BET protein inhibition in macrophages enhances dorsal root ganglion neurite outgrowth in female mice

Abstract Peripheral nerve regeneration is limited after injury, especially in humans, due to the large distance the axons have to grow in the limbs. This process is highly dependent on the expression of neuroinflammatory factors produced by macrophages and glial cells. Given the importance of the epigenetic BET proteins on inflammation, we aimed to ascertain if BET inhibition may have an effect on axonal outgrowth. For this purpose, we treated female mice with JQ1 or vehicle after sciatic nerve crush injury and analyzed target reinnervation. We also used dorsal root ganglion (DRG) culture explants to analyze the effects of direct BET inhibition or treatment with conditioned medium from BET‐inhibited macrophages. We observed that although JQ1 produced an enhancement of IL‐4, IL‐13, and GAP43 expression, it did not have an effect on sensory or motor reinnervation after crush injury in vivo. In contrast, JQ1 reduced neurite growth when interacting directly with DRG neurons ex vivo, whereas conditioned medium from JQ1‐treated macrophages promoted neurite outgrowth. Therefore, BET‐inhibited macrophages secrete pro‐regenerative factors that induce neurite outgrowth, and that may counteract the direct inhibition of BET proteins in neurons in vivo. Finally, we observed an activation of the STAT6 pathway in DRG explants treated with conditioned medium from JQ1‐treated macrophages. In conclusion, this study demonstrates that BET protein inhibition in macrophages provides a mechanism to enhance axonal outgrowth. However, specific targeting of BET proteins to macrophages will be needed to efficiently enhance functional recovery after nerve injury.

During Wallerian degeneration of the distal stump, a network of different cells, including infiltrating macrophages and Schwann cells, participate in axonal and myelin breakdown, through the release of different products. Between these molecules, cytokines and chemokines are essential to regulate the microenvironment that promotes axonal regeneration (Chen, Lin, et al., 2016;DeFrancesco-Lisowitz et al., 2015). Thus, Wallerian degeneration is a cascade of events that reassemble an immune-like reaction, which can be divided in two main phases. The early phase is characterized by the synthesis of inflammatory cytokines and chemokines such as TNFα, IL-1α, IL-1ß, IL-6, and CCL2, inducing the recruitment of immune cells.
The late phase is defined by the increase in anti-inflammatory cytokines such as IL-13, IL-10, and IL-4, and promotes the clearance of debris (Rotshenker, 2011;Vidal et al., 2013). Both processes are indispensable for successful nerve regeneration. For instance, it has been demonstrated that the enhancement of the anti-inflammatory phase attenuates pro-inflammatory cytokine secretion, concludes degeneration, and increases axonal outgrowth (Fregnan et al., 2012;Siqueira Mietto et al., 2015;Vidal et al., 2013). Thus, to provide a favorable environment for axonal growth, it is crucial to achieve a rapid transition from the pro-inflammatory phase toward the anti-inflammatory phase of Wallerian degeneration. Experimental evidence has shown that cytokines trigger JAK/STAT signaling pathways, which are associated with neural plasticity. Particularly, IL-4 and IL-13 activate STAT6, leading to neuroprotection (Bhattarai et al., 2016;Deboy et al., 2006;Vidal et al., 2013). IL-6 and IL-10 activate the STAT3mediated pathway, increasing regeneration and neuroprotection of cortical neurons (Chen, Lin, et al., 2016;Dubovy et al., 2019). BET proteins are epigenetic readers of acetylated lysine residues in histone and nonhistone proteins, and are associated with cellular growth, evasion of apoptosis, and inflammatory response (Filippakopoulos & Knapp, 2014). BET family comprises BRD2, BRD3, BRD4, and the testis-specific protein BRDT. The most studied member of this family is BRD4. BRD4 interacts with an acetylated p65/RELA subunit, triggering the recruitment of p-TEFb complex and stimulating the transcription of NF-kß genes (Hajmirza et al., 2018). Inhibition of BET proteins through the small molecule JQ1 reduces the expression of inflammatory cytokines and prevents TNF-inducible gene expression in macrophages in vitro (Nicodeme et al., 2010). In addition, we have previously found that treatment with the BET inhibitor JQ1 reduces inflammation, increases the presence of anti-inflammatory cytokines, and enhances functional outcome after spinal cord injury (Sanchez-Ventura et al., 2019). Given the importance of BET proteins on inflammatory gene expression, we hypothesized that BET inhibition could favor an early conclusion of Wallerian degeneration through the downregulation of proinflammatory cytokines and the upregulation of anti-inflammatory cytokines. Such a process may enhance axonal growth after PNI via JAK/STAT pathways.
We used the BET inhibitor JQ1 and analyzed its effects after PNI in mice. To further analyze the effects of BET proteins on neuronal outgrowth ex vivo, we treated dorsal root ganglion (DRG) neuronal explants with JQ1. In addition, we used conditioned medium from JQ1-treated macrophages, and assessed its effect on DRG neurite growth. In summary, this study aims to determine the role of BET protein inhibition on axonal regeneration and functional recovery after PNI.

| Animals and surgery
All animal procedures were approved by the Universitat Autònoma de Barcelona Animal Experimentation Ethical Committee (code 10306) and followed the European Commission Directive 2010/63/ EU on the protection of animals used for scientific purposes. Female C57BL/6J mice (Charles River Laboratories) were used in all the experiments, which were kept under a 12-hr light cycle and had access to water and food at libitum.
To perform nerve injury, 7-week-old female mice were anesthetized by intraperitoneal injection of ketamine (90 mg/kg) and xylazine (10 mg/kg) in saline. Sciatic nerve crush injury was performed on the right hind limb at 45 mm distance from the tip of the fourth digit. The lesion consisted on applying a tight pressure sing forceps (Dumont #5) at the sciatic nerve 3 times during 30 seconds, with different orientations each. Then, the muscle and skin were closed in layers. To prevent reopening and later infection of the wound, the skin was secured with staples and iodine was topically applied.
Mice were kept in a warm environment until they recovered from anesthesia. Operated animals were intraperitoneally administered with vehicle (saline with 5% DMSO and 5% Tween-80), or with JQ1 (30 mg/kg, diluted in vehicle) starting at 2 hr (day 0) or at 4 days postoperation (dpo). Each cage of mice contained animals treated with the different conditions, to minimize experimental bias. A total of 27 animals for in vivo studies (15 animals for functional testing and 12 for western blot and qPCR studies) were used.

| Functional testing
Evaluation of axonal regeneration and target reinnervation was conducted through noninvasive electrophysiological tests at 14, 21, and 28 dpo (n = 5 animals per condition). Prior to the functional

Significance
We have found that inhibition of BET proteins in macrophages produce the expression of pro-regenerative factors by these cells. These factors promote neuronal neurite growth. This finding may be important to promote nerve regeneration after peripheral nerve injury.
test, mice were anesthetized by intraperitoneal injection with 0.04 ml of ketamine/xylazine mixture. For assessment of peripheral nerve conduction, the sciatic nerve was stimulated with single electrical pulses of 0.02-ms duration up to supramaximal intensity, delivered by needles inserted percutaneously at the sciatic notch.
Compound muscle action potentials (CMAPs) of the tibialis anterior and plantar muscles were recorded by means of needle electrodes, amplified and displayed on the oscilloscope to measure the amplitude and the latency of the M wave (Navarro, 2016). For sensory nerve conduction, the digital nerve was stimulated similarly, with short-duration electrical pulses of increasing intensity delivered at the sciatic notch. The evoked compound nerve action potentials (CNAPs) were recorded distally at the fourth digit. Functional tests were also performed in the contralateral hind limb as control values for each group.

| Analysis of skin reinnervation
Animals were intraperitoneally anesthetized with 0.2 ml of 1:1 pentobarbital-saline mixture at 28 dpo and perfused with 4% paraformaldehyde (PFA) (n = 5 animals per condition). The distal plantar pads of the hind paw were carefully excised and postfixed with 4% paraformaldehyde for an hour to be later cryopreserved in sucrose 30% solution. Pads 1, referred as the most distal and medial pads that belong to the innervation territory of the sciatic and saphenous nerves, and Pads 2, referred as the most distal and lateral pads that belong only to the innervation territory of the sciatic nerve, were cut longitudinally at 40 μm thickness in a Leica CM190 cryostat (https:// drp8p5tqcb2p5.cloudfront.net/fileadmin/downloads_lbs/Leica%20 CM1950/User%20Manuals/Leica_CM1950_IFU_2v1N_en.pdf, RRID:SCR_018061). Free-floating samples were washed with PBS and PBS-0.3% Tween, and later incubated overnight at 4°C with PBS-0.3%Triton, 1.5% normal donkey serum, and rabbit anti-PGP9.5 antibody (Table 1). Samples were washed with PBS-0.3% Tween and again incubated overnight at 4°C with PBS 0.3% Triton, normal donkey serum 5% and Alexa fluor 594 donkey anti-rabbit (Table 1)
Then, the expression of target sequences was quantified by RTqPCR using SYBR Green QPCR Master Mix (Agilent Technologies) and the corresponding primers (Table 2). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a housekeeping gene.

| Bone marrow-derived macrophage (BMDM) primary culture
Mice were euthanized with pentobarbital diluted in saline (1:1) and cleaned with ethanol 70%. Femur and tibial bones were dissected, bone epiphyses removed, and bone marrows flushed with chilled PBS using a 10-ml syringe and 25G needle. Cells were centrifuged at 500 RCF for 10 min, to be later cultured in 100 ml Petri plates with DMEM/F12 medium that contained 10% fetal bovine serum, 1% penicillin-streptomycin, and L-glutamine. Macrophage colonystimulating factor (M-CSF) was added at 10 ng/ml. Medium was replaced every 3 days with the correspondent addition of M-CSF. At 8 days in vitro (div), adherent cells were reseeded in 6-well plate in a density of 1.6 × 10 6 cells/well with the medium mentioned above, but without the differentiating factor. Cells were treated at 10 div with vehicle containing DMSO or JQ1 1,000 nM for 6 hr (n = 3 independent experiments). Then, conditioned medium was harvested and snap-frozen with liquid nitrogen.
Media from treated-BMDM cells were filtered through 10 kDa ultrafiltration filters (Merck Millipore). Samples were centrifuged at 4°C and 10 000× g for 1 hr. A concentrated medium volume of 50 μl was obtained from each 1 ml medium sample, thus concentrated 20×. Concentrated media were snap-frozen in liquid nitrogen.

| DRG explant culture with or without BMDM-conditioned media
Mice were euthanized with pentobarbital diluted in saline (1:1) and washed in 70% ethanol. All DRGs were extracted using sterilized surgery tools and preserved in cold Gey's balanced solution with 0.6% glucose. DRG roots were removed and nervous tissue was exposed. DRGs were placed in a 24-well plate between two droplets of collagen mixture in Neurobasal-A medium with 1% penicillinstreptomycin, 1x B-27, 0.5% glucose, 0.5% glutamax, and murine β-NGF (100 ng/ml, Peprotech). The collagen mixture used for the culture consisted in 446.43 μl of rat collagen type 1 (Corning), 50 μl of MEM 10X medium, 2 μl of 7.5% sodium bicarbonate (Gibco), and 1.57 μl of PBS. Either JQ1 (500 or 1,000 nM) or DMSO was added into the collagen and into the medium. The culture was maintained for 2 div before fixation and immunohistochemical analysis.
For DRG explant culture with BMDM-conditioned medium, the procedure was the same except for the preparation of the collagen mixture.
In this case, half of the MEM 10x medium was substituted with conditioned medium and the amount of sodium bicarbonate added to the mixture was 8 μl. Provided that JQ1 is a small compound that is filtered during medium concentration, JQ1 or DMSO was also added to collagen to obtain a 500 nM concentration. This procedure resulted in four experimental culture groups: DMSO-treated DRGs with conditioned medium from DMSO-treated macrophages (D + mD), DMSO-treated DRGs with conditioned medium from JQ1-treated macrophages (D + mJQ1), JQ1-treated DRGs with conditioned medium from DMSO-treated macrophages (JQ1 + mD), and JQ1-treated DRGs with conditioned medium from JQ1-treated macrophages (JQ1 + mJQ1) (n = 3 independent experiments). The cultures were also maintained 2 div before fixation.

| Determination of neurite outgrowth in DRG explants
After 2 Figure 1b). A mild increase of IL-4 also was observed when the treatment started at 0 dpo. We further studied the effects of JQ1 in neurotrophic factor expression. We did not observe changes of glial-derived neurotrophic factor (GDNF) (F 2,7 = 0.137, p = 0.874) and brain-derived neurotrophic factor (BDNF) (F 2,7 = 1.833, p = 0.229) expression, whereas we found a reduction of nerve growth factor (NGF) with the treatment starting at 0 dpo (F 2,7 = 14.400, p = 0.003), probably due to the reduced infiltration of macrophages.
We then wanted to ascertain if the transcriptional alterations promoted by JQ1 produced an effect on axonal growth markers expression (Figure 1c). We did not observe an alteration in the superior cervical ganglion-10 protein (SGC-10) (F 2,6 = 1.833, p = 0.229), whereas we found an enhanced expression of the Growth associated protein 43 (GAP43) when the treatment started at 4 dpo (F 2,6 = 6.698, p = 0.0270). Therefore, we planned to further study the effects of axonal regeneration in vivo by a delayed treatment with JQ1.

| BET protein inhibition does not enhance nerve regeneration after sciatic nerve crush
The For the sensory nerve conduction test, CNAPs of the digital nerve were recorded only at 28 dpo. The CNAP latency was increased in the two groups of mice (t 7 = 2.516, p = 0.040). Regarding CNAP amplitude, although JQ1-treated group showed a tendency to be higher (30%) compared to the vehicle-treated group (25%), the difference was not statistically significant ( Figure 2c) (t 7 = 1.126, p = 0.297).
To further study the effects of BET inhibition on axonal regeneration, we analyzed skin reinnervation of the hind paw. Regarding IENF density, JQ1-treated animals showed a tendency of higher number of IENFs/mm (Figure 3b), compared to vehicle-treated animals.
Therefore, the results of in vivo tests showed that JQ1 administration did not have any significant effect on axonal regeneration in the mice.

| BET protein inhibition reduces DRG neurite outgrowth
We analyzed the effects of BET protein inhibition on neurite outgrowth in DRG explants ex vivo, by adding to the medium variable doses of JQ1 or DMSO as control. Two parameters were analyzed in Therefore, BET inhibition produced a significant decrease on DRG neurite outgrowth, that was dependent on the JQ1 concentration. To determine if this decrease was due to compound toxicity, we performed an MTT assay (Figure 4d). Results showed that F I G U R E 2 BET proteins do not affect nerve regeneration after sciatic nerve crush assessed by electrophysiological tests. No significant differences were found between BET-inhibited animals (JQ1) and control animals (DMSO) in the electrophysiological recordings of the latency and amplitude of the CMAP of tibialis anterior muscle (a), CMAP of plantar muscle (b), and CNAP of the digital nerve (c). Data shown as mean ± SD and minimum to maximum in box and whisker graphs 500 or 1,000 nM of JQ1 did not produce cell death, whereas the positive control cisplatin led to 60.07% of cell death (F 3,14 = 12.330, p < 0.001).

| BET protein-inhibited macrophages secrete pro-regenerative factors
Results obtained from in vivo and ex vivo data indicated contradictory effects of BET protein inhibition on axonal growth. Therefore, we hypothesized that these differences might be produced by macrophages, which infiltrate the nerve after lesion in vivo, but were missing in the DRG cultures. To assess the effects that BET-inhibited Therefore, these results indicate that BET-inhibited macrophages secrete pro-regenerative factors that stimulate neurite outgrowth.
Then, we wanted to assess if these pro-regenerative factors were also able to counteract the effects produced by direct BET protein inhibition on DRG neurons. For this purpose, JQ1 was added into the collagen and the medium of DRG explants. When JQ1 was administered in the collagen, it reduced neurite outgrowth, despite the conditioned media added (Figure 5a *p < 0.05, # p < 0.01, % p < 0.001, and $ p < 0.0001 as calculated by one-way ANOVA with Tukey's multiple comparison test and two-way ANOVA analysis followed by Sidak's correction of multiple comparisons. Data shown as mean ± SD or minimum to maximum in box and whisker graphs F I G U R E 5 Conditioned media from BET-inhibited macrophages enhance DRG neurite outgrowth. (a) Representative images from DRG explants at ×20 magnification show that media from BET-inhibited macrophages enhances neuritogenesis. Scale bar = 100 μm.
(b) Conditioned media from BET-inhibited macrophages increases maximum neurite length in vitro. From left to right: DMSO in collagen + DMSO-conditioned medium (D + mD), DMSO in collagen + JQ1-conditioned medium (D + mJ), JQ1 in collagen + DMSOconditioned medium (J + mD), and JQ1 in collagen + JQ1-conditioned medium (J + mJ). (c) Conditioned media from JQ1-treated macrophages increases the number of neurites of different lengths when DMSO is in collagen (upper graph). No significant differences are found between treatments regarding the number of neurites of different lengths when JQ1 is in collagen (lower graph). *p < 0.05, # p < 0.01, and $ p < 0.0001 as calculated by two-way ANOVA with Sidak's correction of multiple comparisons. Data shown as mean ± SD or minimum to maximum in box and whisker graphs (treatment) F 1,34 = 41.86, p < 0.0001) (Figure 5b). Hence, this experiment confirmed our hypothesis that cultured macrophages secreted factors that stimulated neurite growth when treated with JQ1.
To further corroborate our findings, we determined the number of neurites of different lengths. We found that DMSO-treated DRGs with JQ1 conditioned medium (D + mJ) had a significant increase in the number of neurites ranging from 0 μm to 200 μm (Figure 5c, top),

| Conditioned medium from BET-inhibited macrophages induce STAT6 phosphorylation in DRG explants
Our previous experimental study described that BET-inhibited macrophages increase anti-inflammatory cytokine expression of IL-4, IL-10, and IL-13 (Sanchez-Ventura et al., 2019). Moreover, former qPCR analysis demonstrates that JQ1 treatment at 4 dpo after sciatic nerve crush increases IL-4 and IL-13 cytokine transcription. As experimental evidence points that anti-inflammatory cytokines promote neurite outgrowth through acting on JAK/STAT pathways (Vidal et al., 2013), we aimed to determine the activation state of these pathways. For this purpose, DRG explants were treated with conditioned media from JQ1 or DMSO-treated macrophages. We found that ganglia treated with conditioned medium from BET-inhibited macrophages displayed a significant increase in the phosphorylation of STAT6 (Figure 6a,b, Figure S1), being 2.46 times higher than in DRGs treated with conditioned medium from DMSO-treated macrophages (t 3 = 3.278, p = 0.046). However, no significant differences between treatments were found regarding the activation of STAT3 pathway (Figure 6c,d, Figure S1) (t 3 = 1.903, p = 0.153). Thus, conditioned medium from BET-inhibited macrophages led to an increased phosphorylation of STAT6, probably through an enhanced release of IL-4 and IL-13 anti-inflammatory cytokines.

| DISCUSS ION
In the present study we have analyzed the effects of BET inhibition on nerve regeneration in vivo and neurite outgrowth ex vivo.
Overall, BET inhibition after sciatic nerve crush injury in mice did not produce any effect on axonal regeneration.  (Korb et al., 2015;Matzuk et al., 2012). Since no changes on nerve regeneration were observed in mice, to better decipher the action of BET proteins on axonal regeneration in neurons, we used DRG ex vivo cultures.
We observed that JQ1 reduced neurite growth when interacting directly with neurons in culture. The mechanisms producing neurite outgrowth inhibition are not known. Cell viability assay demonstrated that JQ1 was not toxic at the used concentrations, confirming the results of other authors in human mesenchymal stem cells and neurons (Bakshi et al., 2018;Li et al., 2020). It has been previously reported that treatment with JQ1 on cortical neurons produce gene repression of immediate early genes such as Arc and Fos Nr4a1 in response to BDNF external stimulation (Korb et al., 2015).  Mietto et al., 2015), and that the administration of IL-4 and IL-10 enhance axonal regeneration (Atkins et al., 2007;Vidal et al., 2013).
Therefore, our results suggest that BET inhibition leads to an increase of anti-inflammatory cytokines that compensates the detrimental effects that JQ1 has on neurons in vivo by activating pathways able to increase axonal regeneration. In the ex vivo studies, conditioned medium from JQ1-treated macrophages was not able to overcome the inhibitory signaling of JQ1 present in the collagen matrix embedding the DRG, thus acting directly on neurons. We hypothesize that within these cultures, JQ1 is in higher concentrations than in the in vivo study in mice, producing a stronger effect on neuronal BET proteins in the culture than in vivo.
To determine if cytokines are the potential mediators promoting neurite growth of DRG explants, STAT3 and STAT6 phosphorylation was assessed. We did not observe any changes on STAT3 phosphorylation in DRG samples treated with conditioned media from BETinhibited macrophages. STAT3 is activated through IL-6 and IL-10. In fact, IL-6 and IL-10 have been reported to produce neuronal regeneration through phosphorylation and concomitant activation of STAT3 (Chen, Lin, et al., 2016;Vidal et al., 2013). However, since our previous studies demonstrated that JQ1 reduces the expression of IL-6, which may compensate for the IL-10-enhanced expression (Sanchez-Ventura et al., 2019), we did not observe an effect on this pathway to explain the enhanced neuritogenesis. In contrast, we found an enhanced phosphorylation of STAT6 in DRG samples treated with conditioned media from BET-inhibited macrophages. The best known activators of STAT6 are IL-13 and IL-4 (Mori et al., 2016) which are, in turn, strongly induced by STAT6 itself (Czimmerer et al., 2018). In fact, we previously found that BET-inhibited macrophages secrete these cytokines (Sanchez-Ventura et al., 2019), and we confirmed these results after a crush injury in sciatic nerve. Since there are previous reports showing that these cytokines promote axonal regeneration (Vidal et al., 2013), the enhanced neuritogenesis in DRG explants observed by the secreted factors from BET-inhibited macrophages could be attributed, at least in part, to this pathway. It should be also considered that macrophages have two variants of IL-4 receptor (IL-4R) that are able to induce STAT6 phosphorylation. Type I IL-4R can also lead to IRS-2 phosphorylation and consequently to the activation of PIP3/Akt and Erk pathways (Czimmerer et al., 2018), which are also related to axonal regeneration (Hausott & Klimaschewski, 2019;Saijilafu et al., 2013). Therefore, further studies should be performed, to also analyze the implication of the PIP3/Akt and ERK pathways on the neural outgrowth induced by BET-inhibited macrophages.
Finally, it should be also pointed out that we have only used females in this study, which may be a limitation for deciphering potential sex differences. Since differential regenerative capacities between males and females is controversial (Kovacic et al., 2004;Wood et al., 2012), a further study with male mice may be performed to obtain a complete vision of the results obtained.
In conclusion, this study demonstrates that BET protein inhibition in macrophages provides a suitable condition to enhance axonal outgrowth, and brings insight on the relevance of the antiinflammatory cytokines IL-13 and IL-4 and the STAT6 pathway in axonal regeneration. Thus, BET proteins are an effective target to improve axonal regeneration. However, a specific targeting of BET inhibition in macrophages would be needed to efficiently enhance functional recovery after PNI.

DECLARATION OF TRANSPARENCY
The authors, reviewers and editors affirm that in accordance to the policies set by the Journal of Neuroscience Research, this manuscript presents an accurate and transparent account of the study being reported and that all critical details describing the methods and results are present.

ACK N OWLED G M ENTS
The

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.

AUTH O R CO NTR I B UTI O N S
GP and CP performed methodology, investigation and formal analysis. GP, XN and CP performed writing and manuscript preparation.
XN and CP did funding acquisition. CP performed study conception and supervision.

PEER R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/jnr.25036.

DATA AVA I L A B I L I T Y S TAT E M E N T
The raw data can be found in the Universitat Autònoma de Barcelona