Abstract

Background

   Accumulating evidence suggests tRNA-derived fragments (tRFs) play
   important roles in cellular homeostasis. Here we aimed to explore
   aberrant expression of tRFs in CD4^+ T cells from patients with
   systemic lupus erythematosus (SLE) and their potential function in
   the SLE pathogenesis.

Methods

   First, small RNA sequencing was performed on CD4^+ T cells from four
   SLE patients and three healthy controls (HCs). Candidate tRFs were then
   validated in CD4^+ T cells from 97 SLE patients and their relevant
   disease controls using qRT-PCR. Then sequencing was used to investigate
   the profiles of HC-derived CD4^+ T cells transfected with tRF-3009.
   Lastly, tRF-3009 siRNA or tRF-3009 mimics were transfected into CD4^+ T
   cells with/without IFN-α. Changes in oxygen consumption rate (OCR),
   ATP, and ROS production were analyzed.

Results

   We identified 482 differentially expressed tRFs from SLE CD4^+ T cells
   and chose tRF-3009 for further analysis due to its upregulation and the
   positive correlations between its expression and SLEDAI, active lupus
   nephritis and serum IFN-α levels. In vitro, tRF-3009 over-expressing
   CD4^+ T cell profiling and putative analysis linked this product to the
   type I IFN and oxidative phosphorylation (OXPHOS) pathways.
   Interestingly, IFN-α is capable of inducing ROS and ATP production in
   CD4^+ T cells, while knockdown of tRF-3009 reversed this process.
   Overexpression of tRF-3009 in CD4^+ T cells alone was sufficient to
   upregulate OCR, ROS, and ATP production.

Conclusions

   Our study is the first to link aberrant tRF expression and SLE.
   tRF-3009 may participate in metabolic modulation of IFN-α-induced CD4^+
   T cell OXPHOS in lupus.

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s12967-021-02967-3.

   Keywords: tRNA derived fragments (tRFs), Systemic lupus erythematosus
   (SLE), CD4^+ T cells, Oxidative phosphorylation (OXPHOS), Type I
   interferon (IFN)

Introduction

   Systemic lupus erythematosus (SLE) is a complex autoimmune disease with
   a largely undefined pathogenesis [[39]1]. Recently, changes in the
   metabolism of immune cells have been implicated in SLE [[40]2, [41]3].
   For example, SLE CD4^+ T cells were identified by their increased
   glycolysis and mitochondrial oxidative phosphorylation (OXPHOS) [[42]4,
   [43]5]. There is a growing belief that metabolic modulators (metformin
   and/or 2-Deoxy-D-glucose) could restore their cellular function and
   reverse certain disease phenotypes, this theory is supported by several
   studies using mouse models of lupus [[44]4, [45]6, [46]7]. In addition,
   our recent clinical trials suggest that the addition of metformin to
   the standard of care for these patients may reduce the risk of a lupus
   flare [[47]8].

   Transfer RNA-derived fragments (tRFs) are an abundant class of small
   non-coding RNAs (ncRNAs) that exist in most organisms [[48]9–[49]12].
   tRFs can be constitutively produced as a result of various
   pathophysiologies and their sequence is based on the different cleavage
   products produced from the mature and precursor tRNA transcripts
   [[50]13]. Growing evidence suggests that tRFs, like other ncRNAs, can
   serve as active modulators of cellular homeostasis [[51]14, [52]15].
   Recently, tRFs have been implicated in numerous diseases, including
   multiple cancers, neurodegenerative disorders, viral infections, and
   metabolic disorders. Accumulating evidence have shown their biomarker
   potential and determined the molecular mechanisms of tRFs‐mediated
   pathological processes [[53]15, [54]16]. Although their biological
   functions are still poorly understood, recent studies have shown sperm
   tRNA fragments might affect human metabolic disorders [[55]17, [56]18],
   indicating that tRFs may participate in metabolism process.

   Surprisingly, we have found tRFs were abnormally expressed in CD4^+ T
   cells isolated from SLE patients. To the best of our knowledge, tRFs
   have never been studied in SLE. This study aims to explore the
   potential effects of candidate tRFs in immunometabolism on the
   pathogenesis of SLE.

Materials and methods

Patients and study design

   This study included 187 participants and their clinical data are listed
   in Additional file [57]6: Table S1. Four new-onset SLE patients and
   three healthy controls (HCs) were chosen randomly using in the primary
   small RNA profiling study and the validation cohort consisted of 97 SLE
   patients , and 25 disease controls [10 rheumatoid arthritis (RA)
   patients and 15 ankylosing spondylitis (AS) patients] and 20 HCs. An
   additional 38 SLE patients were enrolled for further IFN-α-related
   assay. All patients fulfilled the classification criteria for SLE
   [[58]19], RA [[59]20], or AS [[60]21], respectively, and the enrollment
   was completed at the Department of Rheumatology, Renji Hospital,
   Shanghai Jiaotong University School of Medicine, from June 2017 to
   December of 2018. The SLE disease activity index (SLEDAI) [[61]22] was
   assessed and all patients with  >  500 mg/24 h proteinuria were defined
   as experiencing active lupus nephritis at the time of enrollment. Among
   all enrolled SLE patients, 71 subjects with stable disease activity, 31
   with moderate disease activity and 37 with severe disease activity. The
   SLEDAI score in all SLE patients was 5.77  ±  4.96. And 68 SLE patients
   were experiencing active disease. Based on the prednisone dose that
   patients taking, SLE patients were divided into low/median/high dose (<
    20 mg/d or  ≤  20 mg/d and  <  50 mg/d or  ≤  50 mg/d) group. Any
   subjects with acute or chronic infections were excluded and the study
   protocol was reviewed and approved by the institutional ethics
   committee, and written informed consent was obtained from all
   participants.

Isolation of CD4^+ T cells

   The peripheral blood lymphocytes (PBMCs) were isolated from HCs and SLE
   patients using Ficoll gradient centrifugation within 6 h of collection
   of the samples. CD4^+ T cells were then purified after immunomagnetic
   separation with the human CD4^+ T cell isolation kit (BD Biosciences)
   according to the manufacturer’s instructions. The isolated human
   primary CD4^+ T cells were lysed in Trizol Reagents (Sigma) or used to
   perform in vitro transfection. Samples from the validation cohort were
   used for qRT-PCR.

Cell culture and transfection

   CD4^+ T cells from HCs were cultured in Roswell Park Memorial Institute
   (RPMI)-1640 medium (Hyclone) supplemented with 10% fetal bovine serum
   (FBS) and 100 U/mL IL-2 at 37 °C in a humidified atmosphere containing
   5% CO[2] (Thermo Fisher). For transfection, CD4^+ T cells were seeded
   at a density of 0.5 × 10^6 cells/well and three types of tRF-3009 or
   three types of negative control (single-stranded RNA, ssRNA) (Guangzhou
   RiboBio Co., Ltd.) were transfected using RNAimax (Thermo Fisher)
   according to the manufacturer’s protocols. After 48 h of incubation at
   37 °C, cells were harvested for small RNA sequencing.

   To evaluate the metabolic changes CD4^+ T cells were seeded at a
   density of 1  ×  10^6 cells/well and stimulated with/without 1  ×  10^5
   U/mL IFN-α and transfected with or without 20 µM tRF-3009 siRNA or
   20 µM tRF-3009 mimic for 24 h. Metformin (Met) or 2-Deoxy-D-glucose
   (2-DG) were added at 1 mM for 24 h, and phytohemagglutinin (PHA) was
   added at 1 µg/mL as indicated.

Deep sequencing

   Total RNA was extracted from the primary human CD4^+ T cells using the
   TRIzol Reagent (Invitrogen) and the quality of the RNA samples was
   evaluated by using a Nanodrop 2000c Spectrophotometer (Thermo Fisher
   Scientific Inc.) and then size-fractionated on a 15% polyacrylamide gel
   electrophoresis (PAGE) gel to collect the 18–60nt fraction. The 5′ and
   3′ RNA adapters were ligated to the RNA pool, and then this was
   subjected to RT-PCR to generate the sequencing libraries. PCR products
   were purified and then sequenced using a HiSeq 2000 sequencing system
   (BGI Tech). tRFs were aligned to the tRFdb
   ([62]http://genome.bioch.virginia.edu/trfdb) and the transcripts were
   trimmed and aligned to the human genome (hg19). Differentially
   expressed tRFs and the transcripts were filtered using the following
   criteria: |Fold change  ≥  2.0| and P value  <  0.05 and |Fold change
    ≥  1.5| and FDR  <  0.05, respectively.

qRT-PCR validation of candidate tRFs and other mRNAs

   Quantification of tRFs was completed using the same methods as those
   for analyzing PIWI-interacting RNAs (piRNAs) with some modifications
   [[63]23]. Isolated total tRFs RNA or mRNA (1.0 μg) was polyadenylated
   at 37 °C for 20 min with poly(A) polymerase (NEB) and reverse
   transcription was performed using the Superscript II Reverse
   Transcriptase kit (Takara). RNA transcript levels were measured by
   qRT-PCR using SYBR Green PCR Master Mix. After adding forward and
   reverse primers, the mixtures were incubated at 95 °C for 10 min,
   followed by 40 cycles of 95 °C for 10 s and 58 °C for 10 s. All
   experiments were performed in biological triplicate and the data were
   analyzed using the comparative cycle threshold (Ct) method. U6/β-actin
   was used as an internal control and the relative quantification was
   calculated using the following equation: Amount of target gene = 2^−ΔCt
   where ΔCt = Ct [target genes]  −  Ct [U6/β-actin]. Gene-specific
   primers are listed in Additional file [64]7: Table S2. All samples were
   performed in triplicate.

Serum IFN-α level in lupus patients

   Serum was obtained from SLE patients and serum IFN-α levels were
   measured using human IFN alpha ELISA kits (Raybiotech, lnc) according
   to the manufacturer’s instructions [[65]24].

Bioinformatics analysis

   The functions of the differentially expressed tRF target genes were
   investigated using Gene Ontology (GO) annotation and Kyoto Encyclopedia
   of Genes and Genome (KEGG) pathway analysis. Hierarchical clustering of
   the differentially expressed genes (DEGs) according to their biological
   process (BP), cellular component (CC) and molecular function (MF)
   categories were completed using GO analysis and used to elucidate
   genetic regulatory networks ([66]http://www.geneontology.org). Pathway
   analysis using graphical diagrams was performed to explore the DEG
   pathways using the Pathway database ([67]http://www.genome.jp/kegg/)
   and a co-expression network was built using the Cytoscape software
   (version 3.6.1). Significance was determined by P value and FDR. And
   the protein–protein interaction network (PPI) of the tRF-3009
   interacting proteins was constructed and the module analysis was
   performed using STRING database.

   The target binding sites of the tRFs (targeting both the 3′UTR region
   and promoter regions of all annotated genes) were predicted using the
   miRanda database ([68]http://www.microrna.org/microrna/home.do).
   Significant 3′UTR regions and promoter regions were identified using
   the following criteria: energy ≤  − 15 and score  ≥  140 and
   energy ≤  − 20 and score ≥ 150, respectively.

Measurements of metabolic changes

   Oxygen consumption rate (OCR) was measured using an XF96 Extracellular
   Flux Analyzer under mitochondrial stress test conditions (Seahorse).
   Assay buffer was prepared from non-buffered RPMI medium (Sigma)
   supplemented with 10 mM glucose, 2 mM glutamine, and 1 mM pyruvate.
   Baseline OCR values were averaged between technical replicates for the
   first three successive time intervals. Extracellular lactate production
   was measured using a Glucose-lactic acid analyzer (SBA-40E).
   Mitochondrial membrane potential (△Ψm) was assessed using
   5,5,6,6′-tetrachloro-1,1,3,3′-tetraethyl benzimidazolyl carbocyanine
   iodide (JC-1, green^+ red^− cells%, Sigma). JC-1 is an ideal
   fluorescent probe whose mitochondrial uptake is directly related to the
   magnitude of the mitochondrial membrane potential [[69]25]. The greater
   the mitochondrial uptake, the greater concentration of JC-1 aggregate
   forms which have a red fluorescent emission signal, as opposed to the
   JC-1 monomer that fluoresces green [[70]25, [71]26]. The ratio of red
   to green fluorescence is often applied to measure the ratio of
   mitochondrial depolarization. Intracellular ATP was measured using the
   ATP Determination Kit. ROS were measured using DCFH-DA dye (Thermo
   Fisher), flow cytometry and a Fluorometric Intracellular Ros Kit
   (Sigma) [[72]27]. Blank medium was used as the negative control, and
   PHA was used as the positive control. PHA belongs to high molecular
   glycoprotein family and has the activity of promoting the mitosis of
   PBMCs [[73]28].

RNA pull down

   The healthy donor-derived CD4^+ T cells were transfected with
   5′-biotinylated synthesized tRF-3009 or a random sequence. CD4^+  T
   cells were collected in an extraction buffer 48 h post transfection.
   After separating the lysate by centrifugation at 14000 rpm for 15 min,
   the cell extract was incubated with streptavidin-coated microbeads
   overnight at 4 °C, following the manufacturer’s protocols (μMACS,
   Miltenyi Biotec.). The proteins interacting with the tRF-3009 were
   pulled down using magnetic separator, followed by three washes. The
   proteins were resolved on PAGE, followed by silver staining. Gels
   fragments that contained bands of the protein of interest were excised
   and analyzed by mass spectrometry (LC/MS).

Statistical analysis

   Data are presented as the mean  ±  SD. Two-tailed Student’s t test or
   Mann–Whitney test was used to evaluate the differences between the
   pairs of groups. Correlations between clinical information and tRF
   expression were evaluated using Pearson’s correlation test. All
   statistical analyses were performed using GraphPad Prism (version 6.0)
   and SPSS software (version 21.0). A P value of  <  0.05 was considered
   statistically significant.

Results

Differentially expressed tRFs in CD4^+ T cells from SLE patients and HCs

   To determine the differential expression of tRFs in CD4^+ T cells from
   SLE patients, small RNA sequencing was performed using cells from 4 SLE
   patients and 3 healthy controls. In addition to the regular peaks
   associated with miRNA expression, significantly aberrant expression of
   tRNA-derived small RNAs was observed in these samples (Fig. [74]1A). A
   total of 482 tRFs were significantly differentially expressed in our
   sequencing data (fold change  ≥  2 or  ≤  0.5 and P  <  0.05), of which
   220 tRFs were up-regulated and 262 tRFs were down-regulated. Among the
   482 dysregulated tRFs, sequence analysis of the abnormal tRNAs revealed
   the key enrichment of a common sequence motif: with the majority of the
   upregulated tRFs derived from the 3′ end of mature tRNAs with a ‘CCA’
   tail. While the downregulated tRFs were predominantly produced from the
   3′ end of tRNA precursors (Fig. [75]1B). Based on their sequences,
   these abnormally expressed small RNAs were derived primarily from
   tRNA-Leu, followed by tRNA-Ala, tRNA-Ser, and tRNA-Thr (Fig. [76]1C;
   Additional file [77]8: Table S3).

Fig. 1.

   [78]Fig. 1
   [79]Open in a new tab

   Identification of differentially expressed tRFs in CD4^+ T cells from
   systemic lupus erythematosus (SLE) patients and healthy controls (HCs).
   A Heatmap of differentially expressed tRFs between SLE patients (n  = 
   4) and HCs (n  =  3). The red indicates upregulated genes, and the
   green indicates downregulated genes. B Motif prediction showed the 3′
   end sequence characteristics of upregulated tRFs (upper panel) and
   downregulated tRFs (lower panel). C Sequences of the candidate
   differentially expressed tRFs and their precursors. D Expressions of
   candidate tRFs (tRF-3009, tRF-3021, tRF-1001 and tRF-1035) in CD4^+ T
   cells of SLE patients, rheumatoid arthritis (RA) patients, ankylosing
   spondylitis (AS) patients and HCs. The expression trends of four tRFs
   was consistent with the small RNA sequencing results. qRT-PCR was
   performed on RNA samples from 97 SLE patients, 10 RA patients, 15 AS
   patients and 20 HCs. Data are presented as 2^−ΔCt relative to U6
   expression. **P  <  0.01; ***P  <  0.001

Identification of the candidate SLE related tRFs from CD4^+ T Cells

   To further confirm the expression levels of the tRFs in SLE patients,
   the expression levels of the four most significant candidate tRFs were
   validated by qRT-PCR. Consistent with the tRF-seq data, the expression
   levels of tRF-3009 and tRF-3021 were both significantly higher in SLE
   samples compared with HCs and disease controls (AS or RA patients; P
    <  0.001), while the tRF-1001 and tRF-1035 expression levels were
   significantly lower in these samples (P  <  0.01; Fig. [80]1D). In
   addition, these tRFs expressions were not significantly different
   between the disease control groups and HC group.

tRF-3009 expression correlates with disease activity index, lupus nephritis
and serum IFN-α levels

   Based on the difference in the candidate tRFs in the validation cohort,
   we went on to evaluate the correlations between the expression levels
   of these tRFs in CD4^+ T cells and clinical phenotype. Among the two
   candidate up-regulated tRFs, only the expression level of tRF-3009 in
   CD4^+ T cells was positively correlated with SLEDAI (Fig. [81]2A).
   Interestingly, tRF-3009 expression level was significantly higher in
   patients with active lupus nephritis (LN) than in those without LN
   (Fig. [82]2B). Nevertheless, the other three tRFs showed no distinct
   connection with active SLE (Additional file [83]1: Figure S1). Among
   the SLE patients, we found that serum IFN-α levels were positively
   correlated with tRF-3009 expression in CD4^+ T cells (Fig. [84]2C).
   Furthermore, we have investigated the relationship between treatments
   (type and dose) and the levels of tRF-3009. We found the expression
   level of tRF-3009 showed no significant difference among different type
   of treatments (hydroxychloroquine and different immunosuppressants;
   Additional file [85]2: Figure S2A). Likewise, there was no significant
   difference in the tRF-3009 expression in different doses of prednisone
   groups (Additional file [86]2: Figure S2B). These analyses further
   suggest that tRF-3009 has a strong association with SLE in the
   validation cohort.

Fig. 2.

   [87]Fig. 2
   [88]Open in a new tab

   The association between the expression of tRF-3009 in CD4^+ T cells and
   clinical characteristics in systemic lupus erythematosus (SLE)
   patients. A Correlations between the expression of tRF-3009 and SLEDAI
   index in SLE patients (n  =  59). B The level of tRF-3009 expression
   between SLE patients with active nephritis and no nephritis. **P  < 
   0.01. C Correlations between the expression of tRF-3009 and the level
   of serum IFN-α in SLE patients (n  =  38)

Bioinformatics analysis of tRF-3009 regulated genes

   We used next generation sequencing to explore the molecular function of
   tRF-3009 in SLE CD4^+ T cells. When we compared the tRF-3009 mimics
   with the two negative control groups we were able to clearly identify
   1771 differentially regulated mRNA transcripts. Of these 861 were
   upregulated and 910 were downregulated when evaluated using the
   following criteria: |fold change| ≥  1.5 and FDR  <  0.05 (Fig. [89]3A,
   B; Additional file [90]9: Table S4).

Fig. 3.

   [91]Fig. 3
   [92]Open in a new tab

   Differentially expressed genes (DEGs) of CD4^+ T cell transfected with
   tRF-3009 and negative control (NC). A The Heatmap of DEGs of CD4^+ T
   cell from three healthy controls transfected with tRF-3009 (n  =  3)
   and NC (n  =  6). The red indicates upregulated genes, and the green
   indicates downregulated genes. B The Volcano plot of DEGs of CD4^+ T
   cell transfected with tRF-3009 (n  =  3) and NC (n  =  6). The red
   points indicate upregulated genes (n  =  861), and the blue points
   indicate downregulated genes (n  =  910). C Gene ontology (GO) analysis
   (left panel) and Pathway analysis (right panel) of DEGs. D Relation
   network of KEGG pathways enrichment of DEGs. The red indicates pathways
   that upregulated DEGs significantly enriched, and the blue indicates
   pathways that downregulated DEGs significantly enriched. E The DEGs
   related to type I IFN signaling pathway or respiratory electron
   transport chain. The purple indicates DEGs related to type I IFN
   signaling pathway, the red indicates DEGs related to respiratory
   electron transport chain

   These target genes were then evaluated using GO and pathway analysis.
   The GO analysis indicated that tRF-3009 target genes primarily
   participate in the process of translation, immune and inflammatory
   responses, the type I IFN response and the respiratory electron
   transport chain (ETC) (Fig. [93]3C). Meanwhile, the results of the KEGG
   analysis showed that the tRF-3009 target genes were significantly
   enriched for oxidative phosphorylation, ribosome, ‘systemic lupus
   erythematosus pathway’ and metabolic pathways (Fig. [94]3C; Additional
   file [95]10: Table S5). To further evaluate the associations between
   tRF-3009 target genes and SLE we went on to construct a
   target-gene-pathway network (Fig. [96]3D). The results of the Cytoscape
   analysis also indicate that tRF-3009 target genes link the type I IFN
   pathway and oxidative phosphorylation response. The visual regulation
   network was built up between tRF-3009 and its target genes involved in
   the type I IFN pathway and respiratory ETC (Fig. [97]3E). qRT-PCR
   further validated that over-expression of tRF-3009 up-regulated genes
   involved in respiratory ETC, including ND1 through ND6 and ND4L
   (Additional file [98]3: Figure S3).

IFN-α directly induces tRF-3009 expression and increases OXPHOS in CD4^+ T
cells

   Based on our transcriptome data and the results of the bioinformatics
   analysis, we went on to evaluate the connections between tRF-3009, type
   I IFN and respiratory ETC. First, healthy donor-derived CD4^+ T cells
   demonstrated a significant increase in both tRF-3009 and its precursor
   tRNA-Leu-TAA expression following IFN-α treatment (Fig. [99]4A), which
   suggests IFN-α could directly activate transcription of tRNA-Leu, and
   may also enhance the fragmentation of tRF-3009 at the same time. And
   the production of ATP and ROS (products of respiratory ETC) were also
   increased in these cells after IFN-α treatment (Fig. [100]4B, C). Then
   when CD4^+ T cells where pretreated with metformin or 2-DG and then
   exposed to IFN-α-induction both the OXPHOS and glycolysis responses
   were reduced and when these cells were treated with a combination of
   metformin and 2-DG these effects were completely abolished
   (Fig. [101]4B, C). To address whether tRF-3009 participates in this
   process, we designed three siRNAs against its tRNA precursor, all of
   which were able to block the expression of tRF-3009 when exposed to
   IFN-α-induction (Fig. [102]4D). si-tRF-3009-1 was selected and
   transfected into healthy donor-derived CD4^+ T cells and was shown to
   significantly suppress IFN-α-induced ROS and ATP production
   (Fig. [103]4E, F), showing equivalent to the suppressor metformin or
   2-DG. Therefore, tRF-3009 actively participates in the IFN-α-induced
   respiration metabolism.

Fig. 4.

   [104]Fig. 4
   [105]Open in a new tab

   The metabolism change in healthy donor-derived CD4^+ T cells induced by
   IFN-α following tRF-3009 treatment. All experiments show the mean of
   individual replicates (n  ≥  3) A The expression of tRF-3009 and its
   precursor tRNA-Leu-TAA in CD4^+ T cells treated with IFN-α for 24 h by
   qRT-PCR. Bars show the mean of individual replicates. B The change of
   ATP production in CD4 ^+ T cells with/without 2-DG or/and metformin
   after 24-h IFN-α treatment (n  =  6). C The change of ROS production in
   CD4^+ T cells with/without 2-DG or/and metformin after 24-h IFN-α
   treatment. D The expression of tRF-3009 in CD4^+ T cells with IFN-α
   treatment transfected three siRNA against tRF-3009 precursor. The
   sequences of three si-tRF-3009 and target sites in the tRF-3009 were
   shown. E The change of ATP production in CD4^+ T cells induced by IFN-α
   treatment after transfected si-tRF-3009-1(n  =  6). F The change of ROS
   production in CD4^+ T cells induced by IFN-α treatment after
   transfected si-tRF-3009-1 (n  =  6). **P  <  0.01; ***P  <  0.001

Transfection of tRF-3009 induced oxidative metabolism changes in CD4^+ T
cells

   To further dissect the exact metabolic pathway affected by tRF-3009
   expression, we went on to evaluate the lactate level of treated CD4^+ T
   cells. There were no differences in the lactate levels of these cells
   (Fig. [106]5A), suggesting that tRF-3009 is unlikely to directly
   modulate glycolysis. As we speculated, mitochondrial oxidative
   phosphorylation, which is represented by OCR, was significantly
   increased in CD4^+ T cells transfected with tRF-3009 mimic when
   compared to the negative controls (medium or transfection with ssRNA;
   Fig. [107]5B; Additional file [108]4: Figure S4). The results of the
   JC-1 staining indicate that the mitochondrial membrane potential in
   CD4^+ T cells transfected with tRF-3009 mimic was also significantly
   elevated compared with the two negative controls (Fig. [109]5C).
   Likewise, ATP and ROS production in CD4^+ T cells transfected with
   tRF-3009 mimic were also enhanced (Fig. [110]5D–F). These results
   suggest that tRF-3009 expression is sufficient to directly regulate
   OXPHOS.

Fig. 5.

   [111]Fig. 5
   [112]Open in a new tab

   In vitro transfection of tRF-3009 caused oxidative metabolism changes
   in healthy donor-derived CD4^+ T cells. All experiments show the mean
   of individual replicates (n  ≥  3). A The change of lactate level
   between healthy donor-derived CD4^+ T cells transfected with tRF-3009
   mimic, negative controls or positive control (n  =  6). The medium and
   random ssRNA were used as negative controls, and PHA was used as
   positive control. B The change of oxygen consumption rate (OCR) in
   CD4^+ T cells transfected tRF-3009 mimic in vitro (n  =  6). C The
   change of mitochondrial membrane potential (ΔΨm) as estimated by JC-1
   in CD4^+ T cells transfected tRF-3009 mimic in vitro (n  =  4). The
   greater the mitochondrial uptake, the greater concentration of JC-1
   aggregate forms which have a red fluorescent emission signal, as
   opposed to the JC-1 monomer that fluoresces green. D The change of ATP
   production in CD4^+ T cells transfected tRF-3009 mimic (n  =  3). E The
   change of ROS production as estimated by DCFH-DA dye in CD4^+ T cells
   transfected tRF-3009 mimic (n  =  3). F The change of ROS production as
   estimated by fluorometric analysis in CD4^+ T cells transfected
   tRF-3009 mimic (n  =  3). NS, no difference; **P  <  0.01; ***P  < 
   0.001

The underlying effects of tRF-3009 in the metabolic processes of CD4^+ T
cells

   To address the question of how tRF-3009 affects metabolic processes, we
   performed RNA pull down to explore the immediate “receptor” binding
   molecule of tRF3009. Here CD4^+ T cells were evaluated using synthetic
   biotinylated tRF-3009 to identify interacting proteins. RNA pull down
   products were subsequently identified by mass spectrometry (LC/MS;
   Additional file [113]5: Figure S5A), and the tRF-3009 interacting
   molecules was shown to be closely related to ATP binding and ion
   binding/transportation (Additional file [114]5: Figure S5B). Then the
   STRING database was used to further explore the potential interaction
   network between these molecules (Additional file [115]5: Figure S5C).
   And the Reactome Pathway enrichment analysis of these interacting
   proteins showed that these candidate proteins were enriched in the
   formation of ATP, mitochondrial biogenesis, and respiratory electron
   transport which are closely related to OXPHOS (Additional file [116]5:
   Figure S5D). These data further supported the proposal that tRF-3009
   participates in metabolic modulation in CD4^+ T cell OXPHOS in SLE
   patients.

Discussion

   The original goal of this study was to study small RNAs in CD4^+ T
   cells of SLE patients. Surprisingly, in addition to some miRNAs with
   well-known function, the small RNA sequencing data also showed a series
   of tRNA derived fragments(tRFs). This is the first study to identify
   differentially expressed tRFs in CD4^+ T cells from SLE patients, which
   caught our attention. As a new category of small ncRNAs, the potential
   function of tRFs in the SLE pathogenesis was explored.

   In the sequencing data, expression levels and the motif patterns of the
   tRFs indicated that they are not random products of CD4^+ T cells. This
   prompted us to think about the role of tRFs in the pathogenesis of SLE.
   The correlation between tRNA fragment and clinical indicators of SLE
   was analyzed, amongst these tRNA fragments, tRF-3009, a small fragment
   (18nt long) processed from the 3′ end of mature tRNA-Leu, was
   positively correlated with SLEDAI, active lupus nephritis, and the
   serum IFN-α level, but showed no significant relationship with
   different treatments in SLE. The correlations with SLEDAI and active
   lupus nephritis suggested that tRF-3009 may be related to the severity
   of disease and existence of nephritis. Moreover, the correlation with
   serum IFN-α level suggested that the expression or function of tRF-3009
   may be related to IFN signaling pathway in lupus. Limited by the number
   of patients, we are unable to determine the potential of this molecule
   as a disease biomarker this time, which would be our next research
   goal.

   To further explore potential function of tRF-3009 in CD4^+ T cells of
   SLE, we performed in vitro transfection of tRF-3009 mimic, to analyze
   the transcriptome change induced by tRF-3009 over expression.
   Transcriptome and bioinformatics analysis suggested that multiple
   SLE-related pathways were activated in response to this tRF. More
   specifically, the results of the GO and KEGG analyses identified the
   potential effects of tRF-3009 expression on type I IFN signaling and
   OXPHOS metabolism in CD4^+ T cells. As reported previously [[117]29],
   mitochondrial hyperpolarization has been by several potential sensors
   in lupus T cells, along with increased mitochondrial ROS production and
   diminished ATP synthesis. Therefore, to confirm tRF-3009 may
   participate in OXPHOS, we determine the change of mitochondrial
   membrane potential, ROS and ATP production caused by tRF-3009. Our data
   further demonstrated that IFN-α induces CD4^+ T cell OXPHOS in a
   tRF-3009-dependent manner and that tRF-3009 knockdown suppressed
   IFN-α-induced ROS production and ATP biogenesis in vitro. This result
   reflected tRF-3009 knockdown and metformin or 2-DG have similar
   moderate effect in modulating OXPHOS and glycolysis responses.
   Furthermore, tRF-3009 overexpression upregulates OXPHOS but not
   glycolysis in CD4^+ T cells, indicating that this small RNA may serve
   as a metabolic modulator downstream of the type I IFN pathway, which in
   turn could enhance OXPHOS in CD4^+ T cells of SLE.

   The serum IFN-α level of SLE patients showed positive correlation with
   tRF-3009 expression, IFN-α also could promote tRF-3009 and precursor of
   tRF-3009 expression in vitro, indicating a transcriptional regulation
   of IFN-α in CD4^+ T cells. So here we suspected that IFN-α may be an
   important character in tRF-3009 expression and function. Also, our
   findings raise a ‘chicken or egg’ scenario for the type I IFN response
   and OXPHOS metabolism. A previous study demonstrated that mitochondrial
   oxidative stress is a key source of type I IFN production in SLE,
   evident by enhanced NETosis and oxidized mitochondrial DNA, strongly
   stimulating type I IFN signaling [[118]30]. As for the natural type I
   IFN-producing cells, that is, plasmacytoid dendritic cells (pDC), it
   has been reported that IFN-α can upregulate pDC fatty acid oxidation
   and OXPHOS via an autocrine type I IFN receptor-dependent pathway
   [[119]31]. Moreover, our data suggests that the metabolic OXPHOS in
   CD4^+ T cells, which have type I IFN receptors on their cell surface,
   may be regulated by type I IFN in the context of SLE. Currently
   available data appears to give us more of a snapshot of disease
   progression reflecting different aspects of the complex and dynamic
   interactions between IFN responses and cellular metabolism in different
   immune cells. This means that while the data may seem contradictory it
   is more likely that there is a complex feedback network regulating IFN
   signaling and mitochondrial oxidative stress in SLE. Along with the
   clinical progress of repurposing metabolic modulators, such as
   metformin, in the management of SLE [[120]8, [121]32, [122]33], our
   data may offer additional insight into the immunometabolism of SLE.

   Furthermore, attempts to predict the target genes of tRF-3009 (3′UTR
   region and promoter region) using miRNA prediction databases were a
   failure due to the lack of intersection between predicted target genes
   and transcriptome data (Additional files [123]9, [124]11, [125]12:
   Tables S4, S6, S7). The prediction method for miRNA is probably not
   applicable for tRFs, which might indicate that tRFs function in a
   different way. Additionally, we performed RNP pull down using tRF-3009
   and analyzed the resulting immunoprecipitation products to identify
   possible targets. With further bioinformatics analyses, the results
   suggested that these potential interaction partners were enriched in
   ATP production, mitochondrial biogenesis, and respiratory electron
   transport, which are closely related to OXPHOS. These results were
   preliminary without independent validation, and we are still in the
   early phases of developing a roadmap to evaluate the precise mechanisms
   of tRF-3009 mediated metabolic regulation.

   Still, the major limitation of our study is the lack of a mechanistic
   explanation for the OXPHOS modulating effect of tRF-3009. For example,
   it is still questionable how ATP production is affected by tRF-3009.
   Lupus PBL has been previously demonstrated to exhibit increased O[2]
   consumption through mitochondrial ETC complex I [[126]34]. However, in
   the current study, the Seahorse assay could not permit functional
   analysis of individual ETC complexes. And we failed to determine
   increased ATP production originate from complex I or distributed across
   several complexes in SLE over healthy control or RA T cells. Another
   limitation of our study is the lack of biomarker-scale sample size and
   cross-sectional data, which can’t speculate the potential and
   development of biomarkers. As reported previously, RA subjects T cells
   also display rebalancing glucose utilization and restoring oxidant
   signaling [[127]35]. In view of the metabolic regulatory links of the
   tRF-3009, the conclusion should be further validated in RA patients.
   The clinical significance of tRF-3009 in SLE is still undetermined and
   a larger prospective longitudinal study is pending.

Conclusion

   In summary, we demonstrate that tRF-3009 participate in metabolic
   modulation in IFN-α-induced CD4^+ T cell OXPHOS in SLE patients. This
   means that tRF-3009 may serve as a potential therapeutic target and may
   help to advance our understanding of CD4^+ T cell pathogenesis in SLE.

Supplementary Information

   [128]12967_2021_2967_MOESM1_ESM.tif^ (2.5MB, tif)

   Additional file 1: Figure S1. The association between the expressions
   of tRF-3021, tRF-1001, and tRF-1035 in CD4^+ T cells and clinical
   characteristics in systemic lupus erythematosus (SLE) patients.
   [129]12967_2021_2967_MOESM2_ESM.tif^ (10.9MB, tif)

   Additional file 2: Figure S2. The relationship between treatments and
   the levels of tRF-3009 in SLE patients. A The expression level of
   tRF-3009 among different type of treatments (hydroxychloroquine and
   different immunosuppressants). B The tRF-3009 expression in different
   doses of prednisone groups.
   [130]12967_2021_2967_MOESM3_ESM.tif^ (1.2MB, tif)

   Additional file 3: Figure S3. Validation of tRF-3009 target genes
   expressions in respiratory electron transport chain. After transfection
   of tRF-3009, the change of the candidate target genes expression in
   CD4^+ T cells. Data are presented as 2^−ΔCt relative to β-actin
   expression. Bars show the mean of individual replicates (n = 3).
   [131]12967_2021_2967_MOESM4_ESM.tif^ (3.3MB, tif)

   Additional file 4: Figure S4. In vitro transfection of tRF-3009 induced
   metabolism changes in CD4^+ T cells using Seahorse assay. Proton leak
   (A) and ATP production (B) in CD4^+ T cells transfected tRF-3009 mimic
   or random ssRNA. O[2] consumption (C) and Energy Map (D) of CD4^+ T
   cells transfected tRF-3009 mimic or random ssRNA.
   [132]12967_2021_2967_MOESM5_ESM.tif^ (9.1MB, tif)

   Additional file 5: Figure S5. The potential mechanistic role of
   tRF-3009 in CD4^+ T cells in terms of regulating OXPHOS. A The
   immunoprecipitation products from RNA pull down using tRF-3009 were
   shown by mass spectrometry (LC/MS). B The potential “receptor” binding
   molecule of tRF-3009. C Protein and protein interaction (PPI) network
   of the potential “receptor” binding molecule of tRF-3009. D Rectome
   Pathway analysis of the potential “receptor” binding molecule of
   tRF-3009. The size of the spots represents the number of the potential
   molecule of tRF-3009, and the color of the spots represents the P
   value.
   [133]12967_2021_2967_MOESM6_ESM.xlsx^ (10.5KB, xlsx)

   Additional file 6: Table S1. Clinical charateristics of SLE patients,
   disease controls and healthy controls.
   [134]12967_2021_2967_MOESM7_ESM.xlsx^ (10.2KB, xlsx)

   Additional file 7: Table S2. Primers used for qRT-PCR.
   [135]12967_2021_2967_MOESM8_ESM.xls^ (89KB, xls)

   Additional file 8: Table S3. Differentially expressed tRFs in CD4^+ T
   cells of SLE patients and healthy controls (Has).
   [136]12967_2021_2967_MOESM9_ESM.xls^ (493KB, xls)

   Additional file 9: Table S4. Differentially expressed genes between
   transfection of tRF-3009 and NC groups.
   [137]12967_2021_2967_MOESM10_ESM.xlsx^ (9.9KB, xlsx)

   Additional file 10: Table S5. The KEGG analysis of differentially
   expressed genes in the ‘systemic lupus erythematosus pathway’.
   [138]12967_2021_2967_MOESM11_ESM.xlsx^ (206.8KB, xlsx)

   Additional file 11: Table S6. Predicted target 3′UTR regions of
   tRF-3009.
   [139]12967_2021_2967_MOESM12_ESM.xlsx^ (213.7KB, xlsx)

   Additional file 12: Table S7. Predicted target promoter regions of
   tRF-3009.

Acknowledgements