LNFPIII was synthesized by Dr. Peng George Wang (Georgia State University, Atlanta, GA) (23, 29). LNnT was synthesized by Neose Technologies, Inc. LNFPIII (MW: 853.877 g/mol) and LNnT (MW: 707.60 g/mol) were sent to Dr. Thomas Norberg (Uppsala University, Uppsala, Sweden) for conjugation to aminodextran (DEX, 40 kDa average Mw, from Invitrogen, prod # D1861) using APD linker-spacers. On average, conjugates had 10-12 LNFPIII or LNnT monomers per 40 kDa dextran carrier. LNFPIII accounts for ~17-20% of the molecular weight of the P3DEX conjugate. LNnT accounts for ~15-17% of the molecular weight of the NTDEX conjugate.

6 to 8-week old male C57BL/6 mice were purchased from The Jackson Laboratory and maintained on a 12h light/dark cycle in the University of Georgia’s AALAC-accredited College of Veterinary Medicine Animal Resources Facility with food and water available ad libitum. After one week of acclimation, mice were housed n=3/cage and placed on a high-fat diet (HFD: Bio-Serv Cat. No. F3282) as described in Bhargava et al. (28). Given that the HFD is subject to spoilage, food was replenished 3 times per week. After 6 weeks of HFD, mice were divided into 3 cohorts and injected twice per week for 8 weeks via the intraperitoneal route with 25µg of 40 kDa dextran (DEX), LNFPIII conjugated to 40 kDa dextran (P3DEX), or LNnT conjugated to 40 kDa dextran (NTDEX). An experimental timeline is shown in Figure 1A . DEX, P3DEX, and NTDEX were dissolved in 0.9% NaCl prior to injection in a volume of 200µL. Experiments were performed in two independent mouse cohorts (n=6-8/group). Metabolic studies (GTT/ITT) were performed after 4 weeks of treatment with a 2-week rest period between in vivo assays. We performed a glucose tolerance test (GTT) at W10 of the experiment and an insulin tolerance test (ITT) at W12 (described below). Body weights were measured once/week and fasting blood glucose levels were measured prior to start of HFD (W0), pre-treatment (W6), and post-treatment (W14). Mice were euthanized at the end of W14 via CO₂ asphyxiation followed by cervical dislocation after a 6h fast. Organs and serum samples were collected and stored at -80°C until use.

Mice were placed in a clean plastic container on a tared compact scale (Ohaus, Cat. No. CS200) to acquire body weight measurements (in grams). Body weights were recorded each week on Monday afternoon to minimize daily fluctuations. Body weights were recorded to the nearest hundredth. At sacrifice, organs (i.e. heart, spleen, liver, kidney, AT) were placed on clean Fisherbrand™ Polystyrene Antistatic Weighing Dishes (Fisher Scientific, Cat. No. 08-732-112) on a tared analytical scale (Sartorius, Cat. No. TE313S) for measurement. Organ weights were recorded to the nearest thousandth.

A glucose tolerance test (GTT) were performed after 4 weeks of treatment (W10), according to Beguinot & Nigro with some modifications (24). Mice were fasted for 6h to obtain baseline blood measurements. Various fasting periods have been utilized in published studies involving GTTs and ITTs, but 6h fasts appear to be more physiological than longer 14-16h fasts. Previous studies have shown that prolonged fasting in mice results in increased insulin sensitivity, whereas fasting in humans leads to inhibition of insulin-stimulated glucose uptake (25, 26). Moreover, prolonged fasting in mice may lead to a state of starvation (30). To ensure translatable results, we utilized a 6h fasting period in our experiments. During GTTs, blood was acquired via tail snip and measured using a Bayer Contour^® Next blood glucose monitoring system (Bayer, Parsippany, NJ). Basal glucose measurements were taken before mice were administered 2g glucose/kg (Sigma, Cat. No. G8270) in 10µL/g body weight via oral gavage. Thereafter, blood glucose measurements were taken at 15, 30, 60, 90, 120, 150, and 180-minute timepoints.

An insulin tolerance test (ITT) was performed after 6 weeks of treatment (W12) to allow for a 2-week rest period between metabolic tests. The ITT was also performed according to Beguinot & Nigro with the aforementioned fasting modification (24). During the ITT, mice were fasted for 6h prior to baseline blood glucose measurements and intraperitoneal injection of 0.5IU/kg insulin (Humulin^® R U-100, Lilly, Cat. No. HI-210) diluted in 0.9% NaCl in an injection volume of 3.6µL/g body weight. Thereafter, blood glucose measurements were taken at 15, 30, 60, and 90-minute timepoints. If blood glucose levels dropped below 36mg/dL, mice were rescued via injection of 20% aqueous glucose solution (Sigma, Cat. No. G8270).

The Bio-Plex Pro Mouse Cytokine 23-Plex Assay (Bio-Rad, Cat. No. M60009RDPD) and the Bio-Plex Pro Mouse Diabetes 8-Plex Assay (Bio-Rad, Cat. No. 171F7001M) were performed in multi-plex, according to Bio-Rad’s Technical Note 5975. These two panels can be performed via multiplex without altering assay sensitivity, specificity, and accuracy. Serum samples collected at sacrifice after a 6h fast and frozen at -80°C were prepared according to manufacturer’s instructions. Adipokines, chemokines, and cytokines were detected and quantified using the Bio-Rad Bio-Plex^® 200 System, available at UGA’s Cytometry Shared Resource Laboratory.

WAT and liver samples were preserved in 10% formalin and processed, sectioned, and stained with hematoxylin & eosin (H&E) by the Comparative Pathology Lab in the College of Veterinary Medicine at the University of Georgia. Dr. Tamas Nagy, a board-certified pathologist, then assessed inflammation with WAT sections and lipid accumulation in the liver.

Liver function was assessed via measurement of alanine aminotransferase (AST) and aspartate aminotransferase (ALT). ALT was measured in serum using the Alanine Aminotransferase (ALT or SGPT) Activity Colorimeteric/Fluorometric Assay Kit (Biovision, Cat. No. K752), according to manufacturer’s instructions. AST was measured in serum using the Aspartate Aminotransferase (AST or SGOT) Activity Colorimetric Assay Kit (Biovision, Cat. No K753), according to manufacturer’s instructions. Absorbance was measured using the SPECTROstar Nano Microplate Reader (BMG Labtech).

Statistics were performed using One-Way or Two-Way ANOVA with Dunnett’s Multiple Comparisons Tests in GraphPad Prism 8. * or ** was used to indicate significance (*p<0.05 or **p<0.01). ^ was used to indicate a trend (^p<0.10). Values are presented as mean ± SEM. Experiments were repeated twice for a total n=12-14 per treatment group.

Male C57BL/6 mice were placed on HFD for 6 weeks (W0-6) prior to beginning treatment twice per week with 25ug DEX, P3DEX, or NTDEX (W7-14) via the intraperitoneal route. Body weight was monitored weekly throughout the duration of the experiment (W0-W14). An experimental timeline is shown in Figure 1A . Mice treated with P3DEX weighed less overall than mice treated with the DEX carrier control or NTDEX ( Figure 1B ). Compared to DEX, P3DEX treatment reduced total body weight. This reduction trended towards significance (Two-Way ANOVA with Dunnett’s Multiple Comparisons Test; ^<0.10) at W9, 12, and 13. While it appears that the DIO mice treated with P3DEX begin to show lessened body weight and reduced total weight gain at W5 and W6 of the experiment, the difference is neither trending, nor significant at these timepoints. In Figure 1B , statistical analysis performed at W5 between DEX and P3DEX resulted in a non-trending or non-significant p-value of 0.5482 (One-Way ANOVA with Dunnett’s Multiple Comparisons Test, *p<0.05, ^p<0.10). At W6, the p-value for the statistical test performed between DEX and P3DEX was equal to 0.2719 (One-Way ANOVA with Dunnett’s Multiple Comparisons Test, ^p<0.10, *p<0.05). Treatment with NTDEX did not cause a significant reduction in total body weight at any point of measurement. Furthermore, there was no significant difference in total body weight between the 3 groups prior to treatment. When normalizing the body weights and examining total weight gain over time, it is clear that P3DEX treatment led to a reduction in total weight gain when compared to DEX-treated animals ( Figure 1B ). This trended toward significance at W8, 9, 10, 11, and 12 (Two-Way ANOVA with Dunnett’s Multiple Comparisons Test; *<0.05, ^p<0.10), and became significant at W13 (Two-Way ANOVA with Dunnett’s Multiple Comparisons Test; *p<0.05, ^p<0.10). When we normalized the body weight in Figure 1C , the difference between DEX and P3DEX remained non-trending and non-significant at W5 and W6, with a p-value of 0.4307 and 0.1891, respectively (One-Way ANOVA with Dunnett’s Multiple Comparisons Test, *p<0.05, p<0.10). For these reasons, we are confident that the trending and significant differences we observe at later timepoints are divergent and a result of P3DEX intervention. Again, treatment with NTDEX did not cause a significant reduction in total weight gain at any point of measurement and nearly mirrors the DEX control.

At sacrifice (W14), we collected and weighed subcutaneous (scWAT), visceral (VAT), and brown (BAT) adipose tissue, as well as livers, hearts, spleens, and kidneys. Figure 1D demonstrates that we saw a trend towards reduction of scWAT in P3DEX-treated mice compared to those treated with DEX (One-Way ANOVA with Dunnett’s Multiple Comparisons Test; *p<0.05, ^p<0.10). We saw no effect of NTDEX on scWAT compared to DEX. In terms of vWAT, we noted a significant reduction in P3DEX-treated mice compared to those treated with DEX ( Figure 1E ). We observed no differences in BAT between groups, but this is a smaller AT depot ( Figure 1F ). We did not observe a significant difference in liver, heart, spleen, or kidney weights ( Figures 1G–J ).

We measured 6h-fasting blood glucose levels prior to HFD (W0), prior to treatment (W6), and prior to sacrifice (W14). Prior to HFD (W0), the average fasting blood glucose for all male mice was within normal range (88.5-154.9 mg/dL) (31). This was elevated prior to treatment (W6), demonstrating that the HFD induced metabolic dysfunction. Prior to sacrifice and post-treatment, mice treated with P3DEX had significantly lower fasting blood glucose levels ( Figure 2A ) than those treated with DEX. NTDEX treatment did not result in reductions in fasting blood glucose levels compared to those treated with DEX. When administered an oral gavage of 2g glucose/kg body weight during a GTT, mice treated with P3DEX exhibited a less dramatic increase in blood glucose levels at the 15-minute timepoint compared to those treated with DEX or NTDEX. P3DEX-treated animals were able to return to basal blood glucose levels more quickly than DEX- or NTDEX-treated mice. Mice treated with P3DEX had significantly lower blood glucose levels throughout the duration of the GTT and this was significant at 30, 60, 90, 120, 150, and 180 minutes ( Figure 2B ). When given an intraperitoneal injection of 0.5IU/kg body weight insulin, mice treated with P3DEX exhibited a greater decrease in blood glucose levels overall compared to those treated with DEX or NTDEX. Mice treated with P3DEX had significantly lower blood glucose levels throughout the duration of the ITT and this trended towards significance at 15 minutes ( Figure 2C ). Blood glucose levels of P3DEX-treated mice was significantly lower than DEX-treated mice at the 30-minute timepoint ( Figure 2C ). Thereafter, blood glucose levels between groups appeared to stabilize. This shows that DIO mice treated with DEX or NTDEX were less responsive than mice treated with P3DEX when injected with a bolus of insulin, suggesting the presence of IR in the DEX and NTDEX groups.

Individuals diagnosed with MS are at higher risk of infection and illness, suggesting that the constant presence of low-grade inflammation exhausts and dampens the immune response to other pathogens or insults (32, 33). In order to maintain function, progenitor cells depend on nutrients and hematopoietic signals to proliferate and differentiate into diverse white blood cells (WBC) populations (34, 35). IL-3, for example, promotes proliferation of hematopoietic cells (34, 36). G-CSF and GM-CSF are then responsible for differentiation of cells into specific lineages (i.e. granulocytes and monocytes) and subsequent activation (34, 37). High levels of IL-3, G-CSF, or GM-CSF can lead to increased populations of WBCs, which can exacerbate inflammation and damage normal tissues. High WBC counts have been associated with parameters related to inflammation and well-documented in obese subjects (38–43).

We measured IL-3, C-CSF, and GM-CSF in serum samples collected from fasting DIO mice treated with DEX, P3DEX, or NTDEX. Serum from DIO mice treated with P3DEX or NTDEX trended towards decreased IL-3 levels compared to the DEX control ( Figure 3A ). G-CSF and GM-CSF were also slightly reduced, with the reduction in G-CSF becoming significant following NTDEX treatment ( Figure 3A ). These results suggest that both P3DEX and NTDEX have potential to dampen the excessive immune response associated with the obese state.

WBCs depend on chemoattractant signals for migration to injured sites. In the obese state, expansion of WAT enhances chemokine release and attracts WBCs to AT depots (44). This worsens metabolic inflammation and associated comorbidities.

We measured eotaxin/CCL11, KC/CXCL1, MCP-1/CCL2, MIP-1α/CCL3, MIP-1β/CCL4, and RANTES in serum samples collected from fasting DIO mice treated with DEX, P3DEX, or NTDEX. We saw no differences in circulating levels of eotaxin/CCL11, a chemokine known to attract eosinophils ( Figure 3B ). While circulating eotaxin/CCL11 is known to increase in the obese state, neither P3DEX, nor NTDEX altered these levels (45, 46). This makes sense as we have no evidence that either conjugate alters eosinophil function. KC/CXCL1 is a neutrophil chemoattractant that increases in patients with T2DM, as well as in diabetic-prone db/db mice with evidence of impact on pancreatic islet function (47–49). P3DEX treatment resulted in a significant decrease in serum KC/CXCL1 compared to the DEX control ( Figure 3B ). NTDEX treatment did not reduce serum KC/CXCL1, suggesting mechanistic differences between NTDEX and P3DEX conjugates. It is well-reported in the literature that P3DEX and NTDEX act on macrophages (50–54). MCP-1/CCL2 is instrumental for macrophage recruitment to AT depots and circulating concentrations are increased in the obese state (55). Furthermore, mice deficient in MCP-1 signaling exhibit lessened macrophage infiltration and inflammation in AT depots (56, 57). P3DEX treatment resulted in a significant decrease in serum MCP-1/CCL2 compared to the DEX control ( Figure 3B ). Although NTDEX also appeared to decrease circulating MCP-1/CCL2 levels, this decrease was not significant ( Figure 3B ). Neither P3DEX, nor NTDEX had an effect on MIP-1α/CCL3 or MIP-1β/CCL4 ( Figure 3B ). MIP-1α and MIP-1β are also elevated in genetic (ob/ob, db/db) and DIO mice, as well as obese humans (19, 58, 59). While MIP-1α and MIP-1β both increase in the obese state, there is evidence that they do not alter macrophage infiltration in AT. RANTES is responsible for T cell recruitment and is also increased in the obese state (58, 59). We noted a significant reduction in serum RANTES following P3DEX treatment ( Figure 3B ). NTDEX also appeared to reduce serum concentration of RANTES, but this was not significant ( Figure 3B ). Overall, these results demonstrate that P3DEX is more effective at decreasing circulating chemoattractant signals than NTDEX.

Given their pleiotropic nature, cytokines are difficult to characterize in the context of DIO. We measured numerous innate and adaptive cytokines in serum samples collected from fasting DIO mice treated with DEX, P3DEX, or NTDEX. Once macrophages are recruited to AT depots, IFNγ secretion helps these cells maintain classical (M1) activation (60). Neither P3DEX, nor NTDEX, had a significant effect on circulating concentrations of IFNγ ( Figure 4 ). P3DEX treatment led to a slight, nonsignificant, reduction in IL-1α ( Figure 4 ). IL-1β was unaffected by P3DEX or NTDEX treatment ( Figure 4 ). Both IL-1α and IL-1β are elevated in obese individuals with reductions documented following weight loss (61–63). IL-2, involved in T cell proliferation and activation, was lowered, but nonsignificant, in response to P3DEX and NTDEX ( Figure 4 ). IL-4 trended towards reduction in P3DEX-treated mice and was significant for those treated with NTDEX ( Figure 4 ). This reduction was unanticipated, given that both P3DEX and NTDEX promote T_H2 responses and alternative activation (M2) of macrophages (51–54). P3DEX and NTDEX treatment also led to slight reductions in IL-5, another T_H2 cytokine, which tends to help maintain homeostasis in WAT depots ( Figure 4 ) (64). P3DEX did not impact circulating IL-6 levels, but NTDEX eliminated IL-6 compared to the DEX control ( Figure 4 ). IL-9, a T_H2 cytokine, trended toward decreased circulating levels in DIO mice treated with NTDEX ( Figure 4 ) (65). Neither P3DEX, nor NTDEX treatment, had a significant effect on circulating IL-10 levels ( Figure 4 ). While P3DEX and NTDEX did not impact levels of IL-12(p40), P3DEX treatment led to a significant reduction in IL-12(p70) ( Figure 4 ). This reduction in the bioactive form of IL-12 suggests quelling of the T_H1 immune response, which might contribute to P3DEX’s therapeutic effects (66). We observed no differences in serum IL-13, another T_H2 cytokine that aids in overcoming insulin resistance ( Figure 4 ) (67). IL-17a increased in response to P3DEX and NTDEX treatment, but this increase was not significant ( Figure 4 ). Last, we observed a slight, but nonsignificant, reduction in TNFα in response to P3DEX and NTDEX treatment ( Figure 4 ). Similar to other markers, circulating TNFα is elevated in obese individuals (17). Moreover, neutralization of TNFα led to improved insulin responses (15). Overall, P3DEX treatment seems to decrease circulating cytokine levels, both inflammatory and TH₂-associated. While these cytokine responses are ambiguous when considered on their own, they are indispensable for delineating how P3DEX treatment reduces body weight and improves glucose homeostasis in vivo.

Measurement of AT-specific markers (adipokines), rather than pleiotropic cytokines, might provide more insight on the therapeutic or non-therapeutic effects of P3DEX and NTDEX in the context of DIO. Ghrelin, for instance, is an orexigenic adipokine that stimulates food intake (68). Circulating levels of ghrelin are lower in the obese state and higher in those that are lean (69–71). Ghrelin also quells secretion of markers related to inflammation and inhibits NF-κB signaling (72, 73). Although we were unable to monitor food intake in our studies, we found that P3DEX treatment led to a slight, but nonsignificant, increase in ghrelin ( Figure 5A ). P3DEX treatment also led to a significant decrease in leptin, a hormone generated in AT in proportion to fat content ( Figure 5A ) (74). Circulating leptin concentrations coincide with reductions in AT (75, 76). Decreases in leptin also coincide with increases in orexigenic peptides (i.e. ghrelin), which might explain the slight increase shown in Figure 5A (77, 78). NTDEX treatment led to a trending reduction in circulating levels of resistin, while P3DEX treatment led to a significant decrease ( Figure 5A ). P3DEX treatment seems to have beneficial effects on specific adipokines (i.e. leptin and resistin), which corresponds with its effects on adipose tissue and glucose homeostasis.

During the incretin effect, glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide 1 (GLP-1) are released from the gut to stimulate insulin secretion in response to food intake (79). P3DEX and NTDEX reduced circulating levels of GIP, with the reduction induced via NTDEX being significant ( Figure 5B ). In pancreatic islets, decreases in GIP correspond with decreases in insulin and glucagon secretion (80). Both P3DEX and NTDEX did not have an effect on GLP-1, another incretin hormone ( Figure 5B ). P3DEX treatment led to a slight reduction in glucagon and NTDEX treatment led to a significant reduction ( Figure 5B ). NTDEX treatment led to a significant reduction in GIP, which aligns with the corresponding decrease in glucagon. We observed a trending increase in the ratio of insulin to glucagon in response to NTDEX treatment, an indicator of excess nutrient load ( Figure 5B ) (81). We conclude here that P3DEX treatment has beneficial effects on specific adipokines (i.e. ghrelin, leptin, and resistin), while NTDEX might impact incretin hormones and postprandial insulin release. These findings further support that these two HMO conjugates differ in terms of mechanism and therapeutic potential.

We report various in vivo changes in response to both P3DEX and NTDEX. Upon examination of vWAT tissue, we observed what appeared to decreased immune cell infiltrate and crown-like structures in DIO mice treated with P3DEX compared to DEX and NTDEX ( Figures 5C–E ). DIO mice treated with NTDEX exhibited greater inflammation within vWAT even when compared to the DEX control ( Figure 5E ). In the liver, we observed significant lipid accumulation in DIO mice treated with the DEX control ( Figure 6A ). In contrast, we observed reduced lipid accumulation in the livers of DIO mice treated with P3DEX ( Figure 6B ). DIO mice treated with NTDEX showed greater lipid accumulation than the DEX control, suggesting a worsening of MS ( Figure 6C ). This was corroborated by the increased ratio of aspartate aminotransferase (AST) to alanine aminotransferase (ALT) detected in the serum of DIO mice treated with NTDEX ( Figure 6D ). At the tissue level, it is clear that NTDEX does not induce the same therapeutic benefit as P3DEX.