There is accumulated evidence in humans and several animal models that genetic disruption of melanocortin signaling has a central role in the hypothalamic control of appetite and metabolism, and secondarily in body weight. Deleterious heterozygous RSVs downregulating this signaling pathway can cause dominantly inherited monogenic obesity with incomplete penetrance and variable expressivity, including cases of severe EOO with effects persisting into adulthood (4–20, 28). Alternatively, gain-of-function variants in this pathway can be protective for obesity (29). We have now found that the parental origin of predicted loss-of-function RSVs can explain at least part of the observed variable penetrance and expressivity of the obesity phenotype in children. A meta-analysis of reported cases has documented an increased proportion of maternally inherited heterozygous RSVs (1.66:1) in severe EOO, independently of the sex of the child. The proportion of maternally inherited RSVs is significantly increased in transmitting parents without obesity (1.5:1), but it is much higher (5.2:1) when the transmitting parent also had severe obesity. These data are in line with initial reports indicating that probands with pathogenic RSVs in MC4R, either mono or biallelic, were more likely than non-carriers to have a mother with obesity (30), even though maternal genotypes were not tested. The maternal transmission bias is significant for pathogenic and likely pathogenic RSVs in MC4R, as well as in other multiple genes of the leptin-melanocortin pathway related to EOO when analyzed together.
Considering that the identified RSVs are likely the main contributing factor to the obesity phenotype observed in the probands, the overrepresentation of these variants in mothers indicates that maternal inheritance is an additional risk factor for EOO. RSVs present in both, child and mother, may be acting synergistically to facilitate earlier penetrance and more severe phenotype in children. This parental sex inheritance bias in penetrance and expressivity of the EOO phenotype could have several causes, including decreased paternal reproductive fitness, imprinting effects, fetal and long-term effects of the maternal obesity during gestation, and different maternal perception and/or behavior during postnatal childcare and feeding.
Obesity is known to have a negative effect on reproductive fitness both in males and females. A dysregulated hypothalamic-pituitary-gonadal axis in men can associate low total testosterone and diminished luteinizing hormone pulse amplitude (31). In addition, hyperinsulinemia in males with obesity increases androgen aromatization to estrogen in adipose tissue, resulting in negative feedback in gonadotrophin secretion and gonadal dysregulation affecting spermatogenesis. However, the negative effect of obesity on reproductive fitness is possibly stronger in females, with variable contribution of ovarian, endometrial and hypothalamic–pituitary factors (31). Female obesity is associated with increased time to achieve pregnancy, and with greater risk of recurrent miscarriage (32–33). Therefore, although an effect cannot be fully excluded, decreased paternal versus maternal fitness due to subfertility and social issues related to the phenotype are unlikely the main causes of the observed maternal transmission bias. In addition, this transmission bias is also significant in the absence of parental obesity.
Imprinted genes that affect prenatal growth control and postnatal metabolism display parent-of-origin-specific expression, so the observed maternal over-transmission could also be related with genetic imprinting at some of the loci with pathogenic RSVs. In addition to some imprinting disorders at specific loci such as Prader Willi (15q11.2), Beckwith-Wiedemann (11p15.5) or pseudohypoparathyroidism (GNAS, 20q13.32) syndromes which have marked effects on growth and behavior, there is growing evidence that genomic imprinting plays a significant role in regulating BMI. In mice, two-thirds of the loci associated with obesity have additive effects with dominance and imprinting (34). MC4Rs activate the G protein Gsα encoded by the imprinted GNAS gene, and Gsα signaling is required for controlling energy balance, thermogenesis, and peripheral glucose metabolism. Individuals with heterozygous Gsα-inactivating variants only develop obesity when the GNAS variant is present on the maternal allele because of tissue-specific genomic imprinting (35). GNAS variants have been reported in 1% of children with severe obesity, with or without associated features of pseudohypoparathyroidism, being the variants de novo or maternally inherited and affecting MC4R signaling (36). However, we have excluded probands with genomic or imprinting defects at these loci (15q11.2, 11p15.5, 20q13.32) and RSVs affecting the GNAS gene from this study. Although subtle imprinting effects influencing growth regulation, metabolic programming or feeding behavior affecting in utero and/or postnatal weight gain trajectories cannot be discarded, none of the studied loci, including MC4R, are known to be imprinted.
Children with maternally inherited RSVs had significantly higher birthweight for gestational age than the population average and than children with non-maternally transmitted RSVs, being 43% of them significantly large for gestational age. This observation documents a prenatal onset of the overweight in most cases, which is in contrast with Mcr4 deficient mice, whose weights are not different at birth from their wild-type littermates (37). Maternal obesity during gestation is a known risk factor for increased human fetal growth and child obesity (38), partly due to placental dysfunction and aberrant DNA methylation affecting the adiponectin and leptin systems (39). Therefore, obesity in mothers carrying pathogenic RSVs would be a risk factor in all their pregnancies, but the outcome of these perturbations is also highly dependent upon the genetic background of the individual. Those with an obesity-prone genotype are thus more likely to be affected, with potential synergistic effect and severe EOO in children who inherited the RSV.
Other possible mechanism for EOO in children with maternally inherited RSVs may relate to the maternal attitudes and practices regarding child feeding during the first years of life. There is strong evidence for the influence of genetics in multiple human behaviors including eating and feeding practices (40). Shared genetic variants in family relatives may explain similar attitudes towards feeding. We have analyzed the diet, feeding behaviors, and overall familiar functioning using a specific questionnaire answered by mothers. Interestingly, maternal diet was not significantly different among the groups of mothers, in contrast with the reported binge eating in MC4R carriers (30). Emotional responsiveness associated with binge eating of mothers can affect their feeding behavior to children (41), but this relationship was not assessed since binge eating was not reported. Poor family functioning has also been described as a risk factor for unhealthy parental feeding practices in children with overweight or obesity (42). However, despite some minor divergences, no significant differences were found among groups. Thus, feeding practices of mothers carrying RSVs in the studied genes do not seem to be significantly influenced by dietary habits or family functioning.
Transmitting mothers had significantly lower scores in perceived responsibility for feeding, which would be indicative of a parental role less committed to control children eating habits. As parental involvement during early years is essential for children to develop their eating control mechanisms (43), carelessness in these aspects may facilitate a dysregulation of eating habits in children with maternally inherited variants, with subsequent risk for EOO. The main observed differences in the other subscales were likely related to the intrinsic characteristics of the groups analyzed. The perceived child weight and the concern about child’s weight correlated with the actual children’s BMI. The perceived parent weight revealed the differences expected as most mothers with obese children were affected by obesity while control mothers were not. Similarly, the pressure to eat showed an inverse relationship with lower scores in mothers with obese children. No clear associations were found with factors related with feeding control. Interestingly, despite no statistically significant differences were identified in these subscales, a clear trend was observed with transmitting mothers having the most extreme values on one side of the distribution in all subscales, indicating a relevant role of maternally inherited variants in addition to the extreme phenotype when explaining child feeding attitudes. Our data pointed to a negative association between child’s BMI and pressure to eat and a positive association between child’s BMI and concern about child weight, as previously reported (27).
The main limitations of this study are related to the relatively small sample size with complete parental and prenatal information, and questionnaire participation. Unfortunately, parental genotyping is missing in multiple genetic studies of childhood obesity. Based on the observed transmission bias, we focused exclusively on maternal responses, but the role of the father in influencing the obesity risk of children during early life must not be underestimated (44). Nevertheless, although no causal evidence was clearly identified, some mechanisms are suggested which deserve further analysis in additional and larger epidemiological and experimental studies.
In summary, we have described for the first time a significantly increased proportion of maternally inherited RSVs in genes of the leptin-melanocortin pathway associated with EOO. We propose that synergistic effects of functional RSVs present in both, transmitting mothers and their children, can contribute to EOO in the affected children, acting both pre- and postnatally. Because EOO is associated with an increased morbidity and mortality and because of its resistance to treatment, early diagnosis of the genetic risk factor and its inheritance pattern is required. Then, prevention should begin as soon as possible to minimize the risk for permanent, adverse alterations in neural pathways which control energy homeostasis. Management of at-risk pregnancies and monitoring of early life diet and feeding practices, along with physical activity should be implemented. Pharmacological enhancement of melanocortin signaling with available medications can also eventually have a role in these patients (45).