**3. Discussion**

To date, the BTBRo<sup>b</sup> mouse has been understood as a model of severe T2D secondary to β-cell failure, with a consistent, full disease penetrance by 16 weeks of age in both sexes and earlier in males [4,6]. In this study, we encountered a unique cohort of male BTBRo<sup>b</sup> mice for which environmental factors, including but not limited to diet, completely prevented hyperglycemia without further intervention. We have exploited these findings to study physiological changes as they pertain to islet function in obese, insulinresistant, normoglycemic animals compared to their equally obese and insulin-resistant T2D littermates.

The gu<sup>t</sup> microbiome is of grea<sup>t</sup> interest in the T2D field and is well-known to be related to diet and other environmental conditions. Previous reports on mice and humans have documented alterations in the gu<sup>t</sup> microbiome of T2D individuals compared to healthy controls [8,24]. While changes in the ratio of the dominant phyla, *Firmicutes* and *Bacteroidetes*, have been previously linked with T2D, we found no differences in *Firmicutes* or *Bacteroidetes* composition or relative abundance among groups [24,25]. Instead, the most defining characteristic of the NGOB gu<sup>t</sup> microbiome fingerprint compared to both of the other groups was a change in nondominant phyla composition and abundance driven primarily by *Proteobacteria*. The elevations in *Proteobacteria* in NGOB mouse cecal matter are perplexing, as many reports indicate a higher disease incidence when *Proteobacteria* is elevated [26]. On the other hand, a richer overall microbial diversity has been associated with positive outcomes regarding glucose homeostasis [8,27].

A distinct endocrine profile was observed in the NGOB mouse compared to WT, and this profile was further altered in T2D. Elevated insulin and glucagon are hallmarks of β-cell stress and insulin resistance, and NGOB mice showed clear β-cell compensation, with a robust increase in IGR that was lost in the T2D cohort. The gut-islet relationship, of grea<sup>t</sup> interest in the T2D field, is thought to be modulated in part by the microbial composition of the GI tract. The incretin hormones, GLP-1 and GIP, are primarily secreted from gu<sup>t</sup> enteroendocrine cells and act on the β-cell to promote insulin secretion. GLP-1 and GIP were both elevated in NGOB mice compared to WT. Surprisingly, they were even further elevated in plasma from T2D mice, even though loss of the incretin response is a known pathophysiological defect in T2D [27]. However, gu<sup>t</sup> microbiota alterations have been shown to influence incretin sensitivity in obesity, insulin resistance, and T2D [28,29]; therefore, elevated GLP-1 and GIP levels may be reflective of a compensatory response that ultimately fails to promote β-cell function.

Changes in gu<sup>t</sup> microbiota are also associated with alterations in circulating factors associated with T2D pathology [28,30,31]. This relationship with microbiota dysbiosis is further supported by elevated levels of resistin and PAI-1, adipokines associated with obesity and T2D [30,32,33]. Resistin and PAI-1 production and secretion are stimulated during innate immune response and, therefore, can reflect alterations in gu<sup>t</sup> endothelial cell integrity that impact the systemic metabolic profile of the organism [30,33–36]. In obesity, bacterial byproducts, such as lipo-polysaccharides (LPS), enter circulation as the endothelial

wall of the intestine is degraded: a common consequence of obesity-associated intestinal inflammation [37,38]. LPS induces the release of resistin and PAI-1 from adipocytes, and LPS has also been shown to impair β-cell function via downregulation of the mature β-cell gene expression profile [39,40]. Therefore, it is possible that resistin and PAI-1 may serve as markers for β-cell dysfunction of T2D, independent of any direct influence on the β-cell themselves. More work would be necessary to tease apart this relationship.

The effects of elevated plasma AA and its precursor linoleic acid have been previously implicated in diabetes pathology and correlate with elevated HBA1c levels and hyperglycemia in human subjects [41]. In this work, we confirmed that elevated levels in circulating AA (or its isomers) correlate directly with elevated PGE metabolite levels. However, it has long been known that PGE2 is synthesized and excreted from pancreatic β-cells themselves and that this synthesis is a factor of both substrate availability and the expression of key synthetic enzymes, many of which are upregulated by the pro-inflammatory cytokine IL-1β [3,4,20–22]. In this work, we confirmed that this direct IL-1β-associated effect is of biological consequence, as the PGE2 analog, sulprostone, only influences the GSIS response of islets from T2D BTBROb mice, which already exhibits a prominent secretion defect compared to their nondiabetic controls. However, even though essential PUFAs must be obtained from the diet, diet alone cannot explain the altered ratios of omega-6 and omega-3 PUFAs downstream of linoleic acid and linolenic acid (including the most important to our work, AA and EPA) as equal numbers of T2D mice had also been fed an identical diet. Still, when considered in the context of our previous work demonstrating that BTBROb islets cultured in EPA-enriched media or nonobese diabetic mice fed an EPA-enriched diet significantly improved ex vivo or in vivo β-cell function, respectively [3], our findings certainly support the continued study of PUFA-based dietary interventions for T2D prevention or therapy.

In summary, we found that a host of systemic metabolic alterations are associated with the T2D phenotype of a strong genetic model of the disease and that, at least for the AA-to-PGE2 pathway, have directly implicated a specific class of metabolites in β-cell dysfunction of the disease. These findings have strong implications in the managemen<sup>t</sup> of T2D, as many foods in the Western diet are enriched with AA. Therefore, pharmacological strategies to reduce gut/adipose inflammation in combination with a diet focused on limiting circulating AA may help to facilitate better β-cell function, either alone or in concert with T2D medications, promoting more effective blood glucose control even in the face of chronic obesity and insulin resistance.

#### **4. Materials and Methods**

#### *4.1. Animal Care and Husbandry*

BTBR mice heterozygous for the *Leptino<sup>b</sup>* mutation were purchased from The Jackson Laboratory and bred in house at the UW-Madison Breeding Core Facility to generate homozygous OB mice or wild-type (WT) controls. Experimental mice were singly housed in temperature- and humidity-controlled environments and maintained on a 12:12 h day/light cycle with free access to acidified water (InnoVive, San Diego, CA, USA) and one of two standard mouse chows of nearly identical macronutrient composition and energy density. A direct comparison of the nutrient content and ingredients of the Teklad global soy protein-free extruded 2920X diet (Envigo, Indianapolis, IN, USA) or the Rodent Laboratory Chow 5001 diet (Purina, Neenah, WI, USA) has been previously published [7]. Both wild-type control mice (*n* = 9; 3-Purina and 6-teklad) and OB mice (*n* = 17; 11-Purina and 6-teklad) were given either Purina or Teklad diets. At 10 weeks of age, OB mice that developed hyperglycemia were grouped as T2D/HGOB (*n* = 12). A subset of OB mice was able to maintain normoglycemia, which was grouped as NGOB (*n* = 5). All protocols were approved by the Institutional Animal Care and Use Committees of the University of Wisconsin-Madison and by the William S. Middleton Memorial Veterans Hospital, which are both accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (Project ID: G005181-R01-A03). All animals were treated in accordance with the standards set forth by the National Institutes of Health Office of Animal Care and Use.

#### *4.2. Blood Glucose Measurements and Insulin Tolerance Tests*

Insulin tolerance tests (ITTs) were performed essentially as previously described [42]. Briefly, mice were fasted for 4–6 h and 0.75 U/kg short-acting recombinant human insulin (Humulin® R; Eli Lilly, Indianapolis, IN, USA) was injected intraperitoneally. Blood glucose readings were taken by tail nick with an AlphaTRACK glucometer (Zoetis, Parsippany-Troy Hills, NJ, USA) at baseline and the indicated times after insulin administration. Percent blood glucose change from baseline was determined by normalizing the blood glucose reading at each timepoint to that at baseline for each mouse independently. The blood glucose percent change from baseline was averaged within genotypes at each time point, giving the means ± SEM.

#### *4.3. Ex Vivo Islet Glucose Stimulated Insulin Secretion Assays*

Islets were isolated from experimental mice at 10 weeks of age utilizing a collagenase digestion protocol as previously described [43]. Glucose-stimulated insulin secretion (GSIS) assays were performed as previously described [44] after the indicated treatments. Briefly, islets were washed and preincubated for 45 min in a 0.5% Bovine Serum Albumin (BSA) Krebs Ringer Bicarbonate Buffer (KRBB) containing 1.7 mM glucose. Islets were then incubated for an additional 45 min in either low glucose (1.7mM) or stimulatory glucose (16.7 mM) ± the EP3-selective agonist sulprostone (10 nM). Secretion media was collected, and islets were lysed in equal volume to determine insulin content. Insulin was measured via ELISA as previously described [44].

#### *4.4. Quantitative PCR for Gene Expression Analyses*

RNA isolation, cDNA synthesis, and quantitative PCR using SYBR Green reagen<sup>t</sup> (Bio-Rad) to determine relative mRNA abundance among groups were all performed as previously described [45]. Data were normalized to that of β-actin to calculate ΔCT values. Primer sequences are available upon request.

#### *4.5. Terminal Blood Collection and Plasma Hormone/Metabolite Assays*

Terminal blood collection for plasma samples was performed by retroorbital puncture under anesthesia. Briefly, mice were anesthetized using 2,2,2-tribromoethanol (Sigma-Aldrich, St. Louis, MO, USA; #T48402). Blood was collected retro-orbitally using a heparincoated glass capillary tube and mixed with 5 μM EDTA, 10 nM DDP-4 inhibitor, and 20 nM aprotinin. Plasma was isolated via centrifugation and stored at −80 ◦C until needed. Plasma hormones were measured utilizing the Bio-Plex ProTM Diabetes Assay (Bio-Rad Laboratories, Hercules, CA, USA) following the manufacturer's protocol. The plasma prostaglandin E metabolite was measured utilizing a PGEM ELISA (Cayman Chemical Company, Ann Arbor, MI, USA; no. 514531) following the manufacturer's protocol as previously described [42].

#### *4.6. Gut Microbial DNA Preparation, Sequencing, and Analysis*

Microbial 16s sequencing of cecal matter was performed as previously described [46]. Briefly, 20–50 mg of the cecal matter was collected from WT, NGOB, and T2D mice, and genomic DNA was extracted and cleaned by using the Macherey-Nagel PCR Clean-up kit according to manufacturer's protocol (ThermoFisher Scientific, Waltham, MA, USA). Purified genomic DNA was submitted to the University of Wisconsin-Madison Biotechnology Center. DNA concentration was verified fluorometrically, and samples were prepared and amplified according to Illumina's 16s Metagenomic Sequencing Library Preparation Protocol with few modifications as described before [46]. Following PCR, reactions were cleaned using 0.7× volume of AxyPrep Mag PCR clean-up beads (Corning, Corning, NY, USA). Quality and quantity of the finished libraries were assessed using an Agilent DNA

1000 kit (Agilent Technologies, Santa Clara, CA, USA) and Qubit® dsDNA HS Assay Kit (ThermoFisher Scientific), respectively. In an equimolar fashion, libraries were pooled and appropriately diluted prior to sequencing. After sequencing, images were analyzed using the standard Illumina Pipeline, version 1.8.2. OTU assignments (Illumina, Inc., San Diego, CA, USA). Diversity plots were created using the QIIME (Ver. 1.9.1) analysis pipeline [46,47].

For the pilot experiment using fecal pellets from NGOB and pre-T2D mice, cryopreserved fecal pellets were shipped to Argonne National Labs where sequencing and data analysis were performed according to their standard protocols. Sequencing from the pilot experiment only went to the order level and not the species level.

#### *4.7. FIE-FTCIR MS for Unbiased Plasma Metabolomics*

Plasma samples were collected as described in Section 4.5. Samples were prepared using a methanol extraction procedure, and unbiased FIE-FTICR MS experiments were performed as previously described [7]. Statistical analysis was performed using MetaboScape 4.0 (Bruker, Billerica, MA, USA) and the online software MetaboAnalyst [48]. Putative metabolites were further annotated by METLIN with a 2-ppm mass error cutoff. Details about data analysis were described previously [7].

## *4.8. Statistical Analyses*

Data from all experiments excluding that previously described for microbiome and metabolomics analyses were analyzed using GraphPad Prism v.9 (GraphPad Software Inc., San Diego, CA, USA). Data were analyzed by *t*-test, one-way ANOVA, or two-way ANOVA as described in the figure legends. *p* < 0.05 was considered statistically significant.

## *4.9. Data Availability*

All data contained within this manuscript are available upon reasonable request of the corresponding author. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE [49] partner repository with the dataset identifier PXD022624.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2218-198 9/11/1/58/s1, Figure S1: Workflow of the FIE-FTICR MS-based platform for metabolomics. Table S1: Diet Composition, Table S2: Pilot 16s rRNA sequencing, Table S3: 16S rRNA analysis, Table S4: FIE-FTICR MS Mass List, Table S5: WT vs NGOB metabolites, Table S6: NGOB vs T2D metabolites.

**Author Contributions:** Conceptualization, M.D.S., Y.Z., J.C.N., B.T.L., D.W.L., Y.G. and M.E.K.; data curation, Y.Z., Y.G. and M.E.K.; formal analysis, M.D.S., Y.Z., N.E.R., C.P., I.M.O., R.J.F., J.C.N., E.G., A.R., H.K.S. and M.E.K.; funding acquisition, I.M.O., E.D.C., D.B.D., B.T.L., A.R.B., D.W.L., Y.G. and M.E.K.; investigation, M.D.S., Y.Z., N.E.R., I.M.O., R.J.F., J.C.N., E.G., A.R., H.K.S., M.H.F. and M.E.K.; methodology, M.D.S., Y.Z., N.E.R., D.W.L., Y.G. and M.E.K.; project administration, E.D.C., D.B.D. and M.E.K.; supervision, M.D.S., B.T.L., A.R.B., D.W.L., Y.G. and M.E.K.; visualization, M.D.S., Y.Z., N.E.R. and C.P.; writing—original draft, M.D.S., N.E.R., C.P. and J.C.N.; writing—review and editing, M.D.S., Y.Z., C.P., R.J.F., A.R., D.B.D., B.T.L., D.W.L., Y.G. and M.E.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported in part by Merit Review Awards I01 BX003700 (to M.E.K.), I01 BX004031 (to D.W.L.), and I01 BX003382 (to B.T.L.) from the United States (U.S.) Department of Veterans Affairs Biomedical Laboratory Research and Development (BLR&D) Service. This work was also supported in part by National Institutes of Health (NIH) grants K01 DK080845 (to M.E.K.), R01 DK102598 (to M.E.K.), R01 AG056771 (to D.W.L.), UL1 TR002373 (to A.R.B.), R01 DK104927 (to B.T.L.), R01 DK111848 (to B.T.L.), U01 DK127378 (to B.T.L.), P30 DK020595 (to B.T.L.), and P30 CA014520 (to I.M.O.); American Diabetes Association Grant 1-14-BS-115 (to M.E.K.); and a UW2020 gran<sup>t</sup> from the Wisconsin Alumni Research Foundation (to M.E.K., E.D.C., and D.B.D). This research was conducted while D.W.L. was an AFAR Research Grant recipient from the American Federation for Aging Research. N.E.R. and J.C.N. were supported in part by training grants from the UW Institute on Aging (NIA T32 AG000213).

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Animal Care and Use Committees of the University of Wiscon-sin-Madison and by the William S. Middleton Memorial Veterans Hospital (protocol code: G005181-R01-A03 and date of approval: 6/13/2018).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** All data are contained within the article or supplementary material.

**Conflicts of Interest:** M.D.S., Y.Z., N.E.R., C.P., I.M.O., R.J.F., E.G., H.K.S., M.H.F., E.D.C., D.B.D., B.T.L., A.R.B., Y.G. and M.E.K. declare that they have no conflicts of interest with the contents of this article. J.C.N. is now employed by Novo Nordisk (800 Scudders Mill Rd, Plainsboro, NJ 08536). This work was completed in full during his predoctoral training with M.E.K. and is not related to his current position. D.W.L. received funding from and is a scientific advisory board member of Aeovian Pharmaceuticals, which seeks to develop novel, selective mTOR inhibitors for the treatment of various diseases, including diabetes. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health, the U.S. Department of Veterans Affairs, or the United States Government. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.
