Introduction

Titin is a giant protein in striated muscle and forms an elastic filament in sarcomeres¹, the smallest contractile unit in muscle. One of the functions attributed to titin is to generate passive tension in muscle^2,3, ensuring optimal overlap of the actin-based thin filaments and myosin-based thick filaments for muscle contraction. The elastic properties of titin combined with its arrangement in thesarcomeres, spanning the half-sarcomer, make titin an ideal stress sensor. The mechanosensory properties of titin have been extensively described^4–6 and appear to localize to several signaling hotspots on titin.

In skeletal muscle, the N2A element is the prevalent titin region associated with hypertrophy signaling. The N2A element is known for its interaction with the muscle ankyrin repeat proteins (MARP1-3)^7,8 and calpain3⁹. All MARPs have been implicated in trophicity signaling, with redundancy between the MARP proteins^7,10. MARP1 tethers titin to the thin filament, forming a mechanism for increasing passive tension^11,12. In the heart, MARP1 also interacts with MLP and protein kinase C alpha in intercalated disks, resulting in a maladaptive trophic response in dilated cardiomyopathy¹⁰. MARP2 can be phosphorylated at serine-99 (S69 in mice) by Akt, resulting in reduced differentiation potential of myoblasts¹³. MARP2 has also been linked to the NFkB-pathway in inflammatory responses¹⁴, suggesting MARP2 may have roles in atrophy signaling. MARP3 is the least studied member of the MARPs but appears to play a role in glucose uptake and vascular remodeling in skeletal muscle^15,16. Titin has previously been shown to activate signaling in response to stretch^17,18, providing a potential activation mechanism for N2A to activate trophic signaling.

Muscle hypertrophy is directly associated with titin-based stiffness. High stiffness, as observed in the Ttn^{Δex112–158} and Ttn^{Δex219–225} mouse models^19,20, results in longitudinal hypertrophy, i.e. an increase in serial-linked sarcomeres, to reduce sarcomere length and normalize passive tension. We previously used a surgical model in which we denervated one hemi-diaphragm (unilateral diaphragm denervation, UDD)²¹, inducing passive cyclic stretch of the denervated hemi-diaphragm by the innervated costal to induce (longitudinal) hypertrophy. This hypertrophy was dependent on titin-based stiffness, as higher titin stiffness resulted in exaggerated hypertrophy and lower titin stiffness resulted in attenuated hypertrophy²¹. UDD is a unique model as it allows the in vivo study of passive stretch-induced muscle hypertrophy.

We used UDD as a model for studying titin-mechanosensing and examined the signaling that underlies the hypertrophy response. Our findings support the N2A-MARP signalosome inhibiting longitudinal hypertrophy and show mTOR signaling to be a prominent contributor to longitudinal hypertrophy.

Results

Longitudinal hypertrophy regulates transient trophicity in UDD

A unique aspect of unilateral diaphragm denervation (UDD) is the transient nature of the hypertrophy. This transient nature is likely the result from changes in fractional extension of sarcomeres. Sarcomere length (SL) at end-expiration length was 2.9±0.1 µm in sham, while stretch lengths at end-inspiration were predicted to reach 3.7 µm directly after denervation of the right costal diaphragm (FIG.1A, previously published²¹). Such sarcomere lengths put high strain on titin (FIG.1A, right panel), providing a potent trigger for mechanosensing and subsequent hypertrophy signaling. Our results show that during 6-days of UDD the denervated costal rapidly hypertrophies, followed by slow onset of atrophy (FIG.1B). The mass increase coincides with longitudinal hypertrophy of muscle fibers, addition of serial-linked sarcomeres, adding 952±81 sarcomeres (30.2% increase in total fiber length; p<0.0001; FIG.1C) following 6-days of UDD. This increase in sarcomere number appeared to stabilize by 6-days UDD, as at 12-day UDD sarcomere addition was measured to be 778±87 sarcomeres versus sham levels (not significantly different compared to 6-day UDD). Assuming that fractional extension of sarcomeres during inspiration decreases with longitudinal hypertrophy, 30.2% increase in sarcomere number would reduce the sarcomere length at end inspiration from 3.7 µm to ∼2.8 µm, in close agreement with whole body formaldehyde perfusion experiments which showed sarcomere lengths of 2.7±0.1 µm²¹. As a SL of 2.8 µm falls below the end-expiration SL, the denervated costal effectively no longer experiences “stretch”, reducing titin’s mechanosensing trigger and thus explaining the transient nature of the hypertrophy seen in UDD.

Figure 1.

To confirm that stretch is the trigger for hypertrophy, rather than denervation, we performed bilateral diaphragm denervation in rats (BDD), resulting in complete inactivity of the diaphragm. Note that we attempted BDD in mice but experienced high mortality rates, while survival in rat was >75%. All rats were of similar body weight pre-surgery and showed a slight decrease in body weight after 3-day BDD (Supplemental FIG.1A; p=0.036), and comparable tibial length and soleus weights (Supplemental FIG.1B & C). Similar to mice, rats present with increased denervated costal mass after 3-day UDD: 14.68±0.64 mg/mm versus sham rats 10.82±0.39 mg/mm (FIG.1D; p>0.0001; muscle weight normalized to tibial length). However, 3-day BDD rats did not develop increased mass 11.92±0.89 mg/mm compared to sham rats 10.82±0.39 mg/mm. The increase in costal mass at 3-day UDD results in part from an increase in longitudinal hypertrophy, with denervated costal diaphragm adding 933±301 serial sarcomeres compared to sham animals (FIG.1E; p=0.018), whereas 3-day BDD rats showed no difference in serial sarcomere number (7559±313 versus sham 7740±111). Thus, stretch and not denervation triggers longitudinal hypertrophy.

To support titin’s role as mechanosensor of stretch driven hypertrophy, we performed UDD on Rbm20^ΔRRM mice and Rbm20 ko rats. Rbm20 is an RNA splicing-factor for titin and both the mouse and rat model generate larger, more compliant titin isoforms and should thus be less sensitive to stretch. Rbm20^ΔRRM mice showed relatively less tissue mass increase compared to wildtype (WT) 3, 6, or 12-days UDD (Supplemental FIG.2A), without changes in longitudinal hypertrophy at 6-days UDD (Supplemental FIG.2B) suggesting the difference in mass is a result from radial fiber growth, in agreement with our previous findings²¹. Finally, we compared denervated costal mass from Sprague Dawley (SD) and Rbm20 ko rats 3-day UDD, showing a difference in mass increase of 24.4±3.8% in Rbm20 ko and 35.5±5.9% in SD rats (p=0.02; Supplemental FIG.2C), supporting that stretch evokes hypertrophy in muscle and that titin compliance modulates the extent of hypertrophy across species. Thus, muscle stretch in the costal diaphragm is a potent trigger for longitudinal muscle hypertrophy, with titin being a primary sensor.

Global transcriptome and proteome level studies reveal a distinct subset of genes and proteins involved in UDD stretch hypertrophy

To identify the cellular processes involved in stretch-based hypertrophy we performed both transcriptome wide studies by RNA sequencing (RNAseq) as well as proteome studies by global mass spectrometry (MS) on our rat 3-day sham, UDD and BDD samples. Initial studies focused on separating the effect of denervation versus stretch. Principal component analysis of the top 500 genes showed close clustering of samples by group (Supplemental FIG.3A), suggesting distinct gene programs are active between sham, UDD and BDD. At the protein level, principal component analysis of the top 250 proteins suggests narrow separation of groups, predicting overlap of the BDD and UDD samples (Supplemental FIG.3B). Gene expression studies revealed very close overlap of BDD and UDD profiles with 77.15% overlap in differentially expressed genes (DEG’s: p_adj<0.05) and just 9.4% of DEG’s being specific for UDD (FIG.2A). Volcano plots of UDD>Sham (FIG.2B, left panel) and BDD>Sham (FIG.2B; middle panel) showed the ∼11.000 DEG’s (Supplemental data Tables 1-6) to be nearly equally distributed between up and downregulated DEG’s. Direct comparison of UDD>BDD (FIG.2B, right panel) revealed 850 DEG’s to be uniquely associated with UDD, of which 581 were upregulated (Supplemental data Table 3). MS studies broadly agreed with the RNAseq data, showing UDD and BDD share overlapping expression programs (FIG. 2A & C). Twenty percent of the differentially expressed proteins (DEP’s; P<0.05) detected are unique to UDD (FIG.2C). Volcano plots of UDD>Sham (FIG.2D, left panel) and BDD>Sham (FIG.2D, middle panel) showed 889 and 789 DEP’s, respectively, (Supplemental data Tables 7-12) which were primarily increased. Comparing UDD>BDD revealed 173 DEP’s (FIG.2D, right panel) that are unique to UDD. Most of DEP’s (and DEG’s) found are related to transcription, translation, energetics and folding, with the highest fold DEP’s in rat UDD>BDD being: Med15 (15.5-fold), Eif3i (6.0-fold) and Gtpbp3 (3.9-fold) (Supplemental data Table 9). GOterm enrichment of UDD>BDD, separated by downregulated and upregulated DEG’s (Supplemental FIG.3C, left top and bottom graph, respectively), and DEP’s (Supplemental FIG.3C, right top and bottom graph, respectively), indicated that upregulated DEG/DEP’s are primarily related to muscle development and function consistent with an active hypertrophy program and downregulated targets with metabolism and cellular respiration. To identify the pathways involved in UDD>BDD, we searched the DEG’s against the KEGG PATHWAY database (Supplemental data Table 6), confirming GOterm results and indicating that the DEG’s are involved in muscle remodeling. Thus, UDD has a distinct expression profile from denervation (BDD), focused on remodeling of muscle.

Figure 2.

Exploring if titin-mechanosensing was active in either UDD or BDD we generated heatmaps of titin-binding proteins based on RNAseq (Supplemental FIG.4) and global MS data (FIG.2E). To narrow down proteins that were differential between UDD and BDD we tested (2-way ANOVA p_int<0.1) the z-score and found: MARP2 (Ankrd2; p=0.003), Cryab (p= 0.026), Csrp3 (p= 0.002), Hsp90ab1 (p= 0.09), Hspb1 (p= 0.008) and Smyd2 (p= 0.004), to be unique to UDD. Interestingly, these DEP’s (MARP2, Smyd2, Hsp1b, Hsp90ab and Cryab) are established binding partners of the N2A element of titin.²²

Titin phosphorylation in UDD

With the rat data indicating changes in titin-associated signaling and titin stiffness modifying the hypertrophy response in mice, we evaluated how UDD affects titin post-translationally. We performed MS on phosphorylation enriched peptides from 24-hour UDD diaphragm samples. 24-hour UDD was selected to study the early phosphorylation events. MS revealed 2870 phosphorylated peptides (p<0.05), of which 142-sites were in titin (∼700-sites identified including non-significant sites; Supplemental data Table 13). We opted to define titin phosphorylation by analyzing the phosphorylation at the domain level to study if titin showed regional changes in phosphorylation (FIG.3A and Supplemental data Table 14), as total titin phosphorylation (FIG.3B) was unchanged. Domain level analysis indicated titin phosphorylation was primarily changing in the I-band. This motivated quantifying titin phosphorylation by segment (FIG.3C and Supplemental data Table 15; following established naming conventions in Kolmerer and Labeit¹, and Bang et al²³). Segmental analysis of titin (FIG.3C) revealed a marked increase in phosphorylation of the N2A-element and I/A-junction, and a decrease in the Z/I-junction. Focusing on the increase in phosphorylation of the N2A-element, a known trophicity signaling hub (reviewed in ²²), we assessed the specific domain phosphorylation of the N2A-element (FIG.3C, highlight). In the N2A-element we found 7 phosphorylation sites (FIG.3D-E) of which S9346 & S9350 are located in a linker sequence between I79 and I80, S9483 in I80, S9459 & S9520 in the N2A unique sequence, S9654 in I81 and S9643 in I82). The 2 sites located in the N2A unique sequence were significantly upregulated following UDD (FIG.3D-E). pS9459 and pS9520 currently have no known function but could serve as a recruitment signal (see discussion).

Figure 3.

The MARP proteins regulate hypertrophy in UDD

The MARP1 and MARP2 proteins were strongly expressed both at the RNA level and protein level following UDD (FIG.2E). MARPs have been shown to be important regulators of trophicity and have been proposed as intermediaries between titin and trophic signaling pathways (see discussion). As the MARPs share high homology and possible redundancy⁷, we performed 6-day UDD on both single-and multi-KO combinations of the MARPs. We initially performed UDD on the MARP triple KO mice (MARP tKO; knockout of Ankrd1, -2 and -23 genes) and found a 12% reduction in the hypertrophy response compared to WT (FIG.4A; p=0.0099). This reduction in hypertrophy supports a potential link between titin-mechanosensing and MARP-based trophic signaling. To test if one or a combination of MARPs caused the reduction in hypertrophy, we performed 6-day UDD on MARP1 (Ankrd1), MARP2 (Ankrd2) and MARP3 (Ankrd23) KO mice (FIG.4B-D) and double KO for MARP1/2, MARP1/3 and MARP2/3 (Supplemental FIG.6).

Figure 4.

Whereas MARP1 KO mice did not show a difference in denervated costal diaphragm mass, MARP2 KO mice showed a 14% increase (p=0.003) in mass and MARP3 KO mice showed a 13% reduction (p=0.004) in mass gain. These data suggest that MARP3 attenuates hypertrophy. However, all MARP double KO mice also presented with attenuated hypertrophy (Supplemental FIG.6). This suggests that the separate MARP proteins could have distinct functions in UDD, and there may be a dependency for two MARPs in regulating trophic signaling, further discussed in the discussion section.

MARPs negatively regulate longitudinal hypertrophy

Prompted by the reduction in diaphragm mass gain during UDD in MARP tKO mice and by the complexity of targeting all the single MARP KOs, we assessed longitudinal hypertrophy in 6-day UDD MARP tKO samples. MARP tKO mice at baseline have fewer serial sarcomeres than WT (2476±55 vs. 3157±69, respectively; p<0.0001), however 6-days UDD MARP tKO had similar numbers of sarcomeres to WT, 4081±44 versus 4109±59 (FIG.5A). These data show MARP tKO mice add 653 ± 91.6 more sarcomeres than wildtype mice (FIG.5B; 1605±71 vs. 952±59, respectively; p<0.0001). To discern a possible mechanism of the MARPs inhibiting longitudinal hypertrophy, we probed several candidate proteins of hypertrophy pathways that were prominent in the MS datasets. Western blots for Mapk1/3, Calcineurin, mTor, P70 s6k and 4E-bp1 were performed on costal diaphragms of WT and MARP tKO, 6-day Sham and UDD animals. All samples were normalized to Gapdh and data were presented as relative to WT sham (FIG.5C; representative western blot images in FIG.5D). Mapk1 showed increases in protein level that were comparable between WT and MARP tKO, whereas MAPK3 showed preferential upregulation in MARP tKO (p= 0.0099). Calcineurin was unchanged following UDD, note that in a t-test the WT mice showed a significant increase in UDD. In WT mice, mTor showed a striking increase in expression following 6-day UDD (p= 0.0006), whereas in MARP tKO this was trending (P= 0.08). Downstream proteins of mTORC1; P70 s6k and 4E-bp1 both showed altered regulation, with P70 s6k being similarly upregulated in WT (p= 0.0005) versus MARP tKO (p= 0.001), while 4E-bp1 was only significantly upregulated in MARP tKO (p= 0.033). 2way-ANOVA for the signaling proteins did not indicate specific pathway changes between WT and MARP tKO.

Figure 5.

mTor signaling regulates longitudinal hypertrophy in UDD

Motivated by the data from the mTor western blots in the MARP tKO mice and mTor being a prominent regulator of skeletal muscle hypertrophy, we performed UDD in WT mice treated with rapamycin, an inhibitor of mTORC1 signaling and skeletal muscle hypertrophy. The rat transcriptome studies also indicated calcium signaling (Supplemental data Table 6) to be a prominent pathway in UDD. To test the role of calcium signaling we included a second inhibitor; cyclosporin A, an inhibitor of the calcineurin-NFAT pathway, another hypertrophy regulating pathway. Mice received twice daily intraperitoneal injections of either rapamycin (2.5 mg/kg/day), cyclosporin A (25 mg/kg/day) or vehicle (DMSO), starting 3-days prior to surgery until sacrifice of the mice (schematic in FIG.6A). Cyclosporin A treated mice did not show a change in hypertrophy (FIG.6B) or serial sarcomere addition (FIG.6C), indicating that the calcineurin-NFAT pathway is not a primary mechanism for hypertrophy in UDD. Treatment did not affect the innervated left costal diaphragm (FIG.6D), body weight (FIG.6E) or tibia length (FIG.6F). Interestingly, mice that received rapamycin displayed less hypertrophy of the denervated costal diaphragm (FIG.6B; - 11.2%; p<0.001) compared to vehicle treated mice. This attenuated hypertrophy response coincided with a reduction in serial sarcomere addition (FIG.5C; - 8.5%; p<0.001) compared to vehicle. Note that following 3-day UDD, rapamycin treated mice did not show significant increases in serial sarcomeres compared to untreated sham animals (Δ67 ± 81 sarcomeres, versus vehicle treated mice Δ363 ± 79 sarcomeres). This suggested rapamycin treatment almost completely inhibited longitudinal hypertrophy and that mTORC1 signaling plays a vital role in longitudinal hypertrophy development. We thus propose that mTORC1 signaling positively regulates longitudinal hypertrophy in skeletal muscle and that titin’s N2A-element tunes the extent of longitudinal hypertrophy through the MARPs to prevent excess hypertrophy (graphic summary in FIG. 6G).

Figure 6.

Discussion

Longitudinal hypertrophy as a mechanism for reducing stretch-induced mechanosensing

One of the most striking aspects of the UDD surgical model is the early transient hypertrophy. No other muscle denervation model induces hypertrophy, supporting the notion that stretch, even in inactive muscle, is a potent driver of hypertrophic growth. The extreme nature of the stretch at work in UDD, ∼25% stretch of the denervated costal by the innervated costal at a frequency of 120-230 times a minute (respiration rate)²¹, forms a potent trigger for muscle hypertrophy. Following 6-days UDD the denervated costal diaphragm develops 49.7±10.0% (FIG.1B) increase in mass, which is primarily caused by addition of 952±81 sarcomeres (FIG.1C) and to a lesser extent by radial fiber growth²¹. The transient nature of the hypertrophy can be explained by the reduction in sarcomere strain due to the addition of sarcomeres in series (longitudinal hypertrophy), removing the “trigger” that underlies the hypertrophy signaling. This hypothesis fits the data (FIG.1B-C), where before remodeling UDD costal width (i.e., fiber length) is ∼8.7 mm (3000 sarcomeres x SL 2.9 µm), and when stretched 25%²¹ equals a width of ∼10.8 mm. Following 6-days hypertrophic remodeling (UDD), costal width is ∼10.8 mm (4000 sarcomeres x SL 2.7 µm). This suggests that the costal width increase attenuates the hypertrophy trigger caused by stretching, finally resulting in atrophy (FIG.1B; 35-day UDD). Denervation itself does not appear to induce hypertrophy of the diaphragm, as 3-day BDD in rat showed no hypertrophy (FIG.1D-E). UDD being a denervation model shows that hypertrophy is not necessarily dependent on a muscle’s ability to contract. This indicates that skeletal muscle may derive its signals for hypertrophy during the muscle relaxation-phase when the muscle is at its longest state. Muscle stretch in human is known to benefit muscle growth and is a vital part of exercise routines (reviewed in ^24,25). Perhaps muscle antagonism may provide sufficient stretch to induce hypertrophy. As contracting muscles inadvertently stretch their relaxed antagonists, they create an elegant feedback loop that promotes both muscle growth and the maintenance of muscle mass.

Titin’s response to stretch

Posttranslational changes in titin, particularly phosphorylation, have been studied mostly in relation to passive tension development^26–28. Specific PTMs for titin have not been widely studied and only recently has ubiquitination been shown to recruit autophagic receptors to the kinase domain of titin^29,30. Autophosphorylation of titin kinase at Y170 is considered one of the classic mechanosensing responses in titin resulting in phosphorylation of the autophagic receptor Nbr1 at S115/116, activating autophagy signaling in vitro³¹. We did not observe any signs following UDD that titin kinase was autophosphorylated at Y170. This could be related to phospho-peptide abundance being below the detection limit, or that the titin kinase is inactive³². In total we identified ∼700 phosphorylation sites in titin of which 142 were significantly affected by stretch. These sites were distributed along the entire length of titin with several hot spots in the PEVK region, likely involved in stiffness regulation, and several in the Z-disk, which could be related to signaling or structural interactions. The sites in the N2A-element (FIG.3) were of particular interest as previously it has been suggested that S9540 (S9895 according to diaphragm RNAseq by Brynnel et al²⁰) could serve as a recruitment signal for MARP1³³, a protein that is highly upregulated following UDD (FIG.2E). MARPs have been shown to quench N2A phosphorylation^33,34, which make the phosphorylation sites in the N2A segment tantalizing targets for studying titin-MARP binding. If S9459 and S9520 (FIG.3D-E) play such a role remains to be determined, but such insights could provide future avenues for manipulating titin-MARP interactions. Similarly, we found many titin-associated proteins to show differential phosphorylation in UDD, including 3 sites in the MARPs (Supplemental FIG.5). How these sites contribute to signaling or recruitment to the N2A-element remains to be seen but provide tantalizing targets for follow-up study.

The MARP proteins and muscle trophicity

Global transcriptome and proteome studies from 3-day UDD showed multiple titin-associated proteins being upregulated (FIG.2E). We focused our efforts on the MARP proteins, as they are known interacting partners of titin’s N2A-element and have been shown to be important for hypertrophy regulation in the heart^35,36. We used UDD on MARP KO mice, focusing on the triple KO for MARP1-3 to account for expected redundancy^7,34. MARP tKO showed a 12% reduction in hypertrophy following 6-day UDD (FIG.4A). This suggested stretch-mechanosensing operated through the MARP proteins. In an attempt to isolate a single MARP protein as the main effector, we performed UDD on single MARP knock-out mice (FIG.4B-D). MARP1 KO did not reveal changes in hypertrophy. However, we previously established that MARP1 localizes to the N2A-segment of titin following UDD²¹. We also independently determined that MARP1 cross-links titin to the thin filament to increase passive tension^37,38, suggesting MARP1 plays mechanical roles over trophic regulation in skeletal muscle. MARP2 KO developed an exaggerated hypertrophy and lastly MARP3 KO showed an attenuated hypertrophy response to UDD, suggesting MARP2 and 3 play opposing roles in hypertrophy regulation. MARP2 interacts with Akt2¹³, providing a tentative link with the mTor pathway, discussed below. Additionally, MARP2 interaction with P50-NFκB acting as an analogue for IκB¹⁴ suggests a possible role in inhibiting NFκB atrophy signaling. MARP3 is poorly understood but has been linked to glucose metabolism¹⁵, forming another tentative link to mTor signaling³⁹. To specifically determine if the MARPs affected longitudinal hypertrophy we measured the number of serial sarcomeres across the width of the costal diaphragm in MARP tKO and found that the tKO mice added 653.2 ± 91.55 more sarcomeres than WT following 6-day UDD (FIG.5B). This suggests that the MARP proteins inhibit longitudinal hypertrophy. A possible mechanism could be that following stretch the three MARPs bind and compete for the N2A-binding site^7,34 resulting in translocation of specific MARPs for signaling purposes. The benefit of MARP deletion in cardiomyopathy was shown by Lange et al¹⁰, where deletion of MARP1/2 ameliorated MLP KO induced DCM. MARP KO has not proven detrimental in mice⁴⁰, making this family an attractive target for therapeutic intervention.

mTor and longitudinal hypertrophy

Muscle hypertrophy is regulated through a number of pathways, with the insulin-insulin growth factor (IGF) mediated pathway^41–43 being the most well understood in skeletal muscle and the calcineurin-NFAT pathway in cardiac muscle^44,45. With most studies focused on radial (cross-sectional) hypertrophy, we aimed to gain insight into the regulatory mechanism underlying longitudinal hypertrophy, two types of hypertrophy that are not necessarily mutually exclusive. The mTor signaling pathway was a prime candidate as UDD in MARP tKO mice suggested altered mTor activity (FIG.5C-D). This prompted us to test inhibition of mTor signaling and see if mTor regulates longitudinal hypertrophy. Using rapamycin⁴⁶, an inhibitor that targets the mTORC1 (protein synthesis regulation) complex, we found a reduction in longitudinal hypertrophy following 3-days UDD compared to vehicle treated mice (FIG.6B). This strongly suggests that the mTORC1-pathway is in-part responsible for longitudinal hypertrophy following UDD. Although we focused on mTor signaling in longitudinal hypertrophy development, we do not exclude other pathways being important. mTOR formed an attractive target as previous work showed that longitudinal stretch phosphorylates Akt and upregulates MARP2⁴⁷, forming a tentative link between longitudinal stretch, MARPs and the mTor pathway. Further studies are needed to establish the roles of the various pathways in stretch hypertrophy.

In conclusion, we found that the transient hypertrophy induced by UDD is dependent on muscle fiber length, with longitudinal hypertrophy reducing the trigger for hypertrophy. The hypertrophy coincides with increased phosphorylation of the N2A element. The N2A-associated MARP proteins are strongly upregulated following UDD and deletion of the MARPs increases the extent of longitudinal hypertrophy following UDD, indicating the MARPs serve roles in inhibiting longitudinal hypertrophy. MARP tKO mice show altered regulation in the mTORC1 pathway following UDD and inhibition of mTORC1 by rapamycin shows mTor is a main regulator of longitudinal hypertrophy.

Methods

Animal studies

All experiments were done in accordance with the University of Arizona Institutional Animal Care and Use Committee and followed the US National Institutes of Health Using Animals in Intramural Research guidelines for animal use. We used 3-month-old C57BL/6J mice referred to as wildtype (WT), and 6-month-old Sprague Dawley rats (SD). Knockout models for titin binding proteins MARP1-3 (Ankrd1, Ankrd2 and Ankrd23)^10,40,48 were kindly provided by Dr Ju Chen and Dr Stephan Lange. Homozygous Rbm20^ΔRRM mice ^21,49 and Rbm20 ko rats^50,51, have previously been described. Mice were maintained on a C57BL/6J background, with the data from the MARP mice being on a black swiss background.

Surgical procedure

For unilateral diaphragm surgery (UDD)²¹ or bilateral diaphragm denervation (BDD) studies, mice or rats were anaesthetized with 2-3% isoflurane and a small incision was made in the neck area just above the clavicle. The right phrenic nerve was isolated behind the sternohyoid muscle, and a 3–4 mm section was transected at the height of the supraclavicular nerve branch. For BDD surgery both left, and right phrenic nerves being transected. Sham operated animals underwent the same procedure, except the phrenic nerve was left intact. Animals were sacrificed 1, 3, 6, 12 or 35-days after surgery for morphometric analysis and tissue harvest.

Pharmacological inhibition of hypertrophy

Inhibitor studies with rapamycin (mTOR inhibitor) and cyclosporin A (Calcineurin inhibitor) were performed by injecting mice, twice daily intra-peritoneal (IP), with 2.5 or 25 mg/kg/day, respectively. Each inhibitor was dissolved in Dimethylsulfoxide (DMSO) and diluted to 20% DMSO with saline solution just before injection. Mice received inhibitors or vehicle (20% DMSO in saline) starting 3-days prior to UDD surgery to prime the mice and continued following UDD to inhibit hypertrophy growth (FIG.4C). Mice were sacrificed after 3-days UDD and processed for serial sarcomere measurements as described below.

Serial sarcomere measurements

Previously described in van der Pijl et al²¹. Briefly, mice were anaesthetized with a 140/10 mg/kg ketamine/xylazine solution, a small incision to visualize the jugular vein, which was subsequently cannulated for perfusion. Mice were perfused with a solution consisting of 4% formaldehyde, with 70 U/mL of heparin in phosphate buffered saline (PBS), after which the diaphragm was removed and stored in 4% formaldehyde in PBS overnight for complete fixation. Full length diaphragm midcostal strips were gently dissected and flattened between glass slides, costal width was measured using a caliper and sarcomere lengths were measured using a He/Ne laser diffraction system.

Alternatively, muscle fiber bundels from chemically demembranated full length diaphragm midcostal strips of 3-day mice were flattened between glass slides. Costal width and sarcomere lengths were measured using a Zeiss Axio Imager M1 microscope (Zeiss), at ×50 magnification to measure the length of the muscle bundles and ×640 for sarcomere length along four points of the fiber bundles. Images were captured using AxioCam MRc with Axiovision software (Zeiss) and images were calibrated using a 0.01 mm stage micrometre (Edmund Optics). To determine the number of serial sarcomeres the muscle bundle length (costal width) was divided by the sarcomere length.

Demembranating solution consisted of a relaxing solution (in mM; 20 BES, 10 EGTA, 6.56 MgCl2, 5.88 NaATP, 1 DTT, 46.35 K-propionate, 15 creatine phosphate, pH 7.0), with 1% Triton-X-100 at 4°C, and protease inhibitors (phenylmethylsulfonyl fluoride (PMSF), 0.25 mM; leupeptin, 0.04 mM; E64, 0.01 mM,), after demembranation, samples were stored in just relaxing solution plus inhibitors (without triton X-100) at 4°C.

Transcriptome studies

RNA sequencing (RNAseq) was performed on right costal diaphragm samples collected from 3-day sham and UDD animals and flash frozen in liquid nitrogen until further processing. For RNA extraction, samples were incubated overnight in RNAlater-ICE (Thermo Scientific) and subsequently transferred to RLT buffer for extraction according to the RNeasy Fibrous Tissue Mini Kit (Qiagen). Tissue disruption was achieved using a Bullet Blender (Next Advance) and Green Eppendorf lysis kit tubes (Next Advance), by grinding samples for 4 minutes at setting 10. Thereafter, total RNA extraction was performed following the RNeasy Fibrous Tissue Mini Kit’s instructions and quantified using a Nanodrop ND-1000 spectrophotometer (Thermo Scientific). Each sample consisted of 3 biological replicate sham or UDD samples. Both Library preparation and sequencing was performed by the University of Chicago Genomics Facility, Chicago, USA. Briefly, library preparation: rRNA was depleted from RNA preparations from 1 µg total RNA. Libraries were prepared using an RNA Library Prep Kit from Illumina following the manufacturer’s instructions. Sequencing was performed on an Illumina Hiseq2500 sequencer using 100 bp paired-end sequencing. For RNAseq analysis see²⁰. Briefly, Adapters and low-quality reads were removed with Trim Galore and reads were mapped to the rat genome (Release mRatBN7.2) using STAR⁵² with default settings. Differentially expressed genes were determined with DESeq2⁵³. Genes with population adjusted p-values (p_adj) <0.05 were considered differentially expressed. For titin splicing, percent spliced in index (PSI) was calculated as a measure for determining if an exon is spliced in, following the titin exon annotation by Bang et al.⁵⁴

Preparation of muscle for mass spectrometry analysis

Diaphragm muscle was ground to a fine powder using Dounce homogenizers cooled in liquid nitrogen and acclimated to –20°C for 30 min before continuing. Tissue powder was resuspended at a concentration of 50 mg/ml in a Urea buffer (4 M urea, 1 M thiourea, 25 mM Tris–HCl, 75 mM dithiothreitol, 1.5 % SDS, 25 % glycerol, pH 6.8) with protease inhibitors (0.04 mM E-64, 0.16 mM leupeptin, and 0.2 mM PMSF). The solution was mixed for 4 min, followed by 10 min of incubation at 60°C. Samples were centrifuged at 12.000 rpm and the supernatant flash frozen for storage at −80°C.

In-solution Tryptic Digestion

50 µg of rat costal diaphragm lysate was subjected to acetone precipitation by adding six times the sample volume of pre-chilled 100 % acetone and incubated one hour at -20°C. The precipitates were centrifuged at 16,000 x g for 10 minutes at 4°C and the acetone was removed.

400 µL of pre-chilled 90% acetone was added to the protein pellet, briefly vortexed and centrifuged at 16,000 x g for 5 minutes at 4°C. The remaining acetone was removed, the protein pellets were air dried for 3 minutes, resuspended in 100 µL of 50 mM NH₄HCO₃ and sonicated for 5 minutes. The samples were supplemented with dithiothreitol (DTT) at a final concentration of 5 mM and incubated at 70°C for 30 minutes. Samples were cooled to room temperature for 10 minutes and incubated with 15 mM acrylamide for 30 minutes at room temperature while protected from light. The reaction was quenched with DTT with a final concentration of 5 mM and incubated in the dark for 15 minutes. One µg of Lys-C was added to each sample and incubated at 37° C for 2-3 hours while shaking at 300 rpm followed by the addition of 50 µL of 50 mM ammonium bicarbonate and 2 µg of trypsin and incubation overnight at 37°C while shaking at 300 rpm. 14.7 µL of 40 % FA/1 % HFBA was added to each sample and incubated for 10 minutes (final concentration is 4 % FA/0.1 % HFBA) to stop trypsin digestion. The samples were desalted with Pierce Peptide Desalting Spin Columns per the manufacturer’s protocol (ThermoFisher Scientific, cat no. 89852) and the peptides were dried by vacuum centrifugation. The dried peptides were resuspended in 20 µL of 0.1 % FA (v/v) and the peptide concentration was determined with the Pierce Quantitative Colorimetric Peptide Assay Kit per the manufacturer’s protocol (ThermoFisher Scientific, cat no. 23275). 350 ng of the final sample was analyzed by mass spectrometry.

Phosphoproteomics

To determine global differences in protein phosphorylation abundance between sham or UDD, 1 mL of protein lysate corresponding to 50 mg diaphragm per sample (pooled costal diaphragm of 2 mice) was subjected to in-solution tryptic digestion and phosphopeptide enrichment using sequential enrichment from metal oxide affinity chromatography per manufacturer’s protocol (Thermo Scientific, cat no. A32993 & A32992) similar to as previously described^55,56. The dried peptides were resuspended in 20 µL of 0.1 % FA (v/v) and the peptide concentration was determined with the Pierce Quantitative Colorimetric Peptide Assay Kit per the manufacturer’s protocol. 350 ng of the final sample was then analyzed by mass spectrometry.

Mass Spectrometry

HPLC-ESI-MS/MS was performed in positive ion mode on a Thermo Scientific Orbitrap Fusion Lumos tribrid mass spectrometer fitted with an EASY-Spray Source (Thermo Scientific, San Jose, CA). NanoLC was performed using a Thermo Scientific UltiMate 3000 RSLCnano System with an EASY Spray C18 LC column (Thermo Scientific, 50cm x 75 μm inner diameter, packed with PepMap RSLC C18 material, 2 µm, cat. # ES803); loading phase for 15 min at 0.300 µL/min; mobile phase, linear gradient of 1–34 % Buffer B in 119 min at 0.220 εL/min, followed by a step to 95% Buffer B over 4 min at 0.220 µL/min, hold 5 min at 0.250 µL/min, and then a step to 1 % Buffer B over 5 min at 0.250 µL/min and a final hold for 10 min (total run 159 min); Buffer A = 0.1 % FA/H₂O; Buffer B = 0.1 % FA in 80 % ACN. All solvents were liquid chromatography mass spectrometry grade. Spectra were acquired using XCalibur, version 2.3 (ThermoFisher Scientific). A “TopSpeed” data-dependent MS/MS analysis was performed (acquisition of a full scan spectrum followed by collision-induced dissociation mass spectra of the Top N most intense precursor ions within the 3 second cycle time). Dynamic exclusion was enabled with a repeat count of 1, a repeat duration of 30 seconds, an exclusion list size of 500, and an exclusion duration of 40 seconds.

Label-free Quantitative Proteomics

Progenesis QI for proteomics software (version 2.4, Nonlinear Dynamics Ltd., Newcastle upon Tyne, UK) was used to perform ion-intensity based label-free quantification as previously described⁵⁷. In brief, in an automated format, raw files were imported and converted into two-dimensional maps (y-axis = time, x-axis =m/z) followed by selection of a reference run for alignment purposes. An aggregate data set containing all peak information from all samples was created from the aligned runs, which was then further narrowed down by selecting only +2, +3, and +4 charged ions for further analysis. The samples were then grouped according to treatment. Peak lists of the top ten fragment ion spectra were exported in Mascot generic file (mgf) format and searched against either the 2020_06 Swiss-Prot Rattus norvegicus database (8128 entries), the 2018_11 Swiss-Prot Mus musculus database (17008 entries), or species respective TrEMBL databases using Mascot (Matrix Science, London, UK; version 2.6.0). The search variables that were used were: 10 ppm mass tolerance for precursor ion masses and 0.5 Da for product ion masses; digestion with trypsin; a maximum of two missed tryptic cleavages; variable modifications of oxidation of methionine, phosphorylation of serine, threonine, and tyrosine, and carbamidomethylation of cysteine; 13C = 1. The resulting Mascot .xml file was then imported into Progenesis, allowing for peptide/protein assignment, while peptides with a Mascot Ion Score of <25 were not considered for further analysis. Abundances were normalized to the total ion current (TIC) to correct for differences in sample loading and instrument response. For quantification, proteins must have possessed at least one or more unique, identifying peptide.

Visualization and analysis of transcriptome and proteome data

Figures were generated using online resources, briefly, quantitative Venn diagrams were made using https://www.biovenn.nl/⁵⁸. Lolipop graphs of gene onthology (GO) enritchment analysis by ShinyGO v0.75 http://bioinformatics.sdstate.edu/go/⁵⁹. Heatmaps were generated using a combination of Graphpad Prism v9.1, Perseus v2.0.3.1⁶⁰ and Heatmapper (http://heatmapper.ca/)⁶¹. Perseus v2.0.3.1 was also used to analyze principal components and visualized using Graphpad Prism v9.1.

Western blot

Western blot experiments previously described in van der Pijl et al²¹. Proteins were transferred onto Immobilon-P PVDF 0.45 μm membranes (Millipore) using semi dry transfer (Bio-Rad). Membranes were blocked with Odyssey blocking buffer (Li-Cor Biosciences) for 1 hour, and subsequently probed with primary antibodies at 4°C overnight (See Table 1). Near Infra-Red dyes were used as secondary antibodies for detection with Odyssey CLx Imaging System (Li-Cor Biosciences, United states).

Table 1

Supplemental figures

S Figure 1.

S Figure 2.

S Figure 3.

S Figure 4.

S Figure 5.

S Figure 6.

Additional information

Funding

This work was financially supported by National Institutes of Health grants R01HL121500 (CO), R01AR083233 (HG) and R35HL144998 (HG)