The epigenome as a putative target for skin repair: the HDAC inhibitor Trichostatin A modulates myeloid progenitor plasticity and behavior and improves wound healing

C57BL/6 male mice, 2–4 months old, were euthanized and the femurs were dissected, the marrow cells were obtained by flushing and seeded in α-MEM medium supplemented with 5% fetal bovine serum (FBS). The cells were kept for 1 h in a humid incubator with 5% CO₂ at 37 °C to separate the non-adherent populations. Supernatant cells were then plated at the density of 10⁶ cells per well in α-MEM medium containing 100U/ml penicillin, 100 μg/ml streptomycin, supplemented with 5% FBS and subdivided into the following experimental groups: GM-CSF, GM-CSF + iHDAC (Trichostatin A at 10 nM). Cells were grown in 5% CO₂ humidified incubator at 37 °C for 48 h, 72 h, 7 or 13 days. All the procedures were approved by the local ethical committee under the CEUA protocol 077/18.

For the phenotypic analysis, the supernatant cells were harvested after 24 h, 48 h, 72 h or 7 days of culture, washed twice with Phosphate Buffer Solution (PBS), pH 7.2, containing 3% FBS, quantified and their concentration adjusted to 10⁶ cells/well. We have used only the cells found in the supernatant because upon GM-CSF addition to the culture medium, the myeloid cells adhere to the plastic substrate which may lead to interferences in cell counting. To saturate Fc receptors, all cells were incubated with the Fc blocker antibody (clone 2.4G2 obtained from the Rio de Janeiro Cell Bank, Inmetro, Rio de Janeiro, Brazil) for 10 min before adding specific monoclonal antibodies (mAbs). For cell surface markers we used a mix of the following mAbs: FITC-labeled CD11b and PE-labeled LyC6 PercP-labeled Gr1 and PE-labeled CD115 (all from BD Bioscience, San Jose, CA), at 4 °C for 15 min.

For intracellular staining the cells previously stained with mAbs FITC-labeled CD11b and PE-labeled LyC6 were fixed with 2% paraformaldehyde for 20 min, permeabilized with PBS + 5% BSA (BSA: Bovine serum albumin) + 0.2% Saponin for 15 min and stained with primary polyclonal antibodies anti-H4ac (antibody anti-acetylation of histone four from Millipore) diluted in PBS + 0.2% Saponin solution for 1 h. Cells were washed twice and incubated with the secondary antibodies Alexa conjugated with fluorochrome 647 and Alexa conjugated with fluorochrome 488 (Thermo Fisher Scientific) diluted in PBS + 0.2% Saponin for 1 h. Cells were washed for further acquisition.

Cells were acquired on FACSCanto flow cytometry (BD Biosciences) and analyzed using FACSDiva software (Version 8.0) or FlowingSoftware (version 2.5.1). The gates were set as shown in Additional file 1: Figure S1.

After 13 days in culture, the adherents cells on coverslips were washed in PBS and fixed in 4% paraformaldehyde for 10 min at room temperature (RT), washed with PBS and permeabilized with PBS + 0.2% Triton X-100 for 5 min. The samples were washed with PBS, Fc receptor blockade was performed for 10 min with 200 μl of 24G2 cell line supernatant and the cells were washed with PBS and PBS + 5% BSA for 1 h. Cells were washed with PBS and incubated with 50 μl of the primary antibodies in PBS/BSA at 4 °C overnight. F4/80 (1:100) and H4ac (1:200) (from Millipore) antibodies were used. Upon overnight incubation the cells were washed with PBS and incubated with 50 μl of the secondary antibodies anti-goat Alexa-488 (1:800) and anti-rabbit Alexa-546 (1:800) for 2 h at RT. The nuclei were stained using DAPI. The coverslips were mounted on slides with ProLong Diamond Antifade Mountant (Thermo Fisher Scientific), analyzed by confocal microscope (LEICA TCS SP5) and the images were processed with ImageJ software.

For iNOS protein quantification, cells were maintained in culture for 18 days. For Arginase-1 protein quantification, wound skin were collected at 7 days after wounding. The protein extraction was performed with RIPA buffer plus protease inhibitor and denaturated in sample buffer. The proteins were resolved by SDS-PAGE (4% stacking and 10% running) at room temperature. Protein transfer was performed for 2 h onto a PVDF membrane. Membrane blocking was performed with 5% milk powder in 0.1% PBT at room temperature and further incubated with primary antibodie anti-iNOS 1:200 and anti-tubulin 1:2000 at 4 °C overnight. Membranes were washed with 0.1% PBT, incubated with the HRP-conjugated secondary antibodies (anti-goat, anti-rabbit and anti-mouse) for two hours at room temperature under stirring and then washed again three times with 0.1% PBT. Reaction development was monitored using ImageQuant LAS 500 (G.E. Healthcare) using SuperSignal developer West Pico Chemiluminescent Substrate (Thermo Scientific). The images were processed in ImageJ software and analyzed in GraphPad Prism 5.0 software.

Statistical analysis were performed using GraphPad Prism 5.0 software and using Student’s T test, One-way ANOVA or Two-way ANOVA with Bonferroni correction depending on the experiment. Differences were considered statistically significant at *P < 0.05, **P < 0.01, ***P < 0.001. The results are presented as the means ± SD.

In our previous work we have demonstrated, for the first time, that the inhibition of HDAC activity shifts the myeloid differentiation dynamics leading to the expansion of the CD11b^low/Gr1^low subset [12]. Upon 15 days in culture, we observed the presence of very elongated macrophages harboring a mixed M1/M2 phenotype. In the present work we have used the surface marker Ly6C to better understand the relevance of HDAC activity to the generation of classic/inflammatory CD11b⁺/Ly6C^high or non-classical/resolving CD11b⁺/Ly6C^low monocyte subsets [13] in the absence of any pro or anti-inflammatory stimuli. We hypothesized that HDAC activity blockade would favor the expansion of Ly6C^low progenitors that would give rise to the very elongated macrophages upon 7 days in culture. For this, non-adherent bone marrow cells were cultured in the presence of GM-CSF to induce myeloid differentiation of the hematopoietic progenitors. In addition, the cells were exposed to 10 nM TSA (iHDAC) and cultured up to 7 days. Firstly, we used the surface markers Gr1 (recognizes both Ly6G, granulocytes, and Ly6C, granulocytes and monocytes) and CD115 (M-CSF receptor and identifies monocytes and macrophages) to better understand whether HDAC activity is necessary for monocytic or granulocytic differentiation. Upon 72 h in culture, we observed that HDAC activity blockade led to an increase in percentage of cells in the P15 regions (Gr1^low/CD115⁻), P21 (Gr1^low/CD115^low), and P23 (Gr1^low/CD115^high) when compared to the GM-CSF group (Additional file 1: Figure S1). These results demonstrate that iHDAC target population is not the population that display higher levels Gr1.

Therefore, we ruled out the hypothesis that activity HDAC is required for granulocyte differentiation and hypothesized that the main target for iHDAC would be the monocytic population. Because myeloid progenitors, monocytes and macrophages display different levels of the surface markers CD11b and Ly6C down the differentiation pathway, we used these two markers to track BMMP differentiation in the absence of HDAC activity. Upon 24 h in culture, GM-CSF promoted the differentiation of cells of the myeloid lineage of CD11b⁺ and Ly6C⁺ (Fig. 1a) and iHDAC did not prevent the myeloid differentiation induced by GM-CSF. Interestingly, iHDAC in the presence of GM-CSF promoted an increase in the population in the P4 region (CD11b^int/Ly6C^low; Fig. 1b, c). In the GM-CSF group, however, we found the cells homogeneously distributed between the P4 (CD11b^int/Ly6C^low) and P3 (CD11b^high/Ly6C^low) regions (Fig. 1a, c). After 7 days of culture, we observed that while GM-CSF treated myeloid cells were homogeneously distributed among the P4 (CD11b^high/Ly6C⁻), P5 (CD11b^low/Ly6C⁻), P6 (CD11b^high/Ly6C^low) and P7 (CD11b^low/Ly6C^low) (Fig. 1d, f) regions, iHDAC treated cells were mostly found in the P10 region (CD11b^low/Ly6C^low) (Fig. 1e, f). In fact, the fluorescence intensity of the CD11b and Ly6C markers in the iHDAC group was low along the 7 days of analysis when compared to GM-CSF alone (Fig. 1g–l). These results show that HDAC activity is required for myeloid differentiation, and in synergy with GM-CSF, drives the differentiation of inflammatory monocytes. The iHDAC, in contrast, favors the expansion of myeloid progenitor of the CD11b^low/Ly6C^low subset cells for longer in culture.

We then asked which bone marrow myeloid subset would be the target of HDAC activity. For this BMMP were cultured in the presence of GM-CSF and/or iHDAC for 48 h and chromatin acetylation levels were analyzed by immunostaining. In fact, cells exposed to iHDAC were immuno positive for acetylated histone H4 (H4Ac) when compared to the control group (Fig. 2a–d). To characterize which subpopulation of BMMP would be the specific target for iHDAC, we analyzed the levels of H4Ac along with the surface markers CD11b and Ly6C after 48 h in culture. In fact, it was possible to observe that H4Ac levels were increased in the CD11b^low/Ly6C^low subset only in the iHDAC treated group (Fig. 2e–g). In accordance, MFI showed that CD11b^low/Ly6C^low subset displayed higher levels of acetylated H4 histone in the GM-CSF + iHDAC group (Fig. 2h, i, l). In contrast, the acetylation levels of H4 histone were not different in the CD11b^high/Ly6C^low subset between control and iHDAC treated groups (Fig. 2j, k, l). These data show that of all bone marrow subpopulations, the CD11b^low/Ly6C^low subset is the specific target of iHDAC, which in turn, leads to chromatin changes due to histone H4 hyperacetylation.

In our last work we demonstrated that the blockade of HDAC activity during 15 consecutive days led to the generation of elongated macrophages with mixed phenotype M1/M2 [12]. Then, we hypothesize that HDAC activity would be a key player for the maintenance of cellular plasticity over the time. In fact, regardless the literature has successfully characterized the existence of M1 (pro-inflammatory) and M2 (repairing/resolving) macrophages, it is still not clear whether these functional subsets arise from the same progenitor cell or from different ones. Thus, we hypothesized that HDAC activity could be, at least in part, a key player to proper confer plasticity to myeloid progenitors committed to macrophage differentiation. If in fact, if this hypothesis is corroborated, the use of iHDACs to modulate the behavior and function of macrophages in vivo might become an important strategy for tissue repair in translational studies. To test this hypothesis, we cultured BMMP under two different conditions: in the first BMMP cells were cultured in the presence of GM-CSF alone for 13 days (Fig. 3A’). At this point the macrophages presented a rounded morphology and were H4Ac negative (Fig. 3A). After this time, iHDAC was added to the medium and cells were analyzed 3 days later (Fig. 3B’ day 16). At this point, the macrophages started displaying an elongated morphology that was correlated to H4Ac positive staining (Fig. 3B).

In the second condition, BMMP cells were cultured in the presence of GM-CSF + iHDAC for 13 days (Fig. 3C’). At this point, macrophages displayed a very elongated morphology and were positive for H4Ac (Fig. 3C). After this time, iHDAC was removed from the culture medium and cells were analyzed 3 days later (Fig. 3D’ day 16). At this point, macrophages lost their elongated morphology as well as the H4Ac staining (Fig. 3D). Cells grown in the presence of GM-CSF alone for 13 days presented rounded morphology (Fig. 3A, A’) and after iHDAC addition to the culture medium, they presented a higher elongation index (Fig. 3E, purple bar) but maintained the same levels of iNOS observed in cells grown in the presence of GM-CSF alone (Fig. 3F, purple bar). However, BMMP grown in the presence of GM-CSF + iHDAC for 13 days showed an extremely elongated morphology (Fig. 3C, C’; E, orange bar) but maintained the same levels of iNOS when compared to cells grown only in the presence of GM-CSF (Fig. 3E, blue bar) or GM-CSF plus iHDAC (Fig. 3E, purple bar). In clear contrast, BMMP cultured in the presence of GM-CSF + iHDAC for 13 days and in the presence of GM-CSF alone for further 3 days, presented lower elongation index, when compared to cells grown in the presence of iHDAC for 13 days, due to a shortening in cell length (Fig. 3E, green bar). More interestingly, these cells up regulated the levels of iNOS after iHDAC was removed from the culture medium (Fig. 3F, green bar). This set of experiments shed light on two very important aspects of myeloid cells development and plasticity: the first one is that we can dissect the HDAC activity necessity along the time. At the beginning of BMMP differentiation, HDAC activity is necessary for BMMP progenitor differentiation towards pro-inflammatory subsets (Cd11b⁺/Ly6C^high) and iHDAC leads to the expansion of CD11b^low/Ly6C^low. The second information is that HDAC activity is also necessary for macrophage plasticity and behavior upon differentiation has occurred. In fact, the elongated morphology of iHDAC-treated derived macrophages was tight correlated to macrophage function because more elongated macrophages express as much iNOS as GM-CSF-induced macrophages. However, iHDAC treated macrophages may become pro-inflammatory iNOS producing cells whether cell shape shortening was evoked. So, we hypothesized that this amazing property would connect cell shape control to function and plasticity, turning the HDAC activity into a putative target for tissue repair translational studies.

In accordance to our data, we next speculated whether iHDAC would modulate BMMP in vivo upon acute skin injury. The relevance of monocyte subsets dynamics into the skin wound as well as the relevance of macrophage role during wound healing are very well established. For this reason, we decided to test whether iHDAC topic delivery on skin wounds would recapitulate the BMMP behavior observed in vitro. If so, iHDAC might favor the expansion of CD11b^low/Ly6C^low subset and also modulate macrophage cells shape towards the elongated morphology in vivo, and as consequence, improve wound healing. To assess whether inhibition of HDAC activity could modulate the phenotypic and morphological plasticity of BMMP in vivo, we monitored monocyte subsets dynamics from day 1 to day 7 post wound lesion. Upon wound induction, iHDAC was topically added to wounds once a day during 7 consecutive days and monocytes subsets dynamics was monitored at days 1, 3, 5 and 7 post lesion by flow cytometry. In fact, we did not detect any consistent difference in monocytes subsets from bone marrow, peripheral blood and wound infiltration between control and iHDAC along the time (Additional file 2: Figure S2; Additional file 3: Figure S3; Additional file 4: Figure S4, respectively). We conclude that it is unlike that topic treatment with iHDAC would be modulating the recruitment of central or peripheral monocytes into the wound and that iHDAC is good candidate for topic use as its effects were restricted to the area of delivery. For this reason, we went to characterize the BMMP subsets in the wound. Regardless we have failed to observe a significant difference in the recruitment of the CD11b⁺/Ly6C⁺ subset from peripheral blood to the wound, between control and iHDAC group (Additional file 5: Figure S5), we noticed that the CD11b⁺/Ly6C^low subset was significantly increased in the iHDAC treated wounds (Fig. 4). Then, we asked whether such enhancement in the CD11b⁺/Ly6C^low subset into the wounds would correlate with an improvement in the wound closure. For this we monitored the wound healing dynamics up to 7 days. In fact, after 5 days of topic treatment, the iHDAC group presented a higher percentage of healing when compared to the group of animals treated with solvent solution (Fig. 5A–E, K).

In accordance, histopathological sections of control and iHDAC treated wounds uncovered remarkable qualitative and quantitative differences in cutaneous lesions. While control and iHDAC treated wounds were positive for iNOS, Arginase-1 and H4ac (Fig. 5A–F), we observed the presence of very elongated F4/80⁺/H4Ac⁺ macrophages only in iHDAC treated wounds (Fig. 5J, J1–J4) that correlated with significant decrease in Arginase-1 + cells in the dermal layer and lower levels of Arginase-1 (Fig. 5L). It is also possible to note that iHDAC treated wounds presented full reconstitution of both epithelial and dermal layers (Fig. 5L) while control wounds still presented a discontinuation between both layers (Fig. 5L, white head arrows). Regardless macrophages from control wounds were positive for F4/80, they did not presented an elongated morphology neither were positive for H4Ac (Fig. 5I, I2–I4). These results clearly show that topic iHDAC treatment recapitulated the effect generated in vitro on BMMP subsets dynamics as well as on macrophage cell shape. We conclude that the topic treatment with iHDAC recapitulated the effects observed in vitro leading to macrophage cell shape changes and promoting the expansion of the Ly6C^low subset in vivo.