After a challenging period, bispecific antibodies (bsAbs) are making a comeback in hematology as they become a prominent drug class and a major threat to CARs.
Historically, excitement around bsAbs peaked in 2012 following the Amgen (AMGN)/Micromet acquisition but gradually faded due to disappointing results for Amgen’s Blincyto and other bsAbs. Blincyto (a CD19xCD3 T cell engager) is designed to kill tumor cells by simultaneously binding T cells (via CD3) and leukemia cells (via CD19), bringing them in close proximity and triggering a T cell response. The drug was approved in 2014 and is a growing product ($85M in Q3/19) but it remains the only FDA-approved bispecific T cell engager and uptake has been hampered by safety issues and the need for continuous administration. Replicating Blincyto’s success with other targets proved challenging and other programs based on Micromet’s or competing platforms have not generated attractive data until recently.
With bsAbs not living up to expectations, investors turned to CARs that provided an alternative to T cell engagers. The idea was similar (mounting a potent and specific T cell response) but instead of recruiting T cells in vivo, CARs are engineered ex-vivo and then re-infused to patients.
The CAR space exploded in 2014-2018 based on exciting data in liquid tumors, especially ALL/DLBCL (CD19) and multiple myeloma (BCMA). CARs generate high response rates and are potentially curative in some indications. The graphs below from Yescarta’s pivotal trial in DLBCL demonstrates this potential: Despite the relatively modest PFS of 6 months, 40% of patients were still in remission after more than two years. BCMA CARs are equally impressive with response rates of 80-100% with a median PFS of 12-18 months.
Bispecifics strike back
Recent data readouts from multiple clinical trials have made bsAbs the hottest segment in hematology. At ASH 2019, companies presented results in DLBCL (CD20) and multiple myeloma (BCMA).
On the CD20 front, data from Roche, Regeneron (REGN), Xencor (XNCR) and others demonstrated strong efficacy and an acceptable safety profile with agents that can be given weekly or less frequently. Response rates in DLBCL are in the 38-58% range, including CRs. Durability is still premature but most CRs appear durable, up to 2 years, and importantly, re-treatment upon relapse can lead to subsequent responses.
BCMA bispecifics are 1-2 years behind CD20 programs with BMS (BMY) in the lead following recent update for CC-93269. At the higher doses, response rate and CR rates were 88.9% and 44.4%, respectively, with most responses ongoing at the time of cut off.
These results come on the heels of Amgen’s AMG420 data at ASCO 2019, which demonstrated a 70% response rate at the relevant dose. In contrast to CC-93269, AMG420 is a classic BiTE like Blincyto and has to be given via continuous infusion. Amgen is now focusing on a half-life extended BCMA program, AMG701, which should be at a similar stage to that of CC-93269 but no clinical data have been published. Regeneron presented preliminary (7 patients) data for its BCMA bsAb (REGN5458) demonstrating a 57%-75% response rate.
These data are preliminary and need to be corroborated by larger studies but definitely make BCMA bsAbs a promising class for multiple myeloma, together with BCMA CARs and BCMA ADCs (see below a table from Nature Reviews Drug Discovery). At ASH 2020, investors should expect more mature data from several BCMA programs that are still in dosing optimization.
Nat Rev Drug Discov. 2019 Jul;18(7):481-484
Threat to CARs
Recent data with bsAbs put them on a collision course with CARs, especially in DLBCL and multiple myeloma, where data look promising for both classes. While both are designed to induce T cell responses against cancer cells, they represent very distinct approaches and product profiles. CARs are personalized living therapies typically given once, with complex logistics. Bispecifics are simple off-the-shelf products that are administered repeatedly.
Looking at the parameters in the table below, it is easy to see why physicians and patients would prefer bsAbs over CARs. They are readily available, logistically simpler (no cell collection, no conditioning, no patient-specific shipments, no waiting periods, no production failures etc.). They also allow dosing flexibility (dose titration, treatment holidays etc.) and their manufacturing cost is low. One potential advantage CARs have is the fact they don’t require chronic treatment.
But what about efficacy, especially long term remissions? Both CD19 and BCMA CARs lead to very high response rates including a long survival tale in a significant portion of patients with CD19 CARs. It will be challenging for bsAbs to match response rates seen with CARs and durability is still an open question, but so far there appear to be multiple cases of long term remissions with CD19 bsAbs even after patients are taken off treatment.
In addition, clinical data for CARs are confounded by the fact that they do not include “technical failures”, (patients who could not receive treatment due to technical and clinical issues). These technical failures occur in 10-30% of cases based on pivotal studies and real-world failure rate is probably higher. Including these cases in the data set will significantly diminish CARs’ efficacy metrics.
While the burden of proof is still on bsAbs, I personally think bsAbs will capture a significant market share even if they are less efficacious than CARs, especially in community centers where the majority of patients are treated. The two classes are not mutually exclusive as patients who fail one treatment modality may receive the other after relapse and this has been reported with both modalities with successful outcomes. Allogeneic CARs address some logistical issues but their clinical profile is still unclear (especially persistence) and as a cell therapy they will always be more complex, expensive to manufacture and administer (including conditioning). It will be interesting to see data for several allogeneic CARs in 2020, which could shed light on CARs’ potential to become a widespread treatment option.
bsAbs for solid tumors
So far success with bsAbs has been limited to hematology. Solid tumors are less accessible and have a challenging immune suppressive mileu, which makes it challenging for current T cell engagers to have an effect at tolerable doses. Other efforts to target two “oncogenic” targets (VEGFxAng2, VEGFxDLL4…) or immune checkpoints (PD1xLAG3, PDL1xTGFb…) either failed or are too early to assess.
The most successful bispecific program in solid tumors I am aware of is Zymeworks’ (ZYME) ZW25 (HER2xHER2), which demonstrated encouraging single agent activity but with somewhat limited durability.While ZW25 has clear potential in HER2 expressing GI tumors, the more attractive program in the company’s pipeline is ZW49, an ADC based on ZW25. Targeting two sites on the same receptor may increase potency without having an effect on off-target toxicity, thus expanding the therapeutic window, which is ADC’s main problem. Who knows, perhaps bispecific antibodies could become an important enabling technology for ADCs and vice versa…
I am selling Spark (ONCE) and Audentes (BOLD) following their respective acquisitions. I am also selling a portion of Xenon (XENE) given its disproportional weight (15%) in the portfolio and adding positions in Madrigal (MDGL) and Viking (VKTX). Cash is now almost a third of the portfolio as it has been challenging for me to find new stocks at attractive valuations.
Portfolio holdings – Jan 19, 2020