In the present study, we generated a novel mouse model of LMF1 deficiency using gene-trap mutagenesis. The new model offers advantages over the naturally occurring cld mutant mouse strain used in previous studies [5, 8]. First, in contrast to a truncating mutation in the cld model, the gene-trap insertion characterized here represents a definitive null-allele and allows analysis of the full phenotypic manifestation of LMF1 deficiency. An additional improvement of the new model over cld is that wild-type littermates are viable, which enables characterization of the potential phenotypic consequences of heterozygous LMF1 deficiency. Finally, the gene-trap allele is on an inbred genetic background devoid of the confounding genetic effects associated with the cld mutation .
We unexpectedly detected ubiquitous and relatively high-level LMF1 expression in the mouse embryo. This observation raised the possibility that in addition to its established role in postnatal lipid metabolism, LMF1 may also be an important factor in embryogenesis. While the role of vascular lipases during development remains poorly characterized, embryonic expression of all three LMF1-dependent lipases (i.e. LPL, HL and EL) has previously been documented [14, 15]. To begin to address the potential role of LMF1 in development, we first asked whether LMF1 deficiency affects the viability of embryos. Our results demonstrate that despite widespread expression in the developing embryo, LMF1 is not required for embryonic survival. However, postnatal viability of LMF1-/- pups is severely compromised, as no surviving LMF1-deficient progeny was detected a few days after birth. Neonatal lethality has also been observed in cld mice  and is thought to be a consequence of circulatory problems associated with severe hyperchylomicronemia, which results from the inability to utilize dietary fat during suckling . Indeed, LMF1-/- pups exhibit hypertriglyceridemia and severely diminished post-heparin LPL and HL activities, hallmark features of LMF1 deficiency in the cld mouse model . In addition to hypertriglyceridemia, plasma concentrations of total cholesterol are also elevated in LMF1-/- mice, a likely consequence of diminished catabolism and accumulation of chylomicron particles due to LPL deficiency . In contrast to total cholesterol, HDL-cholesterol levels are unaffected in LMF1-/- plasma. At first glance, this is a surprising observation considering the critical role of LPL in the maturation of HDL particles  and severely reduced HDL-cholesterol in LPL-deficient mice [17, 19, 20]. However, in addition to LPL, LMF1-/- animals are also deficient in active HL and EL , lipases that promote HDL catabolism [21–23]. Thus, we propose that unaltered HDL-cholesterol level in LMF1-/- mice is a result of combined lipase deficiency involving lipases with opposite effects on HDL metabolism.
LMF1-deficient mice developed in this work will allow in-depth investigations of the role of LMF1 in development. However, early lethality is a limitation of this model for metabolic studies in the adult organism. Neonatal lethality in LMF1-/- mice is not unique among mouse models of hypertriglyceridemia. LPL-deficient mice die within a few days after birth, most likely as a consequence of restricted oxygen exchange in lipid-engorged lung capillaries . In contrast, hypertriglyceridemia in mice deficient in GPIHBP1, a protein involved in the trans-endothelial transport of LPL, is not associated with increased mortality owing to the availability of a functional pool of LPL in the neonatal liver [24, 25]. Two strategies have been used to rescue LPL-deficient mice from neonatal lethality. First, transient expression of LPL through adenoviral gene transfer allowed a small fraction of infected LPL-/- progeny to reach adulthood and enabled metabolic characterization of LPL-deficiency [18, 20, 26]. We attempted a similar strategy to rescue LMF1-/- pups using adenovirus expressing LMF1, but have been unable to recover adult LMF1-deficient mice (unpublished observation). Importantly, LMF1-deficient mice are not only devoid of active LPL, but also HL and EL, which may result in more severe morbidity relative to LPL-deficiency only and explain why transient expression of LMF1 is insufficient for rescue. Consistent with this explanation, HL/EL double knock-out mice suffer from neonatal lethality , which raises the possibility that combined HL/EL-deficiency contributes to mortality in LMF1-/- mice. A second strategy that has been successfully applied to rescue LPL-/- mice is based on transgenic complementation of LPL expression in single tissues such as heart, skeletal muscle and liver [27–29]. A similar approach is currently pursued in our laboratory to rescue LMF1-/- mice using a muscle-specific LMF1 transgene .
In conclusion, we validated a new LMF1-deficient mouse model by demonstrating that it recapitulates salient phenotypes of cld mutant mice including neonatal lethality, dyslipidemia and combined lipase deficiency. At the same time, our study also confirms that phenotypes previously observed in the cld model are genuine consequences of LMF1 deficiency, as opposed to unrelated mutations present in the cld genetic background . The fact that our initial characterization of LMF1-/- neonates did not reveal novel phenotypes beyond those already observed in cld mice is consistent with the possibility that cld represents a null-allele of LMF1. However, more detailed phenotypic characterization will be necessary before a definitive conclusion can be reached. The LMF1-deficient mouse model developed in this study will facilitate further analysis of LMF1 function in development and metabolic regulation.