Abstract
Objective:
The velopharyngeal mechanism is comprised of several muscular components that act in a coordinated manner to control airflow through the nose and mouth. Proper velopharyngeal function is essential for normal speech, swallowing, and breathing. The genetic basis of normal-range velopharyngeal morphology is poorly understood. The purpose of this study was to estimate the heritability of velopharyngeal dimensions.
Method:
We measured five velopharyngeal variables (velar length, velar thickness, effective velar length, levator muscle length and pharyngeal depth) from MRIs of 155 monozygotic and 208 dizygotic twin pairs and then calculated heritability for these traits using a structural equation modeling approach.
Results:
The heritability estimates were statistically significant (95% confidence intervals excluded zero) and ranged from 0.19 to 0.46. There was also evidence of significant genetic correlations between pairs of traits, pointing to the influence of common genetic effects.
Conclusions:
These results indicate that genetic factors influence variation in clinically relevant velopharyngeal structures.
Keywords: Genetics, Structural Equation Modeling, Uvula, Soft Palate, MRI
Introduction
The velopharyngeal (VP) mechanism is comprised of several muscular components including the velum (soft palate) and the lateral and posterior walls of pharynx (throat). These muscles act in a coordinated manner to control airflow through the nose and mouth, with the highly mobile velum acting as a seal through contact with the pharyngeal walls. Proper VP function is essential for normal speech, swallowing, and breathing. VP dysfunction (VPD) is a common consequence of orofacial clefts, particularly when these defects extend to the hard and/or soft palate, disrupting the structural integrity of VP apparatus (Woo et al., 2012). In addition, neuromuscular lesions can also impact VP function.
Because the VP mechanism involves the precise spatial integration of numerous morphological components, the formation of these components is likely guided by coordinated genetic signals. Most genetic studies, to date, have focused on the soft palate. Recent molecular studies of soft palate development in mice have revealed a network of cell type-specific genetic pathways involving Wnt, Shh, Tgf-β and Fgf signaling (Janečková et al., 2019; Li et al., 2019). As with other complex morphological traits, many genes are likely involved in determining the size, shape, and position of the various VP components. Moreover, these same genes may impact VP dysmorphology and the resulting dysfunction.
Estimating the heritability of VP morphology is an important first step toward understanding the genetic basis of these traits. Heritability can be defined as the proportion of trait variation attributable to genetics. Several previous family-based studies have shown moderate heritability for general measures of upper airway size (Schwab et al., 2006; Patel et al., 2008; Chi et al., 2014; Al-Qawasmi et al., 2019). A recent twin study revealed evidence of a genetic contribution to several cephalometry-derived measures of the pharynx, with the highest contribution observed for the distance between the base of the tongue and the posterior pharyngeal wall and for the length of the soft palate (Kang et al., 2018).
While these studies provide some evidence that relevant aspects of VP morphology are heritable, they are limited by small sample sizes, lack of detailed phenotyping, and/or technical biases in the heritability estimation. In this study, we measured several key dimensions of the VP apparatus from 3D MRIs in a large sample set of monozygotic (MZ) and dizygotic (DZ) twins and then used this information to calculate heritability estimates and the generic correlations among traits with a structural equation modeling approach.
Materials and Method
Study cohort
For this analysis, we utilize existing data collected as part of the NIH-funded Adolescent Brain and Cognitive Development (ABCD) study, a decade-long 21-site collaborative longitudinal data collection effort (Jernigan et al., 2018). To date, the ABCD study has recruited 11,875 boys and girls at 9 and 10 years of age and has made full-head MRI scans available to the research community through the controlled-access NIMH Data Archive (Uban et al. 2018; Casey et al., 2018). The multi-ethnic dataset also includes detailed demographic data and developmental/medical history information (Barch et al., 2018). A subset of the larger cohort is comprised of monozygotic (MZ) and dizygotic (DZ) twin pairs recruited at four sites: University of Minnesota, University of Colorado, Virginia Commonwealth University, and Washington University. The ABCD study’s twin recruitment target is 860 pairs, recruited through birth registries (Iancono et al. 2018). At the time of the current study, MR images were available for 363 twin pairs (155 MZ and 208 DZ), and this twin subset formed the basis for our heritability analysis. Zygosity was determined by estimating the proportion of the genome identity by descent (IBD) using ABCD whole-genome genotype data in PLINK (Purcell et al., 2007). MZ twins had IBD proportions ranging from 0.94 to 1, and DZ twins had IBD proportions ranging from 0.41 to 0.61. The guidelines of the World Medical Association Declaration of Helsinki were followed, and institutional ethical approval was obtained at each ABCD study recruitment site. Individuals enrolled in the ABCD study provided their informed consent for broad research use and data sharing through the NIMH Data Archive. In addition, institutional ethics approval was obtained to cover the data access and secondary data analyses described in this study at the University of Pittsburgh and East Carolina University.
MR imaging
Structural 3T MRI scans were obtained on each study participant as part of the ABCD study. The MRI acquisition protocol was harmonized to be compatible across three different 3T scanner platforms: Siemens Prisma (Erlangen, Germany), General Electric 750 (Milwaukee, Wisconsin), and Philips (Koninklijke Philips, Netherlands). The structural brain scanning protocol includes 3D T1 and T2 weighted images and diffusion weighted images. A detailed description of the scanning parameters and the scanner harmonization process is described by Casey et al. (2018).
Phenotyping
Digital Imaging and Communications in Medicine (DICOM) raw data were imported into Amira 5.4.0 Visualization and Volume Modeling Software (Visage Imaging, GmbH, Berlin, Germany). Three-dimensional MRI data were analyzed to produce a serial stack of images along the midsagittal and oblique coronal planes for each participant. In this current study, we measured five VP variables: velar length, velar thickness, effective velar length (distance from posterior nasal spine to insertion of the levator veli palatini muscle into the velum), levator veli palatini muscle length, and pharyngeal depth measured as the distance between the posterior nasal spine (PNS) and the posterior pharyngeal wall (PPW). These measures are described in Table 1 and shown in Figure 1. These variables were chosen because of their clinical relevance for speech production. Specifically, these measures are related to VP function and establishing normal oral-nasal balance among speakers. To assess reliability, a primary and secondary rater with experience in MRI data analyses randomly selected and re-measured 35% of the participants five months after the first measures were obtained. Inter-rater and intra-rater reliability ranged from r = 0.82 to r = 0.94 for the velopharyngeal measures. A two-way intraclass correlation coefficient mixed effects model was 0.978 (p < .001) demonstrating excellent correlation between repeated measures.
Table 1.
Descriptions of VP measurements.
| Measure | Description |
|---|---|
| Velar length | Curvilinear length (mm) of the velum extending from the posterior border of the hard palate through the middle of the velum to the inferior tip of the uvula from the sagittal image plane |
|
| |
| Velar thickness | Distance (mm) between oral surface of velum and velar knee at the estimated location of the velar knee (region of greatest velar thickness) and velar dimple (oral location) as seen on the midsagittal image plane |
|
| |
| Effective velar length | Functional portion of the velum measured as the distance (mm) between the posterior border of the hard palate and point of levator muscle insertion into the velum as seen on the sagittal image plane |
|
| |
| Levator muscle length | Average of right and left levator muscle length measures (mm) from origin at cranial base to terminal end of the levator bundle as seen on the oblique coronal image plane |
|
| |
| Pharyngeal depth (PNS-PPW) |
Distance (mm) from the posterior nasal spine (PNS) to the posterior pharyngeal wall (PPW) along the hard palate plane as seen on the sagittal image plane |
Figure 1.
Example MRI showing the VP measurements. PNS-PPW = pharyngeal depth. Levator length is the average of the left and right levator lengths (left panel).
Heritability and genetic correlation analysis
To estimate heritability in the ABCD twin data, we used a structural equation modeling approach in OpenMx ver. 2.17.2 (Neale et al., 2015). This variance components approach partitions the phenotypic variance into additive genetic effects (A), common environmental effects (C), dominance genetic effects (D), and/or unique environmental effects and measurement errors (E). In this modeling framework, the pathways between latent variables representing the additive genetic variance and covariance are set at 50% for DZ twins and 100% for MZ twins, reflecting the average number of shared alleles. The pathways between latent variables representing the dominance genetic variance and covariance are set at 25% for DZ twins and 100% for MZ twins, reflecting the expected value of sharing two alleles per locus. The pathways between latent variables representing the common environmental variance and covariance are set to 100% for both DZ and MZ twins. The pathways between latent variables representing the non-shared, unique environmental variance and covariance are set to 0% for both DZ and MZ twins. We initially considered ACE or ADE models. The ACE model was used to estimate A, C, and E components for traits in which the DZ twin correlation was more than half the MZ twin correlation. This model was chosen as the most parsimonious model given that additive variance usually explains most of the genetic variance for anthropometric traits. The ADE model was used to estimate A, D, and E components for traits in which the DZ twin correlation was less than half the MZ twin correlation, given such a relationship between the DZ and MZ twin correlations suggests there may exist a non-negligible non-additive component to the genetic variance. We extended this variance components framework to model pairs of two traits simultaneously in order estimate the genetic correlation (ρG) between them using the bivariate ACE or ADE model with Cholesky decomposition. Genetic correlation was interpreted as the degree to which a pair of two VP measurements were influenced by common genetic effects. After model fitting using a goodness-of-fit statistic, we determined that C or D components could be omitted from the final models. Our final model included only A and E components (shown in Figure 2).
Figure 2.
(a) Univariate and (b) Bivariate AE models showing the variance components for heritability and genetic correlation. P = phenotypes/traits; A = latent variable for additive genetic effects; E = latent variable for environmental effects; a = variance and covariance of additive genetic effects; e = variance and covariance of environmental effects. The pathways between latent variables representing the additive genetic variance and covariance are set at 1.0 for MZ and 0.5 for DZ twins, and the pathways between latent variables representing the unique environmental variance and covariance are set a 0.0 for both MZ and DZ twins. The univariate AE module allows the estimation of the proportion of variance due to the additive genetic component, which is the heritability. The bivariate AE model with Cholesky decomposition allows estimation of the genetic correlation among the additive genetic components.
Results
The sample characteristics and summary statistics for five VP traits are presented in Table 2. For the five measured VP traits, the DZ and MZ groups had almost identical means and standard deviations. Table 3 shows heritability estimates (h2) and genetic correlations (ρG) among the VP traits. In general, the VP traits showed moderate heritability, with estimates ranging from a low of 0.19 (95% CI: 0.06–0.31) for effective velar length to a high of 0.46 (95% CI: 0.34–0.56) for pharyngeal depth. There was evidence of significant genetic correlation between velar length and several other VP traits. The strongest genetic correlation was observed between velar length and effective velar length (ρG = 0.796).
Table 2.
Characteristics and summary statistics for the Twin sample.
| MZ | DZ | Total | |
|---|---|---|---|
| Twin pairs: n | 155 | 208 | 363 |
| Sex (male/female): n | 150/160 | 191/225 | 341/385 |
| Age in months: mean (range) | 123.6 (108–132) | 122.7 (108–132) | 123.1 (108–132) |
| Velar length: mean (sd) | 28.36 (3.47) | 28.45 (3.66) | 28.41 (3.58) |
| Velar thickness: mean (sd) | 9.44 (1.27) | 9.46 (1.23) | 9.45 (4.66) |
| Effective velar length: mean (sd) | 11.89 (2.65) | 11.85 (2.50) | 11.87 (1.25) |
| Levator muscle length: mean (sd) | 38.96 (2.84) | 38.92 (3.20) | 38.94 (3.05) |
| Pharyngeal depth: mean (sd) | 19.50 (4.38) | 19.47 (4.87) | 19.48 (2.56) |
Table 3.
Heritability estimates and genetic correlations among the VP traits.
| Heritability estimates (h2 (95% CI)) |
Genetic correlation (ρG (95% CI)) |
||||
|---|---|---|---|---|---|
| Velar length | Velar thickness | Effective velar length |
Levator muscle length | ||
| Velar length | 0.43 (0.31–0.54) | ||||
| Velar thickness | 0.35 (0.23–0.47) |
0.316 (0.0765, 0.5754) |
|||
| Effective velar length |
0.19 (0.06–0.31) |
0.796 (0.5874, 1) |
−0.057 (−0.4413, 0.3440) |
||
| Levator muscle length | 0.37 (0.23–0.49) |
0.281 (0.0292, 0.5662) |
0.282 (0.0119, 0.5662) |
1* (−0.1099, 1) |
|
| Pharyngeal depth | 0.46 (0.34–0.56) | 0.087 (−0.1465, 0.2873) |
−0.018 (−0.2619, 0.2290) |
0.313 (−0.0273,0.6019) |
0.104 (−0.1409, 0.3644) |
Bold values indicated statistical evidence that the 95% CI excluded zero.
This correlation between these two variables was not statistically significant because one of the variances (a22) was close to zero.
Discussion
The goal of this study was to measure the genetic contribution to several clinically relevant VP traits in a cohort of twins. While several earlier family-based or twin studies concluded that the overall size of the pharyngeal region is heritable (Schwab et al., 2006; Patel et al., 2008; Chi et al., 2014; Al-Qawasmi et al., 2019), relatively few studies have focused on specific anatomical components of the VP mechanism. Kang et al. (2018) measured several pharyngeal dimensions in their sample of twins and reported heritability estimates, with length of the soft palate showing the highest heritability. The sample size for this study, however, was very small (36 twin pairs) and relied on 2D radiography, which is not optimal for the measurement of soft tissue structures. Using MRIs from a considerably larger sample, we found evidence that all five VP traits in our study were moderately heritable. Similar to Kang et al. (2018), we observed significant heritability estimates for soft palate (velar) length, but in contrast we also reported significant heritability for pharyngeal depth (PNS-PPW). The magnitude of the heritability estimates, which ranged from 19% to 46%, indicate that VP traits in adolescents are moderately heritable. Given our final heritability models included latent variables representing only the additive genetic and unique environmental sources of variation, this finding indicates that a substantial amount of variation in the VP traits was due to unknown sources of variation. We speculate that some of this non-genetic variation may be attributable to fetal exposures, as these have been shown to impact palatal development in experimental model, as well as other childhood exposures, environmental factors, and measurement error.
In addition to the moderate heritability, we observed significant genetic correlations among some of our VP traits, suggesting that these traits may be influenced by some common genetic factors. The high genetic correlation between velar length and effective velar length may be explained by the fact that they are both length measures of the soft palate. Only pharyngeal depth did not show a genetic correlation with any other trait, which suggests that different genetic variants may influence pharyngeal depth compared to the other traits. Pharyngeal depth is the only trait included that involves the posterior pharyngeal wall, and one possible explanation is that this anatomical region may be under separate genetic control compared to the rest of the VP complex. Future gene mapping studies may be able to shed more light on the specific genetic variants involved in different VP traits.
Establishing the heritability of traits is an important step toward understanding their underlying genetic architecture. Our results suggest that the VP traits included here are amenable to large-scale gene mapping efforts, such as genome-wide association studies (GWAS). These kinds of studies can point us to specific genes and pathways involved in the formation of the VP mechanism. This is important because the genes that influence normal VP morphology may also be involved in anatomically-based VP dysfunction and in malformations like orofacial clefting, both of which are characterized by unclear etiology.
It is important to highlight that we did not have access to measures of VP function because the database includes only structural static MRI. Therefore, we must be cautious about extrapolating from the heritability of individual VP structures to the heritability of the VP-related activities and mechanics. To assess these aspects, additional studies that use dynamic MRI methods will be needed. Additional limitations in the current study included sample size, which while large compared to some earlier VP heritability studies, may still be underpowered. As more data become available, we should see improved precision of these heritability estimates.
Acknowledgments
Data used in the preparation of this article were obtained from the Adolescent Brain Cognitive Development (ABCD) Study (https://abcdstudy.org), held in the NIMH Data Archive (NDA). This is a multisite, longitudinal study designed to recruit more than 10,000 children age 9–10 and follow them over 10 years into early adulthood. The ABCD Study is supported by the National Institutes of Health and additional federal partners under award numbers U01DA041048, U01DA050989, U01DA051016, U01DA041022, U01DA051018, U01DA051037, U01DA050987, U01DA041174, U01DA041106, U01DA041117, U01DA041028, U01DA041134, U01DA050988, U01DA051039, U01DA041156, U01DA041025, U01DA041120, U01DA051038, U01DA041148, U01DA041093, U01DA041089, U24DA041123, U24DA041147. A full list of supporters is available at https://abcdstudy.org/federal-partners.html. A listing of participating sites and a complete listing of the study investigators can be found at https://abcdstudy.org/consortium_members/. ABCD consortium investigators designed and implemented the study and/or provided data but did not necessarily participate in analysis or writing of this report. This manuscript reflects the views of the authors and may not reflect the opinions or views of the NIH or ABCD consortium investigators.
This work is supported by grants from the National Institutes of Health.
Footnotes
Conflict of Interest
The authors have no conflicts of interest to declare.
References
- Al-Qawasmi R, Parsons S, Wetherill L. Heritability of the pharyngeal airway volume and dimensions as assessed from siblings with overt malocclusions. Int Orthod 2019;17:660–666. [DOI] [PubMed] [Google Scholar]
- Barch DM, Albaugh MD, Avenevoli S, Chang L, Clark DB, Glantz MD, Hudziak JJ, Jernigan TL, Tapert SF, Yurgelun-Todd D, Alia-Klein N, Potter AS, Paulus MP, Prouty D, Zucker RA, Sher KJ. Demographic, physical and mental health assessments in the adolescent brain and cognitive development study: Rationale and description. Dev Cogn Neurosci 2018;32:55–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Casey BJ, Cannonier T, Conley MI, Cohen AO, Barch DM, Heitzeg MM, Soules ME, Teslovich T, Dellarco DV, Garavan H, Orr CA, Wager TD, Banich MT, Speer NK, Sutherland MT, Riedel MC, Dick AS, Bjork JM, Thomas KM, Chaarani B, Mejia MH, Hagler DJ Jr, Daniela Cornejo M, Sicat CS, Harms MP, Dosenbach NUF, Rosenberg M, Earl E, Bartsch H, Watts R, Polimeni JR, Kuperman JM, Fair DA, Dale AM; ABCD Imaging Acquisition Workgroup. The Adolescent Brain Cognitive Development (ABCD) study: Imaging acquisition across 21 sites. Dev Cogn Neurosci 2018;32:43–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chi L, Comyn FL, Keenan BT, Cater J, Maislin G, Pack AI, Schwab RJ. Heritability of craniofacial structures in normal subjects and patients with sleep apnea. Sleep 2014;37:1689–1698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Iacono WG, Heath AC, Hewitt JK, Neale MC, Banich MT, Luciana MM, Madden PA, Barch DM, Bjork JM. The utility of twins in developmental cognitive neuroscience research: How twins strengthen the ABCD research design. Dev Cogn Neurosci 2018;32:30–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Janečková E, Feng J, Li J, Rodriguez G, Chai Y. Dynamic activation of Wnt, Fgf, and Hh signaling during soft palate development. PLoS One 2019;14:e0223879. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jernigan TL, Brown SA, ABCD Consortium Coordinators. Introduction. Dev Cogn Neurosci 2018;32:1–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kang JH, Sung J, Song YM, Kim YH. Heritability of the airway structure and head posture using twin study. J Oral Rehabil 2018;45:378–385. [DOI] [PubMed] [Google Scholar]
- Li J, Rodriguez G, Han X, Janečková E, Kahng S, Song B, Chai Y. Regulatory mechanisms of soft palate development and malformations. J Dent Res 2019;98:959–967. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Neale MC, Hunter MD, Pritikin JN, Zahery M, Brick TR, Kirkpatrick RM, Estabrook R, Bates TC, Maes HH, Boker SM. OpenMx 2.0: Extended structural equation and statistical modeling. Psychometrika 2016;81:535–549. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Patel SR, Frame JM, Larkin EK, Redline S. Heritability of upper airway dimensions derived using acoustic pharyngometry. Eur Respir J 2008;32:1304–1308. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, Maller J, Sklar P, de Bakker PI, Daly MJ, Sham PC. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007;81:559–575. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schwab RJ, Pasirstein M, Kaplan L, Pierson R, Mackley A, Hachadoorian R, Arens R, Maislin G, Pack AI. Family aggregation of upper airway soft tissue structures in normal subjects and patients with sleep apnea. Am J Respir Crit Care Med 2006;173:453–463. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Uban KA, Horton MK, Jacobus J, Heyser C, Thompson WK, Tapert SF, Madden PAF, Sowell ER, Adolescent Brain Cognitive Development Study. Biospecimens and the ABCD study: Rationale, methods of collection, measurement and early data. Dev Cogn Neurosci 2018;32:97–106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Woo AS. Velopharyngeal dysfunction. Semin Plast Surg 2012;26:170–177. [DOI] [PMC free article] [PubMed] [Google Scholar]