Depletion of Fumarate Hydratase, an Essential TCA Cycle Enzyme, Drives Proliferation in a Two-Step Model

Simple Summary

Fumarate hydratase (FH) is an evolutionary conserved TCA cycle enzyme and has a moonlight function in the DNA damage response (DDR). FH has a contradictory function, as it is pro-survival through its role in the TCA cycle, yet its loss can drive tumorigenesis. Here, we solve this contradiction and show that in both non-cancerous (HEK-293T) and cancerous cell lines (HepG2), the cell response to FH loss is separated into two distinct time frames. During the early stages of FH loss, cell proliferation and DNA damage repair are inhibited. However, over time, cells overcome the FH loss and form proliferative knockout stable clones. Furthermore, we discovered that the impaired DDR induces mutations enriched in central signaling pathways, such as JAK/STAT3, allowing the cells to circumvent FH and TCA cycle loss.

Abstract

Fumarate hydratase (FH) is an evolutionary conserved TCA cycle enzyme that reversibly catalyzes the hydration of fumarate to L-malate and has a moonlight function in the DNA damage response (DDR). Interestingly, FH has a contradictory cellular function, as it is pro-survival through its role in the TCA cycle, yet its loss can drive tumorigenesis. Here, we found that in both non-cancerous (HEK-293T) and cancerous cell lines (HepG2), the cell response to FH loss is separated into two distinct time frames based on cell proliferation and DNA damage repair. During the early stages of FH loss, cell proliferation rate and DNA damage repair are inhibited. However, over time the cells overcome the FH loss and form knockout clones, indistinguishable from WT cells with respect to their proliferation rate. Due to the FH loss effect on DNA damage repair, we assumed that the recovered cells bear adaptive mutations. Therefore, we applied whole-exome sequencing to identify such mutated genes systematically. Indeed, we identified recurring mutations in genes belonging to central oncogenic signaling pathways, such as JAK/STAT3, which we validated in impaired FH-KO clones. Intriguingly, we demonstrate that these adaptive mutations are responsible for FH-KO cell proliferation under TCA cycle malfunction.

1. Introduction

Fumarate hydratase (FH) is a TCA cycle enzyme that catalyzes the hydration of fumarate to L-malate (1). Despite its essential role in central carbon metabolism, FH is inactive or absent in several tumors, such as hereditary leiomyomatosis and renal cell cancer (HLRCC) (2). This lack of functional FH results in the significant rewiring of the cellular metabolism as the cells accumulate fumarate and demonstrate a fermentative metabolism, which includes reduced oxygen consumption and elevated lactate secretion (3). Accumulated fumarate in cells functions as an “oncometabolite” (4), which inhibits α-ketoglutarate (αKG)-dependent dioxygenases and histone DNA demethylases (5) and stabilizes the hypoxia-induced factor 1 alpha (HIF1α) (6). Moreover, accumulated fumarate induces the epithelial-mesenchymal transition (EMT) program (7) that is associated with high-grade tumors (8) by altering the epigenetic landscape of these cells. Together, FH has contradictory cellular functions, as it is pro-survival through its role in the TCA cycle and functions as a tumor suppressor in tumorigenesis. However, a comprehensive cellular mechanism to explain these opposing roles of FH is still missing.

In addition to its localization in the mitochondria, FH is also present in the cytosol/nucleus, where it participates in the DNA damage response (DDR) (9). In the nucleus, FH converts malate to fumarate, which, upon the induction of DNA breaks, inhibits the KDM2B-mediated H3K36me2 demethylation, resulting in non-homologous end-joining repair (10). However, the function of fumarate in the DDR is still unclear, as it also suppresses the DDR by inhibiting KDM4B (11). Here, we show that the effect of FH loss in cells is divided into two time frames. In the early stages of FH loss, the cells are under stress, do not proliferate, and inhibit DDR. Consequently, this loss causes the accumulation of mutations in the FH-KO cells, enriched within genes encoding components of central oncogenic pathways. These adaptive mutations provide the long-term FH-KO cells with the capability to proliferate even in the absence of a functional TCA.

2. Materials and Methods

2.1. Cell Lines and Cell Culture

The cell lines HEK-293T and HepG2 were procured from the American Type Culture Collection (ATCC) and were cultured in DMEM supplemented with 10% FBS (Biological industries, Kibbutz Beit-Haemek, Israel)). All cells were cultured at 37 °C with 5% CO₂.

2.2. Cell Lysis and Immunoblotting

Cells were rinsed twice with ice-cold PBS and lysed with RIPA lysis buffer (20 mM Tris pH 7.4, 137 mM NaCl, 10% glycerol, 1% Triton X-100, 0.5% deoxycholate, 0.1% SDS, 2.0 mM ethylenediaminetetraacetic acid (EDTA, pH 8.0), complete EDTA free Protease inhibitor cocktail tablets (1 tablet per 50 mL RIPA) (Sigma-Aldrich Israel, Rehvot, Israel, #11836170001) and phosphatase inhibitor cocktail (100×) (Bimake, Houston, TX, USA, # BI5001-A and B). The lysates were cleared by centrifugation at 13,000 rpm at 4 °C in a microcentrifuge for 10 min. The protein concentration was determined by Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad, Hercules, CA, USA, #5000006). Proteins were denatured by adding SDS sample buffer (5×) and boiling for 5 min, then resolved by 10% SDS-PAGE, transferred onto a 0.45 µm PVDF membrane (Immobilon-P Transfer membrane) (Merck Millipore, Carrigtwohill, Ireland, Co, #IPV00010), and probed with the appropriate antibodies.

2.3. Exome Sequencing Preparation and Analysis

The DNA was extracted from HEK-293T cells using DNeasy Blood & Tissue Kit (QIAGEN, Hilden, Germany, (#69504)). The samples were prepared and sequenced in Bio Basic Asia Pacific Pte Ltd (Singapore) and analyzed by our faculty bioinformatics unit. Specifically, we applied Mutect2 (from the GATK tools) for calling somatic short mutations (single nucleotide (SNA) and insertions/deletions). Mutect2 uses a Bayesian somatic genotyping model and the assembly-based machinery of GATK’s HaplotypeCaller. We used the “Tumor with matched normal” mode with a joint analysis of multiple samples. Samples of FH KO clones were all defined as tumor samples, while the WT cell line was defined as normal. The resulting VCF was further filtered using the FilterMutectCalls tool with default parameters. Variants that passed the filter were annotated by the wANNOVAR online annotation tool [1], and only those found to be located within exonic regions were kept.

2.4. Antibodies

Antibodies were obtained from the following sources: FH (#4567) and Phospho-Histone H2A.X-Ser139 (20E3 (#9718), p-STAT3-Tyr705 (#9145), STAT3 (#9139), Rabbit β-actin (#4970), mouse β-actin (#3700) from Cell Signaling Technology (Danvers, Ma), and HRP-labeled anti-mouse (Jackson ImmunoResearch Laboratories, (West Grove, Pa), 115-035-003) and anti-rabbit (111-035-144) secondary antibodies from Jackson ImmunoResearch. For immunofluorescence assays: Goat-anti mouse-Alexa fluor 647 (Abcam, (Cambridge, UK), #ab150119), Donkey-anti rabbit-Rhodamine red X (Jackson ImmunoResearch Laboratories, 711-295-152), and Donkey anti-rabbit-Alexa fluor 488 (Jackson ImmunoResearch Laboratories, 711-545-152).

2.5. Virus Production

HEK-293T cells were co-transfected with the pcw107-V5 plasmid expressing CA-STAT3 (Addgene, (Watertown, MA, USA), #64611), VSV-G envelope plasmid, and Δvpr lentiviral plasmid using X-TremeGene 9 Transfection Reagent (MERCK, # 6365779001). The supernatant containing the virus was collected 48 h after transfection and spun for 5 min at 400 g to eliminate cells.

2.6. Fluorescence Microscopy

The HEK-293T (60,000 cells) and HepG2 (50,000 cells) were seeded on polylysine-coated glass coverslips in 12-well tissue culture plates. After 24 h, the slides were rinsed twice with PBS and fixed with 4% paraformaldehyde in PBS for 15 min at room temperature, followed by quenching with ammonium chloride (1% in PBS). The slides were then rinsed three times with PBS, and cells were permeabilized with 0.1% Triton X-100 in PBS (PBS-T) for 10 min. Next, the cells were gently rinsed three times with PBS and incubated for 1 h in the blocking buffer (1% BSA in PBS-T), followed by primary antibodies for 2 h (FH and γH2AX 1:200 each in 1%BSA in TBST) at room temperature. Next, the cells were rinsed three times with PBS, incubated with secondary antibodies (diluted 1:200 in 1% BSA in TBST), and Phalloidin-iFluor 555 Reagent (Abcam, ab176756) for 1 h at room temperature in the dark, and washed three times with PBS. Slides were mounted on glass coverslips using Vectashield (Vector Laboratories, Newark, CA, USA# H-1000). The cells were imaged on Nikon Spinning Disk/high content screening system (Nikon, Tokyo, Japan) using the 60× (NA = 0.95, dry, CFI Plan-Apochromat Lambda) and 100X (NA = 1.4, oil, Plan-Apochromat). This microscope is equipped with a spinning disk Yokogawa W1 Spinning Disk, 2 SCMOS ZYLA cameras, and 405 nm, 488 nm, 561 nm, and 638 nm laser. While the 20× (NA = 0.5, dry, WD 2.1 mm, pH1) images were acquired using the Eclipse NI-U upright microscope (Nikon), equipped with DS-QI2 MONO-cooled digital microscope camera 16 MP. The image analyses were performed using the NIS Elements software package for multi-dimensional experiments and exported as 16-bit. The pictures were slightly adjusted (levels) using Adobe Photoshop (San Jose, CA, USA).

2.7. CRISPR/Cas9-Mediated Knockout Cell Lines

For CRISPR-Cas9-mediated genome editing we used pSpCas9(BB)-2A-Puro (PX459) V2.0 (Addgene Plasmid #62988), and PX458-GFP (Addgene Plasmid #48138). Transfected cells for the long-term experiments were then subjected to single-cell isolation by limiting dilution in 96-well plates. Editing of the FH locus was confirmed by assessing protein levels by Western Blot. Primers used for cloning non-targeting control (NTC), NTC-sgRNA 5′ caccgGCGCTTCCGCGGCCCGTTCAA 3′ and 5′ aaacTTGAACGGGCCGCGGAAGCGg; 3′; FH-gRNA1 5′ caccgACATGATCGTTGGGATGCAC 3′ and 5′ aaacGTGCATCCCAACGATCATGTc 3′; FH-gRNA2 5′ caccgGGTATCATATTCTATCCGGA 3′ and 5′ aaacTCCGGATAGAATATGATACCc 3′; FHgRNA3 5′ caccgAATAATGAAGGCAGCAGATG 3′ and 5′ aaacCATCTGCTGCCTTCATTATTc 3′; DPYD-gRNA1 5′ caccgTGTGCTCAGTAAGGACTCGG 3′ and 5′ aaacCCGAGTCCTTACTGAGCACAc 3′.

2.8. Cell Proliferation Assay

For the IncuCyte live-cell imaging system (Sartorius, Goettingen, Germany), the cells were seeded in 96-well plates at a density of 800 cells/well (HEK-293T or HepG2, WT, FH-KO short time, and stable clones). The cell proliferation rate was monitored every two hours at intervals using the IncuCyte live-cell imaging system. After 48 h for HEK-293T or 70 h for HepG2, the plates were removed, and the proliferation assay results were compiled from six wells from each group. Then, the data were analyzed and plotted as a graph. For Cell Titer-Glo assays, the cells were seeded in white 96-well plates (Greiner, Kremsmünster, Austria) at a density of 800 cells/well. Cell viability was assessed with Cell Titer-Glo (Promega, Madison, WI, USA #G7571) after 1, 3, and 5 days following seeding, and luminescence was measured with Cytation 3 Multi-Mode Reader (Biotek, Winooski, VT, USA).

2.9. RNA Preparation, qPCR Analysis, and RNA-seq

Total RNA was isolated from cells using the NucleoSpin^® RNA Kit (MACHEREY-NAGEL, Duren, Germany), and reverse transcription was performed using the qScript cDNA Synthesis Kit (Quantbio, Beverly, MA, USA). The resulting cDNA was diluted in DNase-free water (1:10) before quantification by real-time quantitative PCR. The mRNA transcription levels were measured using 2× qPCRBIO SyGreen Blue Mix Hi-ROX (PCR Biosystems, London, UK) and StepOnePlus (Applied Biosystems, Waltham, MA, USA). The data are expressed as the ratio between the expression level of IL-6 mRNA and that of Actin. The primers used for the qPCR analysis were obtained from Integrated DNA Technology (

Table 1. qPCR primers used in this study and their sequences.

Gene	Forward	Reverse
IL-6	ACTCACCTCTTCAGAACGAATTG	CCATCTTTGGAAGGTTCAGGTTG
β-Actin	CCAACCGCGAGAAGATGA	CCAGAGGCGTACAGGGATAG
VIM	ACCCGCACAACGAGAAGGT	ATTCTGCTGCTCCAGGAAGCG
TWIST	TGCGGAAGATCATCCCCAGG	ATTCTGCTGCTCCAGGAAGCG
MMP9	TGTACCGCTATGGTTACACTCG	GGCAGGGACAGTTGCTTCT

):

For RNA-seq analysis, the RNA quality was determined using a Tap station. Only RNA samples with 9–10 RIN were subjected to sequencing at our faculty Genomic Applications Laboratory of the Core Research Facility. The RNA-seq results were analyzed by our faculty bioinformatics unit.

2.9.1. Metabolite Supplementation

FH was knocked out from HEK-293T cells, and after 48 h, 1000 cells/well were seeded in white 96-well plates for an additional 24 h. Following that, the cells were treated with DMSO (control), 250 µM diethyl malate (Sigma Aldrich, #W237418), or 250 µM mono-ethyl fumarate (Sigma Aldrich, #12842225G). The metabolites were added twice a day for a total of three days. Cell viability was measured using the Cell Titer-Glo kit, and luminescence was measured with Cytation 3 Multi-Mode Reader.

2.9.2. Comet Assay

The assay was conducted according to the Abcam comet assay kit (#ab238544) protocol. Briefly, 35 μL of suspended cells (1 × 10⁶/mL) was mixed with 150 μL agarose, and 35 µL of this mixture was added to each agar-precoated slide. After incubation, the slides were immersed in pre-chilled lysis buffer for one hour and replaced with a pre-chilled alkaline solution (300 mM NaOH, 1 mM EDTA-Na₂, pH 13) for 30 min, all at 4 °C and in the dark. Then the slides were electrophoresed in TBE buffer for 20 min at 35 volts (1 V/cm). After the run, the slides were dried, immersed twice with pre-chilled DI-H₂O for 2 min, dehydrated in cold 70% ethanol for 5 min, stained with vista green DNA Dye (from the kit) at room temperature for 15 min, and analyzed using a fluorescence microscope (FITC filter) (Nikon, Tokyo, Japan, Eclipse Ni). Quantification was performed using the software Comet Assay IV to determine the average comet tail lengths.

2.9.3. Statistical Analysis

Data are shown as mean ± SD from at least three independent biological experiments. All the statistical analyses were performed using the R (version 4.2.0) or GraphPad Prism (version 8.0) statistical analysis programs. In most figures, the p-values measured the difference between the indicated samples and quantified using the unpaired two-tailed Student’s t-test. Data distribution was assumed to be normal, but this was not formally tested. The significance of the mean comparison is present in each figure. The pie chart was generated using the R package ggplot2, and the p-value was calculated using Fischer’s exact test. Chromosome with mutation location was generated using R package RIdeogram, and the gene list was taken from Piovesan St. al. [2]. One sample proportion test was applied to calculate the significance of the mutation number for each chromosome. Benjamini-Hochberg procedure was used to normalize the result. Information about the chromosomal mutations, including the mutation site, was extracted from HEK293T cells’ exon sequence results, then plotted on the chromosome map using the R package RIdeogram. Mutations with a frequency of four or more are assumed to be a frequent mutation.

3. Results

3.1. Short-Term FH Loss Inhibits Cell Proliferation

We first sought to determine the impact of FH loss on the cell proliferation rate and, in particular, to characterize the difference between short-term and long-term effects. We found that after five days of CRISPR-Cas9 induction in HEK-293T cells (non-cancerous human embryonic kidney cell line) and one day with the HepG2 (hepatocellular carcinoma cell line), FH expression levels were almost completely abolished (Figure 1a). Nevertheless, after a few days of the CRISPR-Cas9 induction, we observed that in the polyclonal population of FH-KO cells, the enzyme expression gradually returned to WT levels. These findings are consistent with the role of FH in proliferation, as cells that have escaped the knockout and still express FH have a growth advantage over those lacking the enzyme (FH-KO). This conclusion is further supported by the short-term knockout of the DPYD, a non-essential metabolic gene [3], which, even after seven days, retained a low expression level (Figure S1a). Remarkably, despite this impact on cell proliferation, we could generate stable FH-KO clones (FH-Clone-1 and 2) for both cell lines (Figure 1a), suggesting a long-term cellular mechanism that enables cells to overcome FH loss.

We conducted real-time monitoring of the effect of FH loss on cell growth using the Incucyte Live-Cell analysis system. We found that in both cell lines, the proliferation rate of the two different short-term FH-KO cells was significantly lower than WT cells (Figure 1b,c and Figure S1b). Moreover, supplementing the medium with fumarate (FH substrate) had no effect (Figure S1c), while malate (FH product) supported FH-KO cell growth (Figure 1d). Therefore, the initial response of cells to FH loss is inhibition of proliferation due to TCA cycle impairment. Nevertheless, several short-term FH-KO cells can overcome this proliferation obstacle and convert into long-term growing clones. Collectively, we reveal that the impact of FH loss on cell proliferation is divided into two phases, based on the time from CRISPR-Cas9 induction: short- and long-term. Furthermore, the ability of cells to proliferate without a functional FH demonstrates that these cells have adapted to the metabolic change.

3.2. Short-Term FH Loss Induces DNA Damage

Recent studies indicated that FH cellular function is more complex than initially thought, as it is also localized in the nucleus, where it functions in the DDR [4,5]. Indeed, the immediate effect of FH KO (short-term) in both HEK-293T (Figure 2a) and HepG2 (Figure 2b) was stepwise induction in H2A.X phosphorylation (γ-H2AX), which is negatively correlated with FH loss. In contrast, this phosphorylation, a sensitive marker of DNA double-strand breaks [6], is absent in the long-term FH-KO cells, suggesting that DNA damage is impaired at the early stages of FH loss. This finding supports our notion that DNA damage functions as an integral part of the cell adaptation to FH loss.

Co-staining of HEK-293T and HepG2 cells with FH and γ-H2A.X antibodies indicates that the DNA damage is specific for the short-term KO cells (Figure 2c,d). Fascinatingly, in this polyclonal population (short-term), only cells that appear as FH-KO (not stained with FH antibodies) exhibit γ-H2A.X foci. In line with these results, we show by comet assay that the short-term FH-KO cells demonstrate a significantly high level of DNA breaks versus WT and long-term clones (Figure 2e). Since the long-term FH-KO cells do not display DNA damage, despite the absence of functional FH, it suggests that the growth of these cells is a product of an adaption process. Hence, identifying the mutations induced by the FH-KO-dependent DNA damage impairment will help understand the adaption process.

3.3. Whole Exome Sequencing Analysis of Long-Term FH KO Clones

To assess the consequence of the impaired DNA damage response, we subjected 11 HEK-293T FH-KO clones and one WT (as a reference) to whole-exome sequencing (Table S1). The mutations identified in the different clones were compared to WT cells using Mutect2, a somatic variant caller [7] (Figure 3a, illustration). This analysis resulted in 14,959 FH-KO-specific variants. Next, we filtered out the variants localized in intronic regions using the wANNOVAR online annotation tool [1]. To avoid the possibility of non-specific variants, we first retained only those with more than ten reads. Then, to increase the probability of true hits, we focused only on the variants that exhibited a high percentage of reads from the total. To this end, for every variant in each FH-KO clone, we calculated the percentage of alteration out of the total reads of this sequence. This analysis allowed us to identify those with at least 30% variation reads, about 12% of the total variants (Figure S2b).

Next, we removed all synonymous mutations, resulting in 1437 variants. We found that the majority of the genes in the list had one mutation per clone (968 variants) (Figure 3b). However, several genes had recurrent mutations in multiple clones, including six mutated in all samples (MTRF1L, IGFALS, CTSD, BMP1, PCDHB13, and FAM196A). Moreover, 97 genes were mutated in two or more locations (Figure 3c). Interestingly, we observed that the mutated genes did not distribute randomly over the chromosomes, as several are significantly more abundant on chromosomes 9, 11, and 19 (Figure S2c). These results portray a novel perception of how FH loss induces genomic mutations, which presumably provide the cells with the ability to overcome the FH-related proliferation impairment.

3.4. Long-Term FH Knockout Clones Harbor Mutations in Central Metabolic and Signaling Pathways

We inquired whether mutations detected in the previous section are associated with common biological processes. Using the Kyoto encyclopedia of genes and genomes (KEGG) pathway database (https://www.genome.jp/kegg/pathway.html (accessed on 1 September 2021)) [8], we discovered that these FH loss-associated mutations are present in genes of multiple metabolic pathways, such as central carbon metabolism, nucleotide biosynthesis, and amino acid metabolism (Figure 4a). Moreover, we identified that FH loss significantly induces mutations in signaling pathways (Figure S3a,b), indicating that they provide a growth advantage. Specifically, these pathways include central pro-survival signaling cascades such as the MAPK, chemokine, Wnt, and JAK/STAT3 signaling (Figure 4b). By applying the KEGG mapper color tool, we present the mutated genes in MAPK-signaling, chemokines, and JAK/STAT3 pathways (Figure S3c). These findings highlight the enrichment of FH-loss-induced mutations in central signaling pathways, indicating that these genomic impairments are not random.

3.5. FH Loss Results in STAT3 Signaling Impairment

To further examine the effect of FH loss on signaling pathways, we chose the well-known JAK/STAT3, which may provide a step in the adaption process. We demonstrated that this signaling pathway regulates HEK-293T growth since expressing a constitutively activated STAT3 (A662C, N664C, V667L, CA-STAT3) [9,10] significantly inhibited its proliferation rate (Figure 5a). Accordingly, FH-long-term loss inhibited STAT3 activity, as reflected by a reduction in tyrosine 705 phosphorylation (Y705) (Figure 5b) and repression of its downstream target, IL-6 (Figure 5c). Moreover, this change in pSTAT3 levels is reduced shortly after FH loss in both HEK-293T and HepG2 (Figure 5d). Finally, RNA seq analysis (Table S2) of the WT and FH-KO clone, followed by Gene Set Enrichment Analyses (GSEA) [11], demonstrated that FH loss significantly reduces the “Hallmark of IL6-JAK-STAT3 signaling”. Additionally, we noticed other STAT3-related hallmarks, such as “Hallmark of the epithelial-mesenchymal transition” and “Hallmark of KRAS signaling-up,” were reduced (Figure 5e). This analysis was further validated by qPCR, where the known STAT3 targets, vimentin (VIM), TWIST, MMP9, and IL-6 [12], are significantly downregulated in FH-clones 1 and 2 (Figure 5f). Interestingly, analysis of patient data using the Cbiopotal web-based (https://www.cbioportal.org (accessed on 1 May 2022)) tool identified two tumor samples; MPCPROJECT_0130 (Prostate Adenocarcinoma) and SD5038 (Cutaneous Melanoma), in which both IL-6 and FH are deleted. Collectively, we demonstrated that in FH-KO cells, the STAT3 signaling is impaired, which restored proliferation.

4. Discussion

The FH enzyme is a critical component of the TCA cycle and thus plays a vital role in cell metabolism and proliferation. However, cells are capable of proliferating without this enzyme, and moreover, mutation or deletion of FH is an apparent driver of various cancers [13]. To illuminate this apparent paradox, we show that the immediate impact of FH loss is the inhibition of cell proliferation. Yet, stable knockout clones that we generated display insignificant differences in their proliferation rate compared to the WT cells. In this study, we set out to resolve this paradox and reveal the molecular mechanism by which cells shift from FH-dependent to FH-independent proliferation.

Addressing this paradox, we describe a comprehensive model that differentiates between the short-term and long-term effects of the FH knockout. We determine that the connection between these two phases of proliferation lies in the DNA damage response, which involves the moonlighting function of FH. In WT cells, FH participates in the TCA cycle, as well as in the DNA damage response [14]. However, upon the initiation of FH loss, the halted cells also become deficient in their ability to repair their DNA and therefore are prone to the accumulation of mutations (Figure 6). We found that these are not random mutations, as they are enriched in genes encoding signaling molecules. The JAK/STAT3 pathway is an example of a signaling cascade that is impaired in these long-term FH-stable clones.

Over the past twenty years, several mechanisms have been proposed to address the FH role in cancer, mainly focusing on accumulated fumarate due to enzyme loss [15,16]. In renal tumors, accumulated fumarate functions as an oncometabolite [17], which stimulates cancer-associated processes. For example, accumulated fumarate inhibits PH (prolyl-hydroxylase) activity and thus stabilizes HIF1α at normal oxygen tensions [18,19]. Additionally, under acidic conditions, fumarate interacts with cysteine thiol groups, forming a post-translation modification termed “succination”, which regulates the activity of pro-oncogenic signaling factors such as KEAP1 (Kelch Like ECH Associated Protein 1) [13]. We found that fumarate intracellular levels did not change significantly at the early stage of FH loss, and the addition of mono-ethyl fumarate esters does not alleviate short-term FH KO effects. This indicates that the proliferation inhibition is not due to vast intracellular oncometabolite accumulation. Previously, we demonstrated that FH is also found in the nucleus, where it functions in the DDR [4,5]. Specifically, FH is localized to the damage site, where it produces fumarate from malate [20,21]. However, it is becoming clear that fumarate may have both pro- and anti-DNA damage repair roles [22,23]. We do not wish to dwell on this question regarding the fumarate function in the DDR, as long as its role in this process is established.

The execution of the EMT program induces significant changes in cell characteristics, such as their morphology and migratory capabilities [24]. Notably, the EMT program outcome is reversible, as the cells can regain their epithelial characteristics by inducing the mesenchymal-epithelial transition program (MET) [25]. Previously, Sciacovelli et al. demonstrated that FH loss in a Fh1^−/− mouse and in human FH-deficient cells (UOK262) induced the EMT program [26]. However, our study found that, in HEK-293T cells, FH loss induces the opposite MET program. Additionally, we demonstrated that FH loss affects the JAK/STAT3 signaling pathway, a central regulator of the EMT program [27]. This inconsistency in FH’s role in the cell’s fate could be due to the different cell line models used. Specifically, we investigated the effect of FH loss on HEK-293T, which is in an epithelial and mesenchymal hybrid state as it expresses both markers [28]. In contrast, the cells used in the other studies examined FH loss in epithelial cells. Thus, it is possible that the FH role in cell plasticity is not limited to the EMT program but has a more complex function.

5. Conclusions

The FH enzyme is a critical component of the TCA cycle and thus plays a vital role in cell metabolism and proliferation. However, cells can proliferate without this enzyme, and, moreover, mutation or deletion of FH is an apparent driver of various cancers. To illuminate this apparent paradox, we show that the immediate impact of FH loss is the inhibition of cell proliferation. Yet, we generated stable knockout clones that display insignificant differences in their proliferation rate compared to the WT cells. In this study, we set out to resolve this paradox and reveal the molecular mechanism by which cells shift from FH-dependent to FH-independent proliferation.

Addressing this paradox, we describe a comprehensive model that differentiates between the short-term and long-term effects of the FH knockout. We determine that the connection between these two phases of proliferation lies in the DNA damage response, which involves the moonlighting function of FH. In WT cells, FH participates in the TCA cycle, as well as in the DNA damage response. However, upon initiating FH loss, the halted cells also become deficient in their ability to repair their DNA damage and therefore accumulate mutations. We learned that these are not random mutations, as they are enriched in genes encoding signaling molecules. The JAK/STAT3 pathway is an example of a signaling cascade impaired in these long-term FH stable clones.