Long-Acting PASylated Leptin Ameliorates Obesity by Promoting Satiety and Preventing Hypometabolism in Leptin-Deficient Lep^ob/ob Mice

Abstract

Body weight loss of Lep^ob/ob mice in response to leptin is larger than expected from the reduction in energy intake alone, suggesting a thermogenic action of unknown magnitude. We exploited the superior pharmacological properties of a novel long-acting leptin prepared via PASylation to study the contribution of its anorexigenic and thermogenic effects. PASylation, the genetic fusion of leptin with a conformationally disordered polypeptide comprising 600 Pro/Ala/Ser (PAS) residues, provides a superior way to increase the hydrodynamic volume of the fusion protein, thus retarding kidney filtration and extending plasma half-life. Here a single PAS(600)-leptin injection (300 pmol/g) resulted in a maximal weight reduction of 21% 6 days after application. The negative energy balance of 300 kJ/(4 d) was driven by a decrease in energy intake, whereas energy expenditure remained stable. Mice that were food restricted to the same extent showed an energy deficit of only 220 kJ/(4 d) owing to recurring torpor bouts. Therefore, the anorexigenic effect of PAS(600)-leptin contributes 75% to weight loss, whereas the thermogenic action accounts for 25% by preventing hypometabolism. In a second experiment, just four injections of PAS(600)-leptin (100 pmol/g) administered in 5- to 6-day intervals rectified the Lep^ob/ob phenotype. In total, 16 nmol of PAS(600)-leptin per mouse triggered a weight loss of 43% within 20 days and normalized hypothermia and glucose homeostasis as well as hepatic steatosis. The beneficial properties of PAS(600)-leptin are substantiated by a comparison with previous studies in which approximately 400 nmol (∼25-fold) unmodified leptin was mandatory to achieve similar improvements.

The adipocyte-derived protein hormone leptin is a key factor in energy balance regulation (1, 2). Leptin receptors in the brain are expressed in regions well characterized for a role in metabolic control (3). Low plasma leptin levels indicate a state of negative energy balance and adipose tissue depletion. Leptin deficiency caused by rare loss-of-function mutations results in hyperphagia and severe obesity (4). In cases of functional leptin deficiency, leptin replacement therapy is a successful approach to restore blood leptin levels and to ameliorate the disease phenotype (5, 6).

Beyond treatment of congenital leptin deficiency, leptin therapy mediates beneficial effects on glucose metabolism, insulin sensitivity, and plasma triglyceride levels, thus raising prospects for intervention of type 1 or type 2 diabetes as well as hypoleptinemic disease states such as hypothalamic amenorrhea and lipodystrophy (7–9). Recombinant human leptin, currently distributed as metreleptin (Myalept; Bristol-Myers Squibb), was approved by the US Food and Drug Administration for treating lipodystrophy in humans. Metreleptin consists of 146 amino acids corresponding to the mature human leptin (UniProt identification P41159) with an additional N-terminal methionine residue (10). Due to its low molecular weight of 16 kDa, it is rapidly removed from circulation by renal filtration (11). Therefore, both in replacement therapy and in animal experiments, leptin has to be administered daily to achieve satisfactory pharmacodynamic effects (12, 13).

A general approach to enhance the bioactivity of small proteins is to extend their plasma half-life by increasing the hydrodynamic volume beyond the pore size (4 nm) of the renal filtration barrier, thus resulting in a sustained blood concentration within the therapeutic window over a longer period of time (14, 15). Here we have applied PASylation, a biological alternative to polyethylene glycol conjugation (PEGylation), which involves the genetic fusion of leptin with a conformationally disordered amino acid polymer of 600 residues comprising proline, alanine and serine (PAS) (16). This PAS(600)-leptin has a much larger globular diameter (18 nm) compared with unmodified leptin (4 nm) but still possesses high receptor binding activity.

Preceding studies in lean C57BL/6J mice revealed a much extended plasma half-life, from less than 30 minutes for the unmodified murine hormone to 20 hours for PAS(600)-leptin (17). In addition, after a single injection of PAS(600)-leptin, a reduction in food intake by up to 60% as well as a loss in body weight of approximately 10%, lasting for more than 5 days, was observed, whereas unmodified leptin was poorly effective for just 1 day. Administration of a PASylated superactive mouse leptin antagonist (PAS(600)-SMLA) led to the opposite physiological effects, demonstrating that the leptin or antagonist moieties, and not the PAS domain, were responsible for the measured influence on energy balance.

Apart from its clinical relevance, PASylated leptin appears promising to refine animal experiments when studying its physiological role by enabling less frequent injections or avoiding pump implantation and thus reducing stress while maintaining a strong pharmacodynamic effect. Lep^ob/ob mice reveal typical characteristics of the human disease condition, making this mouse line an appropriate model to study the metabolic consequences of leptin deficiency as well as leptin treatment in vivo (18, 19). Leptin is known to induce satiety and to stimulate the metabolic rate in Lep^ob/ob mice (12, 20–22). The latter effect is based on the observation that weight loss induced by leptin in Lep^ob/ob mice is larger than in pair-fed mice of the same genotype (23). Further support is delivered by indirect calorimetry, in which leptin was found to stimulate oxygen consumption in Lep^ob/ob mice (12, 20–22). The validity of these findings is limited due to short periods of monitoring and, in fact, has been criticized with regard to pitfalls in data analysis (24, 25).

Here we have exploited the excellent pharmacological properties of PAS(600)-leptin to characterize the relative contribution of anorexigenic and thermogenic effects on body weight loss in Lep^ob/ob mice. Furthermore, we applied unmodified leptin and PASylated leptin side by side in a repeated injection regimen to compare their potencies of correcting obesity and associated disturbances like glucose intolerance and hepatic steatosis in Lep^ob/ob mice.

Research Design and Methods

Preparation of recombinant murine leptin and PAS(600)-leptin

Unmodified and PASylated mouse leptins were produced as previously described (17). The PAS#1 polypeptide (for details see reference 16) with a length of 600 amino acids, starting with a His₆-tag, was genetically fused to the N terminus of mature murine leptin (UniProt identification P41160), which was secreted into the periplasm of Escherichia coli. Both proteins were purified by immobilized metal ion affinity (Ni²⁺ charged HisTrap HP; GE Healthcare) and ion-exchange (Resource Q; GE Healthcare) chromatography, followed by a final size exclusion chromatography (Superdex 200 or 75 column; GE Healthcare) with PBS (4 mM KH₂PO₄; 16 mM Na₂HPO₄; 115 mM NaCl, pH 7.4) as running buffer. Protein concentration was determined via absorption measurement at 280 nm. Typical endotoxin content was below 10 EU/mg protein as quantified by an Endosafe-PTS assay (Charles River).

Animals

Experiments were conducted with permission from the District Government of Upper Bavaria, Germany. Lep^ob/ob and Lep^+/+ mice were from our B6.Cg-Lep^ob/J breeding colony and were housed in a specific pathogen-free animal facility at 22°C temperature and relative humidity of 55% under 12-hours light, 12-hours dark controlled conditions. Either males or females were used in the two experiments to reduce the total number of animals needed. Mice were fed with standard chow diet (number V1124; Ssniff). During the experiments, mice were singly housed in type II long cages (Green line; Tecniplast).

Experimental design

Experiment 1 was devised to investigate the effect of three different PAS(600)-leptin doses after a single sc injection on body weight, body composition, body temperature, food intake, and energy expenditure over 18 days. Eight-week-old female Lep^ob/ob mice with an initial weight of 37.0 ± 5.4 g received injections of either PBS or 30, 100, and 300 pmol/g of PAS(600)-leptin in a volume of approximately 150 μL. During the following days, the PBS and 30- and 300-pmol/g PAS(600)-leptin groups were placed in an indirect calorimetry system to measure energy expenditure. During indirect calorimetry, the 100- and 300-pmol/g PAS(600)-leptin treatment led to a 70% reduction of food intake relative to the PBS injected mice. Accordingly, to induce weight loss by dieting alone, an additional group of PBS injected Lep^ob/ob mice was food restricted by 70%. Food amounts provided were reduced stepwise from day to day to parallel daily food intake observed in the 100- and 300-pmol/g PAS(600)-leptin groups.

In experiment 2, the effects of repeated sc injections of PBS, unmodified leptin, or PAS(600)-leptin were investigated over 20 days. Twelve-week-old male Lep^ob/ob mice with an average weight of 49.0 ± 3.6 g received four injections of PBS, 100 pmol/g leptin, or PAS(600)-leptin every 5–6 days (100 pmol equals 1.7 μg leptin and 6.7 μg PASylated leptin, respectively). Body weight, body composition, rectal body temperature, and food intake were repeatedly assessed. Age-matched (untreated) Lep^+/+ mice were used to determine reference values for all metabolic parameters. On day 20, that is 2 days after the last injection, nonfasted mice were dissected in the morning for subsequent analysis of uncoupling protein 1 (UCP1) protein abundance in adipose tissue, hypothalamic gene expression, and liver histology as well as measuring insulin, alanine aminotransferase (ALT), and antidrug antibodies in blood plasma.

Body temperature and body composition

Body temperature was monitored between 12:30 and 1:30 pm with a rectal probe using an Almemo 2490 thermometer (Ahlborn). Body composition was assessed in a minispec LF50H TD-nuclear magnetic resonance analyzer (Bruker Biospin).

Indirect calorimetry

In experiment 1, an open flow respirometry system (TSE Systems) was used to monitor energy expenditure. Air was drawn from the cages (flow 0.8 L/min), dried in a cold trap, and analyzed for O₂ and CO₂ concentrations in 9-minute intervals. Respiratory exchange ratio (RER) and heat production (HP) were calculated as described previously (26). Average daily metabolic rate (ADMR) represents 24 hours average HP for each day of the experiment. Resting metabolic rate (RMR) for each day was assessed by calculating the mean of the four lowest consecutive HP values. For data evaluation, ADMR and RMR were adjusted to the covariate body weight (27). Therefore, ADMR, RMR, and body weight data before the administration of PASylated leptin (d 0) were used as baseline values for a regression analysis (ADMR: y = 6.32x + 361.4, r² = 0.4394, P = .0367; RMR: y = 7.66x + 150.7, r² = 0.5312, P = .0047). ADMR and RMR values from the following days were analyzed in relation to each linear baseline increase.

Oral glucose tolerance test and plasma insulin quantification

In experiment 2, glucose tolerance in 6-hour-fasted mice was determined on day 19 after oral gavage of glucose. To take into account that the specific metabolic activity of fat is only 20% of lean tissue in mice (27), glucose load was calculated as follows: 2.8 mg glucose × (lean mass + 0.2 × fat mass). Blood was sampled from the tail tip by a small incision. Glucose was quantified with a glucometer (Freestyle Lite; Abbott). Plasma insulin levels were analyzed with a mouse insulin ELISA (Mercodia).

Liver phenotyping

The grade of hepatic steatosis was estimated under a light microscope by two blinded authors using a histoscore ranging from 0 (no steatosis) to 4 (severe steatosis). Representative hematoxylin/eosin-stained sections were selected and photographed. Alanine aminotransferase in plasma was measured with the Piccolo lipid panel plus reagent disk (Abaxis).

UCP1 Western blotting

Inguinal white and interscapular brown adipose tissue depots were homogenized in radioimmunoprecipitation assay buffer supplemented with 1% (wt/vol) protease and 1% (wt/vol) phosphatase inhibitor cocktail (Sigma-Aldrich). After centrifugation, 30 μg of protein per sample from the supernatant was separated by 12.5% SDS-PAGE and blotted onto a nitrocellulose membrane. UCP1 was detected using a polyclonal rabbit antibody raised against hamster UCP1 (28) (Table 2); β-Actin served as loading control and was detected with a commercially available antibody (#MAB1501; Millipore). Bands were visualized with infrared dye-conjugated secondary antibodies (#926–32211/12/22; LI-COR Biosciences) using the Odyssey CLx infrared imaging device (LI-COR Biosciences) (Table 2).

Table 2

Antibody Table

Peptide/Protein Target	Antigen Sequence (if Known)	Name of Antibody	Manufacturer, Catalog Number, and/or Name of Individual Providing the Antibody	Species Raised (Monoclonal or Polyclonal)	Dilution Used
Ucp1	Hamster UCP1	Anti-Ucp1 antibody	Self-made	Rabbit, polyclonal	10 000
β-Actin	Amino acids 50–70	Anti-actin antibody	#MAB1501; Millipore	Monoclonal	5000
Mouse antibodies		Anti-mouse polyvalent Ig/AP conjugate	A0162; Sigma-Aldrich	Goat, polyclonal	1000

Hypothalamic gene expression

As previously described, hypothalami were dissected with a brain blocker (Plastics One) to ensure standardized tissue sampling (29). RNA was extracted with Trisure reagent (Bioline) and was reverse transcribed with the Quantitect cDNA synthesis kit (QIAGEN). Quantitative real-time PCR was performed in technical triplicates using a SYBR Green supplemented SensiMix (Bioline) in a LightCycler (Roche). The amplification efficiency was calculated based on standard curves from 2ⁿ dilution series (Realplex Cycler software; Eppendorf) and used to determine the starting quantity of the cDNA. Expression values of proopiomelanocortin (Pomc), agouti-related protein (Agrp), neuropeptide Y (Npy), and suppressor of cytokine signaling 3 (Socs3) were normalized to the housekeeper β-Actin (Actb) and expressed in percentages with the values for Lep^+/+ mice set to 100%.

Primer pairs were as follows: Actb forward 5′-caccacaccttctacaatga-3′, Actb reverse 5′-gtacgaccagaggcatacag-3′; Pomc forward 5′-ccctcctgcttcagacctc-3′, Pomc reverse 5′-cgttgccaggaaacacgg-3′; Npy forward 5′-ggcaagagatccagccctg-3′, Npy reverse 5′-ccagcctagtggtggcatgc-3′; Agrp forward 5′-tcccagagttcccaggtctaagtc-3′, Agrp reverse 5′-gcggttctgtggatctagcacctc-3′; and Socs3 forward 5′-agcccctttgtagacttcac-3′, Socs3 reverse 5′-gaaacttgctgtgggtgac-3′.

Immunogenicity ELISA

Plasma samples of mice from experiment 2 were analyzed by an ELISA for the presence of antibodies directed against leptin or the PAS#1 polypeptide. Briefly, a MaxiSorp microtiter plate (NUNC) was coated overnight with 50 μL of (1) unmodified recombinant leptin, (2) PAS(600)-leptin, (3) recombinant IL1-ra (30), (4) PAS(600)-IL1-ra, or (5) the recombinant Fab fragment of the humanized anti-HER2 antibody 4D5 (16), each at a concentration of 25 μg/mL in PBS. After washing with PBS containing 0.1% (vol/vol) Tween 20 (PBS/T), 50 μL of a 1:1000 dilution with PBS/T of each serum sample was applied for 1 hour. After washing, bound murine antibodies were detected with 50 μL of a 1:1000 diluted antimouse polyvalent Ig/alkaline phosphatase conjugate (A0162; Sigma-Aldrich) (Table 2). After washing twice with PBS/T and twice with PBS, the enzymatic activity was detected using p-nitrophenyl phosphate as chromogenic substrate. After 15 minutes at 25°C, the absorbance at 405 nm was measured.

Statistics

If not stated otherwise, data are presented as mean with SD. When parameters were assessed over time, a two-way repeated measures ANOVA was conducted with dose/treatment and time as independent variables. Statistical significance in single measurements was determined by one-way ANOVA and Tukey's post hoc analysis, which compares all mean values to test for statistically significant differences between the experimental groups. Outcomes from this multiple comparison test are presented in lowercase letters with nonmatching letters indicating significant differences between groups. Experimental groups assigned with identical letters are not significantly different.

Due to high variation, Western blotting data from inguinal white adipose tissue were analyzed by Grubbs' outlier test prior to statistical evaluation.

Results

Single injection of PAS(600)-leptin dose-dependently ameliorates obesity in leptin-deficient Lep^ob/ob mice

We investigated the effect of different PAS(600)-leptin doses in Lep^ob/ob mice after a single sc injection. Mice receiving the highest PAS(600)-leptin dose (300 pmol/g) exhibited a maximal body weight reduction of 21% at day 6 after injection (pi). The weight difference relative to the PBS group lasted until the end of the measurement (d 18). Mice treated with the low (30 pmol/g) or intermediate dose (100 pmol/g) showed weaker and less sustained weight loss. The 30-pmol/g and the 300-pmol/g group, respectively, returned to control levels 7 days and 12 days pi (Figure 1A. Notably, weight decline in all groups was mainly due to a loss in body fat (Figure 1, B and C).

Figure 1

Effect of different PAS(600)-leptin doses in female Lep^ob/ob mice with an average initial body weight of 37.0 ± 5.4 g (n = 4–5, SD) from experiment 1. The dotted vertical lines indicate sc injection of PBS or PAS(600)-leptin. Body mass (A), fat mass (B), lean mass change (C), food intake (D), cumulative energy intake during days 1–4 (E), and rectal body temperature (F) were assessed daily. Statistical analysis in panels A, B, C, D, and F was conducted by a two-way, repeated-measures ANOVA. P values indicate overall effects of the factors dose and time. Statistical significance in panel E was determined by one-way ANOVA and Tukey's multiple comparison tests. Nonmatching lowercase letters indicate a statistically significant difference among groups (P < .05).

PAS(600)-leptin induces negative energy balance by promoting satiety and preventing hypometabolic torpor bouts

Food intake of Lep^ob/ob mice was significantly lowered after single sc PAS(600)-leptin administration (Figure 1, D and E). The duration as well as the magnitude of this effect varied in a dose-dependent manner with trajectories similar to the body weight reduction. The 30-pmol/g group returned to the food intake levels of PBS treated mice levels 4 days pi, whereas the 100-pmol/g and 300-pmol/g groups approached control levels later, at days 7 and 9, respectively.

Relative to the PBS group, mice treated with 100 or 300 pmol/g PAS(600)-leptin consumed approximately 300 kJ less energy (−70%) during the first 4 days after injection (Figure 1E. Additionally, rectal body temperature was elevated upon treatment with PASylated leptin (Figure 1F.

To determine the impact of PASylated leptin on energy expenditure and metabolic fuel selection, we compared Lep^ob/ob mice injected with PBS or 30 or 300 pmol/g PAS(600)-leptin with mice subjected to 70% food restriction using indirect calorimetry. To this end, the food-restricted group received 132 kJ per 4 days, which is 70% less compared with the 439 kJ consumed by the PBS group during the same time. Although food intake in the food-restricted group was adjusted to a level comparable with the mean anorexigenic effect of PAS(600)-leptin at doses of 100 and 300 pmol/g (Figure 1E, food restriction did not decrease body weight to the same extent. Body weight loss on day 4 was −4.4 ± 0.5 g vs −5.8 ± 1.0 g (P < .05) in food-restricted compared with the 300-pmol/g PAS(600)-leptin-dosed group, respectively.

Notably, food-restricted mice exhibited recurring hypometabolic torpor bouts (Figure 2A. Thus, torpid mice expended in total 80 kJ less energy during 4 days of food restriction than the three treated groups (Figure 2B. During torpor bouts, body temperature dropped to 25.7°C ± 0.2°C. In contrast to the pronounced hypometabolism observed in the food-restricted mice, no reduction in cumulative metabolic rate occurred when mice were treated with PASylated leptin. Only a minor trend toward elevated cumulative metabolic rate after PAS(600)-leptin injection was observed (Figure 2B.

Figure 2

Effects of different PAS(600)-leptin doses in female Lep^ob/ob mice on metabolic rate and RER (n = 4–5, SD) observed in experiment 1. The dotted vertical lines indicate sc injection of PBS, PAS(600)-leptin, or start of food restriction. A, Heat production. Black/white boxes represent dark/light cycle. B, Cumulative metabolic rate during days 1–4. C, Body weight correlated with RMR. Crosses represent mice prior to injection at day 0. Circles and squares represent mice 2 days after injection. Arrows connect data points from individual mice prior and after injection. D, RMR adjusted to the covariate body weight. E, ADMR adjusted to body weight. F, RER. Statistical analysis in panels A, C, D, E, and F was conducted by a two-way, repeated-measures ANOVA. P values indicate overall effects of the factors dose and time. Statistical significance in panel B was determined by a one-way ANOVA and Tukey's multiple comparison test. Nonmatching lowercase letters indicate a statistically significant difference among groups (P < .001).

To further analyze the effect of PAS(600)-leptin treatment on metabolic rate, differences in body mass were taken into account. The RMR of mice at the pretreatment day showed a positive correlation with body mass, which could be approximated by linear regression. Based on this reference relation, we evaluated repeated measurements of RMR and body mass in response to PASylated leptin treatment. For example, RMR recorded in mice 2 days after injection of 300 pmol/g PAS(600)-leptin was clearly elevated above the pretreatment values (see arrows in Figure 2C, resulting in positive residuals, whereas the PBS group did not show a systematic RMR shift (Figure 2C.

For quantitative evaluation, residuals were calculated based on the reference regression line for all repeated measurements to obtain an adjusted RMR for the entire treatment period. This analysis confirmed that PAS(600)-leptin significantly elevated the adjusted RMR in a dose-dependent manner, with the most sustained increase of approximately 130 mW on days 2–4 (+30%) observed for the highest dose (Figure 2D. In line with the result for cumulative metabolic rate (Figure 2B, only a slight, nonstatistically significant elevation of adjusted ADMR in both PAS(600)-leptin treated groups above PBS controls was observed (Figure 2E. Finally, we evaluated the effect of treatment on metabolic fuel partitioning. Indeed, PAS(600)-leptin depressed RER in a dose-dependent relationship, demonstrating a metabolic switch from carbohydrate toward lipid oxidation. Mice treated with 30 pmol/g PAS(600)-leptin resumed the RER level of the PBS group at the last day of calorimetry (d 4), whereas the 300-pmol/g group maintained the low RER and did not return to control values during the course of the measurement (Figure 2F.

Four repeated injections of PAS(600)-leptin distributed over 20 days rectify food intake and obesity

We compared the efficacy between short-acting unmodified (recombinant) leptin and its long-acting version, PAS(600)-leptin, in male Lep^ob/ob mice in a chronic treatment regimen (100 pmol/g per injection). Mice received four repeated sc injections at 5- to 6-day intervals. Remarkably, PASylated leptin induced a final weight loss after 20 days of 43% (body weight (d 0): 48.3 ± 2.5 g vs body weight (d 20): 27.5 ± 1.5 g, P < .001) (Table 1 and Figure 3A. Body weight loss was solely due to a reduction of body fat (Figure 3B. Although the body weight attained in PAS(600)-leptin injected Lep^ob/ob mice was comparable with the one found for healthy Lep^+/+ mice (Figure 3A, they still showed somewhat higher fat and reduced lean mass (Table 1 and Figure 3, B and C).

Table 1

Body Weight Change, Organ Weights, Hypothalamic mRNA, and UCP1 Protein Levels of Male Lep^ob/ob Mice (n = 4–5, SD) from Experiment 2

Parameter (SD)	Lepob/ob			Lep+/+
	PBS	Leptin	PAS-Leptin
Body mass change (d 0–20), g	4.45 ± 0.62a	2.68 ± 0.53a	−20.8 ± 2.73b	—
Organ weights
Interscapular BAT, g	0.27 ± 0.03a	0.22 ± 0.03b	0.10 ± 0.01c	0.06 ± 0.01d
Right inguinal WAT, g	1.58 ± 0.34a	1.60 ± 0.14a	0.57 ± 0.13b	0.11 ± 0.02c
Right epididymal WAT, g	1.71 ± 0.41a	1.98 ± 0.28a	0.74 ± 0.05b	0.14 ± 0.01c
Right quadriceps, g	0.05 ± 0.01a	0.04 ± 0.01a	0.06 ± 0.01a	0.11 ± 0.03b
Intestine, cm	53.0 ± 1.50a	52.4 ± 4.00a	43.4 ± 1.70b	43.1 ± 1.60b
Liver, g	4.27 ± 0.60a	3.54 ± 0.40a	1.43 ± 0.14b	1.37 ± 0.15b
Right kidney, g	0.20 ± 0.02	0.19 ± 0.03	0.17 ± 0.01	0.18 ± 0.01
Heart, g	0.18 ± 0.03	0.17 ± 0.02	0.13 ± 0.01	0.15 ± 0.02
Hypothalamic mRNA levels, %
Agouti-related protein (Agrp)	672 ± 121a	630 ± 299a	252 ± 37b	100 ± 17b
Neuropeptide Y (Npy)	258 ± 33a	262 ± 132a	158 ± 20a,b	100 ± 17b
Proopiomelanocortin (Pomc)	50 ± 8a	66 ± 39a,b	116 ± 19b	100 ± 22a,b
Suppressor of cytokine signaling 3 (Socs3)	86 ± 9	101 ± 28	104 ± 23	100 ± 13
UCP1 proteins levels, %
Interscapular BAT	46 ± 15b	56 ± 12b,c	86 ± 10a,c	100 ± 34a
Inguinal WAT	3.7 ± 8.0	n.d.	105 ± 67	100 ± 98

Abbreviations: BAT, brown adipose tissue; WAT, white adipose tissue; n.d., not detectable. Mice were killed on experimental day 20 after four repeated injections. Statistical significance was determined by a one-way ANOVA and Turkey's multiple comparison test. Nonmatching lowercase letters indicate a statistically significant difference among groups (P < .05). UCP1 levels in inguinal white adipose tissue showed an unusual variation: Grubbs' outlier test detected one data point belonging to the PAS(600)-leptin group with a relative UCP1 expression of 831%. This value was excluded from this table and the statistical analysis. (For Western blot image, see Supplemental Figure 1.)

Figure 3

Effects of four repeated sc injections of PBS, leptin, or PAS(600)-leptin on body mass, fat and lean mass, food intake, and rectal body temperature in male Lep^ob/ob mice vs age-matched lean Lep^+/+ mice (n = 4–5, SD) measured in experiment 2. The dotted vertical lines indicate the time points of injection. Body mass (A), fat (B) and lean mass (C), food intake (D), and rectal body temperature (E) were assessed daily. F, Dosing of leptin and PAS(600)-leptin was normalized to body mass for each injection. Inset in panel C shows representative Lep^ob/ob mice at day 20 treated with either PBS or PAS(600)-leptin. +/+, Reference values of age-matched lean Lep^+/+ mice. Statistical analysis was conducted by a two-way, repeated-measures ANOVA. P values indicate overall effects of the factors treatment and time.

In the Lep^ob/ob mice, the repeated administration of PAS(600)-leptin induced satiety and normalized hypothermia (Figure 3, D and E). The intestine was shortened after PAS(600)-leptin administration and approached the length in Lep^+/+ mice at the end of treatment (Table 1), possibly as a consequence of reduced food intake (31). It should be mentioned that the injected dose was adjusted to the body weight at the day of injection, thus resulting in a decrease of the applied amount of PAS(600)-leptin per mouse toward the end of the treatment period (Figure 3F. This did not alter the efficacy of the treatment as judged from the similar body weight decline after each injection.

Repeated treatment with PAS(600)-leptin normalizes glucose homeostasis, liver steatosis, and UCP1 expression in brown adipose tissue

On day 19, oral glucose tolerance was assessed after glucose gavage (Figure 4, A and B). Interestingly, PAS(600)-leptin improved glucose clearance in Lep^ob/ob even beyond the capacity that was measured in normal Lep^+/+ mice; in fact, 30 minutes after glucose gavage, blood glucose levels of PAS(600)-leptin treated Lep^ob/ob mice were lower than the glucose values of lean Lep^+/+ mice. Furthermore, plasma insulin concentrations were fully normalized after PAS(600)-leptin administration at the end of treatment (Figure 4C.

Figure 4

Effect of repeated sc injections of PBS, leptin, or PAS(600)-leptin on glucose homeostasis in male Lep^ob/ob mice vs age-matched lean Lep^+/+ mice (n = 4–5, SD) observed in experiment 2. A, Oral glucose tolerance was assessed 19 days after the first injection. B, Corresponding areas under the curves. C, Plasma insulin levels measured 20 days after the first injection. Statistical significance in panels B and C was determined by one-way ANOVA and Tukey's multiple comparison tests. Nonmatching lowercase letters indicate a statistically significant difference among groups (P < .05).

Whereas Lep^ob/ob mice injected with PBS or unmodified leptin suffered from hepatic steatosis as indicated by enlarged livers and lipid droplets (Table 1 and Figure 5A, the administration of PAS(600)-leptin fully normalized liver weight. Accordingly, ALT, a blood plasma marker for hepatocellular injury, was restored after PAS(600)-leptin administration (Figure 5C. Moreover, a histological analysis of hematoxylin/eosin-stained liver tissue demonstrated a reduction of lipid droplet size and number by PAS(600)-leptin treatment (Figure 5, A and B).

Figure 5

Effects of repeated sc injections of PBS, leptin or PAS(600)-leptin on liver parameters in male Lep^ob/ob mice vs age-matched lean Lep^+/+ mice observed in experiment 2. A, On day 20 after the first injection, Lep^ob/ob mice were killed and liver morphology was assessed by hematoxylin/eosin staining. Representative tissue sections for each treatment group are shown. B, Histology score derived from hematoxylin/eosin-stained liver sections (n = 4, range). C, ALT level in plasma (n = 3–4, SD). Statistical significance in panel C was determined by one-way ANOVA and Tukey's multiple comparison tests. Nonmatching lowercase letters indicate a statistically significant difference among groups (P < .05).

Furthermore, Western blotting of interscapular brown adipose tissue samples revealed an increase in UCP1 in the PAS(600)-leptin-treated mice to values comparable with those measured for Lep^+/+ mice. UCP1 levels in inguinal white adipose tissue of mice treated with PAS(600)-leptin were elevated compared with leptin and PBS injected mice, as well, but showed strong variation and the difference did not reach statistical significance (Table 1 and Supplemental Figure 1).

Repeated PAS(600)-leptin injections normalize the expression of leptin target genes in the hypothalamus

Looking at the expression level of genes regulated by leptin signaling, leptin deficiency resulted in an imbalance between the expression of anorexigenic Pomc and orexigenic Agrp/Npy transcripts in the hypothalamus. In line with the metabolic improvements, PASylated leptin lowered the mRNA expression levels of Agrp and Npy but elevated Pomc abundance, whereas Socs3 expression was unaffected (Table 1).

PAS(600)-leptin injections do not elicit the formation of antidrug antibodies

Generally, antidrug antibodies (ADAs) can neutralize the pharmacodynamic activity of biopharmaceuticals (32). Therefore, blood plasma samples were finally tested for the presence of ADAs after the end of the study (Figure 6). Notably, in none of the treated mouse groups antibodies against leptin itself or the PAS moiety (either attached to leptin or to IL1-Ra, as an unrelated control protein) were detectable.

Figure 6

Immunogenicity analysis at the end of experiment 2. Plasma samples taken on day 20 from mice treated with PBS, unmodified leptin, or PAS(600)-leptin were analyzed in an ELISA for the presence of antidrug antibodies using a series of reference proteins adsorbed to the microtiter plate. Bound antibodies were detected with an anti-mouse polyvalent Ig/alkaline phosphatase conjugate (n = 4, SD). The recombinant 4D5 Fab fragment served as a positive control reference protein due to some cross-reactivity with the secondary antibody.

Discussion

Due to its inherently short plasma half-life, daily injections of leptin are mandatory to induce pharmacodynamic effects in both experimental and clinical settings. Recently, we developed a novel long-acting leptin fusion protein with enhanced in vivo efficacy using PASylation technology (16, 17).

The first objective of the present study was to determine the contributions of anorexigenic and thermogenic effects to weight loss induced by treatment with PASylated leptin. PAS(600)-leptin injections caused strong reduction of adiposity in Lep^ob/ob mice. Body weight loss is necessarily the consequence of an energy deficit, which may develop either via lower energy intake, reduced assimilation efficiency, increased energy expenditure, or a combination of these three effectors, and is furthermore promoted by a shift of metabolic energy partitioning toward lipid oxidation.

Previous studies already demonstrated that leptin replacement alters energy intake and expenditure in Lep^ob/ob mice (12, 20–22). The respective contributions, however, of the anorexigenic and thermogenic effects to yield a negative energy balance cannot be ascertained from these studies due to the lack of continuous recordings of metabolic rate during sustained treatment (21, 22) and/or inappropriate data handling (24, 25). For instance, dividing energy expenditure by body weight leads to erroneous conclusions on energy balance regulation during leptin treatment (12, 20).

In the present study, continuous monitoring of energy intake and energy expenditure as well as adequate data analysis demonstrated that the anorexigenic effect of PASylated leptin is the predominant cause for negative energy balance. Lower energy assimilation efficiency in response to PAS(600)-leptin may have contributed to weight loss, as well, but was not assessed in our study. During the first days after sc injection, mice treated with 100 or 300 pmol/g PAS(600)-leptin had an energy deficit of around 300 kJ/(4 d) compared with the PBS group. In contrast, the thermogenic action was less obvious. The RMR was elevated, but this resulted only in a marginal elevation of ADMR. Notably, it has been pointed out before that proper evaluation of energy expenditure and its role in weight loss must take into consideration the occurrence of adaptive reductions in energy expenditure (33, 34).

In our study, food restriction was clearly not as effective as the treatment with PASylated leptin because recurring torpor bouts depressed metabolic rate by approximately 80 kJ/(4 d), thus attenuating weight loss. Therefore, torpor in food-restricted mice lowered the energy deficit to 220 kJ/(4 d) compared with a value of 300 kJ/(4 d) for the PAS(600)-leptin-treated group. Torpor is a transient hypometabolic state, which enables the conservation of energy stores and thus decelerates weight loss. It has been suggested that reduced leptin levels represent a signal for torpor induction because leptin replacement can block photoperiod-dependent torpor in the Siberian hamster (35) as well as torpor in food-restricted C57BL/6J (36) and in Lep^ob/ob mice (37).

The stabilizing effect of PASylated leptin on the metabolic rate by preventing torpor bouts promotes weight loss by further intensifying the energy deficit. Therefore, we estimate that the satiety effect of PAS(600)-leptin contributes 75% to body weight reduction, whereas its thermogenic action accounts for 25% by suppressing torpor.

Pertaining to energy partitioning in the body, PAS(600)-leptin mediated a shift in fuel selection toward lipid oxidation as indicated by a drop in the respiratory exchange ratio. Preferential lipid metabolism may be a secondary consequence of the reduced energy intake and/or a primary effect of PASylated leptin on lipid catabolism in adipose tissue, skeletal muscle, and liver (38–40).

Our second objective was to elucidate the efficacy of repeated injections of PASylated leptin to rectify the deranged metabolic phenotype of Lep^ob/ob mice. Remarkably, apart from adiposity, both hyperphagia and hypothermia as well as liver steatosis and glucose intolerance were successfully normalized, whereas no adverse effects were noticed during the intervention.

Neurons expressing the leptin receptor play a pivotal role in the control of glucose homeostasis and body weight (41, 42). Therefore, we quantified the expression of the leptin target genes Agrp, Npy, and Pomc. Indeed, the hypothalamic mRNA expression pattern was shifted toward an anorexigenic/catabolic profile, thus providing a molecular underpinning for the amelioration of the disease phenotype. In addition to the direct effects of leptin on glucose handling by stimulating hypothalamic neurons (42), body weight reduction as an indirect effect of PASylated leptin may also contribute to the normalization of glucose homeostasis. Direct and/or indirect action of PAS(600)-leptin may further explain the correction of liver steatosis. It has been demonstrated that moderate weight loss in obese subjects reduces intrahepatic lipid content and promotes hepatic insulin sensitivity (43, 44). In addition, leptin can acutely stimulate lipid oxidation in the liver (40).

Moreover, repeated PAS(600)-leptin treatment normalized expression of uncoupling protein 1, which is a crucial effector of energy expenditure and body temperature regulation by mediating nonshivering thermogenesis in brown adipocytes (45). Accordingly, brown adipocyte thermogenesis may contribute to the maintenance of metabolic rate observed in experiment 1 and to the normalization of hypothermia.

In line with published findings, we could show here that the development of a negative energy balance in response to leptin is independent of sex (46–48). Additionally, we know from our preceding study that in lean C57BL6/J mice no sex difference exists with regard to the plasma half-life of PAS(600)-leptin (17).

For the longitudinal investigation, we chose a dose of 100 pmol/g PAS(600)-leptin to execute robust metabolic effects. But as shown in experiment 1, even lower doses like 30 pmol/g or less should work successfully in such approaches. Compared with other chronic treatments of Lep^ob/ob mice, PASylation permitted substantial reduction in the total dose of administered substance as well as the number of injections: for comparison, Halaas et al (46) injected in total 325 nmol (5.2 mg) unmodified leptin per mouse in 20 injections over 20 days to induce a weight reduction of 23 g (−35%); Pelleymounter et al (12) applied in total 440 nmol (7 mg) per mouse in daily injections over 28 days to trigger a body weight loss of 20%. We administered only 16 nmol (1 mg, including the PAS tag) of PAS(600)-leptin per individual (four injections every 5–6 d) over 20 days to achieve a weight loss of 21 g (−43%). Therefore, lower amounts (∼25-fold) and largely reduced injection frequencies substantiate the benefit of PASylated leptin.

During experiment 2, we did not observe an attenuation of the pharmacodynamic effect, even though the total amount of PASylated leptin was gradually reduced toward the end of the treatment. Mice responded throughout the intervention, suggesting no occurrence of leptin resistance. In addition, hypothalamic Socs3, a known mediator of neuronal leptin resistance (49), was not elevated. In this context it is important to mention that many protein drugs elicit ADAs with variable consequences, ranging from loss of pharmacodynamic efficiency to severe allergic reactions (32). In principle, a modification such as PAS, although comprising a polymer of natural L-amino acids, represents a novel antigen to the immune system. Due to its natively unfolded conformation, the lack of charged or hydrophobic side chains and the absence of major histocompatibility complex anchor residues, PAS is a weak antigen at best, which was indicated here, as well as in previous studies (16, 50), by a lack of both ADAs against the active protein component and the PAS moiety. Thus, PASylation offers a beneficial alternative to PEGylation, which poses a well-known immunogenicity risk (51). Furthermore, in contrast to the metabolizable PAS polypeptide, PEG is poorly biodegradable; in fact, previous attempts to develop a PEGylated leptin showed strong vacuole formation (52), whereas similar effects were so far not seen for a PASylated protein (16).

In addition to congenital leptin deficiency, metabolic disorders like type 1 and type 2 diabetes, hypothalamic amenorrhea, and, in particular, lipodystrophy are proposed as targets for leptin therapy (8, 9, 53). On the other hand, use of leptin for the treatment of common obesity remained unmet because most obese subjects already show elevated leptin levels and suffer from so-called leptin resistance (54, 55). Recently, it was reported that the treatment of diet-induced obesity with leptin-sensitizing hormones like amylin or a glucagon-like peptide-1/glucagon chimeric peptide in combination with exogenous leptin restored leptin signaling and promoted weight loss (56, 57). Notably, in addition to its half-life-extending properties, the PAS polypeptide may serve as a linker to combine leptin with a sensitizing partner to create a long-acting and bispecific fusion protein, also offering simplified and cost-effective manufacturing as well as pharmaceutical development. Thus, PASylation technology not only permits the design of improved and versatile bioreagents to refine animal studies but also opens new perspectives for treating metabolic diseases.

Acknowledgments

We thank Klaus Wachinger (Lehrstuhl für Biologische Chemie, Technische Universität München) for technical assistance.

Disclosure Summary: A.S. is a managing director and A.S. and M.S. are shareholders of XL-protein GmbH. The other authors have nothing to disclose.

Abbreviations

 
• ADA

  antidrug antibody

 
• ADMR

  average daily metabolic rate

 
• ALT

  alanine aminotransferase

 
• HP

  heat production

 
• PAS

  proline, alanine and/or serine

 
• PBS/T

  PBS containing Tween 20

 
• PEG

  polyethylene glycol

 
• pi

  post injection

 
• RER

  respiratory exchange ratio

 
• RMR

  resting metabolic rate

 
• UCP1

  uncoupling protein 1.