1. Introduction
Macrophages are often the most abundant immune cell found in the tumor microenvironment, but are frequently associated with poor patient prognosis and considered detrimental to mounting an effective anti-tumor response [Citation1,Citation2]. There is, however, a growing recognition that macrophages can have potent anti-tumor effects but this beneficial activity is frequently suppressed by the CD47-SIRPα axis.
CD47 is a ubiquitously expressed transmembrane protein that interacts with several ligands, including signal regulatory protein alpha (SIRPα) on myeloid cells. Engagement of SIRPα by CD47 triggers an anti-phagocytic ‘do not eat’ signal that prevents macrophages from engulfing and destroying target cells [Citation3]. The significance of this anti-phagocytic signal in cancer was first discovered in acute myeloid leukemia (AML), where high CD47 expression was associated with poor clinical outcome and blockade of CD47 exhibited efficacy in AML xenografts. Subsequently, many tumors, both hematologic and solid, were shown to overexpress CD47 and high expression of CD47 is associated with poor prognosis [Citation4,Citation5]. Thus, the CD47-SIRPα axis represents a widespread mechanism of immune evasion.
Preclinical studies have demonstrated that disruption of the CD47-SIRPα axis is efficacious across numerous animal models. Interestingly, these preclinical findings showed that CD47 blockade alone is insufficient to enable tumor cell phagocytosis [Citation6,Citation7]. Macrophages must also receive a pro-phagocytic (‘eat’) signal. Although tumor cells are known to express ‘eat’ signals such as calreticulin, in general, these are too weak to enable robust phagocytosis with CD47 blockade alone. The required activating signal can be provided through the Fc region of the blocking agent, with IgG1 providing maximal potency. Alternatively, the ‘eat’ signal can be delivered by a second agent such as an IgG1-bearing anti-cancer antibody (e.g. rituximab) which can opsonize tumor cells and provide an Fc signal to macrophages. Macrophages are also functionally diverse and can have both pro- and anti-tumor properties depending on the priming conditions. Lin and colleagues demonstrated that in vitro CD47 blockade improved phagocytosis across macrophage subtypes, intriguingly showing that the M2c subtype was particularly sensitive to CD47 blockade, and suggesting that CD47 blockade may benefit all tumor-associated macrophages [Citation8].
Although the primary mechanism of action for CD47 blockade is likely enhanced macrophage-mediated phagocytosis, pre-clinical studies has suggested a role for additional immune cells. Tumor-associated dendritic cells (DC) have increased SIRPα expression with silencing leading to increased IL-12 production [Citation9]. Given the potential for improved DC function, T cells may also indirectly benefit, through cross-priming of CD8 + T cells. Dendritic cells, but not macrophages, were required to trigger an adaptive response and the anti-tumor effect was abrogated in the absence of T cells in a syngeneic murine model with anti-CD47 [Citation10]. Decreased NK cell cytotoxicity is associated with higher expression of CD47 in tumor cell lines [Citation11]. Thus, blockade of CD47 on NK cells may enhance their functionality within the tumor environment.
2. CD47-SIRPα blocking agents in clinical trials
The first agent targeting the CD47-SIRPα axis entered the clinic in 2014, and there are now more than 10 compounds in active clinical development (). The majority of these target CD47, and encompass both antibodies and SIRPα decoy receptors with a range of Fc effector function (potent IgG1, weaker IgG4 or IgG2, and inert, mutated IgG1). More recently, a CD47/CD19 bispecific and anti-SIRPα antibodies have entered human testing. Clinical data for three agents (Hu5F9-G4, TTI-621, and ALX148) have been reported and are summarized below.
Table 1. Clinical trials targeting the CD47-SIRPα axis.
Hu5F9-G4 (5F9) is a high affinity humanized IgG4 anti-CD47 antibody administered intravenously with a 1 mg/kg priming dose followed by weekly maintenance doses up to 45 mg/kg [Citation12]. The unique dosing regimen minimizes red blood cell toxicity by selectively clearing aging red blood cells (RBC), which results in a mild and transient anemia. Pooled efficacy results from Phase 1b+2 of 5F9 and rituximab (n = 75) indicate an objective response rate (ORR) of 49% and a complete response (CR) rate of 21% [Citation13]. These response rates are likely higher than those achievable with rituximab alone in this patient population, the majority of which have previously failed a prior rituximab-containing regimen, although data are lacking. Combining 5F9 with azacitidine, a demethylating agent that upregulates calreticulin expression on tumor cells, is also showing clinical promise, with ORR of 100% in untreated myelodysplastic syndromes (MDS) patients (n = 11, 55% CR rate), and 64% in untreated AML patients (n = 14) [Citation14]. The combination appears to compare favorably with azacitidine monotherapy, although a direct head-to-head comparison was not performed. As a monotherapy, 5F9 has shown relatively low response rates: 10% in relapsed/refractory AML/MDS patients and 5% in solid tumor patients receiving maintenance doses of 20 mg/kg or greater [Citation12,Citation14]. This supports the hypothesis that IgG4-based CD47-blocking agents require a combination partner to deliver the necessary pro-phagocytic signal due to weaker IgG potency.
TTI-621 is a fusion protein composed of the CD47-binding domain of SIRPα linked to an IgG1 Fc region. It has the unusual property of having minimal binding to human RBC, which has been attributed to its affinity for CD47 and the lack of mobility of CD47 within the erythrocyte membrane [Citation7]. A study of intravenous TTI-621 in patients with relapsed/refractory hematologic malignancies has shown that weekly infusions are well tolerated, with the principal safety finding of thrombocytopenia (grade 3 or higher in 18% of patients), likely an on-target effect. Notably, the thrombocytopenia is transient and does not result in an increased risk of bleeding or impact study drug delivery. Monotherapy activity has been observed in patients with cutaneous T-cell lymphoma (CTCL), Sézary Syndrome, peripheral T-cell lymphoma (PTCL) and DLBCL, with response rates ranging from 17-25%, and several CRs documented. Although this single-agent activity is modest, most of these patients received a relatively low dose of the drug (0.2 mg/kg) due to a conservative definition of dose-limiting toxicity that has since been reevaluated. Dosing beyond 0.2 mg/kg is currently in progress. Interestingly, intratumoral injection of TTI-621 has shown remarkable effects in patients with CTCL. Among 22 evaluable CTCL patients, 91% exhibited decreased CAILS scores (a measure of local tumor responses) and 41% achieved a ≥ 50% reduction in CAILS following local delivery, with anti-tumor responses occurring in non-injected lesions in some patients [Citation15]. Such promising single-agent activity with TTI-621 in CTCL has led to an increased focus in this area.
ALX148 is a decoy receptor comprised of a SIRPα domain that has been heavily mutated to confer high-affinity binding to CD47 fused to a mutated, inert IgG1 Fc. Unlike TTI-621, this fusion protein binds to human RBC, but the lack of Fc effector activity reduces the risk of anemia. The absence of an ‘eat’ signal renders the decoy receptor inactive as a monotherapy in preclinical studies and thus ALX148 is being developed as a combination therapy [Citation16]. Results to date indicate ALX148 is well tolerated with the most common treatment-related adverse events including fatigue, increased AST and ALT, anemia, and decreased platelets. While no monotherapy responses have been observed, partial responses were reported in 22% of HER2-positive gastric cancer patients in combination with trastuzumab and 16% of head and neck squamous cell carcinoma patients in combination with pembrolizumab. No responses have been observed in non-small cell lung cancer patients (n = 18) treated in combination with pembrolizumab [Citation17].
Interestingly, 5F9 and ALX148 are administered at substantially higher doses than TTI-621. The differences in dose between TTI-621 and Hu 5F9-G4 may be explained by the size of the peripheral sink for each therapeutic combined with relative effector activity of the Fc moiety. Hu5F9-G4 binds RBC, which accounts for ~40% of the total blood volume at 4.3–5.9 × 1012 cells per liter of blood, effectively marking them as senescent and requiring clearance. TTI-621 binds platelets, which only account for 1.5–4.5 × 1011 cells per liter [Citation7,Citation18,Citation19]. Therefore, circulating platelets represent a much smaller sink for TTI-621 to overcome. In the absence of platelet or RBC binding and without effector activity, ALX148’s dose-dependent toxicities may be associated with different on-target binding or Fc engagement.
3. Expert opinion
A growing awareness of the importance of innate immunity in immuno-oncology and solid foundation of preclinical data have fueled considerable excitement around targeting the CD47-SIRPα axis, with an increasing number of agents in clinical development and encouraging early results. Despite widespread expression of the target, blockade of CD47 is generally well tolerated, with most safety findings relating to mild hematologic toxicity (anemia, thrombocytopenia) that are likely on-target effects. Anti-tumor activity has been observed in patients receiving CD47 blockade alone as well as in combination with agents that provide additional ‘eat’ signals (rituximab, trastuzumab, azacitidine) and the T-cell checkpoint inhibitor pembrolizumab.
While these early clinical findings are highly encouraging, several outstanding questions remain. First, what is the ideal format for blocking CD47 in cancer patients? Both antibodies and decoy receptors with varying levels of Fc effector activity are currently in clinical testing. Preclinical studies indicate that IgG1 provides maximal potency, and indeed the IgG1-based SIRPαFc decoy receptor TTI-621 appears to have the highest level of monotherapy activity in patients. Although the ubiquitous expression of CD47 provides a challenge for the systemic delivery of a high potency agent, this may be overcome by local intratumoral delivery. Second, how does SIRPα blockade compare to targeting CD47? With a more restricted expression pattern, SIRPα may be preferable from a safety standpoint, but CD47 has the advantage of being upregulated on tumor cells. The recent entry of anti-SIRPα antibodies into the clinic should provide data on this point. Third, is CD47-SIRPα blockade better suited to blood versus solid tumors? The majority of anti-tumor activity observed in patients to date has been in hematologic malignancies and hematological toxicities (anemia, thrombocytopenia) appear to be the principal on-target safety findings. Additional clinical data in solid tumors should help to clarify whether blockade of the CD47-SIRPα axis is broadly applicable across both hematologic and solid cancers. Fourth, what is the mechanism of action of CD47 or SIRPα blockade in patients? There are little translational data from ongoing clinical studies in the public domain, and questions remain regarding the contribution of macrophages, DC, NK cells and, of course, anti-tumor T cells. Finally, does blockade of the CD47-SIRPα axis need to be paired with a T-cell checkpoint inhibitor? Neutralization of the suppressive CD47 signal alone is sufficient to trigger robust anti-tumor activity in animal models, but these systems do not replicate the degree of immune suppression observed in humans. Blockade of the CD47-SIRPα axis, an innate immune checkpoint, may ultimately work best when coupled to an adaptive immune checkpoint inhibitor, and several trials are underway to test this concept.
In summary, a multitude of biologic agents targeting the CD47-SIRPα axis are currently in clinical testing. These approaches seek to harness the anti-tumor activity of macrophages, a cell type found in abundance in many tumors that has until recently been largely overlooked in immuno-oncology. The data that have emerged are promising, and further studies over the next several years should answer key questions in the field and determine how this pathway fits within the evolving immune-oncology landscape.
Declaration of interest
R. Uger is an employee of and has a leadership role at Trillium Therapeutics Inc. L. Johnson is employed by Trillium Therapeutics Inc. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
Reviewer Disclosures
A reviewer on this manuscript has disclosed that they work on CD47 blockers in their lab and their institution has filed for protection for them. All other peer reviewers on this manuscript have no relevant financial or other relationships to disclose.
Additional information
Funding
References
- Poh AR, Ernst M. Targeting macrophages in cancer: from bench to bedside. Front Oncol. 2018;8:49.
- Pollard JW. Tumour-educated macrophages promote tumour progression and metastasis. Nat Rev Cancer. 2004;4(1):71–78.
- Veillette A, Chen J. SIRPalpha-CD47 immune checkpoint blockade in anticancer therapy. Trends Immunol. 2018;39(3):173–184.
- Willingham SB, Volkmer JP, Gentles AJ, et al. The CD47-signal regulatory protein alpha (SIRPa) interaction is a therapeutic target for human solid tumors. Proc Natl Acad Sci U S A. 2012;109(17):6662–6667.
- Chao MP, Alizadeh AA, Tang C, et al. Anti-CD47 antibody synergizes with rituximab to promote phagocytosis and eradicate non-hodgkin lymphoma. Cell. 2010;142(5):699–713.
- Weiskopf K, Ring AM, Ho CCM, et al. Engineered SIRPα variants as immunotherapeutic adjuvants to anticancer antibodies. Science. 2013;341(6141):88–91.
- Petrova PS, Viller NN, Wong M, et al. TTI-621 (SIRPalphaFc): a CD47-blocking innate immune checkpoint inhibitor with broad antitumor activity and minimal erythrocyte binding. Clin Cancer Res. 2017;23(4):1068–1079.
- Lin GHY, Chai V, Lee V, et al. TTI-621 (SIRPalphaFc), a CD47-blocking cancer immunotherapeutic, triggers phagocytosis of lymphoma cells by multiple polarized macrophage subsets. PLoS One. 2017;12(10):e0187262.
- Liu Q, Wen W, Tang L, et al. Inhibition of SIRPalpha in dendritic cells potentiates potent antitumor immunity. Oncoimmunology. 2016;5(9):e1183850.
- Liu X, Pu Y, Cron K, et al. CD47 blockade triggers T cell–mediated destruction of immunogenic tumors. Nat Med. 2015;21:1209.
- Kim MJ, Lee JC, Lee JJ, et al. Association of CD47 with natural killer cell-mediated cytotoxicity of head-and-neck squamous cell carcinoma lines. Tumour Biol. 2008;29(1):28–34.
- 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):946–953.
- Advani R, Bartlett NL, Smith SM, et al. The first-in-class anti-CD47 antibody Hu5F9-G4 + rituximab induces durable responses in relapsed/ refractory DLBCL and indolent lymphoma: interim phase 1B/2 results. Hematol Oncol. 2019;37(S2):89–90.
- Sallman DA, Donnellan WB, Asch A, et al. The first-in-class anti-CD47 antibody Hu5F9-G4 is active and well tolerated alone or with azacitidine in AML and MDS patients: initial phase 1b results.2019 ASCO Annual Meeting; Vol. 37. Chicago, IL. J Clin Oncol; 2019.
- Querfeld C, Thompson J, Taylor MH, et al. Intralesional administration of the CD47 antagonist TTI-621 (SIRPαFc) induces responses in both injected and non-injected lesions in patients with relapsed/refractory mycosis fungoides and sézary syndrome: interim results of a multicenter phase I trial. Blood. 2018;132(Suppl 1):1653.
- Kauder SE, Kuo TC, Harrabi O, et al. ALX148 blocks CD47 and enhances innate and adaptive antitumor immunity with a favorable safety profile. Plos One. 2018;13(8):e0201832.
- 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. ASCO Annual Meeting. Vol. 37. Chicago, IL: American Society of Clinical Oncology; 2019. p 2514.
- Velliquette RW, Aeschlimann J, Kirkegaard J, et al. Monoclonal anti-CD47 interference in red cell and platelet testing. Transfusion. 2019;59(2):730–737.
- Giles C. The platelet count and mean platelet volume. Br J Haematol. 1981;48(1):31–37.