Just eat it: A review of CD47 and SIRP-α antagonism

Bryan Oronsky a, Corey Cartera, Tony Reid b, Franck Brinkhaus a, Susan J. Knox c,∗
a EpicentRx, San Diego, California
b Department of Medical Oncology, UC San Diego School of Medicine, San Diego, California
c Department of Radiation Oncology, Stanford University School of Medicine, Stanford, California

a r t i c l e i n f o a b s t r a c t

Article history:
Received 27 February 2020
Revised 4 May 2020
Accepted 4 May 2020

Keywords: CD47 SIRP-α
Antagonists Phagocytosis Tumor
The mammalian immune system consists of two distinct arms, nonspecific innate and more specific adap- tive, with the innate immune response as the first line of defense and protection, which primes and amplifies subsequent adaptive responses. On the basis of this binary immune interplay, stimulation of T cells through checkpoint inhibitors (CIs), which bypasses innate involvement, seems likely to engender suboptimal or incomplete anticancer immunity, given that the successful induction of effect or responses depends on two-way innate/adaptive coordination. Indeed, the majority of patients—70%–80%, do not respond to CIs, which is potentially problematic if access to more optimal standard therapies is with- held or delayed in favor of ineffective or only marginally effective anti-PD-1/PD-L1 treatment. Therefore, stimulation of the innate immune response in combination with CIs (or other inducers of T cell cyto- toxicity) has the potential to make the immune system “whole” and thereby to enhance and broaden the anti-tumor activity of PD-1/PD-L1 inhibitors for example, in relatively nonimmunogenic or “cold” tu- mor types. A critical innate macrophage immune checkpoint and druggable target is the antiphagocytic and “marker of self” CD47-SIRPα pathway, which is co-opted by cancer cells to mediate escape from immune-mediated clearance and checkpoint inhibition. This review summarizes the status of key CD47 antagonists in clinical trials, including the biologics, Hu5F9-G4 (5F9), TTI-621, and ALX148, as well as the small molecule, RRx-001, now in a Phase 3 clinical trial, which has not been previously included in CD47–SIRPα reviews focused on biologics. Hu5F9-G4 (5F9), TTI-621, ALX148, and RRx-001 are chosen as compounds with potentially promising data that have advanced the farthest in clinical development.
© 2020 Elsevier Inc. All rights reserved.

Introduction

New treatment strategies are disparately needed to enhance the efficacy of conventional treatments. The 5 current pillars of can- cer treatment are surgery, radiation, chemotherapy, targeted ther- apy and, more recently, immunotherapy, which is further divided into therapies designed to potentiate either nonspecific innate and more specific adaptive immune responses (Fig. 1).
It is axiomatic that anti-CTLA-4 and anti-PD1/PD-L1 adaptive therapies, either alone or in combination, have revolutionized the field of oncology over the last decade, with dramatic and durable responses—some as long as ten years or more [1] in pre- viously refractory tumors such as metastatic melanoma, leading to the routine use of qualifiers such as “paradigm-shifting”, “game- changing” [2] and “breakthrough” [3]. It is, therefore, less ax- iomatic that in only a subset of tumor histologies (ie, melanoma,
∗ Corresponding author. Department of Radiation Oncology, Stanford Cancer In- stitute, 875 Blake Wilbur Dr. Stanford, CA 94305. Tel.: 650-723-6171.
E-mail address: [email protected] (S.J. Knox).
non–small cell lung cancer (NSCLC), renal cell carcinoma, head and neck squamous cell carcinoma, urothelial carcinoma, microsatellite instability–high colorectal cancer) these therapies have a response rate of 20%–30%. Conversely, 70%–80% of patients do not respond to immune checkpoint blockade, or they initially respond, but their responses are not durable, and they relapse [4]. Tumors recognized as relatively non-responsive to checkpoint inhibitors include ad- vanced pancreatic ductal adenocarcinoma, metastatic prostate can- cer, microsatellite stable colorectal cancer, and malignant pleural mesothelioma.

The reasons for resistance to checkpoint inhibitors are multi- faceted and include low mutational neoantigen burden, the im- munosuppressive tumor microenvironment, and increased expres- sion of other co-inhibitory molecules [5]. In addition, the inability of T-cell checkpoint inhibition to concurrently activate innate path- ways, limits its ability to induce complete antitumor responses, resulting in suboptimal clinical outcomes. Furthermore, approved T cell checkpoint inhibitors are associated with several sequelae including autoimmune toxicities, cytokine release syndrome [6,7] and hyperprogressive disease [8]. Numerous approaches are be- https://doi.org/10.1053/j.seminoncol.2020.05.009 0093-7754/© 2020 Elsevier Inc. All rights reserved. The 5 pillars of cancer treatment. At the present time we envision the ther- apy of cancer as based on five pillars that include the original triad of surgery, radiation and chemotherapy, to which targeted therapy and, more recently, im- munotherapy, have been added. The immunotherapy pillar is further divided into therapies that potentiate nonspecific innate or more specific adaptive immune re- sponses.ing studied to increase the efficacy and mitigate toxicity associated with checkpoint inhibitor therapies so that (1) the therapeutic in- dex is increased for checkpoint inhibitors and more patients can benefit from such therapy, (2) the toxicity profile for checkpoint in- hibitors is acceptable, with a lower rate of grade 3–5 toxicities, and (3) access to other potentially more effective/less toxic standard therapies is not delayed or withheld in favor of ineffective and of- ten monotherapy checkpoint inhibitor use. To date, seven immune checkpoint inhibitors have received US Food and Drug Adminis- tration approval [9]: a CTLA-4 inhibitor (ipilimumab), 3 PD-1 in- hibitors, nivolumab, pembrolizumab, and cemiplimab, and 3 PD-L1 inhibitors, atezolizumab, durvalumab, and avelumab, all of which exclusively stimulate adaptive T cell responses.

As above, natural immune responses are not adaptively one- sided, but rather depend on coordinated innate and adaptive ef- fector cell activity. The tandem operation of the in vivo immune system, which is based on synergistic cooperation, supports the ra- tional combination of the “one-two punch” of adaptive and innate immune checkpoint antagonists vs. T cell CI monotherapy [10]. However, an important caveat is the potential for greater toxici- ties in combination therapy, an example of which was a tripling of the rate of severe toxicities observed with ipilimumab + nivolumab
compared to checkpoint inhibitor monotherapy with these agents. Therefore, it is possible that dual targeting of innate and adaptive checkpoints may have more toxicity than monotherapy [11].

A critical innate macrophage checkpoint is the CD47/Signal- regulatory protein alpha (SIRPα) pathway, a druggable target,
which delivers an antiphagocytic signal to macrophages that in- hibits destruction of cancer cells overexpressing CD47 (Cluster of Differentiation 47). Tumors that overexpress CD47 include acute myeloid leukemia(AML), acute lymphoblastic leukemia, chronic lymphocytic leukemia, multiple myeloma, myelodysplastic syn- drome (MDS), diffuse large B-cell lymphoma (DLBCL), mantle cell lymphoma, and marginal cell lymphoma, as well as bladder, brain, breast, colon, esophageal, gastric, kidney, leiomyosarcoma, liver, lung, melanoma, ovarian, pancreatic, and prostate cancer [12]. Also known as integrin-associated protein, CD47 is a ubiquitously ex- pressed 50 kDa membrane receptor with five-transmembrane do- mains [13] that is co-opted by cancer cells as an anti-phagocytic “don’t eat me” signal. In addition to promoting macrophage- mediated phagocytosis, CD47 antagonism is associated with in- creased dendritic cell [14] and natural killer cell cytotoxicity, which contributes to the heightened interest that CD47/SIRPα antagonism has generated [15]. However, since CD47 is expressed on all cells, particularly red blood cells (RBCs), neutrophils [16] and platelets [17], and CD47 blockade leads to clearance of these cells by splenic macrophages, anemia, neutropenia and thrombocytopenia are ex- pected complications of CD47 or SIRPα antagonists. Two anti-CD47 antibodies, SRF231 (Surface Oncology), and CC-90002 (Celgene) have been deprioritized/discontinued due to significant hemato- logic toxicities. In contrast, the small molecule RRx-001 does not appear to have hematologic toxicity.
In this review, available information on the status of CD47 antagonists, which are intended to enhance tumor rejection, is summarized. Key CD47/SIRPα antagonists in clinical trials, all of

hich are biologics except for the small molecule, RRx-001, in- clude Hu5F9-G4 (5F9), a high affinity humanized IgG4 anti-CD47 antibody, TTI-621, a fusion of the CD47-binding domain of SIRPα and an IgG1 Fc region, ALX148, a SIRPα domain fused to an in- ert IgG1 Fc, and RRx-001, a small molecule that can downregulate both CD47 and SIRPα. These compounds, with potentially promis- ing data, have progressed the farthest in active clinical develop- ment.

CD47/SIRPα in macrophages

In contrast to T-cells that tend to minimally infiltrate solid tu- mors [18], due to high-interstitial fluid pressure, aberrant vascula- ture and dense fibrotic tissue, chemotactic/migratory macrophages are found abundantly in situ [19]. The abundance of macrophages in some tumors affords the opportunity for therapeutic interven- tion, especially given the correlation between a high infiltrate of tumor-associated macrophages (TAMs) and the development of therapeutic resistance with its poor prognosis. Macrophages are classically described in terms of an M1/M2 dichotomy: M1 or “good” macrophages kill tumor cells through generation of reac- tive oxygen/nitrogen species (ROS/RNS) and proinflammatory cy- tokines such as IL-1β, IL-6, and tumor necrosis factor α (TNF- α) [20], whereas M2-like or “bad” macrophages, also referred to TAMs facilitate repair, and hence, their presence is thought to serve as an obstacle to successful anticancer treatment and portend an unfavorable prognosis (Table 1) [21,22]. However, the M1/M2 di- chotomization/polarization is not rigid and unchanging; rather the macrophage phenotype is pliable and dynamic and switches re- versibly in response to microenvironmental cues and context. With the repolarization of TAMs from M2 to M1, their tumor promoting functions that is, angiogenesis, fibrosis and immunosuppression are suppressed in favor of elimination of tumor cells, inhibition of an- giogenesis and less fibrosis [23].
The tumor microenvironment elaborates a number of cytokines and chemokines, such as colony-stimulating factor-1 (CSF-1), vas- cular endothelial growth factor, C-C motif chemokine ligand 2 (CCL2), interleukins -4 and -13 (IL-4, IL-13), transforming growth factor-β (TGF- β), and IL-10 [24], which lead to M2-like dif- ferentiation and a protumor, “do not eat” phenotype.
In contrast, M1 macrophages eliminate cancer cells through phagocy- tosis based on the presence of multiple “eat-me” signals. The latter signals include (1) exposure of phosphatidylserine, (2) the interaction between calreticulin, a molecule overexpressed on cancer cells and the low-density lipoprotein receptor-related protein on macrophages [25], and (3) the interaction between antibody-coated tumor cells and Fc gamma receptors (Fcγ Rs) on macrophages (Fig. 2) [26]. Antibodies of the IgG class possess a variable F(ab) domain, responsible for the binding of antigen, and a constant Fc domain, which engages with the Fcγ R on macrophages and elicits both antibody dependent cell phagocyto- sis (ADCP) and antibody dependent cell cytotoxicity [27]. In terms of relative potencies of ADCP based on affinity of binding to Fcγ R, IgG1=IgG3>>IgG4>IgG2 [28]. For CD47 and SIRPα blocking an- tibodies with effector function competent Fc regions, the Fc IgG

Table 1
Macrophages.

⦁ M1 Macrophages • M2 Macrophages

Phenotype • Eliminate cancer cells through phagocytosis based on the presence of multiple “eat-me” signals
⦁ Proinflammatory phenotype with pathogen-killing abilities
⦁ Pro-tumor, “do not eat”
⦁ Promote cell proliferation and tissue repair

Activation • Classical (Th1) • Alternative (Th2)

Stimuli elaborated by or occurring in the microenvironment that lead to activation
⦁ Interferon-γ (IFN-γ ) [produced by produced by Th1 lymphocytes or natural killer cells]
⦁ LPS [typical for gram-negative bacteria]
⦁ Lipoteichoic acid (LTA) [typical for gram positive bacteria]
⦁ Granulocyte/macrophage colony stimulating factor (GM-CSF)
⦁ Tumor-necrosis factor (TNF) [produced by antigen presenting cells]
⦁ C-C motif chemokine ligand 2 (CCL2)
⦁ Colony-stimulating factor-1 (CSF-1)
⦁ Interleukin-4 (IL-4)
⦁ Interleukin-10 (IL-10)
⦁ Interleukin-13 (IL-13)
⦁ Vascular endothelial growth factor (VEGF)
⦁ Transforming growth factor-β (TGF- β)

Produce pro-inflammatory cytokines • Yes, at high levels • Only low levels

Signaling molecules/transcription factors expressed
⦁ STAT1 alpha/beta
⦁ Interferon-Regulatory Factor (IRF5)
⦁ Btk
⦁ P2Y(2)R
⦁ SOCS3
⦁ Activin A
⦁ HIF1-α
⦁ Nuclear Factor of kappa light polypeptide gene enhancer (NF-KB)
⦁ Activator-Protein (AP-1)
⦁ STAT6
⦁ IRF4
⦁ KLF-4
⦁ NF-κB p50 homodimers
⦁ PPARγ
⦁ HIF-2α
⦁ IL-21
⦁ BMP-7
⦁ FABP4
⦁ LXRα

Initiate immune response/Antigen presentation
⦁ Yes • No – they do not present antigens to T-cells

Phagocytize microbes • Yes • No

Arginine metabolism • Turn on inducible NO synthase (iNOS), the enzyme that produces toxic nitric oxide (NO) and citrulline
⦁ NO and many of its downstream metabolites are very toxic making M1 a killing machine
⦁ iNOS expression is low in human blood
monocyte-derived macrophages, but expressed at high levels in inflamed human tissue macrophages and infiltrating monocyte-derived macrophages.
⦁ Note: NO production occurs primarily in mouse /
rodent macrophages. Human macrophages stimulated the same way do not typically secrete appreciable NO nor kill near as well as mouse macrophages

Function• Phagocytize and kill microbes – produce nitric oxide or reactive oxygen intermediates (ROIs) to protect against bacteria and viruses
⦁Bacterial infections induce polarization of macrophages toward the M1 phenotype, resulting in phagocytosis and intracellular killing of bacteria in vitro and in vivo
⦁Uncontrolled M1 macrophage-mediated inflammatory response can disrupt normal tissue homeostasis, impede vascular repair and lead to a cytokine storm

RRx-001 is a dinitroazetidine-based [33,34] small molecule.The chemical structure of RRx-001 is shown in Fig. 4A. It has multi- ple mechanisms of action (Fig. 4B) that include downregulation of CD47 through the inhibition of the transcription factor MYC, that upregulates CD47 expression [35], as well as the induction of the transition of M2-TAMs to M1-like macrophages (Fig. 4). As such,
RRx-001 does not fit the typical profile of a therapeutic that “bio- logically” mediates SIRPα-CD47 antagonism. In the Phase 1 first in man dose escalation trial, no DLTs were observed, and the maxi- mum tolerated dose (MTD) was not reached with doses up to 166 mg (83 mg/m2) [36]. As a small molecule “Trojan Horse”, RRx- 001 enters and oxidatively alters the milieu of the red blood cell [37], and is selectively taken up in tumors. TAMs phagocytose and catabolize the oxidized RRx-001-bound RBCs, leading to the re- lease of RRx-001-metabolites in the tumor [38,39] microenviron- ment with subsequent inhibition of MYC [40], a negative regulator of CD47 [35].

Clinical biopsies have demonstrated that high macrophage den- sity appears to correlate with RRx-001-mediated clinical benefit, and conversely, that low macrophage density is associated with poorer progression free survival and overall survival, which is an- tithetical to clinical orthodoxy in the sense that TAMs are known to mediate cancer chemotherapy and radiotherapy resistance [24]. Clinical biopsies during RRx-001 treatment also suggest that the
overexpression of CD47 and SIRPα in tumors is predictive of clin-
ical benefit from treatment with RRx-001, an observation that is similarly surprising since positive CD47 status tends to be associ- ated with poor prognosis [41].

On the basis of its favorable safety profile with the notable absence of hematologic toxicities, promising Phase 2 small cell lung cancer (SCLC) data [42] and known overexpression of CD47 in SCLC [43], a Phase 3 trial called REPLATINUM [44] in third line or beyond SCLC was initiated in Q4 2019. Interestingly, in the Phase 2 trial in SCLC, decreased expression of PD-L1 on circulat- ing tumor cells correlated with clinical benefit [45]. Positive outcomes were also observed in a Phase 1 trial, PRIMETIME, with RRx- 001 + nivolumab [46].

Hu5F9-G4
Hu5F9-G4 (5F9) is an intravenously administered humanized IgG4 antibody that targets CD47 with high affinity [47]. A 1 mg/kg priming dose [48], which is used to clear damaged senescent RBCs acutely, results in enhanced reticulocytosis and increased lifes- pan of the new red cells, and is followed by weekly maintenance doses of 30 mg/kg. In the Phase 1 first-in-man trial, fatigue, chills, pyrexia, anemia, hemagglutination (adhesive clusters of cells which can lead to impaired circulation, especially through the spleen, and anemia), headache, lymphopenia, transient hyperbilirubinemia, and myalgias were the most common adverse events. No MTD was reached up to 45 mg/kg and no Grade 3 anemias occurred [47]. Relatively low response rates were demonstrated – 10% in patients with relapsed/refractory AML/MDS and 5% in patients with solid tumors. These low response rates to treatment with a
single agent

Macrophage CD47/SIRP α interaction. The interaction of CD47 on tumor cells with SIRPα on tissue macrophages leads to the polarization of tissue macrophages to the “pro-tumor” don’t eat M2 phenotype. The latter is mediated in part by the activation of SH2-containing tyrosine phosphatase – 1/2 (SHP-1/2) which has an antiphagocytosis effect that inhibits a macrophage’s tumoricidal tendencies. As with the interaction between the PD-1 receptor with its PD-L1 ligand, the SIRPα / C47 interaction provides two targets for therapeutic intervention, the SIRPα receptor and its CD47 ligand. the lack of dose limiting toxicities have been attributed to the reduced affinity of IgG4 for Fcγ receptors on macrophages, which limits ADCP [12] and leads to reduced activity and off target tox- icities such as cytopenias. The low rate of antitumor activity as a single agent provided the rationale for combinatorial dosing to in- crease phagocytic activity [49].

A Phase 1b/2 trial in patients with relapsed/refractory non- Hodgkin’s lymphoma (NHL), the majority of whom had failed to have continued benefit from a prior rituximab-containing regimen- treated 75 patients with a combination of 5F9 plus rituximab. The objective response rate was reported as 49% with a complete re- sponse rate of 21% [50]. A MTD was not reached up to 30 mg/kg [51].
In a Phase 1b study, a combination with azacitidine, a DNA demethylating agent that can upregulate the “eat me” signal, cal- reticulin, on tumor cells [32], the objective response rate was 100% in eleven patients with untreated MDS with a 55% complete re- sponse rate, and 64% in 14 patients with AML who had not been previously treated, although a head-to-head comparison with azac- itidine was not performed [52].
Other trials in progress include combinations with the antiepi- dermal growth factor receptor monoclonal antibody cetuximab, and the PD-L1 targeting antibody avelumab.

TTI-621
TTI-621 is a dual functioning soluble decoy receptor with the CD47-binding domain of SIRPα fused to an IgG1 Fc region that has been dosed intravenously for hematologic malignancies and intra- tumorally for solid tumors and mycosis fungoides [53]. Preclini- cally, activity was reduced when the IgG1 Fc tail of the fusion pro- tein was replaced with an IgG4 tail or with an inert mutated IgG4 tail, which is supportive of the role of the IgG1 Fc as a more potent inducer of macrophage-mediated effects or functions [12].
Unlike 5F9, TT-621 is reported to bind minimally to red blood cells, which reduces the potential for anemia, although 2 patients may have experienced dose-limiting transaminitis and thrombocy- topenia [54]. It is unclear whether the thrombocytopenia was, in fact, dose-limiting since limited information on the Phase 1 trial is available, and thrombocytopenia was subsequently excluded as a DLT on the basis that it was asymptomatic and transient [32].

The manufacturer, Trillium Therapeutics, reports that in cuta- neous T-cell lymphoma, Sézary Syndrome, peripheral T-cell lym- phoma and DLBCL, 0.2 mg/kg TT-621 produced objective response rates of 17 to 25% and these responses included CRs [32]. Promis- ing activity with intratumoral injection in cutaneous T-cell lym- phoma has also been reported. A second candidate, linked to an IgG4 domain instead of IgG1, TTI-622, is in a Phase 1a/1b clinical trial started mid-2018 [55].

ALX-148
ALX148 is a decoy receptor comprised of a SIRPα domain mu- tated for high affinity CD47 binding and an inactive Fc region for the mitigation of hemagglutination and anemia. In the presence of an inert Fc region, the decoy is rendered inactive as monotherapy, which necessitates development in combination with other thera- pies. Results to date indicate ALX148 is generally well tolerated with fatigue, increased AST and ALT, anemia, and decreased platelets as the most common adverse events [56]. While no complete or par- tial responses were observed with monotherapy, combination with trastuzumab resulted in a 22% partial response rate in patients with HER2-positive gastric cancers and a 16% partial response rate in combination with pembrolizumab in patients with head and neck squamous cell carcinomas. No responses were reported in combination with pembrolizumab in eighteen patients with a di- agnosis of non-small cell lung cancer [57].
Conclusion

Immune evasion, which is a hallmark of cancer [58], oc- curs through multiple mechanisms including activation of immune checkpoint pathways. CTLA4 and PD-1/L1 checkpoint inhibitors were developed — and ultimately approved [59] in multiple tumor types — to overcome T cell anergy and dysfunction [60]. The de- velopment of activators, however, has lagged significantly behind anti-PD-1/PD-L1 immunotherapy, the reverse scenario from what occurs in vivo where the innate immune system not only precedes the initiation of adaptive immunity, but also helps to drive and co- ordinate anticancer responses through priming and recruitment of tumor-specific T cells [61]. Moreover, based on their in situ avail- ability/accessibility, and abundance in tumor stroma [62], TAMs are an attractive target for therapeutic intervention.
An innate counterpart to checkpoint inhibitors targeting PD1 and its ligand, PD-L1 are antagonists of SIRPα/CD47, a druggable
innate macrophage immune checkpoint pathway, with several an- tagonists of this pathway in Phase 1–3 clinical trials. SIRPα- CD47 functions as a “don’t-eat-me” signal. Cancer cells overex- pressing CD47, exploit the CD47/SIRPα axis to inhibit macrophage cytoskeletal activity [63] and, thus, evade macrophage-mediated phagocytosis.

Despite the promise of SIRPα/CD47 antagonism seen in preclin-
ical studies and in early clinical trials with Hu5F9-G4 (5F9), TTI- 621, ALX148, and RRx-001, one of the main concerns with these agents is the potential for severe hematologic toxicities, due to the ubiquitous expression of CD47, which may manifest as acute ane- mia, neutropenia and thrombocytopenia. Of note, RRx-001 is not associated with cytopenias, and Hu5F9-G4-induced anemia is re- portedly circumventable with the use of a low priming dose of 1 mg/kg on day 1, followed by a dose of 30 mg/kg weekly for
3 doses, and then a maintenance dose of 30 mg/kg every other week. There is also concern about possible reduced efficacy from the massive erythrocytic “CD47 sink” that may decrease the tumor bioavailability of anti-CD47 therapies, a limitation not expected
with anti-SIRPα blocking agents since they bind to macrophages,
not RBCs.

Other limitations pertaining specifically to SIRPα/CD47 monoclonal antibodies is their large molecular weight, which re- sults in a heterogeneous distribution in tumors with limited pene- tration in poorly vascularized or poorly perfused areas of tumor.
The overall success of immunotherapy, which presently only benefits a minority of patients, depends on the minimization of
de novo and acquired resistance [64]. A possible resistance mech- anism to SIRPα/CD47 antagonism is upregulation of CD47 and SIRPα [65]. Current strategies to overcome/counteract the upreg- ulation of CD47 and SIRPα include the use of chemotherapy and radiotherapy, as well as monoclonal antibodies including ritux- imab, cetuximab, trastuzumab, alemtuzumab, ramucirumab, dara- tumumab, obinutuzumab, and ofatumumab. These strategies en- hance the pro-phagocytic effect of CD47/SIRPα antagonism since chemotherapy and radiotherapy have been reported to increase cell surface exposure of calreticulin, an “eat me” stimulus, [66] and the Fc region of monoclonal antibodies binds to FcRs on macrophages and monocytes, thereby triggering an ADCP [67].

Other strategies to overcome acquired resistance include stimu- lation of both innate and adaptive immunity [68] to modulate the TME and enhance the induction of antitumor responses. Analogous to dual anti-CTLA4 and PD-1/PD-L1 blockade [69], combinations of antagonists against CD47 and SIRPα may demonstrate complemen- tarity, thus yielding better results than monotherapy, provided tox-icities are manageable. Combination therapy with inhibition of the PD-1/PD-L1 axis may also be useful, as suggested by a recent re-port of dual blockade of CD47 and PD-L1 enhancing the efficacy of immunotherapy against circulating tumor cells [70].

One approach to mitigate toxicity overall is to temporally se- quence these agents. For example, on the premise that innate im- munity precedes the development of cell-mediated immunity by 4-7 days, and in accordance with the physiologic immune tem- plate, a CD47 or SIRPα antagonist might be given for 1 week to prime phagocytosis before its discontinuation and the start of an anti-PD-1/PD-L1 therapy, with the possibility of alternating differ- ent CD47 or SIRPα antagonists in the weeks when the PD-1/PD-L1 checkpoint inhibitors are not administered. Further study of these agents and optimization of combination therapy will be an active area of investigation in the future.

In conclusion, antagonism of the CD47/SIRPα axis is a promising and rapidly emerging anticancer strategy that tries to lever- age the ability to counteract or block the antiphagocytic signal de- livered by CD47 through SIRPα [71]. Combination with the other
established therapies such as surgery, radiotherapy, chemother- apy, targeted therapy and anti-PD-L1/PD-1 checkpoint inhibitors has the potential to augment prophagocytic signaling through in- creased tumor cell death with FcR-dependent antigen acquisition and upregulation of calreticulin expression in the hope of substan- tially enhancing overall antitumor responses.

Declaration of Competing Interest

B. Oronsky, C. Carter, S. F. Brinkhaus, and T. Reid are employees of and have leadership roles at EpicentRx Inc. T. Reid is also an em- ployee of the University of California, San Diego. The authors have no other relevant affiliations or financial involvement with any or- ganization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Acknowledgments

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References

[1] Kunitoh H. Promises and challenges of immuno-oncology from a clinical per- spective. J Clin Oncol 2018;48(2):101–2.
[2] Dhariwal R, Pindoria N, Dasgupta P, Khan MS. #Checkmate: could checkpoint inhibitors be the game changer in the fight against metastatic urothelial carci- noma? BJU Int 2019;123(2):203–7.
[3] Couzin-Frankel J. Breakthrough of the year 2013. Cancer immunotherapy. Sci- ence 2013;342(6165):1432–3.
[4] Nowicki TS, Hu-Lieskovan S, Ribas A. Mechanisms of Resistance to PD-1 and PD-L1 blockade. Cancer J 2018;24(1):47–53.
[5] Mitsuka K, Kawataki T, Satoh E, Asahara T, Horikoshi T, Kinouchi H. Expression of indoleamine 2,3-dioxygenase and correlation with pathological malignancy in gliomas. Neurosurgery 2013;72(6):1031–8 discussion 1038-1039.
[6] Riley RS, June CH, Langer R, Mitchell MJ. Delivery technologies for cancer im- munotherapy. Nat Rev Drug Discov 2019;18(3):175–96.
[7] Lee S, Margolin K. Cytokines in Cancer Immunotherapy. Cancers 2011;3(4):3856–93.
[8] Champiat S, Dercle L, Ammari S, et al. Hyperprogressive disease is a new pat- tern of progression in cancer patients treated by anti-PD-1/PD-L1. Clin Cancer Res 2017;23(8):1920–8.
[9] Haslam A, Prasad V. Estimation of the percentage of US patients with can- cer who are eligible for and respond to checkpoint inhibitor immunotherapy drugs. JAMA 2019;2(5):e192535 03.
[10] Lian S, Xie R, Ye Y, et al. Simultaneous blocking of CD47 and PD-L1 increases innate and adaptive cancer immune responses and cytokine release. EBio Med 2019 Apr;42:281–95.
[11] Friedman CF, Navid-Azarbaijani P, Shoushtari AN, et al. Toxicity associated with ipilimumab and nivolumab (Ipi+Nivo) combination therapy in melanoma pa- tients (pts) treated at a single-institution under an expanded-access program (EAP). J Clin Oncol 2016;34(15_suppl):9519.
[12] Petrova PS, Viller NN, Wong M, et al. TTI-621 (SIRPαFc): a CD47-blocking in-
nate immune checkpoint inhibitor with broad antitumor activity and minimal erythrocyte binding. Clin Cancer Res 2017;23(4):1068–79.

[13] Sick E, Jeanne A, Schneider C, Dedieu S, Takeda K, Martiny L. CD47 update: a multifaceted actor in the tumour microenvironment of potential therapeutic interest. Br J Pharmacol 2012;167(7):1415–30.
[14] Liu Q, Wen W, Tang L, Qin C, et al. Inhibition of SIRPα in dendritic cells po-
tentiates potent antitumor immunity. Oncoimmunology 2016;5(9):e1183850.
[15] Kim MJ, Lee J-C, Lee J-J, et al. Association of CD47 with natural killer cell-me- diated cytotoxicity of head-and-neck squamous cell carcinoma lines. Tumour Biol 2008;29(1):28–34.
[16] Oldenborg P-A, Gresham HD, Chen Y, Izui S, Lindberg FP. Lethal autoimmune hemolytic anemia in CD47-deficient nonobese diabetic (NOD) mice. Blood 2002;99(10):3500–4.
[17] Olsson M, Bruhns P, Frazier WA, Ravetch JV, Oldenborg P-A. Platelet homeosta- sis is regulated by platelet expression of CD47 under normal conditions and in passive immune thrombocytopenia. Blood 2005;105(9):3577–82.
[18] Fousek K, Ahmed N. The evolution of T-cell therapies for solid malignancies. Clin Cancer Res 2015;21(15):3384–92.
[19] Chaturvedi P, Gilkes DM, Takano N, Semenza GL. Hypoxia-inducible factor-de- pendent signaling between triple-negative breast cancer cells and mesenchy- mal stem cells promotes macrophage recruitment. Proc Natl Acad Sci U S A 2014;111(20):E2120–9.
[20] Jeannin P, Paolini L, Adam C, Delneste Y. The roles of CSFs on the functional polarization of tumor-associated macrophages. FEBS J 2018;285(4):680–99.
[21] Najafi M, Hashemi Goradel N, Farhood B, et al. Macrophage polarity in cancer: a review. J Cell Biochem 2019;120(3):2756–65.
[22] Ley K. M1 Means Kill; M2 Means Heal. J Immunol 2017;199(7):2191–3 01.
[23] Beatty GL, Chiorean EG, Fishman MP, et al. CD40 agonists alter tumor stroma and show efficacy against pancreatic carcinoma in mice and humans. Science 2011;331(6024):1612–16.
[24] Chen Y, Song Y, Du W, Gong L, Chang H, Zou Z. Tumor-associated macrophages: an accomplice in solid tumor progression. J Biomed Sci 2019;26(1):78.
[25] Gardai SJ, McPhillips KA, Frasch SC, et al. Cell-surface calreticulin initiates clearance of viable or apoptotic cells through trans-activation of LRP on the phagocyte. Cell 2005;123(2):321–34.
[26] Grugan KD, McCabe FL, Kinder M, Greenplate AR, Harman BC, Ekert JE, et al. Tumor-associated macrophages promote invasion while retaining Fc-de- pendent anti-tumor function. J Immunol 2012;189(11):5457–66.
[27] Chen X, Song X, Li K, Zhang T. Fcγ R-binding is an important functional at-
tribute for immune checkpoint antibodies in cancer immunotherapy. Front Im- munol 2019;10:292.
[28] Goulet DR, Atkins WM. Considerations for the design of antibody-based thera- peutics. J Pharm Sci 2020;109(1):74–103.
[29] Pietsch EC, Dong J, Cardoso R, et al. Anti-leukemic activity and tolerability of anti-human CD47 monoclonal antibodies. Blood Cancer J 2017;7(2):e536.
[30] Barclay AN, Van den Berg TK. The interaction between signal regulatory pro- tein alpha (SIRPα) and CD47: structure, function, and therapeutic target. Annu
Rev Immunol 2014;32:25–50.
[31] Chao MP, Jaiswal S, Weissman-Tsukamoto R, et al. Calreticulin is the dominant pro-phagocytic signal on multiple human cancers and is counterbalanced by CD47. Sci Transl Med 2010;2(63):63ra94.
[32] Uger R, Johnson L. Blockade of the CD47-SIRPα axis: a promising approach for
cancer immunotherapy. Expert Opin Biol Ther 2020;20(1):5–8.
[33] Ning S, Bednarski M, Oronsky B, Scicinski J, Saul G, Knox SJ. Dinitroazetidines are a novel class of anticancer agents and hypoxia-activated radiation sensitiz- ers developed from highly energetic materials. Cancer Res 2012;72(10):2600–8.
[34] Oronsky B, Scicinski J, Ning S, et al. Rockets, radiosensitizers, and RRx-001: an origin story part I. Discov Med 2016;21(115):173–80.
[35] Casey SC, Tong L, Li Y, et al. MYC regulates the antitumor immune response through CD47 and PD-L1. Science 2016;352(6282):227–31.
[36] Reid T, Oronsky B, Scicinski J, et al. Safety and activity of RRx-001 in patients with advanced cancer: a first-in-human, open-label, dose-escalation phase 1 study. Lancet Oncol 2015;16(9):1133–42.
[37] Cabrales P, Scicinski J, Reid T, et al. A look inside the mechanistic black box: Are red blood cells the critical effectors of RRx-001 cytotoxicity? Med Oncol 2016;33(7):63.
[38] Oronsky B, Paulmurugan R, Foygel K, et al. RRx-001: a systemically non-toxic M2-to-M1 macrophage stimulating and prosensitizing agent in Phase II clinical trials. Expert Opin Investig Drugs 2017;26(1):109–19.
[39] Cabrales P. RRx-001 acts as a dual small molecule checkpoint inhibitor by
downregulating CD47 on cancer cells and SIRP-α on monocytes/macrophages. Transl Oncol 2019;12(4):626–32.
[40] Oronsky B, Reid TR, Oronsky A, Caroen S, Carter CA, Cabrales P. Brief re- port: RRx-001 is a c-Myc inhibitor that targets cancer stem cells. Oncotarget 2018;9(34):23439–42.
[41] Majeti R, Chao MP, Alizadeh AA, et al. CD47 is an adverse prognostic factor and therapeutic antibody target on human acute myeloid leukemia stem cells. Cell 2009;138(2):286–99.
[42] Morgensztern D, Rose M, Waqar SN, et al. RRx-001 followed by platinum plus etoposide in patients with previously treated small-cell lung cancer. Br J Can- cer 2019;121(3):211–17.
[43] Weiskopf K, Jahchan NS, Schnorr PJ, et al. CD47-blocking immunotherapies stimulate macrophage-mediated destruction of small-cell lung cancer. J Clin Invest 2016;126(7):2610–20 01.
[44] Oronsky B, Reid TR, Larson C, et al. REPLATINUM Phase III randomized study: RRx-001 + platinum doublet versus platinum doublet in third-line small cell lung cancer. Future Oncol 2019;15(30):3427–33.

[45] Tomita Y, Oronsky B, Carter CA, et al. In small cell lung cancer patients treated with RRx-001, a downregulator of CD47, decreased expression of PD-L1 on cir- culating tumor cells significantly correlates with clinical benefit. Int J Cancer 2020 In Press.
[46] Carter C, Oronsky B, Quinn M, Trepel JB, Abrouk ND, Skinner J. Phase 1 pilot study of RRx-001 + nivolumab in advanced metastatic cancer. PRIMETIME, Proc SITC 2019 Abstract p353.
[47] Sikic BI, Lakhani N, Patnaik A, et al. First-in-human, first-in-class phase I trial of the anti-CD47 antibody Hu5F9-G4 in patients with advanced cancers. J Clin Oncol 2019;37(12):946–53.
[48] Advani R, Flinn I, Popplewell L, et al. CD47 Blockade by Hu5F9-G4 and Ritux- imab in Non-Hodgkin’s Lymphoma. N Engl J Med 2018;379(18):1711–21.
[49] Liu J, Wang L, Zhao F, et al. Pre-clinical development of a human- ized Anti-CD47 antibody with anti-cancer therapeutic potential. PloS One 2015;10(9):e0137345.
[50] Advani R, Bartlett NL, Smith SM. The first-in-class anti-Cd47 antibody Hu5f9-G4 + Rituximab induces durable responses in relapsed/refractory Dl- bcl and indolent lymphoma: interim phase 1b/2 results. Hematol Oncol 2019;37(S2):89–90.
[51] https://www.fortyseveninc.com/application/files/6215/4387/4523/ASCO_2018_ Forty_Seven_Presentation_5F9003_NHL_trial_FINAL.pdf. 2018; Accessed Febru- ary 18, 2020.
[52] Sallman DA, Donnellan WB, Asch AS, et al. The first-in-class anti-CD47 antibody Hu5F9-G4 is active and well tolerated alone or with azaciti- dine in AML and MDS patients: Initial phase 1b results. J Clin Oncol 2019;37(15_suppl):7009.
[53] Velliquette RW, Aeschlimann J, Kirkegaard J, Shakarian G, Lomas-Francis C, Westhoff CM. Monoclonal anti-CD47 interference in red cell and platelet test- ing. Transfusion (Paris) 2019;59(2):730–7.
[54] Trillium Therapeutics Provides Update on its TTI-621 and TTI-622 Clinical Programs. Available from: https://ir.trilliumtherapeutics.com/press-releases/ press-release-details/2018/Trillium-Therapeutics-Provides-Update-on-its–TTI- 621- and- TTI- 622- Clinical-Programs/default.aspx. Accessed February 18, 2020.
[55] A. Trial of TTI-622 in Patients With Advanced Relapsed or Refractory Lym- phoma or Myeloma – Full Text View – ClinicalTrials.gov. Available from: https:
//clinicaltrials.gov/ct2/show/NCT03530683. Accessed February 18, 2020.
[56] Kauder SE, Kuo TC, Harrabi O, et al. ALX148 blocks CD47 and enhances in- nate and adaptive antitumor immunity with a favorable safety profile. PloS One 2018;13(8):e0201832.
[57] Chow LQM, Gainor JF, Lakhani NJ, et al. A phase I study of ALX148, a CD47 blocker, in combination with established anticancer antibodies in patients with advanced malignancy. J Clin Oncol 2019;37(15_suppl):2514.
[58] Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 2011;144(5):646–74.
[59] Topalian SL, Hodi FS, Brahmer JR, Gettinger SN, Smith DC, McDermott DF, et al. Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl J Med 2012;366(26):2443–54.
[60] Hodi FS, O’Day SJ, McDermott DF, Weber RW, Sosman JA, Haanen JB, et al. Im- proved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med 2010;363(8):711–23.
[61] Liu Y, Zeng G. Cancer and innate immune system interactions: translational potentials for cancer immunotherapy. J Immunother 2012;35(4):299–308.
[62] Ireland LV, Mielgo A. Macrophages and fibroblasts, key players in cancer chemoresistance. Front Cell Dev Biol 2018;6:131.
[63] Sosale N, Rouhiparkouhi T, Bradshaw A, Dimova R, Lipowsky R, Discher D. Cell rigidity and shape override CD47’s “self”-signaling in phagocytosis by hyperac- tivating myosin-II. Blood 2015;125(3):542–52.
[64] Joshi S, Durden DL. Combinatorial approach to improve cancer immunother- apy: rational drug design strategy to simultaneously hit multiple targets to kill tumor cells and to activate the immune system. J Oncol 2019;2019:5245034.
[65] Chao MP, Weissman IL, Majeti R. The CD47-SIRPα pathway in cancer im-
mune evasion and potential therapeutic implications. Curr Opin Immunol 2012;24(2):225–32.
[66] Obeid M, Panaretakis T, Tesniere A, et al. Leveraging the immune system dur- ing chemotherapy: moving calreticulin to the cell surface converts apoptotic death from “silent” to immunogenic. Cancer Res 2007;67(17):7941–4.
[67] Hogarth PM, Pietersz GA. Fc receptor-targeted therapies for the treatment of RRx-001 inflammation, cancer and beyond. Nat Rev Drug Discov 2012;11(4):311–31.
[68] Criscitiello MF, de Figueiredo P. Fifty shades of immune defense. PLoS Pathog 2013;9(2):e1003110 Epub 2013 Feb7. doi:10.1371/journal.ppat.1003110.
[69] Chae YK, Arya A, Iams W, et al. Current landscape and future of dual an- ti-CTLA4 and PD-1/PD-L1 blockade immunotherapy in cancer; lessons learned from clinical trials with melanoma and non-small cell lung cancer (NSCLC). J Immunother Cancer 2018;6(1):39.
[70] Lian S, Xie R, Ye Y, et al. Dual blockage of both PD-L1 and CD47 enhances immunotherapy against circulating tumor cells. Sci Rep 2019;9(1):4532.
[71] Tsai RK, Discher DE. Inhibition of “self” engulfment through deactivation of myosin-II at the phagocytic synapse between human cells. J Cell Biol 2008;180(5):989–1003.