By: Frank Corden
As I was driving in this morning to the Hynes Convention Center, I was thinking about how lucky we are to have an international meeting of this caliber in Boston. I just love BPI. It’s such a great mix of presentations; the breadth of discussions from the applied science around cell engineering to the underlying business drivers that will carry us into the next decade is amazing. The organizers successfully execute this daunting task with seeming ease, though I’m sure in the background their little duck feet are paddling away.
Those of you who know me have heard the story of how I became a scientist. As a kid growing up in the sixties and early seventies, my interests in science and technology were birthed from watching Jacques Cousteau television specials and the race to the moon. In the years since, I’ve had the pleasure of seeing first hand that those fish on the coral reef are really all those beautiful colors and that NASA used proven technology, kerosene and liquid oxygen (aka Rocket Propellant No. 1) to get our astronauts into space. Fortunately, my interest in all cool things science hasn’t waned with age and that’s why I love BPI.
Yesterday afternoon, I was sitting in on the presentation “Raising the Bar: Advanced Analytics in Upstream Bioprocess Development.” Much of the discussion focused on the use of Liquid Chromatography coupled with Tandem Mass Spectroscopy (LC-MS/MS) to identify and quantitate attributes of various cell culture processes. Without getting overly technical, the technique can be used to map the amino acid sequence of a drug substance as well as the pattern of added sugars (glycosylation) that are bound to an intact monoclonal antibody molecule.
LC-MS/MS works by very accurately measuring the mass of the peptide fragments created from the proteins in the sample during the sample preparation process. By accurately measuring the mass, the LC-MS/MS can identify the amino acid composition of each fragment. This composition and the mix of the fragments, especially unique fragments specific to the proteins of interest, enable the instrumentation to identify with a very high certainty the presence and relative concentration of different proteins in the sample.
In the data shown, the presenter, Chris Yu, demonstrated the practical power of the LC-MS/MS method to characterize drug substance. Data showed that LC-MS/MS could identify and quantitate host cell proteins and, further, could give positive confirmation of glycosylation patterns. But what was most interesting to me was a more basic discovery…
As part of the effort to characterize a particular drug substance, the team identified a low percentage of peptides which differed from the expected amino acid composition, often a swap between serine and asparagine. Initially the thought was that the DNA sequence for a fraction of the cells was different, as a result of either spontaneous mutations or by misincorporation of the DNA sequence during cell engineering. Surprising neither of these possible errors was the root cause.
It turns out that there is a natural error rate for misincorporation of amino acids into the primary sequence of proteins. The authors presented data that showed the error rate was in the same range for both mammalian cells and E. coli. The error rate could also be influenced by the relative availability of the amino acids in the cell culture. Restrict the availability of the intended amino acid relative to the incorrect amino acid and you get higher substitution. Provide a plentiful source of the intended amino acid and the error rate decreases.
On reflection, it’s not surprising that some low level error rate should be expected. After all, these biological processes are driven by chemistry. You can have a preference for a given binding affinity or reaction path, but it’s just a preference even when it’s a very strong preference. As an analogy, byproduct production in a chemical reaction is common and even if the reaction is very strongly directed for a particular outcome, in all but the simplest reactions, some byproducts are created. So, you can view the amino acid substitutions as byproducts.
Whether the protein with the substitution is functionally different than the desired product is unknown. What we do know is that the LC-MS/MS can measure the error rate and that process conditions affect the error rate. So if the error rate turns out to be a critical quality attribute (CQA) of the protein that is the drug substance, we can monitor it and control it. Neither the deeper insight gained about misincorporation nor the understanding of the ability to control the rate of amino acid substitution would have occurred without the applied research of Dr. Yu’s team and the advanced capabilities of the LC-MS/MS instruments.
Gaining that insight into how biology really works is the 21st century equivalent of rocket science - that’s why I love this meeting.
Share this article with your social network, just click below to share now!