A genotype and phenotype analysis of SMAD6 mutant patients with radioulnar synostosis

Abstract Background SMAD6 variants have been reported in patients with radioulnar synostosis (RUS). This study aimed to investigate the genotypes and phenotypes for a large cohort of patients with RUS having mutant SMAD6. Methods Genomic DNA samples were isolated from 251 RUS sporadic patients (with their parents) and 27 RUS pedigrees. Sanger sequencing was performed for the SMAD6 coding regions. For positive probands, co‐segregation and parental‐origin analysis of SMAD6 variants and phenotypic re‐evaluation were performed for their family members. Results We identified 50 RUS probands with SMAD6 variants (13 co‐segregated with RUS in pedigrees and 37 in RUS‐sporadic patients). Based on the new and previous data, we identified SMAD6 mutated in 16/38 RUS pedigrees and 61/393 RUS sporadic patients, respectively. Overall, 93 SMAD6 mutant patients with RUS were identified, among which 29 patients had unilateral RUS, where the left side was more involved than the right side (left:right = 20:9). Female protective effects and non‐full penetrance were observed, in which only 6.90% mothers (vs. ~50% fathers) of SMAD6 mutant RUS probands had RUS. Pleiotropy was observed as a re‐evaluation of SMAD6 mutant families identified: (a) three families had axial skeletal malformations; (b) two families had polydactyly; and (c) eight families had other known malformations. Conclusion SMAD6 was mutated in 42.11% RUS pedigrees and 15.52% RUS sporadic patients. The RUS patients with SMAD6 variants exhibit both non‐full‐penetrance, variable expressivity, pleiotropy, female protective effects, and the left side is more susceptible than the right side.

Further, in our daily work, a previously SMAD6positive RUS family (M2553; Yang et al., 2019) consulted with our laboratory for counseling about the risk of recurrence and the possible explanations, because their newborn baby (II:3) is suffering from CHD but without RUS and carries the same SMAD6 variant as the family RUS proband (Figure 4).
The following points were made based on the situations above: (a) whether the mutant SMAD6 detected in RUS can be replicated; (b) how is the transmitting features of RUS in SMAD6 mutant families if the (a) is correct; (c) the number of SMAD6-positive RUS family that co-existed with other known SMAD6-related disorder; and (d) whether the SMAD6 mutant RUS patients exhibit other skeletal malformation given the loss of Smad6 mice exhibiting both axial skeletal and appendicular malformations.
To answer these questions, we have further collected genomic DNA specimens from 251 RUS sporadic cases (and their family members) and 27 RUS pedigrees and performed Sanger sequencing of SMAD6-coding regions for these newly collected samples. Afterward, by integrating the data (from the cases in Yang et al., 2019 and the present cases), we performed a phenotypic reevaluation and genotypic re-analysis for 61 SMAD6 mutant probands (and their family members) by focusing on RUS. We identified that SMAD6 was mutated in 42.11% RUS pedigrees and in 15.52% RUS sporadic patients and that RUS families with SMAD6 variants exhibited nonfull-penetrance, variable expressivity, pleiotropy, female protective effects, and higher susceptibility at the left side than at the right side.

| Study subjects
Written informed consent was obtained from all probands, their parents, and their available family members. The inclusion criteria involved the diagnosis of RUS in the absence of identifiable syndromes, such as Apert/Crouzon/Pfeiffer syndrome (Schaefer et al., 1998), Holt-Oram syndrome (Wall et al., 2015), William syndrome (Charvat et al., 1991), Ehlers-Danlos Syndrome (Ritelli et al., 2017), or other obvious dysmorphic-syndromes. None proband with amegakaryocytic thrombocytopenia or bone marrow failure (Niihori et al., 2015) were met. Patients with chromosome aneuploidy (tested by GTG banding (Yang et al., 2019) were also excluded from this study.

| Subject classification
According to the RUS family history, we classified RUS into two categories, namely, RUS pedigree and sporadic patient. RUS pedigree indicates that a family has more than one RUS patient. RUS sporadic patient means that the family has only one RUS patient regardless of the presence of other malformations. An overall cohort description was provided in Table S3.
The RUS patients with SMAD6 variants exhibit both non-full-penetrance, variable expressivity, pleiotropy, female protective effects, and the left side is more susceptible than the right side.

K E Y W O R D S
axial skeletal malformations, penetrance, pleiotropy, polydactyly, SMAD6 variants

| Sanger sequencing and bioinformatics analysis
For each subject, genomic DNA was extracted from peripheral blood or oral swabs by using DNA isolation kits (Cat# D3392-02;Omega Bio-Tek, Inc.;or Magbead Swab DNA Kit, CW2507, CoWin Biotech Co., Ltd.) in accordance with the manufacturer's procedures. Sanger sequencing was performed for the exons and intron-exon boundaries (with at least +5 and −5 bp areas were included) of SMAD6 (NM_005585.5). Detection of 5′UTR and 3′UTR variants of SMAD6 did not included in the present study. Polymerase chain reaction (PCR) amplification was performed using genomic DNA as a template by using a Goldstar ® PCR kit (Cat# CW0655M; CoWin Biotech Co., Ltd.). Sanger sequencing was conducted using a BigDye ® Terminator v3.1 cycle sequencing kit (Applied Biosystems, Thermo Fisher Scientific, Inc.) in accordance with the manufacturer's protocol. The amplified PCR products were purified with 70% ethanol (analytically pure) and then run on an Applied Biosystems™ 3500 series genetic analyzer (Applied Biosystems, Thermo Fisher Scientific, Inc.). Details about the primers and PCR conditions in the current study are provided in Table S2.

| Phenotypic investigation for families with proband having a SMAD6 variant
Considering the SMAD6 variants enriched with CHD, BAV, CS, or ID (Calpena et al., 2020;Gillis et al., 2017;Jin et al., 2017;Kloth et al., 2019;Lelieveld et al., 2016) and Smad6 knock-out mice exhibiting axial skeletal malformations (Estrada et al., 2011), patients with SMAD6positive RUS probands and their available family members (those with SMAD6 variants) were invited for counseling regarding the presence of any sign of the above disorders. Patients with possible positive signs were further invited for phenotypic re-evaluation, which was carried out by a physician, surgeon, and geneticist, independently. When necessary, B-ultrasound and x-ray examinations were performed for concerned individuals.

| Resequencing identified 50 SMAD6 variants
We previously performed genetic analysis for patients with RUS (Yang et al., 2019). In the present study, we newly collected genomic DNA (with unknown cause) from 27 pedigrees and 268 sporadic cases (plus their available family members) with RUS. Sanger sequencing was firstly performed for all 295 RUS probands. We determined that all coding regions of SMAD6 were fully covered for each (5′-and 3′-UTR were not included). After filtering procedures, we identified 50 SMAD6 rare variants (Table 1), comprising 37 loss-of-function, 12 damaging missense, and 1 disruptive inframe variants (Table 1).

were detected
For recessive SMAD6 variants reported on complex CHD/ RUS patients (Kloth et al., 2019), we checked if the RUS proband in the present study carries SMAD6 recessive variants. However, results showed that none of the probands carried rare SMAD6 recessive variants (Table 1). Even after using a less stringent filtering condition by adjusting MAF to less than 0.01 (gnomAD_Eas) and expanding the data by adding 24 previously reported sporadic RUS patients, further analysis (Yang et al., 2019) indicated that none of them have SMAD6 recessive variants (data not shown).

| Expressivity was variable
At least three points can prove that SMAD6-mutated RUS patients are associated with variable expressivity. First, in 61 sporadic RUS patients with SMAD6 variants, 19 (31.15%) were unilaterally affected (Table 1, Table S3). Second, in 32 RUS patients from 16 SMAD6 mutated pedigrees, 10 patients have unilateral RUS (31.25%, Figure 1, Table S3). Third, within a pedigree, family members with the same SMAD6 variant can exhibit bilateral or unilateral RUS. For example, in pedigrees, R108 and RJ050, fathers (II:1) had bilateral RUS but their children (III:1) only had unilateral RUS (Figure 1). In family RJ003, father (II:1) had right RUS, but his son (III:1) had bilateral RUS (Figure 1). The Carter effect, also known as female protective effect (Carter, 1961), was identified in SMAD6-mutated RUS patients. First, the number of males was higher than that of females. As previously described, the maleto-female ratio of sporadic RUS was 3:1 (Yang et al., 2019). In the present study, the male-to-female ratio for SMAD6-positive sporadic RUS patients was 4.10:1 (49 males vs. 12 females). Also, in SMAD6-positive RUS pedigrees, RUS males (Jordan et al., 2012a) were more than RUS females (Estrada et al., 2011), with the male to female ratio of 3.6:1. Second, the penetrance of RUS for SMAD6-positive parents (of the probands) varied (maternal: 6.90% vs. paternal: 47.83%). Third, in mutated patients, the RUS was less severe in females than that in males. RUS can be bilateral or unilateral. In the present study, we identified 93 SMAD6-mutated RUS patients (Schwarz et al., 2010). In 72 males, 19 were unilateral (26.39%). In 21 females, 10 were unilateral (47.62%). Therefore, if we define the unilateral RUS as less severe, the females with the SMAD6 variant tend to have less severe RUS. In 29/93 SMAD6-mutated patients with unilateral RUS (Table S3), 20 were affected at the left side, while 9 were affected at the right side, indicating that the left side was more susceptible than the right side. Further, we notified that the left and right differences occurred in sporadic patients. In SMAD6 mutant sporadic patients, 19/61 patients had unilateral RUS, in which 15/19 were affected at the right side, while 4/19 were affected at the right side. In unilateral RUS patients from SMAD6 mutant pedigrees, 5/10 were affected at the left side, while 5/10 were affected at the right side.

| Pleiotropy: Novel (polydactyly, spinal malformations) and known phenotypes identified in SMAD6 mutant families
In combination with data obtained from previous cases (Yang et al., 2019), 77 SMAD6 mutant probands with RUS were identified. We intended to recall all patients and their available relatives for clinical re-evaluation. However, several early patients lost to follow-up. In total, 61 SMAD6 mutant probands (and their family members) participated in our program for the further survey about other possible phenotypes (except for RUS). A total of 13/61 families had other related malformations. This figure was possibly under-estimated because only individuals (or family members) with identifiable symptoms underwent further clinical examination. Three families had axial skeletal malformations. The present study identified three RUS families with mutant SMAD6 had axial skeletal deformities (Figure 2). In family RJ037 (Figure 2a), both the proband (II:2) and his mother I:2 carried the same SMAD6 loss of function variant (c.240delG/p.A80fs), II:2 suffered from RUS and enlarged the fourth rib at the right side (Figure 2b), but his mother (I:2) suffered from bone fusion that occurred between the first and second cervical vertebrae (Figure 2c). In family RJ051 (Figure 2d), III:2, II:1, and II:5 all carried a SMAD6 loss-of-function variant (c.165C>A/p.C55X), III:2, and II:1 had RUS, but II:5 had caudal vertebra dysplasia without RUS. In family M1204 (c.1016A>C/p.H339P, (Yang et al., 2019), Figure 2e), the proband suffered from RUS and spinal malformations, including scoliosis, kyphosis, vertebral bone osteosclerosis, and micro-shrinkage ( Figure  2f,g). We have also observed that another seven SMAD6 mutant members in RUS families had kyphosis, (5/7 were less than 40-years-old) and six SMAD6 mutant members in RUS families had obvious vertebral degeneration.
Considering that the incidence of kyphosis or vertebral degeneration is high in the general population, we cannot define a definite association between the SMAD6 variant and kyphosis or vertebral degeneration at present.
Two RUS-positive families had polydactyly. In family R080, the proband (SMAD6: c.2T>C/p.M1T) suffered from RUS and left thumb polydactyly (extra floating finger) (Figure 3a,b). In family RJ002, the proband II:1 suffered from RUS, but his uncle I:3 suffered from left fifth finger polydactyly (also extra floating finger, Figure 3c,d) but without RUS, and both of them had SMAD6:c.1224delC/p. H408fs.
Nine RUS families had other known phenotypes. Based on the re-evaluation, we identified 6/61 SMAD6-positive RUS families (seven patients) with CHD (with one had BAV, Figure 4). Typically, in a previously reported SMAD6 positive family M2553, I:1, III:1, and III:3 all had a SMAD6 loss-of-function variant (c.1050C>G/p.Y350X). However, the I:1 and III:1 suffered from RUS, but the newly born individual III:3 suffered from CHD (patent ductus arteriosus and mild mitral regurgitation) without RUS (Figure 4). In family R005 (Figure 4), both the individuals I:1 and II:1 carried the same SMAD6 variant (c.1012G>t/p.E338X), I:1 was normal, but his son II:1 suffered from RUS and CHD (mild tricuspid regurgitation). Similar findings for the four other SMAD6 mutant RUS families with CHD are illustrated in Figure 4.
We observed that six SMAD6 mutant family members suffered from skull abnormalities, in which three had frontal bossing, two had plagiocephaly, and one had both frontal bossing and plagiocephaly (note: 4/6 had both RUS and skull abnormalities, Figure 4). Notably, we did not identify any SMAD6 rare variants from four RUS patients with intellectual disability, and none of the SMAD6 mutant members of these 61 families has reached the point that an intelligence test is required.

| DISCUSSION
By focusing on RUS, the present study identified 50 rare SMAD6 variants through the 295 probands obtained from 268 RUS sporadic patients and 27 RUS pedigrees. We first determined that 37/50 variants were deleterious because 37 SMAD6 variants were loss-of-function variants (comprising of 21 frameshift, 14 stop-gain, 2 initiation codon variants, Table 1). These 37 variants tended to produce abnormal mRNA that is generally associated with nonsense-mediated decay and then exerted a haploinsufficiency effect as the mechanism. Second, we determined that 12/50 SMAD6 missense variants were all deleterious on the basis of the following considerations: (a) any of these 12 variants did not exist on genomAD databases and in our 479 in-house  Calpena et al. (2020) designed a model to evaluate in silico the deleteriousness of SMAD6 missense variants (such model was based on functional experiments data). According to Calpena model, if a SMAD6 variant fulfills both DS > 4 and the CADD predicted damaging, such SMAD6 missense should be defined as deleterious. In the present study, all 12 SMAD6 missense variants met the above criteria (Table S2) and were thus defined as deleterious. Only one SMAD6 variant, that is, the c.81_82insGGCGGCGGCGGT:p.S27delinsSGGGG that identified from family RXC001, should be defined as uncertain significance, because this variant was not reported in the gnomadAD database or in our in-house 479 exome database, and the proband with such variant had both RUS and BAV (both two disorders were specific to SMAD6 disruption (Gillis et al., 2017;Yang et al., 2019). Moreover, at the same position of S27, 49 deletion alleles (15-66995669-GGCGGCA-G, inframe deletion, p.Ser27_ Gly28del) are reported in gnomadAD database. Further functional experiments are needed to test the pathogenicity of this variant.
From the 61 SMAD6-mutated probands (who stayed in touch), we identified that 14 families had subordinate clinical findings. Notably, in the three families, SMAD6mutant members had axial skeletal malformations, comprising one patient with cervical vertebrae-fusion, one patient with rib malformation, one patient with caudal vertebral dysplasia, and one patient with scoliosis and kyphosis. Loss of Smad6 in mice leads to defects in both axial and appendicular skeletal development (Estrada et al., 2011). Specifically, Smad6−/− mice exhibited a posterior transformation of the seventh cervical vertebra, bilateral ossification centers in lumbar vertebra, and bifid sternebrae caused by incomplete sterna band fusion (Estrada et al., 2011). Therefore, the SMAD6-mutated individuals who exhibit axial skeletal malformations support the skeletal phenotypes of Smad6−/− mice, suggesting that SMAD6 mutant patients should focus on the presence of axial skeletal malformations. Another novel incidental phenotype on RUS families identified in the present study was polydactyly. Two SMAD6 mutant members had polydactyly, both two additional digits were connected to the fingers just like a nubbin (Figure 3). In one of the affected hands, the extra finger was attached to the thumb (the radial side). In another hand, the extra finger was attached to the small finger (the ulnar side). Considering that the frequency of polydactyly was as rare as 1 in ~700-1000 live births (Jordan et al., 2012a), and SMAD6 was involved in the antagonizing BMP-signaling (such signaling disruption involved in the number of phalanges in animals (Jordan et al., 2012b)), it is unlikely that polydactyly occurred on such two SMAD6 mutant families was a coincidence. Further confirmative experiments are needed.
From the SMAD6-mutated probands, we also identified eight families in which the family members suffered from CHD, BAV, or skull abnormalities. SMAD6 variants enriched in CHD, BAV, or skull malformation have been well described previously (Calpena et al., 2020;Gillis et al., 2017;Jin et al., 2017;Kloth et al., 2019;Lelieveld et al., 2016). The present study confirmed that different phenotypes can occur in different members (with the same mutant SMAD6) within a family. Therefore, CHD, BAV, RUS, skull abnormalities, axial skeletal malformation, and polydactyly should be concluded to SMAD6-related phenotypic spectrum. The Carter effect, which was observed by Cedric Carter in patients with pyloric stenosis in 1961, refers to females that are less commonly affected by pyloric stenosis and are more likely than males with pyloric stenosis to have children affected with the disorder (Carter, 1961;Carter & Evans, 1969). In the present study, we found that the Carter effect is remarkably associated with the RUS phenotype in patients with SMAD6 variants as several obvious female protective effects observed which were described above. However, one point of the present study did not fit to Carter effect. According to the Carter effect, the affected female should carry more severe (or increased number of) variants than that of the affected male, and the relatives of the affected female have more chance to develop the index disorder. In the present study, 3/11 (27.3%) SMAD6 mutant female probands had a family relatives suffered by RUS. By comparison, 13/50 SMAD6 mutant males (26.0%) had family history of RUS, and we did not observe female RUS probands with mutant SMAD6 having more family history of RUS.
In SMAD6 mutant patients with unilateral RUS, the number of left RUS was remarkably higher than that of right RUS. SMAD6 encodes an inhibitory component of BMP/SMAD signaling (Estrada et al., 2011). It is known the lateral plate mesoderm (LPM) forms the progenitor cells that constitute the limb skeleton, heart and cardiovascular system, and others (Prummel et al., 2020) in the developing vertebrate embryo. Considering that the BMP/ SMAD signaling sets a repressive threshold in the LPM essential for the integrity of LEFT/RIGHT signaling (Furtado et al., 2008), SMAD6 haploinsufficiency may affect the integrity of LEFT/RIGHT signaling, causing asymmetric development of the left/right limbs. Vertebrate embryo development is not a complete symmetric event because many organs (such as stomach, heart, spleen, etc.) primary located on the left side, and a more precise BMP signal is needed on the left side development. Therefore, the left side was more susceptible to developing RUS under SMAD6 haploinsufficiency.

DATA AVAILABILITY STATEMENT
The datasets generated and/or analyzed during the current study are available in the National Center for Biotechnology Information (NCBI) ClinVar repository, (https://www.ncbi.nlm.nih.gov/clinv ar/varia tion/SUB98 87746/).