Serum exosomal tsRNA biomarkers: A novel strategy for identifying lupus nephritis

Abstract

Dear Editor, For effective treatment and control of systemic lupus erythematosus (SLE), it is pertinent to be able to precisely identify and predict lupus nephritis (LN).1 The current gold standard for this is renal biopsy, but this is invasive, making it less attractive for dynamic monitoring of disease progression.2, 3 tRNA-derived small non-coding RNAs (tsRNAs), as a novel RNA biomarker, intricately involved in regulating various stages of gene expression, from transcription, translation, RNA processing, to maturation, are implicated in key cellular processes, including self-renewal, differentiation, proliferation and onset of pathological conditions such as immune system disorders and cancer.4-7 Exosomes, which typically range from 30 to 100 nm, derived from cells, serve as pivotal mediators for intercellular communication and also act as crucial carriers for circulating tsRNA.8 Here, we anticipated and demonstrated that serum exosome-encapsulated tsRNAs can serve as biomarkers for executing LN diagnosis noninvasively. The study recruited participants (HC, n = 80; SLE (LN‒), n = 122; LN, n = 131) from Nanjing Drum Tower Hospital between September 2020 and June 2022, in accordance with the 1997 SLE classification criteria9 set by the Rheumatology Society (Table 1). Ethical approval was obtained from the Ethics Committee of Nanjing Drum Tower Hospital, with the approval identification number 2020-327-01. An exosome isolation kit (from serum) (Thermo Fisher Scientific Inc.) was initially used for exosome separation from 100 µL of serum. The exosomes were then characterised by transmission electron microscopy, nanoparticle tracking analysis and Western blotting (Figure S1). RNA from the serum exosomes was extracted using Trizol and treated with an rtStart tRF and tiRNA pre-treatment kit (Arraystar Inc.) prior to sequencing. The volcano plot showed that 88 tsRNAs were increased and 66 tsRNAs were decreased based on the criteria of fold change greater than 2 and a p-value less than 0.01 (Figure 1A). The species and occupancies of the nine types of tsRNAs had no obvious differences between SLE (LN‒) and LN groups (Figure 1B), while the fragment lengths of tsRNAs were mainly enriched in the 20−22 nt and 31−32 nt intervals (Figure 1D). The Venn distribution indicated that 195 tsRNAs existed simultaneously in both groups, 84 tsRNAs were unique to the LN group and 64 tsRNAs were unique to the SLE (LN‒) group (Figure 1C). Ultimately, we plotted the column chart of the top 10 high/low-expression tsRNAs in the two groups (Figure 1E) and preliminarily verified the actual expression of the 10 high-expression tsRNAs in the LN group relative to SLE (LN‒) using RT-qPCR (Figure 1F). Furthermore, tRF-iMet-CAT-1, tRF-Ala-AGC-2-M4 and tRF-Tyr-GTA-1-M2 showed eminently high expression in the SLE (LN‒) group than in the HC group (Figure 2A,C,D). In contrast, tRF-Thr-TGT-4-M3 and tRF-Tyr-GTA-1-M2 were significantly higher expressed in the LN group than in the SLE (LN‒) group (Figure 2B,D). Even so, tRF-Gly-CCC-1-M4 showed no eminent expression differences between the two groups (Figure 2E). To evaluate the diagnostic value of the screened tsRNAs, we generated subject working receiver operating characteristic (ROC) curve and developed a random forest diagnostic model. Among the tested tsRNAs, tRF-Tyr-GTA-1-M2 exhibited the highest diagnostic efficacy in distinguishing HC from SLE (LN‒) with an area under the ROC curve (AUC) of 0.7663 (Figure 2F). tRF-iMet-CAT-1, tRF-Thr-TGT-4-M3, tRF-Ala-AGC-2-M4 and tRF-Tyr-GTA-1-M2 showed high diagnostic values for the diagnosis of LN from HC, with AUCs over 0.86 (Figure 2G). tRF-Thr-TGT-4-M3 and tRF-Tyr-GTA-1-M2 had diagnostic values for the diagnosis of LN from SLE (LN‒), with AUCs of 0.6589 and 0.6499, respectively (Figure 2H). Subsequently, the correlation heatmap of five tsRNAs and 18 clinical indicators showed that tRF-Tyr-GTA-1-M2 had significant positive correlations with the SLE disease activity index score, 24-h urine protein, C-reactive protein and erythrocyte sedimentation rate (Figure 2I). Furthermore, according to the random forest algorithm, estimated glomerular filtration rate (eGFR) exhibited the highest mean decrease Gini in all clinical index (Figure 2J), with joint AUC 0.88 (Figure 2K); tRF-Tyr-GTA-1-M2 and tRF-Thr-TGT-4-M3 exhibited the highest mean decrease Gini in five tsRNAs (Figure 2L), with joint AUC 0.879 (Figure 2M). Finally, a machine learning modeling analysis was conducted by integrating four tsRNAs with six clinical indexes. It was observed that eGFR, tRF-Thr-GTA-4-M3 and tRF-Tyr-GTA-1-M2 exhibited the top mean decrease Gini (Figure 2N). Combined the top 10 diagnostic markers together, the AUC for the diagnosis of SLE (LN‒) and LN reached 0.919 (Figure 2O). Eventually, we predicted the structure and function of the screened tsRNAs using bioinformatics tools. tRF-Thr-TGT-4-M3 and tRF-Tyr-GTA-1-M2 were generated from tRNA-Thr-TGT-4-1 and tRNA-Tyr-GTA-1-1, respectively, both of which belong to the tRF-3b type of tsRNA molecules (Figure 3A,D). Biological process, cellular component and molecular function showed that tRF-Thr-TGT-4-M3 was most significantly enriched in striated muscle cell development (GO: 0055002), striated muscle thin filament (GO: 0005865) and Tat protein binding (GO: 0030957) (Figure 3B). tRF-Tyr-GTA-1-M2 demonstrated enrichment in regulation of epithelial cell proliferation associated with prostate gland development (GO: 0060768), presynaptic active zone cytoplasmic component (GO: 0098831) and RNA polymerase II activating transcription factor binding (GO: 0001102) (Figure 3E). Interestingly, tRF-Thr-TGT-4-M3 was enriched in the vitamin D receptor pathway (WP2877), which is closely associated with the occurrence and development of SLE10 (Figure 3C). tRF-Tyr-GTA-1-M2 was enriched in bacterial invasion of epithelial cells (hsa05100) pathway, which may be a new explanation for bacterial infections in patients with LN (Figure 3F). Overall, we have identified several serum exosomal tsRNA biomarkers for diagnosing patients with SLE. Additionally, we have developed a diagnostic model that can distinguish the presence or absence of nephritis in SLE patients using combined tsRNAs. However, it is crucial to observe that this study is limited to a single-centre sample, and subsequent validation is still required through multi-centre, large-scale investigations across different regions. Methodology and data validation: Ping Yang, Chenlan Wang and Yifan Sun. Formal analysis: Ping Yang, Zhiyang Li. Guidance to the study: Fangfang Jin, Zhiyang Li, Jin Wang and Yanbo Wang. Methodology and writing—original draft: Adeel Khan. Writing—original draft: Ping Yang and Zhiyang Li. Collection of samples and clinical information: Jianming Gong, Yiyuan Han. Chief designer of the whole experiment: Ping Yang. All authors have reviewed and approved the final manuscript. We are grateful to Director Liang Jun (Department of Rheumatology and Immunology, Nanjing Drum Tower Hospital, Nanjing, China) for his guidance and advice on the clinical classification of patients. We extend our heartfelt thanks to the National Natural Science Foundation of China (grant nos. 82202600 and 82003024), the Nanjing Drum Tower Hospital Clinical Research Special Fund Project (nos. 2022-LCYJ-PY-36 and 2023-LCYJ-MS-13), the Natural Science Foundation of Jiangsu Province (no. BK2020041989), Natural Science Foundation of Nanjing University of Chinese Medicine (NZY82003024) and Qing Lan Project for their invaluable financial support for the realization of these research goals. The authors declare they have no conflicts of interest. We adhered to the guidelines of the Declaration of Helsinki, and our research was endorsed by the ethics committee of Nanjing Drum Tower Hospital (ethics code: 2022-466-01, registration number of China Clinical Trial Registration Center: ChiCTR2100048082). Trial registration: Clinical Trial Registration Center, Registered 30 June 2021, https://www.chictr.org.cn/bin/home. Individuals whose faecal samples were part of the research willingly granted informed consent. Not applicable. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts