Accurate Dystrophin Quantification in Mouse Tissue;Identification of New and Evaluation of Existing Methods

Abstract

Duchenne muscular dystrophy (DMD) is a progressive muscle–wasting disorder primarily affecting males. This disorder is caused by mutations in the DMD gene that abolish dystrophin protein function. Many therapeutic approaches for DMD aim at recovery of the dystrophin protein in muscle fibers of affected patients, rendering accurate dystrophin quantification important. Several methods have been reported to detect and quantify dystrophin restoration in preclinical and clinical trials. We here evaluated the applicability of dystrophin specific enzyme-linked immunosorbent assays (ELISA) and a TaqMan protein assay, benchmarking them against Western blotting analysis. Despite numerous optimization attempts, in our hands the background signals in the ELISA and TaqMan protein assays were too high to allow dystrophin quantification. By contrast, the Western blot approach was able to detect dystrophin levels as low as 0.2% in a reproducible manner. We provide a Western blot protocol that allows sensitive and accurate dystrophin quantification in preclinical studies.

INTRODUCTION

Patients suffering from Duchenne muscular dystrophy (DMD) mostly have frame-shift or non-sense mutations in their DMD gene, resulting in premature truncation of protein translation and the synthesis of non-functional dystrophin proteins [1, 2]. The DMD gene is located on the X-chromosome at position Xp21. Dystrophin is associated with other members of the dystrophin-glycoprotein complex [3], which links the actin cytoskeleton to the extracellular basal lamina, providing tensile strength to muscle fibers and protecting them from injury as the muscle contracts and relaxes [4].

Lack of dystrophin makes muscle fibers more sensitive to damage during contractions, leading to chronic breakdown of muscle tissue and loss of muscle function in patients. Eventually, they become wheelchair-dependent in early adolescence and generally die in the 3rd or 4th decade of life in the Western world [5]. Patients suffering from Becker muscular dystrophy (BMD) have mutations in the DMD gene that maintain the open reading frame and allow the production of internally deleted and/or reduced levels of dystrophin [6, 7]. As Becker type dystrophins are partially functional, BMD is associated with a slower disease progression.

Although the genetic defects responsible for DMD and BMD have been identified years ago, most DMD patients only receive symptomatic treatment and care [8, 9]. Potential therapeutic approaches are being developed by many laboratories worldwide. A significant proportion of these approaches aims to restore the production of dystrophin. Of these, a stop codon read-through compound (Translarna) [10, 11] received conditional market authorization from the European Medicine Agency recently (August 2014), while for antisense-mediated (AON) exon skipping [12, 13] marketing authorization applications have been submitted with the Food and Drug Administration (US) and the European Medicine Agency. During preclinical and clinical development, it is important that dystrophin restoration can be measured with a method that is sensitive and quantitative. Immunofluorescence has been used mostly to provide a relative quantification of the level of dystrophin restoration in muscle biopsies (see [14] for a review).

There are several other methods with which protein concentrations can be quantified. Firstly, the Enzyme-linked Immunosorbent Assays (ELISA) have the advantage that they are generally sensitive and relatively easy to implement and standardize. The sandwich ELISA has been previously reported for dystrophin quantification [15–17]. However, except for a modified version of the protocol of Ishikawa et al. [16] and Ferrer et al. [18] the ELISA system has never been reported by other groups in the DMD field, nor have the groups reporting on the system reused it in later publications. Recently, a dystrophin specific indirect ELISA kit became commercially available, but this has not yet been evaluated in published work.

Secondly, protein levels can be quantified with the TaqMan protein assay, a technique which is an adaption from the proximity ligation assay (PLA^TM) technology [19, 20]. It enables quantification of low amounts of protein with high sensitivity and has never been used for dystrophin to our knowledge.

Thirdly, the reverse protein assay has been developed for diagnostic purposes and can be used to assess the levels of dystrophin and related proteins in patients suspected of a muscular dystrophy [21]. However, this assay does not allow quantification of (very) low amounts of proteins, but primarily detects presence or absence of proteins. As such it is less useful for (pre-)clinical studies aiming at dystrophin restoration.

Fourthly, mass spectrometry offers a platform to reproducibly quantify protein concentrations [22]. Combinational use with micro dissection of biopsies rich of fibrotic and fat infiltrates would allow quantification of dystrophin only in muscle tissue. However, at the moment, the lower detection limit was reported at ∼5% of wild type (WT) levels and therefore not suitable in pre-clinical settings where often <  5% dystrophin is produced upon treatment. Furthermore, this method relies on the presence of sophisticated and expensive mass spectrometry equipment, which is not available at most institutes.

Lastly, Western blotting is a much used technique to quantify dystrophin levels [14], allowing both quantitative and qualitative analysis, since it also provides information on the size of the protein. Most laboratories use a protocol adapted from a report by Anderson et al. [23], but with numerous individual variations and revisions. However, in most cases the procedure is laborious, time consuming (>2.5 days) and the reproducibility is generally not optimal.

Our aim was to find a sensitive, quick and reproducible method for the quantification of dystrophin on protein level. In our hands the ELISA and TaqMan protein assays do not allow reliable quantification of dystrophin. The commercially available Trans-Blot Turbo transfer system allows sensitive and reproducible quantification of dystrophin using a relatively quick protocol.

MATERIAL AND METHODS

Samples

Protein lysates used in the experiments were derived from muscles from mdx and mdx/utrn^–/– mice (no dystrophin), Xist^Δhs (WT; 100% dystrophin, kindly provided by Prof Neil Brockdorff, Department of Biochemistry, University of Oxford, UK [24], female mdx-Xist^Δhs [25] and mdx/utrn^–/–-Xist ^{Δ hs} [26] mice. The latter two models were obtained from crossings of mdx and Xist^Δhs mice and mdx/utrn ^+/– and utrn^–/–/Xist^Δhs mice respectively. Mdx-Xist^Δhs and mdx/utrn^–/–-Xist^Δhs mice express variable, low levels of dystrophin (varying from 3–47% ) due to skewed X-inactivation of the WT dystrophin allele. Mice were sacrificed by cervical dislocation and diaphragm, quadriceps and heart muscles were dissected immediately. Furthermore, the diaphragm, triceps and quadriceps were isolated from mdx mice treated twice weekly with 100 mg/kg of a 2’-O-methyl phosphorothioate antisense oligonucleotide that was previously reported [27] targeting exon 23 for eight weeks. Treated mice were sacrificed at different time points (t = 36 hours, 1, 3, 8, 12 and 24 weeks) after the last injection [28]. Mice were housed under standard conditions and had ad libitum excess to regular chow and water. All animal experiments were approved by the animal ethical committee of the LUMC. All tissues were snap-frozen in liquid nitrogen-cooled 2-methylbutane and stored at –80°C until analysis. Human control muscle tissue was obtained after informed consent and frozen in 2-methylbutane cooled in liquid nitrogen.

Protein isolation

Samples were homogenized in zirconium beads (1.4 mm; OPS Diagnostics) or MagNa Lyser green bead tubes with a MagNa Lyser (Roche, Germany) for 2–5 rounds of 20 seconds at speed 7000 or the BBY24M Bullet Blender Storm (Next Advance, Averill Park, NY) in 1 ml treatment buffer (75 mM Tris-HCl, (pH 6.8) and 15% (w/v) sodium dodecyl sulphate (SDS)) or Ripa buffer (50 mM Tris-HCl, (pH 8.0), 150 mM NaCl, 1% IGEPAL CA-630, 0.5% DOC and 0.1% SDS). Protein was also isolated with treatment buffer containing no SDS (used for ELISA and TaqMan protein assays) or with protein quant sample lysis kit containing TritonX100 (ThermoFisher Scientific, USA, used for TaqMan protein assay). Protein of human samples was isolated with the treatment buffer containing 15% (w/v) SDS. Protein concentrations were determined by the bicinchoninic acid (BCA) protein assay kit (ThermoFisher Scientific) using bovine serum albumin as a standard. Lysates were aliquoted to prevent freeze/thaw cycles and stored at –80°C.

Histology

The triceps of AON treated mdx mice was snap frozen in 2-methylbutane cooled in liquid nitrogen. Sections (8 μm) were cut along the entire length of the muscle and pasted on Superfrost Plus slides (ThermoFisher Scientific). Sections were fixated in ice-cold aceton and blocked with 5% horse serum in 1×phosphate buffered saline (PBS) with 0.05% Tween20 (PBST) for one hour. Primary antibodies for spectrin-beta III (anti rabbit, ThermoFisher Scientific, dilution 1:200) and dystrophin (anti goat C20, sc-7461, Santa Cruz, Germany, dilution 1:50) were incubated over night at room temperature (RT). After washing, slides were incubated with donkey-anti-goat alexa 488 and donkey-anti-rabbit alexa 594 at a dilution of 1:1000 for one hour. Pictures were generated at a 20×magnification with the DM5500 (Leica, the Netherlands).

Indirect ELISA

Flat-bottom 96-wells plates (Nunc maxisorp, VWR, the Netherlands or Costar 3590 high binding, USA) were either pre-coated with 30×diluted collagen at 37°C for one hour followed by an overnight incubation with 100 μl WT or mdx protein sample (0.005 – 5 μg/ml in 1×PBS) at 4°C; or only coated with protein samples. Controls that were taken along were water and protein incubated with only primary, secondary or no antibodies. Wells were emptied and blocked with either 100 μl 2% casein, 4% milk, 1% FBS or 2% BSA diluted in 1×PBS, for 30 min at RT with continues shaking of the plate (900 rpm). Plates were washed four times with PBST. The wells were incubated with 100 μl dystrophin specific primary antibody Dy 4 (NCL-DYS1, Novocastra Laboratories, UK); Dy 8 (NCL-DYS2, Novocastra Laboratories); Dy 10 (NCL-DYS3, Novocastra Laboratories) or C20 (Santa Cruz Biotech) in different concentrations (1:10; 1:20, 1:30, 1:50 and 1:100 in blocking buffer) in a shaking (100 rpm) water bath at 37°C for one hour. Plates were washed four times with PBST and incubated with a secondary antibody (goat-anti-mouse or donkey-anti-goat HRP, ThermoFisher Scientific) in different concentrations (1:1000 and 1:5000 in blocking buffer) in a water bath under gentle agitation (100 rpm) at 37°C for one hour. Plates were washed four times with PBST and 100 μl OPD solution with 0.03% H₂O₂ was added and incubated for 15 min. The reaction was stopped by addition of 50 μl 1 M H₂SO₄ after which the plate’s absorption was scanned in a plate reader at 490 nm.

Sandwich ELISA

The protocol has been adapted from Glenn Morris [17]. A flat-bottom 96-wells plate (Nunc, VWR) was pre-coated with 100 μl desalted and purified Mandra1 or Manex7 antibody in 1×PBS (10 - 389 μg/ml, Glenn Morris, CIND, UK). The plate was washed three times (1×PBS, 1% triton X-100 (PBST-X100)) and blocked with 4% BSA in PBST-X100 for 30 min. After three washing steps with PBST-X100, 100 μl protein lysate (20 - 40 μg/ml, mdx or WT) or water was added to each well and incubated for one hour at RT. The plate was washed three times with PBST-X100 and incubated with an in house biotin labelled Manex antibody (Manex7374A, Glenn Morris, CIND, UK, 5 μg/ml) or C20 (Santa Cruz Biotech, Germany, 5 μg/ml) in incubation buffer (1×PBS, 1% BSA, 1% HS, 1% FCS) for two hours followed by washing of the plate with PBST-X100. Finally a peroxidase-avidin formulation (Vectastain Elite ABC kit, Vector Laboratories, US) was performed for 30 min. The plate was then washed three times in PBST-X100 and 50 μl 1M H₂SO₄ was added per well. The plate was scanned at 490 nm in a plate reader. Control wells were either not pre-coated, only incubated with primary, secondary or no antibodies.

Indirect ELISA commercial kit protocol

The protocol has been adapted from the manufacturer (USCNK Life Science, UK). To each well of a 96-wells plate 100 μl of either the standard (ranging from 0.312 – 20 ng/ml), water control or protein sample (0.312×10⁶ – 25×10⁶ ng/ml, mdx, WT and mdx-Xist^Δhs) was added and incubated at 37°C for two hours. The wells were emptied and 100 μl of the detection reagent A was added and incubated at 37°C for one hour. The wells were then washed three times for two min with 350 μl wash solution and incubated with 100 μl detection reagent B at 37°C for 30 min. The wells were washed five times with 350 μl wash solution and incubated with 90 μl substrate solution at 37°C for 15–25 min. Finally, 50 μl stop solution was added to each well and the plate was read at 450 nm in a plate reader.

TaqMan protein assay

The TaqMan protein assay was performed according to manufacturer’s protocols, with some adaptations. In short, four different monoclonal antibody pairs and two polyclonal antibodies were selected; targeting the dystrophin protein with limited distance (<300 amino acids) between them (Table 1); and biotinylated (EZ-Link Sulfo-NHS-LC-Biotin, Pierce, the Netherlands) followed by an overnight dialysis in 1×PBS (during which the solution was refreshed four times). Antibodies provided in phenol red buffer were purified with a desalting column (Slide-A-Lyzer Mini Dialysis Unit (MWCO = 7000), ThermoFisher Scientific) prior to the biotinylation step.

The 5’ and 3’Prox-Oligo’s (provided with the TaqMan Protein Assay’s Open Kit) were bound to the different biotinylated antibodies and their binding capacity was assessed with the forced proximity probe test. The delta Ct value of each antibody was determined by a Real-time PCR on a 7900HT-system. The antibody’s binding capacity was calculated by subtracting the Ct of the negative control from the average Ct of the probes. Antibodies with a delta Ct value >8.5 were accepted.

The assay’s probe sets were separately prepared by combining 200 nM antibody A or B with 5 μl 3’Prox-Oligo or 5’Prox-Oligo respectively and incubated at RT for one hour. Antibodies that were conjugated to the 3’ and 5’Prox-Oligos targeted the down and upstream parts of the protein respectively. Thereafter 90 μl probe storage buffer was added and incubated at RT for 20 min. Probe sets were stored at –20°C.

A serial dilution was prepared from 0.00024 – 5 μg mdx, WT or mdx-Xist^Δhs protein in lysate buffer; 2 μl of each sample well was then transferred to a new plate containing assay probe solution. During optimization, different amounts of assay probe solution (2, 5 or 10 μl) and probe ratios (0.1 μl 3’ and 0.1 μl 5’ or 0.05 μl 3’ and 0.1 μl 5’) were used.

The plate was then either pre-blocked with 200 μl 5% BSA, 5% Casein or 5% milk at 37°C for one hour after which wells were washed three times with 1×PBS, or not pre-blocked. The solutions were mixed and incubated at 37°C for one hour or overnight. Then, 96 μl ligation solution was added and incubated at 37°C for 10 min. For the protease reaction, 2 μl protease solution was added and incubated at 37°C for 10 min; at 95°C for five min and stored at 4°C. The RT-PCR was performed with the LightCycler 480 in a 384-wells plate where 9 μl of the protease treated product was added to 11 μl PCR master mix (provided with the TaqMan Kit). The plate was run (95°C for two min, 40 cycles of 95°C for 15 sec, 60°C for one min), and Ct values were analysed.

Trans-Blot Turbo Western blotting system

Protein samples isolated with treatment buffer were supplemented with 20% (v/v) glycerol, 5% (v/v) β-mercaptoethanol and 0.001% (w/v) bromophenol blue and heated at 95°C for five min. Alternatively, samples were heated in XT sample buffer and XT reducing agent (Bio-Rad Laboratories, the Netherlands). Total protein was loaded on 1.0 mm thick Criterion XT Tris acetate (poly-acrylamide) gels with a linear resolving gel gradient of 3–8% (Bio-Rad Laboratories). The reference serial dilution was made with protein that corresponded to the sample muscle type. A protein mixture of several WT mice was produced and aliquoted to prevent repeated freeze-thaw cycles and protein degradation. Gels were run on ice at 75 V (∼0.07A) for one hour and at 150 V (∼0.12A) for two hours in the provided running buffer (XT Tricine; Bio-Rad Laboratories). The buffer was refreshed to maintain pH when the voltage was changed to 150 V.

The separated proteins were blotted on a nitrocellulose membrane using the ready to use Trans-Blot Turbo transfer packs and the Trans-Blot Turbo transfer system from Bio-Rad at 2.5A (∼25 V) for 10 min (standard Bio-Rad protocol for high molecular weight proteins). See the Supplementary data for an extended protocol. To determine blotting efficiency, the gel was stained with coomassie brilliant blue R250 staining (overnight) and destained with water for three hours during which the solution was refreshed for five times. The gel was dried for two hours at 80°C in a gel dryer.

The nitrocellulose blots were blocked for one hour with 5% (w/v) non-fat dried milk powder (ELK Campina Melkunie, the Netherlands) in a buffer containing 10 mM Tris-HCl (pH 8.0) and 0.15 M NaCl (Tris-buffered saline (TBS)). Membranes were washed three times for 10 min with TBST buffer containing 10 mM Tris-HCl (pH 8.0), 0.15 M NaCl and 0.005% (v/v) Tween20 followed by an overnight incubation with the primary monoclonal dystrophin antibody (NCL-DYS1, Novocastra Laboratories) diluted 1:125 in TBS at RT with gentle agitation. The following day blots were washed three times for 15 min in TBST. Then, blots were incubated with the secondary antibody containing an infrared dye (IRDye 800CW goat-anti-mouse IgG, Li-COR, USA, dilution 1:5000 in TBS) for one hour.

After a final washing step (two times 20 min in TBST and a final wash in TBS for 20 min) signals were visualized and quantified using the Odyssey system and software (Li-COR). Samples that contained air bubbles or other artifacts were excluded from quantification.

Loading controls

Alternative proteins to be used as loading controls were tested using antibodies against alpha-actinin (1:7500; AB72592, Abcam, UK); alpha-tubulin (1:1000, T6199-200UL, Sigma-Aldrich, the Netherlands) and sodium-potassium adenosine triphosphate (anti-Na-K ATPase 1:100, α-F6 from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by the University of Iowa, Department of Biology, Iowa City, IA [29]). In case of double staining with dystrophin, primary antibodies were simultaneously incubated overnight. The secondary antibody used for alpha-actinin was the IRDye 680LT (1:10000 donkey-anti-rabbit IgG, Li-COR, USA) and IRDye 800CW (1:5000 goat-anti-mouse IgG, Li-COR, USA) was used for both alpha-tubulin and anti-Na-K ATPase.

RESULTS

Applicability of three ELISA-based methods

To assess the potential of ELISA to quantify dystrophin levels we performed a series of experiments using an indirect ELISA, sandwich ELISA and a commercially available indirect ELISA kit.

For the indirect ELISA, multiple experiments showed that there was almost no difference between optical densities (OD) of mdx samples (negative control) and WT samples (positive control). We consistently observed that ODs of wells with only solutions and no proteins added were slightly lower than wells containing protein lysates, suggesting that there might be a component in the protein isolation buffer that interfered during the ELISA reactions as sample OD values rarely exceeded 0.5. To rule this out Ripa and treatment buffers were used for protein isolation with different concentrations of SDS (0–20% ). However, ODs remained indistinguishable between mdx and WT samples (data not shown). To reduce background noise and increase specificity we also tested different 96-wells plates, blocking buffers (milk, FBS, BSA and casein), primary antibodies (NCL-DYS1, 2, 3 and C20) and concentrations of primary and secondary antibodies. Even though ODs of samples incubated with C20 were higher than those of samples treated with NCL-DYS1, 2 or 3, we still could not distinguish between mdx and WT samples.

As we could not get the direct ELISA to work in our hands despite our extensive optimization attempts, we instead focused on the sandwich ELISA, following the protocol of Morris et al. [17]. We have tested several protein isolation buffers with varying SDS concentrations and concentrations of capture antibody and protein samples. Even though OD values obtained for this sandwich ELISA were higher than for the indirect ELISA (1.5 – 2 versus <0.5), OD values of non-coated or with water incubated wells were generally higher than sample values, and we could never distinguish between mdx and WT samples. The use of other dystrophin antibodies as capture (Mandra 1 and Manex) and recognition (biotinylated Manex and C20) antibodies did not reduce background noise.

Finally, we investigated the applicability of a commercially available ELISA kit for dystrophin. Despite the fact that the ODs of the dilution series made from the standard protein provided corresponded to the increasing concentrations added, ODs of mice with low dystrophin levels or no dystrophin were comparable or even higher than WT ODs. Notably, ODs of water were higher than expected in all experiments (∼0.3), indicating high background noise. Optimization attempts varying protein isolation lysate and sample concentrations did not decrease background noise.

Applicability of the TaqMan protein assay

The TaqMan protein assay utilizes the recognition of proteins by specific antibody pairs to which a 5’ and 3’ Prox-Oligonucleotide is conjugated that ligates to a third oligonucleotide used for quantification by RT-PCR. As we were primarily interested in detection of full-length dystrophin, four monoclonal antibody pairs recognizing the C-terminal part of the dystrophin protein were selected with <300 amino acids between them (Table 1). Initial optimization attempts were undertaken with the antibody combination of Mandra 5 and 15, which both passed the conjugation test. Unfortunately, independent of the protein isolation buffer (either 0% SDS, Ripa buffer containing 0.1% SDS or protein quant lysis buffer containing TritonX100), or protein concentration used, Ct values of the mdx and WT samples were comparable and were not lower than those of negative controls, suggesting that the third quantification oligonucleotide was unable to bind. We hypothesized that this could result from binding of the antibodies to the plate’s plastic surface instead of the protein and therefore blocked for one hour with either 5% BSA, 5% Casein or 5% milk. As this did not decrease sample Ct values, we also varied relative probe quantities (2, 5 or 10 μl probe mix, and/or unequal amount of probes, i.e. 0.05 μl 3’ and 0.1 μl 5’). Again, these alterations did not decrease the sample Ct values.

To rule out that the antibodies selected were the underlying problem, we tested the three other monoclonal antibody combinations (mandra 1, 5 and 15, manex, Table 1). Additionally, we also improved binding opportunities for the antibodies and probes by increasing the incubation time from one hour to overnight. Again, this did not improve results as the Ct values of both WT and mdx samples remained high (above 28). As a final attempt, we tested polyclonal antibodies recognizing potentially longer epitopes of 49 (C20) and 299 (H300) amino acids, arguing that perhaps this would lead to a better coverage of the dystrophin protein enabling the quantification oligonucleotide to bind. Also with these antibodies Ct values of samples and negative controls were high (>28) and did not allow us to distinguish between them.

Applicability, reproducibility and detection limits of Trans-Blot Turbo system

To assess the potential of the Trans-Blot Turbo system, protein lysates from diaphragm of mdx -Xist^Δhs mice, which have varying levels of dystrophin due to skewed X-inactivation [12] were analyzed (representative blot is shown in Fig. 1A, a flow diagram is shown in Fig. 1B). Serial dilutions from protein lysates of Xist^Δhs mice (100% dystrophin) of the corresponding muscle were used as a reference concentration series to determine dystrophin levels in mdx-Xist^Δhs samples with the Odyssey system. Using the XT sample buffer and XT reducing agent (recommended for the Trans-Blot Turbo system) instead of a homemade treatment buffer (see materials and methods) did not improve the results (data not shown). We loaded 25 μg total protein in 20 μl total volume of the samples. The blot showed the expected pattern (typical double band pattern that is commonly seen using the NCL-DYS1 antibody). As anticipated, a serial dilution of the reference sample from 25 μg to 0.825 μg total protein (100% to 3.3% dystrophin as compared to the amount of protein loaded for the mdx -Xist^Δhs samples) revealed a decrease in intensity of the specific dystrophin bands with lower protein concentrations, while the lowest amount of protein (3.3% dystrophin) was still clearlydetectable.

To test the reproducibility of the Trans-Blot Turbo blotting system, the quadriceps of 22 mdx/utrn^–/–-Xist^Δhs mice expressing low dystrophin levels were tested 3–7 times on individual blots (Fig. 2). The relative standard deviation (% Change) was 30.23% (Table 2).

To assess the upper and lower detection limits, we investigated various total protein concentrations by preparing a reference concentration series of lysates derived from WT mouse heart, varying from 50 to 0.6 μg (Fig. 3A). It can be appreciated that the intensity of the bands from 50 μg to 35 μg are more or less the same. We speculate that this non linearity might be due to incomplete blotting for samples containing >25 μg protein, due to reaching the maximum binding capacity of the blotting membrane. As the Trans-Blot Turbo system is a semi-dry blotting system, incomplete blotting can be visualized by coomassie blue staining of the gel after blotting (i.e. to detect the unblotted protein) (Fig. 3B). Coomassie staining showed that indeed at these high concentrations, some protein at the molecular mass of dystrophin was left behind in the gel after blotting. We maximized loading to 25 μg total protein, resulting in a proper transfer of proteins of the size range of dystrophin and assessed the lower detection limit (Fig. 3C and D). Notably, dystrophin was still detectable in 0.08 μg total protein (0.3% ), suggesting that the Trans-Blot Turbo system is very sensitive. Loading up to 25 μg total protein resulted in a good reference concentration curve, while using the 50 μg reference concentration curve there is a tendency to underestimate protein levels lower than 60% (Fig. 3E).

We further confirmed sensitivity by quantifying dystrophin levels in diaphragm and triceps of mdx mice treated with AONs targeting exon 23 (Fig. 4A-B). Dystrophin levels after eight weeks of treatment were low, therefore we loaded 30 μg total protein and a reference concentration series containing 10% , 3.3% ; 1.1% and 0.4% dystrophin. Despite the very low dystrophin levels in these samples, they could be detected and quantified. In contrast, while dystrophin was detectable by eye on immunostainings of the same triceps muscles (Fig. 4C), quantification could not be performed with software tools available in our Institute and most likely require dedicated software analyses tools such as described by Beekman et al. [30]. To assess reproducibility of the system, we quantified dystrophin levels on three to four individual blots from quadriceps samples from 27 AON treated mdx mice (Table 3). Dystrophin levels were on average 0.79% with a standard deviation of 0.34% . Next, we evaluated whether loading higher total protein concentrations (up to 200 μg) of treated mdx muscle would allow better quantification. However, the abundance of total protein hampered electrophoresis and detection was worse due to wavy bands (data not shown).

To confirm that the Trans-Blot Turbo system can also be used for human lysates, a serial dilution of human control muscle lysate (from 25 to 1.66 μg) was analyzed (Fig. 4D). The immunoreactive bands are present at the same molecular mass as the mouse lysates and again show the typical double band pattern. The lowest dilution could still be easily detected. Thus, normal human dystrophin can be detected as well as mouse dystrophin.

Protein loading controls

Since not all myosin is properly blotted using the Trans-Blot Turbo system (Fig. 3B and D), myosin is clearly unsuitable as a loading control. Therefore we tested multiple proteins to be used instead. Fig. 5A shows the results of a double staining with dystrophin (1:125) and alpha-actinin (1:7500) as primary antibodies and IRDye 800CW (1:5000) and IRDye 680LT (1:10000) respectively as secondary antibodies. An advantage of using the Odyssey for detection and visualization of proteins of interest is that it is possible to perform a double staining with two different colored fluorescent secondary antibodies, raised in different species (dystrophin: Mouse monoclonal, alpha-actinin: Rabbit polyclonal). It can be appreciated from Fig. 5A that the intensity of the alpha-actinin bands is largely similar for the different samples, while the dystrophin levels vary as expected in the mdx- Xist^Δhs samples. To confirm that alpha-actinin was completely blotted we tested for protein residues by coomassie staining of a gel after blotting. Unlike myosin, alpha-actinin was completely transferred at 30 μg total protein (Fig. 5B). Furthermore, alpha-actinin levels were similar between different dystrophic and WT animal models and between different skeletal muscles (data not shown).

We also tested other loading controls (alpha-tubulin (data not shown), anti-Na-K ATPase), but these were less suitable. Alpha-tubulin varied a lot between samples and was differently expressed between WT and dystrophic muscles. In addition, the low molecular weight (50 kDa, compared to 427 kDa for dystrophin) hampers good detection since proteins <55 kDa run off the gel in the protocol we optimized for dystrophin, while alpha-actinin (103 kDa) remains on the gel. Anti-Na-K ATPase was previously published [29] as a loading control for dystrophin Western blotting. Fig. 5C shows our findings with this antibody. Though bands are seen at the expected molecular mass in all tested dilutions, we prefer to use the alpha-actinin as loading control as this can be used at higher dilutions and gives nicer bands in our hands.

DISCUSSION

The two techniques that are routinely used for dystrophin quantification are Western blotting and immunohistochemical analysis [14, 31]. Alternative methods are available for dystrophin quantification [15–17, 22], of which only the sandwich ELISA has been previously used in preclinical settings. However, its use has been reported by others in the Duchenne field only once [18]. Here, we have extensively tested several ELISA methods for dystrophin quantification. Unfortunately, neither the previously published procedures nor the other (commercial) ELISA methods did work in our hands. We speculate that this is due to the stickiness of dystrophin antibodies to plastics and their very low abundance. For the commercial ELISA method the absorbance of the provided standard did correlate with the serial dilutions. However, this standard was not a recombinant full-length dystrophin but rather a shorter peptide which was diluted in buffer rather than protein lysate.

Additionally, we assessed the applicability of the TaqMan protein assay [19]. Despite intensive testing, the technique did not work in our hands, as we were never able to pick up dystrophin levels above the background noise. Since high SDS concentrations (i.e. >10% SDS) hamper the ligation reaction, we were forced to isolate protein in low SDS concentrations. Unfortunately, this reduces the amount of dystrophin isolated by at least seven-fold [22], rendering it even more difficult to detect this low abundant protein. Additionally, a detergent like SDS is warranted to improve accessibility for antibodies, which is especially essential for large proteins like dystrophin. Finally, the negative outcomes of this assay might have partly resulted from the suboptimal epitope length of the antibodies used (generally >100 amino acids is recommend, while all except one epitope tested were <100 amino acids long), however, also the H300 antibody (with an epitope of 299 amino acids) did not amplify signal.

Alternatively, we propose a new, reliable and sensitive Western blot system for dystrophin quantification. When using the Trans-Blot Turbo system, several things need to be taken into account. Firstly, due to the sensitivity of the Trans-Blot Turbo system (indicated by the low detection limit) low amounts of total protein can be loaded. The use of more than 25 μg total protein leads to an overcapacity of the membrane, resulting in an inefficient transfer of the protein from the gel to the membrane during blotting. Some non-linearity can also be seen when 25 to 15 μg total protein is loaded (Fig. 3C). This most likely results partly from improper transfer and competition for the antibody. Therefore, when high dystrophin levels are expected, loading should be limited to 15 μg of total protein, whereas higher total protein concentrations can be used when low dystrophin levels are expected.

Secondly, when performing experiments with delicate patient material, or AON treated mdx mice, containing very low amounts of dystrophin we would recommend a reference concentration series in the dynamic range of the dystrophin levels expected in the samples. The low dystrophin concentrations in treated human and mouse samples, which are probably around the lower detection limit of the membrane, pose some challenges. For mouse samples these are only minor, since we were able to readily detect very low dystrophin levels in AON treated mdx mice when 30 μg total protein was loaded. However, for patients there is an added challenge in that the quality of the muscle is often impaired. Thus when isolating total protein this will often be a combination of protein from fibrotic/adipose tissue and muscle tissue, while only the latter is able to produce dystrophin. Therefore, higher amounts of protein have to be loaded which could be hampered by the limited loading capacity of the Trans-Blot Turbo system (40 μl can be loaded into a slot).

To confirm equal loading of total protein for all samples, a loading control is warranted. Myosin has been often used, but due to its high abundance in muscle (∼40–50% of total protein in muscle protein lysates) it is not completely blotted and therefore unreliable as a loading control. Alternatively, we propose alpha-actinin as a reliable loading control. A big advantage is the possibility to perform double staining of dystrophin and alpha-actinin, resulting in a direct, two colored, read-out on the Odyssey.

In summary, ELISA and TaqMan protein assays were unsuitable to quantify dystrophin levels in our hands. Alternatively, we present a reliable, sensitive and easy to use system for performing Western blots with which dystrophin levels can be assessed in a variety of lysate samples in a standardized manner. This system should be helpful for preclinical research aiming to restore dystrophin.

CONFLICTS OF INTEREST

The authors have no conflicts of interest to report that are related to the work described here.