Purification and Validation of Mouse and Human DM1 Pericytes

Pericytes can be identified by the presence of alkaline phosphatase (ALP).3, 6, 32 In contrast to other pericyte markers such as neuron-glial antigen 2 (NG2) and platelet-derived growth factor receptor beta (PDGFRβ), ALP is not expressed in skeletal muscle fibers, nor in other myogenic progenitors, but it is restricted to the microvasculature of striated muscle in postnatal life6 and is therefore an appropriate selection marker. Expression of this phosphatase by pericytes enabled us to distinguish them from PAX7⁺ or MYOD⁺ satellite cells, which might also be present in the explant cultures. Moreover, pericytes lack endothelial markers such as CD31.3

To obtain ALP⁺ cells from the mixed population of outgrown mouse cells, we sorted these cells via fluorescent-activated cell sorting (FACS) on the presence of ALP and absence of CD31 on day 7 (Figures 1C and 1D; Figures S1A and S1B). Enzymatic ALP staining in all cells after sorting confirmed our selection protocol (Figure 1B). Due to the presence of blood cells in the human cultures, it took 7 days longer for an outgrowth ring of cells to appear. Cell sorting of five cell lines (control-derived lines C1 and C2, and patient-derived lines P1, P3, P6) showed that via replating under pericyte-favorable conditions, we had already established a selection for ALP⁺ and CD31⁻ cells during cell culture (Figures 2C and 2D; Figures S1C and S1D). Consequently, the last three patient-derived lines (P2, P4, P5) were not sorted but were validated via enzymatic ALP stain (Figure 2B). We thus were able to collect pure ALP⁺ cultures from all participants (Table 1).

After sorting of mouse and human ALP⁺ cells, we further confirmed the cell type via immunocytochemistry. A combination of pericyte markers alpha smooth muscle actin (α-SMA), NG2, and PDGFRβ, combined with absence of MHC, clearly demonstrated both identity and purity of our pericyte isolates from both mouse and human muscle (Figure 3).

Open in a separate window

Figure 3

Validation of Purified Pericytes by a Combinatorial Staining Approach

Immunofluorescent staining of mouse and human pericytes for Pdgfrβ/PDGFRβ, α-Sma/α-SMA, and Ng2/NG2 expression. Mhc/MHC, a marker for myogenic differentiation, was not expressed. Control cell lines: DMSXL myoblasts (Pdgfrβ), C25 human control myoblast (α-SMA), SH-SY5Y human cell line (NG2), and DMSXL myotubes (Mhc). Scale bar, 50 μm.

Homozygous (HOM) DMSXL mice show high neonatal mortality, muscle defects, and growth retardation, whereas heterozygous (HET) mice show milder muscle symptoms and no increase in neonatal death.29 As expected from these phenotypes, we found a strong positive correlation (R² = 0.70) between body weight of the mice and the amount of skeletal muscle isolated from the hind limbs (Figure 1E). Unexpectedly, however, no correlation existed between the total weight of the muscle biopsy used for the explant cultures and the number of pericytes obtained after sorting (R² = 0.03; Figure 1F), and comparable numbers of pericytes could be isolated from HOM, HET, and wild-type (WT) mice. Importantly, we cannot exclude that the abundance of pericytes in skeletal muscle tissue in vivo varies between the genotypes.

Human pericytes maintained their diploid karyotype after at least nine passages (Figure 2E) and showed a typical triangular refractile morphology for at least 20 population doublings, with a doubling time of approximately 40 h (Figure 2F). Notably, the proliferation rate of all pericyte cultures was comparable and independent of donor, MIRS, sex, and participants’ age (note that no growth curve was measured for P6, as this biopsy was used to optimize the isolation protocol).

To examine the progenitor cell state at higher passages, we re-sorted mouse pericyte cultures at passage 19. The entire population of cells still expressed Alp (Figures S2A and S2B), so no differentiation toward myoblasts or contamination of additional cell types was encountered. For human pericytes, we performed MHC staining at passage 20 to verify that they did not differentiate toward MHC⁺ myotubes. MHC staining was negative and all cells still expressed ALP (Figure S2C).

Characterization of Disease Miomarkers in DM1 Pericytes

The DM1 mutation is an autosomal-dominant, unstable (CTG)n repeat in the 3′ untranslated region of DMPK. The repeat mutation ends up as a long (CUG)n stretch in pathogenic, expanded transcripts, which initiate a broad range of detrimental downstream effects.33 We investigated the expression of endogenous Dmpk and transgenic DMPK transcripts in pericytes from the DMSXL line. Dmpk expression in proliferating pericytes was similar between cells isolated from WT, HET, and HOM mice, as well as DM500 myoblasts (Figure 4A). DM500 myoblasts are an immortalized cell lineage that was isolated earlier from the same transgenic mouse line, when the (CTG)n repeat still contained only ~500 triplets.17, 28 No significant differences were found in DMPK (CTG)1300 transcript levels between HET and HOM pericytes, carrying one and two transgene copies, respectively. As expected, no DMPK expression was detected in WT pericytes (Figure 4B). DMPK expression in human pericytes was remarkably similar between patients and controls, and also comparable to that in human myoblasts (Figure 4C).

Open in a separate window

Figure 4

Dmpk/DMPK Expression in Mouse and Human Pericytes

qRT-PCR quantification of Dmpk and DMPK expression. Endogenous Dmpk (A) and transgenic DMPK (B) expression in pericytes from WT, HET, and HOM DMSXL mice. DMPK expression was significantly higher in HOM and HET pericytes compared to WT pericytes. Immortalized DM500 mouse myoblasts (MB) were included as reference (n = 2).17 p < 0.05, p < 0.0001. (C) DMPK expression in pericytes from unaffected controls and DM1 patients. Immortalized human DM1 myoblasts (MB) were included as reference (n = 1).50 Error bars indicate SEM.

According to the RNA dominance theory, expression of DMPK RNA with abnormal repeat length may affect proper muscle development and have consequences for regenerative muscle capacity in adulthood.34 Both the timing and the level of expression of pathogenic repeats will influence the type and the extent of down-stream toxicity caused by for example RNP binding or RAN translation products.17, 35 Pericytes isolated from HET and HOM DMSXL mice displayed expanded DMPK transcripts in the form of fluorescent in situ hybridization (FISH)-detectable RNP complexes in the nucleus, generally termed foci, after hybridization with a CAG repeat probe. The number of foci varied strongly between cells, but, on average, 1–2 foci per nucleus were present in pericytes from HET mice versus 5–10 foci in cells from HOM mice (Figure 5). Few foci were detected in some WT mouse cells, which has been reported before for myoblasts and may represent background staining recorded by the automated image analysis.17 Especially noticeable was the high variation in foci count in pericytes from HOM animals, ranging between 0 and >25 per nucleus.

Open in a separate window

Figure 5

DM1 Foci in Transgenic Mouse Pericytes

DM1 foci were visualized by RNA in situ hybridization using a fluorescent (CAG)6 probe in mouse pericytes. (A) Quantification of the number of foci per nucleus for pericytes from WT, HET, and HOM mice. At least three mice per genotype were included, and at least 20 nuclei per mouse were analyzed with a minimum of 125 nuclei per genotype. Statistical analysis was performed between genotypes. There are significantly more foci present in pericytes derived from HET and HOM mice compared to pericytes from WT mice and significantly more foci are visualized in the HOM-derived pericytes compared to pericytes from HET mice. Data points represent the foci number per nucleus; the mean ± SEM is plotted in the graph. *p < 0.05, ***p < 0.0001. (B) Representative RNA FISH images. Scale bar, 20 μm.

We also examined the occurrence of FISH foci in human pericytes. Pericytes from P1, P2, P3, and P5 showed an average of 1–2 foci per nucleus. No foci could be detected in pericytes from P4 and P6, presumably due to the smaller repeat lengths, which reduces the number of fluorescent probes that can bind to one transcript. As expected, no foci were detected in control pericytes (Figure 6; Figure S3). To verify MBNL1 protein accumulation in the (CUG)n RNA foci, we combined FISH with immunofluorescent detection of MBNL1. Nuclei in DM1 patient pericytes, but not controls, consistently showed focal accumulation of MBNL1 protein, which colocalized with (CUG)n RNA FISH foci (Figure 6). The number of MBNL foci counted per nucleus was slightly lower than the number of FISH foci, which is probably due to automated image analysis and thresholding.

Open in a separate window

Figure 6

DM1 Foci in DM1 Patient Pericytes

(A) Automated quantification of nuclear (CUG)n and MBNL1 foci for P1, P2, P3, and P5. At least 60 nuclei per pericyte population were analyzed. (B) Simultaneous detection of DM1 foci by RNA in situ hybridization using a fluorescent (CAG)6 probe and immunofluorescence using a MBNL1 antibody in human pericytes. Error bars indicate SEM. Scale bar, 10 μm.

DMSXL Mice

DMSXL mice (C57BL/6 background) used in this study carry a 45-kb genomic fragment encompassing the DM1 locus of a patient, including a (CTG)1300 repeat.28, 29, 56 Mouse number 4, 5, 11, and 12 were females; all others were males. Transgenic status was assessed by PCR on ear punch DNA. The DMSXL phenotype includes reduced muscle strength, lower motor performances, and respiratory impairment, which is essentially only apparent in HOM mice. On the molecular level, weak splicing abnormalities and RNA foci are present in skeletal muscle from HOM mice.30, 57

Human Participants

DM1 patients and unaffected controls visited Radboudumc (Nijmegen, the Netherlands) where a skeletal muscle biopsy was taken from the quadriceps femoris (vastus lateralis) using a Bergström needle.58 Specifications of the eight participants can be found in Table 1. MIRS was the latest score of clinical evaluation.31 (CTG)n repeat length was determined in blood DNA at diagnosis of the disease.

Cell Culture Media

Aliquots of 80 mL fetal calf serum (FCS, Sigma-Aldrich) were heat inactivated in a water bath at 56°C for 40 min. Horse serum (HS, Sigma-Aldrich) was aliquoted into 10 mL and heat-inactivated at 56°C for 30 min. Proliferation medium for mouse pericytes (DMEM20) was made by combining DMEM (AQmedia, Sigma-Aldrich) with 20% heat-inactivated FCS, 50 IU/mL penicillin, 0.05 mg/mL streptomycin (Sigma-Aldrich), 10 mM nonessential amino acids (Sigma-Aldrich), 1 mM sodium pyruvate, and 50 mM β-mercaptoethanol (Gibco) over a 0.22-μm filter. Proliferation medium for human pericytes (DMEM5) has a similar composition but without sodium pyruvate, only 5% heat-inactivated FCS, and an additional 5 ng/mL human basic fibroblast growth factor (Sigma-Aldrich). Differentiation medium contained DMEM with 2% heat-inactivated horse serum, 50 IU/mL penicillin, 0.05 mg/mL streptomycin, and 1 mM sodium pyruvate. Coating of cell culture surfaces was performed with collagen from calf skin (Sigma-Aldrich) at 1 mg/mL dissolved in 20% glacial acetic acid. Then, 1 mg/mL collagen (20% glacial acid) was applied to the surface and left for at least 15 min at room temperature. The solution was removed and the collagen-coated wells and/or flasks were transferred to a dedicated 30°C oven to dry the surface for 24 h. The next day, the surface was washed twice with 1× PBS and once with a medium containing phenol red to ensure no glacial acid was left behind.

Human and Mouse Explant Cultures

DMSXL mice were euthanized via CO₂ and weighed. Skeletal muscles of both hind limbs (GPS complex [gastrocnemius-plantaris-soleus complex]) of 4- to 5-week-old mice were dissected, weighed, and collected in DMEM20 with as little manipulations as possible to increase the yield of cells. Tissue from human participants was collected in 1× PBS on ice. For both mouse and human skeletal muscle biopsies, the following procedures were performed in a laminar flow hood. Muscle tissue was transferred into a 10-cm dish containing the appropriate medium. Any residues of fat or tendon were carefully removed while the tissue was dissected into ~2-mm³ fragments with a sterile round-shaped scalpel. For human material, around half of the muscle biopsy already consisted of small fragments due to the collection procedure. To avoid contamination with blood as much as possible, tissue fragments were washed repeatedly with 1× PBS. Around 12 muscle fragments were then placed per well in a collagen coated six-well plate. To ensure maximal yield, the muscle fragments were distributed over the surface of the well. 200 μL of prewarmed DMEM5 (human explant) or DMEM20 (mouse explant) was placed on top of all fragments in one well, and an additional 400 μL was dropwise added to the six-well plate. MQ (milliQ [ultrapure water]) was added to the space between wells to ensure a humidified environment and keep the tissue from drying out and detaching. Explants were incubated at 37°C for 18–24 h in a 5% CO₂/5% O₂ incubator. The next day, 1.5 mL prewarmed DMEM20 or DMEM5 medium was slowly added to the side of the well to avoid detachment. Explant cultures were then left undisturbed for 2–3 days. Remaining erythrocytes started to sprout, after which elongated, attached fibroblast-like cells migrated from the fragment, followed by pericytes. Cells initially looked round and small and then attached to the surface. After 4–5 days for mouse explants (7–8 days for human explants), the outgrowth of pericytes was visible and an additional 500 μL of prewarmed medium was added. From here onward fragments were monitored daily to inspect the extent of cell outgrowth from the explants. After 6–7 days for mouse explants (14–15 days for human explants), a 1-cm outgrowth of cells in a wavy pattern in all wells was present, surrounding almost every muscle fragment in the well. Cells were collected by trypsinization for FACS.

FACS

Media and harvested cells were collected over a 70-μm cell strainer to separate cells from the explants. Cells were counted and subsequently resuspended in 200 μL staining medium, i.e., HBSS Ca²⁺ Mg²⁺ (Sigma-Aldrich) supplemented with 2% heat-inactivated FCS and 10 mM HEPES (Sigma-Aldrich). Staining was performed with phycoerythrin-conjugated monoclonal anti-human/mouse ALP, clone B4-78 (R&D Systems), at 10 μl/10⁶ cells and 0.3 μl per sample of viability dye eFluor 780 (eBioscience) for 45 min at 4°C in the dark. Control samples and FACS samples (ideally >1 × 10⁶) were prepared in sterile FACS-suitable capped polystyrene tubes. Thereafter, the cells were washed three times with 1× PBS and filtered over a 70-μm filter to be ready for cell sorting. Cell sorting was performed on a BD FACSAria flow cytometer (BD Biosciences). Gates were set on fluorescence minus one (FMO) (human samples) and negative controls (mouse samples). ALP⁺ pericytes were collected in the appropriate media. Collected cells were cultured at 37°C in a 5% CO₂/5% O₂ humidified incubator. Cell cultures were passaged upon 70%–80% confluence at a 1:4 ratio. Cells were disassociated by Tryple Express (Invitrogen) into a single-cell suspension and replated on collagen-coated wells/flasks.

Characterization of Pericytes

Cells were seeded on glass coverslips coated with calf skin collagen. Coated coverslips were washed with 1× PBS, and 10,000–15,000 cells per cm² were seeded. One day later, coverslips were washed with 1× PBS at room temperature. Fixation methods varied between conditions. Myotube cultures were fixed with 2% paraformaldehyde (PFA) in 0.1 M phosphate buffer (pH 7.4) for 15 min at RT. For immunodetection of NG2 (ab129051, Abcam), α-SMA (ab7817, Abcam) and MHC (MF20 DSHB, University of Iowa) cells were fixed in 4% PFA at RT for 10 min. In case of immunodetection of PDGFRβ (MA5-15143, Thermo Fisher), cells were fixed in 1:1 acetone/methanol for 10 min at −20°C. Cells were subsequently washed three times for 5 min with 1× PBS, after which they were permeabilized with 0.2% Triton X-100 (Sigma-Aldrich) for 10 min. Cells were incubated with blocking buffer (0.1% Triton X-100 [Sigma-Aldrich], 4% normal goat serum [NGS, Sigma-Aldrich], and 0.1% BSA [Sigma-Aldrich] in 1× PBS) for 60 min. Incubation with primary antibodies was done in blocking buffer, overnight at 4°C. Cells were then washed again three times with 1× PBS and incubated with 4 μg/mL goat anti-mouse Alexa Fluor 568 or goat anti-rabbit Alexa Fluor 488 (Thermo Fisher) and 100 ng/mL 4′,6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich) in blocking buffer for 60 min at RT in the dark. Cells were washed with 1× PBS and mounted on slides. Fluorescent images were acquired with a Leica DMI6000B microscope using a ×63 objective.

For enzymatic ALP staining, cells were washed with 1× PBS, fixed for 5 min with 4% PFA, and washed again with 1× PBS before being covered with BCIP/NBT solution (B1911, Sigma-Aldrich). Images were acquired with the ZOE fluorescent cell imager (Bio-Rad Laboratories).

Analysis of Myotube Characteristics and Image Analysis

To assess myogenic capacity via MHC staining, 30,000 pericytes were seeded in eight-well ibidi chambers coated with 1 mg/mL collagen in 20% glacial acetic acid and propagated for 2 days in growth medium until 100% confluency. Culturing in differentiation medium continued for 8 more days. Cells were washed with 1× PBS and fixed with 2% PFA in 0.1 M phosphate buffer for 15 min at RT. After fixation, cells were washed three times with 1× PBS and permeabilized for 10 min with 0.2% Triton X-100 in 1× PBS. Blocking was performed at RT for 1 h with blocking buffer containing 4% NGS (Sigma-Aldrich), 0.1% Triton X-100 (Sigma-Aldrich), and 0.1% BSA (Sigma-Aldrich). Samples were incubated overnight at 4°C with antibody dilution buffer with triton (ADB-T) containing 1:20 diluted anti-MHC antibody (MF20 DSHB, University of Iowa). Cells were washed three times with 1× PBS and incubated with goat anti-mouse Alexa Fluor 488 and 100 ng/mL DAPI (Sigma-Aldrich) in blocking buffer for 1 h at RT in the dark. Cells were washed three times with 1× PBS and stored in 1× PBS. Fluorescent images were acquired using a Leica DMI6000B microscope with a ×20 objective. Four-by-four tile scans were made to fit the large myotubes within the images.

Fiji software (ImageJ v1.52d) was used for analysis of myotube formation. The myogenic fusion index (MFI) was calculated as 100 × the ratio of the number of nuclei inside MHC⁺ myotubes to the number of total nuclei. The number of DAPI-positive nuclei in myotubes was determined by counting the number of nuclei that co-localized with positive MHC staining using the multipoint tool of ImageJ. Total nuclei number was determined by autothresholding the DAPI channel with Huang, applying a watershed on the binary image to separate overlaying nuclei and automatic cell count using analyze particles (settings: 200 to infinity for pixel size). Using the Bezier curve tool, myotube length and width were determined. The longest myotube branch was measured while the broadest part of the myotubes was measured to determine the average width. For these measurements, 20 myotubes per sample were measured from two independent experiments. All quantifications were performed in duplicate by two independent researchers.

qRT-PCR

In human samples, DMPK expression levels were determined at passage 8. To quantify expression levels in cells from WT, HET, and HOM DMSXL mice, qRT-PCR was performed in three biological replicates (three mice per genotype) and two technical replicates. Cell pellets of passage 5–7 were collected and stored in 350 μL of RA1 lysis buffer (Macherey-Nagel NucleoSpin RNA, lot no. 75858). RNA isolation for all samples was performed using the NucleoSpin RNA isolation kit (Macherey-Nagel) according to the manufacturer’s protocol. Next, cDNA synthesis with random hexamers from the iScript cDNA synthesis kit (Bio-Rad Laboratories) was performed using 2 μg of RNA as a template in a reaction volume of 40 μL. For qRT-PCR, cDNA was diluted 40×. The cDNA sample was mixed in a final volume of 25 μL containing GoTaq qPCR Master Mix, distilled H₂O, and 0.4 pmol of each primer. Samples were analyzed using the 7900HT Fast Real-Time PCR System (Thermo Fisher). RT minus and no-template control were included as negative controls. A DM500 myoblast RNA sample17 was included as a positive control for analysis of Dmpk expression, while DM11 cells50 were included as a reference for DMPK expression. Gene-of-interest levels were normalized to Gapdh and Hprt1 levels for mouse pericytes. TBP and HPRT1 were included as reference genes for human pericytes.

PCR primers used were as follows: Gapdh e2–e3, 5′-CGTAGACAAAATGGTGAAG-3′ and 5′-GTTGATGGCAACAATCTC-3′; Hprt1 e6–e7, 5′-GACACTGGTAAAACAATGC-3′ and 5′-CTGTATCCAACACTTCGAG-3′; Dmpk e2–e3, 5′-TTTTGAAGGTGATCGGGCGTG-3′ and 5′-CCTCTCTTCAGCATGTCCCACTTA-3′; DMPK e15-3′, 5′-TGCCTGCTTACTCGGGAAA-3′ and 5′-GAGCAGCGCAAGTGAGGAG-3′; TBP e3, 5′-CCACTCACAGACTCTCACAAC-3′ and 5′-CTGCGGTACAATCCCAGAACT-3′; HPRT1 e7–e9, 5′-TGACAGTGGCAAAACAATG-3′ and 5′-GGTCCTTTTCACCAGCAAGCT-3′.

We are indebted to all participants who engaged in this research and made it possible, and to Joost Raaphorst, Fran Smulders, and Yvonne Cornelissen (Nijmegen, Netherlands) who assisted with recruitment and collection of human muscle biopsy material. We are grateful to Prof. Maurilio Sampaolesi (Leuven, Belgium) and Florence van Tienen (Maastricht, Netherlands), who taught us the technique of pericyte isolation. We thank Prof. Geneviève Gourdon (Paris, France) for the DM500 and DMSXL mouse lines. We thank the members of the department of Cell Biology, especially Prof. Bé Wieringa, for helpful discussions and advice. This work was funded by the Donders lnstitute for Brain Cognition and Behavior, Radboud University Medical Center, and supported by the Prinses Beatrix Spierfonds (grant W.OR18-18; with contribution from the Stichting Spieren voor Spieren).