By Philip Astley-Sparke, CEO of Replimune, as part of the From The Trenches feature of LifeSciVC.
Manufacturing of medicines based on biology is complex. Viral vector manufacturing to produce gene therapy constructs, vaccines and oncolytic viruses is especially intricate. Most early stage companies use contract manufacturing organizations until real clinical proof of principle is achieved as a logical decision inflection point for investing in in-house manufacturing, versus continuing to use a CMO for pivotal studies and launch.
Those developing complex viral vector products do not have that luxury and should secure control over production early in the development cycle.
My manager at Arthur Andersen, where I was first employed, used to tell me “Do something once you’re an expert, do it twice you are an industry leader.” During my career, I have championed and then mandated the construction of three viral vector manufacturing facilities so I wonder what category he might put me in. Of course, this statement is a bit self-promotional and doesn’t align with my British self-deprecating disposition, but I am writing this post in the USA where even that little sandwich shop around the corner from our office declares its food to be “world famous.”
What I can say is the conscious choice to bring our manufacturing capability in-house, first at BioVex, then at uniQure and ultimately now with Replimune, have born some important learnings and insights as others consider the right choice to make.
It was back in the early 2000’s after having joined the company, BioVex, that developed the only FDA-approved Oncolytic virus, TVEC, that I first bore witness to some of the challenges presented by using contract manufacturers to produce complex biologics and, yes, the simple solution often was to change providers, but we ultimately would run into similar issues. There are inevitable mishaps that occur (of course, not to say this can’t happen in our own facility) but in my experience, when they happened, the downstream effects were exponential. Clinical batches failed for asinine reasons like the manager went on holiday, pipe A inserted into pipe D rather than C, chromatography columns allowed to dry out, etc. One batch failed because an operator cut themselves and dripped blood in a batch. I believe at one site (nothing to do with us) in Europe, someone once even fell into a fermenter – that’s a pretty large contaminant. The truth is, a batch fails for any number of reasons leading to the need to produce another batch, however, being one of many clients for the contract manufacturer meant it could take six months to book another production slot. For early stage companies, timelines matter. Funding is tied to timelines and milestones being met. We operate within a finite window to achieve proof of concept or progress through a clinical trial. Being marred by delays in the manufacturing process can mean failure for many companies. Ultimately, during this BioVex experience, we lost 18 months, a whole funding cycle, and the issue almost bankrupted the company. Further, it left us questioning whether we had confidence that our contractor would be able to deliver three consistency batches in a row so we could move forward at some future point with a Biologics License Application.
To progress into pivotal studies only one solution remained – build. In 2005, BioVex relocated from the United Kingdom to the U.S., and we built out a manufacturing facility in Woburn, Mass in about a year with an initial investment of approximately 5 million dollars. We chose a geography that had a large pool of talent with the skillset we required. Even today, after BioVex was acquired by Amgen in 2011, all commercial supplies of TVEC are produced at the facility.
- Lesson Learned:
- Having destiny in your own hands and controlling the process isn’t just about eliminating errors.
- Owning your own manufacturing facilitates the growth of in-house knowledge and expertise to increase yields and perfect release specifications.
My story doesn’t end there. After BioVex was acquired by Amgen, I was given the opportunity to assist uniQure, a leading gene therapy company, with its move from Europe to the U.S., the company’s move to the public markets and listing on NASDAQ and establishment of its U.S. infrastructure. Based on the prior experience with BioVex, I had a mantra which I repeated at each and every IPO meeting – “the process is the product” when it comes to biologics. I think I may have shared enough about my prior experiences and repeated that mantra so frequently that we collectively made the decision to spend proceeds from the IPO on establishing in-house manufacturing capabilities to avoid all the pitfalls I knew so well and to learn to grow at scale. Back in 2013, this raised some eyebrows as the cost of construction had risen to 20 million dollars combined with the overhead burn. We searched for sites in Lexington and ultimately told the broker, “I want this site,” which may have been tied in part to the Irish Pub I had eyed across the road. uniQure was a pioneer, becoming the first gene therapy company to construct a facility. I caught some flak for the overhead the manufacturing facility created when the company, in its darkest days traded at $4, however, we pulled through and the decision was vindicated and since repeated by many to the benefit of hopefully millions of patients. 
- Lesson Learned:
- Manufacturing replication-competent oncolytic viruses is one thing, manufacturing replication-deficient vectors for gene therapy adds a whole new layer of complexity.
- Gene therapy manufacturing at scale is complex and control over the whole process is not only necessary but absolutely essential from the get-go.
So, this brings me to today, manufacturing facility lessons learned number three. After we founded Replimune in 2015, we did so with the knowledge gathered from our prior experience and a commitment to build a manufacturing facility for the company. Not only would it allow us control over the process, but we knew it was a sound investment given our oncolytic virus approach was de-risked to some extent based on the prior clinical success at BioVex and with TVEC. Of course, we know what has transpired in the Boston commercial real estate market given the explosive growth in the biotech industry, leaving the I-28 corridor from Waltham to Woburn crowded and making the search for a suitable site that much more difficult. Our search led us to Andover, but we were confronted with a landlord who was not accustomed to life science companies and struggled with the risk involved in turning their warehouse into a shiny new-age starship. As luck (and strong relationships) would have it, the landlord of uniQure’s Lexington facility had just purchased land in Framingham next to where Sanofi-Genzyme has about a million square feet of biologic manufacturing space. Given our prior relationship, we knew how the other operated and were able to strike a deal in record time just before our IPO. Not surprisingly, resumes flooded in given the skillset of people who had built their lives west of Boston. The facility, now fully constructed, is state of the art and provides us with the ability to grow and scale our manufacturing to produce oncolytic immunotherapies that we believe will transform patient’s lives just like the facilities in Woburn and Lexington already have.
- Lesson Learned:
- Relationships matter, even with your landlord.
- Once again, geographic choices that enable talent sourcing is key.
I doubt there will be a fourth…that would just be silly right?
Thanks to Pamela Esposito and Arleen Goldenberg for editorial comments on drafts of this post.