Obstructive sleep apnea syndrome (OSAS) (OMIM #107650) affects 936 million adults worldwide (Benjafield et al., 2019) and is hallmarked by the interruption of or a significant decrease in airflow in the presence of respiratory effort during sleep, leading to increased sleepiness and significantly increases cardiovascular morbidity and mortality risks (Lévy et al., 2015). While the precise etiological factors for OSAS are not clear, disturbances in ventilatory control, craniofacial anatomy, and being overweight, as well as genetic factors are major contributors to OSAS (Peppard and Hagen, 2018).

Familial clusters of OSAS have been reported and in fact heritability and genetic factors have been estimated to contribute ∼40% to OSAS phenotype (Redline et al., 1995; Patel et al., 2008). Over the past few years, sequence variants in candidate genes seemingly associated with OSAS (as objectively quantified by AHI [(apnea–hypopnea index (AHI), defined as the number of apnea and hypopnea events per hour of sleep (American Academy of Sleep Medicine, 2014) (International Classification of Sleep Disorders, 2014) have been reported (Larkin et al., 2010), as well as variants identified in the course of admixture mapping (Wang et al., 2019) and genome wide association studies (GWAS) (Cade et al., 2021). However, few genes have been reported and validated as “OSAS genes” using these approaches (Mohit, Shrivastava, Chand, 2021). The aim of the current study was to identify pathogenic germline sequence variants that could be associated with predisposition and/or development of OSAS using whole-exome DNA sequencing of individuals from two Brazilian OSAS families.

The study encompassed individuals from two Brazilian families displaying OSAS in at least two generations (Figure 1). These families were referred to the School of Medicine, Universidade Federal de Minas Gerais. They underwent the OSAS evaluation protocol as previously described (Kapur et al., 2017). All OSAS diagnoses were based on the established criteria as specified by the American Academy of Sleep Medicine (2014) (International Classification of Sleep Disorders, 2014). Briefly, each participant underwent complete OSAS-focused clinical evaluation and polysomnography. The inclusion criteria were included: diagnosis of OSAS as defined by the American Academy of Sleep Medicine criteria (2014) and absence of craniofacial dysmorphism, genetic syndromes with OSAS as part of the spectrum of manifestations, drug and alcohol abuse, psychiatric disorders, and dementia; age 18–85 years. Participants with upper airway resistance syndrome (UARS) and central sleep apnea were excluded. In addition, dental data were collected such as the presence of self-reported bruxism, pain in masticatory muscles, noise upon movement of the temporomandibular joint (TMJ), tongue size, floor of mouth (i.e., sublingual space) U-shaped region, bordered inferiorly by the mylohyoid muscle, laterally by the gingiva overlying the lingual surface of the mandible, superiorly by the oral tongue, and posteriorly at the insertion of the anterior tonsillar pillar into the tongue. Genomic DNA was extracted from peripheral blood samples of participants using standard protocols (Lahiri and Nurnberger, 1991). The quality and quantity of each DNA sample were tested by NanoDrop ND -2000 UV-Vis Spectrophotometer (Thermofisher, Waltham, MA). The experimental protocols were approved by the Institutional Review Board at the Universidade Federal de Minas Gerais (CEP UFMG 2.980.453). All participants signed an informed consent.

In this study, three heterozygous sequence variants that seemingly exert a deleterious effect on protein function in three genes (COX20, PTPDC1, and TMOD4) were identified in two Brazilian OSAS families. COX20 is a nucleus-encoded protein, essential for the assembly of the mitochondrial respiratory chain complex IV (CIV)/cytochrome C oxidase (Szklarczyk et al., 2013). COX20 assists in the stabilization and translocation of newly synthesized COX2 C-terminal into the intermembrane space during CIV assembly and promotes the association between COX2 and metallochaperones SCO1/2, essential for the biogenesis of the cooper core in Complex IV (Lorenzi et al., 2018). Consequently, limiting COX20 levels was associated with unstable, degradable COX2 proteins (Bourens et al., 2014), resulting in impaired complex IV biogenesis and reduced respiratory capacity (Otero et al., 2018). Homozygous COX2 gene pathogenic variants lead to mitochondrial complex IV deficiency (OMIM# 220110), a disorder clinically hallmarked by neuromuscular findings such as hypotonia, dystonia, and ataxia (Szklarczyk et al., 2013; Doss et al., 2014). Polysomnography showed that individuals with central apnea and obstructive sleep apnea, display a decreased ventilatory drive in response to hypoxia and/or hypercapnia (Brunetti et al., 2021).

Although not located in the transmembrane domains of COX20, p.Glu9Lys is at a position of considerable importance for the stability and functionality of COX20. This variant is predicted (but shown conclusively) to disrupt protein structure causing the loss of an important hydrogen bond.

In family B, sequence variants in two genes seemingly underlie OSAS pathogenesis. While the precise function of the Protein tyrosine phosphatase domain-containing 1 (PTPDC1) protein is yet to be elucidated, its down regulation has been associated with phenotypes such as obesity (Zhu et al., 2020) and dementia (Wang et al., 2020). Functionally, PTPDC1 protein is involved in cilium assembly and degradation, directing fluid movement (motile cilia), sensing extracellular environment, and controlling signal transduction pathways (primary cilia). Thus, PTPDC1 depletion leads to an increase in cilia length, suggesting that it negatively regulates cilia length via increased anterograde transport (Lai et al., 2011).

The other variant present in family B is TMOD4, one of the four members of the Tropomodulin family, a group of proteins that bind to the pointed ends of actin filaments thereby stabilizing them and inhibiting filament elongation (Fischer & Fowler, 2003). TMOD4 gene is highly expressed in skeletal muscle and heart (Cox & Zoghbi, 2000). The Inactivating biallelic pathogenic sequence variants in the TMOD4 have been associated with myopathy (e.g., nemaline myopathy OMIM #609284) also reported in zebrafish-mutated model (Berger et al., 2014), suggesting that TMOD4 mutations may lead to OSAS relevant myopathy, in the intercostal muscles and/or diaphragm.

An OSAS -like phenotype has previously been reported in cases with nemaline myopathy (Sasaki et al., 1997; Cook & Berkowitz, 2005). Sasaki et al. (1997), reported the results of polysomnography in two nemaline myopathy (NM) cases who had apnea or irregular chest movements with hypercapnia that occurred only during REM sleep. Cook and Berkowitz (2005) also described a milder form of NM in a child who presented at polysomnography, severe obstructive sleep apnea with an oxygen saturation nadir of 55% during REM sleep. Thus, these single reports are in line with the notion that loss of function in TMDO4 may play a role in muscle strength and may contribute to OSAS pathogenesis.

The wild-type amino acids in COX20, PTPDC1 and TMOD4 are highly conserved between species, which demonstrates its evolutionary relevance. Additionally, it is interesting to notice that the variants observed are rare in general population, demonstrating that it could be related to less adaptive subjects.

Noteworthy, no studies have previously identified possible genes related to mendelian forms of OSA and none of the above cited variants or other variants in the three genes reported herein have been identified in previous OSAS-GWAS studies. A search in GWAS catalog (https://www.ebi.ac.uk/gwas/genes) for variants within 150 Kb of the variants reported in the current study led to no results associated with sleep apnea. A possible reason is that SNPs of modest effect are missed as they do not reach a significance threshold (Eichler et al., 2010). Non-etheless, among the 41 OSA-associated GWAS variants, some trace back to genes involved in cytoskeletal structure and dynamic functionality, such as CDH4 (a calcium-dependent cell-cell adhesion glycoprotein that plays a role in muscle development), similarly to the PTPDC1 and TMDO4 genes mutated in Family B, which exhibit mainly structural functions important for muscle cell functionality. Furthermore, many of the GWAS OSA-associated variants display a pattern of participating in neuromuscular physiologic pathways and neurogenesis, such as RMST, CAMK1D and ATP2BA (Cade et al., 2016; Strausz et al., 2021). So far, most of the studies on OSA have focused on the effect of metabolic disorders such as obesity and environmental factors in the pathogenesis of OSA. However, the present finding of a syndromic form of OSA related to three genes with possible important roles in neuromuscular development and functionality, along with the existence of GWAS variants participating in similar pathways suggests that organic alterations and genetic susceptibility may play a pivotal role on the pathogenesis of OSA.

The limitations of our study should be acknowledged, such as lack of functional data, and limitations of exome sequencing for identification of non-coding and structural variants. Furthermore, the number of analyzed individuals from two families is limited and no segregation analysis could be performed in one of the two families, and the lack of replication in a second OSA family with variants in the same gene. Added to this, the results reported herein may only pertain to OSAS cases from Brazil.

In conclusion, the present study suggests that sequence variants in the COX20, PTPDC1, and TMDO4 genes could be associated with OSAS phenotype in Brazilian families. The putative involvement of these genes in OSAS phenotype needs to be validated and expanded using more families of diverse ethnicity.