Impact of the counterion on the solubility and physicochemical properties of salts of carboxylic acid drugs

Aim: Salt formation is a widely used approach to improve the physicochemical and solid state properties of an active pharmaceutical ingredient. In order to better understand the relationships between the active drug, the selected counterion and the resultant salt form, crystalline salts were formed using four different carboxylic acid drugs and a closely related series of amine counterions. Thirty-six related crystalline salts were prepared, characterized and the relationship between solubility and dissolution behaviour and other properties of the salt and the counterion studied. Methods: Salts of four model acid drugs, gemfibrozil, flurbiprofen, ibuprofen and etodolac were prepared using the counterions butylamine, hexylamine, octylamine, benzylamine, cyclohexylamine, tert-butylamine, 2-amino-2-methylpropan-1-ol, 2-amino-2-methylpropan-1,3-diol andtris(hydroxymethyl)aminomethane. Salt formation was confirmed, the salts were characterized and their corresponding solubilities determined and rationalized with respect to the counterions' properties. Results and conclusion: The properties of the salt highly dependent on the nature of the counterion and, although there is considerable variation, some general conclusion can be drawn. For the alkyl amines series, increasing chain length leads to a reduction in solubility across all the acidic drugs studied and a reduction in melting point, thus contradicting simplistic relationships between solubility and melting point. Small, compact counterions consistently produce crystalline salts with high melting point accompanied with a modest improvement in solubility and the nature of hydrogen bonding between the ions has a major impact on the solubility. © 2012 Informa Healthcare USA, Inc.

Studies on the thermal decomposition behaviour, kinetics and electrical conductivity of the non-isothermal decomposition of pyridine mono carboxylic acids and some of their transition metal complexes

The thesis is divided into four chapters. They are: introduction, experimental, results and discussion about the free ligands and results and discussion about the complexes. The First Chapter, the introductory chapter, is a general introduction to the study of solid state reactions. The Second Chapter is devoted to the materials and experimental methods that have been used for carrying out tile experiments. TIle Third Chapter is concerned with the characterisations of free ligands (Picolinic acid, nicotinic acid, and isonicotinic acid) by using elemental analysis, IR spectra, X-ray diffraction, and mass spectra. Additionally, the thermal behaviour of free ligands in air has been studied by means of thermogravimetry (TG), derivative thermogravimetry (DTG), and differential scanning calorimetry (DSC) measurements. The behaviour of thermal decomposition of the three free ligands was not identical Finally, a computer program has been used for kinetic evaluation of non-isothermal differential scanning calorimetry data according to a composite and single heating rate methods in comparison with the methods due to Ozawa and Kissinger methods. The most probable reaction mechanism for the free ligands was the Avrami-Erofeev equation (A) that described the solid-state nucleation-growth mechanism. The activation parameters of the decomposition reaction for free ligands were calculated and the results of different methods of data analysis were compared and discussed. The Fourth Chapter, the final chapter, deals with the preparation of cobalt, nickel, and copper with mono-pyridine carboxylic acids in aqueous solution. The prepared complexes have been characterised by analyses, IR spectra, X-ray diffraction, magnetic moments, and electronic spectra. The stoichiometry of these compounds was ML2x(H20), (where M = metal ion, L = organic ligand and x = water molecule). The environments of cobalt, nickel, and copper nicotinates and the environments of cobalt and nickel picolinates were octahedral, whereas the environment of copper picolinate [Cu(PA)2] was tetragonal. However, the environments of cobalt, nickel, and copper isonicotinates were polymeric octahedral structures. The morphological changes that occurred throughout the decomposition were followed by SEM observation. TG, DTG, and DSC measurements have studied the thermal behaviour of the prepared complexes in air. During the degradation processes of the hydrated complexes, the crystallisation water molecules were lost in one or two steps. This was also followed by loss of organic ligands and the metal oxides remained. Comparison between the DTG temperatures of the first and second steps of the dehydration suggested that the water of crystallisation was more strongly bonded with anion in Ni(II) complexes than in the complexes of Co(II) and Cu(II). The intermediate products of decomposition were not identified. The most probable reaction mechanism for the prepared complexes was also Avrami-Erofeev equation (A) characteristic of solid-state nucleation-growth mechanism. The tempemture dependence of conductivity using direct current was determined for cobalt, nickel, Cl.nd copper isonicotinates. An activation energy (ΔΕ), the activation energy (ΔΕ ) were calculated.The ternperature and frequency dependence of conductivity, the frequency dependence of dielectric constant, and the dielectric loss for nickel isonicotinate were determined by using altemating current. The value of s paralneter and the value of'density of state [N(Ef)] were calculated. Keyword Thermal decomposition, kinetic, electrical conduclion, pyridine rnono~ carboxylic acid, cOlnplex, transition metal compJex.

3,3'-Selenobis(propionic acid)

In contrast to Se[CH2C(O)OH]2 versus S[CH2C(O)OH](2), the title compound, Se[CH(2)CH(2)C(O)OH]2 or C6H10O4Se, is structurally quite similar to its sulfur analogue. The molecule has twofold symmetry. The C-Se-C bond angle is 96.48 (8) degrees and the Se-C bond lengths are 1.9610 (14) Å. The shortest SeO intermolecular distance is 3.5410 (11) Å. The  OO distances in the carboxylic acid dimers are 2.684 (2) Å. The temperature dependence of the IR spectrum suggests tautomerism in the solid state. Formula: C6H10O4Se

Thermo-oxidative stabilization of poly(lactic acid) with antioxidant intercalated layered double hydroxides

Two antioxidant modified layered double hydroxides (AO-LDHs) were successfully prepared by theintercalation of 3-(3,5-di-tert-butyl-4-hydroxyphenyl)propionic acid (IrganoxCOOH) and 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox) in the layered structure of LDH. It was foundthat by anchoring the phenolic moieties to the LDH layers the antioxidant power is retained in the caseof Trolox, and even amplified in the case of IrganoxCOOH. A small amount of the two AO-LDHs wasincorporated into poly(lactic acid), PLA, by solution mixing and melt extrusion. The thermo-oxidativestability of the composites was compared with that of the neat PLA and PLA containing free AOs. SECanalysis indicates that, after a controlled period of ageing, both the AO-LDHs protect the PLA fromchain scission. The oxidation induction time (OIT, DSC) at 230 °C shows also the beneficial effects ofthe presence of the functional filler in the polymer matrix. Further, results from a preliminary migrationtest suggest that the AO species have a low tendency to migrate away from the AO-LDHs embedded inthe polymer matrix thus keeping the AO protected inside the nanofiller layers thereby remaining activefor a longer time.

A new application for transition metal chalcogenides:WS2 catalysed esterification of carboxylic acids

The first application of WS2, a well-known graphene analogue, as a solid acid catalyst for carboxylic acid esterification is reported. WS2 exhibits excellent specific activities and high conversion to methyl esters of (65–90%) for C2–C16 carboxylic acid esterification with methanol under mild conditions, with Turnover Frequencies between 80 and 180 h−1, and outstanding water tolerance even under equimolar water spiking. WS2 also exhibits good stability towards methyl propanoate in the continuous esterification of propanoic acid, and is a promising candidate for biofuels production.

Structural properties of a family of hydrogen-bonded co-crystals formed between gemfibrozil and hydroxy derivatives of t-butylamine, determined directly from powder X-ray diffraction data

We report the formation and structural properties of co-crystals containing gemfibrozil and hydroxy derivatives of t-butylamine H2NC(CH3)3-n(CH2OH)n, with n=0, 1, 2 and 3. In each case, a 1:1 co-crystal is formed, with transfer of a proton from the carboxylic acid group of gemfibrozil to the amino group of the t-butylamine derivative. All of the co-crystal materials prepared are polycrystalline powders, and do not contain single crystals of suitable size and/or quality for single crystal X-ray diffraction studies. Structure determination of these materials has been carried out directly from powder X-ray diffraction data, using the direct-space Genetic Algorithm technique for structure solution followed by Rietveld refinement. The structural chemistry of this series of co-crystal materials reveals well-defined structural trends within the first three members of the family (n=0, 1, 2), but significantly contrasting structural properties for the member with n=3. © 2007 Elsevier Inc. All rights reserved.

A solvent-free matrix application method for matrix-assisted laser desorption/ionization imaging of small molecules

Matrix application continues to be a critical step in sample preparation for matrix-assisted laser desorption/ionization (MALDI) mass spectrometry imaging (MSI). Imaging of small molecules such as drugs and metabolites is particularly problematic because the commonly used washing steps to remove salts are usually omitted as they may also remove the analyte, and analyte spreading is more likely with conventional wet matrix application methods. We have developed a method which uses the application of matrix as a dry, finely divided powder, here referred to as dry matrix application, for the imaging of drug compounds. This appears to offer a complementary method to wet matrix application for the MALDI-MSI of small molecules, with the alternative matrix application techniques producing different ion profiles, and allows the visualization of compounds not observed using wet matrix application methods. We demonstrate its value in imaging clozapine from rat kidney and 4-bromophenyl-1,4-diazabicyclo(3.2.2)nonane-4-carboxylic acid from rat brain. In addition, exposure of the dry matrix coated sample to a saturated moist atmosphere appears to enhance the visualization of a different set of molecules.

Functionalised polymers

Synthetic routes to polymers possessing functional groups were studied. Direct functionalisation of poly(vinyltoluene) by lithiation and carboxylation resulted in the expected carboxylic acid but reaction was complicated by the production of a mixture of products. Reaction occurred both at the polymer backbone and at the pendant methyl group. Reaction with ethyl formate was also difficult to control and a secondary alcohol was formed even when an excess of the carbonyl compounds was employed. Grignard formation of poly(bromostyrene) was successful but once formed, the derivative rearranged resulting in chain scission and degradation of the polymer. Therefore subsequent reactions of the Grignard reagent with carbonyl groups were unsuccessful in producing functionalised polymers. Reactions of vinyltoluene monomer were more successful. Although complications arose when lithiation and carboxylation of the monomer were carried out using lithium diisopropylamide because the carboxylic acid product reacted with the excess lithium diisopropylamide present, metallation by potassium t-butoxide followed by reaction with 2-(3-chloropropyl)-2-methyl-1,3-dioxalane resulted in the formation of 2-methyl-2(4-(vinylphenyl)-butyl-1,3,-dioxalane. The butyllithium initiated anionic polymerisation of this protected monomer resulted in a polymer which had a very narrow molecular weight distribution (Mw/Mn= 1.05) and subsequent hydrolysis of the polymer resulted in poly(6(vinylphenyl)-hexan-2-one) which was derivatised with 2,4 dinitrophenyl-hydrazine. Functionalisation by modification of the siloxane derivative 3-(methylpropenoxycarbonyl)ltrimethoxysilane was unsuccessful. The acid catalysed exchange reactions of this monomer with alcohols such as eugenol, octan-1-ol, pentan-1-ol, and hexan-1-ol were inefficient, resulting in a mixture of products and unreacted starting materials.

Improving the solubility and dissolution of poorly soluble drugs by salt formation and the consequent effect on mechanical properties

The bioavailability of BCS II compounds may be improved by an enhanced solubility and dissolution rate. Four carboxylic acid drugs were selected, which were flurbiprofen, etodolac, ibuprofen and gemfibrozil. The drugs were chosen because they are weak acids with poor aqueous solubility and should readily form salts. The counterions used for salt formation were: butylamine, pentylamine, hexylamine, octylamine, benzylamine, cyclohexylamine, tert-butylamine, 2-amino-2-methylpropan­2-ol, 2-amino-2-methyl propan-1,3-ol and tromethamine. Solubility was partially controlled by the saturated solution pH with the butylamine counterion increasing the solution pH and solubility and dissolution to the greatest extent. As the chain length increased, solubility was reduced due to the increasing lipophilic nature of the counterion. The benzylamine and cyclohexylamine counterions produced crystalline, stable salts but did not improve solubility and dissolution significantly compared to the parent compound. The substitution of hydroxyl groups to tert-butylamine counterions produced an increase in solubility and dissolution. AMP2 resulted in the most enhanced solubility and dissolution compared to the parent drug but using the tris salt did not further improve solubility due to a very stable crystal lattice structure. The parent drugs were very difficult to compress due to orientation effects and lamination. Compacts were prepared of each parent drug and salt and their modulus of elasticity values were measured using a three-point bend (Young’s modulus, E0) were extrapolated to zero porosity and compared. Compressibility and E0 were improved with the butylamine, tert-butylamine, cyclohexylamine and AMP2 counterions. The most significant improvement in compression and E0 was with the AMP2 salts. Mechanical properties were related to the hydrogen bonding within the crystal lattice structure for the gemfibrozil salt series.

The chemistry of some organotellurium compounds

Tbe formation of Pd(TeR)n and (CuTeR)n from the reaction between telluroesters and Pd(II)or Cu(II) suggested that these organa­tellurium reagents may be useful precursors of RTe- ligands in reactions with transition-metal substrates. Also the formation of telluronium salts Me2RTeI- from the reaction between telluroesters and methyl iodide, together with the above, confirm the cleavage of -cõ-Te bonds rather than -C-Te bonds. The formation of a carboxylic acid from the toluene solution of a ditelluride d palladium(O) complex in the presence of light oxygen (from air) is demonstrated. When the solvent employed is p-xylene an aldehyde is formed.The reaction proceeds via the free radical, RTeO, with Pd(PPh3)4 as a catalyst.It has also been shown that the oxidation of aldehydes to carboxylic acids is catalysed by ditelluride. Spin trapping experiments with PhCH=N(O)But (phenyl-t-butyl-nitrone) have provided evidence that the oxidative addition of an alkyl halide (RX=Mei, BunBr, BusecBr, ButBr, BrCH2-CH=CHCH2Br, and Br(CH2)4Br) to diphenyltelluride and reductive elimination of CH3SCN from Ph2(CH3)Te(NCS) proceeds via radical pathways. A mechanism is proposed for oxidative addition which involves the preformation of a charge transfer complex of alkyl halide and diphenyltelluride.The first step is the formation of a charge transfer complex, and the initial product of the oxidative addition is a "covalent" form of the tellurium(IV)compound. When the radical R is more stable, Ph2TeX2 may be the major tellurium(IV)product. The reaction of RTeNa (R=p-EtOC6H4, Ph) with organic dihalides X2(CH2)n (n=1,2,3,4) affords telluronium salts (n=3,4; X=Cl, Br) the nature of which is discussed.For n=l (X=Br, I)the products are formulated as charge transfer complexes of stoichiometry (RTe)2(CH2).CH2X2• For n=2, elimination of ditelluride occurs with the formation of an alkene. Some 125’Te Mõssbauer data are discussed and it is suggested that the unusually low value of 6 (7.58 mm.s-1 ) for  p-EtO.C6H4.Te)2(cH2)cH2Br2 relates to removal of 5's electronsfrom the spare pair orbltal via the charge transfer interaction. 125Te Mossbauer data for (p-EtO.C6H4)Te(CH2)4Br are typical of a tellurium (IV) compound and in particular ∇ is in the expected range for a telluronium salt. The product of the reaction of Na Te (C6H4.OEt), with 1,3-dibromopropane is, from the Mössbauer data, also a telluronium salt.

Heat shock proteins in leukaemia cell differentiation and cell death

When HL60 cells were induced to differentiate to granulocyte-like cells with the agents N-methylformamide and tunicamycin an concentrations marginally below those which were cytotoxic, there was a decrease in the synthesis of the glucose- regulated proteins which preceded the expression of markers of a differentiated phenotype. There was a transient increase in the amount of hsp70 after 36 hours in NMF treated cells but in differentiated cells negligible amounts were detected. Inducers which were known to modulate hsp70 such as azetadine carboxylic acid did not induce differentiation suggesting early changes in the endoplasmic reticulum may be involved in the commitment to terminal differentiation of HL60 cells. These changes in group synthesis were not observed when K562 human chronic myelogenous leukemia cells were induced to differentiate to erythroid-like cells but there was a comparable increase in amounts of hsp70. When cells were treated with concentrations of drugs which brought about a loss in cell viability there was an early increase in the amount of hsp70 protein in the absence of any increase in synthesis. HL60 cells were treated with NMF (225mM), Adriamycin (1μM), or CB3717 (5μM) and there was an increase in the amounts of hsp70, in the absence of any new synthesis, which preceded any loss of membrane integrity and any significant changes in cell cycle but was concomitant with a later loss in viability of > 50% and a loss in proliferative potential. The amounts of hsp70 in the cell after treatment with any of the drugs was comparable to that obtained after a heat shock. Following a heat shock hsp70 was translocated from the cytoplasm to the nucleus, but treatment with toxic concentrations of drug caused hsp70 to remain localised in the cytoplasm. Changes in hsp70 turn-over was observed after a heat shock compared to NMF-treated cells. Morphological studies suggested that cells that had been treated with NMF and CB3717 were undergoing necrosis whereas the Adriamycin cells showed characteristics that were indicative of apoptosis. The data supports the hypothesis that an increase in amounts of hsp70 is an early marker of cell death.

Studies of the stability and metabolism of tumour inhibitory tetrazinones

Temozolomide is an imidazotetrazinone with antineoplastic properties. It is structurally related to dacarbazine. Temozolomide was not metabolized in vitro by liver fractions. Chemical decomposition appears to play an important r^ole in its in vitro and in vivo disposition. In contrast, 3-methylbenzotriazinone, a structural analogue, was metabolized by hepatic microsomes to afford benzotriazinone and a hydrophilic metabolite. The cytotoxicity of temozolomide, dacarbazine, 5-[3-(hydroxy-methyl-3-methyl-triazen-1-yl]imidazole-5-carboxamide (HMMTIC) and 3-monomethyl-(triazen-1-yl)imidazole-4-carboxamide (MTIC) were investigated in TLX5 murine lymphoma cells. Unlike dacarbazine, which was not toxic, MTIC, HMMTIC and temozolomide were cytotoxic in the absence of microsomes. Decarbazine was only cytotoxic in the presence of microsomes. The formation of MTIC from dacarbazine, HMMTIC and temozolomide was determined by reversed phase high performance liquid chromatography in mixtures incubated under conditions identical to those described before. MTIC was generated chemically from temozolomide and HMMTIC metabolically from dacarbazine. Using [14C]temozolomide, it was found that, in mice, the major route of excretion of the drug is via the kidneys. An acidic metabolite (metabolite I) was found in the urine of mice which had received temozolomide but its identity has not been established. 1H NMR, UV and chemical analyses revealed that Metabolite I possesses an intact NNN linkage and the site of metabolism is at the N3 methyl group. A further acidic metabolite (metabolite II) was found in the urine of patients. Metabolite II was unambiguously identified as the 8-carboxylic acid derivative of temozolomide. In vitro cytotoxicity assay showed that ony metabolite II is cytotoxic but not metabolite I. Pharmacokinetic studies of temozolomide and MTIC in vivo were performed on mice bearing TLX5 tumour. Temozolomide was eliminated from the plasma monophasically with a t1/2 of 0.7hr. MTIC was identified as a product of decomposition. MTIC was eliminated rapidly with a t1/2 of 2min. Though temozolomide shares many biochemical and biological similarities with clinically used dacarbazine, the results obtained in this study show that it differs markedly in its pharmacokinetic properties from dacarbazine, as temozolomide produced relatively sustained plasma levels which were reflected by drug concentrations in the tumour.

The importance of Na(plus)-K(plus)-Cl- cotransport in nitrogen mustard induced cell death

The incubation of murine leukaemic L1210 cells in vitro for 4 hours (hr) with 10uM nitrogen mustard (HN2), a bifunctional alkylating agent, inhibited the influx of the potassium congener, 88rubidium+ ( 86Rb+) by the selective inhibition of the Na+-K+-CI- cotransporter. The aim of this project was to investigate the importance of this lesion in HN2-induced cytotoxicity. 86Rb+ uptake in human erythrocytes was inhibited by high concentrations of HN2 (2mM) and occurred in two phases.In the first hour both the Na+/K+ ATPase pump and the Na+-K+-CI- cotransporter were equally inhibited but after 2 hrs exposure to 2mM HN2, the Na+ -K+ -CI- cotransporter was significantly more inhibited than the Na+/K+ ATPase pump. In contrast, both potassium transport systems were equally inhibited in L1210 cells incubated for 10 minutes with 1mM HN2. The selective inhibition of the Na+-K+-CI- cotransporter, after a 3 hrs exposure to 10uM HN2, was not absolved by coincubation with 5ug/ml cycloheximide (CHX), an inhibitor of protein synthesis. Incubation of L1210 cells with concentrations of diuretics which completely inhibited Na+-K+-CI- cotransport did not enhance the cytotoxicity of either HN2 or its monofunctional analogue 2-chloroethyldimethylamine (Me-HN1). The incubation of L1210 cells with a twice strength Rosewell Park Memorial Institute 1640 media did not enhance the toxicity of HN2. An L1210 cell line (L1210FR) was prepared which was able to grow in toxic concentrations of furosemide and exhibited a similiar sensitivity to HN2 as parental L1210 cells. Treatment of L1210 cells with 10uM HN2 resulted in a decrease in cell volume which was concurrent with the inhibition of the Na+-K+-CI- cotransporter. This was not observed in L1210 cells treated with either 1 or O.SuM HN2. Thus, possible differences in the cell death, in terms of necrosis and apoptosis, induced by the different concentrations of HN2 was investigated. The cell cycle of L1210 cells appeared to be blocked non-specifically by 10uM HN2 and in S and G2/M by either 1 or 0.5uM HN2. There were no significant changes in the cytosolic calcium concentrations of L1210 cells for up to 48 hrs after exposure to the three concentrations of HN2. No protection against th_ toxic effects of HN2 was observed in L1210 cells incubated with 5ug/ml CHX for up to 6 hrs. Incubation for 12 or 18 hrs with a non-toxic concentration (5mM) of L-Azetidine-2- carboxylic acid (ACA) enhanced the toxicity of low concentrations (<0.5uM) of HN2.

Antitumour imidazotetrazines: probes for the major groove of DNA

The antitumour imidazotetrazinones are believed to act as prodrugs for the triazene series of alkylating agents, showing a marked pteference for the alkylation of the middle guanine residue in a run of three or more contiguous guanines. However, the. exact nature of the interactions of imidazotetrazinones within the micro~environment of DNA are; as yet unknown. In order to examine such interactions a three pronged approach involving molecular modelling, synthetic chemistry and biological analysis has been undertaken during the course of this project. . Molecular modelling studies have shown that for the 8-carboxamido substituted imidazotetrazinones antitumour activity is dependent upon the. presence of a free NH group which can be involved in the formation of both intramolecular and intermolecular hydrogen bonds, and the presence of a non-bulky substituent with a small negative potential . volume. Modelling studies involving the docking of .mitozolomide into the major groove of DNA in the region of a triguanine sequence has shown that a number of hydrogen bonding interactions are feasible. A series of 8-substituted carboxamide derivatives of mitozolomide have been synthesised via the 8-acid chloride and 8-carboxylic acid derivatives including a number of peptide analogues. The peptide derivatives were based upon the key structural features of the helix-turn-helix motif of DNA-binding proteins with a view to developing agents that are capable of binding to DNA with greater selectivity. An examination of the importance of intramolecular hydrogen bonding in influencing the antitumour activity:of :the imidazotetrazinones has led to the synthesis of the novel pyrimido[4',5' :4,3]pyrazolo[5,1-d]-1,2,3,5-tetrazine ring system. In general, in vitro cytotoxicity assays showed that the new derivatives were less active against the TLX5 lymphoma cell line. than the parent compound mitozolomide despite an increased potential for hydrogen bonding interactions. Due to the high reactivity of the: tetrazinone ring system it is difficult to study the interactions between the imidazotetrazinones and DNA. Consequently a number of structural analogues that are stable under physiological conditions have been. prepared based upon the 1,2,3 triazin-4(3H)-one ring system fused with both benzene and pyrazole rings. Although the 3-methylbenzotriazinones failed to antagonise the cytotoxic activity of temozolomide encouraging results with a 3-methylpyrazolotriazinone may suggest the existence of an imidazotetrazinone receptor site within DNA. The potential of guanine rich sequences to promote the alkylating selectivity of imidazotetrazinones by acting as a catalyst for ring cleavage and thereby generation of the alkylating agent was examined. Experiments involving the monitoring: of the rate of breakdown of mitozolomide incubated in the presence of synthetic oIigonucleotides did not reveal any catalytic effect resulting from the DNA. However, it was noted that the breakdown of mitozolomide was dependent upon the type of buffer used in the incubations and this may indeed mask any catalysis by the oligonucleotides.

Molecular architecture influences on material properties of pharmaceutical compounds

Salt formation has extensively been studied as a strategy to improve drug solubility but it has not been explored as a strategy to improve mechanical properties. A better understanding of which factors of the solid state can have an influence in the mechanical properties of pharmaceutical powders can help to optimise and reduce cost of tablet manufacturing. The aim of this study was to form different series of amine salts of flurbiprofen, gemfibrozil and diclofenac and to establish predictive relationships between architectural characteristics and physicochemical and mechanical properties of the salts. For this purpose, three different carboxylic acid drugs were selected: flurbiprofen, gemfibrozil and diclofenac, similar in size but varying in flexibility and shape and three different series of counterions were also chosen: one with increasing bulk and no hydroxyl groups to limit the hydrogen bonding potential; a second one with increasing number of hydroxyl groups and finally a third series, related to the latter in number of hydroxyl groups but with different molecular shape and flexibility. Physico-chemical characterization was performed (DSC, TGA, solubility, intrinsic dissolution rate, particle size, true density) and mechanical properties measured using a compaction replicator. Strained molecular conformations produce weaker compacts as they have higher energy than preferred conformations that usually lie close to energy minimums and oppose plastic deformation. It was observed that slip planes, which correspond to regions of weakest interaction between the planes, were associated with improved plasticity and stronger compacts. Apart from hydrogen bonds, profuse van der Waals forces can result in ineffective slip planes. Salts displaying two-dimensional densely hydrogen bonded layers produced stronger compacts than salts showing one-dimensional networks of non-bonded columns, probably by reducing the attachment energy between layers. When hydrogen bonds are created intramolecularly, it is possible that the mechanical properties are compromised as they do not contribute so much to create twodimensional densely bonded layers and they can force molecules into strained conformations. Some types of hydrogen bonding network may be associated with improved mechanical properties, such as type II, or R (10) 3 4 using graph-set notation, versus type III, or R (12) 4 8 , columns. This work clearly demonstrates the potential of investigating crystal structure-mechanical property relationship in pharmaceutical materials.