Quantitative measurement method by isotope dilution mass spectrometry

Abstract

In isotope dilution mass spectrometry (IDMS), a corrected signal ratio R' instead of measured signal ratio R is in theory proportional to analyte concentration in a study sample. A reverse IDMS process is used to characterize an analog such as stable isotope labelled (SIL) to obtain an analyte equivalent concentration. With another two accurately predetermined parameters, analyte concentration in an unknown sample can be determined simply following IDMS procedure without the burdens of a calibration curve and matching sample matrix.

Description

[0001] TITLE: Quantitative Measurement Method by Isotope dilution mass spectrometry

[0002] INVENTOR: Bingfang Yue

[0003] Assignee: Disruptive Laboratory Innovations, LLC

[0004] Cross Reference to Related Applications:

[0005] This application claims benefit of and priority to U.S. Provisional Patent Application No. 63 / 572,414, filed April 01, 2024, where permissible incorporated by reference in its entirety.

[0006] FIELD OF THE INVENTION

[0007] The present invention relates generally to mass spectrometry detection of analytes, in particular, to isotope dilution mass spectrometry methodology and system, including quantitative measurement, thereof. Accordingly, this invention involves the field of analytical chemistry as well as related fields.

[0008] BACKGROUND

[0009] Mass spectrometry (MS) has been an essential instrumental setup for a wide range of applications, especially when coupled to gas chromatography (GC-MS) and liquid chromatography (LC-MS). Due to its lower limits of quantification and high selectivity, MS detection has become the gold standard for small molecule analysis, of either endogenous or exogenous origin. Also, the hyphenation of GC or LC to MS improves sample analysis in biological matrices, allowing analyte and matrix separation prior to detection.

[0010] Classically, the calibration strategy is adopted as an intrinsic part of the quantification process. Currently, a multipoint calibration curve is the most widely used strategy to estimate the calibration function in MS analysis. During this process, a set of calibration standards are analyzed to establish an empirical concentration-response function over a desired linear dynamic range. Usually, a model is fitted using the least-squares regression technique and the specimen concentrations are back-calculated from the model.

[0011] This classical calibration strategy encounters various technical difficulties in practice. For example, with endogenous substances in complex biological matrices, the difficulty in obtaining a true blank matrix (i.e., a matrix free of the analyte) hampers a reliable quantification in terms of accuracy and precision. A second example, there is no consensus with regards to selecting an empirical concentration-response model (linear vs. nonlinear, weighting factor of least square regression, the number of calibration points, etc.). Another example, comparison studies between measurement systems often show calibration-related discrepancies. Even more importantly, in clinical laboratory settings, this classical strategy makes fast turnaround time impossible to achieve considering the method run time and the number of calibrators and quality controls to be analyzed before generating results for a patient sample.

[0012] In MS analysis, the measured response for a given analyte can strongly vary. Thus, the analyte relative response must be normalized to compare performance over time or across different instruments. An internal standard (IS) is commonly added to the study and calibration samples at fixed concentrations. The absolute response of the analyte is then normalized by dividing the measured IS signal from the same sample, as a response ratio (R) that reduces the analytical variability.

[0013] There are two primary classes of IS, isotope and structural analogs. The first class is related to the stable isotope (for example1H,13C,15N) labelled form of the analyte of interest (SIL). The second IS class includes compounds that share structural and / or physicochemical similarities with the analyte of interest. For quantification purposes, the use of one IS per target analyte is generally recommended because they are assumed to compensate for specific differences in matrix effect during MS analysis and extraction recovery between the calibration and study samples during analytical workflow.

[0014] Among multi-point calibration strategies, the nature of the analyte, the availability of the study sample matrix and the reference material, define which type of analytical calibration is adopted: surrogate vs. authentic matrix, surrogate vs. authentic analyte, while a surrogate analyte is often a SIL analog. Because of SIL, in recent years, internal calibration (IC) is gaining momentum quickly. IC is performed directly on a study sample, in which one or multiple SIL is used as surrogate analyte, signals from one or multiple isotopologues are used to prepare a calibration curve, and the analyte concentration is then back calculated from the calibration curve. Obviously, it is assumed that same mole quantities of isotopologues originated from either SIL analog or authentic analyte should produce same signal intensities in mass spectrometry (however, this is not universally true). This is also the basis for using a response factor (RF) for IC: RF is stable over the studied concentration range and the sample concentration is directly calculated via the analyte / SIL signal ratio with predetermined RF.

[0015] SUMMARY

[0016] Quantitative mass spectrometry often makes use of stable isotope labelled (SIL) internal standards (IS). Quantitation is based on the ratio between analyte and SIL-IS signals. Internal standards are used to correct for variations in sample preparation, injection, ionization, instrument performance, and the like. Although seldom explicitly stated, certain assumptions are often made about calibration curves, in particular, that the analyte does not interfere with the IS, and the IS does not interfere with the analyte. Although this is often a reasonable assumption, it is seldom strictly true.

[0017] The present disclosure is drawn to the use of a linear relationship between a corrected signal ratio R' = (R-Rs) / (1-R / Ra) instead of the measured signal ratio R and the analyte / IS concentration ratio to represent the true form of the calibration curve. The R' is derived from generic procedure of isotope dilution mass spectrometry (IDMS) and is thus applicable in all IDMS cases, including quantitative mass spectrometry with a SIL-IS. The parameter Raand Rsare purely related to isotopic composition of the analyte and SIL-IS respectively, can be easily determined experimentally. In ideal situations, Ra=00and Rs= 0, it turns out that R' = R and the calibration curve is linear with R. This linearity is assumed in various conventional calibration schemes that use the ratio R instead of R' as the basis of quantitative analysis.

[0018] When Rais relatively small (say <500), meaning an analyte isotope interferes with SIL-IS, the calibration curve with R is nonlinear in nature. Conventionally, nonlinear regression (often quadratic) or linear regression with 1 / x or 1 / x2weighting factor with R is used to fit the calibration curve, usually resulting in limited linear dynamic range. In this case, it is beneficial to increase the quantity of SIL IS to force the value of R to lower values where the nonlinearity resulting from Rais not as extreme. A positive value of Rsequates to a nonzero value for the y intercept in a calibration plot. To mitigate its impact on measuring low concentration level, reduced quantity of SIL-IS is often used.

[0019] In one embodiment of the present invention, a linear calibration curve with R' can be used in quantitative mass spectrometry with a SIL-IS. Such a process allows for more accurate quantitation, large linear dynamic range and is less demanding on SIL-IS quality. As a result, less expensive SIL-IS can be utilized, such as lesser deuterium atom labelling that ensures chromatographic coelution of analyte and SIL-IS. Theoretically, SIL-IS quantity is not relevant to achievable linear dynamic range anymore and has no impact on quantitative accuracy, particularly at low analyte concentration levels.

[0020] In further embodiments, quantitative mass spectrometry can be performed by IDMS approach with R', similar to internal calibrations, such as in-sample calibration curves and isotope pattern deconvolution. In this case, the SIL-IS is characterized using a reference standard solution (certified concentration level Creference, prepared with reference material of highest purity) in a process called reverse IDMS, in which a value for Ceq= Creference / R' is obtained as the analyte equivalent concentration of the SIL-IS (dependent on mole amount used). The analyte concentration Canaiyte in an unknown sample can then be obtained by an IDMS process: Canaiyte = Ceq ■ R'. In this way, accurate quantitative results can be obtained by one analysis without the troubles of a calibration curve and matching sample matrix.

[0021] Additionally, embodiments of the present invention allow for an integrated sample preparation and analysis system capable of random-access, which means the next quantitative analysis can be performed for any analyte in any sample. The system comprises of a sample preparation subsystem and a sample analysis subsystem. Each sample can be processed by the sample preparation subsystem following IDMS protocol: preparing a blend of sample and SIL-IS and performing any necessary clean up; the sample analysis subsystem analyzing the extracted sample according to the IDMS protocol and determine the analyte concentration in the sample with Canalyte — Ceq ’ R' utilizing predetermined Ra, Rsand Ceqvalues. BRIEF DESCRIPTION OF DRAWINGS

[0022] Fig. 1. Depiction of generic isotope dilution mass spectrometry (IDMS) principle.

[0023] Fig. 2. Simulated mass spectrum illustrating isotopic effect using testosterone as an example.

[0024] Fig.3. Plot illustrating regression model effect - data simulated without inclusion of error component.

[0025] DETAILED DESCRIPTION

[0026] It is critical to explain the principle of isotope dilution mass spectrometry (IDMS) to fully understand how various calibration schemes work. In IDMS, a sample with known isotopic composition of the analyte A, but unknown concentration is mixed with a known amount of a spike. The spike contains the analyte in a different isotopic composition S (for example, stable isotope labelled or enriched (SIL) form, or a structural analog, of the analyte). Later subscripts "a" and "s" are used to indicate the analyte A in the sample and the analyte S in the spike. After complete mixing of sample and spike, the so-called sample-spike blend or isotope diluted sample gained a new isotope ratio being between the isotope ratio of the sample and that of the spike. This blend isotope ratio directly reflects the analyte concentration in the sample.

[0027] Note that the analyte A in the sample and the analyte S in the spike are the same chemical species (molecular formula) and the only differences between the analyte A and S are the differences in terms of isotopic composition (the more different, the better), any other physicochemical properties between the two are the same (exceptions exist, such as analyte labelled with a large number of deuteriums with regards to reverse phase chromatographic retention).

[0028] As shown in Fig.l, there are two signal channels 1 (Chi) and 2 (Ch2) of the analyte, corresponding to two different isotopologues (for example, mass to charge ratio, or multiple reaction monitoring (MRM) transition). Let yl and y2 represent the relative signal intensities on channel 1 and 2 respectively for the analyte A in the sample, y3 and y4 for analyte S in the spike, and y5 and y6 for both analyte A and S in the blend, on the basis of mole quantities (in unit of moles). It is easy to see that: y5 = yl x Na+ y3 x Nsand y6 = y2 x Na+ y4 x Nswhere Naand Nsare the respective mole quantities of the analyte A in the sample and the analyte S in the spike. Naand Nsare defined in Fig. 1 together with R, Raand Rs. After defining another constant parameter p as equal to yl / y4, it is easy to arrive at the following equation:

[0029] The constant parameter p equals to the corrected signal intensity ratio with equal mole amounts of the analyte A and S, resulting from the different isotopic composition of the analyte A and S. Clearly, the corrected signal intensity ratio R' as defined above is proportional to the ratio of the mole quantities of the analyte in the sample and the spike.

[0030] Values for both Raand Rscan be determined experimentally by measuring signal intensities of Chi and Ch2 from a standard solution containing either analyte A or S (the solution containing analyte S is called the spike). For analyte A, as indicated in Fig. 1, Ra= yl / y2, while for analyte S, Rs= y3 / y4. To ensure the accuracy of Raand Rs, the concentration of standard solution containing analyte A or S should be sufficiently high to achieve good signal to noise ratios for yl, y2, y3 or y4. The measurements of Raand Rsshould be repeated multiple times, say 20 or 100 times, to achieve high accuracy. When Rais infinitely large, it means that the analyte A does not generate any signal on Ch2; when Rsequals to 0, it means that the analyte S does not generate any signal on Chi. When Rais infinitely large and Rsequals to 0, it turns out that R' = R and this linearity is assumed in various calibration schemes that use the ratio R instead of R' as the basis of quantitative analysis (more details below).

[0031] To determine the constant parameter p for analyte A and S, a reference solution (certified concentration level, prepared with reference material of highest purity) is blended with the spike containing analyte S. This blend is measured same as the experiments used to determine Raor Rs, now R = y5 / y6. The measurements of R should be repeated multiple times, say 20 or 100, to achieve high accuracy of R. R' can then be calculated from measured R, Raand Rs. With R'and known Naand Ns, p can be determined by the equation below on the left.

[0032] In theory, Nscan be calculated with the spike concentration (e.g., mole per Liter or gram per mil li-Liter) or mass fraction, used volume (e.g., Liter or mL) or mass amount, and analyte S molecular weight. In practice, the spike mole quantity of analyte S used is difficult to determine, due to incomplete information (see below). We can define another constant parameter Neq= Ns / p, which is Nsdependent, and the equation above on the left can be simplified and illustrated as in the right. Following the above process, the spike is usually characterized using a reference standard solution and this process is called reverse IDMS, compared to IDMS.

[0033] To determine analyte A mole quantity in an unknown sample, normal IDMS procedure is followed, in which a sample is mixed with a spike and R is measured, the mole quantity of analyte A is calculated by the following equation:

[0034] From the Naand the sample volume used, the sample molar concentration can be calculated, for example in unit of mole per Liter; From the Na, molecular weight of analyte A and the sample volume used, the sample mass concentration can be calculated, for example in unit of gram per Liter; From the Na, molecular weight of analyte A and the sample mass used, the sample mass fraction can be calculated.

[0035] As mentioned above, in practice, the spike mole quantity of analyte S used is difficult to determine, due to incomplete information. The first factor is isotope effect, as illustrated in Fig. 2, in which isotope abundances of testosterone as well as13C labelled (13Cig) and deuterium labelled (Dzs) are shown. With all12C replaced by13C or H by D in testosterone, this example is an extreme case to show how isotope abundance is affected by stable isotope labelling. In Fig. 2, it is obvious that the monoisotope of testosterone-13Cig at 307 has much higher abundance compared to testosterone at 288 (99.1% vs 80.8%). Even worse is the isotope purity, which may include various amounts of13Cis,13Ci613Ci?, and13Cis in case of13Ci9-testosterone. In most cases, accurate chemical purity is not available either. Thus, it is impossible to accurately determine the spike mole quantity and reverse IDMS is preferred to characterize the spike and obtain the analyte A equivalent mole quantities for the analyte S, called Neq. These reasonings also explain that stable isotope labelled (SIL) analyte is often used as internal standard for normalization purposes since same amounts (volume or mass) of same SIL-IS solution are added into calibrators, quality controls and study samples, without concerning its accurate concentration. In practice, the signal ratio can be chromatographic peak area or height ratio, while the concentration or mass fraction or their ratio is often used instead of the mole quantity ratio since they are directly proportional to one another in this case.

[0036] As IDMS only requires signal ratio determination, the advantages compared to other methods become apparent: After equilibration, losses of analyte do not affect the accuracy of the analytical result, because the measurand, the blend ratio, is equal in all sub-samples of the blend. Thus, extended sample preparation schemes can be carried out to mitigate matrix effect as much as possible. The fact that IDMS has the potential to produce analytical results of highest accuracy and precision, metrologically denoted to as smallest uncertainties, makes it most suitable as reference method for reference material characterization.

[0037] Now we discuss the multi-point calibration schemes often adopted so far for quantitative analysis. As clearly pointed out above, all adopted schemes hypothesize that the signal ratio R is proportional to the mole quantity ratio, instead of the corrected signal ratio R' . This assumption of linearity is only valid when Rais infinitely large and Rsequals to 0. Fig. 3 illustrates different calibration schemes in case of extreme mole quantity ratios from 0.001 to 1000. Curves are simulated for an imaginary analyte with Ra= 250, Rs= 0 and p = 1, without inclusion of any error components. It is apparent that despite the theoretical linear curve of y = x with R', quadratic fit with R leads to false nonlinear curve, while the linear fit of R with 1 / x2weighting factor leads to low value of slope and low acceptable linear dynamic range. This explain why the linear and quadratic fit with a 1 / x or 1 / x2weighting factor are most common in literatures. In this case, it is beneficial to increase the quantity of SIL IS to force the value of R to lower values where the nonlinearity resulting from Rais not as extreme. A positive value of Rsequates to a nonzero value for the y intercept in a calibration plot. To mitigate its impact on measuring low concentration level, reduced quantity of SIL-IS is often used. In contrast to multi-point calibration schemes, the internal calibration (IC) approach is distinguished by an analytical calibration curve obtained directly in the study sample to simplify the quantification process and to overcome challenges such as matrix effects and instrument signal drift encountered with conventional multi-point calibration practices. Isotope pattern deconvolution (IPD) is essentially an isotope dilution process using an SIL analog to alter isotopic pattern, thereafter linear deconvolution is performed with multiple signal channels corresponding to isotopic pattern to determine the authentic analyte concentration.

[0038] More general, one or more SIL analogs are used in IC to obtain a calibration curve. Multiple Isotopologue of different abundances are used to provide multiple calibration points. Value assignments for different isotopologues are performed the same way as reverse IDMS in some cases while theoretical values are used in other cases. When one SIL analog is used and no interferences between the analyte and SIL are present (Ra=00and Rs= 0), the analyte-to-SIL response factor (RF) approach can be adopted because R is proportional to the mole quantity ratio as discussed above.

[0039] In an "exact matching" form of IDMS, sample and standard can be prepared to contain equimolar amounts of an analyte and its SIL analog. Ideally, this results in the ratio measurements for both sample and standard being indistinguishable in a mass spectrometer. As the accurate analyte concentration in the standard is known, the analyte concentration in the sample can be easily calculated from its signal ratio. It is believed here that the so called "equimolar amounts" are used to more accurately account for constant parameter p defined above and the imperfect SIL analog used. This method also requires a preliminary analysis of the sample by another method to estimate the concentration of the analyte in the sample.

[0040] As shown above, the linear relationship between the corrected signal ratio 7?'and the quantity of the analyte in the sample (concentration, mass fraction, or ratio) can be used to explain various calibration schemes in practice. A few assumptions are made and believed to be true: SIL analog (analyte S) has the same physicochemical properties as the analyte (analyte A);different isotopologues of SIL and the analyte may have same or different response factors in mass spectrometry (the parameter p does not equal to 1 all the time ); and SIL has very different isotopic composition from the analyte. Reverse IDMS is used to account for imperfections in SIL such as chemical purity, isotopic purity as well as isotopic effect. It is also recognized that ionization suppression may reduce the signal intensity to unacceptable degree and detection saturation may lead to nonlinearity at extremely high signal intensity. Thus, with accurate and predetermined values for Ra, Rs, and Neq, analyte concentration in an unknown sample can be directly determined following IDMS procedure disclosed in this invention, without the burdens of calibration curve or authentic sample matrix. This approach will make a random- access analytical system possible and leads to shortest run time for a quantitative sample analysis.

Claims

CLAIMSWhat is claimed is:

1. A method for measuring a sample quantitively by isotope dilution mass spectrometry, the method comprising: a. providing an analyte A with known isotopic composition and a spike prepared in a solvent or solvent mixture containing the analyte S having a different isotopic composition; b. measuring two signal channels corresponding to two different isotopologue features and calculating a signal intensity ratio by dividing the signal intensity from a first channel by signal intensity from a second channel; c. obtaining from step b a first standard solution containing the analyte A multiple times to obtain a signal intensity ratio mean as Ra; d. obtaining from step b a spike containing the analyte S multiple times having a signal intensity ratio mean as Rs; e. preparing a second standard solution containing both the analyte A and S by mixing the second standard solution of the analyte with an accurate concentration Caand the spike, recording quantities of the spike of the second standard solution; f. obtaining in step b a mixture solution incorporating step e multiple times to obtain a signal intensity ratio mean as Rm; ft _ g. calculating a corrected Rm' using the equation Rm= — ;1- m / Ra h. calculating a Ceqvalue, that is dependent on analyte A and S isotopic( compositions, as well as the quantities used, using the equation Ceq= -2-;;4 Rm i. preparing a sample blend by mixing the spike and a sample of any biological matrix containing the analyte A having unknown concentration Cx, to record the quantities of the sample and spike used; j. applying step b on the sample blend in step i to obtain a signal intensity ratio as k. calculating a corrected R' value for the sample blend using the equation in step g by substituting Rmwith R; and l. calculating the analyte A concentration in the sample where a constant b is a scaling factor resulting from a quantity ratio difference in step e and I such that Ca= b x Ceqx R'.

2. The method of claim 1 wherein the analyte A can be any kind of chemical entity that is an ionic species or organic molecule which can be measured by mass spectrometry.

3. The method of claim 1 wherein the analyte S can be an analog of the analyte A in a different isotopic composition having a stable isotope labelled (SIL) form in an organic molecule.

4. The method of claim 1 wherein the spike containing the analyte S is prepared in a solvent or solvent mixture that keeps the analyte S stable for a long period of time with the volume or mass of sample and spike vary as needed.

5. The method of claim 1 wherein the signal channel can be mass to charge ratio for selected ions in single stage mass spectrometry or fragmented ions for tandem mass spectrometry.

6. The method of claim 1 wherein the measurements may include a sample preparation procedure that renders the sample into a liquid format that is suitable for measurement.

7. The method of claim 6 wherein the sample preparation procedure may be one of or a combination of tissue homogenization, cell lysis, solid phase extraction, liquid-liquid extraction, protein precipitation, dilution or shot.

8. The method of claim 1 wherein the sample can be any type of biological origin.

9. A quantitative mass spectrometry method of quantifying a target analyte in a sample comprising: a. fortifying calibrators and samples with a SIL-IS; b. preparing samples and instrument analysis; c. calculating analyte to SIL-IS signal ratio R; d. calculating corrected R' using the equation R' = with predetermined Rsand Ravalues for each calibrator or sample; e. using linear regression with calibrator R' values to obtain a linear calibration model; and f. calculating an analyte concentration in the sample with sample R' and a linear calibration model.

10. The method of claim 9 wherein Rais the R value obtained with a calibration standard solution of analyte.

11. The method of claim 9 wherein Rsis the R value obtained with a SIL-IS solution that is used in the method of claim 9.

12. An integrated sample preparation and analysis system, comprising: a. a sample preparation subsystem; b. a sample analysis subsystem; and c. a control subsystem.

13. The integrated sample preparation and analysis system of claim 12 wherein the sample preparation subsystem comprises processing an unknown sample.

14. The integrated sample preparation and analysis system of claim 12 wherein the sample analysis subsystem comprises performing sample analysis in a method for measuring a sample quantitively by isotope dilution mass spectrometry of claim 1.

15. The integrated sample preparation and analysis system of claim 12 wherein the sample preparation and analysis are performed in a random-access fashion: next sample analysis can be any analyte in any sample.

Images