In this report, we have addressed two questions concerning immunological memory: the way in which naive and memory T cells recirculate through the body, and the intrinsic rate of division within the naive and memory populations. We identified naive and memory T cells in sheep by their cell surface phenotype and their ability to respond to recall antigen. Memory T cells were CD2hi, CD58hi, CD44hi, CD11ahi, and CD45R-, as pertains in man. T cells that crossed from blood to the tissues of the hind leg and accumulated in the popliteal afferent lymph were all of memory phenotype. Conversely, T cells in efferent lymph, 90% of which entered the lymph node (LN) via high endothelial venules (HEV), were mostly of the naive phenotype (CD2lo, CD58lo, CD44lo, CD11alo, and CD45R+). The marked enrichment of these two phenotypes in different recirculatory compartments indicated that memory T cells selectively traffic from blood to peripheral tissues to LN (via afferent lymph), whereas naive T cells selectively traffic from blood to LN (via HEV). We argue that the differential use of these two recirculation pathways probably optimizes lymphocyte interactions with antigen. The nonrandom distribution of T cell subsets in various recirculatory compartments may be related to the relative proportion of memory cells in each subset. In particular, gamma/delta T cells in blood were almost exclusively of memory phenotype, and accumulated preferentially in afferent, but not in efferent, lymph. Finally, using the bromo-deoxyuridine labeling technique, we found that at least a sizeable proportion of memory T cells, whether in blood or afferent lymph, were a dividing population of cells, whereas naive T cells were a nondividing population. This result supports an alternative model of lymphocyte memory that assumes that maintenance of memory requires persistent antigenic stimulation.

Abstract. CD44 is a ubiquitous surface molecule that exists as a number of isoforms, generated by alternative splicing of 10 "variant" exons. Little is known about the expression and function of the variant isoforms, except that certain isoforms may play a role in cancer metastasis. We produced mAbs against CD44 variant regions encoded by exons 4v, 6v, and 9v, by immunizing mice with a fusion protein spanning variant exons 3v to 10v. A comprehensive analysis of human tissues revealed that CD44 variant isoforms were expressed widely throughout the body, principally by epithelial cells. However there was differential expression of CD44 variant exons by different epithelia. Most epithelia expressed exon 9v, but much fewer expressed 6v or 4v. The regions of epithelia that expressed the highest levels of the variant isoforms were the generative cells, particularly the basal cells of stratified squamous epithelium, and of glandular epithelium. CD44 variant isoforms were also expressed differentially by leukocytes, with CD44-9v expressed at very low levels and CD44-6v and 4v virtually absent. However, CIM4-9v and CD44-6v were the main variants that were transiently upregulated on T cells after mitogenic stimulation and on myelomonocytic cell lines by TNFot and IFN~ treatment. Some epithelial cell lines could preferentially upregulate CIM4-6v upon IFN3, incubation. These results show that CD44 variant isoforms are expressed much more widely than first appreciated, and that expression of the variant isoforms on some cell types can be modulated by particular cytokines.C D44 is a widely expressed cell surface glycoprotein that serves as an adhesion molecule in cell-substrate and cell-cell interactions, including lymphocyte homing, hemopoiesis, cell migration, and metastasis (for reviews see Haynes et al., 1991; Underhill, 1992; Gtinthert, 1993;Lesley et al., 1993). CD44 also has other functions that relate to lymphocyte activation (Haynes et al., 1989) and the binding of certain cytokines to endothelium (Tanaka et al., 1993). CD44 is a proteoglycan with an NH2-terminal region that is structurally related to several hyaluronate binding proteins (Underhill, 1992). CD44 is known to bind hyaluronate and collagen (Carter and Wayner, 1988;Aruffo et al., 1990;Culty et al., 1990;Miyake et al., 1990) and a chondroitin sulfated form of CD44 binds fibronectin (Jalkanen and Jalkanen, 1992). Additional ligands for CD44 may well exist. The numerous functions and molecular interactions of CD44 probably relate to its complex structure. In addition to the "standard" 85-95-kD form (CD44s), severalThe Basel Institute for Immunology was founded and is supported by E Hoffman-La Roche Ltd., Basel, Switzerland.Address all correspondence to Dr. Charles R. Mackay, Leukosite Inc., 800 Huntington Ave., Boston, MA 02115, or Dr. Ursula Giinthert, The Basel Institute for Immunology, Grenzaeherstrasse 487, CH-4005 Basel, Switzerland.larger "variant" isoforms exist (CD44v) t that are generated by alternative splicing of at least 10 exons (Screaton et al....

The protein methyltransferase Set7/9 was recently shown to regulate p53 activity in cancer cells. However, the impact of Set7/9 on p53 function in vivo is unclear. To explore these issues, we created a null allele of Set7/9 in mice. Cells from Set7/9 mutant mice fail to methylate p53 K369, are unable to induce p53 downstream targets upon DNA damage, and are predisposed to oncogenic transformation. Importantly, we find that methylation of p53 by Set7/9 is required for the binding of the acetyltransferase Tip60 to p53 and for the subsequent acetylation of p53. We provide the first genetic evidence demonstrating that lysine methylation of p53 by Set7/9 is important for p53 activation in vivo and suggest a mechanistic link between methylation and acetylation of p53 through Tip60.

A proportion of T cells recirculate in a tissue-selective manner. Recent studies which showed that the skin-tropic subset of T cells was of memory/activated type, led us to examine whether the preferential homing of T cells to the gut also involved memory T cells, and if so whether these memory T cells were phenotypically distinct from other memory T cells. Lymphocytes migrating through the gut and the skin of sheep was collected by cannulating the lymphatic ducts draining these tissues. Both naive and memory T cells were found to recirculate through the gut, although only memory T cells migrated through the skin. However, when T cells from the gut were labeled with fluorescein isothiocyanate and assessed for their migration back to the gut, it was the memory population which showed a tropism for the gut. Gut-tropic memory T cells migrated poorly through the skin, indicating that these cells were distinct from skin-tropic memory T cells. This was confirmed by phenotypic analysis. Gut memory T cells expressed very low levels of the alpha 6 and beta 1 integrins, in contrast to skin memory T cells which expressed high levels. There was no evidence for heterogeneity within the naive T cell population, which migrated preferentially to lymph nodes. This migration pattern could be explained in part by the high expression of the L-selectin (lymph node homing receptor, LAM-1) on naive T cells, in contrast to memory T cells from gut or skin which were mostly L-selectin negative. These results in sheep indicate that subsets of alpha/beta memory T cells show tissue-selective migration patterns, which probably develop in a particular environment following encounter with antigen.

Antigen challenge has profound effects on a regional lymph node (LN); it leads to an increase in blood flow to the node, and a marked increase in lymphocyte output through the efferent lymphatics. We used the isolated LN model developed in the sheep to see if antigen challenge in a LN resembled inflammation in peripheral tissues. Following stimulation with an antigen (purified protein derivative of tuberculin), lymphocyte output from the LN showed the typical periods of "lymphocyte shutdown" and "recruitment". The shutdown phase, when cell numbers in efferent lymph dropped by approximately 80%, affected almost exclusively the naive-type (adhesionlo, L-selectin+) T cell population. The large increase in T cell traffic through the node during the recruitment phase was mostly due to CD4+ memory-type T cells and, moreover, the majority of these T cells were L-selectin-, indicating that these cells were crossing from the blood by a molecular mechanism other than L-selectin interaction with its ligand, the "lymph node vascular addressin" (MECA-79). Examination of LN high endothelial venules revealed the presence of vascular cell adhesion molecule-1 (VCAM-1), an endothelial adhesion molecule which has been reported to bind preferentially memory-type T cells in inflammatory lesions. Within the skin, antigen challenge also induced the rapid expression of VCAM-1 on vascular endothelium. It was purely memory-type T cells (beta 1+, L-selectin+/-) that collected in lymph draining from this tissue. However within chronically inflamed skin, the MECA-79 determinant appeared on vascular endothelium, and a small proportion of T cells draining from chronically inflamed skin were of naive-type. The present results illustrate that there are similarities in the cellular and molecular events that characterize antigen stimulation of a LN and inflammation in a peripheral tissue.