal ESC from blastocysts of ARS2^fl/fl mice that express the tamoxifen-inducible Cre transgene Cre-ERT2 (Fig. S1; [26]). Knockout of exons 2 through 20 of Srrt, the gene coding for ARS2, from these mouse ESC was accomplished by addition of 1μM 4-hydroxytamoxifen (4-OHT) to serum free medium containing MEK and GSK-3α/β inhibitors (2i medium). Four days later ARS2 knockout was confirmed at the mRNA and protein levels (Fig. 1A, S1C). Importantly, 4-OHT treatment of strain- and sex-matched JM8A3.N1 mESC did not alter ARS2 expression.

Figure 1. HPC generation from ARS2^KO ESC.

(A) Expression of the ARS2 transcript (top) and protein (bottom) following 4 day exposure of ARS2^f/f, CreERT2⁺ ESC or strain- and sex-matched Jm8A3.N1 ESC to 4-hydroxytamoxifen (4-OHT, 1 μM). Bars represent relative expression ± standard deviation determined by the ΔCT method relative to Tbp endogenous control, dots represent biological replicates. (B) Schematic depicting in vitro strategy used to generate hematopoietic progenitor cells from ESC (ESC-HPC). (C) Representative bright field microscopy image of ESC-HPC cultures on day 5. (D) Number of loosely attached cells collected from control versus ARS2^KO ESC-HPC cultures on day 5, prior to re-plating at an equal density. Bars represent mean± standard deviation, dots represent biological replicates. **** p < 0.0001. (E) Representative phase contrast image of ESC-HPC cultures on day 8.

Following 4-OHT induced ARS2 knockout or vehicle control treatment of mESC, hematopoietic progenitor cells (HPC) were generated using a slightly modified version of the protocol described by Kučerová-Levisohn et al. (Fig. 1B;[29]). This differentiation protocol relied on co-culture of ARS2^KO or control ESC with GFP-expressing (GFP⁺) OP9-DL1 mesenchymal stroma cells [13] for 5 days followed by re-plating in the presence of recombinant FLT3-ligand (FLT3L; 5ng/mL) and culture for 3 additional days. Microscopic examination and counting of ESC-HPC cultures revealed severely reduced expansion of ARS2^KO cells prior to FLT3L addition (Fig. 1C & 1D). Assessment of cell-cycle revealed significant accumulation of ARS2^KO ESC in S-phase (Fig. S2). This finding was consistent with the known role of ARS2 in promoting replication-dependent histone mRNA biogenesis [18] to support the transit of ESC through the S-phase of the cell cycle. In contrast to day 5 cultures, phase-contrast microscopy of day 8 ARS2^KO cultures revealed an apparent increase of loosely attached luminous cells (Fig. 1E, S3), consistent with the morphology of ESC derived HPC (ESC-HPC). Since cells were re-plated at an equal density (1 x 10⁵ per well) at day 5 concurrent with FLT3L addition, these images suggested that ARS2 knockout may improve FLT3L-driven differentiation toward HPC.

3.2. ARS2^KO ESC-HPC express markers of potent long-term bone marrow HSC

To confirm the apparent increase in ESC-HPC in day 8 ARS2^KO cultures, flow cytometry for a commonly used ESC-HPC cell surface phenotype (CD34⁻, CD41⁺, c-Kit⁺) [30] was performed (Fig. S3). These studies revealed that the frequency of CD34⁻, CD41⁺, c-Kit⁺ ESC-HPC was similar between control and ARS2^KO cultures (Fig. 2A). Since initial flow cytometry was incongruent with phase-contrast images we examined the expression of additional HSC markers. Our flow cytometry analysis revealed that cells in ESC-HPC cultures expressed EPCR, Sca-1, and CD150, while CD48 expression was undetectable (Fig. S4). Moreover, we found a dramatic increase in expression of Sca-1 in ARS2^KO ESC-HPC cultures (Fig. 2B). Mouse bone marrow hematopoietic stem cells (HSC) with high expression of Sca-1 are capable of robust chimerism and long-term multi-lineage reconstitution [27, 31]. Such robust long-term bone marrow HSCs have a gene expression signature recently defined through indexed cell sorting combined with single-cell RNA-seq and single-cell bone marrow reconstitution assays [27]. Because ARS2^KO ESC-HPC displayed a large increase in Sca-1 surface expression, we asked whether ARS2 influenced expression of genes associated with long-term bone marrow HSC, including the Sca-1 gene Ly6a. Of the 12 genes we analyzed, 6 HSC signature genes (Sult1a1, Ly6a, Procr, Gbp8, Mllt3, and Ifitm1) were induced in ARS2^KO ESC-HPC cultures (Fig. 2C). Additionally, 3 genes (Syk, Muc13, and Thoc6) associated with bone marrow progenitors lacking long-term reconstitution ability had reduced expression in ARS2^KO ESC-HPC cultures, while a single non-HSC signature gene (Mrm1) was induced. These data establish that ARS2 represses expression of some HSC signature genes during in vitro ESC to HPC differentiation.

Figure 2. ARS2 limits expression of HSC signature genes and surface markers in ESC-HPC cultures.

(A) Representative overlay of contour plots (left) showing expression of c-Kit on CD34⁻, CD41⁺ cells from day 8 ARS2^KO (red) and control (blue) ESC-HPC cultures (see Fig. S4 for full gating strategy). Bars represent average frequency ± standard deviation (right) of CD34⁻, CD41⁺, c-Kit⁺ cells within the ESC-derived GFP⁻ population (OP9-DL1 feeder cells express GFP) in ARS2^KO versus control day 8 ESC-HPC cultures ± standard deviation. Dots represent biological replicates (n = 3, p > 0.05). (B) Representative histogram (left) depicting Sca-1 surface expression on cells from ARS2^KO (MFI = 1151 ± 175) versus control (MFI = 850 ± 160) ESC-HPC cultures. Bars represent average frequency ± standard deviation (right) of Sca-1^hi cells of day 8 control versus ARS2^KO ESC-HPC cultures, dots represent biological replicates (n = 5). *p = 0.01. (C) Relative expression in day 8 ARS2^KO ESC-HPC cultures of HSC signature genes (red) and genes associated to non-HSC bone marrow progenitor cells (blue) [27]. Bars represent mean log₂: fold change± standard deviation in ARS2^KO cultures relative to control cultures determined by qRT-PCR using the ΔΔCT method of quantification. Dots represent technical replicates from three independent experiments. Tbp was used as an endogenous control. (D) Representative histograms depicting CSMD-HPC in ARS2^KO versus control ESC-HPC cultures. (E) Frequency of CSMD-HPC within the ESC-derived GFP⁻ population of ARS2^KO versus control day 8 cultures. Bars represent mean± standard deviation, dots represent biological replicates (n = 4) **p = 0.009. (F) Calculated number of CSMD-HPC generated from a single control versus ARS2^KO ESC over 8 days of ESC-HPC culture. Bars represent mean± standard deviation, dots represent biological replicates (n = 4). ** p = 0.001.

The cell surface phenotype CD34⁻, c-Kit⁺, CD48⁻, CD150⁺, Sca-1^hi, EPCR^hi identifies bone marrow cells that express HSC signature genes and are capable of long-term hematopoietic stem cell function [27], whereas the cell surface phenotype CD34⁻, c-Kit⁺, CD41⁺ has been widely used to define ESC-HPC [30]. Since gene expression data indicated that ARS2 represses HSC signature genes, including Ly6a that codes for Sca-1 and Procr that codes for EPCR, we stained cells from control and ARS2^KO ESC-HPC cultures with a panel of antibodies that includes HSC and ESC-HPC markers (Table S2). This combination of markers allowed for identification of a population of c-Kit⁺, CD41⁺, CD150⁺, Sca-1⁺, EPCR⁺ cells (Fig. S5), referred to henceforth as CSMD-HPC for cell surface and molecular definition-HPC, that was found at significantly higher frequency in ARS2^KO ESC-HPC cultures (Fig. 2D, 2E).

As a cursory examination of CSMD-HPC phenotype cells we performed an in vitro assay previously used to determine the proliferation and differentiation characteristics of single HSC [27, 32–34]. In this assay, single cells were sorted from bone marrow of C57BL/6J mice using either CSMD-HPC or standard HPC markers (Table S2) and cultured for 10 days in medium supplemented with IL-11 and stem cell factor (SCF). These studies revealed that bone marrow cells expressing CSMD-HPC markers formed larger colonies that were enriched in undifferentiated lineage⁻, Sca-1⁺, c-Kit⁺ (LSK) cells than bone marrow cells expressing standard HPC markers (Fig. S6). When these assays were applied to single cells sorted from wild-type ESC-HPC cultures only small colonies were observed, regardless of surface phenotype. Flow cytometry revealed that single CSMD-HPC phenotype cells from ESC-HPC cultures generated a higher frequency of LSK cells than single standard HPC phenotype cells, regardless of the concentration of stem cell factor (SCF) in culture medium (Fig. S7). These early results demonstrate that CSMD-HPC phenotype cells from ESC-HPC cultures were able to undergo in vitro self-renewal and differentiation at least as well as standard HPC phenotype cells from ESC-HPC cultures. In vivo reconstitution assays are required to confirm in vitro self-renewal and differentiation data but are beyond the scope of the current study.

3.3. Transient ARS2 knockdown enhanced ESC to CSMD-HPC differentiation

Further examination of sorted CSMD-HPC versus standard HPC from wild-type ESC-HPC cultures revealed marked downregulation of ARS2 gene expression in CSMD-HPC (Fig. S8). Consistent with this, knockout of ARS2 increased the frequency of CSMD-HPC phenotype cells in ESC-HPC cultures (Fig. 2E). However, when the number of CSMD-HPC phenotype cells generated per input ESC was calculated, a striking decrease was observed in the ARS2^KO group (Fig. 2F). This decrease in CSMD-HPC number could almost entirely be attributed to defective expansion of differentiating ARS2^KO ESC over the first 5 days of culture (Fig. 1D, S2). Since previous experiments indicated that knocking down ARS2 with siRNA had a limited effect on proliferation compared to ARS2 knockout, we examined the effect of ARS2 siRNA on ESC-HPC cultures (Fig. 3A). To this end, two independent siRNAs targeting murine ARS2 (siARS2–1 and siARS2–4) [35], or a non-targeting control siRNA (siCTL), were transfected into wild type ESC two days prior to initiation of ESC-HPC differentiation cultures. This approach resulted in transient depletion of ARS2 (Fig. 3B) without reduction in the number of cells in ESC-HPC cultures (Fig. 3C). Similar to ARS2^KO ESC-HPC cultures, CSMD-HPC phenotype cells were found at a higher frequency in ARS2 knockdown ESC-HPC cultures (Fig. 3D, 3E). However, because transient ARS2 depletion did not limit ESC expansion, observed increase in CSMD-HPC frequency in ARS2 knockdown ESC-HPC cultures was reflective of a significantly increased number of CSMD-HPC generated from each input ESC (Fig. 3F). Taken together these data demonstrate that transient limitation of ARS2 can enhance the efficiency of in vitro ESC to CSMD-HPC differentiation.

Figure 3. Transient ARS2 depletion enhances efficiency of CSMD-HPC differentiation.

(A) Schematic depicting in vitro strategy used to generate ESC-HPC following siRNA transfection of ESC. (B) Expression of the mRNA coding for ARS2 on day 0 (2 days following siRNA transfection) versus day 5 of ESC-HPC culture. Bars represent mean expression ± standard deviation relative to control siRNA transduced cells at day 0 as determined by qRT-PCR using Tbp as an endogenous control. **** p < 0.001, n.s. = not significant, C = siCTL, 1 = siARS2–1, 4 = siARS2–4. (C) Number of loosely attached cells collected from control siRNA versus ARS2 knockdown ESC-HPC cultures on day 5 (n = 4). (D) Representative contour plot of CSMD-HPC in control siRNA transfected versus ARS2 knockdown day 8 cultures. (E) Frequency of CSMD-HPC within the ESC-derived GFP⁻ population of control versus ARS2 knockdown day 8 cultures (*p = 0.02, **p = 0.007; n = 3 ).(F) Calculated number of CSMD-HPC generated from a single control versus ARS2 knockdown ESC (*p = 0.03, **p = 0.004; n = 3) over 8 days of ESC-HPC culture.

3.4. Unaltered terminal differentiation of CSMD-HPC generated from ARS2 knockdown ESC

Manipulating genes involved in differentiation of ESC has the potential to perturb self-renewal and/or differentiation potential of resulting HPC. Standard methylcellulose colony assays were performed to examine whether ARS2 knockout or knockdown in ESC altered the general ability of cells from ESC-HPC cultures to be further differentiated. Consistent with our previous finding that deletion of ARS2 caused self-renewal defects in bone marrow HSC [26], cells from ARS2^KO ESC-HPC cultures exhibited limited ability to form colonies in methylcellulose (Fig. S9A). In contrast, cells from transient ARS2 knockdown ESC-HPC cultures formed a similar number of colonies in methylcellulose as controls (Fig. S9B), suggesting that transient ARS2 depletion did not affect the bulk differentiation potential of cells from ESC-HPC cultures.

To investigate the differentiation potential of ARS2 knockdown ESC-HPC further, FACS sorted CSMD-HPC were differentiated into T-cells by extending OP9-DL1 co-cultures, supplemented with IL-7 and FLT3L, for an additional 20 days (Fig. 4A) [29]. Flow cytometry analysis of semi-adherent cells on days 20 and 28 following initiation of co-cultures revealed a high frequency of CD3⁺ cells among the GFP⁻ fraction and no CD3 expression on GFP⁺ OP9-DL1 cells, which served as an internal control for T-cell marker staining (Fig. 4B, S10). Importantly, early siRNA-mediated ARS2 depletion had no effect on the frequency or level of CD3 expression. As expected, the majority of CD3⁺ cells from day 20 cultures expressed both CD4 and CD8 T-cell markers (Fig. 4C), similar to developing T-cells in the thymus [36]. Frequency of these double-positive cells was diminished in day 28 cultures. Again, no difference in CD4/CD8 double-positive T-cells was observed between control siRNA and ARS2 siRNA transfected cells at either time point. Consistent with previous reports [29, 37], the decrease in double-positive cells from day 20 to 28 of culture was accompanied by an increase in mature CD8 single-positive (CD8⁺) T-cells (Fig. 4D). Once again, the frequency of CD8⁺ T-cells was not affected by transient ARS2 knockdown. As ARS2 depletion profoundly inhibits in vivo thymic T-cell development [26], these data are consistent with restoration of ARS2 expression following transient siRNA-mediated knockdown over the first few days of co-culture (Fig. 3B). Of particular interest was the observation that transient ARS2 knockdown led to a doubling in the number of mature CD8⁺ T-cells obtained per input ESC (Fig. 4E).

Figure 4. Terminal T-cell differentiation of CSMD-HPC.

(A) Schematic depicting differentiation of mature T-cells from sorted CSMD-HPC. (B) Representative histogram (left) of CD3 surface expression comparing GFP⁻ cells generated from ARS2 knockdown versus control siRNA transfected ESC following 20 and 28 days of T-cell differentiation culture. GFP⁺ OP9-DL1 served as an internal control. Frequency (top right) and mean fluorescence intensity (MFI; bottom right) of CD3 on GFP⁻ cells from day 20 and 28 T-cell differentiation cultures. (C-D) Frequency of CD3⁺ cells from (B) that co-express (C) CD4 and CD8 T-cell markers or (D) only CD8 (CD8 single-positive = CD8⁺). (E) Calculated number of CD8⁺ T-cells generated from a single control versus ARS2 knockdown ESC over 20 or 28 days of T-cell differentiation culture (***p = 0.0002, **p= 0.009, ****p < 0.0001). For (B) - (E) bars represent mean± standard deviation and dots represent biological replicates (n = 7 for siCTL and siARS2–1, n = 4 for siARS2–4).

3.5. Conclusions

Data presented in this study implicate the RNA binding protein ARS2 in the regulation of ESC differentiation toward hematopoietic lineages, a process essential for mammalian embryogenesis. Unlike previous reports that identified positive roles for ARS2 in developmental fate decisions [25, 38, 39], we report here a negative regulatory role of ARS2 in early hematopoietic differentiation. Relief of ARS2's inhibitory influence on ESC diffrentiation through genetic knockout or transient RNAi knockdown induced expression of HSC signature genes and surface markers expressed by potent long-term repopulating HSC. ARS2 knockout produced a strong anti-proliferative response in differentiating ESC (Fig. 1D), while transient ARS2 knockdown had limited impact on proliferation (Fig. 3C). Critically, both ARS2 knockout and knockdown had similar effects on expression of HSC markers (Figs. 2D & 3D), suggesting that the functions of ARS2 in differentiation and proliferation are separable. Moreover, the doubling in efficiency of ESC to CSMD-HPC differentiation in the absence of effects on terminal differentiation capacity following transient ARS2 knockdown suggests that targeting ARS2 or downstream RNA maturation processes may be an attractive strategy to improve in vitro differentiation protocols aimed at generating mature blood cells for therapeutic purposes.

Abbreviations

ESC

embryonic stem cells

HSC

hematopoietic stem cells

HPC

hematopoietic progenitor cells

ESC-HPC

HPC derived by differentiation of ESC

CSMD-HPC

cell surface and molecularly defined HPC