Bile liquid biopsy is a useful modality for molecular diagnosis of cancerous tumors in pancreatic head lesions

doi:10.21203/rs.3.rs-4228560/v1

Cell-free DNA (cfDNA) in the blood is a less-invasive tool to reveal genomic profiles in pancreatic and bile tract cancers (PC and BTC); however, its utility is limited owing to its small amount and fragmentation of DNA for integrated sequences. This study evaluated the quality of in bile juice to clarify the potential of genome profiling for precision treatment. We collected data from 13 patients with pancreatic-biliary tumors invading the bile duct and subsequently underwent surgical resection: 7 cholangiocarcinomas, 1 gallbladder carcinoma, 3 ampullary carcinomas, 1 pancreatic cancer, and 1 intraductal papillary mucinous carcinoma. Whole-exome sequencing-based extracted DNA from resected samples revealed pathogenic alterations of TP53, ABCA8, BRCA2, FGFR2, APC, SMAD4 and NRAS. Of the 13 patients, nine pairs of cfDNA from bile and plasma were analyzed. Bile contained more (and longer fragments) cfDNAs compared with plasma. The most representative alterations in TP53, KRAS, PIK3CA, BRAF and NRAS were detected in bile cfDNA. Thus, bile cfDNA indicated better quality and quantity for sequence analysis and reflected tumor-derived genetic variation. Bile cfDNA is a useful tool for liquid biopsy in genomic profiling and precision medicine.

Biological sciences/Cancer/Cancer genomics

Health sciences/Gastroenterology/Gastrointestinal diseases/Gastrointestinal cancer/Biliary tract cancer

Health sciences/Medical research/Genetics research

Health sciences/Medical research/Translational research

Pancreatic cancer (PC) and biliary tract cancer (BTC) have 10 and 20% of 5-year overall survival, respectively [1,2]. First-line therapy for advanced BTC and PC [3-5] with cytotoxic agents such as gemcitabine, nab-paclitaxel, or cisplatin has limited therapeutic efficacy. Integrated genome sequencing has recently enabled the clarification of pathogenic alterations associated with specific diseases and the introduction of genotype-matched treatments for patients with gastrointestinal caners [6-9]. Approximately 40% of the patients with BTC harbor actionable genomic alterations associated with molecular-target or immune therapy, and the personalized medicine might provide a chance to prolong the survival [10,11]. Moreover, the accumulation or heterogeneity of genomic alterations is involved in the acquisition of resistance to treatment; therefore, the subsequent clonal changes in genomic profiles during treatment might provide hints to overcome treatment resistance [12,13].

Cell-free DNA (cfDNA) has recently attracted attention as a novel biomarker for treatment efficacy and is an important tool for genome profiling [14]. In clinical practice, liquid biopsy using plasma cfDNA is analyzed and enables precision medicine based on their genomic information [15,16]. However, there are several issues regarding the utility of plasma cfDNA in BTC and PC. First, the plasma cfDNA positivity was very low, even in patients with advanced BTC and PC. A previous paper reported 10–40% positivity for cfDNA in blood samples [17-20]. Second, the DNA fragment of plasma cfDNA is smaller than that extracted from resected tumor tissue; therefore, it is occasionally difficult in deep sequence to detect single nucleotide variations (SNVs) and amplification or fusion alterations. Moreover, the alterations in KRAS, SMAD, CTNNB, and TP53 cover 90% of pathogenic alteration in the genomic profiles for PC [9], while the most heterogeneous genetic variations are in BTC [6-8]. Longer DNA fragments are pivotal to identify the full extent of genetic mutations and reveal the mutational landscape of BTC.

This study focused on the utility of bile cfDNA in liquid biopsies. It is unknown whether bile cfDNA can serve as diagnostic marker for the early detection of BTC and PC. In patients with BTC and PC, endoscopic retrograde cholangiography (ERC) is frequently performed for pathological diagnosis or biliary drainage of obstructive jaundice caused by pancreatic-biliary tumors and the collection of bile juice is easy and less invasive through this procedure. The cancer cells of BTC and PC directly drop into the bile juice; therefore, the quality of cfDNA obtained from bile juice is potentially better. The aim of this study was to evaluate the quality of cfDNA in bile juice as a liquid sample and to clarify the potential of cfDNA in genome profiling for precision treatment.

Filtering exome data of the pancreatic-biliary tumors

We obtained exome data of 13 pancreatic-biliary tumors (PBT) invading bile duct (seven cholangiocarcinomas and a gallbladder cancer: BTC, three ampullary cancers: AC, one pancreatic cancer, and one intraductal papillary mucinous carcinoma: PC) in our cohort (Table 1). Only cancerous tissues were available for 10 cases and three cases consisted of pairs of cancerous and non-cancerous tissues for exome sequencing. Fifty-seven percent of the target regions covered over 30 and 99.9% of the reads were mapped (Supplementary Table 1). A summary of the variant calls after filtering the unreliable variant calls using BQSR (see Methods) is shown in Supplementary Table 2. Tumor-only cases showed many more variants than paired samples. The mean numbers of variants were 1205 and 346 for the tumor-only and paired samples, respectively. This suggests that the variant calls from tumor-only samples contain many germline-derived variants that may cause false-positive results in subsequent analyses.

To remove potential germline variants and other artifacts, we annotated all variant calls with 54KJPN [21] and a panel of normal calls generated by merging three non-cancerous exome variants (Figure 1). A summary of the filtered exome variant data is presented in Supplementary Table 3.

Table 1. Clinical characteristics and genomic alterations revealed by whole genomic sequencing in 13 patient samples
Patient No.	Classification	Diagnoses	Age	Gender	T	N	M	Stage	G	NAT	Driver alterations
#1	BTC	Perihilar cholangiocarcinoma	56	Female	2b	1	0	IIIC	2	Yes*	TP53, NRAS, PDE1C
#2	BTC	Perihilar cholangiocarcinoma	67	Male	2a	2	0	IVA	2	No	TP53, ERBB2, KMT2C, PDZRN3
#3	BTC	Perihilar cholangiocarcinoma	67	Male	2b	1	0	IIIC	1	No	MADD
#4	BTC	Distal cholangiocarcinoma	82	Male	2	0	0	IIA	2	No	TP53, PSMC4
#5	BTC	Distal cholangiocarcinoma	78	Male	2	0	0	IIA	2	No	NRAS, PIK3CA, DRD1, ARID1B, SLITRK5, SMAD4, AMER1, TAF1
#6	BTC	Distal cholangiocarcinoma	73	Female	1	0	0	I	3	No	CPZ, KRAS, TP53, SMAD4
#7	BTC	Distal cholangiocarcinoma	90	Male	1	0	0	I	1	No	FBXW7, KMT2C, ERBB3
#8	BTC	Gallbladder carcinoma	78	Male	1b	0	0	IB	1	No	KMT2C, TP53
#9	AC	Ampullary carcinoma	80	Male	3b	2	0	IIIB	3	No	AMER3, RAF1, APC, BRAF
#10	AC	Ampullary carcinoma	79	Female	3b	2	0	IIIB	1	No	PDK4, TP53
#11	AC	Ampullary carcinoma	84	Male	1b	0	0	IB	2	No	DLGAP4
#12	PC	Pancreatic cancer	47	Male	3	1	0	Iib	1	Yes**	RXRG
#13	PC	Intraductal papillary mucinous carcinoma	78	Male	3	0	0	IIA	2	No	DAPK1, WT1
NAT: neoadjuvant therapy, BTC: bile duct cancers, AC: ampullary cancers, PC: pancreatic cacners, TNM classification by UICC 8th
: GCS regimen (gemcitabine + cisplatin + S1) 10 courses, *: GnP regimen (Gemciabine + nab-paclitaxel) 8 courses followed by radiation thearpy (37.5 Gy)

Mutations detected in the pancreatic-biliary tumors invading bile duct

Oncoplots of selected cancer-related genes are shown in Figure 2A. Major cancer driver genes such as TP53 and RAS family genes are mutated in PBTs invading bile duct. Furthermore, SMAD4 and CDKN2A were mutated in three and two cases of PBTs, respectively. However, we did not detect a mutation in the IDH1 gene in the BTCs in our study. The APC gene showed a truncated mutation in the AC case (#9). In addition, this case also showed that a mutation in the BRAF gene may resemble colorectal cancers. There were no apparent oncogenic mutations in the PCs, suggesting that the tumor content of our PC samples might not be sufficient to detect driver gene mutations (Figure 2A). The candidate driver mutations identified in our study are summarized in Supplementary Table 3. We also calculated the tumor mutation burden (TMB) after extensive filtration (see Methods) and found that none of the cases had a TMB > 10 (Figure 2B).

Mutational signature of the pancreatic-biliary tumors invading bile duct

All our samples showed substantial ratios of single-base substitution (SBS)1 (0.27–0.756, Figure 3A, Supplementary Table 4), which is related to aging and appears in almost all cancers. The second most prevalent signature was SBS26 for BTC and AC, which is related to defective DNA mismatch repair. Similarly, SBS6 and 15 (both of which are associated with mismatch repair defects) were observed in BTC and AC. SBS7 is a signature related to ultraviolet exposure that was frequently observed in all the types of cancers analyzed. SBS1 has an unknown etiology mainly observed in hepatocellular carcinoma and was observed in 5 tumor-only cases and most prominently in PCs in this study. SBS30 is related to formalin-fixed paraffin-embedded (FFPE) processing as observed in one case (#10), although the sample was paired with FFPE samples, and the FFPE effects on the sample genomic DNA might be similar to each other. We selected three PBTs with paired non-cancerous tissues for the subtraction of germline variants from the representative data of our mutational signature delineation (Figure 3B).

Table 2. Concentration of cell-free (cf-) DNAs extracted from bile juice and plasma
Samples	Bile cf-DNA (ng/ul)	Plasma cf-DNA (ng/ul)
#1	1875.0	28.5
#2	6165.0	20.4
#3	2223.0	NA
#4	465.0	10.4
#5	3960.0	24.1
#6	1800.0	NA
#7	1411.5	12.1
#8	3150.0	91.2
#9	673.5	40.7
#10	565.5	11.4
#11	134.9	NA
#12	925.5	8.6
#13	0.0	NA

Table 3. Fragment size of cell-free (cf-) DNAs extracted from bile juice and plasma
Samples	Bile cf-DNA (bp)	Plasma cf-DNA (bp)
#1	10202.0	137.0
#2	33755.0	162.0
#3	12969.0	NA
#4	579.0	10.4
#5	1075.0	135.0
#6	1347.0	NA
#7	874.0	119.0
#8	141.0	141.0
#9	9507.0	145.0
#10	15416.0	143.0
#11	13270.0	NA
#12	13749.0	139.0
#13	Not detected	NA

Detecting tumor derived mutations in the bile cfDNA with digital droplet PCR

We subsequently tried to determine whether the cfDNA in the bile juice contained cancer-derived mutations. In only one case (case #13), we could not obtain cfDNA from the bile. We also compared the efficiency of obtaining tumor-derived DNAs from two types of body liquids (bile and plasma) for PBTs. Nine pairs of bile and plasma samples were collected from each patient. Bile contains much more cfDNAs that have longer fragments compared with plasma (median: 1663.7 ng/mL vs. 21.1 ng/mL, p<0.001, 9230.7 vs. 138.6 bp, p<0.001, respectively) (Figure 4, Table 2, 3) and only weak correlations were observed in the concentration of cfDNAs of the two types of liquid (r = 0.250, Table 2). No correlations were observed between the cfDNA concentration and any clinical parameters in our collection (data not shown).

Digital droplet PCR (ddPCR) and Sanger sequencing revealed that the corresponding mutations detected by exome sequencing were present in the bile of the patients (Figure 5, Table 4). The mutation fractions varied among samples. In the case of NRAS p.Q61R in #1 and KRAS p.G12V in #6 (Figure 5A and B), the mutation fractions in ddPCR were 24.9% and 0.29%, respectively. The corresponding allele frequencies of the mutations in the exome of the same samples were 64.4% and 18.4%. We only detected only one case (TP53 p.P151R in #8) with an extra peak in both strands using Sanger sequencing (Figure 5C). The corresponding allelic frequency of the mutation in the exome of the same sample was 10.5%. The target variants of PIK3CA and BRAF were also detected in bile samples of #5 and #9, respectively. These results suggest that bile could contain a relatively high amount of tumor-derived cfDNAs, although the allele frequencies in the next generation sequencing (NGS) data of the DNAs from tissue sections may not be related to that from the bile of the same patients.

Table 4 Validation of target variants in bile cell-free (cf-) DNAs in this study
Samples	cfDNA in bile juice (ng/ml)	Target variants	Method	Validation
#1	1875	NRAS p.Q61R	ddPCR	Succeeded
#4	465	TP53 p.H179L	PCR-Sanger	Failed
#5	3960	NRAS p.G13R	ddPCR	Succeeded
#5	3960	PIK3CA p.H1047L	ddPCR	Succeeded
#6	1800	KRAS p.G12V	ddPCR	Succeeded
#8	3150	TP53 p.P151R	PCR-Sanger	Succeeded
#9	673.5	BRAF p.D594G	ddPCR	Succeeded
#10	565.5	TP53 p.C135F	PCR-Sanger	Failed

Bile cfDNA yielded significantly larger fragments and higher DNA concentrations than plasma cfDNA. Recently, a few reports showed longer fragments of bile cfDNA [22-25], and the length in our cohort was almost consistent with those of previous reports. The size and amount of cfDNA obtained from bile juice showed sufficient quality for integrated sequencing, and genomic profiling could be performed using bile cfDNA rather than plasma cfDNA. The utility of bile cfDNA in cholangiocarcinoma and gallbladder cancer has been reported [22-24]; however, our cohort also included patients with ampullary and pancreatic cancers. Although some ampullary cancers are considered a subtype of intestinal or colon cancer [8], the genetic characteristics of this disease partially share profiles with BTC or PC. Moreover, patients with ampullary cancer, PC, and BTC require clinical intervention for biliary drainage, and bile juice can be easily collected as a liquid sample. To our knowledge, this is the first report demonstrating the usefulness of bile cfDNA for liquid biopsy in any pancreatic-biliary tumors invading bile duct.

Plasma cfDNA is often insufficient in quantity and quality for DNA analysis owing to the minute amount and degradation of plasma DNA, whereas bile juice contains a large amount of tumor-derived DNA because cancer cells are directly spilled from cancerous tissue, and DNA is conserved as long fragments without any degradation [22-24]. The difference in the fragment length of cfDNA between the plasma and bile was owing to the difference in the activity of DNA restriction endonucleases [23]. Furthermore, the reason for the low detection rate of plasma cfDNA is that PC tumors are largely composed of stromal cells with a dense desmoplastic reaction, and a small amount of tumor-derived DNA fragments can reach the blood flow [26]. Although a few reports have indicated that the positivity of plasma cfDNA is sufficient in PC and BTC [26,27] the rate is not as high as in colorectal cancer or other gastrointestinal cancers. In particular, the positive rate of plasma cfDNA in resectable PC is very low, which is approximately 15–50%, compared to other cancers [17,28-30]. In this regard, bile cfDNA is potentially superior to plasma cfDNA in terms of sufficient amount and better quality for DNA analysis of PC and BTC.

Whole-genome sequencing (WGS) of BTC and PC tissues was already reported and indicated their unique characteristics. BTC has heterogeneous alterations and is classified into several types based on genomic profiling [6-8]. In 90% of PC tumors, any of these four representative alterations (KRAS, TP53, SMAD4, or CDKN2A) can be detected [9]. In the current study, characteristic gene alterations in TP53, ABCA8, BRCA2, FGFR2, APC, SMAD4, PIK3CA, BRAF or NRAS were detected in resected BTC tissues, and the most representative alterations in TP53, KRAS, NRAS, PIK3CA or BRAF were also detected in bile cfDNA. Thus, genomic information from bile cfDNA accurately reflects tissue-derived genomic profiles. Generally, the accuracy of pathologic diagnosis of BTC by endoscopic biopsy is relatively low (40–65%) [31,32] and we often face the difficulty of accurate diagnosis by biopsied samples in cases with apparent bile duct stenosis. Regarding ampullary carcinoma, endoscopic biopsy for pathological diagnosis is difficult in cases with non-exposed tumor type [33]. Endoscopic biopsy has recently become common for PC, but this technique is associated with the risk of peritoneal dissemination and perforation [34,35]. In these respects, genomic sequencing of bile cfDNA can provide a less invasive and more accurate diagnosis and assist pathologists in accurate diagnosis. The sensitivity of BTC or PC diagnosis might be improved by the use of bile cfDNA [24].

In plasma cfDNA, the amount of extracted cfDNA or genomic alterations correlates with disease progression in PC [30,36]. In contrast, there was no significant relationship between the amount or fragment size of bile cfDNA and cancer progression in the present study, which is consistent with previous reports [22]. Plasma cfDNA is only detectable in cases of tumors expanding to the subserosal layer, invading lymphatic vessels or microvessels, while bile cfDNA is more likely to be detectable even in the case of a less advanced BTC because the tumor directly drops into bile juice. Furthermore, genome profiling of bile cfDNA suggests a potential for malignancy in cases of bile duct stenosis that can be mistakenly diagnosed as benign [24]. Thus, bile cfDNA is useful to detect precancerous or early-stage PC or BTC lesions. There is also a report on the usefulness of pancreatic juice cfDNA to detect early pancreatic cancer [37].

This study has several limitations. First, we only collected cfDNA from patients with cancer, and the false-positive rate and specificity could not be evaluated. Secondly, the number of objective cohort studies was small. Third, there were a few false-negative cases wherein tissue-derived mutations could not be detected in bile cfDNA (Table 4). The frequency of mutated genes might have been low, which may be the detection limit for ddPCR analysis. Therefore, we should proceed with an integrated analysis using WGS of bile cfDNA. The utility of bile cfDNA in the detection rate of pathogenic alterations compared to plasma cfDNA has not been clarified.

In conclusion, bile cfDNA is more useful than plasma cfDNA, indicating better quality and amount for analysis, and bile cfDNA reflects tumor-derived genetic variation. This report demonstrates the diagnostic utility of cfDNA in genomic profiling and precision medicine.

Preparing exome libraries for NGS

We enrolled 14 patients with tumors in the PBTs invading the bile duct and requiring endoscopic intervention for biliary drainage. After biliary drainage and bile duct stenting, all patients subsequently underwent surgical resections and were pathologically diagnosed with cholangiocarcinoma, gallbladder carcinoma, ampullary carcinoma, PC, or intraductal papillary mucinous carcinoma in seven, one, three, one, and one patients, respectively (Table 1). This study was approved by the Institutional Review Board of our institution and the Ethics Committee of the Tohoku University Graduate School of Medicine (#2023-1-487) and conducted in accordance with the principles of the Declaration of Helsinki. The informed consents were obtained from all patients.

Genomic DNAs from the resected samples was obtained as previously described [38]. Briefly, genomic DNA was extracted from tumor and non-tumor portions of each patient’s FFPE tissue using a QIAGEN kit according to the manufacturer’s protocols. Genomic DNAs was quantified using a Nanodrop (Waltham, MA, USA). The WGS from library construction to variant calls was outsourced to Rhelixa (Tokyo, Japan). The extracted genomic DNAs were subjected to the library construction.

An Agilent SureSelect Human All Exon V6 (58 MB) kit (Agilent Technologies, Santa Clara, CA, USA) was used for DNA target enrichment. Each sample was sequenced using the 150 bp paired-end sequencing method on an Illumina NovaSeq 6000 platform (Illumina, Inc., La Jolla, CA, USA). The quality of raw paired-end sequence reads was assessed using FastQC (version 0.11.7; https://www.bioinformatics.babraham.ac.uk/projects/fastqc/).

Bioinformatics of exome sequencing: mapping, variant call, and filtration

Low-quality (< 20) bases and adapter sequences were trimmed using Trimmomatic software (version 0.38) with the following parameters: ILLUMINACLIP: path/to/adapter. fa:2:30:10 LEADING:20 TRAILING:20 SLIDINGWINDOW:4:15 MINLEN:36 [39]. The trimmed reads were aligned to the indexed reference genome using bwa-mem (v0.7.17-r1188) (arXive. 2013: 1303.3997v1301 q-bio.GN). The generated SAM files were converted into BAM files and sorted using Picard SortSam (version 2.18.11; http://broadinstitute.github.io/picard/). The number of duplicates was calculated using the Picard Mark Duplicate. After removing duplicates, reads were realigned around known indels provided by the Genome Analysis Toolkit (GATK) group [40]. A base quality recalibration table was generated using the GATK BaseRecalibrator (Version 4.0.8.1) with known site files: dbSNP138 from the GATK resource bundle. The resultant base recalibration table was applied using GATK ApplyBQSR (Version 4.0.8.1). Somatic variants were called using GATK MuTect2 in either the tumor-only mode in 10 cases or the normal-tumor pair mode in 3 cases. Variants were filtered using the GATK FilterMutectCalls.

Bioinformatics of the exome sequencing: annotation

The identified variants were annotated using SnpEff (version 4.3t) and ANNOVAR [41] to identify the putative effects on protein translation and high-impact mutations.

The clinical significance of the genetic variants was determined by comparing with ClinVar (https://www.ncbi.nlm.nih.gov/clinvar). We annotated SNP-trait associations from gwasCatalog and 1000 Genome Project data, including African, American, East Asian, European, and South Asian individuals. We used domain information from Interpro and prediction scores from dbNSFP, including CADD, FATHMM, LRT, MetaSVM, MutationAssessor, MutationTaster, Provean, Polyphen-2, Sift, and PhastCons scores to predict the functional consequences of the observed amino acid substitutions.

To subtract potential germline variants from the tumor-only data, we employed the 54 KJPN dataset [21] as a normal panel. The 54 KJPN data were downloaded from jMORP (https://jmorp.megabank.tohoku.ac.jp/downloads: last visit 4th December 2023), and the bcftools norm –m–option was used to separate multiple allelic sites [42]. We then manipulated the dataset with bcftools query options and the ANNOVAR convert2annover.pl command. In the case of subtraction of variants using 54KJPN, we considered that some of the deleterious variants that drive carcinogenesis may be included in 54KJPN. For example, four pathogenic/likely pathogenic variants of TP53 (p.D281N, p.R273C, p.R248Q, and p.G245S) and frequent mutations in the TERT promoter (c.-124C>T) are present in 54KJPN. Therefore, we decided to include the variants found in 54KJPN with minor allele frequencies (MAFs) lower than 0.0002, which is the criterion used in the development of AlphaMissense as weak labelling for benign variants [43]. Furthermore, we annotated using ClinVer (version a 31st December 2022), the cosmic database ver. 91 [44], and the databases were prepared according to the descriptions on the ANNOVAR website (https://annovar.openbioinformatics.org/en/latest/: last visit, 4th December 2023). To further remove potential germline variants and platform-related biases [45], we generated another set of normal from three non-cancerous exome data in our cohort (#2, #6, and #10). After merging the three non-cancerous exome datasets with bcftools, we created a normal panel database using the convert2annovar command in the Annover package.

To calculate the tumor mutation burden, we removed annotated variants using the following databases: 54KJPN with MAF > 0.0002, intronic variants, benign variants in either ClinVar or AlphaMissense, and all synonymous variants divided by the target size of the exome in megabases.

Bioinformatics: Delineating mutational signature analysis for cholangiocarcinoma exome data

We also performed a deconstructSig analysis to trace the mutational signatures detected by Alexandrov et al [46]. To delineate the mutational signatures, we used deconstructSigs [47]. We removed potential non-cancerous variants by subtracting 54KJPN and three non-cancerous exome datasets.

Sample collection of cfDNA from bile juice and plasma

Endoscopic nasobiliary drainage (ENBD) tubes were placed endoscopically for biliary drainage or biopsy of cancerous lesions in the bile duct. We collected 5 mL bile juice from the ENBD tubes. Initially, the bile juice was centrifuged at 1600 g for 15 minutes at room temperature and the supernatants were carefully picked up in tube and stored at -80 ℃. In nine of all patients, we collected 5 mL plasma in the cfDNA collection tube (PAXgene Blood ccfDNA Tubes, Qiagen) and stored at -80 ℃ for analysis. Cell-free DNA was extracted using a QIAamp Circulatin Nucleic Acid Kit (Qiagen) according to the manufacturer’s instructions.

Library preparation, NGS, and bioinformatic analyses of the bile cell-free DNA

Cell-free DNA extracted from plasma or bile juice was subjected to LiquidPlex library construction according to the manufacturer’s instructions. Plasma from three non-cancer patients was used as a control to calculate the background noise. For library construction, 8.2–91.2 ng of Plasma cfDNA, 79.7–300 ng of bile cfDNA was subjected to end-repair, ligation, and addition of molecular barcodes. The cfDNA libraries were amplified using two separate PCR steps, and the index sequences were incorporated into the DNA fragments in the second PCR. The NGS libraries were quantified using the KAPA Universal Library Quantification Kit (KK4824; Roche, Basel, Switzerland) according to the manufacturer’s instructions. MiSeq has also been used for library quantification with an Illumina platform sequencer [48]. Sequencing data acquisition for mutation analysis was outsourced to Rhelixa, Inc. (Tokyo, Japan) using a NovaSeq 6000. A 150 bp paired-end protocol was used to obtain cfDNA sequence data. Sequence read mapping and variant calls were performed using the Archer Analysis platform, a web-based bioinformatics service for LiquidPlex, and other sequencing kits with default options. The GRCh37 coordinates served as the reference genome for mapping. After the variant call, annotations of the called variants were performed using ANNOVAR with a custom-made 54 KJPN variant dataset and COSMIC 94.

Droplet digital PCR of the cfDNAs

A ddPCR Mutation Assay (Bio-Rad, Hercules, CA, USA) was used to select commercially available probes for somatic mutations identified in the cfDNA samples using ddPCR. Two probes were selected to verify whether mutations detected by whole-exome sequencing in the excised sample DNA could be detected in the bile of the same patient. (for KRAS p.G12V: Thermo Fisher Scientific Inc Catalog #: A44177, Assay ID: Hs000000050_rm; for NRAS Q61R: Bio-Rad Laboratories, Inc Catalog #: 10049047, UniqueAssayID: dHsaMDS882187944), and ddPCR was performed using the QX200 AutoDG Droplet Digital PCR IVD system (Bio-Rad). Subsequently, 9.9 μL of the eluted cfDNA solution (average DNA content = 23.3 ng) was used for each ddPCR in a quadlicate manner for each assay points. Genomic DNA was extracted as previously described 15. The total reaction volume was 22 μL (20,000 drops were generated). The PCR conditions were as follows: initial denaturation at 95°C for 10 min; 40 cycle steps at 94°C for 30 s and 55°C or 60°C for 1 min; and final denaturation at 98°C for 10 min for enzyme deactivation. The probe fluorophores were FAM/HEX or VIC, and wild-type and mutant copy numbers were measured. All ddPCR reactions were performed at least twice. Allele frequencies were calculated using Quanta Software (Bio-Rad) according to the manufacturer’s instructions. Because the plasma cfDNAs were scarcely obtained and we used up to the cfDNA for construction of libraries of multi-gene panels, we could not test the plasma cfDNA with ddPCR. The results of plasma multi-gene panel liquid biopsies will be described elsewhere.

Sanger sequencing of cfDNA in the bile

We performed Sanger sequencing of the PCR amplicons at exon 5 of the TP53 gene for three cases (H21_10688_26, p.H179L; H22_161_19, p.P151R; and H23_1363_21, p.C135F) to confirm whether bile cfDNA included tumor-derived cfDNA. The primers for amplicon sequencing were as follows: TP53hg19_chr17_7578593F, 5′-GCCCTGACTTTCAACTCTGTCTC-3′ and TP53hg19_chr17_7578120R, 5′-GGCCACTGACAACCACCCTTAACC-3′ for exons 5 and 6. Sequencing analyses were performed by Fillgen (Nagoya, Japan).

Conflicts of interest: The authors declare no conflicts of interest.

Author Contributions: SA and JY developed the concept. MA and SI designed the experimental studies, analyzed the data, and wrote the manuscript. JZ, NH, RK, KM, SY, HS, AK, MI, KI, MM, KN and MU collected the clinical data, and SA, JY, MK, and SI analyzed and interpreted the data. All authors had access to the study data, and reviewed and approved the final manuscript.

Data availability: The datasets generated and/or analyzed during the current study are available in the SRA repository, https://www.ncbi.nlm.nih.gov/sra. The ID and project name is SUB14391254 and Exome data of cancerous tumors in pancreatic head lesions, respectively. We did not include germline exome data in the dataset submitted to the SRA. They will be available upon request.

Ethics approval and consent to participate: This study was approved by the Institutional Review Board of our institution and the Ethics Committee of the Tohoku University Graduate School of Medicine (#2023-1-487). This study was conducted in accordance with the principles of the Declaration of Helsinki.

Acknowledgements: The authors express their deep gratitude to all surgeons and physicians who performed the endoscopic intervention and resection at Tohoku University Hospital.

Founding information: This work was supported in part by a Japan Society for the Promotion of Science (JSPS) KAKENHI grants (20K17672 to SA and 21K08748 to NK) and a Grand-in-Aid from KUROKAWA CANCER RESEARCH FOUNDATION.

Siegel, R. L., Miller, K. D., Fuchs, H. E. & Jemal, A. Cancer statistics, 2022. CA Cancer J Clin 72, 7-33 (2022). https://doi.org/10.3322/caac.21708
Ilyas, S. I. et al. Cholangiocarcinoma - novel biological insights and therapeutic strategies. Nat Rev Clin Oncol 20, 470-486 (2023). https://doi.org/10.1038/s41571-023-00770-1
Von Hoff, D. D. et al. Increased survival in pancreatic cancer with nab-paclitaxel plus gemcitabine. N Engl J Med 369, 1691-1703 (2013). https://doi.org/10.1056/NEJMoa1304369
Conroy, T. et al. FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med 364, 1817-1825 (2011). https://doi.org/10.1056/NEJMoa1011923
Valle, J. et al. Cisplatin plus gemcitabine versus gemcitabine for biliary tract cancer. N Engl J Med 362, 1273-1281 (2010). https://doi.org/10.1056/NEJMoa0908721
Nakamura, H. et al. Genomic spectra of biliary tract cancer. Nat Genet (2015). https://doi.org/10.1038/ng.3375
Jusakul, A. et al. Whole-Genome and Epigenomic Landscapes of Etiologically Distinct Subtypes of Cholangiocarcinoma. Cancer Discov 7, 1116-1135 (2017). https://doi.org/10.1158/2159-8290.CD-17-0368
Yachida, S. et al. Genomic Sequencing Identifies ELF3 as a Driver of Ampullary Carcinoma. Cancer Cell 29, 229-240 (2016). https://doi.org/10.1016/j.ccell.2015.12.012
Bailey, P. et al. Genomic analyses identify molecular subtypes of pancreatic cancer. Nature 531, 47-52 (2016). https://doi.org/10.1038/nature16965
Bekaii-Saab, T. S., Bridgewater, J. & Normanno, N. Practical considerations in screening for genetic alterations in cholangiocarcinoma. Ann Oncol 32, 1111-1126 (2021). https://doi.org/10.1016/j.annonc.2021.04.012
Lamarca, A., Barriuso, J., McNamara, M. G. & Valle, J. W. Molecular targeted therapies: Ready for "prime time" in biliary tract cancer. J Hepatol 73, 170-185 (2020). https://doi.org/10.1016/j.jhep.2020.03.007
Yachida, S. & Iacobuzio-Donahue, C. A. Evolution and dynamics of pancreatic cancer progression. Oncogene 32, 5253-5260 (2013). https://doi.org/10.1038/onc.2013.29
Yachida, S. et al. Distant metastasis occurs late during the genetic evolution of pancreatic cancer. Nature 467, 1114-1117 (2010). https://doi.org/10.1038/nature09515
Cohen, J. D. et al. Detection and localization of surgically resectable cancers with a multi-analyte blood test. Science 359, 926-930 (2018). https://doi.org/10.1126/science.aar3247
Yang, J. D. et al. DNA Methylation Markers for Detection of Cholangiocarcinoma: Discovery, Validation, and Clinical Testing in Biliary Brushings and Plasma. Hepatol Commun 5, 1448-1459 (2021). https://doi.org/10.1002/hep4.1730
Mody, K. et al. Circulating Tumor DNA Profiling of Advanced Biliary Tract Cancers. JCO Precis Oncol 3, 1-9 (2019). https://doi.org/10.1200/PO.18.00324
Groot, V. P. et al. Circulating Tumor DNA as a Clinical Test in Resected Pancreatic Cancer. Clin Cancer Res 25, 4973-4984 (2019). https://doi.org/10.1158/1078-0432.CCR-19-0197
Lee, B. et al. Circulating tumor DNA as a potential marker of adjuvant chemotherapy benefit following surgery for localized pancreatic cancer. Ann Oncol 30, 1472-1478 (2019). https://doi.org/10.1093/annonc/mdz200
Nakano, Y. et al. KRAS mutations in cell-free DNA from preoperative and postoperative sera as a pancreatic cancer marker: a retrospective study. Br J Cancer 118, 662-669 (2018). https://doi.org/10.1038/bjc.2017.479
Pietrasz, D. et al. Plasma Circulating Tumor DNA in Pancreatic Cancer Patients Is a Prognostic Marker. Clin Cancer Res 23, 116-123 (2017). https://doi.org/10.1158/1078-0432.CCR-16-0806
Tadaka, S. et al. jMorp: Japanese Multi-Omics Reference Panel update report 2023. Nucleic Acids Res 52, D622-D632 (2024). https://doi.org/10.1093/nar/gkad978
Han, J. Y. et al. Liquid Biopsy from Bile-Circulating Tumor DNA in Patients with Biliary Tract Cancer. Cancers (Basel) 13 (2021). https://doi.org/10.3390/cancers13184581
Shen, N. et al. Bile cell‑free DNA as a novel and powerful liquid biopsy for detecting somatic variants in biliary tract cancer. Oncol Rep 42, 549-560 (2019). https://doi.org/10.3892/or.2019.7177
Arechederra, M. et al. Next-generation sequencing of bile cell-free DNA for the early detection of patients with malignant biliary strictures. Gut (2021). https://doi.org/10.1136/gutjnl-2021-325178
Jiang, P. et al. Lengthening and shortening of plasma DNA in hepatocellular carcinoma patients. Proc Natl Acad Sci U S A 112, E1317-1325 (2015). https://doi.org/10.1073/pnas.1500076112
Sausen, M. et al. Clinical implications of genomic alterations in the tumour and circulation of pancreatic cancer patients. Nat Commun 6, 7686 (2015). https://doi.org/10.1038/ncomms8686
Zill, O. A. et al. Cell-Free DNA Next-Generation Sequencing in Pancreatobiliary Carcinomas. Cancer Discov 5, 1040-1048 (2015). https://doi.org/10.1158/2159-8290.CD-15-0274
Takai, E. et al. Clinical Utility of Circulating Tumor DNA for Molecular Assessment and Precision Medicine in Pancreatic Cancer. Adv Exp Med Biol 924, 13-17 (2016). https://doi.org/10.1007/978-3-319-42044-8_3
Hata, T. et al. Circulating tumor DNA as a predictive marker for occult metastases in pancreatic cancer patients with radiographically non-metastatic disease. J Hepatobiliary Pancreat Sci 28, 648-658 (2021). https://doi.org/10.1002/jhbp.993
Patel, H. et al. Clinical correlates of blood-derived circulating tumor DNA in pancreatic cancer. J Hematol Oncol 12, 130 (2019). https://doi.org/10.1186/s13045-019-0824-4
Kawashima, H. et al. Endoscopic management of perihilar cholangiocarcinoma. Dig Endosc 34, 1147-1156 (2022). https://doi.org/10.1111/den.14317
Yang, M. J. et al. Factors affecting the diagnostic yield of endoscopic transpapillary forceps biopsy in patients with malignant biliary strictures. J Gastroenterol Hepatol 36, 2324-2328 (2021). https://doi.org/10.1111/jgh.15497
DeOliveira, M. L., Triviño, T. & de Jesus Lopes Filho, G. Carcinoma of the papilla of Vater: are endoscopic appearance and endoscopic biopsy discordant? J Gastrointest Surg 10, 1140-1143 (2006). https://doi.org/10.1016/j.gassur.2006.05.005
Carrara, S. et al. Pancreatic endoscopic ultrasound-guided fine needle aspiration: complication rate and clinical course in a single centre. Dig Liver Dis 42, 520-523 (2010). https://doi.org/10.1016/j.dld.2009.10.002
Matsumoto, K. et al. Endoscopic ultrasound-guided fine-needle aspiration biopsy - Recent topics and technical tips. World J Clin Cases 7, 1775-1783 (2019). https://doi.org/10.12998/wjcc.v7.i14.1775
Schlick, K. et al. Evaluation of circulating cell-free KRAS mutational status as a molecular monitoring tool in patients with pancreatic cancer. Pancreatology 21, 1466-1471 (2021). https://doi.org/10.1016/j.pan.2021.09.004
Okada, T. et al. Utility of "liquid biopsy" using pancreatic juice for early detection of pancreatic cancer. Endosc Int Open 6, E1454-E1461 (2018). https://doi.org/10.1055/a-0721-1747
Kokumai, T. et al. GATA6 and CK5 Stratify the Survival of Patients With Pancreatic Cancer Undergoing Neoadjuvant Chemotherapy. Mod Pathol 36, 100102 (2023). https://doi.org/10.1016/j.modpat.2023.100102
Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114-2120 (2014). https://doi.org/10.1093/bioinformatics/btu170
McKenna, A. et al. The Genome Analysis Toolkit: a MapReduce framework for analyzing next-generation DNA sequencing data. Genome Res 20, 1297-1303 (2010). https://doi.org/10.1101/gr.107524.110
Wang, K., Li, M. & Hakonarson, H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res 38, e164 (2010). https://doi.org/10.1093/nar/gkq603
Danecek, P. et al. Twelve years of SAMtools and BCFtools. Gigascience 10 (2021). https://doi.org/10.1093/gigascience/giab008
Cheng, J. et al. Accurate proteome-wide missense variant effect prediction with AlphaMissense. Science 381, eadg7492 (2023). https://doi.org/10.1126/science.adg7492
Tate, J. G. et al. COSMIC: the Catalogue Of Somatic Mutations In Cancer. Nucleic Acids Res 47, D941-D947 (2019). https://doi.org/10.1093/nar/gky1015
Miyabe, S. et al. Clinical and genomic features of non-small cell lung cancer occurring in families. Thorac Cancer 14, 940-952 (2023). https://doi.org/10.1111/1759-7714.14825
Alexandrov, L. B. et al. Signatures of mutational processes in human cancer. Nature 500, 415-421 (2013). https://doi.org/10.1038/nature12477
Rosenthal, R., McGranahan, N., Herrero, J., Taylor, B. S. & Swanton, C. DeconstructSigs: delineating mutational processes in single tumors distinguishes DNA repair deficiencies and patterns of carcinoma evolution. Genome Biol 17, 31 (2016). https://doi.org/10.1186/s13059-016-0893-4
Katsuoka, F. et al. An efficient quantitation method of next-generation sequencing libraries by using MiSeq sequencer. Anal Biochem 466, 27-29 (2014). https://doi.org/10.1016/j.ab.2014.08.015

No competing interests reported.

SupplementaryTable20240420.xlsx

Bile liquid biopsy is a useful modality for molecular diagnosis of cancerous tumors in pancreatic head lesions

Status:

Version 1

Abstract

Figures

Introduction

Results

Filtering exome data of the pancreatic-biliary tumors

Mutations detected in the pancreatic-biliary tumors invading bile duct

Mutational signature of the pancreatic-biliary tumors invading bile duct

Detecting tumor derived mutations in the bile cfDNA with digital droplet PCR

Discussion

Materials and Methods

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1