{"paper_id":"bb8be7f2-5f2c-48a6-9acd-f454fea08962","body_text":"1\nThe role of MEGF10 in myoblast fusion and hypertrophic response to overload of skeletal muscle \n \nLouise Richardson1*, Ruth Hughes1*, Colin A Johnson2, Stuart Egginton3, Michelle Peckham1* \n*Equal contribution \n \n1School of Molecular and Cellular Biology, University of Leeds, Leeds, LS2 9JT \n2Faculty of Medicine, University of Leeds, LS2 9JT \n3School of Biomedical Science, University of Leeds, Leeds, LS2 9JT \n \n*Corresponding author: m.peckham@leeds.ac.uk \nORCID ID: 0000-0002-3754-2028 \n \nAcknowledgments \n \nWe would like to acknowledge funding from a Medical Research Council (MRC) research grant MR/M000532/1 \n(to CAJ), a Sir Jules Thorn Biomedical Award JTA/09 (to CAJ). RH was funded by an MRC PhD studentship \naward (1233630). LR was funded by a Biotechnology and Biological Sciences Research Council (BBSRC) White \nRose Doctoral Training Partnership PhD studentship (BB/M011151/1). \n \nAbstract \nBiallelic mutations in multiple EGF domain protein 10 ( MEGF10) gene cause EMARDD (early myopathy, \nareflexia, respiratory distress and dysphagia) in humans, a severe recessive myopathy, associated with reduced \nnumbers of PAX7 positive satellite cells. To better understand the role of MEGF10 in satellite cells, we \noverexpressed human MEGF10 in mouse H-2k\nb-tsA58 myoblasts and found that it inhibited fusion. Addition of \npurified extracellular domains of human MEGF10, with (ECD) or without (EGF) the N-terminal EMI domain to \nH-2k\nb-tsA58 myoblasts, showed that the ECD was more effective at reducing myoblast adhesion and fusion by day \n7 of differentiation, yet promoted adhesion of myoblasts to non-adhesive surfaces, highlighting the importance of \nthe EMI domain in these behaviours. We additionally tested the role of Megf10 in vivo using transgenic mice with \nreduced (Megf10\n+/-) or no ( Megf10-/-) Megf10. We found that the extensor digitorum longus muscle had fewer \nPax7 positive satellite cell nuclei and was less able to undergo hypertrophy in response to muscle overload \nconcomitant with a lower level of satellite cell activation. Taken together, our data suggest that MEGF10 may \npromote satellite cell adhesion and survival and prevent premature fusion helping to explain its role in EMARDD.  \n \nKey words (4-6) \nMEGF10, satellite cells, myogenesis, skeletal muscle, overload model \n \nShort running title: The role of MEGF10 in skeletal muscle  \n \n  \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 2\n \nIntroduction \nMyopathies, diseases of skeletal muscle, impair the ability of a muscle to regenerate in response to damage, \neither directly affecting muscle fibre activity or indirectly through effects on muscle stem (satellite) cells. In the \nrecessive congenital myopathy, EMARDD (early myopathy, areflexia, respiratory distress and dysphagia), the \nability of muscle to regenerate is impaired, but the reason for this is still unclear.  Skeletal muscle fibres in \nEMARDD patients have a reduced diameter, fewer nuclei per fibre and lack Pax7\n+ satellite cells (Logan et al., \n2011). The disorder is caused by mutations in MEGF10 (multiple epidermal growth factor-like domains 10). \nMEGF10, the membrane protein encoded by this gene, has been suggested to be important for satellite cell \ninteraction with the extracellular matrix (Logan et al., 2011). Mutations in MEGF10 that cause EMARDD appear \nto reduce proliferation and migration of activated satellite cells, resulting in fewer myogenic cells that can \neventually fuse together to form new adult myofibres (Holterman et al., 2007; Li et al., 2021; Saha et al., 2017). \nMEGF10 has also been suggested to promote satellite cell proliferation, whilst regulating myogenic differentiation \n(Holterman et al., 2007).  \nIn addition to its role in satellite cells, MEGF10 has been reported to be required for engulfment, a similar role \nto that reported for the C. elegans CED-1 protein (Callebaut et al., 2003; Holterman et al., 2007; Suzuki and \nNakayama, 2007b), an orthologue of MEGF10. Exogenous expression of MEGF10 in Hela cells induces these \ncells to engulf the protein GULP (Hamon et al., 2006). This role in engulfment is particularly important in the \nbrain, where MEGF10 is highly expressed in astrocytes and likely plays a role in synapse trimming (Park and \nChung, 2023). More recently MEGF10 has been found to be enriched in neuromuscular junctions with a possible \nrole in modifying these synapses (Juros et al., 2024).  \nMEGF10 has a large extracellular domain that comprises an N-terminal EMI domain followed by 17 epidermal \ngrowth factor (EGF)-like domains (Fig. 1A). The EMI domain was first described in proteins in the EMILIN \nfamily of glycoproteins and is associated with protein multimer formation (Colombatti et al., 2011). The EMI \ndomain of MEGF10, whilst having the conserved consensus sequence at the C-terminus (WRCCPG(Y/F)xGxxC), \nhas only 6 rather than the 7 cysteine residues found in many EMI domains, and thus is more similar to the EMI \ndomain in multimerin. This domain has been predicted to interact with the membrane phospholipid \nphosphatidylserine (PS), which is exposed on the surface of apoptotic cells as a signal marking the cell for \nphagocytosis via TTR-52 (Tung et al., 2013), consistent with the potential role of MEGF10 in engulfment. \nInterestingly, PS is also exposed on the surface of skeletal muscle myoblasts during fusion (Jeong and Conboy, \n2011; van den Eijnde et al., 2001) and the PS receptor BAL1 promotes myoblast fusion (Hochreiter-Hufford et al., \n2013). Thus, the EMI domain of MEGF10 could also play a role in myoblast-myoblast adhesion in fusing cells, \nvia PS. \nThe EGF domains of MEGF10 comprise two different forms. There are 12 EGF-like domains composed of 7 \nconserved cysteine rich residues and 5 laminin-type EGF-domains that contain 8 conserved cysteine residues \ncapable of forming four disulphide bonds. EGF domains may have a role in mediating intercellular signalling and \nact in receptor-ligand interactions (Wouters et al., 2005). Thus, these extracellular domains mediate a role for \nMEGF10 in cell adhesion (Suzuki and Nakayama, 2007a). \nSatellite cells can contribute to skeletal muscle hypertrophy, in which skeletal muscle fibre diameter increases \nin response to hormonal, endocrine or mechanical stimuli, resulting in increased girth and strength of the muscle \n(reviewed in (Bagley et al., 2023). Hypertrophy commonly results from activities such as resistance training or \nmechanical overload. It results in transient increases of rapamycin complex 1 (mTORC1), which increases muscle \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 3\nprotein synthesis, but increases in satellite cell number and in myonuclear accretion can also be important (Roberts \net al., 2023). Mechanical overload models in mice and rats can mimic this response. One type of overload model \ninvolves surgery to remove the tibialis anterior  (TA) muscle from one leg, which forces the synergistic extensor \ndigitorum longus  (EDL) to undergo sustained stretch, and thereby work harder when the animal is ambulant \n(Egginton et al., 2011). It results in a mild hypertrophy compared to unloaded muscle. This type of muscle \nperturbation is more physiological than inducing acute muscle damage by injecting toxins (Mahdy, 2019). The \nmuscle overload model is also useful in determining how disease states may affect muscle hypertrophy. Previous \nwork using the mdx mouse, which is a model for Duchenne muscular dystrophy (DMD), has shown that \noverloaded EDL muscle is unable to withstand the added strain resulting from the removal of the TA muscle, and \nundergoes accelerated deterioration (Dick and Vrbova, 1993).  \n The exact roles and function of MEGF10 in skeletal muscle are still poorly understood. Here, we have \nexplored the role of MEGF10 in myoblast adhesion and fusion in vitro  using a myoblast clone from the H-2kb-\ntsA58 transgenic mouse (Morgan et al., 1994; Richardson et al., 2022).  We tested the effects of overexpressing \nMEGF10 on fusion and migration. We further tested the effects of MEGF10 by adding expressed and purified \nexogenous MEGF10 domains to cultured myoblasts to determine effects on cell adhesion, migration and fusion. \nWe then explored the role of Megf10 in hypertrophy and satellite cell activation in response to overload in vivo, \nusing a knockout mouse for Megf10. Use of the muscle overload model indicates that in humans, MEGF10 has a \npotential role in satellite cell function that would blunt the overload response, indicating a potential role in the \naetiology of EMARDD. \n  \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 4\n \nResults \n \n \n \n \n \n \nFigure 1: MEGF10 domains, and constructs used in the in vitro experiments. A: The domains of MEGF10, \nshowing positions of known disease mutations (Fujii et al., 2022). B. The three constructs used in the in vitro \nexperiments: MEGF10-eGFP, in which eGFP is fused to the C-terminus of MEGF10; and the ECD and EGF \ndomains, expressed and purified using a mammalian cell system (see methods). \n \nOverexpression of EGFP-MEGF10 reduces fusion and cell motility of cultured myoblasts \nTo explore MEGF10 effect on myoblast differentiation in vitro, we used a single H-2k\nb-tsA58 myoblast clone \n(C1F) derived from satellite cells (Morgan et al., 1994; Richardson et al., 2022).  We generated an adenovirus to \nexpress either eGFP or eGFP-fused to the C-terminus of MEGF10 (Fig. 1B). The resultant constructs were the \nexpected size (Supplemental Fig. 2). Tests for the optimal MOI (multiplicity of infection) demonstrated that an \nMOI of ~100 was optimal (Supplemental Fig. 2).  \nUninfected wild type C1F myoblasts fused into multinucleated myotubes with high efficiency (fusion index of \n~80%, Fig. 2A, B) when cultured under differentiation conditions. Expression of eGFP, using the adenovirus, \nsignificantly reduced fusion to 64% (Fig. 2A, B), suggesting that viral infection alone may reduce fusion.  \nHowever, expression of MEGF10-eGFP reduced the fusion index to very low levels (fusion index of 7.5%, Fig. \n2A, B). Thus, overexpression of MEGF10 inhibits fusion. \nA reduction in myoblast fusion could arise from effects on cell viability or proliferation resulting from \nMEGF10\n/i1 eGFP expression. We found that expression of MEGF10-eGFP significantly reduced cell number at \nday 7 of differentiation (Fig. 2C). This effect is unlikely to be due solely to infection with the adenovirus, as \nexpression of GFP alone using an adenoviral construct did not affect cell number. Thus, expression of MEGF10-\neGFP is likely to reduce cell proliferation or could also affect cell viability. Expression of MEGF10-eGFP also \nsignificantly reduced myoblast motility compared to cells expressing eGFP and uninfected cells (Fig. 2D). This \nreduction in migration may also contribute to the reduction in fusion. \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 5\n \nFigure 2: Effects of overexpression of MEGF10-eGFP on fusion, cell number and migration. A : Images \nof fused cells at day 7 for myoblasts (non-infected cells), myoblasts infected with an eGFP expressing adenovirus \n(MOI 100) and myoblasts infected with MEGF10-eGFP expressing adenovirus (MOI 100). Myotubes were stained \nfor skeletal myosin (red) and DAPI (blue). B: Analysis of fusion using a minimum of 5 fields of view from 3 \nseparate experiments. Fusion index was measured at day 7 (differentiation conditions).  C: total number of cells \npresent at day 7.  D: Cell migration measured after 24 hours of expression. Data shows multiple individual \nmeasurements from three biological replicates. The mean ± the standard deviation (S.D.) is shown together with \nthe results from an ANNOVA. ** P<0.01, **** P<0.001 \n \nThe extracellular domain of MEGF10 reduces fusion of cultured myoblasts \nImaging of C1F myoblasts infected with adenovirus to express MEGF10-eGFP at different MOIs \n(Supplemental Fig. 2B) showed that the higher the MOI, the more MEGF10-eGFP was localised to the Golgi. At \nthe MOI of 100 used in these experiments, MEGF10-eGFP was localised to intracellular vesicles and Golgi and \nplasma membrane, and thus we cannot rule out that intracellular as well as membrane localised MEGF10-EGFP \ncontributes to the effects we observe on cell fusion. To rule out intracellular effects, we carried out additional \nexperiments in which we tested the effects of adding expressed and purified MEGF10 extracellular domain \nconstructs to cultured myoblasts.  We generated two constructs, an extracellular domain comprising all of the EGF \ndomains together with the N-terminal EMI domain (ECD) and a shorter extracellular domain that lacks the N-\nterminal EMI domain and thus comprises the EGF domains only (EGF).  \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 6\nPure MEGF10 extracellular domain (ECD) and EGF domain (lacking EMI domain) (Fig. 1B), was obtained by \nexpression and purification of these constructs in HEK293 cells (Supplemental Fig. 3A, B). The molecular weights \nof these proteins were higher than predicted (Supplemental Fig. 4A, B); the molecular weights of both ECD and \nEGF were found to be approximately 130kDa, whereas the predicted molecular weights are 91kDa and 81kDa, \nrespectively. We confirmed that the expressed and purified protein were the correct MEGF10 constructs by \nwestern blotting for the c-myc tag, which showed a band at the same size as that of the purified protein. Mass \nspectrometry analysis of the purified proteins also confirmed the identity of these two protein domains, providing \n14 unique peptides for EDC (27% coverage) and 31 peptides for EGF (77% coverage).  \nThe increased molecular weights of the expressed and purified EGF and ECD domains are likely to be the \nresult of post-translational modifications such as glycosylation. Mass spectrometry showed that the ECD contained \nO-GlcNac modified residues. Using lectin blots (Supplemental Fig. 3C), we found that the expressed ECD and \nEGF were both highly glycosylated. This both explains the increased molecular weight compared to that expected \nand demonstrates that these domains are likely to be post-translationally modified in a similar way to what would \nbe expected for endogenous MEGF10. \nNext, we used these purified domains to coat non-adherent culture plates, and tested if they could promote cell \nattachment, using gelatin as a control. We found that both ECD and EGF domain constructs increased cell \nadhesion to non-adherent plates, compared to no coating at all (Fig. 3A), suggesting that both domains can \npromote cell adhesion of myoblasts.  \nWe then coated glass surfaces with gelatin, ECD and EGF constructs to test if these domains promoted cell \nadhesion on adherent surfaces and had effects on differentiation as observed for exogenous eGFP-MEGF10 \nexpression. At day 7 of differentiation, the total number of cells per area was significantly reduced for the ECD \nconstruct compared to cells differentiated on uncoated surfaces, and the fusion index was also markedly reduced \n(Fig 3B, D) but there were no significant effects on cell migration (Fig. 3C). Thus, addition of the external ECD \ndomain has somewhat similar effects to expression of eGFP-MEGF10.  However, the EGF construct (which lacks \nthe EMI domain) did not affect number of cells at day 7 of differentiation or fusion, but did reduce cell migration \n(Fig 3B-D).  \nTaken together, these data suggest that MEGF10 expression must be tightly regulated for its correct cellular \nfunction and that the EMI domain is more important in modulating cell adhesion and fusion than the EGF \ndomains. The EMI domain appears to be important in promoting adhesion of cells to a non-adhesive surface, but \nlikely inhibits cell-cell adhesion required for fusion, thus reducing cell number at day 7 of differentiation and \nreducing fusion. \n \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 7\n \nFigure 3: Effects of purified extracellular MEGF10 domains on myoblast fusion attachment and \nmigration. A: Comparison of the ability of no-coat, 0.1% gelatin coating and purified ECD and EGF domains \nfrom MEGF10 to promote C1F myoblast attachment to non-adherent growth surfaces. Attachment was measured \nafter 24 hours. B : Comparison of C1F myoblasts attachment to glass surfaces not coated, or coated with 0.1% \ngelatin, purified ECD or EGF domains after 7 days of differentiation. Individual measurements in A and B \nrepresent the numbers of cells per field of view (minimum of 5) collated from 3 biological replicates.  C: shows \nthe migration (cell speed) of myoblasts on these different coated surfaces 24 hours after plating (individual tracks \nfrom 3 biological replicates) and D: shows fusion on these differently coated glass surfaces at day 7 of \ndifferentiation.  All the data show multiple individual measurements for three biological replicates. The mean ± \nS.D. of the mean is shown together with the results from an ANNOVA. * P<0.05 ** P<0.01, **** P<0.001 E. \nshows example images for myoblast motility/tracking and F shows example images for myoblast fusion on the \ndifferent surfaces at day 7. \n \nEffect of muscle overload on the EDL muscle in wild type, Megf10 +/-  and Megf10 -/-  mice  . \nWe next performed experiments using genotyped wild type, Megf10 +/-  and Megf10 -/-  mice to understand the \npotential in vivo roles of MEGF10. The phenotype of MEGF10 knockout mouse has been reported to be relatively \nmild and exacerbated when crossed to the dystrophin knockout mouse model (mdx) (Saha et al., 2017).  The same \nstudy showed that injection of barium chloride into the tibialis anterior muscle, an approach that results in severe \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 8\nmuscle damage (Hardy et al., 2016), reduced the rate of fibre regeneration in MEGF10 -/- mice. Here, we focussed \non the response of skeletal muscle to stimuli that elicit hypertrophy, a milder treatment. In this case, the TA \nmuscle is removed, to induce overload on the EDL (extensor digitorum longus) muscle, which hypertrophies as a \nresponse. This can occur both as an increase in protein synthesis and though stimulation of satellite cells (Murach \net al., 2021).  In all of these experiments, the tibialis anterior muscle is only removed from one leg, resulting in \noverload (OL) on the EDL muscle in this leg. In the contralateral (CL) muscle, the TA muscle is kept intact, acting \nas a control. \n \n \nFigure 4. Variation in EDL hypertrophy and number of myonuclei following overload as a function of \ntime in wild type mice.  A. The relative weight (EDL muscle weight as a proportion of the total mouse body \nweight) is shown (mean values +/- S.D.) at each time point for unloaded contralateral (CL) and overloaded (OL) \nEDL (n=30 for D0, D6 and D10; n=6 for D14). B. Hypertrophy expressed as the % change in weight between \nunloaded CL and OL EDL. C. The numbers of myonuclei per 1mm of fibre from overloaded and contralateral \nEDL muscle at different days after overload (N=15 fibres, from 3 biological replicates).  All the data was analysed \nusing a 2way ANOVA (Sidák’s multiple comparisons test) **** P<0.0001. *** P  <0.001. ** P <0.01. D. \nRepresentative mages of overloaded muscle fibres after different days of overload. Nuclei stained with DAPI.  \n \nTest experiments revealed that the overload response peaked at 10 days after surgery (Fig. 4: D10). The \nrelative mass of the overloaded EDL muscles increased at each of the three time points sampled: D6, 10 and 14 \ncompared to D0 (Fig. 4A). In contrast, the relative mass of the unloaded contralateral EDL muscles did not \nchange, with no significant difference between contralateral tissue at all time points compared to D0 EDL (Fig. \n4A). Expressing the change in weight as % hypertrophy comparing the overloaded muscle (OL) to the \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 9\ncontralateral muscle (CL) (Fig. 4B) shows that the overloaded muscles hypertrophied by ~40% at D6-14, a \nsignificant increase compared to D0. Muscle hypertrophy appeared to plateau between days 10 and day 14.  \nCounting the numbers of nuclei per fields of view for isolated muscle fibres showed that the numbers of nuclei \nper mm was significantly increased in the overloaded compared to contralateral muscle fibres at D10 and D14. \nThe percentage increase in nuclei per mm in overloaded fibres was significantly increased compared to D0 (Fig. \n4C, D). No significant differences were found when number of nuclei on contralateral fibres were compared \nbetween time points, or to D0 fibres. The increase in numbers of nuclei shows a similar trend to the increase in \nEDL mass following overload. As the largest effects of overload were observed at D10, this time point was used in \nsubsequent experiments. \n \nThe hypertrophy response to overload is reduced in Megf10\n-/- mice \n \nFigure 5. Megf10+/- and Megf10-/- mice show reduced EDL hypertrophy following muscle overload: day \n10.  A. The relative weights for unloaded (contralateral) and overloaded EDL muscles for wild type, heterozygous \nand homozygous knockout Megf10 mice. B: Relative hypertrophy (%) of the overloaded EDL muscle following \n10 days of overload. C. numbers of myonuclei per 1mm of fibre from overloaded and unloaded (contralateral) \nEDL muscles at day 10. (n=15 fibres, minimum of 3 biological replicates) D. Mean fibre cross sectional area of \nmuscle fibres (per mouse). Data was analysed by ANOVA. Error bars represent S.D. ** P <0.01, **** P<0.001, \n**** P <0.0001. ns: non-significant. \n \nAfter 10 days, the weight of the overloaded EDL muscle from wild type and Megf10+/- mice increased \nsignificantly compared to the contralateral, unloaded muscles but was not significantly different for homozygote \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 10\nMegf10-/- mice (Fig. 5A).  However, the percentage change in hypertrophy was lower for overloaded EDL muscles \nfrom Megf10+/- and Megf10-/- mice compared to wild type (Fig. 5B).  Similarly, there was a significant increase in \nmyonuclear number (myonuclear accretion) in overloaded EDL muscles from all three genotypes, which was \nsimilar in magnitude for wild type and MEGF10+/- mice. However, myonuclear accretion was significantly \nreduced in overloaded EDL muscles from Megf10-/- mice (Fig. 5C). These results suggest the complete loss of \nMEGF10 reduces the hypertrophic response to muscle overload. The overall fibre cross sectional area (CSA) per \nmouse did not significantly change, although there was a trend for fibre CSA to increase for wild type and \nMEGF10+/- mice, but not for the Megf10-/- mouse. \n  \nLower numbers of Pax7 + cell nuclei are present at D0 and increases in transcription factor expression \ninduced by overload are reduced in in Megf10+/- and Megf10-/- mice \n \nFigure 6.  Percentage Transcription Factor (TF) expression by satellite cells on fibres from wild type, \nMegf10+/- and Megf10 -/- mice. A: Expression of each TF (Pax7, MyoD and Myogenin) at day 0 for wild type, \nMEGF10+/- and MEGF10 -/- mice.  B-D shows the expression of each TF for contralateral, unloaded and \noverloaded EDL muscles at day 10. Values shown in A-D are the mean values per mouse (using the EDL muscle) \ntogether with the S.D.  Data was analysed by ANOVA.  **** P <0.0001. * P <0.05. ns: non-significant. \n \nPax7 W\nT\nPa\nx7 ME\nGF10 \n+/-\nPax7 M\nEGF10 -/-\nMyoD W\nT\nMyoD ME\nGF10 +\n/-\nMyoD \nMEG\nF10 -/-\nMyogenin WT\nMyogenin ME\nGF10 \n+/-\nMyog\nenin M\nEGF10 \n-/-\n-10\n0\n10\n20\n30% of TF positive nuclei\nDay 0 TF positive nuclei\n W\nT CL\n W\nT O\nL\n MEGF\n10 +/- CL\n M\nEGF\n10 -/- O\nL\nMEGF\n10 -/-\n CL\nMEGF\n10 -/- O\nL\n0\n10\n20\n30\n40\n50\nMyoD day 10\n% of TF positive nuclei\nns\nns\nWT CL\n WT OL\nMEGF10 +/-  CL\n MEGF1\n0 +/-  OL\nMEGF10 \n-/- CL\nMEGF10 -/- \n OL\n0\n10\n20\n30\n40\n50\nDay 10 Pax7  \n% of TF positive nuclei\nns ns\nWT CLWT\n OL\nMEGF10\n +/- C\nL\nMEGF10 -/- OL\nMEGF\n10 -/-\n CL\nMEGF\n10 -/-\n OL\n0\n10\n20\n30\n40\n50\nMyogenin day 10\n% of TF positive nuclei\nns ns\nns\nAB\nCD\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 11\nLoss of MEGF10 expression in humans has been linked to decreased numbers of Pax7+ satellite cells (Logan et \nal., 2011). Decreased numbers of Pax7 + satellite cells could underlie the reduced hypertrophic response observed \nfor Megf10-/- EDL muscle. To evaluate this, fibres were isolated from mice and stained for the transcription factors \nPax7, MyoD and myogenin, at D0 and after 10 days of overload.  At day 0, the numbers of Pax7+ positive satellite \ncells were lower in heterozygote MEGF10+/- and homozygous knockout MEGF10-/- mice compared to wild type \n(Fig. 6A). A similar, but non-significant trend was found for MyoD and myogenin (Fig. 6A). This suggests that \nthe numbers of satellite cells are reduced in EDL from MEGF10+/- and MEGF10-/- mice at D0.  \nAfter 10 days of overload, the percentage of nuclei that stained positive for Pax7, MyoD and myogenin \nincreased significantly for overloaded EDL muscle from wild type mice compared to the contralateral (CL) muscle \n(Fig. 6B-D).  However, the percentage of nuclei positive for Pax7 and myogenin did not increase significantly in \noverloaded EDL muscle from MEGF10+/- and MEGF10 -/- mice compared to the unloaded, contralateral muscle \n(Fig. 6B, D). The number of MyoD positive nuclei for the overloaded EDL muscle did increase significantly for \nMegf10+/- mice but not for Megf10 -/- mice (Fig. 6C). In summary, overload of the EDL muscle in Megf10+/- and \nMegf10-/- mice reduces the increase in transcription factor expression compared to wild type mice. \n \nMean fibre cross-sectional area in the diaphragm from Megf10+/- mice is smaller \n \n \nFigure 7. Comparison of wild type and Megf10 +/- diaphragm. A. Representative images of stained \ndiaphragm cross-sections showing differences in fibre structure. B. Scatter plot showing individual measurements \nof CSA across 3 biological replicates. C. Scatter plot showing thickness of laminin divisions between fibres, \nmeasured from 5 fibres per mouse. Bars show mean ± S.D. Error bars represent S.D. **** P <0.0001. \n*** P <0.001. * P <0.05.   \n \nAs we did not obtain the expected Mendelian ratio for the genotypes of pups at birth, and to further identify \neffects of the loss of Megf10 on muscle, we analysed the diaphragm muscle from 6-week old wild type and \nMegf10+/- mice. The muscle was stained for laminin to outline the muscle fibres and show if muscle fibre size and \norganisation is affected. This revealed strong differences between wild type and Megf10+/- mice (Fig. 7A). The \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 12\nmean fibre CSA of individual fibres from the diaphragm muscles were reduced significantly in Megf10+/- \ncompared to wild type mice (Fig. 7B). The thickness of the laminin in the extracellular matrix between fibres was \nincreased significantly in the Megf10+/- diaphragm compared to wild type (Fig. 7D). This may explain why the \nMendelian ratio was not as expected, as it is possible that the diaphragm of Megf10-/- mice is more strongly \naffected, and mice do not survive long after birth.  \n \nDiscussion \nHere, using a combination of in vitro and in vivo  experiments, we have shown that overexpression of GFP-\nMEGF10 or addition of the extracellular domain of MEGF10 (ECD) inhibits fusion of C1F myoblasts. The \npurified full-length ECD was more effective in attachment to a non-adhesive surface than the shorter EGF domain, \nin which the N-terminal EMI domain was absent. However, it was less effective at maintaining cell-cell \nattachment as cell differentiate into myotubes, reducing both number of cells and fusion at day 7 of differentiation, \nsuggesting a key role for the EMI domain in these processes.  In vivo, loss of Megf10 reduced hypertrophy of the \nEDL muscle in response to overload, as measured by weight change and myonuclear accretion, with the most \nmarked effect for homozygous knockout Megf10\n-/- mice. This reduced response may be accounted for by the \nreduction in Pax7 + satellite cells, and decreased satellite cell response to overload in Megf10 heterozygous \n(Megf10+/-) and homozygous knockout mice. Finally, homozygous knockout mice were born at a lower mendelian \nratio than expected. An analysis of the diaphragm muscle in wild type and heterozygous mice showed more \nvariable and reduced fibre cross-sectional area in the heterozygous mice compared to wild type, and increased \nlaminin deposition between fibres. The lower survival rate in homozygous mice, could arise from increased \ndefects in the diaphragm muscle, which reiterates the main clinical presentation of affected individuals with \nMEGF10-related EMARDD; respiratory distress due to diaphragmatic paralysis. \n During myoblast differentiation into myotubes, previous work demonstrated that that Megf10 is \ndownregulated during differentiation of C2C12 cells, by qPCR (Holterman et al., 2007) and in C1F myoblasts \ndifferentiated on glass surfaces, by RNAseq analysis (Richardson et al., 2022). These changes have not been \nconfirmed at the protein level, as a well-validated antibody to Megf10/MEGF10 is lacking. However, if Megf10 is \nnormally downregulated during myoblast differentiation, then overexpression of MEGF10 in vitro , either via  \nexpression of eGFP-MEGF10 or via addition of the purified ECD of MEGF10 might be expected to interfere with \nfusion and differentiation, as we found here. Our findings for eGFP-MEGF10 are similar to those observed \npreviously for overexpression of HA-tagged Megf10 in C2C12 myoblasts, which also decreased myoblast fusion \nand differentiation in vitro (Holterman et al., 2007).  \nAddition of purified ECD domain also inhibited fusion of myoblasts in vitro and was more effective at \ninhibiting fusion than the EGF domain, which lacks the EMI domain. However, the ECD domain was more \neffective at promoting cell attachment to non-adhesive surfaces than the EGF domain. Thus, the EMI domain \nseems to be important in both inhibiting fusion (which requires cell-cell adhesion) and in promoting cell-surface \nadhesion. The EMI domain is found in many other MEGF isoforms, and also in the worm orthologue of human \nMEGF10 (Callebaut et al., 2003). If the EMI domain is able to recognise and bind to PS, then its presence in \nfusing cells that expose PS on their surface could potentially block binding of other PS receptors that promote \nfusion (Chikazawa et al., 2020; Hochreiter-Hufford et al., 2013; Jeong and Conboy, 2011; van den Eijnde et al., \n2001). This suggests a possible role of MEGF10 in preventing premature myoblast fusion.  Of note, MEGF10 has \nalso been suggested to interact with Notch via its intracellular domain (reviewed in (Vargas-Franco et al., 2022)), \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 13\nand thus the effects of the exogenous ECD and EGF domains would be independent of any downstream signalling \npathway involving Notch. \nOur observation of a hypertrophic response of the EDL muscle to overload for wild type mice is consistent \nwith previous reports (Egginton et al., 2011; Johnson and Klueber, 1991). The increase in mass of the overloaded \nEDL as well as the increase in myonuclei has been reported previously (Bruusgaard et al., 2010; Huey et al., 2016; \nSeiden, 1976; Zhou et al., 1998). Moreover, the activation of satellite cells by overload, with increases in Pax7\n+, \nMyoD+ and myogenin+ satellite cells of wild type mice matches findings from previous reports (Hyatt et al., 2008; \nSakuma et al., 1999). However, we did not find a strong effect of overload on fibre cross-sectional area, although \nthis has been observed previously in other overload conditions (Carter et al., 1995; Rosenblatt et al., 1994; \nSnijders et al., 2020). Overall, our experimental data shows that EDL from wild type mice responds to \nphysiological overload as expected, and that this is a suitable approach to explore the effects of overload in \nMegf10\n+/- and Megf10-/- mice.  \nThe reduction in or loss of Megf10 reduced the hypertrophic response of the EDL muscle to overload for both \nheterozygous Megf10+/- and homozygous knockout Megf10 -/- mice, compared to wild type mice. This reduction \nwas greatest for the homozygous mice. Moreover, the increase in Pax7, MyoD and myogenin positive cells was \nreduced, again with the largest effect in homozygous mice. The contribution of satellite cells to hypertrophy has \nbeen the subject of much discussion (Pallafacchina et al., 2013) and it has been shown that some increase in fibre \nsize is possible without functional satellite cells (Murach et al., 2017). However, Pax7 expression has been \nreported to contribute to the activation and subsequent expansion of satellite cells in response to stimuli such as \noverload (Wang and Rudnicki, 2011). Moreover, our findings agree with a previous report that found Megf10 \ndeficient mice to have reduced expression of Pax7 and MyoD, resulting in inadequate regeneration of EDL muscle \nfollowing acute injury due to barium chloride treatment (Li et al., 2021).  \nIn addition to the reduced hypertrophic response, we also found that muscle fibres in the diaphragm were \nhighly variable in size, and that there was increased fibrosis. Loss of MEGF10 in humans causes respiratory \ndistress (Logan et al., 2011). The change to the structure in the diaphragm of Megf10\n+/- mice we observed could \nalso lead to respiratory distress and could help to account for the lower numbers of homozygous mice that we \nobtained. Further work is needed to confirm if the diaphragm of Megf10\n-/- mice is affected more severely. \nInterestingly, a recent report described defects in the neuromuscular junction (NMJ) in the tibialis anterior and \ndiaphragm muscles in Megf10\n+/- mice, consistent with its role in glial cells in synapse remodelling (Juros et al., \n2024), suggesting altered NMJs may also contribute to the phenotype of the diaphragm muscle seen here. \nHowever, the NMJ of the EDL muscle was not affected in Megf10 -/- mice (Juros et al., 2024), suggesting that \nalterations to the NMJ do not have a major contribution to the reduced hypertrophic response we observed here. \nOverall, our results demonstrate that MEGF10 is likely to be important in myoblast-surface adhesion but can \npotentially block fusion of myoblasts when present at high levels, thus possibly preventing premature fusion of \nsatellite cells. The reduction in numbers of Pax7\n+ satellite cells, and the reduction in their activation in response to \nmuscle overload in Megf10-/- mice broadly supports this idea. Satellite cells could be lost through poor adhesion \nand/or premature fusion. The overall phenotype of the Megf10-/- mice recapitulates the human phenotype observed \nin EMARDD (Logan et al., 2011), resulting from homozygous nonsense mutations in the MEGF10 gene. In \naddition to respiratory distress resulting from paralysis of the diaphragm muscle, the reduced numbers of Pax7 + \nsatellite cells,  reduced skeletal muscle fibre growth and reduced satellite cell activation in the mouse model, are \nreminiscent of the reduced number of Pax7 + satellite cells, small muscle fibres, fibre necrosis and fibre \nreplacement by fibrous or adipose tissue in humans (Logan et al., 2011).   \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 14\n \nMaterials and Methods \nIn vitro experiments: cell culture and staining \nTo explore muscle differentiation in vitro, we used a clone of satellite cells (C1F) derived from the H-2k b-\ntsA58 mouse (Morgan et al., 1994; Richardson et al., 2022). These cells proliferate at 33°C in the presence of \nIFN/i1  in growth medium (Dulbecco’s minimal essential medium (DMEM), high glucose, containing Glutamax \n(Gibco), supplemented with 20% FCS, 1% Penicillin/Streptomycin (v/v) (Gibco), 2% Chick Embryo Extract \n(E.G.G. Technologies)) and are switched to differentiate by changing the medium to DMEM supplemented with \n4% Horse Serum and 1% penicillin/streptomycin, and increasing the temperature to 37°C. Proliferation ceases \nwithin 24 hours, and cells start to fuse at this time point. RNASeq experiments suggested that Megf10 is expressed \nin these cells (Richardson et al., 2022). \nFor immunostaining, cells were plated onto washed 13 mm diameter glass coverslips, coated with 0.1% \ngelatin. Cells were fixed with 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS) for 20 minutes, \nwashed in PBS, permeabilised with 0.5% Triton X-100 diluted in PBS containing 1% (w/v) bovine serum albumin \n(BSA) for 30 minutes prior to incubating with primary antibodies (Supplementary Table 1) diluted into PBS with \n1% (w/v) BSA for 60 minutes at room temperature. Coverslips were washed x5 with PBS, and then appropriate \nanti-mouse or anti-rabbit secondary antibodies (Alex-Fluor conjugated, Invitrogen) diluted 1/400 in PBS with 1% \n(w/v) BSA were added. Fluorescent phalloidin was used to stain filamentous actin and DAPI (4\n′ ,6-diamidino-2-\nphenylindole) was used to stain nuclei (Sigma). Of note, we were unable to validate any commercial antibody to \nMEGF10, and thus could not stain or blot for endogenous protein.  Once stained, coverslips were mounted onto \nglass microscope slides using Prolong gold antifade mountant (Invitrogen).  Cells were imaged using a Zeiss LSM \nAiryscan confocal microscope, using a x40 objective lens (NA 1.4) or an Olympus widefield microscope using a \nx63 objective lens (N.A. 1.4) followed by deconvolution. The resulting images were assembled into figures using \nAdobe photoshop.  \n \nGFP-MEGF10 adenoviral construct generation \nAdenoviral expression constructs for GFP-MEGF10 (full length human MEGF10, 3423 bp was provided by \nColin Johnson) and for GFP were generated by PCR-based cloning into a pDC315 vector (Addgene). For GFP-\nMEGF10, eGFP was placed at the C-terminus of MEGF10, separated from MEGF10 by a short 3x glycine linker, \nand a 6xHis tag was incorporated at the C-terminus after the eGFP (Fig. 1B). Fully sequenced constructs, with no \nsequence errors, were used to generate adenovirus using Ad293 cells (Microbix). Purified DNA plasmids \n(pDC315 alongside pBHGlox\nΔ E1,3Cre) were transfected into Ad293 cells, using Fugene and the resulting \nadenovirus was amplified as described in the manufacturer’s protocol (Wolny et al., 2013). After the final round of \namplification, virus was purified using the Vivapure Adenopack 100 kit (Sartorius Stedim Biotech), purified virus \nwas stored in storage buffer (20 mM Tris/HCl, 25 mM NaCl, 2.5% glycerol (w/v), pH8) at -80°C. Viral titre was \ndetermined by tissue culture infectious dose 50 (TCID50) method. Final titres for GFP-MEGF10 and the GFP \nvirus were 3 x 10\n8 and 6 x 10 8 PFU per ml respectively. This was used to estimate the MOI of infection, when \ninfecting cultured myoblasts. Western blots confirmed the expected sizes for GFP-MEGF10 and GFP \n(Supplemental Fig. 1). \n \nExpression and purification of extracellular domain constructs of MEGF10. \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 15\nExtracellular MEGF10 constructs (human) were generated for mammalian protein expression (Fig. 1B). The \ncDNA for the extracellular domain (ECD: residues Leu35 to Gly861) or the ECD lacking the EMI domain (ECF, \nHis 108 to Gly861) were cloned into the pSecTag2A plasmid (Invitrogen), sequences confirmed, and constructs \nwere transfected into HEK-293 cells using calcium chloride transfections. pSecTag2a contains an N-terminal \nsecretion signal from Ig- κ  for efficient protein secretion in the media, a cytomegalovirus (CMV) promoter for \nhigh-level expression and C-terminal 6xHis and c-myc tags for nickel column purification and antibody detection, \nrespectively, as well as a Zeocin resistance gene for the selection of stably expressing mammalian cell lines. 48 \nhours after transfection, cells were harvested, diluted to 1 x 10 5 cells per ml, and seeded at 2 x 10 4 cell per ml in \nselection medium (DMEM, 10% FCS, 1 % FCS, 200 μ g/ml Zeocin). After 10 days, individual clones, were picked \ninto 24 well plates and allowed to grow. Samples of media from each colony were analysed by dot blot, to isolate \nclones for which expression was highest.  \nFor expression and purification of the MEGF10 extracellular domain constructs, stable cell lines with high \nlevels of expression, were seeded into five 75 cm 3 flasks coated with 20 µg ml -1 poly-L-lysine, grown to 80% \nconfluence in normal growth medium and then the medium was exchanged for OptiMEM low serum medium \n(GIBCO) to reduce contaminants during purification, and cells cultured for a further 3 days. The medium was then \nremoved, centrifuged at 1000 x g rcf, and the supernatant incubated with 1 ml Complete His-Tag Purification \nResin slurry (Roche) and Complete EDTA-free protease inhibitor cocktail tablet (Roche) for 30 min on a roller.  \nThe mixture was then applied to a 5 ml column, and the flow through collected. The resin was washed 5x with \ncolumn wash buffer (300 mM NaCl, 50 mM NaHPO\n4) and eluted with elution buffer (300 mM NaCl, 50 mM \nNaHPO4, 200 mM Imidazole). Eluted protein (in 2 mL) was dialysed into PBS overnight using a Gebaflex Maxi \nDialysis Tube (MWCO = 3.5kDa) (Generon). Purified protein was stored in 200 µL aliquots at -80 °C. Protein \nconcentration was measured using a Cary 50 Bio-UV visible spectrophotometer (Varian) at a wavelength of 280 \nnm. Typical concentrations were 60-100 ng/µl for ECD and 150-200 ng/µl for ECF. Protein identity was \nconfirmed by mass spectrometry. A lectin blot (Biotinylated Lectin Kit I (Vector Laboratories)) was used to \nconfirm that the ECD and EGF constructs had been glycosylated. \n \nAnalysis of protein expression \nSamples of cells used for western blotting were either prepared by scraping cells directly into 2x Laemlli buffer \nprior to boiling at 100 °C for 10 min and freezing aliquots at -80 °C, or freshly pelleted cells were resuspended \ninto 50 µl ice-cold  lysis buffer (150 mM NaCl, 50 mM Tris (pH8), 1% Triton X-100, 1 mM EDTA (pH8)) \ncontaining Halt Protease Inhibitor, single-use cocktail (ThermoScientific), incubated on ice for 30 mins with \nregular vortexing, centrifuged at 17000 x g rcf for 20 mins at 4 °C, and lysates stored at -20 °C. Protein \nconcentration was then quantified by a Pierce micro BCA Protein Assay kit (Thermo Scientific) following the \nmanufacturer’s instructions. Absorbance at a wavelength of 544nm was measured using a Polstar Optima plate \nreader. \n \nAnalysis of myoblast fusion and cell motility \nTo analyse cell motility, cells were imaged for 14 hours, capturing images every 10 minutes, using differential \ninterference contrast (DIC) microscopy (Olympus widefield microscope), x10 objective lens at 37 °C. For eGFP-\nMEGF10 expressing cells, individual wells of a 96 well plate with a borosilicate glass bottom (Iwaki) were seeded \nwith 50 \nμ l of C1F cells at 1x105 cells/ml and infected overnight using an MOI of 100 for each adenoviral construct \nin 500 μ l culture media.  For each condition, 5 fields of view were imaged and the experiment was repeated three \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 16\ntimes (3 biological replicates).  Cell motility was analysed using ImageJ software (MTrackJ plugin), tracking 10 \ncells per field of view.  To determine the effect of the purified ECD and ECF domains on cell motility, individual \nwells of the borosilicate glass were coated with 1.25 μ g of purified protein for 20 min at 37°C. Excess coating was \naspirated and wells seeded with 50 μ l C1F cells at 1x105 cells/ml. Cells were incubated with 50 μ l of culture media \nat 33 °C, 10% CO 2 for 24 hrs, medium made up to 500 μ l and then cells were filmed and motility analysed as \nabove. \nTo estimate the fusion index, C1F cells were seeded onto pre-washed 13 mm diameter coverslips, coated with \n0.1% gelatin and differentiated for seven days. Cells were fixed with pre-warmed 4% PFA for 20 minutes, washed \nwith PBS and then permeabilised with 0.1% Triton X-100 in PBS.  The nuclei were stained for nuclei using DAPI, \nfilamentous actin using fluorescently labelled phalloidin and striated muscle myosin using the A4.1025 antibody \n(Cho et al., 1994; Maggs et al., 2000). The fusion index is calculated from the percentage of nuclei found in \nskeletal myosin positive myotubes as a percentage of the total number of nuclei.  Only skeletal myosin positive \nmyotubes with three or more nuclei were classed as myotubes. For each condition, five fields of view were imaged \nand cells counted from three biological replicates. \n \nCell attachment and cell motility assays \nTo determine the ability of the ECD and ECF proteins to mediate C1F myoblast attachment to a surface a cell \nattachment assay was performed. Briefly, the wells of a non-adhesive 24 well plate (Greiner), were coated with 2.5 \nμ g protein diluted in 150 μ l PBS for 20 min at 37 °C. Harvested myoblast cells (C1F clone) were prepared at \n1x104 cells ml-1 in growth medium and 100 μ l added to each well. Cells were incubated at 33 °C, 10% CO 2 for 30 \nmin before adding 500 μ l of fresh medium. After 24 hr incubation, cells were imaged using a Cytomate \ninstrument, taking 5 different fields of view and counting the number of nuclei from each field. The experiment \nwas repeated three times. Alternatively, a 96 well borosilicate glass plate (Iwaki) was coated for 20 minutes at 37 \n°C with 0.1% gelatin, or with 1.25 µg of the EGF or the ECD domains of MEGF10, or left uncoated, coating was \naspirated and wells seeded with 50 µl of CIF cells at 1 x 10\n5 cells per ml. Cells were allowed to attach, \nsupplemented with additional medium (50 µl), and filmed overnight to analyse their motility. \n \nMegf10 knockout mice \nMegf10tm1(KOMP)Vlcg mice (RRID: MMRRC_048576-UCD, MGI ID: 4454190, background: C57BL/6Tac, \nintragenic targeted knockout deletion, gene ID: 70417) were obtained from the Mary Lyon Centre at the MRC \nHarwell Institute and exported to our laboratory to establish breeding colonies. Briefly, the model was generated \nby replacing exons 1-24 of mouse Megf10 by homologous recombination with an expression selection cassette as \ndetailed by the Knockout Mouse Project (KOMP, University of California Davis, Davis, CA: \nhttps://www.komp.org/geneinfo.php?geneid=68051\n). The Megf10tm1(KOMP)Vlcg mice harbour the Velocigene \ncassette ZEN-Ub1 inserted into the Megf10 gene between positions 57340143 and 57372060 on chromosome 18, \ngenerating a 31918bp deletion that deletes exons 1-24 of MEGF10. The mouse line was rederived in the Harwell \nfacility after its original generation at Regeneron Pharmaceuticals. qPCR has shown no mRNA for MEGF10 is \nexpressed in the cerebellum in homozygote animals (Mouse Genomics Informatics) (Iram et al., 2016) and \nknockouts, hereafter designated Megf10\n-/-, lack MEGF10 protein (Fig. 1f in (Chung et al., 2013)), confirming that \nthe tm1 allele is null. Megf10-/- (homozygous knockout) mice were generated on a C57BL/6NTac background.  \nMegf10-/-, Megf10+/- (heterozygous), and wild type animals used in the experiments were generated by crossing \nMegf10+/- heterozygotes, and all comparisons between genotypes are between age-matched littermates. Male and \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 17\nfemale adult C57BL/6Tac (bred in-house at the University of Leeds) and Megf10tm1(KOMP)Vlcg mice (final body \nmass approx. 25g) were used in this study.  \nAll experimental procedures and sacrifice were conducted with approval of the local animal welfare and ethics \ncommittee, under Home Office project licences 70/8674 and PP1775021. Mice were housed in groups in a \ntemperature-controlled environment with access to food and water ad libitum . Cages underwent 12/12 light/dark \ncycles. Breeding was carried out under service licence PP0237211, with two breeding cages of C57BL/6Tac x \nheterozygous Megf10tm1(KOMP)Vlcg, and one breeding cage of heterozygous Megf10tm1(KOMP)Vlcg x heterozygous \nMegf10tm1(KOMP)Vlcg. Mice were stunned by concussion and killed by cervical dislocation. \n \n \n \nGenotyping \nTo accurately determine the genotype for each mouse, ear biopsies were collected by unit staff at Central \nBiomedical Services (CBS), University of Leeds. Biopsies were placed into 96-well plates, sealed, and shipped to \nTransnetyx for genotyping (Transnetyx Inc. Cordova, TN) via courier. A bespoke PCR assay to determine \ngenotype was designed by the Genetic Services team at Transnetyx based on information provided by KOMP \n(Knockout Mouse Project) mouse repository (Supplementary Fig. 1). Results were obtained within 72 hours of \nsending the samples.  Of note, we did not obtain the expected Mendelian ratio of 1:2:1 (Supplemental Table 2). \n \nHypertrophy model surgery \nUnilateral extirpation (removal) of tibialis anterior (TA) muscle was performed under aseptic conditions and \ninhalation anaesthesia. All instruments were sterilised and work carried out under a dissection microscope. Mice \nwere first anaesthetised with 5% isoflurane in 2Lmin\n-1 O2. The left leg was then shaved and wiped with ethanol to \nsterilize the area and remove surface bacteria. For the remainder of the operation, mice were maintained under \nanaesthetic at 2% isoflurane in 2L/min O 2.  All possible steps were taken to avoid animals suffering at each stage \nof the experiment. \n A single incision was made on the hindleg to expose the TA, tweezers used to lift the superficial distal \ntendon, and the TA removed by making incisions at proximal and distal points of attachment using a scalpel. The \nTA cut end was then held over the wound area for approx. 10s to allow the release of chemokines to aid repair and \nblood clotting. The TA was then discarded and 1-2 drops of 2.5% Baytril (Bayer AG) was applied to the wound \nfor antiseptic protection. The area was swabbed with a cotton bud to remove blood and then the incision was \nsutured with MERSILK\nTM (Ethicon Inc.) braided silk suture, size 5.0. Sutures were intermittent and double-\nknotted to reduce the chance of mice unravelling them post-operatively. 1-2 more drops of Baytril /i1 were applied \nto the closed wound and swabbed with sterile cotton buds, to remove dried blood that may lead to irritation. 0.1ml \n10% Vetergesic (Ceva Animal Health, Ltd) was administered to the scruff of the neck to provide post-operative \nanalgesia. Mice were placed in a heated cage without sawdust for approximately 10 minutes to recover from \nanaesthetic, before being placed back into their original cage. Mice were observed to be normally ambulant, thus \noverloading the extensor digitorum longus (EDL), for a pre-determined length of time before sampling. \n \nEDL isolation  \nChanges in EDL muscle phenotype were assessed in control animals (no overload), as well as animals \noverloaded for 10 days, to observe changes in the muscle.  This interval was chosen as we observed that the \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 18\noverload response had peaked at this time point in test experiments. Mice were killed by Schedule 1 (concussion \nfollowed by cervical dislocation). Muscle was removed as quickly as possible to minimise post-mortem \nbiochemical changes. The leg of a freshly killed mouse was shaven and dabbed with ethanol to promote cutaneous \nvasoconstriction. A small incision from just lateral to the knee to the beginning of the hindfoot aided blunt \ndissection using scissors and forceps to break through the layer of fascia atop the muscle. On the unoperated \n(contralateral) leg, the TA was first removed to access the EDL underneath. The EDL was then accessed and \nremoved in the same way (forceps used to hold the tendon and scalpel used to release it at the base). On the \nipsilateral (overloaded) leg, the EDL was simply removed as described. The EDL, and the whole mouse, were both \nweighed to determine the EDL mass as a percentage of total body mass.   \n \n \nPreparation of muscle samples for imaging. \nSamples of skeletal muscle were additionally harvested for single fibre isolation (as described above), or for \ncryo-sectioning. For accretion measurements, single fibres were fixed and permeabilised, incubated with DAPI \n(1/500) for 90 minutes at room temperature before washing with TBST and finally with PBS. Samples were \nmounted on cleaned glass microscope slides using ProLong Gold and covered by a 20 x 40 mm glass coverslip.  \nFor cryosectioning, intact muscles were trimmed, mounted onto a cork disk with optimum cutting temperature \ncompound (OCT) (Agar Scientific) and immediately snap frozen in isopentane-liquid nitrogen and stored at a \ntemperature of -80 °C. Diaphragm muscle was also prepared for cryo-sectioning, using a similar approach. \nSamples were sectioned (30 µm and 10 µm sections) using a cryostat (Leica) pre-cooled to -20 °C and sections \nplaced onto labelled glass slides and stored at -20°C until ready to fix and stain. \n \nSlides were left at room temperature for ~10 mins to dissipate condensation. A hydrophobic barrier pen was \nthen used to draw around each segment of 3-4 sections joined together. Within the confines of the hydrophobic \nbarriers, tissue was fixed by applying 100-200 µl ice-cold 100% methanol and incubating at room temperature for \n10 mins. Sections were then washed three times with PBS. Non-specific antibody binding was reduced by \nincubating with 5% BSA diluted in PBS at room temperature for 30 min. Primary antibodies were diluted in PBS \nand applied to tissue sections following removal of the blocking solution. Slides were incubated with the primary \nantibody overnight at 4°C. Primary antibody was removed and sections washed three times with wash buffer (1% \nBSA in PBS). Secondary antibodies were diluted in PBS, applied to tissue sections, and incubated for 1hr at room \ntemperature. Secondary antibody was removed, and sections washed three times with wash buffer, before a final \nwash with PBS. Two drops of ProLong Gold antifade mountant was then added to the slide, and a 20 mm x 40 mm \nglass coverslip was placed on top. Slides were left overnight at room temperature in the dark overnight, and then \nstored at 4 °C.  \n \nQuantification of transcription factor expression and fibre cross-sectional area \nStained fibres were imaged using a widefield Olympus IX-70 microscope, using a 40x, N.A. 1.4 objective lens.  \nFor measuring myonuclear accretion, 15 x DAPI stained fibres were imaged per condition (three fibres per EDL \nsample). Numbers of nuclei were counted from 1 mm sections using ImageJ (Fiji).  \nFor transcription factor staining, the number of nuclei positively stained for a transcription factor (Pax7, MyoD \nor myogenin) per 50 myonuclei along the fibre was measured from three fibres per EDL sample, stained for DAPI \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 19\nand the transcription factor. The numbers of positive nuclei per fibre were expressed as a percentage of 50 \nmyonuclei, and the number per fibre was averaged.  \nCross-sectional area of individual muscle fibres was also determined using ImageJ (FIJI). Outlines of each \nindividual fibre were drawn around on each image using the freehand selection tool, and area in µm 2 was \nautomatically calculated.  \n \nMeasurement of laminin thickness \nThe width of the basement membrane between fibres in stained diaphragm cross sections was measured using \nImageJ processing software (NIH). The ‘straight line’ tool was used to take orthogonal measurements of laminin \n(visualised with anti-rabbit IgG Alexa Fluor 488 conjugate) normal to the sarcolemma, and this value recorded. \nFive measurements were taken per fibre, and five fibres were measured for 3 wild-type and 3 Megf10\n+/- mice.  \n \nStatistical analyses \nStatistical tests were performed, and graphs generated, using GraphPad Prism for Mac (GraphPad Software, La \nJolla California, USA, www.graphpad.com). Graphs show mean ± standard deviation (S.D.) for each observation. \nUnpaired t-tests with Welch’s correction and one-way and two-way ANOVAs were carried out to test for any \nstatistically significant differences between conditions. The level of statistical significance is indicated by the \nnumber of asterisks displayed above graphs: **** represents a P value <0.0001. *** represents a P value <0.001. \n** represents a P value <0.01. * represents a P value <0.05. \n \n \nSupplemental Material \n \nThe supplemental material comprises two supplementary tables: a table of the antibodies used in this work \n(Supplemental Table 1), a table of the Mendelian ratio (Supplemental Table 2) an outline of how mice were \ngenotyped, and three supplemental figures: supplemental Fig. 1 the strategy used to genotype the mice, \nSupplemental Fig. 2: expression tests for eGFP and MEGF10-eGFP, and Supplemental Fig. 3: characterisation of \nthe expressed and purified ECD and EGF domains, and their glycosylation. \n \n \nCReDiT \nAll authors contributed to the conceptualization of this work, methodology and writing, review and editing of the \npaper. Investigation: LR, RH. Funding acquisition, resources and supervision: MP, CAJ, SE. \n \nCompeting Interests:  \nThe authors declare no competing interests, except for MP, who is the general editor for the Journal of Muscle \nResearch and Cell Motility.  \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 20\nReferences \nBagley, J.R., L.T. Denes, J.J. McCarthy, E.T. Wang, and K.A. Murach. 2023. The myonuclear domain in adult \nskeletal muscle fibres: past, present and future. J Physiol. 601:723-741. \nBruusgaard, J.C., I.B. Johansen, I.M. Egner, Z.A. Rana, and K. Gundersen. 2010. Myonuclei acquired by overload \nexercise precede hypertrophy and are not lost on detraining. Proc Natl Acad Sci U S A. 107:15111-15116. \nCallebaut, I., V. Mignotte, M. Souchet, and J.P. Mornon. 2003. EMI domains are widespread and reveal the \nprobable orthologs of the Caenorhabditis elegans CED-1 protein. Biochem Biophys Res Commun . \n300:619-623. \nCarter, G.T., M.A. Wineinger, S.A. Walsh, S.J. Horasek, R.T. Abresch, and W.M. Fowler, Jr. 1995. Effect of \nvoluntary wheel-running exercise on muscles of the mdx mouse. Neuromuscul Disord. 5:323-332. \nChikazawa, M., M. Shimizu, Y. Yamauchi, and R. Sato. 2020. Bridging molecules are secreted from the skeletal \nmuscle and potentially regulate muscle differentiation. Biochem Biophys Res Commun. 522:113-120. \nChung, W.S., L.E. Clarke, G.X. Wang, B.K. Stafford, A. Sher, C. Chakraborty, J. Joung, L.C. Foo, A. Thompson, \nC. Chen, S.J. Smith, and B.A. Barres. 2013. Astrocytes mediate synapse elimination through MEGF10 \nand MERTK pathways. Nature. 504:394-400. \nColombatti, A., P. Spessotto, R. Doliana, M. Mongiat, G.M. Bressan, and G. Esposito. 2011. The \nEMILIN/Multimerin family. Front Immunol. 2:93. \nDick, J., and G. Vrbova. 1993. Progressive deterioration of muscles in mdx mice induced by overload. Clin Sci \n(Lond). 84:145-150. \nEgginton, S., I. Badr, J. Williams, D. Hauton, G.C. Baan, and R.T. Jaspers. 2011. Physiological angiogenesis is a \ngraded, not threshold, response. J Physiol. 589:195-206. \nFujii, K., M. Hirano, A. Terayama, R. Inada, Y. Saito, I. Nishino, and Y. Nagai. 2022. Identification of a novel \nmutation and genotype-phenotype relationship in MEGF10 myopathy. Neuromuscul Disord. 32:436-440. \nHamon, Y., D. Trompier, Z. Ma, V. Venegas, M. Pophillat, V. Mignotte, Z. Zhou, and G. Chimini. 2006. \nCooperation between engulfment receptors: the case of ABCA1 and MEGF10. PLoS One. 1:e120. \nHardy, D., A. Besnard, M. Latil, G. Jouvion, D. Briand, C. Thepenier, Q. Pascal, A. Guguin, B. Gayraud-Morel, \nJ.M. Cavaillon, S. Tajbakhsh, P. Rocheteau, and F. Chretien. 2016. Comparative Study of Injury Models \nfor Studying Muscle Regeneration in Mice. PLoS One. 11:e0147198. \nHochreiter-Hufford, A.E., C.S. Lee, J.M. Kinchen, J.D. Sokolowski, S. Arandjelovic, J.A. Call, A.L. Klibanov, Z. \nYan, J.W. Mandell, and K.S. Ravichandran. 2013. Phosphatidylserine receptor BAI1 and apoptotic cells \nas new promoters of myoblast fusion. Nature. 497:263-267. \nHolterman, C.E., F. Le Grand, S. Kuang, P. Seale, and M.A. Rudnicki. 2007. Megf10 regulates the progression of \nthe satellite cell myogenic program. J Cell Biol. 179:911-922. \nHuey, K.A., S.A. Smith, A. Sulaeman, and E.C. Breen. 2016. Skeletal myofiber VEGF is necessary for myogenic \nand contractile adaptations to functional overload of the plantaris in adult mice. J Appl Physiol (1985) . \n120:188-195. \nHyatt, J.P., G.E. McCall, E.M. Kander, H. Zhong, R.R. Roy, and K.A. Huey. 2008. PAX3/7 expression coincides \nwith MyoD during chronic skeletal muscle overload. Muscle Nerve. 38:861-866. \nIram, T., Z. Ramirez-Ortiz, M.H. Byrne, U.A. Coleman, N.D. Kingery, T.K. Means, D. Frenkel, and J. El Khoury. \n2016. Megf10 Is a Receptor for C1Q That Mediates Clearance of Apoptotic Cells by Astrocytes. J \nNeurosci. 36:5185-5192. \nJeong, J., and I.M. Conboy. 2011. Phosphatidylserine directly and positively regulates fusion of myoblasts into \nmyotubes. Biochem Biophys Res Commun. 414:9-13. \nJohnson, T.L., and K.M. Klueber. 1991. Skeletal muscle following tonic overload: functional and structural \nanalysis. Med Sci Sports Exerc. 23:49-55. \nJuros, D., M.F. Avila, R.L. Hastings, A. Pendragon, L. Wilson, J. Kay, and G. Valdez. 2024. Cellular and \nmolecular alterations to muscles and neuromuscular synapses in a mouse model of MEGF10-related \nmyopathy. Skelet Muscle. 14:10. \nLi, C., D. Vargas-Franco, M. Saha, R.M. Davis, K.A. Manko, I. Draper, C.A. Pacak, and P.B. Kang. 2021. \nMegf10 deficiency impairs skeletal muscle stem cell migration and muscle regeneration. FEBS Open Bio. \n11:114-123. \nLogan, C.V., B. Lucke, C. Pottinger, Z.A. Abdelhamed, D.A. Parry, K. Szymanska, C.P. Diggle, A. van Riesen, \nJ.E. Morgan, G. Markham, I. Ellis, A.Y. Manzur, A.F. Markham, M. Shires, T. Helliwell, M. Scoto, C. \nHubner, D.T. Bonthron, G.R. Taylor, E. Sheridan, F. Muntoni, I.M. Carr, M. Schuelke, and C.A. \nJohnson. 2011. Mutations in MEGF10, a regulator of satellite cell myogenesis, cause early onset \nmyopathy, areflexia, respiratory distress and dysphagia (EMARDD). Nat Genet. 43:1189-1192. \nMahdy, M.A.A. 2019. Biotoxins in muscle regeneration research. J Muscle Res Cell Motil. 40:291-297. \nMorgan, J.E., J.R. Beauchamp, C.N. Pagel, M. Peckham, P.  Ataliotis, P.S. Jat, M.D. Noble, K. Farmer, and T.A. \nPartridge. 1994. Myogenic cell lines derived from transgenic mice carrying a thermolabile T antigen: a \nm\nodel system for the derivation of tissue-specific and mutation-specific cell lines. Dev Biol. 162:486-498. \nMurach, K.A., C.S. Fry, E.E. Dupont-Versteegden, J.J. McCarthy, and C.A. Peterson. 2021. Fusion and beyond: \nSatellite cell contributions to loading-induced skeletal muscle adaptation. FASEB J. 35:e21893. \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint \n\n 21\nMurach, K.A., S.H. White, Y. Wen, A. Ho, E.E. Dupont-Versteegden, J.J. McCarthy, and C.A. Peterson. 2017. \nDifferential requirement for satellite cells during overload-induced muscle hypertrophy in growing versus \nmature mice. Skelet Muscle. 7:14. \nPallafacchina, G., B. Blaauw, and S. Schiaffino. 2013. Role of satellite cells in muscle growth and maintenance of \nmuscle mass. Nutr Metab Cardiovasc Dis. 23 Suppl 1:S12-18. \nPark, J., and W.S. Chung. 2023. Astrocyte-dependent circuit remodeling by synapse phagocytosis. Curr Opin \nNeurobiol. 81:102732. \nRichardson, L., D. Wang, R. Hughes, C.A. Johnson, and M. Peckham. 2022. RNA-Seq analysis of a Pax3-\nexpressing myoblast clone in-vitro and effect of culture surface stiffness on differentiation. Sci Rep. \n12:2841. \nRoberts, M.D., J.J. McCarthy, T.A. Hornberger, S.M. Phillips, A.L. Mackey, G.A. Nader, M.D. Boppart, A.N. \nKavazis, P.T. Reidy, R. Ogasawara, C.A. Libardi, C. Ugrinowitsch, F.W. Booth, and K.A. Esser. 2023. \nMechanisms of mechanical overload-induced skeletal muscle hypertrophy: current understanding and \nfuture directions. Physiol Rev. 103:2679-2757. \nRosenblatt, J.D., D. Yong, and D.J. Parry. 1994. Satellite cell activity is required for hypertrophy of overloaded \nadult rat muscle. Muscle Nerve. 17:608-613. \nSaha, M., S. Mitsuhashi, M.D. Jones, K. Manko, H.M. Reddy, C.C. Bruels, K.A. Cho, C.A. Pacak, I. Draper, and \nP.B. Kang. 2017. Consequences of MEGF10 deficiency on myoblast function and Notch1 interactions. \nHum Mol Genet. 26:2984-3000. \nSakuma, K., K. Watanabe, M. Sano, I. Uramoto, K. Sakamoto, and T. Totsuka. 1999. The adaptive response of \nMyoD family proteins in overloaded, regenerating and denervated rat muscles. Biochim Biophys Acta . \n1428:284-292. \nSeiden, D. 1976. Quantitative analysis of muscle cell changes in compensatory hypertrophy and work-induced \nhypertrophy. Am J Anat. 145:459-465. \nSnijders, T., T. Aussieker, A. Holwerda, G. Parise, L.J.C. van Loon, and L.B. Verdijk. 2020. The concept of \nskeletal muscle memory: Evidence from animal and human studies. Acta Physiol (Oxf). 229:e13465. \nSuzuki, E., and M. Nakayama. 2007a. The mammalian Ced-1 ortholog MEGF10/KIAA1780 displays a novel \nadhesion pattern. Exp Cell Res. 313:2451-2464. \nSuzuki, E., and M. Nakayama. 2007b. MEGF10 is a mammalian ortholog of CED-1 that interacts with clathrin \nassembly protein complex 2 medium chain and induces large vacuole formation. Exp Cell Res. 313:3729-\n3742. \nTung, T.T., K. Nagaosa, Y. Fujita, A. Kita, H. Mori, R. Okada, S. Nonaka, and Y. Nakanishi. 2013. \nPhosphatidylserine recognition and induction of apoptotic cell clearance by Drosophila engulfment \nreceptor Draper. Journal of biochemistry. 153:483-491. \nvan den Eijnde, S.M., M.J. van den Hoff, C.P. Reutelingsperger, W.L. van Heerde, M.E. Henfling, C. Vermeij-\nKeers, B. Schutte, M. Borgers, and F.C. Ramaekers. 2001. Transient expression of phosphatidylserine at \ncell-cell contact areas is required for myotube formation. J Cell Sci. 114:3631-3642. \nVargas-Franco, D., R. Kalra, I. Draper, C.A. Pacak, A. Asakura, and P.B. Kang. 2022. The Notch signaling \npathway in skeletal muscle health and disease. Muscle Nerve. 66:530-544. \nWang, Y.X., and M.A. Rudnicki. 2011. Satellite cells, the engines of muscle repair. Nat Rev Mol Cell Biol . \n13:127-133. \nWolny, M., M. Colegrave, L. Colman, E. White, P.J. Knight, and M. Peckham. 2013. Cardiomyopathy mutations \nin the tail of beta-cardiac myosin modify the coiled-coil structure and affect integration into thick \nfilaments in muscle sarcomeres in adult cardiomyocytes. J Biol Chem. 288:31952-31962. \nWouters, M.A., I. Rigoutsos, C.K. Chu, L.L. Feng, D.B. Sparrow, and S.L. Dunwoodie. 2005. Evolution of \ndistinct EGF domains with specific functions. Protein Sci. 14:1091-1103. \nZhou, A.L., S. Egginton, M.D. Brown, and O. Hudlicka. 1998. Capillary growth in overloaded, hypertrophic adult \nrat skeletal muscle: an ultrastructural study. Anat Rec. 252:49-63. \n \n \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 20, 2024. ; https://doi.org/10.1101/2024.10.19.619219doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}