Non‐invasive Biomarkers for Systemic Lupus Erythematosus: A Lookback at 2016
Ling Qin,Chandra Mohan
DOI: https://doi.org/10.1111/1756-185x.13075
2016-01-01
International Journal of Rheumatic Diseases
Abstract:Early diagnosis, accurate assessment of disease activity and monitoring of disease flares in patients with systemic lupus erythematosus (SLE) and lupus nephritis (LN) remain huge challenges due to the lack of validated biomarkers with good sensitivity and specificity profiles. Conventional clinical parameters, such as anti-double stranded DNA (anti-dsDNA), complement C3 and proteinuria, are not sensitive enough for the early diagnosis and monitoring of disease activity, renal flares or long-term outcomes. In recent years, several promising biomarkers for the early diagnosis and monitoring of SLE and LN have been evaluated. Here, we take stock of the largest studies that have been reported over the past year, 2016. SLE is an autoimmune disease with multiple organ involvement, and diverse clinical manifestations. LN is one of the commonest and most ominous clinical manifestations of SLE and accounts for significant morbidity and mortality in SLE patients. Although improvements in 5- and 10-year SLE/LN survival rates have been reported in recent years, early diagnosis and monitoring disease flares remain a big challenge. Traditional clinical and laboratory parameters, such as anti-dsDNA, C3, proteinuria, and estimated glomerular filtration rate (eGFR) are not sensitive for the early diagnosis and monitoring of treatment responses. A renal biopsy is the gold standard for the diagnosis and prognosis of LN, but it is invasive and not practical for routine monitoring of disease activity. An increasing number of studies focus on identifying non-invasive biomarkers for the early diagnosis and monitoring of SLE and LN. Three recent articles published in this journal focus on different aspects of SLE biology. Tozkir et al. have reported that the frequencies of classes I and II anti-human leukocyte antigen (HLA) antibodies were higher in SLE and scleroderma patients than in healthy controls. Interestingly, frequencies of thrombocytopenia, anti-ribonucleoprotein and anti-Ro antibodies were higher in the class I antibody-positive SLE group.1 In a second study, Banica et al. reported that SLE patient peripheral blood mononuclear cells (PBMCs) exhibited a cluster of anergy-related phenotypes, including reduced regulatory T cells (Tregs), Cbl-b and egr2 message, although association to disease activity or SLE manifestations were not examined.2 In a third article, Huang et al. reported that macrophages from SLE patients exhibited reduced phagocytic capacity compared to those from healthy controls.3 The SLE-associated phenotypes described in these three recent articles may also have potential biomarker applications, although the specificity and sensitivity of these phenotypes were not pursued in these studies. Regardless, when the literature beyond this journal was surveyed, the past year has witnessed a flurry of articles in the arena of SLE biomarkers. When we searched PubMed, Medline and Embase databases for original English-language articles published in 2016 using the following search terms (title or abstract), ‘systemic lupus erythematous’ OR ‘SLE’ OR ‘lupus nephritis’ OR ‘LN’ and ‘biomarkers’ OR ‘RNA’ OR ‘protein’ OR ‘metabolites’ that focused only on human studies, 31 articles satisfied these criteria. From these, 17 publications that had utilized the largest sample sizes (n ≥ 70) are summarized in Table 1.4-20 Highlights from these studies are summarized below. Plasmablast signature Neutrophil signature Plasmablast signature correlated with disease activity. It can track with disease activity and predict treatment responses longitudinally. Neutrophil-related signature was associated with lupus nephritis. Different transcriptional responses were noted in response to treatment. miRNA-21 miRNA-196a miRNA-181a All three miRNAs correlated positively with anti-dsDNA and negatively with complement C3 and C4. miRNA-21 and miRNA-196a were independent risk factors for disease progression. SLE = 159 (LN = 57) RA = 40 HC = 100 SLE = 209 (LN = 138) CKD = 75 HC = 32 CKD = 18 HC = 20 IGFBP-2 correlated with both global disease activity and renal disease activity in LN. IGFBP-2 levels positively correlated with concurrent renal pathology chronicity index. SLE = 96 (LN = 65) Serum IL-18 correlated with global disease activity and renal flares in pediatric SLE. Baseline IL-18 predicted long-term renal outcome, including ESRD in pediatric LN. Serum IL-34 was increased in active SLE and correlated with disease activity. Serum IL-34 decreased with remission in active SLE after treatment. Nonrenal SLE = 115 LN = 288 PTX3 correlated with disease activity and renal pathology activity indices in LN. Serum pentraxin decreased with remission in LN. S100A8/A9 S100A12 S100A8/A9 and S100A12 correlated with disease activity in SLE. S100A8/A9 can predict treatment response in SLE and LN. sVCAM-1 showed strong correlation with disease activity sVCAM-1 distinguished the treatment responders from non-responders. sVCAM-1 correlated with numbers of activated memory B cells and plasma cells in SLE blood. Anti-C1q IgG anti-C3b IgG SLE = 114 (LN = 73) Anti-C3b IgG showed a specificity of 98% for LN. Combined with anti-C1q IgG, anti-C3b IgG tracked with LN flare. SLE = 121 (Active LN = 58) HC = 20 RA = 20 DN = 20 Urinary MCP-1 was increased in active LN and correlated with renal SLEDAI and SLEDAI. MCP-1 dropped with treatment and increased before relapse. MCP-1 tracked with disease activity and predicted poor treatment response and relapse in LN. SLE = 121 (Active LN = 58) Urinary OPG was increased in active LN, correlating with renal SLEDAI. Urinary OPG decreased with treatment and increased before relapse of LN in some patients. Non-renal SLE = 115 LN = 288 Urine LFABP, MCP-1, albumin and transferrin were increased in LN. This urine protein panel distinguished patients who developed renal dysfunction from those who maintained renal function. Only one of the 17 publications used a transcriptomics-based biomarker discovery approach. RNA signatures in PBMC were pursued in a longitudinal study of 158 pediatric SLE patients up to 4 years. The authors identified a plasmablast signature as a biomarker of SLE disease activity, comprised of 383 differentially expressed transcripts (DETs) positively correlated with disease activity and 103 DETs negatively correlated with disease activity.4 In addition, they detected a neutrophil signature that was associated with active LN and distinct changes in transcriptomic signatures in responses to treatment, according to LN class. Quantitative set analysis for gene expression showed that neutrophil, myeloid lineage and inflammation-related transcriptomic modules were only enriched in LN, associated with renal involvement. Weighted gene co-expression network analysis was used to identify modules of longitudinally co-expressed transcripts in individual SLE patients. A module/trait correlation matrix was also built for each patient. Use of modules that best correlated with SLE Disease Activity Index (SLEDAI) permitted personalized SLE transcriptional immune-monitoring that could guide individualized therapeutic interventions. The authors also reported that SLE patients can be stratified into seven groups based on their transcriptional correlates of disease activity. These elegant studies raised hope that immune signatures that are correlated with disease activity and/or associated with different genotypes or phenotypes may facilitate accurate disease monitoring and help tailor individualized treatment strategies in SLE. MicroRNAs (miRNAs) play a critical role in multiple organs and systems, including regulating immune cell proliferation, differentiation and regulation. In 2016, one publication assayed plasma miRNAs in 70 SLE patients. Motawi et al. reported that plasma miRNA-21, miRNA-181a and miRNA-196a positively correlated with ds-DNA titers and negatively correlated with complements C3 and C4 in SLE patients. miR-196a exhibited good sensitivity and specificity (area under the curve [AUC] = 0.76) in discriminating SLE patients from healthy controls. Multivariate analysis revealed that miRNA-21 and miRNA-196a were significant independent risk factors for disease progression.5 Proteomics and metabolomics are emerging tools for the screening of serum/plasma biomarkers in SLE, LN and other chronic diseases. An increasing number of serum/plasma proteins and metabolites have been identified as potential biomarkers for the early diagnosis or monitoring of SLE and LN. Eleven articles (each with sample sizes ≥ 70) documented serum/plasma biomarkers in SLE and LN in 2016. Most of them were cross-sectional studies (Table 1). Being a regulatory receptor on B cells, CD72 mediates the proliferation and activation of B cells. It was reported that serum soluble CD72 was elevated in SLE patients, particularly in active LN patients.6 SLE patients with positive anti-dsDNA antibodies or anti-cardiolipin antibodies showed higher sCD72 levels compared to those with negative antibodies. Moreover, sCD72 levels correlated positively with SLEDAI. Another cross-sectional study focused on human epididymis protein 4 (HE4), a tumor biomarker for ovarian cancer. Data from Yang et al. showed that serum HE4 levels were increased in SLE, particularly in those patients with renal involvement. Also, serum HE4 was associated with SLEDAI in SLE and serum HE4 levels distinguished chronic kidney disease (CKD) controls from LN patients.7 Yet another cross-sectional study focused on insulin-like growth factor binding protein-2 (IGFBP-2), a protein known to regulate insulin-like growth factor activity. Serum IGFBP-2 was found increased significantly in LN patients and could discriminate LN patients from other CKD patients and healthy controls.8 In addition, IGFBP-2 was a potential indicator of both global disease activity and renal disease activity in LN patients, and showed positive correlation with serum creatinine and proteinuria. This was one of a very few studies that had also examined serum biomarker levels in LN at the same time that renal biopsy was performed. Study of serum IGFBP-2 and concurrent renal pathology in paired serum/biopsy sample revealed that serum IGFBP-2 correlated with the chronicity index of renal pathology concurrently assayed. Three of the serum protein biomarker papers published recently had a longitudinal component, focusing on interleukin (IL)-10, IL-18 or vascular cell adhesion molecule 1 (VCAM-1). IL-10 possesses an anti-inflammatory role and can inhibit the activation of macrophages and T cells. In a recent longitudinal study of 141 SLE patients, higher serum IL-10 levels were documented in active SLE patients compared to inactive disease, and a significant positive association of serum IL-10 with disease activity was observed.9 Importantly, high first visit IL-10 level was a predictor of high subsequent disease activity. The authors concluded that IL-10 may be a potential biomarker of future disease activity in SLE. As opposed to IL-10, IL-18 is a pro-inflammatory factor, has shown to play an unfavorable role in the pathogenesis of SLE. Recently, serum IL-18 was reported to be positively correlated with SLE disease activity among pediatric-onset Chinese SLE patients. In addition, serum IL-18 at 6 months post-treatment follow-up was the strongest risk factor for poor prognosis, irrespective of the initial renal status. For LN patients, serum IL-18 levels at 6-month post-treatment were also shown to predict long-term renal outcome.10 As an alternative ligand for the colony-stimulating factor-1 receptor (CSF-1R), IL-34 plays an important role in the biology of mononuclear phagocytic cells. Wang et al. showed that serum IL-34 positively correlated with SLE disease activity, anti-dsDNA titers, and C-reactive protein levels and negatively correlated with complement C3 levels. Serum IL-34 decreased after effective treatment.11 The biomarker potential of this interleukin warrants further independent validation. Pentraxin 3 (PTX3) is an acute-phase protein induced by inflammatory triggers. It can regulate the innate immune response and modulate the pathogenesis of autoimmune diseases. Serum pentraxin 3 showed positive correlation with SLEDAI, serum creatinine, renal pathological activity indices and complement C3 in active LN patients in a cross-sectional analysis of 115 SLE and 288 LN patients of Chinese ethnicity.12 S100A8, S100A9 and S100A12 are major calcium-binding pro-inflammatory proteins that participate in immune responses. Tydén et al. reported that serum levels of pro-inflammatory proteins S100A8/A9 and S100A12 were elevated in SLE patients with high disease activity compared to patient samples with low disease activity.13 In particular, increased levels of S100A8/A9 were found in patients with positive anti-dsDNA antibodies and glomerulonephritis before treatment. After receiving immunosuppressive treatment, the patients showed decreased serum S100A8/A9 levels. The authors conclude that serum S100A8/A9 may be used as a biomarker to monitor disease activity and response to treatment in SLE patients, particularly in patients with glomerulonephritis. Triggering receptor expressed on myeloid cells-1 (TREM-1) belongs to the immunoglobulin superfamily of molecules. It has an established role in innate and adaptive immune response. Bassyouni et al. reported that plasma TREM-l levels were positively correlated with SLEDAI, IL-6 and hypoalbuminemia.14 The authors conclude that soluble TREM-1 levels may be a promising marker of disease activity in SLE, although longitudinal studies were not performed. VCAM-1 is induced on endothelial cells by various inflammatory cytokines. In several earlier studies, SLE patients exhibited higher soluble VCAM-1 levels. Changing levels of (delta measurements) soluble VCAM-1 demonstrated stronger correlation with a change in disease activity than conventional SLE biomarkers in a recent longitudinal analysis of 80 SLE patients. Delta soluble VCAM-1 was shown to distinguish SLE responders from non-responders to treatment and performed better than conventional SLE biomarkers.15 Besides serum proteins, one recent study has examined the performance of serum autoantibodies in SLE. Both anti-C1q immunoglobulin G (IgG) and anti-C3b IgG levels were reported to be higher in SLE and LN patients.16 Interestingly, anti-C3b IgG was associated with LN patients who experienced at least one flare. Longitudinal analysis indicated that circulating anti-C1q IgG was elevated at the time of LN flare only in patients who were anti-C3b positive. The authors noted that serum anti-C3b IgG was particularly specific for LN, and may be useful for monitoring LN activity, although it was not clear how these antibodies performed compared to anti-dsDNA IgG. Over the past year, one study examined serum metabolite signatures in 80 SLE patients.17 The documented metabolite changes were related to amino acid turnover or protein biosynthesis, saccharo-metabolism, lipid metabolism and gut microbial metabolism. Glutamate and 2-hydroxyisobutyrate levels were increased while citrate, glycerol, linoleic acid, and propylparaben levels were decreased in active SLE patients compared to inactive SLE patients and healthy controls. Specific metabolites in the metabolomic panel were correlated with SLEDAI scores in SLE. Moreover, when a panel of six endogenous metabolites was combined with C3 and C4, high diagnostic accuracy for distinguishing active SLE from inactive SLE was obtained.17 Further independent validation is warranted. Four recent studies have focused on potential urine markers of SLE/LN in 2016, most of which have already been reported in previous years. Monocyte chemoattractant protein-1 (MCP-1) is a chemokine that can attract mononuclear cells to inflammatory sites and induce inflammatory responses. A longitudinal study indicated that baseline urine MCP-1 was significantly higher in active LN compared to active SLE without renal involvement, rheumatoid arthritis and healthy controls. It showed positive correlation with renal SLEDAI and SLEDAI. Urine MCP-1 performed better than serum MCP-1, anti-dsDNA antibodies, C3 and C4 in differentiating between active LN patients and active SLE patients without renal involvement. The authors conclude that urine MCP-1 levels were decreased after treatment and may be used as a biomarker to monitor the treatment response as well as a predictor of relapse of LN.18 Osteoprotegerin (OPG) is secreted by osteoblasts and other cells in the immune and renal systems. It was recently reported that urinary OPG (uOPG) was significantly elevated in active LN than active non-renal SLE and inactive SLE. uOPG also showed positive correlation with renal SLEDAI and SLEDAI and performed well in differentiating active LN from active non-renal SLE. Longitudinal analysis indicated that uOPG decreased significantly after treatment at all follow-up visits.19 Active LN patients showed higher urinary PTX3 levels compared with patients in remission and normal controls. Significant correlations between urinary PTX3 levels and multiple tubule-interstitial lesions markers, such as urinary kidney injury molecule-1, neutrophil gelatinase-associated lipocalin (NGAL), microalbumin, transferrin levels and renal pathological indices scores, especially interstitial inflammation, were found in active LN patients.12 In another recent study, liver-type fatty acid binding protein (LFABP), albumin, MCP-1 and transferrin levels were higher in LN patients with renal function decline compared to those who maintained renal function. The combination of LFABP, MCP-1, albumin and transferrin showed good predictive accuracy for renal function decline in both pediatric and adult cohorts in a longitudinal analysis.20 2016 has been a very active year, with 31 articles evaluating various biomarkers for SLE and/or LN. Although most of these were cross-sectional in nature, a couple of these reports included longitudinal analyses and renal biopsy-concurrent analyses of blood or urine biomarkers. Most of the studies investigated protein biomarkers, while a few evaluated RNA transcripts or metabolites. Going forth, several issues need to be addressed. First, it appears that most of the studies were biased toward the evaluation of one pre-selected biomarker. It would be important to cast wider screens employing contemporary ‘OMIC’ technologies in order to uncover the most predictive biomarkers. Second, given that several markers are showing biomarker potential, it becomes important to cross-validate leading biomarkers in several ethnically different patient cohorts, and to assemble biomarker panels. Third, the specificity of the examined biomarkers for SLE and LN warrant further characterization. Finally, it is possible that the genetic makeup of individual patients may dictate which biomarkers may be most informative, and this needs to be dissected out systematically. There is no doubt that the coming years will witness a steady growth in biomarker research, and this augurs well for the management of SLE and LN in the clinics. Ling Qin played a role in screening the literature and writing the manuscript. Chandra Mohan designed the review and edited the manuscript. This study was supported by grants from National Natural Science Foundation of China (81571574) and ‘Thousand Foreign Talents Program of China’ (WQ20133100165). None.