The study was approved by Orentreich Foundation for Advancement of Science, Inc's. Institutional Animal Care and Use Committee and performed in accordance with the National Institutes of Health guidelines for use of animals in research laboratories.

Four-week-old male Fischer-344 rats from Taconic Farms (Germantown, NY) were maintained one rat per cage in a controlled 12 h light/dark cycle with free access to water and a standard diet (LabDiet 5001; PMI Nutrition International, LLC, Brentwood, MO) for 2 weeks. The rats were then randomly assigned to one of four dietary groups consuming chemically defined AIN-76-based diets with protein replaced by different amino acid mixtures (Dyets Inc.; Bethlehem, PA) for 12 weeks: 1) Control-fed (CF, 0.86% methionine); 2) methionine-restricted (MR, 0.17% methionine); 3) CF supplemented with 0.5% L-cysteine (CF+Cys); and 4) MR supplemented with 0.5% L-cysteine (MR+Cys). Certificates of analysis indicated that the CF+Cys and MR+Cys diets contained 0.453% and 0.476% (by weight) cysteine, respectively. The unsupplemented control and MR diets were devoid of cysteine and other sulfur amino acids apart from methionine as described previously (²⁹). To compensate for the reduced amino acid content in the MR diet, glutamic acid content was raised on an equal weight basis. Food and water were provided ad libitum.

Body weight was registered at baseline and weekly throughout the experiment. Food intake was measured weekly by weighing residual food at the end of a 48 h period, and the average intake over 24 h was calculated. At baseline and after 6 weeks and 12 weeks on the dietary regimen, nonfasting blood samples were collected from the subclavian vein after anesthetizing the rats using an E-Z Systems apparatus (Palmar, PA).

After completing the dietary regimens for 12 weeks, the rats were euthanized by Euthanex E-Z Systems. White adipose tissue fat pads (inguinal, epididymal, peri-intestinal, retroperitoneal), interscapular brown fat (obtained by dissecting away the white superficial fat and harvesting the deeper brown fat between the shoulder blades), liver, and skeletal (quadriceps) muscle were immediately harvested, weighed, frozen in liquid nitrogen, and stored at −80°C.

Liquid chromatography tandem mass spectrometry was used to analyze total homocysteine (tHcy), methionine, cystathionine, tCys, and total glutathione (tGSH) concentrations (³⁰). Taurine and glutamate were measured by a modification of a previously described liquid chromatography tandem mass spectrometry method (²⁸).

Serum concentrations of glucose, lipids, albumin, and protein were measured on a Beckman Coulter AU400 semi-automated clinical chemistry analyzer (³¹). Commercially available rat ELISA kits were used to measure leptin (Assay Design, Ann Arbor Michigan), IGF-1 (IDS, Fountan Hills, AZ), insulin (Alpco, Salem, NH), and adiponectin (Millipore, Billerica, MA).

Fatty acid profiles in serum total lipids were assayed by gas liquid chromatography with flame ionization detection at AS Vitas, Oslo Innovation Center, Oslo, Norway (www.vitas.no). The method has been validated for quantification of polyunsaturated fatty acids according to US FDA Guidance Document on Bioanalytical Method Validation. In a separate prevalidation study, efficacy of the direct transmethylation step has been found to be 96–104% for the lipid classes cholesterol esters, triglycerides, and phospholipids. Fatty acid content was calculated based on the area % of peaks and response factors relative to stearic acid (C18:0).

Desaturase and elongase activity indices were estimated as product/precursor ratios of fatty acids in serum (³²). SCD1-16 activity was estimated as 16:1n-7/16:0, SCD1-18 activity as 18:1n-9/18:0, Δ⁶-desaturase (D6D) activity as 18:3n-6/18:2n-6; Δ⁵-desaturase (D5D) activity as 20:4n-6/20:3n-6; and elongase activity as 18:0/16:0 (³²).

Total RNA was isolated from inguinal white adipose tissue, liver, and quadriceps muscle tissue using Qiagen's RNeasy kits. RNA concentrations were determined spectrophotometrically and their integrity was assessed by agarose gel electrophoresis. MRNA (mRNA) was transcribed into cDNA using a High-Capacity cDNA Reverse Transcription Kit in a Perkin-Elmer (Wellesley, MA) GeneAmp PCR System 9600 as previously described (²¹). Multiplexed PCR was conducted in a StepOnePlus Real-Time PCR System with the prepared cDNA and commercially available TaqMan primer-probe sets (Life Technologies, Carlsbad CA). Gene expression was assessed by the comparative CT (ΔΔCT) method with β-actin as the reference gene. β-actin CT values were similar in tissues from each group, and efficiencies for β-actin and the genes of interest in the multiplexed reactions were greater than 90% and comparable to each other.

Data is presented as mean (± SEM). Groups were compared by one-way ANOVA followed, where relevant (P ANOVA < 0.05), by pairwise comparisons: MR versus CF, MR+Cys versus CF, CF+Cys versus CF, and MR versus MR+Cys. All tests were 2-tailed and P-values < 0.05 were considered significant. Analyses were performed using the Statistical Package for Social Sciences 12.0 for Windows (SPSS, Chicago. IL).

MR markedly lowered weight gain, as previously reported (²¹, ²², ²⁸), despite food intake/g body weight being consistently higher than CF group throughout the study (Fig. 1A, B). Supplementing the MR diet with cysteine (MR+Cys) completely rescued weight gain despite decreasing food intake/body weight to CF levels, but no increase in weight gain above CF levels was observed in the CF+Cys group. Food intake/body weight in the CF+Cys group tended to be lower than CF throughout the study, the difference being significant at weeks 3 and 7 (Fig. 1A, B).

Fat pad mass expressed as a percent of body weight was reduced in all depots from MR rats, except for brown adipose tissue mass, which was larger than in CF rats (Fig. 1C). Compared with the MR group, MR+Cys rats featured increased white adipose tissue fat pad mass % to, or toward, control levels, and reduced brown fat mass. In contrast, cysteine supplementation of the control diet (CF+Cys) promoted a modest but significant reduction of white adipose tissue mass in all fat depots with no effect on brown fat mass (Fig. 1C).

Overall, supplementing the MR diet with cysteine restored weight gain and adiposity to control levels, whereas adding cysteine to the control diet tended to decrease food intake and modestly decreased adiposity with no effect on total weight gain.

Table 1 shows serum metabolites at baseline and after 12 weeks on the diets. MR had mixed favorable and unfavorable effects on serum lipids. Triglycerides and total cholesterol were lowered compared with control, confirming previous reports (²¹, ²²). However, LDL-C was approximately 25% higher, and HDL-C was approximately 60% lower in MR compared with CF rats. The addition of cysteine to the MR diet (MR+Cys) prevented the triglyceride lowering and LDL elevation induced by MR. LDL-C was reduced in MR+Cys to levels below those observed in MR and CF rats.

Total serum proteins and serum albumin were decreased in MR rats and were restored by cysteine supplementation (MR+Cys) to control levels. No difference in any of the serum lipid or protein parameters was observed between the CF and CF+Cys groups.

In summary, supplementation of the MR diet with cysteine prevented most of the MR effects on serum lipids and protein markers, whereas addition of cysteine to the control diet had no effect on these parameters.

The fatty acid profile in total serum lipids was determined after 12 weeks of dietary intervention. Marked differences were noted between the MR and CF groups, which were essentially blunted by cysteine (MR+Cys) (Table 2). MR lowered the levels of monounsaturated fatty acids palmitoleic (16:1,n-7) and oleic acids (18:1,n-9), and raised stearic acid (18:0), with consequent decreases in estimated indices of SCD1-16 and SCD1-18 desaturase activities (Fig. 2). These effects were reversed by cysteine supplementation of the MR diet (MR+Cys). Furthermore, cysteine supplementation of the MR diet markedly lowered estimated indices of D5D, D6D, and elongase activities (Fig. 2).

MR also raised the omega 6 (n-6) polyunsaturated fatty acids γ-linoleic acid (18:3, n-6) and docosadienoic acid (22:2, n-6), and reduced eicosadienoic acid (20:2, n-6). These effects were also reversed by cysteine. In contrast, cysteine supplementation of the control diet produced no changes in the fatty acid profile (Table 2).

Serum hormones and adipokines were measured at baseline and after 6 and 12 weeks of dietary intervention (Fig. 3). As previously observed (²¹, ²³), MR significantly decreased serum concentrations of insulin, leptin, IGF-1, and raised adiponectin compared with CF. All these effects were reversed by cysteine (MR+Cys) at 12 weeks. Compared with the CF group, CF+Cys rats had lower serum leptin at 6 and 12 weeks (Fig. 3), consistent with the marginal reduction in fat pad mass (Fig. 1). Paradoxically, the CF+Cys group also had lower adiponectin levels compared with CF rats.

Thus, cysteine reversed the MR effects on serum insulin, IGF-1, leptin, and adiponectin, while adding cysteine to the control diet lowered serum leptin.

Serum sulfur amino acids and glutamate were measured at baseline and at 6 and 12 weeks on the diets. As previously reported (²⁸) and shown in Fig. 4, MR markedly decreased serum methionine, cystathionine, tCys, and taurine compared with CF, and elevated tHcy. Serum glutamate was not different in MR compared with CF rats. No difference in serum tGSH was observed among the dietary groups at any time-point (P ANOVA at 12 weeks = 0.64; data not shown).

Supplementation of the MR diet with cysteine (MR+Cys) restored plasma tCys and cystathionine to control levels, but only partly raised taurine levels and partly lowered tHcy. After 12 weeks of intervention, serum tHcy in MR+Cys group (25.5 ± 2.2 μmol/L) was lower (P < 0.001) compared with MR (37.7 ± 3.4 μmol/L), but remained higher (P = 0.016) than CF (18.2 ± 0.6 μmol/L). Serum methionine was not raised by adding cysteine to the MR diet (Fig. 4).

No consistent difference in sulfur amino acid levels was observed between the CF+Cys and CF groups, except for an approximate 25% and 15% increase of serum taurine at 6 weeks and 12 weeks, respectively (P ≤ 0.005 at both time points; Fig. 4). Notably, serum tCys was not higher in CF+Cys versus CF after 6 or 12 weeks on the diets (Fig. 4).

In summary, supplementing the MR diet with cysteine restored serum tCys, cystathionine, tHcy, and taurine to or toward control levels, with no effect on methionine. Adding cysteine to a control diet failed to increase serum tCys but raised serum taurine.

We examined changes in expression of some genes known to be responsive to MR (Fig. 5) (²¹). Consistent with previous reports, MR enhanced hepatic expression of peroxisome proliferator activated receptor g (Pparg), and markedly suppressed Scd1. Cysteine (MR+Cys) reversed or partly reversed these effects. However, cysteine supplementation of the control diet (CF+Cys) had no effect on either gene.

In contrast to hepatic changes, Scd1 expression was greatly induced in white adipose tissue from MR rats, as previously observed (²¹). Cysteine (MR+Cys) lowered adipose tissue Scd1 expression to control levels, and further suppressed adipose tissue Scd1 expression below control values when added to the control diet (CF+Cys). Supplementing the control diet with cysteine (CF+Cys) also decreased the expression of cytochrome c oxidase subunit IV (Cox4i1), Pparg, and fatty acid-binding protein 4 (Fabp4) in adipose tissue.

In skeletal muscle, cysteine reversed a 2-fold upregulation of uncoupling protein 3 (Ucp3) induced by MR, and also modestly decreased expression of carnitine palmitoyl transferase 1b (Cpt1b) (CF+Cys vs. CF and MR+Cys vs. MR) and Cox4i (MR+Cys vs. MR).

Low-methionine dietary rodent models feature lower weight gain and fat mass% and higher energy expenditure and insulin sensitivity concomitant with decreased adipocyte size and profound inhibition of hepatic Scd1 expression (²¹–²³, ²⁵, ²⁶). This phenotype is independent of energy intake as demonstrated by pair-feeding (²³) and by our observation that MR rats consumed more food per gram body weight than CF rats. Decreased BMI despite higher energy intake and adequate dietary protein was also observed in adults on a MR diet for treatment of cancer (³³). It has also been suggested that vegetarianism is a “mild” form of MR (¹⁴, ³⁴), which may partly explain the lower weight gain and diabetes incidence among vegetarians (⁷, ⁸, ¹⁰–¹²), in whom sulfur amino acid intake is typically ∼40–70% of omnivores (³⁵).

The present study extends previous findings of decreased hepatic Scd1 mRNA and protein levels in MR rodents (²¹, ²⁵) by demonstrating decreased serum concentration of the SCD1 products palmitoleic and oleic acids, with concomitant reduction in estimated SCD1 activity indices in MR rats. The observed reduction in serum TG and MR rats is also consistent with Scd1 suppression (³⁶), as synthesis of monounsaturated fatty acids by SCD1 is a prerequisite for their incorporation into triglycerides (²⁷). Cysteine supplementation blocked MR effects on Scd1 expression, SCD1 indices, weight gain, fat mass%, and serum triglycerides and leptin. Serum insulin was higher, and adiponectin markedly lower in MR+Cys compared with MR rats, suggesting that cysteine may prevent the enhanced insulin sensitivity of MR rats (²³). Consistent with this, high plasma tCys was recently reported in insulin resistant subjects (³⁷).

In mice, SCD1 activity index correlates with BMI and fat mass% (³⁸). The reduced adiposity in MR rats and its reversal by cysteine paralleled the changes in Scd1 expression and function, suggesting that SCD1 changes mediate these phenotypes. A shift from lipogenesis toward β-oxidation via activating hepatic AMP-activated protein kinase underlies the hypermetabolism induced by hepatic SCD1 inhibition (²⁷, ³⁹). Further, cysteine reduced skeletal muscle CPT1b mRNA, which is translated into the rate-limiting enzyme in transport of long-chain fatty acid into mitochondria for β-oxidation (⁴⁰). Cysteine also reversed the MR-induced up-regulation of muscle Ucp3. Mice over-expressing Ucp3 are hypermetabolic and lean, at least partly due to uncoupling of substrate oxidation from ATP production (⁴¹), although whether UCP3 has a physiological function in regulating energy expenditure is uncertain (⁴²). Overall, cysteine induced transcriptional changes that favored fatty acid partitioning away from β-oxidation.

The possible contribution of the integrated stress response to the SCD1 and fat mass changes of MR rats deserves investigation. Integrated stress response, triggered by numerous factors including imbalance or deprivation of essential amino acids, involves phosphorylation of the α-subunit of the eukaryotic initiation factor 2 complex (eIF2α). Phosphorylated eIF2α induces subsequent changes in transcription and translation of genes involved in adaptations to dietary stress (⁴³). It was recently shown that MR rats feature enhanced eIF2α phosphorylation in conjunction with upregulation of various components of the integrated stress response (⁴⁴). The same pathway was found to mediate both SCD1 suppression and decreased abdominal fat mass in response to leucine deprivation (⁴⁵). This raises the possibility that the changes in SCD1 and fat mass observed in MR rats, and their reversal by cysteine, may be mediated by elements of the integrated stress response.

One limitation of this study is that SCD1 indices were calculated from fatty acid profiles of total serum lipids, rather than of VLDL, which better reflects hepatic Scd1 expression (³²). However, measuring fatty acid profile in specific lipid fractions is particularly indicated in human studies in which SCD1 indices are used to infer gene expression (³², ⁴⁶). In the present study, Scd1 mRNA was measured in liver and white adipose tissue and the serum fatty acid profile clearly coincided with hepatic gene expression. In contrast to liver, adipose tissue showed enhanced Scd1 expression in MR rats, an effect that was reversed by cysteine supplementation. Cysteine also markedly decreased the activity indices of elongase, D5D, and D6D. Although these serum indices have been linked to BMI, serum lipids, and risk of metabolic syndrome (⁴⁷, ⁴⁸), it has been shown that, unlike SCD1 index, they are of little value in predicting corresponding indices or enzyme gene expression in the liver (³²).

Both serum methionine and tCys are decreased in MR rats (²⁸). Cysteine blocked the effects of MR without increasing serum methionine. This suggests a relatively direct effect of cysteine rather than a methionine-sparing action of cysteine (⁴⁹), and implicates lowered cysteine as the cause of reduced adiposity in MR. However, cysteine is the precursor of many biologically active compounds including taurine, GSH, and sulfate (¹⁶). A recent study showed decrease in weight gain and in hepatic lipogenesis in rats treated with an inhibitor of the rate-limiting enzyme in GSH synthesis, γ-glutamylcysteine synthetase (⁵⁰). We observed no difference in plasma tGSH between the groups, but MR rats feature hepatic GSH depletion (⁵¹) similar to that observed with γ-glutamylcysteine synthetase blockade (⁵⁰). Thus, changes in hepatic GSH synthesis may have contributed to the metabolic impacts of MR and cysteine supplementation.

Cysteine supplementation of the control diet did not induce obesity. Compared with the CF group, CF+Cys rats showed no difference in serum tCys, or in hepatic SCD1 expression or activity indices, and had slightly lower fat mass%. We postulate that this response may have resulted from the marginally decreased food intake in CF+Cys rats, as observed by others (⁵², ⁵³). Kawayami et al. (⁵³) supplemented the same dose of L-cysteine used in the present study (0.5%) to four diets of different protein quality and quantity and found that cysteine reduced or tended to reduce food consumption. Further, cysteine enhanced weight gain only in groups where plasma tCys was increased above the unsupplemented group (⁵³). Similarly, in the present study, supplementary cysteine enhanced SCD1 function and adiposity only in the model in which plasma tCys increased from subnormal (MR) to normal (MR+Cys) levels. Cysteine supplementation in CF+Cys rats neither increased tCys nor adiposity above CF levels, and thus failed to replicate the elevated plasma tCys that is linked to human obesity (¹⁷, ¹⁸). This may result from the role of the liver, demonstrated in rodents (⁵⁴), in disposing of excess dietary cysteine via conversion to taurine, as shown by raised serum taurine in CF+Cys rats.

It is tempting to speculate that the lean phenotype associated with low serum tCys in MR rats (²⁸) and the human obesity associated with elevated plasma tCys (¹⁷, ¹⁸) are related processes. However, the evidence for this remains limited, and different mechanisms may be involved. Our data from rats indicates that MR increases energy expenditure by triggering a cellular response in which lipid synthesis and degradation pathways are simultaneously activated, initiating energy-consuming “futile cycles” (²²). In contrast, mice fed a high cystine diet feature decreased energy expenditure (Elshorbagy et al., unpublished observations), but whether the molecular pathways mirror those of MR remains to be determined. Further, in the present study the effect of dietary cysteine on SCD1 function and fat mass accretion clearly showed a ceiling effect that stopped at control values for these parameters. This is at odds with human findings of a linear dose-response relation between tCys and body fat extending well into the obese range of BMI and fat mass (¹⁷, ¹⁸). The discrepancy could reflect a species difference in body weight response to excess dietary cysteine, or could alternatively indicate that high tCys in humans is not driven by excess cysteine intake.

Apart from the recognized induction of SCD1 in response to glucose oversupply to activate de novo lipogenesis (⁵⁵), available data on nutritional regulation of SCD1 is limited. Our observation that SCD1 expression and activity indices are strongly responsive to dietary sulfur amino acid composition, with corresponding impact on adiposity and metabolic profile, is a key finding linking dietary protein quality with SCD1 regulation. Whether moderate inter-individual differences in sulfur amino acid intake, as observed between vegetarians and omnivores (³⁵), might influence weight gain and metabolic risk (¹⁰, ¹²) through an effect on SCD1 function deserves investigation.

Amany K. Elshorbagy: Concept, study design, statistical analysis and interpretation, preparation of the manuscript. Maria Valdivia-Garcia: Development of liquid chromatography tandem mass spectrometry method for taurine and glutamate assay, conduction of all mass spectrometry assays. Dwight A. L. Mattocks: Conduction of body and tissue data collection as well as gene expression analysis, review of the manuscript. Jason D. Plummer: Primary contact person regarding formulation of the cysteine supplemented diets as well as conduction of body and tissue data collection. A. David Smith: Concept, study design, critical revision of the manuscript. Christian A. Drevon: Study design related to fatty acid metabolism, critical revision of the manuscript. Helga Refsum: Concept, study design, critical revision of the manuscript. Carmen E. Perrone: Study design, coordination of MR studies, conduction of MR experiment, tissue collections and serum hormone assays and critical revision of the manuscript.