Long Noncoding RNA lnc-TSSK2-8 Activates Canonical Wnt/β-Catenin Signaling Through Small Heat Shock Proteins HSPA6 and CRYAB

Congenital heart defects (CHDs) are the most common birth defects worldwide. 22q11.2 deletion syndrome is the most common microdeletion disorder that has been frequently associated with conotruncal malformations. By now, the dosage-sensitive gene TBX1 has been adopted as the major pathogenic gene responsible for 22q11.2 deletion, which is regulated by canonical Wnt/β-catenin signaling pathway in heart outflow tract development. Here, we report the long noncoding RNA (lncRNA) lnc-TSSK2-8, which is encompassed in the 22q11.2 region, that can activate canonical Wnt/β-catenin signaling by protecting β-catenin from degradation, which could result from decreased ubiquitination. Such effects were mediated by two short heat shock proteins HSPA6 and α-β-crystallin (CRYAB), whose expression was regulated by lnc-TSSK2-8 through a competing endogenous RNA (ceRNA) mechanism. In clinical practice, the pathogenesis of copy number variation (CNV) was always attributed to haploinsufficiency of protein-coding genes. Here, we report that the 22q11.2 lncRNA lnc-TSSK2-8 significantly activated canonical Wnt/β-catenin signaling, which has major roles in cardiac outflow tract development and should act upstream of TBX1. Our results suggested that lncRNAs should contribute to the etiology of CNV-related CHD.


INTRODUCTION
Congenital heart defects (CHDs) are the most common birth defects worldwide, with an incidence of nearly 10∼12 per 1,000 live infants (1∼1.2%) (Hoffman and Kaplan, 2002;Van Der Linde et al., 2011). As a consequence of improvement in surgical and medical management, more and more patients with severe forms of CHDs can survive until their 30s and beyond. Genetic variation, environmental causes, and their interactions could contribute to the disruption of heart development, thus leading to formation of CHDs (Pierpont et al., 2018;Zu et al., 2020). Genetic variations associated with CHDs have been revealed extensively in recent years. Single gene disorder and gross chromosomal anomalies/aneuploidy can be identified in up to 5 and 10% CHD cases, respectively. Copy number variations (CNVs) are segmental deletions or duplications ranging widely in size of millions of base pairs; pathogenic CNVs could explain about 3∼25% syndromic CHD and 3∼10% isolated CHD cases (Cowan and Ware, 2015).
Since the function of proteins had been well annotated, most clinical pathogenic CNVs were interpreted based on their effect on gene dosage of protein-coding genes. There are 46 proteincoding genes located in the typical 3M region of 22q11.2 locus, among which most studies of interest relate to TBX1, which encodes a T-box transcription factor. TBX1 was found to be a crucial gene in the LCR22A-LCR22B region; heterozygous loss of function mutations of Tbx1 in the mouse resulted in penetrant defects that are reminiscent of CHD in 22q11.2 (Lindsay et al., 2001;Chapnik et al., 2012). Another gene of interest is DGCR8, which encodes a double-stranded RNA-binding protein that is involved in microRNA (miRNA) biogenesis. Heterozygosity of Dgcr8 results in neuronal deficits (Stark et al., 2008), whereas ablation of both alleles in neural crest cells causes heart defects (Chapnik et al., 2012). Additionally, evidence of other individual protein-coding genes in 22q11.2 region responsible for major cardiac phenotypes of 22q11.2DS has also accumulated. For example, haploinsufficiency of CRKL could account for the cardiac anomalies in individuals with nested distal deletions (Zheng et al., 2015).
Long noncoding RNAs (lncRNAs), which contain >200 nucleotides, have been revealed to participate in the regulation of cellular and tissue function and play a great part in heart development (Devaux et al., 2015). LncRNAs can repress or activate gene expression epigenetically with diverse mechanisms: e.g., most lncRNA molecules localize in the nucleus and regulate gene expression epigenetically; a proportion (∼15%) of lncRNA are located in the cytoplasm where they function posttranscriptionally (Devaux et al., 2015). Several lncRNAs have been identified as key regulatory molecules involved in cardiac development, such as Braveheart (Klattenhoff et al., 2013), Fendrr (Grote and Herrmann, 2013), Upperhand (Anderson et al., 2016), and BANCR (Wilson et al., 2020). Several examples of CHD-associated lncRNAs have also been reported. Gu et al. (2016) indicated that circulating plasma lncRNAs might serve as novel biomarkers for CHD. Jiang et al. (2018) discovered that HOTAIR is upregulated in cardiac tissues and plasma of patients with CHD; thus, it could serve as a potential diagnostic biomarker for CHD. Wang et al. (2018) suggested the relationship between HA117 and TOF, although the molecular basis of HA117 remains unclear. Jiang et al. (2019) reported upregulated SNHG6 expression in fetal heart tissues of patients with VSDs. Ma et al. (2020) reported that TBX5-AS1 should be involved in TOF by affecting cell proliferation by regulating TBX5. This evidence signifies the importance of lncRNA in CHD. Haploinsufficiency of TBX1 is regarded as the major candidate for cardiac OFT malformation. Recently, canonical Wnt/β-catenin has been revealed to play major roles in cardiac OFT development upstream of TBX1 (Racedo et al., 2017). It is unknown whether lncRNAs participate in this process. The canonical Wnt/β-catenin signaling pathway has been revealed to play developmental stage-specific roles in early heart development. The Wnt family proteins serve as secreted signaling molecules that initiate a series of intracellular pathways. In the canonical Wnt/β-catenin signaling pathway, Wnt signaling starts with the generation of Wnt ligands into the extracellular space. The Wnt molecules then bind to the Frizzled receptor (Frz) and its coreceptor LRP5/6 to form trimeric complexes. The formation of such complexes inhibits the glycogen synthase kinase 3β (GSK-3β) destruction complex, which functions in stabilizing the cytosolic β-catenin (Blankesteijn et al., 2008). Cytosolic accumulation of β-catenin leads to its translocation into the nucleus. Subsequently, the Wnt proteins bind to the T-cell factor (TCF)/lymphoid enhancerbinding factor (LEF) transcription factors to activate Wntresponsive genes. Activation of Wnt/β-catenin signaling during embryoid body (EB) formation promotes embryonic stem cell differentiation into cardiomyocytes. Whereas the activation of Wnt/β-catenin signaling in a later stage (during gastrulation) results in inhibition of cardiomyocyte formation (Naito et al., 2006;Ueno et al., 2007).
In the present study, we discovered that the lncRNA lnc-TSSK2-8, which is located in the 22q11.2 region, significantly promotes the stabilization of β-catenin. Therefore, we propose that lnc-TSSK2-8 should contribute to heart development by regulating Wnt/β-catenin signaling pathway. In contrast to earlier studies that emphasize the crucial roles of protein-coding genes, these data suggest that the relationship between disease and noncoding transcripts should be further explored.

Rapid Amplification of Complementary DNA Ends
Here, 5 and 3 rapid amplification of complementary DNA (cDNA) ends (RACE) was performed on RNA isolated from HEK293 cells using a SMARTer RACE 5 /3 Kit (TAKARA, Dalian, China), following the manufacturer's instructions. The PCR products were separated on a 1% agarose gel and validated by Sanger sequencing. The following gene-specific primers were used for PCR: GACTGAAGGAGTAGAAA (5 RACE primer) GTCCAGGTGTCCCTGCCTCCCATTG (3 RACE primer 1) GGGGGAAGCCCACAATGAGCAG (3 RACE primer 2) Isolation of Nuclear and Cytoplasmic RNA Nuclear RNA and cytoplasmic RNA of HEK293 cells were extracted and purified using the PARIS TM Kit (Invitrogen, Thermo Fisher Scientific, Inc., Waltham, MA, United States) according to the manufacturer's instructions.

Cell Culture and Long Noncoding RNA Overexpression
The HEK293 cell line was purchased from the Type Culture Collection of the Chinese Academy of Sciences, Shanghai, China. Cells were cultured with Dulbecco's modified Eagle's medium (DMEM; Gibco, Thermo Fisher Scientific, Inc., Waltham, MA, United States) containing 10% fetal bovine serum (FBS; Gibco, Thermo Fisher Scientific, Inc., Waltham, MA, United States) and 0.1% penicillin/streptomycin (NCM Biotech, Suzhou, Jiangsu, China) at 37 • C with 5% CO 2 .

RNA Sequencing Analysis
Cell samples were lysed in TRIzol (Ambion, Austin, TX, United States). Total RNA was extracted using the RNeasy kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions.
A total amount of 2 µg RNA per sample was used as input material for the RNA preparations. Firstly, ribosomal RNA was removed using Epicenter Ribo-zero TM Kit (Illumina Inc., San Diego, CA, United States), and rRNA-free residue was cleaned up by ethanol precipitation. By using the fragmentation buffer, the mRNA and noncoding RNAs were fragmented into short fragments (about 200∼700 bp), then the first-strand cDNA was synthesized by random hexamer-primer using the fragments as templates. Buffer, dNTPs, RNase H, and DNA polymerase I were added to synthesize the second strand cDNA. Subsequently, A-Tailing Mix (QIAGEN, Hilden, Germany) and RNA Index Adapters were added to perform end-repair, and the resulting cDNA was amplified by PCR. The double-strand cDNA was purified with QiaQuick PCR extraction kit and then used for endpolishing. Sequencing adapters were ligated to the fragments, then the second strand was degraded using Uracil-N-Glyosylase (UNG) finally. The quantity and quality of cDNA libraries were evaluated using Agilent 2100 Bioanalyzer (ABI, New York, NY, United States). The double-stranded PCR products were heat denatured and circularized by the splint oligo sequence to obtain the final library. The single-stranded circular DNA was amplified using phi29 (Thermo Fisher Scientific, Waltham, MA, United States) to generate a DNA nanoball. The library preparations were sequenced on BGISEQ-500 platform, and 150bp paired-end reads were generated.
Raw sequencing reads with low quality (e.g., proportion of bases with sQ ≤ 5 greater than 50%; proportion of N greater >10%; 5 adaptor contamination) were removed using customized scripts. The remaining clean data were aligned to the human reference genome (version GRCh38) using STAR (Dobin et al., 2013) with standard options for long RNA sequencing (RNAseq) pipeline of Encyclopedia of DNA Elements (ENCODE; 1 ): -outFilterType BySJout, -outFilterMultimapNmax 20, -alignSJoverhangMin 8, -alignSJDBoverhangMin1, -outFilterMismatchNmax 999, -outFilterMismatchNoverReadLmax 0.04, -alignIntronMin 20, -alignIntronMax 1000000, -alignMatesGapMax 1000000. Gene level expression was estimated using the featureCounts software (Liao et al., 2014). Differential gene expression analysis was performed on the raw counts using the DESeq2 (Love et al., 2014) package. Genes with fold change greater than 2 and P value lower than 0.05 were identified as differentially expressed genes (DEGs). Functional enrichment analysis was performed with the clusterProfiler (Yu et al., 2012) package; gene set enrichment analysis was performed for Gene Ontology (GO) items [Biological Process (BP), Molecular Function (MF), Cell Component (CC)] and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway items. We used the genes with fold change greater than 1.2 for functional enrichment analysis, since the DEG number is too small for such analysis. BH adjustments were used for multiple testing of functional enrichment analysis; p < 0.05 and q < 0.05 were used as thresholds.

RT-qPCR
Total RNA was extracted with TRIzol reagent (Ambion, Austin, TX, United States) and reverse transcribed into cDNA with PrimeScript RT reagent Kit (Takara, Dalian, China). RT-qPCR was conducted with TB Green Premix Ex Taq II kit (Takara, Dalian, China) on the CFX 9600 Real-Time PCR detection system (Bio-Rad Laboratories, Inc., Hercules, CA, United States). Fold changes of the lncRNA and mRNAs were calculated with small nuclear RNA U6 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as internal controls, respectively, based on 2 − Ct method. All the Bulge-Loop RT primers for both microRNAs and U6 were purchased from RiboBio (Guangzhou, Guangdong, China).

Luciferase Reporter Assay
Lnc-TSSK2-8 wild type/mutation (WT/MUT) and heat shock protein (HSP)A6 UTR WT/MUT were inserted into the BamH I and Hind III sites of the pMIR dual-luciferase vector (Promega, Madison, WI, United States). These plasmids as well as the pMIR vector were transfected with the miR-6721-5p mimics or NCmimics into HEK293 cells for 48 h, respectively. Finally, luciferase activities were detected using the Dual-Luciferase Reporter Assay System (Promega, Madison, WI, United States) according to the manufacturer's protocol.

BrdU Cell Proliferation Assay
The cell proliferation assay was performed using TransDetect EdU Flow Cytometry Kit-647 Fluorophore (TransGen Biotech, Beijing, China). Cells were cultured with EdU solution for 2 h and fixed by 1× EdU permeabilization buffer. The assay was visualized by the FACScan system (Bio-Rad Laboratories, Inc., Hercules, CA, United States).

Statistical Analysis
The SPSS 16.0 software was used for statistical analysis. Differences between two groups were analyzed by implementing Student t-test, and P-value < 0.05 was regarded as statistically significant. All data were expressed as mean ± SD.

Full-Length Characterization and Subcellular Localization of lnc-TSSK2-8
The full-length sequence of lnc-TSSK2-8 was successfully obtained from HEK293 cells by rapid amplification of the 5 and 3 cDNA assays. We confirmed the sequence by performing PCR and Sanger sequencing. The full-length transcript of lnc-TSSK2-8 was 1,128 in size with a polyadenylated tail ( Figure 1A).
To determine the subcellular localization of lnc-TSSK2-8, we separated the nuclear and cytoplasm fractions of HEK293 cells and quantified its expression with RT-qPCR. The results suggested that lnc-TSSK2-8 is located both in the nucleus and cytoplasm, although the cytoplasmic composition was much more abundant ( Figure 1B).

lnc-TSSK2-8 Activates Canonical Wnt/β-Catenin Signaling
We overexpressed the full-length transcript of lnc-TSSK2-8 in HEK293 cells (Figure 2A) and examined the expression of the key components of canonical Wnt/β-catenin signaling pathway. Overexpression of lnc-TSSK2-8 increased cell proliferation by ∼20% ( Figure 2B) and activated the transcription of the LEF (LEF1) (Figure 2C), whereas the transcriptional level of β-catenin (CTNNB1) did not change ( Figure 2D). Further exploration suggested that the protein expression level of β-catenin increased significantly; such an increase should be associated with its decreased phosphorylation (Figures 2E,G). Subsequently, we revealed that the phosphorylation of GSK-3β (Figure 2F) was enhanced and established that ubiquitination of β-catenin could be inhibited by lnc-TSSK2-8.

Transcriptomic Analysis of lnc-TSSK2-8 Overexpression
To explore the molecular basis of Wnt/β-catenin signaling regulation, we performed transcriptomic analysis. Our results indicated that overexpression of lnc-TSSK2-8 resulted in differential expression of 62 genes/lncRNAs, among which HSPA6 was the most significantly upregulated (p = 1.27 × 10 −5 ) (Figures 3A,B). GO gene set enrichment analysis indicated that the DEGs were enriched in MFs such as transmembrane transporter activities ( Figure 3C). KEGG enrichment analysis indicated that the DEGs were enriched in beta-Alanine metabolism (p < 0.05).

lnc-TSSK2-8 Activates Wnt/β-Catenin Signaling Through a Competing Endogenous RNA Mechanism
Since lnc-TSSK2-8 was mainly distributed in the cytoplasm, we explored whether it could exert its function as a competing endogenous RNA (ceRNA) on DEGs. We predicted that miR-6721-5p targets both the HSPA6 3 UTR and the lnc-TSSK2-8 fulllength sequence (Figure 4A) with the miRDB 2 (Liu and Wang, 2019; Chen and Wang, 2020). We then validated the interaction between miR-6721-5p and HSPA6-3 UTR/lnc-TSSK2-8 using a dual-luciferase reporter assay. The results suggested that ectopic expression of miR-6721-5p leads to notably reduced luciferase activity of WT HSPA6 3 UTR, while the luciferase activity could not be changed if the binding sites of HSPA6 3 UTR were mutated ( Figure 4B). Similarly, the interaction of miR-6721-5p and lnc-TSSK2-8 was also validated ( Figure 4C). Overexpression of miR-6721-5p decreased the transcriptional level of lnc-TSSK2-8 ( Figure 4D) but did not affect the gene expression of βcatenin (Figure 4E), while the stabilization of β-catenin was impaired, since its phosphorylation and protein expression were upregulated and downregulated, respectively ( Figure 4F). Furthermore, overexpression of lnc-TSSK2-8 greatly activated the transcription of HSPA6 (Figure 4G). We overexpressed HSPA6 to verify its effect on Wnt/β-catenin signaling. The result indicated that HSPA6 (Figures 4H,I) activated transcription of LEF1 ( Figure 4J) and exerted an inverse effect of miR-6721-5p on β-catenin expression (Figures 4K,L).

DISCUSSION
Wnt proteins are a protein family composed of secreted glycoproteins acting as signaling molecules. They bind to Frizzled and several other coreceptors such as lipoprotein receptor-related protein (LRP)-5/6 to trigger intracellular signal transduction (Logan and Nusse, 2004;Kestler and Kuhl, 2008). In the canonical Wnt/β-catenin signaling pathway, when Wnt signal is absent, β-catenin is phosphorylated by the destruction complex. Such processes include CK1-mediated phosphorylation at Ser45, and GSK3β-mediated phosphorylation at Ser33, Ser37, and Thr41. The phosphorylated β-catenin is degraded by the proteasome through the ubiquitin pathway. When Wnt ligands are available, a cascade initiated by the binding of Wnt to the Frzs is activated. The destruction complex consisting of axin, adenomatous polyposis coli (APC), and GSK3β is disassembled, resulting in the stabilization of β-catenin. The β-catenin molecules subsequently accumulate in the cytoplasm and are imported into the nucleus, where they serve as transcriptional coactivators of the TCF/LEF family transcription factors to regulate gene expression (Rao and Kuhl, 2010). The canonical Wnt/β-catenin signaling pathway has been proven to play important roles in multiple aspects of vertebrate heart development such as cardiac differentiation and right ventricular growth (Ai et al., 2007;Gessert and Kuhl, 2010).
The stability of cytosolic β-catenin is controlled by the destruction complex. As the member of the destruction complex, GSK3β acts as a negative regulatory factor. We found that overexpression of lnc-TSSK2-8 promoted the phosphorylation of GSK3β. Subsequently, the phosphorylation of β-catenin was decreased, which might inhibit its ubiquitination. We showed that lnc-TSSK2-8 promotes the stabilization of βcatenin through regulating the expression of HSPA6 and CRYAB. CRYAB is the most abundant small HSP constitutively expressed in cardiomyocytes. It protects against reperfusion injury/myocardial ischemia and suppresses cardiac hypertrophic responses by attenuating nuclear factor of activated T cells (NFAT) signaling (Kumarapeli et al., 2008). CRYAB was also identified as a downstream effector of calcineurin-induced protection against cardiomyocyte apoptosis (Bousette et al., 2010) and a molecular switch in bypassing mitochondrial pathway of apoptosis during myocardial infarction (Mitra et al., 2013). Zhu et al. (2020) recently revealed that CRYAB physically interacts with β-catenin; it promotes osteogenic differentiation of bone marrow stem cells by protecting β-catenin from ubiquitination and degradation (Zhu et al., 2020). HSPA6 has not been reported as associated with Wnt/β-catenin signaling. Our study showed that HSPA6 and CRYAB not only stabilized β-catenin but also activated the downstream transcription factor LEF1. Since dosage of β-catenin significantly affects OFT development upstream of Tbx1 (Racedo et al., 2017), we propose that the small HSP-mediated stabilization of β-catenin should be a common mechanism involved in heart development.
Significant achievements have been made in revealing the roles of miRNAs and lncRNAs in heart development (Alexanian and Ounzain, 2020;Das et al., 2020). Cytoplasmic lncRNA could act as ceRNAs to sequester miRNAs from their natural mRNA targets; we established that lnc-TSSK2-8 regulates HSPA6 and CRYAB in this way. Whereas lnc-TSSK2-8 also has substantial nuclear expression, the transcriptomic analysis indicated that lnc-TSSK2-8 might be associated with cilia functions. Cilia have an essential role in the pathogenesis of congenital heart disease (Li et al., 2015;Gabriel et al., 2020). We noted that the DEGs were enriched in dynein binding, which is critical for ciliary processes. The MF of lnc-TSSK2-8 in cilia process and heart development needs to be further investigated. Conotruncal defects such as TOF, truncus arteriosus, conoventricular ventricle septum defect, type B interruption of the aortic arch account for about 70% of heart malformations associated with 22q11.2DS (Agergaard et al., 2012). Although several genes including TBX1 have been identified responsible for heart defects resulting from 22q11.2 deletion, noncoding functions of this common CNV in cardiac development have not received enough attention. We demonstrated that the 22q11.2 lnc-TSSK2-8 is involved in the canonical Wnt signaling pathway by stabilizing βcatenin. There were two main limitations of our study. Firstly, our experiments were based on the HEK293 cell line, which is derived from human embryonic kidney. Additional validation using primary cardiac cells derived from model animals or in vitro cardiomyocyte differentiation from human iPS cells would help reveal the cardiac-specific effect of lnc-TSSK2-8. Secondly, loss-of-function validation was not performed for lnc-TSSK2-8, HSPA6, and CRYAB. It would be meaningful to carry out si/shRNA-mediated knockdown of lnc-TSSK2-8, HSPA6, and CRYAB to investigate their roles in the context of heart development, which would help elucidate the contribution of lnc-TSSK2-8 to cardiac phenotypes of 22q11.2DS.
In summary, we established that the 22q11.2 lncRNA lnc-TSSK2-8 acts as an ceRNA to regulate the expression of HSPA6 and CRYAB through sponging miR-6721-5p and miR-491-5p; subsequently, HSPA6 and CRYAB activates canonical Wnt signaling by promoting the stabilization of β-catenin and activating LEF1 expression. Our study provided evidence of a lncRNA that regulates the dosage of β-catenin, which contributes to the OFT anomalies in 22q11.2DS.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm. nih.gov/geo/, GSE165927.

AUTHOR CONTRIBUTIONS
JF, YG, XpZ, and XqZ performed the lab experiments. BW, MQ, and XyZ performed bioinformatic data analysis. BW wrote the manuscript. QF, YG, ZX, and BW conceived the study and contributed to editing. All authors contributed to the article and approved the submitted version.