Evaluation of Optimal Biopsy Location for Assessment of Histological Activity, Transcriptomic and Immunohistochemical Analyses in Patients With Active Crohn's Disease

Gregor Novak; Toer Stevens; Tanja Van Viegen; Peter Bossuyt; Borut Štabuc; Jenny Jeyarajah; Guangyong Zou; Ingrid C. Gaemers; Trevor D. McKee; Fred Fu; Lisa M. Shackelton; Reena Khanna; Gijs R. van den Brink; William J. Sandborn; Brian G. Feagan; Rish K. Pai; Vipul Jairath; Niels Vande Casteele; Geert D'Haens


Aliment Pharmacol Ther. 2019;49(11):1401-1409. 

In This Article


Study Design and Participants

Data were obtained from a prospective cross-sectional study conducted in three centres (Imelda General Hospital, Bonheiden, Belgium; Academic Medical Centre, Amsterdam, the Netherlands [coordinating center]; and Ljubljana University Medical Centre, Ljubljana, Slovenia) between June 2016 and January 2017. The local ethics committee at each of the centres approved the protocol. All patients provided written informed consent prior to study enrolment.

Eligible patients were at least 18 years of age with Crohn's disease of the terminal ileum (L1), colon (L2) or ileocolon (L3) according to the Montreal classification[15] with at least one Crohn's disease-related ulcer >0.5 cm in diameter at the time of ileocolonoscopy. Ileocolonoscopies were performed as part of routine clinical care. Patients receiving systemic anticoagulation therapy and pregnant patients were excluded.


All potentially eligible patients underwent physical examination with measurement of vital signs, and collection of demographic data and medical history prior to ileocolonoscopy. Ileocolonoscopies were video-recorded using a central image management system and endoscopic disease activity was assessed using the Simple Endoscopic Score for Crohn's disease[16] by a single blinded central reader.

Endoscopists were instructed to measure the distance from an ulcer edge with a standard open biopsy forceps (7-8 mm in diameter; EndoJaw Alligator Jaw-step fenestrated with needle, FB220-U, Olympus [Leiderdorp, The Netherlands]).

In addition to procurement of biopsies as part of routine care (if required), six biopsy specimens (three for histopathological and immunohistochemical assessment, and three for mRNA gene expression analysis) were obtained in the following manner: (a) two adjacent to the edge of the largest ulcer present in the colon and/or (neo)terminal ileum; (b) two at a distance of one open forceps (7-8 mm) from the edge of the largest ulcer in non-ulcerated mucosa; (c) two at a distance of three open forceps (21-24 mm) from the edge of the largest ulcer in non-ulcerated mucosa, further referred to as locations 1, 2 and 3 respectively. If ulcers were present in both the (neo)terminal ileum and colon, six biopsies were taken adjacent to the largest ulcer in both the colon and the ileum (12 total). If ulcers were present only in the (neo)terminal ileum or colon, six biopsies were taken adjacent to the largest ulcer (six in total). Endoscopists were instructed to avoid biopsies of ulcers at the anastomosis in the setting of post-operative recurrence of Crohn's disease due to possible ischaemic aetiology. Correct biopsy procurement and subsequent eligibility of samples were confirmed by the same central reader previously mentioned.

Individual biopsies were collected into separate tubes. Those planned for histopathology and immunohistochemistry were stored in 10 mL 10% buffered formalin, whereas those planned for mRNA gene expression analysis were stored in 2 mL Eppendorf tubes pre-filled with RNA stabilisation solution (RNALater®, Ambion Inc, Austin, TX). All biopsies were shipped at room temperature to the central laboratory (Tytgat Institute for Liver and Intestinal Research, Amsterdam UMC, The Netherlands) within 48 h of collection.


After routine processing at the central laboratory, biopsies were sectioned at 4.5 μm, stained with haematoxylin and eosin and digitised prior to review. To avoid interobserver variability,[17] a single-blinded expert gastrointestinal pathologist (RKP) trained in the scoring of histological indices used in the assessment of inflammatory bowel disease (IBD) activity scored each of the digitised slides with the Global Histological Disease Activity Score (GHAS),[8,18] Robarts Histopathology Index (RHI)[19] and the Nancy Histological Index (NHI).[20] The GHAS is a numerical histological disease activity score that assesses eight items. We modified the GHAS to include assessment of the 'percentage of affected surface in a single biopsy' rather than 'number of biopsy specimens affected' as only one biopsy per location was available for assessment as opposed to multiple biopsies in the original study. The RHI is a continuous histological disease activity scale that assesses four items. Scores range from 0 (no disease activity) to 33 (severe disease activity).[19] The NHI is a stepwise ordinal histological disease activity scale that assesses three items, and ranges from a grade of 0 (the absence of significant histological disease activity) to a grade of 4 (severely active disease).[20] Although the RHI and NHI were developed and validated for use in the assessment of ulcerative colitis histological disease activity, we included these outcomes to provide preliminary information on their potential future utility for assessment of histological disease activity in Crohn's disease.


Immunohistochemical staining of slides containing 4.5 μm biopsy sections with antibodies against CD68 (Dako/Agilent, Amstelveen, The Netherlands, M0876 mouse monoclonal αCD68 clone RPM1), CD3 (Thermo Scientific, Ermelo, The Netherlands, RM-9107, rabbit monoclonal αCD3e clone SP7) and myeloperoxidase (Dako/Agilent, Amstelveen, The Netherlands, A0398 rabbit polyclonal α-myeloperoxidase) was performed on a Roche Ventana Benchmark Ultra IHC autostainer (Roche Diagnostics Nederland BV, Almere, The Netherlands) using a Roche Optiview DAB detection kit (Roche Diagnostics Nederland BV, Almere, The Netherlands) as follows: (a) antigen retrieval for 24 minutes (CD68 and myeloperoxidase) or 48 minutes (CD3) at 100°C in Roche CC1 buffer (pH 8.5), followed by (b) incubation with αCD3, αCD68 or α-myeloperoxidase at dilutions of 1:100, 1:200 and 1:4000, respectively, for 32 minutes at room temperature. Counterstaining of sections with haematoxylin was performed automatically on the Benchmark Ultra stainer.

Stained slides were subsequently scanned on a Philips Intellisite Ultrafast scanner (Philips, Eindhoven, The Netherlands) at 40X magnification. Whole slide images in Philips TIFF format were converted for processing into standard TIFF images using the GDAL/OGR Geospatial Data Abstraction software Library, version 2.3.2 (Open Source Geospatial Foundation, http://gdal.org/), and loaded into Definiens Tissue Studio 4.3.1 (Definiens AG, Munich Germany) for brightfield analysis (0.25 μm per pixel, 40 magnification). Images were pre-processed to identify relevant regions of interest on each slide and manual correction was performed to exclude various artifacts, tissue folds, regions that were out of focus or exhibited poor scanning quality, completely unstained large tissue fragments and regions with stained epithelial cells. Colour deconvolution was used to isolate the haematoxylin and DAB stain channels, and nuclei were segmented in the tissue region of interest using a haematoxylin stain intensity threshold of 0.05 and an average nuclear size of 30 μm2. Cell simulation was performed by growing the nuclear regions. Positive cells were detected using DAB intensity thresholds in the cell, where intensities less than 0.3, between 0.3 and 0.4, and greater than 0.4 were denoted as negative, 'low', and 'high' for the marker of interest. Thresholds were chosen to classify the cellular areas, set by visual inspection of multiple tissue sections with positive and negative cells present. Three thresholds corresponding to more permissive (low-stain intensity) to more restrictive (high-stain intensity) were chosen.

Total number of marker-positive cells per mm2 of analysed tissue area was calculated. Positive cells included those classified as both 'low' and 'high'. Analysed tissue area was determined by adding areas of the tissue and stained epithelium regions of interest.

mRNA Analysis

Biopsy samples for mRNA analysis were homogenised using TissueLyser LT (Qiagen, Venlo, The Netherlands). RNA was extracted from the biopsy specimens using the Bioline Isolate II RNA Mini Kit (GC Biotech, Alphen a/d Rijn, The Netherlands), according to the manufacturer's instructions. Total RNA concentration was measured by NanoDrop 1000 Spectrophotometer (Thermo Fisher Scientific, Ermelo, The Netherlands). First strand complementary DNA was synthesised from mRNA using RevertAid reverse transcriptase (Fermentas; Thermo Fisher Scientific), according to the manufacturer's instructions. The expression of five genes previously shown to be significantly up-regulated in colonic and/or ileal mucosa of Crohn's disease patients in the setting of acute inflammation and known to be implicated in the pathogenesis of Crohn's disease was analysed, including interleukin (IL)-6,[14,21] IL-8,[14,22,23] and IL-23 (p19 and p40 subunits),[21,24,25] platelet endothelial cell adhesion molecule (PECAM)-1 (or CD31)[22,26] and S100A9, a subunit of calprotectin.[14,27] Expression levels of mRNA for the genes of interest as well as reference genes (β-actin and cyclophilin) were determined in duplicate by quantitative real-time reverse transcriptase polymerase chain reaction using SensiFAST SYBR® No-ROX Kit (GC-Biotech, Alphen a/d Rijn, The Netherlands) according to the manufacturer's instructions on a LightCycler instrument (Roche, Alphen a/d Rijn, The Netherlands) and gene-specific primers (See Table S1 for primers used). Relative mRNA expression levels for the genes of interest were calculated as a ratio relative to the geometric mean of the endogenous reference genes (β-actin and cyclophilin).

Safety Assessment

Adverse and serious adverse events that occurred during, and 1- or 4 weeks post-ileocolonoscopy, respectively, were recorded. The intensity and potential relationship of both serious and nonserious adverse events to biopsy sampling were determined by the investigators.

Statistical Analysis

A mixed effects model approach was adopted to analyse each of the histopathological disease activity scores (GHAS, NHI and RHI), mRNA expression levels of the genes of interest and number of marker-positive inflammatory cells in the lamina propria, with the latter two analysed on the log scale to account for non-normality of the data. The mixed effects model approach was taken to adjust for within-subject correlations among locations (1, 2 and 3) and segments (colon and ileum).[28] The models included the fixed, categorical effects of locations, segments and their interactions. An unstructured (co)variance structure shared across locations was used to model the within-patient errors. The Kenward-Roger approximation was used to estimate denominator degrees of freedom and adjust standard errors.[29]

The least-square contrasts were used to compare locations in terms of histological disease activity, mRNA expression levels for the genes of interest and proportions of marker-positive inflammatory cells. A two-sided P value of 0.05 was the criterion for statistical significance.

Due to the lack of relevant data in the literature, the sample size was calculated using an estimation approach based upon Cohen's standardised effect size, defined as the mean difference divided by the SD.[30] Assuming a comparison of scores between any two locations within a patient using a paired t test, a sample of 50 patients (pairs) was sufficient to detect an effect size of at least 0.4 with 80% power at the 5% significance level. Sample size distributions for each analysis performed are provided in Table S2.