Sunday, January 25, 2015
Several news articles caught my attention this past week. One dealt with the antibody cocktail for the treatment of Ebola (Andy Pollack was an important source of information here). I note that I have managed to avoid the entire Ebola news cycle since I’m more of an antibiotics person than an epidemiologist or virologist. But Andy’s article reminded me of something I wrote about in my book five years ago. Biotechs, with a few notable exceptions, are notoriously bad at thinking about manufacturing their potential products during the early stages of their development. This oversight can lead to severe problems and long delays later. That is exactly what appeared to happen to ZMapp and its partners including BARDA (Biomedical Advanced Research and Development Authority – an agency of HHS). ZMapp, if you don’t remember, is a biotech company that had discovered a mixture of antibodies that could cure Ebola in a monkey model of infection. Their product was actually used to treat a few patients. Whether it works in humans or not is anybody’s guess since no actual trials were ever carried out and patients do recover from Ebola spontaneously in 30-60% or so of cases. But the company’s idea of manufacturing was to use genetically engineered tobacco plants. This idea is not new – we were talking about this in Wyeth in the 1990s. But so far, no marketed product has ever been manufactured this way since achieving the scale required for commercialization or even decent clinical trials has never yet been achieved. But this is where ZMapp placed its future. When BARDA partnered with ZMapp, one of the first things the BARDA scientists did was to explore other modes for manufacture. All this takes time and it is not even clear that even if they were to use a more traditional method (like CHO cells for e.g.), the resulting antibodies would have the same activity. Therefore, the new antibodies might have to be tested once again in the monkey model. How many years of delay is this? I don’t know, but to me this is another great example of the shortsightedness of biotech when it comes to manufacturing their products.
On the ear infection front, another placebo-controlled trial was just published in the Journal of the American Medical Association. First, I need to remind you that two such trials were published several years ago looking at the clinical cure, relapse and complication rates of antibiotic vs. placebo treatment of acute middle ear infections. Both trials definitively showed that antibiotics have a substantial treatment effect to the point where the FDA now allows non-inferiority trials of ear infections once again. The new data published in JAMA were obtained during trials that were carried out before and during the earlier trials. This is important since it is becoming highly questionable whether it remains ethical to withhold therapy from children with well-documented acute middle ear infections. The trial was carried out in Finland – like one of the earlier trials. These new data show that the antibiotic-treated children had an impressive rate of resolution of the middle ear fluid that builds up during the infection compared to those that received no antibiotic therapy. This also may be associated with a lower rate of hearing loss in the treated children according to the authors. While I agree that these new data are of potential value to physicians considering whether to treat children with antibiotics or not, I think that it is long past time to stop initiating placebo-controlled trials of antibiotic treatment of such children.
Finally, Europe has banned a 120-page list of generic drugs manufactured in India by Abbott Laboratories, Actavis, Dr. Reddy's Laboratories, Mylan Pharmaceuticals, Sandoz, Takeda Pharmaceuticals and others. The reason for this ban is that, according to French inspectors, a contractor called GVK biosciences manipulated the EKG data of subjects during studies to show that the generic drugs were equivalent to the branded products in question. While it is unlikely that this actually would affect the health of patients taking these drugs, it is more evidence that India, as the world’s largest and possibly least expensive manufacturer of drugs, is also the most perilous place to carry out such production.
Wednesday, January 14, 2015
Yesterday, Roche, Meiji and Fedora pharmaceuticals announced a deal that could be worth up to $750 million to develop OPO595, Fedora’s phase 1 DABCO B-lactamase inhibitor. Roche will develop the drug throughout the world with the exception of Japan where Meiji will obtain commercialization rights. I was unable to find the details of the deal – but this is the broad outline according to a Fedora press release.
There are a number of fascinating aspects to this. The President and CEO of Fedora is Christopher G. Micetich, the son of Ron Micetich. Ron was the discoverer of tazobactam, one of the early inhibitors of B-lactamase and marketed by Wyeth as piperacillin-tazobactam – a billion dollar plus product. I was involved in the early days of tazobactam showing that piperacillin would be a good partner for the inhibitor. Later, I helped develop the drug in the clinic and even later was involved with the marketing strategy for piperacillin-tzobactam for Wyeth. The world is so small . . .
Once again, this deal clearly shows that Roche is serious about getting into antibiotics. This will be their second major clinical stage deal after their deal with Polyphor late in 2013. It also leads me once again to question exactly what they are doing.
As far as I can tell from published data from ICAAC in 2013, OPO595 is a DABCO type non-beta-lactam beta-lactamase inhibitor – that is – it is similar to both avibactam and Merck’s MK-7655. Avibactam (partnered with ceftazidime) has completed one phase 3 trial and will complete another soon. The FDA’s advisory committee has already voted to approve avibactam (combined with ceftazidime) mainly based on its phase 2 data. Merck’s MK-7655 is partnered with imipenem and is still in phase 2 trials as far as I can tell. But both of these DABCO inhibitors are years ahead of OPO595 which is only in phase 1. Why would Roche leap into a third DABCO inhibitor. Perhaps they are thinking about its “enhanced” anti-bacterial activity beyond its ability to inhibit beta-lactamases. But my experience (and I have a lot of it) with this class is that some antibacterial activity is typical. Avibactam can have activity against E. coli at as low as 4 ug/ml. Novexel had a DABCO compound, NXL-105, with activity against Pseudomonas aeruginosa that was thought to be partly related to its ability to bind PBPs of Pseudomonas. In fact, the entire DABCO series started at Roussel-Uclaf targeting antibacterial activity through PBP binding and was only later discovered to have potent B-lactamase inhibition activity. So I’m not convinced that OPO595 will have much of an advantage over the other DABCOs based on data published so far.
One potential strategy for Roche/Meiji would be to try and quickly develop a combination with a monobactam similar to the aztreonam-avibactam combo that is in phase 1 trails at AstraZeneca. This combo has the advantage of being active not only against Gram negatives bearing extended spectrum beta-lactmases, but also against those harboring metallo-beta-lactamase like the NDM-1 superbugs. Many of us have been screaming at AZ to develop their version of this quickly for years – but so far – no dice. Here Roche/Meiji might try and beat their competition to market with this combination targeting NDM-1 like superbugs - and go on from there. But will this pay off? We await further developments . . .
Monday, January 12, 2015
Last week (during my vacation), Losee Ling, Kim Lewis and their collaborators published a paper in Nature that was picked up by the world’s press – and deservedly so. But most of the news articles missed the key point – at least to me.
What the press focused on was the new antibiotic discovered and described in the paper, called teixobactin. It turns out that teixobactin represents a new class of antibiotics that binds to the bacterial cell wall and thus inhibits further synthesis of the cell wall. This mechanism is similar to that of vancomycin, teicoplanin, dalbavancin and other glycopeptides, but teixobactin has a completely novel structure. Teixobactin was even active in a mouse model of infection. If it ever gets developed, it will be only available by intravenous (maybe intramuscular) injection and will only be active against Gram-positive pathogens including superbugs like MRSA.
Teixobactin itself, in spite of all the hype, is not so exciting to me. First, Gram-positive pathogens, even the MRSA superbug, are not the area of major medical need right now. We have lots of old and new antibiotics active against these organisms. We need new drugs active against Gram-negative pathogens like the carbapenem-resistant superbugs now plaguing patients and physicians around the world.
Also – antibiotics like teixobactin are a little like old news. We at Wyeth, and many pharmaceutical companies during the 1990s, carried out programs where we re-examined our collections of natural products going back to the 1940s when soil was being collected from around the world and screens based on microorganisms from soil were used to identify new antibiotics. We discovered an old natural product sitting around on our shelves since the 1960s called mannopeptimycin. The extract from the culture of the producing strain was active in vivo in a mouse model just like teixobactin. And mannopeptimycin bound the cell wall in a way that resembles teixobactin. Alas, mannopeptimycin caused severe inflammation of blood vessels in animals both at the site of injection and at distant sites and could never progress into clinical trials. The fate of teixobactin remains to be seen as further studies of the safety of the new product are carried out.
But lets talk about the most exciting aspect of the discovery – the new method for finding new antibiotics. The way Ling et al found teixobactin is novel, simple and may pave the way to a whole new generation of new antibiotics from soil microorganisms. If you think about it, microorganisms from soil were and are the source of most of today’s antibiotics starting with penicillin, cephalosporins, strepotomycin all the way through tetracyclines, erythromycin, vancomycin, daptomycin etc. But its been estimated that you can only grow less than 1% of microbes that live in soil using normal culture media. And to extract antibiotics from the soil, you have to be able to grow them. What Ling et al did was to take soil samples and dilute them such that they would introduce about one bacterial cell in a test tube with a membrane bottom. This membrane would allow nutrients to flow into the test tube. The test tube (in this case actually a small plastic well) was then placed back on the soil from which the original dilutions had been made. This would allow about 50% of the organisms to grow – presumably getting key nutrients from the soil that are not present in normal culture media. Ling et al could then extract the growth media from these test tubes to identify those containing products that would inhibit the growth of other bacteria – that is – antibiotics. They came up with teixobactin. By carrying out such a screening campaign using large automated systems and many soil samples, someday, we might even be able to find new antibiotics active against the Gram-negative pathogens where our medical need is the greatest. To me, this is the key finding of the paper, not so much teixobactin.
My hearty congratulations to Drs. Ling, Lewis and their entire team!