key: cord-0259531-m9bp6bjg
authors: Nalawade, Rohit; Chowdhury, Tamoghna; Chatterjee, Saibal
title: XRN2-mediated regulation of let-7 microRNAs is of critical pathophysiological significance in Humans
date: 2021-09-22
journal: bioRxiv
DOI: 10.1101/2021.08.20.457117
sha: 3b6e1029534bfe059b5d73a78ee12768186d3045
doc_id: 259531
cord_uid: m9bp6bjg

MicroRNAs are critical regulators of diverse developmental and physiological processes, and they themselves get regulated both at the level of biogenesis and turnover. We demonstrate that the ribonuclease XRN2 can degrade the mature forms of certain let-7 family members in multiple human cancer cell lines, without affecting their precursors. XRN2 depletion results in a reduction in the expression of a number of oncogenes, and diminishes the proliferative and metastatic potential of cancer cells. The clinical relevance of these observations is also verified in tumour transcriptomics data from public RNA-sequencing datasets, where XRN2 mRNA expression is inversely correlated with the levels of a large number of miRNAs, including let-7 members, and high XRN2 mRNA levels are associated with poor survival in hepatocellular carcinoma, lung adenocarcinoma, and glioblastoma. We demonstrate that the miRNA is released by an as-yet unidentified proteinaceous ‘miRNA release factor’ from the grasp of Argonaute before its degradation. Our analyses of the patient-derived transcriptomics data also show that XRN2, via its regulation of let-7, affects multiple pathways in a consistent manner across epithelial and glial cell lineages, and thus, is of critical pathophysiological significance.

MicroRNAs (miRNAs) function as critical regulators of diverse developmental and physiological processes in animals 1 . Extensive studies on miRNA abundance and function have demonstrated a consistent link between dysregulation of miRNAs and various diseases, highlighting the importance of robust regulation of miRNA activity 2 . Many miRNAs exhibit tissue-and/or-stage-specific expression patterns 3 and dramatic changes in the abundance of some miRNAs during different developmental stages of animals have also been reported 4 . Global kinetic studies of miRNA metabolism have revealed that regardless of high biogenesis, several miRNAs exhibit reasonably low steady-state levels 5, 6 . It indicated that additional determinants, other than the factors of biosynthesis, play critical roles for the establishment of cellular miRNA homeostasis 5, 6 .

A substantial amount of knowledge has been acquired on the different steps of miRNA biogenesis and their regulation, but much less is known about the miRNA turnover pathways and the constituent molecular machineries 7 . Active turnover of miRNAs in animals was first reported in Caenorhabditis elegans (C. elegans), where 5'-3' exoribonuclease XRN-2 was demonstrated to mediate the degradation of several miRNAs 8 . Later, the paralogous protein, XRN-1, was also described as a 'miRNase' in worms 9, 10 . In mammalian cells, a number of nucleases have been described to act on a given miRNA or a handful of miRNAs. XRN1 has been implicated in the turnover of miR-382 in HEK293 cells 11 , and polynucleotide phosphorylase degrades miR-221, miR-222, miR-106b in human melanoma cells 12 . Tudor-SN was 3 demonstrated to target several functional mature miRNAs, including miR-31 and miR-29b, which control specific mRNAs encoding cell cycle regulatory proteins 13 . Interestingly, Argonaute1 (Ago1)-bound miRNAs in Drosophila are known to be tailed by terminal nucleotidyl transferase 14 and trimmed by the 3' -5' exoribonuclease Nibbler 15 upon binding an exogenously introduced target that harbours extensively complementary binding-sites for the given miRNA 14, 15 . Notably, in mice, an unusual extensively complementary endogenous target has been reported that lead to the 3'-end trimming-mediated destabilization of the cognate miRNA upon their interaction 16 . The terminal uridyl transferases TUT4 and TUT7 are known to participate in tailing of miRNAs across species from C. elegans to humans 17 . A recent work using HEK293T cells has identified a machinery consisting of TUTs and DIS3L2, which could execute the decay of a subset of Argonaute (AGO)-bound mature miRNAs that have exposed 3′-ends 18 .

Very recently, two simultaneous reports have substantially accentuated our understanding on the mechanisms of target-directed miRNA degradation (TDMD), where it was described that pairing of RISC/AGO-loaded miRNA to certain unusual highly complementary targets triggers ZSWIM8 ubiquitin ligase-mediated degradation of AGO, rendering the miRNA susceptible to decay by unknown nuclease(s) 19, 20 . Perturbation of ZSWIM8 led to the accumulation of several miRNAs in different mammalian cells and some other systems (worms and flies). Intriguingly, half-lives of AGO proteins are much longer than majority of the miRNAs 5, 6 , which suggested that ZSWIM8-mediated TDMD may not be the predominant miRNA turnover mechanism, rather, it is dedicated towards miRNAs that interact with unusual highly complementary targets. Thus, it is quite apparent that identification of yet unknown 'miRNases', followed by understanding of the pathways in which they are constituents, and their mechanistic details warrant further investigation.

Notably, although, knowledge about miRNA metabolism emanated from the studies employing worms have been found to be largely conserved, the role of the human ortholog of the worm 'miRNase'-XRN-2 remains unclear. Apart from the critical roles played by XRN2 in multiple RNA-transaction pathways 21 , recent studies have also shown that it can act in an oncogenic capacity and worsen prognoses of specific cancers 22, 23 . Here, we explore the capacity of human XRN2 as a 'miRNase', and elucidate the molecular mechanism(s) underlying the oncogenic capacity of XRN2 in relation to its 'miRNase' activity. In this regard, we focused on the fundamentally important let-7 family of miRNAs that are known to regulate important target mRNAs in multiple tissues, whose perturbation leads to different disease states 24 . We demonstrate that XRN2 specifically acts on the mature forms of most of the let-7 family members. Rapid and efficient depletion of XRN2 leads to the accumulation of AGO-bound let-7 members that in turn downregulate their cognate targets, including proto-oncogenes. Depletion of XRN2 in different cancer cells affect their cellular physiology by reversing critical cancer-parameters, including epithelial-tomesenchymal transition. Our ex vivo biochemical assays indicate that XRN2-mediated degradation of 4 mature let-7 miRNAs happens upon 'release' of miRNAs from AGO by a proteinaceous factor without affecting AGO integrity, and these two steps are kinetically linked. We also demonstrate that our cell linebased observations are of critical significance as they correlate with the clinical data of cancer patients from public transcriptomics datasets. Here, our analyses partially explain the multifarious roles of XRN2 in cancer through its likely participation in miRNA turnover, with the observation that XRN2 influences multiple pathways related to cellular proliferation, development, and signalling, especially related to extracellular matrix (ECM) development, and that these influences are mostly consistent across both epithelial and glial lineages. These analyses also partially explain previous observations of elevated XRN2 mRNA expression being associated with worse survival in lung cancer patients, and extend such observations into the more expansive Cancer Genome Atlas (TCGA) datasets for lung adenocarcinoma, hepatocellular carcinoma, and glioblastoma. Our analyses of these TCGA datasets also suggest that for many of the let-7 family members, XRN2 demonstrates a strong inverse relationship comparable to some of the recently identified factors implicated in miRNA turnover, at an RNA level. Collectively, our study reveals that human let-7 miRNAs are regulated by a two-step turnover pathway, wherein, XRN2 plays the role of a 'miRNase' in various tissues.

In human HEK293 cells, XRN1 has been implicated in the decay of miR-382, whereas depletion of the paralogous XRN2 had no significant effect 11 . The authors performed shRNA-mediated knockdown of XRN1 and XRN2, but did not demonstrate the knockdown efficiencies at the protein level. Additionally, substantial knockdown of XRN2 was not achieved in other related studies 11, 25 , which could be insufficient to exert an effect on all the pathways XRN2 is involved in. Therefore, prior to investigating the role of XRN2 in miRNA degradation in human cells, we optimized a rapid and efficient shRNA-mediated specific knockdown of XRN2 using a lentivirus-based approach (Supplementary Text 1, Figure S1 A-D).

By means of turnover, XRN-2 in C. elegans modulates the mature forms of different members of the let-7 family, which play important roles in development. Dysregulation of their homologues also perturbs development and leads to many disease states in vertebrates 24 . Therefore, we began by examining the level of human let-7a miRNA, an identical homolog of C. elegans let-7, in control and XRN2 knockdown cells.

In HEK293T cells, where the levels of endogenous let-7 miRNAs are very low 26 , XRN2 depletion did not affect the level of let-7a ( Figure 1A) . A possible reason for this could be that XRN2-mediated turnover might not be a predominant factor in maintaining its abundance in HEK293T cells. Contrastingly, ~3-fold accumulation in the level of mature let-7a was observed in XRN2-depleted A549 cells ( Figure 1A ).

In animals, let-7 miRNAs are known to exhibit different levels of abundance in different tissues, which could be determined by different regulatory forces 24 . Thus, turnover mediated regulation of specific miRNA levels might be more important in certain tissues, whereas regulated biogenesis might be more critical in others. Accordingly, we selected additional cell lines from different cell lineages, which are known to express adequately detectable levels of let-7 family members 27 . We incorporated Glioblastoma cell lines LN229 and U87-MG, Hepatocellular carcinoma cell lines HUH7 and HEPG2, Breast cancer cell lines MDA-MB-231 and MCF-7, Lung cancer cell line A549 and Cervical cancer cell line HeLa in our study.

XRN2-depleted lung (A549), hepatic (HUH7, HePG2), and neuronal (LN229, U87-MG) cells showed increase in the levels of their mature let-7a ( Figure 1A , S2A). Conversely, let-7a miRNA did not accumulate upon XRN2 depletion in any of the breast or cervical cancer cell lines we examined ( Figure S2A ). This indicated that XRN2 might be involved in the determination of the levels of mature let-7 miRNAs in multiple, but not all human cancer cell lines.

Furthermore, we checked the effect of XRN2 depletion on other members of the let-7 family. TaqMan analyses revealed that XRN2 depletion affected many members, but not all. As demonstrated through 6 Northern probing, TaqMan analysis also showed 2-3 folds accumulation of mature let-7a, while other members of let-7 family including let-7c, -7e, -7g, and miR-98 were also upregulated to different extents upon XRN2 depletion in A549, HUH7 and U87-MG cells ( Figure 1B) . However, let-7b, -7d, -7f, and -7i did not show any significant upregulation in the levels of their mature forms. These data suggested that XRN2 can specifically affect certain let-7 family members in a cell-or tissue-specific manner.

Additionally, we reckoned that being a ribonuclease, XRN2 could affect miRNA levels either by degradation of the mature forms or by regulating the processing of the corresponding precursor molecules during biogenesis. To answer this query, we determined the levels of the precursor RNAs of let-7a and let-7g, whose mature forms are most accumulated upon XRN2 knockdown. RT-qPCR analyses did not show any significant upregulation in the levels of the pri-let-7a and pri-let-7g ( Figure 1C , S2B). The relative levels of pre-let-7a and pre-let-7g remained unchanged in the XRN2-depleted samples, in comparison to the control samples ( Figure 1D , S2C). Accumulation of mature let-7a and let-7g without any appreciable change in the levels of any of their precursor forms indicated that XRN2 acts specifically on mature let-7 miRNAs leading to their degradation in A549, HUH7, and U87-MG cells.

Some of the let-7 family members in humans are known to act as tumour suppressors, and they get dysregulated in different cancers 24, 28 . The tumour suppression ability of let-7 miRNAs is attributed to the fact that they target several important genes including RAS, HMGA2, MYC, CYCD1 (Cyclin D1), and CDC25A 24 , which are potent oncogenes in humans. We anticipated that if the accumulated let-7 miRNAs in XRN2-depleted cells were not rendered non-functional, the levels of their target mRNAs should get affected. Therefore, we compared the levels of the validated target mRNAs of let-7 family members in control and XRN2-depleted A549, HUH7, and U87-MG cells. We noted a considerable reduction in the mRNA levels of the aforementioned targets in the experimental cells (Figure 2A, S2D) , which clearly suggested that the accumulated let-7 family members are functionally active.

To augment our observations derived from cell culture-based studies, we investigated if the mRNA levels of the validated let-7 targets KRAS, NRAS, MYC, HMGA2 and CDC25A were also accordingly affected in patients displaying high expression of XRN2 mRNA in a public transcriptomics dataset for hepatocellular carcinoma (TCGA-HPCC). Each of these mRNAs was elevated in patients having higher than median XRN2 expression as compared to patients with lower than median expression of XRN2 ( Figure 2B ). Similar trends were also noticed from the analyses of public transcriptomics dataset of clinical glioblastoma (TCGA-GBM) and lung adenocarcinoma (TCGA-LUAD; Figure S3A , S3B) tumor samples. These 7 observations indeed corroborated with our results obtained from the aforementioned XRN2-depleted cell lines. Having observed this dysregulation of the oncogenic let-7 targets, we then investigated the relationship of let-7 family members with XRN2 mRNA levels in the same TCGA-HPCC dataset. Based on the heatmap in Figure 2D , which indicated that let-7a, -7c, -7f, -7g, and miR-98 depict an inverse relationship with XRN2 mRNA expression, we chose to perform all subsequent analyses on a randomlyselected subset of the TCGA-HPCC, TCGA-GBM, and TCGA-LUAD datasets broadly possessing the greatest combined mean fold change for let-7c and let-7g for low over high XRN2 mRNA expression, as well as the strongest combined inverse correlation of let-7c and let-7g with XRN2 mRNA expression (see Methods for detailed selection procedure), hereafter referred to as a 'representative subset'. We expected that this approach would allow us to select a set of samples, where XRN2-mediated regulation of let-7 miRNAs was optimally operational for further comparative studies. We found that in case of HPCC, in addition to let-7c and let-7g, let-7a and let-7f were inversely correlated with XRN2 ( Figure 2E ), and indeed also displayed reduced abundance in case of high XRN2 ( Figure 2F ) with varying degrees of significance in this 'representative subset', which suggests that the dynamics between XRN2 mRNA and let-7 miRNAs from the complete TCGA-HPCC dataset indeed get reflected in the 'representative subset'. Notably, here the global status of let-7a, -7c, and -7g also exhibited equivalence with our cell line-derived observations ( Figure 2D ). Of note, let-7e showed a stronger inverse relationship with XRN2 in GBM than in HPCC or LUAD ( Figure S3C , S3D). These data indicate that the specificity of XRN2-mediated regulatory activities vary in different clinical tissue samples, similar to the way it varied in the above-described defined cancer cell lines.

Notably, a perfect let-7a-target mRNA can also be a target for the other members of the let-7 family. But the magnitude of regulatory effect of a given member would be dependent on its abundance, coherence with the target expression, as well as its subcellular localization; thus making a given target mRNA relatively more specific to a particular family member. We observed that let-7b and let-7i do not show any significant changes in their levels upon XRN2 knockdown in the examined cell lines, and let-7i was also not negatively correlated with XRN2 mRNA in any of the TCGA-HPCC, TCGA-LUAD or TCGA-GBM datasets ( Figures 1B, 2D , 2E, S3C, S3D). To confirm that the unaffected levels of let-7b and let-7i were indeed manifested functionally, we checked the levels of known let-7b and let-7i target mRNAs, namely, KLK6 (let-7i target) 29 , PLK1, and CCNB2 (let-7b targets) 30 . As expected, levels of these mRNAs were not affected upon XRN2 depletion ( Figures 2C, S2E ). KLK6 mRNA levels were additionally also not affected in the clinical tumour sample data for hepatocellular carcinoma, glioblastoma, or lung adenocarcinoma ( Figures 2B, S3A , S3B), further strengthening our observations regarding the specificity of the regulatory activity of XRN2 on miRNAs in vivo. 8 Outcomes of the aforementioned analyses of the patient samples intrigued us to investigate whether the effect of XRN2 extends beyond the family of let-7 miRNAs. Accordingly, we compared the levels of other miRNAs with XRN2 mRNA in the abovementioned patient-derived datasets. Around 10% of all the cellular miRNAs, including some of the highly conserved and functionally important ones, showed a significant inverse relationship with XRN2 mRNA levels in the 'representative subsets' from all three datasets ('representative subsets' for TCGA-GBM and TCGA-LUAD were obtained following the same criteria and process described for TCGA-HPCC) ( Figure 2G , 2H, 2I, left panels). To validate and complement our in silico observations, we selected one 'potential non-let-7 XRN2-substrate' miRNA from each dataset, and checked its level in a cell line of the respective lineage upon XRN2 depletion. We assessed the levels of miR-122 in HUH7 cells (HPCC), miR-124 in U87 cells (GBM), and miR-29a in A549 cells (LUAD) 2,31 .

Indeed, we observed ~4-fold upregulation in case of miR-122 in XRN2-depeleted HUH7 cells ( Figure 2G , right panel), ~3-fold for miR-124 in XRN2-depleted U87 cells ( Figure 2H , right panel), and around 2-fold for miR-29a in XRN2-depleted A549 cells ( Figure 2I , right panel). These observations and results indicate that XRN2 has a significant impact on the total miRNA pools in different cell lineages and therefore, XRN2-mediated miRNA turnover might constitute an important pathway in cellular miRNA metabolism.

Dysregulated XRN2 expression is a common motif in several types of cancers, with XRN2 mRNA expression being increased in tumour tissues of brain, cervix, lung, head and neck, colon, prostate, and testis, as compared to normal tissues 22 . Increased XRN2 mRNA expression has been observed to be associated with increased risks of developing lung cancer and oral squamous cell carcinoma (OSCC), as well as increased severity and poor prognosis in these cancers 23, 25 . Decreased expression of miRNAs of the let-7 family have also been correlated with poor survival of patients suffering from multiple types of cancers 28 . As our results demonstrated XRN2-mediated turnover of let-7 miRNAs in cancer cells from lung adenocarcinoma, hepatocellular carcinoma, and glioblastoma ( Figure Indeed, these analyses further verified that elevated levels of XRN2 mRNA were associated with worsened survival outcomes in the patients who contributed to these datasets, though the relationship between elevated XRN2 expression and worsened survival was weak in LUAD as compared to either HPCC or GBM. let-7 miRNAs are known to repress proliferation and facilitate differentiation by affecting multiple critical pathways in human cells 24 . Previous studies in human lung adenocarcinoma and ovarian cancer cell lines 9 have reported that overexpression of certain let-7 family members results in anti-proliferative effects and reduced survival of cancer cells 24 . Our observations on the levels of mature let-7 miRNAs upon XRN2 knockdown, with a concomitant effect on the levels of crucial target-oncogenes, prompted us to check if XRN2 depletion can lead to the reversal of tumour phenotypes in these cancer cell lines. Notably, cell migration is a crucial process in cancer progression that allows cancer cells to metastasize to different tissues through the circulatory system. To test whether XRN2 depletion can affect the migration of these cancer cells, we performed a wound healing assay, which allowed us to study directional migration of cancer cells in vitro. We observed that the migration of A549 and HUH7 cells undergoing XRN2 knockdown was significantly reduced as compared to the control cells ( Figure 3B , 3C, S4A, S4B). We further performed clonogenic assays, which examines the ability of a single cancer cell to undergo multiple divisions and grow into a colony. This assay performed with the aforementioned two cell lines revealed diminished colony formation ability in XRN2-depleted cells as compared to the control cells ( Figure 3D , 3E, S4C, S4D). Of note, cell migration assay couldn't be performed with U87 cells as they start forming spheroids before reaching confluence. Also, survival of U87 cells was drastically affected upon XRN2 depletion, which made them unsuitable for performing clonogenic assays that require prolonged maintenance. Overall, these results demonstrated that the changes in miRNA levels upon XRN2 depletion are indeed physiologically significant as they affect the proliferation and migration of the cancer cells.

Notably, certain let-7 family members have already been shown to function as vital regulators of the Epithelial-to-Mesenchymal transition (EMT), which is considered to be a prerequisite for acquisition of malignant phenotypes like metastasis by epithelial cancer cells 24 . Accordingly, we checked the expression levels of E-cadherin (CDH1), which is primarily expressed in epithelial cells, and is known to undergo a reduction in its expression during EMT, and the mesenchymal marker Vimentin (VIM), which shows increase in its level during EMT. As expected, depletion of XRN2 in A549 cells, drastically reduced the VIM mRNA expression, while the expression of CDH1 mRNA showed a significant upregulation. Another mesenchymal marker, SNAI1, also showed a decrease in its expression level in XRN2-depleted cells ( Figure   3F ). mRNA levels of certain important matrix metalloproteases such as MMP2 and MMP9, which are vital for metastasis of cancer cells, were also found significantly downregulated in XRN2-depleted cells ( Figure   3G ). Notably, similar results were also obtained with HUH7 cells (Figure S4E, S4F) . Together, these results indicated that XRN2 depletion impedes EMT transition and metastatic progression in cancer cells.

miRNA biogenesis is completed with the generation of an AGO-loaded mature miRNA, as a constituent of a macromolecular complex called miRNA-induced silencing complex (miRISC) 1 . We wanted to check how XRN2 depletion affects the mature miRNAs bound to the AGO proteins. Accordingly, we immunoprecipitated the endogenous AGO2 from either control or XRN2-depleted A549 cells to check the levels of let-7a miRNA associated with immunoprecipitated AGO2. A significantly increased signal was recorded from AGO2 immunoprecipitated from the XRN2-depleted cells ( Figure 4A ). Similar observations were made upon immunoprecipitation of AGO2 from HUH7 and U87 cells ( Figure S5A ). Furthermore, we checked the levels of other let-7 sisters associated with immunoprecipitated AGO2 from A549 cells. Here, we observed that amongst all the let-7 miRNAs, let-7a, -7c, -7e, -7g, and miR-98 showed the highest accumulation in immunoprecipitated AGO2 from XRN2-depleted cells compared to control ( Figure 4B ), and that further indicated to a dislocation in the cascade of events leading to turnover of these miRNAs.

We were intrigued by this observation to check whether those miRNAs undergo a 'release step' from AGO before degradation, as it has been reported elsewhere 8 , and how XRN2 may affect this process.

Accordingly, we performed an ex vivo 'miRNA release assay', as has been described earlier 8 , employing immunoprecipitated endogenous AGO2 from A549 cells ( Figure 4C ). Immunoprecipitated bead-bound AGO2, upon incubation with control lysate and a further recovery, showed a significant loss of let-7a signal, when compared to the signal obtained from the bead-bound AGO2 incubated with buffer ( Figure 4D , which clearly demonstrated that those miRNAs are not associated with AGO2, but still not degraded. This data suggested that a diminished level of XRN2 in knockdown samples might be good enough to perform the 'release activity' but unable to degrade the released let-7a miRNA. Alternatively, XRN2 is kinetically linked to the process of let-7a release from AGO2, performed by a dedicated 'miRNA release factor' that is not present in molar excess in comparison to XRN2. However, since, no XRN2 protein could be detected with immunoprecipitated AGO2 (Figure S5B ), it could be suggested that either the AGO2-XRN2 interaction is highly transient or there is a yet unidentified 'miRNA release factor' that facilitates miRNA release from AGO2 for its degradation. Notably, subjecting the above-mentioned control cell lysate with protein denaturing conditions resulted in a complete loss of its 'miRNA releasing activity', which suggested a proteinaceous nature of the 'miRNA release factor' ( Figure S5C ).

As treatment with total cell lysate resulted in an efficient release of let-7a from immunoprecipitated AGO2, we checked whether other let-7 sisters got similarly released from the AGO2 protein. We observed that let-7a and let-7g underwent release to the extent of 85-90%, while the other sisters showed 60-80% release from AGO2 ( Figure 4E ).

Spanning over a decade, several reports have highlighted an alternative pathway for miRNA degradation, which primarily relies on release or tailing-and-trimming/ trimming-mediated destabilization of miRNAs from miRISC upon binding to target RNAs that possess extensive complementarity with the miRNAs 14, 16, 32, 33 . Accordingly, in vitro-reconstituted AGO2-guide RNA complexes have been demonstrated to unload the guide RNA strand upon incubation with high molar excess of perfectly complementary target RNA sequences 32 . We therefore sought to perform a comparative study between the release of the AGO2-bound miRNAs mediated by the extensively complementary target RNAs, and the above-mentioned total cell lysate devoid of any nucleic acid due to treatment with micrococcal nuclease.

We began by monitoring the release of let-7a from endogenous AGO2 in the presence of increasing concentration of an in vitro-transcribed target RNA, perfectly complementary to let-7a ( Figure 5A-D, S5D) .

We observed release of around 90% AGO2-bound let-7a upon incubation with ~150 folds molar excess of the target RNA at 37°C for 30 minutes ( Figure 5D ). Interestingly, very little or no miRNA was released from AGO2, when the reaction was performed with equimolar concentration of target RNA ( Figure 5C ).

Contrastingly, ~80% of endogenous AGO2-bound let-7a was released upon incubation with total lysate (75 µg), where amount of the lysate was in equivalence with that of the immunoprecipitated AGO2 employed in the 'miRNA release assay' (Figure 5C, 5E) . Notably, with ~5% immunoprecipitation efficiency, the immunoprecipitated AGO2 employed in each miRNA release assay reaction was equivalent to ~75 µg of total lysate ( Figure S5E) , and from which ~20 femtomoles of let-7a could be extracted ( Figure S5F ). Upon incubation with 150 µg of total lysate, the release of let-7a was further facilitated to the extent of ~90% ( Figure 5E ). Similarly, let-7d, which was only 60% released upon treatment with 75 µg total lysate, showed a greater magnitude of release upon treatment with higher concentrations of total lysate ( Figure 5E ). These results clearly demonstrated that the release of miRNA from miRISC that could be achieved at a high molar excess of target RNA, could also be achieved with total protein lysate, but at a physiologically relevant concentration. Of note, in the aforementioned target-mediated release assay, we also measured the magnitude of release for other members of the let-7 family, for whom the employed let-7a-target RNA was only partially complementary. Indeed, the incompletely complementary target failed to efficiently release let-7i (4 mismatches) or let-7d (2 mismatches) from the endogenous AGO2, even when it was present at 150-fold molar excess ( Figure 5D, S5G) . Together, our data suggest that protein-mediated miRNA release from miRISC is an important mechanism for miRNA metabolism in humans.

Interestingly, we observed that let-7b, let-7d and let-7i, unlike other let-7 family members, were not released to a substantial amount from the endogenous AGO2 upon exposure to the total lysate ( Figure 4E ). This could possibly be due to the following three reasons: (i) their release is partially dependent on target RNAs, (ii) the factor(s) responsible for their release from the specific miRISCs were partially inactive in the employed total lysate due to extraction conditions, or (iii) the composition of the respective miRISCs housing let-7b, -7d and -7i were such that they were resistant to the releasing activity, which might confer a greater stability and longer half-life to them. Indeed, by inhibiting RNA Pol II with α-amanitin, we could examine their relative stabilities, and they appeared to be more stable than the other members of the let-7 family ( Figure 5F ). This data suggested that the composition of a specific miRISC or a specific miRNP structure formed by a set of proteins and a given miRNA, may override a general turnover mechanism, and thus play an important role in determining the functionality of the given miRNA.

As previous reports have implicated ZSWIM8 Ubiquitin ligase and Tudor-SN (SND1) as two crucial regulators of miRNA abundance in mammalian cells 13, 19, 20 , we became interested to learn the relative contribution of these candidates towards negative regulation of the expression levels of the let-7 family members in a set of patients, where XRN2-mediated regulation of let-7 family members is favoured. In this direction, we compared the fold changes in the mean levels of let-7 family miRNAs between the low and high expression categories of XRN2, SND1 and ZSWIM8 mRNAs in the 'representative subset' for the TCGA-HPCC clinical tumour sample dataset defined earlier ( Figure 2E -I). In this subset, let-7a, -7c, -7d, -7f, -7g and miR-98 showed a stronger inverse relationship with XRN2 mRNA than with those of the other two enzymes, which is further borne out by their increased accumulation in case of low XRN2 mRNA levels than for the other enzymes considered ( Figure 6A, 6B) . However, in this subset, the expression of SND1 mRNA also possessed an appreciable inverse relationship with some of the let-7 family members.

Additionally, we found that the levels of a liver-specific miRNA, miR-122 2 , were also negatively affected upon increased expression of all three candidates, though maximally for XRN2 ( Figure 6A ). Further we wanted to check the global picture of the inverse relationships of XRN2, SND1, and ZSWIM8 mRNA levels with different let-7 family members. Accordingly, we repeated the comparison using the entire dataset. In the case of TCGA-HPCC, we observed a prominent and comparable negative effect of both SND1 and XRN2 mRNA expression on the levels of many let-7 family members ( Figure S6A ). However, high ZSWIM8 mRNA expression appeared to have a much weaker effect on these miRNAs than the other two candidates ( Figure S6A ). miR-122 was most negatively affected by XRN2 mRNA levels ( Figure S6A ). We also made observations from both the 'representative subsets' and the complete datasets for TCGA-GBM and TCGA-LUAD, employing the same method of analysis as for TCGA-HPCC. All three candidate enzymes had comparable contributions to the negative regulation of let-7 family members in the 'representative subset' for TCGA-GBM, while in the whole dataset, ZSWIM8 appeared to be the most 13 prominent candidate, followed by SND1 and XRN2. This is consistent with previous reports, which suggested that TDMD might be more important phenomenon in the context of neuronal cells 16, 33 . Notably, high mRNA levels of all three enzymes had a negative effect on the level of miR-124, a neuron-specific miRNA 2 , in the whole dataset for TCGA-GBM ( Figure S6B, S6C) . Also, in the complete dataset for TCGA-LUAD, it was ZSWIM8 mRNA that had the strongest negative relationship with let-7 family miRNAs, followed by SND1 and XRN2 mRNA ( Figure S6D) . Although, in case of clinically important miR-29a 31 , XRN2 was the only candidate to have an appreciable negative relationship in this complete dataset ( Figure   S6E ). The above observations suggested that all three candidates have roles in regulating levels of let-7 family miRNAs, and their relative contribution towards maintaining let-7 levels vary considerably, not just in different tissues, but also in different cohorts of patients. XRN2 appears to have a markedly important role in regulating many let-7 family members, especially in hepatocellular carcinoma, and is a contributor to miRNA homeostasis in multiple cancer types, alongside other candidates implicated in miRNA turnover.

Furthermore, since, let-7 family members have been adequately characterised to be involved in regulating several important cellular pathways 24 , we went on to investigate if any other major cellular pathways, aside from those directly associated with cancers, were affected in relation to XRN2 expression in the employed clinical tumour transcriptomics datasets. Gene Ontology (GO) terms are used to annotate genes with their functional roles in pathway databases such as the KEGG 34 and Reactome databases 35 . Therefore, we performed GO term enrichment analysis on genes, which are predicted or validated targets of let-7 family members (as per human TargetScan 7.2). Among these genes, whose mRNAs were also positively correlated with XRN2 mRNA in the relevant TCGA-HPCC dataset were selected for downstream analysis ( Figure 6C ). Given the roles of let-7 family members in regulating pluripotency, especially through the conserved LIN28B axis 24 , it was unsurprising to find that one of the enriched terms was related to regulation of stem cell pluripotency, and other terms related to differentiation. Similarly, due to the known roles of let-7 family members in regulation of genes involved in controlling the cell cycle like CDK6, CDK8, CDC25A and CYCLIN D 24 , we also found terms related to cell cycle processes and mitotic checkpoints to be enriched. Enrichment of terms related to viral infections, especially of RNA viruses such as HTLV-1, measles, and Hepatitis C were also enriched. We additionally observed that regulation of cellular senescence through the p53 and Ras-MAPK/RTK signalling pathways 24 were also enriched terms in the term enrichment analysis. Since, let-7 family members are also known to regulate the TGF-receptor TGFBR1, as well as STAT3 and PI3K-AKT 24 , terms related to TGF-and PDGF signalling were also enriched. Furthermore, it was interesting to observe that terms related to collagen biosynthesis, focal adhesion and syndecan interactions displayed enrichment as well. Notably, this could be because let-7a, let-7b and let-7g are known to regulate the expression of type 1 collagen genes 24 . Outcomes of this analysis 14 suggest a strong involvement of both XRN2 and let-7 family members in regulation of cell proliferation and differentiation, as well as cell and extracellular matrix (ECM) homeostasis.

Very similar GO terms were enriched in the outcome of a similar analysis performed on the TCGA-GBM dataset ( Figure S6F ). The outcome of the GO term enrichment analysis performed on the TCGA-LUAD dataset was less intriguing, since the variety of enriched terms was much lower (unpublished observations).

This could most probably be due to a higher sample heterogeneity in the TCGA-LUAD dataset as compared to the TCGA-HPCC and TCGA-GBM datasets, which might have allowed far fewer GO terms to reach the requisite significance threshold for representation. Notably, we found very similar GO terms enriched in the TCGA-HPCC and TCGA-GBM datasets, as well as cross-enrichment of hepatocellular carcinoma terms in GBM and glioma terms in HPCC. Probably, it indicates that the functional roles played by XRN2, and let-7 family miRNAs are preserved across different cell lineages, and further suggests far-reaching effects of XRN2 in various aspects of epithelial and glial development and homeostasis.

Our results demonstrate that in human cancer cells, turnover of let-7 family of miRNAs is a layered process involving multiple steps (Figure 7) , and it acts in conjunction with other regulatory forces to govern the functionality of miRNAs. We furnish the role of XRN2 as a critical regulator of miRNA abundance with a prominent impact on the physiology of cancer cells. Our findings suggest that the putative 'miRNA release factor' and XRN2-mediated miRNA decay might together constitute an important pathway to regulate the stability and function of let-7 members.

A previous study reported that XRN2 binds pre-miR-10a in a DICER-independent manner and accelerates the maturation of miR-10a in human lung cancer cells 25 , but did not provide insights on the members of let-7 family. Another study demonstrated reduced degradation of a synthetic miRNA by an XRN2-depleted lysate, but did not augment the observation through in vivo evidence 36 . Here, we demonstrate that in a number of human cancer cell lines of lung, hepatic, and neuronal origin, XRN2 depletion led to the accumulation of a number of mature let-7 family members without any effect on their pri-or pre-miRNAs.

It was also observed that XRN2 depletion led to a significant decrease in the levels of oncogenes that are validated let-7 targets in cancer cells. The clinical significance of our findings was reinforced by the analyses of public tumour transcriptomics data from patients with lung adenocarcinoma, hepatocellular carcinoma, and glioblastoma, especially considering the heterogeneity of the patient cohorts. The transcriptomics analyses demonstrated an appreciable inverse relationship between the levels of certain let-7 family members and XRN2 mRNA in the clinical tumour samples, which corroborated with our cell linebased observations. They also revealed that let-7 target genes positively correlating with XRN2 mRNA expression are involved in multiple signalling pathways regulating cell proliferation and differentiation, as well as in biogenesis and maintenance of the extracellular matrix in both epithelial and neuroglial cell lineages. This partially reveals why elevated XRN2 has previously been associated with worse survival outcomes in lung cancer patients, and also accentuates prior observations on overexpressed XRN2 mRNA in spontaneously occurring lung tumours in mice and human tumour microarray data analysed through genome-wide association studies 22 . Apart from let-7 members, XRN2 depletion also affected the levels of certain tissue-specific miRNAs, like miR-122 in hepatocellular carcinoma cells, and miR-124 in case of glioblastoma cells, which are known to have important functions in the developmental processes of liver and neuronal tissues, respectively, and whose dysregulations have been linked to multiple disease states 2 .

Our observations from the analyses of TCGA-HPCC and TCGA-GBM datasets indeed revealed the association of high XRN2 mRNA levels with worse survival in hepatocellular carcinoma and glioblastoma, respectively. Whereas, analysis of TCGA-LUAD dataset further reiterated the association of high XRN2 16 mRNA levels with worse survival in lung adenocarcinoma. Our motivation towards performing bioinformatic analyses employing the TCGA datasets was because of how they encompass samples from a large and diverse patient population for each cancer type and the corresponding dataset, while ensuring that all sequencing data acquisition and processing have been performed using uniform procedures. We expect that our observations, especially considering the large and heterogenous clinical population represented in the TCGA datasets, would advocate for wider applicability of the molecular signatures of XRN2 in determining cancer severity than what could be suggested from smaller or more limited datasets. With the revelation of a clear inverse relationship between the levels of let-7 members and XRN2 expression, we found that for two other important factors implicated in miRNA turnover, such an inverse relationship between the let-7 family members and those concerned mRNAs (SND1 and ZSWIM8) was comparatively weaker in patient-derived samples favouring XRN2-mediated let-7 regulation. In a larger set of samples, we found that XRN2 is a critical player in maintenance of let-7 miRNA levels in hepatocellular carcinoma and contributes to let-7 regulation in glioblastoma and lung adenocarcinoma, alongside the other two critical players implicated in miRNA turnover (Tudor-SN and ZSWIM8). These observations indeed remind us of the fact that complexity of cancer is often an outcome of perturbations of multiple regulators acting in parallel. Therapeutic approaches relying on targeting miRNAs, even from the same family, would be greatly benefited by a more thorough understanding of the landscape of miRNA turnover, potentially allowing the design of drug formulations that can target multiple such pathways at the same time.

It has been previously described that exogenously introduced target RNAs having extensive complementarity to the cognate miRNAs can lead to target RNA-directed tailing and trimming of endogenous miRNAs, and influence miRNA stability and function 14 . Structural studies have demonstrated that the 5'-end of a miRNA remains anchored in a pocket in the MID domain of AGO, and the 3'-end is bound by the PAZ domain. The miRNA remains largely buried in a channel between the two lobes of AGO, except the bases of the miRNA-seed region, which remain partially exposed to the surface of the host protein. However, extensive base pairing of AGO2-bound miRNA with target RNAs disrupts the 3'-end anchoring and renders the 3′-end accessible for enzymatic activity 37 . Also, specific AGO-bound miRNAs that have their 3'-ends exposed, have been demonstrated to undergo an oligo-tailing, followed by degradation by DIS3L2 18 . Since, for most cellular AGO-bound miRNAs, their 3'-ends are protected by the AGO proteins, unless they encounter extensively complementary targets, it is improbable that DIS3L2 mediated miRNA turnover pathway could be the foremost channel for miRNA turnover inside the cell 18 .

Our studies provide evidence that XRN2-mediated degradation of miRNAs is dependent on an upstream event of 'miRNA release' from miRISCs, carried out by a yet unidentified proteinaceous factor(s).

Moreover, our results indicate that protein factor-mediated release of AGO-bound let-7 appears to be far more efficacious compared to complementary target RNA-mediated mechanism 32 . Additionally, this two-step turnover mechanism has the potential to prevent immediate degradation of an AGO-released miRNA, which might enable shuttling of some specific candidates to the yet uncharacterized non-canonical pathways that utilize mature miRNAs without any association with known components of miRNA metabolism like AGO, GW proteins, etc. 38, 39 , whereas, most of the released miRNAs might get delivered to the bona fide 'miRNases' for degradation.

Interestingly, the recently reported 'miRNase'-Tudor-SN co-immunoprecipitates with RISC subunits and appears to be a component of miRISC 13 . The intracellular signalling events marking an AGO-residing miRNA for termination of its function remain largely unknown, and it will be important to know the responsible factors and the mechanism that allows such miRNAs to travel from the grasp of AGO to Tudor-SN for degradation.

Downregulation of endogenous miRNAs has been shown to be induced by certain viral RNAs [40] [41] [42] [43] , as well as a few genome-encoded transcripts 16 , which bear extensive complementarity to specific miRNAs.

Recently, highly complementary target interactions were demonstrated to induce TDMD, independently of tailing and trimming events, by causing destabilization of AGO proteins through the ZSWIM8 ubiquitin ligase 19, 20 . It has been proposed that ZSWIM8 directly recognizes the conformational changes induced in AGO upon extensive base-pairing of the resident-miRNA (miR-7) with a highly complementary target (Cyrano lncRNA), but the mechanism for this recognition remains to be understood. Since, it is well known that 3'-tailing often acts as an important signal for downstream processes 14, 33, 44 in its absence, it is quite possible that there might be yet unknown factor(s) that would facilitate recognition of the conformational changes elicited upon highly complementary miRNA-target interaction by ZSWIM8, in order to mediate TDMD. Additionally, the nucleases participating in the degradation of miRNAs, post their release from AGO undergoing proteolysis, needs to be identified. It is also noteworthy that the half-lives of AGO proteins are longer than most of the miRNAs they host 5, 6 , suggesting that miRNA release and degradation through destabilization of AGO by ZSWIM8 and related factors is unlikely to be the major pathway for miRNA turnover. Intriguingly, a highly complementary viral RNA element-mediated degradation of a host miRNA may be considered an efficient viral strategy to destabilize the host system. However, turnover of a miRNA due to highly complementary interaction with the cognate endogenous target, mediated by the destabilization of the AGO itself, would certainly pose an energy burden on the cellular system. Importantly, such unusual endogenous targets are not plentiful, and thus, TDMD through destabilization of AGO may have evolved as a regulatory measure under special circumstances for a specific set of miRNAs with exclusive functions. Contrary to TDMD, pairing of target mRNA with its cognate miRNA has been shown to exert a protective effect on miRNAs of specific families from cellular degradation in C. elegans 9 .

Furthermore, target mediated protection of a few miRNAs was also reported to be crucial for controlling miRNA levels and their subcellular localization in human cells 45 . Also, in human cells, biogenesis of specific miRNAs was stimulated upon binding to their cognate target mRNAs 46 . Collectively, these reports indicate that probably there is no single principle and mechanism of turnover that determine the stability or vulnerability of miRNAs in different tissues and organisms. Crucially, while TDMD has been suggested to be an important mechanism for miRNA homeostasis in neuronal cells 33 , highly complementary let-7 targets suitable for TDMD are not yet known. Given our observations of ZSWIM8 mRNA levels being prominently associated with a negative effect on let-7 levels in glioblastoma and lung adenocarcinoma, ZSWIM8 might be involved in additional pathways to exert its regulatory effect in patient samples of these cancer types, and further exploration of those regulatory roles would be critical for better understanding of disease states. let-7 members are amongst the highly conserved animal miRNAs that critically regulate expression of many genes governing diverse biological processes 1, 24 . Interestingly, aside from their well-studied roles as tumour suppressors, let-7 family miRNAs are also important for mounting immune responses in vertebrates upon exposure to pathogens 47 . It is known to repress both the pro-inflammatory cytokine interleukin 6 (IL-6) and the anti-inflammatory cytokine IL-10 during Salmonella infection in murine phagocytic RAW264.7 cells, and thereby regulate early and late immune responses 47 . Interestingly, XRN2 depletion in RAW264.7 cells led to the upregulation of certain let-7 family members, and that in turn affected the levels of interleukins (unpublished observations), and this clearly makes XRN2 a potential candidate for modulation of inflammatory responses during infection. Notably, many of the patients infected by SARS-CoV-2 have been reported to show uncontrolled elevation in cytokines with persistent acute inflammation. Failure of the cellular homeostatic machinery to appropriately downregulate these 'cytokine-storms' leads to a vicious cycle of accumulating inflammation and subsequent lethality of the condition 48 . Since, let-7 has been shown to be capable of targeting multiple interleukins, its regulation through XRN2 might play a vital role in such critical situations. Moreover, IL-6 is also implicated in regulating various hypothalamic functions and contributes to the activation of hypothalamo-hypophyseal-pituitary axis, and thus to overall metabolism in vertebrates 49, 50 . Supported by our cell culture-based results employing neuronal cancer cells and patient samples of neuronal tissues, it is possible that in the aforementioned neuro-endocrine tissues, let-7 plays a modulatory role on IL-6, and that in turn gets regulated by XRN2. Overall, these evidence and suggestions indicate that the regulation of let-7 through XRN2-mediated degradation might be conserved across the eukaryotic systems, and have widespread implications ranging from cancer physiology to host responses during infection. Further exploration of the functions of XRN2 and other components of the XRN2mediated miRNA turnover pathway could shed new light on how these crucial regulators of the transcriptome themselves get regulated, and thus, might also provide opportunities for the discovery of new diagnostics and therapeutics. A. Immunoprecipitation using anti-EIF2C2 (anti-AGO2) antibody from total cell lysate of the indicated samples reveal increased signal of let-7a in AGO2 of XRN2-depleted cells. One half of each immunoprecipitate was subjected to northern probing or TaqMan analysis, the other half to anti-AGO2 western probing and served as loading control.

B. TaqMan analyses (n = 3; mean ± SEM) reveal increased accumulation of several let-7 sisters in immunoprecipitated AGO2 from XRN2-depleted samples.

C. Schematic representation of miRNA release assay. Bead-bound immunoprecipitated AGO2/ miRISC were subjected to different treatments as indicated, and recovered back as bead and supernatant fractions.

One half of each re-recovered bead fraction was subjected to northern probing or TaqMan analysis, the 29 other half to anti-AGO2 western probing. Northern probing was performed for equivalent amount of RNA extracted from the supernatant fraction. Loss of signal from the re-recovered intact AGO2/ miRISC subjected to a specific treatment, compared to that of the buffer/high salt treatment, indicates release of AGO2/ miRISC-bound let-7a miRNA.

D. miRNA Release assay with immunoprecipitated AGO2 was performed with the indicated samples.

Release of let-7a is diminished in the assay employing XRN2-depleted total lysate compared to the one with control lysate. E. TaqMan analyses (n = 4; mean ± SEM) of the RNA isolated from re-recovered bead-bound fraction reveal differential release of different let-7 sisters from immunoprecipitated AGO2 subjected to control cell lysate. A. Comparison of the mean fold change in expression of different let-7 family members and the liverspecific miRNA miR-122 between low and high expression groups for XRN2, SND1 (Tudor-SN), and ZSWIM8 mRNAs in the 'representative subset' of the TCGA-HPCC dataset. In this subset, low XRN2 mRNA levels are associated with larger increases in let-7a, -7c, -7d, -7f, -7g, miR-98 and miR-122 levels in comparison to low expression of either SND1 or ZSWIM8 mRNAs. B . Maximum relative accumulation of let-7 sisters with low expression of XRN2, SND1, and ZSWIM8 mRNAs from (A). Six out of nine let-7 members are maximally accumulated in case of low XRN2 mRNA levels over low SND1 or ZSWIM8 mRNA levels. Among these, let-7a and let-7c display the most accumulation for low XRN2 mRNA levels, while let-7b and let-7i are accumulated more for low SND1 33 mRNA levels. let-7e is not appreciably accumulated for low expression of any of the three enzymes considered.

C. Gene Ontology (GO) Term Enrichment analysis indicate the most enriched GO terms for the genes positively correlated with XRN2 expression in the TCGA-HPCC dataset, include candidates from multiple signalling pathways, different types of cancers and viral infections, as well as pathways related to ECM metabolism. "Signalling by Receptor Tyrosine Kinases" is the most significantly enriched GO term (y-axis: significance of enrichment), while "Gene expression (Transcription)" of both the most abundant and enriched GO term (x-axis: fraction of analysed genes associated with a GO term) is also strongly correlated with XRN2 mRNA expression (depicted marker radii on the left box are proportional to the mean Spearman correlation of all the GO-term associated genes with XRN2 mRNA expression). This observation is consistent with many let-7 targets being transcription factors. A two-step miRNA turnover model, where XRN2 acts on the miRNA upon its release from the miRISC/ AGO by a yet unknown 'miRNA release factor'.

Target recognition and regulatory functions

The role of non-coding RNAs in oncology

Control of stress-dependent cardiac growth and gene expression by a microRNA

The let-7 target gene mouse lin-41 is a stem cell specific E3 ubiquitin ligase for the miRNA pathway protein Ago2

Time-resolved small RNA sequencing unravels the molecular principles of microRNA homeostasis

Global analyses of the dynamics of mammalian microRNA metabolism

MicroRNA turnover: When, how, and why

Active turnover modulates mature microRNA activity in Caenorhabditis elegans

Target-mediated protection of endogenous microRNAs in C. elegans

The Decapping Scavenger Enzyme DCS-1 controls microRNA levels in Caenorhabditis elegans

Differential regulation of microRNA stability

Human polynucleotide phosphorylase selectively and preferentially degrades microRNA-221

Proceedings of the National Academy of Sciences of the United States of America

Tudor-SN-mediated endonucleolytic decay of human cell microRNAs promotes G1/S phase transition

Target RNA-directed trimming and tailing of small silencing RNAs

The 3′-to-5′ exoribonuclease Nibbler shapes the 3′ ends of microRNAs bound to Drosophila Argonaute1

MicroRNA degradation by a conserved target RNA regulates animal behavior

Post-transcriptional generation of miRNA variants by multiple nucleotidyl transferases contributes to miRNA transcriptome complexity

AGO-bound mature miRNAs are oligouridylated by TUTs and subsequently degraded by DIS3L2

The ZSWIM8 ubiquitin ligase mediates target-directed microRNA degradation

A ubiquitin ligase mediates target-directed microRNA decay independently of tailing and trimming

The multifunctional RNase XRN2

Genetic variants cis-regulating Xrn2 expression contribute to the risk of spontaneous lung tumor

Upregulation of XRN2 acts as an oncogene in oral squamous cell carcinoma and correlates with poor prognosis

Let -7 as biomarker, prognostic indicator, and therapy for precision medicine in cancer

XRN2 promotes EMT and metastasis through regulating maturation of miR-10a

Lin28 mediates the terminal uridylation of let-7 precursor microRNA

Expression atlas update: From tissues to single cells

MicroRNA expression profiles classify human cancers

let-7i-5p inhibits the proliferation and metastasis of colon cancer cells by targeting kallikreinrelated peptidase 6

Joint bioinformatics analysis of underlying potential functions of hsa-let-7b-5p and core genes in human glioma

MicroRNAs in non-small cell lung cancer and idiopathic pulmonary fibrosis

Highly complementary target RNAs promote release of guide RNAs from human Argonaute2

Potent degradation of neuronal miRNAs induced by highly complementary targets

Kyoto encyclopedia of genes and genomes

The reactome pathway knowledgebase

The human decapping scavenger enzyme DcpS modulates microRNA turnover

Structural basis for target-directed microRNA degradation

miR-328 functions as an RNA decoy to modulate hnRNP E2 regulation of mRNA translation in leukemic blasts

miR-29 acts as a decoy in sarcomas to protect the tumor suppressor A20 mRNA from degradation by HuR

Down-regulation of a host microRNA by a Herpesvirus saimiri noncoding RNA

Murine cytomegalovirus encodes a miR-27 inhibitor disguised as a target

Degradation of cellular miR-27 by a novel, highly abundant viral transcript is important for efficient virus replication in vivo

Selective degradation of host microRNAs by an intergenic HCMV noncoding RNA accelerates virus production

A tale of non-canonical tails: gene regulation by post-transcriptional RNA tailing

Resolving subcellular miRNA trafficking and turnover at single-molecule resolution

Target-dependent biogenesis of cognate microRNAs in human cells

Analysis of the host microRNA response to Salmonella uncovers the control of major cytokines by the let-7 family

Cytokine storm in COVID-19: The current evidence and treatment strategies

Interleukin-6-deficient mice develop mature-onset obesity

Interleukin-6 regulates the expression of hypothalamic neuropeptides involved in body weight in a gender-dependent way

RT-qPCR (n = 4; mean ± SEM) analyses reveal reduction in let-7 target mRNAs, as indicated

Relative mRNA expression levels (arbitrary units) of the indicated let-7 target genes in the high and low XRN2 mRNA expression groups from the TCGA-HPCC RNA-Seq dataset of clinical hepatocellular carcinoma (HPCC) tumour samples. KLK6, a let-7i target that is unchanged upon XRN2 depletion in cell lines (C), is also not affected by different XRN2 expression levels in the clinical tumour dataset

Known let-7b and let-7i targets remain unaffected in XRN2-depleted HUH7 cells (RT-qPCR analyses

Expression patterns of let-7 family members in relation to XRN2 mRNA levels in the TCGA-HPCC dataset. let-7a, -7c, -7f and -7g are visibly inversely correlated with XRN2. Colour Gradient: Yellowhighest expression, Blue -lowest expression (relative to the minimum and maximum expression of the

Volcano plot of significance of correlation (y-axis: -log10(p)) vs

let-7 family members with XRN2 mRNA expression in the 'representative subset' for TCGA-HPCC further confirms the inverse relationship of the four let-7 sisters (let-7a, 7c, -7f, and -7g) with XRN2 mRNA, thus strengthening the conclusion that they are likely to be targeted by XRN2 in HPCC from (E)

While only let-7c was significantly negatively altered (top-left) in the high XRN2 group, let-7a, let-7c, and let-7g were also negatively altered to lower extents, and all of these are likely to be targeted by XRN2 in HPCC

With respect to XRN2 mRNA levels, the number of miRNAs detected to be significantly negatively altered in TCGA-HPCC (G) was 77, which included let-7c and the liver-specific miRNA miR-122, while 153 miRNAs were negatively altered in TCGA-GBM (H) including the neuron-specific miRNA miR-124, and among the 36 miRNAs negatively altered in TCGA-LUAD (I) was the clinically important miR-29a

quantification (n = 4; mean ± SEM) of AGO2-associated let-7a and let-7d (re-recovered beadbound fraction) upon treatment with different concentrations of total cell lysates

RNA was isolated from cells harvested at 0, 12, 24, 36 and 48 hrs following α-amanitin (10 µg/ml) treatment, and TaqMan quantification

We thank Pradipta Kundu for technical help with cell culture. Results presented here are in part based on analyses performed with data generated by the TCGA Research Network: https://www.cancer.gov/tcga, specifically the TCGA-HPCC, TCGA-GBM, and TCGA-LUAD datasets. We are grateful to the specimen donors and research groups, who made the TCGA initiative possible. This work has been supported by the SERB/ DST (Govt. of India). RN received CSIR-Syama Prasad Mookerjee Fellowship (Govt. of India), and TC was supported by KVPY/ MHRD (Govt. of India).