Histone deacetylases in kidney physiology and acute kidney injury

Abstract

Histone deacetylases (HDACs) are part of the epigenetic machinery that regulates transcriptional processes. The current paradigm is that HDACs silence gene expression via regulation of histone protein lysine deacetylation, or by forming corepressor complexes with transcription factors. However, HDACs are more than just nuclear proteins, and they can interact and deacetylate a growing number of non-histone proteins to regulate cellular function. Cancer-field studies have demonstrated that deranged HDAC activity results in uncontrolled proliferation, inflammation, and fibrosis; all pathologies that may also occur in kidney disease. Over the past decade, studies have emerged suggesting that HDAC inhibitors may prevent and potentially treat various models of acute kidney injury. This review focuses on the physiology of kidney HDACs and highlights the recent advances using HDAC inhibitors to potentially treat kidney disease patients.

In 1942 C.H. Waddington coined the term “epigenetics” as a means to characterize the connection between the genotype (inherited material) and the phenotype (adult characteristics) of an organism.¹ Epigenetics is now defined as the nongenetic regulation of gene expression with a potential for inheritance.² There are three known global epigenetic mechanisms that can regulate gene expression: 1) DNA methylation,³ 2) histone modification,^4,5 3) non-coding RNA.⁶ These epigenetic mechanisms involve a number of proteins that modify transcriptional processing. The focus of this review is specifically on the histone modifiers, histone deacetylases (HDACs).

In 1964, Allfrey et al.⁴ published that histone proteins could be modified and speculated that this was a way to regulate RNA transcription. They provided evidence that histone proteins could be dynamically acetylated/deacetylated on their lysine ε-amino group.⁷ In the 1980’s the first acetyltransferases were purified^8–10 and later cloned in 1995.¹¹ This family of enzymes was termed the “histone acetyltransferases” (HAT) because of their functions in regulating histone structure. Shortly after the cloning of HAT1, the first mammalian deacetylase was cloned (name histone deacetylase-1, HDAC1).¹² Currently, there are at least 21 acetyltransferases and 18 deacetylases found in mammals (Figure 1).¹³

Figure 1:

The lysine acetyltransferases and deacetylases found in mammals, grouped by their phylogenetic relationships. Dash lines represent predicted enzymes however whether they directly modify lysines is elusive. From Hyndman et al.¹³

The HDACs are organized into classes based upon phylogeny, structure and localization. Class I, II, and IV HDACs are the zinc dependent enzymes and class III are the NAD+ dependent sirtuins (SIRTs) (Figure 1).¹⁴ Current HDAC inhibitors (HDACis) do not directly affect SIRT activity ¹⁵ and thus SIRTs will not be further discussed in this review. Class I HDACs are predominantly nuclear (Figure 2); however, studies have demonstrated their ability to translocate into the cytosol.¹⁶ Class II and IV HDACs translocate between the nucleus and cytosol (Figure 2). HDACs function in the post-translational modification of lysine residues of over 3500 histone and non-histone proteins.^17–19 Although HDACs are classically defined as epigenetic regulators, recent studies have highlighted that dynamic lysine acetylation controls a variety of processes including metabolism, redox homeostasis, actin cytoskeleton, and ion transport.^17,19,20 The importance of non-histone lysine acetylation was recently reviewed.²⁰

Figure 2:

Histone deacetylases (HDAC) are classified based upon phylogenetic relationship. Class I HDACs are predominantly expressed in the nucleus, while class II and IV HDACs can translocate between the nucleus and cytosol.

The HDACs have garnered a lot of attention because deranged HDAC activity is associated with cancer.²¹ There are four Food and Drug Administration (FDA) approved HDACis used to treat T-cell lymphomas and myelomas and hundreds of cancer based clinical trials using HDACis (Table 1). Because HDACi use can lead to reduced inflammation, angiogenesis, proliferation, and fibrosis; it is hypothesized that perhaps HDACis could be repurposed to prevent or treat cardiovascular disease ^22,23 and kidney diseases.^24–26 Although deranged HDAC activity is associated with variety of disease, the focus of this review is on acute kidney injury.

Table 1:

Food and Drug Administration (FDA) approved histone deacetylase inhibitors.

Inhibitor	Class	FDA treatment approval	target	FDA approval Date	Terminal half life T1/2 (h)	Ref.
Vorinostat/SAHA	hydroxamic acid	T cell lymphoma	pan	2006	1.05-1.44	89.90
Belinostat	hydroxamic acid	T cell lymphoma	pan	2015	0.45-2.9	91,92
Panobinostat	hydroxamic acid	multiple myeloma	pan	2015	13-29.3	50
Romidepsin	cyclic tetrapeptide	cutaneous T cell lymphoma	I	2009	1.1-8.1	93,94
Valproic acid	short chain fatty acids	seizures	I, IIA	1978*	10-20	95
Phenylbutyric acid	short chain fatty acids	hyperammonaemia	I,II	1996*	<1	96

Footnote: Ref. - reference

HDACs and the regulation of transcription

Generally, HDACs are considered transcriptional repressors. The classical paradigm is that HDACs remove acetyl groups from lysine residues of core histone proteins²⁷ thereby restoring the positive charge of the histone lysine. The negatively charged DNA can then bind tightly to the positively charged histone protein. HDACs also form complexes with DNA binding proteins (HDACs do not bind directly to DNA²⁸) such as nuclear receptor corepressor 1 (NCOR1)²⁹ and mammalian SIN3.³⁰ Recruitment of these complexes in the nucleus leads to chromatin remodeling and transcriptional repression.³¹ However, perhaps paradoxical, HDAC activity can also result in transcriptional activation. For example, in the developing kidney, genetic deletion of HDAC1 and HDAC2 from the ureteric bud resulted in 226 transcripts significantly upregulated and 270 transcripts downregulated.³² Greer et al.³³ reported that HDACs positively regulate transcriptional elongation; a key process involved in RNA polymerase transcription of target genes. They demonstrated that HDAC inhibitors cause a loss of enhancer RNA synthesis and elongation repression, which resulted in reduced gene transcription.³³ Furthermore, HDACs can also activate transcription factors in the cytosol through lysine deacetylation and thereby indirectly leading to activation of transcription. An example of this is HDAC9 which interacts and deacetylates the transcription factor Forkhead box O 1 (FoxO1) in the cytosol.³⁴ Once deacetylated, FoxO1 translocates to the nucleus where it increases the transcription of target genes such as the glucocorticoid receptor.³⁴ Thus, HDACs function in variety of pathways to either negatively and positively regulate gene transcription.

HDAC expression and physiological function in the kidney

Kidneys are rich in acetylated lysine proteins^17,19 and express many of the acetyltransferases and deacetylases in the nephron (Table 2).¹³ In RNA sequencing data studies of isolated tubule segments,³⁵ all 11 HDACs have been shown to be expressed in the tubules. At the RNA level, HDACs are enriched in the thick ascending limb through to the inner medullary collecting duct.¹³ Interestingly, rats have a duplicate HDAC1 gene named HDAC1-like (NM_053446), which is also expressed at the RNA level in the loop and collecting duct.¹³ It shares 476/484 identity at the amino acid level with rat HDAC1, but whether this is a functional protein remains to be determined. Similar to the rat, the mouse single-tubule RNA-seq dataset^36,37 and the mouse kidney single cell transcriptomics dataset³⁸ provide evidence of all 11 HDACs in the kidney (Table 2). In human subjects, Class I and IV HDAC expression was also found in all tubule segments by serial analysis of gene expression (SAGE).^39,40 Thus, at least over the short evolutionary scale from rodents to man, the kidney HDAC transcripts are highly abundant and conserved in their expression in the kidney.

Table 2:

Histone deacetylase (HDAC) RNA expression from the rat transcriptome³⁵ or mouse single-cell transcriptome³⁸ in the healthy kidney.

	Glom	Podo	PT	SDL	LDLOM	LDLIM	tAL	mTAL	cTAL	DCT	CNT	CCD	OMCD	IMCD	Endo	Fib	Macro	Neut	B lymph	T lymph	NK
Hdac1	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓		✓	✓	✓
Hdac2	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓			✓	✓	✓
Hdac3	✓	✓	✓	✓		✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓				✓
Hdac4						✓			✓	✓	✓	✓	✓	✓				✓
Hdac5	✓	✓	✓	✓		✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓		✓	✓
Hdac6		✓
Hdac7															✓				✓	✓
Hdac8									✓			✓
Hdac9																✓	✓		✓
Hdac10	✓					✓	✓		✓					✓
Hdac11			✓	✓	✓	✓	✓	✓	✓	✓	✓	✓	✓
Rat Hdac 11			✓	✓	✓	✓	✓		✓	✓	✓	✓	✓	✓

Glom, glomerulus; podo, podocyte; PT, proximal tubule; SDL, short descending limb of the loop of Henle; LDLOM, long descending limb of the loop of Henle in the outer medulla; LDLIM, long descending limb of the loop of Henle in the inner medulla; tAL, thin ascending limb of the loop of Henle; mTAL, medullary thick ascending limb of the loop of Henle; cTAL, cortical thick ascending limb of the loop of Henle; DCT, distal convoluted tubule; CNT, connecting tubule; CCD, cortical collecting duct; OMCD, outer medullary collecting duct; IMCD, inner medullary collecting duct; Endo, endothelium; Fib, fibroblast; Macro, macrophage; Neutro, neutrophil; B lymph, B lymphocytes; T lymph, T lymphocytes; NK, natural killer cells.

HDAC1-10 protein abundance was detected in rat or mouse, freshly isolated inner medullary collecting ducts using protein mass spectrometry (see http://helixweb.nih.gov/ESBL/Database/ The Kidney Systems Biology project of The Epithelial Systems Biology Laboratory, National Heart Lung and Blood Institute). Similarly, the Human Protein Atlas (http://www.proteinatlas.org),⁴¹ provides evidence of the 11 HDACs in the human kidney and localization of all but HDAC7. We recently reported the immunolocalization of the 11 HDACs in the adult mouse kidney.⁴² Compared to the single-cell transcriptomics of the mouse kidney,³⁸ there is a strong agreement between our immunolocalization and the RNA sequencing of different kidney cell types. For example, HDAC1 localizes to the nucleus of nearly all cells in the kidney.⁴² HDAC1 RNA is found in all cells including endothelium, podocytes, epithelium, and interstitial cells (fibroblasts, neutrophils, and lymphocytes).³⁸ HDAC9 immunolocalizes to kidney interstitial cells,⁴² and HDAC9 RNA is found predominantly in fibroblasts, macrophages, and B lymphocytes (Table 2).³⁸ HDACs are expressed in all cell types of the kidney and this would suggest the function in diverse physiological processes.

Surprisingly, information is limited about the substrates that HDACs may regulate, and therefore the physiological pathways under their control in the kidney remain undiscovered. One hypothesis is that HDACs regulation water and ion transport processes in the kidney. For example, HDAC7 is expressed in the cortical collecting duct where it is proposed to be a negative regulator of the epithelial sodium channel (ENaC).⁴³ With an increase in HDAC7, there were significant reductions in ENaC abundance and current that were associated with increased lysine ubiquitination.⁴³ Butler et al.⁴³ speculated that lysine acetylation of ENaC would stabilize the channel and lead to sodium reabsorption. The putative lysines that may be acetylated and whether they are substrates for HDAC7 remains unknown. A recent report demonstrated that pan-HDAC inhibition with trichostatin A (TSA) prevented aldosterone-mediated ENaC activity,⁴⁴ and the authors suggested that lysine acetylation of the aldosterone receptor (mineralocorticoid receptor) disrupts the sodium retaining effect of aldosterone. This finding is in agreement with studies measuring blood pressure, which have proposed that HDACi may prevent the development of salt-sensitive hypertension.⁴⁵ Thus, these studies suggest that specific HDAC isoforms may regulate cellular functions in distinct manners and that pan-HDACi reveal the complexity of these signaling cascades. There remains a large gap in our understanding of the physiological role of kidney HDACs, especially at the individual isoform level. We ¹⁹ and others¹⁷ have sequenced kidney acetylomes that identified substrates for HDACs to deacetylate, but the cell type, HDAC isoform, and ultimately physiological effect of regulating lysine acetylation all are understudied.

HDAC Inhibitors

Although HDAC isoform and cellular specific functions may not be clearly defined in the kidney under physiological conditions, the use of HDACis has demonstrated that HDACs may become dysfunctional or deranged in disease states. Most of what we know about HDACi comes from cancer-focused studies, but over the past decade, the therapeutic potential of inhibiting kidney HDACs has emerged as a result of studying different models of acute kidney injury (AKI). Current commercially available HDACis are generally pan-inhibitors that target Class I, II, and IV; or are Class I-selective inhibitors (Table 1). There are four FDA approved HDACis to treat cancers (Table 1). There are also two other compounds with deacetylase inhibitor activity, valproic acid (VPA) and phenylbutyric acid (PBA), which are FDA approved to treat seizures and hyperammonaemia, respectively. These compounds are also being tested in cancer trials.¹⁵ HDACis are given orally or by i.v. infusions, and each has a very different terminal half-life (Table 1). The pharmacokinetics may also differ between preclinical models and humans. For example, entinostat (also called MS-275) has a half-life of >30 h in humans^46,47 but <1 h in rodents.⁴⁶ It is unknown how entionostat is metabolized and removed from the body.⁴⁸ Entinostat also binds plasma proteins in a species-specific manner,⁴⁶ which may explain the significant differences in the pharmacokinetic profiles. This species-specific variability is important to consider when designing and analyzing preclinical studies.

Another large gap that requires attention is the in vivo specificity and pharmacokinetics of HDACi, especially in kidney disease patients. The main clearance pathways of HDACis are excretion of metabolites in the urine and/or feces.^49–51 Even with different levels of kidney function, the terminal half-life of panobinostat was measured between 27.5-29.5 h.⁵⁰ Only small traces of parent drug were found in the urine, yet up to 50% of the metabolites of panobinostat were found in the urine of all patients including those with impaired creatinine clearance.^50,51 Adverse events were no worse among the subjects with reduced kidney function.⁵⁰ Thus, panobinostat was concluded to be well tolerated and safe for people with reduced kidney function although subjects with end-stage renal disease have not been studied. The adverse events associated with HDACis include vomiting, diarrhea, hyponatremia, hypokalemia, and changes in blood pressure (both hypertension and hypotension),^52–54 which need to be carefully monitored especially in people with reduced kidney function. These types of studies are necessary to determine if HDACis can be repurposed to treat kidney disease.

HDAC dysfunction in models of Acute Kidney Injury

There are many preclinical models of AKI.⁵⁵ In the cisplatin-induced nephrotoxicity model of AKI in male mice, whole kidney mRNA for class I HDACs was increased.⁵⁶ In the unilateral ureteral obstruction model, kidney abundance of HDAC1-6 and -10 were significantly increased 2 weeks after obstruction.⁵⁷ In the bilateral model of ischemia-reperfusion injury (IRI), cortical HDAC abundance and localization was dynamic.⁴² For example, HDAC1 was expressed in the nucleus and within 1 h following IRI nuclear abundance was decreased but then progressively increased over the next 3 days. HDAC4 was found in a low abundance in the healthy kidney, but following IRI, it was clearly present in the nucleus 1 h post ischemia but highly expressed in the proximal tubules24-72 h later.⁴² When healthy to injured mouse kidneys were compared, many of the HDACs were significantly affected in their localization.⁴² The consequences of these changes remains to be determined. Although the specific HDAC isoform and its cellular source (epithelial, endothelia, interstitial cell etc.) that is deranged, in most AKI models is unknown, studies using HDACis have revealed some interesting mechanisms of AKI.

Preventative HDACi approaches in:

Sepsis:

Uncontrolled systemic inflammation or “sepsis”, is a clinically relevant problem that unfortunately can lead to AKI and increased mortality. In mice, sepsis can be modeled by the cecal ligation puncture (CLP) procedure.⁵⁸ Valproic acid (VPA) given at the time of CLP, prevented the rise in plasma creatinine, reduced inflammatory cytokines, and limited kidney damage at 24 h post CLP.⁵⁹ Sepsis can also be modeled by intraperitoneal injections of lippopolysacchride (LPS) leading to AKI. Following LPS injection, there was significant mortality within days, but this was significantly improved by a 5 day pretreatment with MS-275. Preventative MS-275 treatment attenuated the rise in serum creatinine, ameliorated kidney damage, and suppressed kidney reactive oxygen species.⁶⁰ Thus, preventative HDACi use can limit sepsis-induced AKI; however, long-term outcomes and treatment regimens require further investigation.

Unilateral Ureteral Obstruction(UUO):

Ureteral obstruction can occur from kidney stones or scarring and lead to an inability to properly drain urine from the kidney to the bladder. If left untreated it causes profound kidney damage. This is modeled by ligating one of the ureters and leaving the other intact. Within days this model leads to severe kidney damage, inflammation, and fibrosis. When trichostatin A (a pan-HDACi) was given by i.p. daily for 6 days beginning at the time of obstruction, it significantly prevented kidney interstitial fibrosis and prevented epithelial apoptosis.⁶¹ Similarly, pretreatment with VPA (i.p. daily) for 5 days prior to 14 days of UUO, attenuated interstitial fibrosis, kidney inflammation, and kidney transforming growth factor (TGF)-β concentration.⁶²

Cisplatin and other nephrotoxic drugs.

Unfortunately, some drugs can accumulate in the kidney, especially in the proximal tubules, resulting in nephrotoxicity.⁶³ Cisplatin, a chemotherapy drug, is frequently used to model nephrotoxic-induced AKI. In cisplatin-induced AKI, pretreatment with TSA (daily i.p.) significantly attenuated the renal damage and rise in plasma creatinine.⁵⁶ Liu et al.⁶⁴ demonstrated in this same model that TSA was renoprotective via enhancing proximal tubule autophagy. Specifically, HDAC6 has been implicated in being deranged in cisplatin-induced AKI, and inhibition of HDAC6 with tubastatin A also increased autophagy and reduced kidney injury.⁶⁵ This was also true for rhabdomyolysis-induced AKI, which is caused by nephrotoxic levels of myoglobin released from necrotic muscle. Inhibition of HDAC6 with tubastatin A resulted in an attenuation in the rise of serum creatinine and kidney damage 48 h after induction.⁶⁶ This was associated with a decrease in kidney inflammation and oxidative stress and preserving theexpression of superoxide dismutase, a free radicle scavenger.⁶⁶

Ischemia-reperfusion-injury (IRI).

If there is low perfusion of the kidneys, IRI can occur. IRI is prevalent in cardiac, vascular, and cardiopulmonary surgery because blood flow is reduced during the surgical procedures.^67–69 To model this, either one or both of the renal pedicles are clamped resulting in ischemia followed by removal and reperfusion. This leads to kidney damage that is dependent upon the time of ischemia. In the warm (body temperature maintained at 36°C), unilateral model of IRI in female mice, an injection of TSA 16 h prior to and again at time of unilateral IRI surgery (ischemia = 28 min), a quicker recovery of plasma blood-urea-nitrogen (BUN) was found. This ultimately lead a to better 30-day survival than vehicle treated controls.⁷⁰ MS-275 was also administered 16 h prior to and at time of unilateral IRI, and although the 24-96 h post injury recovery of plasma BUN was not improved, kidney interstitial fibrosis was still significantly attenuated.⁷⁰ Interestingly, using HDAC6 null mice, Levine et al.⁷⁰ demonstrated that the change in BUN was similar to control mice and interpreted this as the renoprotective effects of TSA and MS275 on inhibiting Class I HDACs and not Class II. This is distinct from nephrotoxic induced-AKI mentioned above, where deranged HDAC6 activity was inhibiting autophagy and leading to worsened outcomes.

Levine et al.⁷⁰ also demonstrated that in cold IRI kidney transplant studies in female mice, TSA given as a single i.p. to both the donor and the recipient resulted in an improved 30-day survival compared to vehicle-treated mice. This was associated with attenuated interstitial fibrosis.⁷⁰ From this study, we can conclude that two preventative i.p. doses of TSA can improve renal outcomes in both the warm IRI and cold IRI transplant models in females. This is likely due to inhibition of deranged Class I HDACs, although deeper investigation is needed.

We recently took a different approach. To better mimic the human pharmacokinetics of TSA and MS725 (Table 1), we implanted i.p. osmotic minipumps to deliver vehicle, TSA (1 mg/kg/day) or MS275 (20mg/kg/day) for 3 days prior to warm, bilateral IRI in male mice.⁴² Surprisingly, we found that HDACis led to an exacerbated increase in plasma creatinine in the IRI mice. This was associated with greater kidney tubular damage from histological evaluation, and greater kidney cortical proximal tubule injury markers compared to controls or shams.⁴² Although kidneys were evaluated at 72 h after ischemia, we detected a small but significant reduction in interstitial fibrosis in the HDACi treated mice. This is consistent with the aforementioned AKI studies in which HDAC inhibition prevents interstitial fibrosis. Thus, although interstitial fibrosis may be prevented (a benefit) by continual HDACi exposure, it may also lead to exacerbated tubular injury (an adverse effect). We speculate that part of the increased injury was from proximal tubule HDAC4 inhibition (with TSA) leading to an attenuated cell cycle and reduced proliferation.⁴² These findings suggest that changes in HDAC activity following AKI may be both beneficial by promoting cellular repair but also detrimental by potentiating injury and increasing scarring. We hypothesize that the cellular source and HDAC isoform activated leads to divergent functions in the kidney.

Interventional approaches:

The antibiotic gentamicin can cause AKI in larval zebrafish.⁷¹ Using this model, Cianciolo and colleagues⁷¹ demonstrated that their novel compound, methyl-4-(phenylthio)butanoate (m4PTB), significantly improved survival two days after gentamicin injection. Moreover, m4PTB treated zebrafish had increased proximal tubule proliferation that was further enhanced in the AKI larvae.⁷¹ Next, they treated two different IRI mouse models. First was a uninephrectomy followed by contralateral IRI (26 min) and treatment with vehicle of m4PTB 24 h post-surgery. Over the next 2 weeks, the m4PTB mice had significantly improved kidney function as measured by serum creatinine but still presented with significant tubular damage.⁷¹ The second model was a 30 min ischemia, followed by contralateral nephrectomy 8 days later. Mice were treated 24 h after IRI and presented with an attenuated increase in serum creatinine compared to vehicle-treated mice 24 h after nephrectomy. This model also presented with reduced kidney injury and less outer medullary interstitial fibrosis when treated with m4PTB.⁷¹ Similarly, m4PTB treatment 4 days after initiation of aristolochic acid nephropathy (AAN) in male mice improved kidney function, reduced proximal tubule injury, and prevented interstitial fibrosis.⁷² Recently, new analogs of 4-(phenylthio) butanoic acid (PTBA) have also been tested in zebrafish and mouse models.⁷³ Skrypnyk et al.⁷³ demonstrated that the prodrug, UPHD186, was metabolized in the liver to form active PTBA and can accelerate recovery when administered 3 days after injury. Thus, this novel HDAC inhibitor is able to treat AKI; however, the HDACs that it specifically targets have not yet been reported.

Adrianmycin infusion in mice causes focal segmental glomerulosclerosis leading to proteinuria and renal failure.⁷⁴ Ten days after adrianmycin treatment of female mice, when proteinuria was at its peak, VPA treatment (given in the drinking water) lead to a return of proteinuria to baseline levels following 3 days of treatment (controls had significantnt proteinuria). Proteinuria remained suppressed over the two week treatment period in the VPA treated mice.^74,75 This was associated with an attenuated rise in serum creatinine, less interstitial fibrosis, and less kidney inflammation. Ultimately, these studies demonstrate that in different models of established AKI, HDAC inhibitors may attenuate the progression of kidney disease.

In agreement with these preclinical studies, a recent clinical study suggested that VPA may also slow the decline of estimated glomerular filtration rate (eGFR) decline in human subjects. Using data from the Veterans Aging Cohort study (a male cohort), Inoue et al.,⁷⁵ reported that the mean rate in eGFR decline was -0.94 ml/1.72 m²/yr (n= 122,870) in this population. Of these, n= 2269 were exposed to VPA as prescribed for psychiatric disorders. Exposure was defined as “presence of a VPA prescription for more than 50% of the days between 2 consecutive creatinine measurements.” VPA exposure resulted in a 35% reduction in the decline of eGFR (−0.61 ml/1.72 m²/yr). Moreover, in some individuals there were data before and after VPA exposure, and again, the decline in eGFR was slowed (−0.93 ml/1.72 m²/yr and -0.32 ml/1.72 m²/yr, respectively).⁷⁵ Finally, this beneficial effect of VPA exposure was most pronounced in individuals with severe proteinuria.⁷⁵

All of these different preventative and interventional studies in AKI models highlight how drug specificity, dosing, timing of treatment, and pharmacokinetics can all be significant contributors to whether HDACi are renoprotective or perhaps lead to severe adverse events. Considering the excitement and positive analysis from the Veterans Aging Cohort study,⁷⁵ hopefully future retrospective analyses and encouragement leading to prospective kidney disease-focused clinical trials with HDACi.

Mechanisms of the renoprotective effects of HDACi in AKI

The HDACi-dependent mechanisms of renoprotection or potentiating injury are understudied. In all of the preclinical studies presented thus far, common pathways such as anti-fibrosis and anti-inflammation have emerged. Especially evident in the preventative strategy preclinical studies, pan- or Class I- selective HDACis attenuated the production of TGF-β, the master regulator of myofibroblast activation,⁷⁶ reduced myofibroblast markers (e.g. vimentin alpha smooth muscle actin), and reduced extracellular matrix (e.g. collagen, fibronectin). Although it is unclear whether this is via direct or indirect inhibition of HDAC-mediated signaling cascade leading to less interstitial fibrosis, studies have demonstrated that HDACis prevent the phosphorylation of SMAD3 (Mothers against decapentaplegic homolog 3).^61,62,77 in the presence of TGF-β, SMAD3 is phosphorylated, translocates to the nucleus, and increases transcription of target genes encoded for extracellular matrix proteins.⁷⁸ How HDACis prevent the phosphorylation of SMAD3 is not clear. Whether it is through activation of a phosphatase, by directly affecting SMAD3 lysine acetylation, or via another pathway remains to be determined. Direct lysine acetylation of SMADs is an intriguing hypothesis because an ortholog of SMAD3, SMAD7, is lysine acetylated resulting in a stabilized protein that inhibits TGF-β signaling.⁷⁹ SMAD7 acetylation of K161 was detected in the inner medullary collecting duct acetylome.¹⁹ Whether Smad3 is also lysine acetylated and the functional consequence of this posttranslational modification require further investigation.

HDACis also reduce inflammation and proinflammatory cytokines.^{62,72,77,80,81} M4PTB prevented CD68-positive macrophage infiltration at 7 and 21 days post injury, and significantly reduced chemokines in the AAN model.⁷² Likewise, the PTBA prodrug, UPHD25, significantly reduced AKI-dependent M1 but not M2 macrophages suggesting that PTBA likely isn’t involved in macrophage polarization.⁸² TSA significantly reduced kidney macrophage infiltration in the UUO model and attenuated the chemokine, colony-stimulating factor-1.⁸¹ Marumo et al.⁸¹ went on to demonstrate that HDAC1 and HDAC2 activity is necessary for Tumor Necrosis Factor (TNF)-stimulated CSF-1 production; however the mechanism was not elucidated. In studies with dendritic cells, SAHA induced indoleamine 2,3-dioxygenase (IDO), an enzyme that suppresses dendritic cell function, through increasing acetylation of histone H4 at the IDO promoter.⁸³ Whether similar mechanisms occur in kidney immune cells with HDACi is unknown.

Kidney and splenic T-cells are also significantly affected by HDACi and UUO. Daily TSA by i.p. to UUO mice resulted in increased infiltrating T regulatory cells (Tregs; CD4⁺FOXP3⁺) and decreased inflammatory T helper 17 cells (Th17; CD4⁺IL-17⁺).⁸⁴ The HDAC-mediated effects on T cell development were recently reviewed⁸⁵ and revealed a complex system of inhibition and activation of various T cell populations through lysine acetylation, metabolism, and cofactor availability. Detailed studies are needed to understand AKI-induced, HDAC-mediated kidney immune responses.

Stimulation of regenerative pathways in the kidney are critical to promote healing. Extended cell cycle arrest in G2/M leads to injury and kidney fibrosis.⁸⁶ HDACis may enhance the regenerative capacity of the injured kidneys via regulation of the cell cycle. Kidneys from IRI mice treated with m4PTB have significantly higher cell cycle-related gene expression, and more actively cycling cells with fewer cells arrested in the G2/M phase.⁷¹ Likewise, m4PTB treatment to AAN injured mice resulted in better regenerative capacity with less cells arrested in G2/M.⁷² However, this is controversial as class I HDACs may be required for proximal tubule proliferation through activation of the epithelial growth factor receptor (EGFR)/signal transducers and activators of transcription 3 (STAT3).^87,88 We also reported that HDAC4 activation may be necessary for proximal tubule proliferation following IRI.⁴² Thus, the mechanisms of HDAC-mediated cell cycle regulation and regeneration are complex and require deeper investigation.

Perspective

A common criticism of studies utilizing HDACis is that because HDACs regulate transcriptional processes and posttranslational lysine acetylation, intervening could potentially lead to a high number of dysregulated genes and pathways. This may be one reason as to why there are many significant adverse events associated with the current pan- and class I-selective HDACis. This in itself is an important reason for defining the role of HDACs in the body, especially in the kidney. With over 500 clinical trials using HDACi to treat different cancers, increasing numbers of patients will be exposed to HDACi so it is necessary to understand HDAC function in the kidney in order to not cause further harm critically ill patients. HDACi have shown a great deal of promise in studies for preventing and even treating kidney injury and preserving kidney function. Still, there remains large gaps in our understanding of the physiological function of HDACs in the kidney (and other organs). Future studies determining HDAC isoform activity, localization, and downstream targets will be key to fully revealing the therapeutic potential of HDACis in the prevention and treatment of AKI. The resulting information will be useful for developing isoform selective, low toxicity compounds; particularly compounds that do not accumulate in the kidney and lead to nephrotoxicity. Finally, considering biological variables such as sex and time of day of dosing, especially given the varied terminal half-life of HDACis, may be key to developing successful therapeutic regiments to prevent and treat AKI.