Familial combined hyperlipidemia (FCHL), the most common inherited disorder of abnormal blood lipid levels, is believed to share part of its etiology with the metabolic syndrome1. The extent of this overlap has been difficult to define because of important differences in diagnosis. FCHL is typically diagnosed through an index individual affected with marked hyperlipidemia (elevated blood cholesterol or triglyceride levels) and normally excludes individuals with hyperlipidemia secondary to other coronary heart disease (CHD) risk factors, such as central obesity and maturity-onset diabetes (type 2 diabetes). More fundamentally, a diagnosis of FCHL demands that index individuals have a blood relative with primary hyperlipidemia, whereas the metabolic syndrome does not. On page 371–376 of this issue, Päivi Pajukanta and colleagues2 present evidence that a specific haplotype of the gene encoding upstream transcription factor 1 (USF1) located in the chromosome 1q21–23 region, is associated with FCHL. They propose that USF1 is a good candidate for conferring susceptibility to the core features of the metabolic syndrome: insulin resistance, glucose intolerance, type 2 diabetes, central obesity, dyslipidemia (elevated blood triglyceride or low high density lipoprotein-cholesterol) and hypertension2.

Homing in on USF1

Pajukanta et al.3 previously found evidence for linkage of the chromosome 1q21–23 interval to FCHL in a cohort of 31 Finnish families, identified through an index individual with early-onset CHD and blood cholesterol or triglyceride levels greater than or equal to age- and sex-specific 90th percentiles. Subsequent studies supported these data and implicated the same chromosomal region in the etiology of type 2 diabetes4,5,6.

In the present study, Pajukanta et al.2 built on their previous work by systematically examining a stretch of DNA containing about 10 Mb for a genetic lesion conferring susceptibility to FCHL2. This involved tracking the transmission of 56 functionally unbiased single-nucleotide polymorphisms (SNPs) to the affected family members of 42 pedigrees with FCHL, including the 31 families who participated in the original linkage study. They examined six SNPs that had some association with FCHL in an additional 18 extended families. The two SNPs that produced the best evidence for association with FCHL are both located in the gene encoding USF1, a transcriptional activator that regulates a number of genes involved in whole-body lipid and glucose homeostasis7,8. This report prompts two important questions. First, how compelling is the genetic evidence for the involvement of USF1 in FCHL and, by implication, the etiology of the condition? Second, irrespective of the contribution of USF1 to FCHL, do specific USF1 haplotypes increase the risk of type 2 diabetes or the metabolic syndrome?

Regarding the first issue, in the absence of existing linkage data, some investigators9 argue for a statistical criterion of P < 5 × 10−9, whereas others emphasize the importance of replicating results in independent data sets and of biological plausibility10. In the present FCHL study, the alleles at the two associating loci (usf1s1 (exon 11) and usf1s2 (intron 7)) are in strong linkage disequilibrium and, based on family data, have major allele frequencies of about 0.65 in the Finnish population. The lowest P value (9 × 10−7) was obtained for the triglyceride trait in men using the gamete-competition test, which views the transmission of alleles or haplotypes to affected individuals as a contest. These data were supported by the results of a second test statistic (HBAT-o) that determined which specific haplotypes were preferentially transmitted. The haplotype containing the minor alleles at the usf1s1 and usfs2 loci was transmitted less frequently to affected males with the triglyceride trait (P = 4 × 10−3), prompting the authors to suggest that this might protect against FCHL in their families.

The HBAT-o test also produced evidence (P = 7 × 10−4) for preferential transmission of the common USF1 haplotype to males with the triglyceride trait of FCHL, supporting the results from three different analyses: the gamete-competition test; a multilocus geno-PDT and a haplotype-based haplotype relative risk test. But even robust statistical data from a single cohort of families are not always replicated in another population. Because the Finnish families were, necessarily, selected on the basis of a complex phenotype and did not exclude individuals with type 2 diabetes, the size of the effect of the putative USF1 risk haplotype on any single FCHL-related trait is uncertain. Given the relatively high frequency of the risk haplotype, this effect will presumably vary according to an individual's susceptibility to develop FCHL, which we now know is attributable to at least four other genes11.

A SNP in an internal promoter

The evidence for association of USF1 with FCHL prompted Pajukanta et al.2 to sequence this gene in 31 FCHL probands2. They found no amino acid changes in USF1 that could account for association of the putative USF1 risk haplotype with FCHL. Nor could they find any difference in USF1 mRNA levels in fat biopsy samples from individuals with FCHL with the risk haplotype versus those without. They did, however, identify a putative promoter in intron 7 of USF1. Although the usf1s2 polymorphism, which resides in or close to the putative promoter, had no impact on transcriptional activity, its identification is somewhat encouraging. In principle, the putative promoter could enable USF1 to initiate translation from one of two internal AUGs in exon 8 and, in doing so, throw a switch that reverses the normal function of this protein from activation to repression or vice versa, as has been described for a number of other transcription factors12 (Fig. 1). Evidence that the intron 7 promoter is used in vivo will be keenly awaited.

Figure 1: A theoretical scenario to illustrate how the use of a putative promoter in intron 7 of USF1 (orange box) might lead to the generation of a mini-USF1 protein lacking the transactivation domain.
figure 1

Because mini-USF proteins in vitro behave as transdominant inhibitors14,15, there is a substantial incentive to establish whether the putative promoter identified by Pajukanta et al.2 might operate in vivo to downregulate USF1 activity. The two polymorphisms associated with FCHL in Finnish families are indicated. b-HLH-Zip, basic helix-loop-helix-leucine zipper.

Full size image

In summary, the work of Pajukanta et al.2 has identified USF1 as the prime candidate in the chromosome 1q21–23 linkage region for increasing susceptibility to FCHL, type 2 diabetes and metabolic syndrome. The challenges ahead will be to confirm the genetic data through replication, to identify other genes associated with FCHL at additional loci and to determine the roles of the different isoforms of USF1 in the regulation of genes pivotal for whole-body lipid and glucose homeostasis and the maintenance of the arterial wall vasculature13.