From hyperplasia to T cell lymphoma

Enteropathy-type intestinal T cell lymphomas (EITCL) are a recognised complication of coeliac disease (CD).¹ A recent survey confirmed that non-Hodgkin lymphomas, although rare, are the main cause of mortality in CD.² The mechanisms favouring the development of EITCL in CD patients but not in other chronic inflammatory bowel diseases remain elusive, but mounting evidence points to a profound disturbance in intraepithelial lymphocyte (IEL) homeostasis, leading to the emergence of lymphoid malignancies. A link between IELs and EITCL was first advocated in 1988 by Spencer et al who observed that most EITCL expressed the CD103 IEL marker.³ Two complementary observations suggested that EITCL derive from a reactive T cell population present in the intestine of CD patients: thus the same T cell clonal rearrangement was detected by Murray et al in EITCL and in the adjacent non-tumoral flat mucosa,⁴ and by Ashton-Key et al in non-lymphomatous ulcers of ulcerative jejunitis and in lymphomas, which later developed in these patients.⁵ Recent work in refractory sprue (RS) provided a missing link between IELs and lymphomas in CD.

RS is a coeliac-like enteropathy, primary or secondary resistant to a strict gluten free diet (GFD). Several conditions underlie villous atrophy resistant to GFD⁶ (Cellier et al, in preparation) but the majority of RS complicate CD and are associated with massive expansion of IELs with normal cytology but clonal T cell receptor γ (TCRγ) rearrangements and abnormal phenotype.^7,8 The malignant nature of IELs in RS was demonstrated by the frequent association of RS with ulcerative jejunitis (30%), and the outset of EITCL sharing the same clonality and phenotype after several months or years in approximately 20% of cases.^7,9,10 RS can thus be regarded as a “cryptic or low grade T cell lymphoma” derived from IELs, and draws a link between IEL hyperplasia, characteristic of CD, and EITCL. In some CD patients however, EITCL develop directly without this first intermediary step.⁶

The report of seven new patients by Farstad and colleagues¹¹ in this issue of Gut, four with RS without overt high grade lymphomas and three with EITCL associated with TCRγ clonal rearrangement (3/3) and phenotypically abnormal IELs (1/3) away from the tumour, concurs with previous work in RS and EITCL, and highlights two novel findings [see page 372]. One concerns the phenotype of IELs. Cellier et al initially reported that IELs in RS contained intracellular CD3ɛ+ but did not express surface CD3ɛ, CD8, or TCR.⁸ Herein, in 2/4 cases of RS, IELs expressed a TCRβ chain. These results are in keeping with previous reports,^6,9 and our unpublished results, showing TCRβ chain expression in 2/20 patients. Interestingly, flow cytometry in one patient studied herein as in our two cases, showed that TCRβ chains remain exclusively intracellular. Farstad et al also suggest that the lack of CD8 is not a constant feature of abnormal IELs in RS and EITCL.¹¹ Using flow cytometry, we detected weak CD8 expression in a fraction of abnormal IELs in 30% of RS patients but only one and two of 20 cases were positive for CD8 by immunohistochemistry on fixed and frozen tissue sections, respectively⁷ (unpublished data). Daum et al in a recent report observed one case of CD8+ (CD56+) EITCL with CD8+ clonal IELs away from the tumour but concluded that clonality in EITCL and RS is generally associated with loss of CD8 and/or βF1 expression.⁶ Taken together, these data indicate that the phenotype of IELs in RS is very similar to that reported in EITCL, with the majority of cases being CD3+ βF1-CD8-CD4−, but with some cases being CD8+ or βF1+; rare cases are positive for both markers.^4,6 These results emphasise the need to combine phenotype and molecular biology studies to investigate patients suspected of having RS or EITCL. They also support the idea that RS and EITCL share a common origin from IELs deprived of surface CD3-TCR complexes by a mechanism to be deciphered.

A second issue raised by Farstad and colleagues¹¹ concerns the predictive value of CD30 expression. The outcome of RS is variable and not easily predictable. A small number of patients respond to therapy, mainly corticoids and parenteral nutrition, and can be maintained symptom free for years on a GFD in spite of persistent monoclonal IELs, as illustrated in patient Nos 1 and 2 described by Farstad et al. In contrast, some patients develop high grade lymphomas while in others abnormal IELs do not form tumours but disseminate to lamina propria, blood, and eventually to other organs. Furthermore, some patients who do not develop lymphomas, rapidly die from untractable malabsorption. In the two latter cases, classical chemotherapy is ineffective or even deleterious because the abnormal lymphocytes divide too slowly to be efficiently destroyed by drugs interfering with cell proliferation. Discovery of prognosis markers would thus be useful to adjust therapy.

Farstad et al suggest that CD30 allows the early detection of overt lymphoma.¹¹ Firstly, they detected some CD30+ blast-like IELs and lamina propria cells away from the tumours in their three cases of EITCL. Secondly, in one patient with RS without overt lymphoma, the rapidly fatal outcome was associated with the presence of 25% CD30+ blast-like IELs. No CD30+ IELs were however observed in another RS patient with rapid severe outcome, whereas a third patient with some CD30+ IELs was improved by parenteral nutrition and a GFD. Finally, the patient with many CD30+ IELs did not benefit from one attempted cure of CHOP.¹¹ These observations illustrate the difficulties in predicting outcome and in propounding an appropriate treatment in RS patients. Apart from CD30, other immunohistochemical markers may help to detect transformation from low to high grade proliferation, such as proliferation markers or p53 detected by Murray et al on small lymphocytes in the bowel adjacent to EITCL.⁴ Nevertheless, as in patient No 3 reported by Farstad et al, none of these markers may be useful in predicting untractable malabsorption in the absence of overt lymphoma. Functional analysis of abnormal IELs in RS may identify criteria predictive of their aggressiveness for the mucosa or new targets for therapy, a pressing need given the lack of current efficient treatment for severely sick RS patients⁷ and the poor prognosis of EITCL.¹ Insight into the mechanism(s) disturbing IEL homeostasis in CD may help to decipher the links between inflammation and lymphoid malignancies and to design treatments able to prevent or cure these rare but most severe complications of CD.