iso-structurality induced solid phase ... - ACS Publications

Article pubs.acs.org/crystal

Iso-Structurality Induced Solid Phase Transformations: A Case Study with Lenalidomide Ramanaiah Chennuru,#,$ Prakash Muthudoss,# Raja Sekhar Voguri,# Srividya Ramakrishnan,# Peddy Vishweshwar,# R. Ravi Chandra Babu,$ and Sudarshan Mahapatra*,# #

Centre for Excellence in Polymorphism and Particle Engineering, Integrated Product Development (IPDO), Dr. Reddy’s Laboratories Ltd., Bachupally, Hyderabad, Telangana 500090, India $ Department of Chemistry, College of Science, GITAM University, Visakhapatnam, Andhra Pradesh 530045, India S Supporting Information *

ABSTRACT: Lenalidomide (LDM) is a thalidomide analogue known for its immunomodulation, antiangiogenic, and antineoplastic properties. However, to date, only two forms of lenalidomide [Form-1 (anhydrous) and Form-2 (hemihydrate)] are reported in the literature. Through a comprehensive polymorph screening herein, we report five forms of lenalidomide [Form-3 (DMF-solvate), Form-4 (anhydrous), Form-5 (DMSO solvate), Form-6 (acetone solvate), and Form-7 (dihydrate)]. Single crystal structures (for all forms) are established to provide potential knowledge about the intermolecular interactions, three-dimensional structures, and the nature of solvent/water within the lattice. Thermodynamic stability investigations revealed unusual solid state phase transformations which are relatively unexplored to date. It is noteworthy that all solvates upon desolvation convert to Form-1 (thermodynamically stable anhydrous form), whereas all hydrates upon dehydration convert to a metastable Form-4 (novel anhydrous form) which, upon further heating, converts to more stable Form-1. Correlation of results from modeling, single crystal analysis, and nonambient studies established “isostructurality” as one of the major factors leading to such bifurcated phase transformations. Mechanisms of desolvation and dehydration in different forms of LDM are explained by utilizing various analytical techniques such as variable temperature Fourier transform infrared spectroscopy, variable temperature powder X-ray diffraction, differential scanning calorimetry, and hot stage microscopy. A thorough understanding of the relationships between structure and thermodynamic properties is deemed a prerequisite which is considered vital in selecting the most suitable form for drug product development.

1. INTRODUCTION Comprehensive studies on the isolation/characterization of pharmaceutical solids such as polymorphs, salts, hydrates, solvates, cocrystals, etc. are integral parts of the pharmaceutical industry. Further, the selected solid form should meet the basic requirements of optimal stability, reproducibility, and scalability, which will eventually lead to devising a robust and reliable process for its manufacture.1−4 The existence of more than one crystal structure for a chemical substance is referred as polymorphism. Polymorphs of active pharmaceutical ingredients (APIs) have always drawn attention in view of their physical and intellectual property.5−8 Polymorphs can exhibit different physical and chemical properties, including melting point, density, stability, manufacturability, solubility, bioavailability, etc.9,10 Moreover, from a commercial point of view, different solid forms of a drug can be patented, and discovery of novel solid forms by competitors is always a setback for the innovator.8,11,12 The importance of polymorph screening can be visualized by a fact of sudden appearance or disappearance of a polymorphic form during manufacturing and storage. These transformations will lead to a series of serious consequences in terms of patient safety, finance and brand value of © 2016 American Chemical Society

the company, and so on. The probability of obtaining polymorphs of an API depends on the conformational flexibility existing in the molecular structure and diversity in the crystal packing. However, pursuant to McCrone, the probability of obtaining polymorphs of a compound depends on “the time and money spent on research in the compound”.13 This statement is evidenced by the recent discovery of novel polymorphs of aspirin,14,15 maleic acid,16 and thymine.17 The effect of parameters such as temperature, pH, relative humidity, solubility etc. on the stability of the APIs is wellestablished in the literature.18−25 The goal of any specific solid form screening/development is to identify such triggering factors deciding the polymorphic phase stability.18−22 Axitinib is a classic example, which details the importance of the above statement. It crystallizes in number of solid forms including 4 anhydrous and 64 solvates/hydrates.23−25 However, the stable anhydrous form (Form-41) was obtained only when the desolvation pathway was fully understood.23−25 Hence, it is Received: October 5, 2016 Revised: November 28, 2016 Published: December 8, 2016 612

DOI: 10.1021/acs.cgd.6b01462 Cryst. Growth Des. 2017, 17, 612−628

Crystal Growth & Design

Article

2.2. Preparation of LDM Forms. 2.2.1. LDM Form-1 (Anhydrous Form). The input material obtained from Dr. Reddy’s Laboratories Ltd. matches with Form-1 (simulated PXRD pattern of CSD code: AJISES) and confirmed by matching PXRD, DSC, and FT-IR data (see Supporting Information, Figure S1, S8 and S15 respectively). 2.2.2. LDM Form-2 (Hemihydrate). Form-2 (bulk material) was obtained by suspending 100 mg of LDM Form-1 in 30 mL of 10% aqueous methanol and maintained this reaction mass for about 5 to 10 h at 10−15 °C, followed by filtration and drying. Polymorphic conformation of Form-2 was done using data from PXRD, DSC, and Fourier transform infrared (FT-IR) spectroscopy (see the Supporting Information, Figures S2, S9, and S16, respectively). 2.2.3. LDM Form-3 (DMF Solvate). Polycrystalline Form-3 was obtained by suspending LDM Form-1 (1.0 g) in 3 mL of DMF at room temperature for about 24−48 h, followed by filtration and drying the material at 25 °C in a vacuum tray dryer (VTD) for 5 h. Polymorphic purity of the bulk material was confirmed by PXRD, DSC, and FT-IR (see the Supporting Information, Figures S3, S10, and S17, respectively). One-hundred milligrams of LDM Form-3 was dissolved in 1.0 mL of DMF, and the solution was placed in a closed environment of methyl tert-butyl ether (MTBE)/ethyl acetate (EA) (1/1, v/v) under vapor diffusion. Colorless block-type crystals suitable for X-ray diffraction were obtained after 3−4 days. 2.2.4. LDM Form-4 (Anhydrous Form). Bulk material of Form-4 was obtained by desolvation of Form-2 in the air tray dryer (ATD) at 140 °C for 2−6 h. The material was removed from the ATD, and the novelty of Form-4 was confirmed by PXRD, DSC, and FT-IR (see the Supporting Information, Figures S4, S11, and S18, respectively). For the single crystal growth of Form-4, a vapor diffusion technique was used. One-hundred milligrams of LDM Form-4 was dissolved in 1 mL of trifloro ethanol, and toluene was taken as antisolvent. Concomitant crystallization of Form-4 and Form-1 (see the Supporting Information, Figure S22) was observed during the single crystal growth of Form-4. The existence of Form-1 and Form-4 were confirmed by attenuated total reflection (ATR)-FTIR analysis on carefully separated single crystals (Figure S22). 2.2.5. LDM Form-5 (DMSO Solvate). LDM Form-5 (bulk material) was obtained by suspending LDM Form-1 (1 g) in 2 mL of DMSO at room temperature for about 24−48 h followed by filtration and drying at 25 °C in VTD for 12 h. The bulk material was analyzed using PXRD, DSC, and FT-IR (see the Supporting Information, Figures S5, S12, and S19, respectively). One-hundred milligrams of LDM Form-5 was dissolved in 1.0 mL of DMSO, and the solution was placed in a closed environment of MTBE/EA (1/1, v/v) under vapor diffusion. Colorless block-type crystals suitable for X-ray diffraction were obtained after 15−20 days. 2.2.6. LDM Form-6 (Acetone Solvate). Polycrystalline LDM Form-6 was isolated by suspending LDM Form-1 (1 g) in 10 mL of acetone at room temperature for about 24−48 h, followed by filtration and drying at 25 °C in VTD for 1 h. The isolated bulk material was further analyzed by PXRD, DSC, and FT-IR (see Supporting Information, Figures S6, S13, and S20, respectively). One-hundred milligrams of LDM Form-6 was dissolved in 5.0 mL of acetone, and the solution was placed in a closed environment of MTBE/EA (1/1, v/v) for vapor diffusion. Colorless block-type crystals suitable for X-ray diffraction were obtained after 15−20 days. 2.2.7. LDM Form-7 (Dihydrate). Bulk material of Form-7 was obtained by suspending 100 mg of LDM Form-1 in 30 mL of 10% aqueous methanol for 24−48 h at 0−5 °C, followed by filtration and drying. The isolated material was further characterized by PXRD, DSC, and FT-IR (see the Supporting Information, Figures S7, S14, and S21, respectively). Colorless block-shaped single crystals of LDM Form-7 were obtained by dissolving 150 mg of LDM Form-1 in 2 mL of 0.1 N HCl (pH 1.2) and leaving the solution for crystallization for 1 h. The crystallization path for above solid forms of LDM is pictorially shown in Figure 1. It can be visualized from the figure that both hydrates of LDM (Form-2 and Form-7) can be obtained by slurring Form-1 at low temperature in 10% aqueous methanol. However, the hemihydrate isolates initially, and a prolonged slurry maintenance at low temperature leads to dihydrate formation. The remaining three

essential and important that the API solid form is monitored closely to understand the dehydration/desolvation pathways.18−21 Several guidelines and strategies have been developed by regulatory authorities to incorporate series of analytical methodologies to understand polymorphism in hydrate, solvate, and anhydrate forms of a drug.18−20 Single crystal X-ray diffraction (SXRD) and powder X-ray diffraction (PXRD) are considered as the standard techniques which provide fingerprint information on polymorphism.19,20 Various orthogonal techniques such as thermal [differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA)], spectroscopy (Raman, infrared and near-infrared), and hot stage microscopy (HSM) can provide additional insights on polymorphism. There are number of in silico tools which reduce the experimental load by predicting the stable polymorphs of a given compound. This ensures forecasting the issues and recommending the right parameters to stabilize the discovered solid form. It is noteworthy that the form selection activity is not just for easy unit operation but rather is an integral part of entire drug development process. Polymorphism and iso-structurality are inversely related to each other. These two have been reported independently in various literatures for a long time, and it is less explored when it comes to a single compound exhibiting both.26−30 Iso-structural crystals can be referred to as isomorphous (isometric) molecules forming identical packing motifs.31−46 Iso-structural crystals may be expected when molecules with certain functional groups exhibit the same or similar structural roles in crystal structures.28−30 Iso-structurality is more common in multicomponent systems such as cocrystals, solvent inclusion compounds, molecular complexes, etc.31−46 There are a number of reports on iso-structural solvent inclusion compounds with guests of different sizes and shapes, where the overall crystal lattice remain intact. However, iso-structurality in solids of a single component is rare. Single-component isostructural solids may be broadly divided into two types: one with rigid conformational molecules and the other with flexible conformational molecules. Iso-structural crystal structures involving molecules with conformational rigidity and a single component are less in number.47−54 On the other hand, isostructurality in organic solids with flexible conformational molecules is not straightforward. Flexible molecules exhibit changes in their conformation, shape, and hydrogen bonding, leading to a different packing arrangement or polymorphism rather than showing iso-structurality. The main objective of the current work was to discover novel polymorphs of LDM and investigate the existence of isostructurality in them. Thorough investigations were carried out to find the relationship of iso-structurality with dehydration and desolvation behavior. In the process of polymorph screening, various novel forms along with the reported forms of LDM were isolated and analyzed by various analytical techniques. In the current paper, we selectively focused on polymorphism, iso-structurality, and the mechanism of desolvation/dehydration in pseudopolymorphs of lenalidomide.

2. EXPERIMENTAL SECTION 2.1. Materials and Methods. Lenalidomide (API) was provided by Dr. Reddy’s Laboratories Ltd. and was used for research without further purification. All crystallization solvents used during experimentation were of analytical or chromatographic grade. 613



Article

Figure 1. Crystallization pathways for different forms of LDM.

Scheme 1

solvates (Form-3, Form-5, and Form-6) can be isolated by slurring Form-1 (input anhydrous material) in the respective solvent at room temperature over a period of 24 h. The anhydrous Form-4 can be obtained by dehydration of the hydrates. The phase transformations in different solid forms of LDM are described separately in the following section. 2.3. Instrumentation. 2.3.1. Differential Scanning Calorimetry. DSC thermograms were recorded on a Thermal Advantage (TA) Discovery instrument by heating the samples at a rate of 10 °C/min up to 350 °C with continuous purging of dry nitrogen gas (50 mL/min). 2.3.2. Thermogravimetry Analysis. TGA thermograms were recorded on a TA Q 500 series machine by heating the samples at a rate of 10 °C/min up to 350 °C with purging of dry nitrogen gas [balance (40 mL/min) and furnace (60 mL/min)]. 2.3.3. Powder X-ray Diffraction. All X-ray powder diffraction data were collected on a PANalytical X’Pert PRO diffractometer

(X’cellerator detector, Cu-anode, 45 KV, 40 ma, Brag-Brentano geometry) using the 2θ scan range, step size, and exposure time of 3−40°, 0.03°, and 1200 s/step, respectively. 2.3.4. Single Crystal X-ray Diffraction. Single crystal data sets for LDM Forms-3 to 7 were collected in open mounting condition on a Bruker AXS SMART APEX CCD diffractometer; the crystal structure was solved using a direct method, and reasonably good quality data sets were collected for Forms-3 to 7 on carefully chosen crystals. 2.3.5. FT-IR Spectroscopy. FTIR spectra of LDM forms (Forms-1 to 7) were measured in the mid-IR range (4000−400 cm−1) using a Cary 680 FTIR spectrometer (Agilent, Santa Clara, CA) coupled with a room temperature detector, deuterated L-alanine-doped triglycine sulfate (DLaTGS). The spectrometer was coupled with a variabletemperature Specac Golden Gate diamond ATR accessory (Specac Limited (UK) Orpington, Kent). The temperature dependences of the IR spectra in the range of 30−200 °C were obtained with a step of 614



Article

Table 1. Crystallographic Parameters of Different Crystalline Forms of LDM Form-1

Form-2

Form-3

reference

AJISES

AJISES

CCDC 1487163

chemical formula formula weight crystal system space group T (K) a (Å) b (Å) c (Å) α (°) β (°) γ (°) Z V (Å3) Dcalc (g cm−3) reflections collected unique reflections observed reflections R1 [I > 2(I)] wR2 (all) goodness-of-fit

C13H13N3O3 259.26 triclinic P1̅ 294(2) 5.9983(5) 8.9198(8) 11.5785(10) 75.7110(10) 84.6600(10) 86.0380(10) 2 597.04(9) 1.442 5070

2(C13H13N3O3)· (H2O) 527.52 monoclinic P21/c 294(2) 8.4425(7) 22.2874(19) 13.6627(12) 90 101.0690(10) 90 8 2523.0(4) 1.413 7073

2(C13H13N3O3)· (C3H7NO) 591.62 triclinic P1̅ 100.01(10) 11.1854(7) 12.1000(7) 12.2516(7) 62.614(6) 80.069(5) 67.440(6) 2 1359.67(16) 1.445 8342

2092

4430

1975 0.0393 0.0407 1.052

Form-4 CCDC 1487162 C13H13N3O3

Form-5

Form-6

Form-7

CCDC 1487166

CCDC 1487165

CCDC 1487164

259.26 orthorhombic Pbca 294(2) 12.9538(8) 11.1196(7) 16.7332(12) 90 90 90 8 2410.3(3) 1.429 13663

2(C13H13N3O3)· (C2H6OS) 597.66 triclinic P1̅ 100.01(10) 10.9073(6) 12.1823(7) 12.7970(7) 62.010(6) 78.765(5) 65.127(5) 2 1362.24(15) 1.455 5423

2(C13H13N3O3)· (C3H6O) 576.60 triclinic P1̅ 100.01(10) 10.8783(6) 12.1389(7) 12.5246(8) 61.603(6) 82.208(5) 67.607(5) 2 1343.05(16) 1.426 8534

(C13H13N3O3)· 2(H2O) 295.29 monoclinic P21/c 150.00(10) 13.3683(5) 8.7873(3) 11.6761(4) 90 98.926(4) 90 4 1355.00(8) 1.448 4382

5535

2118

5127

5477

3368

3972

4804

1665

4740

4748

2486

0.0362 0.0400 1.021

0.0481 0.0554 1.053

0.0336 0.0474 1.068

0.0565 0.0630 1.019

0.0517 0.0606 1.037

0.0496 0.0727 1.037

Figure 2. Molecular overlay of all molecules from the asymmetric unit of different solid forms of lenalidomide (Forms-1 to 7) showing conformational flexibility in the pyridinedione ring. 20 °C. The data collection parameters were 4 cm−1 spectral resolution, 16 scans, triangle apodization, and no zero-filling. The ATR accessory was preheated to the desired temperature; the sample was placed, and spectra were measured. 2.3.6. Polarized Light Microscopy (PLM). All microscopic images were collected on CARL ZIESS (AXIO SCOPE A1) optical microscope. Fine solid particles of Forms-1 to 7 were spread on a glass slide to obtain a uniform layer, and a coverslip was used to cover the total sample during the analysis. 2.3.7. Hot Stage Microscopy. Hot stage microscopic images were collected on a NIKON-EELIPSE-EOPOL microscope equipped with Mettler Toledo-FP28HT hot stage. Single crystals of hydrates (Forms2 and 7) and solvates (Forms-3, 5, and 6) were placed on the hot stage and heated at a rate of 10 °C/min up to 300 °C with continuous image collection at 5 s intervals.

3. RESULTS AND DISCUSSION A literature (CSD, version 5.36, November 2014) search for LDM resulted in eight hits, out of which two (Refcode: AJISES and AJISIW) are for Form-1 and Form-2,55 and the remaining are cocrystals.56,57 In spite of its pharmaceutical significance from 2004 onward, this compound is hardly considered for polymorphism. We initiated an exhaustive and systematic study in this direction and explored the polymorph screening/ structural landscape for different forms of LDM. This molecule exhibits conformational flexibility and indicates different possibilities of donor−acceptor pairing by pyridinedione ring moiety (Scheme 1), making it a potential candidate for polymorph screening. In the current paper, we discuss two anhydrous 615



Article

Table 2. Hydrogen Bond Metrics for Different Crystal Structures of LDM interaction

D−H (Å)

D−A (Å)

H−A (Å)

D−H···A (deg)

symmetry

Form-1 N1−H1···O2 N3−H2···N3 N3−H3···O1 C3−H3A···O2 C3−H3B···O3 C4−H4···O3

0.88 0.90 0.81 0.97 0.97 0.98

2.8987 3.2596 3.1800 3.2130 3.4606 3.2051

N1A−H(1N)···O3B N1B−H4N···O1B O1W−H(1W)···O2B N3A−H(2N)···O1A N3A−H3···O3A N1A−H1···O1B N3B−H5N···O3A N3B−H6N···O1W C3A−H3A···N3B C4B−H4B···O1A C10B−H10B···O2A

0.89 0.88 0.98 0.90 0.91 0.88 0.89 0.90 0.97 0.98 0.93

2.8921 3.0161 2.9178 3.1538 3.0761 3.0161 3.0808 3.1991 3.4912 3.1713 3.3346

N1A−H1A···O2A N1B−H1B···O1B N3A−H3AA···O3B N3B−H3BA···O1B N3A−H3AB···O1S C6B−H6BB···O3A C4A−H4A···O2B C2B−H2BB···O1A C9B−H9A···O1A

0.91(2) 0.87(2) 0.89(2) 0.83(2) 0.94(4) 0.96(2) 1.03(2) 0.96(2) 0.95(2)

2.795(2) 2.915(3) 2.971(3) 3.312(3) 2.989(3) 3.226(3) 3.353(3) 3.316(3) 3.135(3)

N1A−H1A···O3A N3A−H3AC···O2A N3A−H3AD···O1A C4A−H4A···O3A

0.87 0.87 0.90 0.98

2.8051 3.0290 3.0659 2.8765

N1A−H1A···O1A N1B−H1B···O2B N3A−H3AA···O1A N3B−H3BA···O3A N3B−H3AB···O3B N3B−H3BB···O1S C6A−H6AB···O3B C4B−H4B···O2A C2A−H2AA···O1B C9B−H9B···O1B

0.87(4) 0.83(3) 0.85(4) 0.84(4) 0.86(4) 0.93(4) 0.98(3) 1.02(4) 0.93(3) 0.95(4)

2.901(3) 2.810(3) 3.357(3) 3.016(3) 3.064(3) 3.024(3) 3.165(3) 3.223(3) 3.360(3) 3.148(3)

N1A−H1A···O2A N1B−H1B···O1B N3A−H3AA···O3B N3B−H3BA···O1B N3A−H3AB···O1S N3B−H3BB···O3A C6B−H6BA···O3A C4A−H4A···O2B C9A−H9A···O1A C2B−HBB···O1A

0.89(2) 0.90(2) 0.88(3) 0.88(4) 0.96(4) 0.83(4) 0.97(2) 1.03(3) 0.97(4) 0.92(2)

2.808(2) 2.900(3) 3.008(3) 3.305(3) 3.077(3) 3.230(3) 3.184(3) 3.275(3) 3.140(3) 3.368(3)

N1−H1···O2S O2S−H2SA···O1S O1S−H1SA···O1

0.89(2) 0.89(2) 0.86(3)

2.851(2) 2.725(2) 2.780(2)

2.02 2.62 2.50 2.56 2.58 2.36

179 128 142 125 152 144

−x,−y,−z −x,1−y,−z x,y,1+z 1+x,y,z 1+x,y,z 1−x,−y,1−z

2.00 2.15 1.95 2.29 2.18 2.15 2.22 2.47 2.61 2.39 2.46

177 169 170 160 168 169 163 138 150 136 156

2−x,1/2+y,1/2−z 1−x,−y,−z x,1/2−y,1/2+z −1+x,1/2−y,−1/2+z x,1/2−y,−1/2+z 1−x,−y,−z x,y,z x,y,z x,1/2−y,−1/2+z 2−x,−1/2+y,1/2−z 1+x,1/2−y,1/2+z

1.88(2) 2.06(2) 2.08(2) 2.50(2) 2.18(4) 2.43(2) 2.38(2) 2.40(2) 2.50(2)

176(2) 170(2) 175.4(17) 167(2) 144(3) 139.9(18) 156.4(17) 160(2) 124.7(16)

1−x,1−y,1−z −x,1−y,1−z 1−x,−y,1−z x,1+y,z 1−x,−y,1−z −x,1−y,1−z −x,1−y,1−z x,−1+y,z x,−1+y,z

1.95 2.19 2.20 2.53

167 163 162 101

/2−x,−1/2+y,z /2+x,y,1/2−z 1−x,−y,1−z x,y,z

2.03(3) 1.99(3) 2.52(4) 2.19(4) 2.21(4) 2.16(4) 2.42(3) 2.24(4) 2.48(3) 2.50(4)

172(3) 173(4) 169(3) 168(4) 170(3) 154(3) 132.9(18) 161(3) 156(3) 126(3)

1−x,1−y,−z x,−y,2−z x,−1+y,z −x,1−y,1−z x,y,−1+z −x,1−y,1−z 1−x,−y,1−z 1−x,−y,1−z x,1+y,−1+z x,1+y,z

1.92(2) 2.02(2) 2.13(3) 2.43(3) 2.27(4) 2.47(4) 2.41(2) 2.31(4) 2.45(3) 2.51(2)

176(2) 167(2) 172(2) 168(2) 142(3) 153(3) 135.9(18) 155(2) 128.1(17) 156(2)

1−x,1−y,−z −x,−y,2−z 1−x,−y,1−z x,1+y,z 1−x,−y,1−z x,y,1+z −x,1−y,1−z −x,1−y,1−z x,−1+y,z x,−1+y,1+z

1.960(19) 1.84(2) 1.93(3)

173.7(17) 176(2) 174(3)

x,3/2−y,−1/2+z x,y,z −x,1−y,−z

Form-2

Form-3

Form-4 1 1

Form-5

Form-6

Form-7

616



Article

Table 2. continued interaction

D−H (Å)

D−A (Å)

H−A (Å)

D−H···A (deg)

symmetry

Form-7 O1S−H1SB···O3 N3−H3···O2S O2S−H2SB···O3 C3−H3B···O1 C9−H9···O2

0.89(3) 0.88(2) 0.88(3) 0.99 0.95

2.812(2) 3.116(2) 2.8188(18) 3.220(2) 3.107(2)

1.92(3) 2.24(2) 1.94(3) 2.31 2.41

(Form-1 and 4), two hydrates (Form-2, hemihydrate; Form-7, dihydrate) and three solvates (DMF, Form-3; DMSO, Form-5; and acetone, Form-6) of LDM. We were successful in obtaining the bulk materials and single crystals consistently for all the forms by following a variety of crystallization techniques. The crystallographic details for all solid forms of LDM are listed in Table 1. The structural comparison of all forms reveals three types (Types-1, 2, and 3) of possible motifs through N−H···O interactions (Scheme 1). The chemical structure of LDM (Scheme 1) indicates that the molecule is conformationally flexible and retains the ability to show different conformations in solid state. The conformation of the LDM moiety can be best explained by the six torsion angles τ1 (O1−C1−N1−C5), τ2 (N1−C1−C2−C3), τ3 (C5−C4−N2−C6), τ4 (N2−C4−C5−N1) τ5 (O3−C13−N2−C4), and τ6 (C5−C4−N2−C13). The corresponding angles for all forms are tabulated in Table S1. The maximum difference in the torsion angles (Δθ) of Forms-1, 2, 3, 4, 5, 6, and 7 was found to be ∼50−60°. This is clearly evident by overlaying the molecule from the asymmetric units of all forms (Figure 2). In the following section of the paper, we focused on the single crystal structure of all the seven forms of LDM. This analysis reveals vital information related to the packing motifs and interactions. Solid state phase transformations were studied using variable temperature (VT)-PXRD and VT-FTIR. The behavioral difference between solvates and hydrate is correlated to their crystal structure using Xpac analysis and is described in detail in the following section. The relation of iso-structurality to the solid phase transformation is revealed and described exhaustively in the paper. Physico-chemical stability is an essential part of any solid form development and this stability at biological pH conditions (pH = 1.2, 4.5, and 6.8) is considered to be a fingerprint test for a quick screen. In this regard, stability of all forms of LDM is investigated in the last section of the paper. 3.1. Characterization. 3.1.1. Single Crystal Analysis. LDM Form-1. In Form-1, the crystal packing plays a vital role, where hydrogen bonds (Table 2) control the supramolecular assembly of the molecules. In the crystal packing of Form-1 (Figure 3), the self-complementary pyridinedione ring forms centrosymmetric dimers via N1−H1N···O2 hydrogen bonding with a graph set R22 (8) (Figure 3a).58,59 The terminal amine forms a intermolecular hydrogen bond with O1 of the dione, linking the dimers to generate a cyclic tetramer with graph set R44 (26) (Figure 3b). The detailed crystal structure of LDM Form-1 is discussed by Ravikumar et al.55 LDM Form-2. The crystal structure of LDM Form-2 was solved and refined in a monoclinic crystal system with space group P21/c by Ravikumar et al. (Table 1).55 There are two molecules of LDM (denoted as A and B) and one molecule of water in the asymmetric unit. In the crystal packing of LDM Form-2 (Figure 4a), the self-complementary pyridinedione ring forms centrosymmetric dimers N1B−H4N···O1B (3.0161 Å, 169°) with a graph set of R22 (8).58,59 Molecule A links to

177(3) 173(2) 176(2) 152 130

x,3/2−y,−1/2+z 1−x,1−y,1−z x,−1+y,z x,3/2−y,1/2+z 1−x,2−y,1−z

Figure 3. Three dimensional packing diagram of LDM Form-1 showing graph set of (a) R22(10), R22(8) and (b) R44(26) via N−H···O and C−H···O intermolecular interaction.

molecule B through pyridine atom N1A and carbonyl atom O3B. The water molecule was found to be in a cavity and held by N−H···O and O−H···O intermolecular interaction (Figure 4b). It is interesting to note that the water molecule interacts only with molecule B. Carbonyl atom O2A of the pyridinedione ring is not involved in any conventional hydrogen bonds but participates in an acceptable C−H···O intermolecular interaction (Table 2). The detail crystal structure of LDM Form-2 is discussed by Ravikumar et al.55 LDM Form-3. Block morphology crystals of LDM Form-3 were obtained from a saturated solution of DMF at room temperature. The crystal structure was solved and refined in a triclinic cell with space group P1̅ (Table 1). There are two molecules of LDM (denoted as A and B) and one molecule of DMF in the asymmetric unit. Self-complementary dimers with graph set R22 (8) involving pyridinedione ring are found in symmetry equivalent molecules of Form-3 via N1A-H1A···O2A (2.795(2) Å, 176(2)°) and N1B−H1B···O1B (2.915(3) Å, 170(2)°) intermolecular hydrogen bonds (Figure 5a). It is interesting to note that O2 is involved in the centrosymmetric dimer formation for molecule A, whereas O1 is involved in molecule B. The DMF solvent molecule is held by N−H···O(S) and C−H···O(S) interactions with two symmetry independent LDM molecules. In the three-dimensional packing motif, LDM molecules form channels along the b-axis to accommodate DMF molecules (Figure 5b). 617



Article

in the asymmetric unit (Table 1). In the crystal packing of Form-4, the pyridinedione ring is connected with three other LDM molecules via N−H···O intermolecular hydrogen bonding. The terminal nitrogen atom (N3) from the amine group of LDM forms a hydrogen bond with O1 of the pyridinedione via N3−H3AD···O1 (3.0659 Å 162°) to form a dimer of graph set R22 (22) (Figure 6a). These selfcomplementary tetramers form infinite chains parallel to the c-axis (Figure 6b).

Figure 6. Three dimensional packing motif of LDM Form-4 showing (a) graph set R22 (22) via N−H···O intermolecular interaction and (b) packing diagram showing the self-complementary tetramers forming infinite chains along the c-axis.

Figure 4. Three dimensional packing diagram of LDM Form-2 showing (a) graph set R22 (8) via N−H···O intermolecular interaction and (b) water in the pocket.

LDM Form-5. Block-type crystals of LDM Form-5 were isolated from saturated solution of DMSO at room temperature. The crystal structure was solved in a triclinic cell and refined with P1̅ space group (Table 1). The asymmetric unit consist of two molecules of LDM (denoted as A and B) and one molecule of DMSO. As in LDM Form-3, selfcomplementary dimers graph set R22 (8) involving the pyridinedione ring is found in LDM Form-5 (Figure 7a). The dimer in LDM Form-5 is formed between symmetry equivalent molecules via N1A-H1A···O2A (2.795(2) Å, 176(2)°) and N1B−H1B···O1B (2.915(3) Å, 170(2)°) intermolecular hydrogen bonds (Figure 7a). Also similar to Form-3, O2 is involved in the centrosymmetric dimer formation for molecule A, whereas O1 is involved in molecule B. The DMSO solvent molecule is held by N−H···O(S) and C−H···O(S) interactions with two symmetry independent LDM molecules in the channel along the b-axis (Figure 7b). LDM Form-6. Single crystals of LDM Form-6 (block morphology) were isolated from a saturated solution of acetone at room temperature. The crystal structure was solved and refined in a triclinic cell with space group P1̅ (Table 1). There are two symmetry independent molecules of LDM (denoted as A and B) and one molecule of acetone in the asymmetric unit. In line with Form-3 and Form-5, a self-complementary dimer with a graph set R22 (8) involving the pyridinedione ring is also found in Form-6 (Figure 8a). The dimer involves symmetry equivalent molecules of LDM, and the dimers are formed via N1A-H1A···O2A (2.808(2) Å, 176(2)°) and N1B−H1B···O1B (2.900(3) Å, 167(2)°) intermolecular hydrogen bonds (Figure 8a). As in Form-3 and Form-5, O2 and O1 are involved in the centrosymmetric dimer formation for molecule A and molecule B, respectively. The acetone molecule in the three-dimensional crystal lattice is held by N−H···O(S) and C−H···O(S) intermolecular interactions originated from two symmetry independent

Figure 5. Three dimensional packing motif of LDM Form-3 showing (a) graph set R22 (8) via N−H···O intermolecular interaction and (b) DMF in the channel along the b-axis.

LDM Form-4. Single crystals (block) of LDM Form-4 were obtained by vapor diffusion technique using trifluoro-ethanol as solvent and toluene as antisolvent at room temperature. The crystal structure was solved and refined in the orthorhombic unit cell with space group Pbca, having one molecule 618



Article

in 2 mL of 0.1 N HCl (pH 1.2) and leaving the solution for crystallization for 1 h. The crystal structure was solved and refined in the monoclinic system with space group P21/c (Table 1). There are two molecules of water and one molecule of LDM in the asymmetric unit. It is interesting to note that in Form-7 there is no conventional hydrogen bonding between two LDM molecules. However, water is acting as a bridge between the two LDM molecules (Figure 9a). In the crystal packing of

Figure 7. Three dimensional packing motif of LDM Form-5 showing (a) graph set R22 (8) via complementary N−H···O intermolecular interaction and (b) packing diagram showing DMSO as a space fill model in the channel along the b-axis.

Figure 9. Three dimensional packing motif of LDM Form-7 showing (a) the O−H···O intermolecular interaction holding the water molecules, (b) graph set R22 (18) via C−H···O intermolecular interaction, and (c) the network of water molecules in the crystal lattice.

Form-7, the pyridinedione ring is connected to the water molecule via N1−H1···O2S (2.851(2) Å, 173.7(17)°) and O2S−H2SA···O1S (2.725(2) Å, 176(2)°) hydrogen bonding. Further, the terminal amine group of LDM connects to neighboring water molecules via N3−H3···O2S (3.116(2) Å, 173(2)°) hydrogen bonding and provides additional stability (Figure 9a). A centrosymmetric dimer with a graph set of R22 (18) synthon is formed via C−H···O intermolecular interaction (Figure 9b). The three-dimensional packing motif of Form-7 is shown in Figure 9c where water molecules are represented in the space filling model. 3.1.2. Vibrational Spectroscopy. Vibrational spectroscopy has been frequently acknowledged to be an important analytical technique for the solid state characterization of polymorphic forms of drug substances and drug products.60,61 ATR-FTIR has the capability to provide detailed information in situ on a molecular level, and it is also frequently utilized to study the dehydration/desolvation processes, solid−solid transition, etc. VT-ATR-FTIR has also become a dependable workhorse for characterizing hydrogen bonding pattern changes,61,62 determining the interactions between water and polymer matrixes,63−65 relating the molecular role in the cooperative structural rearrangements on dehydration,66 etc. FT-IR spectra overlay of different crystalline forms of LDM with their unique color code are shown in Figure 10. Set of bands corresponding to unique type of

Figure 8. Three dimensional packing motif of LDM Form-6 showing (a) graph set R22 (8) via complementary N−H···O intermolecular interaction and (b) packing diagram showing acetone along the b-axis in the space fill model.

LDM molecules. LDM Form-6 forms channels along the b-axis to accommodate the acetone molecule (Figure 8b). LDM Form-7. Colorless block shape single crystals of LDM Form-7 were obtained by dissolving 150 mg of LDM Form-1 619



Article

Figure 10. Vibrational spectra overlay of different crystalline forms of LDM zoomed in to regions (a) 4000−2500, (b) 1100−1300, (c) 1100−1300, and (d) 430−480 cm−1.

in hydrogen bonding at the aniline NH2, imide N−H, and carbonyl give rise to distinct changes in the 3500−3300, 3250−3140, and 1650−1750 cm−1 regions of the spectrum, for various forms can be used to discriminate between hydrates, solvates, and anhydrates. There are several other bands that discriminate the individual polymorphs in the aliphatic C−H regions (3000−2800 and 1500−1300 cm−1). Furthermore, the aromatic C−H (3100−3000 cm−1) and CC stretching regions (∼1600 cm−1) display peaks distinct for individual polymorphs, suggesting differential intermolecular interactions between the forms of LDM. The scope of this study is to develop protocols to understand the thermal stability of individual polymorphs as well as provide a mechanistic understanding of the diffusion and structural arrangement or cooperative mechanism of individual polymorphs, which is discussed in the following section. 3.1.3. Powder X-ray Diffraction. Most of the routine synthetic routes employed for drug substances result in polycrystalline powders. Hence, PXRD is considered as a vital

Table 3. Infrared Spectral Assignments of Lenalidomide −1

wavenumber (cm )

functional group

3475−3310 3250−3140 3080−3050 2975−2840 1740−1620 1610−1400 1300−1200 1200−400

aniline N−H stretching imide N−H stretch aromatic C−H stretch aliphatic C−H stretch carbonyl aromatic CC stretch and C−H deformation aniline C−N stretch fingerprint region

vibrations in LDM are tabulated in Table 3. The peaks appearing around 3566 and 3506 cm−1 are characteristic for lattice water (crystalline water peak) that can be used to discriminate hydrates (especially Form-2) from anhydrates. Moreover, the appearance of a strong and wide peak between 3450 and 3200 cm−1 (centered ∼3407 cm−1) along with the 920 cm−1 peak could be assigned to dihydrate.60 The differences 620



Article

with an onset at 266.7 °C. This endotherm for a ΔH = 149.51 J/g represent the melting event, and there was no indication of any phase transformation prior to melting (Figure 12). As all forms

and essential tool for the study of polycrystalline materials and particularly for polymorphs.67,68 The experimental PXRD patterns for all samples were collected on a PANalytical X’Pert PRO diffractometer with scan range of 2θ−3 to 40° for a step size of 0.03 and an exposure of 1200 s per step. Experimental powder X-ray diffraction patterns for all forms of LDM matches with their corresponding simulated powder pattern (see the Supporting Information, Figures S23−S29). A different set of peaks are identified as a characteristic peak position for identification of a particular form (Table 4). The experimental powder X-ray diffraction patterns for bulk samples are shown in Figure 11. Table 4. Characteristic Peaks for Different Crystalline Forms of LDM characteristic peak position (2θ, °) Form-1

Form-2

Form-3

Form-4

Form-5

Form-6

Form-7

7.93 11.34 14.39 7.67 7.93 12.08 7.94 8.7 17.1 10.5 11.66 12.51 7.79 8.85 15.61 7.91 8.76 14.5 12.02 13.65 15.29

(h k l) (0 (0 (0 (0 (0 (1 (0 (0 (2 (0 (1 (1 (0 (1 (0 (0 (0 (1 (1 (1 (0

0 1) 1 1) 1 −1) 1 1) 2 0) 1 −1) 0 1) 1 0) 0 0) 0 2) 1 1) 0 2) 0 1) 0 0) 0 2) 0 1) 1 0) −1 0) 1 0) 1 −1) 0 2)

Figure 12. DSC thermogram overlay of different crystalline forms of LDM.

(LDM Forms-2 to 7) of LDM do undergo a crystal to crystal phase transformation to Form-1 below 269.1 °C (melting point of Form-1), it is hard to find the melting event for them. The desolvation/dehydration and the melting point events are mentioned in Table 5. Phase transformations for different forms of LDM are explained methodically in the subsequent section. 3.1.5. Hirshfeld Surface and 2D Fingerprint Plot. The molecular Hirshfeld surface71−74 represents the potential area of interaction in a molecule. Hence, its analysis gives additional insight into the intermolecular interactions in the crystal structure. Further, Hirshfeld surfaces enable a rapid and easy visual comparison of the polymorphs. Out of seven forms of LDM, only Forms-1 and 4 are anhydrous. Therefore, Hirshfeld molecular surfaces and the associated fingerprint plots were generated for these two forms using Crystal Explorer (version 3.0, revision 1262M, build 3.0.1262M). In both

3.1.4. Thermal Analysis. DSC has unique advantages, including the ease, simplicity, and rapidity of the measurement providing direct information on thermodynamic parameters associated with crystalline/amorphous phase.69,70 Heating of LDM Form-1 in DSC results a sharp endothermic peak at 269.1 °C

Figure 11. Experimental powder X-ray diffraction pattern overlay of different crystalline forms of LDM. 621



Article

Table 5. Dehydration/Desolvation and Melting Events in DSC Thermograms of Different Forms of LDM solvent Form-1 Form-2 Form-3 Form-4 Form-5 Form-6 Form-7

anhydrous (neat) water (hemihydrate) DMF anhydrous (neat) DMSO acetone water (dihydrate)

dehyration/desolvation temperature Tonset/Tpeak (°C) 154.24/159.23 140.29/156.51 196.58/199.26 136.77/143.72 87.77/100.93 and 140.30/158.80

mp Tonset/Tpeak (°C) 266.78/269.10 266.92/269.35 262.84/265.52 268.67/269.93 268.82/271.66 268.89/271.64 270.64/273.20

molecular surfaces of Form-1 and Form-4, regions of N−H···O hydrogen bonds can be visualized by large red circles (Figure 13). The relative contributions from O···H [Form-1 (34.8%) and Form-4 (33.8%)], N···H [Form-1 (3.7%) and Form-4 (4.8%)], and other interactions are shown in Figure 14. Interactions such as N···H, O···H, and C···H comprise 51.6% of the total Hirshfeld surface in Form-1 and 54.9% in Form-4. Contribution due to C···C contacts is negligible (0.4%) for both forms, indicating the absence of π···π interactions. The structure of Form-1 has less N···H and C···H contacts and more O···H contacts compared to those in Form-4. From the crystal structure and Hirshfeld fingerprint plot, it is evident that the interactions in Forms-1 and 4 are different. 3.2. Solid Phase Transformation Study. 3.2.1. Thermal Study. Solid state phase transformation in pseudopolymorphs

Figure 14. Percentage contribution of different types of intermolecular interaction present in LDM Form-1 and Form-4.

(solvate and hydrate) of LDM was studied extensively using different thermal techniques such as DSC, TGA, and HSM. The crystal luster and transparency is found to be lost immediately after desolvation or dehydration in pseudopolymorphs of LDM. The crystal image at different stages of heating in hot stage microscopy is shown in Figure 15. The DSC data of all solvates (Forms-3, 5, and 6) show two endothermic events, one corresponding to the desolvation, and the other matching to the melting event of Form-1. The desolvation temperature is tabulated in Table 5, and the event is further supported by the weight loss in TGA. It is expected that the desolvation and

Figure 13. Hirshfeld surface maps of (a) LDM Form-1 and (b) LDM Form-4. (c) Hirshfeld fingerprint plot of LDM Form-1 and Form-4. 622



Article

Figure 15. DSC, TGA, and hot stage microscopy results for different pseudopolymorphs of LDM.

phase transformation in all of these solvates happen at the same time. This particular event is evident by the VT-PXRD data and is discussed in the following section. Hydrates of LDM (Forms-2 and 7) do not show thermal behavior similar to that of solvates. DSC data of LDM Form-2 (hemihydrate) shows four endothermic events. The endotherm with onset at 154.24 and peak at 159.23 °C represents the dehydration event. There are two endothermic events between 160 and 190 °C, representing the phase transformation. The first endotherm can be a transformation of Form-2 to Form-4, and the second is Form-4 to Form-1 because the last endothermic peak and the corresponding onset matches that of Form-1. On the other hand, LDM Form-7 (dihydrate) shows an initial broad endotherm with onset at 87.77 °C and peak at 100.93 °C, corresponding to a loss of 1.5 mol of water and solid phase transformation to Form-2. This fact is further supported by following the TGA weight loss corresponding to 0.5 mol water at around 150 °C. The DSC thermogram of Form-7 shows another broad endotherm with onset at 140.30 and peak at 158.80 °C. This can be assigned to multiple events of dehydration and phase transformation to Form-1. As in Form-2, the last endotherm matches with that of Form-1. The solid to solid phase transformation and its complexity is further studied with VT-PXRD and VT-FTIR in the subsequent sections. 3.2.2. Variable Temperature Powder X-ray Diffraction Study. Lenalidomide solid forms were subjected to six different temperatures (25, 50, 100, 150, 200, and 250 °C), and corresponding in situ powder X-ray data were collected. The presence or absence of a particular phase in PXRD pattern is trivial, so a set of characteristic peaks were chosen to identify each of the forms (Form-1 to 7). The list of charecteristic peaks against each form with their corresponding Miller indices are given in Table 4. None of the forms show any phase transformations up to 100 °C except for Form-7 (converts to Form-2). In contrast to the solvates (Forms-3, 5, and 6), the hydrates (Forms-2 and 7) show a phase transformation to Form-4 before completely converting to the thermodynamically stable Form-1. The phase transformation and PXRD comparison are pictorially presented in Figure 16, and the corresponding PXRD pattern overlays are given in the Supporting Information, Figures S30−S36).

Figure 16. Pictorial representation of phase transformation obtained using variable temperature powder X-ray diffraction for different crystalline forms of LDM.

3.2.3. Variable Temperature Fourier Transform Infrared Spectroscopy. Five pseudopolymorphs of LDM (Forms-2, 3, 5, 6, and 7) were subjected to VT-FTIR-ATR ranging from 30 to 200 °C to investigate the thermally induced phase transformation. Because selection of universal characteristic peaks to monitor various forms of LDM was not available in literature, the unique peaks were selected based on the starting form, transition form, and the final form, as shown in Table 3. The plots of peak intensity vs temperature are shown in Figures 17a and b and 18a−c for a range of 30−150 °C for hydrates and 30−200 °C for solvates with an interval of 20 °C. From Figure 17a, it is clearly indicated that the thermal dependency of IR spectral intensity of LDM Form-2 was markedly changed with a four-step process within temperature ranges at 30, 50−70, 90−130, and 150 °C (refer to Section 3.2.1). The LDM spectra captured at 30 °C indicates the actual 623



Article

Figure 17. VT-FTIR phase transformation for (a) LDM Form-2 and (b) LDM Form-7, depicting intermittent Form-4 formation.

Figure 18. VT-FTIR phase transformation for (a) LDM Form-3, (b) LDM Form-5, and (c) LDM Form-6, illustrating direct conversion to Form-1.

150 °C. The unique peaks selected for Forms-3, 5, and 6 are 468, 1032, and 1198 cm−1, respectively. The characteristic peak selected for Form-1 tracking in the cases of From-3 and Form-6 is ∼2910 cm−1, whereas it is 3407 cm−1 for Form-5. The plot of peak intensity vs temperature for all three solvates is shown in Figure 18. The results indicate that the main transition event of Form-3 and Form-5 occur around the solvates boiling points, which are ∼156 and 186 °C, respectively. While for acetone, although the boiling point is around 55 °C, the transition happens around 140 °C, which is closer to the Form-1 conversion temperature. 3.3. Dehydration/DesolvationMechanism. The phase change in pseudopolymorphs of LDM is associated with a structural rearrangement translated through the collective changes and is directly related to the loss of water/solvent molecules from the lattice. The coexisting phases for dihydrate during dehydration process are hemihydrate, Form-4, and Form-1, while with hemihydrate it is Form-4 and Form-1. However, with solvates, the phase transition to Form-1 happens once the solvent diffuses out completely from the channel. The single crystal structure has clearly indicated the presence of large channel holding the solvent molecule (Figure 19). The void in the channels allows the solvate to display one-step desolvation for the escape of DMF, DMSO, or acetone in Forms-3, 5, and 6, respectively (Figure 19). The solvent escape leads to empty channels and possibly demonstrates the wellknown diffusion phenomenon. Single crystal structures for the hydrates indicate that the water is held by the LDM molecule and work like a bridge in the crystal structure. Hence, mere escape of the water molecules is impossible, showcasing structural rearrangements visible during water loss to Form-4. 3.3.1. XPac Analysis for the Understanding of Dehydration and Desolvation Behavior. There are several methods for the comparison of two crystal structures, among which XPac is the most promising one in the recent literature to measure

hemihydrate spectra without alteration, confirmed by the presence of bands visible around 3566 cm−1 and absence of bands around 1170 cm−1 (Form-4) and 3407 cm−1 (Form-1). As the temperature is raised, the loss of crystalline water displayed the structural rearrangement viz. Form-4 by the appearance of the peak around 1170 cm−1 (Figure 17a). The formation of Form-4 is clearly visible around 50 °C (Figure 17a). The appearance of Form-4 at 50 °C is inconsistence with the corresponding DSC and VT-PXRD data. However, considering hydrates to be a very strong absorbers of infrared radiation, subtle structural differences are captured well in FTIR. Moreover, the transition from hydrate (Form-2) to Form-4 to Form-1 involves sequential shifting of molecular level hydrogen bonding interactions between the groups. Furthermore, these H-bonding interactions involve strong dipole moment changes; hence, the event discrimination is better with VT-FTIR. However, DSC and PXRD consider the bulk effect; therefore, subtle differences may not be captured well. This could be one of the potential reasons for the observed inconsistency below 50 °C. When the temperature is around 150 °C, the structural water was found to decrease rapidly, and disappearance of Form-4 with the formation of Form-1 is observed. However, with Form-7, the VT-FTIR results were slightly complicated. At 50 °C, the dihydrate initially converted to hemihydrate (tracked by the presence of peak around 3566/3506 cm−1, Figure 17b). The hemihydrate loses the remaining water of crystallization above 150 °C, and intermediate structure Form-4 started appearing. In the later stage, this Form-4 is converted to the anhydrate Form-1. The results of detection of intermediate structures for hydrates using variable temperature-FTIR are in line with those obtained from other orthogonal experiments such as VT-PXRD and DSC. VT-FTIR of solvates to understand the desolvation phenomena was carried out, and it was found that Forms-3, 5, and 6 upon thermal treatment convert to Form-1 above 624



Article

Figure 19. Connolly surface showing a pocket for hydrates (a) LDM Form-2 and (b) LDM Form-7 and a channel for solvates (c) LDM Form-3, (d) LDM Form-5, and (e) LDM Form-6 drawn at grid intervals = 0.4 Å; vdW scale factor = 1.0; Connolly radius = 1.0 Å.

Figure 20. Xpac analysis showing packing similarity between (a) LDM Forms-3 and 1, (b) LDM Forms-5 and 1, (c) LDM Forms-6 and 1, and (d) LDM Forms-2 and 4.

packing similarity in a quantitative manner.75−77 Xpac (version 2.0.2) analysis is carried out for different crystalline forms of LDM. It is interesting to note that LDM Forms-3, 5, and 6

(namely DMF, DMSO, and acetone) solvates have an isostructurality with Form-1 with dissimilarity index X = 8.1, 7.2, and 8.4, respectively, with one-dimensional iso-structurality 625



Article

and Form-4 were calculated using the morphology module of Material Studio 8.0 (BIOVIO) (see the Supporting Information, Table S3). The lattice energy for Form-1 and Form-4 were found to be −103.828 and −429.156 kcal/mol, respectively. This value further supports the FTIR (shift in N−H stretching) and DSC (phase transformation) data.

(Figure 20 and Supporting Information, Figure S37). Form-2 has an iso-structurality with Form-4 with dissimilarity index X = 9.7 with zero-dimensional iso-structurality. Iso-structurality of Forms-3, 5, and 6 could be one of the reasons for the conversion of all solvates (Forms-3, 5, and 6) to Form-1 directly under thermal treatment. At the same time, it also explains that isostructurality in Forms-2 and 4 is one of the reasons for the phase transformation of Form-2 to Form-4. On the other hand, the dissimilarity in iso-structurality between Form-2 and Form-1 does not allow the first form to convert to the second directly in the solid state. The matrix representation of iso-structurality between Forms-1 to 7 is given in the Supporting Information (Table S2). 3.4. Stability of LDM Solid Forms. 3.4.1. Stability of LDM Forms in Different Biological Conditions. The crystalline form stability for different forms of LDM was evaluated by subjecting them to different biological conditions such as pH 1.2, 4.5, and 6.8. In this regard, 150 mg of different crystalline forms of LDM (Forms-1 to 7) were suspended in 2 mL of respective buffer solution. The suspension was then subjected to 150 rpm for 30 min at 37 °C. After adequate vortexing, the suspension was filtered, and the filtrate was analyzed using PXRD. The results for different crystalline forms are shown pictorially in Figure 21. Interestingly in all biological pH

4. CONCLUSIONS Polymorphism of pharmaceuticals is a highly complex and well-recognized phenomena, and not directing significant investigation will have profound effects on finished drug product performance. High-throughput polymorph screening of lenalidomide yielded seven crystalline forms that have been extensively characterized and described in the present study, including their crystal structures. A suite of complementary theoretical (Xpac and Material Sudio BIOVIO) and experimental (PXRD, TGA, DSC, and FTIR) methods were applied and were found to be very beneficial in tracing the mechanism of dehydration and desolvation phenomena and provided valuable insights that are transferable to product development. Interestingly all solvates upon desolvation convert to Form-1 (thermodynamically stable anhydrous form), whereas all hydrates upon dehydration convert to Form-4 (novel anhydrous form), and further heating of the same lead to Form-1. Thorough investigations revealed that iso-structurality is one of the responsible factors for such bifurcated phase transformations. This study brings about the necessity of further experiments to investigate the kinetics of conversion and transition temperature that would be considered a prerequisite information for formulation development.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.cgd.6b01462. PXRD, DSC, and FT-IR results for Forms-1 to 7 (Figures S1−S21); concomitant crystallization of Form-4 with Form-1 (Figure S22), simulated and experimental PXRD overlays of Forms-1 to 7 (Figures S23−S29); VTPXRD data for Forms-1 to 7 (Figures S30−S36); Xpac analysis showing dissimilarity index of LDM psuedopolymorphs (Figure S37); PXRD overlay of different forms at pH values of 1.2, 4.5, and 6 for Forms-1 to 7 (Figures S38−S44); torsion angles of Forms-1 to 7 (Table S1); matrix representation of iso-structurality in Forms-1 to 7 (Table S2); parameters for the lattice energy calculation in material studio 8.0 (BIOVIO) (Table S3) (PDF)

Figure 21. Stability of different crystalline forms of lenalidomide at different biological pH conditions.

conditions, Forms-2 and 7 remain as is. Irrespective of any biological pH conditions, Forms-3 and 5 convert to Form-2 and Form-7, respectively. At pH 1.2, all crystalline forms convert to Form-7 except for Form-3, which converts to Form-2. At pH 4.5 and 6.8, Forms-4 and 6 convert to Form-2, whereas Form-1 remains as is (see the Supporting Information, Figures S38−S44). 3.4.2. Relation of Hydrogen Bond Strength and Lattice Energy with the Physical Stability of Polymorphs. To understand the physical stability of various polymorphs of LDM (Form-1 and Form-4), the NH stretching frequency of the aniline moiety is considered. The NH stretching frequency of Form-1 appears at 3407 cm−1, while for Form-4, the frequency is observed at 3455 cm−1. From the polymorphic characterization, it is well understood that Form-1 displays stronger intermolecular interactions in comparison with those of Form-4. According to the “infrared rule” of Burger and Ramberger,78,79 Form-4 displays higher N−H stretching modes than those of Form-1, implying that the H bonding interactions are weaker and demonstrating larger entropy and the possibility to convert to more stable Form-1. That said, the correlation of N−H stretching frequency to physical stability of pseudopolymorphs could be difficult because the latter displays more than one type of H bond. Further, the lattice energy for Form-1

Accession Codes

CCDC 1487162−1487166 contains the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by emailing [email protected], or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, U.K.; fax: +44 1223 336033.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; Tel.: +91-40-44346040; Fax: +91-40-44346164. ORCID

Sudarshan Mahapatra: 0000-0001-6369-8811 626



Article

(30) Cinčić, D.; Frišcǐ ć, T.; Jones, W. New J. Chem. 2008, 32 (10), 1776−1781. (31) Wood, P. A.; Oliveira, M. A.; Zink, A.; Hickey, M. B. CrystEngComm 2012, 14 (7), 2413−2421. (32) Thakuria, R.; Nangia, A. Cryst. Growth Des. 2013, 13 (8), 3672− 3680. (33) Rubcic, M.; Milić, D.; Pavlović, G.; Cindrić, M. Cryst. Growth Des. 2011, 11 (12), 5227−5240. (34) Rajput, L.; Biradha, K. New J. Chem. 2010, 34 (11), 2415−2428. (35) Nassimbeni, L. R.; Su, H.; Weber, E. New J. Chem. 2008, 32 (10), 1702−1709. (36) Nanubolu, J. B.; Sridhar, B.; Ravikumar, K.; Cherukuvada, S. CrystEngComm 2013, 15 (21), 4321−4334. (37) Galcera, J.; Frišcǐ ć, T.; Molins, E.; Jones, W. CrystEngComm 2013, 15 (7), 1332−1338. (38) Galcera, J.; Frišcǐ ć, T.; Hejczyk, K. E.; Fábián, L.; Clarke, S. M.; Day, G. M.; Molins, E.; Jones, W. CrystEngComm 2012, 14 (23), 7898−7906. (39) Fábián, L.; Kálmán, A.; Argay, G.; Bernáth, G.; Gyarmati, Z. C. Cryst. Growth Des. 2005, 5 (2), 773−782. (40) Ebenezer, S.; Muthiah, P. T.; Butcher, R. J. Cryst. Growth Des. 2011, 11 (8), 3579−3592. (41) Csöregh, I.; Hirano, S.; Toyota, S.; Bombicz, P.; Toda, F. CrystEngComm 2004, 6 (14), 60−69. (42) Cinčić, D.; Frišcǐ ć, T.; Jones, W. Chem. - Eur. J. 2008, 14 (2), 747−753. (43) Childs, S. L.; Wood, P. A.; Rodríguez-Hornedo, N.; Reddy, L. S.; Hardcastle, K. I. Cryst. Growth Des. 2009, 9 (4), 1869−1888. (44) Braun, D. E.; Gelbrich, T.; Kahlenberg, V.; Tessadri, R.; Wieser, J.; Griesser, U. J. Cryst. Growth Des. 2009, 9 (2), 1054−1065. (45) Bhattacharya, S.; Sameena, J.; Saha, B. K. Cryst. Growth Des. 2011, 11 (4), 905−909. (46) Bhattacharya, S.; Saha, B. K. Cryst. Growth Des. 2013, 13 (2), 606−613. (47) Reddy, C. M.; Kirchner, M. T.; Gundakaram, R. C.; Padmanabhan, K. A.; Desiraju, G. R. Chem. - Eur. J. 2006, 12 (8), 2222−2234. (48) Owczarzak, A. M.; Kourkoumelis, N.; Hadjikakou, S. K.; Kubicki, M. CrystEngComm 2013, 15 (18), 3607−3614. (49) Nath, N. K.; Saha, B. K.; Nangia, A. New J. Chem. 2008, 32 (10), 1693−1701. (50) Moorthy, J. N.; Mandal, S.; Venugopalan, P. Cryst. Growth Des. 2012, 12 (6), 2942−2947. (51) Moorthy, J. N.; Venkatakrishnan, P.; Mal, P.; Dixit, S.; Venugopalan, P. Cryst. Growth Des. 2003, 3 (4), 581−585. (52) Asmadi, A.; Kendrick, J.; Leusen, F. J. J. Chem. - Eur. J. 2010, 16 (42), 12701−12709. (53) Anthony, A.; Nangia, A.; Desiraju, G. R.; Jaskólski, M. Chem. Commun. 1998, No. 22, 2537−2538. (54) Anthony, A.; Jaskólski, M.; Nangia, A. Acta Crystallogr., Sect. B: Struct. Sci. 2000, 56 (3), 512−525. (55) Ravikumar, K.; Sridhar, B. Acta Crystallogr., Sect. C: Cryst. Struct. Commun. 2009, 65 (10), o502−o505. (56) Song, J.-X.; Chen, J.-M.; Lu, T.-B. Cryst. Growth Des. 2015, 15 (10), 4869−4875. (57) Song, J. X.; Yan, Y.; Yao, J.; Chen, J. M.; Lu, T. B. Cryst. Growth Des. 2014, 14 (6), 3069−3077. (58) Etter, M. C. Acc. Chem. Res. 1990, 23 (4), 120−126. (59) Bernstein, J.; Davis, R. E.; Shimoni, L.; Chang, N. Angew. Chem., Int. Ed. Engl. 1995, 34 (15), 1555−1573. (60) Threlfall, T. L. Analyst 1995, 120 (10), 2435−2460. (61) Krivoshein, A. Curr. Pharm. Des. 2016, 22 (32), 5029−5040. (62) Ayala, A. P.; Siesler, H. W.; Boese, R.; Hoffmann, G. G.; Polla, G. I.; Vega, D. R. Int. J. Pharm. 2006, 326 (1), 69−79. (63) Banks, S. R.; Sammon, C.; Melia, C. D.; Timmins, P. Appl. Spectrosc. 2005, 59 (4), 452−459. (64) Bajwa, G. S.; Sammon, C.; Timmins, P.; Melia, C. D. Polymer 2009, 50 (19), 4571−4576.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors would like to thank Dr. Reddy’s Laboratories for encouragement and providing the facility to do fundamental research. We would also like to thank Prof. T. N. Guru Row for providing access to the SXRD data collection and Dr. Rajesh Thipparaboina for useful scientific discussion.

■

REFERENCES

(1) Reutzel-Edens, S. M. Curr. Opin. Drug Discovery Dev. 2006, 9 (6), 806. (2) Müller, M.; Meier, U.; Wieckhusen, D.; Beck, R.; Pfeffer-Hennig, S.; Schneeberger, R. Cryst. Growth Des. 2006, 6 (4), 946−954. (3) Morissette, S. L.; Almarsson, Ö .; Peterson, M. L.; Remenar, J. F.; Read, M. J.; Lemmo, A. V.; Ellis, S.; Cima, M. J.; Gardner, C. R. Adv. Drug Delivery Rev. 2004, 56 (3), 275−300. (4) Babu, N. J.; Nangia, A. Cryst. Growth Des. 2011, 11 (7), 2662− 2679. (5) Trask, A. V. Mol. Pharmaceutics 2007, 4 (3), 301−309. (6) Morissette, S. L.; Soukasene, S.; Levinson, D.; Cima, M. J.; Almarsson, O. Proc. Natl. Acad. Sci. U. S. A. 2003, 100 (5), 2180−2184. (7) Byrn, S.; Pfeiffer, R.; Ganey, M.; Hoiberg, C.; Poochikian, G. Pharm. Res. 1995, 12 (7), 945−954. (8) Bernstein, J. Polymorphism in Molecular Crystals; Clarendon Press: Oxford, 2002. (9) Threlfall, T. L. Analyst 1995, 120 (10), 2435−2460. (10) Huang, L.-F.; Tong, W.-Q. T. Adv. Drug Delivery Rev. 2004, 56 (3), 321−334. (11) Datta, S.; Grant, D. J. W. Nat. Rev. Drug Discovery 2004, 3 (1), 42−57. (12) Bauer, J.; Spanton, S.; Henry, R.; Quick, J.; Dziki, W.; Porter, W.; Morris, J. Pharm. Res. 2001, 18 (6), 859−866. (13) McCrone, W. C. Fox, MM Labes A. Weissberger 1965, 725−767. (14) Vishweshwar, P.; McMahon, J. A.; Oliveira, M.; Peterson, M. L.; Zaworotko, M. J. J. Am. Chem. Soc. 2005, 127 (48), 16802−16803. (15) Ouvrard, C.; Price, S. L. Cryst. Growth Des. 2004, 4 (6), 1119− 1127. (16) Day, G. M.; Trask, A. V.; Motherwell, W. D. S.; Jones, W. Chem. Commun. (Cambridge, U. K.) 2006, 1 (1), 54−56. (17) Chennuru, R.; Muthudoss, P.; Ramakrishnan, S.; Mohammad, A. B.; Ravi Chandra Babu, R.; Mahapatra, S.; Nayak, S. K. J. Mol. Struct. 2016, 1120, 86−99. (18) Rodríguez-Spong, B.; Price, C. P.; Jayasankar, A.; Matzger, A. J.; Rodríguez-Hornedo, N. Adv. Drug Delivery Rev. 2004, 56 (3), 241− 274. (19) Qiu, Y.; Chen, Y.; Zhang, G. G. Z.; Liu, L.; Porter, W. Developing solid oral dosage forms: pharmaceutical theory & practice; Academic Press: Cambridge, MA, 2009. (20) Byrn, S. R.; Bates, S.; Ivanisevic, I. Am. Pharm. Rev. 2005, 8 (5), 137. (21) Blagden, N.; De Matas, M.; Gavan, P. T.; York, P. Adv. Drug Delivery Rev. 2007, 59 (7), 617−630. (22) Kawabata, Y.; Wada, K.; Nakatani, M.; Yamada, S.; Onoue, S. Int. J. Pharm. 2011, 420 (1), 1−10. (23) Han, J.; Suryanarayanan, R. Pharm. Dev. Technol. 1998, 3 (4), 587−596. (24) Chekal, B. P.; Campeta, A. M.; Abramov, Y. a.; Feeder, N.; Glynn, P. P.; McLaughlin, R. W.; Meenan, P. a.; Singer, R. a. Org. Process Res. Dev. 2009, 13 (6), 1327−1337. (25) Agbada, C. O.; York, P. Int. J. Pharm. 1994, 106 (1), 33−40. (26) Jones, P. G.; Vancea, F. CrystEngComm 2003, 5 (52), 303−304. (27) Cinčić, D.; Frišcǐ ć, T.; Jones, W. Chem. - Eur. J. 2008, 14 (2), 747−753. (28) Saha, B. K.; Nangia, A. Heteroat. Chem. 2007, 18 (2), 185−194. (29) Gelbrich, T.; Hughes, D. S.; Hursthouse, M. B.; Threlfall, T. L. CrystEngComm 2008, 10 (10), 1328−1334. 627



Article

(65) Sammon, C.; Bajwa, G.; Timmins, P.; Melia, C. D. Polymer 2006, 47 (2), 577−584. (66) Sammon, C.; Li, C.; Armes, S. P.; Lewis, A. L. Polymer 2006, 47 (17), 6123−6130. (67) (a) Stahly, G. P. Cryst. Growth Des. 2007, 7 (6), 1007−1026. (68) Brittain, H. Am. Pharm. Rev. 2002, 5, 74−80. (69) Vippagunta, S. R.; Brittain, H. G.; Grant, D. J. W. Adv. Drug Delivery Rev. 2001, 48 (1), 3−26. (70) Ford, J. L.; Timmins, P. Pharmaceutical thermal analysis: techniques and applications; Halstead Press: Sydney, Australia, 1989. (71) Spackman, M. A.; Jayatilaka, D. CrystEngComm 2009, 11 (1), 19−32. (72) McKinnon, J. J.; Jayatilaka, D.; Spackman, M. A. Chem. Commun. 2007, No. 37, 3814−3816. (73) Fabbiani, F. P. A.; Byrne, L. T.; McKinnon, J. J.; Spackman, M. A. CrystEngComm 2007, 9 (9), 728−731. (74) Bernstein, J. Cryst. Growth Des. 2011, 11 (3), 632−650. (75) Gelbrich, T.; Threlfall, T. L.; Hursthouse, M. B. CrystEngComm 2012, 14 (17), 5454−5464. (76) Gelbrich, T.; Hursthouse, M. B. CrystEngComm 2006, 8 (6), 448−460. (77) Fabian, L.; Kalman, A. Acta Crystallogr., Sect. B: Struct. Sci. 1999, 55 (6), 1099−1108. (78) Burger, A.; Ramberger, R. Microchim. Acta 1979, 72 (3−4), 259−271. (79) Burger, A.; Ramberger, R. Microchim. Acta 1979, 72 (3−4), 259−271.

628
