Background

Leptin, the first discovered adipokine, plays a critical role in the regulation of energy balance and homeostasis of metabolism [1, 2]. Congenital leptin deficiency in humans results in extreme obesity, hyperphagia and many complications such as type 2 diabetes [3]. Leptin administration therapy with metreleptin, a recombinant human leptin analogue, has been approved for the treatment of the metabolic abnormalities linked to dyslipidemia [4]. Metabolic effects of leptin have been studied in rodent animal models [5]. Leptin signaling deficient rodent mutants, such as ob/ob mice, db/db mice and Zucker rats, have been commonly used as animal models in leptin studies [5]. Similar to the rare cases of congenital human leptin deficiency, these rodent mutants display hyperphagia, obesity and an insulin resistant phenotype. Several studies have shown metabolic disorders in ob/ob mice [6,7,8], db/db mice [6, 9, 10] and obese Zucker rats [11] measured by mass spectrometry (MS) or 1H solution nuclear magnetic resonance (NMR). Using a positional isotopomer NMR tracer analysis method, Perry et al. showed that leptin mediates a glucose-fatty acid cycle to maintain glucose homeostasis in starvation in rats [12]. Using a combination of metabolomics and transcriptomics, a recent published paper demonstrates that the carbohydrate, lipid and amino acid metabolic liver responses to glucose administration are broadly different between wild type and ob/ob mice [13].

Leptin and leptin receptor (lepr) are highly conserved and share extensive homology across vertebrates including all mammals and fish and have been studied in many model organisms [14, 15]. However, leptin functions in early development of vertebrates are largely unknown. Notwithstanding many reports indicate that leptin plays a key role in gestational diabetes and fetal development [16,17,18,19,20]. Further understanding of the function of leptin in these processes is hampered by the challenges of using rodent animal models for the study of fetal development. Zebrafish represents a robust animal model for early development because of its external fertilization, transparency of its larvae and large numbers of offspring. Since the zebrafish larvae are independent of feeding in the first five days after fertilization, it offers a great model for comparative leptin studies in fetal development with adult mammals. In zebrafish, there are two orthologs of the human leptin gene, leptin a (lepa) and leptin b (lepb). A zebrafish mutant line with lepa gene deficiency displays a phenotype of obesity and various deviations in behavior and circadian rhythm in the adult stage [21]. It was shown that lepb mutant zebrafish have more visceral fat and higher glucose level in male adult fish [22]. However, a zebrafish mutant for lepr was reported not to exhibit increased obesity in adult fish [23]. In larval studies, we have previously shown that lepb is one of the most affected genes after insulin injection [24]. However, how lepb gene affects the metabolic and transcriptomic level in zebrafish larvae is still unknown.

In this study, we have compared the metabolic changes resulting from leptin deficiency in blood of adult ob/ob mice and extracted and intact zebrafish larvae using MS, solution-state NMR and high-resolution magic-angle-spinning NMR (HR-MAS NMR) spectrometry. HR-MAS NMR is a noninvasive method that can be used for analysis of intact tissues at low temperature. In addition, we have compared the transcriptomic changes resulting from leptin deficiency in ob/ob mice heads, a published dataset for ob/ob mice liver and lepb mutant zebrafish larvae. These comparisons show a remarkable similarity of the effects of leptin knockdown on the metabolomes and transcriptomes of adult mice and zebrafish larvae.

Results

Metabolic profiles of blood from ob/ob and wild type mice measured by MS

We first investigated the metabolic profiles of blood from ob/ob and wild type lean male mice at 14 weeks of age. Mice were kept on a standard diet for 8 weeks, after which the body weight of ob/ob mice was significantly higher than wild type C57BL/6 mice (Additional file 1: Figure S1). Metabolic profiles of the blood of the two groups were obtained by MS. Using a highly standardized platform we could measure 41 small amine-containing compounds. A Partial Least Squares Discriminant Analysis (PLS-DA) scores plot of the 41 identified metabolites showed clear differences between the ob/ob and the wild type mice, indicating metabolic alterations in the metabolism due to leptin deficiency (Fig. 1A). Using a cut-off p value of 0.05, we could classify 30 out of the 41 identified small amine-containing compounds as associated with ob/ob mice. These 30 metabolites were significantly downregulated with a p value < 0.05 in ob/ob mice compared to wild type mice (Fig. 1B, Additional file 1: Table S1). For 25 of these metabolites, we have previously shown that they are biomarkers for Mycobacterium tuberculosis (Mtb)-infected mice (Fig. 1C). Graphs showing the quantification of these 25 common metabolites revealed that both the original and normalized values were decreased in leptin-deficient ob/ob mice (Additional file 1: Figure S2).

Fig. 1
figure 1

Metabolic profiles of blood from ob/ob and wild type C57BL/6 mice measured by mass spectrometry. A PLS-DA analysis of wild type and ob/ob mice, n = 15 in total, each replicate represents one mouse. PLS-DA partial least square discriminant analysis. WT wild type. B Heat map of 30 statistically significant biochemical markers profiled in this mice study. C A Venn diagram showing the overlap of the 30 metabolites of B with the set of wasting syndrome biomarkers published by Ding et al. [26]

Full size image

Metabolic profiles of extracts of lepb deficient and wild type zebrafish larvae measured by NMR

A lepb mutant zebrafish line was generated by CRISPR/CAS methodology [22]. Metabolic profiles of extracted zebrafish larvae from lepb−/− mutant and lepb+/+ wild type siblings were measured by one-dimensional 1H solution NMR. Figure 2A shows the representative 1H NMR spectra of extracted metabolites in the two groups. The assignment was performed based on the peaks of reference metabolites from literature [6, 25] in the library of Chenomx 600 MHz (version 11). A PLS-DA scores plot showed differences between the lepb−/− and lepb+/+ groups (Fig. 2B), suggesting metabolic changes resulting from lepb deficiency in zebrafish larvae. We found 27 metabolites to be significantly changed with a p value < 0.05 in extracted lepb−/− zebrafish larvae compared to lepb+/+ group. For 19 of these metabolites, we have previously shown that they are biomarkers for Mycobacterium marinum-infected zebrafish larvae (Fig. 2C). Quantification of these 19 common metabolites showed that the levels of all the metabolites were decreased in lepb mutant zebrafish larvae (Additional file 1: Figure S3).

Fig. 2
figure 2

One-dimensional 1H NMR spectra and PLS-DA analysis of extracted lepb mutant zebrafish larvae. A The representative spectra of extracted larvae from wild type and lepb mutant groups measured by solution NMR spectrometry. Spectra from chemical shift 0.5–4.4 were assigned to specific metabolites. Acet acetate, Ala alanine, Arg arginine, Asp aspartate, Cho choline, Chol cholesterol, Cit citrulline, Eta ethanolamine, FA fatty acid, Glc glucose, Gln glutamine, Glu glutamate, Gly glycine, Ile isoleucine, Kyn kynurenine, Lac lactate, Leu leucine, Lys lysine, Met methionine, m-Ins myo-inositol, Ser serine, Tau taurine, tCr total creatine (creatine + phosphocreatine), Trp tryptophan, NMR nuclear magnetic resonance. B PLS-DA analysis of wild type and lepb mutant groups, n = 4, each replicate represents 105 pooled larvae. PLS-DA partial least square discriminant analysis. C A Venn diagram is shown of the common 19 metabolites that changed significantly towards lepb deficiency in extracted zebrafish larvae and tuberculosis caused by Mycobacterium marinum infection in extracted zebrafish larvae published by Ding et al. [26]

Full size image

Metabolic profiles of intact lepb deficient and wild type zebrafish larvae measured by HR-MAS NMR

Due to the possibility of degradation and selective loss of compounds because of the extraction method needed for solution NMR, we used HR-MAS NMR as a comparative method on intact zebrafish larvae. Figure 3A showed the comparison of metabolic profiles and the assignments of metabolites of representative spectra in lepb mutant and wild type siblings. It was shown that the intensities of many peaks were lower in the mutant group. A PLS-DA scores plot showed clear discrimination between the lepb−/− and lepb+/+ groups (Fig. 3B). To compare the methods of solution NMR and HR-MAS NMR, we showed a Venn diagram of the significantly changed metabolites in the mutant and control siblings. The result revealed that there were 25 common metabolites significantly changed in both measurements (Fig. 3C). These 25 metabolites include the small amines alanine, asparagine, aspartate, citrulline, cysteine, ethanolamine, glutamate, glutamine, glycine, histidine, isoleucine, kynurenine, leucine, methionine, phenylalanine, putrescine, serine, threonine and tyrosine (Fig. 4A, C). In addition, the concentration of ATP, glucose, mannose, acetate, lactate and myo-inositol were changed significantly (Fig. 4B, D). For 21 of the 25 metabolites, both methods showed the same result: lower measurements of 20 metabolites and higher glucose level in the mutant group. However, kynurenine, tyrosine, ATP and mannose were detected at a decreased level in the mutant group with extracted larvae while at an increased level using intact larvae (Fig. 4).

Fig. 3
figure 3

One-dimensional 1H HR-MAS NMR spectra and PLS-DA analysis of intact lepb mutant zebrafish larvae. A The representative spectra of intact larvae from wild type and lepb mutant groups measured by HR-MAS NMR spectrometry. Spectra from chemical shift 0.5–4.4 were assigned to specific metabolites. Acet acetate, Ala alanine, Asp aspartate, Cho choline, Chol cholesterol, Cit citrulline, Cys cysteine, Eta ethanolamine, FA fatty acid, Glc glucose, Gln glutamine, Glu glutamate, Gly glycine, His histidine, Ile isoleucine, Lac lactate, Leu leucine, Lys lysine, Met methionine, m-Ins myo-inositol, Pu putrescine, Ser serine, Tau taurine, tCr total creatine (creatine + phosphocreatine), Thr threonine, TAMO trimethylamine N-oxide, HR-MAS NMR high-resolution magic-angle-spinning nuclear magnetic resonance. B PLS-DA analysis of intact larvae from wild type and lepb mutant groups, n = 3, three times measurements, each replicate represents 120 pooled larvae. PLS-DA partial least square discriminant analysis. C A Venn diagram is shown of the common 25 metabolites that are significantly changed both in extracted zebrafish larvae measured by 1H solution NMR and intact larvae measured by 1H HR-MAS NMR

Full size image
Fig. 4
figure 4

Quantification of the common 25 metabolites that are significantly changed in zebrafish larvae. A The concentration of amino acids and amines of wild type and lepb mutant in extracted larvae. B The concentration of ATP, carbohydrates and organic acids of wild type and lepb mutant in extracted larvae. C The concentration of amino acids and amines of wild type and lepb mutant in intact larvae. D The concentration of ATP, carbohydrates and organic acids of wild type and lepb mutant in intact larvae. *p < 0.05, **p < 0.01, ***p < 0.0005, ****p < 0.0001

Full size image

A core set of metabolites are markers for leptin deficiency in mice and zebrafish larvae

A common set of 13 metabolites were significantly changed in ob/ob mice blood, extracted lepb mutant and intact lepb mutant zebrafish larvae compared to their respective wild type controls (Fig. 5A). These 13 common metabolites were alanine, citrulline, ethanolamine, glutamine, glycine, histidine, isoleucine, leucine, methionine, phenylalanine, putrescine, serine and threonine (Fig. 5B). The concentrations of these 13 metabolites were reduced in a mutant compared to wild types for all the three metabolomic data sets (Fig. 5B). Of these metabolites, the following six are also reported as markers for tuberculosis infection in human, mice and zebrafish larvae based on mass spectrometry: citrulline, ethanolamine, leucine, methionine, phenylalanine, serine and threonine [26].

Fig. 5
figure 5

Common biomarkers for leptin deficiency in ob/ob mice, extracted and intact zebrafish larvae. A A Venn diagram shows that 13 common metabolites are significantly changed after leptin knockdown in mice blood, extracted and intact zebrafish larvae. B The ratio of leptin mutant versus wild type of the 13 common metabolites in the three metabolomic datasets

Full size image

Lipid profiles of lepb-deficient zebrafish larvae

To investigate whether lipid metabolism is influenced by leptin deficiency at the early stage of zebrafish development, lipids were extracted from pooled 5 days post fertilization (dpf) zebrafish larvae in the lepb mutant and sibling control groups and then measured with 1H solution NMR (Fig. 6A). A PLS-DA scores plot of the tetramethylsilane (TMS) normalized spectra showed a clear separation of the lipid profiles of the two groups (Fig. 6B), which indicated lipid metabolism was altered in lepb mutant zebrafish larvae. Twenty-two lipid signals could be assigned from chemical shift 0.5–5.5 in the spectra of both groups (Fig. 6A, Additional file 1: Table S2). Based on the quantification of normalized peaks, we can conclude that saturated lipids were significantly increased in the lepb mutant zebrafish larvae (Fig. 6C, D). In addition, the polyunsaturated fatty acid (PUFA) docosahexaenoic acid (DHA) was found in a higher abundance in the mutant group (Fig. 6E).

Fig. 6
figure 6

Lipid profiles of lepb mutant zebrafish larvae compared to wild type siblings. A The representative spectra of total lipid extracts from wild type and lepb mutant zebrafish larvae obtained by 1H NMR spectroscopy. The assignments of the peak numbers were shown in Additional file 1: Table S2. NMR nuclear magnetic resonance. B PLS-DA analysis of lepb mutant and wild type zebrafish larvae, n = 3, each replicate represents 105 pooled larvae. PLS-DA partial least square discriminant analysis. C The relative concentration of the signal 14 FA in A. FA fatty acids. D The relative concentration of the signal 18 PC in A. PC phosphatidylcholines. E The relative concentration of the signal 15 DHA in A. DHA docosahexaenoic acid. *p < 0.05

Full size image

Deep sequencing of transcriptomes of leptin deficient mice and zebrafish larvae

We investigated the effects of leptin deficiency at the transcriptome level in mice and zebrafish larvae by using deep RNA sequencing methods. Samples were taken from the same experimental groups as used for the metabolomic analysis described above. Mice heads were taken as a body part of interest because of the known classical signaling of leptin in the brain. A volcano plot showed that 5658 genes significantly regulated at a p value < 0.05 in ob/ob mice compared to wild type C57BL/6 mice (Fig. 7A). A recent paper published by Kokaji et al. reported the transcriptomes of mice liver from ten-week-old male ob/ob mutant and C57BL/6 wild type mice [13]. The comparison of the two mice liver groups showed 6693 genes significantly regulated at a p value < 0.05 (Additional file 1: Figure S4). The two gene sets encompassing 5658 and 6693 genes of the transcriptomes in mice head and mice liver, respectively, showed an overlap of 1865 genes (Fig. 7B). Gene ontology (GO) enrichment analysis using DAVID showed a large group of GO terms. In Fig. 7C, we showed a selected set of GO terms (biological process) with the lowest p adjusted values and the highest numbers of genes representatives. The GO term nervous system development was in line with the function of leptin in the brain. The GO enrichment of the overlap sets in Fig. 7B gave comparable results as with the mouse head GO terms, with the exception of ion transport and nervous system development (Fig. 7C). This could be explained by the relatively large number of neuronal cells in the head compared to liver.

Fig. 7
figure 7

Transcriptome signature sets of mice and zebrafish larvae. A A Volcano plot showing a graphical representation of the significance (p < 0.05) in ob/ob mice head compared to C57BL/6 mice head. The transcripts with fold change over 1.5 are highlighted in red. Fifteen significant genes in mice head out of the fold change in X axis are excluded to make the graph look well. B A Venn diagram showing the comparison of the number of significantly changed genes between ob/ob mice head and mice liver published by Kokaji et al. C The top eight GO terms of biological process (BP) with lowest p adjusted values and highest numbers of genes representatives in mice head and the overlap of B. GO gene ontology. D Number of genes in classification of GO term proteolysis in the signature set of mice head. E A Volcano plot showing a graphical representation of the significance (p < 0.05) in lepb mutant zebrafish larvae compared to wild type siblings. The transcripts with fold change over 1.5 are highlighted in green. Twenty-two significant genes in zebrafish larvae out of the fold change in X axis are excluded to make the graph look well. F A Venn diagram showing the comparison of the number of significantly changed genes from human homologs of the signature gene sets of zebrafish larvae and ob/ob mice head. G The top eight GO terms of BP with lowest p adjusted values and highest numbers of genes representatives in zebrafish larvae and the overlap of F. H Number of genes in classification of GO term proteolysis in the signature set of zebrafish larvae

Full size image

For zebrafish larvae, there were 2718 genes significantly regulated at a p value < 0.05 in lepb mutant zebrafish larvae compared to wild type siblings (Fig. 7E). We validated the mRNA expression level of a few representative genes in the lepb+/+ and lepb−/− zebrafish larvae with qPCR (Additional file 1: Figure S5). The human orthologs of this zebrafish larvae gene set and of the mice head transcriptome ob/ob signature set showed an overlap of 470 genes (Fig. 7F). The GO enrichment analysis of Fig. 7G showed the top eight GO terms (biological process) with lowest p adjusted values and highest numbers of genes representatives in the signature set of zebrafish larvae (Fig. 7G). The GO enrichment of the overlap set gave a similar result as in the zebrafish larvae terms with the exception of DNA replication, cell division and mitotic nuclear division. As shown in the Fig. 7C, G, one of the top GO terms in the signature set of mice heads, zebrafish larvae and the overlap was proteolysis. We also found the GO term proteolysis to be significantly enriched in the overlap of mice head and liver ob/ob signature set (Fig. 7C). The genes linked to this GO term were proteases which could be classified as aspartyl protease, carboxypeptidase, cysteine protease, metallopeptidase, serine protease, and threonine protease (Fig. 7D, H, Additional file 1: Table S3, S4). The pattern of the enriched gene numbers of those proteases in the signature sets of mice heads and zebrafish was similar in the up or down regulated groups (Fig. 7D, H).

As shown in Fig. 6, fatty acids such as DHA were significantly increased in the lepb mutant zebrafish larvae compared to wild type siblings. Lipid metabolism disturbance is possibly associated with inflammation [27]. Obese leptin deficient ob/ob mice show a low-grade chronic inflammation [28]. Interestingly, we found another common enriched GO term using DAVID (KEGG pathway) in the signature sets of mice head, mice liver and zebrafish larvae was arachidonic acid (ARA) metabolism (Additional file 1: Table S5–S7). Arachidonic acid is a pro-inflammatory precursor that can mediate inflammatory responses via transforming into a variety of downstream products such as prostaglandins and leukotrienes. It is also an early indicator of inflammation [27]. Therefore, the human orthologs of the signature sets of mice head, mice liver and zebrafish larvae were projected on the drawn ARA metabolic pathway based on the human wikipathways data using Pathvisio (Fig. 8). As shown in Fig. 8, five genes in the pathway namely ANXA1, ANXA5, ACSL3, MAP2K6, NFE2L2 were altered in all three datasets. Some other genes were significantly changed in only one or two datasets. However, the majority of the gene expression levels of the three datasets visualized in this pathway were not high (Additional file 1: Table S8). This indicates there might be only mild inflammation in the leptin deficient mice and zebrafish larvae.

Fig. 8
figure 8

Genes involved in arachidonic acid pathway in human orthologs of the three transcriptome signature sets. Dashed lines means indirect regulation. Red color represents genes significantly changed in ob/ob mice head compared to control. Blue color represents genes significantly changed in ob/ob mice liver compared to wild type published by Kokaji et al. Green color represents genes significantly changed in lepb mutant zebrafish larvae compared to wild type siblings. COX cyclo-oxygenase, LOX lipoxygenase, EOX epoxygenase, DHA docosahexaenoic acid

Full size image

Discussion

In this study, we have compared the metabolic changes resulting from leptin deficiency in blood of adult mice and extracted as well as intact zebrafish larvae. We studied metabolism using three different technologies: mass spectroscopy (MS), nuclear magnetic resonance (NMR) and high-resolution magic-angle-spinning NMR (HR-MAS NMR) spectrometry. In addition, we have compared the transcriptomic changes resulting from leptin deficiency in ob/ob mice heads and published data sets for ob/ob mice liver and lepb mutant zebrafish larvae using deep RNA sequencing (RNA-seq). These comparisons using very different omics technologies all show a remarkable similarity of the effects of leptin knockdown on the metabolomes and transcriptomes of adult mice and zebrafish larvae. These similarities are surprising because the analyzed samples of this comparative study are in many respects extremely different: (1) Mice and zebrafish are very diverse examples of the vertebrate subphylum, e.g., metabolic rate, body size, body temperature and examined life stages vary greatly; (2) Samples of blood or body tissue, in the case of the mice experiments, are compared with the entire organism in the case of zebrafish larvae; (3) The environmental conditions are different in mice and zebrafish larvae; (4) The genetic variation within the studied populations is highly diverse in zebrafish test samples, whereas a highly inbred population is used in the case of mice; and (5) For zebrafish larvae, there is no feeding of the organism involved and embryos are able to develop normally based on their reserves in the yolk until 5dpf. Nevertheless, also in a previous study, we showed remarkable similarities in small metabolite levels occurring in mice blood and zebrafish larvae after infection by mycobacteria [26]. The observed metabolic changes were mainly comprising a reduction of the levels of various amino acids that were also detected in human tuberculosis patients of several ethnical populations [26, 29, 30].

In the present study, we have also included HR-MAS NMR as a non-invasive method for analysis of metabolites in intact embryos. The results confirm the findings obtained with solution-state NMR analysis of extracted tissues. A few metabolites are changed in different directions measured by the two approaches, namely kynurenine, tyrosine, ATP and mannose. These are detected at a decreased level in the mutant group with extracted larvae while an increased level was detected with intact larvae using HR-MAS NMR (Fig. 4). This might be due to the fact that samples detected by solution-state NMR require extraction and pretreatment. Therefore, solubility with the used extraction solvents plays a key role in the detectable concentration. In addition, some metabolites might get degraded and oxidized during the extraction process. Conversely, these limitations are not present with HR-MAS NMR as it works with natural, unaltered, and intact samples at low temperature. Therefore, it likely better mirrors the underlying biochemical activity and state. In the case of kynurenine, this has been reported to have a significant higher level in blood of tuberculosis patients possibly due to an increased level of the enzyme indoleamine 2,3 dioxygenase 1 (IDO1) that converts tryptophan [29]. Tyrosine and mannose levels were previously also shown to be increased in mice and zebrafish samples using NMR analyses [26, 31]. Considering that zebrafish larvae and mouse and human blood samples are very similar in their metabolite profiles after mycobacterial infection [26], the increased level of kynurenine, tyrosine and mannose seen using HR-MAS NMR indicates an advantage of detecting metabolites directly in intact embryos using non-invasive HR-MAS NMR over extracted metabolites using solution NMR. However, a disadvantage of HR-MAS NMR compared to solution NMR is its lower resolution capacity for lipids.

As it is well known, rodents with leptin signaling deficiency show a typical phenotype of fat accumulation and obesity. Phospholipids and polyunsaturated fatty acids (PUFAs) including arachidonic acid and eicosapentaenoic acid are significantly increased in plasma and liver of ob/ob and db/db mice measured by MS [6]. Another study on obese Zucker and lean rats performed by 1H NMR reported increased concentrations of total fatty acids and triglycerides, while the ratio of PUFAs/monounsaturated fatty acids (MUFAs) was decreased in liver and blood of obese rats [11]. In our larval zebrafish lepb mutant, we also found that many lipid peaks are generally higher and for instance levels of DHA and phosphatidylcholines are significantly increased in lepb mutant larvae compared to the wild type siblings (Fig. 6). These observations demonstrate that lepb deficiency in zebrafish leads to lipid accumulation even at the organismal level at the larval stage. The parental adult lepb mutant zebrafish display distinctly more visceral fat compared to wild type sibling fish measured by magnetic resonance imaging (MRI) [32]. As zebrafish larvae before 5dpf only use yolk as their nutrition supply, which comes from the mother, zebrafish larvae offer a promising model to investigate maternal effects of the adult parents on the metabolic state of their offspring in the absence of a feeding regime. We reported previously that adult lepb mutant zebrafish display features of type 2 diabetes mellitus (T2DM) including higher glucose levels and develop early signs of diabetic nephropathy [32]. In this study, we also found that the concentration of glucose is significantly elevated in lepb−/− zebrafish larvae compared to lepb+/+ group in both 1H NMR and HR-MAS NMR measurements. These observations in adult and larval zebrafish could lead to a better understanding of the effects of parents with gestational diabetes mellitus (GDM) on their offspring. GDM is one type of diabetes characterized by high blood pressure and high levels of glucose occurring only during pregnancy. Children from mothers suffering from GDM have a higher risk to develop obesity and T2DM, but also diabetic complications such as kidney disease. Unfortunately, it is impracticable to investigate maternal effects of GDM on offspring in humans and mammal animal models. Zebrafish larvae are therefore promising to explore the maternal effects of T2DM on their offspring as they develop outside the mother’s body [33, 34].

In this study, we demonstrate that 6 of the 13 amino acid metabolites of which the levels are reduced in both mutant ob/ob mice and lepb−/− zebrafish larvae are also biomarkers for tuberculosis infection in human, mice and zebrafish larvae [26]. As it is well known, tuberculosis is also called a consumption disease with severe wasting syndrome symptoms at a later stage in TB patients. Therefore, the similarities between the deficiency of leptin and tuberculosis could be related to the occurrence of wasting syndrome in both ob/ob mice and lepb mutant zebrafish larvae. In this respect, metabolic changes due to leptin deficiency are also relevant for understanding T2DM that is accompanied by wasting syndrome. Of the 30 amino acids levels that we find reduced in the blood of ob/ob mice, several have been reported to be also changed in diabetic mice models in other studies. A decrease in glucogenic amino acids such as alanine, serine, glycine and glutamine indicates a high level of gluconeogenesis in leptin deficient animals. Plasma levels of glycine and serine were found to be significantly decreased in ob/ob mice and db/db mice compared to their wild type controls [6]. Leucine and isoleucine are two branched-chain amino acids (BCAAs) which are reported to stimulate protein synthesis in muscle [35, 36]. In contrast to our study, BCAAs levels were reported to be increased in ob/ob mice and db/db mice [6]. However, a study of human plasma samples demonstrated that the concentrations of the BCAAs, alanine and glutamine were significantly decreased in the plasma of T2DM patients compared to healthy volunteer groups [37]. The similarity of amino acid level changes resulting from leptin deficiency between mammals and zebrafish larvae provides the potential utility of common metabolites as biomarkers for both diabetic parents and their offspring by providing prognostic markers for the early identification of the risks of GDM.

The similarities in changes in metabolite levels resulting from leptin deficiency in different model organisms provide a way to further investigate the mechanism underlying these changes. In a first step towards further functional studies, we investigated the effect of leptin deficiency on the transcriptomic level. Studies have shown wasting syndrome to occur in obese animals as evidenced by muscle mass reduction due to the activation of proteolytic pathways such as the caspase-3 and the ubiquitin–proteasome proteolytic pathways [38, 39]. We also observed the gene ontology (GO) term proteolysis as one of the top GO terms in the transcriptome signature sets of ob/ob mice heads compared to wild type lean mice heads. This GO term was also enriched in the overlap set of this signature set with a signature set that we derived from a published liver transcriptome study of ob/ob mice compared to wild type mice (Fig. 7C). Genes involved in proteolysis can be classified as six types of proteases (Fig. 7D). Multiple proteolytic pathways are shown to be involved in wasting syndrome, including the following enzyme families: cysteine proteases such as calpains, cathepsins, caspases, ubiquitin peptidase families, metallopeptidases, serine proteases and threonine proteases such as proteasome subunit families [40]. Similar to the results obtained with the ob/ob mice body parts, we found that the expression levels of the genes encoding these proteases are significantly changed in lepb mutant zebrafish larvae compared to their wild type siblings (Fig. 7H). This is an indication that the lepb mutation leads to wasting syndrome even at an early stage of zebrafish larval development. It has been reported that amino acids are key regulators of protein turnover [41] and that the depletion of amino acids stimulates proteolysis in differentiated muscle cells [42]. The mechanisms underlying the observed reduced levels of amino acids in ob/ob mice and lepb mutant zebrafish larvae remains to be determined, but could be explained by protein degradation. The significant decrease of many amino acids in ob/ob mice and lepb mutant zebrafish might be a trigger for protein degradation to compensate for the loss of these amino acids.

In zebrafish larvae, both saturated fatty acids and polyunsaturated fatty acid DHA are increased in the lepb mutant group. DHA is an omega-3 fatty acid which is a precursor of eicosanoids such as resolvins and protectins with potential anti-inflammatory activity [43]. In contrast, omega-6 PUFA arachidonic acid (ARA) is a key precursor for eicosanoids such as prostaglandins, thromboxanes and leukotrienes which mediate inflammatory response [44]. Peak 11 of the spectra (Fig. 6A) could represent the PUFA arachidonic acid. However, the relatively low abundance and the overlap with the peak of eicosapentaenoic acid (EPA) made it hard to quantify the concentration in the two groups. In zebrafish larvae, genes such as PTGS2, PTGIS, involved in the generation of prostaglandins in the cyclo-oxygenase (COX) pathway are downregulated in lepb−/− compared to the lepb+/+ group (Fig. 8, Additional file 1: Table S8). This might be the result of the anti-inflammatory effect of an increased level of DHA observed in lepb mutant zebrafish larvae. In ob/ob mice head and liver, genes like PLA2G4A, ALOX5AP, DPEP1 involved in the release of ARA from cell membrane and lipoxygenase (LOX) pathway are significantly upregulated (Fig. 8, Additional file 1: Table S8). Therefore, more leukotrienes are expected to be produced, which leads to a potential inflammatory state. This is consistent with the generally accepted concept that obesity and type II diabetes are accompanied with chronic, low-grade inflammation [45]. This is in line with the previously shown correlation of leptin deficiency and diabetes with a higher susceptibility to tuberculosis [46]. Furthermore, it has been shown that zebrafish larvae and humans respond in a very similar way to infection with mycobacteria, for instance in the activation of the prostaglandin pathway [47, 48]. Therefore, the opportunities for future studies of the common mechanism underlying wasting syndrome in various disease such as T2DM and infectious disease in zebrafish larvae are extremely promising for leading to understand human diseases.

Conclusion

Leptin deficiency in adult mice and larval zebrafish leads to highly similar metabolic alterations in amino acid levels. These metabolic changes show the same key features as observed during progression of tuberculosis in human patients, rodents and zebrafish larvae. This conclusion is supported by different technologies, namely MS, solution-state NMR and HR-MAS NMR. Moreover, by studying the transcriptome, we found highly similar changes in gene regulation related to proteolysis and arachidonic acid pathways in these two test systems. These results show a remarkable similarity of the effects of leptin knockdown on the metabolomes and transcriptomes of adult mice and zebrafish larvae that might be related to wasting syndrome. Apparently, the metabolic control by leptin is similar in adult and embryonic stages in mammals and fish, respectively.

Materials and methods

Biological materials

Mice

Male ob/ob mice and lean C57BL/6 wild type mice were obtained from Charles River Laboratories at 6 weeks of age (n = 8 per group) and maintained for 8 weeks under specific pathogen free conditions in the animal facility of the Leiden University Medical Center (LUMC). Male mice were chosen because metabolic variation due to the hormonal cycle is limited. Mice were kept on a standard-chow diet with ad libitum access to food and water. One ob/ob mouse had to be sacrificed at an early stage due to malocclusion. Body weight of all mice was measured weekly. Mice were sacrificed at week 14 and blood was collected and heads were snap-frozen in liquid nitrogen and stored at − 80 °C until RNA isolation. Mice heads were taken as a body part of interest because of the known classical signaling of leptin in the brain. Handling of mice was conducted in compliance with European Community Directive 86/609 for the care and use of laboratory animals and in accordance with the regulations set forward by the LUMC animal care committee.

Mouse serum sample preparation

Mouse serum samples were collected from clotted blood tubes and mixed with pre-heated 80% ethanol at a 1:3 ratio (end concentration: 60% ethanol) in polypropylene screwcap tubes. Samples were heated for 10 min at 90 °C and subsequently chilled on ice for 10 min before centrifugation at 13.000 rpm for 10 min at 4 °C. Supernatants were harvested and stored at − 80 °C for LC–MS analysis.

Zebrafish larvae

Zebrafish were handled in compliance with the local animal welfare regulations and maintained according to standard protocols (http://zfin.org). Mutant lepb−/− and wild type sibling lepb+/+ zebrafish lines were generated, screened and raised as described previously [32]. A lepb mutant with a seven base pair deletion encompassing TAGAGGG in exon 2 was used in this study. Zebrafish larvae at 5 dpf from lepb−/− and lepb+/+ groups were collected and stored at − 80℃ until further analysis. For solution-state NMR measurement, four replicate samples per genotype comprised of 105 pooled larvae were taken. From the same batch, three replicate samples per group of 15 pooled larvae were used for RNA isolation and transcriptome analysis. For HR-MAS NMR measurement, three replicates of 120 pooled larvae were used (each sample was measured three times).

LC–MS/MS

Metabolite levels in mice serum were measured in individual replicates using a targeted LC–MS/MS platform as described before [26, 29]. Subject numbers were randomized and run in five batches which included a calibration line, QC samples and blanks. QC samples were analyzed every 10 samples. They were used to assess data quality and to correct for instrument responses.

The amine platform covers amino acids and biogenic amines employing an Accq-Tag derivatization strategy adapted from a previously published protocol [49]. Briefly, 5.0 μl of each sample was spiked with an internal standard solution. Then proteins were precipitated by the addition of MeOH. The supernatant was dried in a speedvac. The residue was reconstituted in borate buffer (pH 8.5) with AQC reagent. 1.0 μl of the reaction mixture was injected into the UPLC-MS/MS system. Chromatographic separation was achieved by an Agilent 1290 Infinity II LC System on an Accq-Tag Ultra column. The UPLC was coupled to electrospray ionization on a triple quadrupole mass spectrometer (AB SCIEX Qtrap 6500). Analytes were detected in the positive ion mode and monitored in Multiple Reaction Monitoring (MRM) using nominal mass resolution. Acquired data were evaluated using MultiQuant Software for Quantitative Analysis (AB SCIEX, Version 3.0.2). The data are expressed as relative response ratios (target area/ISTD area; unit free) using proper internal standards. For analysis of amino acids, their 13C15N-labeled analogs were used. For other metabolites, the closest-eluting internal standard was employed. In-house developed algorithms were applied using the pooled QC samples to compensate for shifts in the sensitivity of the mass spectrometer over the batches. After quality control correction, metabolite targets complied with the acceptance criteria of RSDqc  < 15%. Using this platform, we were able to identify 41 metabolites in blood samples from mice.

MS data analysis

Data was analyzed using the software package MetaboAnalyst 4.0 [50]. MetaboAnalyst offers the possibility to provide automated data reports which we used for archiving data sets. Default settings were used with log transformation and auto scaling of the data for normalization. Naming of the metabolites is based on reference compounds using standard nomenclature of the human metabolome database (https://www.hmdb.ca/).

1H solution NMR measurement of extracted larvae

For 1H solution NMR spectroscopy, metabolites from pooled zebrafish larvae were extracted according to a previous study [26]. Zebrafish larvae were crushed and 1 ml mixture of methanol: water (1:1, v/v) and 1 ml chloroform were immediately added to the sample. The mixture was sonicated for 15 min and then centrifuged at 5000 rpm for 5 min. After centrifugation, two layers were formed: the upper layer is methanol and water containing metabolites, the lower layer is chloroform containing lipids. Those two layers were separately collected and evaporated via nitrogen gas flow. The metabolite pellets were resuspended in 600 μl of 100 mM deuterated phosphate buffer (KD2PO4, PH = 7.0) containing 0.02% trimethyl-silylpropanoic acid (TSP) as a reference and was subsequently centrifuged, and the supernatant was analyzed by solution NMR. The lipid pellets were resuspended in 600 μl deuterated chloroform containing 0.03% TMS which was used as a reference. Metabolites and lipids in zebrafish larvae were measured with a Bruker DMX 600 MHz NMR spectrometer at 4 °C equipped with a 5 mm inverse triple high-resolution probe with an actively shielded gradient coil. The 1H NMR spectra were accumulated with 65,000 data points, a 2-s relaxation delay, a sweep width of 12.4 kHz, and 256 scans which were required to obtain a satisfactory signal-to-noise ratio.

1H HR-MAS NMR measurement of intact larvae

Metabolic profiling by 1H HR-MAS NMR spectroscopy was performed as adapted from previous studies [51,52,53]. Zebrafish larvae from lepb+/+ and lepb−/− groups were carefully transferred to a 4-mm zirconium oxide MAS NMR rotor (Bruker BioSpin AG, Switzerland). As a reference (1H chemical shift at 0 ppm), 10 µl of 100 mM deuterated phosphate buffer (KD2PO4, PH = 7.0) containing 0.1% (w/v) TSP was added to each sample. The rotor was then placed immediately inside the NMR spectrometer.

All HR-MAS NMR experiments were done on a Bruker DMX 600-MHz NMR spectrometer, which was equipped with a 4-mm HR-MAS dual inverse 1H/13C probe with a magic angle gradient and spinning rate of 6 kHz with a proton resonance frequency of 600 MHz. Measurements were carried out at a temperature of 277 K using a Bruker BVT3000 control unit. Acquisition and processing of data were done with Bruker TOPSPIN software 2.1 (Bruker Analytische Messtechnik, Germany).

A rotor synchronized Carr–Purcell–Meiboom–Gill (CPMG) pulse sequence with water suppression was used for one-dimensional 1H HR-MAS NMR spectra. Each one-dimensional spectrum was acquired applying a spectral width of 8000 Hz, domain data points of 16 k, a number of averages of 512 with 8 dummy scans, a constant receiver gain of 2048, an acquisition time of 2 s, and a relaxation delay of 2 s. The relaxation delay was set to a small value to remove nascent short transverse (T2) components due to the presence of lipids in intact embryo samples. All spectra were processed by an exponential window function corresponding to a line broadening of 1 Hz and zero-filled before Fourier transformation. NMR spectra were phased manually and automatically baseline corrected using TOPSPIN 2.1. The total analysis time (including sample preparation, optimization of NMR parameters, and data acquisition) of 1H HR-MAS NMR spectroscopy for each sample was approximately 20 min.

NMR analysis

The one-dimensional 1H solution NMR and HR-MAS NMR spectra obtained from lepb−/− and lepb+/+ group were corrected for baseline, phase shifts and reference using the MestReNova software version 11.0 (Mestrelab Research S.L., Santiago de Compostela, Spain). The region of 4.8–4.9 (solution NMR) was excluded from the analysis to remove the water peak. The spectra were then subdivided in the range between 0 and 10 ppm into buckets of 0.04 ppm. The resulting data matrix was saved as the format of script: NMR CSV matrix (transposed) (*.CSV, *.txt). This was then imported into MetaboAnalyst 4.0 for multivariate analysis using PLS-DA. Correlation coefficients with p < 0.05 were considered statistically significant. Quantification of metabolites was performed using Chenomx NMR Suite 8.6 (Edmonton, Alberta, Canada), which allowed for qualitative and quantitative analysis of an NMR spectrum by fitting spectral signatures from HMDB database to the respective spectrum. Assignment of peaks was based on the chemical shifts of compounds of interest in Chenomx software. The concentration of lipids was calculated by comparing the integral peak intensity of the lipids of interest with that of the reference TMS peak [54]. Statistical analysis (t-tests) of the NMR quantification results was performed with GraphPad Prism 8.0.1 (San Diego, CA, USA) and p-values < 0.05 were considered significant.

RNA isolation

Frozen ob/ob and C57BL/6 mouse heads (n = 4) were thawed in 30 ml of TRIzol Reagent (Life Technologies) and manually crushed in a mortar while zebrafish larvae from lepb+/+ and lepb−/− groups (n = 3) were resuspended and crushed in 0.5 ml of TRIzol Reagent. Subsequently, total RNA was extracted in accordance with the manufacturer’s instructions. Contaminating genomic DNA was removed using DNase I digestion for 15 min at 37 °C. RNA concentration was determined by NanoDrop 2000 (Thermo Scientific, the Netherlands). RNA integrity (RIN) was assessed by bioanalyzer (Agilent) and samples with RIN values > 6 were used for further library construction and sequencing.

Deep sequencing

Mice

Deep sequencing of total RNA samples derived from ob/ob and lean C57BL/6 mice heads was performed at ZF-screens B.V. (Leiden, the Netherlands) as described in a previous study [55]. A total of 3 μg of RNA was used to generate RNA-seq libraries using the Illumina TruSeq RNA Sample Preparation Kit v2 (Illumina Inc., San Diego, USA). In the manufacturer’s instructions two modifications were made: In the adapter ligation step 1 μl instead of 2.5 μl adaptor was used; In the library size selection step, the library fragments were isolated using a double Ampure XP purification with a 0.7 × beads to library ratio. The resulting mRNA-seq libraries were sequenced using an Illumina HiSeq2000 instrument according to the manufacturer’s description with a read length of 50 nucleotides. Image analysis and base calling were done by the Illumina HCS version 1.15.1. At least 15 million reads were obtained that could be mapped to the mouse genome version GRCm38.

Zebrafish larvae

Deep sequencing of the zebrafish larvae RNA from lepb+/+ and lepb−/− groups was performed by GenomeScan B.V. (Leiden, the Netherlands). The NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (NEB #E7760S/L) was used to process the samples. Briefly, mRNA was isolated from total RNA using oligo-dT magnetic beads. After fragmentation of the mRNA, a cDNA synthesis was performed. This was used for ligation of the sequencing adapters and PCR amplification of the resulting product. The quality and yield after sample preparation was measured with Fragment Analyzer. The size of the resulting products was consistent with the expected size distribution (a broad peak between 300 and 500 bp). Clustering and DNA sequencing using the NovaSeq6000 was performed according to manufacturer’s protocols. A concentration of 1.1 nM of DNA was used. For the zebrafish larval samples, data sets of paired end reads of 150 nucleotides were obtained with at least 20 million reads of reads that could be mapped to the zebrafish genome version GRCz11.

Deep sequencing data map** and analysis

Sequencing data of mice heads were aligned and mapped to the mouse genome GRCm38.p6 using Genetiles server [55]. Sequencing data of zebrafish larvae were aligned and mapped to the zebrafish genome GRCz11 using Salmon v1.2.1, and differential gene expression was analyzed using DESeq2 v1.21.1. Gene Ontology (GO) term enrichment and KEGG pathway analysis were performed in DAVID Bioinformatics Resources 6.8 (https://david.ncifcrf.gov/). The arachidonic acid pathway of Fig. 8 was drawn in Pathvisio software based on the wikipathways eicosanoid synthesis, eicosanoid metabolism via cytochrome P450 mono-oxygenases (CYP), prostaglandin synthesis, and omega3 and omega6 fatty acids synthesis [56]. Genes MAP2K6 and Nfe2l2 were added to the pathway based on literature [57, 58].

qPCR

Zebrafish larvae cDNA was generated from the same RNA samples of RNAseq by using iScript cDNA synthesis kit (Bio-Rad). qPCR experiment was performed by following a protocol of SsoAdvanced Universal SYBR® Green Supermix kit (Bio-Rad). qPCR measurement was detected on a CFX96 machine (Bio-Rad). The Cq values of targeted genes were normalized to a zebrafish housekee** gene Tsp as the expression level was not changed due to lepb mutation. The relative expression level were analyzed by using 2−ΔΔCt method. We selected the representative genes based on the fold change, expression level, p adjusted value and the ease to make good primers. The forward and reverse primer sequences of tested genes in zebrafish larvae are showing below. LO018181.1: TGAAGCGACTGGGATGCTG/TGGATCTCTTCGTTCAAGGGTT. Si:dkey-14d8.6: ACTCCTATGATCAGCCCCTG/TTACAGCCAAACTCCCACACC. Amy2al2: AGCACAACCCAAACACGAAA/CTGAACTCCTCCATAGCCGT. Tsp: CCTGCCCATTTTCAGTC/TGTTGTTGCCTCTGTTGCTC.