Intestinal prevents metabolic syndrome in mice

Kanakaraju Kaliannana,1, Sulaiman R. Hamarneha,1, Konstantinos P. Economopoulosa,1, Sayeda Nasrin Alama, Omeed Moavena, Palak Patela, Nondita S. Maloa, Madhury Raya, Seyed M. Abtahia, Nur Muhammada, Atri Raychowdhurya, Abeba Teshagera, Mussa M. Rafat Mohameda, Angela K. Mossa, Rizwan Ahmeda, Shahrad Hakimiana, Sonoko Narisawab, José Luis Millánb, Elizabeth Hohmannc, H. Shaw Warrenc, Atul K. Bhand, Madhu S. Maloa,2, and Richard A. Hodina

aDepartment of Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114; bSanford Children’s Health Research Center, Sanford-Burnham Medical Research Institute, La Jolla, CA 92037; and cInfectious Disease Unit, Department of Medicine and Pediatrics, and dDepartment of Pathology, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114

Edited by Patricia K. Donahoe, Massachusetts General Hospital, Boston, MA, and approved March 15, 2013 (received for review November 22, 2012) Metabolic syndrome comprises a cluster of related disorders that been shown that an HFD causes local intestinal inflammation includes obesity, glucose intolerance, insulin resistance, dyslipide- (11). Systemic and local inflammation lead to overexpression of mia, and fatty liver. Recently, gut-derived chronic endotoxemia has proinflammatory cytokines (12), which cause increased gut per- been identified as a primary mediator for triggering the low-grade meability (13) and an acceleration of endotoxin translocation inflammation responsible for the development of metabolic syn- (14), resulting in a vicious cycle of endotoxemia. A central role of drome. In the present study we examined the role of the small LPS in the pathogenesis of metabolic syndrome is also supported intestinal brush-border enzyme, intestinal alkaline phosphatase by the observation that mice lacking toll-like receptor 4 (TLR4), (IAP), in preventing a high-fat-diet–induced metabolic syndrome the receptor for LPS, are resistant to HFD-induced inflammation, in mice. We found that both endogenous and orally supplemented obesity, and insulin resistance (15). IAP inhibits absorption of endotoxin (lipopolysaccharides) that We and others have shown that the brush-border enzyme in- occurs with dietary fat, and oral IAP supplementation prevents as testinal alkaline phosphatase (IAP) detoxifies a variety of bacterial well as reverses metabolic syndrome. Furthermore, IAP supplemen- fl toxins, including LPS, CpG DNA, and agellin (16). Furthermore, MEDICAL SCIENCES tation improves the lipid profile in mice fed a standard, low-fat it has been reported that inhibition of endogenous IAP by L-phe- chow diet. These results point to a potentially unique therapy nylalanine (Phe) increases serum endotoxin levels (17). Based against metabolic syndrome in at-risk humans. upon these previous observations regarding IAP function, we hy- pothesized that this enzyme could play an important role in pre- atherosclerosis | cytokines | diabetes | dysbiosis | steatosis venting gut-derived systemic inflammation. Here we report that endogenous IAP plays a critical role in reducing endotoxemia, and etabolic syndrome is a complex syndrome composed of oral supplementation with IAP prevents HFD-induced endotox- Ma cluster of disorders that includes obesity, glucose in- emia, as well as metabolic syndrome in mice. We also show that tolerance, insulin resistance, abnormal lipid profile (dyslipidemia), IAP supplementation was able to reverse HFD-induced metabolic fi fatty liver, and hypertension (1, 2). Metabolic syndrome leads to syndrome and improves the lipid pro le in mice fed a standard low- fi type 2 diabetes, atherosclerosis, and nonalcoholic fatty liver dis- fat chow diet (LFD). Taken together, these ndings suggest that ease (1, 2). Approximately 35–39% of the US population suffers oral IAP supplementation may represent a unique therapeutic from the syndrome (3). This epidemic of metabolic syndrome has approach for the prevention or treatment of metabolic syndrome devastating consequences in terms of mortality, morbidity, and in humans. total healthcare expenditures (4). Results Recently, “metabolic endotoxemia” has been proposed to be fi central to the pathogenesis of metabolic syndrome. The Gram- IAP Prevents Endotoxemia. Given that IAP detoxi es LPS and negative bacterial wall component lipopolysaccharide (LPS) other bacterial toxins, we predicted that IAP knockout (KO) is known as endotoxin, and metabolic endotoxemia is defined as mice (18) would suffer from metabolic endotoxemia and also metabolic syndrome. Indeed, we found that IAP-KO mice suffer a two- to threefold persistent increase in circulating endotoxin A concentrations above the normal levels (5). Metabolic endotox- from endotoxemia (Fig. 1 ) as well as overexpression of the fl proinflammatory cytokines TNF-α (Fig. 1B) and interleukin-1β emia leads to low-grade systemic in ammation as evidenced by β A increased serum levels of tumor necrosis factor-alpha (TNF-α), (IL-1 )(Fig. S1 ) IAP-KO mice also had greatly increased gut permeability based on enhanced absorption of orally adminis- interleukin (IL)-1, and IL-6 (5). It is well recognized that chronic C inflammation causes damage to pancreatic beta cells (6), hep- tered dextran-FITC (Fig. 1 ). As expected, this increased gut permeability caused enhanced LPS translocation in IAP-KO atocytes (7), and vascular endothelial cells (8), and dysfunction of D these cells is thought to contribute to metabolic syndrome. mice compared with WT mice (Fig. 1 ). A high-fat diet (HFD) has been shown to cause metabolic endotoxemia in animals and humans (5, 9), but the underly- ing molecular mechanisms remain incompletely understood. Author contributions: M.S.M. and R.A.H. developed the study concept and theory; K.K., M.S.M., and R.A.H. designed research; K.K., S.R.H., K.P.E., S.N.A., O.M., P.P., N.S.M., M.R., Ghoshal et al. (10) demonstrated that intestinal epithelial cells S.M.A., N.M., A.R., A.T., M.M.R.M., A.K.M., R.A., and S.H. performed research; K.K., K.P.E., (enterocytes) internalize LPS from the apical surface, which is S.N., J.L.M., E.H., H.S.W., A.K.B., M.S.M., and R.A.H. analyzed data; and M.S.M. and R.A.H. then transported to the Golgi apparatus where it complexes with wrote the paper. chylomicrons, the lipoproteins that transport the absorbed long- The authors declare no conflict of interest. chain fatty acids in enterocytes. The chylomicron–LPS complex This article is a PNAS Direct Submission. is then secreted into mesenteric lymph and makes its way into the 1K.K., S.R.H., and K.P.E. contributed equally to this work. systemic circulation. Excess chylomicron formation during high- 2To whom correspondence should be addressed. E-mail: [email protected]. fat feeding leads to prolonged chylomicronemia (complexed with This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. LPS) that ultimately induces systemic inflammation. Also, it has 1073/pnas.1220180110/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1220180110 PNAS | April 23, 2013 | vol. 110 | no. 17 | 7003–7008 Downloaded by guest on September 24, 2021 A B C D E IAP-KO:LPS LFD, GTT WT:LPS 1.2 7070 *** 0.400.4 *** 400 1.2 *** 0.3 2.5 400 * **

g/ml)) 2.5 350 1 6060 5 ** 1.0 μ ** (pg/ml) 0.300.3 2.02 300 0.8 5050 0.2 300

0.8 α 1.5 250 40 5 1.5 0.6 40 0.2 200 ** 0.6 0.200.1 1 200 3030 1.0 150

(EU/ml) 0.4 0.4 5 0.5 (mg/dl) 2020 0.1 0.5 100 IAP-KO 0.2 0.100.0 100 0.2 1010 00 50 WT

5 nd Blood glucose Serum endotoxin endotoxin Serum 0.00 00 00 -0.5-0.5 00 Change in serum serum in Change Endotoxin (EU/ml) Endotoxin Intestinal permeability 12 Serum TNF- WT-IAP WT IAP-KO WT12 IAP-KO (dextran-FITC ( WT IAP-KO (min) 04512 (min)015306090120 0 15 30 60 90 120 MaleKO-

F G MaleH I 2.02 IAP-KO 200200 IAP-KO * 2525 IAP-KO ****** 1212 IAP-KO *** WT ** WT 20 WT ** 1010 1.5 150150 20 ** WT 1.5 * * ** 8 * 1515 8 1.01 100100 * 66 *** 1010 *** (mg/dl) 4

gain (g) gain 4 0.50.5 5050 55 2 Body weight weight Body LFD, ITT WAT/BW (%) 2 Blood glucose 0 Serum insulin 0.00 0 00 00 t t s during GTT (ng/dl) GTT during (min) 00306015 30 (min)0 0 30609012 30 60 90 120 Age (wk) 0123456789 4 8 12 16 Subcutaneous Epididymal Visceral ou l fa l fa ne t ym a ta J K L M fa id : : : : isce ra WT HF WT Oil WT Oil cu WT Oil V id

: Ep WT:HF+Phe WT:Oil+IAP (10U) WT Oil+IAP (500U) Sub WT:LPS WT:LF+Phe WT:Oil+IAP (100U) IAP-KO:Oil WT:Oil+LPS WT:LF WT:Oil+IAP (500U) IAP-KO:Oil+IAP (500U) WT:Oil+LPS+IAP (100U) 600 # 2 4 : 600 ## GTT 2.0 WT Oil+LPS+IAP (500U)

*** 5 500500 1.51.5 *** # *** 3 4 400 1 3 400 1.0 *** # 2 2 300300 0.50.5 1 *** 1 *** 0 200200 0.00 -1-1 -0. -2 100100 -0.5 0 -3-3 Change in serum serum in Change Change in serum serum in Change Endotoxin (EU/ml) Endotoxin Endotoxin (EU/ml) Endotoxin 5 -4 Change in serum serum in Change Endotoxin (EU/ml) Endotoxin Blood glucose(mg/dl) 00 -1.0-1 -1-1 -5-5 Time (min)0 153060901215 30 60 90 1200 Time (min) 045 Time (min) 045 Time (min) 045

− − Fig. 1. IAP prevents endotoxemia. Groups of 4- to 17-wk-old C57BL/6 male IAP-KO (Akp3 / ) mice and their WT littermates (n = 5–10 each group) were used in these experiments unless otherwise indicated (SI Materials and Methods). Mice received a low-fat diet (LFD, 14% kcal from fat) unless otherwise indicated. (A) Serum endotoxin levels. (B) Serum TNF-α levels. (C) Intestinal permeability as determined by quantifying the amount of FITC-dextran (70 kDa) levels in the serum after its oral gavage. (D) Change in serum endotoxin levels after oral LPS gavage. (E) Blood glucose levels during glucose tolerance test (GTT). (F) Serum insulin levels during the first 30 min of GTT. (G) Blood glucose levels during insulin tolerance test (ITT). (H) Body weight gain by mice on LFD. (I) White adipose tissue (WAT) distribution. (J) Blood glucose levels during GTT in mice receiving LFD (LF group) or high-fat diet (HFD, 45% kcal from fat) (HF group) treated with L-phenylalanine (Phe, 10 mM in the drinking water), a specific inhibitor of IAP. (K) Inhibition of corn-oil–induced endotoxemia in CD-1 mice by IAP in a dose-dependent manner. (L) Inhibition of corn-oil–induced endotoxemia in WT and IAP-KO mice. (M) Inhibition of corn-oil– plus LPS-induced endotoxemia in CD-1 mice by IAP in a dose-dependent manner. Statistics: data expressed as mean ± SEM. Two-tailed unpaired Student’s t test. For multiple comparisons, analysis of variance with Tukey was used. * or #P < 0.05; ** or ##P < 0.01; *** or ###P < 0.001. The number sign (#) refers to the HF vs. HF + Phe comparison.

The above data established that the IAP-KO mice display ev- oil alone compared with corn oil plus calf IAP, and these in- idence of metabolic endotoxemia. We next investigated whether hibitory effects of IAP were found to be dose dependent (Fig. these mice also display features of metabolic syndrome. We ob- 1K). We also observed higher serum LPS concentrations in IAP- served that IAP-KO mice suffer from glucose intolerance (Fig. 1E KO mice fed with corn oil compared with WT mice (Fig. 1L). and Fig. S1B) and hyperinsulinemia (Fig. 1F). An insulin toler- Oral IAP supplementation prevented corn-oil–induced endotox- ance test (ITT) showed that IAP-KO mice had significantly higher emia in both groups of mice (Fig. 1L) and in a dose-dependent glucose levels at multiple time points, indicating a state of insulin manner was able to reduce endotoxemia when excess LPS was resistance in these animals (Fig. 1G and Fig. S1C). We also ob- administered along with the corn oil (Fig. 1M). served metabolic-syndrome–associated obesity in IAP-KO mice (Fig. 1H) and found that these mice accumulated higher body fat IAP Prevents Chronic High-Fat-Diet–Induced Metabolic Syndrome. We (white adipose tissue, WAT) compared with WT (Fig. 1I), in- next assessed whether over a prolonged period, oral supplemen- cluding much more intraabdominal fat (Fig. S1D). These data tation with IAP could prevent metabolic syndrome. We exposed indicate that endogenous IAP plays a critical role in preventing 15-wk-old male C57BL6 mice to an HFD (45% kcal from fat) ± metabolic endotoxemia and the resultant metabolic syndrome IAP (100 units/mL drinking water) for 11 wk. Control mice (low in mice. fat, LF group) consumed a standard chow LFD (14% kcal from The amino acid Phe specifically inhibits the enzymatic activity of fat) only. As expected, the mice receiving the HFD + IAP (HF + IAP. Mice receiving an HFD and Phe (HF + Phe group) had im- IAP group) exhibited better glucose tolerance (Fig. 2A) and less paired glucose tolerance compared with mice receiving an HFD postprandial hyperinsulinemia (Fig. 2B) and insulin resistance alone (HF group), indicating the preventive role of endogenous (Fig. 2C) than mice receiving the HFD alone (HF group). Fasting IAP (Fig. 1J and Fig. S1E). In addition, the HF + Phe group had blood glucose levels were higher in the HF compared with LF significantly higher serum endotoxin levels (Fig. S1F). Interestingly, mice, whereas there were no differences between LF and HF + the related groups consumed the same amount of food and had IAP groups (Fig. 2 A and D and Fig. S2 A and B). We observed similar weight gain (Fig. S1 G and H). Furthermore, IAP-KO mice higher body weight in the HF group, however no significant dif- receiving an HFD also suffered from type 2 diabetes as evidenced ference between LF and HF + IAP groups (Fig. S2C). The HF by higher fasting glucose levels and abnormal glucose tolerance and HF + IAP groups had nearly equal energy intake (Fig. S2D); tests (Fig. S1 I and J). however, the feed efficiency [FE = weight gained (g)/kcal con- An HFD is associated with enhanced translocation of LPS sumed × 100] was lower, although insignificant, in the HF + IAP from the gut to the systemic circulation through chylomicrons. group (Fig. S2E). We also found that IAP prevented HFD- We found higher serum LPS concentrations in WT mice fed corn induced decrease in pancreatic weight (Fig. 2E). Compared with

7004 | www.pnas.org/cgi/doi/10.1073/pnas.1220180110 Kaliannan et al. Downloaded by guest on September 24, 2021 A B LF C D LF *** GTT * HF * # # HF ### 500 # 3.5 25 ** 350 *** 500 * 3.5 HF+IAP 25 350 HF+IAP ### * 3 300 * 400400 # # * 3.0 * 2020 300 # 2.52.5 250250 300300 ** 2.02 1515 200200 200200 LF 1.51.5 1010 150150

(ng/dl) 1 100 (mg/dl) 1.0 (mg/dl) 100 100100 HF 55 ITT HF+IAP 0.50.5 5050 Serum insulin Blood glucose Blood Blood glucose 00 00 00 00 ) ) )

Time (min)0 1530609012015 30 60 90 120 Fasting After Insulin resistance (HOMA-IR) Index Time (min) 0030609012 30 60 90 120 Fasting 2 h after GTT LF HF HF+IAP=4

2hrs (n =4 (n =4 D E F LF G eh H ** * LF + I + A 0.8 12 HF ** 25 *** V + # 800 *

0.8 12 25 F 800 HF 0.7 HF+IAP H 700 1010 * # 20 0.60.6 ** 20 600600 88 # 0.5 * ## 1515 500 0.40.4 66 ** 400400 10 0.3 44 10 300 0.20.2 200 2 55 200 0.1 WAT/BW (%) 2 100

0 0 0 intake Energy 0 0 0 0 (kcal/group/wk)

Adiposity index Adiposity (total WAT/BW (%)) 0 P F P eh eh Pancreas/BW (%) L LF HF HF+IAP SubcutaneousSubcutaneousEpididymalEpididymalVisceral fatVisceral fa LFLF HF HF+IAP LFHF+VehLF HFHF+cIA HF+IAPP

Fig. 2. IAP prevents chronic HFD-induced glucose intolerance and insulin resistance. Groups of 15-wk-old WT male C57BL/6 mice (n = 5 for each group) were fed an HFD (45% kcal from fat) ± IAP (100 units/mL in drinking water) for 11 wk. A control group of mice received LFD (14% kcal from fat). (A) Blood glucose levels during GTT in mice receiving LFD (LF group), HFD (HF group), and HFD + IAP (HF + IAP group). (B) Serum insulin levels during GTT. (C) Insulin resistance index (homeostasis model assessment of insulin resistance, HOMA-IR). (D) Blood glucose levels during ITT. (E) Weight of pancreas. (F) White adipose tissue (WAT) expressed as percentage of body weight (BW). (G) Adiposity index. (H) Energy intake by different groups. Statistics: data expressed as mean ± SEM. Two-tailed unpaired Student’s t test. For multiple comparisons, analysis of variance with Tukey was used. * or #P < 0.05; ** or ##P < 0.01; *** or ###P < 0.001. The asterisk (*) refers to the LF vs. HF comparison and the number sign (#) refers to the HF vs. HF + IAP comparison.

the HF group, mice in the HF + IAP group also showed an overall Most notably, the HDL-C levels markedly increased by ∼73% in reduction in body fat, most dramatically seen in the case of s.c. fat the HF + IAP group compared with the HF group (Fig. 4D). The (Fig. 2F). As expected, the adiposity index (sum of the weights of HF group receiving IAP supplementation exhibited a great re- adipose depots expressed as the percentage of total body weight) duction in atherogenic propensity including the important choles- was also significantly decreased in the HF + IAP group compared terol ratios and the atherogenic index (Fig. S4 A–F). MEDICAL SCIENCES with the HF group (Fig. 2G), despite the fact that the two groups consumed nearly equal amounts of food (Fig. 2H). We noticed that IAP Prevents HFD-Induced Endotoxemia and Intestinal Permeability. even after just 6 wk of an HFD, mice developed glucose intolerance We next sought to assess some of the factors known to underlie that was prevented in the IAP-treated group (Fig. S2 F–I). metabolic syndrome. Compared with the LF mice, serum endo- toxin levels were much higher in the HF group (Fig. 4F), whereas IAP Prevents HFD-Induced Liver Injury. We next examined the effi- these levels were reduced in the HF + IAP group. In addition, cacy of IAP in preventing HFD-induced liver injury. IAP prevented compared with the HF group, the endotoxin levels were also HFD-induced increase in liver weight (Fig. S3) and also protected greatly reduced in the cecal contents of the IAP-treated group mice from HFD-induced increase in the liver enzymes, aspartate (Fig. S4G). To further explore the mechanism by which IAP aminotransferase (AST), gamma-glutamyl transferase (GGT), and prevents endotoxemia, we sought to determine whether oral alanine aminotransferase (ALT) (Fig. 3 A–C). In addition, com- IAP detoxifies LPS within the intestinal lumen or, alternatively, pared with the HF mice, the IAP-treated animals had lower levels if IAP enters the systemic circulation to detoxify circulating of total liver lipids (Fig. 3D) and triglycerides (TG) (Fig. 3E). Fi- LPS. Accordingly, we treated separate groups of mice with oral or nally, histological analyses showed that HF mice had accumulated i.p. IAP and oral or i.p. LPS. Oral IAP supplementation blocked much higher amounts of hepatic fat, and these changes were not the luminal LPS but was unable to prevent endotoxemia induced seen in the HF + IAP animals (Fig. 3 F and G). by two different doses of i.p. LPS (Fig. 4G and Fig. S4H). In contrast, the i.p. IAP was able to detoxify the i.p. LPS. These IAP Prevents HFD-Induced Dyslipidemia. Lipid profile analyses results support the concept that oral IAP prevents endotoxemia showed that IAP prevents HFD-induced dyslipidemia (Fig. 4 A–E). by detoxifying LPS within the intestinal lumen.

A B C D * * ** ** *** * 140140 2020 6060 100100 120 120 5050 80 100 1515 80 100 4040 8080 6060 1010 3030 6060 40 20 40 40 20 40 55 20 2020 1010 20 0 0 0 (mg/g tissue) 0

0 0 0 Total lipids liver 0 Serum AST(U/l) Serum ALT(U/l) LF123 HF HF+IAP Serum(U/l) GGT LF123 HF HF+IAP LF123 HF HF+IAP LF123 HF HF+IAP

E * F *** *** G 3030 4.5 4.04 2525 3.53.0 2020 3 2.5 15 15 2.02 1010 1.5 1.01 55 0.50 (mg/g tissue) 00 0 Steatosis score

Liver triglycerides Liver LF123 HF HF+IAP LFDLF HFD HF HF+IAP HFD LF HF HF+IAP

Fig. 3. IAP prevents HFD-induced liver injury. (A) Serum aspartate aminotransferase (AST) levels (see Fig. 2 for description of mice). (B) Serum gamma- glutamyl transferase (GGT) levels. (C) Serum alanine aminotransferase (ALT) levels. (D) Total liver lipids. (E) Liver triglyceride levels. (F) Liver steatosis score. (G) Oil Red O staining of frozen liver sections (10× objective). Statistics: data expressed as mean ± SEM. Two-tailed unpaired Student’s t test. For multiple comparisons, analysis of variance with Tukey was used. *P < 0.05; **P < 0.01; ***P < 0.001.

Kaliannan et al. PNAS | April 23, 2013 | vol. 110 | no. 17 | 7005 Downloaded by guest on September 24, 2021

** **

* * A ** B C D E

200200 200200 * 140140 5050 1.01 120 120 4040 0.80.8 150150 150150 100100 8080 3030 0.60.6

100100 100100 60 ratio LDL (mg/dl) 60 2020 0.40.4 (mg/dl) 5050 5050 4040 Triglycerides 1010 0.20.2 2020 HDL : LDL-C (mg/dl) HDL-C (mg/dl)

Total cholesterol 00 00 00 00 00 LF HF HF+IAP LF HF HF+IAP LF123 HF HF+IAP LF123 HF HF+IAP LF123 HF HF+IAP F LFD HFHF+IAP G 123 H I *** *** *** *** *** *** ** 25 10 120120 *** ** 4000040

25 10 α 35000 100100 α 2020 88 3000030 8080 25000 1515 66 6060 2000020 1010 44 (pg/ml) 15000 (EU/ml) (EU/ml) * 40 40 10000 5 2 10

5 2 Serum TNF- 20 20 Cecal TNF- 5000 Serum endotoxin Serum 00 endotoxin Serum 00 00 00 P (pg/mg cecal content) cecal (pg/mg LFD HFD HFD

LF HF HF 12345 LF

LF HF HF+IAP IAP (200 U/ml) Oral Oral i.p. LF HFEH HF+IAP LF HF HF+IAP

− − IA + + LPS (25 μg) − i.p. Oral i.p. i.p. + + + F+V veh IAP HF veh IAP J K L LF H M LF HF HF 3 ** HF+IAP 400 HF+IAP 3.0 2.02 88 400 GTT ** g/ml)) 2.5 16 wk 7 350 *** μ 2.5 6 300 2 1.51.5 6 300 ** 2.0 5 250 *** 1.5 1.0 4 200 ** permeability 1.5 1 4 200 ** 3 150

1 (mg/dl) 0.5 (g) gain 0.50.5 22 100100 0.5 Akp3 mRNA 1.0 weight Body 1 50 nd Blood glucose 00 00 00 00 123 (relative expression)

Intestinal 8 9 10 11 12 (dextran-FITC ( 0 LF HF HF+IAP LF HF HF+IAP Age (wk) 8 9 10 11 12 Time (min) 015 15 306 60 90 120 123 120

Fig. 4. IAP prevents HFD-induced dyslipidemia and endotoxemia. (A) Serum total cholesterol levels (see Fig. 2 for description of mice). (B) Levels of serum triglycerides. (C) Serum low-density lipoprotein cholesterol (LDL-C) levels. (D) Serum high-density lipoprotein cholesterol (HDL-C) levels. (E) Ratio between HDL-C and LDL-C. (F) Serum endotoxin levels. (G) Serum endotoxin levels after i.p. injection of LPS (different groups of mice, see SI Materials and Methods). (H) Serum tumor necrosis factor-alpha (TNF-α) levels. (I) Cecal TNF-α levels. (J) Intestinal permeability as determined by quantifying the amount of FITC-dextran (70 kDa) levels in the serum after its oral gavage (different groups of mice, see SI Materials and Methods). (K) Relative expression of duodenal IAP (Akp3) mRNA as determined by qRT-PCR (different groups of mice, see SI Materials and Methods). (L) Body weight after 4 wk of HFD feeding, which was preceded by 3 wk of oral antibiotics treatment [ampicillin (1 g/L) plus norfloxacin (1 g/L)] (see Fig. S4P for energy intake) (different groups of mice, see SI Materials and Methods). (M) GTT after 4 wk of HFD feeding (see Fig. S4Q for AUC) (see L for description of mice). Statistics: data expressed as mean ± SEM. Two-tailed unpaired Student’s t test. For multiple comparisons, analysis of variance with Tukey was used. *P < 0.05; **P < 0.01; ***P < 0.001. The asterisk (*) refers to the LF vs. HF comparison, and no statistically significant difference was observed between HF and HF + IAP groups (for M).

We found that IAP prevented HFD-induced increased levels of IAP Exerts Beneficial Effects in Treating HFD-Associated Metabolic serum TNF-α and IL-1β (Fig. 4H and Fig. S4I, respectively). To Syndrome. We next investigated whether IAP supplementation further assess the local inflammatory status within the intestine, would be able to reverse any features of metabolic syndrome. We we measured TNF-α levels in the cecal contents and found ap- exposed mice to an HFD for 14 wk to induce metabolic syn- proximately threefold higher levels of TNF-α in the HF compared drome (Fig. S5 A–D) and then treated them with ±calf IAP with the HF + IAP mice (Fig. 4I). We then directly determined supplementation (100 units/mL drinking water) for 6 wk. The the proinflammatory status of the cecal contents by exposing IAP-treated group showed a significant reduction in glucose in- A B mouse leukaemic monocyte macrophage RAW264.7 cells to cecal tolerance and postglucose hyperinsulinemia (Fig. 5 and and E C contents and found that the fluid from the HF + IAP mice was Fig. S5 ). Fig. 5 shows weekly glucose levels during GTT at the fi less inflammatory (Fig. S4J). In addition, we found that the IAP- 15-min time point, indicating the bene cial effects of IAP. Glucose levels were also significantly reduced in the IAP-treated treated animals were protected from the increase in intestinal D F permeability that occurs in response to an HFD (Fig. 4J). We also group during ITT (Fig. 5 and Fig. S5 ). As expected, we ob- served lower endotoxin levels in the cecal contents (Fig. 5E)as studied the acute (10 d) and chronic (16 wk) effects of oral IAP F + supplementation on the expression of IAP (mouse intestinal al- well as in serum (Fig. 5 ) of the HF IAP group. We found that the HF + IAP group had lower levels of serum TNF-α and IL-1β kaline phosphatase Akp3) mRNA (Fig. 4K and Fig. S4K), as (Fig. 5 G and H) and their insulin resistance index was slightly well as the activity of the endogenous IAP enzyme (Fig. S4 L and reduced, but not significantly (Fig. S5G). During this treatment M) and observed no significant differences in IAP expression or period, we observed nearly equal energy intake by the HF and activity among the groups. HF + IAP mice and IAP treatment did not have any effect on Based on the known activity of the IAP enzyme and our data on body weight (Fig. S5 H and I). We saw slight improvement in the luminal LPS and endotoxemia, it appears that IAP may prevent dyslipidemia and extent of liver injury, but in this short experi- metabolic syndrome through a mechanism that involves the lu- ment these differences did not reach statistical significance. minal inhibition of bacterially derived proinflammatory me- diators. To further explore this mechanism for IAP action, we Low-Fat Diet Supplemented with Oral IAP Improves Lipid Profile. We performed an experiment using a 3-wk antibiotic regimen to de- next sought to determine whether oral supplementation with crease the bacterial content in the gut (Fig. S4 N and O). We then IAP would have any beneficial effects in mice exposed to an fed the mice with a high-fat diet ± IAP for 4 wk. Based on LFD. We fed 5-wk-old female mice a LFD ± IAP (100 units/ measures of body weight (Fig. 4L and Fig. S4P) and glucose tol- mL drinking water) for 7 wk. Mice receiving the LFD + IAP erance (Fig. 4M and Fig. S4Q), it appears that IAP was ineffective appeared to show slight improvement in glucose tolerance and in inhibiting the changes of metabolic syndrome in the antibiotic- insulin sensitivity, but these differences were not significant. On treated mice, consistent with our hypothesis that IAP may prevent the other hand, we found dramatic beneficial effects in the case metabolic syndrome by inhibiting ligands derived from luminal of the lipid profile (Fig. 6 A–E). The IAP-treated mice had , such as LPS. higher levels of HDL-C (Fig. 6D) and greatly increased HDL-C:

7006 | www.pnas.org/cgi/doi/10.1073/pnas.1220180110 Kaliannan et al. Downloaded by guest on September 24, 2021 A B * C D 600 * 5 * 600 300 600 * GTT 5 600 HF * 300 * 500500 4 * 250250 ITT 4 550550 400 HF+IAP * 200 400 33 200 * 300300 500500 150150 2

200 (ng/dl) 2 100 200 HF HF 450 100 HF (mg/dl) (mg/dl) (mg/dl) 450 100100 11 5050 HF+IAP Serum insulin HF+IAP HF+IAP 0 0 400 0 Blood glucose Blood glucose Blood glucose 0 0 400 0 Time (min)015306090120 0 15 30 60 90 120 (min) 0 15 300 120 (week)Day03456Week1 012Week2Week3Week4Week5Week (min)0 30609012 30 60 90 120 15 2 1 E F G H 200200 2020 200200 350350 300 α

β 300 150150 1515 * 150150 250250 ** 200200 100100 1010 100100 * 150150 * (pg/ml) (EU/ml) 5050 55 (pg/ml) 5050 100100 50 Serum IL-1 Serum 50 0 0 Serum TNF- 0 0 0 endotoxin Serum 0 0 0

Cecal endotoxin (EU/g cecal content) HF HF+IAP HF HF+IAP HF HF+IAP HF HF+IAP

Fig. 5. IAP cures HFD-induced metabolic syndrome. Groups of 5-wk-old WT male C57BL/6 mice (n = 5 for each group) were fed HFD (45% kcal from fat) for 14 wk to induce metabolic syndrome (Fig. S5). Mice with metabolic syndrome were then treated with or without IAP (100 units/mL drinking water) for 6 wk. (A) Blood glucose levels during GTT. (B) Serum insulin levels during GTT. (C) Weekly blood glucose levels during GTT at 15-min time point. (D) Blood glucose levels during ITT. (E) Cecal endotoxin levels. (F) Serum endotoxin levels. (G) Serum TNF-α levels. (H) Serum IL-1β levels. Statistics: data expressed as mean ± SEM. Two-tailed unpaired Student’s t test. *P < 0.05; **P < 0.01; ***P < 0.001.

LDL-C ratio (Fig. 6E). We also calculated the various athero- determine whether these other AP isozymes also have the capa- genic indices and found that the LF + IAP group had a great bility to protect the host from metabolic syndrome. reduction in the atherogenesis risk (Fig. S6 A–F). IAP exists within an environment that contains a wide range of bacterially derived proinflammatory factors and we have pre- Discussion viously demonstrated that in addition to LPS, other targets for Metabolic syndrome is one of the most important modern global the IAP enzyme include flagellin and CpG DNA (16). In regard

health problems. The etiology of HFD-induced metabolic syn- to metabolic syndrome, although LPS has been the most in- MEDICAL SCIENCES drome is thought to be related to metabolic endotoxemia (19). An tensively studied gut-derived inflammatory mediator, it is likely HFD has also been associated with an imbalance in the normal that it is not the only one because the other mediators also ac- composition and number of microbes in the gut (dysbiosis), tivate the NF-κB pathway and ultimately increase the levels of resulting in barrier dysfunction followed by LPS translocation to proinflammatory cytokines (25). As such, we speculate that the the systemic circulation (20). beneficial effects of IAP may be due to its detoxification of The present study was undertaken based upon work in our a number of proinflammatory factors, not just LPS. Our results laboratory and others that points to the role for IAP in protecting using antibiotics to decrease the luminal bacterial content before the host from bacterial toxins (16). The present data in mice the HFD ± IAP add support to the idea that the mechanism of provide a “proof of principle” that IAP could be an effective oral IAP action largely involves its ability to block luminal, bacterially supplement against endotoxemia, thus protecting the host from derived inflammatory mediators. Of course, we are unable to metabolic syndrome. We have shown that IAP reduces corn-oil– completely rule out the possibility that IAP may also be exerting induced endotoxemia and prevents the inflammation and in- some impact on metabolic syndrome through a mechanism that testinal permeability changes that occur in response to an HFD. is independent of luminal bacterial products. Interestingly, the excess secretion of IAP seen in rats fed an HFD In regard to the gut microflora, we have previously reported (21) could represent a physiological response of the body to that IAP-KO mice display an overall decrease in the number of protect against HFD-associated systemic inflammation. IAP is intestinal bacteria and oral supplementation with IAP in WT primarily expressed in the enterocytes of the proximal small in- mice was able to rapidly restore the normal gut flora in mice testine and is bidirectionally secreted into the intestinal lumen as exposed to antibiotics (24). The prevention of HFD-induced well as the systemic circulation (22). IAP is strictly conserved metabolic endotoxemia in this study could be due in part to the among species (23) and exists within a microbial environment role of IAP in preventing HFD-induced gut dysbiosis, although that highlights its role in the detoxification of bacterial pro- our acute experiments with corn oil and LPS show an inhibitory inflammatory factors as well as in the prevention of dysbiosis impact of IAP that is independent of any chronic changes in the (16, 24). Various isozymes of alkaline phosphatases (APs) exist, flora. The critical role for IAP at the interface between the host namely IAP, placental AP, tissue nonspecific (liver/bone/kidney/ and the intestinal lumen is supported by studies in zebrafish (26), neutrophils) AP, and germ cell AP (23). These various AP who found that IAP expression was induced only when the fish enzymes share significant structural homology as well as func- were exposed to bacteria, whereas the enzyme was absent under tional similarities. It will be interesting in future studies to germ-free conditions. Interestingly, Jang et al. (27) showed that

A B C D E * 120120 140140 8080 3030 * 44 7 wk 7 wk 70 7 wk 7 wk 7 wk 100100 120120 2525 100100 6060 33 8080 *** ** 50 2020 8080 6060 4040 1515 22 6060 30 **

40 10 LDL ratio 40 40 10 (mg/dl) 40 2020 11 (mg/dl) 2020 2020 10 55 Triglycerides Triglycerides LDL-C (mg/dl) HDL-C (mg/dl) 00 00 00 00 HDL : 00

Total cholesterol LF12 LF+IAP LF12 LF+IAP LF12 LF+IAP LF12 LF+IAP LF LF+IAP

Fig. 6. Low-fat diet (LFD) supplemented with IAP improves lipid profile. Groups of 5-wk-old WT female C57BL/6 mice (n = 5) were fed LFD (14% kcal from fat) ± IAP (100 units/mL drinking water) for 7 wk. (A) Serum total cholesterol levels. (B) Serum triglyceride levels. (C) Serum LDL-C levels. (D) Serum HDL-C levels. (E) Ratio between HDL-C and LDL-C. Statistics: data expressed as mean ± SEM. Two-tailed unpaired Student’s t test. *P < 0.05; **P < 0.01; ***P < 0.001.

Kaliannan et al. PNAS | April 23, 2013 | vol. 110 | no. 17 | 7007 Downloaded by guest on September 24, 2021 endogenous IAP levels in rats decrease with age, indicating Materials and Methods “ ” −/− a possible role of loss of IAP as a precipitating cause of met- C57BL/6 IAP-KO (Akp3 ) mice were generated by the Sanford-Burnham abolic syndrome that is known to be common with aging (28). Medical Research Institute (La Jolla, CA) and have been described (18). IAP- In a therapeutic context, Lukas et al. (29) reported on a single KO and WT littermates were bred at the Massachusetts General Hospital arm study in humans with severe ulcerative colitis and showed (MGH) animal facility. All animal work protocols were reviewed and ap- that enterally administered bovine IAP is extremely safe. We have proved by the Institutional Animal Care and Use Committee at MGH. Groups previously shown that oral IAP supplementation in mice increases of age- and sex-matched WT and IAP-KO mice (n = 5) were fed an LFD (14% IAP concentrations in stools in a dose-dependent manner (24), kcal from fat) or HFD (45% kcal from fat) for the specified periods of time suggesting that oral IAP dose could be easily adjusted to achieve (see SI Materials and Methods for details) with or without calf IAP (100 units/ an effective therapeutic level. mL drinking water). Established protocols were followed to determine in- In summary, we have demonstrated that oral IAP supplemen- sulin resistance (5), dyslipidemia (18), endotoxemia (5), and hepatic enzyme levels (15). Frozen liver tissues were submitted to the histology laboratory at tation both prevents an HFD-induced metabolic syndrome and MGH to prepare Oil Red O stained slides. Details of methods are described in reverses the changes associated with an HFD-induced metabolic SI Materials and Methods. syndrome. Furthermore, IAP can also improve the lipid profile during an LFD feeding. Although caution needs to be taken in ACKNOWLEDGMENTS. We thank Dr. Hang Lee (Department of Biostatistics, extrapolating the findings in mouse models of inflammation to MGH) for help with statistical analyses. This work was supported by National human diseases (30), taken together, the present findings suggest Institutes of Health (NIH) Grants R01DK050623 and R01DK047186 (to R.A.H.), a Junior Faculty Award (to M.S.M.) from the MGH Department of Surgery, that oral IAP supplementation in humans could represent a safe, and a Grand Challenge Exploration Grant from the Bill and Melinda Gates effective, and unique approach to the prevention or treatment of Foundation (to M.S.M.). The work was also supported by NIH Grant 2P30 metabolic syndrome. DK43351 to the Center for the Study of Inflammatory Bowel Disease, MGH.

1. Grundy SM, Brewer HB, Jr., Cleeman JI, Smith SC, Jr., Lenfant C; American Heart As- 16. Chen KT, et al. (2010) Identification of specific targets for the gut mucosal defense sociation; National Heart, Lung, and Blood Institute (2004) Definition of metabolic factor intestinal alkaline phosphatase. Am J Physiol Gastrointest Liver Physiol 299(2): syndrome: Report of the National Heart, Lung, and Blood Institute/American Heart G467–G475. Association conference on scientific issues related to definition. Circulation 109(3): 17. Koyama I, Matsunaga T, Harada T, Hokari S, Komoda T (2002) Alkaline phosphatases 433–438. reduce toxicity of lipopolysaccharides in vivo and in vitro through dephosphorylation. 2. Hotamisligil GS (2006) Inflammation and metabolic disorders. Nature 444(7121): Clin Biochem 35(6):455–461. 860–867. 18. Narisawa S, et al. (2003) Accelerated fat absorption in intestinal alkaline phosphatase 3. Ford ES (2005) Prevalence of the metabolic syndrome defined by the International knockout mice. Mol Cell Biol 23(21):7525–7530. Diabetes Federation among adults in the U.S. Diabetes Care 28(11):2745–2749. 19. Moreira AP, Texeira TF, Ferreira AB, Peluzio MdoC, Alfenas RdeC (2012) Influence of 4. Malik S, et al. (2004) Impact of the metabolic syndrome on mortality from coronary a high-fat diet on gut microbiota, intestinal permeability and metabolic endotox- heart disease, cardiovascular disease, and all causes in United States adults. Circula- aemia. Br J Nutr 108(5):801–809. tion 110(10):1245–1250. 20. de La Serre CB, et al. (2010) Propensity to high-fat diet-induced obesity in rats is as- 5. Cani PD, et al. (2007) Metabolic endotoxemia initiates obesity and insulin resistance. sociated with changes in the gut microbiota and gut inflammation. Am J Physiol Diabetes 56(7):1761–1772. Gastrointest Liver Physiol 299(2):G440–G448. 6. Hohmeier HE, Tran VV, Chen G, Gasa R, Newgard CB (2003) Inflammatory mechanisms 21. Mahmood A, Shao JS, Alpers DH (2003) Rat enterocytes secrete SLPs containing al- in diabetes: Lessons from the beta-cell. Int J Obes Relat Metab Disord 27(Suppl 3): kaline phosphatase and cubilin in response to corn oil feeding. Am J Physiol Gas- S12–S16. trointest Liver Physiol 285(2):G433–G441. 7. Gieling RG, Wallace K, Han YP (2009) Interleukin-1 participates in the progression 22. Eliakim R, Mahmood A, Alpers DH (1991) Rat intestinal alkaline phosphatase secre- from liver injury to fibrosis. Am J Physiol Gastrointest Liver Physiol 296(6):G1324– tion into lumen and serum is coordinately regulated. Biochim Biophys Acta 1091(1): G1331. 1–8. 8. Hansson GK (2005) Inflammation, atherosclerosis, and coronary artery disease. N Engl 23. Goldstein DJ, Rogers CE, Harris H (1980) Expression of alkaline phosphatase loci in J Med 352(16):1685–1695. mammalian tissues. Proc Natl Acad Sci USA 77(5):2857–2860. 9. Pendyala S, Walker JM, Holt PR (2012) A high-fat diet is associated with endotoxemia 24. Malo MS, et al. (2010) Intestinal alkaline phosphatase preserves the normal homeo- that originates from the gut. Gastroenterology 142:1100–1101 e1102. stasis of gut microbiota. Gut 59(11):1476–1484. 10. Ghoshal S, Witta J, Zhong J, de Villiers W, Eckhardt E (2009) Chylomicrons promote 25. Kaisho T, Akira S (2006) Toll-like receptor function and signaling. JAllergyClin intestinal absorption of lipopolysaccharides. J Lipid Res 50(1):90–97. Immunol 117(5):979–987, quiz 988. 11. Ding S, et al. (2010) High-fat diet: Bacteria interactions promote intestinal in- 26. Bates JM, Akerlund J, Mittge E, Guillemin K (2007) Intestinal alkaline phosphatase flammation which precedes and correlates with obesity and insulin resistance in detoxifies lipopolysaccharide and prevents inflammation in zebrafish in response to mouse. PLoS ONE 5(8):e12191. the gut microbiota. Cell Host Microbe 2(6):371–382. 12. Park EJ, et al. (2007) Dietary ganglioside inhibits acute inflammatory signals in in- 27. Jang I, Jung K, Cho J (2000) Influence of age on duodenal brush border membrane testinal mucosa and blood induced by systemic inflammation of li- and specific activities of brush border membrane enzymes in Wistar rats. Exp Anim popolysaccharide. Shock 28(1):112–117. 49(4):281–287. 13. Bruewer M, et al. (2003) Proinflammatory cytokines disrupt epithelial barrier function 28. Bechtold M, Palmer J, Valtos J, Iasiello C, Sowers J (2006) Metabolic syndrome in the by apoptosis-independent mechanisms. J Immunol 171(11):6164–6172. elderly. Curr Diab Rep 6(1):64–71. 14. Cani PD, et al. (2008) Changes in gut microbiota control metabolic endotoxemia- 29. Lukas M, et al. (2010) Exogenous alkaline phosphatase for the treatment of patients induced inflammation in high-fat diet-induced obesity and diabetes in mice. Diabetes with moderate to severe ulcerative colitis. Inflamm Bowel Dis 16(7):1180–1186. 57(6):1470–1481. 30. Seok J, et al.; Inflammation and Host Response to Injury, Large Scale Collaborative 15. Tsukumo DM, et al. (2007) Loss-of-function mutation in Toll-like receptor 4 prevents Research Program (2013) Genomic responses in mouse models poorly mimic human diet-induced obesity and insulin resistance. Diabetes 56(8):1986–1998. inflammatory diseases. Proc Natl Acad Sci USA 110(9):3507–3512.

7008 | www.pnas.org/cgi/doi/10.1073/pnas.1220180110 Kaliannan et al. Downloaded by guest on September 24, 2021