Protein Kinase D2 drives chylomicron-mediated lipid transport in the intestine and promotes obesity

Abstract

Lipids are the most energy-dense components of the diet, and their overconsumption promotes obesity and diabetes. Dietary fat content has been linked to the lipid processing activity by the intestine and its overall capacity to absorb triglycerides (TG). However, the signaling cascades driving intestinal lipid absorption in response to elevated dietary fat are largely unknown. Here, we describe an unexpected role of the protein kinase D2 (PKD2) in lipid homeostasis. We demonstrate that PKD2 activity promotes chylomicron-mediated TG transfer in enterocytes. PKD2 increases chylomicron size to enhance the TG secretion on the basolateral side of the mouse and human enterocytes, which is associated with decreased abundance of APOA4. PKD2 activation in intestine also correlates positively with circulating TG in obese human patients. Importantly, deletion, inactivation, or inhibition of PKD2 ameliorates high-fat diet-induced obesity and diabetes and improves gut microbiota profile in mice. Taken together, our findings suggest that PKD2 represents a key signaling node promoting dietary fat absorption and may serve as an attractive target for the treatment of obesity.

Keywords: chylomicron; fat absorption; intestine; obesity; protein kinase D2/PKD2/PRKD2.

Figure 1. PKD2 inactivation protects from diet‐induced obesity

1. A

   Body weight gain of male mice with the specified genotypes under normal (ND) or high‐fat diet (HFD).

2. B

   Quantification of fat, free fluid, and lean mass by nuclear magnetic resonance (NMR) of mice in HFD in panel (A).

3. C

   Organ weight (percentage of total weight) of different fat depots, liver, and quadriceps of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice after 22 weeks in HFD.

4. D

   Quantification of the average adipocyte size in SubWAT and EpiWAT of male mice of the specified genotypes in normal and high‐fat diet.

5. E

   Representative pictures of H&E staining of SubWAT and EpiWAT of indicated male mice fed HFD.

6. F

   Triglyceride content in liver of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice in panel (A). Relative to Pkd2^{wt / wt}.

7. G

   Glucose tolerance test of the specified genotypes after 16 weeks in HFD.

8. H

   Insulin tolerance test of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice after 18 weeks in HFD.

9. I, J

   Triglycerides (I) and free fatty acids (J) in circulation of specified mice after 18 weeks in HFD.

10. K, L

    Energy expenditure (K) and energy intake (L) of mice after 20 weeks in HFD.

Data information: Male mice were subjected to ND or HFD directly after weaning, and the specified metabolic parameters were measured. n = 7 for ND. In HFD, n = 7 for Pkd2^{wt / wt} and n = 8 for Pkd2^{ki / ki}. Data presented as average ± SEM. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Significances were assessed by using a two‐tailed Student's t‐test for independent groups.

Figure EV1. Inactivation of PKD2 does not affect adipose tissue or liver function but increases pancreatic islets size ‐ Related to Fig 1

1. A

   Relative expression of specified genes in SubWAT of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice after 22 weeks in HFD. n = 6.

2. B

   Representative pictures of H&E staining of liver of indicated male mice in panel (A).

3. C, D

   Aspartate aminotransferase (C) and alanine aminotransferase (D) levels in serum of male mice in panel (A). n = 5.

4. E

   Glucose‐stimulated insulin secretion of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice after 8 weeks in HFD. n = 8.

5. F

   Representative pictures of immunofluorescent staining for insulin (red) and DAPI (blue) of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice. n = 3. For each subject, three different sections of the pancreas were taken and a distance of 50 μm was kept between the sections.

6. G

   Mean fluorescence intensity (MFI) for insulin of mice in panel (F). Relative to Pkd2^{wt / wt}. n = 3. MFI was quantified from all the islets found in each of the three sections per experimental subject.

7. H

   Islet area compared to pancreas area of mice in panel (F). Relative to Pkd2^{wt / wt}. All the islets found in each of the three sections per mouse were quantified.

8. I

   Activity of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice after 20 weeks in HFD. Expressed as counts per 12 h of intervals. n = 7 for Pkd2^{wt / wt} and n = 8 for Pkd2^{ki / ki}.

Data information: Data presented as average ± SEM, *P ≤ 0.05, **P ≤ 0.01. Significances were assessed by using a two‐tailed Student's t‐test for independent groups.

Figure 2. Pkd2 knockin mice excrete more energy in feces and present lower absorption of lipids in the intestine

1. A

   Pictures of feces collected from male mice fed HFD and photo of them placed in water.

2. B

   Weight of feces collected in a week.

3. C

   Feces excreted per gram of food consumed.

4. D

   Body weight gained per gram of food consumed.

5. E, F

   Energy content (E) and lipid content (F) per gram of feces.

6. G

   Percentage of energy excreted in feces from the total energy ingested.

7. H

   Metabolizable energy calculated from intake minus excreted and assuming a urinary excretion of 2%.

Data information: For panels (A–H), male mice after weaning were kept in individual cages during 2 weeks fed with HFD. For first 2 weeks, animals were acclimatized in the cages, and then during two more weeks, mice were monitored for food consumption and feces were analyzed for deposition of lipids and energy. n = 7 for Pkd2^{wt / wt} and n = 8 for Pkd2^{ki / ki}. Data presented as average ± SEM. **P ≤ 0.01, ***P ≤ 0.001. Significances were assessed by using a two‐tailed Student's t‐test for independent groups.

Figure EV2. PKD2 does not regulate the expression of genes implicated in alimentary lipid processing in the intestine or the liver ‐ Related to Fig 2

1. Weight of feces collected from Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice fed normal diet per week. n = 6.

2. Pictures of feces from Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice under normal diet. n = 6.

3. Western blot for pancreatic lipase in duodenum of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice 1 h after olive oil gavage. n = 6. The animals were dissected at the indicated time of the day.

4. Relative expression of specified genes in pancreas of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice. n = 6.

5. Relative expression of genes involved in bile acids production, metabolism, and transport in liver and ileum of male mice with indicated genotypes after 1 week of HFD. n = 6.

6. Relative expression of genes involved in fatty acid transport in jejunum of indicated male mice. n = 6.

7. Pkd1 and Pkd3 expression levels in duodenum and jejunum of male mice with indicated genotypes. n = 6.

8. Western blot analysis of protein kinase D1 (PKD1), protein kinase D2 (PKD2), and protein kinase D3 (PKD3) in small intestine of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice. n = 3.

9. Quantification of the Western blots from intestine samples in panel (H). n = 3.

10. Western blot analysis of protein kinase D2 (PKD2) and protein kinase D1 (PKD1) in Caco2 cells expressing shscramble or shPkd2. n = 3.

11. Quantification of the Western blots from Caco2 cells presented in panel (J). n = 3.

Data information: Data presented as average ± SEM, *P ≤ 0.05. Significances were assessed by using a two‐tailed Student's t‐test for independent groups.

Figure 3. PKD2 inactivation decreases lipids' absorption

1. Western blot analysis of phosphorylated PKD (S916/S876) in small intestine of male C57BL/6 mice after overnight fasting, olive oil gavage (or PBS in controls), and dissection at different time‐points.

2. Lipid tolerance test after oral gavage of olive oil (10 µl per gram of body weight). n = 6.

3. Lipid tolerance test 30 min after tyloxapol i.p. injection (250 µg/g body weight). In this case, only 5 µl of olive oil per gram of body weight was gavaged orally. n = 6.

4. Western blot analysis and quantification of phosphorylated PKD2 (S876) and phosphorylated PKD (S916/S876) in small intestine of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice. n = 3.

5. Western blot analysis and quantification of phosphorylated PKD (S916/S876) in Caco2 cells expressing shscramble or shPkd2. n = 3.

6. Representative light microscopy pictures of organoids derived from jejunum crypts of indicated male mice.

7. Percentage of released, uptaken and retained ¹⁴C‐labeled fatty acids from Caco2 cells shscramble or shPkd2 (sequence 1) in a Transwell system. n = 6.

8. Percentage of released ¹⁴C‐labeled fatty acids by Caco2 cells expressing shscramble* or shPkd2* (sequence 2) grown in a Transwell system. n = 6.

9. Transepithelial electric resistance of the Caco2 cells shscramble or shPkd2 (sequence 1). n = 6.

Data information: For panels (B and C), after acclimation, male mice after were treated as specified and triglycerides in circulation measured at the specified time‐points. Data presented as average ± SEM, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Significances were assessed by using a two‐tailed Student's t‐test for independent groups. Source data are available online for this figure.

Figure EV3. Inactivation of PKD2 does not affect the architecture, cellularity, or permeability of the intestine ‐ Related to Fig 3

1. A

   Representative pictures of H&E staining of jejunum of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice.

2. B, C

   Confocal microscopy images of mice‐derived organoids stained for E‐cadherin, DAPI and villin (B), or chromogranin A (C).

3. D

   Dextran concentration in serum after oral gavage of indicated male mice. n = 4

4. E

   Length of small intestine of indicated male mice. n = 7 for Pkd2^{wt / wt} and n = 8 for Pkd2^{ki / ki}.

5. F

   Western blot of PKD2 in Caco2 cell control (shScramble) and depleted from PKD2 by shRNA (sequence 1). n = 3.

6. G

   Western blot of PKD2 in Caco2 cells expressing shscramble or shPkd2 (sequence 2). Labeled as shscramble* or shPkd2*. n = 3.

7. H

   Percentage of retention and uptake of ¹⁴C‐labeled fatty acids in Caco2 cells expressing shscramble* or shPkd2* (sequence 2) grown in a Transwell system for 14 days. n = 6.

8. I

   Absolute mRNA quantification of Pkd1, Pkd2, and Pkd3 in different parts of intestine of C57BL/6 mice. n = 4

9. J

   Western blot of PKD3 in Caco2 cell control (shScramble) and depleted from PKD3 by shRNA. n = 2.

10. K

    Percentage of release, uptake, and retention of ¹⁴C‐labeled fatty acids in Caco2 cells expressing shscramble or shPkd3 grown in a Transwell system for 14 days. n = 5.

Data information: Data presented as average ± SEM. Significances were assessed by using a two‐tailed Student's t‐test for independent groups.

Figure EV4. Deletion of PKD2 in the intestine does not affect neither food intake nor energy dissipation of the mice, and does not influence hepatic function nor the intestine architecture ‐ Related to Figs 4 and 5

1. A

   Relative expression of Pkd2 in specified organs of Pkd2^{flox / flox} and Pkd2^{gutΔ / Δ} male mice. n = 4.

2. B

   Western blot of PKD2 in specified organs of Pkd2^{flox / flox} and Pkd2^{gutΔ / Δ} mice.

3. C, D

   Food intake (C) and energy expenditure (D) of Pkd2^{flox / flox} and Pkd2^{gutΔ / Δ} male mice after 13 weeks in HFD. n = 4.

4. E

   Length of small intestine of indicated male mice. n = 4.

5. F, G

   Representative H&E pictures of jejunum (F) and liver (G) of male mice with indicated genotypes after 17 weeks in HFD.

6. H

   MetaNMDs plots of bacterial composition show specific differences in small intestinal section. n = 7.

7. I

   Pairwise PERMANOVA indicates no significant differences in compositions of jejunum and ileum. n = 7.

8. J

   Beta diversity visualized by metaNMDS plots present differences in the microbial composition in duodenum and ileum & jejunum. n = 7.

9. K

   Relative expression of genes involved in triglyceride and chylomicron synthesis in jejunum of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice after 1 week in HFD. n = 7.

Data information: Data presented as average ± SEM, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Significances in panels (A–E) were assessed by using a two‐tailed Student's t‐test for independent groups. In panels (I–K), statistical testing was performed using PERMANOVA with corrections for multiple testing using the Benjamini and Hochberg method.

Figure 4. Intestinal deletion of PKD2 reduces lipid absorption, protects from HFD‐induced obesity, and improves microbiota profile

1. A

   Western blot and quantification of phosphorylated PKD (S916/S876) in small intestine of Pkd2^{flox / flox} and Pkd2^{gutΔ / Δ} male mice. n = 3.

2. B

   Western blot and quantification of phosphorylated PKD (S744–748) in small intestine of Pkd2^{flox / flox} and Pkd2^{gutΔ / Δ} male mice. n = 3.

3. C

   Lipids tolerance test after oral gavage of olive oil to Pkd2^{flox / flox} and Pkd2^{gutΔ / Δ} male mice after 6 weeks in HFD. n = 10.

4. D

   Body weight gain of male mice of indicated genotypes in HFD. n = 10.

5. E

   Quantification of fat, free fluid, and lean mass by nuclear magnetic resonance of mice in panel (B). n = 10.

6. F

   Glucose tolerance test of male mice with indicated genotypes after 16 weeks in HFD. n = 10.

7. G, H

   (G) Free fatty acids and (H) Triglyceride content in serum of male mice after 17 weeks in HFD. n = 10.

8. I

   Differences in bacterial alpha diversity as indicated by Shannon Index. n = 7.

9. J

   Changes in specific operational taxonomic unit (OTU) in small intestinal samples. n = 7.

Data information: Data presented as average ± SEM. In panels (I and J), central band indicates the median, the lower and upper parts of the box indicate the 25 and 75 percentiles and the whiskers connect data points outside these quartiles. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Significances in panels (A–H) were assessed by using a two‐tailed Student's t‐test for independent groups. In panel (G), statistical testing was performed using Mann–Whitney U‐test. In panel (H), statistical testing was performed using PERMANOVA with corrections for multiple testing using the Benjamini and Hochberg method. Source data are available online for this figure.

Figure 5. PKD2 promotes chylomicron size

1. Western blot (WB) analysis of specified proteins in small intestine of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice after 1 week of HFD. Quantification of the bands for each protein normalized to loading control and relative to Pkd2^{wt / wt}. n = 3.

2. WB of apoliprotein A4 in jejunum of Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice after 1 week of HFD. Quantification of bands normalized to loading control and relative to Pkd2^{wt / wt}. n = 3.

3. WB of apoliprotein A4 in jejunum of Pkd2^{flox / flox} and Pkd2^{gutΔ / Δ} male mice after 1 week of HFD. Quantification of bands normalized to loading control and relative to Pkd2^{flox / flox} n = 7.

4. WB analysis of specified apolipoproteins in 10 µl of serum of overnight fasted, 2 h refed male animals after 1 week in HFD. Quantification of bands relative to Pkd2^{wt / wt}. n = 4.

5. WB of 20 µl of basal medium of Caco2 cells expressing shscramble or shPkd2 in Transwell system after 12 h of stimulation with oleic acid. n = 6.

6. Electron microscopy pictures of chylomicrons obtained by ultracentrifugation of plasma from Pkd2^{wt / wt} and Pkd2^{ki / ki} male mice as specified in methods.

7. Quantification of the diameter of the chylomicrons shown in panel (F). n = 3.

8. WB with specified antibodies after an in vitro kinase assay.

Data information: Data presented as average ± SEM, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Significances were assessed by using a two‐tailed Student's t‐test for independent groups. Source data are available online for this figure.

Figure 6. Inhibition of PKDs by small‐molecule compound reduces the degree of diet‐induced obesity and diabetes

1. Basolateral release of ¹⁴C‐labeled fatty acids in Caco2 cells grown in a Transwell system and treated with CRT0066101 at the indicated concentrations for 24 h. n = 6.

2. Basolateral release of ¹⁴C‐labeled fatty acids in Caco2 cells expressing shscramble or shPkd2 in a Transwell system and treated with CRT0066101 for 24 h. n = 8.

3. Western blot analysis of phosphorylated PKD2 in jejunum of male C57BL/6 mice in HFD and treated with CRT0066101 or control (water) for 13 weeks. n = 3.

4. Energy content per gram of feces of male mice in HFD and receiving daily gavage of CRT0066101 or water as control. n = 6 for control and n = 5 for CRT0066101.

5. Lipid tolerance test after oral gavage of olive oil to male mice after 9 weeks in HFD and receiving daily CRT0066101 or control. n = 6 for control and n = 5 for CRT0066101.

6. WB of Apolipoprotein A4 (APOA4) in jejunum of male C57BL/6 mice in HFD and treated with CRT0066101 or control (water) for 13 weeks. Quantification of bands normalized to GAPDH and relative to control (water) n = 6.

7. Body weight of male C57BL/6 mice in HFD and treated with CRT0066101 or control (water) for 13 weeks. n = 6 for control and n = 5 for CRT0066101.

8. Glucose tolerance test in male C57BL/6 mice which received CRT0066101 or water (control). Measurements after 10 weeks of treatment and HFD feeding. n = 6 for control and n = 5 for CRT0066101.

9. Insulin tolerance test in male C57BL/6 mice which received CRT0066101 or water (control). Measurements after 11 weeks of treatment and HFD. n = 6 for control and n = 5 for CRT0066101.

Data information: Data presented as average ± SEM, *P ≤ 0.05, **P ≤ 0.01. Significances in panels (A and B) were assessed by one‐way ANOVA followed by the post hoc Tukey test. In panels (C–I), significances were assessed by using a two‐tailed Student's t‐test for independent groups. Source data are available online for this figure.

Figure EV5. Oral dosing of PKD inhibitor (CRT0066101, 10 mg/kg of body weight) does not affect PKD activity in the liver or adipose tissue ‐ Related to Figs 6 and 7

1. A

   Western blot and quantification of phosphorylated PKD (S916/S876) in liver of male C57BL/6 mice in HFD which received CRT0066101 or water (control) for 13 weeks. n = 3.

2. B

   Western blot and quantification of phosphorylated PKD (S916/S876) in EpiWAT of male C57BL/6 mice fed HFD which received CRT0066101 or water (control) for 13 weeks. n = 3.

3. C–E

   Energy intake (C), energy expenditure (D), and activity (E) of male C57BL/6 mice which received CRT0066101 or water (control). Measurements after 7 weeks of treatment and HFD. n = 4.

4. F

   Insulin levels of male C57BL/6 mice fed with HFD which received CRT0066101 or water (control) for 13 weeks. Samples taken after 4 h fasting and 1‐h refeeding. n = 6 for control and n = 5 for CRT0066101.

5. G

   Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels in serum of male C57BL/6 mice fed HFD which received CRT0066101 or water (control) for 13 weeks. n = 6 for control and n = 5 for CRT0066101.

6. H

   Triglyceride content in the liver of male C57BL/6 mice fed HFD which received CRT0066101 or water (control) for 13 weeks. Relative to control. n = 6 for control and n = 5 for CRT0066101.

7. I

   Representative H&E pictures of brown adipose tissue (BAT) of male C57BL/6 mice in HFD which received CRT0066101 or water (control) for 13 weeks.

8. J

   Intestinal permeability measured by dextran gavage of mice in Fig 7A. n = 8.

9. K

   Anthropometric, metabolic, and biochemical data from human female patients in Fig 7F–I. n = 7.

Data information: Data presented as average ± SEM, *P ≤ 0.05. Significances were assessed by using a two‐tailed Student's t‐test for independent groups.

Figure 7. Reduction in PKD activity in intestine ameliorates pre‐established obesity in mice and correlates with healthier blood profile in obese patients

1. Body weight gain of male C57BL/6 mice that, after 7 weeks in HFD, started receiving treatment with CRT0066101 or water. HFD continued along with treatment. n = 8.

2. Organ weight (expressed as percentage of total body weight) of different fat depots and liver of male mice in panel (A).

3. Quantification of fat, free fluid, and lean mass by nuclear magnetic resonance (NMR) of male mice in panel (A).

4. Glucose tolerance test 8 weeks after starting the treatment of mice in panel (A).

5. Insulin tolerance test 10 weeks after starting the treatment of mice in panel (A).

6. Western blot analysis of PKD phosphorylation in proximal jejunum of morbidly obese female patients. Samples are loaded according to fasting triglyceride levels (lower to higher). n = 7.

7. Scatter plot of PKD phosphorylation in jejunum (normalized to GAPDH) vs. triglycerides of morbidly obese female patients. n = 7.

8. Scatter plot of PKD phosphorylation in jejunum (normalized to GAPDH) vs. percentage of glycated hemoglobin (HbA1c) of morbidly obese female patients. n = 7.

9. Scatter plot of PKD phosphorylation in jejunum (normalized to GAPDH) vs. high‐density lipoprotein (HDL) of morbidly obese female patients. n = 7.

Data information: Data presented as average ± SEM, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001. Significances were assessed by using a two‐tailed Student's t‐test for independent groups. In panels (G–I), r ² = square of the Pearson product–moment correlation coefficient. p determined for Pearson correlation coefficient for the given degrees of freedom. Source data are available online for this figure.