All procedures were in accordance with the Canadian Council of Animal Care Standards and were approved by the Animal Care Committee of the University of Toronto with the study protocol number 20011526. The mice used for all experiments were male and 12 to 14 weeks old. Ear clips were obtained for genotyping purposes. Genotyping was conducted using the Wisent Genotyping Kit, and DNA was extracted according to the manufacturer’s protocols. The PCR primers and representative micrographs of genotyping results are shown in Supplementary Figure 1 . Myeloid-specific Nod1-knockout (KO) mice were generated by crossing Nod1 floxed mice on a C57BL/6 background obtained from Dr. Dana J. Philpott (25) with mice expressing lysozyme 2 (Lyz2) Cre (Jackson Laboratory #004781) also on a C57BL/6 background. Nod1 floxed (Flox) and Lyz2Cre (Cre) mice were used as controls. Both controls were either littermates or mice from the same breeding colony. For all experiments, heterozygous Lyz2Cre mice were used.

All animals were housed in the University of Toronto’s Department of Comparative Medicine (DCM) facility, exposed to a 12-hour light-dark cycle, and had free access to water and standard rodent chow, which contained 58% carbohydrate, 24% protein, and 18% fat by calories (Harlan Tekland #2018).

Mice were anesthetized using isoflurane (3%–3.5% for induction, 1.5%–2% for maintenance). Analgesics (Buprenorphine SR, 1 mg/kg) were administered pre-operatively to maintain therapeutic serum levels during recovery. A cannula made from 10–11 cm of polyethylene tubing (PE-10) and 1.5 cm of silastic tubing was inserted into the jugular vein, reaching the right atrium. The silastic tubing was tapered for easier insertion. The cannula was flushed with heparinized saline (40 units/mL) to prevent clotting, exteriorized at the back of the neck, and secured with tape. The mice were then placed in individual cages and allowed to recover for 3–5 days with free access to food and water. After surgery, the mice were continuously monitored until they regained consciousness and were fully ambulatory. Post-operative monitoring was conducted twice daily by both the surgeon and the DCM veterinarian to check for signs of distress or infection. A detailed log was maintained to document animal behavior, incision sites, and cannula condition.

Our model for inducing β-cell lipotoxicity through prolonged elevation of plasma palmitate levels in vivo involves infusing mice with ethyl palmitate (PAL) at a rate of 0.12 μmol/min for 48 hours (26). Direct infusion of unbound palmitate is highly toxic due to its detergent effects; however, adding an ethyl group to the carboxylic end of palmitate neutralizes this toxicity. Ethylpalmitate is hydrolyzed into ethanol and palmitate by ethyl esterases in mouse plasma (26), therefore, ethanol served as the vehicle control (VEH). Mice were briefly anesthetized with 1.5% isoflurane to connect the inserted catheter to our infusion setup, which allowed them full mobility in their cages. We used syringe pumps (model # PHD2000) from Harvard Apparatus Inc. for the infusions. Throughout the 48-hour infusion, the mice had access to food and water ad libitum. After the infusion period, a hyperglycemic clamp was conducted to assess in vivo β-cell function.

The mice were fasted for 5 hours prior to the start of the clamp, after which the mice were moved into the restrainer for the duration of the experiment. Blood glucose was elevated and maintained at ~20mM by an intravenous glucose infusion. A 37.5% glucose solution was used as the infusate to achieve the hyperglycemic target while minimizing the infusion volume, and it was co-infused with ethylpalmitate or vehicle using a 2-way connector. Blood glucose was measured every 10 minutes during the 2-hour clamp via tail vein blood sampling, and the glucose infusion rate (Ginf) was adjusted accordingly. A higher Ginf represents better glucose tolerance, which is determined by increased ß-cell secretion and decreased insulin clearance to compensate for palmitate-induced insulin resistance. At time points 0, 100, and 120 minutes, 50–100μL of plasma was collected from the tail vein to measure insulin and C-peptide levels.

Mouse plasma glucose concentrations were measured using the HemoCue Glucose 201 Analyzer.

Plasma obtained at times 0, 100, and 120 min of the clamp was used to determine levels of insulin and C-peptide. These samples were analyzed by enzyme-linked immunosorbent assays that are sandwich-type immunoassays obtained from ALPCO (Catalog #80-INSMS-E01, E10 for insulin, Catalog #80-CPTMS-E01 for C-peptide). Standard solutions were run in duplicate, while samples were run in singlicate. Both inter- and intra-assay variations are below 5% for the insulin assay and below 7% for the C-peptide assay. The insulin assay had a sensitivity of 10.34 pM with a detection range of 32.41 to 61,189.66 pM, while the C-peptide assay had a sensitivity of 7.6 pM with a detection range of 60 to 3,000 pM. Basal plasma (0 min) was used for free fatty acid (FFA) determination. The samples were analyzed by an enzymatic colorimetric assay obtained from FUJIFILM Wako Diagnostics (Catalog #999-34691, #995-34791, #991-34891, #993-35191). Standard solutions and samples were run in duplicate as per the manufacturer’s protocol. The enzymatic reaction generated a purple pigment, quantified at 550 nm using a microplate reader. FFA concentrations were determined using a standard curve, with a detection range of 0.07–2.50 mEq/L and a sensitivity of 0.04 mEq/L.

Bone marrow-derived macrophages (BMDMs) were isolated from the femurs and tibiae of untreated Nod1flox, Lyz2Cre, and Nod1-KO mice. A standard Fc receptor-mediated phagocytosis assay was performed as described previously (32, 33). Medium-sized (5 µm) microspheres were opsonized with human immunoglobulin G (IgG) and incubated with differentiated macrophages for 30 minutes at 37°C. After incubation, cells were fixed with paraformaldehyde (PFA) and labeled with Alexa647-conjugated anti-human IgG to stain external (non-phagocytosed) microspheres. The cells were then washed, permeabilized with 0.1% Triton X-100 for 5 minutes, and labeled with Alexa488-conjugated anti-human IgG to identify internalized microspheres. DAPI staining was used for nuclear visualization, and images were acquired using a spinning disk confocal microscope (4–5 fields per condition). Internalized microspheres appeared green, while external microspheres were purple. Quantification was performed by counting the number of internalized microspheres in 50 cells per condition.

Data are represented as mean ± SEM. Given a coefficient of variation of 32% in our primary endpoint, i.e., Ginf, and an effect size of 68%, an n ≥ 4/group gave us a power ≥ 0.8 with 2α=0.05. Statistical differences were evaluated with one-way ANOVA and the Holm–Sidak post hoc test for multiple comparisons. P<0.05 was considered significant. The analysis was performed using GraphPad Prism 9.

Following 48 hours of infusion, plasma FFAs were significantly increased in PAL-infused mice compared to the respective VEH-treated groups ( Figure 1A ). Plasma glucose levels increased to ~20mM in all groups during the last 20 min of the clamp ( Figure 1B ). The Ginf necessary to achieve and maintain the hyperglycemic target is a measure of glucose tolerance. As expected, the flox control PAL mice but not the Nod1-KO mice had reduced Ginf compared to VEH ( Figure 1C ); however, Ginf was not reduced with PAL in the Cre control group. This suggests that myeloid Cre expression alone was enough to induce an anti-inflammatory phenotype, thus protecting against palmitate-induced glucose intolerance.

During the clamp, plasma insulin and C-peptide levels were lower in the VEH-treated Cre and KO mice compared to the respective PAL-treated groups ( Figures 2A, B ), indicating reduced absolute insulin secretion. The insulin clearance index (C-peptide to insulin molar ratio) was significantly greater in the VEH-infused Cre mice during the basal period and tended to be greater during the steady-state clamp phase, compared to the PAL-treated Cre mice ( Figure 2C ). In the KO group, the trend was reversed, with a significantly greater insulin clearance index during the clamp phase and a tendency to increase during the basal period.

The SI (Ginf/plasma insulin) decreased in all groups infused with PAL compared to VEH ( Figure 3A ). Although these SI results align with expectations, it is noteworthy that VEH-treated Cre mice showed higher SI compared to all other groups except the VEH-treated KO, suggesting that myeloid Cre expression in the absence of fat had an anti-inflammatory insulin-sensitizing effect. This insulin-sensitizing effect corresponded to significantly lower insulin secretion and increased insulin clearance (as described in Figure 2 ), suggesting compensatory changes in insulin secretion and clearance to maintain glucose homeostasis. In the presence of fat, however, insulin sensitivity was impaired, insulin secretion correspondingly increased, and the effect of myeloid Cre expression to protect against ß-cell lipotoxicity (anti-inflammatory effect on ß-cell function) manifested. This is evidenced by the DI, which, as expected, decreased in the PAL-treated flox control mice compared to VEH, while this decrease was prevented by myeloid Lyz2Cre expression alone or in combination with Nod1 deficiency ( Figure 3B ).

A phagocytosis assay with fluorescent microspheres was used to assess the effect of Lyz2Cre expression on the phagocytic ability of BMDMs. The results demonstrated a significant reduction in the phagocytic activity of BMDMs expressing Cre compared to their Cre-negative Nod1 floxed counterparts ( Figure 4 ). Specifically, Lyz2Cre+ macrophages internalized significantly fewer microspheres, as evidenced by a reduced number of green-labeled (internalized) microspheres per cell ( Supplementary Figure 2A ). Although the phagocytic activity of BMDMs of Nod1-KO mice (also Cre-positive) was not significantly different from that of either Nod1 floxed or wild-type Lyz2Cre mice, by combining the two Cre-positive groups, we still observed a significant difference compared to the Cre-negative mice ( Supplementary Figure 2B ). These findings suggest that Lyz2Cre expression impairs macrophage function.