Fig 1.
Western blot analysis to demonstrate the specificity of the 32–5600 (12H2) antibody to recognise Claudin-2 protein transcript.
40 μg of each cell lysate and 40 ng recombinant human GST tagged claudin-2 was loaded. Lane order is as follows: (A & I) magic marker, (B) HT29 cell lysate, (C) T84 cell lysate, (D & F) empty lane, (E) recombinant GST-Claudin-2 protein, (G) CHO K1 (overexpressing GFP-CLDN2 construct) cell lysate and (H) CHO K1 (untransfected control) cell lysate. 12H2 clearly detects recombinant GST-CLDN2 protein (lane E, 52kDa). CHO-K1 cells do not express CLDN2, as seen in lane H, however upon transient transfection of CHO-K1 cells with a GFP-CLDN2 construct (53kDa), the antibody detects the overexpressed construct. 12H2 is able to detect endogenous levels of Claudin-2 protein (MW ~20kDa) in the HT29 & T84 cell lysates (B & C respectively), as confirmed by mass spectrometry (S2 Fig).
Table 1.
Outcome of the five commercial antibodies in the screening process.
Fig 2.
Lead candidate antibody staining patterns for claudin-2 in cell models.
Fig 2 shows claudin-2 staining, using (A) the lead candidate antibody (12H2) or (B) isotype control in diseased UC. Panels (C-E) show the claudin-2 staining patterns in ICC FFPE cell pellets using 12H2. Specifically, (C) CLDN2 overexpressing CHO K1 cells, (D) untransfected CHO K1 cells and (E) T84 cells. The scale bars represent 50 μm.
Fig 3.
Comparison of claudin-2 ISH signal in affected and unaffected UC colon biopsies.
Fig 3 shows in situ hybridisation (ISH) staining in CLDN 2 overexpressing CHO K1 cells using QuantiGene ViewRNA (A) claudin-2 probe (VA1-12003-01) and (B) SENSE probe (VA1-14363-01) from Panomics. Adjacent sections from unaffected UC colon biopsy material were H&E stained (C) or claudin-2 ISH stained (D); adjacent sections from affected UC colon biopsy material were H&E stained (E) or claudin-2 ISH stained (F). The scale bars represent 50 μm.
Fig 4.
Comparative analysis of claudin-2 protein and mRNA localisation in adjacent sections from affected UC colon.
(A) H&E staining, (B) IHC claudin-2 staining and (C) ISH claudin-2 staining. The localisation of claudin-2 protein (IHC staining) correlates to observed mRNA (ISH staining) pattern. The scale bars represent 50 μm.
Fig 5.
Assay variability for the Claudin-2 ISH method using FFPE cell controls.
Three analysts each performed 2 independent analytical runs using positive (HT29) (A and B) and negative (CHO K1) (C and D) Formalin Fixed Paraffin Embedded (FFPE) cell lines. Cells were stained with claudin-2 (Cldn-2) probe (VA1-12003-01) (A and C) or Sense probe (VA1-14363-01) (B and D) from Panomics. Variability within & between analytical batches (performed by a single analyst) was minimal. Moreover, no analyst bias was observed across the 3 analysts’ performance. Data presented are ISH scores assigned by a single trained analyst.
Fig 6.
Effect of prolonged time intervals (> 1 day), following biopsy section pre-treatment, on ISH signal outcome.
Samples were analysed in the same analytical run. (A) shows the variation in claudin-2 (Cldn-2) ISH signal by (eye) & by (auto) computer algorithm on sequential days (from 1 to 4) after slide pre-treatment step. Minimal variability is observed over the time period.
Fig 7.
Investigating the potential variability of biomarker signal for adjacent and non-adjacent sequential biopsy sections.
One micro-tomed 4 μM tissue section was generated (per datapoint) and ISH stained for claudin-2 mRNA distribution. Variability in the claudin-2 signal was investigated in (A) adjacent sections (3–6 novel sections per tissue) generated from a pancolitic (completely) active subject, and active & non-active UC material from two further subjects and in (B) the mean of two (adjacent) sections taken from four non-adjacent (≥ 24 μm distance apart).sites within a tissue biopsy, generated from 3 subjects who had non, low & moderate activity. In total, nine biopsies were analysed, and signal variability appears to be minimal within a specific biopsy.
Fig 8.
Representative grades of ISH claudin-2 staining in UC colon biopsies as disease severity increases.
The level of disease severity was determined, by a trained pathologist, as (A) 0, none (B) 1, low, (C) 2, slight, (D) 3, moderate, (E) 4, marked and (F) 5, severe. The scale bars represent 50 μm.
Fig 9.
Heat-map (with sample frequency) showing the association between the Claudin-2 ISH score and a microscopical histopathological activity index.
ISH scores and the activity index evaluations by the scoring pathologist were measured in the same biopsy sections from in 47 (affected and unaffected) UC biopsy sections used in the validation of the assay.
Fig 10.
Heat-map (with sample frequency) showing the association between the Claudin-2 ISH score and the Geboes score.
ISH and Geboes scores were measured in adjacent biopsy sections from in 132 (affected and unaffected) UC biopsies.