Actin nemaline myopathy mouse reproduces disease, suggests other actin disease phenotypes and provides cautionary note on muscle transgene expression

Abstract

Mutations in the skeletal muscle α-actin gene (ACTA1) cause congenital myopathies including nemaline myopathy, actin aggregate myopathy and rod-core disease. The majority of patients with ACTA1 mutations have severe hypotonia and do not survive beyond the age of one. A transgenic mouse model was generated expressing an autosomal dominant mutant (D286G) of ACTA1 (identified in a severe nemaline myopathy patient) fused with EGFP. Nemaline bodies were observed in multiple skeletal muscles, with serial sections showing these correlated to aggregates of the mutant skeletal muscle α-actin-EGFP. Isolated extensor digitorum longus and soleus muscles were significantly weaker than wild-type (WT) muscle at 4 weeks of age, coinciding with the peak in structural lesions. These 4 week-old mice were ~30% less active on voluntary running wheels than WT mice. The α-actin-EGFP protein clearly demonstrated that the transgene was expressed equally in all myosin heavy chain (MHC) fibre types during the early postnatal period, but subsequently became largely confined to MHCIIB fibres. Ringbinden fibres, internal nuclei and myofibrillar myopathy pathologies, not typical features in nemaline myopathy or patients with ACTA1 mutations, were frequently observed. Ringbinden were found in fast fibre predominant muscles of adult mice and were exclusively MHCIIB-positive fibres. Thus, this mouse model presents a reliable model for the investigation of the pathobiology of nemaline body formation and muscle weakness and for evaluation of potential therapeutic interventions. The occurrence of core-like regions, internal nuclei and ringbinden will allow analysis of the mechanisms underlying these lesions. The occurrence of ringbinden and features of myofibrillar myopathy in this mouse model of ACTA1 disease suggests that patients with these pathologies and no genetic explanation should be screened for ACTA1 mutations.

Figure 1. Nemaline bodies were a prominent feature of Tg(ACTA1)^D286G-EGFP muscle.

(A) Gomori trichrome staining revealed that nemaline bodies (see insert) were a prominent feature of Tg(ACTA1)^D286G-EGFP muscles, especially at 1 month of age. Serial sections showed that the location of these nemaline bodies corresponded with regions of intense EGFP aggregates. Note also the presence of internal nuclei, the large variation in fibre size in muscle of Tg(ACTA1)^D286G-EGFP mice and the absence of EGFP signal in spindle fibres (arrows). These first two characteristics were not observed in WT muscle (Figure 1.E: WT EDL muscle stained with Gomori trichrome). ACTA1(D286G)-EGFP α-actin was incorporated into the sarcomeres as seen by the striated EGFP signal but also formed intensely fluorescing aggregates in quadriceps muscles (B, C). The EGFP-positive aggregates labeled with antibodies to the Z-disk proteins α-actinin (B) and myotilin (C). (D) Electron-dense nemaline bodies were observed as extensions of the Z-disk. These were most frequently observed in the SOL muscle of 1-month old and the EDL and gastrocnemius muscles of 4-month or older Tg(ACTA1)^D286G-EGFP mice. Scale bars  =  50 µm (A, E), 25 µm (B, C) and 2 µm (D).

Figure 2. Tg(ACTA1) ^D286G-EGFP mice were weak and less active at 1 month of age than WT control mice.

(A) sP_o was significantly lower in EDL (n = 9, p = 0.026) and SOL muscles (n = 10, p<0.0001) of Tg(ACTA1)^D286G-EGFP mice compared to that of WT muscle (EDL: n = 11, SOL: n = 8). (B) In the SOL muscles from Tg(ACTA1)^D286G-EGFP mice, the force–frequency relationship displayed a significant shift to the right compared to WT. (C) The force-frequency relationship was shifted to the left in the Tg(ACTA1)^D286G-EGFP EDL muscles compared to WT EDL (NOTE: for some data points the SEM bars are too small to be seen). (D) One-month old Tg(ACTA1)^D286G-EGFP mice (n = 8) used voluntary running wheels less than WT mice (n = 9), and traveled 30% less distance on average, per day. (E) Average and (F) maximum speeds were not different between Tg(ACTA1)^D286G-EGFP and WT mice. *p<0.05, **p = 0.005, ***p<0. 001.

Figure 3. Expression of the ACTA1(D288G)-EGFP fusion α-actin protein changes with MHC fibre type and was associated with an increased contribution of MHC hybrid fibres.

(A) A time-course of EGFP-expression in the gastrocnemius muscle indicated that during the early PN period all fibres expressed similar levels of ACTA1(D288G)-EGFP. By PN d7 expression was slightly reduced and by PN d25 a proportion of the fibres expressed comparatively low levels of ACTA1(D288G)-EGFP. (B) IHC of serial sections of SOL and EDL muscles from 1- and 4-month old Tg(ACTA1)^D286G-EGFP mice revealed that in the SOL, the fibres expressing the most ACTA1(D288G)-EGFP were MHCI and MHCI/IIA hybrids (1 month) or pure MHCI fibres (4 months). In the EDL, the EGFP signal was most intense in the MHCIIB fibres for both timepoints. By 4 months of age in the EDL, MHCIIA fibres were essentially negative for EGFP by comparison to MHCIIB and MHCIIX fibres. Scale bar  =  50 µm. Note the EGFP images were not all taken at the same exposure times for gastrocnemius, EDL and SOL muscles but were instead taken at the settings required to achieve sufficient EGFP signal to facilitate visualisation of the fibres.

Figure 4. The EGFP signal intensity varied between muscles and diminishes with age.

The EGFP signal imaged under the same exposure settings clearly demonstrated that the amount of mutant skeletal muscle α-actin-EGFP varied between different muscles, with intense EGFP aggregates most frequent in the EDL, gastrocnemius and SOL muscles at 1 month of age. By 4 months of age the intensity of the EGFP signal was reduced in all muscles compared to the 1-month time point. This was most noticeable in the SOL muscle. The diaphragm was largely EGFP-negative at both time-points. Scale bar  =  50 µm.

Figure 5. Fibre-type proportions and fibre size were significantly different in Tg(ACTA1)^D286G-EGFP muscle.

(A) At 1 month of age, Tg(ACTA1)^D286G-EGFP SOL muscle (n = 8) contained significantly less MHCIIA fibres and a concurrent increase in the percentage of MHCI/IIA hybrid fibres than in 1-month old WT SOL muscle (n = 9). This change persisted at 4 months, and in addition the Tg(ACTA1)^D286G-EGFP SOL (n = 5) also contained a higher percentage of MHCI fibres than WT SOL (n = 5). At 1 month of age a significant increase in the percentage of hybrid fibres was observed in Tg(ACTA1)^D286G-EGFP EDL muscle (n = 7) compared to WT EDL (n = 10). By 4 months of age the contribution of hybrid versus pure fibres was not significantly different for WT (n = 6) and Tg(ACTA1)^D286G-EGFP (n = 5) EDL muscle. (B) Morphometry of EDL and SOL muscle fibres at 1 and 4 months of age revealed that there is a trend towards all Tg(ACTA1)^D286G-EGFP fibres being smaller than WT. This feature was more striking at the 1-month time-point for both the EDL and SOL fibres. *p<0.05, **p<0.01, ***p<0.001 compared to WT.

Figure 6. Internally-nucleated fibres were a persistent feature of skeletal muscle from Tg(ACTA1)^D286G-EGFP mice post-weaning.

(A) Representative images taken of Hoechst-labeled EDL and SOL muscles from PN d14, 1 and 4-month old Tg(ACTA1)^D286G-EGFP mice showed the occurrence of internally-nucleated fibres in post-weaned Tg(ACTA1)^D286G-EGFP mice. A representative image of a WT EDL muscle labelled with phalloidin-FITC and Hoechst is also shown. (B) The percentage of internally-nucleated fibres in whole EDL (1m, 4m WT: n = 7, n = 6; Tg: n = 10, n = 5) and SOL (1m, 4m WT: n = 9, n = 5; Tg: n = 8, n = 5) muscles was determined for WT and Tg(ACTA1)^D286G-EGFP mice, with a significant increase seen in Tg(ACTA1)^D286G-EGFP muscle. However there was no significant difference in the number of internally nucleated fibres in EDL and SOL muscles (WT: n = 7, Tg: n = 4) of Tg(ACTA1)^D286G-EGFP mice at PN d14. *p<0.05 compared to WT, **p<0.01 compared to WT.

Figure 7. SOL muscle of 1-month old Tg(ACTA1)^D286G-EGFP mice displayed, in addition to nemaline bodies, a number of alterations at the ultrastructural level that are found in rod-core and myofibrillar myopathies.

(A) Aggregates of sarcoplasmic reticulum profiles punctuated by scattered autophagosomes, mitochondria and intermediate filaments. (B) An intermyofibrillar aggregate of fine filaments and electron dense rods of varying dimensions. (C) A collection of anastomosing, irregular electron dense rods into which fine filaments are inserting. The electron dense rods have a faintly fibrillar substructure. (D) Fibrillogranular electron dense material accumulating along the plane of the Z-bands. Note that the Z-bands are fragmented and there is disorganisation of the sarcomeric myofilaments. (E) Transverse view of electron dense fibrillogranular material. Labelling of SOL muscle sections from Tg(ACTA1)^D286G-EGFP mice revealed altered localisation and expression levels of desmin and αB-crystallin (G) compared to that seen for WT muscle (F). Desmin expression was more diffuse throughout the muscle fibres in Tg(ACTA1)^D286G-EGFP muscle compared to WT muscle. Fibres containing intense EGFP-positive aggregates also showed elevated levels of αB-crystallin compared to neighbouring fibres. Scale bars  =  2 µm (A-E), 20 µm (F–G).

Figure 8. Ringbinden fibres were a prominent feature of skeletal muscle from adult Tg(ACTA1)^D286G-EGFP mice.

(A) Confocal imaging of EDL muscle from a 4-month old Tg(ACTA1)^D286G-EGFP mouse showed the presence of numerous ringbinden fibres (arrows) that were best seen when labelled with an antibody to α–actinin. (B) Ringbinden fibres were most frequent in EDL and gastrocnemius muscles of adult Tg(ACTA1)^D286G-EGFP mice (≥ 4-months old) and were prominent by Gomori trichrome and toluidine blue staining (arrows). At the ultrastructural level, ringbinden fibres consisted of 2 to 7 peripheral myofibrils arranged concentrically around the central myofibres. (C) Fibre-typing of gastrocnemius muscle using DAB of consecutive sections, revealed that the ringbinden fibres were exclusively MHCIIB fibres.