Myotome adaptability confers developmental robustness to somitic myogenesis in response to fibre number alteration

Summary Statement Homeostatic interactions between muscle stem cells and fibres during myogenesis ensure the correct muscle size is formed independent of fibre number in zebrafish Abstract Balancing the number of stem cells and their progeny is crucial for tissue development and repair. Here we examine how muscle stem/precursor cell (MPC) numbers are tightly regulated during zebrafish somitic muscle development. MPCs expressing Pax7 are initially located in the dermomyotome (DM) external cell layer, adopt a highly stereotypical distribution and thereafter a proportion of MPCs migrate into the myotome. Regional variations in the proliferation and terminal differentiation of MPCs contribute to growth of the myotome. To probe the robustness of spatiotemporal regulation of MPCs, we compared the behaviour of wild type (wt) MPCs with those in mutant zebrafish that lack the muscle regulatory factor Myod. Myodfh261 mutants form one third fewer multinucleate fast muscle fibres than wt and show a significant expansion of the Pax7+ MPC population in the DM. Subsequently, myodfh261 mutant fibres generate more cytoplasm per nucleus, leading to recovery of muscle bulk. In addition, relative to wt siblings, there is an increased number of MPCs in myodfh261 mutants and these migrate prematurely into the myotome, differentiate and contribute to the hypertrophy of existing fibres. Thus, homeostatic reduction of the excess MPCs returns their number to normal levels, but fibre numbers remain low. The GSK3 antagonist BIO prevents MPC migration into the deep myotome, suggesting that canonical Wnt pathway activation maintains the DM in zebrafish, as in amniotes. BIO does not, however, block recovery of the myodfh261 mutant myotome, indicating that homeostasis acts on fibre intrinsic growth to maintain muscle bulk. The findings suggest the existence of a critical window for early fast fibre formation followed by a period in which homeostatic mechanisms regulate myotome growth by controlling fibre size.


Introduction 25
How tissue size is regulated is largely unknown, but depends on both the number of cells 26 and their size. When the 'correct' size is reached, growth ceases. Although signalling 27 pathways such as IGF, BMP, TOR and Hippo have been implicated in tissue size control 28 However, quantitative mechanistic understanding of how DM cell dynamics are controlled 9 within the somite and relate to later fibre formation is lacking. 10 We have previously shown that the zebrafish myotome rapidly increases in volume during 11 the pre-and post-hatching period, growing threefold between 1 and 5 days post-12

Growth of zebrafish muscle 32
Muscle fibre cross sectional area was determined in embryonic, larval and adult zebrafish 1 (Fig. 1). Mean fibre size increased dramatically in the embryonic period, less rapidly during 2 larval life, slowly beyond 5 months and appeared to plateau after 1 year of age (Fig. 1A). 3 In adults, fibre types were distinguished by myosin heavy chain (MyHC) content (Fig. 1B). 4 As reported previously (Patterson et al., 2008), slow fibres were smaller than the adjacent 5 intermediate fibres, with the more numerous fast fibres in the deep myotome being the 6 largest ( Fig. 1A). Paralleling the rapid increase in fibre size from 1-5 dpf, somites increase 7 in mediolateral width ( Fig. 1C; p=0.011). Growth also involves increase in fibre number 8 (Fig. 1D). The smallest fibres in developing somites are located near the DM, particularly 9 at the epaxial and hypaxial somitic extremes, suggesting that new fibres arise from DM 10 cells (Stellabotte et al., 2007). Although fibre number increases between 1-3 dpf (Fig. 1D  11 and see below), mean fibre size doubles, despite the lowering effect on mean fibre size of 12 small newly-added fibres (Fig. 1A,D). As around five new slow fibres are formed between 13 1-3 dpf (Barresi et al 2001), the remaining new fibres must be fast. Counts of nuclei within 14 the myotome also show a 20% increase (Fig. 1D). The increase in myotome nuclei is 15 sufficient to yield five mononucleate slow fibres and twenty extra fast fibres, but does not 16 double like fibre size (Fig. 1D). As shown below, these trends continued until at least 6 17 dpf. Thus, both increase in fibre volume per nucleus and addition of nuclei to fibres by 18 precursor myoblasts contribute to myotome growth. 2014), we analysed the changing numbers of Pax7 + cells in defined somitic zones (Fig.  1   S1). Immunolabelling of Pax7 in larvae prior to 4 dpf revealed that most Pax7 + cells were 2 located at myotome borders (DE,HM,VM), with the remainder in the superficial CP, the 3 central DM (Figs 1E and 2A). Subsequently, Pax7 + cells appeared deep within the somite 4 ( Fig. 2A). As no temporal differences in Pax7 + cell behaviour in the epaxial and hypaxial 5 somite were noted at any stage examined, and as numbers of Pax7 + cells per epaxial 6 somite did not vary detectably along the rostrocaudal axis from somites 14-22 (Fig. S2), 7 we chose to explore changes in Pax7 + cell number in the epaxial domain of somites 15-20. 8 Between 3 and 6 dpf, the total number of Pax7 + cells per epaxial somite increased by 9 about 50%, from ~40 to ~60 cells ( Fig. 2B; p = 0.003). Strikingly, Pax7 + cell numbers did 10 not change significantly in the superficial DM; the increase in Pax7 + nuclei was accounted 11 for by a rise deep within the somite ( Fig. 2A,B; p < 0.001). 12 To understand how Pax7 + cells arise in the deep somite, the locations of Pax7 + cells were 13 characterized at successive stages. In 3 dpf larvae, about half the Pax7 + cells were at the 14 superficial VM, mostly oriented with their long axes parallel to the VM ( Fig. 2A-C). Some 15 Pax7 + cells were in the superficial CP and DE regions, and a few were located deep within 16 the somite at the HM ( Fig the Pax7 + cells that invade CP myotome rapidly differentiate and contribute to myotome 20 growth.
We conclude that although terminal differentiation removes Pax7 + cells, 21 proliferation and migration is sufficient to replenish the Pax7 + cell pool. 22

Altered growth in myod fh261 mutants 23
Embryos lacking myod function show a 50% reduction in muscle at 1 dpf accompanied by 24 a twofold excess of MPCs. This followed by rapid growth resulting in recovery by 5 dpf 25 muscle. Analysis of fibre number and size revealed that at 5 dpf myod fh261 mutants have 27 33% fewer fibres, but these are about 30% larger than those in wt ( Fig. 5A-C). Thus, 28 recovery compensated for the reduced fibre number by increased fibre growth. 29 Analysis of fibre number and size during muscle recovery showed significant defects in 30 myod fh261 mutants. In wt embryos, fibre numbers increased by about 15% from 1-5 dpf. 31 No increase was observed in myod fh261 mutants (Fig. 5B). In contrast, fibre size was 32 comparable in wt and myod fh261 mutants at 1 dpf, but fibre size increased faster in mutants 1 so that, by 5 dpf, fibres were larger, thereby compensating for the reduction in fibre 2 number (Fig. 5C). 3 Myod fh261 mutants have an increased number of Pax3/7 + cells at 1 dpf, paralleling the 4 reduction in muscle differentiation (Hinits et al., 2011). The number of Pax7 + cells remains 5 elevated at 3 dpf, but returns almost to normal by 5 dpf, accompanied by a rise in the 6 number of myonuclei in the myotome (Fig. 5D,E). 7 Do myod fh261 mutants recover by fusion of the excess myoblasts into the pre-existing 8 fibres? Both the maximal number and the mean number of nuclei in single fast muscle 9 fibres of myod fh261 mutants was higher than in siblings (Figs 5F,G and S6), suggesting that 10 the excess myoblasts contribute to the growth of existing myotomal fibres. However, when 11 the fibre size per nucleus was calculated (by dividing the cross-sectional area of each fibre 12 by its nuclear number), myod fh261 mutants had a clear increase in effective nuclear domain 13 size (Figs 5H and S6). Thus, fusion of excess Pax7 + DM cells into pre-existing fibres 14 during recovery of myod fh261 mutants accompanies hypertrophy -an increase in fibre 15 volume per nucleus. 16 17

Excess Pax7 + cells in the deep myotome of 3 day myod fh261 mutants 18
To understand the contribution of DM cells to the recovery of myod fh261 mutants the 19 number and location of Pax7 + and Myog + cells were determined (Fig. 6). In 3 dpf 20 myod fh261 mutants there were approximately 25% more Pax7 + cells than in wt (p=0.025, 21 myotome of myod fh261 mutants than in siblings and wt fish, which accumulate such cells 24 from 4 dpf (Fig. 3). 25 The extra Pax7 + cells in 3 dpf myod fh261 mutants appear to be undergoing differentiation. 26 More Myog + cells were observed in the deep CP of myod fh261 mutants at 3 dpf, compared 27 with their siblings (Fig. 6F-H). Moreover, these increases were observed specifically in the 28 deep central myotome (Fig. 6G,H), and not at other locations of the myotome. Comparing 29 myod fh261 mutants and siblings, a similar fraction of Pax7 + cells were also Myog + and the 30 ratio of Pax7 + Myog + to Pax7 -Myog + cells was unaltered between genotypes. These 31 findings suggest that, once in the deep central myotome, myod fh261 mutant Pax7 + cells 32 progress to terminal differentiation in the normal manner. At 5 dpf, no significant 1 difference in either Pax7 + or Myog + cell numbers persisted (Fig. S7A-C). These data argue 2 that the premature appearance of Pax7 + cells in the deep central myotome does not reflect 3 a failure of differentiation in myod fh261 mutant, but rather an adaptive process contributing 4 to increase in nuclei/fibre and muscle mass recovery. 5 If the increase in MPCs in the deep central myotome reflects an adaptive process, it could 6 arise either from increased migration or proliferation of Pax7 + cells. To examine this issue, 7 myod fh261 was crossed onto the pax7a:GFP transgene to permit tracking of cell dynamics. 8 Profiles suggesting migration of cells from the vertical myosepta into the deep myotome 9 were more common at 3 dpf in myod fh261 mutants than in siblings ( Fig. 6B and data not 10 shown). Moreover, EdU labelling showed that cell proliferation was similar in the deep 11 central myotome in mutants and siblings ( Fig. S7D-F). We conclude that there is an 12 increased migration of Pax7 + cells into the deep myotome of 3 dpf myod fh261 mutants. 13 14

Blockade of GSK3 reduces accumulation of Pax7 cells in the myotome 15
A small molecule screen for pathways that affect Pax7 + cell behaviour revealed that GSK3 16 signalling may regulate migration (Fig. 7A CP myotome compared with vehicle (Fig. 7A). Quantification revealed that the number of 22 GFP + cells was significantly decreased in the deep CP of BIO-treated larvae, but relatively 23 unaffected elsewhere (Fig. 7B). The numbers of differentiating Pax7 + cells and Myog + cells 24 were also significantly reduced in both superficial and deep CP (Fig. S8). Importantly, BIO 25 also blocked the premature entry of Pax7 + cells into the myotome in myod fh261 mutants 26 (Fig. 7C). Thus, BIO prevents migration and/or accumulation of Pax7a + cells in the deep 27 CP and alters terminal differentiation. 28 29 BIO fails to reduce compensatory muscle fibre growth 30 As GSK3 inhibition led to fewer Pax7 + cells migrating into the deep myotome, it was 1 possible to investigate the importance of this migration in recovery of the muscle in 2 myod fh261 mutants. When myod fh261 mutant embryos were treated with BIO at 2 dpf, thus 3 blocking the premature ingression of Pax7 + cells to the deep CP, the increase in muscle 4 fibre volume triggered by the loss of Myod still occurred (Fig. 7D). Interestingly, BIO 5 caused a slight rise in fibre number in wt siblings, but in myod fh261 mutants there was no 6 significant change in fibre number (Fig. 7E). These data indicate that hypertrophy of fibres 7 by increase in volume per nucleus provides the robust recovery of muscle in myod fh261 8 mutants and migration of Pax7 + cell from DM is not required for this recovery. 9 10 Discussion 11 The present work demonstrates that Pax7 + MPCs contribute to larval muscle growth and 12 contains four major findings relevant to myogenesis. First, we show that Myod function is 13 required for formation of the correct number of fast muscle fibres. Second, we describe In summary, we examined the role and regulation of Pax7 + MPCs in larval muscle growth. 25 We observed tight regional control of MPC numbers, distribution and behaviour within the 26 somite and myotome. Perturbations that alter muscle size and MPC number were rapidly 27 corrected, suggesting the existence of a homeostatic mechanism that senses muscle size 28 and ensures robust development in the face of environmental and genetic insults. 29 30

Zebrafish lines and maintenance 32
Genetically-altered Danio rerio (listed in Table S1) on a primarily AB background were 1 reared at King's College London on a 14/10hr light/dark cycle at 28.5°C (Westerfield, 2 2000). BIO (0.5 µM) or DMSO vehicle were added to fish water. 3 4 Immunodetection and S-phase labelling 5 Fibre sizes on photomicrographs of cryosections either unstained or after 6 immunoperoxidase detection of MyHC were quantified with OpenLab (Improvision). For 7 wholemounts, larval pigmentation was suppressed with 0.003% 1-phenyl-2-thiourea 8 (Sigma) added at 12 hpf. Larvae were fixed with 2% PFA for 25 minutes, washed with 9 PBTx (PBS, 0.5% or 1% (4 dpf+) Triton-X100) and incubated in primary antibody (see 10   Table S2)  Fixed fish were imaged using the 10x/0.3 air or 40x/1.1 water immersion objectives. Three 3 to nine somites around the anal vent were imaged from lateral using the tile scan Z-stack 4 function. Short stack maximum intensity projections, specific slices or cross-sectional 5 views were exported as tiffs. Nuclear number was determined from three equi-spaced 6 transverse images from somite 17 of each embryo. Cells were counted in original ZEN 7 stacks and allocated to regions (Fig. S1) in confocal stacks of epaxial somites of 8 wholemount fish by scanning through in the XZ direction while toggling channels. 9 Xanthophores were excluded from Pax7 counts based on nuclear shape, location and  (Fig. 1A,B), KL (Fig. 1D), SK 30 ( Fig. 3A,B) and RDK (Fig. 7A,B). RDK proposed and initiated the BIO experiments. SMH 1 conceived the project, provided advice and wrote the manuscript with input from all 2 authors.