Developmental Changes in Cardiac Expression of KCNQ1 and SCN5A Spliceoforms: Implications for Sudden Unexpected Infant Death

Abstract

Background

Sudden, Unexpected Infant Death (SUID) occurs unpredictably and remains unexplained after scene investigation and autopsy. Approximately 1 in 7 cases of SUID can be related to a cardiac cause, and developmental regulation of cardiac ion channel genes may contribute to SUID.

Objective

The goal of this study was to investigate the developmental changes in the spliceoforms of SCN5A and KCNQ1, two genes implicated in SUID.

Methods

Using RT-qPCR, we quantified expression of SCN5A (Adult and Fetal) and KCNQ1 (KCNQ1a and b) spliceoforms in 153 human cardiac tissue samples from decedents that succumbed to SUID (‘Unexplained’) and other known causes of death (‘Explained Non-cardiac’).

Results

There is a stepwise increase in the Adult/Fetal SCN5A spliceoform ratio from <2 months (4.55±0.36, n=51) through infancy and into adulthood (17.41±3.33, n=5). For KCNQ1, there is a decrease in the ratio of KCNQ1b to KCNQ1a between the <2 months (0.37±0.02, n=46) and the 2-4 months (0.28±0.02, n=52) age groups. When broken down by sex, race, or cause of death, there were no differences in SCN5A or KCNQ1 spliceoform expression, except for a higher ratio of KCNQ1b to KCNQ1a at 5-12 months of age for SUID females (0.40±0.04, n=9) compared to males (0.25±0.03, n=6) and at <2 months of age for SUID White (0.42±0.03, n=19) compared to Black (0.33±0.05, n=9) infants.

Conclusion

This study documents the developmental changes in SCN5A and KCNQ1 spliceoforms in humans. Our data suggest that spliceoform expression ratios change significantly throughout the first year of life.

INTRODUCTION

There are roughly 3,500 cases of Sudden Unexpected Infant Death (SUID) in the United States each year, and about 75% remain unexplained after autopsy and extensive post-mortem investigation¹. While up to 20% of autopsy-negative Sudden Infant Death Syndrome (SIDS; a subset of SUID) cases have underlying cardiac channelopathy- or cardiomyopathy-causing mutations, it is difficult to distinguish the pathogenicity of specific genetic variants as it is likely that multiple factors (genetic, environmental, developmental) interact to create the ideal environment for SUID to occur^2-4.

The incidence of sudden, unexpected death is much higher in the infant (<12 month), as compared to the non-infant (>12 month), population, peaking at 2-4 months of age^5,6. There is evidence suggesting that developmental changes in ion channel genes may play a role in SUID^7,8. Two ion channel genes that have been implicated in SUID include the cardiac sodium channel gene SCN5A and the potassium channel gene KCNQ1, both of which encode two main isoforms that can be incorporated into their channel structures. Functional studies have shown that the non-canonical isoforms result in different electrophysiological properties, and it has been hypothesized that these alternatively spliced isoforms may lead to increased arrhythmogenicity^8,9.

The two spliceoforms of SNC5A include the canonical “Adult” transcript, which includes exon 6B, and the alternatively spliced “Fetal” transcript, which includes exon 6A and differs by 7 amino acids⁸. Previous functional studies revealed that the Fetal isoform results in slower channel activation and inactivation kinetics, decreased current amplitude, and increased late sodium current^8,10,11. Thus, when the Fetal isoform is highly expressed, the heart may be predisposed to fatal arrhythmias when combined with mutations or other factors that create a favorable environment for sudden cardiac death to occur. The two spliceoforms of KCNQ1 include KCNQ1a, the full-length, canonical spliceoform, and KCNQ1b, the alternatively spliced, N-terminally truncated spliceoform that does not include exon 1¹². In comparison to KCNQ1a, expression of KCNQ1b elicits functional differences in human cardiac tissue, resulting in decreased potassium currents, prolonged repolarization, and longer action potential duration^13,14. However, the developmental changes in each KCNQ1 spliceoform has yet to be investigated.

We tested the hypothesis that KCNQ1 and SCN5A spliceoform expression levels are age-dependent and aimed to investigate whether there are differences in spliceoform expression levels between infants with Unexplained versus Explained Non-cardiac causes of death. Given that the rates of SUID are higher in males and the Non-Hispanic Black population, we also aimed to compare spliceoform expression levels based on sex and race⁵.

METHODS

Study Population

De-identified data from 580 decedents were collected in collaboration with Forensic Medical Management Services, the regional forensic center for Middle Tennessee. Autopsy reports were reviewed for key characteristics, including age at death, race, sex, and details surrounding death, as indicated in Table 1. ‘Unexplained’ subjects were defined as previously healthy decedents with sudden, unexplained deaths and unremarkable autopsy findings, and therefore served as “cases.” ‘Explained Non-cardiac’ samples were defined as decedents with explained, known non-cardiac causes of death (i.e., trauma, acute illness, homicide), and these subjects served as “controls.” Samples from patients with known cardiac causes of death (i.e., congenital heart disease, hypertrophic cardiomyopathy, etc.) were excluded. Finally, ‘Possible Unexplained’ samples were defined as decedents with sudden, unexplained, normal autopsy death, with contributory circumstances including possible asphyxiation or suffocation, and were included in overall age group analyses, but not in “case-control” comparisons (Fig 1). ‘Sudden’ death was defined as death within 24 hours of the first symptom, or death in the hospital after resuscitation from an out-of-hospital cardiac arrest. ‘Unexpected’ was defined as the death of someone who was believed to be in good health or had a stable, chronic condition, or had an acute illness that would not be expected to cause death. The majority of samples were procured from decedents under 1 year of age, as we focused primarily on spliceoform expression during the first year of life in our experiments. All non-infant (child >12 months and adult >18 years) decedents had known, non-cardiac causes of death. The project was reviewed by the Vanderbilt University Institutional Review Board, which determined the study does not qualify as “human subject” research based on HHS regulation 45 CFR 46.102(f) (IRB# 121853). The research reported in this paper adhered to guidelines of the Helsinki Declaration as revised in 2013.

Table 1.

Clinical characteristics of study population at autopsy.

Characteristic	Unexplained	Explained Non-Cardiac	Possible Unexplained
Age at Death
Fetal	-	8	-
Infant (<2 months)	31	10	3
Infant (2-4 months)	35	9	14
Infant (5-12 months)	15	9	6
Child (13-24 months)	-	7	-
Adult	-	6	-
Total	81	49	23
Sex
Fetal
Male	-	2 (25%)	-
Female	-	6 (75%)	-
Infant (0-12 month)
Male	44 (54%)	14 (50%)	17 (74%)
Female	37 (46%)	14 (50%)	6 (26%)
Child and Adult
Male	-	5 (38%)	-
Female	-	8 (62%)	-
Race
Fetal
Black	-	3 (38%)	-
White	-	5 (62%)	-
Infant (0-12 month)
Black	26 (32%)	7 (25%)	5 (22%)
White	53 (65%)	20 (71%)	18 (78%)
Other/Unknown
Child and Adult	2 (3%)	1 (4%)	-
Black	-	4 (31%)	-
White	-	9 (69%)	-
Details surrounding death (Infants only)
Co-sleeping/unsafe sleep environment*	52 (64%)	3 (11%)	22 (96%)
Recent illness**	3 (4%)	7 (25%)	2 (9%)
Premature birth	14 (17%)	4 (14%)	4 (17%)

^*Unsafe sleep environment defined as sleep environment deviating from safe sleep recommendations from the American Academy of Pediatrics, including placing infant alone in crib or bassinet on back on a firm surface, without crib bumpers, blankets, pillows, stuffed animals, or other bedding.

^**recent illness defined as known, documented illness within one week of death.

Figure 1.

Overview of experimental design and study group definitions.

Sample Collection

Human cardiac tissue samples along with de-identified autopsy reports were obtained from Forensic Medical Management Services (Nashville, TN). All decedents’ samples were collected between 2013 and 2017. Each sample was harvested within 24-48 hours of death and stored in an airtight test tube in a −80°C freezer until ready to be analyzed. Three of the adult cardiac tissue samples were obtained from unmatched donor hearts with consent for research use through Tennessee Donor Services and were collected at the time of organ harvest and stored in the same manner.

RNA Isolation

Cardiac tissue was homogenized in Qiazol reagent (Qiagen, Hilden, Germany) using a microbead homogenizer (Omni, Kennesaw, Georgia). Chloroform extraction and purification of RNA were performed using the RNeasy Mini Kit (Qiagen, Hilden, Germany) on the QIAcube according to manufacturer’s instructions. RNA Integrity Number (RIN) was measured for each sample using Qubit RNA HS Assay Kits and Qubit Fluorometer protocols as directed (Thermo Fisher, Waltham, Massachusetts). All samples included in this study had a RIN of 5.0 or greater.

qRT-PCR

2ug of total RNA was used for cDNA synthesis using the QuantiTect Reverse Transcription kit (Qiagen, Hilden, Germany). The cDNA was then stored at 4°C. qRT-PCR was carried out using the Bio-Rad CFX384, with initial denaturation at 95°C for 10 minutes, followed by 39 cycles of denaturation at 95°C for 1 minute, primer annealing at 60°C for 30 seconds, and primer extension at 72°C for 5 minutes. qRT-PCR reactions were run in triplicate per sample, then averaged and normalized to GAPDH. Expression levels were analyzed using the 2-ΔΔCt method to calculate relative fold changes between groups. Samples with Ct values greater than 35 were considered undetected and were therefore not included in analysis. In addition, samples with fold changes more than 2 standard deviations outside of the mean for that group were also excluded from analysis. In the SCN5A experiments, 136 samples from decedents aged 0-12 months were included in final analyses, along with 7 samples from decedents aged 13-24 months and 5 from adult decedents. In the KCNQ1 experiments, 127 samples from decedents aged 0-12 months were included in final analyses, along with 6 samples from decedents aged 13-24 months and 6 from adult decedents.

For SCN5A expression studies, we designed primers that were sequence-validated to distinguish between the Fetal versus Adult spliceoforms, flanking exons 5 and 6A or exons 5 and 6B, respectively (Fig 2A): Forward primers were common to both spliceoforms: 5’-CTTCACCGCCATTTACACCT-3’; Reverse (Fetal-only): 5’-AAGAGCCGACAAATTGCCTA-3’; Reverse (Adult-only): 5’-CCCAGGTCCACAAATTCAGT-3’. For KCNQ1 expression studies, we designed primers that were sequence-validated to distinguish between the full-length (KCNQ1a) versus the truncated (KCNQ1b) spliceoform (Fig 3A): Forward (KCNQ1a-only): 5’-GCCGCGTCTACAACTTCCT-3’; Forward (KCNQ1b-only): 5’-TTTCTGGCTCTCGGGAATTT-3’; Reverse primers were common to both spliceoforms: 5’-GACAGCACGCTGAAGATGAG-3’. Primers used to detect GAPDH were: Forward 5’-TGGAAGGACTCATGACCACA-3’; Reverse 5’-GAGGCAGGGATGATGTTCTG-3’.

Figure 2. Fetal SCN5A decreases with age.

A) Alternative splicing pattern schematic of the Fetal and Adult SCN5A spliceoforms. Locations of qRT-PCR primers used to detect each spliceoform are indicated in red. B) qRT-PCR quantification of SCN5A spliceoforms across all age groups. 2-way analysis of variance (ANOVA) Kruskal-Wallis; P<0.0001, N=140. C) Fold change of SCN5A spliceoforms for Unexplained vs Explained Non-cardiac samples in all infant age groups (0-12 months). Unpaired Student’s T-test; Mean ± SEM = 7.321 ± 0.8626, 6.359 ± 0.3296; P=0.5986; N=27, 79. Fold change of SCN5A spliceoforms in Unexplained samples broken down by D) sex and E) race. N = as indicated by numbers within bars. Unpaired Student’s T-tests. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. Error bars = SEM.

Figure 3. Truncated KCNQ1 (KCNQ1b) expression is lowest at 2-4 months of age.

A) Alternative splicing pattern schematic of the two KCNQ1 spliceoforms, KCNQ1a and KCNQ1b. Locations of qRT-PCR primers used to detect each spliceoform are indicated in red. B) qRT-PCR quantification of KCNQ1 spliceoforms for all age groups. ANOVA, Kruskal-Wallis; P=0.0031, N=139. Fold change of KCNQ1 spliceoforms for Unexplained vs Explained Non-cardiac samples in infant age groups: C) 0-12 months (Unpaired Student’s T-test; Mean ± SEM = 0.3106 ± 0.02368, 0.3279 ± 0.01685; P=0.7983; N=27, 75) and D) 0-4 months (Unpaired Student’s T-test; Mean ± SEM = 0.3001 ± 0.01948, 0.3256 ± 0.01947; P=0.9346; N=16, 60). Gray dashed line indicates 0.5 fold change. Fold change of KCNQ1 spliceoforms in Unexplained samples broken down by E) sex and F) race. N = as indicated by numbers within bars. Unpaired Student’s T-tests. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. Error bars = SEM.

Statistical analysis

We focused on the change in expression levels during the first year of life, and separated decedents into the following age groups according to the risk of SD: <2 months, 2-4 months, 5-12 months, 13-24 months, and adult (>18 years). Mann-Whitney t-tests were used for comparisons between 2 groups, and Kruskal-Wallis ANOVA with multiple comparisons was performed for comparisons between multiple groups. Prism software (GraphPad, San Diego, California) was utilized for statistical analyses. P<0.05 was considered statistically significant. All fold change values are reported as Mean ± SEM.

RESULTS

We obtained 580 human cardiac tissue samples with matching de-identified autopsy reports from Forensic Medical Management Services (Fig 1). We isolated RNA from the left ventricle of 222 of the samples that met the selection criteria; 162 samples had a RIN >5.0 and 153 samples had adequate qPCR results¹⁵ (Fig 1). We used GAPDH as an endogenous reference to normalize the sample loading variation. It has been reported that GAPDH is one of the stably expressed housekeeping genes in normal and diseased human heart.¹⁶ Our results also suggest that the GAPDH expression level is similar among different age groups as well as between cases and controls (data not shown). Table 1 summarizes the demographics and clinical characteristics of these subjects.

Comparing the relative expression of the Adult and alternatively spliced Fetal SCN5A spliceoforms (Fig 2A), we detected a significant stepwise increase in Adult over Fetal (Adult/Fetal) spliceoform expression with age. The average fold changes of Adult/Fetal SNC5A were significantly higher in the 5-12 month (9.66±0.71, N=30), the 13-24 month (12.10±1.06, N=7), and the adult (17.41±3.33, N=5) age groups in comparison to the <2 month (4.55±0.36, P<0.0001, N=51) and the 2-4 month (6.65 ±0.29, P<0.05, N=55) age groups (Fig 2B). Further, the average fold change in the 2-4 month age group was significantly higher than in the <2 month age group (P=0.0001, Fig 2B). No significant differences were detected between any age groups across the 5-24 month to adult age ranges (P>0.9). The significant increases in fold change were driven by the predominant expression of the Adult spliceoform, which is consistently expressed at the same level across all infant and child age groups and into adulthood, combined with the steady decrease in abundance of the Fetal spliceoform, which peaks in expression during the fetal stage and trends downward postnatally. There was no significant difference in SNC5A Adult/Fetal spliceoform expression between Unexplained (6.36±0.33) and Explained Non-cardiac (7.32±0.86, P=0.5986) samples in all infant age groups (Fig 2C). There were also no significant differences in Adult/Fetal SCN5A expression levels in Unexplained infant samples when separated by sex and race (P>0.05, Fig 2D and 2E).

Comparing the relative expression of KCNQ1a and the alternatively-spliced KCNQ1b variant (Fig 3A), we detected a significant decrease in the average fold changes of KCNQ1b over KCNQ1a (KCNQ1b/KCNQ1a) between the <2 months (0.37±0.02, N=46) and the 2-4 months (0.28±0.02, P=0.0026, N=52) age groups, but no significant differences between the 2-4 month age group and the 5-12 month (0.32±0.03, P>0.99, N=29), the 13-24 month (0.35±0.05, P>0.99, N=6), or the adult (0.44±0.08, P=0.29, N=6) age groups, nor between the <2 month age group and the 5-12 month (P=0.623), the 13-24 month (P>0.99), or the adult (P>0.99) age groups (Fig 3B). When assessed further, the fluctuations in fold change were determined to be driven by concurrent, opposing alterations in KCNQ1 expression levels – in <2 month samples, KCNQ1b expression trends downward, while KCNQ1a expression trends upward, in 2-4 month samples, expression levels of both remain stable, and in 5-12 month samples, KCNQ1b expression trends upward while KCNQ1a expression trends downward. These trends are responsible for the overall decrease in the KCNQ1b/KCNQ1a ratio during the 2-4-month age range. There was no significant difference between KCNQ1a and KCNQ1b expression in the Unexplained (0.32±0.02) versus Explained Non-cardiac (0.31±0.02, P=0.798) infant groups throughout 0-12 months (Fig 3C). When examined during the 0-4 month time period, during which the rates of SUID are highest and KCNQ1b/KCNQ1a expression hits a nadir, there was still no significant difference in KCNQ1b/KCNQ1a fold changes between Unexplained (0.30±0.02) and Explained Non-cardiac (0.33±0.02, P=0.935) infant groups (Fig 3D). However, there was a subset of Unexplained samples that had a KCNQ1b/KCNQ1a fold change greater than 0.5 (Fig 3D). When separated by sex, females had significantly higher KCNQ1b/KCNQ1a expression in the 5-12 month age group (Males = 0.25±0.03, Females = 0.40±0.04, P=0.0176; Fig 3E). When separated by race, White subjects had significantly higher KCNQ1b/KCNQ1a expression than Black subjects in the Unexplained <2 month age group (Black = 0.33±0.05, White = 0.42±0.03, P=0.0371; Fig 3F).

DISCUSSION

It has been hypothesized that there are important functional differences related to the expression of different isoforms of channelopathy- and cardiomyopathy-related proteins, which may contribute to the higher rates of sudden death observed specifically in the infant population. Because of the importance of both SCN5A and KCNQ1 in cardiac conduction, investigating the differences in the age-related spliceoform expression of these genes may give insight into why infants are uniquely susceptible to sudden death.

SCN5A plays a vital role in cardiac depolarization and has been implicated in Long QT Syndrome, Brugada Syndrome, and sudden cardiac death. SCN5A encodes a Fetal (includes exon 6A) and an Adult (includes exon 6B) isoform (Fig 2A). Exon 6 of SCN5A encodes the voltage-sensing domain, and 6A codes for a positively charged lysine at position 211, while 6B codes for a negatively charged aspartic acid at this position¹¹. This positive charge alteration, present in the Fetal isoform, leads to decreased peak sodium current and slower channel activation and inactivation, and creates a potential substrate for arrhythmia. It has previously been shown that a common polymorphism of SCN5A, H558R, significantly potentiates the sodium current in the presence of the Fetal spliceoform, however only a few studies have examined the developmental expression of the Fetal and Adult SCN5A spliceoforms^8,10. In 2012, Murphy, et al. first reported that the Fetal spliceoform was expressed ~1.5-fold higher than the Adult spliceoform in fetal human hearts, and rapidly decreased to 1:1 in infant hearts. In adulthood, the ratio of the two spliceoforms reversed, and the Adult spliceoform was expressed ~7.5-fold higher than the Fetal spliceoform ⁸. In 2018, Pang, et al. also described the transition of SCN5A spliceoform expression in mouse hearts, as exon 6A was detected at higher levels in the fetal mouse heart, while exon 6B was detected at higher levels in the adult mouse heart. While these prior studies illustrate the developmental switch from Fetal to Adult spliceoform expression in healthy cardiac tissue early in postnatal life, sample sizes were minimal and were not further broken down by age within the first year of life – the study by Murphy, et al. included 4 fetal, 4 infant, and 24 adult human samples, including cardiac samples only from White subjects with known causes of death.

This study is the largest investigation into developmental SCN5A spliceoform expression and supports previously reported findings, while further refining expression levels based on age, sex, and race^8,10. We detected a significant, stepwise increase in Adult/Fetal spliceoform expression with age, driven by decreased expression in Fetal SCN5A over developmental time. Further, the risk for SD during childhood drops after 4 months of age, corresponding to the decrease in Fetal SCN5A. According to our data, the Adult/Fetal SCN5A ratio reaches approximately 10:1 at 5-12 months of age and does not change significantly throughout childhood and into adulthood. Additional studies are needed to provide further evidence for the role of Fetal SNC5A in sudden cardiac death in infancy. While we originally hypothesized that higher Fetal SCN5A expression in the Unexplained group could further predispose a subset of infants to sudden cardiac death, we did not observe a significant difference in expression between Unexplained and Explained Non-cardiac groups in infancy.

KCNQ1, which plays an important role in cardiac repolarization, has also been associated with Long QT Syndrome and sudden cardiac death. The alternatively spliced isoform, KCNQ1b, has a dominant-negative effect on cardiac KCNQ1, leading to prolonged action potentials that could contribute to cardiac arrhythmias^13,14. Changes in the expression and relative ratio of KCNQ1 spliceoforms during the first year of life have not previously been reported. Our study shows that the expression ratio of KCNQ1b/KCNQ1a is lowest during 2-4 months of infancy, corresponding to the peak incidence of SUID (Fig 3B). Thus, in contrast to SCN5A, where we observe higher ratios of the more arrhythmogenic Fetal isoform in infancy, for KCNQ1 we observe lower levels of the more arrhythmogenic KCNQ1b isoform during the peak age for SUID. It is possible that the developmental changes in KCNQ1 expression does not contribute to SUID.

This study analyzes SCN5A and KCNQ1 expression levels in cardiac tissues from a large number of well-characterized subjects with thoroughly reviewed autopsy information, allowing for careful classification of subjects based on multiple characteristics including age, sex, race, and cause of death (Table 1). While we originally hypothesized that higher-than-baseline Fetal SCN5A and KCNQ1b may contribute to some cases of SUID, we failed to detect a difference among our “case” (Unexplained) and “control” groups (Explained Non-cardiac). However, the inability to detect a significant difference in spliceoform expression between our “cases” and “controls” could be due to several factors including: 1. low availability of samples with known, explained, non-cardiac causes of death to act as controls due to limited autopsy procedures with these decedents, 2. the inability to identify which ~20% of Unexplained death samples were due to sudden cardiac death and can therefore serve as proper “cases,” or 3. a lack of a biological difference between “cases” and “controls.” Continued collection of available Explained Non-cardiac samples from the Medical Examiner’s office, combined with identification of true sudden cardiac death samples within the Unexplained death population (sequencing cardiac genes to identify mutations) will be necessary to more fully investigate this hypothesis.

Limitations

An estimated 20% of SUID cases occur due to an underlying cardiac cause, however we do not know which of our Unexplained samples came from patients that died due to sudden cardiac death. Further, protein levels can differ from RNA levels, due to differences in translation rates, protein stability, and post-translational modifications, making it difficult to determine the exact functional alterations that result in vivo upon differential expression of KCNQ1 and SCN5A spliceoforms. Additionally, the stoichiometry of the KCNQ1-KCNE1 heterotetrameric protein complex is flexible and has been found to be dependent on the relative expression densities of both KCNQ1 and KCNE1, adding another layer of complexity to the role of each KCNQ1 spliceoform in overall ion channel function.¹⁷ Our cohort of unexplained SD victims were predominantly < 4 months of age and recent data suggest that rare variants in cardiac genes are more common in the 4-12 month age range.¹⁸

CONCLUSION

This is the largest study detailing the changes in expression of cardiac SCN5A and KCNQ1 spliceoforms throughout infancy and adulthood in humans. The SCN5A Adult/Fetal fold change increases steadily throughout infancy and childhood to adulthood. Developmental changes in the SCN5A Fetal isoform may contribute to the high incidence of sudden cardiac death in infants. The KCNQ1b/KCNQ1a fold change drops during the 2-4 month age range, corresponding with the peak incidence of SUID. Given the known functional differences resulting from the differential expression of the spliceoforms of SNC5A and KCNQ1, our data support the idea that the fetal and infant heart may be uniquely predisposed to sudden cardiac death when combined with cardiac mutations and/or other pro-arrhythmogenic factors.