Immunometabolic Regulation of Anti-Tumor T-Cell Responses by the Oncometabolite

Since the discovery of gain-of-function mutations in the tricarboxylic acid (TCA) cycle enzyme isocitrate dehydrogenase (IDH) and the resulting accumulation of the metabolite D-2-hydroxyglutarate (D-2HG) in several tumor entities (such as glioma, acute myeloid leukemia (AML), and cholangiocarcinoma) about 10 years ago research has focused on the tumor cell-intrinsic consequences. D-2HG acting as an oncometabolite was shown to promote proliferation, anoikis, tumorigenesis, and differentiation block of hematopoietic cells in an autocrine fashion. Although the prognostic value of the different types of IDH mutations remains controversial the development of inhibitors against mutated IDH is flourishing. On the other hand, serum levels of D-2HG proved to be a more robust adverse prognostic marker in AML and glioma. Surprisingly, until recently only few studies on the paracrine effects of this oncometabolite on the tumor microenvironment with particular focus on the innate or adaptive immunity were available. Now, three recent publications focused on the paracrine effects of tumorderived D-2HG on T-cells in the context of AML and glioma. It was shown that T-cells are capable of efficiently taking up D-2HG in vitro, which was mirrored by 2HG-enriched T-cells exclusively found in samples from patients with IDH-mutated AML and glioma. Furthermore, all three studies describe an impairment of T-cell activation (although to different extents). The published effects could be at least partly attributed to metabolic alterations evoked by D-2HG influencing amongst others mTOR signaling, Hif-1α protein stability, the balance between aerobic glycolysis and oxidative phosphorylation, and the abundance of ATP (with according changes of AMPK activation). In the context of glioma it was further shown that IDH mutations and high D-2HG levels lead to reduced T-cell migration and consequently lowered T-cell infiltration at the tumor site. Moreover, two of the studies showed an increased frequency of FoxP3+ Tregs. Nevertheless, effects on downstream mechanisms and consequences have been differently addressed in the independent studies, and taken together the findings shed more light on the potentially targetable sites for Open Access Received: 13 May 2019 Accepted: 04 July 2019 Published: 08 July 2019 Copyright © 2019 by the author(s). Licensee Hapres, London, United Kingdom. This is an open access article distributed under the terms and conditions of Creative Commons Attribution 4.0 International License. Immunometabolism 2 of 15 improving therapeutic approaches. While the work by our group demonstrated profound impairments of Th17 polarization resulting from D-2HG-triggered Hif-1α protein destabilization the comprehensive analyses by Bunse et al. highlighted an inhibitory effect of D-2HG on the intracellular calcium signaling (downstream of the T-cell receptor) and an activation of AMPK signaling with resulting NFAT inhibition leading to the aforementioned activation defects. The summarized results of all three studies emphasized the importance of D-2HG not only as an autocrine but also a paracrine oncometabolite capable of shaping the tumor microenvironment in several facets adding to the common concept of tumor immune escape mechanisms. Those findings could lead to further improvements of the current targeted treatment strategies applied to patients harboring IDH mutations especially in view of the increasing importance of (combined) immunebased therapeutic approaches.

others and most frequently in gliomas [1], glioblastomas [2], and acute myeloid leukemia [3]. Those mutations which occur in the cytosolic IDH1 and the mitochondrial IDH2 isoform lead to a loss-of-function in terms of isocitrate conversion to α-ketoglutarate (ααKG). Importantly, they also lead to a gain-of-function reducing αKG to the D-enantiomer of 2HG (D-2HG) [4,5] (Figure 1). Numerous studies could demonstrate autocrine effects of D-2HG on the tumor cells through involvement in tumorigenesis/leukemogenesis and tumor progression. Acting as a competitive inhibitor of αKG-dependent dioxygenases it blocks the activity of histone and DNA demethylases as well as the TET family of 5mC hydroxylases [6,7]. Together with the inhibition of AlkB proteins prohibiting DNA damage repair mechanisms [8] this leads to profound alterations in the genetic and epigenetic landscape. While some studies demonstrate an inhibitory effect of D-2HG on prolyl hydroxylases leading to e.g., impaired collagen maturation [9] the prolyl hydroxylase PHD2 (fostering the proteasomal degradation of Hif-1α) was found to be activated by D-2HG [10]. Furthermore, IDH mutations and D-2HG have been reported to change the intracellular redox milieu by different mechanisms including alterations in the NADP/NADPH ratio [11] and depletion of glutathione [12]. Taken together, the complex interactional pattern aiding in proliferation and anoikis as well as differentiation block Immunometabolism. 2019;1:e190007. https://doi.org/10.20900/immunometab20190007 that enable tumorigenesis, progression, and dissemination are already quite well understood. Although the prognostic value of different IDH mutations in the distinct tumor entities is still a matter of debate, several inhibitors of mutant IDH are undergoing clinical trials [13][14][15] or have already been approved by the FDA such as Ivosidenib and Enasidenib for AML [16,17]. The occurrences and significances of IDH mutations with the consequential production and accumulation of D-2HG, as well as their resulting effects in the context of different tumor entities have been comprehensively reviewed elsewhere [18][19][20]. produced by neomorphic mutant IDH1 (cytosolic) or IDH2 (mitochondrial) through reduction of αKG thereby consuming NADPH. Together with depletion of GSH this leads to changes in the intracellular redox equilibrium. Newly produced D-2HG further inhibits αKG-dependent enzymes including DNA and histone demethylases, and the TET family of 5mC hydroxylases as well as AlkB proteins leading to alterations of the genetic/epigenetic landscape. While prolyl hydroxylases involved in collagen maturation are inhibited, PHD2 that fosters proteasomal degradation of Hif-1α is activated. Consequently, D-2HG is capable of inducing and promoting tumorigenesis and tumor progression. For further details see reviews [18][19][20]. Mitochondrial IDH3 converts isocitrate to αKG in a NADH-dependent manner. Mutations in the IDH3 isoform have not been reported yet.  has been shown to be highly elevated (up to 100-fold) not only in the IDH-mutant tumor cells but also in the tumor microenvironment [21] as well as serum [22] and cerebrospinal fluid [23]. The possibility of D-2HG detection in those biofluids (using mass spectrometry) [23][24][25] as well as in the tumor tissue (using MRI) [26] provides a valuable tool for non-invasive diagnostics and proved to be more robust in terms of prognostic value.

D-2HG IN THE TUMOR MICROENVIRONMENT
Surprisingly, studies that aimed to elucidate paracrine effects of D-2HG on the tumor microenvironment and particularly the innate and adaptive  [29].

IMMUNITY
The concept of cellular metabolism as a determinant for the phenotype, differentiation, and function of a cell has been shown and established throughout the past years. In T-cells both developmental status and differentiation program are characterized by distinct metabolic profiles optimized for their energetic demands under the given circumstances ( Figure 2A). While the cellular metabolism of naïve as well as memory T-cells is characterized by a modest, mixed-fuel oxidative phosphorylation (OXPHOS) for ATP production, activated effector T-cells undergo a glycolytic shift increasing their aerobic glycolysis activity (Warburg effect) while maintaining OXPHOS. Although aerobic glycolysis is less efficient in terms of ATP production per mole glucose it is faster and provides the necessary components for the biosynthesis of other intermediates (such as lipids, nucleic acids and proteins) needed in highly proliferating cells [30]. This increased utilization of glucose together with enhanced secretion of lactate is accompanied by an induction of glutaminolysis as well [31]. The changes in the metabolic phenotype are majorly governed by the enhanced/diminished expression of regulating, nutrient-sensing enzymes such as AMPK, mTOR, and HIF-1α. Distinct from naïve T-cells, memory T-cell metabolism is capable of a rapid recall of aerobic glycolysis after rechallenge [30]. While T-helper effector subsets (Th1, Th2, and Th17) are all favored by a glucose-rich environment due to reliance on aerobic glycolysis, regulatory T-cells are less dependent on glucose and are primarily characterized by fatty acid oxidation and OXPHOS [31].

DIRECT EFFECTS OF D-2HG ON T-CELL ACTIVATION
Recently, three publications investigating direct effects of exogenous, tumor-derived D-2HG in the context of AML and glioma on T-cells [34][35][36] have been issued almost simultaneously.
In fact, we and others could show that T-cells are capable of taking up exogenously supplied membrane-impermeable D-2HG [34,35].

DIRECT EFFECTS OF D-2HG ON T-CELL METABOLISM
The ability of D-2HG to impair TCR signaling and downstream effects as well as the fact that D-2HG is structurally similar to αKG raises the question if the obtained results are mediated on an immunometabolic level.
We found a dose-dependent, reversible induction of glucose utilization, which was accompanied by a reduced secretion of lactate indicating an increased flux of glucose to the TCA cycle instead of aerobic glycolysis. In a metabolomic profiling Bunse et al. also found elevated levels of hexoses in D-2HG treated CD4 + and CD8 + T-cells irrespective of the activation status.
Indeed, intracellular amounts of TCA cycle intermediates were found to be increased after D-2HG treatment in both studies. It can be speculated here that this may be caused by the physiological function of D-2HG dehydrogenase (D2HGDH) converting D-2HG to αKG leading to anaplerosis of the TCA cycle. Moreover, glutaminolysis as another anaplerotic switch for the TCA cycle was shown to be enhanced as exemplified by the increased expression of the rate-limiting enzyme glutaminase (GLS) [34] as well as elevated levels of intracellular L-glutamine [35]. On the other hand, expression of PDK1 (that blocks the entry of pyruvate into the TCA cycle) and LDHA (that converts pyruvate to lactate) was significantly reduced by D-2HG [34]. Accordingly, we found an increase of mitochondrial respiration and a decrease of aerobic glycolysis by D-2HG as measured by Mammalian target of rapamycin (mTOR) holds a key role in cellular metabolism regulating mitochondrial oxygen consumption [37], glucose uptake [38], and nutrient utilization [39] by dictating the expression and or activity of metabolic enzymes [40]. In fact, enhanced mTOR signaling was found in mutant IDH-expressing cell lines mediated by the inhibition of the αKG-dependent enzyme KDM4A [41]. In line with the observed metabolic phenotype, T-cells stimulated in the presence of D-2HG also showed enhanced mTOR signaling (illustrated by the phosphorylation of both mTOR and its down-stream target 4EBP1) despite unaltered MTOR gene expression. Actually, a negative impact of the oncometabolite on the protein expression of the endogenous mTOR inhibitor DEPTOR (via the described competitive inhibition of αKG-dependent KDM4A) was found to be the responsible driver [34,41]. In fact, D-2HG induced a mTORdependency as rapamycin renders T-cells vulnerable to D-2HG treatment in terms of glucose uptake and proliferation. Another metabolic master regulator and sensor of AMP/ADP:ATP ratio, AMPK [42], was found to be phosphorylated and thus activated in D-2HG treated T-cells due to a lack of total intracellular ATP [35]. Increased AMPK signaling has been shown to be amongst others responsible for catabolic processes such as enhanced glucose uptake and utilization, fatty acid uptake and oxidation, mitochondrial biogenesis and enhanced mTOR signaling [43].
Additionally, AMPK was shown to inhibit ornithine decarboxylase 1 (ODC1) activity, the rate-limiting enzyme of polyamine biosynthesis. As a consequence, intracellular levels of ornithine accumulated while its decarboxylation products putrescine and spermidine were reduced.
Interestingly, those products were proven to have potent immunostimulatory effects [44] and their lack can at least partly explain the observed defects in T-cell proliferation and activation [45].
The transcriptional activity of HIF-1α is also known to be regulated by mTOR [46]. However, D-2HG has been shown to destabilize Hif-1α protein by the induction of αKG-dependent prolyl hydroxylases leading to its proteasomal degradation [10]. Accordingly, we found a decrease of intracellular Hif-1α protein caused by D-2HG treatment in activated T-cells despite enhanced mTOR activity and unaltered HIF1A gene expression.
Consequently, Hif-1α target genes including the aforementioned PDK1 and CD4 + naïve and reduced CD4 + memory T-cell transcripts in a differential expression profiling of IDH wt and IDH mutant gliomas [35]. Yet, other factors apart from tumor-secreted D-2HG cannot be completely excluded as causatives for the differential expression.
Beyond that, Hif-1α has already been linked to the formation of Th17 cells [50]. Consequently, the already mentioned reduction of Hif-1α protein by D-2HG resulted in a diminished expression of target genes such as RORC and IL17A as well as in a reduction of IL-17 secretion. Accordingly, frequencies of Th17 cells were also reduced by D-2HG treatment in vitro.
Ultimately, in the context of AML Th17 frequencies were significantly reduced ex vivo in samples derived from patients carrying an IDH2 mutation as compared to non-mutated AML cases [34]. However, the role of Th17 cells for prognosis and progression of AML is still under debate. Importantly, all observed effects were shown to be specific for the Denantiomer of 2HG [34,35] and not evoked by epigenetic alterations in the T-cells [35].

CONCLUSIONS AND OUTLOOK
Taken together, the results of the three recent studies emphasize the importance of D-2HG not only as an autocrine but also paracrine oncometabolite capable of shaping the tumor microenvironment in several facets adding to the common concept of tumor immune escape mechanisms ( Figure 3). However, it becomes also clear that D-2HG certainly has entity-and niche-specific effects, which also depend on the developmental and activation status of the target T-cells. Interestingly, this can potentially add to the conflictive prognostic impact of IDH mutations in AML and glioma. In fact, mutant IDH-expressing glioma patients have a better prognosis compared to wild-type counterparts of the same histological grade [51]. This might seem contradictory to the presented findings of immunometabolic effects by D-2HG on T-cells and T-cell infiltration at the tumor site. However, it is speculated that IDH-wildtype gliomas might harbor additional mutations and alterations leading to the worse prognosis [28] which certainly needs more experimental evidence.
Immunometabolism. 2019;1:e190007. https://doi.org/10.20900/immunometab20190007 Moreover, there are indications that IDH-mutated malignant cells are surveyed by immune cells but locally elevated D-2HG concentrations at the tumor site break this immune surveillance [35]. Nevertheless, those recent findings will have a great impact especially in view of targeted therapies against mutant IDH in combination with immune-based approaches as shown by the improved outcomes using mutant IDH inhibitors together with PD-1 blockade in pre-clinical approaches. Additionally, the reduction of pro-inflammatory Th17 cells accompanied by an increased Treg frequency in the context of AML could highlight the importance of that oncometabolite for anti-tumor immunity as well as infections that are a major cause of mortality and morbidity during disease progression.  T-cells are capable to take up D-2HG via solute-carrier proteins. Intracellular accumulation of D-2HG leads to Hif-1α destabilization resulting in an increased influx of pyruvate to the TCA cycle and a diminished Th17 formation. The TCA cycle is further fueled by enhanced glucose utilization due to mTOR signaling activation and anaplerotic replenishment of αKG. At the same time, the total ATP pool in the cell is reduced leading to AMPK activation and dysfunctional ATP-dependent TCR downstream signaling. Both effects result in a prohibited calciumdependent NFAT signaling leading to an impairment of target gene expression. Additionally, AMPK inhibits polyamine biosynthesis leading to proliferation defects. All effects taken together lead to an impaired antitumor immunity.