Multiplexing Determination of Small Cell Lung Cancer Biomarkers and

Jun 19, 2014 - A multiplex method for the determination of the small cell lung cancer (SCLC) markers progastrin releasing peptide (ProGRP) and neuron ...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/ac

Multiplexing Determination of Small Cell Lung Cancer Biomarkers and Their Isovariants in Serum by Immunocapture LC-MS/MS Silje B. Torsetnes,† Maren S. Levernæs,† Marianne N. Broughton,‡ Elisabeth Paus,‡ Trine G. Halvorsen,† and Léon Reubsaet*,† †

Department of Pharmaceutical Chemistry, School of Pharmacy, University of Oslo, Oslo, NO-0316, Norway Department of Medical Biochemistry, Oslo University Hospital, Radiumhospitalet, Oslo, 0310, Norway



S Supporting Information *

ABSTRACT: A multiplex method for the determination of the small cell lung cancer (SCLC) markers progastrin releasing peptide (ProGRP) and neuron specific enolase (NSE) is presented, which involves coextraction by immunoaffinity (IA) beads and codetermination by selected reaction monitoring (SRM). The performance was compared with two IA SRM methods which were recently validated for individual marker determination. The multiplexing method reduces sample volume, handling time per sample, and reagent consumption and shows good linearity, recovery, quantitative measurements, and sensitivity with lower limit of detection (LLOD) values of 7.2 pM (=90 pg/mL) and 4.5 pM (=210 pg/mL) and lower limit of quantitation (LLOQ) values of 24 pM (=300 pg/mL) and 15 pM (=700 pg/mL), for total ProGRP and γ-NSE, respectively. The novel aspect of this approach is the multiplexing of ProGRP and NSE with the additional ability to perform fingerprinting by the selective determination of ProGRP isoform 1, ProGRP isoform 3, and total ProGRP, as well as the α- and the γ-subunit of NSE isoenzymes. Six serum samples from patients with SCLC were analyzed to demonstrate the methods feasibility to simultaneously differ between and individually quantify ProGRP, NSE, and their isoform and isoenzyme variants, respectively. Both the presence of and variation between all the isoforms and isoenzymes, as well as covarying results with the conventional immunometric assays for total ProGRP and γ-NSE, were seen in the analyses of patient serum samples.

M

for the mass spectrometer since it, either via SRM or high resolution MS, has a high degree of specificity.21−23 In addition, regarding the resources needed to raise high quality antibodies, the flexible MS technology is considered as a tool for relieving this bottleneck in marker verification throughput as evolvement in “omics” technologies continues to further the demand for interfaces between biomarker discovery and clinical validation. This is highly specific and enables both short lead time and multiplexing capacity. The main limitations of MS are the sensitivity and the general need for preanalysis purification. For proteins occurring at the μg/mL or nM concentrations, this is straightforward and little or no sample-enrichment is needed.24 However, most of the interesting biomarkers occur at the low to ultralow abundance level (fg−ng/mL or aM−pM level), making it challenging to measure them in complex biological matrixes.25 Thus, to exploit the mass sensitivity of the mass spectrometer and to enable sufficient sensitivity for biomarker determination, both sample cleanup and biomarker enrichment are generally necessary prior to analysis. Most successful is the use of immunoaffinity (IA)-

easurements of changes in biomarker concentration in biological samples are often used as support to other clinical tests for evaluation of pathological states.1 The clinical value of such measurements relates to its probability to discriminate between the target condition and health, which depends on its clinical sensitivity (true positive rate) and specificity (true negative rate).2,3 Investigations of multiple disease indicators are often used to increase reliability for clinical interventions. Similarly, combined determination of biomarkers with complementary pathological information may be of additional aid in the decision of clinical intervention and management of disease. For this purpose, bottom-up proteomic selected reaction monitoring (SRM)-based LC-MS/MS technology has been suggested as a putative biomarker tool due to certain limitations of immunometric assays.4,5 The immunometric assays are extensively used as clinical routine methods due to features of being fast, sensitive, and highly automatable. However, they have limitations: their general inability to differentiate between protein variants,6 the possibility of being affected by interferences,7−9 and the possible long lead time and high costs for development.6 With a targeted MS approach, proteins can also be determined at low concentrations,10−17 and in addition, it is possible to distinguish between their variants.10,12,18 Cross-reactivity, an interference that may be a problem in immunoassays,19,20 is of less relevance © XXXX American Chemical Society

Received: March 18, 2014 Accepted: June 19, 2014

A

dx.doi.org/10.1021/ac500986t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Table 1. Primary Structures of Reported Isovariants for ProGRP and NSEa

a Each letter represents one amino acid residue. The differing AA-residues between the two different subunits (α- and γ-enolase) for NSE isoenzymes and the three isoforms of ProGRP are marked green, and the chosen signature peptides are emphasized by frames. The sequences are according to UniProtKB/Swiss-Prot: P07492 (for all three isoforms), P06733.2 (α-enolase), and P09104.3 (for γ-enolase).

immunofluorometric assay (ProGRP TR-IFMA)29 and the immunoradiometric assay (NSE IRMA).30 Validated immunocapture LC-MS methods have earlier been presented for ProGRP8 and NSE.6 The ProGRP method was able to determine and quantify the reported forms for ProGRP; isoform 1, isoform 3, and total ProGRP.8 The method has also been used to study the isoform expression in patient serum samples.31 The NSE method was validated upon quantification of one of the subunits for NSE, the γ-enolase, while the other subunit, the α-enolase, could be identified.6 See Table 1 for the similarities and differences in the primary structure of isoforms of ProGRP and subunits of NSE isozymes. In this paper, a strategy for a multiplexing method is presented, where these two IA MS methods are fused into one single codetermining method. Our assumption is that this multiplexing method will simultaneously provide the conventional marker concentration information as well as all presented isoforms and isoenzymes for the two complementary SCLC markers. The clinical SCLC sensitivity of the separate markers are reported to be 72% and 67% (at 95% specificity), for ProGRP and NSE, respectively,32 and an improved sensitivity of parallel ProGRP and NSE measurement would be approximately 91% if calculated as described by Vitzthum et al.33 To our knowledge, this is the first multiplex IA MS method which allows conventional protein quantification as well as isovariant quantification.

based sample preparation before enzymatic digestion and SRM determination.4,26,27 Early research has been carried out using IA extraction with antibodies immobilized in wells in the 96-plate format.14,28 Although this allowed robust determination of biomarkers in complex human serum samples, the maximal amount of sample volume per well restricted the quantification limits and disabled determination at ultralow concentration levels. Antibodies immobilized on the surface of magnetic beads circumvent this volume limitation and have shown promising results with respect to low detection limits for different protein variants after extraction from a larger volume of patient serum samples.10,12,18 The disadvantages of the referred methods with immunoaffinity (IA) extraction and bottom-up SRM-based LC-MS/MS determination are, however, the current requirement of large sample volumes, a labor intensive procedure, low sample throughput, and the dependence on antibodies with good marker affinity. The aim of the work described in this paper was to combine sample preparation and LC-MS-based determination of several biomarkers occurring in the pg−ng (pM) level in human serum into one single analysis. In this way, the sample throughput and utilization of the available samples are assumed to be improved without compromising the separate methods performance. The small cell lung cancer (SCLC) biomarkers progastrin releasing peptide (ProGRP) and neuron specific enolase (NSE), which occur in various isoforms and isoenzymes, respectively, were chosen as model proteins. Both markers occur in low levels, in pg/mL to ng/mL (corresponding to pM levels) in serum and plasma,29−32 and total ProGRP and γ-NSE are routinely determined using immunoassays such as the time-resolved-



MATERIALS AND METHODS Chemicals and Protein and Peptide Standards. Recombinant ProGRP products, ProGRP isoform 1 (AA 1125) and ProGRP isoform 3 (AA 1-115), were cloned from B

dx.doi.org/10.1021/ac500986t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

were added to a 1 mL sample in an Eppendorf Protein LoBind vial (Eppendorf, Hamburg, Germany). The vial was subsequently rotated and shaken for about 1 h to facilitate antigen− antibody interaction. The beads were collected by DynaMag-2 (Life Technologies Corporation Grand Island, NY, USA) and sequentially washed with a series of solutions as described elsewhere.8 Predigest Conditions and Trypsin Digestion of Captured SCLC Markers. An amount of 70 μL of 50 mM freshly prepared ABC-buffer was added to the washed beads obtained from the immunoextraction. Subsequently, the predigest treatment involved reduction, by adding 5 μL of 100 mM DTT, denaturation, by incubation at 60 °C for 15 at 800 rpm, and alkylation after cooling down the solution to room temperature, by adding 5 μL of 400 mM iodoacetic acid (IAA, freshly prepared in ABC-buffer) and incubating for 15 min in the dark. Then, internal standard (50 nM ELPLY[R_13C6_15N4] and/or NLLGLIEA[K_13C615N2]) was added, resulting in a total sample volume of 90 or 100 μL. Digestion was initiated by adding 10 μL of 50 μg/mL trypsin (freshly prepared in ABC-buffer) to the sample which was incubated overnight at 800 rpm at 37 °C. The solution for injection was prepared by the following two steps: removal of solution from the beads by DynaMag-2 and a spin-down process, as described further elsewhere.8 Finally, a volume of 50 μL was transferred to a vial and placed in the autosampler keeping a temperature of 4 °C, both to stop the digestion and for stability matters for the short-term storage. A volume of 40 μL was injected into the LC-MS system. LC-MS/MS Method. LC-MS was carried out using a triple quadrupole (QqQ) mass spectrometer for quantification purposes. The mass spectrometer was coupled to Dionex Ultimate 3000 chromatographic systems. Chromatographic Parameters. The Dionex Ultimate 3000 chromatographic systems consisted of LPG-3400 M pumps with degasser, WPS-3000TRS autosampler, and FLM-3000 flowmanager (all Dionex, Sunnyvale, CA, USA). The LC system was controlled by Chromeleon v.6.80 SR6 (Dionex). Chromatographic separation was carried out with the following guard and analytical column, respectively: Aquasil C18 Guard Cartridge column (Thermo Scientific, Rockford, IL, USA) (100 Å, 5 μm, 10 mm × 1 mm) and Aquasil C18 column (Thermo Scientific) (100 Å, 3 μm, 50 mm × 1 mm). The chromatographic program was as described earlier elsewhere,8 except from the flow rate, which was 45 μL/min, the temperature of the column oven, which was held at 45 °C, and the LC-gradient, where the composition of the mobile phase during the analysis was as follows: first, 100% of mobile phase A was kept constant for 1 min, second, a linear gradient was run up to 15% of mobile phase B in 8 min, and third, a linear gradient was run up to 85% of mobile phase B in 10 min (see also Initial Test of Feasibility section) Mass Spectrometry Parameters. A TSQ Quantum Access MS-detector (Thermo Scientific) was used for the quantification of the signature peptides. Xcalibur version 2.0.7 software (Thermo Scientific) was used to operate the system and to perform data acquisition for the triple quadrupole (QqQ) mass spectrometer. Electrospray ionization was used and operated in positive mode. The final method consisted of five segments, adjusted to the retention time for each signature peptide, with optimized transitions and collision energies for the SRMtransitions, as listed in Supplementary Table 1, Supporting Information.

human cDNA (OriGene Technologies, Rockville, MD, USA), expressed in Escherichia coli (Promega Corporation, Fitchburg, WI, USA) using pGEX-6P-3 constructs (GE Healthcare, Little Chalfont, UK). The recombinant ProGRP products were then bound to a GSTrap FF column (Amersham Biosciences), second eluted by digestion with PreScission Protease (Amersham Biosciences) and, finally, further purified on a MonoQ anion exchange column using an Ä kta Explorer chromatography system (Amersham Pharmacia Biotech, Uppsala, Sweden) as described elsewhere.34 The purity of the ProGRP in these stem solutions was evaluated on SDS-PAGE, and concentration was determined by absorbance at 280 nm (A280) using individually calculated extinction coefficients. Stock solutions of ProGRP were prepared by diluting the isoforms with 50 mM ammonium bicarbonate buffer (ABC-buffer) and stored at a maximum of −4 °C. Enolase-standards, purified recombinant human non-neuronal enolase (purity >90% by SDS-PAGE, NNE, αα-standard), expressed in Escherichia coli by the manufacturer were purchased from Fitzgerald Industries International (North Acton, MA, USA). Recombinant NSE purified from human brain (purity ≥95% by SDS-PAGE, γγ-standard) was purchased from Scripps Laboratories (San Diego, CA, USA). Stock solutions of NNE and NSE were mainly prepared by diluting the standards with a solution consisting of a 1:1 ratio (v/v) of ethylene glycol solution: 5% BSA in PBS and stored at maximum of −80 °C. Working solutions and standard samples were mainly prepared by diluting stock solutions of enolase with 5% BSA in PBS and stored at a maximum of −4 °C. This was strived for to avoid impairment of the enolase standards. However, as in-solution digests demanded dilution with 50 mM ABC-buffer for stock and working solutions, these were stored in smaller aliquots at a maximum of −80 °C and thawed upon rapid handling and use. Internal standards, both NLLGLIEA[K_13C615N2] and ELPLY[R_13C6_15N4] (AQUA Peptides with purity >95%, SigmaAldrich, St. Louis, MO, USA), were diluted according to the Custom AQUA Peptides Storage and Handling Guidelines by SigmaAldrich for relatively hydrophobic peptides and aliquoted for one single thaw cycle before use. The used monoclonal antibodies (mAbs), mAbE146 also termed anti-ProGRP, with its belonging epitope (AA 48-5235) found in all known ProGRP isoforms, and mAbE21 also termed anti-NSE and anti-γ, specific for the γsubunit,36 were prepared as described previously.35 Coating Magnetic Beads with mAbs. Twenty milligrams of mAbs (mAbE21 or mAbE146) was coupled to 1 g of magnetic beads (Dynabeads M-280, Life Technologies Corporation Grand Island, NY, USA) activated with p-toluene sulfonyl chloride (Life Technologies Corporation). The antibody solution was titrated to pH 2.5 with 5 M HCl and held at this pH for 1 h on ice before the pH was adjusted to 9.5 with 5 M NaOH, and the solution was mixed with the activated beads. The acid treatment was performed to optimize the orientation of the antibodies on the surface. The rest of the procedure was similar to the procedure described elsewhere.35 The resulting bead solution for both anti-ProGRP and anti-NSE was 1 mg of beads/ mL containing approximately 15 μg of antibody/mg of beads. Sample Preparation. Immunoaffinity Extraction of SCLC Markers. The anti-NSE (mAbE21) coated magnetic beads and anti-ProGRP (mAbE146) coated magnetic beads were prewashed as described elsewhere.8 Immunocapture was then performed as follows: one or two volumes of 20 μL of prewashed mAb coated magnetic beads solution (10 mg of beads per mL containing approximately 15 μg of antibody per mg of beads) C

dx.doi.org/10.1021/ac500986t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 1. Illustration of the main steps in the IA MS methods. The numbers correspond to the following: (1) Magnetic beads coated with antimarker mAbs are added to the sample to selectively bind the target markers. (2) A magnetic device withholds the beads with the target markers attached which allows for several washing cycles and enrichment of target markers. (3) Both AQUA-peptide internal standard(s) and trypsin are added the sample. (4) The beads are withheld, and the digested sample is transferred to another vial which is centrifuged prior to analysis of the supernatant by use of LC-SRMMS.

Table 2. Key Parameters for the IA MS Methods for ProGRP and NSEa the individual ProGRP method quantifiable markers sample matrix immunocapture postimmunocapture treatment LC-MS a

the multiplexing method for both NSE and ProGRP

the individual NSE method

ProGRP: isoform 1, isoform 3, and total ProGRP ProGRP-depleted serum mAbE146-coated magnetic beads

NSE: the γ-subunit

Trypsin digest

reduction, heat, alkylation, and trypsin digest

5% BSA mAbE21-coated magnetic beads

Aquasil C18 column, standard gradient elution Aquasil C18 column, standard gradient elution with 40 μL/min flow at 30 °C, ESI-SRM-MS with 40 μL/min flow at 30 °C, ESI-SRM-MS in positive mode in positive mode

ProGRP: isoform 1, isoform 3, and total ProGRP NSE: the α- and the γ-subunit ProGRP-depleted serum mAbE146-coated magnetic beads plus mAbE21-coated magnetic beads reduction, heat, alkylation, and trypsin digest Aquasil C18 column, 2-step gradient elution with 45 μL/min f low at 45 °C, ESI-SRMMS in positive mode

The overview shows the main parameters for the three methods, and the differing parameters between them are in italic and underlined.

methods consist of an IA extraction step to purify and enrich the sample followed by a bottom-up approach and an SRM-based determination (see Figure 1). IA extraction is performed by directly adding magnetic beads coated with antimarker mAbs to the sample, followed by incubation, wash cycles and, finally, digest treatment to produce signature peptides/proteotypic peptides for LC-MS/MS determination. Table 2 shows the similarities and differences between the separate IA MS methods, as well as the multiplex method. The following sections describe the proof of principle, the effect of coextraction and codetermination of the performance, how to harmonize the calibration, and how to include the α-isozyme of NSE in the method. Finally, the method’s applicability is tested on six patient samples.

The TSQ-data was processed by Xcalibur’s QuanBrowser (Thermo Scientific), and MS-response, based on the peak area, was used. Automatic processing with genesis was used as the peak detection algorithm. Peaks with a signal-to-noise(S/N)-ratio above 5 and with retention time and ion ratios corresponding to reference samples at high concentrations (using the relative signal from qualifier transitions in relation to quantifier) were integrated.



RESULTS AND DISCUSSION The intention was to develop a format for simultaneous IA extraction and LC-MS determination for a multiplex method. To evaluate this, two separate methods for individual measurements of the SCLC markers ProGRP8 and NSE6 were combined. Both D

dx.doi.org/10.1021/ac500986t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 2. MS/MS chromatogram of serum sample with human NSE and recombinant ProGRP analyzed by the multiplexing method. The chromatogram shows simultaneous IA extraction using mAb coated beads for each of the two SCLC markers and the detection of all assigned signature peptides for ProGRP isoform 1, ProGRP isoform 3, and total ProGRP and the two NSE subunits, α- and γ-enolase. See Supplementary Table 1, Supporting Information, for SRM transitions and Table 1 for the method parameters.

flow at 45 °C. As shown in Figure 3, alterations in the method were done both to avoid coelution of a possible interference and the signature peptide LSAPGSQR (for ProGRP isoform 1) and to better separate the signature peptides in MS segments. Effect of IA Coextraction. As mentioned, the intention was to combine individual methods without compromising their separate performance. Simultaneous immunocapture of both markers, involving copresence of IA beads, was one of the changed steps compared to the individual methods. Therefore, the effects possibly following changes for the extraction process, the tryptic digest, and the LC-MS analysis conditions had to be investigated and evaluated. Effect of IA Coextraction on the Yield. As a start to review whether the extraction yield of the two markers would be affected by the presence of the other marker’s mAb coated beads, IA extractions with both the individual extraction and coextraction was performed and compared. Thus, immunocapture with only anti-ProGRP magnetic beads, only anti-γ magnetic beads, and coaddition of anti-ProGRP and anti-γ magnetic beads was performed for both serum and 5% BSA samples. In addition, the effect of the presence and absence of the other marker was tested by adding both or either of the standards isoform 1 for ProGRP and γγ-enolase for NSE to the different series of test samples (see Supplementary Table 2, Supporting Information, for an overview). Figure 4 shows a typical result from one sample: presence of anti-ProGRP beads has no effect on the yield of

Initial Test of Feasibility. First, an initial test to evaluate the feasibility of simultaneous extraction of both markers and their variants was carried out by simultaneously adding the IA beads to the same samples. Volumes of 20 μL anti-ProGRP and 20 μL anti-NSE magnetic beads were added to a serum sample containing detectable endogenous levels of NSE (both the αand γ-subunits) and spiked ProGRP isoform 1 and isoform 3 standards. The sample was pretreated according to the NSE method (see Table 2), involving denaturation with DTT and heat and alkylation with IAA, which previously has been shown to have no negative effect on the ProGRP determination.37 The SRM transitions for the signature peptides from both the individual methods were monitored (see Supplementary Table 1, Supporting Information), and as illustrated in Figure 2, all these signature peptides were detected which confirmed feasible coextraction of their parent markers isoforms and isoenzymes. In the initial study, an adaptation was made for the codetermination of ProGRP and NSE by adjustment of the LC-MS method to ensure reproducible determination of all signature peptides. The LC programs for the two separate validated methods were identical (Figure 3), however, for the combined method; the LC program had to be further optimized due to increased complexity of the samples to be analyzed. The differences were that the individual methods had standard gradient elution with 40 μL/min flow at 30 °C, while the multiplex method had a 2-step gradient elution with 45 μL/min E

dx.doi.org/10.1021/ac500986t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 3. Chromatograms obtained using two different LC-MS methods to analyze the IA extracts of serum samples. The chromatogram to the left shows the results using the LC program, the individual methods, and a combination of their MS method (in the box of solid lines), while the chromatogram to the right shows the results from the combined marker determination (in the box of dashed lines). The differences between the LC programs are displayed in the top graph; a slight alteration in the gradient (straight line belongs to the graph to the left, and dashed line is for the graph to the right) and an increase in both the temperature of the column and the flow rate (the highest values belong to the graph to the right).

Figure 4. The effect on detection yield of signature peptides by the presence of the other IA beads and/or the other marker. ProGRP and/or NSE were extracted from either 5% BSA or ProGRP-depleted human serum and the signature peptides to represent that their markers are LSAPGSQR for isoform 1 of ProGRP and ELPLYR for the γ-subunit of NSE. The illustrated yield is relative to the extraction yield from a sample spiked with only the IA targeted marker and extracted with its respective antibody (n = 2 for all samples) (for details, see Supplementary Table 2, Supporting Information).

signature peptide ELPLYR for γ-NSE. The same was shown for the yield of ProGRP isoform 1 signature peptide, LSAPGSQR, when anti-γ beads were present. The data for ProGRP isoform 1 were confirmed by the additional monitoring of NLLGLIEAK, the signature peptide for total ProGRP, as shown in Supplementary Table 2, Supporting Information. This indicates that coextraction of NSE and ProGRP can be carried out without affecting each other’s measurement, both when the levels of the other marker are elevated and carried out in the presence of the other marker’s antimarker beads. Effect of Coextraction on the Linearity of Response. To extensively study if NSE and ProGRP concentrations could be measured independent of the level of the other marker, three situations were simulated for marker extraction from serum. In the first situation, ProGRP and NSE concentrations covaried (NSE between 5 and 500 ng/mL = 11−11000 pM and total ProGRP between 10 and 3000 pM). Second, either ProGRP concentration was kept constant (30 pM), while the NSE

concentration was varied, or NSE concentration was kept constant (20 ng/mL = 420 pM) while the total ProGRP concentration was varied. The results from these experiments show comparable linear regressions for the ProGRP isoforms and total ProGRP (Figure 5A,C) from the two different conditions, indicating that signature peptides can be measured quantitatively linear from ProGRP independent of very different levels of NSE. The same can be seen for NSE (Figure 5B,D). A statistical experiment was performed to confirm the outcome of the data shown in Figure 5. Again, samples containing both NSE and ProGRP at the various concentration combinations were analyzed (n = 3, Table 3). Here, the indifference of high and low levels of ProGRP on the measurement of the NSE signature peptide was tested (at high and low NSE concentration levels) by an unpaired two-sided t test with nonpooled standard deviations (n = 3), an α-limit of 0.05, and a H0 hypothesis of equality. The same was performed for the effect of NSE on ProGRP signature peptides (n = 2, Table F

dx.doi.org/10.1021/ac500986t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 5. The effect on linearity by the presence of the other marker in coextraction. The markers are extracted and measured at both stagnant levels of the other marker and covarying levels in serum. The stagnant levels were 30 pM (=380 pg/mL) for ProGRP and 20 ng/mL (=420 pM) for NSE; the covarying levels are listed in Supplementary Table 3, Supporting Information. All samples had two to three parallels (n = 2−3). Symbol description: gray ■, NLLGLIEAK; gray ◆, LSAPGSQR; gray ▲, DLVDSLLQVLNVK; gray ×, ELPLYR.

Table 3. Testing the Significance of Effects on the Signal Yield of Each Signature Peptide by Varied Concentration Levels of the Other Markera the measured signature peptide NLLGLIEAK LSAPGSQR DLVDSLLQVLNVK ELPLYR

concentration of measured marker

low concentration of the other marker

high concentration of the other marker

P-values for the unpaired twosided t test

90 pM 300 pM 30 pM 100 pM 30 pM 100 pM 50 ng/mL (=1100 pM) 500 ng/mL (=11 000 pM)

50 ng/mL (=1100 pM) 20 ng/mL (=420 pM) 50 ng/mL (=1100 pM) 20 ng/mL (=1100 pM) 50 ng/mL (=1100 pM) 20 ng/mL (=420 pM) 90 pM 90 pM

500 ng/mL (=11 000 pM) 50 ng/mL (=1100 pM) 500 ng/mL (=11 000 pM) 50 ng/mL (=1100 pM) 500 ng/mL (=11 000 pM) 50 ng/mL (=1100 pM) 3000 pM 3000 pM

0.432 0.095 0.979 0.376 0.660 0.064 0.099 0.710

Serum is added to γγ- and ProGRP-standards (isoforms 1, 2, and 3) in high and low concentrations, and the multiplex IA MS method is used to analyze the samples. An unpaired two-sided t test is used to statistically measure if different concentrations of the other marker will give a significant differing effect on the yield or not. The test was performed with nonpooled standard deviations (n = 3), α-limit of 0.05, and H0 hypothesis of equality. The H0 hypothesis could then only be rejected with P-values