Titin-based mechanosensing modulates muscle hypertrophy

Abstract

Background: Titin is an elastic sarcomeric filament that has been proposed to play a key role in mechanosensing and trophicity of muscle. However, evidence for this proposal is scarce due to the lack of appropriate experimental models to directly test the role of titin in mechanosensing.

Methods: We used unilateral diaphragm denervation (UDD) in mice, an in vivo model in which the denervated hemidiaphragm is passively stretched by the contralateral, innervated hemidiaphragm and hypertrophy rapidly occurs.

Results: In wildtype mice, the denervated hemidiaphragm mass increased 48 ± 3% after 6 days of UDD, due to the addition of both sarcomeres in series and in parallel. To test whether titin stiffness modulates the hypertrophy response, RBM20^ΔRRM and Ttn^ΔIAjxn mouse models were used, with decreased and increased titin stiffness, respectively. RBM20^ΔRRM mice (reduced stiffness) showed a 20 ± 6% attenuated hypertrophy response, whereas the Ttn^ΔIAjxn mice (increased stiffness) showed an 18 ± 8% exaggerated response after UDD. Thus, muscle hypertrophy scales with titin stiffness. Protein expression analysis revealed that titin-binding proteins implicated previously in muscle trophicity were induced during UDD, MARP1 & 2, FHL1, and MuRF1.

Conclusions: Titin functions as a mechanosensor that regulates muscle trophicity.

Keywords: Denervation; Diaphragm; Hypertrophy; Mechanosensing; Muscle stretch; Titin.

Figure 1

(A) EMG signal from implanted telemetry device showing electrical activity in the right hemi‐diaphragm pre‐UDD (top) and absence of activity post‐UDD (bottom). Data are from the same mouse and confirm successful denervation. Note that both before and after UDD the telemetry device recorded bleed‐through EMG signal from the myocardium. (B) Diaphragm right costal weight normalized to tibia length (TL) from Sham and UDD operated WT mice. UDD induces a strong increase in tissue mass (TM) in the right costal hemidiaphram 6 days post‐UDD. (C) B‐mode image from the right hemidiaphragm, the vena cava was used as reference. The green dotted line denotes the position of strain measurement in the right costal diaphragm. Inset shows an abdominal view of the diaphragm and the black line the plane of B‐mode imaging. (D) Strain analysis showing maximum strain in the contracting right hemidiaphragm (top image; shortening) in Sham mice and stretching in the denervated hemidiaphragm (bottom image; lengthening) in UDD mice. Zero percent represents length at end‐expiration. (E) Maximum strain in the right hemidiaphragm before and after UDD, indicating that after UDD the hemidiaphragm stretches by ~25% during inspiration. (F) Representative M‐mode image of the right costal diaphragm during whole body formaldehyde perfusion, showing arrest during end‐expiration. (G) Sarcomere length in the perfused diaphragm fibres was ~2.9 μm (middle bar), and the strain data (in e) were used to calculate end‐inspiration sarcomere length before (left bar) and after UDD (right bar). (H) Schematic representation of the length changes in the diaphragm after UDD. Bars, mean ± SEM.

Figure 2

(A) M‐mode image of the left hemidiaphragm during normal inspiration. (B) Using the average inspiration time and the sarcomere length change during inspiration (0.8 μm), contractile velocity of the left hemidiaphragm was determined. (C) We assumed that after UDD the contractile velocity of the left hemidiaphragm equals the stretch velocity of the right hemidiaphragm, and used the derived value (in B) in a stretch‐release protocol imposed on single diaphragm fibres to mimic the kinetics of in vivo stretching of the right hemidiaphragm after UDD. Top: the length change imposed on the muscle fibre during the protocol of 20 stretches; middle: the corresponding sarcomere length change from 2.9– to 3.7 μm; bottom: passive force generated by the muscle fibres (data from the 20th cycle, to account for hysteresis, was used to determine passive tension. (D) Average passive tension (n = 6; 20 fibres) during the 20th stretch. Note that passive tension during this in vivo dynamic protocol is higher compared to that during a static stepwise fibre stretch protocol (grey curve, n = 12; 54 fibres); for protocol comparison see Supporting Information, Figure S1.

Figure 3

(A) Domain structure of titin in skeletal muscle. Immunoglobulin (Ig)‐like domains in red; fibronectin type III (Fn) domains in white; PEVK in yellow; titin kinase in black; unique sequence in blue. The exons targeted for insertion in RBM20^ΔRRM model (orange exons, approximation of RBM20 splice targets; underlined blue) and deletion in the Ttn^ΔIAjxn mouse model (underlined red). (B) These deletions result in a difference in titin (Ttn) migration, relative to nebulin (neb) on SDS‐AGE (note that Ttn^ΔIAjxn sample was not co‐migrated next to WT sample, hence the black line separation, however they were run on the same gel. For titin size estimation, see text. (C) Maximum strain in the denervated hemidiaphragm was not different between RBM20^ΔRRM, Ttn^ΔIAjxn, and wt mice after UDD. (D) Similarly, respiration rate (i.E. strain rate) in the denervated hemidiaphragm was not different between conscious RBM20^ΔRRM, Ttn^ΔIAjxn, and wt mice after UDD. Tidal and minute volume were also unaffected by the genotype, see Supporting Information, Figure S2. (E) Passive tension, as determined by the stretch‐release protocol, was significantly lower in the RBM20^ΔRRM mice (n = 4; 13 fibres) and higher in the Ttn^ΔIAjxn (n = 4; 13 fibres) mice compared to WT mice (n = 6; 19 fibres). (F) the increase in diaphragm right costal (denervated) tissue mass (TM) after UDD is attenuated in the RBM20^ΔRRM mice and exaggerated in the Ttn^ΔIAjxn mice (n = Sham; UDD). Graphs represent mean ± SEM.

Figure 4

(A) Fibre type composition before and after UDD. Note that UDD induces a shift toward slower fibre types (WT: n = 21 mice; RBM20^ΔRRM: n = 14; Ttn^ΔIAjxn: n = 11); (B) representative cross sections of right costal diaphragm in Sham and UDD mice. Red = ɑ‐laminin, blue = ɑ‐type 2x fibres, green = ɑ‐type 1 & 2a fibres. (C) Changes (%) in MinFeret after UDD in WT (n = 20), RBM20^ΔRRM (n = 13) and Ttn^ΔIAjxn (n = 9) mice compared to Sham mice. Note the differential response in the two titin models. (D) Effect of UDD on the number of sarcomeres in series; top right: representative fibre segment used for assay. (E) Schematic illustration showing the nature of the diaphragm fibre hypertrophy after UDD. (F) The cumulative hypertrophy (longitudinal combined with cross‐sectional hypertrophy) accounts for the mass increase observed in Figure 3F. Graphs represent mean ± SEM.

Figure 5

(A) Left panel: Several titin‐binding proteins are upregulated after UDD (grey bars; levels normalized to GAPDH levels; n = 15–18). Middle panel: MARP1 and CAPN3 are differentially upregulated in the two titin models, with exaggerated upregulation in the RBM20^ΔRRM (blue bars) and attenuated upregulation in the Ttn^ΔIAjxn model (red bars); n = 9 for the titin models. Right panel: Representative images of the corresponding western blots. We also probed for FHL2, MuRF2 and MYPN, but these were undetectable in both Sham and UDD samples. (B) Upregulation of z‐disc proteins after UDD (left panel, grey bars; levels normalized to gapdh levels; n = 17). Note that none of these proteins was differentially upregulated in the two titin models (middle panel). Left panel shows the corresponding western blots.

Figure 6

Cross sections stained for MARP1 (red), BF‐35 (green; MyHC2x exclusion) and DAPI (blue), showing absence of MARP1 signal in Sham‐operated mice (A; n = 4) and myofilament/cytosol localization in mice after UDD (B; n = 4). No MARP1 nuclear localization was observed in Sham‐operated mice (C) and in mice after UDD (D); plot profiles of representative images illustrate the absence of MARP1 in nuclei in Sham‐operated mice (E) and in mice after UDD (F). Using SIM, we determined that after UDD MARP1 (red) localizes to the N2A‐segment of titin (blue), visualized as doublets flanking the z‐disc (α‐actinin; green). Representative images of Sham‐operated mice (G–J) and mice after UDD (I‐O), with plot profiles [(K) and (P)] showing the co‐localization of MARP1 and N2A (n = 3).

Figure 7

Titin‐based mechanosensing model for muscle trophicity, wherein hypertrophy signalling is induced by muscle stretch and regulated through titin‐binding proteins in the I‐band region. TK denotes kinase‐domain.