Suppression of the CD8 T cell response by human papillomavirus type 16 E7 occurs in Langerhans cell-depleted mice.

Human papillomavirus (HPV) is an epitheliotropic virus that is the primary causal agent for cervical cancer. Langerhans cells (LC) are skin antigen presenting cells that are reduced in number in HPV-infected skin. The aim of this study was to understand the immune-modulatory effects of HPV16 E7 on LC and on the CD8 T cell response to a skin-expressed antigen. To test this, HPV16 E7 was expressed in mouse skin keratinocytes with the model antigen ovalbumin (Ova). Similar to what is observed in HPV-infected human skin, LC numbers were significantly reduced in E7-expressing mouse skin. This shows that expression of the E7 protein alone is sufficient to mediate LC depletion. Expression of E7 with Ova in keratinocytes strongly suppressed the Ova-specific CD8+ T cell response in the skin draining lymph node. When tested in LC-ablated mice, the CD8 T cell response to skin-expressed Ova in control mice was not affected, nor was the T cell response to Ova restored in E7-expressing skin. These data indicate a role for E7 in regulation of LC homeostasis in the skin and in suppression of antigen specific CD8 T cell expansion, but suggest that these two effects occur independent of each other.

Antibodies against listerial protein 60 act as an opsonin for phagocytosis of Listeria monocytogenes by human dendritic cells.

Human-monocyte-derived dendritic cells (MoDC) are very efficient in the uptake of Listeria monocytogenes, a gram-positive bacterium which is an important pathogen in humans and animals causing systemic infections with symptoms such as septicemia and meningitis. In this work, we analyzed the influence of blood plasma on the internalization of L. monocytogenes into human MoDC and compared the uptake of L. monocytogenes with that of Salmonella enterica serovar Typhimurium and Yersinia enterocolitica. While human plasma did not significantly influence the uptake of serovar Typhimurium and Y. enterocolitica by human MoDC, the efficiency of the uptake of L. monocytogenes by these phagocytes was strongly enhanced by human plasma. In plasma-free medium the internalization of L. monocytogenes was very low, whereas the addition of pooled human immunoglobulins resulted in the internalization of these bacteria to a degree comparable to the highly efficient uptake observed with human plasma. All human plasma tested contained antibodies against the 60-kDa extracellular protein of L. monocytogenes (p60), and anti-p60 antibodies were also found in the commercially available pooled immunoglobulins. Strikingly, in contrast to L. monocytogenes wild type, an iap deletion mutant (totally deficient in p60) showed only a minor difference in the uptake by human MoDC in the presence or the absence of human plasma. These results support the assumption that antibodies against the listerial p60 protein may play an important role in Fc-receptor-mediated uptake of L. monocytogenes by human MoDC via opsonization of the bacteria. This process may have a major impact in preventing systemic infection in L. monocytogenes in immunocompetent humans.

Targeting of lymphotoxin-alpha to the tumor elicits an efficient immune response associated with induction of peripheral lymphoid-like tissue.

A recombinant antibody-lymphotoxin-alpha fusion protein induced an adaptive immune response protecting mice from melanoma. Importantly, this fusion protein elicited the formation of a lymphoid-like tissue in the tumor microenvironment containing L-selectin+ T cells and MHC class II+ antigen-presenting cells, as well as B and T cell aggregates. Furthermore, PNAd+/TCA4+ high endothelial venules were observed within the tumor, suggesting entry channels for naive T cell infiltrates. Over the course of therapy, a marked clonal expansion of certain TCR specificities occurred among tumor-infiltrating lymphocytes that displayed reactivity against melanoma cells and the TRP-2(180-188) peptide. Consequently, naive T cells may have been recruited to as well as primed and expanded in the lymphoid-like tissue induced by the lymphotoxin-alpha fusion protein at the tumor site.

Expression of multilectin receptors and comparative FITC-dextran uptake by human dendritic cells.

Dendritic cells (DC) are potent antigen-presenting cells and understanding their mechanisms of antigen uptake is important for loading DC with antigen for immunotherapy. The multilectin receptors, DEC-205 and macrophage mannose receptor (MMR), are potential antigen-uptake receptors; therefore, we examined their expression and FITC-dextran uptake by various human DC preparations. The RT-PCR analysis detected low levels of DEC-205 mRNA in immature blood DC, Langerhans cells (LC) and immature monocyte-derived DC (Mo-DC). Its mRNA expression increased markedly upon activation, indicating that DEC-205 is an activation-associated molecule. In Mo-DC, the expression of cell-surface DEC-205 increased markedly during maturation. In blood DC, however, the cell-surface expression of DEC-205 did not change during activation, suggesting the presence of a large intracellular pool of DEC-205 or post-transcriptional regulation. Immature Mo-DC expressed abundant MMR, but its expression diminished upon maturation. Blood DC and LC did not express detectable levels of the MMR. FITC-dextran uptake by both immature and activated blood DC was 30- to 70-fold less than that of LC, immature Mo-DC and macrophages. In contrast to immature Mo-DC, the FITC-dextran uptake by LC was not inhibited effectively by mannose, an inhibitor for MMR-mediated FITC-dextran uptake. Thus, unlike Mo-DC, blood DC and LC do not use the MMR for carbohydrate-conjugated antigen uptake and alternative receptors may yet be defined on these DC. Therefore, DEC-205 may have a different specificity as an antigen uptake receptor or contribute to an alternative DC function.

Dendritic cell activation by danger and antigen-specific T-cell signalling.

Recent transplantation, animal and in vitro studies suggest a dependence of some immune reactions on tissue damage. Although many factors involved in enhancing immune responses through tissue damage have yet to be identified, recent data suggests that one of the targets of these cellular stress factors is the bone marrow derived dendritic cell (DC). DC are potent initiators of primary immune responses and hold the key to immune reactions through their ability to sense changes in their local environment and respond appropriately to induce T-cell immunity, or possibly tolerance. In the lymph node, DC are also influenced by antigen-specific signalling from T cells, which may extend and amplify DC antigen presenting capabilities, especially for the stimulation of cytotoxic responses. It now appears that both tissue damage and antigen-specific T-cell derived signals act together on the DC to promote the appropriate immune reaction to antigen. Thus DC antigen presenting behaviour is not only dependent on the context of antigen encounter in the periphery, but also on the availability of antigen-specific T cells and their T-cell receptor specificities.

Human decidua contains potent immunostimulatory CD83(+) dendritic cells.

Dendritic cells (DCs) are sentinel cells of the immune system important in initiating antigen-specific T-cell responses to microbial and transplantation antigens. DCs are particularly found in surface tissues such as skin and mucosa, where the organism is threatened by infectious agents. The human decidua, despite its proposed immunosuppressive function, hosts a variety of immunocompetent CD45 cells such as natural killer cells, macrophages, and T cells. Here we describe the detection, isolation, and characterization of CD45(+), CD40(+), HLA-DR(++), and CD83(+) cells from human early pregnancy decidua with typical DC morphology. CD83(+) as well as CD1a(+) cells were found in close vicinity to endometrial glands, with preference to the basal layer of the decidua. In vitro, decidual CD83(+) cells could be enriched to approximately 30%, with the remainder of cells encompassing DC-bound CD3(+) T cells. Stimulation of allogeneic T cells in a mixed leukocyte reaction by the decidual cell fraction enriched for CD83(+) cells, was similar to that obtained with blood monocyte-derived DCs, demonstrating the potent immunostimulatory capacity of these cells. Decidual DCs with morphological, phenotypic, and functional characteristics of immunostimulatory DCs might be important mediators in the regulation of immunological balance between maternal and fetal tissue, leading to successful pregnancy.

Functions of myeloid and lymphoid dendritic cells.

The bone marrow derived dendritic cell (DC) is an essential antigen presenting cell (APC) for the initiation of primary, T cell based immune responses. DC are a heterogenous haematopoietic lineage, in that many subsets from different tissues show different surface phenotypes, but the ability to stimulate antigen specific naïve T cell proliferation appears to be shared between these DC subsets. It has been suggested that the so called myeloid and lymphoid-derived subsets of DC perform distinct stimulatory or tolerogenic functions. However, recent data has blurred this apparent distinction of DC subset function and shown that both subsets are at least capable of stimulatory and possibly even tolerogenic functions. Thus, the immunoregulatory potential of DC may depend less on ontology than on recent activatory or downregulatory stimuli.

Differential susceptibility to CD95 (Apo-1/Fas) and MHC class II-induced apoptosis during murine dendritic cell development.

Disappearance of antigen presenting cells (APC) from the lymph node occurs following antigen specific interactions with T cells. We have investigated the regulation of CD95 (Apo-1/Fas) induced apoptosis during murine dendritic cell (DC) development. Consistent with the moderate levels of CD95 surface expression and low, or absent, MHC class II expression, immature DC in bone marrow cultures were highly sensitive to CD95 induced apoptosis, but insensitive to class II mediated apoptosis. In contrast, mature splenic, epidermal and bone marrow derived DC were fully resistant to CD95 induced cell death, but sensitive to class II induced apoptosis. Although caspase 3 and 8 activation was detected in immature DC undergoing CD95L-induced apoptosis, the pan-caspase inhibitor zVAD-fmk did not inhibit the early events of CD95-induced mitochondrial depolarisation or phosphatidyl serine exposure and only partially inhibited the killing of immature DC. In contrast, zVAD-fmk was completely effective in preventing CD95L mediated death of murine thymocytes. Collectively, these data do not support a major role of CD95: CD95L ligation in apoptosis of mature DC, but rather emphasise the existence of distinct pathways for the elimination of DC at different stages of maturation.

Induction of dendritic cell costimulator molecule expression is suppressed by T cells in the absence of antigen-specific signalling: role of cluster formation, CD40 and HLA-class II for dendritic cell activation.

Full activation of T lymphocytes by dendritic cells (DC) during antigen presentation is known to require the interaction of several inducible receptor-ligand pairs. We have postulated that the reciprocal activation of DC by T lymphocytes is also important. Potential signalling molecules that might increase the stimulatory capacity of DC during antigen presentation to T lymphocytes were tested using an in vitro model. Fresh human blood DC were cocultured with CD4+ and CD8+ allogeneic or with autologous T lymphocytes plus Staphylococcus superantigen A (SEA). Surprisingly, costimulator expression on DC cocultured with T lymphocytes was reduced in comparison to DC cultured alone. However, the minority (10-30%) of DC clustering with T lymphocytes showed antigen-specific up-regulation of the CD40, CD80 and CD86 costimulator molecules, whereas the non-clustered DC (70-90%) had less up-regulation than control DC cultured alone and did not respond to antigen-specific triggering. Monoclonal antibodies (mAb) to CD40 ligand (CD40L) and human leucocyte antigen (HLA)-DR, but not lymphocyte function-associated antigen-1 (LFA-1), LFA-3 or HLA-class I, significantly inhibited the T-lymphocyte induction of DC costimulator expression. Since HLA-class II, but not HLA-class I mAb, inhibited allogeneic T-lymphocyte-mediated activation of DC, CD4 T lymphocytes appear to be the main subset activating DC in the mixed lymphocyte reaction. Cross-linking of CD40, but not HLA-class II, up-regulated DC or B-cell costimulator expression. Although direct class II signalling does not appear to play a role in DC activation, antigen-specific T-cell recognition contributes via other mechanisms to regulate DC activation.

Human dendritic cells express a 95 kDa activation/differentiation antigen defined by CMRF-56.

Despite the unique functions of dendritic cells (DC), only two cell surface antigens (CMRF-44 and CD83) with relatively restricted expression on human DC have been described to date. We describe a third mAb, CMRF-56, which recognizes another DC early activation/differentiation antigen with limited expression on other haemopoietic cell populations. Circulating blood leukocytes did not express the CMRF-56 antigen and, following either in vitro culture or activation of PBMC populations, CMRF-56 antigen expression was detected only on DC and a subpopulation of CD19+ lymphocytes. Circulating blood DC were CMRF-56 but induced expression within 6 h of in vitro culture. This, together with the finding that tonsil and synovial fluid DC upregulate the antigen following short-term in vitro culture, confirmed that CMRF-56 recognizes an early activation antigen on DC. Isolated Langerhan's cells, dermal DC, migratory dermal DC and monocyte derived DC (GM-CSF/IL-4/TNFalpha) also express the CMIRF-56 antigen. Antigen modulation studies demonstrated that the amount of cell surface bound CMRF-56 and CMRF-44 (but not CD83) mAb was dramatically reduced by short-term incubation at 37 degrees C. This effect was not due to internalization and the reduction in CMRF-56 binding was a reversible, temperature-dependent process. In contrast, the decrease in CMRF-44 binding was irreversible, suggesting that following ligation the CMRF-44 antigen undergoes an irreversible conformational change or shedding at 37 degrees C. Western blotting confirmed that CMRF-56 recognizes a previously undescribed 95 kDa activation antigen whose cellular distribution and expression kinetics overlaps with, but is clearly distinguishable from, that of the CD83 and CMRF-44 antigens. CMRF-56 therefore provides a useful additional marker for studies on human DC.

Dermal dendritic cells associated with T lymphocytes in normal human skin display an activated phenotype.

The CMRF-44 and CD83 (HB15) antigens are associated with functional maturation and activation of blood dendritic cells (DC). We describe the expression of these antigens on freshly isolated epidermal Langerhans cells and dermal DC as well as the distribution of CD83+/ CMRF-44++-activated DC within sections of normal human skin. Fresh Langerhans cells were prepared by standard techniques and large numbers of enriched (25%-55%), viable dermal DC were obtained using an improved collagenase treatment protocol with density gradient enrichment. Freshly isolated Langerhans cells and dermal DC had similar costimulator and activation antigen expression, and both stimulated moderate levels of allogeneic T lymphocyte proliferation as determined in the 7 d mixed leukocyte reaction. In situ labeling of DC within skin sections revealed a population of CD83 and CMRF-44 positive dermal cells of which most (approximately 75%) were in intimate contact with CD3+ T lymphocytes, especially in the adnexal regions. In contrast, only 25%-30% of the more numerous CD1a++ dermal DC population were directly apposed to T lymphocytes. The CMRF-44++ dermal DC population stimulated an allogeneic mixed leukocyte reaction, confirming their identity as DC. These data, plus comparative data obtained for migratory dermal DC, suggest that only a small proportion of dermal DC have been triggered to a more advanced state of differentiation or activation. The striking association of the activated dermal DC population with T lymphocytes suggests that communication between these two cell types in situ may occur early in the immune response to cutaneous antigen.

Human dendritic cells express functional interleukin-7.

Interleukin-7 (IL-7) supports the proliferation of mature T lymphocytes, however, the cellular source of IL-7 for T lymphocyte activation has not been well established. We therefore investigated whether human peripheral blood dendritic cells (DC) produce IL-7 as a contribution towards T lymphocyte activation. Human CMRF-44+/CD14-/CD19- low density DC, purified after overnight tissue culture, contained IL-7 transcripts, detected by direct cell reverse transcription-polymerase chain reaction. Intracytoplasmic staining confirmed IL-7 protein in at least a subpopulation of cultured low density DC. In contrast, resting/immature DC, isolated directly by immunodepletion of lineage marker positive cells, contained no IL-7 mRNA. Thus, the expression of IL-7 by DC follows the pattern described previously for CD80, CD86 and CD40. However, tissue culture of purified resting/immature DC, in contrast to CD80, CD86 and CD40, failed to induce IL-7 transcripts. The functional importance of DC IL-7 expression was demonstrated in an allogeneic mixed leukocyte reaction (MLR). Neutralising mAb to IL-7 significantly inhibited T lymphocyte proliferation when low DC numbers were used, but at higher stimulator numbers, anti-IL-7 mAb failed to inhibit an allogeneic MLR. This suggests, that when DC are in excess, other co-stimulatory pathways can compensate for the lack of IL-7. Addition of IL-7 to a MLR caused a significant increase in the proliferative response stimulated by monocytes and B lymphocytes but not by DC. These data support the concept of an initial phase of antigen uptake by DC followed by the optimisation of DC co-stimulatory potential. The co-stimulatory repertoire expressed, including IL-7, may be regulated by exogenous stimuli, thereby ensuring DC flexibility in mounting a response appropriate to the environmental changes.

Isolation of human blood dendritic cells using the CMRF-44 monoclonal antibody: implications for studies on antigen-presenting cell function and immunotherapy.

Dendritic cells (DC) are potent antigen-presenting cells (APC) with the capacity to stimulate a primary T lymphocyte immune response and are therefore of interest for potential immunotherapeutic applications. Freshly isolated DC or DC precursors may be preferable for studies of antigen uptake and the potential control of APC costimulator activity. In this report, we report that the monoclonal antibody CMRF-44 can be used to detect early DC differentiation. The majority of DC circulating in blood do not express any known DC lineage specific markers, but can be identified by CMRF-44 labeling after a brief period of in vitro culture. The sequential acquisition of DC activation antigens allows the identification of two stages of DC maturation/activation. Cytokines, especially granulocyte-macrophage colony-stimulating factor (GM-CSF) and tumor necrosis factor (TNF)alpha, enhance both phases of this process, whereas CD40-ligand trimer preferentially enhances the final DC maturation to a fully mature, activated phenotype. DC positively selected using CMRF-44 possess potent allostimulatory activity and are efficient at the uptake, processing, and presentation of soluble antigens for both primary and secondary immune responses. CMRF-44+ DC are also more potent than other APC types at restimulation of a chronic myeloid leukemia peptide specific T-cell clone. The use of a purified population of freshly isolated DC may be advantageous in attempts to initiate, maintain, and direct immune responses for immunotherapeutic applications.

Identification of a novel dendritic cell surface antigen defined by carbohydrate specific CD24 antibody cross-reactivity.

Dendritic cells (DC) are characterized as leucocytes that lack mature lineage specific markers and stimulate naive T-lymphocyte proliferation in vitro and in vivo. The mouse heat stable antigen (HSA) participates in T lymphocyte co-stimulation and is expressed by DC isolated from thymus, skin and spleen. The human HSA homologue, CD24, is predominantly expressed by B lymphocytes and granulocytes, but its expression on DC has not been studied in detail. CD24 clearly participates in B-lymphocyte signalling but co-stimulatory activity for T lymphocytes has not yet been described. We have examined the expression of CD24 on human peripheral blood DC populations isolated directly or following in vitro culture. The CD24 antigen was absent from blood DC however, cross-reactive sialylated carbohydrate epitopes were detected on DC with some CD24 monoclonal antibodies (mAb). These CD24 mAb define a protein surface antigen, which is expressed by an immature or resting subpopulation of peripheral blood DC and is down-regulated following activation differentiation in vitro.

Human dendritic cells activate T lymphocytes via a CD40: CD40 ligand-dependent pathway.

The CD40:CD40 ligand (CD40L) interaction provides T lymphocyte-mediated help for B lymphocyte and monocyte function but has also been shown to serve as a co-stimulus for T lymphocyte activation. In this report, we studied the regulation of CD40 expression and its functional relevance for the human dendritic cell (DC) stimulation of T lymphocytes. Only a small subpopulation of directly isolated blood DC expressed CD40. However, CD40 was rapidly up-regulated by culture, and its expression was further enhanced by interleukin (IL)-1 alpha, IL-1 beta, IL-3, tumor necrosis factor-alpha and granulocyte/macrophage-colony-stimulating factor. Expression of CD40L on DC was not detected. The proliferation of T lymphocytes in an allogeneic mixed leukocyte reaction, stimulated by blood DC or epidermal Langerhans cells, was significantly reduced in the presence of the CD40 immunoglobulin (CD40Ig) fusion protein or CD40L monoclonal antibodies. Cross-linking of CD40 on directly isolated DC with mouse CD40L trimer (mCD40LT) markedly augmented CD80 and CD86 up-regulation. Nevertheless, the same cross-linking mCD40LT inhibited DC stimulated T lymphocyte proliferation. When CD40Ig was added simultaneously with CTLA-4Ig, only minimal and variable additional inhibition of DC-stimulated allogeneic T lymphocyte proliferation and IL-2 secretion was observed, compared to each fusion protein alone. These results suggest that both CD80/CD86-dependent and -independent components of DC-T lymphocyte CD40:CD40L co-stimulation exist and further emphasize that the majority of blood DC have to differentiate or be activated to express co-stimulatory molecules.

Intercellular adhesion molecule-3 is the predominant co-stimulatory ligand for leukocyte function antigen-1 on human blood dendritic cells.

Dendritic cells (DC) are potent stimulators of primary T lymphocyte responses to foreign antigen. The initial DC-T lymphocyte interaction involves the binding of the adhesion molecule leukocyte function antigen-1 (LFA-1; CD11a/CD18) on the T lymphocyte to an intercellular adhesion molecule (ICAM) on the DC. Although blood and tonsil DC express ICAM-1 (CD54) and ICAM-2 (CD102) on their surface, anti-ICAM-1 and anti-ICAM-2 monoclonal antibodies (mAb) have little inhibitory activity on the DC-stimulated mixed leukocyte reaction (MLR). We therefore examined the expression of the more recently identified LFA-1 ligand, ICAM-3 (CD50), in comparison to ICAM-1 and ICAM-2 on blood DC and sought a functional role for ICAM-3 in DC-mediated T lymphocyte responses. Resting blood DC expressed significantly more ICAM-3 than ICAM-1 or ICAM-2 as assessed by flow cytometry. Treatment of resting DC with interferon-gamma led to increased expression of ICAM-1; however, ICAM-2 and ICAM-3 levels remained relatively constant. Solid-phase recombinant chimeric molecules ICAM-1-, ICAM-2- and ICAM-3-Fc were able to co-stimulate CD4+ T lymphocyte proliferation in conjunction with suboptimal solid-phase CD3 mAb 64.1. However, the anti-ICAM-3 mAb CAL 3.10 inhibited a DC-stimulated MLR to a greater extent than anti-ICAM-1 or anti-ICAM-2 reagents and appeared to act by blocking the DC ICAM-3- T lymphocyte LFA-1 interaction. As ICAM-3 is the predominant LFA-1 ligand on resting blood DC, we postulate that DC may utilize ICAM-3 for initial DC- T lymphocyte interactions, and that ICAM-1, which is up-regulated upon DC activation, and/or ICAM-2, may contribute to DC migration or later phases of the T lymphocyte activation process.

Activation of human peripheral blood dendritic cells induces the CD86 co-stimulatory molecule.

Maximal T lymphocyte responses require presentation of antigen by major histocompatibility complex molecules and delivery of one or more co-stimulatory signals. Interaction of the CD28 molecule on T lymphocytes with its ligands on antigen-presenting cells (APC) initiates a critical co-stimulatory pathway inducing T lymphocyte proliferation and cytokine secretion. Dendritic cells (DC) are potent APC for a primary T lymphocyte response and potential CD28/CTLA-4 ligands on DC are, therefore, of particular functional relevance. In these experiments, the expression and function of the CD28/CTLA-4 ligands B7.1 (CD80) and B7.2 (CD86) were examined on human blood DC. Resting DC populations directly isolated by immunodepletion of lineage marker-positive cells lacked cell membrane expression of CD80 and expressed little or no CD86, although CD86, but not CD80 mRNA was detected by reverse transcription-polymerase chain reaction analysis. In contrast, low-density DC isolated after culture in vitro strongly expressed CD86 surface protein, but expressed limited or no CD80, although mRNA for both molecules were detected. Short-term culture of directly isolated DC up-regulated both CD80 and CD86 expression. Analysis of the kinetics of CD28/CTLA-4 ligand induction showed that surface CD86 was present within 8 h, whereas CD80 antigen was first detected after 24 h of culture. The functional importance of CD28/CTLA-4 ligand up-regulation on DC during T lymphocyte interactions was demonstrated by the ability of both CTLA-4Ig and CD86 monoclonal antibodies (mAb), but not CD80 mAb, to block an allogeneic mixed lymphocyte reaction stimulated by DC populations initially negative for CD80 and CD86. These results demonstrate that CD86 is both the earliest and functionally the predominant co-stimulatory CD28/CTLA-4 ligand on DC.

Isolation of human blood dendritic cells by discontinuous Nycodenz gradient centrifugation.

The most potent antigen presenting cell present in peripheral blood, lymphoid and non-lymphoid tissue is the dendritic cell (DC). The study of human DC has been restricted by their low frequency in the tissues and the lack of a truly DC specific surface marker to assist in identification and isolation. Standard techniques for the isolation of blood DC generally employ a period of in vitro culture followed by flotation on dense albumin gradients, or more recently, discontinuous gradients of metrizamide. Dense albumin gradients are time consuming to prepare, giving low and variable yields of DC. Metrizamide is more convenient, although exposure of monocytes to metrizamide can decrease the expression of CD14 and alter the accessory cell properties of antigen presenting cells. Here we demonstrate that Nycodenz gradient centrifugation of 16 h cultured, T lymphocyte depleted, peripheral blood mononuclear cells (PBMC) reliably yields a population of low density cells that is highly enriched for DC. Most B and residual T lymphocytes are depleted and NK cell numbers are reduced two-fold from the interface cell population. The high density pellet fraction exhibits very little allostimulatory activity, indicating that few DC pass into the pellet. The low density fraction contains a significant population (20 +/- 5 (SD)%, n = 8) of cells which fail to stain for the lineage markers CD3, CD11b, CD14, CD16, CD19 and CD57. Nycodenz exhibits low toxicity, does not alter the allostimulatory activity of antigen presenting cells, and is therefore ideal for the isolation of cultured DC.