Page 1 of 57 Diabetes
1 Deficiency of ZnT8 promotes adiposity and metabolic dysfunction by
2 increasing peripheral serotonin production
3 Running title: ZnT8 regulates peripheral serotonin production
4
5 Zhuo Mao1, Hui Lin1, Wen Su1, Jinghui Li1, Minsi Zhou1, Zhuoran Li1, Beibei
6 Zhou2, Qing Yang1, Mingyan Zhou1, Ke Pan2, Jinhan He3, * and Weizhen Zhang 1,4, *
7
8 1. Center for Diabetes, Obesity and Metabolism, Department of Physiology,
9 Shenzhen University Health Science Center, Shenzhen, Guangdong province, China.
10 2. Institute for Advanced Study, Shenzhen University, Shenzhen, Guangdong
11 province, China.
12 3. Department of Pharmacy, State Key Laboratory of Biotherapy and Cancer
13 Center, West China Hospital of Sichuan University, Chengdu, Sichuan, China.
14 4. Department of Physiology and Pathophysiology, School of Basic Science,
15 Peking University Health Science Center, Beijing, China.
16
17 Correspondence to
18 Dr. Weizhen Zhang
19 Center for Diabetes, Obesity and Metabolism, Department of Physiology,
20 Shenzhen University Health Science Center, Shenzhen, Guangdong province, China.
21 Department of Physiology and Pathophysiology, School of Basic Science, Peking
22 University Health Science Center, Beijing, China.
Diabetes Publish Ahead of Print, published online April 1, 2019 Diabetes Page 2 of 57
23 Email: [email protected]
24 Tel: +86-15010909001
25
26 Dr. Jinhan He
27 Department of Pharmacy, State Key Laboratory of Biotherapy and Cancer Center,
28 West China Hospital of Sichuan University, Chengdu, Sichuan, China.
29 E-mail: [email protected]
30 Tel.: +86-28-85426416
31
32 Word count: 7071
33 Figures: 8 (color)
34 Supplemental Figures: 11
35 Supplemental Tables: 1 Page 3 of 57 Diabetes
37 Abstract
38 ZnT8 is a zinc transporter enriched in the pancreatic beta cells and its
39 polymorphism is associated with increased susceptibility to type 2 diabetes. However,
40 the exact role of ZnT8 in systemic energy metabolism remains elusive. In this study,
41 we found that ZnT8 knockout mice displayed increased adiposity without obvious
42 weight gain. We also observed the intestinal tract morphology, motility and gut
43 microbiota were changed in ZnT8 knockout mice. Further study demonstrated that
44 ZnT8 was expressed in enteroendocrine cells, especially in 5-HT positive
45 enterochromffin cells. Lack of ZnT8 resulted in an elevated circulating 5-HT level
46 owing to enhanced expression of tryptophan hydroxylase 1. Blocking 5-HT synthesis
47 in ZnT8 deficient mice restored adiposity, high-fat diet-induced obesity and glucose
48 intolerance. Moreover, overexpression of human ZnT8 diabetes high risk allele
49 R325W increased 5-HT levels relative to the low risk allele in RIN14B cells. Our
50 study revealed an unexpected role of ZnT8 in regulating peripheral 5-HT biogenesis
51 and intestinal microenvironment, which might contribute to the increased risk of
52 obesity and type 2 diabetes.
53
54 Key words: ZnT8; 5-HT; obesity; adiposity; type 2 diabetes, intestinal hormones;
55 energy metabolism Diabetes Page 4 of 57
57 ZnT8 is a zinc transporter which is closely associated with both type 1 and type 2
58 diabetes mellitus (DM). It is an important autoantigen in T1DM patients [1].
59 Meanwhile its gene polymorphism has also been identified as a risk factor for T2DM
60 [2], suggesting an important physiological function of ZnT8 in metabolic disease
61 progression. ZnT8 is highly abundant in pancreatic beta cells [3]. Several colonies of
62 global and beta cell or alpha cell specific ZnT8 knockout mice have been generated to
63 investigate its effects on insulin granule morphology, insulin secretion and systemic
64 glucose metabolism [4-6]. Global ZnT8 knockout mice exhibit an exacerbation of
65 diet-induced obesity and glucose intolerance compared to wild-type mice [4, 6].
66 Unexpectedly, this phenotype was not observed in mice that lack ZnT8 specifically in
67 beta cells or alpha cells [6]. This discrepancy strongly implies the presence of ZnT8 in
68 non-beta/alpha cell or extrapancreatic tissues plays a critical role in organism energy
69 homeostasis. Since ZnT8 is negligible in hypothalamus, fat and skeletal muscle [6],
70 we therefore speculate that ZnT8 may be expressed in other endocrine tissues or cells.
71 Gastrointestinal (GI) tract contains the largest number of various endocrine cells.
72 Many GI hormones play critical roles in glucose homeostasis [7].
73 5-hydroxytryptamine (5-HT, serotonin) is the most prevalent GI hormone which
74 exhibits both central and peripheral functions. More than 90% of 5-HT is synthesized
75 in and released from the enterochromaffin cells (ECCs). Other tissues, such as
76 neurons, adipose tissues and pancreas, only produces a small amount of 5-HT [8]. The
77 initial and rate-limiting step of 5-HT synthesis is catalyzed by the tryptophan
78 hydroxylase (TPH). There are two isoforms of TPHs, TPH1 in the peripheral tissues Page 5 of 57 Diabetes
79 and TPH2 in the central nervous system (CNS) and enteric neurons [9]. 5-HT system
80 possesses complex bioactivities mediated by different types of 5-HT receptor (5HTR)
81 expressed in various tissues. In the CNS, 5-HT acts as a neurotransmitter to regulate
82 appetite, emotions, sleep, and systemic metabolism through sympathetic nervous
83 system (SNS) [10]. The peripheral 5-HT function is relatively less clear compared to
84 its central role. The classical action of peripheral 5-HT includes regulation of GI
85 functions such as motility, secretion, sensation, modulation of platelet coagulation and
86 bone density [11]. Interestingly, obesity increases peripheral 5-HT level [12]. And
87 genetic polymorphism of TPH1 is associated with obesity [13]. Recent studies have
88 found that peripheral 5-HT promotes white adipose tissue (WAT) lipogenesis and
89 inhibits brown adipose tissues (BAT) adaptive thermogenesis [14, 15]. Genetic
90 deficiency or pharmacological inhibition of 5-HT synthesis enzyme TPH1 in mice
91 leads to a resistance to diet-induced obesity and glucose intolerance [14, 15],
92 suggesting that peripheral 5-HT is an important regulator of lipid metabolism and
93 systemic energy homeostasis.
94 In this study, we generated a new strain of ZnT8 knockout mice using TALEN
95 technology. We examined the presence of ZnT8 in enteroendocrine cells (EECs) and
96 its role in 5-HT biogenesis and lipid metabolism using cell biological and transgenic
97 techniques. We also observed an unexpected change in colon morphology, function
98 and microbiota in ZnT8 deficient mice which may contribute to the increased
99 sensitivity of diet-induced obesity and T2DM. Diabetes Page 6 of 57
100 Research Design and Methods
101 Animals
102 ZnT8 knockout mice (Slc30a8-/- mice) were generated by Cyagen Biosciences
103 Inc. (Guangzhou, China). Exon 3 of Slc30a8 gene was selected as the target site.
104 TALEN mRNAs generated by in vitro transcription were then injected into fertilized
105 eggs from C57BL/6N strain for knockout mouse production. The founders were
106 genotyped by PCR followed by DNA sequencing analysis. The positive founders
107 were breeding to the next generation which was genotyped by PCR and DNA
108 sequencing analysis.
109 For high-fat diet (HFD) treatment experiment, male mice (aged 6-8 weeks old)
110 were fed either a normal chow diet (ND) or an HFD (45 % fat calories, Research
111 Diets D12451 or 60 % fat calories, Research Diets D12492). For TPH inhibitor
112 injection experiment, PBS or 4-Chloro-DL-phenylalanine methyl ester hydrochloride
113 (PCPA, C3635, Sigma-Aldrich) (300 mg/kg BW) was administered as a daily
114 intraperitoneal injection. All animal experiments were undertaken with the approval
115 of the Scientific Investigation Board of Health Science Center of Shenzhen University
116 (Shenzhen, Guangdong Province, China).
117 Antibody
118 The rabbit anti-ZnT8 polyclonal antibody was generated against the synthetic
119 mouse ZnT8 peptide (KPVNKDQCPGDRPEHPEAGGIYH, 29-51 aa). The
120 antibodies against insulin, chromogranin A, UCP1, Beta3 AR, GIP and GLP-1 were
121 from Abcam. Anti TH antibody was from Millipore. The tubulin, actin and GAPDH Page 7 of 57 Diabetes
122 antibodies were from Proteintech. The Alexa Fluor 488 and Alexa Fluor 594 dye were
123 obtained from Molecular Probes. The horseradish peroxidase (HRP)-labeled
124 secondary antibody was purchased from Amersham BioSciences (GE Healthcare).
125 Histology and Immunofluorescent staining
126 The dissected tissues were fixed with 4% paraformaldehyde (PFA) in PBS for 16
127 hrs at 4°C. The samples were sequentially dehydrated and embedded in paraffin. Then
128 the tissue samples were sectioned at a 6-μm thickness and were used for standard
129 H&E staining and quantification. The quantification was determined by the Image J
130 software. Immunochemistry/immunofluorescent staining were performed following
131 general protocols. Images were obtained by Nikon Eclipse Ti microscope.
132 Oil Red O staining and PAS staining
133 The frozen liver sections were washed in PBS once and were fixed with 4% PFA
134 in PBS for 15 min at room temperature and then washed three times with PBS. The
135 sections were incubated in the 60% isopropyl alcohol and then stained with filtered
136 Oil Red O solution (1.5 mg/mL) for 30 min and rinsed twice with distilled water.
137 Periodic Acid Schiff （PAS）staining for glycogen in the liver and goblet cells in
138 the colon were performed using the commercial kit (Solarbio, Beijing) following the
139 manual instruction.
140 Oral/intraperitoneal glucose tolerance test (OGTT/IPGTT)
141 ZnT8 group mice were first starved for 16 hrs, followed by an oral glucose
142 infusion or intraperitoneal glucose injection (1.5 g/kg body weight). The blood
143 glucose levels were measured from the tail vein before and at 15, 30, 60, 90 and 120 Diabetes Page 8 of 57
144 min after injection using a glucometer (Accu-check, Roche).
145 Insulin tolerance test (ITT)
146 ZnT8 group mice were fasted for 4 hrs prior to the ITT test and then received an
147 injection of human regular insulin (0.5 U/kg body weight). Blood glucose levels were
148 recorded before and at 15, 30, 60, 90 and 120 min after injection using a glucometer
149 (Accu-check, Roche).
150 Blood biochemistry and ELISA
151 The serum total triglyceride and total cholesterol were measured by using
152 commercial kits (Biosino) and NEFA were measured using the Wako kit (Wako
153 Chemicals). Serum 5-HT levels were measured using the mouse ELISA kit (ENZO
154 Life Sciences). Insulin levels were analyzed using the mouse ultrasensitive insulin kit
155 (Alpco) and catecholamine levels were measured using the ELISA kits (Biovision)
156 according to the manufacturer’s instruction.
157 Cell culture and transfection
158 The RIN14B cells (Obtained from Zeye Bio-Tech, Shanghai) were cultured in
159 high-glucose DMEM medium (Gibco) containing 10% fetal bovine serum (FBS,
160 Gibco) in a humidified incubator with 5% CO2 at 37 °C. Cells were plated at optimal
161 densities and grown for 24 h, then transfected with plasmids using Lipofectamine
162 3000 (Thermo Fisher Scientific, US) according to the manufacturer's instruction.
163 Total gastrointestinal transit time
164 A solution of 6% carmine red (300 μL, Sigma-Aldrich) suspended in 0.5%
165 methylcellulose (Sigma-Aldrich) was administered by gavage through a 21-gauge Page 9 of 57 Diabetes
166 round-tip feeding needle. The time at which gavage took place was recorded as T0.
167 After gavage, fecal pellets were monitored at 10-minute intervals for the presence of
168 carmine red. Total GI transit time was considered as the interval between T0 and the
169 time of first observance of carmine red in stool.
170 Fecal microbiota sequencing
171 Stool samples freshly collected from each mouse were immediately frozen at −20
172 °C and transported to the laboratory with ice pack. Metagenomic sequencing and
173 analysis were performed by Novogene Bioinformatics Technology Co., Ltd (Beijing,
174 China).
175 Metabolic cage studies and body composition
176 Measurements of energy expenditure, respiratory exchange ratio, indirect
177 calorimetry, and physical activity using metabolic cages (Columbus Instruments)
178 were performed by the Biomedical Research Institute of Nanjing University (Nanjing,
179 China). Whole body compositions of ZnT8 group mice were analyzed by Echo
180 MRITM.
181 DTZ staining
182 The isolated islets were incubated in the 0.1 mg/mL dithizone (DTZ) (Sigma)
183 solution at 37°C for 15 mins. After washing with HBSS, the islets were examined
184 with the stereo-microscope.
185 Western blotting
186 The tissues or RIN14B cells were quickly harvested, rinsed with cold PBS and
187 homogenized in cold RIPA buffer (150 mM NaCl, 1% Triton X-100, 1% sodium Diabetes Page 10 of 57
188 deoxycholate, 0.1% SDS, 50 mM Tris-HCl, and 2 mM EDTA, pH 7.4) supplemented
189 with protease inhibitor cocktail (Roche). A total of 40 μg proteins were loaded onto
190 SDS-PAGE gels and electrophoretically transferred to polyvinylidene fluoride
191 membranes (Bio-Rad). Transferred membranes were blocked with 5% non-fat milk in
192 TBS/0.1% Tween-20 (TBS-T), then incubated with primary antibodies at 4℃
193 overnight. After washing with TBS/T, membranes were incubated with secondary
194 antibodies and developed with SuperSignal™ West Pico Chemiluminescent Substrate
195 (Thermo). Signals were detected by the Amersham Imager 600 (GE).
196 RNA extraction and qRT-PCR
197 Total RNAs were isolated from mouse tissues or RIN14B cells using Direct-zol
198 RNA miniprep (ZYMO Research, R2052). 1 μg of total RNAs was used for reverse
199 transcription using PrimeScriptTM RT Master Mix (Takara). SYBR Green-based
200 real-time PCR was performed using Light Cycler 96 (Roche) with SYBR Premix Ex
201 TaqⅡ(Takara). The quantity of mRNA was calculated using the ΔΔ Ct method. All
202 reactions were performed as duplicates. The primers used for qPCR were listed
203 in supplementary Table 1.
204 Statistical analysis
205 All the results are presented as mean ± SEM. Data were analyzed with Student’s t
206 test or one-way ANOVA followed by Bonferroni’s multiple comparison test. P value
207 < 0.05 was considered as statistically significant.
208 Results
209 Slc30a8-/- mice display increased adiposity. Page 11 of 57 Diabetes
210 In order to investigate the mechanism by which ZnT8 increases T2DM sensitivity,
211 we generated ZnT8 knockout mice (Slc30a8-/- mice) on the C57BL/6N genetic
212 background using the TALEN technology (Figure S1A) by deleting 2 bp in exon 3.
213 We verified the newly generated ZnT8 knockout line by analyzing the pancreatic
214 expression and function of ZnT8. Loss of ZnT8 mRNA and protein expression were
215 first confirmed in ZnT8 knockout pancreas (Fig S1B, C). Then DTZ staining of
216 isolated islets from Slc30a8-/- mice demonstrated a marked reduction of intensity
217 representing zinc depletion, indicating the functional inactivity of ZnT8 protein in
218 pancreatic islets (Figure S1D).
219 The overall morphology and growth curve were similar between wild-type and
220 Slc30a8-/- mice (Fig 1A). However, we observed a significant increase in the fat mass
221 in both male and female Slc30a8-/- mice while the lean mass showed a trend of
222 reduction (Fig 1B). The WAT mass, including epididymal WAT (eWAT) and
223 subcutaneous WAT (scWAT), were significantly increased in Slc30a8-/- mice (Fig.
224 1C-D). This alteration was resulted from the increased adipocyte size (Fig 1E-F).
225 Further, the expression levels of genes related to lipid synthesis, fatty acid uptake, as
226 well as lipolysis were significantly increased, whereas lipogenesis and fatty acid
227 oxidation relevant genes remained unchanged (Fig 1G-K).
228 In addition to WAT expansion, there was an increase of lipid accumulation in
229 brown adipose tissue (BAT) evidenced by increased mass and unilocular fat droplets
230 in ZnT8 knockout mice (Fig 2A, B). Most genes related to thermogenesis, including Diabetes Page 12 of 57
231 Ucp1 and Pgc1α, remained unchanged (Fig 2G, H). The lipogenesis, lipid synthesis
232 genes were increased while the fatty acid oxidation related genes were decreased (Fig
233 2C-F). Unexpectedly, we found a significant reduction of β3 adrenergic receptor (β3
234 AR) protein level in BAT (Fig 2H). The neuronal marker tyrosine hydroxylase (TH)
235 remained unchanged, suggesting an intact neuron innervation to the BAT (Fig 2H).
236 Total sympathetic tone was normal as measured by serum catecholamine (included
237 dopamine, norepinephrine, and epinephrine) (Figure S2).
238 Liver is another important organ for lipid metabolism. Although previous study
239 did not detect excessive lipid or glycogen deposition in Slc30a8-/- liver [16], we
240 found that the lipid and glycogen contents were significantly increased in our ZnT8
241 knockout liver (Fig 3A-C). The hepatic TG level but not TC level was also elevated
242 (Fig 3D, E). All these data suggest that ZnT8 deficiency increases fat accumulation in
243 adipose tissues and liver, which may contribute to increased propensity to metabolic
244 derangement under stress conditions.
245 Increased adiposity results from increased energy intake or decreased energy
246 expenditure. Interestingly, metabolic profiling of ZnT8 littermate mice showed that
247 the food consumption and energy expenditure characterized by oxygen consumption,
248 total activity and heat generation remained unchanged (Fig S3). Moreover, the serum
249 inflammation markers (Fig S4) and most obesity/diabetes-related hormones (Fig S5)
250 remained unaltered in ZnT8 knockout mice.
251 We also examined glucose homeostasis. The fasting blood glucose and insulin Page 13 of 57 Diabetes
252 tolerance were comparable between these two genotypes of mice (Fig 4A, B).
253 Unexpectedly, the oral glucose tolerance was significantly impaired in Slc30a8-/-
254 mice, whereas the intraperitoneal glucose tolerance remained unaltered (Fig 4C, D).
255 This result suggested a possible involvement of gastrointestinal (GI) tract in lipid and
256 glucose derangement in Slc30a8-/- mice.
257 ZnT8 knockout mice have altered intestinal morphology and motility.
258 When we carefully examined the GI tract, we found that the intestine tract from
259 ZnT8 knockout mice was significantly thickened, especially the proximal colon part
260 (Fig 5A, B), while no differences of the whole GI tract length was observed (Fig S6A,
261 B). Proliferation analysis of PCNA staining showed a remarkable increased number
262 and intensity of positive stained cells (Fig 5C). And TUNEL assay for apoptosis
263 identified a smaller number of apoptotic cells in the mucosal layer of ZnT8 knockout
264 mice (Fig 5D). These results suggested that ZnT8 deficiency promoted proliferation
265 and inhibited apoptosis in ZnT8 knockout colon. Interestingly, we also observed an
266 increased volume of mucosal goblet cells and observed a significant increase of the
267 enzyme carbonic anhydrase 1 (CA1) in ZnT8 knockout colon (Fig 5E, F). CA1 is a
268 zinc metalloenzyme which catalyzes the reversible hydration of carbon dioxide.
269 Increased CA1 may disturb the colonic acid-base balance in ZnT8 knockout mice. In
270 addition to the morphology and biochemical changes, ZnT8 knockout mice also had a
271 slower intestinal motility under physiological condition (Fig 5G).
272 Gut microbiota is an important factor regulating intestinal tract microenvironment Diabetes Page 14 of 57
273 and energy homeostasis. Therefore, we assessed the effects of ZnT8 deficiency on the
274 gut microbiota composition by sequencing the fecal bacterial 16S rRNA V3+V4
275 region. ZnT8 deficiency significantly decreased richness and diversity of gut
276 microbiota indicated by decreased OTU number and the observed species (Fig
277 S7A-D). However, the microbiota change pattern was a protective pattern for host
278 metabolism, with increased relative abundance of Bacteroidetes and Verrucomicrobia,
279 together with reduced ratios of Proteobacteria and Deferribacteres. The Firmicutes
280 abundance was not significantly changed (Fig S7E, F). Taken together, these results
281 suggested that deficiency of ZnT8 resulted in a significant change in colon
282 morphology, biochemical environment, motility as well as microbiota.
283 ZnT8 is expressed in enteroendocrine cells and regulates peripheral 5-HT
284 levels.
285 Next, we sought to examine whether ZnT8 was present in the intestinal tract by
286 immunofluorescent staining. ZnT8 immunoreactivity was observed in the epithelial
287 layer, but not in the smooth muscle layer or lamina propria in the intestine. The
288 positively stained cells were triangular and scattered in the enterocytes, mainly in
289 proximal colon (Fig 6A). We co-stained EEC marker chromogranin A (CgA) and
290 other hormone markers with ZnT8 antibody. ZnT8 signal was observed mainly in
291 cells positive for CgA and 5-HT (Fig 6 B, S8A), whereas a small proportion of ZnT8
292 co-localized with GIP (gastric inhibitory polypeptide) (Fig S8B). We did not observe
293 a colocalization between ZnT8 and GLP-1 (glucagon-like peptide 1) or PYY Page 15 of 57 Diabetes
294 (polypeptide YY) in the intestine of wild-type mice. And no ZnT8 signal was
295 observed in the intestinal tract of ZnT8 knockout mice (Fig S8A).
296 5-HT is the most prevalent GI hormones and GI tract is the major source for
297 peripheral 5-HT. More than 90% of circulating 5-HT is synthesized and secreted from
298 GI tract [17]. Notably, the staining intensity and number of 5-HT cells in the colon
299 were remarkably increased in ZnT8 knockout mice (Fig 6C, D). The serum 5-HT
300 level in ZnT8 knockout mice was also significantly higher than wild-type mice (Fig
301 6E). TPH1 is responsible for synthesizing 5-HT in the peripheral tissue. Both mRNA
302 and protein levels of TPH1 in the proximal colon were significantly increased (Fig 6F,
303 G). These results suggested that ZnT8 was expressed in the enteroendocrine cells and
304 deficiency of ZnT8 increased peripheral 5-HT by elevating TPH1 protein level.
305 Reversal of metabolic dysfunction by inhibition of 5-HT synthesis in ZnT8
306 deficient mice
307 To determine whether peripheral 5-HT contributes to the metabolic effects of
308 ZnT8 deficiency, we treated ZnT8 group mice with the specific TPH inhibitor,
309 p-chlorophenylalanine (PCPA). Daily injection of PCPA for 4 weeks attenuated the
310 increase in body weight and oral glucose intolerance, which was markedly elevated in
311 Slc30a8-/- mice fed with ND (Fig 7A, B). ZnT8 knockout mice fed with HFD showed
312 a significant increase in body weight, fat mass and hepatic steatosis compared to the
313 wild-type mice (Fig S9A, B). The hyperglycemia, hyperinsulinemia and
314 hyperlipidemia were also observed (Fig S9C-F). PCPA effectively rescued the Diabetes Page 16 of 57
315 diet-induced obese and glucose intolerance phenotypes in Slc30a8-/- mice. It
316 prevented the increase in body weight, fat mass and hepatic steatosis, as well as
317 glucose intolerance (Fig 7C-F, Fig S10 A-F). PCPA demonstrated negligible effects
318 on food intake in wild-type or Slc30a8-/- mice whether they were fed with ND or
319 HFD (Fig 7G). These results suggested that reducing peripheral 5-HT level was able
320 to, at least partially, rescue the increase in adiposity and related glucose intolerance in
321 ZnT8 deficient mice fed with either ND or HFD.
322 ZnT8 deficiency increases 5-HT biosynthesis through elevating TPH1 level
323 in RIN14B cells.
324 To investigate how ZnT8 regulates 5-HT level, we used RIN14B cells as the in
325 vitro cellular model. RIN14B cell is a rat delta cell line and has been used for
326 studying 5-HT secretion as a proper cellular model [18]. Delta cells are known to
327 produce somatostatin and are expressed in the pancreatic islets, as well as intestine.
328 We detected the presence of ZnT8 mRNA by Q-PCR analysis in RIN14B cells. When
329 endogenous ZnT8 was knocked down by siRNA transfection (Fig 8A), the TPH1
330 protein level was increased almost 3 folds (Fig 8B). ELISA analysis showed that
331 silence of ZnT8 significantly increased 5-HT level in the cell lysate as well as in the
332 medium (Fig 8C, D). These results suggested that ZnT8 deficiency increased the
333 expression of TPH1 and hence the synthesis of 5-HT.
334 ZnT8 mediates the zinc transport across cell membranes [19]. We therefore
335 analyzed whether altered zinc concentration affected levels of TPH1 and 5-HT. Page 17 of 57 Diabetes
336 Addition of ZnSO4 in the medium significantly increased the expression level of
337 TPH1 and 5-HT in the medium. Depletion of zinc ions by zinc chelator
338 N,N,N’,N’-tetrakis (2-pyridinylmethyl)-1,2-ethanediamine (TPEN) reduced TPH1
339 mRNA and protein level, although 5-HT secretion in the medium was not
340 significantly changed (Fig 8E-G). These results suggested that ZnT8 deficiency
341 increased TPH1 level probably by changing the intracellular level of zinc ion.
342 Genomic studies have identified that the risk C allele (rs13266634) of ZnT8 gene
343 which encodes an arginine (R) in place of a tryptophan (W), is associated with T2DM
344 and body mass index [20, 21]. Recent studies found that T2DM-risk R325 ZnT8
345 variant had a higher zinc transport activity [22] and human islets with the R325 ZnT8
346 variant had a higher zinc content [23]. Therefore, we tested whether the human T2DM
347 risk allele altered 5-HT biosynthesis. We generated human ZnT8 wild type and
348 mutant (R325W) plasmids and transfected them into RIN14B cells. Interestingly,
349 ZnT8 R325W mutant overexpression significantly increased 5-HT level in the
350 medium (Fig 8I), but no change of 5-HT or TPH1 level in the cell lysate was detected
351 (Fig 8H, Fig S11). These results suggested that the diabetes-risk ZnT8 allele could
352 affect 5-HT level but with a distinctive mechanism.
353 Discussion
354 Our present study demonstrates that ZnT8 regulates 5-HT biogenesis which is
355 critical for organism lipid and energy metabolism (Fig 8J). This conclusion is
356 supported by following observations: (1) Deficiency of ZnT8 increased lipid Diabetes Page 18 of 57
357 accumulation in adipose tissues and liver without obvious body weight gain. (2) ZnT8
358 was detected in the EEC cells of the gastrointestinal tract. (3) Deficiency of ZnT8
359 increased peripheral 5-HT level through elevating TPH1 protein level. (4)
360 Pharmacological inhibition of 5-HT synthesis effectively reversed diet-induced body
361 weight increase and glucose intolerance in ZnT8 knockout mice. (5) In RIN14B cell,
362 down-regulation of ZnT8 promoted 5-HT synthesis and secretion. Interestingly,
363 overexpression of diabetes high-risk allele of human SLC30A8 increased 5-HT
364 secretion in the medium. These effects could be mimicked by administration of zinc
365 ion.
366 ZnT8, a transmembrane protein mediating the transport of zinc ion, is mainly
367 expressed in pancreatic islets. Its physiological function has been related to the
368 secretion of islet hormones and glucose homeostasis [24]. Genomic analyses have
369 shown that the risk C allele (rs13266634) of ZnT8 gene is associated with body mass
370 index and T2DM [2, 20, 21, 25]. The metabolic function of ZnT8 has long been
371 proposed to occur through its regulation on the secretion of islet hormones,
372 specifically insulin. This concept has been recently challenged by studies using
373 transgenic techniques. Although global deletion of ZnT8 leads to significant
374 diet-induced obesity and impaired glucose tolerance, the exacerbation was not
375 observed in beta cell specific ZnT8 knockout mice [6]. These results indicate an
376 alternative mechanism underlying the regulation of glucose metabolism by ZnT8. Our
377 study revealed that ZnT8 contributed to the modulation of 5-HT biosynthesis in the
378 intestinal EEC cells, which subsequently altered systemic glucose and lipid Page 19 of 57 Diabetes
379 metabolism. Deficiency of ZnT8 led to a significant increase in colonic TPH1 and
380 circulating 5-HT level. Since over 90% of circulating 5-HT is derived from the GI
381 tract, we assume that the elevation of circulating 5-HT is primarily resulted from
382 deficiency of ZnT8 in the gut rather than other tissues such as adipose tissue and
383 enteric neurons.
384 The influence of ZnT8 on lipid metabolism remains largely unexplored.
385 Deficiency of ZnT8 in mice leads to an increase of adiposity under normal chow diet,
386 which is profoundly aggravated under high-fat diet feeding. Our observations
387 suggested that elevated circulating 5-HT contributed to the obesity and glucose
388 dysfunction. Consistently, recent studies have demonstrated that gut-derived 5-HT
389 inhibits lipolysis and thermogenesis in eWAT and BAT, promotes hepatic glucose
390 production, leading to obesity and dysglycemia [14, 15]. Our study further confirmed
391 that blockade of TPH reversed the increased adiposity and glucose intolerance in
392 ZnT8 knockout mice. In our ZnT8 deficient mice, we also observed an increase in
393 lipid and glycogen deposition in the liver, although Tamaki et al [16] did not show
394 significant changes in liver. Previously, several groups have reported that ZnT8
395 deficient mice have substantial hypersecretion of insulin from pancreatic beta cells [5,
396 16]. It is reasonable to speculate that the insulin directly flows into the liver through
397 portal vein, causing lipid and glycogen deposition in the liver. The discrepancy may
398 be due to the different mouse model with different genetic background or other
399 environmental factors, causing distinct liver sensitivity to the elevated insulin. Diabetes Page 20 of 57
400 Up to now, there are six ZnT8 knockout mice studies [4-6, 26-28]. There was a
401 discrepancy of metabolic phenotypes, including glucose tolerance, insulin sensitivity
402 and insulin secretion capability, under either ND or HFD conditions. Other factors
403 must influence the effect of Slc30a8 deletion. Environmental conditions, such as diets
404 and intestinal microbial composition, might be important. In our ZnT8 knockout
405 mouse model, we observed a significant change in intestinal tract morphology and
406 motility change. The altered gut microenvironment could lead to increased harvest
407 from the diet, altered fatty acid metabolism and modulation of gut 5-HT secretion,
408 and modulation of intestinal barrier. Moreover, intestinal 5-HT production is also
409 affected by gut microbes [29, 30]. Until now, we still do not know whether it is a
410 direct effect of intestinal ZnT8 on 5-HT regulation or an indirect effect caused by the
411 disturbed gut microenvironment or microbiota. Moreover, we cannot exclude the
412 existence and effects of ZnT8 in other tissues, such as immune cells, central nervous
413 system or enteric nervous system, which may contribute to the disturbed 5-HT
414 biogenesis and systemic metabolic dysregulation. To exclude these confounding
415 factors, the intestinal specific ZnT8 deletion mouse could be a better model for
416 investigating the role of ZnT8 in the intestinal tract.
417 There are several plausible mechanisms through which intestinal ZnT8 regulate
418 circulating levels of 5-HT. First, ZnT8 regulates zinc level within EEC cells. Zinc is a
419 critical co-factor for more than 300 proteins and enzymes. It plays essential roles in
420 various biochemical processes including tryptophan metabolism [31]. ZnT8 is a zinc
421 transporter which mediates zinc export from the intracellular to the extracellular or Page 21 of 57 Diabetes
422 cellular organelles [19]. Previous studies reported that ZnT8 is expressed on the
423 plasma membrane and may mediate bi-directional transport of zinc ions [32].
424 Ablation of ZnT8 alters intracellular or vesicular zinc level in beta cells [5]. We
425 speculate that a similar increase of cytosolic zinc level occurs in EEC cells after ZnT8
426 deletion, and hence stimulating the 5-HT biosynthesis. In support of this concept,
427 TPH1 is significantly increased in intestinal mucosa of ZnT8 knockout mice and
428 silence of ZnT8 gene increases 5-HT levels in RIN14B cells. Another potential
429 mechanism relates to the secretion of 5-HT granules. Lack of zinc ion in the granule
430 may lead to a looser crystalline structure which probably affects the storage and
431 exocytosis process of 5-HT.
432 Although a number of ZnT8 knockout rodent models consistently displayed
433 increased sensitivity of T2DM and impaired glucose insulin secretion in vivo [33].
434 Several loss-of-function mutations of SLC30A8 in human confer protection to human
435 T2DM [34]. Later transgenic mouse models which harbor the human ZnT8 truncation
436 mutant increased insulin secretion [35]. Increased 5-HT level may account for this
437 discrepancy. 5-HT increases insulin secretion via modulating insulin granule
438 exocytosis [36, 37]. During pregnancy, the expression of islet 5-HT is profoundly
439 increased and regulates pancreatic beta cell mass and increased glucose induced
440 insulin secretion [38, 39]. Upon the metabolic stress condition such as the high-fat
441 diet feeding, 5-HT is also significantly increased [12] and enhances glucose induced
442 insulin secretion [40]. Consequently, elevated insulin level leads to lipid deposition in
443 the adipose tissue and lipid/glycogen production in liver. Diabetes Page 22 of 57
444 Interestingly, overexpression of diabetes high-risk allele ZnT8 R325 variant
445 increased 5-HT secretion in RIN14B cells. However, TPH1 protein level and 5-HT in
446 the cell lysate were not significantly changed. This suggests that different from total
447 loss of function in ZnT8 deficiency, the mutant form of human ZnT8 may interfere
448 with the exocytosis rather than the biosynthesis of 5-HT. The exact mechanism by
449 which diabetes-risk ZnT8 mutant alters the exocytosis process still needs further
450 investigation.
451 Normal weight obesity means higher fat mass but normal body mass index (BMI).
452 These groups of patients have received more and more attention since they have
453 similar risks for serious metabolic disorders as obese people, but they are easily
454 ignored by the clinicians or themselves. Interestingly, two recent genome-wide
455 association studies have shown that the human SLC30A8 risk allele confers higher
456 risk of developing T2DM in a lower BMI population or non-obese subjects [21, 41].
457 These findings suggest that ZnT8 and 5-HT level are important factors that determine
458 the percentage of body fat regardless of body weight. Our findings further support this
459 concept. ZnT8 knockout mice demonstrated a significant increase of fat mass without
460 body weight gain, which was reminiscent of normal weight obesity in human. Since
461 the lean mass remained unchanged, one possibility of increased fat mass without body
462 weight gain is due to loss of bone density as 5-HT elevation has been known to cause
463 osteoporosis [42].
464 In summary, our study identified the presence of ZnT8 in the 5-HT positive EEC Page 23 of 57 Diabetes
465 cells. Deficiency of ZnT8 enhanced 5-HT biosynthesis through modulating TPH1
466 level. GI-derived 5-HT acted on the peripheral organs to promote lipid deposition and
467 obesity. Targeting intestinal ZnT8 may provide an alternative strategy for the
468 intervention of obesity and its associated metabolic dysfunction such as hepatic
469 steatosis and T2DM.
470
471 Acknowledgement and Funding
472 We acknowledge Mr. Chaowei Zhu, Ms. Weiqi Wu and Ms. Ruolu Bao for
473 technical support. This work was funded by the National Key R&D Program of China
474 (2017YFC0908900), the National Natural Science Foundation of China (81500619，
475 81730020, 81870405), the Natural Science Foundation of Guangdong Province
476 (2016A030310040), the Shenzhen Science and Technology Project
477 (JCYJ20160422091658982, JCYJ20160422153856130).
478
479 Author contributions
480 Z.M., J.H. and W.Z. developed the study rationale and wrote the manuscript. Z.M.
481 designed and performed most of the experiments. H.L., W.S., J.L., M.Z., Z.L., B.Z.
482 and K.P. performed the experiments and assisted with data analysis. Z.M. and W.Z.
483 supervised the study.
484 W.Z. serves as the guarantor. Diabetes Page 24 of 57
485
486 Competing interests
487 The authors declare no competing interests.
488 Page 25 of 57 Diabetes
490 References
491 1. Wenzlau, J.M., et al., The cation efflux transporter ZnT8 (Slc30A8) is a major autoantigen in 492 human type 1 diabetes. Proc Natl Acad Sci U S A, 2007. 104(43): p. 17040-5. 493 2. Sladek, R., et al., A genome-wide association study identifies novel risk loci for type 2 diabetes. 494 Nature, 2007. 445(7130): p. 881-5. 495 3. Chimienti, F., et al., Identification and cloning of a beta-cell-specific zinc transporter, ZnT-8, 496 localized into insulin secretory granules. Diabetes, 2004. 53(9): p. 2330-7. 497 4. Lemaire, K., et al., Insulin crystallization depends on zinc transporter ZnT8 expression, but is 498 not required for normal glucose homeostasis in mice. Proc Natl Acad Sci U S A, 2009. 106(35): 499 p. 14872-7. 500 5. Nicolson, T.J., et al., Insulin storage and glucose homeostasis in mice null for the granule zinc 501 transporter ZnT8 and studies of the type 2 diabetes-associated variants. Diabetes, 2009. 502 58(9): p. 2070-83. 503 6. Hardy, A.B., et al., Effects of high-fat diet feeding on Znt8-null mice: differences between 504 beta-cell and global knockout of Znt8. Am J Physiol Endocrinol Metab, 2012. 302(9): p. 505 E1084-96. 506 7. Moran, G.W., et al., Enteroendocrine cells: neglected players in gastrointestinal disorders? 507 Therap Adv Gastroenterol, 2008. 1(1): p. 51-60. 508 8. Spohn, S.N. and G.M. Mawe, Non-conventional features of peripheral serotonin signalling - 509 the gut and beyond. Nat Rev Gastroenterol Hepatol, 2017. 14(7): p. 412-420. 510 9. Zhang, X.D., et al., Tryptophan hydroxylase-2 controls brain serotonin synthesis. Science, 511 2004. 305(5681): p. 217-217. 512 10. Berger, M., J.A. Gray, and B.L. Roth, The expanded biology of serotonin. Annu Rev Med, 2009. 513 60: p. 355-66. 514 11. Mawe, G.M. and J.M. Hoffman, Serotonin signalling in the gut--functions, dysfunctions and 515 therapeutic targets. Nat Rev Gastroenterol Hepatol, 2013. 10(8): p. 473-86. 516 12. Bertrand, R.L., et al., A Western diet increases serotonin availability in rat small intestine. 517 Endocrinology, 2011. 152(1): p. 36-47. 518 13. Kwak, S.H., et al., Association of Variations in TPH1 and HTR2B with Gestational Weight Gain 519 and Measures of Obesity. Obesity, 2012. 20(1): p. 233-238. 520 14. Crane, J.D., et al., Inhibiting peripheral serotonin synthesis reduces obesity and metabolic 521 dysfunction by promoting brown adipose tissue thermogenesis. Nat Med, 2015. 21(2): p. 522 166-72. 523 15. Oh, C.M., et al., Regulation of systemic energy homeostasis by serotonin in adipose tissues. 524 Nat Commun, 2015. 6: p. 6794. 525 16. Tamaki, M., et al., The diabetes-susceptible gene SLC30A8/ZnT8 regulates hepatic insulin 526 clearance. J Clin Invest, 2013. 123(10): p. 4513-24. 527 17. Namkung, J., H. Kim, and S. Park, Peripheral Serotonin: a New Player in Systemic Energy 528 Homeostasis. Molecules and Cells, 2015. 38(12): p. 1023-1028. 529 18. Nozawa, K., et al., TRPA1 regulates gastrointestinal motility through serotonin release from 530 enterochromaffin cells. Proc Natl Acad Sci U S A, 2009. 106(9): p. 3408-13. 531 19. Chimienti, F., et al., In vivo expression and functional characterization of the zinc transporter 532 ZnT8 in glucose-induced insulin secretion. J Cell Sci, 2006. 119(Pt 20): p. 4199-206. 533 20. Scott, L.J., et al., A genome-wide association study of type 2 diabetes in Finns detects multiple Diabetes Page 26 of 57
534 susceptibility variants. Science, 2007. 316(5829): p. 1341-5. 535 21. Wu, Y., et al., Common Variants in CDKAL1, CDKN2A/B, IGF2BP2, SLC30A8, and HHEX/IDE 536 Genes Are Associated With Type 2 Diabetes and Impaired Fasting Glucose in a Chinese Han 537 Population. Diabetes, 2008. 57(10): p. 2834-2842. 538 22. Merriman, C., et al., Lipid-tuned Zinc Transport Activity of Human ZnT8 Protein Correlates 539 with Risk for Type-2 Diabetes. J Biol Chem, 2016. 291(53): p. 26950-26957. 540 23. Wong, W.P., et al., Exploring the Association Between Demographics, SLC30A8 Genotype, and 541 Human Islet Content of Zinc, Cadmium, Copper, Iron, Manganese and Nickel. Sci Rep, 2017. 542 7(1): p. 473. 543 24. Lichten, L.A. and R.J. Cousins, Mammalian Zinc Transporters: Nutritional and Physiologic 544 Regulation. Annual Review of Nutrition, 2009. 29: p. 153-176. 545 25. Fuchsberger, C., et al., The genetic architecture of type 2 diabetes. Nature, 2016. 536(7614): 546 p. 41-+. 547 26. Pound, L.D., et al., Deletion of the mouse Slc30a8 gene encoding zinc transporter-8 results in 548 impaired insulin secretion. Biochem J, 2009. 421(3): p. 371-6. 549 27. Wijesekara, N., et al., Beta cell-specific Znt8 deletion in mice causes marked defects in insulin 550 processing, crystallisation and secretion. Diabetologia, 2010. 53(8): p. 1656-68. 551 28. Pound, L.D., et al., The physiological effects of deleting the mouse SLC30A8 gene encoding 552 zinc transporter-8 are influenced by gender and genetic background. PLoS One, 2012. 7(7): p. 553 e40972. 554 29. Yano, J.M., et al., Indigenous bacteria from the gut microbiota regulate host serotonin 555 biosynthesis. Cell, 2015. 161(2): p. 264-76. 556 30. Reigstad, C.S., et al., Gut microbes promote colonic serotonin production through an effect of 557 short-chain fatty acids on enterochromaffin cells. Faseb j, 2015. 29(4): p. 1395-403. 558 31. Krężel, A. and W. Maret, The biological inorganic chemistry of zinc ions. Archives of 559 Biochemistry and Biophysics, 2016. 611: p. 3-19. 560 32. Huang, Q., et al., Coupling of Insulin Secretion and Display of a Granule-resident Zinc 561 Transporter ZnT8 on the Surface of Pancreatic Beta Cells. J Biol Chem, 2017. 292(10): p. 562 4034-4043. 563 33. Davidson, H.W., J.M. Wenzlau, and R.M. O'Brien, Zinc transporter 8 (ZnT8) and beta cell 564 function. Trends Endocrinol Metab, 2014. 25(8): p. 415-24. 565 34. Flannick, J., et al., Loss-of-function mutations in SLC30A8 protect against type 2 diabetes. Nat 566 Genet, 2014. 46(4): p. 357-63. 567 35. Kleiner, S., et al., Mice harboring the human SLC30A8 R138X loss-of-function mutation have 568 increased insulin secretory capacity. Proc Natl Acad Sci U S A, 2018. 115(32): p. E7642-e7649. 569 36. Paulmann, N., et al., Intracellular serotonin modulates insulin secretion from pancreatic 570 beta-cells by protein serotonylation. PLoS Biol, 2009. 7(10): p. e1000229. 571 37. Bennet, H., et al., Serotonin (5-HT) receptor 2b activation augments glucose-stimulated 572 insulin secretion in human and mouse islets of Langerhans. Diabetologia, 2016. 59(4): p. 573 744-54. 574 38. Kim, H., et al., Serotonin regulates pancreatic beta cell mass during pregnancy. Nat Med, 575 2010. 16(7): p. 804-8. 576 39. Ohara-Imaizumi, M., et al., Serotonin regulates glucose-stimulated insulin secretion from 577 pancreatic beta cells during pregnancy. Proc Natl Acad Sci U S A, 2013. 110(48): p. 19420-5. Page 27 of 57 Diabetes
578 40. Kim, K., et al., Functional role of serotonin in insulin secretion in a diet-induced 579 insulin-resistant state. Endocrinology, 2015. 156(2): p. 444-52. 580 41. Cauchi, S., et al., The genetic susceptibility to type 2 diabetes may be modulated by obesity 581 status: implications for association studies. BMC Med Genet, 2008. 9: p. 45. 582 42. Yadav, V.K., et al., Lrp5 controls bone formation by inhibiting serotonin synthesis in the 583 duodenum. Cell, 2008. 135(5): p. 825-37. 584
585
586 Diabetes Page 28 of 57
588 Figure legend
589 Figure 1. Slc30a8-/- mice display increased adiposity.
590 (A) Body weight of male and female mice (n = 6 per genotype) over the course of
591 the study. (B) Body composition of 12-week-old male mice (n = 7 per genotype) and
592 12-week-old female mice (n = 6 per genotype). (C) The weight to body weight ratio
593 of epidydimal white adipose tissue (eWAT) and subcutaneous white adipose tissue
594 (scWAT) in wild-type and Slc30a8-/- mice (n = 5 per genotype). (D) Representative
595 hematoxylin and eosin (H&E) images of eWAT from wild-type and Slc30a8-/- mice.
596 (E-F) Average (E) and distribution (F) of eWAT fat cell size from 8-week-old male
597 wild-type and Slc30a8-/- mice (n = 5 per genotype). (G-L) The mRNA expression
598 level of lipogenesis (G), lipid synthesis (H), fatty acid oxidation (I), lipolysis (J) and
599 fatty acid uptake (K) in eWAT from wild-type or Slc30a8-/- mice. Data are
600 represented as mean ± SEM. Statistical analysis via unpaired Student’s t test: *, P <
601 0.05, **, P < 0.01, ***, P < 0.001.
602
603 Figure 2. Slc30a8-/- mice have increased lipid deposition in brown adipose
604 tissue.
605 (A) The weight to body weight ratio of brown adipose tissue (BAT) in wild-type
606 and Slc30a8-/- mice (n = 5 per genotype). (B) Representative H&E images of BAT
607 from wild-type and Slc30a8-/- mice. (C-G) The mRNA expression level of
608 lipogenesis (C), lipid synthesis (D), lipolysis (E), fatty acid oxidation (F), and
609 thermogenesis (G) in BAT from wild-type and Slc30a8-/- mice (n = 5 per genotype). Page 29 of 57 Diabetes
610 (H) Western blotting and quantification of UCP1, Beta 3 AR and TH in wild-type and
611 Slc30a8-/- mice. Data are represented as mean ± SEM. Statistical analysis via
612 unpaired Student’s t test: *, P < 0.05, NS, non-significant.
613
614 Figure 3. Slc30a8-/- mice have increased lipid and glycogen deposition in the
615 liver.
616 (A) Gross morphology (upper panel) and representative H&E images (lower panel)
617 of liver from wild-type and Slc30a8-/- mice. (B) Representative images and
618 quantitative analysis of Oil red O staining. (C) Representative images and quantitative
619 analysis of PAS staining. (D-E) Hepatic total triglyceride (TG) (D) and total
620 cholesterol (TC) (E) levels in wild-type and Slc30a8-/- mice. Data are represented as
621 mean ± SEM. Statistical analysis via unpaired Student’s t test: *, P < 0.05.
622
623 Figure 4. Slc30a8-/- mice have normal IPGTT but abnormal OGTT.
624 (A) Fasting blood glucose level (n = 8 per genotype). (B) Insulin tolerance test
625 (ITT) and area under curve (AUC) analysis (n = 6 per genotype). (C) Intraperitoneal
626 glucose tolerance test (IPGTT) and AUC analysis (n = 6 per genotype). (D) Oral
627 glucose tolerance test (OGTT) and AUC analysis (n = 6 per genotype). Data are
628 represented as mean ± SEM. Statistical analysis via unpaired Student’s t test: *, P <
629 0.05.
630
631 Figure 5. Slc30a8-/- mice have altered intestinal tract morphology and
632 motility. Diabetes Page 30 of 57
633 (A) Representative H&E images of proximal colon from wild-type and Slc30a8-/-
634 mice. (B) Quantitative analysis of proximal colon diameter (n = 5 per genotype). (C)
635 Representative immunostaining and quantitative analysis of PCNA in wild-type and
636 Slc30a8-/- proximal colon. (D) Immunohistochemical staining and quantification of
637 TUNEL assay in wild-type and Slc30a8-/- proximal colon (n = 3 per genotype). Red
638 arrowheads indicate positive staining signals. (E) Representative H&E (upper panel)
639 and PAS staining (lower panel) of goblet cells in mucosal layer from wild-type and
640 Slc30a8-/- proximal colon. (F) Representative immunostaining and quantitative
641 analysis of CA1 in wild-type and Slc30a8-/- proximal colon. (G)Total intestinal
642 transit time of wild-type and Slc30a8-/- mice (n = 5 per genotype). Data are
643 represented as mean ± SEM. Statistical analysis via unpaired Student’s t test: *, P <
644 0.05, **, P < 0.01, ***, P < 0.001.
645
646 Figure 6. ZnT8 is expressed in enteroendocrine cells and regulates
647 peripheral 5-HT level.
648 (A) Representative immunofluorescent staining of ZnT8 in duodenum and
649 proximal colon. (B) Representative immunofluorescent staining and quantification of
650 overlap between 5-HT-expressing cells and ZnT8-expressing cells in proximal colon.
651 (C) Representative immunohistochemical staining and quantitative analysis of 5-HT
652 intensity in proximal colon. (D) Representative immunofluorescent staining and
653 quantitative analysis of 5-HT positive cell number in proximal colon. (E) Serum 5-HT
654 level (n = 5 per genotype). (F) mRNA analysis of 5-HT metabolism related genes. Page 31 of 57 Diabetes
655 36B4 was used as the referenced gene. (G) Western blot analysis and quantification of
656 TPH1 and CgA in wild-type and Slc30a8-/- proximal colon. Tubulin was used as the
657 loading control. Data are represented as mean ± SEM. Statistical analysis via unpaired
658 Student’s t test: *, P < 0.05, **, P < 0.01, ***, P < 0.001.
659
660 Figure 7. Inhibition of 5-HT synthesis in ZnT8 deficient mice reverses
661 metabolic dysfunction.
662 (A-B) Wild-type and Slc30a8-/- mice fed with normal chow diet (ND) were
663 received daily intraperitoneal PCPA injection (300 mg/kg BW) for 4 weeks, starting
664 at the age of 7-week-old. (A) Body weight over the course of study. (B) OGTT and
665 AUC analysis after PCPA injection for 4 weeks (n = 5-7 per genotype). (C-F)
666 Wild-type and Slc30a8-/- mice fed with HFD (60 kcal %) starting at the age of
667 6-week-old and were received daily intraperitoneal PCPA injection (300 mg/kg BW)
668 for 4 weeks starting at the age of 12-week-old (n = 5-7 per genotype). (C) Body
669 weight over the course of study. (D) Body weight change after PCPA injection from
670 the age of 12-week-old to 16-week-old. (E) Serum 5-HT level. (F) IPGTT and AUC
671 analysis after PCPA injection for 3 weeks. (G) Average food intake of wild-type and
672 Slc30a8-/- mice as indicated. Data are represented as mean ± SEM. Statistical analysis
673 via unpaired Student’s t test: *, P < 0.05, **, P < 0.01, ***, P < 0.001, NS,
674 non-significant.
675
676 Figure 8. Depletion of ZnT8 in RIN14B cells increases TPH1 expression and Diabetes Page 32 of 57
677 5-HT level.
678 (A-D) RIN14B cells were transfected with scramble or siRNA against ZnT8 for
679 48 hrs and the cell lysates were collected for further analysis. (A) mRNA expression
680 analysis of Slc30a8 gene. (B) Western blotting analysis and quantification of TPH1.
681 Tubulin was used as the loading control. (C) 5-HT level in the cell lysates. (D 5-HT in
682 the medium. (E-G) RIN14B cells were treated with zinc chelator, TPEN, or different
683 concentration of zinc sulfate for 24 hrs and then collected for further analysis.3 (E)
684 Western blotting analysis and quantification of TPH1. Actin was used as the loading
685 control. (F) mRNA expression analysis of Slc30a8 and Tph1 gene in RIN14B cells.
686 (G) 5-HT level in the medium. (H-I) RIN14B cells were transfected with human ZnT8
687 wild-type (WT), ZnT8 polymorphism variant R325 (Mut) or pCDNA vector for 48
688 hrs. (H) 5-HT level in the cell lysate. (I) 5-HT in the medium. (J) Proposed
689 mechanisms by which ZnT8 deficiency regulates systemic energy balance through
690 peripheral serotonin production in enteroenchromaffin cells. Data are represented as
691 mean ± SEM. Statistical analysis via unpaired Student’s t test or one-way ANOVA
692 test: *, P < 0.05, **, P < 0.01, ***, P < 0.001, NS, non-significant.
693 Page 33 of 57 Diabetes
Figure 1. Slc30a8-/- mice display increased adiposity. (A) Body weight of male and female mice (n = 6 per genotype) over the course of the study. (B) Body composition of 12-week-old male mice (n = 7 per genotype) and 12-week-old female mice (n = 6 per genotype). (C) The weight to body weight ratio of epidydimal white adipose tissue (eWAT) and subcutaneous white adipose tissue (scWAT) in wild-type and Slc30a8-/- mice (n = 5 per genotype). (D) Representative hematoxylin and eosin (H&E) images of eWAT from wild-type and Slc30a8-/- mice. (E-F) Average (E) and distribution (F) of eWAT fat cell size from 8- week-old male wild-type and Slc30a8-/- mice (n = 5 per genotype). (G-L) The mRNA expression level of lipogenesis (G), lipid synthesis (H), fatty acid oxidation (I), lipolysis (J) and fatty acid uptake (K) in eWAT from wild-type or Slc30a8-/- mice. Data are represented as mean ± SEM. Statistical analysis via unpaired Student’s t test: *, P < 0.05, **, P < 0.01, ***, P < 0.001. Diabetes Page 34 of 57
Figure 2. Slc30a8-/- mice have increased lipid deposition in brown adipose tissue. (A) The weight to body weight ratio of brown adipose tissue (BAT) in wild-type and Slc30a8-/- mice (n = 5 per genotype). (B) Representative H&E images of BAT from wild-type and Slc30a8-/- mice. (C-G) The mRNA expression level of lipogenesis (C), lipid synthesis (D), lipolysis (E), fatty acid oxidation (F), and thermogenesis (G) in BAT from wild-type and Slc30a8-/- mice (n = 5 per genotype). (H) Western blotting and quantification of UCP1, Beta 3 AR and TH in wild-type and Slc30a8-/- mice. Data are represented as mean ± SEM. Statistical analysis via unpaired Student’s t test: *, P < 0.05, NS, non-significant. Page 35 of 57 Diabetes
Figure 3. Slc30a8-/- mice have increased lipid and glycogen deposition in the liver. (A) Gross morphology (upper panel) and representative H&E images (lower panel) of liver from wild-type and Slc30a8-/- mice. (B) Representative images and quantitative analysis of Oil red O staining. (C) Representative images and quantitative analysis of PAS staining. (D-E) Hepatic total triglyceride (TG) (D) and total cholesterol (TC) (E) levels in wild-type and Slc30a8-/- mice. Data are represented as mean ± SEM. Statistical analysis via unpaired Student’s t test: *, P < 0.05. Diabetes Page 36 of 57
Figure 4. Slc30a8-/- mice have normal IPGTT but abnormal OGTT. (A) Fasting blood glucose level (n = 8 per genotype). (B) Insulin tolerance test (ITT) and area under curve (AUC) analysis (n = 6 per genotype). (C) Intraperitoneal glucose tolerance test (IPGTT) and AUC analysis (n = 6 per genotype). (D) Oral glucose tolerance test (OGTT) and AUC analysis (n = 6 per genotype). Data are represented as mean ± SEM. Statistical analysis via unpaired Student’s t test: *, P < 0.05. Page 37 of 57 Diabetes
Figure 5. Slc30a8-/- mice have altered intestinal tract morphology and motility. (A) Representative H&E images of proximal colon from wild-type and Slc30a8-/- mice. (B) Quantitative analysis of proximal colon diameter (n = 5 per genotype). (C) Representative immunostaining and quantitative analysis of PCNA in wild-type and Slc30a8-/- proximal colon. (D) Immunohistochemical staining and quantification of TUNEL assay in wild-type and Slc30a8-/- proximal colon (n = 3 per genotype). Red arrowheads indicate positive staining signals. (E) Representative H&E (upper panel) and PAS staining (lower panel) of goblet cells in mucosal layer from wild-type and Slc30a8-/- proximal colon. (F) Representative immunostaining and quantitative analysis of CA1 in wild-type and Slc30a8-/- proximal colon. (G)Total intestinal transit time of wild-type and Slc30a8-/- mice (n = 5 per genotype). Data are represented as mean ± SEM. Statistical analysis via unpaired Student’s t test: *, P < 0.05, **, P < 0.01, ***, P < 0.001.
205x161mm (300 x 300 DPI) Diabetes Page 38 of 57
Figure 6. ZnT8 is expressed in enteroendocrine cells and regulates peripheral 5-HT level. (A) Representative immunofluorescent staining of ZnT8 in duodenum and proximal colon. (B) Representative immunofluorescent staining and quantification of overlap between 5-HT-expressing cells and ZnT8- expressing cells in proximal colon. (C) Representative immunohistochemical staining and quantitative analysis of 5-HT intensity in proximal colon. (D) Representative immunofluorescent staining and quantitative analysis of 5-HT positive cell number in proximal colon. (E) Serum 5-HT level (n = 5 per genotype). (F) mRNA analysis of 5-HT metabolism related genes. 36B4 was used as the referenced gene. (G) Western blot analysis and quantification of TPH1 and CgA in wild-type and Slc30a8-/- proximal colon. Tubulin was used as the loading control. Data are represented as mean ± SEM. Statistical analysis via unpaired Student’s t test: *, P < 0.05, **, P < 0.01, ***, P < 0.001. Page 39 of 57 Diabetes
Figure 7. Inhibition of 5-HT synthesis in ZnT8 deficient mice reverses metabolic dysfunction. (A-B) Wild- type and Slc30a8-/- mice fed with normal chow diet (ND) were received daily intraperitoneal PCPA injection (300 mg/kg BW) for 4 weeks, starting at the age of 7-week-old. (A) Body weight over the course of study. (B) OGTT and AUC analysis after PCPA injection for 4 weeks (n = 5-7 per genotype). (C-F) Wild-type and Slc30a8-/- mice fed with HFD (60 kcal %) starting at the age of 6-week-old and were received daily intraperitoneal PCPA injection (300 mg/kg BW) for 4 weeks starting at the age of 12-week-old (n = 5-7 per genotype). (C) Body weight over the course of study. (D) Body weight change after PCPA injection from the age of 12-week-old to 16-week-old. (E) Serum 5-HT level. (F) IPGTT and AUC analysis after PCPA injection for 3 weeks. (G) Average food intake of wild-type and Slc30a8-/- mice as indicated. Data are represented as mean ± SEM. Statistical analysis via unpaired Student’s t test: *, P < 0.05, **, P < 0.01, ***, P < 0.001, NS, non-significant. Diabetes Page 40 of 57
Figure 8. Depletion of ZnT8 in RIN14B cells increases TPH1 expression and 5-HT level. (A-D) RIN14B cells were transfected with scramble or siRNA against ZnT8 for 48 hrs and the cell lysates were collected for further analysis. (A) mRNA expression analysis of Slc30a8 gene. (B) Western blotting analysis and quantification of TPH1. Tubulin was used as the loading control. (C) 5-HT level in the cell lysates. (D 5-HT in the medium. (E-G) RIN14B cells were treated with zinc chelator, TPEN, or different concentration of zinc sulfate for 24 hrs and then collected for further analysis.3 (E) Western blotting analysis and quantification of TPH1. Actin was used as the loading control. (F) mRNA expression analysis of Slc30a8 and Tph1 gene in RIN14B cells. (G) 5-HT level in the medium. (H-I) RIN14B cells were transfected with human ZnT8 wild-type (WT), ZnT8 polymorphism variant R325 (Mut) or pCDNA vector for 48 hrs. (H) 5-HT level in the cell lysate. (I) 5-HT in the medium. (J) Proposed mechanisms by which ZnT8 deficiency regulates systemic energy balance through peripheral serotonin production in enteroenchromaffin cells. Data are represented as mean ± SEM. Statistical analysis via unpaired Student’s t test or one-way ANOVA test: *, P < 0.05, **, P < 0.01, ***, P < 0.001, NS, non-significant. Page 41 of 57 Diabetes 1 Supplementary Figure Legends
2 Figure S1. Generation and characterization of Slc30a8-/- mice.
3 (A) Schematic diagram of the targeting strategy for Slc30a8-/- mice using
4 TALEN technology. (B) mRNA analysis of Slc30a8 gene expression in islets from
5 wild-type and Slc30a8-/- mice (n = 3 per group). (C) Western blot analysis of ZnT8
6 protein in pancreas from wild-type and Slc30a8-/- mice (n = 3 per group). (D)
7 Immunofluorescent staining of ZnT8 (green) and insulin (red) in islets from wild-type
8 and Slc30a8-/- mice. Nuclei are stained blue by 4,6-diamidino-2- phenylindole
9 (DAPI). (E) DTZ staining of isolated islets from wild-type and Slc30a8-/- mice
10 showed depleted zinc staining in Slc30a8-/- islets.
11
12 Figure S2. Serum catecholamine in ZnT8 group mice.
13 Serum catecholamine ELISA test showed unchanged sympathetic output in
14 Slc30a8-/- mice.
15
16 Figure S3. Metabolic parameters in ZnT8 group mice.
17 Oxygen consumption (A-B), Carbon dioxide production (C-D), respiratory
18 exchange ratio (RER) (E-F), total activity (G), ambulatory activity (H), heat (I-J) and
19 daily food intake (K) of wild-type and Slc30a8-/- mice at 8-weeks of age (n = 5 per
20 group). Statistical analysis via unpaired Student’s t test: *, P < 0.05.
21
22 Figure S4. Serum cytokine level in ZnT8 group mice. 23 Serum cytokines including interleukin 10 (IL-10) (A), interleukin 17A (IL-17A)
24 (B), interleukin 1β (IL-1β) (C), interleukin 6 (IL-6) (D), monocyte chemoattractant
25 protein 1 (MCP-1) (E), tumor necrosis factor α (TNF-α) (F) in wild-type and
26 Slc30a8-/- mice were measured by U-PLEX biomarker assay (MSD) (n = 4 per
27 group). Statistical analysis via unpaired Student’s t test.
28
29 Figure S5. Serum diabetes and obesity related hormones in ZnT8 group
30 mice.
31 Serum diabetes and obesity related hormones including ghrelin (A), gastric
32 inhibitory polypeptide (GIP) (B), glucagon-like peptide-1 (GLP-1) (C), insulin (D),
33 leptin (E), plasminogen activator inhibitor-1 (PAI-1) (F), resistin (G) and glucagon (H)
34 in wild-type and Slc30a8-/- mice were examined by the Bio-Plex ProTM mouse
35 diabetes immunoassays (n = 13 per group). Statistical analysis via unpaired Student’s
36 t test.
37
38 Figure S6. Gross morphology of GI tract in ZnT8 group mice.
39 Gross appearance of GI tract of 2-week-old (A) and 8-week-old (B) wild-type
40 and Slc30a8-/- mice.
41
42 Figure S7. Altered gut microbiota in Slc30a8-/- mice.
43 Metagenomic analysis of gut microbiota: (A-B) OTU number. (C) Observed
44 species. (D) ACE index. (E-F) Phylum-level classification of bacteria identified in 45 individual stool samples (E) and group samples (F).
46
47 Figure S8. Immunostaining of ZnT8 and enteroendocrine cell markers in
48 mouse colon.
49 (A) Representative immunofluorescent staining of ZnT8 (red) and
50 enteroendocrine cell marker chromogranin A (green) in proximal colon from
51 wild-type and Slc30a8-/- mice. Arrows indicate the cells co-expressing ZnT8 and CgA
52 in wild-type colon. Arrowheads indicate the CgA-expressing cells does not express
53 ZnT8 in Slc30a8-/- colon. (B) Representative immunofluorescent staining of GIP and
54 ZnT8 in serial paraffin sections of wild-type mouse colon. Nuclei are stained blue by
55 DAPI. Scale bar, 50 μm.
56
57 Figure S9. Exacerbation of glucose and lipid dysfunction by HFD in
58 Slc30a8-/- mice.
59 (A) Body weight of wild-type and Slc30a8-/- mice fed with HFD (n = 6 per
60 genotype). (B) Representative H&E staining of liver, eWAT and BAT from wild-type
61 and Slc30a8-/- mice fed with HFD. (C) OGTT and AUC analysis of wild-type and
62 Slc30a8-/- mice fed with HFD (n = 6 per genotype). (D-F) Serum insulin (D), TG (E)
63 and NEFA (F) of wild-type and Slc30a8-/- mice fed with HFD (n = 6 per genotype).
64 Data are mean ± SEM, statistical analysis via unpaired Student’s t test: *, P < 0.05,
65 **P < 0.01.
66 67 Figure S10. Metabolic parameter in PCPA treated HFD mice.
68 (A) Body weight. (B) Fasting glucose. (C-F) The weight of eWAT (C), scWAT
69 (D), BAT (E) and liver (F) (n = 4-6 per group). (G) Gross appearance of abdomen
70 cavities and adipose tissues. Arrows indicated the fibrosis of eWAT. Data are mean ±
71 SEM, statistical analysis via one-way ANOVA analysis: NS, non-significant.
72
73 Figure S11. Effect of human diabetes-risk ZnT8 mutant on TPH1 level in
74 RIN14B cells.
75 RIN14B cells were transfected with wild-type (WT) form or diabetes-risk mutant
76 (Mut) form of human ZnT8 or pCDNA vector for 48 hrs. The cell lysate was collected
77 and subjected to Western blotting analysis for TPH1. Representative immunoblot and
78 quantification of TPH1 were shown. GAPDH was used as the loading control. Data
79 are mean ± SEM, statistical analysis via one-way ANOVA test.
80 A B C mRNA Pancreas 40.25kb Forward strand 1.5 kDa +/+ -/- 1.0 35 ZnT8 Exon 3 55 β-actin Selected as target site 0.5 WT:tatcctcactgatgcggctcatct Slc30a8-/- (-2 bp):tatcctcactgat---ggctcatct 0.0
Relative mRNA abundance Relative mRNA WT Slc30a8-/- D ZnT8 Insulin DAPI Overlay
WT
Slc30a8-/- 50 μm
E WT Slc30a8-/-
Figure S1. Generation and characterization of Slc30a8-/- mice. Serum
800
NS 600
400
200
0 Catecholamine (pg/mL) WT Slc30a8-/- Figure S2. Serum catecholamine in ZnT8 group mice A B C D
VO2 VO2 VCO2 6000 VCO2 5500 WT Slc30a8-/- 6000 5500 5000 5000 5000 * (ml/kg/h) * 2
4500 (ml/kg/h) 4000 2
(ml/kg/h) 4500 2 (ml/kg/h) 4000 * 2 VCO VCO VO 4000
VO 4000 3000 3500 3500
2000 2000 3000 Dark Light Total 3000 Dark Light Total 20:00 8:00 20:00 8:00
E F G H RER Total activity Ambulatory activity 1.2 RER 2000 800
) 1.2 2 ) 2 600
/VO 1500
2 *
1.0 /VO
2 1.0 1000 400 RER(VCO Beam breaks Beam 0.8 0.8 breaks Beam 200
RER(VCO 500
0 0 0.6 Dark Light Total Dark Light Total 0.6 Dark Light Total 20:00 8:00
I K 0.8 HEAT J HEAT Food intake 0.8 0.20 0.6 0.6 0.15
0.4 0.4 0.10 KCAL/hr KCAL/hr 0.2 0.2 0.05 Food intake (g/kg BW/d)intake Food 0.0 0.0 0.00 Dark Light Total 20:00 8:00 WT Slc30a8-/-
Figure S3. Metabolic parameters in ZnT8 group mice A B IL-10 IL-17A 40 4
30 3
20 2
10 1 Serum IL-10 (pg/mL) Serum Serum IL-17A (pg/mL) Serum 0 0 WT KO WT KO
C IL-1β D IL-6 6 300
4 200
2 100 Serum IL-6 (pg/mL) Serum Serum IL-1b (pg/mL) Serum
0 0 WT KO WT KO E F MCP-1 TNF-α 100 60
80 40 60
40 20 20 Serum TNF-a (pg/mL) TNF-a Serum Serum MCP-1 (pg/mL) MCP-1 Serum 0 0 WT KO WT KO
Figure S4. Serum cytokine level in ZnT8 group mice A B C D Ghrelin GIP GLP-1 Insulin 5000 1000 280 10000
4000 800 260 8000 240 3000 600 6000 220 2000 400 4000 200 1000 200 2000 Serum GIP (pg/mL) GIP Serum 180 Serum insulin (pg/mL) insulin Serum Serum GLP-1 (pg/mL) GLP-1 Serum Serum Ghrelin (pg/mL) Ghrelin Serum 0 0 160 0 WT KO WT KO WT KO WT KO
E F G H Leptin PAI-1 Resistin Glucagon 8000 40000 200000 1100
6000 30000 150000 1000
4000 20000 100000 900
2000 10000 50000 800 Serum leptin (pg/mL) leptin Serum Serum PAI-1 (pg/mL) Serum Serum resistin (pg/mL) resistin Serum Serum glucagon (pg/mL) glucagon Serum 0 0 0 700 WT KO WT KO WT KO WT KO
Figure S5. Serum diabetes and obesity related hormones in ZnT8 group mice
A B 500 *
450
400 OTU number
350 WT KO
C D
*
E F
Figure S7. Altered gut microbiota in Slc30a8-/- mice. A ZnT8 CgA Merge
WT
Slc30a8-/-
50 μm
B GIP DAPI
ZnT8 DAPI
50 μm
Figure S8. Immunostaining of ZnT8 and enteroendocrine cell markers in mouse colon. A C
OGTT: area under curve Blood glucose (mmol/L) 10 15 20 25 Body weight(g) 0 5 10 20 30 40
(mmol*min/L) 0 0 1000 1500 2000 3 500 0 4 Figure S9. Exacerbation of glucose andlipiddysfunctionbyHFDin Figure S9.Exacerbationofglucose * THDKO-HFD WT-HFD 5 KO-HFD WT-HFD 30 * 6 Body weight Body 7 Time(min) Age (weeks) OGTT 8 60 * 9 10 11 * 90 12 * ** 13 D 14 **
Serum insulin (ng/mL) 120 0.0 0.2 0.4 0.6 0.8 1.0 15 THDKO-HFD WT-HFD Insulin B eWAT Liver BAT * E Serum TG (mM) 2.5 0.0 0.5 1.0 1.5 2.0 WT-HFD THDKO-HFD WT-HFD ** TG * Slc30a8-/-mice. F KO-HFD Serum NEFA (mEq/L) 0.8 0.0 0.2 0.4 0.6 THDKO-HFD WT-HFD 50 μm NEFA * A B Body weight Fasting glucose 50 10 NS 40 NS 8 30 6 20 4 10 2 Body weight (g)
0 Blood glucose (mM) 0
WT-HFD WT-HFD
WT-HFD-PCPAKO-HFD-PCPA WT-HFD-PCPAKO-HFD-PCPA
C 3 D 1.5 E 0.15 F 1.5 NS NS 2 1.0 0.10 1.0 NS NS 1 0.5 0.05 0.5 BAT mass (g) BAT Liver weight (g) eWAT mass (g) eWAT scWAT mass (g) scWAT 0 0.0 0.00 0.0
WT-HFD WT-HFD WT-HFD WT-HFD WT-HFD-PCPAKO-HFD-PCPA WT-HFD-PCPAKO-HFD-PCPA WT-HFD-PCPAKO-HFD-PCPA WT-HFD-PCPAKO-HFD-PCPA
G WT-HFD WT-HFD-PCPA KO-HFD-PCPA
scWAT
eWAT
BAT
Figure S10. Metabolic parameter in PCPA treated HFD mice. TPH1 1.5
pCDNA hZnT8 (WT) hZnT8 (Mut) 1.0 TPH1 55 kDa 0.5 35 kDa GAPDH 0.0 Relative protein level protein Relative
pCDNA
hZnT8 (WT)hZnT8 (Mut)
Figure S11. TPH1 level in RIN14B cell Supplementary Table 1. Primers used for real-time quantitative PCR.
Gene Species 5' sequence 3' sequence
Slc30a8 mouse AGCCACCAAGATGTACGCC CTTGCTTGCTCGACCTGTT
Pparg mouse GCCCTTTGGTGACTTTATGGAG GCAGCAGGTTGTCTTGGATG
Srebp1 mouse TGACCCGGCTATTCCGTGA CTGGGCTGAGCAATACAGTTC
Fasn mouse GGAGGTGGTGATAGCCGGTAT TGGGTAATCCATAGAGCCCAG
Acc mouse GGGAGCCTGACAGCAAGAAG CGGACAGACCAGTGGTATAAGTC
Dgat1 mouse TCCGTCCAGGGTGGTAGTG TGAACAAAGAATCTTGCAGACGA
Dgat2 mouse GCGCTACTTCCGAGACTACTT GGGCCTTATGCCAGGAAACT
Gpam mouse TCCAGAAGGTGAAAAGGAAAGC GGCAAAAGAGGATGAAGGTGAG
Scd1 mouse TTCTTGCGATACACTCTGGTGC CGGGATTGAATGTTCTTGTCGT
Mcad mouse AGGGTTTAGTTTTGAGTTGACGG CCCCGCTTTTGTCATATTCCG
Lcad mouse TCTTTTCCTCGGAGCATGACA GACCTCTCTACTCACTTCTCCAG Vlcad mouse CTACTGTGCTTCAGGGACAAC CAAAGGACTTCGATTCTGCCC
Lipe mouse ACGAGCCCTACCTCAAGAACTG ATCTGGCACCCTCACTCCATAG
Atgl mouse AGACAGAGCTTTCTCCCAGTGAA CCCCGTGAAGCCCAACT
Mgll mouse CGGACTTCCAAGTTTTTGTCAGA GCAGCCACTAGGATGGAGATG
Slc27a1 mouse CTGGGACTTCCGTGGACCT TCTTGCAGACGATACGCAGAA
Fabp4 mouse AAGAAGTGGGAGTGGGCTTTG CTGTCGTCTGCGGTGATTTC
Cd36 mouse ATGGGCTGTGATCGGAACTG GTCTTCCCAATAAGCATGTCTCC
Ucp1 mouse AGGCTTCCAGTACCATTAGGT CTGAGTGAGGCAAAGCTGATTT
Pgc1a mouse TATGGAGTGACATAGAGTGTGCT CCACTTCAATCCACCCAGAAAG
Pgc1b mouse TCCTGTAAAAGCCCGGAGTAT GCTCTGGTAGGGGCAGTGA
Cox4 mouse ATTGGCAAGAGAGCCATTTCTAC CACGCCGATCAGCGTAAGT
Nrf1 mouse AGCACGGAGTGACCCAAAC TGTACGTGGCTACATGGACCT
Tph1 mouse AAGAAATTGGCCTGGCTTC GTTTGCACAGCCCAAACTC Tph2 mouse GCCATGCAGCCCGCAATGATGATG CAACTGCTGTCTTGCTGCTC
Cga mouse AGGAGCATGGGATTCCACAG TGGCTTTTCTGGCTTGCTG
Maoa mouse GGAGAAGCCCAGTATCACAGG GAACCAAGACATTAATTTTGTATTCTGAC
Slc6a4 mouse TATCCAATGGGTACTCCGCAG CCGTTCCCCTTGGTGAATCT
Htr4 mouse AGTTCCAACGAGGGTTTCAGG CAGCAGGTTGCCCAAGATG
36B4 mouse AGATTCGGGATATGCTGTTGGC TCGGGTCCTAGACCAGTGTTC
Slc30a8 rat AAGTGGAGACTCTGTGCTGCTTCA GGCCTCGATGACAACCACAAAGAA
Tph1 rat TTTTGTGACTGCGACATCAA CGTGGTGTGGGACTTCAGC
Gapdh rat ACAGCAACAGGGTGGTGGAC TTTGAGGGTGCAGCGAACTT