Nanomechanics of cell-derived matrices as a functional read-out in collagen VI-related congenital muscular dystrophies

Abstract

Changes in the mechanical properties of the extracellular matrix (ECM) are a hallmark of disease. Due to its relevance, several in vitro models have been developed for the ECM, including cell-derived matrices (CDMs). CDMs are decellularized natural ECMs assembled by cells that closely mimic the in vivo stromal fibre organization and molecular content. Here, we applied atomic force microscopy-force spectroscopy (AFM-FS) to evaluate the nanomechanical properties of CDMs obtained from patients diagnosed with collagen VI-related congenital muscular dystrophies (COL6-RDs). COL6-RDs are a set of neuromuscular conditions caused by pathogenic variants in any of the three major COL6 genes, which result in deficiency or dysfunction of the COL6 incorporated into the ECM of connective tissues. Current diagnosis includes the genetic confirmation of the disease and categorization of the phenotype based on maximum motor ability, as no direct correlation exists between genotype and phenotype of COL6-RDs. We describe differences in the elastic modulus (E) among CDMs from patients with different clinical phenotypes, as well as the restoration of E in CDMs obtained from genetically edited cells. Results anticipate the potential of the nanomechanical analysis of CDMs as a complementary clinical tool, providing phenotypic information about COL6-RDs and their response to gene therapies.

1. Introduction

The extracellular matrix (ECM) is a three-dimensional and highly dynamic acellular component present in all tissues. It is formed by a variety of compounds including minerals and a mesh of macromolecules that are produced intracellularly by resident cells. While the ECM provides physical support and controls tissue homeostasis, its dysregulation results in a number of human conditions [1]. In particular, changes in the mechanical properties of the ECM are a hallmark of severe diseases such as pulmonary fibrosis and cancer [2,3], for which a chronic stiffening has been reported and associated with increased secretion of collagens and collagen cross-linkers [4].

Given the relevance of the ECM in disease progression, several in vitro models have been developed to study its biochemical and mechanical properties, ranging from the use of decellularized native tissues to the engineering of ECM analogues of different complexity. Decellularized tissue slices can provide a faithful mimic of tissue-specific architecture, yet they cannot be bioengineered. Alternatively, cell-derived matrices (CDMs), produced in vitro from patients’ cells, closely mimic the in vivo stromal fibre organization and molecular content of the ECM [5–8], as well as allowing for their controlled manipulation at different stages to produce tunable outcomes in matrix production [9]. In CDMs, cells can be genetically engineered to tailor their properties or they can be cultured on substrates with different topographies that are then transferred to the final matrices [10]. In this line, our group has previously published on the production of engineered anisotropic CDMs to study cell growth and directed three-dimensional cell migration [11,12]. More recently, we produced CDMs from patients with collagen type VI-related congenital muscular dystrophies (COL6-RDs) to study the morphology and organization of protein fibres in the ECM. In agreement with previous results obtained in collagen type VI-deficient CDMs from stable COL6A1 knockdown cell lines [13], we found that CDMs from COL6-RD patients presented alterations in structure and composition, showing a significant decrease in COL6 secretion, also affecting the organization of matrix proteins fibronectin and fibrillin-1 [14].

COL6-RDs comprise a heterogenous group of incurable and rare disorders that present different severities of clinical manifestation, spanning from the milder Bethlem myopathy (BM) to the severe Ullrich congenital muscular dystrophy (UCMD), with a range of intermediate cases [15–17]. Underlying all these forms are mutations in any of the three major collagen VI (COL6) genes, COL6A1, COL6A2 and COL6A3 [18]. These genes encode three α chains, α1(VI), α2(VI) and α3(VI), respectively, which assemble first to form a triple-helix monomer before further association to dimers and then tetramers via disulfide bond formation. These tetramers are finally secreted in the ECM producing an extended network of beaded microfilaments [19]. In COL6-RDs, more than 200 mutations in COL6 genes have been identified, predominantly affecting COL6A1 and COL6A2. Exon skips or point mutations proximal to the N-terminal region of the triple helix domain that still allow tetramer assembly and secretion result in the incorporation of aberrant COL6 into the ECM. Conversely, those mutations that lead to very low to no expression of the affected α-chains or that prevent tetramer formation, result in COL6 being almost absent in patients’ muscle and fibroblast cultures [18,20,21]. Current COL6-RD diagnosis includes the genetic confirmation of the disease and categorization of the phenotype based on maximum motor ability [22], as correlations between genotype and phenotype are still very difficult to identify [20]. Therefore, new tools able to recognize phenotype traits can significantly contribute to the diagnosis and prognosis of COL6-RDs.

COL6 was found fundamental in defining a healthy ECM with optimal stiffness. Stress-strain tests conducted in muscle tissue of a total knockout mouse model (Col6a1^−/−) resulted in a significant decrease in stiffness compared with the WT counterparts, which was attributed to the lack of COL6 in the ECM, compromising the in vitro and in vivo activity of satellite cells and affecting muscle regeneration [23]. Based on these previous results and taking advantage of the reductionist model provided by the CDMs, in the present work we use atomic force microscopy-force spectroscopy (AFM-FS) to evaluate the stiffness in CDMs produced from patients diagnosed with COL6-RDs. We further investigated the effects of gene editing on the mechanical and morphological properties of CDMs, demonstrating the potential of mechanical analysis of patient CDMs as a functional biomarker of COL6-RD progression and therapeutic response.

2. Experimental

2.1. Human fibroblasts

Forearm dermal fibroblasts were used in this study [24]. Dermal fibroblasts were obtained from the Hospital Sant Joan de Deu biobank for five representative COL6-RD patients covering a spectrum of mutations and phenotypes, and two controls unaffected by any neuromuscular condition (table 1). Primary cells were collected following a standard skin biopsy procedure. All patients had a genetically or pathologically confirmed diagnosis of COL6-RD, with their phenotypes subsequently categorized based on maximum motor ability [22]. Written informed consent was obtained from individuals and/or their parents or guardians. Biological samples were stored and managed by the Hospital Sant Joan de Déu (HSJD) Biobank.

Table 1.

Description of fibroblast donors included in this study with their relevant genetic background.

donor	phenotype	mutation	gene (pattern of inheritance)	sex	age at time of biopsy (years)	passage number
patient 1	BM	c.877G>A (Gly293Arg) Exon 10	COL6A1 (AD)	female	9	P9
patient 2	intermediate	c.877G>A (Gly293Arg) Exon 10	COL6A1 (AD)	male	8	P11
patient 3	intermediate	c.2329T>C (Cys777Arg) Exon 26	COL6A2 (AD)	female	12	P4
patient 4	UCMD	c.901−2A>G Intron 7	COL6A2 (AD)	female	3	P5
patient 5	UCMD	c.930+189C>T Intron 11	COL6A1 (AD)	male	4	P9
control 1	—	—	—	male	1	P9
control 2	—	—	—	male	2	P11

The present study was performed in accordance with the Declaration of Helsinki. All experimental protocols were approved by the Fundació Sant Joan de Déu Ethics Committee on Clinical Research (project identification code: PIC-24-23). Methods were carried out in accordance with relevant guidelines and regulations.

2.2. Genetically edited cells

Genetically edited cells from López-Márquez et al. [25] were used. In summary, CRISPR/Cas9-based genome editing was performed on dermal fibroblasts of two patients presenting either the BM (patient 1) or intermediate phenotype (patient 2), from passages P11 and P4, respectively. Two different CRISPR RNA guides, crRNA_1 (nucleotide sequence: 5’-CCTGGTACCCAACAGGTCTG−3’) and crRNA_2 (nucleotide sequence: 5’-CCCGGGGACCTCAGACCTGT−3’), were used to silence the dominant-negative COL6A1 Het. c.877G > A; p.Gly293Arg pathogenic variant present in both individuals. Dermal fibroblast cultures were transfected with the ribonucleoprotein complex formed by the Cas 9 endonuclease and the duplex RNA from the union of each of the crRNAs and a fluorescently labelled transactivating crRNA (tracrRNA-ATTO550). Transfection efficiency was measured by fluorescence-activated cell sorting (FACS) using the tracrRNA-ATTO550 fluorophore, resulting in greater than 80% of ATTO550 positive cells for any of the two guides used. Editing success was evaluated by analysing the DNA reads from transfected cells using Mosaic-Finder [26]. WT COL6A1 allele resulted almost unaltered, while the mutant allele was reduced more than 50%, demonstrating the specificity of the process. Allele-specific expression of COL6A1 was analysed by droplet digital polymerase chain reaction (ddPCR) using specific probes for WT or mutant cDNA. The normalized expression of the mutant allele was reduced, reaching greater than 80% in patient 1 treated with crRNA_2 (BM_Edited2). Please refer to the study by López-Márquez et al. [25] for details.

2.3. Cell culture

Donor fibroblasts from passages P4 to P11 were used in the present study. Cells were maintained at 37°C and 5% CO₂ in a high glucose Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen) supplemented with 10% fetal bovine serum (FBS) (Sigma-Aldrich), 1% penicillin and streptomycin (Invitrogen) and 1% sodium pyruvate (Sigma-Aldrich). The growth medium was replaced every 2 days until 80% confluence was reached, at which point cells were harvested. To achieve detachment, cells were incubated with 0.25% trypsin-EDTA (Life Technologies) at 37°C and 5% CO₂ for 5 min. Supplemented growth medium was then added to inactivate trypsin, and the cell suspension was transferred to a 15 ml tube for centrifugation at 500 r.p.m. for 5 min. The supernatant was discarded, and the pellet resuspended in 10 ml of supplemented growth medium. Using a Neubauer chamber, the number of cells and thus the concentration per ml of suspension could then be calculated for subsequent CDM deposition.

2.4. Cell-derived matrix production

CDMs were generated as previously described [11,12,14]. Precleaned glass microscope coverslips (Ø 18 mm, Superior Marienfeld) were incubated with a sterile 1% gelatin (from porcine skin Sigma-Aldrich) solution in phosphate buffered saline (PBS, Sigma-Aldrich) at 37°C. A cross-linking reaction for the gelatin layer was initiated with 1% filtered glutaraldehyde in PBS (Sigma-Aldrich) and then quenched with 1 M sterile glycine (Sigma-Aldrich) solution in PBS. The coated coverslips were stored in the fridge with 1% penicillin and streptomycin in PBS for a maximum of two weeks until use.

To produce the CDMs, growth medium containing 4 × 10⁴ cells cm⁻² of primary fibroblasts was added to the coated substrate. After the fibroblasts had formed a confluent monolayer, 50 μg ml⁻¹ ascorbic acid in PBS (Sigma-Aldrich) was added every day to stimulate collagen production. Cell media was replaced every 3 days. After 7 days of ascorbic acid treatment, samples were decellularized with a filtered extraction buffer: 0.5% Triton X-100 (Sigma-Aldrich) and 300 mM of ammonium hydroxide solution (NH₄OH, Sigma-Aldrich) in PBS. After the removal of cell media, this solution was added gently on top of the samples for 4 min before being diluted in 2 ml of PBS. Cell detachment can be monitored under light microscopy [27]. Matrices were stored overnight at 4°C and then washed carefully multiple times to remove any cell debris. To evaluate the effectiveness of the decellularization process, the absence of cellular structures such as cell nuclei was confirmed by fluorescent staining with Hoechst dye (Hoechst 33342, Life Technologies) (1 : 1000), as previously described [14].

2.5. Immunostaining

CDMs were fixed with formalin 10% (Sigma-Aldrich) for 20 min, before further treatment with ammonium chloride 50 mM (Sigma-Aldrich) in PBS for 20 min to quench any free aldehyde groups. Blocking and permeabilization were then carried out with a solution of 1% bovine serum (BSA, Sigma-Aldrich) and 0.1% saponin (Sigma-Aldrich) in PBS for 10 min. The matrices were next incubated with the primary antibodies rabbit polyclonal anti-Fibronectin antibody (ab2413, Abcam) (1 : 200) and mouse monoclonal anti-collagen type VI antibody, clone VI-26 (MAB3303, Millipore) (1 : 400) in a PBS solution with 1% BSA and 0.1% saponin for 1 h at room temperature. Samples were carefully rinsed before addition of the fluorophore-conjugated secondary antibody: Alexa 568 goat anti-rabbit (Thermo Fisher Scientific) (1 : 1000), and Alexa 488 goat anti-mouse (Thermo Fisher Scientific) (1 : 1000) in PBS with 1% BSA and 0.1% saponin for a further 1 hour at room temperature. Staining with Hoechst 33 342 (Molecular Probes) (1 : 1000) was also performed to detect any nuclear remains as a verification of successful decellularization. Samples were finally rinsed with PBS and then mounted with Fluoromount (Sigma-Aldrich) on 76 mm × 26 mm microscope slides (RS FRANCE). The samples were allowed to set overnight at room temperature before imaging.

2.6. Confocal microscopy

The Zeiss LSM 800 (Carl Zeiss Microscopy GmbH, Jena, Germany) confocal laser scanning microscope with a 20× air (0.8 numerical aperture) Plan-Apochromat objective lens was used to acquire 16-bit immunofluorescence (IF) images. Diode lasers with wavelengths 405, 488 and 561 nm were used to excite the relevant fluorophores with emission detection bandwidths configured at 400−500 nm (Hoechst 33342), 400−570 nm (fluorochrome Alexa 488) and 570−700 nm (fluorochrome Alexa 568). Z-stacks were acquired in increments of 1 μm for an average of 20 to reconstruct the three-dimensionality of the CDM. In total COL6 data were acquired for a minimum of three fields of view per replicate with at least two replicates per donor. Laser and detector settings were maintained constant across all samples. Confocal stacks were visualized as maximum intensity projections using Fiji software [28]. Curvelet Transform—Fiber Extraction (CT-FIRE) algorithm was then used to segment individual protein fibres and calculate biophysical parameters, as previously described [14].

2.7. Atomic force microscopy-force spectroscopy

AFM-FS experiments were conducted on the CDMs, at room temperature under liquid environment (PBS buffer solution pH 7.4, 10 mM phosphate, 2.7 mM KCl and 137 mM NaCl), following a well-established method to quantify the elastic properties of ECM models, including decellularized tissue ECM [29]. Conducting measurements at room temperature simplifies the experimental approach and minimizes the introduction of additional noise into the recordings, while ensuring the comparability of all samples. AFM-FS experiments were done with an MFP three-dimensional atomic force microscope (Asylum Research) using a V-shaped Si₃N₄ cantilever with a 10 μm diameter, polystyrene colloidal probe and a nominal spring constant of 0.06 N m⁻¹ (NOVASCAN). The sensitivity of the photodiode was acquired by recording a deflection-Z position curve on a mica substrate before the cantilever spring constant was calibrated according to the thermal oscillation technique [30]. AFM-FS measurements were performed by approaching and retracting the colloidal probe to the sample at a constant velocity of 3 μm s⁻¹, to a maximum applied force around 600 pN. Force-indentation curves were recorded in force map mode acquiring 36 curves within a total scan size of 60 × 60 μm². A minimum of five distinct sample regions were selected per sample for the force mapping with at least two replicates per donor measured. Measurements were performed on all donors except patient 5 and control 2 (table 1). The Young’s elastic modulus E value was determined from these curves by fitting the Hertz model using IGOR Pro 6.22A software to indentation curves (electronic supplementary material, figure S1), using a Poisson ratio ν = 0.5, without the need of any correction factor. The fitting was done to a maximum force below 400 pN, which ensured an indentation of less than 15% of the sample thickness, estimated to be around 6−10 µm (electronic supplementary material, figure S2) [14], thus avoiding any mechanical influence from the underlying substratum [29,31–33]. Curves were individually analysed and discarded when excessive adhesion between the sample and tip prevented a good quality of fit for the model. Measurements performed on CDMs from patients 2 and 3 from intermediate phenotype and corresponding to passages P11 and P4, showed no significant differences in E (electronic supplementary material, figure S3), indicating that any possible changes in matrix stiffness due to cell passaging would not be affecting E within the range of cell passages used in the present study.

2.8. Statistical analysis

All statistical analysis and plotting of quantitative data were performed with GraphPad Prism 9. Normality tests were performed before any statistical comparison. For normal distributions, significant differences were assessed with the two-tailed Student’s t‐test or Tukey’s multiple comparisons test for comparison of more than two groups. When data were significantly drawn from a normal distribution, significant differences were assessed with the Mann–Whitney test or Kruskal–Wallis test for comparison of more than two groups. Statistical significance was assumed for α < 0.05. At least three replicates per condition and two independent experiments were performed in each case (N = 2). When results are grouped together into a single phenotype obtained from more than one patient it is because there are no significant differences among the individual patients.

3. Results

3.1. Mechanical analysis of cell-derived matrices from COL6-RDs patients

In the present study, human dermal fibroblasts were used. Although producing inferior levels of ECM when compared with fibro/adipogenic progenitor cells (FAPs), which are the main ECM source [34–36], dermal fibroblasts are easily accessible, expandable cells in culture that express high levels of the three major COL6 genes [37].

Human fibroblasts were isolated from forearm skin biopsies of five representative patients covering a spectrum of mutations and phenotypes, and two healthy donors. Their COL6-RD phenotypic classification and genetic background are summarized in table 1. The missense variant c.877G > A (Gly293Arg) replaces glycine, which is a small, neutral and non-polar amino acid (AA), with arginine, which is basic and polar at codon 293 in the canonical Gly-X-Y repeat of the triple helix domain of COL6A1. This is expected to alter the normal folding and function of the protein, being an established mechanism of disease. Gly-missense mutations mostly correspond to patients with moderate phenotypes [38–40]. Similarly, the variant c.2329T>C (Cys777Arg) replaces cysteine, which is a medium size, neutral and slightly polar AA, with the large size, basic and polar arginine, at codon 777 of COL6A2, also affecting protein folding [41].

No biochemical information is available for the variant c.901-2A>G Intron 7 [42]. The c.930+189C>T variant affecting intron 11 of COL6A1 corresponds to a de novo deep intronic mutation leading to a dominantly acting in-frame pseudoexon insertion. This mutation is generally associated with a severe form of UCMD, which is characterized by a delay of early symptoms and subsequent accelerated progression of the disease. The mutant α1 chain includes a stretch of 24 AA residues that disrupt the Gly-X-Y repeat in the N-terminal region. This still allows the assembly into tetramers, which exert a dominant-negative effect by disrupting the polymerization of wild-type collagen VI tetramers in the ECM [43].

CMDs were produced as described in previous publications from our group [11,12,14]. Briefly, confluent cultures of plated fibroblasts were treated with ascorbic acid for 7 consecutive days stimulating the secretion of collagen. Cellular components were subsequently removed, leaving a three-dimensional scaffold representative of individual COL6-RD cases.

To evaluate the mechanical properties of the CDMs of the different patients (and phenotypes), we measured the CDM’s resistance to deformation and calculated the Young’s elastic modulus, E using AFM-FS. AFM-FS is a suitable technique to evaluate the mechanical characteristics of ECM models at the micro- and nano-metre scale, where mechanical interactions between cell membrane receptors and matrix ligands take place [29].

Owing to the heterogenous microenvironment of CDM models, high variability within the same sample can occur for the mechanical response at the nanometre scale. Thus, we selected a spherical geometry (diameter = 10 μm) for the AFM tip to maximize tip–surface contact area and enable the mechanical contribution of multiple matrix fibres to be averaged in a single measurement. Additionally, this prevents the cantilever from contacting the sample on rough surfaces. Polystyrene 10 μm diameter tips showed lower adhesion to the sample, when compared with borosilicate colloidal or silicon nitride sharp probes, facilitating the obtention of reliable force curves. From this, we obtain an effective mesoscopic elasticity to compare different samples and conditions, that is, more relevant to clinics and diagnostics than that of individual nanoscale components [29,44–46]. Additionally, selecting a spherical tip is advisable for very soft samples, to generate small strains [47].

Force–indentation curves of the different COL6-RDs CDMs displayed clear differences with respect to the contact region response and thus deformability of the matrix (representative curves are shown in figure 1a). Before quantification of the CDM’s elasticity, the curves were assessed individually to ensure excessive adhesion between the probe and sample was not present. Interestingly, this phenomenon was encountered more frequently for UCMD samples, with many pulling events observed in the retraction curves (figure 1a). These are typical patterns for polymeric mesh/network pulling, indicating a less integrated network of fibres for UCMD phenotypes. Despite showing total COL6 production not significantly different from control samples, CDMs derived from the UCMD phenotype show a lower number of COL6 fibres, suggesting a lower degree of polymerization (electronic supplementary material, figure S4 and methods). The Hertz contact model was subsequently applied to the appropriate indentation curves, using E as a fitting parameter (electronic supplementary material, figure S1). We found the E value for CDM from healthy donors (control) to be 1736 ± 100 Pa, with COL6-RD patients CDMs displaying significant deviations from this ‘healthy’ range (figure 1b). While an increase in stiffness with E of 3313 ± 200 Pa (mean ± s.e.m.) was observed for the intermediate phenotype, CDMs E value was reduced to 659 ± 50 and 152 ± 6 Pa correspondingly for BM and UCMD patients. E for intermediate patients 2 and 3, were not significantly different, with values of 3347 ± 163 and 3019 ± 349 Pa, respectively (electronic supplementary material, figure S3). Interestingly, a fivefold increase in E was calculated for intermediate patient 2 over BM patient 1 CDMs, despite both sharing the same mutant variant. The more frequent observation of pulling events in the BM and, especially, UCMD samples along retraction of the AFM probe (figure 1a) suggests a less tight mesh, making it easier to separate and stretch fibres out of the ECM. Therefore, the mechanical properties of the CDMs from COL6-RD patients are significantly different from those of healthy donors, but also among the different phenotypes.

Figure 1.

Mechanical analysis of CDMs. (a) Representative force-indentation curves for each condition tested (dark: approach, light: retraction). (b) Values of the Young’s modulus (E) for CDMs obtained from healthy donors (control) and COL6 RD patient cells: Intermediate (patients 2 and 3), BM (patient 1) and UCMD (patient 4) phenotypes. Results (bars graphs) are the mean ± s.d. ****p < 0.0001.

3.2. Cell-derived matrices derived from patient fibroblasts genetically edited

Genetically edited cells from López-Márquez et al. [25] were used. Dermal fibroblasts from patients bearing the pathogenic variant COL6A1 Het. c.877G>A, presenting either the BM (patient 1, table 1) or intermediate (patient 2, table 1) phenotype were edited by CRISPR/Cas9-based gene editing. Two different CRISPR RNA guides, crRNA_1 and crRNA_2, were used to silence the dominant-negative COL6A1 c.877G>A; p.Gly293Arg pathogenic variant. The effectiveness of the editing strategy was analysed at the genomic DNA level by next generation sequencing, observing the high specificity of both RNA guides, since the WT allele was not edited at significant levels. Furthermore, most of the allelic variants generated after editing the mutant allele were due to out-of-frame indels that would predictably lead to specific silencing of the mutant allele [25]. To demonstrate this, specific probes were designed to measure mRNA expression levels of the COL6A1 mutant and WT allele. Using ddPCR, it was concluded that the reduction in expression of the mutant allele was at least 60%, being 80% in the case of patient 1 treated with crRNA_2. The WT allele was expressed at normal levels compared with the unedited fibroblasts [25]. To examine the response of patient’s fibroblasts (patient 1 and patient 2) to the gene silencing of their mutant allele, CDMs were produced and analysed by immunofluorescence and image processing, and by AFM-FS.

3.2.1. Immunofluorescence analysis of cell-derived matrices derived from treated patient fibroblasts

The secreted COL6 was stained in the CDMs. Representative confocal z-projections obtained for COL6 staining in CDMs from control and COL6-RD samples, produced by edited and unedited cells, are shown in figure 2a. Visual inspection of the images shows fewer and more segmented fibres for patients with COL6-RDs compared with the control sample, as well as some degree of morphological recovery in the edited samples, especially for the intermediate phenotype. Accordingly, the results from image processing (figure 2b) show that fewer, shorter and thinner COL6 microfilaments are produced in patient samples compared with controls, and that they are significantly less aligned for the intermediate phenotype. Gene editing for the intermediate phenotype when using the crRNA_2 guide (Edited2), produces a levelling towards control values for the length and alignment of COL6 microfilaments present in CDMs. However, COL6 control levels are not fully recovered, as expected since COL6A1 allele is partially silenced.

Figure 2.

Effect of gene editing on COL6 organization in CDMs. (a) Representative confocal z-projections of immunostained COL6 in the CDMs: from fibroblasts of healthy donors (Control) and from COL6-RD patients 1 and 2 with BM and intermediate phenotypes, respectively, before and after gene editing using either crRNA_1 (Edited1) or crRNA_2 (Edited2). Scale bar = 50 µm. (b) Graphs displaying various characteristics of COL6 microfilaments obtained by image processing using CT-FIRE algorithm. Results are the mean ± s.d. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Although there appears to be a trend, the use of the crRNA_1 guide (Edited1) does not produce significant differences with respect to the unedited sample. In this regard, López-Márquez et al. also reported variations in patient response as a function of the RNA guide used, with crRNA_2 often recovering more COL6 signal [25].

Alterations in COL6 arrangement after gene editing were less evident for the BM phenotype. In this case, a similar tendency to level control values was observed in samples edited with cRNA_1, although the increase was only statistically significant for the width of secreted COL6 microfilaments. Collagen VI interacts with multiple key components of the ECM such as major fibrillar collagens I, II, III and V, glycosaminoglycans, proteoglycans and fibronectin (FN), among others [48]. Collagen VI filaments interconnect with FN fibrils, affecting their organization in the ECM [49]. Studies in cultured fibroblasts from Col6a1 null mutant mice, which lack Col6 assembly and the ability to secrete Col6 into the ECM, show a loss of three-dimensional organization of FN fibrils, indicating that FN assembly is predictive of collagen VI dysfunction [49,50]. Accordingly, when we examined FN organization in CDMs from patients with COL6-RDs, we observed that similar to COL6 fibres, control samples contained significantly more FN fibrils, whereas these were longer, wider and more aligned in patient CDMs [14].

Here, the morphological analysis of FN in the CDMs of the edited samples shows that in contrast to COL6, for the intermediate phenotype, especially when using the crRNA_2 guide (Edited2), there is a departure from control values, producing a lower number and shorter FN fibrils.

For the BM phenotype, editing with the crRNA_1 guide (Edit1) results in similar behaviour, causing a deviation from the control values and resulting in significantly fewer FN fibrils, which are also thinner. Strikingly, the editing using the crRNA_2 guide in BM phenotype leads to significantly longer FN fibrils (figure 3a,b).

Figure 3.

Effect of gene editing on FN fibrilar organization in CDMs. (a) Representative confocal z-projections of immunostained FN in the CDMs: from fibroblasts of healthy donors (Control) and from COL6-RD patients 1 and 2 with BM and intermediate phenotypes, respectively, before and after gene editing using either crRNA_1 (Edited1) or crRNA_2 (Edited2). Scale bar = 50 µm. (b) Graphs displaying various characteristics of FN fibrils obtained by image processing using CT-FIRE algorithm. Results are the mean ± s.d. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

3.2.2. Mechanical analysis of cell-derived matrices derived from treated patient fibroblasts

Markedly, matrices derived from gene-edited cells all obtained values of E comparable to the healthy donor (control), independently of the RNA guide used (figure 4a). For BM_Edited1 and BM_Edited2, we found the value of E to be 1420 ± 90 and 1894 ± 80 Pa (mean ± s.e.m.), respectively, displaying a greater than twofold increase relative to CDMs from unedited cells of BM patient 1. An increase in compliancy of matrices from the non-treated cells of intermediate patient 2 was additionally achieved, with E calculated as 1380 ± 70 Pa considering Intermediate_Edited1 and 1593 ± 100 Pa for Intermediate_Edited2. In all cases, no significant difference between edited samples and the control (healthy donor) could be highlighted (figure 4a). Figure 4b shows representative force–indentation curves for the edited samples.

Figure 4.

Mechanical analysis of CDMs derived from edited COL6-RD patient fibroblasts. (a) Values of the Young’s modulus (E) for CDMs from fibroblasts of healthy donors (Control) and from COL6-RD patients 1 and 2 with BM and intermediate phenotypes, respectively, before and after gene editing using either crRNA_1 (Edited1) or crRNA_2 (Edited2). Results (bars graphs) are the mean ± s.d. ****p < 0.0001, ns = no significant difference. (b) Representative force–indentation curves for CDMs from COL6-RD patients 1 and 2 with BM and intermediate phenotypes, respectively, using either crRNA_1 (Edited1) or crRNA_2 (Edited2) (dark: approach, light: retraction).

Although gene editing does not completely restore COL6, the changes in elasticity of CDMs match that of control samples, thus defining a healthy range for E that excludes the different COL6-RDs phenotypes (estimated as pale grey areas in figure 5).

Figure 5.

Correlation between morphological and mechanical characteristics. Representation of the COL6 parameters determined by IF and image processing against the CDM’s stiffness (E) measured by AFM-FS. The pale grey areas represent the recovery towards control values in the edited samples. Values are the mean ± s.d.

4. Discussion

We used AFM-FS to investigate the mechanical properties of CDMs from COL6-RDs patients’ fibroblasts. COL6-RDs represent a group of neuromuscular disorders affecting the ECM of connective tissues. Patients affected by COL6-RDs produce a lower amount of COL6 with altered assembly and interaction with other components of the ECM [24,51]. The ECM’s mechanical response is determined by the organization and relative disposition of its different components [52]. We therefore assessed whether the observed alterations in ECM assembly of COL6-RD patients translate to particular mechanical fingerprints. Recent studies postulate CDMs as a reductionistic in vitro model capable of recapitulating the native architecture and composition of the ECM, allowing the identification of the different pathophysiological characteristics of the tissue [9,53]. In the present work, we use AFM-FS to obtain the Young’s modulus (E) as representative of the stiffness in CDMs from patients diagnosed with COL6-RDs.

We found the stiffness of CDMs from COL6-RD patients to significantly deviate from those produced by healthy cells, presenting either an increase or a reduction in E depending on the phenotype.

A softening of the matrix for BM sample compared with controls was identified. This agrees with previous results conducted in a total knockout mice model (Col6a1^−/−), for which a decrease in muscle stiffness was observed [23]. COL6-defficient mice, in general, behave similarly to BM phenotypes [50].

The CDMs from UCMD presented the lowest E values. Additionally, pulling events frequently observed in the retraction force curves on UCMD samples can be attributed to a less densely integrated network of fibres (electronic supplementary material, figure S4), which would be compatible with a possible dominant-negative effect of the intronic mutation that was preventing COL6 polymerization in the ECM [43]. This COL6-RD phenotype is associated with more severe muscle weakness.

Counterintuitively, the intermediate phenotype does not produce E values between the UCMD and BM phenotypes but results in significantly higher E values. In this context, we have previously described that patients with the intermediate phenotype produce straighter and more aligned FN fibres [14], which could possibly contribute to increased CDM stiffness, as it is known that aligned FN fibres act as a template for the assembly of collagens [54]. However, figure 3 does not show any characteristic behaviour in the intermediate phenotype that differentiates it from BM, with respect to the control values. Both produce a smaller number of FN fibrils, which are less thick and more aligned. Also, the morphological analysis of FN shows that edited samples do not match the control values, so we cannot assign the AFM-FS results to FN organization. On the other hand, in studies performed with advanced confocal microscopy on the ECM of the patient with the intermediate phenotype, a significant number of abnormal aggregates positive for COL6 were observed [25], presumably the product of deficient assembly of the protein. The abundance of these aggregates could also contribute to an increase in the stiffness of the ECM [55]. In any case, it seems that the values of E are hardly assignable to an individual contribution, and are affected by multiple interactions between ECM components, which have yet to be elucidated.

Stiffness can produce different mechanical stimuli, which can regulate cell activity both at local and systemic levels through the process of mechanotransduction [56]. As a result, any perturbations to matrix rigidity will generate responses in cells, which could affect their survival, proliferation and migration [57]. Our findings for E might therefore help explain impaired processes previously recorded for COL6-RD patients including increased rates in mitochondrial autophagy and apoptosis in skeletal muscle from mice featuring a knockout of the Col6a1 locus [19], and an accumulation of endosomes–lysosomes and altered mitochondria and Golgi apparatus morphology in primary COL6-RD human fibroblasts [58].

While no effective treatment is currently available for COL6-RDs, novel nucleic acid therapeutics (NATs) have recently emerged as a strategy to target different classes of mutation responsible for the disease [25,37]. Here, we take advantage of CDMs combined with AFM-FS to evaluate the effects of a CRISPR system recently developed by López-Márquez et al., which was able to silence the mutant allele responsible for substituting glycine 293 for an arginine at the N-terminal in the triple helix of COL6. Two CRISPR RNA guides, namely, crRNA_1 and crRNA_2, directed Cas9-mediated endonuclease activity to this mutant allele-inducing indels, a proportion of which significantly reduced its expression [25]. Matrices derived from gene-edited patient fibroblasts were analysed in comparison with those produced by the same unedited cells. The immunofluorescence study of secreted COL6 shows that for the intermediate phenotype edited with the crRNA_2 guide (Edited2), there is a significant recovery of the length and alignment of COL6 microfilaments towards the control values. For the BM phenotype, a similar tendency to level control values was observed in samples edited with crRNA_1 (Edited1), although the increase was only statistically significant for the width of secreted COL6 microfilaments. This is in agreement with previous results of López-Márquez et al., achieving higher percentage increases of COL6 intensity in the ECMs of the edited intermediate phenotypes [25].

The effect of genetic correction is also noticeable with respect to the morphology of FN fibrillar organization. In our previous work, when we examined FN in CDMs from patients with COL6-RDs, we observed that similar to COL6 fibres, control samples contained significantly more FN fibrils, whereas these were longer, wider and more aligned in patient CDMs [14]. For the same guides for which the most evident changes in COL6 are observed, i.e. crRNA_2 (Edited2) for the intermediate phenotype and crRNA_1 (Edited1) for BM, a lower number of FN fibrils and a shortening of these for the intermediate phenotype are observed, as well as less and thinner fibrils for BM, with respect to the control values.

From AFM-FS studies, we demonstrate how fibre organization affects matrix stiffness and that it is possible to define a healthy range of E that excludes different COL6-RD phenotypes. After genetic correction, the stiffness determined for patient matrices shows no significant difference with respect to control (healthy donors) CDMs, suggesting a restoration of ECM mechanical properties, regardless of phenotype or crRNA guide used. This indicates that, although gene editing does not completely restore COL6 levels, the resulting changes are sufficient to significantly improve the mechanical characteristics of the matrix. Thus, analysis of the samples by AFM-FS would provide a more global view of the changes produced by gene editing, highlighting the mechanical characteristics of the matrix as a sensitive biomarker [59] for the evaluation of NATs applied to COL6-RDs.

In conclusion, using CDMs from COL6-RDs patients’ fibroblasts and AFM-FS, it has been possible to establish a healthy range of Young’s modulus for the CDM models, from which the different diseased phenotypes deviate and are distinguishable from each other. We evaluated the response of different patient cells to a recently developed CRISPR/Cas 9 gene correction strategy for which we observed a restoration of matrix stiffness to control values. Thus, variations in the stiffness of CDMs measured with AFM-FS appear to reflect the cascade of events starting from a specific mutation to the final phenotypic expression. We were able to effectively compare and differentiate phenotypes of patients and healthy donors in vitro, as well as to assess genetic editing with high resolution, all without requiring the evaluation of whole tissue samples. Our study provides a functional assessment linked to the elasticity of muscles, tendons and ligaments, as well as the clinical condition of the patients, correlations that require further investigation. Moreover, its ability to evaluate genetic editing in vitro is of great significance. This approach offers more clinically relevant insights into therapy efficacy compared with merely measuring COL6 levels. The presented results therefore suggest that the mechanical analysis of patients’ CDMs could help either in early stratification of patients, as well as contribute to the evaluation of new therapeutic strategies.

Ethics

The authors state that they have obtained appropriate institutional review board approval and have followed the principles outlined in the Declaration of Helsinki for all human experimental investigations. Written informed consent was obtained from individuals and/or their parents or guardians. Biological samples were stored and managed by the Hospital Sant Joan de Déu (HSJD) Biobank.

Data accessibility

Data is available on [60].

Supplementary material is available online at [61].

Declaration of AI use

We have not used AI-assisted technologies in creating this article.

Authors’ contributions

T.W.: formal analysis, investigation, methodology, visualization, writing—original draft; A.L.-M.: formal analysis, investigation, methodology, resources, writing—review and editing; C.B.: investigation, methodology, writing—review and editing; C.J.-M.: funding acquisition, methodology, project administration, resources, supervision, writing—review and editing; J.S.: conceptualization, funding acquisition, project administration, supervision, writing—review and editing; M.I.G.: conceptualization, formal analysis, funding acquisition, methodology, project administration, supervision, visualization, writing—original draft, writing—review and editing; A.L.: conceptualization, formal analysis, funding acquisition, methodology, project administration, supervision, visualization, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

This study was financially supported by ‘Plan Complementario de Biotecnología aplicada a la Salud’, coordinated by Institute for Bioengineering of Catalonia (IBEC) within the framework of ‘Plan de Recuperación, Transformación y Resiliencia (C17.I1)—financed by the European Commission—NextGenerationEU’. Also financial support from ‘Torrons Vicens-RAC1’ grant and Spanish Ministry of Science and Innovation (project PID2022-140459OB-I00 funded by MCIN/AEI/10.13039/501100011033/ and by FEDER A way of making Europe), Department of Research and Universities of the Generalitat de Catalunya (2021 SGR 01545, 2021 SGR 01410), Instituto de Salud Carlos III (ISCIII, grants PI19/0122 and PI22/01382), CERCA Programme/Generalitat de Catalunya and Networking Biomedical Research Center (CIBER) of Spain. CIBER is an initiative funded by the VI National R&D&i Plan 2008–2011, Iniciativa Ingenio 2010, Consolider Program, CIBER Actions and the Instituto de Salud Carlos III, with the support of the European Regional Development Fund (ERDF).