Abstract
Post-translational modifications of coagulation factors in the liver are essential for function. The vitamin K dependent coagulation proteins (VKCPs) require vitamin K to undergo gamma carboxylation of the glutamic residues in their Gla domain by gamma-glutamyl carboxylase [GGCX]. The vitamin K is then recycled by the action of epoxide reductase [VKORc1] and/or quinone reductase [NQO1]. The hemostatic importance of the vitamin K “cycle” is evidenced by patients who may suffer bleeding complications when anticoagulated with warfarin, which targets the vitamin K cycle. Moreover, the ability of a variety of VKCPs to secrete a biologically active product depends on the removal of their propeptide by the action of the intracellular endoprotease furin [FURIN gene]. Previous in vitro work on recombinant coagulation Factor IX, which is used for hemophilia B treatment, has connected these two processing steps by showing that endogenous VKORc1 as well as FURIN can be limiting factors in high-yield expression systems. In vivo, skeletal muscle (in contrast to liver) has been utilized to express low levels of coagulation Factor IX in the first hemophilia B gene therapy clinical trial. However, our experiments in mice demonstrated that the specific activity of muscle-synthesized Factor IX via gene transfer decreased at the high levels of FIX expression by a limited muscle area (Schuettrumpf J. et al., Blood 2005). These results suggest that in vitro and in vivo expression of biologically-active VKCPs outside the liver may be limited by the host cell post-translational modification machinery. Here, we performed a systematic study to determine the expression profiles of the vitamin K cycle and furin endoprotease genes in human liver and muscle, compared to the mouse. We also established these profiles in two hemophilic dogs, given the extensive use of this animal model in gene-based hemophilia therapies. RNA from liver and skeletal muscle was used as a template for reverse transcription and the subsequent relative quantification of the GGCX, VKORc1, NQO1, and FURIN genes by qPCR in each tissue using a housekeeping reporter gene. For this, a variety of housekeeping genes were investigated in all three species to identify ones with similar transcript levels in both liver and muscle tissue. We identified the housekeeping genes HPRT1, beta actin, and 18s rRNA as equivalently expressed in the liver and skeletal muscle of human, mouse, and dog, respectively. The relative mRNA transcript quantification of the vitamin K cycle genes in humans showed that the transcript levels of GGCX were similar in liver and muscle. In contrast, both VKORc1 and NQO1 were under-expressed in muscle vs. liver (69.5 ± 4.9% and 67.8 ± 12.5%, respectively, P<0.01). In the mouse, VKORc1 transcript levels in the muscle were reduced to 73.8 ± 9.9% vs. liver (P<0.05), while GGCX and NQO1 exhibited similar transcript levels in both tissues. In the dog, we observed a dramatic reduction in VKORc1 and GGCX transcript levels in the muscle vs. liver (11.8 ± 4.2% and 29.5 ± 15.8%, respectively, P<0.01). Surprisingly, NQO1 transcript levels were 253.8 ± 156.7% higher in muscle than liver (P<0.05). Lastly, in all three species tested, transcript levels for FURIN were similar in both muscle and liver. Our results indicate that VKORc1, a key enzyme in the vitamin K cycle, is consistently under-expressed in the skeletal muscle of humans as well as in mice and hemophilic dogs. In contrast, FURIN transcripts are similarly abundant in the liver and muscle of all three species tested. These suggest that the vitamin K cycle but not propeptide processing by furin can be a limiting factor in the secretion of biologically active muscle-expressed VKCPs. As a result, our observations provide (1) a plausible explanation for the inverse relationship between specific activity and Factor IX expression levels in mice following Factor IX gene transfer, and (2) further support for the mouse and dog as useful models for therapies that depend on the muscle-derived expression of VKCPs.
Disclosures
No relevant conflicts of interest to declare.
Vitamin K2 Needs an RDI Separate from Vitamin K1