Longitudinal assessment of reactivity and affinity profile of anti-Jo1 autoantibodies to distinct HisRS domains and a splice variant in a cohort of patients with myositis and anti-synthetase syndrome

Background To address the reactivity and affinity against histidyl-transfer RNA synthetase (HisRS) autoantigen of anti-Jo1 autoantibodies from serum and bronchoalveolar lavage fluid (BALF) in patients with idiopathic inflammatory myopathies/anti-synthetase syndrome (IIM/ASSD). To investigate the associations between the reactivity profile and clinical data over time. Methods Samples and clinical data were obtained from (i) 25 anti-Jo1+ patients (19 sera with 16 longitudinal samples and 6 BALF/matching sera at diagnosis), (ii) 29 anti-Jo1− patients (25 sera and 4 BALF/matching sera at diagnosis), and (iii) 27 age/gender-matched healthy controls (24 sera and 3 BALF/matching sera). Reactivity towards HisRS full-length (HisRS-FL), three HisRS domains (WHEP, antigen binding domain (ABD), and catalytic domain (CD)), and the HisRS splice variant (SV) was tested. Anti-Jo1 IgG reactivity was evaluated by ELISA and western blot using IgG purified from serum by affinity chromatography. In paired serum-BALF, anti-Jo1 IgG and IgA reactivity was analyzed by ELISA. Autoantibody affinity was measured by surface plasmon resonance using IgG purified from sera. Correlations between autoantibody reactivity and clinical data were evaluated at diagnosis and longitudinally. Results Anti-Jo1 IgG from serum and BALF bound HisRS-FL, WHEP, and SV with high reactivity at the time of diagnosis and recognized both conformation-dependent and conformation-independent HisRS epitopes. Anti-HisRS-FL IgG displayed high affinity early in the disease. At the time of IIM/ASSD diagnosis, the highest autoantibody levels against HisRS-FL were found in patients ever developing interstitial lung disease (ILD) and arthritis, but with less skin involvement. Moreover, the reactivity of anti-WHEP IgG in BALF correlated with poor pulmonary function. Levels of autoantibodies against HisRS-FL, HisRS domains, and HisRS splice variant generally decreased over time. With some exceptions, longitudinal anti-HisRS-FL antibody levels changed in line with ILD activity. Conclusion High levels and high-affinity anti-Jo1 autoantibodies towards HisRS-FL were found early in disease in sera and BALF. In combination with the correlation of anti-HisRS-FL antibody levels with ILD and ILD activity in longitudinal samples as well as of anti-WHEP IgG in BALF with poor pulmonary function, this supports the previously raised hypothesis that the lung might have a role in the immune reaction in anti-Jo1-positive patients. Supplementary Information The online version contains supplementary material available at 10.1186/s13075-022-02745-6.

Register for IIM (Swemyonet) (2), and the Euromyositis register (3). The total improvement score was calculated according to the IMACS response criteria by comparing the IMACs core set measures at baseline and at the last available longitudinal sample. The total improvement score was used to define the response as minimal, moderate or major according to American College of Rheumatology (ACR response) (4). C-reactive protein (CRP) levels were collected within 1 month of diagnosis and at any time of longitudinal sampling as a proxy for disease activity.

ELISA assay for full-length, domains and variant of HisRS
Anti-Jo1 ELISA protocol was followed as described in (5) with slight modifications. Briefly, 384 well plates were coated with 25 ng of streptavidin (Sigma) overnight at 4°C, washed with PBS 0.05% Tween (PBST), and blocked with PBST 0.1% BSA for 1hour at room temperature (RT). Biotinylated recombinant HisRS antigens (HisRS-FL, WHEP domain, CD, ABD, and SV) diluted in PBST 0.1% BSA were added at 10 µg/mL and let to incubate in the plates for 1hour at RT. Patient samples tested in the ELISA (duplicates) were the following: i) IgG purified from serum at 10 µg/mL, for detection of anti-Jo1 IgG, ii) undiluted BALF, and iii) serum collected at the same time as the BALF and diluted 1:500. Total IgG, total IgA, anti-Jo1 IgG, and anti-Jo1 IgA were measured in BALF and matching serum samples. All samples were incubated in the plates for 2h at RT. Plates were then washed and incubated for 1h at RT with either 1:1000 diluted alkaline-phosphatase (ALP) monoclonal antibody MT78 (Mabtech) for IgG detection, or ALP monoclonal antibody MT20 (Mabtech) for IgA detection in BALF and serum. For anti-Jo1 reactivity analysis of total IgG, the peroxidase F(ab'2) fragment goat anti-Human IgG (Jackson Laboratories) diluted 1:10 000 in blocking buffer was employed (1h at RT). After the last washing step, pNPP substrate for ALP antibody detection or TMB for peroxidase antibody detection (Mabtech and Sigma) were added. The reaction with TMB was stopped after 15 minutes by adding 1M H2SO4. When using pNPP substrate, the OD was measured every 15 minutes at 405 nm and when using TMB substrate, the OD was measured one time at 450 nm (SoftMax Pro version 4.8, Molecular Devices).

ELISA control experiments
To evaluate whether we could detect the reactivity of anti-Jo1 IgG against HisRS variants we initially screened diluted serum (1:1000) and total IgG (10 µg/mL) enriched from the matching serum. Two representative patient samples (Patient 8 and 17) are displayed in Supplementary   Figure 3. Although reactivity against HisRS-FL could be detected in serum, anti-WHEP and -ABD response could only be detected in the total IgG fraction. Therefore, the serum dilution utilized to test the reactivity against HisRS-FL, splice variants and domains in patients whom provided corresponding BALF samples was decreased to 1:500. Anti-Jo1 IgG purified from total IgG isolated from a sera pool of 38 IIM/ASSD patients was employed as a standard curve in order to calculate antibody levels and correlate with clinical data (5) (Supplementary Figure   3B). To be able to measure both low and high anti-Jo1 reactivities within the same dilution (10 µg/mL), the standard curve is only within the linear range between ≈5 and ≈100 ng/mL. Anti-Jo1 reactivity towards biotinylated and non-biotinylated HisRS-FL was compared to determine whether biotin interfered with the assay (Supplementary Figure 3C). Biotinylated HisRS-FL and non-biotinylated HisRS-FL were added to streptavidin pre-coated or high binding multiwell plates, respectively at a concentration of 1 µg/mL. The remaining protocol follows that described above. for 60 minutes in Glycine buffer) and blocked for 60 minutes with 5% milk in phosphatebuffered saline 0.05% Tween (blocking buffer). The different membranes, each containing a different HisRS protein variant, were individually incubated with anti-Jo1 IgG from 19 anti-Jo1 + patients, 2 anti-Jo1patients and 3 healthy controls (at 1 µg/mL diluted in blocking buffer).
Secondary antibody rabbit anti-human IgG HRP (diluted 1:1000 in blocking buffer, sc-2769 Santa-Cruz Biotechnology) was employed. Images were acquired and bands were quantified on a ChemiDoc XRS + System using Image Labä Software (Bio-Rad). The correlation between ELISA and WB data reported in Supplementary Figure 5C was calculated by plotting OD 450 nm and mean intensity values, respectively (GraphPad Prism version 8).

Surface plasmon resonance
Affinity measurements were performed on serum-derived IgG from the 19 anti-Jo1 + patients using Biacore T200 biosensor instrument (Cytiva) and singe cycle kinetics. An anti-human Fab antibody (Cytiva) was covalently coupled to a Series S CM5 amine sensor chip (Cytiva) according to the manufacturer's instructions and used as capture molecule. Protein G purified IgG, diluted to a concentration of 5-10 µL/mL, were caught on the surface, followed by the addition of 0.08, 0.4, 2, 10, and 50 nM recombinant HisRS-FL at a flow rate of 30 µL/min. The anti-Fab antibody surface was regenerated with 10 mM Glycine-HCl pH 2.1. The experiment was performed at 25 °C using HBS-EP+ (Cytiva) with 1% BSA as running buffer. Response curves were obtained by subtracting a reference surface (no antibody captured) and a reference cycle (running buffer instead of HisRS-FL). The Biacore T200 evaluation 3.1 software (Cytiva) was used for analyses of the response curves and reaction rate kinetics calculated using the predefined 1:1 Langmuir binding model. Four anti-Jo1samples was selected as negative controls and run against HisRS in the same set-up as described above. In addition, the anti-Jo1 + IgG was also tested against another aminoacyl tRNA synthetase (ThrRS) as control protein.
Two recombinant IgG, J-HARS-58 and L-TARS-11 (6), binding HisRS and ThrRS respectively, were used as positive controls. Representative sensorgrams of controls are presented in Supplementary Figure 10. ThrRS was produced as previously described (6). Of note, considering that the average affinity is the component being measured, there are some 6 facts to take into consideration; firstly, the analyzed total IgG samples are polyclonal, thereby likely containing a mixture of specific anti-Jo1 autoantibodies of different affinities, most likely present in very different concentrations. Here, an average affinity is measured, and the binding profile of an individual antibody clone within a polyclonal mix may be very different to what we report. For example, it is possible that high affinity antibodies in the sample mask low affinity antibodies. Secondly, the antigen HisRS-FL is a homodimer and this may give rise to an avidity effect, i.e. an apparent increase in affinity. Despite these complicating factors, the analysis was made using the predefined 1:1 Langmuir binding model to get comparable values among the patient samples.

Total IgG purified from serum of anti-Jo1 + patients collected longitudinally
Of the 19 anti-Jo1 + patients included in this study, we could withdraw longitudinal sera from 16. However, not all of those 16 anti-Jo1 + patients had longitudinal samples available at all and the same time-points. Therefore, we pooled the 7 anti-Jo1 + individuals who provided sera both closer to diagnosis (from -0.25 years up to date of diagnosis) and in the consecutive years (up

Anti-HisRS reactivity multivariate data analysis
The anti-HisRS reactivity data (HisRS-FL, WHEP, SV, CD, and ABD, described in Results section 1.1.), was subjected to PCA analysis (resulting in two components, R 2 =0.89, Q 2 =0.85).  Figure 8C, it is evident that HisRS-FL and WHEP are strongly associated with ILD status with a significant difference of p=2.9E-17 comparing ILD + and ILDanti-Jo1 + patients. Interestingly, a similar trend can be observed for ILD + and ILDanti-Jo1patients (p=1.5E-1). Compared to HisRS-FL and WHEP, the impact of ABD and CD on ILD status is less prominent for the anti-Jo1 + patients (p=2.2E-4) and the trend is obviated (p=3.5E-1) for the anti-Jo1patients, Supplementary Figure 8B.

Anti-HisRS reactivity and clinical information data analysis
Principal component analysis (PCA) analyses were performed on the first available sample obtained from the patients whose demographics are listed in Table 1 (cohort 1: 25 anti-Jo1and 19 anti-Jo1 + patients) as well as on the data obtained from the patients whose demographics are listed in Supplementary Table 1 (cohort 2: 4 anti-Jo1and 6 anti-Jo1 + ). Similar to the PCA scores plot of the anti-HisRS reactivity data (Supplementary Figure 8A), both models indicated a clustering of the anti-Jo1patients and/or ILDpatients ( Figure 5A and 5C). In order to identify which factors correlated with anti-Jo1 and ILD status, two OPLS-DA models (per data set) were generated and plotted against each other ( Figure 5B and 5D). As expected, the correlation between anti-Jo1to anti-Jo1 + and ILDto ILD + in both models is strongly positive (cohort 1: R 2 =0.92 and cohort 2: R 2 =0.76). In addition to the anti-HisRS reactivity data which strongly correlated with both anti-Jo1 + and ILD + , also ASSD, MSA, SSA and arthritis correlated with anti-Jo1 + and ILD + in cohort 1 ( Figure 5A and 5B).

SUPPLEMENTARY TABLES
Supplementary Table 1 Clinical and laboratory features, pulmonary function, and BALF characteristics from patients (cohort 2) who underwent BAL * at time of serum sampling.   Each colored line represents one patient (P).