International Journal of

Molecular Sciences Article

An Automated Micro-Total Immunoassay System for Measuring Cancer-Associated α2,3-linked Sialyl N-Glycan-Carrying Prostate-Specific Antigen May Improve the Accuracy of Prostate Cancer Diagnosis Tomokazu Ishikawa 1,2 , Tohru Yoneyama 1,3, *, Yuki Tobisawa 1 , Shingo Hatakeyama 1 , Tatsuo Kurosawa 2 , Kenji Nakamura 2 , Shintaro Narita 4 , Koji Mitsuzuka 5 , Wilhelmina Duivenvoorden 6 , Jehonathan H. Pinthus 6 , Yasuhiro Hashimoto 3 , Takuya Koie 1 , Tomonori Habuchi 4 , Yoichi Arai 5 and Chikara Ohyama 1,3 1

2 3 4 5 6

*

Department of Urology, Hirosaki University Graduate School of Medicine, Hirosaki 036-8562, Japan; [email protected] (T.I.); [email protected] (Y.T.); [email protected] (S.H.); [email protected] (T.K.); [email protected] (C.O.) Diagnostics Research Laboratories, Wako Pure Chemical Industries, Hyogo 661-0963, Japan; [email protected] (T.K.); [email protected] (K.N.) Department of Advanced Transplant and Regenerative Medicine, Hirosaki University Graduate School of Medicine, Hirosaki 036-8562, Japan; [email protected] Department of Urology, Akita University Graduate School of Medicine, Akita 010-8543, Japan; [email protected] (S.N.); [email protected] (T.H.) Department of Urology, Tohoku University Graduate School of Medicine, Sendai 980-8574, Japan; [email protected] (K.M.); [email protected] (Y.A.) Department of Surgery, McMaster University, Hamilton, ON L8S4L8, Canada; [email protected] (W.D.); [email protected] (J.H.P.) Correspondence: [email protected]; Tel.: +81-172-39-5091

Academic Editor: Carsten Stephan Received: 1 February 2017; Accepted: 18 February 2017; Published: 22 February 2017

Abstract: The low specificity of the prostate-specific antigen (PSA) for early detection of prostate cancer (PCa) is a major issue worldwide. The aim of this study to examine whether the serum PCa-associated α2,3-linked sialyl N-glycan-carrying PSA (S2,3PSA) ratio measured by automated micro-total immunoassay systems (µTAS system) can be applied as a diagnostic marker of PCa. The µTAS system can utilize affinity-based separation involving noncovalent interaction between the immunocomplex of S2,3PSA and Maackia amurensis lectin to simultaneously determine concentrations of free PSA and S2,3PSA. To validate quantitative performance, both recombinant S2,3PSA and benign-associated α2,6-linked sialyl N-glycan-carrying PSA (S2,6PSA) purified from culture supernatant of PSA cDNA transiently-transfected Chinese hamster ovary (CHO)-K1 cells were used as standard protein. Between 2007 and 2016, fifty patients with biopsy-proven PCa were pair-matched for age and PSA levels, with the same number of benign prostatic hyperplasia (BPH) patients used to validate the diagnostic performance of serum S2,3PSA ratio. A recombinant S2,3PSAand S2,6PSA-spiked sample was clearly discriminated by µTAS system. Limit of detection of S2,3PSA was 0.05 ng/mL and coefficient variation was less than 3.1%. The area under the curve (AUC) for detection of PCa for the S2,3PSA ratio (%S2,3PSA) with cutoff value 43.85% (AUC; 0.8340) was much superior to total PSA (AUC; 0.5062) using validation sample set. Although the present results are preliminary, the newly developed µTAS platform for measuring %S2,3PSA can achieve the required assay performance specifications for use in the practical and clinical setting and may improve the accuracy of PCa diagnosis. Additional validation studies are warranted.

Int. J. Mol. Sci. 2017, 18, 470; doi:10.3390/ijms18020470

www.mdpi.com/journal/ijms

Int. J. Mol. Sci. 2017, 18, 470 Int. J. Mol. Sci. 2017, 18, 470

2 of 15 2 of 15

Keywords: prostate-specific antigen; α2,3-linked sialyl N-glycan; Maackia amurensis lectin (MAA) lectin; biomarker Keywords: prostate-specific antigen; α2,3-linked sialyl N-glycan; Maackia amurensis lectin (MAA) lectin; biomarker

1. Introduction 1. Introduction Serum prostate-specific antigen (PSA) is widely used as a powerful biomarker for detecting prostate-specific antigen (PSA) is widely used as a powerful biomarker for detecting prostateSerum cancer (PCa) [1,2]. However, PSA-based PCa screening has resulted in over-diagnosis, prostate (PCa) [1,2]. However, PSA-based PCa screening has over-diagnosis, leading leading to cancer unnecessary prostate biopsies and overtreatment of resulted indolentincancer [3–5]. Therefore, to unnecessary prostate biopsies and overtreatment of indolent cancer [3–5]. Therefore, novel, more novel, more specific screening methods are urgently needed, especially for young healthy men. screening methods isoforms are urgently especially healthy Of the various Ofspecific the various molecular of needed, PSA, proPSA is for oneyoung of the mostmen. promising potential molecular [6–10]. isoforms PSA, proPSA is one theidentifying most promising potential biomarkers [6–10]. multibiomarkers Aofmulti-biomarker testoffor aggressive prostate cancer thatAcombines biomarker test for identifying aggressive prostate cancer that combines total PSA, free PSA, intact total PSA, free PSA, intact PSA, and human kallikrein-2 has been developed [11,12]. In contrast, PSA, and human kallikrein-2 has been developed [11,12]. In contrast, we focused on the cancerwe focused on the cancer-associated glycan alterations that have frequently been observed during associated glycan alterations that have frequently been observed during carcinogenesis [13,14]. More carcinogenesis [13,14]. More importantly, some glycans, for example α-fetoprotein [15] and human importantly, some glycans, for example α-fetoprotein [15] and human chorionic gonadotropin [16], chorionic gonadotropin [16], have been found to have specific cancer-associated carbohydrate have been found to have specific cancer-associated carbohydrate alterations compared with their alterations compared with their normal counterparts. The glycoprotein PSA has an N-glycosylation normal counterparts. The glycoprotein PSA has an N-glycosylation site on its 45th amino acid from site on its 45th amino acid from the N-terminus. In a prior study, we cleaved a PSA-specific sequence the N-terminus. In a prior study, we cleaved a PSA-specific sequence (Ile-Arg-Asn-Lys, IRNK) that (Ile-Arg-Asn-Lys, IRNK) that includes the glycosylated-Asn (N) and performed an intensive structural includes the glycosylated-Asn (N) and performed an intensive structural analysis of the glycan analysis glycan of PSA using laser desorption/ionization time-of-flight profileofofthe PSA usingprofile matrix-assisted lasermatrix-assisted desorption/ionization time-of-flight (MALDI-TOF) mass (MALDI-TOF) mass spectrometry [17]. This resulted in our identification of a PCa-associated aberrant spectrometry [17]. This resulted in our identification of a PCa-associated aberrant glycosylation of glycosylation PSA, which produces α2,3-linked is readily observed on free PSA PSA, whichofproduces α2,3-linked sialyl N-glycansialyl that N-glycan is readily that observed on free PSA (S2,3PSA), (S2,3PSA), α2,6-linked sialyl N-glycan free PSA (S2,6PSA) is exclusive to benign prostatic whereas whereas α2,6-linked sialyl N-glycan on freeonPSA (S2,6PSA) is exclusive to benign prostatic hyperplasia (BPH) [18] hyperplasia (BPH) [18](Figure (Figure1). 1).

Figure 1. Prostate cancer-associatedaberrant aberrant glycosylation glycosylation of antigen Figure 1. Prostate cancer-associated of N-glycan N-glycanon onprostate-specific prostate-specific antigen (PSA). normal PSA, theterminal terminalsialic sialicacids acids link link to to galactose linkage whereas (PSA). In In normal PSA, the galactoseresidues residueswith withananα2,6 α2,6 linkage whereas prostate cancer (PCa)-associatedPSA, PSA, the the linkage linkage between galactose in in prostate cancer (PCa)-associated between the theterminal terminalsialic sialicacid acidand and galactose residues is an α2,3 linkage [18]. Carbon linkage positions are denoted by the bond position drawn onon residues is an α2,3 linkage [18]. Carbon linkage positions are denoted by the bond position drawn each monosaccharide. Ile-Arg-Asn-Lys, (IRNK): N-glycosylation site of PSA. each monosaccharide. Ile-Arg-Asn-Lys, (IRNK): N-glycosylation site of PSA.

We have also previously developed a magnetic bead-based S2,3PSA assay (Luminex method) that more accurately diagnoses early PCa than the conventional PSA test [19]. However, this method cannot measure benign S2,6PSA and does not have enough quantitative capability for clinical application. With the aim of overcoming these issues, we investigated using microfluidic technology, also referred

Int. J. Mol. Sci. 2017, 18, 470

3 of 15

We have also previously developed a magnetic bead-based S2,3PSA assay (Luminex method) that more accurately diagnoses early PCa than the conventional PSA test [19]. However, this method cannot measure benign S2,6PSA and does not have enough quantitative capability for clinical Int. J. Mol. Sci. 2017, 18, 470 3 of 15 application. With the aim of overcoming these issues, we investigated using microfluidic technology, also referred to as micro-total analysis systems (μTAS system) [20–23], to quantitate both S2,6PSA and in patient samples with a [20–23], one-timetomeasurement. immunoassays to asS2,3PSA micro-total analysisserum systems (µTAS system) quantitate bothμTAS S2,6PSA and S2,3PSA incorporating an automated platform can achieve the required assay performance specifications in patient serum samples with a one-time measurement. µTAS immunoassays incorporatinginan clinical laboratories efficiently existing In this study, we developed an automated platformmore can achieve the than required assaymethodologies. performance specifications in clinical laboratories automated μTAS than system for measuring serum In S2,3PSA ratiowe (%S2,3PSA) settings. more efficiently existing methodologies. this study, developedinanclinical automated µTAS system for measuring serum S2,3PSA ratio (%S2,3PSA) in clinical settings. 2. Results 2. Results 2.1. Preparation of FLAG-Tag-Fused Recombinat S2,3 and S2,6 PSA in Chinese Hamster Ovary 2.1. Preparation of FLAG-Tag-Fused Recombinat S2,3 and S2,6 PSA in Chinese Hamster Ovary (CHO) Cells (CHO) Cells Toconstruct construct S2,3PSA S2,6PSA recombinant standard protein, FLAG-tag-fused To S2,3PSA andand S2,6PSA recombinant standard protein, FLAG-tag-fused humanhuman PSA PSA cDNA (PSA–FLAG) was transiently transfected into CHO cells. Results of immunoblotting cDNA (PSA–FLAG) was transiently transfected into CHO cells. Results of immunoblotting of PSA– of PSA–FLAG-transfected CHO culture are supernatant are shown in Figure 2a. cylindracea Using an FLAG-transfected CHO culture supernatant shown in Figure 2a. Using an Agrocybe Agrocybe cylindracea (ACG) lectin affinity column and gel filtration column chromatography, S2,3PSA (ACG) lectin affinity column and gel filtration column chromatography, S2,3PSA and S2,6PSA and S2,6PSA containing asialo-type N-glycan carrying recombinant PSA standard protein was containing asialo-type N-glycan carrying recombinant PSA standard protein was purifiedpurified from from culture supernatants of human PSA expressed by CHO cells (Figure 2b,c). Lectin array analysis culture supernatants of human PSA expressed by CHO cells (Figure 2b,c). Lectin array analysis then then revealed the putative glycan structure of purified and S2,6PSA recombinant revealed the putative glycan structure of purified S2,3 andS2,3 S2,6PSA recombinant standard standard (Figure (Figure 2d).intensity Signal intensity of recombinant S2,3PSA against Mackkia amurensis lectin (MAL)that that 2d). Signal of recombinant S2,3PSA against Mackkia amurensis lectin (MAL) preferentially bound the sialic acid α2,3-linked galactose structure was much higher than that preferentially bound the sialic acid α2,3-linked galactose structure was much higher than that ofof recombinantS2,6PSA. S2,6PSA.Signal Signal intensity intensity of Nigra lectin (SNA) and recombinant of recombinant recombinantS2,3PSA S2,3PSAagainst againstSambucus Sambucus Nigra lectin (SNA) Sambucus Sieboldiana lectin (SSA)(SSA) that preferentially boundbound the sialic α2,6-linked galactose structure and Sambucus Sieboldiana lectin that preferentially theacid sialic acid α2,6-linked galactose was much lower than that of recombinant S2,6PSA. These results suggest that the culture supernatant structure was much lower than that of recombinant S2,6PSA. These results suggest that the culture of PSA–FLAG-transfected CHO cells contained bothcontained α2,3- and both α2,6-linked sialylα2,6-linked N-glycan-carrying supernatant of PSA–FLAG-transfected CHO cells α2,3- and sialyl recombinant PSA and could be utilized as a standard protein of µTAS system measuring S2,3PSA N-glycan-carrying recombinant PSA and could be utilized as a standard protein of μTAS system ratio (%S2,3PSA). measuring S2,3PSA ratio (%S2,3PSA).

Figure 2. Cont.

Int. J. Mol. Sci. 2017, 18, 470 Int. J. Mol. Sci. 2017, 18, 470

4 of 15 4 of 15

Figure 2. 2. Preparation of of α2,3-linked sialyl N-glycan-carrying PSA (S2,3PSA) and α2,6-linked sialyl NFigure Preparation α2,3-linked sialyl N-glycan-carrying PSA (S2,3PSA) and α2,6-linked glycan-carrying PSA (S2,6PSA) in Chinese hamster ovary (CHO)-K1 cells transfected with PSA– sialyl N-glycan-carrying PSA (S2,6PSA) in Chinese hamster ovary (CHO)-K1 cells transfected with FLAG. CHO-K1 cells were transfected with FLAG-tag-fused human PSA cDNA (PSA–FLAG). PSA–FLAG. CHO-K1 cells were transfected with FLAG-tag-fused human PSA (PSA–FLAG). (a) Culture Culture supernatant serum freefree media werewere blotted onto polyvinylidene fluoridefluoride (PVDF) (a) supernatant(CS) (CS)with with serum media blotted onto polyvinylidene membrane and probed with anti-FLAG antibodies; (b) Chromatograms obtainedobtained using Agrocybe (PVDF) membrane and probed with anti-FLAG antibodies; (b) Chromatograms using cylindracea (ACG) (ACG) lectin lectin column chromatography forforα2,3-linked N-glycan-carrying Agrocybe cylindracea column chromatography α2,3-linked sialyl N-glycan-carrying recombinantPSA PSA (S2,3rPSA) collected eluted fraction and other glyco-isoform free recombinant collectedfrom from0.2 0.2MMlactose lactose eluted fraction and other glyco-isoform PSAPSA containing α2,6-linked sialylsialyl N-glycan (S2,6rPSA) or asialo-type N-glycan collected from washed free containing α2,6-linked N-glycan (S2,6rPSA) or asialo-type N-glycan collected from fraction.fraction. S2,3rPSA and S2,6PSA was further purified by gel column chromatography; (c) washed S2,3rPSA and S2,6PSA was further purified by filtration gel filtration column chromatography; Washed fraction (lane 1) and 0.2 M0.2 lactose-eluted fraction (lane 2)(lane (indicating chromatogram) (b) were (c) Washed fraction (lane 1) and M lactose-eluted fraction 2) (indicating chromatogram) blotted PVDF membrane and probed anti-FLAG antibodies;antibodies; (d) Lectin array profiling of (b) wereonto blotted onto PVDF membrane andwith probed with anti-FLAG (d) Lectin array profiling S2,3rPSA andproteins. S2,6rPSA proteins. Purified S2,3rPSA and S2,6rPSA applied the S2,3rPSAofand S2,6rPSA Purified S2,3rPSA and S2,6rPSA was appliedwas to the lectintoarray, lectin array, including spots. The glass scanned usingaa GlycoStationReader1200 including triplicate triplicate spots. The glass slidesslides werewere scanned using GlycoStationReader1200 (Glycotechnica, (Glycotechnica,Sapporo, Sapporo,Japan). Japan). MWM: MWM: molecular molecular weight weight markers; markers; MAL: MAL: Mackkia Mackkia amurensis amurensis lectin; lectin; SNA: SNA:Sambucus SambucusNigra Nigralectin; lectin;SSA: SSA:Sambucus Sambucussieboldiana sieboldianalectin. lectin.

2.2.S2,6PSA S2,6PSAand andS2,3PSA S2,3PSASeparation SeparationininaaMicrofluidic MicrofluidicChannel ChannelFilled Filledwith withElectrophoresis ElectrophoresisLeading LeadingBuffer Buffer 2.2. Containing ContainingMaackia MaackiaAmurensis AmurensisLectin Lectin Simultaneous Simultaneousquantitative quantitativemeasurement measurementof ofS2,6PSA S2,6PSAand andS2,3PSA S2,3PSAwas wasachieved achievedby byperforming performing microchip microchipcapillary capillaryelectrophoresis electrophoresisand andliquid-phase liquid-phasebinding binding assay assay (LBA) (LBA) on on aa µTAS μTASWako Wakoi30 i30auto auto analyzer microfluidic chip chip IO6 (Wako Pure Chemical Industries, Industries, Hyogo, Japan) for electrokinetic analyzerwith with microfluidic IO6 (Wako Pure Chemical Hyogo, Japan) for analyte transport assay (EATA) (Figure Total(Figure assay time of µTAS measuring %S2,3PSA electrokinetic analyte transport assay3a–c). (EATA) 3a–c). Total system assay time of μTAS system ratio was 9 %S2,3PSA min, substantially the previously Luminex developed method (4 Luminex h). The measuring ratio wasshorter 9 min, than substantially shorterdeveloped than the previously reaction during transporttime of DNA conjugate and bound free PSA analyte channels methodtime (4 h). The reaction during transport of DNA conjugate andthrough bound the freechip PSA analyte isthrough short (25–50 s), and the binding kinetics increased isotachophoresis stacking the chip channels is short (25–50ares),greatly and the bindingbykinetics are greatly(ITP) increased by and concentration of the reactants (Figure 3b) [24,25]. Figure 3c shows a typical electropherogram of isotachophoresis (ITP) stacking and concentration of the reactants (Figure 3b) [24,25]. Figure 3c shows S2,6PSA S2,3PSA in a microfluidic separated lectin affinity electrophoresis. a typicaland electropherogram of S2,6PSAchannel and S2,3PSA in aby microfluidic channel separated Because by lectin S2,3PSA reactive to Mackkia amurensis lectin is(MAA), electrophoretic the sandwich affinity iselectrophoresis. Because S2,3PSA reactive to Mackkia migration amurensisof lectin (MAA), 0 0 complex of DNA–labeled PSA Fab (DNA–Fab ), S2,3PSA,anti-total and HiLyte 647–labeled electrophoretic migration anti-total of the sandwich complex of DNA–labeled PSAFluor Fab′ (DNA–Fab′), 0 (HiLyte–Fab0 ) in the separation channel is slower than that of the immunocomplex anti-free S2,3PSA,PSA andFab HiLyte Fluor 647–labeled anti-free PSA Fab′ (HiLyte–Fab′) in the separation channel is of DNA–Fab , S2,6PSA, and HiLyte–Fab0 . of DNA–Fab′, S2,6PSA, and HiLyte–Fab′. slower than 0that of the immunocomplex

Int. J. Mol. Sci. 2017, 18, 470 Int. J. Mol. Sci. 2017, 18, 470

5 of 15 5 of 15

Figure 3.3. %S2,3PSA test using using micro-total micro-total immunoassay immunoassay systems systems (µTAS)-based (μTAS)-based microcapillary microcapillary Figure %S2,3PSA test electrophoresis. (a) Molded plastic chip with top wells for reagents and precision microchannels on electrophoresis. (a) Molded plastic chip with top wells for reagents and precision microchannels the bottom. Red line indicates microchannels. Highlight green line indicates microcapillary separation on the bottom. Red line indicates microchannels. Highlight green line indicates microcapillary channel. Fluorescence signals are signals detected through thethrough thin film theclosing channels the bottom separation channel. Fluorescence are detected theclosing thin film theonchannels on of the chip;of (b)the Schematic of adiagram chip showing the showing electrokinetic analyte transport assay (EATA) the bottom chip; (b)diagram Schematic of a chip the electrokinetic analyte transport method. Waste wells Waste (WWs), trailing buffer (TB)buffer well, (TB) leading (LB), handoff (HO), DNA– assay (EATA) method. wells (WWs), trailing well,buffer leading buffer (LB), handoff (HO), 0 0 labeled anti-total PSA Fab′ (DNA–Fab′) (DW), serum sample and HiLyte Fluor 647–labeled anti-free DNA–labeled anti-total PSA Fab (DNA–Fab ) (DW), serum sample and HiLyte Fluor 647–labeled PSA Fab′PSA (HiLyte–Fab′) mixture0 )(SW), and(SW), stacking (ST) wells(ST) are shown. Vacuum to anti-free Fab0 (HiLyte–Fab mixture and buffer stacking buffer wells are shown.applied Vacuum the WWs loads the loads reagents sample-to-chip channel segments and voltage applied to the WWs the and reagents and sample-to-chip channel segments and applied voltage between applied cathode cathode and anode sample reagents for the binding reaction. Switching the voltage between andmixes anodethe mixes the and sample and reagents for the binding reaction. Switching the from the TB the wellTBtowell the to HO from from isotachophoresis (ITP)(ITP) stacking mode to capillary gel voltage from thewell HOswitches well switches isotachophoresis stacking mode to capillary electrophoresis (CGE) mode. Separation ofofS2,3PSA occurs in in gel electrophoresis (CGE) mode. Separation S2,3PSAfrom fromother otherglyco-isoform glyco-isoform of of free PSA occurs theMackkia Mackkiaamurensis amurensislectin lectin(MAA)-filled (MAA)-filledCGE CGEseparation separationchannel channelprior priorto to laser-induced laser-inducedfluorescence fluorescence the (LIF)detection. detection. (c) (c) Typical Typicalelectropherogram electropherogramof ofS2,6PSA S2,6PSAand andS2,3PSA S2,3PSApeak peakseparation. separation.Fluorescent Fluorescent (LIF) markers markers have have been been coelectrophoresed coelectrophoresed to to identify identify the the PSA PSA protein protein peak peak positions positions (data (data not not shown). shown). RFU: relative fluorescence units. RFU: relative fluorescence units.

When MAA MAA was was eliminated eliminated from from leading leading buffer buffer (LB) (LB) in inthe theseparation separation channel, channel, both both S2,6PSA S2,6PSA When and S2,3PSA immunocomplex co-migrated to the same position (Figure 4a). Furthermore, using and S2,3PSA immunocomplex co-migrated to the same position (Figure 4a). Furthermore, using affinity-basedseparation separationwith with MAA lectin against aberrant glycosylation, also confirmed affinity-based MAA lectin against aberrant glycosylation, we alsowe confirmed a capturea capture efficiency of applied S2,3rPSA standard of almost 100% in this assay (Figure 4b,c). efficiency of applied S2,3rPSA standard of almost 100% in this assay (Figure 4b,c).

Int. J. Mol. Sci. 2017, 18, 470 Int. J. Mol. Sci. 2017, 18, 470

6 of 15 6 of 15

Int. J. Mol. Sci. 2017, 18, 470

6 of 15

Figure 4. 4. Capture in μTASassay. assay. (a)Electropherogram Electropherogram Figure Captureefficiency efficiencyofofapplied applied S2,3PSA S2,3PSA standard standard in of of Figure 4. Capture efficiency of applied S2,3PSA standard inμTAS µTAS assay.(a) (a) Electropherogram of S2,6rPSA and S2,3rPSA without MAA lectin. S2,6rPSA (b) and S2,3rPSA (c) standard were applied S2,6rPSA and S2,3rPSA without MAA lectin. S2,6rPSA (b) and S2,3rPSA (c) standard were applied to to S2,6rPSA and S2,3rPSA without MAA lectin. S2,6rPSA (b) and S2,3rPSA (c) standard were applied to thethe detection ofof peak channelfilled filledwith withleading leading buffer (LB) detection peakseparation separationrespectively respectively in in aa microfluidic microfluidic channel buffer (LB) the detection of peak separation respectively in a microfluidic channel filled with leading buffer (LB) containing MAA lectin (+) or not (−). The capture efficiency was evaluated by the peak mobility shift containing MAA lectin (+) or not (−). The capture efficiency was evaluated by the peak mobility shift containing MAA lectin (+) or not (−). The capture efficiency was evaluated by the peak mobility shift based onon thethe specific based specificinteraction interactionwith withlectin lectin reactivity. reactivity. based on the specific interaction with lectin reactivity.

2.3.2.3. Assay Linearity and μTAS System System Assay Linearity andSensitivity Sensitivityofof%S2,3PSA %S2,3PSA Test Test by μTAS 2.3. Assay Linearity and Sensitivity of %S2,3PSA Test by µTAS System demonstratethat thatour ourassay assay isis robust robust for determining determining %S2,3PSA, ToTo demonstrate %S2,3PSA,recombinant recombinantstandard standard To demonstrate that our assay is robust determining %S2,3PSA, recombinant standard samples samples wereserially serially diluted with controlfor PSA protein and %S2,3PSA determined atata aconstant samples were diluted with control PSA protein and %S2,3PSA determined constant were serially diluted with control protein and %S2,3PSA determined at a constant concentration of concentration free PSA. AsPSA shown in Figure Figure 5a,b, there was relationship between concentration ofoffree PSA. As shown in 5a,b, there was aa linear linear relationship between free PSA. As shown in Figure 5a,b, there was a linear relationship between percentage of S2,3PSA and percentage of S2,3PSA and fluorescence intensities in the two prepared samples tested. The assay’s percentage of S2,3PSA and fluorescence intensities in the two prepared samples tested. The assay’s fluorescence intensities in the two prepared samples tested. Thewith assay’s sensitivity was determined by sensitivity was determined by testing samples of buffer buffer spiked serially diluted S2,3PSA with thethe sensitivity was determined by testing samples of spiked with serially diluted S2,3PSA with means 2 standard deviations (SD)being being calculated calculated for As ininFigure 5c,5c, 0.05 testing of buffer spiked(SD) with serially diluted S2,3PSA with the means ± 2 standard deviations means ±samples 2± standard deviations for five five replicates. replicates. Asshown shown Figure 0.05 ng/mL S2,3PSA was clearly detectable over the zero sample, there being no overlap of the 2SD range (SD) being calculated for five replicates. As shown in Figure 5c, 0.05 ng/mL S2,3PSA was clearly ng/mL S2,3PSA was clearly detectable over the zero sample, there being no overlap of the 2SD range with zero. The reproducibility of thebeing peak area detection for2SD the 0.05 ng/mL levelThe wasreproducibility within 15% detectable zero sample,of there overlap of the with zero. with zero. over The the reproducibility the peak no area detection for therange 0.05 ng/mL level was within 15% coefficient of variation (CV) for %S2,3PSA, indicating that the limits of detection of quantitation of the peak area detection for the 0.05 ng/mL level was within 15% coefficient of variation (CV)offor coefficient of variation (CV) for %S2,3PSA, indicating that the limits of detection of quantitation of the assayindicating are 0.05 ng/mL. %S2,3PSA, that the limits of detection of quantitation of the assay are 0.05 ng/mL. the assay are 0.05 ng/mL.

Figure 5. Cont.

Int. J. Mol. Sci. 2017, 18, 470 Int. J. Mol. Sci. 2017, 18, 470

7 of 15 7 of 15

Figure Assaylinearity linearityfor formeasurement measurementofofS2,3PSA S2,3PSAratio ratio(%S2,3PSA) (%S2,3PSA)and andlimit limitofofdetection detectionofof Figure 5. 5.Assay S2,3PSA measurement. (a) Sample was a prepared sample, the original sample having PSA S2,3PSA measurement. (a) Sample was a prepared sample, the original sample having ananPSA concentration of 1.5 ng/mL and %S2,3PSA of 50.0%; (b) Sample was prepared by spiking with PSA concentration of 1.5 ng/mL and %S2,3PSA of 50.0%; (b) Sample was prepared by spiking with PSA (0.4 ng/mL)and and%S2,3PSA %S2,3PSAatat50%. 50%.The Thetwo twosamples sampleswere wereserially serially diluted control PSA containing (0.4 ng/mL) diluted byby control PSA containing S2,6 sialylation or asialo theconcentration. same concentration. PSA concentrations andwere %S2,3PSA were S2,6 sialylation or asialo at the at same PSA concentrations and %S2,3PSA determined by the μTAS The limit of detection (LOD) wasthis 0.05level ng/mL, this level showing bydetermined the µTAS system; (c) Thesystem; limit of(c) detection (LOD) was 0.05 ng/mL, showing no overlap no overlap 2 standard ranges for S2,3PSA concentration and negative between the 2between standardthe deviation (SD) deviation ranges for (SD) S2,3PSA concentration and negative control. control.

2.4. Assay Reproducibility of %S2,3PSA Test Using the µTAS System 2.4. Assay Reproducibility of %S2,3PSA Test Using the μTAS System Assay reproducibility was examined by measuring two samples at each concentration with total reproducibility was 10 examined measuring two at each total of free Assay PSA and %S2,3PSA, using replicateby measurements of samples each sample. Weconcentration confirmed thewith assay’s of free PSA and %S2,3PSA, using 10 replicate measurements of each sample. We confirmed the very good reproducibility. The CV was calculated to within 3% for free PSA and 4% for %S2,3PSA for reproducibility. The CV was calculated to within 3% for free PSA and 4% for allassay’s rangesvery testedgood (Table 1). %S2,3PSA for all ranges tested (Table 1). Table 1. Assay reproducibility of S2,3PSA test. Table 1. Assay reproducibility of S2,3PSA test. Sample Number 1 2 3 4 5 6 7 8 9 10 Ave. 1 SD 2 CV 3 1 1

1.0 ng/mL Free PSA 5.0 ng/mL Free PSA 1.0 ng/mL Free PSA 5.0 ng/mL Free PSA Sample Number Free PSA Free PSA%S2,3PSA %S2,3PSA Free Free PSAPSA %S2,3PSA %S2,3PSA 1 1.04 1.04 50.5 5.06 38.1 50.5 5.06 38.1 50.750.7 38.1 2 0.99 0.99 5.08 5.08 38.1 49.449.4 38.3 3 1.04 1.04 5.00 5.00 38.3 1.05 50.1 5.07 37.9 4 1.05 50.1 5.07 37.9 1.04 47.7 5.05 37.7 5 1.04 47.7 5.05 37.7 1.03 50.4 5.24 37.7 6 1.13 1.03 5.24 5.13 37.7 50.350.4 38.3 7 1.10 1.13 5.13 5.01 38.3 46.850.3 37.9 8 1.05 1.10 5.01 5.04 37.9 47.146.8 38.1 9 1.01 1.05 5.04 5.21 38.1 48.547.1 37.9 49.148.5 38.0 10 1.04 1.01 5.21 5.09 37.9 1.5149.1 0.21 1.04 5.09 0.08 38.0 Ave. 1 0.03 3.1%1.51 0.6% 0.03 0.08 1.6% 0.21 SD 2 2.8% 2 SD, standard 2.8% deviation; 3.1%3 CV: coefficient 1.6% variation. 0.6% CV 3 Ave.: average;

Ave.: average; 2 SD, standard deviation; 3 CV: coefficient variation.

2.5. Validation of %S2,3PSA Test 2.5. Validation of %S2,3PSA Test Relevant clinical details of patients from whom samples were obtained are shown in Table 2. clinical details from whom shown in50Table In this Relevant study, age and PSA levelofofpatients 100 matched serum samples samples were were obtained assessed, are comprising from2. In this with study, ageand and50PSA level of cases 100 matched were6cassessed, comprising 50was from patients PCa from BPH (Table 2,serum Figuresamples 6a). Figure shows that %S2,3PSA patients with PCa and 50 from BPH cases (Table 2, Figure 6a). Figure 6c shows that %S2,3PSA was significantly higher in patients with PCa than in patients with BPH (p < 0.0001). Total PSA level was significantly higher in patients with than (Figure in patients BPH (p < 0.0001).characteristic Total PSA level was not significantly different between thePCa groups 6d).with Receiver operating curve not significantly different between the groups (Figure 6d). Receiver operating characteristic curve analyses were then used to compare the diagnostic potential of total PSA and %S2,3PSA (Figure 6e). analyses werethe then used to compare diagnostic of total anddid %S2,3PSA 6e). The area under curve (AUC) showedthe that results of potential conventional PSAPSA testing not differ(Figure between The area under the curve (AUC) showed results of conventional PSAwas testing did not differ patients with BPH and PCa (AUC 0.5062, 95% that CI 0.3922–0.6202), whereas there a good separation between patients with BPH and PCa (AUC 0.5062, 95% CI 0.3922–0.6202), whereas there was good for %S2,3PSA (AUC 0.8340, 95% CI 0.7555–0.9125, p < 0.0001) without any correlation between aeach separation for %S2,3PSA (AUC 0.8340, 95% CI 0.7555–0.9125, p < 0.0001) without any correlation between each assay (Figure 6b). The optimum cutoff point giving high specificity (72.0%) at 80%

Int. J. Mol. Sci. 2017, 18, 470

8 of 15

assay (Figure 6b). The optimum cutoff point giving high specificity (72.0%) at 80% sensitivity was Int. J. Mol. Sci. 2017, 18, 470 8 of 15 determined to be 42.20% of %S2,3PSA, with positive and negative predictive values of 75.5% and 78.7%,sensitivity respectively, much superior totalofPSA specificity 80% sensitivity was determined was and determined to be 42.20% %S2,3PSA, with(14.0%) positiveatand negative predictive values to be of 4.45 ng/mL of total PSA, with positive and negative predictive values of 48.2% and 41.2%, 75.5% and 78.7%, respectively, and much superior total PSA specificity (14.0%) at 80% sensitivity respectively. In this validation sample set, 54% = 27/50) patients with PCa were found to be in was determined to be 4.45 ng/mL of total PSA, (n with positiveof and negative predictive values of 48.2% andGrade 41.2%, Group respectively. this with validation set, 54% (n = 27/50) patients with2 PCa found the low (GG)In[26], GG 1sample (Gleason Score, GS 3 +of3) and GG (GSwere 3 + 4) tumors. to be in46% the low Group (GG) [26], with GG 1 (Gleason GSa3greater + 3) andthan GG GG 2 (GS 3 + 4)4 + 3) In addition, (n =Grade 23/50) of patients with PCa were found Score, to have 3 (GS In addition, 46%(Pbx) (n = 23/50) of 2). patients with PCashows were found to have a greater than GG 3 (GS tumortumors. by prostate biopsy (Table Figure 6g,f the total PSA and %S2,3PSA ratio of 4 + 3) tumor by prostate biopsy (Pbx) (Table 2). Figure 6g,f shows the total PSA and %S2,3PSA ratio PSA PCa patients classified by the prostate biopsy grade group (Pbx GG). The %S2,3PSA and total of PCa patients classified by the prostate biopsy grade group (Pbx GG). The %S2,3PSA and total PSA level showed no significant difference between Pbx GG ≤ 2 and Pbx GG ≥ 3. In the case of greater level showed no significant difference between Pbx GG ≤ 2 and Pbx GG ≥ 3. In the case of greater than than 50% S2,3PSA ratio, %S2,3PSA of Pbx GG ≥ 3 patients (range 50.7%–71.7%)(47%, n = 11/23) was 50% S2,3PSA ratio, %S2,3PSA of Pbx GG ≥ 3 patients (range 50.7%–71.7%)(47%, n = 11/23) was significantly higher thanthan thatthat of Pbx GGGG ≤ ≤2 2patients 50.3%–52.6%)(26%, (26%, = 7/27) = 0.0019). significantly higher of Pbx patients (range (range 50.3%–52.6%) n =n7/27) (p = (p 0.0019).

Figure 6. Serum %S2,3PSA and total PSA level in the validation sample set. (a) Fifty patients with PCa

Figure 6. Serum %S2,3PSA and total PSA level in the validation sample set. (a) Fifty patients with PCa were pair-matched for age and PSA level with the same number of patients with BPH; (b) Correlations were pair-matched for age and PSA level with the same number of patients with BPH; (b) Correlations between %S2,3PSA and total PSA value in patient serum samples; (c) Serum %S2,3PSA was between %S2,3PSA and total PSA value in patient serum samples; (c)grand Serum %S2,3PSA was significantly higher in patients with PCa than BPH (p < 0.0001). The mean represents thesignificantly overall highermean in patients with PCa than BPH (p < 0.0001). The grand mean represents the overall of the Y variable; (d) Serum total PSA was not significantly different between PCa and BPHmean (e) of the Y Receiver variable;operating (d) Serum total PSA was not curve significantly between and BPH (e)areas Receiver characteristics (ROC) analysis different for detection of PCaPCa showed that the undercharacteristics the curve (AUC) for %S2,3PSA and conventional PSAshowed tests were 0.8340 and 0.5062, operating (ROC) curve analysis for detectiontotal of PCa that the areas under the (f) %S2,3PSAand among PCa patients classified by thewere Pbx GG; (g) and total 0.5062, PSA level of PCa curverespectively; (AUC) for %S2,3PSA conventional total PSA tests 0.8340 respectively; patients classified by the Pbx GG. Red triangle indicates %S2,3PSA 50%; (h)of Comparison between (f) %S2,3PSA among PCa patients classified by the Pbx GG; (g) totalover PSA level PCa patients classified ≤ 2Red and triangle Pbx GG ≥indicates 3 patients%S2,3PSA that had anover %S2,3PSA ratioComparison over 50% (p =between 0.0019). Pbx GG ≤ 2 and by thePbx PbxGG GG. 50%; (h) Pbx GG ≥ 3 patients that had an %S2,3PSA ratio over 50% (p = 0.0019).

Int. J. Mol. Sci. 2017, 18, 470

9 of 15

Table 2. Age and total PSA BPH and PCa patients in the validation study. Characteristics

BPH a

PCa b

n = 100

50

50

Age, median (range) PSA ng/mL, median (range) %S2,3PSA, median (range) pbx GG 3 GG 1 GG 2 GG 3 GG 4 GG 5

66.5 (51–85) 6.45 (1.9–20.4) 38.55 (22.9–59.1)

67 (51–86) 6.6 (1.5–21.4) 45.70 (34.7–71.7) n (%) 15 (30%) 12 (24%) 8 (16%) 9 (18%) 6 (12%)

2,

1 b

no significant difference; prostate cancer.

2

total PSA;

3

pbx GG: prostate biopsy grade group;

p (a vs. b )

a

ns 1 ns 1

An Automated Micro-Total Immunoassay System for Measuring Cancer-Associated α2,3-linked Sialyl N-Glycan-Carrying Prostate-Specific Antigen May Improve the Accuracy of Prostate Cancer Diagnosis.

The low specificity of the prostate-specific antigen (PSA) for early detection of prostate cancer (PCa) is a major issue worldwide. The aim of this st...
4MB Sizes 0 Downloads 8 Views