Characterization of DMPK and MBNL1 expression in cell models of myotonic dystrophy: a platform for drug screening

Abstract

Myotonic dystrophy type I (DM1) is caused by CTG repeat expansions in the DMPK gene leading to mRNA toxicity and sequestration of the splicing regulator MBNL1, affecting many tissues. We have developed an in vitro screening platform based on ddPCR and in-cell western to quantify these mRNAs and proteins and characterized >20 cell models to define DM1 biomarkers that could be useful for drug screening. DMPK protein levels were reduced in DM1-immortalized myoblasts and myotubes, but not in fibroblasts, while MBNL1 protein was consistently lower in all DM1 myogenic cultures, whether primary or immortalized. Myogenic differentiation of cultures led to an increase in DMPK mRNA expression, which was translated into increased MBNL1 sequestration in foci. We further corroborated the platform’s ability to assess therapeutic outcomes, evaluating the effect of a DMPK gapmer ASO and one siRNA: while the gapmer increased MBNL1 protein levels, the siRNA had no significant effect on MBNL1 release. Our platform and the in-depth characterization of some of the most used models would be of use to the DM1 research community.

Graphical Abstract

Graphical Abstract.

Introduction

Myotonic dystrophy type I (DM1) is a complex autosomal disease caused by expansion of CTG trinucleotide repeats in the 3′ untranslated region of the DMPK gene. It is the most common form of adult muscular dystrophy, affecting 1 in 8000 people [1]. DM1 is characterized by a variety of multisystemic symptoms caused by a complex molecular pathogenesis, including toxicity of DMPK mRNA transcripts and nuclear sequestration of muscleblind-like (MBNL) proteins, leading to deregulation of multiple signalling pathways [2–4]. The MBNL family are RNA-binding proteins that play a key role in the regulation of RNA processing, particularly in alternative splicing. MBNL1 is widely expressed, and it is essential for post-natal remodelling of the skeletal muscle [5]. In DM1, MBNL proteins, mainly MBNL1, bind with high affinity to the repeat tract of the DMPK expanded transcript and accumulate in the nucleus. This leads to the depletion of free and functional MBNL1 in the nucleus and cytoplasm [6, 7]. The lack of free MBNL1 induces an alternative splicing transition from adult-to-foetal, altering the splicing of multiple targets, such as CLCN1,DMD,BIN1, and INSR. The deregulation of these transcripts is associated with some of the many DM1 symptoms, including myotonia, the dystrophic process, muscle weakness and insulin resistance, respectively [8].

Various therapeutic strategies such as small molecules, antisense oligonucleotides (ASOs) and gene therapies, have been tested in DM1 in vitro, in vivo, and in a handful of clinical trials. Most therapies, through different mechanisms of action, aim to reduce RNA toxicity by decreasing DMPK mRNA transcripts or preventing MBNL1 binding to the expanded transcript [9].

DM1’s complex molecular pathogenesis is compounded by the large number of tissues affected by the disease. While numerous studies have been conducted to understand the molecular mechanism of the disease and to test therapeutic strategies, the limited availability of patient-derived cell lines poses a challenge to research. Fibroblasts, which are easily obtained from a skin biopsy, are preferred for cell culture studies over myoblasts, which require more invasive muscle biopsies. However, fibroblasts have not been extensively studied as DM1 cell models and further characterization is needed to identify appropriate biomarkers [10–13].

Current methods for characterizing cell models and assessing therapeutic response focus on quantifying foci and alternative splicing defects, but there is no consensus on the protocols for this assessment and results can vary between research groups. Although technical developments in recent years have allowed more automated analysis of foci and improved splicing analysis through transcriptomics, a simple and straightforward platform could accelerate treatment screening. Furthermore, it remains unclear which levels of foci reduction and splicing improvement analysed in vitro are needed to improve a DM1 patient phenotype. While some therapeutic approaches have progressed to clinical trials, only a few patients have experienced symptomatic improvement. With many small molecules and ASOs available for further testing, rapid in vitro evaluation of different therapeutic approaches is essential for the development of effective treatments [14–18].

In light of all this, in this work we have optimized a platform based on in-cell western (ICW), a quantitative immunofluorescence assay that enables direct protein quantification in cell cultures, integrating the specificity of western blots with the efficiency and scalability of an ELISA. Indeed, it is a versatile technique with multiple applications in different fields [19, 20]. In the field of neuromuscular disorders, our group previously developed a myoblast/myotube ICW (or myoblot) to quantify dystrophin [21], and myoblots have since been used to quantify dystrophin, utrophin, and other proteins after exon skipping experiments in Duchenne muscular dystrophy cultures [22], and to characterize cell models newly generated by gene edition [23]. Thus, we built on this expertise to quantify DMPK and MBNL1 expression in DM1 cell models, characterize these models, and identify useful biomarkers for in vitro drug screening. We complemented the protein validation of this platform with the absolute quantification of corresponding RNAs by digital droplet polymerase chain reaction (ddPCR). In this manuscript, we describe this platform, we use it to characterize a wide range of cell culture models, and we validate it further by testing two potential DM1 therapeutic strategies: a DMPK gapmer previously described [24] and a DMPK dsiRNA.

Materials and methods

Myoblasts cultures

All cell cultures are detailed in Table 1. Immortalized myoblasts were immortalized by the MyoLine platform (Paris) and kindly provided by Dr. Furling (Institut of Myology) and Dr. Nogales-Gadea from Germans Trias I Pujol Research Institute (IGTP). DM-ImM-1 culture was previously characterized [25], and (CTG)n repeats were sized by Southern blot (see Table 1). DM-ImM-2, DM-ImM-3, and DM-ImM-4 cultures were characterized in [26]. (CTG)n repeats were evaluated by small-pool PCR followed by Southern blot. In this paper, DM-ImM-2 corresponds to JCC-DM1, DM-ImM-3 to GPM-DM1, and DM-ImM-4 to ADE-DM1 in [26]. Primary myoblasts were purchased from Cook MyoSite (USA) which also provided with the range of (CTG)n repeats, quantified in the patient’s blood samples by Southern blot.

Table 1.

Myoblasts and fibroblasts cultures used in this publication

Cell line	Status	Sex	Age of biopsy	MIRSa	(CTG)n	Biopsy tissue	Origin
Immortalized myoblasts
CTRL-ImM-1	Control	Female	12			Paraspinal	MyoLine
CTRL-ImM-2	Control	Male	25			Gracilis	[26]
CTRL-ImM-3	Control	Male	38			Quadriceps	[26]
CTRL-ImM-4	Control	Male	16			Paravertebral	[26]
DM-ImM-1	DM1	Female	11	–	2600	Gastrocnemius	[25]
DM-ImM-2	DM1	Female	36	2	875–1950b	Biceps	[26]
DM-ImM-3	DM1	Female	46	1	581–1028b	Biceps	[26]
DM-ImM-4	DM1	Female	39	4	1505–3075b	Biceps	[26]
Primary myoblasts
CTRL-PM-1	Control	Male	38			Vastus lateralis	Cook Myosite
CTRL-PM-2	Control	Female	17			Vastus lateralis	Cook Myosite
CTRL-PM-3	Control	Male	21			Vastus lateralis	Cook Myosite
DM-PM-1	DM1	Female	18	1	90–100	Vastus lateralis	Cook Myosite
DM-PM-2	DM1	Male	43	2	200–350	Vastus lateralis	Cook Myosite
DM-PM-3	DM1	Male	19	2	230–330	Vastus lateralis	Cook Myosite
Primary fibroblasts
CTRL-PF-1	Control	Male	48			Skin	[27]
CTRL-PF-2	Control	Female	48			Skin	[27]
CTRL-PF-3	Control	Male	27			Skin	[27]
DM-PF-1	DM1	Male	56	3	588–1378	Skin	[27]
DM-PF-2	DM1	Female	45	2	243–1425	Skin	[27]
DM-PF-3	DM1	Male	56	3	671–1287	Skin	[27]
Myo-inducible immortalized fibroblasts
CTRL-ImF-1	Control	Male	16			Skin	Myoline
CTRL-ImF-2	Control	Male	16			Skin	Myoline
DM-ImF-1	DM1	Female	11	–	2600	Skin	Myoline
DM-ImF-2	DM1	Male	29	–	600–1600	Skin	Myoline

^aMuscular Impairment Rating Scale or MIRS describes muscle affectation in DM1 patients at biopsy age. Values range from 0 (no muscle impairment) to 5 (severe proximal muscle affectation) [53].

^bThe two most frequent (CTG) n values, defining the mode range, are represented.

All myoblast cell lines were maintained in growth medium, consisting of 1:1 Skeletal Muscle Cell Growth Medium (SMM, PELOBiotech, Germany) and Dulbecco’s Modified Eagle medium (DMEM, Corning^®, USA) supplemented with 10% foetal bovine serum (FBS), 2% GlutaMax, and 1% Penicillin/Streptomycin (PenStrep). All supplements were purchased from Gibco^™ (USA). Cell cultures were only used up to passage 10, to avoid changes in expansion size due to cell passage. Myoblast cultures were incubated at 37ºC in 5% CO₂ and medium was refreshed every 48–72 h. For myogenic differentiation, myoblasts were seeded in growth medium in plates coated with 1% Matrigel (Corning^®) until they reached 80%–90% confluence, usually after 24–48 h. Medium was then replaced with differentiation medium (DM). DM was prepared with high-glucose DMEM containing GlutaMax (Gibco^™), 1% PenStrep, 1% KnockOut^™ Serum Replacement (Gibco^™), and 1% insulin-transferrin-selenium-ethanolamine (ITS-X from Gibco^™). Myoblasts were cultured in DM until myotubes were observed: typically, they were harvested on day 6 for mRNA studies, and on day 7 for protein studies: pelleted for quantification by Jess simple western and fixed for ICW, FISH, and immunofluorescence.

Fibroblasts cultures

Primary fibroblasts (see Table 1) were kindly provided by Dr. Lopez de Munain (Biogipuzkoa HRI). Muscular impairment rating scales (MIRS) were evaluated at age of biopsy and CTG expansions were sized by small-pool PCR and Southern blot [27]. Myo-inducible immortalized fibroblasts had been generated by the MyoLine platform (Paris) and were kindly provided by Dr. Furling (Institut of Myology). DM-ImF-1 and DM-ImM-1 were derived from biopsies from the same patient. DM-ImF-2 repeat size was measured by small-pool PCR and Southern blot.

Primary and immortalized fibroblast cultures were maintained in DMEM (Corning^®) supplemented with 10% FBS (Gibco^™), 1% PenStrep (Gibco^™), and 2% GlutaMax (Gibco^™). They were incubated at 37ºC in 5% CO₂ and the medium was refreshed every 48–72 h. For MyoD induction of myo-inducible immortalized fibroblasts, cells were maintained in fibroblast medium until cell passage, then, they were resuspended in DM containing 0.02% of doxycycline (Merck, USA) and cultured at 37ºC and 5% CO₂ for 5 days. The medium was refreshed every 48 h, and doxycycline was added to the medium just before it was added to the cells.

ASOs treatment

Two different kinds of ASOs were used, dsiRNAs and gapmers. All were purchased lyophilized from Integrated DNA Technologies (IDT, USA). DMPK13.1 dsiRNA and its scramble control were resuspended in duplex-buffer as per manufacturer’s instructions, while gapmers ASOs were resuspended in MiliQ water in sterile conditions. Sequences are detailed in Table 2.

Table 2.

Antisense oligonucleotides used in this publication

Treatment	Sequence (5→3)	Dose
Dicer-substrate short interfering RNA
DMPK13.1 dsiRNA	UrGrCrArArArGrCrUrUrUrCrUrUrGrUrGrCrArUrGrArCGC rGrCrGrUrCrArUrGrCrArCrArArGrArArArGrCrUrUrUrGrCrArCrU	0.1, 1, and 10 nM
Scramble dsiRNA	Sequences not provided	0.1, 1, and 10 nM
RNAseH1-mediated gapmers
DMPK gapmer	rC*rG*rG*rA*rG*C*G*G*T*T*G*T*G*A*A*rC*rU*rG*rG*rC	10 nM and 100 nM
Scramble gapmer	rG*rA*rC*rG*A*C*G*A*C*G*A*C*G*A*rC*rG*rA*rC	10 nM and 100 nM

r indicates RNA, if not present, DNA. Underline indicates 2′O-Me modifications and * are phosphorothioate linkages.

All ASOs were transfected into myoblasts using Lipofectamine 3000 (Invitrogen^™, USA) 24 h after seeding, at 70%–80% confluence, following the manufacturer’s protocol. Fresh growth medium was added 4 h after transfection and was replaced with DM when cultures reached 80%–90% confluence, typically 24–48 h after transfection. Myoblasts were then incubated at 37ºC in 5% CO₂ for 6 or 7 days until myotubes were visually observed. Medium was refreshed with fresh DM every 2–3 days until harvest or fixation.

Fluorescent in situ hybridization

Fluorescent in situhybridization (FISH) is the standard method used to visualize foci in cultures [28]. Briefly, cells were fixed with 10% formalin solution (Sigma) for 10 min followed by three washes of 5 min with PBS. Then, cells were pre-hybridized with 30% formamide in 2xSSC for 10 min at RT. Meanwhile, an hybridization buffer was prepared containing 40% formamide, 10% saline sodium citrate buffer (SSC) 20×, 20% bovine serum albumin (BSA) 1%, 0.1g/ml dextran sulphate, 10% vanadyl complex 20 mM, 10% yeast tRNA (10 μg/ml), 10% herring sperm DNA, and a (CAG)₇-Cy3 probe (1/100) purchased from IDT. The samples were hybridized overnight at 37ºC. The following day, samples were washed twice with 2× SSC in 30% formamide for 15 min at 42ºC in an incubation oven. The hybridized samples were then used for immunocytochemistry.

Immunocytochemistry

All samples (hybridized samples from FISH or fixed samples for direct immunofluorescence) were permeabilized with 0.1% Triton X-100 in PBS and blocked with Intercept^® (PBS) Blocking Buffer (LI-COR) for 2 h. After blocking samples were incubated overnight with primary antibodies at 4ºC at the appropriate dilutions: MB1a (The MDA Monoclonal Antibody Data Source), anti-desmin (Abcam), or anti-DMPK (HPA007164 from Sigma). The following day, coverslips were washed with 0.1% Tween in PBS and incubated with a secondary antibody mixture [Alexa Fluor^™ 647 goat anti-Rabbit IgG (H + L) Secondary Antibody and Alexa Fluor^™ 488 goat anti-Mouse IgG (H + L) Secondary Antibody]. All secondary antibodies were purchased from Life Technologies. Coverslips were washed again with 0.1% Tween in PBS and incubated for 10 min with Hoescht 3342 (at 1/1000) from Life Technologies. Coverslips were then mounted with ProLong^™ Diamond Antifade Mountant. Images were captured using a Zeiss LSM 880 ID SCAN microscope at ×40 for DMPK antibody screening and ×63 for foci and MBNL1 quantification and analysed using ZEN black software and Fiji.

In-cell western

For fibroblasts analysis, 4000 fibroblasts/well were seeded into a well of 96-well plates. Cells were cultured for 72 h prior to fixation. For myoblasts and myotubes, 7500 cells/well were seeded into a well of 96-well plate. Myoblasts were cultured for 48 h prior to fixation for myoblast protein detection. For myotube differentiation, myoblasts were cultured for 24–48 h before the addition of differentiation medium until they reached 80%–90% confluence. Myotubes were then fixed on day 7 of differentiation. Myo-inducible fibroblasts were cultured for 5 days prior to fixation, with the differentiation medium replaced with fresh doxycycline every 48 h. Cells were fixed with 10% formalin solution for 10 min and three washes of 5 min with PBS 1×. After fixation, plates were permeabilized with 0.1% Triton X-100 in PBS and blocked for 2 h with Intercept^® (PBS) Blocking Buffer (LI-COR). Primary antibodies were incubated overnight at 4ºC at the respective dilutions (see “Immunocytochemistry” section). The following day, plates were washed with 0.1% Tween in PBS and incubated with the corresponding secondary antibody mixtures consisting in Cell Tag 700 Stain and either IRDye^® 800CW Goat anti-Mouse IgG Secondary Antibody or IRDye^® 800CW Goat anti-Mouse IgG Secondary Antibody. Secondary antibody dilutions were described in the manufacturer’s protocol. All reagents were supplied by LI-COR Biosciences. Plates were washed again with 0.1% Tween in PBS and 200 μl/well of PBS were added before scanning the plates in an Odyssey^® M Imager (LI-COR Biosciences). Data were then obtained with LI-COR Acquisition Software and analysed with Empiria Studio^® Software before statistical analysis.

Jess simple western

Protein was extracted from myotube cultures following RIPA extraction and lysis buffer (Thermo Scientific^™, USA) supplemented with protease and phosphatase inhibitor cocktails (Roche, Switzerland) following manufacturer’s protocol and later quantified using a Pierce^™ BCA Protein Assay Kit (Thermo Scientific^™) following the manufacturer’s protocol. Quantification was performed in an Infinite^® 200 PRO plate reader (TECAN) at 562 nm. Lysate concentrations and antibody dilutions for DMPK targeting were optimized after assessing the linear range of detection of three lysate concentrations (2, 0.4, and 0.08 mg/ml) at three antibody dilutions (1:10, 1:50, and 1:250). For DMPK immunodetection, 3 μl of 0.4 mg/ml sample were loaded into a 12-220 kDa Separation Module (Bio-Techne, France), together with the appropriate antibodies, MANDM5 (1:10, The MDA Monoclonal Antibody Data Source) or anti-DMPK (1:50, HPA007164 from Sigma), and blocking buffer. To detect MBNL1, we followed the conditions used by Cerro-Herreros et al. 2024 [29]. 3 μl of 0.2 mg/ml sample were loaded into a 12-220 kDa Separation Module (Bio-Techne, France) together with the appropriate antibody, MB1a (1:50, The MDA Monoclonal Antibody Data Source), and blocking buffer. All reagents were purchased from Bio-Techne. The cartridge was then read on a Jess instrument and the data were analysed using Bio-Techne’s Compass SW software.

Digital droplet PCR

RNA was extracted from cell pellets using the RNeasy Micro Kit for pellets collected from 12-well plates and the RNeasy Mini Kit for pellets collected from 6-well plates, following manufacturer’s protocols (QIAGEN, Germany). Once extracted, RNA was quantified using a Nanodrop apparatus and 100 ng to 1 μg of the extracted RNA was used as template for reverse transcription with SuperScript^™ IV Reverse Transcriptase (Invitrogen^™), following manufacturer’s guidelines. The resulting cDNA was diluted to 0.05 ng/μl in nuclease-free water for the ddPCR protocol. For gene expression analysis, 2 μl of cDNA were used in 20 μl of PCR reactions, that included 10 μl of ddPCR^™ Supermix for Probes (No dUTP) from Bio-Rad, 7 μl of nuclease free-water, 0.5 μl of PrimeTime^™ qPCR Probe Assay (IDT) marked with FAM, 0.5 μl of PrimeTime^™ qPCR Probe Assay (IDT) marked with HEX. Each reaction included a PrimeTime^™ qPCR Probe Assay (IDT) targeting HPRT1 or TBP as loading control. Housekeeping genes in each case were selected based on expression levels. For splicing assessment, instead of PrimeTime^™ qPCR Probe Assays, 250 nM of specific primer pair, 125 nM of exon-IN probe marked with FAM and 125 nM of exon-OUT probe marked with SUN (Supplementary Table S1) were added.

To generate droplets, 20 μl of the prepared ddPCR reactions and 70 μl of the Droplet Generation Oil for Probes (Bio-Rad) were added to the 8-channel droplet generator cartridge (Bio-Rad) sealed with a DG8 gasket (Bio-Rad) and placed in the QX200 droplet generator (Bio-Rad). Forty microliters of the droplet mixture were collected from the cartridge and transferred to a 96-well semi-skirted ddPCR plate (Bio-Rad), sealed with a pierceable foil and amplified on a deep-well thermal cycler using the following conditions: enzyme activation for 10 min at 94ºC followed by 40 cycles of denaturation for 30 s at 94ºC, 1 min at 59ºC for annealing and extension, and heat deactivation for 10 min at 98ºC. Plates containing the amplified droplets were then analysed in the QX200 droplet reader. Results were analysed using QuantaSoft^™ Software and QuantaSoft^™ Analysis Pro (Bio-Rad) software. Droplet number was used as a quality control and wells with >10 000 droplets were included in the analysis.

Statistical analysis

All results were expressed as mean ± standard error of the mean (SEM). Statistical analysis was performed using GraphPad Prism 10.1.2 software. Data distribution was assessed using the Shapiro–Wilk test. If the normality test was passed, outliers were detected using the ROUT method (Q = 1%) and heteroscedasticity was assessed using the Fisher test. Normal data with equal variance was assessed by parametric method, one-way analysis of variance (ANOVA) followed by Bonferroni’s multiple correction test. Data that did not follow normal distribution or equality of variances was analysed by non-parametric tests, Kruskal–Wallis test followed by Dunn’s post hoc test. P-values used in this study to determine statistical significance where as follows: *P-value < 0.05, **P-value < 0.01, ***P-value < 0.001, ****P-value < 0.0001.

Results

Optimization of in-cell western protein quantification

An ICW assay is a quantitative immunofluorescence assay performed in 96-well plates, as in our case, or 384-well plates. Cells are seeded in a 96-well plate and fixed, permeabilized and stained in the same well. The protein is thus detected in situ, in its natural conformation and intracellular location, without the need for denaturing (Fig. 1). Protein quantification by ICW assays requires accurate cell seeding density, which needs to be within a linear range and permits the quantification of cell density in the same well as target protein. To define the linear range of detection of the cell number stain used for this purpose, we seeded different cell densities of control-derived immortalized myoblasts, and we stained with Cell Tag 700 stain (Supplementary Fig. S1).

Figure 1.

ICW workflow and experiment design. (A) Schematic representation of cell culture and ICW procedure for protein quantification in fibroblasts and myotubes. (B) Example of plate set-up for cell culture characterization or ASO testing at three different concentrations (C1, C2, and C3). No treatment refers to cells no subjected to transfection and lipo only refers to cell transfected without ASO (negative control). Created in BioRender. Arechavala, V. (2025) https://BioRender.com/mohbb6n

As Odissey M scan, used for the fluorescence detection of ICW assays, do not have the resolution of microscopy nor the possibility of confirming the size of the protein detected, antibodies used need to be thoroughly validated.

While DMPK transcript levels have previously been assessed in DM1 to understand the role of DMPK-expanded transcripts in DM1 pathogenesis, fewer studies have focused on quantifying DMPK protein in cell culture [30, 31]. Difficulties in assessing the role of DMPK protein implications in different features of the disease are due in part to the lack of validated antibodies [32]. Several DMPK antibodies with different targeting epitopes are commercially available, but there is no consensus on which are the best ones for targeting DMPK in cell culture. Therefore, we tested different DMPK antibodies on control myotubes to perform a parallel validation of their targets. After discarding some antibodies for unspecific binding, we evaluated HPA007164 from Sigma and MANDM5 from the Monoclonal Antibody Database from the Wolfson Centre for Inherited Neuromuscular Disease (CIND). Both showed the expected values described in the literature for DMPK recognition: correct localization under the microscope (cytoplasm, endoplasmic reticulum, and mitochondria) [33] (Fig. 2A), and a size, when evaluated by Jess simple western blotting, of 82 kDa , consistent with the molecular weight reported in the literature (79 kDa) [34] (Fig. 2C). We selected the Sigma as it showed a more robust signal. The complete Jess simple western run is represented in Supplementary Fig. S2.

Figure 2.

Antibody selection for DMPK and MBNL1 quantification. (A) Immunofluorescent staining of control immortalized myotubes with two DMPK-targeting antibodies: CTRL-ImM-1 myotubes differentiated for seven days were stained with HPA007164 (Sigma) and MANDM5 DMPK antibodies (green), nuclei (blue), and desmin (red) or MHC (red) showing the expected localization. (B) Immunofluorescent staining of control (CTRL-ImM-1) and DM1 immortalized myotubes (DM-ImM-1) with MBNL1 antibody MB1a and (CAG)₇ probe. CUG foci are seen as small red points in DM1 cultures, MBNL1 is stained green, desmin is cyan, and nuclei are blue). (C) Representative image of Jess simple western lane visualization of DMPK-targeting antibodies in CTRL-ImM-1 and MB1 antibody targeting MBNL1 in CTRL-ImM-1 and DM-ImM-1 myoblasts. (D) MB1a normalized data by Cell Tag 700 stain signal in CTRL-ImM-1 and DM-ImM-1 immortalized myoblasts after fixation with different methods. Data are represented as mean ± SEM of three technical replicates per experiment in three independent experiments. Significant differences were determined by two-way ANOVA and Bonferroni’s multiple comparison test (****P < 0.0001).

The quantification of MBNL1 protein, on the other hand, has been of great interest in the development of drugs aimed at increasing MBNL1 expression or releasing the protein from the DMPK-expanded transcripts. Detection of the MBNL1 protein signal is often performed by immunocytochemistry or western blot after treatment with potential therapeutic compounds to assess MBNL1 release from foci and to study treatment efficacy [35, 36]. The most commonly used antibody, MB1a (clone 4A8) from the Wolfson Centre for Inherited Neuromuscular Diseases (CIND) monoclonal antibody database, targets an epitope encoded by exon 3 of the MBNL1 protein, which is constitutively expressed in all MBNL1 isoforms [37]. We corroborated the specificity of this antibody testing it on DM1 and control immortalized myotubes by FISH and immunocytochemistry and Jess simple western, as represented in Fig. 2B and C (respectively). It showed the expected distribution of MBNL1, co-localizing with RNA foci, which was also detected by a (CAG)₇ probe in the nucleus (Fig. 2B). This antibody was used to analyse protein extracts from both control and DM1 cultures by Jess and it showed a band at the expected molecular weight, 48 kDa (Fig. 2C) [38].

As the entrapment of MBNL1 in foci is at the core of the main pathogenesis theories in DM1 and its release is the mechanism of several potential treatments, we were interested to know if we were able to measure free or “soluble” MBNL1 by ICW. ICW relies on standard immunohistochemical techniques but lack the resolution of confocal microscopies. Recently, it was suggested that fixation of DM1 and control cell lines with 4% PFA (paraformaldehyde) or 50% methanol–50% acetone could affect MBNL1 protein detection [39]: methanol–acetone fixated cells show a preference for insoluble MBNL1 binding to DMPK-expanded transcripts, whereas PFA-fixed cells show a more dispersed pattern, corresponding to insoluble MBNL1, but also soluble MBNL1 in the nucleus, which is also present in control cell lines [39]. We, therefore, decided to test this hypothesis and compared MBNL1 distribution by ICW when DM1-ImM-1 and CTRL-ImM-1 immortalized myotube cultures where fixed with formalin, ice-cold methanol (methanol) and 50% methanol–50% acetone (methanol–acetone) fixation methods (Fig. 2D). The resolution of the ICW data, obtained by Odissey M plate scanner, is 100 μm, which defines the distance by which the instrument distinguishes between two points of signal intensity. As shown in Fig. 2B, the cell nucleus is ∼20 μm, suggesting that we are unable to distinguish different points within the nucleus, i.e. foci. The intensity of the foci is dispersed by the intensity of the surroundings, suggesting that the MBNL1 signal detected by ICW reflects the diffuse MBNL1 signal observed in control and DM1 myotubes and not the foci-specific signal from DM1 cultures. Quantification by ICW using the Odissey M scan shows statistically significant differences between control and DM1 samples for all fixation methods. However, the signal is higher in formalin-fixed wells compared to methanol and methanol–acetone wells, suggesting that we are detecting the diffuse MBNL1 signal previously observed after 4% PFA fixation [39]. Lower levels of diffuse MBNL1 are observed when fixed with methanol and methanol–acetone, similar to methanol–acetone fixed cells in [39].

Validation of the platform on immortalized control and patient cultures

Once the ICW methodology was optimized, it was combined with absolute RNA quantification by digital droplet PCR and this platform was used to characterize control and DM1 patient cultures described in [26] by assessing DMPK and MBNL1 expression at RNA and protein levels (Fig. 3).

Figure 3.

Quantification of DMPK and MBNL1 protein in control and DM1 cell cultures. DMPK (A), MBNL1 (C), and MYH3 (E) transcripts quantification in control and DM1-derived immortalized myotubes. Data are represented as mean ratio ± SEM between DMPK or MYH3 transcripts and HPRT1 transcripts (as loading control) or MBNL1 transcripts and TBP transcripts of three technical replicates per experiment in a single experiment. DMPK (B), MBNL1 (D), and MHC (F) relative protein quantification in immortalized control and DM1-derived immortalized myotubes by ICW assay. Sigma, MB1a and anti-MHC (MF20) antibody signal is normalized by Cell Tag 700 stain signal in each well. Data are represented as mean ± SEM of three technical replicates per experiment in three independent experiments. Significant differences were determined by two-way ANOVA and Bonferroni’s multiple comparison test (**P < 0.01, ***P < 0.001).

While no differences were observed at transcript level for DMPK and MBNL1 (Fig. 3A and C, respectively), DMPK and MBNL1 protein signal was lower in DM1 immortalized myotubes compared to control cultures (Fig. 3B and D). To account for any differences in differentiation, MHC was also assessed by digital PCR and ICW, and, although differences were found, these were not significant. As described in Nuñez-Manchón et al. (2024), myoblast differentiation into myotubes is variable between different cell lines, despite being control or DM1 patient-derived cell cultures, primary or immortalized. Therefore, in this paper no statistically significant differences in MHC quantification were observed between control and patient-derived cell cultures, although variability is observed. The characterization of these cultures is consistent with the sequestration of MBNL1 protein in DMPK-expanded transcripts and the splicing alterations previously described by [26].

DMPK and MBNL1 quantification in control and DM1-patient derived cultures

Several cell models have been used to understand the pathogenic mechanisms of DM1 and to test potential therapeutic compounds, with immortalized myotubes [26, 38, 40] and primary fibroblasts [41–43] being the most commonly used cell models. We aimed to assess these models, to evaluate their adequacy to evaluate new therapies.

In addition to the six immortalized cultures characterized in the previous figure (Fig. 3), we extended our study to three DM1 and three control primary myoblast cultures purchased from Cook MyoSite, two DM1 and two control immortalized fibroblasts and three DM1 and control-derived primary fibroblasts described in [27]. The differences are summarized in Fig. 4 and the individual results are included in Supplementary Fig. S3 (immortalized myoblasts), 4 (primary myotubes), 5 (immortalized fibroblasts), and 6 (primary fibroblasts).

Figure 4.

Summary of the differences in quantification of DMPK and MBNL1 between DM1 and controls, in all the cultures characterized in this project. While no statistical differences were found at RNA levels between DM1 and controls, protein expression in myogenic lines is consistently downregulated in the DM1 patients, and ICW can pick up early downregulation of MBNL1 in primary myotubes. Created in BioRender. Arechavala, V. (2025) https://BioRender.com/yrbi56u.

DMPK protein signal was lower in DM1 immortalized myoblasts and myotubes compared to controls, while no difference was observed in primary myotubes nor immortalized or primary fibroblasts. On the other hand, MBNL1 protein signal was consistently lower in all DM1 myoblast and myotube cultures (primary and immortalized), whereas it was not altered in DM1 fibroblast cultures (neither primary nor immortalized). DMPK and MBNL1 mRNA transcripts were unaltered in all cases, suggesting that the observed differences are downstream from mRNA transcription.

The importance of differentiation in the expression of DMPK and MBNL1

To further understand the differences described between the cell models evaluated, we assessed DMPK mRNA expression in control primary fibroblasts and immortalized fibroblasts compared to immortalized myoblasts. While still undifferentiated, immortalized myoblasts show 3–10 times more DMPK mRNA transcripts compared to fibroblasts (Fig. 5A) and, when these myoblast cultures were differentiated, and evaluated at different time points (Fig. 5B), DMPK mRNA increased over time. We hypothesized that DMPK transcripts are physiologically increased in myoblasts models compared to fibroblasts, so we evaluated control and DM1 myo-inducible fibroblasts with and without MyoD induction: we observed how the induction of MyoD alone was enough to increase DMPK transcripts in these cultures (Fig. 5C). However, DMPK protein was only downregulated in the patient’s DM1 myo-inducible cultures (Fig. 5D), suggesting an increase on the expanded transcript expression and therefore, lower protein translation levels.

Figure 5.

DMPK and MBNL1 expression and myogenic differentiation. Expression of DMPK (A and B) and MBNL1 (E and F) was evaluated by ddPCR in control fibroblasts (primary and immortalized) compared to immortalized myoblasts before (A and E) or during myogenic differentiation of the myogenic cultures (B and F). (C) DMPK mRNA and (D) protein expression in control myo-inducible fibroblasts with and without MyoD induction assessed by ddPCR and ICW. Transcript data are represented as mean ratio ± SEM between DMPK transcripts and HPRT1 transcripts (loading control) of two technical replicates per experiment in three independent experiments (A and E) and two technical replicates per experiment in two independent experiments (C). Protein data are represented as mean ± SEM of four technical replicates per experiment in three independent experiments. Anti-DMPK antibody signal (800 nm) is normalized by Cell Tag 700 stain signal (700 nm). Significant differences were determined by two-way ANOVA and Bonferroni’s multiple comparison test (****P < 0.0001). Quantification of MBNL1 aggregates (G) and foci (H) in DM1 myo-inducible fibroblasts with and without MyoD induction assessed by FISH and immunocytochemistry. Significant differences were determined by Kruskal–Wallis test followed by Dunn’s post hoc test (*P < 0.05, **P < 0.01, ****P < 0.0001).

The increase in DMPK mRNA expression seen in myoblasts versus fibroblasts in Fig. 5A was accompanied by a slight but statistically significant reduced level of MBNL1 mRNA in myoblasts compared to fibroblasts (Fig. 5E). The MBNL1 mRNA expression showed no change during the myoblast maturation process (Fig. 5F).

The upregulation of DMPK after MyoD induction suggested an increased sequestration of MBNL1 protein in expanded transcripts and, to evaluate it, we quantified MBNL1 aggregates (Fig. 5G) and foci (Fig. 5H) in myo-inducible DM1 fibroblasts before and after MyoD induction. Indeed, MBNL1 aggregates and foci increased after MyoD induction, suggesting that the induction increased the number of DMPK transcripts, including the expanded transcripts, and therefore caused higher MBNL1 sequestration and accumulation in foci.

DMPK and MBNL1 quantification after treatment with DMPK-targeting antisense oligonucleotides

Most therapies targeting DM1 aim to decrease the toxicity of the DMPK repeats by decreasing DMPK transcripts (expanded and wild-type) or by avoiding the binding of MBNL1 protein to the expanded transcripts [9]. However, in most cases there’s no certainty if a potentially therapeutic compound is acting on the wild-type or the expanded transcripts as, unless a specific and uncommon SNP is found in the repeated allele, it is not possible to discriminate between both alleles [44]. DMPK expanded transcripts are thought to be retained inside the nucleus and not transferred into the cytoplasm for further translation, as opposed to WT transcripts [45]. Therefore, we hypothesized that by quantifying DMPK protein, we could assess how the DMPK wild-type transcript is affected by DM1 potential treatments.

For proper validation of DMPK quantification, we assessed by ddPCR and ICW a dsiRNA targeting a DMPK region in exon 15 (before the 3′UTR region) (Fig. 6A).

Figure 6.

DMPK dsiRNA response quantification after differentiation. (A) Representation of DMPK siRNA targeting in the DMPK transcript. Created in BioRender. Arechavala, V. (2025) https://BioRender.com/y26xt54 (B) Quantification of DMPK transcripts by ddPCR 6 days after transfection of DM1-ImM-2 myotubes with the DMPK13.1 dsiRNA and its corresponding scramble control at increasing concentrations. Data from three technical replicates per experiment in two independent experiments is represented as mean ratio ± SEM between DMPK transcripts and HPRT1 transcripts (as loading control). Quantification by ICW of DMPK (C) and MBNL1 protein (D) 7 days after transfection of DM1-ImM-2 myotubes with the DMPK13.1 dsiRNA and scramble controls at three concentrations. Anti-DMPK and MB1a antibody signals (800 nm) are normalized by Cell Tag 700 stain signal (700 nm). Data from four technical replicates per experiment in a single experiment are represented as mean ± SEM. Significant differences were determined by two-way ANOVA and Bonferroni’s multiple comparison test (*P < 0.05, ***P < 0.001). (E) Exon inclusion quantification by ddPCR of MBNL1, DMD, and SOS1 transcripts in control myotubes and DM1 myotubes at day 6 after transfection of scramble or DMPK dsiRNA at 10 nM (both). Data are presented as mean ± SEM of two technical replicates in a single experiment. Significant differences were determined by Kruskal–Wallis test followed by Dunn’s post hoc test (*P < 0.05). While DMPK expression in reduced by the dsiRNA both at RNA and protein levels, with a clear dose– response at protein levels, there is no difference in the amount of MBNL1 protein quantified or in the splicing assessment.

We are able to quantify a decrease in DMPK mRNA (Fig. 6B) and protein level (Fig. 6C) in DM1 immortalized myotubes after siRNA treatment, but this was not accompanied by a change in the signal of MBNL1 protein (Fig. 6D) or a correction of the assessed splicing alterations (Fig. 6E).

We also evaluated a gapmer targeting DMPK previously described in [24] (Fig. 7). This DMPK gapmer ASO was designed to promote the degradation of DMPK transcripts through RNase H1 by targeting a sequence in exon 15 of DMPK mRNA, the 3′UTR region, nearby the region targeted by the siRNA evaluated in Fig. 6. In their publication (Fig. 7A), 24–48 h after treatment of the cultures with the DMPK gapmer at 100 nM, the authors observed by qPCR a decrease in DMPK transcripts, by immunohistochemistry a decrease of MBNL1 density in foci, and an improvement of overall missplicing. We completed their evaluation by quantifying DMPK transcripts by ddPCR and DMPK and MBNL1 protein by ICW. Our quantification aligns with their results and shows a decrease in DMPK transcripts at all concentrations (Fig. 7B), a moderate decrease in DMPK protein signal (Fig. 7C) and an increase of up to twice the amount of MBNL1 protein, compared to the scramble control (Fig. 7D), possibly indicating the release of MBNL1 protein from foci. This hypothesis is supported by the partial improvement of the three splicing alterations assessed (Fig. 7E).

Figure 7.

DMPK gapmer effect on DM1 myoblasts and myotubes. (A) Representation of DMPK gapmer targeting in the DMPK transcript and summary of the response to treatment with 100 nM of this gapmer of DM-ImM-1 myoblasts as previously reported by [24]. Created in BioRender. Arechavala, V. (2025) https://BioRender.com/p39kat4 (B) Quantification by ddPCR of DMPK transcripts 6 days after transfection of DM1-ImM-1 myotubes with the DMPK gapmer and its corresponding scramble control at increasing concentrations. Data of three technical replicates per experiment in a single experiment are represented as mean ratio ± SEM between DMPK transcripts and HPRT1 transcripts (as loading control). (C) Quantification by ICW of DMPK and MBNL1 protein (D) 7 days after transfection of DM1-ImM-2 myotubes with the DMPK gapmer and scramble control at different concentrations. Anti-DMPK and MB1a antibody signals (800 nm) are normalized by Cell Tag 700 stain signal (700 nm). Data of three technical replicates per experiment in two independent experiments is represented as mean ± SEM. Significant differences were determined by two-way ANOVA and Bonferroni’s multiple comparison test (*P < 0.05, **P < 0.01, ****P < 0.0001). (E) Exon inclusion quantification by ddPCR of MBNL1, DMD, and SOS1 transcripts in control myotubes and DM1 myotubes at day 6 after transfection of scramble or DMPK gapmer at 100 nM (both). Data are presented as mean ± SEM of two technical replicates in a single experiment. Significant differences were determined by Kruskal–Wallis test followed by Dunn’s post hoc test (*P < 0.05). The DMPK gapmer was succesful at both reducing DMPK expression and increasing the amount of MBLN1 protein detected, in a dose-dependent manner.

Discussion

Myotonic dystrophy type 1 (DM1) is a multisystemic disease with molecular dysfunctions affecting various tissues. Research is hindered by the limited availability of patient cell lines: fibroblasts, more easily obtained than myoblasts, are preferred for their accessibility and robust culture conditions, and, although they might reproduce some of the splicing defects that are hallmarks of the disease, other hallmarks may not be easily found in this model. Foci, containing RNA repeats and MBNL1 protein, are a key element of DM1 pathology and are widely used as an outcome measure by many research groups. Therefore, proper detection of RNA and MBNL1 foci is critical in DM1 research. Many therapies in DM1 research target DMPK transcripts and aim reducing MBNL1 binding and therefore, reducing MBNL1 sequestration by either reducing the expanded DMPK transcripts or displacing MBNL1 from them. Being able to quantify both the DMPK transcripts and free MBNL1, that is, the MBNL1 protein that it is not trapped in the expanded transcripts, particularly in a high-throughput format is a useful tool to deepen in the biology of the disease, characterize the models used in the field and screen for potential treatments.

We have developed a platform for quantifying DMPK and MBNL1 expression in various cell models, with the goal of identifying biomarkers for in vitro drug screening and enhancing therapeutic development for DM1. With this methodology, we have characterized a large cohort of cell models widely used in the field, including immortalized and primary myoblasts, myotubes, and fibroblasts, to explore how DM1 may alter DMPK and MBNL1 expression in these cultures. DMPK protein levels were reduced in immortalized myoblasts and myotubes but remained unchanged in primary myotubes and fibroblasts. The lack of DMPK depletion in primary myotubes might be due to the longer expansions in the DM1 immortalized cultures versus the primary ones. It is likely that transcript sequestration is more prominent in samples where there are either longer expansions or a larger amount of these mRNAs, as seen in myotubes, where longer repeat lengths might exacerbate transcript sequestration [46]. Furthermore, DMPK mRNA levels were higher in immortalized myoblasts compared to fibroblasts, and these transcripts increased during myogenic differentiation. Additionally, in myo-inducible models, protein levels did not correlate with the increase in DMPK transcripts, especially in DM1 cells, where DMPK protein levels were reduced, suggesting higher transcript sequestration.

Similarly, we also observed a reduction in MBNL1 protein levels in DM1 myoblasts and myotubes compared to controls, consistent with previous reports. No difference was observed between DM1 and control fibroblasts. Interestingly, MBNL1 protein levels were lower in myoblasts compared to fibroblasts and no changes were observed during myogenic differentiation. These findings suggested that DM1 myoblasts and myotubes may experience greater alterations due to the combined effects of higher DMPK transcript levels and reduced MBNL1 availability. Further analysis in myo-inducible models showed increased MBNL1 aggregates and more RNA foci in DM1 myotubes compared to fibroblasts, indicating higher accumulation of DMPK-expanded transcripts in muscle cells.

Our platform enabled the assessment of DMPK and MBNL1 as potential outcome measures for drug testing in DM1. The results indicated that myoblasts and myotubes were more suitable than fibroblasts for screening molecules targeting DMPK and MBNL1. Moreover, the quantification of DMPK mRNA and protein could distinguish between compounds that affect expanded versus non-expanded DMPK transcripts, providing insight into therapeutic mechanisms. Treatment with DMPK-targeting molecules, such as the gapmer and dsiRNA, showed different effects: the DMPK gapmer decreased DMPK transcripts without reducing protein levels but increased MBNL1 protein and improved splicing alterations, while the dsiRNA reduced both DMPK transcripts and protein, with no effect on MBNL1 or altered splicing. We hypothesize that these differences are due to the distinct mechanisms of action of both ASOs, with the gapmer acting at the pre-mRNA level in the nucleus to release MBNL1 from sequestration and the dsiRNA acting at post-transcriptional mRNA level in the cytoplasm [47, 48]. In DM1, siRNAs have previously been tested and have shown an effect on MBNL1 release [47, 49–51]. However, this effect is dependent on the delivery method, as siRNAs require an active transport (e.g. lentiviral transduction) or nuclear permeabilization to reach the nucleus [52]. In our hands, in the absence of active nuclear transport, we have only observed how DMPK silencing targets the DMPK wild-type transcript. Thus, reducing DMPK protein and not showing any effect on MBNL1 release, as it is not able to target the expanded DMPK transcript in the nucleus.

Regarding the differences observed in DMPK transcript knockdown between the DMPK gapmer and the DMPK siRNA, we hypothesize these are also due to their mechanism of action and the methods used for its evaluation. The DMPK gapmer targets both pre-mRNA and mRNA in the nucleus and cytoplasm, while with the dsiRNA, only targets mRNA in the cytoplasm [45]. As the probe used for amplification by ddPCR binds to the exon 8-9 span, it will only detect mRNA that has already undergone splicing. Therefore, the knockdown results from both the dsiRNA and the gapmer are not strictly comparable: all the molecules that the dsiRNA can target (mRNA) could potentially be detected, but not all the molecules that the gapmer can target (pre-mRNA and mRNA).

In conclusion, in this manuscript, we describe an effective platform for assessing DMPK and MBNL1 in cell culture, we characterize a large cohort of DM1 cell culture models widely used in preclinical research and corroborate the screening capabilities of our platform by confirming and completing the effect of previously reported potential therapies. We expect our results to be of use to the DM1 research community.

Supplementary data

Supplementary data is available at NAR Molecular Medicine online.

Conflict of interest

None declared.

Funding

This work was supported by by Instituto de Salud Carlos III (ISCIII, Spain) [PI18/00114] and Government of the Basque Country [2019111010 and 2022111014]. A.L.-M acknowledges funding from the FPU Program of Spanish Ministry of Science, Research and Universities [FPU20/00912]. V.A.-G. acknowledges funding from Ikerbasque (Basque Foundation for Science). G.N.-G acknowledges funding from “Consolidación investigadora” MCIN, grant CNS2022-135519 by MICIU/AEI/ 10.13039/501100011033 and, by the “European Union NextGenerationEU/PRTR.

Data availability

All relevant data can be found within the article and its supplementary information. All materials and further information of this study are available upon request.