Introduction

Cytosolic dsDNAs are recognized drivers of chronic inflammation in a broad range of human diseases1. A major pathway involved in detecting cytosolic dsDNA and triggering the subsequent activation of inflammatory responses, relies on the cyclic GMP-AMP (cGAMP) synthase (cGAS) pathogen recognition receptor2,3,4,5. The interaction of cytosolic dsDNAs with cGAS leads to the production of the 2’3’-cGAMP second messenger, which in turn interacts with the stimulator of interferon genes (STING) adaptor protein, promoting the recruitment and activation of transcription factors, such as the Interferon regulatory factor 3 (IRF3), that drive a transcriptional programme ultimately leading to the production of type I Interferons (IFN) and inflammatory cytokines2,3,4,5. Regulation of the cGAS-STING signalling axis at several levels has been reported, including STING degradation6, or regulators of cGAS catalytic activity[7](https://www.nature.com/articles/s41467-025-65713-z#ref-CR7 “Burleigh, K. et al. Human DNA-PK activates a STING-independent DNA sensing pathway. Sci. Immunol. 5, https://doi.org/10.1126/sciimmunol.aba4219

(2020).“),8. In contrast, although the detection of dsDNA by cGAS is a key rate-limiting step in the activation of cGAS-STING signalling, few regulators of this interaction have been identified to date9. Those include the Three-prime exonuclease 1 (TREX1), the ablation of which feeds cGAS-STING-dependent signalling through endogenous cytosolic DNA accumulation, and underlying the most severe form of Aicardi-Gouttière Syndrome (AGS)10. This highlights the importance of the regulation of the cytosolic DNA-cGAS interaction for the regulation of pro-inflammatory signalling.

Pathological activation of cGAS-STING signalling has been documented in a wide range of neurological disorders11, mostly highlighting a link between chronic STING signalling and neurodegeneration12,13,14. Similarly, chronic inflammatory responses and immunological dysfunction have been reported in the Rett syndrome neurodevelopmental disorder (RTT; OMIM 312750). RTT is a rare genetic disorder frequently associated with de novo mutations in the Methyl-CpG-binding protein 2 (MECP2) X-linked gene, giving rise to dysfunctional MeCP215. Neurological manifestations of RTT have been shown to be reversible16 and accompanied by immunological dysfunction as well as chronic low-grade inflammation17,18,19. Indeed, the absence of MeCP2 sensitizes immune cells to pro-inflammatory stimulation20. While inflammation has been proposed to be associated with RTT disease progression and severity21,22, there is no mechanism explaining the onset of chronic inflammation in RTT.

MeCP2 is a broadly expressed nuclear protein known to operate as a major transcription regulator23. MeCP2 was initially described to bind and repress the expression of methylated regions of the genome23,24, but the functions of MeCP2 in the regulation of the mammalian genome have expanded in recent years. Indeed, MeCP2 is now recognized to play context-dependent roles in gene expression and chromatin architecture regulation through DNA methylation-independent mechanisms25,26,27. MeCP2 was shown to associate with and repress genomic locations corresponding to endogenous retroelements, such as the Long Interspersed Nuclear Element-1 (LINE-1)28. Interestingly, increased LINE-1 activity has been shown to be sufficient to generate DNA substrates for cGAS-STING activation29.

Altogether, the current state-of-the-art suggests a link between MeCP2 and dsDNA-mediated activation of the cGAS-STING pathway. Thus, we here investigated the potential role of MeCP2 as a regulator of dsDNA-associated inflammatory responses.

Results

MeCP2 interacts with cytosolic dsDNA

Previous work has shown that several actors of cytosolic dsDNA recognition, including cGAS, are primarily nuclear in the absence of cytosolic dsDNA8,30, suggesting that dsDNA binders are capable of interacting with dsDNAs regardless of their subcellular localisation. We thus hypothesized that MeCP2 may also interact with cytosolic dsDNA. To test this hypothesis, we performed a series of DNA pulldown experiments.

First, whole cell extracts from wild-type (WT) mouse embryonic fibroblasts (MEF) were incubated with streptavidin beads alone, or streptavidin beads on which either 80nt-long 5’biotin-bearing ssDNA (b-ssDNA) or dsDNA (b-dsDNA) were immobilized (in vitro pulldowns), prior to assessment of bound proteins by Western blot (WB) (Fig. 1a). Both DNA probes are capable of inducing the expected cGAS-STING pathway activation, as characterized by increased cGas levels, Sting degradation, phosphorylating activation of Sting and Irf3 (Supplementary Fig. 1a) and increased expression of Interferon β (Ifnβ) and Interleukin 6 (Il6), as well as IFN response genes, such as the C-X-C motif chemokine ligand 10 (Cxcl10), 2’−5’-oligoadenylate synthetase 1 (Oas1) and cGas (Supplementary Fig. 1b). No significant modulation of Mecp2 transcriptional levels were observed (Supplementary Fig. 1b). In vitro DNA pulldowns performed in WT-MEF using these probes showed robust recruitment of cGas and MeCP2 to dsDNA (Fig. 1b). Congruent with previous work, MeCP2 and cGas interaction with ssDNA was less efficient when compared to that with dsDNA (Fig. 1b)31. Thus, MeCP2 preferentially interacts with immune-stimulatory dsDNA.

Fig. 1: MeCP2 interacts with cytosolic dsDNA.

a Experimental scheme for in vitro DNA pull-downs. b Whole cell extracts (WCE) from WT-MEF were incubated with streptavidin beads alone (mock) or with streptavidin beads-bound biotinylated dsDNA (b-dsDNA) or ssDNA (b-ssDNA) prior to in vitro DNA pull-down. Input and eluates were analyzed by Western blot (WB) using the indicated antibodies. c As in a, except that WCE from WT-RAW264.7 were used. d In vitro pulldown was performed as in c, except that beads-bound biotinylated dsRNA were also used. e In vitro pulldown was performed as in b, except that two different b-dsDNA were used. f Experimental scheme for in-cell DNA pulldowns. g WT-MEF were transfected with b-dsDNA or not (mock) before WCE preparation and streptavidin-affinity pull-down. Input and eluates were analyzed by WB using the indicated antibodies. h As in g, except that WCE were from WT-RAW264.7 cells transfected or not with b-dsDNA. i Immunofluorescence analysis was conducted on WT-MEF transfected or not with dsDNA for 6 h using anti-MeCP2 antibody, anti-dsDNA antibody and DAPI nuclear staining. BF, bright field. Images are representative of 3 independent experiments. Scale bar: 10 µm, except for Zoom: 5 µm. j Violin plots show the % MeCP2 intensity in the cytosol in experiments performed as in i (n = 123 for Mock and for 120 dsDNA-transfected cells). k Pearson’s correlation coefficient was calculated on the cytosolic dsDNA and MeCP2 signals in WT-MEF treated as in j. l Immunofluorescence analysis of WT-MEF transfected or not with dsDNA and ssDNA for 6 h was performed using anti-MeCP2 antibody (enhanced signal or not) and DAPI nuclear staining. BF, bright field. Scale bars: 10 µm; Scale bars for Zoomed images: 5 µm. Images are representative of 3 independent experiments. m Violin plots show the % MeCP2 intensity in the cytosol and nucleus in experiments performed as in l; n = 105 cells per condition. WB and images are representative of at least 3 independent experiments. Significance was assessed using two-sided Student t-test. ns: non-significant. *P < 0.05, **P < 0.01, and ****P < 0.0001. Source data are provided as a Source Data file.

We next wished to verify if the interaction of MeCP2 with dsDNA can be recapitulated in other cell types. To this aim, we performed in vitro DNA pulldowns using whole cell extracts from the WT-RAW264.7 murine myeloid cell line using bead-immobilized b-ssDNA and b-dsDNA. Both MeCP2 and cGas were recovered as dsDNA binders in this assay, while neither cGas nor MeCP2 was recovered as bound to ssDNA (Fig. 1c). To further assess the specificity of the binding of MeCP2 to dsDNA, we performed similar in vitro binding assays using bead-immobilized biotinylated dsRNA. We found that MeCP2 does not bind to dsRNA (Fig. 1d). Finally, we analysed whether the binding of MeCP2 to dsDNA may be sequence-specific by using another 80nt-long dsDNA synthetic probe. In vitro pulldown performed using whole cell extracts from WT-RAW264.7 showed a similar level of recruitment of MeCP2 and cGas to the 2 tested probes (Fig. 1e). Together, these experiments suggest that the ability of MeCP2 to interact with immune-stimulatory dsDNA is a conserved mechanism across cell types.

Next, to assess whether the recruitment of MeCP2 to dsDNA occurs in cells, WT-MEF were transfected or not with b-dsDNA for 6 hours, a time point at which immunofluorescence analyses showed that transfected dsDNA probes are prominently cytosolic (Supplementary Fig. 1c). Whole cell extracts were prepared and used in pulldowns using streptavidin affinity beads (in-cell pulldown, Fig. 1f). WB analysis of pulled-down material showed that MeCP2 and cGas interact with dsDNA in cells (Fig. 1g). Similar results were obtained in WT-RAW264.7 (Fig. 1h) and in the THP-1 human myeloid cell line (Supplementary Fig. 1d) when in-cell pulldowns were performed 6 hours after transfection of b-dsDNA. These data thus show that MeCP2 can interact dsDNA in MEF, as well as in immune murine and human cell lines. These data further suggest that the interaction takes place in the cytosol.

To assess where the interaction between MeCP2 and dsDNA probes takes place, we performed in vitro DNA pulldowns using cytosolic and nuclear extracts from WT-MEF (Supplementary Fig. 1e). WB analyses showed an enrichment of MeCP2 bound to dsDNA in both fractions (Supplementary Fig. 1f), supporting an interaction of cytosolic MeCP2 with dsDNA. In addition, we performed immunofluorescence assays where WT-MEF were challenged with dsDNA prior to staining using MeCP2 and dsDNA-specific antibodies. We observed that dsDNA challenge led to MeCP2-specific staining in the cytosol (Fig. 1i, j), with a significant overlap with the dsDNA signal (Fig. 1k). To assess whether MeCP2 cytosolic staining is specific to dsDNA as suggested by Fig. 1b–e, we next challenged WT-MEF with ssDNA or dsDNA prior to analysis of MeCP2 subcellular localization. Immunofluorescence analyses showed an increase of MeCP2 cytosolic levels, coupled to a decrease in MeCP2 nuclear levels upon challenge with dsDNA, and not with ssDNA (Fig. 1l, m). Transfection of dsRNA did not lead to an increase of MeCP2cytosolic staining (Supplementary Fig. 1g) while MeCP2 cytosolic staining was observed regardless of the dsDNA sequence used (Supplementary Fig. 1h). Finally, dsDNA transfection in WT-RAW264.7 also led to MeCP2 cytosolic staining (Supplementary Fig. 1i).

Altogether, these data support that the presence of MeCP2 in the cytosol is triggered by the presence of cytosolic dsDNA.

MeCP2 is actively exported from the nucleus in the presence of cytosolic dsDNA

MeCP2 is mostly known for its nuclear localization and functions32. Thus, its presence in the cytosol is surprising. This led us to question whether dsDNA stimulation may trigger MeCP2 export from the nucleus.

To test whether increased cytosolic levels of MeCP2 result from active export from the nuclear compartment, we performed time course analyses of MeCP2 subcellular localisation following dsDNA stimulation for 3, 6, and 16 hours in WT-MEF. Immunofluorescence analyses showed that dsDNA stimulation led to increased cytosolic MeCP2 staining, accompanied by a decrease of MeCP2 nuclear staining (Fig. 2a, b). To control for the specificity of the cytosolic Mecp2 staining, similar experiments were performed in MEF expressing a MeCP2-targeting gRNA (MEFgMecp2) (supplementary Fig. 2a, b), as well as in co-cultures of control gRNA expressing MEF and MEFgMecp2 (supplementary Fig. 2c), showing that dsDNA stimulation induced MeCP2 cytosolic foci formation only in MeCP2 proficient cells.

Fig. 2: Challenge with dsDNA triggers MeCP2 export.

a WT-MEF were stimulated with dsDNA for 3, 6, and 16 h prior to immunofluorescence analyses using an MeCP2 targeting antibody and DAPI nuclear staining. Scale bars: 10 µm. b Violin plots show the number of MeCP2 foci intensity in the cytosol and in the nucleus in experiments performed as in a; n > 50 cells per condition. c Immunofluorescence analysis was performed on WT-MEF treated or not with 20 nM of Leptomycin B (LMB) prior to dsDNA transfection for 3 h using anti-MeCP2 antibody (enhanced signal or not) and DAPI nuclear staining. Scale bars: 10 µm; Scale bar for Zoomed images: 5 µm. Images are representative of two independent experiments. d Violin plots show the % of MeCP2 intensity in the cytosol and nucleus in cells treated as in c; n = 95 cells per condition. e Immunofluorescence analysis was conducted on WT-MEF and MEFcGas−/− cells transfected or not with dsDNA for 6 h using anti-MeCP2 antibody, anti-dsDNA antibody and DAPI nuclear staining. BF, bright field. Images are representative of 3 independent experiments. Scale bar: 20 µm. f Immunofluorescence analysis was performed on WT-MEF and MEFTrex1-/- using anti-MeCP2 antibody and DAPI nuclear staining. Scale bars: 20 µm. Images are representative of two independent experiments. g Violin plots show the % of MeCP2 intensity in the cytosol and nucleus of cells treated as in f, n = 97 cells per condition. Significance was assessed using a two-sided Student t-test. Source data are provided as a Source Data file.

Next, we performed subcellular fractionation experiments where the cytosolic and nuclear soluble fractions were isolated following challenge with dsDNA. WB analyses showed an enrichment of MeCP2 in cytosolic fractions of WT-MEFs (Supplementary Fig. 2d). Similar subcellular fractionation was conducted in WT-RAW264.7, showing that dsDNA challenge induces increased MeCP2 signal in the cytosolic fraction (Supplementary Fig. 2e). Next, we performed immunofluorescence analyses of MeCP2 subcellular localization in WT-MEF upon treatment with the Leptomycin B nuclear export inhibitor prior to stimulation or not with dsDNA. Leptomycin B treatment led to an efficient block of nuclear-cytosolic transport, as shown by immunofluorescence of RAN Binding Protein 1 (Ranbp1) (Supplementary Fig. 2f). As expected, Leptomycin B treatment also abolished dsDNA-induced cGas export (Supplementary Fig. 2g)30. We found that Leptomycin B treatment reduced the levels of MeCP2 in the cytosol following dsDNA stimulation (Fig. 2c, d). Conversely, Leptomycin B treatment led to increased retention of MeCP2 in the nucleus following dsDNA challenge (Fig. 2c, d). These data support that cytosolic MeCP2 accumulation upon dsDNA challenge is driven by active export of MeCP2.

That MeCP2 is not displaced in the cytosol following ssDNA or dsRNA stimulations (Fig. 1l and Supplementary Fig. 1g) suggests that inflammatory signalling does not drive MeCP2 export. To confirm this observation, we assessed MeCP2 subcellular localization in a context where dsDNA-associated inflammatory responses are abrogated. We thus challenged WT-MEF and cGas-knockout MEFs (MEFcGas-/-) with dsDNA. Immunofluorescence analyses of MeCP2 subcellular localization in response to dsDNA stimulation showed that MeCP2 is exported to the cytosol regardless of cGas expression (Fig. 2e). These data confirm that MeCP2 export is triggered by the presence of dsDNA in the cytosol and is not a response to cGas-Sting signalling.

Finally, we questioned whether cytosolic accumulation of endogenous DNA may trigger MeCP2 export. To this aim, we analyzed MeCP2 subcellular localization in MEFs harbouring an invalidating mutation in the Trex1 gene, leading to chronic accumulation of cytosolic DNAs33, notably owing to aberrant activity of the long interspersed element-1 (LINE-1) endogenous retroelements. Immunofluorescence analyses of control and Trex1-deficient MEFs (MEFTrex1-/-) showed the presence of MeCP2 in the cytosol (Fig. 2f, g). Furthermore, treatment with the Tenofovir reverse transcriptase inhibitor, known to inhibit LINE-1 activity34, led to decreased cytosolic dsDNA, cGas (Supplementary Fig. 2h) and MeCP2 stainings (Supplementary Fig. 2i, j). Therefore, the presence of endogenous cytosolic dsDNA is sufficient to trigger MeCP2 export.

MeCP2 and cGas interact with the same dsDNA moieties in the cytosol

Experiments performed in Fig.1 and Supplementary Fig. 1 show that MeCP2 and cGas are both capable of binding cytosolic dsDNAs. We thus questioned whether they may bind the same dsDNA moieties.

To this aim, MEFs engineered to stably express an enhanced green fluorescence protein (EGFP)-tagged cGas allele (MEFEGFP-cGas) were transfected or not with dsDNA prior to immunofluorescence analyses of MeCP2 and EGFP-cGas co-localization using Pearson’s correlation coefficient. This showed that dsDNA transfection increased the colocalization between cGas and MeCP2 in the cytosol (Fig. 3a, b), suggesting a tripartite interaction between dsDNA, cGas and MeCP2.

Fig. 3: Absence of MeCP2 increases cGas interaction with cytosolic dsDNA.

a Immunofluorescence analysis was conducted on MEF stably expressing a EGFP-cGAS construct (MEFEGFP-cGAS) transfected or not with dsDNA for 6 h using anti-MeCP2 antibody and DAPI nuclear staining. Images are representative of 3 independent experiments. Scale bar: 20 µm; Scale bar for Zoom: 10 µm. (Right) Graph presents mean ± SEM. b Pearson’s correlation coefficient values for co-localization of cGas and MeCP2. p values were determined by Student’s t test. ****p < 0.0001; n = 12. c WT-MEF were transfected or not for 10 min, 30 min, 1 h, 3 h or 6 h with b-dsDNA before whole-cell extract preparation and pull-down using streptavidin-affinity beads. Input and eluates were analyzed by WB using the indicated antibodies. d WT-MEF were transfected or not with dsDNA before whole-cell extract preparation and immunoprecipitation using control IgG or a MeCP2-specific antibody. Input and immunoprecipitated material were analyzed by WB using the indicated antibodies. e MeCP2 knockout RAW264.7 were engineered to stably express FLAG-tagged MeCP2 prior to stimulation with dsDNA for 1, 3, 6 and 16 h. Whole cell extracts were subjected to FLAG immunoprecipitation prior to analysis of inputs and eluates by WB using the indicated antibodies. f Whole cell extracts prepared from WT-MEF or MEFcGas-/- cells were incubated with streptavidin beads alone or with streptavidin bead-bound b-dsDNA prior to pull-down. Input and eluates were analyzed by WB using the indicated antibodies. g Whole cell extracts prepared from MEF expressing a control non-targeting gRNA (MEFgCTRL) or a MeCP2-targeting gRNA (MEFgMecp2) were incubated with streptavidin beads alone or with streptavidin bead-bound b-dsDNA prior to pull-down. Input and eluates were analyzed by WB using the indicated antibodies. All WB are representative of 3 independent experiments. Source data are provided as a Source Data file.

We next interrogated the dynamics of the interaction of MeCP2 and cGas with cytosolic dsDNA. To this aim, WT-MEFs were transfected with b-dsDNA for up to 6 hours prior to whole cell extraction and in-cell DNA pulldowns using streptavidin affinity beads. WB analyses showed MeCP2 and cGas interaction with dsDNA as early as 30 min post transfection (Fig. 3c). This suggests similar interaction dynamics of MeCP2 and cGas with dsDNA. This was further tested by performing MeCP2 immunoprecipitation in WT-MEFs, transfected or not with dsDNA prior to assessment of cGas co-immunoprecipitation. We found that dsDNA transfection led to co-immunoprecipitation of cGas with MeCP2, as compared to mock transfection (Fig. 3d). Next, RAW264.7 expressing a MeCP2-targeting gRNA (RAW264.7 gMecp2) were engineered to stably express a FLAG-tagged MeCP2allele (FLAG-MeCP2) (Supplementary Fig. 3a), prior to transfection for 1, 3, 6 or 16 hours with dsDNA and FLAG immunoprecipitation. WB analyses showed that the interaction between MeCP2and cGas increased over the time course (Fig. 3e). As a control, we monitored levels of Isg15, which increased in input material, but did not co-immunoprecipitate with FLAG-MeCP2 (Fig. 3e), supporting the specificity of cGas recruitment. Thus, when taken together, these data support that the presence of cytosolic dsDNA triggers the formation of MeCP2, cGas and dsDNA-containing complexes.

The recruitment of cGas and MeCP2 to the same dsDNA molecules in the cytosol raises the possibility that they may compete for interaction. To test this, we first used WT-MEF and MEFcGas-/- to perform in vitro DNA pulldowns. WB analyses showed that the absence of cGas enhanced the recruitment of MeCP2 to dsDNA (Fig. 3f and Supplementary Fig. 3b). Conversely, we assessed whether MeCP2 regulates cGas recruitment to dsDNA. To this aim, we used MEFs expressing control (MEFgCTRL) or MeCP2-targeting gRNAs (MEFgMecp2) to perform in vitro dsDNA pulldowns. We found that reduced levels of MeCP2led to increased cGas recruitment to dsDNA (Fig. 3g). Finally, we performed in vitro pulldowns using Sting knockout MEFs. We found that the absence of Sting did not modify MeCP2 nor cGas recruitment dsDNA (Supplementary Fig. 3c). Together, these data show that MeCP2 and cGas can regulate their respective recruitment to dsDNAs, further supporting that cGas and MeCP2 are recruited to the same dsDNA molecules.

MeCP2 inhibits cGas-dependent Type I Interferon responses

Our data show that the absence of MeCP2 leads to an increase of cGas recruitment to cytosolic dsDNA (Fig. 3), suggesting that the absence of MeCP2 may promote enhanced dsDNA-induced cGas-dependent signalling. While it was previously reported that the absence of MeCP2 correlated with increased inflammatory responses17,18,19,20,21,22, a role of MeCP2 in regulating cGas- or dsDNA-associated inflammatory responses was not reported.

To assess whether MeCP2 may regulate cGas activity, we challenged MEFgCTRL and MEFgMecp2 (Supplementary Fig. 4a) with dsDNA prior to assessment of cGas-dependent pathway activation. Immunofluorescence analyses of cGas subcellular localization in these cell lines showed that dsDNA stimulation enhanced cGas levels in the cytosol in MEFgMecp2 as compared to MEFgCTRL (Fig. 4a). Assessment of intracellular 2’3’-cGAMP levels in MEFgCTRL and MEFgMecp2 cell lines showed that low levels of MeCP2 led to higher levels of dsDNA-associated 2’3’-cGAMP (Fig. 4b). Analyses of the expression of genes classically associated with cGas activation, such as Ifnβ, Il6, Cxcl10 and C-C motif chemokine ligand 5 (Ccl5) (Fig. 4c), and of inflammatory cytokine production (Fig. 4d, e) showed that MeCP2 knockout led to their increased levels. Similarly, RAW264.7gMecp2 (Supplementary Fig. 4b) presented heightened Ifnβ, Isg15, Cxcl10 and cGas upon dsDNA stimulation, as compared to control cells (RAW264.7gCTRL) (Supplementary Fig. 4c), accompanied by upregulation of pTbk1 and pSting, attesting to activation of Sting, as well as increased pStat1 levels, that attests to the production of bioactive IFNs (Supplementary Fig. 4d). Importantly, the inflammatory responses witnessed upon dsDNA stimulation in cells expressing a MeCP2-targeting gRNA is decreased by treatment with the H-151 Sting inhibitor (Fig. 4f), suggesting that the witnessed inflammatory signature is, at least in part, Sting-dependent. Together, these data support the conservation of an inhibitory role of MeCP2 in different cell lines.

Fig. 4: Absence of MeCP2 primes cGas-dependent Inflammatory responses.

a Immunofluorescence analysis was conducted on MEFgCTRL or MEFgMecp2 after challenge or not with dsDNA for 6 h, using anti-cGas antibody and DAPI nuclear staining. Scale bar: 20 µm. Images are representative of 3 independent experiments. b Intracellular 2’3’-cGAMP levels were measured by ELISA following transfection or not with the dsDNA of MEFgCTRL or MEFgMecp2. Graph presents fold increase 2’3’-cGAMP levels from 5 independent experiments. c MEFgCTRL or MEFgMecp2 were challenged or not with dsDNA for 6 h prior to gene expression analyses. Graph presents mean (±SEM) Ifnβ, Cxcl10, Il6 and Isg15 mRNA levels (n = 3 independent experiments). d MEFgCTRL or MEFgMecp2 were challenged with dsDNA for 24 h prior to collection of the supernatant and analyses using proteome profiler antibody arrays. The heat map presents data obtained from duplicate measurements. e representative images of arrays from d. f MEFgCTRL or MEFgMecp2 were challenged or not with dsDNA for 6 h in the presence or not of the H-151 Sting inhibitor. Graphs present mean (±SEM) Ifnβ, Il6, Ccl5, Oas1, and Ifit2 mRNA levels and mean centroid analysis (n = 3 independent experiments). One-way ANOVA with Sidak’s multiple comparison. g MEF overexpressing WT-MeCP2 (eMeCP2) or not (Empty) were transfected or not with dsDNA for 6 h prior to analysis of Ifnβ, Il6, Cxcl10, and Mecp2 mRNA levels. Graphs present mean (±SEM) from 3 independent experiments. h MEFgCTRL or MEFgMecp2 were challenged or not with poly(I:C) for 6 h prior to gene expression analyses. Graphs present mean (±SEM) Ifnβ, Oasl1, Cxcl10, and Isg15 mRNA levels (n = 3 independent experiments). Significance was assessed using a two-sided Student t-test, except when otherwise stated. Source data are provided as a Source Data file.

Next, we wished to confirm the direct role of MeCP2 in negatively regulating dsDNA-associated inflammatory responses. Using 2 different gRNAs targeting MeCP2 (Supplementary Fig. Se), we confirmed that absence of MeCP2 enhanced inflammatory gene expression (Supplementary Fig. 4f). Conversely, when MEFs were transfected with either control (Empty) or MeCP2-expressing vector (eMeCP2) (Supplementary Fig. S4g), we found that overexpression of MeCP2 prior to dsDNA stimulation led to decreased expression of Ifnβ, Il6, and Cxcl10 as compared to Empty vector-expressing cells (Fig. 4g). We next questioned the specificity of the inhibitory role of MeCP2 and challenged MEFgCTRL and MEFgMecp2 with the poly(I:C) synthetic RNA agonist that stimulates cGas-independent inflammatory pathways. Absence of MeCP2 did not modulate the expression of Inflammatory genes following poly(I:C) stimulation (Fig. 4h). This confirms the role of MeCP2 as a negative regulator of cGas-associated inflammatory responses.

Taken together, these data (Figs. 3 and 4) show that the absence of MeCP2 leads to increased cGas recruitment to cytosolic dsDNA in cells and subsequent activation of cGas-dependent inflammatory responses.

Absence of MeCP2 enforces an antiviral state

That MeCP2 inhibits dsDNA-associated inflammatory responses suggests that the absence of MeCP2 may lead to chronic low-grade type I IFN responses. The latter are known to lead to the establishment of protective antiviral states. This led us to question the interplay between viral infections and MeCP2 (Fig. 5a).

Fig. 5: Absence of MeCP2 enforces an antiviral state.

a Experimental scheme for b–h. b WT-MEF infected or not with a GFP-expressing HSV-1 KOS64 at 0.5 or 5 multiplicity of infection (MOI) were analyzed by immunofluorescence using anti-MeCP2 antibody and DAPI nuclear staining. Images are representative of 3 independent experiments. Scale bar: 50 µm. c Percent MeCP2 intensity in the cytosol in cells infected as in (b); n = 30 cells per condition. d As in c, except that cells were infected with the EGFP-expressing HSV-1 McKrae at 1 MOI; n > 50 cells per condition. e MEFgCTRL or MEFgMecp2 were infected or not with 5 MOI of HSV-1-KOS64 for 6 and 16 h prior to analysis of expression of indicated genes. Graphs present the mean (±SEM) of 3 independent experiments. f Mean (±SEM) plaques per cm2 after 72 h of infection of MEFgCTRL or MEFgMecp2 with 1 MOI of HSV-1 KOS64. 8 replicates, representative of 3 independent experiments. g WT-MEF were infected with conditioned media from HSV-1 KOS64-infected MEFgCTRL or MEFgMecp2. The graph presents the mean (±SEM) percentage GFP-positive cells in recipient cells (n = 3 independent experiments). h Representative images of cells in g. Scale bar: 400 µm. i Gene Set Enrichment Analysis (GSEA) was performed looking for Biological Process on DESeq2 results (log2foldchange > 0.01. GSEA p-value cut off = 0.05) from RNAseq data from RAW264.7gCTRL or RAW264.7gMecp2. j Gene expression analysis was performed in the livers of male Mecp2*+/y* and Mecp2*-/y* mice. Graph presents mean (±SEM) fold increase gene expression in Mecp2*-/y* mice as compared to Mecp2*+/y*; n = 4 mice per group. k Mean centroid expression was calculated on the expression of genes indicated in j (n = 4 mice per group). l Significant DEGs involved in inflammation, interferon-beta, virus response, STING, and innate immunity between hippocampi from control and Mecp2-silenced mice. X-axis: the False Discovery Rate (FDR). Gene symbols are reported for genes relevant to innate immunity. m Example Gene Set upregulated in patients with severe RTT symptoms. Significance was assessed using a two-sided Student T-test except for h, j and k. Source data are provided as a Source Data file.

We first questioned whether the delivery of DNA in the cytosol through infection with a DNA virus would induce accumulation of MeCP2 in the cytosol, as witnessed upon dsDNA transfection (Fig. 2). We used two molecular clones of the Herpes simplex virus 1 (HSV-1), namely KOS64 and McKrae, harbouring dsDNA genomes. We found that infection of WT-MEF with HSV-1 led to the presence of MeCP2 in the cytosol (Fig. 5b-d and Supplementary Fig. 5a). Similar experiments were conducted using a molecular clone of the Vesicular stomatitis virus (VSV), a virus harbouring an ssRNA genome. In contrast to what was visualized following infection with HSV-1, we found that infection with VSV did not lead to cytosolic staining of MeCP2 (Supplementary Fig. 5b). These experiments, combined with data in Fig. 1, show that the delivery of dsDNA in the cytosol is necessary and sufficient to induce MeCP2 nuclear export.

Next, we assessed whether the presence or absence of MeCP2 can influence inflammatory responses following infection with HSV-1. To this aim, MEFgCTRL and MEFgMecp2 were infected with HSV-1-KOS64 prior to assessment of the expression of Ifnβ at 6 hours and of antiviral IFN-stimulated genes such as Cxcl10, as 2’-5’-Oligoadenylate Synthetase-Like (Oasl1), 2’-5’-oligoadenylate synthetase 3 (Oas3) and Il6 at 16 hours post infection. Absence of MeCP2 led to increased expression of those genes following HSV-1 infection (Fig. 5e). Thus, lower levels of MeCP2 lead to increased expression of inflammatory genes following infection with HSV-1, confirming the inhibitory impact of MeCP2 on cGas-associated signalling (Fig. 4).

We next assessed the capacity of HSV-1 to infect and replicate in MeCP2-deficient and proficient cells. To this aim, MEFgCTRL and MEFgMecp2 were infected with HSV-1-KOS64 for 72 hours prior to assessment of plaque formation. Decreased levels of MeCP2 led to decreased plaque formation (Fig. 5f and Supplementary Fig. 5c), attesting to reduced ability of HSV-1 to infect these cells as compared to control cells. In contrast, when similar experiments were conducted using VSV, we found that the absence of MeCP2 did not significantly alter viral infection (Supplementary Fig. 5d). Next, conditioned media from HSV-1-KOS64-infected MEFgCTRL and MEFgMecp2 were used to infect WT-MEF cells prior to quantification of infected cells. We found that infection of WT-MEF with the supernatant collected from MEFgMecp2 led to a decreased number of infected cells as compared to supernatant collected from MEFgCTRL (Fig. 5g, h). Altogether, these data show that the absence of MeCP2 fosters an antiviral state that hinders infection by HSV-1. This further suggests that the absence of MeCP2 is sufficient to foster an antiviral state that is efficient towards DNA virus infection.

Since the presence of a type I IFN antiviral signature was not previously reported in MeCP2 deficiency, we finally assessed whether MeCP2 knockout is sufficient to promote the expression of a type I IFN signature. We first performed RNA sequencing (RNAseq) on RAW264.7gCTRL and RAW264.7gMecp2. Gene Set Enrichment Analysis (GSEA) was conducted on genes upregulated in RAW264.7gMecp2 as compared to RAW264.7gCTRL, revealing an upregulation of processes related to stress response, immune response and biological interactions between organisms, including responses to external and biotic stimuli (Fig. 5i). We next assessed the presence of such an antiviral signature in the liver of mouse models of MeCP2 deficiency (Mecp2**y/-) as compared to WT littermates (Mecp2**y/+). Analyses of the expression levels of type I IFN response genes and antiviral genes (Oas1b, Oas2, Oasl2, Oas3, Ccl5, Schlafen 5 (Slfn5), MX dynamin-like GTPase 2 (Mx2), and Isg15 showed a tendency for increased expression (Fig. 5j). Mean centroid analysis of those antiviral genes showed a significant increase in MeCP2-deficient animals (Fig. 5k), confirming the presence of an antiviral type I IFN response in MeCP2 deficiency. Similarly, metanalyses were conducted on public datasets35, generated from the hippocampus of adult mice where MeCP2 expression was disrupted by RNA interference. Analyses focused on the Gene Ontology (GO) terms with the keywords “inflammation”, “interferon-beta”, “viral infection”, “sting”, and “innate immunity” in the description. We found that the differentially expressed genes (DEGs) comprise genes involved in the aforementioned groups (Fig. 5l and Supplementary Fig. 5e, f).

Finally, we wished to identify whether such an innate immune response signature could be observed in RTT patients presenting with MeCP2 deficiency. To this aim, we considered the samples taken from patients with RTT with different severity of symptoms36. GSEA showed enrichment of pathways related to response to IFN alph