Serum samples from patients and healthy volunteers were obtained from the Rheumatology Service of the Hospital de Clínicas José de San Martín, Buenos Aires, Argentina. All studies were conducted, and all samples were obtained with written, informed consent under Institutional Review Board on Human Ethics approved protocols from the mentioned Hospital. The study has been performed in accordance with the ethical standards laid down in the 1964 World Medical Association Declaration of Helsinki.

Animals were used in compliance with the research animal use guidelines established by the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET-Argentina) and with the EC Directive 86/609/EEC [34]. The protocol was approved by the Institutional Review Board on Ethics of Animal Experiments of the Instituto de Estudios de la Inmunidad Humoral (IDEHU) from CONICET-University of Buenos Aires.

A cohort of 169 individuals were examined at the Rheumatology Service of the Hospital de Clínicas José de San Martín, as part of a previous larger study [24], then 10 of those patients were selected for this study, according to their clinical characteristics and laboratory findings. Diagnoses of SLE, SS, Rheumatoid Arthritis (RA) and CP were made according to the criteria adopted in the abovementioned previous study.

Twenty milliliters (ml) of serum were obtained from each of the 10 patients and stored at –20˚C until the purification processes began. Sera from 10 healthy subjects (normal human serum, NHS) were used as control; 1 ml of each one was pooled and stored at –20˚C. All 10 NHS obtained were tested individually for ANA, anti-Ro/SS-A 52 and anti-Ro/SS-A 60 kDa, being the results always negative. Immunoglobulin G (IgG) concentration was measured in sera from the 10 patients and also in the NHS pool by radial immunodiffusion (Diffu-Plate, Biocientífica, Argentina).

Recombinant human Ro/SS-A 52 kDa and Ro/SS-A 60 kDa (Ro52 and Ro60 hereafter) were cloned and expressed in an E-coli system, where a N-terminal 6× Histidine-tag was added to the proteins and then they were purified using an immobilized metal ion adsorption chromatography, as described before [24]. Aabs against either Ro52 or Ro60 proteins were isolated from sera by affinity chromatography, using the corresponding recombinant protein coupled to CNBr-activated Sepharose 4B (GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom) according to the manufacturer’s instructions. Briefly, 1 ml of CNBr-activated resin was mixed with 2 ml of protein (1 mg/ml) for 2 h at room temperature using an end-over-end mixer; next unreacted groups were blocked with Glycine 0.2 M pH 8, for 2 h at room temperature and finally the resin was washed with phosphate buffered saline (PBS). Afterwards, 2 ml of patient’s serum were allowed to recirculate through the affinity column for 1 h, in order for the specific Aabs to bind to their antigen. Washing steps were performed first with PBS, then with PBS-NaCl 0.5 M. Anti-Ro52 or anti-Ro60 Aabs were eluted with Glycine 0.1 M pH 3.5 and then pH 2.7, using an ÄKTA purifier system (GE Healthcare). Eluted fractions (1 ml each) were immediately neutralized with 1 M Tris-HCl pH 8 and subsequently tested for anti-Ro reactivity by a heterogeneous non-competitive ELISA, coated with the recombinant proteins, as described before [24]. High reactivity fractions were pooled (one pool for fractions eluted at pH 3.5 and another for 2.7) and dialyzed against PBS. Pooled fractions were titrated by ELISA and their protein concentration determined by BCA Protein Assay Kit (Pierce, Rockford, IL, USA) using the extended protocol suggested by the manufacturer, in order to increase the sensitivity of the detection. Finally they were stored at –80˚C in glycerol 10% until the flow cytometric analysis began.

Reactivity against native Ro proteins was also evaluated for the pooled fractions, using an indirect immunofluorescence technique with commercially available Hep-2 cells (Kallestad, BIORAD, Houston, TX, USA), according to the manufacturers’ instructions.

A human keratinocyte cell line, HaCaT, was kindly provided by Dr N. E. Fusenig (German Cancer Research Center, Heidelberg, Germany) [35]. HaCaT cells were cultured in plastic dishes and subconfluent cells were irradiated with 50 mJ/cm² of UVB light as previously described [36,37] in order to obtain a high percentage of apoptotic cells. At least one dish was not exposed to UVB irradiation, but was handled in the same fashion as the irradiated cells, that is to say the culture medium was replaced by PBS before the irradiation. After irradiation, the PBS was immediately replaced with DMEM. All detached and attached cells were harvested and collected 24 h after UVB irradiation and they were incubated with PBS-total rabbit serum 1/10, for 2 h at 4˚C before the labeling for flow cytometry, to block Fc receptors in order to prevent unspecific binding of Igs.

Aabs binding to cell surface was measured by flow cytometry, incubating total irradiated cells (1 × 10⁶ cells) with patients’ sera or affinity-purified anti-Ro52 or antiRo60 Aabs for 2 h at 4˚C, followed by rapid fixation with p-formaldehyde 2%. Using this method, binding to cell surface is allowed and subsequent internalization of Aabs is prevented. Sera were used at a dilution of 1:100 in PBS-1% BSA and purified Aabs were assayed undiluted. This was done in order to make the anti-Ro reactivity equivalent both in sera and purified Aabs for every patient. NHS (1:100), human purified IgG (hIgG, 0.1 mg/ml) and secondary antibody alone (UBC—unspecific binding control) were used as controls. Sera and NHS IgG masses used were equal, hIgG control was used in a concentration equivalent to one of the purified Aabs fraction which presented the highest mass. After washing twice with cold PBS, cells were incubated with secondary antibody FITC-conjugated goat anti-human IgG H + L (Bethyl, Montgomery, TX, USA) at a dilution of 1:100 in PBS-1% BSA, for 2 h at room temperature. After the final washing with PBS cells were analyzed with a PAS-III PARTEC flow cytometer (Münster, Germany). In each experiment, the same first antibody labeling (sera, purified Aabs, NHS and hIgG) was also performed on non-irradiated cells.

In vivo assays were performed using male and female Crl: SKH-1-hrBR Hairless mice of 8 weeks of age, purchased from Charles River Laboratories. Animals were housed in quarters with 12/12 h light/dark cycle and maintained with water and food ad libitum.

Mice were divided into 4 groups of 10 animals, and a conventional intraperitoneal immunization schedule was performed with 100 µg of each protein in complete Freund’s adyuvant (CFA). Immunizations (total of 6) were made in 20-day intervals and sample taking (total of 3) in 40-day intervals. Groups were immunized with recombinant human Ro52 (Group 1), recombinant human Ro60 (Group 2), recombinant human Ro52 and Ro60 (Group 3), and the vehicle buffer in CFA (Group 4). Blood samples were taken by facial vein puncture and tested for reactivity against recombinant human Ro proteins by ELISA, as described before [24], and also against mice spleen extract in order to test autoreactivity [38].

When the animals’ anti-Ro autoreactivity values became different from the control (group 4) in a statistically significant way, each group was divided in two, one group was irradiated (5 mice) and the other was mockedirradiated (5 mice). They were irradiated on their back with one minimal erythema dose (200 mJ/cm²) of UVB, as described before [36]. After irradiation, animals were kept in the same conditions for 24 h and they were subsequently sacrificed. Non-irradiated mice were used as control and they were handled in the same fashion as the irradiated animals. Mice were sacrificed using a CO₂ gas chamber, then dorsal skin tissue was obtained, one part was used for histological analysis and the rest was used to prepare epidermal homogenates.

In order to prepare epidermal homogenates total dorsal skin was immersed 30 s in water at 60 ˚C, then the epidermis was scraped from the dermis using a blade and homogenized in 2 ml of PBS, with 0.5 mg/ml EDTA and 0.174 µg/ml PMSF (Sigma Chemical Co) in a glass homogenizer with Teflon pestle, then centrifuged at 4˚C for 15 min at 12,000 rpm. Afterwards, total protein concentration in the supernatant was measured with the BCA Protein Assay kit (Pierce), using bovine serum albumin (BSA) as standard. Finally, the epidermal extracts were stored at −70˚C until the assays were done.

Cytokine quantification was performed by ELISA. Mouse TNF-α (epidermis extracts) was quantified using a non-competitive ELISA assay (BD Biosciences, San José, CA, USA), values were expressed as pg/mg of protein. All assays were repeated at least three times.

Skin samples for histological analysis were fixed with 4% neutral formalin, and embedded in white paraffin. Samples were cut in 3 µm sections and stained with hematoxylin-eosin (H-E).

Flow cytometric data was analyzed using Cyflogic software (version 1.2.1, CyFlo ltd, Turku, Finland). Graphical and statistical analyses were performed with GraphPad Prism 5.0 (GraphPad Software Inc., San Diego, CA, USA). For the in vitro model Analysis of Variance (ANOVA) followed by a Dunnet postest was performed to make multiple comparisons against a control value and a Student’s t-test to compare means of two groups. For the in vivo model ANOVA followed by a Dunnet postest was performed in case of homoscedasticity and KruskalWallis (non parametric ANOVA) followed by a Dunn postest in case of heteroscedasticity. Results are expressed as mean ± SD; asterisks indicate significant at ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.

The study group consisted of 10 patients selected from a larger previous study; Table 1 shows their clinical characteristics and laboratory findings. There were 4 patients with CP (2 SLE and 2 sSS, and 6 patients without CP (1 SLE, 2 pSS and 3 sSS). Sjögren’s Syndrome was found secondary to SLE or RA. All patients had anti-Ro52 and anti-Ro60 appropriate titers; except one (p4) which did not present reactivity for Ro60. The average disease evolution among all 10 patients was of 5 years and their median age was 48.5 years old (between 24 and 83 years old), with 90% being female.

Anti-Ro52 Aabs were purified from 2 ml of serum using an affinity column with Ro52 protein coupled and anti-Ro60 Aabs were purified from another 2 ml of serum using an affinity column with Ro60 coupled. Subsequent

figures show patient num 6 (p6) as an illustrative example.

Figure 1(a) shows the purification profile of p6 serum, where the reactivity by ELISA of each eluted fraction can be seen. The efficiency of the purification procedure was evaluated by titration of the serum by ELISA, before and after the affinity-purification procedure. Figure 1(b) shows the titration sigmoidal curves and titers of p6 serum, which needed to be re-purified in order to obtain the majority of the anti-Ro52 Aabs present in the serum, but purified only once for anti-Ro60 Aabs. Several sera needed a second purification procedure; and high reactivity fractions from both chromatographies of each serum were afterwards pooled. The criteria used to decide if a serum sample did not need to be re-purified was reaching a titer ≤2.3 (1/200 dilution) or a decrease in the titer ≥4 dilutions with respects to the original sample.

For each patient, one pool was made from high reactivity fractions eluted at pH 3.5 and another pool from the ones at pH 2.7. Figure 1(c) shows the corresponding titers of the pooled fractions, none of them presented cross-reactivity against the other protein (e.g. anti-Ro52 pH 3.5 pool did not exhibit reactivity against Ro60— Data not shown). Dialyzed pooled pH 3.5 fractions lost some anti-Ro reactivity with respects to the corresponding serum, but all of them retained enough reactivity in order to be used afterwards. Dialyzed pooled pH 2.7 fractions lost almost all of their anti-Ro reactivity, so these fractions were not used afterwards. This reactivity decrease could be due to damage of the Aabs because of both low pH and manipulation, and also to sample dilution. Titers of all sera and the corresponding purified Aabs (pH 3.5 pool) can be observed in Table 1. It is worth noticing that p6 is a special case for anti-Ro52, it has a significant high titer of those Aabs and it is the only patient whose pooled dialyzed pH 2.7 fractions retained an adequate reactivity, as it can be seen in Figure 1(c). For all the other 9 patients the situation is similar to what happens with dialyzed pH 2.7 fractions of anti-Ro60 from p6.

To confirm that purified Aabs were able to recognize the native antigen present in the cells, samples were tested by indirect immunofluorescence on Hep-2 cells. This is shown in Figure 2 for p6 as representative of the whole cohort, where it can be seen that the three samples were able to bind to the cells, presenting the typical speckled anti-Ro pattern. This was tested for all patients’ sera and their corresponding purified Aabs, confirming their ability to recognize the native antigen. Controls (NHS, hIgG and UBC) are also shown, with no positive images, as expected.

In a previous study on HaCaT cells we have determined that irradiation with 50 mJ/cm² of UVB and evaluation of the cells 24 h post irradiation are the best conditions which produce the highest percentage of apoptotic cells with the lowest amount of living or necrotic ones [37]. Aabs binding to apoptotic keratinocyte’s cell surface was evaluated by flow cytometry for the serum of each patient and the corresponding purified anti-Ro52 and antiRo60 Aabs. Figure 3(a) shows histograms with the

marker used to select the FITC-positive cells. As p6 serum was the one with the highest anti-Ro52 titer, and showed a very high fluorescence when incubated with irradiated cells (Irr), it was used in all experiments as a

positive control in order to set the FITC-positive cells marker and to standardize subsequent data processing. Figure 3(b) shows the histograms corresponding to controls: NHS, hIgG and UBC on irradiated as well as on non-irradiated cells (nIrr). Figure 4(a) shows the binding capacity to apoptotic keratinocyte’s cell surface of all 10 serum samples, expressed as Binding Index (BI) in arbitrary units. This BI was used to standardize variations between experiments and was calculated as follows; considering the mean fluorescence intensity (MFI) and the percentage of cells (%) under the selection marker:

Figures 4(b) and (c) show the BI of anti-Ro52 purified Aabs and anti-Ro60 respectively, for the 10 patients under study. BI for control NHS and hIgG were included in the graphs as a cut off; they were calculated as the mean of the BI of 5 independent experiments. These experiments prove that serum samples of the 10 patients did bind to keratinocytes’ cell surface, showing BI which were statistically significant greater than the BI for control NHS. Their corresponding purified anti-Ro Aabs were also able to bind to the irradiated cells, both antiRo52 and anti-Ro60, most of them having a BI statistically significant different from the BI for control hIgG. When patients were grouped according to the presence or absence of CP (Figure 4, lower row), there were no differences found in the mean BI from the two groups.

Immunized SKH-1 Hairless mice were tested for serum reactivity against recombinant human Ro proteins and also against mice spleen extract in order to test autoreactivity, by ELISA. By the end of the immunization schedule all animals from groups 1, 2 and 3 were able to react against both types of proteins (Figure 5), in a statistically significant different way compared to group 4 (control).

Mice with circulating anti-Ro Aabs were divided in two groups, one was irradiated and the other mockedirradiated (on day 144 of the immunization schedule). They were sacrificed 24 h later and histological analysis and cytokine quantification were performed. Figure 6 shows the skin sections of one representative mouse from each group of animals, where it can be seen that there were no differences between immunized non irradiated groups and the corresponding irradiated ones, except for the known effects of UVB irradiation on the skin [36,39]. Only one group showed a slight difference with respects to the others: when irradiated, group 1 presented a slight inflammatory infiltrate not seen in the other groups. TNF-α production quantified in epidermal extracts can be seen in Figure 7. It is observed that TNF-α is barely not detectable in non irradiated mice, while its levels increased in all irradiated animals as expected, but showing no statistically significant differences between groups.