metabolites

The present invention relates to methods for the ultra-sensitive determination of one or more catecholamine(s) and metabolites thereof in a biological sample, comprising steps of phenylboronic acid (PBA) complexation, solid phase extraction (SPE) on hydrophilic-lipophilic-balanced (HLB) sorbents, liquid chromatography (LC) on pentafluorophenyl (PFP) columns, and tandem mass spectrometry (MS/MS), optionally using summation of multiple reaction monitoring (MRM) data.

Catecholamines play a vital role in the interactions between the nervous and immune systems and their dysfunctions are implicated in various autoimmune and neurological diseases. However, accurate quantitation of catecholamines in the immune system presents a special analytical challenge due to their low concentrations and complicated matrix.

For a long time, the central nervous system has been considered immune privileged. However, a growing body of evidence has suggested an extensive cross-talk between the nervous and immune systems. The neurotransmitter's endogenous production and releasing mechanism, function in the immune system, and regulation of various immune-related diseases, has become an emerging new field of multidisciplinary research. Catecholamines, including norepinephrine (NE), epinephrine (E) and dopamine (DA), one major category of neurotransmitters, has gained special attention. Extensive evidence has suggested that dysfunction of sympathoadrenergic and dopaminergic pathways has been implicated in a wide range of autoimmune and neurological diseases, such as multiple sclerosis, rheumatoid arthritis, schizophrenia, Parkinson's disease, depression, etc. The quantitation of catecholamines has been considered vital to evaluate the dysfunctions of these two important pathways and monitor diverse diseases as well as facilitate new therapeutic drug development. However, compared to widely studied catecholamine analysis in urine, plasma and tissues, detection of catecholamines in the immune system has rarely been explored. The occurrence of catecholamine in the immune system was first proved using capillary electrophoresis. Catecholamines were detected in lymphocytes of human cerebrospinal fluid and subsequently in human peripheral blood mononuclear cells (PBMC). The existence of NE and DA in human PBMC was further confirmed by using high-performance liquid chromatography combined with electrochemical detection (HPLC-ECD). Endogenous levels of NE, E and DA in human PBMC were determined to be 0.206, below detection and 0.121 pmol/10 million cells, respectively, whereas 4.7, 5.4 and 4.1 pmol/10 million cells were reported in another study. The presence of catecholamines in immune cells evidenced by these pioneer studies clearly supported the communication between the nervous and immune systems. However, the analytical methods used for the catecholamine measurement had several drawbacks. In particular, a high limit of detection is associated with capillary electrophoresis methods; further, HPLC-ECD methodology generally requires laborious sample clean-up and a long chromatographic run to minimize the potential interferences. No sample purification was performed in these endeavors.

The determination of catecholamines in PBMC presents a unique analytical challenge. First, the high polar nature of catecholamines makes them weakly retained on the traditional reverse-phase C18 column, thus substantially reducing the mass spectrometry (MS) ionization efficiency. Second, the basal levels of catecholamines in PBMC are extremely low indicated by the fact that the highly sensitive HPLC-ECD method lacked sufficient sensitivity for this application, thus demanding an ultra-sensitive detection. Third, the instability derived from possible degradation of catechol limits the choices of sample treatment. Last but not the least, the requirement of high volumes of blood (e.g., 150 to 180 imL for two to four experiments) makes the detection of catecholamines in PBMC more challenging compared to other biological matrices (e.g., usually 0.5 imL is sufficient for urine and plasma). Given the clinical significance of the endogenous catecholamines in human PBMC and limitations of the existing analytical methods, the need for an alternative method with high sensitivity and selectivity and an effective sample preparation is clear. Tandem mass spectrometry (MS/MS) coupled with liquid chromatography (LC) (LC- MS/MS) has been increasingly employed as an alternative analytical tool for the catecholamine measurement in biological matrices among a wide range of detection technologies in view of its superior sensitivity and specificity. Pioneering work demonstrated the advantage of LC-MS/MS by the specific determination of catecholamines in urine samples after applying phenylboronic acid (PBA)-based liquid-liquid extraction in the presence of drugs that commonly interfere with HPLC analysis. However, the sensitivity was low (e.g., 2.5 ng/mL for E), and the liquid- liquid extraction was time-consuming.

Several different approaches have been reported to increase the sensitivity and selectivity for LC-MS/MS detection of catecholamines in biological matrices, including: derivatization, ion-pair reagents and hydrophilic interaction chromatography (HILIC) column and different solid phase extraction (SPE). Chemical derivatization is a common strategy to improve chromatographic separation and increase mass spectrometric detection of analytes by converting the high polar analytes to less polar and easier ionizable counterparts. An intriguing reductive ethylation labeling of catecholamines was reported for LC-MS/MS analysis of NE and E in plasma samples with considerably increased sensitivity to 20.0 pg/mL and 5.0 pg/mL, respectively. However, the key reductive ethylation labeling was a laborious process involving a highly toxic chemical, which required to be performed in a vacuum hood and thus hampered its application in general clinical laboratories. Additionally, the detection of another important catecholamine, DA, was not explored. Utilization of ion-pair reagents was another effective strategy to improve sensitivity, but at the expense of causing MS contamination and LC peak problems. HILIC was applied as an alternative approach for the plasma catecholamine analysis coupled with a sample clean-up on a 96-well weak cation exchange (WCX) Elution plate. Although better retention was obtained, the sensitivity of 50 pg/mL for E was insufficient to detect the low endogenous levels of E in PBMC. A significant progress was recently made by a fast and sensitive UPLC-MS/MS method coupled with SPE on a manufactured 96-well alumina microplate, enabling simultaneous quantification of catecholamines in human plasma without derivatization. However, inadequate recovery of 38 to 48% and a substantial high matrix effect of 178 to 221 % were observed in this assay. Therefore, despite the notable progress that has been achieved for the detection of catecholamines, the reported methods either required extensive sample preparation and/or lack of sensitivity, no methods are available for sensitive and fast detection of trace amount of catecholamines in human PBMC.

Determination of catecholamines and metanephrines in human urine has been commonly used for the diagnosis of pheochromocytoma and neuroblastoma along with a wide range of other pathological conditions. Numerous liquid chromatography (LC) methods combined with different detectors have been reported for the determination of catecholamines and metanephrines in human urine, such as: electrochemical detector (ECD), fluorescent detector and mass spectrometry (MS)- based detector. LC-Tandem MS (LC-MS/MS) has gradually been used due to its inherent selectivity and sensitivity. Although LC-MS/MS offers improvements over other detectors, several conditions of analysis make its use cumbersome. An effective sample cleanup for the complicated urine matrix prior to analysis is crucial, including liquid-liquid extraction (LLE), solid phase extraction (SPE) and others. A high volume of urine was generally required (e.g., 0.5 - 1 .0 imL) for quantitative analysis due to the low physiological concentrations of catecholamines and metanephrines. The pH of the pre-treated urine sample played a vital role on the extraction recovery in the reported assays. Hence, to achieve reliable extraction, it is essential to adjust the pH of each sample to a narrow pH range using a base or acid. However, the pH of individual urine could vary from 4.5 - 8.5, making the pH adjustment cumbersome and time-consuming. For example, a pH value of about 9.5 is necessary in a liquid-liquid extraction before LC-MS/MS analysis, and WCX, the most common sorbent used in SPE for catecholamines and metanephrines, requires a narrow pH range of 6 - 7 to ensure that both analytes and sorbent are in their ionized forms. Additionally, the instability of the catechol moiety in strong basic conditions makes the pH adjustment more risky and challenging, and could easily cause data uncertainty or errors. In previous studies, an LC-MS/MS catecholamine method incorporated with SPE on HLB with PBA complexation was described. Since the catecholamines instantly formed a stable complex with PBA, the instability was not an issue during the cleanup. However, pH check and mild adjustment for all the samples were necessary due to the high volume of urine {i.e., 0.6 imL) used and the required pH range of pH 7.5 - 9.5 to obtain good extraction efficiency for this protocol. Therefore, laborious sample preparation for the available assays has significantly affected the speed and throughput of the assay, which has become the bottleneck for routine testing of catecholamine in urine.

Although many assays have been reported for the detection of urinary catecholamines or metanephrines, the majority of the available methods were designed only for either catecholamines or metanephrines, not both. For example, HPLC-ECD was utilized for urinary catecholamine analysis while an LC-MS/MS method was applied for metanephrine determination. The measurement of both catecholamines and metanephrines in a single run has been far less explored. Such simultaneous LC-MS/MS analysis would not only reduce costs for routine testing, but also facilitate better clinical interpretation. However, available assays are not sensitive enough to monitor low levels of E and/or have a narrow dynamic range for high concentrations of dopamine (DA) and metanephrines in a single run. Furthermore, sample cleanup with regular SPE 30 mg/1 imL sorbent required substantial consumption of isotopically-labeled internal standard, reagent, solvent and waste disposal.

To address the limitations of the previously reported methods for catecholamine analysis, the primary goal of the present invention was to develop and validate a fast, highly sensitive and selective LC-MS/MS method incorporating a simple sample preparation for the simultaneous determination of catecholamines in biological samples such as e.g. human PBMC lysate or urine. To date, no such MS/MS-based methods are available for this purpose.

Therefore, the technical problem underlying the present invention is the provision of methods for the easy, fast, and cost-efficient ultra-sensitive determination of catecholamines.

The solution to the above technical problem is achieved by the embodiments characterized in the claims. In particular, the present invention relates to a method for the determination of one or more catecholamine(s) in a biological sample, comprising the steps of:

(a) treating the biological sample with a buffer comprising phenylboronic acid

(PBA) under alkaline conditions,

(b) subjecting the treated biological sample obtained in step (a) to solid phase extraction (SPE) on a hydrophilic-lipophilic-balanced (HLB) sorbent extraction plate, HLB sorbent cartridge, or HLB sorbent column,

(c) subjecting the eluate obtained in step (b) to liquid chromatography (LC),

(d) subjecting column effluent from the LC to tandem mass spectrometry (MS/MS), and

(e) determining the presence and/or amount of the one or more catecholamine(s) in the biological sample based on the acquired MS/MS data.

As used herein, the term "determination of one or more catecholamine(s)" relates to qualitative detection, i.e., detection of the presence or absence of a given catecholamine in the biological sample, as well as quantitative detection, i.e., detection of the specific amount of a given catecholamine in the biological sample.

The type of biological sample that can be used in the methods of the present invention is not particularly limited, provided that (i) it is a biological sample that is suspected of comprising one or more catecholamines, and (ii) it can be brought into solution so that it can be subjected to SPE. Preferably, the biological sample is selected from the group consisting of plasma, serum, urine, a tissue sample, and a cell sample, wherein plasma and urine samples can be present as dried plasma and urine spots. In a particularly preferred embodiment, the biological sample is a cell sample comprising peripheral blood mononuclear cells (PBMC). In case the biological sample is a tissue sample, said tissue sample is first homogenized to obtain a single cell suspension. In another particularly preferred embodiment, the biological sample is urine.

In case of a homogenized tissue sample, or in case of cell samples, e.g. cell samples comprising PBMC, the method of the present invention preferably comprises, prior to step (a), the step of lysing the cells in the biological sample. Preferably, the cells, e.g. PBMC, are lysed in an acidic lysis buffer, preferably in 0.1 to 2.0 N acetic acid (HOAc), more preferably 0.1 to 0.5 N HOAc, more preferably about 0.2 N HOAc. Preferably, the lysis buffer is used ice-cold. Catecholamines that can be determined with the method of the present invention are not particularly limited. Preferably, they are selected from the group consisting of norepinephrine (NE), epinephrine (E), dopamine (DA). Further, catecholamine metabolites that can be determined with the method of the present invention are not particularly limited. Preferably, they are selected from the group consisting of metanephrine, normetanephrine, 3-methoxytyramine, 3,4-dihydroxyphenylalanine (DOPA), 3,4-dihydroxyphenylacetic acid (DOPAC), 3,4-dihydroxyphenylglycol (DHPG), 3,4-dihydroxymandelic acid (DOMA), homovanillic acid (HVA), 4-hydroxy- 3-methoxyphenylglycol (MHPG), vanillylmandelic acid (VMA) and serotonin (5-HT). In step (a) of the method of the present invention, the biological sample is treated with a buffer comprising phenylboronic acid (PBA) under alkaline conditions. In this context, the term "treated (... ) under alkaline conditions" encompasses combining the sample with the buffer under conditions that allow for the reaction of catecholamines present in the sample with PBA, wherein the pH of the combined solution is an alkaline pH. Suitable conditions in this respect are known in the art. In particular embodiments, the above buffer contains 0.1 to 1 .0% (w/v) PBA, preferably 0.1 to 0.5% (w/v) PBA, and has a pH of between 7.5 and 1 1 , preferably between 7.5 and 9.5. In a particularly preferred embodiment, the buffer contains about 0.2% (w/v) PBA and has a pH of about 8.5.

In step (b) of the method of the present invention, the PBA-treated biological sample is subjected to solid phase extraction (SPE) on a hydrophilic-lipophilic-balanced (HLB) sorbent extraction plate, HLB sorbent cartridge, or HLB sorbent column. SPE procedures are not particularly limited and are known in the art. Preferably, SPE comprises steps of loading the biological sample to the optionally pre-conditioned and equilibrated extraction plate (or cartridge or column), washing the plate (or cartridge or column), drying the plate (or cartridge or column), and then eluting plate- bound material or material bound on the cartridge or column. HLB sorbent extraction plates, cartridges or columns that can be used in the method of the present invention are not particularly limited and are known in the art. In particular, HLB sorbents are macroporous copolymers made from a balanced ratio of lipophilic divinylbenzene monomers and hydrophilic N-vinylpyrrolidone monomers. Respective sorbents and plates, cartridges or columns can be obtained commercially, e.g. from Waters (Milford, MA, USA) under the trade-name Oasis ^® HLB, from Phenomenex (Torrance, CA, USA) under the trade-name of Strata™-X, or under different names from other manufacturers. Respective plates can e.g. be in the form of 96-well plates or 96-well microplates, wherein the latter are particularly preferred. However, sorbents can also be used in the present invention also in the form of individual HLB cartridges or columns.

Preferably, the SPE in step (b) of the method of the present invention comprises the steps of washing the extraction plate, cartridge or column one time with an aqueous wash buffer and subsequently two times with a wash buffer comprising between 20 and 40% (v/v) MeOH. In a preferred embodiment, the SPE does not comprise any further washing steps. The aqueous wash buffer used in this respect preferably is 0.1 to 0.5 N NH ₄ CI-NH ₄ OH, more preferably about 0.2 N NH ₄ CI-NH ₄ OH. The wash buffer comprising MeOH used in this respect preferably is a wash buffer comprising 10 to 50% (v/v) MeOH, more preferably 20 to 40% (v/v) MeOH, more preferably, about 30% (v/v) MeOH or about 20% (v/v) MeOH. Preferably, said wash buttfer is 10 to 50% (v/v) MeOH/ 0.1 to 1 .0 N NH ₄ CI-NH ₄ OH, more preferably 20 to 40% (v/v) MeOH/ 0.1 to 0.5 N NH ₄ CI-NH ₄ OH, more preferably about 30% (v/v) MeOH/ about 0.2 N NH ₄ CI-NH ₄ OH.

In a preferred embodiment, SPE in step (b) of the method of the present invention comprises the step of eluting plate-bound material (or material bound on a cartridge or column) in a volume of at most 100 μΙ, preferably at most 75 μΙ, more preferably at most 50 μΙ, more preferably of about 50 μΙ, per sample. The eluent used in this respect is preferably 0.1 to 2.0 N acetic acid (HOAc), more preferably 0.1 to 0.5 N (HOAc), more preferably about 0.2 N HOAc. In step (c) of the method of the present invention, the eluate obtained from the SPE is subjected to liquid chromatography (LC). Methods for performing LC are not particularly limited and are known in the art. Preferably, said LC is an LC using a pentafluorophenyl (PFP) column. Respective columns are not particularly limited and are known in the art. They can be obtained commercially, e.g. from Phenomenex (Torrance, CA, USA) under the trade-name of Luna ^® PFP(2) columns. Suitable columns can have e.g. a length of 150 mm, an inner diameter of 2.1 mm, and a particle size of 3 μιη. Preferably, the LC of step (c) of the method of the present invention uses a gradient elution profile using as a first mobile phase (mobile phase A) 0.005 to 0.1 % (v/v) HCOOH in water, preferably 0.005 to 0.05% (v/v) HCOOH in water, more preferably about 0.01 % (v/v) HCOOH in water, and as a second mobile phase (mobile phase B) 0.005 to 0.1 % (v/v) HCOOH in MeOH, preferably 0.005 to 0.5% (v/v) HCOOH in MeOH, more preferably about 0.01 % (v/v) HCOOH in MeOH. An exemplary gradient profile is 5% B for 1 min, linear increase to 22% B over 1 .5 min, ramp-up to 97% B in 0.2 min and maintained for 1 .5 min, and then equilibration with 5% B for 2.5 min, using a flow rate of 0.45 imL/min. In step (d) of the method of the present invention, column effluent from the LC is subjected to tandem mass spectrometry (MS/MS). In preferred embodiments, in order to improve sensitivity, multiple reaction monitoring (MRM) data is acquired for each catecholamine or metabolite thereof to be determined by summation of 3 or more, e.g. 3, 4, 5, 6, 7, 8, 9, 10, or more, preferably 3 or 5, identical MRM transitions for each catecholamine or metabolite thereof. In this embodiment, step (e) of the methods of the present invention is based on the acquired MRM data. Due to the extremely low analyte levels in PBMCs, this embodiment is of particular use when using cell samples comprising PBMCs, or when using samples where analyte levels are expected to be extremely low. For urine samples, this embodiment is usually not required, but may nevertheless be used.

In this context, MRM is a mode for operating a triple quadrupole (Q) instrument, wherein an ion of a given m/z fragments or dissociates in Q1 to give a product ion of specific m/z in Q3, and the Q1 -Q3 transition is used for quantitation. Summation of MRM transitions can be performed in a single analysis, e.g. by using summation of multiple ions via the software controlling the mass spectrometer. Preferably, during MS/MS in step (d) of the method of the present invention, for NE, the transitions 152.0 > 107.0 (quantifier) and 152.0 > 135.0 (qualifier) are determined, and/or for E, the transitions 166.0 > 107.0 (quantifier) and 166.0 > 135.0 (qualifier) are determined, and/or for DA, the transitions 154.0 > 91 .0 (quantifier) and 154.0 > 1 19.0 (qualifier) are determined and/or for NMN, the transitions 166.0 > 121 .0 (quantifier) and 166.0 > 106.0 (qualifier) are determined, and/or for MN, the transitions 180.0 > 165.0 (quantifier) and 180.0 > 148.0 (qualifier) are determined. In this context, the quantifier transition is the MRM transition used for the quantitation, whereas the qualifier transition is the MRM transition used for confirmation.

Procedures for performing MS/MS and MRM analysis are not particularly limited and are known in the art. The same applies for procedures for performing SPE and LC as defined above. Preferably, the mass spectrometer is operated in positive electrospray ionization (ESI) mode. Exemplary ion source parameters are outlined in section "Material and Methods" below.

In step (e) of the method of the present invention, the presence and/or amount of the one or more catecholamine(s) in the biological sample is determined based on the acquired MS/MS or MRM data. To this end, suitable standards are included in MS/MS as known in the art. Data analysis and determination of the results can be performed with any suitable software known in the art, e.g. the software Analyst (Version 1 .6.2) available from AB Sciex (Foster City, CA, USA).

In a particularly preferred embodiment, the present invention relates to a method for the determination of one or more catecholamine(s), selected from the group consisting of NE, E, and DA, in a biological cell sample comprising PBMC, the method comprising the steps of:

(a) lysing said PBMC in about 0.2 N HOAc (b) treating the lysed biological sample with a buffer comprising about 0.2% (w/v) PBA and having a pH of about 8.5,

(c) subjecting the treated biological sample obtained in step (b) to SPE on a hydrophilic-lipophilic-balanced (HLB) sorbent extraction microplate, wherein said SPE comprises the steps of:

(i) washing the extraction plate once with about 0.2 N NH ₄ CI-NH ₄ OH, subsequently

(ii) washing the extraction plate twice with about 30% (v/v) MeOH/ about 0.2 N NH ₄ CI-NH ₄ OH, and subsequently

(iii) eluting plate-bound material in a volume of about 50 μΙ of about 0.2 N HOAc,

wherein said SPE does not contain any further washing steps,

(d) subjecting the eluate obtained in step (b) to liquid chromatography (LC) using a pentafluorophenyl (PFP) column, said LC using a gradient elution profile using as first mobile phase (mobile phase A) about 0.01 % (v/v) HCOOH in water, and as second mobile phase (mobile phase B) about 0.01 % (v/v) HCOOH in MeOH,

(e) subjecting column effluent from the LC to tandem mass spectrometry (MS/MS), wherein multiple reaction monitoring (MRM) data is acquired for each catecholamine to be determined by summation of 5 identical MRM transitions for each catecholamine, and wherein

for NE, the transitions 152.0 > 107.0 (quantifier) and 152.0 > 135.0 (qualifier) are determined, and/or

for E, the transitions 166.0 > 107.0 (quantifier) and 166.0 > 135.0 (qualifier) are determined, and/or

for DA, the transitions 154.0 > 91 .0 (quantifier) and 154.0 > 1 19.0 (qualifier) are determined, and

(f) determining the presence and/or amount of the one or more catecholamine(s) in the biological sample based on the acquired MRM data.

Additional experimental details concerning the method of the present invention and preferred embodiments thereof can be found under "Material and Methods" below. Any particular embodiments described therein can expressly be combined with any particular embodiments described above.

As used herein, the term "about" is intended to be a modifier of ± 10%, preferably ± 5% of the specified value. As an example, the term "about 0.2" is intended to encompass the range of 0.18 to 0.22, preferably 0.19 to 0.21 .

The terms "comprising/comprises", "consisting of/consists of", and "consisting essentially of/consists essentially of" are used herein in an interchangeable manner, i.e., each of said terms can expressly be exchanged against one of the other two terms.

The present invention provides the first LC-MS/MS method for the determination of catecholamines in biological samples such as human peripheral blood mononuclear cells (PBMC) with significantly improved sensitivity, selectivity and throughput with no need for derivatization, evaporation and ion-pairing reagents. The proposed method should lend a strong support for the advance of the new emerging arena of studies of the neural-immune network. In particular, the lower limit of quantification of 1 pg/mL for E and 5 pg/mL for NE and DA represents a significant sensitivity improvement over available methods. Less than 8.7% of intraday and interday precision, 91 .8 to 1 1 1 .3% of accuracy and successful assessment of the reference interval for 40 healthy donors demonstrates good reproducibility and reliability of the assay. The novel PBA-HLB-PFP-MRM summation approach allows rapid, sensitive and reliable determination of catecholamines in PMBC in high throughput, which will facilitate better understanding of the new arena of neuro-immune network. Additionally, the significantly improved method can be applied to quantify catecholamines and metabolites in other biological matrices.

Regarding the determination of one or more catecholamine(s) in urine samples, only 10 μΙ_ of urine are necessary for the simultaneous LC-MS/MS detection of catecholamines and metanephrines after SPE on an HLB μΕΙυΐίοη plate upon PBA complexation. Compared to the high volume urine sample (e.g., 0.5 - 1 .0 imL) generally required, the minimized sample volume reduces the pH variation derived from urine sample, and promotes the instant formation of a stable PBA-complex, thus significantly simplifying the pH adjustment while improving the throughput. Moreover, the decrease of sample volume required for testing allows quantitation for limited samples or more experiments to be conducted with the remaining sample volume, which is particularly important for pediatric urine samples and other animal model biological studies. Additionally, the miniature SPE platform substantially reduces the internal standard, reagent and solvent consumption and waste disposal, as well as the analysis cost, which would be beneficial for implementation in clinical routine analysis.

The figures show:

Figure 1 :

Optimization of SPE conditions: (A) effect of lysis buffer; (B) effect of washing buffer: Wash 1 , three washes of 30% MeOH in NH ₄ CI buffer; Wash 2, one wash of NH ₄ CI buffer followed by two washes of 30% MeOH in NH ₄ CI buffer; Wash 3, two washes of NH ₄ CI buffer followed by one wash of 30% MeOH in NH ₄ CI buffer; Wash 4, one wash of NH ₄ CI buffer followed by two washes of 30% (acetonitrile: isopropanol = 1 : 1 ) in NH ₄ CI buffer; (C) effect of elution buffer.

Figure 2:

Assessment of HCOOH concentration as additive in mobile phase A. (A) 0.1 %; (B) 0.05%; (C) 0.01 %; (D) 0%.

Figure 3:

Baseline LC chromatographic separation of catecholamines and metanephrines in a standard solution. (A) NE; (B) E; (C) normetanephrine; (D) DA; (E) metanephrine. Figure 4:

Comparison of catecholamine quantitation with a single MRM transition (top panels) and summation of five MRM transitions (bottom panels) for a representative PBMC lysate. (A) and (D), NE MRM quantifier 152.0 > 107.0, 147 pg/mL; (B) and (E), E MRM quantifier 166.0 > 107.0, 3.2 pg/mL; (C) and (F), DA MRM quantifier 154.0 > 91 .0, 18.7 pg/mL.

Figure 5:

Endogenous concentrations of catecholamines in apparently healthy population (n = 40): 0.0658 - 0.8063 pmol/10 million cells with a mean of 0.3593 pmol/10 million cells for NE, 0.0020 - 0.0363 pmol/10 million cells with a mean of 0.01 12 pmol/10 million cells for E and 0.0033 - 0.0493 pmol/10 million cells with a mean of 0.0203 pmol/10 million cells for DA. Figure 6:

Optimization of SPE conditions: (A) & (B) effect of washing buffer: three washes of 30%, 20% and 10% MeOH in NH ₄ CI buffer; (C) effect of urine loading volume; (D) effect of PBA volume.

The present invention will be further illustrated by the following examples without being limited thereto.

Examples

Material and Methods: Reagents and Materials

Catecholamine and metanephrine standard solutions were provided by Cerilliant (Round Rock, Texas, USA). Isotopically-labeled internal standards (IS) including d6- NE, d3-E, d4-DA, d3-NMN and d3-MN were purchased from CDN isotopes (Pointe- Claire, Quebec, Canada). Endocrine urine controls were obtained from Chromsystems (Grafelfing, Munich, Germany). Mass Spect Gold® Urine was supplied by Golden West Biologicals (Temecula, CA, USA), and was confirmed with no detectable analytes to be used as blank urine. Positive pressure-96 processor and Oasis HLB 96-well pElution plate (2 mg/30 pm) were obtained from Waters (Milford, MA, USA). LC-MS grade of water, water with 0.1 % formic acid (HCOOH), methanol (MeOH) and acetic acid (HOAc), analytical grade of 28% ammonium hydroxide solution (NH4OH), ammonium chloride (NH4CI), perchloric acid (HCIO4), Dulbecco's phosphate-buffered saline (DBPS) and 2-Aminoethyl diphenyl borate were obtained from Sigma-Aldrich (St. Louis, MO, USA). Heparin tubes for blood collection were obtained from Becton, Dickinson and Company (BD, Franklin Lakes, NJ).

LC conditions

Chromatographic measurement was performed on an ultrafast liquid chromatography (UFLC-XR) system equipped with a binary pump, an autosampler, a degasser and a column compartment from Shimadzu Corporation (Columbia, MD, USA). Analyte separation was achieved on a Luna PFP (2) column (150 mm, 2.1 mm i.d., 3 pm) guarded with a pre-column from Phenomenex (Torrance, CA, USA). The column oven was kept constant at 30 °C and auto sampler was set at 4 °C. The analytes were eluted by a gradient mobile phases containing 0.01 % HCOOH in water (Mobile phase A) and 0.01 % HCOOH in MeOH (Mobile phase B).

For urine samples, after five microliters of sample was loaded on the column, the initial LC gradient of 5% B was linearly increased to 35% within 2.5 min. Then the gradient was quickly ramped to 95% B in 0.1 min and held for 1 .2 min, finally the column was equilibrated at 5% B for 1 .2 min. The flow rate was 0.45 imL/min and the cycle time was 5.5 min.

For PBMC samples, the initial LC gradient of 5% B was held for 1 min prior to linearly increasing to 22% over 1 .5 min. Then the gradient was ramped to 97% B in 0.2 min and maintained for 1 .5 min, finally the column was equilibrated at 5% B for 2.5 min with a turnaround time of 7 min. The auto sampler was kept at 4 °C and twenty microliters of sample was injected.

MS/MS apparatus and conditions

Analyte detection was performed by an AB Sciex 5500 triple quadrupole mass spectrometer (Foster City, CA, USA) in positive electrospray ionization (ESI) mode, optionally with multiple reaction monitoring (MRM). For urine samples, only the LC effluent from 0.6 min to 4.0 min was subjected to the mass spectrometer utilizing a diverter valve. For PBMC samples, only the column effluent from 0.5 to 2.5 min was introduced. Analyst software version 1 .6.2 from AB Sciex was used for instrument control, data acquisition and analysis. The ionspray voltage was optimized to be 2200 V at 600 °C and curtain gas was maintained at 35 psi. The nebulizer gas was set at 50 psi, while the heater gas was maintained at 55 psi and CAD gas was kept at 9 psi. To improve sensitivity for detecting catecholamines in PBMC samples, due to extremely low endogenous levels, summation of five identical MRM transitions was utilized for analyte quantitation. However, MRM summation was not applied for the analysis of urine samples. The MRM transitions and compound parameters were individually optimized, and are summarized in Tables 1 A (PBMC) and 1 B (urine) below.

Table 1 A: Optimal MRM parameters (PBMC)

Analyte MRM transition (m/z) DP (V) EP (V) CE (eV) CXP (V)

NE (quantifier) 5 x 152.0 > 1 07.0 1 10 10 23 10

NE (qualifier) 152.0 > 135.0 1 10 10 23 12 d6-NE 158.0 > 1 1 1 .0 101 10 25 10

E (quantifier) 5 x 166.0 > 1 07.0 130 10 27 10

E (qualifier) 5 x 166.0 > 135.0 130 10 21 12 d3-E 169.0 > 107.0 130 10 27 10

DA (quantifier) 5 x 154.0 > 91 .0 55 10 32 1 1

DA (qualifier) 5 x 154.0 > 1 19.0 55 10 24 12 d4-DA 158.0 > 95.0 55 10 32 1 1

DP: Declustering potential; EP: Entrance potential; CE: Collision energy; CXP: Collision cell exit potential

Table 1 B: Optimal MRM parameters (urine)

Analyte MRM transition (m/z) DP (V) EP (V) CE (eV) CXP (V)

NE (quantifier) 152.0 > 107.0 1 10 10 23 10

NE (qualifier) 152.0 > 135.0 1 10 10 23 12 d6-NE 158.0 > 139.0 101 10 25 10

E (quantifier) 166.0 > 107.0 130 10 27 10

E (qualifier) 166.0 > 135.0 130 10 21 12 d3-E 169.0 > 107.0 130 10 27 10

DA (quantifier) 154.0 > 91 .0 55 10 32 1 1

DA (qualifier) 154.0 > 1 19.0 55 10 24 12 d4-DA 158.0 > 95.0 55 10 32 1 1

NMN (quantifier) 166.0 > 121 .0 120 10 23 15

NMN (qualifier) 166.0 > 106.0 95 10 25 10 d3-NMN 169.0 > 137.0 95 10 23 12

MN (quantifier) 180.0 > 165.0 121 10 23 16

MN (qualifier) 180.0 > 148.0 121 10 23 15 d3-MN 183.0 > 151 .0 130 10 25 24

DP: Declustering potential; EP: Entrance potential; CE: Collision energy; CXP: Collision cell exit potential Creatinine concentration was determined by an in-house method using Roche COBAS INTEGRA 400 plus (CA, USA).

Preparation of calibration standard and quality control (QC) samples

A stock solution of the five catecholamines and metanephrines at 10 pg/mL was prepared in 0.2M HOAc. A working calibration stock solution containing 100 ng/mL of each analyte was created by dilution of the stock standard solution with 0.2N HOAc. A calibration curve was established by serial dilution of the working calibrator solution with pooled PBMC lysate samples at the following concentrations: 1 , 2, 5, 10, 25, 50, 100, 250, 500, 1000 and 2500 pg/mL. Due to the presence of endogenous catecholamines in the pre-screened pooled matrix, the basal levels of catecholamines were extrapolated for the calibration curve. An IS working solution containing 1 ng/mL of α¾-ΝΕ, α¾-Ε and ck-DA was prepared in 0.2N HOAc by serial dilution of a stock IS solution containing 100 ng/mL of these ISs. The stock solutions were stored at -20°C until use.

For urine samples, the calibration stock solution was diluted with 0.2M HCI acidified blank urine to yield a working calibrator solution at a concentration of 1250 ng/mL for each analyte. Calibration curves in the concentration range of 0.50 - 1250 ng/mL were established by serial dilution of the working calibrator solution with the acidified blank urine. A mixture of internal standard (IS) working solution including d6-NE (5 ng/mL), d3-E (1 ng/mL), d4-DA (10 ng/mL), d3-NMN (10 ng/mL) and d3-MN (10 ng/mL) was prepared in water by ten times dilution of combined IS stock solution in 0.2M HOAc. LLOQ, low, medium and high QC samples were independently constructed by spiking calibrators to in-house prepared pooled urine samples. All the stocks were stored at -20°C until analyzed. Sample preparation procedure

1 . PBMC lysate sample preparation

Venous blood samples were drawn into 10 ml_ heparinized tubes and centrifuged using LeucoSep tube (Monroe, NC, USA) according to manufacture instruction. The supernatant containing PBMC was collected and potential contaminating red blood cells (erythrocytes) were removed by 10 min digestion with an NH ₄ CI solution (STEMCELL Technologies Inc., Vancouver, BC, Canada). After washing PBMC three times with Dulbecco's phosphate buffered saline (DPBS), cell numbers were determined by ViCell XR Cell Viability Analyzer (Beckman Coulter, Chaska, MN, USA) with viability > 99%. Finally, PBMC were lysed by ice-cold 0.2N HOAc at a final concentration of 40 million cells/mL and stored at -80°C overnight prior to use.

2. Solid phase extraction procedure (PBMC lysate)

250 μΙ_ PBMC lysates, standards or QC samples were pre-mixed with 50 μΙ_ of 1 ng/mL IS working solution, 50 μΙ_ 7% NH ₄ OH and 250 μΙ_ 0.2% PBA buffer (pH 8.5). The entire mixture was loaded slowly to an Oasis ^® HLB 96-well μΕΙυίίοη extraction plate, which was pre-conditioned and equilibrated by 200 μΙ_ of MeOH and 0.2N NH ₄ CI-NH ₄ OH buffer (pH 8.5), respectively. After loading, each well was first washed by 200 μΙ_ of 0.2N NH ₄ CI-NH ₄ OH buffer (pH 8.5) followed by washing twice with 200 μΙ_ of 30% MeOH in 0.2N NH ₄ CI- NH ₄ OH (pH 8.5). After that, the wells were completely dried by vacuum before 50 μΙ_ of 0.2N HOAc was applied to elute the retained catecholamines and ISs for LC-MS/MS analysis. Detailed protocols regarding the preparation of the SPE buffers is provided below. a) 2N NH ₄ CI buffer at pH 8.5

NH ₄ CI (106.98 g, 2 mol) was dissolved with 900 ml water and the pH of the solution was adjusted to 8.5 with NH ₄ OH (28% NH3 in H2O). Then water was added to reach a total volume of 1000 ml and mixed well before use.

b) 0.2N NH ₄ CI-NH ₄ OH buffer at pH 8. 5 (NH ₄ CI buffer)

100 imL of 2N NH ₄ CI buffer (pH 8.5) was diluted with 900 ml water and the solution was mixed well before use. c) 0.2% PBA complexation buffer at pH 8. 5

2-Aminoethyl diphenylborinate (1 .0 g) and EDTA.2Na.2H2O (2.5 g) were dissolved in 2N NH ₄ CI buffer (500 imL, pH 8. 5) with stirring until a homogeneous solution was observed. The pH was adjusted to 8.5 with NH ₄ OH. d) 30% MeOH in 0.2N NH ₄ CI-NH ₄ OH buffer at pH 8.5

300 ml MeOH was mixed well with 700 ml of 0.2N NH ₄ CI buffer (pH 8.5) before use. e) 0.2N Acetic acid elution buffer

Acetic acid (12 imL) was mixed thoroughly with water (898 ml) and then filtered through 0.22 μιη filter. Solid phase extraction procedure (urine)

A sample mixture for loading was prepared by mixing 10 pL of urine sample, standard or QC sample, 100 μΙ_ of IS working solution, 50 μΙ_ 0.2% PBA buffer (pH 8.5) and 100 μΙ_ of 0.7% NH ₄ OH solution. An Oasis HLB 96-well pElution extraction plate was conditioned with 200 μΙ_ of MeOH and equilibrated by 0.2M NH ₄ CI-NH ₄ OH buffer at pH 8.5 (NH ₄ CI solution). After slowly loading the pre-mixed sample mixture to the pre-treated extraction plate, each well was washed by three washes of 200 μΙ_ 20% MeOH in NH ₄ CI buffer. Afterwards, the wells were completely dried by vacuum and the remaining analytes and corresponding ISs were eluted by 100 μΙ_ of 0.2M HOAc, which was directly injected for LC-MS/MS analysis.

Assay performance (PBMC)

The assay was validated for linearity, limit of detection (LOD), and lower limit of quantification (LLOQ), precision, accuracy, extraction recovery, matrix effect, carryover, analyte stability and reference interval according to the guidelines of Clinical and Laboratory Standards Institute (CLSI).

1 . Linearity, LOD and LLOQ

The standard curve was established by linear regression of the peak area ratio of catecholamine to the corresponding IS versus nominal catecholamine concentration with a weighting factor of 1 /x. The linearity was evaluated by analyzing PBMC lysate samples prepared in the concentration range of 1 to 2500 pg/mL for E and 5 to 2500 pg/mL for NE and DA at five different days. The coefficient of determination (r) should be greater than 0.9990. LLOQ was determined as the lowest concentration at which the signal-to-noise ratio was greater than ten with accuracy and precision less than 20%. LOD was defined as the lowest concentration at which chromatographic peak of catecholamine was present with a signal-to-noise ratio at least three. 2. Selectivity

The selectivity of the assay was first assessed during optimization of LC chromatographic separation of E with its isobaric compound, normetanephrine. Additionally, the selectivity was assessed by analyzing 14 authentic PBMC samples to ensure no interfering peaks occurred in the retention time of the analytes and ISs. Precision and accuracy

The intraday and interday assay precision and accuracy were determined by analyzing six replicates of four levels of QC samples by spiking known amount of catecholamines at three different days. Precision of the method was expressed as coefficient of variance (CV%). Assay accuracy was assessed in terms of recovery, which was demonstrated as a percentage of catecholamine concentration change compared with the true added value. The acceptance criterion for accuracy was set within ± 15% from the true value except the LLOQ level at ± 20%. To ensure the accuracy of the assay, additional recovery experiments were conducted by spiking of five different PBMC samples with known amount of catecholamine standards since neither standard method nor certified reference materials were available for method comparison. Recovery of extraction procedure and matrix effect

The extraction recovery and matrix effect of the method of the present invention was assessed as proposed in the art. In this regard, five different PBMC samples were processed with spiking of catecholamine standards at 40, 400 and 1600 pg/mL before and after extraction. Due to the inherent endogenous catecholamines in the matrix, the endogenous contribution determined from the blank sample was subtracted. The extraction recovery in percentage was calculated as the mean increased peak area of analyte spiked before the extraction divided by that of the samples after extraction. The matrix effect was determined by the mean increased peak area of analyte spiked after the extraction divided by the peak area of the neat standard. Additionally, to evaluate possible matrix effect on the quantitation, matrix induced signal suppression or enhancement (SSE) was evaluated by comparison of the slopes of calibration curves generated in the matrix-matched PBMC lysate samples and neat standard solution (n = 2). The measurement is considered not affected by the matrix if the SSE value is within 80% - 120%. 5 Stability

To assess the stability of catecholamines in PBMC lysate, freshly prepared PBMC samples at low and high concentrations were aliquoted and stored under different conditions: 20 ^Q C for four days, 4 ^Q C for four days and -80 ^Q C for three weeks. Additionally, freeze-thaw stability was examined by freezing and thawing PBMC samples after four cycles. Samples were considered stable if less than ±15% difference was obtained for the catecholamine concentrations determined in duplicate on the day of collection and under the specified storage conditions.

Assay validation (urine) The calibration curve for each analyte was established on five different days by extraction of blank urine spiked with five analytes in the concentration range of 0.5 to 1250 ng/mL under the optimized SPE condition followed by LC-MS/MS determination. The linear fit of the peak area ratio of analyte and IS versus corresponding analyte concentration with 1 /x weighting was used for concentration quantitation. Lower limit of detection (LOD) was defined as the lowest concentration with the signal-to-noise ratio greater than three in blank urine sample based on the Clinical and Laboratory Standards Institute (CLSI) guidelines. Lower limit of quantification (LLOQ) was determined as the lowest concentration in blank urine sample can be measured within ±20% of imprecision and accuracy at signal-to- noise ratio greater than ten.

To evaluate assay precision and accuracy, LLOQ, low, medium and high QC samples prepared in blank urine were extracted and analyzed five times on three different days. Additionally, two levels (e.g., normal and pathological) of commercial available QC samples from Chromsystems were processed alongside under the optimal conditions, and the results obtained by this method were compared with the expected mean values provided by the supplier. To lend further support, the accuracy was assessed by performing analytical recovery studies in five random urine samples spiked with three concentration levels of the combined calibrators. Recovery was calculated as a percentage of [(final concentration - initial concentration) / added value].

Statistical analysis

Graph Pad Prism 6 (Graph Pad Software Inc., CA) was used for statistical analysis. EP evaluator release 9 (Data Innovations LLC, VT) was applied for the determination of reference intervals.

Example 1 :

Solid phase extraction (SPE) optimization

Due to the low concentrations of catecholamines in complicated biological matrices, an efficient sample pretreatment for biological samples is essential. SPE has been widely employed to remove interfering compounds and enrich the analytes with different types of sorbents, such as: boric gel, C18, HLB, weak cation exchange, alumina and PBA, etc. However, the reported procedures typically resulted in low sensitivity and/or required laborious evaporation and reconstitution steps. Micro devices of alumina and WCX were recently utilized for the sample clean-up with considerably simplified procedures, however, the extraction recovery was relatively low. Online extraction of catecholamines in urine samples eliminated laborious sample pretreatment, however, these methods required relatively long LC cycle time (14 to 30 min) and additional cost for the automation equipment. HLB (hydrophilic- lipophilic-balanced) sorbent combined with PBA complexation was illustrated as an alternative strategy based on the specific covalent interaction between PBA and catecholamines under alkaline pH. Satisfactory recovery was obtained for HPLC- ECD detection, but these protocols involving MS incompatible elution were not suitable for MS detection. The SPE was modified for LC-MS/MS with acidic elution, however, the sensitivity was inadequate for the detection of catecholamines in PBMC. In addition, a narrow pH range of 7.5 to 9.5 for the pre-treated mixture was crucial to achieve a good extraction recovery. As such, the process became tedious if the pH had to be adjusted for each sample due to substantial pH variation of individual sample, which was a drawback of the PBA-based process. In contrast, the fluctuation of the pH in PBMC lysate sample was negligible since the sample was reconstituted in a lysis buffer, thus the tedious pH adjustment was eliminated. To achieve the simultaneous sample clean-up and analyte enrichment, improve extraction throughput and reduce SPE solvent usage and waste disposal, Oasis HLB in a 96-well Elution format was employed for SPE, which has not been explored.

1 . Effect of lysis buffer

The selection of lysis buffer was investigated first since it was not only a key element for SPE optimization, but also important for maximal lysis of PBMC to release catecholamines for accurate determination. Four different lysis buffers were examined for lysing a pooled PBMC sample including: 1 .2 imM Tris in 70% MeOH at pH 7.4 (Tris-70% MeOH), 0.1 N HCIO ₄ , 0.2N HOAc and 0.4N HCIO4. As illustrated in Figure 1 A, dramatically reduced signal intensity for all the analytes is observed by the use of Tris-70% MeOH buffer for direct SPE. The high MeOH content of the lysis buffer was probably responsible for the decreased response, thus tedious solvent evaporation and reconstitution in an appropriate buffer for SPE were necessary. Among the three acidic buffers, comparable intensities are obtained for NE, whereas 0.2N HOAc provides substantially higher signal response than the HCIO4 buffers regarding E and DA. Therefore, 0.2N HOAc was selected as lysis buffer in terms of the signal response. In addition to the maximal response generated, using the mild 0.2N HOAc as lysis buffer allowed direct SPE, which considerably simplified the SPE protocol and enhanced the throughput of the assay. 2. Effect of washing buffer

Washing buffer plays a vital role to efficiently remove possible interferences in complicated biological matrix while keeping the analyte of interest retained on the sorbent. Preliminary screening revealed that 30% MeOH was the highest organic content in aqueous washing buffer (data not shown). Four sequential washings containing 30% organic solvent were examined with three total washings. The extraction recovery was determined using a pooled PBMC sample spiked with 200 pg/mL of calibrators before and after extraction, and the results are shown in Figure 1 B. The wash 2 involving a 100% aqueous washing before the 30%-MeOH washing substantially increases the extraction recovery for all the analytes compared to the wash 1 without full aqueous washing, indicating the beneficial role of the aqueous washing. However, further increasing aqueous washing to two times

(Wash 3) decreases the recovery. Additional efforts to replace MeOH with other organic solvents (acetonitrile and isopropanol) dramatically decreases extraction efficiency as shown in the washing 4. Therefore, the sequential wash 2 containing one wash of 0.2N NH ₄ CI-NH ₄ OH and two washes of 30% MeOH-0.2N NH ₄ CI-NH ₄ OH was chosen for the optimal washing. Effect of elution buffer

To maximize the pre-concentration factor of the SPE, the volume of the elution solution should be minimized. Therefore, the effect of elution buffer was assessed by varying elution volume and times. Figure 1 C illustrates that the decreased elution of 50 μΙ_ is advantageous over 100 μΙ_ in terms of signal response while no obvious difference is observed by splitting elution into two smaller portions. Considering the easier operation of one elution with relatively larger volume, elution with 50 μΙ_ of 0.2N HOAc was selected as the final elution. Of note, the elution in a low volume of MS compatible solution allows fivefold pre-concentration and direct injection for LC-MS/MS analysis. Example 2:

Optimization of liquid chromatoqraphy-tandem mass spectrometry (LC-MS/MS) conditions 1 . Screening of pentafluorophenyl (PFP) columns

PFP columns have been employed to increase LC retention of catecholamines, however, the sensitivities in these studies were insufficient to detect trace levels of catecholamines in PBMC. To enhance the sensitivity, it was important to select an optimal column among various commercially available PFP columns. Hence, a variety of commercial PFP-type columns were screened, including ultra PFPP column (150 mm, i.d. 2.1 mm, 3 μιη) and (100 mm, i.d. 2.1 mm, 3 μιη) from Restek (Bellefonte, PA, USA), Luna PFP (2) column (150 mm, i.d. 2.1 mm, 3 μιη), Kinetex (150 mm, i.d. 2.1 mm, 2.6 μιη) and F5 (150 mm, i.d. 2.1 mm, 3 μιη) from Phenomenex (Torrance, CA, USA), and Hypersil

PFP column (150 mm, i.d. 2.1 mm, 3 μιη) from Thermo Scientific. Luna PFP (2) column (150 mm, i.d. 2.1 mm, 3 μιη) was found to generate maximum signal response and achieve separation in a short time (data not shown), which was used for further optimization.

2. Selection of additives and mobile phases

Additives in the mobile phase was another significant factor influencing signal response. Assessment of the effect of HCOOH concentration from 0 to 0.1 % in mobile phase A was examined using 1000 pg/mL of catecholamine calibrators in neat buffer. Figure 2 illustrates that notable signal intensity gain is achieved by decreasing the HCOOH concentration with a maximal intensity increase ca. threefold at the concentration of 0.01 % compared to 0.1 % HCOOH. However, separation problem occurred with further drop to 0% HCOOH, indicating the crucial role of HCOOH for the protonation of the basic catecholamines. To find the best solvent system, methanol and acetonitrile with 0.01 % HCOOH as mobile phase B were further examined, and methanol with 0.01 % HCOOH was found to be the optimal based on less elution power and greater signal response. Therefore, 0.01 % HCOOH in water and 0.01 % HCOOH in methanol were chosen as final mobile phase A and B, respectively. A gradient elution profile using the optimal mobile phases was established at a flow rate of 0.45 imL/min and catecholamines were eluted within 1 .7 min with a total run time of 7.0 min (Figure 3).

Optimization of mass spectrometry parameters

Since lack of sensitivity was the bottleneck for monitoring trace levels of catecholamines, mass spectrometry parameters were optimized to achieve greatest signal response, including MRM transitions, gas source variables, ion spray voltage, source temperature, etc. For example, to maximize the signal response, various MRM transitions including these reported were scrutinized under the conditions according to the present invention. After careful evaluation, MRM transitions 152.0-107.0, 166.0-107.0 and 154.0-91 .0 were used for NE, E and DA quantification, respectively (Table 1 ). Representative MRM chromatograms for an apparently healthy individual are illustrated in Figures 4 A to C. To further improve sensitivity for the measurement of low levels of E and DA, introduction of extra sample preparation steps or/and purchase of new instruments with advanced technology are typically needed, thus a much more complicated and expensive analysis is expected. In contrast, summation of multiple MRM transitions has been emerged as a valuable and direct approach to enhance sensitivity in LC-MS/MS assays. Threefold and 1 .5-fold sensitivity increase have been reported by using MRM summation for analysis of a polyethylene glycated protein and measurement of serum estradiol, respectively. In this invention, the application of the summation of five MRM transitions yields three to fourfold rise in sensitivity compared to the corresponding single MRM transition, and substantially reduces the LLOQ thus allows the simultaneous quantification of the trace levels of catecholamines in PBMC (Figure 4 D to F). A possible reason of the sensitivity enhancement is because the increase of the analyte peak area is greater than the gain for the randomized noise around the retention time of the analyte peak. The notable sensitivity gain achieved by the simple MRM summation approach demonstrated in this work provides a strong support to the suitability of this simple approach as an alternative strategy to address sensitivity and selectivity challenges.

Example 3:

Assay validation results

1 . Linearity, LOD and LLOQ

As illustrated in Table 2, the developed assay exhibited a wide dynamic linear range from 5.0 to 2500 pg/mL for NE and DA, and 1 .0 - 2500 pg/mL for E with a mean r≥ 0.9996 for all the analytes. The mean slope of the calibration curves at five different days was 0.0368, 0.0223 and 0.0477 for NE, E and DA, respectively. The LLOQ for NE, E and DA was determined to be 5.0, 1 .0 and

5.0 pg/mL with the respective LOD of 2.0, 0.5 and 2.0 pg/mL. Of significance, 250 - 2500 fold sensitivity enhancement of E was achieved comparing with the sensitivity of 2.5 ng/mL and 0.25 ng/mL by the previously reported PBA- based SPE protocols. The sensitivity obtained by the method of the present invention represents the most sensitive LC-MS/MS method for catecholamine analysis in biological samples. The ultra-sensitivity for E and DA obtained in this invention allowed the accurate quantitation of extremely low abundant catecholamines in human PBMC lysate by LC-MS/MS for the first time. Table 2: Linearity and calibration curve characteristics (n = 5)

Linearity range Slope Intercept r

Analyte

(pg/mL) Mean SD CV (%) Mean Mean SD CV (%)

NE 5.0 - 2500 0.0368 0.001 6 4.31 -0.0473 0.9996 0.0003 0.03

E 1 .0 - 2500 0.0223 0.0005 2.07 0.01 64 0.9996 0.0001 0.01

DA 5.0 - 2500 0.0477 0.0035 7.44 -0.1 086 0.9998 0.0001 0.01 Selectivity

Figure 3 illustrates that E is completely separated by chromatography from normetanephrine, which is crucial for the accurate quantitation since the isobaric compounds cannot be differentiated by MS detection. Figure 4 D to F shows representative MRM chromatograms of catecholamines in PBMC for a relative healthy donor. No interfering peaks for the real biological matrix were observed at the retention times of the catecholamines and ISs, suggesting the high selectivity of the assay.

Precision and accuracy

Table 3 presents the results of the assay precision and accuracy obtained from four concentration levels of QC samples on three different days. As can be seen in Table 3, the intraday assay precision and accuracy for NE, E and DA ranges from 1 .5 to 8.7% and 95.9 to 1 1 1 .3%, respectively. The respective interassay precision and accuracy were in the range of 0.0 - 5.5% and 91 .8 - 108.5%. In addition, the accuracy of the developed method was evaluated by performing recovery experiments of spiking catecholamines at three levels in different PBMC samples (n = 5), and the range of 97.1 to 105.4% is illustrated in Table 4. These data lent a strong support to the reproducibility and reliability for the accurate determination of all the analytes in PBMC samples either within the same day or on different days.

Table 3: Intraday and interday assay precision & accuracy

Intraday Interday sample

(pg/mL) Mean CV Accuracy Mean CV Accuracy

SD SD

(pg/mL) (%) (%) (pg/mL) (%) (%)

Blank 0 84.9 2.5 2.9 84.9 3.4 4.0

LQC 20 107.2 3.2 3.0 1 1 1 .3 106.1 5.7 5.4 105.9

NE MQC1 50 137.8 5.4 3.9 105.9 135.7 6.6 4.9 101 .6

MQC2 400 499.8 14.9 3.0 103.7 495.0 4.4 0.9 102.5

HQC 1600 1685.0 74.8 4.4 100.0 1680.0 31 .1 1 .9 99.7

Blank 0 2.7 0.3 10.0 - 2.6 0.1 2.2 -

LQC 5 7.7 0.7 8.7 101 .7 7.7 0.0 0.0 102.2

MQC1 20 22.2 0.7 3.2 97.7 22.6 0.5 2.1 99.9

MQC2 50 53.6 1 .6 3.0 101 .8 54.4 0.7 1 .3 103.6

HQC 400 421 .2 20.5 4.9 104.6 41 1 .7 15.3 3.7 102.3

Blank 0 4.6 0.4 7.9 - 4.9 0.6 12.6 -

LQC 20 23.8 0.7 2.9 95.9 23.3 1 .2 5.3 91 .8

DA MQC1 50 56.9 0.8 1 .5 104.5 54.9 3.0 5.5 100.0

MQC2 400 443.0 24.6 5.5 109.6 438.8 5.6 1 .3 108.5

HQC 1600 1683.3 33.9 2.0 104.9 1699.2 60.4 3.6 105.9

Table 4: Analytical recovery of catecholamines in different lots of PBMC (n

NE E DA

Added

(pg/mL) Measured CV Accuracy Measured CV Accuracy Measured CV Accuracy

(pg/mL) (%) (%) (pg/mL) (%) (%) (pg/mL) (%) (%)

40 40.7 4.1 101 .7 39.3 5.6 98.3 38.9 5.5 97.1

400 418.3 2.8 104.6 409.6 4.6 102.4 401 .6 4.5 100.4

1600 1686.3 2.2 105.4 1642.4 2.1 102.6 1600.0 2.3 100.0

Recovery of extraction procedure and matrix effect

Table 5 shows that the extraction recovery of 93.2 to 100.5%, 88.1 to 97.3% and 81 .0 to 86.9% for NE, E and DA is obtained at low, medium and high levels of catecholamines in five different PBMC samples. Good extraction efficiency is critical for reproducible and reliable measurement of trace amount of catecholamines in complicated biological samples albeit the extraction recovery was typically low for available LC-MS/MS methods. The significantly improved extraction efficiency achieved in this work is due to the combination of the specific PBA complexation and selective SPE on pElution HLB extraction under optimal condition.

Table 5: Extraction recovery in five different lots of PBMC

NE E DA

Added

(pg/mL) Mean Mean Mean

SD CV (%) SD CV (%) SD CV (%) (%) (%) (%)

40 100.5 9.3 9.3 88.1 9.0 10.2 86.3 12.4 14.3

400 91 .6 10.2 1 1 .1 92.6 10.0 10.9 81 .0 10.7 13.2

1600 93.2 8.7 9.4 97.3 12.6 13.0 86.9 2.3 2.6

As can be seen from Table 6, slightly enhanced matrix effect is observed for NE and E, while substantial ion suppression is associated with DA. Highly enhanced matrix effects were reported for catecholamine measurement in plasma by other LC-MS/MS methods. Nonetheless, the observed matrix effect can be compensated by correction of IS, which is the unique advantage of the MS technology compared to other conventional bio-analytic techniques. The respective SSE value of 94.2%, 99.1 % and 105.2% were determined for NE,

E and DA, which suggested that the matrix was compensated by IS thus did not affect the quantitation of catecholamines.

Table 6: Matrix effect in five different lots of PBMC

NE E DA

Added

(pg/mL) Mean Mean Mean

SD CV (%) SD CV (%) SD CV (%) (%) (%) (%)

40 1 18.5 15.3 12.9 126.4 20.5 16.2 58.0 8.0 13.8

400 1 1 1 .0 6.3 5.7 122.6 28.9 23.6 57.5 5.8 10.1

1600 103.5 7.6 7.3 125.2 14.7 1 1 .8 50.0 6.6 13.3 Carryover effect

To examine the carryover effect between two consecutive injections, an extracted highest standard sample and a blank sample were successively injected. No signal response enhancement was observed, suggesting no carryover effect of the assay.

6. Stability The stability study was designed to cover the anticipated conditions for handling real PBMC samples. Table 7 presents that less than 10% of catecholamine concentration change is observed under the following conditions: 4 days at 20 ^Q C, 4 days at 4 ^Q C, 3 weeks at -80 ^Q C and 4 freeze/thaw cycles. Given the acceptance criterion of ±15% difference, the analytes were considered stable under the various storage conditions.

Table 7: Stability of catecholamines in PBMC lysate

Stability NE E DA

Mean (pg/mL) CV Change Mean (pg/mL) CV Change Mean (pg/mL) CV Change

±SD ^a (%) (%) ±SD ^a (%) (%) ±SD ^a (%) (%)

Fresh (0 day) 35.1 ±1.3 3.8 19.3 ±0.4 2.2 14.6 ±0.4 2.4

1425.0 ±7.1 0.5 - 1345.0 ±63.6 4.7 - 1295.0 ±21.2 1.6 -

20 ^a C (4 days) 34.4 ± 0.8 2.3 -2.1 19.0 ±1.3 6.7 -1.6 14.1 ±0.3 2.0 -4.1

1500.0 ±0.0 0.0 5.3 1290.0 ±70.7 5.5 -4.1 1260.0 ±42.4 3.4 -2.7

4 ^a C (4 days) 34.4 ±2.1 6.0 -2.1 19.9 ±1.4 7.1 3.1 14.6 ±0.4 2.4 -1.0

1490.0 ±70.7 4.7 4.6 1350.0 ±56.6 4.2 0.4 1335.0 ±21.2 1.6 3.1

-80 ^a C (3 weeks) 34.4 ± 1.0 1.6 -2.0 19.5 ±1.0 1.6 0.8 13.5 ±0.4 3.1 -8.2

1305.0 ±21.2 2.0 -8.4 1275.0 ±21.2 2.0 -5.2 1195.0 ±77.8 6.5 -7.7

Freeze-thaw (4 Cycles) 36.1 ±0.8 2.2 2.7 19.3 ±0.4 1.8 -0.3 15.4 ±0.4 2.3 4.4

1465.0 ±21.2 1.4 2.8 1350.0 ±28.3 2.1 0.4 1275.0 ±21.2 1.7 -1.5 ^a SD, standard deviation; %change = ((mean level at each condition - mean value at day 0)/mean value at day 0) ^* 100 Example 4:

Assay application

To assess the applicability of the method of the present invention for the quantitative analysis of authentic PBMC samples, forty relatively healthy people were recruited with written consent and PBMC lysate samples were analyzed. They were 31 women and 9 men ranging from 23 to 62 years-old with a median age of 39 years- old. No catecholamine-related medications, hormonal therapies, vigorous exercise, alcohol and excessive liquid intake were allowed. The study was performed following a protocol approved by internal clinical ethics committee.

The parametric reference interval (95 ^th percentile) was determined to be 0.0658 to 0.8063 pmol/10 million cells (corresponding to 44.5 to 545.1 pg/mL in 40 million cells/mL) for NE, 0.0020 to 0.0363 pmol/10 million cells (1 .5 to 26.5 pg/mL) for E and 0.0033 to 0.0493 pmol/10 million cells (1 .5 to 30.1 pg/mL) for DA, respectively. The determined NE value is in accordance with earlier reported results, whereas the E and DA values are considerably lower than those of reported in the literature. This indicates that the HPLC-ECD analysis may have been affected by the endogenous interfering compounds, thus high selectivity and sensitivity provided by LC-MS/MS is necessary to quantify the trace levels of catecholamines in PBMC. Of note, the concentrations of catecholamines in PBMC lysate determined in this invention agree with the general trend reported in plasma, whereas substantially higher levels of E and DA in PBMC were obtained by the previous HPLC-ECD detection. Additionally, a small amount of blood (< 0.5 mL) typically is needed for catecholamine analysis in plasma, in contrast, approximately 10 mL blood (20 fold) is required for the detection of catecholamines in PBMC. The larger volume of blood for PBMC assay poses another challenge on the sensitivity of the PMBC assay, and thus demands an ultra-sensitive LC-MS/MS assay. Example 5:

Optimization of LC conditions

One of challenges for catecholamine and metanephrine determination is the poor retention of catecholamines on the reversed-phase C18 columns due to their high polarity, especially NE was fast eluted on the column front. Another difficulty is the isomeric E and NMN have to be baseline separated on LC column to avoid cross contamination because they cannot be distinguished only by MS. In a previous LC- MS/MS catecholamine assay in urine, a Restek PFPP column (150 mm, 2.1 mm, 3 μίπ) was used, however, this column had severe column to column variations regarding retention time shift and peak shape, which was known for the PFP-column thus made it impossible to be implemented into routine testing. To address the aforementioned challenges, a variety of commercial PFP-type columns were tested, including ultra PFPP column (150 mm, i.d. 2.1 mm, 3 μιη) and (100 mm, i.d. 2.1 mm, 3 μιη) from Restek (Bellefonte, PA, USA), Luna PFP (2) column (150 mm, i.d. 2.1 mm, 3 μιη), Kinetex (150 mm, i.d. 2.1 mm, 2.6 μιη) and F5 (150 mm, i.d. 2.1 mm, 3 μιη) from Phenomenex (Torrance, CA, USA), and Hypersil PFP column (150 mm, i.d. 2.1 mm, 3 μιη) from Thermo Scientific. The major criteria for the column selection was the retention for NE, separation of E and NMN, signal response and back- pressure, etc. Luna PFP (2) column (150 mm, i.d. 2.1 mm, 3 μιη) was found to be the best column (data not shown). After additional optimization of LC solvent and temperature, the five analytes were completely separated within 2.3 min with a total run time of 5.5 min, which was faster than the reported methods using PFP column.

Example 6:

Optimization of MS conditions

Based on the low E level in urine in physiological conditions, particularly in the early morning, a sensitivity higher than 1 ng/mL was generally required, which most available assays lacked. On the other hand, relatively high concentration of DA in healthy and metanephrines in tumor population were expected. However, narrow linearity ranges were often reported, which may have to dilute the high levels of samples for re-testing, resulting in extra-time and cost. Therefore, to enhance the sensitivity for E and obtain high dynamic range for DA and metanephrines to achieve one single run covering all the levels of analytes in urine samples, mass spectrometric parameters were optimized systematically. For this purpose, individual MRM transitions, gas source variables, ion spray voltage and source temperature were assessed. To maximize the signal response for detection of E, numbers of MRM transitions were carefully compared, including the original quantifier m/z 184.0 > 107.0 and qualifier m/z 184.0 > 107.0 and water loss Q1 of m/z 166.0 associated different fragmentations, such as m/z 166.0 > 107.0 and m/z 166.0 > 135.0. The MRM m/z 166.0 > 107.0 was selected as the quantifier for E according to the most intensive response. For DA, NMN and MN, special attention was paid to achieve at least 1250 ng/mL upper level of limit without signal saturation. The optimal MRM quantifiers for NE, E, DA, NMN and MN were determined to be m/z 152.0 > 107.0, 166.0 > 107.0, 154.0 > 91 .0, 166.0 > 121 .0 and 180.0 > 165.0, respectively. Detailed MRM parameters are provided in Table 1 B.

Example 7:

Optimization of SPE conditions

To address the issues of non -satisfactory extraction efficiency and/or substantial matrix effect for some analytes, pooled urine samples were used for the optimization of several important SPE parameters on a 96-well HLB Elution plate, including washing solution, urine loading volume and PBA buffer.

1 . Effect of washing solution

An effective SPE washing solution would be as strong as possible to wash away other interfering compounds in the complicated urine matrix while not strong enough to affect the analyte of interest on the sorbent. Three different washing solutions with 10%, 20% and 30% of MeOH in NH ₄ CI solution (pH 8.5) were assessed in 50 μΙ_ of pooled urine samples. As illustrated in Figure 6 A, decreasing the strength of the washing solution from 30% to 20% MeOH content exhibits considerable influence on the extraction recovery of DA, NMN and MN according to the respective 19.6%, 31 .9% and 28.3% increase of the mean recovery, whereas it slightly affects extraction recovery for NE and E based on a < 6% increase. Further decreasing the MeOH from 20% to 10% gives minimal improvement of the recovery (Figure 6 A), however, the corresponding matrix effect for NE starts to increase substantially (7.3%) although other analytes are not affected (Figure 6 B). Based on these observations, the washing solution with 20% MeOH in NH ₄ CI solution was selected as the optimal washing for the following optimizations. Effect of sample volume

The need of pH adjustment of urine samples during the sample preparation as reported in the available literature made the clean-up process cumbersome and time-consuming, which is one of the most important drawbacks. This is due to the relatively large variation of pH in human urine and the high urine volume used (typically over 500 μΙ_). It was hypothesized that the pH variation caused by the individual urine sample could be negligible if very small amounts of urine samples were applied for the sample preparation, given it has sufficient sensitivity. To test this hypothesis, 10 μί of urine were compared with 50 μΙ_ for the SPE process. The extraction recovery data is illustrated in the Figure 6 C. As can be seen, comparable or slightly better recovery has been achieved by the decrease of the urine amount. Therefore, the minimized amount of 10 μΙ_ of urine was chosen. Effect of PBA buffer

The concentration of PBA played a vital role for the formation of complex between the catechol moieties with the PBA reagent. The PBA amount was investigated as well. Thus, three different amount of PBA were tested using 10 μΙ_ of pooled urine samples, including: 250 μΙ_ PBA, 100 μΙ_ PBA + 150 μΙ_ water and 50 μΙ_ PBA + 200 μΙ_ water. As exhibited in Figure 6 D, the extraction recovery of NE, E and MN is comparable under the three PBA conditions, whereas using 50 μΙ_ provides substantial higher recovery for NE and DA. Hence, the optimal volume of PBA was determined to be 50 μΙ_. The final optimized SPE condition was described above. Briefly, a mixture of 10 pL sample, 100 μΙ_ of IS working solution, 50 μΙ_ 0.2% PBA buffer (pH 8.5) and 100 μΙ_ of 0.7% NH40H solution was loaded on a pre-conditioned Oasis HLB

96-well pElution extraction plate and washed three times with 200 μΙ_ 20% MeOH in NH4CI buffer before eluated with 100 μΙ_ of 0.2M HOAc.

Due to the minimized urine usage, the contribution of the urine to the pH of the sample mixture was negligible, thus eliminating the tedious pH check and adjustment. This is the first report that pH adjustment is avoided for SPE of urine samples for the catecholamine determination, which has been considered as the bottleneck of the sample preparation and/or the entire catecholamine determination in routine analysis. Additionally, with the reduced sample volume, the associated IS usage was decreased accordingly, which could make the analysis less expensive because of the high cost of isotopically-labeled ISs. Meanwhile, considering much diluted matrix for each injection, the LC column lifetime was expected to be longer than under the regular conditions. Furthermore, the solvent and waste were notably reduced by applying the HLB microplate compared to the regular 1 imL cartridge or 96- well plate. Therefore, the proposed assay is beneficial for the advancement of the current status of catecholamine and metanephrine routine analysis in urine, leading to a simple, rapid, sensitive, cost-effective and robust assay.

Example 8:

Method validation

Linearity, LLOQ and LOD

As demonstrated in Table 8 below, the extracted calibration curves are linear over the concentration range of 2.5 - 500 ng/ml for NE, 0.5 - 500 ng/ml for E and 2.5 - 1250 ng/ml for DA, NMN and MN with mean correlation coefficient (r) greater than 0.9995. The CV% of the slope for the analytes are within the range of 1 .18 - 2.85% with a narrow standard deviation (SD) < 0.00222.

Table 8: Linearity, calibration curve parameters, LOD and LLOQ (n

Linearity range Slope Intercept r LOD LLOQ

Analyte

(ng/mL) Mean SD CV (%) Mean Mean (ng/mL) (ng/mL)

NE 2.5 - 500 0.1078 0.00222 2.06 0.01 80 0.9995 1 .0 2.5

E 0.5 - 500 0.0962 0.001 13 1 .1 8 0.0024 0.9998 0.25 0.5

DA 2.5 - 1250 0.01 00 0.00029 2.85 0.0012 0.9997 1 .0 2.5

NMN 2.5 - 1250 0.0042 0.0001 0 2.35 -0.0002 0.9999 1 .0 2.5

MN 0.5 - 1250 0.0070 0.00012 1 .76 0.0006 0.9998 0.25 0.5

Table 8 illustrates that the LODs for NE, DA and NMN are 1 .0 ng/mL, whereas 0.25 ng/mL LODs are determined for E and MN. The LLOQs for NE, DA and NMN are 2.5 ng/mL, and 0.50 ng/mL for E and MN. It is noteworthy that the majority of available assays have relatively high quantitation limit for E and/or a narrow linearity range of DA. An LLOQ below 1 .0 ng/mL for E and a wide linearity range for DA are necessary to quantify low level of E and high level of DA in human urine demonstrated by the reference range. The sensitivity for E and wide dynamic range for DA in the proposed assay enables the simultaneous analysis of all the analytes in urine samples in a single analysis.

Precision and accuracy

Tables 9 and Table 10 below illustrate that the intraassay and interassay imprecisions for the spiked QC samples are < 9.4%, and < 5.2% are for the two levels of commercial QC samples. The intra-recoveries for the spiked QC samples are determined to be in the range of 97.2% - 1 12.5% with a mean recovery of 102.2%, whereas inter-recoveries are between 95.9% and 104.0% with a mean recovery of 100.4%. As for the commercial QC samples, 84.3% - 108.7% intra-recoveries with a mean recovery of 102.2% and 85.8% - 108.3% intra-recoveries with a mean recovery of 102.2% are observed. Finally, 97.9% - 1 14.5% recovery with a mean recovery of 105.3% in the five random urine samples at three concentration levels further confirmed the assay to be precise and accurate. Table 9: Intraday and interday assay precision & accuracy (spiked QC samples)

Intraday Interday

Analyte Sample Added

(ng/mL) Mean acy Mean

SD CV Accur

SD cv Accuracy (ng/mL) (%) (%) (ng/mL) (%) (%)

LLOQ 2.5 2.6 0.2 7.8 103.4 2.4 0.2 7.2 95.9

LQC 5.0 5.3 0.1 2.2 106.3 5.1 0.2 3.9 101 .9

NE

MQC 25 24.3 0.7 2.9 97.2 24.9 0.6 2.3 99.5

HQC 400 391 .8 18.3 4.7 98.0 390.1 15.6 4.0 97.5

LLOQ 0.5 0.5 0.0 8.1 105.8 0.5 0.0 6.0 99.2

LQC 2.5 2.5 0.1 3.8 100.2 2.5 0.1 3.6 98.3

MQC1 5.0 5.1 0.2 4.5 101 .9 5.0 0.1 2.4 100.9

MQC2 25 25.7 0.8 3.0 102.8 25.1 0.5 2.2 100.3

HQC 400 409.8 5.9 1 .4 102.5 405.2 6.8 1 .7 101 .3

LLOQ 2.5 2.8 0.2 7.0 1 12.5 2.6 0.2 7.1 104.0

LQC 5.0 5.1 0.3 6.7 102.6 5.0 0.2 3.0 99.3

MQC1 25 25.6 0.8 3.3 102.4 25.3 0.3 1 .1 101 .3

MQC2 400 412.4 9.2 2.2 103.1 409.3 4.2 1 .0 102.3

HQC 1000 984.0 21 .3 2.2 98.4 987.2 1 1 .9 1 .2 98.7

LLOQ 2.5 2.6 0.2 7.8 105.0 2.5 0.1 3.2 101 .3

LQC 5.0 5.0 0.2 5.0 100.0 5.1 0.2 4.8 102.1

NMN MQC1 25 24.5 0.4 1 .8 97.9 24.8 0.8 3.4 99.3

MQC2 400 399.4 14.0 3.5 99.9 397.5 5.2 1 .3 99.4

HQC 1000 1016.6 28.5 2.8 101 .7 999.4 21 .3 2.1 99.9

LLOQ 0.5 0.5 0.0 9.4 104.3 0.5 0.0 2.8 102.1

LQC 2.5 2.6 0.1 4.6 103.3 2.6 0.1 5.2 102.2

MQC1 25 25.4 0.3 1 .1 101 .7 25.3 1 .0 4.1 101 .1

MQC2 400 405.0 6.8 1 .7 101 .3 406.4 10.0 2.5 101 .6

HQC 1000 1008.4 26.9 2.7 100.8 1006.2 4.7 0.5 100.6

Table 10: Intraday and interday assay precision & accuracy (commercial external

QC samples)

Expected Expected Intraday Interday

Analyte Sample Range Value

Mean CV Acc Mean CV Acc (ng/mL) (ng/mL) SD SD

(ng/mL) (%) (%) (ng/mL) (%) (%)

Normal 51 .0-76.6 63.8 56.7 3 5.2 88.8 57.1 1 .1 1 .9 89.5

NE

Pathological 163-245 204 172 2.4 1 .4 84.3 175 5.5 3.1 85.8 p Normal 9.1 -13.7 1 1 .4 1 1 .1 0.4 3.9 97.4 1 1 0.1 0.9 96.8

Pathological 39.4-59.0 49.2 48.7 0.7 1 .4 99.1 47.4 1 .3 2.8 96.4

Normal 153-230 192 200.8 6.4 3.2 103.4 196.3 3.9 2 102.2

DA

Pathological 341 -512 426 463.2 8.5 1 .8 108.7 461 .5 8.8 1 .9 108.3

Normal 212-318 265 236.4 9.8 4.2 89.2 232.4 3.8 1 .7 87.7

NMN

Pathological 767-1 151 959 904.6 20.1 2.2 94.3 899.4 44.9 5 93.8

Normal 1 17-175 146 137.6 3.6 2.7 94.2 134.6 3.1 2.3 92.2

MN

Pathological 278-418 348 348.6 6.1 1 .8 100.2 345.9 12.9 3.7 99.4

Specificity

To eliminate possible interference between the isomeric E and NMN, chromatographic condition was optimized to achieve the baseline chromatographic separation of the isomers. An additional interference study was performed to assess 34 substances as listed in Table 1 1 below, including their structurally-related metabolites, such as 3,4-Dihydroxyphenylacetic acid (DOPAC), Homovanillic acid (HVA), 3,4-Dihydroxyphenylglycol (DHPG), 3,4- Dihydroxymandelic acid (DOMA), Vanillylmandelic acid (VMA), and common drugs or supplements such as Labetalol and Paracetamol reported to interfere with HPLC-ECD assays. In this regard, the potential interfering compounds at 20 mg/L were spiked into pooled urine samples, and the spiked samples were processed alongside with the non-spiked urine samples. Less than 10% difference of the measured concentrations in the spiked urine samples compared with the non-spiked urine samples indicated that the quantitation was not affected by the variety of substances. Table 1 1 : Specificity and potential interferences

Potential interference Mw Potential interference Mw

Structurally similar compounds

3,4-Dihydroxyphenylglycol (DHPG) 170.2 Metanephrine (MN) 197.2

3,4-Dihydroxymandelic acid (DOMA) 184.2 4-Hydroxy-3-methoxy-phenylglycol (M HPG) 184.2

L-DOPA 197.2 3-Methoxytyramine (3-MT) 167.2

3,4-Dihydroxyphenylacetic acid (DOPAC) 168.2 Norepinephrine (NE) 169.2

Dopamine (DA) 153.2 Normetanephrine (NMN) 183.2

Epinephrine (E) 183.2 2-Phenylethylamine 121 .2

5-Hydroxyindoleacetic acid (5-HIAA) 191 .2 Serotonin 176.2

Homovanillic acid (HVA) 182.2 Tryptamine 160.2 a-Methyldopamine 167.2 Tyramine 137.2

Methyldopa 21 1 .2 Tyrosine 181 .2 a-Methylnorepinephrine 183.2 Vanillylmandelic acid (VMA) 198.2

Drugs or supplements

Acetylsalicylic Acid 180.2 Octopamine 153.2

Caffeic acid 180.2 Paracetamol 151 .2

Caffeine 194.2 Phenylalanine 165.2

Carbidopa 226.2 Phenylephrine 167.2

Isoproterenol 21 1 .3 Theophylline 180.2

Labetalol 328.4 Tryptophan 204.2

4. Extraction efficiency and matrix effect

The extraction efficiency and matrix effect are critical parameters for reproducibility and reliability of routine analysis of catecholamines and metanephrines in complicated urine samples, which were often problematic in the available assays. Herein the factors were investigated by the traditional post-extraction method proposed in the art. For this purpose, five different urine samples were spiked with the combined calibrators at three different concentration levels, and the spiked urine samples were extracted and analyzed along with the non-spiked urine samples. As shown in Table 12 below, good extraction recoveries of 90.5 - 96.3%, 89.2 - 93.7%, 89.9 - 97.3%, 74.1 - 82.4%, 79.3 - 82.8% for NE, E, DA, NMN and MN are achieved at the three concentration levels in the five urine samples. Table 12: Extraction recovery, matrix effect and analytical recovery in five random urines

10 91 .4 16.9 18.5 86.1 5.6 6.5 103.3 4.9 4.7

NE 50 96.3 1 1 .5 1 1 .9 84.1 5.1 6.1 100.2 6.6 6.6

500 90.5 5.9 6.6 85.4 1 .3 1 .5 99.3 2.5 2.5

10 89.2 7.9 8.9 1 1 1 .8 2.8 2.5 105.5 5.3 5.0

E 50 92.9 4.1 4.5 108.9 6.5 6.0 105.5 5.2 4.9

500 93.7 2.4 2.5 95.4 4.0 4.2 109.1 7.3 6.7

13.

10 97.3 9.3 9.6 108.2 16.3 15.1 97.9 12.7

0

DA 50 93.1 7.4 8.0 106.4 4.7 4.4 107.0 8.9 8.3

500 89.9 9.6 10.7 95.8 6.5 6.7 100.3 4.5 4.5

10 82.4 5.4 6.6 1 19.0 6.7 5.7 1 14.5 4.5 3.9

NMN 50 78.4 0.6 0.8 1 18.3 2.4 2.0 106.1 5.2 4.9

500 74.1 4.5 6.1 1 10.5 5.0 4.5 102.5 3.3 3.2

10 82.8 6.6 8.0 106.3 7.0 6.6 1 1 1 .6 8.6 7.7

MN 50 79.3 2.8 3.5 109.6 5.8 5.3 109.4 4.9 4.5

500 79.3 6.7 8.5 107.9 3.4 3.1 107.9 3.2 3.0

Table 12 exhibits that the developed assay has minimal matrix effect according to the 84.1 - 1 1 1 .8% mean matrix effect for catecholamines and 106.3 - 1 19.0% for metanephrines. The high extraction recovery and negligible matrix effect are partially attributed to the dramatically reduced urine volume and effective HLB extraction on a microplate sample upon the specific PBA complexation.

5. Carryover effect

The possible autosampler carryover between two consecutive injections was assessed by sequential injections of extracted highest standard urine sample followed by a blank urine sample. Negligible signal response enhancement was obtained in the blank sample, revealing no carryover effect with the method.

Analyte stability

The stability study was designed based on the conditions to be encountered during routine sample handling of authentic urine samples. Freshly prepared urine samples at low and high concentration levels were aliquoted and reanalyzed after storage at 20 ^Q C and 4 ^Q C for seven days and 30 days at -20 ^Q C. To evaluate the freeze-thaw stability of the analytes, the samples were measured after four cycles of thawing and freezing. The acceptance criterion was ±15% difference between the analyte concentration determined in duplicate on the day of collection and the specific storage condition. As illustrated in Table 13 below, the change of all the analyte concentrations is below 7.9% under all the tested conditions, indicating satisfactory stability for the analytes in urine samples under these conditions according to the criterion of ±15% change. In addition, maximal 8.8% concentration change is observed in the post-extract solution in 4 ^Q C autosampler within 48 hours, enabling possible sample re-test in routine without the need to re-perform SPE.

Table 13: Analyte stability

NE E

Stability

Mean % Mean %

SD CV% SD CV%

(ng/mL) Change (ng/mL) Change

Fresh (0 day) 71.8 1.3 1.9 - 53.3 1.1 2.0 -

410.0 2.8 0.7 - 419.5 16.3 3.9 -

20 ^S C (7 days) 69.3 1.7 2.4 -3.5 52.5 2.1 4.0 -1.5

419.5 2.1 0.5 2.3 436.5 12.0 2.8 4.1

4 ^S C (7 days) 70.1 2.5 3.6 -2.4 53.9 0.3 0.5 1.1

417.0 35.4 8.5 1.7 452.5 6.4 1.4 7.9

-20 ^S C (30 days) 74.7 4.9 6.6 4.0 53.9 1.9 3.5 1.0

381.0 9.9 2.6 -7.1 430.5 3.5 0.8 2.6

Freeze-thaw 71.1 4.7 6.6 -1.0 56.7 2.1 3.6 6.3

(4 Cycles) 420.0 21.2 5.1 2.4 440.5 0.7 0.2 5.0

4 ^S C Post-extract 69.2 1.7 2.5 -3.6 54.7 2.0 3.6 2.6

(48 hours) 398.5 17.7 4.4 -2.8 431.5 9.2 2.1 2.9

DA NMN

Stability

Mean % Mean %

SD CV% SD CV%

(ng/mL) Change (ng/mL) Change

Fresh (0 day) 241.0 5.7 2.3 - 58.2 1.8 3.2 -

1210.0 42.4 3.5 - 1025.0 35.4 3.4 -

20 ^S C (7 days) 242.5 7.8 3.2 0.6 62.6 1.4 2.3 7.6

1165.0 21.2 1.8 -3.7 1002.5 10.6 1.1 -2.2

4 ^S C (7 days) 253.5 6.4 2.5 5.2 60.6 0.6 0.9 4.1

1220.0 14.1 1.2 0.8 1025.0 21.2 2.1 0.0

-20 ^S C (30 days) 240.0 14.1 5.9 -0.4 59.2 2.6 4.4 1.6

1215.0 7.1 0.6 0.4 1006.5 19.1 1.9 -1.8

Freeze-thaw 249.0 14.1 5.7 3.3 59.8 1.1 1.9 2.7

(4 Cycles) 1150.0 42.4 3.7 -5.0 1020.0 0.0 0.0 -0.5

4 ^S C Post-extract 251.0 12.7 5.1 4.1 63.3 0.8 1.3 8.8

(48 hours) 1140.0 14.1 1.2 -5.8 1020.0 14.1 1.4 -0.5 MN

Stability

Mean %

SD CV%

(ng/mL) Change

Fresh (0 day) 65.0 1 .3 2.1 /

1045.0 49.5 4.7 /

20 ^S C (7 days) 64.5 1 .8 2.9 -0.8

1045.0 7.1 0.7 0.0

4 ^S C (7 days) 64.0 0.5 0.8 -1 .6

1045.0 21 .2 2.0 0.0

-20 ^S C (30 days) 66.4 2.1 3.1 2.1

1055.0 7.1 0.7 1 .0

Freeze-thaw 66.3 2.0 3.0 2.0

(4 Cycles) 1050.0 14.1 1 .3 0.5

4 ^S C Post-extract 69.5 1 .0 1 .4 6.9

(48 hours) 1040.0 42.4 4.1 -0.5

Application of the method to biological samples

To evaluate the practical feasibility of the developed method for the quantitative analysis of authentic urine samples, 88 second-morning urine samples after overnight fasting of apparently healthy people were processed under the optimal conditions as aforementioned. They were 49 women and 39 men with the age from 18 to 72 years-old with a median age of 41 years-old. The participants were refrained from taking catecholamine-related medications, hormonal therapies, vigorous exercise, alcohol and excessive liquid intake. The study was conducted according to a protocol approved by internal clinical ethics committee.

Using the EP evaluator release 9, the parametric reference interval (95th percentile) was determined to be 5.9 - 105.7 ng/mL with a mean of 38.5 ng/mL (8.8 - 55.0 μg/g Cr with a mean of 26.3 μg/g Cr) for NE, 0.8 - 35.1 ng/mL with a mean of 8.0 ng/mL (0.6 - 17.2 μg/g Cr with a mean of 5.6 μg/g Cr) for E, 39.8 - 585.5 ng/mL with a mean of 253.2 ng/mL (65.7 - 283.7 μg/g Cr with a mean of 174.7 μg/g Cr) for DA, 4.9 - 60.7 ng/mL with a mean of 24.1 ng/mL (6.5 - 33.6 μg/g Cr with a mean of 17.1 μg/g Cr) for NMN and 3.9 - 51 .3 ng/mL with a mean of 20.0 ng/mL (5.4 - 31 .1 μg/g Cr with a mean of 14.3 μg/g Cr) for MN, respectively. The values determined by the developed method are in agreement with previous reported data, which lend further support for the validity of the method in routine analysis.

Example 9:

Determination of catecholamines in plasma

Accurate quantitation of plasma catecholamine remains a difficult analytical challenge due to their extremely low endogenous levels. In this study, the SPE and LC-MS/MS conditions for PBMC were applied to plasma samples. Preliminary results are summarized in Table 14 below.

Table 14: Preliminary recovery of catecholamines in blank plasma

Analyte Added (pg/mL) Accuracy (%)

NE 40 97.8

400 101.5

E 40 104.6

400 100.7

DA 40 85.5

400 83.5

Table 14 shows that the initial recovery of the catecholamines at two levels spiked in blank plasma. As can be seen, good accuracy is obtained: NE and E (97.8 - 104.6%), whereas ~ 85% is observed for DA. Additionally, to check the suitability in authentic plasma, 18 plasma of apparently healthy people were successfully determined by the method. The concentration range was determined to be 166 - 622 pg/mL with a mean of 377.5 pg/mL for NE, 12.2 - 45.3 pg/mL with a mean of 26.5 pg/mL for E and 7.2 - 34.2 pg/mL with a mean of 12.5 pg/mL for DA, which is in accordance with other reported ranges, These data indicate that the PBMC catecholamine assay is working for plasma.

Conclusion:

An integrated PBA-HLB-PFP approach was developed and validated to quantify endogenous catecholamines in human PBMC or urine. The method is advantageous over the reported methods due to its ultra-sensitivity and good selectivity obtained without the need of derivatization, evaporation and reconstitution, ion-paring reagents and HILIC column. The unique combination of PBA-complexation and SPE on a 96-well HLB Elution plate enabled purification and fivefold enrichment of catecholamines in a single and straightforward extraction. The utilization of a Luna PFP column with 0.01 % HCOOH as additive yielded optimal LC chromatographic separation with enhanced signal response. The summation of optimized MRM transitions further gained notable sensitivity, enabling simultaneous quantification of catecholamines in PBMCs with ultra-sensitivity as low as 1 pg/mL. The minimized urine sample volume (e.g., 10 μί) significantly reduced the pH variation derived from urine sample, which eliminated the laborious and risky- associated pH adjustment in the reported assays, and thus significantly improved the efficiency of sample preparation and throughput of the entire assay. The multiplexing assay of five catecholamines and metanephrines in a single run facilitated by the good sensitivity for E and a wide linearity for DA and metanephrines substantially reduced the cost per analyte and aided better clinical interpretation. Additionally, the internal standard, reagent, solvent consumption and waste disposal were reduced due to the miniature SPE platform, substantially reduced the analysis cost, which would be beneficial for clinical routine analysis. The excellent validation results paired with successful application in the authentic samples justify the reliability and suitability of the first LC-MS/MS method to address the challenging measurement of catecholamine in PBMC or urine, which will be beneficial for a better understanding of the new arena of neural-immune communication and routine analysis. Additionally, the integrated platform of the present invention, being simple, fast, sensitive and selective, can be explored to quantify catecholamines and metabolites in other biological matrices (e.g. plasma, serum, tissue), which will have a great impact in broad fields of clinical, pharmaceutical, neuroscience and basic research.

Abbreviations:

CE Collision energy

CV Coefficient of variation

CXP Collision cell exit potential

DA Dopamine

DP Declustering potential

DPBS Dulbecco's phosphate buffered saline

E Epinephrine

EP Entrance potential

ESI Electrospray ionization

HILIC Hydrophilic interaction chromatography

HLB Hydrophilic-lipophilic-balanced

HPLC High-performance liquid chromatography

HPLC-ECD High-performance liquid chromatography combined with electrochemical detection

IS Internal standard

LC Liquid chromatography

LC-MS Mass spectrometry coupled with liquid chromatography

LC-MS/MS Tandem mass spectrometry coupled with liquid chromatography

LLOQ Lower limit of quantification

LOD Limit of detection

MN Metanephrine

MRM Multiple reaction monitoring

MS Mass spectrometry

MS/MS Tandem mass spectrometry

NE Norepinephrine

NMN Normetanephrine

PBA Phenylboronic acid

PBMC Peripheral blood mononuclear cells

PFP Pentafluorophenyl

QC Quality control SD Standard deviation

SPE Solid phase extraction

SSE Signal suppression or enhancement

UPLC Ultra performance liquid chromatography

UPLC-MS/MS Tandem mass spectrometry coupled with ultra-performance liquid chromatography

WCX Weak cation exchange