This cohort sample incorporates a rare combination of childhood measures, including adiposity, APOE genotype, blood lipid profile, and PA and CRF. Using this study, we have identified that APOE4 appears to be associated with reduced adiposity in children. Additionally, this effect is modified by CRF; in children of lower CRF, APOE4 seems to predispose to lower adiposity, but also to a more atherogenic serum lipid profile. These findings show the importance of examining APOE, adiposity and blood lipids in the context of related PA and fitness.
There is a well-established association between APOE
genotype and blood lipids in both adults and children  
. In the TIHS, as expected, APOE4
-carrying individuals had significantly higher LDL-C levels than those carrying other alleles. We also observed an association between APOE
genotype and adiposity. Individuals with an APOE4
-containing genotype had lower mean BMI, truncal fat mass and waist circumference than those with other genotypes. This suggests that the E4 allele may have a protective effect against increased adiposity. Little evidence for a conversely detrimental effect of E2 on adiposity was observed.
Insight into the possible mechanism(s) through which APOE
regulates fatness might be gleaned from consideration of the regulation of lipid flux by adipocytes. Adipocytes act as storage compartments for TGs. TGs are either produced endogenously by adipocytes from internalised free fatty acids that are the product of triglyceride rich lipoprotein (TGRL) breakdown in the circulation, or from internalised TGRLs themselves, which are broken down endogenously. In the circulation, ApoE carries TGRLs, such as VLDLs and chylomicrons, to adipose cells, and thus ApoE inefficiency has the capacity to reduce clearance of lipid from the circulation into adipose tissue. EKO mice demonstrate smaller adipocytes, and less adipose tissue 
. Recent studies have demonstrated that endogenous ApoE modulates the TG content of adipocytes, independently of exogenous ApoE 
. Further, when mice with either knocked-in human APOE3
, or APOE4
, were subjected to high fat diets for 8 weeks, E4 knock-ins were significantly more resistant to obesity than E3 knock-ins 
. Taken together, these data suggest that a possible mechanism through which E4 may protect against adiposity is via a reduction in the ability of adipocytes to accumulate TG.
PA was recently demonstrated to modify the effect of obesity related genes in adults 
. PA has also been shown to modify the effect of APOE
genotype on lipids, particularly LDL-C, in children and adults 
. We saw little evidence of such an interaction when considering both an objective, and a subjective measure of PA, although the ability to fully test this was limited by the fact that we collected objective PA data at only one time-point in a single day. However, a novel aspect of this study was our ability to also examine the effect of CRF on the child APOE
-adiposity association. This is of relevance because CRF appears more important in controlling cardiovascular disease risk factors than PA 
. Our data concurs with recent work demonstrating that the association between CRF and PA is generally weak 
, especially in children 
. It is likely that CRF is determined not only by a number of environmental factors that includes PA, but also by genetic factors 
. Thus there is a clear need to consider PA and CRF separately. We found no interaction between PA and APOE
genotype in determining BMI. However, we did find that APOE4
was strongly associated with lower BMI only in those of lower CRF, and had little effect on BMI in those of higher CRF. This suggests that the mode of action through which APOE4
exerts its apparent protective effect against adiposity may only be relevant in those who are less fit.
We also examined the relationship between blood lipids and BMI in children grouped by APOE variant and fitness. There was little association between lipids and BMI, except in the low-fit, non-E4 allele group, where BMI was associated with TG, LDL-C, and HDL/Total-C ratio. We suggest that in children of higher fitness, lipid may be more efficiently removed from circulation via, for example, muscle metabolism, and in those carrying APOE4, storage of lipid by adipocytes is reduced. Only in children who are of lower fitness and who lack APOE4 may be free to take up and store lipids at a rate that is mainly dependent on circulating lipid levels.
The relationship between APOE4
, circulating LDL-C, cardiovascular risk, and BMI warrants further scrutiny. It is well established that APOE4
predisposes to higher circulating LDL-C, along with higher cardiovascular risk 
. In the TIHS, we confirmed the association of E4 with higher LDL-C, however the study design precluded assessment of cardiovascular risk. The lower BMI seen in E4 carriers appears somewhat counterintuitive to higher cardiovascular risk that might be predicted to be present in these individuals. But the existence of a direct relationship between high BMI, as one of a web of metabolic syndrome factors, and increased cardiovascular risk, is unclear 
. Barriers to the deposition of lipid into adipocytes may result in lower BMI, but may also drive higher circulating levels of lipids such as LDL-C and thus a more detrimental atherogenic lipid profile and higher cardiovascular risk. Thus, it is possible that while APOE4
might confer protection from fatness, it may also confer higher risk of cardiovascular disease.
It is also pertinent to mention that, despite recent large scale genome-wide association analyses 
, association of APOE
genotype with human adiposity has not, to our knowledge, been reported previously. We suggest that this may be because genome-wide association studies have by-and-large focussed on adult obesity. We also suggest that the effects of APOE
on adiposity may not be strongly and replicably apparent without both specific consideration of the E4 allele, and the interaction of E4 with CRF. Additionally, the lack of consistency of association of APOE
genotype with various lipids such as TGs and HDL-C across numerous studies 
may be due to the interacting effect of BMI and CRF, the latter of which is rarely taken into consideration in such analyses. Our findings require independent confirmation in cohorts where data on both APOE
genotype and CRF is available. However, we suggest our data demonstrates the importance of the consideration of environmental factors in tandem with genetic factors, in order to continue to progress our understanding of the architecture of complex diseases and phenotypes 
It is interesting to speculate as to whether APOE
may be a ‘thrifty’ gene during times of food abundance. Thrifty genes are historically advantageous in times of famine, but are rendered detrimental by modern high fat diets and sedentary lifestyles 
. In Caucasians, as in most populations, the APOE
3 allele is the most frequent, and is in high frequency in the non-APOE4
-containing genotype carriers to which we compared our E4-containing carriers. Our data suggests that E3 (and also possibly E2) may provide more efficient fat deposition than E4, and may be the thrifty allele(s) that in the face of high fat diets and low fitness levels, now predispose to higher BMI. The modern APOE4
allele might protect against fat deposition, however it appears to also predispose to a more atherogenic lipid profile 
, thus leading to a higher risk of cardiovascular disease in APOE4
In conclusion, in 8 yr old children from the TIHS, APOE4 is associated with lower BMI. The relationship between APOE4 and lower BMI is potentiated by lower CRF. We propose that, in those of lower CRF, APOE4 compensates for the lack of the protective effect of fitness from increased adiposity. Based on animal model studies, the mechanism through which APOE4 may protect against increased adiposity may involve a decreased ability of E4-expressing adipocytes to take up, synthesise, and/or store TGs. At a future clinical level, we suggest that assessment of child CRF and APOE4 genotype may potentially define a subgroup that is more vulnerable to increased adiposity, and therefore more at risk of later obesity-related disease. Although our sample size is relatively small, our data provide strong impetus for further examination of the role of APOE in determining adiposity in larger studies of both children and adults.