Transplantation studies reveal internuclear transfer of toxic RNA in engrafted muscles of myotonic dystrophy 1 mice.

BACKGROUND: Stem cell transplantation represents a potential therapeutic option for muscular dystrophies (MD). However, to date, most reports have utilized mouse models for recessive types of MD. Here we performed studies to determine whether myotonic dystrophy 1 (DM1), an autosomal dominant type of MD, could benefit from cell transplantation.
METHODS: We injected human pluripotent stem (PS) cell-derived myogenic progenitors into the muscles of a novel mouse model combining immunodeficiency and skeletal muscle pathology of DM1 and investigated transplanted mice for engraftment as well as for the presence of RNA foci and alternative splicing pattern.
FINDINGS: Engraftment was clearly observed in recipient mice, but unexpectedly, we detected RNA foci in donor-derived engrafted myonuclei. These foci proved to be pathogenic as we observed MBNL1 sequestration and abnormal alternative splicing in donor-derived transcripts.
INTERPRETATION: It has been assumed that toxic CUG repeat-containing RNA forms foci in situ in the nucleus in which it is expressed, but these data suggest that CUG repeat-containing RNA may also exit the nucleus and traffic to other nuclei in the syncytial myofiber, where it can exert pathological effects. FUND: This project was supported by funds from the LaBonte/Shawn family and NIH grants R01 AR055299 and AR071439 (R.C.R.P.). R.M-G. was funded by CONACyT-Mexico (#394378).

Evidence before this study

We and others have previously demonstrated the beneficial outcome of cell transplantation in mouse models of recessive types of muscular dystrophies. In this context, donor-derived engrafted myonuclei are able to provide fully functional proteins that overcome the phenotype of absent or mutated endogenous proteins, such as dystrophin in the case of Duchenne muscular dystrophy (DMD). However, to date, limited literature exists in the context of cell transplantation for dominant types of muscular dystrophies, and none for myotonic dystrophy 1 (DM1).

Added value of this study

Here we assessed for the first time whether DM1 could be a potential candidate for cell-based therapy. Unexpectedly, our findings revealed that toxic RNAs produced in the endogenous nuclei of DM1 recipient mice are transmitted to donor-derived nuclei, leading to alterations in the alternative splicing of donor-derived transcripts.

Implications of all the available evidence

Our findings suggest a pathomechanism for DM1, in which toxic RNAs are able to exit the nucleus in which they were synthesized to exert pathological effects in other myonuclei within the syncytium. Although expression of toxic RNA in the DM1 mouse model is different from DM1 patients, one cannot exclude the possibility of foci acquisition by normal cells in the scenario DM1 patients receive cell transplantation. This is also relevant for alternative strategies, such as gene editing, as non-corrected myonuclei producing toxic RNAs may affect the alternative splicing of successfully gene-corrected myonuclei within the syncytium.

Alt-text: Unlabelled Box

Introduction

Myotonic Dystrophy 1 (DM1) is an autosomal dominant multi-systemic disorder with an estimated prevalence of 1/10,000 worldwide, although this varies among populations [1,2]. Several tissues are affected in DM1, and in particular skeletal muscle undergoes progressive weakness and wasting, along with myotonia [3,4]. DM1 is caused by an expansion of the CTG triplet repeats contained in the 3’UTR of the dystrophia myotonica protein kinase (DMPK) gene [5]. DMPK transcripts containing expanded CUG triplets fold into extended stem-loop structures that accumulate as intranuclear RNA foci [[6], [7], [8], [9]]. Most of the clinical manifestations of DM1 have been related to the sequestration of the alternative splicing factor muscleblind-like protein 1 (MBNL1) by the RNA foci, which alters the proper alternative splicing of MBNL1 target genes [[10], [11], [12]].

Several mouse models have been developed to study DM1 pathology [[13], [14], [15]]. Among these, the HSALR mouse model [14], which expresses 250 CUG repeats in the context of the human skeletal actin gene, has been widely used to study DM1 skeletal muscle phenotype and potential treatment strategies aiming to recover the mis-splicing alterations. Despite progress, clinical trials tested so far, such as the use of modified antisense oligonucleotides to eliminate toxic RNA foci [[16], [17], [18]], have failed to provide significant improvement in DM1 patients mainly due to limited drug uptake in muscle (ClinicalTrials.gov, NTC02312011). Thus, therapeutic strategies that ameliorate the muscle disease phenotype in DM1 are warranted.

Cell-based therapy has emerged as a promising treatment strategy for recessive muscular dystrophies [[19], [20], [21]]. For instance, it has been shown that transplantation of myogenic precursors in mouse models of Duchenne muscular dystrophy (DMD) leads to the generation of donor-derived new myofibers and hybrid myofibers containing donor-derived myonuclei, which are able to provide dystrophin, thus resulting in functional improvement [[22], [23], [24], [25], [26]]. However, the feasibility of cell-based therapy for dominant MDs is complicated by the fact that the muscle fiber is a syncytium. This means that if mutant nuclei are not completely eliminated, there will always be some background expression of the dominant disease-causing gene.

To address this question, here we performed transplantation studies using a novel immunodeficient DM1 mouse model. Successful engraftment was achieved, but unexpectedly, we observed the presence of pathogenic RNA foci in donor-derived nuclei of hybrid myofibers. These results shed light on a novel dynamic behavior of toxic RNA foci among myonuclei, with potential implications for the efficacy of cell therapy for DM1.

Materials and methods

Mice

Animal handling was performed according to protocols approved by the University of Minnesota Institutional Animal Care and Use Committee. Breeding pairs of homozygous HSALR mice [14] were kindly provided by Dr. Charles Thornton (University of Rochester, NY). NSG [27,28] and CAG:H2B-EGFP [29] mice were purchased from Jackson Laboratories (005557 and 006069, respectively). HSALR were bred to NSG mice to obtain a homozygous HSALR (+/+) NS(+/+)G(+) colony. HSALR transgene homozygosity and CTG length confirmation were carried out as previously described [30].

Peripheral blood FACS analysis

Peripheral blood from facial vein of HSALR, NSG and NSG-HSALR mice was collected onto EDTA-containing tubes. Samples were sequentially incubated with RBC lysis buffer (Biolegend) for 5 min at room temperature (RT), and FcR blocking antibody for 5 min on ice. Then, cells were resuspended in FACS buffer (10% FBS in PBS) and incubated with appropriate antibodies for 20 min on ice. Antibodies are listed in Table S1. FACS analysis was performed using a FACSAria (BD Biosciences). Data were analyzed using FlowJo software.

Cell transplantation

For transplantation studies, we used myogenic progenitors derived from the previously reported human iPAX7 PLZ iPS cell line [26,31]. Mouse myogenic progenitors were generated from inducible Pax3 (iPax3) mES cells, as previously described [[32], [33], [34]]. Mouse satellite cells were isolated from several muscles, as reported [35]. For the transplantation of mES cell-derived myogenic progenitors into HSALR mice, these mice received daily intra-peritoneal (IP) injections of the immunosuppressant agent Tacrolimus (AKSci) at a dose of 5 mg/kg. Treatment began two days before transplantation and ended by the day of euthanasia [36]. mES cell-derived myogenic progenitors, hiPS cell-derived myogenic progenitors and human skeletal myoblasts (HSKM) (Gibco) were injected at 7·5 × 105 cells, and H2B-eGFP satellite cells at 5 × 103 cells. Cells resuspended in 15 μl of PBS were injected into the TA muscles that had been pre-injured 24 h prior to cell transplantation with either 15 μl of 10 μM cardiotoxin, 15 μl of 1·2% barium chloride in saline solution, or by freeze injury [26,37]. Freeze injury was performed by applying a liquid nitrogen-cooled spatula directly on the exposed muscle twice for 10 s with a 1 min interval in-between [38]. TA muscles were dissected four weeks after transplantation and resulting cryosections (10 μm) were evaluated, as described below.

RNA-FISH/DNA-FISH/immunofluorescence

To identify nuclei containing human DNA, we performed DNA FISH using human Cot-1 as a probe [39,40]. Cot human DNA (Roche) was labelled using the Atto647N NT labeling kit (Jena Bioscience) following the manufacturer's instructions. Labelled probe was then purified using QIAquick PCR purification kit (QIAGEN). DNA FISH protocol was performed as follows: muscle cryosections on glass slides were thawed to RT and hydrated with PBS for 5 min. Then, sections were fixed with 4% PFA for 30 min, permeabilized with 0·3% Triton X-100 for 15 min, rinsed with 2× SSC and then incubated with denaturation buffer (70% formamide in 2× SSC) for 10 min at 80 °C in a humid chamber. Samples were then incubated with probe solution (2× SSC, 50% formamide, 2 mM ribonucleoside vanadyl complex, 75 ng/μl tRNA and 4 ng/μl of labelled probe) for 7 min at 80 °C, followed by overnight incubation at 37 °C. Probe solution was heated for 10 min at 80 °C prior to incubation. The next day, slides were washed with 50% formamide in 2× SSC for 5 min at 41 °C and then washed again with 2× SSC at RT. Slides were then processed for RNA FISH-immunofluorescence starting with the pre-hybridization buffer incubation step (see RNA FISH protocol below).

To detect the intranuclear RNA foci, muscle cryosections were fixed with 4% PFA for 10 min, permeabilized with 0·3% Triton X-100 for 15 min, and incubated with pre-hybridization buffer solution (2 x SSC and 30% formamide in DEPC water) for 10 min at RT. Samples were incubated with hybridization solution (2× SSC, 30% formamide, 0·02% BSA, 2 mM ribonucleoside vanadyl complex, 66 μg/ml yeast tRNA and 0.1 ng/ul of a Cy3-labelled (CAG)7 probe in DEPC water) for 2 h at 37 °C in a humid chamber. Samples were washed twice with pre-hybridization buffer solution for 7 min at 42 °C and two more times with 1× SSC in DEPC water for 5 min at RT. Samples were then processed for immunostaining: first slides were blocked with 3% BSA in PBS for 30 min, followed by incubation overnight with selected primary antibodies at 4 °C. The following day, samples were washed three times with PBS for 5 min and incubated with proper secondary antibodies for 1 h at RT. Samples were washed again 3× with PBS for 5 min and mounted with ProLong Gold antifade reagent with DAPI (Invitrogen). Analysis was done by confocal microscopy (Nikon NiE C2 upright confocal microscope). Antibodies are listed in Table S1.

RT-PCR

Samples were collected with TRIzol™ Reagent (Invitrogen) and RNA was purified using the Direct-zol™ RNA Miniprep Plus kit (Zymo Research) following manufacturer's instructions. Purified RNA was quantified with NanoDrop 2000 (Thermo Scientific) and reverse-transcribed using the SuperScript VILO cDNA Synthesis Kit (Invitrogen) following manufacturer's instructions. cDNA was used as template for PCR using GoTaq Flexi DNA polymerase (Promega). Primers that have been previously reported were used to analyze Clcn1 exon 7A [41], Nfix exon 7 [41], Serca1 exon 22 [42], Gapdh [41], human specific SERCA1 exon 22 [43] and human specific LDB3 exon 11 [44]. Primers for ACTB: Fwd 5’ CATGTACGTTGCTATCCAGGC 3′ and Rev. 5’ CTCCTTAATGTCACGCACGAT 3′; and human specific CACNA1S exon 29: Fwd 5’ CTGATTGTCATTGGCAGCAT 3′ and Rev. 5’ AGGGTTCGCACTCCTTCTG 3′.

Statistical analysis

Statistical comparisons were performed using the unpaired Student's t-test in Prism 7 software (GraphPad). p values <0·05 were considered significant.

Results

Generation of an immunodeficient mouse model for DM1

To provide an ideal environment for the transplantation of mouse and human cells, we have crossed the HSALR mutant DM1 strain [14] onto the NSG (NOD/SCID; IL2 receptor gamma) background [28]. NSG mice lack all functional classes of lymphocytes (T, B, and NK cells) [45,46], and therefore are widely used for xenotransplantation studies [24,28,47]. Upon obtaining NSG-HSALR homozygous at all loci, the peripheral blood of these mice was characterized for the lack of T, B and NK cells, whereas the skeletal muscle was tested for the presence of RNA foci. FACS analysis confirmed the absence of NK cells (NK 1.1+CD49b+ cells), as well as T and B lymphocytes (CD3e+ and CD19+ cells, respectively) in NSG-HSALR mice, in contrast to immunocompetent HSALR mice (Fig. 1a), which contained all these lymphocyte subtypes.

Fig. 1

Characterization of immunodeficiency and DM1 features in NSG-HSALR mice. (a) FACS profile for circulating NK (CD49b/NK1.1), T (CD3e) and B (CD19) cells in HSALR, NSG, and NSG-HSALR mice. (b) Representative images show RNA-FISH analysis (foci, red) in combination with immunostaining for MBNL1 (green) and pan-dystrophin (DYS, white) in muscle sections of NSG-HSALR mice. DAPI stains nuclei (blue). Asterisks indicate co-localization between foci and MBNL1. Scale bar is 20 μm. Maximum projection is shown. (c) Alternative splicing pattern analysis of Clcn1 exon 7A, Nfix exon 7 and Serca1 exon 22 in TA muscles from NSG, HSALR and NSG-HSALR mice. (d) Bar graphs represent densitometric quantification of Clcn1 exon 7A inclusion, Nfix exon 7 inclusion, or Serca1 exon 22 exclusion from three independent replicates (from c). Data are shown as mean + SEM. ***p < 0·001. N = 3 muscles, one per mouse. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

To verify the presence of typical DM1 molecular hallmarks, we performed RNA FISH, using a labelled probe that detects the expanded CUG repeats, combined with immunostaining for MBNL1. Similar to homozygous HSALR mice (Fig. S1), NSG-HSALR mice displayed intranuclear RNA foci that co-localized with MBNL1 labelling, confirming the sequestration of MBNL1 by the mutant RNA aggregates (Fig. 1b). As anticipated, these features were not found in control non-DM1 immunodeficient mice (Fig. S1). Furthermore, we also confirmed significant alterations in alternative splicing of MBNL1 target genes, including Clcn1 exon 7A inclusion, Nfix exon 7 inclusion and Serca1 exon 22 exclusion (Fig. 1c-d). These results validate this novel mouse model as NSG-HSALR mice combine the immunodeficiency features of NSG mice with the DM1 molecular features of HSALR mice.

Engraftment of iPS cell-derived myogenic progenitors in NSG-HSALR mice

To determine whether cell transplantation could be used to overcome the splicing defects observed in DM1, we injected human iPS cell-derived myogenic progenitors [26] into cardiotoxin pre-injured TA muscles of NSG-HSALR mice. A cohort of NSG mice was used as control recipients. Four weeks post-transplantation, we assessed human engraftment by performing immunostaining of muscle cryosections with antibodies specific to human Lamin A/C (hLMNA/C) and human Dystrophin (hDYS). We detected donor-derived myofibers (hLMNA/C + hDYS+) in both NSG-HSALR and NSG mice (Fig. 2a), and at similar levels (Fig. 2b), corroborating the usefulness of the NSG-HSALR mouse model for the transplantation of human cells. Importantly, no labelling signal was detected in TA muscles that had been injected with PBS (control), thus confirming the human specificity of used antibodies (Fig. S2).

Fig. 2

Engraftment of myogenic progenitors upon their intra-muscular transplantation into NSG-HSALR mice. (a) Representative images show human engraftment as demonstrated by immunostaining for human DYSTROPHIN (hDYS) and human LAMIN A/C (hLMNA/C) in cryosections of TA muscles from NSG (upper panel) and NSG-HSALR (lower panel) mice that had been injected with human iPS cell-derived myogenic progenitors. Nuclei (blue). Scale bar is 100 μm. (b) Quantification of engraftment, as shown by the number of double-positive hDYS+/hLMNA/C+ myofibers (from a). Data are shown as mean + SEM, N = 4 transplanted muscles, one per mouse. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Identification of RNA foci in donor-derived myonuclei

To determine whether engraftment reduced the presence of toxic foci, we performed RNA-FISH on muscle sections of transplanted NSG-HSALR mice. To our surprise, we detected the foci not only in the endogenous host mouse nuclei, but also in human donor-derived myonuclei (hLMNA/C+ nuclei within hDYS+ myofibers) (Fig. 3a, upper panel). This result was specific to NSG-HSALR mice since no such foci could be detected in control NSG recipients transplanted with the same cell population (Fig. 3a, lower panel).

Fig. 3

Detection of toxic RNA foci in donor-derived engrafted myonuclei. (a) Representative images show RNA-FISH analysis (foci in red) in combination with immunostaining for hDYS (white) and hLMNA/C (green) in transplanted TA muscles of NSG-HSALR (upper panel) and NSG (lower panel) mice. Nuclei (blue). Yellow asterisks in the upper panel indicate human donor-derived myonuclei positive for the foci (NSG-HSALR recipients), and white asterisk (lower panel) shows human donor-derived myonuclei negative for RNA foci in engrafted NSG mice muscles. Scale bar is 20 μm. Mid Z projection is shown. (b) Representative images confirm the human origin of engrafted hLMNA/C+ myonuclei (green) containing RNA foci (red), as evidenced by detection of human DNA content (hDNA, purple) by DNA-FISH analysis in transplanted NSG-HSALR. Asterisks show NSG-HSALR endogenous myonucleus (white) and donor-derived myonucleus (yellow) Scale bar is 5 μm. Mid Z projection is shown. (c) Pie charts show the quantification hLMNA/C+ myonuclei positive for hDNA and RNA foci in TA muscles of NSG-HSALR mice transplanted with hiPSC-derived myogenic progenitors (N = 6 transplanted muscles, one per mouse; 160 total nuclei). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

We reasoned that the presence of foci in donor-derived nuclei could be a result of fusion of non-affected human myogenic progenitors with pre-existing DM1 mouse muscle fibers, in which human Lamin A/C proteins would be synthesized and directed to mouse nuclei within the fiber, thereby falsely identifying host nuclei as donor-derived. To rule out this hypothesis, we performed immunostaining to hLMNA/C in combination with DNA FISH using human Cot-1 as a probe to specifically label nuclei containing human DNA (hDNA) [40]. Importantly, no hDNA signal was observed in PBS-injected muscles (Fig. S2), thus confirming the reliability of the combined methodologies to identify cells of human origin. Lastly, we incorporated RNA-FISH to assess the number of human myonuclei containing foci. We observed that from the total of hLMNA/C+ myonuclei, 67% were also positive for hDNA. From these double positive myonuclei, 65% were also positive for RNA foci (Fig. 3b-c) (Statistical analysis is shown in Table S2). Of note, within the same fiber, we detected the presence of mouse nuclei, absent of hDNA or hLMNA/C labelling, but positive for RNA foci (Fig. 3b). These results indicate that non-affected human myogenic nuclei acquired intranuclear RNA foci, likely through the shared cytoplasm of a chimeric fiber. In accordance, RNA foci transfer was also observed upon the transplantation of mouse embryonic stem (mES) cell-derived myogenic progenitors [32,33] into tacrolimus-treated immunocompetent HSALR mice, as shown by the presence of foci in donor-derived H2B-GFP+ nuclei (nGFP) (Fig. S3, upper panel).

To determine whether RNA foci transfer was related in some way to cardiotoxin, we performed cell transplantation in the context of other types of muscle injury, including freezing injury and BaCl2 injection. The same outcome was obtained, regardless of the pre-injury method used (Fig. S4A).

Interestingly, when we performed a reverse experiment, in which patient-specific DM1 myogenic progenitors [48] or HSALR satellite cells were transplanted into TA muscles of control NSG mice, we did not detect RNA foci in host myonuclei (data not shown). These results may be due to the low abundance of engrafted myonuclei providing mutant transcripts in comparison with host myonuclei within the myofiber of HSALR mice.

RNA foci present in donor-derived human myonuclei are pathogenic

To understand the biological relevance of RNA foci detected in engrafted human myonuclei, we first analyzed MBNL1 distribution and observed its sequestration by the acquired foci (Fig. 4a), thus suggesting a toxic effect. Next we determined whether MBNL1 sequestration would have subsequent consequences in the alternative splicing pattern of human SERCA1 exon 22, LDB3 exon 11 and CACNA1S exon 29, which have been shown to be altered in DM1 [11,44,49]. RT-PCR analysis was performed using primers specific to human transcripts, as shown by analysis of control PBS-injected TA muscles as well as non-injected gastrocnemius muscle (Fig. 4b). While transplantation of non-affected human myogenic progenitors into NSG mice resulted in a splicing pattern similar to non-affected human muscle tissue, transplantation into NSG-HSALR mice resulted in alterations of alternative splicing in human transcripts, similar to the ones reported in DM1 i. e. exclusion of SERCA1 exon 22, inclusion of LDB3 exon 11, and exclusion of CACNA1S exon 29 (Fig. 4b and d). To rule out the possibility that this finding was exclusive to the transplanted cell type, we performed similar intramuscular transplantation studies using primary human skeletal myoblasts (HSkM). Again, we observed hLMNA/C+ nuclei containing RNA foci in NSG-HSALR, but not in NSG mice (Fig. S4b), and this was accompanied by alternative splicing alterations of donor-derived transcripts (Fig. 4c-d), thus corroborating our previous observations with PS cell-derived myogenic progenitors.

Fig. 4

Pathogenic features of RNA foci detected in donor-derived engrafted myonuclei. (a) Representative images show MBNL1 (green) sequestration by RNA foci (red) in human-derived myonuclei expressing hLMNA/C (white). Asterisks indicate NSG-HSALR endogenous myonucleus (white) or donor-derived engrafted myonucleus (yellow). Scale bar is 5 μm. Mid Z projection is shown. (b-c) RT-PCR shows alternative splicing analysis of human-specific SERCA1 exon 22, LDB3 exon 11 and CACNA1S exon 29 in TA muscles from NSG or NSG-HSALR mice injected with PS cell-derived myogenic progenitors (b) and human skeletal myoblasts (c). Controls included PBS-injected TA muscles, gastrocnemius as a non-injected muscle, and human muscle cDNA as a reference of non-affected alternative splicing. *Samples not quantified due to low amplicon detection. (d) Bar graphs show densitometric analysis of human specific alternative splicing events from (b) and (c) in transplanted muscles of NSG (N = 4 transplanted muscles, one per mouse) and NSG-HSALR (N = 6 transplanted muscles, one per mouse) mice. Data are shown as mean + SEM. * p ≤ 0·05 **p ≤ 0·01 ***p ≤ 0·001. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

In another experimental cohort, we isolated satellite cells from CAG:H2B-eGFP mice [29], which allowed us to identify donor-derived nuclei through nuclear GFP labelling (as shown in Fig. S3 for mES-derived myogenic progenitors), and transplanted these into NSG-HSALR mice. Consistent with previous results, nGFP+ nuclei were also found positive for RNA foci (Fig. 5a), and accordingly, no changes in splicing pattern was observed in transplanted muscles (Fig. 5b-c).

Fig. 5

Detection of toxic RNA foci in donor-derived myonuclei upon satellite stem cell transplantation. (a) Upper panels show representative images for the engraftment of satellite cells isolated from a H2B-eGFP reporter mouse, as shown by the presence of nGFP (green) positive for RNA foci (red). Lower panels show PBS-injected muscles, which are negative for nGFP and positive for the foci. DYS (white) allows myofiber visualization. Asterisks indicate donor-derived engrafted myonuclei. Scale bar is 20 μm. Mid Z projection is shown. (b-c) RT-PCR analysis shows alternative splicing pattern of Clcn1 exon 7A, Nfix exon 7 and Serca1 exon 22 in NSG-HSALR muscles injected with satellite cells or PBS (b), and respective quantification (c). Data are shown as mean + SEM. N = 3 transplanted muscles, one per mouse. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Discussion

Although several attempts have been made to improve the muscle pathology in DM1, to date, there is no effective treatment available. Therefore, there is still a need to develop novel therapeutic approaches with the potential to improve patient's quality of life. In this regard, cell-based therapy has demonstrated promising results in the recovery of muscle function when applied to recessive muscular dystrophies (e.g. DMD), diseases in which donor-derived cells are able to provide a functional molecule when the endogenous one is missing or defective as a consequence of a given genetic condition. This can occur through the generation of new myofibers and/or by fusion with preexistent ones. In an autosomal dominant MD context however, this may be more challenging as engrafted cells need to overcome the molecular hurdles associated with the dominant allele that continuously damage the muscle fiber. In this regard, there has been limited discussion about the importance of disease etiology on cell transplantation efficiency. In DM1, sequestration of MBNL1 by RNA foci leads to alternative splicing alterations in genes that are important to maintain muscle structure and function. If toxic RNAs were retained in their nuclei of origin, one would expect that cell transplantation in DM1 would provide mRNAs with proper alternative splicing, leading to fully functional proteins that would dilute the endogenous ones, thus contributing to phenotype recovery. Here we show this is not the case since upon successful engraftment of hiPS cell-derived myogenic progenitors in immunodeficient DM1 mice, we found that donor-derived myonuclei are positive for RNA foci, resulting in disrupted alternative splicing of donor-derived transcripts. This finding was consistent regardless of the myogenic cell type transplanted (i. e. human skeletal myoblasts, mouse satellite cells, and mES-derived myogenic progenitors) or the injury method used prior to transplantation [48].

Our results suggest a dynamic behavior of the RNAs containing CUG expansions, by which these are able to migrate to other nuclei within a muscle fiber and form RNA foci. Notably, we did not reliably detect RNA foci in the cytoplasm of DM1 mouse myofibers, which supports the hypothesis that toxic RNAs migrate from the nucleus as complexes less structured than foci, or alternatively, as single molecules that could not be detected by the resolution of RNA-FISH. Although DM1 RNA foci have been widely reported to be retained within the nucleus, there has been limited literature aimed at understanding foci molecular dynamics [[50], [51], [52], [53], [54], [55], [56]]. Of note, RNA foci have also been observed in the cytoplasm of different cell types, such as patient-derived DM1 myoblasts and fibroblasts [54,55], but this has been mainly attributed to cell division causing the aggregates to remain outside the nucleus after nuclear membrane disassembly, a process that would not occur in non-dividing cells, such as myotubes. Nonetheless, a potential mechanism in which toxic RNA molecules are actively exported has not been completely ruled out [55,57]. For instance, it has been shown that the RNA-binding protein Staufen1 (STAU1) interacts with and promotes the nuclear export of DMPK 3’UTR transcripts containing expanded CUG triplets [50]. This becomes relevant as in the same study, the authors also showed that STAU1 is increased in skeletal muscle tissue of DM1 patients and DM1 mouse models, such as HSALR [50]. Notably, STAU1 was not found to co-localize with foci, in line with the hypothesis that the toxic RNA export may occur as individual molecules not detectable by RNA-FISH. This premise is corroborated by a study showing that unwinding of CUG-repeats secondary structures by the DEAD-box RNA helicase DDX6 also increases the abundance of cytoplasmic mutant transcripts [54]. Moreover, it has been shown that an individual CUGn-containing mRNA is sufficient to interact with MBNL1, a protein known to undergo a dynamic nucleus-cytoplasmic transport, thus enabling a possible nuclear export mechanism by means of single molecules of MBNL1 carrying single molecules of RNA [51,58,59]. Interestingly, a previous study documented that in in vitro-differentiated heterokaryons consisting of human DM1 myonuclei and control mouse myonuclei, alternative splicing of mouse transcripts is affected by human DMPK mutant mRNAs [60]. Of note, no foci were observed in mouse nuclei, despite detection of mouse splicing alterations. This could be related to the time of heterokaryons formation, i. e. in vitro hybrid myotubes were analyzed five days after differentiation while our in vivo transplantation experiments were analyzed 30 days post-transplantation. This finding goes in line with our in vivo observations suggesting that DMPK mutant transcripts exit the nucleus in which they were synthesized to cause toxicity in other myonuclei within the syncytium.

Similar to other muscular dystrophies, mouse models of DM1 that reliably resemble the muscle phenotype and course of the disease while harboring the human mutation are lacking [61,62]. In spite of this limitation, the HSALR mouse model has been successfully used in the investigation of the i) mechanisms underlining DM1and the ii) testing of therapeutic approaches, such as the use of antisense oligonucleotides [63]. For the purposes of this study, it is essential to consider that the HSA mouse model expresses toxic RNA at higher levels than in the context of human DM1, which could lead to saturation and re-aggregation of RNA foci among myonuclei. This possibility remains to be tested by performing cell transplantation studies using DM1 mouse models expressing the CTG expansions in the context of the DMPK gene [13]. Nonetheless, our results provide the first precedent of a potential outcome that would need to be considered for DM1 cell therapy. Moreover, from the molecular point of view, we show evidence not only of toxic RNA export to the cytoplasm, but also of the ability of this RNA to be imported into other nuclei within the same muscle fiber and having pathological consequences. This is particularly relevant in the design of a therapy for DM1 since alternative approaches currently under investigation, such as genome editing, may encounter similar challenges. It is possible that within a myofiber, toxic RNA from non-gene edited myonuclei may be toxic to the successfully corrected ones.

Although mechanistic studies are still required to decipher the dynamic behavior of DM1 toxic RNA in myofibers, our cell transplantation results shed new light on toxic RNA nucleus/cytoplasm transport and opens a new perspective on DM1 molecular pathogenesis. Furthermore, our findings suggest a potential communication among nuclei within a myofiber via RNA export/import, a possibility that, to our knowledge, has not been foreseen.

The following are the supplementary data related to this article

Supplementary Table 1

List of antibodies used in this study.

Supplementary Table 2

Statistical analysis of the quantification of engrafted myonuclei (hLMNA/C+) analyzed for hDNA and RNA foci.

Supplementary material

Supplementary Fig. 1. RNA foci analysis in muscles from HSALR mice. Representative images show RNA-FISH (foci in red) in combination with immunofluorescence for MBNL1 (green) in TA muscle sections from control NSG (upper panel) and HSALR (lower panel) mice. DYS is shown in white. Asterisks indicate NSG myonuclei negative for foci (white) or HSALR myonuclei positive for foci and MBNL1 sequestration (yellow). Scale bar is 20 μm. Maximum projection is shown.

Supplementary Fig. 2. Validation of human-specific staining. Representative images show immunostaining for hLMNA/C and hDYS, as well as DNA-FISH for human DNA, in TA muscles from NSG-HSALR mice that had been injected with PBS. Scale bar is 100 μm.

Supplementary Fig. 3. Transplantation of mES-derived myogenic progenitors into immunosuppressed HSALR mice also results in the presence of RNA foci in donor-derived myonuclei. Representative images show nGFP immunostaining and RNA FISH analysis in TA muscles of HSALR that had been transplanted with mES-derived myogenic progenitors (upper panel). NSG mice served as control recipients (lower panel). Scale bar is 20 μm. Maximum projection is shown

Supplementary Fig. 4. RNA foci are detected in donor-derived engrafted myonuclei regardless of muscle pre-injury method used prior to transplantation. (a) Representative images show RNA-FISH analysis (foci, red) in combination with immunostaining for hLMNA/C (green) and hDYS (white) in TA muscles that had been pre-injured by freezing (upper panel) or Barium Chloride (BaCl2) injection (lower panel) 24 h prior to transplantation with human iPS-cell derived myogenic progenitors. Yellow asterisks indicated donor-derived engrafted myonuclei positive for RNA foci. Scale bar is 10 μm. Mid Z projection is shown. (b) Representative images show engrafted human skeletal myoblasts hMLNA/C+ (green) hDYS+ (white) positive for RNA foci (red) in NSG-HSALR muscles (upper panel, yellow asterisks) or negative in NSG mice (lower panel, white asterisks). Scale bar is 20 μm. Mid Z projection is shown.

Funding sources

This project was supported by funds from the LaBonte/Shawn family and NIH grants R01 AR055299 and AR071439 (R.C.R.P.). R.M-G. was funded by CONACyT-Mexico (#394378). Funding sources were not involved in the study design, collection, analysis or interpretation of data, or decision to submit the work for publication.

Author contributions

R.M-G. designed and performed experiments, analyzed the data and wrote the manuscript; K.A. and S.S. designed and performed experiments, and analyzed the data, A.Y. performed experiments and analyzed the data. R.C.R.P. supervised the study, contributed with experimental design and interpretation of the data, and wrote the manuscript.

Declaration of Competing Interest

The authors declare no competing financial interests.