Reduced Crystallization Temperature Methodology for Polymer

Subscriber access provided by NJIT | New Jersey Institute of Technology

Brief Article

Reduced Crystallization Temperature methodology for polymer selection in amorphous solid dispersions – Stability Perspective Chandan Bhugra, Chitra Telang, Robert Schwabe, and Li Zhong Mol. Pharmaceutics, Just Accepted Manuscript • DOI: 10.1021/acs.molpharmaceut.6b00315 • Publication Date (Web): 14 Jul 2016 Downloaded from http://pubs.acs.org on July 20, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Molecular Pharmaceutics is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Reduced Crystallization Temperature methodology for polymer selection in amorphous solid dispersions – Stability Perspective Chandan Bhugra*, Chitra Telang, Robert Schwabe, Li Zhong

Pharmaceutical Development, Boehringer-Ingelheim Pharmaceuticals Inc., Ridgefield, CT. * Corresponding author Chandan Bhugra, Ph.D. Boehringer-Ingelheim Pharmaceuticals Inc. 900 Ridgebury Road, Ridgefield, CT - 06877 Email: [email protected] Tel: +1-203-798-5174 Fax: +1-203-791-6197

1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 24

Abstract: API-polymer interactions, used to select the right polymeric matrix with an aim to stabilize an amorphous dispersion, are routinely studied using spectroscopic and/or calorimetric techniques (i.e. melting point depression). An alternate selection tool has been explored to rank order polymers for formation of stable amorphous dispersions as a pragmatic method for polymer selection in a resource friendly manner. Reduced crystallization temperature of API, a parameter introduced by Zhou et.al.1, was utilized in this study for rank-ordering interactions in API-polymeric systems. The trends in reduced crystallization temperature monitored over polymer concentration range of up to 20 % polymer loading were utilized to calculate “crystallization parameter” or CP for two model systems (nifedipine and BI ABC). The rank order of CP, i.e. a measure of API-polymer interaction, for nifedipine followed the order PVP > PVP-VA > Soluplus > HPMCAS > PV Ac > PAA. This rank ordering was correlated to published results of molecular interactions and physical stability for nifedipine. A different rank ordering was observed for BI ABC: PAA > PVP > HPMCAS > Soluplus > PVPV-VA > PVAc. Interactions for BI ABC were not as differentiated as for nifedipine as observed in CP trends. BI ABC dispersions at drug loadings between 40 – 60 % were physically stabile for prolonged periods under ICH conditions as well as accelerated stress. We propose that large CP differences among polymers could be predictive of stability outcomes. Acceptable stability at pharmaceutically relevant drug loadings would suggest that the relative influence of downstream processes, such as polymer solubility in various solvents, process suitability & selection, and more importantly super-saturation potential, should be higher compared to stability considerations while developing compounds like BI ABC.

Key Words: Amorphous, Crystallization, Physical Stability, Reduced Crystallization Temperature, Polymer Selection Introduction: The amorphous state of a drug substance is a metastable high energy state which is attractive due to its higher kinetic solubility when compared with its crystalline counterpart, thereby improving chances of better bioavailability through improved dissolution. This finds application when developing BCS Class II/IV compounds, i.e. compounds with high dose and poor aqueous solubility of the API (active pharmaceutical ingredient), where use of crystalline drug may not result in clinically relevant exposure. While this is the main rationale for selecting an amorphous route for dosage form, it is common 2 ACS Paragon Plus Environment

Page 3 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


knowledge that there are complexities associated with developing drug candidates as amorphous products, one among them being tendency to crystallize. Due to its high energy, the amorphous state has increased reactivity both chemical and physical (re-crystallization tendency; either during stability or dissolution), which may offset the benefits in certain scenarios

2,3

. Since physical instability is generally

not predictable, a significant effort has been put into this area of science and has resulted in an increase in research publications investigating methods to mitigate risks associated with physical instability4-6. Much literature is available with predictive tools for both single and multi-component amorphous systems5,7. However more recently, efforts have been directed to studying physical stability of amorphous dispersions that could be developed as drug products

8-10

. Amorphous dispersions are

relatively more stable than the amorphous API alone with contributions from interaction of API with the polymeric carrier or kinetic stabilization. Interactions which are mostly due to hydrogen bonding can increase the kinetic (through reduced molecular mobility) and the thermodynamic barrier to crystallization thereby increasing stability. An increase in understanding of factors affecting physical stability of amorphous dispersion however needs to be balanced with understanding of stability during in vivo dissolution. Therefore selection of the right polymeric matrix for API is important to successfully advance the amorphous phase in drug product development. Interactions between components of the amorphous dispersions are routinely studied via FTIR and RAMAN spectroscopy, where shifts in peak positions or formation of shoulders & changes in peak intensity and shape for interacting groups, provides a means to quantitatively assess interaction11,12. Although such changes in the IR spectra can be used to indicate interaction, it is also known that interpretation of IR data is subjective and may require experience to draw conclusions 13. As a result, use of the above techniques during early stages of development has limited applicability and scientists rely on real time stability testing to select dispersion formulation composition; which defeats the objective of pre-selecting the right matrix based on interaction. Eerdenbrugh et. al. suggested that by rankordering hydrogen bond donors and acceptors, one could qualitatively select interacting polymer-drug combinations for formation of stable amorphous dispersions6. This approach could possibly provide a means to select interacting components in early development. Other methods suggested in the literature include measuring miscibility of the API with the polymer and/or solubility of API in polymers at higher temperatures utilizing differential scanning calorimetry to measure drug dissolution in polymers as a means to assess interaction14,15. Although promising, experimental measurement of solubility and/or miscibility is not straight forward, often requiring extrapolation of high temperature 3 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 24

data to project behavior at room temperature. The objective of this research therefore, is to explore an alternative selection tool for rank ordering polymers for formation of stable amorphous dispersions. Nifedipine and an internal BI compound (referred as BI ABC) were selected as model compounds for this work (See table 1 for structures). Nifedipine was selected as published research provides background information on interactions of nifedipine with various polymers (measured via IR, melting point depression) and also its physical stability in some of the selected polymer systems

13,16,17

.

Additionally since structurally both nifedipine and felodipine are similar (both belong to chemical class of dihydropyridines and therefore have the same backbone structure with NH type donor and C=O acceptor groups), one would expect similar interaction between either molecule and polymers, providing an even broader database on documented interactions 13. Some examples are captured below. Sun et.al. published a study describing the solubility of nifedipine and indomethacin in various polymers 16

. In their study, drug-polymer interactions were analyzed (via scanning DSC) using melting point

depression of the drug in the presence of polymer. The authors concluded that for both indomethacin and nifedipine, polyvinylpyrrolidone (PVP) had the most dissolving capacity which was followed by polyvinylpyrrolidone-vinylacetate (PVP-VA64) and polyvinyl acetate (PVAc) in that order. The choice of polymers in the above study suggests the importance of polymer structure in dictating the level of interaction as measured via melting point depression, i.e. interaction with PVP-VA64 (a copolymer of PVP and PVAc) was intermediate to that of PVP and PVAc. Results from a more recent publication, by Khushboo, et.al., ranked the molecular mobility, polymer interactions, and physical stability of nifedipine dispersions as measured by dielectric spectroscopy, FT-IR, and XRPD respectively. The authors concluded that the drug polymer hydrogen bonding, the structural relaxation time and the physical stability followed the order: PVP > HPMCAS > PAA with PVP systems forming the most stable dispersions due to strong hydrogen bonding and therefore long structural relaxation times17. In another study, Rumondor et.al., evaluated drug polymer miscibility for felodipine with two polymers namely, polyvinylpyrrolidone (PVP) and polyacrylic acid (PAA) and concluded that interactions in felodipine-PVP were much stronger when compared with the interactions in felodipine-PAA system13. As expected from similar structural motifs in nifedipine and felodipine, the interaction rank ordering for these APIs appears to follow the same pattern. The authors studied interactions using IR, DSC and PDF analysis. Another study complemented the findings of the Rumondor study, where the data showed felodipinePAA system as phase separated, i.e. samples crystallized during the course of the study and were therefore less physically stable when compared to the felodipine-PVP system18. The results from these


Page 5 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


various studies provide a basis of selecting nifedipine as our model API with the following rank order for stability, deduced from literature PVP>PVP-VA 64> HPMCAS~PVAc>PAA. In addition, we selected another model compound BI ABC in order to apply the method to a development candidate with different expected structure-based interactions (see Figure 1) with polymers. BI ABC was also chosen because of its high crystallization potential, a relatively modest Tg (glass transition temperature) of 77°C, and a reduced crystallization temperature (discussed later) of 0.4 (and similar to that of nifedipine). When compared to nifedipine both compounds have a similar Tg /Tm (Tm is melting point) ratio of ~0.71 and both compounds crystallize rapidly from the amorphous state. In order to reduce risk of physical instability, we developed a rapid polymer selection method with an aim to provide a relative rank order of beneficial interactions between different combinations of API and polymers that could potentially impact long-term stability. Towards this objective, we utilized the parameter developed by Zhou et.al.1 In their study, Zhou introduced a parameter, reduced crystallization temperature, to rank order crystallization tendency of an amorphous API using data from a simple DSC scan. By measuring the temperature of crystallization onset on re-heating (Tc) and crystal melt (Tm), relative to the Tg of the API they ranked the relative ease or difficulty of crystallization from the amorphous state via a kinetic (non-isothermal) DSC experiment. In their study, the authors also complemented the data on ease of crystallization, i.e. reduced crystallization tendency, to thermodynamic and kinetic characteristics of the compound (namely the configurational free energy and entropic barrier to crystallization). We wanted to extend the utilization of reduced crystallization temperature (Rc) tool to API polymeric systems and explore if this approach could be used for rank ordering drug-polymer interactions from a stability perspective. Given that the measurement is nonisothermal, the onset temperature for the crystallization event in presence of polymer in comparison to the event from the amorphous API alone or with that in systems of different polymers can provide information reflecting API-polymer interactions across polymers which will result in different positioning of thermal events. These events are driven by API/polymer structures, Tg of polymer and miscibility of API in polymer. Note the Tm in mixed systems will be influenced by dissolution of crystal in molten polymer and this phenomenon is also captured in the Rc measurement. Additionally ease of crystallization, i.e. kinetics, as reflected in relative positioning of crystallization exotherm from the Tg of the system indicates stability of the API in polymer matrix in supercooled state. It is important to note that the applicability of this methodology would be limited to low polymer loadings, as at higher polymer concentrations, crystallization from the amorphous state would 5 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 24

be completely hindered during the time span of DSC experiment. Presence of beneficial interaction, such as hydrogen bonding or ionic interactions, are expected to further inhibit this process by indirectly impacting the molecular mobility in a given system. In this study, by utilizing nifedipine and BI-ABC as model compounds at high API loadings, we have explored the utility of reduced crystallization temperature to rank order API-polymer interaction. Further, by selecting nifedipine as a model system we were able to compare our findings with those in the literature using other well established techniques such as IR spectroscopy and miscibility analysis that could lend support to our observations. Materials: Crystalline nifedipine was purchased from Sigma Aldrich. BI ABC was sourced internally. The melting point for nifedipine and BI ABC is 172˚C and 207.5˚C respectively. Nifedipine and BI ABC have a glass transition temperature of 43.9˚C and 73.6˚C as determined via standard DSC measurement when heating at a rate of 2˚C/minute. Poly acrylic acid (PAA) and Poly vinyl acetate (PvAc) (MW ~100,000, glass beads) were purchased from Sigma-Aldrich. Povidone (PVP, Kollidon 25), Copovidone (PVPVA 64) and Soluplus were purchased from BASF. HPMCAS-MF was purchased from Shin-Etsu. All compounds were used without further purification.

Methods: Preparation of API/ Polymer Mixtures: API/ polymer mixtures were prepared using the solvent evaporation method. This method was used to intimately mix the API and polymer; the amorphous nature of the resulting mixture was not relevant at this stage. First, nifedipine and polymer were dissolved in 50 mL methanol at a 10% solids loading. 10 mL of acetone (and 40 mL methanol) was used for HPMCAS mixtures. A water bath was used to dissolve all solids before solvent evaporation. For mixtures with PAA loadings greater than or equal to 20%, water was added to dissolve the polymer. The same process was used for BI ABC with polymers The solvent was evaporated on a Buchi Rotovap and API-polymer solution was run to dryness using a temperature of 75°C and rotovap speed of 80 rpm. The resultant solid was dried overnight under conditions of vacuum and temperature (60°C) to remove remaining solvent. Differential Scanning Calorimetry:


Page 7 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


DSC was used in standard mode to determine the glass transition temperature (Tg), melting temperature (Tm), and the peak crystallization temperature (Tc) for all dispersions. For this purpose, the solvent evaporated samples (as described in the previous section) were heated in a pinhole DSC pan to a temperature above the Tm, to obtain an amorphous sample with no residual crystallinity. The melt was then quenched to below room temperature and heated at a constant rate of 2˚C/minute to determine Tc, Tg, and Tm of the melt quenched dispersion thus obtained. Reduced crystallization temperature (Rc) was analyzed according to the equation 1 1:

ܴܿ =

்௖ି்௚ ்௠ି்௚

Equation 1

Where Rc is the reduced crystallization temperature. Note, all experiments were run in triplicate and the reported data is an average of three independent DSC runs. X-Ray Powder Diffraction (XRPD) X-Ray powder diffraction (XRPD) was conducted using a Bruker AXS X-ray Powder Diffractometer (Discovery D8) with a CuKα X-ray source and a Vantec 2000 detector. For each sample, XRPD pattern was collected from 2θ of 5° to 40° using a total of 2 minutes scan. Stability Studies The stability study of dispersions was performed over 6 months, under 40°C/75%RH conditions with controlled humidity by selecting appropriate packaging configuration (in this case poly-propylene bottles with desiccant). Similarly, stability was also monitored at 25°C/60%RH for a period of up to 3 years. Additionally short term stress stability testing was performed at 50 ˚C at various humidity conditions. Results and Discussion: Reduced Crystallization Temperature of nifedipine as a function of polymer concentration: Figures 1 and 2, show DSC thermograms for nifedipine-PVP system and nifedipine-PAA system respectively, with increasing polymer loading. It can be seen that on increasing PVP concentration in nifedipine-PVP dispersions, the crystallization peak temperature increases from ~ 92˚C to ~ 145˚C for 2% and 20% polymer loading respectively. In contrast, the crystallization peak temperature in nifedipinePAA system, for the same polymer concentration, only changes from 92˚C for dispersion containing 2% polymer to 95˚C for a dispersion containing 20% polymer.

Polymorphism in nifedipine result in 7

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 24

additional thermal events of conversion to more stable polymorph either during recrystallization or prior to melt and are seen in these thermal profiles to various extents with different polymer composition 17. The prominent effect on Tg increase in PVP systems has also been marked in figure 1, clearly differentiating it from a negligible increase of Tg in the PAA system as marked in figure 2. This can be primarily attributed to the different strength of interactions in the two dispersion systems compared. This comparison of the interactions has been described well in a recent publication from Khushboo et. al. where the authors note strong hydrogen bonding between nifedipine-PVP system and no hydrogen bonding interaction between nifedipine-PAA system.17 The PAA system, therefore, show a single Tg confirming single amorphous phase with negative deviation of observed Tg 19. We attribute the small trend in Tc noted in the DSC thermograms for nifedipine-PAA system at 20 % polymer loading to dilution effect that the polymer may exert on the API molecules resulting in a small but noticeable delay in Tc peak. Overall, one could conclude looking at this data, that nifedipine-PVP dispersion (80:20) with a reduced crystallization temperature of 0.81 has potential to form more a physically stable dispersion than the nifedipine-PAA dispersion (80:20) with a reduced crystallization temperature of 0.40. In order to explore if the above trends in reduced crystallization temperature can be utilized to provide a rank ordering of stabilization potential of a polymer, we expanded our list of polymers to test the hypothesis. In addition to PVP, and PAA, we included PVP-VA64, Soluplus, HPMCAS, and PVAc. Note: PAA and PVAc, have not been used in drug product development as also shown by a literature survey of amorphous drug development projects

20

. Figure 3 shows the trend in reduced crystallization

temperatures as a function of polymer concentration for four of the selected polymers for nifedipine (entire data set is included in Table 2). As evident from the data in the plot (see discussion above), PVP has the most potential to retard crystallization of nifedipine when compared with dispersions with other polymers. In contrast, PAA and PVAc are unable to delay the crystallization event as effectively as the other two polymers. HPMCAS (MF Grade) on the other hand has an intermediate potential. The circle on the y-axis represents reduced crystallization temperature for nifedipine alone. Crystallization trends with all systems (nifedipine-polymer) up to the limiting polymer of about 20% concentration were observed to be linear with excellent correlation coefficients. Therefore a linear fit between reduced crystallization temperature and percent polymer loading with various polymers was used to define a ‘crystallization parameter’ or CP for a binary system. We use this term to quantitatively rank order polymer choices. The data in the Table 2 for nifedipine, shows that for three of the six polymers, i.e. PVP, PVP-VA64, and soluplus have approximately the same CP, followed by HPMCAS, PVAc, and PAA in that order. Interestingly PVP, PVP VA64, and soluplus also have similar functional groups which may be 8 ACS Paragon Plus Environment

Page 9 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


responsible for similar interaction and therefore CP for nifedipine. Note that the R2 values for the linear fit are very close to 1 increasing the confidence in CP. In fact, our rank-ordering correlates well with the results from a recent study which used FT-IR and structural relaxation times to arrive at the same conclusion17. The authors, based on FT-IR data, noted shoulder peaks at lower wavenumber for PVP and at higher wavenumber for HPMCAS when compared with nifedipine alone. No interaction was noted with the PAA system. Knowing that a shoulder at lower wavenumber indicates stronger API-polymer interaction than the intermolecular interaction between drug molecules the authors concluded that the interactions of these polymers with nifedipine follow the rank order: PVP > HPMCAS > PAA. Additionally, based on the work of Sun et.al.; where solubility and Flory-Huggins interaction parameters of nifedipine in various polymers systems was measured16, the authors conclude that, PVP was the best solvent for nifedipine, followed by PVP-VA64, and PVAc. Note that solubility of API crystals in polymers is an indicator of the interaction with the polymer and therefore the thermodynamic driving force for stability against crystallization. Their solubility data correlated very well with the measured drugpolymer interaction parameters (Flory-Huggins Model) and range from -2.5 for PVP to -0.02 with PVAc suggesting that the API/polymer interaction is much stronger with PVP when compared to that with PVAc. Based on the discussion above, it appears that the interaction rank ordering is similar to the rank ordering observed in this study via reduced crystallization temperature experiments. Reduced Crystallization Temperature of BI-ABC as a function of polymer concentration: The current work was extended to another structurally unrelated proprietary compound BI ABC. The plot of reduced crystallization temperature as a function of polymer loading can be seen in figure 4. An interesting rank ordering trend emerges for BI ABC where polymer PAA (least effective for nifedipine) interacts very strongly with BI ABC. Also contrasting with nifedipine, are the lower numerical values of CPs for most BI ABC polymer dispersions. Significant changes in reduced crystallization temperature were not observed with BI ABC and some polymers, as a function of change in polymer concentration, thereby resulting in scatter and thus lower CP values. An examination of CP for different polymers shows that PAA is the only polymer that is differentiated. In fact, it is important to note that it was only with PAA (at 20% polymer level) that an amorphous dispersion resulted after solvent evaporation with BI ABC; an indication that the polymeric stabilizer (PAA) interacted with API to cause stabilization of the dispersion, an observation not made with other polymer systems for BI ABC. 9 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 24

BI-ABC is a weak base with a pKa of 2.5. Structurally the molecule contains a sulfoxide group, a carbonyl group and a fluorine moiety (all acceptor groups) and a NH group (donor) as shown in Table 1. The three acceptor groups would make the BI-ABC molecule primarily an acceptor molecule and one would anticipate BI-ABC to interact well with polymers having good donor potential. PAA is the only polymer that has groups which are strong donors and all other polymers have functional groups which are either strong or medium level acceptor groups6. Based on the data and discussion presented thus far, it appears that reduced crystallization temperature measurement can be an alternative methodology for selection of polymers in early drug development. The experiment is material sparing, and simple to perform. As demonstration of long-term ICH stability is a pre-requisite in formulation development, we tested the Rc method for its predictive capability by correlating to a 2 year long term stability outcome. In addition, stress stability testing at high temperature and humidity was also performed for selected systems. Correlation to physical stability: (a) Long term stability study: A survey of literature data shows lack of emphasis on stability testing at industrially practical drug loadings, and conditions reflecting real time stability with practical packaging configurations in order to accelerate crystallization. In this work, we aimed to overcome this gap and performed stability testing at more relevant drug loadings to evaluate if such thermal method correlate to long term stability for solid dispersions with drug loadings and packaging configurations used in formulation development. Table 3a and Figure 5 present long-term stability data for BI ABC dispersions prepared with PVP, PVP-VA64, and HPMCAS at 60% drug loading. The stability study of dispersions was performed over 6 months at 40°C/75%RH (Table 3a) and over 35 months at 25°C/60%RH (Figure 5), with dispersions stored in polypropylene bottles in presence of desiccants. As shown in the data, no crystallization was observed during the timeframe of the study. The stability data suggests that even polymers (such as HPMCAS, PVP, and PVP-VA64) with undifferentiated CP values in the range of 0.0028 to 0.0095 can stabilize dispersions (see figure 5). Note the stability study was performed as part of development work with pharmaceutically suitable polymers and therefore no long term stability data with PAA is presented. However, as mentioned previously an indication of enhanced stability with PAA was obtained during dispersion preparation itself wherein only the BI ABC – PAA dispersion was a single phase system after solvent evaporation (whereas others dispersion systems required cycling through the melt). 10 ACS Paragon Plus Environment

Page 11 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In this work nifedipine was not set up for stability due to availability of short term stress stability and long term stability on this model compound in the literature17,18. For example: Caron et. al. showed that nifedipine:PVP (95:5) dispersion was stable at 25 °C and 35 °C (under desiccation) for more than 200 days8. Igor Ivensevic in his work, published stability data of nifedipine-PVP dispersions wherein stability was monitored at ambient temperatures at 60% drug loading for up to 400 days without observation of crystallization. The data shows that Nifedipine-PVP systems (interacting, with high CP) form physically stable dispersions. More recently, Khushboo et.al. monitored kinetics of crystallization for nifedipine dispersions with PVP, HPMCAS, and PAA at 10 % w/w polymer loading at 70 ˚C. The authors concluded that nifedipine-PAA system fully crystallized in about 400 minutes whereas crystallization was about 40 % and about 15 % complete for HPMCAS and PVP based dispersions respectively17. This published stability data for nifedipine-polymer systems correlates with the thermal rank ordering observed using CP in this study. Although a trend for CP was seen with BI ABC, the numerical values for CPs in individual polymer systems were not as differentiated as seen in the nifedipine-polymer systems. Therefore, individual Rc values were also examined. For a given polymer concentration the values of Rc are not very different across both drugs for polymers that are practically used in industry to develop solid dispersions (Rc for PVP, PVP-VA64, HPMCAS, and soluplus lies between 0.51 to 0.62 at 10 % polymer loading for both BI ABC and nifedipine dispersions). This indicates that the extent of stabilization provided by polymer is expected to be similar since the amorphous APIs had similar Rc values. Furthermore for BI ABC, a significant change in Rc as a function of polymer concentration was also not observed for these polymers (see Table 2) resulting in undifferentiated and negligible CP values. Therefore, similar long term stability trends are anticipated for BI ABC dispersions. (b) Short term humidity stress stability: As samples over long-term storage were not showing trends for polymer differentiation for BI-ABC, a short-term stress stability study was also conducted at different drug loads and humidity for a period of up to a week with the polymers PVP, PVP-VA64 and HPMCAS-L. An in-house qualitative ranking from AAAA (completely amorphous as evidenced by lack of XRPD peaks or PLM birefringence) to CCCC (clearly crystalline) was used to tabulate results (Table 3b). A few trends emerged from this study and have been summarized for 60% drug load namely, (i) at 75 % RH, dispersion with PVP crystallizes fastest; (ii) in general, a drop in Tg from dry SDD followed the order of PVP > PVP-VA64 > HPMCAS. This observation was most noticeable for 51% RH and is a result of polymer hygroscopicity which is known to 11 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 24

be highest for PVP and least for HPMCAS. In presence of humidity, Tg decrease is most conspicuous for PVP containing dispersions and Tg being the indicator of global mobility, would therefore make the PVP systems most mobile and prone to crystallization as also confirmed in the study. This phenomenon is well known in the area of amorphous pharmaceuticals and therefore the results of humidity stress are not elaborated further. No correlation of crystallization was observed with CP due to the overwhelming influence of polymer hygroscopicity. As expected, when humidity drops (in this case, stability results at 30% RH), polymers are no longer differentiated within the duration of study and this observation is correlated to non-differentiating CP results which were also obtained in long term stability study. Similar results were also obtained in another high temperature stress study (at 60˚C) for two months where no differentiation from physical stability perspective for the three BI-ABC-polymer systems studied was observed (data not shown). Overall, we show that the interaction rank ordering developed using our methodology, i.e. crystallization parameter or CP, matches with the crystallization behavior shown for nifedipine via high temperature stress stability testing17. Available data for both nifedipine (with PVP)8,18 and our data on BI-ABC (with PVP, PVP-VA64, and HPMCAS) show that on long-term storage, physically stable dispersions were produced. The available long term stability results are reflected in the nondifferentiating CP values for BI-ABC and similar Rc values for these polymer-API systems at a given polymer concentration. While PAA and PVAc would lend additional support/strengthen the predictability of this thermal method, these polymers were not included in the study due to their irrelevance to pharmaceutical solid dispersions. These similar CPs and acceptable stability as shown at pharmaceutically relevant drug loadings emphasize that the relative influence of downstream processes, such as polymer solubility in various solvents, process suitability & selection, and more importantly super-saturation potential, should be higher when compared to stability considerations for compounds like BI ABC.

Summary: A simple thermal method has been presented to rank-order polymers for amorphous solid dispersions. Crystallization parameter (CP) monitored via reduced crystallization temperature offers a semiquantitative alternative to spectroscopic methods to probe interactions. We report a correlation for nifedipine-polymer CPs to physical stability as assessed in short-term high temperature stability study 12 ACS Paragon Plus Environment

Page 13 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


and the available long term stability data for nifedipine-PVP system. No crystallization was noted with BI ABC dispersions over a 3 year time frame under ICH conditions. We speculate this to be the result of lack of differentiation in CP values among the three polymers used in this stability study. PVAc and PAA polymer dispersions were not part of the development goals for BI ABC (as these polymers not widely used in pharmaceutical dispersions for various reasons) and therefore not monitored over long term stability. Comparing across BI ABC and nifedipine, we believe as CP differences become larger more differentiation in stability outcomes may be expected. Therefore, this thermal approach may be applied in eliminating polymer choices when CP differences are markedly different. Although this work was limited to only two model compounds, extending this thermal approach to include structural diversity will test and strengthen its applicability.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 24

Reference List

1. Zhou D, Zhang GGZ, Law D, Grant DJW, Schmitt EA 2002. Physical stability of amorphous pharmaceuticals: Importance of configurational thermodynamic quantities and molecular mobility. J Pharm Sci 91:1863-1872. 2. Murdande SB, Pikal MJ, Shanker RM, Bogner RH 2011. Solubility advantage of amorphous pharmaceuticals, part 3: Is maximum solubility advantage experimentally attainable and sustainable? J Pharm Sci 100:4349-4356. 3. Bhugra C, Pikal MJ 2008. Role of thermodynamic, molecular, and kinetic factors in crystallization from the amorphous state. J Pharm Sci 97:1329-1349. 4. Bhugra C, Rambhatla S, Bakri A, Duddu SP, Miller DP, Pikal MJ, Lechuga-Ballesteros D 2007. Prediction of the onset of crystallization of amorphous sucrose below the calorimetric glass transition temperature from correlations with mobility. J Pharm Sci 96:1258-1269. 5. Van Eerdenbrugh B, Baird JA, Taylor LS 2010. Crystallization tendency of active pharmaceutical ingredients following rapid solvent evaporation - Classification and comparison with crystallization tendency from undercooled melts. J Pharm Sci 99:3826-3838. 6. Van Eerdenbrugh B, Taylor LS 2011. An ab initio polymer selection methodology to prevent crystallization in amorphous solid dispersions by application of crystal engineering principles. CrystEngComm 13:6171-6178. 7. Bhugra C, Shmeis R, Krill SL, Pikal MJ 2008. Prediction of onset of crystallization from experimental relaxation times. II. comparison between predicted and experimental onset times. J Pharm Sci 97:455-472. 8. Caron V, Bhugra C, Pikal MJ 2010. Prediction of onset of crystallization in amorphous pharmaceutical systems: Phenobarbital, nifedipine/PVP, and phenobarbital/PVP. J Pharm Sci 99:3887-3900. 9. Kestur US, Van Eerdenbrugh B, Taylor LS 2011. Influence of polymer chemistry on crystal growth inhibition of two chemically diverse organic molecules. CrystEngComm 13:6712-6718. 10. Van Eerdenbrugh B, Taylor LS 2010. Small scale screening to determine the ability of different polymers to inhibit drug crystallization upon rapid solvent evaporation. Mol Pharm 7:1328-1337. 11. Taylor LS, Langkilde FW, Zografi G 2001. Fourier transform Raman spectroscopic study of the interaction of water vapor with amorphous polymers. J Pharm Sci 90:888-901. 12. Tang XC, Pikal MJ, Taylor LS 2002. A spectroscopic investigation of hydrogen bond patterns in crystalline and amorphous phases in dihydropyridine calcium channel blockers. Pharm Res 19:477483.


Page 15 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


13. Rumondor ACF, Ivanisevic I, Bates S, Alonzo DE, Taylor LS 2009. Evaluation of drug-polymer miscibility in amorphous solid dispersion systems. Pharm Res 26:2523-2534. 14. Tao J, Sun Y, Zhang GGZ, Yu L 2009. Solubility of small-molecule crystals in polymers: D-Mannitol in PVP, indomethacin in PVP/VA, and nifedipine in PVP/VA. Pharm Res 26:855-864. 15. Marsac PJ, Li T, Taylor LS 2009. Estimation of drug-polymer miscibility and solubility in amorphous solid dispersions using experimentally determined interaction parameters. Pharm Res 26:139-151. 16. Sun YE, Tao J, Zhang GGZ, Yu L 2010. Solubilities of crystalline drugs in polymers: An improved analytical method and comparison of solubilities of indomethacin and nifedipine in PVP, PVP/VA, and PVAc. J Pharm Sci 99:4023-4031. 17. Kothari K, Ragoonanan V, Suryanarayanan R 2015. The Role of Drug−Polymer Hydrogen Bonding Interactions on the Molecular Mobility and Physical Stability of Nifedipine Solid Dispersions. Mol Pharm 12:162-170. 18. Ivanisevic I 2010. Physical stability studies of miscible amorphous solid dispersions. J Pharm Sci 99:4005-4012. 19. Matsumoto T, Zografi G 1999. Physical properties of solid molecular dispersions of indomethacin with poly(vinylpyrrolidone) and poly(vinylpyrrolidone-co-vinyl-acetate) in relation to indomethacin crystallization. 16:1722-1728. 20. Newman A, Knipp G, Zografi G 2012. Assessing the performance of amorphous solid dispersions. J Pharm Sci 101:1355-1377.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 24

Table 1: Model compounds (BI ABC and Nifedipine) with structures showing donor and acceptor groups along with information on interacting polymers based on the literature 6

Model Compound

Key Hydrogen Bonding CONTIBUTOR Groups (API)

Compound BI ABC Tg = 77 C

ACCEPTOR

Nifedipine Tg = 42 C

Interacting Polymer Based on Groups Polymers: Strong Acceptor Groups in PVP & PVP VA will form good hydrogen bonding with reasonable donors in selected compounds Medium strength acceptor Groups in HPMCAS will form relatively weaker hydrogen bonding with reasonable donors in selected compounds PAA have strong donor groups and therefore may not interact with nifedipine however may have interaction potential with Compound A with a reasonable acceptor molecule

DONOR


Page 17 of 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 2: Crystallization Parameter and R2 values for Nifedipine and BI-ABC with selected polymers as determined from a linear fit of the plot between reduced crystallization temperature and percent polymer loading. Crystallization parameter is also the slope of the linear fit. Polymer

Nifedipine

BI-ABC

Crystallization Parameter

R2

Crystallization Parameter

R2

PVP-K25

0.0247

0.9925

0.0095

0.9558

PVP VA64

0.024

0.9778

0.0028

0.5423

HPMCAS (MF)

0.013

0.9997

0.0064

0.9978

Soluplus

0.023

0.9945

0.0035

0.8159

PVAc

0.0085

0.9807

0.0002

0.0029

PAA

0.0016

0.8232

0.0296

0.961



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 24

Table 3a: Long term stability results of dispersions after 6 months at 40˚C/75% RH in polypropylene bottles containing desiccant BI ABC:PVP K12 (60:40)

BI ABC:PVP VA64 (60:40)

BI ABC:HPMCAS (60:40)

XRPD

PLM

XRPD

PLM

XRPD

PLM

Amorphous

Some birefingence

Amorphous

Minor birefingence

Amorphous

Minor birefingence

Table 3b: Short term humidity stress stability of BI ABC Solid Dispersion at 50°C (75%, 50% & 30%RH) 3 day

1 day XRPD

PLM

XRPD

PLM

7 day

DSC (Tg˚C) XRPD

PLM

BI ABC:PVPK12 60DL Tg: 85.7˚C

75%RH 51%RH 30%RH

AAAA AAAA *

AACC CCCC CCCC AAAC AAAA AACC AAAC AAAC *

No Tg 33.8 51.4

NA NA AAAC ACCC AAAC *

BI ABC:VA64 60DL Tg: 86˚C

75%RH 51%RH 30%RH

AAAA AAAA *

AAAC ACCC CCCC AAAC AAAA AAAC AAAC AAAC *

40.7 45.5 62.9

* * AAAC AACC AAAC *

BI ABC:HPMCAS 60DL Tg 77˚C

75%RH

*

AAAC AAAA ACCC

41.3

AACC

51%RH

*

AAAA

*

AAAA

48.7

AAAA AACC

30%RH

AAAA

AAAC

*

AAAC

60.4

AAAA AAAC

NA

*: For completely amorphous/crystalline samples as measured via Polarized Light Microscopy, XRPD analysis was not conducted. Similarly if a higher humidity condition sample showed completely amorphous/crystalline XRPD profile, then a lower humidity sample was not measured.


Page 19 of 24

Figure 1: DSC thermogram showing heat flow versus temperature for nifedipine as a function of polymer loading from 2% PVP (green dashes) to 20% PVP (red solid line).

Nifedipine:PVP (98:02)

Tc Peak moving with DLNifedipine:PVP (95:05)

0.1 Nifedipine:PVP (90:10)

Nifedipine:PVP (80:20)

Heat Flow (W/g)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Tg

0.0

-0.1

-0.2 20 Exo Up

40

60

80

100

120

Temperature (°C)

140

160

180

Universal V4.2E TA Instruments



Figure 2: DSC thermogram showing heat flow versus temperature for nifedipine as a function of polymer loading from 2% PAA (green dashes) to 20% PAA (red solid line).

0.2

Tc Peak moving with DL Nifedipine:PAA (98:02)

Nifedipine:PAA (95:05)

Tg

Nifedipine:PAA (90:10) Nifedipine:PAA (80:20)

Heat Flow (W/g)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 24

-0.3

-0.8 20 Exo Up

40

60

80

100

120

Temperature (°C)

140

160

180

Universal V4.2E TA Instruments


Page 21 of 24

Figure 3: Reduced Crystallization Temperature as a function of polymer concentration for nifedipine for PVP (blue diamonds), HPMCAS (triangles), PVAc (circles), and PAA (squares). The open circle on the yaxis represents reduced crystallization temperature for amorphous nifedipine. The fit lines along with slopes and R2 is provided in the plot itself

PVP y = 0.0247x + 0.3097 R² = 0.9925 HPMCAS y = 0.013x + 0.3796 R² = 0.9997

0.9 Reduced Crystallization Temperature (Tc-Tg)/(Tm-Tg)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


0.8 0.7 0.6

PVAc y = 0.0085x + 0.3841 R² = 0.9807

0.5 0.4

PAA y = 0.0016x + 0.37 R² = 0.8232

0.3 0.2 0.1 0.0 0

5

10

15

20

25

Polymer Concentration w/w



Figure 4: Reduced Crystallization Temperature to polymer concentration for BI-ABC for PAA (triangles), PVP (diamonds), PVP VA64 (circles), and PVAc (squares). The open circle on the y-axis represents reduced crystallization temperature for amorphous BI ABC. The fit lines along with slopes and R2 is provided in the plot itself

1.0

Reduced Crystallization Temperature (Tc-Tg)/(Tm-Tg)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 24

PAA y = 0.0296x + 0.4359 R² = 0.961

0.9 0.8

PVP y = 0.0095x + 0.4867 R² = 0.9558

0.7 0.6

PVP-VA64 y = 0.0028x + 0.4947 R² = 0.5423

0.5

PVAc y = -0.0002x + 0.4557 R² = 0.0029

0.4 0.3 0.2 0.1 0.0 0

5

10

15

20

25

Polymer Concentration w/w


Page 23 of 24

Figure 5 (a, b, c): XRPD patterns of BI ABC dispersions with HPMCAS, PVP, PVP-VA64. The ratio of active to polymer, i.e. BI ABC:Polymer is 3:2. The XRPD patterns were taken at time zero and at time 35 months for samples stored at 25˚C/60% RH in poly-propylene bottles with desiccant. Note Figure 5a is dispersion with PVP, 5b is dispersion with PVP-VA64, and 5c is dispersion with HPMCAS 5a BI ABC:PVP K12 (6:4) SDD 1400

1300

Time 0 1200

35mo stored@25C/60%RH 1100

1000

Lin (Counts)

900

800

700

600

500

400

300

200

100

0 4

10

20

30

2-Theta - Scale

5b BI ABC : PVP VA64 (6:4) SDD 1000

900

Time 0 35mo stored@25C/60%RH

800

700

Lin (Counts)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


600

500

400

300

200

100

0 4

10

20

30

2-Theta - Scale

5c 23 ACS Paragon Plus Environment


BI ABC : HPMCAS (6:4) SDD 900

Time 0

800

35mo stored@25C/60%RH 700

600

Lin (Counts)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 24

500

400

300

200

100

0 4

10

20

30

2-Theta - Scale


Reduced Crystallization Temperature Methodology for Polymer

Recommend Documents