Age-related decline in HMGB1-neutralizing IgM autoantibody response impairs resistance to high-fat diet in mice

Abstract

As we age, it becomes increasingly important to reduce the consumption of fatty foods. In mice, we also find that after consuming a high-fat diet, older mice develop insulin resistance more easily than young mice. But how aging renders both humans and mice more vulnerable to the detrimental effect of fatty foods is not completely known. Fatty food consumption has been shown to increase extracellular HMGB1, a key player in driving sterile inflammation. In this study, we show in mice that aging impairs the ability to produce, after stimulation of HMGB1, a neutralizing anti-HMGB1 IgM autoantibody that controls the extracellular HMGB1 level. This impairment in eliciting the anti-HMGB1 IgM response renders mice, regardless of age, more susceptible to the development of insulin resistance after consuming high-fat diet. The cause of this impairment lies within the B-1 cells that produce the autoantibody. As they age within the mice, these B-1 cells become less sensitive to the HMGB feedback stimulation mediated via TLR4 signaling. As a result, the mice fail to upregulate the anti-HMGB1 IgM autoantibody in response to the increase in extracellular HMGB1 following fatty food consumption. These findings point to age-related decline in eliciting the anti-HMGB1 IgM response as one of the factors contributing to age-related loss of tolerance to fatty foods. The possibility to explore this immune axis as a therapeutic target emerges.

Introduction

It is unrealistic to completely avoid consuming fatty foods throughout our lifetime. It is thought that as we age, our bodies become more vulnerable to the detrimental effects of fatty foods. Although it has been shown, in mice, that increased age potentiates the deleterious effects of consuming fatty foods,^1,2 it is still open to question whether aging renders the body more susceptible to the harmful effects of fatty foods, and if so, what causes such an increase in susceptibility.

Consumption of fatty foods has been shown to elevate the level of extracellular high mobility group box 1 (HMGB1), which is, evolutionally, a highly conserved proinflammatory mediator that plays a major role in driving sterile inflammation.³ A homeostatic serum level of extracellular HMGB1 promotes immune defense, tissue repair, and tissue regeneration.⁴ However, excessive extracellular HMGB1 causes uncontrolled inflammation and promotes insulin resistance (IR) via activating the JNK and NF-κB pathways, which causes inhibitory serine phosphorylation of IRS-1 and impairs insulin signaling.^5–9

Our previous study reveals that the level of extracellular HMGB1 is controlled at the humoral level by an equally conserved anti-HMGB1 IgM autoantibody that binds to an epitope called HMW4.^10,11 HMW4 stands for HMGB1-derived Weak Epitope 4, a name derived from a bioinformatics algorithm that we previously used to identify the epitope.¹¹ As a well-known ligand for TLR4, extracellular HMGB1 activates the TLR4 signaling pathway.⁴ Furthermore, our recent study shows that extracellular HMGB1 acts as a BCR/TLR4 dual ligand in B-1 cells, and that concomitant triggering of BCR/TLR4 in B-1 cells by extracellular HMGB1 stimulates the production of anti-HMGB1 IgM autoantibody, which in turn neutralizes endogenous extracellular HMGB1.¹⁰ Taken together, anti-HMGB1 IgM autoantibody functions as a novel feedback loop mechanism regulating the homeostasis of extracellular HMGB1.

In the present study, we found that the feedback loop limiting the level of extracellular HMGB1 is impaired in aging. During aging, B-1 cells become less sensitive to the HMGB feedback stimulation mediated via TLR4 signaling, and consequently, they fail to upregulate anti-HMGB1 IgM autoantibody in response to the increase in extracellular HMGB1 following consumption of high-fat diet (HFD), resulting in a more rapid onset of IR. Our findings show that aging renders the body more susceptible to the harmful effects of fatty food consumption, and that such increase in the susceptibility is partly attributed to age-related impairment of B-1 cells in the anti-HMGB1 IgM feedback loop mechanism.

Materials and methods

Mice and diets

Aged C57BL/6 mice were provided by the National Institute on Aging and were used with approval from the Biologic Resource Committee of the University of Illinois College of Medicine Rockford. The HFD was obtained from Research Diets (D12492); the low-fat diet (LFD) was from Harlan Laboratories (Harlan Teklad 7001 diet). The same batches of diets were used for all HFD versus LFD experiments. C57BL/6 mice were initially fed the LFD and switched to the HFD as indicated. The mice were categorized by age as young (8-week-old), middle-aged (9-month-old), older (14-month-old), and aged (20-month-old).

Reagents

Most of the reagents used in this study have been described previously.^10,11 BV510 anti-mouse CD19, BV421 anti-mouse CD23, and FITC anti-mouse F4/80 were from BioLegend. APC-Cy7 streptavidin (SA) was from BD Biosciences. PE anti-phospho-NF-κB p65 (Ser536) was from Cell Signaling Technology. Hybridoma-serum free medium (SFM) was from Thermo Fisher Scientific. B-1 cell staining mix was formed from 2 μg of Fc blocker, 6 μg of mouse IgG, 0.5 μg of BV510 anti-mouse CD19, 0.25 μg of BV421 anti-mouse CD23, 0.25 μg of FITC anti-mouse F4/80, 0.4 μg of APC-Cy7 SA, 4 μg of SA, and 0.07 μg of biotin. The final volume was adjusted to 60 μl in blood buffer (PBS/2 mM EDTA/0.5% BSA). For staining, 2 μl of the mix was used per million total B cells.

Metabolic studies

All blood glucose levels were determined from tail venous blood with a glucose monitor (Contour plus). Intraperitoneal glucose tolerance test (IPGTT) was performed following 12 h fast. In brief, glucose (2 g/kg body weight) was injected i.p., and blood glucose was measured at before and defined timepoints after glucose injections. Intraperitoneal insulin tolerance test (IPITT) was performed following 6 h fast. In brief, insulin (1 U/kg body weight) was injected (i.p.), and blood glucose was measured before and at specific timepoints after insulin injections. Plasma insulin was measured using an ELISA kit (Abcam, United Kingdom).

ELISAs

Methods have been described by us previously for detecting plasma anti-HMGB1 IgM,¹¹ plasma anti-HMW4 IgM,¹⁰ and plasma HMGB1.¹¹ The method for detecting plasma anti-HMGB1 IgM in the presence of HMW4 (as specific blocker) or raHMW4 (as nonspecific blocker) has also been described earlier.¹¹ “% reduced” on the y-axis in Fig. 1B was calculated as: $\mathrm{\%}\mathrm{r}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{e}\mathrm{d}=(\mathrm{a}-\mathrm{b})/\mathrm{a}\times 100\mathrm{\%}$, where a and b are the levels of anti-HMGB1 detected in the absence and presence of a blocker, respectively. To detect the HMGB1:anti-HMW4 IgM immune complex formed in cell culture, a Greiner Microlon High-Binding Plate (catalog number: 655081) was coated with goat anti-mouse IgM in coating buffer (0.05 M sodium bicarbonate, pH 9.6), blocked with 0.5% BSA, and loaded with spent medium. The immune complex was bound to the plate, whereas free recombinant HMGB1 was removed through washing. Biotinylated 3E8 anti-HMGB1 mAb, which does not bind to the HMW4 epitope as we showed previously,¹¹ was added along with 200 ng/ml of mouse IgG2b (as a blocker), followed by detection using high-density streptavidin–horseradish peroxidase.

Figure 1.

HFD challenge upregulates the anti-HMW4 IgM response in young mice. (A) Young (8-week-old) C57BL/6 males were fed a high-fat diet (HFD) or low fat-diet (LFD); at indicated time points on the diet, levels of plasma anti-HMGB1 IgM were analyzed by ELISA against plate-bound HMGB1. Shown (as mean ± SD) are the data from 12 mice per group. *P ≤ 0.02 (t-test), HFD vs LFD. (B) Plasma pooled from the mice fed HFD for 4 to 12 weeks in (A) was tested by ELISA, using HMW4 or raHMW4 (randomized peptide of the same amino acid composition) as specific or nonspecific blocker, respectively. Readings without any blocker were set as 100%. Shown (as mean ± SD) are data of 4 experiments. *P ≤ 0.02 (t-test), HMW4 vs raHMW4.

Detection of TLR4 signaling in peritoneal B-1 cells and macrophages

Mouse peritoneal cavity cells were obtained nonelicited from lavage and resuspended in 50 μl of hybridoma-SFM containing the TLR4 agonist monophosphoryl lipid A (MPLA) of various doses and incubated at room temperature for 1 h. The cells were blocked and stained with the B-1 cell staining mix for 10 min on ice. The cells were fixed in 4% paraformaldehyde and permeabilized in 0.2 M glycine/0.5% Tween-20 (each of these steps was performed at room temperature for 10 min). After permeabilization, the cells were stained with PE anti-phospho-NF-κB p65 (1:100 dilution in blood buffer) at 4 °C overnight. On the next day, the cells were analyzed via flow cytometry. Peritoneal B-1 cells and macrophages were identified as FSC^largerCD19⁺CD23⁻ cells and F4/80⁺ cells, respectively. Mean fluorescence intensities (MFIs) of PE within the gated B-1 cells and macrophages were quantified as phosphorylation of NF-κB p65 (“p-p65”) in these cells.

Detection of TLR4 signaling in HMW4-specific peritoneal B-1 cells

Peritoneal cells were resuspended in 50 μl of hybridoma-SFM with or without 1 μg/ml of MPLA and incubated at room temperature for 1 h. The cells were stained sequentially with the B-1 cell staining mix and HMW4 tetramer as described previously.¹⁰ They were then fixed, permeabilized, and stained with PE anti-phospho-NF-κB p65 as described earlier. HMW4-specific B-1 cells were identified as FSC^largerCD19⁺CD23⁻tetramer⁺ cells.¹⁰ ΔMFI is the MFI of the MPLA-stimulated – MFI of the nonstimulated.

Depleting HMW4-specific B-1 cells in young male mice

Depleting HMW4-specific B-1 cells in young C57BL/6 male mice was done as we previously described.¹⁰

Stimulating the TLR4 axis in middle-aged male and older female mice

Middle-aged C57BL/6 males were injected (i.p.) weekly with a low dose (1 μg/mouse) of MPLA or PBS (as control) at the first day of a week for a total of 4 weeks while being fed HFD. Older C57BL/6 females were also given the same treatments as middle-aged males while being fed HFD for 4 weeks. Fasting (12 h) blood glucose levels were monitored before treatment and biweekly after treatment. On the last day of the last week, blood anti-HMW4 IgM and cytokine levels were analyzed via ELISA and BD Cytokine Bead Array (BD Biosciences), respectively.

Statistical analysis

Data for 2 independent groups were analyzed by the unpaired 2-tailed t-test. Bonferroni correction was applied when multiple t-tests were performed simultaneously. Data for more than 2 groups were analyzed by one-way ANOVA followed by post hoc Tukey honestly significant difference test. Results with a P value less than 0.05 were considered statistically significant.

Results

HFD challenge upregulates the anti-HMW4 IgM response in young mice

We previously reported that western-type diet challenge upregulates anti-HMGB1 IgM response in atherosclerotic Apoe^−/− mice.¹¹ In the present study, we first assessed whether HFD challenge induces the same response in wild-type C57BL/6 mice. To that end, young (8-week-old) C57BL/6 male mice were fed an HFD (60% calories from fat) for 12 weeks; as control, mice were fed an LFD (10% calories from fat). During that period, plasma anti-HMGB1 autoantibodies (autoAbs) were monitored by ELISA.¹¹ Anti-HMGB1 IgM autoAb was induced in the first week of HFD and peaked after 4 weeks (Fig. 1A), whereas no IgG subclasses were detected (data not shown). The majority (~75%) of the anti-HMGB1 IgM was directed to the HMW4 epitope (HMGB1_98–112) because it could be blocked by the HMW4 peptide (Fig. 1B). Thus, the anti-HMGB1 IgM belongs to the evolutionarily conserved anti-HMW4 IgM autoAb we reported previously.¹⁰ Collectively, these results show that HFD challenge upregulates the anti-HMW4 IgM response in young mice.

Middle-aged mice are impaired in the anti-HMW4 IgM response to HFD challenge

When we repeated the HFD challenge in middle-aged (9-month-old) mice, we noticed an intriguing difference. After 4 weeks of HFD, while the young mice upregulated anti-HMW4 IgM, the middle-aged mice failed to do so, but rather maintained a low, static level of anti-HMW4 IgM (Fig. 2A, left panel). The same results were obtained when ELISA plates were coated with whole HMGB1 protein (Fig. 2A, right panel); thus, failure of the middle-aged mice to upregulate anti-HMW4 IgM was not due to an epitope change. Also, HFD challenge resulted in increases in circulating HMGB1 level in young and middle-aged mice (Fig. 2B); thus, the impairment in eliciting the anti-HMW4 IgM response to HFD was not due to a lack of HMGB1 either. These results indicate that middle-aged mice are impaired in eliciting the anti-HMW4 IgM response to HFD challenge.

Figure 2.

Middle-aged mice are impaired in the anti-HMW4 IgM response to HFD challenge. Young (8-week-old) or middle-aged (9-month-old) C57BL/6 males were fed an HFD or LFD. At 4 weeks on the diet, the mice were analyzed by ELISA for the plasma levels of anti-HMW4 and anti-HMGB1 IgM antibodies (A) and HMGB1 (B). Data shown (as mean ± SD) are from 12 mice per group. *P ≤ 0.002; **P ≤ 0.04 (t-test).

Middle-aged mice show decrease in short-term tolerance to HFD

We next determined whether the age-related difference in the anti-HMW4 IgM response in Fig. 2 would affect the tolerance of mice to HFD, ie their ability to delay the onset of IR after HFD consumption. As expected, HFD feeding resulted in increase in body weight in young and middle-aged mice (Fig. 3A). In the second week on HFD, middle-aged mice already showed significant increase in fasting blood glucose level compared to LFD-fed mice; in contrast, such increase was not observed for even up to 4 weeks in young mice (Fig. 3B). Focusing on IR as the direct cause of glucose dysregulation, we found that, at 3 to 4 weeks on HFD, middle-aged mice showed significant (~3-fold) increase in fasting insulin level, whereas young mice did not show such an increase (Fig. 3C). Upon HFD challenge, when compared to young mice, middle-aged mice showed decreases in glucose tolerance (Fig. 3D) and insulin sensitivity (Fig. 3E). The middle-aged mice, even when fed an LFD, showed decreases in glucose tolerance and insulin sensitivity. The data suggest that IR can develop due to aging, independently of an HFD. Collectively, these results reveal an age-related difference in short-term tolerance to HFD, with middle-aged mice being less tolerant to HFD challenge (with earlier onset of IR) than young mice. Taken together, our findings (Figs. 2 and 3) indicate that middle-aged mice are impaired in eliciting the anti-HMW4 IgM response to HFD, which correlates with their loss of short-term tolerance to HFD.

Figure 3.

Middle-aged mice show decrease in short-term tolerance to HFD. C57BL/6 males of different ages were fed the indicated diets and measured for body weight (A) and fasting (12 h) glucose level (B). *P ≤ 0.0004, HFD vs LFD. **P ≤ 0.003, HFD-fed young vs HFD-fed middle-aged mice. At 3 to 4 weeks on the diet, the mice were measured for fasting (12 h) insulin (C); IPGTT (12 h fasting) (D) and IPITT (6 h fasting) (E) were performed. *P ≤ 0.03; **P ≤ 0.05. Data shown (as mean ± SD) are from 12 mice per group and were analyzed by t-test.

Raising anti-HMW4 IgM in middle-aged mice improves their short-term tolerance to HFD

Having revealed the correlation between the upregulation of anti-HMW4 IgM and short-term tolerance to HFD, we then wanted to determine whether anti-HMW4 IgM played a causal role in diabeto-protection. First, adopting the “gain-of-function” approach, we assessed whether raising anti-HMW4 IgM in middle-aged mice can increase their tolerance to HFD. We showed previously that anti-HMW4 IgM was produced by the B-1 cells in the peritoneal cavity.¹⁰ These cells produced the Ab when activated through both of their cognate BCR and TLR4 concurrently.¹⁰ We therefore immunized middle-aged male mice by injecting (i.p.) HMW4 and the TLR4 agonist MPLA, as described previously.¹⁰ We did not include control groups injected with either HMW4 or MPLA in this experiment, as we had previously shown that both HMW4 and MPLA are required for immunization to elicit the anti-HMW4 IgM response.¹⁰ The mice were then switched to HFD and, subsequently, analyzed for their tolerance to HFD. As expected, the immunization had effectively raised the level of plasma anti-HMW4 IgM (Fig. 4A). No significant intergroup difference in body weight was seen (data not shown). Importantly, the immunization increased short-term (up to ~4 weeks) tolerance to HFD (Fig. 4B, C) compared to the nonimmunized control. These results indicate that raising anti-HMW4 IgM in middle-aged mice via immunization renders them tolerant to short-term HFD challenge.

Figure 4.

Raising anti-HMW4 IgM in middle-aged mice improves their short-term tolerance to HFD. Middle-aged C57BL/6 males were injected (i.p.) with PBS (“Control”) or a mixture of HMW4 (25 μg) and MPLA (2 μg) (“Test”), as we previously described,¹⁰ on days 0 and 7. On day 14, the mice were switched to HFD for 12 weeks. (A) Level of plasma anti-HMW4 IgM was determined at 4 weeks on HFD. (B) Fasting (12 h) glucose level was measured at different times on HFD. (C) Fasting (12 h) insulin level was determined at 4 weeks on HFD. Shown are the data from 12 mice per group. *P ≤ 0.006 (t-test), “Test” vs “Control.”

Diminishing anti-HMW4 IgM response in young mice impairs their tolerance to HFD

We wanted to confirm the causal role of anti-HMW4 IgM in diabeto-protection. We thus took the “loss-of-function” approach and determined whether diminishing the anti-HMW4 IgM response in young mice can reduce their tolerance to HFD. To that end, anti-HMW4 IgM-producing B-1 cells in young mice were depleted using the immune complex “DOSC” (depletion of stained cells) we designed¹⁰; then, the mice were challenged with HFD. Compared to the (mock-treated) control mice, DOSC-treated mice showed an ~80% reduction in plasma anti-HMW4 IgM (Fig. 5A). Critically, at 4 weeks on HFD, while the control mice still exerted normal function in glucose metabolism, DOSC-treated mice developed hyperglycemia (Fig. 5B) and showed significant increase in fasting insulin level (Fig. 5C).

Figure 5.

Diminishing anti-HMW4 IgM response in young mice impairs their tolerance to HFD. Young C57BL/6 males were depleted of HMW4-specific B cells by injection (i.p.) of immune complex DOSC (“Test”) or mock-treated (“Control”). One week postdepletion, the mice were challenged with HFD for 4 weeks and then analyzed. (A) Plasma anti-HMW4 IgM levels were analyzed via ELISA. (B) Fasting (12 h) blood glucose levels were measured. (C) Fasting (12 h) blood insulin levels were analyzed by ELISA. Data shown (as mean ± SD) are from 12 mice per group. *P ≤ 0.0001 (t-test), “Test” vs “Control.”

To summarize, the results from these complementary approaches identify the anti-HMW4 IgM response as one of the factors contributing to short-term protection against HFD-induced IR.

B-1 cells of middle-aged mice are impaired in upregulating anti-HMW4 IgM in response to HMGB1

Because peritoneal B-1 cells are the major producer of anti-HMW4 IgM,¹⁰ we next assessed the age-related difference in eliciting the anti-HMW4 IgM response in cultured peritoneal B-1 cells. While young and middle-aged mice possessed a similar fraction of the HMW4-specific B-1 cells, identified via staining with the HMW4 tetramer as we reported,¹¹ the cells from the latter failed to upregulate anti-HMW4 IgM upon stimulation by (exogenous) recombinant HMGB1 (rHMGB1) (Fig. 6, on a log scale). The result demonstrates that B-1 cells of middle-aged mice have an impairment in producing anti-HMW4 IgM in response to stimulation by HMGB1.

Figure 6.

B-1 cells of middle-aged mice are impaired in upregulating anti-HMW4 IgM in response to HMGB1. Peritoneal B-1 cells from young or middle-aged C57BL/6 males were cultured in the presence or absence of rHMGB1 (10 μg/ml) for 7 days. Secreted anti-HMW4 IgM was analyzed by ELISA as HMGB1:IgM immune complex. Data shown (as mean ± SD) are from 4 experiments. *P = 0.001 (t-test).

HMW4-specific B-1 cells show age-related decline in TLR4 signaling

Previously, we reported that extracellular HMGB1 stimulated the production of anti-HMW4 IgM by B-1 cells via co-triggering BCR and TLR4.¹⁰ Consistent with that finding, we showed that immunizing middle-aged mice with HMW4 and (the TLR4 agonist) MPLA resulted in strong upregulation of anti-HMW4 IgM and rendered the mice tolerant to subsequent short-term HFD challenge (Fig. 4). Our findings, and previous studies showing age-related decrease in TLR signaling in immune cells such as dendritic cells and macrophages,^12,13 compelled us to determine whether the impairment in producing anti-HMW4 IgM is attributed, at least partly, to age-related decline in TLR4 signaling in HMW4-specific B-1 cells (producing the IgM).

Given the rather low percentage of HMW4-specific B-1 cells, we assessed their TLR4 signaling (upon triggering by MPLA) using an intracellular phosphor-flow method.¹⁴ We began with total peritoneal B-1 cells. As shown in Fig. 7A, an age-dependent decline in TLR4 signaling was most evident when these cells were stimulated with 1 μg/ml of MPLA (over the tested dose range of 0.1–10 μg/ml). At this optimal MPLA dose (1 μg/ml), we analyzed HMW4-specific B-1 cells, identified using the HMW4-tetramer (“tetramer”), as we described previously.¹¹ The HMW4-specific (tetramer⁺) B-1 cells showed an age-related decline in TLR4 signaling, similarly to that observed in total B-1 cells (Fig. 7B). Because B-1 cells are classified as innate (or “innate-like”) immune cells, we performed the same experiment with peritoneal macrophages, another type of innate cells that are abundant in the peritoneal cavity. At any MPLA concentrations tested, the decrease in TLR4 signaling was not observed (Fig. 7C). These argue against a global decline in TLR4 signaling, at least in the innate arm. Collectively, these results point to an age-related defect in TLR4 signaling in B-1 cells.

Figure 7.

HMW4-specific B-1 cells show age-related decline in TLR4 signaling. (A) Peritoneal total B-1 cells in young, middle-aged, and aged (20-month-old) C57BL/6 males were stimulated for 1 h with MPLA at indicated doses and analyzed for the level of phospho-NF-κB p65 by flow cytometry; ≥60,000 B-1 cells (per age group) were analyzed. Top panel depicts the MFI analysis of phospho-NF-κB p65. *P = 0.001; **P = 0.0009; ***P = 0.0001 (one-way ANOVA). (B) Peritoneal HMW4-specific B-1 cells were incubated with or without MPLA (1 μg/ml) for 1 hour and then analyzed for phospho-NF-κB p65. ΔMFI (phospho-NF-κB p65) was calculated by subtracting the MFI of a nonstimulated control from the MFI of an MPLA-stimulated sample; ≥1,000 HMW4-specific (tetramer⁺) B-1 cells were analyzed. *P = 0.0001; **P = 0.00007; ***P = 0.00001 (one-way ANOVA). (C) Peritoneal macrophages were similarly analyzed for phospho-NF-κB p65; ≥50,000 macrophages were analyzed. Data shown (as mean ± SD) in this figure are from 6 mice per group.

Enhanced stimulation of TLR4 in middle-aged mice improves short-term tolerance to HFD challenge

To further demonstrate that the impairment in producing anti-HMW4 IgM was attributed, at least partly, to the age-related defect in TLR4 signaling, we next determined whether enhanced TLR4 signaling in middle-aged mice can restore their anti-HMW4 IgM response and tolerance to HFD. To that end, middle-aged male mice were injected (i.p.) weekly with a low dose (1 μg/mouse) of MPLA or PBS (as control) while being fed HFD. After 4 weeks, while the control mice had minimal (or basal) level of anti-HMW4 IgM, MPLA-treated mice robustly produced anti-HMW4 IgM (Fig. 8A). The treatment also lowered plasma levels of the proinflammatory cytokines TNF-α and IL-6 (Fig. 8B) and restored the short-term tolerance to HFD in the middle-aged mice (Fig. 8C).

Figure 8.

Enhanced stimulation of TLR4 in middle-aged mice improves short-term tolerance to HFD challenge. (A–C) Middle-aged C57BL/6 males were injected (i.p.) weekly with PBS (control) or MPLA (1 μg/mouse) while being fed HFD. Anti-HMW4 IgM in blood was analyzed via ELISA at 4 weeks on HFD (A). *P ≤ 0.0001, MPLA vs PBS. Levels of plasma TNF-α and IL-6 were analyzed by BD Cytokine Bead Array (B). *P ≤ 0.0001, PBS vs MPLA. Fasting (12 h) glucose was analyzed biweekly (C). *P ≤ 0.0001, PBS vs MPLA. Data shown (as mean ± SD) in this figure are from 10 mice per group and were analyzed by t-test.

Age-related decline in the anti-HMGB1 IgM response impairs tolerance to HFD in females

To determine whether the findings made in male mice also hold true for female mice, we next performed the same experiments in female mice. Young C57BL/6 female mice are relatively resistant to the diabetogenic effect of HFD,^15,16 likely due to the effect of estrogen,^15–17 which positively regulates B cell homeostasis and Ab production.^18,19 Consistent with those findings, young females (not shown) and middle-aged (9-month-old) females (Fig. 9A) remained tolerant to HFD challenge for >4 weeks, whereas older (14-month-old) females (having reached reproductive senescence) became relatively susceptible to HFD challenge (Fig. 9A). Critically, while middle-aged females upregulated anti-HMW4 IgM upon HFD challenge, older females were relatively impaired in eliciting the anti-HMW4 IgM response (Fig. 9B). Thus, in older females the impairment in eliciting the anti-HMW4 IgM response to HFD challenge, again, correlates with loss of diabeto-protection.

Figure 9.

Age-related decline in the anti-HMW4 IgM response impairs tolerance to HFD in females. (A) Middle-aged (9-month-old) and older (14-month-old) C57BL/6 females were fed HFD, and the mice were measured for fasting (12 h) glucose level at different times on HFD. *P ≤ 0.04 (t-test), older vs middle-aged. Shown (mean ± SD) are data from 12 mice per group. Of note, fasting glucose levels in age-matched, LFD-fed control mice were 53–87 mg/dl during the period (not depicted). (B) At 4 weeks on the indicated diet, plasma anti-HMW4 IgM levels were analyzed by ELISA. *P ≤ 0.001 (t-test). Shown (mean ± SD) are data from 12 mice per group. (C) Peritoneal HMW4-specific B-1 cells in young, older, and aged C57BL/6 females were analyzed for phospho-NF-κB p65 by flow cytometry; ≥1,000 HMW4-specific (tetramer⁺) B-1 cells were analyzed. *P = 0.0001; **P = 0.00007; ***P = 0.00001 (one-way ANOVA). Data shown (as mean ± SD) are from 6 mice per group. (D and E) Older C57BL/6 females were given the same treatment of MPLA and switched to HFD, as described in Fig. 8A. Anti-HMW4 IgM in blood was analyzed via ELISA at 4 weeks on HFD (D). *P ≤ 0.0001, MPLA vs PBS. Fasting (12 h) glucose was analyzed biweekly (E). *P ≤ 0.0001, MPLA vs PBS. Data shown in D and E (as mean ± SD) are from 10 mice per group and were analyzed by t-test.

To assess potential defect in TLR4 signaling of the B-1 cells in females, we used older mice in place of middle-aged mice, as females showed the impairment and vulnerability to HFD at older age. HMW4-specific B-1 cells, again, showed age-dependent decrease in TLR4 signaling (Fig. 9C), whereas peritoneal macrophages did not show such a decrease (data not shown). In older females, the MPLA treatment also restored the anti-HMW4 IgM response (Fig. 9D) and tolerance to HFD (Fig. 9E). Collectively, these revealed that a similar age-related impairment in producing anti-HMW4 IgM, attributed to the defect in TLR4 signaling of the B-1 cells, leads to the loss of tolerance to HFD in female mice.

Discussion

We have shown that the anti-HMW4 IgM response plays a role in diabeto-protection, and that middle-aged male and older female mice are impaired in producing anti-HMW4 IgM autoAbs in response to HFD challenge, which is attributed partly to a defect in TLR4 signaling in B-1 cells of those mice. We have shown that such impairment in upregulating anti-HMW4 IgM is one of the factors contributing to age-related loss of tolerance to fatty foods.

A young, healthy body adjusts metabolism to fuel supply or demand and adapts to metabolic challenges such as fatty foods. During aging, metabolic function and adaptability decline.^20,21 Aging enhances HFD-related adverse effects,^22,23 including metabolic stress.^1,24 While it is well accepted that we should eat healthfully throughout life, fatty foods have been pervading modern dietary habits. Despite the studies showing that aging potentiates the harmful effects of fatty food consumption, it remains unclear why aging renders the body more vulnerable to metabolic stress from fatty foods. Our study identifies the anti-HMGB1 IgM response of B-1 cells as a humoral mechanism regulating tolerance to HFD, providing new insight into the immunometabolic crosstalk during metabolic stress.

In the present study, aging negatively skews the TLR4 signaling response to HMGB1 in peritoneal B-1 cells, resulting in impaired upregulation of the anti-HMGB1 IgM upon HFD challenge. Interestingly, we found that the TLR4 signaling activity in peritoneal macrophages did not change with age. It should be emphasized that we examined nonelicited (resident) macrophage in this study. The phenotype and functions of these cells differ from thioglycolate-elicited macrophages that infiltrate the peritoneal cavity under an inflammatory condition.²⁵ Previous studies of the latter found age-related impairment in TLR4 activity in peritoneal macrophages,^26,27 likely due to these differences. Our results suggest that the age-related loss of TLR4 activity in peritoneal B-1 cells may occur sooner than in peritoneal resident macrophages.

Unlike B-2 cells and macrophages, B-1 cells maintain homeostasis throughout life primarily through self-renewal and, thus, age alongside the body. Considering this characteristic, we speculate that the age-related defect in TLR4 signaling may be unique to B-1 cells and that their responses to other TLR ligands may be similarly impaired by aging, likely to varying degrees depending on the specific ligand. While further research will be necessary to confirm or refute this hypothesis, the present study suggests that B-1 cell senescence, partly evidenced by the impaired ability to elicit an anti-HMGB1 IgM response, may contribute to the loss of short-term tolerance to fatty food challenges during aging. Although this tolerance is short-term, it allows occasional or even frequent fatty food challenges to remain well below the threshold for triggering IR in a young body. In contrast, the same metabolic stress may push an aged body past this threshold.

To solidify that the impairment in producing anti-HMW4 IgM was attributed (at least partly) to the age-related defect in TLR4 signaling, we determined whether enhancing TLR4 signaling in middle-aged male as well as older female mice using the TLR4 agonist MPLA can restore their anti-HMW4 IgM response and, more critically, tolerance to HFD. One caveat was that triggering TLR4 in vivo may have prodiabetic, proinflammatory effects, which can confound the result interpretation. TLR4 signaling in insulin target tissues (such as adipose tissue) was shown to activate proinflammatory kinases (JNK, IKK, p38) that impair insulin signaling via inhibitory phosphorylation of IRS-1; also, TLR4 activation was shown to cause upregulation of proinflammatory mediators (cytokines, chemokines, etc), which further promotes insulin desensitization (reviewed in²⁸). Such effect was observed with LPS.^29,30 As a LPS derivative with much (up to 1,000-fold) lower toxicity, MPLA was shown to have weaker proinflammatory activity,³¹ which was, likely, due to its TRIF-bias resulting in reduced production of proinflammatory mediators.^32,33 Despite this notion, we did a preliminary study to evaluate the effect of MPLA in our experimental setting. Our results showed that the proinflammatory effect of MPLA at the maximal dose used for the present study (2 μg/mouse, i.p.) is weak, transient, and confined locally (unpublished results). MPLA was shown to cause metabolic reprograming in macrophages.^34,35 In those studies, however, MPLA was used at higher doses, eg 20 μg/mouse;³⁴ also, the reprogramming of macrophages, while increasing their antimicrobial activity, did not increase production of proinflammatory cytokines.³⁴

Our study does not exclude the potential role of other molecular axes (such as Wnt/β-catenin, BAFF, APRIL, CD148-Lyn, etc),^36–39 and the interplay between TLR4 and these axes, in eliciting the anti-HMGB1 IgM response to metabolic stress. Studies are needed to address this issue. Besides producing the neutralizing anti-HMGB1 IgM autoAb, the body also has other ways to control the activity, availability, and signaling of HMGB1 (and, likely, other damage-associated molecular patterns [DAMPs]). These include the mechanisms regulating the redox forms of HMGB1⁴⁰; binding of HMGB1 to CD24, which offsets HMGB1-induced dendritic cell activation via mobilizing Siglec-G (Siglec-10 in humans)⁴¹; binding of the N-terminal domain of thrombomodulin to HMGB1 and thrombin to promote thrombin-mediated HMGB1 degradation⁴²; and so forth. The role played by these mechanisms in modulating the body’s tolerance to metabolic stress in aging warrants future investigation.

The specific effects of HFD on B-1 cells are relatively understudied. Our study has demonstrated that HFD activates B-1 cells via the HMGB1:anti-HMW4 IgM feedback loop. Previously, it was shown that B-1a cells are reduced in frequency in obese HFD-fed mice.⁴³ HFD was found to impair phosphatidylcholine-specific B-1 cell phagocytosis in obese mice, inducing a transition of B-1 cells into macrophages.⁴⁴ The composition of dietary fat also influences B cell responses. For example, a reduction in IgM production by B-1 cells was shown to be reversible with n-3 polyunsaturated fatty acid supplementation.⁴⁵ These findings underscore the potential of B-1 cells as therapeutic targets.

We observed that plasma HMGB1 levels were increased in middle-aged mice compared to young mice, even when fed an LFD (Fig. 2B). This finding aligns with previous studies in the literature. As a prototypical DAMP, extracellular HMGB1 increases with age,⁴⁶ and its elevation has been implicated in various age-related traits (reviewed in⁴⁷). As shown in Fig. 6, aging mice on an LFD exhibit decreased sensitivity of B-1 cells to HMGB1, which compromises the production of anti-HMW4 IgM. This is partly due to age-related decline in TLR4 signaling (Fig. 7). These changes result in a weakened feedback response that is insufficient to fully counteract the rising HMGB1 levels in aging mice. We also observed increased levels of anti-HMW4 IgM in middle-aged mice compared to young mice, even when fed an LFD (Fig. 2A). A plausible interpretation is that the higher circulating HMGB1 in middle-aged mice stimulates a relatively stronger anti-HMW4 IgM response.

Our study provides a simple method for exploring anti-HMW4 IgM-producing B-1 cells as therapy target. These B-1 cells can be raised via immunization, using MPLA as adjuvant. In this regard, to avoid off-target effect of TLR4 signaling, the use of MPLA may need to be optimized; for instance, strategy can be designed to target MPLA to B-1 cells specifically, as done with dendritic cells and liver cells.^48,49 Alternatively, anti-HMW4 IgM may be explored directly as “injectable” drug for ameliorating inflammatory, age-related diseases including IR/type 2 diabetes.

Data availability

The data underlying this article are available in the article.