GRK6 regulates ROS response and maintains hematopoietic stem cell self-renewal

G protein-coupled receptor kinases (GRKs) are critically involved in immune response through regulation of cytokine receptors in mature leukocytes, but their role in hematopoiesis is largely unknown. Here, we demonstrate that GRK6 knockout (GRK6−/−) mice exhibit lymphocytopenia, loss of the hematopoietic stem cell (HSC) and multiple progenitor populations. GRK6 deficiency leads to compromised lymphoid differentiation, largely owing to the impairment of HSC self-renewal. Transcriptome and proteomic analysis suggest that GRK6 is involved in reactive oxygen species signaling. GRK6 could interact with DNA-PKcs (DNA-dependent protein kinase, catalytic subunit) and regulate its phosphorylation. Moreover, reactive oxygen species scavenger α-lipoic acid administration could partially rescue the loss of HSC in GRK6−/− mice. Our work demonstrates the importance of GRK6 in regulation of HSC self-renewal and reveals its potential role in participation of stress response.

G-protein-coupled receptor kinases (GRKs) are kinases that phosphorylate and desensitize agonist-bound G proteincoupled receptors (GPCRs). 1 To date, GRKs have been shown to have critical regulatory roles in the neuronal, 2,3 cardiac 4 and immune systems, 5,6 via GPCR-dependent and GPCR-independent mechanisms. Accumulating evidence implicates the importance of GRKs in regulating embryonic formation and development of key organs, including heart and brain. GRK2 knockout in mice are embryonic lethal due to marked cardiac abnormalities. 7,8 GRK2 mediates Smoothened-Hedgehog signaling desensitization, 9 which regulates neural tube formation and muscle development in zebrafish and mice. 10 Our earlier studies showed that GRK2 regulates cyclin B-dependent transcription, and downregulation of GRK2 in zebrafish embryos results in developmental early arrest and abnormal eye, midbrain and blood island formation. 11 Recently we showed that GRK5 regulates neuronal morphogenesis and memory formation via a GPCR-independent mechanism. 12 GRKs, especially GRK6, are highly expressed in vertebrate immune organs and peripheral blood cells. 13,14 GRK6 knockout (GRK6 − / − ) mice show increased severity of acute inflammatory arthritis 15 and colitis 16 because of enhanced granulocyte chemotaxis, and develop autoimmune diseases due to impaired macrophage engulfment. 17 GRK6 regulates chemotaxis through SDF/CXCLs-CXCR4, 18,19 leukotriene B4-induced CGRP receptor 20 and BLT receptor 21 activation. Moreover, It has been reported that the expression and activity of GRK6 change during differentiation of the promyelocytic cell line HL-60, 22 suggesting the potential involvement of GRK6 in earlier leukocyte development. However, the role of GRK6 in the development of blood lineages, from hematopoietic stem cells (HSCs) to differentiated immunoreactive leukocytes, is unclear.
The balance of cellular quiescence and proliferation of HSCs is critical for the preservation of their capacity for self-renewal and differentiation. In this study, we demonstrate that GRK6 regulates of self-renewal and lymphoid differentiation of HSCs, and underscore its underlying importance in the maintenance of reactive oxygen species (ROS) homeostasis.

Results
GRK6 knockout leads to lymphocytopenia. To investigate the potential role of GRK6 in steady-state hematopoiesis, we first analyzed peripheral blood of 8-12 weeks old wild-type (WT) and GRK6 − / − mice. Compared with WT littermates, GRK6 − / − mice showed prominent decrease in lymphocyte number, and a slight reduction in red blood cells (RBCs) and platelets (Figures 1a-d). T-and B-cell percentages in GRK6 − / − mice both dropped, which further indicates lymphoid deficiency (Figures 1f and g). The CD4 + population of T cells was smaller than that of WT, suggesting immunosuppression ( Figure 1h), but the total granulocyte number was slightly increased (Figure 1e), consistent with elevated CD11b + Gr-1 + granulocyte population (Figure 1i). The skewed peripheral blood constitution caused by GRK6 ablation suggests that it may participate in regulation of hematopoiesis and lymphoid differentiation.
GRK6 knockout reduces HSC and progenitor populations. Dissected GRK6 − / − femurs and tibiae were paler, with fewer cell number than age-matched WT, suggesting bone marrow suppression (Figures 2a and b). Femoral H&E staining showed fewer cells in the marrow cavity (Figure 2c, left), markedly decreased RBCs, more early erythroid elements (Figure 2c, middle, arrowhead), while Wright-Giemsa revealed a notable increase in ring-like nuclei of immature myeloid cells in bone marrow smears (Figure 2c, right, white arrowhead) of GRK6 − / − bone marrow (BM). These results indicate that GRK6 knockout leads to distorted hematopoietic cell maturation.
We then examined hematopoietic stem cell and progenitor populations to find out the fractions affected by GRK6 ablation. The results showed decreased HSC populations, LSK (Lin − Sca-1 + cKit hi ), and side population (Hoechst 33342 lo LSK) in GRK6 − / − BM (Figures 2d-f). The HSC sub-populations, long-term HSC, and short-term HSC, as well as multi-potent progenitor in GRK6 − / − BM dropped dramatically (Figures 2d and g-i). Loss of lymphoid competent HSC 23,24 and lower common lymphoid progenitor (CLP) population were observed (Figures 2d, j and k), indicating lymphoid hematopoiesis deficiency in GRK6 − / − BM. Common myeloid progenitors and megakaryocyte/erythrocyte progenitors in GRK6 − / − BM were also reduced, while the number of granulocyte/monocyte progenitors did not change (Figures 2d and l). The loss of these progenitor populations is consistent with decreased peripheral RBCs and lymphocytes, indicating the importance of GRK6 in HSC maintenance and differentiation. GRK6 is essential for HSC self-renewal. To investigate the function of GRK6 in hematopoiesis, WT or GRK6 − / − BM was transplanted along with EGFP + BM as competitor (Figure 3a). The results show that peripheral reconstitution of granulocytes was comparable ( Figure 3b). However, higher percentage of EGFP + cells was observed in total peripheral reconstitution, as well as T cells and B cells of GRK6 − / − BM recipients than that of WT BM recipients ( Figure 3b). Moreover, higher percentage of EGFP + HSCs was observed in bone marrow of GRK6 − / − recipients than those of WT recipients (Figure 3c), indicating lower ability of GRK6 − / − HSCs to reconstitute. These data suggest that GRK6 may regulate HSC maintenance.
Serial transplantation of sorted WTand GRK6 − / − LSKs was carried out to assess the potential role of GRK6 in HSC homeostasis ( Figure 3d). To evaluate the reconstituting capability of WT and GRK6 − / − HSC, we quantified the recipient-derived residual GFP + cells after serial transplantation. The secondary recipients of GRK6 − / − LSK cells exhibited higher EGFP + percentage of WBCs, granulocytes, T cells and B cells than those received WT LSK cells (Figure 3e), and consistently, smaller GRK6 − / − -derived HSC population (Figure 3f) was observed, indicating stem cell exhaustion. The above data indicate that GRK6 regulates HSC-intrinsic self-renewal.
GRK6 regulates differentiation of hematopoietic progenitor cells. Loss of lymphocytes in GRK6 − / − mice suggests that GRK6 may also participate in regulating differentiation. Serial peripheral blood sampling after a single 5-fluoruracil injection revealed slower rebound of number for WBCs, lymphocytes, and neutrophils in GRK6 − / − mice (Figure 4a     Enrichment map 25,26 analysis was used to sum up significantly altered pathways, which suggest that receptor signaling, transcription regulation, post-translational regulations, and notably stress and cell cycle related pathways were involved in GRK6-dependent mechanisms in hematopoiesis ( Figure 5a). We then asked if oxidative stress is involved in the phenotypic defects caused by GRK6 ablation. ROS is a major source of oxidative stress. DCF-DA staining showed that GRK6 ablation resulted in elevated ROS level (Figure 5b) in HSC and CLP populations. Moreover, DNA damage-induced γH2AX phosphorylation was significantly aggravated in GRK6 − / − HSC ( Supplementary Figures 2a and b). These data indicate that GRK6 may participate in cellular ROS cleavage and DNA damage repair.
In vivo and in vitro antioxidant treatment was utilized to see if increased ROS is causal. The results show that 50 mg/kg α-lipoic acid (LA) treatment significantly increased HSC count and CLP was slightly increased CLPs in GRK6 knockout mice  Figure S1c).
In parallel, we investigated the effect of GRK6 knockdown in Jurkat cells. Lentivirus-based shRNA was designed to target common exons of GRK6 ( Supplementary Figures 3a and b). To gain insight into molecular mechanism by which GRK6 regulates stress-related response, we tried to identify GRK6 interacting proteins with immunoprecipitation and mass spectrometry (IP-MS) analysis (Figure 6a). Interestingly, besides known substrates such as HSP90AA1, HSP90AA2 and HSP90AB1, proteomic screening revealed association of GRK6 with members of phosphatidylinositol-3-kinase-related kinase (PIKK) family, including ATM, ATR, and DNA-PKcs. Especially, 220 peptide fragments from DNA-PKcs were detected. DNA-PKcs is known to mediate non-homologous end joining and lymphocyte-specific V(D)J recombination, 27 and is recently reported as an 'ROS sensor'. 28 In response to H 2 O 2 treatment (200 μM), reduced phosphorylation of DNA-PKcs (pT2056), but not pATM (pS1981) was observed (Figure 5d). Moreover, DNA damage-induced γH2AX phosphorylation was significantly aggravated in GRK6-deficient cells (Figures 6c and d). Taken together, the interaction between DNA-PKcs and GRK6 might be essential for DNA-PKcs phosphorylation, which mediates cellular protection against oxidative stress, and possibly participates in lymphoid differentiation.

Discussion
GRKs have shown broad distribution in various tissues and act on a variety of substrates. GRK6 was known to critically control Here we report a novel aspect of GRK6 function in hematopoietic stem cell maintenance. We found that GRK6 ablation lead to pronounced lymphocytopenia, fewer HSCs, and smaller common lymphoid progenitor population. We also proved that GRK6 is essential to HSC self-renewal. Increased ROS level and DNA damage in GRK6 − / − HSC and CLP suggest involvement of GRK6 in maintaining redox homeostasis, and antioxidant treatment could at least partially rescue the loss of HSC and clonogenic ability of GRK6deficient bone marrow. Our data suggest an indispensable role of GRK6 in regulating hematopoietic stem cell renewal. Regulation of ROS level is critical in maintaining stem cell self-renewal and differentiation, as well as in treatment of stem cell associated diseases. High level of ROS has been long suggested detrimental. [29][30][31] However, several studies have shown that low physiological level of ROS and other stressors operate as intracellular signaling molecules and promote stem  [32][33][34] Redox regulation of HSC quiescence and self-renewal, lymphoid and myeloid balance, 34 and T-cell differentiation 35 have been reported in recent literatures. 36 We found that GRK6-related redox disturbance could be rescued by a relatively low dose of α-lipoic acid (Figures 5, 6 and Supplementary Figure 2). Interestingly, in healthy mice, HSC count was unexpectedly reduced by in vivo LA administration (Figure 5c). Furthermore, in WT bone marrow culture, whereas CFU-C ability was improved by 30 μM LA, LA supplementation seems destructive to Pre-B growth (Figure 5d) and growth of healthy Jurkat cells (Supplementary Figure 3d). These data implicate that differentiation of myeloid and lymphoid progenitors require distinct ROS level, and ROS seems essentially promotes lymphoid growth and differentiation. Furthermore, excessive antioxidants administration could be hazardous to normal objects, and should be administrated with caution.
In our IP-MS results, 220 peptides from DNA-PKcs were detected, in contrast to the few fragments found from ATM and ATR, the other 22 PIKK family members, indicating specific and strong protein association. DNA-PKcs mediates nonhomologous end joining pathway and lymphocyte-specific V(D)J joining by recruiting its complex components to the sites of DNA double-strand break. DNA-PKcs knockout or Thr2609 cluster mutation leads to premature-aging and immunodeficiency due to blocked V(D)J rearrangement and DNA damage accumulation in mice. 37 DNA-PKcs was also reported to be an 'ROS sensor'. In response to ROS signal, DNA-PKcs undergoes phosphorylation at Thr2056 and subsequently phosphorylates p53 as well as other downstream transcription factors. 38 Interestingly, although both HSC and CLP exhibit higher ROS level and altered cellular stress pathway in GRK6 − / − mice, they responded differentially to LA dose and treatment, and CLP number as well as lymphoid differentiation was not restored by ROS scavenger. It is likely that redoxindependent functions of DNA-PKcs, like non-homologous end joining in lymphocytes, are involved. One thing to note is that GRK6 has many splice variants which differ in terms of their regulation by carboxyl-terminal post-translational modification. 39 These variants exhibit differential subcellular and tissue distribution. 40 The GRK6 knockout mice were generated through deletion of common exons 3-9. Therefore, the GRK6 ablation phenotype was derived from functional inactivation of all splice variants. Clarifying the dominant isoform of GRK6 in hematopoiesis would help to further the understanding of GRK6 function. Further mechanisms, like how GRK6 mediates phosphorylation of DNA-PKcs and whether it depends on kinase activity of GRK6, needs to be further elucidated.
Previous studies indicated that GRK6 participates in cytokine and chemokine-mediated immune responses. It was demonstrated that perturbed CXCR4 signaling leads to abnormalities of mature and immature hematopoiesis, while GRK6 negatively modulates ligand-induced CXCR4 desensitization. Thus enhanced CXCR4-mediated neutrophil chemotaxis and impaired responsiveness to G-CSF was observed in GRK6-deficient mice. 41 In transcriptome analysis, we also identified alterations in receptor and ligand expression in this well-reported regulatory axis in GRK6-deficient HSC and CLP (Figure 5a). However, in spleen clonogenic assay, comparable clonogenic ability of GRK6 − / − and WT bone marrow cells was observed, indicating dispensable migration and reconstitution of hematopoietic progenitor cell activity in GRK6 ablation (Supplementary Figure 4). In combination with other data presented, we believe that the novel function of GRK6, in regulating DNA-PKcs phosphorylation and maintaining redox homeostasis, is a critical pathway that maintains HSC selfrenewal and differentiation.
In conclusion, our study provides evidence that GRK6 is essential to the maintenance of self-renewal of HSC, and acts as a regulator of redox homeostasis. Given that GRK6 knockout mice are viable, and prominent growth arrest of lymphoid leukemia cell line was observed upon GRK6 knockdown, it might be feasible that GRK6 be used as a potential target for treatment of leukemia.

Materials and Methods
Animals. GRK6 knockout strain backcrossed to C57BL/6 background was kindly provided by RJ Lefkowitz and RT Premont (Duke University Medical Center, Durham, NC, USA). WT and GRK6 knockout (GRK6 − / − ) littermates were produced by GRK6 heterozygote crossing and genotyped by triplex PCR amplification using tail tip genomic DNA. 42 Transgenic mice with ubiquitous somatic expression of enhanced green fluorescent protein (C57BL/6-Tg(CAG-EGFP)1Osb/J strain, hereafter referred to as EGFP mice) 43 was purchased from Model Animal Research Center of Nanjing University and maintained by homozygote crossing. Mice were on a reversed 12h light/dark cycle with food and water available ad libitum. Mice used for experiments were 8-12 weeks old unless otherwise specified. All experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and approved by Animal Care and Use Committee of Shanghai Medical College, Fudan University.
Hematology analysis. Peripheral blood was obtained by retro-orbital puncture into EDTA·K 2 -treated tubes. Complete blood counts were performed on Sysmex KX-21N or XS-800i hematology analyzers (Sysmex Corporation, Kobe, Japan). Bone marrow transplantation. EGFP mice were used as transplantation recipients. In competitive transplantation assay, BM cells from 8-12 weeks old WT or GRK6 − / − littermates were 1:1 mixed with EGFP + BM cells and 1 × 10 6 cells were intravenous injected into each EGFP recipient. In non-competitive serial transplantation assay, a mixture of 1000 LSK cells from WT or GRK6 − / − BM and 1 × 10 5 EGFP + Sca-1 − cells from BM of EGFP mice were intravenously injected into each EGFP + recipient. Three months later, BM cells from these recipients were used for secondary transplantation. The recipients were GFP + transgenic mice and the donors were WT or GRK6 − / − mice which carry no GFP fluorescence. The GRK6 regulates HSC self-renewal Q Le et al recipient-derived residual GFP + cells were quantified to evaluate the reconstitution capability of WT and GRK6 − / − HSCs. All recipient mice received 8.5 Gy ionizing radiation (IR). Transplantation was performed within 6 h post irradiation. EGFP + percentage of each peripheral and bone marrow population was analyzed by flow cytometry as indicated.
5-Fluoruracil assay. 5-Fluoruracil was administered intraperitoneally at 150 mg/kg on day 0. Retro-orbital bleeding was performed at indicated time and blood samples were assayed with a Sysmex XS-800i hematology analyzer.
Mice α-Lipoic acid treatment. For α-lipoic acid treatment, eight-week old mice were given daily i.p. 50 mg/kg α-lipoic acid administration for two weeks, and mice were sacrificed 24 h. after the last injection.
Immunofluorescence and histological microscopy. HSCs were sorted onto poly-L-lysine coated slides with and fixed with 4% paraformaldehyde for 15 min. Cells were permeablized with 0.3% Triton X-100, blocked with 1% FBS +0.1% BSA in PBS, incubated with γH 2 AX antibody (JBW301) overnight at 4°C and counterstained with DAPI (4,6 diamidino-2-phenylindole). Images were captured with LSM510 META confocal microscope (Carl Zeiss , Oberkochen, Germany). For each cell, z-stack of optical sections, 0.75 μm each and 15 μm in total thickness was performed and merged into one with Aim Image Browser (Carl Zeiss). Dots within DAPI positive area were enumerated with Image-Pro plus 6.0 (Media Cybernetics Inc., Rockville, MD, USA). At least 100 cells were calculated in each population. For H&E staining, femur from WT and GRK6 − / − mice were fixed in 4% paraformaldehyde, and decalcified in 14% EDTA, sectioned and stained with hematoxylin and eosin (H & E). Bone marrow smears were prepared and subjected to Wright-Giemsa staining. Images were captured with Olympus BX41 microscope.
Plasmid construction. Plasmids encoding Flag-tagged human GRK6 splice variant 1 and GFP were constructed into pCDNA3.0. Construction of shRNA plasmids for human GRK6 was performed as described. 44 Briefly, common coding sequence of human GRK6 mRNA sequence was analyzed by BLOCK-iT RNAi Designer and shRNAs were picked out, synthesized and constructed into FG12-hU6-shRNA vector. The shRNA lentivirus system was obtained from Dr. Immunoprecipitation and western blotting. For immunoprecipitation, 293T cells was transfected with pCDNA3.0-GRK6-Flag construct or pCDNA3.0-GFP and harvested 48 h later. Cells were washed with ice-cold PBS and lysed in IP buffer (50 mM HEPES, pH 8.0, 250 mM NaCl, 0.5% NP-40, 10% glycerol, 2 mM EDTA, 1 mM Na 3 VO 4 , 10 μg/ml aprotinin, 10 μg/ml benzamidine, and 0.2 mM PMSF) for 1.5 h as described. 46 The supernatant was incubated with anti-FLAG M2 affinity gel at 4°C for 8 h. The beads were subsequently washed, and the proteins bound to the beads were subjected to western procedures.
Reverse Transcription and real-time polymerase chain reaction. PrimeScript RT reagent Kit (Takara Biotechnology (Dalian) Co, LTD., Dalian, China) was used for regular reverse transcription. For HSC and CLP populations, cells were sorted and subjected to reverse transcription and cDNA amplification according to Smart-seq2 method. Briefly, collected cells were subjected to reverse transcription, template switching and preamplification for 20 cycles, and validated for housekeeping gene expression. Then amplified cDNA was purified, and used for RT-PCR detection. PCR primers were designed with NCBI Primer Blast and tested for PCR band specificity and excluded for genomic DNA amplification. Real-time PCR was done on an Eppendorf Realplex 2 real-time PCR cycler using the Takara SYBR Green PCR kit. All samples were run in duplicate. Relative expression was determined using the 2 − ΔΔCT method. See Supplementary Table 1 for primers used.
RNA sequencing and data analysis. Sample preparation was performed as described. 47,48 Briefly, sorted cells were individually transferred into lysis buffer with a mouth pipette. Reverse transcription reactions were directly performed on the whole-cell lysate. Poly (dT) primer was used to specifically reverse transcribe mRNA from the lysate. Terminal deoxynucleotidyl transferase was used to add a poly (A) tail to the 3′ end of the first-strand cDNA. The unincorporated primer was digested by exonuclease and 20+10 cycles of PCR was applied to amplify the cDNA. Sorted cells (13-15) from 3-4 mice were individually amplified and pooled together as cDNA libraries. Library preparation and sequencing were performed by WuXi AppTec. We obtained 20 million 100-bp PE reads for each lineage. Data analysis was performed with Genome assembly GRCm37/mm9. Data were aligned and calculated for differential expression with the Tuxedo suite. 49,50 Network analysis of over-represented pathways in RNA-seq by Enrichment Map. Gene sets were organized in a network, where each set was a node and edges represented gene overlap between sets. Automated network layout groups related gene sets into network clusters, indicated by grey shadows. Terms in each cluster were summarized by WordCloud.