Hammer Stock Blog

AVE1642 is a “naked” antibody that binds insulin-like growth factor (IGF-1R). It is currently evaluated in a phase I clinical trial that started in October of last year. Interestingly, this antibody is not attached to a drug payload, and is most likely intended to be used in combination with chemotherapy.

IGF-1R is postulated to be a very important target in several types of cancers such as colorectal, lung and breast cancers. This receptor has been shown to contribute to the development and progression of tumors, as its activation triggers a cascade of signals ultimately leading to survival and proliferation. An antibody targeted at IGF-1R may serve as an anticancer agent by preventing the growth factors from binding the receptor or by inducing an immune response against cells that express IGF-1R. IGF-1R is also expressed by normal cells, including blood vessels, which offers an explanation to why Sanofi-Aventis decided not to arm AVE1642 with a deadly payload. This is a good example for cases where ADCs cannot be used, because they will probably lead to unbearable side effects.

Since targeting IGF-1R by monoclonal antibodies seems very promising, several other companies, including Pfizer (PFE) and Imclone (IMCL) are actively pursuing this pathway. Both companies have already published results from phase I clinical trials, showing some clinical activity and a very good safety profile, which makes Sanofi Aventis a little late to the party, but eventually, demonstrating clinical activity is the top priority for AVE1642.

Author is long IMGN

This Antibody-drug conjugate was created by ImmunoGen and licensed to Sanofi-Aventis. AVE9633 consists of the huMy9-6 antibody, which binds specifically to the CD33 antigen found on acute myeloid leukemia cells, and Immunogen’s DM4 cell-killing agent. There expected to be more than 13,000 new cases of AML this year in the US alone, and around 9,000 americans are expected to die as a result of the disease. Although during the last decade, an increase in survival rates was achieved due to the introduction of new treatments, most patients will die less than 5 years after diagnosis. The high likelihood of disease relapse is especially unsatisfactory, despite the relatively high portion of complete responses achieved by chemotherapy and Wyeth’s (WYE) Mylotarg^®, the sole approved antibody-drug conjugate to date. CD33 antigen is present in approximately 90% of AML patients, which makes it a very attractive target. More importantly, the concept of targeting CD33 has been validated by the impressive activity of Mylotarg in AML. On he other hand, AVE9633 will have to be show at least the same activity and safety profile in order to be approved. This a relatively high bar, and according to preliminary results, chances are pretty low.

AVE9633 entered phase I in 2005, where the compound was dosed once per three weeks at doses up to 260 mg/m2, without encountering dose-limiting toxicities. Since there was no substantial clinical activity, Sanofi-Aventis decided to launch 2 additional phase I trials where AVE9633 is dosed more frequently. Although data is yet to be reported from this trial, the company defines results “encouraging”. Clinical findings from this trial are expected to be presented in ASH 2007 as well. The comparison to Mylotarg is inevitable, since both compounds are ADCs that target CD33. In pre-clinical trials, AVE9633 was found to be more active than Mylotarg, however, a quick glance at the dosing profile of the two agents reveals a staggering difference. Mylotarg is dosed twice at 9 mg/m², with 14 days between the first and the second dose, and achieves impressive clinical response, including 20-30% complete responses. AVE9633 could not achieve an objective response at a single dose of 260 mg/m². What is even more discouraging is the fact that according to several trials, Mylotarg reaches complete saturation of CD33 sites present in the bloodstream and 42% to 90% saturation in the bone marrow at a dose of 9 mg/m². In other words, there is no use to administer additional amount of drug since it has no target to bind. Therefore, unless there is something we are totally missing here, something went very wrong with AVE9633.

Author is long IMGN

The story behind this compound perhaps explains the considerable skepticism surrounding Immunogen’s platform, and should serve as a demonstration for the risks and uncertainties in drug development. huC242-DM4 is comprised of the humanized antibody huC242 , which binds to CanAg, which is expressed on colorectal, pancreatic and gastric cancer cells. The first ADC based on huC242 was huC242-DM1, which is identical to huC242-DM4 with one difference – the linker.

huC242-DM1 entered 2 phase I clinical trials, in which the vast majority of patients had colorectal cancer. Unfortunately, the compound’s performance was rather weak, leading to some minor responses but no objective responses (50% regression in tumor burden lasting at least 6 weeks). There were two possible explanations for the disappointing activity. First, it is well known that colorectal cancer is relatively resistant to antimicrotubule agents, such as paclitaxel (Taxol^®) and maytansine, (which is the drug Immunogen uses for its compounds). Second, the compound had a rather poor stability in the bloodstream, which led to lower drug concentrations shortly after administration. In parallel to the phase I, the company started evaluating ADCs based on the same antibody, but with a different linker - huC242-DM4. These evaluations revealed that the substitution of the linker led to an increase of more than 2-fold in the ADC’s stability in comparison to huC242-DM1.

Subsequently, on October 2004, the company announced that it had decided to replace huC242-DM1 by huC242-DM4. That meant, of course, a new phase I trial must take place, in order to assess the safety of the new compound. The phase I clinical trial was initiated in mid 2005, with results announced in June of 2007. In this trial, the vast majority of patients were also advanced-stage colorectal cancer patients, so the chances didn’t look too good in the first place. Although there were no objectives responses, huC242-DM4 indeed proved to be more stable than the earlier version of this ADC. It was quite clear that huC242-DM4 will not be effective against colorectal cancer, so the company decided to assess the compound’s activity for the treatment of gastric cancer, as approximately half of all gastric cancer tumors express CanAg. The company recently initiated a phase II trial in advanced gastric cancer patients, who have failed front-line chemotherapy. During the second half of the 20^th century, the incidence of gastric cancer has dramatically decreased in developed countries, but it is still a leading cause of death, responsible for over 11,000 deaths annually in the US alone. Just like many other types of cancer, gastric cancer is curable in its early stages, however, unfortunately, in many cases it is diagnosed in its advanced stages, where the only option is chemotherapy treatments. The company takes pride in the fact that gastric cancer has been found to be highly sensitive to huC242-DM4 in preclinical studies, but we also know that was the case for several colorectal cancer cell lines. Fortunately, last year, the antimicrotubule agent, docetaxel (Taxotere^®) was approved by the FDA for the treatment of gastric cancer in combination with other drugs. Additional clinical trials, including trials published this year, also support the use of Taxotere for the treatment of gastric cancer. This is in striking contrast to colorectal cancer, for which antimicrotubule agents are not approved. In addition, both Taxotere and Taxol demonstrated clinical activity in gastric cancer as single agents. These are very encouraging news, since this time, huC242-DM4 actually stands a chance of showing some sort of clinical response, especially in light of the higher potency of DM4 relatively to Taxotere. Nevertheless, although Taxotere was the first FDA approved drug to demonstrate a survival advantage in gastric cancer in more than a decade, it did not improve survival dramatically. The fact that patients enrolled to the phase II trial had failed at least one chemotherapy treatment makes it even more challenging.

In order to get a reliable assessment of huC242-DM4’s efficacy with minimal allocation of resources, Immunogen will first evaluate only 17 patients, hoping that there will be at least one objective response (50% regression in tumor size lasting at least 6 weeks). Statistically, if no clinical response is observed among these 17 patients, it is likely that the drug won’t be very effective for gastric cancer. That means that interim results from this study may be published in the coming months. In our opinion, this trial is the most important one for Immunogen, as even one partial response can open the door for recruitment of additional patients and possibly for a phase III trial already in 2009. The market expectations regarding this compound are pretty low, which will make any positive indication an extremely positive surprise. Bearing in mind this compound is fully owned by Immunogen, clinical success in the huC242-DM4 front, will lead to substantial appreciation in the stock price.

There is no doubt that Immunogen’s most high-profiled candidate is Herceptin-DM1. Its development started back in 2000, as it looked like the perfect candidate for ADC development. Herceptin is an approved block-buster antibody for the treatment of breast cancer. It recognizes and binds the Her2/ErbB2 receptor, which has been strongly validated as a specific and efficient for the targeting of breast cancer cells. In addition, the superiority of Herceptin-DM1 can be easily demonstrated by administering it to patients who do not respond to Herceptin. If Herceptin-DM1 demonstrates a clinical effect among these “Herceptin resistant” patients, it may be the ultimate proof of concept for Immunogen’s platform.

Herceptin-DM1 has recently entered a phase II clinical trial, following a phase I trial which was launched in April 2006. The study was designed to evaluate Herceptin-DM1 among patients who had initially responded to Herceptin but then relapsed. In other words, in the beginning, these patient benefited from Herceptin but for some reason, their tumors eventually became resistant to Herceptin. It is important to note that the tumors which became resistant to Herceptin were still recognized by the antibody despite being unaffected by its binding. This leads us to one of the major advantages immunoconjugates have over “naked” antibodies.

One of the biggest problems with developing antibodies for cancer is that most of them have no therapeutic effect by themselves, even if they target and bind cancer cells specifically. As a result, for every Herceptin out there, there are dozens, if not hundreds of antibodies who can bind and recognize specific antigens on cancer cells, but have no therapeutic use. Attaching drugs to those antibodies may turn many of them into very potent agents, regardless of their inactivity as naked antibodies. All of the sudden, those “useless” antibodies turn into an infinite pool of potential drugs, so there is no need to develop new antibodies. That is why, in our opinion, technologies such as Immunogen’s do not only advance the field of cancer antibodies toward safer and more effective treatments, but actually represent a true revolution, as it shortens time-to-market of potential drugs. Since there are so many well studied antibodies that have already been developed and characterized, the resources and time required for bringing such immunoconjugates to the clinic become dramatically lower.

Back to our Herceptin-DM1, Genentech started the phase I in April 2006 and disclosed initial results in December 2006, followed by more updates in 2007. The goal of phase I trials is primarily to evaluate whether the drug is safe, by examining side effects and several doses. When dealing with a potent agent such as DM1, there is always a chance for something to go wrong, but luckily, Herceptin-DM1 did not have any special side effects, even in relatively high doses. In this specific phase I trial, the maximum tolerated dose (MTD) was set at 3.6 mg/kg every three weeks. At that dose level, 5 out of 15 patients who received Herceptin-DM1 had a partial response (tumor load went down at least 50% and stayed in remission for at least 6 weeks). Actually, another patient who received a lower dose of 2.4 mg/kg also had a partial response, so if we include her as well, the objective response rate for the two doses is 37.5%.

Since the patients had stopped responding to naked Herceptin, we can safely assume that the broad clinical effect was a result of the potent DM1 infiltrating into the cancer cells following the binding of the Herceptin moiety. Based on historical figures, the overall amount of DM1 that was used in this trial could not have led to this clinical response on its own, so we know the DM1 was targeted to tumors specifically.

In addition, Herceptin-DM1 had a favorable safety profile, since a clinical effect was achieved with relatively minor side effects. This implies the linker which glue the antibody to the drug is fairly stable in the blood stream.

Finally, the patients who participated in this trial had undergone numerous chemotherapy treatments and saw their cancer relapse prior to receiving Herceptin-DM1. Thus, their disease was in a very advanced and aggressive mode. It is likely that the drug would be more effective in earlier disease stages and less pretreated patient population. Although we view these results as very positive, we must admit we had expected a more potent activity by this promising candidate.

The reason for our extremely high expectations was the high potency of DM1 and promising pre-clinical results of Herceptin-DM1. When Herceptin-DM1 and Herceptin were evaluated for the treatment of mice bearing Herceptin-resistant tumors, Herceptin alone had no effect on tumor growth, while Herceptin-DM1 caused >90% tumor reduction in all mice examined. In addition, earlier pre-clinical studies showed that DM1 conjugates were +1000-fold more potent than naked Herceptin against breast cancer cell lines. Herceptin is dosed at 2 mg/kg per week, while in the phase I trial, T-DM1 was injected every 3 weeks at a dose of 3.6 mg/kg. This puts Herceptin-DM1’s clinical activity in a somewhat unflattering light.

Make no mistake, Herceptin-DM1 still has very promising prospects as it managed to show a significant effect where all other alternatives failed. Obviously, small phase I trials’ results are not reliable for deciding whether a drug candidate is effective or not. Still, the fact that the candidate had an effect among 37.5% of Herceptin-resistant patients is an extremely important indication. Although not as bright as we had expected, Herceptin-DM1 is still the jewel in the crown of the cancer antibodies field.

Immunogen realized that the critical point in developing ADCs is the linker that conjugates the antibody to the chemo drug. There are plenty of antibodies that target cancer cells specifically as well as plenty of effective chemo agents, but the biggest challenge is gluing them together. Most importantly, for the drug to be safe, the linker must be very stable inside the blood stream. Then, in order for it to be effective, the drug must be released in its active form once inside the cancer cell. It might sound simple in theory but doing it properly is anything but simple, and for that reason, we believe that Immunogen’s proprietary technology is very appealing.

Unlike other groups who develop ADCs, Immunogen does not use traditional chemotherapy drugs such as Taxol and Doxorubicin, but decided to use derivatives of maytansine, which is a much more powerful agent. Because maytansine is such a toxic compound, it cannot be used “as is” due to unbearable side effects. However, when properly linked to an antibody, such a deadly compound may even do a better job killing cancer cells. The derivatives used by Immunogen are patented and can be linked to antibodies via various mechanisms, which should be tailored for each case individually.

Investors should take note of the company’s attractive business model, where it develops its own agents on top of licensing its technology to other companies. In an industry where most products fail, it is crucial for any drug development company to be involved in as many projects as possible in order to reduce the risks involved. Immunogen is engaged in the development of its own ADCs and currently has 2 such agents in clinical trials. In addition, they license their products and technology to other companies such as Genentech (DNA) and Sanofi-Aventis (SNY). Such partnerships usually involve milestone payments of several tens of millions per product, manufacturing and R&D costs coverage and royalties from future sales. A small company such as Immunogen is limited in the resources it can allocate to each clinical program. Therefore, we see such collaborations as an ideal strategy, which enables the development of a large number of candidates based on Immunogen’s technology, with all the development and clinical costs being covered by large partners. Furthermore, the milestone and R&D payments it receives from its partners decreases the need for additional fund raising and dilution, throughout the clinical evaluation process.

Owing to its strategy, Immunogen has an impressive pipeline, with five candidates in clinical development and many more in pre-clinical stages. Notably, only two of the five candidates are exclusively owned and developed by Immunogen (huC242-DM4 and huN901-DM1) with the rest developed and financed by its partners (Herceptin-DM1, in development by Genentech; AVE9633 and AVE1642, in development by Sanofi-Aventis).

Author is long IMGN

In addition to the five candidates, Immunogen expects to have 1-3 additional compounds in clinical stages by June of 2008. Potential candidates may be compounds which are being developed in partnership with Biogen Idec, Centocor, and of course Genentech and Sanofi-Aventis. 2 promising compounds that have demonstrated great pre-clinical results have been discussed in recent scientific publications. The first is an anti CD79 ADC, based on immunogen’s DM1, which is developed by Genentech for the huge market of NHL treatment. Interestingly, the DM1-based compound was compared head to head with a compound based on Seattle Genetics’ (SGEN) technology and there were no substantial differences between the two compounds regarding their activity in mice. It remains to be seen, whether Genentech decides to advance any of the compounds into clinical trials. The second article was published by Centocor and evaluated 3 different ADCs based on Immunogen’s technology that target ανintegrins, which are present on several tumors including lung cancer and melanoma. One of these ADCs, CNTO 365, showed the most promising results.

In summary, we believe that not only does Immunogen have an impressive and diverse pipeline, they also have an attractive platform for developing new ADCs, either independently or via its partnerships. This platform will hopefully lead to a constant flow of candidates into the clinic in the coming years, without putting too much pressure on the company’s expense line. Even after acknowledging that statistically, the majority of its evaluations will probably fail, we view Immunogen as a very attractive long-term play.

Author is long IMGN

Metaphorically, antibodies can be described as unarmed guided missiles, which have extraordinary precision and targeting abilities, but once they hit the target, they inflict minimal damage. Chemotherapy and radiotherapy can be described as artillery, very powerful, but unguided. In order to optimally use the two, the most logical step is arming those unarmed missiles with a variety of explosives. Using the same reasoning, there is a true need to develop anti-cancer therapies which have an antibody-like specificity as well as chemo/radio-therapy-like potency. Doing so enables us to take advantage of the selectivity of antibodies and the potent toxic activity of chemo/radio-therapy, thus creating superior cancer treatments. The antibody binds the target on the tumor, delivers its payload and kills the cell. Arming antibodies with effector molecules like chemotherapy agents and radio-isotopes results in a hybrid agent referred to as an Immunoconjugate. An antibody which is not conjugated to an effector is referred to as “naked” antibody.

  (more…)

Antibodies are naturally occurring proteins that help our body fight bacteria, viruses and cancer. Upon its release into the blood stream, an antibody can identify and bind a specific pathogen, and by doing so, it can neutralize the pathogen or “flag” it for attack by the immune system. Our body is able to generate antibodies against a virtually infinite number of targets, thanks to brilliant biological mechanisms developed throughout the course of evolution. Once a pathogen enters our body, our immune system conducts a high-throughput screen of all the antibodies it can generate (around 10 billion different antibodies). After the appropriate antibodies are selected, special white blood cells called B-cell lymphocytes enter a mass production phase in which large amounts of the selected antibodies are secreted into the blood stream “in search” of their specific target. The combination of diversity on the one hand and specificity on the other hand, makes antibodies a crucial component of our defense mechanism. This combination also makes antibodies an extremely popular platform among drug companies.

  (more…)