Axial - Observations #2
Life sciences reflections
Every Friday morning (only if there is something unique to write about) a set of ideas and observations from a week’s worth of work analyzing businesses and technologies.
I was listening to quite a few Roy Vagelos talks - he talks about Merck and drug development, but one topic he’s been harping on for at least a decade is drug sales and the Internet. Most biopharma companies have large sales and marketing teams annually spending well over $20B. The premise is how the Internet can disintermediate this market by directly selling medicines to physicians. The Internet has been around for a few decades and nothing has really changed about the industry practices.
Excitedly, there are some early signals and business that allude to a much larger opportunity for a new company. With companies such as Hims, Roman, Lemonaid Health building viable businesses around direct-to-consumer (DTC) advertising through the Internet for generics creates an initial playbook to sell new medicines. DTC advertising for drugs has been around for over 3 decades - first made legal in the US in 1985 and taking off in 1997 once the FDA made less restrictions on detailed side-effects for TV commercials. Spending on drug advertisements at the time were in the $100Ms with the market now worth well over $20B. With a major shift of selling products in general online in particular on Instagram, there are quite a few examples of selling medicines through existing social networks. Overall, this is a massive opportunity to create new business models in life sciences and could help improve transparency and reduce costs for new medicines.
Right now, pricing in healthcare is as opaque as you can get. Healthcare in the US at least is not a free market. There are only two payors - insurers and biopharma. In between new inventions and patients are a lot of middlemen who create some value but disproportionally take too much. A great piece to read on the history of the field is from Julie Donohue at the University of Pittsburgh -https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2690298/ The piece sets up a tension between supporters who argue that advertisements help patients make better decisions and bring more transparency while critics fear that biopharma while manipulate their ads to sell bad drugs. For drug advertisements, there are three major types:
Help-seeking - just describing a disease without naming a drug (i.e. marketing how many people have ED or herpes to compel a patient to look into treatments)
Product claim - drug brand name along with risks/benefits
Reminders - just the brand name (i.e. Viagra)
Product claims and reminders are pretty heavily regulated by for the FDA and FTC - https://www.goodrx.com/blog/prescription-drug-advertising-regulation-united-states/ Whereas help-seeking ads are not regulated by the FDA and can be on Instagram or Twitter. However, Instagram has banned paid marketing for pharmaceuticals. So companies have been going through various loopholes to comply.
Advertising medicines directly to patients is truly an American thing and New Zealand too - the only two countries that allow drug ads. Given how mature the Internet is and the massive overhead costs of having massive sales teams in biopharma, there is going to be more and more interesting opportunities to reinvent how medicines, both new and old, are sold to patients.
Climate change and biology
While getting breakfast with some buddies we ended up talking about solutions to climate change. I believe most of the companies/products promoted to mitigate any potential future problems with our environment are superficial. We ended up talking about where biology can make an impact. We discussed materials and energy, but I’m not an expert enough to do much in these fields. I still have trouble building out a complete framework to understand exactly where biology can have a major impact on our environment. It’s a lot easier to think through how a new discovery in say neuroscience can lead to a better medicine for AD or PD. However, it’s much harder to figure out how a breakthrough in say non-canonical amino acid production in microbes will control emissions. Roughly, after breakfast and walking through a botanical garden, there are three major parts in the problem:
Ecosystem - general human organization and how it impacts the environment so companies like Impossible Foods reduces the land use of cows and Pivot Bio reduces nitrogen runoffs on farms
Atmosphere - general emissions of mainly carbon - companies like Solugen help make a dent in this problem and hopefully work around CO2 carbon gets successfully translated: https://cen.acs.org/articles/94/i46/Synthetic-biology-pulls-CO2-atmosphere.html
Plastics - successfully degrading plastics and maybe converting them into something useful; this is probably the hardest problem: https://www.findaphd.com/phds/project/new-enzymes-for-degrading-plastics-a-combined-structural-and-synthetic-biology-approach/?p94582
Ultimately, all three of these part are connected to each other through petroleum. This recent post has some interesting data - http://www.bioeconomycapital.com/posts/2019/9/23/seeing-the-end-of-oil Biology can help along with a lot of other new inventions (i.e. energy, materials) have a good shot to make a dent in this transforming our global economy (getting away from the petro-dollar system) and reduce overall carbon emissions from 900-1000 GtC to 700-800 GtC.
History of biotechnology
History is really fun to read. It’s valuable in spots; reading futuristic pieces are pretty important as well. UC Berkeley has an incredible repository of interviews of great individuals in biotechnology - https://bancroft.berkeley.edu/ROHO/projects/biosci/oh_list.html - from Paul Berg, Arthur Kornberg to David Goeddel, Herb Boyer to the team at Chiron.
Lupus (systemic lupus erythematosus - SLE) has a personal place in my life - a close family friend’s mother suffered from this horrible disease. The disease is driven by the production of autoantibodies against a set of antigens that lead to chronic inflammation and eventual organ failure. Unfortunately, only belimumab in 2011 (targeting B lymphocyte stimulator - BLyS) from GSK has been approved for SLE over the last 6 decades or so. Scary stuff for lupus patients and probably autoimmunity patients overall as well. In 1955, hydroxychloroquine and corticosteroids were specifically approved for lupus. With a SLE found in around 70-80 people per 100K, this is no small market. The key driver for this massive failure from biopharma to meet this demand has been from clinical trial design. Better patient recruitment strategies will help as well as bringing a concept that has taken oncology by storm to autoimmunity - treating the genetic drivers instead of just the disease.
Patients with SLE are very diverse - heterogeneity around the affected organs and various subtypes. Many lupus trials will exclude patients with progressive kidney or CNS disease in the hopes that approval becomes more likely. But this leads to exclusion of patients who need better treatments. These drugs end up failing because they are not upfront design to treat severe cases. Lupus nephritis (LN) could be considered a subtype where it is characterized by kidney inflammation. For LN, biopsies are made with a lot of variation that lead to inefficient recruitment and lackluster clinical results. Over the last decade of so there have been various initiatives, conferences, and a lot of fancy looking terms to solve this problem. The FDA finally gave guidance in 2005 for developing medicines for lupus leading to the formation of a variant of methods to stratify patients - such as British Isles Lupus Assessment Group (BILAG) index, SLE Responder (SRI-4) index, and BILAG-Based Composite Lupus Assessment (BICLA). Essentially these indexes categorized the disease activity across organs for a given SLE patient. All of these measures have done a pretty poor job improving clinical trial design. New methods are in the pipeline to emerge to combine new data sets (i.e. genomic) to better match a candidate with SLE and its subtypes and create new endpoints/outcomes to improve the ability to detect the difference between treatment groups and placebo especially over a reasonable time scale and recruitment class. Recently and to great news to lupus patients, three new medicines met their endpoints:
Anifrolumab - AstraZeneca
Cenerimod - Idorsia
Telitacicept - RemeGen (relying on SRI-4)
Unfortunately, lupus still has outstanding problems in trial design that is a limiting factor in bringing cures to highly progressive patients. New modalities like cell therapies and tools like autoantigen discovery platforms will help validate new biomarkers/endpoints and treat SLE as a genetic disease to target bringing curative therapies to patients:
What it would take to make a pet dragon?
The concept of a killer product for synthetic biology has been explored for awhile. This idea to make a pet dragon has been appealing to me for a few years and I know it’s not too original because I’ve spoken to many people about the idea who came to similar conclusions.
The starting point is fairly obvious. The endpoint is not so clear. This means that synthetic biology has a lot of room grow but also that the ability to create truly unique consumer experiences with biology is still in the early days. This is why most synthetic biology companies end up converging to making a new medicine.
The three main components are:
Dragonlike appearance - that’s not that hard
Flight - really hard
Fire breathing - difficult but possible
The first part is driven by the template - either an avian species, a mammal like a bat, maybe a lizard, or even an exist species that could be resurrected like a pterodactyl. The second part, flight, is the hardest part due to maintain a dragonlike appearance while agreeing with the square-cube law. As a result, a bird is probably the best path forward maybe a bat. However, this means the dragon would probably have a beak and some feathers. I haven’t looked deeply into the genetic linkage for beak/feather phenotypes but they seemed tightly coupled to the overall fitness of a bird. For fire, the dragon would need oxygen, a fuel like methane or propane (and accessories), and an igniter. For an avian species, they suck up oxygen through their air sacs and a fuel could be over-expressed through genetically encoding methods or channeled differently from the gut to the sacs. The igniter is the hard part - maybe just doing surgery on each dragon to insert one is the easiest way to accomplish fire-breathing. Thinking through how to make dragons with synthetic biology is not only fun but important to understand the upper limits of the toolkit and drive the march forward toward truly unique experiences with biology.
Cell therapy manufacturing - medical centers versus industry
There was a recent article around solving the bottleneck of cell therapy manufacturing - https://www.statnews.com/2019/11/20/car-t-drugs-academic-medical-centers-save-billions/ - it had one interesting concept of bringing the capability closer to patients. This should reduce costs, improve efficacy, and hopefully reduce variance. However, the first two are easier to think through, but the article did not acknowledge the benefits of biopharma controlling process - quality control. Academic/medical centers at a scale beyond say 10 patients have a really bad track record of bringing new medicines to market at scale. The field right now is caught in a weird spot - the data being release for CAR-T and beyond are really exciting for curative therapies. But the costs are often prohibitive, responses not clear, and not enough of the product can get made. This is an obvious problem where new technologies can provide tremendous amount of value and support the creation of new businesses to scale up cell therapies and bring them closer to patients without sacrificing quality. New companies are emerging, it’s an exciting time to not only design cell therapies but manufacture them; think oil and railroads in the last 19th century, like Indee Labs and more.