Return to Cover of Wave of the Future

Near-Infrared Imaging and Spectroscopy in Stroke Research:

Lipoprotein Distribution and Disease

University of Kentucky Medical Center

Lexington, KY 40536-0082

(1) Present address: Department of Neurosurgery, University of Wisconsin Clinical Science Center, 600 Highland Ave., H4/338, Madison, WI 53792-3232.

(2) Department of Pathology, University of Kentucky Medical Center, Lexington, KY 40536-0093.

(3) Author to whom correspondence should be addressed. Telephone: (606)257-9232

Email: Lodder@pop.uky.edu

INTRODUCTION

Stroke is a critical problem in the U.S. that affects more than 500,000 people annually. Research into the causes of stroke and testing of drug therapies to reduce ischemic and postischemic damage to the brain is hampered by an inability to continuously follow the physical and chemical events that occur during ischemia and reperfusion in vivo. Near-infrared (near-IR) spectrometry has recently been used to observe stroke-induced changes in the lipids and proteins of whole brain samples in vitro and in vivo (1). The examination of whole brains was made possible by a combination of hardware and software techniques designed to make the sample presentation to the spectrometer more reproducible. Near-IR spectrophotometry of gerbil brain tissue discriminated between adult (3-4 months of age) and aged (18-20 months of age) brains as well as between brains exposed to 5 and 10 min ischemia. The near-IR analytical method has many applications in aging and stroke research, including the noninvasive determination of age from brain spectra obtained transcranially, simultaneous multicomponent analysis of lipids and proteins, and quantification of stroke-induced edema. More recently, near-IR research has moved to the events preceding ischemic injury.

Arterial disease contributes to most of the deaths in the United States. Epidemiological studies performed over a period of years have indicated that reduction of blood cholesterol levels significantly reduces the risk of atherosclerosis, ischemia, myocardial infarction and death. For some time these data have been cited in experimental attempts to prevent arterial disease. One study estimates that 60 million people, or 36% of all adults between the ages of 20 and 74, have cholesterol levels high enough to warrant medical advice and intervention (2). A 1% reduction in plasma cholesterol concentration in individuals at risk for cardiovascular disease has been shown to reduce the risk of cardiac events in these individuals by approximately 2%. More recent data indicate that lowering cholesterol improves the condition of coronary arteries partially blocked by atherosclerotic lesions, actually effecting a regression of the disease process (3). These data have been used to make a case for creating a target level for total blood cholesterol of 200 mg/dl (4), a level below the average of the U.S. population.

The main carrier for cholesterol in the blood stream is low density lipoprotein (LDL) and it is the primary source of cholesterol deposits in the arteries at risk. Several studies have correlated an increase in serum LDL with an increased risk of atherosclerosis (5). Evidence suggests that the oxidation of LDL may increase its involvement in atherosclerotic lesions. It is known that oxidatively modified LDL increases the lysolecithin content and increases its negative charge. The protein's density is increased and the content of polyunsaturated fatty acid is decreased leading to the fragmentation of apolipoprotein B (6). These changes lead to changes in the electrophoretic mobility and the aggregability of LDL as well as the in the interaction of LDL with the arterial wall. The oxidized species is increasingly recognized by the acetylated LDL (scavenger) receptor on the macrophages and on the endothelial cells. Further, oxidized LDL is capable of vascular cell injury as well as the modification of cellular products of gene expression. The oxidation of LDL is hypothesized to occur through a free-radical mechanism (7). The endogenously generated free radicals, in turn, may be related to continued generation of leukotrienes by a lypoxygenase mechanism (8). Recent developments have suggested that the oxidation of LDL is a critical step in the acceleration of atherosclerosis to clinically significant levels (5).

Oxidized LDL (oxLDL) also induces a monocyte binding activity in endothelial cells and is chemotactic to monocytes. OxLDL affects cytokine gene expression. (9) With the attraction of monocytes and their accumulation of oxidized LDL via the scavenger receptor pathway, foam cell formation is noted in the arterial wall (5,9). The foam cell precursors to fatty streaks are hypothesized to arise from monocytes and macrophages, which migrate into the intima, although smooth muscle cell involvement is postulated as well (5). The oxLDL may initially act as a cytotoxic agent, as shown in cell culture studies (10).

The role of lipid oxidation in atherosclerosis has been inferred primarily by the examination of the naturally occurring and synthetic antioxidants. Ascorbic acid and bilirubin have been shown to be effective in protecting lipids from oxidative damage (11). Alpha-tocopherol acetate (Vitamin E) has been shown to be protective in the lipid profile of atherogenesis (12), is active as an antioxidant (13), and is believed to share the high affinity receptor for LDL in cultures of fibroblasts (14).

The lipophilic antioxidant probucol has been shown to decrease both the oxidation of lipoproteins as well as their cytotoxicity (15). In addition to altering the oxidative modification of LDL, studies in Watanabe Rabbits show that probucol pretreatment inhibits the degradation of LDL in fatty streaks. Probucol pretreatment significantly lowered the rate of atherosclerotic progression in such animals even when correcting for the lipid lowering effects of probucol (16,17) These data suggest that antioxidants may exert an effect by limiting the oxidized LDL component of atherosclerotic progression beyond that to be expected from lowering native LDL levels alone.

Stereotypic locations for atherosclerotic development have been identified and include the coronaries and carotid vessels (18-22). The carotid arteries have been particularly accessible to noninvasive study due to their subcutaneous location and progressively better understood role in the pathogenesis of cerebral ischemia and stroke (20,22). In addition, noninvasive sonographic determination of atherosclerotic plaque present at the carotid bifurcation has been correlated with overall ischemic risk and systemic atherosclerosis, and suggests carotid atherosclerosis as a marker of overall atherosclerotic insult to vessels (23).

While considerable interest has been placed on the geometry of the carotid artery atherosclerotic plaques and their relationship to ischemia (20,22), the actual biochemical composition of the plaque is of increasing importance. Approximately 30% of total fatty acids present in arterial walls lipids are readily oxidized (21).

Reference methods for lipoprotein and apoprotein determinations in plaques involving ultracentrifugation are destructive of the sample, cumbersome and/or expensive. Electrophoretic methods, immunofluorescence, and radioimmunoassays are far slower than a purely spectrometric method. Many times, even radioimmunoassays have coefficients of variation as high as 5-10% (24). The dynamic range and reproducibility of electrophoretic methods is even worse (25). The cost of immunoreagents and the time required to complete an analysis are prohibitive for many screening applications. Radioimmunoassays are complicated even further by the need to handle radioactive compounds. None of these methods are well suited to preserve the spatial relationship of compounds in the plaque.

There is currently no accurate nondestructive in vivo reference assay for HDL, LDL, or apolipoproteins immobilized in the walls of living arteries in humans. Fiber-optic catheters have been used to locate atherosclerotic lesions, but most spectrometric techniques can do no more than distinguish lesions from healthy arterial tissue (i.e., a detailed breakdown of constituent proteins is not possible). Research currently underway in this laboratory employs InSb focal plane arrays (FPA's) and PtSi CCD near-IR video cameras and tunable light sources to identify lesions in living arteries of human patients and map their chemical constituents in three spatial dimensions. Tunable light sources based on blackbody emission and tunable filters and monochromators, as well as a Nd:YAG-pumped KTP/OPO near-IR laser system, are used for different imaging experiments depending upon the light intensity required. Chemical analysis of lesions in vivo permits the kinetic study of atherogenesis and contributes to the understanding of lesion formation and growth. As new processes (e.g., oxidation of LDL) are identified as playing key roles in the initiation and progression of lesions, better treatment programs can be designed that focus on these mechanisms.

Near-IR spectroscopy has been used industrially in lipid analysis for years to determine saturation of unsaturated fatty acid esters (26). More recently, near-IR spectrometry was used in our laboratories to examine lipids in vitro and in vivo in gerbil brains following experimentally induced stroke and to identify nine different saturated and unsaturated fatty acids found in the gerbil brain (1). Near-IR spectrometry has also been used in our laboratories to analyze HDL, LDL, and cholesterol in the blood vessels of rats (27). Additionally, near-IR spectroscopy has been used to determine fat content of commercial meat products (28). Analytes including glucose, lactate, and many others have been determined simultaneously using near-IR spectrometry (29). FT-Near-IR imaging of lipid and protein in primate brain tissue has been described (30). In humans, near-IR spectroscopy has been used noninvasively to analyze deoxyhemoglobin in blood and to determine whole body fat (31,32). In our laboratories, near-IR imaging has been used in human stroke patients to locate atherosclerotic plaque by identifying and locating oxidized lipoprotein spectral signatures (33). A major advantage of near-IR spectral analysis is its chemical imaging ability. Additionally, near- IR spectral imaging provides information on details of various internal structures including muscle, bone, and arteries (34).

Near-IR methods are used in the following study to test the hypothesis that nondestructive spectrometric imaging of plaque lipoproteins relevant to atherosclerosis and stroke is as precise as protein extraction, ultracentrifugation and gel electrophoresis. In the process, the analytical power of near-IR cameras and MPP (massively parallel processor) supercomputing are demonstrated.

EXPERIMENTAL SECTION

Lipoprotein reference analyses. LDL was isolated from fresh plasma (< 24 hr old) and tissue plaques of all endarterectomy patients according to the method of Havel (35). A blood sample from each patient was withdrawn and plasma was isolated by centrifugation at 3,000g (4 C). For tissue plaque reference analysis, excised plaques were quickly rinsed in Kreb's physiological salt solution to remove adherent LDL and oxLDL before being frozen in liquid nitrogen. An aliquot of plasma or tissue extract was diluted with a 50 mM phosphate buffer containing the following preservatives (buffer A): 2.7 mM EDTA, 2 mM benzamidine, 10 &micro;M probucol, 1 &micro;M PPACK, 0.01% aprotonin, 0.008% chloramphenicol, and 0.008% gentamycin, 1 mM PMSF, 1 mM leupeptin, and 40 &micro;M elastinal. For the tissue plaque extraction, plaque segments were cut into small pieces (~ 1 mm) in a cooled, nitrogen-purged glove box while under the de-gassed extraction buffer using a custom-made immersible tissue chopper. Minced tissue was then extracted overnight at 4 C under nitrogen using an orbital shaker (10 rpm) in 0.14 M NaCl/0.01 M phosphate buffer, pH 7.2, containing the preservatives as in buffer A. The extract was collected by low-speed centrifugation at 4 C, washed once with extraction buffer, and the supernatants combined. The supernatant was transferred into 12-ml tubes, overlayered with 0.5 ml of water, and centrifuged at 100,000g for 30 min at 5 C. LDL (density, 1.019-1.063 g/ml) was isolated from plasma and plaque extracts by density gradient ultracentrifugation over a potassium bromide gradient at 200,000g (4 hours at 22 C). The LDL fraction (density of 1.019 - 1.063, determined by light illumination) was obtained and dialyzed overnight at 4 C against 0.14 M NaCl/0.01 M phosphate buffer (pH 7.4) containing 0.27 mM EDTA and 1 mM PMSF. After dialysis, the LDL fraction was sterile filtered (0.45 &micro;m) and stored under nitrogen at 4 C. Protein content was determined according to the 1971 method of Bradford. For comparison, commercial LDL (Sigma, St Louis) and oxLDL (LDL oxidized in the presence of 10 &micro;M copper sulfate at 37 C for 24 hrs) were used as reference standards. A calibration curve was constructed for the near-IR spectra of freshly prepared ox-LDL to allow quantitative determination of oxLDL as well as to identify spectral peaks that correlate with oxLDL concentration.

Reference assays for plasma and tissue extracts consisted of SDS-PAGE (36) (index of alterations in electrophoretic mobility) and measurement of thiobarbituric acid reactive substances (TBARS). Before electrophoresis, the samples were prepared in a buffer containing 0.063 M Tris-HCl, 2% SDS, 10% glycerol, 10 &micro;M BHT, and 0.001% bromphenol blue (pH 6.8) and heated for 3 min in a boiling water bath. SDS-PAGE was performed using 4-12% gradient gels (100 V, 30 mA for 75 min; Protean 2 mini gel electrophoresis system, Biorad, CA) in a running buffer containing 0.025 M Tris, 0.19 M glycine buffer, pH 8.3, containing 0.1% SDS. Coomasie brilliant blue or silver stains were used to visualize protein bands. Each stained gel containing extracted plaque lipoproteins was digitized using a Si CCD camera, and absorbance values were calculated for the bands in each lane. The bands on the gels were identified and quantified using molecular weight markers (broad range, Biorad, CA), commercial LDL, and oxLDL standards on each gel. Additional details and illustrations of these methods are available on the WWW at http://kerouac.pharm.uky.edu/.

The in vitro near-IR scanning of excised plaques (for comparison with in vivo scans obtained in the operating room) was conducted in a nitrogen-purged glove box at 4 C. The optical window in the glove box was maintained at -20 C for the frozen plaques, which were scanned immediately before chopping and biochemical extraction).

Constructing images from near-IR spectra. The BEST (Bootstrap Error-adjusted Single- sample Technique) calculates distances in multidimensional asymmetric nonparametric central 68% confidence intervals in spectral hyperspace (roughly equivalent to standard deviations). The BEST metric can be thought of as a "rubber yardstick" with a nail at the center (the multidimensional mean) (33). The stretch of the yardstick in one direction is therefore independent of the stretch in the other direction. This independence enables the BEST metric to describe odd shapes in spectral hyperspace (spectral-point clusters that are not multivariate normal, like the calibration spectra of many biological systems). BEST distances can be correlated to sample composition to produce a quantitative lipoprotein calibration, or simply used to identify regions with lipoprotein distributions similar to plaque in a spectral image. The BEST automatically detects samples and situations unlike any encountered in the original calibration, making it more accurate in biomedical analysis than typical regression approaches to near-IR analysis. The BEST produces accurate distances even when the number of calibration samples is less than the number of wavelengths used in calibration, in contrast to other metrics that require matrix factorization. Unlike its predecessor, the BEAST (37), the BEST retains the direction vector of a standard deviation in hyperspace throughout all calculations, an essential characteristic for multicomponent quantification of sample composition.

The BEST calculates the integral of a probability orbital in hyperspace by starting at the centroid of the orbital and working outward in all directions at a uniform rate. The distance between the center of a plaque orbital and a sample spectrum is proportional to the concentration(s) of the plaque constituent(s) responsible for the vector connecting the central and sample- spectral points. The direction of the vector identifies the constituent(s) of the plaque. The BEST direction and distance are typically used to create color contour plots of the spatial distribution of lipoproteins (33). In such plots, the contours are drawn at sequential distances in SD's, and RGB (red-green-blue) colors are used to denote class membership based on vector direction. The intensity of the color is proportional to the amount of substance present. Shades of blue are used to represent sample spectra similar to those already in the calibration set, while shades of red are used to represent sample spectra that contain the selected analyte. Shades of green are used to represent a second analyte or possible interfering effect. The BEST offers superior performance as an assimilation method (a method that progressively increases its analytical performance by incorporating previously unknown samples into its calibration). The calibration samples are analyzed by another reference method (such as those listed earlier) in the same manner that Beer's Law is used to develop a conventional spectrophotometric calibration.

In the BEST, a population P in a hyperspace R represents the universe of possible spectrometric samples (the rows of P are the individual samples, while the columns are the independent information vectors, such as wavelengths or energies). P* is a discrete realization of P based on a calibration set T, which has the same dimensions as P* and is chosen only once from P to represent as nearly as possible all the variations present in P.

P* is calculated using a bootstrap process by an operation (T), and P* has parameters B and C, where C = E(P) and B is the Monte Carlo approximation to the bootstrap distribution. The expectation value, E( ), of P is the center of P, and C is a row vector with as many elements as there are columns in P.

eq 1

Each new sample spectrum X is analyzed by an operation (T,B,X,C) (27), which projects a discrete representation of the probability density of P in hyperspace by many-one mapping onto the vector connecting C and X. X and C have identical dimensions. The directional standard deviation (SD) is found from the projected probability density in eq 1. The integral over the hyperspace R is calculated from the center of P outward. The calculation of a skew adjusted is based on a comparison of the expectation value C=E(P) and C=med(T), the median of T in hyperspace (with the same dimensions as C) projected on the hyperline connecting C and X in eq 2.

eq 2

The result of the corrected projection is an asymmetric that provides two measures of the standard deviation along the hyperline connecting C and X.

eq 3

eq 4

eq 3 in the direction of X in hyperspace, and eq 4 in the opposite direction along the hyperline connecting C and X. Skew adjusted SDs can be used to calculate mean distances between spectra of different samples.

The use of these equations in both quantitative and qualitative analysis of plaque provides a number of advantages over all other methods of analysis:

1. No analytical assumptions are required to make the problem solvable. Other chemometric methods assume that no discriminating variable (wavelength) is a linear combination of other discriminating variables, that the covariance matrices for all spectral groups are approximately equal, and that each group is drawn from a population that is normally distributed on the discriminating variables. None of these assumptions are usually true and violations of these assumptions increase the likelihood of producing incorrect quantitative and qualitative analytical results.

2. This nonparametric assimilation method can be used with full spectra of samples, which often include a thousand or more variables that describe each sample. The large global memory on the newest supercomputers has made possible the manipulation of images involving tens of thousands of spectra at thousands of wavelengths. No wavelength selection procedures or data compression techniques, such as principal-axis transformation or Fourier transformation, are required by this assimilation method in order to analyze complete spectra. Collinearity in the discriminating variables (wavelengths) does not degrade the results. (Collinearity disturbs conventional matrix techniques like the Mahalanobis method, which relies on matrix inversion or factorization to produce a distance in SDs.)

3. The vector CX in the assimilation method identifies the sample components. The metric is calculated by a highly parallel code that can be distributed across as many processors as are available, and can be used in computerized searches of spectral libraries for qualitative analysis of plaque samples. The length of the vector is proportional to the concentrations of the plaque constituents. Thus, quantitative analytical capabilities are also provided by the same assimilation method, which can still be distributed across as many processors as are available.

4. The BEST metric is not only more accurate and precise than the Mahalanobis metric (the metric commonly used in near-IR spectrometry), the BEST metric is often calculated more rapidly as well. The matrix inversion required by the Mahalanobis metric is usually accomplished by algorithms whose complexity (in terms of number of operations required) increases as the number of wavelengths cubed. In contrast, the complexity of the BEST metric increases linearly with the number of wavelengths.

The assimilation model developed by the BEST can be compressed if desired with eigenvalue / eigenvector or singular value decomposition procedures to conserve memory, and can be converted into a hash table and hash function by calculating the distance and direction from the center of the calibration set to each of the replicates, and used on laboratory PC's.

Supercomputer. The supercomputer used to run the BEST is a 4 hypernode (32 processor) Convex Exemplar SPP1200 system with 7 Gigabytes of memory and 80 Gigabytes of disk storage. The unit processor in the Exemplar SPP1200/XA is Hewlett-Packard's PA-RISC 7200 processor with 240 MFLOPS peak performance. The SPP1200/XA can have up to 16 hypernodes, for a total of 128 processors, with a peak performance of 30.7 GFLOPS.

Carotid ultrasound duplex scanning. All patients were studied with carotid ultrasound duplex scanning utilizing a Hoffrel Duplex Scanner with a 7.5 MHz scanning probe. Frequency distribution, disruption in flow column and the geometry of any atherosclerotic plaque were recorded. Geometry measurements of the plaque included residual lumen of vessel, maximum thickness of plaque and percent stenosis as compared to normal carotid artery lumen. All measurements were recorded preoperatively and within 3 months postoperatively for assessment of completeness of plaque removal and evidence of carotid restenosis. Both carotid and vertebral arteries were assessed. Carotid arteries were studied in two planes. Frequency analysis was correlated to angiographic and intraoperative findings. The primary data point was the maximum thickness of intra-arterial plaque measured to within 0.1 mm at the level of the carotid bifurcation flow divider.

Carotid endarterectomy. All patients were pre-operatively evaluated for medical suitability after obtaining cardiac and medical clearance and operative consent. Carotid endarterectomy was carried out under general anesthesia. Operations were done unilaterally under EEC and continuous arterial monitoring. Selective placement of intravascular shunting was done at the discretion of the surgeon. Atherosclerotic plaque was removed with the guidance of a Zeiss operating microscope to maintain a meticulous media plane of dissection. Plaque was removed from the distal centimeter of the common carotid artery and the dissection extended distally in the internal carotid artery for the full extent of the plaque. Microscopic inspection of the remaining vessel was carried out for breaks in the wall or any evidence of a residual intimal flap or atherosclerotic plaque. Plaque was then removed and rapidly frozen for analytical and near-IR assessment. Arteries were closed with running 6.0 suture proline with microscopic placement of sutures. Ultrasound assessment of postoperative integrity was carried out during the postoperative period of study.

In vivo near-IR imaging. Two to four light sources were employed for spectrometric imaging to reduce shadows and achieve the best possible S/N. Both a near-infrared camera (InSb focal plane array or PtSi CCD) and a visible-light camera (Si CCD) were located outside the sterile field, approximately 1 m from the patient. The near-IR and visible-light cameras were operated side-by-side. Images were obtained transarterially as soon as the vessel were exposed (before shunting or opening the vessel for endarterectomy). The number of wavelengths recorded in the images was determined by available time: 8 sec of signal integration were used at each wavelength (the frame rate was 51.44 frames/sec), and spectral images were collected at each wavelength with the near-IR light sources off and again with them on to correct for the presence of other lights in the room and for blackbody photon emission from the sample. The total time allotted for imaging in a normal patient is 7-8 min. BEST contour plots were used to depict the near-IR results. Fig. 1 shows a visible-light image of the exposed carotid bifurcation in one of the patients. Fig. 2 is the corresponding near-IR contour plot. The inset white box in Fig. 2 surrounds the location of the plaque inside the artery.

Whenever near-IR cameras were used to collect spectra, two spherical silicon dioxide reflectance standards (one high reflectance and one low reflectance) were placed in each image to control for variations in light intensity and direction. Images collected on different days and with the light sources in different locations were made comparable by adjusting the gain and offset by multiplicative scatter correction on the images so the intensities on the standards were identical. The specular reflectance on the standards was used to pinpoint the locations of the light sources and to provide a means to calibrate reflected intensity from wet surfaces in tissue samples. Diffuse reflectance from the curved surfaces of the two standards was used to calibrate shaded areas and sloping surfaces in the images.

RESULTS AND DISCUSSION

Results of ultracentrifugation and gel electrophoresis measurement of extracted plaque proteins. The SDS-PAGE of LDL and oxLDL have been published (33). In the present study, extracted plaque lipoproteins were monitored on the gels from 8 to 200 kD in 2 kD increments by capturing images of the gels on a CCD camera. Fig. 3 shows a few of the significant correlations (f test on model residuals, p=0.05) between the gel data and the medical histories of the patients (n_c=29, n_v=29). Other significant correlations noted have been previously described (33). The white bars show the value of the Pearson correlation coefficient and the black bars show the percentage of the total spectral variation accounted for by the principal components showing the correlation. Coronary artery disease (CAD) and bypass grafts (CABG) have similar correlations and contributions to variation. Fig. 4a-d shows the relative contribution of the lipoproteins in the gel to the observed correlations. (These contributions were calculated by inverse principal axis transformation. The y-axis on the gels was z-scored, so the average gel appears as a flat line centered at zero through the graph. The actual average gel for the set of patients appears as Fig. 5 , and represents the content of the mean spectrum of the area inside the inset box in Fig. 2 ). The difference between the gels of plaques in patients with CAD and CABG and the gels of plaques in patients without these characteristics is the difference between the solid and the dashed lines, respectively. While the CAD and CABG patients overall had less protein than average in the 24- 200 kD range in their plaques, the plaques were slightly enhanced in their content of the proteins at 16, 28-38, and 66 kD. A smaller enrichment of the proteins at 84, 96, and 125-140 kD was also observed. The gels of patients experiencing speech problems showed a similar overall pattern, with a slightly different band region from 30-40 kD. The pattern of proteins in progression from CAD to CABG to stroke and Speech problems to Major Surgery (such as bypass grafts or previous carotid endarterectomy on the opposite side) suggests free radical / oxidation reactions that break the proteins into smaller and smaller fragments, causing an overall increase in the proteins <24 kD and an overall decrease in proteins >24 kD. The finding that proteins of mass ~130 kD (and to a lesser extent, 100 kD) correlated positively with increasing severity of disease symptoms suggests these proteins may be more-stable by-products of free- radical reactions. Alternatively, the 100 and 130 kD proteins may be specific protein involved in plaque growth or correlated disease processes. A medical history of previous major surgery was also associated with a reduction in the amount of larger proteins in the plaque and an increase in the amount of smaller proteins. The crossover point was about 45 kD, and the peak features at 66 and 130 kD were reminiscent of those in the LDL standard samples (33).

There were also strong correlations between the proteins in the plaque and examination of the plaque by pathologists. Fig. 6 shows a few of the significant correlations between the gels and pathology reports. Other significant correlations noted are given in Dempsey, 1996 (33). Microhemorrhage (see Fig. 7a ) and microulceration (33) produced the strongest correlations to protein content, and these correlations accounted for major fractions of the total variation in the gels. The most visible protein feature of microhemorrhage involved the proteins at 130 kD. Overall, the loss of proteins in the patient plaques was similar to that seen in the cases of medical histories of CAD, CABG, and major surgery.

The presence of necrosis in the plaque (see Fig. 7b ) in the report of the pathologists' microscopic examination of the plaque correlated significantly to certain proteins in the plaque. A major protein feature in necrosis was an increase in the amount of proteins with masses 38 kD and 85 kD, and from a 50 kD wide band centered at 125-130 kD. These proteins were also easily identifiable features in patients with medical histories of CAD, CABG, and major surgery. The severity of disease appeared to be a function of the concentration of these proteins in the plaque. Specifically, the peaks were larger in size in the patients in which CAD has progressed to the point that a bypass graft was necessary.

Results of Near-IR Spectrometry of Carotid Plaque. Near-IR spectra correlated significantly (f test, p=0.05, n_c=23, n_v=21) to the concentration of most of the 93 protein molecular weights monitored by gel electrophoresis (protein peaks at 20, 162, and 180 kD were the only exceptions). In other words, the appearance of a gel of extracted and centrifuged plaque lipoproteins could be calculated from nondestructive near-IR spectrometry of the whole plaque. The correlation was strongest in the proteins that produce the largest peaks in the electrophoresis. Near-IR spectra are generally far less noisy and have a much larger linear dynamic range than gelelectrophoresis (up to 6 orders of magnitude for near-IR spectrometry, compared to perhaps 2 for gel electrophoresis). In fact, gel electrophoresis is used as a calibration method for near-IR spectrometry only because it is widely employed, fairly well understood, and permits quantification of a large number of proteins in a sample simultaneously. The eventual replacement of gel techniques with near-IR ones will not only speed the process of plaque analysis but will make it more precise as well.

Quantification of proteins in plaque by near-IR spectrometry compared favorably to values obtained by ultracentrifugation and gel electrophoresis when these separation methods were used for calibration and validation. A single near-IR spectrum of each plaque was run through an assimilation model to simultaneously predict the concentration of each of the ten protein peaks selected from the gels. The calibration model (see Fig. 8 ) was designed with bias=0 so all error appears as RSD (black bars). The results were about as accurate as quantification by gel electrophoresis itself. The fact that the error did not depend upon the magnitude of the near-IR signal (white bars), combined with the fact that error was a little higher at the ends of the gel, suggests that the gel electrophoresis was the limiting error source.

Inverse principal axis transformation of near-IR spectra permitted calculation of the approximate spectrum of each protein in its natural plaque matrix. The spectrum of the 128 kD protein (see Fig. 9 ) showed increasing absorbance with increasing concentration at 1750 and 2310 nm, which suggested that the protein was a lipoprotein or had substantial lipophilic character. A similar spectral pattern was noted for the other proteins, with the exception of the 40 kD protein, which did not appear to possess lipophilic character.

CONCLUSIONS

Near-IR spectrometry has recently been used to observe stroke-induced changes in the lipids and proteins of whole brain samples in vitro and gerbil brains in vivo. In this study, near-IR spectrometric imaging was used to examine events in blood vessels that might precede ischemic brain injury in humans. Near-IR imaging methods were shown to provide spatially resolved nondestructive analysis of plaque lipoproteins relevant to atherosclerosis and stroke at least as precise as ultracentrifugation and gel electrophoresis. In the process, the analytical power of near-IR imaging cameras and MPP (massively parallel processor) supercomputing were demonstrated. These spectrometric techniques permit enormous quantities of data on the protein and lipid composition of carotid plaque to be obtained quickly with high S/N, and make the testing of hypotheses about atherogenesis and the progression of disease easy and nondestructive. The eventual extension of this technology to noninvasive transcutaneous laser measurements in patients should increase greatly our understanding of the disease process, and may even make a rational assignment of symptomatic patients to drug and/or surgical interventions possible.

REFERENCES

1. Carney, J.M. Lodder, R.A., et al., Anal. Chem., 1993, 65, 1305-1313. Back to text

2. Sempos, C. et al. J. Am. Med. Assoc., 262, 45-52, 1989. Back to text

3. Blankenhorn, D. H. et al. J. Am. Med. Assoc., 257, 3233-3240, 1987. Back to text

4. McManus, B.M.et al., J. Am. Med. Assoc., 262, 83-88, 1989. Back to text

5. Esterbauer H, et al., Chemical Research in Toxicology 3:77-92, 1990. Back to text

6. Steinberg D, et al., The New England Journal of Medicine 320:915-924, 1989. Back to text

7. Sato K, Niki E, Shimasaki H: Archives of Biochemistry and Biophysics 279:402-5, 1990. Back to text

8. Sun AY: Biochimica et Biophysica Acta 266:350-60, 1972. Back to text

9. Chisolm GM: Clin Cardiol 14:I 25-30, 1991. Back to text

10. Hessler JR, Morel DW, Lewis LJ, Chisolm GM: Arteriosclerosis 3:215-222, 1983. Back to text

11. Frei B, Stocker R, Ames BN: Proc Natl. Acad. Sci 85:9748-9752, 1988. Back to text

12. Paul J, Bai NJ, Devi CL: Internat J Vit Nutr Res 59:35-39, 1989. Back to text

13. Burton GW, Traber MG: Vitamin E: Annu Rev Nutr 10:357-82, 1990. Back to text

14. Esterbauer H, Dieber-Rotheneder M, Streigl C, Waeg G: Am J Clin Nutr 53:314S-21S, 1991. Back to text

15. Chilsolm III GM, Morel DW: Am J Cardiol 62:20B-26B, 1988. Back to text

16. Steinberg D, Parthasarathy S, Carew TE: Am J Cardiol 62:6B-12B, 1988. Back to text

17. Carew TE, Schwenke DC, Steinberg D: Proc NatI Acad Sci 84:7725-7729, 1987. Back to text

18. Bassiouny HS, Davis H, Massawa N, Gewertz BL, Glagov S, Zarins CK: Vasc Surg 9:202-12, 1989. Back to text

19. The Dutch TIA Trial Study Group: N Eng J Med 325:1261-6, 1991. Back to text

20. North American Symptomatic Carotid Endarterectomy Trial Collaborators: N Eng J Med 325:445-53, 1991. Back to text

21. Harman D: In Free Radicals in Molecular Biology, Aging, and Disease, D.Armstrong et al. (ed.) Raven Press, New York, 1984, 1-41. Back to text

22. Mayberg MR, Wilson SE, Yatsu F, Weiss DC, Messina L, Hershey LA, Colling C, Eskridge 3, Deykin D, Winn HR: JAMA 266:3289-3294, 1991. Back to text

23. Dempsey RJ, Diana AL, Moore R: Neurosurgery 27:343-348, 1990. Back to text

24. Ginsberg, H. N. et al. New England Journal of Medicine, 322, 574-579, 1990. Back to text

25. Jorgenson, J. W. Anal. Chem., 58, 743A-760A, 1986. Back to text

26. Wetzel, D. Anal Chem 55, 1165A, 1983. Back to text

27. Cassis, L. A. and Lodder, R. A. Anal. Chem., 1993, 65, 1247-1256. Back to text

28. Wold, JP, Krane, J, Krane, L J. Food. Sci. 61(1), 74, 1996. Back to text

29. Chung, H, Arnold, MA, Rhiel, M, Murhammer, DW. Appl Spectrosc 1996, 50, 270-276. Back to text

30. Lewis, EN, Gorbach, AM, Marcott, C, Levin, IW. Appl Spectrosc 1996, 50, 263-269. Back to text

31. Conway, J., Norris, K. and Bodwell, C. Amer. J. Clin. Nutr. 40:1123, 1984. Back to text

32. Ozaki, Y., Matsunage, T. and Miura, T. Applied Spect. 46:180-182, 1992. Back to text

33. Dempsey, R. J. Davis, D. G. Buice, R. G. Jr. and Lodder, R. A. Appl. Spectrosc. 1996, 50, 18A-34A. Back to text

34. VandeVan, M., French, T., Fishkin, J. and Gratton, F. Biophys. J. 59:167a, 1991. Back to text

35. Havel LJ, Eder HA, Brogdon JH: J. Clin. Invest. 34: 1345-1353, 1955. Back to text

36. Yla-Herttuala, S., Palinski, W. Rosenfeld, M.E. J. Clin. Invest., 84:1086-1095, 1989. Back to text

37. Lodder, R.A.; Hieftje, G.M. Appl. Spectrosc., 1988, 42, 1351-1365. Back to text

CREDIT

The clinical research was supported by funds from the NIH under grants HL45143 and RR08242 and the American Heart Association (Kentucky Affiliate), and the computational research by funds from the NSF under grant CHE 9257998.

Return to Wave of the Future Table of Contents

Return to Cover of Wave of the Future