Transcriptomic profile investigations highlight a putative role for NUDT16 in sepsis

Abstract Sepsis is an aberrant systemic inflammatory response mediated by the acute activation of the innate immune system. Neutrophils are important contributors to the innate immune response that controls the infection, but harbour the risk of collateral tissue damage such as thrombosis and organ dysfunction. A better understanding of the modulations of cellular processes in neutrophils and other blood cells during sepsis is needed and can be initiated via transcriptomic profile investigations. To that point, the growing repertoire of publicly accessible transcriptomic datasets serves as a valuable resource for discovering and/or assessing the robustness of biomarkers. We employed systematic literature mining, reductionist approach to gene expression profile and empirical in vitro work to highlight the role of a Nudix hydrolase family member, NUDT16, in sepsis. The relevance and implication of the expression of NUDT16 under septic conditions and the putative functional roles of this enzyme are discussed.

(NETs)-related thrombosis and organ dysfunction. 1 Thus, a better understanding of how cellular processes in neutrophils and other contributing leukocytes are modulated during sepsis is needed. 4 Such an endeavour can be initiated via the investigation of transcriptomic expression profiles. To that point, the growing repertoire of publicly accessible transcriptomic datasets, like those on the Gene Expression Omnibus (GEO), serves as a valuable resource for discovering novel predictors, diagnostic markers and disease progression markers. Diverse reductionist approaches have been used to successfully identify putatively novel genes/functions [5][6][7][8][9] and perform system-level re-analyses. [10][11][12] The anti-inflammatory and immunosuppressive state of latephase sepsis resemble endotoxin tolerance in which the TLR4dependent pathway is non-responsive to gram-negative bacterial lipopolysaccharide (LPS; reviewed in 13 ). The underlying molecular mechanisms are centred on mRNA-decapping enzymes (DCP1) and the assembly of the translational repressor complex, which facilitate the degradation of TNFα and IL-6 mRNA and the formation of p-bodies in monocytic cells. 14 NUDT16 is a member of the Nudix hydrolase family and catabolizes nucleoside triphosphates, nonnucleoside polyphosphates and capped mRNAs, 15 initiating RNA turnover. 16 NUDT16 also possesses important DNA protective roles and can impact mediators of DNA repair. 17 In this study, NUDT16 was identified as a suitable candidate gene for a reductionist investigation of transcriptomic profiles based on the following criteria: 1) being consistently upregulated in septic patients compared with healthy individuals across multiple transcriptomic datasets, and 2) the absence of overlap between NUDT16 and sepsis in the current literature.

| An exploratory reductionist approach
Public repository of articles and data, such as PubMed and GEO, represents a vast resource but can be complicate to explore. Here, we present a logical reductionist approach to investigate putative novel biomarkers in sepsis. The approach can be applied to any field of research.
The steps consist of (1) identifying a gene of interest based on its differential expression in the pathological/physiological context of interest, (2) validating the reproducibility of the initial observation, (3) determining the current body of literature linking the gene and topic, (4) extracting the known biological concepts regarding to the gene and (5) inferring putative novel roles to the gene with literature support.

| NUDT16 expression in blood cells
A dataset that we have previously generated and deposited in GEO (GSE60424) was used to assess the expression of NUDT16 across leukocyte populations. The dataset consists of RNA-Seq profiles of neutrophils, monocytes, B cells, CD4+ T cells, CD8+ T cells and NK cells isolated from blood of healthy controls, patients with type 1 diabetes, amyotrophic lateral sclerosis, sepsis or multiple sclerosis prior to and 24 hours post treatment with interferon beta (up to 20 subjects/cell type). NUDT16 expression profile is accessible for this dataset via the GXB data browsing web application (link).
Significant differences (p < 0.0001) were determined by one-way ANOVA with Tukey's multiple comparisons test.

| Independent datasets for concordance
Datasets were obtained from GEO to be used to conduct independent validation of the initial finding in a relevant clinical setting.
Validation data were selected without a priori knowledge of NUDT16 expression levels and consists of only human studies in which transcriptome profiles were generated in septic patients and compared with uninfected controls (Table 1). Other information was retrieved from each GEO entry, such as the geographic locations of the patient populations under study which spanned four continents, and the type of biological samples, which included purified neutrophils, PBMCs, and whole blood. The studies included neonate, paediatric and adult populations. Finally, these data were generated using two different microarray platforms. Further sepsis-related datasets can be found and interactively explored in the database recently created by our group, SysInflam HuDB (link). 18

| In vitro stimulation assay
Heparinized diluted whole blood (WB) from healthy donors was exposed to media (control) or combined LPS/PGS at 100 ng/ml and 5 ug/ml, respectively, to simulate septic conditions. Samples were collected at 6 h post-stimulation for total RNA and overnight for protein expression analyses. Briefly, whole blood (maximum total of 4 ml) from each donor was collected via venipuncture into a heparinsulphate vacutainer (Becton Dickinson) and mixed at a 1:1 ratio with RPMI (Gibco). 500 ul of WB:RPMI mixture was added to cryovial or microwell already containing the specific stimulation. The samples were incubated in tissue culture incubator at 37°C with 5% CO 2 for the desired time. For transcript abundance, the 6 h incubation was optimally selected based on previously reported time-course data for NUDT16 expression (GSE3284). 19 The experiment was repeated twice.

| Flow cytometry
Cultured whole blood phenotyping and immunostaining for surface  Mann-Whitney test was used to identify significant differences between groups.

| RE SULTS
We identified NUDT16 as the candidate gene based on its expression in healthy neutrophils exposed to plasma from septic patients compared with healthy individuals. This observation was made in a dataset generated in our previous work, which is publicly available in GEO under accession GSE49755, 23 and can be visualized on the Gene Expression Browser (GXB, link). In brief, the study was After 6 h of exposure to the serum, neutrophils were lysed, and total RNA extracted for profiling on the Illumina HT12 Bead arrays. We downloaded the data and processed them for outlier identification (ROUT method). NUDT16 expression profile ( Figure 1A) showed significant differences in both the expression levels (4.9-fold change) and variance between septic and control serum treated groups (t- We found that NUDT16 is expressed in several immune cell types with the highest expression level in granulocytes and monocytes, suggesting a greater contribution from these cell types to the ex-  Table 1). Significant (p < 0.01) increases in NUDT16 expression in septic compared with nonseptic patients were observed in six datasets (Table 1, select datasets presented in Figure S2B). Differences in the magnitude of the change may be attributed to the in vivo design and the mixed cellular populations in whole blood or peripheral blood mononuclear cells.
Additional sepsis-related datasets are available for rapid and interactive exploration in SysInflam HuDB (Link). 18 Using SysInflam HuDB, we found that NUDT16 expression was consistently increased in pa- We then set up in vitro experiments to have a closer look at NUDT16 expression in whole blood. We measured the NUDT16 levels in cultured whole blood (obtained from healthy adult volunteers, n = 7) exposed to a combination of lipopolysaccharide (LPS) and peptidoglycan (PGN) or culture medium (control); we observed a 6.6-fold increase in NUDT16 expression with stimulation to unstim ( Figure 1D). A portion of the cultured whole blood samples mentioned above was analysed using multi-parameter flow cytometry. We observed contrasting NUDT16 abundance-we saw a 1.5-fold increase in the frequency of NUDT16+ leucocytes but a significant 1.85-fold decrease in the frequency of CD3+ T lymphocytes relative to the change in the respective cell population with stimulation ( Figure 1E).
These trends were consistent among the different individual donors, despite the different baseline frequencies ( Figure S2C). Hence, the calculated fold changes (stimulation/control) showed significance for both NUDT16+ leukocytes and lymphocytes compared with their respective parent populations ( Figure 1F); representative dot plots of the relative frequency of NUDT16+ cells before and after stimulation are shown in Figure S2D. The intrinsic level of NUDT16 in other identifiable cell subsets did not show significant changes with stimulation and was not included in the analyses (data not shown).
To gain a general overview of the biological functions of NUDT16, the literature mining approach was re-utilized. Biological concepts were manually extracted from the NUDT16 literature by using the NUDT16 query presented earlier, but this time restricting   Figure 1H).

| DISCUSS ION
The NUDT16 protein is primarily associated with 'RNA decapping' and 'ADP-ribose/ribosylation'; those concepts were presented in 35 and 38 of the articles retrieved. The conceptual link between decapping activity and sepsis/inflammation is supported by the role of the mRNA-decapping subunit 1 (DCP1), which is a known mediator of pro-inflammatory cytokine inhibition during sepsis. 14 In addition, NUDT16 has recently been shown to mediate the selective degradation of Rift Valley fever virus mRNA, implicating the enzyme in immune response. 24 The absence of articles among the sepsis literature relating to RNA decapping highlights a putative knowledge gap with regards to translational regulation of immune proteins, such as cytokines and chemokines, that are critical for sepsis development.

ACK N OWLED G EM ENT
The authors would like to thank the healthy blood donors and Sidra Medicine's Deep Phenotyping Core for their assistance with flow cytometry.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

E TH I C A L A PPROVA L
The study was approved by Sidra Medicine's Institutional Review Board. Written informed consents were obtained from the healthy donors.

CO N S E NT FO R PU B LI C ATI O N
All the authors consent for publication.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are derived from the following resources available in the public domain: NCBI Gene