The immune receptor CD300e negatively regulates T cell activation by impairing the STAT1-dependent antigen presentation

CD300e is a surface receptor, expressed by myeloid cells, involved in the tuning of immune responses. CD300e engagement was reported to provide the cells with survival signals, to trigger the expression of activation markers and the release of pro-inflammatory cytokines. Hence, CD300e is considered an immune activating receptor. In this study, we demonstrate that the ligation of CD300e in monocytes hampers the expression of the human leukocyte antigen (HLA) class II, affecting its synthesis. This effect, which is associated with the transcription impairment of the signal transducer and activator of transcription 1 (STAT1), overcomes the capacity of interferon gamma (IFN-γ) to promote the expression of the antigen-presenting molecules. Importantly, the decreased expression of HLA-II on the surface of CD300e-activated monocytes negatively impacts their capacity to activate T cells in an antigen-specific manner. Notably, unlike in vitro- differentiated macrophages which do not express CD300e, the immune receptor is expressed by tissue macrophages. Taken together, our findings argue against the possibility that this molecule should be considered an activating immune receptor sensu stricto. Moreover, our results support the notion that CD300e might be a new player in the regulation of the expansion of T cell-mediated responses.

In vitro differentiation of monocytes to macrophages is accompanied by the down-regulation of CD300e expression 3 . However, we recently reported that the expression of the immune receptor in monocyte-derived macrophages can be rescued by the down-modulation of miR-4270 7 . In addition, Pagliari and colleagues revealed that the activation of rescued-macrophages by the CD300e agonistic antibody, led to a decreased exposure of HLA-DR on the plasma membrane 7 .
These data raised the intriguing possibility that CD300e might elicit both an activating and an inhibitory pathway, and prompted us to investigate in human monocytes, a cell subset that strongly expresses both CD300e and HLA-II molecules, the mechanisms used by CD300e to hinder the expression of HLA-II molecules.

Methods
Monocyte isolation and cell treatment. Monocytes were purified from buffy coats of healthy donors by density gradient protocol, as described previously 8 . Alternatively, monocytes were purified by negative selection using Classical Monocyte Isolation Kit (Miltenyi Biotec, Bergisch Gladbach, North Rhine-Westphalia, Germany), following the manufacturer's instructions. Cells were maintained in RPMI 1640 10% FBS, 50 μg/ml gentamicin and 4 mM HEPES.
Cross-linking of CD64 was performed as described elsewhere 9 . Briefly, 24-well plates were coated with 50 µg/ ml of goat anti-mouse Fc-specific F(ab′) 2  Referring to the experiments with conditioned media, monocytes were treated as above. After 24 h, media were collected, centrifuged and transferred on freshly isolated monocytes seeded on 24-well plate. Antibodydepleted media were obtained by incubating 1.5 ml of conditioned medium with 10 µl of protein G agarose (ThermoFisher, Waltham, MA, USA) at RT for 2 h, in rotation. Cells were processed for subsequent analyses 24 h after receiving media.
For the experiments including IFN-γ stimulation, monocytes were activated with anti-CD300e, as above. After 6 or 16 h, medium was changed with a fresh one containing 10 ng/ml IFN-γ. After 6 or 8 h, cells were processed for qRT-PCR and Flow Cytometry. Six h of activation with anti-CD300e antibody plus 6 h with IFN-γ was the timing adopted for qRT-PCR, while 16 h plus 8 h was the timing adopted for Flow Cytometry. For the evaluation of pSTAT1 by western blot, cells were harvested at 5, 10 and 15 min after receiving IFN-γ-containing medium.
When indicated, monocytes were simultaneously stimulated with 10 ng/ml IFN-γ and 10 µg/ml anti-CD300e and harvested after 5, 10 and 15 min for the evaluation of pSTAT1 by western blot.
Patients. Cells and tissues were collected from patients in accordance with the principles of the Helsinki Declaration and written informed consent was obtained from each patient, according to the study protocol approved by the Ethical Committee of the Universities of Padova, Brescia and Florence.
Stromal vascular fraction (SVF). Cells were resuspended in FACS buffer and incubated for 15 min at RT with 10% human serum to saturate Fc receptors. 1 × 10 6 cells were stained with the monoclonal antibody anti-CD300e, followed by a goat anti-mouse Alexa Fluor 488 antibody (ThermoFisher). The cell viability dye eFluor780 (Ebiosciences) was used to exclude dead cells from the analysis. Cells were fixed with 3.7% formaldehyde and permeabilized with PermWash solution (1% FBS, 0.2% saponin in PBS). Cells were stained with PE-conjugated anti-CD68 monoclonal antibody, resuspended in FACS buffer and analysed by FACSCanto II. Forward and side scatter light were used to identify cell populations. Values were expressed as the percentage of CD300e positive cells inside the gate of CD68 positive cells. A sample labelled only with the goat anti-mouse Alexa Fluor 488 antibody (ThermoFisher) was used to define the background. The percentage of cells expressing CD300e with MFI ≤ of background was considered negative. Based on the evidence that the intensity of CD300e expression identified two distinct CD68-positive cell populations, we arbitrarily defined them as CD300e bright and CD300e dim . All data were analysed using FlowJo software.
Flow cytometry-based internalization and recycling assay. For the internalization assay, monocytes were incubated with biotinylated anti-HLA-DR (clone L243, BioLegend, San Diego, CA, USA) in chilled FACS buffer. After 30 min, cells were washed and incubated in RPMI 10% FBS at 37 °C in rotation, in presence of medium alone, anti-CD300e antibodies or control IgG. The amount of tagged-HLA-DR remaining on the cell surface was determined by incubating the cells with Alexa 488-conjugated streptavidin (ThermoFisher) on ice. Cells were analysed by flow cytometry and the MFI at each time point was expressed as % of the value at time 0 (set as 100%). HLA-DR recycling assay was performed incubating monocytes with biotinylated anti-HLA-DR mAb (clone L243, BioLegend) at 37 °C for 30 min. Cells were washed in ice-cold FACS buffer and the surface-exposed antibodies were blocked using the Alexa 555-conjugated streptavidin (ThermoFisher). Cells were washed in FACS buffer and then cultured in pre-warmed complete medium at 37 °C in rotation, in presence of medium alone, control IgG or anti-CD300e for indicated time points. The recycled HLA-DR reappeared on the cell surface was detected by staining the cells with Alexa 488-conjugated streptavidin (ThermoFisher) on ice. Cells were analysed by flow cytometry and the MFI at each time point was expressed as % of the value at 120 min (set as 100%).
ELISA. Culture supernatants of stimulated monocytes were collected and stored at − 80 °C for subsequent quantification of cytokine content. ELISA kits, specific for IL-1β, IL-6, IL-8 and TNF-α (eBiosciences), were used following the manufacturers' instructions.

T-cell proliferation assay. Monocytes isolated by CD14 positive selection (CD14 MicroBeads, Mylteni
Biotec) from 10 patients with melanoma, were seeded as above on plates coated with anti-CD300e mAb or with isotype IgG. After 24 h of stimulation, cells were harvested and co-cultured in triplicate with CD4 + T lymphocytes, purified from the same patients, in presence of the melanoma associated antigen 3 (MAGE-A3), for 5 d. At 16 h before harvesting, 1 μCi/well of 3 [H]-TdR was added and the radionuclide uptake was measured, as reported 11 .

Results
CD300e activation reduces HLA-II total cell content in monocytes. Based on the evidence that the activation through CD300e affected the capacity of macrophages to expose HLA-DR antigens on the plasma membrane 7 , we wondered whether this pattern was recapitulated in monocytes which, unlike in vitro-differentiated macrophages, highly express the immune receptor 2 . Monocytes purified from buffy coat by density gradient were stimulated with the monoclonal antibody agonistic for CD300e (clone UP-H2) 3 . After 24 h we determined the expression of HLA-DR in all monocytes subsets, CD14 + CD16 − (classical), CD14 + CD16 + (intermediate), and CD14 low CD16 + (nonclassical). In line with our previous observations, with respect to monocytes exposed to the isotype-matched control, the engagement of CD300e maintained the amount of HLA-DR low in all subsets, even if for nonclassical monocytes the effect did not reach the significance (Fig. 1). The low expression of HLA-DR was reproduced in classical monocytes purified by negative selection (data not shown). The latter result ruled out the possibility that the reduction of HLA-DR was due to remaining lymphocytes. Twenty-four h after purification, among the cells seeded on uncoated or antibody-coated plate and identified as CD16 − , a proportion become CD14 low , as reported elsewhere 12 . We verified that in this subset the expression of HLA-DR was also kept low by CD300e engagement. It is noteworthy that the proportion of CD14 low CD16 − cells was much higher in monocytes activated through CD300e than in control cells (Fig. 1). This effect, which certainly deserves further investigation, suggested that CD300e might be involved in monocyte differentiation towards a macrophage-or dendritic cell-like phenotype 12 .
According to the finding that the activation of monocytes through CD300e maintained at basal level the expression of HLA-DR on the plasma membrane of all monocyte subsets, we decided to perform the subsequent experiments without distinguishing among the subsets. Similar to HLA-DR, HLA-DP and -DQ antigens were also affected by the crosslinking of the immune receptor ( Supplementary Fig. S1a,c).
In spite of the drop of HLA-II and in accordance with former studies 2,4 , activated monocytes up-regulated the co-stimulatory molecule CD86 and released pro-inflammatory cytokines (Supplementary Fig. S2). The increased expression of CD68 and CD206, further supported the hypothesis on the role of CD300e in promoting the differentiation of monocytes towards a macrophage-like phenotype (Supplementary Fig. S2). Oddly, the expression of the macrophage marker CD163 was unaffected following the activation through the immune receptor (data not shown).
CD300e compromises the ability of monocytes to elicit an HLA-II-restricted T-cell response. We moved to investigate whether the down-regulation of HLA-II was sufficient to hamper the ability of monocytes to stimulate the antigen-specific proliferation of CD4 + T lymphocytes, despite the upregulation of the co-stimulatory molecule CD86 2 (Supplementary Fig. S2). Monocytes isolated from melanoma patients were activated by the anti-CD300e antibody for 24 h and then used as antigen presenting cells towards autologous T lymphocytes, in presence of the melanoma antigen MAGE-A3 13 . As shown in Fig. 2 www.nature.com/scientificreports/  www.nature.com/scientificreports/ exposed to the isotypic antibody, efficiently activated T cells, the ability of CD300e-ligated monocytes to induce T cell proliferation was significantly impaired.

CD300e does not interfere with the intracellular trafficking of HLA-II. The expression of HLA-II
on the surface of antigen presenting cells is not a static phenomenon, since the complexes are internalized and can recycle back from early endosomes to the cell surface, reaching an equilibrium 14,15 . Therefore, we hypothesized that the activation through the immune receptor could reduce the exposure of HLA-II by interfering with its intracellular trafficking. To explore this possibility, we examined the kinetics of HLA-DR internalization and recycling by flow cytometry-based assays 10 . In control monocytes, either untreated or exposed to isotype IgG, HLA-DR internalized rapidly and reached an equilibrium of 50% internalization after 60 min (Fig. 3a). The assessment of the HLA-DR recycling in control monocytes confirmed that the internalized molecules could traffic from intracellular compartments to the plasma membrane (Fig. 3a). Essentially identical results were obtained when examining the kinetics of internalization and recycling of HLA-DR in CD300e-activated monocyte. These findings ruled out the possibility that CD300e interferes with the trafficking of HLA-DR complexes. We reasoned about the possibility that the ligation of CD300e could impact on HLA-DR expression. As revealed by western blot, the total amount (membrane + intracellular) of HLA-DR in stimulated monocytes, was actually reduced by ~ 50% (Fig. 3b).
We verified that the drop in HLA-DR expression did not depend on factors released in the medium (Supplementary Fig. S3). Moreover, the evidence that the crosslinking of Fc γ receptor I (CD64), which is also an activating receptor, did not influence HLA-DR expression, further disclosed the distinctive features of CD300e ( Supplementary Fig. S3). www.nature.com/scientificreports/

CD300e affects the synthesis of HLA-II in monocytes.
Based on the notion that class II major histocompatibility complex transactivator (CIITA) is the master regulator for the expression of HLA-II genes 16 , we assessed whether the abatement of HLA-II molecules in monocytes might be due to the down-regulation of CIITA.
Monocytes were treated as above and, after 3, 6 and 12 h, the expression of mRNA for CIITA was determined. Results revealed that, while in control monocytes (either untreated or exposed to the isotypic antibody) the www.nature.com/scientificreports/ expression of CIITA increased over the time, in cells activated for CD300e it remained at the basal level (Fig. 4a).
In accordance with the role of CIITA in the transcription of HLA-II genes, the expression level of mRNAs for HLA-DR, -DP-and -DQ were maintained low in monocytes, following the engagement of CD300e (Fig. 4b,c and Supplementary Fig. S1b,d).
Although this evidence reinforced the notion that the activation through CD300e hampered the expression of HLA-II, the final proof was obtained by comparing the HLA-DR protein synthetic rate in CD300e-activated monocytes vs control monocytes (Fig. 4d).

CD300e interferes with the IFN-γ pathway. Interferon-γ (IFN-γ) is a cytokine known to strongly
promote the transcription of CIITA and, hence, the expression of HLA-II 16 . Based on the observation that the activation through CD300e maintained the expression of CIITA at low levels in monocytes, we wondered whether this phenomenon was recapitulated even in presence of IFN-γ. Monocytes were stimulated with the anti-CD300e antibody. After 6 h, cells were left on the antibody-coated plate, but the medium was substituted with a fresh one containing IFN-γ and after further 6 h incubation, the expression of mRNA for CIITA was determined. As shown in Fig. 5a, the amount of mRNA for CIITA was up-regulated in control monocytes exposed to the cytokine while it was maintained low in CD300e-activated monocytes, regardless the presence of IFN-γ (Fig. 5a). Accordingly, the ligation of CD300e strongly affected the IFN-γ-induced HLA-DR expression (Fig. 5b). This finding suggests that the signalling pathway triggered by the engagement of the immune receptor could interfere with that activated by IFN-γ.
Activation through CD300e affects STAT1 expression. Binding of IFN-γ to its membrane receptor (IFNGR) induces the activation of Janus kinase 1 (JAK1) and 2 (JAK2), which leads to the phosphorylation of the signal transducer and activator of transcription 1 (STAT1) in the cytoplasm. Phosphorylated STAT1 forms a homodimer that translocates into the nucleus and binds to the IFN-γ-activated site (GAS)-E-box motif in the target genes, triggering their expression. Among these genes there is that encoding CIITA 16,17 . To evaluate if the activation through CD300e hindered the phosphorylation of STAT1, monocytes were activated for 16 h on anti-CD300e coated-plate and then stimulated with IFN-γ for 5, 10 and 15 min to induce the activation of STAT1. As shown in Fig. 6a, IFN-γ triggered the phosphorylation of STAT1 in control monocytes, while it was maintained www.nature.com/scientificreports/ low in cells previously activated through CD300e. However, the most intriguing finding was the reduced amount of unphosphorylated STAT1 in monocytes activated for CD300e, compared to controls. Notably, the IFN-γ -induced phosphorylation of STAT1 was not affected in monocytes that were simultaneously activated through the immune receptor ( Supplementary Fig. S4). Collectively, these data support the notion that the CD300e engagement could affect the expression of STAT1 rather than its activation. In concordance with this hypothesis, mRNA for STAT1, which increased over the time in control monocytes, did not augment in CD300e-activated monocytes (Fig. 6b). Even in cells exposed to IFN-γ that stimulates STAT1 gene expression 18 , the level of mRNA was strongly affected by the pre-activation through the immune receptor (Fig. 6c).
Tissue macrophages express CD300e. In vitro differentiation of human monocytes towards macrophages is accompanied by the downregulation of CD300e 3 . However, in vitro differentiated macrophages represent an oversimplification of tissue resident macrophages which consist of variably mixed populations of embryonic origin and bone marrow-derived blood monocytes. As a result of their complex origin, distribution and responses to endogenous and exogenous stimuli, these cells express marked phenotypic heterogeneity 19,20 . By immunohistochemistry we assessed the expression of CD300e in macrophages infiltrating 3 different tissues, colon mucosa, lung and liver. The analysis revealed that, dependently on the tissue, a different proportion of macrophages, was positive for CD300e (Fig. 7). With respect to the lung, in which all macrophages strongly expressed the immune receptor, in the liver ~ 20% of macrophages were positive for CD300e and displayed a weak pattern of expression. An intermediate scenario was observed in colon mucosa where within the 30-40% of macrophages expressing CD300e it was possible to clumsily distinguish between macrophages with strong expression of CD300e from others with a weak expression. A similar heterogenicity was recapitulated in macrophages infiltrating the adipose tissue from obese subjects (Supplementary Fig. S5). These findings are congruent with those from other studies demonstrating the heterogenous profile of tissue macrophages 19 and once more confirm that using in vitro differentiated macrophages is a simplistic and inadequate approach to study the functional profile of macrophages.

Discussion
Most of the immune receptors of the CD300 family, including CD300e, possesses a transmembrane domain and a short cytoplasmic tail. To propagate the downstream signals, the transmembrane domain associates with adaptor proteins containing immunoreceptor tyrosine-based activating motifs (ITAMs) 21 . The two main  www.nature.com/scientificreports/ ITAM-containing adaptors expressed by myeloid cells are DAP12 and FcRγ, the former suggested as adaptor for human CD300e 3 , while both for the mouse homologue 4 . Although for decades ITAM-and ITIM-containing molecules were thought to display opposite functions, this paradigm has been questioned by several evidence showing that the same ITAM-coupled receptor can trigger either an activating or an inhibitory signalling cascade [22][23][24] .
In line with this finding, we demonstrated in monocytes that CD300e elicits both a positive and a negative signalling pathway. The latter hampers the expression of HLA-II molecules thus, affecting the capacity of the cells to activate a specific T cell response. We are aware that monocytes are minimally involved in the antigen www.nature.com/scientificreports/ presentation in vivo, with respect to macrophages for example, but they represented an essential cell model to study the pathway triggered by CD300e, since they strongly express the immune receptor. On the other hand, macrophages differentiated from monocytes in vitro lose CD300e 3 . Interestingly, the ability of CD300e in interfering with HLA-II expression in monocytes overcomes the effect of IFN-γ, potent inducer of antigen-presenting molecules 25 .
The IFN-γ-induced HLA-II expression depends on the activation of the STAT1 pathway. Following phosphorylation, STAT1 dimerizes and enters the nucleus. There, it promotes the expression of target genes which include STAT1 itself and CIITA, the master regulator of HLA-II genes (Fig. 8a). Here, we report that the engagement of CD300e in monocytes hinders the expression of STAT1 and CIITA, regardless the presence of the cytokine (Fig. 8b,c). As expected, in cells activated by IFN-γ this leads to a reduction of the amount of STAT1 available for phosphorylation (Fig. 8c).
The existence of an intersection between the signalling cascade triggered by IFN cytokines and that involving receptors associated with ITAM-containing adaptors has been already suggested 26,27 . Several data agree about a positive cooperation between the two signalling cascades which augments the activation of STAT1 and thus the expression of inflammatory target genes 28,29 . Even if the role of the ITAM-containing adaptor DAP12 in the signalling cascade triggered by CD300e remains to be established, our results open the possibility that ITAMbearing adaptors might also negatively modulate the STAT1 pathway activated by IFN-γ.
It is worthy to highlight that CD300e affects the expression of STAT1, and that of its target CIITA, even in the absence of IFN-γ. Although this may sound curious because the expression of STAT1 target genes is known to require the transcription factor in its phosphorylated form, it has been demonstrated that unphosphorylated STAT1 mediates the basal expression of several target genes 30 , including CIITA and STAT1 31 (Fig. 8b).
Despite in vitro-differentiated macrophages do not express CD300e, we got preliminary evidence that the immune receptor is actually expressed by a proportion of tissue macrophages. Moreover, in accordance with the notion that tissue macrophages are a heterogenous population, with different function, we want to emphasize the evidence that CD300e is not equally expressed by all tissue macrophages and varies among different tissues. The fact that CD300e could modulate the antigen presentation capacity also in tissue macrophages and in other antigen presenting cells, such as dendritic cells, is an aspect that still needs to be further investigated. Furthermore, the evidence of a strong expression of CD300e by the glands of colon mucosa and by pneumocytes opens the possibility that the immune receptor could have a function also in non-immune cells.
Conscious that, beside the identification of the ligand of CD300e, many mechanistic details are still missing, including if truly CD300e cooperates with an ITAM-bearing protein adaptor and how its activation inhibits the synthesis of STAT1, we propose the existence of a pathway triggered by CD300e that negatively modulates the expression of STAT1.
Overall, our data support the notion that CD300e is not an activating receptor sensu stricto and open the possibility that this receptor might be a novel immune check point that contributes to the regulation of the expansion of T cell-mediated responses. Moreover, the induction of the immune receptor expression on the surface of phagocytic cells could represent a strategy for bacterial pathogens to evade the host immune response, as already suggested for Helicobacter pylori 7 .

Data availability
The data that support the plots and other findings of this study are available from the corresponding author upon reasonable request.