Immunophenotyping of Sézary cells

The article by Bernengo and colleagues from Turin offers the possibility of a reliable immunophenotypic method for identifying patients with cutaneous T-cell lymphoma (CTCL) who have haematological involvement. This could provide a valuable addition to the methods that are currently used to diagnose Sézary syndrome (SS) and could also be used to provide prognostic information in patients with SS and possible mycosis fungoides (MF). The test relies upon loss of the CD26 antigen, which is an accessory cell surface molecule involved in T-cell activation and proliferation. The great majority of circulating CD4+ cells are also CD26+. FACS analysis of patients with either SS or MF with haematological involvement reveals that over 30% of circulating CD4+ cells are CD29–, whereas healthy donors or erythrodermic patients with inflammatory skin disease (EISD) fall below this figure.

The key message of this article is contained in Fig. 2, which demonstrates that none of 21 patients with SS and only one of 14 MF patients with haematological involvement had less than 30% of CD4+ CD26– cells in their peripheral blood.

Although these figures are small and need to be reproduced, the CD26 marker certainly seems superior to other immunophenotypic methods that have been used to distinguish patients with SS from patients with inflammatory erythrodermic diseases. In particular, loss of CD7 antigen has been proposed as a marker for circulating Sézary cells.¹ In fact, there is considerable overlap in the percentage of CD4+ CD7– cells in patients with SS and EISD. Furthermore, other workers have demonstrated that the clonal population in SS can be detected in both CD7+ and the CD7– subsets of the same patient.² Bernengo specifically addresses this question in her paper and confirms that loss of CD26 is superior to loss of CD7 as a marker for Sézary syndrome.

How can this new marker be incorporated into clinical practise? Currently there is no consensus as to the diagnostic criteria for SS. The EORTC proposed fairly rigorous haematological criteria that included a CD4 : CD8 ratio greater than 10, a Sézary count greater than 1000 mm⁻³ and a clonal population, as demonstrated by TCR gene analysis.³ If these criteria are fulfilled then there is little chance of misdiagnosing Sézary syndrome. However, it also defines a group with a high tumour burden and a poor survival prognosis of only 2 years.⁴ Other workers have proposed criteria that would identify SS patients at an earlier stage in the disease process. Thus, Russell-Jones and Whittaker suggested erythroderma, compatible skin histology, a clonal T-cell population in the peripheral blood and more than 5% circulating Sézary cells.⁵ 5% was chosen because it corresponds to the original B1 definition of blood involvement in MF⁶ and has been shown to carry prognostic significance in erythrodermic CTCL.⁷ However, the consensus document produced by the International Society for Cutaneous Lymphoma (ISCL) rejected this definition because of uncertainty about the specificity of TCR gene analysis and their preference for a higher tumour burden.⁸

Bernengo’s data does not resolve this dilemma because all the SS patients in her study had Sézary counts greater than 1000 mm⁻³. One cannot therefore draw any conclusion as to the diagnostic value of immunophenotyping in patients with a low tumour burden. One might have gained some insight from the cohort of 14 patients with MF and haematological involvement but again, Bernengo uses a high tumour burden to define her patients with B1 (greater than 10% circulating leukocytes and greater than 1000 mm⁻³).

In consequence, the value of CD26 staining has only been demonstrated in situations where SS is already easy to diagnose by other methods. Even so, it could prove valuable. The majority of patients with SS have the small-cell variant and these cells can be difficult to distinguish on routine blood smears from reactive lymphocytes. Indeed, many haematology departments do not even offer Sézary cell counts because of uncertainty about the reproducibility of morphological examination alone.

TCR gene analysis of peripheral blood goes some way to resolving this dilemma. Thus, in patients with MF we have shown that PCR positivity in blood is an independent prognostic marker in patients with skin stages T1–T3.⁹

We have also shown that TCR gene analysis will distinguish patients with Sézary syndrome from patients with reactive forms of erythroderma and circulating Sézary cells.⁵ However, gene analysis does not provide an accurate measure of high tumour burden as PCR-based techniques will detect a clonal population representing as low as 1% of peripheral blood lymphocytes, while Southern blot analysis has a detection sensitivity of around 5%.⁹ Bernengo’s paper offers the possibility of a more reproducible method of staging patients with higher levels of haematological involvement. In particular, patients with erythrodermic CTCL have recently been divided into five stages, designated H0 to H4 according to the haematological findings.¹⁰ H0 patients have no haematological involvement, while H1 patients are PCR-positive but exhibit less than 5% circulating Sézary cells. Essentially, H0 and H1 fulfil the diagnosis of erythrodermic mycosis fungoides. H2 patients are Southern blot-positive or PCR-positive with more than 5% circulating Sézary cells. H3 patients have more than 1000 Sézary cells mm⁻³ and H4 greater than 10 000 cells mm⁻³. These categories are known to carry prognostic significance in erythrodermic CTCL.^7,10 Employing this staging system in combination with the immunophenotypic method proposed by Bernengo could improve the reproducibility of the system and stage H2–H4 patients more accurately.

Further studies are needed to establish whether the ‘H’ system could be extended to all patients with MF. If so, it could provide a unifying system for accurately staging blood involvement in CTCL and avoid semantic arguments over terminology.