Article Text

Differentiating ulcerative colitis from Crohn’s disease: false dawn for CD44
  1. Department of Histopathology,
  2. UMDS St Thomas’s Campus,
  3. Lambeth Palace Road,
  4. London SE1 7EH, UK

Statistics from

See article on page 375

It is unusual for a manuscript with negative conclusions to be highlighted in a commentary. However, the study by Reinischet al (see page 375) is the latest of a number of reports to comment on the ability of an immunohistochemical stain for CD44 variant expression to discriminate between ulcerative colitis and Crohn’s disease. It is a thorough study which puts earlier work into perspective, and as such is of considerable value.

The subject of Reinisch et al’s study is epithelial expression of CD44, a receptor for matrix components including hyaluronic acid, which exists in multiple isoforms generated by exon splicing.1 2 The tissue distribution of CD44 isoforms in normal and malignant tissues was first described in 1994 by Fox et al who observed weak expression of CD44v3 and CD44v6 isoforms on crypt epithelium in the gastrointestinal tract.3 Subsequently, Rosenberg et al,4 using archival paraffin wax embedded specimens and a microwave system to reveal the epitopes recognised by the antibodies, observed a noticeable increase in the expression of CD44v3 and CD44v6 on crypt epithelium in ulcerative colitis compared with normal colon and colonic Crohn’s disease. They proposed that this difference could form the basis of a diagnostic test to discriminate between these diseases. Such a test would clearly be of enormous value as the histopathologists’ contribution to the disease classification, which continues to rely on a panel of classic histopathological features, sometimes fails to reach a satisfactory conclusion. Despite the resemblance between ulcerative colitis and Crohn’s disease at many levels, they are presumed to be different diseases in terms of their aetiology and pathogenesis. The clinical management required is different, and accurate diagnosis is of considerable importance. Rosenberg et al’s study was confirmed by two other groups5 6 but a third study7failed to reproduce these results. The two groups which had agreed had used frozen sections, whereas the group which had failed to discriminate between ulcerative colitis and Crohn’s disease had used paraffin wax sections. These studies all used the same antibodies.

The study by Reinisch et al concentrates on the expression of CD44v6. They first produced a high affinity antibody to CD44v6, which they studied alongside the 2F10 anti-CD44v6 monoclonal used in previous studies. They then stained a panel of paraffin wax specimens using a microwave technique to reveal the antigens, and assessed the reactivity in terms of the intensity of expression and the number of positive cells. Like Rosenberg et al, they observed that CD44v6 expression was higher in ulcerative colitis than in Crohn’s disease in both the intensity of the staining and the number of positive cells. Both antibodies gave the same pattern of staining though the higher affinity antibody stained more cells, and stained more intensely. These authors state, however, that the staining would not constitute a diagnostic test because of deficiencies in specificity and sensitivity. Reinischet al therefore confirm both the original observation of an increase in CD44v6 staining in ulcerative colitis and also the doubts of others about the diagnostic value of the observation.

The very fact that there was debate on the usefulness of CD44v6 staining stymied its prospects as a diagnostic test. In modern pathology laboratories, diagnostic immunohistochemical techniques are regularly carried out using paraffin wax sections. Different laboratories use different antigen retrieval systems to reveal epitopes masked by the processes of fixation and embedding. Alternative techniques are trypsinisation, microwave or pressure cooking. Though different antigens may be optimally revealed by different processes, selection of technique used is often based on personal experience. Variability in staining can depend on factors such as the local formalin (buffered or not) and the size of the piece of tissue (resection or biopsy). To achieve success as a diagnostic test, the antibody has to overcome these variables and bind to its target reliably and reproducibly. The failure of Papadogiannakiset al to observe any difference in CD44v6 expression between ulcerative colitis and Crohn’s disease could have been due to any of these variables.7

From the pathologist’s point of view, the most difficult immunohistochemical reactivity to assess must be that which depends on degree of positivity. When Rosenberg et al’s paper was first published,4 Ilyas and colleagues8 pointed out to the authors that CD44v6 did in fact stain normal colonic crypt epithelium,3 8 albeit weakly, a point which was dismissed by Rosenberg as possibly semantic.9 This weak staining may well have been easy for those who see many such specimens stained in the same way to ignore, but to a diagnostic pathologist, it would be important to understand that weakly positive staining might be a negative result. CD44v6 staining would have to be examined alongside specimens with high intensity and weak staining, and the shade of the positivity considered in this context. Failure of a specimen to stain using an immunohistochemical technique, with all of the variables which can affect staining mentioned earlier, is not interpretable. In addition, epithelial cells frequently show weak non-specific background staining; a fact which could add to the confusion.

In conclusion, although scientifically the increase in CD44v6 expression in ulcerative colitis appears to be a sound and interesting observation, the fact that it relies on a differential expression over a variable background makes it useless for routine diagnosis. The differential diagnosis of Crohn’s disease and ulcerative colitis must remain a multidepartmental collaborative effort, at least for the time being.

See article on page 375


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles