The quinolone antibacterials have over fifty years of use in humans, although the first generation of them was not all that impressive. Most people either don’t know where they came from, or will answer “naldixic acid“, which was the first of the breed to make it to market (1962, from Sterling). That one, so the story goes, was developed from an impurity isolated during a synthesis of the antimalarial chloroquine, and was found to have antibacterial properties.
But “so the story goes” is not much of historical foundation. This new open-access article (from Gregory Bisacchi at AstraZeneca) is probably the most comprehensive look at the early days of the quinolones, and it tries to fill in a lot of the missing details. There are many. I found the story itself to be interesting, although those days in drug research now have a sort of otherwordly feel to them compared to how we work now. It’s a bit like reading Reminiscences of a Stock Operator, which is basically a biography of Jesse Livermore under another name and looking at today’s stock market. The actions are all recognizable, but appear to have taken place on another planet.
Adding to the problem is that even in the earliest days, the origins of this drug class were not clear:
Considering the vast and still growing literature on (fluoro)quinolone antibacterials and the medical and commercial importance of this class, it may seem surprising that the lines of research leading to the identification of nalidixic acid had been nearly a complete mystery to the drug discovery community for several decades following its first disclosures. David Greenwood stated in his 2008 book “Antimicrobial Drugs: Chronicle of a twentieth century medical triumph” that “Sterling-Winthrop was reticent about revealing details of the discovery”
There were references to at least two more detailed papers “to be published” that never appeared (so that’s certainly not a new phenomenon!) But the switch from the original chloroquine impurity (which had a quinolone structure) to naldixic acid (which is a 1,8-naphthyridone) has been mysterious. What this new paper shows, though, is that quinolones had already been identified as potential antibacterials in three patents from ICI, which seem to be almost forgotten now. This almost certainly influenced Sterling’s own research and patent filings, but the exact details are beyond recovery.
There are some Australian academic papers mentioned compounds in this class from even before the ICI patents, but they have no mention of biological activity. And there are some Indian publications in between the ICI and Sterling work, which cite the ICI patent. But Sterling’s own patents don’t get around to citing any of the above. As Bisacchi put it, “Conceivably, patents filed during this earlier time were governed by less rigorous standards regarding broad prior art citation than we are accustomed to today.” Searching the literature was, of course, more of a pain back then than it is today, but at the same time, there was a lot less literature, and it’s hard to imagine that Sterling’s team was unaware of all this prior art – especially since the move to naldixic acid itself looks very much like an attempt to get around it. For their part, as this paper notes, ICI never published their own quinolone work in any journal, and that along with Sterling’s relentless lack of interest in mentioning it has left it almost entirely forgotten.
Bisacchi mentions that he has tried to dig into the AstraZeneca archives (which stretch back to the ICI days) to see what happened with that project, but that he’s been unsuccessful. It does appear that the company missed a significant opportunity by not following up on their early quinolone work, though. The whole story brings up a larger problem in drug discovery and in scientific programs in general: it can be very difficult to reconstruct their history. Decisions get taken that made sense at the time, based on limited data, but in retrospect become hard to understand. Even the people directly involved can end up rationalizing such things and making them more of a coherent narrative in their own memories, and that doesn’t even take into account more conscious forms of tale-telling. Very few research programs are ever retold in all their shaggy, Brownian-motion glory after the fact. The difference is roughly that of natural human speech, as transcribed from a tape, and dialog as written for a play.
But this isn’t just a problem in science. It’s a problem of history in general. What looks from the outside like an imposing structure of facts and eyewitness accounts can turn out to be a rickety structure when you start poking around inside it. A story about Sir Walter Raleigh, who began his “History of the World” during his imprisonment in the Tower of London, might be appropriate here. As it’s told – and here’s some perhaps not fully documented history as well – Raleigh heard some sort of major disturbance going on one day, well out of his sight, and he couldn’t make out what had happened. In the coming days, he heard several completely different versions of what had taken place, and didn’t know which one to believe. This, it’s said, got him thinking about his own manuscript – how was he to write a history of the world, when he couldn’t even be sure about what had just happened within his own earshot? Historians have faced this problem ever since. Journalist Ron Rosenbaum has written about this several times, describing the tone that you can hear in an interviewer’s voice in an old tape recording, as the last surviving witness to some event is questioned. This is the only person left who can clear up the story – if they don’t know, it will turn into yet another mystery. So it’s not surprising that we in drug discovery have plenty of our own.