2.1. UPF1 in Cancer
UPF1 is downregulated in numerous cancers, which correlates with poor prognosis and low overall survival (OS) rates. Its low expression causes the dysfunction of NMD, increased levels of toxic transcripts and, consecutively, tumor initiation and progression
[25][31][32][34][37][39][40][41][42][43][44][47][48]. Additionally, UPF1 acts in various signaling pathways and promotes undifferentiated stem cell phenotypes, proliferation and metastases. Along with different levels of expression, epigenetic and genetic alterations, as well as the aberrant splicing of UPF1
[31][48], have been demonstrated. Epigenetic alterations, such as the hypermethylation of promoter regions of tumor suppressor genes, lead to the downregulation of their expression and influence tumorigenesis
[59][60]. The UPF1 putative promoter region possesses an enriched CpG island in gastric cancer (GC), which was proved to be hypermethylated
[37]. Hypermethylation was also detected in hepatocellular cancer (HCC). Treatment with the DNA-demethylating drug 5-Aza-2′-deoxycytidine increases UPF1 mRNA and protein levels in both GC and HCC
[37].
As transcription regulation is disrupted in cancer cells, the accumulation of aberrant transcripts has been observed in tumors
[61], detected additionally in such high levels due to the downregulation of the NMD factor, most importantly UPF1, in various cancer types
[25]. In cell lines derived from non-small cell lung cancer (NSCLC), such aberrant splicing isoforms have been identified as potential templates for producing neoantigens, as some of those forms were proved to be translated into several peptides
[44]. Clinical lung cancer specimens were also analyzed for the search of aberrant forms and neoantigens in cancer cells in vivo, with each examined patient possessing such isoforms. A total of 2021 novel isoforms were identified. Such a large number has been proposed to be the result of impaired NMD, as ~30% of those isoforms contained PTCs that should be targeted by UPF1
[44]. An impairment in splicing can also be due to mutations in splicing-related factors, such as U2AF1 (the U2 auxiliary factor 35 kDa subunit) and SF3B1 (splicing factor 3B subunit 1), frequently found in several types of solid tumors. U2AF1 and SF3B1 are the core components of spliceosomes that have been proposed as novel therapeutic targets for cancer
[62].
Numerous somatic mutations of the
UPF1 gene have been ascertained in pancreatic adenosquamous carcinoma (ASC) and stated as the first known unique molecular markers of ASC
[48]. The perturbation of NMD resulting from those tumor-specific mutations significantly increases the number of aberrant mRNAs that should be targeted by UPF1-mediated NMD. Notably, other NMD factors, namely,
UPF2, UPF3A and
UPF3B genes, did not display any detectable mutations in analyzed ASC samples. The point mutations gathered in two regions, one embracing exons 10 and 11 and intron 10 in RNA the helicase domain, and a second one consisting of exons 21–23 coding the SQ domain and the ST-Q motif and introns 21 and 22. These mutations are equally distributed among exons and introns, triggering the alternative splicing of
UPF1 pre-mRNAs by disrupting intrinsic splicing enhancers (ISEs) and exonic splicing enhancers (ESEs)
[48]. Those unique mutations are beneficial for ASC diagnosis and create the possibility for NMD substrate-targeted therapies.
The epithelial–mesenchymal transition (EMT) contributes to tumor metastasis, and is regulated by various signaling pathways
[33][63]. One of the most important is the TGF-β (transforming growth factor beta) signaling pathway, which induces the EMT through activating Smad signaling
[64][65]. Overexpressed UPF1 inhibits TGF-β signaling component genes,
MIXL1 and
SOX17. Upregulated UPF1 decreases the expression of Smad2/3 proteins, which, in turn, leads to the inhibition of TGF-β signaling
[46]. Furthermore, UPF1 alters Smad2/3 phosphorylation, which is required for signal transduction
[42]. In colorectal cancer (CRC) tissues and cell lines, UPF1 has been found to be significantly upregulated and exhibit a positive correlation with lymph node metastasis and shorter survival, thus, acting as an oncogene. UPF1 knockout (KO) in colorectal cell lines increases the number of cells in the S phase, therefore, UPF1 promotes cell cycle progression. Furthermore, UPF1 KO promotes apoptosis via increasing DNA damage and inhibiting cell migration, invasion and EMT, as it leads to a higher expression of the epithelial marker E-cadherin and decreased levels of the mesenchymal marker vimentin
[33]. UPF1 in CRC can act as a promising diagnostic marker and target for novel therapies.
The upregulation of UPF1 in CRC also leads to chemoresistance in vivo and in vitro to oxaliplatin, a third-generation platinum coordination complex, used for treatment in several types of cancer. Chemoresistance results from the SMG1-dependent phosphorylation of the human topoisomerase II-α (TOP2A) and the maintenance of cell stemness
[32]. TOP2A organizes the genome structure, promotes chromosome segregation and is overexpressed in multiple tumors, leading to aggressive phenotypes of the disease and poor prognosis
[65][66]. Post-translational modifications, such as phosphorylation, ubiquitination and SUMOylation, regulate TOP2A activity
[67]. SMG1 directly phosphorylates UPF1 and possibly induces TOP2A phosphorylation through UPF1. Moreover, UPF1 enhances the stem phenotype of CRC cells in a TOP2A-dependent manner. The underlying mechanism of oxaliplatin chemoresistance possibly arises from the attenuation of DNA damage resulting from TOP2A, induced by oxaliplatin as the phosphorylation of TOP2A increases its enzymatic activity. TOP2A was also proved to be upregulated in CRC. Notably, this chemoresistance was subverted with TOP2A silencing
[32]. These results pointed out new possible targets for decreasing oxaliplatin chemoresistance in CRC patients.
LncRNA and microRNA can interact with UPF1, resulting in its tumor-suppressive functions in various cancers
[68]. In HCC, UPF1 interacts with lncRNA SNHG6 (small nucleolar RNA host gene 6) and suppresses cell proliferation and migration through inhibiting the TGF-β/Smad pathway. SNHG6 represses Smad7 expression and, in turn, induces Smad2/3 phosphorylation
[40][42]. Moreover, in CRC, SNHG6 KO led to decreased UPF1 and p-Smad2/3 levels
[34]. Another lncRNA engaged in the TGF-β/Smad pathway is SNAI3-AS1, which mediates cell invasion. After the direct interaction of SNAI3-AS1 with UPF1, the tumorigenesis of HCC is suppressed. Additionally, SNAI3-AS1 KO significantly decreases the levels of p-Smad2/3
[69]. Upregulated in HCC miR-1468, which promotes cell proliferation and colony formation, targets UPF1, leading to its downregulation in HCC. The plasmacytoma variant translocation 1 (PVT1) lncRNA upregulated in breast cancer (BC) has been proposed to act as an oncogene through binding miR-128-3p and UPF1 and promoting EMT and, thus, proliferation and metastasis
[31]. Sponging miR-128-3p via competitive binding with PVT1 leads to the upregulation of FOXQ1, which is responsible for inducing EMT through e-cadherin repression
[70]. Additionally, miR-128-3p was found to be downregulated in BC
[31]. Therefore, PVT1 can serve as a potential therapeutic target in BC. In GC, where UPF1 expression is downregulated due to promoter hypermethylation, a negative correlation between the lncRNA MALAT 1 (metastasis-associated lung adenocarcinoma transcript 1) has been observed. The UPF1 inhibition of gastric cancer progression was reduced by high levels of MALAT1, demonstrating a promising target for gastric cancer treatment
[37]. Moreover, the miR-seq data show that the degradation of nearly 50% of potential TumiD substrates in human T24 bladder cancer cells was enhanced through UPF1. UPF1 causes the dissociation of miRNA from their mRNA targets, making them vulnerable to TumiD. One example of targeted miRNA is miR-31-5p, which has been correlated with aggressive tumor phenotypes, cell invasion and poor prognosis in breast and bladder cancer. In the case of BC, miR-31-5p is epigenetically silenced, while, in bladder cancer, TumiD plays a regulatory role in determining the miR-31-5p levels. By reducing miRNA amounts, oncogenic genes can be upregulated, which, in turn, leads to tumorigenesis and metastasis
[30].
2.2. UPF1 in Neurological Disorders
NMD targets are expressed throughout the brain, with more than 90% identified in the cerebral cortex, which is responsible for cognitive functions, such as attention, memory and overall consciousness. Many of those targets are significantly increased by UPF1 knockdown (KD), and are mutated or misregulated in neuronal diseases, such as spastic paraparesis, amyotrophic lateral sclerosis (ALS), frontotemporal dementia (FTD) and autism spectrum disorder
[27]. Recently, the fragile X mental retardation protein FMRP, of which the loss of function is the main cause of intellectual disability and autism spectrum disorders in fragile X syndrome (FXS)
[71][72], has been identified as a direct NMD-activated phosphorylated UPF1 interactor, influencing its activity by acting as its repressor. That interplay led to the conclusion that the downregulation of substantial neuronal mRNA in FXS is caused by the stimulation of NMD through FMRP loss
[27]. The neuroprotective function through RNA binding and helicase activity of UPF1 independent of the NMD pathway has been shown in several in vitro and in vivo models of ALS
[26]. Overexpressed UPF1 mitigates the neurotoxicity of a G
4C
2 hexanucleotide repeat expansion in the
C9orf72 gene, ascertained to be the most common factor inducing familial and sporadic ALS and FTD
[51][52].
Spinal muscular atrophy (SMA) is a neurodegenerative disorder caused by mutations in the
SMN1 (
survival motor neuron 1) gene, affecting splicing and leading to the production of PTC harboring less stable transcripts extensively targeted by NMD, aggravating the disease phenotype
[53][54]. Some truncated proteins, despite being able to partially retain their functionality, are degraded by NMD, which causes haploinsufficiency
[73].
Moreover, the link between NMD and epileptogenesis has been examined
[55]. Mooney et al. investigated UPF1 and its phosphorylated form, as well as UPF2 and UPF3B levels in a mouse hippocampus after status epilepticus. They described an increase in UPF1, phosphorylated UPF1 and UPF2 in their model, and a mainly neuronal distribution of UPF1. They also ascertained higher levels of UPF1 in human hippocampi from patients with temporal lobe epilepsy (TLE). As miR-128, which targets the NMD system, including UPF1, is decreased in human epilepsy
[74], it can be the cause of increased UPF1 levels. Additionally, the increased binding of UPF1 to the 3′ UTR regions of transcripts in mice has been presented. Mice treated with an NMD inhibitor, NMDI14
[75], had less spontaneous seizures and lower daily seizure rates. All those results suggested an elevated NMD associated with status epilepticus in the hippocampus, leading to epilepsy emergence and progression. The NMD system could possibly act as a target for seizure prevention.
2.3. UPF1 in Viral Infections
As stated above, UPF1 plays an essential role in viral infection development. In 2022, Fang et al. analyzed the interactome of the Ebola virus (EBOV) polymerase, and found that UPF1, collectively with another mRNA decay factor, GSPT1 (G1 to S phase transition protein 1 homolog), interacts with EBOV polymerase to promote viral replication. At the onset of the infection, UPF1 leads to a reduction in vRNA and mRNAs levels and fewer cells that are infected, while GSPT1 KD decreased the vRNA level but increased mRNA levels. In later infection, UPF1 and GSPT1 are hijacked, and promote viral replication
[56]. In a similar manner,
human immunodeficiency virus (HIV) exploits UPF1 to increase its infectivity. The depletion of UPF1 through siRNA reduces the infectivity of HIV virions by altering their reverse transcription. UPF1 is crucial for virus assembly and its function seems to be independent from NMD
[57]. In the future, mutants of UPF1 with impaired ATP-binding or hydrolysis activity could serve as inhibitors of HIV virion infectivity. Additionally, genomes of viruses can also be targets to alternative UPF1-mediated SMD and SRD pathways
[28].
2.4. UPF1 in Antiprion Systems
Prions are differed, infectious forms of native proteins. They were first discovered in sheep, where they cause fatal, transmissible encephalopathy named scrapie
[76]. In the human infectious form of PrP, protein causes a neurodegenerative disorder called Creutzfeldt–Jakob disease and its variants by forming insoluble amyloid plaques
[77]. Yeast possesses numerous variants of single-prion proteins
[78], which, in their native form, function in nitrogen metabolism (Ure2), translation termination (Sup35) and provide better adaptation to environmental conditions by facilitating amyloid formation (Rnq1). The infectious forms created by those proteins are called (URE3), (PSI+) and (RNQ+), respectively
[79][80]. Yeasts are an easy, cheap and safe model to study prions, creating a good alternative to animal models that sometimes can raise ethical questions. Parallelly, yeast has numerous antiprion systems, reducing amyloid formation and curing variants that do form prions. Son and Wickner discovered the link between Upf1 and (PSI+) levels with general screening for antiprion proteins. In the absence of either Upf1, Upf2 or Upf3, the generation of (PSI+) was elevated 10- to 15-fold. A direct interaction was proved, and the mechanism by which prions are cured was proposed. Upf1 monomers compete with the amyloid fibers of Sup35 or bind to the filament ends, preventing fiber growth
[58]. NMD is conserved from yeasts to humans, and detailed analyses of the analogs of homologous systems can be applied to the treatment of neurodegenerative prion disorders in humans by enhancing native interactions
[29][58]. Additionally, those studies suggest rather prion- or even variant-specific interactions of factors engaged in antiprion systems.