Spatiotemporal regulation of microRNA-mediated gene silencing in cells
2026 年 51 巻 2 号 p. 277-288
詳細
2026 年 51 巻 2 号 p. 277-288
MicroRNAs (miRNAs) are small non-coding RNAs that regulate gene expression by silencing mRNAs with complementary sequences. Over the past few decades, genetic and biochemical studies have revealed the molecular machinery and mechanisms underlying miRNA function. However, the current understanding of their cellular biology remains relatively limited. While the conventional view simplifies the miRNA journey as transcription in the nucleus and function in the cytoplasm, the precise spatiotemporal regulation of miRNAs within the intracellular environment is only beginning to emerge. Accumulating evidence suggests that miRNAs do not merely float in the cytosol, but instead associate with various organelles and biomolecular condensates, potential sites of miRNA regulation. In this review, we summarize the current knowledge regarding how these intracellular compartments relate to miRNA function. Furthermore, we introduce cutting-edge methodologies designed to explore this expanding frontier of miRNA cell biology.
Key words: microRNA, mRNA, translation, organelles, single-molecule imaging
Graphical Abstract
MicroRNAs (miRNAs) are a class of small non-coding RNAs, typically ~22 nt in length, that regulate gene expression by silencing mRNAs with complementary sequences (Bartel, 2018). The field was initiated in 1993 by the discovery of the first miRNA, lin-4, in C. elegans, which down-regulates lin-14 mRNAs via sequence complementarity (Lee et al., 1993; Wightman et al., 1993). In 2000, the identification of the second miRNA, let-7, demonstrated that miRNAs are broadly conserved across diverse species, including humans (Reinhart et al., 2000; Pasquinelli et al., 2000). This triggered a massive expansion in miRNA research, leading to the rapid cloning and annotation of thousands of miRNAs across various organisms (Lagos-Quintana et al., 2001; Lau et al., 2001; Lee and Ambros, 2001). Currently, the human genome is estimated to include more than 500 miRNA genes (Kozomara et al., 2019; Clarke et al., 2025), which are predicted to regulate more than 60% of human mRNAs (Friedman et al., 2009). Accordingly, miRNAs regulate diverse physiological processes and have been implicated in a wide range of human diseases (Fujimoto et al., 2012; Shenoy and Blelloch, 2014; Rupaimoole and Slack, 2017).
Given their profound significance, miRNAs have been a subject of intensive research over the past few decades. However, the vast majority of these studies have relied on genetic and biochemical approaches. Consequently, while the molecular machinery and mechanisms of miRNA-mediated gene silencing are well understood, their cell biological aspects, such as when and where they function within the intracellular environment, remain poorly understood. To bridge this gap, this review, after providing a brief overview of how miRNAs are generated and function, focuses on the current understanding of the interplay between miRNAs and intracellular compartments, such as P-bodies, endosomes, and the endoplasmic reticulum (ER). Here, we focus primarily on miRNAs in animals. Furthermore, we introduce emerging methodologies designed to explore this expanding frontier of the cell biology of miRNAs.
miRNAs are encoded in the genome (Fig. 1; Kim et al., 2009, 2025). Most miRNAs are intragenic, residing more frequently within protein-coding genes and, to a lesser extent, within non-coding genes, and are predominantly located in introns rather than exons (Rodriguez et al., 2004; Kim and Kim, 2007; Marsico et al., 2013; França et al., 2016). A smaller fraction of miRNAs is intergenic, located between annotated genes. Regardless of their genomic origin, the majority of miRNA genes are transcribed by RNA polymerase II into long primary miRNAs (pri-miRNAs) (Lee et al., 2002, 2004; Cai et al., 2004), which contain characteristic stem-loop structures. The Microprocessor complex, consisting of the RNase III enzyme Drosha and its cofactor DGCR8 (Pasha in flies), recognizes and cleaves the stem-loops of pri-miRNAs to generate ~65-nt short hairpin RNAs called precursor miRNAs (pre-miRNAs) (Lee et al., 2003; Denli et al., 2004; Gregory et al., 2004; Han et al., 2004; Landthaler et al., 2004; Partin et al., 2020; Jin et al., 2020; Garg et al., 2024). Subsequently, pre-miRNAs are exported from the nucleus to the cytoplasm by Exportin-5 and Ran GTPase (Yi et al., 2003; Lund et al., 2004; Bohnsack et al., 2004; Okada et al., 2009). In the cytoplasm, the loop regions of the pre-miRNAs are removed by another RNase III enzyme Dicer, in association with the double-stranded RNA-binding proteins TRBP or PACT (Loquacious in flies) (Bernstein et al., 2001; Grishok et al., 2001; Hutvágner et al., 2001; Ketting et al., 2001; Chendrimada et al., 2005; Haase et al., 2005; Lee et al., 2006; Liu et al., 2018), yielding ~22-nt mature miRNA duplexes.
Biogenesis and function of miRNAs
miRNA genes are transcribed by RNA polymerase II into pri-miRNAs, which are processed by the Microprocessor complex containing Drosha and DGCR8 to generate pre-miRNAs. Exportin-5 and Ran GTPase transport pre-miRNAs to the cytoplasm, where Dicer, together with TRBP or PACT, generates mature miRNA duplexes. The Hsp70/Hsp90 chaperone machinery facilitates duplex loading into AGO, with the guide strand (magenta) retained and the passenger strand (green) discarded to form RISC. RISC binds to target mRNAs through base pairing and mediates translational repression followed by mRNA decay. The scaffold protein GW182 recruits downstream silencing effectors, including the CCR4–NOT deadenylase complex and DDX6, to mediate these processes. Deadenylated mRNAs are decapped by DCP1/DCP2 and degraded from the 5' end by XRN1.
miRNAs cannot function alone. To execute gene silencing, the miRNA duplexes must be loaded into Argonaute (AGO) proteins to form the RNA-induced silencing complex (RISC) (Fig. 1; Kobayashi and Tomari, 2016; Iwakawa and Tomari, 2022; Nakanishi, 2022). This loading process is not a spontaneous binding but is actively driven by the Hsp70/Hsp90 chaperone machinery in an ATP-dependent manner, transiently opening the conformation of AGO to accommodate the bulky miRNA duplexes (Yoda et al., 2010; Iki et al., 2010; Iwasaki et al., 2010; Miyoshi et al., 2010; Iwasaki et al., 2015; Tsuboyama et al., 2018; Lee et al., 2026). Following the loading, one strand of the duplex (the guide strand) is preferentially selected and retained within AGO, while the other (the passenger strand) is ejected and degraded (Schwarz et al., 2003; Khvorova et al., 2003), completing RISC assembly. Here, we focus on its established roles in the cytoplasm, although recent studies have also implicated miRNAs in nuclear functions (Hu et al., 2023).
Once assembled, guided by the single-stranded miRNA, RISC binds to target mRNAs through complementary base pairing (Bartel, 2018). In general, RISC targets the 3' UTR, where it is less likely to encounter translating ribosomes (Gu et al., 2009). Upon binding to the target mRNA, RISC silences gene expression through translational repression followed by mRNA decay (Bazzini et al., 2012; Djuranovic et al., 2012; Béthune et al., 2012). To orchestrate these events, AGO cooperates with the scaffold protein GW182 (TNRC6 in humans) (Rehwinkel et al., 2005; Meister et al., 2005; Liu et al., 2005a; Jakymiw et al., 2005), which recruits downstream silencing effectors such as the CCR4–NOT deadenylase complex (Behm-Ansmant et al., 2006; Braun et al., 2011; Chekulaeva et al., 2011) and DDX6 (Dhh1 in yeast) (Chu and Rana, 2006; Eulalio et al., 2007a; Chen et al., 2014; Mathys et al., 2014; Rouya et al., 2014). Once deadenylated, mRNAs undergo decapping by DCP1 and DCP2 enzymes (Rehwinkel et al., 2005; Behm-Ansmant et al., 2006), followed by mRNA decay from the 5' end by XRN1 (Braun et al., 2012).
As outlined above, while the molecular machinery governing miRNA function is well characterized, its spatiotemporal regulation across the complex intracellular compartments remains a compelling frontier. In the following sections, we summarize the current knowledge of the interplay between miRNAs and intracellular compartments.
In 2005, using fluorescence microscopy, multiple groups reported the localization of miRNAs, AGO, and target mRNAs to P-bodies, a class of phase-separated condensates (Sen and Blau, 2005; Liu et al., 2005b; Pillai et al., 2005), which are involved in mRNA regulation (Sheth and Parker, 2003; Hubstenberger et al., 2017). Because P-bodies also enrich factors required for miRNA-mediated gene silencing, such as DCP1, DCP2, and DDX6, these condensates initially attracted significant attention as the potential site where miRNAs work (Jakymiw et al., 2007). However, this model was soon challenged by two independent groups, which showed that miRNAs normally silence their targets even when P-body formation is disrupted (Chu and Rana, 2006; Eulalio et al., 2007b). Consistent with this, although the punctate localization of AGO to P-bodies is visually striking under the microscope, quantitative measurements have shown that only ~1.3% of cytoplasmic AGO is localized to P-bodies (Leung et al., 2006). Together, these findings led to the view that P-body formation is not the cause, but a consequence, of miRNA-mediated gene silencing. Over a decade later, however, the relationship between miRNAs and P-bodies regained attention through a study demonstrating that phase separation driven by AGO and GW182 accelerates the deadenylation of target mRNAs in vitro (Sheu-Gruttadauria and MacRae, 2018). Whether this phase separation truly contributes to miRNA function under physiological conditions in cells remains an important question for future investigation. Notably, while mRNAs targeted by miRNAs are sequestered into P-bodies, recent transcriptome-wide analysis showed little correlation between P-body enrichment and poly(A) tail length (Pessina et al., 2026), suggesting that P-body localization is not necessarily associated with preferential deadenylation.
When subjected to environmental stresses, such as heat shock, oxidative stress, or viral infection, cells dynamically assemble another major class of phase-separated condensates, stress granules, where a subset of mRNAs and translational machinery are sequestered (Kedersha et al., 2005; Khong et al., 2017; Riggs et al., 2020). Upon stress induction, miRNAs and AGO rapidly localize to stress granules (Leung et al., 2006; Pare et al., 2009). Intriguingly, however, their silencing partner, GW182, largely remains within P-bodies (Kedersha et al., 2005; Majerciak et al., 2023). This spatial separation of AGO from GW182 might constitute a molecular switch to temporarily pause or modulate silencing activity in response to environmental cues, given that cellular stress has been shown to relieve miRNA-mediated gene silencing (Bhattacharyya et al., 2006).
Historically, mammalian AGO was identified as a membrane-associated protein (Cikaluk et al., 1999). Following this observation, two independent groups suggested in 2009 that membranous organelles are linked to the regulation of miRNA function (Gibbings et al., 2009; Lee et al., 2009). In the endocytic pathway, internalized molecules progress from early endosomes to late endosomes/multivesicular bodies (MVBs) before being delivered to lysosomes (Fig. 2). Both groups proposed that MVBs may serve as sites of miRNA-mediated gene silencing. Using density gradient fractionation of human monocytes, Gibbings et al. reported that AGO and GW182 showed partial overlap with the MVB fractionation profile. Interpreting N-rhodamine-labeled phosphatidylethanolamine as an MVB marker, the authors also reported its colocalization with AGO and GW182. Based on partial colocalization between GW182 and Lamp1 in fly cells, Lee et al. similarly suggested that GW182 associates with MVBs. Notably, both studies reported that depletion of ESCRT components, which are required for MVB formation, compromised miRNA-mediated gene silencing. Even though the ESCRT machinery participates in a diverse array of cellular processes beyond MVB biogenesis (Hurley, 2015), these studies raised the possibility that MVBs may influence miRNA-mediated gene silencing, either directly as sites of miRNA regulation or indirectly through their broader cellular function. Because the depletion of ESCRT components significantly increased GW182 protein levels (Gibbings et al., 2009), one possibility is that this accumulation disrupts the stoichiometry of functional RISC complexes.
Intracellular localization of miRNAs, AGO, and GW182
miRNAs and their machinery, including AGO and GW182, have been reported to associate with diverse membranous and membraneless compartments, based on microscopy-based colocalization and fractionation-based co-sedimentation analyses. However, the reported localization patterns often differ across studies, highlighting the limitations of conventional analyses for exploring when and where miRNAs function.
From early endosomes, instead of proceeding to late endosomes/MVBs and lysosomes, certain cargo is sorted into recycling endosomes to be returned to the plasma membrane (Fig. 2). In neurons, AGO has been shown to partially colocalize with recycling endosomes (Antoniou et al., 2014). This colocalization is decreased by depletion of the BAR-domain protein, PICK1, which links AGO to recycling endosomes. Because the depletion of PICK1 slightly enhances miRNA-mediated gene silencing in neurons, this interplay between AGO and recycling endosomes may negatively regulate miRNA function.
Interestingly, in neurons, fluorescently labeled pre-miR-181a can be transported to distal axons, in association with late endosomes or lysosomes, where it is processed into mature miRNAs in response to the axon guidance cue. The newly generated miR-181a then represses translation of TUBB3 mRNA (Corradi et al., 2020), highlighting local mRNA regulation by spatially restricted miRNAs. How broadly such a mechanism operates beyond neurons is an intriguing question for future investigation.
AGO loads not only miRNAs but also small interfering RNAs (siRNAs) to trigger RNA interference (RNAi), in which target mRNAs are cleaved (Kobayashi and Tomari, 2016). In 2013, the ER emerged as a plausible platform for RNAi mediated by siRNAs (Stalder et al., 2013). In human cells, colocalization analysis with multiple organelle markers showed that endogenous AGO partially colocalizes not only with P-bodies but also with the ER, ER-Golgi intermediate compartment, and Golgi. In addition to conventional density gradient fractionation, this study utilized membrane floatation analysis and membrane immunoprecipitation analysis to clearly demonstrate that AGO associates with the cytosolic surface of ER membranes. By introducing siRNA duplexes into cells by lipofection, Stalder et al. showed that immunoprecipitation of AGO from ER fractions co-immunoprecipitated both guide and passenger strands, suggesting that the ER acts as a site for RISC assembly at least for siRNAs. Moreover, siRNA-cleaved mRNAs were detected in the same fractions, suggesting that target binding and cleavage by siRNAs also occur at the ER. Because depletion of TRBP or PACT, which facilitated AGO targeting to the ER, slightly delayed target cleavage by siRNAs, ER localization of AGO may enhance RNAi efficiency. This naturally raises the question: Does the ER play a role in miRNA function as well? A following study by another group showed that reporter mRNAs with miRNA sites were detected in ER fractions in an AGO-dependent manner (Barman and Bhattacharyya, 2015), yet it remains unclear whether the following steps, such as translational repression, occur on the ER. Although the mechanisms of miRNA-mediated gene silencing differ considerably between animals and plants, the ER has been suggested to contribute to miRNA function in plants as well (Li et al., 2013). Together, these findings suggest that the ER likely functions as a platform for key steps in miRNA action, particularly for RISC assembly and target mRNA binding.
Mitochondria are the major source of ATP, which is required for RISC assembly, intrinsically linking them to miRNA-mediated gene silencing. Indeed, inhibition of mitochondrial function by CCCP has been reported to impair miRNA function (Huang et al., 2011). Overexpression of Ucp2 and depletion of Mfn2, both of which perturb mitochondrial function, also reduced the level of AGO-associated miRNAs (Bose et al., 2020). Beyond this, miRNAs have also been reported to localize to mitochondria (Kren et al., 2009). Density gradient fractionation showed that a purified mitochondrial fraction from rat liver contains several miRNAs. Notably, miR-21, which is highly abundant in the liver, was not detected in this fraction, suggesting that certain miRNAs are preferentially localized to mitochondria, rather than simply reflecting their cellular abundance. Another group reported similar results using mouse liver and showed that a purified mitochondrial fraction also contains AGO (Bian et al., 2010), although the detection of mitochondrial AGO may depend on the experimental approach (Ro et al., 2013). A subsequent study using human cells also detected miRNAs and AGO in a mitochondrial fraction and observed partial colocalization of AGO with mitochondrial markers (Bandiera et al., 2011). Notably, in all three studies, mitochondrial fractions were treated with RNase, suggesting that the detected miRNAs were localized within, rather than on the surface of, mitochondria. What is the role of miRNAs inside mitochondria? Unlike other organelles, mitochondria possess a unique genetic system, including their own genome and transcriptional and translational machinery (Rackham and Filipovska, 2022). COX1 mRNA, which is transcribed and translated within mitochondria (Richter-Dennerlein et al., 2026), is targeted by miR-181c (Das et al., 2012). Overexpression of miR-181c decreased the protein level of COX1 without affecting its mRNA level, suggesting translational repression. This was accompanied by increased ROS generation, indicative of mitochondrial dysfunction. Supporting a function of AGO in mitochondria, a recent study demonstrated that exogenously transfected siRNAs can trigger RNAi of mitochondrial mRNAs (Gao et al., 2021).
Unexpectedly, during myogenic differentiation, overexpression of miR-1, which also targets COX1 mRNA, was reported to increase COX1 protein levels without affecting its mRNA abundance, suggesting translational activation rather than repression (Zhang et al., 2014). Although miRNAs generally repress translation, a few studies have reported their ability to activate translation under specific conditions (Vasudevan et al., 2007; Iwasaki and Tomari, 2009). How miRNAs determine whether to repress or activate translation within mitochondria, and the molecular mechanisms underlying these opposing outcomes, remain intriguing questions for future investigation.
Interestingly, miRNA-free empty AGO is subject to degradation across species. Inhibition of Hsp90, which is required for miRNA loading into AGO, was reported to reduce AGO protein levels in human cells (Suzuki et al., 2009; Johnston et al., 2010). Inhibition of miRNA biogenesis was also shown to reduce AGO levels in mice and flies, as well as in plants (Derrien et al., 2012; Martinez and Gregory, 2013; Smibert et al., 2013). In particular, Smibert et al. showed that this reduction is due to increased degradation rather than reduced translation, using a translation inhibitor. How does empty AGO, but not miRNA-loaded AGO, preferentially undergo degradation, and what is the biological rationale behind this? In flies, empty AGO was shown to be ubiquitinated by the E3 ubiquitin ligase Iruka (Kobayashi et al., 2019a). Notably, depletion of Iruka, which compromised elimination of empty AGO, resulted in accumulation of dysfunctional AGO, thereby impairing successful miRNA-mediated gene silencing. This suggests that the degradation of empty AGO is a quality control process to ensure miRNA function—the sustained empty state appears to serve as a surveillance signal of dysfunctional AGO.
While protein degradation is mediated by two major systems, the proteasome and autophagy, accumulating evidence supports that the degradation of empty AGO is mediated by autophagy (Derrien et al., 2012; Martinez and Gregory, 2013; Kobayashi et al., 2019b; Kobayashi and Tomari, 2020). Whereas the proteasome degrades relatively small proteins, autophagy, in which autophagosomes dynamically engulf cytoplasmic components for delivery to lysosomes, can handle large clients such as protein aggregates (Dikic, 2017). Given that AGO forms high-molecular-weight complexes (Olejniczak et al., 2013; La Rocca et al., 2015) and is aggregation-prone (Sen and Blau, 2005; Liu et al., 2005b; Pillai et al., 2005; Jakymiw et al., 2005), autophagy may be more suitable for the degradation of AGO. Beyond empty AGO, the steady-state abundance of AGO, as well as GW182, is also regulated by autophagy (Gibbings et al., 2012; Zhang and Zhang, 2013; Pircs et al., 2018). Interestingly, Ghosh et al. recently showed that autophagy can degrade AGO together with its associated target mRNAs upon leucine deprivation in flies (Ghosh et al., 2024).
Notably, a recent study proposed that the cytosolic surface of autophagosomes could provide a platform for AGO-containing complex formation (Santovito et al., 2020). Thus, the interplay between the miRNA machinery and autophagy might extend beyond simple degradation, with autophagosomes potentially serving as platforms for miRNA machinery, a possibility that warrants further investigation.
As discussed above, previous studies have mapped the intracellular localization of miRNA machinery (Fig. 2), often based on partial colocalization and co-sedimentation with specific compartments, proposing these compartments as potential platforms for miRNA-mediated gene silencing. However, the reported localization of miRNA machinery has varied considerably across studies, with some observations appearing to contradict one another (Cikaluk et al., 1999; Eystathioy et al., 2002; Gibbings et al., 2009; Stalder et al., 2013). As such, definitive evidence regarding when and where miRNAs actually work remains elusive. A fundamental limitation underlying this issue is that conventional microscopy-based colocalization or fractionation-based co-sedimentation cannot distinguish whether the miRNA machinery is merely “present” or actively “functioning” there. To overcome this long-standing hurdle, imaging-based methods for monitoring miRNA function in cells have recently emerged.
In 2022, a three-color single-molecule method was developed to image distinct steps in miRNA-mediated gene silencing: RISC binding to mRNAs, translational repression, and mRNA decay in fixed cells (Fig. 3; Kobayashi and Singer, 2022). In this system, a reporter mRNA harboring the SunTag-coding sequence in the ORF and miRNA-binding sites in the 3' UTR is expressed in cells and subsequently analyzed by single-molecule imaging (Kobayashi and Singer, 2025). Individual reporter mRNAs are imaged by single-molecule fluorescence in situ hybridization (smFISH), in which multiple fluorescent probes hybridize to the target mRNA (Femino et al., 1998; Raj et al., 2008). Translation is monitored by the fluorescence intensity of SunTag on individual mRNAs—SunTag, consisting of 24 repeats of an antibody epitope (Tanenbaum et al., 2014), enables detection of nascent peptides on translating mRNAs by immunofluorescence (Wu et al., 2016). RISC binding is monitored by the fluorescence intensity of AGO on individual mRNAs—8 repeats of miRNA-binding sites enable robust detection of endogenous AGO on mRNAs via immunofluorescence. This system revealed that RISC bound to target mRNAs immediately after nuclear export, preferentially selecting translating mRNAs rather than untranslated ones. Subsequently, RISC repressed translation within ~30 min of binding, followed by mRNA decay by ~60 min.
Methods to image miRNA-mediated gene silencing in cells
(A) A three-color single-molecule method for imaging distinct steps of miRNA-mediated gene silencing in fixed cells, including RISC binding to mRNAs, translational repression, and mRNA decay (Kobayashi and Singer, 2022). The reporter mRNA carrying a SunTag sequence in the ORF and miRNA-binding sites in the 3' UTR is detected by smFISH, while translation and RISC binding are monitored by immunofluorescence against SunTag and AGO, respectively. This method enables imaging of miRNA-mediated gene silencing by endogenous miRNAs and AGO, but is applicable only to fixed cells. (B) A three-color single-molecule method for imaging AGO-mediated translational repression in living cells (Cialek et al., 2022). The reporter mRNA contains MS2 stem-loops for mRNA detection, FLAG tag repeats in the ORF for translation imaging, and BoxB stem-loops in the 3' UTR to tether fluorescently labeled AGO. This method enables live-cell imaging of AGO-mediated translational repression, while the tethering approach bypasses the endogenous dynamics of AGO–mRNA interactions.
Cialek et al. developed a conceptually similar, yet complementary, single-molecule method in living cells (Fig. 3; Cialek et al., 2022). The reporter mRNA contains 24 repeats of MS2 stem-loop, which recruit fluorescently labeled MS2 coat proteins (MCP), enabling single-molecule detection. The ORF of the reporter harbors 10 repeats of FLAG tag to image translation through fluorescently labeled anti-FLAG Fab fragments. The 3' UTR of the reporter carries 15 repeats of BoxB stem-loop to recruit λN-fused fluorescently labeled AGO. Although artificial tethering bypasses the endogenous dynamics of AGO–mRNA interactions, including their strength, timing, and intracellular location, this approach enables live-cell imaging of AGO-mediated translational repression, providing a clear advantage over the fixed-cell approach. Notably, this advantage revealed that mRNAs with low levels of AGO tethering occasionally restart translation, suggesting that AGO-mediated translational repression is reversible. Moreover, by following mRNAs over longer timescales, 8–12 h, Cialek et al. observed that translationally repressed AGO-tethered mRNAs accumulated in P-bodies.
Over the past few decades, our view of miRNA-mediated gene silencing has shifted from a simple cytoplasmic event to a spatiotemporally regulated process involving various intracellular compartments (Kim et al., 2014; Leung, 2015). However, conventional localization analyses have made it difficult to identify the functional sites of miRNA-mediated gene silencing, limiting our understanding of miRNA cell biology. The emergence of these single-molecule methods, which allow us to image miRNA-mediated gene silencing inside cells, will help directly answer when and where miRNAs work, and which intracellular compartments actually serve as sites of miRNA regulation.
This work was supported by JST PRESTO Grant Number JPMJPR20E7, JST FOREST Program Grant Number JPMJFR230U, AMED Grant Number JP223fa627001 (UTOPIA Young Researcher Development Program), JSPS KAKENHI Grant Number 23H04251, 25K02208, 25H01439/25K24896, 25H01438/25K24895, The Nakajima Foundation, The Uehara Memorial Foundation, The Sumitomo Foundation, the Daiichi Sankyo Foundation of Life Science, the Kato Memorial Bioscience Foundation, and the Medical Research Center Initiative for High Depth Omics (to HK).
Conflict of Interest StatementAll authors declare no conflicts of interest.
Data Availability StatementNo new data were created or analyzed in this review article. All data discussed are available from the published sources cited in the reference list.
Author Contribution StatementAS, HI, HA, and HK wrote the manuscript. All authors have read and agreed to the final version of the manuscript.
Ethics Approval and Consent to ParticipateNot applicable.
Patient Consent for PublicationNot applicable.
We are grateful to all members of the Hotaka Kobayashi laboratory for insightful discussions and constructive comments.
