Mitochondrial Fus1/Tusc2 and cellular Ca2+ homeostasis: tumor suppressor, anti-inflammatory and anti-aging implications

FUS1/TUSC2 (FUSion1/TUmor Suppressor Candidate 2) is a tumor suppressor gene (TSG) originally described as a member of the TSG cluster from human 3p21.3 chromosomal region frequently deleted in lung cancer. Its role as a TSG in lung, breast, bone, and other cancers was demonstrated by several groups, but molecular mechanisms of its activities are starting to unveil lately. They suggest that Fus1-dependent mechanisms are relevant in etiologies of diseases beyond cancer, such as chronic inflammation, bacterial and viral infections, premature aging, and geriatric diseases. Here, we revisit the discovery of FUS1 gene in the context of tumor initiation and progression, and review 20 years of research into FUS1 functions and its molecular, structural, and biological aspects that have led to its use in clinical trials and gene therapy. We present a data-driven view on how interactions of Fus1 with the mitochondrial Ca2+ (mitoCa2+) transport machinery maintain cellular Ca2+ homeostasis and control cell apoptosis and senescence. This Fus1-mediated cellular homeostasis is at the crux of tumor suppressor, anti-inflammatory and anti-aging activities.


INTRODUCTION
Execution of cellular programs including proliferation, differentiation, apoptosis, senescence is based on fine-tuned signal transduction cascades. Growth receptor signaling (such as receptor tyrosine kinases or RTK) should be tightly controlled at the spatial and temporal levels to execute appropriate cell decisions. Uncontrolled RTK signaling (via EGFR, HER2/neu, or VEGFR) could be oncogenic and conducive to a variety of other pathologies [1,2]. Indeed, tumor suppressors integrate various signaling networks pivotal for tissue homeostasis [1,2]. They control DNA repair (e.g., BRCA1, MSH2), cell cycle progression (RB1, CDKN2A, TGFB1), angiogenesis (VEGF, ANGPTL4), apoptosis (TP53, BCL2), and cell adhesion (CADM1, FAT, CDH) [1][2][3][4][5][6][7][8]. A separate class of tumor suppressors is miRNAs, small hairpin RNAs regulating gene expression (miR-7, miR-29, miR-145) [5,6,9] and https://bioinfo.uth. edu/TSGene/. Loss of TSG functions in malignant cells occurs at the DNA, transcriptional and post-transcriptional levels and may be achieved via different mechanisms. According to the Knudson's two-hit paradigm (stemming from clinical observations on Rb deficiency), first hit comes from the loss of heterozygosity (LOH) due to deletion or loss-of-function mutation in one of two TSG alleles that can be inherited as a recessive mutation. Second hit causes a biallelic mutation eliminating the function of the remaining TSG allele. This transforms the cell towards malignancy. The effect of haploinsufficiency, when mutation of a single gene copy is sufficient for malignant transformation, has been described for some TSGs (CDKN1B, TP53, NF1, PTEN) [7,8]. Later, importance of epigenetic silencing via promoter hypermethylation or histone modification was also recognized [7,8]. At the protein level, TSG products can be functionally inactivated via proteasomal degradation (MDM2/p53 axis) or through altered cellular compartmentalization (mislocalization of SMAD4 in cytosol vs. nucleus) [5]. Transcriptional silencing of TSGs in tumors has also been reported (TWIST/CDH1 repression axis) [5].
In this review, tracing the history of the tumor suppressor Fus1/ Tusc2 studies "from bench to bedside", we discuss what is known of its role in tumor development, inflammation, and aging. Initially characterized as one of several genes belonging to potential 3p21.3 TSG cluster [10][11][12], recently FUS1 gene enters clinical trials as a gene-based therapy in the form of plasmid DNA encapsulated in cation lipid nanopartciles (REQORSA) in combination with other drugs for patients with non-small cell lung cancer (NSCLC) [13].
Reported ability of Fus1/Tusc2 to inhibit receptor (EGFR, PDGFR, c-Kit) and non-receptor (c-Abl, Akt) tyrosine kinases as well as promote apoptosis [14] makes Fus1/Tusc2 a promising adjuvant for successful anti-tumor chemotherapy (https://clinicaltrials.gov/ ct2/show/NCT01455389, https://clinicaltrials.gov/ct2/show/NCT0 4486833). By affecting the listed above upstream pathways, Fus1/Tusc2 can potentailly prevent cancer cell evolution towards drug-resistant variants. During the last years, ongoing revolution in immunotherapy including series of Abs targeting immune checkpoint (IC) molecules altered therapeutic landscape in tumor treatment and significantly improved therapeutic outcomes [15][16][17]. However, broad spectrum of immune evasion mechanisms let cancer cells escape from immune surveillance stimulated by using IC inhibitors in clinics (e.g., expression of alternative IC molecules such as TIM-3 after anti-PD-1 therapy) [18,19]. This issue was suggested to be solved by using REQORSA as Fus1/Tusc2-based genetic therapy in mouse model of lung cancer reduced expression of array of IC molecules (TIM-3, CTLA-4, PD-1) concomitant with anti-tumor response [20].

FUS1/TUSC2 GENE DISCOVERY AND FUNCTION
The pioneering concept describing the importance of loss of chromosomal segments responsible for tumor suppression was put forward by Theodor Boveri in 1914 [26]. In the 1980s, chromosomal abnormalities identified in small cell lung carcinoma (SCLC) demonstrated that SCLC cell lines and tumor biopsies shared the deletion of 3p14-23 region [27,28]. Further, this chromosomal segment was narrowed down to 3p [21][22][23]29] and 3p21 [30] as two most common overlapping deletions in SCLC and other lung cancers. In parallel, it was established that the 3p deletions followed the LOH pattern and matched the Knudson's two-hit paradigm [31,32]. The localization of TSGs in the 3p21 region was reinforced by detection of deletions in this area in other solid (e.g., breast, cervix, and renal carcinoma) and hematopoietic (chronic myeloid leukemia) tumors [33][34][35][36][37][38][39][40][41]. Few research groups used a microcell fusion technique to transfer 3d chromosome into tumor cells. Such hybrid tumor cells displayed signs of senescence, growth arrest, and decreased tumorigenicity in athymic nude mice. This demonstrated a tumor suppressor potential for the entire chromosome 3 [42] and regions 3p21 [43] and 3p21.3 [44].
To perform detailed mapping of the 3p21 region, initial analysis of growth suppression of human-mouse tumor hybrid cells was done and resulted in the isolation of a subclone HA [3] BB9F carrying a 2-Mb fragment of human 3p21-22 chromosome with tumor suppressor function [43]. Further, this region was narrowed down to the 3p21.2-21.3 area using specific chromosomal markers [45]. A long-range physical map spanning about 1.8-Mb DNA over the deleted region helped to construct a 700-kb clone. This library (23 cosmids and one PI phage) included a genetically defined 370kb segment containing lung cancer TSGs from the 3p21.3 fragment [40]. This led to the identification of a 220 kb segment deleted in primary breast cancer. By overlapping chromosomal fragments deleted in SCLC cells, the TSGs area was narrowed down to the minimal 120 kb deletion nested within three small homozygous deletions in the 3p21.3 segment (3p21.3 C or LUCA, 3p21.3 or CER1, and 3p21.3 T or AP20). Within this 120 kb deletion were identified eight genes: HYAL2, FUS1(TUSC2), RASSF1, BLU/ ZMYND10, NPR2L, 101F6, PL6, and CACNA2D2 [11]. The original name of TUSC2, FUS1, was introduced based on its position at the junction (FUSion) of cosmids LUCA12 and LUCA13. FUS1 remains to be commonly used even though it is sometimes confused with an unrelated gene, FUS [10].
Shortly after gene identification, it was demonstrated that overexpression of the FUS1/TUSC2 transgene in FUS1/TUSC2deficient lung cancer cells suppressed proliferation, blocked G1/ S or G2/M transition, and increased doubling time implying a tumor suppressor role for FUS1/TUSC2 [46,47]. Likewise, intratumoral adenoviral delivery of the Fus1/Tusc2 transgene suppressed tumor growth and lung metastases in mice [46].
For further insights into the biological role of FUS1, Fus1/Tusc2 −/− mice were developed. These animals showed increased frequencies of lupus-like autoimmune conditions (vasculitis, glomerulonephritis, anemia, circulating autoantibodies) and spontaneous vascular tumors [22], confirming tumor suppressor properties of Fus1. Moreover, these mice exhibited increased susceptibility to irradiation [48,49], enhanced response to A. baumanii infection [21], premature aging [23], hearing loss [50], and olfactory and spatial memory impairments [51]. These pleiotropic effects of Fus1/Tusc2 loss expanded its role beyond tumor suppression activities. Also, these data suggested that Fus1/Tusc2-dependent therapeutic approaches could alleviate/treat different diseases associated with mitochondrial dysfunction, which is linked to inflammation, infection, metabolic imbalance, and aging.
At the beginning of 3p21.3 cluster analysis, it was expected that TSGs from this region would possess cancer-associated spectrum mutations. Surprisingly, extensive studies found little or no mutations in the genes from 3p21.3 cluster, and in the FUS1/TUSC2 gene, except for a few lung cancer cell lines with gene deletions and stop mutations. Overall, the mutational rate in this area did not exceed 5% [10]. The identified stop mutation (28 bp deletion at the 3' terminus of the FUS1/TUSC2 exon 2) resulted in the expression of nonfunctional C-terminal truncation at position 82 [47].
In normal tissues, FUS1/TUSC2 is ubiquitously expressed with the highest levels in all brain regions, followed by blood vessels, stomach, esophagus, colon, adrenal, pituitary, and thyroid glands, skeletal muscle, kidney, spleen, lung, testis, and fallopian tubes (https://gtexportal.org/home/gene/TUSC2). FUS1/TUSC2 mRNA was detected in lung and sarcoma cancer cell lines and sarcoma tissues, but no protein expression was detected in SCLC or non-small cell lung carcinoma (NSCLC) cells [2,47,52,53]. Low or no protein expression was explained by its reduced half-life in tumor cells due to the loss of myristoylation, a post-translational modification that prevents FUS1/TUSC2 from proteasome-mediated degradation [53]. Detailed information on systemic, cellular and molecular manifestations of Fus1 decrease/ loss/increase in normal and tumor tissues is presented in Table 1.
Cancer-specific epigenetic mechanisms involved in FUS1 gene suppression have also been reported. Generally, loss of a gene expression in tumors occurs~10 times more frequently due to CpG islands hypermethylation of the promoters than due to mutations [62]. Histone H3K9 methylation patterns (H3K9me1-3) are also responsible for gene silencing [63], whereas H3 acetylation leads to the opposite effect [64]. As for FUS1, no CpG methylation in the promoter was demonstrated in lung, nasopharyngeal, and breast cancers, even though promoter hypermethylation in the neighboring gene RASSF1 was detected [65,66,47,10]. However, partial promoter methylation of the FUS1/TUSC2 gene was reported in head-and-neck [67] as well as in 20% of NSCLC cancers [68]. As for histone modifications, predominance of H3 acetylation (H3K9ac) over methylation (H3K9me3) was reported [66].
Physiological stimuli affecting FUS1/TUSC2 mRNA levels vary in their origin. The expression is reported to be downregulated by reactive oxygen species (ROS) [69,70], and upregulated by hypoxia [58] and the factors stimulating differentiation [71]. Summary of mechanisms regulating Fus1 levels is presented in Fig. 2. However, a fuller understanding of how FUS1/TUSC2 expression is regulated by physiological and pathological stimuli is still lacking.

PROTEIN MOTIFS AND POSTTRANSLATIONAL MODIFICATIONS OF FUS1/TUSC2
Human FUS1/TUSC2 is a small (110 aa) protein with an estimated MW of 12 kD. The protein is basic and predicted to have a pI (isoelectric point) of 9.69 [10]. According to a computer-based modeling, FUS1/TUSC2 lacks transmembrane domains, is highly hydrophobic and contains helix-coil domain secondary structures [53]. A 3D model for FUS1/TUSC2 protein is shown in Fig. 1B. Tertiary FUS1/TUSC2 structure is still debatable. Below are the functional motifs identified in the protein sequence 1. At the N-terminus, FUS1/TUSC2 contains a myristoylation signal (Met-Gly-X-X-X-Ser/Thr), and experiments confirmed that FUS1 is myristoylated [53] (Fig. 1C). 2. Motif-based profile scanning found a protein kinase A interaction site, an A-kinase anchoring protein interaction site, and a PDZ class II domain [53]. 3. In silico analysis revealed that protein fragment 54-65 aa is highly homologous to the EF-hand Ca 2+ -binding domains found in calmodulin as well as in mitochondrial proteins MICU1 and LETM1 [24,25] (Fig. 1C).  Myristoylation-deficient FUS1/TUSC2 loses its abilities to induce apoptosis and suppress tumor cell proliferation in vitro and promotes tumor growth and metastases in vivo [53] Murine osteoclasts from bone marrow. Normal Tusc2 levels.
Overexpression of Tusc2 positively regulates osteoclast differentiation induced by RANKL. Tusc2 induces activation of Ca 2+ -dependent RANKL-mediated NF-κB and CaMKIV/CREB signaling cascades. [71] Human NSCLC cell lines A549 and H1299, normal human lung fibroblast 32D P210 cells. (TUSC2 expression is lower in all cancer cells than in a normal cell line) Deletion of 83-110 aa at the FUS1/TUSC2 C-terminus leads to the loss of its ability to inhibit tyrosine kinase c-Abl [98] Human NSCLC cell lines H1299, H460, A549, H322 and normal human bronchial epithelial cells HBEC. (TUSC2 expression is lower in all cancer cells than in a normal cell line) Co-expression of FUS1/TUSC2 and p53 synergistically increased apoptosis associated with down-regulation of MDM2 and activation of Apaf/caspase-3 [113] Anaplastic thyroid cancer cell line 8505 C and papillary thyroid cancer cell line TPC-1 (Fus1 level is decreased) Overexpression of FUS1/TUSC2 increases levels of Smac/ Diablo, suppressor of IAPs blocking caspase-and cytochrome c-mediated apoptosis [115] Cell-free (Protein Chip array and SELDI-TOF mass spectrometry) Myristoylation is extremely important for cell localization and functioning of the FUS1/TUSC2 protein [53]. N-myristoylation is a post-translational modification consisting of the removal of the N-terminal methionine from a protein and ligation of the released NH2 group of glycine to the residue of myristic acid, a 14-carbon saturated fatty acid (Fig. 1D) [73]. The resulting lipid tail participates in protein folding and anchoring myristoylated proteins to different membrane compartments, where they execute their functions (Fig. 1E) [74,73,75]. A mutant form of FUS1/TUSC2 missing the myristoyl tail has a shorter half-life (6 h vs 12 h) due to improper protein folding and increased proteasome degradation [53]. Also, myristoylation-deficient FUS1/TUSC2 loses its characteristic mitochondria/ER localization and its abilities to induce apoptosis and suppress tumor cell proliferation in vitro. Importantly, it also acquires the abilities to Fig. 2 Transcriptional, post-transcriptional, and post-tranlsational regulation of Fus1/Tusc2. Fus1/Tusc2 gene is regulated on transcriptional level by different physiological and pathological factors such as ROS, hypoxia, and differentiation molecules (e.g., RANKL). In some tumors (e.g., NSCLC), DNA methylation may lead to down-regulation of TUSC2 gene transcription. Binding of microRNAs (miRs) to 3' and 5' untranslated regions (UTR) of TUSC2 mRNA suppresses its translation. mRNA for TUSC2 pseudogene (TUSC2P) sequesters miRNAs and prevents their binding to TUSC2 mRNA. During translation, 5' UTR can adopt conformation preferring stoppage of ribosomes along TUSC2 mRNA observed in NSCLC tumor cells. After effective translation, Fus1/Tusc2 undergoes myristoylation (Myr) which is necessary for a protein stability in normal cells; in the absence of Myr tail (in tumor cells), protein is degraded at faster rate via proteasome machinery.  Increased ROS production, up-regulation of antioxidant defense proteins (Prdx1) at steady-state, age-dependent decrease in the expression of Sod2 and Prdx1 [24,50,70] Primary mouse embryonic fibroblasts and immortalized mouse kidney epithelial cells from Fus1 KO mice (Fus1 is deleted) Defects in respiration (significantly decreased maximal mitochondrial respiration and respiratory reserve capacity) [23] Cochlear cells from Fus1 KO mice (Fus1 is deleted) Pathological alterations in antioxidant (AO) and nutrient and energy sensing pathways (mTOR and PTEN/AKT) and downregulation of PINK1, a sensor of mitochondrial quality control occur in cochleae of young Fus1 KO mice before major hearing loss. Short-term anti-oxidant treatment corrected these pathological molecular changes and delayed hearing loss. [50] R. Uzhachenko et al. promote tumor growth and metastases in vivo [53]. Therefore, loss of myristoylation may be considered a key event that leads to insufficiency of Fus1 function, when compared to mutations and hypermethylation [14,53]. The predicted Ca 2+ EF hand binding motif and hydrophobic binding pocket of FUS1/TUSC2 prompted it to be classified as a novel calcium/myristoyl switch protein (Fig. 1D) [24]. Its involvement in Ca 2+ -dependent signaling have been reported for different cell types including CD4 + T lymphocytes, kidney epithelial cells, osteoclasts, and mouse embryonic fibroblasts [71,23,24,23,25]. The presence of myristoylation and Ca 2+ binding motifs in one protein is pivotal for regulation of cellular processes. In has been established that Ca 2+ /myristoyl switch proteins, in response to Ca 2+ binding to EF-hand motifs, release their lipid myristoyl tail from a protein hydrophobic pocket and anchor it to membranes during Ca 2+ elevations (Fig. 1E) [74,75].

FUS1/TUSC2 IN MITOCHONDRIAL CA 2+ REGULATION
Mitochondrial Ca 2+ transport: main players MitoCa 2+ transport is balanced by Ca 2+ import and export. The main route for mitoCa 2+ entry is mitochondrial Ca 2+ uniporter holocomplex (MCU cx ), consisting of membrane channel protein MCU (mitochondrial Ca 2+ uniporter) and MCU regulatory proteins MICU1-3, MCUb, MCUR1, EMRE [79][80][81]. MCU is an inner mitochondrial membrane protein that forms a pentameric channel with a selectivity filter for Ca 2+ ions (a DIME motif) (Fig. 3A) [82,83]. MCU becomes permeable for Ca 2+ when ion concentration reaches 1-10 μM, which is only possible at the contact with the endoplasmic reticulum (ER) where Ca 2+ is released via inositol-1,4,5-triphosphate receptors (IP3Rs). The 1-10 μM Ca 2+ threshold is set by the Ca 2+ -binding EF-hand motif of the mitochondrial protein MICU1 (mitochondrial Ca 2+ uptake1) (Fig. 3A) [84,85]. In the absence of MICU1, MCU transports Ca 2+ at much lower cytoplasmic concentrations (hundreds of nM range) Fig. 3 Role of intracellular Ca 2+ signaling in cell decisions and potential place of Fus1/Tusc2 in Ca 2+ signaling network. A Mitochondrial calcium transport at the endoplasmic reticulum (ER)/mitochondria interface and its regulation by tumor suppressors and protooncogenes. Ca 2+ released from ER via inositol triphosphate receptors (IP3Rs) is taken up by mitochondria via Ca 2+ uniporter (MCU). MICU1 stimulates opening of the MCU channel after Ca 2+ binding to its EF-hand motif and fine-tunes Ca 2+ currents. Ca 2+ accumulated in mitochondria can be transported back into inter-organellar space by the mitochondrial Na + /Ca 2+ exchanger (mNCX) and captured by Ca 2+ ATPases in ER (SERCA). IP3Rs activity is negatively regulated by pro-survival/pro-tumorigenic Akt-mediated phosphorylation. This effect is counteracted by tumor suppressors: [1] PML which binds to IP3R and [2] PTEN which hydrolyzes phosphoinositides necessary for Akt/PKB activation. p53 inhibits SERCA activity, which leads to cytosolic Ca 2+ retention potentially promoting senescence and cell death. Another pro-survival mediator, miR-25, down-regulates MCU and reduces cell susceptibility to apoptosis. B Cell death induced by Ca 2+ overload. Massive Ca 2+ uptake by mitochondria results in overstimulation of the tricarboxylic acid (Krebs) cycle (TCA) supplying the respiratory chain (RC) with redox components. This process induces reactive oxygen species (ROS) overproduction leading to a redox modification of MCU that intensifies Ca 2+ uptake, therefore forming a positive feedback loop. Accumulation of Ca 2+ along with enhanced ROS production triggers opening of the mitochondrial permeability transition pore and release of cytochrome C via voltage-dependent anion channel (VDAC). Mitochondrial disruption and interaction of cytochrome C with Apaf-1 leads to apoptosome formation triggering cell death. C Proposed mechanism of action of Fus1/Tusc2 protein. At low Ca 2+ , Fus1/Tusc2 maintains its inactive state by adopting a conformation with the myristoyl tail hidden inside of the hydrophobic pocket. This conformation keeps mtCU inactive and mNCX active, thus preventing Ca 2+ rise in mitochondrial matrix. When Ca 2+ is elevated in the intermembrane space (IMS) or matrix, Fus1/Tusc2 releases its lipid tail that anchors to the inner mitochondrial membrane (IMM), thus maintaining the MCU at the open state and preventing mNCX activation. Overall, these actions help to increase Ca 2+ in mitochondrial matrix to the levels necessary for TCA stimulation. D A schematic diagram demonstrating how deficiency in mitochondrial Ca 2+ accumulation affects cytosolic Ca 2+ . At the steady state, adequate and coordinated mtCU and mNCX activities allow Ca 2+ to accumulate in mitochondria and maintain cytosolic Ca 2+ at the moderate levels. However, inhibition/loss of mtCU and activation of mNCX would result in cytosolic Ca 2+ retention and alterations in the pattern of activation of Ca 2+ -dependent proteins.
The importance of MCU in tumor growth is based on its involvement in cell death. Malignant cells benefit from reduction of mitoCa 2+ uptake. This is especially important as tumor cells experience increased ROS production. Interestingly, MCU contains ROS sensing cysteines in its N-terminus. Their oxidation promotes MCU cx assembly, persistent channel activity, and mitoCa 2+ overload following by opening the permeability transition pore, which triggers cell death (Fig. 3B) [80]. However, for some breast cancers it was reported that activation of MCU promotes cancer cells motility, invasion, and growth [79]. So, it is not surprising that mitoCa 2+ accumulation undergoes regulation by tumor suppressors. Thus, p53 directly interacts with the Ca 2+ -ATPase in ER, promotes its negative oxidative modification, and stimulates enhanced Ca 2+ release from ER with further Ca 2+ transfer to mitochondria via the MCU protein (Fig. 3A) [88]. The lipid/protein phosphatase PTEN binds IP3R and, thus, counteracts PKB/Akt phosphorylating and inhibiting IP3R; this results in a higher mitochondrial Ca 2+ accumulation and apoptosis (Fig. 3A) [89]. On the other hand, down-regulation of the MCU mRNA by miR-25 is accompanied by increased cancer cell survival and apoptosis resistance (Fig. 3A) [90].
Extrusion of Ca 2+ from mitochondria into cytosol is mediated by NCLX (Na + /Ca 2+ /Li + exchanger), a mitochondrial form of Na + / Ca 2+ exchanger (mNCX), and Ca 2+ -binding EF-hand motif containing LETM1, a 2H + /Ca 2+ exchanger (Fig. 3A) [79,91]. Diminished NCLX activity leads to mitoCa 2+ overload and elevated ROS production due to the activation of Krebs cycle [92]. In a few tested non-tumor cell models, these events led to cell death. In cancer cells, however, although increased ROS blocked cell proliferation, they also induced metastasis via the ROS/HIF1α signaling axis [93]. Therefore, Ca 2+ transients in mitochondria play a decisive role in the normal and tumor cell biology.
Initially, it was proposed that MICU1 regulates MCU-mediated Ca 2+ currents via Ca 2+ /myristoyl switch mechanism [96]. However, the MICU1 and MCU interaction is mediated via direct protein binding. At the MICU1 C-terminus, KQRLMRGL peptide represents an MCU-binding domain interacting with with DIME motif via salt bridges [97]. Structural analysis of the MCU N-terminal domain (NTD) surprisingly revealed unidentified lipid molecule bound to a hydrophobic protein surface formed by residues in the L1 loop, two helices (α2 and α3) and C-terminal tail. This lipid was described as a linear lipid-like structure consisting of 13-16 carbon atoms that is similar to a tetraethylene glycol molecule [67]. It is noteworthy that myristic acid (a substrate for N-terminal myristoylation) is a saturated linear long-chain lipid with a 14-carbon backbone complying with the features of an unidentified lipid. Therefore, Fus1/Tusc2 could regulate the activity of MCU via a novel mechanism of interprotein lipid tail exchange. We suggest that the myristic acid residue of FUS1/TUSC2 protein released from the hydrophobic pocket after Ca 2+ binding to EF-hands could interact with the hydrophobic surface of MCU and maintain its Ca 2+ transporting function (Fig. 4A). This is likely in view of the fact that Fus1/Tusc2 has the amino acid sequence KARGLWPF resembling MCUbinding domain of MICU1-3 proteins (Figs.1, 4B and C). Moreover, FUS1/TUSC2 can potentially interact with MCU via its interprotein interaction site at the C-terminus (positions 81-96) (Figs. 3  and 4C). This mechanism needs further elucidation by functional and structural studies.
Putative Fus1 protein-protein interaction via myristoyl tail exchange An intriguing opportunity for protein-protein interactions was demonstrated by mutational analysis of FUS1/TUSC2 binding to a pro-tumorigenic tyrosine kinase c-Abl [98,99]. Initiated by discovery of a truncated form of FUS1/TUSC2 (deletion of C-terminal 83-110 aa) in tumor cells, Lin et al. showed that the FUS1/TUSC2 81-96 aa C-terminal peptide linked with stearate inhibits c-Abl activity and facilitates its degradation. On the contrary, the FUS1/TUSC2 N-terminal peptide 1-80 aa, although capable of binding to c-Abl, was unable to inhibit its kinase activity [98]. Like FUS1/TUSC2, c-Abl is a myristoylated protein. It is activated by releasing the myristoyl tail from the hydrophobic pocket and conforming to an active open state, which allows c-Abl to bind its substrates via the SH2-domain [100]. FUS1/TUSC2 mutational analysis showed that while the N-portion of FUS1/ TUSC2 binds to myristoyl-binding pocket of c-Abl, the C-terminal portion of FUS1/TUSC2 (containing c-Abl inhibitory peptide 81-96 aa) interacts with the ATP-binding site in the N-lobe of c-Abl. Alternatively, FUS1/TUSC2 could interact with an activation loop in the C-lobe of c-Abl [99].
Initially, it was thought that myristoylation is exclusive for targeting proteins to certain membrane compartments but, nowadays, this posttranslational modification is appreciated to play roles in protein folding and prevention from premature degradation [100,101,53]. A synthetic ligand for myristoyl binding pocket named GNF-5 maintains closed inactive conformation of c-Abl apo form or convert open conformation to a closed one [102]. Therefore, one could speculate that the myristoyl tail of FUS1/TUSC2 inserted into the hydrophobic pocket of c-Abl inhibits its tyrosine kinase activity. This tail-pocket interaction can be an important feedback loop in the network of interactions initiated by RTKs. Thus, plateletderived growth factor receptor (PDGFR) stimulates signaling molecules such as PLCγ1, c-Src, and c-Abl. In turn, PLCγ1 generates IP3 and promotes Ca 2+ release from the ER [103].
We suggest that activation of FUS1/TUSC2 by Ca 2+ leading to the release of its myristoyl tail could suppress c-Abl activity. Noteworthy, c-Abl-deficient B cells exhibited reduced Ca 2+ flux in response to antigen receptor or CD19 stimulation [104] reinforcing the concept of mutual regulatory loops between FUS1/ TUSC2 and c-Abl.
Compromised capability of Fus1 −/− adult stem cells to repopulate tissues was reported in aged Fus1 −/− animals (e.g., hair follicles, thymus) [23] or younger mice after exposure to radiation (e.g., GI crypt epithelial cells, melanocyte stem cells) [49]. In vitro data suggest that Fus1 is involved in bone remodeling shaped by bone deposition (osteoblasts) and resorption (osteoclasts) [106]. Silencing of Fus1/Tusc2 gene in bone marrow precursor cells in vitro diminished RANKL-induced osteoclast differentiation without affecting osteoblast formation [71]. Thus, it is possible that Fus1/Tusc2 loss in vivo could also affect bone remodeling, shifting it towards bone deposition that would reduce tissue repair potential [106] corroborating with accelerated aging and development of age-related disorders in Fus1 −/− mice [23].
Thresholds in the mitogenic signal transduction demand tight control of activation of proteins involved in cell proliferation. Malfunction of key tumor suppressors (e.g., E3 ligase PML) may switch cell fate from senescence to malignant transformation [50,107]. The ability of Fus1/Tusc2 to calibrate Ca 2+ responses would translate into adequate cell responses based on the characteristics of input signals such as signal strength. In this regard, Fus1/Tusc2 reminds other tumor suppressors (e.g. PTEN, Spry), which prevent cells from over-stimulation by mitogenic signals maintaining their survival and responsiveness to proliferation signals [108][109][110][111][112].

REGULATION OF APOPTOSIS BY FUS1/TUSC2
Numerous studies demonstrated that FUS1/TUSC2 overexpression in cancer cells that lack 3p21.3 or FUS1/TUSC2 gene/expression induces cell death [2,46,47,53,113,114]. Co-expression of FUS1/ TUSC2 and p53 synergistically increased apoptosis in NSCLC cells. Fig. 4 Hypothetical mechanisms of Fus1/Tusc2 and MCU (mitochondrial Ca 2+ uniporter) interaction. A A proposed mechanistic model of inter-protein interaction between Fus1/Tusc2 and MCU. At low Ca 2+ levels, Fus1/Tusc2 maintains a closed conformation with its myristoyl tail hidden in a hydrophobic pocket that promotes protein folding, stabilizes inactive state, and prevents premature protein degradation. At high Ca 2+ levels, myristoyl tail is released triggered by Ca 2+ binding to EF-hand of Fus1/Tusc2. Two possible scenarios could be proposed after the tail release: [1] the lipid tail anchors Fus1/Tusc2 to the membrane followed by the Fus1/MCU interaction in the mitochondrial intermembrane space (IMS) or matrix side of the inner membrane (IMM), or [2] the lipid tail of Fus1/Tusc2 binds the hydrophobic surface of the MCU N-terminal domain (NTD) and affects activity/kinetics of the mitochondrial channel. B The Fus1/Tusc2 protein motif, which may interact with the MCU DIME motif (Ca 2+ selectivity filter). The 7-15 aa fragment from Fus1/Tusc2 was compared with motifs of the MICU1-3 proteins responsible for binding to MCU in the IMS. Critical amino acids responsible for formation of salt bridges with the DIME motif in MCU are shown in red. Blue/green color highlights identical residues in similar positions in Fus1/Tusc2 as compared to MCU1-3 proteins. C Docking simulation of predicted interaction between the MCU/EMRE complex and Fus1/Tusc2 protein. In the configuration 1, Fus1/Tusc2 interacts with MCU close to its DIME sequence (Ca 2+ selectivity filter) in the mouth of the channel. Lys (K) and Arg (R) from the Fus1/Tusc2 motif homologous to MCU-binding motif of MICU1-3 (KxRxxRGx) (see Fig. 4B) are positioned against the DIME motif. Critical Ser and Asp residues in the DIME motif required for formation of salt bridges as well as potential Fus1/Tusc2-binding sequence (including K and R) are marked with licorice sticks. (see 3B and text). In the configuration 2, Fus1/Tusc2 interacts with the MCU/EMRE complex via the 81-96 aa fragment (highlighted in blue) that was experimentally shown to be involved in the inter-protein interactions. Docking simulation has been performed using the ClusPro server [143] and visualized using PyMOL software (The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC; https://pymol.org/2/#page-top). The MCU sequence was retrieved from the Protein Data Bank (acc # 6O5B).
This synergistic effect was associated with the ability of Fus1/ Tusc2 to down-regulate expression of MDM2, an E3 ubiquitin ligase that suppresses p53, and thereby stabilizes p53 levels. Importantly, combined effect of Fus1 and p53 co-expression required Apaf-1 that mediates mitochondria-and caspase3dependent apoptosis (113) . Additionally, in thyroid cancer cell lines, overexpression of FUS1/TUSC2 increased levels of Smac/Diablo that blocks caspase regulatory inhibitors of apoptosis proteins and cytochrome c [115]. Protein Chip array and SELDI-TOF mass spectrometry revealed direct interaction between PDZ domains of FUS1/TUSC2 and Apaf [116] corroborating the link of FUS1 with the Apaf-1-mediated apoptosis.

FUS1/TUSC2 IN CELLULAR SENESCENCE Cellular senescence: main players
Cellular senescence refers to cellular aging characterized by a highly stable cell cycle arrest accompanied by biochemical and morphological alterations [117,118]. The phenomenon of a limited number of cell divisions discovered by Leonard Hayflick in 1961 became known as the Hayflick limit [119]. A role of p53 in senescence is evident from its effect on cell cycle arrest usually mediated by up-regulation of p21, an inhibitor of cyclindependent kinases CDK1, CDK2, and CDK4/6 required for the G1/S transition [120,121]. Senescence can be induced by oncogene overactivation, oxidative stress, genotoxic drugs, radiation, CDK suppression, demethylating and acetylating agents, etc. [121]. Cell cycle arrest accumulates unphosphorylated form of Rb protein, upregulating the lysosomal compartment and subsequently stimulating mTOR signaling (Fig. 5) [122]. mTOR signaling, in turn, up-regulates mitochondrial biogenesis via TF PGC1α. Increased mass and dysfunctional state of mitochondria enhance intracellular ROS and DNA damage. Downstream of this cascade, DNA damage kinases (i.e., ATM) further activate the AKT/mTOR signaling axis stimulating PGC1α and creating a positive feedback loop stabilizing senescent state (Fig. 5) [117,118].

Ca 2+ signaling and senescence
Fine-tuned control of cellular Ca 2+ signaling by tumor suppressors and oncogenes [127] established that increased cellular Ca 2+ (cytosolic and/or mitochondrial) results in senescence or apoptosis, whereas moderate levels of Ca 2+ favor cell proliferation. Ca 2+ regulates SASP via activation of calpain that converts pro-IL-1α into functional IL-1α (128) . Another Ca 2+ effector is NFAT, a TF regulated by calmodulin/calcineurin complex that has both proproliferative and pro-senescent effects. In particular, NFAT induces p53 by suppressing expression of ATF3, a p53 negative regulator. NFAT acts as a pro-senescent factor by activating expression of IP3R2 that mobilizes Ca 2+ from ER elevating cytosolic and mitoCa 2+ levels [128].
IP3Rs are instrumental in cancer cells, where the number of oncogenes and tumor suppressors cross their paths [127]. Prosurvival/pro-tumorigenic PKB/Akt phosphorylates and inactivates IP3R3. Tumor suppressor protein PML binds to IP3R3 and recruits protein phosphatase PP2A to dephosphorylate IP3R3 leading to IP3R3-marked opening (Fig. 3A). Consequently, Ca 2+ accumulation in mitochondria would increase ROS production leading to a senescent state (Fig. 3B). It is noteworthy that abrogation of MCU helps cells avoid senescence [128] and apoptosis, and promotes uncontrolled proliferation, migration, and metastases [80].
Fus1/Tusc2-dependent processes mediating cellular senescence Proliferation. The role of Fus1/Tusc2 in the control of senescence appears to be a part of its role as a TSG. FUS1/TUSC2 overexpression in tumor cells is associated with cell cycle arrest at G1/S [47] and G2/M [115] transition checkpoints. Deletion or silencing of Fus1/Tusc2 accelerates proliferation of activated mouse CD4 + T cells [24] and human tumor cells [129]. Upregulation of miR-197 miRNA that increases FUS1/TUSC2 levels and, thus, suppresses tumor metastasis was attributed to the ability of FUS1 to regulate proliferation of human glioblastoma cells [129].
Genotoxic stress. In several models of cell injury, Fus1 −/− cells showed increased levels of genotoxic stress and senescence markers compared to wild-type counterparts. For example, after intraperitoneal injection of asbestos, infiltrating cells from Fus1 −/− mice had higher levels of γH2AX, DNA damage response molecule, and phosphorylated pro-inflammatory NFκB and ERK1/2. This was associated with increased number of macrophages and accelerated accumulation of granulocytes in the peritoneal cavity. Among others, IL-1β, a signature cytokine of SASP, was up-regulated in peritoneal Fus1 −/− cells.  5 Senescence-associated secretory phenotype (SASP). G1/S cell cycle block (marked by a cross) driven by p53 is accompanied by accumulation of Rb protein, a critical suppressor of G1/S transition. Rb-triggered up-regulation of lysosomal compartment stimulates mTOR. In turn, mTOR activates PGC1α, a transcription factor involved in mitochondrial biogenesis. Up-regulated mitochondrial compartment accompanied by increase in respiration leads to increased ROS formation. It is followed by DNA damage, a main trigger of the ATM-mediated activation of NFκB pathway, p53, and PGC1α. As a result, positive feedback loops are forming that lead to senescence. NFκB-regulated transcription program triggers elevation of cytokines, chemokines, metalloproteinases (MMPs), and other molecules secreted by senescent cells (senescence-associated secretory phenotype, SASP). The secreted molecules attract immune cells, which induce inflammation in tissues populated with senescent cells.
The increased sensitivity to irradiation in Fus1 −/− mice was accompanied by accelerated cell cycle arrest, aberrant mitosis, lack of proper DNA repair (mitotic catastrophe), early activation of p53, and death of gastrointestinal crypt cells, which are especially susceptible to ionizing radiation [49]. Thereby, Fus1-deficient mice demonstrated enhanced cell damage upon challenging stimuli like other TSG deficiency models (i.e., p27) [8].
Clearance of senescent cells. The most striking alteration in the chemokines/cytokines profile of Fus1-deficient mice was downregulation of RANTES/CCL5, a chemokine for T and NK cells [70]. Since T and NK cells are necessary for clearance of senescent cells [130], lack of CCL5 production in Fus1 −/− mice could lead to accumulation of senescent cells during the course of lifetime and result in chronic inflammation, aging-related diseases, and malignant transformation.
CONSEQUENCES OF FUS1/TUSC2 DEFICIENCY AT ORGANISMAL LEVEL Fus1/Tusc2-associated systemic pathologies Several Fus1-dependent pathologies developing in mice have been reported following the targeted Fus1 inactivation. These could be divided into two groups: Spontaneous pathologies developed in young or middle-aged mice and progressing with time.
a. Chronic systemic inflammation [23]. b. Progressive development of SLE-like autoimmune disease in some mice (incomplete penetrance) (vasculitis, glomerulonephritis, anemia, circulating autoantibodies) [22]. c. Increased frequency of spontaneous vascular tumors [22]. d. Preponderance of aging signs that include lordokyphosis, absence of vigor, diminished hair regrowth, reduced sperm count and motility, enlarged seminal vesicles, and compromised stem cells self-renewal [22,23]. e. Early development of aging-associated diseases: 1. Premature progressive hearing loss (higher threshold for sound intensity, longer latency to respond to sound, and smaller amplitude in auditory brainstem responses (ABR) waves, compared to wild-type mice [50]. The hearing loss corresponded with impaired PTEN/Akt/mTOR pathways in cochlear cells and was ameliorated by administration of N-acetyl cysteine, an antioxidant agent [50]. 2. Impairments in olfactory and spatial memory at a relatively young age (4-5 months old), as indicated by habituation test, hidden cookie test, and Morris watermaze test [51].
Injury-induced pathologies in young mice.
Most likely, defects in common Fus1-dependent mechanism(s)/ pathways that are discussed below underlie these pathologies.
The mTOR pathway activation Fus1 −/− mice showed prominent mTOR signaling activation and oxidative stress, signature hallmarks of senescence and early aging. In young Fus1 −/− mice, cochlear cells are distinguished with reduction of antioxidant enzymes (mitochondrial SOD2, PRDX1) and concomitant activation of the Akt/mTOR pathway (decrease in PTEN levels, up-regulation of phospho-Akt and S6). This molecular pattern was associated with low-grade chronic inflammation observed in bone marrow cells in the temporal bone surrounding the cochlea [50]. During senescence, activation of Akt/mTOR (for example, via DNA damage/ATM pathway) stimulates PGC1α followed by up-regulation in mitochondrial biogenesis. Mitochondria-derived ROS damage DNA and, thereby, maintain mTOR activation. Additionally, mTOR regulates translation of mRNA related to SASP via MAP kinase-activated protein kinase 2 phosphorylating protein ZFP36L1 responsible for mRNA degradation (117) . Therefore, it is not surprising that chronic inflammation accompanies senescent phenotype of cochlear Fus1 −/− cells [50].

Oxidative stress
Increased ROS stabilize p21, a p53 activator and key suppressor of cell proliferation in senescence, and inhibit autophagy 23. Increased ROS production was evident in FUS1/TUSC2-deficient head-and-neck cancer cells JHU012 [70] and in mouse splenocytes [24]. Additionally, cochlear cells from Fus1 −/− mice demonstrated up-regulation of antioxidant defense proteins (Prdx1) in steadystate epithelial cells as well as gradual decrease in the expression of Sod2 and Prdx1 [50]. Further, Fus1 −/− primary mouse embryonic fibroblasts and immortalized kidney epithelial cells showed defects in respiration such as significantly decreased maximal mitochondrial respiration and respiratory reserve capacity, likely due to down-regulation of mitochondrial respiratory proteins [127]. However, upon challenging conditions (i.e., irradiation) Fus1deficient cells displayed delayed up-regulation of Sod2, a mitochondrial form of antioxidant superoxide dismutase [49]. Treatment with antioxidants (Tempol, pyroxidamine) restored Sod2 expression and significantly improved survival of whole-body irradiated Fus1-deficient mice [48]. Another antioxidant, Nacetylcysteine, rescued expression of Prdx1 and respiratory chain compounds, restored normal mitochondrial morphology, and prevented progression of hearing loss in Fus1 −/− animals [50].

Disruption of mitochondrial Ca 2+ homeostasis
Increased ROS production in Fus1/Tusc2-deficient tissues may derive from Fus1 ability to regulate mitoCa 2+ transport [23][24][25]. Indeed, inhibition of MICU1 via Akt/PKB phosphorylation increased ROS production and downstream Akt/PKB activation [85,90]. The deficiency in mitoCa 2+ leads to elevated cytosolic Ca 2+ and activation of Ca 2+ -dependent proteins in the cytosol. For example, Ca 2+ -dependent stimulation of Miro1 results in the remodeling of long filamentous mitochondria into globe-shaped mitochondria, step prerequisite for autophagosomal degradation [80]. Indeed, presence of globule-shaped giant mitochondria in Fus1-deficient epithelial and cochlear cells point to preferred mitochondrial fission in these cells [24,51]. This process should activate mitophagy, but in Fus1 −/− mice it rather leads to accumulation of nonfunctional mitochondria. Fus1 −/− cochlear cells displayed a major down-regulation in the expression of PTENinduced kinase-1 (PINK1) [51], a sensor of mitochondrial quality control, which directs dysfunctional mitochondria towards autophagy/mitophagy [131]. Importantly, PINK1 protects cells from oxidative stress and premature senescence [132]. Therefore, Fus1deficient cells accumulate dysfunctional mitochondria leading to further increase in ROS production and senescence [50].

CONCLUSIONS AND CLINICAL RELEVANCE
The Fus1-mediated cellular homeostasis is at the crux of its tumor suppressor, anti-inflammatory, and anti-aging activities. Developing FUS1-based genetic therapies for cancer patients became an apparent step after revealing its TSG properties.
In preclinical studies, intra-tumoral injection of cationic liposome nanoparticles complexed with plasmid DNA encoding FUS1/ TUSC2 gene (REQORSA or quaratasugene ozeplasmid formely known as Oncoprex) significantly inhibited growth of human NSCLC cells H1299 and A549 in subcutaneously inoculated mice. Administered intravenously, REQORSA suppressed metastases and extended survival of tumor-bearing mice [114]. Phase-I clinical trial reported the effective dosage and safety of REQORSA [13].
Conventional therapy for metastatic NSCLC currently uses epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors such as erlotinib, gefitinib, and osimertinib [137]. Beside toxicity, this type of therapy often leads to treatment-resistant cancer as a result of selecting tumor cell clones carrying mutations. Thus, continuous treatment with erlotinib prescribed to NSCLC patients with overexpression or mutations in EGFR (deletion in 19th exon or L858R point mutation), leads to accumulation of cell clones with T790M mutation in EGFR or up-regulation in other growth receptor signaling pathways (e.g., HGFR, RBB3/PI3K) leading to drug resistance. Osimertinib allows to overcome this issue as it can block EGFR mutated at T790M site; however, resistance to this treatment can also develop after losing tumor cells with T790M [137]. Therefore, combination of EGFR inhibitors with other treatments affecting upstream signaling pathways involved into drug resistance is highly desirable. Ability of Fus1/Tusc2 to suppress receptor and non-receptor kinases as well as sensitize NSCLS cells to chemotherapeutics [14,98,138,139] makes this tumor suppressor a promising therapeutic candidate for combinatorial therapy. Currently, two clinical trials based on combination of Fus1-gene based drug REQORSA and EGFR inhibitors erlotinib (https://clinicaltrials.gov/ct2/show/NCT01455389) and osimertinib (Acclaim-1, https://clinicaltrials.gov/ct2/show/ NCT04486833) are underway for treatment of NSCLC patients.
Ongoing revolution in immunotherapy started with a clinical introduction of checkpoint inhibitors significantly improved therapeutic outcomes in patients with different types of cancer (melanoma, breast cancer, colon cancer, Hodgkin lymphoma, etc.) [15][16][17]. However, effectiveness of immunotherapeutics is limited due to evasion of tumor cells from immunosurveillance mechanisms (e.g., up-regulation of TIM-3 immune suppressive molecules on cancer cells after PD-1 blocking therapy) eventually leading to tumor resistance similar to chemotherapy [18,19]. Thus, treatments with drugs targeting pathways upstream of drug resistance would be beneficial.
Thus, studies initiated about 40 years ago to dissect chromosomal aberrations in lung cancer cells culminated in a new promising gene therapy aimed to battle most aggressive lung cancer stages. In addition, Fus1/Tusc2-mediated anti-inflammatory and anti-aging activities provide avenues for development of new approaches to fight conditions of chronic inflammation, infections, premature aging and geriartic diseases.