Generation of myeloid-specific KO mice Myeloid-specific HDAC3 KO mice

C57BL/6J floxed mice with LoxP sites on either side of exon 7 of HDAC3 (HDAC3f/f, originally developed by Dr. Scott W. Hiebert) were obtained from Dr. McGee-Lawrence and bred in our colony [34]. These mice were crossed with LysM Cre mice (Jackson Laboratory Stock No. 004781) to generate MΦ and microglia cell-specific HDAC3 KO (LysM Cre; HDAC3f/f or M-HDAC3−/−) mice. Mice genotyping and characterization are shown in Supplementary Fig. S1.

Myeloid-specific A1 KO mice

C57BL/6J floxed mice that have LoxP sites for A1 (A1f/f) were crossed with LysM Cre mice (Stock No. 004781) to generate myeloid-specific A1 KO (M-A1−/−) mice, which were used for studies with isolated MΦ.

Mouse retinal ischemia-reperfusion (IR) injury model and RGFP966 administration

All experiments were approved by the UAMS Institutional Animal Care and Use Committee (IACUC). M-HDAC3−/−, HDAC3f/f, and wild-type (WT) C57BL/6J mice (10–12 weeks old) were anesthetized using a ketamine/xylazine mixture and then subjected to retinal ischemia followed by reperfusion to achieve IR injury as described by us and others [10, 35,36,37]. A needle connected to a raised saline bag was inserted into the anterior chamber of the right eye to raise the intraocular pressure to 110 mmHg (calculated based on the height of the saline bag). Ischemia was induced for 60 min followed by needle removal to allow reperfusion. The left eye served as a sham control since we did not detect differences between contralateral eyes and those from uninjured mice (Supplementary Fig. S2) [10, 37]. Animals with a drop in pressure due to saline leakage out of the eye during the procedure were excluded. Mice were deeply anesthetized and sacrificed by transcardial perfusion or cervical dislocation at the various time points shown in Table 1 and Supplementary Fig. S3 based on our previous studies and existing literature.

Table 1 Time points of the main outcome measures in the in vivo IR studies

For studies on WT mice, the specific HDAC3 inhibitor RGFP966 (Sigma, St. Louis, MO) was dissolved in DMSO and then diluted in 70% polyethylene glycol (PEG) 200, 30% acetate buffer for intraperitoneal (i.p.) injection as previously described for a final DMSO concentration of ∼ 10% [44]. RGFP966 was administered (10 mg/kg i.p.) at 1 h after IR and every 48 h thereafter based on previous literature [45].

Optical coherence tomography (OCT)

The eyes of mice were dilated with 1% tropicamide (Akron Pharmaceuticals, IL) and retinas were scanned with an OCT Ophthalmic Imaging System (Bioptigen Inc., Durham, NC) under ketamine/xylazine anesthesia. Mice were placed on a mouse holder to fix the animal’s posture for scanning. The retina scans were acquired in a rectangular volume mode (3 frames/scan, 1000 A scan/B scan × 100 B scan, 1.4 mm × 1.4 mm). Automated analysis of scans to obtain retinal thicknesses was conducted using InVivoVue software (Bioptigen Inc., Durham, NC).

Electroretinography

A full scotopic electroretinogram (ERG) was recorded on day 14 post-IR as previously described with some modification [46]. Briefly, after 16 h of dark adaptation, mice were anesthetized using a ketamine/xylazine mixture, and the pupils were dilated with 1% tropicamide (Akron Pharmaceuticals, IL) eye drops. Mice were then placed on the Celeris electroretinogram system (Diagnosys LLC, Cambridge, UK) with temperature control (37 °C), and dark-adapted ERGs were recorded. Amplitudes and implicit times of ERG waveforms were measured at a series of flash intensities (0.01, 0.1, 1 cd.s/m2) based on a standard machine protocol and literature [46].

Retinal vascular permeability Albumin extravasation

Vascular permeability was examined by measuring the extravasation of albumin to the retina on day 2 after IR injury following transcardial perfusion to remove intravascular albumin as we and others previously described [35, 47]. Briefly, a 20-gauge perfusion cannula was introduced into the left cardiac ventricle of the anesthetized mice. Mice were perfused with phosphate-buffered saline (PBS) for 5 min to rinse out blood cells and proteins. Drainage of the blood in the PBS was conducted by opening the right atrium. Retinas were then lysed, and Western blotting was performed to quantify retinal vascular leakage by measuring extravascular albumin in the whole retina.

Retinal fluorescein angiography

Retinal vascular permeability in vivo was assessed using fluorescein angiography as previously described [48, 49]. Briefly, two days post-IR injury, mice were anesthetized using a ketamine/xylazine mixture, and pupils were dilated with 1% tropicamide (Akorn Pharmaceuticals, IL) eye drops. The anesthetized mice were mounted on a maneuverable imaging platform, and their corneas were lubricated with a thin layer of GenTeal moisturizing eye gel (Alcon, Geneva, Switzerland) to keep surface moisture during the procedure. Each mouse received a subcutaneous injection of 50 µL of AK-FLUOR® fluorescein sodium (10%; Long Grove Pharmaceuticals, Rosemont, IL). Rapid image acquisition ensued over a period of approximately 5 min. Fluorescein leakage was demonstrated by indistinct vascular borders progressing to diffusely hazy fluorescence.

Evans blue leakage

In vivo retinal vascular permeability was assessed by imaging of intravascular Evans Blue dye extravasation at day two post-IR injury as previously described with some modification [50]. Briefly, anesthetized mice received transcardial injections of Evans Blue (200 µL, 2% in normal saline), euthanized after 5 min while under anesthesia, and eyes were fixed with 4% paraformaldehyde (PFA) for 24 h. Retinal flat mounts were mounted and cover-slipped with Vectashield antifade mounting medium (Vector Laboratories, Newark, CA), and images were captured using a fluorescence microscope. Images of the entire retina flat mount were processed using ImageJ. The red color (Evans Blue) was thresholded in each image, the background was subtracted, and the fluorescence area was binarized. Finally, the leakage area, represented as a binary area in the entire retina, was measured using ImageJ.

Retinal vasculature trypsin digestion and counting of acellular capillaries

Eyeballs were isolated on day 14 after IR injury and fixed overnight in 4% PFA. The retinal vasculature was isolated by trypsin digestion as we and others have reported [6, 35, 51]. The vasculature was air-dried on silane-coated slides and stained with periodic acid-Schiff and hematoxylin. Acellular capillaries were counted in random fields of the mid-retina under the microscope. The number of acellular capillaries was divided by the field area to get the number of acellular capillaries per 1 mm of the retina [2].

Fluorescent immunolabeling

Eyeballs were fixed in 4% PFA and then dissected into retinal flat mounts. The flat mounts were permeabilized in 0.1% triton X-100 and then blocked in 10% donkey serum and 3% bovine serum albumin (BSA) for 30 min. Subsequently, the flat mounts were incubated in primary antibodies including Iba1 (FUJIFILM Wako, Cat. #019-19741) and NeuN (MilliporeSigma, Cat. #BN78MI) overnight at 4 °C followed by washing in PBS and incubation of secondary antibodies at room temperature for 4 h as described earlier [35].

Confocal microscopy and image analysis

Confocal imaging was performed using an LSM 880 Airyscan Zeiss Laser inverted microscope equipped with 405 nm, 488 nm, 561 nm, and 640 nm laser lines. Identical laser intensity settings were applied to all samples and Z-stacks images (resolution: 1024 × 1024 Pixels) were taken. Neurodegeneration and myeloid cell proliferation were determined as the NeuN+ and Iba1+ area, respectively [52]. After the acquisition, a maximum intensity projection of the Z-stack was applied using ZEN Blue software (Zeiss) or ImageJ. Quantitative analysis was performed using ZEN software or ImageJ on single-slice confocal images. Morphological analysis of myeloid cells utilized Sholl analysis (Fiji Sholl analysis plugin), as previously described [53]. Random Iba1+ cells in the ganglion cell layer (GCL) were chosen from three different fields of view and subjected to Sholl analysis. Criteria for cell selection included: (i) non-overlapping with other labeled cells, ensuring clear distinction of cell body and processes from neighboring cells, and (ii) all cellular processes within the field of view. The analysis determined the complexity of myeloid (Iba1+) cell processes by analyzing branch intersections at increasing radial distances from the cell body.

Microscopy-based in vivo efferocytosis assay on retinal flat mounts

For microscopy-based quantification of in vivo efferocytosis, double immunolabeling of the microglia/MΦ marker, Iba1, and TUNEL staining of retinal flat mounts was conducted on day 2 after IR injury using the Click-iT™ Plus TUNEL Assay (Invitrogen™) following the manufacturer’s instructions. The number of Iba1+ positive cells associated with TUNEL+ apoptotic bodies was counted in multiple fields of view in Z-stack images taken at the ganglion cell layer using an LSM 880 Zeiss confocal microscope (Supplementary Fig. S4A). The efferocytosis index (% of dead/dying cells engulfed by myeloid cells) was calculated using the following equation: ([number of Iba1+TUNEL+ cells ÷ total number of TUNEL+ cells] × 100).

Flow cytometry analysis of retinal immune cell populations

Following animal euthanasia by transcardial perfusion, eyes were removed by orbital dissection and retinas were isolated. Retinal tissues (2 to 4 retinas pooled per preparation) were processed for flow cytometry and analysis as described by us [54]. Briefly, retinas were digested in a solution containing 5% filtered fetal bovine serum (FBS, Gibco; Thermo Fisher, NY) 10 mM HEPES, 0.5 mg/mL of liberase (Sigma Aldrich, St. Louis, MO ), and 0.1 mg/mL of DNase (Sigma Aldrich, St. Louis, MO). Cells were strained through a 40 μm cell strainer and then incubated with a viability dye (Fixable Viability Dye eFluor™ 450, eBioscience™) for 30 min. After rinsing the viability dye, cell suspensions were blocked with 1 µg/mL of Fc block anti-mouse CD16/32 and 20% normal rat serum for 10 min at room temperature. Subsequently, retinas were incubated with labeled antibodies that included PerCP-Cy5.5-conjugated rat anti-mouse CD11b monoclonal antibody (1:100, Clone M1/70, BD Bioscience, Cat. #BDB550993), APC-Cy7 rat anti-mouse CD45 monoclonal antibody (1:100, Clone 30-F11, BD Bioscience, Cat. #BDB557659), PE-conjugated rat anti-mouse Ly6C monoclonal antibody (1:100, Clone HK1.4, eBioscience, Cat. #50-245-507), and FITC-conjugated rat anti-mouse Ly6G (Gr-1) monoclonal antibody (1:100, Clone RB6-8C5, eBioscience, Cat. #50-991-9). After incubation with antibodies, retinas were rinsed 3 times with cold PBS, fixed with 0.4% PFA, and analyzed using a BD LSRFortessa flow cytometer (BD Biosciences) and FlowJo software (Tree Star Inc., San Carlos, CA). Retinal cell populations initially were gated using the common leukocyte marker CD45 and the myeloid lineage marker CD11b. Based on the expression of CD45, myeloid cells were gated as CD11b+/CD45low microglia and CD11b+/CD45hi myeloid leukocytes. Myeloid leukocytes were gated further based on Ly6C expression and the granulocyte/neutrophil marker Ly6G. Myeloid leukocytes were subdivided into classical monocytes (CD11b+/CD45hi/Ly6Chi/Ly6Gneg), intermediary monocytes (CD11b+/CD45hi/Ly6Clow/Ly6Gneg), and non-classical monocytes (CD11b+/CD45hi/Ly6Cneg/Ly6Gneg).

Flow cytometry-based in vivo efferocytosis assay

In subsequent experiments, mice were treated with PSVue 550 eyedrops (Molecular Targeting Technologies, Cat. #P-1005) one day before sacrifice to assess efferocytosis by flow cytometry (Supplementary Fig. S4B). PSVue dye allows in vivo fluorescent labeling of dead cells since it binds to externalized phosphatidylserine (PtdSer, an “eat-me” signal) on apoptotic cells in live mice [55]. Retinas were then incubated with labeled antibodies that included PerCP-Cy5.5-conjugated rat anti-mouse CD11b monoclonal antibody (1:100, Clone M1/70, BD Bioscience, Cat. #BDB550993), APC-Cy7 rat anti-mouse CD45 monoclonal antibody (1:100, Clone 30-F11, BD Bioscience, Cat. #BDB557659) and eFluor 660 conjugated CD68 monoclonal antibody (eBioscience, Cat. #50-0681-82) with the latter used as a marker of phagocytic cells. Phagocytic microglia and myeloid leukocytes were gated from CD11b+/CD45low and CD11b+/CD45hi populations, respectively, as CD68+/PSVue+ cells engaged in efferocytosis of apoptotic cells.

Cell isolation, cell lines, and culture Bone marrow-derived macrophages (BMDMs)

Bone marrow cells were isolated and differentiated into MΦ based on our published protocol [6]. In brief, both femurs and tibias were harvested and flushed with 20 − 25 ml sterile PBS using a 27-gauge needle. Flushed cells in PBS were spun down and resuspended in differentiation medium (Dulbecco’s modified Eagle’s medium, DMEM, high glucose, Gibco; Thermo Fisher, NY)) containing 20% FBS (Gibco; Thermo Fisher, NY), 20% L929 conditioned media, and 1% penicillin-streptomycin (Pen-Strep). Cells were subsequently plated on uncoated 100-mm dishes. For studies of in vitro efferocytosis, BMDMs were plated on 35-mm glass bottom dishes (MatTek Corporation, Ashland, MA) with a 14-mm glass diameter. Media were replaced with fresh differentiation media on day 4 after plating.

K562 lymphoblast cells

Suspended K562 cells were cultured in RPMI 1640 medium (Gibco; Thermo Fisher, NY) containing 2 mM L-glutamine supplemented with 10% FBS (Gibco; Thermo Fisher, NY) and 100 IU/ml Pen-Strep (Gibco; Thermo Fisher, NY), and maintained in a 5% CO2 incubator at 37 °C [56].

R28 retinal neuronal-like cells

The R28 cell line (Kerafast, Boston, MA) was maintained and differentiated as described previously [52]. Cells were cultured in DMEM supplemented with 10% FBS (Gibco; Thermo Fisher, NY), 100 U/ml Pen-Strep, and maintained under standard culture conditions (37 °C, 5% CO2). The medium was changed completely every other day and cultures were passaged at 90% confluence. To induce differentiation, cells were passaged onto laminin-coated plates and supplemented with 250 µM 8-(4-Chlorophenylthio)adenosine 3′,5′-cyclic monophosphate sodium salt (pCPT-cAMP) for overnight incubation.

Induction of apoptosis in K562 and R28 cells in vitro

K562 cells or R28 cells (apoptotic and non-apoptotic) were labeled with the Vybrant® CFDA SE Cell Tracer Kit (Invitrogen, Carlsbad, CA) for 30 min at 37 °C, washed twice with complete media, and resuspended at 5 × 105 cells/ml. BMDM were labeled with CM-DiI Dye® (Invitrogen, Carlsbad, CA) in pre-warmed PBS and incubated for 15 min at 37 °C followed by another incubation for 30 min at 37 °C with DMEM (supplemented with 100 IU/ml Pen-Strep) before rinsing three times with PBS.

Apoptosis of K562 or R28 cells was induced by subjecting the cells to UV-B irradiation for 15 min using a UV crosslinker (Spectrolinker™ XL-1500, Spectronics Corporation, Melville, NY). Afterward, the cells were resuspended in DMEM containing 10% FBS (Gibco; Thermo Fisher, NY) and incubated at 37 °C for up to two hours and monitored for apoptosis induction (visible cell blebbing under the microscope). Apoptosis was also confirmed by Annexin V labeling (Supplementary Fig. S4C) [56]. UV-B was chosen as the apoptosis induction method due to its rapid action and widespread use [56, 57].

Evaluation of macrophage efferocytosis of apoptotic cells in vitro

Three sets of experiments using the K562 and R28 cells prepared as described above were conducted to evaluate in vitro efferocytosis:

Experiment 1

CM-DiI labeled BMDMs derived from HDAC3f/f and M-HDAC3−/− mice were plated on 35-mm dishes and treated with or without the A1 inhibitor, 2(S)-amino-6-boronohexanoic acid (ABH) (100 µM). Then cells were incubated with CFDA-labeled K562 cells (apoptotic or non-apoptotic) at a 1:1 ratio for 45 min at 37 °C followed by PBS washing to remove non-engulfed K562 cells. Subsequently, the dishes either were imaged under a confocal microscope to detect the efferocytic MΦ or were trypsinized and immediately analyzed by the BD Accuri™ C6 Plus Flow Cytometer (BD Bioscience, San Jose, CA). Cells that were double positive for CM-DiI/CFDA were regarded to represent phagocytic BMDMs that engulfed apoptotic cells. Efferocytosis was calculated as a fold change by dividing CM-DiI+ CFDA+ (phagocytic BMDMs) by the total number of CM-DiI+ BMDMs.

Experiment 2

CM-DiI labeled BMDMs derived from A1f/f and M-A1−/− mice and plated on 35-mm dishes were incubated with CFDA-labeled R28 cells (apoptotic or non-apoptotic) for 45 min at 37 °C followed by PBS washing to remove non-engulfed cells. The dishes were then processed for imaging under the microscope to detect the efferocytic MΦ.

Experiment 3

CM-DiI labeled BMDMs derived from A1f/f and M-A1−/− mice that were plated on 100-mm dishes were incubated with CFDA-labeled K562 cells (apoptotic and non-apoptotic) for 45 min at 37 °C. After washing the non-engulfed cells, the adherent BMDMs either were imaged by a confocal microscope or trypsinized and immediately analyzed by flow cytometry.

Real time-PCR to detect A1 expression in BMDMs

For RNA isolation, 4 × 105 BMDMs derived from HDAC3f/f and M-HDAC3−/− mice were seeded onto 6-well plates and incubated with 7.5 × 105 K562 cells (apoptotic or non-apoptotic) for 45 min at 37 °C. Following treatment, K562 cells were removed by several PBS washes. BMDMs were then incubated for another 5 hours, and total RNA was isolated using TRIzol reagent (Invitrogen, CA) and converted to cDNA using M-MLV reverse transcriptase (Invitrogen, CA). The reverse-transcriptase qPCR assay was performed using a Verso cDNA Synthesis Kit (Fisher Scientific, NJ). Gene expression was performed using the PowerTrack™ SYBR Green Master Mix (Applied Biosystems™) and a CFX96 Touch Real-Time PCR Detection System (BioRad). Data were analyzed using the comparative Ct method with GAPDH as the reference gene. For A1, the following primer sequences were used: Forward, CAGAAGAATGGAAGAGTCAG; Reverse, CAGATATGCAGGGAGTCACC.

Preparation of BMDMs subjected to efferocytosis for Western blotting

BMDMs derived from HDAC3f/f and M-HDAC3−/− mice were seeded into 6-well plates and incubated with 7.5 × 105 K562 cells (apoptotic or non-apoptotic) for 45 min at 37 °C. BMDMs were then washed several times with PBS to rinse the non-engulfed K562 cells and further incubated for 18 h before protein extraction in RIPA buffer prior to Western blotting.

Western blotting of frozen retinas and cells

Retinas or cells were collected, snap-frozen, and then stored in a -80oC freezer for further processing. Tissues or cells were homogenized in RIPA buffer (Thermofisher) and Western blotting was conducted as previously described [6]. The primary antibodies were as follows: HDAC3 (BD Bioscience, San Jose, CA, Cat. #611,124), albumin (Proteintech, Cat. #16475-1-AP), A1 (GeneTex, Cat. #GTX109242), and β-actin (Sigma, Cat. #A5441). Secondary antibodies (Invitrogen) were prepared in 5% milk in a 1:2000 dilution.

Human eye sections and transcriptome

Postmortem human eye paraffin-embedded sections were obtained from the National Disease Research Interchange (NDRI, Philadelphia, PA). Sections from patients diagnosed with DR and controls were deparaffinized in xylene (Fisher Scientific, NJ), and rehydrated in graded ethanol baths (100%, 90%, 70%, 50%) followed by a final incubation in distilled water. Antigen retrieval was achieved by microwaving the sections in Tris-EDTA buffer (10 mM Tris Base, 1 mM EDTA, 0.05% Tween 20, pH 8.0). Sections were incubated in 0.2% Triton X-100 in PBS for 15 min followed by blocking with 3% normal donkey serum (MilliporeSigma, Cat. # 5,058,837) and 3% BSA for 1 h. Sections were then incubated overnight at 4 °C with diluted primary antibodies in a blocking buffer that included Iba1 (FUJIFILM Wako, Cat. # 011-27991), and HDAC3 (Invitrogen, Cat. #PA5-29026). After primary antibody incubation, sections were incubated with the appropriate secondary antibody, then washed with cold PBS and cover-slipped with Vectashield antifade mounting medium containing 4′,6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Newark, CA) to mark nuclei.

Information on HDAC3 expression in the human eye was obtained by searching the Human Eye Transcriptome Atlas, a publicly available database https://www.eye-transcriptome.com/search_latest.php.

Statistical analysis

Statistics were performed and graphs prepared using GraphPad Prism 10 software and data were presented as mean ± standard deviation (SD). A p-value < 0.05 was considered statistically significant. All statistical analyses were performed using the Student’s t-test (for two group comparisons) or analysis of variance (ANOVA) with Tukey’s post hoc test (for comparison of multiple groups).