Introduction

Pathogenic germline variants in APC (MIM# 611,731) cause familial adenomatous polyposis syndrome (FAP; MIM# 175,100), a rare autosomal dominant-inherited syndrome characterized by the development of multiple colorectal adenomas and a very high risk of colorectal cancer [1,2,3,4]. In classic FAP, patients develop hundreds to thousands of colorectal adenomatous polyps, while in attenuated FAP (AFAP), patients develop fewer adenomas (< 100) at a later age than those with classical FAP [5,6,7,8]. A subset of patients with multiple colorectal adenomas and no APC germline variants have been found to carry biallelic variants in the base excision repair gene MUTYH (MIM# 604,933), causing MUTYH-associated polyposis (MAP; MIM# 608,456) [9]. In addition, a number of other genes associated with adenomatous polyposis, such as POLE, POLD1, NTHL1, MSH3 and MLH3, have recently been reported [10,11,12,13]. The detection rate of APC variants in FAP patients depends on phenotype and methods. In classic FAP, APC germline variants can be detected in up to 85% of patients [14, 15]; however, the detection rates of APC germline variants in patients with fewer colorectal adenomatous polyps (AFAP patients) are lower, ranging from 10 to 30% of patients [14, 16], suggesting that a proportion of pathogenic variants remain undetected by routine methods [17,18,19]. Mosaic APC variants and deep intronic variants localized in regions not covered by PCR-based diagnostics were previously identified as additional causal factors. Using RNA-based assays and next-generation sequencing (NGS), it has been shown that a proportion of variant-negative FAP patients harbor molecular changes in deep intronic regions of APC [19, 20]. These studies identified deep intronic APC variants that result in pseudoexon formation [19, 20]. Through the use of sensitive techniques, somatic APC mosaicism has been demonstrated in a minority of adenomatous polyposis patients [21,22,23,24,25,26]. In addition, using deep sequence analysis of APC in DNA isolated from multiple adenomas, mosaic variants were identified in 9 of 18 patients with 21 to 100 adenomas; in some of these cases, NGS also detected the variants in leukocyte DNA at low frequency [27]. In this study, we investigate the role of deep intronic germline APC variants and mosaic APC variants in leukocyte DNA as possible genetic causes of colorectal polyposis in a Dutch cohort of unexplained patients with more than 50 polyps.

Materials and methods

Patients

A total of 80 index patients with more than 50 colorectal polyps (Table 1) were selected from a previously described cohort [28,29,30,31]. The cohort included patients previously screened for germline mosaic APC variants by denaturing gradient gel electrophoresis (DGGE) [17], the protein truncation test (PTT) [17] and high resolution melting analysis (HRMA) [21]. All cases tested negative for pathogenic germline variants in APC, MUTYH, POLE, and POLD1 and for NTHL1 hotspot variants. Clinicopathological data included date of birth, gender, age at diagnosis of colorectal polyps/adenomas, cumulative number of polyps, location and histology of polyps/adenomas, information on CRC and presence of polyps/CRC in first-degree family members. Since the term serrated adenomas is currently preferred over hyperplastic polyps, we lumped together polyps described as such under the term sessile serrated lesions with or without dysplasia. Three controls were included in this study. Leukocyte DNA from this cohort was available for the study. The study was approved by the medical ethics committee of Leiden University Medical Center, protocol P01-019.

Table 1 Clinical characteristics of the colorectal polyposis patients (n = 80)
Full size table

APC intronic variant screening

Leukocyte DNA of the patients was screened for the intronic APC variants in Table 2 using Sanger sequencing. Primers were designed using Primer3 software http://primer3.ut.ee/ and were obtained from Eurofins Genomics (Ebersberg, Germany). The following primers with universal M13 tails were used: c.1408 + 731C > T, c.1408 + 735A > T and c.1408 + 729A > G; forward: 5′-TGTAAAACGACGGCCAGTATCATGCTGAACCATCTCAT-3′ and reverse: 5′ CAGGAAACAGCTATGACCAAATGACGAATGAAACGATG-3′. For c.532-941G > A; forward: 5′ TGTAAAACGACGGCCAGTAGAGGGTTTGGGAAGTGGAG-3′ and reverse: 5′ CAGGAAACAGCTATGACCTCTGTGTGCCCTTAGAAAACTG-3′. Sanger sequencing of the PCR amplified fragments was performed by Macrogen (Amsterdam, Netherlands). The sequencing results were analyzed using Mutation Surveyor software (Sofgenetics, State College PA, USA).

Table 2 Summary of the germline pathogenic APC intronic variants
Full size table

Next-generation sequencing and data analysis

Deep APC sequencing was performed using a previously described custom APC panel [27]. The complete sequencing panel consisted of 115 amplicons (11,216 bp), covering 99.3% of the coding regions of APC. Libraries were prepared with Ion Ampliseq™ 2.0 Kit (Thermo Fisher Scientific, Bleiswijk, The Netherlands) according to the manufacturer’s instructions and were sequenced on the Ion Torrent Proton Platform (Thermo Fisher Scientific, Bleiswijk, The Netherlands). Sequence data were analyzed as described previously [27]. Variants were annotated to the GenBank reference sequence NM_000038.4. The Integrative Genomics Viewer (IGV) (https://www.broadinstitute.org/igv/) was used to visualize the read alignment and the presence of variants against the reference genome.

Results and discussion

In this study, we attempt to identify the genetic causes of colorectal polyposis in unexplained patients with colorectal polyposis. Deep NGS of APC was performed to identify possible undetected pathogenic mosaic variants. Furthermore, APC intronic germline variants described previously [19, 20] were studied to evaluate their role. A high-risk cohort was selected for this study, consisting of 80 index patients with ≥ 50 colorectal polyps (Table 1), of whom many had a relatively early onset, which increases the probability of finding undiscovered mosaic or intronic variants. The mean age at diagnosis of colorectal polyps was 49 years (range 12–80). The majority of patients (n = 51, 63.8% with a mean age of 51 years at diagnosis) had a cumulative polyp count between 50 and 100, while 29 patients (36.2% with a mean age of 46 years at diagnosis) showed more than 100 polyps. Forty-five percent of the patients displayed only adenomatous polyps, while 47.5% of the patients displayed a mixed phenotype with adenomas and sessile serrated lesions with or without dysplasia. CRC was found in 27 patients (33.8%, with a mean age of 56 years, range 37–80). The clinical characteristics of the patients are summarized in Table 1.

First, we screened the leukocyte DNA of 80 patients for the following deep intronic heterozygous germline variants in APC: c.1408 + 731C > T, p.(Gly471Serfs*55), c.1408 + 735A > T, p.(Gly471Serfs*55), c.1408 + 729A > G, p.(Gly471Serfs*55) and c.532-941G > A, p.(Phe178Argfs*22). We did not detect any of these variants in our cohort. The study by Spier et al. [19] was the first to describe APC-related pseudoexons in FAP patients from Germany. These pseudoexons were caused by three heterozygous germline variants with a combined frequency of 6.4% (8/125); APC c.532-941G > A was identified in five patients, APC c.1408 + 731C > T was identified in two patients, and APC c.1408 + 735A > T was identified in one patient [19]. In a second study by Nieminen et al. [20], two additional intronic variants were identified in a cohort of 54 patients from Finland: APC c.1408 + 729A > G and APC c.646-1806 T > G and the variant identified previously by Spier et al., APC c.1408 + 731C > T. The overall reported frequency of these variants from the study by Nieminen et al. was 5.5% (3/54). The reported frequency of these intronic variants from both studies is approximately 6%. Nevertheless, we could not detect these variants in our cohort, possibly because either the frequency of intronic variants is lower in the Dutch population and the sample size of our cohort is not large enough or because these variants are local founder variants.

Subsequently, we performed deep APC sequencing of leukocyte DNA from 80 colorectal polyposis patients. Our positive controls were two previously described cases with mosaic APC variants [27]; APC c.4110_4111delAA was reported to be present in leukocyte DNA with 4% variant allele frequency (VAF), and APC c.2493dupA was reported with a VAF of 3% in leukocyte DNA. The APC mosaic variant c.4057G > T served as a negative control, as the variant was detected previously [27] in normal colonic mucosa and was absent in leukocyte DNA. Both positive controls, APC c.4110_4111delAA (Fig. 1) and APC c.2493dupA, were clearly present, while APC c.4057G > T was absent in the negative control. No additional APC mosaic variants were detected in our cohort. A limitation of this study is that we used only leukocyte DNA for mosaicism screening due to the scarcity of available DNA from patient adenomas. Mosaicism might remain undetectable or be overlooked if the molecular analysis is limited to blood, even when sensitive techniques are applied, due to very low or even absent presentation of the mutated allele [23, 27]. Peripheral blood cells arise from the mesoderm, and when the variant occurs after mesoderm and endoderm specification (early postzygotic mutation), the mosaicism is likely restricted to the colon and is difficult to detect the variant in leukocyte DNA [23, 27, 32, 33]. In a previous study, it was recommended to test at least two or more adenomas to detect mosaic variants [27].

Fig. 1
figure 1

Integrative Genomics Viewer (IGV) images of next-generation sequencing (NGS) data of mosaic APC c.4110_4111delAA variant detected in the leukocyte DNA of the positive control sample

Full size image

A recent systematic review of current APC mosaicism studies recommends testing adenomas from the polyposis patients without APC germline variant to allow the detection of low allele frequency mosaicism as well as mosaicism confined to colon [33]. Consequently, in our study, APC mosaic variants confined to the colon could have been missed because we could not screen the DNA from the adenomas of the patients.

In conclusion, we did not detect any of the four previously reported APC intronic variants in our cohort. We also did not detect mosaic APC variants in our cohort using deep sequencing analysis in blood. This finding suggests that the benefit of using targeted amplicon-based NGS to further scrutinize the APC gene in unexplained cases of polyposis is limited. Analyzing DNA from adenomas in addition to leukocyte DNA is recommended to detect a possible underlying mosaicism. Also, other approaches, such as whole genome sequencing or transcriptome sequencing, could be employed to detect undiscovered intronic or promoter variants or other regulatory variants.