NMR method for monitoring changes in the core of lipoprotein particles in metabolism and disease

Abstract

A method is disclosed for measuring the properties of protein and lipoprotein elements in a sample. The method includes the of placing a small volume of a sample into a NMR instrument tuned to measure a particular nucleus, applying a series of radiofrequency pulses with intermittent delays in order to measure spin-spin and/or spin-lattice relaxation time constants from the time-domain decay of the signal, without the use of chemical shifts and without converting data into the frequency domain by Fourier transform or other means, at least partially suppressing the water signal prior to the beginning of a sequence used to record relaxation time constants in the time domain, optionally utilizing relaxation contrast agents or other chemical additives to perturb the solvent water or other elements of the sample, analyzing the exponentially decaying NMR signal in the time domain using multi-exponential analysis, and comparing differences in the relaxation time constants for lipoprotein- or protein-specific elements within a single human subject, or between subjects, to assess normal and abnormal metabolism reflective of increased disease risk or active disease.

Claims

1. A method for assessing cardiovascular disease risk based upon the properties of protein and lipoprotein elements in a sample, the method comprising the steps of: conducting a low-speed centrifugation of a human blood sample obtained after clotting; placing a small volume of the unfractionated human serum sample into a low field TD-NMR spectrometer tuned to measure hydrogen-1; applying a series of CPMG spin echo pulse sequences with intermittent delays to the sample in order to measure spin-spin relaxation time constants from the time-domain decay of the signal, without the use of chemical shifts and without converting data into the frequency domain by Fourier transform or other means; at least partially suppressing the water signal from the sample by inserting a 180-degree inversion pulse followed by a delay prior to the CMPG sequence used to record relaxation time constants in the time domain; analyzing the exponentially decaying NMR signal from the sample in the time domain using multi-exponential analysis; and comparing differences in the relaxation time constants for lipoprotein classes selected from the group consisting of LDL, HDL, and VLDL in the sample to measure mobility differences in the oil phases within the core compartment of the selected lipoprotein particles which are in turn reflective of increased cardiovascular disease risk or active disease and in which the comparison of relaxation time constants is made to other relaxation time constants within a group consisting of samples from the same human subject and samples among different human subjects.

2. A method according to claim 1 comprising using between about 0.2 and 0.6 mL of the sample.

3. A method according to claim 1 further comprising normalizing the viscosity of the sample.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a very general schematic diagram of an LDL particle and its internal oil phase.

(2) FIG. 2 is a TD-NMR T.sub.2 profile for purified human low-density lipoprotein at 25 C.

(3) FIG. 3 is a plot of metabolic remodeling of LDL core lipids during metabolism following a meal.

(4) FIGS. 4-6 are CONTIN profiles for pure triolein and for two lipid cores.

(5) FIG. 7 is a plot of T.sub.2V values for albumin, LDL plus albumin, and VLDL plus albumin.

(6) FIG. 8 is a plot of T.sub.2V values for three samples of whole human serum.

(7) FIG. 9 is a plot of T.sub.2V values for lipoprotein particles at various times after a meal.

DETAILED DESCRIPTION

(8) The methods of the invention resolve individual lipoprotein particle classes (e.g., low-density lipoprotein or LDL) by detecting differences in core lipid mobility, which is influenced by the relative amount of cholesteryl ester to triglyceride molecules within each particle's core. Variability in core mobility and core composition within a particle class, such as LDL, can result from patient-to-patient differences, or from particle remodeling within an individual subject as occurs during metabolism following a meal. Changes in lipoprotein particle core mobility and core composition are monitored using a time-domain nuclear magnetic resonance (TD-NMR) analysis. A hallmark of this approach is that the analysis is performed without Fourier transformation and without the use of frequency-domain information such as chemical shifts. Unlike frequency-domain Fourier-transform NMR, this time-domain NMR analysis can be performed at low magnetic fields (60 MHz for hydrogen) in a low-cost, bench-top instrument configuration, although it can also be performed in conventional high-field NMR spectrometers.

(9) The general principles of time domain pulse NMR are generally well understood and familiar to persons of ordinary skill in the art and need not be discussed in detail. In brief, however, a sample is positioned in an external magnetic field provided by a permanent magnet. This aligns the magnetic moments of the hydrogen atoms with (or against) the permanent magnetic field. Then, a radio frequency pulse is applied in a direction that provides a secondary (temporary) magnetic field perpendicular to the permanent magnetic field. This moves the magnetic moments of the hydrogen atoms away from their equilibrium state. The time duration of the pulse determines how far the magnetic moments move. The combined movement of many spins (many hydrogen atoms) generates a small but detectable oscillating magnetic field that in turn induces an alternating voltage that is measured as the NMR signal by a detection coil.

(10) At the end of the pulse, the protons in the sample give up excess energy to their surroundings and relax back to the equilibrium state with respect to the permanent magnetic field. This relaxation takes a certain amount of time, so that the NMR signal remains detectable for a period of time that can range from several milliseconds to several seconds.

(11) Furthermore, the relaxing component of the NMR signal will be characteristic of individual mobility domains, which in turn, help identify the molecules involved in the motions. For example, cholesterol molecules are more internally rigid than triglyceride molecules and will tend to give lower T.sub.2 and T.sub.1 values.

(12) Additionally, the data resolution of the pulse time domain NMR technique of the invention is on the order of particle size. In comparison, Fourier transfer NMR resolves data on an atomic scale. As a result, the time domain technique makes fewer technical demands (so to speak) on the instrument and can provide useful data at the available resolution.

(13) According to the invention, it is been determined that time decay constants are sensitive to both particle size and particle mobility.

(14) The method is also tolerant of multiple phases or mixed phases; i.e., solids and liquids in many circumstances.

(15) As part of the correlation discoveries of the invention, it is now been determined that LDL particles with a higher triglyceride/cholesterol molecular ratio in the core have a longer spin-spin relaxation time (T.sub.2) and particles with a lower triglyceride/cholesterol ratio have a shorter T.sub.2.

(16) Although the inventors do not wish to be bound by a particular theory, it appears that this may result from the characteristics of an LDL particle as not being solid in the same sense as a solid homogeneous composition would be. Instead, the LDL particle has an internal oil phase (FIG. 1). In turn, the oil phase moves (tumbles) differentlyand typically fasterthan the remainder of the particle. This faster internal tumbling increases the spin-spin relaxation time.

(17) In one embodiment, the hydrogen spin-spin relaxation rate constants (or time constants) are measured using a low-field bench-top time-domain NMR analyzer, and the relaxation rate constants for individual lipoprotein classes are resolved through a multi-exponential deconvolution algorithm. Another key feature of this analysis is that measurements can be made directly on intact body fluids (e.g., serum, plasma or blood) without the need for separation or fractionation of individual lipoprotein classes by ultracentrifugation, electrophoresis, chromatography or other time-consuming, sample-perturbing methods. Because of the relative simplicity and low cost, this method has potential application to clinical testing for the detection of unique dyslipidemias and for the early detection and risk assessment of cardiovascular disease, diabetes and cancer.

(18) The measurements can, of course, be made in conventional high-field NMR spectrometers, but as set forth herein, the use of Benchtop instruments offers a number of clinical advantages.

(19) In the invention, time-domain NMR resolves individual lipoprotein classes by measuring mobility differences in the oil phases within the core compartment of lipoprotein particles. The invention is also based on the discovery that TD-NMR is sensitive to changes in the particle core within a lipoprotein class. For example, the LDL particles in diabetic subjects tend to be richer in triglyceride, which makes the particle core more mobile.

(20) The mobility differences are monitored by measuring relaxation rate constants (or time constants) without chemical shifts. Chemical shifts are the centerpiece of conventional high-field, frequency-domain NMR. By contrast, time-domain NMR does not require chemical shifts for frequency domain resolution and does not require high magnetic field strength or field homogeneity. This approach is fundamentally different from conventional NMR spectroscopy in both methodology and instrumentation requirements.

(21) In one embodiment, the invention is a process for measuring the spin-spin or transverse relaxation time constants (T.sub.2) for the lipid core compartments in unfractionated human serum. The human serum is obtained in a conventional manner from a low-speed centrifugation of human blood after clotting. Approximately 0.6 mL of unmodified serum is pipetted into a 10 mm NMR tube, and the tube is placed into the bore of the magnet of a bench-top TD-NMR analyzer, typically operating at 10, 20, 40 or 60 MHz resonance frequency for hydrogen. In the examples described here, 20 MHz and 40 MHz data were collected using Bruker benchtop mq20 and mq40 TD-NMR instruments (Bruker BioSpin Corporation, Billerica, Mass., USA).

(22) A Carr-Purcell-Meiboom-Gill (CPMG) pulse sequence is used to measure the exponential time-decay curve. This pulse sequence effectively eliminates chemical shifts and magnetic field inhomogeneity, permitting the measurement of T.sub.2 values. Although T.sub.2 measurements can be linked with chemical shifts and measured in the frequency domain, the present TD-NMR method measures T.sub.2 in the time domain without chemical shifts. This provides a distinct advantage with respect to instrument simplicity and cost.

(23) The resulting T.sub.2 decay curve for human serum is multi-exponential, so the individual exponential terms are deconvoluted and resolved with the use of an inverse Laplacian transform. An implementation of this mathematical calculation is provided in the public domain software CONTIN, authored by Steven Provencher (http://s-provencher.com/pages/contin.shtml; accessed Mar. 11, 2013). Under the proper experimental conditions with excellent signal-to-noise, the CONTIN calculation can resolve up to 8 different exponential terms in TD-NMR T2 profiles of human serum. Because human serum has abundant quantities of lipoprotein core lipids and soluble proteins, and because these assemblies are relatively large, the protein and lipoprotein components dominate the T.sub.2 profile.

(24) One experimental issue involves solvent suppression, because an intense water signal can overshadow the contributions of lipoprotein components and lead to artifacts such as radiation damping. The solvent water can be partially suppressed using a number of NMR schemes. In this embodiment, a 180-degree pulse and delay is inserted prior to the CPMG sequence. This achieves partial relaxation (and partial suppression) of the water with full recovery of the lipoprotein components by the start of the CPMG pulse scheme. Although there are many sophisticated NMR methods for suppressing water, the goal of this invention was to develop the simplest, most inexpensive method for measuring lipoprotein core properties in unmodified human serum.

(25) This embodiment is further illustrated using the figures and tables. FIG. 2 shows a time-domain T.sub.2 NMR profile of normal fasting human serum, and FIG. 3 is the TD-NMR T.sub.2 profile for purified human low-density lipoprotein at 25 C.

(26) The profile of FIG. 2 (not to be confused with a conventional NMR spectrum) was obtained by performing an inverse Laplacian transform of the T.sub.2 decay curve. In turn, the T.sub.2 curve was measured using a modified CPMG experiment on a Bruker 40 MHz TD-NMR analyzer. The profile resolves 7 distinct T.sub.2 components, here represented as peaks in the profile. Higher T.sub.2 values represent more mobile elements. The solvent water peak in serum, not shown in this plot, is observed at T.sub.2 values of approximately 600 ms. The peaks at approximately 200 and 70 ms represent two distinct mobility domains in LDL, as assigned from control samples containing only LDL.

(27) These two peaks are not observed in the other control samples containing fractionated serum proteins or lipoproteins and appear to provide a unique signature for the core lipid mobility of LDL. The peaks at lower T.sub.2 values have contributions from both serum proteins and lipoproteins.

(28) FIG. 9 illustrates the variation in T.sub.2 measurements in response to metabolic changes following a meal. Shown are T.sub.2 profiles for whole human serum at fasting, 2 hours, 4 hours and 6 hours post-prandial. The T.sub.2 values for the LDL-specific peaks increase and peak at 4 hours, reflecting the remodeling of the LDL core as the content of triglyceride increases relative to cholesteryl ester.

(29) These preliminary results demonstrate the feasibility of obtaining particle-specific measurements of the core mobility of LDL in whole human serum. The data also demonstrate that T.sub.2 measurements obtained from TD-NMR are sensitive to metabolic remodeling and patient-to-patient variability.

(30) Furthermore, the invention requires neither high magnetic field instrumentation nor a frequency-domain analysis. Instead, it uses a time-domain analysis. Unlike (for example) the Otvos approach, the methods of the invention can be performed on inexpensive low-field bench-top instruments, because high field strength and field homogeneity is not required. The key measurables are relaxation rate constants rather than chemical shifts. Differences in relaxation rate constants are used to resolve lipoprotein classes (not chemical shifts, as in Otvos and Kremer). Also, the derived parameter in our invention is lipoprotein particle core mobility or fluidity, rather than particle number or particle size. In summary, the instrumentation, data processing, measurables and derived parameters of our invention are different from those of Otvos.

(31) In contrast to (for example) U.S. Pat. No. 7,550,971, the present invention does not measure analyte concentrations. Rather it measures lipoprotein particle properties, specifically the mobility or fluidity (squishiness) of the oily lipid core found within lipoprotein particles. Also, the invention is not restricted to low-field, bench-top NMR instruments, but can also be performed on conventional high-resolution NMR instruments as well.

(32) Overall, the method of the invention is much simpler and can be performed on inexpensive low-field benchtop NMR analyzers, and the particle core mobility provides diagnostic information different from particle size and concentration distribution.

ADDITIONAL EXAMPLES

(33) In the following Examples, all aqueous samples are prepared in a 9.1 D2O/H2O saline buffer, concentrated to a viscosity of 1.20 cP at 37 C. The raw data are in the form of a multi exponential decay curve. The individual relaxation time constants are deconvoluted using an inverse Laplacian transform calculation as implemented in the public domain program CONTIN.

(34) Lipoprotein Lipid Core Mixtures

(35) A CONTIN profile of triolein, the most abundant TG in lipoproteins, is shown in FIG. 4. Each T.sub.2 (or T.sub.2V) corresponds to a mobility domain of the triglyceride molecule. A VLDL like lipid core with 80% TG and 20% CE shows a similar profile (FIG. 5) but is shifted slightly to the left, indicating a less mobile, more vicious environment. The LDL like lipid core (FIG. 6) composed of 80% CE and 20% TG exhibited lower T.sub.2 or T.sub.2V values indicating reduced mobility. This trend resembles that observed with physiological lipoprotein particles with differences in TG/CE ratios. T.sub.2V values for lipid mixtures are summarized in Table 1.

(36) TABLE-US-00001 TABLE 1 (T.sub.2V times in ms) Fast Medium Slow Other 100% TO 347 155 79 8 80% to 20% CL 303 141 70 25, 5 (VLDL Core) 20% to 80% CL 183 92 45 19, 9, 4 (LDL Core)

(37) Fractionated Lipoproteins and Serum Proteins

(38) FIG. 7 shows the T.sub.2V profiles for albumin (HAS), LDL plus albumin and VLDL plus albumin. Although all of the profiles display fast decaying components in the range of 2-50 ms, there are T.sub.2V values above 50 ms that are unique to individual lipoprotein classes

(39) Whole Human Serum

(40) De-identified samples of whole human serum, representing various metabolic and disease states were obtained from Pitt County Memorial Hospital (Greenville, N.C., USA). As seen in FIG. 8 variability was observed in lipoprotein T.sub.2V values. Patient samples with a higher HbA1c, indicative of poorly managed Type 2 Diabetes Mellitus, have increased T.sub.2 or T.sub.2V values suggesting that the lipoprotein particles have increasingly mobile, TG-rich lipid cores versus non-diabetic patients.

(41) Results: FIG. 9 shows the remodeling healthy non-diabetic subject ingested a liquid meal that contained 50 grams of lipid following a 16 hour fast, after which blood was drawn every hour for 8 hours. The NMR data are shown in FIG. 4b at 0 hour (fasting) and other time points after the meal. As lipoprotein remodeling occurs and the particles become TG rich, the T.sub.2V peak shifts to the right indicating an increase in lipid core mobility. Table 2 shows standard lipid analysis for these samples.

(42) TABLE-US-00002 TABLE 2 0 Hr. 2 Hr. 3 Hr. 4 Hr. 5 Hr. 6 Hr. 7 Hr. 8 Hr. Triglycerides 87 184 181 194 197 231 188 169 Cholesterol 197 204 200 199 193 187 184 187 HDL-C 54 55 52 52 49 47 47 48 Non-HDL-C 148 149 148 147 144 140 137 139 LDL-C 125 112 112 108 105 94 99 105

(43) Benchtop TD-NMR appears to provide unique information about LDL and VLDL particle properties reflective of different states of normal and abnormal metabolism. This approach holds promise for translation from the research lab into the clinical setting as the measurements are performed on whole human serum and are relatively simple, inexpensive and non-invasive.

(44) In the drawings and specification there has been set forth a preferred embodiment of the invention, and although specific terms have been employed, they are used in a generic and descriptive sense only and not for purposes of limitation, the scope of the invention being defined in the claims.