In vitro expansion and characterization of dendritic cells derived from human bone marrow CD34+ cells.

Dendritic cells (DC), as professional antigen-presenting cells, play a major role in stimulating naive T cell responses in vivo and in vitro, and may exacerbate or modulate T lymphocyte-mediated reactions, such as interactions between a hematopoietic graft and the recipient, eg GVHD and graft-versus-leukemia. Here, we describe a two-stage cell culture system for expansion of functionally active human DC from CD34+ marrow precursors. Optimal outgrowth was achieved by initially culturing CD34+ cells for 5 days in medium containing GM-CSF, MGF and TNF-alpha. Substitution of CD40L and IL-4 for TNF-alpha during a subsequent 5-day subculture increased DC content, such that by 10 days the cultures contained approximately 40% DC as determined by immunophenotype and morphology. An increase in DC purity to 84% at 10 days was achieved by immunomagnetic separation for CD1a+ cells from 5-day cultures and subculturing these cells in medium with IL-4 and CD40L. Reversing the sequence of growth factors during culture and subculture decreased the yield and purity of DC. Expression of CD80 and CD86 was enhanced by adding CD40L and IL-4, and the DC showed stimulatory activity in MLC. In conclusion, we have described a simple two-stage culture system to generate functional DC from CD34+ marrow precursors.

Function and cell surface phenotype of dendritic cells from rat cornea.

PURPOSE: To isolate dendritic cells (DC) from rat corneas and to examine their functions and surface markers in vitro. METHODS: Cells were isolated enzymatically from dissected rat corneas and cultured for various intervals of time. Dendritic cells were enriched immunomagnetically from corneal cell preparations using monoclonal antibodies against DC surface antigens and tested for functional activity in lymphocyte stimulation assays. In vivo migration of DC was induced by traumatizing the corneal epithelium. Wholemounts of epithelial sheets were stained immunofluorescently with anti-DC antibodies and examined by confocal microscopy. Dendritic cells isolated from traumatized corneas were tested for functional activity. RESULTS: Corneal DC exhibited the properties of other members of the DC family, i.e., low buoyant density, lymphoid DC-specific markers, and lymphostimulatory function. In fresh unfractionated cell preparations of normal cornea, no functional activity was detected. However, DC immunomagnetically purified from fresh preparations were functionally active. Injury to the corneal epithelium induced the migration of DC from the periphery to the central cornea; DC measured in this situation showed significantly increased functional activity. Finally, IL-1 beta and GM-CSF enhanced the functional activity of corneal DC. CONCLUSIONS: Corneal DC have lymphostimulatory capacity in situ, but they may be maintained in a state of latency by the suppressive influence exerted by neighboring cells. Injury to the corneal epithelium results in functional activation of the corneal DC, which may be caused by cytokines such as IL-1 beta or GM-CSF. Thus, corneal DC may be important in the immune regulation of the anterior segment.

Characterization and functional activity of dendritic cells from rat choroid.

Cell preparations from the posterior eye cup of the eye cultured for 2 days exhibited accessory activity for T-cell responses to a mitogenic treatment and stimulatory activity in a mixed leukocyte reaction (MLR), two functions characteristic of dendritic cells (DC). These activities both partitioned with cells having a low buoyant density, another characteristic of DC. Immunomagnetic separations with monoclonal antibodies against lymphoid dendritic cell surface antigens revealed that the accessory activity of the low-density cells was entirely associated with a small population of cells positively selected by these antibodies. Immunofluorescent staining with these same antibodies also revealed a small subpopulation of low-density cells having the morphology of DC. On cryostat sections of eye tissue the positively-stained cells were localized in the choroid and were not observed within the sclera or the retina. Based on these results we conclude that there are functional DC in the choroid, and we speculate that they may have a significant role in the inflammatory process during posterior uveitis.

High stimulatory activity of dendritic cells from diabetes-prone BioBreeding/Worcester rats exposed to macrophage-derived factors.

Dendritic cells (DC) present antigen and initiate T cell-mediated immune responses. To investigate the possible association of autoimmunity with DC function, we compared the accessory activity of splenic DC from Wistar/Furth (WF) and diabetes-prone (DP) BioBreeding (BB) rats. The latter develop autoimmune diabetes and thyroiditis. DC function was quantified in vitro by measuring T cell proliferation in mitogen-stimulated and mixed lymphocyte reactions. When purified without macrophage coculture, WF and DP DC displayed similar levels of accessory activity. In contrast, when purified by a method involving coculture with macrophages, DC from DP rats consistently displayed greater accessory activity. This finding could not be explained by morphological or phenotypic differences between DP and WF DC. In accessory activity assays performed after reciprocal DC cocultures with DP and WF macrophages, DP DC exhibited higher accessory activity irrespective of macrophage donor strain. We also compared the accessory activity of WF and DP DC cultured in the presence of conditioned medium and a mixture of IL-1 and GM-CSF. In all assays, DP DC exhibited higher accessory activity. In studies of (WF x DP) F1 hybrids, the high accessory activity of DP DC was observed to be heritable, and studies of WF and DP radiation chimeras indicated that the effect was an intrinsic property of the DP hematopoietic system. We conclude: (a) splenic DC from DP and WF rats possess similar basal levels of accessory potency; (b) after interaction with macrophages, DC of DP origin are capable of greater stimulatory activity than are WF DC; and (c) the mechanism responsible for this phenomenon involves differential responsiveness of DP and WF DC to macrophage-derived factors such as IL-1 and GM-CSF.

Conditioned medium from activated rat macrophages and the recombinant factors, IL-1 beta and GM-CSF, enhance the accessory activity of dendritic cells.

Low density lymph node cells (LD-LNC; 5% of total unfractionated LNC) contain 95% of the accessory activity required for responses of T lymphocytes to mitogens. Significantly greater responses to mitogens occur when T lymphocytes are added to LD-LNC that have been exposed overnight to silica, in comparison to responses occurring with LD-LNC incubated without silica. Conditioned medium (CM) from silica-treated LD-LNC is itself able to mediate enhanced responses; i.e., when LD-LNC are exposed overnight to CM alone and mitogen-treated T lymphocytes added the next day. The enhancing activity found in CM from LD-LNC exposed to silica is produced by macrophages; however, their low accessory activity is not enhanced by CM. In contrast, dendritic cells isolated from LD-LNC exposed to silica or to CM show significantly increased accessory activity, but dendritic cells do not produce the enhancing activity found in CM. CM lacks IL-2 activity and does not have any effect on the responses of untreated or mitogen-treated T lymphocytes alone. Thus, macrophages produce the enhancing activity and dendritic cells respond to it. Maximum enhancement of dendritic cell accessory activity requires overnight exposure to CM; once induced, accessory activity is not further modulated after continued incubation in the presence or absence of CM. LD-LNC, adherent peritoneal exudate cells, and adherent thioglycollate-induced peritoneal exudate cells produce enhancing activity after exposure to silica, LPS, and silica plus LPS. After gel filtration of a CM produced by silica plus LPS, enhancing activity shows a broad molecular weight distribution between 20 and 55 kD. IL-1 is present in CM and shows a more narrow molecular weight distribution that falls within the lower molecular weight range for enhancing activity. Silica treatment by itself produces CM containing little IL-1, but abundant enhancing activity; gel filtration of this CM shows that the distribution of enhancing activity is confined more narrowly to the higher molecular weight range, suggesting that IL-1 is one of several factors that enhances the accessory activity of dendritic cells. Recombinant human IL-1 beta does have enhancing activity, but of the other recombinant factors tested only mouse GM-CSF also has enhancing activity. Human IL-1 alpha, tumor necrosis factor alpha, IL-4, rat IL-3 and rat IFN-gamma, as well as L cell-conditioned medium containing M-CSF, lack enhancing activity.

Dendritic cell ontogeny.

Abundant evidence indicates that dendritic cells arise from the bone marrow. In vitro, precursors that differ phenotypically from mature dendritic cells divide several times to form functional dendritic cells. A soluble factor(s) produced in the supernatants of ConA-stimulated spleen cells enhances the production of dendritic cells. This factor(s) has not been fully characterized. Further maturation of dendritic cells occurs after they are released from the bone marrow; species differences exist. Interrelationships between various types of dendritic cells need to be elucidated.

Dendritic cells from rat lung are potent accessory cells.

Accessory cells are required for the induction of lymphocyte proliferation in response to mitogens or antigens. Rat pulmonary cells were tested for the presence of accessory activity in lymphocyte proliferation induced by sodium periodate. Buffer-perfused lungs were excised, minced, and enzymatically digested. The resulting pulmonary cells (PC) were separated into high density (HD-PC, 32 to 57%) and low density (LD-PC, 9 to 32%) fractions in a discontinuous density gradient of bovine plasma albumin (BPA). Both macrophages and dendritic cells were observed in the LD-PC by light microscopy. HD-PC, LD-PC, adherent LD-PC, nonadherent LD-PC, and a purified preparation of pulmonary dendritic cells (DC-P) were tested for accessory activity in the presence of periodate-treated, lymph-node-derived lymphocytes as responders. Most of the accessory activity was found in the LD-PC. Increasing numbers of LD-PC stimulated proliferation of responder lymphocytes in a linear, dose-dependent manner; higher numbers had a suppressive effect. Nonadherent LD-PC containing dendritic cells also produced a dose-dependent increase in periodate-induced lymphocyte proliferation, whereas adherent LD-PC, morphologically identified as macrophages, were suppressive. Removal of phagocytic macrophages from nonadherent LD-PC resulted in an eightfold increase in both the percent of DC-P present and the amount of accessory activity in the LD-PC. We conclude that pulmonary dendritic cells are potent accessory cells for periodate-induced lymphocyte proliferation.

The effect of silica treatment on accessory cell-dependent rat T lymphocyte proliferation.

Previous work has shown that purified rat macrophages lack both accessory activity for T lymphocyte responses to mitogens and stimulatory activity in a mixed leukocyte reaction, in marked contrast to the potent activity of dendritic cells. This study was designed to re-evaluate macrophages as accessory cells by treating various cell preparations with either silica or L-leucine methyl ester, which have been reported to be toxic to macrophages, and then determining the effect of the treated cells on responses to the mitogens, sodium periodate or concanavalin A. These studies indicated that treatment with L-leucine methyl ester failed to kill rat macrophages or dendritic cells, whereas silica was specifically toxic for rat macrophages. The studies therefore focused on silica. Co-culturing mitogen-treated lymph node cells with silica over a wide range of concentrations had no effect on responses. The same results were obtained if mitogen-treated lymphocytes were enriched with lymph node macrophages and dendritic cells and then co-cultured with silica. Preparations containing both macrophages and dendritic cells were incubated with silica for 24 h to ensure the death of virtually all macrophages; upon the addition of mitogen-treated lymphocytes, the macrophage-depleted accessory cells induced vigorous proliferative responses. Peritoneal exudate cells showed variable, but low accessory activity that increased after incubation with silica. Elimination of more than 90% of the macrophages from peritoneal exudate cells, as determined by staining for non-specific esterase, failed to eliminate this accessory activity. Taken together, these findings confirm and extend the conclusion that rat macrophages lack or have exceedingly low accessory activity.

Accessory cell dependent T lymphocyte proliferation: potent activity of dendritic cells.

The proliferative response of T lymphocytes to Concanavalin A (Con A) and the oxidative mitogens, sodium periodate (NaIO4) and neuraminidase plus galactose oxidase (NGO), requires the participation of dendritic cells (DC). High density cells (HDC) recovered from the fractionation of lymph node cells on a discontinuous gradient of bovine plasma albumin did not respond to NaIO4, but responded well above background levels to NGO or Con A. Addition of DC elevated these responses further. By an indirect panning technique, the HDC were exhaustively depleted of cells expressing Ia surface antigens. Ninety-nine percent of these HDC displayed T cell surface antigens. These Ia- T cells did not respond to any of the three mitogen treatments until DC were added, whereupon the proliferative responses were restored in a concentration-dependent manner, with maximal levels attained at a DC to T cell ratio of 1:200. The addition of a purified preparation of IL2 to untreated or mitogen-treated Ia- T cells increased the proliferative responses slightly and was unable to substitute for the potent activity of dendritic cells. Only the addition of DC was able to stimulate mitogen-induced proliferation to maximal levels. The limiting factor in these responses was the number of dendritic cells, which controlled the induction of both the release of IL2 and responsiveness to IL2 for the oxidative mitogens and Con A. Thus, DC function as potent accessory cells for each of the three mitogens.

Differentiation of dendritic cells in cultures of rat bone marrow cells.

Although dendritic cells (DC) originate from bone marrow, they were not observed in fresh preparations of bone marrow cells (BMC). Likewise, accessory activity was barely measurable in a sensitive assay for this potent function of DC. However, both DC and accessory activity developed when BMC were cultured for 5 d. Based on fractionation before culture, nearly all of the accessory activity could be attributed to only 5% of the total BMC recovered in a low-density (LD) fraction. The LD-DC precursors differed from mature DC in a number of important respects. Removal of Ia+ cells from the LD fraction by panning did not decrease the production of DC when the nonadherent cells were cultured. Thus, the cell from which the DC is derived does not express or minimally expresses Ia antigens, in contrast to the strongly Ia+ DC that is produced in bone marrow cultures. Irradiation of LD cells before culture prevented the development of DC. When irradiation was delayed by daily intervals, progressive increases in the number of DC resulted, up to the fifth day. These findings, together with preliminary autoradiographic data, indicate that cell division has occurred, in contrast to the DC, which does not divide. We conclude that bone marrow-derived DC arise in culture from the division of LD, Ia- precursors.

Canine dendritic cells from peripheral blood and lymph nodes.

Canine dendritic cells were prepared from peripheral blood or lymph nodes using a series of steps including fractionation on bovine plasma albumin (BPA), irradiation with 4000 R, incubation for 16-18 hours, and refractionation on BPA. Dendritic cells were recovered in the low density (LD) fraction containing approximately 0.6% of the unfractionated cells. Measured by the incorporation of 3H-thymidine, the response of the high density (HD) cells to neuraminidase-galactose oxidase (NGO) was lower than that of the unfractionated lymph node cells (LNC) but increased in a concentration dependent manner after the addition of a population of cells enriched for dendritic cells (30-70% by morphologic criteria). Cooperation between HD- and LD- cells was not restricted to identity of the major histocompatibility complex. Canine dendritic cells also displayed stimulatory activity higher than unfractionated peripheral blood mononuclear cells (PBMC) in a one way mixed leukocyte culture (MLC). Canine dendritic cells were nonadherent to plastic, were of low density, and remained viable and functional after irradiation. For the first time, canine dendritic cells have been identified in peripheral blood and lymph nodes and have been shown to act as accessory cells in the response of lymphocytes to NGO and as stimulator cells in a MLC.

Dendritic cell identification in head and neck lymphoid tissue. Newly recognized cells control T-lymphocyte functions.

The physiologic measurements of a subpopulation of mononuclear cells derived from head and neck lymphoid tissues are similar to those of dendritic cells are described. Dendritic cells are a subpopulation of bone marrow-derived leukocytes that were originally identified in rodents and now described in man as having central control of T-lymphocyte functions. We describe a technique for the enrichment of dendritic cells obtained from tonsils utilizing a bovine serum albumin (BSA) gradient and note that they have the light and electron microscopic appearance of dendritic cells. The measured oxidative mitogenic response and interferon-gamma production in complete leukocyte cultures was compared with BSA gradient-separated preparations. The denser cells, comprised mostly of normal appearing lymphocytes, would not undergo a mitogenic response nor produce normal amounts of interferon when stimulated unless the dendritic cell-rich, less-dense fraction, was added back. The dendritic cells derived from tonsils seem to behave as a potent accessory cell for these T-lymphocyte-associated functions.

Removal of rat dendritic cells from single cell suspensions by passage through columns of Sephadex G-10.

Rat T lymphocytes require dendritic cells as accessory cells in order to respond to the mitogen sodium periodate. Passage of a lymph node cell suspension through a column of Sephadex G-10 reduced the mitogenic response by greater than 90%, despite a cell recovery of 75%. An even greater reduction in the response occurred after a second passage over Sephadex G-10; addition of purified dendritic cells restored the response. Lymph node cell suspensions and preparations enriched in dendritic cells were nearly depleted of these cells by passage over Sephadex G-10. After passage of lymph node cells over Sephadex G-10, a limited number of retained cells could be recovered; these included both dendritic cells and macrophages. Enumeration of macrophages in the various passed and retained fractions by non-specific esterase staining of smears confirmed that macrophages were effectively removed from lymph node cell suspensions by repeated passage over Sephadex G-10. Cell preparations that pass through Sephadex G-10 are therefore depleted of both dendritic cells and macrophages.

The use of radioactive cysteine methyl ester for labeling glycosylated molecules oxidized by periodate or neuraminidase plus galactose oxidase.

Treatment of rat lymph node cells with periodate or neuraminidase plus galactose oxidase initiates blast transformation and cell division of T lymphocytes. Either treatment introduces aldehyde functions onto glycosylated molecules of the plasma membrane. Reduction of the aldehydes with borohydride leads to a concentration-dependent inhibition of the mitogenic response. Cysteine methyl ester (Cys(Me], which can form a stable thiazolidine adduct with aldehydes, also inhibits mitogenesis in a concentration-dependent manner. Maximum inhibition is achieved at Cys(Me) concentrations about 10-fold lower than those required for borohydride (0.4 and 5 mM, respectively). [35S]Cys(Me) has been synthesized and compared with [3H]borohydride as a labeling reagent for molecules on the plasma membrane oxidized by periodate or neuraminidase plus galactose oxidase. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of labeled whole cell lysates or of crude membrane fractions prepared from labeled cells revealed that the same oxidized molecules are specifically labeled with both reagents. Homogenates of cells labeled with either radioactive reagent were fractionated by differential and isopycnic centrifugation. The fractions were analyzed for radioactivity and for a number of marker constituents localized in various subcellular organelles. Following treatment with either reagent, the radioactivity that was covalently incorporated into macromolecules was primarily associated with sedimentable components that distributed among the fractions like plasma membrane markers. When compared with [3H]borohydride, Cys(Me) offers several advantages as a surface labeling reagent for glycosylated plasma membrane molecules, chiefly the possibility of preparing reagents labeled with isotopes other than tritium, including those like 35S, which are much stronger radioactive emitters.

Characterization of rat lymphocyte cell membranes by analytical isopycnic centrifugation.

Lymph node cell homogenates were fractionated by differential or isopycnic centrifugation and the fractions analyzed for biochemical markers with particular focus on plasma membrane constituents. Markers for the nucleus (DNA), mitochondria (cytochrome oxidase), and lysosomes (acid hydrolases) showed the expected distributions which were different from those of membrane-bound enzymes. 5'-Nucleotidase, alkaline phosphodiesterase, gamma-glutamyltranspeptidase, and cholesterol were membrane-bound and distributed identically after isopycnic centrifugation with peaks at 1.15. The distributions of the enzymes were all shifted to higher densities by digitonin treatment, confirming their association with plasma membrane-derived elements. The distribution of galactosyltransferase (ovalbumin acceptor), largely overlapped those of plasma membrane markers but it was only slightly shifted by digitonin, suggesting its localization in Golgi apparatus. The distribution of mannosyltransferase (dolichyl phosphate acceptor) also overlapped those of plasma membrane and Golgi markers but it was centered at higher density (1.18) and was unaffected by digitonin. It is a useful marker for endoplasmic reticulum. 50% of the activity was in low speed "nuclear" sediments where it was associated with the nuclear membrane. A number of other putative and previously used markers for the endoplasmic reticulum of lymphocytes were shown not to be localized in these membranes. In particular, NADH-cytochrome c reductase was only partly associated with the endoplasmic reticulum (56%) and the remainder of the activity was in mitochondria (44%). The results show the heterogeneity in equilibrium density of plasma membrane vesicles and the considerable overlap of their distribution with those of other cellular membranes; they should provide a basis for the more rational design of preparative procedures for the lymphocyte plasma membrane.