Hypoxia-mediated regulation of mRNA metabolism: from transcription to stability.

The transcription of DNA into mRNA is essential for gene expression and function. Hypoxia exposure induces transcriptional changes involving various aspects of the transcription process includes transcription initiation, elongation, termination, and posttranscriptional modification [ 17 ]. In eukaryotic cells, RNA polymerase II (RNA Pol II) mediates gene transcription by recognizing regulatory elements such as promoters and terminators. The dynamics of its function significantly affect cellular physiological activities and developmental processes. Under normal physiological conditions, prior to transcription initiation, RNA Pol II interacts with general transcription factors (GTFs), including TFIIA, TFIIB, TFIID, TFIIE, TFIIF, and TFIIH. RNA Pol II recognizes and binds to promoter regions upstream of the transcription start site and assembles to form the preinitiation complex, which has a volume that exceeds that of the ribosome. Then, RNA Pol II catalyzes transcription initiation [ 18 , 19 ]. Hypoxic environments can significantly remodel gene transcription initiation patterns. Among them, HIF-1 is a core molecule that regulates the expression of hypoxia-responsive genes [ 20 ]. Under normoxia, proline residues in the C-terminal oxygen-dependent degradation domain (ODD) of HIF-1 are hydroxylated by prolyl hydroxylases (PHDs) and interact with von Hippel–Lindau tumor suppressor protein (pVHL) to recruit the E3 ubiquitin ligase complex, which mediates HIF-1α ubiquitination, ultimately leading to HIF-1α degradation by the proteasome [ 21 , 22 ]. The inhibition of PHD activity under hypoxia leads to the stable accumulation and translocation of the HIF-1α protein to the nucleus, where it interacts with HIF-1β, which is also known as aryl hydrocarbon receptor nuclear translocator (ARNT), to form a heterodimeric HIF-1 complex [ 23 – 25 ]. Hydroxylation happens on proline residues facilitates HIF-1 recognition and binding to hypoxia response elements (HREs), which are located in the promoters of target genes. This occurs through interactions with transcriptional cofactors, such as P300/CREB-binding protein (CBP) [ 26 ]. Upon binding to HREs, HIF-1 can unlock stalled RNA Pol II in the promoter region, collaborate with GTFs (e.g., TFIID) to remodel the transcription initiation complex, and ultimately drive the transcriptional activation of hypoxia-responsive genes [ 26 – 28 ] (Fig.  1 ). Fig. 1 Transcription processes under normoxia and hypoxia. Created in BioGDP ( https://biogdp.com ). Under normoxia, RNA Pol II interacts with GTF before transcription begins as follows: TFIID recognizes and binds to the TATA box or other core promoter elements in the promoter region through its subunit TBP (TATA-binding protein), which is subsequently stabilized by TFIIA [ 29 ]. Moreover, TFIIB mediates the binding of RNA Pol II to the promoter region. TFIIF further stabilizes the interaction between the two [ 29 ]. Finally, TFIIEβ binds to the preinitiation complex and recruits TFIIEα and TFIIH to the start site. TFIIH induces conformational changes by unwinding DNA double strands through ATPase activity and phosphorylating the RNA Pol II carboxy-terminal structural domain (CTD). During this process, TFIIEα is released from the preinitiation complex prior to DNA opening, whereas TFIIEβ dissociates after DNA opening to initiate transcription [ 30 , 31 ]. GTFs: general transcription factors; TFII: transcription factor II; DSIF: DRB sensitivity-inducing factor; NELF: negative elongation factor; CDK7: cyclin-dependent kinase 7; CE: capping enzyme; CPSF: cleavage and polyadenylation specificity factor; CBP: CREB-binding protein; TRIM28: tripartite motif-containing 28; CDK9: cyclin-dependent kinase 9 Transcription processes under normoxia and hypoxia. Created in BioGDP ( https://biogdp.com ). Under normoxia, RNA Pol II interacts with GTF before transcription begins as follows: TFIID recognizes and binds to the TATA box or other core promoter elements in the promoter region through its subunit TBP (TATA-binding protein), which is subsequently stabilized by TFIIA [ 29 ]. Moreover, TFIIB mediates the binding of RNA Pol II to the promoter region. TFIIF further stabilizes the interaction between the two [ 29 ]. Finally, TFIIEβ binds to the preinitiation complex and recruits TFIIEα and TFIIH to the start site. TFIIH induces conformational changes by unwinding DNA double strands through ATPase activity and phosphorylating the RNA Pol II carboxy-terminal structural domain (CTD). During this process, TFIIEα is released from the preinitiation complex prior to DNA opening, whereas TFIIEβ dissociates after DNA opening to initiate transcription [ 30 , 31 ]. GTFs: general transcription factors; TFII: transcription factor II; DSIF: DRB sensitivity-inducing factor; NELF: negative elongation factor; CDK7: cyclin-dependent kinase 7; CE: capping enzyme; CPSF: cleavage and polyadenylation specificity factor; CBP: CREB-binding protein; TRIM28: tripartite motif-containing 28; CDK9: cyclin-dependent kinase 9 The well-characterized HIF isoforms, HIF-1 and HIF-2, share high homology within their oxygen-dependent degradation domains and both activate transcriptional programs under hypoxia. However, they exhibit marked differences in their expression profiles, oxygen sensitivity, activation kinetics, and target gene specificity [ 32 , 33 ]. HIF-1α is ubiquitously expressed across most adult mammalian tissues, whereas HIF-2α demonstrates a more restricted expression pattern, primarily localizing to endothelial cells, hepatocytes, and epithelial cells of the intestine, pancreas, and alveoli [ 34 , 35 ]. Their induction is differentially regulated by the severity and duration of hypoxic exposure. HIF-1α is predominantly responsible for the acute hypoxia response, accumulating rapidly and to high levels upon oxygen deprivation [ 33 ]. In contrast, HIF-2α is stabilized during chronic hypoxia and is identified as the principal mediator of adaptation to prolonged low-oxygen conditions, such as those encountered at high altitudes [ 36 ]. In contrast to the well-defined roles of HIF-1 and HIF-2, the function of HIF-3α is less established. Emerging evidence suggests a dual role, wherein HIF-3α can bind HREs but is a weak transcriptional activator. Certain isoforms, most notably the inhibitory PAS domain protein, a splice variant of HIF-3α, function as dominant-negative regulators by suppressing HIF-1 and HIF-2 signaling [ 37 ]. When the nascent transcript (the product of transcription) reaches a length of approximately 25 nucleotides, the preinitiation complex undergoes a conformational transition to form a stable elongation complex, also known as a transcription bubble. At this point, transcription enters the elongation phase [ 38 ]. Gene transcription elongation requires an initiation‒elongation transition phase, so it can be divided into two subphases: ‘early elongation’ and ‘productive elongation’ [ 39 ]. Following transcription initiation, the DRB sensitivity-inducing factor (DSIF) binds to RNA Pol II, and the negative elongation factor (NELF) is recruited to chromatin. This causes Pol II to pause at the proximal region of the promoter [ 40 ]. Pausing is a key regulatory mechanism that controls the rate of RNA synthesis [ 41 ]. The pause can then be released to allow continued extension, or it can attenuate gene expression by terminating transcription prematurely at the promoter-proximal end [ 42 ]. The stagnation of early elongation allows the capping enzyme to recruit the capping mechanism [ 39 ]. The serine at position 5 of the RNA Pol II C-terminal domain (CTD) is phosphorylated by the TFIIH kinase subunit cyclin-dependent kinase 7 (CDK7) [ 43 – 45 ]. The phosphorylated RNA Pol II CTD cooperates with the suppressor of Ty 5 (Spt5) subunit of DSIF to synergistically recruit capping enzymes to promote 5' end capping [ 39 , 44 ]. However, under hypoxic conditions, RNA Pol II pauses 30–60 nt downstream of the transcription start site after transcription initiation because of its interaction with NELF and DSIF. This step is rate limiting for 40%−70% of gene transcription [ 46 – 48 ]. Additionally, hypoxic signaling induces hexamethylene bis-acetamide inducible protein 1 (HEXIM1), an inhibitor of P-TEFb, as well as histone deacetylase (HDAC)-dependent deacetylation of cyclin-dependent kinase 9 (Cdk9) and Cyclin T1. These changes alter the functional homeostasis of P-TEFb, leading to transcriptional repression [ 49 ]. Furthermore, under hypoxia, HIF-1α and tripartite motif-containing 28 (TRIM28) form a DNA-dependent protein kinase heterotrimer with catalytic activity. This heterotrimer then phosphorylates TRIM28 at serine 824, enabling the recruitment of CDK9. CDK9 subsequently phosphorylates serine 2 (Ser2) of the Pol II large subunit CTD as well as NELF to release paused Pol II, thereby stimulating productive transcriptional elongation [ 50 ]. HIF-1α also induces binding of CDK8-Mediator and the Super Elongation Complex (SEC), containing ALF transcription elongation factor 4 (AFF4) and CDK9, to alleviate RNA Pol II pausing [ 51 ]. This mechanism guarantees specific gene transcription under hypoxic conditions. Histone modifications are closely related to transcriptional regulation and participate in the regulation of cell fate, development, and the occurrence and development of diseases. In tumor cells, KAP1 (KRAB-associated protein 1)/TRIM28 regulates the continuous entry and exit of RNA polymerase II from the paused state by binding to the hypoacetylated histone 4 tails at the promoter, thereby achieving CDK9-dependent, gene-selective pause release [ 52 ]; In addition, there is a positive correlation between the activity of P300/CBP and histone acetylation. The activity of P300/CBP can promote the recruitment of RNA polymerase II (RNAPII) and the release of pausing mediated by histone acetylation [ 53 , 54 ]. The Integrator-PP2A complex (INTAC) can regulate the pause-release transition of RNA polymerase II by removing histone H3K4 methylation [ 55 , 56 ]. It has also been found in cancer cells such as human colon cancer that the phosphorylation of the CTR1/CTR2 region of CDK9 can coordinately control transcriptional pausing. In addition, the phosphorylation of the RNA—binding linker of SPT5 can promote RNAPII to end transcriptional pausing at the promoter—proximal region [ 57 , 58 ]. Currently, there are relatively few studies on the involvement of chromatin modifications in the regulation of transcriptional pausing under hypoxic conditions. Further in—depth exploration of the relationship between the two in the future may provide new therapeutic directions for the treatment of cancers and other related diseases. Primary transcripts need to be processed and modified into mature mRNAs to function properly. Most posttranscriptional processing is highly synergistic with the transcriptional process (also known as cotranscriptional processing). This includes capping at the 5' end of eukaryotic genes, splicing, and adding a poly(adenylate) (poly(A)) tail at the 3' end [ 42 ]. Alternative splicing is regulated by other splicing regulators that interact with the spliceosome. One of the most typical regulatory factors is serine-arginine-rich (SR) proteins [ 59 ]. SR proteins can interact with U2 auxiliary factor heterodimers and U1 70 K proteins through their C-terminal arginine–serine structural domains, linking both sides of the exon and binding to exonic splicing enhancers to facilitate exon recognition [ 60 , 61 ]. Heterogeneous nuclear ribonucleoproteins (hnRNPs) constitute another well-characterized group of splicing regulators [ 62 ]. The binding of hnRNPs to exonic splicing silencers inhibited the binding of SR proteins to exonic splicing enhancers and covered the splice site [ 63 ]. Some transcription and splicing factors are also activated under hypoxia. Although HIF-1 plays an important role in the cellular response to hypoxia, studies have demonstrated that alternative splicing of HIF-1α and HIF-3α pre-mRNAs in humans indicates that the HIF-mediated response to hypoxia is regulated at both the transcriptional and posttranscriptional levels [ 64 , 65 ]. An increasing number of genes whose splicing depends on oxygen tension are being identified. Previous studies have confirmed that pre-mRNA splicing plays an important role in cellular adaptation to hypoxia [ 66 ]. Among these studies, HIF-1 has been shown to be indirectly involved in hypoxia-dependent splicing regulation [ 27 , 61 ]. Research has shown that an increase in the transcription level of the HIF-1-regulated CDC-like kinase 1 ( CLK1) gene causes an increase in CLK1 protein levels in hypoxic cells. This increases the expression of the SR protein kinase and alters the activity of the SR protein by increasing its phosphorylation level, mainly by excessively phosphorylating its arginine-serine domain [ 67 ]. In hypoxia, hyperphosphorylated SR proteins appear to be recruited to pre-mRNA molecules, where they participate in splicing to produce mRNAs [ 68 ]. The increased CLK and SR protein kinase expression levels detected in hypoxic HeLa cells indicate that increased SR protein kinase activity is necessary for cells to adapt to reduced oxygen tension and change pre-mRNA splicing. Therefore, hypoxia can enhance the interaction between SR proteins and specific RNA sequences, leading to the need for cells to synthesize products to adapt to changes in the surrounding environment [ 67 ]. These sequences do not interact or interact very inefficiently in normoxic cells [ 61 , 67 ]. In addition, there are several reports on hypoxia and alternative splicing. It is divided into two parts: HIF-dependent and HIF-independent. The following studies have been conducted on HIF dependency: HIF-1α-dependent alternative splicing of carcinoembryonic antigen-related cell adhesion molecule 1 is an essential regulatory mechanism for ischemic stress resistance in liver transplantation [ 69 ]. miR-155 deletion facilitates the activation of fructose metabolism in hypoxic cardiac fibroblasts through regulating alternative splicing of HIF-1α pre-mRNA and thus circHIF-1ɑ formation [ 70 ]. HIF-1α-dependent alternative splicing of carcinoembryonic antigen-related cell adhesion molecule 1 is an essential regulatory mechanism for ischemic stress resistance in liver transplantation [ 69 ]. In addition to HIF-dependent studies, there are also HIF-independent studies. ROS-dependent remodeling of the nuclear architecture can promote production of splicing variants that facilitate adaptation to hypoxia [ 71 ]. Hypoxia-induced loss of serine and arginine rich splicing factor 2-dependent DNA methylation promotes CCCTC-binding factor-mediated alternative splicing of VEGFA in breast cancer [ 72 ]. Currently, how pre-mRNAs dependent on oxygen tension are regulated remains unclear. More studies are needed to identify genes and factors involved in hypoxia-dependent alternative pre-mRNA splicing regulation [ 61 ]. mRNA translation relies on the involvement of eukaryotic translation factors, termed eukaryotic translation initiation, elongation and termination/release factors (eIFs, eEFs, and eRFs), which transiently associate with the ribosome [ 73 ]. Under hypoxic conditions, the metabolic rate decreases, greatly inhibiting the rate of protein translation, at which point the cell maintains ATP homeostasis by regulating the translation rate to ensure survival [ 12 , 74 – 78 ]. Notably, during severe hypoxia, the ATP demand for protein synthesis decreases to approximately 7% of that in normoxic cells, which is correlated with a dramatic reduction in protein translation rates [ 79 ]. On the other hand, under hypoxic conditions, the translation efficiency of genes involved in pathways such as glycolysis increases to ensure cellular survival and functional maintenance. Ribosome profiling results also demonstrate that the increase in ribosome densities at 5' UTR is positively correlated with the presence of upstream open reading frames in the 5' UTR of mRNAs under hypoxic stress [ 80 ]. However, some gene translation is maintained under hypoxia, even higher than under normoxia. These findings indicate that the effect of hypoxia on translation is more complex than that of energy regulation alone. Translation initiation can be broadly divided into two separate modes: canonical and noncanonical. In the canonical mode, cap-dependent translation is the primary mechanism for cellular protein synthesis [ 81 , 82 ]. This process begins with the recognition and binding of the 5' cap (m7GpppN) domain of mRNA by the eIF4E complex, which comprises eIF4E, eIF4G, and eIF4A, in a cap-dependent manner [ 83 – 85 ]. The assembly of the eIF4F complex is the rate-limiting step in cap-dependent translation [ 86 ]. After binding to the cap, eIF4E recruits the 43S preinitiation complex to the 5' end of the mRNA, activating mRNA translation [ 83 – 85 ]. The 43S preinitiation complex consists of a ternary complex (eIF2•Met-tRNAi•GTP), which includes Met-tRNAi bound to eIF2-GTP (but not eIF2-GDP), as well as eIF1, eIF1A, eIF3, and the 40S ribosomal subunit [ 12 , 83 , 84 ]. After recruitment, the 43S preinitiation complex scans the mRNA from 5' to 3' until it recognizes the AUG start codon [ 83 – 85 ]. The large ribosomal subunit (60S) then joins to form an elongation-competent ribosome (80S), which comprises the 60S and 40S ribosomal subunits and facilitates peptide elongation [ 87 ]. This process results in the release of eIFs and termination of the initiation process [ 12 ]. The noncanonical mode is not mediated by the eIF4F complex. Instead, it is achieved through the binding of ribosomes to mRNA transcripts that contain an IRES [ 81 , 82 , 86 ]. This mechanism of translation is particularly relevant under stressful conditions, such as hypoxia, when global cap-dependent protein synthesis is inhibited [ 88 ]. While the mechanisms underlying noncanonical translation initiation vary in terms of categories involving different eIFs, they share common features such as cap recognition and ribosome scanning, as well as other conditions [ 82 ]. Additionally, literature reports indicate that HIF function does not depend on the IRES. Its mRNA 5' and 3' UTR regions interact with non-coding RNAs or RNA-binding proteins to jointly regulate translation efficiency, thereby ensuring the normal execution of its functions [ 89 ]. In addition to noncanonical translation initiation mediated by IRESs, other known approaches include m 6 A translation initiation, eIF3d translation initiation, and ribosome shunting initiation [ 90 ]. The initiation of mRNA translation is the most crucial step in regulating protein synthesis in hypoxia [ 12 ]. Studies have shown that hypoxia primarily prevents eukaryotic translation initiation through two distinct pathways: one propagated by stress-responsive protein kinase R (PKR)-like endoplasmic reticulum kinase (PERK) and the other by the mammalian target of rapamycin (mTOR) [ 91 – 96 ]. During short-term hypoxia, PERK is rapidly activated and hyperphosphorylates eIF2α. This prevents the formation of the ternary complex (eIF2, Met-tRNAi, and GTP) and inhibits the initiation of mRNA translation. However, eIF2α phosphorylation partially recovers after four to eight hours [ 91 , 97 ]. Short-term inhibition of mRNA translation under severe and moderate hypoxia has been attributed to eIF2α phosphorylation [ 77 , 97 , 98 ]. Phosphorylated eIF2α blocks translation initiation by inhibiting the GDP‒GTP exchange catalyzed by eIF2B and impeding the release of initiation factors from the ribosome. Specifically, under prolonged hypoxia, the global inhibition of protein synthesis may be related to PERK-independent eIF2α phosphorylation [ 12 ]. Prolonged hypoxia disrupts the eIF4F complex, promoting increased binding of 4E-BP to eIF4E. This inhibits the association of mTOR with eIF4G, ultimately preventing cap-dependent translation [ 94 , 95 , 99 , 100 ] (Fig.  2 ). Fig. 2 Canonical translation process under normoxia and hypoxia. Created in BioGDP ( https://biogdp.com ). The translation process begins with the recognition and binding of the 5' cap (m7GpppN) domain of mRNA by the eIF4E complex, which comprises eIF4E, eIF4G, and eIF4A, in a cap-dependent manner [ 83 – 85 ]. The assembly of the eIF4F complex is the rate-limiting step in cap-dependent translation [ 86 ]. After binding to the cap, eIF4E recruits the 43S preinitiation complex to the 5' end of the mRNA, activating mRNA translation [ 83 – 85 ]. The 43S preinitiation complex consists of a ternary complex (eIF2•Met-tRNAi•GTP), which includes Met-tRNAi bound to eIF2-GTP (but not eIF2-GDP), as well as eIF1, eIF1A, eIF3, and the 40S ribosomal subunit [ 12 , 83 , 84 ]. After recruitment, the 43S preinitiation complex scans the mRNA from 5' to 3' until it recognizes the AUG start codon [ 83 – 85 ]. The large ribosomal subunit (60S) then joins to form an elongation-competent ribosome (80S), which comprises the 60S and 40S ribosomal subunits and facilitates peptide elongation [ 87 ]. mTOR: mammalian target of rapamycin; 4E-BP: eIF4E-binding protein; eIF: eukaryotic translation initiation factor; eEF: eukaryotic translation elongation factor; eRF: eukaryotic translation termination/release factor; GTP: guanosine triphosphate; GDP: guanosine diphosphate; eEF2K: eEF2 kinase; PHD: prolyl hydroxylase; RMB4: RNA binding motif protein 4; PERK: PKR-like endoplasmic reticulum kinase Canonical translation process under normoxia and hypoxia. Created in BioGDP ( https://biogdp.com ). The translation process begins with the recognition and binding of the 5' cap (m7GpppN) domain of mRNA by the eIF4E complex, which comprises eIF4E, eIF4G, and eIF4A, in a cap-dependent manner [ 83 – 85 ]. The assembly of the eIF4F complex is the rate-limiting step in cap-dependent translation [ 86 ]. After binding to the cap, eIF4E recruits the 43S preinitiation complex to the 5' end of the mRNA, activating mRNA translation [ 83 – 85 ]. The 43S preinitiation complex consists of a ternary complex (eIF2•Met-tRNAi•GTP), which includes Met-tRNAi bound to eIF2-GTP (but not eIF2-GDP), as well as eIF1, eIF1A, eIF3, and the 40S ribosomal subunit [ 12 , 83 , 84 ]. After recruitment, the 43S preinitiation complex scans the mRNA from 5' to 3' until it recognizes the AUG start codon [ 83 – 85 ]. The large ribosomal subunit (60S) then joins to form an elongation-competent ribosome (80S), which comprises the 60S and 40S ribosomal subunits and facilitates peptide elongation [ 87 ]. mTOR: mammalian target of rapamycin; 4E-BP: eIF4E-binding protein; eIF: eukaryotic translation initiation factor; eEF: eukaryotic translation elongation factor; eRF: eukaryotic translation termination/release factor; GTP: guanosine triphosphate; GDP: guanosine diphosphate; eEF2K: eEF2 kinase; PHD: prolyl hydroxylase; RMB4: RNA binding motif protein 4; PERK: PKR-like endoplasmic reticulum kinase In addition to these two main pathways, other factors are also associated with translation initiation under hypoxia. Liang et al. [ 101 ] reported that the cap-binding translation initiation factor eIF4EHP, a homolog of human eIF4E2, plays a key role in 3' UTR-dependent translation under hypoxic conditions and is necessary for the 3' UTR-dependent translation of Ldh mRNA under hypoxic conditions. Under hypoxia, HIF-2α not only performs its canonical transcription factor function but also participates in regulating the cap-dependent translation initiation process [ 102 ]. Specifically, hypoxia induces HIF-2α to form a special translation initiation complex with the RNA-binding protein RMB4(RNA binding motif protein 4) and the eukaryotic translation initiation factor eIF4E2 [ 102 ]. This complex recognizes specific mRNA molecules by binding to the rHRE sequence (characterized as CG(G)) in the 3' UTR of the mRNA and subsequently interacts with the cap structure at the 5' end of the mRNA [ 102 ]. On this basis, the complex binds to translation initiation factors, such as eIF4A and eIF4G3. These factors assemble to form the eIF4F complex, which ultimately mediates ribosome recruitment and translation initiation [ 102 – 104 ]. The translational extension phase requires the translocation of the ribosome along the mRNA, the decoding of mRNA codons, and the formation of a peptide bond to extend the newly synthesized polypeptide chain [ 105 , 106 ]. The process relies on the transfer RNA (tRNA) molecules to carry amino acids to make up the growing polypeptide chain [ 82 ]. eEF1α mediates the entry of aminoacyl-tRNA into the vacant A (aminoacyl) site of the ribosome. As the ribosome moves in a 5' to 3' direction along the mRNA, the tRNA at the A site is moved to the P (peptidyl) site. Subsequently, eEF1β catalyzes the conversion of GDP bound to eEF1α into GTP and reconstitutes the dimer of eEF1α and eEF1β (eEF1). Then, it moves to the E (exit) site. eEF2 has translocase activity and is fed by hydrolyzed GTP, which causes the ribosome to move down one codon along the mRNA. This allows the next codon on the mRNA to enter the A site, where it continues to participate in elongation and peptide bond formation [ 82 ]. Under hypoxia, protein synthesis is regulated not only at the initiation stage but also dynamically during translation elongation through the oxygen-sensitive phosphorylation of eEF2 [ 13 , 107 ]. As a ribosome translocase, eEF2 undergoes phosphorylation at its Thr 56 residue, which is mediated by eEF2 kinase (eEF2K), a process tightly controlled by PHD and oxygen levels [ 108 , 109 ]. Under normoxia, PHD2 suppresses eEF2K activity via hydroxylation, maintaining eEF2 in a low-phosphorylation state to sustain normal elongation. However, hypoxia inactivates PHD2, leading to eEF2K activation, rapid eEF2 phosphorylation, and impaired ribosome binding, thereby blocking elongation to reduce energy expenditure [ 13 , 109 ]. eEF2K, a calcium (Ca 2+ )/calmodulin (CaM)-dependent α-kinase, integrates signals from multiple pathways: Ca 2+ signaling induces autophosphorylation via CaM binding; energy stress-activated AMP-activated protein kinase (AMPK) or glycogen synthase kinase 3 (GSK3) enhances eEF2K activity by phosphorylating specific residues; conversely, mTORC1 and extracellular regulated protein kinases (ERK) phosphorylate eEF2K to inhibit its function [ 110 ]. The translation termination process is synergistically regulated by the eukaryotic release factors eRF1 and eRF3. When a termination codon (UAA, UGA, or UAG) enters the ribosomal A site, eRF1 specifically recognizes and binds to the termination signal via its highly conserved N-terminal NIKS (aspartate-isoleucine-lysine-serine) sequence. The GTPase eRF3 is subsequently recruited to the eRF1 complex. eRF3 provides the necessary energy for the hydrolysis of the ester bond between the tRNA at the end of the P site and the intact polypeptide strand. This results in the release of the nascent polypeptide from the ribosome, ultimately leading to translation termination. Notably, eRF1 function depends on hydroxylation. Jumonji domain-containing protein 4 (Jmjd4) oxygenase catalyzes this modification, which is required for stop codon recognition and translation termination [ 82 , 103 , 106 ]. Under hypoxia, eRF1 hydroxylation is significantly reduced, leading to a decrease in its ability to recognize the termination codon, which in turn inhibits translation termination and triggers continued extension of the aberrant polypeptide chain [ 103 ]. It has also been found in the acute ischemic stroke model that inhibiting the expression of eRF1 through the METTL3/miR-355/eRF1 axis can alleviate the symptoms of early acute ischemic stroke [ 111 ]. In addition, some studies have found that reducing the abundance of eRF1 by degrading it can suppress premature termination codons, providing a new therapeutic strategy for various genetic diseases caused by nonsense mutations [ 112 , 113 ]. There are multiple posttranscriptional modifications of RNA. To date, more than 100 different types of modifications have been identified [ 114 ]. These different RNA modifications can regulate a variety of biological processes, including RNA stability, nuclear export, translation efficiency, mRNA degradation, and splicing. Under hypoxia, the RNA modification landscape is dynamically regulated through modulation of the writer–eraser–reader system, thereby reprogramming posttranscriptional gene regulatory networks. Several common mRNA posttranscriptional modifications are reviewed below. In general, mRNA methylation is the addition of one or two methyl groups to specific nucleotide residues in mRNAs, such as N 1 -methyladenosine (m 1 A), 7-methylguanosine (m 7 G), 5-methylcytosine (m 5 C), N 6 -methyladenosine (m 6 A), N 6 ,2’-O-dimethyladenosine (m 6 Am), and 5-hydroxymethylcytosine (hm 5 C) [ 115 – 118 ]. The most common of these modifications are m 6 A, m 5 C and m 1 A [ 118 ]. m 6 A, the most common type of mRNA methylation, is synergistically regulated by three functional proteins: methyltransferases (writers), m 6 A-binding proteins (readers), and demethylases (erasers) [ 116 , 119 – 121 ]. Cotranscriptional modification of m 6 A is mediated by the methyltransferase complex [ 119 ]. The core components of the methyltransferase complex include methyltransferase-like 3 (METTL3, catalytic subunit), METTL14 (structural scaffold), and Wilms’ tumor 1-associated protein (WTAP) [ 122 – 124 ]. METTL3 exerts its methyl transfer activity by binding S-adenosylmethionine (SAM), whereas METTL14 enhances its catalytic efficiency by stabilizing the METTL3 conformation despite lacking the SAM-binding domain [ 119 , 125 , 126 ]. During methylation initiation, METTL3 binds to SAM, targets the DRACH shared sequence and adds a methyl group at position 6 of the adenosine in this sequence, whereas METTL14 recognizes the substrate and forms a stable heterodimer with METTL3, regulating the stability of the complex and maintaining the conformation required for METTL3 enzymatic activity [ 126 – 130 ]. The role of m 6 A in cells is related mainly to the VHL gene. Under normoxia, VHL controls m 6 A modification by coordinating the assembly of METTL3 and METTL14, thereby stabilizing phosphoinositide-3-kinase regulatory subunit 3 ( PIK3R3 ) mRNA [ 131 , 132 ]. Xue et al. [ 133 ] reported that Forkhead box M1 (FOXM1) increased the m 6 A methylation level of apolipoprotein A1 (APOA1) by inhibiting METTL3 transcription, which is involved in myopic scleral remodeling. m 6 A-binding proteins are a class of proteins that recognize m 6 A methylation sites on RNA and bind mRNAs to facilitate their nuclear export, stabilization, or degradation [ 119 ]. The canonical reader proteins YTH domain family protein 1–3 (YTHDF1-3) and YTH domain containing protein 1–2 (YTHDC1-2) are localized to the cytoplasm and nucleus, respectively [ 116 , 119 , 121 ]. In the nucleus, YTHDC1 facilitates the nuclear export of m 6 A-modified pre-mRNA by recruiting the serine/arginine-rich splicing factor 3 (SRSF3)-NXF1 (nuclear export factor 1) complex [ 119 ]. Subsequently, cytoplasmic YTHDF1 enhances the translation of m 6 A-modified mRNAs through coordinated recruitment of the small ribosomal subunit with eIF3, whereas YTHDF2 accelerates the degradation of target mRNAs by recruiting RNA decay enzymes or adaptor proteins [ 119 ]. For example, YTHDF2 may directly bind to the m 6 A modification site of the 3' UTR of epidermal growth factor receptor (EGFR) and promote the degradation of EGFR mRNA in hepatocellular carcinoma (HCC) cells, thereby inhibiting cell proliferation and growth [ 134 ]. Under hypoxia, HIF-1α upregulates YTHDF1 expression by directly binding to its promoter region. YTHDF1 subsequently binds to m 6 A sites on autophagy-related gene 2 A (ATG2A) and ATG14 mRNAs, promoting the translation of autophagy-related proteins and thereby exacerbating the malignant phenotype of HCC [ 121 ]. Under hypoxia, HIF-1α induces the upregulation of the long noncoding RNA Six-transmembrane epithelial antigen of prostate 3 antisense RNA 1 (STEAP3-AS1) . This hypoxia-induced lncRNA competitively inhibits YTHDF2 function, leading to the dissociation of YTHDF2 from STEAP3 mRNA. Consequently, STEAP3 mRNA is protected from m 6 A-mediated degradation, resulting in increased STEAP3 protein expression. This upregulation activates the Wnt/β-catenin pathway, thereby promoting metastasis in colorectal cancer [ 135 ]. In addition, researchers have recently identified a new m 6 A-binding protein present in endothelial cells, proline-rich coiled-coil 2B (PRRC2B), which plays an indispensable role in hypoxia-induced vascular remodeling [ 136 ]. RNA demethylation occurs under specific conditions and is mostly stimulus dependent [ 119 ]. The ‘eraser’ of m 6 A is a demethylase that converts m 6 A to A and may also act on the exon of m 6 A. The m 6 A demethylation enzymes include fat mass and obesity-associated protein (FTO) and AlkB homolog H5 (ALKBH5) [ 119 – 121 ]. Under hypoxia, HIF signaling activates ALKBH5, directing it to target the 3' UTR of the pluripotency factor NANOG mRNA. ALKBH5 then upregulates NANOG expression through m 6 A demethylation, thereby promoting the expansion of breast cancer stem cells [ 119 , 137 ]. In addition, hypoxia-induced ALKBH5 exacerbates the inflammatory response in rheumatoid arthritis by stabilizing cholesterol-25-hydroxylase ( CH25H ) mRNA [ 138 ] or promotes tissue recovery after myocardial infarction by maintaining Erb-B2 receptor tyrosine kinase 4 (ErbB4) mRNA stability in fibroblasts [ 139 ]. Hypoxia also induces ALKBH5 expression via a YTHDF2-m 6 A-HDAC4 (histone deacetylase type 4)-HIF-1α positive feedback loop, leading to a decrease in global m 6 A levels, which in turn enhances glycolysis and invasiveness in pancreatic cancer cells [ 140 ]. Additionally, bioinformatics analysis of single-cell RNA sequencing data of murine livers also revealed a significantly elevated expression level of the RNA demethylase ALKBH5 in the peri-central low oxygen region [ 141 ]. When exposed to external stimuli or environmental changes, different cells respond differently to m 6 A in order to perform their functions or survive [ 142 ]. It has been confirmed in multiple cancer cell lines that HIF-1α can bind to two major HREs in the METTL3 promoter, inducing its expression under hypoxic conditions. The expression of ALKBH5 and FTO increases significantly under hypoxic conditions, thereby promoting cancer progression and metastasis [ 143 – 145 ]. However, in ischemic heart disease and myocardial infarction, the expression of FTO is significantly down-regulated, suggesting its protective role in cardiac injury [ 146 ]. In addition, the regulation of YTHDF1 shows dynamic changes under hypoxic conditions, which can induce disease deterioration by increasing or decreasing its expression [ 121 , 147 ]. Overall, different cells have different requirements for different proteins to perform their functions, and the dynamic regulation of m 6 A plays an important role in this process. By increasing or decreasing the expression of related proteins such as FTO, YTHDF1 and ALKBH5, the transcription and translation efficiency of downstream target genes are changed, which in turn affects the occurrence and development of diseases. In mammalian mRNAs, the abundance of m 1 A modification sites is extremely low, approximately one-tenth the level of m 6 A modifications [ 148 , 149 ]. However, a m 1 A site with high methylation levels is present in the NADH dehydrogenase 5 ( ND5) gene encoded by the mitochondrial genome [ 148 ]. m 1 A modification is widely distributed in tRNAs and rRNAs, and recent studies have confirmed the presence of such modifications in eukaryotic mRNAs [ 116 , 119 ]. Specifically, m 1 A in mRNAs is enriched predominantly near the translation initiation site and the first splice site, with each transcript containing, on average, only one m 1 A modification site [ 116 , 149 ]. Studies addressing m 1 A modification under hypoxia are still limited. The level of m 1 A modification of mRNAs in zebrafish brain tissues was increased under normoxia and significantly decreased after hypoxia treatment, resulting in abnormal mRNA splicing and the activation of endogenous retroviral elements, suggesting that m 1 A may be involved in the posttranscriptional regulation of hypoxic stress [ 150 ]. m 5 C is a methylation of the 5th carbon atom of cytosine residues in RNA molecules and is enriched mainly in the coding sequence (CDS) and CG high pyrimidine regions of mRNAs, particularly just after start codons and before stop codons [ 151 – 153 ]. Recent studies have revealed novel functions of m 5 C modifications in the hypoxic tumor microenvironment. HIF-1α indirectly upregulates pyruvate kinase isozyme type M2 (PKM2) expression by activating the m 5 C-binding protein Aly/REF export factor (ALYREF). ALYREF then stabilizes PKM2 mRNA and binds to the m 5 C site of the 3' UTR, promoting glucometabolic reprogramming in bladder cancer [ 154 ]. m 5 C-modified pro-neurexophilin 4 ( NXPH4 ) mRNA is degraded by RNA autophagy escape and enhances HIF-1α stability by competitively inhibiting PHD4, which in turn drives malignant tumor progression [ 155 ]. m 7 G modification is a core component of the mRNA 5' end cap structure, which is highly evolutionarily conserved and is widely involved in the whole life cycle regulation of mRNA transcription processing, nucleoplasm translocation, stability maintenance and cap-dependent translation [ 119 , 151 ]. In addition to the classical 5' end cap structure, a methyltransferase complex consisting of METTL1-WD repeat structural domain 4 catalyzes m 7 G modification of the internal region of mRNAs with tRNA-like structures [ 156 ]. m 7 G modification affects gene expression by regulating key steps such as pre-mRNA splicing, nuclear export and cap-dependent protein synthesis [ 119 ]. Recent studies have revealed the potential function of m 7 G modification in hypoxic stress. HIF-1α in colorectal cancer cells can bind to the HRE in the promoter region of the m 7 G regulator METTL1, leading to transcriptional repression of METTL1, which in turn decreased the abundance of N7-methyleguanosine (m 7 G) modification in tRNA [ 157 ]. Studies on m 7 G-modified mRNAs under hypoxia are still extremely limited, and their associations with pathological processes such as metabolic reprogramming and therapeutic resistance still need to be explored in depth. N 4 -acetylcytidine (ac 4 C) is the only known acetylation in eukaryotic RNA and is widely distributed in tRNAs, 18S rRNA and mRNAs [ 158 – 161 ]. This modification is catalyzed by the N-acetyltransferase 10 (NAT10) enzyme or its homolog and is produced by the addition of an acetyl group at the 4-position of its cytidine residue [ 160 – 162 ]. NAT10 is a multifunctional enzyme endowed with both lysine acetyltransferase activity and RNA-binding capacity [ 158 , 161 , 163 ]. ac 4 C modification exerts bidirectional regulation of mRNA function through position-specific mechanisms: modifications in the CDS increase translational elongation, whereas modifications in the 5' UTR may impede ribosome scanning to suppress translational initiation [ 151 , 158 – 161 ]. Recent studies revealed a dynamic regulatory network of ac 4 C modifications in the hypoxic microenvironment. Under hypoxia, HIF-1α can transcriptionally activate NAT10 and mediate septin 9 (SEPT9) mRNA ac 4 C modification, which in turn enhances HIF-1 signaling pathway activity and promotes glucose metabolism adaptation and a hypoxia-tolerant phenotype in gastric cancer cells [ 164 ]. In cardiac ischemia‒reperfusion models, p53 transcriptionally upregulates NAT10 expression, leading to elevated ac 4 C modification levels in cardiomyocytes and exacerbating iron death-dependent injury, whereas NAT10 inhibitors significantly ameliorate this pathological process [ 165 ]. In vitro experiments revealed that the knockdown of NAT10 in renal tubular epithelial cells significantly reduced the overall level of ac 4 C modification and effectively alleviated hypoxia/reoxygenation-induced cell injury and iron death [ 162 ]. Research has shown that remimazolam ameliorates myocardial infarction by suppressing cardiomyocyte pyroptosis via inhibition of the NAT10-mediated ac4C acetylation of Nek7 (one of eleven NEK kinases found in vertebrates) [ 166 , 167 ]. The primary form of RNA editing involves the chemical modification of adenosine to inosine (A-to-I), an important posttranscriptional modification (cotranscriptional covalent modification) found in mRNAs as well as in tRNAs of both eukaryotes and prokaryotes [ 168 – 170 ]. A-to-I editing primarily converts A to I in double-stranded RNA substrates through the hydrolytic deamination of adenine bases [ 171 , 172 ]. The enzymes that catalyze this reaction are the adenosine deaminase (ADAR) family of enzymes that act on RNA [ 171 , 173 ]. All ADAR proteins share a highly conserved catalytic deaminase structural domain in their C-terminal half and multiple double-stranded RNA-binding structural domains in their N-terminal half. These domains are responsible for recognizing substrates [ 174 , 175 ]. ADAR can regulate mRNA splicing by generating or disrupting the splicing of donor or acceptor sites [ 151 ]. Since ADAR proteins recognize double-stranded RNA, the substitution of inosine for adenosine may alter the stability of the RNA structure and affect the accessibility of the splice site [ 170 ]. During acute inflammation, large amounts of inosine-containing mRNAs are produced by the upregulation of ADAR1-mediated RNA editing, which may affect inflammatory and immune responses by regulating protein production [ 176 ]. HIF-1α is an important target for A-to-I editing [ 169 ]. Under hypoxia, two negative regulators of HIF-1α, the natural antisense transcript HIF1α antisense RNA 2 (HIF1α-AS2) and the ubiquitin ligase scaffold LIM domains containing 1 (LIMD1), are directly and differentially regulated by ADAR1 [ 169 ]. HIF1α-AS2 antagonizes HIF-1α expression in the early stages of hypoxia, possibly through convergent transcriptional competition with cis-ADAR1, which in turn inhibits the transcriptional progression of the antisense gene [ 169 , 177 ]. In contrast, ADAR1 post transcriptionally regulates LIMD1 expression by impeding the cytoplasmic translocation of LIMD1 mRNA, thereby inhibiting its protein translation [ 169 ]. This multilayered regulation orchestrated by ADAR1 promotes the stabilization and accumulation of HIF-1α during hypoxia and enhances target gene induction and downstream angiogenesis [ 169 ] (Table  1 ). Table 1 Changes in mRNA modifications and the functions of related genes under hypoxic conditions (‘↑’ represents upward adjustment, ‘↓’ represents downward adjustment.) Epigenetic modification Regulation of the HIF pathway Genes that change under hypoxia Modification changes under hypoxia Changes in mRNA related to modifications under hypoxia Regulation of cellular functions Reference m6A HIF-1α↑ YTHDF1↑ ATG2A and ATG14 m6A↑ ATG2A and ATG14 mRNA translation↑ facilitating autophagy and autophagy-related malignancy of HCC [ 121 ] HIF-1α↑ YTHDF2↑ STEAP3 m6A↑ STEAP3 mRNA degradation↓ facilitating the proliferation and metastasis of colorectal cancer cells both in vitro and in vivo [ 135 ] HIF-1α and HIF-2α↑ ALKBH5↑ NANOG m6A↓ NANOG mRNA stability↑ enhancing the breast cancer stem cells enrichment [ 137 ] HIF-1α↑ ErbB4 m6A↓ ErbB4 mRNA stability↑ positively regulating post-myocardial infarction healing [ 139 ] HIF-1α↑ HDAC4 m6A↓ HDAC4 mRNA stability↑ promoting glycolytic metabolism and migration of pancreatic cancer cells [ 140 ] m1A - - m1A↓ - affecting the RNA splicing and zebrafish endogenous retroviruses [ 150 ] m5C HIF-1α↑ ALYREF↑ - PKM2 mRNA stability↑ promoting the glucose metabolism of bladder cancer and bladder cancer cells proliferation [ 154 ] ac4C HIF-1α↑ NAT10↑ SEPT9 ac4C↑ SEPT9 mRNA stability and translation efficiency↑ enhancing hypoxia tolerance in gastric cancer cells by promoting glycolysis addiction [ 164 ] - Mybbp1a ac4C↑ Mybbp1a mRNA stability↑ promoting cardiomyocyte ferroptosis to exacerbate cardiac ischemia‒reperfusion injury [ 165 ] - NCOA4 ac4C↑ NCOA4 mRNA stability↑ contributing to ischemia–reperfusion injury-induced ferroptosis in tubular epithelial cells [ 162 ] - Nek7 ac4C↑ Nek7 mRNA stability↑ promoting pyroptosis of myocardial cells after myocardial ischemia–reperfusion [ 166 ] Changes in mRNA modifications and the functions of related genes under hypoxic conditions (‘↑’ represents upward adjustment, ‘↓’ represents downward adjustment.) mRNA stability is a key link in the regulation of gene expression and directly affects the dynamic balance of cellular transcription levels [ 178 ]. mRNA stability is associated with a variety of factors, including RNA sequence elements, RNA secondary structure in transcripts, DNA elements in promoter regions, and mRNA levels [ 179 , 180 ]. Among them, the synergistic interaction of the CDS with the 3' UTR is considered to be the central mechanism determining mRNA stability [ 181 ]. Specifically, double-stranded structures at the 3' end and its neighbors, including those mediated by poly(A) tails and U-rich sequences in the 3' UTR, play a key role in maintaining mRNA stability [ 182 ]. Under hypoxia, structural changes mediated by poly(A) tails and U-rich sequences in the 3' UTR may significantly affect mRNA stability. For example, Kondo et al. [ 183 ] reported that hypoxia significantly increases the mRNA level of connective tissue growth factor ( CTGF/CCN2 ) by remodeling 3' UTR-mediated mRNA stability without requiring de novo protein synthesis. Similarly, Hsiao et al. [ 184 ] reported that under hypoxia in patients with endometriosis, the synergistic action of AU-rich element-binding factor 1 (AUF1) and microRNA-148a destabilizes DNA methyltransferase 1 ( DNMT1 ) mRNA. The stability of mRNA is associated with its structural alterations, which are largely mediated through interactions with mRNA-binding proteins that recognize specific sequences or structural motifs [ 182 ]. For example, Yang et al. [ 185 ] identified the hypoxia-induced long noncoding RNA CBSLR (cystathionine beta-synthase mRNA-destabilizing lncRNA) and demonstrated that CBSLR interacts with YTHDF2 to form a CBSLR/YTHDF2/CBS signaling axis. This axis reduces the stability of CBS mRNA by enhancing the binding of YTHDF2 to the m 6 A-modified CDS region of CBS mRNA, thereby inhibiting ferroptosis in gastric cancer cells and promoting their chemoresistance. These studies suggest that hypoxia may indirectly regulate mRNA stability by altering the CDS structure. Short-term hypoxia may enhance cellular stress tolerance by stabilizing the overall level of mRNA, whereas long-term hypoxia leads to severe changes in the entire gene expression profile of tumor cells, which in turn drives the evolution of a malignant phenotype [ 186 ]. Ju et al. [ 187 ] reported that protein arginine methyltransferase 9 suppresses the de novo synthesis of the HIF-1α protein by destabilizing HIF-1α mRNA, thereby negatively regulating cellular hypoxic responses. In addition, several studies have revealed the role of hypoxia-associated lncRNAs in the regulation of mRNAs in tumors. Ma et al. [ 188 ] reported that hypoxia-induced lncRNA BSG antisense RNA 1 (BSG-AS1) could promote HCC proliferation and migration by maintaining the stability of BSG mRNA. Qiu et al. [ 189 ] reported that the hypoxia-responsive lncRNA MIR155 Host Gene ( MIR155HG) promotes the expression of programmed cell death ligand 1 in HCC cells by increasing the stability of HIF-1α mRNA, which in turn mediates immune escape from HCC. Lin et al. [ 190 ] demonstrated that lncRNA- PMAN (plasmacytoma variant translocation 1 (PVT1) MYC-associated non-coding RNA) was highly expressed in peritoneal metastases of gastric cancer and that HIF-1α upregulated PMAN and enhanced the stability of solute carrier family 7 member 11 ( SLC7A11) mRNA by promoting the cytoplasmic localization of embryonic lethal abnormal vision-Like 1 (ELAVL1), which increased the level of l-glutathione, thereby inhibiting reactive oxygen species and iron accumulation. In contrast, Zhong et al. [ 191 ] reported that HIF-1α could increase NOD-like receptor family pyrin domain containing 3 (NLRP3) stability and promote chronic rhinosinusitis by antagonizing hypoxia-mediated inflammation-induced degradation of the NLRP3 mRNA. Together, these studies suggest that hypoxia plays an important role in modulating cellular function by regulating mRNA stability. RNA degradation is essential for RNA maturation, quality control and expression [ 192 ]. mRNA degradation, whereby the cell recycles ribonucleotides and swiftly destroys unnecessary transcripts, is primarily mediated by the combined action of endo-ribonucleases and exoribonucleases. This process may be accompanied by cotranscriptional events [ 180 ]. The half-life of an mRNA is closely related to its degradation rate and is determined by both translation-dependent and nontranslation-dependent mechanisms, and translation-dependent decay occurs primarily via deadenylation-dependent mechanisms [ 193 ]. Shorter half-life mRNAs are more easily degraded, whereas longer half-life mRNAs are relatively stable [ 193 ]. The classical degradation pathways include the deadenylation-dependent pathway and the deadenylation-independent pathway. The major decay pathway for endogenous mRNAs is deadenylation-dependent decapitation, which involves removal of the poly(A) tail from the 3' end of the mRNA [ 194 – 196 ]. Deadenylation-independent mRNA decay pathways, including deadenylation-independent decapitation and nucleic acid endonucleotide cleavage, may also lead to decay in some cases [ 193 ]. Deadenylation is achieved through deadenylases, including the poly(A) nuclease 2 (PAN2)-PAN3 and CCR4-NOT (carbon catabolite metabolite deterrent 4-TATA negative) deadenylase complexes [ 197 , 198 ]. Following deadenylation by decarboxylase (Dcp1/DCp2), the mRNA decapping complex is recruited to the 5'‐terminus of the mRNA, which functions to hydrolyze the N 7 ‐methylguanosine 5'‐cap structure [ 197 , 199 ]. The decay procedure for most mRNAs involves direct recruitment of the CCR4-NOT complex to the 3' UTR of the target mRNA via various 3' UTR-binding proteins, degrading the 3'-poly(A) tail, which is thought to act as a switch for gene expression [ 197 , 200 ]. In this context, deadenylation catalyzed by one of the catalytic subunits of the CCR4-NOT complex, CNOT6/CNOT6L (CCR4-NOT transcription complex subunit 6/subunit 6-like), is considered the rate-limiting step in mRNA decay [ 200 ]. It is therefore thought that the 3'-poly-(A) tail protects the mRNA from decay and improves translation efficiency [ 200 ]. Under hypoxia, the serine/threonine kinase homeodomain-interacting protein kinase 2 (HIPK2) acts as an integrator of stress signals to sense hypoxia and regulate mRNA degradation [ 201 ]. HIPK2 restricts CCR4-NOT mRNA decay activity [ 198 ] (Fig.  3 ). Fig. 3 Deadenylation-dependent mRNA degradation mechanism under hypoxia. Created in BioGDP ( https://biogdp.com ). Under normoxic conditions, mRNAs undergo two possible degradation pathways, one of which involves degradation from the 3'-end to the 5'-end under the action of exosomes. Another common degradation pathway involves removal of the cap structure by Dcp1/DCp2, which is then degraded by 5'−3' exoribonurelease 1 (Xrn1) from the 5'-end to the 3'-end [ 202 ]. Under hypoxia, the serine/threonine kinase HIPK2 acts as an integrator of stress signals to sense hypoxia and regulate mRNA degradation [ 201 ]. After hypoxic stimulation, cells return to normoxic conditions, and the mRNAs containing AREs produced during the hypoxic phase are degraded. AREs are well-represented regulatory motifs that can be recognized by RBPs that control the deadenylation and decay of transcripts [ 203 ]. CCR4-NOT: carbon catabolite metabolite deterrent 4-TATA negative; HIPK: homeodomain-interacting protein kinase; DCP: decapping enzyme; Xrn1: exoribonurelease 1; AREs: AU-rich elements; RBPs: RNA-binding proteins. Deadenylation-dependent mRNA degradation mechanism under hypoxia. Created in BioGDP ( https://biogdp.com ). Under normoxic conditions, mRNAs undergo two possible degradation pathways, one of which involves degradation from the 3'-end to the 5'-end under the action of exosomes. Another common degradation pathway involves removal of the cap structure by Dcp1/DCp2, which is then degraded by 5'−3' exoribonurelease 1 (Xrn1) from the 5'-end to the 3'-end [ 202 ]. Under hypoxia, the serine/threonine kinase HIPK2 acts as an integrator of stress signals to sense hypoxia and regulate mRNA degradation [ 201 ]. After hypoxic stimulation, cells return to normoxic conditions, and the mRNAs containing AREs produced during the hypoxic phase are degraded. AREs are well-represented regulatory motifs that can be recognized by RBPs that control the deadenylation and decay of transcripts [ 203 ]. CCR4-NOT: carbon catabolite metabolite deterrent 4-TATA negative; HIPK: homeodomain-interacting protein kinase; DCP: decapping enzyme; Xrn1: exoribonurelease 1; AREs: AU-rich elements; RBPs: RNA-binding proteins. AU-rich elements (AREs) are well-represented regulatory motifs that can be recognized by RNA-binding proteins (RBPs), which control the deadenylation and decay of transcripts. In AREs containing RNA, ARE-mediated decay is a mechanism that leads to the rapid degradation of mRNA due to the presence of oxygen after cells return to normoxic gene expression [ 203 ]. In addition, other stress-related subcellular organs also regulate mRNA degradation. Throughout their lifecycle, cellular mRNAs are coated with various proteins and exist as messenger ribonucleoprotein (mRNP) complexes. The mRNP composition determines the fate of the mRNA and controls different processes such as transport, degradation and translational regulation. Within cells, mRNP complexes can organize to form microscopically visible cytoplasmic RNA granules. To date, the best characterized RNA granules are stress granules (SGs) and processing bodies (P bodies) [ 204 ]. SGs are membrane-less organelles formed through phase separation. When eukaryotic cells are exposed to environmental stress conditions such as hypoxia, they rapidly form in the cytoplasm [ 204 , 205 ]. Components of SGs include stalled 48S preinitiation complexes consisting of translationally arrested mRNAs, small ribosomal subunits, translation initiation factors (e.g., eIF4E, eIF4G, and eIF4A), and many RNA-binding proteins such as PolyA-binding protein (PABP), Ras-GTPase-activating protein-binding protein 1 and 2 (G3BP1/2), T cell intracellular antigen-1 (TIA1), and human antigen R (HuR) [ 204 , 205 ]. When cells are suddenly exposed to stress stimuli such as hypoxia, these components interact with each other and jointly drive the formation of stress granules with liquid–liquid phase separation characteristics [ 204 , 205 ]. The formation process involves multiple signaling pathways and post-translational modifications of SGs components [ 206 ]. Acute hypoxia can promote the expression of SGs markers such as G3BP1, TIA1, and FUS, thereby facilitating the formation of SGs in the cytoplasm [ 207 , 208 ]. In addition, the lactate modification of high—mobility group box 1 (HMGB1) induced by hypoxia promotes its complex formation with TIAR and subsequent export from the nucleus, playing a positive role in the formation of SGs [ 209 ]. On the other hand, in cancer, G3BP1 may further undergo arginine methylation or deacetylation, which is beneficial for the formation of SGs [ 206 ]. P-bodies are cytoplasmic membraneless organelles that consist of ribonucleoprotein complexes and are formed by phase separation [ 210 ]. In response to environmental stimuli such as hypoxia, the mutual recruitment of translation-arrested mRNAs and specific proteins in the cytoplasm leads to the formation of P-bodies [ 211 ]. Edc3 is a P-body resident protein that plays an important role in the formation and maintenance of P-bodies [ 212 ]. P-bodies can also play a key role in regulating hypoxia-related mRNA degradation by integrating deadenylases and decapping complexes [ 197 , 213 ]. Researchers have also reported that RNA promotes phase separation of glycolytic enzymes into glucose (glycolytic) bodies (G bodies) in hypoxia [ 50 , 214 ]. G bodies formation is a conserved, adaptive response [ 214 ]. Glycolysis enzymes are typically diffuse and soluble in the cytoplasm under normoxic conditions. However, under hypoxic conditions, these enzymes become compartmentalized into cytoplasmic structures [ 215 ]. During hypoxic stress, when oxidative phosphorylation is inhibited, glycolysis enzymes coalesce into membraneless cytoplasmic granules called G bodies, and then enhance the rate of glycolysis [ 214 , 216 ]. Previous studies have shown that hypoxia can affect mRNA degradation through autophagy. Hypoxia downregulates p62 mRNA by inhibiting nuclear factor-E2 related factor 2 (Nrf2), which promotes autophagic degradation [ 217 ] (Fig.  4 ). Fig. 4 Other RNA degradation modes under hypoxia. Created in BioGDP ( https://biogdp.com ). Autophagosomes may promote mRNA degradation, whereas SGs may inhibit mRNA degradation. P bodies may play a pivotal role in hypoxia-associated mRNA decay. However, there are currently no studies reporting whether G bodies mediate mRNA degradation under hypoxia. P-body: processing body; G-body: glucose (glycolytic) body; ATG-9: autophagy-related protein 9; G3BP1: GTPase-activating protein-binding protein 1; GW182: glycine-tryptophan protein of 182 kDa; HK2: hexokinase 2 Other RNA degradation modes under hypoxia. Created in BioGDP ( https://biogdp.com ). Autophagosomes may promote mRNA degradation, whereas SGs may inhibit mRNA degradation. P bodies may play a pivotal role in hypoxia-associated mRNA decay. However, there are currently no studies reporting whether G bodies mediate mRNA degradation under hypoxia. P-body: processing body; G-body: glucose (glycolytic) body; ATG-9: autophagy-related protein 9; G3BP1: GTPase-activating protein-binding protein 1; GW182: glycine-tryptophan protein of 182 kDa; HK2: hexokinase 2

Oxygen is essential for normal life activities in a wide range of organisms, especially mammals. An adequate oxygen supply is essential for the metabolic activities of cells and tissues. Recent studies have shown that oxygen acts as a signaling molecule and that an insufficient oxygen supply alters the transcription and translation of the cellular genome [ 1 , 2 ]. Oxygen is also involved in regulating various molecular processes [ 3 ]. High-altitude environments, tumors, and ischemic diseases are characterized by a lack of oxygen. A low level of oxygen content and pressure in cells is called hypoxia [ 4 , 5 ]. The central aim of the hypoxia response is to restore oxygen homeostasis [ 6 ]. A hypoxic environment significantly activates transcription factors in cells and promotes hypoxic adaptation through the transcription and protein expression of many genes. Among these genes, hypoxia-inducible factor (HIF) is the main transcription factor that regulates the hypoxic response [ 7 ]. The research approach was jointly proposed by Nobel Laureates in Physiology or Medicine—William Kaelin Jr, Sir Peter Ratcliffe, and Gregg Semenza. HIF-1 activates gene expression to restore cellular homeostasis after hypoxia. It induces proinflammatory responses (e.g., NF-κB and TLR), promotes metabolic shifts (e.g., HK1/2 and LDHA), induces angiogenesis (e.g., EPO and VEGF), and promotes apoptosis [ 8 – 10 ]. Additionally, hypoxia is related to energy metabolism [ 11 ]. Hypoxia inhibits protein translation by reducing metabolic rates to maintain intracellular ATP homeostasis [ 12 , 13 ]. During hypoxia, global protein synthesis decreases to conserve ATP. However, the translation of critical hypoxic response factors, such as HIF-1α and VEGF, is maintained via an internal ribosomal entry mechanism [ 14 ]. These findings indicate that upon transcriptional and energy-dependent regulation of gene expression, hypoxia affects the posttranscriptional regulation of mRNAs in other ways [ 15 ]. Studies have demonstrated that the half-life of VEGFA mRNA is extremely short. During hypoxia, aspartyl-tRNA synthetase 1 antisense 1 (DARS-AS1) recruits METTL3/METTL14 to bind to the N 6 -methyladenosine (m 6 A) modification of DARS mRNA and enhance translation, thereby promoting hypoxia-induced cytoprotective autophagy in cervical cancer cells [ 16 ]. These studies indicate that hypoxia regulates the mRNA life cycle through various processes, including synthesis, processing, translation, and degradation. However, there has not yet been a systematic discussion of changes in mRNA homeostasis under hypoxia. To better understand the regulation of mRNA homeostasis under hypoxia, this review summarizes the changes and their functional implications.

In summary, hypoxia influences mRNA metabolism at various levels, including transcription, translation, modification and degradation. This enables cells to adapt to lower oxygen levels. Dysregulation of the fate of mRNAs plays a critical role in the occurrence and development of hypoxia-related diseases. As previously mentioned, the detailed regulation of the fate of mRNAs is gene- and context dependent. A comprehensive understanding of the changes in the molecular network induced by hypoxia will be helpful in exploring potential targets for disease treatment. Hypoxia has a significant effect on the process of gene transcription, including transcription initiation, early elongation and alternative splicing. This is primarily mediated by the regulation of the HIF signaling pathway. Other transcription factors also respond to hypoxic stimuli in a HIF-independent manner. These transcription factors and their regulators provide potential intervention strategies for hypoxia-induced cellular changes. Hypoxia significantly affects the mRNA translation process through a multilevel regulatory mechanism, which includes a synergistic effect on the initiation, extension, and termination stages of translation to maintain cellular energy homeostasis and adapt to hypoxia. These regulatory mechanisms collectively represent the adaptive strategies of cells under hypoxia: by dynamically balancing translation efficiency and energy consumption, they prioritize the synthesis of essential survival proteins while inhibiting the expression of nonessential proteins. However, the molecular mechanisms underlying this process are not fully understood. For example, it is unclear whether cap-dependent and cap-independent translation have different pathways of influence under hypoxic conditions. In addition to its role in gene expression, hypoxia is associated with various mRNA modifications, including methylation (m 6 A, m 1 A, m 5 C, and m 7 G), acetylation, and A-to-I editing. Most studies have shown that the protein molecules involved in hypoxia induction are primarily writers, erasers, and readers related to these modifications. While more than 170 RNA modifications have been identified, the functions and hypoxic effects of these modifications are not fully understood. However, further studies are needed. Hypoxia regulates mRNA stability through structural changes mediated by poly(A) tails and U-rich sequences in the 3' UTR and altering the CDS structure, RBP interactions, and lncRNA-mediated mechanisms [ 179 , 180 ], driving adaptive or pathological outcomes in cancer and inflammation. Targeting these pathways (e.g., YTHDF2 and HIF-1α-stabilizing lncRNAs) has therapeutic potential. Notably, RNA-binding proteins are involved in all the reviews mentioned above. Targeting RBPs has significant implications for the future development of medicine. For instance, the RBP human antigen R (HuR) stabilizes and enhances the translation of pro-survival mRNAs under cancer-related stresses such as hypoxia, radiotherapy, and chemotherapy [ 218 ]. Therefore, developing small molecules that inhibit or alter the localization of HuR could offer a novel approach for treating hypoxia-related diseases and improving therapeutic responses. Similarly, YTHDF1, an m6A-binding protein linked to high-altitude adaptation, is amplified in cancers like non-small cell lung cancer [ 147 ]. Targeted inhibition of YTHDF1 and its paralog YTHDF2 may thus provide a significant prognostic benefit in oncology. Overall, hypoxia can regulate mRNA homeostasis. Understanding the relationship between low-oxygen-related diseases and mRNA homeostasis is important for developing new RNA-targeting therapies. We can diagnose diseases and estimate their stage or extent on the basis of specific gene expression levels. We can treat diseases in the early stages and prevent them from becoming difficult to treat in the middle and late stages by regulating the stability or degradation of certain mRNAs. We can develop preconditioning exercises to increase tolerance to hypoxia by regulating specific mRNA modifications. The use of RNA-targeting therapies, such as mRNA stabilizers, splicing regulators, and HIF inhibitors/activators, has expanded the scope and strategies of drug research. Future research combining new molecular methods, such as single-cell sequencing and spatial transcriptomics, can further elucidate the spatiotemporal specificity of mRNA dynamic changes in the hypoxic microenvironment to optimize precise treatment strategies.