Evidence of a Sjögren’s disease–like phenotype following COVID-19 in mice and humans

Sjögren’s Disease (SjD) is a systemic autoimmune disease characterized by lymphocytic inflammation of the lacrimal and salivary glands (SG), dry eyes and mouth, and systemic symptoms. SARS-CoV-2 may trigger the development or progression of autoimmune diseases. To test this, we used a mouse model of SARS-CoV-2 infection and convalescent patients’ blood and SG in order to understand the development of SjD-like autoimmunity after infection. First, SARS-CoV-2–infected human angiotensin-converting enzyme 2 (ACE2) transgenic mice exhibited decreased salivation, elevated antinuclear antibodies (ANA), and lymphocytic infiltration in the lacrimal and SG. The sera from patients with COVID-19 sera showed increased ANA (i.e., anti-SSA [Sjögren’s-syndrome-related antigen A]/anti-Ro52 and anti-SSB [SS-antigen B]/anti-La). Male patients showed elevated anti-SSA compared with female patients, and female patients exhibited diverse ANA patterns. SG biopsies from convalescent COVID-19 patients were microscopically similar to SjD SG with focal lymphocytic infiltrates in 4 of 6 patients and 2 of 6 patients exhibiting focus scores of at least 2. Lastly, monoclonal antibodies produced in recovered patients blocked ACE2/spike interaction and cross-reacted with nuclear antigens. Our study shows a direct association between SARS-CoV-2 and SjD. Hallmark features of SjD-affected SGs were histologically indistinguishable from convalescent COVID-19 patients. The results implicate that SARS-CoV-2 could be an environmental trigger for SjD.


Animal models and LD50 determination:
The homozygous K18-hACE2 mice were bred at the University of Florida (Breeding Project ID #3276).Mating pairs of hemizygous K18-hACE2 mice were purchased from Jackson Laboratory (034860-B6.Cg-Tg(K18-ACE2)2Prlmn/J). F1 mice were anesthetized for ear punching for DNA extraction and subjected to zygosity analysis.The F1 zygosity (homozygosity) was confirmed via a SNP assay (The Jackson Laboratory's Protocol 38275: Sanger sequencing assay -Chr2_rs13476660-SEQ) by Transnetyx Inc.This assay detects a SNP between C57BL6/J and SJL/J (homozygous mutant), located approximately 4.7 kb from the transgene integration site.Once the homozygosity of the F1 mice was confirmed, mice between 8-10 weeks old were moved into micro-isolator cages, 5 mice per cage, under pathogen-free conditions for a challenge in the ABSL-3 laboratory.After being acclimated in the ABSL3 laboratory for three days, the mice were anesthetized with 87.5 mg/kg of ketamine and 12.5 mg/kg of xylazine of body weight by intraperitoneal (ip) injection.Once fully anesthetized, each group of mice was intranasally inoculated with 20 µL of SARS-CoV-2 inoculum drop-by-drop into both nostrils until fully inhaled.The mice received 860 PFU of SARS-CoV-2 WA1/2020 (BEI Resources, NIAID, NIH: SARS-Related Coronavirus 2, Isolate USA-WA1/2020, Recombinant Infectious Clone (icSARS-CoV-2-WT), NR-52281) (n=26).The naïve control mice (n=10) were inoculated with 20 µL of Dulbecco's Modified Eagle Medium (DMEM) supplemented with 2% fetal bovine serum (FBS).The mice were observed twice daily until the end of the study, 21 days.The mice were euthanized when moribund.The LD50 value was calculated using Probit analysis.

Making viral inoculum
One 150 cm 2  off and a freshly prepared 25 mL of MEM Eagle supplemented with 2% heat-inactivated FBS was added.The flask was incubated at the same growth condition as above.The monolayer was checked daily for signs of cytopathic effect.After 4 days of the infection, the viruses were harvested from the culture medium.The culture medium containing cells and the viruses was collected in a 50 mL conical tube.The tube was centrifuged at 4000 x g for 5 minutes to pellet the cells.The supernatant was collected and aliquoted into 2 mL vials before being stored in -80 o C freezer until used.A solid double overlay plaque assay was used to quantify the viral particles as previously described (1).

Confirmation of infection by real-time PCR
15 mg of the lung tissues from SARS-CoV-2 infected K18-hACE2 mice were grinded using the Closed Tissue Grinder System (Fisher Scientific, Huston, TX).The single-cell suspension was obtained through a 70um cell strainer (Fisher Scientific, Huston, TX).The RNA was extracted through QIAamp® Viral RNA Mini Kit (QIAGEN, Germantown, MD), and 700 ng of extracted RNA was used for the Luna® SARS-CoV-2 RT-qPCR Multiplex Assay Kit (NEB, Ipswich, MA) to detect for nucleocapsid (N)1 and N2 genes of SARS-CoV-2 simultaneously.The samples were performed in duplicate.A Ct value of less than 40 was considered positive according to the manufacturer's guidelines.The copy number was calculated based on a manufacturer-provided positive control

Histological examination of salivary and lacrimal glands
Whole mouse heads were fixed in 10% phosphate-buffered formalin for at least 48 hours, then salivary and lacrimal glands were excised, embedded in paraffin, sectioned at a thickness of 5 μm, and stained for B Cells (BD, Franklin Lakes, NJ, 550286), T Cells (CD3: Santa Cruz Biotechnology, sc-1127; CD8: Santa Cruz Biotechnology, Dallas, TX, sc-20041), and macrophages (Genetex, Irvine, CA, GTX73723) as described previously (2).Stained samples were visualized using 400x magnification of Nikon Ti-E fluorescent microscope with an exposure of 200 milliseconds.
Human MSG were fixed immediately after biopsy in 10% neutral-buffered formalin for 24 hours, embedded in paraffin, sectioned at a thickness of 5 μm, and stained with hematoxylin and eosin.Histopathological assessment was performed and reviewed by a board-certified surgical pathologist (DK) and an oral and maxillofacial pathologist (BMW).When available, immunohistochemistry was performed to characterize immune infiltrates in the MSG using CD3

Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) and Caspase-3 immunofluorescence stains
Paraffin-embedded glands were deparaffinized via pressure cooking in Trilogy (Cell Marque, Rocklin, CA) for 10 minutes.Following a PBS-T wash, the sections were treated with proteinase K for 30 minutes.After a PBS-T wash, the positive control was incubated with 0.5 mg DNase I (Sigma Aldrich, Saint Louis, MO) for 10 minutes.Click-iT TUNEL Alexa Fluor Imaging Assays was performed as instructed by the manufacturer (ThermoFisher, Waltham, MA).Briefly, slides were blocked with a Tris-buffered solution, then incubated with terminal deoxynucleotidyl transferase (TdT) and Biotin-11-dUTP for 1 hour at 37 °C, before being stopped with sodium citrate butter.Slides were incubated with AF647 streptavidin and mounted with Vectashield DAPImounting medium.For caspase-3 detection, slides were blocked for an hour with donkey serum, followed by overnight incubation with 1:1,000 anti-caspase3 (Novus Biologicals, Littleton, CO, 9661T).Secondary antibody treatment consisted of 1:100 AF488 (ThermoFisher, Waltham, MA, A21206).Stained sections were mounted using a Vectashield DAPI-mounting medium.Stained

Detection of antinuclear antibodies
Antinuclear antibody detection in the sera of mice and humans and mAbs was performed per the manufacturer's instructions (Immuno Concepts, Sacramento, CA).Briefly, sera were diluted 1:40-1:320 in PBS and incubated on HEP-2 ANA slides for 30 minutes.For mAbs, undiluted samples were added where concentrations were all below 1.5 mg/ml.Goat anti-mouse IgG AF488 (Invitrogen, Waltham, MA, A11001) or manufacturer-provided anti-human IgG AF488 were added as secondary antibodies.Slides were sealed with Vectashield DAPI medium (Vector Laboratories, Burlingame, CA) and a glass coverslip was mounted.ANA staining pattern was observed at 400x with a Nikon Ti-E fluorescent microscope with an exposure of 200 ms (Nikon, Tokyo, Japan).

Measurement of saliva flow
To measure stimulated flow rates of saliva, individual mice were weighed and given an intraperitoneal (ip) injection of 100 μl of a mixture containing isoproterenol (0.2 mg/1 ml of PBS) (Sigma-Aldrich, St. Louis, MO) and pilocarpine (0.05 mg/1 ml of PBS) (Sigma-Aldrich, St. Louis, MO).Saliva was collected for 10 min from the oral cavity of individual mice using a micropipette starting one minute after injection of the secretagogue.The volume of each saliva sample was measured and calculated in relation to the mouse's weight to produce saliva flow rate (SFR).

Immunohistochemical evaluation of SARS-CoV-2
Slides were deparaffinized and rehydrated, then antigens were retrieved using Antigen Unmasking Solution (Vector Laboratories, Burlingame, CA) per the manufacturer's instructions.
Cloned plasmids were sequenced through GENEWIZ (Azenta, Chelmsford, MA) and aligned by NCBI IgBLAST to ensure sequences of chains were intact.To produce monoclonal antibodies (mAbs), FreeStyle™ CHO-S cells (ThermoFisher Scientific, Waltham, MA, R80007) were transfected with corresponding H/L expression vector DNA through FectoCHO® Expression System (PolyPlus, Vectura, France) as instructed by the manufacturer.On day 7 posttransfection, the cell culture was centrifuged and purified through EconoFit UNOsphere SUPrA on NGC™ Chromatography Systems according to the manufacturer (Bio-Rad, Hercules, CA).

Protein concentration was quantified by Bradford protein assay (Quick Start™ Bradford 1x Dye
Reagent, Bio-Rad, Hercules, CA) and the integrity of the Ig fractions with joined H/L chains was determined by Mini-PROTEAN TGX Stain-Free Precast Gels (Bio-Rad, Hercules, CA) by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS/PAGE).antibody and virus mixture was inoculated onto confluent Vero E6 monolayers seeded in 6-well plates.Viral inocula were allowed adsorb for 1 h at 37°C, 5% CO2 before a primary overlay containing MEM supplemented with a final concentration of 0.5% agar (w/v) and 5% FBS (v/v) was added to each well.A secondary overlay comprised of MEM with a final concentration of 5% agar, 1% FBS, and 0.007% neutral red solution was overlaid two days later.Plaques were counted with the naked eye the following day.

Detection of SjD-associated autoantibodies in sera
Detection of human and mouse anti-SSA/Ro52, anti-SSA/Ro60, and anti-SSB/La was performed as previously described( 7) Absorbance values of 450/630 nm were determined with a Tecan Infinite M200 Pro-spectrophotometric plate reader (TECAN, Mannedorf, Switzerland).

Detection of autoantibodies using INNO-LIA ANA Update Test strips
To analyze whether broad autoimmunity was preset in sera samples, INNO-LIA ANA Update Test strips (FujireBio Diagnostics, Tokyo, Japan) were utilized on human control (n=5 males, n=5 females) and COVID-19+ patient sera (n=10 males, n=10 females).In brief, sera were diluted 1:320 in sample diluent and 2 mL was loaded into each trough, with 2 mL Cut-off Control for the positive control strip.After incubating on rocker for one hour, LIA strips were washed and Conjugate Working Solution was added.After 30 minutes of rocking, LIA strips were washed and Substrate Solution was added.After 30 minutes of rocking, solution was aspirated and Stop Solution was added to LIA strips.After 15 minutes, LIA strips were dried and image was captured with a Gel Doc XR+ (BioRad, Hercules, CA).In the ImageLab 5.2.1 Software (BioRad, Hercules, CA), "lanes" (i.e.LIA strips) and "bands" were detected; every individual antigen band on every LIA strip was normalized against the Control LIA strip and correction was made for the altered sera concentration.Sera was considered positive if over the threshold (1± 2 standard deviations) utilizing the intensity of the internal controls provided on every LIA strip.

Table S1: Focal scores in salivary and lacrimal glands
Table S1: Incidence of exocrine-gland SjD symptoms in SARS-CoV-2 mice.The number of mice positive for a focal score (FS) in the salivary or lacrimal glands is provided, where an aggregate of ≥ 50 lymphocytes qualified as a single focus.Additionally, lymphocytic infiltration of B and/or T cells which did not qualify as a focus is reported for both the salivary and lacrimal glands.A McNemar's test was performed to compare the divergence of symptomology between the control and infected groups, reported as the χ 2 , with p-values provided.against the noted antigen, where significant binding occurs at a signal of 0.78 (1±2 standard deviations) or more.
flasks (T-150) containing 75-80% confluent monolayer of Vero E6 (BEI Resources, NIAID, NIH: African Green Monkey Kidney Epithelial Cells (Vero E6) Expressing High Endogenous Angiotensin-Converting Enzyme 2, NR-53726) cells grown in Minimum Essential Medium (MEM) Eagle supplemented with 10% heat-inactivated FBS was brought from the BSL-2 laboratory into a BSL3 laboratory two days before the experiment and incubated at 37 o C with 5% CO2.A frozen stock of 2 mL of SARS-CoV-2 WA1/2020 culture (BEI Resources, NIAID, NIH: SARS-Related Coronavirus 2, Isolate USA-WA1/2020, Recombinant Infectious Clone (icSARS-CoV-2-WT), NR-52281, Centers for Disease Control and Prevention)) was thawed at 37 o C.After aspirating the culture medium from the T-150 flask, the thawed viral culture was inoculated onto the monolayer of the Vero E6 cells and was incubated for 1 hour.The viral inoculum was aspirated

(
clone 2GV6, 790-4341, Roche.Indianapolis, IN), CD20 (clone L26, 760-2531, Roche.Indianapolis, IN), CD4 (clone SP35, 790-4423, Roche.Indianapolis, IN), and CD8 (clone SP57, 790-4460, Roche.Indianapolis, IN).The intensity and pattern of the infiltrates were interpreted.All immunohistochemical studies used on anatomic pathology specimens were performed in the NCI Center for Cancer Research, Laboratory of Pathology, a laboratory approved by the College of American Pathologists Laboratory Accreditation Program.The included antibodies were clinically validated to the satisfaction of the College of American Pathologists' checklists and inspectors.Positive and negative tissue controls for each antibody are regularly checked and maintained in the Laboratory of Pathology.
samples were visualized at 10X magnification on Nikon Ti-E fluorescent microscope with an exposure of 200 milliseconds.Nikon NIS-Elements software was used to detect the threshold intensities utilizing the ROI function; threshold intensities were kept consistent throughout the experiment.

Figure S1 :
Figure S1: Frequencies of B and T lymphocytes in salivary and lacrimal glands of the infected mice.A) Identification of infiltrating cells in the salivary glands and lacrimal glands, where representative immunofluorescent staining of CD3 + T cells and B220 + B cells are displayed with blue DAPI nuclei staining at 40X magnification.B) Enumeration of B and T cells using.One-

Figure S3 :
Figure S3: Monoclonal antibodies from recovered COVID-19 patients showed reactivity against self-antigens.A) Competitive ELISA demonstrated the comparable inhibition ability of

Figure S5 .
Figure S5.hACE2 expression levels in the salivary and lacrimal glands.Expression of human ACE2 (hACE2) mRNA in salivary glands (SG, black) and lacrimal glands (LG, grey) of

Figure S7 :
Figure S7: Determinination of viral loads in the lungs post infection.Mice received SARS-CoV-2 inoculum were randomly selected (n=5 with 3 females and 2 males).Viral RNAs were isolated from the lung tissues.Realtime PCR was performed to determine the copy number of