Enhanced migration and adhesion of peripheral blood neutrophils from SAPHO patients revealed by RNA-Seq

Six SAPHO patients and six healthy volunteers were recruited between March 2017 and June 2018. The historical disease status and clinical manifestation of each patient are shown in Table 1. All six patients suffered from bone lesions, with five on anterior chest wall, six on axial skeleton, two on peripheral joints and one on mandible. Skin involvement was observed in five patients, among which palmoplantar pustulosis is most commonly presented (5/6). Patient #1 was the only one without cutaneous lesions, whose osteoarticular manifestations were similar to the others except for an additional involvement of mandible. As shown in Table 2, there was no significant difference in age or gender between SAPHO group and Control group. Although the levels of peripheral white blood cells and neutrophils are slightly higher in patients with SAPHO syndrome, no statistical significance was found (p = 0.101 and p = 0.117, respectively). The mean visual analogue scale (VAS) of recruited patients is 4.50, indicating a middle-to-severe pain of infected bones. Compared with healthy controls, high-sensitivity C reactive protein (hsCRP) is slightly higher in patients. Besides, three of the patients exhibited faster erythrocyte sedimentation rate (ESR), while two had a lower level of serum osteocalcin, although the average levels of ESR and osteocalcin are within the normal range.

To systematically analyze mRNAs related to SAPHO syndrome, approximately 1.15 billion reads for 12 samples from 6 patients and 6 healthy controls were sequenced using RNA-Seq (Additional file 1: Table S1). Experimental and computational scheme for the RNA-Seq process is shown on Fig. 1a. Firstly, the total RNA from neutrophils was isolated and reverse-transcribed into cDNA library. RNA sequencing was then performed on Illumina HiSeq 2000 Platform. Clean data was mapped to GRCh38 with HISAT2 [17] and then assembled into transcripts. After data homogenization and quality control, the expression level of each transcript was calculated by fragments per kilobase of exon per million fragments mapped (FPKM). The boxplot of FPKM values of all transcripts indicated similar whole transcriptome expression in each sample (Fig. 1b) and principal component analysis (PCA) revealed a relatively clear distinction between samples in the two groups along the first principal component of 28.29% variance (Fig. 1c).

The differentially expressed mRNAs were identified between SAPHO group and Control group using “P < 0.05, fold change > 2” as a cutoff (Fig. 2a). The Hierarchical Clustering analysis was conducted to illustrate the distinguishable expression pattern between the two groups (Fig. 2b). A total of 442 differentially expressed genes were identified, among which 294 were up-regulated and 148 down-regulated (Additional file 2: Table S2). As reported before, tumor necrosis factor-α (TNF-α) signaling plays important role in the pathogenesis of SAPHO syndrom [18], while IL6, IL17 and IL18 were reported to be elevated in the sera of SAPHO patients [19‐21]. We found that TNF-α receptor encoding genes TNFRSF1B, IL6 receptor subunit IL6ST, IL17 receptor gene IL17RA and IL18 receptor genes IL18R1 were up-regulated in peripheral blood neutrophils of SAPHO patients (Fig. 2c). These results indicated that peripheral neutrophils of SAPHO patients were in a more active status than those of the healthy controls.

Five differentially expressed genes of interest were selected for the validation by qRT-PCR (Fig. 3a) in a larger population (12 patients and 12 gender-age matched healthy controls). Among them, S100 Calcium Binding Protein A12 (S100A12) has been detected in salivary proteome profile of SAPHO syndrome [22]. Prostaglandin-endoperoxide synthase 2 (PTGS2 or COX-2) was down-regulated in chronic inflammation [23], whose inhibitors are considered as a potential risk factor for bone healing [24]. In the meanwhile, growth-arrest-specific 7 gene (GAS7) is required for osteoblast differentiation [25], which may play a role in bone manifestations in SAPHO syndrome. Besides, a feedback loop between inflammation and Zn uptake has been observed in rheumatoid arthritis [26], while Zn transporters, also known as SLC30 families, are related to different skin disorders [27] as well as systemic inflammatory response syndrome [28]. Besides, myeloid-associated differentiation marker (MYADM) drew our attention because of its important role in cell spreading and migration [29], which is consistent with our findings in GO term enrichment analysis. The expression levels of S100A12 and SLC30A1 were up regulated while those of PTGS2, GAS7 and MYADM were down-regulated in SAPHO patients, which was in accordance with the RNA-Seq data (Fig. 3b). However, not all the differences in gene expression were statistically significant, and the verification in a larger population is still necessary. To further investigate the biological significance of differentially expressed genes, their correlation with clinical parameters was analyzed. Interestingly, a positive correlation was observed between S100A12 and peripheral hsCRP (Fig. 3c), indicating a role in inflammation. Besides, the down-regulated gene MYADM was positively correlated to osteocalcin (Fig. 3d), which is essential for bone metabolism.

Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis showed that differentially expressed genes were mainly enriched in pathways about autoimmunity, infection, cancer and metabolism (Fig. 4). Among them, the involvement of system lupus erythematosus (SLE) drew our attention. Similar to SAPHO syndrome, SLE is an autoimmune disease with diverse presentations ranging from rash and arthritis to anemia, serositis, nephritis, seizures, etc., while incidence of both diseases is higher in women aged 15–50 [30]. Anti-nucleosome antibodies as well as anti-dsDNA antibodies are present and essential in the diagnosis of patients with SLE [31], while mRNAs encoding histone-derived peptides such as H2A, H2B, H3 and H4 were up-regulated in SAPHO group. The activation of SLE pathway in patients may implicate an immune-mediated etiology in SAPHO syndrome. Besides, extracellular matrix (ECM)-receptor interaction is also among the top 20 significant pathways, which was commonly associated with poor prognosis of cancer [32], and recent study also revealed its association with osteoarthritis [33]. Genes enriched in each pathway are shown in Additional file 3: Table S3.

The top 10 significantly enriched terms of biological process, cellular component and molecular function in Gene Ontology (GO) analysis are shown in Fig. 5. In terms of biological process, differentially expressed genes were predicted to be involved in GO terms such as cell adhesion, platelet degranulation and peripheral nervous system development. The most significantly indicated terms were extracellar matrix in cell component and IgE binding in molecular function. Interestingly, differentially expressed genes were enriched in cell migration (14 genes enriched) and cell adhesion (28 genes enriched), which were both important for neutrophils recruitment in response to inflammation. Genes enriched in each GO term are shown in Additional file 4: Table S4.

To gain insight into how neutrophil adhesion and migration are regulated in patients with SAPHO syndrome, the network of protein-protein interaction (PPI) among genes enriched in cell adhesion and cell migration was constructed based on STRING database (Fig. 6a). The up-regulation of hub genes in the network indicated an over-activated cell adhesion and migration in patients with SAPHO syndrome (Fig. 6b). Among them were components of laminin (LAMA4, LAMB1 and LAMC1) and melanoma cell adhesion molecule (MCAM), which functions as a receptor for LAMA4 [34], promoting cell adhesion and slow-rolling of neutrophils [35]. Up-regulation of tenascin-C (TNC), an endogenous ECM molecule required in neutrophil recruitment [36], was also observed in SAPHO group.

As previously described, the recruitment of neutrophils consists of priming, rolling, arrest, adhesion, crawling, and migration [37]. Specifically, resting neutrophils in peripheral blood are primed by stimuli such as cytokines and interaction with activated endothelial cells, which leads to active transcription and expression of receptors or other molecules [38], among which the P-selectin glycoprotein ligand 1 (PSGL-1) expressed by neutrophils binds in high on-and-off-rate to P-selectins on the endothelial cells and mediates the rolling step [39]. Afterwards, the endothelial cells present chemokines which activate lymphocyte function-associated antigen-1 (LFA-1) and macrophage-1 antigen (Mac-1) on neutrophils, promoting neutrophil arrest [40]. Neutrophils then crawl along the surface of the endothelium to a suitable extravasation site (Fig. 7a). Besides, as we further scanned the differentially expressed genes, several additional genes were found to play important roles in neutrophil recruitment, including the up-regulated gene ITGB2, which is component of LFA-1 and Mac-1. Another increased gene, C-C chemokine receptor-like 2 (CCRL2), has also been proved to be crucial in neutrophil recruitment at sites of inflammation [41]. Other genes such as CD302, a C-type lectin receptor involved in cell adhesion and migration [42] and CD151, known to complex with integrins and other transmembrane 4 superfamily proteins [43], were also up-regulated in SAPHO syndrome (Fig. 7b).

This study was approved by the Institutional Review Board of Peking Union Medical College Hospital (protocol number: ZS-994). We recruited 12 patients with SAPHO syndrome (informing consent obtained) between March 2017 and June 2018. As controls, we included 12 healthy volunteers in the same gender and age range. 20 ml peripheral blood sample was collected from each participant and neutrophils were purified with Polymorphprep™ by gradient centrifugation (neutrophil purities > 90%).

The total RNA was isolated from neutrophils using Trizol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA). Qualified RNA (rRNA removed) was reverse transcribed into cDNA after fragmentation. RNA sequencing was constructed on Illumina HiSeq 2000 Platform. Clean data was mapped to the reference database GRCh38 with HISAT2 [17], while Bowtie2 [50] and eXpress [51] were used to assemble the transcripts. The expression level of RNA was calculated by fragments per kilobase of exon per million fragments mapped (FPKM). Differentially expressed mRNAs between Control group and SAPHO group were determined by DEseq [52] with “p < 0.05, fold change > 2” as the threshold for significance.

Relative expression levels of selected genes were determined by quantitative real-time PCR (qRT-PCR). 200–500 ng of total RNA was reverse transcribed to cDNA using a qPCR RT Master Mix with gDNA Remover Kit (TOYOBO). QRT-PCR was performed on Biorad iQ5 machine using a SYBR Green Realtime PCR Master MixPremix (TOYOBO) according to the manufacturer’s instruction. All experiments were conducted in three technical replicates. The relative expression level of each target gene was normalized to the expression of β-actin with the 2^−ΔΔCt method. Primers are shown in Additional file 5: Table S5.

Differentially expressed genes were selected for functional annotation. Terms of biological process, cellular component and molecular function were analyzed according to Gene Ontology (GO) database (http://​amigo.​geneontology.​org/​amigo/​landing) and functional pathway analysis were based on Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://​www.​kegg.​jp/​). Hypergeometric distribution test was used to determine the significance of differentially expressed mRNA enrichment in each GO term or KEGG pathway. The enrichment score was evaluated as follows:

Here, N is the number of GO-annotated or KEGG-annotated mRNAs in all genes, while n is the number of GO-annotated or KEGG-annotated mRNAs in differentially expressed genes in N. M is the number of mRNAs annotated to a particular GO term or KEGG pathway in all genes while m is the number of genes annotated to a particular GO term or KEGG pathway in differently expressed genes.