British Journal of Radiology 74 (2001),1097-1102 © 2001 The British Institute of Radiology
Stabilization of the hydrophilic sphere of non-ionic monomers: are all protected in a similar way?
D Violon, MD
Acht Eeuwenlaan 73, 2650 Edegem, Belgium
 |
Abstract
|
|---|
The present study attempts to investigate whether incorporation of a methyl group as second substituent in the tertiary amido group of the two benzamide side chains of iobitridol (Xenetix®) increases the stability of the hydrophilic sphere around this molecule as claimed by its manufacturer, and whether this hydrophilic sphere is unique to this particular molecule or to what extent other monomer non-ionic contrast media show this feature. Five non-ionic monomer contrast medium molecules, ioversol, iohexol, iobitridol, ioxilan and iopromide, were studied. Barriers to the rotation of acetanilide and benzamide side chains, and to the rotation of the amide bond in the benzamide chains, were calculated at a semi-empirical quantum mechanical level of theory with the Mopac/Ampac computer program. The five studied contrast medium molecules showed very similar energy barriers to the rotation of complete substituent chains (benzamide and acetanilide) around their bond with the benzene ring. The magnitude of the barriers fell in the range of values mentioned in the literature. In conclusion, introduction of a methyl group as second substituent in the tertiary amido group of the benzamide substituent chains of iobitridol does not increase the studied rotation barriers, i.e. it does not seem to stabilize the hydrophilic sphere to a greater extent compared with similar monomer non-ionic molecules. The sphere is shown to exist in all five analysed molecules. Introduction of a group with a more hydrophobic character has to be considered and eventually questioned, bearing in mind that a more hydrophilic molecule interacts to a lesser degree with body systems.
|
Introduction
|
|---|
Iodinated contrast media have been intensively used for many years in various radiological examinations and, over time, their safety has increased. Contrast media used to be ionic but the newer generations of contrast media are non-ionic. Water solubility of ionic contrast media is achieved by salt formation. The introduction of a cation for each iodine-containing anion doubles the number of particles in solution and increases the osmolality. Non-ionic contrast media do not depend on salt formation; hydrophilic groups, mostly hydroxyalkyl groups, incorporated around the benzene ring achieve water solubility. Toxicity of non-ionic contrast media is lower than that of ionic contrast media because they possess no charge, and consequently there is no interaction with body electrolytes and neurophysiological mechanisms. The number of particles in a solution determines its osmolality. The observed osmolality of non-ionic monomers is lower than theoretically calculated owing to association of molecules, mainly by hydrophobic interaction; osmolality appears to be inversely proportional to hydrophilicity. Viscosity of a solution depends on molecular structure, with smaller molecules showing the lowest viscosity. Increasing molecular volume by lengthening the moieties that provide hydrophilicity leads to increased viscosity. It is obvious that osmolality, viscosity, hydrophilicity and solubility are interrelated and cannot all be optimized together. Nevertheless, adverse reactions are still feared, even with the most recent products [1]. Among these, anaphylactoid reactions are unpredictable and potentially lethal. Many hypotheses have been linked to the possible mechanisms of these adverse reactions. In spite of the absence of a firmly established mechanism, some facts are beginning to emerge. One fact is that, all other parameters being equal, the more hydrophilic a molecule, the better it will be tolerated in clinical use [2]. Contrast media interact with biological sites on a hydrophobic basis [3–5]. It is well known that contrast media should possess a high hydrophilic nature for minimal chemotoxicity [2], and should therefore show the lowest possible hydrophobic characteristics to reduce interactions with body systems as much as possible.
A non-ionic monomeric contrast medium (Figure 1
) consists of a triiodobenzene ring with three substituent side chains bearing hydrophilic hydroxyl groups. Hydrophilic protection of the hydrophobic portion of the contrast medium molecule (essentially the triiodobenzene ring) was formalized in the facial/lateral protection model [6]. The facial/lateral protection model requires the hydrophilic components of the three substituent chains to be evenly distributed around the hydrophobic parts of the molecule to shield them from interaction with important human body molecules such as proteins, lipoproteins, enzymes and membranes. This facial/lateral protection model is highly correlated with the octanol/water partition coefficient (LogP) [7, 8], an experimental measure of the hydrophilicity of a molecule, or in other words, a measure of the relative absence of hydrophobic characteristics.
View larger version (12K):
[in this window]
[in a new window]
|
Figure 1. Model of triiodobenzene contrast media. Rn (n=1–6) represents any possible substituent. The six carbon atoms of the benzene ring are numbered.
|
|
It has been suggested that hydrophilic protection of the molecule's hydrophobic zones has to be permanent, i.e. a hydrophilic sphere that will not change its configuration under the influence of a hydrophobic environment should surround the molecule to prevent exposure of the triiodobenzene ring by hydrophobic forces [9]. Consequently, iobitridol (Xenetix®; Laboratoire Guerbet, Roissy Charles de Gaulle, France) (Figure 2) was designed to contain a tertiary amido group (disubstituted) in each of the two benzamide side chains—a methyl group was chosen as supplemental N-substituent—to decrease their flexibility [9]. The presence of the methyl group is intended to increase the barrier to rotation of the substituent chain around its bond with the benzene ring by causing a supposedly important steric hindrance. This concept was demonstrated with the help of small models of contrast medium molecules [9].
View larger version (9K):
[in this window]
[in a new window]
|
Figure 2. Formula of iobitridol. The methyl groups included in the teriary amido groups in the two benzamide side chains are indicated (arrows).
|
|
The purpose of this communication is to study the barriers to the rotation of substituent chains, this time not in a small model but in five complete monomer non-ionic contrast media molecules in clinical use, namely ioversol, iohexol, iobitridol, ioxilan and iopromide. These results will be compared with assumptions [9] regarding the so-called stabilized hydrophilic sphere of iobitridol to explore whether the situation of iobitridol is unique in this matter. Incorporation of two methyl groups, being of a hydrophobic nature, influences the hydrophilicity of the global molecule. Should the barriers to rotation of the substituent chains not increase significantly in the presence of these methyl groups, incorporation of two hydrophobic methyl groups could at least be questioned.
|
Materials and methods
|
|---|
The structural formulae of the five studied molecules, ioversol, iohexol, iobitridol (Figure 2
), ioxilan and iopromide, are available in the literature [9–12]. They are commercially available as Optiray® (Mallinckrodt Medical Inc., St Louis, MO), Omnipaque® (Nycomed Inc., Oslo, Norway), Xenetix®, Oxilan® (Cook Inc., Bloomington, IN) and Ultravist® (Schering, Berlin, Germany), respectively. Initially, two-dimensional structural models of the molecules were drawn with a molecule drawing and visualization program (ALCHEMY III; Tripos Ass. Inc., St Louis, MO). The ALCHEMY III output files were converted, with a file type converting program (BABEL Version 1.06; [13]), to MOPAC (Molecular Orbital PACkage) [14] input files. This detour was necessary because Alchemy III output files cannot be read immediately as MOPAC input files.
MOPAC is a semi-empirical energy program that can optimize the structure of molecules, i.e. find the minimal energy configuration (the most stable configuration possessing the lowest energy being the most probable form in the given experimental conditions), find transition structures and calculate harmonic vibrational frequencies [15]. Semi-empirical programs (including MOPAC) use a mixture of theoretical models and empirical data to build the equation (containing the Hamiltonian operator) defining the studied system.
The three substituent chains on the triiodobenzene ring consist basically of one acetanilide group (generally called position 5) and two occasionally but not always identical benzamide groups (position 1 and 3) (Figure 1).
As far as the chiral carbons are concerned, the configurations of the five analysed molecules were: ioversol, S-form in substituent groups in positions 1 and 3; iohexol, S-form in substituent groups in positions 1, 3 and 5; iobitridol, S-form for both chiral carbons in the two identical side chains; ioxilan, S-form in the chiral carbon atom in side chain in position 1 and 5; iopromide, R-form in substituent group 1 and S-form in substituent group 3.
The energy minimizations in the first part of the experiment (rotation of the bond between substituent chain and benzene ring) were carried out with the MOPAC program available for PC use with MSDOS. This program uses dedicated keywords to perform a precise task. The MOPAC keywords used in this case were "MNDO" and "MMOK", which signifies that the MNDO (modified neglect diatomic overlap) [16] Hamiltonian operator in the system-describing equation was used. The molecular mechanics correction term MMOK for amide bonds was included. When the system contains a peptide linkage, this keyword allows application of a molecular mechanics correction term to increase the barrier to rotation in order to induce a more correct simulation of reality. In fact this molecular mechanics term ameliorates the semi-empirical description fitting in with experimental data.
All calculations were executed on a Pentium II 450 MHz computer. The dihedral angle of the acetanilide group with the benzene ring was increased manually in steps of 10°, so that 37 input files were created. The benzamide substituent chain was then allowed to rotate (automatically driven) with steps of 10°, by setting the flag to the appropriate dihedral in the 37 MOPAC input files. In the case of non-identical benzamide side chains, as in iopromide, it was the chain containing the tertiary amido group that was permitted to rotate and, as in ioxilan, it was the aminopropanediol chain that was allowed to rotate. With this method the 37 files ran at convenience and used CPU time when available. This replaced the implemented STEP 1 and STEP 2 procedure and yielded 37 x 37 (total 1369) configurations. Indeed, the implemented method with STEP 1 and STEP 2 runs the entire computation at once in one time-consuming job.
Each of the 1369 configurations was minimized as described and its heat of formation (kcal mol-1) was tabulated, resulting in a 37 x 37 array, representing the heats of formation in function of the two torsion angles varying in steps of 10°. This procedure was repeated for the five studied molecules.
The next part of the experiment concerned only ioversol, iobitridol and iopromide. The lowest energy configuration of the acetanilide chain was chosen for each molecule on the basis of the grid generated as previously described (Figure 3). In this part of the experiment, the amide bond in the benzamide substituent chains was allowed to rotate over 360° in steps of 10°. These energy minimizations were performed with a commercial semi-empirical program analogous to MOPAC (AMPAC; Semichem Inc., Shawnee, KS). The same keywords were used as for the MOPAC calculations. AMPAC was used as it was felt that in local experimental conditions AMPAC runs faster than MOPAC (personal communication, Dr A Holder, Semichem Inc., KS).
View larger version (20K):
[in this window]
[in a new window]
|
Figure 3. For the five studied molecules, each box contains bold figures corresponding to the heat of formation of the four calculated minima conformers. Between minima, hilltops are given (italics). Outside the boxes, the differences or rotation barriers can be found between the minima and hilltops. The benzamide rotation is given in the columns, and the acetanilide rotation results in the rows. All included figures are expressed in kcal mol-1.
|
|
|
Results
|
|---|
The first part of the experiment with the combined rotation of the acetanilide chain and one benzamide chain around their bond with the benzene ring resulted in 1369 values of the heat of formation. Heat of formation is plotted against the two torsion angle (Figure 4). This energy surface yields four minimum energy conformers (heats of formation values), together with a saddle point situated between each pair of minima. This process was repeated for all the studied molecules. The four other molecule computations resulted in surface forms with four minima very similar to the energy surface of iobitridol in Figure 4. The heats of formation of these four minima per molecule are given in Figure 3, together with the values of the hilltops. The energy differences between minima and hilltops are also indicated, representing the respective barriers. In Figure 3, values for the benzamide rotations can be found in the columns, and values regarding the acetanilide rotations in the rows.
View larger version (61K):
[in this window]
[in a new window]
|
Figure 4. Iobitridol: energy surface plot (heat of formation values kcal mol-1) of the rotation of the benzamide chain plotted against the the rotation of the acetanilide chain. Four valleys (four minimal energy conformations) are easily distinguished. Between two minima, a hilltop or saddle point can be located.
|
|
Results of the calculations regarding the rotation of the amide linkage in the benzamide chains can be found in Table 1 (ioversol, iobitridol and iopromide). In Table 2, the LogP values of the studied molecules are tabulated, as retrieved from the literature [6, 17].
View this table:
[in this window]
[in a new window]
|
Table 1. The calculated heat of formation (HOF) of the minimum energy conformers for ioversol, iobitridol and iopromide obtained when the the amide link in the benzamide chain of the respective molecules is permitted to rotate over 360°. The hilltop between each pair of minima together with the energy differences or energy barriers to rotation are provided. All figures are expressed in kcal mol-1
|
|
|
Discussion
|
|---|
Monomeric non-ionic contrast media contain a central hydrophobic part, the benzene ring and three iodine atoms in positions 2, 4 and 6 (Figure 1), which must be shielded from the environment to avoid interaction with body systems. This is achieved by strategically placed peripheral hydrophilic groups, mostly hydroxyl groups, evenly distributed around the molecule. It is argued that the three substituent chains with hydrophilic elements could be deformed or rotate to a less protective position when the molecule is placed in a hydrophobic environment, permitting the contrast medium to interact with the environment that was to be avoided. The idea behind the design of iobitridol [9] was that incorporation of a methyl group in the tertiary amido group in both benzamide chains, close to the benzene ring (Figure 2), would make this substituent side chain less prone to rotation and thus less deformable in a hydrophobic environment, essentially by steric hindrance, i.e. by increasing the barrier to the rotation of the substituent chain. In this way the protective effect would be assured even in "bad" conditions. In other words, a "stabilized hydrophilic sphere" would be created surrounding the hydrophobic parts of the molecule [9]. The experiment described here calculates the barrier (kcal mol-1) to the rotation of the acetanilide and one benzamide side chain around the bond with the benzene ring in five non-ionic monomer contrast medium molecules, including iobitridol with its benzamide side chain containing a tertiary amido group with a methyl substituent.
These molecules were represented by one configuration with complete substituent chains, instead of a small model. Indeed, the part of the study proposing the concept of a stabilized hydrophilic sphere surrounding iobitridol [9] concerned rotation barriers of small models of contrast medium molecules, where the substituents Rn in Figure 1 corresponded either to a methyl group or a hydrogen atom. The energy of all the designed molecules was minimized with a semi-empirical computer program (MOPAC/AMPAC). Minimum energy arrangements of the system correspond to stable states and are the most likely to exist. The MNDO Hamiltonian operator, which is a method of describing the studied system was chosen. This choice was supported by other work [18–22]. One isomer/molecule was studied for reasons of processor time availability and because of the known phenomenon of combinatory explosion.
Unsurprisingly, two minima are computed for each of the two substituent chains. For steric reasons, a minimum is found when the chain extends above the benzene ring and the two neighbouring iodine atoms, and a second one is observed for the same chain extending beneath the ring and two iodine atoms. Between them, a hilltop or barrier has to be overcome. The combination of these two minima with the two minima of the other substituent chain yields four global minima and four hilltops or saddle points in between, permitting eight rotation barriers to be calculated for the investigated molecule (Figure 3). The existence of the four minima (four valleys) can readily be seen in the energy surface plot (Figure 4) of the corresponding values of iobitridol. Analysis of the other four molecules gave very similar types of energy surfaces with four minima.
In the aforementioned study [9], rotation of the benzamide bond with the benzene ring showed a barrier ranging from 12.9 kcal mol-1 to at least 20.0 kcal mol-1, and for the acetanilide bond with the benzene ring a range from 11.1 kcal mol-1 to 20.0 kcal mol-1 was calculated. The results of the present study correspond with the magnitude of the barriers of the smaller experimental models but are mostly slightly longer, which is not surprising given the experimental conditions. Furthermore, in the present study, barrier values were highly comparable for the five observed molecules in any direction of rotation, whether the amido group was secondary or tertiary with or without a methyl group as second substituent. Thus, the intentionally incorporated hydrophobic methyl group did not appear to increase the rotation barrier in the particular case of iobitridol.
Barriers to the rotation of the bond in the amide link in the benzamide chains of ioversol, iobitridol and iopromide are also very similar. However, incorporation of the methyl group appears to increase the rotation barrier in the amide link by 2–3 kcal mol-1. In the complete molecules this barrier is smaller (±50%) than the barrier reported for the model (14.2–15.1 kcal mol-1) [9].
The so-called hydrophilic sphere surrounding iobitridol is shown to exist not only in iobitridol, as previously suggested [9], but also to the same extent in the other four studied molecules and with levels of rotation barriers in the same range. This is achieved in ioversol, iohexol and ioxilan without the incorporation of a hydrophobic methyl group. Moreover, incorporation of a methyl group as a second substituent in the tertiary amido group of the benzamide chain does not perceptibly increase the rotation barriers in the present experiment when complete contrast medium molecules are studied and when the entire benzamide chain is considered. The barrier to rotation of the amide linkage in the same chain is only very slightly increased (in the order of 2–3 kcal mol-1), as can be seen in the second part of the experiment with ioversol, iobitridol and iopromide. This type of barrier is even slightly higher in iopromide than in iobitridol (1 kcal mol-1). This very small difference is considered to be of no importance.
Consequently, it is concluded from this experiment that the introduction of a methyl group with a hydrophobic character as a second substituent in the tertiary amido group in the two benzamide chains of iobitridol should be carefully reconsidered and eventually questioned when this is done with the aim of increasing the stability of the so-called hydrophilic sphere. Indeed, iobitridol, with two hydrophobic methyl groups and a secondary nitrogen in position 5 (protecting the hydrophobic zones to a lesser degree than a tertiary nitrogen in this position) shows a LogP higher than the LogP of ioversol and iohexol (Table 2). The smaller (more negative) the value of the LogP coefficient, the more hydrophilic the compound will be. LogP is a good laboratory measure of the hydrophilicity of a molecule. This conclusion is reached with the experimental restriction that only one isomer per molecule was studied, although when the chosen isomers are considered as a group, their distribution of hydrophilic substituents shows a high correlation with their respective LogP [7, 8] so that this group can be considered representative of the hydrophilic characteristics. In addition to the hydrophilicity problem, incorporation of the methyl group in the two benzamide chains can also be considered to cause a safety hazard. In the amide series of substituents, monosubstituted amides are generally safer than disubstituted amides [23].
In summary, this communication enables the radiologist, when choosing a contrast medium from the commercially available range, to consider important aspects concerning hydrophilicity and safety.
Received for publication May 25, 2000.
Revision received August 2, 2001.
Accepted for publication August 13, 2001.
|
References
|
|---|
-
Katayama H, Yamaguchi K, Kozuka T, Takashima T, Seez P, Matuura K. Adverse reactions to ionic and nonionic contrast media. A report from the Japanese Committee on the Safety of Contrast Media. Radiology 1990;175:621–8.[Abstract/Free Full Text]
-
Sovak M, Ranganathan R, Lang JH, Lasser EC. Concepts in design of improved intravascular contrast agents. Ann Radiol 1978;21:283–9.
-
Prange T, Gaudey F, Ohanessian J, Avenel D, Neuman A, Corot C, et al. Angiography contrast agents interact with serine proteinases. The molecular structure of the model system elastase/iohexol. FEBS Lett 1995;357:247–50.[Medline]
-
Prange T, Neumann A, Corot C, Meyer D. Study of the complex between the contrast agent iobitridol (Xenetix®) and elastase (PPE): a model for hydrophobic site protection in drug-protein interactions. Pharm Res 1997;14:1713–7.[Medline]
-
Lang J, Lasser EC. Inhibition of adenosine triphosphatase and carbonic anhydrase by contrast media. Invest Radiol 1975;10:314–6.[Medline]
-
Bonnemain B, Meyer D, Schaefer M, Dugast-Zrihen M, Legreneur S, Doucet D. New iodinated, low-osmolar contrast media. A revised concept of hydrophilicity. Invest Radiol 1990;25(Suppl. 1):S104–6.
-
Violon D. Correlation of some geometrical parameters, derived from molecular models of monomer nonionic contrast medium molecules, with their octanol/water partition coefficient. Invest Radiol 1996;31:569–76.[Medline]
-
Violon D. Application of the lateral/facial protection model to a new molecule, iobitridol. Invest Radiol 1996;31:668–70.[Medline]
-
Meyer D, Fouchet M-H, Petta M, Carrupt P-A, Gaillard P, Testa B. Stabilization of the hydrophilic sphere of iobitridol, an iodinated contrast agent, as revealed by experimental and computational investigations. Pharm Res 1995;12:1583–91.[Medline]
-
Wilkins RA, Whittington JR, Brigden GS, Lahiri A, Heber ME, Hughes LO. Safety and pharmacokinetics of ioversol in healthy volunteers. Invest Radiol 1989;24:781–8.[Medline]
-
Foster S, Sovak M. Isomerism in iohexol and ioxilan. Analysis and implications. Invest Radiol 1988;23(Suppl. 1):S106–9.
-
Muetzel W, Speck U. Pharmacochemical profile of iopromide. Am J Neuroradiol 1983;4:350–2.[Abstract]
-
Walters P, Stahl M. BABEL Version 1.06. Dolata Research Group, Department of Chemistry, University of Arizona, Tucson, AZ.
-
QCPE (Quantum Chemical Exchange Program) MOPAC6/PC. QCMP number 137. Department of Chemistry, Indiana University, 800 East Kirkwood Avenue, Bloomington, IN 47405–7102.
-
Luke BT. http://members.aol.com/btluke/mopac7.htm
-
Dewar MJS, Thiel W. Ground states of molecules. 38. The MNDO method. Approximations and parameters. J Am Chem Soc 1977;99:4899–907.
-
Beaufils H, Idee JM, Berthommier BS, Balut C, et al. Iobitridol, a new nonionic low-osmolality contrast agent, and Iohexol. Impact on renal histology in the rat. Invest Radiol 1995;30:33–9.[Medline]
-
Besler BH, Merz KM, Kollman PA. Atomic charges derived from semiempirical methods. J Comp Chem 1990;11:431–9.
-
Orozco M, Luque FJ. On the use of AM1 and MNDO wave functions to compute accurate electrostatic charges. J Comp Chem 1990;11:909–23.
-
Merz KM. Analysis of a marge data base of electrostatic potential derived atomic charges. J Comp Chem 1992;13:749–67.
-
Bonati L, Cosentino U, Fraschini E, Moro G, Pitea D. Molecular electrostatic potential of substituted aromatic compounds: factors affecting the differences between ab initio and semiempirical results. J Comp Chem 1992;13:842–50.
-
Alkorta I, Villar HO, Arteca GA. Comparative study between ab initio and semiempirical electrostatic potentials on molecular surfaces. J Comp Chem 1993;14:530–40.
-
Sovak M. Radiocontrast agents. New York: Springer-Verlag 1984.
This article has been cited by other articles:
|
|
|
|
 
P-A Carrupt and B Testa
Stabilization of the hydrophylic sphere of iobitridol
Br. J. Radiol.,
February 1, 2003;
76(902):
147 - 147.
[Full Text]
[PDF]
|
|
|
|
|
|
|
D Violon
Author's reply
Br. J. Radiol.,
February 1, 2003;
76(902):
147 - 148.
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
