CD40.FasL inhibits human T cells: evidence for an auto-inhibitory loop-back mechanism.

A chimeric CD40.FasL (CD40-CD95L) protein was designed with the combined capacities to bind to two surface receptors on activated T cells, CD40 ligand (CD40L; CD154) and Fas receptor (CD95). CD40.FasL, once tethered to the cell surface via one of its ends, can transmit a signal via its other end. In principle, simultaneous triggering from both ends is possible, and thus there is the intriguing potential for 'auto-inhibition' if such dual triggering occurs on the same cell itself. Several lines of evidence support this mechanism: (i) CD40.FasL is cytotoxic to Fas receptor-positive cell lines of different cell lineages, (ii) CD40.FasL's function is potentiated when there is enforced expression of CD40L on target cells, (iii) CD40.FasL inhibition does not require intercellular contact, as demonstrated by soft agar clone formation and cell dilution analysis and (iv) introduction of exogenous CD40 into the system interferes with CD40.FasL inhibition. Taken together, these data are consistent with a 'loop-back' inhibitory mechanism within individual activated (CD40L and Fas receptor expressing) T cells causing suicide of these T cells. Significantly, this type of fusion protein provides a unique way to confine immunoinhibition to activated T cells.

CTLA-4.FasL inhibits allogeneic responses in vivo.

CTLA-4.Fas ligand (CTLA-4.FasL), a paradigmatic 'trans signal converter protein (TSCP)', can attach to APC (via CTLA-4 binding to B7) and direct intercellular inhibitory signals to responding T cells (via FasL binding to Fas receptor), converting an activating APC-to-T cell signal into an inhibitory one. Our previous studies established that CTLA-4.FasL inhibits human primary mixed lymphocyte reactions (MLR) and induces alloantigen-specific hyporesponsiveness ex vivo. The present study extends this to an in vivo context. Using splenocytes from MHC-mismatched C57BL/6 and Balb/c mice, we demonstrated that his(6)CTLA-4.FasL, effectively inhibits murine MLR. Moving in vivo, we demonstrated that subcutaneously administered his(6)CTLA-4.FasL modulates the in vivo response of infused allogeneic splenocytes. his(6)CTLA-4.FasL reduces the number of cells in each cell division, and increases the percentage of dead cells in each division. These findings are consistent with an antigen-induced cell death of the alloreactive cells, and bolsters recombinant TCSP promise as a therapeutic for transplantation diseases.

A rheostatic mechanism for T-cell inhibition based on elevation of activation thresholds.

The activation of discrete T-cell responses depends on the triggering of individualized threshold numbers of T-cell receptors (TCRs). The results of this study indicate that the lipocalin placental protein 14 (PP14), a T-cell inhibitor produced by cells of the reproductive and hematopoietic systems, mediates its anti-inflammatory activity by elevating the T-cell activation threshold, thereby rendering T cells less sensitive to stimulation. Significantly, the data demonstrate hierarchical sensitivity of selected cytokine responses to PP14-mediated inhibition, with the hierarchy reflecting their respective activation thresholds. These findings suggest a novel paradigm for immunoinhibition wherein negative regulators can finely tune, rather than inactivate, T-cell responses, and thereby skew the cytokine output of immunologic responses.

Induction of antitumor immunity via intratumoral tetra-costimulator protein transfer.

Our group recently described a novel two-step Fc(gamma1) fusion protein transfer method, which entails the docking of Fc(gamma1) fusion proteins onto cells precoated with chemically palmitated protein A (pal-prot A). In the present study, we have adapted this protein transfer method, originally used in an ex vivo context, for in situ tumor cell engineering, and in so doing, we have evaluated its utility for the induction of antitumor immunity via combinatorial costimulator protein transfer on to tumor cell surfaces. The feasibility of "painting" cells with preformed conjugates of a murine B7-1 costimulator derivative, B7-1.Fc(gamma1), and pal-prot A in a single step was first established ex vivo. Next, B7-1.Fc(gamma1):pal-prot A transfer was accomplished in vivo by directly injecting the preformed conjugates into highly aggressive L5178Y-R lymphomas grown intradermally in syngeneic mice. The presence of cell surface-associated B7-1 epitopes on cells of the injected tumors was documented by flow cytometric analysis of cells recovered subsequently from the injected tumors. B7-1.Fc(gamma1), along with Fc(gamma1) fusion protein derivatives of three additional costimulators (Fc(gamma1).4-1BBL, CD48.Fc(gamma1), and Fc(gamma1).CD40L) geared toward a variety of immune effectors, were together preconjugated with pal-prot A and injected directly into tumor beds. Significantly, this "tetra-costimulator" combination, delivered intratumorally, induced complete tumor regression in approximately 45% of treated mice, whereas control injections of pal-prot A alone had no therapeutic effect. Furthermore, there was evidence for systemic antitumor immunity in that tumor-specific CTLs were detected in spleens recovered from cured mice, and these mice were uniformly protected against tumor rechallenge at distant tumor sites. Hence, combinatorial costimulator transfer, coupled to intratumoral delivery, may have special advantages for the induction of antitumor immunity.

Properties of exogenously added GPI-anchored proteins following their incorporation into cells.

Isolated glycosylphosphatidylinositol (GPI)-anchored proteins, when added to cells in vitro, incorporate into their surface membranes and, once incorporated, exert their native functions. Virtually any protein of interest, if expressed as a GPI-reanchored derivative, can be modified to acquire this capacity. Such transfer of proteins directly to cells, termed "protein engineering" or "painting" constitutes an alternative to conventional gene transfer for manipulating cell surface composition that has many potential applications. Previous studies with incorporated GPI-anchored proteins have focused almost entirely on their extracellular functions. In this study, biotinylated human erythrocyte (E(hu)) decay accelerating factor, E(hu) acetylcholinesterase, and GPI-reanchored murine B7-1 and B7-2 were used as GPI-anchored reporters to characterize their plasma membrane organization and cell signalling properties following addition to Hela or Chinese hamster ovary cells. For each reporter, three types of cell-association were documented; (1) nonphysiological attachment and/or incomplete insertion, (2) uncomplexed membrane integration, and (3) organization into TX-100-resistant microdomains. Transit from the first two compartments into the third, i.e., microdomains, progressed slowly, continuing even after 24 to 36 h and was associated with the acquisition of cell signalling capacity. All four reporters, incorporated in two different detergents, behaved similarly. When organized in microdomains, caveolin and other GPI proteins co-isolated with the incorporated reporter. These results have implications for protein engineering of cells in general, and in particular, for cells such as modified tumor cell immunogens administered to patients for therapeutic purposes.

CTLA-4-Fas ligand functions as a trans signal converter protein in bridging antigen-presenting cells and T cells.

Co-stimulator blockade and trans inhibitory signaling, using agents such as CTLA-4-Ig and Fas ligand (FasL) respectively have been invoked as alternative strategies for suppressing pathogenic T cells. This study describes a novel hetero-bifunctional fusion protein, CTLA-4-FasL, designed to combine within a single protein both co-stimulator blocking and trans inhibitory signaling potentials. A chimeric expression cassette, in which the ectodomain coding sequences for CTLA-4 and FasL were linked in-frame, was used to produce a CTLA-4-FasL fusion protein. CTLA-4-FasL binding to both B7-1/B7-2-expressing Daudi B cells and Fas-expressing Jurkat T cells was documented by immunofluorescence and flow cytometry. The capacity of CTLA-4-FasL to induce apoptosis in Jurkat targets was markedly enhanced by the addition of Daudi and other B7-1/B7-2(+) B cell lines, which provided a membrane platform for the otherwise soluble CTLA-4-fusion protein. Moreover, in dual-chamber experiments, Daudi cells pre-coated with CTLA-4-FasL demonstrated Jurkat inhibitory activity that was cell-contact dependent. Significantly, when used to inhibit in vitro cellular proliferation of peripheral blood mononuclear cells, CTLA-4-FasL was approximately 1000-fold more potent than the extensively characterized CTLA-4-Ig fusion protein. Furthermore, the degree of inhibition induced by CTLA-4-FasL substantially surpassed that observed for CTLA-4-Ig and a soluble FasL when used in combination. CTLA-4-FasL represents the first of a novel class of fusion proteins, designated here as 'trans signal converter proteins', that combine trans signal masking and direct trans signaling functions.

alpha2-macroglobulin modulates the immunoregulatory function of the lipocalin placental protein 14.

Human placental protein 14 (PP14; also known as glycodelin and progesterone-associated endometrial protein) is an immunosuppressive protein of the lipocalin structural superfamily. Mechanisms regulating serum PP14's immunosuppressive activity remain to be elucidated. In the present study, an interaction between PP14 and a major serum protein carrier, alpha(2)-macroglobulin (alpha(2)M), was documented for the first time. Using native gel electrophoresis, we showed that PP14, as well as its alternative splice variant PP14.2, binds to both alpha(2)M and methylamine-activated (MA)-alpha(2)M. Cross-competition studies demonstrated that the variants compete for binding to alpha(2)M. PP14 bound to alpha(2)M and MA-alpha(2)M with K(d) values of 167+/-70 and 221+/-56 nM (means+/-S.D.) respectively, as determined by surface plasmon resonance. Significantly, the addition of alpha(2)M or MA-alpha(2)M to a T-cell proliferation assay strongly potentiated the inhibitory capacity of PP14. On the basis of these findings, alpha(2)M emerges as the first serum protein that can physically associate with, and thereby regulate, PP14. Moreover, this represents the first documented interaction between the protein carrier alpha(2)M and a lipocalin protein.

Glycosyl-phosphatidylinositol reanchoring unmasks distinct antigen-presenting pathways for CD1b and CD1c.

Human CD1 proteins present lipid and glycolipid Ags to T cells. Cellular trafficking patterns of CD1 proteins may determine the ability of differing isoforms of CD1 to acquire, bind, and present these Ags to T cells. To test this hypothesis, glycosyl-phosphatidylinositol (GPI)-modified variants of CD1b and CD1c were engineered by chimerization with a GPI modification signal sequence derived from decay-accelerating factor (DAF). GPI reanchoring was confirmed by demonstrating the phosphatidylinositol-specific phospholipase C sensitivity of the CD1b. DAF and CD1c. DAF fusion proteins expressed on transfectant cell surfaces. Using cytotoxicity and cytokine release assays as functional readouts, we demonstrated that CD1c. DAF is as efficient as native CD1c in presenting mycobacterial Ags to the human CD1c-restricted T cell line CD8-1. In contrast, CD1b. DAF, although also capable of presenting Ag (in this case to the CD1b-restricted T cell line LDN5), was less efficient than its native CD1b counterpart. The data support the idea that CD1c. DAF maintains the capacity to access CD1c Ag-loading compartment(s), whereas CD1b. DAF is diverted by its GPI anchor away from the optimal CD1b Ag-loading compartment(s). This constitutes the first GPI reanchoring of CD1 proteins and provides evidence that CD1b and CD1c have nonoverlapping Ag-presenting pathways, suggesting that these two Ag-presenting molecules may have distinct roles in lipid Ag presentation.

Hierarchical costimulator thresholds for distinct immune responses: application of a novel two-step Fc fusion protein transfer method.

Activation of T cells is dependent upon coordinate engagement of Ag and costimulator receptors on their surfaces. In the case of the Ag receptors (TCRs), activation thresholds have been defined, with the number of TCRs that must be triggered to stimulate cytokine secretion by individual activated T cells differing for the various cytokines. In the present study, we have determined whether comparable activation thresholds exist for the costimulator receptors on T cells. To facilitate this type of quantitative costimulator analysis, we developed a novel two-step protein transfer approach that permits delivery of graded amounts of proteins to APC surfaces. By adding a human B7-1. Fcgamma1 (Fc domain of human IgG1) fusion protein to cells precoated with palmitated protein A, fine titration of the B7-1 extracellular domain was achieved. The B7-1. Fcgamma1 reincorporated into cell membranes by this method retained costimulator function, as measured by an in vitro proliferation assay. The degree of proliferation was dependent on the surface density of B7-1. Fcgamma1. Significantly, the threshold B7-1. Fcgamma1 density required for cytokine production differed between IFN-gamma and IL-2 and mirrored the hierarchy (IFN-gamma < IL-2) described previously for the TCR activation threshold. Hence, this study invokes a novel protein transfer strategy to establish that the levels of surface costimulator on APCs can dictate both the magnitude and the quality of evoked T cell responses. The notion of costimulator receptor activation thresholds emerges.

The imprinted H19 gene is a marker of early recurrence in human bladder carcinoma.

AIMS: To investigate the expression of the imprinted oncofetal H19 gene in human bladder carcinoma and to examine the possibility of using it as a tumour marker, similar to other oncofetal gene products. METHODS: In situ hybridisation for H19 RNA was performed on 61 first biopsies of bladder carcinoma from Hadassah Medical Centre in Jerusalem. The intensity of the reaction and the number of tumour cells expressing H19 in each biopsy were evaluated in 56 patients, excluding biopsies with carcinoma in situ. The medical files were searched for demographic data and disease free survival. RESULTS: More than 5% of cells expressed H19 in 47 of the 56 (84%) biopsies. There was a decrease in the number of cells expressing H19 with increasing tumour grade (loss of differentiation) (p = 0.03). Disease free survival from the first biopsy to first recurrence was significantly shorter in patients with tumours having a larger fraction of H19 expressing cells, controlling for tumour grade. This was also supported by the selective analysis of tumour recurrence in patients with grade I tumours. CONCLUSIONS: It might be possible to use H19 as a prognostic tumour marker for the early recurrence of bladder cancer. In addition, for the gene therapy of bladder carcinoma that is based on the transcriptional regulatory sequences of H19, the expression of H19 in an individual biopsy could be considered a predictive tumour marker for selecting those patients who would benefit from this form of treatment.

Protein transfer of glycosyl-phosphatidylinositol (GPI)-modified murine B7-1 and B7-2 costimulators.

The feasibility of using protein transfer as a means for enhancing the immunogenicity of murine tumor cells was evaluated. Glycosyl-phosphatidylinositol (GPI)-modified variants of the murine costimulators B7-1 (CD80) and B7-2 (CD86), designated B7-1.GPI and B7-2.GPI, respectively, were immunoaffinity-purified from CHO-K1 cells transfected with glutamine synthetase amplification/expression constructs encoding each of these chimeric proteins. The proteins, once purified in detergent-depleted pseudomicelles, were exogenously incorporated into the membranes of several different murine tumor lines (EL-4, SMUCC-1, BW5147.3, P815, Ag104A, and EMT6). Successful membrane painting with the B7.GPI proteins was documented by immunofluorescence and flow cytometry, and membrane integration was verified by demonstrating that the reincorporated proteins were phosphatidylinositol-phospholipase C-sensitive, glycosyl-phosphatidylinositol-phospholipase D-resistant, and refractory to removal with dimyristylphosphatidylcholine vesicles. Significantly, B7-1.GPI and B7-2.GPI could be together copainted onto EL-4 cell surfaces with no interference observed between the two. A standard in vitro proliferation assay was used to show that both of the B7.GPI proteins retained costimulator function after membrane reincorporation. These findings further validate the therapeutic potential of protein-transferred costimulator.GPIs and pave the way for their combinatorial use in animal tumor models.

The expansion of human gammadelta T cells in response to Daudi cells requires the participation of CD4+ T cells.

The Burkitt's lymphoma cell line Daudi is a potent inducer of human gammadelta T-cell expansion. Using an in vitro culture system comprised of irradiated Daudi cells as stimulators and normal human lymphocytes as responders, the cellular determinants of this response were investigated. Three of four monoclonal antibodies (mAbs 1-1C4, L243, and 9.3F10) directed against disparate epitopes of human major histocompatibility complex (MHC) class II, as well as a mAb with specificity for CD4 (OKT4), inhibited the expansion of gammadelta T cells in response to Daudi cell stimulators. mAbs with a specificity for CD74 and CD8 were non-inhibitory. Lymphocyte depletion experiments demonstrated a critical role for the CD4+ T-cell subset in the expansion of gammadelta T cells. Other data pointed towards requirements for direct cell contact in this system, and the addition of exogenous recombinant interleukin (IL)-2, IL-4, and IL-12 failed to reconstitute gammadelta T-cell expansion in CD4+ lymphocyte-depleted cultures. These results complement previous findings in murine infectious disease and mycobacterial systems, providing a direct demonstration that CD4+ T cells play a role in gammadelta T-cell expansion through an interaction with human leucocyte antigen (HLA) class II on Daudi cells. The data point towards important functional links between the acquired and natural immune systems.

Expression of the imprinted H19 oncofetal RNA in epithelial ovarian cancer.

STUDY: To examine the expression of the imprinted maternally expressed H19 gene in benign, low malignant potential (borderline) and malignant surface epithelial ovarian tumors. DESIGN: In situ hybridization for H19 RNA using S-labeled and digoxigenin-labeled probes was performed on paraffin sections of ovarian surface epithelial tumors. The serous tumors included nine section cystadenomas, twelve serous tumors of low malignant potential and twenty serous carcinomas, grade I-IIII (FIGO classification). A smaller group included two mucinous cystadenomas, four mucinous tumors of low malignant potential and two mucinous cystadenocarcinomas. RESULTS: H19 expression was found to be positive in 6/9 (67%) serous cystadenomas, 9/12 (75%) of serous tumors of low malignant potential and 13/20 (65%) of invasive serous carcinomas. Expression in mucinous tumors was confined to the stroma beneath the epithelial lining. CONCLUSION: H19 is expressed in the majority of serous epithelial tumors. Taking into consideration the high percentage of H19 expressing serous ovarian neoplasms we suggest that H19 RNA may be used as an adjuvant tumor marker for the diagnosis and mainly for staging and follow-up of patients with serous ovarian carcinoma.

Placental protein 14 functions as a direct T-cell inhibitor.

Human placental protein 14 (PP14, also referred to as glycodelin and progesterone-associated endometrial protein) inhibits phytohemagglutinin (PHA)-stimulated T-cell proliferation and monokine secretion within PBMC populations. However, the mechanisms underlying these and other PP14 immunoinhibitory activities remain unclear. In the present study, we asked whether PP14's T-cell inhibitory effect is a direct one or, alternatively, an indirect consequence of accessory cell (AC) perturbation. Using either immunopurified PP14 or first-trimester amniotic fluid (AF) as a rich source of PP14, we documented inhibition of the proliferation of highly purified peripheral blood T-cells when stimulated with anti-CD3 mAbs or PHA in the presence of paraformaldehyde-fixed AC. Significantly, PP14 inhibited T-cell proliferation and IL-2 secretion induced by immobilized anti-CD3 and anti-CD28 mAbs in the absence of AC. PP14 depletion (via immunoprecipitation) abrogated AF's T-cell inhibitory activity, indicating that the PP14 within the amniotic fluid is required for this immunoregulatory effect. These findings establish that PP14 can inhibit T-cell proliferation in the absence of AC and thus add PP14 to the relatively restricted set of immunoinhibitory proteins that are known to target T-cells directly. Additional data demonstrate that PP14's inhibitory effect can be overridden by stimuli which circumvent early events during T-cell receptor (TCR) activation, namely, protein kinase C activators in combination with Ca2+ ionophores. These latter results suggest that PP14 inhibits early events in the TCR signaling pathway.

A receptor for the lipocalin placental protein 14 on human monocytes.

Human placental protein 14 (PP14), a member of the lipocalin structural superfamily, is an abundant amniotic fluid glycoprotein with documented immunoinhibitory activities. While receptors have been characterized for several other lipocalins, none have been reported to date for PP14. In the present study, two-color immunofluorescence and flow cytometry was used to screen peripheral blood mononuclear cell subpopulations for their capacity to engage fluoresceinated recombinant PP14. The tagged PP14 bound strongly in a specific and saturable fashion to CD14+ (monocyte lineage) cells, but not to CD20+ (B cell lineage) or CD3+ (T cell lineage) cells. This binding was both pH- and temperature-sensitive, and was reduced by proteolytic pre-digestion of the cells with trypsin or proteinase K. Scatchard analysis demonstrated a single class of receptors on CD14+ cells, with a K(D) of approximately 1 x 10(-8) and approximately 10-35,000 receptors per cell. These findings constitute the first report of a cell surface-associated binding protein for PP14 and set the stage for exploring the molecular mechanisms of PP14-mediated signaling and immunomodulation.

Negative signaling by anti-HLA class I antibodies is dependent upon two triggering events.

mAb with specificity for the alpha3 domain of HLA class I antigens, such as mAb TP25.99 and W6/32, are capable of inhibiting the proliferation of stimulated T cells in vitro by binding to their surface HLA class I antigens. The inhibitory potential of another HLA class I alpha3 domain-specific mAb, A1.4, was evaluated. In contrast to mAb TP25.99 and W6/32, which routinely inhibited superantigen (SEB) stimulation of T cells by >90%, mAb A1.4 at equivalent concentrations demonstrated only 20-50% inhibition. Univalent Fab fragments of all three mAb lacked inhibitory activity. Interestingly, however, by combining univalent W6/32 (or TP25.99) Fab fragments with intact, bivalent mAb A1.4 (at a non-inhibitory, sub-threshold concentration of 1 microg/ml), significant inhibition of SEB-driven T cell proliferation was obtained. Inhibition by the anti-HLA class I mAb W6/32 and TP25.99 was evident even when SEB was used in conjunction with paraformaldehyde-fixed HLA class I-, class II+ Daudi cells, suggesting that the inhibitory activity of these mAb results from direct HLA class I epitope engagement on the T cell. These findings suggest that effective antibody-mediated induction of the HLA class I inhibitory pathway within T cells is dependent upon two separable molecular triggers at the T cell surface. The first can be delivered by univalent mAb derivatives that engage one or more critical HLA class I epitope(s). The second requires intact mAb, though seems to be less selective as to the HLA class I specificity. This model may explain why some, but not all, anti-HLA class I mAb are inhibitory when used singly. Achieving synergies between a wider array of anti-HLA class I mAb and their derivatives may provide a path for more effectively tapping into the HLA class I inhibitory pathway in a therapeutic context.

Antisense modulation of the ICAM-1 phenotype of a model human bone marrow stromal cell line.

Efficient stable gene transfer was achieved in a model human bone marrow stromal cell line, KM-102, using both Epstein-Barr virus and BK virus episomal expression vectors. Using this episomal expression system, effective overexpression and inhibition of ICAM-1 expression was achieved in stably transfected KM-102 cells by sense and antisense RNA gene transfer, respectively. Loss of surface ICAM-1 on antisense KM-102 transfectants did not significantly affect adhesion to LFA-1-bearing JY hematopoietic cells. However, KM-102 ICAM-1 overexpressors demonstrated enhanced binding (2.5-fold) to phorbol ester-treated, but not untreated, LFA-1-bearing JY cells. The increased binding could be blocked with anti-ICAM-1 antibodies. These findings suggest that while ICAM-1 is not required for basal adhesion between stromal and hematopoietic cells, stromal ICAM-1 may contribute to stromal:leukemic cellular interaction when bound to the phorbol ester-dependent high-avidity state of hematopoietic LFA-1.

Non-glycosylated human B7-1(CD80) retains the capacity to bind its counter-receptors.

Though the cell surface-associated costimulator B7-1(CD80) is known to be highly N-glycosylated, the functional significance of this N-glycosylation has not been evaluated. Two experimental approaches were taken to assess the influence of N-glycosylation on human B7-1 function. First, stable K562 transfectants expressing human B7-1 were treated with the N-glycosylation inhibitor tunicamycin. This treatment reduced the levels of B7-1 at the cell surface as judged by both indirect immunofluorescence/flow cytometry and immunoprecipitation analyses. Significantly, the non-glycosylated cell surface-associated B7-1 on tunicamycin-treated cells retained the capacity to bind CTLA-4 x Ig, a soluble derivative of the CTLA-4(CD152) counter-receptor. Second, experiments were performed with bacterially-produced non-glycosylated derivatives of human B7-1, comprising either the complete B7-1 extracellular domain (hB7-1 x ed) or the membrane-proximal IgC-homologue domain of B7-1 in isolation (hB7-1 x IgC). While the hB7-1 x IgC derivative failed to bind to CTLA-4, the larger hB7-1 x ed derivative associated with CTLA-4 x Ig in cell-free binding assays. Futhermore, recombinant hB7-1 x ed effectively blocked B7-1-mediated costimulation in an in vitro T cell proliferation assay, suggesting that this soluble non-glycosylated B7-1 derivative is capable of engaging CD28, the B7 counter-receptor implicated in T cell activation. Taken together, these data indicate that the N-glycosylation of B7-1 is not required for its association with counter-receptors. Moreover, the findings pave the way for the therapeutic use of recombinant bacterial B7-1 derivatives as competitive inhibitors of B7-mediated signals.

Differential effects of chondroitin sulfates A and B on monocyte and B-cell activation: evidence for B-cell activation via a CD44-dependent pathway.

At inflammatory sites, proteoglycans are both secreted by activated mononuclear leukocytes and released as a consequence of extracellular matrix degradation. Chondroitin 4-sulfate proteoglycans constitute the predominant ones produced by activated human monocytes/macrophages. In this study, we show that two chondroitin 4-sulfate forms, CSA and CSB, can activate distinct peripheral blood mononuclear cell types. Whereas CSA activates monocytes (to secrete monokines), CSB activates B-cells (to proliferate). In contrast, the chondroitin 6-sulfate CSC and heparin do not exert these functional effects. We further show that CD44 monoclonal antibodies block CSB-induced B-cell proliferation. These findings point to glycosaminoglycans, and specifically chondroitin 4-sulfates, as a novel class of immunological mediators at inflammatory sites. Furthermore, the data link CD44 to B-cell activation, paralleling the established roles of CD44 in T-cell and monocyte activation.