The Borlaug Report #3: Bioprocessing

The Secular Shift in How Drugs are Made (Part 1)

This article is about bioprocessing, the process behind advanced drug production. Several large companies make the equipment and consumables that go into next-generation drugs, getting paid for manufacturing volumes regardless of if a single drug is successful or not. At the right price, they all have the potential to be interesting investments. 

While anyone can appreciate the excitement of a 1-day pop on a drug approval or clinical success, the magnitude of opportunity available to investors without taking binary clinical trial risk via the picks and shovels here is too great to ignore.  

Recap: Why Healthcare is a Secular Growth Industry

Healthcare is a secular growth trend driven by the following factors: 

• People are living longer - the global population aged 60+ is set to double to 2B by 2050 (2.5% CAGR). Even if the broader population growth slowed, this population will not. 

• As people grow older, they require more medical care / pharmaceuticals. 

• Obesity has been on the rise and nutrition on the decline for decades and causes people to require chronic medical care more early in life.

• Industry is highly regulated, preventing competition from creating significant cost deflation for innovative care. 

Together, this allows the total demand for healthcare to grow a few points above GDP on its own, from a base that is already high. Most of the annual US GDP spend is for services (hospital stays, surgery, your Jell-O after the surgery, etc), and ~20% is drugs, and another ~6% is the profit insurers make. Generally, the volume of drugs grows 3-5% a year - faster than most other industries and faster than GDP spend. 

What is a “Biologic”? 

Most scientists say “biologics” refers only to antibody drugs (which have been around and popular for a decade), and that fancier things like gene therapies get their own category, such as “advanced therapies”, while investors generally just say “biologics” an all-encompassing umbrella term.  

Let’s assume that all large molecules are “biologics” for now, because they all require bioprocessing as a method of manufacturing, which we’ll get to later. This includes monoclonal antibodies (“mabs”), gene therapy, cell therapy, mRNA drugs/vaccines, RNA drugs, etc. 

R&D Trends: The Shift to Large Molecules in Slow-Motion

IQVIA data (given in the company’s annual R&D report) has shown the shift toward biologics in slow motion over the years. In oncology for example, 2012 still was majority-owned by small-molecule. Below is a chart showing the breakdown of biopharma pipeline candidates from phase 1 to submission. “Targeted biotech” is composed mostly of antibodies. Interestingly, small molecules have still grown, just not as quickly as their biologic counterparts. 

Oncology is the most-funded disease research area today, but another large-and-growing one is neurology. This is where a lot of the longer-term growth can come from - historically, small molecules have dominated neurology drug research because larger molecules can’t pass the blood-brain barrier (BBB). Lots of novel approaches have made it so that this may not be the case in the future, such as temporarily weakening the BBB or delivering a drug directly past it via cannula.  Below is the neuro pipeline from 2022: 

Almost 700 candidates, roughly 500 still in small-mol land. If this transitioned to look more like oncology over time, heading to 50%, it would add ~6% to the total number of biologic drugs in the pipeline, a number that itself is already growing faster than total biologic pipeline growth (8% CAGR 2017-2022). With Eli Lilly’s recent data on Alzheimer’s, it looks like the first big Alzheimer’s drug will indeed be an antibody drug, which could accelerate this shift. 

As an aside, our understanding of the brain from a genomic standpoint has inflected since the adoption of single-cell, and more recently, spatial genomics. This could drive the above shift as well. 

Background - Small Molecules and How Almost Every Drug Was Made 15 Years Ago

A small molecule drug is basically any of the pills in your average household medicine cabinet - they are chemicals, synthesized and pressed into a solid tablet or coated in dissolvable pill plastic. Most of the drugs that dit this description are very popular - meaning lots of them need to be made. This has traditionally been done in a stainless-steel fermenter - such as the one below:  

The process for making drugs this way, in large batches, is fairly simple logistically - you are combining active pharmaceutical ingredients and blending them to create your drug, adding things like excipients along the way as you remove moisture, mill, and blend some more. Finally, you remove everything, press the substance into pills, and coat the pill as needed. Run complete! 

Before the next run, one must thoroughly sterilize these large steel tanks with cleaning chemicals. Logistically, this is a perfectly acceptable way of manufacturing chemicals at scale, and versions of this have been done for decades. 

However, things are changing. The future of the pharmaceutical industry is looking increasingly biological in nature, and producing biologics requires a little more TLC (and money). 

How Larger Molecules are Made

While most of small molecule manufacturing can be done with just a couple of discrete pieces of equipment, making a biologic drug has a plethora of steps that can be divided into upstream and downstream bioprocessing. I would add a third category to make clear that parts of “upstream” are generally for R&D purposes only.    

R&D (Scale-Up): Before biologic drugs are commercially manufactured, there is a manufacturing component - scientists have to figure out how to ensure the drug can be scalably manufactured without compromising the effectiveness or safety profile of the drug itself. This is usually called “scale up”. The process is basically a guess-and-check exercise of finding cells that excel in a smaller (150mL) bioreactor, and keep finding the best cells as they multiply and test them in larger and larger bioreactors until you’re up to 1-2,000 liters or more. Other conditions, like what goes into the cell culture media, how much oxygen is let in, temperature, stirring speeds, etc. are all tinkered with here. Once the “process” is defined, R&D is over and production can proceed. This has become a key value proposition of a lot of contract manufacturers, because it can be extremely hard to do, especially in gene therapy. 

Upstream: Basically, what you are doing in upstream bioprocessing is taking a bunch of cells (the “active ingredient” per se) and putting them in a soup of nutrients (media) that stimulates them to multiply at high rates based on the R&D process you tested out. In doing so, you are getting to a giant vat of soup that has an adequate volume of those cells (the drug) floating inside. In the image below, most of this is “production” in the manufacturing process itself. It’s actually a lot like the stainless steel process up until here, given everything is going into a bioreactor - the reactor is just smaller in this case. 

Downstream: Once you have your cell soup, you engage in the “downstream” half of the process which separates those cells from all of the things that you don’t want in your final product. Once you’ve purified and filtered everything, it goes into a freezer (“cryo-preservation”) and is then shipped elsewhere to be put into the right delivery mechanism (IV bags, syringe vials, etc.) and boxed/packaged - the “fill-finish” process. This is the part that is fundamentally different - in small molecule production, you’re much closer to the finished product when things come out of the bioreactor. In biologics, you are separating the active ingredient a lot more carefully from the other stuff you put in the soup. 

Most of these drugs are made in smaller batches - they often serve more targeted populations of people than some of the small molecule blockbuster drugs of old. The exception here is antibody drugs, which are still finding themselves going after large populations. Cell and gene therapies, however, are a much different story. After all, healthcare was never going to be one-size-fits-all. If you tried to apply the old method of making drugs to this new reality, you’d realize quickly how much time you are spending cleaning the tanks after every run. 

The Single-Use “Innovation” 

As mentioned above, the economics of manufacturing small batches of a drug in a stainless steel tank stops making sense very quickly when you have to shut down the process afterward to follow strict sterilization protocols, using lots of water, chemicals, and energy just to be able to start the process up again using the same equipment. Fortunately, the industry has already adapted by commercializing single-use technology. 

Single-Use Saves Money 

Instead of cleaning out the fermenter every time you use it, you can just line it with a disposable bag made of a fancy polymer that guarantees the same level of sterility. Kind of like using a trash bag instead of washing out the trash can under your sink every time you empty it. The same goes for all of the tubing connecting each subsequent piece of equipment in the workflow, as well as the cartridges, capsule and columns within the machines themselves. After a run is over, downtime can be short - just replace everything and start over. 

Turns out, at lower batch sizes, net of energy/water/sterilization costs, this can actually be significantly cheaper, both on COGS and capital investment. Below is a sample set of those numbers cobbled together by UBS using Pharmaceutical Engineering data:

As you can see, the key savings items in both cases are on labor (cleaning and quality control) and utilities, at the cost of more money spent on consumables. 

Flexibility is Valuable Too

Even at higher batch sizes, the flexibility of single-use is valuable - this is what drove Moderna and Pfizer to use single-use tech in production of the COVID vaccine. While COVID was a special situation, and drug sales probably won’t vary as much as COVID vaccines did from year to year, most new drug manufacturers have to choose between capex and outsourcing to support new drugs.

• If investing in one’s own manufacturing, building a stainless steel facility to support a drug requires a level of conviction that is hard to have, and once the facility is operational, it burdens the PnL in a massive way that is unacceptable to most companies. 

• The CDMOs they outsource to also have to make a similar choice - and CDMOs in the US/Europe are predominantly building in single-use. There is little reason not to as the contract manufacturer, because building out a large stainless steel facility in hopes that a customer needs that much capacity makes little to no sense in the long run, because if you’re wrong, you can still support it using a single-use manufacturing footprint. 

And ESG is built in!

On top of savings and flexibility, companies get the added benefit of ESG improvements, given the lower utility/chemical/water usage. Single-use consumables are made from polymers that are generally burned in a furnace down to a compostable ash after they are used. 

Why it Matters - An Investable Long-Term Trend

So far, we’ve established the following: 

• Healthcare is an area of secular growth

• Clinical trials increasingly study larger, more complex molecules every year on average

• Given smaller patient population sizes, larger molecules will be increasingly manufactured with single-use technology 

You should care because this is an easily investable trend for few key reasons: 

Durable Usage Trends: Manufacturing in biopharma is different from the R&D tools themselves - there is no “fad” factor like you might see in genomics, for example, where researchers will crowd into a hot new space and use the relevant technology until the next thing comes along. These changes can be quick and violent. You know what doesn’t change? The bag you line the bioreactor with and the tubes that connect it to the clarification system. That’s the same regardless of whether someone invents a new gene therapy, a cell therapy, an antibody, or an mRNA drug. 

Companies selling this technology don’t benefit from one type of therapy - they benefit from the complexity of all therapies moving through the clinic. 

Highly recurring revenue with deep moats: Once you file a drug with the FDA, a lot of things get set in stone - one of these things is the manufacturing process and the equipment that goes into it, specified down to the vendor. Recently, companies have been specifying second sources from a second vendor into these filings to deal with supply chain risks, but the fact remains that once something is “spec’d” into the process, it’s painfully difficult to remove or change it. 

This discourages new entrants to the market because the only share you can win is for clinical-scale dosage for new drugs - meaning your initial “TAM” is extremely small. In bioprocessing, scale is a massive barrier to entry and the FDA is a massive barrier to scale. 

Oligopolistic industry: While there are lots of smaller mom-and-pop-shop companies serving the bioprocessing space, the industry itself is dominated by just a handful of players for 3 key reasons: 

• Price is not (really) and object

• Reliable & consistent supply is king 

• Hard to switch once you enter clinical trials 

The calculus is simple - why wouldn’t you pay a premium to the larger players because you know they’ll ensure you get what you need to manufacture, and that the company will still exist in 10 years if your drug works? Even now as drug companies specify backup suppliers, the choice is the same. Furthermore, many of them bundle products, so one-stop-shop economics probably beat or match savings vs. point solutions at individual discounts.

What this gives the few incumbents, however, is pricing power. In addition to volumes, bioprocessing tools companies raise prices 1%+ per year. Price hikes also flexed with inflation last year - in 2022 and 2023 many of these companies raised prices 5%+.

The Players

So who’s playing in this industry? Below are the tools companies, listed by percent revenue exposure to bioprocessing (approximate - some numbers are estimates and/or slightly stale).

• Repligen (RGEN, 100% of sales exposed to bioprocessing)

• Sartorius Stedim Biotech  (SDMHF, 96%)

• Sartorius (Parent of Stedim, SARTF, 80%)

• West Pharma (WST, 42%)

• Stevanato Group (STVN, 30%)

• Danaher (DHR, 27%)

• Avantor (AVTR, 22%) 

• Merck KGaA (MRK.DE, 11%)

• Thermo Fisher (TMO, 10%)

The best part? Effectively all of them are large, profitable public companies. The only one on the chart below that isn’t, Polypus, recently got acquired by Sartorius, which is.

I will cover CDMOs themselves as well as input makers for advanced therapies like cell & gene therapy in subsequent writeups. These companies include Catalent, Lonza, and Wuxi Biologics.

These companies don’t compete as directly as the above list might lead one to think. Here is a list that does a better job of characterizing each section of the workflow and includes other larger players that happen to have some exposure as well: 

In general, the market opportunity is arguably bigger in the downstream part of the workflow, because as mentioned above, there are applications for single-use downstream technology even in the absence of a single-use upstream (bioreactor) process.  

Where are We in This Shift?

Sartorius believes that single-use penetration is hovering around ~35% today. Over time, given the obvious advantages of using it, single use could easily reach 75%+ penetration over time. It is less simple than this, however, as the early steps in the process adopted single-use bags a lot faster than some of the downstream steps did. Overall, according to an Alphawise survey, about 50% of new manufacturing capacity added is for single-use technology.

As a reference point, ~15% of legacy installed capacity is single-use. 

Generally, penetration is more of an output from a different growth driver than the final story - as biologics become more popular, this number naturally rises.

Regardless, the pipeline data shows us we have a long way to go before the biologics wave is over - most of this new capacity is just to supply clinical trials!  

What is Happening Now - COVID, Stocking, R&D, Alzheimer’s, RSV   

2023 is shaping up to be the noisiest year in bioprocessing history. While the secular trends are crystal-clear, in my view, the hangover from COVID, both on the COVID vaccines/drugs themselves and the R&D funding that low rates brought us, has created a lot of debate about what the next couple of years could look like for the industry.  

COVID vax/drugs is a simple concept - a big, 1-time bolus of demand for commercial levels of vaccine and drug substance manufacturing came as a result of the pandemic. What happened next is what we’re working through now. 

The bolus of COVID demand caused all biologics manufacturers to realize that there were risks to their supply chain as a spike in demand of the size we saw in 2021 was unprecedented. This caused them to engage in “stocking”, where companies over-purchase to ensure they have what they need to make non-COVID drugs. 

Furthermore, low rates and interest in biopharma R&D led to a spike in R&D activity more broadly. A lot of questionable biologics science was funded, and some even made it to phase 2 trials. Last year, funding seized up like it did in other sectors as rates rose. 

Consider the fact that the COVID vaccines/drugs use the same equipment that was being stocked by other manufacturers. Now, consider that COVID vaccine volume estimates have come down sharply, in many cases more quickly than expected. On top of that, a lot of demand for clinical trial supply has been going away in real time as the tide goes out and bad science fails to get more funding. That clinical trial supply generally ran through contract manufacturers, who don’t give the tools companies the best picture of what demand will be because they themselves are struggling to predict cancellations. 

All told, there is almost no short-term visibility into how far biologics growth could slow or decline this year. Without knowing how much demand goes away, how can you know when inventories will return to normal levels? This led to an interesting earnings season, where Danaher and Repligen said two very different sets of things on their earnings calls. If you put the mosaic together, the rest of the year looks messy. 

If you are a long-term investor, however, who cares? The trend isn’t going away just because companies overestimated demand and overbought equipment over a pair of years. 

To make things even more interesting, there has also been a stream of good news. Eli Lilly just published extremely promising data for its Alzheimer’s drug, which is a monoclonal antibody. Biogen has a similar one right behind it. To treat about a quarter, just 25%, of the global Alzheimer’s population with its new drug, Eli Lilly will need to employ about 250,000 liters of capacity. This compares to ~3.5M liters of global demand, according to Biotrak. This could mean that volume demand increases at least 6-7% on one approval, and we are getting 2 competitors in the space. There is always a risk that these drugs are ultimately flops on safety, but they are more upside scenarios than base-case estimates for total demand anyway. 

GSK also just got approval for its RSV vaccine. Both of this and the Alzheimer’s drugs serve massive patient populations, are biologic in nature, and are the first of their kind to market (in a meaningful way - Aduhelm has not made much a splash). They are both likely to use single-use equipment at many steps of their manufacturing process. It is not clear, however, if they will be made in stainless steel reactors (probably likely), but downstream processes have moved to single-use tech even in the absence of a single-use upstream process. 

Timing-wise, the demand these two create probably won’t be meaningful this year, but it definitely supports significant growth over the next 5 years. 

Conclusion 

Hopefully this primer has been helpful in better understanding why the bioprocessing space is getting so much well-deserved attention. It is intended to set the backdrop for future articles. 

As far as specific companies go, I covered Avantor in my last article, where I outlined why I think that there were better uses of Avantor’s M&A dollars than to be #5/6/7 in bioprocessing. The top players, such as Danaher/Sartorius/Repligen/West/etc, are all compelling businesses at the right price. Stay tuned for deep-dives on each over the next 12 months.