Seattle Genetics’ technology – The Arms Merchant

The capability of developing antibodies for cancer can be found at most pharma companies’ R&D centers, either as a result of internal R&D efforts or M&A activity, such as the acquisitions of Cambridge Antibody Technology and Abgenix by AstraZeneca (AZN) and Amgen (AMGN), respectively. Therefore, there is nothing unique about a company that can develop cancer antibodies, even though there are many other differentiating factors between the companies. The crucial element in developing an ADC is linking the antibody to the drug payload. As simple as this concept may sound, its realization is highly complex and challenging, and in our opinion represents the main entry barrier to the field. As ADCs are also termed “armed antibodies”, companies like Seattle Genetics can be viewed as the arms merchants of the antibody industry.

As an arms merchant, the company focuses on two areas: Technologies for conjugating antibodies to toxic drugs and potent toxic compounds that will be attached to the antibodies. The ability to develop highly potent drugs and conjugation technologies is Seattle Genetics’ main asset, since this is the ideal way to differentiate itself and to broaden the company’s pipeline through partnership deals. In an industry where the vast majority of candidates fail, it is imperative for companies like Seattle Genetics to have as many candidates as possible, even if eventually most of the revenues go to the partners. At this stage, with the limited resources Seattle Genetics has, betting on few wholly owned candidates is statistically unfeasible. Although the company has had its share of failures over the years, we believe the advances made both in terms of linkers and drugs will finally enable it to generate a constant flow of candidates into the clinic, whether independently or in collaboration with partners. In order to look at the progress that has been made so far, the best place to start is the failure of Seattle Genetics’ flagship product, SGN-15, an antibody linked to the chemo agent Doxorubicin, whose development was discontinued in mid 2005 after a series of discouraging clinical trials. On top of the usual uncertainties related to drug development, there were probably two main factors that severely sabotaged this candidate’s prospects.

The first factor was the use of an approved chemotherapy drug such as Doxorubicin as the conjugated drug. Chemotherapy agents that are conventionally administered to patients are distributed across the body and affect healthy cells as well as cancer cells, leading to the so typical side effects of chemo. Consequently, approved chemo drugs represent a fine balance between two needs: They must be strong enough in order to kill cancer cells, but not too strong, so the damage caused to normal tissues is acceptable. In contrast, when chemo drugs are linked to an antibody, they can be targeted to tumors specifically, since the antibody guides them. This enables the use of much more potent drugs, otherwise impossible to use in conventional administration. Furthermore, since only a small fraction of the administered antibodies eventually accumulate in cancer cells, it is critical that the few antibodies that do reach the tumors carry a very potent payload. This can be accomplished by two approaches: The antibody must either be loaded with a large amount of drug molecules or a small amount of very potent drug molecules. Although there are efforts on both fronts, the latter approach is more practical, at least for now. Bottom line, in order to have an effective ADC, drug developers should use chemo drugs that are too toxic to be generally administered. This approach was validated by the only FDA-approved ADC, Mylotarg, which utilizes Calicheamicin, a drug that is too toxic on a stand alone basis. Both Seattle Genetics and Immunogen (IMGN) are currently using such compounds as the basis for their ADC platforms: Seattle Genetics picked auristatin, while Immunogen focuses on maytansine. The second disadvantage in SGN-15 is linker instability. An ideal linker should be very stable in the bloodstream but also readily degradable once inside cancer cells, so it would release the free drug only inside target cells. For SGN-15, Seattle Genetics uses an acid-labile linker, which is relatively stable in neutral environment (bloodstream) and very unstable in acidic environment (present in certain compartments inside cells). This kind of linker is used very successfully in Mylotarg for the treatment of acute myelogenous leukemia [AML], making Seattle Genetics’ pick very reasonable at the time. However, SGN-15’s stability in patients proved to be pretty low, mainly as a result of premature linker degradation in the bloodstream, before reaching the tumors. Mylotarg had a great success despite being based on an acid-labile linker because it attacks a blood-borne malignancy and the antibody can find its target quickly, before linker degradation and drug release. In contrast, the dense mass of solid tumors makes them far less accessible compared to blood cancers. Therefore, the ADC must be present in the bloodstream for longer periods at higher concentrations, necessitating highly stable linkers.

By the time SGN-15 was scrapped, Seattle Genetics already had its next generation of ADC technology up and running. On the drug front, the company licensed a potent drug called auristatin E from Arizona State University, which was found to be almost 200-fold more potent than Doxorubicin, and used it as a basis for its own proprietary drug, MMAE. This drug is a very potent anti-tubulin inhibitor that can be synthesized cheaply in very large quantities and subsequently be conjugated to a virtually unlimited number of different antibodies. Another appealing attribute of Seattle Genetics’ conjugation technology is the highly homogeneous population of ADCs, as oppose to other methods, including that of Immunogen. On the linker front, Seattle Genetics chose a peptide-based linker which is cleaved by enzymes that are present inside cells but not in the bloodstream. Upon cancer cell binding, ADCs are trafficked to a special compartment called lysosome, where there is an abundance of enzymes that cleave the linker and release the drug inside the cell. Seattle Genetics’ peptide linker has demonstrated an increase of more than 3-fold in stability in the bloodstream, which, combined with the high potency of MMAE, puts the company’s candidates in a better starting point.

It is crucial to understand that ADCs are not commodity products, but highly complex systems that require a great deal of customization and optimization. Multiple factors, including (but not limited to) cancer type, the target on cancer cells, exact binding site, type of linker, efficiency of drug release, mechanism of conjugation, type of drug and amount of drug payload affect the performance of each candidate. The number of variations for each ADC is high but it is impossible to predict the optimal combination in advance. Thus, the exact antibody-linker-drug combination should be tailored specifically for each ADC candidate, perhaps even for each condition the candidate is aimed at treating. In order to stay relevant, Seattle Genetics must constantly develop new linkers and drugs, in addition to developing antibodies and identifying attractive cancer related targets. It is not surprising though, that the company is currently developing next generation linkers and drugs that will possibly be employed in future projects.

Author is long SGEN

2 thoughts on “Seattle Genetics’ technology – The Arms Merchant

  1. Ohad,

    You raise an interesting point that I was not aware of, that SGEN have more homogenous ADC populations than IMGN.

    You say SGEN use peptide-linkers, while IMGN use non-peptide linkers, the disulfide linker (SPDB-DMx) and the thioester linker (SMCC-DMx).

    I’m not the best chemist, but are you saying that, IMGN attach the linker to any-old amine residue (lysine) on the MAb, as shown in the Figure below:

    One other thing, you mention that DNA appear to prefer uncleavable linkers, once again I’m not the best chemist, but which of the above linkers are cleavable and which are uncleavable.

    Herceptin-DM1 seems to use the thioester linker (SMCC-DMx).

    Lastly, in your very informative articles you mention you are long IMGN, but you do not mention being long SGEN. I’m in the same boat, but wondered the reasons for your preference.


  2. Hi Stephen

    I will try to address the important issues you brought up:

    Regarding the homogenicity issue, obviously the more homogenous a drug is the better, but I am not sure how crucial it is when comparing SGEN’s ADC technology to that of IMGN’s. I think that eventually both companies tend to use around 4 drug molecules per antibody. In fact, I believe SGEN came to the conclusion that an ADC with 4 drug molecules has a better therapeutic window than and ADC armed with 8 drug molecules.

    Regarding the different linkers, u said you are not the best of chemist, well, that makes two of us ;).
    To the best of my knowledge, IMGN cannot fully control to which lysine residues the drug is conjugated, but i m no expert on this front.
    Although the two ADCs it has in the clinic utilizes the vc peptide linker, SGEN also uses non-peptide linkers, similiar to the SMCC (thioether) linker used by IMGN.

    thioether linkers are not cleaveable. The disulfide linkers are cleaved by the reducing environment in the lysosome while the peptide based linkers are cleaved by lysosomal proteases, mainly cathepsin B.
    From what I see, Genentech prefers thioether linkers (which are noncleavable, of course) from both companies, but I could be wrong.

    I hold both SGEN and IMGN believe both have a bright future ahead of them, although there will surely be many frustrating failures down the road. Right now, I am more excited about SGEN’s pipeline as a whole, but Herceptin-DM1 is still the most important ADC out there, imo.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s