Two different functions for CD44 proteins in human myelopoiesis.

CD44 is important during myelopoiesis, although the contributions of variant CD44 proteins are unclear. We show here that in human long-term bone marrow culture antibodies recognizing a CD44 NH2-terminal epitope (mab 25-32) or a CD44v6 epitope (mab VFF18) inhibit myelopoiesis. However, mab 25-32 but not mab VFF18 affects myeloid colony formation. These data suggest that an early precursor cell compartment is the target for the 25-32 antibody, whereas the mab VFF18 targets later stages in myelopoiesis. Since the bulk of hemopoietic precursor cells are negative for the v6 epitope and only a minor subset of myeloid cells express the v6 epitope, we have used several human myeloid progenitor cell lines to unravel the function of different CD44 proteins. These cell lines produce variant CD44 proteins, predominantly a new variant CD44v4-v10, when stimulated towards myeloid differentiation. Features that can be acquired by the expression of CD44v4-v10 are an increased hyaluronate (HA) and a de novo chondroitin sulphate A (CS-A) binding. Although, the expression of CD44v4-v10 per se is necessary for HA and CS-A binding, the protein backbone seems to require appropriate glycosylation. HA binding results in CD44-mediated cellular self-aggregation and adhesion to the stromal cell line MS-5. In summary, our data suggest that different CD44 proteins are important for at least two different steps in myelopoiesis.

Expression of CD44 variant isoforms in peripheral blood leukocytes in malignant lymphoma and leukemia: inverse correlation between expression and tumor progression.

In a variety of human tumors, including high grade Non-Hodgkin's lymphoma (hgNHL), a linkage between expression of CD44 variant isoforms (CD44v) and tumor progression has been described. In search of an easily accessible diagnostic parameter, expression of CD44 standard (CD44s) and CD44 variant isoforms (exons v5, v6, v7 and v10) in peripheral blood lymphocytes (PBLs) of patients with hematological malignancies was evaluated by fluorescence activated cell scanning. The analysis of 30 blood samples of healthy donors and patients with non-malignant diseases and of 183 blood samples of patients with malignant hematological disorders revealed that only in patients with malignant disorders did a measurable proportion of PBLs express CD44 variant isoforms, mostly exons v5, v6, v7 and, less frequently, exon v10. Elevated levels of CD44v expression were noted in PBLs of patients with acute and chronic myeloid leukemia (AML: 16%, CML: 25%), Hodgkin's disease (HD: 17%), multiple myeloma (MM: 22%), polycythemia vera (PV: 33%), acute lymphoid leukemia (ALL: 23%) and, most frequently, in PBLs of patients with non-Hodgkin's lymphoma (NHL:54%). CD44v expression was not restricted to the malignant phenotype, but instead was also noted in T cells, B cells and monocytes, preferentially in a subpopulation of large cells. Furthermore, expression of CD44v in PBLs was not linked to the histological grading or clinical staging. There was, however, an inverse correlation with tumor progression, whereas response to therapy was frequently accompanied by upregulation of CD44v. Thus, expression of CD44v in the PBLs of patients with NHL mainly reflected immune responsiveness. Since NHL manifests itself primarily in lymphoid organs, its progression is difficult to follow. Monitoring of CD44v in PBLs could be used as an additional and convenient parameter for surveying the course of disease.

Autoreactive antibodies in thymus and spleen of neonatal and young adult BALB/c mice: influence of prenatal tolerization.

The repertoire of autoantibody-producing B cells was evaluated in a collection of spleen- and thymus-derived hybridomas from 6- and 28-day-old BALB/c mice, which were untreated or prenatally tolerized with trinitrobenzenesulphonic acid (TNBS). MoAb were tested for their reactivity with TNP-BSA and the autoantigens thyroglobulin (TG), myoglobin (MG), actin (AC), cytochrome C (CY), collagen (CO), transferrin (TF), single-stranded DNA (ssDNA), double-stranded DNA (dsDNA), and bromelain-treated mouse red blood cells (BrMRBC). More than 10% of spleen cell (SC)-derived MoAb from 6- and 28-day-old control mice did bind to AC, ssDNA, dsDNA, MY, and TG, the frequency of MoAb reacting with MY, TG, and BrMRBC increasing with age. Thymus cell (TC)-derived hybridomas contained autoreactive clones too, but only few of them produced multireactive MoAb. MoAb from prenatally TNBS-treated mice were more frequently autoreactive than MoAb from control mice, especially if derived from TC hybridomas. The most remarkable difference in the reactivity pattern as compared with MoAb from untreated mice consisted of a significant increase in the frequency of TG-, My-, ssDNA- and above all dsDNA-reactive MoAb, all TC-derived multireactive MoAb binding to dsDNA. The differences in autoreactivity between MoAb from prenatally untreated and TNBS-treated mice as well as age- and organ-related variations support the interpretation that part of the repertoire of naturally activated B cells is not random but is influenced by and responding to the available panel of self antigens.

Evidence for regulation of naturally activated autoreactive B cells.

A large fraction of naturally activated B cells in the neonate displays degenerate specificity, including reactivity with autoantigens. Transgenic mouse models of autoreactive B cells are mainly concerned with monospecific B cells of high avidity, and the fate of naturally activated autoreactive B cells is still a matter of debate. To pursue this question further, we chose an IgM autoantibody with a recurrent idiotype (Id), i.e. Sp6, because transgenic mice expressing this IgM also were available. In a first approach monoclonal antibodies (mAb) derived from untreated, antigenically stimulated and transgenic mice were used to test whether there were indications for deletion or for Id regulation of naturally activated autoreactive B cells. Over 90% of thymus and spleen cell derived hybridomas from 6-day-old Sp6-transgenic mice were trinitrophenyl (TNP) reactive, carried the Sp6-Id and bound to a panel of self antigens, including mouse albumin. We failed to obtain B cell hybridomas from the thymus of 28-day-old Sp6-transgenic mice. Furthermore, we could not detect any mAb carrying an anti-Sp6 Id, but Sp6 did weakly bind to itself. About 25% of mAb derived from control mice displayed degenerate specificity, the majority of them also were TNP reactive. The Sp6 Id was found at a low frequency and a comparable number of mAb carried an anti-Sp6 Id. Prenatal manipulation at the antigen level (trinitrobenzenesulfonic acid treatment) led to a transient expansion of TNP- and autoreactive mAb. The number of mAb carrying the Sp6 Id was not increased, but mAb carrying an anti-Sp6 Id were observed at high frequency. Those mAb also displayed degenerate specificity. Since Sp6-transgenic mice were perfectly healthy, it is concluded that this particular autoreactive antibody of degenerate specificity cannot be harmful for the developing organism, which may possibly be due to its self-binding capacity. Furthermore, some process of down-regulation was indicated by the absence of B cells expressing the transgene in the thymus of young adult mice. Autoreactivity of untreated and prenatally antigen-treated mice was, in addition, regulated at the Id level. In particular, mAb recognizing the Id of Sp6 were significantly expanded in antigenically stimulated mice. The data were interpreted in the sense that autoreactive B cells appearing early during ontogeny were rather strictly controlled either by (functional) clonal deletion or by idiotypic connectivity.

Idiotypic profile of natural autoantibodies in newborn and young adult BALB/c mice.

Idiotypic profiles of autoreactive monoclonal antibodies (MoAb) were evaluated by their reactivity with a panel of alkaline phosphatase (AP)-coupled detector MoAb derived from the same fusions. Attention was given to the question of whether differences exist between MoAb derived from spleen cells (SC) or thymocytes (TC) and whether ID profiles would change during post-natal development. In the newborn, natural autoantibodies and MoAb which did not react with any one of eight autoantigens displayed different ID profiles, autoreactive MoAb being characterized by the expression of a restricted pattern of ID. During post-natal development, changes of ID expression were only observed with autoreactive MoAb. Many ID which were detected on MoAb derived from 6-day-old mice were not detected on SC-derived MoAb from young adults, while a few ID were significantly over-represented. Furthermore, especially with TC-derived MoAb, a clear linkage between certain idiotypes and autoantigen specificities could be demonstrated. Thus, in contrast to non-autoreactive MoAb, natural autoantibodies in the young adult were characterized by expressing only a selected number of ID at high frequency. Furthermore, the B-cell environment apparently played a role, since there were marked differences between ID profiles of TC- versus SC-derived MoAb. The data are interpreted in the sense that expansion and maturation of naturally activated autoreactive B cells are controlled rather than being random processes.