In this issue of the Journal of Bone and Mineral Research, Deickmann and colleagues1 report that different isoforms of the human Apolipoprotein E gene (APOE) significantly influence bone turnover and bone mass in mice, and they present some data to suggest the same might be true in man.1
The preclinical experiments were performed using genetically engineered mice in which the endogenous coding sequences of ApoE had been replaced with one of the human ApoE2, ApoE3, or ApoE4 isoforms (so called knock-in mice). These three isoforms differ from each other in the amino acid sequence at codons 118 and 158. The most common variant is ApoE3, which contains a cysteine residue at codon 118 and an arginine at codon 158. The ApoE2 variant has cysteine residues at both sites, and the ApoE4 variant has arginine residues at both sites.
When compared with the other isoforms, ApoE2 has reduced affinity for binding hepatic low-density lipoprotein receptors and, as a result, between 1% and 4% of homozygotes for this isoforms develop type III hyperlipoproteinemia. APOE plays a major role in lipid homeostasis and regulates the risk of cardiovascular disease through this mechanism. APOE has also been implicated in the pathogenesis of Alzheimer's disease, which is strongly associated with carriage of ApoE4; however, the mechanisms by which the ApoE4 variant predisposes to Alzheimer's are incompletely understood.2
The experiments of Deickmann and colleagues1 were prompted by the observation that deletion of ApoE in mice is associated with high bone mass due to increased bone formation3 and by genetic association studies indicating that common polymorphic variants of ApoE may be genetic risk factors for osteoporosis and fragility fractures.4, 5 Although many association studies have been performed in various populations, and evidence has been sought for an association between ApoE isoforms and bone phenotypes, the results have been conflicting, probably because the studies have been underpowered to detect the very modest effects on bone mineral density (BMD), bone turnover markers, and fracture that would be expected for variants of this type.6 For example, in some studies, trends were observed for reduced BMD and increased fracture risk in carriers of the ApoE4 allele,4, 7, 8 which is the opposite of the results reported by Deickmann and colleagues1 in mice; however, in other studies, no associations between ApoE alleles and BMD or fractures were observed.9, 10 Moreover, in a recent meta-analysis of genome-wide association studies that involved more than 100,000 subjects, the ApoE locus did not emerge as a significant determinant of BMD or fracture.7
These results contrast with the quite substantial 25% reduction in trabecular BMD that Deickmann and colleagues1 observed in mice homozygous for the ApoE2 allele. They argue that the lack of association observed by others may be attributable to the fact that most investigators have failed to study ApoE2 homozygotes as a separate group, and instead have elected to perform an allele-specific analysis in which the (relatively rare) ApoE2 homozygotes were combined with ApoE2/E3 heterozygotes. In order to circumvent this problem, Deickmann and colleagues1 studied biochemical markers of bone turnover in a small group of male ApoE2 homozygotes, and compared them with markers found in homozygotes for the ApoE3 and ApoE4 alleles. The serum receptor activator of NF-κB ligand (RANKL) levels and the RANKL/osteoprotegerin (OPG) ratio were higher in the ApoE2 group, replicating the observations found in the ApoE2 knock-in mice; however, there were no differences in other biochemical markers of bone turnover. Unfortunately, this study did not provide information on BMD because the cohort they used was primarily designed to look at cardiovascular rather than bone endpoints. Only one previous study used a similar approach to that of Deickmann and colleagues,1 and that study found no difference in BMD values or biochemical markers of bone turnover in ApoE2 homozygotes (n = 18) as opposed to ApoE4 homozygotes (n = 18).11 No doubt, however, promoted by Deickmann and colleagues'1 results, other investigators who have studied these polymorphisms in reasonably large numbers of individuals9, 10 may go back to their original data to determine if they can detect a significant association between ApoE2 homozygosity and BMD or biochemical markers of bone turnover.
Although the data of Deickmann and colleagues1 clearly show that human ApoE2 does influence bone mass and bone turnover when expressed in the homogeneous genetic background of a knock-in mouse, the situation is rather more complex in humans, where hundreds of genetic variants contribute to the regulation of bone mass and bone turnover.7 Although one cannot exclude the possibility that common ApoE isoforms play a role in regulating bone metabolism in man, it seems probable that they do not have a large enough effect size for associations to be consistently observed in human studies.