Intestinal intraepithelial lymphocytes: the plot thickens

O ver the last few years, the characterization of T lyraphocytes within the intestinal epithelium (ilEL) has revealed phenotypic and functional heterogeneity that rivals only that observed within the thymus. Of note in this context is that thymic and intestinal epithdia share the same embryologic origin in that they are both derived from endoderm. Several lines of evidence have led to the conclusion that T lymphopoiesis occurs within the intestinal epithelium. This observation raises several fundamental questions some of which have been recently addressed in this and other journals. Specifically, are all iIEL of gut origin, and do the phenotypic and functional peculiarities of the various ilEL subsets reveal novel pathways of T cell development? Origin of ilEL. The phenotypic and functional heterogeneity of iIEL has been recently reviewed (1, 2). Briefly, as in the thymus, all ilEL are T cells expressing either TCK"y/g or TCR-odfl, although the proportions of these two lineages vary among species. In the mouse, a high proportion of iIEL express TCR-~//8, the majority of which coexpress the c, chain of CDS. Those iIEL expressing TCK-cdfl can be further distinguished based on the differential expression of the accessory molecules CD4, CD8oe, and CDSfl. Two of the TCR-odfl + ilEL subsets are phenotypically identical to mature thymus-derived T cells: they express CD4, or the heterodimeric form of CD8; they are depleted of cells expressing autoreactive TCK; and they have functional antigen receptor complexes in that they can be stimulated by nominal antigen-, superantigen-, and TCK-specific mAbs (3, 4). Two additional TCK-od/3 + ilEL subsets are peculiar to the intestinal epithelium: one expresses only the CD8 c~ chain, while the other coexpresses this chain and CD4. Although all TCK-c~/B ~ ilEL subsets express a level of TCR/CD3 similar to that of mature thymus-derived T cells (2), the CD4+8c~+Band CD8o~+/5 ilEL subsets are unresponsive to the above set of stimuli (3, 4), and moreover contain high frequencies of auto-reactive cells (5). In this context the functional status of these four ilEL subsets correlates with membrane expression of CD28 (6). However, all ilEL subsets respond to concanavalin A (7). There is consensus that all TCK-3//r + ilEL develop in situ. They are found in both congenitally athymic nude mice (8), as well as in athymic radiation chimeras (7). The controversy surrounds the origin of the various subsets of ilEL expressing TCR-~//5. Specifically, Guy-Grand et al. have interpreted the differential repe~oire expressed by CD8~+/3 and CD8ol +/3 + ilEL as reflecting the thymic origin of the latter, and the gut origin of the former. Furthermore, in order to explain the high frequency of CD8ol + ~ilEL expressing autoreactive TCRs, these investigators invoke distinct selection pressures applied to gut-derived T cells (5, 8). This brings us to the first new piece of information published recently by the same group (9). In this study, adult animals lacking the recombination activating gene RAG-2 were either thymectomized or not, sublethally irradiated (400 rad), and then reconstituted with bone marrow from H-2 matched (C57B1/6) nude donors. The observation is that euthymic recipients contain a normal number of ilEL phenotypically similar to those observed in 6-8-wk-old conventional animals. In contrast, the total number of ilEL in athymic recipients was reduced roughly fourfold, and the authors highlight that these recipients are devoid of TCR-~/[3 + ilEL expressing either CD4, or CD8B. In support of their original premise, the TCK-c~/~ + ilEL found in these recipients expressed CD8c~ exclusively. They conclude that these results unequivocally demonstrate the thymic origin of TCR-c*/[3 + ilEL expressing CD4 or CD8oefl, and the gut origin of TCR-odfl + iIEL expressing only the c, chain of CD8. Further support for this tenet is derived from the observation that CD4 + and CDSoefl + iIEL subsets are found in the gut of euthymic nonirradiated RAGdeficient animals injected with lymph node T cells from conventional animals (9). The concern with the authors' conclusion rests with the extent of reconstitution of the iIEL compartment of athymic RAG 2-deficient recipients. Whereas the reduction in total ilEL was only fourfold, the proportion of TCR-3,/g + iIEL was 20% of that observed in euthymic recipients. More strikingly, only 1-2% of iIEL in athymic recipients expressed TCRodB. Thus, one could argue that the development of a large proportion of TCR-~//g + ilEL and of virtually aU TCRo~/f3 + ilEL, is thymus dependent. Consistent with this latter hypothesis is the observation that nude mice are devoid of TCR-ol/B + iIEL, and contain a reduced number of TCR-~/8 + ilEL (8). Inconsistent with this hypothesis is the observation that all TCR-cff/3 + and TCR-',//g + iIEL subsets are generated in normal numbers in athymic radiation chimeras, and are derived from donor stem ceUs (7, 10). In addition, ectopic f~tal intestine grafts reconstitute the peripheral T cell pool of athymic radiation chimeras as ef~dently as do ectopic fetal thymic grafts (11).