Pharmaceutical Salts and Co-crystals

By Johan Wouters, Luc Quéré

The Royal Society of Chemistry

Copyright © 2012 Royal Society of Chemistry
All rights reserved.
ISBN: 978-1-84973-158-4

Contents

Chapter 1 Pharmaceutical Salts and Co-crystals: Retrospect and Prospects Gautam R. Desiraju, 1,
Chapter 2 Fundamental Aspects of Salts and Co-crystals Andrew D. Bond, 9,
Chapter 3 Role of Fluorine in Weak Interactions in Co-crystals Seetha Lekshmi Sunil, Susanta K. Nayak, Venkatesha R. Hathwar, Deepak Chopra and Tayur N. Guru Row, 29,
Chapter 4 Polymorph Prediction of Small Organic Molecules, Co-crystals and Salts Frank J. J. Leusen and John Kendrick, 44,
Chapter 5 Shape and Polarity in Co-crystal Formation: Database Analysis and Experimental Validation L. Fábián and T. Friscic, 89,
Chapter 6 Role of Co-crystals in the Pharmaceutical Development Continuum Nate Schultheiss and Jan-Olav Henck, 110,
Chapter 7 Solid Forms and Pharmacokinetics N. Biswas, 128,
Chapter 8 Application of Mechanochemistry in the Synthesis and Discovery of New Pharmaceutical Forms: Co-crystals, Salts and Coordination Compounds Tomislav Friscic and William Jones, 154,
Chapter 9 Co-crystallization in Solution and Scale-up Issues E. Gagnière, D. Mangin, S. Veesler and F. Puel, 188,
Chapter 10 Analytical Techniques and Strategies for Salt/Co-crystal Characterization Susan M. Reutzel-Edens, 212,
Chapter 11 Co-crystal Solubility and Thermodynamic Stability L. Roy, M.P. Lipert and N. Rodríguez-Hornedo, 247,
Chapter 12 Application of Phase Diagrams in Co-crystal Search and Preparation Timo Rager and Rolf Hilker, 280,
Chapter 13 Limits of the Co-crystal Concept and Beyond Gerard Coquerel, 300,
Chapter 14 Co-crystals: Commercial Opportunities and Patent Considerations Marcel Hoffman and Jeffrey A. Lindeman, 318,
Chapter 15 Concluding Remarks using Piracetam as a Learning Model Johan Wouters, Anaelle Tilborg and Luc Quéré, 330,
Chapter 16 Monographs of most Frequent Co-Crystal Formers Johan Wouters, Sandrine Rome and Luc Quéré, 338,
Subject Index, 383,

CHAPTER 1

Pharmaceutical Salts and Co-crystals: Retrospect and Prospects

GAUTAM R. DESIRAJU

Solid State and Structural Chemistry Unit, Indian Institute of Science, Bangalore 560012, India

Every new field in chemistry needs a link to an application of commercial and practical use to sustain interest. Each such field generates a whole new set of ideas, paradigms and models. These concepts need to be tested in as wide a variety of forums as possible because their generality has to be proved. The industrial enterprise has always provided an excellent testing ground for new ideas in the chemical sciences. Many fundamental concepts took root because of an impetus from industry, the most spectacular ones being the discovery of stereochemistry by Pasteur, Haber’s process for nitrogen fixation, and the birth of polymer chemistry starting with the production of synthetic rubber from isoprene. The subject of crystal engineering appeared in its modern manifestation in the late 1980s and early 1990s. Two important branches of this subject emerged. The field of co-ordination polymers quickly found its practical application in the gas absorption properties of metal-organic framework compounds. The field of organic crystal engineering found its practical application, a little later, in the area of pharmaceutical co-crystals and salts. The chapters in this book illustrate the tremendous growth in this area during the past decade.

Co-crystals have been around in the chemical literature ever since Wöhler described quinhydrone in 1844. Pfeiffer’s monumental work in 1922, Organische Molekulverbindungen now has a worthy successor in Herbstein’s two volume magnum opus, Crystalline Molecular Complexes and Compounds, nearly a century later. What crystal engineering did was to provide a context wherein these compounds were given a name, co-crystals, however contentious it was. Secondly, it provided the formalism of the supramolecular synthon with which these compounds could be described, designed and deconvoluted. Finally it defined a setting in which these compounds could be assessed, namely in the pharmaceutical industry. The chapters in this book cover all the above aspects of research.

One might ask “What’s in a name?”. But names are important in chemistry, and there is usually much disputation about chemical names especially in areas that are still evolving. Some of us have defended or questioned the very term co-crystal. Coquerel, in his chapter in the present book, advises us in somewhat harsh terms when he says, “Therefore, rather than creating questionable terms and maintaining endless semantic debates with poor added values, the scientific community should concentrate more on the three long lasting problems …”. Now, I would tend to both agree and disagree with this view. While no discussion about nomenclature should be endless and there are indeed real practical scientific issues that need to be addressed, a part of this discussion about naming something in chemistry is about refining one’s scientific ideas. Names in incompletely developed subjects keep changing and evolving. In the end, a name that survives, survives. Pfeiffer’s Molekulverbindung and Herbstein’s Molecular Compound are one and the same, but they are separated by a century of structural chemistry, an enormous time period in science. Bond’s definition of a co-crystal as a multi-component molecular crystal has the great advantage of scientific accuracy with reasonable brevity but it lacks the pep and vigour of the term co-crystal. At the same time, one should not get carried away with style and ascribe scientific value to the term co-crystal. It has none. It is a bit like the term pseudopolymorph. In a recent perspective article, Bernstein levels criticism at this latter term, as being loose and lacking in scientific value. But the very same criticism can be applied to co-crystal, a term that the same author uses without flinching. My view is that both co-crystal and pseudopolymorph are of similar standing as far as the philosophy of nomenclature is concerned. Neither term has any real scientific value. Both terms are commonly used and in most cases there is little disagreement about whether or not a particular compound is a co-crystal or a pseudopolymorph. They are trivial names, as opposed to systematic names and, as I have written elsewhere, the trivial name sometimes survives. Few call acetic acid ethanoic acid. We shall have to wait and see if the term co-crystal is still extant 100 years from now.

What about pharmaceutical co-crystals? Zaworotko and co-workers defined these as “co-crystals that are formed between a molecular or ionic active pharmaceutical ingredient (API) and a co-crystal former that is a solid under ambient conditions”. Bernstein, in the above-mentioned perspective article, criticises this term too and asks rhetorically if we would call something a pharmaceutical polymorph. But this is begging the question because in calling something a pharmaceutical co-crystal, one is moving away from science and going towards patent litigation. As Hoffman and Lindeman have argued in their chapter, it does not matter that every inventor uses the same definition of what is a co-crystal or, for that matter, a pharmaceutical co-crystal. A patentee may choose to define a term in any way in a patent specification and the patentee’s definition controls the meaning of the term in the claim. For example, a co-crystal may be defined to exclude solvates, salts and amorphs. Therefore, terms in the patent literature not only include certain examples and situations but also deliberately exclude others. The issue of nomenclature is therefore not just a strictly scientific matter but is also of legal concern. The term pharmaceutical co-crystal needs to be looked at from this perspective. Both scientists and patent attorneys need to approach this matter with caution and care.

A major reason for the popularity of pharmaceutical co-crystals in industry is that they lend themselves well to patent protection. They admirably satisfy the three criteria of patentability, namely novelty, non-obviousness and utility. A co-crystal almost always satisfies the novelty criterion because it is a new composition of matter. Non-obviousness is provided by the fact that the identification of the co-former is hardly ever routine, unlike say salt formation wherein an acid is obviously required to make a salt from a base. Utility is generally the only criterion that must be established but it is often easy to demonstrate — usually it is the lack of a particular attribute (solubility, bio- availability, dissolution profile, good shelf life) that has led to the identification of a pharmaceutical co-crystal. With respect to patentability, co-crystals offer opportunities vis-à-vis polymorphs. They are clearly new substances, problems of inherent anticipation are not likely to arise so often and more of them can be made for any given API, expanding the pharmaceutical space around it and consequently the types of advantageous properties that may be accessed.

The issue of non-obviousness is particularly attractive: the design of a co-crystal using synthon theory has all the elements of design and strategy. In this sense, the matter is predictable and the choice of a co-former is neither arbitrary nor random. The necessary experimentation is executable and manageable. At the same time, there is no guarantee that every co-crystal that is designed retrosynthentically will actually be obtained in practice. This lack of inevitability strongly supports non-obviousness and in this respect co-crystals are unlike salts. There are other methods of design. Fábián and Fášcic, in their chapter, discuss shape and polarity descriptors and identify the dihydric phenol orcinol as a promising co-former. Quite independently, in a high-throughput exercise, we have recently confirmed that orcinol is indeed an excellent co-former. High-throughput methods are of obvious relevance. We have stated elsewhere, in a paper on the drug–drug co-crystal of lamivudine and zidovudine that a combination of logic-driven synthon-based design and high-throughput crystallisation is what might actually be needed.

If crystal engineering required a concept known as the supramolecular synthon, pharmaceutical co-crystals needed the heterosynthon. The synthon was sensed before it was identified and named. Similarly, we knew about heterosynthons long before Zaworotko coined the term in 2003. This structural unit was highlighted in the early 1990s as an example of molecular recognition in the very extensively studied melamine–barbituric acid molecular complex and related compounds. Interestingly, this particular co-crystal is described in the chapter by Biswas on pharmacokinetics; it was the cause of lethal poisoning through kidney disease because its solubility is so much less than either of its constituents. Identifying the heterosynthon as a particular type of synthon was important because it helped to focus design strategies for pharmaceutical co- crystals. These ideas are elaborated by Bond in greater detail in his chapter.

A notable aspect of co-crystal research has been the development of high-throughput crystallisation screens. A review from the TransForm group appeared as early as 2004, with the phrase high-throughput crystallisation in the article title. It was recognised that a large number of factors influence crystallisation outcome and that the theoretical basis for predicting such outcomes is poorly developed. Accordingly, there is a need for high-throughput crystallisation methods that sample variables such as temperature, solvent, concentration, additives, vessel design, time, heating and cooling rates, pH and mixing rates. Such research is also important for crystal engineering itself, in a general sense. A nagging worry in any experimental study is that all possible crystal forms of a single- or a multi-component system have not been isolated. Some of the conclusions we draw about structures and structure design could be biased by the fact that we are not dealing with a statistically significant number of examples. This is of even greater concern today, with our just emerging ideas about crystal energy landscapes and structural landscapes, and the notion that a crystal structure of a compound is just that, a data point. It is not the crystal structure of that compound. High-throughput crystallisation, accompanied by the related technique of high-throughput crystallography will go a long way to reduce these concerns.

A drug molecule is identified after a long and arduous process that begins with target identification, virtual screening, lead optimisation and process chemistry. Issues such as solubility, dissolution profiles and bioavailability normally come later in the drug design process. Issues pertaining to absorption, distribution, metabolism, excretion and toxicity (ADMET) are taken up even later. Often, drugs are acids or bases and it is possible to convert them easily into salts, wherein many of the above mentioned properties are favourable. However, current in silico strategies for lead molecule identification and optimisation are biased toward target binding and therefore lead more often than not to the development of lipophilic molecules as drugs. These molecules cannot generally be converted into salts so easily. The chapters by Schultheiss and Henck, and by Biswas address the issue of whether solid form properties need to be taken into account earlier in the drug design continuum. The traditional approach has been to address problems of solubility, dissolution and bioavailability, and ADMET in general at the formulation stage. Schultheiss and Henck state that early identification of a pharmaceutical co-crystal has the potential to minimise the need for multiple changes of the solid form of an API during drug development, which in turn reduces costs directed towards in vitro and in vivo studies. Whether or not solid forms should be taken into account earlier in the drug development cycle, or whether problems relating to solubility and bioavailability should be addressed during formulation, is a complex corporate decision involving scientists, research managers and financial experts. However, the identification of the drug co-crystal as a legitimate extension of pharmaceutical space has certainly increased the flexibility of decision makers in this regard.

The great majority of co-crystals are constructed with strong hydrogen bonds and there is the possibility that the proton involved in the hydrogen bonding interaction is actually transferred from the donor (acid) to the acceptor (base) to form a salt. The salt to co-crystal continuum has been discussed extensively and a common rule of thumb is it that when the difference in pKa between the acid and the conjugate acid of the base is greater than 3 units, salt formation is expected. A number of chapters discuss these salt and co-crystal variations in structure. It is expected that salts are generally more soluble than co-crystals. Whether or not a salt is formed could have patent implications with respect to non-obviousness because salt formation could always be considered to be an operation that could be performed by a person skilled in the art. We found, for example, in a screen of the relatively strong acid saccharin with several basic APIs, that salt formation was the almost inevitable outcome (the solitary exception was piroxicam).

Any crystal structure of a co-crystal or a salt appears to be reasonable in terms of hydrogen bonds and electrostatic interactions, if it is examined post facto. Predicting the interaction patterns for a given API-co-former combination a priori is quite another matter. Crystal structure prediction (CSP) is the most demanding type of crystal engineering because it seeks to predict fine details of crystal packing. The addition of salts, co-crystals and hydrates (generally multi-component crystals) to the list of compounds given in the Cambridge Crystallographic Data Centre (CCDC) sponsored blind test for CSP has significantly added to the difficulty of this exercise. Gratifyingly it was found that in the latest (2010) blind test, good success was obtained in the CSP of both salts and hydrates. Leusen and Kendrick, who were part of a group that made a number of accurate predictions, have reviewed this topic in their chapter. We found, in an earlier blind test, that it is necessary in CSP to know whether or not a particular binary crystal exists as a co-crystal or a salt before one embarks upon rigorous computation.

CSP is closely allied to the subject of polymorphism. A frequently asked question is whether or not a pharmaceutical co-crystal is less likely to exhibit polymorphism when compared to the API itself. In a more general context, are co-crystals less prone to polymorphism than single component crystals? The current consensus seems to be that the number of examples available at hand is far too small to permit any reliable answer to such questions. That a co-crystal is inherently less prone to polymorphism than is a single component crystal appears to be intuitive, but this hypothesis cannot have a direct answer because to prove the absence or lesser incidence of polymorphism is tantamount to proving the negative. More specific is the matter of synthon polymorphism in a co-crystal. Various types of polymorphism are possible but the most direct manifestation of this phenomenon occurs when the primary synthons in the forms are different. Here, the data, although limited, seem to indicate a trend. Zaworotko and co-workers report that, of the 38 pairs of polymorphic organic co-crystals in the Cambridge Structural Database (CSD), polymorphism in 35 of these cases is a result of conformational flexibility and/or minor structural changes in the packing. The hydrogen bonded heterosynthons persist. True synthon polymorphism, or a deep seated difference at the core of the structure, occurs only thrice. In this context, we isolated synthon polymorphs in a fourth example, namely the 1:2 co-crystals of 4,40-bipyridine and 4-hydroxybenzoic acid.

(Continues…)Excerpted from Pharmaceutical Salts and Co-crystals by Johan Wouters, Luc Quéré. Copyright © 2012 Royal Society of Chemistry. Excerpted by permission of The Royal Society of Chemistry.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.