Although some of the aforementioned patients of Pakistani background have now reached adulthood, they have all started leptin replacement during childhood. Therefore, the effects of leptin deficiency in human adults who underwent their entire childhood and puberty without leptin were unknown. In 1998, the first leptin-deficient adults in the world were reported (18). Due to a severe obese phenotype (body mass index [BMI] of 55.8 kg m−2) and hypogonadotropic hypogonadism, a 22-year-old adult man (patient A) entering the Turkish mandatory military service was suspected as being leptin-deficient. Extremely low serum leptin levels (0.9 ng mL−1) were found not only in this patient, but also in another 34-year-old adult woman (patient B (1.6 ng mL−1) and in a 6-year-old girl (1.1 ng mL−1) from the same family, who were also morbidly obese and hypogonadal (the adult woman). Subsequently, a 30-year-old woman (patient C) and a 5-year-old boy (patient D) from the same pedigree were diagnosed as leptin-deficient (17,21) (Fig. 1). The leptin-deficient girl (patient E) died of sepsis at age 9, before the initiation of leptin replacement, possibly due to immune defects related to leptin deficiency. Our group has been studying these unique patients since 1998, and the effects of leptin replacement have been evaluated for 10 years. In this review, we summarize the findings from previous studies, which were all approved by the local Ethics Committees at various institutions. For all of the studies, the patients signed informed consents.
Treatment with recombinant methionyl human leptin was initiated at ages 5 (boy), 27 (adult man), 30 and 40 (women). Doses started at 0.02–0.04 mg kg−1 d−1 at 6 pm and were designed to achieve a normal leptin concentration based on a body fat of 30% in adult women and 20% in adult men. The child's dose was initially calculated to achieve a peak serum leptin concentration of 70 ng mL−1, which is equivalent to 10% of the child's predicted normal serum leptin concentration (based on age, sex and body composition) (24,25). Subsequently, doses were decreased as patients lost weight to avoid excessively rapid weight loss. We chose a daily subcutaneous injection in the evening in order to mimic leptin's circadian rhythm, of which the peak occurs at night (26).
During this prolonged follow-up, minimum dose adjustments were made, based on clinical grounds (mainly weight gain or excessive loss). The adults' initial mean dose was 4.1 ± 1.2 mg d−1 (2.8 mg for the man, 4.2 mg for the younger woman and 5.3 mg for the older woman), and the current dose is 1.4 ± 1.9 mg d−1 (0.3 mg for the man, 0.45 mg for the younger woman and 3.6 mg for the older woman). The child's initial dose was 1.36 mg d−1, and the current dose is 0.95 mg. Significant dose decreases were possible for the man and for the younger woman. However, the older woman is on a considerably higher dose on leptin. Besides being the oldest patient, she is the most obese and was diagnosed with type 2 diabetes before treatment. Therefore, it is possible that this patient also presents common obesity as an underlying disease and therefore leptin resistance. The child is on a higher dose, as compared to the man and the younger woman. Decreases in dose led to weight gain and to growth deceleration (from the 50th to the 10th percentile of the growth curve, over 2 years). As dose was increased, the child is now between the 10th and the 25th percentile, and will likely achieve his targeted stature (Fig. 2).
Body composition, food intake and energy expenditure
Phenotypically, all of the patients were morbidly obese. With treatment, the adults' mean BMI decreased from 51.2 ± 2.5 kg m−2 to 26.9 ± 2.1 kg m−2 after 18 months of treatment. At the eighth year of treatment, the mean BMI was 29.5 ± 3.4 kg m−2. The patients' BMIs most likely reflect those of commonly obese or overweight non-leptin deficient humans, and not incomplete dosing of recombinant methionyl human leptin. Patients A and B complained of excessive anorexia when higher doses were given, and Patient C is already taking the maximum physiological dose that determines normal leptin levels. Over the course of treatment, weight was fairly stable, except for clinically advised and closely monitored periods of leptin withdrawal in the adults. During those periods, leptin replacement was withdrawn for up to 6 weeks, making the patients gain from 6 to 10 kg. At the 10th year, after a temporary leptin withdrawal, mean BMI was 33.1 ± 5.3 kg m−2. Weight promptly returned to pre-withdrawal levels after leptin was reintroduced. The child's BMI decreased from 39.6 kg m−2 to 24.4 kg m−2.
The decrease in BMI occurred mainly due to fat mass loss, and at a smaller extent, to lean mass loss, as measured by dual energy X-ray absorptiometry (16). After 18 months of leptin replacement, the absolute fat mass decreased from 62.5 kg to 9.5 kg in Patient A, from 54.9 kg to 16.8 kg in Patient B and from 68.4 kg to 28.3 kg in Patient C. Relative to total body weight, total body fat decreased from 46.6% to 14.8% in Patient A, from 50.6% to 27.8% in Patient B and from 51.7% to 38.1% in Patient C. In addition, the total lean mass decreased from 71.6 kg to 54.8 kg in Patient A (24% of total weight loss), from 53.5 kg to 43.6 kg in Patient B (21% of total weight loss) and from 64.0 kg to 46.0 kg in Patient C (31% of total weight loss). Since the fat mass decreased at a larger extent, the lean mass relative to total body weight increased in all patients (from 53.4% to 85.3% in Patient A, from 49.3% to 72.1% in Patient B and from 48.3% to 61.9% in Patient C). At baseline and after 18 months, the women's measured fat masses were close to the predicted fat masses according to their total body weights. The man's fat mass was 33% above the predicted mass before treatment, and decreased to levels close to the predicted after 18 months (27).
In the first month of treatment food intake markedly decreased in all adult patients, from 2384 ± 946 kcal d−1 to 1179 ± 790 kcal d−1. Subsequently, caloric intake increased and stabilized at approximately 1907 ± 739 kcal d−1 (as measured 7 years after treatment). In the first 6 months, the decrease in caloric intake resulted in a temporary increase of fat intake and in a decrease in carbohydrate intake, both relatively to the total caloric intake. That was reversed after 12 months, with an increase in carbohydrate and a decrease in fat consumption (28). In the younger woman, leptin withdrawal for 2 weeks did not change the proportion of fat, carbohydrate and protein in the daily diet, but did decrease the intake of vitamin C, pyridoxine, pantothenic acid, folate, potassium, magnesium, copper, chromium, threonine, lysine, histidine, tryptophan, methionine and cystine (29). In the child, after 2 years, caloric intake also decreased from 2709 ± 370 to 2194 ± 292 kcal d−1, which is 106% of the recommended caloric intake for a boy this age, height and weight (17).
The observed alterations in food intake resulted from changes in the motivation to eat. After 15 weeks, leptin replacement reduced overall food consumption, rate of eating and duration of eating at each course of every meal. Leptin slowed the rate of eating and reduced the time spent eating each food item. In addition, patients reported markedly less hunger, less desire to eat and greater fullness, both before and after the meals. After receiving leptin, participants found the foods to be equally appealing in taste, texture, aroma and the filling qualities (30).
Activity levels, measured by actigraphy, also increased with leptin replacement (16). At baseline, the three adults had 24-h energy expenditure and 24-h fat oxidation similar to those of normal controls. After treatment, energy expenditure was still normal, and its decrease associated with weight loss was less pronounced in the leptin-deficient subjects than that of controls under a weight loss programme. Similarly, fat oxidation was also higher in the treated leptin-deficient patients. These results show that leptin does not increase absolute energy expenditure, but it does prevent the weight loss-induced reduction in metabolic rate (31).
Bone mineral density was normal in all patients, except for the adult man, who had low bone mineral density of the spine (bone mineral density of L2–L4, 0.924 g cm−2; T-score −1.96; Z-score −2.36) (21). These changes in bone mineral density are in contradiction to those that are observed in ob/ob mice, and are more likely to be explained by the hypogonadism, rather than leptin deficiency. At baseline, parathyroid hormone levels were above reference values for all women (21). Later on, we identified low levels of 25-hydroxyvitamin D in these patients. We assume that they had secondary hyperparathyroidism due to low sun exposure, but we cannot exclude a direct effect of the absence of leptin on the vitamin D metabolism.
Lipids and glucose metabolism
Before treatment, all patients were hyperinsulinemic and dyslipidemic (low high-density lipoprotein cholesterol and normal-high or high triglycerides). The older woman had hyperglycaemia and was diagnosed with type 2 diabetes, which onset had probably occurred before the diagnosis of leptin deficiency (16,21). Leptin replacement normalized serum lipids, insulin and glucose levels, leading to the resolution of type 2 diabetes in the oldest patient (16).
The male adult was submitted to meal tolerance tests before and 1 week, 18 months and 24 months after leptin replacement was initiated. During each test, blood was collected every 7 min during 24 h for measurements of glucose, insulin and C-peptide. Deconvolution of C-peptide, modified insulin kinetics models and minimal glucose models were employed to determine insulin secretion, hepatic extraction and sensitivity, respectively. After 1 week, leptin replacement increased insulin hepatic extraction by 2.4-fold (resulting in a 1.8-fold decrease of posthepatic insulin delivery), without significant changes in insulin secretion, and also increased insulin sensitivity by 10%. After 24 months, we showed that leptin replacement increases insulin by at least 10-fold, with additional decreases in insulin secretion and hepatic extraction (32).
Paradoxically, after leptin withdrawal, insulin sensitivity further increased, as measured by euglycaemic hyperinsulinemic clamps (33). However, we hypothesize that this increase in insulin sensitivity is attributed to the quick gain in glucose-absorbing fat mass following leptin withdrawal. This increase in insulin sensitivity when off leptin was clinically evident when the adult man developed severe hypoglycaemia during an oral glucose tolerance test.
During euglycaemic hyperinsulinemic clamps, we showed that the levels of nonesterified fatty acids are decreased by hyperinsulinemia due to insulin's stimulatory effects on lipogenesis and inhibitory actions on lipolysis. After leptin withdrawal, this decrease in nonesterified fatty acids was further enhanced, showing that leptin counteracts insulin's effects on lipogenesis and lipolysis (34).
Gonadal and reproductive function
Hypogonadotropic hypogonadism was observed in all of the adults, with normal gonadotropin responses to GnRH stimulation. The adult man had no signs of puberty – no beard, scanty hair in the pubic and axillary areas, bilateral gynecomastia, small penis, small testis size and azoospermia. His baseline measurements of testosterone and free testosterone were below the normal range for his age. The older woman had had spontaneous menarche at age 35, and had scarce hair in the pubic and axillary areas, no mammary tissue and small uterus and ovaries. The younger woman had spontaneous menarche at age 29, with normal hair in the pubic and axillary areas, diminished mammary tissue and small ovaries and borderline uterus. Both women had a luteal phase defect with low midluteal phase progesterone levels (21). The spontaneous onset of puberty shows that leptin is important (35), but not obligatory for the start of puberty in female humans.
Menstrual periods became regular after leptin replacement in both patients, with serial midluteal phase progesterone measurements >10 ng mL−1, which are indicative of ovulation. The adult man's testosterone and free testosterone levels reached normal values for adults, and the mean levels of luteinizing hormone (LH) and testosterone, measured over a 24-h period, also increased from 0.75 ± 0.04 mU mL−1 and 2.61 ± 0.06 ng mL−1 to 2.75 ± 0.07 mU mL−1 and 7.50 ± 0.07 ng mL−1, respectively. There were no changes in the number of pulses of LH and testosterone, but there was an increase in their amplitude. All patients fully developed secondary sexual characteristics (16).
Adrenal, sympathetic tone and biomarkers of cardiovascular disease
Although it has been shown that leptin concentrations are inversely related to adrenocorticotropic hormone (ACTH) and cortisol and modulate the levels of endogenous cortisol in healthy humans (26), hypercortisolemia was not observed in our patients. Levels of free urinary cortisol were normal, and were suppressed with 1 mg of dexamethasone (21). In the adult man, treatment increased the mean levels of serum cortisol measured over 24 h, from 4.04 ± 0.22 µg dL−1 to 5.97 ± 0.30 µg dL−1, with fewer pulses (25 vs. 19) of greater amplitude, greater morning rise and increased pattern regularity (16).
Low sympathetic tone was shown in the adults and in the girl, evaluated by cold pressor response tests, by orthostatic hypotension tests and by skin response tests. Sympathetic function was normal in heterozygous and wild-type subjects. When compared to 15 age- and sex-matched healthy subjects, systolic and diastolic cold pressor responses were significantly lower in the leptin-deficient patients (z = −4.7; P < 0.001 for both systolic and diastolic responses) (21).
The changes in biomarkers of inflammation, coagulation, fibrinolysis and platelet aggregation did not follow a pattern after leptin replacement. However, leptin withdrawal determined changes in most of these biomarkers towards a decreased state of thrombogenesis and increased fibrinolysis (36). This suggests that the absence of leptin may protect against cardiovascular diseases, even in a morbidly obese state.
Growth hormone (GH) responses to both insulin-induced hypoglycaemia and exercise tests were absent in the man and in the younger adult woman, even though heights were normal. This blunted response is likely to be attributed to obesity (21). All of the insulin-like growth factor (IGF)-related parameters were within the normal range, except the postprandial insulin-like growth factor binding protein (IGFBP1), which was below the lower limit of normal (16). After 2 and 18 months of leptin replacement therapy, significant increases in IGFBP1 levels at the 18th month, both while fasting (21.1 ± 1.3 at baseline, 30.8 ± 7.1 at the second month, 140 ± 42 at the 18th month; reference range 13–120 ng mL−1; P = 0.02) and postprandial (3.7 ± 0.5 at baseline, 8.6 ± 1.9 at the second month, 18.1 ± 2.6 at the 18th month; reference range 10–30 ng mL−1; P = 0.0001). We also observed significant increases in mean insulin-like growth factor binding protein (IGFBP2), from 528 ± 54 to 936 ± 53 ng mL−1 at the 18th month, P < 0.001 (16). These changes are possibly attributed to the decrease in insulin levels (37).
Thyroid function tests were normal for the adults. However, the girl had elevated TSH levels, with negative titres of antithyroid antibodies and normal free T4 and T3. Her response to thyrotropin-releasing hormone (TRH) was exaggerated, reflecting primary hypothyroidism (21). The 5-year-old boy had normal thyroid function tests (17), showing that the thyroid phenotype is heterogeneous in leptin deficiency (38). Changes in thyroid hormones have also been observed in the patients of Pakistani (15) and Austrian backgrounds (20), and the changes in thyroid function may be attributed to the lack of leptin. Leptin replacement did not increase free T4 or T3 (38), as previously observed in other studies (15). We have first demonstrated that leptin has a highly organized and distinct diurnal and circadian rhythm (26), which has a synchronous pattern with TSH (39). In healthy Caucasians, the rhythms of leptin and TSH were synchronized, with a nadir in late morning and a peak in the early morning hours. Heterozygous subjects showed a significant correlation between leptin and TSH rhythmicity, although weaker. Contrarily, the leptin-deficient adult man presented dysregulated patterns of TSH pulsatile and circadian rhythms.
Those data suggest that leptin has a role in regulating TSH secretion in humans, and its absence may lead to thyroid dysfunction. However, it is still unclear why the thyroid phenotype is heterogeneous, and why some patients have primary or secondary thyroid dysfunction.
All adult patients had high plasma C3C levels, and the girl had low C4 levels, as well as a low total T-cell count. The adult women had elevated IgG and IgA levels (21). Immunoglobulin levels were normal in the boy, except for elevated IgE, indicating atopy. He also had normal blood cells count, with decrease in the absolute lymphocyte count 6 weeks after leptin was initiated, from 3.5 × 103 µL−1 to 2.9 × 103 µL−1. By flow cytometry, we determined that the decrease in lymphocytes corresponded to decreases in CD3, CD4 and CD19 cells.
In the boy, lymphocyte proliferative response to mitogens phytohaemagglutinin, concanavalin A, pokeweed, tetanus and candida antigens was normal at baseline, except for tetanus. Two and six weeks after treatment, responses increased significantly for all substances, except for tetanus antigen (40). Therefore, leptin enhanced T-cell responsiveness, even in a normal setting.
Brain structure and function
Eighteen months of leptin replacement increased grey matter concentration in three brain regions of the adults: the anterior cingulate gyrus, parietal lobe and medial cerebellum (41), which are implicated in neural circuits regulating hunger and satiation. Unpublished results show that annual withholding of replacement for several weeks reverses this effect in the same brain structures, and treatment re-initiation partially restores grey matter.
In functional studies, leptin replacement reduced activation of regions linked to hunger (insula, parietal and temporal cortex) and enhanced activation of regions linked to inhibition and satiety (prefrontal cortex), as well as the cerebellum. These results show that leptin has extra-hypothalamic effects in the regulation of food intake (42).
Procognitive effects of leptin were demonstrated in the boy, who presented improvements of several subtests of the Differential Ability Scale, a measure of general verbal and non-verbal functioning, and of selected subtests from the neuropsychological assessment (NEPSY), a measure of neuropsychological functioning in children (17). At baseline, the adult patients did not reach scores compatible with anxiety or depression on the Hamilton scale, and no changes during treatment were observed. Nevertheless, we noted that the patients' behaviour changed from very docile and infantile to assertive and adult-like, within 2 weeks of the onset of leptin treatment, before weight loss occurred (16).