Anal. Chem. 1998, 70, 1041-1044
Correspondence
Lipid Concentration Profile across the Wall of Pseudoatherosclerotic Synthetic Arterial Prostheses Using FTIR Microspectroscopy Diego Mantovani, Patrick Vermette, Robert Guidoin, Suzanne Bourassa, and Gae´tan Laroche*
Quebec Biomaterials Institute, Saint-Franc¸ ois d’Assise Hospital and Laval University, 10, rue de l’Espinay, Quebec City, Quebec, G1L 3L5, Canada
FTIR microscopy is a versatile technique successfully used to probe the subcellular chemical composition of atherosclerostic arterial walls. To design new vascular substitutes that resist lipid uptake (the major cause of the phenomenon referred to as atherosclerosis-like), identifying and understanding lipid distribution within the pseudoatherosclerosed arterial prostheses is of prime importance. Until now, the amount of lipids present within arterial prostheses that had been explanted from either animals (during in vivo trials) or humans (after the failure of vascular grafts) or had been submitted to in vitro investigations could only be measured through the use of histological techniques or radioactive labeling methods. We present here a novel method to quantitatively measure the lipid concentration profile within the wall of arterial prostheses by means of Fourier transform infrared microspectroscopy. Essentially, prostheses are fixed in a 1% osmium tetraoxide aqueous solution under vacuum and radially cut with a 5-µm thickness with a microtome. The sections are then placed onto BaF2 windows and observed with a microscope attached to a FTIR spectrometer with a 30 µm × 50 µm sampling area. The lipid concentration profile is obtained by scanning the prosthesis wall from the inner to the outer surface and reporting the corresponding integrated absorbance between 2700 and 3100 cm-1 against a calibration curve. The application of this technique constitutes the first quantitative measurement of the concentration of biological molecules within the wall of artificial arterial substitute. Fourier transform infrared (FTIR) spectroscopy is largely recognized as a valuable tool for the characterization of both biological molecules and polymer structures and is currently used to probe the subcellular chemical composition of atherosclerostic arterial walls.1,2 In implanted material research, FTIR spectroscopy provides new insight as a biomedical research tool, because * Corresponding author: (phone) (418) 525-4417; (fax) (418) 525-4372; (email)
[email protected]. (1) Koenig, J. L. Adv. Polym. Sci. 1983, 54, 87-154. S0003-2700(97)01061-5 CCC: $15.00 Published on Web 01/27/1998
© 1998 American Chemical Society
the information that is provided is essentially related to the molecular structure of the biochemical compounds of interest, and the information sought may be obtained using relatively small amounts of material.3 These two features add to the attractiveness of using FTIR spectroscopy to study failed human implanted devices after retrieval from the implantation site.4 First, the design of blood-compatible surfaces for cardiovascular applications requires the knowledge of the interaction between the device and the surrounding human tissues.5,6 Second, the fact that a few cubic millimeters of volume are sufficient to obtain high-quality FTIR spectra constitutes an important advantage: less material is used and more analyses may be performed. Materials explanted from humans are crucial to research as they contain vital information regarding the actual effects of the implantation; however, only a limited quantity of materials are available, while several investigations of any kind must be performed.7 In previous studies, we showed that FTIR spectroscopy, in the attenuated total reflectance mode (ATR), can be successfully used to investigate lipid uptake on the internal and external surfaces of microporous PTFE arterial prostheses explanted from humans.8,9 In addition, we suggested that lipid uptake be taken into account as an explanation for the poor healing response observed in synthetic arterial substitutes by revealing that lipids strongly adhered to both the external and luminal surfaces of the walls of expanded poly(tetrafluoroethylene) (ePTFE) arterial prostheses when implanted in humans. Unfortunately, these studies were limited by the fact that the FTIR-ATR technique did not allow us to investigate the lipid macromolecular transport across the arterial prosthesis wall. (2) Winter, S.; Gendreau, R. M.; Leininger, R. I.; Jakobsen, R. J. Appl. Spectrosc. 1982, 36, 404-409. (3) Gendreau, R. M. Trends Anal. Chem. 1986, 5, 68-71. (4) Giroux, T. A.; Cooper, S. L. J. Colloid Interface Sci. 1991, 146, 179-194. (5) Afanasyeva, N., Feng, H., Han, Y., Huang, L., Eds. Polymers and Biomaterials; Elsevier Science: Amsterdam, 1991; pp 337-341. (6) Gendreau, R. M.; Jakobsen, R. J. J. Biomed. Mater. Res. 1979, 13, 893906. (7) Kaplan, S. S. Med. Prog. Technol. 1994, 20, 209-230. (8) Mantovani, D.; Be´dard, E.; Marois, M.; Guidoin, R.; Laroche, G. J. Mater. Sci.-Mater. Med. 1996, 7, 40-45. (9) Mantovani, D.; Vermette, P.; Guidoin, R.; Laroche, G. Atherosclerosis, submitted for publication.
Analytical Chemistry, Vol. 70, No. 5, March 1, 1998 1041
Understanding the atherosclerosis-like process requires knowing the lipid concentration profile across the arterial wall, as it allows identification of macromolecular transport parameters related to the prosthesis’ boundary conditions. These parameters are of major importance in the development of a physicochemical treatment that can protect the artificial arterial prosthesis from the lipid uptake by specifically modulating the lipid transport process. Histological and radioactive labeling methods are the two major techniques used today to investigate biologic molecules present across the wall of synthetic microporous blood conduits. In synthetic arterial prostheses explanted from humans, methods directly derived from histological techniques are successfully used to investigate the presence of cellular components.10 Lipids in tissues can be specifically targeted using various well-known techniques, the most popular being the Sudan and the Oil Red-O methods.11,12 The objective of these many techniques is to reveal the presence of lipids by coloring them on the tissue sections. Sudan shows lipids by coloring them in black, while Oil Red-O colors lipids in red. As may be suspected, these techniques do not allow quantification of the amount of lipids within the microporous media; only a qualitative analysis may be carried out, with the more colored the sections, the higher the amount of lipids present. Quantifying the macromolecular transport profile across the wall of different semipermable conduits is a challenge that has concerned research for many years.13-15 The problem has been partially solved by introducing radioactive labeling of biological molecules. This technique is currently used in the quantitative analysis of macromolecular transport16,17 and consists basically in measuring the radioactivity of various thin, longitudinal sections from the external to the internal side of the artery. However, this method presents three major drawbacks. First, the molecules to be studied must be radioactively marked prior to being introduced into the body. Second, the use of radioactive materials requires that the experiments be performed in vitro or in vivo in animal models. Third, a concentration profile is obtained from subsequent longitudinal sections accurately cut with a microtome. FTIR microscopy, on the other hand, produces all of the information with only a single section sample. For these reasons, the use of radiolabeling techniques must be ruled out for studies that must be conducted on synthetic and biological materials implanted in humans and explanted after long periods of time. In fact, the main advantage of FTIR microspectroscopy is that the quantitative analysis of any biological molecule concentration may be done on virtually any types of in vitro or in vivo experiments without using any intrinsic or extrinsic probe. (10) Guidoin, R.; Chakfe´, N.; Maurel, S.; How, T.; Batt, M.; Marois, M.; Gosselin, C. Biomaterials 1993, 14, 678-693. (11) Drury, R. A. B.; Wallington, E. A. Carleton’s histological technique; Oxford Medical Publications: Oxford, U.K., 1980; pp 285-297. (12) Bancroft, J. D.; Stevens, A.; Turner, D. R. Theory and practice of histological techniques; Churchill Livingstone: Edinburgh, Scotland, 1990; pp 215-244. (13) Fry, D. L.; Vaishnav, R. M. In Basic hemodynamics and its role in disease processes; Patel D. J., Vaishnav R. M., Eds.; University Park Press: Baltimore, MD, 1980, pp 425-482. (14) Deng, X.; Marois, Y.; Guidoin, R.; Merhi, Y.; Stroman, P.; King, M. W.; Douville Y. Artif. Organs 1996, 20, 1208-1214. (15) Kim, W. S.; Tarbell, J. M. J. Biomech. Eng. 1994, 116, 156-163. (16) Caro, C. G.; Lever, M. J.; Laver-Rudich, Z.; Meyer, F.; Liron, N.; Ebel, W. Parker, K. H.; Winlove, C. P. Atherosclerosis 1980, 37, 397-411. (17) Deng, X.; King, M. W.; Guidoin, R. ASAIO J. 1995, 41, 5 8-67.
1042 Analytical Chemistry, Vol. 70, No. 5, March 1, 1998
While FTIR microspectroscopy has already been used for microimaging of atherosclerotic rabbit arteries,18 it has yet to be used to create three-dimensional images representing macromolecular transport across the wall of synthetic semipermeable membranes. The present paper accurately addresses this aspect by identifying a lipid concentration profile across the prosthesis wall with a 30 µm × 50 µm spatial resolution by means of Fourier transform infrared microspectroscopy. With this objective in mind, we undertook to develop a quantitative method to measure the lipid concentration profile within the wall of ePTFE arterial prostheses. EXPERIMENTAL SECTION Sample Preparation. Expanded ePTFE prostheses with a wall thickness of approximately 1 and 0.5 mm were kindly provided by Baxter HealthCare Corp. (Laguna Hills, CA) and GoreTex Corp. (Flagstaff, AZ), respectively. The prostheses had an internal diameter of 10 mm and presented a microporous structure similar to that previously described for other similar prostheses made of microporous ePTFE.19 Validation of the analytical method was achieved by wetting a small piece of ePTFE arterial prosthesis with a mixture of 1% egg phosphatidylcholine in a 87:3 chloroform-methanol solution. Following a 2-h immersion under stirring in this solvent, the prosthesis was placed in a vacuum oven at 40 °C for 24 h in order to promote a paraboliclike lipid distribution in the prosthesis wall. In addition, the effect of the time exposure to a lipid dispersion under identical flow rate and pressure conditions was evaluated by mounting two 5-cmlong prosthesis samples on a continuous pulsatile flow system that has been described elsewhere.20 Following these procedures, lipids trapped within the prosthesis wall were fixed in a 1% osmium tetraoxide solution, which reacted with the lipid ester carbonyl groups and unsaturated CdC bonds in hydrocarbon chains, thus preventing the solubilization of the lipid molecules during the subsequent washing in solvents. The exposure of the prosthesis to the osmium tetraoxide solution was made under vacuum, thus allowing penetration of the aqueous media within the hydrophobic structure of the poly(tetrafluoroethylene). The prosthesis was then returned to the vacuum oven for 48 h at 40 °C to completely remove the water for subsequent embedding in paraffin. Five micrometer-thick radial sections were made using a microtome, placed on a 2-mm-thick barium fluoride window, and then deparaffinized by subsequent dipping in toluene, ethanol, and distilled water. Identification of the Lipid Concentration Profile by FTIR Microspectroscopy. The prosthesis samples were investigated using a Magna-550 Fourier transform infrared spectrometer (Nicolet Instrument, Madison, WI) equipped with a germaniumcoated potassium bromide beam splitter. Spectra were recorded using a NIC-Plan infrared microscope by sizing the area of analysis to 30 µm × 50 µm using the redundant aperture system. The infrared radiation was focused on the prosthesis sample through a 15× Cassegrainian objective, collected back by means of a 10× Cassegrainian beam condenser, and directed to the (18) Kodali, D. R.; Small, D. M.; Powell, J.; Krishan, K. Appl. Spectrosc. 1991, 45, 1310-1317. (19) Guidoin, R.; Maurel, S.; Chakfe´, N.; How, T.; Zhang, Z.; Therrien, M.; Formichi, M.; Gosselin, C. Biomaterials 1993, 14, 694-704. (20) Vermette, P.; Thibault, J.; Laroche, G. Artif. Org., in press.
mercury-cadmium-telluride (MCT-A) detector located inside the microscope. One hundred scans with an optical retardation of 0.25 cm were coadded, triangularly apodized, and Fourier transformed to yield a 4-cm-1 spectral resolution. Lipid concentration within the prosthesis wall was determined by reporting the integrated infrared absorbance in the 2700-3100cm-1 spectral region on a seven-data-point calibration curve which was built by putting a 15-µL drop of a egg phosphatidylcholine solution in chloroform-methanol (87:3) with concentrations ranging from 1 to 11 mg/mL on 17-µm-thick polyethylene IR cards (3M, St. Paul, MN). These sampling windows were chosen because of their constant thickness over that of ePTFE IR cards from the same manufacturer. This is because interference fringes may easily be detected with polyethylene sampling windows while these features were not observed with the ePTFE substrate. The surface area of the drops were then precisely measured using BioQuant IV software (R&M Biometrics Inc., Nashville, TN), thus allowing us to calculate the volumetric lipid concentrations within the polyethylene windows. Because polyethylene strongly absorbs in the 2700-3100-cm-1 spectral region, the calibration curve was constructed using the integrated absorbance of the lipid ester carbonyl band multiplied by a factor of 4.22, corresponding to the ratio of the integrated absorbance of the features due to the methylene and methyl spectral mode vibration (between 2700 and 3100 cm-1) over that of the ester carbonyl band between 1680 and 1800 cm-1, as determined by the transmittance infrared spectrum of a lipid film deposited on a BaF2 window. This calculation was made because, as mentioned above, the fixation of the lipid molecules within the prosthesis’ structure is made with osmium tetraoxide, which reacts with the lipid ester carbonyl groups, thereby impeding the detection of these moieties under infrared spectroscopy. This calibration procedure enable us to obtain a linear calibration that correlated to the Beer-Lambert law with a correlation coefficient of 0.98. Finally, the absorbance axis of the calibration curve was multiplied by 5/17 to account for the various optical paths of the prosthesis sections and polyethylene windows. RESULTS AND DISCUSSION Validation of the Analytical Method. As mentioned above, pumping over the ePTFE arterial prosthesis after dipping it in a solution of egg phosphatidylcholine in an organic solvent produced a parabolic-like distribution of the lipid molecules within the prosthesis wall. Indeed, the evaporation of the solvent molecules from the prosthesis structure involved their migration from the center of the prosthesis wall to either the internal or external surface of the device. During this alteration, the solute molecules were carried in a similar fashion, although continuously precipitated as the amount of the solvent decreased. The resulting lipid concentration profile that was expected exhibited a low lipid amount in the center of the prosthesis wall, with continuously increasing amounts from the center to both the internal and external surfaces, as depicted in Figure 1. Figure 2 illustrates the procedure used to determine the lipid concentration profile within a radial section of an ePTFE arterial prosthesis wall. As illustrated in Figure 2B, each surface area of the prosthesis radial section analyzed by FTIR microspectroscopy gave rise to an infrared spectrum revealing features indicating the presence of both PTFE and lipids within the prosthesis
Figure 1. Description of the lipid movement within the ePTFE prosthesis wall following wetting in a 1% lipid solution in chloroformmethanol (87:3) and subsequent drying under vacuum.
Figure 2. Description of the analytical technique for the measurement of the lipid concentration profile within the ePTFE arterial prosthesis where a parabolic-like profile was induced. The infrared spectra were first recorded on a 5 µm-thick radial section by scanning the prosthesis wall from the internal to the external side with a 30 µm × 50 µm sampling area (A). Typical spectrum exhibits features characteristic of PTFE and lipids (B). Reporting the integrated absorbance between 2700 and 3100 cm-1 on a calibration curve allowed us to obtain a 3-D map to quantitatively describe the lipid concentration within the prosthesis wall (C).
structure. Reporting the integrated intensity between 2700 and 3100 cm-1 on a previously built calibration curve produced a map describing the lipid concentration within the prosthesis wall. As shown in Figure 2C, the lipid concentration profile obtained using FTIR microscopy was in agreement with the prediction presented in Figure 1, thus confirming the ability of the measurement method to produce a quantitative map of the amount of lipid molecules through the prosthesis wall. Lipid Concentration Profile Dependence with Time. Parts A and B of Figure 3 show the effect of exposing a ePTFE prosthesis to a 2.5% lipid dispersion under a flow rate of 2 L/min and a pressure of 80 mmHg for periods of 2 h and 1 week, Analytical Chemistry, Vol. 70, No. 5, March 1, 1998
1043
internal side of the prosthesis wall to approximately 600 µm with an evident accumulation on the external surface of the prosthesis wall up to a concentration of 0.09 mg/mm3. These results concur with previously published FTIR results on ePTFE arterial prostheses explanted from humans showing that lipid accumulation occurs rapidly after implantation and is highly dependent on the duration of implantation of the synthetic device within the body. Other experiments are currently being performed in order to model and mathematically describe the effect of parameters such as flow rate, pressure, time, and pulsatile conditions on the lipid distribution profile within the arterial prosthesis wall. In conclusion, FTIR microspectroscopy may be successfully used to determine quantitatively the concentration profile of lipids across the wall of polymeric blood conduits. The technique is far simpler, safer, and more accurate than are radiolabeling techniques or the more generally used histopathological analyses. The technique developed here may be successfully adapted to a wide variety of biomedical applications requiring the quantitative study of the macromolecular transport of species such as lipids, proteins, enzymes, and lipoproteins across a permeable or semipermeable membrane.
Figure 3. Lipid concentration profile in the ePTFE arterial prosthesis wall exposed to a 2.5% lipid dispersion under a flow rate of 2 L/min at a pressure of 80 mmHg for (A) 2 h and for (B) 1 week.
respectively. As can be seen in Figure 3A, exposing the prosthesis to a continuous lipid dispersion flow for 2 h resulted in an almost flat lipid concentration profile, not exceeding a lipid concentration of 0.03 mg/mm3 throughout the prosthesis wall. A longer exposure to the lipid dispersion (Figure 3B) produced a lipid concentration that generally exceeded 0.03 mg/mm3 from the
1044 Analytical Chemistry, Vol. 70, No. 5, March 1, 1998
ACKNOWLEDGMENT This work was supported by NSERC-Canada, FRSQ-Quebec (G.L.) and the Quebec Biomaterials Institute. D.M. and P.V. are the recipients of Ph.D. and M.Sc. scholarships, respectively, from FCAR (Quebec). Our thanks to Mrs. Claire Kingston for her technical assistance.
Received for review September 24, 1997. December 4, 1997. AC971061R
Accepted
