James Frahill, Research Analyst, Pfizer Process Development Centre (Cork, Republic of Ireland) spoke to LCGC about the role of the chromatographer in the pharmaceutical process development group at Pfizer.
Q. What are the main objectives of your research group?
A: We are a pharmaceutical process development group. We are tasked with developing new synthetic routes to active pharmaceutical ingredients (APIs) that are already established products in the market place. The group uses novel synthetic chemistry, new technology, and innovation tools to design and deliver these new processes. This approach has led us in the past to deliver processes that include flow chemistry technology, enantioselective enzyme reactions, phase transfer catalysts, and more.
Q. How do you use analytical chemistry to develop new synthetic routes to APIs?
A: The analytical support underpinning the process development activities is a critical part of delivering these new processes. The analytics that support the process development gives us our understanding of what is happening in the process. We monitor reaction rates, product formation and purity, impurity profiles, and so on. In order to achieve a high standard of process understanding, we are continually developing new analytical methods, both spectroscopic and chromatographic, to cope with the changing matrix of the samples generated by the developing chemistry. The analytical activities in the group cover a range of tasks such as quality testing of isolated intermediates and APIs, identification and structural elucidation of unknown compounds or impurities formed in the new chemistry, analysis of enantiomers, and identification and control of any possible genotoxic impurities (GTIs) and on-line process analytical technology such as UV and IR measurements made at the reaction itself.
Typically, the methods developed during the process development are evolved into the analytical test methods that are validated and are filed with regulatory bodies during the process filing.
In the drive to get better process understanding the analytical group leverages any useful technology available in the market place to deliver the results required. We have recently expanded our collection of detectors by adding a charged aerosol detector (CAD) and a quadrupole time-of-flight (Q-TOF) mass spectrometer to our array of orthogonal chromatography systems which include reverse-phase high performance liquid chromatography (HPLC) and ultrahigh-pressure liquid chromatography (UHPLC), non aqueous reverse-phase HPLC, normal phase HPLC, and supercritical fluid UHPLC.
Q. What is the focus of your research at the present time?
A: At the moment I am working on impurity identification. I am analysing samples generated by our new process chemistry using LC–QTOF-MS. I use the high resolution of the instrument in MS mode to generate the molecular formulae for the peaks of interest. We then conduct MS–MS experiments on each of the impurities and determine the structure of the impurities by assigning structures to the fragment ions generated in the MS–MS experiment.
Nobody can claim to be able to do de novo structural elucidation using MS and MS–MS data alone, but given the limited amount of transformations possible in a pharmaceutical chemistry reaction, we can elucidate structures from the data with a high degree of confidence. We then synthesize the proposed structure and confirm its identity under the original analytical technique.
Q. Have you encountered challenges when attempting to synthesize a proposed structure and how have you overcome them?
A: We have encountered a number of situations where the synthesis of the proposed structure has been a challenge. In this situation we isolate the impurity using either flash or preparative chromatography and derive the structure of the isolated compound using nuclear magnetic resonance (NMR) spectroscopy. Our group has developed a streamlined approach to this workflow and compiled this as what we call a “knowledge map”. This is a decision tree with all of the guidance required at each step hyperlinked to each decision point. This saves the analyst time when they run into problems because the knowledge and the experience they can use is already recorded somewhere in our documentation, the knowledge map then links them directly with that documentation. We find this useful in guiding the analyst through the process of impurity identification.
Once the nature of the impurity is confirmed, we develop control strategies around the compound to limit its presence in our process to the appropriate levels determined from its possible toxicity. In the past this has led to control of impurities to the PPM level where they have been identified as possible genotoxic impurities.
I am also currently conducting method development around one of our processes. We have opted to develop these methods on an UHPLC basis, as we know that the manufacturing site that will ultimately receive this manufacturing process has UHPLC capability. However, we must bear in mind that external companies who may, at any point during the process development or commercial production, be called on to support the process may not have UHPLC capability. With this in mind we have chosen to develop all of our chromatography on superficially porous stationary phases for the first time. This should smooth the transition to HPLC if the method needs to be scaled up for other laboratories. So far the work is promising but high flow rates at HPLC can result when scaling from an equivalent UHPLC method, for example 0.4mLs/min on a 2.1 mm, 2.6-mm particle column up to 2.0 mLs/min on a 4.6 mm, 5-mm particle column.
Q. What analytical technology do you utilize and how do you improve efficiency in the laboratory?
A: We try to leverage any technology that we feel would be useful in our analytical support of the process development. This covers all aspects including new column chemistries (as above), various modes of chromatography (hydrophilic interaction chromatography [HILIC], non aqueous reverse phase), HPLC, UHPLC, supercritical fluid chromatography (SFC), and GC, as well as an array of detectors from QTOF down to variable wavelength detection (VWD) and refractive index detection (RI).
We try to maximize the potential of the instrumentation we have. We recently placed a number of chromatography systems on trolleys. This removes the restriction of using only the detectors built into their stacks and allows us to move the chromatography stacks to the less flexible detectors such as our mass spectrometer, charged aerosol detector, evaporative light scattering detector, and electrochemical detector. The ability to move the systems that offer us orthogonal approaches to chromatography allows us to interface these with a wide range of detectors, therefore giving us the best chance of success when trying to develop methods for difficult compounds or sample matrices.
We have also streamlined our method development approach to HPLC by incorporating switching valves into one of our stacks. This method development “super system” uses the switching valves to incorporate six different columns and up to 16 different mobile phases into a single screen. This approach allows us to run intensive screens which we design on a Quality by Design (QbD) basis. This helps us to reach our method development goals much faster than we would if we used a One Factor At a Time (OFAT) approach.
We also try to push the envelope with our analytical instrumentation. At the moment we are conducting an evaluation into using a preparative HPLC system as a flow chemistry reactor. If this trail shows some promise we may develop this into a two-dimensional HPLC approach where the flow chemistry is carried out in the first dimension and a heart-cutting approach is used to sample the flow chemistry and perform a HPLC analysis.
With all of this technology in mind we also try not to forget about the good old-fashioned techniques in analytical chemistry. We still use titrations and recently used derivatization to test for the low level presence of aldehydes in one of our samples, and even last week we used the barium chloride precipitation test to identify an unknown filtrate as a sulphate.
Q. What technical difficulties have you faced in your research recently and how did you overcome these challenges?
A: From an analytical point of view I have encountered quite a few challenges during our process development, particularly over the last 18 months to two years. The projects we are currently working on have a range of challenges from sampling non-homogeneous reactions, poor detection attributes of product or impurity molecules, both thermally sensitive and water-sensitive compounds, as well as the routine challenges of method development and impurity identification.
We are currently using dinitrophenylhydrazine derivatization of one compound with a very poor UV response to determine its levels in our reaction. While the derivatization technique is very sensitive, it is tunnel-visioned in that we can only detect the presence of aldehydes. We have demonstrated the detection of compounds other than aldehydes in our samples using LC–MS. To overcome this detection issue we are developing a CAD that gives us clear visibility of the compounds in the reaction similar to the profile obtained by LC–MS. This allows us to properly evaluate the impurity profile of our reactions and make informed decisions about what the next stage of process development should examine.
The same process offered analytical challenges at earlier steps in the chemistry. One of our early stage reaction products had an excellent UV response but was very sensitive to hydrolysis. This was evident from poor stability data and poor precision testing on reverse phase HPLC. It is also unfortunate that the product in the next step of the process is the hydrolysis product itself. This situation led to poor estimation of reaction conversions and the inability to run prolonged sequences. The compound is also incompatible with GC as its molecular weight prevents it from flying under GC analysis. This presented the opportunity for SFC to shine. The non-aqueous technique delivered a very robust and precise means of analysis for this part of the process. Detection was still achieved by UV, which means excellent sensitivity was achieved for both the compound of interest and the hydrolysis product in the next step.
Aside from the chemistry issues we face like every other laboratory, we are striving to get more and more efficient with our workload and lab activities. The impurity identification and isolation knowledge map referred to earlier gives an example of these efforts. We have taken this approach on a number of fronts in the analytical lab. Another of these approaches was taken with achiral HPLC method development. The workflow developed for this activity begins with a paper assessment of the method development goals.
This paper exercise is then extended to the compound of interest where the pKas and the best modes of detection are assessed. The ensuing screening runs are then tempered by the results of the paper exercise. Only buffers of appropriate pH are selected for the column screen and a selection of orthogonal stationary phases is made. By developing a broad screen on columns and mobile phases we develop the method under QbD principles. The screen can have a large number of runs that is automated using our method development “super system” referred to above. This screening system allows us to identify the appropriate conditions for the separation and therefore points us to the right area of our design space for the method development. This approach to method development has saved us significant time that would otherwise be tied up with an OFAT approach. It also avoids any assumptions around stationary phases that may be made to speed up the OFAT development approach.
A similar approach has been taken with other activities such as genotoxic impurity assessment activities, GC method development, and material characterization. All of these approaches have helped us to make smarter decisions when applying our analytical chemistry. However, one of our greatest assets is our close relationship with the synthetic chemists in our group. This relationship has developed into a two-way connection where the chemist can clearly define the problems that they are encountering and help the analyst to brainstorm and troubleshoot solutions to difficult analytical problems. Equally, the analyst consults the chemists on the findings of the analytical work to give the chemist tangible theories about what is happening in the chemical reactions. This synergistic approach to our work allows us to make well-informed decisions about where our process development goes next and, ultimately, delivers reliable, quality processes to the manufacturing sites.
Q: Where will your research take you in the future?
A: At the moment polymer analysis is a challenging area for us. We have seen that some of our reactions can lose yield if some of the reactants form oligomers or polymers. This is an interesting area I am keen to get more familiar with. I would like to build up a level of knowledge in the analysis of polymers such that we can deliver both qualitative and quantitative results to our chemist. Detection of polymers is not really an issue but their analytical or chromatographic separation is. A lot of this work can readily be done with LC–MS where co-elution can be afforded using the mass spectrometer. However, we have seen polymers that are isomers of each other and developing LC–MS–MS methods for a small portion of our project may not be justifiable. Perhaps this calls out for another Knowledge Map to make the process more efficient and speed up the delivery.
To finish I would like to acknowledge all of my colleagues working in the Process Development Centre. All of the Analysts, Chemists and Engineers working here make the Process Development Centre an exciting, dynamic and enjoyable place to work. The group is always ready for the next challenge and enjoys finding solutions that can really deliver for Pfizer’s customers.
James Frahill graduated BSc (Hons) in Analytical Chemistry with Quality Assurance (ACQUA) from Cork Institute of Technology (CIT). After conducting 5 years of analytical research into mass spectrometric analysis of mycotoxins James began employment for the Pfizer Process Development Centre (PDC) in 2009.