Innovation and trends in analytical development: XRPD
The relationship between powder X-ray diffraction and pharmaceutical R&D goes back a long way. For most of the way, the technique was used for a strictly defined purpose: to identify crystalline phases based on fingerprinting. By the end of the 90s, many patents started expiring and XRPD started experiencing a revival in the pharmaceutical industry. Nowadays, this relatively cheap, simple, fast and nondestructive technique is applied in a surprising number of steps in the development and production of pharmaceuticals . Active Pharmaceutical ingredients (APIs) are routinely undergoing high-throughput polymorph screening and their crystal structures are increasingly solved from XRPD data. Crystal structures of APIs and their XRPD patterns are commonly used as references for polymorph monitoring in stability tests, production, as well as for intellectual property protection. Development of pharmaceutical formulations and finished products often relies on XRPD to determine crystallite domain size of a component in intact solid oral dosages, and to quantify crystalline components, even if they are present in low amounts.
With the modern technological advances, it is not difficult to envisage that XRPD can give an excellent push to the emerging research fields, such as characterization of amorphous systems, understanding of crystallization processes, phase transformations, and “Quality by Design” approach in the development of pharmaceutical formulations . Are we allowed to dream that XRPD will enable us to:
The place where the dream gets closer to the reality is clear: a synchrotron source.
Pharmaceutical R&D and quality control at a synchrotron source? Although eager to fulfill their XRPD dreams, pharmaceutical industry may not find it so easy to get a regular access to a synchrotron source. One can easily imagine a number of reasons for it: nondisclosure requirements, handling of highly-potent and controlled substances, or data processing and interpretation. However, the way synchrotrons are perceived and used by the pharma industry is rapidly changing. The main catalysts in this change are the companies that lower the inertia to access to synchrotron analyses by offering synchrotron radiation-based analytical services.
Dr. Fabia Gozzo (Photo: Scanderbeg Sauer Photography)
Curious about how synchrotron XRPD advances pharma R&D projects, we met with Fabia Gozzo, an expert in the field. The CEO of the company Excelsus Structural Solutions and a former beamline scientist herself, Fabia answered practical questions and highlighted several examples where synchrotron XRPD was used to:
Why does the pharma industry use synchrotron sources?
The majority of pharma companies have an advanced XRPD diffractometer in-house. Alternatively, they use the services of analytical companies, such as Solvias, that guarantee an access to modern laboratory XRPD instrumentation and expertise. There are, however, cases where the quality of the XRPD data collected in a laboratory is insufficient to understand a problem. For example: what causes a deviation in the dissolution rate of a formulation at an early stage of a stability test? In such cases, the XRPD technique needs to be pushed to the extreme. Tasks that are challenging and extremely time-consuming in the lab are easy and fast with synchrotron XRPD, and result in a superior data quality often by several orders of magnitude.
How often do pharmaceutical companies use synchrotron services (either by directly contacting the source or through a company like Excelsus)?
It is hard to give a quantitative answer to this question. Compared to the routine use of laboratory diffractometers, the use of synchrotron XRPD is very small. However, the situation is changing, as companies like Excelsus significantly reduce the difficulty and cost of accessing synchrotron facilities. So, why not access synchrotron XRPD for routine data analysis of superior quality at comparable costs?
Where does your passion for pharmaceutical research come from?
A long time ago, when Bernd Schmitt and his team were developing the MYTHEN detector at the Paul Scherrer Institute, the idea was to have an extremely efficient and fast detector for time-resolved experiments. And, indeed, MYTHEN is capable of performing data acquisition in the microsecond time regime. As a beamline scientist at the Materials Science (MS) beamline of the Swiss Light Source, I had the idea to use the detector for analyzing materials that suffer from radiation damage. The PSI detector group and the beamline staff worked hard on the optimization of the detector and the beamline for fast XRPD measurements. With a series of complex calibrations and data reprocessing, we have managed to achieve unprecedented data quality. In the end, the results were just fantastic.
What is a difference between offering services as a beamline scientist and offering services as a company?
As a spin-off of the Paul Scherrer Institute, Excelsus is tightly connected to its infrastructure and services. Hubbed at the Park Innovaare, we have an easy access to the MS beamline, and we are using the MYTHEN detector for most of our measurements. On top of that, as a company, we make a commitment to solve a problem, either with our own internal resources or through more appropriate resources within our scientific network. We take responsibility for all steps during an industrial project: defining an experimental plan, estimating costs, preparing sample, executing the experiments, reprocessing the data, analyzing the data, and interpreting the results. A beamline scientist can only occasionally do all of this, as her/his tasks include the development and maintenance of complex instrumentation, their own research projects as well as the assistance to a vast scientific user community. Furthermore, as a company, we can guarantee continuity in the development of a project by accessing more than one synchrotron facility. A beamline scientist would not be able to do this.
2. Synchrotron XRPD on Pharmaceuticals
What are the major challenges in modern pharmaceutical R&D?
One of the major challenges is improving the stability of poorly crystalline and amorphous drugs. Amorphous solid dispersions are a smart solution for preventing recrystallization, but the mechanisms that govern such undesired processes are not well understood yet.
What problems in pharmaceutical R&D have you tackled?
Some issues that we have worked on include qualitative and quantitative analysis of traces, support on matters relating to Intellectual Property (IP) and stability studies, such as disproportionation in salts. We have also worked on developing new methodologies for the structural analysis and quantification of non-crystalline materials.
What projects of yours would you highlight as the most relevant or the most successful?
In a general sense, I would point to our ability to perform a reliable quantitative phase analysis of traces in both drug substances and drug products. There have been several specific projects where we have put our expertise at the service of the industry, but I will mention two, which we developed with Dr. Arnaud Grandeury at Novartis Pharma in Basel:
1) We were able to assess the evidence of salt disproportionation in an active ingredient (i.e., the conversion from the ionized state back to its neutral state) that occurs in pharmaceutical formulations, and we did this indirectly, by studying its reaction products of the excipients. The free base that eventually formed was, in fact, amorphous in nature, and therefore not directly accessible in spite of our very low Level of Detection.
2) We directly identified and quantified traces of an unexpected polymorphic form in pharmaceutical drugs that were causing important deviations in dissolution rate tests.
These two cases, which we will soon publish, are also good examples of how efficient collaboration and communication with our industrial partners can enhance the output of our investigations. Without our sense of observation or without the sharp critical spirit of Dr. Grandeury, we might not have found a solution to these problems.
What is the lowest LOD you can achieve?
The achievable level of detection (LOD) is a function not only of the elemental composition of the compound under investigation, but also of its degree of crystallinity. Having said that, we have been able to detect residuals down to 0.01% wt in crystalline pharmaceutical mixtures, while the LOD of a lab source is typically 100 fold higher. See the following article:
Are you involved in the analysis of formulations and finished products?
Yes, very much so, and our ability to perform reliable qualitative and quantitative analysis of traces in pharmaceutical formulations is surely part of our core business.
What is the greatest challenge in performing XRPD at synchrotron sources?
We cope with both small and big difficulties every day. We have to navigate the requirements and complexity of the synchrotron community on one side and the pharmaceutical industry on the other side. In doing all of that, we have to find the solution systematically and reliably every single time.
3. Synchrotron Pair Distribution Function (PDF) analysis of Pharmaceuticals
Why is pharma interested in PDF?
Pharma companies are constantly looking for APIs and formulations that can be patented or have better functional properties (stability, bioavailability, etc.). Poorly crystalline, nano-crystalline and amorphous drugs, which are more soluble than crystalline forms, are very interesting candidates for the new developments. However, their structural characterization requires new approaches, such as Pair Distribution Function (PDF) analysis. This, in turn, requires more complex measurement setups.
How difficult is it to collect and interpret PDF data collected from pharmaceutical samples?
In general, Pair Distribution Function is a histogram that presents the distribution of atom-atom distances in an investigated sample. In order to properly characterize a sample, many distances between many atoms need to be reliably determined. The extent to which one can distinguish atom-atom distances is called PDF resolution and depends on both the measurement and the data correction.
The measurement is in fact a XRPD experiment, performed at high X-ray energies and/or high 2 theta angles. This way, the scattering vector Q can be pushed to high values and thus increase the data resolution. The PDF is obtained by applying a Fourier Transform to the scattered intensities collected in the XRPD experiment. The process is rather complex, as the pattern needs to be correct for all extrinsic incoherent/inelastic contributions.
Compared to inorganic materials, pharmaceutical compounds produce a poor signal at high Q values, and a larger inelastic/incoherent (Compton) scattering. This makes PDF analysis of pharmaceuticals challenging. Even achieving a simple PDF with a good level of reliability can be difficult. Then there is the complexity of interpreting PDF data that is proportional to the complexity of the compound, particularly in the intra-inter-molecular region of the PDF curve. The installment of artifacts, in the form of ripples that often result from a poor correction of the extrinsic and/or inelastic/incoherent contribution to the data and/or the insufficiently high Qmax value (so-called truncation errors) also makes interpreting PDF data particularly difficult, especially for pharmaceuticals.
Which sample types are usually candidates for PDF analysis: APIs or formulations?
In principle, both drug substances and products can be analyzed using the PDF approach, provided that we can achieve a correction or modeling of the placebo’s contribution. Obviously, PDF analysis performed on formulations constitutes a higher level of complexity.
What project of yours related to PDF analysis would you highlight as the most relevant or the most successful?
We have a successful collaboration with the U.S. company Improved Pharma, which specializes in the analysis of amorphous materials via Pair Distribution Function techniques and synchrotron radiation. Through this collaboration, we have been able to expand the field of expertise we can offer to our customers. That same benefit applies to Improved Pharma as well.
4. Detecting the future
The MS beamline at the SLS is equipped with the SAXS/WAXS setup. Will you also offer these services?
Possibly. So far, at the MS beamline, we have been dealing with particles whose size is inappropriate for the specifications of the SAXS-WAXS method, but there might be an opportunity in the future.
You recently started a collaboration with Irene Margiolaki, collecting data on protein powder samples. What is the data quality?
When dealing with proteins, even relatively small ones, the requirements regarding angular (i.e., FWHM) resolution are very high. At Excelsus, we have identified beamline optics conditions that allow us to reach impressive angular resolutions below 30 degrees in 2 theta. With the MYTHEN detector, we were able to collect data of a quality that is comparable to that of data obtained with a multicrystal analyzer, but with much better counting statistics. Prof. Margiolaki was able to detect polymorphic impurities on protein samples that were believed, based on previous measurements, to be pure. A possibility to see this level of detail opens new gateways to the use of synchrotron XRPD on proteins. We are intensifying our collaboration with Prof. Margiolaki and embarking on a long-term common R&D project!
Can one really solve protein structures using XRPD data?
Yes, but it is an art of its own, and it requires years of experience. Prof. Margiolaki has been developing this technique for almost two decades now. Through a tight collaboration with her and her expert team, we intend to bring this great expertise to the industry by identifying and developing well-defined and efficient industrial analytical services.
What is the future of XRPD on proteins? Is the pharma industry ready to adopt these new methodologies?
Pharma companies are extremely attentive to the development of new techniques and methodologies and the modalities of accessing them. Innovation is very important, and there is great awareness of the impact that new techniques can have on the quality of pharmaceutical products. Our first task is to demonstrate that synchrotron XRPD is suitable for the characterization of protein-based drugs, because it provides data with high-resolution and high-counting statistics, and induces minimal radiation damage of the sample. With a good proof of concept, I am confident that we will gain great attention from the pharmaceutical industry.
What are your plans for the immediate future?
Excelsus needs to consolidate its established business and extend it to a larger pharmaceutical community. In parallel, we will always pursue our search for ways to improve the quality of our analytical services and to develop new ones. Right now, our focus is on applying techniques for characterization of biological drugs and biosimilars to proteins. We are also developing advanced methodologies for the quantification of amorphous phases in amorphous mixtures, improving the accuracy of quantification of traces, and working on structural analysis of amorphous drugs. In addition, we have recently started delivering theoretical and practical tutorial lectures on XRPD and data analysis, tailored to the specific needs of the interested company. The positive feedback from the companies to whom we have offered this additional service encourages us to pursue this new project further.
The Excelsus team (left to right): Pamela Whitfield, Mathilde Reinle-Schmitt, Michael Morin, Fabia Gozzo
We wish Dr. Fabia Gozzo and her team all the best in their next projects, and we are certainly looking forward to hearing more success story from this company.
In case you are interested to solve your pharmaceutical problem using the synchrotron radiation and the MYTHEN detector, please contact Excelsus directly. For those who would like to get their non-industrial samples measured at the MS beamline, keep an eye on the call for proposals. Alternatively, check for similar setups at the Aichi, Alba, Australian Synchrotron, Beijing Light Source, Diamond Light Source, LNLS, Shanghai Light Source, Soleil , Spring-8 (BL02B2 and BL44B2), Taiwan Photon Source or Thailand Light Source .
As last, do not forget that the MYTHEN2 detector is implemented in many laboratory diffractometers, and is shown to deliver high quality data. For more information, check the following links:
Diffractometer manufacturers: equipment, test measurements, and fun time:
 Thakral, N. K. et al. (2018) J. Pharm. Sci. 107(12), 1-14.