Using protein microarray reveals clinical correlation between self-perception of patient and the apoptosis-related proteins in rheumatoid arthritis

Background: The most severe effects of rheumatoid arthritis (RA) are loss of physical function, which may have a signicant impact on self-perception of patient (SPP). However, the inherent relationship between SPP and the key proteins is not clear. The aim of this study was to get an insight into SPP of RA in connection with the the apoptosis-related proteins. Methods: We set out to investigate changes of the apoptosis-related proteins expression in the peripheral blood mononuclear cells (PBMCs) of RA. Additionally, we aimed to correlate the apoptosis-related proteins expression proles with SPP and clinical indexes. To this end, we employed antibody microarrays of the the apoptosis-related proteins in PBMCs from four RA patients and seven healthy controls. We used bioinformatics to screen several the apoptosis-related proteins. To validate key protein candidates, we performed Enzyme linked immunosorbent assay (ELISA) on 30 RA patients and 30 healthy controls. Results: We found the expression of ten the apoptosis-related proteins (caspase3, CD40, SMAC, HSP27, HTRA, IGFBP-1, IGFBP-6, sTNF-R1, sTNF-R2, TRAILR-3) were signicantly altered in PBMCs of RA patients. Receiver operating characteristic (ROC) curve analysis suggested that these ten the apoptosis-related proteins are potential biomarkers of RA. Spearman Correlation analysis and Logistic-regression analysis revealed that the 10 selected the apoptosis-related proteins correlated with SPP and clinical indexes. Conclusion: Therefore, we highlight some the apoptosis-related proteins may serve as potential biomarkers in prediction of SPP for RA patients, although the underlying mechanisms need to be further explored.

QOL in RA, likewise, patients with more comorbidities and extraarticular manifestations show worse QOL 12 .
The exact causes of in RA is unknown. Recently, the number of studies revealing the important role of epigenetics in the pathogenesis of RA has increased 13 14 . However, we do not know whether it is related to decreased SPP in RA. These days, although there are advances in the treatment of RA, we still need to discovery the pathogenesis of RA. The RA clinical diagnosis is mainly based on causes of RA, clinical symptoms, signs, laboratory tests, and clinical imaging 15 . Until now there is no gold standard for early diagnostic approach for RA, and patients are usually comprehensively evaluated by serological tests for rheumatoid factor (RF) and anti-citrullinated protein antibodies (ACPA) and for in ammatory markers such as hypersensitive C-reactive protein (CRP), an erythrocyte sedimentation rate (ESR) test, and an imaging test if symptoms appear 16 17 . However, RF-positive rate and ACPA-speci city can be detected in approximately 50-80% of RA patients, which suggesting that the overall detection e ciency is lower 18 . Apoptosis plays a key role in RA, and the inhibition of apoptosis may provide a novel treatment method for RA diseases 19 20 . Therefore, there is an increasingly growing interest in identifying biomarkers for diseases, in which apoptosis is involved. Studies toward novel diagnostic biomarker discovery, using protein microarray assay, have been increasing 21 22 . A proteomics approach using protein microarray assays can be used to determine the pattern of proteins and compare their absolute levels between diseased and healthy controls 23 .
In present study, we obtained peripheral blood mononuclear cells (PBCMs) from four RA patients and tried to discover diagnostic biomarkers in informative blood samples using protein microarray assays. In the discovery and validation set, candidate biomarkers were selected and veri ed. Additionally, the relationship between these candidate biomarkers with SPP and clinical indexes of RA patients was also analyzed.  24 . In addition, 30 age-and sex-matched healthy controls with no clinical history of tumors, trauma, infectious diseases or autoimmune diseases, who underwent routine Physical Examination Center in the same hospital during the same period. All samples from RA patients and healthy controls were obtained with written informed consent. This study was approved by the ethics committee of the First A liated hospital of Anhui University of Traditional Chinese Medicine.

Measurements
We collected demographic and clinical data from the subjects, including age, gender, disease duration, blood chemistry. All clinical measurements were performed by the clinical laboratory staff of our hospital. The clinical laboratory data such as erythrocyte sedimentation rate (ESR), high-sensitivity C-reactive protein (CRP), rheumatoid factor (RF), anti-cyclic citrullinated peptide antibody (CCP) and clinical characteristics are determined. PBMCs preparation and total RNA extraction An amount of 5 mL of whole blood was obtained from RA patients and healthy controls and PBMCs were isolated through Ficoll-Paque density gradient centrifugation (GE Healthcare, Uppsala, Sweden). The concentration of cells was adjusted to 5-7×10 6 cells per ml and reserved at − 80 °C until use.
Total RNA was extracted from PBMCs of all samples using Trizol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) according to the manufacturers instructions. The RNA was then reverse transcribed to single stranded cDNA, which was used as a template to synthesize the second cDNA strand. Aliquots of total RNA samples were used to determine the RNA concentration and purity using the Nano Drop ND-1000 spectral photometer (peqlab). The RNAs were selected on the basis of a combination of p-value, fold change, raw intensity and type. In addition, RNAs with miRNA response elements (MREs) related to RA reported in the literatures were selected preferentially. All qPCR assays were performed on the Viia7 Real-Time PCR System, each sample was replicated three times. Library quality was assessed on the Agilent Bioanalyzer 2100 system. The relative expression levels of RNAs were calculated using the 2-ΔΔCt method.
The apoptosis-related protein microarray analysis Capture antibodies were spotted onto a glass slide at a volume of 350 pL per spot at a pitch of 500 mm using a microarray printer + . Biotinylated goat anti-bovine IgG antibody was also spotted as a detection control. Each capture antibody was printed in quadruplicate spots within a subarray. Each glass slide contained 16 identical subarrays separated by a 16-well gasketed hybridization chamber to prevent sample cross-contamination. Antibody arrays were stored at -20 ℃ until use. The raw signal data was extracted using MapPix 6.0 software and quantitative data were extracted and analyzed with software for the multiplexed antibody array.

Bioinformatics Analysis
The raw data obtained by the chip scan was subjected to chip background removal and inter-chip normalization processing by Raybiotech software.2.3.2. After the raw data were normalized by the software, the resulting data were selected for analysis. DEPs with P<0.05 were rst retained and then further screened by Foldchange (expression difference multiple). The selection conditions were as follows: Foldchange≤0.83 or Foldchange ≥1.2; and Fluorescent signal > 150. For cluster analysis, the heatmap.2 function and gplots package from R/bioconductor were used. The distance between two samples was calculated as the Euclidean distance; the distance between the two clusters was calculated with the furthest neighbor method (complete), and the distance between classes was de ned as the maximum distance. Fisher's exact test and the clusterPro ler package from R/bioconductor were used.
For selection, the number of genes that differed on a certain GO term and KEGG pathway were ≥ 2, and P < 0.05. The normalized data were exported to SIMCA-p 11.5 for principal component analysis (PCA), partial least-squares discriminant analysis (PLSDA), and orthogonal partial least-squares discriminant analysis (OPLSDA).

Validation with ELISA
According to the manufacturers'instructions, the differentially expressed proteins levels in serum were quanti ed using ELISA kit (eBioscience, Inc., San Diego, CA, USA). Brie y, whole-blood samples were centrifuged at 4,000 g for 10 min to collect serum. After that, conditioned media or supernatants of lysed samples were collected and added to 96-well plates pre-coated with appropriate capture antibody, followed by incubation with the appropriate biotinylated detection antibody. Lastly, Streptavidin horseradish-peroxidase (HRP) was added to each well and incubated for 1 h at room temperature. The results were detected using an ELISA microplate reader (BioTek) at 450 nm.
Statistical Analysis SPSS 23.0 software for Windows was used for statistical analysis. Continuous variables are expressed as means ± standard deviation and categorical variables are expressed as numerals. When normally distributed, independent samples will be compared for numerical variables between the two groups using Student's t-test. Mann-Whitney U-test was used for the data not conforming to a normal distribution. Categorical data were compared between groups using the Chi-square test. Correlations between the variables were performed using Spearman Correlation Analysis and Logistic-regression. A p value < 0.05 was statistically signi cant (*p < 0.05, **p < 0.01, ***p < 0.001).

Results
Basic characteristics of RA patients and healthy controls Thirty RA patients (2 males, 28 females, mean age: 55.63 ± 11.41 years, course of disease: 15.67 ± 8.53 years) and thirty healthy controls (2 males, 28 females, mean age: 54.33 ± 12.59 years) were used as the study group. No signi cant differences in age or gender were identi ed between the two groups (p > 0.05).
Comparisons of SPP and clincial indexes between RA patients and healthy controls As showed in Table 1, compared with healthy control, PF, RP, GH, MH of RA were signi cantly lower, while DAS28, VAS, SAS, SDS, BP were signi cantly higher than healthy control (P < 0.05). Meanwhile, ESR, CRP, RF, CCP of RA were signi cantly higher than HC (P < 0.05). To identify the apoptosis-related differential proteins in RA, we performed antibody microarrays analysis of the the apoptosis-related differential proteins in the PBMCs from four patients with RA and seven healthy people. After the raw data were normalized, the the apoptosis-related proteins were identi ed between the two groups. We identi ed 30 proteins (17 upregulated and 13 downregulated) that were differentially expressed between the two groups (Fig. 1A). Scatter plots showed 10 abundance of the apoptosis-related differential proteins by screening for log2-fold-changes greater than 1 and P < 0.05 (Fig. 1B). A heat map was constructed to group the the apoptosis-related differential proteins based on their expression levels among the samples (Fig. 1C). OPLS-DA clearly separated tested samples into two blocks according to their apoptosis-related differential proteins in both RA and healthy controls (Fig. 1D).
Function of the apoptosis-related differential proteins in RA patients To explore the the apoptosis-related differential proteins, Gene Ontology (GO) terms, Kyoto Encyclopedia of Genes and Genomes (KEGG) and functional enrichment analyses were performed. We placed emphasis on GO analysis of proteins found that regulation of apoptotic process involved in morphogenesis, regulation of apoptotic process involved in deveiopment, positive regulation of proteolysis were in Biological Process (BP) ( Fig. 2A), varicosity, receptor complex and neuronal cell body were in Cellular Component (CC) (Fig. 2B), TRAIL binding, death receptor binding, death receptor activity were in Molecular Function (MF) (Fig. 2C). Many pathways connected with functions of proteins in RA were de ned by the KEGG analysis, TNF signaling pathway, MAPK signaling pathway, Apoptosis were in proteins (Fig. 2D).

Protein-protein interaction network describing RA
To better understand the cellular networks that are altered in the PBMCs of RA patients, the relationship between proteins was analyzed using different cutoff points with similar results (Fig. 3A). Additionally, we created protein-protein interaction network models of RA-specifc using the STRING database (Fig. 3B).
Veri cation of the apoptosis-related differential proteins in RA patients A total of 10 the apoptosis-related differential proteins by screening for log2-fold-changes greater than 1 and P < 0.05 in RA (Table 2), including 7 upregulated and 3 downregulated proteins were selected.
Serum from 30 RA patients and 30 healthy controls were used for veri cation by ELISA. We selected 10 the apoptosis-related differential proteins from the most signi cant proteins for further veri cation. The expression pro le of caspase3, CD40, SMAC (p < 0.01) in the RA group was apparently lower than in healthy control (Fig. 4A, B, G). The expression pro le of HSP27, HTRA, IGFBP-1, IGFBP-6, sTNF-R1, sTNF-R2, TRAILR-3 (P < 0.01) in the RA group were apparently higher than healthy control (Fig. 4C, D, E, F, H, I, J).

ROC curve analysis of the apoptosis-related differential proteins in RA patients
To further assess the biological functions, we performed ROC curve analysis to evaluate the diagnostic value of the 10 the apoptosis-related differential proteins. As showed in

Spearman Correlation analysis of the apoptosis-related differential proteins with SPP and clinical indexes in RA patients
To clarify the relationship between the the apoptosis-related differential proteins with SPP and clinical indexes, we performed Spearman Correlation analysis. The results of Spearman Correlation analysis suggested that caspase3 was positively correlated with age, ESR, RF ( Fig. 6A-C). HSP27 was negatively correlated with CCP (Fig. 6D). HTRA was positively correlated with VT (Fig. 6E). IGFBP-1 was positively correlated with SAS (Fig. 6F). IGFBP-6 was negatively correlated with VAS (Fig. 6G). SMAC was positively correlated with CRP, negatively correlated with RF ( Fig. 6H-I). sTNF-R2 was positively correlated with ESR, negatively correlated with MH ( Fig. 6J-K).
Logistic-regression analysis of SPP, clinical indexes related the apoptosis-related differential proteins In order to identify risk factors of the the apoptosis-related differential proteins in RA patients, we conducted Logistic-regression analysis. Signi cant differences in caspase3 was found between RA patients with ESR (p = 0.002), CCP (p = 0.027), SAS (p = 0.012), indicating that ESR, CCP, SAS were risk factors for caspase3. Signi cant differences in HPS27 was found between RA patients with VAS (p = 0.034), MH (p = 0.038), indicating that VAS and MH were risk factors for HPS27. Signi cant differences in SMAC was found between RA patients with RF (p = 0.023), DAS28 (p = 0.050), VT (p = 0.001), indicating that RF, DAS28, VT were risk factors for SMAC.

Discussion
The early diagnosis of RA is very challenging because there are no specifc diagnostic indicator. Antibody microarray assay-based proteomics is a state-of-the-art analytical technique that enables the diagnosis and treatment of diseases. Proteomics has been a subject of interest in recent research, there has been a recent expansion in proteomics research on a number of diferent rheumatic diseases. Many factors contributed to decreased SPP for those living with RA 25 . With the change of medical model from a biomedical to a bio-psycho-social one and with the development of value-based medicine, SPP should be considered in making clinical decision 26 . Therefore, medical workers are increasingly concerned about SPP of RA. However, it is necessary to seek out new biomarkers and explore their functions for details of the mechanisms of decreased SPP in RA remain ambiguous.
Protein chips play an important role in scienti c research due to its richer detect target factors and smaller sample size requirement. It is widely used in the research of disease mechanisms. Recently, Mun S producted a analysis of changes in serum protein expression pro les of RA patients by SCIEX TripleTOF 560, which found that serum amyloid A4 and vitamin D binding protein could be potential biomarkers related to the in ammatory response and joint destruction that accompany RA. This nding provides an important reference for further research on RA 27 . To identify predictive biomarkers in patients with RA, 50 the apoptosis-related proteins pro ling was conducted individually with 7 PBMCs from the normal group and 9 from the patient group with RA. Analysis of proteins pro ling revealed the ten apoptosis-related proteins in the discovery set that could be considered as potential candidate biomarkers. The ten candidate biomarkers, namely cysteine aspastic acid-speci c protease 3 (caspase3), heat shock protein 27 (HSP27), tumor necrosis factor receptor superfamily member 5 (TNFRSF5, CD40), second mitochondria-derived activator of caspase, SMAC), HtrA serine peptidase 1 (HTRA), insulin-like growth factor binding protein − 1 (IGFBP-1), insulin-like growth factor binding protein − 6 (IGFBP-6), tumor necrosis factor receptor superfamily member 1A (sTNF-R1), tumor necrosis factor receptor superfamily member 1B (sTNF-R2), tumor necrosis factor receptor superfamily member 10C (TRAILR-3) with > 1.5 of fold change in RA compared to healthy controls. Functional annotation of the the apoptosis-related proteins into GO terms and KEGG pathways was performed, which revealed that these the apoptosisrelated proteins involved in many pathophysiological processes, including regulation of apoptotic process, receptor complex, death receptor binding. A PPI network was constructed, which showed that these the apoptosis-related proteins are closely related in terms of their functions.
ELISA experiments for ten the apoptosis-related proteins were performed to further verify the diferential expression level. The expression pro le of caspase3, CD40, SMAC (p < 0.01) in the RA group was apparently lower than in healthy control. While the expression pro le of HSP27, HTRA, IGFBP-1, IGFBP-6, sTNF-R1, sTNF-R2, TRAILR-3 (P < 0.01) in the RA group were apparently higher than healthy control. ROC curve analysis revealed the 10 the apoptosis-related differential proteins as diagnostic predictors of RA ( all AUC value is greater than 0.900). Caspase 3 is involved in signalling pathways leading to apoptosis, which plays signi cant roles in the pathogenesis of RA 28 . Hsps have been implicated in the RA 29 .
Sedlackova L found that signi cantly increased Hsp27 and Hsp90a mRNA levels in RA synovial tissues, which indicated that Hsps could be new diagnostic approach to RA patients 30 . The anti-apoptotic activity of inhibitors of apoptosis proteins (IAP) proteins can be blocked by the SMAC, which is liberated into the cytoplasm in response to proapoptotic stimuli. Lattuada D found that SMAC was associated with caspase 8 and caspase 3 activities, and induced signi cant apoptosis in all RA-FLS samples. In addition, SMAC signi cantly upregulated IGFBP-5, a protein involved in differentiation, apoptosis, and osteoblastic activation, so they included that SMAC may represent a new therapeutic approach to RA treatment 31 .
Finally, we would like to discuss our ndings in regard of their SPP and clinical indexes. We performed Spearman Correlation analysis suggested that caspase3 was positively correlated with age, ESR, RF, HSP27 was negatively correlated with CCP, HTRA was positively correlated with VT, IGFBP-1 was positively correlated with SAS, IGFBP-6 was negatively correlated with VAS, SMAC was positively correlated with CRP, negatively correlated with RF, sTNF-R2 was positively correlated with ESR, negatively correlated with MH. Additionally, results from Regression Analysis indicated that ESR, CCP, SAS were risk factors for caspase3, VAS and MH were risk factors for HPS27, RF, DAS28, VT were risk factors for SMAC. Liao H used 2-dimensional liquid chromatography-coupled tandem mass spectrometry to identify protein biomarkers of disease severity in the synovial uid and serum of RA patients, they concluded that CRP, S100A8, S100A9 and S100A12 proteins could be identi ed for prognosis of the erosive form of RA 32 . However, there were no analysis between potential proteins and clinical indicators, nor did it predict the risk factors affectted potential proteins.
This study addresses important questions in understanding the important problem area in persons with decreased SPP in RA. The novelty in this study is that we have used microarray analysis combined with SPP to investigate their relationship. The main limitation of this study is a small number and inhomogeneity of respondents. Further in vitro and animal studies should be performed to evaluate comprehension of the detailed mechanism and speci c functions of genes in decreased QOL in RA. To sum up, the ten apoptosis-related differential proteins are of potentially signi cant prediction value in decreased SPP in RA patients, but there remains a need to further study the mechanisms of these proteins.

Conclusions
In this work, we focus on the relationship between self-perception of patient (SPP) and the apoptosisrelated proteins in RA patients. We choosed four RA patients and seven healthy control for antibody array. Ten apoptosis-related proteins were veri ed by ELISA.We recruited 30 patients with RA and 30 healthy controls to ll in the ve quentionnaires, including DAS28, VAS, SDS, SAS, SF-36. Finally, Spearman The data used to support the ndings of this study are available from the corresponding author upon request.

Ethics approval and consent to participate
The study was approved by the Ethics Committee of the First A liated Hospital of Anhui University of Traditional Chinese Medicine and carried out under the Helsinki Declaration. Before participating in the study, the patients lled in a written informed consent forms. A written informed consent was obtained from all the study participants.

Consent for publication
Not applicable.

Con icts of interest
The authors have no competing interests to declare.   ROC curve analysis of the apoptosis-related differential proteins.

Figure 6
Correlation between the apoptosis-related differential proteins with SPP and clincial indexes. with ESR, a close negative correlation of the sTNF-R2 with MH.