A METHOD OF DIAGNOSING OR PROGNOSING PSORIATIC ARTHRITIS

Abstract

The present invention relates to methods of diagnosing or prognosing arthritis, specifically methods of diagnosing or prognosing psoriatic arthritis. Also disclosed are methods of diagnosing or prognosing rheumatoid arthritis, and methods of differentiating psoriatic arthritis from rheumatoid arthritis. Specifically, the methods involve determining the quantitative or qualitative level of one or more biomarkers in a biological sample from a subject.

Claims

1. A method of diagnosing or prognosing psoriatic arthritis in a subject, the method comprising the steps of: (a) determining the quantitative or qualitative level of one or more biomarkers in a biological sample from the subject; and (b) diagnosing or prognosing psoriatic arthritis in the subject based on the quantitative or qualitative level of the or each biomarker in the biological sample; wherein the or each biomarker is selected from: Rheumatoid factor C6 light chain; Leucine-rich alpha-2-glycoprotein; Alpha-1-antichymotrypsin; Complement C4-B; Coagulation factor XI; Haptoglobin; Haptoglobin-related protein; and Thrombospondin-1.

2. A method according to claim 1, wherein the or each biomarker is further selected from: Alpha-1-acid glycoprotein 1; Alpha-1-antitrypsin; Insulin-like growth factor-binding protein complex acid labile subunit; Antithrombin; C4b-binding protein alpha chain; Ceruloplasmin; Complement factor B; Clusterin; Platelet basic protein; Extracellular matrix protein 1; Inter-alpha-trypsin inhibitor heavy chain H4; Kininogen-1; Lipopolysaccharide-binding protein; Pigment epithelium-derived factor; Vitamin K-dependent protein C; and Prothrombin.

3. A method according to claim 1, wherein the or each biomarker is further selected from: Gelsolin; Filamin-C; Complement component C9b; Peroxiredoxin-2; Plasma serine protease inhibitor; Adenosine deaminase 2; Pregnancy zone protein; Myomegalin; Apolipoprotein D; Glycocalicin; Afamin; Plasma protease C1 inhibitor; Inter-alpha-trypsin inhibitor heavy chain H3; Insulin-like growth factor-binding protein 3; Galectin-3-binding protein; Alpha-2-HS-glycoprotein chain B; and Antithrombin-Ill.

4. A method of diagnosing or prognosing rheumatoid arthritis in a subject, the method comprising the steps of: (a) determining the quantitative or qualitative level of one or more biomarkers in a biological sample from the subject; and (b) diagnosing or prognosing rheumatoid arthritis in the subject based on the quantitative or qualitative level of the or each biomarker in the biological sample; wherein the or each biomarker is selected from: Rheumatoid factor C6 light chain; Leucine-rich alpha-2-glycoprotein; Alpha-1-antichymotrypsin; Complement C4-B; Coagulation factor XI; Haptoglobin; Haptoglobin-related protein; and Thrombospondin-1.

5. A method according to claim 4, wherein the or each biomarker is further selected from: Alpha-1-acid glycoprotein 1; Alpha-1-antitrypsin; Insulin-like growth factor-binding protein complex acid labile subunit; Antithrombin; C4b-binding protein alpha chain; Ceruloplasmin; Complement factor B; Clusterin; Platelet basic protein; Extracellular matrix protein 1; Inter-alpha-trypsin inhibitor heavy chain H4; Kininogen-1; Lipopolysaccharide-binding protein; Pigment epithelium-derived factor; Vitamin K-dependent protein C; and Prothrombin.

6. A method according to claim 4, wherein the or each biomarker is selected from: Gelsolin; Filamin-C; Complement component C9b; Peroxiredoxin-2; Plasma serine protease inhibitor; Adenosine deaminase 2; Pregnancy zone protein; Myomegalin; Apolipoprotein D; Glycocalicin; Afamin; Plasma protease C1 inhibitor; Inter-alpha-trypsin inhibitor heavy chain H3; Insulin-like growth factor-binding protein 3; Galectin-3-binding protein; Alpha-2-HS-glycoprotein chain B; and Antithrombin-Ill.

7. A method of differentiating psoriatic arthritis from rheumatoid arthritis in a subject, the method comprising the steps of: (a) determining the quantitative or qualitative level of one or more biomarkers in a biological sample from the subject; and (b) differentiating psoriatic arthritis from rheumatoid arthritis in the subject based on the quantitative or qualitative level of the or each biomarker in the biological sample; wherein the or each biomarker is selected from: Rheumatoid factor C6 light chain; Leucine-rich alpha-2-glycoprotein; Alpha-1-antichymotrypsin; Complement C4-B; Coagulation factor XI; Haptoglobin; Haptoglobin-related protein; and Thrombospondin-1.

8. A method according to claim 7, wherein the or each biomarker is further selected from: Alpha-1-acid glycoprotein 1; Alpha-1-antitrypsin; Insulin-like growth factor-binding protein complex acid labile subunit; Antithrombin; C4b-binding protein alpha chain; Ceruloplasmin; Complement factor B; Clusterin; Platelet basic protein; Extracellular matrix protein 1; Inter-alpha-trypsin inhibitor heavy chain H4; Kininogen-1; Lipopolysaccharide-binding protein; Pigment epithelium-derived factor; Vitamin K-dependent protein C; and Prothrombin.

9. A method according to claim 7, wherein the or each biomarker is further selected from: Gelsolin; Filamin-C; Complement component C9b; Peroxiredoxin-2; Plasma serine protease inhibitor; Adenosine deaminase 2; Pregnancy zone protein; Myomegalin; Apolipoprotein D; Glycocalicin; Afamin; Plasma protease C1 inhibitor; Inter-alpha-trypsin inhibitor heavy chain H3; Insulin-like growth factor-binding protein 3; Galectin-3-binding protein; Alpha-2-HS-glycoprotein chain B; and Antithrombin-Ill.

10. A method according to claim 7, wherein differentiating subjects suffering from psoriatic arthritis from subjects suffering from rheumatoid arthritis is based on the quantitative or qualitative level of the or each biomarker in the biological sample.

11. A method according to claim 1, wherein the determining step (a) comprises determining the quantitative or qualitative level of all of the biomarkers in the biological sample from the subject.

12. A method according to claim 1, wherein the determining step (a) comprises determining the quantitative or qualitative level of each of the biomarkers in the biological sample from the subject.

13. A method according to claim 1, wherein the or each biomarker is a protein defined by a UniProt Accession Number selected from: A0N5G1; P02750; P01011; P0C0L5; P03951; P00738; P00739; and P07996.

14. A method according to claim 2, wherein the or each biomarker is a protein defined by a UniProt Accession Number selected from: P02763; P01009; P35858; Q8J001; P04003; P00450; P00751; P10909; P02775; Q16610; Q14624; P01042; P18428; P36955; P04070; and P00734.

15. A method according to claim 3, wherein the or each biomarker is a protein defined by a UniProt Accession Number selected from: P06396; Q14315; P32119; P05155; P05154; P02748; P20742; Q5VU43; P05090; P07359; P43652; Q06033; Q08380; P02765; and P01008.

16. A method according to claim 1, wherein the or each biomarker is a protein comprising an amino acid sequence selected from any one of SEQ ID NOs: 1-18.

17. A method according to claim 2, wherein the or each biomarker is a protein having an amino acid sequence selected from any one of SEQ ID NOs: 19-37.

18. A method according to claim 3, wherein the or each biomarker is a protein having an amino acid sequence selected from any one of SEQ ID NOs: 38-55.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0120] Embodiments of the present invention will now be described with reference to the following non-limiting examples and accompanying drawings, in which:

[0121] FIG. 1 illustrates the association of protein signatures with diagnosis as shown by (A) unsupervised hierarchical cluster analysis (HCA), (B) supervised HCA and (C) Principal component analysis, wherein plots were generated on differentially expressed proteins between PsA (n=30) and RA (n=30) patients (p≤0.01, Benjamini Hochberg FDR);

[0122] FIG. 2 illustrates the association between protein expression and differential diagnosis (A) ROC analysis of LC-MRM data (n=60), (B) MRM and MS/MS spectrum for CRP, (C) Levels of CRP analysed by ELISA (p≤0.009) and MRM (p≤0.006) (D) Pearson correlation between ELISA and CRP measurements (R.sup.2 0.8345);

[0123] FIG. 3 illustrates ROC analysis of (A) LC-MS/MS (n=60), (B) SOMAscan (n=36), (C) Luminex (n=64), (D) RNA seq (n=63), and (E) Combined matched (n=36) data;

[0124] FIG. 4 illustrates serum proteins measured by Luminex analysis were significantly differently expressed between PsA and RA patients, wherein Luminex analysis of serum samples revealed (A) IL-18 (p=0.001) II-18 BPa, HGF and FAS (p≤0.05) were differentially expressed between PsA (n=32) and RA(n=32); and

[0125] FIG. 5 illustrates serum miRNAs measured in PsA and RA patients, wherein 10 miRNAs were significantly differentially expressed PsA Vs RA and RA GR Vs NR patients respectively (p=0.05).

EXAMPLES

[0126] Materials and Methods

[0127] Patients

[0128] A total number of 64 patients were recruited, and a full description of the cohort is described in Szenpetery et al., “Striking difference of periarticular bone density change in early psoriatic arthritis and rheumatoid arthritis following anti-rheumatic treatment as measured by digital X-ray radiogrammetry”. Rheumatology (Oxford), 2016. 55(5): p. 891-896. Recent-onset (symptom duration<12 months), treatment naïve PsA and RA patients with active joint inflammation, aged 18 to 80 years were enrolled consecutively. PsA patients (n=32) fulfilled the CASPAR criteria according to Taylor, W., et al., “Classification criteria for psoriatic arthritis: development of new criteria from a large international study”. Arthritis Rheum, 2006. 54(8): p. 2665-73. and patients with RA (n=32) met the 2010 ACR/EULAR classification criteria for RA according to Aletaha, D., et al., “Rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative”. Arthritis Rheum, 2010. 62(9): p. 2569-81. Exclusion criteria were pregnancy, diseases of bone metabolism, previous treatment with disease-modifying anti-rheumatic drugs (DMARDs) or biologic agents, and treatment with anti-resorptive medications, parathyroid hormone or strontium ranelate 6 months prior to the study. The use of calcium and vitamin D supplements and a stable dose of steroids of less than 10 mg/day were permitted during the study.

[0129] Label Free nLC-MS/MS Analysis

[0130] Prior to proteomic analysis, serum samples were depleted of 14 high abundant proteins (HAP) using the Agilent Multiple Affinity Removal System comprising a Hu-14 column (HuMARS14) (4.6×100 mm; Agilent Technologies, 5188-6557) on a Biocad Vision Workstation and subsequently trypsinized. Samples were run on a Thermo Q Exactive mass spectrometer according to the manufacturer's instructions.

[0131] Bioinformatic Data Analysis

[0132] nLC-MS/MS data were visually inspected using Xcalibur software (2.2 SP1.48). MaxQuant (1.4.12) was then used for quantitative analysis of the Thermo Scientific .raw files while Perseus software (1.5.0.9) supported statistical analysis of the data.

[0133] SOMAscan Analysis

[0134] Individual patient serum samples were subjected to a multiplexed aptamer-based assay (SOMAscan) developed by Gold et al. to measure the levels of 1129 proteins as described by McArdle, A., et al., “Developing clinically relevant biomarkers in inflammatory arthritis: A multiplatform approach for serum candidate protein discovery”. Proteomics Clin Appl, 2016. 10(6): p. 691-8.

[0135] Luminex Analysis

[0136] Individual serum samples were subjected to in-house developed and validated multiplexed immunoassays measuring 48 analytes using Luminex xMAP proteomics technology (Austin, Tex., USA). This analysis was undertaken by the Multiplex Core Facility Laboratory of Translational Immunology LTI, in the University Medical Centre Utrecht. The assays were performed as previously described by McArdle, A., et al., “Developing clinically relevant biomarkers in inflammatory arthritis: A multiplatform approach for serum candidate protein discovery”. Proteomics Clin Appl, 2016. 10(6): p. 691-8.

[0137] RNAseq Analysis

[0138] Serum RNA was isolated using the miRNeasy serum/plasma kit (Qiagen) according to the manufacturer's instructions. RNA concentration was measured using the NanoDrop Spectrophotometer. For each sample, 1.5 μL of RNA was reverse transcribed using the miScript reverse transcription kit (Qiagen) according to the manufacturer's instructions. Reverse transcription is based on a poly-A tailing of mature miRNAs followed by tailed oligo-dT reverse transcription. As such, all mature miRNAs in the RNA sample are reverse transcribed and amenable for qPCR detection. Individual cDNA samples were pooled, followed by a miRNA-specific pre-amplification and quantification using qPCR. In total, 2402 individual miRNAs were profiled using version 20 of the miRNome platform. Assays are spotted across 7×384-well plates. The qPCR mix contained a synthetic PCR template (PPC) that is used to assess PCR performance. The PPC assay was measured in duplicate for each sample on each miRNome assay plate. The number of detected miRNAs was determined by applying a Cq detection cut off of 29 cycles. miRNA analysis was carried out by Biogazelle, Gent, Belgium.

[0139] MRM Design and Optimisation

[0140] The development and optimisation of MRM assays was performed using Skyline software (version 3.6.0.1062) (MacCoss laboratory, Washington, D.C.). Assays were developed to prototypic peptides for all proteins of interest according to the following criteria: no missed cleavages or ‘ragged ends’, sequence length between 4-25 amino acids. Where possible, peptides sequences with reactive (C) or methionine (M) residues were avoided but not excluded. A working MRM was determined based on the dot product n.8, signal to noise 10, data points under the curve and percentage coefficient of variance (retention time 1%, area 20%).

[0141] Sample Preparation for LC-MRM Analysis

[0142] Crude serum (2 μL) was added to the wells of a 96 deep-well plate (Thermo) and diluted 1 in 50 with NH.sub.4CO.sub.3. Rapigest™ SF surfactant/denaturant (Waters) was re-suspended in 50 mM NH.sub.4CO.sub.3 to give a stock solution of 0.1% w/v. The stock solution was added to each sample so that the final concentration of Rapigest™ was 0.05%. Plates were covered with adhesive foil and samples were incubated in the dark at 80° C. for 10 min. After incubation, plates were centrifuged at 2000 rcf, 4° C. for 2 min to condense droplets. Following this, DTT was added to each sample at a final concentration of 20 mM. Samples were then incubated at 60° C. for 1 hr followed by centrifugation at 2000 rcf, 4° C. for 2 min. Next, IAA was added to each sample to give a final concentration of 10 mM and plates were incubated at 37° C. in the dark for 30min. Again, plates were centrifuged at 2000 rcf at 4° C. for 2 min and samples were next diluted with LC-MS/MS grade H.sub.2O to give a final concertation of 25 mM NH.sub.4CO.sub.3. Trypsin was then added to each sample so that the protein enzyme ratio was 25:1. The reaction was stopped with the addition of 2 μL of neat TFA to each sample and incubation for a further 30 min at 37° C. In order to pellet Rapigest™, digests were transferred from 96-well plates to 1.5 mL to bind eppendorfs (Eppendorf) and centrifuged for 30 min at 12000 rcf. Supernatants were removed and transferred into clean eppendorfs and lyophilised by speed vacuum at 30° C. for 2 hr. Lyophilised samples were stored at −80° C. until further use.

[0143] LC-MRM Analysis

[0144] MRM analysis was performed using an Agilent 6495 QqQ mass spectrometer with a JetStream electrospray source (Agilent) coupled to a 1290 Quaternary Pump HPLC system. Peptides were separated on an analytical on a Zorbax Eclipse plus C18, rapid resolution HT: 2.1×50 mm, 1.8 um, 600 Bar column (Agilent) before introduction to the QqQ. A linear gradient of 3-75% over 17 mins was applied at a flow rate of 0.400 μL/min with a column oven temperature of 50° C. Source parameters were as follows; gas temp: 150° C., gas flow 15 L/min, nebuliser psi30, sheath gas temp 200° C. sheath gas flow 11 L/min. Peptide retention times and optimised collision energies were supplied to MassHunter (B0.08 Agilent Technologies) to establish a dynamic MRM scheduling method based on input parameters of 800 ms cycle times and 2 min retention time windows. Percentage coefficient of variance (% Cv) of biological and technical replicates was used as a measure of variance and was calculated using the standard calculation of % Cv=(standard deviation/mean)*100.

[0145] Enzyme linked Immunosorbent Assay Analysis

[0146] CRP levels were evaluated using the “gold” standard clinical grade assay in St Vincent's University Hospital, Dublin. 125 μL of serum from each patient was analysed for levels of CRP using an automated CRPL3 Tina-quant C-Reactive Protein assay (Roche Diagnostics, GmbH).

[0147] Statistical Analysis

[0148] Graph pad Prism software package (7.00) were used to investigate the statistical significance of Luminex and miRNA data whereas SOMAsuite (1.0) was used to analyse SOMAscan data. The ability of quantified proteins/peptides and miRNAs to predict the diagnosis (PsA or RA) of individual patients was assessed using the random Forest package in R (version 3.3.2). The most important variables in providing the area under the receiver operating curve were selected by use of the variable importance index and the Gini decrease in impurity was used to assess the importance of each variable. All AUC values were determined using the ROCR package in R (version 3.3.2).

Example 1

Patient Sample Characterisation and Study Design

[0149] Serum samples were collected from 32 PsA and 32 RA patients. The demographic and clinical features of patients are summarised in Table 1.

TABLE-US-00001 TABLE 1 Baseline demographics and clinical parameters of 64 patients with early inflammatory arthritis. Total PsA RA (n = 64) (n = 32) (n = 32) (n = 32) Age (years) 43.58 ± 13.25 39.56 ± 11.14 47.59 ± 14.13 Female/Male 37(58)/27(42) 15(47)/17(53) 22(69)/10(31) n(%) aCCP [+] n(%) 33 (52) 7 (22) 26 (81) (normal 0-6.9) RF [+] n(%) 25 (39) 0 25 (78) (normal 0-25) ESR (mm/h) 19.4 ± 16.8  12 ± 8.1 26.7 ± 20   CRP (mg/L) 14.4 ± 19.8 6.6 ± 8.3 22.2 ± 24.6 (normal <5) DAS28-CRP 4.2 (1.66-6.88) 3.7 (2.1-5.8) 4.9 (1.7-6.9) TJC (0-28 joints) 6 (0-23) 4 (0-20) 8.5 (0-23) SJC (0-28 joints) 2 (0-12) 1 (0-5) 3.5 (0-12) Dactylitis n(%) 10 (31) BMI (kg/cm.sup.2) 28.1 ± 6.27 27.97 ± 6.32  28.24 ± 6.32  PASI 3.35 (0-27.7)

[0150] Unbiased nLC-MS/MS Based Protein Analysis

[0151] To investigate differences in serum protein expression between patients, individual depleted samples were analysed by nLC-MS/MS on a QExactive mass spectrometer. A total of 451 proteins were identified across all samples analysed.

[0152] To identify proteins that were differentially expressed between patients with PsA from those with RA (a) univariate analysis was applied to 121 commonly identified proteins in these patient samples and (b) multivariate analysis was applied to the complete data set. Univariate analysis (student T test using a Benjamini Hochberg FDR 0.01) revealed that 66 proteins were significantly differentially expressed between PsA and RA patients.

[0153] Hierarchical cluster and principle component analysis was carried out on these 66 proteins and this demonstrated in an unbiased manner, the overall differences/similarities between expression levels in the individual PsA and RA patients. Clear within group clustering and between group separations could be observed (see FIG. 1).

[0154] Random forest analysis on the 451 proteins revealed patients could be segregated with an AUC of 0.94 (see Table 3; ROC plot of FIG. 3A). Thus, this data clearly reflected a difference in the serum protein profile between PsA and RA patients.

TABLE-US-00002 TABLE 3 Pattern of expression changes in peptides measured by MRM and LC-MS/MS. Peptides were analysed in PsA (n = 30) and RA (n = 30) patient samples during LC-MS/MS analysis of depleted serum) and MRM analysis of crude serum. UniProt Gene Pattern of Expression MRM Vs LC-MS/MS # Accession ID Name Protein Peptide Concordance Discordance  1 A0N5G1 A0N5G1 V-kappa-1 Rheumatoid factor C6 light chain ASSLESGVPSR ↑RA  2 P02763 A1AG ORM1 Alpha 1 acid glycoprotein SDWYTDWK ↑RA  3 P01009 A1AT SERPINA1 Alpha 1 antitrypsin SVLGQLGITK ↑RA  4 P01009 A1AT SERPINA1 Alpha 1 antitrypsin LSITGTYDLK ↑RA  5 P02750 A2GL LRG1 Leucine rich alpha 2 glycoprotein VAAGAFQGLR x  6 P01011 ACT AACT alpha-1-antichymotrypsin ADLSGITGAR ↑RA  7 P35858 ALS IFGALS Insulin-like growth factor binding LEYLLLSR x complex acid labile subunit  8 Q8J001 AT3 AT3 Antithrombin SLNPNR NA NA  9 P04003 C4BPA C4BPA C4b-binding protein alpha chain LSLEIEQLELQR x 10 P00450 CERU CP Ceuroplasmin ALYLQYTDETR ↑RA 11 P00450 CERU CP Ceuroplasmin GAYPLSIEIGVR ↑RA 12 P00751 CFAB CFB Complement factor B VSEADSSNADWVTK x 13 P10909 CLUS CLU Clusterin TLLSNLEEAK x 14 P0C0L5 CO4B C4B Complement C4-B GSSTWLTAFVLK ↑PsA 15 P02775 CXCL7 PPBP Platelet basic protein NIQSLEVIGK x 16 Q16610 ECM1 ECM1 Extracellular matrix protein AWEDTLDK ↑PsA 17 P03951 FA11 F11 Coagulation factor X DIYVDLDMK ↑PsA 18 P00738 HPT HP Haptoglobin VTSIQDWVQK ↑RA 19 P00739 HPTR HPR Haptoglobin-related protein GSFPWQAK ↑RA 20 Q14624 ITH4 ITH4 Inter-alpha-trypsin inhibitor heavy SIQNNVR ↑RA chain 21 P01042 KNG1 KNG1 Kininogen YFIDFVAR x 22 P18428 LBP LBP Lipopolysaccharide-binding protein ITLPDFTGDLR ↑RA 23 P36955 PEDF SERPINF1 Pigment epithellum-derived factor SSFVAPLEK ↑PsA 24 P04070 PROC PROC Vitamin K-dependent protein C TFVLNFIK NA NA 25 P04070 PROC PROC Vitamin K-dependent protein C SGWEGR NA NA 26 P00734 THRB F2 Prothombin ETWTANVGK ↑PsA 27 P07996 TSP1 THBS1 Thrombospondin FVFGTTPEDILR ↑RA

TABLE-US-00003 TABLE 4A SEQ UniProt ID # Accession ID Protein Peptide NO:  1 A0N5G1 A0N5G1 Rheumatoid factor C6 light ASSLESGVPSR  1 chain  2 P02750 A2GL Leucine-rich alpha-2- VAAGAFQGLR  2 glycoprotein  3 P02750 A2GL Leucine-rich alpha-2- ADLSGITGAR  3 glycoprotein  4 P02750 A2GL Leucine-rich alpha-2- TLDLGENQLETLPPDLLR  4 glycoprotein  5 P01011 AACT Alpha-1-antichymotrypsin ADLSGITGAR  5 His-Pro-less  6 P01011 AACT Alpha-1-antichymotrypsin EIGELYLPK  6 His-Pro-less  7 P01011 AACT Alpha-1-antichymotrypsin ITLLSALVETR  7 His-Pro-less  8 P0C0L5 CO4B Complement C4-B GSSTWLTAFVLK  8  9 P0C0L5 CO4B Complement C4-B GLEEELQFSLGSK  9 10 P03951 FA11 Coagulation factor XI DIYVDLDMK 10 11 P03951 FA11 Coagulation factor XIa light DSVTETLPR 11 chain 12 P00738 HPT Haptoglobin VTSIQDWVQK 12 13 P00738 HPT Haptoglobin VGYVSGWGR 13 14 P00738 HBB1 Hemoglobin subunit gamma LLVVYPWTQR 14 15 P00738 HBB1 Hemoglobin subunit beta VNVDEVGGEALGR 15 16 P00738 HPT Haptoglobin VGYVSGWGR 16 17 P00739 HPTR Haptoglobin-related protein GSFPWQAK 17 18 P07996 TSP1 Thrombospondin-1 FVFGTTPEDILR 18

TABLE-US-00004 TABLE 4B SEQ UniProt ID # Accession ID Protein Peptide NO:  1 P02763 A1AG Alpha-1-acid glycoprotein 1 SDVVYTDWK 19  2 P01009 A1AT Alpha-1-antitrypsin SVLGQLGITK 20  3 P01009 A1AT Alpha-1-antitrypsin SITGTYDLK 21  4 P35858 ALS Insulin-like growth factor-binding LEYLLLSR 22 protein complex acid labile subunit  5 Q8J001 Q8J001 Antithrombin SLNPNR 23  6 P04003 C4BPA C4b-binding protein alpha chain LSLEIEQLELQR 24  7 P04050 CERU Ceruloplasmin ALYLQYTDETFR 25  8 P04005 CERU Ceruloplasmin GAYPLSIEPIGVR 26  9 P00751 CFAB Complement factor B VSEADSSNADWVTK 27 10 P10909 CLUS Clusterin VSEADSSNADWVTK 28 11 P02775 CXCL7 Platelet basic protein NIQSLEVIGK 29 12 Q16610 ECM1 Extracellular matrix protein 1 AWEDTLDK 30 13 Q14624 ITH4 Inter-alpha-trypsin inhibitor heavy SIQNNVR 31 chain H4 14 P01042 KNG1 Kininogen-1 YFIDFVAR 32 15 P18428 LBP Lipopolysaccharide-binding protein ITLPDFTGDLR 33 16 P36955 PEDF Pigment epithelium-derived factor SSFVAPLEK 34 17 P04070 PROC Vitamin K-dependent protein C TFVLNFIK 35 18 P04070 PROC Vitamin K-dependent protein C SGWEGR 36 19 P00743 THRB Prothrombin ETWTANVGK 37

TABLE-US-00005 TABLE 4C UniProt SEQ # Accession ID Protein Peptide ID NO:  1 P06396 GELS Gelsolin EVQGFESATFLGYFK 38  2 Q14315 FLNC Filamin-C NDNDTFTVK 39  3 P02748 CO9 Complement component C9b TSNFNAAISLK 40  4 P32119 PRDX2 Peroxiredoxin-2 TDEGIAYR 41  5 P05155 IPSP Plasma serine protease inhibitor QLELYLPK 42  6 P05154 IPSP Adenosine deaminase 2 IGHGFALSK 43  7 P02748 CO9 Complement component C9b LSPIYNLVPVK 44  8 P20742 PZP Pregnancy zone protein SSGSLLNNAIK 45  9 Q5VU43 PDE4DIP Myomegalin IYFLEER 46 10 P05090 APOD Apolipoprotein D VLNQELR 47 11 P07359 GP1BA Glycocalicin LTSLPLGALR 48 12 P43652 AFM Afamin FLVNLVK 49 13 P05155 IC1 Plasma protease C1 inhibitor LLDSLPSDTR 50 14 Q06033 ITIH3 Inter-alpha-trypsin inhibitor heavy ALDLSLK 51 chain H3 15 Q06033 ALS Insulin-like growth factor-binding FLNVLSPR 52 protein 3 16 Q08380 LG3BP Galectin-3-binding protein SDLAVPSELALLK 53 17 P02765 FETUA Alpha-2-HS-glycoprotein chain B AHYDLR 54 18 P01008 AT3 Antithrombin-III VGDTLNLNLR 55

Example 2

SOMAscan and Luminex Targeted Protein Analysis

[0155] To extend the breadth of proteome coverage afforded by nLC-MS/MS, samples were also analysed on 2 alternative and complementary protein biomarker discovery platforms. SOMAscan analysis supported the quantification of 1129 proteins in a subset of patient samples PsA (n=18) and RA (n=18). Univariate analysis of these data revealed that 175 proteins were significantly differentially expressed between PsA and RA patients (see Table 2).

TABLE-US-00006 TABLE 2 Determination of protein signatures to predict diagnosis in patients with early PsA and RA. Area under the curve (AUC) values were generated using the predicted probabilities from the random forest model used to discriminate between the groups. Platform n Correctly predicted AUC LC-MS/MS 60 55/60 0.94 Luminex 64 43/64 0.69 SOMAscan 36 26/36 0.75 miRNA 63 36/63 0.55 Combined Omic 36 31/36 0.90

[0156] Multivariate analysis revealed that it was possible to discriminate PsA from RA patients with an AUC of 0.73 (Table 3; ROC plot of FIG. 3B).

[0157] Based on their known importance in PsA and RA, 48 proteins were selected for analysis using the Luminex assay. Of the 48 proteins targeted, 23 were identified in every sample. T-tests revealed that 4 proteins; IL-18 (p≤0.001)11-18 BPa, HGF, and FAS (p≤0.05) were significantly differentially expressed between PsA and RA samples (see FIG. 4).

[0158] Random forest analysis of the Luminex data demonstrated patients could be segregated with an AUC of 0.64 (see Table 3; FIG. 3C). In comparison to the LC-MS/MS analysis, the targeted approach to protein discovery yielded data sets with reduced predictive power.

Example 3

RNAseq Based miRNA Analysis

[0159] The miRNAome of baseline PsA (n=31) and RA (n=32) samples were analysed using a miRNA array. A total of 376 miRNAs were identified of which 178 were commonly expressed in each sample. Using a Mann Whitney U-test it was found that of the 178 common miRNAs analysed 10 were significantly differentially expressed between PsA and RA (see Table 3; FIG. 5).

[0160] Random forest analysis of the 376 miRNA data set revealed it was possible to correctly classify only 36/63 patients resulting an AUC of 0.55 (see ROC plot of FIG. 3D). This data indicated that the serum miRNA profile between PsA and RA patients was not different.

Example 4

Multivariate Analysis of Combined Omic data

[0161] In an attempt to directly compare platforms, the combined matched data set (i.e. from the same 36 samples analysed on each platform) were analysed. Results showed that it was possible to distinguish PsA from RA patients with an AUC of 0.90 (see Table 3; ROC plot of FIG. 3E). The weighted variable importance, which was assessed by a mean decrease in Gini, demonstrated that, out of the top 30 analytes contributing to this AUC, 28 were proteins (19 of which were identified by nLC-MS/MS and 9 by SOMAscan) and 2 were miRNAs. These same proteins and miRNAs were observed as statistically significant during uni- and/or multivariate analysis of the individual protein and miRNA datasets. Taken together, the LC-MS/MS data emerged as the most promising and the proteins identified by nLC-MS/MS were prioritised for further evaluation as the most time- and cost-effective strategy.

Example 5

LC-MRM Evaluation of nLC-MS/MS Identified Biomarkers

[0162] A total of 233 proteins represented by 735 peptides and 3735 transitions (5 per peptide) were brought forward for MRM assay development. These candidates included proteins identified by uni-/multi-variate analysis of the discovery data described here in addition to proteins identified during previous studies in pooled patient samples. Of the proteins brought forward, it was possible to develop an assay for 150 of them represented by 299 peptides. These peptides were measured in the 64 clinical samples in a randomised run order. Random forest analysis of the data revealed it was possible to discriminate between PsA and RA patients with an AUC of 0.79 (see Table 3; FIG. 2A).

[0163] The top 27 most important peptides in providing this AUC were selected by use of the variable importance index, here the Gini decrease in impurity was used to assess the importance of each variable.

[0164] Peptide expression changes observed during LC-MRM analysis were next compared to those observed during nLC-MS/MS analysis. Comparisons could be made for 24/27 peptides since for 3 peptides nLC-MS/MS data was not available. Thus, it was found that for 17/24 peptides, expression changes in PsA and RA patients were in agreement when analysed by both MRM and nLC-MS/MS (5 upregulated in PsA and 12 upregulated in RA) supporting their genuine value as putative biomarkers. For the remaining 7/24 peptides, a potential reason for discordance in observations may be due to false discoveries introduced during the initial LC-MS/MS analysis whereby workflows employed were less robust compared to those used during MRM analysis (Table 3). Finally, a MRM assay was developed to CRP (see FIG. 2B) with analysis of this protein by standard lab assay serving as a comparator. It was not surprising to find that serum levels of CRP were significantly upregulated in patients with RA (n=30) as compared to those PsA (n=30) when measured by both ELISA (p≤0.0009) and MRM (p≤0.0006) (see FIG. 2C) and these measurements could be correlated (R.sup.2=0.8345) (see FIG. 2D).

[0165] The present invention identifies biomarkers for the differentiation of patients with PsA from those with RA. Importantly, the invention is based on multiplexed analysis of serological markers in patients with early onset PsA. Here it was established that patients with PsA could be differentiated from those with RA based on molecular signatures identified in serum. Multi-omic analysis revealed it was possible to discriminate PsA from RA patients with an AUC of 0.94 (nLC-MS/MS), AUC 0.69 (Luminex), AUC 0.73 (SOMAscan) and AUC 0.55 (miRNA), while combining data from a group of matched patients resulted in an AUC of 0.90.

Example 6

Independent Evaluation of Candidate Serum Protein Biomarkers for Differentiation of Psoriatic from Rheumatoid Arthritis

[0166] To further identify serological protein biomarkers for the stratification of patients with psoriatic arthritis (PsA) from those with rheumatoid arthritis (RA) at early stages of the disease; the serum proteome of patients with PsA and RA was interrogated using liquid chromatography mass spectrometry (LC-MS/MS). Multiple reaction monitoring (MRM) assays were developed to 206 proteins and subsequently analysed using a triple quadrupole mass spectrometer.

[0167] Recent-onset (symptom duration<12 months), treatment-naïve PsA and RA patients with active joint inflammation, aged 18 to 80 years, were enrolled consecutively. PsA patients (n=94) fulfilled the CASPAR criteria and patients with RA (n=72) met the 2010 ACR/EULAR classification criteria for RA. Exclusion criteria were pregnancy, diseases of bone metabolism, previous treatment with disease-modifying anti-rheumatic drugs (DMARDs) or biologic agents, and treatment with anti-resorptive medications, parathyroid hormone or strontium ranelate 6 months prior to the study. The use of calcium and vitamin D supplements and a stable dose of steroids of less than 10 mg/day were permitted during the study.

[0168] The development and optimisation of MRM assays was performed using Skyline software (MacCoss laboratory, Washington, D.C.). Assays were developed to proteotypic peptides for all proteins of interest according to the following criteria: no missed cleavages or ‘ragged ends’, sequence length between 4-25 amino acids. Where possible, peptides sequences with reactive (C) or methionine (M) residues were avoided but not excluded. A working MRM was determined based on the dot product n.8, signal to noise 0, data points under the curve 0 and percentage coefficient of variance (retention time 1, area 20%).

[0169] Serum samples were collected from 94 PsA and 72 RA patients. The demographic and clinical features of patients are summarised in Table 4.

TABLE-US-00007 TABLE 4 Demographic and clinical features of patients: Discovery and verification: PsA (n = 32) RA (n = 32) Age (years) 39.56 ± 11.14 47.59 ± 14.13 Female/Male n(%) 15(47)/17(53) 22(69)/10(31) aCCP [+] n(%) (normal 0-6.9) 7 (22) 26 (81) CRP (mg/L) (normal <5) 6.6 ± 8.3 22.2 ± 24.6 Dactylitis n(%) 10 (31)  — PASI  3.35 (0-27.7) — Validation: PsA (n = 95) RA (n = 72) Age 52.52 +/− 6.59 55.08 +/− 9.62  Female/Male (%) 51(54)/44(46) 38(53)/34(47) aCCP [+] n(%) 1 (1) 49 (74) CRP 4.74 +/− 6.66 20.96 +/− 34.16 Dactylitis 46 (52) — PASI 2.69 (0-14) — * n = 90 # n = 66 ** n = 86 ## n = 71

[0170] Crude serum (2 μl) was added to the wells of 96-deep-well plates and digested with trypsin. Tryptic digestion was performed in a flat-bottom polystyrene 96-well plate following an in-house developed standard operating procedure (SOP18A). For protein denaturation, 25 μL denaturant solution (50% trifluoroethanol (TFE) in 50 mM NH.sub.4HCO.sub.3 with 10 mM dithiothreitol (DTT)) were added to 2 μL serum in each well. The 96-well plate was covered with a sterile adhesive foil and incubated for 45 min at 60° C. To remove any condensation from the foil, samples were allowed to cool down to room temperature and centrifuged for 2 min at 4000 g. 10 μL of 120 mM iodoacetamide (IAA) solution was added to each sample, the plate was sealed, vortexed and incubated for 30 min protected from light. To quench the excess of IAA, 10 μL of 50 mM DTT was added to each well. The plate was then re-sealed, vortexed and incubated for 30 min protected from light before diluting samples by adding 190 μL of 12.5 mM NH.sub.4HCO.sub.3 solution. For each well 5.5 μL trypsin solution (0.2 mg/mL sequencing grade modified trypsin (Promega) re-suspended 1:1 in trypsin resuspension buffer (Promega) and 50 mM NH.sub.4HCO.sub.3) was used. After 18 h incubation at 37° C., 5 μL of 25% formic acid (FA) was added to each well in the 96-well plate. The digestion plates were stored at −80° C. once the digestion process was complete.

[0171] A total of 206 proteins represented by 423 peptides were used for the MRM assay, which was applied to 166 patient samples. MRM analysis was performed using an Agilent 6495 triple quadrupole (QqQ) mass spectrometer with a JetStream electrospray source (Agilent) coupled to a 1290 Quaternary Pump HPLC system. Peptides were separated on an analytical Zorbax Eclipse plus C18, rapid resolution HT: 2.1×50 mm, 1.8 μm, 600 Bar column (Agilent) before introduction to the QqQ. A linear gradient of 3-75% over 17 mins was applied at a flow rate of 0.400 μl/min with a column oven temperature of 50° C. Source parameters were as follows; gas temp: 150 ° C., gas flow 15 l/min, nebuliser psi 30, sheath gas temp 200° C. and sheath gas flow 11 l/min. Peptide retention times and optimised collision energies were supplied to MassHunter (B0.08 Agilent Technologies) to establish a dynamic MRM scheduled method based on input parameters of 800 millisecond (ms) cycle times and 2 min retention time windows. The percentage coefficient of variance (% Cv) of biological and technical replicates was used as a measure of variance and was calculated using the standard calculation of % Cv=(standard deviation/mean) 100.

[0172] The ability of quantified proteins/peptides to predict the diagnosis (PsA or RA) of individual patients was assessed using the random Forest package in R (version 3.3.2). The most important variables in providing the area under the receiver operating curve were selected by use of the variable importance index and the Gini decrease in impurity was used to assess the importance of each variable. All area under the curve (AUC) values were determined using the ROCR package in R (version 3.3.2).

[0173] Multivariate analysis of the data revealed it was possible to discriminate PsA from RA patients with an area under the curve (AUC) of between 0.844 and 0.901. The most important peptides in providing this AUC were selected by use of the variable importance index≥the Gini decrease in impurity was used to assess the importance of each variable.