Single-cell antibody nanowells: a novel technology in detecting anti-SSA/Ro60- and anti-SSB/La autoantibody-producing cells in peripheral blood of rheumatic disease patients

Participants underwent extensive serologic evaluations as standard of care. The serological tests were analyzed using ELISA for anti-SSA/Ro (QUANTA Lite® SS-A, catalog number 708570; Inova Diagnostics, San Diego, CA, USA); anti-SSB/La QUANTA Lite® SS-B, catalog number 708575; Inova Diagnostics); rheumatoid factor (Inova Diagnostics); and anti-Smad, anti-ribonucleoprotein, anticentromere, anti-Scl-70, and double-stranded DNA (Bio-Rad Laboratories, Hercules, CA, USA). Indirect immunofluorescence was used for Hep2 cells (antinuclear antibodies [ANAs]; Bio-Rad Laboratories), chemiluminescence immunoassay for anticardiolipin (Inova Diagnostics), and immunoturbidimetry for complement components C3 and C4. PBMCs and blood (n = 9 healthy control subjects aged 17–65 years, with 8 females and 1 male; n = 13 pSS/SLE patients with clinical profiles presented in Table 1) were either generously given by the Sjögren’s International Collaborative Clinical Alliance (SICCA) or obtained at the University of Florida Department of Pediatric Rheumatology. PBMCs from the SICCA registry were frozen samples and stored in cryoprotective media containing 90 % fetal calf serum and 10 % dimethyl sulfoxide. To maintain consistency, blood samples obtained at the University of Florida were first processed by Ficoll-Paque PLUS as instructed by the manufacturer (GE Healthcare Life Sciences, Piscataway, NJ, USA). Isolated PBMCs were frozen in cryoprotective media. Before being screened for autoantibody production, PBMCs were thawed and rested for 1 h in complete media (RPMI 1640 medium, 10 % FBS, 2 mM l-glutamine, 0.05 mM β-mercaptoethanol, 37 °C, 5 % CO₂). Cell viability was determined to be 70–85 % using 0.4 % Trypan Blue dye (Bio-Rad Laboratories) and read with an automated cell counter (TC20; Bio-Rad Laboratories). All procedures were reviewed and approved by the University of Florida Health Institutional Review Board (IRB201400079). Written informed consent was obtained from study participants before enrollment in the study.

Nanowells were fabricated using SYLGARD 184 silicone elastomer base (PDMS; Dow Corning, Midland, MI, USA) and a curing agent as described previously [17]. A suspension of 2 × 10⁵ cells in 100 μl of complete culture media was stained with CellTrace Calcein Violet, AM, for live cells and anti-CD19-Alexa Fluor (AF) 488 for 30 minutes on ice (Life Technologies, Carlsbad, CA, USA). Stained cells were washed in PBS and resuspended in 300 μl of culture media. The suspension of cells was loaded into an array of 50-μm nanowells. The cells were allowed to settle via gravity for 5 minutes. Excessive cells were rinsed off with media, and a LifterSlip coverslip (Fisher Scientific, Pittsburgh, PA, USA) was placed on top to prevent evaporation from the nanowells. The arrays were imaged using an automated epifluorescence microscope (Nikon Eclipse Ti; Nikon Instruments, Melville, NY, USA) equipped with a motorized stage, phase contrast, and 405-nm and 488-nm wavelength filter sets using Nikon NIS Elements Advanced Research image capture software.

The array of nanowells was submerged in media to be rewetted after imaging. The array was then gently placed in the chamber base of a hybridization chamber (Agilent Technologies, Santa Clara, CA, USA), and excess liquid was aspirated off using a glass pipette. The face of each dry glass slide treated with poly-l-lysine as described previously [17] was coated with polyclonal donkey antihuman immunoglobulin G (IgG) (25 μg/ml; Jackson ImmunoResearch, West Grove, PA, USA) was placed on top of the array, which was placed inside the chamber base. The assembly was secured by a finger-tightened screw and incubated at 37 °C for 1 h. After incubation, the glass slides were carefully removed from the array and immediately placed in 1× PBS. The glass slides were processed using the HS 400 Pro Hybridization system (Tecan, Männedorf, Switzerland) with the following protocol: 15-minute hybridization with 3 % nonfat milk in PBS with 0.5 % Tween 20, washed twice for 1 minute each time, and incubation for 45 minutes with a 1:1000 dilution of goat antihuman IgG-AF647 (Jackson ImmunoResearch), AF488-labeled SSA/Ro60, and AF550-labeled SSB/La using DyLight antibody labeling kits per the manufacturer’s instructions (Thermo Scientific, Rockford, IL, USA). Ro60/SS-A and La/SS-B were purchased from a commercial source (DIARECT, Freiburg, Germany). The slides were vacuum-dried and scanned using a GenePix 4400 microarray scanner (Molecular Devices, Sunnyvale, CA, USA) with specific gain and power to maintain consistency between fluorescence channels as well as among subsequent slides.

Microarray micrographs and microscopic images were processed to identify nanowells containing single cells with corresponding secretion of IgG antibody against Ro60/SS-A and La/SS-B proteins. In brief, GenePix Pro-7.0 software (Molecular Devices) was used to locate positive features on the scanned images of printed microarrays using a custom GenePix Array List designated with feature indicators. Once all the features were found, each position in the array was analyzed to extract the mean fluorescence intensity (MFI) for each channel corresponding to immunoglobulin. The data were extracted on the basis of specific criteria for each channel. This included setting the percentage of saturation to any value less than 2, the coefficient of variation to a maximum of 100 to indicate signal uniformity, the signal-to-noise ratio to greater than or equal to 1, and the SD above the background (% > B + 2 SD) at a minimum of 50. These criteria ensured that affirmative and uniform signals above the background noise were selected and reduced the chances for false-positive results. Analysis of the images of the cells recorded by automated epifluoresence microscopy were inspected by using a custom software program to determine the number of cells present in each well and the MFI in each of the fluorescent channels. These data were matched with the corresponding antibodies detected by microengraving according to the unique location identification of each nanowell. This combined dataset was then filtered for analysis to include only wells occupied with single live cells.

A standard curve for the antibodies produced by each cell was constructed by applying a series of concentrations (e.g., 1 nM to 10 μM) of the corresponding fluorochrome-conjugated antihuman IgG (antihuman IgG-AF488 for SSA/Ro60 and antihuman IgG-AF550 for SSB/La) to the set of replicate microarrays and then measuring the fluorescence intensities of captured anti-SSA/Ro60 and anti-SSB/La as a function of concentration [18, 19]. Based on the standard curve, the concentrations of anti-SSA/Ro60 and anti-SSB/La were calculated using MFI of individual signals on the microarray from each CD19⁺ B cell in nanowells. Based on the unknown concentrations and the MFI values, the linear regression analysis generated for anti-SSA/Ro60 was y = 0.0008x + 0.0122, R ² = 0.97609, and for anti-SSB/La it was y = 205,706x + 124.22, R ² = 0.99995.

Statistical evaluations were performed by using the Mann-Whitney U test generated using InStat software (GraphPad Software, La Jolla, CA, USA). A one-tailed p value less than 0.05 was considered significant.

To demonstrate the feasibility of SCAN methodology, PBMCs from pSS/SLE patients and healthy control subjects were isolated. As presented in Fig. 1, the fluorescently labeled cells were dispersed into the nanowells and imaged to precisely locate specific nanowells that contained individual live B cells. Capture slides containing human anti-IgG were used to hybridize the arrays containing cells. To identify specific captured Ig, detection reagents, including antihuman IgG-AF647, SSA/Ro60-AF488, and SSB/La-AF550, were used to determine the antigen specificity of B cells isolated from patients and healthy control subjects. As demonstrated in Fig. 2a, with SCAN technology we were able to identify an individual live CD19⁺ B cell in each nanowell, but, more importantly, we were able to analyze the IgG-specific autoantibodies produced by these individual CD19⁺ B cells, as represented by IgG, anti-SSA/Ro60, and anti-SSB/La. On the basis of the microarrays of the secreted autoantibodies, nanowells containing a single CD19⁺ B cell were identified to be either negative for anti-SSA/Ro60 and anti-SSB/La or positive for anti-SSA/Ro60 or anti-SSB/La, but not both. This observation points to the astute specificity of SCAN technology in detecting anti-SSA/Ro60 and anti-SSB/La autoantibodies.

The application of SCAN technology to PBMCs of pSS/SLE patients or healthy control subjects shows a quantitative difference between binding of SSA/Ro60 and SSB/La antigens. As presented in Fig. 2b, when normalized against nanowells that contained single live cells, patients showed a higher number of CD19⁺ B cells producing IgG in comparison to healthy control subjects. In addition, CD19⁺ B cells of patients produced a higher frequency of anti-SSA/Ro60 and anti-SSB/La autoantibodies than those of healthy control subjects. Combining the two datasets indicated that patients’ CD19⁺ B cells secreted significantly elevated levels of anti-SSA/Ro60 and anti-SSB/La autoantibodies with IgG isotype in comparison to control subjects. The striking aspect of these results is the detection of anti-SSA/Ro60 and anti-SSB/La autoantibodies in healthy control subjects, which were determined to be negative for autoantibodies by ELISA. These results demonstrate that SCAN is a sensitive and multiparametric technology with high specificity that can be used to profile the isotype and autoantibody specificity of individual B cells in patients with pSS and patients with SLE.

The standard curves were generated to calculate the concentration of each anti-SSA/Ro60 and anti-SSB/La signal. As presented in Fig. 3a, individual CD19⁺ B cells of all patients produced various concentrations of anti-SSA/Ro60 based on individual microarray spots. Interestingly, three of nine (patient number 2, 1626, and OF) of the healthy subjects were shown to produce anti-SSA/Ro60 autoantibody, but at significantly lower concentrations (calculated by spot signal intensity) than the patients (Fig. 3b). Six of thirteen of the patients (patient numbers 2.4, 2.24, 14, 15, 24, and 25 in Table 1) were positive for anti-SSA/Ro60 by SCAN technology, even though they tested negative for anti-SSA/Ro60 by ELISA. Similarly, individual B cells of pSS/SLE patients secreted different amounts of anti-SSB/La autoantibody (Fig. 3c). Some of the anti-SSB/La-negative patients (6 [46 %] of 13) (Table 1) produced large amounts of anti-SSB/La autoantibody determined by using SCAN technology. Analysis of the healthy control subjects revealed that six of nine healthy subjects secreted significant levels of anti-SSB/La autoantibody with overall concentration equivalent to that of the patients (Fig. 3d). The results indicate that SCAN is capable of precisely quantifying the amount of anti-SSA/Ro60 and anti-SSB/La autoantibodies produced by individual B cells. In addition, the data indicate that patients produced higher concentrations of anti-SSA/Ro60 autoantibodies in comparison to healthy control subjects. Last, the secreted levels of anti-SSA/Ro60 autoantibodies by individual B cells appeared to be highly elevated in pSS/SLE patients; however, there was no change in the amount of anti-SSB/La autoantibody secreted by individual B cells in both cohorts.