Rubinstein–Taybi Syndrome (RSTS, Broad thumb–hallux syndrome, MIM 180849) is a rare congenital malformation syndrome that affects approximately 1 in 100,000–125,000 newborns. RSTS is characterized by the broadening of the thumbs and halluces, developmental delay, moderate to severe intellectual disability, proneness to keloids, and distinctive facial features such as a large beaked nose and a low-hanging columella (Roelfsema and Peters, 2007).

Genetic research on RSTS has mainly focused on mutations in the genes encoding CREB binding protein (CREBBP, MIM 600140) and its homolog, E1A-binding protein P300 (EP300, MIM 602700). CREBBP is located at 16p13.3, while EP300 is located at 22q13.2. Both CBP and p300 are highly homologous transcriptional coactivators sharing conserved protein–protein interaction domains including the enzymatic histone acetyltransferase (HAT) domain. Acetylation disrupts the DNA–histone interaction by neutralizing the positively charged lysine residues in histones and enabling increased accessibility for transcription factors to activate gene expression (Kalkhoven, 2004; Oliveira et al., 2006). Among patients with RSTS, roughly 50–60% harbor pathogenic variants in CREBBP (RSTS type I), while only 3–8% of the patients harbor mutations in EP300 (RSTS type II) (López et al., 2018). The vast majority (about 99%) of RSTS cases occur sporadically from de novo heterozygous CREBBP mutations, although vertical transmission has also been documented in a handful of cases (Bartsch et al., 2010). The spectrum of reported genetic alterations regarding RSTS includes point mutations (i.e., frameshift, nonsense, missense, and splice-site mutations), intragenic or large deletions, translocations and inversions (Negri et al., 2019).

Whole-exome sequencing (WES), which utilizes the “shotgun-based” approach of next-generation sequencing, is known for its efficiency and effectiveness in detecting single-nucleotide polymorphisms (SNPs) and small insertions/deletions (indels). Structural variants and large deletions, on the other hand, remain challenging for WES due to its reliance on the short-read approach (Chong et al., 2015; Chiu et al., 2021). Multiplex ligation-dependent probe amplification (MLPA), on the other hand, although unable to recognize SNPs and indels, shows strengths in detecting genomic variations, such as copy number variations (CNVs) and large-spanning deletions by comparing the differences in PCR-amplified fluorescently labeled primers binding to the probes. Although molecular diagnosis can reveal the genetic alterations in most RSTS cases, a sizeable subset of patients (30%) remain undiagnosed when this method is used (Tajir et al., 2013). In this study, we leveraged the distinct advantages of WES and MLPA by combining the two methods in our genetic diagnostic protocol for RSTS. Here, we report three patients with pathogenic CREBBP mutations and one with a large deletion spanning CREBBP in a five-patient cohort of clinically diagnosed RSTS individuals.

Genetic alterations relevant to RSTS identified in this study by both MLPA and WES are summarized in Figure 1. The characteristic clinical manifestations of RSTS for all five cases are summarized in Figure 2. As a whole, among the patients enrolled, MLPA showed a large 16p13.3 DNA deletion spanning the CREBBP gene in 1/5 (20%) patient, while causative CREBBP variants were detected in 3/5 (60%) patients by WES and were confirmed by Sanger sequencing. Through MLPA and WES, causative variants were identified in 4/5 (80%) of the RSTS patients and are summarized in Table 1.

Case 1

was a 4-year-old female showing typical clinical features, including a high-arched palate, broadened thumbs and halluces, and a tendency to keloid development (Figure 2A). Congenital cyanotic heart disease were also found, including hypoplastic aortic isthmus with mild coarctation of the aorta, perimembranous ventricular septal defect, large right patent ductus arteriosus (PDA), and anomalous left pulmonary artery arising from left side PDA. MLPA of DNA from the proband and the parents revealed a de novo heterozygous 16p13.3:3,728,096-3,962,938 deletion that spanned the loci of TRAP1 and CREBBP (Figures 3A, 4A). To validate the results, we also performed WES on the proband’s DNA, and a visualization of the WES results in IGV revealed low read coverage across the mutation span (Figure 4B). Surgical repair of the patient’s congenital cyanotic heart disease was performed to alleviate the cardiac symptoms, and diflucortolone ointment was prescribed for the treatment of the developed keloid. The developmental delay was managed by follow-up treatments in the outpatient department and diet education.

Case 2

was a 15-year-old male with typical clinical features suggesting RSTS (Figure 2B). MLPA of case 2 showed no detectable deletions in CREBBP. WES of the proband’s germline DNA revealed a novel CREBBP heterozygous nonsense mutation (NM_004380:c.C2608T:p.Gln870Ter) with a significant pathogenicity score (CADD:21.4, DANN:0.9986). Further segregation analysis by Sanger sequencing confirmed the variant detected through WES and revealed both parents to harbor homozygous wild-type alleles at the mutation site, confirming that the proband’s mutation had occurred de novo (Figures 3B, 5A). Since that the patient also presented with intellectual disability and mental retardation, the patient was referred to neurologists and psychiatrists at our hospital for diagnostic interviews and further psychological assessments.

Case 3

was a 9-year-old female with typical clinical features suggesting RSTS (Figure 2C). MLPA did not reveal any large deletions spanning CREBBP. WES showed a 2-bp deletion leading to a frame-shift in the CREBBP locus (NM_004380:c.4404_4405del:p.Gly1469AlafsTer9), which had been reported in another RSTS patient (Murata, 2001). Segregation analysis showed a wild-type genotype for both parents at the mutation site (Figures 3C, 5B), indicating that the mutation had occurred de novo.

Case 4

a 10-year-old male, also showed features of RSTS (Figure 2D). No large deletions were found in the patient’s DNA by MLPA, but WES revealed a CREBBP nonsense mutation (NM_004380:c.C3649T:p.Gln1217Ter) with a significant pathogenicity score (CADD:44, DANN:0.9987) that had been reported in a previous RSTS patient (Saettini et al., 2020). Further segregation analysis showed that the mutation had occurred de novo (Figure 5C). The patient also showed microcephaly and white matter hyperintensities under magnetic resonance imaging (MRI) examination. Rehabilitation programs and psychological assessments were arranged on a regular time course for the neurological symptoms.

Case 5

was a 4-year-old boy also with typical features of RSTS (Figure 2E). MLPA showed no large deletions in CREBBP. Unexpectedly, WES also failed to identify pathogenic variants within the CREBBP and EP300 loci (Figure 3E). Therefore, the causative genetic or genomic variant for case 5 remains unknown. Since the patient also manifested congenital glaucoma, trabeculectomy including peripheral iridectomy under the microscope was performed.

Of the five patients whose clinical presentation suggested RSTS, pathogenic variants in CREBBP were identified in four patients (80%) using an MLPA–WES genetic diagnostic workflow. However, the causative variant for 1/5 (20%) patient, remains uncertain. Among the sequencing methods used in this study, although WES efficiently detects small indels in protein-coding (exonic) DNA regions, the detection of copy number variations (CNVs), structural variants, and homologous regions of the genome remains challenging for WES. Given that pathogenic variants in RSTS comprise single-nucleotide or oligonucleotide variations as well as large deletions, genomic assays that can screen for large genomic variants (e.g., MLPA) are highly desirable. In this study, we integrated MLPA and WES in our molecular diagnostic protocol. PCR-based Sanger sequencing was also utilized to confirm the variants identified by WES. Indeed, a similar workflow has been applied to a Korean cohort (Choi et al., 2021), leading to the detection of causative variants in 80% of their patients, which is similar to our diagnostic yield.

Given that RSTS is a rare disorder arising from de novo mutations, the number of reported cases remain limited. In the Taiwanese population, one study reported the clinical and molecular data of 10 RSTS patients (Hou, 2005). In that study, chromosomal deletions over the 16p13 region (responsible for the CREBBP gene) were detected by fluorescence in situ hybridization (FISH). Only 30% of the RSTS patients were detected as having interstitial submicroscopic deletions in the 16p13 region. The low diagnostic rate from the use of FISH alone might lie in its ability to resolve only large DNA deletions. Thus, pathogenic single-nucleotide variants would be neglected. Similarly, among the five RSTS patients collected in our study, only 1/5 (20%) patient showed a large DNA deletion spanning CREBBP (case 1). Our study highlights the potential gains from harnessing various genetic assays.

Mutation hotspots in the CREBBP and EP300 loci have yet to be documented. In a previous report, mutations in the highly conserved HAT domain were regarded as leading to high pathogenicity (López et al., 2018). However, no mutations resided within the HAT domain among our patients. Previous studies have further reported that frameshift mutations are the most prevalent type of mutation in RSTS patients (López et al., 2018; Choi et al., 2021). Although limited in number, more nonsense mutations (cases 2, 4) than frameshift mutations (case 3) were observed in the current study.

In our diagnostic workflow, the causative genetic alteration in 1/5 (20%) RSTS patient was not identified. Given that the pathogenesis of RSTS is not yet fully understood, the possibility remains that mutations occurring outside the CREBBP and EP300 loci may contribute to the RSTS phenotype. CREBBP and EP300 are both epigenetics-associated factors that alter acetyltransferase activity. In a previous study, experiments on a developmental animal model for RSTS were conducted using mice deficient in CREB-binding protein (CBP), and these CBP^+/− mice were found to exhibit an abnormal RSTS skeletal pattern (Tanaka et al., 1997). Furthermore, transgenic mice generated by Korzus et al. expressing reversible CBP that lacked HAT activity showed deficits in long-term memory, indicating that the HAT activity of CBP is essential for brain information processing and cognitive function (Korzus et al., 2004; Korzus, 2017). Although the genetic and epigenetic functions of CREBBP genes in rodents have been thoroughly investigated, the complete pathogenicity of RSTS remains to be explored.

Here, we summarize the molecular genetic diagnostic progress of five patients clinically diagnosed with RSTS. Diagnostic workflow combining MLPA and WES was carried out to identify causative genetic alterations in CREBBP and EP300. Four of the five patients were found to harbor causative mutations in CREBBP. As the exact pathogenesis of RSTS has yet to be completely elucidated and genetic alterations in roughly 20% of the patients have yet to be found, more research is needed on the roles of genetic and epigenetic alterations in RSTS.