Cutting Edge: Generation of Splenic CD8+ and CD8– Dendritic Cell Equivalents in Fms-Like Tyrosine Kinase 3 Ligand Bone Marrow Cultures1
http://www.100md.com
免疫学杂志 2005年第11期
Abstract
We demonstrate that functional and phenotypic equivalents of mouse splenic CD8+ and CD8– conventional dendritic cell (cDC) subsets can be generated in vitro when bone marrow is cultured with fms-like tyrosine kinase 3 (flt3) ligand. In addition to CD45RAhigh plasmacytoid DC, two distinct CD24high and CD11bhigh cDC subsets were present, and these subsets showed equivalent properties to splenic CD8+ and CD8– cDC, respectively, in the following: 1) surface expression of CD11b, CD24, and signal regulatory protein-; 2) developmental dependence on, and mRNA expression of, IFN regulatory factor-8; 3) mRNA expression of TLRs and chemokine receptors; 4) production of IL-12 p40/70, IFN-, MIP-1, and RANTES in response to TLR ligands; 5) expression of cystatin C; and 6) cross-presentation of exogenous Ag to CD8 T cells. Furthermore, despite lacking surface CD8 expression, the CD24high subset contained CD8 mRNA and up-regulated surface expression when transferred into mice. This culture system allows access to bona fide counterparts of the splenic DC subsets.
Introduction
The delineation of dendritic cell (DC)4 subsets has highlighted their unique functional specializations, which play important roles in the control of immune responses (1). Three functionally distinct subsets can be defined in steady-state mouse spleen and include the plasmacytoid pre-DC (pDC), CD8+ and CD8– conventional DC (cDC) subsets, which appear to be distinct and not precursor-product related in the steady-state (2, 3, 4). The diminutive numbers of DC, and difficulty in their isolation, has often precluded study of their development and function. For example, at most 1 x 106 CD8+ cDC can be recovered with high purity from one mouse spleen after an elaborate purification protocol (5). A culture method for generating higher yields of these subtypes would make them more accessible.
There are several well-established procedures for generating DC in culture from bone marrow (BM) precursors or from blood monocytes using GM-CSF and IL-4 (GM/IL-4 DC) (6). However, GM/IL-4 DC do not seem to show the heterogeneity in DC phenotype and function found with splenic DC. In fact, it is not clear whether GM/IL-4 DC have any counterparts among steady-state DC in vivo. In relation to this, a recent study found the development of steady-state DC was Stat3 dependent, whereas GM/IL-4 DC development from monocytes was Stat3 independent, suggesting that these two groups of DC have distinct developmental origins (7).
Culture of BM with fms-like tyrosine kinase 3 (flt3) ligand (FL) is a more recent method that allows the generation of both cDC and pDC in large numbers (now referred to as FL-DC) (8, 9, 10). The dependency on FL for DC generation in this culture system correlates with the observations that FL–/– mice have reduced DC numbers (11), that injection of mice with FL greatly increases DC numbers (12, 13), and that only early BM progenitors that express flt3 are effective precursors of DC (14). Some initial observations suggested that FL-cDC may further divide into subsets, but their relationship to splenic DC subsets was not further investigated (10, 15).
In this study, we show that despite the presence of some CD11b on all cDC from FL cultures, and the absence of surface CD4 and CD8 expression, FL-DC could be clearly segregated by surface markers into the equivalents of steady-state splenic pDC, CD8+ cDC, and CD8– cDC. The shared properties between the FL-DC subsets and the splenic DC subset counterparts included surface marker expression, transcription factor expression and dependence for development, ability to cross-present cellular Ag to CD8 T cells, expression of TLR and chemokine receptors, and production of cytokines and chemokines in response to TLR stimulation. This system should allow access to large numbers of the DC subsets for further study. In particular, up to 25 x 106 of the CD8+ cDC equivalents can be generated from culturing the BM of one mouse with FL.
Materials and Methods
ELISAs were performed as previously described (21) with capture mAbs to IL-12 p40/p70 (C16.6 or R29A5), IL-6 (MP5-20F3), RANTES (53433), MIP1 (39624.11), or IFN- (RMMA-1), and detection via biotinylated mAbs to IL12 p40/p70 (C17.8), IL-6 (M35-32C11), RANTES (R&D Systems), MIP-1 (R&D Systems), or polyclonal rabbit Ab to IFN- (PBL).
Results and Discussion
Phenotypic characterization of FL-DC and splenic DC subsets
A total of 5 x 106 DC is usually isolated from one mouse spleen, whereas 50 x 106 DC could be generated from culturing the BM of one mouse with FL (Fig. 1A). To assess phenotype, DC derived from FL-stimulated BM cultures were compared in surface Ag expression to freshly isolated splenic DC. Both FL culture-derived and splenic CD11c+ DC contained distinct CD45RA– cDC and CD45RA+ pDC populations (Fig. 1A), as others have noted previously (8, 9, 16). However, because the cDC component had been shown to express both surface CD11b and CD205 (8), but not CD4 or CD8, the markers normally used to segregate the splenic CD8+ and CD8– cDC subtypes were not applicable for FL-cDC subset discrimination (1, 17) (Fig. 1).
To assess whether FL-cDC nevertheless contained subsets, we analyzed for other surface markers differentially expressed by spleen cDC. These included CD24, a surface molecule selectively expressed by CD8+ cDC (2, 5, 22), and SIRP-, recently identified by microarrays as being selectively expressed by CD8– cDC (M. Lahoud, A. Proietto, and K. Shortman, manuscript in preparation) and differentially expressed among rat DC (23). FL-cDC expressed these markers, as well as CD11b, at different levels, which allowed their further division into two subtypes: CD24highSIRP-lowCD11blow and CD24lowSIRP-highCD11bhigh (Fig. 1A, top panels). We found that four-color staining for CD11c, CD45RA, CD11b, and CD24 was the ideal combination for discrimination of the three FL-DC subtypes.
An almost identical separation of subsets using the same combination of markers was found with splenic DC, but with differences in the relative proportions of each (Fig. 1A, lower panels). In particular, there was a larger proportion of CD24high cells among the FL-DC, the putative CD8+ cDC equivalent. Also, whereas the splenic cDC subsets are selective in their surface expression of CD11b and CD205 in the steady state (5), all FL-cDC expressed some levels of each (Fig. 1A) (8, 9). This is not surprising considering that all splenic cDC up-regulate CD11b and CD205 upon in vitro culture (17). In view of these similarities in aspects of surface phenotype, we tested whether the FL-DC and splenic DC populations were developmentally and functionally equivalent. The terminology used hereafter for the three subsets is pDC, CD24high DC (CD8+ cDC equivalents), and CD11bhigh DC (CD8– cDC equivalents).
DC subset differences in IRF-8–/– mice
IRF-8 transcription factor knockout mice retain splenic CD8– cDC but have reduced numbers of pDC and CD8+ cDC (Fig. 1B, lower panel) (24). To assess whether a related defect occurred within FL-DC, we cultured BM from IRF-8–/– and wild-type (WT) mice with FL and compared the DC produced. Among the IRF8–/– FL-DC, there was comparable production of CD11bhigh DC but greatly reduced numbers of pDC and CD24high DC compared with WT FL-DC (Fig. 1B, top panel, and our unpublished data). Interestingly, a converse deficit in DC subset production has been reported using IRF-4–/– mice, which have an absence of CD4+ DC in vivo and an absence of CD11bhigh DC in FL BM cultures (15).
IRF expression
Not only do the different spleen DC subsets require particular IRFs for their normal development, they also vary in their expression of these transcription factors in the steady state: the CD8+ DC mainly express IRF-8, CD8– DC mainly express IRF-4, whereas pDC express both (15, 25). The FL-DC subsets were examined for IRF-4 and IRF-8 mRNA. Indeed, the expression pattern of IRFs in FL-DC subsets correlated with those in splenic DC (Fig. 2A), in line with previous studies (15).
Expression of CD4 and CD8
Even though unstimulated FL-DC do not express surface CD4 or CD8, unlike their putative in vivo counterparts, mRNA expression of these molecules was tested. CD24high DC and pDC did express some CD8 transcript, whereas CD11bhigh DC did not. Only pDC expressed CD4 transcript (Fig. 2A). To see whether surface expression might be dependent on the normal in vivo DC environment, we transferred each purified FL-DC subset in vivo. After 3 days, we compared host and donor-derived splenic DC for surface CD4 and CD8 expression (Fig. 2B). Almost 75% of the recovered CD24high DC had up-regulated CD8, but very few expressed CD4. Transferred FL-pDC up-regulated both CD4 and CD8, so they then resembled their in vivo counterparts (2). However, transferred CD11bhigh DC up-regulated very little surface CD4 or CD8 in this time (15).
Ag presentation and cross-presentation
To assess the capacity of the cDC subsets to activate naive T cells, DC were coated with synthetic OVA peptides for MHC I or II, and then incubated with OT-I CD8 T cells or OT-II CD4 T cells, respectively. FL-DC showed comparable efficiency to their splenic DC counterparts in stimulating the proliferation of naive CD8 (Fig. 3A) and CD4 T cells (B). Cross-presentation, a process whereby exogenous Ags are taken up and presented via MHC I, is conducted efficiently by splenic CD8+ cDC but not CD8– DC nor pDC (19, 26). To examine whether any of the FL-cDC populations were able to cross-present, we incubated each subset with OT-I cells and cellular Ag in the form of OVA protein-coated gamma-irradiated splenocytes from bm1 mice (unable to present OVA to OT-I cells). Of the FL-DC, only the CD24high subset was able to cross-present with efficiency similar to splenic CD8+ cDC (Fig. 3C).
CyC expression
In the steady state, the splenic CD8+ cDC are unique among the DC subsets in their expression of the cysteine protease inhibitor CyC (20), although the precise role of CyC in CD8+ cDC biology is still controversial (20, 27). We found that only CD24high FL-DC expressed significant levels of CyC (Fig. 3D).
TLR expression
The splenic DC subsets have unique TLR expression patterns, which enable them to directly respond to particular TLR ligands (28, 29). Splenic pDC express TLR7 and -9 but not TLR3 or -4; CD8+ cDC express TLR3, -4, and -9 but not TLR7; CD8– cDC express TLR4, -7, and -9, but not TLR3. Using real-time PCR, an almost identical pattern of TLR mRNA expression was found between the FL-DC and splenic DC subset equivalents (Fig. 4A).
Chemokine receptor expression
The spleen DC subsets show quantitative differences in chemokine receptor expression, with pDC being the highest expressors of CXCR3 and CCR9, and CD8– cDC the highest expressors of CCR6 and CX3CR1, whereas CD8+ cDC express some CXCR3 and CX3CR1 (29). We found comparable mRNA expression of most of these receptors between the FL-DC and splenic DC (Fig. 4B). One striking difference was the lack of CCR6 expression in CD11bhigh FL-DC compared with splenic CD8– cDC, which suggests an in vivo factor may be required for its normal expression.
Cytokine and chemokine production
A major functional difference between the splenic DC subsets is their capacity to produce particular cytokines. Cytokine production in response to TLR engagement depends on the TLR expression pattern of the DC subset, the TLR ligand, and on other factors such as the cytokine environment. Moreover, the splenic DC subsets differ in the cytokines they produce to the same TLR ligand (21, 29). For example, only pDC and CD8– cDC express TLR7 and produce cytokines in response to the TLR7 ligand, R848. In contrast, all splenic DC express TLR9, but when stimulated with the TLR9 ligand, CpG DNA, pDC produce the highest levels of IFN-, whereas CD8+ cDC produce the highest levels of IL-12 p70. FL-DC subsets were compared with their splenic DC counterparts for cytokine production, in response to R848, CpG 2216, or LPS, in conditions known to induce cytokine production, and the supernatants were analyzed by ELISA. The patterns of cytokine production by the FL-DC subsets reflected their splenic DC equivalents (Fig. 5A).
Chemokine production in response to the TLR7 ligand R848 was comparable between FL-DC and splenic DC (Fig. 5B). In particular, CD11bhigh DC were the major producer of RANTES, whereas pDC were the major producers of MIP-1. The CD24high DC subsets showed negligible chemokine or IL production in response to R848, correlating to their lack of TLR7 expression.
Conclusions
We have shown that FL BM cultures produce DC subsets that, despite differences in CD4 and CD8 expression, are close equivalents of the steady-state splenic pDC, CD8+ cDC, and CD8– cDC subtypes. Importantly, key functional differences are conserved between the FL-DC and spleen DC equivalents. The major novel finding in this study is the generation of a functionally distinct spleen CD8+ cDC equivalent, and a strategy for its isolation in high yield. We thus propose that, in contrast to GM/IL-4 DC, FL cultures better facilitate the study of steady-state DC subset ontogeny and function. Finally, FL-stimulated cultures may be useful for identifying and producing human DC counterparts to the murine DC subsets.
Acknowledgments
We thank Katya Gray, Jaclyn Carneli, and Kelly Trueman for outstanding animal husbandry, Viki Milovac, Catherine Tarlinton, Jenny Garbe, and Carley Young for first-rate flow cytometry, and Dorothee Neukirchen for excellent technical assistance.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by the Australian Government’s Cooperative Research Centres (CRC) Program and the National Health and Medical Research Council. S.H.N. is supported by scholarships from the University of Melbourne and the CRC for Vaccine Technology, and A.I.P. is supported by an Australian Postgraduate Award.
2 S.H.N. and A.I.P. contributed equally to this work.
3 Address correspondence and reprint requests to Dr. Shalin H. Naik or Dr. Li Wu, Walter and Eliza Hall Institute, 1G Royal Parade, Parkville, Victoria, Australia. E-mail addresses: naik{at}wehi.edu.au and wu{at}wehi.edu.au
4 Abbreviations used in this paper: DC, dendritic cell; pDC, plasmacytoid pre-DC; cDC, conventional DC; BM, bone marrow; GM/IL-4 DC, GM-CSF and IL-4 DC; FL, fms-like tyrosine kinase 3 (flt3) ligand; MHC I, MHC class I; MHC II, MHC class II; IRF, IFN regulatory factor; CyC, cystatin C; SIRP, signal regulatory protein; WT, wild type.
Received for publication March 15, 2005. Accepted for publication April 12, 2005.
References
Shortman, K., Y. J. Liu. 2002. Mouse and human dendritic cell subtypes. Nat. Rev. Immunol. 2: 151-161.
O’Keeffe, M., H. Hochrein, D. Vremec, I. Caminschi, J. L. Miller, E. M. Anders, L. Wu, M. H. Lahoud, S. Henri, B. Scott, et al 2002. Mouse plasmacytoid cells: long-lived cells, heterogeneous in surface phenotype and function, that differentiate into CD8+ dendritic cells only after microbial stimulus. J. Exp. Med. 196: 1307-1319.
Naik, S., D. Vremec, L. Wu, M. O’Keeffe, K. Shortman. 2003. CD8+ mouse spleen dendritic cells do not originate from the CD8– dendritic cell subset. Blood 102: 601-604.
Kamath, A. T., J. Pooley, M. A. O’Keeffe, D. Vremec, Y. Zhan, A. M. Lew, A. D’Amico, L. Wu, D. F. Tough, K. Shortman. 2000. The development, maturation, and turnover rate of mouse spleen dendritic cell populations. J. Immunol. 165: 6762-6770.
Vremec, D., J. Pooley, H. Hochrein, L. Wu, K. Shortman. 2000. CD4 and CD8 expression by dendritic cell subtypes in mouse thymus and spleen. J. Immunol. 164: 2978-2986.
Sallusto, F., A. Lanzavecchia. 1994. Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/macrophage colony-stimulating factor plus interleukin-4 and downregulated by tumor necrosis factor-. J. Exp. Med. 179: 1109-1118.(Shalin H. Naik2,3,*, Anna)
We demonstrate that functional and phenotypic equivalents of mouse splenic CD8+ and CD8– conventional dendritic cell (cDC) subsets can be generated in vitro when bone marrow is cultured with fms-like tyrosine kinase 3 (flt3) ligand. In addition to CD45RAhigh plasmacytoid DC, two distinct CD24high and CD11bhigh cDC subsets were present, and these subsets showed equivalent properties to splenic CD8+ and CD8– cDC, respectively, in the following: 1) surface expression of CD11b, CD24, and signal regulatory protein-; 2) developmental dependence on, and mRNA expression of, IFN regulatory factor-8; 3) mRNA expression of TLRs and chemokine receptors; 4) production of IL-12 p40/70, IFN-, MIP-1, and RANTES in response to TLR ligands; 5) expression of cystatin C; and 6) cross-presentation of exogenous Ag to CD8 T cells. Furthermore, despite lacking surface CD8 expression, the CD24high subset contained CD8 mRNA and up-regulated surface expression when transferred into mice. This culture system allows access to bona fide counterparts of the splenic DC subsets.
Introduction
The delineation of dendritic cell (DC)4 subsets has highlighted their unique functional specializations, which play important roles in the control of immune responses (1). Three functionally distinct subsets can be defined in steady-state mouse spleen and include the plasmacytoid pre-DC (pDC), CD8+ and CD8– conventional DC (cDC) subsets, which appear to be distinct and not precursor-product related in the steady-state (2, 3, 4). The diminutive numbers of DC, and difficulty in their isolation, has often precluded study of their development and function. For example, at most 1 x 106 CD8+ cDC can be recovered with high purity from one mouse spleen after an elaborate purification protocol (5). A culture method for generating higher yields of these subtypes would make them more accessible.
There are several well-established procedures for generating DC in culture from bone marrow (BM) precursors or from blood monocytes using GM-CSF and IL-4 (GM/IL-4 DC) (6). However, GM/IL-4 DC do not seem to show the heterogeneity in DC phenotype and function found with splenic DC. In fact, it is not clear whether GM/IL-4 DC have any counterparts among steady-state DC in vivo. In relation to this, a recent study found the development of steady-state DC was Stat3 dependent, whereas GM/IL-4 DC development from monocytes was Stat3 independent, suggesting that these two groups of DC have distinct developmental origins (7).
Culture of BM with fms-like tyrosine kinase 3 (flt3) ligand (FL) is a more recent method that allows the generation of both cDC and pDC in large numbers (now referred to as FL-DC) (8, 9, 10). The dependency on FL for DC generation in this culture system correlates with the observations that FL–/– mice have reduced DC numbers (11), that injection of mice with FL greatly increases DC numbers (12, 13), and that only early BM progenitors that express flt3 are effective precursors of DC (14). Some initial observations suggested that FL-cDC may further divide into subsets, but their relationship to splenic DC subsets was not further investigated (10, 15).
In this study, we show that despite the presence of some CD11b on all cDC from FL cultures, and the absence of surface CD4 and CD8 expression, FL-DC could be clearly segregated by surface markers into the equivalents of steady-state splenic pDC, CD8+ cDC, and CD8– cDC. The shared properties between the FL-DC subsets and the splenic DC subset counterparts included surface marker expression, transcription factor expression and dependence for development, ability to cross-present cellular Ag to CD8 T cells, expression of TLR and chemokine receptors, and production of cytokines and chemokines in response to TLR stimulation. This system should allow access to large numbers of the DC subsets for further study. In particular, up to 25 x 106 of the CD8+ cDC equivalents can be generated from culturing the BM of one mouse with FL.
Materials and Methods
ELISAs were performed as previously described (21) with capture mAbs to IL-12 p40/p70 (C16.6 or R29A5), IL-6 (MP5-20F3), RANTES (53433), MIP1 (39624.11), or IFN- (RMMA-1), and detection via biotinylated mAbs to IL12 p40/p70 (C17.8), IL-6 (M35-32C11), RANTES (R&D Systems), MIP-1 (R&D Systems), or polyclonal rabbit Ab to IFN- (PBL).
Results and Discussion
Phenotypic characterization of FL-DC and splenic DC subsets
A total of 5 x 106 DC is usually isolated from one mouse spleen, whereas 50 x 106 DC could be generated from culturing the BM of one mouse with FL (Fig. 1A). To assess phenotype, DC derived from FL-stimulated BM cultures were compared in surface Ag expression to freshly isolated splenic DC. Both FL culture-derived and splenic CD11c+ DC contained distinct CD45RA– cDC and CD45RA+ pDC populations (Fig. 1A), as others have noted previously (8, 9, 16). However, because the cDC component had been shown to express both surface CD11b and CD205 (8), but not CD4 or CD8, the markers normally used to segregate the splenic CD8+ and CD8– cDC subtypes were not applicable for FL-cDC subset discrimination (1, 17) (Fig. 1).
To assess whether FL-cDC nevertheless contained subsets, we analyzed for other surface markers differentially expressed by spleen cDC. These included CD24, a surface molecule selectively expressed by CD8+ cDC (2, 5, 22), and SIRP-, recently identified by microarrays as being selectively expressed by CD8– cDC (M. Lahoud, A. Proietto, and K. Shortman, manuscript in preparation) and differentially expressed among rat DC (23). FL-cDC expressed these markers, as well as CD11b, at different levels, which allowed their further division into two subtypes: CD24highSIRP-lowCD11blow and CD24lowSIRP-highCD11bhigh (Fig. 1A, top panels). We found that four-color staining for CD11c, CD45RA, CD11b, and CD24 was the ideal combination for discrimination of the three FL-DC subtypes.
An almost identical separation of subsets using the same combination of markers was found with splenic DC, but with differences in the relative proportions of each (Fig. 1A, lower panels). In particular, there was a larger proportion of CD24high cells among the FL-DC, the putative CD8+ cDC equivalent. Also, whereas the splenic cDC subsets are selective in their surface expression of CD11b and CD205 in the steady state (5), all FL-cDC expressed some levels of each (Fig. 1A) (8, 9). This is not surprising considering that all splenic cDC up-regulate CD11b and CD205 upon in vitro culture (17). In view of these similarities in aspects of surface phenotype, we tested whether the FL-DC and splenic DC populations were developmentally and functionally equivalent. The terminology used hereafter for the three subsets is pDC, CD24high DC (CD8+ cDC equivalents), and CD11bhigh DC (CD8– cDC equivalents).
DC subset differences in IRF-8–/– mice
IRF-8 transcription factor knockout mice retain splenic CD8– cDC but have reduced numbers of pDC and CD8+ cDC (Fig. 1B, lower panel) (24). To assess whether a related defect occurred within FL-DC, we cultured BM from IRF-8–/– and wild-type (WT) mice with FL and compared the DC produced. Among the IRF8–/– FL-DC, there was comparable production of CD11bhigh DC but greatly reduced numbers of pDC and CD24high DC compared with WT FL-DC (Fig. 1B, top panel, and our unpublished data). Interestingly, a converse deficit in DC subset production has been reported using IRF-4–/– mice, which have an absence of CD4+ DC in vivo and an absence of CD11bhigh DC in FL BM cultures (15).
IRF expression
Not only do the different spleen DC subsets require particular IRFs for their normal development, they also vary in their expression of these transcription factors in the steady state: the CD8+ DC mainly express IRF-8, CD8– DC mainly express IRF-4, whereas pDC express both (15, 25). The FL-DC subsets were examined for IRF-4 and IRF-8 mRNA. Indeed, the expression pattern of IRFs in FL-DC subsets correlated with those in splenic DC (Fig. 2A), in line with previous studies (15).
Expression of CD4 and CD8
Even though unstimulated FL-DC do not express surface CD4 or CD8, unlike their putative in vivo counterparts, mRNA expression of these molecules was tested. CD24high DC and pDC did express some CD8 transcript, whereas CD11bhigh DC did not. Only pDC expressed CD4 transcript (Fig. 2A). To see whether surface expression might be dependent on the normal in vivo DC environment, we transferred each purified FL-DC subset in vivo. After 3 days, we compared host and donor-derived splenic DC for surface CD4 and CD8 expression (Fig. 2B). Almost 75% of the recovered CD24high DC had up-regulated CD8, but very few expressed CD4. Transferred FL-pDC up-regulated both CD4 and CD8, so they then resembled their in vivo counterparts (2). However, transferred CD11bhigh DC up-regulated very little surface CD4 or CD8 in this time (15).
Ag presentation and cross-presentation
To assess the capacity of the cDC subsets to activate naive T cells, DC were coated with synthetic OVA peptides for MHC I or II, and then incubated with OT-I CD8 T cells or OT-II CD4 T cells, respectively. FL-DC showed comparable efficiency to their splenic DC counterparts in stimulating the proliferation of naive CD8 (Fig. 3A) and CD4 T cells (B). Cross-presentation, a process whereby exogenous Ags are taken up and presented via MHC I, is conducted efficiently by splenic CD8+ cDC but not CD8– DC nor pDC (19, 26). To examine whether any of the FL-cDC populations were able to cross-present, we incubated each subset with OT-I cells and cellular Ag in the form of OVA protein-coated gamma-irradiated splenocytes from bm1 mice (unable to present OVA to OT-I cells). Of the FL-DC, only the CD24high subset was able to cross-present with efficiency similar to splenic CD8+ cDC (Fig. 3C).
CyC expression
In the steady state, the splenic CD8+ cDC are unique among the DC subsets in their expression of the cysteine protease inhibitor CyC (20), although the precise role of CyC in CD8+ cDC biology is still controversial (20, 27). We found that only CD24high FL-DC expressed significant levels of CyC (Fig. 3D).
TLR expression
The splenic DC subsets have unique TLR expression patterns, which enable them to directly respond to particular TLR ligands (28, 29). Splenic pDC express TLR7 and -9 but not TLR3 or -4; CD8+ cDC express TLR3, -4, and -9 but not TLR7; CD8– cDC express TLR4, -7, and -9, but not TLR3. Using real-time PCR, an almost identical pattern of TLR mRNA expression was found between the FL-DC and splenic DC subset equivalents (Fig. 4A).
Chemokine receptor expression
The spleen DC subsets show quantitative differences in chemokine receptor expression, with pDC being the highest expressors of CXCR3 and CCR9, and CD8– cDC the highest expressors of CCR6 and CX3CR1, whereas CD8+ cDC express some CXCR3 and CX3CR1 (29). We found comparable mRNA expression of most of these receptors between the FL-DC and splenic DC (Fig. 4B). One striking difference was the lack of CCR6 expression in CD11bhigh FL-DC compared with splenic CD8– cDC, which suggests an in vivo factor may be required for its normal expression.
Cytokine and chemokine production
A major functional difference between the splenic DC subsets is their capacity to produce particular cytokines. Cytokine production in response to TLR engagement depends on the TLR expression pattern of the DC subset, the TLR ligand, and on other factors such as the cytokine environment. Moreover, the splenic DC subsets differ in the cytokines they produce to the same TLR ligand (21, 29). For example, only pDC and CD8– cDC express TLR7 and produce cytokines in response to the TLR7 ligand, R848. In contrast, all splenic DC express TLR9, but when stimulated with the TLR9 ligand, CpG DNA, pDC produce the highest levels of IFN-, whereas CD8+ cDC produce the highest levels of IL-12 p70. FL-DC subsets were compared with their splenic DC counterparts for cytokine production, in response to R848, CpG 2216, or LPS, in conditions known to induce cytokine production, and the supernatants were analyzed by ELISA. The patterns of cytokine production by the FL-DC subsets reflected their splenic DC equivalents (Fig. 5A).
Chemokine production in response to the TLR7 ligand R848 was comparable between FL-DC and splenic DC (Fig. 5B). In particular, CD11bhigh DC were the major producer of RANTES, whereas pDC were the major producers of MIP-1. The CD24high DC subsets showed negligible chemokine or IL production in response to R848, correlating to their lack of TLR7 expression.
Conclusions
We have shown that FL BM cultures produce DC subsets that, despite differences in CD4 and CD8 expression, are close equivalents of the steady-state splenic pDC, CD8+ cDC, and CD8– cDC subtypes. Importantly, key functional differences are conserved between the FL-DC and spleen DC equivalents. The major novel finding in this study is the generation of a functionally distinct spleen CD8+ cDC equivalent, and a strategy for its isolation in high yield. We thus propose that, in contrast to GM/IL-4 DC, FL cultures better facilitate the study of steady-state DC subset ontogeny and function. Finally, FL-stimulated cultures may be useful for identifying and producing human DC counterparts to the murine DC subsets.
Acknowledgments
We thank Katya Gray, Jaclyn Carneli, and Kelly Trueman for outstanding animal husbandry, Viki Milovac, Catherine Tarlinton, Jenny Garbe, and Carley Young for first-rate flow cytometry, and Dorothee Neukirchen for excellent technical assistance.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by the Australian Government’s Cooperative Research Centres (CRC) Program and the National Health and Medical Research Council. S.H.N. is supported by scholarships from the University of Melbourne and the CRC for Vaccine Technology, and A.I.P. is supported by an Australian Postgraduate Award.
2 S.H.N. and A.I.P. contributed equally to this work.
3 Address correspondence and reprint requests to Dr. Shalin H. Naik or Dr. Li Wu, Walter and Eliza Hall Institute, 1G Royal Parade, Parkville, Victoria, Australia. E-mail addresses: naik{at}wehi.edu.au and wu{at}wehi.edu.au
4 Abbreviations used in this paper: DC, dendritic cell; pDC, plasmacytoid pre-DC; cDC, conventional DC; BM, bone marrow; GM/IL-4 DC, GM-CSF and IL-4 DC; FL, fms-like tyrosine kinase 3 (flt3) ligand; MHC I, MHC class I; MHC II, MHC class II; IRF, IFN regulatory factor; CyC, cystatin C; SIRP, signal regulatory protein; WT, wild type.
Received for publication March 15, 2005. Accepted for publication April 12, 2005.
References
Shortman, K., Y. J. Liu. 2002. Mouse and human dendritic cell subtypes. Nat. Rev. Immunol. 2: 151-161.
O’Keeffe, M., H. Hochrein, D. Vremec, I. Caminschi, J. L. Miller, E. M. Anders, L. Wu, M. H. Lahoud, S. Henri, B. Scott, et al 2002. Mouse plasmacytoid cells: long-lived cells, heterogeneous in surface phenotype and function, that differentiate into CD8+ dendritic cells only after microbial stimulus. J. Exp. Med. 196: 1307-1319.
Naik, S., D. Vremec, L. Wu, M. O’Keeffe, K. Shortman. 2003. CD8+ mouse spleen dendritic cells do not originate from the CD8– dendritic cell subset. Blood 102: 601-604.
Kamath, A. T., J. Pooley, M. A. O’Keeffe, D. Vremec, Y. Zhan, A. M. Lew, A. D’Amico, L. Wu, D. F. Tough, K. Shortman. 2000. The development, maturation, and turnover rate of mouse spleen dendritic cell populations. J. Immunol. 165: 6762-6770.
Vremec, D., J. Pooley, H. Hochrein, L. Wu, K. Shortman. 2000. CD4 and CD8 expression by dendritic cell subtypes in mouse thymus and spleen. J. Immunol. 164: 2978-2986.
Sallusto, F., A. Lanzavecchia. 1994. Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/macrophage colony-stimulating factor plus interleukin-4 and downregulated by tumor necrosis factor-. J. Exp. Med. 179: 1109-1118.(Shalin H. Naik2,3,*, Anna)