Fourier Transform Infrared Spectroscopic Determination of the

Department of Pharmaceutics, Faculty of Pharmacy, Utrecht Institute for Pharmaceutical Sciences (UIPS), Utrecht University, P.O. Box 80 082, 3508 TB U...
2 downloads 0 Views 56KB Size
3466

Langmuir 2002, 18, 3466-3470

Fourier Transform Infrared Spectroscopic Determination of the Hydrolysis of Poly(ethylene glycol)Phosphatidylethanolamine-Containing Liposomes E. A. A. M. Vernooij,§ J. J. Kettenes-van den Bosch,*,‡ and D. J. A. Crommelin†,§ Department of Pharmaceutics, Faculty of Pharmacy, Utrecht Institute for Pharmaceutical Sciences (UIPS), Utrecht University, P.O. Box 80 082, 3508 TB Utrecht, The Netherlands, Department of Biomolecular Mass Spectrometry, Utrecht Institute for Pharmaceutical Sciences (UIPS) and Bijvoet Center for Biomolecular Research, Utrecht University, P.O. Box 80 082, 3508 TB Utrecht, The Netherlands, and OctoPlus BV, Zernikedreef 12, 2333 CL Leiden, The Netherlands Received October 24, 2001. In Final Form: March 5, 2002 For poly(ethylene glycol)-derivatized liposomes (PEG-liposomes), the chemical stability of both the phospholipid ester groups and the urethane anchor between the PEG chain and the phospholipid are two important factors determining their long-term stability. High-performance liquid chromatography (HPLC) quantification of PEG-derivatized phospholipids such as PEG-distearoylphosphatidylethanolamine (PEG(DS)PE) is hampered by the molecular mass distribution of the PEG polymer. We describe a novel Fourier transform infrared (FTIR) method to monitor the hydrolysis of (PEG-)phospholipids. The integrated intensities of the CdO stretching bands of the various components of hydrolyzed (PEG-)phospholipid samples are determined by curve fitting. Separation of intact lipids and hydrolysis products is not required. To validate the method, the hydrolysis of dipalmitoylphosphatidylcholine/distearoylphosphatidylethanolamine (DPPC/DSPE) liposomes was monitored both by HPLC and FTIR. The results from both methods are in excellent agreement. Under the experimental conditions (pH 9, 70 °C), the hydrolysis rate constant of esters in liposomes was hardly affected by the presence of hydrophilic PEG chains at the surface of the liposomes. The stability of the PEG-phospholipid linkage could also be assessed. Although the urethane group is probably located in a more hydrophilic environment than the ester groups, hydrolysis of the PEG-PE anchor is still much slower than that of the esters.

Introduction In recent years, poly(ethylene glycol)-derivatized phospholipids, in particular poly(ethylene glycol)-derivatized phosphatidylethanolamines (PEG-PEs), have been used extensively as components of liposomes for drug formulations. By derivatization with PEG, the lifetime of the liposomes in the blood circulation is extended, because they are not recognized by the immune system and thus not prematurely removed by phagocytosis.1-3 A prolonged lifetime results in an increased probability of liposomes to reach their target sites, areas of enhanced vascular permeability, for example, at tumors and at sites of infection and inflammation.4-6 In addition, PEG-derivatized liposomes, because of their longevity, can be used * To whom correspondence should be addressed. Tel: +31 30 2536796. Fax: +31 30 2518219. E-mail: [email protected]. † Department of Pharmaceutics, Faculty of Pharmacy, Utrecht Institute for Pharmaceutical Sciences (UIPS), Utrecht University. ‡ Department of Biomolecular Mass Spectrometry, Utrecht Institute for Pharmaceutical Sciences (UIPS) and Bijvoet Center for Biomolecular Research, Utrecht University. § OctoPlus BV. (1) Woodle, M. C.; Lasic, D. D. Biochim. Biophys. Acta 1992, 1113, 171-199. (2) Storm, G.; Belliot, S. O.; Daemen, T.; Lasic, D. D. Adv. Drug Delivery Rev. 1995, 17, 31-48. (3) Long Circulating Liposomes: Old Drugs, New Therapeutics; Woodle, M. C., Storm, G., Eds.; Springer-Verlag: Berlin, 1998. (4) Storm, G.; ten Kate, M. T.; Working, P. K.; Bakker-Woudenberg, I. A. J. M. Clin. Cancer Res. 1998, 3, 111-115. (5) Bakker-Woudenberg, I. A. J. M.; Lokerse, A. F.; ten Kate, M. T.; Mouton, J. W.; Woodle, M. C.; Storm, G. J. Infect. Dis. 1993, 168, 164171. (6) Boerman, O. C.; Oyen, W. J. G.; van Bloois, L.; Koenders, E. B.; van der Meer, J. W. M.; Corstens, F. H. M.; Storm, G. J. Nucl. Med. 1997, 38, 489-493.

for vascular imaging agents for both nuclear medicine and magnetic resonance imaging. For drug formulations, the chemical stability of their components must be assessed in connection with shelf life and storage conditions. The stability of phospholipids is partially determined by the stability of the phospholipid ester groups. Another factor for PEG-PE is the stability of the anchor between PEG chain and phospholipid. Often, the PEG chain is linked to a phosphatidylethanolamine molecule via a urethane bond. In a homogeneous aqueous environment, urethane linkages are more stable than ester bonds.7 However, in the nonhomogeneous environment of a liposomal bilayer this may not be so, because of the variation in the dielectric constant. The dielectric constant varies from 2 in the hydrophobic region of the membrane to 5-40 in the more hydrophilic phospholipid headgroup region.8,9 The acyl ester groups are at the boundary of the hydrophobic and the hydrophilic region of the liposomal bilayer;10 the urethane group is in the hydrophilic region. Therefore, we also tried to get an indication of the stability of the urethane linkage in PEG-PE incorporated in liposomes. Usually, the stability of a component, in the formulation itself or in a model system, is measured as the decrease in concentration as a function of time and parameters such as pH and temperature. For common phospholipids, stability studies have been carried out with high-performance liquid chromatography (7) Gombotz, W. R.; Pettit, D. K. Bioconjugate Chem. 1995, 6, 332351. (8) Bellemare, F.; Fragata, M. J. Colloid Interface Sci. 1980, 77, 243252. (9) Kimura, Y.; Ikegami, A. J. Membr. Biol. 1985, 85, 225-231. (10) Scherer, J. R. Biophys. J. 1989, 55, 957-964.

10.1021/la011589y CCC: $22.00 © 2002 American Chemical Society Published on Web 04/06/2002

FTIR Evaluation of PEG-Phospholipid Hydrolysis

Langmuir, Vol. 18, No. 9, 2002 3467

Table 1. Band Maxima, Widths and Shapes of Carbonyl Stretch Vibrations As Used for Curve Fitting (Mathematically Derived from the Spectra of (Phospho)lipid Standards) carbonyl type intact di-esters lysophospholipids fatty acid monomer fatty acid dimer + PEG-DSPE urethane carboxylate anion

band position bandwidth fraction (cm-1) (cm-1) Lorentzian 1733.404 1729.350 1726.359 1709.267

27.641 28.948 50.442 20.101

0.17312 0.19163 0.00250 0.53176

1680.408

23.456

0.00040

(HPLC).11-13 However, such an HPLC method cannot be applied for PEG-derivatized phospholipids, because of the molecular mass distribution of the PEG polymer. We now describe a novel and simple Fourier transform infrared (FTIR) method to monitor the hydrolysis of PEGphosphatidylethanolamines: the integrated intensities of the CdO stretching bands between 1750 and 1650 cm-1 of the various components of partially hydrolyzed PEGPE samples are determined by curve fitting. Materials and Methods Materials. 1,2-Dipalmitoyl-sn-glycero-3-phosphocholine (dipalmitoylphosphatidylcholine or DPPC) was a gift from Lipoid GmbH (Ludwigshafen, Germany). N-(Carbonyl-methoxypoly(ethylene glycol) 2000)-1,2-distearoyl-sn-glycero-3-phosphoethanolamine, sodium salt (PEG-(DS)PE), was purchased from Genzyme Pharmaceuticals (Cambridge, MA). 1,2-Distearoylsn-glycero-3-phosphoethanolamine (distearoylphosphatidylethanolamine or DSPE) and 1-palmitoyl-2-hydroxy-sn-glycero3-phosphocholine were obtained from Avanti Polar Lipids (Alabaster, AL). The purity of these lipids was g99%. Palmitic acid 99% was purchased from Sigma (St. Louis, MO). The solvents used were of analytical grade, and water was of reversed osmosis quality. Methods. Liposomes were prepared by the film method. DPPC/DSPE 93/7 (mol/mol) and DPPC/PEG-DSPE 93/7 (mol/ mol) mixtures were dissolved in methanol/chloroform 1/1 (v/v) in a round-bottomed flask. The organic solvent was removed under reduced pressure. Then the film was hydrated with a 500 mM ammonium chloride buffer, pH 9. The final phospholipid concentration was 20 mM. For FTIR and HPLC analysis, separate samples were prepared from liposomes of the same batch. Liposomes were filled into ampules under nitrogen and sealed. The ampules were kept in a 70 °C water bath for hydrolysis studies. Ampules were removed from the water bath at appropriate time intervals and stored in the freezer until analysis. All experiments were carried out in duplicate (n ) 2). For FTIR analysis, 500 µL of methanol, 500 µL of chloroform, and 100 µL of 4 M hydrochloric acid were added to 500 µL of the liposome dispersion.14 From the resulting chloroform layer, containing the phospholipids and their hydrolysis products, 400 µL was removed, and the solvent was evaporated under nitrogen. The residue was redissolved in 500 µL of chloroform. FTIR spectra were acquired with a Bio-Rad FTS6000 (Bio-Rad Laboratories, Cambridge, MA) equipped with a dTGS detector. Reference samples of a lysophospholipid (lyso-PC) and a fatty acid (palmitic acid) were dissolved in chloroform and measured. The phospholipid solutions (ca. 12 mg/mL) were analyzed in the transmission mode. A transmission liquid cell with KBr windows and a path length of about 200 µm was used. Sixty-four scans were averaged at 2 cm-1 resolution, smoothed with a 9-point Savitzky-Golay smoothing function to remove white noise, and (11) Grit, M.; Underberg, W. J. M.; Crommelin, D. J. A. J. Pharm. Sci. 1993, 82, 362-366. (12) Grit, M.; Zuidam, N. J.; Underberg, W. J. M.; Crommelin, D. J. A. J. Pharm. Pharmacol. 1993, 45, 490-495. (13) Grit, M.; de Smidt, J. H.; Struijke, A.; Crommelin, D. J. A. Int. J. Pharm. 1989, 50, 1-6. (14) Bligh, E. G.; Dyer, W. J. Can. J. Biochem. Physiol. 1959, 37, 911-917.

Figure 1. FTIR spectrum of a hydrolyzed DPPC/PEG-DSPE liposome sample in chloroform (t ) 3000 min). The magnification shows the carbonyl stretching region, 1790-1650 cm-1. The solid line is the measured spectrum, the dotted lines are the fitted bands, and the dashed line is the resulting curve fit. baseline-corrected between 1790 and 1650 cm-1. Band positions and shapes used for curve fitting were determined from the spectra of phospholipid and fatty acid standards. Band shapes were approximated by mixed Gaussian and Lorentzian functions, in such a way that the χ2 parameter was minimal.15 The results are summarized in Table 1. All calculations were performed with Bio-Rad Win-IR version 4.11 (GRAMS-386 based) software. The hydrolysis rate constants were calculated from the area of the 1733 cm-1 band (intact di-esters) as the slope of the plot of ln(% intact di-esters remaining) against time(s). The DPPC/DSPE liposome samples were also analyzed with HPLC. To 100 µL of the liposome dispersion, 250 µL of methanol, 250 µL of chloroform, and 100 µL of 1 M hydrochloric acid were added. The chloroform layer was removed. To 100 µL of this chloroform phase, 400 µL of eluent was added. The final lipid concentration was about 2 mM. The samples were transferred into a Spark Basic-Marathon 816 autosampler (Spark, Emmen, The Netherlands), fitted with a 200 µL loop. The injection volume was 100 µL. The solvent delivery system was a Waters 515 HPLC pump (Waters Associates, Milford, MA). The HPLC system consisted of a Lichrosphere endcapped RP-18 (5 µm, 2 mm × 250 mm) column; the mobile phase was acetonitrile/methanol/ triethylamine 550/1000/25 (w/w/w), and the flow rate was 1 mL/ min.16 The DPPC in the phospholipid mixture was quantified with a Waters 410 Differential Refractometer (Waters Associates, (15) Czarnecki, M. A.; Ozaki, Y. Spectrochim. Acta, Part A 1996, 52, 1593-1601. (16) Brouwers, J. F. H. M.; Gadella, B. M.; van Golde, L. M. G.; Tielens, A. G. M. J. Lipid Res. 1998, 39, 344-353.

3468

Langmuir, Vol. 18, No. 9, 2002

Vernooij et al.

Figure 2. Chemical changes of (PEG-)phospholipid carbonyl groups during hydrolysis. For clarity, hydrolysis is presented only for the ester group in the 2 position but may also occur at the 1 position. Milford, MA). Calibration curves obtained with phospholipid standards were linear in the range of 0-2 mM.

Results and Discussion Figure 1 shows the FTIR spectrum obtained for partially hydrolyzed DPPC/PEG-DSPE liposomes. To monitor the hydrolysis of (PEG-)liposomes, we studied the carbonyl region of the spectrum. The band corresponding to the carbonyl stretch vibration is found at 1750-1650 cm-1 and is well separated from other spectral features. Quantification of phospholipid hydrolysis from other regions in the spectrum is difficult: each individual functional group shows several peaks resulting in extensive overlap. A phospholipid molecule contains two ester carbonyl groups. In PEG-DSPE, the PEG chain is coupled to the amine group of distearoylphosphatidylethanolamine via a urethane linkage. Therefore, this compound contains an additional carbonyl group, the urethane carbonyl (Figure 2). During hydrolysis, several changes of the carbonyl groups may occur (see Figure 2). The acyl esters in (PEG-)phospholipids are hydrolyzed to (PEG-)lysophospholipids and free fatty acids. Depending on the pH,

the free fatty acids are protonated or in their anionic form. During hydrolysis of the urethane group in PEG-DSPE, HO-PEG and the carbamic acid of DSPE are formed. The latter product rapidly decarboxylates to yield DSPE and CO2. Hydrolysis of the urethane group shows up in the IR spectrum as a decrease of the peak area of the urethane carbonyl group. In summary, during hydrolysis of DPPC/PEG-DSPE liposomes, the following changes in the carbonyl groups of (PEG-)phospholipid samples are expected: ester carbonyl groups of the intact (PEG-)phospholipids decrease during hydrolysis; ester carbonyl groups of the (PEG-)lysophospholipids show an initial increase during hydrolysis and then decrease; free fatty acid carbonyl groups increase during hydrolysis; urethane carbonyl groups of the PEG-(lyso)phospholipid decrease during hydrolysis. All carbonyl groups show IR maxima at different wavenumbers. A mixture of DPPC/DSPE 93/7 (sample t ) 0) gives a single band with a maximum at 1733 cm-1. The carbonyl of (PEG-)lysophopholipids, one of the hydrolysis products, was fitted with a band with its maximum at 1729 cm-1, the value obtained from the

FTIR Evaluation of PEG-Phospholipid Hydrolysis

Figure 3. Hydrolysis of liposomes at pH 9 and 70 °C: DPPC/ DSPE 93/7 as determined with HPLC (9) and FTIR (4) and DPPC/PEG-DSPE 93/7 as determined with FTIR (O). Data points represent the mean of two measurements. The correlation coefficient of the DPPC/DSPE HPLC and FTIR data is 0.997. Typically, the kobs values from duplicate experiments differed by less than 10%.

reference spectrum of a lyso-PC. Free fatty acids show a strongly asymmetric carbonyl band, consisting of three unresolved bands: a strong one at 1709 cm-1, assigned to a CdO stretching vibration of the dimeric form of the acid, and a weaker one at 1726 cm-1, attributed to a CdO stretching vibration of the monomeric form of the acid.15 The third band at 1680 cm-1 is usually assigned to the carboxylate anion. For DPPC/PEG-DSPE 93/7, the carbonyl band is broader than that of a DPPC/DSPE 93/7 mixture. This broadening was attributed to the presence of the urethane carbonyl. The maximum of this urethane band was found at 1712 cm-1, by subtracting the DPPC/ PEG-DSPE and DPPC/DSPE spectra. It was fitted not separately but together with the fatty acid carbonyl at 1709 cm-1. A summary of band positions, widths, and shapes as used for curve fitting is given in Table 1. Mixed Lorentzian and Gaussian curves were used since these were found to give the best fits. Initially, the band positions and shapes were allowed to vary within a small range around the average position determined from spectra of the control samples. However, this led to erroneous curve fitting results. By narrowing the constraints and fixing the band positions and shapes to the values as determined from the spectra of phospholipid and fatty acid standards, good fits were obtained between observed and fitted bands for (mixtures of) standards and hydrolyzed phospholipid samples (see Figure 1). Hydrolysis of DPPC/DSPE Liposomes As Determined with HPLC and FTIR. To establish the hydrolysis of DPPC/DSPE 93/7 liposomes, the area of the 1733 cm-1 band was monitored by curve fitting. To validate the FTIR method for the determination of phospholipid kobs values, the FTIR results obtained for DPPC/DSPE 93/7 liposomes were compared with those obtained from HPLC analysis. The results are shown in Figure 3. With HPLC, only the peak area of DPPC was evaluated. The FTIR method, however, determines the sum of the hydrolysis of both DPPC and DSPE in the liposomes. Other studies17 indicated that the kobs values of the two phospholipid classes in the liposomes may be different. Whether (17) Vernooij, E. A. A. M.; Kettenes-van den Bosch, J. J.; Underberg, W. J. M.; Crommelin, D. J. A. J. Controlled Release 2002, 79, 299-303.

Langmuir, Vol. 18, No. 9, 2002 3469

Figure 4. Band broadening of the 1733 cm-1 carbonyl stretch vibration (di-esters) of DPPC/DSPE 93/7 (4) and DPPC/PEGDSPE 93/7 (O) liposome samples as a function of hydrolysis time (pH 9, 70 °C). The broadening includes the lysophospholipid, fatty acid, and, for DPPC/PEG-DSPE mixtures, also the urethane carbonyl stretching bands and is indicated by T.

or not this was the case in the present liposomes could not be determined: the relative amount of DSPE in the liposomes was too low to allow adequate quantification by HPLC. The contribution of the (possibly) deviating kobs for DSPE was probably negligible: the hydrolysis profiles as obtained by HPLC and FTIR (Figure 3) were in excellent agreement. The kobs values were 4.1 and 4.0 × 10-6 s-1 determined from HPLC and from FTIR data, respectively. These values are similar to those found for DPPC/1,2dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) 3/1 liposomes at pH 9 and 70 °C.17 Hydrolysis of DPPC/PEG-DSPE Liposomes. HPLC quantification of PEG-phospholipids is hindered by the large number of peaks due to the molecular mass distribution of the PEG polymer. Therefore, it is difficult to separate the PEG-phospholipid hydrolysis products from the intact lipids. For FTIR measurements, separation of intact and degraded PEG-phospholipids is not required: the spectral features of degraded PEG-phospholipid samples do not interfere. To monitor the hydrolysis of DPPC/PEG-DSPE 93/7 liposomes, the area of the 1733 cm-1 band was determined by curve fitting. The results are shown in Figure 3. The kobs for DPPC plus DSPE in DPPC/DSPE and for DPPC plus PEG-DSPE in DPPC/PEG-DSPE liposomes, as determined with FTIR, is 4.0 and 4.8 × 10-6 s-1, respectively. Apparently, the attachment of hydrophilic PEG chains to the surface hardly affects the kobs of the phospholipids in the liposomal membrane. Hydrolysis of PEG-DSPE Urethane. In nonhydrolyzed DPPC/DSPE liposomes, the ester carbonyl band is symmetric. During hydrolysis, the bandwidth increases (the band broadens) with time as a result of the appearance of lysophospholipids and fatty acid. The carbonyl absorption band of DPPC/PEG-DSPE has its maximum at the same frequency as DPPC/DSPE, but the band is broader due to the presence of the urethane carbonyl. To get an indication of the stability of the PEG-PE urethane linkage, band broadening was monitored over time in both DSPC/DSPE and in DSPC/PEG-DSPE liposomes. The difference in “broadening” was found to be constant over at least 2 half-lives (see Figure 4). From Figure 3, it was concluded that the kobs values of the ester groups in DPPC/DSPE and DPPC/PEG-DSPE liposomes are approximately the same. Therefore, for both types of

3470

Langmuir, Vol. 18, No. 9, 2002

Vernooij et al.

liposomes the same contribution of lysophospholipids and fatty acids to the band broadening is anticipated at all time points. The constant broadening difference between DPPC/DSPE and DPPC/PEG-DSPE liposomes, therefore, indicates that even after 2 half-lives of the phospholipid ester groups, no significant hydrolysis of the urethane linkage has occurred. Thus, it was concluded that although the dielectric constant at the site of the urethane group in DPPC/PEG-DSPE liposomes is probably higher than that near the ester groups, the urethane linkage is still much more stable than the ester bonds.

HPLC quantification chromatographic separation of intact PEG-phospholipids and their degradation products is an essential requirement, no separation is needed with the FTIR method. The presence of hydrophilic PEG chains at the surface of the liposomes hardly affected the kobs of the phospholipid ester groups in the membrane at pH 9. Also, the stability of the PEG-phospholipid linkage could be determined. The dielectric constant near the urethane group is probably higher than that at the ester groups, but the PEG-phospholipid anchor is still much more stable than the phospholipid ester groups.

Conclusion

Acknowledgment. The authors thank Lipoid GmbH (Ludwigshafen, Germany) for their generous gift of phospholipids and for their financial support of this research project. We also thank Dr. M. van de Weert for valuable discussions.

The kobs values for the phospholipids in DPPC/DSPE liposomes obtained with the novel FTIR technique described above are in excellent agreement with data obtained by HPLC. Therefore, FTIR is a useful tool for evaluation of PEG-phospholipid hydrolysis: whereas for

LA011589Y