In the November 2013 publication of Nature Biotechnology, we authored an article titled “In search of dry powder” to serve as a resource for busy entrepreneurs seeking to move efficiently through the VC fundraising process. Published as part of Nature’s “Bioentrepreneur” series, the article builds on a previous PureTech blog post which ‘crowdsourced’ a database on active VC funds by creating a living document, containing both traditional information about funds (past investments, fund size, vintage year) and input from the ‘crowd’ of relevant industry stakeholders. Given the rapid changes in the VC community in general, the table in the Nature article (last updated in September) is already out of date.
We’ve created a new living document to continue to crowdsource updates on the active healthcare VC community, and to update the data on presented in the article. Our aim is to allow for a virtual peer review process that improves accuracy and keeps the data current. As part of the community, we welcome your feedback and encourage you to contribute to updating and correcting this dataset. To do so, please either directly contribute to the database or reach out to us at firstname.lastname@example.org, or via Twitter at @PureTechH
Does it have to be a biotech or pharma company that figures out a way to monetize fecal transplants? Or will hospitals find a way to make them appealing to patients without help from industry?
Given the potential hurdles involved, many in the popular press have questioned whether FT can be successfully commercialized (me too). Regardless, people will try to find a way. They will need to come up with a viable business model and navigate numerous challenges.
Commercializing Fecal Transplantation (FT): who can stomach the challenges?
Patient perceptions: the really desperate patients have bigger worries than the “ick factor”
The visceral distaste the general population has for the procedure may not necessarily change with new data from well controlled, randomized trials of FT. However, the views of the general population are not necessarily representative of how CDI patients at the end of the rope feel about the procedure (think patients that repeatedly relapse or experiment the fulminant form of the disease, for whom the alternative can be emergent resection of the colon). This topic has been covered at length by others (See for example Maryn McKenna’s story in Scientific American) so I won’t elaborate here.
Intellectual Property: little room to get broad patents
There are decades of prior art that would preempt broad claims around pharmaceutical compositions consisting essentially of feces, and around the FT procedure. Going into this in depth deserves a whole separate piece. Suffice it to say that the basic steps of obtaining a sample from a pre-screened donor, FT preparation (use of preservatives like saline, homogenization, filtration), recipient preparation (lavage, pre-treatment with antibiotics), and means of administration (nasogastric and nasoduodenal tubes, colonoscope, or retention enema) have been described in the art and are off the table. Novel methods to select donors or improvements on methods to process and administer samples could well be patentable, but I am not convinced these would clear the bar for IP protection that Pharma expects (specific compositions of well-characterized microbial consortia will likely be more attractive to Pharma, more on them in a future entry).
Uncertainty on which regulatory framework will apply
Will FTs be regulated as Biologic Drugs, Human Cell & Tissue Products (HCT/Ps), medical procedures, or none of the above? In principle, FTs meet the FDA’s definition of Live Biotherapeutic Products, LBPs (a biological product that: 1) contains live microorganisms, such as bacteria or yeast; 2) is applicable to the prevention, treatment, or cure of a disease or condition of human beings; and 3) is not a vaccine). However, meeting FDA’s GMP standards for LBPs would be unworkable for an FT product, which can contain thousands of bacterial species. Characterizing each of the strains in a FT to GMP standards and with all the reproducibility, validation and documentation required is out of the question. Not only it is technically unfeasible, but nobody would be able to afford the cost of an FT produced by these standards.
Alternatively, the current regulatory framework for HCT/Ps could provide some guidance. Cell therapies pose challenges to the traditional Investigational New Drug Application (IND) or Biologics License Application (BLA) US regulatory approval pathways similar to those that FT can expect to encounter (donor eligibility, tests for infectious agents, manufacturing requirements, regulation on manipulations that might alter the biologic function of the composition, storage and distribution protocols, etc). An obvious difference is that an FT is composed of microbial rather than human cells (although what should and should not be considered “human” in light of the evolving understanding of humans as “superorganisms” has philosophical undertones best left to others to debate). Either way, the FDA’s Tissue Reference Group apparently has recommended against regulating FTs as HCT/Ps (“Microbiota isolated from fecal matter of a donor is not an HCT/P, as defined under 21 CPR 1271.3(d)”). So scratch HCT/Ps.
FTs also share attributes with medical proceduressuch as bone marrow transplants, human organ transplants, blood transfusions, and transfusions of blood-derived products. By way of an analogy, reconstituting the gut microbiota via a fecal transplant can be considered akin to reconstituting the hematopoietic system via hematopoietic stem cell transplantation (HSCT). There are some non-trivial differences between the two that in my view make HSCT a far more dangerous and complex procedure: immunological matching will probably not be required for FT, which substantially expands the pool of donors; conditioning regimens are much simpler for FT than for HSCT, which usually requires the recipient’s immune system to be destroyed with chemo or radiation; and fatal complications, e.g. graft-versus-host disease in HSCT, are unheard of for FT. Blood transfusion procedures are also loosely analogous to FT. Donor screening protocols (health history, presence of infectious agents, etc) are comparable –with the notable exception, again, that FTs don’t require immunological matching. The composition of blood transfusions for different individuals cannot be standardized, neither can that of FTs. Different blood groups occur, and it has been proposed that different microbiome types may occur in the human population too, although this remains a very controversial topic among academics in the field.
I think it is conceivable that the FDA would allow hospitals to offer FT as a medical procedure. This could be rationalized on the basis that FT manufacture involves minimal manipulation of donor samples and that a fecal transplant directly replaces the function for which it was biologically intended (the transplant ‘repopulates’ a previously damaged recipient’s microbiota). Under this scenario, FTs would not be considered commercial products, would be overseen instead by state Medical Boards, and would face less regulatory scrutiny than LBPs or HCT/Ps. I doubt, however, that the FDA would let companies commercialize FTs as medical procedures. Unless all the operations of preparing a FT are conducted at the hospital and in a short period of time, it is difficult to claim that the FT has not been manipulated and its stability has not changed. Any step such as freezing, thawing, preparation of a defined dose, packaging, storage, or shipping, could be seen as a manipulation. If the experience with processed blood products and cell therapies is of any guidance, the FDA might sooner or later designate companies manufacturing FTs as biological drug manufacturers and require from them a BLA. An illustrative example of an FDA-approved blood product for which manufacturers are required to get a BLA is Intravenous immunoglobulin (IVIG). Among other things, IVIG manufacturers are asked to specify any additives and modifications made to the final product, guarantee the product is free from certain harmful contaminants, and have in place suitable quality control release tests. Similar requirements could be demanded from FT products.
It will be interesting to see what balance the FDA strikes between safety and encouraging translation of FTs. As far as I am aware, the agency hasn’t really made its thinking known, but it could do a good service to this field by emphasizing safety standards over product testing standards, since there are some serious questions over the ability of existing technologies to ensure product consistency and quality of FTs.
Plausible Business Models: Individualized vs Off-The-Shelf Treatments
Potential business models for commercializing FT may involve hospital medical procedures, commercial products and services, or a combination of both.
Two potential business models would be individualized and off-the-shelf approaches, both of which are currently being pursued for stem cell transplants. An individualized procedure (Figure 1a) could be performed entirely at the hospital or transplant center: a stool sample would be obtained from a pre-screened donor; the sample would be minimally manipulated (homogenization, filtration); and finally, a physician would take the fresh sample and instill it into a recipient. The sample would never leave the premises of the hospital, which would minimize the time delay between harvest and instillation. Hospitals could potentially generate revenue streams from testing donors, from labor involved in collecting the stool sample, processing it, and instilling it, and from monitoring and testing the recipient (not all of these procedures may be reimbursed though, see section below). Disposables such as lavage solutions or nasogastric tubes could be sold individually or as kits, and would generate revenue for existing suppliers of these products. This model could face the lowest regulatory hurdles assuming hospitals were allowed to offer FT as a medical procedure exempt of regulatory approval. Variability of the results due to donor heterogeneity and non-uniform processes across hospitals could be a potential concern.
A first variant of this individualized model would involve donor stool samples being sent to a central facility(Figure 1b)which could do all the necessary testing, process the samples, freeze them, optionally pre-package them into an application device such as a colonoscope, and store them. Any such additional steps would likely invite stricter regulatory oversight. The product would be shipped to the physician, who would instill it into the recipient. This alternative would enable performing more complex manipulations of the sample offsite (e.g. expanding certain microbes within the sample, reducing other microbes with antibiotics, etc) although at present there is no science that indicates that such manipulations would be desirable. Each donor sample would be processed as an individual batch, requiring individual acceptance testing to determine the quality of the raw materials, in-process testing, and release testing to establish the quality of the final product. Downsides of this model would include more strict regulatory oversight, higher cost of goods, introduction of a time delay between harvesting and instillation, and more complex logistics.
A second variant of the individualized model could involve preemptive banking (Figure 1c) of a patient’s stool for use in the future by the same patient (akin to umbilical cord blood banking). Patients could send stool samples for banking in a central facility while they are healthy. The banked samples would be available to “repopulate” the patient if in the future he needs to undergo aggressive antibiotic courses or chemotherapy that might damage his microbiota.
An “Off-the-Shelf” business model (Figure 2) could involve obtaining samples from a limited number of “qualified” donors. These subjects would be pre-screened with standardized tests, and their stool samples would be processed and banked in a central facility, and later shipped on demand when requested by a physician. There would be an additional cost for maintaining the stool bank, but that could be offset by lower cost of goods of the product since each donor could provide several batches, which could be screened as one pool, resulting in less frequent acceptance, in-process, and release testing. As in 1b and 1c, manipulation steps such as freezing, storage, or addition of agents could also attract stricter regulatory oversight.
The viability of each of these models hinges on the regulatory framework applied to them and on how FT gets reimbursed in the future. The uncertainty is likely to keep many on the fence, but biotech startups with an appetite for risk may take the challenge. There is evidence that this is already the case, since a couple of startups have entered the FT stage (some info on them here).
Reimbursing FT: are we turning a corner?
The economic burden of CDI is huge and growing. According to the CDC, 14,000 American deaths each year are linked to CDI, and it costs on the order of $1 billion annually to treat infections. The rate of infections is at an all-time high and the emergence of a hypervirulent strain has contributed to an increase of 400% in deaths between 2000 and 2007. The numbers are just plain scary.
The current standard of care is oral administration of antibiotics such as vancomycin and metronidazole, which cost on average ~$400 per episode (but can cost up to 3x as much). However, the biggest burden to the healthcare system comes from associated inpatient care costs, which are on average $11,000 per episode. Payers ought to be thinking seriously about the prospect of a curative treatment that could take a chunk of these patients out of the healthcare system in one shot.
The lack of reimbursement for FT has been pointed to as a factor impeding broader use of the procedure. Until recently, physicians performing FT could get reimbursed only for the colonoscopy procedure (if that was the method of instillation used), but not for donor and patient pretesting (e.g. testing donor for Hep A, B, and C, HIV, parasites, etc), which can cost several hundred dollars, or for the labor in preparing the sample. A recent development indicates that attitudes toward reimbursement are evolving though. The Centers for Medicare & Medicaid Services (CMS) have just created a new code for CY2013 for FT (HCPCS Code G0455) which covers the work of preparation and instillation of the microbiota by any method. This is a first step. It is still unclear if any of the necessary donor and patient pretesting will be covered. The pretesting is not cheap, but it would make economic sense for payers to cover it considering the potential savings in pharmacy and inpatient care costs versus the current standard of care.
In its current form, FT may not attract the attention of pharmaceutical companies, which are more likely to focus on second-generation products (e.g., well-defined and characterized cocktails of strains). However, standardization of the FT procedure, rigorous donor screening, and more comprehensive reimbursement could enable hospitals to broaden its appeal among physicians and patients. That is, of course, assuming regulators allow it.
Some background first. Fecal transplantation (FT) involves obtaining a stool sample from a healthy donor and administering it to a recipient by enema, via a nasogastric tube, or via a colonoscope. The procedure has been used primarily as a last resort treatment for patients with Clostridium difficile infections (CDI) that stop responding to antibiotic courses. CDI is a major problem in hospital settings. Just in the US, where diagnosis rates have increased four-fold since 2000, it causes 14,000 deaths a year. Often the FT donor is a relative that has been pre-screened to rule out carriage of certain enteric pathogens, parasites, and viruses, and usually the recipient is prepared for the transplant with a gut lavage, and then given one or more stool infusions. There isn’t much science yet to rationalize the donor screening criteria (it is unclear if relative and non-relatives are equally desirable), the preparation steps (e.g., it is still unclear whether pre-treatment with antibiotics helps the transplant ‘graft’), the administration route (all three mentioned above seem to work), the dose, dose frequency, and numerous other aspects of the procedure. But it works. Sort of like when you hit ‘reset’ when something is wrong with your computer and 90% of the time it does the trick.
FT also seems to do the trick around 90% of the time. As of a few months ago, there were over 300 published cases showing 90% resolution. Gastroenterologists and infectious disease specialists have been aware of the growing anecdotal evidence of efficacy. Still, combine the fact that published cases did not include well controlled, randomized trials with disbelief around the notion that anything good can come from feces and the result has been a mixture of surprise, amusement, and some skepticism.
This is why this first randomized, controlled trial with FT is a key development for the field. Van Nood and colleagues compared infusion of donor feces after vancomycin with vancomycin alone in CDI patients and found an 81% response after one FT infusion vs. 31% for the vancomycin arm. With a second infusion, the cure rate with FT was 94%. The trial was terminated early after an interim efficacy analysis when only 30% of the recruitment was completed because the safety and data monitoring board deemed it would be unethical to withhold FT from the remaining patients. When you take into account this trial as well as the previous accumulated clinical studies, the track record with the procedure is very impressive.
This time it is not just correlation
The most frequently heard cautionary warning around microbiome-related work has been that the numerous emerging associations between microbiome alterations and human disease reflect correlation, not necessarily causation. This is often a fair criticism and press releases that ignore the difference are doomed to be demolished by @phylogenomics, the official Twitter policeman of microbiome-related press blunders. As I’ve noted previously, I too think that some of the correlative connections emerging in the literature will turn out to have no role in driving disease. But the evidence that alterations in the microbiome in CDI aren’t just correlatively associated to the pathology, but are actually a driver of the pathology, is now very compelling. A randomized-placebo controlled trial is not absolute proof, but it ranks among the most reliable forms of scientific evidence when it comes to ruling out issues of spurious causality. Perhaps most importantly, it is this type of evidence that affects policy and clinical practice. If you weren’t taking this field of research seriously, after the NEJM paper, you have to.
…Or not. See the reaction on social networks
Of course, there is something deeply fascinating about jokes involving feces that no NEJM paper is ever going to change. The reactions on Twitter following news of the paper were hilarious. Here are some pearls of wisdom in 140 characters or less (see the full conversation here):
@adamfeuerstein: NEJM study: Infusion of shit from a healthy donor works better than vancomycin in treating c. diff. Um, I’d rather have c. diff, thx.
In the media frenzy following the NEJM paper there has been plenty of lighthearted commentary. Just about every imaginable play on words that could be tweeted or printed has been. Embedded among the puns, however, are a few of the issues that will determine if FT becomes a mainstream procedure. In short, The Good: credible science and good track record of clinical efficacy and safety. The Bad: patient and physician perceptions of the procedure (“ick factor”), the inability to get broad IP protection on the approach, lack of standardization (donor selection protocols, patient pre-treatment protocols, mode of delivery), and uncertainty around the regulatory route that will apply to FT. Science writers @marynmck (Wired, Scientific American), @edyong209 (Nature), and @carlzimmer (NYTimes) are among a few who have been discussing the topic in depth for some time now, and highlighted some of these issues.
Follow-on indications and a word of caution
The publication of the first randomized, controlled trial with FT has lent credibility to FT and to the prospects of manipulating the human microbiome for treating human disease. Some experts are starting to wonder out loud if FT should become first line of treatment for CDI. The clinical success in CDI is already leading to experimentation with the procedure in other diseases. CDI was a natural starting point for the procedure because the role of microbiome dysbiosis in driving pathology in CDI is unambiguous, doctors run out of options to treat desperate patients that relapse after first line antibiotic treatments, and published case reports of FT in CDI going back decades made it easier for investigators to get institutional review board approval for clinical studies. A natural follow-on indication which we will now see pursued more vigorously is IBD. A search on Clinicaltrials.gov shows 13 ongoing FT trials worldwide, 7 of which are recruiting IBD patients. IBD shares common mechanisms of pathology with CDI (a subset of CDI patients develop colitis); among gastroenterologists, the role of microbiome dysbiosis in driving IBD pathology is well accepted; and perhaps most importantly, several small human studies have already shown promising results of FT in IBD (e.g., here).
As it often happens, increased attention begets quackery and snake oil salesmen as well. Claims of astonishing results in unproven applications such as depression, multiple sclerosis, or what have you, will crop up.People will try FT at home. And somebody may drop the ball and transplant a contaminated sample. This is something to be concerned about. It only takes one self-taught cowboy passing as a physician to poison the well for the rest of reputable clinical researchers. Unsupervised FT outside of the hospital setting should be discouraged.
Clinicians working on FT are getting some long overdue respect. It’s time to take FT seriously and it’s time for regulators, payers, and the biotech industry to join the conversation. It remains to be seen if anyone can make a business out of FT, but it’s worth starting a discussion about it. My thoughts on that next week in this blog.
Alzheimer’s Disease (AD) R&D might be the most challenging therapeutic area that faces medicine today. It is known that AD is caused by the build-up of proteins into clusters that clog up nerve cells in the brain. These “plaques” break down nerve cells which, in turn, results in a decrease in the ability of these cells to function thereby leading to the familiar AD symptoms of memory loss and erosion of cognitive skills.
How these plaques form is a matter of scientific debate. One theory is that AD is caused by the build-up of the protein, beta-amyloid, and that if you can prevent deposition of this protein in the brain, you can slow plaque progression. Unfortunately, two compounds designed to do just that, PfizerPFE-0.56%’s bapineuzumab and Lilly’s solanezumab, failed in late stage clinical trials, thus casting doubt on this hypothesis. A second theory is that AD is caused by the build-up of another protein, known as tau, inside neurons where tau causes protein tangles. Interestingly, post-mortem samples of AD patients show that there is a correlation between tau protein tangles and the degree of AD severity.
What further complicates finding a cure for AD is that it is unclear how best to conduct clinical trials to determine the efficacy of a new experimental medicine. It is not known if a patient needs to be treated long before AD is fully apparent or if people need to be treated at the first signs of disease. Once the AD process starts, can it be halted or reversed? Should people with a perceived genetic predisposition to AD be treated to slow or prevent the onset of the disease? All of these questions need to be answered. However, to do this, long and expensive clinical trials are required as was done with bapineuzumab and solanezumab. Thus, the challenges that exist in AD go well beyond the normal R&D process of having a hypothesis and finding a compound safe and potent enough to test the hypothesis in patients. Unlike a disease like cancer where clinical trial protocols are well established, in AD even the clinical trial design is at an experimental stage.
“I think we have to do a lot more basic science work to understand what’s going on. We really, at best, partially understand the cause of the disease. It’s hard to come up with meaningful targets.”
One can’t blame Viehbacher for taking this position. After all, there are a lot of diseases where new and/or better treatments are needed, diseases whose pathology is better understood and where the chances for success are greater. My guess is that his shareholders will support his decision to make safer R&D bets. But, as the lifespan of the world’s population increases, so will cases of AD. Furthermore, the increased financial burden of this disease will have an enormous impact on healthcare budgets. The fact is that our best hope in bringing new AD medicines to market rests with the major pharmaceutical companies as they are the only ones with the necessary resources to be able to perform the key clinical trials that will ultimately yield a breakthrough. If Sanofi’s stance is mimicked by other companies, the chances of an AD breakthrough greatly diminish.
“The most meaningful impact in Alzheimer’s might involve targeting multiple pathways and using combinations of drugs.”
Lilly’s plan is to go after AD by combining drugs that prevent beta-amyloid deposition and also tau protein tangles. To help them in doing this, Lilly has bought the rights to tau-measuring tests from Siemens . As Skovronsky explained:
“The whole field has been amyloid-centric, amyloid-driven. But we need more than that. That’s why we’re investing in tau.”
There is no doubt that Lilly has taken a much bolder move in AD compared to Sanofi, and it’s a move that will be challenged by industry analysts. But Lilly has stepped up to the plate on a health issue that has major consequences for us all. I, for one, am very happy that they are making these important, but high risk, investments.
*Our initial list includes primarily life science VCs but we encourage our collaborators (you) to add tech and other sectors
As institutional/serial entrepreneurs (PureTech is a venture creation firm, not a venture capital firm), we actively interact with the venture capital community. As a member of this ecosystem, we wanted to address an elephant in our collective room and get the conversation started on a resource for entrepreneurs.
A challenge entrepreneurs face in today’s environment is figuring out which VCs are actively investing versus those that are fundraising and/or not actively investing. In order to protect their reputation and ability to fundraise, venture capital firms can sometimes be evasive about their ability to actively invest. As entrepreneurs, we can try to pick up on code words or phrases that might indicate a lack of capital (a topic for another blog post). One thing we want to avoid is pitching (or worse, jumping through diligence hoops) for a venture firm that actually is not in a position to make an investment.
The term “dry powder” is often used to describe money available for both new and follow-on investments. However, for entrepreneurs seeking out a new firm, the important variable is whether the firm has the ability to invest in their company. Simply put, we would like to know which VCs are actively seeking to make investments in new companies, i.e. companies they have not previously made an investment in. Let’s call it “Relevant Powder.”
One way to figure this out is to look at when firms have last raised a fund and when they have made investments in new firms (via Thomson One or a similar database). We can assume that firms that have either just raised a fund recently and/or have recently invested in a new company are likely to have relevant powder remaining. On the other hand, we can guess, in the absence of raising a new fund recently or investing in new funding rounds, that those firms have no/limited relevant powder. Thus, these two data points can be very helpful. Our own lists of VCs that have recently raised funds, based in large part on this list from Jay Caplan’s blog, and made recent investments (based on a search on Thomson) as well as similar efforts that look at recent investments begin to provide part of the picture.
We think this data needs to be supplemented by a peer review process, by asking people in the community. The folks at the Coelyn Group got a start on this second approach. They surveyed a group of life science executives about the funding capacity of VC’s, and then surveyed the VC firms themselves to confirm/deny the results. “Actively” seeking investments was the operating criteria. Of course, “actively” is not as specific a term as we would like, and they do not provide additional information on who they interviewed to get the data, but at least it’s a start. When a member of our team tweeted about the Coelyn list in October, an active discussion was started, which inspired us to get a more comprehensive resource together for the entrepreneurial and venture community.
We believe that a crowdsourced approach may be the best way to tackle this problem. To this end, we have created a first pass at combining the existing data available into a living document. If we got something wrong, then please correct it by participating in the project. Here is the Actively Investing VC Spreadsheet, a live and working document for tracking the ability of a given firm to make investments in new companies (i.e., companies they haven’t previously invested in). We have seeded the list initially with Life Science investors but would like the crowd to expand the project to tech as well. Firms are categorized as “Active” if data points (last fund raised and recent investments in outside companies) plus word-on-the-street consistently support that. They are categorized as “Some activity” if there is some mixed data, and as “Limited/No Activity” if all data points support that.
So, take a look. If you think there are inconsistencies, use this form and let us know what they are and why (and if possible tell us who you are, as anonymity can sometimes be used to disguise animosity or bolster one’s own firm!) In the interest of transparency, the feedback will be available in its raw form here and we will revise the list as we get feedback. We hope you find it useful, until someone figures out how to accurately measure cash available for new investments on a firm-by-firm basis, and publish it in a public and transparent way.
A phrase by David Relman in a recent article in the San Francisco Chronicle sums up best how researchers in the microbiome field seem to view commercially available probiotics: “these somewhat irrelevant microorganisms”. Data generated to date testing probiotics in a long list of diseases has occasionally shown some beneficial effects. For example, some studies have shown that certain probiotics such as L. rhmanosus GG, L. reuteri, L. casei Shirota, and B. lactis Bb12 can shorten rotavirus diarrhea by ~1 day, and a small study showed a mixture of Lactobacillus, Bifidobacterium, and Streptococcus strains can reduce pouchitis flares. However, as a whole, the clinical efficacy data with probiotics has not been very convincing. In contrast, their safety track record has been excellent, and many physicians are comfortable suggesting them (often patients ask for them). The Food industry certainly deserves credit for having helped establish this safety record.
For years, probiotics have been promoted because of their alleged effects in “wellness”. Global sales of probiotics (food + supplements) are forecasted to reach ~$30B by 2015. Much has been written about the failure to support marketing claims on probiotics and the no-nonsense approach that the EFSA has taken towards evaluating such claims. For all the harsh criticism, there is actually some scientific rationale behind the use of probiotics in foods. It just happens to be too simplistic.
The story goes like this. At the beginning of the last century, Elie Metchnikoff observed that in rural areas of Bulgaria people who had a diet rich in milk fermented by lactic-acid bacteria seemed to live longer. Metchnikoff hypothesized that the lactic acid bacteria could be the reason behind the longevity, and proposed they worked by colonizing the gut, lowering the pH, and inhibiting “bad” proteolytic bacteria such as Clostridia. The theory proved to be far too simple. Already in the 1920’s other researchers showed that Metchnikoff’s probiotic (Lactobacillus delbrueckii subsp. Bulgaricus) could not colonize the human gut , and what Metchnikoff dubbed as “bad” bacteria now appear to be essential components of the healthy human microbiome.
The next iteration of the theory came in the 1930’s. Researchers argued that it would be better to isolate bacteria from the human gut, which makes sense. Attention turned to human-derived Lactobacilli and Bifidobacteria organisms such as L.acidophilus, L. rhamnosus, L. casei, L. johnsonii or B. lactis . The narrow focus on Lactobacilli seems now a historical vestige, since at the time it was already known that Metchnikoff’s theory had holes. The focus on Bifidobacteria, on the other hand, was based on observations by Henry Tissier around 1900 showing that these bacteria are dominant in breast-fed infants. However, neither Lactobacilli nor Bifidobacterium are major colonizers of the adult human gut (Bifidobacteria are dominant in babies during the first months of life, and the rationale for using them in babies may turn out to be sound in some cases).
Yet, somehow, these two genera include, with few exceptions, most of the strains currently used in foods and supplements and which have obtained GRAS (US) or QPS (EU) status. These two groups are considered “safe” based on their history of use, which surely has helped attract attention to them. But it’s hard to make a case why many other human commensals shouldn’t be just as safe. Yet, they have been systematically ignored. I wonder if what lead to that was a perception in the industry that it’d be far easier to commercialize close relatives of established safe strains (that had ambiguous health benefits) than to venture into exploring other gut commensals that might be more relevant to human biology. Whatever the reasons, here we are, one hundred years after Metchnikoff’s original idea, still making probiotics based on a conceptual framework that we know is flawed.
You’re not invited
Currently marketed probiotics don’t permanently colonize the gut. Some of these strains were originally isolated from dairy foods or fermented foods, and are strangers in the human gut. And the human-derived species of Lactobacilli and Bifidobacteria mentioned above are often only transiently present in the human gut. Typically they can be found in amounts of 105 to 108 bacteria per gram of feces , which is 1,000 times lower than other dominant species of the gut ecosystem. When given exogenously in clinical trials, detecting colonization has been a challenge, since they disappear from the feces soon after administration , . They have a hard time crashing the party in the gut. They are outcompeted by a synchronized home team of species that are more efficient at harvesting nutrients and better adapted to the complex human gut ecosystem. And while evidence is starting to emerge that shifts in other dominant commensals are associated with human disease (e.g. , , ,  ), there is still no compelling association that I am aware of between deficits of any commercial probiotics and disease.
For all these reasons, I think the current paradigm of the single-strain Lactobacilli or Bifidobacteria probiotic will continue to generate unimpressive data in the clinic (it may remain a marketing success in foods). At the other end of the spectrum, fecal transplants (yes, gross) have shown very promising results in human trials but are likely far too complex to commercialize (they contain undefined communities of 1,000s of bacterial species). Somewhere in the middle, potential future products based on simpler communities of dominant commensals seem worth exploring. Interestingly, the only probiotic approved by the FDA (that I’m aware of) as an animal drug (for prevention of Salmonella infections in poultry) is a mixture of 29 indigenous bacterial strains (PREEMPT).
Designing probiotics. You get what you select for
Protein engineers learn by heart the first law of directed evolution: “You get what you select for” (you may not know what the principles that govern the stability or structure of a protein are, but if you apply a selective pressure for, say, thermostability -e.g. a screen that selects proteins that retain activity at high Temperature-, you will get a thermostable protein). Keeping the distances, you can make an analogy with probiotics. Industry players have selected probiotics without any knowledge of the principles that govern the structure of the human microbiome (which the research community is just starting to figure out). They looked for microbes that would be culturable, tolerant to the acidic environment of fermented milks and yogurts (which Lactobacilli can tolerate, unlike other human intestinal species); that would be resilient in environments that have oxygen (unlike most gut anaerobes, many Lactobacilli have some tolerance to oxygen and can survive food packaging processes); and that would be resilient to the shear forces encountered during food manufacture (Lactobacilli and Bifidobacteria have thick cell walls that can sustain high-shear food production processes such as blending ). And they deserve a lot of credit for overcoming the technical hurdles of producing live organisms. But they got exactly what they screened for: products with favorable technological properties, but which may be irrelevant to human biology. Somewhat irrelevant. The good news is that we have only seen the tip of the iceberg. New, powerful tools will now enable to look beneath the surface.
Why Drugs that Modulate the Microbiome Could be Translated to the Clinic Faster than You Think
The last few years have seen an explosion of academic work in the field of the human microbiome and the NIH has made a big bet that this field can transform medicine. To make a long story short, scientists are finding that the bugs that colonize your gut, airways, skin, etc. do a whole lot of good things without which human life would be pretty miserable (see for example the collection of articles in the most recent issue of Science Translational Medicine, or the flood of articles in Nature 1,2,3and PLoS a couple of days ago) including protecting us from opportunistic infections, keeping allergies and autoimmune diseases at bay, or supplying key nutrients. The tricks that our microbes use to make all this goodness happen and the chaos that results from messing with them hold critical clues for developing a whole new class of therapies. Future microbiome modulators could be (among others) orally administered commensal microbes, or natural or synthetic molecules that mimic their effects. For full disclosure, at PureTech Ventures we strongly believe in this field too, and have started newco initiatives to mine this space including this company.
Other promising therapeutic modalities that captured the public’s imagination at the time they emerged such as recombinant DNA (rDNA) and RNAi have taken a long time to materialize (or have not been translated into products yet, in the case of RNAi). Given all the attention the microbiome field is getting now, and the lessons of the past, striking a note of caution when considering translation of microbiome therapies to the clinic seems wise.
Not your average new therapeutic modality
However, there are several reasons why drugs that modulate the microbiome could be translated to the clinic significantly faster than other new therapeutic modalities such as rDNA and RNAi:
· No fundamental technological barrier: Notwithstanding the technical complexity of the microbiome field, which requires integration of many technologies from genome sequencing, new bioinformatics tools, unique animal models, and metabolite profiling tools (more on this later), there isn’t one fundamental technological barrier that needs to be resolved to enable clinical translation. That was not the case for rDNA and RNAi therapeutics. Manufacturing stood in the way of translating rDNA drugs to the clinic. Delivery stands in the way of RNAi drugs. Initially, protein products were only available in miniscule quantities or were very costly to produce in large-scale. MAbs were not feasible until hybridoma technologies came around. Early pioneers of the rDNA industry had to figure out ways to express protein products in recombinants hosts, at high yields, with correct folding, with the right glycosylation patterns, at large scales, and with the right purity. That took time. Similarly, delivering RNAi therapies systemically to where they are needed in the body remains stubbornly complicated, despite some progress in using lipids and nanoparticles to address targets in the liver.
In contrast, the technology needed to manufacture products based on live microbes already exists. With some tweaks, essentially the same cell banking, fermentation, and analytical methods already in use to manufacture biologics can be applied to manufacture live organisms. Purifying a microbe from a fermentation broth is trivial in comparison to purifying a protein. And over millions of years of evolution, our microbes have evolved intelligent ways of delivering themselves into our gut via the oral route and without the need for fancy delivery systems. Humans are born sterile, and we get colonized by microbes that use a number of strategies to reach their natural niche in the human body, such as forming spores.
· Established proof of principle that the microbiome can be modulated and that this is of value to patients: OK, so there is exciting data coming out of microbiome research projects – but are there any simple applications of this therapeutic modality that have worked in the clinic? RNAi, stem cells, and genomics companies did not have a satisfactory answer to this question. In contrast, transplantation of the microbiota in humans (“fecal transplantation” … yep, eating somebody else’s poop) has been performed for decades with very promising results. At present, it is only used to treat refractory patients with selected conditions such as C. difficile infections or IBD. Whether the procedure is repulsive is besides the point. It works, and it establishes proof of principle that manipulating the microbiome (1) can be done and (2) is of clinical value.
· You can’t change your “human” genome, but you CAN change your microbiome: In fact, it happens all the time. It changes depending on what you eat, your hygiene, your age, or what drugs you take (particulary antibiotics). This provides much more flexibility for intervening. Identification of microbiome anomalies points to alterable fates.
· Safety: humans have life-long exposure to massive amounts of commensal microbes (we have on the order of a hundred trillion microbes in our gut). So you’d think we have evolved to tolerate a wide range of doses of certain live organisms. Yes, rDNA drugs too were expected to be safer than small molecules because they are more “natural” but they proved to be immunogenic sometimes. That has not been a problem with probiotics used as supplements and food ingredients, which have an extremely safe track record of human use.
It can’t possibly be that simple
Sure, it probably will not be simple. Some notes of caution often heard these days merit attention:
“Mapping out the biological pathways is a scientific process that takes a long time”. True, no matter what therapeutic modality is being used. But has this been the rate limiting step for RNAi therapies to make it to the clinic? And while mapping what role our microbiome plays in a highly complex disease like cancer will take time, there is low hanging fruit to go after in other therapeutic areas in infectious disease and inflammation, just as certain hormones and growth factors were a logical starting point for rDNA therapies.
“It is not always clear if microbiome changes are cause or effect, or what caused the microbiome changes in the first place”. While often it is not known whether microbiome alterations were caused by hygiene, diet, antibiotic use, genotype, immune responses, enteric infections, or combinations of the above, this has not stopped clinicians from intervening and shifting the microbiome of patients back to a ‘healthy state’ (See fecal transplantation comments above). Surely, some of the correlative connections coming up in the literature will prove to have no role in driving disease (they may still lead to very useful biomarkers). But there already is very compelling evidence that in other cases alterations of the microbiome are a driver in pathology, regardless of whether they were the initial disease trigger. Numerous experiments have demonstrated that transplanting the microbiota of diseased animals into healthy hosts transfers disease phenotypes and influences onset and/or progression of obesity and chronic inflammatory diseases such as allergies and autoimmune diseases. It is also well accepted among clinicians, and supported by animal experiments, that microbial communities promote pathology in IBD.
“We don’t know how much variation there is in the human microbiome across individuals and within individuals”. Actually we are starting to get a good sense of this variation thanks to $140M from the NIH and a growing suite of open source tools (bottom line: yes, there is variation in species but their functions, e.g. metabolic pathways, vary little across individuals). In certain instances, this variation will uncover subtypes of individuals within a complex disease spectrum, improving the odds of success in clinical trials.
“Probiotics in foods and supplements have shown little to substantiate their health claims and don’t even colonize the gut”. Again, true in most cases. In making probiotics, just as in making antibodies, you get what you select for. And food companies have focused on selecting strains based on their technological properties (manufacturability, ability to formulate in food matrixes, acid tolerance, shelf live, etc), NOT based on their functions in human biology. Which is a theme for a whole separate blog entry, coming up soon…
We had a few weeks off-blog, but are diving back in this week to share some interesting biotech conversations happening on Twitter.
This week, we noticed a lot of discussion dedicated to what might broadly be called “access to science information.” You may be thinking about one specific case when reading that phrase, but we noticed that theme pop up in quite a few contexts. It is certainly telling to see that conversations on core philosophy of information dissemination and preserving scientific value can garner the same discourse as stock value and IPOs (especially during the week of a certain large IPO).
We have written on the Open Access (OA) trend previously, and this topic continues to be hot. One specific piece of press hit the airwaves this week as a potential milestone in the OA movement: UCSF announced that they are embracing a fully open access policy for all faculty. However, the announcement did not fall quietly on Twitter, as journalists and scientists debated whether UCSF was just giving lip service to the topic: http://peerin.co/ppQ2
Ever since the release of #asco12 abstracts last week, there has been an ongoing debate on social media regarding some media personnel’s access to embargoed scientific data (conference presentations, clinical trial data, scientific pubs, etc.) ahead of public releases. Embargoes are a long-standing tradition that presumably enable the benefit of real-time commentary and analysis (and with it free buzz) upon a data announcement. However, this is clearly not universally held as a positive point in the industry, especially when investment and stock prices are involved, referenced by these few lively back-and-forths:
A third context where this open access theme was debated related to how that information is communicated to the lay public. Science writers have long struggled with the task of properly representing the complex nuance of the scientific concepts they cover while maintaining a flavor that is palatable and approachable for the non-scientifically trained. A NYAS #Sackler Symposium on Scientific Communication predictably spurred many Twitter conversations.
A particularly interesting exchange nucleated from the quoted stat that the most trusted person in Canada was a scientist, with ensuing discussion regarding the degree to which this sheds light on America’s science literacy or communication issues (Aside: would Athletes and Entertainers top our list?).
One increasingly valuable function of conversations occurring on social media platforms like Twitter is the ability to get crowd-sourced, technical news that is rapidly fact-checked and analyzed. While many view Twitter as a layman’s tool, it draws increasingly from deep, industry-specific experts. This surfaces surprisingly deep analyses of data and underlying assumptions associated with news stories, and dissemination of that information to a broad audience. (The #arseniclife Twitter discourse comes to mind as one of the most game-changing. If you’re unfamiliar, Carl Zimmer wrote a great synopsis of the entire fiasco last year.)
The ability to view full Twitter conversations puts these expert-driven exchanges on display in the wake of high-profile announcements. A few examples from this week below:
1. Big news was made this week on the approval of Protalix’s ELELYSO for Gaucher’s disease, with the majority of buzz around the fact that Pfizer/Protalix have made clear they will compete with the marketed Cerezyme by offering their new drug at a 25% lower price point. However, a few interesting conversations popped up to pressure-test the real value of this market, including this discourse following fully-transparent sharing by Jason Napodano, CFA (@jnapodano) of his market model for estimating $PLX’s share-price:
Additional references were made to a similarly-transparent analysis made on SeekingAlpha
2. The Forbes MidasList was released this week, ranking the top venture investors for 2012 and highlighting their key deals. The biotech network immediately reacted to the low representation of life science investors, and highlighted examples of folks who should have made the list (spurring discussions about Forbes’ methodology):
3. RBC Capital’s bullish analysis of Pharmacyclics’s stock was debated, calling into question Pharmacyclics’s position in the competitive landscape and ability to capture value across untested indications:
Aided by our newly-prototyped tool called Appeering (first described here and soon to be in closed beta), we begin this week’s biotech roundup in the UK, which led the news and provoked some lively, nuanced conversations around healthcare economics and policy:
First, some of UK’s leading health experts weight in on recent news around whether Novartis’ Avastin should be used as a cheaper alternative to Lucentis for wet macular degeneration.Sparked by a post by Lancet Editor-in-Chief Richard Horton, experts like Claire Gerada (Chair of the Royal College of General Practitioners) and Alan Maynard (health economics professor at York University) debated the role of UK regulatory agencies in changing protocols to allow more off-license use of meds, and argued whether MD advocacy groups should be pushing MDs to prescribe off-label given evidence that patients may benefit. A link to that conversation can be found here: http://www.appeering.com/lifesciences-shared/index.php/conv/194668472465231872
AstraZeneca was in the news earlier in the week with their acquisition of Ardea Biosciences for over $1B. It seems that the company is working to drive change and fill their pipeline..
…but apparently not fast enough for their critics. This brings us to the strategic implications of David Brennan’s exit from his post as AZ’s CEO. Our perspective, mirrored by many of the conversations, is that analysts could be more thoughtful about the results of their critique of senior management of companies such as AstraZeneca. In particular, it is prudent to be realistic about how long it takes for new initiatives to bear fruit. One case in point is the entrepreneurial “innovative medicines units” or “iMeds” (championed by the well-regarded Martin Mackay) to drive R&D productivity. Why is there such a major disconnect between the time it takes to launch new drugs and Wall Street’s expectations of managing for the near term?
Staying in the UK, another discussion touched on the implications of new UK legislation which allows nurses and pharmacists to prescribe controlled drugs:
Finally, back in Boston… We are liking Scott Kirsner’s cool new series “The Friday Five” where he wraps up the top 5 stories in tech and biotech. This week his guest host was our own Daphne Zohar: http://t.co/QRTsJ8t5