Cancer immunotherapy has been advancing on several fronts, most strikingly in the direction of checkpoint inhibition and chimeric antigen receptor (CAR) T-cell therapy. Another front, however, is about to see its share of action. Here, newly engineered bispecific and multispecific antibodies will be put to the test. Such antibodies may engage two or more antigens at once, serving as force multipliers that can exploit opportunities beyond the reach of monospecific antibodies, whether they are deployed solo or in teams.
Although monospecific antibodies are beginning to show their limitations, they should be recognized as part of a sequence of antibody-based cancer immunotherapy developments, a sequence that reaches back at least as far as the Nobel Prize–winning efforts of James P. Allison, PhD, and Tasuku Honjo, MD, PhD. Allison’s work on the CTLA-4 led to the first FDA-approved checkpoint inhibitor drug, ipilimumab (Yervoy, Bristol-Myers Squibb), whereas Honjo’s discovery of PD-1 led to the development of anti-PD-1 drugs such as pembrolizumab (Keytruda, Merck). These drugs and other checkpoint inhibitors have profoundly impacted the treatment of cancer.
An alternative cancer immunotherapy approach, namely CAR T-cell therapy, has also demonstrated its potential to combat cancer. In this approach, T cells are engineered to launch sustained attacks on tumors. Although CAR T-cell therapies clearly have fight in them, they may cede some anticancer glory to bispecific antibodies (bsAbs). The first FDA-approved bsAb to directly compete with CAR-T was the CD19/CD3 drug blinatumomab (Blincyto, Amgen). It was introduced in 2014 for indications in B-cell precursor acute lymphoblastic leukemia.
Even while monospecific antibody–based checkpoint inhibition therapies and CAR T-cell therapies continue to be improved, bispecific and multispecific antibodies are shaping up as a cancer immunotherapy options that may provide significant advantages. At present, companies such as Amunix Operating, Invenra, Glycotope, and Xencor are working independently and in collaboration with larger pharmaceutical companies, such as Novartis, Daiichi Sankyo, and Roche, to bring bispecific and higher-order antibodies into the cancer immunotherapy market. Fundamentally, their engineered expression platforms focus on streamlining novel antibody development, reducing the risk factors to patients, and optimizing tumor destruction.
bsAbs emerged with the technologies developed by two pioneering companies Amgen and MacroGenics. Amgen introduced the BiTE platform; MacroGenics, the DART platform. Despite the availability of such platforms, it can still be a challenge to produce bsAbs that incorporate an Fc domain, suggests John Desjarlais, PhD, senior vice president of research and CSO at Xencor. “If you don’t have an Fc domain,” he says, “you have a very short half-life,” necessitating low and frequent injections or continuous infusion in patients.
Xencor’s solution was to build a robust and GMP-scalable bispecific platform that includes an engineered Fc domain for the antibody, ensuring that antibodies produced with this platform would have a longer half-life in vivo. Xencor’s XmAb Fc platform increases this efficiency of heterodimer Fc formation to 95% out of the gate.
“If I want to make a heterodimeric Fc domain, one that is different on either side,” he says of a traditional process, “I’m going to get a mixture of 50% of the heterodimer, and 25% of the different homodimers by comparison.”
To improve efficiency yet further, Xencor has engineered an additional feature in the Fc domain. “We perturb the isoelectric point on either side of the Fc heterodimer through substitutions in the Ch3 domains,” Desjarlais details. “The idea behind that was, we would have an ability to very easily separate out the small amount of contaminating homodimers just by using ion exchange chromatography.”
Xencor is exploring bsAbs that act as dual checkpoint inhibitors, such as anti-PD-1/CTLA-4 and CTLA-4/LAG-3. The field has learned that cancer evolves to suppress the immune system by engaging different pathways meant to protect the body against autoimmunity.
Single checkpoint blockers on the market such as nivolumab (Opdivo; anti-PD1) and ipilimumab (Yervoy; anti-CTLA-4) have been used in combination to improve antitumor activity, but this approach, says Desjarlais, comes at the cost of increased toxicity. Dual-targeting antibodies may promote less toxicity by more selectively targeting the tumor reactive T cells. “The idea is to turn off the brakes,” he explains, “and the more brakes you can hit at the same time, the more you can activate those tumor T cells.”
In addition to checkpoint inhibitors, Xencor has been successful in establishing two Phase I trials in collaboration with Novartis involving T-cell-engaging bsAbs; one that has an AML indication and binds to CD123 on AML blasts and CD3 on T cells, and a second that binds to CD20 on malignant B cells and CD3 on T cells. The company has a third wholly owned bsAb that binds CD3/SSTR2 (somatostatin receptor 2). Currently in Phase I trials, this bsAb is being explored with dose escalation in neuroendocrine tumors.
“CD3 bispecifics would be considered direct competitors to CAR-T,” asserts Desjarlais. CAR T-cell therapies require weeks of preparation including cellular extraction from a patient, engineering in vitro, culturing, speculative dosing, and continued growth in vivo. In contrast, Desjarlais points out, “a bispecific is something in a vial that you have in the pharmacy.”
“With a bispecific,” he emphasizes, “you know exactly what you’re putting in.”
Volker Schellenberger, PhD, president and CEO of Amunix, affirms that the challenge of the CAR T-cell therapies is that they must be individually created for each patient. “Another challenge,” he says, “is that you are injecting live cells into a patient. So, it is very difficult to control what happens to them. They can even multiply in that person.”
“We need to somehow mitigate the toxicity of these T-cell engagers,” insists Schellenberger. “If you have a protein-based drug, then you could give it right away, instead of after the several weeks it takes to develop an individualized CAR T-cell therapy; that would be a big benefit to the patient.”
Amunix has developed a new format of bispecific T-cell engagers that can be delivered in a low dose with lower toxicity using XTEN technology, an alternative to PEGylation. “The T-cell engager,” Schellenberger explains, “works like an adaptor molecule. It bridges the tumor and the T cell.” XTEN is a protein polymer that is engineered to behave like polyethylene glycol (PEG) which is attached to bsAbs to increase their half-life in vivo without the need for an Fc domain.
“XTEN has evolved into kind of a Lego kit for pharmaceuticals,” Schellenberger notes. “It allows us to make very complex molecules which by other means we just couldn’t produce.”
The company’s lead XTENylated bsAb, AMX-268, is in preclinical development. It is a T-cell engager that binds to CD3, a T-cell receptor (TCR), and EpCAM, an adhesion molecule overexpressed in 80% of solid tumors.
“We give the drug in an inactive form and convert it to the active form only when it is in the tumor environment,” Schellenberger says. The company’s pro-drug is activated by the inflammatory process found primarily within the tumor microenvironment, reducing off-target toxicity and increasing antitumor selectivity, “so that if our molecule finds that target in a healthy organ, it will still leave it alone.”
The active form of the drug is smaller than typical Fc-containing intact antibodies, allowing it to be removed easily and rapidly through the kidney. Schellenberger’s data suggests that AMX-268 may have lower immunogenicity and a lower toxicity profile among other potential EpCAM-targeting T-cell engagers such as Removab (Fresenius Biotech) and the investigational MT110 (Amgen).
Moving from mono- to bispecific antibodies
One company that is leveraging its success in developing monospecific antibodies into bi- and trispecific antibodies is Glycotope. According to Anika Jäkel, PhD, the company’s director of preclinical pharmacology and cancer immunology, “Glycotope has strong expertise in glycobiology and focuses on the generation of antibodies against tumor-specific glycoepitopes.”
The company’s first-in-class mAb, Gatipotuzumab, targets the tumor-specific epitope TA-MUC1, a novel combined carbohydrate/peptide conformational epitope on the tumor marker MUC1 (mucin-1). This antibody shows broad therapeutic potential in 80–100% of its main solid tumor indicators (that is, ovarian, lung, and breast cancers).
“Our most advanced pipeline bispecific is a TA-MUC1-targeting T-cell engager (PankoMab-CD3-GEX),” Jäkel points out. “It was designed to combine the high tumor specificity of Gatipotuzumab with activation of polyclonal T cells independent of MHCI engagement upon simultaneous binding of TA-MUC1 and CD3 on T cells.”
A second molecule in development at Glycotope is PankoMab-PDL-GEX, which combines binding to TA-MUC1 with immune checkpoint molecule PD-L1 attached to a glycol-optimized functional Fc domain. PankoMab-PDL-GEX is designed to direct checkpoint blockade to the tumor and thereby enhance tumor cell killing.
Glycotope’s GlycoExpress (GEX®) technology platform is used for screening and production of biopharmaceuticals, such as those described above, and other glycoproteins for fully human optimized glycosylation. “It consists of a toolbox of proprietary human cell lines generated by glycoengineering,” says Jäkel. “It is biotechnologically optimized for product improvement as well as fast, reproducible, and high-yield glycoprotein production.”
“We do not use a standard platform approach for our bispecific programs,” Jäkel continues, suggesting that by focusing on GlycoTargets, the company has positioned itself to screen several construct formats for each bispecific product idea. “We can produce classical IgGs but also bispecific formats in our GlycoExpress system,” she asserts. “We can test different glycosylation variants for identification of a lead candidate with highest antitumor efficacy.”
Although Glycotope is not exclusively focusing on the bsAb market, Jäkel suggests that there are many possible advantages to targeting two epitopes over monospecific antibodies, including increased specificity and/or avidity, increased inhibition of tumor growth, enhanced local tumor cell killing, and blockade of immune checkpoint inhibitors.
In immuno-oncology, a well-trod path is the redirection of tumor T cells. A less-well-traveled path is being explored by Invenra, which seeks to activate functional processes that require a novel mechanism of action through bispecific and higher-order antibody binding.
“A good example is agonist antibodies for the tumor necrosis factor [TNF] receptor superfamily,” says Bonnie Hammer, PhD, vice president of biologic development at Invenra. “The ligands for that family are trimeric. To get good activity, you need at least three receptors coming together, but it is even better if you have even higher-order clustering.”
Antibodies that drive this type of receptor clustering are the focus of Invenra’s ARCHER (Agonistic Receptor Clustering by High-order Exogenous Rearrangement) technology. One of the receptors in the TNF superfamily, OX-40, is the target of an Invenra bsAb in lead selection.
To engage the higher-order clustering, Invenra used its B-Body multispecific antibody development platform to produce a bispecific with a two by one (2 × 1) format. “The bispecific has three Fab domains,” Hammer notes. “But two Fab domains bind to one epitope, and the other Fab domain binds to a different epitope.”
“Traditional monoclonal antibodies for OX-40 have suffered in the clinic,” Hammer says, pointing out that they are dependent on having Fc engagement to provide the secondary crosslinking needed for activity. In contrast, she continues, Invenra’s OX-40 agonist has allowed the company “to achieve activity in the absence of any additional crosslinking by targeting multiple epitopes.” Although the OX-40 agonist has yet to see the clinic, Hammer suggests that the agonist “will provide higher activity than has been previously seen with monospecific antibodies.”
A bacteriophage library that consists of wholly human Fab fragments and that matches the natural diversity found in the human repertoire can provide the starting point for selecting Fabs of interest used in Invenra’s B-Body platform, Hammer says. A domain-substitution strategy with a few orthogonal chain mutations allows for highly specific light chain–heavy chain pairing and enables high-throughput production and purification of bispecific and multispecific antibodies.
“We found that you can predict some things [during antibody design],” she reports, “but a lot of it is through empirical testing. The affinities for the antibodies, the geometry, and the epitopes that you’re hitting matter.” One other group of multispecific antibodies in Invenra’s pipeline consists of discovery candidates that create higher specificity through the targeting of more than one antigen. “These candidates are the bispecific antibodies we call the SNIPERsTM,” says Hammer. Currently a regulatory T cell–depleting SNIPER molecule is in lead selection.
Ian Clift Ph.D. is Scientific Communications Consultant, Biomedical Associates and Clinical Assistant Professor, Indiana University