Browsing by Author "Hirschey, Matthew"

In recent years, our understanding of the scope and diversity of protein post-translational modifications has rapidly expanded. While mitochondrial proteins in particular have been studied in detail and are decorated with an array of acyl groups that can occur non-enzymatically, cytosolic proteins are also known to be acylated. Interestingly, these modifying chemical moieties are often associated with intermediary metabolites from core metabolic pathways. We looked to explore the emerging links between the intrinsic reactivity of metabolites, non-enzymatic protein acylation, and possible signaling roles for these metabolites. Previously HMG-CoA was identified as a reactive metabolite that non-enzymatically modifies proteins when its utilization is blocked. We investigated whether statins, which inhibit the HMG-CoA consuming enzyme HMG-CoA reductase (HMGCR), would result in HMGylation as well. This is particularly relevant because statins are a class of drug widely prescribed for the prevention of cardiovascular disease, with pleiotropic cellular effects. Using mass spectrometry based metabolomics, we found that in both cells and mice treated with statins, HMG-CoA levels increased. Surprisingly, this increase in HMG-CoA led to a single protein being HMGylated, which is in contrast to prior work showing a multitude of protein modifications occurring when acyl-CoA consuming enzymes were altered. Using mass spectrometry, we identified the modified protein as fatty acid synthase (FAS), which implied a new connection between two lipid biosynthetic pathways. We thoroughly characterized the chemical nature of the modification through in vitro chemical treatments. These investigations revealed the modification on FAS to be labile and dynamic compared to previously described acyl modifications. The modification was susceptible to heating, reducing agents, as well as displacement by other acyl-CoAs. Additionally, the modification was rapidly induced and removed in cells in response to the addition or removal of statin. This pointed to a modification that was not lysine bound, which is the commonly reported acyl modification, and is very stable. We developed mass spectrometry methods that would stabilize the modification on FAS and were able to locate the site of HMGylation. The modification occurs on both the serine residue responsible for accepting acyl groups, as well as on the prosthetic group that facilitates the growing of the fatty acyl chain. These modifications occur on active site residues and when tested using purified HMGylated FAS, we found the activity to be inhibited in proportion with the amount of HMG-CoA present. We then tested the impact of statin-treatment on the production of fatty acids in both cells and mice. Surprisingly, we found no change in the lipogenesis rates during statin-treatment, despite the in vitro evidence indicating FAS is inhibited when HMGylated. Recent discoveries have shown a role for FAS outside of the canonical production of lipids for energy storage. With the knowledge that there are subpools of FAS localized within the cell, some of which perform signaling functions, we predicted that the HMGylation occurred on a subpool of FAS in close proximity to HMGCR and endoplasmic reticulum (ER) membrane and conveyed a signal. Utilizing cell fluorescence microscopy, we confirmed the interaction of FAS and HMGCR. To look for any possible changes in cellular signaling we utilized label-free quantitative proteomics and identified pathways changing due to the HMGylation of FAS. While the specific mechanism and nature of these changes remains to be discovered, we identified several signatures of ER and Golgi pathways changing in response to FAS HMGylation and discuss possible areas of interest for future investigation. Finally, we used this new information to begin an investigation into broader non-lysine linked modifications. By artificially inducing acylation in cells, we show that heat or reducing agents play a large role in the detection of acyl-PTMs by Western blotting across HMGylation, glutarylation, and acetylation. In summary, we show how the inhibition of a single enzyme can result in the targeted modification of a protein in an adjacent pathway. This modification occurs on a serine residue and prosthetic group, allowing for dynamic modification and regulation of the enzyme. Furthermore, the discovery of these non-lysine residues reveals the possibility for widespread non-lysine acyl PTMs, waiting to be discovered following the revision of current standard methodology.

Identifying metabolic vulnerabilities of cancer cells remains a subject of investigation for the identification of potential metabolically based therapies for cancer. It is well known that proliferating cells become largely dependent on glucose and glutamine for their growth. Interestingly, we find that lipid oxidizing genes are consistently downregulated across a wide variety of cancers while lipid synthesizing genes are elevated. This indicates that lipid oxidation may be refractory to cancer cell growth. Studies have shown beneficial effects of carbohydrate restriction in various forms in the treatment of cancer. For example, the use of ketogenic diets, which contain high levels of fat and protein with very low levels of carbohydrates, have shown efficacy in decreasing tumor growth in glioma, colon, prostate, and gastric cancers. A major challenge facing the use of these diets in cancer therapy is that the mechanism by which they show efficacy in cancer remains unclear. By using octanoate, the most well-known ketogenic fatty acid, we are able to drive fatty acid oxidation and ketogenesis to study these processes in proliferating cells. We found that supplementation of octanoate into complete culture medium causes a dramatic, dose-dependent and reversible suppression of proliferation across numerous cell lines and significant changes to anabolic cellular metabolism. Importantly, we have found that ketone production from octanoate had no effect on cell proliferation but that the overall cellular response to the lipid causes inhibition of cell growth.

Nutrients and metabolites are sensed by the cell at many levels and the cellular response to metabolites is critical to proliferation and survival of a cancer cell. One way in which the cell responds to glucose, the major fuel source in cancer cells, is by increasing histone acetylation to promote gene expression. Wellen et al., found that upon glucose addition there was a specific gene expression pattern characterized by the upregulation of genes involved in glucose metabolism. In this way the cell promotes glucose-derived fatty acid synthesis, a rate-limiting process for cancer cell proliferation. This is one way in which the metabolic response to nutrients is integrated into cellular signaling and the epigenome. Remarkably, we have found that lipids can promote a feed-forward mechanism of lipid metabolism by inducing histone acetylation and increasing gene expression of lipid metabolizing genes. We find that upon treatment of cells with octanoate there is an inhibition of both glucose and glutamine metabolism and that octanoate-derived carbon becomes the major fuel source in the cell. We then found that histones were hyperacetylated after octanoate treatment and remarkably, close to 90% of the carbon on histones was octanoate derived. In addition, octanoate is a weak HDAC inhibitor which further promotes octanoate-derived acetyl-CoA being deposited onto histones. A gene array from octanoate treated samples finds that fatty acid metabolism is the top pathway in our gene ontology analysis. This provides evidence that the cell responds to nutrient sources in a specific manner depending on the nature of the carbon source. Finally, we find the most negatively regulated pathway upon octanoate treatment is DNA replication. Consistently, we find that octanoate causes an accumulation of cells in G1 phase of the cell cycle and induction of apoptosis.

Here we describe a mechanism for how fatty acids are sensed and how they communicate with the nucleus to alter gene expression. We show that the cell responds to lipids via a coordinated response to promote lipid metabolism and induce histone acetylation. This feed-forward mechanism of lipid metabolism consists of a reprograming of anabolic metabolism, and promotion of gene expression changes culminating in inhibition of cell growth and apoptosis.

Lysine acetylation and its regulation by the mitochondrial NAD+-dependent protein deacetylase sirtuin 3 (SIRT3) are emerging as important regulators of cardiac pathophysiology. SIRT3 serves as a negative regulator of cardiac hypertrophy, and several studies demonstrate decreased SIRT3 expression and increased mitochondrial acetylation in the setting of heart failure. However, the mechanisms that enable the cardio-protective activities of SIRT3 are not well understood. Furthermore, the mechanisms of how acetylation translate to repression of cardiomyocyte metabolism are under studied. Of interest is the genetic cardiomyopathy model that mimics Friedreich’s Ataxia (FRDA) cardiomyopathy–the frataxin knock-out (Fxn-/-, a.k.a. FXNKO) mouse–a model of primary mitochondrial dysfunction due to complex I-III deficiency and/or inhibition that exhibits progressive mitochondrial protein hyperacetylation. We predicted that complex I deficiency in the FXNKO heart may lead to mitochondrial protein hyperacetylation, altered NAD+ homeostasis, and SIRT3 inactivity. We further hypothesized that the hyperacetylation and SIRT3 inactivity in the FXNKO heart are pathogenic and contribute to maladaptive shifts in carbon and energy metabolism known to occur with disease progression in this model and in FRDA patients. Thus, the goal of my dissertation work was to elucidate the role of SIRT3 and acetylation in the pathogenesis of FXNKO and FRDA heart failure. We developed a multi-pronged strategy to realize this goal.

The bulk of my dissertation work focuses on our first approach which involved examining the cardiac intrinsic role of SIRT3 in a cardiac/skeletal-muscle-specific ablation model, as well as the role of SIRT3 in the context of the FXNKO heart failure (see Chapter 3). For these studies, we genetically manipulated SIRT3 in wild-type and FXNKO mice to evaluate the role of cardiac SIRT3 in normal and pathophysiological states. In these studies, we found a cardiac intrinsic role for SIRT3 in maintenance of normal hypertrophic response and contractile function. These data may explain the increased susceptibility of Sirt3 deficient mouse models to cardiac stress, as observed in the literature. Additionally, we explored NAD+ precursor supplementation as a therapeutic means to modulate SIRT3 activity for FXNKO cardiomyopathy. Remarkably, NMN administered to FXNKO mice restored cardiac function to near-normal levels. These effects of NMN required SIRT3, as the improvement in cardiac function upon NMN treatment in the FXNKO were lost in a SIRT3KO/FXNKO (dKO) knockout model. Coupled with cardio-protection, we found that chronic NMN supplementation resulted in improvements in both cardiac and extra-cardiac metabolic function and energy metabolism, and that many of these improvements required cardiac SIRT3. Taken together, these results serve as important preclinical data for a potential strategy of NMN supplementation or SIRT3 activator therapy in FRDA patients.

As a second approach, we conducted preliminary studies to investigate the development of hyperacetylation in the FXNKO heart, with the aim to assess the metabolic source of acetyl-CoA (i.e., the acetylating agent) throughout disease progression (see Appendix A). With these studies, we found that the contribution of various metabolic substrates to the acetyl-CoA pool led to differential acetylation of cardiac proteins in the FXNKO. Additionally, we sought to manipulate the degree of acetylation and test if reducing global acetylation could improve disease outcome. Preliminary results here suggest that the degree of global acetylation is less contributory to disease progression than acetylation at specific sites on proteins. Together, these studies support the notion that shifts in cardiac metabolism with disease progression result in the accumulation of mitochondrial protein hyperacetylation. Additionally, these findings suggest that key sites of hyperacetylation are most critical to the development of cardiomyopathy in the FXNKO, rather than global hyperacetylation.

Lastly, the above findings could serve as preliminary preclinical data for either SIRT3 activation or inhibition of acetylation as therapeutic approaches to treat FXNKO and perhaps FRDA patients. For the results of these studies to have clinical importance, parallel observations of mitochondrial protein hyperacetylation and/or SIRT3 inactivity in the human FRDA heart would need to be assessed. Therefore, we evaluated the cardiac acetylome of autopsied FRDA patient hearts (see Appendix B). Alongside these samples, we included non-failing hearts as a negative control and two other types of human heart failure (i.e., non-ischemic failing and ischemic failing hearts) as positive controls. We found significant hyperacetylation of cardiac proteins in all three types of human heart failure, as expected based on the literature. Importantly, we found that the FRDA patients, like the preclinical FXNKO mouse, develop marked hyperacetylation of cardiac proteins at end-stage disease. Altogether, these preliminary data provide some clinical relevance of acetylation and/or SIRT3 inactivity in the FRDA heart. Additionally, these findings strongly support furthering efforts to develop therapeutic approaches to manipulate acetylation and/or SIRT3 activity as a prospective treatment for FRDA.

Overall, our work provides insight into the role of SIRT3 and acetylation in the heart under basal and stressed conditions. These data validate an important role for SIRT3 in the heart as a mediator of the hypertrophic response. Furthermore, these data identify a novel role of SIRT3 in mediating contractile function in the heart, a finding that may explain the increased sensitivity of Sirt3 ablation mouse models to cardiac stress. Finally, these data provide early mechanistic insight into the contribution of acetylation at key SIRT3 regulated sites to the pathogenesis of cardiomyopathy in FXNKO mice and FRDA patients. This insight is necessary to fully realize the potential of metabolic therapy via manipulation of SIRT3/acetylation in the FRDA heart and in other types of human heart failure.

Sirtuin 5 (SIRT5) is one of three mitochondrial proteins that belongs to the sirtuin family of NAD+-dependent deacylases. Mitochondrial sirtuins (SIRT3-5) control metabolism in physiological and pathophysiological conditions by their deacylation activity. SIRT5 possesses demalonylase, desuccinylase, and deglutarylase activity. While the enzymatic activity of SIRT5 has been well characterized, the physiological role of SIRT5 is less understood. Recent evidence suggests that SIRT5 may have a role in responding to cardiac stress. Given that succinylation is abundant in the SIRT5KO heart, it is important to understand the role of SIRT5 mediated desuccinylation in the heart. Since it appears that there are no defects in cardiac function in SIRT5KO mice under basal conditions, I hypothesized that a stress would be required to determine a protective effect of SIRT5 in the heart. Additionally, I hypothesized that multiple sites of lysine succinylation would contribute to the overall phenotype observed.

To address this hypothesis, genetic mouse models were exposed to a well characterized model of pressure overload induced cardiac hypertrophy—transverse aortic constriction (TAC). Two main mouse models were used: 1) a whole body SIRT5KO mouse and 2) a cardiomyocyte tamoxifen-inducible heart specific SIRT5KO mouse. In order to characterize cardiac structure and function following TAC, the methods of serial echocardiogram and pressure volume loops with inferior vena cava suppression were used. Mechanistic studies included metabolomic and succinyl- proteomic profiling of heart samples. Additionally, Western Blot and RT- qPCR were used to further gain mechanistic insight.

Chapter 3 of this dissertation characterizes the response of the whole body SIRT5KO mouse to pressure overload induced hypertrophy compared to littermate controls. We found that exposure to chronic TAC significantly increases mortality in SIRT5KO mice. Analysis of cardiac morphology and function after 4 weeks of TAC shows that SIRT5KO that have survived to this point have similar cardiac morphology and function compared to WT TAC mice. We predict that impaired oxidative metabolism is a major contributor to accelerated death in SIRT5KO mice.

To specifically investigate the role of SIRT5 in cardiomyocytes, a heart-specific, inducible SIRT5KO mouse was generated and exposed pressure overload via TAC surgery. We find that the phenotype of increased mortality in whole-body SIRT5KO mice is not recapitulated under the conditions tested in heart-specific SIRT5KO mice. However, the two genetic mouse models show differences in protein succinylation, leading us to perform succinyl-proteomics in this model. The results of this investigation are presented in Chapter 4. Collectively, the results of these studies provide new insight into the role of SIRT5 mediated desuccinylation in regulating cardiac function and metabolism.