Muscle pathology from stochastic low level DUX4 expression in an FSHD mouse model

Abstract

Facioscapulohumeral muscular dystrophy is a slowly progressive but devastating myopathy caused by loss of repression of the transcription factor DUX4; however, DUX4 expression is very low, and protein has not been detected directly in patient biopsies. Efforts to model DUX4 myopathy in mice have foundered either in being too severe, or in lacking muscle phenotypes. Here we show that the endogenous facioscapulohumeral muscular dystrophy-specific DUX4 polyadenylation signal is surprisingly inefficient, and use this finding to develop an facioscapulohumeral muscular dystrophy mouse model with muscle-specific doxycycline-regulated DUX4 expression. Very low expression levels, resulting in infrequent DUX4 + myonuclei, evoke a slow progressive degenerative myopathy. The degenerative process involves inflammation and a remarkable expansion in the fibroadipogenic progenitor compartment, leading to fibrosis. These animals also show high frequency hearing deficits and impaired skeletal muscle regeneration after injury. This mouse model will facilitate in vivo testing of therapeutics, and suggests the involvement of fibroadipogenic progenitors in facioscapulohumeral muscular dystrophy.Facioscapulohumeral muscular dystrophy is a severe myopathy that is caused by abnormal activation of DUX4, and for which a suitable mouse model does not exist. Here, the authors generate a novel mouse model with titratable expression of DUX4, and show that it recapitulates several features of the human pathology.

Fig. 1

DUX4 poly A signal read through and the effect of removing the SV40 pA signal. a Schematic of transgene integration in iDUX4[2.7] mice and primers/probe used for detection of transcript proximal to and distal of the DUX4 pA signal. Note the SV40 poly A signal downstream of the DUX4 poly A signal. b RTqPCR with reverse primers on either side of the DUX4 pA signal in PDGFRα⁺ integrin α7^– (fibroadiopgenic) and PDGFRα^– integrin α7⁺ (myogenic) sorted cultured primary muscle cells. n = 3 technical replicates. Note the large amount of transcript that is detected by the primer downstream of the DUX4 pA, indicating that the DUX4 poly A is not efficient and allows a high level of read through. In the iDUX[2.7] mouse, these read through transcripts are stabilized by the SV40 poly A, potentiating higher basal levels of DUX4 in the off state. c The transgene is targeted 5ʹ of HPRT, is regulated by a second generation tetracycline-response element, includes DUX4 and 3ʹ UTR including the DUX4 poly A signal. d Mice were made in which the SV40 pA was removed, leaving only the endogenous, less efficient, DUX4 pA. Viability is now near-normal for male carriers. Summary of 174 genotypes of 3-week old mice from 21 litters of iDUX4pA females backcrossed to males with a WT chrX. Results are not significantly different from expected for the hypothesis of no selection against iDUX4pA ( χ², two tailed p = 0.2956), but are strongly diverged from expected (p < 0.0001) for the hypothesis that iDUX4pA is selected against at the same level as observed against iDUX4[2.7]. Thus, we conclude that the presence of the DUX4 poly A makes the transgene less toxic, with viability to 3 weeks not different from WT in the absence of dox

Fig. 2

Phenotypes of iDUX4pA mice in the basal “off” state. a Weekly average of the weight of iDUX4 and wild type (WT) male and female siblings (n = 3–25). b Normalized mass to the body weight (BW) of various muscles in iDUX4pA males. Extensor digitorum longus (EDL), tibialis anterior (TA), gastrocnemius and soleus (G/S), quadriceps (Quad), triceps brachii (Triceps), pectoralis, psoas major (Psoas), and masseter from 10 to 15 week old mice and age-matched siblings are analyzed (n = 10). c Representative example of hanging wire, grip strength and rotarod analyses of 6-week-old iDUX4pA male and female mice (n = 4). d Specific force, time to reach peak twitch force (TPT) and time to reach 1/2 relaxation during twitch (1/2 TPT) of EDL in 6–10-week-old iDUX4pA males and age-matched siblings (n = 8). e Representative FACS analyses for Lin^neg (CD45^neg; CD31^neg) and PDGFRα⁺ or Sca1⁺ cells in muscle from iDUX4pA males and sibling controls. Percent of PDGFRα⁺ or Sca1⁺ cells in pooled muscle digests comprising TA, gastrocnemius, soleus, quadriceps, pectoralis, and triceps is shown at right (n = 7). f RTqPCR analyses for DUX4, DUX4 targets, myogenic genes, and genes involved in fibrosis in pectoralis from iDUX4pA male mice. Results are presented as fold difference to GAPDH (n = 4). g Auditory brainstem response (ABR) in iDUX4 males, females and WT sibling controls (n = 5, 6–8 week old mice). Note that iDUX4pA males do not respond to sounds of 16 kHz and above and iDUX4pA females to sounds of 32 kHz. All data are presented as mean±SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by T-test or g two-way ANOVA followed by Tukey’s post hoc test

Fig. 3

DUX4 induces skeletal muscle atrophy. a Mass of various muscles normalized to the body weight (BW) of iDUX4pA male mice treated with doxycycline over doses and time ranges (n = 3). b Specific isometric, concentric and eccentric forces and 1/2 relaxation time (1/2 TPT) of EDL in 10-week-old iDUX4pA male mice induced with 5 mg/kg doxycycline for 28 days (n = 4). c Immunostaining for DUX4 and laminin on quadriceps of iDUX4pA;HSA-rtTA males induced with 5 and 100 mg/kg doxycycline for 14 days. DAPI was used for nuclear staining. d RTqPCR analyses for DUX4 and DUX4 target genes Myo1g, Wfdc3, Azi2, Cdkn1a, and Zscan4c in pectoralis muscle and liver of iDUX4pA;HSA-rtTA male mice induced with 25 mg/kg doxycycline for 14 days. Results are presented as fold difference to GAPDH (n = 3). e Hematoxylin and Eosin (H&E) staining and Sirius red/fast green staining of quadriceps from iDUX4pA;HSA-rtTA male mice induced with 5 and 100 mg/kg doxycycline for 14 days (left). Quantification of fibrous tissue in control and induced (100 mg/kg) quadriceps (n = 3, right). f RTqPCR for indicated fibrogenic markers in iDUX4pA;HSA-rtTA males mice induced with 5 mg/kg doxycycline for 14 days. Results presented as fold difference to GAPDH (n = 3). g Representative FACS analyses for hematopoietic inflammatory cells expressing CD45 and CD68, CD206 or Ly6G in muscle of iDUX4pA;HAS-rtTA males or control siblings (WT or lacking rtTA) induced with 25 mg/kg doxycycline for 14 days (left), and summary data (right, n = 4). Data are presented as mean±SEM; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 by T-test except a, b two-way ANOVA or g one-way ANOVA, with Tukey’s post hoc test

Fig. 4

DUX4 impairs recovery after injury. a Hematoxylin and eosin (H&E), Sirius red/fast green and immunostaining for DUX4, Laminin and DAPI nuclear staining of uninjured and injured TA muscle sections from iDUX4pA;HSA-rtTA males and controls lacking rtTA 12 days post injury. Mice were induced with 25 mg/kg doxycycline for 12 days. b Gross morphology of injured/uninjured TA muscles with/without DUX4 induction 12 days post injury. c Ratio of weight of post injury vs. uninjured contralateral TA from iDUX4pA;HSA-rtTA males, and siblings lacking rtTA or fully WT 12 days post injury (n = 3). d Ratio of post injury vs. uninjured contralateral TA from iDUX4pA;HSA-rtTA male and female mice 12 days post injury. Mice were induced with 25 mg/kg doxycycline (n = 3). Data are presented as mean±SEM; **p < 0.01, ****p < 0.0001 by c one-way ANOVA with Tukey’s post hoc test or d T-test