Caloric restriction prevents the development of airway hyperresponsiveness in mice on a high fat diet

We have previously shown that high fat diet (HFD) for 2 weeks increases airway hyperresponsiveness (AHR) to methacholine challenge in C57BL/6J mice in association with an increase in IL-1β levels in lung tissue. We hypothesize that obesity increases AHR via the IL-1β mechanism, which can be prevented by caloric restriction and IL-1β blockade. In this study, we fed C57BL/6J mice for 8 weeks with several hypercaloric diets, including HFD, HFD supplemented with fructose, high trans-fat diet (HTFD) supplemented with fructose, either ad libitum or restricting their food intake to match body weight to the mice on a chow diet (CD). We also assessed the effect of the IL-1β receptor blocker anakinra. All mice showed the same total respiratory resistance at baseline. All obese mice showed higher AHR at 30 mg/ml of methacholine compared to CD and food restricted groups, regardless of the diet. Obese mice showed significant increases in lung IL-1 β mRNA expression, but not the protein, compared to CD and food restricted mice. Anakinra abolished an increase in AHR. We conclude that obesity leads to the airway hyperresponsiveness preventable by caloric restriction and IL-1β blockade.

SCiEntifiC RePoRTS | (2019) 9:279 | DOI: 10.1038/s41598-018-36651-2 groups, respectively] or food restricted to match their weight to the control group on a chow diet [HFD(R), HFD + HFr(R), and HTFD + HFr(R) groups, respectively], and subsequently AHR and pulmonary inflammation were measured. In the second experiment, C57BL/6J mice were fed with a HFD ad libitum for 8 weeks and treated with an IL-1β receptor blocker or placebo during the last 2 weeks of the experiment followed by the same measurements.
Metabolic measurements showed that the HFD(O) and HFD + HFr(O) groups had higher fasting blood glucose levels compared to CD (p < 0.05, Table 1). Fasting blood glucose was lower in HTFD + HFr(R) group compared to CD group but not different between CD, HFD(R) and HFD + HFr(R) group (Table 1). Serum insulin and leptin levels were significantly higher in HFD(O), HFD + HFr(O) groups compared to the CD group. There was no difference in adiponectin levels across the groups (Table 1). Serum free fatty acids (FFAs) were higher in HFD + HFr(R) and HTFD + HFr(R) group than CD group. There was no difference in serum FFAs between HFD(O), HFD + HFr(O), HTFD + HFr(O) and CD groups. The HFD(R) group had lower triglycerides as compared to CD group and HFD (O) group. HFD(O), HFD + HFr(O) and HTFD + HFr(O) groups showed a greater than 2-fold increase in the bronchoalveolar lavage (BAL) cellularity with the predominance of macrophages (Table 2). Neither eosinophils nor basophils were observed in BAL, and the proportion of neutrophils were less than 1% in all groups. Lung triglycerides were higher in HFD + HFr(O) than CD group, but the values were similar between HFD(O), HTFD + HFr(O), caloric restricted groups and CD group (Table 2). There was no difference in lung FFAs across the groups (Table 2). Lung TNF-α and IL-6 were not significantly different between CD, HFD(O) and HFD(R) group (Suppl. Fig. 1). Such pro-inflammatory cytokines as IL-4, IL-5, IL-13, IL-17, IL-18, IL-21, IL-23 were not detected. There were a 3.1 ± 1.0-fold, 4.7 ± 1.6-fold and 8.7 ± 5.2-fold increases in IL-1 β mRNA levels in the lung tissue of HFD(O), HFD + HFr(O) and HTFD + HFr(O) mice compared to the CD group. In contrast, IL-1 β mRNA expression in lung tissue of the HFD(R), HFD + HFr(R), HTFD + HFr(R) groups were not elevated (Fig. 3A). However, IL-1 β protein levels in lung lysates were not significantly different across all the obese and caloric restricted groups, either (Fig. 3B). Flow cytometry also did not show any difference in cell-specific IL-1β production from pulmonary macrophages, neutrophils or lymphocytes between the groups (Suppl. Fig. 2). IL-1 β protein secretion by pulmonary macrophages was similarly not different between the CD, HFD(O) and HFD(R) groups (Fig. 3C). IL-1β mRNA levels in epididymal fat and inguinal fat tissues were not increased in mice fed on HFD ad libitum as compared to CD group ( Table 2).
In the second experiment we examined the effect of an IL-1β blocker anakinra on metabolic parameters and airway responsiveness in mice fed HFD. The placebo and anakinra groups gained similar amount of weight after feeding with HFD for 8 weeks (Fig. 4). Serum insulin levels were lower in the anakinra group compared to the placebo group (Table 3). There was no significant difference in fasting glucose, leptin, adiponectin, FFAs or triglyceride levels between two groups (Table 3). Anakinra had no effect on Rrs at baseline, 0.65 ± 0.03 cmH 2 O.s/mL vs 0.67 ± 0.03 cmH 2 O.s/mL in the placebo group. There was no difference between the groups at 3 mg/ml of methacholine. However, at 30 mg/mL of methacholine the AHR in the anakinra group was significantly lower compared to the placebo group (2.9 ± 0.9 vs 5.1 ± 1.4 respectively, p = 0.01) (Fig. 5). In fact, the AHR in obese mice treated with anakinra was identical to the lean CD mice and the HFDR mice from Experiment 1. The proportion of lymphocytes in BAL was lower in the anakinra group compared to the placebo group. The lung volumes, serum and lung FFAs, triglycerides and pro-inflammatory cytokines, total BAL cell count and other than lymphocyte cell content were not different between placebo and anakinra groups (Table 4).

Discussion
The main finding of our study was that, in mice on a high fat hypercaloric diet, caloric restriction prevented the development of airway hyperresponsiveness and upregulation of IL-1β gene expression in lung parenchyma, regardless of the diet. Moreover, IL-1β receptor blockade also prevented and maybe even reversed the development of airway hyperresponsiveness in obese mice, despite persistent metabolic abnormalities.
Our study was designed to determine whether obesity per se or diet lead to asthma. A positive correlation between high fat diet intake and asthma has been previously observed 9 . Wood et al. suggests that the abundance of saturated fatty acids and lack of antioxidants in HFD can induce inflammation by activating toll-like receptors and hence stimulating the NF-κB inflammatory cascade 10 . HFD also induces the proliferation of invasive bacteria and eliminates the protective bacteria in the gut which can induce inflammation 11,12 . A high fructose diet has also been associated with asthma 8 . Our data demonstrated that two different types of HFD fed ad libitum induced pulmonary inflammation with increased macrophages in BAL, especially in the HTFD + HFr group. However, our food restriction protocol showed that food restricted mice did not develop AHR, despite being fed the same diet. These results lead to the conclusion that the airway hyperresponsiveness is a consequence of obesity rather than high fat or high fructose diets. We recently reported that HFD feeding for 2 weeks led to an increase in AHR of the similar magnitude as HFD feeding for 8 weeks, despite much more significant weight gain in a longer term experiment 7 . Taken together these data suggest that even mild obesity can lead to airway hyperresponsiveness.
Several physiological and immunological mechanisms have been implicated in the pathogenesis of obese asthma. Excessive adiposity can have a restrictive effect on the lung decreasing functional residual capacity and expiratory reserve volume 13 . Radial traction around the distal airway is decreased at low lung volume contributing to airway narrowing 14,15 . A reduction in initial airway caliber allows a greater increase in resistance for a given absolute reduction in smooth muscle shortening, which manifests as increased airway reactivity 16 . Obesity's effect on lung and chest wall compliance also lessens the effectiveness of a bronchoprotective deep breath to dilate airways 17 . In addition to these mechanical effects, our current study highlights an even more important role of pulmonary inflammation in the pathogenesis of obesity-induced airway hyperresponsiveness in mice.
Our previous study showed that HFD feeding for two weeks increases AHR in association with increased IL-1 β gene expression in the lung and augmented IL-1 β secretion by pulmonary macrophages 7 . The current study showed that the expression of IL-1 β mRNA was increased in the mouse lungs after HFD feeding for 8 weeks and this increase was prevented by caloric restriction. The IL-1β receptor blockade also prevented both obesity-induced AHR and pulmonary inflammation, supporting the concept that up-regulation of IL-1β gene expression in the lung could be a mechanism linking obesity and asthma 5 . Given that obesity increases AHR early in the time course 7 and that anakinra was administered only during last two weeks of the 8-week experiment, our data suggest that IL-1β receptor blockers not merely prevent, but also reverse obese asthma. Mechanisms of obesity-induced up-regulation of IL-1β have been linked to the NLRP3 inflammasome, which was examined in detail by Umetsu et al. 5 . However, we did not detect an increase in adipose IL-1β in mice on a HFD. We found an increase in lung triglycerides only in the HFD + HFr (O) group, whereas the AHR was increased in all obese groups. Nevertheless, the lung is an important organ of triglyceride rich lipoprotein clearance 18 and it is conceivable that particular species of fatty acids, e.g. long chain fatty acids, which were not
Other potential mechanisms contributing to obesity-induced pulmonary inflammation and AHR are hyperleptinemia 21 , hyperglycemia 22 and insulin resistance 23,24 , which were observed in obese mice on a HFD, but not in control chow fed mice or food restricted mice on a HFD 25 . Of note, the IL-1β receptor blocker anakinra has been shown to improve hyperglycemia and insulin secretion in type 2 diabetes 26 . In our study, anakinra lowered fasting serum insulin levels without a significant change in fasting glucose levels indicating increased insulin sensitivity. Our finding suggests that anakinra could contribute to improvement in AHR due to its off-target effects by improving glucose metabolism.
The most important novel finding of our study was that caloric restriction prevented the development of airway hyperresponsiveness. The data from several trials showed that weight loss in asthmatics through caloric restriction can lead to clinical improvement 6,[27][28][29] . We agree and further propose that caloric restriction may be beneficial for asthma in obese individuals, possibly by suppressing the IL-1β response.
Our study had several limitations. First, we were unable to demonstrate an increase in IL-1β protein. We have previously shown that IL-1β secretion by pulmonary macrophages is dramatically increased after 2 weeks of high fat diet 7 . After 8 weeks of HFD and development of severe obesity, IL-1β mRNA levels remained elevated, but the IL-1β protein levels and secretion by pulmonary macrophages were no longer increased. These data may suggest that IL-1β peaked early in the time course of HFD -induced obesity. Second, there was an apparent discrepancy between the lack of increase in IL-1β protein levels and beneficial effects of the IL-1β receptor blocker on AHR. Besides possible fluctuations of IL-1β levels over the time course, anakinra could exert a therapeutic effect in the absence of an increase in IL-1β by blocking the receptor and downstream inflammation. It is also possible that anakinra had non-specific off-target beneficial effects, for instance by alleviating insulin resistance as discussed above. Third, other than IL-1β mechanisms by which obesity may increase AHR were not addressed. Recent literature suggest that long chain fatty acids may induce inflammation by altering macrophage lipid metabolism indirectly via the toll receptor 4 30 . Finally, we did not investigate sex differences by including female mice.  Figure 5. Anakinra injection decreased total resistance of the respiratory system (Rrs) in response to methacholine in HFD induced obese mice. The Rrs values were normalized to baseline (no significant difference between groups at baseline). * Denotes p < 0.05.

Conclusions and Implications
Diet induced obesity increases airway hyperresponsiveness and the effects of obesity are preventable by caloric restriction and IL-1β blockade. Taken together our data suggest that caloric restriction should be used for prevention of obese asthma and that IL-1β blockade may be considered as an adjunct therapy.

Methods
Animals and study design. In caloric-restricted groups, mice were provided with the same HFD, HTFD and fructose but the amount of food was restricted to match body weight to the CD group. The composition of HFD has been described previously 7 .
Diets were refrigerated at 4-8 °C before it was added to the cages. 16 C57BL/6J male mice were used in the second experiment. The second experiment consisted of 2 groups, placebo group (n = 8) and anakinra group (n = 8).
Anakinra was a gift from Sobi (Stockholm, Sweden). Both groups were fed with HFD ad libitum for 8 weeks.
During last 14 days of the experiment, the anakinra group was injected subcutaneously at 50 mg/kg in 250 µl of saline daily and placebo group was injected subcutaneously with 250 µl of saline daily.
Physiological measurements. Mice were anesthetized with ketamine/xylazine i.p., tracheostomized and the total respiratory resistance (Rrs) was measured by forced oscillation technique (Flexivent, SCIREQ Québec, Canada) at baseline and after methacholine aerosol challenge at 3 and 30 mg/mL as described 7,31 . Blood was collected from the aorta, BAL was performed with 2 × 0.8 mL of sterile phosphate-buffered saline (PBS) through a tracheal cannula. The thorax was opened, and the right lung was tied off, dissected free and immediately frozen in liquid nitrogen and stored at −80 °C. Inguinal fat and epididymal fat tissue were collected, immediately frozen in liquid nitrogen and stored at −80 °C.  monocytes, neutrophils, alveolar and interstitial macrophages were gated with characteristic low forward scatter/ side scatter, using a FACSAria instrument and FACSDiva for data acquisition (Becton Dickinson) and Flowjo for analysis (Tree Star Inc.) as previously described 33 .
Statistical analysis. The statistical analysis was done using STATA version 12. All values were reported as means ± SEM. All the data in the study was checked for normality with Jarque-Bera test. Statistical analysis of normally distributed variables was determined by student t-test or one-way analysis of variance test (ANOVA) with repeated measures when appropriate. Non-normally distributed values were analyzed by Kruskal-Wallis rank test. A p-value of <0.05 was considered significant.