Low dystrophin variability between muscles and stable expression over time in Becker muscular dystrophy using capillary Western immunoassay

Becker muscular dystrophy (BMD) is the milder allelic variant of Duchenne muscular dystrophy, with higher dystrophin levels. To anticipate on results of interventions targeting dystrophin expression it is important to know the natural variation of dystrophin expression between different muscles and over time. Dystrophin was quantified using capillary Western immunoassay (Wes) in the anterior tibial (TA) muscle of 37 BMD patients. Variability was studied using two samples from the same TA biopsy site in nine patients, assessing nine longitudinal TA biopsies, and eight simultaneously obtained vastus lateralis (VL) muscle biopsies. Measurements were performed in duplicate with two primary antibodies. Baseline dystrophin levels were correlated to longitudinal muscle strength and functional outcomes. Results showed low technical variability and high precision for both antibodies. Dystrophin TA levels ranged from 4.8 to 97.7%, remained stable over a 3–5 year period, and did not correlate with changes in longitudinal muscle function. Dystrophin levels were comparable between TA and VL muscles. Intra-muscle biopsy variability was low (5.2% and 11.4% of the total variability of the two antibodies). These observations are relevant for the design of clinical trials targeting dystrophin production, and may urge the need for other biomarkers or surrogate endpoints.


Scientific Reports
| (2021) 11:5952 | https://doi.org/10.1038/s41598-021-84863-w www.nature.com/scientificreports/ mediated exon skipping, gene therapy or stop codon readthrough), dystrophin quantification has become central to show the pharmacodynamic effects of these drugs. The availability of novel technologies based on both immunofluorescence analysis 14,15 and Western blot 16 has enabled the refinement of the experimental procedures and provided new ways to quantitatively assess dystrophin levels. Application of the highly sensitive and quantitative capillary Western immunoassay (Wes) showed that DMD patients produce trace levels of dystrophin in muscle biopsies up to 7% of healthy controls, and that patients with a BMD diagnosis could have dystrophin levels from low (~ 10% of healthy controls) up to 100% of healthy controls 16 . These data support a differentiation between DMD and BMD based on dystrophin levels with a potential threshold effect at about 10% dystrophin, but also challenge the classical dichotomization between DMD and BMD and demonstrate a more fluid distribution with some overlap across the dystrophinopathy spectrum and healthy controls. Analysis of dystrophin from muscle biopsies has been performed inconsistently across different muscle groups, without paired observations enabling comparison of dystrophin levels within and between muscles 17 . It is currently also unknown whether dystrophin levels are stable over time. Disease progression in BMD is typically slow and longitudinal studies are usually not long enough to observe a sufficient number of events (such as loss of ambulation) to support convincing associations between dystrophin levels and clinical progression. Also, several attempts have been made to identify potential associations between dystrophin levels and clinical performance in BMD patients 18,19 , but a clear and linear relationship has not been found.
We present data from an observational longitudinal study in BMD patients showing how dystrophin levels are stable across muscles and over time, but strongly vary between patients independent of mutation type and age, denoting a patient-specific capacity to synthesize dystrophin.

Materials and methods
Inclusion of participants and muscle biopsies. Participants consisted of ambulant and non-ambulant men aged 18 years or older, registered with a BMD diagnosis in the Dutch Dystrophinopathy Database (DDD) 20 and children form the Leiden University Medical Center (LUMC) outpatient clinic. Adult patients had participated in two BMD natural history studies, one performed in 2011 (referred to as the 2011 study) and one starting in 2014 (referred to as the 2014 study).
Recruitment of participants for the 2011 study has been described previously 19 . Muscle strength was assessed via bilateral measurement of handgrip, flexion and extension of the knee, hip and elbow using quantitative muscle strength testing (QMT) with the quantitative muscle assessment (QMA) system (Aeverl Medical LLC; www. QMASy stem.com: Gainesville, GA USA). Maximal voluntary isometric contraction (MVIC) was calculated as described by Hogrel et al. 21 . Functional tests were not part of this study protocol.
Participants in the 2014 study were also recruited via the DDD 22 . This study had a follow-up of 4 years and consisted of the following yearly functional tests: the six-minute walk test, North Star Ambulatory Assessment (NSAA) and timed tests (10 m walk/run, rise from the floor and four step ascend and descend test) in ambulant patients, and the Performance of Upper Limb version 1.2 in all participants. Next, muscle strength was assessed similar to the 2011 study. All assessments were done by two observers with help of a physiotherapist, trained previously as part of a clinical trial protocol. All participants in this natural history study were asked permission for muscle biopsies to be taken from the right TA and/or vastus lateralis muscle (VL) either at baseline or year 1, on the same day after all functional and muscle strength assessments had been performed. We chose to biopsy the clinically more severely affected right VL muscle for intramuscular comparison of dystrophin levels. An exclusion criterium for these biopsies was the use of anticoagulants.
The muscle biopsies from the pediatric BMD patients taken as part of the diagnostical work-up were included if material was available after completion of all diagnostic procedures (i.e. residual tissue). Functional assessments in these children had not been performed at the same time as the biopsies and were therefore not included in this study.
The 2011 and 2014 study protocol were both approved by the LUMC medical ethical committee. The clinical data and human tissue have been obtained, stored and handled in strict accordance with relevant guidelines and regulations. Written informed consent was obtained from all study participants. The residual muscle biopsy tissue of the pediatric BMD patients followed in the outpatient clinic at the LUMC is defined as tissue obtained for diagnostic purposes and that is residual after the diagnosis. The use of residual muscle tissue was in accordance with the codes of conduct as outlined by the Federation of Dutch Medical Scientific Societies. Oral consent from the children's legal representatives was obtained by their treating physician, which is in accordance to regulations of the Medical Research Involving Human Subjects Act regarding research with residual human tissue.
Muscle biopsies and dystrophin quantification. All muscle biopsies were obtained using a conchotome and performed according to standard protocol 23 . The skin was infiltrated with lidocaine 1%. A 1.5-2 cm skin incision was made and the fascia penetrated with a scalpel blade. A total of 3-4 samples were collected through the same incision. All samples were inspected for size and quality, transferred for transport into a plastic specimen bottle and snap-frozen in 2-methylbutane cooled in liquid nitrogen and stored at − 80 °C at the pathology department of the LUMC, all within 30 min.
The lysates were diluted to 25 μg/ml, resulting in a loading amount of 0.125 μg per well/capillary. All lysates were analyzed by Wes twice, in separate experiments. To control for differences in chemiluminescence signal between experiments, a 6-point calibration curve of a healthy muscle reference sample was routinely included, usually ranging from 0.008 to 0.25 μg. This reference sample was selected from a panel of healthy controls based on it displaying average dystrophin levels with both ab154168 and Mandys106 as previously described 16 . Wes runs were considered valid if the calibration curve displayed an R 2 > 0.99, based on 4-6 points in a relevant concentration range.
For α-actinin the calibration curve also ranged from 0.008 to 0.25 μg of the same healthy control. Since α-actinin abundance was too high to get reliable chemiluminescence values from the same lysate concentration required to measure dystrophin (substrate depletion), we reduced HRP activity by diluting the standard Protein-Simple secondary anti-rabbit-HRP antibody with 1/60 unconjugated anti-rabbit antibody (#ab6702, Abcam). The final muscle content-corrected dystrophin values were generated by first expressing both the dystrophin and α-actinin signals as percentage of control (% CTRL) using the following formula: and then dividing the dystrophin % CTRL value by the α-actinin % CTRL value.
Statistical analysis. All statistical analyses were performed using R (version 3.6.0) 24 . Comparison of the values of repeated measurements of the same antibody on each lysate was performed using scatter plots and computing Pearson's correlation coefficient; the measurement error was estimated using a linear regression model with both measurements of the same antibody as response and lysate as covariates (measurement error = 1 − R 2 ).
Paired measurements of Mandys106 and ab154168 were compared using scatter plots and computing Pearson's correlation coefficient; the average difference between the two antibodies was computed and its significance tested using Wilcoxon's paired samples test.
Comparison of dystrophin levels from adjacent samples from the same biopsy was achieved visually with a bar plot and the proportion of biological variability attributable to the different samples was estimated using a linear % CTRL = sample protein equal to × µg/ml CTRL (from cal.curve) × µg/ml sample loaded www.nature.com/scientificreports/ regression model with measurements from both samples as response and patient IDs as covariates (biological variability due to the pieces = 1 − R 2 ). Comparison of dystrophin levels in TA and VL was visualized with a bar plot. The mean difference between the two muscles was estimated and then tested using Wilcoxon's paired samples test.
Changes of dystrophin levels over time were visualized by comparing the levels of dystrophin in the same patient over a time period ranging between 4 and 5 years. The average yearly change was computed and tested using Wilcoxon's signed rank test.
The association between changes in patients' performance and baseline dystrophin measures was estimated using a linear mixed model 25 with changes in performance from baseline as response, baseline dystrophin values as covariate and a patient-specific random intercept included to account for the correlation between repeated measurements from the same individual. The significance of the associations was tested using a T test with Satterwhite's approximation for the degrees of freedom, as implemented in the R package lmerTest 26 .

Results
Participants and available muscle biopsies. From 28 participants of the 2011 study, stored lysates and/ or muscle tissue from the right TA muscle were available and included for re-analysis. Twenty-six participants had an in-frame mutation in the DMD gene and one participant had an out-of-frame mutation (deletion exon 3-7) with a mild disease course (age at wheelchair dependence 45 years). In one participant, no mutation had been found in the past using MLPA and high resolution melting curve analysis. Diagnosis had been made elsewhere by an experienced neuromuscular specialist, based on reduced dystrophin levels in a diagnostic muscle biopsy, raised serum creatine kinase levels, and calf hypertrophy.
Of the thirty-eight BMD patients who consented for participation in de 2014 study, two participants in whom a BMD diagnosis had been made elsewhere based on clinical characteristics and a diagnostic muscle biopsy were excluded after re-analysis of a newly obtained muscle biopsy and sequencing of the DMD gene. A gene panel was performed in both and revealed a homozygous mutation in exon 21 of the CAPN3 gene for the first and a homozygous mutation in exon five of the ANO5 gene for the second. For 35 of the remaining participants the diagnosis was confirmed by an in-frame mutation in the DMD-gene and one participant (who also participated in the 2011 study) had an out-of-frame mutation (deletion exon 3-7).
Of the 36 participants in the 2014 study, 13 consented for one muscle biopsy of the right TA muscle and of these 13 participants, 8 consented for an additional muscle biopsy of the right VL muscle. Nine of the 13 participants had also participated in the 2011 study.
Additionally, diagnostic muscle biopsies taken from the right TA muscle of 5 children (median age 6 years; range 6-12 years) with BMD were also included. Diagnoses in these children were based on an in-frame mutation in the DMD gene in 3 children, and an exon 3-7 deletion with a mild disease course in 2 children.
For the analysis participants were assigned into different study cohorts for analysis depending on their participation in one of the natural history studies or both, and type of muscle (TA and/or VL) biopsied (for an overview see Fig. 2). Cohort 1 (cross-sectional analysis) was defined by participants of whom a biopsy from at least one point in time was available: 28 from the 2011 study, 4 from the 2014 study, and 5 pediatric BMD patients adding to a total of 37 right TA biopsies for cross-sectional analysis.
Cohort 2 (longitudinal analysis) consisted of 9 participants of whom a right TA biopsy was available from the 2011 and the 2014 study.
Cohort 3 (inter-muscular analysis) was defined by 8 participants from the 2014 study of whom biopsies from the right TA and VL muscle were available obtained on the same day.
Finally, to allow comparison of dystrophin levels within the same muscle at the same point in time, two adjacent samples were assessed that had been taken from the same TA muscle site of 9 participants (8 participants from the 2011 and 1 of the 2014 study). Clinical characteristics of all participants per cohort are summarized in Table 1. Mutational data of all participants can be found in the Supplementary Table S1. None of the participants had ever used corticosteroids or were on (nocturnal) respiratory support. Dystrophin quantification. All dystrophin measurements were performed in duplicate with 2 independent antibodies, namely Mandys106 and ab154168 (see "Methods"). Dystrophin levels were below quantification in 1 TA and 3 VL muscles. Of the quantifiable samples, median level and range of absolute dystrophin levels in the TA using Mandys106 were 31.3% of control (CTRL) (4.8-97.7%) for cohort 1 and 33.0% of CTRL (21.6-97.7%) for cohort 2 at baseline (dystrophin % from biopsies of the 2011 study). In cohort 3, median dystrophin levels were 34.8% (18.6-86.4%) for the TA and 37.0% (12-103.4%) for the VL. In the five pediatric cases (part of cohort 1), dystrophin levels in the TA ranged from 4.8% to 18.1% and were lower than in the adults (range 12.5-97.7%) (Mann-Whitney U p < 0.001). Baseline dystrophin levels in the TA of all participants are displayed in Supplementary Table S1.
Dystrophin quantification showed a strong and highly significant correlation for measurements obtained with each antibody and across antibodies. The value of the Pearson's correlation coefficient was equal to 0.897 (p < 0.0001, Fig. 3A) between replicates of Mandys106 and 0.940 (p < 0.0001) (Fig. 3B) between replicates of ab154168. Moreover, the correlation between the averaged values of Mandys106 and those of ab154168 was 0.972 (p < 0.0001) (Fig. 3C). The measurement error, estimated using the linear regression model, was found to be small for both antibodies. It accounted for 5.2% of the total variability of Mandys106 and for 3.1% of the total variability of ab154168. Dystrophin quantification by ab154168 yielded overall higher levels of dystrophin compared to quantification with Mandys106 (mean difference between ab154168 and Mandys106 = 7.9%, p-value of the paired samples Wilcoxon test < 0.0001). www.nature.com/scientificreports/ Analysis of 2 adjacent muscle samples collected from the right TA at the same point in time (intra-biopsy analysis) showed a high correlation of dystrophin levels in all 9 study participants when analyzed with both antibodies independently (Fig. 4A). The biological variability was estimated to be equal to 5.2% of the total variability of Mandys106 and 11.4% of the total variability of ab154168.
In cohort 2, dystrophin was quantified in follow-up right TA biopsies from 9 participants obtained in the 2011 and 2014 study (biopsy procedure interval 3-5 years). Dystrophin levels in all biopsies were quantified twice with both antibodies for each given year. The mean and standard deviation of all measurements are displayed in Fig. 4B. Dystrophin levels were comparable across time for all participants. The average yearly change was not significant for both Mandys106 (average increase of 1.6%, p-value of the Wilcoxon signed rank test = 0.10) and ab154168 (average increase of 1.7%, p = 0.13).
In cohort 3, as expected the biopsy of the right VL was not informative for 3 out of 8 participants because it mainly contained fat, resulting in α-actinin levels below level of quantification. For the remaining 5 participants the analysis provided comparable levels of dystrophin for TA and VL biopsies. The average difference was not statistically different between TA and VL for both antibodies, Mandys106 (p-value of the paired samples Wilcoxon test = 0.84) and ab154168 (p = 0.84), respectively (Fig. 4C).
To test whether yearly change in performance was affected by baseline dystrophin levels (either from the 2011 or the 2014 study, depending on when the first TA biopsy had been taken) we focused on knee extension as exemplar strength readout, and on the North Star Ambulatory Assessment (NSAA) as functional test. Available knee strength and NSAA outcome measures of participants organized by age and dystrophin range are presented in Fig. 5. Data of knee strength and NSAA were available for 24 and 13 participants respectively. Use of a linear mixed model allowed assessment of potential associations between TA dystrophin levels and performance changes. However, no significant association was found, irrespective of the antibody used to quantify dystrophin   www.nature.com/scientificreports/ (knee extension: p = 0.66 using Mandys106 and p = 0.84 using ab154168; NSAA: p = 0.85 using Mandys106 and p = 0.66 using ab154168) (Fig. 5A,B).

Discussion
In this study we assessed dystrophin levels in detail in a cohort of BMD patients with different mutations and variable disease severity. First, we showed that even though dystrophin levels vary considerably between BMD patients (ranging from 5 to more than 90%), their levels remain stable over a 3-5 year period. Second, dystrophin levels were comparable between two differently affected leg muscles, namely the right TA and VL muscle, www.nature.com/scientificreports/ as well as between independent samples obtained from the same TA muscle. Finally, we did not identify a clear relationship between dystrophin levels and changes in muscle strength (over 3-8 years of follow up) and function (over 3-5 years of follow up). This is the first time that dystrophin levels have been analyzed longitudinally. Although results are from a limited number of samples, the fact that dystrophin levels remain stable over 3-5 years implies that progression of the disease is not caused by a progressive reduction in the dystrophin synthesis capacity.
The low variation in dystrophin levels observed between two differently affected leg muscles (TA and VL) collected during the same biopsy procedure suggests that, at least in the lower limbs, dystrophin levels do not determine the extent to which muscles are affected. However, due to the high fat infiltration observed in the VL, it is advisable to biopsy the more accessible and less affected TA muscle to obtain samples for protein quantification, even in severely affected patients. The availability of adjacent independent samples taken during the same biopsy procedure allowed to show that no differences in dystrophin levels are to be expected based on the sampling location. The overall lack of variation in dystrophin observed across muscles and over time raises new questions such as what factors are responsible for the individual-specific dystrophin levels and whether these apply to other proteins in muscle as well. Both cis-and trans-acting elements such as genetic variants, miRNAs and translation factors may be playing a role and deserve further research.
We also found that dystrophin levels remain relatively constant in stored muscle biopsies. We reanalyzed lysates prepared from stored (at − 80 °C) biopsies of a panel of 25 BMD patients from our previous study. Dystrophin levels were found to range from 10 to 90% of control in the previous study 16 compared to 12-98% when reanalyzed for this study.
The lack of variation observed further supports the use of the Wes for dystrophin quantification. This technique used for dystrophin quantification has been proven to be highly sensitive, reproducible and quantitative over a large dynamic range with a lower limit of quantification; low enough to even quantify trace dystrophin levels 16 . Consistent with these findings, results in the present study confirmed this low technical variability. The percent coefficient of variation (CV%) between 2 replicate measurements of the same muscle lysate using 2 different anti-dystrophin antibodies was 5.2% for Mandys106 and 3.1% for ab154168, which is well below the accepted CV% for quantitative assays according to FDA/EMA regulations 27 . The advantage of the Wes high sensitivity is that it requires > 100-fold less muscle tissue compared to regular Western blot. Consequently, very small patient biopsies can be used including needle biopsies, instead of more invasive open biopsies that have a higher patient burden, especially in the pediatric population 28 . www.nature.com/scientificreports/ At last, we also performed longitudinal muscle strength and functional tests to assess potential correlations between dystrophin levels, age and disease severity and progression. The disease severity in our BMD cohort showed the expected variability 5,6,19 . We were not able to demonstrate a direct relation between dystrophin levels and functional decline over a period of 3-5 years not even in the population with the same genotype (Deletion exon 45-47). This is in line with previous studies where no correlations between dystrophin levels and disease severity could be established when above 10-20% of normal 5,19 . Nonetheless, levels below 3-5% are still part of diagnostic criteria for DMD 11,16,19,29 suggesting a dystrophin threshold that is relevant to the clinical phenotype. In the present study the lowest level of dystrophin among patients who were still ambulant by the age of 16 years was 12%. These dystrophin levels are still above those currently reported after longterm treatment with antisense oligonucleotides in DMD 30,31 . Despite the finding of relatively constant dystrophin levels over time, this level does not seem to be an adequate marker for disease severity or to predict the clinical course. Therefore there is a necessity to further explore disease modifying elements such as possible differences in regenerative capacity, epigenetic and/or environmental factors.
In summary, dystrophin expression can be quantified reliably by the state-of-the-art Wes capillary immunoassay and is relatively constant with an interval of 3-5 years in BMD patients with different disease severity. Sample variability within the same muscle and between the TA and VL muscle are both limited, although quantification is hampered in severely affected (VL) muscles due to fat infiltration. Similar to earlier studies, we were not able to demonstrate a relationship between dystrophin levels and age, nor between dystrophin and the rate of decline in muscle force or function, although there may be a threshold that is relevant to the dystrophinopathy phenotype. The large variability in dystrophin levels in BMD with a relatively slow progression in part of the patients, even over 3-5 years, is an important factor to consider in the development of clinical trials and may urge the need for other biomarkers or surrogate endpoints.