Intratumoral heterogeneity of intrahepatic cholangiocarcinoma

Abstract

No personalized therapy regimens could demonstrate a benefit in survival of intrahepatic cholangiocarcinoma (iCCA). Since genetic heterogeneity might influence single biopsy based targeted therapy or the outcome of clinical trials, aim of the present study was to investigate intratumoral heterogeneity of iCCA by whole exome sequencing. Therefore, samples from tumor center and tumor periphery of large iCCA lesions as well as a control from healthy liver tissue were obtained from four patients and whole exome sequencing was performed. Mutations that occurred only in the tumor center or periphery were defined as private, whereas mutations present in both samples were regarded as common. A mean of 3 non-synonymous private mutations (range 0-14) per sample compared to 33,3 common mutations per sample (range 24-41) was identified. Mean percentage of non-synonymous private mutations per sample was 12% (range 0-58). In all samples of patient 1-3 as well as the central sample of patient 4 ≤ 10% private mutations were found, whereas 58% of private mutations were identified in the peripheral sample of patient 4. In this sample a private mutation in the DNA mismatch repair protein MSH6 could be identified most likely causing the high amount of private mutations. No substantial intratumoral heterogeneity was found in copy number variation analysis. In conclusion, iCCA show a small but distinct intratumoral heterogeneity. Somatic mutations in mismatch repair proteins might contribute significantly to increased spatial tumor burden and thereby may influence clinical management.

Keywords: MSH6; cholangiocarcinoma; exome sequencing; intratumoral heterogeneity; microsatellite instability.

Figure 1

(A) Scheme of the tumor regions samples were taken from for whole exome sequencing. (B) Mutational spectra of all four patients. (C) Substitution pattern of common and private mutations. N-Tu: Non-tumor sample, Tu-p: peripheral tumor sample, Tu-c: central tumor sample.

Figure 2. Overview of all recurrent (light grey), common (dark grey) and private (black) synonymous and non-synonymous mutated genes of this study

Tu-c: central tumor sample, Tu-p: peripheral tumor sample.

Figure 3

Representative images of HE and MSH6 stained FFPE slides of tumor center (A: 60×, B: 200×, C: 400×) and tumor front (E: 60×, F: 200×, G: 400×) as well as corresponding pyrogram including detected mutation frequency (D: tumor center, H: tumor front) of macrodissected tissue of patient 4 with C>T missense mutation at position 48010469, chromosome 2. In the tumor center, a glandular growth pattern with clear MSH6 expression and no mutation was observed in contrast to a dedifferentiated growth pattern, no MSH6 expression and 25% of mutant allele frequency at the tumor front.

Figure 4. Representative copy number variations (CNV) on chromosomes 5-7 of patient 1-4

Of each patient CNV of non-tumor sample (black line), central tumor sample (red line) and peripheral sample (blue line) is shown. In some fractions of the genome CNV between both tumor samples and control were observed (dotted arrow), whereas in chromosome 6 the copy numbers between both tumor samples varied. X-axis: chromosome length (megabases), y-axis: deviation from expected coverage, upper dotted line = +50%, lower dotted line = −50%.