Spectrum of mutations in the RB1 gene in Vietnamese patients with retinoblastoma.

Purpose: Retinoblastoma (RB) is a rare childhood malignant disorder caused by the biallelic inactivation of the RB1 gene. Early diagnosis and identification of carriers of heritable mutations in RB1 can improve disease outcome and management. In this study, we present the spectrum of mutations in the RB1 gene in Vietnamese patients with RB.
Methods: Tumor RNA from 50 probands with RB, including 12 bilateral and 38 unilateral cases, was extracted. cDNA, after reverse transcription, was sequenced to identify the RNA mutation of the RB1 gene. At the genomic DNA level, mutational analysis of all RB1 exons, exon-intron boundaries, and the promoter region was conducted using PCR and direct sequencing. Multiplex ligation-dependent probe amplification (MLPA) analysis was performed for patients for whom the first two results were negative. For patients for whom either the sequencing or MLPA results were positive for a tumor mutation, patients' and their parents' blood DNA was analyzed to determine the germline mutation.
Results: Forty-one different kinds of RB1 tumor mutations were identified in 41 probands (82.0%), including 11 of 12 bilateral cases (91.7%) and 30 of 38 unilateral cases (78.9%). The majority of the detected mutations were nonsense (15 different kinds), followed by frameshift (11 kinds), and splice site mutations (nine kinds). Each splice site mutation was confirmed to create a deletion of the corresponding exon with RNA sequencing. The single promoter mutation c.-197G>A was reported previously; however, both missense mutations identified in exon 6 (c.601G>C: p.A201P) and exon 22 (c.2264T>C: p.F755S) were novel. Gross deletions were detected with MLPA in three probands. The detection rate of germline mutations in bilateral and unilateral cases with mutations were 81.8% and 30.0%, respectively. Only one father out of the 20 parents tested was positive for a germline mutation. Conclusions: Mutations in the RB1 gene in Vietnamese patients were heterogeneous and highly prevalent with pathogenic truncated mutations. With advancement in therapeutics, early detection of RB is important for eye salvation.

Introduction

Retinoblastoma (RB) is the most common intraocular neoplasm of childhood with an estimated incidence of 1/15,000 to 1/20,000 live births [1], and is responsible for 5% of blindness and 1% of overall childhood cancer deaths, respectively [2]. The majority of RB cases are diagnosed by 3 years of age with the most common presenting sign of leukocoria, followed by strabismus and proptosis [3].

Most clinical phenotypes of RB can be explained by hereditary or non-hereditary mutational inactivation of the RB1 (Gene ID: 5925; OMIM 614041) gene [4]. Hereditary disease, caused by a constitutional mutation of RB1, predisposes to RB and other cancers later in life, and is transmitted as an autosomal dominant trait with high penetrance (90%) [5]. The non-hereditary form of RB is initiated by acquisition of two somatic mutations in the RB1 gene of retinal progenitor cells [6]. Genetic testing to identify RB1 mutations is not only essential for defining hereditary and non-hereditary forms of the disease but also critical for developing strategic care and management of patients with RB post-surgery and in screening for mutation carriers in related family members [1,7]. Identification of carriers of RB1 mutations can initiate important genetic counseling related to family planning and to the strategic reduction of risk for development of other types of cancers later in life. Moreover, in some special cases, due to the risk of developing RB during pregnancy, children who carry mutations can be purposely delivered prematurely to initiate early treatment of potential macular tumors [8].

Because of the immense disparities that exist between developed and developing countries in access to and quality of healthcare systems, mortality and morbidity rates related to RB are likewise disturbingly divergent in these separate settings. The mortality rates of RB in Asia are estimated to be around ten-fold higher than in Europe, Canada, and the United States (approximately 40% versus 3–5%) [9]. In addition, in developed countries, detection and clinical management of RB often occur much earlier than in developing countries, leading to higher rates of successful treatment and preservation of patients’ vision in the former. Ideally, patients at high risk for RB should undergo clinical screening for tumors using imaging techniques beginning in the first year of life. Because of the complex nature and expensive cost of these screening procedures, determination of genetic carrier status and risk stratification are important in avoiding such unnecessary clinical screenings. Most previously published studies of RB in Vietnam focused on the clinical presentation and application of imaging techniques to the diagnosis of RB. In contrast, the aim of the present study was to identify the spectrum of mutations in the RB1 gene in 50 unrelated Vietnamese patients with RB, and to demonstrate the importance of genetic testing in disease management of patients with RB in Vietnam.

Methods

Patients

Samples for the study were collected from 50 unrelated patients with RB, diagnosed by using standard ophthalmologic and histological criteria, at the Eye Hospital of Ho Chi Minh City, Vietnam, between November 2014 and April 2016. Samples included 46 tumor and blood-matched samples (11 bilateral and 35 unilateral) and four tumor-only samples (one bilateral and three unilateral). Because RB is a childhood disease, written informed consents for mutation analyses were obtained from the next of kin, caretakers, or guardians on behalf of the children enrolled in this study. The study was adhered to the ARVO statement on human subjects and was approved by the ethics committee of the University of Medicine and Pharmacy at Ho Chi Minh City.

DNA and RNA isolation

Tumor samples: Tumor-rich areas of archived formalin-fixed, paraffin-embedded (FFPE) tissues containing at least 50% tumor cells were marked on a hematoxylin and eosin slide by a pathologist, before manual microdissection of a corresponding unstained FFPE tissue section using a 21G needle. DNA was extracted using the ReliaPrep™ FFPE gDNA Miniprep System kit (Promega, Madison, WI).

Blood samples: Fresh peripheral blood samples were collected from patients for the purposes of the study. Total genomic DNA was isolated from EDTA-treated blood using the ReliaPrep™ Blood gDNA Miniprep System (Promega). RNA extraction: RNA was extracted from fresh tumor tissues using the RNeasy Mini Kit (Qiagen, Germantown, MD) following the manufacturer’s protocol.

PCR and Sanger sequencing

Reverse transcription PCR

Approximately 3–4 µg of RNA were reverse-transcribed into single-stranded cDNA using the PrimeScript™ 1st strand cDNA Synthesis Kit (Takara Bio, Shiga, Japan). All primers used in this study were newly designed. Five sets of overlapping primers were then used to amplify the 2,787 bp RNA coding region of RB1 (Supplementary Table 1, Appendix 1). Exons 1–5 were amplified using a touchdown PCR protocol: initial denaturation at 95 °C for 5 min, and 15 cycles of 95 °C for 30 s, 60 °C for 30 s (with a decrease of 1 °C per cycle), 72 °C for 90 s, followed by 30 cycles of 95 °C for 30 s, 58 °C for 30 s and 72 °C for 90 s, and a final elongation of 72 °C for 10 min. PCR amplification of exons 5–27 involved initial denaturation at 95 °C for 5 min, then 45 cycles of 95 °C for 30 s, 58 °C for 30 s, 72 °C for 90 s, with a final elongation of 72 °C for 5 min. PCR products were checked for size and purity using 2% agarose gel electrophoresis.

Genomic PCR

Twenty-two sets of primers (Supplementary Table 2, Appendix 2) were used to amplify all 27 exons, their flanking regions, and the promoter of RB1 in separate 25 µl reactions consisting of 1X PCR Buffer, 1.5 mM MgCl2, 200 µM each dNTP, 0.5 U Taq Hot Start Polymerase (Takara Bio), 0.1 µM each forward and reverse primer, and 25–50 ng of genomic DNA. PCR involved initial denaturation at 98 °C for 5 min followed by 40 cycles of 98 °C for 10 s, 60 °C for 30 s, and 72 °C for 2.5 min, with a final elongation of 72 °C for 10 min. PCR products were checked for size and purity using 1.5% agarose gel electrophoresis.

Sequencing

PCR products were purified enzymatically using the ExoSAP IT™ PCR Product Cleanup Reagent (Thermo Scientific, Waltham, MA) for removal of excess primers and dNTPs before Sanger sequencing using the BigDye Terminator v3.1 Kit and the ABI 3500 Genetic Analyzer (Applied Biosystems, Foster City, CA). PCR fragments were sequenced and analyzed in both directions. The primers used for sequencing are listed in Appendix 1 and Appendix 2. Sequences were compared to the reference sequence of the RB1 gene (GenBank accession number: NG_009009).

Blood DNA sequencing

DNA from the blood of patients with RB tumors found to contain potential mutations was PCR-amplified and sequenced to determine whether the mutations were somatic or germline. Only regions corresponding to the location of mutations found in the tumor were amplified and sequenced. If a germline mutation was confirmed, the patient’s parents were counseled and tested for their carrier status.

MLPA analysis

Multiplex ligation-dependent probe amplification (MLPA) analysis was performed for probands with no detectable mutation after screening with the sequencing techniques. MLPA was performed using the SALSA MLPA P047 RB1 probemix (MRC-Holland, Amsterdam, the Netherlands) and 75 ng of DNA following the manufacturer’s protocol. MLPA products were run on an Applied Biosystems 3500 Genetic Analyzer under standard settings. The results were visualized and analyzed using Coffalyser.Net software (MRC-Holland) according to the manufacturer’s instructions.

Results

Clinical characteristics of patients

All 50 patients with RB were diagnosed with advanced stage (group D or E) disease, 12 with bilateral and 38 with unilateral tumors, and underwent primary enucleation without any treatment before enucleation. Age at diagnosis ranged from 1 to 66 months, with an average age of 18.3 months for bilateral cases and 26.6 months for unilateral cases. In the unilateral cases, the distribution of left and right ocular lesions was similar (18 versus 20, respectively). A summary of cases and their clinical characteristics is shown in Table 1. These data are similar to other studies in which the bilateral cases were smaller in proportion to the unilateral cases, and were diagnosed earlier [10-12]. In the present study, the male to female ratio in general, unilateral, and bilateral RB was 30/20, 23/15, and 7/5, respectively. In the unilateral cases, the distribution of left and right ocular lesions was similar (18/20).

Table 1

Clinical characteristics of RB patients.

Characteristics	Unilateral (n=38)	Bilateral (n=12)
Mean age (months)	26.6	18.3
Gender
Male (n=30)	23	7
Female (n=20)	15	5
Side of RB
Right eye	20	n/a
Left eye	18	n/a
Somatic mutation (n=50)
Mutation (n. and %)	30 (78.9%)	11 (91.7%)
No mutation (n. and %)	8 (21.1%)	1 (8.3%)
Germline mutation (n=41)
Mutation (n. and %)	9 (30.0%)	9 (81.8%)
No mutation (n. and %)	21 (70.0%)	2 (18.2%)

Somatic mutation spectrum of RB1

A total of 41 different somatic alterations in RB1 were identified, including point mutations and chromosomal rearrangements (Figure 1 and Supplementary Table 3, Appendix 3). The majority of the point mutations, detected with Sanger sequencing, were nonsense mutations (15/41, 36.6%) followed by frameshift (11/41, 26.8%), splicing (9/41, 22.0%), missense (2/41, 4.9%), and promoter (1/41, 2.4%) mutations. Every splicing mutation created the corresponding exon deletion on RNA mutational analysis. MLPA detected three gross deletions (3/41, 7.3%), including deletion of exon 7, deletion of exon 17, and deletion of exons 6 to 10. Twenty-eight (56.0%) tumors carried biallelic mutations in RB1 gene.

Figure 1

Distribution of mutations in RB1 identified in this study. Gross deletions are represented as black lines. The underlined mutations are two missense novel mutations. Splicing mutations are indicated in squares. Nonsense mutations are above the coding sequence, while frameshift mutations are below.

Germline mutations in RB1

Among 41 patients with mutations with matched tumor and blood samples, germline mutations were identified in nine out of the 30 (30.0%) unilateral RBs and nine out of the 11 (81.8%) bilateral RBs. There were nine mutations in coding regions and nine splice site mutations leading to a premature stop codon on RNA transcripts (Table 2). Ten parent-couples of patients who carried germline mutations agreed to undergo carrier status analysis. Genetic transmission was found in only one male parent out of the 20 subjects involving a heterozygous splice site mutation (father of patient ID RB-Y55). This individual refused to undergo an eye examination.

Table 2

Germline mutations and parental carrier status of RB patients.

Patient ID	Patient	Father	Mother
RB-Y4	c.1399C>T	WT	WT
(Bilateral)
RB-Y5	c.1363C>T	ND	ND
RB-Y7	c.1150C>T	WT	WT
(Bilateral)
RB-Y13	c.2559delT	ND	ND
RB-Y14	IVS12+1G>A	WT	WT
RB-Y15	IVS12+1G>A	ND	ND
(Bilateral)
RB-Y16	IVS24+1G>T	WT	WT
(Bilateral)
RB-Y19	c.763C>T	ND	ND
RB-Y20	c.1940–1941delTT	ND	ND
RB-Y23	IVS13+1G>A	WT	WT
(Bilateral)
RB-Y26	c.601G>C	WT	WT
RB-Y30	c.1666C>T	ND	ND
RB-Y32	IVS6+1delG	WT	WT
(Bilateral)
RB-Y36	c.1841delA	WT	WT
(Bilateral)
RB-Y50	IVS12+1G>A	ND	ND
RB-Y53	IVS6+1G>T	ND	ND
RB-Y54	IVS15+1delG	WT	WT
(Bilateral)
RB-Y55	IVS13–2A>G	IVS13–2A>G	WT
(Bilateral)		(hetero-)

WT, wild-type; ND, not done; hetero-, heterozygous

Discussion

This is the first comprehensive study to describe the spectrum of somatic and germline mutations in RB1 in Vietnamese patients with RB. To the best of our knowledge, only one previous study investigated the genetics of RB tumors in Vietnam; however, that analysis was limited to investigation of germline mutations in RB1 [13]. In that study, Nguyen et al. reported a high frequency of germline mutations in unilateral (four out of nine) and bilateral (21 out of 25) cases [13].

In the present study, among the somatic mutations in RB1, nonsense and frameshift mutations were the most common ones likely causing truncation of the pRB protein. Similar results have been reported in other studies [14,15]. The mutation p.R358X in exon 11 was detected in four cases, suggesting that it is recurrent in Vietnamese patients with RB. This finding differs from that of a Chinese study of patients with RB in which the p.R579X in exon 18 was considered the most recurrent nonsense mutation [6]. Splice site mutations that disrupt RNA splicing play an important role in generating inactive pRB protein [16-18]. Accordingly, we found nine intronic mutations, all of which were associated with deletion of the involved exons, as revealed by sequencing of matched RNAs. All exon deletions induced shifts in the reading frame, and finally, introduction of a stop codon that would be expected to translate to truncation of pRB proteins without a C-terminal.

Almost all mutations detected in this study have been previously reported, except the two missense mutations identified in exon 6 (c.601G>C: p.A201P) and exon 22 (c.2264T>C: p.F755S). The first novel missense mutation c.601G>C (p.A201P) was found in the tumor and blood of a 21-month-old male infant with unilateral RB. The PolyPhen-2 score of this mutation is 1.0, suggesting a pathogenic ability; however, a functional study is needed to clarify this possibility. The other novel missense mutation, c.2264T>C (p.F755S), was found only in the tumor sample of a 42-month-old female child with unilateral RB. Codon 755 has been reported previously with the pathogenic p.F755N mutation [8]. In general, missense mutations contribute to only a small percentage of RB cases, and are usually related to low penetrance [19].

The cDNA screening analysis used in this study has some advantages. First, it can identify abnormalities that may occur during transcription. For example, in the tumor sample from patient RB-Y2, cDNA sequencing revealed substitution of exon 23 by a small segment of intron 22, but when the tumor DNA was sequenced, we could not find any mutation in exons 22 to 24 and their flanking regions. Similar phenomena were observed in samples from patients RB-Y8, RB-Y27, RB-Y32, and RB-Y34 where mutations were detected in RNA, but not at the DNA level. One possible theory for this discordance in results is the damage of transcriptional machinery in cancer cells. Second, cDNA screening helps to predict the location of intronic mutations (Appendix 3). Curiously, cDNA sequencing showed wild-type RB1 transcripts in samples from patients RB-Y1, RB-Y11, and RB-Y29, whereas gross deletions were detected in these samples using MLPA. It is possible that the wild-type transcripts could be downregulated in these samples due to genomic gross deletions. A well-designed real-time PCR can resolve this issue. Despite the advantages of combined RNA and DNA analyses in the present study, we did not investigate promoter methylation status and loss of heterozygosity. Consequently, the percentage of biallelic somatic mutations in this study was relatively low.

The ultimate goal of RB therapy is to ensure a high survival rate while minimizing adverse events and recurrence of the disease. Promising new therapies for RB, including platforms for intraocular delivery of drugs, have been developed in recent years [20]. However, additional disease management may be gained from genetic testing and effective genetic counseling of family members. Approximately 15% of individuals with sporadic unilateral retinoblastoma carry a germline mutation in RB1 [21,22], and these individuals can develop RB in the remaining eye and transmit the RB predisposition to their offspring. Germline mutations in the present study comprised 9/38 (24%) of the mutations in the unilateral cases, which is higher, although perhaps not significantly, than the average 15% generally noted in other studies. Similar increased rates of germline mutations in Vietnamese patients were noted in another study [13], possibly underscoring the added benefit of genetic testing and counseling for this population. Moreover, it has been shown that mutational screening of RB1 of relatives at risk for RB is more cost-effective compared to conventional ophthalmological examinations [23]. Prenatal diagnosis could also be offered to parents in the future based on the first successful report of preimplantation genetic testing [24]; this preventive intervention appears promising for future disease management in families at risk for having children with inherited RB. Because nonsense mutations comprise the majority of reported point mutations in RB, a gene therapy approach, designed to allow read-through of premature termination codons, offers possible future targeted therapy for RB [25]. Information on mutational profiles of patients with RB will further aid in development of targeted therapeutics, ensuring effective disease management and life-long follow-up.

In conclusion, this study is the first comprehensive investigation of somatic and germline mutations in Vietnamese patients with RB. We described a wide spectrum of mutations, including two novel mutations that appear not to have been described previously. A relatively high rate of germline mutations in this cohort of patients with RB indicates the benefit that would occur for implementation of genetic testing and genetic counseling of families of Vietnamese patients with RB.