Activation of Sphingolipid Pathway in the Livers of Lipodystrophic Agpat2-/- Mice.

A several fold increase in triacylglycerol is observed in the livers of lipodystrophic Agpat2-/- mice. We have previously reported an unexpected increase in the phosphatidic acid (PA) levels in the livers of these mice and that a few specific molecular species of PA were able to transcriptionally upregulate hepatic gluconeogenesis. In the current study, we measured the metabolites and expression of associated enzymes of the sphingolipid synthesis pathway. The entire sphingolipid pathway was activated both at the gene expression and the metabolite level. The levels of some ceramides were increased by as much as ~eightfold in the livers of Agpat2-/- mice. Furthermore, several molecular species of ceramides were increased in the plasma of Agpat2-/- mice, specifically ceramide C16:0, which was threefold elevated in the plasma of both the sexes. However, the ceramides failed to increase glucose production in mouse primary hepatocytes obtained from wild-type and Agpat2-/- mice, further establishing the specificity of PA in the induction of hepatic gluconeogenesis. This study shows elevated levels of sphingolipids in the steatotic livers of Agpat2-/- mice and increased expression of associated enzymes for the sphingolipid pathway. Therefore, this study and those in the literature suggest that ceramide C16:0 could be used as a biomarker for insulin resistance/type 2 diabetes mellitus.

The 1-acylglycerol-3-phosphate O-acyltransferase 2 knockout (Agpat2) mouse was generated based on the genetic observations that we made in human subjects with congenital generalized lipodystrophy, type 1 [1]. Agpat2 mice recapitulate all the features of human lipodystrophy, including hyperinsulinemia, diabetes mellitus, hypertriglyceridemia, and hepatic steatosis [2]. The AGPAT2 enzyme converts lysophosphatidic acid (LPA) to phosphatidic acid (PA) [3]. PA is a precursor for several phospholipids and neutral lipids. Livers of these mice are engorged with lipids, including phospholipids and neutral lipids. We have also reported a several fold increase in hepatic triacylglycerol levels in Agpat2 mice [2].

Currently, 11 AGPAT isoforms are known, both in humans and mice, encoded by different genes with different tissue expression patterns [4]. Some of these AGPATs have additional acyltransferase activities. For example, AGPAT10 is reported to have glycerol phosphate acyltransferase activity [5]. Similarly, AGPAT8 and AGPAT9 also possess lysocardiolipin acyltransferase 1 [6, 7] and lysophosphatidylcholine acyltransferase 1 [8] activity, respectively. Although almost all of these AGPATs have been studied to define their substrate specificities and subcellular localizations [4], only AGPAT1 and AGPAT2 isoforms are more widely studied. AGPAT1 and AGPAT2 isoforms are close homologs [3], but their tissue expression patterns are different: AGPAT1 is ubiquitously expressed, whereas AGPAT2 is more tissue-restricted [3]. In vitro substrate specificities for the AGPAT1 and AGPAT2 recombinant proteins are very similar, with few differences [3]. Both the enzymes prefer sn-1-C18:1-LPA and C18:1-coenzyme A as acyl acceptor and donor, respectively [3]. However, AGPAT1 may also use odd-chain acyl-coenzyme A (C15:0) as donor [3].

In a previous report, we measured the hepatic levels of LPA and PA in Agpat2 mice [9]. Although the LPA levels were elevated as expected, surprisingly we also found the PA levels to be increased. We also showed that some specific molecular species of PA increase hepatic gluconeogenesis when tested in cultured mouse primary hepatocytes. This increased hepatic gluconeogenesis is due to the transcriptional activation of glucose-6-phosphatase (G6pase) and phosphoenolpyruvate carboxykinase(Pepck) enzymes [9].

In the steatotic liver, lipids of all classes are increased, which includes phospholipids such as LPA, PA, phosphatidylcholine, and phosphatidylethanolamine; neutral lipids such as monoacylglycerol, diacylglycerol, and triacylglycerol; and sphingolipids such as ceramides [10]. Previous studies have reported increased ceramide levels in the livers of ob/ob mice [10, 11]. In seipin-deficient Bscl2 mice, another mouse model of lipodystrophy, the total hepatic ceramide levels were lower upon 4-hour fasting compared to those of nonfasting mice [12]. The pathway for the synthesis of sphingolipid, including ceramide and its metabolites, is shown in Fig. 1. The synthesis of sphingolipids is a complex process and requires >28 distinct enzymes [14, 15]. Since the livers of Agpat2 mice have increased de novo lipogenesis (fatty acid synthesis), it seems likely that some of the synthesized fatty acids might condense with the amino acid serine, the first and rate-limiting step in the de novo sphingolipid biosynthetic pathway.

Figure 1.

Lipogenesis, sphingolipid, and PA synthesis pathways. Glucose is the main intracellular substrate for the generation of fatty acids (FA). In the livers of Agpat2 mice, there is an increase of de novo lipogenesis [2]. Shown are some of the key steps involved in the conversion of glucose to FAs (reviewed in [13]). The fate of FAs most likely depends on their intracellular concentrations and could enter different pathway(s), including glycerophospholipid synthesis or the sphingolipid synthesis pathway. In sphingolipid synthesis, first palmitoyl Co-A condenses with the amino acid serine to generate 3-ketosphingosine, which is then reduced to form sphinganine. In the next two sequential steps, ceramide synthase converts sphinganine to dihydroceramide, followed by its conversion by desaturase to ceramide. Further modifications of ceramide are also depicted. In glycerophospholipid synthesis, glycerol-3-phosphate is acylated to produce LPA and an additional acylation to generate PA. The enzymes of these pathway are as follows: Sptlc, serine palmitoyltransferase; Kdsr, 3-ketosphingosine reductase; Sphk, sphingosine kinase (note Sphk has much broader substrate specificity phosphorylating both sphinganine and sphingosine); CerS, ceramide synthase; Degs, dihydroceramide desaturase; Sgms, sphingomyelin synthase; Cerk, ceramide kinase; Ugcg, UDP-glucose ceramide hexosyltransferase or hexosylceramide synthase; B4galt, UDP-Gal:betaGlcNAc β1,4-galactosyltransferase or lactosylceramide synthase. Most of the enzymes for this pathway have multiple isoforms; for example, CerS has six known isoforms. Key enzymes for de novo lipogenesis: Gkrp, glucokinase regulatory protein; Acyl, ATP citrate lyase; Fasn, fatty acid synthase; Acs, Acyl-coenzyme A synthetase; TCA cycle, tricarboxylic acid cycle. Key enzymes for glycerophospholipid synthesis: Gpat, glycerol-3-phosphate acyltransferase; Agpat, 1-acylglycerol-3-phosphate O-acyltransferase. Those metabolites shown in orange and those enzymes shown in blue were measured in this study.

In this study, we analyzed several metabolite groups and various associated enzymes of the sphingolipid biosynthesis pathway in the plasma and livers of Agpat2 and wild type mice. Moreover, we tested whether the elevated ceramide levels increase glucose output in cultured mouse primary hepatocytes.

1. Materials and Methods

A. Animals

The development and phenotype of Agpat2 mice have been reported previously [2]. These mice are of mixed genetic background (C57B6/S129Sv) and were maintained in a pathogen-free environment, constant temperature, and were fed ad libitum a regular chow diet and kept in a 12- hour light–dark cycle. All animal studies were approved by the Institutional Use and Care of Animals and BioSafety Committee at the University of Texas Southwestern Medical Center, Dallas, TX.

B. Sphingolipid Extraction

Total sphingolipids were extracted as follows: 40 mg flash-frozen liver from 4-month-old mice was homogenized in 2.0 mL organic extraction solvent (isopropanol:water:ethyl acetate, 25:10:65; v:v:v). Immediately afterward, 20 µL internal standard solution was added. The mixture was vortexed and sonicated in an ultrasonic bath for 40 minutes at 40°C. Next, the samples were allowed to reach room temperature, and 1.5 mL aqueous sucrose buffer (25 mM, pH 7.0) was added. Two-phase liquid extraction was performed, the supernatant was transferred to a new tube, and the pellet was re-extracted. Supernatants were combined and evaporated under nitrogen. The dried residue was reconstituted in 200 µL high-performance liquid chromatography (HPLC) solvent B (methanol/formic acid 99:1; v:v containing 5 mM ammonium formate) for liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis. Plasma samples were processed using a similar methodology requiring 25 µL sample.

C. Sphingolipid Quantification

Quantitative analysis of sphingolipids was achieved by LC-MS/MS electrospray ionization using selective reaction monitoring scan mode. Lipid separation was achieved by reverse-phase liquid chromatography on a 2.1 (internal diameter) × 150-mm Kinetex C8, 2.6-μm core-shell particle (Phenomenex, Torrance, CA) column under a gradient elution over 20 minutes, using three different mobile phases, as follows: eluent A consisting of CH3OH/H2O/HCOOH, 58/41/1, v/v/v with 5 mM ammonium formate; eluent B consisting of CH3OH/HCOOH, 99/1, v/v with 5 mM ammonium formate; and eluent C consisting of CH3OH/CH2Cl2 35/65, v/v with 5 mM ammonium formate. Gradient description: 0 minutes, 60% A, 40% B, 0.5% C, hold for 1 minute; 4 minutes, 25% A, 75% B, 0% C; 13 minutes, 100% B, hold for 3 minutes; 16 minutes 100% C, hold for 2 minutes; 18 minutes, 60% A, 40% B, 0% C. The instrument consisted of a Shimadzu Prominence HPLC system equipped with a CBM-20A controller, a DGUA3 degasser, three HPLC solvent delivery modules LC-ADXR, a CTO-20 AC column oven/chiller maintained at 30°C, and a SIL-20ACTHTautosampler. The HPLC system is coupled to an API 5000 LC-MS/MS mass spectrometer (AB SCIEX, Framingham, MA) equipped with a Turbo V ion source operating the electrospray ionization probe in positive mode. The concentration of each metabolite was determined according to calibration curves using the peak–area ratio of analyte vs corresponding internal standard [16-19]. Calibration curves were generated using serial dilutions of each target analyte. The mass/charge ratios (m/z) and retention time for each of the metabolites are presented in Supplemental Table 1, and product ion scans for representative metabolites for each class are presented in Supplemental Fig. 1(A) and 1(B). Sphingolipid standards and the Internal Standard Cocktail Mix II were purchased from Avanti Polar Lipids (Alabaster, AL). C22:0 hexosylceramide was purchased from Matreya (Pleasant Gap, PA). This analysis was carried out at University of Texas Southwestern Metabolic Phenotyping Core Facility.

D. Preparation of Liposomes With Ceramides and Transfection of Primary Mouse Hepatocytes

Liposomes were prepared fresh for each experiment, as described earlier [9]. A total of 100 µM various ceramides and 0.1 µCi [3H]-oleoyl-LPA (PerkinElmer, Waltham, MA) was dried and resuspended in phosphate-buffered saline. Our previous data show that LPA has no effect on glucose production from hepatocytes [9]. The mixture was sonicated for 15 minutes, and the liposomes were used in cell culture experiments. Control cells received the liposomes without the ceramides.

E. Isolation of Mouse Primary Hepatocytes

Primary hepatocytes were isolated from 1-month-old nonfasted mice of both sexes by a two-step collagenase perfusion method as described earlier [9, 20, 21]. Briefly, mice were anesthetized with isoflurane, and a catheter (24 × 3/4” safelet catheter; Fisher Scientific, Waltham, MA) was inserted in the portal vein. The liver was perfused with the liver perfusion medium (Life Technologies, Carlsbad, CA) maintained at 37°C using a peristaltic pump at a rate of 3 mL/min until the liver blanched. The liver was then perfused with the liver digest medium (Life Technologies) at a rate of 3 mL/min for 20 minutes, excised from the animal, and placed in a petri dish containing the liver digest media at 37°C. Hepatocytes were released by removing the hepatic capsule and then passed through 100 µm cell strainer. The cells were washed three times in low-glucose Dulbecco’s modified Eagle medium (DMEM) containing 100 mM HEPES by centrifuging at 25g for 2 minutes at 4°C. The cell pellet was resuspended in low-glucose DMEM containing 10% fetal bovine serum and 1% antibiotics. Approximately 1 × 106 cells were seeded into each well of a collagen-coated six-well plate (Fisher Scientific) and allowed to attach overnight. The following day, cells were prepared for glucose output assay.

F. DNA Extraction

Approximately 50 mg liver from 4-month-old wild-type (WT) or Agpat2 mice was homogenized in 100 mM Tris-HCl, pH 7.5, containing 10 mM NaCl. Cellular debris was removed by centrifugation at 3000g for 5 minutes at 4°C, and an aliquot was used to extract DNA. DNA was extracted with Easy-DNA kit (Invitrogen, Carlsbad, CA) following the manufacturer’s protocol.

G. Glucose Output Assay

Glucose output was measured, as described previously [9]. Briefly, hepatocytes isolated from the livers of 1-month-old WT or Agpat2 mice were plated on collagen-coated six-well plates and allowed to attach overnight. Cells were starved for 1 hour by incubating in phenol red–free, glucose-free DMEM (Sigma-Aldrich, St. Louis, MO). Cells were incubated in media containing sodium pyruvate and sodium lactate for 6 hours. Before the end of 6 hours, the cells were incubated with 100 µM PA (C16:0/18:1), used in this study as a positive control, or 100 µM ceramides C16:0, C18:0, or C24:1 for 30 minutes. Culture media was collected, and glucose was measured using a colorimetric assay kit from Cayman Chemical (Ann Arbor, MI). The glucose output was normalized to cellular protein content. Protein concentration was determined by a commercially available colorimetric assay (Bio-Rad Laboratories, Hercules, CA).

H. Quantitative Reverse Transcriptase Polymerase Chain Reaction (RT-qPCR) in Livers of WT and Agpat2−/− Mice

Total RNA was extracted from mouse livers (100–200 mg) using RNA STAT-60 (Tel-Test, Friendswood, TX). Total RNA, in equal quantity, was pooled from six livers of each genotype and sex, and RT-qPCR was carried out in a 20 µL reaction volume. A total of 20 µg RNA was DNase I treated using the DNase-free kit from Ambion (Grand Island, NY). Complementary DNA was made using 2 µg DNase I-treated RNA using reverse-transcription kit from ABI (Carlsbad, CA). RT-qPCR was performed in duplicate using 2.5 µM primers, 20 ng complementary DNA, and SYBR Green. All RT-qPCR were carried out in 96-well plates using the StepOnePlus real-time polymerase chain reaction system (Applied Biosystems, Foster City, CA). RT-qPCR was performed three times and in duplicate, and the transcript levels were normalized to cyclophilin. The ΔCt value for each sample was calculated as ΔCt = [Ct (gene of interest) − Ct (cyclophilin)]. The ΔΔCt value for each gene of interest was calculated as ΔΔCt = [ΔCt (Agpat2) − ΔCt (WT)]. The fold change was calculated as fold change = 2−Ct. Primers used for gene amplification were obtained from Harvard primer bank [22-24] and Integrated DNA Technologies (Coralville, IA). Primers used are listed in Table 1.

Table 1.

Primers Used in This Study

Gene	GenBank Accession No.	PrimerBank ID	Forward Primer	Reverse Primer
Sptlc1	NM_009269.2	29244577a1	ACGAGGCTCCAGCATACCAT	TCAGAACGCTCCTGCAACTTG
Sptlc2	NM_011479.4	6755656a1	AACGGGGAAGTGAGGAACG	CAGCATGGGTGTTTCTTCAAAAG
Sptlc3	NM_175467.3	28202041a1	TCTGAACGACAGTGCTGTTAC	ATGCCTTCCTATTTTGCTGGG
CerS1	NM_138647.3	20149718a1	CCACCACACACATCTTTCGG	GGAGCAGGTAAGCGCAGTAG
CerS2	NM_029789	—	Mm.PT.58.10718121 from Integrated DNA Technologies
CerS3	NM_001164201.1	255958166c1	CAGGCGAGGAGTATCCTGTG	CTCTCCGACCAGAACCATTTTC
CerS4	NM_026058.4	31541930a1	TACCCACATCAGACCCTGAAT	TGAAGTCCTTGCGTTTGACATC
CerS5	NM_028015.2	21312638a1	CGGGGAAAGGTGTCTAAGGAT	GTTCATGCAGTTGGCACCATT
CerS6	NM_172856.3	27370296a1	GATTCATAGCCAAACCATGTGCC	AATGCTCCGAACATCCCAGTC
Degs1	NM_007853.4	6681175a1	GAATGGGTCTACACGGACCAG	CGAGAAGCATCATGGCTACAA
Degs2	NM_001171002.1	27754029a1	AGCGACTTCGAGTGGGTCTA	TCCCCGTACTAACCAGCAGG
Asah1	NM_019734.3	9790019a1	CGTGGACAGAAGATTGCAGAA	TGGTGCCTTTTGAGCCAATAAT
Asah2	NM_018830.1	9055168a1	GCAAAGCGAACCTTCTCCAC	ACTGGTAACAAACAAGAGGGTGA
Acer1	NM_175731.4	28376625a1	TCTGAGGTGGATTGGTGTGAG	TGAGGGGTCCAAAGATGAGGA
Acer2	NM_139306.3	21314858a1	TGTGGCATATTCTCATCTGCCT	CAATAAAAGCCCATTTCTCGCTG
Acer3	NM_025408.2	21313584a1	TGTGATTCACTGAGGAACTTTCG	AGAAACTTCACTTTTGGCCTGTA
Sphk1	NM_011451	3659692a1	AAAATACTGAGAAACTCGGTCGG	GCATCGCTTCTTAAAGTCCAGA
Sphk2	NM_020011	31981070a1	CACGGCGAGTTTGGTTCCTA	CTTCTGGCTTTGGGCGTAGT
Sgpl1	NM_009163.4	31543694a1	CTGAAGGACTTCGAGCCTTATTT	ACTCCACGCAATGAGCTGC

All the primers were from Harvard PrimerBank (HPB) except CerS2, which was a predesigned primer from Integrated DNA Technologies (Coralville, IA). The HPB primers were validated by amplification plots and dissociation curves, and analyzed on 2% agarose gel analysis, sequencing, and BLAST analysis.

Abbreviations: Acer, alkaline ceramidase; Asah1, acid ceramidase; Asah2, nonlysosomal ceramidase; CerS, ceramide synthase; Degs, dihydroceramide desaturase; Sgpl, sphingosine 1-phosphate lyase; Sphk, sphingosine kinase; Sptlc, serine palmitoyltransferase.

I. Administration of Myriocin in Agpat2−/− Mice

The 17.5-week-old Agpat2 mice of both sexes were administered 0.5 mg/kg myriocin or vehicle every other day for 14 days by subcutaneous injection. The myriocin was prepared in 25% dimethyl sulfoxide, 25% propylene glycol, and 50% phosphate-buffered saline. Animals were killed at the end of the experiment, plasma was collected by cardiac puncture, and glucose was measured.

J. Statistical Analysis

Values are given as mean ± standard error of the mean or mean ± standard deviation. Statistical significance was calculated by two-tailed Student t test using GraphPad Prism version 6.04 for Windows. A P value ≤0.05 was considered statistically significant.

2. Results

A. Increased Levels of Sphingolipids in the Livers of Both Sexes of 4-Month-Old Agpat2−/− Mice

As shown in Table 2, we observed a robust increase in the levels of several sphingolipid subspecies in both sexes of Agpat2 mice compared with WT. The increases ranged from twofold to a robust increase of up to ~18- to 22-fold, but this increase was noticed only in the livers of male mice. Female mice showed an increase of ~one- to fivefold. This sexual dimorphism was expected because hepatic steatosis is more severe in male mice [2]. Most of the monitored sphingolipid species were found to be elevated in Agpat2 mice livers. Ceramide C16:0 increased ~eightfold and C18:0 increased ~19-fold in males, but only ~two- to fivefold in female mice. Other notable increases were found for HexCer-C18:0 (~23-fold) and sphingosine 1-phosphate (~fivefold). In females, HexCer-C18:2 increased by ~threefold and sphingosine by ~fourfold. This suggests that the entire sphingolipid biosynthetic pathway is activated in the livers of Agpat2 mice.

Table 2:

Sphingolipid Pathway Metabolite Levels in the Livers of Wild Type (WT) and

Name	Males				Females
	Liver (pg/µg DNA)				Liver (pg/µg DNA)
	WT (n = 6)	Agpat2−/− (n = 6)			WT (n = 6)	Agpat2−/− (n = 6)
	Mean ± SD	Mean ± SD	Fold Change	P Value	Mean ± SD	Mean ± SD	Fold Change	P Value
Sphinganine	1.53 ± 0.27	4.37 ± 1.54a	2.86b	<0.01	1.98 ± 0.23	5.22 ± 2.09a	2.64b	0.01
Sphinganine 1P	0.63 ± 0.15	3.19 ± 1.43a	5.06c	<0.01	0.93 ± 0.10	3.37 ± 0.97a	3.62c	<0.01
Deoxy sphinganine	0.18 ± 0.04	1.07 ± 0.74a	5.94b	0.03	0.23 ± 0.01	1.15 ± 0.35a	5.00b	<0.01
DHC16:0	0.08 ± 0.02	0.27 ± 0.14a	3.38b	0.02	0.17 ± 0.03	0.28 ± 0.09a	1.65b	0.01
DHC18:0	0.12 ± 0.02	0.87 ± 0.62a	7.25b	0.03	0.56 ± 0.18	0.89 ± 0.37a	1.59	0.03
DHC 24:0	0.90 ± 0.28	1.30 ± 0.75	1.44	0.27	1.03 ± 0.11	0.88 ± 0.32	−1.17	0.60
DHC 24:1	1.97 ± 0.41	5.97 ± 2.92a	3.03b	0.02	3.50 ± 0.11	5.75 ± 2.16	1.64b	0.01
SM 16:0	519.28 ± 69.29	1799.03 ± 1114.97a	3.46b	0.04	829.97 ± 177.88	2131.58 ± 780.92a	2.57b	<0.01
SM 18:0	69.62 ± 13.13	561.88 ± 403.31a	8.07b	0.03	248.59 ± 78.62	623.81 ± 164.58a	2.51	<0.01
SM 24:0	758.66 ± 281.60	678.38 ± 537.07	−1.12	0.75	826.84 ± 177.82	650.98 ± 266.62	−1.27	0.21
SM 24:1	976.91 ± 354.31	3105.95 ± 1871.75a	3.18b	0.04	1668.29 ± 347.37	3277.25 ± 1153.51a	1.96b	0.02
Cer16:0	1.26 ± 0.34	9.76 ± 5.83a	7.75c	0.02	2.30 ± 0.55	10.40 ± 2.50a	4.52c	<0.01
Cer18:0	0.36 ± 0.05	6.80 ± 3.25a	18.89c	<0.01	1.75 ± 0.61	8.00 ± 2.13a	4.57c	<0.01
Cer20:0	7.68 ± 1.48	20.89 ± 10.75a	2.72b	0.03	5.52 ± 1.38	17.46 ± 4.35a	3.16b	<0.01
Cer22:0	72.74 ± 4.99	90.56 ± 48.24	1.24	0.41	31.13 ± 5.29	78.29 ± 24.84a	2.51b	<0.01
Cer24:0	78.27 ± 19.86	140.34 ± 90.59	1.79	0.16	68.95 ± 6.04	94.23 ± 40.29	1.37	0.26
Cer24:1	82.72 ± 11.84	356.45 ± 156.74a	4.31b	<0.01	100.57 ± 17.59	322.47 ± 108.60a	3.21b	<0.01
HexCer-16:0	14.47 ± 7.12	39.01 ± 33.78	2.70	0.14	22.32 ± 2.33	40.14 ± 20.18a	1.80b	0.04
HexCer-18:0	0.39 ± 0.14	8.92 ± 6.00a	22.87c	0.02	6.09 ± 2.00	17.35 ± 9.77a	2.85c	0.03
HexCer-22:0	105.41 ± 16.45	136.05 ± 107.20	1.29	0.52	80.95 ± 21.50	115.91 ± 56.80	1.43	0.10
HexCer-24:1	26.63 ± 6.49	100.48 ± 58.98a	3.77b	0.03	47.06 ± 5.29	115.50 ± 47.64a	2.45b	<0.01
LacCer-16:0	7.50 ± 2.11	22.00 ± 11.68a	2.93b	0.03	6.95 ± 1.39	25.33 ± 12.13a	3.64b	<0.01
LacCer-24:0	7.19 ± 3.45	14.81 ± 8.07	2.06	0.07	6.57 ± 1.41	13.36 ± 6.39a	2.03b	0.02
Sphingosine	7.72 ± 1.36	22.23 ± 14.45	2.88	0.06	8.14 ± 0.78	23.60 ± 7.97a	2.90b	<0.01
Sphingosine 1P	2.29 ± 0.72	12.21 ± 5.78a	5.33b	<0.01	3.17 ± 0.83	12.05 ± 4.33a	3.80b	<0.01
Deoxy sphingosine	0.019 ± 0.006	0.04 ± 0.03	2.11	0.14	0.01 ± 0.00	0.048 ± 0.01a	4.80b	<0.01

Shown are the mean ± SD, fold change, and P value for metabolites of the sphingolipid synthesis pathway in the livers of Agpat2 mice compared with those of WT mice.

Abbreviations: Cer, ceramide; DHC, dihydroceramide; HexCer, hexosylceramide; LacCer, lactosylceramide; SD, standard deviation; SM, sphingomyelin; sphinganine-1P, sphinganine-1-phosphate; sphingosine 1P, sphingosine-1-phosphate.

Statistically significant increases.

Statistically significant increase in Agpat2 when compared with WT.

Increased in both liver and plasma of both sexes of Agpat2 mice and are statistically significant. Those in bold represent significant increase in both sexes.

B. Increased Levels of Sphingolipids in the Plasma of Both Sexes of 4-Month-Old Agpat2−/− Mice

We also observed a general elevation of sphingolipids in the plasma of Agpat2 mice (Table 3). However, these increases were not as pronounced as those observed in liver. Ceramide C16:0, C18:0, HexCer-C18:0, and sphingosine-1-phosphate were found to be increased ~three- to eightfold. Two previous studies also observed an increase in ceramide C16:0 in the liver, as well as in the white and brown adipose tissues in murine models of ceramide synthase (CerS) deficiency [25, 26]. A CerS isoform 2 (CerS2) haploinsufficient mouse model resulted in a compensatory increase in the hepatic ceramide C16:0, resulting in diet-induced steatohepatitis and insulin resistance [26]. Deletion of CerS isoform 6 (CerS6) either globally (CerS6), or tissue specifically in brown adipose tissue (CerS6) or in liver (CerS) resulted in improved glucose tolerance [25]. These previous studies and the current findings point to the notion that plasma ceramide levels reflect changes in hepatic ceramide levels and are valuable markers of hepatic metabolic health [27].

Table 3:

Sphingolipid Pathway Metabolite Levels in the Plasma of Wild Type (WT) and

Name	Males				Females
	Plasma (pg/mg Protein)				Plasma (pg/mg Protein)
	WT (n = 6)	Agpat2−/− (n = 6)			WT (n = 6)	Agpat2−/− (n = 6)
	Mean ± SD	Mean ± SD	Fold Change	P Value	Mean ± SD	Mean ± SD	Fold Change	P Value
Sphinganine	5.23 ± 3.79	11.21 ± 6.54	2.14	0.09	5.09 ± 1.46	13.85 ± 6.01a	2.72c	0.01
Sphinganine 1P	170.68 ± 122.21	452.66 ± 230.57a	2.65d	0.03	140.26 ± 30.89	547.98 ± 135.79a	3.91d	<0.01
Deoxy sphinganine	ND		ND	—	ND		ND	—
DHC16:0	4.67 ± 2.79	8.70 ± 10.48	1.86	0.40	5.31 ± 1.22	6.38 ± 3.68	1.20	0.44
DHC18:0	4.05 ± 3.05	6.68 ± 3.98	1.65	0.23	5.51 ± 0.74	7.31 ± 1.40	1.33	0.07
DHC 24:0	8.54 ± 4.75	8.22 ± 7.97	−1.04	0.93	8.75 ± 2.06	7.74 ± 5.70	−1.13	0.87
DHC 24:1	17.07 ± 10.36	17.87 ± 15.80	1.05	0.92	17.11 ± 2.01	17.03 ± 7.57	1.00	0.91
SM 16:0	12031.24 ± 6861.02	9815.96 ± 3788.08b	−1.23	0.50	13224.75 ± 2572.75	14732.00 ± 4583.52	1.11	0.50
SM 18:0	1000.23 ± 543.78	1748.32 ± 578.59a	1.75c	0.04	2242.75 ± 766.98	2744.98 ± 759.47	1.22	0.28
SM 24:0	6049.39 ± 3782.27	1820.08 ± 736.49b	−3.32c	0.04	4193.05 ± 1348.32	1847.55 ± 882.63b	−2.27c	<0.01
SM 24:1	18146.13 ± 9661.52	18393.44 ± 7812.37	1.01	0.96	22895.58 ± 8096.35	21529.06 ± 9774.50	−1.06	0.79
Cer16:0	24.64 ± 13.34	63.92 ± 15.52a	2.59d	0.01	25.95 ± 2.97	92.48 ± 5.90a	3.56d	<0.01
Cer18:0	5.20 ± 2.53	42.45 ± 11.07	8.16d	<0.01	15.82 ± 4.47	76.82 ± 8.27a	4.86d	<0.01
Cer20:0	31.56 ± 16.35	25.38 ± 9.81	−1.24	0.44	13.93 ± 3.34	37.82 ± 10.27a	2.72c	<0.01
Cer22:0	344.94 ± 147.80	107.11 ± 44.79b	−3.22c	<0.01	89.52 ± 20.30	147.04 ± 60.39	1.64	0.14
Cer24:0	343.67 ± 6.04	171.79 ± 72.63b	−2.00c	0.04	218.36 ± 38.67	183.22 ± 75.24	−1.19	0.20
Cer24:1	373.76 ± 150.84	376.11 ± 169.36	1.01	0.98	264.99 ± 87.13	528.62 ± 179.40a	1.99c	0.04
HexCer-16:0	220.84 ± 166.50	269.99 ± 116.53	1.22	0.57	288.81 ± 68.12	432.24 ± 130.77	1.50	0.06
HexCer-18:0	28.73 ± 19.32	235.03 ± 126.93a	8.18d	0.01	263.47 ± 54.42	399.94 ± 55.60a	1.52d	<0.01
HexCer-22:0	2217.93 ± 1272.21	560.24 ± 292.83a	−3.96c	0.02	1052.80 ± 278.14	707.20 ± 336.13	−1.49	0.16
HexCer-24:1	771.82 ± 550.44	723.97 ± 447.56	−1.07	0.87	957.65 ± 347.43	914.02 ± 427.44	−1.05	0.89
LacCer-16:0	63.77 ± 41.29	36.59 ± 21.36	−1.74	0.19	28.01 ± 13.46	35.54 ± 22.78	1.27	0.60
LacCer-24:0	18.90 ± 9.09	11.69 ± 4.49	−1.62	0.12	11.48 ± 2.38	10.50 ± 3.83	−1.09	0.83
Sphingosine	7.44 ± 3.26	19.48 ± 8.80a	2.62c	0.02	7.29 ± 1.38	20.13 ± 7.20a	2.76c	<0.01
Sphingosine 1P	563.10 ± 284.93	799.13 ± 309.10	1.42	0.20	470.25 ± 76.75	927.75 ± 240.80a	1.97c	<0.01
Deoxy sphingosine	ND		ND	—	ND		ND	—

Shown are the mean ± SD, fold change, and P value for metabolites of the sphingolipid synthesis pathway in the plasma of Agpat2 mice compared with those of WT mice.

Abbreviations: ND, not detected.

Statistically significant increases.

Statistically significant decreases.

Statistically significant increase in Agpat2 mice when compared with WT.

Increased in both liver and plasma of both sexes of Agpat2 mice and are statistically significant. Those in bold represents significant increase in both sexes.

C. Increased Expression (Messenger RNA) of Enzymes Involved in Sphingolipid Biosynthesis in the Livers of 4-Month-Old Agpat2−/− Mice

The elevated levels of liver sphingolipids correlate with an increase in the expression of the enzymes involved in sphingolipid biosynthesis. We measured the gene expression of key enzymes with the results shown in Fig. 2. Each of these enzymes has several isoforms except sphingosine-1-phosphate lyase (see Fig. 1). Although the increased gene expression is not as robust as observed for liver sphingolipid levels, in a steady state it is likely the metabolites accumulate over a period of time. The substrate specificities for some of these enzymes are now established (reviewed by Wegner et.al. [28]). For example, there are six different CerS, 1–6. Only CerS2 and CerS4–6 are expressed in the liver, and the increase in expression is in a range of one- to twofold. CerS2 prefers very long chain fatty acids to generate ceramide C22:0 to ceramide C24:0. This corresponds to an increased level of ceramide C24:1. CerS4 generally synthesizes ceramide with fatty acids C18:0 to C22:0, which we also see in the liver, whereas CerS5 and CerS6 prefer fatty acid C16:0. The increased gene expression of these CerS correlates well with the observed elevated levels of the corresponding ceramide.

Figure 2.

Hepatic expression of messenger RNA for enzymes of sphingolipid synthesis and metabolism. Expression of messenger RNA for Sptlc1–3, CerS1–6, Degs1–2, Asah1 (acid ceramidase), Asah2 (nonlysosomal ceramidase), Acer1–3 (alkaline ceramidase), Sphk1–2, and Sgpl (sphingosine 1-phosphate lyase) in the liver from WT and Agpat2 male (upper panel) and female (lower panel) mice. Expression was analyzed by RT-qPCR and normalized to cyclophilin. Shown are the fold changes compared with WT liver (expressed as mean ± standard error of the mean, n = 3). Each measurement was carried out in pooled (n = 6) samples in duplicate and was considered as one sample. *P value ≤0.05. The expression of most of the enzymes for the sphingolipid synthesis and metabolism in the liver has been noted before (www.biogps.org) except those for Sptlc2, Sptlc3, CerS1, CerS3, CerS4, CerS6, Degs2, Asah2, Acer1, Acer3, and Sphk1. Our data showed the expression of Sptlc2, CerS4, CerS6, Asah2, Acer3, and Sphk1 genes in liver.

D. Ceramides Do Not Increase Glucose Output in Mouse Primary Hepatocytes Obtained From WT or Agpat2−/− Mice

We next tested whether the ceramides, as those observed for PAs [9], also increase gluconeogenesis in hepatocytes. Among all the monitored ceramides, ceramide C16:0 and C18:0 were elevated in both sexes. Thus, we selected these two ceramides for our in vitro studies. Ceramide C24:1 was also included because this ceramide bears one unsaturation in the fatty acid chain [Fig. 3(a–c)]. As shown in Fig. 3(d) and 3(e), when testing for glucose production in mouse primary hepatic cells, we observed an increase with C16:0/18:1 PA, as reported earlier [9], which was used in this study as a positive control, but no statistically significant increase with the above-mentioned ceramides was observed. This suggests that ceramides do not increase hepatic gluconeogenesis in cultured mouse primary hepatocytes. Because the chemical structures of PA and ceramide 1-phosphate are very similar [Fig. 3(a–c)], we also tested ceramide C16:0-1-phosphate for its ability to increase glucose output. Again, we did not see an increase in glucose output (n = 1, data not shown). We also tested the plasma glucose levels in whole animals while administering the Sptlc1 inhibitor myriocin and observed no decrease in plasma glucose level in Agpat2 mice [Fig. 3(f)].

Figure 3.

Glucose output remained unchanged in primary mouse hepatocytes in the presence of ceramide, and Agpat2 mice do not show a decrease in plasma glucose levels upon administering myriocin, an inhibitor of ceramide synthesis. (a–c) A representative chemical structure of PA (C16:0/16:0) (a), ceramide (C16:0) (b), and ceramide 1-phosphate (C16:0) (c). Note a near structural similarity between PA and ceramide 1-phosphate. Glucose output was measured in the culture medium from the WT (d) and Agpat2 (e) primary mouse hepatocytes, when tested with the following ceramides: C16:0, C18:0, and C24:1. Glucose is expressed as mg/dL/µg protein. C16:0/18:1 PA was included as a positive control. Bar represents mean ± standard error of the mean obtained from three independent experiments performed in duplicate. *P value ≤0.05. (f) Shown are the individual plasma glucose levels in WT and Agpat2 mice. For individual hepatocytes or in vivo experiments, either male or female mice were used. For analysis, the data from both the sexes were combined. Filled circles represent males, and unfilled circles represent females.

3. Discussion

The role of ceramide is widely established as an insulin desensitizer [29]. In this work, our goal was to determine whether ceramides also have any role in de novo hepatic glucose synthesis, and, as tested in mouse primary hepatocytes, they did not increase glucose production. It is interesting to note that chemical structures of PA and ceramides are similar but not identical. The lipids essentially differ in their skeletal backbone: PA has a glycerol backbone, whereas ceramide has a serine backbone group. This study further establishes the lipid specificity for the induction of hepatic glucose synthesis by PA, as reported earlier [9], but not by ceramides (this study) or other phospholipids reported earlier [9]. A previous study, reported that ceramide C2:0 (tested at 20–50 µM) increased glucose production from mouse primary hepatocytes [30]. However, in our study, we did not detect ceramide C2:0 in the livers of Agpat2 mice. These investigators also showed that intraperitoneal administration of ceramide C16:0 (10 mg/kg/d for 1 week) resulted in glucose intolerance in intestine-specific farnesoid X receptor knockout (Fxr) mice [30]. However, our data show no effects of ceramide C16:0, as well as C18:0 and C24:1 (the predominant ceramide species in the livers of Agpat2 mice) on glucose output from mouse primary hepatocytes.

Another interesting observation is the increased plasma levels of ceramides in Agpat2 mice. Although not studied directly, it has been widely reported that some of the ceramides could be secreted as part of very-low-density lipoprotein secretion from the liver [31]. Additionally, as demonstrated in cultured HepG2 cells, increased hepatic ceramide levels resulted in increased secretion of liver ceramides [32]. It has been a long-standing goal to identify plasma biomarkers for the development of type 2 diabetes (reviewed in [27]). In a recent study, it was reported that plasma sphingolipids were increased in nonhuman primates fed a Western-style diet [33], specifically the molecular species of ceramide C14:0, C16:0, C22:0, and C24:0, which correlated well with insulin resistance. In another study, ceramide C22:0 was also increased in obese female children and adolescents with type 2 diabetes [34]. Ceramides C16:0 and C22:0 were also found to be elevated in the current study. Thus, it seems that these studies corroborate the increased plasma levels of ceramides C16:0 and C22:0 in type 2 and lipodystrophic diabetes.

Does the activation of the sphingolipid synthesis pathway have any additional function(s)? The livers of Agpat2 mice have elevated fatty acid synthesis [2] and phospholipids [2] despite the fact that the main route for its synthesis is genetically deleted and the total AGPAT enzymatic activity remains ∼10% of the WT mice. We have shown that conversion of diacylglycerol to PA by diacylglycerol kinase can occur in the livers of Agpat2 mice [9]. It is also possible that the downstream product of ceramide, sphingosine 1-phosphate, might provide an additional substrate for synthesis of phospholipids as a first step in converting sphingolipids to phospholipids [35]. The enzymes for these conversions have recently been identified in yeast [36] and in Drosophila [37] (reviewed in [38]) (see Fig. 4). The enzyme, sphingosine 1-phosphate lyase (Sgpl1), which synthesizes hexadecenal and ethanolamine-phosphate, is present in the livers of mice. Although the messenger RNA expression of Sgpl1 in this study did not increase in the Agpat2 mice compared with the WT mice, this does not necessarily mean that Sgpl1 has no function in synthesizing phospholipid from sphingosine. In a steady state condition, it might result in this conversion, albeit at a very low rate. In fact, a decrease in phosphatidylethanolamine and phosphatidylserine in the livers of Sgpl1 mice has been reported [39]. Additional experiments are needed to determine whether this route for the synthesis of phospholipids in the livers of Agpat2 mice is functional.

Figure 4.

Pathway for conversion of sphingosine to glycerophospholipid. The generation of sphingosine 1-phosphate is shown in Fig. 1. The enzyme sphingosine 1-phosphate lyase cleaves sphingosine 1-phosphate to generate phosphoethanolamine and hexadecenal. As shown, both the molecules trace two independent pathways. Phosphoethanolamine is converted to CDP-ethanolamine, followed by its sequential conversion to phosphatidylethanolamine. Hexadecenal enters the route to be converted to palmitoyl-coenzyme A via a series of intermediate steps, and could then re-enter the sphingolipid synthesis pathway. Sgpl, Sphingosine 1-phosphate lyase; Pcyt2, phosphoethanolamine cytidylyltransferase; and Cept, cytidine 5′-diphosphate-ethanolamine phospholtransferase. The metabolites shown in orange and the enzymes shown in blue were measured in this study.

As mentioned above, an increased de novo lipogenesis in the livers of Agpat2 mice [2] and the disruption of hepatic phospholipid synthesis might shift the equilibrium of fatty acids toward their usage in the sphingolipid synthesis pathway. Although this study shows that when ceramide was tested in isolated hepatocytes it did not increase hepatic gluconeogenesis, future studies will reveal the role of other sphingolipids in gluconeogenesis. We have previously demonstrated that, despite the deficiency of one of the Agpats, Agpat2, which results in ~90% decrease in the hepatic enzymatic activity [2], the PA levels were not decreased. This is a result of hepatic upregulation of diacylglycerol kinase [9]. Therefore, previous and current studies suggest additional experimental undertaking in resolving the role of each of the sphingolipid metabolite(s) of this pathway for hepatic gluconeogenesis.

In summary, we have shown that the entire pathway for sphingolipid synthesis is activated for both the expression of respective enzymes and the associated metabolites. This study further determined the specificity of PA (reported earlier [9]) toward increasing glucose output from primary hepatocytes since tested ceramides did not increase the hepatic glucose output. Lastly, we can speculate from this study that increased hepatic sphingosine 1-phosphate and the enzyme Sgpl1 might divert some of it toward the synthesis of phospholipids. However, this needs further confirmation.