Model Cell Differentiation from a Human Cell Line Progenitor To

Evidence for Distinct Intracellular Signaling Pathways in CD34 + Progenitor to Dendritic Cell Differentiation from a Human Cell Line Model This information is current as of September 28, 2021. Daniel C. St. Louis, Juliana B. Woodcock, Guido Fransozo, Patrick J. Blair, Louise M. Carlson, Maria Murillo, Mark R. Wells, Amanda J. Williams, Douglas S. Smoot, Sumesh Kaushal, Janelle L. Grimes, David M. Harlan, John P. Chute, Carl H. June, Ulrich Siebenlist and Kelvin P. Lee Downloaded from J Immunol 1999; 162:3237-3248; ; http://www.jimmunol.org/content/162/6/3237

References This article cites 75 articles, 50 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/162/6/3237.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists by guest on September 28, 2021

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts Errata An erratum has been published regarding this article. Please see next page or:

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Evidence for Distinct Intracellular Signaling Pathways in CD34؉ Progenitor to Dendritic Cell Differentiation from a Human Cell Line Model1

Daniel C. St. Louis,*† Juliana B. Woodcock,* Guido Fransozo,‡ Patrick J. Blair,* Louise M. Carlson,‡ Maria Murillo,* Mark R. Wells,* Amanda J. Williams,* Douglas S. Smoot,* Sumesh Kaushal,*† Janelle L. Grimes,* David M. Harlan,*§ John P. Chute,*§ Carl H. June,*§ Ulrich Siebenlist,‡ and Kelvin P. Lee2*§

Intracellular signals that mediate differentiation of pluripotent hemopoietic progenitors to dendritic cells (DC) are largely unde-

fined. We have previously shown that protein kinase C (PKC) activation (with phorbol ester (PMA) alone) specifically induces Downloaded from differentiation of primary human CD34؉ hemopoietic progenitor cells (HPC) to mature DC. We now find that cytokine-driven granulocyte-macrophage CSF and TNF-␣) CD34؉ HPC3DC differentiation is preferentially blocked by inhibitors of PKC) activation. To further identify intracellular signals and downstream events important in CD34؉ HPC3DC differentiation we have characterized a human leukemic cell line model of this process. The CD34؉ myelomonocytic cell line KG1 differentiates into dendritic-like cells in response to granulocyte-macrophage CSF plus TNF-␣, or PMA (with or without the calcium ionophore ␣ ionomycin, or TNF- ), with different stimuli mediating different aspects of the process. Phenotypic DC characteristics of KG1 http://www.jimmunol.org/ /dendritic-like cells include morphology (loosely adherent cells with long neurite processes), MHC I؉/MHC IIbright/CD83؉/CD86؉ -CD14؊ surface Ag expression, and RelB and DC-CK1 gene expression. Functional DC characteristics include fluid phase mac romolecule uptake (FITC-dextran) and activation of resting T cells. Comparison of KG1 to the PMA-unresponsive subline KG1a reveals differences in expression of TNF receptors 1 and 2; PKC isoforms ␣, ␤I, ␤II, and ␮; and RelB, suggesting that these components/pathways are important for DC differentiation. Together, these findings demonstrate that cytokine or phorbol ester stimulation of KG1 is a model of human CD34؉ HPC to DC differentiation and suggest that specific intracellular signaling pathways mediate specific events in DC lineage commitment. The Journal of Immunology, 1999, 162: 3237–3248.

ow multipotential hemopoietic stem cells become ma- B cells (1). In humans, dendritic cells are a heterogeneous popu- by guest on September 28, 2021 ture blood cells is a paradigm of cellular differentiation lation shown in vitro to arise from multiple distinct hemopoietic and lineage commitment. Complex hemopoiesis under- progenitor cells (HPC) along distinct differentiation pathways. H ϩ ϩ lies the generation of each component of the immune system, in- These progenitors include CD14 monocytes (2–5) and CD34 cluding T, B, and professional APCs (such as monocyte/macro- bone marrow cells (in humans a heterogeneous population of mul- 3 phages and dendritic cells (DC) ). Evidence now suggests that DC tipotential progenitors of which the true pluripotent stem cell is a play the central and critical role in both initiating and controlling subset). CD34ϩCD10ϩ HPC give rise to the lymphoid DC lineage the primary immune response via cognate interactions with T and (6), whereas CD34ϩCD86ϩ HPC are progenitors for myeloid DC and monocyte/macrophages (7). *Immune Cell Biology Program, Naval Medical Research Institute, Bethesda, MD Broadly characterized from a cellular standpoint, there are two 20889; †The Henry M. Jackson Foundation for the Advancement of Military Medi- stages of DC differentiation (8). The first stage involves differen- cine, U.S. Military HIV Research Program, Bethesda, MD 20889; ‡Laboratory of Immunoregulation, National Institute of Allergy and Infectious Disease, National In- tiation of multipotential HPC to immature (or unactivated) DC, stitutes of Health, Bethesda, MD 20892; and §Department of Internal Medicine, Uni- cells with high capacity for Ag uptake but relatively poor ability to formed Services University of the Health Sciences, Bethesda, MD 20889 activate T cells. This differentiation is induced in vitro by exoge- Received for publication August 11, 1998. Accepted for publication December nous cytokines (GM-CSF and TNF-␣ (with or without other cy- 11, 1998. tokines) for CD34ϩ HPC (9), and GM-CSF and IL-4 for mono- The costs of publication of this article were defrayed in part by the payment of page ϩ charges. This article must therefore be hereby marked advertisement in accordance cytes (2–5)), CD40 receptor cross-linking (CD34 HPC) (10), or with 18 U.S.C. Section 1734 solely to indicate this fact. calcium ionophore alone (monocytes) (11). The second stage in- 1 This work was supported by the Naval Medical Research and Development Com- volves maturation of immature to mature DC, cells that have de- mand, Research Task 62233N.MM33C30.1402.HD402.1412. Views presented in this paper are those of the authors; no endorsement by the Department of Navy, the creased Ag uptake capability but are much more potent in activat- Department of the Army, or the Department of Defense has been given or should be ing T cells. An alternative viewpoint is that DC maturation is inferred. actually activation of resting DC. In vivo this second stage is prob- 2 Address correspondence and reprint requests to Dr. Kelvin P. Lee, Immune Cell ably triggered by “danger signals,” which may include live bacte- Biology Program, Immune Suppression Branch, Building 17, Room 214, Naval Med- ical Research Institute, 8901 Wisconsin Ave., Bethesda, MD 20889-5067. E-mail ria and components (LPS, DNA), viral infection, and inflammatory address: [email protected] cytokines (reviewed in Ref. 1). In vitro, maturation/activation can 3 Abbreviations used in this paper: DC, dendritic cell; HPC, hemopoietic progenitor be induced by a number of stimuli, including cytokines (TNF-␣ cell; CD34ϩ HPC, CD34ϩ hemopoietic progenitor cell; GM-CSF, granulocyte-macrophage CSF; PKC, protein kinase C; DLC, dendritic-like cell; IMDM, Iscove’s mod- with or without other cytokines, reviewed in Ref. 12), monocyte- ified Dulbecco’s medium; SCF, stem cell factor; bis, bisindolylmaleimide I. conditioned medium (13), or CD40 receptor cross-linking (10).

We have previously reported that the phorbol ester PMA alone although DC-specific markers were not available at that time. The induces differentiation of human CD34ϩ HPC into mature and KG1a subline of KG1 was indifferent to the effects of PMA (and fully functional DC, suggesting that direct activation of protein cytokines) and is characterized as less differentiated than KG1 (29, 30, kinase C (PKC) by itself is sufficient to trigger this lineage com- 32). KG1 and KG1a have qualitative differences in PKC activation mitment (14). PMA-driven CD34ϩ HPC3DC differentiation is and substrate phosphorylation (30, 33, 34). We now report that phor- specific for DC (no other lineages are generated) and does not bol ester or cytokine (GM-CSF and TNF-␣) induces KG1 to differ- involve proliferation (differentiation is complete within 7 days), entiate into dendritic-like cells (DLC) based on morphology, surface ϩ although about 50% of the input number of CD34 HPC are lost Ag phenotype, function, and gene expression. In contrast, KG1a do (in part via apoptosis). All these characteristics distinguish PMA not differentiate. Characterization of stimuli required to induce DLC from receptor-mediated differentiation signals (cytokines, CD40 differentiation as well as comparison to KG1a provide evidence that receptor cross-linking). Since exogenous stimuli (including distinct intracellular signaling pathways are involved and mediate spe- TNF-␣, IL-1␤, IL-4, and CD40) (15–19) used to induce cific components of differentiation. CD34ϩHPC3DC differentiation also activate PKC as part of their downstream signaling cascade, it is likely that PMA is activating the component of this cascade that specifically initiates DC differ- Materials and Methods entiation. Consistent with this, we now report that inhibition of Cells and culture PKC activation suppresses generation of DC by GM-CSF and ϩ TNF-␣ stimulation of CD34ϩ HPC. CD34 HPC were isolated from organ donor bone marrow by immuno-

magnetic selection as previously described (35). Purified cells were typi- Downloaded from Aside from a potential role for PKC activation, relatively little cally Ͼ95% CD34ϩ as determined by FACS analysis with a second non- is known about what intracellular signaling pathways (and the ge- cross-blocked anti-CD34 mAb. For generation of DC, CD34ϩ HPC were netic programs they initiate) are involved in HPC3DC differen- cultured for 14 days at 5 ϫ 104 cells/ml in complete culture medium tiation. Signaling via intracellular calcium flux alone (by calcium (IMDM, Life Technologies, Grand Island, NY) supplemented with 10% ionophore) induces monocyte3DC differentiation (11), although heat-inactivated FCS (HyClone, Logan UT), 100 mM L-glutamine, and 100 ϩ U/ml penicillin/streptomycin solution (Life Technologies)) with GM-CSF we have not observed this for CD34 HPC (D.C.S.L and K.P.L, (200 U/ml) and TNF-␣ (2.5 ng/ml; all from R&D Systems, Minneapolis, unpublished observations). The downstream components involved MN). Where indicated, staurosporine (Calbiochem, San Diego, CA) was http://www.jimmunol.org/ in this calcium signaling are uncharacterized. Another signaling added at 0.1 ng/ml, a dose sufficient to block PMA-induced proliferation of pathway involves the lipid second messenger ceramide, which me- human T cells. At the end of the culture period, nonadherent cells and diates down-regulation of Ag uptake during cytokine-driven mat- adherent cells were resuspended (3 mM EDTA), washed, and concentrated by centrifugation. uration/activation of DC (20). Downstream of these signaling path- KG1 and KG1a cells were obtained from the American Type Culture ways, the transcription factor RelB (a member of the NF␬B Collection (Manassas, VA) and cultured in IMDM/20% FCS/100 mM 4 family) appears to play an important role in DC differentiation L-glutamine/100 U penicillin/streptomycin at 5 ϫ 10 cells/ml. Cultures based on apparent function (21–24) and the phenotype of RelB were stimulated alone or in combination with PMA (10 ng/ml; Sigma, St. ␣ knockout mice (25, 26). RelB expression is up-regulated in vitro Louis, MO), GM-CSF (200 U/ml), TNF- (10 ng/ml), ionomycin (100

ng/ml; Calbiochem). The four-cytokine combination of GM-CSF (2 ng/ by guest on September 28, 2021 during cytokine- or PMA-driven DC differentiation (2, 14). What ml), IL-3 (5 ng/ml), IL-6 (5 ng/ml), stem cell factor (SCF; 120 ng/ml) has genes RelB specifically regulates during DC differentiation is cur- previously been found to induce robust proliferation of primary CD34ϩ rently unknown. HPC (35). Where indicated, bisindolylmaleimide I (bis; Calbiochem) was Even using agents that directly activate signal transduction path- used at 5 ␮M. ways, a number of obstacles makes study of the intracellular and genetic events involved in DC differentiation difficult. The rarity of Flow cytometric analysis and mAbs CD34ϩ HPC (0.1–1% of bone marrow mononuclear cells) makes sufficient isolation for larger scale studies laborious. In addition, Adherent and loosely adherent cells were harvested with 3 mM EDTA, CD34ϩ HPC are a heterogeneous population of progenitors al- washed twice, and resuspended in staining medium (PBS, 5% FCS, 2% BSA, and 0.1% sodium azide). Phenotypic analysis of cells (2 ϫ 105) was ready committed to different lineages (for example, see Ref. 6), performed by flow cytometry using saturating concentrations of the fol- which makes analysis of bulk starting populations problematic. lowing mAbs: CD1a (clone SK9), CD2 (S5.2), CD4 (SK3), CD13 (L138), Receptor-mediated CD34ϩ HPC3DC differentiation involves cell CD14 (MOP9), CD34 (8G12), CD80 (L307.4), and CD154 (89-76; all proliferation, which complicates identification of differentiation- from Becton Dickinson, San Jose, CA); CD10 (HI10A), CD33 (WM53), specific processes. These stimuli also generate mixed populations CD40 (5C3), CD45RA (HI100), CD86 (IT2.2), and CD90 (5E10; all from PharMingen, San Diego, CA); CD83 (HB15, Immunotech, Westbrook, of cells (DC, monocytes, neutrophils) that require additional pu- ME); MHC class I (H58A) and MHC class II (H42A, both from Veterinary rification of DC before definitive analysis (reviewed in Ref. 12). Medical Research Development, Pullman, WA); c-fms (3-4A4), TNF re- Finally, CD34ϩ HPC have traditionally been difficult to transfect/ ceptor 1 (N-20), and TNF receptor 2 (C-20; all from Santa Cruz Biotech- transduce (27), making genetic manipulation involving these ap- nology, Santa Cruz, CA). Appropriate conjugated isotype-matched Abs proaches less effective. were used as controls. Ten thousand cells were analyzed on a Coulter Elite flow cytometer (Coulter, Hialeah, FL) through a viable cell gate as deter- Many of these problems could be circumvented in a cell line model mined by forward and right angle light scatter parameters to exclude sub- of DC differentiation. Czerniecki et al. have suggested that the pro- cellular particles. myelocytic HL-60 cell line differentiates into DC with calcium iono- Apoptotic cells were identified by annexin V staining as previously phore, thus representing a potential model of monocyte3DC differ- described (36). Briefly, cells were washed twice with cold PBS and then ϩ resuspended in a binding buffer (from the manufacturer) containing HEPES entiation (11). It has recently been reported that committed CD34 ϫ 6 with 2.5 mM CaCl2 at a concentration of 1 10 /ml. One hundred mi- precursors for myeloid DC and macrophages are CD86 positive (7). croliters of this suspension was reacted with 10 ␮l of annexin V-FITC (10 ϩ We have found that the CD34 human myeloblastic cell line KG1 ␮g/ml (R&D Systems)) and 10 ␮l of propidium iodide reagent (50 ␮g/ml (28) also expresses CD86. KG1 is a cytokine-responsive human in PBS). The mixture was gently vortexed and then incubated for 15 min ϩ ␮ CD34 myelomonocytic cell line derived from a patient with eryth- at room temperature in the dark. Following incubation, 400 l of the binding buffer (1ϫ concentration) was added to each tube. Analysis by FACS roleukemia. KG1 was originally described to differentiate to the was conducted within1hofassay completion for optimal results. Controls monocyte/macrophage lineage (based on morphologic and histo- of unreacted cells, propidium iodide-stained cells, and annexin-V-FITC- chemical assays) when stimulated with phorbol ester (29–31), stained cells were run first to optimize settings. The Journal of Immunology 3239

FITC-dextran uptake Results ϩ Analysis of micropinocytosis by FITC-dextran uptake has been previously Inhibition of PKC activation blocks cytokine-driven CD34 described (37). Activated KG1 and KG1a cells were resuspended at 106/ml HPC3DC differentiation in IMDM/20% FCS. FITC-dextran (Sigma) was added at a final concen- We have previously shown that PMA activation of PKC induces tration of 1 mg/ml, and the mixture was incubated at either 4 or 37°C for ϩ 30 min. The cells were washed four times with cold PBS containing 1% human CD34 HPC3DC differentiation, which could be blocked FCS and analyzed by flow cytometry. by the PKC inhibitor staurosporine. To assess whether cytokine- ϩ 3 T cell activation induced CD34 HPC DC differentiation also involves PKC intracellular signaling, we cultured CD34ϩ HPC in GM-CSF and Human PBMC were obtained by leukapheresis from normal healthy adult TNF-␣ for 14 days with or without 0.2 nM staurosporine, a dose donors. T cells were purified by negative selection as previously described (38) and resuspended in RPMI 1640 (Life Technologies) supplemented 3.5-fold less than the IC50 of PKC (0.7 nM; Calbiochem Technical with 10% heat-inactivated FCS, 2 mM L-glutamine, 100 U/ml penicillin, Report) but at least 10- to 200-fold lower than the IC50 for protein 100 mg/ml streptomycin, and 20 mM HEPES. T cells (1 ϫ 105/well) were kinase A and tyrosine kinases (40). Cell viability at 14 days was cultured in the absence (in medium) or the presence of the indicated num- Ͼ 137 90% in both culture conditions, but there was a greater expan- ber of gamma-irradiated (3000 rad Cs), in vitro-generated DC or DLC. ϫ ϫ Cultures were incubated for 3 days (DC) or 5 days (DLC) at 37°C in a sion of cell number without staurosporine (3 input number vs 1 with staurosporine). As can be seen in Fig. 1A, CD34ϩ HPC cul- humidified 5% CO2 in air atmosphere. T cell proliferation induced by freshly isolated allogeneic PBMC was used as a reference. Proliferation tured in cytokine alone developed characteristic DC morphology was assessed after the addition of 0.5 ␮Ci/well [methyl-3H]thymidine (New (loosely adherent large cells with irregular shape, prominent den- England Nuclear, Boston, MA) for the final 24 h of culture. Cells were

dritic processes, and hair-like cytoplasmic projections). In contrast, Downloaded from harvested using a 96-well cell harvester, and [methyl-3H]thymidine incorporation was measured using a Beta Plate scintillation counting system cultures containing staurosporine lacked cells with neurite pro- (Pharmacia/LKB, Gaithersburg, MD). All determinations were performed cesses and had greatly reduced numbers with hair-like cytoplasmic in triplicate and expressed as the mean counts per minute Ϯ 1 SD. projections. RT-PCR and Northern blot analysis Staurosporine treatment also resulted in the loss of cells express- ing characteristic DC surface markers, CD83 (DC lineage specif- Total RNA was isolated from cultured cells by the RNA STAT-60 (Tel- ic), MHC class II (bright), and the costimulatory molecules CD80 Test) according to manufacturer’s recommendations. Ten micrograms of http://www.jimmunol.org/ purified RNA was incubated at 37°C for 15 min with 100 U of RNase-free and CD86 (Fig. 1B). Of particular significance were the loss of DNase in a 50-␮l reaction, followed by phenol/chloroform extraction and CD83- and CD80-positive cells, as these Ags are not expressed on ethanol precipitation. Samples of 4 ␮g of DNase-treated RNA were reverse monocytes (resting, in the case of CD80) (41, 42). In contrast to transcribed to cDNA by random primer extension in the presence of 0.1 ϩ 32 CD83 and CD80, there was only a modest decrease in CD14 ␮Ci of [␣- P]dCTP using Superscript-II Reverse Transcriptase (Life ϩ Technologies) following the manufacturer’s instructions. The RT reaction (monocyte marker) cells and no loss of CD33 (myeloid marker) was terminated by heating at 70°C for 5 min, and radiolabeled cDNA was cells. Cells from both culture conditions were CD34 negative and then purified on Spin-50 columns (International Mould Engineering, Oden- CD13 bright, demonstrating that staurosporine did not nonspecifi- ton, MD) by centrifugation at 3500 rpm for 3 min. To normalize the input cally block all differentiation. These results suggest that stauro- in all cDNA samples, equal amounts of cDNA counts were loaded in each ϩ sporine preferentially inhibited generation of CD83 DC com- by guest on September 28, 2021 PCR reaction containing 0.5 ␮M gene-specific primer pairs. Oligonucleo- ϩ tide primers specific for human DC-CK1 (sense, AGT CCC ATC TGC pared with CD14 monocytes. TAT GCC CAG; antisense, TAC GAA GAG TGG AAG GGA AAG) and Given that DC are the most effective APC in activating T cells, the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase the decreased numbers of DC in the staurosporine-treated cultures (GAPDH; sense, ATG GGG AAG GTG AAG GTC GGA GTC AAC should be reflected in a similar decrease in functional potency. GGA; antisense, AGG GGG CAG AGA TGA TGA CCC TTT TGG CTC) ϩ ␣ were used. PCR was performed at 94°C for 1 min, 55°C for 1 min, and CD34 HPC cultured in GM-CSF and TNF- with staurosporine 72°C for 2 min for 30 cycles followed by a final extension time of 10 min were significantly less effective in inducing allogeneic T cell pro- at 72°C. The amplified products (15 ␮l) were then incubated at 94°C for 5 liferation than those cultured in cytokine alone, particularly at a min with 1 ϫ 105 cpm of purified [␥-32P]dATP-labeled gene-specific low APC/T cell ratio (Fig. 1C). However, both were still superior probes followed by an incubation at 55°C for 10 min. Hybridized PCR to an unselected population of PBMC. products were then resolved on 6% PAGE followed by quantitation on a Together these data suggest that inhibition of PKC activation PhosphorImager scanner using ImageQuant software (Molecular Dynam- ϩ ics, Sunnyvale, CA). preferentially suppressed generation of DC from CD34 HPC by Northern blot analysis was performed as previously described (39). GM-CSF and TNF-␣, findings consistent with the differentiating Briefly, total RNA from unstimulated and stimulated KG1 and KG1a cells effects of phorbol ester. However, the scarcity, heterogeneity, and was isolated at the times indicated and equalized by serial dilution and ϩ ethidium bromide visualization. Equal amounts of RNA were separated on difficulty of manipulating primary CD34 HPC made further anal- a formaldehyde/agarose gel, transferred to nylon membranes, and hybrid- ysis of the relevant signaling and genetic events difficult. To over- ized against a human RelB cDNA probe. The same blots were then probed come these obstacles we sought to establish a human cell line for actin to assess equal loading and RNA integrity. model of CD34ϩ HPC3DC differentiation. Although a heteroge- Western blot analysis neous lineage, DC share a constellation of defining characteristics, including morphology, lineage markers, function, and gene expres- Cell lysates were made from KG1 and KG1a cells cultured under the conditions and times indicated. Samples containing equal amounts of protein sion, that can be used to identify potential DLC. were separated by SDS-PAGE (4% stacking/7.5% resolving); electroblotted to nitrocellulose; probed with Abs against RelB and PKC␣,-␤I, -␤II, KG1 develop characteristic DC phenotype (morphology, surface and -␮ (all from Santa Cruz Biotechnology, Santa Cruz, CA); and visual- Ag expression) in response to cytokines or PMA ized by chemiluminescent detection (ECL, Amersham, Aylesbury, U.K.). We first examined the PMA-responsive CD34ϩ myeloblast cell Electromobility shift assays line KG1. KG1 cells in medium alone were typically round non- Cell lysates were made from KG1 and KG1a cells cultured under the con- adherent cells (Fig. 2). When cultured in GM-CSF and TNF-␣ ditions and times indicated. Samples containing equal amounts of protein (with or without IL-4), PMA or PMA plus TNF-␣ (and PMA plus were incubated with a 32P-labeled oligonucleotide primer containing multiple NF␬B binding sites with or without anti-RelB Ab or anti-NF␬B (p50). ionomycin; not shown), KG1 became loosely adherent and a sub- Samples were then separated on a 4% polyacrylamide gel and exposed to set developed long neurite processes and hair-like cytoplasmic film. projections (best seen by Wright staining; not shown), morphology 3240 A CELL LINE MODEL OF HUMAN DENDRITIC CELL DIFFERENTIATION Downloaded from http://www.jimmunol.org/ by guest on September 28, 2021

FIGURE 1. Primary CD34ϩ HPC cultured for 14 days in GM-CSF (200 U/ml) and TNF-␣ (2.5 ng/ml) with or without staurosporine (0.2 nM). A, Morphology. Photomicrograph of the cultures. Black dots are magnetic beads used in the initial CD34ϩ purification. Arrows point to the characteristic neurite processes. B, Surface Ag phenotype. CD34ϩ HPC were cultured as described above, and adherent and nonadherent cells were analyzed by flow cytometry with the Abs indicated (solid line) vs isotype control (dotted line). The x-axis is log fluorescence, and the y-axis is cell number. C, Induction of allogeneic T cell proliferation. CD34ϩ cells cultured with or without staurosporine as described above were harvested on day 14, washed three times in PBS, plated in triplicate in the numbers indicated, and gamma irradiated (3000 rad 137Cs). Allogeneic irradiated PBMC were used for comparison. Purified CD4ϩ T cells (1 ϫ 105) were added to each well. T cells were purified from leukapheresis collections of healthy donors by negative selection (38) and were free of accessory cells as demonstrated by Con A unresponsiveness. After 54 h of culture, cells were pulsed with 0.5 ␮Ci/well [methyl- 3H]thymidine before being harvested at 72 h. [3H]TdR incorporation was measured using a Beta Plate scintillation counting system. reported to be characteristic of DC (12). Morphologic changes (B7-1) and CD86 (B7-2), and the DC lineage marker CD83 (re- began within 30 min of stimulation, were fully manifested within viewed in Ref. 12). Unstimulated KG1 cells expressed MHC I, 48 h, and were stable over at least 2 wk in culture. Development MHC II (bright), the myeloid markers CD13 and CD33, CD34, of the dendrite/neurite morphology was progressively more prev- CD45RA, and CD86 (dim) but not CD40, CD80, the DC marker alent and pronounced with GM-CSF plus TNF-␣3PMA3PMA CD83, or the monocyte marker CD14 (Table I). KG1a had lower plus TNF-␣ (and PMA plus ionomycin). In contrast to these stim- expression of MHC class II and no CD86 staining. We also found uli, there was no effect of individual cytokines (GM-CSF, IL-4, or that KG1 and KG1a differed markedly in expression of TNF-␣ TNF-␣), a cytokine combination that does not induce DC differ- receptors 1 and 2 (Fig. 3), which may in part underlie the differentiation (GM-CSF, IL-3, IL-6, and SCF), ionomycin alone, or ences in TNF-␣ responsiveness between the two cell lines. ionomycin plus TNF-␣ (data not shown). While KG1a also be- KG1 stimulated with GM-CSF plus TNF-␣, PMA, PMA plus came adherent when similarly stimulated, they did not change TNF-␣, or PMA plus ionomycin up-regulated CD83 expression morphology. with the same temporal kinetics as the morphology changes (Fig. Mature DC are typically characterized by expression of MHC 4). The level of CD83 expression has been correlated with the class I, MHC class II (high), the costimulatory molecules CD80 degree of DC maturation (43), and CD83 up-regulation in KG1 The Journal of Immunology 3241

FIGURE 2. Morphology of KG1 stimulated with cytokines and/or PMA. Photomicrographs of day 7 cultures of KG1 in medium alone, GM-CSF (200 U/ml) plus TNF-␣ (2.5 ng/ml), PMA (10 ng/ml), or PMA plus TNF-␣ (10 ng/ml). Arrows point to neurite processes. KG1 in medium alone were nonadherent; in all other conditions they were adherent. Downloaded from

was typically highest with PMA and TNF-␣ or PMA and iono- cin, but did up-regulate (slight) CD83 in response to PMA and http://www.jimmunol.org/ mycin. PMA also up-regulated CD40 expression (not shown) and TNF-␣ (not shown). modestly down-regulated MHC II, which was reversed by addition 3 of TNF-␣ or ionomycin. Up-regulation of CD86 was most pro- Signals that induce KG1 DLC differentiation also suppress nounced with addition of ionomycin, and ionomycin alone could proliferation and trigger programmed cell death up-regulate CD86 without inducing CD83 expression (data not In primary CD34ϩ HPC we found the two hallmarks of PMA- shown). PMA plus ionomycin also induced expression of CD80, driven DC differentiation are the complete inhibition of prolifera- although typically at low levels (not shown). The combination of tion (even to exogenous cytokines) and induction of apoptosis ␣ PMA, ionomycin, and TNF- was no different from PMA plus (14). Proliferation is also absent in monocyte3DC differentiation by guest on September 28, 2021 TNF-␣ or PMA plus ionomycin alone, suggesting that TNF-␣ and driven by GM-CSF, IL-4, and TNF-␣ (4) or calcium ionophore ionomycin deliver the same signal. Individual cytokines (GM- (11). As shown in Fig. 5, KG1 proliferation was suppressed by CSF, IL-4, TNF-␣) or the cytokine combination GM-CSF, IL-3, factors that induce DC/DLC differentiation (PMA (consistent with IL-6, and SCF did not up-regulate CD83 or CD86 expression on the findings presented in Ref. 44), TNF-␣, and most significantly KG1. None of the stimuli tested induced expression of CD1a or the by PMA plus TNF-␣), but not by a cytokine combination (GM- monocyte marker CD14. In contrast to KG1, KG1a did not re- CSF, IL-3, IL-6, and SCF) that does not generate DC. Baseline spond to GM-CSF plus TNF-␣ or PMA with or without ionomy- proliferation of KG1a was 2.6-fold higher than that of KG1 and was unaffected by these stimuli (data not shown). Because the day 5 viability of KG1 was decreased in PMA (75% viable), PMA plus ␣ Table I. Surface Ag expression of unstimulated KG1 and KG1a a TNF- (70%), and PMA plus ionomycin (67%), we examined the induction of apoptosis by these stimuli (Table II). In KG1, apo- KG1 KG1a ptosis was triggered within4hbyPMAplus iono and within 24 h by PMA plus TNF-␣. Smaller increases in apoptotic cell numbers ϩϩϩϩ ϩϩϩϩ MHC class I were seen at 24 h in cultures treated with PMA or TNF-␣ alone, MHC class II ϩϩϩϩ Ϯ CD1a ϪϪ CD2 ϪϪ CD10 ϪϪ CD13 ϩϩ ϩϩϩ CD14 ϪϪ CD28 ϪϪ CD33 ϩϩ ϩ CD34 ϩϩϩϩ ϩϩϩϩ CD40 ϪϪ CD45RA ϩϩ CD80 (B7–1) ϪϪ CD83 ϪϪ CD86 (B7–2) ϩϪ CD90 (Thy-1) ϪϪ CD152 (CTLA4) ϪϪ CD154 (CD40L) ϪϪFIGURE 3. KG1 and KG1a expression of TNF-␣ receptors 1 and 2. c-fms ϮϪUnstimulated KG1 and KG1a were analyzed by FACS for TNF receptor 1 a Each ϩ indicates an increase of 0.5 logs in mean fluorescent intensity vs. isotype and 2 expression as described in Materials and Methods. The isotype- matched control. matched control is the dotted line, and Ag-specific Ab the solid line. 3242 A CELL LINE MODEL OF HUMAN DENDRITIC CELL DIFFERENTIATION

Table II. Induction of apoptosis in KG1 and KG1a a

% Apoptotic Cells

KG1 KG1a

Condition 4h24h4h24h

Media 5.3 5.8 11.9 9.4 PMA 7.2 9.9 9.8 5.6 Iono 6.8 4.4 9.6 8.5 TNF-␣ 5.5 7.9 9.5 9.6 PMA ϩ TNF-␣ 8.7 17.9 7.4 8.7 PMA ϩ iono 14.4 22.4 8.7 10.1

a Cells were stimulated as indicated and assessed for induction of apoptosis (by annexin V binding) at the times indicated.

tional characteristic of both DC and CD34ϩCD86ϩ DC precursors (7, 37). Ag capture (as measured by FITC-dextran uptake) changes from a high capacity in immature/unactivated DC to a lower ca- Downloaded from pacity in mature/activated DC (37). Unstimulated KG1 took up FITC-dextran, and PMA did not affect uptake even though it induced other characteristic DC changes (Fig. 6). However, TNF-␣ alone or in combination with PMA down-modulated macromolecule uptake, suggesting that TNF-␣ (but not PKC activation) de- FIGURE 4. Surface Ag phenotypes of KG1 and KG1a in response to livers a distinct signal that mediates the decreased Ag uptake ca- cytokine and/or PMA stimulation. KG1 cells were cultured in medium pability associated with DC maturation. http://www.jimmunol.org/ alone, GM-CSF (200 U/ml) plus TNF-␣ (2.5 ng/ml), PMA (10 ng/ml), PMA plus TNF-␣ (10 ng/ml), or PMA plus ionomycin (200 ng/ml). KG1a KG1-DLC induce allogeneic T cell proliferation were cultured in PMA (10 ng/ml). Cells were harvested and analyzed by The most distinctive functional characteristic of DC is the potent FACS. Ag-speciﬁc staining is indicated by the solid line, and the isotype- matched control is shown by the dotted line. The time of maximal surface ability to activate T cells (45). We found that PMA-generated KG1 expression varied depending on the stimulus, shown here are GM-CSF plus TNF-␣ and PMA plus ionomycin (day 3 in culture), and medium, PMA, and PMA plus TNF-␣ (day 7 in culture). by guest on September 28, 2021 correlating with a lesser inhibition of cell proliferation by these single factors. Also consistent with the proliferation data was the failure of any stimuli to induce apoptosis in KG1a (the greater baseline percentage of apoptotic cells was probably due to the higher proliferative rate).

Fluid phase macromolecule uptake by KG1 and KG1-DLC Ag uptake (via receptor-mediated endocytosis (Fc␥, mannose receptors) or macropinocytosis of ﬂuid phase molecules) is a func-

FIGURE 5. Inhibition of KG1 proliferation. KG1 were cultured at the FIGURE 6. Fluid phase macromolecule uptake by KG1 and KG1-DLC. cell numbers and in the conditions noted. After 54 h of culture, cells were KG1 were unstimulated (medium) or cultured in PMA, TNF-␣,orPMA pulsed with 0.5 ␮Ci/well [methyl-3H]thymidine before being harvested at plus TNF-␣ for 5 days. FITC-dextran uptake was then assayed as described 72 h. [3H]TdR incorporation was measured using a Beta Plate scintillation in Materials and Methods and analyzed by ﬂow cytometry. FITC-dextran counting system. uptake at 4°C by unstimulated KG1 served as the negative control. The Journal of Immunology 3243

early DLC differentiation in KG1a (based on CD83 expression), these findings suggest that DC-CK1 expression is maximal during early/intermediate stages of DC maturation/activation (induced by PMA in KG1, PMA plus TNF-␣ in KG1a) with much less expression in uncommitted progenitors (KG1a) or fully mature/activated DC (PMA- plus TNF-␣-treated KG1). Together, these findings demonstrate that in response to specific stimuli (GM-CSF plus TNF-␣, PMA with or without TNF-␣,or ionomycin) KG1 undergo differentiation to cells that have morphology, surface Ag phenotype, function, and gene expression characteristic of DC. In addition, specific aspects of this differentiation appear to be mediated by specific signals. Role of PKC in PMA-driven KG13DLC differentiation Although PMA is a stable 2,3-diacylglycerol analogue and specific FIGURE 7. T cell proliferation in response to KG1 and KG1a cultured in a variety of conditions. KG1 and KG1a cells were cultured for 7 days in activator of the classical and new isoforms of PKC (47), it is for- medium, ionomycin (50 ng/ml), PMA (10 ng/ml), and PMA plus ionomy- mally possible that PMA’s effects in KG1 are due to activation of cin. Cells were thoroughly washed and added to purified resting allogeneic a non-PKC phorbol ester receptor (such as Unc-11 or n-chimaerin) T cells in flat-bottom microtiter plates at an APC:T ratio of 1:10. Freshly (48–50). To more rigorously prove specific PKC activation, KG1 Downloaded from isolated allogeneic PBMC were used as APC for comparison. Proliferation were cultured in PMA with or without TNF-␣ with the highly was measured by [3H]TdR incorporation. selective PKC inhibitor bis. Bis inhibited adhesion and morphologic changes in KG1 cultured in PMA without affecting cell viability (Fig. 9A). By FACS, bis inhibited up-regulation of CD83 DLC were capable of stimulating allogeneic T cell proliferation and CD86 in response to both PMA and PMA plus TNF-␣ (Fig.

(Fig. 7) and to a comparatively greater extent than unselected 9B). Bis also prevented cell death induced by PMA plus TNF-␣ http://www.jimmunol.org/ PBMC (as a reference population of APC). KG1 cultured in PMA (91% viable with PMA, TNF-␣, and bis on day 5 vs 37% viable and PMA plus ionomycin were significantly more potent at induc- with PMA plus TNF-␣ alone). Finally, the addition of bis inhibited ing T cell proliferation than KG1 cultured in medium, KG1a cul- the development of allostimulatory capacity in KG1 cultured in tured in PMA with or without ionomycin, or unselected PBMC. PMA or PMA plus ionomycin (Fig. 9C). The fact that bis inhibi- Consistent with the morphology and surface phenotype findings tion is considerably less effective with the addition of ionomycin is above, ionomycin appeared to drive further maturation/activation probably due to a lowered PKC activation threshold caused by over PMA alone as measured by alloproliferation. Lack of signif- increased intracellular calcium. icant T cell proliferation with PMA-treated KG1a cells demon- To further determine the nature of PKC-mediated signal trans- strates the lack of PMA carryover from the original KG1/KG1a duction KG13DLC differentiation, we asked whether the differ- by guest on September 28, 2021 differentiation cultures. ing PMA responsiveness of KG1 vs KG1a was due to differing PKC isoform expression. There are at least 11 different isoforms of KG1-DLC express characteristic DC genes PKC (47), with individual isoforms appearing to mediate specific Characterization of DC has classically relied on morphology, sur- biological function (51, 52). As seen in Fig. 10, unstimulated KG1 face Ag expression, and function. Recent molecular studies have expressed substantially more PKC␣ than KG1a. Unstimulated identified genes that are expressed predominantly or exclusively KG1a also did not express PKC␤I, -␤II, and -␮, in contrast to by DC. These include the RelB transcription factor (see below) and KG1. Both KG1 and KG1a expressed PKC⑀, and neither expressed the chemokine DC-CK1. The DC-CK1 chemokine is expressed PKC␥ (data not shown). only by DC and not other APC, and appears to function as a che- 3 moattractant for naive T cells (46). Unstimulated KG1 expressed Induction of RelB expression during PMA-driven KG1 DLC low levels of DC-CK1 that were substantially up-regulated by differentiation PMA (Fig. 8A). Unexpectedly, the addition of TNF-␣ inhibited RelB is a member of the NF␬B transcription factor family that is DC-CK1 up-regulation in KG1. The converse response was seen in highly expressed in DC but not monocytes (2) and may play an KG1a (Fig. 8B), where PMA has no effect on DC-CK1 expression important role in DC differentiation and function. Because up-reg- but PMA plus TNF-␣ results in up-regulation, arguing against a ulation/activation of RelB by PMA may represent the pathway by direct inhibitory effect of TNF-␣. Since PMA plus TNF-␣ induces which PKC-mediated signaling gains access to the nucleus, we

FIGURE 8. Up-regulation of DC-CK1 expression during PMA-driven KG13DLC differentiation. KG1 and KG1a cells were stimulated with PMA with or without TNF-␣, and DNase-treated RNA were made at the times indicated. Equal amounts of cDNA from mRNA were amplified by PCR with gene-specific primers, hybridized to internal DC-CK 1-specific oligonucleotide, separated by PAGE, exposed, and visualized on a phosphorimager. Expression of the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in the same samples was used as a control. 3244 A CELL LINE MODEL OF HUMAN DENDRITIC CELL DIFFERENTIATION

FIGURE 10. Western blot analysis of PKC isoform expression in KG1

and KG1a cells. Cell lysates were made from unstimulated KG1 and KG1a Downloaded from cells. Samples containing 20 ␮g of protein were separated by SDS-PAGE (5% stacking/10% resolving), electroblotted to nitrocellulose, probed with Abs against speciﬁc PKC isoforms, and visualized by chemiluminescent detection. The same results were obtained using equal numbers of cells per lane. http://www.jimmunol.org/ relatively slow turnover rate. By electromobility shift assays the levels of free (i.e., not bound to I␬B) RelB heterodimers capable of binding to NF␬B sites were also up-regulated by PMA in KG1, while levels in KG1a were comparatively stable (Fig. 11C, upper panels). Examination of other NF␬B family members in KG1 and KG1a revealed that nearly all protein capable of binding NF␬B sites contained p50 (NF␬BI) as one subunit (Fig. 11C, lower panels) as evidenced by the nearly complete supershifting by anti-p50

Abs. This suggests that in KG1/KG1a RelB exists as a RelB/p50 by guest on September 28, 2021 heterodimer.

Discussion FIGURE 9. The effect of the PKC inhibitor bis on PMA-driven Why hemopoietic stem cells differentiate (or do not) into a specific KG13DLC differentiation. KG1 and KG1a cells were cultured for 7 days mature lineage remains one of the enduring biological riddles. ␮ ␣ in medium; bis (1 M); PMA (10 ng/ml); PMA plus bis; PMA plus TNF- Seminal work by a number of groups (53–68) have identified the ␣ (10 ng/ml); PMA, TNF- , and bis; PMA plus ionomycin (50 ng/ml); and critical receptor-mediated stimuli (cytokines, CD40) that induce PMA, ionomycin, and bis. A, Morphology. Photomicrographs of KG1 cul- ϩ lineage commitment of human multipotential CD34 HPC to DC tured in PMA or PMA plus bis are shown. The focal plane is the bottom of the well (adherent cells). Arrows point to characteristic neurite pro- in vitro. We have sought to further characterize the intracellular cesses. B, Surface Ag phenotype. Ten thousand day 7 cells were analyzed signaling pathways that connect these receptor/membrane events as described in Materials and Methods. Dotted lines show the isotype to the nucleus. We have previously shown that the PKC agonist ϩ control; solid lines show the Ag-specific Ab. C, Induction of allogeneic T PMA by itself directly triggers differentiation of human CD34 cell proliferation. KG1 and KG1a were cultured as indicated, thoroughly HPC to DC, suggesting a critical and specific role of this pathway washed, and added to purified resting allogeneic T cells in flat-bottom in lineage commitment to DC (14). This is supported by observa- microtiter plates at an APC:T ratio of 1:10. Freshly isolated PBMC were tions that the receptor-mediated stimuli (GM-CSF, IL-4, TNF-␣, 3 used for comparison. Proliferation was measured by [ H]TdR incorpora- and CD40 cross-linking) that induce DC differentiation all can tion. The data in this figure are a subset of those for the experiment pre- activate PKC as part of their intracellular signaling pathways (15– sented in Fig. 7. 19, 69). We now find that the PKC inhibitor staurosporine blocks cytokine-driven CD34ϩ HPC3DC differentiation, indicating that PKC is part of the signaling cascade initiated by GM-CSF and examined the kinetics of RelB expression following PMA activa- TNF-␣. Although staurosporine can also block tyrosine kinase ac- tion of KG1 and KG1a. As can be seen in Fig. 11A, RelB gene tivity, this is typically at a substantially higher dose than that used expression in KG1 was up-regulated within 24 h, plateaued in these experiments (40). We have also found that more specific through 72 h, and then declined. Unexpectedly, unstimulated PKC inhibitors (e.g., bis) similarly inhibit CD34ϩHPC3DC dif- KG1a constitutively expressed RelB message, which also declined ferentiation (manuscript in preparation). Staurosporine did not glo- after 72 h in culture with PMA. Protein expression paralleled the bally block differentiation as all cells lost the CD34 marker, and inducible KG1 or constitutive KG1a gene expression of RelB (Fig. the generation of CD14ϩ cells (most likely monocytes) was com- 11B). Of note is the rapidity with which RelB protein was induced paratively less affected. Rather, the generation of CD83ϩ DC was in KG1, appearing within 30 min of PMA stimulation. The per- preferentially inhibited. When considered with our previous find- sistence of RelB protein after gene down-regulation suggests a ings that PMA treatment of CD34ϩ HPC only generates DC (and The Journal of Immunology 3245

no other myeloid lineages), these data indicate that the intracellular pathway that specifically initiates DC lineage commitment goes through PKC. This “wiring” may be unique to CD34ϩ HPC, as CD34 negative progenitors and monocytes have different responses to PMA (K. P. Lee and K. Schlienger, unpublished observations). To further study the intracellular events involved in DC differentiation and to circumvent the problems of studying this in primary CD34ϩ HPC, we established a human leukemic cell line model of CD34ϩ HPC3DC differentiation. Work demonstrating that primary chronic myelogenous leukemia cells differentiated in vitro to functional DC suggested that this was feasible (70). Ex- amination of the PMA responsiveness of a panel of CD34ϩ leukemic cell lines revealed that KG1 differentiated into cells with DC characteristics. KG1 was isolated from a patient with erythroleu- kemia (FAB M7) undergoing myeloblastic relapse (28). Serial pas- sage of KG1 gave rise to the KG1a subline (71), which by a number of criteria (surface phenotype, gene expression, and growth Downloaded from factor unresponsiveness) is thought to be arrested at a less differentiated stage than KG1 (32, 72, 73). In response to PMA KG1 was initially characterized to differentiate into macrophages (74), although these original morphologic (adherent cells with long pseudopodia), histochemical (nonspecific esterase-positive, my- eloperoxidase-negative), surface Ag (MHC II- and Fc receptor- positive), and functional (phagocytosis) findings have since been http://www.jimmunol.org/ found to also be characteristic of DC (reviewed in Ref. 12). PMA does not induce differentiation in KG1a, and although KG1 and KG1a have similar numbers of phorbol ester binding sites, they differ in PKC activity, translocation, and substrate phosphorylation (30, 33, 34, 74). We believe that recharacterization of the KG1 response to PMA reveals it to actually recapitulate differentiation of the committed ϩ CD34 myeloid DC/macrophage progenitor to DC. Like the com- by guest on September 28, 2021 mitted DC/macrophage progenitor (7), KG1 express MHC II, CD34, and CD86, whereas KG1a are MHC IIϪCD34ϩCD86Ϫ. Unstimulated KG1 express little or no CD40, but up-regulate expression following culture in PMA. Stimuli shown to induce CD34ϩHPC3DC differentiation (GM-CSF plus TNF-␣, PMA) also induce differentiation of KG1 to DLC (KG1-DLC), while non-DC cytokine(s) (GM-CSF or TNF-␣ alone or GM-CSF, IL-3, IL-6, and SCF) do not. Unlike monocytes, ionomycin alone does not induce DC differentiation in KG1 (or primary CD34ϩ HPC), and likewise, PMA (as a single agent) does not drive DC differentiation in monocytes. Suppression of KG1 cell proliferation and induction of apoptosis by PMA are also characteristics we reported for primary CD34ϩ HPC3DC differentiation driven by PMA. Stimulation of KG1 with GM-CSF and TNF-␣ or PMA (with or FIGURE 11. Expression of RelB in PMA-stimulated KG1 and KG1a. without TNF-␣ or ionomycin) generates cells with a constellation KG1 and KG1a were cultured in 10 ng/ml PMA for the times indicated. A, of DC characteristics. Morphologically, KG1-DLC are adherent RelB mRNA expression. Cells were harvested and mRNA isolated as de- with long neurite processes in culture and have hair-like processes/ scribed in Materials and Methods. Total mRNA was equalized between veils on cytospin. By surface Ag phenotype, KG1-DLC have char- samples by ethidium bromide visualization and serial dilution. Samples acteristic DC expression of MHC I, MHC II (high), CD13, CD33, were separated on formaldehyde agarose gels, transferred to nylon, and the DC-specific lineage marker CD83, and the costimulatory li- 32 probed with P-labeled RelB cDNA probe. The same blot was then re- gand CD86 (reviewed in Ref. 12). They do not express CD1a or probed for actin expression to verify mRNA equalization and integrity. B, the monocyte marker CD14. Expression of CD80 is present but RelB protein expression. Cell lysates from the indicated times points were low, suggesting that KG1-DLC are not fully mature or activated made, and equal amounts of protein were loaded onto a SDS-PAGE gel and (12). Functionally, unstimulated KG1 can take up fluid phase Ag analyzed by Western blotting with the anti-RelB Ab C-19. C, RelB and p50 ␣ electromobility shift assay. Cell lysates were made at the times indicated. as measured by FITC-dextran uptake. TNF- appears to deliver a Equal amounts of protein were incubated with a labeled primer containing signal that down-regulates Ag uptake, a hallmark of DC matura- the NF␬B binding site with or without anti-RelB or anti-p50 mAb. Samples tion/activation (37) consistent with the other effects (morphology, were separated on 4% polyacrylamide gel and exposed to film. Complexes surface Ag phenotype) that TNF-␣ has on PMA-driven that specifically contain the labeled primer, Rel B (or p50), and anti-RelB KG13DLC differentiation. Most importantly, KG1-DLC can (or p50) Ab are indicated by the supershift arrow. function to activate T cells as evidenced by proliferation in an 3246 A CELL LINE MODEL OF HUMAN DENDRITIC CELL DIFFERENTIATION allo-MLR. This proliferation is substantially higher than that in- some single signals (PMA or CD40 cross-linking in primary duced by either allogeneic PBMC or similarly cultured KG1a. It is CD34ϩ HPC, calcium ionophore in monocytes) are capable of interesting to note that even though unstimulated KG1 express driving complete DC differentiation/maturation from progenitors. MHC II and CD86, they do not stimulate significant T cell pro- One possibility is that those particular pathways in those particular liferation. Because MHC II and CD86 expression are not substan- progenitors are directly linked to the genetic DC differentiation tially up-regulated by PMA, this suggests the enhanced allostimu- program. Another possibility is that the primary signal in these latory properties of KG1-DLC are due to other components not specific progenitors triggers autocrine secretion of additional cy- measured. Assessment of cytokine production (especially IL-12) is tokines/factors that drive terminal maturation. We are currently currently underway. Finally, KG1-DLC express genes that are pre- examining whether PMA induces TNF-␣ expression or calcium dominantly (RelB) or exclusively (DC-CK1) found in DC. To- flux in primary CD34ϩ HPC. gether, these four levels of characterization all point to the con- Regardless of the specific pathway, differentiation signals have clusion GM-CSF plus TNF-␣ or PMA stimulation of KG1 to cross into the nucleus to initiate the requisite genetic programs. generates cells of the DC lineage. Such bridges are very often preformed transcription factors, of The inability to induce DLC differentiation in KG1a allows for which the NF␬B family is one of the best characterized (reviewed comparison with KG1 of potential signaling pathways. Blockade in Ref. 78). Critical involvement of NF␬B-responsive genes in DC of DC differentiation in both KG1 and primary CD34ϩ HPC by differentiation has been found in chicken bone marrow cell lines PKC inhibitors establishes a central role for PKC. PKC activation (79). Involvement of rel/NF␬B family transcription factors in DC most likely plays a direct role and does not simply induce auto- differentiation is further suggested by the observations that RelB crine secretion of cytokines; for example, we find no evidence (by (80, 81) is expressed at high levels in murine DC (82) and that Downloaded from PCR) for induction of TNF-␣ expression in KG1. However, the RelB knockout mice have significant reductions in mature DC difference between KG1 and KG1a cannot simply be the absence number and APC function (83, 84). In human DC nuclear local- of PKC signaling in KG1a given that both cell lines have bio- ization of RelB is associated with enhanced APC function (85). chemically measurable PKC activity (75). Likewise, we have Both PMA and TNF-␣ up-regulate the activity of NF␬B family found that although both CD34ϩ HPC and CD34Ϫ HPC respond members (including RelB by PMA) (86), and RelB up-regulation ϩ Ϫ ϩ to PMA, only CD34 HPC differentiate to DC while CD34 HPC is seen during CD34 HPC3DC (14) and monocyte3DC (2) http://www.jimmunol.org/ become macrophages. This suggests that there is a qualitative com- differentiation. This is also true for KG13DLC differentiation. ponent of PKC activation. Consistent with this is the differing ex- However, constitutive expression of RelB in KG1a suggests that pression of specific PKC isoforms in the cell lines. KG1a express DC differentiation may not simply be due to activation of RelB. It less PKC␣ and no PKC␤I, -␤II, or -␮ compared with KG1. Given is also possible that tonic expression of RelB somehow blocks that specific PKC isoforms have been associated with specific bi- induction of NF␬B-responsive genes as has been suggested for ological events in other systems (e.g., PKC␪ in T cell activation conditional v-rel expression (79). It is unclear why RelB expres- (51) and PKC␤ in macrophage differentiation of HL-60 (52)), sion is not regulated in KG1a, but this suggests that some negative these findings suggest that the same may be true for DC differen- feedback signal has been lost. The rapid up-regulation of RelB tiation. Examination of the roles of specific PKC isoforms is cur- expression induced by PMA in KG1 may indicate the involvement by guest on September 28, 2021 rently underway. of a more immediate transcription factor, although NF␬B can me- KG1a also differ from KG1 in the response to TNF-␣; this is diate its own transcription (78). most evident in the induction of apoptosis. This difference may be Together, the data presented in this report corroborate an im- due to comparatively lower expression of both TNF receptors 1 portant role for PKC activation in DC differentiation from CD34ϩ and 2 on KG1a, and this membrane “defect” is consistent with HPC. Characterization of the KG1 cell line model of previous reports demonstrating that apoptotic responses in KG1a CD34ϩHPC3DC differentiation allows for a more detailed study can be restored by directly manipulating the sphingomyelin-cer- of the signal transduction pathways involved. Our initial studies amide pathway triggered by TNF-␣ (76, 77). The roles of ceramide indicate that the PKC, TNF-␣, and/or intracellular calcium signal- (and intracellular calcium flux) in both apoptosis and DC differing pathways mediate specific phenomenon during this differenti- entiation are under study. ation. A potential connection of these pathways to the RelB tran- In addition to the differences between KG1 and KG1a, we found scription factor may provide a doorway into uncovering the genes that distinct signals appeared to mediate distinct, nonoverlapping that control DC differentiation and underlie the unique function of biological phenomena during KG13DLC differentiation. PKC ac- this lineage. tivation by PMA appears to be a requisite inductive signal, initiating morphology changes, surface Ag expression, and allostimu- References latory capacity for T cells. However, the addition of a TNF-␣ or calcium flux signal is necessary for full maturation/activation (as 1. Banchereau, J., and R. M. Steinman. 1998. Dendritic cells and the control of immunity. Nature 392:245. assessed by morphology, surface Ag expression, and T cell pro- 2. Akagawa, K. S., N. Takasuka, Y. Nozaki, I. Komuro, M. Azuma, M. Ueda, liferation) and inhibition of cell proliferation/induction of apopto- M. Naito, and K. Takahashi. 1996. Generation of CD1ϩ RelBϩ dendritic cells and tartrate-resistant acid phosphatase-positive osteoclast-like multinucleated gi- sis. Down-regulation of soluble macromolecule uptake appears to ant cells from human monocytes. Blood 88:4029. be mediated entirely by TNF-␣ and to be independent of PKC 3. Pickl, W. F., O. Majdic, P. Kohl, J. Stockl, E. Riedl, C. Scheinecker, activation. The possibility that TNF-␣ and ionomycin are signaling C. Bello-Fernandez, and W. Knapp. 1996. Molecular and functional characteristics of dendritic cells generated from highly purified CD14ϩ peripheral blood through the same pathway is suggested by the lack of an additive monocytes. J. Immunol. 157:3850. effect when the two are combined (PMA, TNF-␣, and ionomycin) 4. Zhou, L. J., and T. F. Tedder. 1996. CD14ϩ blood monocytes can differentiate ϩ and is being examined. Together, these data suggest that there are into functionally mature CD83 dendritic cells. Proc. Natl. Acad. Sci. USA 93: 2588. discrete signaling pathways involved in specific aspects of DC 5. Kiertscher, S. M., and M. D. Roth. 1996. Human CD14ϩ leukocytes acquire the differentiation; PKC activation is required to initiate differentia- phenotype and function of antigen-presenting dendritic cells when cultured in tion, and TNF-␣ (which initiates several intracellular pathways) GM-CSF and IL-4. J. Leukocyte Biol. 59:208. 6. Galy, A., M. Travis, D. Cen, and B. Chen. 1995. Human T, B, natural killer, and and/or intracellular calcium flux is required for terminal matura- dendritic cells arise from a common bone marrow progenitor cell subset. Immu- tion/activation and apoptosis. However, this does not explain why nity 3:459. The Journal of Immunology 3247

7. Ryncarz, R. E., and C. Anasetti. 1998. Expression of CD86 on human marrow leukemia cell line KG-1 and its phorbol ester-resistant subline KG-1a. Cancer CD34(ϩ) cells identifies immunocompetent committed precursors of macro- Res. 47:6363. phages and dendritic cells. Blood 91:3892. 34. Lehrer, R. I., L. E. Cohen, and H. P. Koeffler. 1983. Specific binding of [3H]phor- 8. Winzler, C., P. Rovere, M. Rescigno, F. Granucci, G. Penna, L. Adorini, bol dibutyrate to phorbol diester-responsive and -resistant clones of a human V. S. Zimmermann, J. Davoust, and P. Ricciardi-Castagnoli. 1997. Maturation myeloid leukemia (KG-1) line 1. Cancer Res. 43:3563. stages of mouse dendritic cells in growth factor-dependent long-term cultures. 35. Davis, T. A., D. H. Robinson, K. P. Lee, and S. W. Kessler. 1995. Porcine brain J. Exp. Med. 185:317. microvascular endothelial cells support the in vitro expansion of human primitive 9. Caux, C., C. Dezutter-Dambuyant, D. Schmitt, and J. Banchereau. 1992. GM- hematopoietic bone marrow progenitor cells with a high replating potential: re- CSF and TNF-␣ cooperate in the generation of dendritic Langerhans cells. Nature quirement for cell-to-cell interactions and colony-stimulating factors. Blood 85: 360:258. 1751. 10. Flores-Romo, L., P. Bjork, V. Duvert, C. Van Kooten, S. Saeland, and 36. Koopman, G., C. P. Reutelingsperger, G. A. Kuijten, R. M. Keehnen, S. T. Pals, ϩ J. Banchereau. 1997. CD40 ligation on human cord blood CD34 hematopoietic and M. H. van Oers. 1994. Annexin V for flow cytometric detection of phos- progenitors induces their proliferation and differentiation into functional dendritic phatidylserine expression on B cells undergoing apoptosis. Blood 84:1415. cells. J. Exp. Med. 2:341. 37. Sallusto, F., M. Cella, C. Danieli, and A. Lanzavecchia. 1995. Dendritic cells use 11. Czerniecki, B. J., C. Carter, L. Rivoltini, G. K. Koski, H. I. Kim, D. E. Weng, macropinocytosis and the mannose receptor to concentrate macromolecules in the J. G. Roros, Y. M. Hijazi, S. Wu, S. A. Rosenberg, et al. 1997. Calcium iono- major histocompatibility complex class II compartment: downregulation by cy- phore-treated peripheral blood monocytes and dendritic cells rapidly display tokines and bacterial products. J. Exp. Med. 182:389. characteristics of activated dendritic cells. J. Immunol. 159:3823. 38. June, C. H., J. A. Ledbetter, M. M. Gillespie, T. Lindsten, and C. B. Thompson. 12. Hart, D. N. 1998. Dendritic cells: unique leukocyte populations which control the 1987. T-cell proliferation involving the CD28 pathway is associated with cyclos- primary immune response. Blood 90:3245. porine-resistant interleukin 2 gene expression. Mol. Cell. Biol. 7:4472. 13. Reddy, A., M. Sapp, M. Feldman, M. Subklewe, and N. Bhardwaj. 1997. A 39. Lee, K. P., C. Taylor, B. Petryniak, L. A. Turka, C. H. June, and C. B. Thompson. monocyte conditioned medium is more effective than defined cytokines in me- 1990. The genomic organization of the CD28 gene: implications for the regula- diating the terminal maturation of human dendritic cells. Blood 90:3640. tion of CD28 mRNA expression and heterogeneity. J. Immunol. 145:344. 14. Davis, T. A., A. A. Saini, P. J. Blair, B. L. Levine, N. Craighead, D. M. Harlan, 40. Meggio, F., A. Donella Deana, M. Ruzzene, A. M. Brunati, L. Cesaro, B. Guerra,

C. H. June, and K. P. Lee. 1998. Phorbol esters induce differentiation of human Downloaded from ϩ T. Meyer, H. Mett, D. Fabbro, and P. Furet. 1995. Different susceptibility of CD34 hemopoietic progenitors to dendritic cells: evidence for protein kinase protein kinases to staurosporine inhibition: kinetic studies and molecular bases C-mediated signaling. J. Immunol. 160:3689. for the resistance of protein kinase CK2. Eur. J. Biochem. 234:317. 15. Arruda, S., and J. L. Ho. 1992. IL-4 receptor signal transduction in human mono- 41. Zhou, L. J., and T. F. Tedder. 1995. Human blood dendritic cells selectively cytes is associated with protein kinase C translocation. J. Immunol. 149:1258. express CD83, a member of the immunoglobulin superfamily. J. Immunol. 154: 16. Harnett, M. M., M. J. Holman, and G. G. Klaus. 1991. IL-4 promotes anti-Ig- 3821. mediated protein kinase C translocation and reverses phorbol ester-mediated pro- 42. June, C. H., J. A. Bluestone, L. M. Nadler, and C. B. Thompson. 1994. The B7 tein kinase C down- regulation in murine B cells. J. Immunol. 147:3831. and CD28 receptor families. Immunol. Today 15:321. 17. Finney, M., G. R. Guy, R. H. Michell, J. Gordon, B. Dugas, K. P. Rigley, and 43. Fearnley, D. B., A. D. McLellan, S. I. Mannering, B. D. Hock, and D. N. Hart. http://www.jimmunol.org/ R. E. Callard. 1990. Interleukin 4 activates human B lymphocytes via transient 1997. Isolation of human blood dendritic cells using the CMRF-44 monoclonal inositol lipid hydrolysis and delayed cyclic adenosine monophosphate generation. antibody: implications for studies on antigen-presenting cell function and immu- Eur. J. Immunol. 20:151. notherapy. Blood 89:3708. 18. Schutze, S., T. Machleidt, and M. Kronke. 1994. The role of diacylglycerol and 44. Territo, M. C., and H. P. Koeffler. 1981. Induction by phorbol esters of macro- ceramide in tumor necrosis factor and interleukin-1 signal transduction. J. Leu- phage differentiation in human leukaemia cell lines does not require cell division. kocyte Biol. 56:533. Br. J. Haematol. 47:479. 19. Ren, C. L., S. M. Fu, R. S. Geha, and T. Morio. 1994. Protein tyrosine kinase 45. Thomas, R., L. S. Davis, and P. E. Lipsky. 1993. Comparative accessory cell activation and protein kinase C translocation are functional components of CD40 function of human peripheral blood dendritic cells and monocytes. J. Immunol. signal transduction in resting human B cells: signal transduction via CD40 in- 151:6840. volves activation of Lyn kinase and phosphatidylinositol-3-kinase, and phosphorylation of phospholipase C␥2. J. Exp. Med. 179:673. 46. Adema, G. J., F. Hartgers, R. Verstraten, E. de Vries, G. Marland, S. Menon, J. Foster, Y. Xu, P. Nooyen, T. McClanahan, et al. 1997. A dendritic-cell-derived 20. Sallusto, F., C. Nicolo, R. De Maria, S. Corinti, and R. Testi. 1996. Ceramide by guest on September 28, 2021 C-C chemokine that preferentially attracts naive T cells. Nature 387:713. inhibits antigen uptake and presentation by dendritic cells. J. Exp. Med. 184: 2411. 47. Nishizuka, Y. 1995. Protein kinase C and lipid signaling for sustained cellular 21. Carrasco, D., R. P. Ryseck, and R. Bravo. 1993. Expression of RelB transcripts responses. FASEB J. 9:484. during lymphoid organ development: specific expression in dendritic antigen- 48. Kazanietz, M. G., N. E. Lewin, J. D. Bruns, and P. M. Blumberg. 1995. Char- presenting cells. Development 118:1221. acterization of the cysteine-rich region of the Caenorhabditis elegans protein 22. Boehmelt, G., J. Madruga, P. Dorfler, K. Briegel, H. Schwarz, P. J. Enrietto, and Unc-13 as a high affinity phorbol ester receptor: analysis of ligand-binding in- M. Zenke. 1995. Dendritic cell progenitor is transformed by a conditional v-Rel teractions, lipid cofactor requirements, and inhibitor sensitivity. J. Biol. Chem. estrogen receptor fusion protein v-RelER. Cell 80:341. 270:10777. 23. Feuillard, J., M. Korner, A. Israel, J. Vassy, and M. Raphael. 1996. Differential 49. Ahmed, S., J. Lee, R. Kozma, A. Best, C. Monfries, and L. Lim. 1993. A novel functional target for tumor-promoting phorbol esters and lysophosphatidic acid: nuclear localization of p50, p52, and RelB proteins in human accessory cells of rac the immune response in situ. Eur. J. Immunol. 26:2547. the p21 -GTPase activating protein n-chimaerin. J. Biol. Chem. 268:10709. 24. Lernbecher, T., B. Kistler, and T. Wirth. 1994. Two distinct mechanisms con- 50. Ahmed, S., I. N. Maruyama, R. Kozma, J. Lee, S. Brenner, and L. Lim. 1992. The tribute to the constitutive activation of RelB in lymphoid cells. EMBO J. 13:4060. Caenorhabditis elegans unc-13 gene product is a phospholipid-dependent high- 25. Burkly, L., C. Hession, L. Ogata, C. Reilly, L. A. Marconi, D. Olson, R. Tizard, affinity phorbol ester receptor. Biochem. J. 287:995. R. Cate, and D. Lo. 1995. Expression of RelB is required for the development of 51. Monks, C. R. F., H. Kupper, I. Tamir, A. Barlow, and A. Kupfer. 1997. Selective thymic medulla and dendritic cells. Nature 373:531. modulation of protein kinase c-␪ during T-cell activation. Nature 385:83. 26. Weih, F., D. Carrasco, S. K. Durham, D. S. Barton, C. A. Rizzo, R. P. Ryseck, 52. Macfarlane, D. E., and L. Manzel. 1994. Activation of ␤-isozyme of protein S. A. Lira, and R. Bravo. 1995. Multiorgan inflammation and hematopoietic kinase C (PKC␤) is necessary and sufficient for phorbol ester-induced differen- abnormalities in mice with a targeted disruption of RelB, a member of the NF- tiation of HL-60 promyelocytes: studies with PKC␤-defective PET mutant. ␬B/Rel family. Cell 80:331. J. Biol. Chem. 269:4327. 27. Sekhar, M., H. Kotani, S. Doren, R. Agarwal, G. McGarrity, and C. E. Dunbar. 53. Inaba, K., M. Inaba, M. Deguchi, K. Hagi, R. Yasumizu, S. Ikehara, 1996. Retroviral transduction of CD34-enriched hematopoietic progenitor cells S. Muramatsu, and R. M. Steinman. 1993. Granulocytes, macrophages, and den- under serum-free conditions. Hum. Gene Ther. 7:33. dritic cells arise from a common major histocompatibility complex class II-neg- 28. Koeffler, H. P., and D. W. Golde. 1978. Acute myelogenous leukemia: a human ative progenitor in mouse bone marrow. Proc. Natl. Acad. Sci. USA 90:3038. cell line responsive to colony-stimulating activity. Science 200:1153. 54. Inaba, K., M. Inaba, N. Romani, H. Aya, M. Deguchi, S. Ikehara, S. Muramatsu, 29. Koeffler, H. P., M. Bar-Eli, and M. C. Territo. 1981. Phorbol ester effect on and R. M. Steinman. 1992. Generation of large numbers of dendritic cells from differentiation of human myeloid leukemia cell lines blocked at different stages of mouse bone marrow cultures supplemented with granulocyte/macrophage colo- maturation. Cancer Res. 41:919. ny-stimulating factor. J. Exp. Med. 176:1693. 30. Kiss, Z., E. Deli, M. Shoji, H. P. Koeffler, G. R. Pettit, W. R. Vogler, and 55. Caux, C., C. Dezutter-Dambuyant, D. Schmitt, and J. Banchereau. 1992. GM- J. F. Kuo. 1987. Differential effects of various protein kinase C activators on CSF and TNF-␣ cooperate in the generation of dendritic Langerhans cells. Nature protein phosphorylation in human acute myeloblastic leukemia cell line KG-1 360:258. and its phorbol ester-resistant subline KG-1a. Cancer Res. 47:1302. 56. Santiago-Schwarz, F., E. Belilos, B. Diamond, and S. E. Carsons. 1992. TNF in 31. Ferrero, D., S. Pessano, G. L. Pagliardi, and G. Rovera. 1983. Induction of dif- combination with GM-CSF enhances the differentiation of neonatal cord blood ferentiation of human myeloid leukemias: surface changes probed with mono- stem cells into dendritic cells and macrophages. J. Leukocyte Biol. 52:274. clonal antibodies. Blood 61:171. 57. Reid, C. D., A. Stackpoole, A. Meager, and J. Tikerpae. 1992. Interactions of 32. Koeffler, H. P., R. Billing, A. J. Lusis, R. Sparkes, and D. W. Golde. 1980. An tumor necrosis factor with granulocyte-macrophage colony-stimulating factor undifferentiated variant derived from the human acute myelogenous leukemia cell and other cytokines in the regulation of dendritic cell growth in vitro from early line (KG-1). Blood 56:265. bipotent CD34ϩ progenitors in human bone marrow. J. Immunol. 149:2681. 33. Shoji, M., P. R. Girard, P. A. Charp, H. P. Koeffler, W. R. Vogler, and J. F. Kuo. 58. Rosenzwajg, M., B. Canque, and J. C. Gluckman. 1996. Human dendritic cell 1987. Effects of phorbol ester on translocation and down-regulation of protein differentiation pathway from CD34ϩ hematopoietic precursor cells. Blood 87: kinase C and phosphorylation of endogenous proteins in human acute myeloid 535. 3248 A CELL LINE MODEL OF HUMAN DENDRITIC CELL DIFFERENTIATION

59. Caux, C., S. Saeland, C. Favre, V. Duvert, P. Mannoni, and J. Banchereau. 1990. 73. Fukuda, M., H. P. Koeffler, and J. Minowada. 1981. Membrane differentiation in Tumor necrosis factor-␣ strongly potentiates interleukin-3 and granulocyte-mac- human myeloid cells: expression of unique profiles of cell surface glycoproteins rophage colony-stimulating factor-induced proliferation of human CD34ϩ hema- in myeloid leukemic cell lines blocked at different stages of differentiation and topoietic progenitor cells. Blood 75:2292. maturation. Proc. Natl. Acad. Sci. USA 78:6299. 60. Scheicher, C., M. Mehlig, R. Zecher, and K. Reske. 1992. Dendritic cells from 74. Koeffler, H. P., M. Bar-Eli, and M. C. Territo. 1981. Phorbol ester effect on mouse bone marrow: in vitro differentiation using low doses of recombinant differentiation of human myeloid leukemia cell lines blocked at different stages of granulocyte-macrophage colony-stimulating factor. J. Immunol. Methods 154: maturation. Cancer Res. 41:919. 253. 75. Shoji, M., P. R. Girard, P. A. Charp, H. P. Koeffler, W. R. Vogler, and J. F. Kuo. 61. Markowicz, S., and E. G. Engleman. 1990. Granulocyte-macrophage colony- 1987. Effects of phorbol ester on translocation and down-regulation of protein stimulating factor promotes differentiation and survival of human peripheral kinase C and phosphorylation of endogenous proteins in human acute myeloid blood dendritic cells in vitro. J. Clin. Invest. 85:955. leukemia cell line KG-1 and its phorbol ester-resistant subline KG-1a. Cancer 62. Witmer-Pack, M. D., W. Olivier, J. Valinsky, G. Schuler, and R. M. Steinman. Res. 47:6363. 1987. Granulocyte/macrophage colony-stimulating factor is essential for the viability and function of cultured murine epidermal Langerhans cells. J. Exp. Med. 76. Bettaieb, A., M. Record, M. G. Come, A. C. Bras, H. Chap, G. Laurent, and ␣ 166:1484. J. P. Jaffrezou. 1996. Opposite effects of tumor necrosis factor on the sphin- 63. Szabolcs, P., M. A. Moore, and J. W. Young. 1995. Expansion of immunostimu- gomyelin-ceramide pathway in two myeloid leukemia cell lines: role of trans- latory dendritic cells among the myeloid progeny of human CD34ϩ bone marrow verse sphingomyelin distribution in the plasma membrane. Blood 88:1465. precursors cultured with c-Kit ligand, granulocyte-macrophage colony-stimulat- 77. Bezombes, C., N. Maestre, G. Laurent, T. Levade, A. Bettaieb, and ing factor, and TNF-␣. J. Immunol. 154:5851. J. P. Jaffrezou. 1998. Restoration of TNF-␣-induced ceramide generation and 64. Young, J. W., P. Szabolcs, and M. A. Moore. 1995. Identification of dendritic cell apoptosis in resistant human leukemia KG1a cells by the P-glycoprotein blocker colony-forming units among normal human CD34ϩ bone marrow progenitors PSC833. FASEB J. 12:101. that are expanded by c-Kit-ligand and yield pure dendritic cell colonies in the 78. Siebenlist, U., G. Franzoso, and K. Brown. 1994. Structure, regulation and func- presence of granulocyte/macrophage colony-stimulating factor and tumor necro- tion of NF-␬B. Annu. Rev. Cell Biol. 10:405. ␣ sis factor . J. Exp. Med. 182:1111. 79. Boehmelt, G., J. Madruga, P. Dorfler, K. Briegel, H. Schwarz, P. J. Enrietto, and 65. Romani, N., S. Gruner, D. Brang, E. Kampgen, A. Lenz, B. Trockenbacher, M. Zenke. 1995. Dendritic cell progenitor is transformed by a conditional v-Rel Downloaded from G. Konwalinka, P. O. Fritsch, R. M. Steinman, and G. Schuler. 1994. Prolifer- estrogen receptor fusion protein v-RelER. Cell 80:341. ating dendritic cell progenitors in human blood. J. Exp. Med. 180:83. 66. Caux, C., B. Vanbervliet, C. Massacrier, B. Dubois, C. Dezutter-Dambuyant, 80. Ryseck, R. P., P. Bull, M. Takamiya, V. Bours, U. Siebenlist, P. Dobrzanski, and ϩ R. Bravo. 1992. RelB, a new Rel family transcription activator that can interact D. Schmitt, and J. Banchereau. 1995. Characterization of human CD34 derived ␬ dendritic/Langerhans cells (D-Lc). Adv. Exp. Med. Biol. 378:1. with p50-NF- B. Mol. Cell. Biol. 12:674. 67. Szabolcs, P., D. Avigan, S. Gezelter, D. H. Ciocon, M. A. Moore, 81. Tewari, M., P. Dobrzanski, K. L. Mohn, D. E. Cressman, J. C. Hsu, R. Bravo, and R. M. Steinman, and J. W. Young. 1996. Dendritic cells and macrophages can R. Taub. 1992. Rapid induction in regenerating liver of RL/IF-1 (an I␬B that mature independently from a human bone marrow-derived, post-colony-forming inhibits NF-␬B, RelB-p50, and c-Rel-p50) and PHF, a novel ␬B site-binding

unit intermediate. Blood 87:4520. complex. Mol. Cell. Biol. 12:2898. http://www.jimmunol.org/ 68. Flores-Romo, L., P. Bjork, V. Duvert, C. van Kooten, S. Saeland, and 82. Carrasco, D., R. P. Ryseck, and R. Bravo. 1993. Expression of relB transcripts ϩ J. Banchereau. 1997. CD40 ligation on human cord blood CD34 hematopoietic during lymphoid organ development: specific expression in dendritic antigen- progenitors induces their proliferation and differentiation into functional dendritic presenting cells. Development 118:1221. cells. J. Exp. Med. 2:341. 83. Burkly, L., C. Hession, L. Ogata, C. Reilly, L. A. Marconi, D. Olson, R. Tizard, 69. Mufson, R. A. 1997. The role of serine/threonine phosphorylation in hematopoi- R. Cate, and D. Lo. 1995. Expression of RelB is required for the development of etic cytokine receptor signal transduction. FASEB J. 11:37. thymic medulla and dendritic cells. Nature 373:531. 70. Choudhury, A., J. L. Gajewski, J. C. Liang, U. Popat, D. F. Claxton, K. O. Kliche, 84. Weih, F., D. Carrasco, S. K. Durham, D. S. Barton, C. A. Rizzo, R. P. Ryseck, M. Andreeff, and R. E. Champlin. 1997. Use of leukemic dendritic cells for the S. A. Lira, and R. Bravo. 1995. Multiorgan inflammation and hematopoietic generation of antileukemic cellular cytotoxicity against Philadelphia chromo- abnormalities in mice with a targeted disruption of RelB, a member of the NF- some-positive chronic myelogenous leukemia. Blood 89:1133. ␬B/Rel family. Cell 80:331. 71. Koeffler, H. P., R. Billing, A. J. Lusis, R. Sparkes, and D. W. Golde. 1980. An

undifferentiated variant derived from the human acute myelogenous leukemia cell 85. Pettit, A. R., C. Quinn, K. P. A. MacDonald, L. L. Cavanagh, G. Thomas, by guest on September 28, 2021 line (KG-1). Blood 56:265. W. Townsend, M. Handel, and R. Thomas. 1997. Nuclear localization of RelB is 72. Furley, A. J., B. R. Reeves, S. Mizutani, L. J. Altass, S. M. Watt, M. C. Jacob, associated with effective antigen-presenting cell function. J. Immunol. 159:3681. P. van den Elsen, C. Terhorst, and M. F. Greaves. 1986. Divergent molecular 86. Olashaw, N. E. 1996. Inducible activation of RelB in ﬁbroblasts. J. Biol. Chem. phenotypes of KG1 and KG1a myeloid cell lines. Blood 68:1101. 271:30307.