Molecular and spatial analysis unveils the functional basis of tertiary lymphoid structures in Sjogren’s syndrome[bulk RNA-seq]

The key role of tertiary lymphoid structures in autoimmune and non-autoimmune conditions has been recently appreciated. While many of the molecular mechanisms involved in tertiary lymphoid structure formation have been identified, their cellular sources and temporal and spatial relationship remain unknown. Using single-cell RNA-sequencing, spatial transcriptomics and proteomics of minor salivary glands of patients with Sjogren’s disease and Sicca Syndrome, ex-vivo and in vivo functional studies, we construct a cellular and spatial map of key components involved in the formation and function of tertiary lymphoid structures. We confirm the presence of a fibroblast cell state and identify an undescribed pericyte/mural cell state with potential immunological functions. The identification of novel cellular properties associated with these structures and the molecular and functional interactions identified by this analysis provide key therapeutic cues for tertiary lymphoid structures associated conditions in autoimmunity and cancer. The selected tissue samples for microdissection processing were acquired from patients recruited within the OASIS cohort at the University of Birmingham, under ethics number 10-018. Preparation of the samples for microdissection: all surfaces of the cryostat and lab bench were thoroughly cleaned with 70% Etoh and sterilised with ultraviolet (UV) light for at least 30 min. Containers used to keep store the collected samples and the filter paper used to separate slides (Membrane slides, Carl Zeiss™ 1.0 PEN NF 41590-9081-000) must also be treated with UV for 30 min and kept sterile before handling the tissue. To cut tissue for microdissection, the UV sterilised membrane slides were maintained at a cool temperature within the cryostat prior to sample cutting. The tissue was mounted on a cryostat block with as little mounting media (OCT, TissueTeK) as possible to minimise interference with laser cutter during microdissection. Sections were cut at 8 um thickness and transferred to the membrane slides. Once the section had been mounted onto the membrane slides, the membrane slide was stored in its carrier box and in dry ice until final storage at -80oC prior to staining with Cresyl violet acetate (SIGMA). Preparation of Cresyl violet acetate is required one day prior to staining the mounted tissues. Preparation of the CresylViolet (1% solution) included dissolving the powder in 80% ethanol which had been made using sterile RNase-free water in a MSC Class II ventilator system. This solution was mixed overnight on a shaker at 4oC to ensure that the CresylViolet was completely dissolved, and the solution was then passed through a 70µm strainer before use. The slides were first dipped in RNase-free water and suspended for 2 min in 70% Etoh, also prepared in RNase-free water, to facilitate the removal of OCT on the tissue. The slides were then suspended in the CresylViolet 1% solution for 30 secs to 1 min and excess solution was tapped off. The slides were then processed for 1 min in 100% Etoh and this step was repeated three times. The PALM slides were then left to dry until the membrane was observed to be completely dry. The PALM slides were then stored at -800C in a parafilm sealed 50 ml falcon until the samples were ready for microdissection. Prior to microdissection, the PALM slides were bought to room temperature for 10 min. Microdissection was completed using the PALM Robo Software V.4.6 software using the brightfield “AxioCam CC1” setting. The objective lens used to micro dissects SG sections was 10X whilst tonsil was set at 5X. The area of interest was then identified using the ocular eye piece. Once the region had been selected the laser was aligned to the section that was to be cut. The collection tubes required for the capture of the microdissected tissue contained 20 ul of RLT buffer with B-mercaptanol, ensuring the lid was kept dry, and tube was kept on dry ice prior to use. The collection tubes were then mounted onto the Collector Set 200 CMII and set to the correct position to collect the microdissected sample as they were cut by applying the calibration procedure to collect tissue in the centre of the lid of the collection tube which were stored at -80oC prior to RNA extraction. RNA extraction from the microdissected samples was completed as per the protocols provided using the RNeasy FFPE kit (Qiagen). Library preparation and sequencing from extracted RNA was completed by Genomics Birmingham. Initial libraries were prepared using the SMART-Seq v4 Ultra Low Input RNA Kit. Further samples were processed using the NEBNext rRNA Depletion Kit v2 and NEBNext Ultra II Directional RNA Library Prep Kit for Illumina with NEBNext Multiplex Oligos for Illumina. All libraries were sequenced using a NextSeq 500 sequencer. Adapters were trimmed with trimmomatic v0.39, reads aligned to the hg38 transcriptome with STAR v2.7.2b, and counts summed using featureCounts in Subread v2.0.1 using the University of Birmingham BlueBEAR High Performance Computing service. R v4.0.3 and DESeq2 v1.30.1 were used for downstream analysis. To minimise batch effects during differential expression analysis batch was included as a factor in the DESeq2 model and for principal component analysis and visualisations batch effects were minimised in variance stabilising transformed data using the limma v3.46.0 function removeBatchEffect.