Histone H3 Serine 10 Phosphorylation Facilitates Endothelial Activation in Diabetic Kidney Disease

Posttranslational histone modifications have emerged as pivotal mediators of diabetes complications by either permitting or preventing cellular injury. Most of the evidence that associates histone modifications with the development and progression of diabetes complications comes from the study of histone (de)methylation or histone (de)acetylation (1). However, several other modifications also can affect histone proteins, including phosphorylation, ubiquitination, O-GlcNAcylation, ADP-ribosylation, and sumoylation (1). Although these alternative modifications have important biological functions, their potential contribution to the development of diabetes complications has largely been overlooked.

Diabetes is an inflammatory disease. Release of inflammatory cytokines by resident cells within the diabetic kidney, for instance, facilitates the recruitment of leukocytes that in turn contribute to the progressive fibrosis that characterizes later-stage nephropathy. Diabetes is also an endothelial disease, and the enhanced expression of endothelial adhesion molecules that facilitate leukocyte recruitment has long been linked to the development of diabetes complications (2). The endothelial expression of cell surface adhesion molecules that facilitate leukocyte recruitment is termed endothelial activation (3). One of the prototypical endothelial adhesion molecules indicative of endothelial activation is vascular cell adhesion protein 1 (also called vascular cell adhesion molecule 1 [VCAM-1]) (3). VCAM-1 primarily functions as the ligand for the β1-integrin subfamily member very late antigen-4 (VLA-4, α4β1) present on the leukocyte plasma membrane, and its upregulation has been reported to occur in the kidneys of both diabetic mice (4) and humans with diabetes (5).

In the current study, we explored the mechanisms that control endothelial activation within the kidney glomerulus in diabetes. We started from the premise that the ordinary functioning of the kidney glomerulus depends on the orchestrated interaction of its cellular constituents. In particular, paracrine communication between podocytes lining the urinary space and endothelial cells lining the glomerular capillary walls maintains the permselective integrity of the glomerular filtration barrier (6). We hypothesized that factors secreted by glomerular podocytes under conditions of high glucose may promote glomerular endothelial cell (GEC) activation characterized by VCAM-1 upregulation. While testing this hypothesis, we discovered a pivotal role for the phosphorylation of histone protein H3 on serine residue 10 (phospho-histone H3Ser10) in facilitating glomerular endothelial activation in diabetic kidney disease.

Experiments were conducted in conditionally immortalized human podocytes (provided by M. Saleem, University of Bristol, Bristol, U.K.) (7) and in primary cultured human renal GECs (hGECs) (ScienCell Research Laboratories, Carlsbad, CA) (8,9). hGECs were cultured under control conditions (5.6 mmol/L glucose) or with the addition of 19.4 mmol/L glucose (final concentration 25 mmol/L, high glucose) or 19.4 mmol/L mannitol for 16 h. To generate human podocyte–conditioned medium, differentiated human podocytes were incubated under control conditions (5.6 mmol/L glucose [hpod_CM]) or high-glucose conditions (25 mmol/L [hpod_HGCM]) for 48 h. The Human Cytokine 41-Plex Discovery Assay was performed by Eve Technologies (Calgary, Alberta, Canada). Neutralizing antibody experiments were performed by incubating hGECs for 16 h in high glucose or hpod_HGCM that had been preincubated with an anti–C-C motif ligand 2 (CCL2) neutralizing antibody (#16-7096-81; Thermo Fisher Scientific, Rockford, IL) at a concentration of 20 μg/mL (10) for 1 h. Recombinant angiopoietin 1, angiopoietin 2, or endothelin 1 were applied to hGECs for 16 h at the following concentrations: angiopoietin 1 100 ng/mL (11) (#923-AN; R&D Systems, Minneapolis, MN), angiopoietin 2 100 ng/mL (#623-AN; R&D Systems), and endothelin 1 10 nmol/L (12) (#1160; Tocris Bioscience, Bristol, U.K.). Recombinant human CCL2 (#RPA087Hu01; Cloud-Clone Corp., Katy, TX) was applied to hGECs at a concentration of 0.5 ng/mL (13,14) for 16 h. For CCR2 antagonism, hGECs were incubated with RS504393 (IC₅₀ <100 nmol/L [15]; Tocris Bioscience) at a concentration of 10 μmol/L (16). For inhibition of p38 mitogen-activated protein kinase (MAPK), hGECs were incubated with SB203580 (IC₅₀ 0.6 μmol/L [17]; Tocris Bioscience) at a concentration of 10 μmol/L (18). For inhibition of mitogen- and stress-activated protein kinase 1/2 (MSK1/2), hGECs were incubated with SB-747651A (IC₅₀ 11 nmol/L [19]; Tocris Bioscience) at a concentration of 5 μmol/L (19).

Immunoblotting was performed with antibodies in the following concentrations: VCAM-1 1:1,000 (#sc-8304; Santa Cruz Biotechnology, Dallas, TX), β-actin 1:10,000 (#A1978; Sigma-Aldrich, Oakville, Ontario, Canada), CCL2 1:1,000 (#NBP1-07035; Novus Biologicals, Littleton, CO), CCR2 1:1,000 (#NBP2-35334; Novus Biologicals), intracellular adhesion molecule 1 1:1,000 (#4915S; Cell Signaling Technology, Danvers, MA), E-selectin 1:1,000 (#ab18981; Abcam, Cambridge, MA), P-selectin 1:1,000 (#ab59738; Abcam), phospho-p38 MAPK threonine 180/tyrosine 182 (Thr180/Tyr182) 1:1,000 (#9216; Cell Signaling Technology), total p38 MAPK 1:1,000 (#9212; Cell Signaling Technology), phospho-histone H3Ser10 1:1,000 (#ab5176; Abcam), and total histone H3 1:1,000 (#9715; Cell Signaling Technology).

In study 1, male C57BL/6 mice (wild type [WT]) (The Jackson Laboratory, Bar Harbor, ME) and Ccr2 knockout mice (CCR2^−/−) (The Jackson Laboratory) age 8 weeks were made diabetic by administration of a daily intraperitoneal injection of streptozotocin (STZ) (55 mg/kg) in 0.1 mol/L citrate buffer (pH 4.5) (or citrate buffer control) after a 4-h fast for 5 consecutive days. Mice were followed for 14 weeks from the first intraperitoneal injection of STZ (WT, n = 9; CCR2^−/−, n = 6; STZ-WT, n = 8; STZ-CCR2^−/−, n = 7). Blood glucose was determined using OneTouch UltraMini (LifeScan Canada, Burnaby, British Columbia, Canada). Glomerular VCAM-1 was determined in frozen kidney sections after immunostaining with a VCAM-1 antibody 1:50 dilution (#550547; BD Biosciences, San Jose, CA) and horseradish peroxidase–conjugated donkey anti-rat IgG (H&L) 1:100 dilution (CLAS10-1115; Cedarlane, Burlington, Ontario, Canada). Glomerular VCAM-1 immunostaining was quantified using ImageScope 11.1 software (Leica Microsystems, Concord, Ontario, Canada) in an average of seven glomerular profiles per mouse kidney and is represented as fold change relative to WT. In study 2, diabetes was induced in male WT and endothelial nitric oxide synthase knockout (eNOS^−/−) mice (The Jackson Laboratory) age 8 weeks by five daily intraperitoneal injections of STZ 55 mg/kg (or citrate buffer) as previously described (8,20). Mice were followed for 3 weeks from the first injection of STZ (WT, n = 12; STZ-WT, n = 14; eNOS^−/−, n = 11; STZ-eNOS^−/−, n = 9). Urine CCL2 excretion was determined by ELISA (#MJE00; R&D Systems) after 24-h metabolic caging. All experimental procedures adhered to the guidelines of the Canadian Council on Animal Care and were approved by the St. Michael’s Hospital Animal Care Committee.

For chromatin immunoprecipitation (ChIP), hGECs were incubated in the presence or absence of 5 μmol/L SB-747651A for 1 h. ChIP was performed using a Magna ChIP kit (EMD Millipore, Etobicoke, Ontario, Canada) with an antibody against phospho-histone H3Ser10 (1:50 dilution) (#ab5176; Abcam) or an equal concentration of normal rabbit IgG (Santa Cruz Biotechnology) as previously described (21,22). Quantitative real-time PCR was performed using primers specific for the human VCAM1 promoter (forward 5′-GAGCTTCAGCAGTGAGAGCA-3′, reverse 5′-CCTTCAAGGGGAAACCCAGG-3′) in hGECs or for the mouse Vcam1 promoter (5′-ATCTCTGTCTTTGCTGTCAC-3′, reverse 5′-CTCTCCTGAAAAGATGATTG-3′) in the kidneys of male WT and CCR2^−/− mice age 22 weeks (n = 4/group).

RNA was isolated from snap-frozen kidney tissue or cell lysates using TRIzol reagent (Thermo Fisher Scientific), and cDNA was reverse transcribed from 1 μg RNA using SuperScript III Reverse Transcriptase (Thermo Fisher Scientific). Primer sequences (Integrated DNA Technologies, Coralville, IA) were as follows: human VCAM1 forward 5′-ATTTCACTCCGCGGTATCTG-3′, reverse 5′-CCAAGGATCACGACCATCTT-3′; human RPL13a forward 5′-AGCTCATGAGGCTACGGAAA-3′, reverse 5′-CTTGCTCCCAGCTTCCTATG-3′; mouse Vcam1 forward 5′-CCCAAGGATCCAGAGATTCA-3′, reverse 5′-TAAGGTGAGGGTGGCATTTC-3′; and mouse β-actin forward 5′-AGAGGGAAATCGTGCGTGAC-3′, reverse 5′-CAATAGTGATGACCTGGCCGT-3′. For determination of micro RNA (miR)-93 levels, RNA isolation was performed using TRIzol reagent; poly(A) tailing was performed using Poly(A) Polymerase, Yeast (#E017; Applied Biological Materials, Richmond, British Columbia, Canada); and cDNA was synthesized using an miRNA cDNA synthesis kit (Applied Biological Materials). Primers for hsa-miR-93 and U6 small nuclear RNA were from Applied Biological Materials (#MPH02022 and #MPH00001). SYBR green–based quantitative RT-PCR was conducted using a ViiA 7 real-time PCR system (Thermo Fisher Scientific), and data analysis was performed using the Applied Biosystems Comparative C_T method.

Archived kidney tissue was examined from eight individuals without diabetes (control subjects) and nine individuals with diabetic kidney disease (22). Tissue had been obtained at the time of nephrectomy for conventional renal carcinoma and was examined from regions of the kidney unaffected by tumor. Immunohistochemistry was performed with an antibody against phospho-histone H3Ser10 (1:200 dilution) (#ab5176; Abcam), and the ratio of positively immunostaining glomerular nuclei to total glomerular nuclei was calculated in 10 glomeruli per kidney section. All histological analyses were performed by an investigator masked to the study groups. The study was approved by the Nova Scotia Health Authority Research Ethics Board (Halifax, Nova Scotia, Canada) and the Research Ethics Board of St. Michael’s Hospital and was conducted in accordance with the Declaration of Helsinki. A waiver of consent was provided by the Nova Scotia Health Authority Research Ethics Board on the basis of impracticability criteria.

In situ hybridization for Vcam1 was performed with RNAScope (Advanced Cell Diagnostics, Hayward, CA) according to the manufacturer’s instructions and using custom software as previously described (23). Briefly, sections of formalin-fixed paraffin-embedded mouse or human kidney tissue were baked for 1 h at 60°C before deparaffinization, dehydration, and air drying. Slides were treated with a peroxidase blocker before boiling in target retrieval solution for 15 min. Protease plus was applied for 30 min at 40°C, and target probes were hybridized for 2 h at 40°C before signal amplification and washing. Hybridization signals were detected using Fast Red, and RNA staining was identified as red puncta on light microscopy. For immunofluorescence microscopy, in situ hybridization was followed by immunostaining for nephrin, CD31, or phospho-histone H3Ser10 using the following antibodies: mouse nephrin 1:200 (#AF3159; R&D Systems), secondary antibody Alexa Fluor 647 donkey anti-goat 1:100 (#A21447; Thermo Fisher Scientific); mouse CD31 1:100 (#ab124432; Abcam), secondary antibody Alexa Fluor 488 donkey anti-rabbit 1:100 (#A21206; Thermo Fisher Scientific); human nephrin 1:100 (#ab136894; Abcam), secondary antibody Alexa Fluor 488 donkey anti-rabbit 1:100 (#A21206; Thermo Fisher Scientific); human CD31 1:100 (#3528; Cell Signaling Technology), secondary antibody Alexa Fluor 647 donkey anti-mouse 1:100 (#A31571; Thermo Fisher Scientific); and human phospho-histone H3Ser10 1:200 (#ab5176; Abcam), secondary antibody Alexa Fluor 488 donkey anti-rabbit 1:100 (#A21206; Thermo Fisher Scientific). DAPI was from Cell Signaling Technology and was used at a concentration of 1:10,000. Slides were viewed using a Zeiss LSM 700 confocal microscope (Carl Zeiss Canada, Toronto, Ontario, Canada) with a ×63 optic.

Data are expressed as mean ± SD. Statistical significance was determined by one-way ANOVA followed by Fisher least significant difference post hoc test for more than two groups and two-tailed Student t test for two-group comparisons. Statistical analyses were performed using GraphPad Prism for Mac OS X (GraphPad Software, La Jolla, CA).

In our first series of experiments, we set out to determine which, if any, podocyte-derived cytokines or chemokines promote the expression of VCAM-1 by GECs under high-glucose conditions. By immunoblotting, we observed that VCAM-1 is expressed by cultured hGECs and that the magnitude of VCAM-1 expression is unaffected by exposure of hGECs to high (25 mmol/L) glucose or mannitol (osmotic control) alone (Fig. 1A). Likewise, when we exposed hGECs to hpod_CM, we similarly observed no change in hGEC VCAM-1 expression (Fig. 1B). In contrast, VCAM-1 expression was significantly increased when hGECs were incubated in hpod_HGCM (Fig. 1B). To determine which cytokines or chemokines are secreted by human podocytes in culture and which of these are altered by high glucose, we performed a multiplex array (Table 1). The most abundant cytokines/chemokines present in the culture media of podocytes under normal conditions were C-X-C motif ligand 1 (CXCL1)/CXCL2/CXCL3 (pan-GRO), interleukin-6 (IL-6), IL-8, CCL2, and vascular endothelial growth factor A. Of all 41 cytokines/chemokines surveyed, CCL2 was the only one to be significantly upregulated with the addition of high glucose to the culture medium (Table 1). Incubating hGECs in hpod_HGCM together with an anti-CCL2 neutralizing antibody resulted in a significant lowering of VCAM-1 mRNA (Fig. 1C) and protein (Fig. 1D) levels, confirming that CCL2 contributes to VCAM-1 upregulation.

We recognized that podocytes secrete other factors that were not included in our cytokine/chemokine multiplex array (e.g., angiopoietin 1, angiopoietin 2, endothelin 1) and that some of these factors have been implicated in podocyte-endothelial communication (6,24). However, when we exposed hGECs to recombinant angiopoietin 1, angiopoietin 2, or endothelin 1, we observed that each recombinant protein actually downregulated VCAM-1 expression (Supplementary Fig. 1). By immunoblotting, we observed that hGECs express the principal receptor for CCL2 CCR2 (as well as CCL2 itself) and that neither CCR2 nor CCL2 were altered in their expression by high glucose in hGECs (Fig. 1E and F). To determine whether CCR2 regulates glomerular VCAM-1 expression in vivo, we examined the kidneys of nondiabetic and diabetic WT and CCR2^−/− mice 14 weeks after the initial induction of diabetes with STZ. Although elevated blood glucose levels were unaffected by CCR2 knockout (Fig. 1G), glomerular VCAM-1 upregulation was significantly attenuated in diabetic CCR2^−/− mice (Fig. 1H). By immunofluorescence microscopy we observed VCAM-1 transcript to be present in CD31+ GECs in the kidneys of both mice and humans (Supplementary Fig. 2). However, GECs were not the only cells to express the adhesion molecule; VCAM-1 mRNA also was detectable in nephrin-positive podocytes (Supplementary Fig. 2).

Having discovered that CCL2 regulates VCAM-1 expression in cultured hGECs and that knockout of the CCL2 receptor CCR2 diminishes glomerular VCAM-1 upregulation in diabetic mice, we next set out to determine the pathways through which CCL2/CCR2 signaling controls VCAM-1. We observed that exposure of hGECs to recombinant CCL2 more than doubled VCAM-1 protein levels and that this increase was negated by antagonism of CCR2 (Fig. 2A). In support of a relative specificity for the regulation in the expression of VCAM-1 by CCL2, we found that the expression of other adhesion molecules (i.e., intracellular adhesion molecule 1, E-selectin, P-selectin) was unaffected by treatment of hGECs with recombinant CCL2 (Supplementary Fig. 3). We recognized the importance of p38 MAPK as a downstream regulator of CCR2 signaling (25) and found that recombinant CCL2 increases hGEC p38 MAPK Thr180/Tyr182 phosphorylation (Fig. 2B), indicative of p38 MAPK activation. As expected, pretreatment of hGECs with the CCR2 antagonist RS504393 negated the increase in p38 MAPK phosphorylation induced by CCL2 (Fig. 2C). We observed that the p38 MAPK inhibitor SB203580 (17) prevented hGEC VCAM-1 upregulation induced by CCL2 (Fig. 2D), confirming that p38 MAPK activation is required for hGEC VCAM-1 expression. Next, we considered how p38 MAPK induces VCAM-1 upregulation. Two downstream kinases that are activated by p38 MAPK are MSK1 and MSK2. We preincubated cells with the MSK1/2 inhibitor SB-747651A (19) and observed that like p38 MAPK inhibition, MSK1/2 inhibition prevented the upregulation in hGEC VCAM-1 induced by CCL2 (Fig. 2E).

MSK1/2 is known to regulate gene expression by directly phosphorylating histone protein H3, including phospho-histone H3Ser10, which is a mark of active gene transcription. Accordingly, we next probed to see whether phospho-histone H3Ser10 levels are altered by CCL2 in hGECs. Aligned with this hypothesis, CCL2 induced an increase in H3Ser10 phosphorylation levels in hGECs, and this effect was negated by antagonism of CCR2 (Fig. 3A) or inhibition of either p38 MAPK (Fig. 3B) or MSK1/2 (Fig. 3C). Furthermore, by using ChIP, we observed enrichment of H3Ser10 phosphorylation at the promoter region of VCAM1 in hGECs, whereas this enrichment was diminished (although not negated) by MSK1/2 inhibition (Fig. 3D). To determine whether H3Ser10 phosphorylation is also enriched at the Vcam1 promoter in vivo and whether this is affected by upstream CCR2-regulated signaling, we performed ChIP experiments in the kidneys of WT and CCR2^−/− mice. Although H3Ser10 phosphorylation was enriched at the Vcam1 promoter in WT mouse kidneys, enrichment was approximately two-thirds lower in the kidneys of CCR2^−/− mice (Fig. 3E). miR-93 recently has been implicated in regulating podocyte MSK-mediated H3Ser10 phosphorylation in diabetic kidney disease (26), but we saw no change in miR-93 levels in hGECs after CCL2 treatment (Supplementary Fig. 4).

Having identified a role for H3Ser10 phosphorylation in facilitating CCL2/CCR2-mediated VCAM-1 upregulation, we set out to determine whether H3Ser10 phosphorylation levels are altered in diabetic kidney disease. For these experiments, we chose to study both a mouse model of diabetic kidney disease that is characterized by endothelial dysfunction and podocytopathy (STZ-diabetic eNOS^−/− mice) (9) and the glomeruli of humans with diabetic kidney disease (22). Compared with nondiabetic mice, 3 weeks after the first intraperitoneal STZ injection, STZ-diabetic eNOS^−/− mice exhibited renal enlargement and heavy albuminuria (Supplementary Table 1) that were accompanied by increased urinary excretion of CCL2 (Fig. 4A), increased renal H3Ser10 phosphorylation (Fig. 4B), and increased renal VCAM-1 mRNA (Fig. 4C and D) and protein (Fig. 4E) levels.

To explore the relationship between H3Ser10 phosphorylation and VCAM-1 expression in human diabetic kidney disease, we studied kidney tissue from individuals with histopathologically confirmed diabetic glomerulosclerosis and compared it with kidney tissue from individuals without diabetes. The clinical characteristics of the individuals from whom kidney tissue was obtained have been reported before (22). In brief, we examined kidney tissue from eight control subjects (five male, three female; age, 69 ± 11 years; serum creatinine, 85 ± 11 μmol/L; estimated glomerular filtration rate, 74 ± 11 mL/min/1.73 m²) and nine individuals with diabetic kidney disease (five male, four female; age, 67 ± 10 years; serum creatinine, 107 ± 39 μmol/L; estimated glomerular filtration rate, 61 ± 25 mL/min/1.73 m²). Five of the individuals with diabetic kidney disease had stage 3 chronic kidney disease or worse. We observed that in the kidney sections from humans with diabetic glomerulosclerosis, there was an approximately threefold increase in the proportion of glomerular nuclei positively immunostaining for phospho-histone H3Ser10 (Fig. 5A), including H3Ser10 phosphorylation in VCAM1-expressing glomerular cells (Fig. 5B).

Every cell that lies within the kidney glomerulus is affected by diabetes, and every cell that lies within the kidney glomerulus is affected by the actions of its neighbors. In the current study, we unearthed a signaling cascade that regulates expression of the adhesion molecule VCAM-1 by GECs. Specifically, ligand binding by the receptor CCR2 expressed by GECs induces VCAM-1 upregulation through a pathway that is regulated by the MSK1/2-dependent phosphorylation of H3Ser10. Heightened phospho-histone H3Ser10 levels in experimental and human diabetic kidney disease and recent improvements in MSK1/2 inhibitor specificity (19) should galvanize efforts to explore the modulation of histone phosphorylation as a means of attenuating kidney disease in diabetes.

As a marker of endothelial activation, we focused on the regulation of expression of VCAM-1, an immunoglobulin superfamily member that is expressed on the surface of endothelial cells in response to proinflammatory cytokines. VCAM-1 promotes the firm adhesion and spreading of leukocytes on the endothelium, enabling their transmigration across the endothelial barrier. Several studies have linked circulating VCAM-1 levels to diabetic kidney disease or mortality risk (27–29), and likewise, a number of reports have described an association between renal expression or urinary excretion of the CCR2 ligand CCL2 and the extent of diabetic kidney disease (30–32). However, even though CCL2, CCR2, and VCAM-1 often are considered together in the same context of inflammation, this description that CCL2/CCR2 binding can directly trigger glomerular endothelial VCAM-1 upregulation is the first to our knowledge.

CCL2 (also called MCP-1) is a member of the CC chemokine family. Although CCL2 is best known for its function as the ligand for the receptor CCR2, which is expressed on the surface of monocytes and macrophages, the actions of CCL2 and CCR2 are not limited to inflammatory cells, and the relationship between CCL2 and CCR2 is not monogamous. For instance, podocytes themselves are known to express both CCL2 and CCR2 (33,34), and we observed that hGECs also express both ligand and receptor. In terms of ligand-receptor specificity, CCL2 also binds to CCR4 (35), and CCR2 also may be bound by CCL7, CCL8, CCL12 (mouse only), CCL13, and CCL16 (human only) (36). In a similar nonreductionist context, although we focused on VCAM-1 upregulation as a marker of endothelial activation (3), it is noteworthy that other glomerular cells, including both podocytes (37) and mesangial cells (38), also are capable of expressing VCAM-1. In the current study, we observed that 1) VCAM-1 levels were increased in hGECs incubated in culture medium that had been conditioned by podocytes exposed to high glucose, 2) secretion of CCL2 by podocytes into the culture medium was upregulated by high glucose, 3) an anti-CCL2–neutralizing antibody diminished hGEC VCAM-1 expression, and 4) recombinant CCL2 induced hGEC VCAM-1 upregulation in a CCR2-dependent manner. Thus, whereas CCL2/CCR2 signaling to hGECs could be paracrine in origin, autocrine in origin, or both and the relationship between CCL2 and CCR2 is not exclusive, the evidence herein demonstrates that CCR2 signaling regulates glomerular VCAM-1 expression, including CCL2-induced VCAM-1 upregulation by hGECs.

In unraveling the cascade by which signaling through CCR2 induces VCAM-1 upregulation in hGECs, we discovered important roles for p38 MAPK and MSK1/2 and an associated enrichment of the phospho-histone H3Ser10 mark at the VCAM1 promoter. Chromatin modifications, such as phospho-histone H3Ser10 rarely control gene activation or repression in isolation. Rather, an interplay exists whereby histone marks function alongside other epigenetic regulators, other histone marks, and canonical transcription factors to coordinate gene expression in an integrated manner (39). For instance, H3Ser10 phosphorylation (like histone acetylation) can facilitate gene activation by affecting the electrostatic charge relationship between histone proteins and DNA, associating with open chromatin during interphase, and allowing access to DNA by the transcriptional machinery. Separately, H3Ser10 phosphorylation also may promote gene transcription by virtue of its proximity to other histone marks. For instance, the histone acetyltransferase Gcn5 can acetylate lysine residue 14 (K14) on histone H3 more effectively when H3Ser10 is phosphorylated, with H3K14ac being found at actively transcribed promoters (40). A number of kinases have been reported to phosphorylate histone H3 on serine residue 10, but the best characterized of these is MSK1/2, which is a substrate for p38 MAPK (41), itself activated by CCR2 (25). The observation that MSK1/2 inhibition reduces but does not negate H3Ser10 phosphorylation at the VCAM1 promoter may suggest an additional role for other kinases (e.g., calcium/calmodulin-dependent protein kinase II [42,43]). Similarly, H3Ser10 is not the only substrate of MSK1/2, with the transcription factor CREB also being phosphorylated by the kinase (44). Indeed, VCAM-1 expression induced by tumor necrosis factor-α in endothelial cells has been reported to involve p38 MAPK-mediated CREB phosphorylation (45). Moreover, transcription factor binding at specific sites on the genome itself depends on histone modifications and is both histone modification and protein family specific (46). Thus, aligned with the current perspective of coordinated interplay between epigenetic modifications and canonical transcription factors, CCR2-regulated VCAM-1 expression by the glomerular endothelium likely involves both effects that are mediated through histone protein posttranslational modifications and effects that are regulated by associated transcription factor responses. Nonetheless, a role for H3Ser10 phosphorylation in regulating endothelial activation in diabetes is supported by increased H3Ser10 phosphorylation at the VCAM1 promoter and a reduction with MSK1/2 inhibition that coincides with a decrease in VCAM-1 protein levels in hGECs. H3Ser10 phosphorylation could regulate VCAM-1 expression by directly affecting CCR2-mediated signaling, or it could have parallel effects, facilitating canonical transcription factor–mediated gene transcription (Fig. 6).

Consistent with its contributory role to the development of diabetic kidney disease, we observed increased levels of H3Ser10 phosphorylation both in the kidneys of STZ-eNOS^−/− mice (a model considered to mimic human disease more closely [47]) and in the glomeruli of humans with diabetic kidney disease. We studied STZ-eNOS^−/− mice soon after the induction of diabetes because we previously found that the heavy albuminuria that these mice develop coincides with the onset of hyperglycemia (9). Even at this early time point, we observed increased urine CCL2 excretion in STZ-eNOS^−/− mice that coincided with increases in both renal H3Ser10 phosphorylation and VCAM-1 expression. Of note, however, is that distinct from its role in transcriptional activation, H3Ser10 phosphorylation also marks highly condensed chromatin during mitosis. Thus, whether the increased kidney cell H3Ser10 phosphorylation in diabetes is indicative of mitotic cell division, a generalized shift in the epigenomic landscape that supports transcriptional activation, or a combination of the two is unclear. Also of note, the findings are aligned with a recent study that reported increased global H3Ser10 phosphorylation levels in podocytes exposed to high glucose and in glomerular cells of type 2 diabetic db/db mice (26).

As already highlighted, our study has limitations. Of note, paracrine podocyte-derived CCL2 may not be the only activator of hGEC CCR2 signaling, H3Ser10 phosphorylation by MSK1/2 may not be the sole means through which CCR2 signaling regulates VCAM-1 expression, and increases in kidney cell H3Ser10 phosphorylation in experimental and human diabetic kidney disease will not solely reflect the changes occurring at the VCAM-1 promoter. On the basis of our initial experiments, despite containing appreciable levels of CCL2, media of podocytes grown under normal glucose conditions did not induce hGEC VCAM-1 upregulation, suggesting the presence of other nonquantified factors in the culture media. For instance, we observed a reduction in VCAM-1 expression by exposure of hGECs to recombinant angiopoietin 1, which is constitutively expressed by podocytes and downregulated in diabetes (48). Thus, overall effects of podocyte-conditioned medium likely reflect the overall balance of its constituents, which are not limited to proteins but may also involve bioactive lipids, nucleic acids, and microparticles (6).

Despite their limitations, the current experiments provide important new insights. First, they demonstrate how proinflammatory cytokines/chemokines can have direct effects on the glomerular endothelium, which could have implications for the interpretation of the mechanism of action of anti-inflammatory therapies recently trialed in the treatment of diabetic kidney disease (49,50). Second, they highlight the emerging role for H3Ser10 phosphorylation and its regulatory kinases MSK1/2 in facilitating the activation of genes important to the development of kidney disease in diabetes, specifically the expression of VCAM-1 by GECs. Moreover, the elucidation of these actions in cultured cells of human origin and the observation of heightened glomerular cell H3Ser10 phosphorylation in human diabetic kidney disease lend weight to the significance of the findings in an era when the value of rodent models is under scrutiny.

In summary, ligand binding by CCR2 initiates an intracellular signaling cascade in GECs that involves the p38 MAPK, MSK1/2-regulated phosphorylation of H3Ser10, facilitating the expression of the inducible proinflammatory adhesion molecule VCAM-1, a marker of endothelial activation. Histone protein phosphorylation should be placed alongside previously better-studied histone modifications when considering potentially druggable candidates suitable for targeted intervention in diabetic kidney disease.

Acknowledgments. The authors thank Dr. Hana Klassen Vakili (currently affiliated with the Department of Pathology, University of Texas Southwestern Medical Center, TX) for technical assistance.

Funding. T.A.A. is supported by a King Abdullah Foreign Scholarship. S.N.B. was supported by the Kidney Foundation of Canada through a Keenan Family Foundation Kidney Research Scientist Core Education and National Training Program Post-Doctoral fellowship, a Heart and Stroke/Richard Lewar Center of Excellence fellowship, and a Banting & Best Diabetes Centre, University of Toronto, postdoctoral fellowship. M.J.H. is a recipient of a St. Michael’s Hospital scholarship from the Research Training Centre and a Banting & Best Diabetes Centre-Novo Nordisk studentship. S.M. and A.A. were supported by a Diabetes Canada postdoctoral fellowship and diabetes investigator award, respectively. These studies were supported by a Canadian Institutes of Health Research operating grant (MOP-133631) and in part by a Heart and Stoke Foundation of Canada grant-in-aid (G-17-0018231) to A.A.

Duality of Interest. A.A. has received research support through his institution from AstraZeneca and Boehringer Ingelheim and is listed as an inventor on an unrelated patent application by Boehringer Ingelheim. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. T.A.A. designed and performed the experiments, analyzed the data, and wrote the manuscript. S.N.B. and S.M. designed and performed the experiments and analyzed the data. M.J.H., V.G.Y., Y.L., B.B.B., and S.L.A. performed the experiments. L.G. and F.S.S. contributed to the human data. A.A. designed the experiments, supervised the study, and wrote the manuscript. A.A. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this work were presented at the World Diabetes Congress, Vancouver, British Columbia, Canada, 30 November–4 December 2015, and the 78th Scientific Sessions of the American Diabetes Association, Orlando, FL, 7–11 June 2018.