Penelope A. Bedford

Transfer of antigen between dendritic cells in the stimulation of primary T cell proliferation

Primary proliferative T cell responses require stimulation with antigen‐pulsed dendritic cells (Ag‐DC). Here we show that for optimal stimulation, dendritic cells (DC) not exposed directly to antigen are also required. Ag‐DC added to DC‐depleted T cells caused negligible primary stimulation; adding back DC resulted in stimulation. These effects were seen using the contact sensitizer fluorescein isothiocyanate (FITC), FITC conjugated to ovalbumin (FITC‐OVA) or influenza virus as antigens. DC co‐cultured with Ag‐DC (using FITC or FITC‐OVA) acquired antigen indicating that antigen was transferred between DC. DC that acquired antigen secondarily were separated by cell sorting and stimulated primary T cell proliferation directly. DC were also pulsed with FITC, washed thoroughly and incubated overnight. Super natants contained shed antigen since DC incubated in these supernatants acquired antigen as indicated by flow cytometry. DC acquiring the shed antigen also stimulated T cell proliferation although the stimulation was not as effective as that seen when cell contact between DC and antigen‐bearing DC occurred. Thus, in primary stimulation, activation of T cells may occur when there is an antigen gradient between Ag‐DC and DC and the mechanisms underlying these effects are now being sought. We propose that this unique interaction between antigen‐presenting cells may be a paradigm for self/non‐self discrimination.

Induction of autoimmunity with dendritic cells: Studies on thyroiditis in mice

The initiation and maintenance of thyroid autoimmunity by professional antigen-presenting cells were assessed by observing thyroiditis and induction of IgG antibodies to thyroglobulin (Tg). Dendritic cells (DC) were purified from spleens of CBA mice and T cells removed with anti-Thy 1 and complement. Some DC were pulsed with 25-500 micrograms/ml of mouse Tg in vitro and normal syngeneic mice received injections of 10(5) cells intravenously. In untreated animals only 1 thyroid out of 40 showed a lymphocyte infiltrate and antibody to Tg was rarely seen. In animals receiving normal DC without Tg, lymphocyte infiltration was seen 2-6 weeks later in 5 out of 33 thyroids and some animals produced low levels of antibody to thyroglobulin (8 of 33 animals). DC pulsed with 500 micrograms Tg/ml in vitro caused thyroid infiltration in 6 out of 15 animals but did not increase the incidence of anti-Tg antibodies. Lower doses had no effect. When 10(5) DC were given from animals with experimental allergic thyroiditis (EAT, induced with Tg in complete Freunds adjuvant, CFA) more than half of the recipient animals showed thyroiditis (8 out of 15) and autoantibody production (12 of 15 animals). DC may therefore play a role in the initiation and maintenance of autoimmunity by providing a stimulus for antigen-specific T cells.

Human blood dendritic cells: binding to vascular endothelium and expression of adhesion molecules

To investigate the binding properties of dendritic cells (DC) to vascular endothelium, a comparative analysis was undertaken of DC, monocytes and lymphocytes isolated from the blood of 25 healthy subjects using monolayers of human umbilical vein endothelial cells as the adherence substrate. More blood DC (mean 24% adherence) were adherent to endothelial monolayers than monocytes (mean 18%; P < 0.001) and lymphocytes (mean 12%; P < 0.001). When the monolayers were pretreated with tumour necrosis factor‐alpha (TNF‐α) all leucocyte populations exhibited an increased attachment, but there was still greater binding of DC (mean 37% adherence) in comparison with monocytes (mean 23%; P < 0.001) and lymphocytes (mean 18%; P < 0.001). Flow cytometric analysis revealed that in relation to monocytes and lymphocytes the DC had a higher surface expression of the adhesion molecules CD11a (P < 0.05), CD11c (P < 0.05) and CD54 (P < 0.05) but a lower prevalence of cells bearing CD49d (mean 38%; P < 0.05) and the homing receptor CD62L (mean 14%; P < 0.001). CD1a was present on 22% of DC and virtually absent from the surface of monocytes and lymphocytes. The intensity of expression of the β1‐integrins, CD49c, CD49d and CD49e was greater on DC than lymphocytes and monocytes (P < 0.05). Antibody blocking studies demonstrated that DC binding to untreated and TNF‐α‐treated endothelium was dependent upon the expression of CD11a, CD18 and CD49d, and the simultaneous application of anti‐CD18 and anti‐CD49d antibodies produced an approximate 70% inhibition of adhesion (P < 0.001). Thus, the expression of both β1‐ and β2‐integrins contributes to the adhesive interaction between DC and endothelium.

Adipose tissue of human omentum is a major source of dendritic cells, which lose MHC class II and stimulatory function in Crohn's disease

Adipose tissue is reported to contain monocyte‐like pre‐adipocytes, which may mature into macrophages, contributing to local inflammation. Dendritic cells (DC) can be derived from monocytes and initiate and regulate primary immune responses. We hypothesized, therefore, that adipose tissue may provide DC involved in local immune activity. To test this, we studied cells from human omental adipose tissue samples from 17 patients with benign gynecological disease. The hypothesis that adipose tissue DC are involved in inflammatory disease was tested by comparing these cells with those from 18 patients with Crohns disease, where hypertrophy of adipose tissue suggests involvement in disease. A high proportion of the 1.33 ± 0.12 × 105 CD45‐positive cells/mg, obtained from control omenta, expressed CD11c, CD1a, and CD83; costimulatory molecules CD40, CD80, and CD86; and major histocompatibility complex (MHC) Class II but little CD14, CD16, or CD33. Omental cells showing morphological characteristics of DC were also observed. Metrizamide gradient‐enriched DC from these populations were potent stimulators of primary proliferation of allogeneic T cells in mixed leukocyte reactions. Increased numbers of CD45+ cells from omentum of Crohns patients (4.50±1.08×105 CD45+ cells/mg) contained higher percentages of CD11c+ and CD40+ cells (80.8±3.8% vs. 63.4±6, P=0.032; 77.9±4% vs. 58.8±6.5, P=0.029, respectively), but MHC Class II and stimulatory capacity were almost completely lost (P=<0.001), suggesting innate activation but lost capacity to stimulate adaptive immune responses. Granulocytes were also present amongst the omental cells from Crohns patients. Results indicated that omentum may provide DC, which could “police” local infections and contribute to and/or reflect local inflammatory activity.

CD4 expression on dendritic cells and their infection by human immunodeficiency virus.

Infection of dendritic cells (DC) by human immunodeficiency virus (HIV) has been disputed. Employing a fluorescence-activated cell sorter, DC, identified by the absence of membrane markers for T, B, natural killer (NK) and monocytic cells and by high levels of MHC class II DR antigen, were shown to express low levels of CD4. Immunomagnetic beads were used to separate blood low density cells, which are enriched for DC, into CD4-positive and -negative populations. Examination of these cells by electron microscopy showed an increase in the percentage of cells with DC morphology in the CD4-positive fraction and a reduction in the CD4-negative fraction. Electron microscopy of semi-purified DC preparations infected in vitro for 5 days with HIV-1 revealed morphologically distinct veiled DC with mature virions on the cell surface and virus budding through the cell membrane. Further evidence for the growth of HIV in DC was provided by experiments in which DC were extensively depleted of contaminating lymphocytes and monocytes prior to infection. Estimation of provirus load by a nested PCR indicated that after 5 days an infection level of one provirus copy per five cells could be achieved. After 7 days the provirus copy number could exceed the cellular genome copy number, suggesting that some cells had more than one provirus. Infectious virus could not be demonstrated in these cultures after 24 h but was detected after 5 or 7 days. Infection of DC in the presence of antibodies against CD4 was inhibited and suggests infection occurs via a CD4-dependent pathway. These results confirm that DC are susceptible to HIV infection in vitro. The immunological consequences of DC infection in vivo may be significant in the pathogenesis of AIDS.

IL-10-Producing B220+CD11c− APC in Mouse Spleen

APC acting at the early stages of an immune response can shape the nature of that response. Such APC will include dendritic cells (DCs) but may also include populations of B cells such as marginal zone B cells in the spleen. In this study, we analyze APC populations in mouse spleen and compare the phenotype and function of B220+CD11c− populations with those of CD11c+ spleen DC subsets. Low-density B220+ cells had morphology similar to DCs and, like DCs, they could stimulate naive T cells, and expressed high levels of MHC and costimulatory molecules. However, the majority of the B220+ cells appeared to be of B cell lineage as demonstrated by coexpression of CD19 and surface Ig, and by their absence from RAG-2−/− mice. The phenotype of these DC-like B cells was consistent with that of B cells in the marginal zone of the spleen. On bacterial stimulation, they preferentially produced IL-10 in contrast to the DCs, which produced IL-12. Conventional B cells did not produce IL-10. The DC-like B cells could be induced to express low levels of the DC marker CD11c with maturational stimuli. A minority of the B220+CD11c− low-density cells did not express CD19 and surface Ig and may be a DC subset; this population also produced IL-10 on bacterial stimulation. B220+ APC in mouse spleen that stimulate naive T cells and preferentially produce IL-10 may be involved in activating regulatory immune responses.

Adhesion molecules are upregulated on dendritic cells isolated from human blood

This study investigated whether the high expression of adhesion molecules on enriched preparations of circulating dendritic cells (DCs) was an intrinsic property of the cells or whether it was a consequence of the procedure used to isolate them from blood. Expression of the &bgr;1, &bgr;2 integrins (CD11/CD18 family) and other adhesion molecules on DCs in whole blood was compared with that on isolated DCs. Dendritic cells were identified by flow cytometry as leucocytes that were positive for human leucocyte antigen (HLA)‐DR, but negative for CD3, CD14, CD16, CD19 and CD56. In contrast to a minority of DCs in whole blood, the majority of isolated DCs expressed the &bgr;2 integrins and there were a greater number of cells bearing CD44, CD54 and some of the &bgr;1 integrins (notably CD49b, CD49d, CD49e and CD29). An increase in the proportion of DCs bearing adhesion molecules was generally apparent at the isolation stage when mononuclear cells, which had been incubated overnight, were centrifuged on a metrizamide gradient to enrich for cells of low density. Inclusion of an inhibitor of protein glycosylation and exocytosis (brefeldin A) at all stages of separation partially prevented an increase in the percentage of DCs bearing CD18, C29 and C54 whereas the inclusion of cycloheximide (an inhibitor of polypeptide synthesis) interfered with increases in the percentage of cells bearing CD29 and CD54. Neither of these antagonists had an effect on the intensity of adhesion molecule expression. We suggest that some of the adhesion‐dependent functions of isolated DCs are caused, in part, by an upregulation of surface adhesion molecules induced by the enrichment procedure.

The role of dendritic cells in the initiation of immune responses to contact sensitizers: II. Studies in nude mice

Twenty-four hours after skin painting nude mice with picryl chloride, there was an increase in the number of dendritic cells (DC) isolated from the draining lymph nodes. This increased inflow or retention of DC in lymph nodes following skin painting is therefore unlikely to depend on interaction of DC with T cells. The DC obtained initiated primary proliferative responses in vitro in lymph node cells from congenic euthymic mice. Contact sensitivity developed in congenic mice when they received footpad injections of 60,000 DC from the lymph nodes of nude mice skin sensitized 1 day previously with picryl chloride or oxazolone. The initiation of delayed hypersensitivity was therefore independent of T-cell contamination within the donor DC.

Developing dendritic cells become ‘lacy’ cells packed with fat and glycogen

On maturation, dendritic cells (DCs) become highly active cells equipped for antigen uptake, migration and clustering and activation of T cells. We therefore asked whether DCs acquire fat and glycogen stores as they mature. DCs were generated from mouse bone marrow stem cells by culturing with granulocyte–macrophage colony‐stimulating factor (GM‐CSF) for 7–8 days. Stimulation of the DCs with lipopolysaccharide (LPS) for the last 24 hr of culture, or exposure to 1–15 ng/ml of interleukin (IL)‐4 during development, resulted in production of DCs not only with an increased ability to stimulate T cells but also with an increasingly lacy appearance on transmission electron microscopy, with multiple unstained areas in the cytoplasm. This changed morphology was associated with the presence of increasing amounts of fat and glycogen, identified by Sudan Black and periodic acid leukofushin/Schiff (PAS) staining, respectively. Lacy DCs up‐regulated type 1 and type 2 scavenger receptors, providing possible mechanisms contributing to these changes. Lacy DCs were found occasionally amongst freshly isolated splenic and lymph node DCs. DCs can be isolated from human adipose tissue, and we tested whether lacy DCs acquiring fat and glycogen were present in mouse omentum. CD45+ cells migrating from the omentum expressed specific DC markers CD11c and 33D1, costimulatory molecules and major histocompatibility complex (MHC) class II, and most showed darkly staining fat inclusions. Thus, during development, DCs can acquire large amounts of fat and glycogen, accumulation of which is promoted by antigen exposure and modulated by the cytokine milieu and location, and which may act as a link between energy stores and immune function.

Interleukin-4 can induce interleukin-4 production in dendritic cells

The presence of interleukin‐4 (IL‐4) during the generation of dendritic cells (DC) from precursor cells results in measurable increases of IL‐12 in supernatants but IL‐4 secretion has not been reported. However, DC have IL‐4 receptors and are able to make IL‐4. We therefore sought evidence for autocrine effects of IL‐4 on DC. IL‐4 gene expression was low in DC generated from bone‐marrow stem cells in the presence of granulocyte–macrophage colony‐stimulating factor but was up‐regulated by exposure of the developing DC to IL‐4. Exposure to IL‐4 also induced intracellular IL‐4 production in DC. The intracellular IL‐4 induced in the presence of IL‐4 was increased following further DC maturation with tumour necrosis factor‐α. By contrast, in supernatants of DC, IL‐4 was rarely detected and only at late culture periods. However, after exposure of DC to IL‐4, cell‐bound IL‐4 was detected transiently, which suggested binding and internalization of the cytokine. Binding via IL‐4 receptor‐α was indicated from phosphorylation of the signal transducer and activator of transcription (STAT) protein 6, which is known to mediate IL‐4 function. Cytokine persisting within the supernatants of the cells may therefore be unrepresentative of the actual production and function of IL‐4 in the cells; IL‐4 may be produced in DC in response to exposure to IL‐4 but may then be lost from the supernatants during cell binding and activation of the cells.