Troponin T3 expression in skeletal and smooth muscle is required for growth and postnatal survival: Characterization of Tnnt3^{tm2a(KOMP)Wtsi} mice

Summary

The troponin complex, which consists of three regulatory proteins (troponin C, troponin I and troponin T), is known to regulate muscle contraction in skeletal and cardiac muscle, but its role in smooth muscle remains controversial. Troponin T3 (TnnT3) is a fast skeletal muscle troponin believed to be expressed only in skeletal muscle cells. To determine the in vivo function and tissue specific expression of Tnnt3, we obtained the heterozygous Tnnt3+/flox/lacZ mice from Knockout Mouse Project (KOMP) Repository. Tnnt3^lacZ/+ mice are smaller than their WT littermates throughout development, but do not display any gross phenotypes. Tnnt3^lacZ/lacZ embryos are smaller than heterozygotes, and die shortly after birth. Histology revealed hemorrhagic tissue in Tnnt3^lacZ/lacZ liver and kidney, which was not present in Tnnt3^lacZ/+ or WT, but no other gross tissue abnormalities. X-gal staining for Tnnt3 promoter-driven lacZ transgene expression revealed positive staining in skeletal muscle and diapharam, and smooth muscle cells located in the aorta, bladder, and bronchus. Collectively, these findings suggest that troponins are expressed in smooth muscle, and are required for normal growth and breathing for postnatal survival. Moreover, future studies with this mouse model can explore TnnT3 function in adult muscle function using the conditional-inducible gene deletion approach.

INTRODUCTION

The troponin (Tnn) complex contains a number of regulatory proteins that control skeletal and cardiac striated muscle contraction. The Tnns are a cooperative complex of three subunits comprised of Troponin T (Tropomyosin interacting subunit, TnnT), Troponin C ([Ca2+] interacting subunit, TnnC), and Troponin I (Inhibitory subunit, TnnI) that function with tropomyosin (Tm) to regulate the interaction of Tm and actin filaments during force generation (el-Saleh, Warber et al. 1986; Szczesna and Potter 2002). Different combinations of Tnn subtypes are expressed in different muscle cells based on their fast or slow muscle phenotype. For example: fast cardiac muscle cells express Tnnt2, Tnni3, and Tnnc1; slow skeletal muscle cells express Tnnt1, Tnni1, and Tnnc1; and fast skeletal muscle cells express Tnnt3, Tnni2, and Tnnc2. In contrast, smooth muscle contraction is generally believed to be independent of Tnn regulation based on biochemical studies that failed to indentify these proteins in smooth muscle tissues (Johansson 1987; Jiang and Stephens 1994; Owens 1995; Darland and D'Amore 2001; Watanabe, Yoshino et al. 2003). However, it was recently reported that the fast skeletal Tnn genes are required to assemble a functional fast-twitch troponin complex, and their expression and association with tropomyosin-actin filaments in vascular smooth muscle cells has been demonstrated by immunofluorescence microscopy (Moran, Garriock et al. 2008). Thus, the nature of Tnn expression and function in smooth muscle in various tissues is an area of active investigation.

Our interest in Tnn expression and function in smooth muscle was derived from our quest to elucidate the mechanism of flare in rheumatoid arthritis (RA). Using longitudinal contrast enhanced (CE) MRI to study the natural history of knee flare in the tumor necrosis factor-transgenic (TNF-Tg) mouse model of RA (Keffer, Probert et al. 1991), we observed a predictive behavior of the adjacent popliteal lymph nodes (PLN), which increases in size and CE during a prolonged asymptomatic “expanding” phase, prior to a sudden decrease in CE and PLN “collapse” during the onset of arthritic flare (Proulx, Kwok et al. 2007; Proulx, Kwok et al. 2007; Li, Kuzin et al. 2010; Li, Zhou et al. 2011; Li, Ju et al. 2012). As this PLN collapse and knee flare, defined as a significant increase in synovial volume, occurs in the absence of any detectable changes in cellularity or autoimmunity, we performed a microarray gene expression analysis of mRNA isolated from wild-type (WT), TNF-Tg expanding, and TNF-Tg collapsed PLN. The most remarkable finding was the ~100-fold decrease in fast skeletal muscle gene expression, including Tnnt3, in expanding PLN versus WT and collapsed PLN. This finding intrigued us, particularly since TnnT3 deficiency in Sheldon-Hall syndrome is characterized by contractures of the distal joints of limbs (Sung, Brassington et al. 2003; Sung, Brassington et al. 2003; Stevenson, Carey et al. 2006; Robinson, Lipscomb et al. 2007; Zhao, Jiang et al. 2011). As subsequent attempts to confirm these differences in mRNA and protein expression in PLN tissue via in situ hybridization and immunohistochemistry failed to produce credible results due to technical issues with probe sensitivity and specificity, we chose to investigate transgene expression of the Tnnt3^{tm2a(KOMP)Wtsi} mouse, obtained from the Knock Out Mice Program (KOMP), as a surrogate marker of endogenous Tnnt3 gene expression. Since this targeted Tnnt3 gene produces a null mutation, we also analyzed the phenotype of Tnnt3 heterozygous (Tnnt3^lacZ/+) and homozygous (Tnnt3^lacZ/lacZ) mutant mice.

RESULT AND DISCUSSION

The Tnnt3^lacZ/+ (Tnnt3^{tm2a(KOMP)Wtsi}) mice in a C57BL/6 background were generated as part of the KOMP using the gene targeting strategy summarized in Figure 1. Central to the purposes of our studies, the targeting vector contains the lacZ expressing cassette with polyA tail inserted between Exon 9 and Exon 10, generating a null allele that produces β-galactosidase in cells with active Tnnt3 gene transcription (Fig. 1A). Although we did not attempt subsequent TnnT3 gain and loss of function studies with this line, the resulting genotypes following crossing with Flp recombinase-expressing (Fig. 1B) and Cre recombinase expressing mice (Fig. 1C) are shown to illustrate the potential of these Tnnt3^{tm2a(KOMP)Wtsi} mice.

Fig 1. Genomic structure of Tnnt3^{tm2a(KOMP)Wtsi} mice.

(A) The Tnnt3/lacZ knock-in allele (Tnnt3^{tm2a(KOMP)Wtsi}) is shown to illustrate the mutated Tnnt3 gene in which the lacZ and neomycin expression cassette was inserted between Exon 9 and Exon 10, resulting in a Tnnt3 null gene, and is flanked by FRT (Flippase Recognition Target) sites. This cassette contains the splice acceptor of mouse engrailed 2 exon 2 (En2 SA), an internal ribosome entry sequence (IRES) to initiate lacZ translation, and polyadenylation (pA) to terminate transcription after the lacZ gene. The neo gene is driven by human beta actin promoter (hBactP) and contains it own pA. Additionally, Exon 10 is flanked by loxP sites. (B) The resulting Tnnt3 conditional knockout gene structure that is generated by crossing Tnnt3^{tm2a(KOMP)Wtsi} mice with mice that express the Flp recombinase is shown to illustrate removal of the lacZ and neomycin cassette. (C) The resulting Tnnt3 knockout allele generated after crossing with Cre expression mice is shown.

Standard breeding of Tnnt3^{tm2a(KOMP)Wtsi} to C57BL/6 mice produced WT and Tnnt3^lacZ/+ with normal Mendelian distribution. However, an obvious growth delay was observed in Tnnt3^lacZ/+ mice assessed as adults at 2 months of age (Fig. 2A), and the mutant mice had a significantly smaller mass as determined by body weight out to 9-weeks (Fig. 2B). In contrast, Tnnt3^{tm2a(KOMP)Wtsi} self breeding failed to produce any Tnnt3^lacZ/lacZ adult mice in >20 litters, although Tnnt3^lacZ/lacZ pups were found dead a few hours after birth. A gross assessment of embryos harvested on E18.5 revealed that the Tnnt3^lacZ/lacZ knockouts are smaller than their Tnnt3^lacZ/+ littermates (Fig. 2C & D). These results demonstrate that TnnT3 is required for normal growth and postnatal survival.

Fig 2. Tnnt3 is required for postnatal survival, normal growth, and tnnc2 and tnni2 expression.

(A) The gross phenotype of a representative (n>4) Tnnt3^lacZ/+ mouse at 2 month of age is shown next to its WT (+/+) littermate to illustrate the smaller sized of the mutant animals. Photographs of homozygous (lacZ/lacZ) mutant mice could not be obtained since these mice die shortly after birth. (B) The body weights of Tnnt3 ^lacZ/+ and their WT littermates were measured weekly from 3 to 9 weeks (wk) after birth, and the data are presented as the mean +/− SD (n=4, *p<0.05). A photograph (C) and radiograph (D) of representative Tnnt3^lacZ/lacZ, Tnnt3^lacZ/+ and Tnnt3^+/+ embryos harvested at E18.5 (n≥3) is presented to illustrate the gene dose effect on growth. QRT-PCR was performed on RNA obtained from muscle of WT, Tnnt3^lacZ/+ or Tnnt3^lacZ/lacZ adult mice (E), or from E18.5 embryos (F), and the data are presented as the mean +/− SD (n=3, *p<0.05).

To confirm Tnnt3gene deletion, tnnt3mRNA levels were assessed in WT and Tnnt3^lacZ/+ mice (Fig. 2E), and WT, Tnnt3^lacZ/+ and Tnnt3^lacZ/lacZ E18.5 embryos (Fig. 2F). As we expected, tnnt3 were significantly decreased in Tnnt3^lacZ/+ vs. WT mice. Interestingly, mRNA levels of the other fast skeletal troponin components, tnnc2 and tnni2, were also down-regulated in Tnnt3^lacZ/+ mice. Moreover, mRNA levels for tnnt3, tnnc2 and tnni2, were all dramatically decreased in E18.5 embryos, suggesting that these genes are co-regulated at this developmental stage, and that there is no compensation for the loss of Tnnt3.

To investigate the expression pattern of Tnnt3, X-gal staining was performed on whole mount new born pups (Fig. 3A), and tissue sections (Fig. 3 B–P) of Tnnt3^lacZ/+ and WT littermates. The whole mount staining revealed β-galactosidase activity throughout the animal except the tissues that are not covered by skeletal muscle (Fig. 3A), such as the brown fat (indicated by *), galea aponeurotica (indicated by #) and temporal fascia (indicated by ↑). We also observed very strong β-galactosidase activity in skeletal muscle of the lower limb (Fig. 3G) and the diaphragm (Fig. 3K) from Tnnt3^lacZ/+ mice, which is consistent with high levels of Tnnt3 expression in these skeletal muscle cells. Moreover, β-galactosidase activity was observed in smooth muscle cells from aorta (Fig. 3H), bladder (Fig. 3I) and bronchus (Fig. 3J) tissues isolated from the Tnnt3^lacZ/+ mice. In contrast, there was no positive staining or β-galactosidase activity in spleen (Fig. 3L), heart (Fig. 3M), brain (Fig. 3N), kidney (Fig.3O), and liver (Fig. 3P) tissues from Tnnt3^lacZ/+mice, or any tissues from the WT littermates (Fig. 3B–F).

Fig. 3. The tissue specific expression pattern of Tnnt3.

(A) Whole mount was performed on 1-day old Tnnt3^lacZ/+mice and their WT littermates (n=2). The photograph illustrates the broad X-gal staining in skeletal muscle throughout the Tnnt3^lacZ/+ animal except in brown fat tissue (*), galea aponeurotica (indicated by #) and temporal fascia (indicated by ↑). No β-galactosidase activity was detected in Tnnt3^+/+ mice. (B–P) X-gal staining of tissues from 2 month-old Tnnt3^lacZ/+ mice and WT littermates (n>3), and representative micrographs are show to illustrate the tissue specific expression of the transgene. Note the intense positive staining in skeletal muscle cells next to leg(G) and diaphragm (K), and smooth muscle cells in the aorta (H), bladder (I) and bronchus (J) in the tissues from Tnnt3^lacZ/+ mice. In contrast there was no positive staining in spleen (L), heart (M), brain (N), kidney (O), and liver (P) tissues from Tnnt3^lacZ/+ mice, or any tissues from the WT littermates.

To further explore the delayed growth phenotype of Tnnt3^lacZ/+ mice (Fig. 2), we examined the skeletons of WT and Tnnt3^lacZ/+ mice (Fig. 4). X-gal staining revealed that Tnnt3 is not expressed in bone, bone marrow or joint connective tissue (Fig. 4A), but is highly expressed in the adjacent skeletal muscle (Fig. 4B). Alcian-blue/alizarin-red staining of adult skeletons revealed a subtle phenotype in which the Tnnt3^lacZ/+ mice had more mineralized bone than WT, especially in the ankle, tail, elbow and skull (Fig. 4C). Similarly, Tnnt3^lacZ/+ embryos have more mineralized bone in the skull and spine versus their WT littermates (Fig. 4D). By taking a closer look at the forelimbs (Fig. 4E) and hindlimbs (Fig. 4F), we found that the mutants have normal joint development, with a very modest decrease in limb length. Quantification of the growth plates in E18.5 embryos confirmed this mild phenotype, as the WT was significantly longer vs. TnnT3 ^lacZ/+ (898.87 +/− 14.65 vs.709.39 +/− 67.18 µm (n=4; p=0.008). Similarly, histologic assessment of spinal tissue from WT and Tnnt3 ^lacZ/+ embryos revealed a mild growth phenotype, but no remarkable differences in anatomy, mineralization or matrix composition (Fig. 4G & H). Collectively, these findings are consistent with the mild growth delay in Tnnt3^lacZ/+ mice, which could be caused by improper skeletal loading due to defects in skeletal and some smooth muscles.

Fig. 4. Tnnt3 heterozygous mice have a mild skeletal phenotype.

X-gal staining was performed on long bones from 2-month-old WT (A) and Tnnt3^lacZ/+ (B) mice (n=3). The micrographs show that the transgene is not expressed in bone, bone marrow or joint connective tissue, but is highly expressed in the adjacent skeletal muscle. (C–F) Alcian-blue/ Alizarin-red staining was performed on the skeletons of 2-month-old (C), and E18.5 embryos (D) of Tnnt3^lacZ/+ and WT mice (n=2) to assess differences in the mineralized (red) and unmineralized (blue) tissue. A closer comparison of the forelimbs (E) and hindlimbs (F) reveals a modest decreased limb length in Tnnt3^lacZ/+ embryos. (G, H) Alcian-blue/Orange G/Alizarin-red staining was performed on histology slides of spinal tissue from E18.5 embryos of WT (G) andTnnt3^lacZ/+ (H) mice. No remarkable differences were observed.

To elucidate the postnatal cause of death of Tnnt3^lacZ/lacZ mice we performed H&E stained histology on heart, lung, liver and kidney from WT (Fig. 5A–D), Tnnt3^lacZ/+ (Fig. 5E–H) and Tnnt3^lacZ/lacZ (Fig. 5I–L) embryos harvested at E18.5. No abnormalities were detected in any of the heart (Fig. 5I) and lung (Fig. 5J) tissues examined. In contrast, anomalous hemorrhagic tissues were detected in Tnnt3^lacZ/lacZ liver (Fig. 5K) and kidney (Fig. 5L). However, based on mouse physiology, it is unlikely that these lesions could cause immediate death after birth (Turgeon and Meloche 2009). Therefore, given the predicted Tnnt3 expression in diaphragm (Fig. 3K) and bronchus (Fig. 3J) tissue based on the X-gal staining, we conclude that the death of Tnnt3^lacZ/lacZ mice shortly after birth is due to respiratory distress.

Fig. 5. Histologic evidence of hemorrhage in Tnnt3^lacZ/lacZ liver and kidney.

H&E stained histology was prepared from WT, Tnnt3^lacZ/+ and Tnnt3^lacZ/lacZ embryos harvested on E18.5. Representative micrographs of heart (A, E, I), Lung (B, F, J), liver (C, G, K) and kidney (D, H, L) tissue are shown. Note the normal appearances of the heart (I) and lung (J) tissue, and the hemorrhagic tissue (arrows) in Tnnt3^lacZ/lacZ liver (K) and kidney (L). No abnormalities were detected in any of the tissue harvested from WT and Tnnt3^lacZ/+ embryos. Representative histology of diaphragm tissues from WT (M), Tnnt3^lacZ/+ (N) and Tnnt3^lacZ/lacZ (O) embryos stained with X-gal are also presented to illustrate the markedly thinner diaphragm muscle in the Tnnt3^lacZ/lacZ embryo.

To follow up on our provocative microarray results, we performed X-gal staining of Tnnt3^lacZ/+ PLN and its surrounding tissue. To our chagrin, the results clearly indicate that Tnnt3 is not expressed in PLN, but is highly expressed in the adjacent skeletal muscle cells that are physically attached by connective tissue (Fig. 6). While this negative find strongly suggests that our microarray results were caused by skeletal muscle tissue contamination of the WT and collapsed PLN, which are harder to dissect than expanding PLN, it also highlights an inconsistency of conventional lymph node anatomy that posits that PLN are embedded in fat (Standring 2005). In contrast, we find that one third of murine PLN are physically attached to skeletal muscle.

Fig. 6. Tnnt3 is highly expressed in the skeletal muscle attached to the PLN.

X-gal staining of the PLN and surrounding fat and muscle tissue from Tnnt3^lacZ/+ mice reveals that positive staining is only detected in the skeletal muscle.

In summary, we find that Tnnt3 is a muscle specific gene in mice, which is required for normal growth and postnatal survival. Although these finding were predicted based on our current knowledge of fast skeletal muscle troponins, we found that the utility of the KOMP derived Tnnt3^{tm2a(KOMP)Wtsi} mouse provided us with the most cost-effective approach to bring our preliminary microarray findings to a timely conclusion. Additionally, these mice allowed us to investigate an emerging controversy in the field of muscle biology, and provide indirect evidence of Tnnt3 expression in smooth muscle cells in aorta, bladder and bronchus. This novel finding warrants follow up studies to formally demonstrate TnnT3 protein expression in smooth muscles cells, and the use of the Tnnt3^{tm2a(KOMP)Wtsi} mouse with an appropriate smooth muscle specific Cre expressing line (Grcevic, Pejda et al. 2012) to assess TnnT3 function in this tissue via conditional-inducible gene deletion.

MATERIALS AND METHODS

Animals

Heterozygous Tnnt3^{tm2a(KOMP)Wtsi} mice in a C57BL/6 background were generated by the Knock Out Mice Program (KOMP) at the University of California (Project ID# CSD41352), Davis, using the targeting vector described in Figure 1. We obtained these founder mice, and performed the following experiments on their WT, Tnnt3^lacZ/+ and Tnnt3^lacZ/lacZ progeny under protocols approved by the University of Rochester Committee for Animal Resources.

RNA Analyses

Muscle tissues were collected from the legs of adult (>3-months) WT and Tnnt3^lacZ/+ mice (n=3); and WT, Tnnt3^lacZ/+ and Tnnt3^lacZ/lacZ E18.5 embryos (n=3); from which total RNA was extracted using the Qiagen RNeasy® Fibrous Tissue Mini Kit (Qiagen, Valencia, CA). The cDNA was prepared by reverse-transcription of the total RNA using the iScript cDNA synthesis kit (BioRad, Hercules, CA). SYBR Green FastMix® (Quanta, Gaithersburg, MD) was used to detect the DNA synthesis in real-time PCR reactions performed in triplicate using the Rotor-Gene Q platform (Qiagen, Valencia, CA). The annealing temperature for the PCR was 59 °C, and the Ct values for the fast skeletal troponins were normalized to β-actin. The PCR primer pairs were:

• β-Actin, F- AGATGTGGATCAGCAAGCAG; R- GCGCAAGTTAGGTTTTGTCA;

• tnnt3, F- CCCTCATTGACAGCCACTTT; R- CCTCCTCTCTTCTGGCCTTC;

• tnnc2, F- AGGTAGGACCTGGCCTCAG; R- CACCTTTGGGTGGTGGAGT;

• tnni2, F- GCTTGAGATCTCAGGATG, R-TCCATGCCAGACTTCTCC;

Tissue Histology

Analysis of E18.5 embryos and whole mount X-gal staining to detect β-galactosidase activity was performed as previously described (Hilton, Tu et al. 2007). Briefly, embryos were harvested and placed in ice-cold PBS, followed by fixation in 0.2% glutaraldehyde overnight at 4°C. The embryos were then incubated overnight at 4°C in the X-gal staining solution (Gold Biotechnology, St Louis, MO), which contains the substrate (0.25 mg/ml), 2mM MgCl2, 0.01% Na Deoxycholate, 0.02% NP-40, 5mM K4Fe(CN)6, and 5mM K3Fe(CN)6.

X-gal staining of frozen tissue sections was performed as previously described (Hilton, Tu et al. 2007). Briefly, the harvested tissues were prefixed in 0.2% glutaraldehyde overnight in 4°C. For adult bone and joint tissues, decalcification was performed in 14% EDTA for 10 days in 4°C. All samples were transferred to a 15% sucrose PBS solution overnight, followed by 30% sucrose PBS solution overnight in 4°C. Afterwards, the samples were mounted in O.C.T mounting medium (Fisher Scientific), snap-frozen in Isopentane/liquid nitrogen, and then stored at −80°C before sectioning. Cryostat sectioning was performed to cut 6 µm thick sections that were post-fixed with 0.2% glutaraldehyde at 4°C for 10min, stained with X-gal solution for 4 hours at room temperature, and counterstained with fast red (Sigma, St. Louis, MO).

For general histology, organ tissues from E18.5 embryos were fixed in 10% neutral buffered formalin, decalcified, paraffin processed and embedded as previously described (Kohn, Dong et al. 2012). Three µm tissue sections were cut, and the sections were stained with hematoxylin and eosin (H&E). Photomicrographs were obtained at 10× magnification with Zeiss Axioskop 40 microscope.

Skeletal prep

The skin and internal organs of adult mice 2-months of age or embryos at E18.5 were removed, and the remaining skeletal and soft tissues were fixed overnight in 95% ethanol, followed by an overnight incubation in acetone, as previously described (Dong, Jesse et al. 2010). Afterwards, the samples were incubated in Alcian Blue/ Alizarin Red staining solution, transferred into 2% KOH, and stored in a 1:1 glycerol/ethanol until they were photographed.

Statistics

Simple pair wise comparisons of WT vs. Tnnt^3lacZ/+ were calculated as mean +/− standard deviation in Microsoft Excel, and the significance was determined using the Student’s t-test. For multiple comparisons, ANOVA was performed using GraphPad prism (GraphPad Software, Inc., La Jolla, CA). For all tests, p<0.05 was considered significant.