Purpose Anti-phospholipid antibodies (aPL) analyzed by line immunoassay (LIA) can recognize beta -glycoprotein I (β GPI) 2 2 domain 1 (D1) epitopes depending on β GPI binding to distinct phospholipids. The aPL LIA was compared with consensus ELISA to investigate whether both techniques can discriminate anti-phospholipid syndrome (APS) patients from aPL- positive, systemic autoimmune rheumatic diseases (SARD) patients without clinical symptoms of APS and controls. Methods Thirty-four APS patients (14 arterial/venous thrombosis, 16 pregnancy morbidity, and 4 both), 41 patients with SARD lacking clinical APS criteria but demonstrating positivity for anti-β GPI (aβ GPI) IgG, and 20 healthy subjects (HS) 2 2 were tested for aPL to cardiolipin (aCL), phosphatidic acid, phosphatidylcholine, phosphatidylethanolamine, phosphatidyl- glycerol (aPG), phosphatidylinositol, phosphatidylserine, β GPI, prothrombin, and annexin V by LIA. Samples were also tested for aCL, aβ GPI, aβ GPI-domain 1 (aD1), and aβ GPI-domains 4–5 (aD4–5) by ELISA and for lupus anti-coagulant. 2 2 2 Results Comparison of LIA with ELISA revealed a good agreement for the consensus criteria aPL aβ GPI and aCL IgG (kappa = 0.69, 0.68, respectively) and a moderate agreement for IgM (kappa = 0.52, 0.49, respectively). Regarding ELISA, aD1/aD4–5 demonstrated the best performance of differentiating APS from asymptomatic SARD [area under the curve (AUC): 0.76]. aPG IgG had the best performance by LIA (AUC: 0.72) not significantly different from aD1/aD4–5. There was a good agreement for aPG IgG with aD1/aD4–5 (kappa = 0.71). Conclusions aD1/aD4–5 (ELISA) and aPG IgG (LIA) differentiate APS from SARD patients. PG appears to interact with β GPI of APS patients and exposes D1 thereof for disease-specific aPL binding in LIA. Keywords Anti-phospholipid syndrome · Beta2 glycoprotein I · Anti-phospholipid antibody · Domain 1 · Phosphatidylglycerol Introduction rheumatic disease (SARD) such as, e.g., systemic lupus erythematosus (SLE). The APS could be associated with a Anti-phospholipid syndrome (APS) is an autoimmune disor- high risk of death in the rare catastrophic anti-phospholipid der, clinically characterized by arterial and/or venous throm- syndrome, a rapid and simultaneous multi-organ failure bosis as well as pregnancy-related complications [1, 2]. The due to generalized thrombosis . Apart from one clini- APS can be primary or secondary, depending on the absence cal criterion (vascular thrombosis and/or adverse obstetric or presence of any other related systemic autoimmune event), the revised classification criteria require the persis - tent detection of anti-phospholipid antibodies (aPL) such as anti-beta glycoprotein I (aβ GPI), anti-cardiolipin (aCL), 2 2 Cecilia Nalli and Valentina Somma shared first authorship. and/or autoantibodies interfering with coagulation [lupus anti-coagulant (LAC)] for the diagnosis of APS. Dirk Roggenbuck and Angela Tincani shared senior authorship. The recommended method to detect aβ GPI and aCL is the enzyme-linked immunosorbent assay (ELISA) using a poly- * Dirk Roggenbuck email@example.com styrene solid phase for autoantigen immobilization. However, aPL testing by ELISA still represents a challenge because of Extended author information available on the last page of the article Vol.:(0123456789) 1 3 6 Page 2 of 11 Autoimmunity Highlights (2018) 9:6 the difficulties in the inter- and intra-assay reproducibility [4 , Methods 5]. Furthermore, aβ GPI antibodies detected by ELISA have been reported in healthy adults and children. These data sup- Patients and control subjects port the hypothesis that “innocent”, non-disease associated aPL could exist, too [6, 7]. The subgroup of pathogenic aβ GPI Thirty-four patients with primary APS including 14 with antibodies seems to be mainly directed versus domain 1 (D1) arterial and/or venous thrombosis, 16 females with obstetric and not to domains 4 and 5 (D4–5), and their pathogenicity APS suffering from pregnancy-related complications, and 4 appears to be dependent on their Fc glycosylation [8–12]. having both clinical symptoms were diagnosed by character- Indeed, the former are involved in thrombotic events charac- istic international clinical and serological consensus criteria teristic of APS, whereas the latter do not interfere with the (Table 1). The patients were selected from a cohort routinely coagulation process, neither are they associated with other followed at the university hospital in Brescia. All patients clinical APS manifestations . demonstrated elevated levels of aβ GPI IgG antibodies by New assay techniques based on chemiluminescence (CIA) an in-house ELISA. This inclusion criterion was chosen to or fluorescence enzyme immunoassays for the detection of study the specificity of these antibodies against different APS-specific aPL have emerged [4 , 14]. Especially, the CIA β GPI domains. system has been shown to reduce the inter-laboratory vari- As disease controls, 41 patients with SARD and no anam- ability . nestic thrombotic and adverse pregnancy events but positiv- Of note, a novel line immunoassay (LIA) offering the ity for aβ GPI IgG [11 with SLE, 2 with systemic sclero- opportunity to test for several aPL has been reported [16, 21]. sis (SSc), 2 with Sjögren syndrome (SjS), 3 with SLE and This LIA appeared to detect preferably aPL to D1 (aD1) of the secondary SjS, 15 with undifferentiated connective tissue patient’s β GPI bound to distinct negatively charged phospho- disease (UTCD), 1 with discoid lupus erythematosus (DLE), lipids. Furthermore, aPL not related to aβ2GPI were detected, 4 with dermato/polymyositis (DM/PM), and 3 patients with too . Altogether, the former and these “non-criteria” IgG primary biliary cholangitis (PBC)] were enrolled. This group and IgM aPL to phosphatidylserine (aPS), phosphatidylinosi- was chosen due to the comparability with the disease group. tol (aPI), phosphatidylcholine (aPC), phosphatidylethanola- Furthermore, 20 healthy subjects (children) (HS) were mine (aPE), phosphatidic acid (aPA), phosphatidylglycerol included as non-diseased controls. All children were aPL (aPG), annexin V (aAnV), and prothrombin (aPT) could be negative. The study was approved by the local ethical com- used for aPL profiling and might be helpful in the clinical dif- mittee after a written informed consent from each patient. ferentiation of APS patients [17–20]. Nevertheless, the clini- All sera were stored at − 20 °C. cal meaning of “non-criteria” antibodies is still debated and β GPI is generally accepted as the major autoantigenic target ELISA for the detection of antibodies to cardiolipin recognized by APS-specific aPL [2 ]. and β GPI The novel LIA used a hydrophobic membrane for the immobilization of different phospholipids and co-factors [21, To detect classification criteria IgG and IgM antibodies to 22] (Fig. 1). In particular, negatively charged phospholipids CL and β GPI in the patient sera, commercially available could bind the patient’s own β2GPI which in turn interacted solid-phase ELISAs employing purified human β GPI in with APS-specific aD1. Thus, the specificity of this new mul- complex with CL and human β GPI alone were used, respec- tiplex reaction environment was reported to be superior to aPL tively (GA Generic Assays GmbH, Dahlewitz, Germany). consensus ELISA . In a recent study comparing aPL test- Assessment of aPL antibodies was conducted according to ing by LIA with ELISA in APS patients, asymptomatic aPL- the instructions of the manufacturer . The sera with a positive carriers, and infectious patients, the LIA demonstrated concentration equal or more than 10 U/mL for IgG and IgM, a better specificity, too . respectively, was considered positive. The same serum sam- The appearance of aPL in SARD patients without charac- ples were also analyzed by in-house assays and the results teristic clinical signs of APS is poorly understood yet. Thus, were comparable with the commercial ELISA (data not we wondered whether aPL detected by LIA or consensus shown). criteria ELISA could discriminate primary APS from SARD Research ELISAs for aD1 and aD4–5 IgG developed by without clinical symptoms of APS. Inova Diagnostics (San Diego, US) were performed as previ- ously described . A ratio of aD1 and aD4–5 with a cutoff of 1.5 was used to test sera for aPL positivity. 1 3 Autoimmunity Highlights (2018) 9:6 Page 3 of 11 6 Fig. 1 Preferential binding of anti-phospholipid antibodies (aPL) the binding of the patient’s β GPI (a) and consequently of the β GPI- 2 2 to domain 1 (D1) of patient’s beta2-glycoprotein I (β GPI) in the dependent aPL (c). After binding of β GPI to the immobilized ani- 2 2 line immunoassay (LIA). In contrast to the planar solid phase used onic PL by domain 5 (D5, containing the PL-binding site), D1 forms in enzyme immunoassays, the porous hydrophobic LIA membrane the accessible top of the induced fish-hook-like β GPI structure (b). incorporates the hydrophobic phospholipid (PL)-tail during immo- Due to the high density of negatively charged PL heads on the mem- bilization. This shields the by far larger tail of the amphiphatic PL brane, the formation of a β GPI layer with a unique D1 epitope struc- molecule from the reaction environment and, thus, prevents unspe- ture is assumed. The layer formation seems to hinder aPL binding to cific interactions. Number, orientation, and accessibility of anionic β GPI epitopes close to D5  phosphate groups of the differing hydrophilic PL heads may influence LAC testing Line immunoassay for the detection of aPL antibodies The analysis of lupus anti-coagulant (LAC) was performed according to the international recommendations . Antibodies to CL, PA, PC, PE, PG, PI, PS, the protein co- Thus, the LAC testing comprised a three-step procedure: factors β GPI, AnV, and PT were detected using a commer- cially available LIA according to the recommendations of Demonstration of a prolonged phospholipid-dependent the manufacturer (GA Generic Assays GmbH) . Pro- clotting time as screening test of hemostasis by dilute cessed LIA strips were read out densitometrically employing Russell viper venom time (dRVVT) or activated partial a scanner with the evaluation software Dr. Dot Line Ana- thromboplastin time (aPTT or lupus aPTT) analysis. lyzer (GA Generic Assays GmbH) and a grayscale calibra- Mixing patient plasma with normal plasma fails to cor- tion card for standardization. The grayscale calibration card rect the prolonged screening test(s). was provided on the template of the kit. Values were read Addition of excess phospholipid shortens or corrects off as optical density (OD) units and OD values equaling or the prolonged coagulation test (demonstration of phos- above 50 were scored positive. This cutoff was determined pholipid dependence). by calculating the 99th percentile of 150 apparently healthy 1 3 6 Page 4 of 11 Autoimmunity Highlights (2018) 9:6 Table 1 Demographic, clinical, and laboratory baseline characteristics of the 75 anti-beta glycoprotein I (aβ GPI) IgG-positive patients and 20 2 2 healthy subjects (children) Primary thrombotic or obstet- SARD (n = 41) Healthy subjects (n = 20) ric APS (n = 34) Sex, no. (%) female 31/34 (91%) 36/41 (88%) 9/20 (55%) Autoimmune disease 34/34 (100%) 41/41 (100%) 0 (0%) Primary APS 34/34 (100%) 0 (0%) 0 (0%) SLE 0 (0%) 11/41 (27%) 0 (0%) SSj 0 (0%) 2/41 (5%) 0 (0%) SLE + SSj 0 (0%) 3/41 (7%) 0 (0%) DLE 0 (0%) 1/41 (2%) 0 (0%) PBC 0 (0%) 3/41 (7%) 0 (0%) SSc 0 (0%) 2/41 (5%) 0 (0%) DM/PM 0 (0%) 4/41 (10%) 0 (0%) UCTD 0 (0%) 15/41 (37%) 0 (0%) Thrombosis 19/34 (56%) 0 (0%) 0 (0%) Arterial 7/19 (37%) 0 (0%) 0 (0%) Venous 12/19 (63%) 0 (0%) 0 (0%) Obstetric manifestations 20/34 (59%) 0 (0%) NA Pregnancy loss 15/20 (75%) 0 (0%) NA Preeclampsia 5/20 (25%) 0 (0%) NA Laboratory features LAC positivity 24/34 (71%) 18/41 (44%) NP aβ GPI IgG, median OD (25–75th percentile) 1.470 (0.929–1.747) 1.004 (0.655–1.298) 0.139 (0.047–0.444) aβ GPI IgM, median OD (25–75th percentile) 0.350 (0.165–0.576) 0.450 (0.202–0.838) 0.088 (0.049–0.132) APS anti-phospholipid syndrome, aCL anti-cardiolipin antibodies, DLE discoid lupus erythematosus, DM/PM dermato/polymyositis, LAC lupus anti-coagulant, NA not applicable, NP not performed, OD optical density, PBC primary biliary cirrhosis, SARD systemic autoimmune rheumatic disease, SLE systemic lupus erythematosus, SSc systemic scleroderma, SjS Sjögren syndrome, UCTD undifferentiated connective tissue disease individuals as recommended by the international classifi- including 41 asymptomatic patients suffering from SARD cation criteria for aPL testing and Clinical and Laboratory and 20 HS (Table2). Comparative analysis of the consensus Standards Institute (CLSI) guideline C28-A3 . criteria aPL aCL and aβ GPI in 95 sera detected by LIA and ELISA demonstrated good agreement for IgG [kappa = 0.69, Statistical analysis 95% confidence interval (CI) 0.55–0.84; 0.68, 95% CI 0.54–0.83, respectively] and moderate concordance for IgM Fisher’s exact test was performed with two-tailed probabil- (kappa = 0.52, 95% CI 0.35–0.69; 0.49, 95% CI 0.32–0.66, ity to detect the differences between groups as appropriate respectively). There was a significant difference according using Medcalc statistical software (Medcalc, Mariakerke, to McNemar’s test for aCL and aβ GPI IgM (difference: 15.8 Belgium). Inter-rater agreement statistics ware applied for and 13.7%, p < 0.05, respectively), whereas the correspond- within-group comparison. The two-tailed, Kruskal–Wallis ing IgG analyses did not reveal a significant difference for test was used to test for statistically significant die ff rences of both methods. The consensus ELISA testing for aCL and independent samples. p values of less 0.05 were considered aβ GPI is not significantly more specific than aPL analysis significant. by LIA covering 10 aPL with regard to the false positives in HS (1/20 vs 3/20, p = 0.605). Results Comparison of aPL testing in APS patients and healthy controls Comparison of aPL analysis by ELISA and LIA Both LIA and ELISA showed significantly higher preva- To identify the aPL antibody profiles by ELISA and LIA, lences of positive consensus criteria aPL (aCL and aβ GPI we tested 34 sera from patients with APS and 61 controls IgG as well as IgM) in APS patients (n = 34) compared to 1 3 Autoimmunity Highlights (2018) 9:6 Page 5 of 11 6 HS (n = 20) (p < 0.05, respectively, Table 2). In addition, the LIA revealed significantly more prevalent aPA and aPS IgG as well as IgM, and further aPG IgG, aPI IgG, and aPT IgG in APS patients (p < 0.05, respectively). Comparison of qualitative aPL testing in APS patients and disease controls The comparison of APS patients (n = 34) with asymptomatic SARD patients (n = 41) revealed significantly higher preva- lences in APS patients for the criteria aPL aCL and aβ GPI IgG detected by ELISA (p < 0.05, respectively, Table 2). The, frequency of LAC positivity was also significantly elevated in APS, whereas the frequency of triple positivity demonstrated a tendency only (p = 0.0312, 0.0713, respec- tively). In terms of aPL testing by LIA, aPG IgG, aβ GPI IgG, and aPT IgG as well as aPS IgG and IgM were sig- nificantly higher prevalent in patients suffering from APS in contrast to SARD patients. Of note, aPG IgG showed a significantly lower prevalence of 9.8% in asymptomatic SARD patients (4/41) compared to 52.9% in APS patients (18/34, p < 0.0001). The same holds true for aPT IgG with a prevalence of 17.1% in SARD (7/41) vs 50.0% in APS (17/34, p = 0.0002). Comparison of quantitative aPL testing in APS patients and controls Quantitative assessment revealed significantly different aPL IgG and IgM levels in the study cohorts regarding all consensus criteria aPL (aCL and aβ GPI) by ELISA (Kruskal–Wallis, p < 0.05, respectively (Fig. 2). In terms of LIA testing, IgG and IgM to CL, β GPI, PA, PS, PT, and PG as well as IgM to AnV demonstrated significantly different values (Kruskal–Wallis, p < 0.05, respectively). Regarding the differentiation of APS patients from asymptomatic SARD patients by consensus criteria ELISA, only aCL and aβ GPI IgG revealed significantly different quantitative levels (post hoc analysis, p < 0.05, respectively) (Fig. 2). In addition to aCL and aβ2GPI IgG detected by LIA, IgG to PA, PS, PG, PT, and IgM to PS revealed signifi- cantly higher levels in APS patients, too (post hoc analysis, p < 0.05, respectively) (Fig. 3). Comparison of the assay performance of aPL detected by ELISA and LIA To compare the diagnostic performance for the differentia- tion of APS from SARD, receiver-operating characteristic (ROC) curve analysis was performed for aPL detected by ELISA and LIA (Fig. 4). The ratio of D1 to D4–5 reactiv- ity demonstrated the best performance with an area under the curve (AUC) of 0.76 when compared with consensus 1 3 Table 2 Anti-phospholipid antibody (aPL) positives by line immunoassay (LIA) and consensus criteria enzyme-linked immunosorbent assay (ELISA) in 95 subjects including 34 patients with anti-phospholipid syndrome (APS), 41 with systemic autoimmune rheumatic disease (SARD), and 20 healthy subjects (children) (HS) as control groups ELISA LAC TP aD1/aD4–5 LIA ACL aβ GPI Any aPL aCL aPA aPC aPE aPG aPI aPS aAnV aβ GPI aPT Any aPL 2 2 G M G M G M G M G M G M G M G M G M G M G M G M APS 24 21 30 18 31 24 14 28 27 13 24 10 0 0 0 0 18 1 7 2 30 16 2 5 27 17 17 5 33 n = 34 SARD 18*** 19 25** 23 33 18*** 8 26 26 12 22 8 0 0 0 0 4* 0 3 2 27*** 9*** 0 3 22*** 12 7** 3 32 n = 41 # # ## # # ## # #### # #### # ## # # ## # HS NA 0 0 1 1 NA NA NA 1 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 1 0 1 1 3 n = 20 Comparison of APS vs SARD: *p < 0.001, **p < 0.01, ***p < 0.05 # ## ### #### Comparison of APS vs HC: p < 0.0001, p < 0.001, p < 0.01, p < 0.05 LAC and TP testing was conducted for 33 APS and 39 SARD patients only aβ GPI anti-beta glycoprotein I, aCL anti-cardiolipin, aD1 anti-domain 1, aD4–5 anti-domains 4 and 5, aPA anti-phosphatidic acid, aPC anti-phosphatidylcholine, aPE anti-phosphatidylethan- 2 2 olamine, aPG anti-phosphatidylglycerol, aPI anti-phosphatidylinositol, aPS anti-phosphatidyl-serine, aAnV anti-annexin V, aPT anti-prothrombin, LAC lupus anti-coagulant, TP triple positivity (aCL positive, aβ GPI positive, LAC positive), NA not available 2 6 Page 6 of 11 Autoimmunity Highlights (2018) 9:6 Fig. 2 Consensus criteria anti-phospholipid antibodies detected with systemic autoimmune rheumatic disease (SARD) and no adverse by enzyme-linked immunosorbent assays (ELISAs) in 95 subjects APS events, as well as 20 healthy subjects (children) (HS) as control including 34 patients with anti-phospholipid syndrome (APS), 41 group. aβ GPI anti-beta -glycoprotein I, aCL anti-cardiolipin 2 2 criteria aPL ELISAs. The AUC of this ratio was significantly [kappa = 0.71, 95% confidence interval (CI) 0.52–0.89] and higher than the AUCs of aCL and aβ GPI IgM (p < 0.05, no significant difference (McNemar’s test: difference = 5.3%, respectively) (Table 3). However, there was no significant 95% CI − 6.8 to 15.7%, p = 0.4807). difference in the prevalence of the aD1/aD4–5 ratio in APS In contrast, the strength of agreement of the aD1/aD4–5 in contrast to SARD using the cutoff of 1.5 established in ratio with all four consensus criteria aPL determined by another study previously . In accordance with the ROC ELISA was only fair (kappa < 0.4, aCL IgM, and aβ GPI curve analysis for the aD1/aD4–5 ratios of this study, a cut- IgG) or poor (kappa < 0.2, aβ2GPI IgM, and aCL IgG). off of 4.6 instead of 1.5 for the optimal differentiation of APS and SARD was required. Applying this new cutoff, there were 21/34 (61.8%) positive APS patients in contrast Discussion to 28/43 (82.4%) with the old cutoff of 1.5. Accordingly, the new cutoff lowered the prevalence of positives in the asymp- The persistent occurrence of aPL was the serological hall- tomatic SARD cohort from 26/41 (63.4%) to 5/41 (12.2%). mark of APS and was defined as a mandatory classification Consequently, the new prevalence of the APS cohort was criterion . It is a well-accepted consensus that APS- significantly higher in contrast to the one of the SARD specific aPL interact with phospholipid-binding proteins cohorts (p < 0.0001). such as β GPI or complexes thereof with phospholipids. Furthermore, there was no significant difference of the Among aPL, the correlation of aβ GPI with clinical symp- AUC for the aD1/aD4–5 ratio to the AUC of aPG IgG toms appeared to be the strongest one [2, 20, 27]. In this detected by LIA. The latter demonstrated in turn the best context, aPL binding to D1 and not to D4–5 of β GPI has performance among the aPL IgG determined by LIA and been the basis for the detection of disease-specific aPL [9 , was significantly higher than the AUC of aCL IgG (Table 3). 28]. There has been no single assay to assess all different In accordance with inter-rater agreement statistics, there was aPL subpopulations, and thus, aCL, aβ GPI, and LAC test- a good agreement for aPG IgG with the aD1/aD4–5 ratio ing have been recommended to identify all the potential aPL. 1 3 Autoimmunity Highlights (2018) 9:6 Page 7 of 11 6 Fig. 3 Anti-phospholipid IgG antibodies analyzed by line immuno- (children) (HS) as control group. aβ GPI anti-beta -glycoprotein I, 2 2 assay (LIA) in 95 subjects including 34 patients with anti-phospho- aCL anti-cardiolipin, aPG anti-phosphatidylglycerol, aPI anti-phos- lipid syndrome (APS), 41 systemic autoimmune rheumatic disease phatidylinositol, aPS anti-phosphatidylserine, aPT anti-prothrombin, (SARD), and no adverse APS events, as well as 20 healthy subjects OD optical density Fig. 4 Receiver-operating characteristics curve analysis of anti-phos- reactivity (aD4–5) by ELISA was compared with criteria aPL deter- pholipid antibodies (aPL) detected by enzyme-linked immunosorb- mined by ELISA (a) and aPL IgG by LIA (b). aβ GPI anti-beta - 2 2 ent assay (ELISA) and line immunoassay (LIA) in 34 patients with glycoprotein I, aCL anti-cardiolipin, aPG anti-phosphatidylglycerol, anti-phospholipid syndrome and 41 disease controls without clinical aPI anti-phosphatidylinositol, aPS anti-phosphatidylserine, aPT anti- APS symptoms. The ratio of anti-domain 1 (aD1) and D4–5 antibody prothrombin 1 3 6 Page 8 of 11 Autoimmunity Highlights (2018) 9:6 Table 3 Receiver-operating characteristics curve analysis of anti- occurring in patients with other autoimmune disorders like phospholipid antibodies (aPL) detected by enzyme-linked immuno- SARD not demonstrating clinical symptoms of APS. sorbent assay (ELISA) and line immunoassay (LIA) in 34 patients The agreement of aPL testing by LIA with consensus with anti-phospholipid syndrome and 41 disease controls criteria aPL by ELISA was good (IgG aPL) to moderate aPL AUC SE 95% CI (IgM aPL) and, thus, was in line with previously published comparative data [16, 21]. In addition, the favorable spec- ELISA ificity of aPL testing by LIA could also be confirmed in aD1/aD4–5 0.760 0.0597 0.647–0.851 this study revealing no significant difference for all ten aPL aβ GPI IgG 0.705 0.0617 0.588–0.805 $,§ tested by LIA compared with the four consensus criteria aβ GPI IgM 0.534*, 0.0679 0.416–0.651 aPL by ELISA. aCL IgG 0.725 0.0598 0.609–0.821 $,§ In terms of the occurrence of aPL in APS patients com- aCL IgM 0.559*, 0.0679 0.440–0.674 pared with that in SARD patients without clinical symptoms LIA of APS, only IgG consensus criteria aPL demonstrated a aβ GPI IgG 0.691 0.0638 0.574–0.793 significant difference. Although LAC testing revealed sig- aCL IgG 0.660 0.0651 0.541–0.765 nificantly different prevalences too, triple positivity analysis aPS IgG 0.716 0.0613 0.600–0.814 did not differentiate APS patients from those with SARD. aPG IgG 0.723 0.0630 0.608–0.821 Of interest, aPL IgG by LIA also revealed significantly aPT IgG 0.701 0.0628 0.584–0.801 higher prevalences in APS compared to SARD without clini- Area under the curve (AUC) was determined for the ratio of anti- cal symptoms of APS. In contrast to aCL IgG by ELISA, domain 1 of beta glycoprotein I (aD1) and domains 4–5 (D4–5) anti- however, the difference of aCL IgG by LIA did not reach body reactivity (D4–5) by ELISA and compared with those for crite- significance. Of note, aPS testing demonstrated for both ria aPL determined by ELISA and aPL IgG by LIA immunoglobulin isotypes significantly different prevalences. aβ GPI anti-beta glycoprotein I, aCL anti-cardiolipin, aPG anti- 2 2 phosphatidyl-glycerol, aPI anti-phosphatidylinositol, aPS anti-phos- Provided that the positive aPS result was due to interaction phatidyl-serine, aPT anti-prothrombin, CI confidence interval, SE of patient’s aPS with serum β GPI of the patient sample standard error having bound to immobilized PS on the LIA membrane, AUC comparison of ELISA this might add further evidence to the assumption that β GPI *p < 0.05 for the comparison to the AUC of the ratio of aD1 to D4–5 binding to negatively charged phospholipids induces spe- p < 0.05 for the comparison to the AUC of aβ GPI IgG cific conformational changes unique for each distinct phos- p < 0.05 for the comparison to the AUC of aCL IgG pholipid. Since D1 binding by aβ GPI was preferred in the AUC comparison of LIA LIA reaction environment , the differing accessibility p < 0.05 for the comparison to the AUC of aCL IgG of respective epitopes on D1 could determine the specific- ity of such aPL reactivity. Thus, the significantly reduced Triple positivity has been considered a risk factor and could prevalence of aPG IgG in SARD without clinical symptoms be used for stratification of APS patients . of APS (9.8%) compared to APS (52.9%) in this study is of A novel aPL assay technology employing a hydrophobic particular interest in this context. Of note, CL also referred membrane for aPL profiling by LIA was reported recently to as diphosphatidylglycerol represents a dimer of PG and . The LIA membrane provided a unique matrix allowing CL’s head bears two phosphate groups forming a dianion phospholipids to mimic their natural conformation required for β GPI binding . This particular setting could induce for co-factor binding as reported for other amphiphatic non- a β GPI configuration enabling sensitive binding of aPL protein antigenic molecules [30–32]. In particular, D1 of but, obviously, did not provide a reaction environment for patient’s serum β GPI appeared to be presented in the LIA the discrimination of aPL occurring in APS and asympto- reaction environment for APS-specific aPL binding more matic SARD patients. However, quantitative aPL testing did favorably than the corresponding D4–5 after the interac- reveal significantly different levels of aCL IgG by LIA like tion of serum β GPI with the immobilized phospholipids did quantitative aPA, aPS, aPG, and aPT IgG analysis by (Fig. 1). Interestingly, complexes of the patient’s β GPI with this method. Of note, quantitative aPS IgM testing did cor- differing immobilized phospholipids demonstrated different roborate the significant difference of the qualitative one. In aD1 reactivity. Altogether, this seemed to support the differ - contrast, consensus criteria’ aPL IgM analysis by ELISA entiation of disease-specific aPL in APS patients from aPL did not reveal significantly different aPL levels in APS and found in individuals with infectious disease or in asympto- asymptomatic SARD patients. This further highlights the matic carriers . Thus, we attempted to ascertain whether specificity of the LIA reaction environment for aPL analysis this novel reaction environment for the multiplex detection and the putative role of differing β GPI configurations for of aPL can discriminate aPL in APS patients from those specific aPL binding. 1 3 Autoimmunity Highlights (2018) 9:6 Page 9 of 11 6 Human and animal rights The study was approved by the local ethics Recently, the ratio of aD1 to aD4–5 was reported as committee and was conducted in accordance with the Helsinki Declara- a useful marker for APS . Surprisingly, the recom- tion of 1964 (revised 2008). mended cutoff of aD1/D4–5 did not enable differentiating APS from SARD patients in this study. Only after apply- Informed consent We hereby declare that all the patients were exam- ined in the outpatient clinic at the Division of Rheumatology of the ing ROC curve analysis and readjusting the cut-off to 4.6, University of Brescia in Italy, and they have all consented to donate significantly different prevalences in both groups were their sera for the purpose of this publication. Furthermore, this study determined. Furthermore, the ROC curve analysis of the was approved by the local ethical committee aD1/aD4–5 ratio revealed the best performance compared with the consensus criteria aPL determined by ELISA. Open Access This article is distributed under the terms of the Crea- The assay performance of aPG IgG analysis being the tive Commons Attribution 4.0 International License (http://creat iveco best amongst the aPL detection by LIA was not signifi - mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- tion, and reproduction in any medium, provided you give appropriate cantly different from the performance of the aD1/aD4–5 credit to the original author(s) and the source, provide a link to the ratio. In addition, there was no significant difference and a Creative Commons license, and indicate if changes were made. good agreement between qualitative aPG IgG analysis by LIA and D1/D4–5 ratio assessment. Thus, the LIA reac- tion environment consisting of immobilized PG interacting with the specimen’s β GPI might favor the specific binding References of aD1 as shown for aPL IgG testing recently (Fig. 1) . 1. Ruiz-Irastorza G, Crowther M, Branch W, Khamashta MA (2010) Furthermore, a specific conformation of the bound β GPI Antiphospholipid syndrome. Lancet 376:1498–1509 as a single molecule or of a β2GPI layer could minimize 2. Meroni PL, Borghi MO, Raschi E, Tedesco F (2011) Pathogenesis the binding of aD4–5 and, thus, bring about the high speci- of antiphospholipid syndrome: understanding the antibodies. Nat ficity of aPG analysis. As a fact, aD4–5 did not seem to Rev Rheumatol 7:330–339 3. Asherson RA, Cervera R, DeGroot PG, Erkan D, Boffa MC, Piette be associated with the APS phenotype . A similar LIA JC, Kamashta MA, Shoenfeld Y (2003) Catastrophic antiphospho- reaction environment with CL (diphosphatidylglycerol) lipid syndrome: international consensus statement on classifica- instead of PG demonstrated a significantly poorer assay tion criteria and treatment guidelines. Lupus 12:530–534 performance which underlines the importance of the phos- 4. Bertolaccini ML, Amengual O, Andreoli L, Atsumi T, Chighizola CB, Forastiero R, deGroot P, Lakos G, Lambert M, Meroni P, pholipid nature for specific aPL binding. Ortel TL, Petri M, Rahman A, Roubey R, Sciascia S, Snyder M, Our study has certain limitations. The relatively low Tebo AE, Tincani A, Willis R (2014) 14th International Congress number of samples results in large confidence interval (CI) on Antiphospholipid Antibodies Task Force. Report on antiphos- in particular regarding the comparison of AUC values by pholipid syndrome laboratory diagnostics and trends. Autoimmun Rev 13:917–930 ROC analysis. Furthermore, there could be a selection 5. Tebo AE, Willis R, Jaskowski TD, Guerra M, Pierangeli SS, bias for patients with SARD, since these patients were Salmon J, Petri M, Branch DW (2016) Clinical significance and recruited by aβ GPI IgG positivity. We employed research 2 correlations between anti-beta2 glycoprotein I IgA assays in ELISA for aD1 analysis which might differ from the pres- antiphospholipid syndrome and/or systemic lupus erythematosus. Clin Chim Acta 460:107–113 ently available commercial aD1 assays. 6. Nalli C, Piantoni S, Andreoli L, Motta M, Tincani A (2012) In summary, aPG IgG analysis by LIA and assessment Antiphospholipid syndrome and antiphospholipid antibodies in of the aD1/aD4–5 ratio by ELISA enabled the discrimina- children: the two sides of the coin. Isr Med Assoc J 14:310–312 tion of aPL in patients with APS from those with asymp- 7. Andreoli L, Nalli C, Motta M, Norman GL, Shums Z, Encabo S, Binder WL, Nuzzo M, Frassi M, Lojacono A, Avcin T, Meroni tomatic SARD. This finding could be helpful in clinical PL, Tincani A (2011) Anti-beta(2)-glycoprotein I IgG antibodies practice, mainly because patients at risk of adverse APS from 1-year-old healthy children born to mothers with systemic events could be identified and managed more appropri- autoimmune diseases preferentially target domain 4/5: might ately by primary prophylaxis and more frequent clinical it be the reason for their ‘innocent’ profile? Ann Rheum Dis 70:380–383 controls . 8. deLaat B, Derksen RH, Urbanus RT, DeGroot PG (2005) IgG antibodies that recognize epitope Gly40-Arg43 in domain I of beta 2-glycoprotein I cause LAC, and their presence correlates strongly Compliance with ethical standards with thrombosis. Blood 105:1540–1545 9. Andreoli L, Chighizola CB, Nalli C, Gerosa M, Borghi MO, Preg- nolato F, Grossi C, Zanola A, Allegri F, Norman GL, Mahler Conflict of interest DR has a management role and is a shareholder M, Meroni PL, Tincani A (2015) Clinical characterization of of GA Generic Assays GmbH and Medipan GmbH. Both companies antiphospholipid syndrome by detection of IgG antibodies are diagnostic manufacturers. MM is employed at Inova Diagnostics against beta2-glycoprotein i domain 1 and domain 4/5: ratio of selling autoantibody assays. TB is employed at GA Generic Assays anti-domain 1 to anti-domain 4/5 as a useful new biomarker for selling autoantibody assays. VS is an employee of Medipan. All the antiphospholipid syndrome. Arthritis Rheumatol 67:2196–2204 other authors declare that they have no financial competing and non- financial conflicts of interest. 1 3 6 Page 10 of 11 Autoimmunity Highlights (2018) 9:6 10. Mattia E, Ruffatti A, Meneghel L, Tonello M, Faggian D, Hoxha 22. Roggenbuck D, Egerer K, Feist E, Burmester GR, Dorner T A, Fredrigo M, Punzi L, Plebani M (2015) A contribution to (2012) Antiphospholipid antibody profiling—association with detection of anticardiolipin and anti-beta2glycoprotein I antibod- the clinical phenotype in APS? Arthritis Rheum 64:2807–2808 ies: comparison between a home-made ELISA and a fluorescence 23. Roggenbuck D, Somma V, Schierack P, Borghi MO, Meroni PL enzyme immunoassay. Clin Chim Acta 446:93–96 (2014) Autoantibody profiling in APS. Lupus 23:1262–1264 11. Mahler M, Norman GL, Meroni PL, Khamashta M (2012) 24. Pengo V (2012) ISTH guidelines on lupus anticoagulant testing. Autoantibodies to domain 1 of beta 2 glycoprotein 1: a promis- Thromb Res 130(Suppl 1):S76–S77 ing candidate biomarker for risk management in antiphospholipid 25. Devreese KM, Pierangeli SS, de Laat Tripodi A, Atsumi T, Ortel syndrome. Autoimmun Rev 12:313–317 TL (2014) Testing for antiphospholipid antibodies with solid 12. Fickentscher C, Magorivska I, Janko C, Biermann M, Bilyy R, phase assays: guidance from the SSC of the ISTH. J Thromb Nalli C, Tincani A, Medeghini V, Meini A, Nimmerjahn F, Schett Haemost 12:792–795 G, Muñoz LE, Andreoli L, Herrmann M (2015) The pathogenicity 26. Miyakis S, Lockshin MD, Atsumi T, Branch DW, Brey RL, Cer- of anti-β2GP1-IgG autoantibodies depends on Fc glycosylation. J vera R, Derksen RH, de Groot PG, Koike T, Meroni PL, Reber Immunol Res 2015:638129 G, Shoenfeld Y, Tincani A, Vlachoyiannopoulos PG, Krilis SA 13. Mahler M, Albesa R, Zohoury N, Bertolaccini ML, Ateka-Bar- (2006) International consensus statement on an update of the clas- rutia O, Rodriguez-Garcia JL, Norman GL, Khamashta M (2016) sification criteria for definite antiphospholipid syndrome (APS). J Autoantibodies to domain 1 of beta 2 glycoprotein I determined Thromb Haemost 4:295–306 using a novel chemiluminescence immunoassay demonstrate 27. Meroni PL (2016) Anti-beta-2 glycoprotein I epitope specificity: association with thrombosis in patients with antiphospholipid from experimental models to diagnostic tools. Lupus 25:905–910 syndrome. Lupus 25:911–916 28. Pengo V, Ruffatti A, Tonello M, Cuffaro S, Banzato A, Bison E, 14. Meneghel L, Ruffatti A, Gavasso S, Meneghel L, Tonello M, Sal- Denas G, Padayattil Jose S (2015) Antiphospholipid syndrome: van E, Pengo V, Punzi V (2015) Detection of IgG anti-domain antibodies to Domain 1 of beta2-glycoprotein 1 correctly classify I beta2 Glycoprotein I antibodies by chemiluminescence immu- patients at risk. J Thromb Haemost 13:782–787 noassay in primary antiphospholipid syndrome. Clin Chim Acta 29. Pengo V, Banzato A, Denas G, Jose SP, Bison E, Hoxha A, Ruf- 446:201–205 fatti A (2013) Correct laboratory approach to APS diagnosis and 15. Erkan D, Lockshin MD (2012) APS ACTION-antiphospholipid monitoring. Autoimmun Rev 12:832–834 syndrome alliance for clinical trials and international networking. 30. Schoenherr G, Roggenbuck D, Seifert M, Jahn S, Porstmann T Lupus 21:695–698 (1996) Technical problems arising from the use of the immu- 16. Roggenbuck D, Borghi MO, Somma V, Büttner T, Schierack P, noblot for determination of the reactivity of natural antibodies Hanack K, Grossi C, Bodio C, Macor P, von Landenberg P, Boc- with different lipopolysaccharides (LPS). J Immunol Methods cellato F, Mahler M, Meroni PL (2016) Antiphospholipid antibod- 190:185–188 ies detected by line immunoassay differentiate among patients 31. Seifert M, Schoenherr G, Roggenbuck D, Marx U, von Baehr with antiphospholipid syndrome, with infections and asympto- R (1996) Generation and characterization of a human monoclo- matic carriers. Arthritis Res Ther 18:111 nal IgM antibody that recognizes a conserved epitope shared by 17. Meroni PL, Biggioggero M, Pierangeli SS, Sheldon J, Zegers I, lipopolysaccharides of different gram-negative bacteria. Hybri- Borghi MO (2014) Standardization of autoantibody testing: a doma 15:191–198 paradigm for serology in rheumatic diseases. Nat Rev Rheumatol 32. Conrad K, Schneider H, Ziemssen T, Talaska T, Reinhold D, 10:35–43 Humbel RL, Roggenbuck D (2007) A new line immunoassay for 18. Pengo V, Biasiolo A, Pegoraro C, Cucchini U, Noventa F, Iliceto the multiparametric detection of antiganglioside autoantibodies S (2005) Antibody profiles for the diagnosis of antiphospholipid in patients with autoimmune peripheral neuropathies. Ann N Y syndrome. Thromb Haemost 93:1147–1152 Acad Sci 1109:256–264 19. Roggenbuck D, Egerer K, von Landenberg P, Hiemann R, Feist 33. Sathappa M, Alder NN (2016) The ionization properties of cardi- E, Burmester GR, Dörner T (2012) Antiphospholipid antibody olipin and its variants in model bilayers. Biochim Biophys Acta profiling—time for a new technical approach. Autoimmun Rev 1858:1362–1372 11:821–826 34. Chighizola CB, Ubiali T, Meroni PL (2015) Treatment of throm- 20. Hoxha A, Ruffatti A, Mattia E, Meneghel L, Tonello M, Salvan E, botic antiphospholipid syndrome: the rationale of current man- Pengo V, Punzi L (2015) Relationship between antiphosphatidyl- agement—an insight into future approaches. J Immunol Res serine/prothrombin and conventional antiphospholipid antibod- 2015:951424 ies in primary antiphospholipid syndrome. Clin Chem Lab Med 53:1265–1270 Publisher’s Note Springer Nature remains neutral with regard to 21. Egerer K, Roggenbuck D, Buettner T, Lehmann B, Kohn A, von jurisdictional claims in published maps and institutional affiliations. Landenberg P, Hiemann R, Feist E, Burmester GR, Dörner T (2011) Single-step autoantibody profiling in antiphospholipid syndrome using a multi-line dot assay. Arthritis Res Ther 13:R118 Affiliations 1 2 1 3 4 5 Cecilia Nalli · Valentina Somma · Laura Andreoli · Thomas Büttner · Peter Schierack · Michael Mahler · 2,4 1 Dirk Roggenbuck · Angela Tincani Cecilia Nalli Valentina Somma firstname.lastname@example.org email@example.com 1 3 Autoimmunity Highlights (2018) 9:6 Page 11 of 11 6 Laura Andreoli University of Brescia, Brescia, Italy firstname.lastname@example.org Research and Development Department, Medipan GmbH, Thomas Büttner Dahlewitz, Berlin, Germany email@example.com Research and Development Department, GA Generic Assays Peter Schierack GmbH, Dahlewitz, Berlin, Germany firstname.lastname@example.org Institute of Biotechnology, Faculty Environment and Natural Michael Mahler Sciences, Brandenburg University of Technology email@example.com Cottbus-Senftenberg, Universitätsplatz 1, 01968 Senftenberg, Germany Angela Tincani firstname.lastname@example.org Inova Diagnostics, San Diego, CA, USA 1 3
Autoimmunity Highlights – Springer Journals
Published: May 29, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera