INTRODUCTION
Endogenous retroviruses (ERVs) are remnants of ancient infections that comprise ~8% of the human genome. Although normally silenced, ERVs can be reactivated by inflammatory stimuli
1. Upon reactivation, ERV-derived nucleic acids trigger sustained type I interferon (IFN) and inflammatory cytokine production that drives chronic inflammation and inflammatory diseases
2,3. Innate immune cells such as macrophages, as the major first responders to a wide range of inflammatory insults
4, are particularly vulnerable to ERV derepression. Aberrant ERV reactivation in macrophages can be sensed as viral mimicry, creating a vicious cycle of chronic inflammation, and contributes to autoimmunity and tissue pathology
4,5. Yet, the mechanisms by which macrophages preserve ERV silencing to maintain immune homeostasis under fluctuating inflammatory pressures remain largely undefined. Elucidating how macrophages suppress ERVs may reveal mechanisms driving sterile inflammation, a key factor in most non-communicable diseases
6.
ERV silencing relies on multiple epigenetic mechanisms, including histone modifications, DNA methylation, and RNA methylation
7-9. A key aspect of these epigenetic controls is their tight coupling to cellular metabolism, as nearly all epigenetic enzymes utilize metabolites as cofactors or substrates
10. Notably, macrophage activation elicits a profound metabolic reprogramming, exemplified by the induction of immune-responsive gene 1 (IRG1)
11,12. The metabolic gene encodes aconitate decarboxylase 1 (ACOD1), a mitochondrial enzyme that catalyzes the production of the immunomodulatory metabolite itaconic acid (ITA)
12,13. Studies using
Irg1-deficient mice or exogenous ITA have established the immunomodulatory role of
Irg1/ITA in diverse inflammatory contexts
12,14. Previous studies have shown that ITA exerts its immune modulatory effects through diverse mechanisms, such as inhibition on succinate dehydrogenase to cause succinate accumulation and metabolic reprogramming
15, protein alkylation to induce electrophilic stress and activation of the KEAP1–NRF2 signaling pathway
16, impairment of aerobic glycolysis
17, and inhibition of TET2-mediated DNA demethylation to dampen acute inflammatory responses
18. However, it remains unclear whether ITA acts as a metabolic safeguard against aberrant ERV reactivation to preserve immune balance in macrophages.
In this study, we demonstrate that ITA competitively displaces α-ketoglutarate (α-KG) to inhibit ALKBH5, potentiating m6A-mediated degradation of ERV transcripts. This mechanism prevents aberrant ERV reactivation and the cytoplasmic nucleic acid accumulation that provokes sustained proinflammatory responses in macrophages. Our findings thus uncover a homeostatic function for the ITA–ALKBH5 axis in restraining ERV-driven sterile inflammation, and suggest that ALKBH5 inhibition may offer a viable strategy against chronic inflammatory pathologies.
RESULTS
Loss of IRG1 promotes endogenous retroelement expression in macrophages
To explore the potential role of
Irg1 induction in regulating ERVs during immune stimuli, we performed in-depth analysis of gene expression in bone marrow-derived macrophages (BMDMs) from wild-type (WT) and
Irg1−/− mice following stimulation with two inflammatory stimuli: lipopolysaccharide (LPS) and vesicular stomatitis virus (VSV). Gene expression profiling revealed significant enrichment of viral protein–cytokine receptor interaction pathways in
Irg1−/− macrophages compared with WT macrophages under both stimulation conditions (Fig. 1a, b). Notably, transcripts from multiple ERV families were consistently upregulated in
Irg1−/− BMDMs following LPS stimulation or VSV infection compared with WT BMDMs (Fig. 1c, d). Among ERV classes, intracisternal A-particle (IAP) elements, a relatively young and transcriptionally active group of ERVs
19, exhibited a pronounced induction in
Irg1−/− macrophages (Fig. 1c, d). Quantitative RT-PCR confirmed increased
IAP transcripts upon LPS stimulation and VSV infection in WT macrophages, which was further amplified in
Irg1−/− cells (Fig. 1e). RNA fluorescence
in situ hybridization (FISH) demonstrated enhanced cytoplasmic accumulation of
IAP transcripts in
Irg1−/− macrophages compared to WT controls after LPS stimulation (Fig. 1f). In addition, LPS stimulation also induced expression of
HERVK, a young and transcriptionally active human ERV family, and this effect was further enhanced by
IRG1 knockdown in human macrophages (Fig. 1g, h), suggesting a conserved role of IRG1 in restraining ERV activation during innate immune responses.
Irg1 promotes m6A-dependent degradation of ERV RNAs
ERVs are normally kept silent by multiple layers of epigenetic regulation under homeostatic conditions but can be reactivated in response to various cellular stress or infection
1. Such reactivation is often associated with perturbations in chromatin-based mechanisms, including histone modifications
7, DNA methylation
8, and RNA methylation
9. To investigate the chromatin context underlying ERV reactivation, we re-analyzed published transposase-accessible chromatin using sequencing (ATAC-seq) data
20. This analysis revealed that macrophage activation by inflammatory stimuli, including Toll-like receptor agonists such as LPS, cytokines such as IFNγ, and microbial infections such as
Listeria or LCMV, is commonly accompanied by widespread chromatin opening (Supplementary Fig. S1a), providing a permissive environment for ERV transcriptional induction. Consistent with this notion, our ATAC-seq analysis showed LPS stimulation led to the gain of 5,708 and loss of 4,336 accessible chromatin peaks in mouse BMDMs compared with unstimulated controls (Fig. 2a). Notably, no significant difference in chromatin accessibility was observed between
Irg1−/− and
Irg1+/+ BMDMs following 24 h of LPS stimulation (Fig. 2a). Furthermore, analysis of nascent RNA using 4-thiouridine (4sU) labeling revealed no proportional increases in newly synthesized ERV transcripts in
Irg1−/− BMDMs compared to WT cells post-LPS challenge (Fig. 2b), suggesting that the observed upregulation of ERVs in activated
Irg1−/− macrophages is unlikely to result from altered chromatin accessibility or transcriptional activation. Instead, when macrophages were treated with transcriptional inhibitor actinomycin D, we found that the stability of
IAP RNA was increased in
Irg1−/− BMDMs (from 1.0 to 1.6 h upon LPS stimulation, and from 1.9 to 2.6 h following vesicularstomatitis virus (VSV) infection; Fig. 2c). Together, these results suggest that IRG1 suppresses ERVs through a post-transcriptional mechanism.
We previously reported that ITA inhibits the enzymatic activity of TET2 DNA demethylase, thereby suppressing the expression of proinflammatory genes and dampening inflammation
18. TET2 has been reported to be recruited to chromatin by Paraspeckle component 1 (PSPC1) to promote
MERVL transcripts degradation via 5-hydroxymethylcytosine (5hmC) modification
21. To determine whether TET2 mediates IRG1/ITA-promoted ERV RNA turnover, we performed RNA decay assays in BMDMs derived from catalytically inactive
Tet2HxD knock-in (KI) mutant mice. The HxD mutant harbors double substitutions (H1295Y and D1297A) in mouse
Tet2, corresponding to H1382Y and D1384A in human
TET2, respectively, which occur recurrently in human myeloid malignancy. These mutations disrupt binding to the essential cofactor Fe
2+ and thus abolish TET2 catalytic activity
22. Upon ITA treatment, we found that
IAP RNA stability was reduced in both
Tet2+/+ and
Tet2HxD macrophages with similar decay rates (Supplementary Fig. S1b), suggesting that IRG1/ITA-driven ERV RNA destabilization does not rely on TET2 catalytic function.
m
6A, the most abundant internal modification in eukaryotic mRNAs, is a critical determinant of RNA stability and turnover
23. To determine whether the IRG1/ITA pathway influences m
6A methylation, we examined
Irg1 expression in macrophages under different immune stimuli.
Irg1 expression, as expected, was robustly induced in BMDMs stimulated with LPS, IFNγ, herpes simplex virus (HSV), or VSV (Supplementary Fig. S1c), resulting in intracellular ITA accumulation reaching millimolar concentrations (Supplementary Fig. S1d). These stimuli increased global RNA m
6A methylation in BMDMs (Supplementary Fig. S1e), consistent with previous findings
24,25. Strikingly, global m
6A levels were significantly lower in
Irg1−/− BMDMs than WT controls under all four inflammatory conditions examined (Fig. 2d), indicating that IRG1-derived ITA is required to sustain m
6A methylation during macrophage activation.
To delineate transcriptome-wide m
6A dynamics, we employed glyoxal and nitrite-mediated deamination of unmethylated adenosines (GLORI), which enables quantitative single-nucleotide resolution mapping of m
6A sites
26. Following LPS stimulation,
Irg1−/− BMDMs exhibited significantly less m
6A deposition across both protein-coding (pcRNAs) and non-coding RNAs (ncRNAs), with a more pronounced reduction observed in ncRNAs (Fig. 2e). Among these, repeat RNAs (reRNAs) showed the most extensive reduction of m
6A modification (Fig. 2f), particularly at long terminal repeat (LTR) loci (Supplementary Fig. S1f). Within the LTR class, ERVs constituted the predominant fraction (~86%) (Supplementary Fig. S1g) and displayed marked reductions in m
6A deposition upon
Irg1 deletion (Supplementary Fig. S1h). Notably, down-regulated m
6A sites within the top 20 ERV elements, exemplified by the
IAPEz-int, showed substantial loss of m
6A methylation while retaining the canonical DRACH (D = G/A/T, R = A/G, H = T/C/A) motif, supporting the reliability of m
6A site calling (Supplementary Fig. S1i−j). To complement this site-specific analysis, we further performed m
6A RNA immunoprecipitation (RIP)-seq to obtain a broader view of m
6A distribution across
IAP elements. Consistent with the GLORI-seq results, we found that m
6A signals were predominantly enriched in the 5′UTR of
IAP transcripts and that
Irg1 deficiency led to a marked reduction in m
6A levels within this region (Fig. 2g). In comparison, several
LINE1 elements were transcriptionally upregulated in
Irg1−/
− BMDMs upon LPS stimulation (Fig. 1c, d), but only
Penelope1_Vert exhibited detectable m
6A sites, and neither of the identified sites showed a significant change upon
Irg1-deficiency in LPS-activated macrophages (Supplementary Fig. S1k), suggesting that alternative regulatory mechanisms govern
LINE1 activation. Together, these results suggest that IRG1/ITA facilitates m
6A deposition on LTR elements, including ERV transcripts, thereby accelerating their RNA turnover during immune activation.
To further investigate how IRG1/ITA influences m
6A modification, we examined and found that mRNA levels of key components of the m
6A methyltransferase complex (
Mettl3, Mettl14, and
Wtap) and demethylases (
Alkbh5 and
Fto) remained largely unchanged in macrophages stimulated with LPS, IFNγ, HSV, or VSV (Supplementary Fig. S2a). These results suggest that IRG1/ITA regulates m
6A dynamics independently of transcriptional control of the m
6A modification machinery in cells. We next conducted a chemical pulldown using a biotinylated ITA probe incubated with RAW264.7 cell lysates, followed by streptavidin-mediated enrichment and mass spectrometry analysis (Supplementary Fig. S2b). This approach identified 435 proteins significantly enriched in the biotin-ITA pulldown compared to the control pulldown using biotin (Supplementary Fig. S2c). Among the top hits in the biotin-ITA pulldown was ALKBH5 (Fig. 2h), a major m
6A demethylase that regulates RNA stability and metabolism
24,27. The ITA-ALKBH5 interaction was further confirmed by Western blotting (Fig. 2i), reaffirming the molecular link between ITA and the m
6A demethylation machinery.
Next, we evaluated the functional significance of ITA–ALKBH5 interaction. In WT BMDMs, LPS stimulation induced a robust increase in global m6A levels, an effect that was not seen in Alkbh5−/− macrophages (Fig. 2j). In addition, Alkbh5−/− BMDMs displayed elevated basal m6A levels but failed to further increase m6A upon stimulation with IFNγ, HSV, or VSV (Fig. 2k), indicating that ALKBH5 is required for stimulus-induced, dynamic m6A regulation during immune activation. GLORI-seq in Irg1f/f, Lysm-cre:Irg1f/f, Lysm-cre:Irg1f/f/Alkbh5f/f(Irg1/Alkbh5 double knockout, DKO) BMDMs revealed that m6A levels at the previously identified sites within the 5′UTR of IAPEz transcripts (Supplementary Fig. S1j) were reduced in Irg1-deficient BMDMs after LPS challenge, whereas this reduction was restored in Irg1/Alkbh5 DKO cells (Supplementary Fig. S2d), establishing a direct link between ALKBH5 and IRG1/ITA-dependent control of ERV RNAs.
Moreover, we measured ERV RNA decay kinetics in BMDMs derived from myeloid-specific Irg1 KO or Irg1/Alkbh5 DKO mice. Strikingly, DKO BMDMs exhibited reduced IAP RNA stability compared to Irg1−/− macrophages even in the absence of exogenous ITA (Fig. 2l; Supplementary Fig. S2e), indicating that loss of ALKBH5 is sufficient to destabilize ERV transcripts. While addition of exogenous ITA to Irg1−/− macrophages destabilized ERV RNAs, this effect was completely abolished in Irg1/Alkbh5 DKO cells (Fig. 2l; Supplementary Fig. S2e). Furthermore, depletion of the m6A reader Ythdf2, but not the other readers examined, eliminated the difference in IAP RNA abundance between Irg1−/− and WT BMDMs following LPS stimulation (Fig. 2m; Supplementary Fig. S2f−h). RIP assays demonstrated that Ythdf2 directly binds IAP RNA, and this interaction was reduced in Irg1-deficient macrophages (Supplementary Fig. S2i, j). Collectively, these results support a model in which ITA inhibits ALKBH5, leading to increased m6A deposition on ERV transcripts and their selective clearance via Ythdf2-mediated RNA decay.
Itaconate directly binds to and inhibits ALKBH5 activity
We previously reported that ITA is a structural analog of α-KG and competitively inhibits α-KG-dependent TET2 DNA demethylase
18. Given that ALKBH5 is an α-KG-dependent m
6A demethylase (Fig. 3a), we hypothesized that ITA might competitively inhibit ALKBH5 activity. To test this, we conducted
in vitro enzymatic assays using purified recombinant human ALKBH5 protein, and found that ITA potently inhibited the catalytic activity of ALKBH5 in a dose-dependent manner, as determined by dot-blot (Fig. 3a) and LC-MS analysis (Supplementary Fig. S3a, b), with a half-maximal inhibitory concentration (IC
50) of 261 μM (Fig. 3b). This inhibitory concentration is well below endogenous ITA levels (~2 mM) observed in mouse macrophages following LPS challenge (Supplementary Fig. S3c, d). Moreover, the inhibitory effect of ITA on ALKBH5 was completely reversed by the addition of excess α-KG
in vitro (Fig. 3c), consistent with the competitive inhibition mechanism.
We also examined whether ITA affects the activity of FTO, the other α-KG-dependent m
6A demethylase. In contrast to its potent inhibitory effect on ALKBH5, ITA had no effect on FTO enzyme activity, regardless of whether m
6A-modified RNA or
N7-methylguanosine-capped (m7G cap) RNA were used as a substrate
28 (Supplementary Fig. S3e, f). This result supports a specific molecular mechanism by which IRG1/ITA regulates m
6A dynamics by inhibiting ALKBH5, but not FTO, in activated macrophages.
Furthermore, we determined the co-crystal structure of human ALKBH5 in complex with ITA at a resolution of 2.1 Å (Fig. 3d). Structural comparison with the previously reported ALKBH5−α-KG complex revealed that ITA occupies the ALKBH5 active site in a binding mode nearly identical to α-KG, engaging key catalytic residues
29 (Fig. 3d). Specifically, the C5 carboxyl group of ITA forms a salt bridge with Arg277 and a water-mediated hydrogen bond with Tyr195, the same residues engaged by the C5 carboxyl group of α-KG (Fig. 3d). In addition, the C1 carboxyl group of ITA forms a hydrogen bond with His266, replacing the ionic interactions formed by α-KG’s C1 and C2 groups. In this way, ITA competes with Fe(II) and α-KG for binding at the active site of ALKBH5 but cannot be decarboxylated, thereby inhibiting ALKBH5 catalytic activity. Microscale thermophoresis (MST) assays confirmed direct binding of ITA to ALKBH5, yielding a dissociation constant (
Kd) of approximately 23 μM, comparable to that of α-KG (
Kd = 56 μM) (Fig. 3e). Of note, sequence alignment revealed that H266 and R277 are strictly conserved across the ALKBH family, and Y195 is conserved in most family members except ALKBH7 and FTO (Supplementary Fig. S3g). Alanine substitution of these three residues (R277, H266, or Y195) abolished binding of both ITA and α-KG to ALKBH5 (Supplementary Fig. S3h), and individual mutation of R277 or Y195 eliminated ALKBH5 demethylase activity (Fig. 3f). Consistent with these structural and biochemical findings, ITA treatment increased global m
6A levels in WT macrophages, but had no effect in
Alkbh5−/− cells (Fig. 3g). Moreover, supplementation with dimethyl-α-KG reversed the ITA-induced increase of m
6A in BMDMs (Fig. 3h). Collectively, these results clearly demonstrate that ITA functions as an inhibitor of ALKBH5 by competitively displacing α-KG at the catalytic site.
Itaconate-mediated inhibition of ALKBH5 suppresses ERVs and attenuates renal fibrosis
ERVs encode reverse transcriptase that converts viral RNA into complementary DNA (cDNA) and subsequently synthesizes a second strand to form double-stranded DNA (dsDNA). These cytosolic nucleic acid species, including RNA:DNA hybrids and dsDNA, act as DAMPs
3 and engage innate nucleic acid sensors to trigger type I IFN responses
30. Consistent with this model, time-course analysis of ERV expression revealed that
IAP RNA levels remained persistently elevated at 24–48 h post-LPS in
Irg1−/− BMDMs compared to WT controls (Supplementary Fig. S4a). Notably,
Ifnb1 expression displayed a biphasic pattern in
Irg1−/− macrophages.
Ifnb1 expression in
Irg1−/− macrophages was reduced at 4 h post-LPS, in line with a recent study
31, showing that loss of
Irg1 reduces mitochondrial RNA (mtRNA) release and thereby attenuates the early and acute induction of
Ifnb1. However,
Ifnb1 expression was elevated by approximately two-fold during the late phase, at 24–48 h after LPS challenge, suggesting a delayed and enhanced type I IFN signaling response following ERV reactivation. Moreover, IFNβ itself induced
IAP expression (Supplementary Fig. S4b), supporting a feed-forward loop in which initial ERV derepression drives IFNβ production, which in turn further enhances ERV expression and progressively amplifies chronic inflammation over time (Supplementary Fig. S4c).
To define the sensing pathway responsible for enhanced type I IFN responses in Irg1-deficient macrophages, we first assessed the contribution of RNA sensing. Depletion of Mavs failed to attenuate the enhanced induction of Ifnb1 in Irg1−/− macrophages following LPS stimulation (Supplementary Fig. S4d, e), arguing against a dominant role for RNA-sensing pathways. In contrast, DNA FISH analysis using IAP-specific probes revealed significantly more cytosolic IAP DNA in LPS-stimulated Irg1−/− BMDMs compared to WT cells (Supplementary Fig. S4f). This accumulation was abolished by treatment with a reverse transcriptase inhibitor (RTi), Tenofovir and Emtricitabine, confirming the retroelement origin of cytosolic DNA species. In accord, RTi treatment markedly reduced the expression of Ifnb1 and interferon-stimulated genes (ISGs) in LPS-stimulated Irg1−/− macrophages (Supplementary Fig. S4g), directly linking ERV-derived DNA to IFN induction. Mechanistically, Irg1−/− BMDMs exhibited prolonged phosphorylation of STING and TBK1 beginning at 12 h post-LPS (Supplementary Fig. S4h), indicating sustained activation of the cGAS–STING pathway. Correspondingly, late-phase induction of Ifnb1 and ISGs, including Ifit1, was significantly upregulated in Irg1-deficient macrophages at 24 h post-LPS, and this exaggerated response was absent in cGas−/− cells (Supplementary Fig. S4i, j), suggesting that the interferon response is cGAS-dependent.
To further establish ALKBH5 as the functional target of ITA in silencing ERV expression, we examined the impact of exogenous ITA on ERV RNA levels in Irg1-KO and Irg1/Alkbh5-DKO BMDMs. ITA supplementation significantly reduced IAP RNA levels in Irg1-KO cells, but not in Irg1/Alkbh5-DKO cells (Supplementary Fig. S4k), indicating that ALKBH5 is essential for ITA-mediated ERV repression. Consistent with this regulatory mechanism, exogenous ITA reduced the late-phase expression of Ifnb1 and Ifit1 in Irg1-KO macrophages post-LPS, but this inhibitory effect was completely lost in Irg1/Alkbh5-DKO cells (Supplementary Fig. S4l). Collectively, these findings delineate an IRG1/ITA–ALKBH5–ERV regulatory axis that mitigates aberrant type I IFN responses triggered by ERV-derived DNA in macrophages.
Chronic innate immune activation driven by nucleic acid sensing promotes inflammatory cell infiltration, tubular epithelial injury, and fibroblast activation, ultimately leading to fibrotic remodeling
32,33. It was previously reported that aberrant ERV activation can elicit fibroinflammatory responses and drive the development of kidney diseases in mice
34. Macrophages are major contributors to chronic kidney disease (CKD), as their accumulation within the kidney strongly correlates with renal damage and has been shown to predict patient responses to immunosuppressive therapy
35,36. Collectively, these findings suggest that macrophage-intrinsic mechanisms that restrain ERV activity may be essential for limiting renal inflammation and disease progression.
To investigate whether loss of IRG1 in macrophages disrupts ERV silencing and thereby exacerbates CKD pathogenesis, we examined kidney biopsies from patients with immunoglobulin A nephropathy (IgAN, also known as Berger’s disease) (Fig. 4a), the most common form of primary glomerulonephritis characterized by mesangial IgA deposition and a major cause of renal failure
37. Immunostaining revealed robust HERVK protein expression in renal tissues from human individuals with IgAN (Fig. 4b). Notably, HERVK expression inversely correlated with IRG1 expression in CD68
+ renal macrophages (Fig. 4c), supporting a potential regulatory relationship between IRG1 and ERV expression. HERVK expression in renal macrophages also negatively correlated with estimated glomerular filtration rate (eGFR), a key clinical indicator of kidney function (Fig. 4d), whereas IRG1 expression in renal macrophages positively correlated with eGFR (Fig. 4e). Together, these findings imply that IRG1 may help maintain renal homeostasis in CKD by suppressing ERV expression and limiting ERV-driven immune activation.
To comprehensively profile ERV activation in myeloid cells from CKD patients, we performed Ribo-minus RNA-seq on circulating CD14
+ monocytes derived from CKD patients and healthy donors. This analysis revealed upregulation of multiple ERV families, accompanied by induction of a smaller subset of
LINE1 elements (Supplementary Fig. S5a). Among the most significantly induced ERVs was
HERVK (HML-2), one of the most recently integrated human ERV families, which retains intact open reading frames encoding viral proteins such as Gag, Pol, Env, and Pro
38. Consistent with these findings, we found that
HERVK-gag expression was significantly increased in CD14
+ monocytes from CKD patients compared with healthy donors (Fig. 4f). Of note,
HERVK-gag expression remained elevated despite conventional therapies (Fig. 4f), including renin-angiotensin system inhibitor (RASi) and sodium-glucose cotransporter-2 inhibitor (SGLT-2i)
39,40, predicting a sustained inflammatory insult. Accordingly, the expression of
IFNB1 and downstream ISGs, such as
IFIT1, was not reduced by these therapies (Supplementary Fig. S5b). In contrast, treatment with either exogenous ITA or RTi markedly suppressed
HERVK-gag and ISG expression in patient-derived monocytes (Supplementary Fig. S5c, d). Combined ITA and RTi treatment showed no additive effect (Supplementary Fig. S5d), suggesting that ITA restrains its associated interferon response in CKD patients through controlling aberrant endogenous retroelement activation.
To assess whether the IRG1/ITA–m
6A regulatory axis on ERVs is conserved in humans, we performed MeRIP-qPCR analysis on CD14
+ monocytes from CKD patients, as previously described
41. We identified m
6A-enriched regions on
HERVK transcripts and designed MeRIP-qPCR primers targeting these regions (Supplementary Fig. S5e). Using this approach, we found that ITA treatment significantly increased m
6A enrichment at peaks #1, #2, #3, and #5 on
HERVK transcripts in patient-derived monocytes (Supplementary Fig. S5e, f). Collectively, these findings indicate that IRG1/ITA-m
6A regulation on ERVs occurs in human cells.
Next, we examined whether IRG1/ITA–ALKBH5 axis can restrain ERV-driven inflammation
in vivo. To this end, we employed a folic acid (FA)-induced nephropathy model, in which high-dose FA precipitates as crystals in renal tubules, mimicking crystal nephropathy and tubulointerstitial fibrosis
42. Notably, ITA was undetectable in renal macrophages under basal conditions but was robustly induced following FA treatment, reaching intracellular concentrations of approximately 90 μM (Supplementary Fig. S6a). By day 21 after FA injection, kidneys from
Lysm-cre:Irg1f/f mice with myeloid-specific conditional deletion of
Irg1 exhibited markedly increased expression of fibrosis-associated genes, including
Col1a1,
Col3a1, and
Fn1 (Supplementary Fig. S6b). Consistently, Masson’s trichrome staining illustrated extensive fibrotic expansion in the kidneys of
Lysm-cre:Irg1f/f mice following FA challenge (Supplementary Fig. S6c, d). Functionally, these mice displayed elevated blood urea nitrogen (BUN) levels and increased urinary albumin-to-creatinine ratios (Supplementary Fig. S6e), indicative of impaired renal function.
In accord, single-cell RNA sequencing (scRNA-seq) of kidney tissues at different time points after FA injury revealed that the profibrotic signaling remained persistently up-regulated in fibroblasts of Lysm-cre:Irg1f/f mice post FA-injection (Supplementary Fig. S6f). Cell–cell communication analysis showed that fibroblasts predominantly received macrophage-derived signals (Supplementary Fig. S6g), and macrophage−fibroblast interactions were potentiated by the myeloid-specific conditional Irg1 deletion in the Lysm-cre:Irg1f/f mice at 21 days post FA-injection, particularly through profibrotic ligand–receptor pairs, such as Tgfb–Tgfbr and Pdgfb–Pdgfbr (Supplementary Fig. S6g). Interestingly, renal macrophages exhibit the most pronounced ERV upregulation among all analyzed cell types at day 21 following FA treatment, an induction that was further enhanced in Lysm-cre:Irg1f/f mice (Supplementary Fig. S6h, i). Consistent with the notion that the cGAS-STING pathway is highly expressed and readily activated in macrophages, this pathway-related gene signature (cGAS-score) exhibited the highest expression in renal macrophages (Supplementary Fig. S6j). Moreover, renal macrophages isolated from Lysm-cre:Irg1f/f mice expressed significantly higher levels of IAP, Ifnb1 and ISGs, including Ifit1 and Ifit3, than those from WT controls (Fig. 4g; Supplementary Fig. S6k). Together, these findings demonstrate that myeloid-derived Irg1 suppresses ERV reactivation and the downstream interferon response, thereby mitigating inflammation-driven fibrosis and preserving renal function following injury.
To further validate our findings, we employed a separate CKD model induced surgically by unilateral ureteral obstruction (UUO), which elicits sustained tubular stress, inflammation, and fibrosis distinct from crystal deposition in the FA-induced nephropathy model
43. Histological analyses demonstrated that
Lysm-cre:Irg1f/f mice with myeloid-specific
Irg1 deletion developed more severe renal pathology, characterized by pronounced tubular injury, increased inflammatory cell infiltration, and exacerbated interstitial fibrosis (Supplementary Fig. S7a). Consistently,
Lysm-cre:Irg1f/f mice post-UUO exhibited markedly elevated
IAP and
Ifnb1 expression in renal macrophages compared with those from WT controls (Supplementary Fig. S7b). More importantly, treatment with RTi effectively alleviated these pathological changes in
Irg1-deficient mice post-UUO (Fig. 4h). In accord, immunohistochemical analysis revealed that either RTi or ITA monotherapy attenuated renal fibrosis in
Lysm-cre:Irg1f/f mice post-UUO, as evidenced by decreased α-SMA-positive areas in the kidney (Supplementary Fig. S7c,e). Interestingly, combined ITA and RTi treatment showed no additive anti-fibrotic effect in kidneys from
Lysm-cre:Irg1f/f mice post-UUO (Supplementary Fig. S7c,e), reinforcing a causal role for IRG1/ITA mitigating inflammation-associated renal fibrosis by limiting ERV activation in macrophages.
Finally, we examined whether ALKBH5 mediates the reno-protective effect of IRG1/ITA in vivo. To this end, exogenous ITA was administered to Lysm-cre:Alkbh5f/f and control Alkbh5f/f mice following UUO (Fig. 4i). In Alkbh5f/f mice, exogenous ITA markedly suppressed ERV and Ifnb1 expression in renal macrophages post-UUO, whereas this effect was not seen in Lysm-cre:Alkbh5f/f mice (Fig. 4j). Consistently, ITA administration markedly decreased the expression of fibrotic marker genes (e.g. Col1a1, Col3a1, and Fn1) in control animals post-UUO but had no effect in Lysm-cre:Alkbh5f/f mice (Fig. 4k). Immunohistochemical analysis confirmed that ITA attenuated fibrotic activation, as evidenced by decreased α-SMA-positive areas in kidneys of WT controls, but not in Lysm-cre:Alkbh5f/f mice post-UUO (Fig. 4l). Collectively, these results demonstrate that ITA inhibits ALKBH5 and suppresses ERV-driven interferon responses, thereby protecting against fibrotic kidney injury (Supplementary Fig. S7f).
DISCUSSION
Our study uncovers a previously unrecognized role of ITA/IRG1 in ERV suppression in activated macrophages for maintaining immune homeostasis. By preventing aberrant ERV reactivation and the production of cytoplasmic nucleic acids — significant sources of endogenous DAMPs — this mechanism prevents sustained sterile inflammation, protects cells from chronic activation, and mitigates downstream tissue damage. These findings also provide new insights into how dysregulated ERV expression in macrophages can drive chronic inflammatory diseases such as CKD.
Our findings reconcile the apparently context-dependent effects of IRG1/ITA on type I interferon responses. A recent study showed that loss of
Irg1 reduces mitochondrial RNA release and attenuates the early, acute induction of
Ifnb1 after LPS stimulation
31. Consistent with this notion, we observed reduced
Ifnb1 expression in
Irg1-deficient macrophages at early time points (Supplementary Fig. S4a). However, at later stages of inflammatory stimulation,
Irg1-deficiency led to the accumulation of ERV-derived nucleic acids and sustained activation of the cGAS–STING pathway, resulting in enhanced
Ifnb1 and ISG expression. This late-phase interferon response was markedly suppressed by reverse transcriptase inhibitors or genetic deletion of
cGas (Supplementary Fig. S4), whereas mtDNA release was not increased in
Irg1-deficient macrophages at 24 h after LPS challenge (data not shown). These observations thus support a biphasic model in which IRG1 may promote early mtRNA-associated IFN-I induction, while restraining prolonged ERV-driven cGAS–STING activation and chronic interferon signaling during sustained immune stimulation.
Mechanistically, we show inflammation-induced ITA competitively inhibits ALKBH5 RNA demethylase by displacing α-KG at the enzyme’s catalytic site, thereby stabilizing m
6A marks on ERV transcripts to promote their decay. ITA inhibits ALKBH5 with an IC
50 of 261 μM, a concentration well below the millimolar levels that ITA can reach endogenously in activated macrophages (Supplementary Fig. S3c). This inhibition is specific, as ITA has no effect on the related α-KG-dependent RNA demethylase FTO. Our structural and mutational analyses elucidate the molecular basis for this discrimination between ALKBH5 and FTO. It will be important to elucidate the physiological significance of ITA-mediated selective inhibition of ALKBH5, but not FTO. Actually, the specificity of ITA is not limited to RNA demethylases; we have previously demonstrated that ITA also potently inhibits TET DNA demethylases while sparing α-KG-dependent JmjC histone demethylases
18. These findings thus suggest a division of labor in ITA-mediated epigenetic regulation: in macrophages, ITA inhibits TET2 to attenuate acute pro-inflammatory transcriptional programs, while its inhibition of ALKBH5 curtails ERV-dependent chronic inflammatory responses. Whether additional metabolites selectively regulate histone demethylases or the RNA demethylase FTO remains an open question. Collectively, our findings expand the role of metabolites in epigenetic regulation and highlight a remarkable degree of biochemical specificity within this regulatory network.
Although a subset of
LINE1 elements was upregulated in
Irg1-deficient macrophages after LPS stimulation, this was not accompanied by a global reduction in m
6A levels on these elements. Among the upregulated
LINE1-associated elements analyzed, only
Penelope1_Vert RNA contained detectable m
6A sites, and these sites were not significantly altered by
Irg1 deficiency (Supplementary Fig. S1k). This contrasts with
IAP elements, whose m
6A levels were markedly reduced upon
Irg1 loss in an
Alkbh5-dependent manner (Supplementary Fig. S2d). Together with previous evidence that
LINE1 m
6A is predominantly regulated by FTO
44, an enzyme not inhibited by ITA, these findings suggest that
LINE1 activation in
Irg1-deficient macrophages may occur through mechanisms largely independent of the IRG1/ITA–ALKBH5–m
6A axis. Thus, while ERV-derived nucleic acids represent a major driver of the sustained cGAS–STING–IFN-I response in our system, different retroelement families may be governed by distinct layers of regulation.
The pathological consequences of disrupting the ITA–ALKBH5–ERV axis underscore the biological relevance. In two independent murine models of nephropathy, myeloid-specific
Irg1 deletion led to elevated
IAP levels, heightened type I IFN signaling, and aggravated renal fibrosis. Conversely, pharmacological suppression of ERV activity by using RT inhibitors or exogenous ITA effectively mitigated fibrosis and restored renal function, supporting a causal role for ERV activity. It has to be noted that mouse
IAPs are more active than most human endogenous retroviruses (HERVs), which are generally older, fragmented, and epigenetically constrained. Nevertheless, our findings provide a rationale for further investigation in human settings. In support of clinical relevance, macrophages from IgAN patients show an inverse correlation between HERVK levels and IRG1 expression, consistent with ERV-associated inflammation in human CKD. Longitudinal cohort studies have reported that CKD is associated with roughly double the risk of cognitive dysfunction compared with age-matched populations, along with elevated risk of stroke and other cerebrovascular complications
45. These findings underscore the systemic impact of CKD and suggest that chronic inflammation, at least in part driven by dysregulated macrophage-ERV activity, may contribute to secondary complications in the brain and heart. In fact, standard therapies such as RAS or SGLT2 inhibitors, primarily target hemodynamic or metabolic pathways and do not directly address inflammation. Given the FDA-approved status of RT inhibitors and the relative safety of ITA, therapeutic strategies targeting the ITA–ALKBH5 axis warrant further exploration as potential approaches for ERV-associated CKD.
In addition to CKD, aberrant ERV activation has also been implicated in a wide range of autoimmune, neuroinflammatory, and degenerative disorders, including systemic lupus erythematosus (SLE)
46, multiple sclerosis (MS)
47, rheumatoid arthritis
48, type 1 diabetes, inflammatory bowel disease (IBD)
49,50, as well as neurodegenerative conditions such as Alzheimer’s disease (AD) and Parkinson’s disease (PD)
51,52, and neuropsychiatric disorders including schizophrenia, bipolar disorder, and autism spectrum disorders
53-55. Macrophages are commonly activated in these pathological conditions. We show that
IRG1 is robustly induced in macrophages by diverse stimuli (i.e. LPS, IFNγ, HSV, VSV), and also demonstrate that myeloid-specific
Irg1 deletion can heighten ERV levels by affecting their m
6A methylation and RNA turnover. These findings suggest that ITA-mediated ALKBH5 inhibition is a general mechanism for ERV suppression beyond CKD, implying a broad therapeutic potential of targeting the ITA–ALKBH5 axis to treat other ERV-driven inflammatory diseases.
METHODS
Animals and study approval
Irg1−/− mice (JAX stock #029340) were purchased from the Jackson Laboratory. Animals were backcrossed for more than 7 generations onto the C57BL/6J background. Tet2HxD KI mutant mice on C57BL/6J background were constructed by using CRISPR/Cas9 system and were kindly provided by Dr. Jiayu Chen and Dr. Shaorong Gao (School of Life Sciences and Technology, Tongji University, Shanghai). Myeloid-specific Irg1-deficient mice (Lysm-cre:Irg1f/f) on C57BL/6J background was generated by Shanghai Model Organisms Center. Myeloid-specific Alkbh5-deficient mice (Lysm-cre:Alkbh5f/f) were kindly provided by Dr. Huabing Li (Shanghai Jiao Tong University, Shanghai). All mice were housed, bred, and maintained under specific pathogen-free (SPF) conditions and were given unrestricted access to a standard diet and tap water. Animal experiments were performed at Fudan Animal Center in accordance with animal welfare guidelines, and the procedures were approved by the Ethics Committee of the Institutes of Biomedical Sciences (IBS), Fudan University (approval number: DSF-2020-066).
Mouse genotyping
For the genotyping of Irg1−/− mice, we followed the instruction from the Jackson lab. Briefly, three primers were used for PCR followed by agarose gel electrophoresis: GTGGGGAGG GGAACTATGAG (forward primer for both WT and mutant Irg1), ATTTGGAGGAACCCCATGAC (reverse primer for WT), CAGCCTCTAAGCCAGACAGC (reverse primer for mutant). The sizes of PCR fragments are 157 bp and 220 bp for the WT and mutant Irg1, respectively.
For the genotyping of Lysm-cre:Irg1f/f mice, two primers were used for the genotyping of flox: CAGGAAGGCCAGTGCTCAGTAATC (forward primer) and ACCTCCTCGCACCCCCTTTGTATG (reverse primer). The sizes of PCR fragments are 331 bp and 388 bp for the WT and mutant Irg1, respectively. For genotyping of Lysm-cre:Alkbh5f/f mice, two primers were used for the genotyping of flox: GCACAGTGGAGCACATCATG (forward primer) and CAGAGGGCAAGCAACCACAC (reverse primer). The sizes of PCR fragments are 243 bp and 300 bp for the WT and mutant Alkbh5, respectively. PCR amplification yielded products of 243 bp for the WT allele and 300 bp for the floxed allele. Moreover, four primers were used to genotype Lyz2-Cre: GGAGGATGCTTAAATAGCAGG (transgene forward primer), CATCACTCGTTGCATCGACC (transgene reverse primer), AGTGGCCTCTTCCAGAAATG (internal positive control forward primer), TGCGACTGTGTCTGATTTCC (internal positive control reverse primer). The sizes of PCR fragments are 150 bp and 521 bp for Cre+ and control, respectively.
Human samples
Kidney biopsy specimens from IgAN patients (n = 23) were collected by the Department of Nephrology, Zhongshan Hospital of Fudan University. Informed consent was received from each patient before surgery. This study was approved by the Ethics Committee of Zhongshan Hospital of Fudan University (approval number: B2021-067).
The whole blood from CKD patients receiving therapy or not were collected in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of Zhongshan Hospital, Fudan University (approval number: B2021-346R). Informed consent was obtained from all participants. Information about clinical samples used in this study is listed in Supplementary Tables S1–4.
Isolation of PBMCs
Peripheral blood (7.5 mL) from healthy donors was diluted with 22.5 mL PBS at room temperature to a total volume of 30 mL, and then carefully layered over 15 mL Ficoll-Paque PLUS (Cytiva) in a 50 mL tube. Samples were centrifuged at 800× g for 20 min at 22 °C using an Eppendorf centrifuge with brake setting at 0–1. Approximately 5 mL of the cloudy interphase containing PBMCs was gently collected and diluted to 40 mL with PBS, followed by centrifugation at 800× g for 5 min. Freshly isolated PBMCs were resuspended in MACS Rinsing Solution Buffer (Miltenyi) and passed through a 40 μm cell strainer to remove clumps. CD14+ monocytes were then isolated using CD14 MicroBeads (Miltenyi Biotec) according to the manufacturer’s instructions. Purified monocytes were seeded at 1 × 106 cells per well in low-attachment 6-well plates (Corning) and cultured in RPMI 1640 supplemented with 10% fetal bovine serum and 100 ng/mL recombinant human M-CSF (Novoprotein). Cells were differentiated into macrophages over 7 days.
Cell culture
RAW264.7 and HEK293T cells (ATCC) were maintained in DMEM with 10% fetal bovine serum and 1% penicillin–streptomycin at 37 °C in a 5% CO2 atmosphere. To generate BMDMs, mice were euthanized by CO2 asphyxiation, followed by cervical dislocation. Bone marrow cells were flushed from femurs and tibias using PBS, filtered through a 40 μm cell strainer. Cells were cultured in DMEM supplemented with 10% fetal bovine serum, 1% penicillin–streptomycin, and 100 ng/mL M-CSF (novoprotein) for seven days to allow macrophage differentiation, with medium changes on days 3 and 5. On day 7 of differentiation, BMDMs were passaged at ~80% confluency using Versene solution (Gibco) and reseeded for downstream assays. M-CSF was used throughout the differentiation process. Unless otherwise stated, cells were pretreated with pH-adjusted ITA (pH 7.0) for 3 h before stimulation with LPS (100 ng/mL).
siRNA transfection and lentivirus infection
For siRNA transfection of macrophages, cells were transfected on day 5 of differentiation. Mouse BMDMs were transfected using the Lonza Amaxa 2b Nucleofector™ system, and human monocyte-derived macrophages were transfected with Lipofectamine™ RNAiMAX (Invitrogen), following the manufacturers’ instructions.
Lentiviral particles were produced by transfecting HEK293T cells with transfer and packaging plasmids (psPAX2, pMD2.G) using Lipofectamine 2000 (Invitrogen). Viral supernatants were collected at 48 h, filtered, and stored or used immediately. Target cells (i.e. RAW264.7) were transduced with virus plus 8 μg/mL polybrene for 24 h. Antibiotic selection (puromycin or blasticidin) began 48 h after infection.
RNA extraction, cDNA synthesis, and quantitative PCR
Total RNA was extracted using the TransZol Up reagent (TransGen) according to the manufacturer’s instructions. cDNA was synthesized from 1 μg of total RNA using the EZscript Reverse Transcription Mix II (EZBioscience) with oligo (dT) and random primers. Quantitative PCR was performed using ChamQ Universal SYBR qPCR Master Mix (Vazyme) on a QuantStudio™ 6 Real-Time PCR System (Applied Biosystems). Gene expression levels were calculated using the ΔΔCt method and normalized to Gapdh or Rps18 as internal references. The primers are listed in Supplementary Table S5.
Cell fractionation
Cells were fractionated as follows: 5 × 106 to 1 × 107 cells were washed with cold 1× PBS/1 mM EDTA buffer and pelleted by centrifugation at 500× g. The pellet was resuspended in 200 μL ice-cold lysis buffer (10 mM Tris-HCl, pH 7.5, 0.1% NP-40, 150 mM NaCl) and incubated on ice for 5 min. The lysate was layered on a sucrose cushion (24% sucrose in lysis buffer) and centrifuged at 15,000× g for 10 min at 4 °C. The supernatant was collected as the cytoplasmic fraction. For the nuclear fraction, the pellet was resuspended in 100 μL ice-cold glycerol buffer (20 mM Tris-HCl, pH 7.9, 75 mM NaCl, 0.5 mM EDTA, 0.85 mM DTT, 0.125 mM PMSF, 50% glycerol). Cold nuclei lysis buffer (10 mM HEPES, pH 7.6, 1 mM DTT, 7.5 mM MgCl2, 0.2 mM EDTA, 0.3 M NaCl, 3 M urea) was added in double volume, and the mixture was vortexed for 4 × 5 s. After a 2-min incubation on ice, the sample was centrifuged at 15,000× g for 2 min at 4°C. The supernatant was collected as the nucleoplasm, and the pellet was washed with cold 1× PBS/1 mM EDTA and kept as the chromosome-associated fraction.
GLORI-seq and data analysis
Following rRNA depletion from chromatin-associated RNAs (caRNAs) using the Ribo-off rRNA Depletion Kit (Vazyme, N406) per the manufacturer's instructions, sequencing libraries were prepared from 50 ng of purified caRNA as per the GLORI method
26 and sequenced on a DNBSEQ-T7 platform (Majorbio Bio-Pharm Technology Co. Ltd., Shanghai, China).
The Illumina sequencing reads (R2) were processed through the following pipeline. First, adapter sequences and low-quality bases were trimmed using Trim Galore (v0.6.7) with the command: trim_galore -q 20 --stringency 1 -e 0.3 --length 35. Subsequently, PCR duplicates were removed using Seqkit (v.2.8.2) based on the 10-base-pair Unique Molecular Identifier (UMI) located at the 5' end of the R2 reads, executing the command: seqkit rmdup -s. The UMI sequences were then excised from these deduplicated reads with FASTX-Toolkit (v.0.0.14) using: fastx_trimmer -f 15. Finally, the processed reads were subjected to downstream analysis with GLORI-tools (github.com/liucongcas/GLORI-tools) under its default parameters.
ATAC-seq and data analysis
ATAC-seq for BMDMs was performed following a published protocol
56 and the transposition reaction was using ATAC-seq Kit from ActiveMotif. In brief, 50,000 BMDM cells (> 90% viability) were harvested at 500×
g at 4 °C for 5 min and washed once with cold PBS buffer. The cell pellet was resuspended in 50 µL cold ATAC-resuspension buffer (10 mM Tris-HCl, pH7.4, 10 mM NaCl, 3 mM MgCl
2, 0.1% Tween-20, 0.1% NP-40, 0.01% Digitonin) and incubated on ice for 3 min. We added 1 mL ATAC-resuspension lysis without NP-40 and digitonin into lysis and invert tube 3 times to mix. Nuclei pellet were collected at 500×
g for 10 min at 4 °C. The nuclei were resuspension in 50 µL transposition mix with 25µL 2×TD buffer (20 mM Tris-HCl, pH7.5, 10 mM MgCl
2, 20% dimethyl formamide), 2 µL transposase, 16.5 µL PBS, 0.5 µL 1% digitonin, 0.5 µL 10% Tween-20, 5 µL H
2O and the mixture was incubated at 37 °C for 30 min in a thermomixer with 1,000 rpm mix. The transposed DNA fragments were purified by DNA FragSelect XP Magnetic Beads (Changzhou Boyi Biotech, MP1301-02) and amplified through PCR with different indexed primers. Sequencing was carried out on an Illumina Novaseq 6000 platform with 150 bp paired-end mode.
Raw reads were trimmed by trim-galore package (version: 0.4.4) under default parameters, clean reads shorter than 35 bp were discarded. Clean reads were aligned to the mm10 genome using Bowtie2 with only uniquely mapped reads retained for downstream analysis. Reads aligned to the mitochondrial genome and PCR duplicated reads were removed. ATAC-seq peaks were identified by MACS2 (version: 2.1.2) with ‘-nomodel’ parameter. All peaks were merged into a united peak set by bedtools, and ATAC-seq read counts were quantified from BAM files and normalized using edgeR (version: 4.6.2) to correct for library size and compositional bias. Effective library sizes were used for CPM and differential peak analysis. Genome coverage bigwig files for heatmap were generated by deeptools (version: 3.5.6)
57 bamCoverage with the parameter ‘--normalizeUsing None --scaleFactor’. Heat maps were generated by deeptools (version: 3.5.6)
57 computeMatrix and plotHeatmap.
RNA-seq and data analysis
Total RNA was extracted using the TransZol Up reagent (TransGen) according to the manufacturer’s instructions. Total of 10 µg RNA was used to remove ribosomal RNA using DNA oligonucleotides hybridization and RNase H digestion as previously described with minor modification
58. In brief, equal number of probes (10 µg) was added into the mixture in 1× hybridization buffer (200 mM Tris-HCl, pH7.5, 1 M NaCl) in a final volume of 25 µL at 95 °C for 2 min and slowly ramped the temperature (‒0.1°C/s) to 22 °C. Then, 20 µL RNase H reaction mix that containing 25 U RNase H (NEB, M0297S) and 1× RNase H buffer were added into the mixture and incubated at 37 °C for 30 min and the mixture was placed on ice. We added 10 U DNase I (NEB, M0303S) and 1× DNase I buffer into the mixture to remove additional DNA. The mixture was incubated at 37 °C for 30 min and was placed on ice. We purified the ribo-minus RNA using acidic phenol-chloroform extraction. [2.2 × RNA Clean beads (Vazyme)] Library preparation was performed using ABclonal Fast-RNA-seq Lib Prep Kit V2 (RK20306) following the manuscript’s protocol, and the samples were sequenced using Novaseq 6000 platform with 150 bp paired-end mode.
For RNA-seq analysis, low-quality bases and adapter-containing reads were first trimmed from all RNA-seq raw data using a trim-galore package (version: 0.4.4) under default parameters, clean reads shorter than 35 bp were discarded. Trimmed reads were aligned to the mm10 genome using STAR
59 (version: 2.7.11) with parameters ‘--outFilterMultimapNmax 1 --outFilterMismatchNmax 999 --outFilterMismatchNoverLmax 0.02’. To analyze global TE expression in RNA-seq experiments, we use SQuIRE
60 (Software for Quantifying Interspersed Repeat Elements). SQuIRE quantifies expression at the subfamily level. It outputs read counts and fragments per kilobase transcript per million reads FPKM. Mapped reads on transcripts of mm10 Refseq gene annotation were counted using featurecounts
61 (version: 2.0.4), and the differential expression analysis was performed using DEseq2 packages
62 (version: 1.48.1) under the parameters: an absolute fold change > 1.5 and
P value < 0.05.
Nascent RNA-seq and data analysis
BMDM cells from WT and
Irg1−/− were seeded in the 10-cm plates with the same number. After LPS induction for 24 h, 4-thiouridine (4sU) was added to 500 mM at 15 min before cell collection. Total RNA was purified by TRIzol reagents (Thermo Scientific) and equal amount of total RNA (~40 µg) was first fragmented and then biotinylated, following by streptavidin beads capture as previously described
63. About 40 µg 4sU-labeled spike-in RNA from
Drosophila S2 cells were added to each sample reaction before RNA fragmentation. The mixture was fragmented with 0.2 M NaOH for 14 min and 0.2 M Tris-HCl (pH 6.8) was added to stop the fragmentation. Biotinylation of labeled RNA was carried out in 150 µL of biotinylating mix for 30 min (40 µg fragmented RNA, 10 mM HEPES-KOH, pH 7.5, 1 mM EDTA, 0.167 mg/mL MTSEA-biotin), and was purified with pre-washed Streptavidin MagPoly Beads (Smart-Lifesciences) in the binding and wash buffer and mixed for 15 min at room temperature (100 mM Tris-HCl, pH 7.5, 1M NaCl, 10 mM EDTA, 0.05% Tween-20 in DEPC H
2O). The biotinylated RNA was washed in binding and wash buffer for 5 times and eluted with 100 mM DTT, and was purified with VAHTS RNA clean beads (Vazyme, N412-01). The captured RNA is then processed into RNA sequencing library using ABclonal Fast-RNA-seq Lib Prep Kit V2 (RK20306) following the manuscript’s protocol. Samples were sequenced using Novaseq 6000 platform with 150 bp paired-end mode.
Raw reads were trimmed by trim-galore package (version: 0.4.4) under default parameters, clean reads shorter than 35 bp were discarded. Trimmed reads were aligned to the mm10 genome using STAR
59(version: 2.7.11) with parameters ‘--outFilterMultimapNmax 1 --outFilterMismatchNmax 999 --outFilterMismatchNoverLmax 0.02’. Read counts on genes were calculated by featureCounts
61 (version: 2.0.4), and the differential expression analysis was performed using DEseq2 packages
62 (version: 1.48.1) under the parameters: an absolute fold change > 1.5 and
P value < 0.05. For TE expression analysis, the reads were mapped to the mouse genome mm10 using STAR
59 (version: 2.7.11) with parameters ‘--outFilterMultimapNmax 100 --winAnchorMultimapNmax 200’ and the counts for each TE family were calculated using the TEtranscripts
64 (version: 2.2.3). DESeq2 was used for data normalization and differential expression analysis with parameters: an absolute fold change > 1.5 and
P value < 0.05.
Poly(A)+ RNA isolation and m6A quantification by LC-MS
Poly(A)+ RNA was purified from 50‒100 μg of total RNA using Oligo(dT)25 magnetic beads (New England Biolabs). Poly(A)+ RNA was eluted in nuclease-free water and quantified. For m6A quantification, 100–200 ng of poly(A)+ RNA was digested into nucleosides by sequential treatment with Nuclease P1 (New England Biolabs) at 37 °C for 2 h in a 16-μL reaction volume, followed by Fast AP (Thermo Fisher) digestion at 37 °C for an additional 2 h in a 20-μL final volume. Digested products were diluted threefold with nuclease-free water and subjected to LC–MS analysis. RNA m6A levels were quantified by LC–MS/MS and normalized to total nucleoside content.
Cloning and protein expression
The gene fragment encoding human ALKBH5 (aa 71–298) was cloned into the pET-28a vector with an N-terminal 6× His tag at the N-terminus. The recombinant plasmid was transformed into Escherichia coli Rosetta (DE3) cells for protein expression. Protein expression was induced with 0.5 mM IPTG at 14 °C overnight. Cells were harvested by centrifugation at 5,000× g for 15 min at 4 °C. For protein purification, the cell pellet was resuspended in lysis buffer containing 20 mM MES (pH 6.5), 500 mM NaCl, 5 mM imidazole, and 5% (v/v) glycerol. After cell lysis, the lysate was clarified by centrifugation, and the supernatant was loaded onto a Ni-NTA affinity column for purification. Target proteins were eluted using imidazole and further purified by size-exclusion chromatography using a Superdex 75 pg 16/60 column equilibrated in buffer containing 20 mM MES (pH 6.5) and 150 mM NaCl. The final purified protein was concentrated to 12 mg/mL for downstream applications.
Protein crystallization and structure determination
To obtain the ALKBH5–ITA complex, purified ALKBH5 (aa 71–298, at 12 mg/mL) was incubated with ITA at a molar ratio of 1:15 on ice. Initial crystallization trials were carried out using the sitting-drop vapor diffusion method at 18 °C, by mixing 0.5 µL of the protein-ligand mixture with 0.5 µL of reservoir solution. Crystals of the ALKBH5–ITA complex were obtained from a reservoir solution containing 0.18 M ammonium citrate dibasic and 20% (w/v) polyethylene glycol 3350. To ensure adequate ITA binding, crystals were further soaked in the crystallization solution supplemented with 50 mM ITA for 20 h. Prior to data collection, crystals were transferred to a cryoprotectant solution consisting of the reservoir buffer supplemented with 20% (v/v) glycerol, then flash-cooled in liquid nitrogen.
Diffraction data for the ALKBH5–ITA complex were collected at beamline BL02U1 of the Shanghai Synchrotron Radiation Facility (SSRF) at 100 K. The data were processed using CCP4
65. The structure of ALKBH5 was determined by molecular replacement using the program PHASER
66 with the crystal structure of ALKBH5 (PDB ID: 4O61) as the search model
29. Structure refinement was performed using REFMAC
67 and model building was carried out with Coot
68. Statistics of the diffraction data, the structure refinement and the quality of the final structure models are summarized in Supplementary Table S6.
MST assay
MST was employed to assess the interaction between ALKBH5 and ITA. A total of 100 µL of purified ALKBH5 protein (5–10 µM) was mixed with 98 µL of 1× PBS-T buffer and 2 µL of fluorescent dye, and incubated in the dark for 30 min. The labeled sample was then centrifuged at high speed to remove aggregates. Labeling efficiency was assessed by measuring fluorescence intensity, with an optimal target of approximately 200–2,000 arbitrary units (a.u.). For preliminary binding assessment, the labeled protein was incubated with the highest concentration of ITA (500 µM) for 30 min at room temperature. A labeled protein sample without ITA was used as a negative control. Each sample was loaded into standard MST capillaries for measurement. For quantitative binding analysis, labeled protein (10 µM) was titrated against a serial dilution series of ITA. After incubation at room temperature, the samples were loaded into capillaries and analyzed by MST to determine the binding affinity (Kd).
In vitro ALKBH5 activity assay by dot-blot
Recombinant human ALKBH5 (0.2 μM), purified from E. coli, was incubated at room temperature for 1 h in a 10 μL reaction containing 50 mM HEPES-NaOH (pH 7.5), 50 μM 2-oxoglutarate (2-OG), 100 μM sodium ascorbate (VC), 50 μM Fe2+, 1 mM TCEP, 10 μM synthetic single-stranded m6A RNA [5′-Biotin-AUUGUGG(m6A)CUGCAGC-3′ sangon] and varying doses of pH-adjusted ITA (pH 7.0). After incubation, the reaction was diluted 5-fold with DEPC-treated water, and 1 μL was used for RNA dot blot analysis.
1 μL of RNA sample was spotted onto a positively charged nylon transfer membrane (GE Healthcare) and air-dried, followed by UV crosslinking at 1200 J/m2. The membrane was then blocked for 30 min at room temperature in RNase-free 3% BSA in TBST, and subsequently incubated overnight at 4 °C with anti-m6A antibody (1:3,000, Synaptic Systems). After washing, the membrane was incubated for 1 h at room temperature with HRP-conjugated goat anti-rabbit IgG secondary antibody (Proteintech), and signals were detected using enhanced chemiluminescence (ECL) substrate (Meilunbio).
In vitro ALKBH5 activity assay by LC-MS
After the in vitro demethylation reaction (10 μL), 10 μg of Proteinase K (Takara) was added and incubated at 50 °C for 10 min to inactivate ALKBH5. The sample was then diluted in pull-down buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% Triton X-100), and 5 μL of Dynabeads™ MyOne™ Streptavidin C1 (Thermo Fisher) was added for incubation at room temperature for 40 min. Beads were subsequently washed twice with pull-down buffer and twice with DEPC-treated water. After washing, the beads were heated at 95 °C for 3 min to inactivate residual Proteinase K. The samples were then subjected to digestion with Nuclease P1 (NEB) in a 16 μL reaction at 37 °C for 2 h, followed by treatment with Fast AP (Thermo Fisher) in a 20-μL reaction at 37 °C for an additional 2 h. Finally, the digested products were diluted threefold with DEPC-treated water and analyzed by LC-MS.
Quantification of intracellular metabolites by LC-MS
Cells were washed with ice-cold PBS and lysed in 80% methanol (pre-chilled to –80 °C) for intracellular metabolite extraction. Lysates were centrifuged at 15,000× g for 20 min at 4 °C. Supernatants were collected, vacuum-dried, and reconstituted in double-distilled water (ddH2O). Cell extracts were analyzed by ultrahigh performance liquid chromatograph (Acquity UPLC I-Class, Waters) coupled to a Triple Quadrupole Mass Spectrometer (Xevo TQ-XS, Waters). Cells were assumed to have a spherical shape, and intracellular metabolite concentration was calculated using external standard curves and taking into account cellular diameter (d, micrometers) and cell number using the following equation: [metabolite] = metabolite quantity (moles)/[(4/3000)π(d/2)3 × cell number]. The data were analyzed using MassLynx v4.2.
Immunofluorescence staining
Paraffin-embedded tissue sections were deparaffinized in xylene and rehydrated through a graded ethanol series. Antigen retrieval was performed in 10 mM sodium citrate buffer (pH 6.0) at 95 °C for 20 min, followed by cooling to room temperature. Sections were then permeabilized and blocked in 3% BSA in PBS containing 0.3% Triton X-100 for 30 min at room temperature. The sections were incubated overnight at 4 °C with the following primary antibodies: anti-HERVK (rabbit polyclonal, 1:200; Thermo Fisher), anti-CD68 (mouse monoclonal, 1:200; HuaBio), and anti-IRG1 (D6H2Y, rabbit monoclonal, 1:200; Cell Signaling Technology). After three washes in TBST, sections were incubated with Alexa Fluor 488- or 594-conjugated secondary antibodies (Invitrogen) for 1 h at room temperature in the dark. Nuclei were counterstained with DAPI (Sigma) for 5 min. Slides were mounted using antifade mounting medium and imaged with a Leica SP8 laser scanning confocal microscope (Leica Microsystems). Quantitative fluorescence intensity analysis was performed using Fiji (ImageJ).
Immunohistochemistry
Paraffin-embedded tissue sections were prepared as described for immunofluorescence staining above. Endogenous peroxidase activity was quenched with 3% hydrogen peroxide in methanol for 10 min. Sections were blocked with 3% BSA in PBS containing 0.3% Triton X-100 for 30 min at room temperature. Sections were incubated overnight at 4 °C with anti-α-smooth muscle actin (α-SMA) antibody (rabbit monclonal, 1:400; CST). After three washes in PBST, sections were incubated with HRP-conjugated secondary antibody (Recordbio) for 30 min at room temperature. Immunoreactivity was visualized using 3,3′-diaminobenzidine (DAB) substrate (Zsbio) and counterstained with hematoxylin. Slides were dehydrated, cleared, and mounted with a permanent mounting medium. Images were acquired using ZEISS Axioscope 5 microscope, and α-SMA-positive areas were quantified using Fiji (ImageJ).
ITA-biotin synthesis
ITA-biotin (CG016955), chemically identified as 2-methylene-4,18-dioxo-22-((3aS,4S,6aR)-2-oxohexahydro-1H-thieno[3,4-d]imidazol-4-yl)-5,8,11,14-tetraoxa-17-azadocosanoic acid, was synthesized via a three-step route.
To a solution of tert-butyl (2-(2-(2-(2-hydroxyethoxy) ethoxy) ethoxy) ethyl) carbamate (400 mg, 1.36 mmol, 1 eq) in MTBE (5 mL) was added 3-methylenedihydrofuran-2,5-dione (152.8 mg, 1.36 mmol, 1 eq), PCL (85 mg, 13.1 μmol, 0.01 eq) at 25 °C. The reaction mixture was stirred at 25 °C for 16 h. LCMS showed the reaction was completed. The reaction mixture was filtered, and the filtrate was concentrated under reduced pressure to give the desired intermediate 2,2-dimethyl-20-methylene-4,18-dioxo-3,8,11,14,17-pentaoxa-5-azahenicosan-21-oic acid (536.3 mg, yield: 97%) as a colorless oil. The crude was used next step without further purification. MS (ESI) m/z = 404.20 [M-H]－.
To a solution of 2,2-dimethyl-20-methylene-4,18-dioxo-3,8,11,14,17-pentaoxa-5-azahenicosan-21-oic acid (536 mg, 1.32 mmol, 1.0 eq) in dioxane (3 mL) was added HCl/dioxane (5 mL) at 25 °C. The reaction mixture was stirred at 25 °C for 2 h. The reaction mixture was concentrated under reduced pressure to give the 1-amino-15-methylene-13-oxo-3,6,9,12-tetraoxahexadecan-16-oic acid (398.6 mg, yield: 99%) as a colorless oil. The crude was used next step without further purification. MS (ESI) m/z = 304.20 [M-H]－.
To a solution of 1-amino-15-methylene-13-oxo-3,6,9,12-tetraoxahexadecan-16-oic acid (150 mg, 516.7 μmol, 1 eq) in DMF (3 mL) was added 2,5-dioxopyrrolidin-1-yl 5-((3aS,4S,6aR)-2-oxohexahydro-1H-thieno[3,4-d]imidazol-4-yl)pentanoate (176.4 mg, 516.7 μmol, 1 eq), TEA (156.8 mg, 1.55 mmol, 3 eq) at 25 °C. The reaction mixture was stirred at 25 °C for 12 h. LC/MS showed the reaction was completed. The reaction mixture was adjusted to pH = 1 with 4 M HCl/dioxane, the reaction mixture was concentrated under reduced pressure to give a residue. The residue was purified by preparative HPLC (0.1% TFA in H2O: acetonitrile = 95:5 to 50:50) to give the desired product (CG016955) (37.6 mg, yield: 14%) as a colorless oil. 1H NMR (400 MHz, DMSO-d6) δ 12.52 (br s, 1H), 7.83 (s, 1H), 6.42 (br s, 2H), 6.17 (d, J = 1.5 Hz, 1H), 5.78 (d, J = 1.0 Hz, 1H), 4.36 – 4.26 (m, 1H), 4.17 – 4.10 (m, 3H), 3.62 – 3.51 (m, 10H), 3.40 (t, J = 5.9 Hz, 2H), 3.33 (s, 2H), 3.22 – 3.15 (m, 2H), 3.13 – 3.05 (m, 1H), 2.86 – 2.79 (m, 1H), 2.58 (d, J = 12.2 Hz, 1H), 2.07 (t, J = 7.5 Hz, 2H), 1.69 – 1.21 (m, 6H). MS (ESI) m/z = 532.3 [M+H]+.
ITA-biotin pull-down assay
Cells were rinsed once with ice-cold PBS and lysed in NP-40 lysis buffer (300 mM NaCl, 50 mM Tris-HCl, pH 7.5, 0.5% NP-40, with protease inhibitor cocktail; 1 mL per 60-mm dish) for 1 h at 4 °C. Lysates were cleared by centrifugation at 13,000 rpm for 15 min at 4 °C and incubated with ITA-biotin (100 µg mL−1) for 3 h at 4 °C. Streptavidin beads were added and rotated for another 3 h at 4 °C. Beads were washed three times with wash buffer (20 mM NaH2PO4, 0.15 M NaCl, 0.5% NP-40, pH 7.4, with protease inhibitors) and eluted with 10 µL 0.1 M glycine-HCl (pH 2.5) for 10 min at 4 °C. Eluates were analyzed by LC-MS or western blotting.
RIP
BMDMs (1.5 × 107 cells) were rinsed once with ice-cold PBS and lysed in 1 mL RIP lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM DTT, 0.6% Igepal CA-630, supplemented with 40 U mL−1 RNasin (Yeasen) and protease inhibitor cocktail (TaegetMol)) for 20 min at 4 °C. Lysates were clarified by centrifugation at 13,000 rpm for 15 min at 4 °C, and 50 µL of the supernatant was collected as input. The remaining lysates were incubated with 5 µg anti-Ythdf2 antibody (Proteintech) or control IgG overnight at 4 °C with rotation. Subsequently, 50 µL protein A/G agarose beads (Yeasen) were added and incubated for an additional 2 h at 4 °C. Beads were washed three times with RIP lysis buffer, and the immunoprecipitated RNA was extracted using TRIzol reagent for subsequent qPCR analysis.
m6A RIP assay
Fragmented RNA (1 µg) was incubated in a 100 µL immunoprecipitation reaction containing 1 µg anti-m
6A antibody (Synaptic Systems), 40 U mL
−1 RNasin (Yeasen), 10 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 0.1% NP-40 for 4 h at 4 °C with rotation. Subsequently, 5 µL Dynabeads protein A/G (Thermo Fisher Scientific) were added and incubated for an additional 2 h at 4 °C. Beads were then washed three times with IP buffer (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, and 0.1% NP-40). Finally, RNA was extracted from the beads using TRIzol reagent for downstream analysis. For MeRIP-seq, library preparation was performed as described above. Trimmed reads were aligned to the mm10 genome using STAR
59 (version: 2.7.11). Reads overlapping with
IAPEz elements were extracted and mapped to the full-length
IAPEz consensus sequence using Bowtie 2
69. Coverage along the consensus sequence was quantified using bedtools
70 and normalized to the total number of mapped reads.
Western blot analysis
For western blotting, cells were directly lysed in SDS loading buffer and proteins were separated by SDS-PAGE, then transferred onto nitrocellulose membranes (GE Healthcare). Membranes were blocked with 5% non-fat milk in TBST for 1 h at room temperature, followed by overnight incubation at 4 °C with primary antibodies. After washing, membranes were incubated with HRP-conjugated secondary antibodies (Proteintech) for 1 h at room temperature. Protein bands were visualized using ECL reagents (Meilunbio) and imaged with the LAS-4000 imaging system. HSP90 or Tubulin was used as a loading control.
Antibodies
The antibodies used in this study are summarized below: rabbit monoclonal anti-IRG1 antibody (CST, 77510s); rabbit polyclonal anti-ERVK antibody (ThermoFisher, PA5-49515); mouse monoclonal anti-CD68 antibody (Huabio, HA601115); mouse monoclonal anti-Flag (HRP) antibody (Sigma A8592); rabbit polyclonal anti-m6A antibody (Synaptic System, 202 003); rabbit monoclonal anti-phospho-Sting (Ser365) antibody (CST, 72971); rabbit monoclonal anti-Sting antibody (CST, 13647); rabbit monoclonal anti-phospho-Tbk1 (Ser172) antibody (CST, 5483); rabbit monoclonal anti-Tbk1 antibody (CST, 3013); rabbit anti-Tubulin antibody (CST, 2144); rabbit polyclonal anti-Alkbh5 antibody (Sigma, HPA007196-25UL); rabbit anti-Hsp90 antibody (CST, 4874); rabbit monoclonal anti-cGas antibody (CST, 15102); rabbit anti-α-Smooth Muscle Actin antibody (CST, 19245); rabbit anti-Ythdf2 antibody (Proteintech, 24744-1-AP); mouse anti-Mavs antibody (Santa, sc-365334).
Fluorescence in situ hybridization (FISH)
BMDMs cultured in 35 mm glass-bottom dishes (Cellvis) were used for both RNA and DNA FISH. For RNA-FISH, cells were fixed with 4% paraformaldehyde for 15 min at room temperature, followed by permeabilization with 0.2% Triton X-100 in PBS for 10 min. After discarding the permeabilization buffer, cells were washed 3 times with 1% BSA in PBS to prepare for hybridization. For DNA-FISH, cells were fixed with 4% paraformaldehyde for 20 min, then incubated in 75% ethanol for 24 h. Cells were subsequently treated with 0.2 N HCl for 10 min and digested with proteinase K (2 μg/mL in PBS) for 5 min at room temperature. After rinsing, cells were treated with 100 μg/mL RNase A (Thermo Scientific) in 2× SSC at 37 °C for 1 h, followed by denaturation in 70% formamide in 2× SSC at 72 °C for 10 min. Dehydration was performed using a graded series of pre-chilled ethanol solutions. Hybridization was performed by adding 200 μL of hybridization buffer (20% formamide, 10% dextran sulfate, 100 μg/mL yeast tRNA, 100 μg/mL salmon sperm DNA, 2× SSC) containing 600 ng of Cy3-labeled IAP-specific probe. The sequence of the IAP-specific probe is provided in Supplementary Table S5. Samples were incubated at 37 °C overnight. The next day, cells were washed three times with pre-warmed wash buffer (20% formamide in 2× SSC) at 37 °C for 10 min each. Nuclear counterstaining was performed using DAPI. Images were acquired using a Leica SP8 confocal microscope with a 63× oil immersion objective.
UUO and drug treatment
UUO surgery was performed in 8-week-old mice under isoflurane anesthesia. The ureter was exposed via a flank incision and ligated at two distinct points. Mice were sacrificed 7 days post-surgery for tissue collection and molecular analysis. For ITA treatment, mice received intraperitoneal injections of neutralized ITA (500 mg/kg) once daily after surgery until the day of sacrifice. For RTi treatment, mice were administered Tenofovir disoproxil fumarate (100 mg/kg) and Emtricitabine (60 mg/kg) by oral gavage once daily. RTi treatment commenced two days prior to UUO surgery and was continued throughout the experimental period, as previously described
30.
FA-induced nephropathy model
FA-induced nephropathy was established in 8-week-old mice by a single intraperitoneal injection of FA (125 mg/kg) dissolved in 300 mM sodium bicarbonate. Mice were sacrificed at either day 3 or day 21 post-injection for histological and molecular analyses. For renal function assessment, blood and urine samples were collected on day 21 following FA injection. Serum blood urea nitrogen (BUN) levels were measured using a commercial assay kit (Solarbio), according to the manufacturer’s instructions. Meanwhile, urine samples were collected from the same mice to quantify albumin and creatinine levels using by commercial kits (Sangon Biotech), following the manufacturers’ protocols. The urinary albumin-to-creatinine ratio (ACR) was calculated as an index of renal injury.
Isolation of renal macrophages
Mouse kidneys were harvested, finely minced, and digested using the High Activity General Tissue Enzymatic Digestion Kit (RWD) following the manufacturer’s instructions. The resulting cell suspension was filtered through a 40 μm strainer, and red blood cells were lysed. Renal macrophages were then enriched using Anti-F4/80 MicroBeads UltraPure (Miltenyi Biotec), according to the manufacturer’s protocol.
scRNA-seq analysis
Single-cell RNA sequencing libraries were generated using the NovelCyto Single-Cell Analysis System (NovelBio Co., Ltd, China) according to the manufacturer’s instructions as previously described
22. Briefly, single-cell suspensions were captured with barcoded beads in microwells, followed by reverse transcription and whole-transcriptome amplification (WTA). Libraries were quantified using a Bioanalyzer 4200 (Agilent) and Qubit High Sensitivity DNA Assay (Thermo Fisher Scientific), and sequenced on a DNBSEQ-T7 platform (MGI) with 150-bp paired-end reads.
Cell Ranger 1.6 (10xgenomics.com) was used to process Chromium single cell 5’ RNA-seq output. Once the gene-cell data matrix was generated, cells with < 200 unique genes expressed genes were excluded. Only genes expressed in 3 or more cells were used for further analysis. Cells were also discarded if their mitochondrial gene percentages were over 5%, resulting in 22,271 genes across 82,180 cells The data were natural log transformed and normalized for scaling the sequencing depth to a total of 10,000 molecules per cell, followed by regressing-out the number of UMI using Seurat package (version 5.1.0). During the dimension reduction process, we first removed the heterogeneity among the samples using the CCAIntegration method. Then, we selected the Top 30 representative PCA components were selected as input for UMAP. Finally, we visualized the results using UMAP. Ultimately, all the cells were divided into 35 clusters. Using the classic cell markers
71, 35 clusters of cells were annotated into 14 major cell types. CellChat R package (version 1.5.0) was used to analyze the changes in communication between cells. All the analyses and visualizations were conducted in R (version 4.3.2). The cGAS–STING pathway activity score was calculated using AddModuleScore with cGAS-STING pathway-related genes. The score reflects the average expression of these genes, normalized against control gene sets matched for expression distribution.
To quantify transposable element (TE) expression in distinct cell populations, cell barcodes corresponding to each annotated cell cluster were extracted from the integrated Seurat object based on the clustering results. Reads associated with each cluster were isolated from the Cell Ranger-generated BAM files using the subset-bam utility, generating cluster-specific BAM files. These BAM files were initially converted into bulk RNA-seq–like datasets using the bedtools bamtofastq command. Subsequently, FASTQ files were generated from the cluster-specific BAM files using Samtools (version 1.3.1) and realigned to the mouse reference genome (GRCm39) using STAR (version 2.7.8) under bulk RNA-seq alignment settings to generate aligned BAM files. TE expression quantification was then performed on the realigned BAM files using SQuIRE with default parameters. The resulting TE expression matrices were imported into Seurat and mapped back to the corresponding cell clusters for downstream analyses and visualization.
DATA AVAILABILITY
Sequencing data are available at the Genome Sequence Archive (accession No. PRJCA066331 for human sequencing data, accession No. PRJCA048805 for all other sequencing data).
GLORI-seq(Irg1/Alkbh5DKO)
Ribominus RNA-seq (Human monocytes)
The Author(s) 2026. Published by Higher Education Press. This is an Open Access article distributed under the terms of the CC BY license (https://creativecommons.org/licenses/by/4.0/).
