Abstract
To determine the role for mutations of MECP2 in Rett Syndrome, we generated isogenic lines of human iPSCs (hiPSCs), neural progenitor cells (NPCs), and neurons from patient fibroblasts with and without MECP2 expression in an attempt to recapitulate disease phenotypes in vitro. Molecular profiling uncovered a reduction of 5hmC, increased expression of subtelomeric genes including TERRA (a long non-coding RNA), and shortening of telomeres in the absence of MECP2 in hiPSCs, NPCs, and Neurons. Neurons made without MECP2 show signs of stress, including induction of gamma-H2aX, p53, and senescence, which are typical molecular responses to telomere shortening. The induction of p53 appeared to affect dendritic branching in Rett neurons, as p53 inhibition restored dendritic complexity. Examination of Rett patient brains uncovered similar molecular phenotypes suggesting that our disease-in-a-dish model yielded insights into human Rett Syndrome patient phenotypes and point towards a role for MECP2 in regulating telomere function.
Introduction
Rett Syndrome is a disease associated with loss of function mutations in the gene MECP2, which was originally identified as encoding a methylated DNA binding protein1–3. Patient symptoms include microcephaly, intellectual disability, facial dysmorphia, and seizure activity4,5. Studies in murine models recapitulate many of the patient phenotypes and have recently identified a role for MECP2 particularly in inhibitory neurons6–9. These studies demonstrated that loss of MECP2 can lead to defects in transcription10–12, dendritic branching13, nuclear size3, and AKT signaling14.
MECP2 is known to bind methlylated DNA (both 5mC and 5hmC)1,2,15,16, and the loss of MECP2 was shown to affect 5hmC levels in at least one portion of the murine brain. MECP2 has also been described as a transcription factor with specific targets10,11,13, and more broadly as either a transcriptional activator14 or repressor17–20. However, despite decades of research on MECP2, it is still unclear how mutations in this protein lead to patient symptoms3,14,21–23. To confirm findings made in other models and further study these in a human system, some have turned to modeling Rett Syndrome in vitro by taking advantage of Disease in a dish approaches. This involves making hiPSCs from patient somatic cells, or using genome engineering to introduce mutations into WT human pluripotent stem cells. In either case, the pluripotent stem cells created are then differentiated toward the neural lineage, and then comparisons can be made between cells that express MECP2 or lack it.
Some of these studies have even taken advantage of isogenically controlled lines to identify both transcriptional and electrophysiological effects of loss of MECP2 in human in vitro models14,24. In the current study, we also sought to mitigate the effect of genetic background and variability of differentiation by taking advantage of several isogenic lines of hiPSCs that either express the WT allele or the mutant allele leading to cells that express or lack MECP225. This allowed for detailed molecular analyses of hiPSCs, NPCs and neurons with and without MECP2 under the same genetic background. In addition, several lines were made and analyzed in each category to avoid variance in differentiation potential amongst isogenic lines. Furthermore, isogenic lines were made from two independent patients with different mutations to highlight only those phenotypes associated with loss of MECP2 expression and not genetic background or variance in hPSC differentiation. Finally, we validated many of these findings using siRNA silencing of MECP2 in WT cells of a distinct genetic background.
In comparing multiple lines of cells, it is clear from our data that loss of MECP2 leads to profound molecular alterations specifically towards the ends of chromosomes due to a decrease of 5-hydroxymethylation, induction of subtelomeric gene expression, and shortening of telomeres. The telomere defects that arise in neurons appear to be related to defects in dendritic branching that are a hallmark of the patient disease. Together, these results define a heretofore unappreciated role for MECP2 in molecular regulation towards the ends of chromosomes.
Results
An isogenic model of Rett Syndrome in vitro
To determine how loss of MECP2 expression leads to defects in the nervous system we generated a disease-in-a-dish model using iPSCs. Cognizant of the fact that differentiation from hPSCs is highly variable across individual lines, culture conditions, and time, we developed an isogenic model to study Rett Syndrome in vitro to remove the confound of genetic background25. Because female patients with Rett Syndrome are usually heterozygous for mutant alleles of MECP2, fibroblasts isolated from these patients display a mosaic pattern where roughly half the cells express either the mutant or WT allele. This is shown in Figure 1A, where fibroblasts isolated from two patients with distinct mutant alleles of MECP2 (R982 and R567) showed that roughly half the cells express MECP2 while the other half lacked detectable amounts of this protein. One of these mutant alleles is predicted to lead to a premature stop codon, while the other leads to failed transcriptional termination. Reprogramming to iPSCs using a small set of transcription factors has been shown to happen at the clonal level, such that individual reprogramming events in single fibroblasts generate isolated hiPSC clones26. Therefore, reprogramming of mosaic fibroblast cultures from two different patients generated single hiPSC clones that either expressed MECP2 protein or lacked it (Fig 1B) (Method described in a previous study27). In addition, our work and that of others has shown that under standard conditions, the inactive X chromosome in human fibroblasts does not reactivate upon reprogramming to the pluripotent state25,27,28, which is distinct from murine reprogramming29.
Thus, we were able to create multiple lines of hiPSCs with and without MECP2 from individual patients and thereby control for differences in genetic background (shown in Fig 1B are clones made from patient 982, clones from 567 look similar). The hiPSCs generated from fibroblasts of both patients appeared to be unaffected by the lack of MECP2, expressed all appropriate markers, and successfully generated teratomas upon injection into the testes of immunocompromised mice, consistent with previous hiPSC models for loss of MECP2 (Fig 1-figure supplement 1)14,30–32. Lack of MECP2 in patient-derived cells and specificity of antibody was also confirmed by western blot (Fig 1-figure supplement 2A). Importantly, we never observed reactivation of the silenced X chromosome that would have resulted in re-expression of the WT allele of MECP2 in any cultures regardless of differentiation status or passage. This is consistent with previous data showing that despite evidence for erosion of isolated portions of the silenced X chromosome33, the portion containing the MECP2 locus was not affected by reprogramming or differentiation.
As Rett Syndrome primarily afflicts the nervous system and MECP2 is most highly expressed in neurons, we first generated neural progenitor cells from all of the hiPSCs lines following standard protocols34. Across at least two lines per patient with and without MECP2, we measured the rate of neuralization, the morphology of NPCs, and expression of typical marker genes. We were unable to detect consistent differences in these properties between multiple clones of both WT and MECP2-lines derived from both patients (Fig 1C and 1-figure supplement 2B). Furthermore, the growth rate of NPCs with and without MECP2 was not consistently different in NPCs made from either patient (Fig 1-figure supplement 2C). Next, the NPCs were further differentiated by a non-directed differentiation approach that yields both neurons and glia (growth factor withdrawal35) (Fig 1D). Both the neurons and glia made from NPCs adopted typical morphologies regardless of MECP2 expression, and all NPCs from both patients produced neurons and glia at the same rate (Fig 1-figure supplement 2D and figure supplement 2E).
Previous studies have also shown that loss of MECP2 in neurons can lead to a decrease in AKT signaling14. A similar pattern was observed here in mutant neurons generated from Rett patient hiPSCs as measured by phosphorylation of AKT and S6, while hiPSCs themselves did not seem to be affected by loss of MECP2 (Fig 1E). Dendritic complexity has been shown extensively to be reliant on MECP2 expression in various models of Rett Syndrome, and we found a statistically significant decrease in complexity in neurons made in the absence of MECP2 by Sholl assay (Fig 1F). In addition, we observed qualitative differences in basic neuronal morphology between WT and mutant neurons, where the neurons lacking MECP2 had shorter, thicker processes, and their soma was not as well defined.
Loss of MECP2 leads to disruption of hydroxymethylation of DNA
Because MECP2 is a well-established methylated DNA binding protein particularly for 5-hydroxymethylcytosine (5hmC)16, we analyzed patterns of this mark across the genome in the presence or absence of MECP2 with Methylation-dependent Immunoprecipitation (MEDIP)41–44 in hiPSCs. We used a stringent criterion to identify differentially hydroxymethylated regions (DhmRs), whereby the indicated regions had to differ by 0.2 per million reads per base pair (Fig 2A). We still observed a large number of differentially hydroxymethylated regions (DhmRs) due to the loss of MECP2 in two clones each from two independent patients (982.15 and 982.17 vs 982.16 and 982.18; 567.24 and 567.26 vs 567.25).
The loss of MECP2 led to many more hypomethylated regions than hypermethylated regions, a strong bias that indicated that MECP2 somehow promotes or stabilizes hydroxymethylation (Fig 2B). This hypomethylation is more clearly identified by plotting the Delta Methylation between the WT and MUT clones from both patients. Both patients showed a dramatic shift towards loss of methylation across both clones (Fig 2B). Mapping 5hmC-DhmRs relative to genomic features indicated a de-enrichment away from intergenic regions and enrichment at coding exons (Fig 2C indicated by *). When mapping the 5hmC-DhmRs across chromosomal locations, they were highly enriched towards the ends of chromosomes (Fig 2D). In addition, the effect of loss of MECP2 on 5hmC levels was strong enough to be observed by immunostaining in hiPSCs made from both patients (Fig 2E and F).
Loss of MECP2 affects the transcriptome of neurons
It has been suggested that loss of MECP2 only affects gene expression in neurons as opposed to the hPSCs and NPCs from which they were derived14. Coupled with the fact that 5hmC levels appear to be disturbed in MECP2 null hiPSCs, we sought to determine whether gene expression was affected in hiPSCs, NPCs or neurons in this patient derived in vitro model. We therefore proceeded with RNA seq (>120 million reads per sample) of hiPSC, NPC and interneuron cultures. With such sequencing depth, it was possible to analyze the RNA-seq reads for the known mutations present in the patients from which these lines were made (Fig 3-figure supplement 1). This analysis demonstrated that each line studied expressed strictly either the WT or mutant allele of MECP2, and that XCI status was unchanged even after extensive differentiation to neurons.
To optimize the search for molecular effects of loss of MECP2 in hiPSCs, NPCs or neurons, we generated more defined neuronal cultures by following the newly established 3i (three inhibitor) method to create populations of human interneuron progenitors (Fig 3-figure supplement 1A) and interneurons (Fig 3-figure supplement 1B)45. Interneurons are particularly relevant in the study of Rett Syndrome as interneuron-specific deletion of Mecp2 in mice recapitulates many of the disease symptoms6,8,46,47. We validated the purity and quality of differentiation at each step by immunostaining for markers typical of particular cell types (SOX2, SOX1 and NESTIN as well as FOXG1 and NKX2.1 for NPCs; and Tuj1, MAP2 and GABA for interneurons) in both WT and MUT cultures followed by quantification (not shown). We first assessed whether interneurons lacking MECP2 also showed diminished dendritic branching. In fact, in patient-derived interneurons made by 3i, defects in dendritic branching as measured by the number of endpoints were clearly observed (Fig 3A).
First, we quantified the expression level of MECP2 in WT cells across these three stages of development and found that the average RPKM was 3.1 for hiPSCs, 4.3 for NPCs, and 7.75 for interneuron cultures. This is consistent with consensus that MECP2 is enriched in neuronal cells, but also demonstrates that it could potentially be relevant to hiPSC and NPC physiology as well. However, high stringency analyses (FDR <0.05) of the RNA-seq data yielded very few gene expression changes due to loss of MECP2 in hiPSCs or NPCs derived from Rett patients (Fig 3B), consistent with Li et al14. On the other hand, interneuron cultures made from patient 982 showed many gene expression changes when comparing two individual WT and MUT clones (Fig 3B). Gene ontology analysis uncovered many neuronal physiology-related pathways were downregulated due to loss of MECP2 in neurons, while genes associated with extracellular remodeling and cell migration appeared to be induced (Fig 3C).
We then mapped the interneuron DEGs according to chromosomal location and found an interesting pattern whereby genes that many of the upregulated genes in the absence of MECP2 were enriched towards the ends of chromosomes (Fig 3D). Moreover, the same pattern emerged in analysis of low stringency DEGs (p value <0.05) from hiPSCs and NPCs lacking MECP2 (Fig S4C and S4D) for the upregulated genes. We did not observe this pattern for the downregulated genes in hiPSCs, NPCs and neurons. These data suggested that MECP2 could play a role in gene regulation particularly at the ends of chromosomes, and was consistent with the pattern observed for hypomethylation of 5hmC in the absence of MECP2. Finally, there was also a small, but statistically significant overlap of 5hmC-DhMRs with DEGs in hiPSCs suggesting a link between the two (Fig 3E).
We validated a number of the subtelomeric gene expression changes induced by loss of MECP2 by RT-PCR in independent preparations of hiPSCs, NPCs and Neurons (Fig 4A). Many of the subtelomeric genes upregulated are typically not expressed at all in hiPSCs, NPCs or neurons, thus the loss of MECP2 led to an aberrant expression pattern as opposed to a reinforcement or suppression of a typical pattern in these cell types. Furthermore, the subtelomeric genes that were induced in the absence of MECP2 were still typically present at less than 1 RPKM, clouding the potential physiological consequence of the induction of these genes.
To determine whether these effects were specific to the genetic background of the cells used or whether defects in reprogramming to the pluripotent state in the absence of MECP2 affected the downstream gene expression pattern, we silenced MECP2 in WT NPCs and assessed gene expression patterns in this context. Several different siRNA targeting oligos were assayed for their ability to silence MECP2 by RT-PCR, western blot, and immunostaining (Fig 4-figure supplement 1). Silencing of MECP2 in either WT-NPCs derived from other pluripotent stem cells or WT-NPCs derived from 15 week old fetal Medial Ganglionic Eminence (MGE) brain tissue, a source of cortical interneuron progenitors35, led to strong induction of expression of subtelomeric genes (Fig 4B and C).
Recently, it was discovered that a long non-coding RNA (lncRNA) is also transcribed from the subtelomeric domain into the telomeric sequence itself48–52. This lncRNA is both known to be induced by telomere shortening and to potentially negatively regulate telomere length by competing for telomere priming within telomerase51–53. TERRA transcripts are difficult to detect by RNA-seq because they contain mostly telomeric repeat sequences. This also makes it difficult to design PCR primers that are specific to a single TERRA transcript. However, we used established RT-PCR primers51,52 to show that, similar to many subtelomeric genes, several TERRAs were strongly induced in the absence of MECP2 in isogenic derivatives (hiPSCs, NPCs, interneurons) (Fig 4D). In addition TERRA was induced in WT-NPCs derived from pluripotent stem cells or from tissue (Fig 4D).
Loss of TET activity phenocopies loss of MECP2
To assess the possibility that MECP2 regulation of 5hmC levels is linked to the regulation of subtelomeric gene expression, we targeted TET enzymes by siRNA. TET enzymes convert 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (5hmC), and their deletion or downregulation is known to severely diminish levels of 5hmC44,54. According to RNA-seq, the three TET enzymes are expressed similarly in both WT and MUT patient neurons (Average RPKM: TET1, 4.1; TET2, 1.6; TET3, 3.6). RT-PCR demonstrated the ability to silence TET 1, 2, and 3 isoforms using a combination of siRNA oligos in neurons (Fig 4E). Assaying for subtelomeric genes, including TERRA transcripts demonstrated that TET inhibition led to strong increases in not only subtelomeric coding genes, but also TERRA transcripts (Fig 4F), in a similar manner as in loss of MECP2, suggesting a link between 5hmC and subtelomeric gene expression. Together, these data confirm that loss of MECP2 can lead to induction of subtelomeric genes, including TERRA, and that this dysregulation could be due to an effect on 5-hydroxymethylation. Left unclear is how the loss of MECP2 leads to changes in 5hmC levels.
Loss of MECP2 leads to telomeric abnormalities
One of the most established functions for TERRAs is their ability to interfere with telomerase function by acting as a competitive inhibitor for telomere priming55, leading to shortened telomeres. In addition, others have shown that decreased 5hmC in murine embryonic stem cells can also lead to telomere shortening56, and, as shown here, loss of MECP2 led to decreased 5hmC in hiPSCs (Fig 2). Therefore, we attempted to determine the physiological consequence of TERRA induction in the absence of MECP2. qPCR for telomere length in fact showed that cells without MECP2 showed shorter telomeres, regardless of their stage of differentiation (Fig 5A) (a complete list of cell lines used for these analyses is provided in Supplemental Table 1). NPCs with transient siRNA knockdown of MECP2 also showed shorter telomeres, indicating that this effect was not simply due to defects during reprogramming or differential expansion of cell lines (Fig 5A). Quantitative fluorescence in situ hybridization (qFISH) was used to determine telomere length at the single cell level. qFISH demonstrated that NPCs without MECP2 showed significantly shorter telomeres relative to centromeric regions (Fig 5B). To further validate these findings we performed southern blot with a telomere probe on NPCs with and without MECP2 and found telomere shortening in cells lacking MECP2 (Fig 5C). These data point towards telomere erosion in the absence of MECP2.
Telomere dysfunction is known to be present in some cancers, and also in cells driven to senescence due to telomere shortening, but this process has yet to be implicated in Rett Syndrome etiology57–60. Telomere dysfunction is characterized by short telomeres, induction of PML, gammaH2aX, and p5361. We assayed for evidence of telomere dysfunction by immunostaining for gamma-H2aX (Fig 6A) and PML (Fig 6B) in NPCs. WT NPCs with silencing of MECP2 by siRNA and neurons lacking MECP2 also showed clear induction of these marks (data not shown), consistent with telomere dysfunction induced by the absence of MECP2.
Induction of P53 and senescence pathways in the absence of MECP2
It is well established that shortening of telomeres puts significant stress on cells, which can lead to senescence or even apoptosis55,62–64. As Rett Syndrome is caused by neuronal defects specifically, we determined how neurons lacking MECP2 respond to telomere shortening at the molecular and physiological level. Cells under stress due to telomere shortening are known to induce p53, which can then activate various response pathways downstream such as DNA repair, senescence, and apoptosis65. Interestingly, p53 induction due to telomere shortening was previously shown to cause defects in dendritic branching64,66, which is also the dominant phenotype in Rett Syndrome. Immunostaining for p53 in neurons with and without MECP2 showed a strong increase in p53 protein in the absence of MECP2 (Fig 6C). p21, a transcriptional target gene of p53, was also induced in MECP2 null neurons at the protein level (Fig 6C). In addition, telomere shortening in NPCs due to overexpression of the Progerin allele, which is associated with accelerated aging also induced p53 expression (Fig 6-figure supplement 1E).
Because telomere shortening is known to also drive cellular senescence, we looked for signs of defective proliferation in vitro. While attempting to make clones of fibroblasts from patients with Rett syndrome, we repeatedly found that clones lacking MECP2 did not expand well after a passage (14 MECP2 null clones were created, none expanded), while clones expressing the WT allele expanded without problem (42 MECP2+ clones were created, and 4 out of 4 all expanded). To determine whether MECP2 null fibroblasts encounter senescence, we performed assays to detect endogenous beta-galactosidase, which is known to be a hallmark of this process67. Indeed, MECP2 null fibroblasts showed strong activity in this senescence assay (Fig 6D).
We did not encounter such difficulties with clonal expansion once hiPSCs or hiPSC-derived NPCs were made from patients, presumably because during reprogramming, telomerase is strongly induced to restore telomere length at least beyond the critical threshold68–73. In fact, our RNA-seq data showed that hiPSCs made from patients had very high expression of TERT, and NPCs still expressed moderate levels, while neurons did not express appreciable levels (average RPKM for TERT: hiPSC, 8.8; NPC, 1.6; neuron, 0.006). Importantly, the same endogenous galactosidase activity assay on interneurons showed a dramatic increase in senescence activity in neurons lacking MECP2 (Fig 6E). On the other hand, similar assays on NPCs lacking MECP2 did not show any induction of senescence (data not shown). Together, these data indicate that loss of MECP2 leads to the generation of neurons that show evidence of telomere dysfunction.
Probing RNA-seq data, we also found that MECP2 null interneuron cultures showed a strong increase in a group of genes that are known to be induced by senescent cells, known as the Senescence Associate Secretory Program (SASP). Fig 6-figure supplement 1A shows that SASP genes were strongly induced in MECP2 mutant neurons, providing further evidence of a senescence phenotype. These senescence phenotypes are also intriguing in light of the transcriptional data suggesting an increase in aging-related genes by gene ontology analysis (Fig 3C). The only previous report linking MECP2 loss to senescence was performed by partial silencing of this protein in mesenchymal stem cells, but the results were consistent with those shown here for patient derived MECP2 null fibroblasts74.
To demonstrate whether the induction of senescence and p53 observed here was due to telomere shortening as opposed to other molecular phenotypes due to loss of MECP2, we deliberately shortened telomeres in otherwise wildtype NPCs. We took advantage of the progerin allele of the Lamin-A gene. This truncated allele is similar to what is found in patients suffering from Progeria, a premature aging disorder typified by telomere shortening75,76. Induction of the progerin allele by lentiviral infection of cDNA in WT NPCs showed a significant telomere shortening as expected (Fig 6-figure supplement 1B). In addition, induction of the progerin allele caused an increase in expression of the same subtelomeric genes and TERRA transcripts that were induced by the loss of MECP2 (Fig 6-figure supplement 1C and D). This was presumably due to the Telomere Position Effect, whereby telomere shortening is known to lead to induction of subtelomeric gene expression77,78. Importantly, progerin expression also led to a strong induction of p53 expression (Fig 6-figure 1E), consistent with what was observed in MECP2 null neurons.
Blocking induction of P53 can rescue dendritic branching defects due to loss of MECP2
Previous evidence from a murine model of telomere shortening as a result of loss of telomerase complex (TERT) led to defects in dendritic branching, and this effect was strictly dependent on induction of p5364. A more recent study also showed that experimentally aging the neural lineage with telomerase inhibition led to neurons with signs of aging, including reduced dendritic branching79. Therefore, we posited that inhibition of P53 in MECP2 null neurons with shortened telomeres could potentially restore appropriate dendritic branching.
To determine whether blocking the action of P53 could improve dendritic branching in MECP2 null interneurons, we took advantage of Pifithrin-α, a potent inhibitor of P53 target gene activation80. Treatment of MECP2 null interneurons with Pifithrin-α showed evidence of p53 inhibition as measured by RT-PCR for GADD4565, a target gene important for DNA repair (Fig 6F). After 24-48 hours of p53 inhibition by Pifithrin-α, MECP2 null interneurons appeared to adopt an improved neuronal morphology typified by increased physical distinction between the soma and neurites, longer, thinner neurites, as well as increased dendritic branching as shown and quantified in Fig 6F. These data provide evidence that neurons with shortened telomeres due to loss of MECP2 respond by inducing P53 activity, which then inhibits the formation of complex neuronal processes. In summary, our in vitro model in human neurons suggests that loss of MECP2 leads to aberrant molecular regulation at the ends of chromosomes, leading to telomere shortening and a resulting induction of cell stress pathways such as p53 and senescence (Fig 6G).
Rett patient brains show evidence of telomeric dysfunction
To determine whether any of the phenotypes discovered in this in vitro model of Rett Syndrome have relevance to patients afflicted with the disease, we acquired tissue specimens from Rett patients and aged matched controls. We first quantified the degree of chimerism of female Rett patient neurons due to skewing of X chromosome inactivation to determine the relative ratio of neurons that express MECP2 versus those that did not. One of the Rett patient brains showed roughly 75% of its neurons lacking MECP2, while others appeared to have less than 25% MECP2 null neurons (Fig 7A). Southern blotting of the patient brain with 75% mutant neurons compared to an aged-matched control demonstrated that this Rett brain had shorter telomeres (Fig 7B). We then measured telomere length by PCR from genomic DNA isolated from small specimens of brain tissue from a group of Rett patients, and found that some aged matched Rett patient brains showed small decreases in average telomere length (Fig 7C), though not in every case (data not shown). Because of both the limited availability of Rett patient brains, and the variable chimerism of WT and MUT neurons within these Rett brains, perhaps it is not surprising that we were unable to detect trends across all brains analyzed using a method that cannot distinguish between WT and mutant cells. This chimerism, coupled with the known variability of telomere length across even normal individuals and even across brain regions and cell types precludes an accurate assessment of general telomere length differences in Rett brain until more samples become available for study.
On the other hand, we did find that the TERRA transcript was induced in nearly all Rett patient brains as measured by RT-PCR (Fig 7D), suggesting that all the Rett patients display dysregulation at the ends of their chromosomes. We did identify two Rett patient brains with a high proportion of MECP2 null neurons and subjected these to further investigation for signs of telomere dysfunction in situ. In patients 1815 and 5784, MECP2 null neurons showed a strong increase in both P53 and PML levels compared to adjacent neurons that expressed MECP2 (Fig 7E and F). This is consistent with the response to telomere shortening due to loss of MECP2 observed in vitro. These data are particularly intriguing in light of data showing that telomere shortening diminishes dendritic branching in various types of neurons and that this process can be dependent on p53 activity64,66.
Discussion
Taken together, these data demonstrate that loss of MECP2 leads to telomere shortening, which in neurons results in clear signs of stress such as H2aX induction, p53/p21 induction, and initiation of a senescence program, all of which suggest that neurons in Rett Syndrome could be in suboptimal health, leading to neurophysiological defects such as dendritic arborization13,22. Many of these phenotypes first observed in the in vitro model also appeared to be consistent with what could be observed in Rett patient brains, suggesting disease relevance for these findings.
It is curious that telomere defects have not been reported in previous models of Rett Syndrome. While one paper suggested that RNAi-mediated silencing of MECP2 could affect the telomeres of mesenchymal cells74, decades of work on Rett Syndrome have not uncovered a role for MECP2 in relation to telomeres in a wide variety of models such as various transgenic mouse line, human patient post-mortem analyses, in vitro human models. Our study certainly benefited from analyses of multiple isogenically controlled cells from two patients and from the single cell analyses of patients with both WT and MUT neurons in the same area of the brain. This allowed for high confidence comparisons without having to correct for genetic background, or differences in tissue preparation. In addition, the study of telomeres in MECP2 mutant mice could be hampered by the simple fact that telomeres from inbred mouse strains typically used for these studies are on average much longer than human telomeres. Therefore, it is possible that telomere shortening in murine models does not proceed to such an extent by which one would expect induction of p53 in murine models.
Patients with Rett Syndrome are typically characterized by a normal development at birth and subsequent failure to thrive leading to microcephaly and intellectual disability that develops with age. As a result, Rett Syndrome is thought to be caused by experience-dependent loss of neuronal function, which would correlate with data suggesting that MECP2 regulates activity dependent gene expression10,13,37. The microcephaly has been proposed to be a function of decreased nuclear size and dendritic arborization of affected neurons13,22. Could the telomere shortening induced senescence described here underlie patient phenotypes? Several studies have looked at the effects of telomere shortening specifically in the neural lineage and found consistently that shortened telomeres leads to upregulation of p53 and decreased dendritic arborization64,66,81, a phenotype widely described to afflict MECP2 null neurons in vitro and in vivo (Fig 1).
These results presented here raise the question of whether telomere defects could be common to the etiologies of other ID syndromes. The phenotypes described here show a striking similarity to those observed in hiPSCs and neural derivatives made from patients with Immunodeficiency, centromeric region instability, facial anomalies syndrome (ICF) Syndrome82–84. Two independent studies showed that ICF patient-derived hiPSCs displayed subtelomeric hypomethylation, induction of subtelomeric gene expression, TERRA induction and telomere shortening that was coupled to senescence of somatic derivatives such as fibroblasts. ICF Syndrome only partially overlaps with Rett Syndrome in terms of patient phenotypes, but is caused by mutations in DNMT3B, a de novo DNA methyltransferase85. These findings together are highly relevant as DNMT3B is a key de novo methyl transferase to create methylated DNA (5mC), which is the substrate for Tet oxigenases to create 5-hydroxmethylated DNA (5hmC). Recently, another study showed that deletion of Tet enzymes, which are critical to generate the 5hmC mark, led to shortened telomeres56,86. Together, these studies demonstrate that DNA hydroxymethylation is important in the regulation of telomere length, and our data suggest that MECP2 is potentially an important mediator of this effect.
ATRX-related syndrome shares many phenotypic features with Rett Syndrome (ID, seizures, and microcephaly), and the causative gene, ATRX, is known associate with MECP2 both genetically as well as biochemically87–92. In murine models of loss of ATRX, telomere shortening and reduced lifespan are observed93. The fact that the causative mutations of ICF, Rett and ATRX syndromes are in genes whose products are thought to interact in the regulation of DNA methylation and all possess telomeric defects suggests that the similarities of patient phenotypes could be the result of neuronal response to telomere shortening. This molecular overlap could even form the basis of novel therapeutic strategies that either reverse telomere shortening or block the response of the cell to telomere defects with agents such as Pifithrin, as shown in Fig 6F.
Considering the phenotypes of ICF derived cells, and those of Tet-deleted cells, it seems reasonable to suggest that telomere deficiency could be related to intellectual disability. In addition, subtelomeric and telomeric dysfunction has been implicated in up to 10% of all intellectual disability syndromes94–96. Another ID syndrome, Hoyeraal-Hreidarsson, is caused by mutations in RTEL1 (regulator of telomere elongation helicase 1), a factor that interacts with shelterin complex and is critical for telomere elongation94. These patients are characterized by low birth weight, microcephaly and immunological dysfunction. Therefore, mutations that specifically result in telomere shortening lead to disease phenotypes similar to those found in patients with loss of MECP2. As a result, we cannot exclude the possibility that telomere shortening during in utero development generates neurons that are less well equipped to deal with post-natal stimulation. Furthermore, our analysis of the single male Rett brain specimen available to us suggested that brains completely lacking MECP2 also had shorter telomeres.
Another possible interpretation of these data is that instead of a failure to mature, Rett Syndrome neurons instead show aspects of premature aging. The fact that MECP2 null neurons have shorter telomeres (Fig 5), show induction of aging related genes including p53 (Fig 3 and 6), and show senescence (Fig 6) are all consistent with this idea97. In addition, the fact that WT-NPCs and neurons transduced with the Progerin allele98, which is known to cause premature aging, show similar phenotypes as neurons lacking MECP2 is also consistent with this idea. On the other hand, while Rett patients suffer from a post-natal cognitive decline, and long term survivors show phenotypes associated with Parkinson’s disease99, the typical phenotypes presented in young female patients are not consistent with premature aging. Whether the physiological response to loss of MECP2 is truly akin to premature aging or whether patients suffer from the effects telomere dysfunction that is unrelated to aging is worthy of continued investigation.
Regardless, it is tempting to speculate that treatments that could relieve telomere dysfunction or abrogate the p53 mediated stress response could potentially ameliorate patient outcomes. Pifithrin-α has already been shown to be an effective treatment to restore neuronal function in murine models of injury or stroke100–102. Significant future effort will be devoted to determining both whether telomere dysfunction is a common trigger for ID Syndromes, and whether telomere restoration could potentially help patients.
Materials and Methods
Generation of isogenic Rett Syndrome iPSCs
Two primary fibroblast lines GM17567 (1461A>G in the gene encoding methyl-CpG binding protein 2 (MECP2)), and GM07982 (frameshift mutation, 705delG, in the gene encoding methyl-CpG binding protein 2 (MECP2)), from patients with Rett Syndrome were obtained from Coriell Cell Repositories. 1 x 105 fibroblasts were plated in a gelatin coated well of a 6 well plate in MEF media (DMEM/F12 + 10% FBS). After 8-12 hours, the cells were infected with reprogramming lentivirus that harbors polycystronic human Yamanaka factors (Oct4, Klf4, Sox2, cMyc) in DMEM medium containing 10ug/ml of polybrene and incubated overnight at 37°C in 5% CO2 incubator. The following day, the viral media was aspirated, replaced with MEF media and cultured for 3 additional days. Cells were re-plated on the 5th day onto irradiated MEFs in MEF media. On day 6, the culturing media was changed to human ES media containing DMEM/F12 supplemented with L-glutamine, nonessential amino acids (NEAA), penicillin-streptomycin, knockout serum replacement (Invitrogen), and 10 ng/ml basic FGF. Cells were cultured in hiPSC media until iPSC-like colonies were formed. Reprogrammed colonies were further identified by live immunofluorescence staining with TRA-1-81 (Chemicon) then mechanically isolated. Individual colonies were isolated and maintained for at least 2 passages before genotyping analysis. For early passages, the iPSCs were propagated mechanically, whereas collagenase was used for subsequent passaging. hiPSCs were cultured as described previously in accordance with the UCLA ESCRO.
Generation of teratomas
Generation of teratoma was previously described103. Briefly, a single incision was made in the peritoneal cavity of adult SCID mice and the testis was explanted through the incision site. Approximately 60,000 iPSC in a volume of 50 ml 0.5X Matrigel (BD) were transplanted into the testis using a 27-gauge needle. Four to six weeks after surgery, mice were euthanized and the tumors removed for histology. Surgery was performed following Institutional Approval for Appropriate Care and use of Laboratory animals by the UCLA Institutional Animal Care and Use Committee (Chancellor’s Animal Research Committee (ARC)).
Differentiation in vitro and analysis
Neural specification with neural rosette derivation, neuroprogenitor (NPC) purification, and further differentiation to neurons and glia were performed as described previously 34,104,105. Relative neuralization efficiency was analyzed by counting the number of neural rosette containing colonies over total number of iPSC colonies. 6-12 35 mm wells were analyzed over four separate experiments. The proliferation efficiency of NPCs was determined by at days 1, 3, and 5 by the total number of cells present in 35mm wells seeded at 200,000 cells on day 0. The cells were detached from the plates using accutase (Millipore) then total number of cells per well analyzed using Z1 Coulter particle counter (Beckman Coulter).
For spontaneous terminal neuronal differentiation by growth factor withdrawal, NPC cultures were subjected to growth factor withdrawal (removal of EGF and FGF) and cultured in basic medium (DMEMF12 + N2 + B27) with three quarter exchange of media every three days. Cells were cultured up to 20 weeks. Neural differentiation efficiency was analyzed four weeks after growth factor withdrawal by counting the number of cells positive for neuronal markers (MAP2 and Tuj1) over the total number of cells visualized by DAPI. NPCs were transfected with DCX-GFP reporter one day prior to differentiation using Lipofectamine 2000 (Invitrogen). Sholl analysis of DCX-GFP positive neuronal neuritis were also measured using ImageJ. All data values were presented as mean +/- SEM. Student’s t-tests were applied to data with two groups. ANOVA analyses were used for comparisons of data with greater than two groups.
For directed differentiation of interneurons, iPSCs were grown on plates coated with matrigel (Corning) until 80% confluency with mTeSR (Stem Cell Technologies). Cells were then treated with DMEM/F12 (GIBCO) containing NEAA (GIBCO), GlutaMAX (GIBCO), bovine serum albumin (Sigma-Aldrich), ß-mercaptoethanol (Sigma-Aldrich), N2 (GIBCO), B27 (GIBCO), SB431542 (10uM; Cayman Chemical), LDN-193189 (100nM; Cayman Chemical) and XAV939 (2uM; Cayman Chemical) later transitioning to the media containing sonic hedgehog (20ng/mL; R&D) and purmorphamine (1uM; Cayman Chemical) as previously described (Maroof et al., 2013). Cells were further differentiated into interneurons with neurobasal medium (GIBCO) containing N2 (GIBCO), B27 (GIBCO), ascorbic acid (Sigma-Aldrich), GlutaMAX (GIBCO), bovine serum albumin (Sigma-Aldrich), ß-mercaptoethanol (Sigma-Aldrich), neurotrophin-3 (10ng/mL; R&D), brain-derived neurotrophic factor (10ng/mL; R&D), and glial cell-derived neurotrophic factor (10ng/mL; R&D).
Neuronal activation
8 weeks in vitro differentiated neuronal culture were subjected depolarizing stimulation with 55mM of KCl in basic media for 0hr, 1hr, 5hr and 7hr then isolated for RNA analysis and coverslips fixed with 4% paraformaldehyde for immunostaining.
Western blot
Cells were lysed on ice with RIPA buffer (Pierce) that contains Halt Protease Inhibitor Cocktail (Thermo Fisher Scientific) and Halt Phosphatase Inhibitor Cocktail (Thermo Fisher Scientific). The total protein concentration was measured using BCA Protein Assay Kit (Thermo Fisher Scientific) following the manufacturer’s protocol. The lysates containing the equal amount of total protein were mixed with NuPAGE sample buffer (Invitrogen) with 5% mercaptoethanol and denatured at 95 °C for 10 min. Supernatant was electrophoresed onto NuPAGE 4-12% Bis-Tris Protein Gels (Invitrogen) using MOPS running buffer (Invitrogen). Gels were then electroblotted using semi-dry transfer apparatus onto a membrane. The membrane was blocked with 5% non-fat milk for 1 hr and incubated overnight with primary antibodies at 4°C. The next day the membrane was washed and incubated with HRP-conjugated secondary antibodies for 1 hr at room temperature. The membrane was then incubated with ECL Western Blotting Substrate and developed.
Immunofluorescence and image quantification
Cells on coverslips were washed with PBS, fixed in 4% paraformaldehyde for 15 min at room temperature, blocked for 1 hr at room temperature with 10% serum and 0.1% Triton-X-100, then incubated overnight at 4 °C with primary antibodies. Frozen tissue sections were thawed to room temperature, fixed in 4% paraformaldehyde for 15 min at room temperature, permeabilized with 0.2% Triton-X-100 for 15 min at room temperature and blocked with 10% serum at 4 °C overnight, followed by incubation with primary antibodies at 4 °C for 16-24 hr. Following primary antibody incubation, the coverslips were incubated with Alexa Fluor (Invitrogen) or Jackson immunoresearch secondary antibodies at room temperature for 1 hr. Cells were counterstained with DAPI and mounted in Prolong Gold (ThermoFisher). Antibodies used include the following: mouse anti-OCT3/4 (1:100, Santa Cruz Biotechnology Inc.), rabbit anti-SOX2 (1:300, Cell Signaling Technology), rabbit anti-Nanog (1:100, Cell Signaling Technology), mouse anti-Tra-1-81 (1:250, Chemicon), mouse anti-NESTIN (1:1000, Neuromics), chicken anti-MAP2 (1:2000, Biolegend), chicken anti-GFAP (1:2000, Abcam), rabbit anti-Tubulin β3 (1:500, Covance), mouse anti-p53 (1:500, Cell Signaling), rabbit anti-p21 (1:250, Santa Cruz), mouse anti-PML (1:100, Santa Cruz), mouse anti-phospho-Histone H2A.X (1:2000, EMD Millipore), rabbit anti-5hmc (1:100, Active Motif), rabbit anti MECP2 (1:1000, Diagenode), rabbit anti Foxg1 (1:1000, Abcam), and mouse anti NKX2.1 (1:300, Novocastra). Secondary antibodies conjugated with Alexa 488, 568, 594, 647 (1:500, Life Technologies) were used. Imaging was performed on Zeiss Axio Imager A1 or Zeiss LSM780 confocal microscope using a 40X or 63X objective on randomly selected cells. Mean intensity or a number of foci were quantified using ImageJ (http://rsb.info.nih.gov/ij/). At least 100 cells per condition were used for each independent experiment.
RT-qPCR
RNA from cultured cells was collected using the RNeasy Mini Kit from Qiagen according to the manufacturer’s instructions. The concentration and purity of RNA were measured using nanodrop spectrophotometers (Thermo Scientific). RNA with an A260/A280 ratio in between 1.8 and 2.0 as well as an A260/A230 ratio in between 2.0 and 2.2 was used. RNA was then reverse transcribed using the Super Script III First-Strand cDNA Synthesis kit with Random Hexamers (Invitrogen) according to the manufacturer’s instructions. Quantitative PCR was performed using SYBR Green master mix (Roche). Primers were used at a final concentration of 1 uM. Reactions were performed in duplicate and duplicate CT values were averaged and then used for standard ΔΔCT analysis. Expression levels were normalized to beta actin. See Supplementary Table 2 for qPCR primer sequences.
Data collection and statistical analysis
All the experimental data (RT-qPCR, qPCR assay for telomere length, immunostaining, ß-Galactosidase Senescence Assay) were presented as mean +/- SEM based on at least three biological replicates from independent experiments using the cell lines indicated in Supplementary Table 1. Student’s t tests were applied to data with two groups. ANOVA analyses were used for comparisons of data with greater than two groups. A p value < 0.05 was considered as statistically significant.
siRNA gene silencing
All knockdown experiments were performed using trilencer siRNAs (from OriGene Technologies) and RNAimax (ThermoFisher) in Opti-MEM media (ThermoFisher). Trilencers were used at a concentration of 20 nM. Transfection media was prepared and then 500,000 cells were plated on top of the transfection media in 6-well plates. The medium was changed to normal NPC media the next day and cells were collected for analysis at the time points indicated.
ß-Galactosidase Senescence Assay
ß-Galactosidase Senescence Assay was performed using the Senescence β-Galactosidase Staining Kit from Cell Signaling according to manufacturer’s instructions. Briefly, the cells were fixed on coverslips, incubated with X-gal overnight at 37°C, then mounted on glass slides and imaged using a brightfield microscope. The number of blue cells and number of total cells were quantified using the Cell Counter plugin in ImageJ.
Quantitative fluorescence in situ hybridization
Cells were fixed with 4% paraformaldehyde for 15 min at RT and permeabilized with 0.5% Triton X for 10 min at RT. After dehydration series of 80, 95, and 100% cold ethanol, cells were then treated with RNase (100ug/mL in 2xSSC) for 30 min at 37°C. After washing and another dehydration series, cells were denatured at 85°C for 15 min in the hybridization mix (70% DI formamide, 10 mM Tris-HCl, pH 7.5, 2xSSC, 0.1ug/mL salmon sperm DNA) containing PNA probes (TelC-FITC and Cent-Cy3, Panagene) at 1ug/mL and then incubated for 2 hr at RT in dark. After hybridization, cells were washed 3X for 5 min in 2xSSC/50% DI formamide, in 2xSSX with 0.1 Tween20, and in 1xSSC. Cells were then counterstained with DAPI and mounted with ProLong Gold (ThermoFisher).
Southern blot analysis of terminal restriction fragments (TRF)
Total genomic DNA was extracted using DNeasy Blood and Tissue Kit (Qiagen) following the manufacturer’s instruction. Fresh genomic DNA with high purity (an A260/A280 ratio of 1.8) was used for experiments. The integrity of genomic DNA was determined by gel electrophoresis. Southern blotting was carried using a TeloTTAGGG assay kit (Roche Applied Sciences) following the manufacturer’s protocol with some modifications (Kimura et al., 2010).
Quantitative PCR assay for average telomere length measurement
Total genomic DNA was extracted using DNeasy Blood and Tissue Kit (Qiagen) following the manufacturer’s instruction. The concentration and purity of genomic DNA were measured using nanodrop spectrophotometers (Thermo Scientific). Fresh genomic DNA with high purity (an A260/A280 ratio of 1.8) was used for experiments. The integrity of genomic DNA was determined by gel electrophoresis. QPCR was performed as previously described (Cawthon, 2002) with further modifications (Jodczyk et al., 2015). Briefly, two qPCR reactions were carried using either primers for single copy reference gene (albumin) or telomeres at a final concentration of 900nM (See Supplementary Table 2 for primer sequences). In each reaction, a standard curve was made by serially diluting one reference DNA by 2 fold ranging from 1.56 ng to 50 ng in order to ensure the PCR efficiency of 90-100% with the linear correlation coefficient greater than 0.98 for each independent experiment. 5ng of sample genomic DNA was added in a 384 well plate with a total reaction volume of 15uL. Each reaction was performed in duplicate and duplicate CT values were averaged and then used to calculate for relative telomere copy number to single gene copy number (T/S) ratio.
Quantification of Dendritic Arborization
Neuronal cultures were immunostained for Tuj1 in order to identify mature neurons and visualize entire cells. The stained cells were then imaged at 20x and dendritic arbors of individual cells were traced using the Simple Neurite Tracer plugin for ImageJ. The number of process ends per cell were counted using the Cell Counter plugin for ImageJ. Only mature neurons—identifiable by their thin processes and condensed somas—were used for analysis. The number of process ends per cell are presented as mean ends per cell +/- SEM. Means were compared using the Student’s T-Test for data with two groups. A p value <0.05 was used as the cutoff for significance.
RNA expression profiling
RNA purification was performed with Qiagen RNAeasy kit following the manufacturer’s instruction. Libraries were prepared according to the manufacturers guidelines using The TruSeq V2 kit (Illumina). Microarray profiling was performed with Affymetrix Human HG-U133 2.0 Plus arrays as described 106. For RNA sequencing, the datasets were mapped with RASER and HISAT2. The reads were counted per gene. Genes were defined by the exon union from the hg19 ensembl annotations. The function of DESeq in DESeq2 package was used to first normalize the gene read counts data and then identified the differentially expressed genes. The MA plot was generated with the function of plotMA in DESeq2 package. Q-value of 0.05 is regarded as the stringent cutoff of calling DEGs while p-value less than 0.05 is regarded as the low stringency cutoff. For the meta-chromosome plot of DEGs, all the chromosomes (except chromosome Y) were first divided equally into 20bins with different length, and then the number of DEGs in each bin was counted. GO analysis was performed using DAVID.
Analysis of 5hydroxymethyl-cytosine 5hmC capture (hmC-Seal)
Chemical labeling-based 5hmC enrichment was described previously107. In brief, five microgram of genomic DNA was sonicated to 100-500 bp, and then mixed with 100 μl reaction buffer (50 mM HEPES at pH 8.0, 25 mM MgCl2, 250 μM UDP-6-N3-Glu and 2.25 μM wild-type β-glucosyltransferase (β-GT)). Reactions were incubated at 37°C for 1 hour, and DNA substrates were purified by Qiagen DNA purification kit. 150 μM dibenzocyclooctyne modified biotin were mixed with β-GT-modified DNA. The labeling reaction was performed at 37°C for 2 hours. The biotin-labeled DNA was then enriched by Strepavidin-coupled Dynabeads (Dynabeads MyOneTM Streptavidin T1, Life Technologies) and purified by Qiagen DNA purification kit. The quantity and quality of purified DNA were analyzed by Qubit 3.0 Fluorometer (Invitrogen) and Agilent 2100 BioAnalyzer using DNA high sensitivity analysis kits (Agilent Technologies).
MeDIP-seq
Methylated DNA Immunoprecipitation (MeDIP) experiments were performed according to the manufacturer’s protocol (Active Motif) and described previously 108. In brief, five microgram of genomic DNA were sonicated to 100-500 bp, and incubated with 5mC-specific antibody (Active Motif) at 4°C overnight. Enriched methylated DNA will be extensively washed and purified by Qiagen DNA purification kit. The quantity and quality of purified DNA were analyzed by Qubit 3.0 Fluorometer (Invitrogen) and Agilent Bioanalyzer using DNA high sensitivity analysis kits (Agilent Technologies).
Library Preparation and High-throughput Sequencing
Enriched DNA from MeDIP and hME-Seal was subjected to library construction using the NEBNext ChIP-Seq Library Prep Reagent Set for Illumina according the manufacturer’s protocol. In brief, 25 ng of input genomic DNA or experimental enriched DNA were utilized for each library construction. 150-300 bp DNA fragments were selected by AMPure XP Beads (Beckman Coulter) after the adapter ligation. An Agilent 2100 BioAnalyzer were used to quantify the amplified DNA, qPCR were applied to accurately quantify the library concentration. 20 pM diluted libraries were used for sequencing. 50-cycle single-end sequencing reactions were performed. Image processing and sequence extraction were done using the standard Illumina Pipeline.
Analysis
Bowtie software was used to map the sequenced reads back to the human genome (hg19) with the parameter of allowing up to two mismatches. Only the uniquely mapped reads were then used to generate the piled-up genome coverage of methylation signals. The methylation signals were further normalized by per million mapped reads for following analysis. The human genome was then segmented into bins of 1kilo-base pairs, which allows for the identification of bins which shows most dramatic methylation signals differences between wild-type and mutant samples. The delta methylation signal of 0.2 per million reads was chosen as the cutoff of calling the Differential Methylation Regions. For the meta-chromosome plot, all the chromosomes (except chromosome Y) were divided equally into 20 bins with different length. We then summarized the total methylation signal within each bin and subtracted the signal between wild-type samples and mutant samples. The subtracted signal was then plotted as the meta-chromosomal plot of methylation differences.
Figure 1-figure supplement 1. hiPSCs lacking MECP2 are pluripotent
Teratoma assay was performed to establish pluripotency of hiPSCs made from Rett patient fibroblasts. The resulting tumors each showed evidence of differentiation towards all three embryonic germ layers.
Figure 1-figure supplement 2. Similarity of NPCs generated with and without MECP2
A, NPCs were produced from isogenic hiPSCs of Rett patient, and assessed by western blot to validate loss of MECP2 and specificity of antibody. Top panel shows that the antibody only recognizes MECP2. Bottom panel shows that in NPCs from both patients, individual clones either express or lack MECP2. B, The ability of hiPSCs to generate NPCs was assayed in Rosette formation assay. Lack of MECP2 did not affect rosette formation across multiple lines from both patients. N=4 independent experiments. *p value < 0.05 according to student’s t test (for patient R567) or ANOVA (for patient R982). Bar graphs represent mean +/- SEM. C, Growth curves show that loss of MECP2 does not affect proliferation of NPCs made from either patient. D, 3 weeks of growth factor withdrawal drives NPCs to differentiate into neurons and glia as measured here by immunostaining for MAP2/Tuj1 or S100/GFAP in patient 567 derived cultures. There was no consistent difference in differentiation potential across lines from either patient. N=2 independent experiments. Bar graphs represent mean +/- SEM. E, Patient 982 derived cultures also do not show dramatic differences in the presence of neurons or astrocytes as measured by MAP2 and S100. N=3 independent experiments. Bar graphs represent mean +/- SEM.
Figure 3-figure supplement 1. RNA-seq analysis to determine the relative ration of WT versus MUT transcripts of MECP2 in Rett patient derived lines. Detection of WT and MUT transcripts from each of the lines indicated demonstrated a clear bias towards individual alleles in each patient derived line. This analysis indicates XCI status for each allele, and demonstrates that XCI status is unchanged, even after differentiation to neurons.
Figure 3-figure supplement 2. A, Immunofluorescence of interneuron progenitors from WT (top) or MECP2null (bottom) clones. B, Immunofluorescence of interneurons generated from a MECP2null hiPSC clone generated by 3i protocol. C, Volcano plots of lower stringency DEGs in hiPSCs and NPCs between MECP2+ versus MECP2-clones (p-value < 0.05). D, Mapping of low stringency DEGs in hiPSCs and NPCs across metachromosome to measure enrichment of DEG location.
Figure 4-figure supplement 1. Silencing MECP2 by siRNA
MECP2 was downregulated by RNA interference, quantified by RT-PCR (left), for protein by western blot (middle), and as demonstrated by immunostaining for MECP2 (right). N=3 independent experiments. *p value < 0.05 according to student’s t test. Bar graphs represent mean +/- SEM.
Figure 6-figure supplement 1. Transduction of Progerin leads to phenotypes similar to loss of MECP2. A, Cells undergoing senescence are known to induce and secrete a group of genes called SASP. RNA-seq data from neurons were mined for SASP genes, and shown are those SASP genes that were differentially expressed between patient derived neurons with and without MECP2. B, qPCR for telomere length in WT NPCs showed that Progerin infected cells have on average shorter telomeres. N=3 independent experiments. *p value < 0.05 according to student’s t test. Bar graphs represent mean +/- SEM. C, RT-PCR from progerin infected NPCs showed an induction of various subtelomeric genes. N≥3 independent experiments. *p value<0.05 according to student’s t test. Bar graphs represent mean +/- SEM. D, RT-PCR for TERRA transcripts following progerin transduction. N≥3 independent experiments. *p value < 0.05 according to student’s t test. Bar graphs represent mean +/- SEM. E, Immunostaining for p53 following progerin expression. Quantification of p53 in infected cells (either Nuclear-GFP or Progerin-GFP) is shown on the right. *p value < 0.05 according to student’s t test. Bar graphs represent mean +/- SEM.
Supplemental Table 1.
Provided is a list of all cell types analyzed across the RT-PCR and Telomere qPCR experiments performed in this manuscript.
Supplemental Table 2.
Provided is a list of all the primers used in this study.
Acknowledgements
We are grateful to Lorenz Studer for sharing the Progerin-GFP and Nuclear-GFP constructs. Human tissue was obtained from the NIH Neurobiobank at the University of Maryland (Baltimore, MD). AZX and TPW are partially supported by Connecticut Regenerative Medicine Foundation (13-SCB-YALE-12). This work was funded by training grants to MO (NIH-Virology and Gene Therapy, UCLA), PL (CIRM, UCLA), CS (CIRM-Bridges, Cal-State-Northridge), DA (HHURP, UCLA). PJ was supported by grants from NIH (KP was supported by NS051630, NS079625 and MH102690). AZX and TPW were supported by the Connecticut Regenerative Medicine Foundation (13-SCB-YALE-12). WEL was supported by a Rose Hills Scholar award through the Eli and Edythe Broad Center for Regenerative Medicine. WEL and KP were supported by NIH (P015P01GM099134). This research was also supported by the Allen Distinguished Investigator Program, through The Paul G. Allen Frontiers Group.
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.
- 19.
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.
- 37.↵
- 38.
- 39.
- 40.
- 41.↵
- 42.
- 43.
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.
- 50.
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.
- 59.
- 60.↵
- 61.↵
- 62.↵
- 63.
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.
- 70.
- 71.
- 72.
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.
- 89.
- 90.
- 91.
- 92.↵
- 93.↵
- 94.↵
- 95.
- 96.↵
- 97.↵
- 98.↵
- 99.↵
- 100.↵
- 101.
- 102.↵
- 103.↵
- 104.
- 105.
- 106.↵
- 107.↵
- 108.↵