Alpha-T-catenin is expressed in peripheral nerves as a constituent of Schwann cell adherens junctions

Alpha-catenins are a family of adherens junction proteins that organize individual cells into tissues through an ability to tether the cadherin/β-catenin cell–cell adhesive complex to actin filaments. They are encoded by separate genes and historically named according to tissue-types where first identified: αE^pithelial-catenin/CTNNA1, αN^eural-catenin/CTNNA2, αT^estes-catenin/CTNNA3 (Chiarella et al., 2018). αEcat (CTNNA1) is now appreciated as broadly expressed and essential for the development and homeostasis of most tissue types (Janssens et al., 2001), leading to it being the most studied form of α-catenin (Herrenknecht et al., 1991; Hirano et al., 1992; Torres et al., 1997; Lien et al., 2006; Vasioukhin et al., 2001; Sheikh et al., 2006). αNcat (CTNNA2) expression is largely restricted to the brain, where it plays critical roles in neuronal synapses required for full brain development (Park et al., 2002; Uemura and Takeichi, 2006). αT-cat (CTNNA3) is the most recently evolved α-catenin, but in contrast to its ancestral relatives (αE-cat and αN-cat) appears developmentally dispensable, as Ctnna3 knock-out mice are viable and fertile (Li et al., 2012; Vite et al., 2015). Despite its name, αTcat is best known for its role in the heart (Vite et al., 2015), where it contributes to the interspersed alignment of adherens junctions and desmosomes at the intercalated disk of cardiomyocytes (Janssens et al., 2003; Franke et al., 2006; Li and Radice, 2010; Li et al., 2019). This organization is critical for long-term cardiac function, since Ctnna3 knock-out mice develop age-related cardiomyopathy, similar to patients with loss-of-function mutations in αTcat (Janssens et al., 2003; van Hengel et al., 2013). These data have led to the notion that αT-cat plays an important but highly restrictive role in cardiomyocyte functionality.

Curiously, recent genetic association studies have linked mutations, non-coding polymorphisms and copy number variants (CNV) in CTNNA3 to a wide spectrum of diseases, including asthma, (Kim et al., 2009; Bernstein et al., 2013; McGeachie et al., 2015; Perin and Potocnik, 2014), food allergy (Li et al., 2015), autism spectrum disorder (Wang et al., 2009; O'Roak et al., 2012; Bacchelli et al., 2014), multiple sclerosis (Vilarino-Guell et al., 2019), diabetes (Zhang et al., 2021) and Alzheimer's (Smith et al., 2006; Miyashita et al., 2007). These disease associations are challenging to rationalize in context of the supposed restricted expression of αT-cat to cardiomyocytes and testes, suggesting αT-cat /CTNNA3 disease linkages may occur via an unappreciated cell/tissue-type. Moreover, while some of the aforementioned genetic studies show genotype/mRNA abundance phenotype correlations (Kim et al., 2009; McGeachie et al., 2015), far fewer show causal changes at the αT-cat protein level, where cross-reactivity of commercially available antibodies between α-catenin family members has been problematic.

Our group previously validated αT-cat's linkage to steroid-resistant asthma using the house-dust mite model, where Ctnna3 knock-out mice showed greatly attenuated airway hyperreactivity (Folmsbee et al., 2016a). Despite our observing cardiomyocytes of pulmonary veins as the major αT-cat-expressing cell in lung (Folmsbee et al., 2015; Folmsbee and Gottardi, 2017), the cell type through which αT-cat loss leads to reduced airway hyperreactivity has remained elusive. Here, we show that αT-cat protein is abundantly expressed in peripheral nerves (e.g. vagus and sciatic), specifically within the Schwann cell component. Since nerves innervate most organ systems and tissue types, we reason that many αT-cat /CTNNA3 disease linkages should be considered via its role in neuroglial cell types.

Early work on αT-cat/CTNNA3 used a human cDNA Rapid-Scan^TM panel to identify heart and testis as major tissue-types expressing CTNNA3, although low levels of RNA were also detected in brain, kidney and liver (Janssens et al., 2001). To better assess the full range of cell- and tissue-types expressing αT-cat /CTNNA3, we interrogated the Genotype-Tissue Expression (GTEx)-database, an NIH Common Fund resource to study relationships between genetic variation and gene expression across multiple reference tissues (GTEx Consortium, 2013). This database also serves as a convenient resource to validate RNA abundance across a greater range of tissues (∼50-types), especially those not typically harvested in earlier studies (e.g. nerve, sub-regions of brain). This tissue-wide bulk-RNA sequencing database reveals CTNNA3 expression in tibial nerve, spinal cord and various brain regions (substantia nigra, hippocampus, amygdala) at levels comparable to those observed in heart and testes (https://gtexportal.org/home/gene/CTNNA3). Additionally, single-cell expression within this database reveals CTNNA3 enrichment in the Schwann cell component of multiple tissues including skeletal muscle, esophagus, prostate and heart. Thus, the GTEx resource reveals that in addition to established expression of in myocytes, CTNNA3 is also abundantly expressed in the glial component of peripheral nerve.

Since spinal cord and brain register the highest CTNNA3 expression levels across human tissues, we sought to determine whether this expression might be through a related glial cell-type. To interrogate CTNNA3 expression in both central and peripheral nervous system, we examined more sensitive RNA-sequencing datasets from enriched (i.e. flow-sorted) cell populations. The Barres Lab RNA-sequencing dataset analyzed gene expression in the cells from the central nervous system, revealing abundant CTNNA3/Ctnna3 expression in oligodendrocytes and their precursor cells in both humans and mice (Zhang et al., 2014, 2016). Additionally, the Sciatic Nerve Atlas (SNAT) examined peripheral nerve cell types within mouse sciatic nerve, showing prominent Ctnna3 expression in Schwann cells (Gerber et al., 2021). Since oligodendrocytes and Schwann cells use cytoplasmic myelin to sheath and thereby insulate neurons of both central and peripheral nervous systems respectively, these databases suggest that αT-cat may support a common role in these functionally related cell types.

In an attempt to visualize αT-cat expression in cell- and tissue-types not optimally captured by typical paraffin-embedded thin-section methods, we generated a fluorescent reporter mouse in collaboration with Northwestern University's Transgenic Mouse Core. Using CRISPR-Cas9 gene editing, a membrane anchored enhanced green fluorescent protein (eGFPcaax) (Hines et al., 2015) was knocked into exon 2 of the Ctnna3 gene. The Ctnna3 promoter drives GFP expression while simultaneously knocking out native expression of αT-cat/Ctnna3 (Fig. 1A). We analyzed eGFPcaax-reporter expression in tissues known to express αTcat protein: heart and testes. We confirmed GFP protein expression in heart tissue from heterozygous (Ctnna3^GFP/+), but not wild-type (Ctnna3^WT) mice by immunoblot analysis (Fig. 1B). As expected, Ctnna3^GFP/+ and Ctnna3^{KO (GFP/GFP)} mice showed allele-dependent loss of αT-cat protein with complementary changes in GFP abundance in both heart and testes (Fig. S1A). No GFP is detected in tissues lacking Ctnna3 expression, such as liver or kidney (Fig. S1B). While GFP expression from this Ctnna3^GFP/+ reporter was sufficiently abundant to detect by immunoblotting (Fig. 1B), we were unable to detect GFP by immunofluorescence analysis with an anti-GFP antibody previously validated by our group for use on paraffin-embedded or whole mount samples, even in Ctnna3^KO(GFP/GFP) with two copies of GFP (Fig. 1C; GFP staining not shown). Although our GFP reporter contains a membrane-targeting lipid modifiable -CAAX motif to improve membrane enrichment (Hines et al., 2015), we reason the ability of αTcat to enrich at cadherin/β-cat junctions improves its detection in comparison with the GFP-caax, which distributes across various membrane compartments in the cell. Nonetheless, these data show we have generated a GFP-reporter mouse for Ctnna3 expression (Ctnna3^GFP/+) that also effectively deletes the endogenous αT-cat protein when crossed to homozygosity (Ctnna3^{KO (GFP/GFP)}).

A number of genome-wide association studies have linked CTNNA3 to chemical-induced occupational asthma (Kim et al., 2009; Bernstein et al., 2013), steroid refractory asthma (Perin and Potocnik, 2014) and asthmatic exacerbations (McGeachie et al., 2015). Indeed, our team validated αT-cat's linkage to asthma using the house-dust mite model, where Ctnna3 knock-out mice showed greatly attenuated airway hyperreactivity (Folmsbee et al., 2016a). However, despite our observing cardiomyocytes of pulmonary veins as the major αT-cat-expressing cell in lung (Folmsbee et al., 2016a, 2015; Folmsbee and Gottardi, 2017), the cell type through which αT-cat loss leads to reduced airway hyperreactivity has remained elusive. Given evidence via the GTEx database that CTNNA3 is abundantly expressed in peripheral nerves, where parasympathetic peripheral nerve inputs are known to innervate airways and control smooth muscle contractility/reactivity responses (Lewis et al., 2006; McGovern and Mazzone, 2014; McAlexander et al., 2015), we sought to determine whether αTcat protein might be expressed in the vagus. We harvested small (∼0.5 cm) segments of the vagus nerve and nodose ganglion just anterior to the thoracic cavity of wild-type (Ctnna3^WT) and knock-out (Ctnna3^{KO (GFP/GFP)}) mice for total protein and RNA extraction (Fig. 2A, schematic). We confirmed αTcat expression in both nodose ganglion and vagus nerve in wild-type mice; knock-out tissue showed no αTcat (Fig. 2B). Remarkably, we detected Ctnna3 RNA expression in vagus and nodose at levels comparable to and even greater than Ctnna3 levels in heart (Fig. 2C). The lack of correlation between Ctnna3 RNA and protein abundance raises the possibility that tissues may differentially regulate αTcat at the protein level. Moreover, since αTcat/Ctnna3 RNA is as abundant in vagus segments as in nodose ganglion, where the latter contains neuronal cell bodies, we reason that the bulk of Ctnna3 signal may be due to its expression in Schwann cells, consistent with database resources showing little Ctnna3 RNA in neurons (Zhang et al., 2014, 2016).

To specifically address whether αTcat protein is expressed in the Schwann cell component of peripheral nerves, we turned to the sciatic nerve due to its large size and ease of dissection (Fig. 3A, Movie 1). We validated αTcat protein expression in sciatic nerve lysates harvested from Ctnna3^WT but not Ctnna3^{KO (GFP/GFP)} mice; the ubiquitous αEcat protein remained largely unchanged (Fig. 3B). Immunofluorescence staining of wild-type sciatic nerve revealed αTcat localization to paranodal regions of glial-axon junctions, as well as structures reminiscent of myelin incisures, thin regions of Schwann cell cytoplasm excluded from compact myelin regions (Fig. 3C). Also known as Schmidt-Lanterman incisures, this structure results from successive concentric wraps of Schwann cell cytoplasm around an axon, where each wrapping is held together by adherens junction proteins in a uniquely characteristic intracellular ‘autotypic’ junction (Fannon et al., 1995; Tricaud et al., 2005). As expected, αTcat protein co-localizes with known adherens junction markers, E-cadherin and F-actin (Fig. 4). Importantly, Ctnna3^{KO (GFP/GFP)} mice showed no detectable αTcat immunostaining (Fig. 3C, bottom and 3D, right). We observed no obvious structural defect in these nerves, as overall E-cadherin and F-actin staining patterns were similar. Disappointingly, we were also unable to rely on GFP immunostaining as a convenient reporter for Ctnna3 expression in this tissue, even in Ctnna3^{KO (GFP/GFP)} mice with two copies of eGFPcaax (not shown). In summary, these data show that αTcat protein localizes to Schwann cell autotypic and heterotypic junction structures with familiar partners, E-cadherin and F-actin.

αTcat is the most recently evolved member of the alpha catenin family and best known for its role in cardiomyocyte junctions of heart, where its loss leads to an age-related cardiomyopathy that phenocopies disease in patients with αTcat loss-of-function mutations (Vite et al., 2015). These data have led to the notion that αT-cat plays an important but highly restrictive role in cardiomyocyte functionality. However, numerous genetic association studies have linked CTNNA3 to a wide spectrum of diseases, ranging from asthma and food allergy to autism, multiple sclerosis, diabetes and Alzheimer's (Kim et al., 2009; Bernstein et al., 2013; McGeachie et al., 2015; Perin and Potocnik, 2014; Li et al., 2015; Wang et al., 2009; O'Roak et al., 2012; Bacchelli et al., 2014; Vilarino-Guell et al., 2019; Zhang et al., 2021; Smith et al., 2006; Miyashita et al., 2007). These disease associations have been puzzling to rationalize in context of established restricted expression of αT-cat to cardiomyocytes (or testes), suggesting αT-cat/CTNNA3 disease linkages may occur via an unappreciated cell/tissue-type. Guided by the human tissue gene expression (GTEx) database, we demonstrate αTcat protein is also highly expressed in peripheral nerves (e.g. vagus and sciatic), specifically within the Schwann cell component. CTNNA3/Ctnna3 mRNA is also abundantly expressed in oligodendrocytes of the central nervous system (Zhang et al., 2014, 2016). Together with previous evidence that αTcat localizes to apical junctions of ependymal cells (Folmsbee et al., 2016b), we propose a generalized role for αT-cat in neuroglial functions. Since neuroglia participate in central and peripheral nervous systems, where the latter innervate all organs of the body, we speculate that the wide range of αT-cat/CTNNA3 disease linkages may be considered via its role in neuroglial cell types.

Anticipating the need to visualize Ctnna3 gene expression in a cell type challenging to resolve by thin section imaging (e.g. Schwann cells), we generated a novel fluorescent reporter Ctnna3^GFP/+ knock-out mouse. While we validated this reporter line for GFP expression and consequent loss of αTcat, we were unable to visualize the membrane-anchored GFP reporter by immunofluorescence analysis. Whether this is due to sequences in Ctnna3 that enhance the translation efficiency of αTcat, or the ability of αTcat to become highly enriched at cell–cell junctions compared with a general membrane-localized GFP is not clear. Nonetheless, this reporter may ultimately prove valuable for future studies seeking to characterize consequences of Ctnna3/αTcat loss in neuroglial cell types.

It is intriguing that αTcat/CTNNA3 expression is restricted to Schwann cells of peripheral nerves and oligodendrocytes in brain, two cell-types known to play critical roles in the myelination and insulation of neurons. Our evidence that αTcat specifically localizes to paranodes and myelin incisures of Schwann cells was not entirely surprising, given previous evidence that both structures comprise modified adherens junctions enriched for cadherins and catenin proteins (Fannon et al., 1995; Tricaud et al., 2005). Paranodes reflect regions of heterotypic glial cell-axon contact, whereas Schmidt-Lanterman myelin incisures reflect an alignment of successive wraps of Schwann cell cytoplasm around an axon, where each wrap is held together by adherens junction proteins (Fannon et al., 1995). While forced-expression of dominant-inhibitory versions of cadherin-catenin complex components can perturb this structure (Tricaud et al., 2005), whether polymorphisms or mutation in members of this cadherin-catenin adhesive complex can alter the critical function of Schwann cells (and related oligodendrocytes) relevant to disease remains unknown.

Our evidence that αTcat protein is detected in the vagus, which innervates the majority of visceral tissues throughout the body, offers an attractive lens through which we may view genetic linkages between αTcat/CTNNA3 and a range of diseases (reviewed in (Chiarella et al., 2018). With regards to CTNNA3 linkages to steroid-resistant asthma, our team previously validated this association showing that Ctnna3^KO mice display reduced airway hyperreactivity in response to methacholine challenge. Our evidence that αTcat is expressed in the vagus, a known regulator of airway smooth muscle responses (Lewis et al., 2006; McGovern and Mazzone, 2014; McAlexander et al., 2015), suggests the intriguing possibility that αTcat/CTNNA3 contributes to asthmatic airway responses through its role in the Schwann cell-myelinating component of peripheral nerves. Indeed, even CTNNA3's linkages to heart disease merit revisiting, where recent linkages to atrial fibrillation/Brugada Syndrome (Maltese et al., 2019) may be due to a peripheral nerve rather than a cardiomyocyte junction defect (Teodorovich et al., 2016).

We knocked-in a membrane associated eGFP construct into exon2 of the Ctnna3 locus using CRISPR gene editing in collaboration with Northwestern's Transgenic and Targeted Mutagenesis Laboratory. By design, the Ctnna3 promoter drives eGFP-caax expression from the endogenous locus, while knocking out expression of the gene. Note, addition of an exogenous bGH polyA immediately downstream of the stop codon was required to improve eGFP-caax expression from the endogenous Ctnna3 exon 2 locus, given the extremely large size of the Ctnna3 gene. We nucleofected the following into murine B6N-derived embryonic stem (ES) cells: Ctnna3-targeting CRISPR reagents, an eGFP repair plasmid designed to insert into the Ctnna3 exon2 locus, and a non-targeting PGK-plasmid harboring a puromycin selection cassette. Puromycin resistant ES cell clones were selected, propagated and genotyped for correct construct insertion. Targeted clones were expanded and validated, with select clones injected into albino B6J blastocyst stage embryos. Briefly, donor females (albino B6J) are hormone treated, mated, and plugged. Recipient (foster) females are also set up. ES cells are injected into the cavity of an expanded blastocyst stage embryo and injected embryos are surgically transferred into the reproductive tract of recipient females to generate chimeric mice. Chimeric mice were genotyped for correct insertion of the construct into the Ctnna3 locus. Two chimeric males (90% and 95%) were generated and bred to C57BL/6J mice, transmitting GFP to the F1 generation with initial genotype validation by endpoint PCR and currently via real-time PCR with Transnetyx (Cordova, TN). GFP expression was validated by immunoblot/ immunostaining analysis of target tissues. Detailed mouse report with PCR-gel validation will be provided upon request. This line will be available at Jax labs (stock number 037394) under the name C57BL/6-Ctnna3em1Cgot. All experimental protocols were approved by the Institutional Animal Care and Use Committee at Northwestern University.

Mice were euthanized and perfused with 10 ml of HBSS. For heart dissection, tissue was fixed in 4% PFA overnight at 4°C, then transferred into 70% ethanol for dehydration, paraffin embedding and ∼4 µm sectioning via microtome. For vagus nerve and nodose ganglion dissection, skin, salivary glands, and masticatory muscles were removed, followed by identification of the trachea, carotid artery, and glossopharyngeal nerve. The vagus nerve and nodose ganglion were identified, excised, and stored at −20°C for further processing as previously described (Norgren and Smith, 1994). For sciatic nerve dissection, skin and muscle of upper hind legs were removed, and a ∼3 mm segment of sciatic nerve was cut and placed into Zamboni's fixative for 10 min, followed by 15% glycerol/PBS solution (v/v) for 24 h at 4°C. Nerves were then incubated in successive glycerol solutions (45%, 60%, 66% glycerol in PBS) for 18-24 h each at 4°C. Tissue was stored at −20°C until ready for further processing. Nerve bundles were further dissected into individual fibers using a Leica MZFLIII dissecting microscope. Briefly, nerves were placed into 10 cm dishes containing HBSS. Nerves were teased apart using 30G needles and individual fibers were placed onto poly-L-lysine-coated Bond380 slides. Slides were then dried and stored at −20°C.

For paraffin-embedded heart slides: Slides were de-paraffinized by submerging in xylene twice for 5 min each. Slides were then gradually rehydrated in 100%, 90%, and 70% EtOH solutions twice for 2 min followed by three washes in PBS for 5 min each. Slides were subsequently quenched in a 10 mM glycine solution for 15 min at room temperature. Slides were simmered in a citrate-based antigen retrieval buffer for 30 min at 95°C. After cooling to 25°C, slides were briefly rinsed in PBS before immunostaining. For heart and sciatic nerve tissue, slides were blocked in a solution of 10% normal goat serum (NGS) in 0.3% TritonX-100 PBS for 30 min at 25°C. Slides received primary antibody prepared in a solution containing 3% NGS in 0.3%TritonX-100 PBS and incubated for 1 h at 25°C. Following primary antibody, slides were rinsed in PBS and then incubated in fluorescence conjugated secondary antibody for 30 min at 25°C. Tissue sections were briefly rinsed in water and mounted using ProLong Gold anti-fade mounting media. Heart tissue was imaged on a Zeiss Axioplan2 epifluorescence microscope. Sciatic nerves were imaged at Northwestern University’s Center for Advanced Microscopy core using a Nikon W1 spinning disk confocal microscope. 20× z-stack images were taken in 0.5 µm steps whereas 60× z-stack images were taken in 0.25 µm step sizes. All image files were pseudo-colored in FIJI for presentation in figures. Three-dimensional (3D) imaging of sciatic nerve was analyzed using Imaris software. Orthogonal views were obtained from maximum intensity projection (MIP) images. Videos were created via 3D reconstruction of z-stacks (10-25 µm depth; 0.25 µm step size).

Heart, vagus/sciatic nerve and nodose ganglion tissues were collected and snap-frozen in liquid nitrogen. Heart tissue was homogenized on ice using Tissue Tearor at medium speed for 1 min in lysis buffer (RIPA plus 0.1% SDS with Roche EDTA-free protease inhibitor) and allowed to sit on ice 10 min followed by sonication using Branson sonifier at 10% amplitude, 1 s on/off intervals for 10 rounds total. Lysates were centrifuged at 14,000 g for 10 min at 4°C and the supernatant collected and stored at −80°C. Vagus nerve and nodose tissues were processed as described except that tissues were pipetted up and down 20 times in lysis buffer using a large bore tip and allowed to sit on ice 10 min before sonication and centrifugation. Protein concentrations were determined via Bradford assay. Samples were diluted in 2× SDS loading buffer and boiled at 95°C for 5 min prior to loading. 50 µg of protein was run on a 4-20% gel at 120 V for 90 min. Following gel transfer, nitrocellulose membranes were rinsed in Ponceau S to confirm presence of protein. Membranes were blocked for 1 h at room temperature in 5% milk/TBS-T, followed by incubation in primary antibody in 5% BSA/TBS-T overnight with rocking at 4°C. After primary antibody incubation, membranes were washed three times for 10 min in TBS-T. Membranes received secondary antibody (HRP or fluorescent conjugated) prepared in a 5% milk/TBS-T solution for 1 h at room temperature with rocking. Membranes were then washed three times for 15 min in TBS-T before either incubation with Pierce ECL Plus solution for 1 min or direct fluorescence imaging on Licor Odyssey Fc scanner for 2-10 min.

RNA was extracted from snap-frozen heart, vagus nerve, and nodose ganglion tissue using the Qiagen RNeasy Plus Mini Kit as per manufacturer instructions. Briefly, tissue samples were lysed in 600 ml RLT buffer using sterile Beadbug homogenization tubes with 3 mm ceramic beads and Benchmark Beadblaster shaker run at power setting 4, 30 s shake with 30 s intervals, for four cycles and repeated the program twice. Homogenized sample was centrifuged at 14,000 g for 3 min. Supernatant was applied to Qiagen gDNA Eliminator spin columns before continuing with Qiagen protocol. RNA was eluted in 40 ml nuclease-free water. First strand cDNA was reverse transcribed from equal amounts of total RNA using iScript cDNA synthesis kit (Biorad, Hercules, CA, USA). Specific primers for amplification of the mouse Ctnna3 message: forward primer (FP) 5'-GGTTACTACCCTGGTGAATTGTCC-3′, reverse primer (RP) 5'-CTCTTTTCGAACTTCCTGGAGTGC-3′. Real-time PCR was performed using iQ SYBR Green Supermix according to the manufacturer instructions (Biorad, Hercules, CA, USA). PCR was carried out in 96-tube plates using the MyiQ Single Color Real-Time PCR Detection System and software (Bio-Rad). All reactions were done in triplicate with negative controls. Gapdh was used as the internal control. The relative change in gene expression was calculated using the 2^ΔΔCt method.

We thank Bruce Appel (University of Colorado, Denver) for the membrane-anchored eGFPcaax construct and Lynn Doglio, Eugene Wyatt and Rajesh Awatramani at Northwestern University’s Transgenic Mouse Core for advice and design of the Ctnna3^GFP/+ reporter mouse line. We thank Robert P. Schleimer (Allergy and Immunology, Northwestern) and Bradley Undem (Johns Hopkins University) for discussions regarding vagal contributions to asthma.