A defect in myoblast fusion underlies Carey-Fineman-Ziter syndrome

Abstract

Multinucleate cellular syncytial formation is a hallmark of skeletal muscle differentiation. Myomaker, encoded by Mymk (Tmem8c), is a well-conserved plasma membrane protein required for myoblast fusion to form multinucleated myotubes in mouse, chick, and zebrafish. Here, we report that autosomal recessive mutations in MYMK (OMIM 615345) cause Carey-Fineman-Ziter syndrome in humans (CFZS; OMIM 254940) by reducing but not eliminating MYMK function. We characterize MYMK-CFZS as a congenital myopathy with marked facial weakness and additional clinical and pathologic features that distinguish it from other congenital neuromuscular syndromes. We show that a heterologous cell fusion assay in vitro and allelic complementation experiments in mymk knockdown and mymk^insT/insT zebrafish in vivo can differentiate between MYMK wild type, hypomorphic and null alleles. Collectively, these data establish that MYMK activity is necessary for normal muscle development and maintenance in humans, and expand the spectrum of congenital myopathies to include cell-cell fusion deficits.

Figure 1. MYMK-CFZS genetics.

(a) Pedigree schematics. Filled symbol=affected; arrow=proband. Mutation status is indicated under each enrolled individual: +=WT allele, M1–M5=mutant alleles as defined in b, red and blue fonts denote null and hypomorphic alleles, respectively. Individuals’ numbers (Ind. 1–8) are boxed in black. (b) Top: Human MYMK gene schematic. Arrows denote nucleotide position of M1-M5 mutations. Bottom: multispecies protein MacVector alignment of amino acids surrounding each substitution: dark grey, light grey, and white shading indicate conserved, partially conserved, and nonconserved residues, respectively. (c) 2D structure of the human 220 amino acid MYMK protein (previously referred to as the 7 transmembrane domain (TM) TMEM8C protein) (modified from Millay et. al.18), with starting and ending amino acid residues of each TM numbered. The red-filled residues denote the predicted locations of the CFZS amino acid substitutions. M4 is predicted to preclude initiation of wildtype translation; if translation were initiated instead by the next methionine codon (Met32), the first TM would be excluded from the protein. M2 and M3 alter residues buried deeply in TM7 and TM4, respectively. M3 alters Gly100, which is highly conserved in MYMK, TMEM8A and TMEM8B. M2 alters a less conserved Cys185, but introduces a charged arginine into a hydrophobic domain. This model locates M1 and M5 at the beginning of TM4 and five residues inside TM5, respectively, and both substitute a hydrophobic amino acid with threonine. The red ‘V’ in residue 106 in TM4 corresponds to the amino acid disrupted by the one-base-pair frameshift-causing insertion generated in the zebrafish model (mymk^insT). TM, transmembrane domain; M, mutation as per (b) followed by corresponding amino acid substitution. See also Supplementary Fig. 1.

Figure 2. Clinical and radiological and pathological features of MYMK-CFZS.

(a) Front (top) and profile (bottom) facial photos of individuals (Ind.) 1–8 highlight facial weakness with flattened nasolabial folds, thin elongated face, midface hypoplasia, prominent nose, micro/retrognathia, and in some, low-set ears (also lagophthalmos on attempted lid closure, Ind. 8, top). (b) Map of left thigh muscles: vastus lateralis (vl), black; vastus medialis (vm), orange; rectus femoris (rf), light green; sartorius (sa), yellow; adductor longus (al), dark green; adductor magnus (am), light blue; gracilis (g), pink; semitendinosus (st), grey; gluteus maximus (gm), purple. Right: left thigh control MRI with sa and am indicated by yellow and blue arrow, respectively. (c) Left thigh MRIs of Ind. 2, 3, 6, and 7 at ages 37, 19, 7.5, and 7 years, respectively, reveal variable muscle involvement with disproportionate atrophy and fatty infiltration of the sa (yellow arrows) and am (blue arrows). (d) Compared to control quadriceps (left) Ind. 2 quadriceps biopsy (right, vastus lateralis) shows variable fiber size (top, hematoxylin and eosin (H&E)), atrophic/hypotrophic type-I fibers (middle, slow myosin heavy chain (MHC), red), hypertrophic type-II fibers (bottom, fast MHC, red). Scattered type IIc fibers (slow and fast MHC double-positive) were present in the control (shown here) and Ind. 2. Laminin (green). Scale bar, 50 μm. (e) Relative type-I (dashed lines) and type-II (solid lines) myofiber size frequency in control (blue), Ind. 2 (red). (f–i) Ind. 2 versus averaged cohorts of adolescent and young adult male controls (vastus lateralis, age 16–28, n=383) revealed: (f) significantly increased type II fiber cross-sectional area of 17,077±4770 μm², which is 330% of measured (4,842±1,288 μm²) and reported controls and corresponding type II fiber minimum Feret diameters of 185% of controls (type I/type II: 60±31/122±21 μm versus 57±16/66±11 μm); (g) significantly reduced myofibers per mm², and (h) increased type-I fiber proportion (60%), but (i) only slightly decreased overall cross-sectional area composition of type I versus type II fibers (average fiber area × relative composition), consistent with a compensatory response. Significance of fiber size (cross-sectional area and minimum Feret diameter) and number and nuclei number were assessed with Student’s t-tests and analysis of variance (ANOVA). Mean±s.d.; ***P<0.0001. See also Supplemental Figs 2, 3.

Figure 3. CFZ mutations cause impaired fusion in ectopic expressing cell lines and patient derived cell lines.

(a–c) Fibroblast-myoblast cell-cell fusion experiments: (a) Schematic of heterologous cell–cell fusion system. Fibroblasts were co-infected with GFP plasmid and FLAG-tagged WT or CFZS-mutant MYMK construct, co-cultured with C2C12 myoblasts for 4 DIV, and immunostained with myosin antibody, a myocyte marker. If fibroblast-myoblast fusion occurs, chimeric myotubes appear yellow/orange from colocalization of GFP (green, fibroblast origin) and myosin (red, myoblast origin). (b) Fibroblasts co-infected with mouse or human WT MYMK-FLAG constructs and GFP fuse with C2C12 cells. Hypomorphic CFZ mutations M1 (p. P91T) and M5 (p. I154T) retain fusogenic activity, while fibroblasts infected with empty vector or null CFZ mutations M2-M4 do not. White arrows indicate GFP⁺ myosin⁺ yellow fused myotubes, whereas white arrowheads highlight unfused myosin⁺ myocytes. (c) Quantification of heterologous fusion between infected fibroblasts and C2C12 cells. The number of GFP-positive myotubes was counted and normalized by the total number of Myo32 positive myotubes. (d–g) CFZS primary myoblasts fusion experiments: (d) Differentiation index, calculated as number of nuclei in myosin positive cells divided by the total number of nuclei in all cells per field, shows no difference between control and Ind. 2 primary myoblast cell lines. (e) Quantifications of the number of nuclei in myosin⁺ cells reveal a significant increase in single nucleated cells and decrease in multinucleated cells in Ind. 2 derived cells compared to control. (f) The cell diameter of Ind. 2 myoblasts is significantly reduced compared to control myoblasts (n=45). (g) Image of control myoblasts (left) reveals greater width, more cytoplasm, and more nuclei per fiber (white arrow) than Ind. 2 myoblasts (right—white arrowhead) at 7 DIV post differentiation. Blue: DAPI stain of nuclei. Red: myosin heavy chain A4.1025 antibody. Analyses of fusion assay and differentiated myoblasts were each performed on at least 6 and 10 random fields per experiment, respectively, and 3 independent experiments. Statistical analyses in: (c) calculated by one way ANOVA with Bonferroni correction for multiple testing and shown as comparison to the empty vector (EV); (d–f) using two-tailed Student’s unpaired t test. Mean±s.e.m.; **P<0.001; ns, not significant. Scale bars in b,g, 50 μm. See also Supplementary Fig. 5.

Figure 4. mymk^insT/insT zebrafish embryos lack fast-twitch myoblast fusion and adults have CFZS-like myopathic features.

(a) 24 h.p.f. embryos stained with 488-conjugated phalloidin (white) and DAPI (red) to label the forming myofibers and nuclei, respectively. WT mymk^wt/wt embryo myofibers are multinucleated, while mymk^insT/insT embryo myofibers elongate and differentiate, but fail to fuse. (b) Upper panel: 48 h.p.f. embryos stained with DAPI show the distribution of myonuclei in fused mymk^wt/wt (left) and unfused (right) mymk^insT/insT fast-twitch myofibers. Compared to the random appearance of WT myonuclei, the single nuclei in unfused mymk^insT/insT myofibers pathologically align at the center, equidistant between two myosepta. Lower panel: Merged images showing phalloidin (red), DAPI (white) and F59 (green, anti-myosin heavy chain) confirm that affected myofibers are fast-twitch fibers because they do not express F59, a slow-twitch cellular marker in zebrafish. (c) A 6-month-old male mymk^wt/wt zebrafish (top) compared to an age- and sex-matched tmem8c^insT/insT fish (bottom). Male and female mutant zebrafish are small and have a flattened/retrognathic jaw (right, indicated by dotted line and black arrowhead) not appreciated during larval and early juvenile stages. By 3 months of age, jaw weakness prohibits mymk^insT/insT zebrafish from fully closing their mouths. (d,e) Adult mymk^insT/insT zebrafish (n=3) are significantly shorter (d) and weigh less (e) than age and sex-matched WT siblings (n=7). (f) Hematoxylin-Eosin (H&E) staining of caudal transverse sections of WT (top) and mymk^isnT/insT (bottom) of 6-month-old male zebrafish siblings at three magnifications. Zebrafish mymk-expressing fast-twitch myofibers are located centrally, while mymk-negative slow-twitch myofibers are located near the body wall and stain slightly darker with H&E. The mutant fish have reduced body width (compare the red line extending from the dorsal artery to the body wall in WT versus mutant fish in left photos), and fat infiltration (thick black arrows, middle photo) that is absent in the WT fish. Fast-twitch myofibers appear smaller compared to WT (right). Statistics by two-tailed Student’s unpaired t-test; mean±s.e.m.; **P<0.002. Scale bars, (a,b) main images 50 μm, insets 10 μm. (f) left and center panels 500 μm, right panels 50 μm. See also Supplementary Figs 6,7 and Supplementary Video.

Figure 5. Ectopic expression of MYMK-WT and -hypomorphic alleles can fully or partially rescue the mymk^insT/insT fusion phenotype.

(a) Approach to evaluate rescue of mymk^insT/insT 48 h.p.f. embryo fusion phenotype following co-injection of WT or mutant human MYMK mRNA. Nuclear dispersion is qualitatively stated to be low (no rescue), minimal (slight rescue), moderate (partial rescue), or high (rescue). (b) Optical sections of laterally mounted 48 hpf mymk^wt/wt embryo injected with mCherry, and six mymk^insT/insT embryos injected as per a. WT-, M1-, M5-MYMK mRNA injected embryos show partial rescue. M2-, M3-, and mCherry alone do not. Red: membrane bound mCherry. White: DAPI. Scale bar=50 μm. (c) Qualitative analysis of nuclear organization of embryos in b according to colour key in a. (d) Digitized images of the distribution of randomly generated non-overlapping xy positions (left), mymk^wt/wt myonuclei (middle), mymk^insT/insT myonuclei (right); left-right arrow d denotes distance between two nuclei; circles a, b, c are ∼10, 25, 50 μm from center nucleus, respectively. (e) Frequency of distances from each nucleus to all others. 2D random (grey), mymk^wt/wt (black) and mymk^insT/insT+WT mRNA (green) have the same trend, while mymk^insT/insT (red) exhibits differences at distances a, b, c (circles in d). Increased frequency at distance a reflects the increased number of near neighbors. Decreased frequency at distance b reflects the lack of nuclei between the rows of aligned nuclei in mymk^insT/insT embryos. (f) mymk^insT/insT embryos injected with mutant MYMK mRNA M2 (yellow) or M3 (orange) have normalized frequencies most similar to mymk^insT/insT (red dashed line) versus mymk^wt/wt (black dashed line), consistent with lack of rescue. (g) mymk^insT/insT embryos injected with mutant MYMK mRNA M5 (turquoise) or M1 (blue) have normalized frequencies that fall between mymk^insT/insT (red dashed line) versus mymk^wt/wt (black dashed line), consistent with partial rescue. Shaded bands=±SEM. (h) Two sample analyses for statistical significance (Kolmogorov-Smirnov) between couples of distributions in regions a, b. Top: Statistics for nuclei that fall in a distance range of 5 μm<d<10 μm (circle a). Bottom: Statistics for nuclei that fall in a distance range of 18 μm<d<28 μm (circle b). Grey box: NS (P_KS>0.05); Yellow box: *0.01<P_KS<0.05; Red box: **P_KS<0.01. See also Supplementary Figs 6,7.