Dynamics of pax7 expression during development, muscle regeneration, and in vitro differentiation of satellite cells in rainbow trout (Oncorhynchus mykiss)

Essential for muscle fiber formation and hypertrophy, muscle stem cells, also called satellite cells, reside beneath the basal lamina of the muscle fiber. Satellite cells have been commonly identified by the expression of the Paired box 7 (Pax7) due to its specificity and the availability of antibodies in tetrapods. In fish, the identification of satellite cells remains difficult due to the lack of specific antibodies in most species. Based on the development of a highly sensitive in situ hybridization (RNAScope®) for pax7, we showed that pax7+ cells were detected in the undifferentiated myogenic epithelium corresponding to the dermomyotome at day 14 post-fertilization in rainbow trout. Then, from day 24, pax7+ cells gradually migrated into the deep myotome and were localized along the muscle fibers and reach their niche in satellite position of the fibres after hatching. Our results showed that 18 days after muscle injury, a large number of pax7+ cells accumulated at the wound site compared to the uninjured area. During the in vitro differentiation of satellite cells, the percentage of pax7+ cells decreased from 44% to 18% on day 7, and some differentiated cells still expressed pax7. Taken together, these results show the dynamic expression of pax7 genes and the follow-up of these muscle stem cells during the different situations of muscle fiber formation in trout.


Introduction
Skeletal muscle consists of multinucleated cells called myofibers that result from the fusion of muscle precursors cells called myoblasts. Myofibers formation (hyperplasia) occurs during the embryonic and fetal period and subsequently is restricted to muscle regeneration in adult mammals, whereas in large growing fish, hyperplasia persists throughout the post-larval period 1  Initially, satellite cells were identified by electron microscopy, based on their anatomical location between the myofiber membrane and the basal lamina, in numerous mammalian species, but only in a few fish such as carp 6 and zebrafish 7 . Subsequently, satellite cells have been localized by the expression of many specific genes such as desmine, M-Cadherin, myf-5, etc 2 . Among these satellite cell markers, the Paired box 7 (Pax7) is the most widely used marker due to its specificity and the availability of antibodies 8 . In mammals, pax7 is expressed in quiescent and activated satellite cells and is essential for satellite cell survival, cell fate and self-renewal 8 . In zebrafish, pax7a and pax7b are expressed by satellite cells and are required for white muscle regeneration 7 . In fish, the identification of satellite cells remains difficult due to the lack of specific antibodies in most species. In zebrafish, Pax7 immunolabeling partially recapitulates the pattern of pax7 expression visualized in a transgenic line expressing GFP under the control of the pax7a promoter 9 .
Surprisingly, very rare Pax7 positive cells were identified in the white muscle of adult zebrafish 7 .
Using a heterologous antibody, Steinbacher et al (2007) were able to localize Pax7 positive cells in the myotome of brown trout embryos 10 . In giant danio (Devario cf. Aequipinnatus), Pax7 immunolabeling shows that almost all the cells (>99%) freshly extracted from white muscle are positive for Pax7, whereas Pax7-positive cells were no longer detected 24h later 11 . In rainbow trout, despite numerous attempts with appropriate controls, we were unable to obtain a specific pax7 signal in muscle or in vitro using the same antibody and protocol as Froehlich et al., (2013).
Furthermore, using in situ hybridization, pax7 expression was readily detected in the dermomyotome during embryonic development 12,13 , but not in the white muscle of juvenile trout (data not shown). Therefore, the dynamics of migration of pax7-positive cells to their final niche in fish remains unknown.
Based on the development of a highly sensitive in situ hybridization technique 14 (RNAScope ® ), this work aims to describe the dynamics of pax7 expression during embryonic and larval stages, muscle regeneration and in vitro satellite cell differentiation in rainbow trout.

Animals
All the experiments presented in this article were developed in accordance with the current legislation governing the ethical treatment and care of experimental animals (décret no. 2001-464, May 29, 2001), and the muscle regeneration study was approved by the Institutional Animal Care

Muscle regeneration experiment
This experiment was carried out at the INRAE facility PEIMA. Briefly, 1530 ± 279 g rainbow trout (O. mykiss) were anesthetized with MS-222 (50 mg/l) and using a sterile 1.2 mm needle, the left side of each fish was injured by a puncture behind to the dorsal fin and above the lateral line. The right side of each fish was used as a control. White muscle samples were collected from both sides (in the injured region and contralateral) at 0, 1, 2, 4, 8, 16, and 30 days post-injury using a sterile scalpel after euthanasia by an overdose of MS-222 (200 mg/l). The obtained samples were properly stored in liquid nitrogen until further processing for gene expression analysis. No infection was detected during the experiment and the survival rate was 100%. Another muscle regeneration experiment was performed on 30 g trout and samples were collected 18 days after injury and fixed with 4% paraformaldehyde overnight at 4°C and embedded in paraffin.

Trout satellite cell culture
Satellite cells from trout white muscle (200-250g body weight) were cultured as previously described 15 . Briefly, muscle tissue was mechanically and enzymatically digested (collagenase, C9891 and trypsin, T4799) prior to filtration (100µm and 40µm). Cells were seeded on poly-Llysine and laminin precoated Lab-Tek™ II Chamber Slide™ (#154534, ThermoScientific, 8 wells) at a density of 80,000 cells/ml and incubated at 18ºC in DMEM (D7777) with 10% fetal bovine serum to stimulate cell proliferation and differentiation. Cells were washed twice with PBS and fixed with ethanol/glycine (pH2) from day 2 to day 7 of culture.

RNA extraction, cDNA synthesis, and quantitative PCR analyses
Total RNA was extracted from 100 mg of muscle using TRI reagent (Sigma-Aldrich, catalog no. T9424), and its concentration was determined using the NanoDrop ND-1000 spectrophotometer.
One µg of total RNA was used for reverse transcription (Applied Biosystems kit, catalog no. were designed at exon-exon junctions to avoid amplification of genomic DNA. Secondary structure formation in the predicted PCR product were determined with the mFOLD software. Quantitative PCR analyses were performed with 5 µl of cDNA using SYBR© Green fluorophore (Applied Biosystems), according to the manufacturer's instructions, with a final concentration of 300 nM of each primer. The PCR program used was as follows: 40 cycles of 95 °C for 3 s and 60 °C for 30 s.
The relative expression of target cDNAs within the sample set was calculated from a serial dilution (1:4-1:256) (standard curve) of a cDNA pool using StepOneTM software V2.0.2 (Applied Biosystems). Real-time PCR data were then normalized to elongation factor-1 alpha (eF1α) gene expression as previously described 16 .

In situ hybridization
For the detection of pax7 expression in trout embryos, samples were fixed with 4% paraformaldehyde overnight at 4°C and embedded in paraffin. Cross-sections (7 µm) of muscle were then cut using a microtome (HM355; Microm Microtech, Francheville, France) and in situ hybridization was performed using RNAscope Multiplex Fluorescent Assay v2 (Bio-Techne #323100) according to the manufacturer's protocol. Sections were briefly baked at 60°C for 1 hour, deparaffinized, and air dried. After 10 min in hydrogen peroxide solution ( Bio-Techne #322335), sections were treated with 1X Target Retrieval (#322000; Bio-Techne ) for 15 min at 100°C, followed by 25 min with Protease Plus solution (#322331; Bio-Techne) at 40°C. Due to the presence of three pax7 genes in the trout genome 17 , we designed a set of probes targeting pax7a1, pax7a2 and pax7b mRNA (see supplementary Figure 2). This probe set was hybridized for 2 hours at 40°C. All steps at 40°C were performed in an Bio-Techne HybEZ II hybridization system (#321720). The RNAscope Multiplex Fluorescent Assay allows simultaneous visualization of up to three RNA targets, with each probe being assigned to a different channel (C1, C2 or C3) with its own amplification steps. For embryo sections, pax7 transcripts were targeted with the fluorescent dyes Opal 520 (#FP1487001KT; Akoya Biosciences). At the end of the in situ hybridization protocol, embryo sections were incubated with a primary antibody against salmon Collagen 1 (#520171; Novotec, France) and then with a secondary antibody conjugated with an Alexa Fluor 594 fluorescent dye (#A11005; ThermoFisher). Nuclei were counterstained with DAPI (0.5µg/ml) and sections were mounted with ProlongGold (P36930, Invitrogen).
Detection of pax7 expression was also performed using the chromogenic RNAscope® 2.5HD detection reagent RED kit (#322360; Bio-Techne) in sections of trout embryos and white muscle from 4 g and 9 g fish. As previously described, samples were fixed in 4% paraformaldehyde overnight at 4°C, embedded in paraffin and cross sections (7 µm) were made. Pax7 in situ hybridization was performed as described above and chromogenic detection using Fast Red, was performed according to the manufacturer's protocol. Nuclei were counterstained with DAPI (0.5 µg/ml) and sections were mounted with EcoMount (EM897L, Biocare medical). In this chromogenic RNAscope assay, the red signal can be observed with either a white light or fluorescence microscope.
For multiplex RNAscope in situ hybridization, fixed cells were hybridized using the RNAscope Multiplex Fluorescent Assay v2 (Bio-Techne #323100) according to the manufacturer's protocols.
Wheat germ agglutinin (WGA) conjugated with Alexa 488 (Molecular Probes # W11261) was used to visualize connective tissue and basal lamina 18 . WGA is a lectin-based molecule specifically binds to N-acetyl-D-glucosamine and N-acetylneuraminic acid (sialic acid) residues. After four washes with PBS, sections were stained with WGA (5 µg/ml) for 3 hours at room temperature.
All the images were taken with a Canon digital camera coupled to a Canon 90i microscope.

Automated quantification of pax7 + and mmx + cells
To automatically quantify the number of cells expressing pax7, mmx or pax7 and mmx, we adapted a macro-command 19 on Fiji software 20 to quantify puncta corresponding to the RNAscope labeling, per cell. Due to the presence of one or two puncta in some cells with the negative control (data not shown; DapB probe #504081, Bio-Techne), a cell was considered positive when at least 3 puncta were detected. Our quantification method is available on https://gitlab.univ-nantes.fr/SJagot/fijimacro_rnascopecells.

Statistical analyses
Data were analyzed using the non-parametric Kruskal-Wallis rank test followed by the post-hoc Dunn test. All analyses were performed with the R statistical package (version 4.2.1).

Pax7 positive cells were detected in the deep myotome before hatching
To determine the stage of appearance of pax7 + cells in the deep myotome, we performed highly sensitive in situ hybridization (RNAscope technology) with pax7 probes on cross sections of 14-28 dpf embryos (Figure 1). At day 14, pax7 + cells were detected in the neural tube and in the undifferentiated myogenic epithelium corresponding to the dermomyotome. At day 19, no pax7 + cells were detected in the fully differentiated deep myotome, excepted near to the horizontal septum, where rare pax7 + cells were observed. Two days later (D21), the majority of pax7 expressing cells were still localized in the dermomyotome, and very rare pax7 + cells appeared scattered throughout the deep myotome. From day 24 to day 28, pax7 + cells were readily detected in the deep myotome, while a slight decrease in pax7 expression in the dermomyotome epithelium was observed. In situ hybridization of longitudinal sections of 28-days-old embryos showed that pax7 + cells were localized along the muscle fibers of the deep myotome (Figure 2A, B) and not in the myoseptum.
The final location of satellite cells is beneath the basal lamina of the muscle fibers. To determine when pax7 + cells reach their niche, we stained the basal lamina with Alexa 488-conjugated wheat germ agglutinin. In Figure 3, pax7 + cells were easily detected at day 37, when the basal lamina was very thin and did not completely surround the muscle fibers. At day 47 (100mg), the basal lamina was thicker and surrounded most of the fibers and some of the pax7 + cells. At day 112 (4g) and 136 (9g), all pax7 + cells were located beneath the basal lamina surrounding the muscle fibers.

Pax7 positive cells accumulate at the lesion site during regeneration
In vertebrates, muscle regeneration requires the activation, proliferation and differentiation of satellite cells (pax7 + ). To determine whether the expression of pax7 genes (pax7a1, pax7a2 and pax7b) is upregulated during muscle regeneration in trout, we examined the kinetics of regeneration at 0, 1, 2, 4, 8, 16, and 30 days after injury. QPCR analysis showed that the expressions of pax7a1, pax7a2 and pax7b were all upregulated after injury with a maximal expression at day 8 with an increase of 8-, 12-and 5-fold, respectively. Thereafter, the expression of pax7a1 and pax7a2 tended to decrease while that of pax7b remained stable until day 30 post-injury ( Figure 4A, 4b, 4C).
To determine whether pax7 + cells accumulate at the site of injury during regeneration, we performed in situ hybridization for Pax7 on sections of regenerated muscle at day 18 ( Figure 4D). At the site of injury, we observed accumulation of nucleus (DAPI) and extracellular matrix (WGA labeling).
Consistent with qPCR results, pax7 + cells accumulated at the site of injury in contrast to the uninjured muscle area, where few pax7 + cells were observed.

The proportion of Pax7 + cells decreases during in vitro differentiation of myogenic cells
To determine the evolution of the proportion of pax7 + cells during in vitro differentiation, we performed in situ hybridization for pax7 and myomixer (mmx) on cultured muscle cells. Figure 5A shows the presence of mononucleated cells expressing pax7 and few cells expressing mmx at day 2 of culture. At day 7, we observed the presence of multinucleated myotubes with a strong expression of mmx and mononucleated cells expressing pax7 + . Surprisingly, at day 7 we observed few cells expressing both pax7 and mmx, although the intensity and the number of red dots (pax7 labeling) were lower compared to what was observed at day 2. Quantification of the proportion of pax7 + , mmx + and pax7 + /mmx + cells is shown in Figure 5B. On day 2, the culture was composed of 44% pax7 + cells, 5% of mmx + cells and 6% of pax7 + /mmx + cells. The percentage of pax7 + cells decreased to 18% on day 7, while the percentage of mmx + and pax7 + /mmx + cells increased up to 19% and 31%, respectively.

Discussion
Adult muscle stem cells (satellite cells, pax7 + cells) are essential for fiber growth and muscle regeneration. Although the existence of these cells has been demonstrated in fish, the lack of highly sensitive and specific tools to label satellite cells has limited the study of their function. The aim of this work was to describe the dynamics of satellite cells (pax7 + ) during embryonic and larval stages, regeneration and in vitro differentiation in rainbow trout.
In most fish, the observation of satellite cells in adult muscle is difficult due to their small size and number, and the available tools are not sensitive and specific enough. In this context, we decided to take advantage of the recent improvement in in situ hybridization sensitivity using the novel technology RNAScope ®,14 , to localize pax7 + cells in trout muscle. In this species, we have previously identified three pax7 genes (pax7a1, pax7a2 and pax7b) resulting from the salmonidspecific whole-genome duplication 17,21 . The coding sequences of these 3 genes are highly similar, sharing more than 88% of sequence identity. Given this high sequence identity, our probe set is able to detect the three pax7 mRNAs as shown in Supplementary Figure 2, and thus all the cells expressing at least one pax7 gene. Using this method, we were able to observe a pax7 signal in embryos, white muscle and isolated satellite cells with no background (see Supplementary Figure   1).
In fish, pax7 expression has previously been detected by classical in situ hybridization in the dermomyotome of zebrafish 22 and trout 13,23 until to the end of somitogenesis. Using RNAScope ® technology, our results confirmed the presence of pax7 + cells at the periphery of the myotome corresponding to the dermomyotome at day 14, assessing the specificity of the pax7 probes. At the end of segmentation (D19-D21), pax7 + cells gradually appeared in the deep myotome from day 24 post fertilization. In addition, pax7 + cells were observed scattered throughout the somite, suggesting that pax7 + cells directly migrated directly from the dermomyotome to the deep myotome by moving between superficial muscle fibers. Thus, our results support the data obtained in zebrafish 24 and brown trout 25 showing that pax7 + cells also migrate between muscle fibers and not around their ends. Double labeling of pax7 and the basal lamina, indicated that in trout, pax7 + cells are located beneath the basal lamina from day 112 post-fertilization, in line with observations in zebrafish, where pax7 + cells are also surrounded by the basal lamina as early as 6 days post-fertilization 7,9 . In addition, in situ hybridization for pax7 performed in muscle sections clearly showed the presence of pax7 + cells scattered in white muscle in juveniles (D136, ~9g), in agreement with with data obtained with heterologous antibody in salmon 26 but in contrast to zebrafish, where very few pax7 + cells are detected in white muscle 7 . This difference may be due to the persistence of a high rate of muscle hyperplasia in salmon in contrast to zebrafish 27 , which should require a high number of pax7 + cells.
In mammals, satellite cells are known to be required for muscle hyperplasia and hypertrophy, but also for muscle regeneration after injury. In trout, we have previously shown that mechanical injury induces the resumption of myogenesis as evidenced by the upregulation of myogenin, myomaker and myomixer and the formation of new small fibers 28,29 . The present results showed that during muscle regeneration the three pax7 genes (pax7a1, pax7a2 and pax7b) showed a peak of expression 8 days after injury, well before the peak of expression of myogenin and the formation of new fibres (D30) 28 . In addition, in situ hybridization with pax7 probes at day 18, showed a large number of pax7 + cells at the wound site compared to the uninjured area. These results are consistent with previous work in adult 7 and larval stage 9 zebrafish showing that white muscle regeneration requires activation and proliferation of pax7 + cells. These results strongly suggest that in trout pax7 + cells are required for muscle regeneration in trout and that the three pax7 genes are involved. Indeed, in zebrafish, the two pax7 genes (pax7a and pax7b) have been shown to have distinct functions, whereas in trout the specific roles, if any, of the three pax7 genes remain unknown.
In mammals, pax7 has been shown to be mainly expressed in quiescent and activated satellite cells 2 .
Our results showed that 2 days after muscle cell extraction, 50% of the cells expressed pax7 (44% pax7 + and 6% pax7 + /mmx + ) and were thus myogenic progenitors. In agreement with this result, using an antibody against MyoD, we observed that 60-70% of the extracted muscle cells were positive for MyoD 15 . The small difference may be due to the different markers used and to the fact that the trout in the present study are heavier (250g versus 5-10g). During the cell culture, the percentage of pax7 + cells decreased from 44% to 18% on day 7, while the percentage of mmx + cells increased from 6% to 19%, in agreement with previous qPCR data 29  In conclusion, the in situ RNAScope ® technology allowed us to localize the pax7 + cells with high sensitivity and specificity. We showed that pax7 + cells migrate into the deep myotome at the end of segmentation and reach their niche after hatching. In addition, we observed an accumulation of pax7 + cells at the wound site, suggesting their requirement for muscle regeneration and that pax7 expression decreased in differentiating myogenic cells. Further work is needed to understand the effect of experimental conditions (fasting, aging, temperature) on the number of satellite cells.    The asterisk indicates significant differences between treatments at a given point. Statistical significance (p < 0.05) was determined by the Kruskal-Wallis rank test followed by the Dunn test.

Figure Legends
Localization of pax7 + cells (red) in 18-day injured muscle was performed by in situ hybridization (D). The extracellular matrix was stained with wheat germ agglutinin (WGA, green) and the nucleus was stained with DAPI (blue). The asterisks indicate the injury site and the scale bar corresponds to 50 µm.