Signal regulatory protein beta 2 is a novel positive regulator of innate anticancer immunity

In recent years, the therapeutic (re)activation of innate anticancer immunity has gained prominence, with therapeutic blocking of the interaction of Signal Regulatory Protein (SIRP)-α with its ligand CD47 yielding complete responses in refractory and relapsed B cell lymphoma patients. SIRP-α has as crucial inhibitory role on phagocytes, with e.g., its aberrant activation enabling the escape of cancer cells from immune surveillance. SIRP-α belongs to a family of paired receptors comprised of not only immune-inhibitory, but also putative immune-stimulatory receptors. Here, we report that an as yet uninvestigated SIRP family member, SIRP-beta 2 (SIRP-ß2), is strongly expressed under normal physiological conditions in macrophages and granulocytes at protein level. Endogenous expression of SIRP-ß2 on granulocytes correlated with trogocytosis of cancer cells. Further, ectopic expression of SIRP-ß2 stimulated macrophage adhesion, differentiation and cancer cell phagocytosis as well as potentiated macrophage-mediated activation of T cell Receptor-specific T cell activation. SIRP-ß2 recruited the immune activating adaptor protein DAP12 to positively regulate innate immunity, with the charged lysine 202 of SIRP-ß2 being responsible for interaction with DAP12. Mutation of lysine 202 to leucine lead to a complete loss of the increased adhesion and phagocytosis. In conclusion, SIRP-ß2 is a novel positive regulator of innate anticancer immunity and a potential costimulatory target for innate immunotherapy.


Introduction
Development of anticancer immunity critically depends on the interplay between the innate and adaptive immune system.One of the most important mechanisms of the innate immune system to facilitate cancer immunity is phagocytic uptake of target cells by phagocytes, such as macrophages and dendritic cells, and the subsequent presentation of antigens to the adaptive immune system (1).Phagocytosis is, among others, regulated by the signal regulatory protein (SIRP)-family.
The SIRP-family is a paired receptor family with an important immunoregulatory function in innate immunity that comprises of inhibitory, stimulatory and non-signaling receptors (2).The most prominent SIRP family member is the inhibitory protein SIRP-alpha (SIRP-a), which binds to its ligand CD47.The CD47/SIRP-a axis is a well-established innate immune checkpoint, in which overexpression of CD47 on cancer cells triggers SIRP-a-mediated inhibition of cancer cell phagocytosis (3).Herewith, removal and immunogenic processing of cancer cells by macrophages and dendritic cells is limited.In line with this, high CD47 expression is associated with poor clinical prognosis in various malignancies (4)(5)(6)(7).
Based on these features, the CD47/SIRP-a axis has become a prominent therapeutic target in cancer.Indeed, combination therapy of a CD47 antagonist with positive phagocytic stimulators such as the monoclonal antibody Rituximab (RTX) yielded complete responses in lymphoma patients refractory to RTX (8).Further, combination with the epigenetic drug azacytidine has yielded prominent clinical activity in Myelodysplastic Syndrome (9).In contrast, monotherapy with antagonistic antibodies has yielded few clinical responses (10), suggesting the balance between anti-phagocytic and pro-phagocytic signals is not sufficiently disturbed.These findings clearly highlight the therapeutic relevance of innate immunoregulatory processes for activating functional anti-cancer immunity.
In recent years, additional negative as well as positive regulators on innate anti-cancer immunity have been identified.For example, members of the leukocyte immunoglobulin-like receptor B (LILRB) family are known to inhibit signals in macrophages (11).Expression of LILRB1 on macrophages inhibits phagocytosis of cancer cells upon CD47 antibody treatment, an effect reversed by inhibition of LILRB1 (12).In addition, expression of LILRB2 on macrophages also inhibits phagocytosis of cancer cells.JTX-8064, a highly specific and potent blocker of this LILRB member, reprograms macrophages from an immunosuppressive to an immunostimulatory phenotype and thereby stimulates and activates T cells (11).Reversely, expression of the pro-phagocytic receptor SLAMF7 was reported to be crucial for execution of macrophage-mediated phagocytosis of cancer cells (13), although cancer cell-expression of SLAMF7 is not required for CD47 antibody treatment (13).Another example of a pro-phagocytic receptor is the G protein-coupled receptor (GPCR), specifically GPR84.GPR84 is found to be upregulated in AML leukemic stem cells and expression of this GPR member is correlated with significantly overall survival of patients diagnosed with acute myeloid leukemia (AML) (14,15).In this context it is worthwhile to note that several of the other much less researched SIRP family members may also be of therapeutic interest.For instance, SIRP-g is an immunostimulatory receptor expressed on T cells and its inhibition using an anti-SIRP-g antibody downregulated IFNgamma secretion and ameliorated Graft versus Host Disease (16).Further, although the exact role of SIRP-beta 1 (SIRP-ß1) is as yet unknown, this family member is expressed on myeloid cells and can trigger costimulatory signaling through the immunoreceptor tyrosine-based activation motif (ITAM)-containing adaptor protein DAP12.Specifically, SIRP-ß1 has a charged lysine residue in the transmembrane domain responsible for recruiting DAP12 (2).DAP12 is involved in immunostimulatory signaling and plays a central role in an extended array of receptors in NK cells, granulocytes, monocytes/macrophages and DCs (17).Finally, the family member SIRP-beta 2 (SIRP-ß2) has so far not been investigated, with only the full-length cDNA encoding SIRP-ß2 being reported to date (18).
Here, we investigated SIRP-ß2 and delineated its expression patterns, its immunostimulatory signaling pathways and its biological role in innate immune responses against cancer in vitro.Our results position SIRP-ß2 as a novel positive regulator of innate anticancer immunity and as a possible new clinically relevant member of the SIRP-family.
Expression of SIRP-ß2 at the protein level was confirmed in monocytes, PMNs and monocyte-derived macrophages.SIRP-ß2 co-localized with the plasma membrane protein CD11b for PMNs (R 2 = 0.79), macrophage type 0 (R 2 = 0.73) and macrophage type 1 (R 2 = 0.81) and had a weak co-localization on monocytes (R 2 = 0.23) (Figure 1E; Supplementary Figure 1C).In line with these results, endogenous expression of SIRP-ß2 was detected on a subset of donors in monocytes, monocyte-derived macrophages as well as PMNs, using flow cytometry (Figures 1F, G).Taken together, SIRP-ß2 was endogenously expressed in various myeloid effector cell types in a subset of healthy donors on the cell surface.
In conclusion, SIRP-ß2 promoted cell adhesion and facilitated monocyte and granulocyte adhesion and differentiation of THP-1 and CB-derived macrophages, increased CD11b expression, and increased macrophage-mediated phagocytosis of cancer cells upon antibody-mediated opsonization.

SIRP-ß2 expression potentiated HPV-E7-specific T cell responses
Since MHC molecules are directly responsible for antigen presentation, expression of SIRP-ß2 and subsequent upregulation of MHC molecules might potentiate antigen presentation and increase activation of specific T cell responses.To test this hypothesis, a Jurkat.NFAT-luciferase reporter assay with a Human Papilloma Virus (HPV) E7 peptide specific TCR, yielding luminescence upon TCR triggering, was utilized to investigate TCR signaling (27).Upon E7 peptide pulsing, the THP-1.SIRP-ß2 as well as the SIRP-ß2.K202L cells significantly increased NFAT activation compared to THP-1.EV at Effector-to-Target (E:T) ratios of 1:1, 5:1 and 10:1 (Figure 5I).When primary E7-TCR specific T cells were co-cultured with THP-1.SIRP-ß2 and E7 peptide, a significant increase in CD69 and CD25 was observed in both CD8 and CD4 positive T cells of different donors compared to co-cultures with THP-1.EV and E7 peptide (Figure 5J).In these primary T cells the co-culture with THP-1.SIRP-ß2.K202L did not enhance T cell activation (CD69 or CD25) compared to THP-1.EV (Figure 5J).Finally, IFN-g secretion by the primary E7-TCR specific T cells was E7 dependent (Supplementary Figure 2D) and significantly increased only in co-cultures with THP-1.SIRP-ß2 compared to THP-1.EV (p< 0.05, Figure 5K).In conclusion, SIRP-ß2 might enhance antigen presentation by upregulation of MHC-I class and increase T cell activation.

Discussion
In the current study, we identified that SIRP-ß2 is expressed under normal physiological conditions in macrophages and granulocytes at the mRNA and protein level and that endogenous expression of SIRP-ß2 on PMNs correlated with trogocytosis of cancer cells.Furthermore, ectopic expression of SIRP-ß2 in the THP-1 monocytic cell line and in primary cord blood-derived macrophages increased adhesion, differentiation, and cancer cell phagocytosis.SIRP-ß2 recruited the immune activating adaptor protein DAP12 to positively regulate innate immunity, with a mutation of the charged lysine responsible for DAP12 interaction abrogating functional activity.Finally, ectopic expression of SIRP-ß2 on the THP-1 model enhanced surface expression of MHC-I molecules, enhanced T cell activation as seen by increased NFAT activation in a Jurkat report system and the upregulation of activation markers CD69 and CD25 and IFN-g secretion on primary T cells.These findings position SIRP-ß2 as a novel positive regulator of innate immunotherapy.
SIRP-ß2 was first documented by Ichigotani et al., who isolated full-length cDNA encoding SIRP-ß2 (18).SIRP-ß2 has high homology with SIRP-ß1, with 75.5% of the amino acid sequence being identical.The main difference is in the extracellular domain, in which SIRP-ß1 has one V-and two C1-set Ig domains, whereas SIRP-ß2 contains two V-set Ig domains (2).Both proteins lack any tyrosine-phosphorylation sites, but contain a single basic lysine residue within the hydrophobic transmembrane domain.For SIRP-ß1, this charged residue was shown to associate with the adaptor protein DAP12, which through its cytoplasmic single ITAM can induce activating pathways, such as increased adhesion and phagocytosis (25,28,29).Therefore, DAP12-mediated activation was also the proposed mechanism for SIRP-ß2 (2, 18).The results presented here confirm this hypothesis, with SIRP-ß2 signaling being clearly dependent on DAP12 recruitment.Indeed, a mutation of the charged lysine residue on SIRP-ß2 lead to a complete loss of the increased adhesion and phagocytosis observed in THP-1.SIRP-ß2 and CB.SIRP-ß2 cells.In addition, DAP12 was found to be associated with SIRP-ß2 (IP), in SIRP-ß2 but not EV THP-1 cells.Finally, protein expression of DAP12 was increased in THP-1.SIRP-ß2 but not SIRP-ß2.K202L or EV cells.Together, this indicates that SIRP-ß2 signaling is mediated through DAP12, with the DAP12-SIRP-ß2 interaction facilitated by a single lysine residue in the transmembrane region.Interestingly, the HPV16-EV TCR Jurkat.NFAT.luciferasemodel in Figure 5I showed a significant difference between THP-1.EV and THP-1.SIRP-ß2.K202L, whereas in the primary T cell model these differences were not found (Figure 5J).However, in here we are looking at two different read-out experiments: NFkB signaling in a THP-1 model and surface expression of CD69 and CD25 on primary T cells.Primary T cells experiments are more sensitive to activation stimuli, also to time (best window for CD25 and CD69 can vary over time), different T cell populations and donor variability.The EV control in primary T cells express already high levels of activation markers, so this might be the reason why we don't see the window with the SIRP-ß2.K202L variant.Also, we have previously encountered a similar difference between the Jurkat.NFAT.luciferasemodel system and primary cell type (27).
Whereas the immune-inhibitory receptor SIRP-a as well as the T cell expressed SIRP-g both bind to CD47, SIRP-ß2 did not bind to CD47 (Supplementary Figure 2E, F).Interestingly, DAP12 signaling required (antibody-mediated) crosslinking of SIRP-ß1, which has also been reported for other ITAM containing adaptor proteins (17, 25, 28-30).However, SIRP-ß2 appeared to be active without secondary cross-linking, which could indicate that SIRP-ß2 does not require a ligand to operate or has the ability to self-cluster.Alternatively, the ligand was present in either the medium or expressed intracellularly.The latter seems unlikely, as immunoprecipitation did not reveal any suitable ligand candidates.Notably, whereas SIRP-ß2 already activates signaling in the current assays, it would be of interest to investigate whether monoclonal agonistic crosslinking of SIRP-ß2 could further enhance activity.In this respect, SIRP-ß1 has recently been exploited as immunotherapeutic target, with targeting of SIRP-ß1 on macrophages promoting phagocytic activity, promoting polarization towards M1 phenotype, and inducing killing of murine bladder cells (31).If SIRP-ß2 responds to agonistic targeting, it could be exploited as a co-stimulatory therapeutic target of the innate immune system.Endogenous expression of SIRP-ß2 on the protein level was only found in a subset (~45% Figure 1F) of donors on myeloid cells and, interestingly, was rapidly down regulated during trogocytosis in PMNs of different donors.Trogocytosis was previously reported to require active signaling via Fc-greceptors that have bound opsonizing antibodies, required CD11b/CD18 integrin and was inhibited by CD47/SIRP-a signaling (32,33).Notably, for SIRP-ß2, but not SIRP-ß1, a clear correlation was found between endogenous expression levels expression levels on PMNs and the level of trogocytosis.Together with the active down-regulation during trogocytosis, this positions SIRP-ß2 as a stimulatory receptor modulating trogocytosis.The mechanisms behind the downregulation of SIRP-ß2 during trogocytosis requires further investigation, but might be due to internalization or to shedding of the extracellular domain.In general, ectodomain shedding plays an important role in various processes, such as, migration, adhesion and different immune responses.Such shedding has been reported previously for SIRP-a (34, 35).Neural activity is involved in synapse maturation, which is activity-dependent on the ectodomain shedding of SIRP-a (36).The reason of shedding of the extracellular domain of SIRP-a is thought to be the enhanced binding capacity to the presynaptic receptor CD47 (36).In addition, the shedding of ectodomain SIRP-a in THP-1 monocytes and lung epithelia cells was regulated by metalloproteinase domaincontaining protein 10 (ADAM10).In here, the shedding enhanced immune activation signaling in response to inflammatory signaling (37).
Interestingly, whereas HLA-A2 was upregulated on the protein level this was not mirrored by mRNA levels, which were comparable for SIRP-ß2, SIRP-ß2.K202L and EV THP-1 cells.Thus, the increased HLA expression is not due to transcriptional upregulation.MHC-I complexes are dynamic molecules that undergo dynamic endocytic cycling and turnover (38,39).Only small amounts are normally exported to the cell surface, while most class I molecules are retained in the ER.Furthermore, once at the cell surface MHC class I molecules are continually removed by endocytosis and either recycled or degraded (40).We hypothesize that SIRP-ß2 influences the surface stability of MHC class I molecules or reduces degradation via direct interaction with the complex, as IP showed that both HLA-A and HLA-C potentially associated with SIRP-ß2.A reverse IP targeting HLA-A and HLA-C could provide more evidence.
In conclusion, SIRP-ß2 is expressed on myeloid cells at protein level and enhanced adhesion, differentiation, cancer cell phagocytosis and T cell activation in an in vitro setting.DAP12 is responsible for SIRP-ß2 signaling, and is recruited through a single, positive lysine residue in the transmembrane domain of SIRP-ß2.Thus, SIRP-ß2 appears to be a positive regulator of innate anticancer immunity.
Umbilical cord blood (CB) was derived from healthy full-term pregnancies after informed consent from the Obstetrics department of the Martini Hospital and the University Medical Center Groningen (UMCG), the Netherlands (protocol code NL43844.042.13, 6 January 2014).Mononuclear cells (MNCs) were isolated by density gradient centrifugation using lymphoprep (Alere Technologies AS, Oslo, Norway) and CD34+ cells were selected using the MACS CD34 microbeads kit on autoMACS (Miltenyi Biotec, Leiden, The Netherlands).Purity of >96% CD34+ cells after isolation was confirmed using Cytoflex (Beckman Coulter, Brea, CA, USA).

Lentiviral production
EV, SIRP-ß2 short and long isoform, SIRP-ß2.K202L were cloned in the pRRL vector, containing GFP separated by an IRES sequence.On day 1, T75 cultures flasks were coated for 2 h with 0.1% gelatin, afterwards 3x10 6 HEK293T cells were plated in 10 mL DMEM (Lonza, Basel, Switzerland) supplemented with 10% FCS.After overnight culture, the HEK293T cells were transfected with a mix of 3 μg packaging construct, 0.7 μg glycoprotein envelop plasmid, 3 μg vector construct and FuGENE HD transfection reagent (FuGENE ® HD transfection Reagent, Promega, Madison, WI, USA) was added in a DNA construct:FuGENE ratio of 1:7, by dropwise addition.After 18 h medium was replaced with DMEM medium without FCS, and incubated for 24 h.Afterwards, the virus containing supernatant was collected, spun down at 450 g for 5 min and passed over a 0.45 μM filter using a syringe.THP-1 and HL-60 cells were plated 5x10 5 /ml in complete medium in a 6-wells plate.100 μl of virus was added to the cells, after 24 h of incubation, virus was washed 3x with PBS, supplemented with 2% FCS, by spinning down at 450 g for 5 min.

In vitro THP-1 phagocytosis assay
THP-1 transduced with GFP expressing vector, SIRP-ß2 and SIRP-ß2.K202L cells were sorted with the cell sorter (Sony SH800, FACS facility, UMCG) and cultured in complete medium.THP-1 cells were seeded at 1.0x10 4 cells/well in a 24-well plate and incubated with 1 μM PMA (Thermo Scientific Waltham, MA, USA), for 72 h.After 72 h, PMA was washed away from the THP-1 cells.Tumor cells were labeled with Incucyte Cytolight Red (Essen BioScience, Ann Arbor, MI, USA) according to manufacturer's instructions.The labeled tumor cells were added to the THP-1 like macrophages at an effector to target ratio of 1:2, after which anti-CD20 or anti-CD47 antibody (1 μg/mL) was added and incubated for 3 h at 37°C.Subsequently, tumor cells were gently removed from the THP-1 like macrophages by washing 2-3 times with PBS and THP-1 macrophages were detached from the plate using TrypLE (Thermo Fisher Scientific, Waltham, MA, USA).Phagocytosis was analyzed by taking fluorescent pictures of the phagocytosis plate by use of the Incucyte microscope (S3 Live-Cell Analysis System, Ann Arbor, MI, USA) or determined by flow cytometry.The percentage of THP-1 mediated phagocytosis was calculated by counting the number of macrophages containing red tumor cells inside per 100 macrophages.Each condition was quantified by evaluating three randomly chosen fields of view.For flow cytometry analysis, the percentage of phagocytosis was determined by quantifying the percentage of double-positive (GFP+/incucyteRed+) THP-1 cells.

In vitro CB derived macrophage phagocytosis assay
Freshly isolated CD34+ cells from CB were first expanded in Stem Line medium (Miltenyi Biotech, Leiden, The Netherlands) supplemented with: 100 ng/mL SCF (Sigma Aldrich, St. Louis, MO, USA), 50 ng/mL FLT-3 (Sigma Aldrich, St. Louis, MO, USA), 30 ng/ mL GM-CSF (ImmunoTools, Friesoythe, Germany) and 10 ng/mL IL-6 (Sigma Aldrich, St. Louis, MO, USA).After expansion, these cells were transduced with GFP expressing vector, SIRP-ß2 short, long isoform and SIRP-ß2.K202L.To differentiate the CD34+ towards macrophages, cells were cultured in RPMI 1640 supplemented with 20% FCS and 50 ng/mL M-CSF (ImmunoTools, Friesoythe, Germany) for 14 days.Hereafter, macrophages were polarized into M1 like macrophages using 20 ng/mL IFNy and 50 ng/mL LPS (ImmunoTools, Friesoythe, Germany) for 24 h.For CB-derived macrophage phagocytosis assay, macrophages were detached by TrypLE and pre-seeded at 1.0x10 4 cells/well in a 96-well plate.Tumor cells were labeled with Incucyte Cytolight Red according to manufacturer's instructions.Labeled tumor cells were mixed with the macrophages at an effector to target (E:T) ratio of 1:5 and treated with anti-CD20 (1 μg/mL) for 3 h at 37°C.Tumor cells were gently removed from the macrophages by washing 2-3 times with PBS.Phagocytosis was analyzed by taking fluorescent pictures using the IncuCyte S3 Live-Cell Analysis System.The percentage of phagocytosis was calculated by counting the number of macrophages containing red tumor cells inside per 100 macrophages.Each condition was quantified by evaluating three randomly chosen fields of view.Experimental phagocytosis was calculated using phagocytosis with anti-CD20 subtracted by the phagocytosis of the medium control.
RT-qPCR data was analyzed as followed: first deltaCT was calculated using the CT-values of gene of interested minus housekeeping gene (RPL27), whereafter the calculation of 2^-deltaCT was made.

Adhesion assay
To determine adhesion of THP-1.EV, THP-1.SIRP-ß2 and THP-1.SIRP-ß2.K202L, 7.5x10 5 cells were added into a 6-well plate and incubated for 4 or 24 h in the presence or absence of 2.5 μg/mL PMA.After incubation, the wells were gently washed twice with PBS and stained with crystal violet (Sigma Aldrich, St. Louis, MO, USA) (1:3 dilution) for 1 h at RT.The crystal violet staining buffer was removed and wells were washed six times with PBS.Photos were taken on day 2, 4 and 7 to screen for adhesion, using an auto screen machine (AID EliSpot reader).Adhesion assay using xCELLigence, was performed as described previously (41).In short, 96-well E-plates (E-plate 96) (ACEA Bio, catalog number: 5232368001) were coated with extracellular matrix (ECM) molecule fibronectin (bovine plasma) (Sigma-Aldrich, catalog number: 341631) for 1 h at 37°C, whereafter excess coating was removed by washing twice with PBS.To block the non-specific binding of the THP-1 and HL-60 cells, the plate was coated for 1h at 37°C with 100 μl 0.1% BSA.For the experiment, 3x10 4 cells were plated in an end volume of 150 μl complete medium and measured for 4h in real time.

Trogocytosis
PMNs were isolated from apheresis blood of healthy individuals using ficoll density gradient using lymphoprep, as described above.SIRP-ß2 antibody was used to detect the expression level of endogenous SIRP-ß2 on PMNs.After 1 h of incubation with SIRP-ß2 antibody on ice, cells were washed thrice with PBS and incubated with Goat-anti-Rabbit-488 for 30 minutes on ice.Cells were washed thrice with PBS and analyzed using flow cytometry.
For trogocytosis by PMNs, SUDHL-6, SUDHL-10 and Ramos cells were labeled with Incucyte Cytolight Red according to manufacturer's instructions and mixed with PMNs in a 1:1 ratio (5x10 4 : 5x10 4 ) with/without anti-CD20 antibody (1 μg/mL) for 3 h at 37°C.Analysis was performed using flow cytometry.For analysis, the percentage of trogocytosis was determined using incucyteRed+ tumor cells within the PMN gate.Experimental trogocytosis (trogocytosis in presence of RTX subtracted by trogocytosis medium control).

Primary T cell activation
Primary T cells were obtained by isolation of PBMCs from buffy coats of HLA-A2 negative donors by ficoll density gradient lymphoprep, followed by pan-T cell isolation according to manufacturer's protocol (139-096-535, Miltenyi Biotec, Leiden, The Netherlands).T cells were lentivirally transduced with EV or
Accession numbers and -10logP values were used for further data processing.First, the acquired accession numbers of THP-1.EV and THP-1.SIRP-ß2 mass spectrometry were compared and only unique SIRP-ß2 accession numbers were kept.Next, all unique hits with an occurrence of 20% or more in the CRAPome database (42) were filtered out.CRAPome is a large publicly available database of standardized negative controls, obtained from various leading lab specializing in affinity purification mass spectrometry.The cellular location of the remaining proteins was determined, using the accession numbers and the UniProt database.Lastly, only the membrane-bound proteins were selected and ranked based on -10logP values.

1
FIGURE 1 Expression pattern, genomic arrangement and transcript variants of the SIRP family.(A) Schematic representation of the protein coding and noncoding transcript variants of the SIRP-ß2 locus.(B) Quantitative RT-qPCR analysis of mRNA expression in T cells and monocytes of SIRP family transcripts.(C) mRNA SIRP-ß2 expression in hierarchical differentiation tree.(D) Colony percentage of BFU-E (Burst Forming Unit E), CFU-GM (Colony-Forming Unit for granulocytes and macrophages) after 14days.1x10 3 CD34+ cells derived from cord blood (CB) plated, transduced with NT, EV and SIRP-ß2.(E) Confocal images of the surface expression of CD11b and SIRP-ß2 on monocytes.(F) Flow cytometry of SIRP-ß2 levels on monocytes (n=11), monocyte-derived macrophages (n=8) and polymorphonuclear neutrophil (PMN) (n=11), including different donors.(G) SIRP-ß2 expression pattern on flow cytometry from 2 different donors.