Stem Cell Rejuvenation by Restoration of Youthful Metabolic Compartmentalization

Introduction

Stem cell function decline in old age is considered a hallmark of aging. ¹ Stem cell decline is often referred to as stem cell exhaustion, although the excessive proliferation-related connotations of this term tend to obscure other important mechanisms for stem cell dysfunction such as alterations of the stem cell niche or milieu and epigenetic drift. Bone marrow-derived mesenchymal stem cells (MSCs) help regenerate bone and are multipotent, being capable of differentiating into adipocytes, chondroblasts and osteocytes, the latter being especially important to maintaining bone homeostasis. ^{2 –4} With aging, MSCs have reduced ability to differentiate into osteocytes and chondrocytes, which may explain the higher risk of osteoporosis and fractures observed in the elderly as well as aging-associated bone dysfunction in old mice. ^5,6

Although accumulation of cellular damage and imperfect maintenance of the epigenome are thought to play key roles in the decline of stem cell function, ¹ the precise mechanisms involved are yet to be elucidated. The great complexity of epigenetic changes that occur involve interactions between transcription factors and other DNA-binding proteins, post-translationally modified histones, and methylated cytosines within the DNA. Chromatin modifications that control cell identity and function, and the not insignificant cell-to-cell variation even in cells of the same type, present significant hurdles to understanding even epigenomic regulation in young animals, never mind aging animals. ⁷

However, the epigenome is intimately connected to intracellular metabolism, as chromatin-modifying enzymes require key metabolites such as acetyl-CoA, NAD, or ATP to function. ⁸ To understand the changes that occur in cell states during aging requires understanding metabolic compartmentalization and the communication that occurs between the nucleus, cytoplasm, and mitochondria. Increasingly complex interactions between localized intracellular metabolism and the epigenome are being reported. ⁹

Altered Mitochondrial Metabolite Transport Causes Stem Cell Dysfunction

In an important article, Pouikli et al. ¹⁰ link aging-associated adult stem cell dysfunction to chromatin structure to altered citrate transport from the mitochondria and ultimately to mitochondrial stress.

Pouikli et al. isolated MSCs from bone marrow endosteum from relatively young 3- to 5-month-old mice, and old 18- to 22-month-old mice followed by in vitro culture for 4 weeks under conditions that attempt to emulate the stem cell niche, including using physiological 2% O₂. The cells from old mice proliferated less than those from young mice, but tended to differentiate more frequently into adipocytes and less frequently into osteocytes than those from young animals. Interestingly, the expression of biomarkers for MSC was similar between the young and old cells, but by using transcriptome profiling changes consistent with the skewing of differentiation toward adipogenesis and the reduced structural integrity of the trabecular bone compartment were uncovered. ¹⁰

Changes in the Epigenome of Aging MSC

To characterize the key changes to the epigenome, Pouikli et al. ¹⁰ used “Assay for Transposase-Accessible Chromatin using sequencing” (ATAQ-seq), in which only open chromatin regions are permissive for integration of bacterial transposon Tn5. ¹¹ The transposon-containing sites are then sequenced using transposon-specific primers. Interestingly global chromatin accessibility declined, which is not what was previously observed in a landmark article using yeast, ¹² but has been observed in adipocyte-derived stem cells ¹³ and oocytes. ¹⁴ Because inaccessible chromatin correlates with genes that are not expressed, bioinformatic analysis using the NucleoATAC algorithm to map nucleosomes revealed that there was higher nucleosome density at the promoter regions; however, no nucleosome fuzziness in location was observed, contrary to that previously reported in old yeast. ¹⁰

The increased number of nucleosomes maintained their positioning. Supporting the possibility that changes to the epigenome were important to MSC dysfunction in older animals, gene ontology analysis showed reduced promoter accessibility for genes involved in MSC differentiation, signaling, and cell-matrix adhesion, which strongly correlated with reduced transcriptional output as determined by transcriptome analysis and metabolic labeling with 5-ethynyl uridine. They then observed decreased H3 and H4 histone acetylation in the old MSCs. Because chromatin accessibility is controlled in part by chromatin acetylation, the decreased histone acetylation was hypothesized to be a probable cause of the altered differentiation and proliferation observed in old MSC. ¹⁰

To define a detailed picture of the chromatin changes, H3K27 acetylation, which correlates with active enhancers, was determined by CUT&RUN-seq. In this case, nuclei are isolated and bound to beads, a H3K27ac-specific antibody is bound to chromatin and a protein-A-micrococcal nuclease fusion protein is bound to the antibody. ¹⁵ The nuclease is activated by calcium, and then the released DNA is sequenced. Enhancers that lost H3K27ac marks correlated with genes involved in the function of the MSCs, suggesting that the loss of these chromatin marks was important for the altered differentiation potential of the old MSC. Those cells that gained H3K27ac marks were involved in glucose metabolism and metabolic signaling, suggesting an adaptation to the decreased glycolytic activity observed in old MSC.

Finally, they created a map of H3K27me3 marks, which are associated with gene silencing, by using CUT&RUN-seq. These marks did not change near promoters, but were slightly increased across gene bodies especially in genes involved in metabolism and cell adhesion. Interestingly, H3K27me3 marks were reduced in genes involved in neuronal differentiation. However, overall, they concluded that these genes were not much changed with old age. Bioinformatic analysis suggested that chromatin remodeling transcription factors CCCTC-binding factor (CTCF) was more active in old MSC and E2F5, which helps regulate proliferation, and DLX5, which helps regulate osteogenesis, were more active in young MSC, consistent with the differentiation data. ¹⁰

Acetyl-CoA Mediates a Novel Mechanism of Epigenome Remodeling

Pouikli et al. then attempted to characterize the mechanism that underlies the altered epigenome in old MSC. First they determined that the expression of the key histone acetyltransferases (HATs) CREB-binding protein (CBP) and GCN5 were unchanged, ruling them out as causative. Then they characterized acetyl-CoA levels. Acetyl-CoA is necessary for acetylation of many proteins, including histones, among many key roles it plays in metabolism. ⁹ Surprisingly, they discovered that acetyl-CoA levels were higher in old MSCs, which seemed to be inconsistent with the reduced acetylation they observed in old MSCs. Because acetyl-CoA plays a key role in lipid catabolism and anabolism, they investigated whether the neutral lipid content of old MSC was altered.

Indeed, they found lower amounts of neutral lipids by staining with a lipid-specific dye. They then determined that the lower neutral lipid levels were due to reduced levels of the ACC1 enzyme, which is involved in the initiation of fatty acid synthesis, and not due to any increase in beta-oxidation of fatty acids, which is involved in fatty acid catabolism. They concluded that acetyl-CoA was underutilized for fatty acid synthesis and instead accumulated in the cytoplasm or the mitochondria. ¹⁰ But ultimately they needed to explain the loss of acetylation in the nucleus, which could only be explained if somehow the acetyl-CoA was not present there.

It is important to understand that acetyl-CoA is combined with oxaloacetate to form citrate as part of the citric acid cycle. That reaction is reversible by such enzymes as ATP-citrate lyase (ACLY) in the cytoplasm, such that citrate can act as a kind of chemical carrier for acetyl-CoA throughout the cell. First, Pouikli et al. determined that levels of ACLY or acetyl-CoA synthetase (ACS) were unchanged in old MSCs, ruling out these obvious candidates. To test the possibility that mitochondrial transport was the culprit, they indirectly determined if high levels of acetyl-CoA were present in the mitochondria by assessing whether mitochondrial proteins were excessively acetylated, which occurs nonenzymatically when acetyl-CoA is at high levels.

Indeed there was a strong shift of the acetyl-lysine signal from the nucleus in young MSCs to the mitochondria of old MSCs, consistent with acetyl-CoA being trapped in the mitochondria, and would explain the altered chromatin acetylation and perhaps the stem cell dysfunction. ¹⁰ So how was the acetyl-CoA trapped in the mitochondria? Pouikli et al. realized that the obvious candidate was the citrate carrier protein CiC that transports citrate from the mitochondria to the cytoplasm. The cell can use citrate to carry acetyl-CoA. They determined that CiC levels were strongly reduced in old MSCs. ¹⁰ So, was this the cause of stem cell dysfunction? Yes, ectopic expression of mitochondrially localized CiC using a lentivirus vector increased CiC levels, rescued acetyl-CoA transport through citrate, restored histone acetylation in old MSCs, and improved their osteogenic differentiation potential, essentially rejuvenating the old MSCs. ¹⁰

As might be expected, treating young MSC with benzenetricarboxylate (BTA), a noncleavable analog of citrate widely employed in in vitro assays to block CIC transport activity, mimicked the acetylation patterns observed in the old MSCs as well as inhibiting osteogenic differentiation similar to that observed in old MSC.

Moreover, Pouikli et al. attempted a second way to rejuvenate the old MSCs. They added supplemental acetate to the medium. The ACS enzyme uses acetate as a substrate along with CoA and ATP to synthesize acetyl-CoA, which can then diffuse into the nucleus. Sodium acetate treatment in culture restored histone acetylation to normal levels in old MSCs. Acetate restored H3ac levels to young levels as assessed by CUT&RUN-seq and H3ac marks were present near key skeletal and cartilage differentiation genes such as BMP2 and MDK1. Importantly, acetate restored osteoblast differentiation frequency to levels observed in young MSCs. ¹⁰

But what is the mechanism by which CiC levels decrease? RNA expression levels of the SLC25A1 gene that encodes CiC are unchanged in old MSCs relative to young MSCs, as is the chromatin region of the SLC25A1 gene. There was no change observed in the levels of mitochondrial proteases, nor of their targets between young and old MSCs. In addition to proteolysis, mitochondrial protein levels are regulated by autophagy and lysosomal degradation. As might have been expected, an autophagy inhibitor, Bafilomycin A1 (BafA1), a macrolide antibiotic did not increase CiC levels. However, E64D, a cysteine protease and lysosomal degradation inhibitor, significantly raised the level of CiC expression. Consistent with this lysosomal-based degradation mechanism, CIC colocalized with LAMP2 a lysosomal marker in E64D-treated cells. ¹⁰

Because it is known that mitochondrial-derived vesicles (MDVs) are involved in lysosomal degradation of inner mitochondrial membrane proteins as an early step in mildly stressed mitochondria, ^{16 –18} Pouikli et al. used electron microscopy to detect a large number of double-membrane vesicles about 70–150 nm size near the mitochondria of old but not young MSCs. They hypothesize that this not only explains the mechanism of decreased CiC in old MSCs, presumably the MDVs carry CiC to lysosomes for destruction, but also suggests that the mitochondria of old MSCs are mildly stressed. Consistent with this contention, old MSCs had greater mitochondrial fragmentation. Of the two classes of MDVs, old MSCs contained TOMM20+/PDH− MDVs that were also CiC+ as determined by immunostaining, but not TOMM20−/PDH+ MDVs. ¹⁰

Are these observations relevant for humans? Old human MSCs derived from total hip arthroplasty or hemiarthroplasty expressed lower CiC levels, lower histone H3 acetylation levels, and lower cytoplasmic acetyl-CoA than young human MSCs similar to the murine MSCs. ¹⁰

Medical Implications

Metabolic restoration of a defect in citrate/acetyl-CoA transport was sufficient to rejuvenate the epigenome and stem cell function, specifically osteogenic differentiation, of old MSCs. Citrate was previously noted to be necessary for osteogenic differentiation through its role in production of alpha-ketoglutarate through the citric acid cycle. ¹⁹ However, it is clear from the study of Pouikli et al. that citrate transport to the cytoplasm and nucleus is necessary as well.

If acetate is capable of restoring MSC function, what might be a good way to increase it?

Alcohol consumption results in increased plasma acetate levels as the final step in ethanol oxidation. Perhaps this explains some of the basis of studies that have claimed that modest alcohol consumption is health-promoting, as it has been reported to raise histone acetylation levels in the brain. ²⁰