Mercury Chloride Impacts on the Development of Erythrocytes and Megakaryocytes in Mice

Inorganic mercury (Hg2+) is a highly toxic heavy metal. The aim of this study was to investigate the impact of Hg2+ on the development of erythrocytes and megakaryocytes. B10.S mice (H-2s) and DBA/2 mice (H-2d) were administrated with 10 μM HgCl2 or 50 μM HgCl2 via drinking water for four weeks, and erythro-megakaryopoiesis was evaluated thereafter. The administration of 50 μM HgCl2 increased the number of erythrocytes and platelets in B10.S mice, which was not due to a reduced clearance for mature erythrocytes. The administration of 50 μM HgCl2, but not 10 μM HgCl2, increased the number of progenitors for erythrocytes and megakaryocytes in the bone marrow (BM) of B10.S mice, including erythroid-megakaryocyte progenitors (EMPs), burst-forming unit-erythroid progenitors (BFU-Es), colony-forming unit-erythroid progenitors (CFU-Es), and megakaryocyte progenitors (MkPs). Moreover, 50 μM HgCl2 caused EMPs to be more proliferative and possess an increased potential for differentiation into committed progenies in B10.S mice. Mechanistically, 50 μM HgCl2 increased the expression of the erythropoietin receptor (EPOR) in EMPs, thus enhancing the Jak2/STAT5 signaling pathway to promote erythro-megakaryopoiesis in B10.S mice. Conversely, 50 μM HgCl2 did not impact erythro-megakaryopoiesis in DBA/2 mice. This study may extend our current understanding for hematopoietic toxicology of Hg.


Introduction
Inorganic mercury (Hg 2+ ) is a highly toxic heavy metal that broadly exists in human living environments. Hg 2+ is derived from both the natural environment and human activities, the latter of which mainly include the combustion of fossil fuel and rubbish, mining activities, and sewage discharge; therefore, humans are exposed to Hg 2+ typically through the inhalation of contaminated air or the intake of contaminated food or water [1][2][3].
Hg 2+ has been shown to be toxic to erythrocytes, or red blood cells (RBC), which is likely due to a direct action of Hg 2+ on erythrocytes. For instance, Hg chloride (HgCl 2 ) increased the generation of reactive oxygen species (ROS) and reactive nitrogen species (RNS) to cause hemoglobin (HGB) oxidation in human erythrocytes in vitro [4]. HgCl 2 suppressed the activation of the glucose 6-phosphate dehydrogenase (G6PD) enzyme in erythrocytes in rats both in vitro and in vivo [5]. In addition, an in vivo study reported that long-term treatment with HgCl 2 on rats reduced the number of erythrocytes, caused lipid oxidation and chromosomal aberrations, and altered the activity of anti-oxidative Toxics 2021, 9,252 3 of 20 and humidity and an artificial 12-hour (hr) dark-light cycle at Fudan University. A total of five to six mice were housed in each cage with free access to food and drinking water.
Adult mice aged from 6 to 8 weeks (w) were used, and both males and females were equally used. Mice were administrated with 10 µM (2 mg/L Hg) or 50 µM (10 mg/L Hg) HgCl 2 (Sigma, St Louis, MO, USA) via drinking water for 4 w; regular drinking water was used as a vehicle control. The doses and time period of Hg treatment were according to our previous publications [19,28,29,34,35]. One criterion for dose selection was that the dose must be comparable to that of human exposure. The other criterion for dose selection was that the dose should be able to cause autoimmunity to test whether Hg impacts on erythro-megakaryopoiesis were related to autoimmunity. Our preliminary experiments suggested that the blood Hg concentration of mice treated with 10 µM or 50 µM HgCl 2 was comparable to that of human subjects living in mining areas [36,37]. Our previous study indicated that 50 µM HgCl 2 treatment was able to induce autoimmunity in mice carrying the H-2 s haplotype [28]. During the period of HgCl 2 administration, no differences in water consumption were observed.
Assays performed on mice were conducted twice or thrice, and therefore the number of mice used for each individual experiment was not exactly the same.
Fudan University Animal Care and Use Committee approved this work. All animal experiments complied with the ARRIVE guidelines and were carried out in accordance with the National Institutes of Health guide for the care and use of Laboratory animals, USA (NIH Publications No. 8023, revised 1978).

Routine Blood Test (RBT)
Blood was harvested intracardially after anesthesia with carbon dioxide (CO 2 ) and immediately placed in anti-clot tubes coated with ethylene diamine tetraacetic acid (EDTA). RBT was performed for fresh blood using an automatic blood cell analyzer, according to the instruction provided by the manufacturer (BC-2800vet, Shenzhen, China).

Hg Measurement
Blood was harvested with anti-clot tubes as described above. BM homogenate was harvested as we have previously reported [19,20,38,39]. Briefly, BM from each leg was flushed out with 200 µL of a homogenate buffer containing IGEPAL CA-630, Tris-Cl, NaCl, EDTA, sodium (Na)-orthovanadate, Na fluoride (NaF), and proteinase inhibitors (Sigma-Aldrich, St Louis, MO, USA). A bicinchoninic acid (BCA) kit (Pierce, Rockford, IL, USA) was used to quantify the total protein concentration of the homogenate, and the protocol was according to the manufacturer's instructions. Total Hg was measured using atomic fluorescence spectrometry (Agilent, Santa Clara, CA, USA), as others have previously reported [40,41]. Hg concentration in the blood and BM homogenate was reported as µg/L and µg/kg protein (part per billion, ppb), respectively. Mature erythrocytes were lysed from splenic and BM cells using an ammonia-based lysis buffer. Live cells were incubated with Fc block on ice for 20 min, and surface staining was performed thereafter by incubating cells with fluorescein-conjugated Ab on ice in the dark for 30 min. Nuclear staining for transcription factors was performed after surface staining using a kit from BD Biosciences; the protocol was according to the instructions provided by the manufacturer. The surface markers used for identifying numerous erythroid and megakaryocyte progenitors as well as erythroblasts are summarized in Table 1, according to previously published literature [42][43][44][45][46]. The stained cells were analyzed through a BD LSRFortessa instrument. Table 1. Gating strategies and abbreviations for erythro-and megakaryocyte progenitors and erythroblasts.

Cell Type Abbreviation Gating Strategy
Erythrocyte-megakaryocyte progenitors EMPs

RBC Clearance Assay
Fresh blood harvested from B10.S mice was placed in EDTA-coated anti-clot tubes and then incubated with sulfo-NHS-LC-biotin Na (0.2 mg/ml; MedChemExpress, Monmouth Junction, NJ, USA) at 37 • C for 30 min to label with biotin, according to a previous report [16]. After washing with phosphate buffer saline (PBS), 200 µL of biotin-labeled blood was intravenously injected into B10.S mice that had been treated with 50 µM HgCl 2 or control for 4 w. The recipient mice were sacrificed to analyze the biotin + RBCs in the blood and biotin + MΦ in the spleen three days after the blood transfer; HgCl 2 or control treatment was continuously provided during this period.

CFU Assay
A total of 2 × 10 4 (2e 4 ) BM cells that had been lysed with RBCs were placed in each 35 mm tissue culture Petri-dish containing 1mL of a semi-liquid methocellulose medium (MethoCult, GF M3434, Stem Cell Technologies, Vancouver, BC, Canada). The dishes were incubated at 37 • C and 5% CO 2 with saturated humidity for 6 d, and the number of CFU-Es was quantified thereafter under a microscope, as per a protocol we have previously reported [24,38].

Hematoxylin-Eosin (H&E) Staining
Legs were harvested from mice treated with 50 µM HgCl 2 or the control for 4 w, and the legs were thereafter fixed with 4% paraformaldehyde (PFA) for 24 h. The legs were then decalcified, embeded with paraffin, and sliced. After that, the sections were coated on charged slides and regular H&E staining was performed.
A total of 500 FACS-purified EMPs from the BM of control B10.S mice were placed in each well of 96-well plates with 200 µL of an erythro-megakaryocyte differentiation medium as described above in the presence or absence of HgCl 2 (200 µg/L) and/or EPO (100 ng/mL). After incubation at 37 • C and 5% CO 2 for 16 h, the number of BFU-Es, CFU-Es, and MkPs and the activation of the Jak2/STAT5 signaling pathway were measured.

Real-Time Quantitative Polymerase Chain Reaction (q-PCR)
RNA was extracted from the kidney of B10.S mice as well as FACS-purified EMP from the BM of B10.S mice. The extracted RNA was reversely transcribed into cDNA. The q-PCR with SYBR green was used to quantify the expression of mRNA for cyclin-dependent kinase (CDK)2, CDK4, EPO, and EPOR using kits and an instrument purchased from Thermo Fisher Scientific (Walsham, MA); the expression of β-actin mRNA was used as an internal control. The results reported as the fold change relative to the control were calculated by the ∆∆CT method. The information for forward and reverse primers for target genes was summarized in Table 2.

Enzyme-Linked Immunosorbent Assay (ELISA)
ELISA kits were purchased from Biolegend to quantify the concentration of EPO in serum and BM homogenate, and protocol was according to the manufacturer's instructions.

Statistics
Data were presented as mean or mean ± standard deviation (StD). Data comparisons between the two groups were analyzed with Student's t-tests. Data comparisons among multiple groups were first analyzed with a one-way analysis of variance (ANOVA), and if there were significant differences, data were further analyzed using Student-Newman-Keuls tests to compare between the two groups. The value p < 0.05 was considered as the level of a significant difference.

Hg Concentration in the Blood and BM
To evaluate the level of Hg burden in mice, we measured total Hg concentration in the blood and BM using atomic fluorescence spectrometry. The concentration of Hg in the blood of B10.S mice treated with 10 µM or 50 µM HgCl 2 was 26.5 µg/L (ppb) and 242.5 µg/L (ppb), respectively; the concentration of Hg in the blood of DBA/2 mice treated with 50 µM HgCl 2 was 487.2 µg/L (ppb) ( Table 3). Although Hg was not detectable in the BM of B10.S mice treated with 10 µM HgCl 2 , Hg was detectable in the BM of mice treated with 50 µM HgCl 2 ; the level of Hg in the BM of B10.S mice and DBA/2 mice treated with 50 µM HgCl 2 was 2500.0 µg/kg protein (ppb) and 2200.0 µg/kg protein (ppb), respectively ( Table 3).

HgCl 2 Increases the Number of Mature RBCs and Platelets in the Blood of B10.S Mice
To test the impact of HgCl 2 on erythro-megakaryopoiesis, we performed RBT on B10.S mice after exposure to HgCl 2 . We found that treatment with 50 µM HgCl 2 , but not 10 µM HgCl 2 , increased the number of mature RBCs in the blood of B10.S mice ( Figure 1A). Accordingly, treatment with 50 µM HgCl 2 , but not 10 µM HgCl 2 , also increased the concentration of hemoglobin (HGB) in the blood of B10.S mice ( Figure 1B). HgCl 2 treatment did not influence the mean corpuscular hemoglobin (MCH), the MCH concentration (MCHC), or the mean corpuscular volume (MCV) in the blood of B10.S mice ( Figure 1C-E). Consistently, treatment with 50 µM HgCl 2 , but not 10 µM HgCl 2 , increased the number of platelets (PLTs), but not the mean platelet volume (MPV) in the blood of B10.S mice ( Figure 1F,G). Therefore, treatment with 50 µM HgCl 2 increased the number of mature RBCs and platelets in the blood of B10.S mice.

HgCl 2 Does Not Impact the Clearance of Mature RBCs in B10.S Mice
Mature RBCs are mainly cleared by MΦ in mice [16,47]. As treatment with 50 µM HgCl 2 increased the number of mature RBCs in the blood of B10.S mice, we sought to test whether the increased number of mature RBCs by HgCl 2 exposure was caused by repressed clearance. To test this hypothesis, biotinylated mature RBCs were transferred into 50 µM HgCl 2 -treated B10.S mice or control B10.S mice through intravenous injection, and the clearance of the donated RBCs was measured thereafter (Supplementary Figure S1A). We found that B10.S mice treated with 50 µM HgCl 2 or the control had comparable residual biotinylated RBCs in their blood (Supplementary Figure S1B,C). Consistent with this, B10.S mice treated with 50 µM HgCl 2 or the control had a similar number of splenic MΦ (gated on CD11b + ly6G -F4/80 + ) devouring the donor RBCs (biotin + MΦ) (Supplementary Figure S1D). Thus, we concluded that the increased number of mature RBCs by 50 µM HgCl 2 treatment was not due to a reduced clearance of RBCs in B10.S mice. Each dot represents one mouse, and a total of 7 to 11 mice were used for each group. Asterisk indicates a significant difference compared to the counterpart control group. p < 0.05 was considered as the level of significant difference.

HgCl2 Does not Impact the Clearance of Mature RBCs in B10.S Mice
Mature RBCs are mainly cleared by MΦ in mice [16,47]. As treatment with 50 μM HgCl2 increased the number of mature RBCs in the blood of B10.S mice, we sought to tes whether the increased number of mature RBCs by HgCl2 exposure was caused by re pressed clearance. To test this hypothesis, biotinylated mature RBCs were transferred into Each dot represents one mouse, and a total of 7 to 11 mice were used for each group. Asterisk indicates a significant difference compared to the counterpart control group. p < 0.05 was considered as the level of significant difference.

HgCl 2 Increases the Development of Erythrocytes and Megakaryocytes/Platelets in the BM of B10.S Mice
Erythro-megakaryopoiesis primarily occurs in the BM after birth in mammals [8]. As HgCl 2 did not reduce the clearance of mature RBCs in B10.S mice, we sought to determine whether HgCl 2 promoted the generation of erythrocytes and megakaryocytes in the BM of B10.S mice. To test this, we evaluated the potential of BM cells to generate CFU-Es. We found that treatment with 50 µM HgCl 2 increased the number of CFU-Es derived from the BM cells of B10.S mice ex vivo (Figure 2A), indicating that HgCl 2 drove erythropoiesis in the BM of B10.S mice. During erythro-megakaryopoiesis, EMPs differentiate into both erythrocyte and megakaryocyte lineages. Chronologically, the erythrocyte lineage includes BFU-Es (also known as pre-CFU-Es), CFU-Es, erythroblast 1 (E1), E2, E3, and mature RBCs, and the megakaryocyte lineage includes MkPs, megakaryocytes, and platelets, as schematized in Supplementary Figure S2A. We further measured the number of different types of progenitors for erythrocytes and megakaryocytes in the BM of B10.S mice after exposure to HgCl 2 . We found that treatment with 50 µM HgCl 2 , but not 10 µM HgCl 2 , increased the number of progenitors for erythrocytes, including EMPs, BFU-Es, and CFU-Es in the BM of B10.S mice ( Figure 2B-E). In addition, treatment with 50 µM HgCl 2 , but not 10 µM HgCl 2 , also increased the number of MkPs in the BM of B10.S mice ( Figure 2B,F). Representative histological images for H&E staining of BM sections from B10.S mice treated with 50 μM HgCl2 or the control. Yellow arrows and blue arrows point to megakaryocytes and RBCs, respectively. L: Quantification of MK for each mm 2 of the BM field as indicated in K. Each dot represents one mouse, and a total of 8 to 11 mice were used for each group. Asterisk indicates a significant difference compared to the counterpart control group. p < 0.05 was considered as the level of significant difference. At the late stage of erythropoiesis, CFU-Es differentiate into erythroblasts (here refers to E1, E2, and E3, according to their surface expression of CD71 and Ter119 as summarized in Table 1) and eventually become mature RBCs, as schematized in Supplementary Figure S2A. We measured the number of erythroblasts in the BM of B10.S mice after exposure to HgCl 2 . As expected, treatment with 50 µM HgCl 2 , but not 10 µM HgCl 2 , increased the number of E1, E2, and E3 in the BM of B10.S mice ( Figure 2G-J). In addition, we performed histological staining for the BM of B10.S mice treated with 50 µM HgCl 2 or the control, and we observed morphologically that treatment with 50 µM HgCl 2 increased the number of megakaryocytes as well as mature RBCs in the BM ( Figure 2K,L).

HgCl2 Drives the Proliferation and Differentiation of EMPs in the BM of B10.S Mice
In mouse models, increased erythropoiesis in the BM was associated with suppressed erythropoiesis in the periphery, e.g., the spleen, and vice versa [16]. As treatment with HgCl 2 drove erythro-megakaryopoiesis in the BM, we sought to determine whether HgCl 2 suppressed erythropoiesis in the spleen of B10.S mice. As inferred, in contrast to the increased erythropoiesis in the BM by treatment with 50 µM HgCl 2 , the number of E1, E2 and E3 in the spleen was indeed reduced in B10.S mice after treatment with 50 µM HgCl 2 , but not treatment with 10 µM HgCl 2 (Supplementary Figure S3A-D).
We therefore concluded that treatment with 50 µM HgCl 2 promoted the development of erythrocytes and megakaryocytes/platelets in the BM of B10.S mice.

HgCl 2 Drives the Proliferation and Differentiation of EMPs in the BM of B10.S Mice
EMPs are recognized as the ancestors for both erythrocytes and megakaryocytes/platelets. As HgCl 2 promoted erythro-megakaryopoiesis in the BM of B10.S mice, we sought to determine whether HgCl 2 increased the potential for EMP differentiation in B10.S mice.
To test this, we first evaluated the proliferation of EMPs in the BM of B10.S mice after exposure to HgCl 2 . We found that treatment with 50 µM HgCl 2 , but not 10 µM HgCl 2 , increased the expression of Ki67 in EMPs in the BM of B10.S mice ( Figure 3A,B), indicating that treatment with 50 µM HgCl 2 increased the proliferation of EMPs. In accordance with this, treatment with 50 µM HgCl 2 increased the expression of mRNA for CDK2 and CDK4, which are kinases driving cell cycling [48,49], in FACS-purified EMPs from the BM of B10.S mice ( Figure 3C,D). Moreover, a confocal imaging assay revealed that treatment with 50 µM HgCl 2 increased the expression of Ki67 in Lin -EPOR + cells in the BM of B10.S mice ( Figure 3E), further supporting the inference that 50 µM HgCl 2 drove the proliferation of erythrocyte lineage progenitors in B10.S mice.
Next, we directly tested the potential for EMPs to differentiation using an ex vivo assay. FACS-purified EMPs from the BM of B10.S mice treated with 50 µM HgCl 2 or the control for 4 w were placed in a differentiation medium to evaluate their potential for generating progenitors for erythrocytes and megakaryocytes (Supplementary Figure S4A). As surmised, FACS-purified EMPs from B10.S mice treated with 50 µM HgCl 2 gave rise to more BFU-Es, CFU-Es, and MkPs than FACS-purified EMPs from the control B10.S mice did ( Figure 3F-H), indicating that treatment with 50 µM HgCl 2 increased the potential for EMPs to give rise to more progenies in the BM of B10.S mice.
Taken together, we concluded that treatment with 50 µM HgCl 2 drove the proliferation and differentiation of EMPs in the BM of B10.S mice.
3.6. HgCl 2 Activates the Jak2/STAT5 Signaling Pathway to Promote EMP Differentiation in the BM of B10.S Mice To investigate the signaling pathway involved in erythro-megakaryopoiesis that was impacted by HgCl 2 exposure, we analyzed numerous signal molecules and transcription factors that were known to impact erythropoiesis in the BM of B10.S mice. We measured the expression of PU.1 and IRF1, two nuclear transcription factors that were reported to suppress erythropoiesis in mice [16]. We found that treatment with 50 µM HgCl 2 did not affect the expression of PU.1 or IRF1 in EMPs in the BM of B10.S mice ( Figure 4A,B). The Jak/STAT signaling pathways have been recognized as crucial for erythropoiesis [12]. We therefore measured the activation of Jak/STAT signaling pathways in EMPs in the BM of B10.S mice after exposure to HgCl 2 . We found that treatment with 50 µM HgCl 2 increased the expression of p-Jak2, but not p-Jak1, in the EMPs in the BM of B10.S mice ( Figure 4C). Interestingly, while treatment with 50 µM HgCl 2 did not influence the expression of p-STAT1, treatment with HgCl 2 reduced the expression of p-STAT3 and increased the expression of p-STAT5 in the EMPs in the BM of B10.S mice ( Figure 4D).   Each dot represents one mouse or one sample, and a total of 5 to 12 mice were used for each group. Asterisk indicates a significant difference compared to the counterpart control group. p < 0.05 was considered as the level of significant difference. Activation of the Jak2/STAT5 signaling pathway has been suggested to be crucial in driving erythropoiesis in mice [12,13]. To test whether HgCl 2 promoted erythropoiesis via activation of this pathway, we used an ex vivo assay to test the role of its activation in EMPs in erythro-megakaryopoiesis in the BM of B10.S mice during HgCl 2 exposure (Supplementary Figure S4B). We confirmed that the increased expression of p-STAT5 in EMPs was mediated by the activation of Jak2 in EMPs in the BM of B10.S mice during HgCl 2 exposure, as the increased expression of p-STAT5 in EMPs purified from the BM of HgCl 2 -treated B10.S mice was diminished in the presence of the Jak2 specific inhibitor fedratinib ( Figure 4E). In addition, we found that the increased potential for FACS-purified EMPs from the BM of B10.S mice treated with 50 µM HgCl 2 to give rise to BFU-Es, CFU-Es, and MkPs was recovered in the presence of the Jak2-specific inhibitor fedratinib or the STAT5-specific inhibitor STAT5-IN-1 ex vivo ( Figure 4F-H).

A Direct Action of HgCl2 on EMPs Suppresses Their Differentiation in the BM of B10.S Mice
Taken together, these observations suggested that treatment with 50 µM HgCl 2 drove erythro-megakaryopoiesis via enhanced activation of the Jak2/STAT5 signaling pathway in the BM of B10.S mice.

A Direct Action of HgCl 2 on EMPs Suppresses Their Differentiation in the BM of B10.S Mice
As activation of the Jak2/STAT5 signaling pathway drove the differentiation of EMPs during HgCl 2 exposure, we sought to determine whether the activation of this pathway was caused by a direct action of HgCl 2 on EMPs. To test this, FACS-purified EMPs from the BM of B10.S mice were treated with HgCl 2 in vitro to evaluate the activation of this pathway as well as their differentiation potential (Supplementary Figure S4C). Intriguingly, treatment with HgCl 2 on EMPs in vitro reduced their expression of p-Jak2 and p-STAT5 ( Figure 5A,B), indicating that a direct action of HgCl 2 suppressed the activation of the pathway in EMPs. In line with this, treatment with HgCl 2 on EMPs in vitro suppressed the potential for EMP differentiation, since FACS-purified EMPs from the BM of B10.S mice tended to give rise to fewer BFU-Es, CFU-Es, and MkPs in the presence of HgCl 2 in vitro ( Figure 5C-E). These observations indicated that a direct action of HgCl 2 on EMPs suppressed their potential for differentiation in the BM of B10.S mice.

HgCl 2 Does Not Impact EPO Production, but Increases the Expression of EPOR to Enhance the Jak2/STAT5 Signaling Pathway in EMPs in the BM of B10.S Mice
The Jak2/STAT5 signaling pathway can be activated by a variety of physiological factors during erythropoiesis, of which the best studied is EPO [12,13]. EPO interacts with EPOR to activate this pathway to drive erythro-megakaryopoiesis in mice [12]. Therefore, we hypothesized that HgCl 2 might drive erythro-megakaryopoiesis via increased EPO-EPOR signaling in the BM of B10.S mice. To test this, we quantified the expression of EPO production in B10.S mice during HgCl 2 exposure. Unexpectedly, we found that treatment with 50 µM HgCl 2 did not impact the expression of EPO protein in serum ( Figure 6A) or BM ( Figure 6B) in B10.S mice. Consistent with this, treatment with 50 µM HgCl 2 on B10.S mice did not influence the expression of EPO mRNA in the kidney ( Figure 6C), which was the major origin for EPO in vivo [50]. Thus, HgCl 2 did not impact EPO production in B10.S mice. Next, we measured the expression of EPOR in EMPs in the BM of B10.S mice during HgCl 2 exposure. Interestingly, treatment with 50 µM HgCl 2 increased the surface expression of EPOR in EMPs in the BM of B10.S mice ( Figure 6D). We further confirmed that HgCl 2 increased the expression of EPOR in EMPs by measuring the expression of EPOR mRNA in FACS-purified EMPs from the BM of B10.S mice treated with 50 µM HgCl 2 or the control ( Figure 6E).
To test the function of the EPO-EPOR interaction in the activation of the Jak2/STAT5 signaling pathway, FACS-purified EMPs from the BM of control B10.S mice were treated with EPO protein in vitro, and the pathway was evaluated thereafter. We confirmed that treatment with EPO protein increased the expression of p-Jak2 ( Figure 6F) and p-STAT5 ( Figure 6G) in FACS-purified EMPs in vitro, indicating that EPO activated the Jak2/STAT5 signaling pathway in EMPs in B10.S mice. As HgCl 2 increased the expression of EPOR in EMPs, we sought to determine whether EMPs from the BM of B10.S mice treated with 50 µM HgCl 2 were more sensitive to the EPO protein-induced Jak2/STAT5 signaling pathway. To test this, EMPs from the BM of B10.S mice treated with 50 µM HgCl 2 or the control were purified, and the FACS-purified EMPs were cultured in vitro thereafter to evaluate their responses to the EPO protein and the involvement of the Jak2/STAT5 signaling pathway. As expected, while EPO treatment increased the expression of p-STAT5 by 1.15 folds in purified EMPs from control B10.S mice, EPO treatment increased the expression of p-STAT5 by 1.29 folds in EMPs from B10.S mice treated with 50 µM HgCl 2 ( Figure 6H), indicating that EMPs in the BM of B10.S mice treated with HgCl 2 were more sensitive to the EPO-induced activation of the Jak2/STAT5 signaling pathway. Moreover, we confirmed that Jak2 mediated the increased expression of p-STAT5 in EPO-treated EMPs, as the Jak2-specific inhibitor fedratinib significantly suppressed the expression of p-STAT5 in EPO-treated EMPs from the BM of B10.S mice treated with 50 µM HgCl 2 ( Figure 6H).
While a direct action of HgCl 2 on EMPs suppressed their differentiation, the activation of the Jak2/STAT5 signaling pathway induced by EPO drove EMP differentiation in B10.S mice treated with 50 µM HgCl 2 , indicating that the EPO-induced activation of this pathway was dominant in erythro-megakaryopoiesis during HgCl 2 exposure. To test this, FACS-purified EMPs from the BM of control B10.S mice were cultured in vitro in the presence or absence of HgCl 2 and/or EPO. Although a direct action of HgCl 2 on EMPs decreased the expression of p-Jak2, EPO significantly increased the expression of p-Jak2 in the presence of HgCl 2 ( Figure 6I), indicating that EPO was dominant in activating the Jak2/STAT5 signaling pathway in B10.S mice during HgCl 2 exposure.
Collectively, these experiments suggested that treatment with 50 µM HgCl 2 increased the expression of EPOR in EMPs to enhance the activation of the Jak2/STAT5 signaling pathway in the BM of B10.S mice. Figure 6. HgCl2 does not impact the production of EPO, but increases the expression of EPOR in EMPs in B10.S mice. EPO protein in the serum and BM, EPO mRNA in the kidney, and EPOR in EMPs in the BM in B10.S mice treated with 50 μM HgCl2 or the control for 4 w were measured. FACS-purified EMPs from the BM of B10.S mice were treated with HgCl2 or vehicle or EPO in vitro, and the Jak2/STAT5 signaling pathway and differentiation of EMPs were measured thereaf ter. A: EPO protein (IU/L) in the serum of B10.S mice treated with 50 μM HgCl2 or water for 4 w. B: EPO protein (IU/g protein) in the BM of B10.S mice treated with 50 μM HgCl2 or water for 4 w C: EPO mRNA (fold change) in the kidney of B10.S mice treated with 50 μM HgCl2 or water for 4 w. D: The expression of EPOR (MFI) in EMPs in the BM of B10.S mice treated with 50 μM HgCl2 water for 4 w. E: The expression of EPOR mRNA (fold change) in FACS-purified EMPs from the BM of B10.S mice treated with 50 μM HgCl2 or water for 4 w. F: EMPs were purified from the BM of control B10.S mice, and the FACS-purified EMPs were treated thereafter with EPO or vehicle in vitro for 16 h. After that, the expression of p-Jak2 (MFI) in EMPs was measured. G: EMPs were purified from the BM of control B10.S mice, and the FACS-purified EMPs were treated thereafter with EPO or vehicle in vitro for 16 h. After that, the expression of p-STAT5 (MFI) in EMPs was measured. H: EMPs were purified from the BM of B10.S mice treated with 50μM HgCl2 or water for 4 w, and the EMPs were treated thereafter with EPO or vehicle in the presence or absence of the Jak2 inhibitor fedratinib for 16 h in vitro. After that, the expression of p-STAT5 (MFI) in the EMPs was measured. I: EMPs were purified from the BM of control B10.S mice, and the FACSpurified EMPs were treated thereafter with HgCl2 and/or EPO or vehicle in vitro for 16 h. After that, the expression of p-Jak2 (MFI) in EMPs was measured. Each dot represents one mouse or on sample, and a total of 5 to 12 mice or samples were used for each group. Asterisk indicates a significant difference compared to the counterpart control group. p < 0.05 was considered as the leve of significant difference.

HgCl2
Does not Impact Erythro-Megakaryopoiesis in DBA/2 Mice After that, the expression of p-Jak2 (MFI) in EMPs was measured. (G) EMPs were purified from the BM of control B10.S mice, and the FACS-purified EMPs were treated thereafter with EPO or vehicle in vitro for 16 h. After that, the expression of p-STAT5 (MFI) in EMPs was measured. (H) EMPs were purified from the BM of B10.S mice treated with 50µM HgCl 2 or water for 4 w, and the EMPs were treated thereafter with EPO or vehicle in the presence or absence of the Jak2 inhibitor fedratinib for 16 h in vitro. After that, the expression of p-STAT5 (MFI) in the EMPs was measured. (I) EMPs were purified from the BM of control B10.S mice, and the FACS-purified EMPs were treated thereafter with HgCl 2 and/or EPO or vehicle in vitro for 16 h. After that, the expression of p-Jak2 (MFI) in EMPs was measured. Each dot represents one mouse or one sample, and a total of 5 to 12 mice or samples were used for each group. Asterisk indicates a significant difference compared to the counterpart control group. p < 0.05 was considered as the level of significant difference.

HgCl 2 Does Not Impact Erythro-Megakaryopoiesis in DBA/2 Mice
Finally, we tested whether HgCl 2 increased erythro-megakaryopoiesis in DBA/2 mice, a strain resistant to Hg-induced autoimmunity [32,33]. Different from B10.S mice, treatment with 50 µM HgCl 2 did not change the number of mature RBCs, the concentration of HGB, or the number of platelets in the blood of DBA/2 mice (Figure 7A-C). Accordingly, treatment with 50 µM HgCl 2 did not alter the number of EMPs, BFU-Es, CFU-Es, or MkPs in the BM of DBA/2 mice (Figure 7D-G). Treatment with 50 µM HgCl 2 did not impact the surface expression of EPOR in EMPs in the BM of DBA/2 mice ( Figure 7H). Consistently, treatment with 50 µM HgCl 2 did not impact the expression of p-Jak2 or p-STAT5 in EMPs in the BM of DBA/2 mice ( Figure 7I,J). We therefore concluded that treatment with 50 µM HgCl 2 did not impact erythro-megakaryopoiesis in DBA/2 mice. Accordingly, treatment with 50 μM HgCl2 did not alter the number of EMPs, BFU-Es, CFU-Es, or MkPs in the BM of DBA/2 mice ( Figure 7D-G). Treatment with 50 μM HgCl2 did not impact the surface expression of EPOR in EMPs in the BM of DBA/2 mice ( Figure  7H). Consistently, treatment with 50 μM HgCl2 did not impact the expression of p-Jak2 or p-STAT5 in EMPs in the BM of DBA/2 mice ( Figure 7I-J). We therefore concluded that treatment with 50 μM HgCl2 did not impact erythro-megakaryopoiesis in DBA/2 mice.

Discussion
In the present study, we revealed that HgCl2 promoted erythro-megakaryopoiesis in the BM of B10.S mice, but not in the BM of DBA/2 mice.
Erythrocytes are turned over periodically in vivo, depending on the balance between the clearance of senescent erythrocytes in the periphery and newly generated erythrocytes in the BM [16,18]. HgCl2 increased the number of progenitors for erythrocytes and

Discussion
In the present study, we revealed that HgCl 2 promoted erythro-megakaryopoiesis in the BM of B10.S mice, but not in the BM of DBA/2 mice.
Erythrocytes are turned over periodically in vivo, depending on the balance between the clearance of senescent erythrocytes in the periphery and newly generated erythrocytes in the BM [16,18]. HgCl 2 increased the number of progenitors for erythrocytes and megakaryocytes in the BM but did not impact the clearance of RBCs in B10.S mice, indicating that the increased number of mature RBCs and platelets in the blood was due to increased erythro-megakaryopoiesis in B10.S during HgCl 2 exposure. This was supported by the observation that the number of erythroblasts was also increased in the BM of B10.S mice after treatment with 50 µM HgCl 2 . Notably, while HgCl 2 drove erythropoiesis in the BM, HgCl 2 suppressed erythropoiesis in the spleen of B10.S mice, suggesting that an opposite process of erythropoiesis between the BM and periphery occurred during HgCl 2 exposure, which was in accordance with the observation that mice with increased BM erythropoiesis had reduced peripheral erythropoiesis [16].
In the BM, EMPs are able to differentiate into both erythrocytes and megakaryocytes/platelets [7][8][9]. As EMPs are the ancestors for both erythrocyte and megakaryocyte lineages, the increased number of EMPs was sufficient to result in increased erythromegakaryopoiesis in the BM of B10.S mice after treatment with HgCl 2 . In addition, EMPs from the BM of B10.S mice treated with HgCl 2 had increased capacity to give rise to more BFU-Es, CFU-Es, and MkPs, indicating that HgCl 2 also promoted the potential for EMP differentiation. The increased number of EMPs was associated with the increased proliferation of EMPs in the BM of B10.S mice, suggesting that EMPs might self-renew via increased proliferation to some extent during HgCl 2 exposure.
Intriguingly, a direct action of Hg 2+ on FACS-purified EMPs from the BM of B10.S mice actually suppressed their potential for differentiation in vitro. Thus, Hg 2+ increased erythromegakaryopoiesis likely through an indirect action on EMPs in the BM of B10.S mice. It should be noted that the suppressive effect of Hg 2+ on EMP differentiation was tested using in vitro assays, while the Hg 2+ increase of this differentiation was measured using in vivo assays.
Previously, we reported that B10.S mice treated with 50 µM HgCl 2 increased the expression of IFNγ in the BM [19]. In contrast to increased erythro-megakaryopoiesis, IFNγ is actually known to suppress erythropoiesis, which was partially through induction of the PU.1/IRF8 pathway [16]. Thus, the increased erythro-megakaryopoiesis in the BM of B10.S mice after exposure to HgCl 2 was likely independent of IFNγ signaling.
EPO is a protein primarily produced by the kidney; during hematopoiesis in the BM, EPO plays an essential role in promoting erythro-megakaryopoiesis, thus leading to the generation of more erythrocytes and megakaryocytes/platelets [50,51]. For instance, EPO drives the proliferation and differentiation of erythrocyte progenitors in the BM of mice [12,50]. Treatment with HgCl 2 did not impact EPO production, as the expression of EPO mRNA in the kidney was not changed in B10.S mice treated with HgCl 2 . Moreover, the EPO protein concentration in the BM and serum was not impacted by HgCl 2 exposure in B10.S mice, indicating that the increased amount of EMPs was not caused by increased EPO protein in the BM. Interestingly, HgCl 2 increased the expression of EPOR in EMPs in the BM of B10.S mice, thus making EMPs more sensitive in response to EPO stimulation. Indeed, some physiological factors do promote erythro-megakaryopoiesis via an increase of EPOR signaling on erythrocyte progenitors, e.g., GM-CSF [52,53]. In terms of the mechanism for differential Hg 2+ influences in the regulation of EPOR expression in EMPs in the BM of B10.S mice and DBA/2 mice, although further studies are required to address this mechanism, we propose that Hg 2+ might induce an autoimmune environment to promote certain inflammatory cytokines, which acted on EMPs to increase their expression of EPOR and thereby drove erythro-megakaryopoiesis in the BM of B10.S mice. Hg did not induce autoimmunity in DBA/2 mice, which might account for the observation that Hg did not influence erythro-megakaryopoiesis in DBA/2 mice.
Regarding the observation that Hg 2+ suppressed EMP differentiation in vitro while promoting it in vivo in B10.S mice, one explanation was that, although a direct action of Hg on EMPs had a suppressive effect on it, the EMPs in B10.S mice had increased EPOR expression and were thus more sensitive to EPO-induced signaling. Relative to the suppressive effect of Hg 2+ on EMP differentiation, the enhanced EPO-EPOR signaling in EMPs dominated the increased erythro-megakaryopoiesis in B10.S mice during Hg 2+ exposure.
A variety of signaling pathways have been suggested to impact erythro-megakaryopoiesis, of which the Jak2/STAT5 signaling pathway activated by EPO interaction with EPOR is crucial [12,13,15]. Indeed, EPO activated this pathway in EMPs in the BM of B10.S mice. HgCl 2 promoted the potential for EMP differentiation in the BM of B10.S mice via the induction of this pathway, which was in accordance with the increased expression of EPOR in EMPs. Therefore, it was suggested that treatment with HgCl 2 activated the pathway by increasing the expression of EPOR in EMPs to promote their potential for differentiation in the BM of B10.S mice.
Our study suggested that the Hg 2+ increase in erythro-megakaryopoiesis in the BM was related to Hg-induced autoimmunity. It has been well known that Hg-induced autoimmunity is critically dependent on the genetic background of mice, in particular the H-2 haplotypes [25,30]. HgCl 2 promoted erythro-megakaryopoiesis in B10.S mice, a strain sensitive to Hg-induced autoimmunity, but not in DBA/2 mice, a strain resistant to Hginduced autoimmunity, indicating that the impact of Hg 2+ on the development of erythrocytes and megakaryocytes is also likely related to H-2 haplotypes and an autoimmune environment in the BM. In terms of the mechanism for the differential effects of Hg on erythro-megakaryopoiesis in B10.S mice and DBA/2 mice, although it is still unclear, we propose that certain cytokines that were produced in the Hg-induced autoimmune microenvironment in the BM of B10.S mice, but not in DBA/2 mice, induced the increased expression of EPOR in EMPs, thus causing increased activation of the Jak2/STAT5 signaling pathway to drive the differentiation of EMPs. Notably, a previous study reported that exposure to HgCl 2 for 90 d reduced the number of mature erythrocytes in the blood of rats [6], which in fact does not contradict the observations in our study. Indeed, in addition to genetic background, long-term exposure to Hg was actually able to suppress the autoimmunity of hosts [25].
The aberrant accumulation of erythrocytes and platelets may lead to severe clinical symptoms; the most studied is polycythemia vera, which is characterized by the accumulation of morphologically normal erythrocytes, platelets and myeloid cells [54,55]. Clinically, erythrocytosis is the most common and obvious manifestation for polycythemia vera. Notably, while the expression of EPO is not recognized as a good marker for diagnosis, the erythroid progenitors are more proliferative and are hypersensitive to EPO in polycythemia vera patients [55,56], which is similar to the process that EMPs are more responsive to EPO, are more proliferative, and are able to generate more mature erythrocytes in B10.S mice by HgCl 2 exposure. To date, the etiology of polycythemia vera remains to be defined. To our knowledge, the impacts of environmental factors on polycythemia vera are largely unknown. Thus, as a pilot study, our study may provide a clue for identifying potential environmental risk factors for polycythemia vera in the future.
Notably, it should be mentioned that previous reports by others showed that Hg 2+ directly interacted with mature RBCs to impact their redox state and enzyme activation in both humans and animals [4,5,57], which is not discrepant from our results, as we focused on the quantity of RBC generation. Although we did not perform experiments to test the impacts of HgCl 2 on the function of mature erythrocytes in B10.S mice or DBA/2 mice, we expected the function of mature erythrocytes to be impaired by a direct action of HgCl 2 , probably in both B10.S mice and DBA/2 mice.
Our study may contribute to the current literature on the hematopoietic toxicology of Hg and benefit future risk assessments of Hg in environmental and occupational health.
The limitations of our study include the absence of both an examination of the effect of HgCl 2 on an induced anemia mouse model and of the quantification of Hg concentration in the bone.

Conclusions
In the present study, as summarized in Figure 8, we have found that HgCl 2 promoted erythro-megakaryopoiesis via the increased expression of EPOR and the resultant enhanced activation of the Jak2/STAT5 signaling pathway in the BM of B10.S mice; HgCl 2 did not impact the erythro-megakaryopoiesis in DBA/2 mice. enhanced activation of the Jak2/STAT5 signaling pathway in the BM of B10.S mice; HgCl2 did not impact the erythro-megakaryopoiesis in DBA/2 mice.

Figure 8.
A schematic model for the impact of Hg 2+ on erythro-megakaryopoiesis in mice. In hosts carrying certain H-2 haplotypes sensitive to Hg-induced autoimmunity (e.g., B10.S mice, H-2 s ), Hg 2+ induces an autoimmune environment, which leads to the increased expression of EPOR in EMPs and the resultant hypersensitivity of EMPs in response to EPO in the BM. Consequently, the Jak2/STAT5 signaling pathway is overactivated by EPO interaction with EPOR in EMPs, thus resulting in the increased production of RBCs and platelets in the BM. Hg 2+ does not impact erythromegakaryopoiesis in mice carrying H-2 haplotypes resistant to Hg-induced autoimmunity (e.g., DBA/2 mice, H-2 d ).
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Figure S1: HgCl2 does not impact the clearance of erythrocytes in B10.S mice, Figure S2: A schematic model for erythro-megakaryopoiesis, Figure S3: HgCl2 reduces the number of erythroblasts in the spleen of B10.S mice, Figure S4: Schematic models for in vitro assays.

Conflicts of Interest：
The authors declare no conflict of interest. Figure 8. A schematic model for the impact of Hg 2+ on erythro-megakaryopoiesis in mice. In hosts carrying certain H-2 haplotypes sensitive to Hg-induced autoimmunity (e.g., B10.S mice, H-2 s ), Hg 2+ induces an autoimmune environment, which leads to the increased expression of EPOR in EMPs and the resultant hypersensitivity of EMPs in response to EPO in the BM. Consequently, the Jak2/STAT5 signaling pathway is overactivated by EPO interaction with EPOR in EMPs, thus resulting in the increased production of RBCs and platelets in the BM. Hg 2+ does not impact erythro-megakaryopoiesis in mice carrying H-2 haplotypes resistant to Hg-induced autoimmunity (e.g., DBA/2 mice, H-2 d ).
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/toxics9100252/s1, Figure S1: HgCl 2 does not impact the clearance of erythrocytes in B10.S mice, Figure S2: A schematic model for erythro-megakaryopoiesis, Figure S3: HgCl 2 reduces the number of erythroblasts in the spleen of B10.S mice, Figure S4: Schematic models for in vitro assays.