Individual crypt genetic heterogeneity and the origin of metaplastic glandular epithelium in human Barrett's oesophagus

Abstract

Objectives: Current models of clonal expansion in human Barrett's oesophagus are based upon heterogenous, flow-purified biopsy analysis taken at multiple segment levels. Detection of identical mutation fingerprints from these biopsy samples led to the proposal that a mutated clone with a selective advantage can clonally expand to fill an entire Barrett's segment at the expense of competing clones (selective sweep to fixation model). We aimed to assess clonality at a much higher resolution by microdissecting and genetically analysing individual crypts. The histogenesis of Barrett's metaplasia and neo-squamous islands has never been demonstrated. We investigated the oesophageal gland squamous ducts as the source of both epithelial sub-types.

Methods: Individual crypts across Barrett's biopsy and oesophagectomy blocks were dissected. Determination of tumour suppressor gene loss of heterozygosity patterns, p16 and p53 point mutations were carried out on a crypt-by-crypt basis. Cases of contiguous neo-squamous islands and columnar metaplasia with oesophageal squamous ducts were identified. Tissues were isolated by laser capture microdissection and genetically analysed.

Results: Individual crypt dissection revealed mutation patterns that were masked in whole biopsy analysis. Dissection across oesophagectomy specimens demonstrated marked clonal heterogeneity, with multiple independent clones present. We identified a p16 point mutation arising in the squamous epithelium of the oesophageal gland duct, which was also present in a contiguous metaplastic crypt, whereas neo-squamous islands arising from squamous ducts were wild-type with respect to surrounding Barrett's dysplasia.

Conclusions: By studying clonality at the crypt level we demonstrate that Barrett's heterogeneity arises from multiple independent clones, in contrast to the selective sweep to fixation model of clonal expansion previously described. We suggest that the squamous gland ducts situated throughout the oesophagus are the source of a progenitor cell that may be susceptible to gene mutation resulting in conversion to Barrett's metaplastic epithelium. Additionally, these data suggest that wild-type ducts may be the source of neo-squamous islands.

Figure 1. Laser capture microdissection. A(i) Individual crypts were identified using haematoxylin & eosin (H&E) slides A(ii) Cytokeratin 7 immunostaining of serial sections showing glandular differentiation of columnar lined oesophagus. A(iii), (iv) Individual crypts were microdissected from serial sections stained with methylene green and mounted on laser capture slides. B(i) Squamous islands were identified histologically using H&E slides. B(ii) Cytokeratin 13 immunostaining of serial sections showing mature squamous cell differentiation. B(iii) Squamous island after laser capture showing selective dissection of only squamous tissue. C(i), (ii) Individual structures were dissected evenly from across the block and numbered.

Figure 2. Oesophageal biopsy clonal map. Individual crypts, squamous islands and areas of lamina propria were microdissected from oesophageal biopsies and analysed for tumour suppressor gene loss of heterozygosity (LOH). An unstained serial section of the whole biopsy was then scraped into proteinase K using a clean scalpel blade, and the lysate analysed. LOH of individual alleles is denoted by a blue- or red-coloured box. The table is the clonal map obtained from a single biopsy and each column within the table represents genetic analysis of a single crypt or area of lamina propria. In each case the shortest allele is referred to as A1 and the longest A2. Non-informative markers are denoted by a black box. Individual crypts often demonstrated LOH not detectable on the whole biopsy section lysate probably as a consequence of a diluting effect of wild-type stroma in the whole biopsy lysate. Other biopsy clonal maps are presented in the supplementary information (supplementary fig 1). H&E, haematoxylin & eosin.

Figure 3. Clonal maps of two blocks from patient 1. Each table is the clonal map for the oesophagectomy specimen pictured and each column within the table represents the genetic analysis of a single crypt, squamous island or area of lamina propria. p16 point mutations are denoted by green- or yellow-coloured boxes as per the key. White boxes are wild-type. Clonal analysis revealed regional similarities in loss of heterozygosity (LOH) patterns correlating with the observed phenotypic differences; however, there was no evidence of tissue-wide selective sweeps and no fixed founder mutations indicating a common ancestral precursor. Two different, independent p16 point mutations were identified in the different blocks suggesting at least two distinct clones. Only one squamous island contained a mutation and this is comparable with the results described by Paulson et al Clonal maps from other patients are presented in the supplementary information (supplementary fig 2).

Figure 4. (A) Oesophageal gland squamous ducts give rise to neo-squamous islands. The haematoxylin & eosin (H&E) slides show a squamous island originating from an oesophageal gland squamous duct and encroaching onto a field of Barrett’s dysplasia. Dissection of the continuous dysplastic tissue revealed a p53 non-sense mutation, which was not present in the separately dissected squamous tissue and squamous duct. It has been previously noted that squamous islands are often associated with oesophageal glands, and this demonstrates the oesophageal gland duct as the source of these wild-type islands. (B) Oesophageal gland squamous ducts give rise to metaplastic columnar epithelium. The H&E slides show a metaplastic glandular crypt arising from a contiguous squamous duct with a clear transition from squamous to columnar epithelium (black arrow). The same mutation was found in both the squamous and columnar epithelial tissue suggesting a clonal origin.

Figure 5. Clonal evolution models in Barrett’s oesophagus. (A) The current model of clonal evolution adapted from Maley et al. A founder mutation (red cross) occurs in a single progenitor and provides a growth advantage that predisposes to a selective sweep. Successive selective sweeps result in progression along the metaplasia–dysplasia pathway. Clone bifurcation is responsible for clonal heterogeneity in this model. (B) Newly proposed model of evolution based on the mutation of multiple progenitor cells situated in the oesophageal gland squamous ducts located throughout the length of the oesophagus (red crosses). Multiple independent clones then arise which evolve separately. The presence of multiple different clones gives rise to a mosaic interdigitating clonal pattern of the Barrett’s segment represented here by the striped areas.