As Aβ plaques accumulate in the brain, microglia are increasingly called upon to clean up the mess. This relentless phagocytic feasting requires enormous amounts of energy in the form of adenosine triphosphate, but by boosting glycolysis, the microglia fail to provide enough ATP, according to a study published October 6 in Nature Metabolism. Scientists led by Jie Zhang, Xiamen University in China, reported that hexokinase 2—a pivotal enzyme in glucose metabolism—ramps up in microglia in the AD brain. Curiously, deleting or inhibiting HK2 in mouse models of amyloidosis boosted microglial ATP production. The cells mustered the extra energy by transforming into fat-burning machines, the scientists found. They cranked up expression of lipoprotein lipase and other lipid metabolism genes. These lipid-utilizing microglia were supreme consumers of Aβ plaques and other neuronal debris, and spared AD mice from memory loss. The findings support the idea that metabolic shifts and functional states in microglia are intimately intertwined.

  • In the AD brain, microglia boost expression of the glycolytic enzyme hexokinase 2.
  • Blocking HK2 in mice improved microglial phagocytosis and plaque clearance.
  • HK2 inhibition switched on lipid metabolism, boosting ATP production.

“This new report is an exciting new addition to a growing number of studies implicating the regulation of microglia metabolism as a central driver in Alzheimer’s disease pathogenesis,” commented Lance Johnson of the University of Kentucky in Lexington.

Russell Swerdlow of Kansas University Medical Center in Kansas City wrote that the study answers important questions while raising many others. “I certainly hope it will help to further focus the field on the relevance of energy metabolism in AD, whether it is in neurons, astrocytes, or microglia.” (See full comment below.)

Glucose metabolism wanes in the brain with age, and even more so in neurodegenerative disease. As an energy-hogging organ, the brain consumes vast amounts of glucose, both via the fast-burning path of aerobic glycolysis in the cytosol, then via the slower, more lucrative burn of oxidative phosphorylation, which churns out even more ATP within the mitochondria. Yet cells also have an energy source that is wholly independent from glucose–fatty acids. Derived from the processing and oxidation of lipids, these provide another substrate for energy production via oxidative phosphorylation, and a slew of studies have implicated lipid processing and metabolism as central in microglial function and AD pathogenesis (Aug 2019 newsNov 2021 news; Sep 2022 conference news).

As Aβ plaques and other pathological insults build up in the brain with age, how do microglia rally their metabolisms to respond? First author Lige Leng and colleagues addressed this by first looking for expression of hexokinases, which catalyze the first rate-limiting step of glycolysis, namely converting glucose into glucose-6-phosphate. One isoform of this enzyme—HK2—was elevated in postmortem brain samples from people with AD, and in the 5xFAD mouse model of amyloidosis. The researchers pinned most of this uptick on microglia. Notably, in 5xFAD mice and in the human AD brain, microglia surrounding plaques expressed the highest levels.

Was this revved HK2 expression helpful, or harmful? To find out, the researchers conditionally deleted HK2 from microglia in wild-type and 5xFAD mice at 5 months of age, then examined them over the following two months. Wild-type animals suffered no obvious ill effects. The 5xFAD had a lower Aβ plaque load, more synaptic proteins, and better memory than 5xFAD controls. Leng found that their microglia extended processes, transforming from a hunched, amoeboid shape into a ramified one resembling microglia in wild-type mice. Two hexokinase inhibitors—lonidamine or 3-BP—had similar effects. In 5xFAD mice, and in cultured, human embryonic stem cell-derived microglia, 3-BP improved microglial phagocytosis, ramped up lysosomal degradation of internalized Aβ peptides, and dampened microglial secretion of pro-inflammatory cytokines.

How might hobbling glycolysis improve microglial function? Surprisingly, the scientists found the 3-BP-treated cells had more ATP, not less. In contrast, neurons or astrocytes suffered an energy deficit when glycolysis was blocked. To investigate why only microglia thrived, Leng examined their metabolomes and transcriptomes. Microglia treated with 3-BP as well as Aβ42 peptides ramped up lipid metabolism, producing more fatty acids, triglycerides, and acylcarnitine, while lactate, a product of glycolysis, declined.

Similarly, transcriptomics indicated that microglia had cranked up expression of genes involved in lipid metabolism, signaling a shift in their prime source of ATP from glucose to lipids. Upregulation of one gene in particular—lipoprotein lipase (Lpl)—explained much of the enhanced lipid processing in response to the HK2 blockade. Primarily expressed by microglia in the brain, lpl processes lipids into fatty acids. Its upregulation has been identified in microglial gene expression signatures in AD as well as other neurodegenerative diseases (Loving and Bruce, 2020). 

To understand how HK2 expression relates to lipid processing and microglial transcriptional states, the scientists analyzed data from a prior, single-cell RNA sequencing study of an APP knock-in mice model of amyloidosis (Sala Frigerio et al., 2019). In short, they found that microglia expressing the least HK2 expressed the most Lpl, and the strongest DAM signature. In all, the findings suggest that the switch to fat-burning mode would go hand in hand with a transition into disease associated states, which many researchers now view as beneficial (Sep 2022 conference news). As such, the authors proposed HK2 inhibition as a potential therapeutic strategy for AD.

Flipping the Fat Switch. In the AD brain, or in a mouse model of amyloidosis (left), microglia rev up expression of HK2, deriving much of their ATP from glycolysis. Blocking HK2 (right) switches microglia to lipid metabolism, which produces more ATP and enhances phagocytosis and reduces release of pro-inflammatory cytokines. [Courtesy of Choi and Mook-Jung, News and Views, Nature Metabolism, 2022.]

To Swerdlow, the findings highlight a fundamental point about metabolism: that the pathways that make or break energy are all interconnected. “If you alter one, others will change, and glucose, amino acid, and fatty acid-fueled bioenergetics are intertwined.” Swerdlow added that microglia are known to have different functional states that are determined by their bioenergetic infrastructures, including mitochondrial phenotypes. “In that regard, I would have liked to have seen data that informed the state of mitochondria in the microglia assessed,” he wrote.

In an editorial accompanying the paper, Inhee Mook-Jung and Hayoung Choi of Seoul National University College of Medicine made a related point. They noted that the health of mitochondria will be pivotal to the success of any lipid-metabolism promoting therapy, because fatty acid oxidase, the rate-limiting enzyme for metabolism of fatty acids, resides within these powerhouse organelles.

If lipids provide the best fuel for favorable microglial functions, why do microglia increasingly rev up expression of HK2 in the face of amyloidosis? While that remains unanswered, Zhang speculated that microglia, like other cells in the brain, sense waning glucose metabolism and ramp up HK2 to compensate.

Mark Mattson of Johns Hopkins University in Baltimore noted that the 5xFAD mice used in the study were fed a high-carb diet. “To begin with, they were not in a lipid-using state,” he said. In this glucose-dependent context, microglia may upregulate HK2 as a compensatory response to protect their own survival, he suggested. He wondered if HK2 would have ramped up to the same extent had the mice had been given a ketogenic diet or fed intermittently—both scenarios that promote a shift from glucose to lipid metabolism in the brain. “Under these conditions, perhaps HK2 inhibition would not have provided an added benefit,” he predicted.

The Aβ-induced uptick in HK2 expression observed by Leng dovetails with previous observations that microglia revitalize their use of glycolysis in response to Aβ accumulation (Baik et al., 2019). Other studies have found that while the brain gradually shifts from fast-burning glycolysis to the slower, more productive process of oxidative phosphorylation with age, some regions— for example, the Aβ-plaque prone default mode network—may rekindle glycolysis in response to growing stresses (Aug 2017 newsSep 2010 news).

Leng’s findings also jibe with a recent report that microglia contribute substantially to signals on FDG-PET scans, which measure glucose uptake in the brain (Oct 2021 news).

Róisín McManus of the German Center for Neurodegenerative Diseases in Bonn noted that Leng’s findings also resonate with studies implicating activation of the NLRP3 inflammasome in instigating glycolysis in microglia (Finucane et al., 2019). Aβ oligomers and aggregates are well-known triggers of the inflammasome, suggesting that the glycolytic shift seen in Aβ-exposed microglia could be facilitated by this inflammatory pathway (Lučiūnaitė et al., 2019). "Because HK2 depletion or inhibition was so effective at reducing Aβ pathology, this would have extensive downstream consequences, reducing long-term microglial activation by removing the very trigger (i.e. Aβ) that chronically activates these cells," she wrote (comment below).—Jessica Shugart

Comments

  1. This is a very interesting, as well as a very complex, paper. It is worth keeping in mind that its attempts to define mechanisms derive from models, and the fidelity between the models and common AD is unclear. Partly for this reason, it is hard to know what, in the way of microglial changes, is truly adaptive versus disease-driving. That point aside, the study is commendable on several fronts, including its spatial focus on brain microglia.

    A number of the reported findings make sense, such as the switch-over to lipid oxidation when confronted by a block in the proximal glycolysis pathway. It also emphasizes what I suspect are a number of fundamental points, the key one being that the pathways that make or break energy are all interconnected. If you alter one, others will change, and glucose, amino acid, and fatty-acid-fueled bioenergetics are intertwined. Another fundamental point, noted by others, is that microglia are known to have different functional states that are determined by their bioenergetic infrastructures. The changes reported in this study tend to fit in with what is known about microglial mitochondrial phenotypes, such as the M0, M1, and M2 mitochondrial phenotypes. In that regard, I would have liked to have seen data that informed the state of mitochondria in the microglia assessed, as well as more data on the state of the pentose phosphate shunt.

    In sum, this is a really nice study that answers some questions while raising others. I certainly hope it will help to further focus the field on the relevance of energy metabolism in AD, whether it is in neurons, astrocytes, or microglia.

  2. This study by Leng and colleagues uses in vitro and in vivo models of Alzheimer’s disease to examine the impact of Aβ on microglial metabolic activity, which is an underexamined area of neurodegenerative research. Interestingly, they found that Aβ pathology increases microglial Hexokinase 2 (HK2, a key enzyme in glycolysis), which impacts microglial migration and phagocytosis. Indeed, genetic depletion or pharmacological inhibition of HK2 resulted in a significant reduction in Aβ deposition throughout the brains of their 5X-FAD mouse model. Strikingly, just seven days of daily treatment with HK2 inhibitors also reduced Aβ levels throughout the brain. I found these results to be particularly interesting, because while many groups have shown the importance of metabolic intermediates on macrophage function (Mills et al., 2016; Liu et al., 2017), how microglia adapt their metabolic status in neurodegenerative diseases and the impact it has on cellular function is still relatively unknown.

    Leng and colleagues have addressed this experimental question really well and I commend their efforts to confirm the role of HK2 in AD. They set up an extensive dataset comparing in vitro with in vivo findings, even showing how HK2 changes with age and the severity of AD pathology, in addition to comparing murine and human disease models. They also ensured that genetic depletion of HK2 matched the pharmacological inhibition of this enzyme, using not just one but two different inhibitors of HK2. While real-time metabolic assays using an instrument such as the Seahorse Bioanalyzer also would have added value to their findings, Leng and colleagues used a combination of metabolomics and glucose isotope flux analysis to confirm that HK2 regulates the switch between glycolysis and fatty acid metabolism.

    HK-dependent glycolysis can mediate NLRP3 inflammasome activation and, correspondingly, NLRP3 activation can increase the levels of HK and other glycolysis enzymes (Moon et al., 2016; Finucane et al., 2019). Since Aβ is a well-established trigger of the microglial NLRP3 inflammasome (Halle et al., 2008; Heneka et al., 2013; McManus, 2022), it is surprising that the authors did not determine whether NLRP3, or its components, such as ASC or cleaved-caspase-1, were altered in their models. Leng et al. did observe a reduction in the protein levels of IL-1β, thus, whether this was due to altered cell priming or reduced inflammasome activity would be important to establish.

    Because HK2 depletion or inhibition was so effective at reducing Aβ pathology, this would have extensive downstream consequences, reducing long-term microglial activation by removing the very trigger (i.e., Aβ) that chronically activates the cells. Future work could examine this in detail to determine if daily inhibition of HK2 is necessary to maintain these effects, and whether targeting HK2 in advanced disease stages is still effective at preventing or rescuing cognitive decline. As one of the HK inhibitors used in this study is already in clinical trials (Lonidamine, as a cancer therapy), the results of extended datasets at later disease stages would be very exciting for scientists and patients alike.

    References:

    . Succinate Dehydrogenase Supports Metabolic Repurposing of Mitochondria to Drive Inflammatory Macrophages. Cell. 2016 Oct 6;167(2):457-470.e13. Epub 2016 Sep 22 PubMed.

    . α-ketoglutarate orchestrates macrophage activation through metabolic and epigenetic reprogramming. Nat Immunol. 2017 Sep;18(9):985-994. Epub 2017 Jul 17 PubMed.

    . mTORC1-Induced HK1-Dependent Glycolysis Regulates NLRP3 Inflammasome Activation. Cell Rep. 2015 Jul 7;12(1):102-115. Epub 2015 Jun 25 PubMed.

    . The NLRP3 inflammasome modulates glycolysis by increasing PFKFB3 in an IL-1β-dependent manner in macrophages. Sci Rep. 2019 Mar 11;9(1):4034. PubMed.

    . The NALP3 inflammasome is involved in the innate immune response to amyloid-beta. Nat Immunol. 2008 Aug;9(8):857-65. PubMed.

    . NLRP3 is activated in Alzheimer's disease and contributes to pathology in APP/PS1 mice. Nature. 2013 Jan 31;493(7434):674-8. Epub 2012 Dec 19 PubMed.

    . The Role of Immunity in Alzheimer's Disease. Adv Biol (Weinh). 2022 May;6(5):e2101166. Epub 2022 Mar 7 PubMed.

  3. This new report by Leng et al. is an exciting new addition to a growing number of studies implicating the regulation of microglia metabolism as a central driver in Alzheimer’s disease pathogenesis.

    The authors first show cell-specific expression of the various hexokinase (HK) isoforms within the mouse and human AD brain, with microglia predominantly expressing the HK2 isoform. They then conditionally knock out HK2 in microglia (and macrophages) using CX3CR1-Cre. Remarkably, this reduced plaque numbers by ~half when crossed to the 5XFAD model of amyloidosis. Importantly, this elimination of HK2 in microglia led to an increase in synaptic markers and a complete rescue of 5XFAD-associated cognitive deficits.

    Excitingly, treatment with HK inhibitors, and in particular the HK2-specific inhibitor 3-bromopyruvate (3-BP), showed similar protective effects. This prompted the authors to move in vitro, where they show that HK2 inhibition results in an increase in microglial phagocytosis (an energetically demanding process). Because hexokinase catalyzes the first step of glucose metabolism, inhibition of the enzyme should reduce ATP levels in the cell (as shown previously in multiple cell types). Paradoxically, Leng et al. show the opposite effect in primary microglia and BV2 cells. The authors go on to suggest that this increase in ATP is due to transcriptional activation of lipid metabolism and increased fatty acid beta-oxidation, perhaps downstream of an upregulation of the DAM gene lipoprotein lipase (Lpl). This is also quite interesting given that the basic tenets of immunometabolism state that “homeostatic” myeloid cells rely heavily on fatty acid beta-oxidation.

    This study comes as the field gains a new appreciation for the metabolic demands of microglia (Xiang et al., 2021), and adds to a growing body of literature implicating the microglial metabolic response to plaques as a contributing event in AD pathogenesis (Baik et al., 2019March-Diaz et al., 2021Grubman et al., 2021Nguyen et al., 2020). This includes studies implicating AD risk genes such as TREM2 and APO4E, further highlighting the importance of microglial metabolism in the onset and progression of AD (Konttinen et al., 2019Lee et al., 2022). 

    References:

    . Microglial activation states drive glucose uptake and FDG-PET alterations in neurodegenerative diseases. Sci Transl Med. 2021 Oct 13;13(615):eabe5640. PubMed.

    . A Breakdown in Metabolic Reprogramming Causes Microglia Dysfunction in Alzheimer's Disease. Cell Metab. 2019 Sep 3;30(3):493-507.e6. Epub 2019 Jun 27 PubMed.

    . Hypoxia compromises the mitochondrial metabolism of Alzheimer’s disease microglia via HIF1. Nature Aging. 1, 2021, pp.385–99. Nat. Aging.

    . Transcriptional signature in microglia associated with Aβ plaque phagocytosis. Nat Commun. 2021 May 21;12(1):3015. PubMed.

    . APOE and TREM2 regulate amyloid-responsive microglia in Alzheimer's disease. Acta Neuropathol. 2020 Oct;140(4):477-493. Epub 2020 Aug 25 PubMed. Correction.

    . PSEN1ΔE9, APPswe, and APOE4 Confer Disparate Phenotypes in Human iPSC-Derived Microglia. Stem Cell Reports. 2019 Oct 8;13(4):669-683. Epub 2019 Sep 12 PubMed.

    . APOE modulates microglial immunometabolism in response to age, amyloid pathology, and inflammatory challenge. bioRxiv. May 20, 2022 bioRxiv

  4. The manner in which specific energy substrates impact other aspects of cell biology is complicated and fascinating. It is now quite clear that many energy substrates—including glucose—boost mTOR, the activity of which suppresses autophagy and related processes. It is important to consider these "related processes," because many of the proteins required for autophagy also play a role in LC3-associated phagocytosis and LC3-associated endocytosis, of which the latter has been shown to participate in microglial degradation of Aβ (Heckmann et al., 2019). 

    One aspect of glycolysis that has gained interest in cancer biology is its contribution to the synthesis of amino acids. Glycine and serine levels are substantially depleted when glycolysis is restricted, and such amino acids are necessary for full mTORC1 activity (Takahara et al., 2020). Thus, glycolysis inhibition may suppress mTOR and thus activate autophagy or related processes to speed Aβ removal.

    Relevant to this hypothesis is the interaction of the APOE allele with autophagy-related processes (Simonovitch et al., 2016). Based on the ability of ApoE4 to suppress an enhancer element present in many autophagy-related genes (Parcon et al., 2018; Lima et al., 2020), it might be predicted that an oxidative shift in the glycolysis-OXPHOS balance would be more effective in individuals lacking an APOE ε4 allele. And, indeed, that has been borne out in human trials Reger et al., 2004; Henderson et al., 2009; Torosyan et al., 2018). 

    References:

    . LC3-Associated Endocytosis Facilitates β-Amyloid Clearance and Mitigates Neurodegeneration in Murine Alzheimer's Disease. Cell. 2019 Jul 25;178(3):536-551.e14. Epub 2019 Jun 27 PubMed. Correction.

    . Amino acid-dependent control of mTORC1 signaling: a variety of regulatory modes. J Biomed Sci. 2020 Aug 17;27(1):87. PubMed.

    . Impaired Autophagy in APOE4 Astrocytes. J Alzheimers Dis. 2016;51(3):915-27. PubMed.

    . Apolipoprotein E4 inhibits autophagy gene products through direct, specific binding to CLEAR motifs. Alzheimers Dement. 2018 Feb;14(2):230-242. Epub 2017 Sep 22 PubMed.

    . Electrochemical detection of specific interactions between apolipoprotein E isoforms and DNA sequences related to Alzheimer's disease. Bioelectrochemistry. 2020 Jun;133:107447. Epub 2019 Dec 23 PubMed.

    . Effects of beta-hydroxybutyrate on cognition in memory-impaired adults. Neurobiol Aging. 2004 Mar;25(3):311-4. PubMed.

    . Study of the ketogenic agent AC-1202 in mild to moderate Alzheimer's disease: a randomized, double-blind, placebo-controlled, multicenter trial. Nutr Metab (Lond). 2009;6:31. PubMed.

    . Changes in regional cerebral blood flow associated with a 45 day course of the ketogenic agent, caprylidene, in patients with mild to moderate Alzheimer's disease: Results of a randomized, double-blinded, pilot study. Exp Gerontol. 2018 Oct 1;111:118-121. Epub 2018 Jul 10 PubMed.

Make a Comment

To make a comment you must login or register.

References

News Citations

  1. Newly Identified Microglia Contain Lipid Droplets, Harm Brain
  2. Do Lipids Lubricate ApoE's Part in Alzheimer Mechanisms?
  3. Shooting Themselves in the Foot? Microglia Block “Good” State with ApoE4
  4. As Youth Fades, So Does the Fire of Glycolysis in the Brain
  5. Brain Aβ Patterns Linked to Brain Energy Metabolism
  6. What FDG PET ‘Sees’ in AD: Angry Microglia, Not Just Neurons

Research Models Citations

  1. 5xFAD (C57BL6)
  2. APP NL-G-F Knock-in

Paper Citations

  1. . Lipid and Lipoprotein Metabolism in Microglia. Front Physiol. 2020;11:393. Epub 2020 Apr 28 PubMed.
  2. . The Major Risk Factors for Alzheimer's Disease: Age, Sex, and Genes Modulate the Microglia Response to Aβ Plaques. Cell Rep. 2019 Apr 23;27(4):1293-1306.e6. PubMed.
  3. . A Breakdown in Metabolic Reprogramming Causes Microglia Dysfunction in Alzheimer's Disease. Cell Metab. 2019 Sep 3;30(3):493-507.e6. Epub 2019 Jun 27 PubMed.
  4. . The NLRP3 inflammasome modulates glycolysis by increasing PFKFB3 in an IL-1β-dependent manner in macrophages. Sci Rep. 2019 Mar 11;9(1):4034. PubMed.
  5. . Soluble Aβ oligomers and protofibrils induce NLRP3 inflammasome activation in microglia. J Neurochem. 2019 Dec 23;:e14945. PubMed.

Further Reading

Papers

  1. . Microglia energy metabolism in metabolic disorder. Mol Cell Endocrinol. 2016 Dec 15;438:27-35. Epub 2016 Sep 28 PubMed.

Primary Papers

  1. . Microglial hexokinase 2 deficiency increases ATP generation through lipid metabolism leading to β-amyloid clearance. Nat Metab. 2022 Oct;4(10):1287-1305. Epub 2022 Oct 6 PubMed. Correction.
  2. . Lipid fuel for hungry-angry microglia. Nat Metab. 2022 Oct;4(10):1223-1224. PubMed.