Monoclonal antibodies targeting CD47/SIRPα
The mAbs of CD47 are divided into three main categories based on their form of action. The first class binds to CD47 on the surface of tumor cells, erythrocytes, and platelets, represented by Forty Seven’s Hu5F9-G4 (5F9, Magrolimab). Hu5F9-G is a humanized IgG4 antibody that targets CD47 with high affinity, but its binding to erythrocytes and platelets can cause anemic reactions [
146]. Currently, 5F9 is in clinical trials for patients with solid tumors (NCT02216409). Based on 100% erythrocyte CD47 saturation tolerance and receptor occupancy studies to avoid the hematotoxicity of 5F9, in a first phase I clinical trial of Hu5F9-G4 in patients with advanced cancer, 1 mg/kg was selected as the starting dose to trigger the clearance of aged erythrocytes in vivo and stimulate reticulocyte maturation and differentiation to produce fresh young erythrocytes. In the subsequent group, patients received maintenance doses ranging from 3 mg/kg to 45 mg/kg, with most toxicity being mild to moderate. The most common toxicity was an expected, targeted, mild, transient anemia that resolved when administered at the initial and maintenance doses in Cycle 1. In addition, a clinical trial is ongoing for 5F9 in combination with Cetuximab for solid tumors and advanced CRC (NCT02953782); another ongoing program is 5F9 in combination with Avelumab for the treatment of patients with solid tumors that have progressed within 6 months after prior platinum-based chemotherapy and patients with ovarian cancer that is primary resistant to immune checkpoint inhibitors (NCT03558139) [
147,
148].
The second category is antibodies obtained by specific screening that do not bind or bind weakly to erythrocytes. Compared with Class I antibodies, these antibodies have a theoretically higher safety profile and do not require modified clinical dosing regimens to reduce the associated hematological toxicity. This group is represented by antibodies such as AK117 (Ligufalimab) and TJC4 (Lemzoparlimab). AK117 is a novel IgG4 mAb targeting CD47 that uses the Ig4 isoform and is engineered to minimize Fc effects [
149]. Compared with other anti-CD47 mAbs, AK117 binds more weakly to human erythrocytes than to tumor cells [
150]. This gives it a potential prophagocytic activity and a good hematological safety profile. AK117 has demonstrated potential anti-tumor activity in both
in-vitro and
in-vivo nonclinical studies. The
in-vitro tests did not cause significant agglutination of human erythrocytes, while
in-vivo tests in a macaque exhibited minimal effects on anemia, with only minor erythrocyte changes [
150]. In a phase I study evaluating the safety of AK117 in advanced or metastatic solid tumors (NCT04349969), 15 patients were enrolled in phase 1a as of February 15, 2021, for DLT evaluation of the ongoing 30 mg /kg cohort. The results revealed that AK117 was safe and well tolerated up to 20 mg/kg QW; no IRRs or severe TRAEs were observed; and no hematological TRAEs occurred except in patients with baseline G1 anemia taking 10 mg/kg AK117 [
151]. Another phase I study demonstrated similar results in 27 solid tumors receiving CD47 mAbs, AO-176 [
152]. This study also reported that 7 patients had SD as a best response, with 2 patients (endometrial carcinoma, gastric cancer) on study for more than 6 months.
The third category is CD47mAbs that do not have binding capacity in blood and circulation, such as the ADG153 SAFEbody. When these antibodies enter the TME, they can be cleaved to expose binding epitopes and activate their binding activity under the cleavage of tumor-specific enzymes [
153]. However, preclinical and clinical trials of these antibodies have not been conducted in solid tumors.
In conclusion, the safety of mAbs, targeting CD47 in the treatment of solid tumors has been demonstrated. However, the efficacy of CD47 mAbs alone is not very prominent in solid tumors. Further studies are required to confirm this and the selected clinical trials of mAbs, targeting CD47/SIRPα are shown in Table
2.
Table 2
Clinical trials of CD47/SIRPα targeted agents in solid tumors
Hu5F9-G4 (Magrolimab) | Humanized, IgG4 | Solid Tumor | Monotherapy | 88 participants | Phase I | NCT02216409 |
AK117(Ligufalimab) | Humanized, IgG4 | Neoplasms Malignan | Monotherapy | 162 participants | Phase I | NCT04728334 |
| | Neoplasms Malignan | Monotherapy | 159 participants | Phase I | NCT04349969 |
| | Metastatic and locally advanced TNBC | Combination | 80 participants | Phase II | NCT05227664 |
| | Advanced malignant tumors | Combination | 130 participants | Phase I/II | NCT05235542 |
| | Advanced malignant tumors | Combination | 114 participants | Phase I/II | NCT05229497 |
| | Advanced malignant tumors | Combination | 160 participants | Phase I/II | NCT05214482 |
| | Metastatic colorectal cancer | Combination | 114 participants | Phase II | NCT05382442 |
AO-176 | Human, IgG4 | Solid tumor | Combination | 183 participants | Phase I/II | NCT03834948 |
SGN-CD47M | Humanized, IgG4 | Solid tumor | Monotherapy | 16 participants | Phase I | NCT03957096 |
CC-9002 | Humanized, IgG4 | Hematologic neoplasms | Monotherapy, or combination | 60 participants | Phase I | NCT02367196 |
Bispecific antibodies targeting CD47/SIRPɑ
BsAbs are synthetic antibodies with two specific antigen binding sites. The China Drug Clinical Trial and Information Disclosure Platform (
http://www.chinadrugtrials.org.cn/) presents more than 80 BsAbs, while a clinical trial website platform (
https://clinical.trials.gov/) showcases more than 200. Among them, BsAbs targeting the CD47/SIRPα axis are the main research focus [
154]. Anti-CD47/SIRPα BsAbs are classified according to the antigen that they target; thus, they can be classified as BsAbs targeting tumor antigens (PD-L1, CD20, CD19, mesothelin [MSLN], Claudin18.2, and Her2), BsAbs targeting immune cells (PD-1, CD40, and 41BB), BsAbs targeting cytokines or receptors (CSF-2 receptor/VEGF), and BsAbs targeting cytokines or receptors (CSF-2 receptor/VEGF) among others. Most BsAbs targeting the CD47/SIRPα axis reduce anemia by modulating the affinity of the two binding structural domains, as demonstrated by the relative reduction in affinity for targeting CD47.
Partly based on targeting the CD47/SIRPα axis, BsAbs are also currently in preclinical or clinical trials for the treatment of solid tumors (Tables
3,
4,
5). These include CD47/PD-1 (PD-L) BsAb, CD47/HER2 BsAb, CD47/GPC3 BsAb, CD47/EGFR BsAb, SIRPα/VEGFR1 BsAb, CD47/MSLN BsAb, SIRPα/CD40L BsAb, SIRPα/CTLA4 BsAb, CD47/DLL3 BsAb, and Claudin18.2/CD47 BsAb [
155‐
165].
Table 3
Preclinical studies of CD47 BsAb in solid tumors
CD47/GPC3 | HCC | |
CD47/EGFR | EGFR + tumors | |
SIRPα/VEGFR1 | NSCLC, Glioblastoma | |
CD47/MSLN | MSLN + tumors | |
SIRPα/CD40L | Solid tumors (mouse CT26 tumor model) | |
SIRPα/CTLA4 | Solid tumors | |
CD47/DLL3 | SCLC Neuroendocrine cancers | |
Claudin18.2/CD47 | gastric cancer Gastroesophageal junction (GEJ) cancer Pancreatic cancer | |
Table4
Clinical trials of CD47 BsAb in solid tumors
HX009 | IgG4 | CD47 + PD-1 | Advanced solid tumor | Monotherapy | 21 participants | Phase I | NCT04097769 |
| | | Advanced solid tumor | Monotherapy | 210 participants | Phase II | NCT04886271 |
IBI322 | | CD47 + PD-L1 | Advanced solid tumor | Monotherapy | 36 participants | Phase I | NCT04912466 |
| | | Advanced malignant tumors lymphomas | Monotherapy | 51 participants | Phase I | NCT04338659 |
| | | Advanced malignancies | Monotherapy | 218 participants | Phase I | NCT04328831 |
| | | Small cell lung cancer | Combination | 40 participants | Phase II | NCT05296603 |
| | | Non small cell lung cancer | Combination | 80 participants | Phase II | NCT05296278 |
| | | Myeloid tumor | Combination | 124 participants | Phase I | NCT05148442 |
6MW3211 | | CD47 + PD-L1 | Advanced malignant neoplasm | Monotherapy | 272 participants | Phase I/II | NCT05048160 |
PF-07257876 | IgG1 | CD47 + PD-L1 | NSCLC SCCHN and ovarian cancer | Monotherapy | 90 participants | Phase I | NCT04881045 |
IBC0966 | | CD47 + PD-L1 | Advanced malignant tumors | Monotherapy | 228 participants | Phase I/II | NCT04980690 |
IMM2902 | IgG1 | CD47 + HER2 | HER2 + advanced solid tumors | Monotherapy | 40 participants | Phase I | NCT05076591 |
Table 5
Clinical trials of SIRPα targeted agents in solid tumors
CC-95251 | Human | Advanced solid and hematologic cancers | Monotherapy, or Combination | 230 participants | Phase I | NCT03783403 |
BI 765,063 (OSE-172) | Humanized IgG4 | Solid tumors | Combination | 18 participants | Phase I | NCT04653142 |
| | Solid tumor, adult | Combination | 116 participants | Phase I | NCT03990233 |
| | HNSCC melanoma NSCLC | Combination | 22 participants | Phase I | NCT05068102 |
| | HNSCC | Combination | 150 participants | Phase I | NCT05249426 |
Evorpacept (ALX148) | mutated SIRPα-Fc IgG1 | HNSCC | Combination | 168 participants | Phase II | NCT04675333 |
| | HNSCC | Combination | 183 participants | Phase II | NCT04675294 |
| | HER2 + gastric cancer | Combination | 450 participants | Phase II/III | NCT05002127 |
| | Metastatic cancer; solid tumor; advanced cancer; NHL | Combination | 174 participants | Phase I | NCT03013218 |
| | Microsatellite stable metastatic colorectal cancer | Combination | 80 participants | Phase II | NCT05167409 |
| | HER2-expressing cancers | Combination | 93 participants | Phase I/II | NCT05027139 |
TTI-621 | SIRPα-Fc IgG1 | R/R solid tumors and mycosis fungoides | Monotherapy,or combination | 174 participants | Phase I | NCT02890368 |
| | Hematologic malignancies solid tumor | Monotherapy, or combination | 250 participants | Phase I | NCT02663518 |
| | Leiomyosarcoma | Combination | 80 participants | Phase I/II | NCT04996004 |
TTI-622 | SIRPα-Fc IgG4 | Platinum-resistant ovarian cancer | Combination | 50 participants | Phase I/II | NCT05261490 |
CD47/PD-1 BsAb: The interaction of anti-PD-1/PD-L1 with anti-CD47 in immunosuppression mediated by TAMs implies targeting both innate and adaptive dual immune checkpoints, suggesting that combined therapies improve survival more than monotherapy [
166]. Thus, CD47/PD-1(PD-L1) BsAb may maximize the effect of anti-tumor therapy and trigger longer-lasting therapeutic responses [
167].
The recombinant humanized antibody fusion protein HX009 is a BsAb that binds to both CD47 and PD-1, and Phase I (NCT04097769) and Phase II (NCT04886271) clinical trials of the drug are underway. The results of the phase I trial indicated that treatment-related adverse reactions occurred in 10 of the 21 patients treated (47.6%), and most AEs were grade 1 or 2. Furthermore, three patients had PR, and six patients had SD, which demonstrated satisfactory treatment results [
155].
CD47/PD-L1 BsAb: IBI322, a BsAb targeting CD47 and PD-L1, selectively binds to CD47 and PD-L1 co-expressing tumor cells on the one hand, thereby attenuating CD47 activity in monovalent binding and blocking PD-L1 activity in bivalent binding, stimulating strong anti-tumor activity. On the other hand, IBI322 effectively accumulates in PD-L1 positive tumors, blocks PD-1 and PD-L1 binding, activates CD8
+ T cells, stimulates adaptive immune responses, and exhibits synergistic activity in inducing complete tumor regression in vivo [
156,
168]. PD-L1 is expressed in tumor cells and IBI322 has a stronger affinity for PD-L1 than CD47, suggesting that IBI322 can bind more selectively to tumor cells than to normal cells. Preliminary studies suggest that IBI322 binds to PD-L1-positive tumor cells more than to red blood cells; therefore, it does not induce hemagglutination and has a stronger safety profile [
169]. In addition, IAB, another “Knobs-into-holes”–based dual-targeting fusion protein that targets CD47 and PD-L1, also demonstrated to have high safety and anti-tumor activity in tumor-bearing mice [
170].
CD47/HER2 BsAb: Previous studies have demonstrated that trastuzumab in combination with CD47 mAb completely eliminates tumors in a mouse model of human HER2
+ breast cancer transplantation tumors [
171,
172]. Based on this finding, the investigators (ImmuneOnco Shanghai Biomedical Co., Ltd.) prepared CD47/HER2 BsAb-IMM2902 using the “mAb Trap” technique. IMM2902 allows the drug to bind preferentially to tumor cells through the high-affinity activity of HER2, and it simultaneously exhibits the characteristics of not binding to human erythrocytes and avoiding “antigenic sink,” thus greatly enhancing the specific synergistic effect of the dual target on tumors. In nonhuman primates, different doses of IMM2902 had no effect on hemagglutination and no significant hematotoxicity [
157]. The established BT-474 breast cancer and NCI-N87 gastric xenograft tumor models exhibited complete tumor elimination in
in-vivo efficacy studies, even at doses as low as 3.5 mg/kg. Significantly, in a Herceptin-resistant breast tumor model, HCC-1954 tumors, IMM2902 also yielded strong anti-tumor activity at a dose of 10 mg/kg [
157]. IMM2902 has been approved to enter clinical trials for the primary indication of HER2-positive breast, gastric, lung, and other advanced solid tumors to assess its efficacy and safety in HER2
+ advanced solid tumors (NCT05076591).
CD47/GPC3 BsAb: GPC3 is an HCC-associated antigen specifically expressed in HCC, while the expression of CD47 in HCC inhibits its phagocytosis. As a result, a new BsAb, namely GPC3/CD47 biAB, was generated. In-vitro and in-vivo experiments revealed the safety profile of GPC3/CD47 biAb, its long half-life, and its more pronounced affinity for dual-antigen-expressing tumor cells, which highlight the advantages of its anti-tumor activity [
158]. In an hCD47/hSIRPɑ humanized mouse model, the serum half-life of GPC3/CD47 biAb was prolonged without hematological toxicity; in in-vitro experiments, GPC3/CD47 biAb enhanced the Fc-mediated effector function against dual antigen-expressing HCC cells. Moreover, in a xenogeneic HCC model, GPC3/CD47 biAb was superior to monotherapy as well as in combination with anti-CD47 and anti-GPC3 mAbs. These results suggest that GPC3/CD47 biAb will further improve cancer treatment with GPC3/CD47 dual antigen expression [
158].
CD47/EGFR BsAb: In addition to being expressed on tumor cells, CD47’s widespread expression on normal cells limits the clinical efficacy of CD47 mAbs [
173]. EGFR is a cell surface target antigen that conducts oncogenic signals and is overexpressed in various malignancies. The generation of BsAb CD47xEGFR-IgG1 blocks CD47 expressed on the surface of cancer cells in an EGFR-directed manner, effectively reducing the targeting/nontumor effects [
159]. Studies demonstrated that BsAb CD47xEGFR has enhanced overall affinity for CD47/EGFR double-positive cancer cells and selectively induces the phagocytosis and immune antigenic processing of double-positive cancer cells; it also enhances the elimination of tumor cells and promotes adaptive anti-cancer immune responses, thereby improving the selectivity and therapeutic efficacy of the CD47/SIRPα checkpoint inhibition approach in EGFR-overexpressing malignancies [
159].
SIRPα/VEGFR1 BsAb: The association of VEGF/VEGFR inhibitor VEGFR1-Fc with CD47-blocking fusion protein produces synergistic anti-tumor efficacy. Targeting CD47 was demonstrated to trigger the macrophage-mediated clearance of recurrent NSCLC cells, and targeting both VEGF and CD47 via the VEGFR1-SIRPα fusion protein induces macrophage infiltration and enhances the ability to destroy anti-tumor cells and sensitize NSCLC to angiogenesis inhibitors and CD47 blockade [
160]. VEGFR1D2-SIRPɑD1, consisting of the second extracellular structural domain of VEGFR1 (VEGFR1D2) and the first extracellular structural domain of SIRPɑ (SIRPɑD1), exerts its potential anti-tumor effects in glioblastoma treatment by inhibiting VEGF-induced angiogenesis and activating macrophage-mediated phagocytosis [
174].
CD47/MSLN BsAb: MSLN is a cell surface glycoprotein that is overexpressed in a variety of solid malignancies, including gastric, lung, mesothelioma, pancreatic, and ovarian cancers [
175]. BsAbs are formed by combining the high-affinity binding arm of MSLN with the blocking CD47 arm. They are designed to target MSLN and CD47 dual-positive tumor cells. In an in-vitro phagocytic killing assay, BsAb targeting MSLN/CD47 exhibited stronger ADCP activity by targeting the proximal epitope of the MSLN membrane than the distal region of the membrane, optimized ADCC activity by enhancing FcγR-IIIA activation, and enhanced ADCP through a more effective blockade of CD47/SIRPα. This BsAb also exhibited superior anti-tumor activity in a xenograft tumor model [
161].
SIRPα/CD40L BsAb: Preclinical studies demonstrated that CD40 signaling enhances CD47/SIRPɑ blockade on the phagocytosis of tumor cells by macrophages and the cross-presentation of tumor antigen CD8
+ T cells by DCs [
176]. The novel BsAb SIRPɑ-Fc-CD40L is a two-sided fusion protein capable of binding the extracellular structural domains of SIRPɑ and CD40L and also of binding a central Fc structural domain [
162]. SIRPɑ-Fc-CD40L, which binds CD47 and CD40 with high affinity, potently enhances anti-tumor immunity by synergizing type I IFN responses through CD40 stimulation with CD47/SIRPɑ blockade. In a cynomolgus macaque model, SIRPɑ-Fc-CD40L stimulated the elevation of multiple serum cytokines and the marginalization of CD40
+ B cells in a dose-dependent manner; however, no signs of hemolysis, hemagglutination, or thrombocytopenia were observed in vitro or in nonhuman primates [
162]. Furthermore, mouse-derived SIRPɑ-Fc-CD40L exhibited superior anti-tumor activity and long-term immune effects over CD47 mAbs and CD40 mAbs in a mouse CT26 tumor model. SIRPɑ-Fc-CD40L synergized with PD-1 and CTLA4 ICIs to increase DC activity, upregulate type I interferon-gamma response, and enhance macrophage-mediated phagocytosis in vitro, as demonstrated by the significantly higher phagocytic activity of SIRPɑ-Fc-CD40L in combination with rituximab in lymphoma cell lines and mouse tumor models (CD20
+WEHI3 and A20) compared with in combination with CD47 mAb and rituximab [
162].
SIRPα/CTLA4 BsAb: CTLA-4 is an immune checkpoint protein highly expressed on Tregs in the TME [
177,
178]. An anti-CTLA-4 antibody targeting Tregs binds a heterodimer of SIRPα that selectively blocks CD47 on Tregs in tumors by blocking “don’t eat me” signaling and enhancing “eat me” signaling to deplete Tregs in an Fc-dependent depletion of Tregs. In MC38 and CT26 mouse colon cancer models, anti-CTLA-4 × SIRPα preferentially depleted immunosuppressive Tregs of ICOS
high in the TME and enhanced immunity against solid tumors. This BsAb has a lower affinity for individual targets and therefore exhibits lower toxicity than anti-human CTLA-4 [
163]. In the MC38 mouse colon cancer model, a single low dose administered systemically for 5 days had a half-life of more than 21.4 h and was preferentially concentrated in tumor tissues compared with normal tissues and organs. In addition, it promotes IFN-γ-dependent T cell responses, which reduces the tumor burden in mice [
163].
CD47/DLL3 BsAb: Recently, the FDA granted orphan drug designation to PT217 as a potential treatment option for patients with small cell lung cancer (SCLC). PT217 is a first-in-class BsAb with Fc effect, designed to target DLL3 and CD47 in patients with SCLC and other neuroendocrine cancers. DLL3 is highly expressed restrictively in SCLC cells and could be an attractive target for SCLC immunotherapy. PT217 is intended to mediate potential antibody-dependent cytotoxicity of NK cells against tumor cells and to block the interaction of CD47 with SIRPα [
164]. PT217 exhibited potential inhibitory activity in preclinical mouse xenograft models, and toxicity studies in nonhuman primates and rats supported the first human clinical trials. At PT217 doses of 3, 10, and 30 mg/kg, the erythrocyte, hemoglobin, leukocyte, and reticulocyte counts remained within normal limits during the first 21 days of treatment in the nonhuman primate model. In a presentation at the 2022 AACR Annual Meeting, researchers presented the structure of a phase 1 dose-escalation trial for studying the role of PT217 in patients with DLL3 positive SCLC, large cell neuroendocrine carcinoma, and neuroendocrine prostate cancer. A dose-escalation-guided 3 + 3 design will evaluate PT217 at five different dose levels administered weekly [
164].
Claudin18.2/CD47 BsAb: On June 15, 2022, Phanes Therapeutics announced that its Claudin18.2/CD47 BsAb PT886 had received FDA clinical trial approval to conduct a phase I clinical trial in patients with gastric cancer, gastroesophageal junction cancer, and PDAC. PT886 is an anti-CLDN18.2/anti-CD47 BsAb with a natural IgG structure [
165]. It is achieved through two mechanisms of tumor killing: (1) the CD47/SIRPα axis is blocked and macrophages are stimulated to phagocytose tumor cells; and (2) the functional Fc of BsAbs mediates the potential effects of NK cell antibody-dependent cellular cytotoxicity (ADCC) and macrophage antibody-dependent cellular phagocytosis (ADCP) [
165]. PT886 has a high affinity for CLDN and a low affinity for CD47, which gives it a high safety profile and the ability to bind specifically to CLDN-expressing tumor cells, while binding weakly or not to CD47-expressing normal cells. In
in-vitro phagocytosis assays, CLDN18.2 binding resulted in a stronger stimulation of phagocytic activity by PT886. The anti-tumor activity of PT886 was also demonstrated in an
in-vivo pancreatic cancer xenograft model, where PT886 treatment resulted in complete tumor clearance at doses as low as 1 mg/kg. PT886 also has a good safety profile in nonhuman primate and demonstrates manufacturability similar to conventional monoclonal antibodies [
165].
In summary, most of the clinical trials of BsAbs, targeting CD47 are in phase I clinical trials, and their value in solid tumors has yet to be verified.