CXCL4 Exposure Potentiates TLR-Driven Polarization of Human Monocyte-Derived Dendritic Cells and Increases Stimulation of T Cells

Stored in α-granules of platelets and released upon platelet activation, CXCL4 is a 7.8-kDa chemokine (1, 2). CXCL4 is also produced and released by various immune cells, including mast cells (3), dendritic cells (DCs) (4, 5), monocytes (6), and activated T cells (7). It has been described that CXCL4 plays a role in several physiological processes (8–10). However, increased levels of CXCL4 have been implicated in pathological conditions such as cancer (11) and infectious (12, 13) and inflammatory diseases (14–19). Indeed, a strong correlation was found between elevated CXCL4 levels in the circulation and the clinical features of patients with systemic sclerosis (SSc) (5). Monocytes and professional APCs are essential players in both innate defense and the initiation of adaptive immune responses, thereby contributing to both immune activation and the maintenance of immune peripheral tolerance (20, 21). Accordingly, upon detection of pathogen-associated molecular patterns (PAMPs) or danger-associated molecular patterns, DCs undergo further maturation into potent APCs now able to prime and induce the clonal expansion of Ag-specific T cells (22).

The imbalance of homeostasis because of the presence of inflammatory mediators such as CXCL4 leads to the modulation of phenotype and function of immune cells (23–32). In monocytes, for instance, CXCL4 not only functions as a chemoattractant mediator but also promotes survival, the production of TNF-α, release of reactive oxygen species, and differentiation into macrophage-like phenotype cells (26, 31, 33). Gleissner et al. (34) found that the exposure of monocyte-derived macrophages to CXCL4 induces unique transcriptomic changes, in comparison with M1 and M2 macrophages, and named these M4 macrophages. Genes involved in inflammatory responses, Ag presentation, and lipid metabolism were overexpressed in M4 macrophages. In addition, monocytes exposed to CXCL4 and IL-4 alone or in the presence of GM-CSF for 6 d were shown to result in a functionally distinct APC (24, 25, 31).

To date, it is not clear how CXCL4 might affect DC function, and thus influence innate and adaptive immune responses. In this study, we hypothesized that CXCL4 may modulate the phenotype and potentiate the innate function of DCs, as triggered by recognition of danger-associated molecular patterns and PAMPs. Such recognition occurs via germline-encoded immune receptors, including TLRs, which represent the frontline of innate defense. Indeed, dysfunction of TLR-mediated responses has been associated with immune and nonimmune cell reprograming (35, 36) and several autoimmune diseases, such as atherosclerosis (37), rheumatoid arthritis (38), psoriasis (39), and SSc (40–45). Notably, CXCL4 has been described as being involved in the same set of diseases (5, 14, 17, 19).

We found that CXCL4 reprograms monocytes as they differentiate into DCs, imprinting a more mature phenotype and an augmented responsiveness to TLR ligands: CXCL4–monocyte-derived DCs (moDCs) showed a dramatic increase in IL-12 and TNF-α production upon stimulation with TLR3, TLR7/8, and TLR8 ligands. Moreover, CXCL4-moDCs were more potent at inducing the proliferation of polyclonal CD4⁺ T cells and CD8⁺ T cells and cytokine production. Finally, Ag processing was impaired in CXCL4-moDCs, as they exhibited a superior ability to cross-present endocytosed Ags to human CMV (HCMV)-specific CD8⁺ T cells. Altogether, in this article we reveal novel functions of CXCL4 both in innate and adaptive immune responses.

Blood from healthy volunteers (HV) was obtained following institutional ethical approval. PBMCs were isolated from heparinized venous blood by density-gradient centrifugation over Ficoll Paque Plus (GE Healthcare). Fresh monocytes were isolated by Ab-based positive separation according to the manufacturer’s protocol, using anti-CD14 magnetic beads and auto-MACS assisted cell sorting (Miltenyi Biotec). Purity of isolated monocytes was >95% for all the independent samples. Negative fractions following monocyte isolation consisting of PBLs were cryopreserved in FCS containing 20% (v/v) DMSO and thawed after 6 d to be used for CD3⁺ T cell isolation and autologous coculture experiments.

Monocytes were cultured at a density of 1 × 10⁶ cells per milliliter using complete medium: RPMI 1640 with GlutaMAX (Life Technologies), supplemented with 10% (v/v) heat-inactivated FCS (Biowest) and 1% (v/v) antibiotics (penicillin and streptomycin) (both from Life Technologies). To generate moDCs, recombinant human IL-4 (500 U/ml; R&D Systems) and GM-CSF (800 U/ml; R&D Systems) were added to the medium, in the presence or absence of recombinant human CXCL4 (10 μg/ml; PeproTech). The moDCs were differentiated for 6 d at 37°C in the presence of 5% CO₂. At day 3, medium supplemented with the same concentration of IL-4, GM-CSF, and CXCL4 was added.

Nunc Lab-Tek II chamber slides (Thermo Scientific) were precoated with 1% (w/v) Alcian blue 8GX (Klinipath) in PBS for 30 min at 37°C, washed with PBS, and air-dried inside the culture hood prior to moDC differentiation culture (as described in 4MoDC differentiation). On day 6, the chamber slide was spun for 2 min at 500 × g. Next, 75% of the culture medium was removed, and cells were incubated with 500 μl Fixation/Permeabilization solution (eBioscience) for 30 min at room temperature. Fixed cells were washed twice with Permeabilization buffer (eBioscience) and incubated with phalloidin-labeled FITC (0.5 mg/ml; ENZO) and Hoechst 33342 (1 μM; Invitrogen) in Permeabilization buffer for 30 min in the dark (room temperature). Afterward, cells were washed once with Permeabilization buffer and a last wash with 1% (w/v) BSA and 0.1% (v/v) sodium azide (NaN₃; Sigma-Aldrich) in cold PBS (designated as FACS buffer). Finally, chambers were removed, and dried slides were mounted in Mowiol (Sigma-Aldrich) and coverslipped. Slides were left at 4°C overnight (O/N) until measurement. Acquisition of imaging data was performed on a Zen2009 LSM 710 (Zeiss) confocal microscope. To determine the cell area and perimeter, confocal images were obtained with the ×63 1.40 oil objective and analyzed using ImageJ software.

A total of 50 × 10³ immature moDCs (day 6 of differentiation) were plated in a 96-well flat-bottom plate (Thermo Scientific) in medium (0.5 × 10⁶ moDCs per milliliter) and rested O/N. Next, cells were left unstimulated or stimulated for 24 h at 37°C with the following TLR ligands: Pam3CSK4 (5 μg/ml), polyinosinic-polycytidylic acid [poly(I:C); 25 μg/ml], LPS (LPS; 100 μg/ml), flagellin (2 μg/ml), resiquimod (R848; 1 μg/ml), loxoribine (500 μM), imiquimod (R837; 3 μg/ml), thiazoloquinoline (CL075; 0.3 μg/ml), and CpG-B (ODN684; 5 μM), all purchased from InvivoGen. Surface expression of maturation markers and MHC molecules on moDCs after TLR stimulation was measured by flow cytometry. Cell-free supernatants were stored at −20°C for measurement of cytokine levels by Luminex technology, as described before (46) at the MultiPlex Core Facility of the Laboratory of Translational Immunology, University Medical Center of Utrecht.

One day prior to moDC coculture with polyclonal T cells, 50 μl anti-CD3 Ab (0.01 μg/ml) in PBS (clone OKT3; eBioscience) was immobilized to the surface of a 96-well round-bottom plate (Thermo Scientific) at 37°C O/N. Unbound Ab was removed by washing the wells three times with PBS. Autologous CD3⁺ T cells were purified by positive selection according to the manufacturer’s protocol using anti-CD3 magnetic beads and autoMACS-assisted cell sorting (Miltenyi Biotec). Purity of CD3⁺ T cells was >95% for all the samples. CD3⁺ T cells were labeled with CellTrace Violet (CTV) fluorescent dye (1.5 μM; Invitrogen) and cocultured with moDCs or CXCL4-moDCs (1:5 ratio) in a final volume of 150 μl. After 5 d of coculture, CD4⁺ T cell and CD8⁺ T cell proliferation and cytokine production were assessed by flow cytometry. The division index was calculated as a measure of proliferation, following FlowJo guidelines and previous publications (47, 48).

The moDCs and CXCL4-moDCs were pulsed with BSA-labeled Alexa Fluor 647 (0.1 μg/ml; Invitrogen) or DQ-Green BSA (0.1 μg/ml), a self-quenched dye conjugate of BSA (Life Technologies), for 10 min at 37°C to measure specific uptake or processing, respectively; or at 4°C to assess nonspecific cell surface binding. Cells were subsequently washed with cold medium and chased at 37°C for 10, 20, 40, or 100 min or left at 4°C. Next, cells were washed twice with FACS buffer and analyzed by flow cytometry. Ag uptake and processing were determined by analysis of median fluorescence intensity (MFI) for Alexa Fluor 647– or FITC-expressing cells. MFI measured at starting point t = 0 (0 min chase) was established as 100%. To calculate the percentage of BSA uptake or processing, the MFI for the time points 10, 20, 40, and 100 min of chasing was normalized to the respective t = 0.

MoDCs from HLA-A2⁺ HV were differentiated as described above, with or without CXCL4. Direct presentation of recombinant NLVPMVATV (NLV)-pp65 peptide (ProImmune) or cross-presentation of soluble recombinant HCMV-pp65 protein (Miltenyi Biotec) to HCMV-specific CD8⁺ T cell clones that we have generated was performed as previously described (49). Briefly, 50 × 10³ moDCs or CXCL4-moDCs were loaded either with the peptide NLV-pp65 (10°–10⁻³ μM) or with 30 μg/ml of the full protein HCMV-pp65 in 96-well round-bottom plates (on a final volume of 100 μl) O/N at 37°C. Where indicated, cells were pretreated with MG132 (2 μM; Calbiochem), hydroxychloroquine (50 μM; Sigma-Aldrich), brefeldin A (2.5 μg/ml; Sigma-Aldrich), or DMSO (Sigma-Aldrich) as control vehicle, for 30 min before Ag loading. The next day, cells were washed vigorously with complete medium and cocultured with 50 × 10³ HCMV-specific CD8⁺ T cells in the presence of GolgiStop (1/1500; BD). After 5 h of coculture, activation of CD8⁺ T cells was assessed by flow cytometry analyses.

Prior to Ab staining, moDCs were incubated with fixable viability dye eFluor780 or T cells with eFluor506 (eBioscience) in PBS, to allow exclusion of dead cells. After washing with FACS buffer, cells were treated with 10% (v/v) mouse serum (Fitzgerald) in FACS buffer to prevent nonspecific Ag binding. The moDC sets were stained for 20 min at 4°C with the following anti-human fluorochrome-conjugated mAbs: CD1a (clone HI149), CD11b (clone ICRF44), CD205 (clone DEC-205), CD206 (clone 19.2), CD80 (clone L307.4), HLA-DR (clone G46-6), HLA-ABC (clone G46-2.6), obtained from BD; CD14 (clone M5E2), CD1c (BDCA1; clone L161), CD83 (clone HB15e), CD86 (clone IT2.2), obtained from BioLegend; CD11c (clone 3.9) and CD40 (clone 5C3), obtained from eBioscience; and CD141 (BDCA3; clone AD5-14H12), obtained from Miltenyi, or the isotype control-matched Ab.

To measure intracellular cytokine expression by polyclonal T cells after 5-d coculture with moDCs, cells were restimulated for 5 h with PMA (50 ng/ml; Sigma-Aldrich) and ionomycin (1 μg/ml; Sigma-Aldrich) in the presence of GolgiStop (1/1500; BD). After incubation with fixable viability dye and mouse serum, cells were stained for 20 min with the following anti-human mAbs: CD3 (UCHT1; eBioscience), CD4 (clone RPA-T4; eBioscience), and CD8 (clone RPA-T8; BioLegend) in FACS buffer. To analyze intracellular cytokine production, cells were fixed and permeabilized for 30 min using Fixation/Permeabilization solution (eBioscience), according to the manufacturer’s instructions, and washed twice with Permeabilization buffer (eBioscience). Intracellular staining for IL-10 (clone JES3-19F1) and IL-4 (clone MP4-25D2), obtained from BD; IL-22 (clone IL22JOP) and IFN-γ (clone 4S.B3), obtained from eBioscience; and IL-13 (clone JES10-5A2), obtained from BioLegend, was performed for 30 min in Permeabilization buffer. Finally, cells were washed twice with Permeabilization buffer, and a last wash with FACS buffer. Autologous CD4⁺ T cell and CD8⁺ T cell proliferation and cytokine production were measured by flow cytometry. To analyze the cytokine expression by HCMV-specific CD8⁺ T cells after coculture with pp65-loaded moDCs, we used a staining protocol to similar to the above. The following anti-human mAbs were used for extracellular staining: CD3 (clone UCHT1; BioLegend), CD8 (clone RPA-T8; BD), and CD107a (LAMP1; clone H4A3; BD), followed by intracellular staining of IFN-γ (clone 4S.B4; BD) and TNF-α (clone MAB11; Sony Biotechnology). The cell acquisition of flow cytometry data was performed using LSR Fortessa (BD), and FlowJo software (version 7.6.5; Tree Star) was used for data analyses. In all flow cytometry analyses, cell debris was first excluded, then CD3⁺CD4⁺ T cells and CD3⁺CD8⁺ T cells were gated and analyzed for the expression of activation markers, or dilution of CTV in proliferation experiments. Data were represented as MFI or the percentage of positive cells for a specific cell marker, as mentioned in the figures.

Graphs and statistical analyses were performed with GraphPad Prism software (version 6.0). Paired t tests were used to compare two groups, and one-way ANOVA was used when more than two groups were compared. In all cases, the significance was defined as p ≤ 0.05.

To examine the contribution of CXCL4 to moDC differentiation, we generated moDCs for 6 d, with or without CXCL4 (Fig. 1A).

We observed that CXCL4-treated moDCs formed plate-adherent cell aggregates, with very heterogeneous morphology and extended dendrites, in contrast to moDCs (Fig. 1B). As a quantitative measurement to determine the cell size and formation of branched protrusions, we analyzed the area and perimeter of the cells based on the expression of f-actin. We observed that moDCs exposed to CXCL4 display a bigger perimeter (Fig. 1C) and area (Fig. 1D).

In addition, CXCL4-moDCs showed downregulation of the lipid-presenting molecules CD1a and CD1c, whereas the integrin CD11b involved in cell adhesion and migration, myeloid marker CD141, endocytic receptor CD205, maturation markers CD86 and CD83, and MHC class I (MHC-I) and class II molecules were expressed at higher levels (Fig. 1C, 1D). No differences were observed for the lineage marker CD11c; the mannose receptor CD206; the costimulatory molecules CD80 and CD40; and macrophage markers CD14, CD64, and CD163 (data not shown). These results demonstrate that CXCL4 exposure alters the differentiation and maturation of moDCs, suggesting that CXCL4-moDCs might perform differently on innate and adaptive immune responses in comparison with moDCs.

As DCs from SSc patients have been shown to display augmented response to TLR agonists, we next explored whether this could be attributed to the increased levels of CXCL4 observed in these patients (41). With this purpose, moDCs and CXCL4-moDCs were challenged for 24 h with TLR agonists or left unstimulated (Fig. 2A). Stimulation with poly(I:C) (TLR3), LPS (TLR4), and R848 (TLR7/8) resulted in upregulation of costimulatory and MHC molecules on both moDCs and CXCL4-moDCs (Fig. 2B). However, these TLR ligands induced upregulation of CD86, CD83, and MHC-I expression to a greater extent on CXCL4-moDCs than on moDCs (Fig. 2B). We compared the ability of CXCL4-moDCs to produce proinflammatory cytokines in response to a panel of nine different TLR ligands. Stimulation with poly(I:C), R848, and CL075 resulted in a markedly higher production of IL-12 (an average 64-fold, 5-fold, and 6-fold increase, respectively) and TNF-α (an average 15-fold, 7-fold, and 6-fold increase, respectively) by CXCL4-moDCs as compared with moDCs. Stimulation with Pam3CSK4 (TLR2), LPS, flagellin (TLR5), loxoribine, or imiquimod (both TLR7) did not result in significant differences in IL-12 and TNF-α production (Fig. 2C, 2D). Altogether, these results point toward CXCL4 as a modulator of TLR3- and TLR8-mediated innate responses in moDCs.

DCs are crucial for the initiation of T cell responses; thus they function as a bridge between the innate and adaptive immune systems. To evaluate whether CXCL4-moDCs were more potent in activation of T cell responses, immature moDCs and CXCL4-moDCs were cocultured with autologous CD3⁺ T cells. CXCL4 strongly potentiated the proliferation of CD4⁺ T cells and CD8⁺ T cells (Fig. 3A, 3B), in this study determined as division index (Fig. 3C). In addition, the frequencies of T cells expressing CD4⁺IFN-γ⁺ (Fig. 3D, 3E) as well as CD8⁺IFN-γ⁺ and CD8⁺IL-4⁺ (Fig. 3F, 3G) were significantly higher after coculture with CXCL4-moDC, in comparison with conventional moDCs. No differences were observed for CD8⁺IL-10⁺, CD8⁺IL-22⁺, and CD8⁺IL-13⁺ T cells (data not shown). On the basis of these results, we propose CXCL4-moDCs as potent inducers of both CD4⁺ T cell and CD8⁺ T cell responses in an Ag-independent manner.

To study whether CXCL4 could affect Ag presentation, we analyzed the ability of moDCs to uptake and process Ag, using BSA as a model. Uptake of Alexa Fluor 647 BSA by both moDC and CXCL4-moDCs was comparable over the 100-min duration of the pulse-chase experiment, with the greatest increase on uptake seen at 40 min of Ag exposure (Fig. 4A). Moreover, CXCL4-moDCs displayed impaired processing of Ag (DQ-BSA) compared with moDCs at 10, 20, and 40 min of chase. The highest Ag-processing level was observed at the time point 20 min (Fig. 4B). Taken together, whereas uptake of Ag by moDCs is not affected by exposure to CXCL4, exogenous Ag, when taken up by early stages of maturated DCs, might be retained intracellularly longer, thus kept preserved for slower degradation, implying that CXCL4 might modulate Ag cross-presentation capacities of moDCs.

CXCL4-moDCs display increased expression of MHC-I, and our data propose that these cells process Ag more efficiently, suggested by the longer Ag preservation on processing compartments. To investigate whether CXCL4-moDCs may be superior on direct Ag presentation or cross-presentation, we analyzed the response of HCMV-specific CD8⁺ T cell clones to their cognate peptide (NLV-pp65 or HCMV-pp65 protein after processing into peptides), either pulsed or endogenously processed by the moDCs. The moDCs and CXCL4-moDCs were first loaded with NLV-pp65 peptide or HCMV-pp65 protein and then cocultured with HCMV-specific CD8⁺ T cells (Fig. 5A). CD8⁺ T cells in the absence of NLV-pp65 or HCMV-pp65 protein did not produce IFN-γ and TNF-α (Fig. 5B, 5D). To control for differences in MHC-I expression, we assessed the direct presentation of NLV-pp65 peptide. We observed a trend for it to be increased, but not significantly so, upon CD8⁺ T cell activation, which showed as frequency of CD8⁺IFN-γ⁺ T cells and CD8⁺TNF-α⁺ T cells after presentation of NLV-pp65 (10⁰–10⁻³ μM) by CXCL4-moDCs (Fig. 5B, 5C). Strikingly, CXCL4-moDCs induced potent activation of HCMV-specific CD8⁺ T cell responses after cross-presentation of HCMV-pp65 protein, measured as the production of IFN-γ and TNF-α (Fig. 5D–F).

Next, to better understand the effects of CXCL4 on Ag-processing pathways, moDCs were treated with proteasome or endosomal inhibitors prior to Ag loading. The moDCs and CXCL4-moDCs conserved their ability to directly present NLV-pp65 peptide and stimulate CD8⁺ T cells after treatment with the inhibitors (data not shown). The moDCs and CXCL4-moDCs treated with proteasome (MG132) or endosomal (hydroxychloroquine) inhibitors displayed a reduced ability to stimulate HCMV-specific CD8⁺ T cells (Fig. 5G). Nevertheless, CXCL4-moDCs were still superior in activating CD8⁺ T cells, despite the inhibition of processing in both compartments. Blocking the transport of newly synthesized molecules from the endoplasmic reticulum to the golgi with brefeldin A completely abrogated Ag cross-presentation by moDCs and CXCL4-moDCs (Fig. 5E). Overall, these results indicate that CXCL4-moDCs potentiate the activation of Ag-specific CD8⁺ T cells.

Inflammatory mediators such as cytokines, chemokines, or PAMPs are found in circulation and localized at affected tissues from patients suffering from autoimmune diseases or infections, driving inflammation and modulating immune responses. CXCL4, a chemokine involved in several physiological processes, has been implicated in many diseases, and was recently proposed as a biomarker in SSc (5, 50, 17). Indeed, evidences of its immune-modulatory function on both immune and nonimmune cells have emerged in the past decade.

In the current study, we explored the effects of CXCL4 on moDC differentiation and on the initiation of innate and adaptive immune responses.

We showed by confocal microcopy images that in comparison with conventional differentiated moDCs, most of the CXCL4-moDCs are very heterogeneous, are bigger, and display branched protrusions. These observations suggest that CXCL4 may have an effect on the remodeling of cytoskeleton components, such as f-actin. As a result, CXCL4-moDCs might perform differently in important APC functions—for instance, in cell mobility and migration but in cell-cell contact, as well.

Besides the effects on moDC morphology, we also observed that CXCL4 exposure results in downregulation of the lipid-presenting molecules CD1a and CD1c. On the contrary, CD141, CD11b, CD205, maturation markers, and MHC molecules were upregulated on CXCL4-moDCs. Partially, these effects were described in previous studies, in which different approaches were used (24, 25, 31). To our knowledge, this is the first study to describe that CXCL4 sensitizes moDCs to triggering with several TLR ligands. Expression of maturation and MHC-I molecules were additionally upregulated on CXCL4-moDCs, specifically when triggered with poly(I:C) and R848. Moreover, we also found that CXCL4-moDCs are more potent producers of proinflammatory cytokines upon triggering with poly(I:C), R848, and CL075, compared with moDCs. Proliferation and cytokine production by autologous CD4⁺ T cells and CD8⁺ T cells were potentiated by CXCL4-moDCs, in an Ag-independent manner. In a previous study by Xia et al. (24), opposite effects were shown on the activation of CD4⁺ T cell responses, in a very distinct allogeneic culture system. We expanded these findings, showing that CXCL4-moDCs displayed impaired Ag processing in comparison with moDCs, and suggesting that CXCL4 might affect the functional ability of cells to perform Ag cross-presentation. CXCL4-moDCs loaded with HCMV-pp65 were able to potentiate the activation of HCMV-specific CD8⁺ T cells, in contrast to moDCs. These results suggest that CXCL4-moDCs display improved ability to activate adaptive immune responses.

When molecules resulting from tissue damage and inflammation are sensed by innate recognition receptors such as TLRs, the IFN signaling pathway is activated. Accumulating evidences from both human and mouse studies implicate TLR and type I IFN signaling in the pathogenesis of SSc and may be one of the causes leading to and sustaining autoimmunity and fibrosis (5, 36, 40, 41, 43, 44, 51). On the basis of this, we hypothesized that CXCL4 could alter TLR-mediated responses and contribute to immune-mediated diseases such as SSc. In this study, we found that CXCL4 exposure enhances TLR3- and TLR8-mediated moDC maturation and proinflammatory cytokine production.

Several studies have suggested that activation of TLR responses, like TLR3 (Myd88 independent) and MyD88 signaling on APCs can improve cross-priming and Ag cross-presentation (52–55). Therefore, our findings showing that CXCL4 exposure modulates TLR signaling may point toward a contribution of these innate effects on the adaptive immune responses.

In addition, the greater stimulating signals provided by the CXCL4-moDCs and T cells during coculture, such as TCR signaling, CD40-CD40L, and other costimulating signals, as well as cytokines released by DCs, most likely contribute to the potentiation of T cell responses by CXCL4-moDCs (56, 57). We also found that CXCL4 exposure upregulated the expression of CD11b. Previous studies showed that CD11^high cells, besides expressing a mature phenotype, are potent chemokine-producing cells, both in homeostasis and after stimulation with airway Ag or TLR ligands, and efficiently prime T cell responses (58, 59).

Internalization, processing, and presentation of Ag or dangerous molecules resulting from dying cells by DCs as peptide/MHC complexes is critical to the priming of T cells against tumors, virus infections, and inflammation. In our experimental setting, several molecules involved in these adaptive mechanisms were upregulated on CXCL4-moDC. For instance, we found upregulation of CD141 and CD205 expression on CXCL4-moDCs, in comparison with moDCs. DCs expressing these molecules were described to efficiently uptake Ags or molecules derived from dying cells, process, and cross-present via peptide-loaded MHC molecules (60–65). Of note, CD205 plays a crucial role in these processes as endocytic or nonendocytic receptor on both immature and mature cells (66). These features also most likely facilitate the efficient activation of Ag-specific CD8⁺ T cell responses.

Although it has been described that micropinocytosis is downregulated in mature DCs, the capacity to capture Ag by moDCs and CXCL4-moDCs was not affected, as seen in other studies (67, 68). We propose that owing to the mature phenotype of CXCL4-moDCs and the capacity to store Ag for a prolonged time, these cells may more efficiently process Ag for continuous supply to MHC-I, as well as promote the stability of these molecules, leading to potentiation of cross-presentation capacities (69–71). In addition, it was described that increase of immunoproteasome subunit expression during DC maturation (72) and modifications of proteasome and TAP activities can promote the presentation of peptide/MHC-I complexes (73, 74). We hypothesize that these pathways may be more activated in CXCL4-moDCs, because of their maturated phenotype, thus contributing also to higher ability for the activation of Ag-specific CD8⁺ T cells. Attempts to understand the contribution of endosomal and proteasome compartments for Ag processing by CXCL4-moDCs were confounded by the reduction to the same extent of CD8⁺ T cell activation when moDCs and CXCL4-moDCs were treated with MG132 and hydroxychloroquine. Clarification of how CXCL4 affects Ag processing on early or late endosomes, lysosomes, and proteasome compartments requires further investigation because these pathways are crucial for the effective cross-presentation of exogenous Ag to CD8⁺ T cells and activation of adaptive immune responses (75). Further research could make use of more specific inhibitors on cross-presentation assays, perhaps including selective immunoproteasome inhibitors (76, 77).

Although the sensitization of DCs by CXCL4 might improve responses against dangerous molecules resulting from dying cells and pathogens, and efficient priming of T cells contributes to protection against infections and tumors, on the other side, the homeostatic balance of these immune responses has a critical role in the maintenance of self-tolerance (78, 79).

Taken together, our findings corroborate and expand upon earlier studies showing that increased circulating levels of CXCL4 modulate inflammatory immune responses (5, 24–27, 29, 32).

We thank Lenny van Bon, Maarten van der Linden, Willemijn Janssen, Thijs Flinsenberg, Tessa Kempen, Maud Plantinga, and the Flow Core Facility for very helpful work discussions; and the Multiplex Core Facility of the Laboratory of Translational Immunology (University Medical Center Utrecht) for performing cytokine measurements.

The authors have no financial conflicts of interest.