Sodium Butyrate Attenuated Diabetes-Induced Intestinal Inflammation by Modulating Gut Microbiota

Background Diabetes mellitus (DM) continues to be one of the world's most costly and complex metabolic disorders. Accumulating evidence has shown that intestinal dysbiosis and associated inflammation can facilitate the onset and progression of DM. In this work, our goal was to investigate how sodium butyrate (SB) controls the gut microbiota to reduce the intestinal inflammation brought on by diabetes. Methods Male KK-Ay mice were randomized into two groups: the DM model group (intragastric administration of 0.9% normal saline) and the SB treatment group (intragastric administration of 1,000 mg/kg/d SB). The C57BL/6J mice were used as the control group (intragastric administration of 0.9% normal saline). These mice were administered via gavage for 8 weeks. Results The results revealed that SB-treated mice significantly reduced fasting blood glucose (FBG), body weight, 24 h food and water intake, and improved islet histopathology in DM model mice. SB reduced TNF-α, IL-1β, and iNOS, whereas it enhanced the expression of the anti-inflammatory Arg-1 marker on intestinal macrophages and the secretion of anti-inflammatory IL-10. Specifically, SB was linked to a marked drop in the expression of the Th17 marker RORγt and a substantial increase in the expression of the Treg marker Foxp3. SB treatment was associated with significant reductions in the levels of Th17-derived cytokines such as IL-17 and IL-6, whereas anti-inflammatory Treg-derived cytokines such as TGF-β were increased. Additionally, the analysis results from 16S rDNA sequencing suggested that SB significantly reversed the variations in intestinal flora distribution and decreased the relative abundance of Weissella confusa and Anaerotruncus colihominis DSM 17241 at the species level as well as Leuconostocaceae, Streptococcaceae, and Christensenellaceae at the family, genus, and species levels. These distinct florae may serve as a diagnostic biomarker for DM-induced intestinal inflammation. In addition, the heat map of phylum and OTU level revealed a close relationship between DM-induced intestinal inflammation and intestinal microbiota. Conclusions The present study suggested that SB may reduce DM-induced intestinal inflammation by regulating the gut microbiota.


Introduction
Diabetes mellitus (DM) is a chronic low-grade inflammatory, metabolic disease in which patients present with symptoms including insulin resistance and elevated blood glucose levels. DM is one of the primary causes of death, with approximately 451 million cases among adults >18 years old in 2017, and this number is expected to rise to 693 million by 2045 [1,2]. Growing evidence suggests intestinal dysbiosis and associated inflammation may facilitate DM onset and progression [3][4][5]. According to the reported studies, the intestinal inflammation observed in db/db mice has been associated with abnormal enteric glucose sensor functionality that results in inaccurate neuronal signaling and a consequent failure to increase hypothalamic NO release [6]. e role of intestinal inflammation as a mediator of DM progression has also been supported by work using mice in which the innate immune receptor tolllike receptor 5 (TLR5) had been knocked out [7,8]. Moreover, high-fat diet (HFD) consumption is known to cause significant increases in total gut permeability, gastrointestinal inflammation, and oxidative stress. DM is frequently associated with changes in the composition or permeability of the intestinal barrier [9][10][11][12]. e reported studies have suggested that discovering novel methods for decreasing intestinal inflammation in DM patients may be a promising option for slowing disease progression or alleviating related symptoms [13]. In addition, accumulating data suggests that gut immunity plays a vital role in regulating glucose homeostasis [14,15]. In recent decades, it has been revealed that inhibition of colonic pro-inflammatory macrophage infiltration might prevent HFD-induced insulin resistance from relieving DM [16]. ese findings suggested that intestinal inflammation and immunological responses may significantly regulate type 2 DM. However, the underlying regulatory mechanism is unknown.
Furthermore, it has been suggested that gut flora disorders accelerate the onset of inflammatory and chronic metabolic diseases [17][18][19][20]. erefore, targeting gut bacteria may be an effective treatment for diabetes-induced intestinal inflammation. It was found that DM patients have been shown to exhibit significantly reduced levels of butyrateproducing bacteria within the gut lumen relative to healthy controls [21], indicating that butyrate may be one crucial mediator of DM progression. Consistent with this hypothesis, previous research suggests that dietary supplementation with butyrate can improve gut integrity and protect against DM in animal model systems [22]. Sodium butyrate (SB) supplementation has also been shown to reduce inflammation and slow disease progression in the db/ db murine model of DM [23]. ese results suggested that SB could significantly improve DM.
In addition, SB has potential anti-inflammatory properties, affects the intestinal barrier, and plays a role in satiety and oxidative stress [24]. e SB significantly reduced pathological intestinal damage, lowered intestinal inflammation, and repaired intestinal flora disruption in mice with necrotizing enterocolitis (NEC) [25]. Furthermore, SB alters the intestinal flora composition and improves the gut barrier in HFD mice [26]. Additionally, colon cancers in hosts and the composition of the gut flora are also affected by SB. However, the effect of SB on the gut microbiota in treating DM-induced intestinal inflammation is unknown [27]. In this view, the current study used the KK-Ay spontaneous DM mouse model system with features consistent with human T2DM to evaluate the mechanism of action of SB in improving diabetes-induced intestinal inflammation by regulating gut microbiota. e research may offer novel ideas for SB to treat diabetes-induced intestinal inflammation from the perspective of new mechanism exploration. Before the trial, these experimental mice received two weeks of adapted feeding. e Nanjing University of Chinese Medicine's Animal Ethics Committee approved the current study (approval no. ACU-13 (20161011)).

Experimental
Design. KK-Ay mice were fed a high-fat, high-sugar diet for 8 weeks, and C57BL/6J mice were fed a standard chow diet for 8 weeks. Male KK-Ay mice were randomized into 2 groups (n � 6/group): the DM model group (intragastric administration of 0.9% normal saline) and the SB treatment group (intragastric administration of 1,000 mg/kg/d SB), while the C57BL/6J mice as the control group (intragastric administration of 0.9% normal saline). For 8 weeks, the treatment was applied orally to the mice. Treated animals were orally administered SB or an equivalent volume of saline for 8 weeks. Control mice were fed a standard chow diet during this period, whereas all other animals were fed an HFD (60.5% standard chow, 24% lard, 10% sugar, 0.2% cholesterol, and 5% egg yolk powder). After 4 and 8 weeks, we measured FBG in all of the above mice using blood collected from the tail vein after a 12 h fast. Before sacrificing study animals, we collected blood samples in heparin-coated tubes. Serum was collected by spinning these tubes for 20 minutes at 3,000 rpm at 4°C before storage at −20°C. Following blood collection, animals were sacrificed, and islet tissue samples were isolated and fixed using 10% formalin. ey were then stored at 4°C before use. Colon samples were additionally collected from these animals, were rinsed using a saline solution, and were stored at −80°C.

Histopathological Evaluation.
Islet tissue sections were fixed in 10% formalin, embedded in paraffin, sliced up 5 μm sections, and stained with hematoxylin and eosin. en, islet histology was assessed using a microscope (400×); ImageJ is used to measure the islet area.

Cytokine Level
Measurements. TNF-α, IL-6, IL-1β, TGFβ, IL-17, and IL-10 levels in colon samples were assessed via ELISA based on provided instructions. Briefly, the samples were exposed to primary antibodies for an hour in a 96-well plate. Next, the wells were washed 5 times, and 100 μL of the chromogenic substrate was added for 15 minutes at 37°C. en the absorbance was measured at 450 nm within 15 minutes of adding 50 μL of stop solution to each well.

Western
Blotting. Samples of murine colon tissue were collected, homogenized using RIPA buffer containing protease inhibitors, and spun for 20 minutes at 12, 000 × g at 4°C. A BCA kit was then used to assess the amount of protein in each sample, after which 40 μg of protein per sample was separated via 10% SDS-PAGE and transferred to methanolactivated PVDF membranes. Next, 5% BSA was used to block these membranes for 2 h, and the membranes were incubated overnight with antibodies specific for iNOS, Arg-1, ROR gamma(t), and FoxP3 (all 1:1,000 in 5% BSA) at 4°C. en the blots were washed thrice using TBST, followed by 2 h incubation with the HRP-linked secondary antibody (1: 10,000). After additional washes, the blots were visualized using an enhanced chemiluminescence reagent. β-Actin, as a normalization control and ImageJ software, was used to assess band densitometry. ® soil DNA kit (mega Bio-tek, Norcross, GA, United States) was used to collect fecal microbial DNA. e purity of the DNA was assessed using 1% agarose gel electrophoresis, while the concentration of DNA was determined through a UV-vis spectrophotometer (Nano-Drop 2000, ermo Scientific, Wilmington, USA). Next, the V3-V4 hypervariable regions of the bacterial 16S rDNA gene were amplified by PCR using primer 338F (5′-ACTCC-TACGGGAGGCAGCAG-3′).

Illumina Miseq Sequencing and Bioinformatics Analysis.
PCR products were extracted from a 2% agarose gel and further purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA), followed by quantification using the Quantifluor-ST system (Promega, USA). e pure amplified fragments were assembled into a PE2 * 300 library using the standard operating protocols for the Illumina MiSeq platform (Illumina, San Diego, USA). Raw readings were stored in the NCBI Sequence Readings archive database. Illumina double-terminal reads were joined, filtered, and then conducted by the Quantifluor-ST software package. en the sequences with ≥97% similarity were allocated to the same operational taxon (OTU). e typical sequences of each OTU were filtered out, and the differences between the dominating species were then examined. the significance threshold. To demonstrate the relationship between various flora, FBG, and intestinal inflammatory cytokines, SPSS determined Pearson's correlation coefficient. e differential flora and the above parameters were plotted on a heat map using Pearson's correlation coefficient through GraphPad software.

SB Treatment Improves Common Symptoms in DM Model
Mice. After 0, 4, and 8 weeks of body weight, 24 h water, food intake, and FBG values, all experimental animals were measured. e body weight, 24-hour food and water consumption, and FBG levels of DM model mice were significantly higher than those of control mice. In contrast, the same variables in animals treated with SB were substantially lower than those in DM model mice (Figures 1(b)-1(e)).

SB Treatment Improves Islet Histopathology in DM Model
Mice. We analyzed the islet area to assess how SB treatment affected islet histopathology in DM model mice. Islet area was significantly reduced in DM model animals relative to controls ( Figure 2; P < 0.05), and SB treatment was associated with significant increases in islet area relative to DM model mice (P < 0.05).

SB Treatment Alleviates Intestinal Inflammation in DM Model Mice.
To assess the impact of SB on intestinal inflammation in DM model mice, we evaluated TNF-α, IL-β, IL-6, IL-17, IL-10, and TGF-β levels in colon samples from these animals. Relative to controls, DM model animals exhibited significantly increased levels of pro-inflammatory factors such as TNF-α, IL-β, IL-6, and IL-17, as well as significantly reduced levels of anti-inflammatory cytokines such as IL-10 and TGF-β ( Figure 3; P < 0.05). SB treatment was associated with significant reductions in TNF-α, IL-β, IL-6, and IL-17 levels, as well as with substantial increases in IL-10 and TGF-β levels in treated mice relative to those in the DM group (P < 0.05).

SB Promotes an Anti-Inflammatory Intestinal Microenvironment in DM Model
Mice. Staining for crucial marker proteins, such as iNOS and Arg-1, expressed by macrophages; RORct, expressed by 17 cells; and Foxp3, which Tregs express, allowed researchers to determine how SB affected the anti-inflammatory microenvironment in the intestines of DM model mice. Relative to control animals, DM model animals exhibited elevated expression of proinflammatory iNOS in intestinal macrophages and RORct + 17 cells, while levels of anti-inflammatory Arg-1 and Foxp3 + Tregs were decreased in these mice ( Figure 4; P < 0.05).

SB Treatment Modulates the Imbalance of Intestinal Flora in DM Model
Mice. Herein, the regulating effect of SB on the DM-associated alteration of intestinal flora was examined using 16S rDNA sequencing. e Shannon index was used to determine intestinal flora's alpha diversity (α-diversity). e DM mice model demonstrated a considerably lower Shannon index than controls. e Shannon index's dynamic variations revealed that intestinal flora's abundance was significantly higher in the SB group than in the DM model group ( Figure 5(a)). Moreover, PCA indicated that the composition of intestinal flora was different in each group ( Figure 5(b)). e findings of the hierarchical clustering tree analysis suggested significant clustering between the control group and the DM model group. Furthermore, the SB treatment group exhibited a better tendency for separation ( Figure 5(c)). In addition, LEfSe analysis was carried out to validate bacterial phenotypes with particular variations from phylum to genus to investigate the differences among control, DM model, and SB groups. Moreover, the LDA score (log 10 > 2) demonstrated remarkable alterations in 76 bacterial strains in the 3 groups ( Figure 5(d)).
e biodiversity of the intestinal flora varied significantly between groups.

SB Treatment Reversed Changes in Microbial Distribution in DM Mice. Leuconostocaceae, Streptococcaceae, and
Christensenellaceae were significantly increased in DM model mice relative to controls at the family level. At the same time, SB treatment was associated with a significant decrease relative to DM model mice ( Figure 6(a); P < 0.05). At the genus level, Collinsella, Weissella, Streptococcus, and Family XIII AD3011 group were substantially more prevalent in DM model mice compared to controls, while SB therapy was linked to a significantly lower prevalence when compared to DM model mice ( Figure 6(b); P < 0.05). Interestingly, 2 of the 3 groups among the 76 microorganisms were significantly different, including Weissella confusa (Weissella) and Anaerotruncus colihominis DSM 17241 (Anaerotruncus). Remarkably, Weissella confusa and Anaerotruncus colihominis DSM 17241 were increased dramatically in DM model mice relative to controls, while SB treatment was associated with a significant decrease relative to DM model mice ( Figure 6(c); P < 0.05). Given these results, SB restored the alterations in the underlined microorganisms' distribution and suggested that they could serve as novel biomarkers for identifying intestinal inflammation induced by diabetes.

Correlation Analysis between Differential Flora and Diabetes-Induced Intestinal Inflammation.
e relationship between intestinal flora, FBG, intestinal inflammatory cytokines, diabetes-related intestinal inflammation, and intestinal flora was examined using heat map analysis. In these heat maps, the more color is away from blue, the more negatively associated the two parameters are, while the more color is away from red, the more positively correlated the two parameters are. As shown in Figure 7(a), the heat map of phylum level indicated that Deferribacteres was positively related to FBG, pro-inflammatory cytokines, including TNF-α, IL-1β, IL-6, and IL-17, which were negatively associated with IL-10 and TGF-β. Saccharibacteria was positively related to pro-inflammatory cytokines, that is, TNF-α, IL-1β, IL-6, and IL-17, and negatively related to TGF-β. According to the heat map of OTU level (Figure 7(b)), FBG was negatively associated with Akkermansia, Bacteroidaceae, and the Bacteroidales S24-7 group. Furthermore, Akkermansia and Bacteroidales S24-7 group were negatively associated with TNF-α. In addition, Lactobacillaceae and Lachnospiraceae NK4A136 group were positively correlated to IL-6, while Akkermansia and Bacteroidales S24-7 group were negatively related to IL-6. Lachnospiraceae NK4A136 group was utterly associated with IL-17, while Akkermansia and Bacteroidales S24-7 group were negatively related to IL-17. IL-10 was positively correlated with Akkermansia, the Bacteroidales S24-7 group, and Bacteroides, while IL-10 was negatively correlated with the Lachnospiraceae NK4A136 group. Consequently, Akkermansia, Bacteroides, and the S24-7 group of bacteria positively correlated with TGF-β.

Discussion
DM is a group of metabolic diseases characterized by chronic hyperglycemia resulting from defects in insulin secretion, insulin action, or both. Furthermore, it affects 422 million adults across the globe, which is ∼8.5% of the world's population [5,28,29]. Several studies suggest that gut dysfunction can lead to diabetes by affecting glucose metabolism, triggering immunological responses, and increasing insulin resistance and low-grade inflammation [30]. According to a recent study, STZ increased pathological damage to the small intestine. is study indicated that high glucose toxicity might cause intestinal epithelium injury, and diabetic mice exhibited a significant drop in the expression of MFG-E8 and an increase in the expression of p-MLKL and HMGB1 in the ileum. e underlined data suggest intestinal inflammation plays a crucial role in DM [31]. Another study found that hyperglycemia could drive intestinal epithelial barrier function by interfering with intestinal epithelial cells in diabetic mice, contributing to the spread of irritating microbial metabolites throughout the body and enhancing the spread of intestinal infections [32].
Furthermore, several studies suggest that gut microbiota plays a crucial role in regulating DM. It has also been found that the low-grade inflammation in DM may be caused by changes in the intestinal flora [33]. ese findings suggested that targeting intestinal inflammation and gut microbiota might be potential approaches to treating diabetes. is study also validated the significance of intestinal Evidence-Based Complementary and Alternative Medicine inflammation and gut microbiota in KK-Ay diabetic mice. We showed a significant link between intestinal flora dysfunction and DM-induced intestinal inflammation. e KK-Ay spontaneous DM model was employed in this study to assess the impact of intestinal inflammation and gut microbiota on DM development. KK-Ay mice experienced typical DM symptoms such as polyphagia, polydipsia, weight loss, and elevated FBG levels, which were associated with decreased islet area. is was in line with the fact that people with DM frequently exhibit symptoms, including high blood glucose [34]. Significant intestinal inflammation was also observed in DM model mice, accompanied by increased pro-inflammatory cytokines and decreased antiinflammatory cytokines. Moreover, DM model mice exhibited elevated pro-inflammatory iNOS in intestinal macrophages and RORct + 17 cells, while anti-inflammatory Arg-1 and Foxp3+ Tregs decreased levels. e underlined data suggested that pro-inflammatory macrophages and 17 cells significantly regulate DM-induced intestinal inflammation, which was consistent with the previous studies [31][32][33]. 16S rDNA sequencing of fecal intestinal flora revealed that DM model mice exhibited a significant disruption in the intestinal flora. e results of Shannon indexes demonstrated that the abundance and diversity of intestinal flora were dramatically decreased in DM model mice. Interestingly, the LDA score (log 10 > 2) revealed that the distinct microbial lineages in the DM model group comprised 36 bacteria, including Enterobacteriaceae, Gammaproteobacteria, Enterobacteriales, Candidatus saccharimonas, Escherichia Shigella, and so on. ese characteristic florae may serve as potential markers for diagnosing DM. Substantial evidence has shown that gut microbiota was crucial in regulating DM. Recent studies have reported that Deferribacteres [35,36] and Saccharibacteria [37,38] at the phylum level were significantly increased in DM, suggesting that Deferribacteres and Saccharibacteria may play a crucial role in regulating DM  Evidence-Based Complementary and Alternative Medicine development. Interestingly, the correlation analysis of gut microbiota with diabetes-induced intestinal inflammation showed that Deferribacteres were positively related to FBG, pro-inflammatory cytokines such as TNF-α, IL-1β, IL-6, and IL-17, which were negatively associated with IL-10 and TGFβ. Furthermore, Saccharibacteria was positively related to the pro-inflammatory cytokine such as TNF-α, IL-1β, IL-6, and IL-17 and negatively related to TGF-β. According to our preliminary findings, Deferribacteres and Saccharibacteria may significantly contribute to triggering intestinal inflammation and driving the onset of DM. Additionally, the correlation analysis between the OTU heat map and intestinal inflammation revealed a significant relationship between gut microbiota, intestinal inflammation, and DM. In contrast to Akkermansia, Bacteroidaceae, and the Bacteroidales S24-7 group, Escherichia Shigella showed a positive correlation with FBG on the heat map of OTU level. Moreover, Akkermansia and Bacteroidales S24-7 group were negatively correlated to TNF-α. Lactobacillaceae and Lachnospiraceae NK4A136 group were positively correlated to IL-6, while Akkermansia and Bacteroidales S24-7 group were negatively related to IL-6. Lachnospiraceae NK4A136 group was utterly associated with IL-17, while Akkermansia and Bacteroidales S24-7 group were negatively related to IL-17. Akkermansia, Bacteroidales S24-7 group, and Bacteroides were positively related to IL-10, while the Lachnospiraceae NK4A136 group was negatively associated with IL-10.
Bacteroidales S24-7 group, Bacteroides, and Akkermansia were positively correlated to TGF-β. ese distinct florae could serve as a biomarker for diagnosing intestinal inflammation caused by diabetes. Butyrate is considered the most potent SCFA for treating DM, even though it accounts for just 15% of total SCFAs [39]. Fasting blood glucose and free fatty acid levels were inversely related to circulating butyrate levels. Once butyrate concentrations rose above physiological levels, butyrate was shown to have anti-inflammatory effects [40]. NaB could improve T2DM development in db/db mice by promoting glycogen metabolism in hepatocytes to maintain blood glucose homeostasis through the GPR43-AKT-GSK3 signaling pathway [41]. Besides, SB could also inhibit the PERK-CHOP pathway of endoplasmic reticulum stress (ERS) to improve the diabetic model rats induced by streptozotocin combined with a high-fat diet [42]. In addition to significantly lowering DM, SB also can reduce intestinal inflammation. Butyrate is produced in high quantities in the colon by bacteria. It can inhibit histone deacetylases within intestinal macrophages, suppressing    Evidence-Based Complementary and Alternative Medicine their patterns of pro-inflammatory gene expression in a DSS-induced colitis model [43]. In vitro, SB significantly inhibited 5-Fluorouracil (5-FU) induced inflammatory responses in THP-1 and Caco-2 cells and maintained the integrity of tight junctions in intestinal mucosal epithelial cells [44]. Moreover, SB significantly reduced pathological intestinal damage, attenuated intestinal inflammation, and improved intestinal flora disturbance in NEC mice [25]. is was the first study to evaluate how SB could be used to treat diabetes by lowering intestinal inflammation and increasing the gut microbiota. e present findings indicated that SB reduced blood glucose levels, increased islet area, and improved signs and symptoms of DM, including polydipsia and polyphagia. Together, these results confirmed the ability of SB to protect against or treat DM. SB may ameliorate DM by improving intestinal inflammation. We also found that SB treatment significantly reduced pro-inflammatory iNOS expression and promoted anti-inflammatory Arg-1 expression in the intestinal macrophages of treated animals. In line with these results, SB treatment was associated with significant reductions in TNF-α and IL-β levels. It increased IL-10 levels, further emphasizing the ability of this metabolite to promote an anti-inflammatory microenvironment within the gastrointestinal tract. SB treatment was also associated with decreases in the expression of RORct, a marker of pro-inflammatory 17 cells responsible for producing inflammatory factors such as IL-6 and IL-17, both of which were decreased in samples from SB-treated DM model animals.
e expression of the Treg-associated transcription factor FoxP3 was also increased in animals treated with DM. is increase coincided with increases in levels of the Treg- associated anti-inflammatory cytokine TGF-β. e results suggested that SB might ameliorate diabetes-induced intestinal inflammation via suppressing intestinal macrophage-mediated inflammation. Also, further modulating the Treg/ 17 balance to promote an anti-inflammatory intestinal microenvironment. SB significantly reduced the abundance of Leuconostocaceae, Streptococcaceae, and Christensenellaceae in the DM mice at the family level. is was consistent with the enrichment of Leuconostocaceae [45,46], Streptococcaceae [47,48], and Christensenellaceae [49] in the pathological state of diabetes reported in the previous literature. ese results suggested that SB may target Leuconostocaceae, Streptococcaceae, and Christensenellaceae at the family level to ameliorate diabetes-induced intestinal inflammation. In the DM mice, SB considerably decreased the genus level abundance of Collinsella, Weissella, Family XIII AD3011 group, and Streptococcus. Additionally, Weissella confusa and Anaerotruncus colihominis DSM 17241 were significantly reduced by SB at the species level. is was consistent with the enrichment of Collinsella [50,51], Weissella genus [45] and its species Weissella confusa [52], and Anaerotruncus colihominis DSM 17241 [52] in the pathological state of diabetes reported in the previous literature. ese results also suggested that SB may target Collinsella, Weissella, Streptococcus, and Family XIII AD3011 group at the genus level and Weissella confusa and Anaerotruncus colihominis DSM 17241 at the species level to contribute to alleviating diabetes-induced intestinal inflammation. Moreover, the heat map correlation analysis between phylum or OTU level and intestinal inflammation further indicated that gut microbiota and intestinal inflammation played a crucial role in driving DM. ese results suggested that SB may reduce diabetes-induced intestinal inflammation via modulating gut microbiota. However, further research is needed to extensively evaluate how SB reduces intestinal inflammation caused by diabetes via regulating the underlined differential microbiota indicators.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that there are no conflicts of interest.

Authors' Contributions
Liping Liu and Yuping Chen contributed equally to this work and should be considered co-first authors.