Some ideas from Macrogenics: B7-H3, DARTs, ADCs and more

Screen Shot 2014-03-10 at 12.54.13 PM

Macrogenics is a very interesting company whose next 2 or 3 years will surely determine its’ future. And it’s probably about time, as the company has 14 years under its belt. The company is public and trades on NASDAQ as MGNX.

We’ve watched many companies approach a critical time horizon, when possible futures begin appearing. Some companies become great, others wither away, but it is no longer possible to tread water. Macrogenics knows this well, and they are aggressively working to carve a successful path forward. CEO Scott Koenig and CFO James Karrels rang me up the other day, spending about an hour giving me a look under the hood. Our discussion was part history lesson, part update and all very interesting. What follows is just my personal opinion about what we discussed. I’ve tried to place my thoughts in the larger context of the antibody and immunotherapy landscape in the treatment of cancer.

The Macrogenics’ story includes an enhanced Fc technology for cytotoxic antibodies, a bi-specific antibody technology known as DART (see below), a series of partnerships with pharmaceutical companies, and some new initiatives. I’m not going to discuss the past type 1 diabetes effort and the teplizumab antibody as that story is well known.

Lets start with a very interesting program that has been built around the HER2 enhanced Fc antibody, margetuximab. Results from the Phase 1 trial were presented at ASCO last year, and there were early signs of therapeutic activity. The early data showed activity at doses ranging from 1-6 mg/kg qw and 10 mg/kg q3w. This is in the dose range already used for Herceptintm (trastuzumab; 4mg/kg qw or 8mg/kg q3w) and Kadcylatm (trastuzumab emtansine; 3.6 mg/kg q3w). Macrogenics’ hopes that margetuximab can eliminate cancer cells expressing lower levels of Her2 than tumors addressable by trastuzumab (which targets Her2 high expressing cancer cells). Program success hinges on this hypothesis.

Macrogenics is aggressively moving this antibody into Phase 2 and 3 programs including breast, bladder and gastroesophageal cancers with relatively low expression of Her2. The planned phase 3 trial of margetuximab plus chemotherapy in gastroesophageal cancer should start by the end of this year. A phase 2a trial in metastatic breast cancer should wrap up in the second half of this year, and we should see data in 2015. This is important because in the absence of compelling clinical data it will remain difficult for this program to drive further valuation. This is because the Her2 space is very crowded, and it may be hard to establish differentiation and gain leverage. The broad attack against diverse cancer types therefore makes sense. So while I like this program it is going to require compelling phase 2 data to really generate buzz.

The other day I wrote about immune-checkpoint programs         (http://www.sugarconebiotech.com/?p=522) and in this area Macrogenics has nicely positioned its Fc-enhanced anti-B7-H3 antibody MGA-271. This first-in-class program is optioned to Servier, and has advanced to Phase 1 in solid tumors. The company was lucky to pick up this antibody target when it acquired Raven. It is one of a number of B7 family proteins with ill-defined biology. Importantly, Macrogenics focused on the expression pattern of this protein, that is, they treated it as a cell surface target and not as a biology target. This turned out to be a smart move, as this protein is highly and preferentially expressed on a wide variety of tumor cells, and Macrogenics is going after this antigen with a cytotoxic antibody. This type of immune targeting antibody is well positioned for combination with other therapies that “unleash” immune cell responses, particularly the NK cell and macrophage release strategies. We talked about a few of these last week (http://www.sugarconebiotech.com/?p=524). This program looks very interesting and has nothing but upside assuming we do not see a toxicity signal as the clinical trials move forward.

Macrogenics is smartly building out this space. Scott and Jim explained that some of the money recently raised in their stock offering is going to support best-in-class antibody process and manufacturing, and that this will include antibody-drug-conjugates (ADCs), which are toxin payloaded antibodies. Importantly, DARTs directed to appropriate targets will internalize – a prequisite for good ADC-mediated cell death. I asked them if they had chosen a partner for the ADC work (like Seattle Genetics or Immunogen) and they hedged just a bit, saying that different linker-payload combinations might be employed for different antibodies and targets. That’s certainly a reasonable approach, if heavy on the downstream process and formulation steps, so we’ll see what they decide.

An important part of the Macrogenics portfolio is certainly the DART technology. DART is an acronym of Dual Affinity ReTargeting (or Redirected T-cell), depending on the compound. So, there are a couple of different plays here. One is a T cell engagement technology. The furthest developed technology in this class is Micromet/Amgen’s Bi-specific T-cell Engager (BiTE). The CD3 x CD19 BiTE, blinatumamab, recruits T-cells through CD3 and directs them to kill CD19 positive cancer cells. Blinatumamab is reportedly active at concentrations of 100pg/ml or less and certainly in some leukemia settings induces very impressive and durable therapeutic responses, although the side effect profile includes CNS toxicity, including encephalopathy. We will talk much more about the BiTE technology in the next post.

I asked Scott and Jim about the encephalopathy toxicity, specifically whether this is a class effect due to T cell activation. Scott pointed to the way Micromet BiTEs are constructed, and the “floppiness” of the two arms, suggesting that this could have a different impact on a responding T cell than the covalently “locked” DART construct. A second point is that this could be a toxicity that is only seen with the CD19 x CD3 bi-specific and not other bi-specifics. Scott mentioned that similar toxicity has been seen in the setting of CAR-T CD19-directed therapy (CAR19). Honestly, we can’t really judge at this point and will have to await clinical results before we actually know the efficacy/toxicity profiles of various T cell recruitment and activation technologies will compare.

Preclinically, Macrogenics has done a nice job of differentiating itself from BiTEs. In an in vitro study comparing CD19 x CD3 bi-specific formats using the DART technology and the BiTE technology the DART compound was active at much lower concentrations, including against patient derived chronic lymphocytic leukemia cells       (http://bloodjournal.hematologylibrary.org/content/117/17/4403.full). This may not matter so much just yet, as Macrogenics has chosen different bi-specific pairings, and for the moment will not compete directly with the CD3 x CD19 modalities, whether blinatumamab or CAR19.

Macrogenics DART technology has been validated in the partnership space. Boehringer Ingelheim (BI), Servier, Pfizer, and Gilead have all bought into the DART story with partnership deals. BI signed for up to 10 targets across diverse therapeutic areas and modalities and recently choose a DART compound to advance into preclinical development. Pfizer also signed a DART technology deal in 2010 for two cancer targets. Most recently, Gilead acquired rights to four pre-clinical DART programs for cancer indications. Gilead will fully fund research activities for all four programs and will receive global rights to three of the programs. Servier has rights to three DART programs and recently exercised an option to develop the MGD006 DART molecule in development for hematologic malignancies. This antibody is bispecific for CD123 (expressed on leukemias and lymphomas) and CD3 (expressed on T cells). Pre-clinical studies showed that the compound killed CD123-expressing leukemia cells at very low concentrations. A Phase 1 study in relapsed and refractory acute myeloid leukemia will start in the second quarter of 2014. This is the first study of a DART in the clinic.

So while this falls short of actual clinical success, the fact that diverse companies have lined up here is promising. Additional deals should be expected. Macrogenics also mentioned that they have an NK cell retargeting platform as well (one that would compete with the BiKE platform) and it will be interesting to see if deals are made on this technology as well.

Lets take a closer look at the DART targets.

The T cell engagement targets are CD123 (for MGD006) a target on acute myeloid leukemia (AML) cells, and gpA33 (for MGD007) a target on colorectal cancer cells. CD123 expression on AML is a target for ADCS, Bi-specific technologies, and CAR-T technology (CART123). So, this is an important target to understand.

CD123, a subunit of the IL-3 receptor, is over-expressed on AML tumor cells (and other hematopoietic tumor types). It is also expressed on normal hematopoietic stem cells, at a somewhat lower level. Recently, cancer stem cells (CSC) have been highlighted as sources of resistance to therapy. These are stem-like tumor cells that are very resistant to chemotherapy or irradiation, and are hypothesized to be a component of relapse in various tumor types. AML CSC are CD123 positive. Ideally then, therapeutics targeting CD123 will deplete AML tumor cells, deplete AML CSCs and hopefully not deplete normal cells. Because of the uniqueness of the CSC hypothesis, agents targeting CD123 and other CSC markers have gotten a lot of attention.

Lets start with CART123. Just a quick reminder, with CAR-T we are talking about the transduction of patient T cells with a modified TCR, a CD3 subunit and the 4-1BB signaling domain. Very nice preclinical data were presented at ASH last year by the group at The Children’s Hospital of Philadelphia         (https://ash.confex.com/ash/2013/webprogram/Paper60937.html) using technology that has been licensed to Novartis. However, there has been no further news on this target. Very recently, a second CAR-T/CD123 program was described by investigators in Italy and the UK (http://www.nature.com/leu/journal/vaop/ncurrent/abs/leu201462a.html). The preclinical data were compelling, and there did not seem to be an effect on normal cells. We’ll have much more on CAR-T technologies in a separate post.

Further along are competing anti-CD123 antibodies. Xencor developed two anti-CD123 antibodies that were then licensed to CSL limited. The first, CSL360, failed to show signs of clinical activity in a Phase 1 AML trial. The second, CSL362, had excellent cytotoxic activity in preclinical models and a Phase 1 trial in AML is recruiting. In December of 2013, CSL Limited licensed this program to Janssen/Johnson&Johnson. So, this program now has some real muscle behind it. An interesting note on the trial, it is being run in patients currently in remission. If I think this through I think this means the therapeutic hypothesis is two-fold. One, to drive the leukemia to MRD status (minimal residual disease = below the limit of detection); two, to eliminate CD123+ CSCs.

Stemline (NASDAQ: STML) has brought an anti-CD123-ADC antibody into the clinic. They had several presentations at ASH last year, including a phase 1 trial in Blastic Plasmacytoid Dendritic Cell Neoplasm, a rare cancer with high expression of CD123. They showed 5/5 patients responded, with 4/5 having a complete                     response (https://ash.confex.com/ash/2013/webprogram/Paper64672.html). This will be an interesting therapeutic to watch.

Macrogenics presented preclinical data on MGD006 at ASH         (https://ash.confex.com/ash/2013/webprogram/Paper55959.html).  As mentioned above the phase 1 study in AML will start later this year. A second CD123 x CD3 bispecific is being developed by the Cancer Research Institute at Scott & White Healthcare in Texas (they already have an IL-3 fusion protein). There is also a tri-specific targeting CD33, CD123 and CD16 (to activate NK cells). This has made it to the clinic (http://www.translational-medicine.com/content/11/1/289).

Gpa33 is an antigen that is highly overexpressed on colon cancer. This antigen was targeted a few years ago using a humanized anti-gpa33 monoclonal antibody. However the humanization effort did not work well and the therapeutic was highly immunogenic. As far as I can tell Macrogenics is alone in this space.

An earlier stage bi-specific technology at Macrogenics targets multiple antigens. Their CD32 x CD79 bispecific cross links these receptors on B cells and stops cell activation.

As discussed above much of the effort ongoing at Macrogenics is directed to their many partnerships. We do not know the targets for most of these, but one can imagine the direction that Gilead might take, or perhaps Pfizer. The deal with Servier is very interesting in the context of the Servier CAR-T technology deal. As reported in mid-February Servier will collaborate with Paris-based Cellectis on UCART19, an engineered T cell with a chimeric antigen receptor targeting CD19, plus 5 other programs all in leukemia and lymphoma. The company plans to develop combination therapies with immunotherapeutic monoclonal antibodies, small molecules, etc.

What we are seeing then is the co-development of bi-specific modalities directed to the same targets as CAR-T modalities, sometimes by the same company. This latter point is critical and fascinating. Could DARTs and BiTES compete not only with ADCs but also the CAR-T technologies? I don’t know, but we may find out in just a few more years. And we do have to be patient – Macrogenics has no clinical news scheduled until 2015. One way the company could make a splash is if it were to do something big on the corporate side – a buyout, merger, acquisition. Certainly their promotional deck (link here: http://bit.ly/1qpPQWL) makes the point on the last slides that Macrogenics is flexible as it has plenty of capital.

I really like this company – their technology is gutsy and innovative and I wish them the best. Now it’s time for clinical execution: the data will guide us from there. In the meantime there are a few really interesting ways to think about the technology opportunity and the underlying equity value.

cheers, and stay tuned for some thoughts about BiTEs, BiKEs, CARTs and KITEs.