Tamoxifen therapy in a murine model of myotubular myopathy

Abstract

Myotubular myopathy (MTM) is a severe X-linked disease without existing therapies. Here, we show that tamoxifen ameliorates MTM-related histopathological and functional abnormalities in mice, and nearly doubles survival. The beneficial effects of tamoxifen are mediated primarily via estrogen receptor signaling, as demonstrated through in vitro studies and in vivo phenotypic rescue with estradiol. RNA sequencing and protein expression analyses revealed that rescue is mediated in part through post-transcriptional reduction of dynamin-2, a known MTM modifier. These findings demonstrate an unexpected ability of tamoxifen to improve the murine MTM phenotype, providing preclinical evidence to support clinical translation.

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1

Tamoxifen treatment prolongs survival and improves muscle strength in Mtm1 knockout mice. a High-dose Tamoxifen (TAM) treatment starting at 21 days promotes significant improvement in Mtm1 KO survival [median survival 48 days (n = 20) vs. 39 days for untreated KOs (n = 7), *p = 0.05]. High-dose treatment at 30 days does not improve survival in aggregate but does in the subset of treated animals surviving beyond the first 2 days of treatment [subset median survival 66 days (n = 5) vs. 41 days for untreated KOs (n = 3), *p = 0.03]. b Low-dose TAM treatment starting 14, 21, or 30 days significantly improves Mtm1 KO survival. Low-dose TAM treatment starting 14 days has a median survival of 71 days (n = 9) vs. 42 days for untreated KOs, ***p = 0.001. Low-dose TAM treatment starting 21 days median survival is 57 days (n = 14) vs. 42 days for untreated KOs (n = 5), ****p = 3.1 × 10−5. Low-dose TAM treatment starting 30 days has a median survival of 54 days (n = 8) vs. 44 days for untreated KOs (n = 4), **p = 0.004 c. Photomicrograph of WT, Mtm1 KO, and high-dose TAM treated WT and KOs. Untreated KOs have splayed hind limbs (arrowhead indicates limb weakness). High-dose TAM treated KOs have normal limb positioning and posture. d Grip strength was measured at 35 days of age. High-dose TAM treated KOs (n = 6) have a mean grip strength of 107 ± 8%, vs. 65 ± 8% for untreated KOs (n = 4), ****p < 0.0001; high-dose TAM treated WT having 118 ± 7% (n = 7) vs. 100 ± 8% (n = 7) for untreated WT. Low-dose TAM KOs (n = 9) mean grip strength = 112 ± 8.9%, vs. 66 ± 10% for untreated KOs (n = 5), ****p < 0.0001, while treated WT = 116 ± 9% (n = 10), vs. 100 ± 10% for untreated WT (n = 6). All values normalized to untreated WT. Statistical analysis by log-rank (Mantel-Cox) test, two-way ANOVA (Tukey’s multiple comparisons/Fisher’s least common differences post-test) or unpaired Student's t-test

Fig. 2

Tamoxifen treatment improves Mtm1 knockout mouse muscle structure. a Tibialis anterior muscle stained with hematoxylin and eosin (H&E) and succinate dehydrogenase (SDH), and by anti-dysferlin immunofluorescence (IF). Untreated Mtm1 KO muscle has centrally located nuclei (arrows in H&E), mitochondrial aggregation on SDH staining, and abnormal distribution of dysferlin staining. High-dose TAM treatment in KOs results in: (1) reduction of central nuclei (arrowheads); (2) improvement in myofiber size; (3) resolution of mitochondrial aggregation; and (4) restoration of dysferlin to the sarcolemmal membrane. Low-dose TAM improves dysferlin localization but does not resolve central nucleation and myofiber hypotrophy. All drug treatments started at 21 days, and histopathology performed at 36 days of age. b High-dose TAM treatment increases myofiber size of Mtm1 KOs. TAM treated WT (29 ± 3 μm, n = 4) vs. untreated Mtm1 KO (19 ± 3 μm, n = 4, ****p < 0.0001 vs. WT), vs. TAM treated KO (28 ± 3 μm, n = 6, **p = 0.0065 vs. untreated KO, not significant vs. WT + TAM). c High-dose TAM treatment reduces % central nuclei in Mtm1 KOs (per 100 fibers): TAM treated WT (0%, n = 4) vs. untreated Mtm1 KOs (3 ± 0.8%, n = 4, ***p = 0.0004 vs. WT), vs. TAM treated KOs (1.1 ± 0.09%, n = 5, **p = 0.0024 vs. KOs and n.s. vs. WT + TAM). d High and low-dose TAM treatment restores dysferlin localization to cell membrane. Untreated Mtm1 KO = 60 ± 13% cytoplasmic (n = 3, ***p = 0.0002 vs. WT), KO + high TAM = 9 ± 1% cytoplasmic (n = 3, ***p = 0.0005 vs. KO, ns compared to WT + TAM). For low-dose TAM, Mtm1 KO = 66 ± 11% cytoplasmic, (n = 3, ****p = < 0.0001 vs. WT) vs. KO + low TAM = 12 ± 4% cytoplasmic (n = 4, ****p  < 0.0001 vs. KO, ns compared to WT + TAM). All data points for b–d presented in Supplementary Table 2. Statistical analyses were conducted by two-way ANOVA, followed by Tukey’s multiple comparisons post-test or Fisher’s Least Common Differences post-test. For direct two sample comparison, unpaired, parametric two-tailed Student's t-test was performed. Scale bars = 20 µm

Fig. 3

Early low-dose tamoxifen treatment prolongs survival and improves muscle strength and structure of Mtm1 KO mice. a Photomicrograph of littermates at 28 days of WT, and Mtm1 KO treated with low-dose tamoxifen starting 14 days of age. KO treated with low-dose tamoxifen have normal hindlimb positioning and appear similar in size to WT littermates. b Plots of body weight show that low-dose treated KOs (n = 5) starting at 14 days have similar weight as compared to low-dose treated WT (n = 3) and untreated KOs. c Early TAM treated KOs have improvement in muscle strength when analyzed at 36 and 50 days of age. At 36 days, TAM treated KOs (n = 8) have a mean grip strength of 107 ± 4% vs. TAM treated WT of 100 ± 2% (n = 6, p-value non-significant). At 50 days, TAM treated KOs (n = 4) have a mean grip strength of 95 ± 6% vs. TAM treated WT of 100 ± 4% (n = 3, p-value non-significant). Data normalized to the mean grip strength of WT + TAM littermates. d Early TAM treatment restores Mtm1 KO muscle structure. Cross-sections from tibialis anterior muscle tissue taken at 36 days and stained for H&E, SDH, and dysferlin (scale bars 20 μm). Early TAM treatment improves overall appearance and myofiber size but not % central nuclei. e Myofiber size for KO + TAM = 24 ± 3 μm (n = 3) vs. WT + TAM = 30 ± 4 μm (n = 3, p non-significant). f Average percent of central nuclei (per 100 fibers) in KO + low-dose TAM = 2 ± 0.4% vs. WT + low-dose TAM = 0.2 ± 0.1% (n = 3, *p = 0.03). g Electron microscopy reveals increased triad number in early treated Mtm1 KOs vs. untreated KOs (scale bars 500 nm). However, triad number is decreases as compared to early treated WT. h Quantification of number of triads per field for WT + low-TAM = 19 ± 1.0 (n = 3) vs. KO + low TAM = 10 ± 2.0 (n = 3, *p < 0.02 compared to WT + TAM). For reference, untreated Mtm1 KOs have 5 ± 1.5 (n = 5) (Fig. 3b)

Fig. 4

Tamoxifen treatment restores triad structure and function in Mtm1 KO mice. a Electron microscopy reveals normal triad structure in WT (arrowheads), reduction of triads in Mtm1 KOs (arrows), and restoration of triads in high-dose TAM treated KOs (arrowheads). Triad number was not improved in low-dose TAM treated KOs (arrows). b Quantification of triad number per field for WT = 17 ± 2 (n = 5), KO = 5 ± 1.5 (n = 5, ****p <0.0001 compared to WT), high-dose TAM treated WT = 14 ± 2 (n = 4), and high-dose TAM treated KOs = 11 ± 2 (n = 5, *p = 0.03 vs. KOs and non-significant vs. WT + TAM). Triad number for low-dose TAM treated KOs = 9.4 ± 5 (n = 4), untreated KOs = 4.8 ± 3 (n = 3, *p < 0.05 compared to WT), WT = 18 ± 5 (n = 3) and WT + low-dose TAM = 20 ± 5 (n = 3). c Calcium transients are increased in dissociated muscle fibers from Mtm1 KO mice treated with high-dose TAM in response to electrically evoked twitch and tetanus stimulation. Representative Mag-Fluo-4 fluorescence traces of FDB fibers during five successive single electrically evoked (Twitch) stimuli and Tetanus (500 ms at 100 Hz) stimulation from either Mtm1 KO, or KO treated with high-dose TAM. d Twitch stimulation is significantly reduced in Mtm1 KOs (0.21 ± 0.01, n = 51 fibers, ****p < 0.0001 vs. WT = 0.38 ± 0.08, n = 73 fibers) and improved with high-dose TAM (0.30 ± 0.01, n = 102 fibers, ****p < 0.0001 vs. KO). e Tetanus stimulation is significantly reduced in KOs compared to WT and improved with high-dose TAM (KO + TAM = 0.60 ± 0.03, n = 103 fibers, ***p < 0.001 vs. untreated KO = 0.40 ± 0.05, n = 51 fibers, ****p < 0.0001 vs. WT = 0.8 ± 0.04, n = 73 fibers, and vs. WT + TAM = 0.70 ± 0.05, n = 53 fibers). Statistical analyses were conducted by two-way ANOVA, followed by Tukey’s multiple comparisons post-test or Fisher’s least common differences post-test. For direct two sample comparisons, unpaired, parametric two-tailed Student's t-test was performed. Scale bars = 500 nm

Fig. 5

Estradiol, but not fulvestrant, improves survival in Mtm1 KO mice. a 17β-estradiol treatment starting at 21 days (1 µg/day) improves survival from 42 days for KOs to 56 days (n = 5, *p < 0.05), with longest survival to 59 days. b Fulvestrant treatment starting at 21 days (1 mg/kg, n = 5) does not improve survival of Mtm1 KOs, with median survival of 41 days for both KO + fulvestrant (n = 5) and KO (n = 4) alone. cMtm1 KO body weight is not restored with either estradiol or fulvestrant. d 17β-estradiol treatment of Mtm1 KOs does not improve muscle strength. Estradiol treated KOs (n = 5) mean grip strength is reduced to 69 ± 6%, untreated KOs to 74 ± 7% (n = 3) (**p = 0.0020 compared to WT), while estradiol treated WT are 93 ± 7% (n = 5). Values are normalized to WT + DMSO (100 ± 6%, n = 5). Fulvestrant treatment of KOs also does not improve muscle strength. Fulvestrant treated KOs (n = 4) mean grip strength = 75 ± 9%, untreated KOs = 56 ± 8% (n = 3) (**p = 0.0021 compared to WT), fulvestrant treated WT = 101 ± 8% (n = 4). Values normalized to WT + miglyol (100 ± 7%, n = 2). e Estradiol and fulvestrant treatment do not improve Mtm1 KO muscle structure. As with untreated KOs, muscle from KOs treated with either estradiol or fulvestrant show decreased myofiber size, increased mitochondrial aggregation, and abnormal distribution of dysferlin staining. f–i Quantification of muscle structural parameters. There was no significant change in Mtm1 KO myofiber size with estradiol or fulvestrant treatment f, and no significant change in dysferlin distribution g or triad number with estradiol treatment. The full enumeration of all data points is presented in Supplementary Table 3. Statistical analyses were conducted by two-way ANOVA, followed by Tukey’s multiple comparisons post-test or Fisher’s least common differences post-test. For direct two sample comparison, unpaired, parametric two-tailed Student's t-test was performed. Scale bars = 20 µm

Fig. 6

Tamoxifen treatment does not alter the Mtm1 knockout transcriptome. a–c Representative western blots for ERα (with β-actin loading control) for WT, WT + TAM, Mtm1 KO, and tamoxifen (TAM) treated Mtm1 KO mice (molecular weight markers indicated in kDa). d–f Quantification of protein levels of the 55 kDa isoform of ERα as determined by densitometry with standardization to β-actin and represented as fold difference from the average of WT (n = 5 mice per group for high-dose TAM, n = 4 mice per group for low-dose TAM and estradiol. Statistical analyses were conducted by two-way ANOVA, followed by Tukey’s multiple comparisons post-test *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Complete data points presented in Supplementary Table 4. g Heat map indicating overall mRNA expression profiles (derived from RNA-seq) of the four experimental groups, blue = high expression, yellow = low expression. Mtm1 KO and Mtm1 KO + TAM heat maps cluster together and are distinct from WT and WT + TAM. h Table indicating the total number of gene transcripts in each comparison with significantly differentially expression (n = 3 mice per group for WT, n = 4 for WT + TAM, n = 3 for KO, and n = 4 for KO + TAM). Note that only 29 genes are differentially expressed (p ≤ 0.01, DEseq and EdgeR R/Bioconductor pairwise comparisons) in Mtm1 KO vs. Mtm1 KO + TAM. i Box plot indicating the average expression of 21 estrogen receptor-responsive genes within each experimental group; no overall differences are observed with TAM treatment. j Individual expression box plots of Esr1 and Pgr1, the only significantly changed estrogen response genes, and of Mtm1. Box plots show the median, the 25th and 75th percentiles, and outliers. k Immunofluorescence with anti-ERα (red; indicated by a white arrow) on tibialis anterior muscle sections from WT, WT + TAM, Mtm1 KO, and Mtm1 KO + TAM (blue = DAPI) (n = 3 mice per group for WT and WT + TAM, n = 4 mice per group for KO and KO + TAM). ERα is expressed primarily in the nucleus in WT muscle (arrow) but found at the sarcolemmal membrane in Mtm1 KOs (arrow). Overall ERα expression and its membrane localization are reduced with TAM treatment. Scale bars = 20 µm

Fig. 7

Tamoxifen treatment does not affect the levels or activity of PIK3C2B. a Representative western blot for PIK3C2B, with HSP90 loading control, in Mtm1 KO mice treated with high-dose tamoxifen (TAM), b low-dose TAM, and c 17β-estradiol; position of the molecular weight markers indicated (in kDa). Samples for each drug trial derive from the same experiment, and blots were processed in parallel for each drug trial. d–f PIK3C2B levels were determined by densitometry, standardized to HSP90 and are represented as the fold difference from the average of the WT (n = 5 mice per group for high-dose TAM, n = 3 for low-dose TAM, n = 4 for 17β-estradiol); graphs represent, 5, 3, and 4 technical replicates respectively and mean ± SEM). g, h Graphs depicting mRNA levels of PIK3C2B in g gastrocnemius and h hamstring of WT and Mtm1 KO mice ± tamoxifen (TAM) treatment. mRNA expression was determined by real-time qPCR and analyzed by the 2−ΔΔCT method. Values were normalized to actin and represent the fold difference from the average of the WT (n = 3 mice per group for gastrocnemius; n = 6 mice per group for hamstring, graphs represent three technical and biological replicates for gastrocnemius; three technical and biological replicates for hamstring and ±SEM). Statistical analyses for a–h were conducted by two-way ANOVA followed by Tukey’s multiple comparisons post-test. No significant differences were observed. i Immunohistochemistry against PI3P in WT and Mtm1 KO mice, and TAM treated WT and Mtm1 KO mice show no overall difference in PI3P levels following treatment with high-dose TAM. j, k Graphs of PIK3C2B in vitro kinase assay depicting the percent inhibition of PIK3C2B by wortmannin (control; IC50 = 500 nM) and tamoxifen (IC50 = 40 µM). Scale bars = 20 µm

Fig. 8

Tamoxifen reduces dynamin-2 protein level in vivo and in vitro. a–d Representative western blots for dynamin-2 (DNM2) (β−actin loading control) for Mtm1 KO mice ± tamoxifen (TAM). a, b 17β-estradiol c and fulvestrant d. e–h DNM2 protein levels are increased in Mtm1 KO and reduced by TAM and estradiol treatments. Quantification determined by densitometry (standardized to β-actin) and represented as fold difference from WT (n = 4 mice per group for high and low-dose TAM; n = 3 for 17β-estradiol and fulvestrant); graphs represent 4 and 3 technical replicates respectively, and mean ± SEM). i–q Representative western blots for DNM2 in various cell culture models, with accompanying quantification (presented as fold difference vs. average for untreated cells). i–l DNM2 is increased in human MTM patient fibroblasts (n = 3 cell types per group for control and MTM) and transdifferentiated myotubes (n = 3 cell types per group for control and n = 2 for MTM) and decreased with TAM; graphs represent three technical replicates and mean ± SEM. k, l DNM2 can be reduced with TAM in C2C12 myotubes m, o and MCF-7 cells q, s both of which express ERα (Supplementary Figure 12A), but not in HEK293T cells n, p which do not (Supplementary Figure 12B); graphs represent three technical replicates and mean ± SEM. r Representative western blot for DNM2 for MCF-7 cells treated with 10 µM MG-132 and 30 nM Bortezomib (BZ) ±10 µM TAM; graphs represent 6–15 technical replicates and mean ± SEM. t Both UPS inhibitors block TAM’s effect on DNM2, as revealed by densitometric quantification (values standardized to β-actin and represented as the fold difference from the average of untreated control). Statistical analyses were conducted by two-way ANOVA for mouse experiments, and one-way ANOVA for in vitro experiments, followed by Tukey’s multiple comparisons post test. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001. Of note, a full listing of all data points for a–r is presented in Supplementary Table 5. u Schematic representing proposed mechanism of action of TAM in Mtm1 KO muscle