Impact of a Single Hydrogen Substitution by Fluorine on the Molecular

Subscriber access provided by YORK UNIV

Article

Impact of a single hydrogen substitution by fluorine on the molecular interaction and miscibility between sorafenib and polymers Chengyu Liu, Cong-Qiao Xu, Junguang Yu, Yipshu Pui, Huijun Chen, Shan Wang, Alan (Donghua) Zhu, Jun Li, and Feng Qian Mol. Pharmaceutics, Just Accepted Manuscript • DOI: 10.1021/ acs.molpharmaceut.8b00970 • Publication Date (Web): 04 Dec 2018 Downloaded from http://pubs.acs.org on December 9, 2018

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 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 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.

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 23 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

Impact of a single hydrogen substitution by fluorine on the molecular interaction and miscibility between sorafenib and polymers

Chengyu Liu1†, Congqiao Xu2†, Junguang Yu1, Yipshu Pui1, Huijun Chen1, Shan Wang1, Alan (Donghua) Zhu3, Jun Li2, Feng Qian1

1School

of Pharmaceutical Sciences, Collaborative Innovation Center for Diagnosis and Treatment of Infectious Diseases, and Key Laboratory of Bioorganic Phosphorus Chemistry & Chemical Biology (Ministry of Education), Tsinghua University, Beijing 100084, China. 2Department

of Chemistry and Key Laboratory of Organic Optoelectronics & Molecular Engineering of Ministry of Education, Tsinghua University, Beijing 100084, China. 3China

CMC, Small Molecule Pharmaceutical Development, Janssen Research & Development, Johnson & Johnson, Shanghai, China †These authors contributed equally to this work.

* To whom correspondence should be addressed: Dr. Alan Zhu: [email protected] Prof. Jun Li: [email protected] Prof. Feng Qian: [email protected]

Manuscript for Molecular Pharmaceutics

ACS Paragon Plus Environment

Molecular Pharmaceutics 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

Abstract We aim to understand the potential impact of a modest chemical modification of a drug molecule on the downstream design of its amorphous solid dispersion (ASDs) formulation. To this end, we used sorafenib (SOR) and its fluorinated form, regorafenib (REG), as model drugs, to assess the impact of a single hydrogen substitution by fluorine on the molecular interaction and miscibility between drug and PVP or PVP-VA, two commonly used polymers for ASDs. In this study, we observed that the Tg values of PVP or PVP-VA based ASDs of SOR deviated positively from Gordon-Taylor prediction that assumes ideal mixing, yet the Tg of REG ASDs deviated negatively from or matched well with ideal mixing model, suggesting much stronger drug-polymer interactions in SOR ASDs compared with the REG ones. Using solution NMR and computational methods, we proved that a six-member-ring formed between the carbonyl groups on the polymers and the uramido hydrogen of SOR or REG, through intermolecular hydrogen bonding. However, steric hindrance resulted from fluorination in REG caused weaker interaction between REG-polymer than SOR-polymer. To further confirm this mechanism, we investigated the molecular interactions of other two uramido-containing model compounds, triclocarban (TCC) and gliclazide (GCZ), with PVP. We found that TCC but not GCZ formed hexatomic ring with PVP. We concluded that PVP based polymers can easily interact with N,N’-disubstituted urea compounds with trans-trans structure in the form of hexatomic rings, and the interaction strength of the hexatomic ring largely depended on the chemistry of drug molecules. This study illustrated that, even a slight chemical modification on drug molecules could result in substantial difference in drug-polymer interactions, thus significantly impact polymer selection and pharmaceutical performance of their ASD formulations.

Keywords: sorafenib, regorafenib, drug-polymer interactions, PVP, PVP-VA, amorphous solid dispersion.


Page 2 of 23

Page 3 of 23 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


Abstract figure：



Page 4 of 23

1. Introduction Drug candidates obtained through combinatorial chemistry against the same pharmacological target often preserve a highly similar core structure, while their pharmacological distinctions are manifested by modest chemical modifications to achieve better structural-activity relationship (SAR) and physiochemical properties. In the past decades, increasing usage of high throughput screening and combinatory chemistry have led to more and more poorly water soluble compounds1-3, which often requires amorphization of drug molecules using amorphous solid dispersion (ASD) technology to enhance solubility and achieve satisfactory oral bioavailability4-7. Different from traditional crystalline formulations wherein drug exists as crystalline particles, extensive drug-polymer interactions exists in ASDs since amorphous drug disperses in molecular forms and is intimately mixed with surrounding polymer molecules. It has been known that drug-polymer interaction plays crucial roles in physical stability, dissolution performance, and in vivo bioavailability of ASDs8-16. Therefore, the impact of chemical modification of a drug molecule on its downstream design of ASD formulation is an important consideration for medicinal chemists and formulation scientists. In this research, we used sorafenib (SOR) and its fluorinated form, regorafenib (REG), as model compounds, to study the impact of a small chemical modification (i.e., a single hydrogen substitution by fluorine) on the molecular interaction between them and PVP or PVP-VA, two commonly used polymers in ASD formulations. Pharmacologically, SOR is a diaryl urea multi-target kinase inhibitor that inhibits several kinases involved in tumor proliferation and angiogenesis, and was approved by US FDA for the treatment of hepatocellular carcinoma (HCC), primary kidney cancer (advanced renal cell carcinoma), and thyroid cancer17. Various sorafenib analogues have been synthesized based on structure-activity relationships18-20, among which, REG is the most successful one that offers the only systemic treatment with proven survival benefit in HCC patients progressing on SOR treatment

21.

Interestingly, the only tiny chemical difference between SOR and REG is that, a


Page 5 of 23 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


single hydrogen on the benzene ring of SOR was substituted by fluorine (Figure 1), which makes REG more potent than SOR while maintaining similar safety profile. Through the comparison of the Tg values of SOR/polymer and REG/polymer systems measured experimentally and predicted by Gordon-Taylor equation assuming ideal mixing, we observe significant differences in the strength of molecular interaction and drug-polymer mixing in PVP or PVP-VA based ASDs of SOR and REG, where SOR interacts with both polymers more strongly compared with REG. We attribute this to the steric hindrance caused by F atom on REG. We also used triclocarban (TCC) and gliclazide (GCZ) to further prove that PVP based polymers could easily interact with N,N’-disubstituted urea compounds with trans-trans structure in the form of hexatomic rings, and the interaction strength of the hexatomic rings was largely dependent on the chemistry of the drug molecules. This study illustrated that a slight chemical modification on drug molecules could result in substantial difference in drug-polymer interactions. Since carbamido group frequently appears in many drug molecules, the knowledge regarding what type of carbamido group might interact with commonly used pharmaceutical polymers such as PVP, and the impact of side groups on the pattern and strength of these drug-polymer interaction, could guide rationale design of ASD formulations for this class of compounds.

2. MATERIALS AND METHODS 2.1. Materials. Sorafenib free base (SOR), regorafenib free base (REG), triclocarban (TCC) and gliclazide (GCZ) were purchased from Ouhe Company (Beijing, China). PVP-VA (Kollidon VA 64) and PVP (poly-(vinylpyrrolidone), Kollidon 30) were provided by BASF Chemical Company Ltd. (Ludwigshafen, Germany). The chemical structures of the drugs and polymers are summarized in Figure 1. Methanol (MeOH), dichloromethane (DCM) and all buffer salts were purchased from Beijing Chemical Works (Beijing, China). Deuterated chloroform (CDCl3) and deuterated water (D2O)



Page 6 of 23

were obtained from Innochem science＆Technology Co., Ltd (Beijing, China).

28 27

F

F

26

30

28 32

F 29

1

Cl 3 4

14

31

O

6

2 5

N H

15

N H

8

7

10

16

O 17

13

12

22

18 19

11

27

O

25

24

21

23

N20

N H

CH3 Cl 30

F

F

1

26

32

F 29 14

31

6

4

5

9

N H

N H

8

19

Cl

1

Cl

2

6

4

3

14 15

O

N H

N H

8

10

13

Cl 18

12 11

9

3

2

21

2 3

8

O 1

CH3 25

F

19

16

O O 4 S N 7 N 11 N 13 O H H 10

15

18

14

12

Gliclazide β

α

n

4

N20

N H

20 5

β

N

19

24 23

33

Triclocarban α

12

18

6 1

9

7

21

Regorafenib

16

5

O

17

11

9

7

22

O

13

10

3

Sorafenib 17

15

O

2

16

n

N

O

4

PVP

m

O

1 3

β'

α'

O

1' 2'

2

PVP-VA

Figure 1. Chemical structures of sorafenib (SOR), regorafenib (REG), triclocarban (TCC), gliclazide (GCZ), polyvinylpyrrolidone (PVP), poly(vinylpyrrolidone-co-vinylacetate) (PVP-VA)

2.2. Preparation of Amorphous Solid Dispersions (ASDs) and Amorphous Drugs PVP and PVP-VA based amorphous solid dispersions of sorafenib and regorafenib were prepared by spray drying a 5wt % solution in DCM/methanol (1/1, v/v) cosolvent using a Yamato spray dryer (ADL311S, Yamato Scientific Co., Ltd., Tokyo, Japan). The solution flow rate was 6-8mL/min, inlet temperature was 85 °C, outlet temperature was ∼50°C, and atomizing N2 pressure was 0.1MPa. Amorphous drugs of SOR and REG were prepared by spray drying 2wt % solution in DCM/MeOH (1/1, v/v) using a Büchi spray dryer (B-90, Büchi Labortechnik AG, Postfach, Switzerland). The inlet temperature was 80 °C and the outlet temperature was ∼35°C. The spray-dried ASDs and pure drugs were vacuum-dried at room temperature for at least 24 h to remove all residual solvents before further use. The


Page 7 of 23 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


solid dispersions were confirmed to be amorphous by PXRD and differential scanning calorimetry (DSC) (data not shown).

2.3. Differential Scanning Calorimetry (DSC). A TA Instruments DSC Q2000 (New Castle, DE) was used to measure the Tg of different ASDs using a heating rate of 10 °C/min. The samples were heated from 40°C to 180 °C after isothermal at 105 °C for 3 min. The samples (4-8 mg) were analyzed under dry nitrogen purge (40 mL/min) in sealed aluminum pans with pinholes. The Tg was reported as the mid-point of the glass transition event. All determinations were made in triplicate. The instrument was calibrated for temperature using indium.

2.4. Solution NMR. 1D 1H NMR, 2D 1H NMR spectra and 13C NMR spectra of various solutions were acquired at 298K by a Bruker AV-600 NMR spectrometer (Bruker BioSpin GmbH, Rheinstetten, Germany) operating at a proton resonance frequency of 600 MHz. For 1H

NMR spectra, the chemical shifts were referenced with respect to TMS (0 ppm).

For nuclear Overhauser effect spectroscopy (NOESY) study, the mixing time was set to 0.2 s, and a relaxation delay of 3 s was used between the scans. For

13C

NMR

analysis, the chemical shifts were referenced with respect to CDCl3 (77.5 ppm).

2.5. Computational Methods. Theoretical calculations regarding drug-polymer interactions were performed using the Density Functional Theory (DFT) as implemented in the Gaussian 09 program package22, with hybrid B3PW91 functional and 6-311++g(2d, 2p) basis sets23, 24. All the geometries were fully optimized without any constraint, and the vibrational frequencies were also calculated to confirm that they are the true local minima. The Gibbs free energies (at 298.15 K and 1 atm) were obtained considering the zero point energies (ZPE). NMR spectra were simulated by using the gauge-independent atomic orbital (GIAO) method25. The solvent effect of chloroform was taken into account by



Page 8 of 23

applying a SCRF (self-consistent reaction filed) through the polarizable continuum model (PCM)26. PVP was simplified to N-Ethyl-2-pyrrolidone (NEP) as the computational model. Energy decomposition analysis (EDA) was carried out by using the Amsterdam Density Functional (ADF 2016.101) program27-29. The B3PW91 functional program was used, together with the DZP Slater basis sets. The scalar relativistic effects were taken into account by the zero-order-regular approximation (ZORA)30. Natural population analysis (NPA) was conducted to show partial atomic charges and better understand the properties of the molecules.

3. Results 3.1. Assess the drug-polymer interactions between SOR/REG and PVP/PVP-VA using thermal analysis The Tg values of amorphous SOR and REG are 87°C and 89°C, respectively. When analyzing all ASDs prepared in this study, single Tg was detected in all systems, indicating the drug and polymers are uniformly mixed. The Tg values of the ASDs were compared with those predicted by Gordon-Taylor equation (Eq. 1)31, an equation based on the additivity of free volumes of each individual components:

Tg 

w1Tg1  K GT w2Tg 2 w1  kGT w2

where K GT 

d1Tg1 d 2Tg 2

(1)

where Tg, w, and d represent the glass transition temperature, weight fraction and density, respectively, and KGT is a fitting parameter. The determined and predicted Tg values of binary systems with various drug-polymer ratios were plotted against drug loading in Figure 2. For SOR, the experimental values showed positive deviation from the predicted ones in both PVP and PVP-VA systems (Figure2A). However, for REG-PVP and REG-PVP-VA systems, the measured Tg values deviated negatively from or matched the theoretical prediction (Figure 2B).


Page 9 of 23 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


Figure 2. Glass transition temperatures of (A) ASDs of SOR with PVP and PVP-VA; (B) ASDs of REG with PVP and PVP-VA. The dashed, and solid lines represent the predictions based on the Gordon-Taylor equation. (d was estimated to be ~0.95 times of the crystal density ， crystal density was obtained according to the single crystal structure reported in reference32, 33)

3.2. Molecular interactions between SOR/REG and PVP/PVP-VA analyzed by solution NMR 3.2.1. 1D 1H NMR spectra. In order to investigate the molecular interaction between the drug and PVP or PVP-VA, SOR solution was titrated by each polymer, with the polymer: drug (w/w) ratio increased from 0:1 to 4:1, and the corresponding 1D 1H NMR spectra were acquired and compared in Figure 3 (A and B). Figure S1 shows the peak assignments for each proton of SOR and REG in CDCl3, and Figure S2 shows the 1D 1H NMR spectra of REG titrated by PVP or PVP-VA.



Figure 3. 1D 1H spectra of 1mg/mL SOR in CDCl3, titrated by increasing amount of PVP (A) and PVP-VA(B) from 0.125 to 4 mg/mL. Chemical shift of active hydrogen H7 (C) and H9 (D) on carbamido of SOR was plotted against the molar ratio between SOR and the -VP groups on PVP and PVP-VA; (E) Substitution efficiency of H7, H9, and H24 by D2O. 1D 1H spectra of SOR-PVP solution in CDCl3 (0.1%D2O) at different time points, SOR and PVP concentration is 5 mg/mL and 30 mg/mL respectively.


Page 10 of 23

Page 11 of 23 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


The addition of PVP or PVP-VA into SOR (or REG) solution caused different chemical shift behaviors of three active hydrogen atoms (H7, H9 and H24) (Figure 3 and Figure S2 for REG system). H24 shifted upfield indicating dipole-dipole interaction exists between polymer and part of the drug molecule, while H7 and H9 shifted downfield in parallel, from 8.02 ppm and 8.31 ppm (without polymer), to 8.74 ppm and 8.94 ppm (PVP/drug=4/1), or 8.52 ppm and 8.71 ppm (PVP-VA/drug = 4/1), respectively, indicating simultaneously weakening of electron clouds of H7 and H9 with the introduction of PVP or PVP-VA, likely due to hydrogen bonding formation with carbonyl groups on the -VP or -VA groups. However, as shown in Figure 3C and 3D, the chemical shifts of both H7 and H9 were found to be mostly influenced by the molar ratio between SOR and the -VP groups of either polymer, we concluded that the –VA groups on PVP-VA did not participate the H-bonding formation with SOR. Furthermore, substitution efficiency of the three active hydrogen (i.e., H7，H9，and H24 of SOR and REG) by D2O was studied. As shown in Figure 3E, 0.1% D2O was added into the SOR-PVP (1/6) solution in CDCl3. After only 15 min, the signal of H7 and H9 diminished significantly, while that of H24 only showed minimal decrease after 4 hours. This result demonstrated that H7 and H9 were active and could be substituted by D2O, while H24 was non-active and difficult to be substituted. Same results were observed in REG-PVP system (Figure S3). It’s worth noting that although D2O is not completely miscible with CDCl3 at high concentration, 0.1% D2O was able to form a clear and uniform solution with CDCl3. Also, the NMR results remained similar when we further reduced the amount of D2O in CDCl3 to 0.05% to ensure miscibility of the two solvents. Thus we confirmed that there was no interference caused by possible phase separation occurred in NMR solvent. Extra evidences are provided in the computational simulation part, where we show that H24 forms intramolecular hydrogen bonds that are difficult to dissociate. 3.2.2. 1D 13C NMR spectra. The effect of SOR on the chemical shifts of carbonyl carbon of PVP or PVP-VA was investigated using 1D

13C

NMR. As shown in Table S1, the carbonyl group in

-VP shifted 0.17ppm downfield when 5 mg/mL SOR was added into the 30 mg/mL



PVP-VA solution, while the one in -VA groups didn’t change. The REG/PVP and REG/PVP-VA systems showed similar results (Table S2).These results again proved that SOR or REG mainly interacted with the -VP groups but not the -VA groups. 3.2.3. 2D 1H NMR spectra: NOESY. The NOESY spectra of SOR-PVP and REG-PVP were shown in Figure 4. The peaks for each proton of PVP and PVP-VA were assigned according to literature report34. For SOR/PVP systems, the NOESY spectrum clearly confirmed the existence of interaction between H7, H9 of SOR and the H2, H3, H4 of VP group according to the spatial distance relationship, in which stronger interaction corresponds to shorter distance from carbonyl group (Figure 4A, red dotted square). Also H1, H15 and H22 of SOR showed much weaker interaction with the -VP groups (Figure 4A, blue and black dotted square), which could be resulted from the dipole-dipole interactions between them. As a comparison, the NOESY spectrum of REG-PVP was collected and shown as Figure 4B. In REG-PVP systems, intermolecular interactions were only observed between H7, H9 of REG and H2 of -VP group because H2 is the closest hydrogen beside carbonyl group of VP, which is consistent with the previous conclusion that REG formed weak interaction with PVP. Because H24 is located next to 25CH3, so the interaction signal in NOESY spectra is strong (green dashed circle in Figure 4A and 4B). Also it can be confirmed that H24 does not interact with the -VP groups (see the empty red solid circles in Figure 4A and 4B), which is consistent with the computational simulation results to be discussed later, where H24 is found to form intramolecular hydrogen bond.


Page 12 of 23

Page 13 of 23 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


Figure 4. 1D 1H NMR spectrum and contour plot of NOESY in CDCl3 system for (A)SOR/PVP systems and (B)REG/PVP systems.(The drug and polymer concentrations are 5 mg/mL and 30 mg/mL respectively).

3.3 Investigate the drug-polymer interactions between SOR/REG and PVP/PVP-VA using computational methods. The computational results showed that drug-polymer complexes were formed through the formation of OH bonds, as PVP approached SOR or REG. The optimized structures are shown in Figure 5, and the details of Cartesian coordinates are listed in Table S3. The bond distance between the O atoms of PVP and the H atoms of the carbamido groups was determined to be 1.90~2.06 Å, within the scale of hydrogen bonds. There are strong hydrogen bond interactions between SOR, REG and PVP. The average bond distance of O-H7 and O-H9 for SOR (2.00 Å) is slightly



larger than REG (1.98 Å), which leads to stronger Pauli repulsion between REG and PVP. Our calculated results showed that the Gibbs free energy of SOR-PVP is -3.31 kcal/mol, lower than that of REG-PVP of -3.01 kcal/mol, indicating that the interactions between SOR and PVP are stronger compared with REG and PVP. Besides, as shown in Table 2, energy decomposition analysis (EDA) of drug + PVP  drug-PVP process illustrates that the total bonding energy of SOR-PVP is 0.68 kcal/mol stronger than that of REG-PVP. The total bonding interactions can be decomposed into electrostatic interaction, orbital interaction and Pauli repulsion. Sum of the electrostatic and orbital interactions can be used to describe the hydrogen bonding between REG, PVP and PVP.

Figure 5. Optimized geometric structures of SOR-PVP and REG-PVP, with the Gibbs free energy (G) changes for formation of SOR-PVP and REG-PVP complex (red dotted line in the structure represent hydrogen bond). Hexatomic ring interaction mode existed in both systems was shown in the enlarged picture.

Normally speaking, substitution of atom with strong electronegativity (such as F) on the adjacent benzene ring of SOR will make the uramido hydrogen more likely to form hydrogen bonding. The NPA charge of the uramido hydrogen (listed in Table S4) for REG and REG-PVP is larger than SOR and SOR-PVP due to the strong electronegativity of F atom nearby, making the hydrogen bonding of REG-PVP stronger than SOR-PVP. Table 1 provides further evidence of this result as Sum (Elstat+Orb) of REG-PVP (-28.85 kcal/mol) is stronger than SOR-PVP (-28.53


Page 14 of 23

Page 15 of 23 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


kcal/mol). However, compared with SOR-PVP, the Pauli repulsion between REG and PVP is also stronger because of the larger steric hindrance. As the Pauli repulsion dominates the total bonding energy, the total bonding interaction of SOR-PVP is still stronger than REG-PVP. Table 1. Energy decomposition analysis (EDA) for drug + PVP  drug-PVP process at the B3PW91/DZP level of theorya Electrostatic Drug

Orbital

Pauli repulsion

Total Sum(Elstat+Orb)b

interaction

interaction

energy

SOR

13.04

-18.59

-9.94

-28.53

-15.49

REG

14.04

-18.59

-10.26

-28.85

-14.81

a All

energies in kcal/mol. represents sum of the electrostatic interaction and orbital interaction.

b Sum(Elstat+Orb)

The calculated and experimental 1H NMR chemical shifts of SOR-PVP and REG-PVP are in good agreement with each other, shown in Table S5. The relative large difference between the calculated and experimental chemical shifts of pure drugs are likely resulted from the non-ignorable intermolecular hydrogen bonds. Consequently, we considered di-sorafenib (di-SOR) and di-regorafenib (di-REG) complexes to study the effects of intermolecular interactions on the NMR properties. As listed in Table S5, the interactions between the drugs can also result in significant shifts in the H NMR spectra, making the calculated results closer to the experimental ones.

3.4. Investigate the drug-polymer interactions between GCZ/TCC and PVP/PVP-VA using solution NMR Two other molecules triclocarban (TCC) and gliclazide (GCZ) which also have uramido groups were used to better understand the interaction pattern between this class of small molecules and PVP based polymers. The chemical structures of TCC and GCZ were shown in Figure 1, and the NMR peak assignments for each proton were shown in Figure S1.



With the addition of PVP, two active hydrogen atoms (H7 and H9) on TCC shifted downfield in parallel, from 7.35ppm and 7.17ppm (polymer/drug=0/1) to 8.42ppm and 8.34ppm (PVP/drug=4/1) (Figure 6A). While for GCZ, the chemical shift of active hydrogen H8 didn’t change and the peak was wider with the addition of polymer. Meanwhile, the chemical shift of active hydrogen H12 slightly shifted downfield (Figure 6B). The two uramido hydrogen atoms of GCZ locate in different chemical environments when interact with PVP. These results suggested that PVP formed hexatomic ring interaction with TCC but not GCZ.

Figure 6. 1D 1H spectra of (A) with 2 mg/mL TCC in CDCl3-CD3CN(9-1), and (B) 4 mg/mL GCZ in CDCl3 titrated by increasing amount of PVP from drug:polymer ratio of 1:0.5 to 1:4.

4. Discussion 4.1 The relationship between Tg plots and the strength of drug-polymer interaction The Gibb’s free energy of mixing (△Gmix) must be negative for complete miscibility, namely △Gmix= △ Hm-T △ Sm < 0. Three thermodynamic terms that contribute to △Gmix are intermolecular interactions, free volume effect, and combinatorial entropy 35. Since contribution of combinatorial entropy is always negligible, miscibility between drug and polymer is predominantly dictated by the enthalpy effects. Thus, hetero-contact interaction has to overcome the energy of homo-contact interaction. Stronger hetero-contact will decrease the “free volume” within the blend, and increase


Page 16 of 23

Page 17 of 23 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


the blend Tg above the values predicted by the additivity rules (Gordon-Taylor equation). In contrast, systems with weaker interaction will increase the “free volume” thus the blend Tg may coincide with or locate below the predicted values36-38. In this research, SOR forms strong interaction with PVP and PVP-VA so the Tg plot shows positive deviation compared with the predictions by Gordon-Taylor equation, while REG forms weak interaction with PVP and PVP-VA so that the experimental Tg values of the ASDS deviate negatively from or matched well with the predicted ones. The interaction mode and interaction strength can also be confirmed by the solution NMR and computational simulations. 4.2. Small molecules with trans-trans N,N’-disubstituted urea structure can form hexatomic rings with PVP based polymers Numerous methods have been used to characterized the interaction between drug and polymer, such as thermal analysis (melting depression)39, characterizations8, simulation43,

44.

9,

solid state NMR41,

42,

40,

spectroscopic

solution NMR43, and molecular

Apart from these methods, positive deviation of Tg plot from

Gordon-Taylor equation has also been used as an indication of strong interaction between drug and polymer44, 45. PVP and PVP-VA possess the same -VP groups, many studies have shown that the -VA groups were less likely to participate in the intermolecular interactions with small molecules8, 12 because of the steric hindrance by the neighboring methyl groups. In this study, we can conclude that SOR/REG mainly interact with VP group but not VA group (see Figure 5, and Table S1, S2). Based on the comparison of four compounds (i.e., SOR, REG, TCC and GCZ), it appeared that small molecules with symmetrical structures or symmetrical electron density neighboring the uramido group were prone to form hexatomic rings with the -VP groups. However, exceptions indicating otherwise also existed46 (Figure S4), suggesting that symmetrical substituent group or electron density around the uramido group is not the prerequisite. Associative patterns of N,N’-disubstituted ureas have been widely reported in the past, which can function as building units for the assembly of hydrogen-bonded crystal chains47. The hydrogen-bonded motif is composed of two -NH proton donors



and one -C=O proton acceptor in the bifurcated hydrogen bonds. The structure of thiourea is similar to that of urea and it has also been widely used as building blocks for the assembly of crystalline frameworks46, 48. Although the thiocarbonyl group is a weaker hydrogen-bond acceptor than the carbonyl one, the –NH donor in thioureas is more acidic. Ureas are present predominantly in the trans-trans conformation (Figure S5A), whereas thioureas usually exist as mixtures of trans-trans and trans-cis rotamers (Figure S5B). Compounds with urea group exist in the form of trans-trans crystal structure will also form hexatomic ring interaction mode with itself or the PVP based polymers. On the other hand, in the trans-cis structure, two urea hydrogens exist in different directions, and could not interact with proton acceptor simultaneously. In this present work, the GCZ molecule exhibits trans-cis structure (Figure S6D), which is determined by steric hindrance of the molecule. So GCZ could not form hexatomic ring interaction mode with itself or the VP group. For SOR, REG and TCC, they are present in the trans-trans configuration in the crystal structure (Figure S6A-C), and they can form hexatomic ring interaction mode with themselves as well as PVP based polymers. Trans-trans structures of the N,N’-disubstituted urea compounds could be an indicator of forming hexatomic ring association pattern with PVP based polymers. The mechanism was also applied to thiourea compounds. Drug molecules containing thiourea are not common, but the interaction mechanism and similar formulation approach could be applied to the thiourea compounds in the future.

4.3 Implication of structure modification of drug molecules to the downstream formulation design Structure-activity-relationship (SAR) is an important concept in medicine chemistry, which provides design guidance on structure optimization of lead compounds in drug discovery stage49, 50. Many drug molecules developed based on SAR have structural similarity, such as common core structure with modified side groups to improve the potency. However, the relationships between chemical


Page 18 of 23

Page 19 of 23 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


structure of drug molecules and the downstream formulation design and their pharmaceutical performance remain elusive, despite the strong desire in pharmaceutical industry to be able to make such connections or prediction early from drug discovery. The structure of small molecules is the determining factor of their physiochemical properties that are critical for pharmaceutical development, such as crystal structure, LogP, pKa, melting point, solubility and etc. Chemical modification on a drug molecule, even subtle ones like a single proton substitution as we demonstrated in this study, could still change the mode and strength of drug-excipient interactions substantially, thus impact the pharmaceutical performance of formulations where drug and excipients are interacting at molecular level.

4. Conclusion The molecular interaction and miscibility between a drug molecule (SOR, and its fluorinated form, REG), and a pharmaceutical polymer (PVP, and PVP-VA) can be significantly affected by a minor chemical modification on the drug molecule. In short, SOR and REG interact with PVP based polymers in the hexatomic ring association pattern. Uramido of compounds and carbonyl group of VP were found to be the main interaction sites. Although the hydrogen bonding of REG-PVP should be stronger than SOR-PVP only based on the electron cloud density, the total bonding interaction between SOR and PVP was found to be stronger because of the dominant Pauli repulsions between REG and PVP. The differences in these drug-polymer interaction and miscibility were evidenced by the Tg-composition profiles, NMR analysis and computational simulation. We concluded that compounds with uramido or thiourea exhibiting trans-trans structures might tend to interact with PVP based polymers in the hexatomic ring mode, and the strength of interaction could be influenced by the neighboring side groups. These interaction patterns could have practically implications on the stability, dissolution, or even in vivo performance of drug-polymer amorphous solid



dispersions, or other formulations wherein drug and polymer are interacting at molecular level. Therefore, the potential relationship between the chemical structure of a drug molecule and the future formulation design should be assessed based on the molecular interaction mechanism between the drug and commonly used pharmaceutical excipients.

Acknowledgements This research is supported by Beijing Advanced Innovation Center for Structural Biology, China National Nature Science Foundation (Project Number 81573355), and Janssen Pharmaceuticals, Inc., a pharmaceutical company of Johnson & Johnson.

Supporting Information Proton NMR assignment for sorafenib, regorafenib, triclocarban and gliclazid in CDCl3; 1H spectra of REG in CDCl3, titrated by increasing amount of PVP or PVP-VA; 1H spectra of REG-PVP solution in CDCl3

at different time points;

Crystal structure of model compound that forms trans-trans structure; Hydrogen-bonding motifs typically observed for disubstituted ureas and disubstituted thioureas in the solid state; Crystal structure of SOR, REG, TCC, and GCZ; Influence of SOR on the chemical shift of carbonyl group of PVP or PVP-VA; Influence of REG on the chemical shift of carbonyl group of PVP or PVP-VA; Cartesian coordinates of SOR, REG, SOR-PVP and REG-PVP; NPA charge of H7, H9 and F33 atoms Calculated and experimental NMR chemical shifts for H7, H9 and H24

References

1.

Lipinski, C., Poor aqueous solubility - An industry wide problem in drug discovery. 2002; Vol. 5,

p 82-85.


Page 20 of 23

Page 21 of 23 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


2.

Lipinski, C. A.

Drug-like properties and the causes of poor solubility and poor permeability.

Journal of Pharmacological and Toxicological Methods 2000, 44, (1), 235-249. 3.

Lipinski, C., Physicochemical properties and the discovery of orally active drugs: technical and

people issues. 2002; Vol. 13. 4.

Chiou, W. L.; Riegelman, S. Pharmaceutical Applications of Solid Dispersion Systems. J. Pharm.

Sci. 1971, 60, (9), 1281-1302. 5.

Newman, A.; Knipp, G.; Zografi, G.

Assessing the performance of amorphous solid dispersions.

J Pharm Sci 2012, 101, (4), 1355-77. 6.

Qian, F.; Wang, J.; Hartley, R.; Tao, J.; Haddadin, R.; Mathias, N.; Hussain, M.

Solution

behavior of PVP-VA and HPMC-AS-based amorphous solid dispersions and their bioavailability implications. Pharmaceutical research 2012, 29, (10), 2765-76. 7.

Qian, F.; Huang, J.; Hussain, M. A.

Drug-polymer solubility and miscibility: Stability

consideration and practical challenges in amorphous solid dispersion development. J Pharm Sci 2010, 99, (7), 2941-7. 8.

Chen, Y.; Liu, C.; Chen, Z.; Su, C.; Hageman, M.; Hussain, M.; Haskell, R.; Stefanski, K.; Qian,

F.

Drug-polymer-water interaction and its implication for the dissolution performance of amorphous

solid dispersions. Mol. Pharmaceutics 2015, 12, (2), 576-89. 9.

Taylor, L. S.; Zografi, G.

Spectroscopic characterization of interactions between PVP and

indomethacin in amorphous molecular dispersions. Pharmaceutical research 1997, 14, (12), 1691-1698. 10. Marsac, P. J.; Li, T.; Taylor, L. S.

Estimation of drug-polymer miscibility and solubility in

amorphous solid dispersions using experimentally determined interaction parameters. Pharmaceutical research 2009, 26, (1), 139-51. 11. Khougaz K, C. S.

Crystallization inhibition in solid dispersions of MK-0591 and

poly(vinylpyrrolidone) polymers. J Pharm Sci 2000, 89, 1325-1334. 12. Chen, Y.; Wang, S.; Wang, S.; Liu, C.; Su, C.; Hageman, M.; Hussain, M.; Haskell, R.; Stefanski, K.; Qian, F.

Initial Drug Dissolution from Amorphous Solid Dispersions Controlled by Polymer

Dissolution and Drug-Polymer Interaction. Pharmaceutical research 2016. 13. Vasconcelos, T.; Sarmento, B.; Costa, P.

Solid dispersions as strategy to improve oral

bioavailability of poor water soluble drugs. Drug discovery today 2007, 12, (23-24), 1068-75. 14. Yu, L.

Amorphous pharmaceutical solids: Preparation, characterization and stabilization.

Advanced drug delivery reviews 2001, 48, (1), 27-42. 15. Vo, C. L.-N.; Park, C.; Lee, B.-J.

Current trends and future perspectives of solid dispersions

containing poorly water-soluble drugs. European Journal of Pharmaceutics and Biopharmaceutics 2013, 85, (3), 799-813. 16. Li, Y.; Pang, H.; Guo, Z.; Lin, L.; Dong, Y.; Li, G.; Lu, M.; Wu, C.

Interactions between drugs

and polymers influencing hot melt extrusion. Journal of Pharmacy and Pharmacology 2014, 66, (2), 148-166. 17. NEXAVAR (sorafenib) tablets instructions,approval:2005, revised:2011. 18. Yao, J.; Chen, J.; He, Z.; Sun, W.; Xu, W.

Design, synthesis and biological activities of thiourea

containing sorafenib analogs as antitumor agents. Bioorganic & medicinal chemistry 2012, 20, (9), 2923-9. 19. Hwang, S. H.; Wecksler, A. T.; Zhang, G.; Morisseau, C.; Nguyen, L. V.; Fu, S. H.; Hammock, B. D.

Synthesis and biological evaluation of sorafenib- and regorafenib-like sEH inhibitors.



Page 22 of 23

Bioorganic & medicinal chemistry letters 2013, 23, (13), 3732-7. 20. Jiao, Y.; Xin, B. T.; Zhang, Y.; Wu, J.; Lu, X.; Zheng, Y.; Tang, W.; Zhou, X.

Design, synthesis

and evaluation of novel 2-(1H-imidazol-2-yl) pyridine Sorafenib derivatives as potential BRAF inhibitors and anti-tumor agents. European journal of medicinal chemistry 2015, 90, 170-83. 21. Bruix, J.; Qin, S.; Merle, P.; Granito, A.; Huang, Y.-H.; Bodoky, G.; Pracht, M.; Yokosuka, O.; Rosmorduc, O.; Breder, V.; Gerolami, R.; Masi, G.; Ross, P. J.; Song, T.; Bronowicki, J.-P.; Ollivier-Hourmand, I.; Kudo, M.; Cheng, A.-L.; Llovet, J. M.; Finn, R. S.; LeBerre, M.-A.; Baumhauer, A.; Meinhardt, G.; Han, G.

Regorafenib for patients with hepatocellular carcinoma who

progressed on sorafenib treatment (RESORCE): a randomised, double-blind, placebo-controlled, phase 3 trial. The Lancet 2017, 389, (10064), 56-66. 22. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A., Gaussian 09, Revision C.01. Gaussian, Inc.: Wallingford, CT, 2010. 23. Becke, A. D.

Density‐Functional Thermochemistry. III. The Role of Exact Exchange. J. Chem.

Phys. 1993, 98, 5648-5652. 24. Perdew, J. P.; Wang, Y.

Accurate and Simple Analytic Representation of the Electron-Gas

Correlation Energy. Phys. Rev. B 1992, 45, 13244-13249. 25. Wolinski, K.; Hinton, J. F.; Pulay, P.

Efficient Implementation of the Gauge-Independent

Atomic Orbital Method for NMR Chemical Shift Calculations. 1990, 112, 8251-8260. 26. Scalmani, G.; Frisch, M. J.

Continuous Surface Charge Polarizable Continuum Models of

Solvation. I. General Formalism. J. Chem. Phys. 2010, 132, 114110. 27. ADF 2016.101. http://www.scm.com. 28. Te Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Fonseca Guerra, C.; van Gisbergen, S. J. A.; Snijders, J. G.; Ziegler, T.

Chemistry with ADF. J. Comput. Chem. 2001, 22, 931-967.

29. Fonseca Guerra, C.; Snijders, J. G.; te Velde, G.; Baerends, E. J.

Towards an order-N DFT

Method. 1998, 99, 391-403. 30. Lenthe, E. v.; Baerends, E. J.; Snijders, J. G.

Relativistic Regular Two-Component

Hamiltonians. J. Chem. Phys. 1993, 99, 4597-4610. 31. Gordon, M.; Taylor, J. S.

Ideal copolymers and the second ‐ order transitions of synthetic

rubbers. I. Non‐crystalline copolymers. Journal of Chemical Technology and Biotechnology 1952, 2, (9), 493-500. 32. Ravikumar, K.; Sridhar, B.; Rao, A. K. S. B.; Reddy, M. P.

Sorafenib and its tosylate salt: a

multikinase inhibitor for treating cancer. Acta Cryst. 2011, C67, o29-o32. 33. Sun, M.-Y.; Wu, S.-X.; Zhou, X.-B.; Gu, J.-M.; Hu, X.-R.

Comparison of the crystal structures

of the potent anticancer and anti-angiogenic agent regorafenib and its monohydrate. Acta Cryst. 2016, C72, 291-296. 34. Roscigno, P.; Asaro, F.; Pellizer, G.; Ortona, O.; Paduano, L.

Complex formation between

poly(vinylpyrrolidone) and sodium decyl sulfate studied through NMR. Langmuir : the ACS journal of surfaces and colloids 2003, 19, 9638-9644. 35. Patterson, D.; Robard, A.

thernodynamic of polymer compatibility. Macromolecules 1978, 11,

(4), 690-695. 36. Schneider, H. A.

The Gordon-Taylor equation. Additivity and interaction in compatible polymer

blends. Makromol. Chem. 1988, 189, 1941-1955. 37. Schneider, H. A.

Glass transition behaviour of compatible polymer blends. polymer 1989, 30,


Page 23 of 23 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


771-779. 38. Schneider, H. A.

Conformational Entropy Contributions to the Glass Temperature of Blends of

Miscible Polymers. J. Res. Natl. Inst. Stand. Technol. 1997, 102, (2), 229-248. 39. Tao, J.; Sun, Y.; Zhang, G. G. Z.; Yu, L.

Solubility of small-molecule crystals in polymers:

D-Mannitol in PVP, indomethacin in PVP/VA, and nifedipine in PVP/VA. Pharmaceutical research 2009, 26, (4), 855-864. 40. Sun, Y.; Tao, J.; Zhang, G. G.; Yu, L.

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 2010, 99, (9), 4023-31. 41. Yuan, X. D.; Xiang, T. X.; Anderson, B. D.; Munson, E. J.

Hydrogen Bonding Interactions in

Amorphous Indomethacin and Its Amorphous Solid Dispersions with Poly(vinylpyrrolidone) and Poly(vinylpyrrolidone-co-vinyl

acetate)

Studied

Using

C-13

Solid-State

NMR.

Molecular

pharmaceutics 2015, 12, (12), 4518-4528. 42. Song, Y.; Yang, X.; Chen, X.; Nie, H.; Byrn, S.; Lubach, J. W.

Investigation of drug-excipient

interactions in lapatinib amorphous solid dispersions using solid-state NMR spectroscopy. Molecular pharmaceutics 2015, 12, (3), 857-66. 43. Nie, H.; Mo, H.; Zhang, M.; Song, Y.; Fang, K.; Taylor, L. S.; Li, T.; Byrn, S. R.

Investigating

the Interaction Pattern and Structural Elements of a Drug-Polymer Complex at the Molecular Level. Molecular pharmaceutics 2015. 44. Gupta, P.; Thilagavathi, R.; Chakraborti, A. K.; Bansal, A. K.

Role of Molecular Interaction in

Stability of Celecoxib-PVP Amorphous Systems. Molecular pharmaceutics 2005, 2, (5), 384-391. 45. Tajber, L.; Corrigan, O. I.; Healy, A. M.

Physicochemical evaluation of PVP-thiazide diuretic

interactions in co-spray-dried composites--analysis of glass transition composition relationships. European journal of pharmaceutical sciences : official journal of the European Federation for Pharmaceutical Sciences 2005, 24, (5), 553-63. 46. Cazacu, A.; Tong, C.; Lee, A. v. d.; Fyles, T. M.; Barboiu, M.

Columnar Self-Assembled

Ureido Crown Ethers: An Example of Ion-Channel Organization in Lipid Bilayers. J . Am. Chem. SOC 2006, 128, 9541-9548. 47. C.Etter, M.; Urbakzyk-Lipkowska, Z.; Zia-Ebrahimi, M.; Panunto, T. W.

Hydrogen Bond

Directed Cocrystallization and Molecular Recognition Properties of Diarylureas. J . Am. Chem. SOC 1990, 112, 8415-8426. 48. Custelcean, R.

Crystal engineering with urea and thiourea hydrogen-bonding groups. Chem.

Commun. 2008, (3), 295-307. 49. Willson, T. M.; Cobb, J. E.; Cowan, D. J.; Wiethe, R. W.; Correa, I. D.; Prakash, S. R.; Beck, K. D.; Moore, L. B.; Kliewer, S. A.; Lehmann, J. M.

The Structure-Activity Relationship between

Peroxisome Proliferator-Activated Receptor γ Agonism and the Antihyperglycemic Activity of Thiazolidinediones. J. Med. Chem. 1996, 39, 665-668. 50. K.Y, R.; Zee-Cheng; Cheng, C. C.

Antineoplastic Agents. Structure-Activity Relationship Study

of Bis(substituted aminoalkylamino)anthraquinones. J. Med. Chem. 1978, 21, (3), 291-294.


Impact of a Single Hydrogen Substitution by Fluorine on the Molecular

Recommend Documents