Immunogen and Seattle Genetics: Verging on an Inflection Point

by: Ohad Hammer

This year’s ASCO annual meeting should be a very exciting event for anyone who has been following the field of antibody-drug conjugates [ADCs]. During the conference, investigators will present impressive clinical data generated by ADCs powered by Immunogen’s (NASDAQ:IMGN) and Seattle Genetics’ (NASDAQ:SGEN) technologies. The data includes studies for Genentech’s (Private:DNA) T-DM1, Seattle Genetics’ SGN-35 and Curagen’s (CRGN) CR011-vcMMAE . These data will put ADCs on the verge of transitioning from a remote niche to one of the hottest areas in oncology.

Here is a summary of everything you always wanted to know about ADCs but were afraid to ask:

  1. Cancer is a disease caused by uncontrolled growth and division of cells in the body. These cells are very similar to healthy cells, which makes developing effective anti-cancer drugs very challenging. Fortunately, cancer cells can often be distinguished from normal cells based on structural elements they present on their outer surface. These cancer specific elements are generally termed tumor-associated antigens or TAAs.
  2. Antibodies are proteins that recognize and bind specific structural elements. These properties can be harnessed for the development of targeted anticancer drugs that affect cancer cells while sparing healthy ones.
  3. In the past decades, technologies that enable the creation, production and evaluation of antibodies against tumor associated antigens, led to the development and commercialization of anti-cancer antibodies. These antibodies, such as Rituxan and Erbitux bind TAAs presented on cancer cells with minimal side effects.
  4. Upon binding cancer cells, antibodies can exert their therapeutic effect via a number of mechanisms such as recruitment of the immune system and disruption of growth signals.
  5. Binding cancer cells does not necessarily lead to an anti-cancer effect, as most cancer-specific antibodies just attach to cancer cells without causing any damage. These antibodies might be useful for diagnosis or imaging purposes but not as drugs. Thus, the approved anti-cancer antibodies represent a tiny fraction of available cancer-specific antibodies.
  6. the need to boost antibodies’ potency by coupling them with effector molecules such as chemotherapy drugs. An ADC is an antibody that is linked to a toxic payload of chemotherapy drug.

Some like to describe antibodies as guided missiles because they can find and attack specific targets without causing collateral damage to their surroundings. Using the same analogy, although these guided missiles can hit the target, the damage they inflict is limited, so there is a need to arm them with more potent warheads. This is the rationale behind developing ADCs – utilizing antibodies’ ability to identify cancer cells with the ability of chemotherapy drugs to kill them.

Although the concept of ADCs is very intuitive, the actual development of these agents has been proven to be excruciatingly difficult. As a result, most companies stayed away from this field, focusing on developing “naked” antibodies that are not linked to an effector molecule. Today, however, technologies for arming antibodies are finally mature enough to allow drug developers to pursue this promising path.

The two most prominent technologies for creating ADCs were developed by Immunogen and Seattle Genetics for over a decade. If proven effective, these platforms can be utilized for developing an unlimited number of drug candidates for a broad spectrum of cancer types. Unlike other emerging fields, the growth and adoption of ADCs may be very rapid thanks to the vast experience and insight gained in the antibody industry. For over 30 years, investigators have identified clinically relevant targets, produced cancer specific antibodies and developed model systems for evaluating efficacy. The big splash ADCs are about to make could not have come at a more opportune time for the pharma industry, which is going through a severe innovation crisis. The company that could benefit the most from ADCs is no other than Genentech, whose antibody pipeline looks like a pale shadow of the great innovative pipeline it had in the 90’s. Genentech has invested a lot of effort and created one of the broadest collection of antibodies against a plethora of well characterized tumor associated antigens, so within several years, Genentech can theoretically have a dozen of ADCs in the clinic at minimal cost.

Back to this year’s ASCO meeting: Positive data from three different clinical programs will be presented. Genentech will present two phase I trials where T-DM1 (which utilizes Immunogen’s technology) was given to breast cancer patients. Seattle Genetics will give an update from the phase I trial of SGN-35 in late stage Hodgkin’s Lymphoma. Curagen will publish results of CR011-VCMMAE in metastatic melanoma patients.

Genentech presented results from two dose escalation studies of T-DM1, which is comprised of Genentech’s blockbuster antibody, Herceptin®, and Immunogen’s linker and effector molecules. Of note, the patients enrolled to the trials had previously progressed during treatment with Herceptin in combination with chemotherapy. These patients are considered “Herceptin-resistant” so any clinical effect demonstrated by T-DM1 in this patient population implies that the arming of Herceptin boosts its potency and thus can serve as a strong validation for Immunogen’s ADC technology.

The two phase I studies differed only in the dosing schedule. In the first trial, T-DM1 was administered every 3 weeks whereas in the second trial, it was administered on a weekly basis. In the case of the three-weekly study, most of the data was already disclosed in previous conferences, including objective responses (significant reduction in tumor burden) in six out of 15 patients who received the maximal tolerated dose (3.6 mg/kg). One piece of new data was the median progression-free survival (NYSE:PFS) of those fifteen patients, which was 9.6 months, substantially higher than patients who received lower doses of T-DM1. These results led to the initiation of a phase II trial that will evaluate T-DM1 in the same dosing schedule in 100 patients.

Data from the second phase I trial, with the weekly dosage of T-DM1, is still very preliminary. Nevertheless, this trial has the potential to demonstrate even better results because this dosing schedule may be more appropriate for ADCs. While naked antibodies stay in the bloodstream for weeks, most ADCs are cleared out the body within a week after administration. In the case of Herceptin, for instance, substantial amounts of the antibody remain in the bloodstream for over a month. Arming Herceptin with a toxic payload of DM1 dramatically reduces the circulation time to less than a week. Therefore, the patients who received T-DM1 every three weeks actually spend two weeks of every cycle without significant levels of the T-DM1 in their bloodstream. Using a weekly administration schedule, it may be possible to maintain stable levels of the ADC in patients’ bloodstream, if, of course, sufficient doses can be given every week without leading to severe side effects. If so, patients could receive higher cumulative doses of the drug. Because there is not a lot of experience and knowledge about the safety profile of ADCs, all initial evaluations were done at a three-weekly schedule. Now that Genentech showed that T-DM1 can be administered at a relatively high dose (3.6 mg/kg every three weeks) with a good safety profile, the move to a more frequent dosing regimen was obvious.

The weekly trial was a typical dose escalation study, with the aim of finding the highest dose that can be given to patients. So far, Genentech disclosed data only for 7 patients who received escalating doses of T-DM1. The lowest dose was 1.2 mg/kg per week, which is equivalent to the MTD from the first phase I trial (3.6 mg/kg). The next cohort of 3 patients received a dose of 1.6 mg/kg per week without any signs of substantial toxicity. Of note, the equivalent dose in the three-weekly trial was found to be too toxic. Therefore, even if the MTD is eventually set at 1.6 mg/kg per week, this trial already enables the evaluation of higher cumulative doses. It does not mean that this regimen will necessarily be more efficacious, but it certainly increases the probability for a stronger effect. The seventh patient was given a dose of 2 mg/kg per week, the same dose at which Herceptin is given. If this dose level is proven safe, that would serve as a great sign with regard to T-DM1’s safety profile. On the efficacy side, of the seven patients, 4 (57%) achieved a partial response. This response rate is encouraging, but in order to show some sort of an advantage over the three-weekly regimen, there must be data on more patients. Nevertheless, seeing that number of responses before establishing an MTD is quite encouraging. Because these results were submitted more than four months ago, it is reasonable to assume that Genentech will present more data with respect to the MTD and the response rate from the trial, so it looks like there is a lot to look forward to.

Seattle Genetics will report results from a dose escalation phase I trial of its wholly-owned ADC, SGN-35, in heavily pretreated Hodgkin’s Lymphoma patients. Preliminary data from this trial was already reported last year, and included multiple partial responses. Since then, the company recruited additional patients at higher dose levels (up to 2.7 mg/kg), but still did not reach MTD. The fact that even at a dose of 2.7 mg/kg every three weeks, there were no severe side effects, implies that Seattle Genetics is capable of creating very safe ADCs as well. According to the data submitted to ASCO in the beginning of the year, 9 out of 28(32%) evaluable patients achieved a partial response. In addition, there was a clear dose dependent response, as at dose levels of 1.2 mg/kg or higher, 7 (54%) of 13 patients had a PR. This compares favorably to only two cases of PR among the 15 patients receiving doses lower than 1.2 mg/kg. Since submitting the data, Seattle Genetics has recruited patients at higher doses and expects to present more responses at the conference. The company initiated a trial for evaluating weekly doses SGN-35 and plans to evaluate this schedule in combination with chemotherapy (gemcitabine) going forward. This is the place to point out that Hodgkin’s Lymphoma represents a limited opportunity, with only 8200 new cases expected this year in the US, the vast majority of whom will be cured with available treatments. Including additional patient populations who may be suitable for SGN-35 treatment, the worldwide potential market for SGN-35 is still less than 10 thousand per year. Therefore, SGN-35 is more of a technology “validator” rather than a substantial commercial opportunity.

CR011- vcMMAE

Results from another ADC powered by Seattle Genetics’ technology, CR011-vcMMAE, will be presented by Curagen. Curagen is expected to report updated results from a phase I dose escalation trial among 27 melanoma patients. It has previously disclosed data from 23 patients, so the new data is only for the last five patients. I wrote about this drug candidate in the past and gave my somewhat skeptical point of view on this trial. My skepticism stemmed from the poor historical success rate in metastatic melanoma and the limited activity the ADC demonstrated (although there was an apparent dose dependent response). The updated results include data from higher doses of CR011-vcMMAE and show a clear dose dependent response, as CR011-vcMMAE managed to achieve one (11%) partial response and led to disease stabilization in 7 (78%) of the 9 patients who received the 3 highest doses. Among the remaining 18 patients who received the lower doses, there was a SD rate of only 39%. These results triggered a phase II trial in additional 32 patients for the evaluation of CR011-vcMMAE given at the MTD.

Bearing in mind that in its naked form, CR011 did not have an effect even on cell cultures, this study can be seen as yet another validation for Seattle Genetics’ ADC technology. The response rate may appear low compared to other oncology programs, but metastatic melanoma is notorious for being resistant to most chemo drugs, which usually demonstrate even worse results. Nevertheless, I am still not a big fan of this clinical program because metastatic melanoma is characterized by the worst ever success rates and the overall market is quite limited. By default, investors should stay away from any drug candidate for metastatic melanoma , with one exception to the above rule: Synta Pharmaceuticals’ (SNTA) elesclomol.

On a more encouraging note, Curagen recently announced plans to initiate another phase II trial in metastatic breast cancer, one of the largest oncology markets. The presence of CR011-vcMMAE’s target, GPNMB, has been observed in many cases of advanced breast cancer. More interestingly, this target has been suggested to play an important role in disease progression and the development of metastases. It is also reasonable to assume that CR011-vcMMAE was found to be active against breast cancer tumors in preclinical systems before the expensive phase II trial was announced. There is still a great deal of uncertainty as to the merits of this study, as GPNMB is not a validated target for breast cancer. However, because the trial will involve only the maximum tolerated dose for CR011-vcMMAE, Curagen should have an answer within 6 months of trial initiation.

Following the positive results for the three ADCs, the next step is carving out a path to FDA approval for each candidate. Navigating a drug candidate towards registration may become an extremely complex task, not only due to the need to show meaningful benefit in the form of PFS or survival in larger, placebo controlled studies, but also because of the need to factor in multiple considerations. These considerations include timing issues, competing products in development, intellectual property and costs. This is where a company’s strategic planning abilities come into play.

The path for SGN-35 is relatively simple, as the Hodgkin’s Lymphoma market is small, with very few drug candidates in the clinic. Because the unmet clinical need is limited, Seattle Genetics will probably first try to get SGN-35 approved in heavily pretreated, relapsed patients, either as mono-therapy or in combination with approved therapies. From that point, the company could try to advance SGN-35 into earlier stage patients, but the commercial rationale for such label expansion is uncertain. The widely used regimens for early stage HL are curative in more than 85% of cases and contain primarily generic compounds. Thus, market acceptance for a treatment that slightly improves likelihood for remission but costs several tens of thousands dollars might be modest.

The situation for T-DM1 is much more complex, with multiple dilemmas at hand, as the market for HER2 positive breast cancer is very active. Adding more complexity is Hercpetin’s strong presence in this market throughout the different stages of the disease. Among the decisions that need to be made are whether and when T-DM1 will eventually replace Herceptin, whether it should be combined with other drugs and which indications should be pursued in the near term future. On the one hand, the fastest route to approval is a phase III trial in Herceptin resistant patients. The advantage in going after this indication is the short trial duration and the option to get the drug approved with less stringent endpoints. However, there are multiple agents in advanced clinical stages for this patient population, including combinations with Herceptin. Two recent examples could be Genentech’s pertuzumab and Kosan’s (KOSN) tanespimycin, both of which showed encouraging activity when combined with Herceptin in Herceptin-resistant patients. A registrational trial for this indication will probably involve the recently approved Tykerb+Xeloda combination as a control arm. If the PFS data of 9.6 months is indicative of the real activity of T-DM1 in larger populations, it might not be enough to show a statistically significant advantage over the 8.5 months of PFS Tykerb®+ Xeloda® demonstrated in their registrational phase III trial. In order to improve chances for approval, Genentech might want to evaluate T-DM1 combined with a chemo agent or in more frequent dosing schedule, but that will translate into a delay of more than 1 year, as every type of combination must be first evaluated for safety in a dose escalation trial.

As long as Herceptin remains the mainstay treatment in the market, Genentech should be in no hurry to replace it with its armed version, especially because it wisely cornered the field of anti HER2 antibodies using patent protection. However, Herceptin’s position may be threatened by other agents that target HER2. The most significant player in that space is Tykerb, which is currently being evaluated head-to-head against, as well as in combination with Herceptin in several different studies. Tykerb’s activity in Herceptin-resistant breast cancer, combined with the fact that it is an oral agent, make it a real threat to Herceptin. Ironically, the heat Genentech is feeling with respect to Herceptin is the best thing that could happen to Immunogen. Genentech is probably bullish on ADCs and T-DM1 regardless of market dynamics, but when a product that has $4.5 in annual sales is at jeopardy, one can count on Genentech to do whatever it takes to defend its turf.

Although the T-DM1 and SGN-35 data presented at ASCO are from small uncontrolled clinical trials, the two trials serve as a strong validation for the technology of each company due to of the patient population in both trials. In the case of T-DM1, all the patients in the trials had previously progressed during Herceptin regimens and are considered Herceptin resistant. This means that naked Herceptin can no longer control the disease, let alone lead to tumor shrinkage.

In the case of SGN-35, the evidence is less striking but I still find it very persuading. SGN-35 is comprised of the antibody, SGN-30, and Seattle Genetics’ linker and toxic payload. SGN-30, the naked antibody has been evaluated in similar patient population and showed only a minor effect, while SGN-35 proved to be highly effective in these patients, at substantially lower doses than those of SGN-30. Since T-DM1 and SGN-35 were evaluated as mono-therapy, the only plausible explanation for the clinical activity is the arming of the two antibodies.

Just to make it clear, it does not mean that both agents will succeed in future clinical trials, but in my opinion, it means that both Immunogen and Seattle Genetics have brilliant technologies that will eventually serve as a basis for multiple approved drugs. As platform companies, the value of these companies should be derived not only from their promising pipelines but also from the huge value embodied in their platforms.

So far, Seattle Genetics has been favored by the stock market and is currently worth four times Immunogen’s market cap. The reason for this difference is probably Seattle Genetics’ focus on naked antibodies for blood cancers, as opposed to Immunogen who chose to focus on ADCs for solid tumors. While solid tumors represent over 90% of the oncology market, success rates in the blood cancer field are substantially higher. Bearing in mind the exploratory nature of ADCs and the disappointing results they have had to date, it is easy to understand why most investors stayed away from Immunogen. Now that ADCs are finally ready for primetime and facing wide industry adoption, Immunogen is becoming very attractive for the same reasons that made it the black sheep in the family: Its focus on ADCs and solid tumors.

For the sake of full disclosure, I am bullish on both companies and do not suggest that Seattle Genetics is a bad investment going forward, as I have expressed my thoughts on its pipeline in previous articles. Nevertheless, Immunogen currently has the lion’s share of the ADC market, as it will have eight compounds in clinical development this year, most of which are being developed and financed by its partners. These compounds include T-DM1 in collaboration with Genentech, three wholly owned compounds (IMGN242, IMGN901, and IMGN388), two compounds in collaboration with Sanofi-Aventis (AVE9633, SAR3419), one compound with Biogen-IDEC (BIIB015) and one compound in collaboration with Biotest (BT-062). There is also a lot of activity behind the scenes with Genentech and Sanofi-Aventis (NYSE:SNY), which should result in a growing flow of candidates into the clinic. Needless to say, most clinical trials in oncology fail, and there is no reason to believe that ADCs will be an exception. For example, the development of IMGN242 will probably be terminated this year, after 8 years of clinical development. The AVE9633 program is another one that should be discontinued, the sooner the better. I know this sentence just cost my inbox tens of angry e-mails but the way I see it, Sanofi is wasting patients’ time and investors’ money with this one. Failures are a natural part of drug development, and this is why companies should strive to cast the widest net in order to get statistics on their side.

SGEN currently has 2 ADCs in the clinic, and is not expected to file an IND for in-house developed ADCs, until next year. Hopefully, partnerships with MedImmune and Agensys [now part of AstraZeneca (NYSE:AZN) and Astellas (OTCPK:ALPMF), respectively] will help it expand the number of ADCs in its pipeline this year.

In summary, the summer of 2008 may be remembered as the inflection point of antibody drug conjugates with the two most promising platforms being those of Immunogen and Seattle Genetics. But ADCs are not the only thing happening in the field of antibodies. Micromet (NASDAQ:MITI), a tiny German-based company represents another promising antibody platform with the potential of transforming the cancer antibodies market as we know it. Although this platform, termed BiTE, is at a very early stage of its evolution, it already showed very impressive signs of activity, as discussed here and here. Micromet is expected to report updated data for its flagship product, MT103 at ICML in Switzerland , two days after the conclusion of ASCO.

Disclosure: Author is long IMGN, SGEN, MITI and SNTA