Plasmacytoid dendritic cells
Cutaneous lymphocyte antigen
Monocyte-derived dendritic cells
Plasmacytoid dendritic cells (PDC) are a small population of leukocytes specialized in the production of type I IFN. It has been shown that PDC have a potent T cell stimulatory capacity in allogeneic mixed lymphocyte reaction, However, their role in initiating primary immune responses remains elusive. We report that blood PDC efficiently prime naive CD8+ lymphocytes specific forthe melan-A26–35 epitope to become IFN-γ producing cells in vitro. In addition, we found that CD40L-stimulated PDC induce expression on primed melan-A-specific T cells of cutaneous lymphocyte antigen and L-selectin (CD62L), homing receptors that allow the migration of effector cells to the inflamed skin. Finally, we show that PDC can be found in the peri-tumoralarea of most primary cutaneous melanomas in vivo and that type I IFN-containing supernatants derived from PDC increase melanoma cell surface expression of CD95 and MHC class I and class II molecules in vitro. Our results suggest a new immunomodulatory role for tissue infiltrating PDC, which may prime tumor-specific T cell responses and affect tumor growth via soluble factors.
Dendritic cells (DC) are the most potent antigen-presenting cells (APC) unique in their ability to efficiently prime both CD4+ helper and CD8+ cytotoxic T cell responses following encounter with antigens and maturation 1. In human blood, at least two subsets of primary DC can be identified, which are morphologically, phenotypically and functionally distinct 2: myeloid DC CD11c+, and plasmacytoid DC (PDC) CD11c–. PDC show a typical plasma cell-like morphology, lack expression of myeloid markers such as CD13, CD33 and the activating receptor ILT1. In contrast, PDC express high levels of IL-3-receptor α chain (IL-3Rα; CD123), BDCA-2, CD68, CD36 and the inhibitory receptor ILT3 3–6. Cells with similar morphology, defined by pathologist as "plasmacytoid monocytes", are found around high endothelial venules (HEV) of inflamed lymph nodes and their numbers are increased in several diseases 7. Expression of the chemokine receptor CXCR3 and L-selectin (CD62L) may allow PDC migration from the blood stream into inflamed lymph nodes through HEV 5.
Upon stimulation with viruses, PDC produce high amounts of type I IFN 5, 8–10 and undergo maturation. Mature PDC efficiently stimulate T cells and drive a type I-polarized T cell response, mediated by the synergistic effect of type I IFN and IL-12 11.
It has been shown that mature PDC induce proliferation of allogeneic CD4+ CD45RA+ naive T cells 4, 5, 12. However, their role in inducing antigen-specific CD8+ T cell responses remains elusive. In this study, we have for the first time investigated the capacity of peripheral blood PDC to in vitro prime CD8+ lymphocytes specific for the melanoma antigen melan-A/MART1, that encodes for the HLA-A*0201-restricted immuno-dominant epitope 26–35 13. In addition, we showed infiltration of primary melanoma lesions with PDC, which may therefore have an important immuno-modulatory role in vivo.
2.1 Priming of type I-polarized melan-A-specific CD8+ lymphocytes by PDC
Naive melan-A tetramer+ cells can be detected in the peripheral blood of about 50% of HLA-A2+ healthy blood donors 14, 15. These cells can be efficiently expanded into effector CD8+ CTL by professional APC, such as monocyte-derived dendritic cells (Mo-DC), pulsed with low dose of peptide 16. We therefore asked whether blood-derived PDC (Fig. 1) behave as bona fide DC in expanding antigen-specific CD8+ T cells and compared the priming capacity of PDC to that of Mo-DC. We observed that immature and CD40L-matured PDC expanded a significant population of melan-A tetramer+ CD8+ cells (Fig. 2A). About 30% of the tetramer+ cells expressed CD62L, a receptor involved in T cell homing to the lymph nodes and to inflamed tissues (Fig. 2B). Co-expression of CD62L and of the skin homing receptor cutaneous lymphocyte antigen (CLA) was observed on one third of the melan-A tetramer+ T cells primed by CD40L-matured PDC but not by immature PDC (Fig. 2B). Similar results were obtained with immature versus CD40L-stimulated Mo-DC (Fig. 3A and B). In five out of nine donors both DC types induced equal numbers of melan-A tetramer+ CD8+ T cells, while in the other four donors a larger expansion of antigen-specific CD8+ T cells was obtained using Mo-DC as APC (data not shown).
To assess their capacity to present viral antigens through the MHC class I pathway, we infected PDC with the replication defective MVA virus encoding the melan-A26–35 epitope. Although less efficient than PDC pulsed with 100 nM peptide, MVA-infected PDC induced a fivefold expansion of melan-A tetramer+ T cells, that all became CD45RO+ (Table 1).
We next investigated the cytokine producing capacity of the expanded CD8+ T cells as a measure of their acquired effector function. Most of the CD8+ T cells expanded by immature as well as CD40L-activated PDC produced IFN-γ upon peptide stimulation (Fig. 2C). As a control, we observed that mature Mo-DC also induced a strong type I polarization of melan-A tetramer+ T cells, while immature Mo-DC were less effective in this respect and lower amounts of IFN-γ were secreted on a per cell basis (Fig. 3C). The differences between immature PDC and Mo-DC, as well as the variability in the overall expansion of antigen-specific CD8+ T cells can be explained by spontaneous PDC activation occurring in some donors. Recently activated CD40L+ CD4+ cells, that may be present in some donors but not in others, induce PDC maturation and enhance proliferation and polarization of antigen-specific CD8+ T cells.
|% Melan-A+ CD8+||Phenotype|
2.2 Type I IFN and IL-12 induce CLA expression on melan-A tetramer+ CD8+ T cells
It has been shown that PDC produce type I IFN after stimulation with viruses and CD40L 5, 8–10. In addition, both PDC and Mo-DC produce bioactive IL-12 after stimulation with CD40L 11, 17. We therefore asked whether IL-12 and/or type I IFN could induce CLA expression on the antigen-specific CD8+ T cells primed by CD40L-matured PDC and Mo-DC. To test this hypothesis, we primed melan-A-specific T cells using peptide-pulsed PBL as APC, either in the presence or in the absence of IL-12 and IFN-α. We have previously shown that a high concentration of peptide (100 μg/ml) is required to prime naive melan-A-specific T cells in the absence of professional APC 16. Although peptide-pulsed PBMC were capable of priming melan-A-specific T cells even in the absence of IL-12 and IFN-α, these cytokines were required to obtain CLA expression (Fig. 4A and B). Notably, concentration of IFN-α higher than 20 UI/ml had an inhibitory effect on the expression of melan-A-specific T cells (data not shown). The percentages of melan-A tetramer+ T cells CD62L+CLA+ induced in the presence of exogenous IL-12 and IFN-α were always lower than those induced by mature PDC, suggesting that additional factors or other members of the IFN family may selectively expand this population. Both IL-12 and IFN-α induced IFN-γ-secreting melan-A-specific T cells (Fig. 4C), IL-12 being generally more effective than IFN-α, as reported 18. In contrast, in the absence of these cytokines melan-A-specific T cells remained largely unpolarized.
Finally, we assessed production of bioactive IL-12 and type I IFN in the supernatants of CD40L-treated PDC and Mo-DC (Fig. 5). While PDC produced both IL-12 and type I IFN in response to CD40 ligation, Mo-DC only secreted IL-12, although six-to-ten times more than PDC. Very high amounts of type I IFN were produced by PDC challenged with viral stimuli, such as infection with influenza virus.
2.3 Activated PDC up-regulate MHC and CD95 expression on melanoma cells
To investigate the effect of PDC activation on melanoma cells, we incubated a panel of melanoma lines with supernatants from immature and mature PDC for 48 h. We observed that supernatants of CD40L- and influenza virus-matured PDC induced a significant up-regulation of MHC class I and class II molecules on melanoma cells (Fig. 6 and Table 2). This effect was only in part mimicked by exogenous type I IFN (Table 2), suggesting that other members of the IFN family, such as either IFN-ω, the recently described IFN-λ 19, 20, or additional soluble factors may be responsible. Incubation with PDC supernatants up-regulated CD95 expression on a large proportion of the melanoma lines tested (Fig. 6 and Table 3). However, we observed that the level of CD95 up-regulation varied for different cell lines and in cells with no basal level of CD95 expression, despite a marked increase of MHC class I molecules, CD95 was not expressed upon exposure to PDC supernatants or recombinant IFN-α (Table 3 and data not shown).
Altogether, our results suggest that by increasing CD95 and MHC class I and class II expression, PDC might render tumor cells more susceptible to CTL recognition.
|Cell line||Treatment||MHC class I||MHC class II|
|MFJ||% positive cells||MFI|
|IFN-α 1,000 U/ml||1,151||44||391|
|IFN-α 1,000 U/ml||3,118||100||1,374|
|IFN-α||PDC||PDC CD40L||PDC Flu|
|SK Mel-29 (19)||53||16||184||80|
|SK Mel-31 (424)||38||10||n.d.||30|
2.4 Plasmacytoid DC and CLA+ T cells are found in melanoma lesions
To assess the in vivo relevance of our in vitro priming results, we looked for PDC in primary cutaneous melanomas by immunohistochemistry. PDC were identified as medium size, round cells with pale cytoplasm, oval to round nuclei and fine chromatin (data not shown), as previously observed in reactive lymphadenitis and tonsil 21. They expressed CD123/IL-3Rα (Fig. 7A), CD68 and CLA/HECA452 (data not shown). We confirmed their identity by staining for BDCA-2 (Fig. 7C), a recently identified specific marker for tissue PDC 6, which was co-expressed with CD123 on serial sections (data not shown). PDC were predominantly located in peritumoral areas, but, in some cases, they were found in close association with the malignant melan-A+/MART-1+ cells (compare arrows in Fig. 7A and B). While PDC were absent in normal skin and nevi (data not shown), they were detected in the majority of melanomas, their numbers being higher in infiltrating and metastatic specimen as compared to in situ tumors. These results show that PDC might be recruited to tumor sites and their close association with melanoma cells in the tissue specimen suggests a possible role of PDC in modulating the anti-tumor immune response.
In addition to PDC, a CD8+ T cell infiltrate is evident within the melanoma lesion (Fig. 7D). Notably, a fraction of the CD8+ cells also co-express the CLA-skin homing receptor (Fig. 7E), conferring an in vivo relevance to the phenotype observed on the in vitro primed CD8+ T cells.
To extend our in vitro studies on the up-regulation of CD95 expression by melanoma lines, we analyzed by immunohistochemistry five primary cutaneous melanomas. We observed that CD95 expression in primary cutaneous melanomas was often weak and cytosolic, rather than on the cell membrane (Fig. 8B and insert). In contrast, CD95 was strongly and homogeneously expressed both in the cytoplasm and on the cell membrane in a melanocitic naevus and Burkitt lymphoma cells (Fig. 8A and C). These results are consistent with the possibility that a weak CD95 expression by melanoma cells in vivo could be due to a combination of poor activation state of intra-tumor PDC 22 and tumor evasion mechanisms to escape CD95-dependent lysis by CTL 23, 24.
In this study we have for the first time investigated the capacity of immature and mature blood-derived PDC to prime antigen-specific naive CD8+ T cells. As a model antigen we used melan-A/MART1, which is expressed in normal melanocytes and in most melanomas and encodes for the HLA-A*0201-restricted immuno-dominant epitope 26–35. We have previously analyzed the ability of Mo-DC to prime melan-A-specific naive T cells 16. We have now extended these results by showing that peptide-pulsed PDC are in many cases as efficient as Mo-DC in expanding melan-A tetramer+ T cells (Fig. 2 and 3). We demonstrated for the first time that virus-infected PDC are able to present an endogenous antigen through the MHC class I pathway, as defined by their ability to expand naive melan-A-specific T cells upon infection with the replication defective MVA virus encoding the melan-A26–35 epitope (Table 1). The lower level of expansion induced by virus-infected PDC as compared to peptide-pulsed PDC is likely due to the anti-proliferative effect of high levels of type I IFN secreted by infected PDC. To this regard, high titers of type I IFN may be relevant to antiviral protection but inhibit T cell growth, while lower amounts (such as after CD40L engagement) may promote PDC and T cell survival without affecting T cell proliferation 25, 26.
We showed that CD40L-matured PDC significantly, and even more efficiently than Mo-DC, expand antigen-specific CD8+ T cells co-expressing CLA and CD62L, homing receptors which enable T cell migration to the inflamed skin 27 (Fig. 2). It has been reported that skin T cells are highly enriched for the CLA+ CD62L+ memory/ effector subset 28 and we showed that a fraction of the CD8+ cells infiltrating melanoma lesions co-express CLA, therefore confirming the in vivo relevance of our in vitro priming results (Fig. 7).
There is evidence that type I IFN and IL-12 have a synergistic effect in inducing CD62L and CLA expression on CD4+ cells 29, 30. In agreement with these data, we showed that melan-A tetramer+ CD8+ T cells expanded in the presence of exogenous IL-12 and IFN-α acquired expression of CLA and CD62L (Fig. 4). We therefore suggest that CD40L-matured PDC may influence homing-receptor expression on T cells by secreting both cytokines 5, 17.
In humans, type I IFN and IL-12 are strong inducers of type 1 immune responses 31. Consistently, both immature and mature PDC induced IFN-γ-secreting antigen-specific CD8+ T cells in our system (Fig. 2) while a strong type I polarization was induced by mature DC, as reported 16, 32 (Fig. 3). Neither PDC nor Mo-DC pulsed with 100 nM of peptide expanded autologous IL-4- or IL-10-producing melan-A tetramer+ T cells in any of the donors tested (data not shown).
There is a growing interest in using DC as natural adjuvants in the design of cancer vaccines 33, 34, as they are powerful priming reagents for humoral and cellular immune responses. Infiltration of DC within tumor masses has been observed in several types of human tumors 35, although this does not necessarily correlate with an effective anti-tumor immune response. Indeed, tumor-associated DC often have an immature phenotype and their antigen-presenting function may be inhibited by soluble factors released by some tumors 36. We found BDCA-2+, CD123+ PDC in the peritumoral area of the majority of primary and metastatic melanomas, and, occasionally, close association of PDC with the malignant cells (Fig. 7). PDC were not present in normal tissues and their numbers increased with the severity of the disease 22. The close association of PDC and melanoma cells in the tissue specimen suggests that PDC, whenever appropriately activated, might contribute to the anti-tumor immune response following antigen uptake through the BDCA-2 receptor 6 or CD36 37, although previous work has shown that PDC have a poor endocytic activity 4, 38.
PDC recruitment to the tumor site in the absence of inflammation is likely to be due to chemokines secreted by the melanoma cells 39. Freshly isolated melanoma cells 40 and some melanoma lines 41 secrete SDF-1, and CXCR4-positive PDC can effectively migrate in response to SDF-1 42, 43. Indeed, wehave been able to detect SDF-1 constitutive expression in about one third of our melanoma lines (data not shown and 22).
Most of the tumor-associated PDC were CD83–, expressed high levels of BDCA-2 and low levels of type I IFN 22, which may be indicative of poor functional activation and lack of maturation. To this regard, we have observed increased expression of CD95 and MHC class I and class II molecules on melanoma cells exposed in vitro to supernatants of activated PDC (Fig. 6 and Table 2 and 3) and we speculated that in vivo this could render melanoma cells more susceptible to CTL recognition 44. However, consistent with a sub-optimal in vivo PDC activation, we detected weak and only cytoplasmic CD95 expression on cutaneous melanomas (Fig. 8). This result confirms published data, suggesting progressive lost of CD95 expression as a possible tumor escape mechanism and a negative prognostic factor 23, 24.
In conclusion, our results show that PDC efficiently prime type I-polarized antigen-specific CD8+ T cells, that also express homing receptors that enable them to reach the inflamed skin. PDC are present in the majority of melanoma lesions, where they could influence the tumor growth, antigen presentation as well as survival of memory T cells through type I IFN secretion 45. In addition, type I IFN could limit tumor growth through a well-known anti-proliferative effect 46 and could potentiate the anti-angiogenic effect of T cell-derivedIFN-γ 47. We therefore speculate that therapies aimed at enhancing both PDC and myeloid DC recruitment and activation at the tumor sites may best accomplish efficient anti-tumor immunity, by bridging innate and adaptive responses 48.
4 Materials and methods
4.1 Cell lines and cultures
The medium used throughout was RPMI 1640 supplemented with 2 mM L-glutamine, 1% nonessential amino acids, 1% pyruvate, 50 μg/ml kanamycin, 5×10–5 M 2ME (Gibco Laboratories, Grand Island, NY) and 10% FCS (Hyclone). Melanoma lines Mel-Brunner, Mel-1477, MZ2, D10, SK mel-31 and SKmel-29 were a gift of Dr. Elisabetta Padovan (Kantonspital, Basel, Switzerland). Melanoma NW-37 was a gift of Dr. A. Knuth (Medizinische Klinik, Frankfurt, Germany); Mel Smith was a gift of Dr. M. Collins (Guy's Hospital, London, GB). Melanomas MM9 and AFJ were generated in our laboratory. Recombinant human IL-2 and IL-4 were produced in our laboratory as described 49. CD40L-transfected J558 cells were provided by P. Lane, Birmingham, GB 50.
4.2 Peptides and tetramers
Melan-A26–35 peptide ELAGIGILTV 51 was purchased from Genosys (Sigma). Melan-A/HLA-A2 tetrameric complexes were synthesized as described 51. Tetramers were validated against specific CD8+ CTL clones and background levels of staining (<0.02%) were defined on PBMC of HLA-A2-negative healthy donors.
4.3 Generation and stimulation of DC and PDC
Blood was purchased from the UK National Blood Service and screened for HLA-A2 expression by FACS analysis. Monocytes were purified by positive sorting using anti-CD14 conjugated magnetic microbeads (Miltenyi, Bergisch Gladbach, Germany) and DC were generated as described 16, by culturing monocytes in RPMI-10% FCS supplemented with 50 ng/ml GM-CSF (Leucomax, Novartis Pharma, Switzerland) and 1,000 U/ml IL-4. Plasmacytoid cells were purified from PBMC by positive sorting using anti-BDCA-4 conjugated magnetic microbeads (Miltenyi). The recovered cells were over 90%pure as determined by flow cytometry with the anti-CD123 antibody (PharMingen) or anti-CD11c (PharMingen) and BDCA-2 (Miltenyi) (Fig. 1). PDC were cultured for 72 h in RPMI-10% FCS supplemented with 10 ng/ml IL-3 (R&D Systems, Minneapolis, Minnesota). DC and PDC (250,000/well in a 48-well plate) were stimulated for 24 h by addition of irradiated (9,000 rad) CD40L-transfected J558 cells (at a 1:5 ratio).
4.4 T cell priming
APC were irradiated (3,000 rad) and pulsed for 3 h with 100 ng/ml A26–35 peptide in serum-free medium. Cells were thoroughly washed and incubated with autologous PBMC at a 1:10 ratio in RPMI-10% FCS. Cells were expanded in IL-2-containing medium and analyzed at day 10–15. In some experiments, recombinant human IL-12 (5 ng/ml; R&D Systems, Minneapolis, MN) or IFN-α (20 UI/ml;Roferon-A, Roche) was added at the beginning of the cultures. PDC were infected with the replication-defective pox-virus MVA, encoding the melan-A26–35 epitope 52 at a MOI of 5, and 3 h later cells were matured with CD40L transfectants for 24 h.
4.5 FACS analysis
Cells were stained in PBS with allophycocyanin-labeled melan-A tetramer at 37° for 20 min, washed at room temperature and incubated on ice with the following antibodies: CD8-PerCP (Becton Dickinson, Mountain View, CA), CLA-FITC and CD62L-PE (PharMingen). The samples were analyzed on a FACSCalibur (Becton Dickinson) using Cell Quest Software. Melanoma cells were stained with the followingmonoclonal antibodies: W6/32 (anti class I, ATCC), L243 (anti class II, ATCC), CD95 (PharMingen), followed by a PE-conjugated affinity purified goat anti-mouse antibody (Southern Biotechnology Associates, Inc., Birmingham, AL). For intracellular cytokine detection, cells were stimulated for 6 h with 20 μg/ml A26–35 peptide as described 16. Staining was performed with anti-CD8-PerCP (Becton Dickinson), anti-IFN-γ-FITC and anti-IL-4-PE (PharMingen).
4.6 Cytokine determination
IL-12 p75 was measured by ELISA (PharMingen). The sensitivity of the assay was 30 pg/ml. Type I IFN was measured evaluating inhibition of Daudi cell proliferation against a standard IFN-α reference curve 32. The sensitivity of the assay was 0.2 U/ml.
Immunohistochemical analysis of freshly frozen tumor samples was performed as described 21. Serial sections were stained with monoclonal antibodies to CD123 (PharMingen, 1:40), Mart-1 (BioGenex, San Ramon, CA, USA, and 1:50), BDCA-2 (AC144, IgG1, Miltenyi, 1:30), CD8 (Becton Dickinson, 1:50) and CLA/Heca452 (PharMingen 1:100). Double immunofluorescence or immunoperoxidase techniques were used to reveal positive cells.
Analysis of CD95 expression (clone B-10, Santa Cruz Biotechnology, 1:50) was performed on formalin-fixed, paraffin-embedded sections following antigen retrieval by microwave in citrate buffer,by using an indirect immunoalkaline phosphatase (Chem-mate alkaline phosphatase detection kit, Dako). An isotype-matched irrelevant antibody (pan-cytokeratin) was used as negative control. Surgicalspecimens were obtained from five patients affected by primary cutaneous melanomas with different thickness and Clark's level and one melanocytic nevus. As positive control, a case of Burkitt lymphoma, known to express CD95, was used.
The authors thank Drs. J. Schmitz (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany), for his generous gift of monoclonal antibody anti-BDCA-2, P. Lane (Birmingham Medical School, Birmingham, GB), A. Knuth (Medizinische Klinik, Frankfurt, Germany), M. Collins (Guy's Hospital, London, GB) and E. Padovan (Kantonspital, Basel, Switzerland) for cell lines. The technical expertise of Mrs. Silvana Festa was also appreciated. This work was supported by funds from the Cancer Research UK and the Cancer Research Institute (USA).