CD47 cross-dressing by extracellular vesicles expressing CD47 inhibits phagocytosis without transmitting cell death signals

Version of Record

Accepted for publication after peer review and revision.

Download
Cite
Share
CommentOpen annotations (there are currently 0 annotations on this page).

Version of Record published: December 1, 2022 (This version)
Accepted: November 15, 2022
Preprint posted: October 11, 2021 (Go to version)
Received: September 7, 2021

1. Of interest
Mouse gingival single-cell transcriptomic atlas identified a novel fibroblast subpopulation activated to guide oral barrier immunity in periodontitis

Takeru Kondo, Annie Gleason ... Ichiro Nishimura

Research Article Nov 28, 2023
Further reading

Abstract
Editor's evaluation
Introduction
Results
Discussion
Materials and methods
Data availability
References
Decision letter
Author response
Article and author information
Metrics

Abstract

Transgenic CD47 overexpression is an encouraging approach to ameliorating xenograft rejection and alloresponses to pluripotent stem cells, and the efficacy correlates with the level of CD47 expression. However, CD47, upon ligation, also transmits signals leading to cell dysfunction or death, raising a concern that overexpressing CD47 could be harmful. Here, we unveiled an alternative source of cell surface CD47. We showed that extracellular vesicles, including exosomes, released from normal or tumor cells overexpressing CD47 (transgenic or native) can induce efficient CD47 cross-dressing on pig or human cells. Like the autogenous CD47, CD47 cross-dressed on cell surfaces is capable of interacting with SIRPα to inhibit phagocytosis. However, ligation of the autogenous, but not cross-dressed, CD47 induced cell death. Thus, CD47 cross-dressing provides an alternative source of cell surface CD47 that may elicit its anti-phagocytic function without transmitting harmful signals to the cells. CD47 cross-dressing also suggests a previously unidentified mechanism for tumor-induced immunosuppression. Our findings should help to further optimize the CD47 transgenic approach that may improve outcomes by minimizing the harmful effects of CD47 overexpression.

Editor's evaluation

This study by Li et al., describes an important molecular mechanism that limits the rejection of xenotransplant or allogenic tissue. By presenting compelling evidence, that describes a mechanism to cross-dress xenogeneic cells with human CD47. Li et al., investigated if this process can improve grafted tissue survival by halting phagocytosis of xenogeneic cells by host human monocytes while also diminishing SIRP-a mediated apoptosis. This work will be of interest to basic and transplant immunology.

https://doi.org/10.7554/eLife.73677.sa0

Introduction

CD47 is ubiquitously expressed and acts as a ligand of signaling regulatory protein (SIRP)α, a critical inhibitory receptor on macrophages and dendritic cells (DCs). Emerging evidence indicates that the CD47-SIRPα signaling pathway plays an important role in regulation of macrophage and DC activation, offering a promising intervention target for immunological disorders. CD47KO cells are vigorously rejected by macrophages after infusion into syngeneic wild-type (WT) mice, demonstrating that CD47 provides a ‘don’t eat me’ signal to macrophages (Oldenborg et al., 2000; Wang et al., 2007a). Xenotransplantation using pigs as the transplant source has the potential to resolve the severe shortage of human organ donors, a major limiting factor in clinical transplantation (Yang and Sykes, 2007). We reported that the strong rejection of xenogeneic cells by macrophages (Abe et al., 2002) is largely caused by the lack of functional interaction between donor CD47 and recipient SIRPα (Wang et al., 2007b, Ide et al., 2007; Navarro-Alvarez and Yang, 2014). These findings led to the development of human CD47 transgenic pigs that have achieved encouraging results in pig-to-nonhuman primate xenotransplantation (Tena et al., 2017; Nomura et al., 2020; Watanabe et al., 2020). In addition to macrophages, a sub-population of DCs also expresses SIRPα (Wang et al., 2007a, Guilliams et al., 2016). Importantly, CD47-SIRPα signaling also inhibits DC activation and their ability to prime T cells, and plays an important role in induction of T cell tolerance by donor-specific transfusion (DST) and hepatocyte transplantation (Wang et al., 2007a, Wang et al., 2014; Zhang et al., 2016). Thus, transgenic expression of human CD47 in pigs may also attenuate xenoimmune responses by ameliorating DC activation and antigen presentation. More recently, transgenic overexpression of CD47 was also applied for reducing allogenicity and generating hypoimmunogenic pluripotent stem cells (Han et al., 2019; Deuse et al., 2019).

It has become increasingly evident that the CD47-SIRPα pathway plays a critical role in containing anti-tumor immune responses. CD47 upregulation was detected in various cancer cells, serving a powerful mechanism of evading macrophage killing (Jaiswal et al., 2009; Chan et al., 2009). Accordingly, treatment with CD47 blockade could inhibit tumor growth via macrophage-mediated mechanism (Jaiswal et al., 2009; Willingham et al., 2012; Weiskopf et al., 2016; Chao et al., 2010; Liu et al., 2015a). More recently, the antitumor activity of CD47 blockade was found to be associated with CD11c⁺ DC activation and largely T cell-dependent (Liu et al., 2015b, Chen et al., 2020; Li et al., 2020). Taken together, these studies revealed clearly that the CD47-SIRPα pathway provides a powerful negative regulation for both innate and adaptive immune responses and is increasingly considered as an effective intervention target for protecting against transplant rejection and unleashing immune responses to cancer.

Although negative regulation of immune responses is predominantly mediated by inhibitory CD47-SIRPα signaling in macrophages and DCs, it remains largely unknown how transgenic CD47 on pig or human pluripotent stem cells and upregulated CD47 on tumor cells interact with its receptor and ligands. In the present study, we identified an alternative source of cell surface CD47. We found that extracellular vesicles (EVs), including exosomes (Exos) from cells transgenically overexpressing CD47 or tumor cells overexpressing endogenous CD47, could induce CD47 cross-dressing on pig or human cells. CD47 cross-dressed on cell surfaces can interact with SIRPα to inhibit phagocytosis. However, unlike the autogenous CD47 that, upon ligation, induces cell apoptosis and senescence (Mateo et al., 1999; Gao et al., 2016; Meijles et al., 2017), ligation of CD47 cross-dressed on cell surfaces is not harmful to cells. This study provides deeper insight into the effect of CD47 overexpression, which needs to be considered when designing strategies for gene-editing in pigs for xenotransplantation or in human pluripotent stem cells for cell replacement therapy, and for developing CD47 blockade-based cancer immunotherapy.

Results

Transgenic hCD47 cross-dressing in pig cells

CD47 cross-dressing was first identified by detecting hCD47 on pig cells that was cocultured with pig cells expressing transgenic hCD47. Two different hCD47 isoforms were used to ensure that findings are not isoform-specific. In these experiments, cell cocultures were performed using cell line cells derived from porcine aortic cells (PAOC; Figure 1—figure supplement 1). First, we cocultured parental PAOCs (expressing pig CD47; referred to as PAOC/CD47^p) with PAOCs that were genetically modified to express hCD47 isoform 2 (referred to as PAOC/CD47^p/h2) or isoform 4 (referred to as PAOC/CD47^p/h4). Two different CD47 isoforms were used to confirm the findings are not isoform-specific. Flow cytometry analysis using anti-CD47 antibodies recognizing both human and pig CD47 revealed that PAOC/CD47^p/h2 or PAOC/CD47^p/h4 cells expressed a markedly increased level of CD47 compared to PAOC/CD47^p cells, and that PAOC/CD47^p cells showed significantly increased CD47 staining after coculture with PAOC/CD47^p/h2 or PAOC/CD47^p/h4 (Figure 1A). These results suggest that PAOC/CD47^p cells were cross-dressed by CD47, likely transgenic hCD47, from PAOC/CD47^p/h cells during cultures. To confirm this possibility, we made CD47-defficient PAOC cells (via targeted deletion using CRISPR-Cas9- technology; referred to as PAOC/CD47^null) and PAOC47^null cells that express transgenic hCD47 isoform 2 (PAOC/CD47^h2) or isoform 4 (PAOC/CD47^h4). When the PAOC cell line cells were cocultured for 24 hr, we found that PAOC47^null cells became positively stained by both anti-h/pCD47 (Figure 1B, top) and anti-hCD47 (Figure 1B, bottom) antibodies. To rule out the possibility that, during the coculture, the CD47^null cells did not become CD47⁺, but PAOC/CD47^h2 cells reduced hCD47 expression, we labeled PAOC/CD47^null (Figure 1C, Left) or PAOC/CD47^h2 (Figure 1C, Right) cells with florescence Celltrace violet and then cocultured the labeled cells with unlabeled PAOC/CD47^h2 or PAOC/CD47^null cells, respectively. This experiment, in which fluorescence-labeling allowed for better distinguishing between the two cell populations in the cocultures, further confirmed that PAOC/CD47^null cells can be cross-dressed by CD47 after coculture with PAOC/CD47^h2 cells (Figure 1C).

Figure 1 with 2 supplements see all

Download asset Open asset

Transgenic hCD47 cross-dressing on pig cells.

(A) Porcine aortic cell (PAOC)/CD47^p cells and PAOC/CD47^p/h2 or PAOC/CD47^p/h4 cells were cultured alone or cocultured for 24 hr, and hCD47 cross-dressing on gated PAOC/CD47^p cells was assessed by flow cytometry using anti-h/pCD47-PE mAb (clone CC2C6, reacting with both human and pig CD47). Shown are representative histogram profiles (left; the numbers in the figure indicate the median fluorescent intensity [MFI] of gated PAOC/CD47^p cells), and average MFI (right; mean ± SDs; n=6 replicates per group) of gated PAOC/CD47^p cells in the indicated cell cultures. ****, p<0.0001 (two-tailed unpaired t-test). Results shown are representative of three independent experiments. (B) PAOC/CD47^null, and PAOC/CD47^h2 or PAOC/CD47^h4 were cultured alone or cocultured for 24 hr, and analyzed for hCD47 cross-dressing on gated PAOC/CD47^null cells by flow cytometry using anti-h/pCD47-PE mAb (*top*) or anti-hCD47-BV786 mAb (*bottom*). Shown are representative histogram profiles (left panel; the numbers in the figure indicate the MFI of gated PAOC/CD47^null cells), and average MFI (right panel; mean ± SDs; n=3 replicates per group) of gated PAOC/CD47^null cells in the indicated cell cultures. **, p<0.01; ****, p<0.0001 (two-tailed unpaired t-test). Results shown are representative of three independent experiments. (C) Celltrace violet-labeled PAOC/CD47^null (left panel) or PAOC/CD47^h2 (right panel) was cultured alone (Violet CD47^null or Violet CD47^h2) or cocultured with unlabeled PAOC/CD47^h2 (Violet CD47^null + CD47^h2) or PAOC/CD47^null (CD47^null or Violet CD47^h2), respectively, then the cells were stained by anti-h/pCD47-PE mAb. Shown are representative flow cytometry profiles (n=3 replicates). Results shown are representative of two independent experiments. (D) Pig lymphoma cell line (LCL) and hCD47-tg LCL (LCL/CD47^p/h) cells were cultured alone or cocultured for 24 hr, and analyzed for hCD47 cross-dressing on gated LCL cells (the numbers in the figure indicate the MFI of gated LCL cells). The staining control of ‘mixed at staining’ indicates the two types of cells were cultured separately and mixed immediately prior anti-CD47 staining. Two independent experiments were performed, and each experiment had two replicates per group. Representative flow cytometry profiles are shown. (E) PAOC47^null cells were cocultured with bone marrow cells from hCD47-tg miniature swine for 2 days and analyzed for hCD47 cross-dressing on gated PAOC47^null cells by flow cytometry using anti-h/pCD47-PE mAb (the numbers in the figure indicate the MFI of gated PAOC/CD47^null cells). Two independent experiments were performed, and each experiment had two replicates per group. Representative flow cytometry profiles are shown.

PAOC cells express SIRPα (Figure 1—figure supplement 2, left), and pig SIRPα is reported to interact with human CD47 (Boettcher et al., 2019). Furthermore, previous studies have shown that CD47-lentiviruses are more effective in transducing SIRPα⁺ non-phagocytic tumor cells, suggesting that the interaction of CD47 with SIRPα may facilitate lentiviral engulfment by target cells (Sosale et al., 2016). Thus, to determine whether hCD47 cross-dressing is mediated by binding of hCD47 to pig SIRPα, we performed cocultures with pig lymphoma cell line (LCL) cells that do not express SIRPα (Figure 1—figure supplement 2, right). LCL cells cocultured with LCL cells that express hCD47 (LCL/CD47^p/h) (Ide et al., 2007; Wang et al., 2011), but not those cultured alone or mixed with LCL/CD47^p/h immediately prior to flow cytometry analysis, were positively stained by anti-hCD47 antibodies (Figure 1D). Although we cannot rule out of the role of SIRPα in hCD47 cross-dressing, our data indicate that hCD47 cross-dressing can occur in an SIRPα-independent manner.

We then wished to determine if cells other than the PAOC cell lines could be a source of CD47 for cross-dressing. Toward this end, PAOC47^null cells were cocultured with bone marrow cells (BMCs) from hCD47-transgenic miniature swine. Like pig cells cocultured with hCD47-transgenic cell lines (Figure 1A–D), PAOC47^null cells also became positively stained by anti-hCD47 antibodies after being cocultured with hCD47-tg swine cells (Figure 1E).

Cross-dressing with native CD47 from human T cell leukemia cells

We next determined whether cells can be cross-dressed by native CD47 using human T-cell leukemia Jurkat cells that express a higher level of CD47 than normal hematopoietic cells (Figure 2A). In order to clearly identify cross-dressed CD47 on cell surface, CD47-defficient Jurkat cells were generated using the CRISPR-Cas9 technique (Figure 2—figure supplement 1) and cocultured for 24 hr with the parental WT Jurkat cells or with hCD47-tg pig LCL cells (Figure 2B). Flow cytometry analysis revealed that CD47KO Jurkat cells were clearly stained positive by anti-hCD47 antibodies after coculture with parental WT Jurkat cells or hCD47-tg pig LCL cells compared to those cultured alone or mixed immediately before staining (Figure 2B). Furthermore, pig LCL cells also became positive for human CD47 staining after coculture for 24 hr with WT Jurkat cells (Figure 2B). More than half of the cross-dressed hCD47 (determined by MFI) remained on pig LCL cells 3 days after being separated from hCD47-expressing cells (Figure 2C), indicating that hCD47 cross-dressing was relatively stable. We also noted that not only hCD47, but other membrane proteins of WT Jurkat cells were also cross-dressed on pig LCL cells (Figure 2—figure supplement 2). While to a lower level than untreated cells, hCD47 cross-dressing was clearly detected on LCL cells that were pretreated with cytochalasin D, a cell-permeable potent inhibitor that disrupts actin microfilaments (Cooper, 1987), after cocultured with cytochalasin D-pretreated WT Jurkat cells (Figure 2—figure supplement 3A), indicating that a functional cytoskeleton is not absolutely required for, but may facilitate hCD47 cross-dressing. These results indicate that human CD47 cross-dressing could be induced by not only hCD47-transgenic cells but also tumor cells that express only the native CD47.

Figure 2 with 3 supplements see all

Download asset Open asset

CD47 expression and CD47 cross-dressing on human T cell leukemia Jurkat cells.

(A) CD47 expression on Jurkat cells and normal human CD3⁺, CD19⁺, and CD14⁺ peripheral blood cells (the numbers indicate median fluorescent intensity [MFI] of human CD47 staining). (B) CD47 expression on wild-type (WT) Jurkat cells, pig lymphoma cell line (LCL)/CD47^p/h cells, CD47KO Jurkat cells, CD47KO cells mixed with WT Jurkat cells (mixed at the time of staining), CD47KO Jurkat cells cocultured (24 hr) with WT Jurkat or pig LCL/CD47^p/h cells, pig LCL cells, and LCL cells cocultured (24 hr) with WT Jurkat cells. (C) PKH67-labeled pig LCL cells were cocultured with PHK26-labeled LCL/CD47^p/h (left) or WT Jurkat (right) cells for 1 day, PHK67-labeled pig LCL cells are sorted and treated with mitomycin C (2 µg/ml) for 30 min (to stop cell division), then analyzed for hCD47 staining on pig LCL cells immediately (Day 0) and at the indicated times after cultured in media. The numbers in the figure indicate MFI of CD47 staining on gated cells.

CD47 cross-dressing by extracellular vesicles and exosomes

We next investigated whether CD47 cross-dressing can be induced by EVs or requires direct cell-cell interaction. We analyzed hCD47 cross-dressing on CD47KO Jurkat cells, pig LCL cells, and PAOC/CD47^null cells in the absence or presence of EVs prepared from PAOC/CD47^h2 cells. Flow cytometry analysis showed that CD47 cross-dressing occurred in both CD47KO human T-cell leukemia Jurkat cells (Figure 3A) and pig B-lymphoma LCL cells (Figure 3B) after incubation for 2 or 6 hr with PAOC47^h2 cell-derived EVs. To a less extent, both CD47KO Jurkat and LCL cells were also positively stained by anti-hCD47 after incubation with Exos released by PAOC47^h2 cells (Figure 3A and B). Similarly, hCD47 cross-dressing was detected in PAOC^null cells after incubation with EVs from PAOC47^h2 cells (Figure 3C). In this experiment, PAOC^null cells were labeled with florescence Celltrace violet prior to incubation with EVs to ensure there was no contamination by PAOC^h2 cells in the prepared EVs. After incubation with PAOC47^h2 EVs, a significant proportion of violet-labeled PAOC^null cells became positive for hCD47 (Figure 3C). Of note, the frequency of hCD47⁺ PAOC^null cells at 42 hr was lower than that at 18 hr, which is most likely due to PAOC^null cell proliferation and EV exhaustion/degradation. Pig LCL cells were also stained positive for hCD47 after incubated with EVs from hCD47-tg pig LCL cells or WT Jurkat cells, in which hCD47 cross-dressing was detected 1 hr after incubation with EVs and the levels remained stable or increased at 24 hr (Figure 3—figure supplement 1). These results indicate that CD47 cross-dressing could be induced independently of cell-cell contact by EVs, including Exos. Again, a functional cytoskeleton is not absolutely required for hCD47 cross-dressing, as hCD47 was clearly detected on cytochalasin D-pretreated LCL cells after incubation with EVs from hCD47-tg LCL cells, though the level was moderately reduced (Figure 2—figure supplement 3B). Furthermore, cross-dressing was not diminished by treatment with trypsin/EDTA, indicating that membrane fusion is involved in EV cross-dressing (Figure 3—figure supplement 2).

Figure 3 with 2 supplements see all

Download asset Open asset

CD47 cross-dressing by extracellular vesicles (EVs) and exosomes (Exos).

(**A–B**) CD47KO Jurkat cells (A) or pig lymphoma cell line (LCL) cells (B) were cultured in the absence (*left*) or presence (*right*) of EVs (*top*) or Exos (*bottom*) prepared from PAOC/CD47^h2 cell culture supernatants for 2 hr or 6 hr, and analyzed for hCD47 cross-dressing by flow cytometry using anti-hCD47-BV786 mAb. Representative flow cytometry profiles of three independent experiments were shown. (C) Celltrace violet-labeled PAOC/CD47^null cells were cultured in the absence (*left*) or presence (*right*) of EVs prepared from PAOC/CD47^h2 cells for 18 hr or 42 hr, and analyzed for hCD47 cross-dressing by flow cytometry using anti-hCD47-BV786 mAb.

Protection against phagocytosis by cross-dressed CD47

We next determined whether cross-dressed CD47 can act as a marker of self to protect the cells against phagocytosis. We first investigated the binding potential of cross-dressed hCD47 with human SIRPα. CD47KO Jurkat cells were cocultured without or with EVs from PAOC/CD47^h2 cells for 5 hr, washed, and incubated with recombinant human SIRPα-Fc chimera for 1 hr. Binding of human SIRPα fusion protein to CD47KO Jurkat cells was then measured using fluorochrome-labeled anti-human IgG Fc antibody. Flow cytometry analysis showed that human SIRPα fusion protein was able to bind CD47KO Jurkat cells cultured with EVs but not those cultured without EVs (Figure 4A). The data indicate that cross-dressed hCD47 on CD47KO Jurkat cells can bind human SIRPα.

Figure 4

Download asset Open asset

Protection against phagocytosis by cross-dressed CD47.

(A) CD47KO Jurkat cells were cocultured without (*left*) or with (*right*) extracellular vesicles (EVs) from PAOC/CD47^h2 cells at 37°C for 5 hr, then washed and incubated with recombinant human SIRPα-Fc chimera at 37°C for 1 hr. The binding of SIRPα-Fc proteins to CD47KO Jurkat cells was visualized by staining with APC-conjugated mouse anti-human IgG Fc mAb. Representative flow cytometry profiles of two independent experiments are shown. (B) PAOC47^null and PAOC47^h2 cells were cultured alone (*left* and *middle*) or together (*right*) for 48 hr and stained using anti-hCD47-BV786 mAb, then PAOC47^null (R1), PAOC47^h2 (R3), and hCD47⁺ (i.e. hCD47 cross-dressed) PAOC47^null (R2) cells sorted from cocultures were used immediately for phagocytic assay. (C) PAOC47^null (R1), PAOC47^h2 (R3), or sorted hCD47 cross-dressed PAOC47^null (R2) cells were labeled with Celltrace violet and incubated with human macrophages for 2 hr, then phagocytosis was determined by flow cytometry using anti-human CD14 mAb. Shown are representative flow cytometry profiles (left) and levels (right; mean ± SDs; n=3) of phagocytosis (i.e. percentages of human macrophages that have engulfed violet + target cells (CD14⁺violet⁺) in human CD14⁺ macrophages). Representative results of three independent experiments are shown. (D) CD47KO Jurkat, wild-type (WT) Jurkat, and CD47KO Jurkat cells pre-incubated with EVs from PAOC/CD47^h2 cells were labeled by Celltrace violet and cocultured with human macrophages for 2 hr, then phagocytosis was analyzed by flow cytometry. Shown are representative flow cytometry profiles (left) and levels (right; mean ± SDs; n=3) of phagocytosis (i.e. percentages of hCD14⁺violet⁺ in total hCD14⁺ macrophages). Representative results of two independent experiments are shown. *, p<0.05; **, p<0.01; ***, p<0.001; ****, p<0.0001; ns, not significant (two-tailed unpaired t-test).

We then performed phagocytic assay to determine the potential of cross-dressed hCD47 to protect pig cells or CD47KO human leukemia cells against phagocytosis by human monocyte-derived macrophages. PAOC47^null and PAOC/CD47^h2 cells were cultured for 48 hr, then hCD47 cross-dressed PAOC47^null cells were sorted out (Figure 4B) and their susceptibility to phagocytosis by human macrophages was determined in comparison to PAOC47^null and PAOC/CD47^h2 cells that were cultured separately (Figure 4C). As expected, PAOC47^null cells were significantly more sensitive than PAOC/CD47^h2 cells to phagocytosis (Figure 4C). However, hCD47 cross-dressing effectively reduced the susceptibility of PAOC47^null cells to phagocytosis by human macrophages, to a level comparable to that of PAOC/CD47^h2 cells (Figure 4C). Human CD47 cross-dressing protects not only xenogeneic pig cells, but also human leukemia cells, against phagocytosis by human macrophages. In phagocytic assays where CD47KO Jurkat cells showed significantly greater phagocytosis than WT Jurkat cells, pre-incubation of CD47KO Jurkat cells with EVs released by PAOC/CD47^h2 cells was found highly effective in reducing their phagocytosis by human macrophages (Figure 4D). These results indicate that human CD47 cross-dressing can act as a functional ligand for human SIRPα and deliver ‘don’t eat me’ signals to human macrophages.

Ligation of autogenous but not cross-dressed CD47 induces death in Jurkat cells

Ligation of cell surface CD47 by its ligand thrombospondin-1 (TSP-1) (Saumet et al., 2005), CD47-binding peptides of TSP-1 (Martinez-Torres et al., 2015) or CD47 antibodies (Mateo et al., 1999) has been shown to induce death in varying types of cells. In line with these reports, we observed that human SIRPα-Fc fusion proteins could induce cell death in a dose-dependent manner in WT, but not CD47KO, human T-cell leukemia Jurkat cells (Figure 5—figure supplement 1; Figure 5A and B). The cell death observed in WT Jurkat cells was induced by CD47 ligation with hSIRPα-Fc proteins, as cell death was minimally detectable in WT Jurkat cells that were cultured simultaneously without hSIRPα-Fc proteins. To determine whether cross-dressed CD47 on Jurkat cells may also induce cell death, we compared the susceptibility to cell death induced by SIRPα-Fc proteins among WT, CD47KO, and hCD47 cross-dressed CD47KO Jurkat cells. CD47 cross-dressing was performed on GFP⁺ CD47KO Jurkat cells by incubation for 2 hr with PAOC/CD47^h2 EVs. To induce cell death, GFP⁺ CD47KO or hCD47 cross-dressed GFP⁺ CD47KO Jurkat cells were cocultured, respectively, with an equal number of control WT Jurkat cells (5×10⁴ each) in the presence of human SIRPα-Fc for 1 hr. The cocultured cells were then stained with anti-hCD47 (BV786) mAb, and cell death in WT (hCD47⁺GFP^-), CD47KO (CD47^-GFP⁺), and hCD47 cross-dressed CD47KO (CD47^lowGFP⁺) Jurkat cells were measured. Of note, binding of hSIRPα-Fc proteins to cell surface CD47 (either native or cross-dressed) could partially block subsequent staining with anti-hCD47 antibodies and thus, the cells cultured with hSIRPα-Fc showed relatively lower hCD47 staining than those cultured without (Figure 5—figure supplement 2, Figure 5). We found that SIRPα-Fc proteins induced significant cell death in WT Jurkat cells regardless of whether they were cocultured with CD47KO (Figure 5C and D) or with hCD47 cross-dressed CD47KO (Figure 5E and F). CD47 cross-dressing did not increase the sensitivity to cell death induced by SIRPα-Fc proteins, and cell death was minimally detectable in both CD47KO (Figure 5C and D) and hCD47 cross-dressed CD47KO (Figure 5E and F) Jurkat cells. However, CD47KO Jurkat cells regained the sensitivity to apoptosis induced by CD47 agonists after being transfected to express CD47 (Figure 5—figure supplement 3). These results indicate that, unlike autogenous CD47, cross-dressed CD47 on Jurkat cells does not induce cell death upon ligation with SIRPα-Fc proteins.

Figure 5 with 3 supplements see all

Download asset Open asset

Ligation of autogenous but not cross-dressed CD47 induces cell death.

(**A–B**) 5×10⁴ wild-type (WT) Jurkat cells (GFP⁻) were incubated in the presence (A) or absence (B) of 2.5 µg/ml hSIRPα-Fc proteins at 37°C for 1 hr and stained with anti-hCD47 mAb and Propidium Iodide (PI). Representative flow cytometry profiles show dead cells (PI⁺) in hCD47⁺ WT Jurkat population. (**C–D**) CD47KO Jurkat (GFP⁺) and WT Jurkat cells were mixed (at 1:1 ratio; 5×10⁴ each) and cocultured in the presence of 2.5 µg/ml human SIRPα-Fc for 1 hr, then stained with anti-hCD47 mAb and PI. Shown are representative flow cytometry profiles (C) and percentages (D; mean ± SDs) of PI⁺ dead cells in gated CD47KO (R1, CD47^-GFP⁺) and WT (R2, hCD47⁺GFP⁻) Jurkat cells. (**E–F**) CD47KO Jurkat (GFP⁺) cells were incubated for 2 hr with PAOC/CD47^h2 extracellular vesicles (EVs), then mixed with WT Jurkat cells (at 1:1 ratio; 5×10⁴ each) and cocultured in the presence of 2.5 µg/ml human SIRPα-Fc for 1 hr. The cells were stained with anti-hCD47 mAb and PI. Shown are representative flow cytometry profiles (E) and percentages (F; mean ± SDs) of PI⁺ dead cells in gated EV-treated CD47KO (R1, CD47^lowGFP⁺) and WT (R2, hCD47⁺GFP^-) Jurkat cells. ****, p<0.0001 (two-tailed unpaired t-test). Of note, the cells cultured with hSIRPα-Fc proteins (**A, C and E**) showed reduced hCD47 staining (as detailed in Figure 5—figure supplement 2). Results shown are representative of three independent experiments.

Discussion

Given its strong inhibitory effects on macrophage activation and phagocytosis, transgenically overexpressed CD47 is considered an effective means of preventing transplant rejection. A major obstacle impeding the translation of xenotransplantation into clinical therapies is vigorous xenograft rejection (Yang and Sykes, 2007), and the lack of functional interaction in CD47-SIRPα pathway is a key mechanism triggering macrophage xenoimmune responses (Yang, 2010). Studies have shown that the use of gene-edited pigs carrying human CD47 is effective in protecting against xenograft rejection by macrophages in non-human primates (Ide et al., 2007; Tena et al., 2017; Watanabe et al., 2020). Recently, transgenic overexpression of CD47 was also successfully used in combination with other approaches, such as deletion of HLA molecules, to generate hypoimmunogenic pluripotent stem cells (Han et al., 2019; Deuse et al., 2019). Here we found that EVs and Exos released from pig cells transgenically overexpressing hCD47 can mediate hCD47 cross-dressing on surrounding pig cells. Furthermore, hCD47 cross-dressed on pig cells can interact with human SIRPα and inhibit phagocytosis by human macrophages. Such CD47 cross-dressing occurs not only in the pig-to-pig combination, but also in the human-to-human, pig-to-human, and human-to-pig directions. These results provide a new mechanism for the inhibition of phagocytosis by the approach of transgenically expressing CD47.

In both xenogeneic and allogeneic settings, a high level of transgenic CD47 expression was found to be essential for its immune inhibitory effect. CD47 is not only a ligand of SIRPα, but also a signaling receptor that mediates a variety of functions, including apoptosis, cell cycle arrest, and senescence. It was reported that deletion of CD47 improves survival, proliferation, and function of endothelial cells, leading to increased angiogenesis and neovascularization both in vitro and in vivo (Meijles et al., 2017; Gao et al., 2016; Gao et al., 2017). In line with these observations, CD47 deletion from organ grafts was reported to ameliorate renal ischemia/reperfusion injury (Isenberg and Roberts, 2019) and cardiac allograft rejection (Chen et al., 2019). Using a pig-to-baboon kidney xenotransplantation model, a recent study suggested that widespread expression of hCD47 in the pig kidney was associated with increased vascular permeability and systemic edema, presumably due to upregulated TSP-1, and hence, TSP-1-CD47 signaling in the graft (Takeuchi et al., 2021). Our study showed that, although CD47 cross-dressed on cells may bind CD47 ligands, the ligand engagement does not induce CD47 signaling to cause apoptosis. We found that ligation of hSIRPα-Fc proteins with autogenous, but not cross-dressed, CD47 on Jurkat cells induces cell death. This study suggests that using a pig vascularized organ with hCD47 overexpression in some cells, which are more sensitive to macrophage attack but relatively resistant to CD47 signaling-induced deleterious effects, may improve the outcomes of xenotransplantation.

Emerging evidence indicates that CD47-SIRPα signaling plays an important role in regulating DC activation, and hence, T cell priming. The initial evidence for a role of CD47 in controlling DC activation was obtained in a mouse model of DST, in which DST using WT cells induces donor-specific tolerance, but DST using CD47-deficient cells paradoxically induces SIRPα⁺ DC activation and augments anti-donor T cell responses (Wang et al., 2010). A similar finding was made in a mouse model of hepatocyte allotransplantation where WT hepatocytes promote allograft survival, but CD47KO hepatocytes exacerbated rejection (Zhang et al., 2016). Although it was not tested directly, it is conceivable that CD47 cross-dressed on cells, which induces sufficient SIRPα signaling to inhibit macrophages, may suppress SIRPα⁺ DC activation. Thus, in addition to inhibition of phagocytosis, CD47 cross-dressing may also attenuate anti-donor T cell responses when transplants are performed using CD47-overexpressing donors.

CD47-SIRPα signaling is an important component of the tumor immune microenvironment. CD47 upregulation was found in many types of cancer cells, in which CD47 provides an important mechanism to evade macrophage killing (Jaiswal et al., 2009; Chan et al., 2009). CD47 upregulation in tumors also significantly contributes to tumor-induced T cell suppression, as the antitumor activity of CD47 blocking treatment is associated with CD11c⁺ DC activation and is largely T cell-dependent (Liu et al., 2015b, Chen et al., 2020; Li et al., 2020). In the present study, CD47 cross-dressing was found in human T-cell leukemia Jurkat and pig B-lymphoma LCL cells, suggesting a possible involvement of CD47 cross-dressing in the formation of a tumor immunosuppressive microenvironment. In addition, CD47 on EVs and CD47 cross-dressed on tumor cells may also neutralize CD47 ligands, such as TSP-1 that has been shown to induce CD47 activation leading to apoptosis in tumor and endothelial cells (Martinez-Torres et al., 2015), hence, favoring tumor growth.

While the mechanisms of EV-mediated exchange of biological information and materials between cells remain poorly understood (Raposo and Stahl, 2019), EV-induced antigen cross-dressing has been reported to play an important role in the regulation of immune responses, including alloantigen recognition and allograft rejection (Zeng and Morelli, 2018; Gonzalez-Nolasco et al., 2018). Earlier studies have shown that CD47, as a ‘don’t eat me’ signal, is essential for EVs to elicit biological function by preventing their clearance by macrophages (Kamerkar et al., 2017). Here we report that EV-induced CD47 cross-dressing possesses partial activity of the autogenous CD47, such as the ability to initiate inhibitory SIRPα signaling, and therefore offers means of separating the desired and harmful effects of CD47.

Materials and methods

Cell culture

Request a detailed protocol

Jurkat (J.RT3-T3.5) cell line was purchased from ATCC (named WT Jurkat to distinguish from CD47KO Jurkat cells) and the identity has been authenticated by STR profiling provided by China Center for Type Culture Collection (CCTCC). Jurkat cells were grown in Dulbecco's Modified Eagle's Medium supplemented with 10% Fetal Bovine Serum (FBS) (Atlanta Biologicals), GlutaMax and 100 U/ml penicillin and streptomycin. Pig LCL cell is a pig B cell lymphoma cell line (LCL-13271) derived from Swine Leukocyte Antigen AD (SLA^AD) miniature swine with post-transplantation lymphoproliferative disease (PTLD), which is kindly provided by Christene Huang (Harvard Medical School). The markers used for characterization of pig LCL cells are CD2, CD25 and anti-mu heavy chain (details were published previously) (Cho et al., 2007). Human CD47 (hCD47) transgenic (tg) pig B-LCL cell line (hCD47-tg LCL) and control pig LCL cell line were generated by transfecting LCL cells with pKS336-hCD47 or empty pKS336 vector, respectively, as described previously (Ide et al., 2007). Pig LCL cells were grown in RPMI 1640 Medium supplemented with 10% FBS, GlutaMax, MEM Non-Essential Amino Acids Solution, sodium pyruvate, 5 μM 2-mercaptoenthaol and 100 U/ml penicillin and streptomycin. Human CD47-tg porcine BMCs were harvested from SLA-defined miniature swine with hCD47 transgene (Watanabe et al., 2020), and BMCs were grown in Dulbecco's Modified Eagle's Medium supplemented with 10% FBS, GlutaMax and 100 U/ml penicillin and streptomycin. Pig aortic endothelial cells (PAOC) immortalized with SV40 were purchased from ABM (catalog # T0448), and the endothelial cell characterization was determined by expression of VE cadherin and CD31, and uptake of fluorescent Dil-Ac-LDL (measured by DiI-Ac-LDL Kit; Cell Applications, Cat # 022K). All PAOC cell lines were grown in Porcine Endothelial Cell Growth Medium (Cell Applications, Cat # P211-500) supplemented with 10% FBS. All cell lines were confirmed negative for mycoplasma contamination by MycoStrip mycoplasma detection kit (InvivoGen, Cat# rep-mys-10). Trypsin/EDTA was used for cell dissociation. Media and reagents for cell culture were purchased from GIBCO. Where indicated, cells were treated with mitomycin C (2 µg/mL for 30 min; Sigma-Aldrich) or cytochalasin D (2 µg/mL for 60 min; Thermo Fisher).

Generation of CD47KO Jurkat and PAOC sublines

Request a detailed protocol

CRISPR small guide RNA (sgRNA) for disrupting hCD47 in Jurkat cells was designed using the online tools (https://crispr.mit.edu), with sequence targeting the exon 2 of hCD47 (CTACTGAAGTATACGTAAAG-TGG [PAM]). The sgRNA was cloned into the pL-CRISPR.EFS.GFP lentiviral vector which was a gift from Benjamin Ebert (Addgene plasmid no. 57818) for co-expression with Cas9 (Heckl et al., 2014). Lentiviral particles were produced by co-transfection of a three-plasmid system consisting of the pL-CRISPR.EFS. GFP vector and packaging plasmids (pVSV-G and pΔ) using CaCl₂ into 293T cells in 175 cm² flasks. Lentivirus supernatant was collected 48 hr post-transfection, concentrated by ultracentrifugation at 22,000 rpm for 2.5 hr (Beckman Coulter, Optima XE-90) and stored at –80°C until use. GFP⁺ cells were sorted 3 days after lentivirus transduction, then sorted GFP⁺ cells were expanded and assessed for CD47 expression by staining with BV786-conjugated anti-hCD47 mAb B6H12 (BD Bioscience) and PE-conjugated anti-hCD47 mAb CC2C6 (Biolegend). CD47-negative Jurkat cells were established by four rounds of cell sorting. We also generated Jurkat cells that express only transgenic hCD47 by transduction of CD47-negative Jurkat cells with pLVX lentiviral vector encoding hCD47 isoform 2 cDNA.

PAOC cells were immortalized with Lenti-hTERT virus (ABM; cat no. G200) following manufacturer’s instructions and clonal sorting/expansion. Alpha GAL KO pAOC-SV40-hTERT (GTKO) cell line (PAOC/CD47^p) was created via nucleofection of pAOC-SV40-hTERT with plasmid expressing Cas9 protein (GeneART CRISPR Nuclease Vector, Invitrogen) and guide RNA targeting GGTA-1 gene (aGal protein, guide RNA sequence used: TCATGGTGGATGATATCTCC) using Lonza 4D-Nucleofector, followed by clonal sorting/expansion. PAOC/CD47^null was created via nucleofection of PAOC/CD47^p cells with plasmid expressing Cas9 protein and guide RNA targeting the pig CD47 gene (guide RNA sequence used: TCACCATCAGAATTACTACA) using Lonza 4D-Nucleofector. PAOC/CD47^null and PAOC/CD47^p cell lines expressing hCD47 short (305aa, 16aa short intracellular domain, NM_198793, NP_942088, PAOC/CD47^h2, and PAOC/CD47^p/h2) or hCD47long isoforms (323aa, 34aa long intracellular domain, NM_001777, NP_001768, PAOC/CD47^h4, and PAOC/CD47^p/h4) (Reinhold et al., 1995) were created via nucleofection with plasmid expressing Cas9 protein (GeneART CRISPR Nuclease Vector, Invitrogen) and guide RNA targeting AAVS1 safe harbor site using Lonza 4D-Nucleofector, followed by clonal sorting/expansion.

Flow cytometric analysis

Request a detailed protocol

CD47 expression or cross-dressing on cells was determined by direct staining with BV786- or AF647-conjugated anti human specific CD47 mAb B6H12 (abbreviated anti-hCD47; BD Bioscience) or PE-conjugated anti-human CD47 mAb CC2C6 (with cross-reactivity to pig CD47 and thus, referred to as anti-h/pCD47; Biolegend). The level of cell surface CD47 is expressed as median fluorescent intensity. FITC(fluorescein isothiocyanate)-conjugated anti-pig SIRPα mAb (clone BL1H7) was from Abcam; recombinant human SIRPα/CD172a Fc chimera protein, CF (cat no. 4546-SA-050) was from R&D system; APC-conjugated anti-human IgG Fc (clone HP6017), PE-conjugated anti-human CD90 (clone 5E10), AF647-conjugated anti-human HLA-ABC (clone W6/32) were from Biolegend; PKH26 and PKH67 were from Sigma-Aldrich. For analysis of CD47 expression on human PBMCs, single-cell suspensions were incubated with anti-hCD47 mAb B6H12 in combination with fluorochrome-conjugated anti-human CD45 (clone HI30), CD19 (clone HIB19), CD3 (clone SK7), and CD14 (clone M5E2; all from Biolegend). Dead cells were identified by staining with propidium iodide or 7-AAD. All samples were collected on Flow Cytometer (Fortessa and Celesta, Becton Dickinson), and data were analyzed by Flowjo software (Tree Star).

Purification of extracellular vesicles

Request a detailed protocol

EVs and Exos from cell culture supernatants were purified by a standard differential centrifugation protocol as previously reported (Chen et al., 2018). In brief, bovine Exos were depleted from FBS by overnight centrifugation at 100,000 g, PAOC/CD47^h2 cells were cultured in media supplemented with 10% Exos-depleted FBS for EVs and Exos purification from cell culture supernatants. Supernatants collected from 48-hr cell cultures were centrifuged at 2000 g (3000 rpm) for 20 min to remove cell debris and dead cells. EVs were pelleted after centrifugation at 16,500 g (9800 rpm) for 45 min (Beckman Coulter, Optima XE-90) and resuspended in PBS. The pelleted Exos from above supernatants were further centrifuged at 100,000 g (26,450 rpm) for 2 hr at 4°C (Beckman Coulter, Optima XE-90) and resuspended in PBS. EVs and Exos from total 1.3×10⁷ PAOC/CD47^h2 cells cultured for 48 hr were concentrated in 250 ul and 400 ul PBS, respectively, 10 ul of each was used for CD47 cross-dressing.

Preparation of human macrophages

Request a detailed protocol

Blood from healthy volunteers was used to prepare peripheral blood mononuclear cells (PBMCs) by density gradient centrifugation. PBMCs were added at 3×10⁶ per well in a 24-well plate, and unattached cells were removed from the plate on the second day. Attached cells were then differentiated to macrophages by 8–9 days of culture in Iscove's Modified Dulbecco's Medium (IMDM) (Gibco) + GlutaMax (Thermo fisher scientific) supplemented with 10% AB human serum (Gemini Bio-products, Inc), containing 10 ng/ml human M-CSF (PeproTech) and 100 U/ml penicillin and streptomycin (Gibco). The use of human blood samples was approved by the Institutional Review Board of Columbia University Medical Center.

Flow cytometry-based phagocytic assay

Request a detailed protocol

Macrophages generated as above were harvested from plates using Trypsin-EDTA (Thermo fisher scientific). The indicated target cells were labeled with Celltrace violet (Thermo fisher scientific) according to the manufacturer’s protocol, and phagocytic assay was performed by co-culturing 6×10⁴ Celltrace Violet-labeled target cells with 3×10⁴ human macrophages for 2 hr in ultra-low attachment 96-well flat bottom plates in IMDM + GlutaMax without antibiotics or serum added. All cells were harvested after coculture, and phagocytosis was determined by flow cytometry analyses, in which the phagocytic ratio is calculated as the percentage of macrophages that engulfed target cells (human CD45⁺CD14⁺Celltrace violet⁺) among total macrophages (human CD45⁺CD14⁺).

Statistical analysis

Request a detailed protocol

Data were analyzed using GraphPad Prism (version 8; San Diego, CA) and presented as mean value ± SDs. The level of significant differences in group means was assessed by student’s t-test, and a p-value of ≤0.05 was considered significant in all analyses herein.

Data availability

Figure 1, 4, 5; Figure 2-figure supplement 2, Figure 2-figure supplement 3, Figure 3-figure supplement 1, Figure 3-figure supplement 2, Figure 5-figure supplement 1, Figure 5-figure supplement 3 - Source Data contain the numerical data used to generate the figures.

References

1. Abe M
2. Cheng J
3. Qi J
4. Glaser RM
5. Thall AD
6. Sykes M
7. Yang YG
(2002) Elimination of porcine hemopoietic cells by macrophages in mice
Journal of Immunology 168:621–628.
https://doi.org/10.4049/jimmunol.168.2.621
- PubMed
- Google Scholar
1. Boettcher AN
2. Cunnick JE
3. Powell EJ
4. Egner TK
5. Charley SE
6. Loving CL
7. Tuggle CK
(2019) Porcine signal regulatory protein alpha binds to human CD47 to inhibit phagocytosis: implications for human hematopoietic stem cell transplantation into severe combined immunodeficient pigs
Xenotransplantation 26:e12466.
https://doi.org/10.1111/xen.12466
- PubMed
- Google Scholar
1. Chan KS
2. Espinosa I
3. Chao M
4. Wong D
5. Ailles L
6. Diehn M
7. Gill H
8. Presti J
9. Chang HY
10. van de Rijn M
11. Shortliffe L
12. Weissman IL
(2009) Identification, molecular characterization, clinical prognosis, and therapeutic targeting of human bladder tumor-initiating cells
PNAS 106:14016–14021.
https://doi.org/10.1073/pnas.0906549106
- PubMed
- Google Scholar
1. Chao MP
2. Alizadeh AA
3. Tang C
4. Myklebust JH
5. Varghese B
6. Gill S
7. Jan M
8. Cha AC
9. Chan CK
10. Tan BT
11. Park CY
12. Zhao F
13. Kohrt HE
14. Malumbres R
15. Briones J
16. Gascoyne RD
17. Lossos IS
18. Levy R
19. Weissman IL
20. Majeti R
(2010) Anti-CD47 antibody synergizes with rituximab to promote phagocytosis and eradicate non-Hodgkin lymphoma
Cell 142:699–713.
https://doi.org/10.1016/j.cell.2010.07.044
- PubMed
- Google Scholar
1. Chen G
2. Huang AC
3. Zhang W
4. Zhang G
5. Wu M
6. Xu W
7. Yu Z
8. Yang J
9. Wang B
10. Sun H
11. Xia H
12. Man Q
13. Zhong W
14. Antelo LF
15. Wu B
16. Xiong X
17. Liu X
18. Guan L
19. Li T
20. Liu S
21. Yang R
22. Lu Y
23. Dong L
24. McGettigan S
25. Somasundaram R
26. Radhakrishnan R
27. Mills G
28. Lu Y
29. Kim J
30. Chen YH
31. Dong H
32. Zhao Y
33. Karakousis GC
34. Mitchell TC
35. Schuchter LM
36. Herlyn M
37. Wherry EJ
38. Xu X
39. Guo W
(2018) Exosomal PD-L1 contributes to immunosuppression and is associated with anti-PD-1 response
Nature 560:382–386.
https://doi.org/10.1038/s41586-018-0392-8
- PubMed
- Google Scholar
1. Chen M
2. Wang Y
3. Wang H
4. Sun L
5. Fu Y
6. Yang Y-G
(2019) Elimination of donor CD47 protects against vascularized allograft rejection in mice
Xenotransplantation 26:e12459.
https://doi.org/10.1111/xen.12459
- PubMed
- Google Scholar
1. Chen H
2. Cong X
3. Wu C
4. Wu X
5. Wang J
6. Mao K
7. Li J
8. Zhu G
9. Liu F
10. Meng X
11. Song J
12. Sun X
13. Wang X
14. Liu S
15. Zhang S
16. Yang X
17. Song Y
18. Yang Y-G
19. Sun T
(2020) Intratumoral delivery of CCL25 enhances immunotherapy against triple-negative breast cancer by recruiting CCR9+ T cells
Science Advances 6:eaax4690.
https://doi.org/10.1126/sciadv.aax4690
- PubMed
- Google Scholar
1. Cho PS
2. Lo DP
3. Wikiel KJ
4. Rowland HC
5. Coburn RC
6. McMorrow IM
7. Goodrich JG
8. Arn JS
9. Billiter RA
10. Houser SL
11. Shimizu A
12. Yang YG
13. Sachs DH
14. Huang CA
(2007) Establishment of transplantable porcine tumor cell lines derived from MHC-inbred miniature swine
Blood 110:3996–4004.
https://doi.org/10.1182/blood-2007-02-074450
- PubMed
- Google Scholar
1. Cooper JA
(1987) Effects of cytochalasin and phalloidin on actin
The Journal of Cell Biology 105:1473–1478.
https://doi.org/10.1083/jcb.105.4.1473
- PubMed
- Google Scholar
1. Deuse T
2. Hu X
3. Gravina A
4. Wang D
5. Tediashvili G
6. De C
7. Thayer WO
8. Wahl A
9. Garcia JV
10. Reichenspurner H
11. Davis MM
12. Lanier LL
13. Schrepfer S
(2019) Hypoimmunogenic derivatives of induced pluripotent stem cells evade immune rejection in fully immunocompetent allogeneic recipients
Nature Biotechnology 37:252–258.
https://doi.org/10.1038/s41587-019-0016-3
- PubMed
- Google Scholar
1. Gao Q
2. Chen K
3. Gao L
4. Zheng Y
5. Yang Y-G
(2016) Thrombospondin-1 signaling through CD47 inhibits cell cycle progression and induces senescence in endothelial cells
Cell Death & Disease 7:e2368.
https://doi.org/10.1038/cddis.2016.155
- PubMed
- Google Scholar
1. Gao L
2. Chen K
3. Gao Q
4. Wang X
5. Sun J
6. Yang Y-G
(2017) Cd47 deficiency in tumor stroma promotes tumor progression by enhancing angiogenesis
Oncotarget 8:22406–22413.
https://doi.org/10.18632/oncotarget.9899
- PubMed
- Google Scholar
(2018) Emerging role of exosomes in allorecognition and allograft rejection
Current Opinion in Organ Transplantation 23:22–27.
https://doi.org/10.1097/MOT.0000000000000489
- PubMed
- Google Scholar
1. Guilliams M
2. Dutertre C-A
3. Scott CL
4. McGovern N
5. Sichien D
6. Chakarov S
7. Van Gassen S
8. Chen J
9. Poidinger M
10. De Prijck S
11. Tavernier SJ
12. Low I
13. Irac SE
14. Mattar CN
15. Sumatoh HR
16. Low GHL
17. Chung TJK
18. Chan DKH
19. Tan KK
20. Hon TLK
21. Fossum E
22. Bogen B
23. Choolani M
24. Chan JKY
25. Larbi A
26. Luche H
27. Henri S
28. Saeys Y
29. Newell EW
30. Lambrecht BN
31. Malissen B
32. Ginhoux F
(2016) Unsupervised high-dimensional analysis aligns dendritic cells across tissues and species
Immunity 45:669–684.
https://doi.org/10.1016/j.immuni.2016.08.015
- PubMed
- Google Scholar
1. Han X
2. Wang M
3. Duan S
4. Franco PJ
5. Kenty JHR
6. Hedrick P
7. Xia Y
8. Allen A
9. Ferreira LMR
10. Strominger JL
11. Melton DA
12. Meissner TB
13. Cowan CA
(2019) Generation of hypoimmunogenic human pluripotent stem cells
PNAS 116:10441–10446.
https://doi.org/10.1073/pnas.1902566116
- PubMed
- Google Scholar
1. Heckl D
2. Kowalczyk MS
3. Yudovich D
4. Belizaire R
5. Puram RV
6. McConkey ME
7. Thielke A
8. Aster JC
9. Regev A
10. Ebert BL
(2014) Generation of mouse models of myeloid malignancy with combinatorial genetic lesions using CRISPR-Cas9 genome editing
Nature Biotechnology 32:941–946.
https://doi.org/10.1038/nbt.2951
- PubMed
- Google Scholar
1. Ide K
2. Wang H
3. Tahara H
4. Liu J
5. Wang X
6. Asahara T
7. Sykes M
8. Yang YG
9. Ohdan H
(2007) Role for CD47-sirpalpha signaling in xenograft rejection by macrophages
PNAS 104:5062–5066.
https://doi.org/10.1073/pnas.0609661104
- PubMed
- Google Scholar
1. Isenberg JS
2. Roberts DD
(2019) The role of CD47 in pathogenesis and treatment of renal ischemia reperfusion injury
Pediatric Nephrology 34:2479–2494.
https://doi.org/10.1007/s00467-018-4123-z
- PubMed
- Google Scholar
1. Jaiswal S
2. Jamieson CHM
3. Pang WW
4. Park CY
5. Chao MP
6. Majeti R
7. Traver D
8. van Rooijen N
9. Weissman IL
(2009) Cd47 is upregulated on circulating hematopoietic stem cells and leukemia cells to avoid phagocytosis
Cell 138:271–285.
https://doi.org/10.1016/j.cell.2009.05.046
- PubMed
- Google Scholar
1. Kamerkar S
2. LeBleu VS
3. Sugimoto H
4. Yang S
5. Ruivo CF
6. Melo SA
7. Lee JJ
8. Kalluri R
(2017) Exosomes facilitate therapeutic targeting of oncogenic KRAS in pancreatic cancer
Nature 546:498–503.
https://doi.org/10.1038/nature22341
- PubMed
- Google Scholar
1. Li Y
2. Zhang M
3. Wang X
4. Liu W
5. Wang H
6. Yang Y-G
(2020) Vaccination with CD47 deficient tumor cells elicits an antitumor immune response in mice
Nature Communications 11:581.
https://doi.org/10.1038/s41467-019-14102-4
- PubMed
- Google Scholar
1. Liu X
2. Pu Y
3. Cron K
4. Deng L
5. Kline J
6. Frazier WA
7. Xu H
8. Peng H
9. Fu YX
10. Xu MM
(2015a) Cd47 blockade triggers T cell-mediated destruction of immunogenic tumors
Nature Medicine 21:1209–1215.
https://doi.org/10.1038/nm.3931
- PubMed
- Google Scholar
1. Liu J
2. Wang L
3. Zhao F
4. Tseng S
5. Narayanan C
6. Shura L
7. Willingham S
8. Howard M
9. Prohaska S
10. Volkmer J
11. Chao M
12. Weissman IL
13. Majeti R
(2015b) Pre-clinical development of a humanized anti-CD47 antibody with anti-cancer therapeutic potential
PLOS ONE 10:e0137345.
https://doi.org/10.1371/journal.pone.0137345
- PubMed
- Google Scholar
1. Martinez-Torres AC
2. Quiney C
3. Attout T
4. Boullet H
5. Herbi L
6. Vela L
7. Barbier S
8. Chateau D
9. Chapiro E
10. Nguyen-Khac F
11. Davi F
12. Le Garff-Tavernier M
13. Moumné R
14. Sarfati M
15. Karoyan P
16. Merle-Béral H
17. Launay P
18. Susin SA
(2015) Cd47 agonist peptides induce programmed cell death in refractory chronic lymphocytic leukemia B cells via PLCγ1 activation: evidence from mice and humans
PLOS Medicine 12:e1001796.
https://doi.org/10.1371/journal.pmed.1001796
- PubMed
- Google Scholar
1. Mateo V
2. Lagneaux L
3. Bron D
4. Biron G
5. Armant M
6. Delespesse G
7. Sarfati M
(1999) Cd47 ligation induces caspase-independent cell death in chronic lymphocytic leukemia
Nature Medicine 5:1277–1284.
https://doi.org/10.1038/15233
- PubMed
- Google Scholar
1. Meijles DN
2. Sahoo S
3. Al Ghouleh I
4. Amaral JH
5. Bienes-Martinez R
6. Knupp HE
7. Attaran S
8. Sembrat JC
9. Nouraie SM
10. Rojas MM
11. Novelli EM
12. Gladwin MT
13. Isenberg JS
14. Cifuentes-Pagano E
15. Pagano PJ
(2017) The matricellular protein TSP1 promotes human and mouse endothelial cell senescence through CD47 and Nox1
Science Signaling 10:eaaj1784.
https://doi.org/10.1126/scisignal.aaj1784
- PubMed
- Google Scholar
1. Navarro-Alvarez N
2. Yang Y-G
(2014) Lack of CD47 on donor hepatocytes promotes innate immune cell activation and graft loss: a potential barrier to hepatocyte xenotransplantation
Cell Transplantation 23:345–354.
https://doi.org/10.3727/096368913X663604
- PubMed
- Google Scholar
1. Nomura S
2. Ariyoshi Y
3. Watanabe H
4. Pomposelli T
5. Takeuchi K
6. Garcia G
7. Tasaki M
8. Ayares D
9. Sykes M
10. Sachs D
11. Johnson R
12. Yamada K
(2020) Transgenic expression of human CD47 reduces phagocytosis of porcine endothelial cells and podocytes by baboon and human macrophages
Xenotransplantation 27:e12549.
https://doi.org/10.1111/xen.12549
- PubMed
- Google Scholar
(2000) Role of CD47 as a marker of self on red blood cells
Science 288:2051–2054.
https://doi.org/10.1126/science.288.5473.2051
- PubMed
- Google Scholar
1. Raposo G
2. Stahl PD
(2019) Extracellular vesicles: a new communication paradigm?
Nature Reviews. Molecular Cell Biology 20:509–510.
https://doi.org/10.1038/s41580-019-0158-7
- PubMed
- Google Scholar
1. Reinhold MI
2. Lindberg FP
3. Plas D
4. Reynolds S
5. Peters MG
6. Brown EJ
(1995) In vivo expression of alternatively spliced forms of integrin-associated protein (CD47)
Journal of Cell Science 108 (Pt 11):3419–3425.
https://doi.org/10.1242/jcs.108.11.3419
- PubMed
- Google Scholar
(2005) Type 3 repeat/C-terminal domain of thrombospondin-1 triggers caspase-independent cell death through CD47/alphavbeta3 in promyelocytic leukemia NB4 cells
Blood 106:658–667.
https://doi.org/10.1182/blood-2004-09-3585
- PubMed
- Google Scholar
1. Sosale NG
2. Ivanovska II
3. Tsai RK
4. Swift J
5. Hsu JW
6. Alvey CM
7. Zoltick PW
8. Discher DE
(2016) “ Marker of self” CD47 on lentiviral vectors decreases macrophage-mediated clearance and increases delivery to SIRPA-expressing lung carcinoma tumors
Molecular Therapy. Methods & Clinical Development 3:16080.
https://doi.org/10.1038/mtm.2016.80
- PubMed
- Google Scholar
1. Takeuchi K
2. Ariyoshi Y
3. Shimizu A
4. Okumura Y
5. Cara-Fuentes G
6. Garcia GE
7. Pomposelli T
8. Watanabe H
9. Boyd L
10. Ekanayake-Alper DK
11. Amarnath D
12. Sykes M
13. Sachs DH
14. Johnson RJ
15. Yamada K
(2021) Expression of human CD47 in pig glomeruli prevents proteinuria and prolongs graft survival following pig-to-baboon xenotransplantation
Xenotransplantation 28:e12708.
https://doi.org/10.1111/xen.12708
- PubMed
- Google Scholar
1. Tena AA
2. Sachs DH
3. Mallard C
4. Yang Y-G
5. Tasaki M
6. Farkash E
7. Rosales IA
8. Colvin RB
9. Leonard DA
10. Hawley RJ
(2017) Prolonged survival of pig skin on baboons after administration of pig cells expressing human CD47
Transplantation 101:316–321.
https://doi.org/10.1097/TP.0000000000001267
- PubMed
- Google Scholar
1. Wang H
2. Madariaga ML
3. Wang S
4. Van Rooijen N
5. Oldenborg PA
6. Yang YG
(2007a) Lack of CD47 on nonhematopoietic cells induces split macrophage tolerance to cd47null cells
PNAS 104:13744–13749.
https://doi.org/10.1073/pnas.0702881104
- PubMed
- Google Scholar
1. Wang H
2. VerHalen J
3. Madariaga ML
4. Xiang S
5. Wang S
6. Lan P
7. Oldenborg PA
8. Sykes M
9. Yang YG
(2007b) Attenuation of phagocytosis of xenogeneic cells by manipulating CD47
Blood 109:836–842.
https://doi.org/10.1182/blood-2006-04-019794
- PubMed
- Google Scholar
1. Wang H
2. Wu X
3. Wang Y
4. Oldenborg PA
5. Yang YG
(2010) Cd47 is required for suppression of allograft rejection by donor-specific transfusion
Journal of Immunology 184:3401–3407.
https://doi.org/10.4049/jimmunol.0901550
- PubMed
- Google Scholar
1. Wang C
2. Wang H
3. Ide K
4. Wang Y
5. Van Rooijen N
6. Ohdan H
7. Yang Y-G
(2011) Human CD47 expression permits survival of porcine cells in immunodeficient mice that express SIRPα capable of binding to human CD47
Cell Transplantation 20:1915–1920.
https://doi.org/10.3727/096368911X566253
- PubMed
- Google Scholar
1. Wang Y
2. Wang H
3. Bronson R
4. Fu Y
5. Yang Y-G
(2014) Rapid dendritic cell activation and resistance to allotolerance induction in anti-CD154-treated mice receiving CD47-deficient donor-specific transfusion
Cell Transplantation 23:355–363.
https://doi.org/10.3727/096368912X661346
- PubMed
- Google Scholar
1. Watanabe H
2. Ariyoshi Y
3. Pomposelli T
4. Takeuchi K
5. Ekanayake-Alper DK
6. Boyd LK
7. Arn SJ
8. Sahara H
9. Shimizu A
10. Ayares D
11. Lorber MI
12. Sykes M
13. Sachs DH
14. Yamada K
(2020) Intra-bone bone marrow transplantation from hcd47 transgenic pigs to baboons prolongs chimerism to > 60 days and promotes increased porcine lung transplant survival
Xenotransplantation 27:e12552.
https://doi.org/10.1111/xen.12552
- PubMed
- Google Scholar
1. Weiskopf K
2. Jahchan NS
3. Schnorr PJ
4. Cristea S
5. Ring AM
6. Maute RL
7. Volkmer AK
8. Volkmer J-P
9. Liu J
10. Lim JS
11. Yang D
12. Seitz G
13. Nguyen T
14. Wu D
15. Jude K
16. Guerston H
17. Barkal A
18. Trapani F
19. George J
20. Poirier JT
21. Gardner EE
22. Miles LA
23. de Stanchina E
24. Lofgren SM
25. Vogel H
26. Winslow MM
27. Dive C
28. Thomas RK
29. Rudin CM
30. van de Rijn M
31. Majeti R
32. Garcia KC
33. Weissman IL
34. Sage J
(2016) CD47-blocking immunotherapies stimulate macrophage-mediated destruction of small-cell lung cancer
The Journal of Clinical Investigation 126:2610–2620.
https://doi.org/10.1172/JCI81603
- PubMed
- Google Scholar
1. Willingham SB
2. Volkmer JP
3. Gentles AJ
4. Sahoo D
5. Dalerba P
6. Mitra SS
7. Wang J
8. Contreras-Trujillo H
9. Martin R
10. Cohen JD
11. Lovelace P
12. Scheeren FA
13. Chao MP
14. Weiskopf K
15. Tang C
16. Volkmer AK
17. Naik TJ
18. Storm TA
19. Mosley AR
20. Edris B
21. Schmid SM
22. Sun CK
23. Chua MS
24. Murillo O
25. Rajendran P
26. Cha AC
27. Chin RK
28. Kim D
29. Adorno M
30. Raveh T
31. Tseng D
32. Jaiswal S
33. Enger PØ
34. Steinberg GK
35. Li G
36. So SK
37. Majeti R
38. Harsh GR
39. van de Rijn M
40. Teng NNH
41. Sunwoo JB
42. Alizadeh AA
43. Clarke MF
44. Weissman IL
(2012) The CD47-signal regulatory protein alpha (sirpa) interaction is a therapeutic target for human solid tumors
PNAS 109:6662–6667.
https://doi.org/10.1073/pnas.1121623109
- PubMed
- Google Scholar
1. Yang Y-G
2. Sykes M
(2007) Xenotransplantation: current status and a perspective on the future
Nature Reviews. Immunology 7:519–531.
https://doi.org/10.1038/nri2099
- PubMed
- Google Scholar
1. Yang Y-G
(2010) Cd47 in xenograft rejection and tolerance induction
Xenotransplantation 17:267–273.
https://doi.org/10.1111/j.1399-3089.2010.00601.x
- PubMed
- Google Scholar
1. Zeng F
2. Morelli AE
(2018) Extracellular vesicle-mediated MHC cross-dressing in immune homeostasis, transplantation, infectious diseases, and cancer
Seminars in Immunopathology 40:477–490.
https://doi.org/10.1007/s00281-018-0679-8
- PubMed
- Google Scholar
1. Zhang M
2. Wang H
3. Tan S
4. Navarro-Alvarez N
5. Zheng Y
6. Yang Y-G
(2016) Donor CD47 controls T cell alloresponses and is required for tolerance induction following hepatocyte allotransplantation
Scientific Reports 6:26839.
https://doi.org/10.1038/srep26839
- PubMed
- Google Scholar

Decision letter

Simon Yona

Reviewing Editor; The Hebrew University of Jerusalem, Israel
Satyajit Rath

Senior Editor; Indian Institute of Science Education and Research (IISER), India
Dennis Discher

Reviewer; University of Pennsylvania, United States

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "CD47 cross-dressing by extracellular vesicles expressing CD47 inhibits phagocytosis without transmitting cell death signals" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Satyajit Rath as the Senior Editor. The following individuals involved in the review of your submission have agreed to reveal their identity: Dennis Discher (Reviewer #3).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential revisions:

1) Figure 1

The rationale for the use of CD47 isoforms 2 and 4 requires further detail a full explanation.

Please provide kinetics or the dynamic of CD47 acquisition in recipient cells.

Is cross-dressing blocked at 4deg?

Does cross-dressing require cell-to-cell contact?

Does cross-dressing require a functional cytoskeleton?

Do other membrane proteins get transferred as well or is it a peculiarity of CD47h2/h4 to be embedded in cross-dressing vesicles released from donor cells?

Does cross-dressed CD47 resist to EDTA washes (one might anticipate that tethered CD47 would not resist while CD47 actually transferred by membrane fusion would)?

What is the persistence of crossdressed CD47 on recipient cells (post-separation from donor cells)?

b) Why "ligation of the autogenous, but not cross-dressed, CD47 induced cell death." (ii) Please show the extracellular vesicles and/or exosomes, do or do not transfer hCD47 mRNA or even DNA (e.g. plasmid, even episomal). (iii) Determine the reason(s) and implications for function (e.g. phagocytosis) for the very broad distribution of intensities on cells that supposedly have 'all' received hCD47. (iv) Show whether "florescence Celltrace violet" transfers from labelled cells to unlabeled cells (Figure 1c). (v) Determine why results for h2 and h4 spliceforms are variable; sometimes the same (e.g. Figure 1a) versus sometimes different (Figure 1b).

c) In Figure 1A-B: How did the authors gate on the PAOCp in the coculture? Congenic marker, cell tracker? 1C is the right way to do it as surface markers/ membrane stains can't be relied upon. The histograms may be misleading as the shifted curves can include both the acceptor cells that capture CD47h and donor cells that had reduced CD47h. 1D: It's not possible to say that it's fully SIRP-a independent as the increase in the CD47 cross-dressing looks more modest compared to 1-A,B,C. However, this may be due to an intrinsic difference between the starting levels of CD47 on tg LCL and PAOC. Yet, this reviewer recommends changing the statement at line 131 as such: "…., indicating that hCD47 cross-dressing can occur in a SIRPα-independent manner."

2) Figure 2

In Figure 2 Please address labelling and gating the cells in the coculture (see comments 1c).

3) Figure 3

The titration of EVs/Exos should be performed to highlight the efficiency of the cross-dressing process.

The yield of EV/Exos production is explained in order to get an idea if the concentration/quantity of EVs correspond to a reasonable amount of cells and how this number compares to the number of cross-dressed "recipient" cells.

Does membrane acquisition upon EV/Exos involve membrane fusion?

(this can be addressed using fluorescent membrane probes).

In Figure 3C: Why did the authors cocultured cells for 18 and 42 hours? At 42h PAOC cells line would be expected to divide and dilute the CTV, why didn't they? Also, the fitness of the cells might be impaired at the late time points, have the authors performed a viability staining and gated out the dead cells?

4) Figure 4

It is unclear if CD14+CTV+ events all correspond to phagocytosis or doublets formation upon tethering of both cell types. Imaging experiments using confocal microscopy on sorted CD14+CTV+ events could easily quantify this. Please add quantitative imaging data.

This assay requires rigorous viability staining, a robust gating strategy to identify the donor and acceptor cells of the co-culture and controls with single cells. Without those, how do the authors claim that the CTV+ cells in the gate are monocytes that had phagocytosed the CTV-labeled cells?

4B: The population sorted is likely a mix of donor and acceptor cells and contains CD47 expressing cells too. This should have been carefully carried out by labeling the cells with cell trackers prior to culture and sorting them according to the label.

4C: What does the CD14 staining of PAOC cells before and after coculture look like? In systems where cross-dressing occurs, it's hard to rely on the surface markers as they have mobility between cells. Have the authors considered labelling donor and acceptor cells with two different cell trackers to set the gates properly and independent of surface markers?

5) Figure 5

SIRPa-Fc binding on should be assessed on EV-exposed KO cells.

The assay could be performed using additional cross-linking of SIRPa-Fc protein by anti-Fc Fab'2 or plate-bound SIRPa-Fc protein.

Why cross-dressing does not happen between Jurkat WT and CD47ko, thereby sensitizing the CD47ko Jurkat to SIRPa-induce apoptosis?

Complementation of CD47ko clones with CD47 would provide a nice control ruling out possible artefacts of CRISPR targeting.

6) PMID: 28053997 needs to be properly cited, and the author's need further data to support their claim that "hCD47 cross-dressing is SIRPα-independent". Specifically, they need to quantify 'cross-dressing' by non-phagocytic cells that express a suitable hSIRPa in the presence or absence of SIRPa blocking Ab.

7) While Figure S5 nicely shows that SIRPa-Fc causes apoptosis on the cells that express CD47 in a dose-dependent manner, the actual Figure 5 fails to support the claim that EV-based crossdressing protects the cells from apoptosis. Crossdressing doesn't seem to cause a notable increase in hCD47 expression on R1 cells, so the difference in apoptosis between R1 and R2 is likely due to the comparably low expression of CD47 in R1 despite the EV treatment.

https://doi.org/10.7554/eLife.73677.sa1

Author response

Essential revisions:

1) Figure 1

The rationale for the use of CD47 isoforms 2 and 4 requires further detail a full explanation.

There are 4 alternatively spliced CD47 isoforms that differ from each other at their intracytoplasmic carboxy termini (Reinhold et al., 1995). Among these isoforms, isoform 2 is most widely expressed in all cells, including both hematopoietic and nonhematopoietic cells, and isoform 4 is the isoform with longest intracellular domain. Because the difference among 4 CD47 isoforms is at the intracytoplasmic carboxy termini, these isoforms are all capable of interacting with SIRPα to inhibit phagocytosis. The purpose to test two isoforms was to rule out the possibility that our findings are isoform-specific, and choosing isoforms 2 and 4 were because of their relatively unique features (most wide expression for isoform 2 and longest intracytoplasmic tail for isoform 4), and both were used for making transgenic expression of human CD47 to prevent rejection (Ide et al., 2007, Hosny et al., 2021). We have now pointed this in the revised manuscript (in the Results section).

Please provide kinetics or the dynamic of CD47 acquisition in recipient cells.

The data presented in the previous manuscript (Figure 3A and 3B) showed that the level of CD47 cross-dressing on recipient cells remained unchanged from 2 to 6 hours. New experiments were performed to further determine the kinetics of CD47 acquisition, in which pig LCL cells were cultured with EVs from hCD47-transgenic LCL cells or WT Jurkat cells. We found that the process of CD47 cross-dressing is fast, significant CD47 acquisition was detected 1 hour, and the levels remained stable or increased at 24 hours after culturing with EVs (Figure 3—figure supplement 1 in the revised manuscript).

Is cross-dressing blocked at 4deg?

To address this question, we performed a new experiment to compare CD47 crossdressing on pig LCL cells that were incubated with EVs of hCD47-LCL cells at 37°C or 4°C, and the results showed clearly that CD47 cross-dressing was significantly blocked at 4°C (Author response image 1).

Author response image 1

Download asset Open asset

CD47 cross-dressing is significantly blocked at 4 °C. EVs from hCD47-LCL cells were labeled with PKH26, and recipient WT pig LCL cells were labeled with PKH67, EVs and recipient cells were cocultured for 6h and washed, then analyzed for PKH26 (A) and hCD47 (B) cross-dressing on recipient cells by flow cytometry using anti-hCD47-AF647 mAb.Results shown are representative of three independent experiments.

Does cross-dressing require cell-to-cell contact?

Cell-to-cell contact is not required for cross-dressing, as CD47 cross-dressing also occurred when cells were incubated with EVs or Exos.

Does cross-dressing require a functional cytoskeleton?

To address this question, we performed new experiments to determine whether CD47 cross-dressing can be blocked by treatment with Cytochalasin D, a cell-permeable potent inhibitor of actin polymerization that disrupts actin microfilaments. We found that CD47 cross-dressing was clearly detected on cells pre-treated with Cytochalasin D, but to lower levels compared to untreated cells. The results indicate that a functional cytoskeleton is not absolutely required for, but may facilitate CD47 cross-dressing. We have now presented these new data in the revised manuscript (Figure 2—figure supplement 3 in the revised manuscript).

Do other membrane proteins get transferred as well or is it a peculiarity of CD47h2/h4 to be embedded in cross-dressing vesicles released from donor cells?

Other membrane proteins on EVs should be detected on recipient cells cross-dressed by EVs, and it is unlikely that EVs express only CD47. In support of this possibility, our new data showed that, not only hCD47, but also hCD45, hCD90 and HLA-ABC proteins were detected on pig LCL cells after incubated with WT Jurkat cells. We have now pointed this out in the revised manuscript (Figure 2—figure supplement 2).

Does cross-dressed CD47 resist to EDTA washes (one might anticipate that tethered CD47 would not resist while CD47 actually transferred by membrane fusion would)?

New experiments were performed to address this question, in which donor and recipient cells that were labeled with different fluorescence dyes and cultured for 24 hours, following by washing with PBS buffer or PBS buffer containing trypsin/EDTA, then analyzed gated recipient cells for fluorescence acquired from the donor cells. We found that treatment with trypsin/EDTA did not reduce the fluorescence intensity, indicating that membrane fusion is involved in CD47 cross-dressing. We have now included these data and discussed this issue in the revised manuscript (Figure 3—figure supplement 2).

What is the persistence of crossdressed CD47 on recipient cells (post-separation from donor cells)?

To address this question, CD47-transgenic LCL or WT Jurkat cells were labeled with PHK26 and co-cultured with PKH67 labeled pig LCL cells for 1 day, then recipient cells (PKH67+) are sorted and treated with mitomycin C (2ug/ml) for 30min (to prevent dilution of cross-dressed proteins by cell division), and cultured in media to determine the persistence of cross-dressed hCD47. Although the level (MFI) of hCD47 on recipient LCL cells gradually decreased, >50% remained 3 days after culturing, indicating that CD47 cross-dressing is quite stable and may persist for several days. We have now pointed this out in the revised manuscript (Figure 2C).

b) i) Why "ligation of the autogenous, but not cross-dressed, CD47 induced cell death."

Autogenous CD47 contains transmembrane and intracytoplasmic domains that are required for transferring death signals. Although this study does not tell the structure of hCD47 cross-dressed on the recipient cells, the failure of cross-dressed CD47 to induce apoptosis indicates that cross-dressed CD47 cannot effectively deliver the intracytoplasmic signaling, leading to activation of apoptotic signal cascades.

(ii) Please show the extracellular vesicles and/or exosomes, do or do not transfer hCD47 mRNA or even DNA (e.g. plasmid, even episomal).

We have analyzed and found that the EVs contain hCD47 DNA and RNA, as well as housekeeping gene GAPDH RNA (Author response image 2). Although we do not have direct evidence, these data suggest that DNA and RNA within EVs might also be transferred to the recipient cells. Indeed, unlike protein delivery that can be tracked by fluorescent dye labeling, progress on understanding DNA and RNA cargo delivery by EVs has been hampered by the lack of efficient and specific tracking methods(O’Brien et al., 2020). We agree that elucidating this issue, including the efficacy of DNA and RNA transfer, their persistence and functional significance, will be helpful in understanding the insights into EV cross-dressing, and would like to perform these studies in the future. Nonetheless, colocalization of CD47 staining and cell-tracker dye fluorescence on the recipient cells (Author response image 3) confirmed that the majority of CD47 on the recipient cells were cross-dressed CD47 proteins and not derived from DNA or RNA cargo conveyed in EVs.

Author response image 2

Download asset Open asset

EVs from pig LCL/CD47^p/h cells contain both hCD47 DNA and RNA, and GAPDH RNA.

DNA and total RNA are extracted from hCD47 transgenic pig LCL and their MVs, and WT pig LCL cells respectively, then cDNA was synthesized from total RNA using a reverse transcription kit, and hCD47 and GAPDH sequences are amplified by PCR. Shown are Agarose gel electrophoresis of hCD47 genomic DNA (A), hCD47 cDNA (B) and GAPDH cDNA PCR products..

Author response image 3

Download asset Open asset

Colocalization of CD47 proteins and EVs on pig LCL cells.

PKH26-labeled EVs from LCL/CD47^p/h cells were co-cultured with pig LCL cells for 6h, then stained by anti-huCD47-AF647 antibody and analyzed by confocal microscope. Shown are CD47 staining (blue; left), EVs (PKH26 red fluorescence, mid) and overlayed (right) images.

(iii) Determine the reason(s) and implications for function (e.g. phagocytosis) for the very broad distribution of intensities on cells that supposedly have 'all' received hCD47.

Although the distribution of cross-dressed CD47 is relatively broader than autogenous CD47, the distribution patterns in most experiments were similar between these cells. However, different patterns were seen in some experiments (Figure 1E; Figure 3). While we do not fully understand the reasons, the difference may possibly be due to variable hCD47 on the donor cells for the experiment in Figure 1E, in which primary hCD47-tg pig BM cells that express highly variable levels of hCD47 were used as the donor cells; and for the experiments shown in Figure 3, insufficient EVs or Exos in cultures might be attributable to it. Furthermore, functional heterogeneity in proliferation and senescence may also affect the efficacy of crossdressing. Nonetheless, we agree that the level of CD47 cross-dressing should have an impact on the degree of protection against phagocytosis.

(iv) Show whether "florescence Celltrace violet" transfers from labelled cells to unlabeled cells (Figure 1c).

It is expected that fluorescence dye that labeling membrane proteins should also be transferred with EVs from labeled cells to unlabeled cells, but it was difficult to see it in Figure 1C, in which no negative controls (i.e., uncultured cells) were analyzed at the same time. To clarify this, we performed a new experiment, in which we used the cells that were mixed before flow cytometry analysis as the negative control. As shown in Author response image 4, fluorescence dye transfer was clearly detected in co-cultured cells compared to the controls (i.e., a mixture of labeled and unlabeled cells that were mixed before flow cytometric analysis), but the labeled cells could be easily separated from cross-dressed unlabeled cells on flow cytometric plots. In these experiments, we also stained the cells with anti-hCD47 antibody, and the results confirmed that both fluorescence dye and anti-hCD47 staining can clearly separate the donor cells from cross-dressed recipient cells, and the results obtained by these methods are identical (this also address the reviewer’s comment #1c below). Furthermore, anti-CD47 staining and fluorescence dye used are in different channels (i.e., different colors), so fluorescence dye-crossdressing is not a confounding factor for the experiments in which cells were also stained by anti-hCD47 antibody.

Author response image 4

Download asset Open asset

Florescence dye cross-dressing.

PKH26-labeled hCD47-LCL (A) or PKH26labeled WT Jurkat (B) as donor cells were co-cultured, respectively, with PKH67-labeled pig LCL for 24h, and then analyzed for fluorescent dye cross-dressing by flow cytometry using anti-hCD47AF647 mAb. Two types of cells mixed immediately prior to flow cytometry analysis (without co-culture) were used as negative controls. Shown are flow cytometry profiles (left panels) and median fluorescence intensities (right panels). Results shown are representative of two independent experiments.

(v) Determine why results for h2 and h4 spliceforms are variable; sometimes the same (e.g. Figure 1a) versus sometimes different (Figure 1b).

As described in the manuscript, different donor cells were used in the experiments shown in Figure 1A and Figure 1B. In Figure 1A, the donor cells were PAOCs that were genetically modified to express hCD47 isoform 2 (i.e., expressed both pig and human isoform 2 CD47 and referred to as PAOC/CD47^p/h2) or isoform 4 (i.e., i.e., expressed both pig and human isoform 4 CD47 and referred to as PAOC/CD47^p/h4). However, in Figure 1B, the donor cells were CD47defficient PAOC (PAOC/CD47^null) cells that were genetically modified to express transgenic hCD47 isoform 2 (PAOC/CD47^h2) or isoform 4 (PAOC/CD47^h4), which express only human CD47 (isoform 2 or isoform 4) but no pig CD47. It is perceivable that different transgenic cell lines may express different levels of the transgene. Indeed, the results for each cell line were considerably stable throughout the experiments.

c) In Figure 1A-B: How did the authors gate on the PAOCp in the coculture? Congenic marker, cell tracker? 1C is the right way to do it as surface markers/ membrane stains can't be relied upon. The histograms may be misleading as the shifted curves can include both the acceptor cells that capture CD47h and donor cells that had reduced CD47h. 1D: It's not possible to say that it's fully SIRP-a independent as the increase in the CD47 cross-dressing looks more modest compared to 1-A,B,C. However, this may be due to an intrinsic difference between the starting levels of CD47 on tg LCL and PAOC. Yet, this reviewer recommends changing the statement at line 131 as such: "…., indicating that hCD47 cross-dressing can occur in a SIRPα-independent manner."

We agree with the reviewer, and that is why we performed the experiment in Figure 1C to validate the results shown in Figure 1A-B. As we discussed in the manuscript (line 106-112 in the revised manuscript), “To rule out the possibility that, during the co-culture, the CD47^null cells did not become CD47+, but PAOC/CD47^h2 cells reduced hCD47 expression, we labeled PAOC/CD47^null (Figure 1C, top) or PAOC/CD47^h2 (Figure 1C, bottom) cells with florescence Celltrace violet, and then cocultured the labeled cells with unlabeled PAOC/CD47^h2 or PAOC/CD47^null cells, respectively. This experiment, in which fluorescence-labeling allowed for better distinguishing between the two cell populations in the cocultures, further confirmed that PAOC/CD47^null cells can be cross-dressed by CD47 after coculture with PAOC/CD47^h2 cells (Figure 1C).” The results from the experiment in Figure 1C confirmed that PAOC^null and PAOC^h2/h4 cells can be clearly separated by anti-hCD47 antibody staining. This conclusion is also supported by the results of our new experiments that were performed to address the reviewer’s comments on fluorescence dye transfer (Comment #1b-iv).

We agree with the reviewer that our data cannot rule out the possibility of SIRPα-mediated crossdressing, and therefore, as the reviewer suggested, revised our statement to “Although we cannot rule out of the role of SIRPα in hCD47 cross-dressing, our data indicate that hCD47 cross-dressing can occur in a SIRPα-independent manner” in the revised manuscript.

2) Figure 2

In Figure 2 Please address labelling and gating the cells in the coculture (see comments 1c).

In the experiments shown in Figure 2, CD47KO Jurkat and pig LCL recipient cells were gated by anti-hCD47 staining. As discussed in our responses to Comment #1c and #1b-iv above, we have firmly confirmed that anti-hCD47 staining can clearly distinguish cross-dressed recipient cells from donor cells by using fluorescence dye labeling as the reviewer suggested.

3) Figure 3

The titration of EVs/Exos should be performed to highlight the efficiency of the cross-dressing process.

The yield of EV/Exos production is explained in order to get an idea if the concentration/quantity of EVs correspond to a reasonable amount of cells and how this number compares to the number of cross-dressed "recipient" cells.

Does membrane acquisition upon EV/Exos involve membrane fusion?

(this can be addressed using fluorescent membrane probes).

We thank the reviewer for pointing out this question. In our study, we used a simple centrifugation method to prepare EVs and Exos, which may not efficiently recover entire and pure EVs or Exos, so the results may not be very helpful in addressing the reviewer’s question. However, in the cell co-culture experiments, the donor to recipient cell ratios ranged from 1:1 to 1:3, which are considered reasonable and can be expected to occur in real in vivo situation, as for both organ/cell transplantation and cancer conditions, the numbers of graft- or tumor-infiltrating recipient cells are expected to be less than the numbers of graft or tumor cells. Thus, our results are of clinical significance.

The results of our new experiments indicate that membrane fusion is involved in membrane acquisition of EVs/Exos. We found that treatment with Trypsin/EDTA did not reduce crossdressed fluorescence dye in the recipient cells. We have now included these new data and discussed this issue in the revised manuscript (Figure 3—figure supplement 2).

In Figure 3C: Why did the authors cocultured cells for 18 and 42 hours? At 42h PAOC cells line would be expected to divide and dilute the CTV, why didn't they? Also, the fitness of the cells might be impaired at the late time points, have the authors performed a viability staining and gated out the dead cells?

PAOC^null cells are adherent endothelial cell line cells, and their proliferation is slower than tumor cell lines used (i.e., Jurkat and LCL cells). We found that EV acquisition by POAC cells is slower than Jurkat and LCL cells, assumably due to their different properties (for example the adherent property) and therefore, these cells were cultured for a longer period than Jurkat and LCL cells. The relatively lower hCD47 staining at 42 hours than 18 hours may be, at least partially, attributed to dilution by cell proliferation, as pointed out by the reviewer. Nonetheless, hCD47 cross-dressing was clearly detected in POAC cells in our experiments, and the data support our conclusion. We have confirmed that PAOC cells were in good conditions with 95% alive at 48h.

Also, dead cells were gated out by PI or 7-AAD staining in all flow cytometry analysis.

4) Figure 4

It is unclear if CD14+CTV+ events all correspond to phagocytosis or doublets formation upon tethering of both cell types. Imaging experiments using confocal microscopy on sorted CD14+CTV+ events could easily quantify this. Please add quantitative imaging data.

This assay requires rigorous viability staining, a robust gating strategy to identify the donor and acceptor cells of the co-culture and controls with single cells. Without those, how do the authors claim that the CTV+ cells in the gate are monocytes that had phagocytosed the CTV-labeled cells?

This flow cytometry-based phagocytosis assay is well established and used routinely in the lab, in which all dead (PI+ or 7-AAD+) cells are excluded from the first gating, then doublets are excluded by gating on FSC-A vs FSC-H plots (Author response image 5). This method has also been validated by imaging experiments using confocal microscopy as the reviewer suggested, confirming that CVT+CD14+ cells are macrophages that have engulfed CTV+ target cells (Author response image 6).

Author response image 5

Download asset Open asset

Gating strategy for analyzing phagocytosis by macrophages.

All PI+ cells are excluded from the first gating, then doublets are excluded by gating on FSC-A vs FSC-H. Macrophages were gated on FSC-A vs SSC-A, then by CD45 and CD14 expression.

Author response image 6

Download asset Open asset

Confocal microscopic analysis of phagocytosis.

CFSE-labeled CD47KO Jurkat cells (green) were cocultured with human macrophages for 2h, floating cells were washed out by PBS, then macrophages were harvested by trypsin/EDTA digestion and stained by anti-huSIRPα-APC (red) antibody and DAPI (blue). The samples were subjected to confocal laser scanning microscopy, and the results confirmed that target cells were engulfed by human macrophages. Shown are representative confocal images.

4B: The population sorted is likely a mix of donor and acceptor cells and contains CD47 expressing cells too. This should have been carefully carried out by labeling the cells with cell trackers prior to culture and sorting them according to the label.

As discussed in our responses to the reviewers’ comments (#1b-iv and #1c) above, we have validated this method using fluorescent CTV labeling and confirmed that anti-hCD47 staining can reliably separate donor cells from recipient cells with hCD47 cross-dressing. Thus, the gated population should have extremely low contamination by donor cells. Indeed, this was also confirmed by the donor (PAOC47^h2) cell alone control included in this experiment. As shown in Figure 4B (middle), there were only <1% of cells in the gated region (R2) in the cultures of PAOC47^h2 only.

4C: What does the CD14 staining of PAOC cells before and after coculture look like? In systems where cross-dressing occurs, it's hard to rely on the surface markers as they have mobility between cells. Have the authors considered labelling donor and acceptor cells with two different cell trackers to set the gates properly and independent of surface markers?

We have confirmed that anti-human CD14 antibody does not cross-react with pig cells, and POAC cells after co-cultured with human macrophages remained negative relevant to human macrophages (Author response image 7).

Author response image 7

Download asset Open asset

Anti-human CD14 antibody does not cross-react with PAOC cells.

PAOC cells were cocultured with human macrophages and analyzed for human CD14 and CD45 expression. Representative flow cytometry profiles showing staining of anti-hCD14 and anti-hCD45. The results conformed that after coculture, POAC cells and human macrophages could be clearly separated by staining with antihuman CD14 and CD45 antibodies on flow cytometry.

5) Figure 5

SIRPa-Fc binding on should be assessed on EV-exposed KO cells.

We have assessed SIRPα-Fc binding on EV-exposed CD47KO Jurkat cells as the reviewer suggested (Figure 4A). We also assessed phagocytosis by human macrophages (Figure 4) and SIRPα-Fc-induced apoptosis (Figure 5E-F) of EV-exposed KO cells.

The assay could be performed using additional cross-linking of SIRPa-Fc protein by anti-Fc Fab'2 or plate-bound SIRPa-Fc protein.

Although we do not have direct evidence to support this, we would agree with the reviewer that additional cross-linking may possibly enhance apoptosis. Since SIRPα-Fc without additional cross-linking induced significant apoptosis in our experiments, we did not perform any experiments using additional cross-linking.

Why cross-dressing does not happen between Jurkat WT and CD47ko, thereby sensitizing the CD47ko Jurkat to SIRPa-induce apoptosis?

The purpose of this experiment (Figure 5C) was to determine whether SIRPα-Fc may induce apoptosis of CD47KO cells without hCD47 cross-dressing (as controls for the experiment shown in Figure 5E, 5F). In this experiment, two types of cells were cocultured for only 1h in the presence of SIRPa-Fc proteins, so there should be very minimal cross-dressing during the culture period. For determining whether hCD47+ EV cross-dressing may induce apoptosis, CD47KO cells were pre-exposed to hCD47+ EVs before mixed with WT cells for induction of apoptosis by SIRPa-Fc (Figure 5E, 5F) and the results showed that CD47 cross-dressing does not transmit death signaling induced by SIRPα-Fc.

Complementation of CD47ko clones with CD47 would provide a nice control ruling out possible artefacts of CRISPR targeting.

We thank the reviewer for the thoughtful suggestion. To address this comment, we transfected CD47KO Jurkat cells with CD47, and found that CD47KO Jurkat cells transfected with CD47 became susceptible to apoptosis induced by SIRPα-Fc. We have now included these data in the revised manuscript (Figure 5—figure supplement 3).

6) PMID: 28053997 needs to be properly cited, and the author's need further data to support their claim that "hCD47 cross-dressing is SIRPα-independent". Specifically, they need to quantify 'cross-dressing' by non-phagocytic cells that express a suitable hSIRPa in the presence or absence of SIRPa blocking Ab.

We thank the reviewer for bring this reference to our attention (which has now been cited in the revised manuscript), and we agree that the interaction of CD47 with SIRPα may facilitate transduction of SIRPα+ non-phagocytic cells with CD47-Lenti. In our study, hCD47 cross-dressing occurred in cells that do not express SIRPα, indicating cross-dressing can occur independent of SIRPα. However, our data cannot rule out the role of SIRPα in hCD47 crossdressing, and therefore, we have revised our statement of "hCD47 cross-dressing is SIRPαindependent" to “hCD47 cross-dressing can occur in a SIRPα-independent manner." (Please refer to our response to a similar comment raised by another reviewer (see Comment #1c above)).

7) While Figure S5 nicely shows that SIRPa-Fc causes apoptosis on the cells that express CD47 in a dose-dependent manner, the actual Figure 5 fails to support the claim that EV-based crossdressing protects the cells from apoptosis. Crossdressing doesn't seem to cause a notable increase in hCD47 expression on R1 cells, so the difference in apoptosis between R1 and R2 is likely due to the comparably low expression of CD47 in R1 despite the EV treatment.

As we explained in the manuscript that the low staining of hCD47 in Figure 5 was due to the fact that prior incubation with hSIRPα-Fc proteins can block subsequent staining by antihCD47 mAb (data are presented in the previous Figure S5, which is Figure 5—figure supplement 2 in the revised manuscript). Please see page 15 of the manuscript, where we point out that “Of note, binding of hSIRPα-Fc proteins to cell surface CD47 (either native or cross-dressed) could block subsequent staining with anti-hCD47 antibodies and thus, the cells cultured with hSIRPα-Fc showed relatively lower hCD47 staining than those cultured without (previous Figure S5, which is Figure 5—figure supplement 2 in the revised manuscript, Figure 5).” Data in Figure 4A also showed that EV treatment resulted in a significant hCD47 cross-dressing.

References

Hosny, N., Matson, A. W., Kumbha, R., Steinhoff, M., Sushil Rao, J., Elabaseri, T. B., Sabek, N. A., Mahmoud, M. A., Hering, B. J. & Burlak, C. 2021. 3'UTR enhances hCD47 cell surface expression, self-signal function, and reduces ER stress in porcine fibroblasts. Xenotransplantation, 28, e12641.

Ide, K., Wang, H., Liu, J., Wang, X., Asahara, T., Sykes, M., Yang, Y. G. & Ohdan, H. 2007. Role for CD47-SIRPà signaling in xenograft rejection by macrophages. Proc Natl Acad Sci USA, 104, 5062-5066.

O’brien, K., Breyne, K., Ughetto, S., Laurent, L. C. & Breakefield, X. O. 2020. RNA delivery by extracellular vesicles in mammalian cells and its applications. Nature Reviews Molecular Cell Biology, 21, 585-606.

Reinhold, M. I., Lindberg, F. P., Plas, D., Reynolds, S., Peters, M. G. & Brown, E. J. 1995. in vivo expression of alternatively spliced forms of integrin-associated protein (CD47). J Cell Sci, 108 ( Pt 11), 3419-25.

https://doi.org/10.7554/eLife.73677.sa2

Article and author information

Author details

Yang Li
1. Key Laboratory of Organ Regeneration and Transplantation of the Ministry of Education, The First Hospital, and Institute of Immunology, Jilin University, Changchun, China
2. Columbia Center for Translational Immunology, Columbia University Medical Center, New York, United States
Contribution
Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review and editing

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0003-4623-4489
Yan Wu

Columbia Center for Translational Immunology, Columbia University Medical Center, New York, United States

Contribution
Data curation, Investigation, Writing – review and editing

Competing interests
No competing interests declared
Elena A Federzoni

Lung Biotechnology PBC, Silver Spring, United States

Contribution
Data curation, Formal analysis, Validation, Investigation, Writing – review and editing

Competing interests
No competing interests declared
Xiaodan Wang

Key Laboratory of Organ Regeneration and Transplantation of the Ministry of Education, The First Hospital, and Institute of Immunology, Jilin University, Changchun, China

Contribution
Data curation, Investigation, Writing – review and editing

Competing interests
No competing interests declared
Andre Dharmawan

Lung Biotechnology PBC, Silver Spring, United States

Contribution
Data curation, Investigation, Writing – review and editing

Competing interests
No competing interests declared
Xiaoyi Hu

Columbia Center for Translational Immunology, Columbia University Medical Center, New York, United States

Contribution
Data curation, Formal analysis, Investigation, Writing – review and editing

Competing interests
No competing interests declared
Hui Wang

Columbia Center for Translational Immunology, Columbia University Medical Center, New York, United States

Contribution
Data curation, Investigation, Writing – review and editing

Competing interests
No competing interests declared
Robert J Hawley

Columbia Center for Translational Immunology, Columbia University Medical Center, New York, United States

Contribution
Resources, Supervision, Funding acquisition, Writing – review and editing

Competing interests
No competing interests declared
Sean Stevens

Lung Biotechnology PBC, Silver Spring, United States

Contribution
Resources, Supervision, Funding acquisition, Writing – review and editing

Competing interests
No competing interests declared
Megan Sykes

Columbia Center for Translational Immunology, Columbia University Medical Center, New York, United States

Contribution
Conceptualization, Resources, Supervision, Funding acquisition, Writing – review and editing

Competing interests
No competing interests declared
Yong-Guang Yang
1. Key Laboratory of Organ Regeneration and Transplantation of the Ministry of Education, The First Hospital, and Institute of Immunology, Jilin University, Changchun, China
2. International Center of Future Science, Jilin University, Changchun, China
Contribution
Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Investigation, Writing – original draft, Project administration, Writing – review and editing

For correspondence
yongg@jlu.edu.cn

Competing interests
No competing interests declared

"This ORCID iD identifies the author of this article:" 0000-0003-4985-1247

Funding

National Institutes of Health (AI045897)

Megan Sykes

The First Hospital of Jilin University

Yong-Guang Yang

NIH (PO1 AI045897)

Megan Sykes

NIH Shared Instrumentation (1S10RR027050)

Megan Sykes

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

This work was supported by NIH PO1 AI045897 and a fund from The First Hospital of Jilin University. The CCTI Flow Cytometry Core used was funded in part through an NIH Shared Instrumentation Grant (1S10RR027050).

Senior Editor

Satyajit Rath, Indian Institute of Science Education and Research (IISER), India

Reviewing Editor

Simon Yona, The Hebrew University of Jerusalem, Israel

Reviewer

Dennis Discher, University of Pennsylvania, United States

Version history

Received: September 7, 2021
Preprint posted: October 11, 2021 (view preprint)
Accepted: November 15, 2022
Version of Record published: December 1, 2022 (version 1)

Copyright

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.