There are several interesting features of the fatty acid compositional data. First, there are no differences in the levels of DHA, the most abundant fatty acid in retinal lipids, in any of the retinas we examined. This is in sharp contrast to what we have previously reported for retinas from other mutant and transgenic rodent retinal degeneration models.
30,31 For example, retinas from P23H and S334ter rhodopsin mutant rats have lower DHA levels than their littermate controls.
31 Lines of P23H rats with the fastest rate of degeneration have lower levels of DHA compared with controls and those that degenerate slower. Mice with
Rds 30 and
Rhodopsin G90D
32 mutations have lower retinal DHA levels than appropriate controls. Of the various mutant rodent models we have analyzed, only those with ELOVL4 mutations or transgene expression of mutant protein, especially TG2
+ retinas which undergo rapid degeneration, do not have lower retinal DHA levels. Second, although there were different levels of expression of wild-type
Elovl4 in
Elovl4+/− and
Elovl4+/mut mouse retina, the levels of VLC-PUFA were not different between the two groups at this age point, both having ∼50% compared with controls. This reduction was expected, although we also expected a greater reduction in
Elovl4+/mut retinas due to expression of the mutant protein. However,
Elovl4+/mut skin appears to produce less VLC-FA, which are primarily found in ceramides, than
Elovl4+/− skin (
Fig. 3B).
Elovl4 is a high-expressing gene in the retina with expression levels 10- to 20-fold higher than in skin.
1,22 This is probably the reason that a small difference (10%) in the protein levels (enzyme activity) does not reflect a comparable difference in product levels in the retina between
Elovl4+/mut and
Elovl4+/− , whereas the skin enzyme expression is not as abundant as in retina and thus their products are more affected. Several labs, including ours, showed a dominant negative effect of the mutant protein on the localization and function of wild-type protein in in vitro and cell culture studies.
7,8,25 It was surprising for us not to see a dominant negative effect of mutant ELOVL4 on the function of wild-type ELOVL4 in
Elovl4+/mut retinas at this younger age. It will be interesting to follow this with age as the
Elovl4+/mut mice are known to develop retinal pathology slowly and with aging.
11,16 Third, there was a progressive reduction in the formation of longer chain length VLC-PUFA in retinas of both
Elovl4+/− and
Elovl4+/mut mice (C32 > C34 > C36). The same was also noted in the skin profiles, where the levels of 30:0 were more affected than 28:0 in
Elovl4+/− and
Elovl4+/mut mice. This suggests that the enzyme activity of ELOVL4 is highly sensitive to the availability of substrates, which become more limited with increasing chain length. Fourth, the presumed expression of different levels of mutant human ELOVL4 in the TG mouse retinas (the mRNA levels were increased; however, there is no antibody to mutant human or mouse ELOVL4 so the level of protein expression cannot be determined) did not affect the retinal levels of VLC-PUFA (compare TG1
+, TG2
+, WT1
+, and TG
− in
Fig. 4). This shows that the mutant human protein, if expressed, does not affect the enzymatic activity of the endogenous mouse ELOVL4. This conclusion is supported by the discussion above regarding the lack of an effect of mutant protein in
Elovl4+/mut mice on levels of VLC-PUFA compared with
Elovl4+/− retinas. This apparent lack of a dominant negative effect of mutant protein on the wild-type protein in mouse retinas raises questions and suggests that further studies are necessary to understand the mechanism of ELOVL4 mutation-mediated STGD3 in humans. Recent studies from our laboratory suggest that VLC-PUFA are important for rod function and survival, as we found that a reduction of VLC-PUFA (>90%) by conditionally deleting
Elovl4 in both rods and cones leads to a loss of rods at 6 months of age and a reduction in rod function (ERG) (Marchette LD, et al.
IOVS 2012;53:ARVO E-Abstract 4654). Thus, the loss of VLC-PUFA could be a contributing factor since these mice do not express mutant ELOVL4, meaning the photoreceptor cell loss cannot be explained by the presence of mutant protein. This can explain the slower development of the mild phenotype in older
Elovl4+/mut mice; the efficiency of ELOVL4 synthesis may go down with aging, which may result in a reduction in the VLC-PUFA levels in these retinas, contributing to the rod cell loss and loss of rod function.
10,11 However, it does not explain a dominant early onset human STGD3 development.