- Top of page
- Experimental procedures
Smith–Lemli–Opitz syndrome (SLOS) is a complex hereditary disease caused by an enzymatic defect in the last step of cholesterol biosynthesis. Progressive retinal degeneration occurs in an AY9944-induced rat model of SLOS, with biochemical and electroretinographic hallmarks comparable with the human disease. We evaluated alterations in the non-sterol lipid components of the retina in this model, compared with age-matched controls, using lipidomic analysis. The levels of 16:0–22:6 and 18:0–22:6 phosphatidylcholine molecular species in retinas were less by > 50% and > 33%, respectively, in rats treated for either 2 or 3 months with AY9944. Relative to controls, AY9944 treatment resulted in > 60% less di-22:6 and > 15% less 18:0–22:6 phosphatidylethanolamine molecular species. The predominant phosphatidylserine (PS) molecular species in control retinas were 18:0–22:6 and di-22:6; notably, AY9944 treatment resulted in > 80% less di-22:6 PS, relative to controls. Remarkably, these changes occurred in the absence of n3 fatty acid deficiency in plasma or liver. Thus, the retinal lipidome is globally altered in the SLOS rat model, relative to control rats, with the most profound changes being less phosphatidylcholine, phosphatidylethanolamine, and PS molecular species containing docosahexaenoic acid (22:6). These findings suggest that SLOS may involve additional metabolic compromise beyond the primary enzymatic defect in the cholesterol pathway.
Smith–Lemli–Opitz syndrome (SLOS) is an autosomal recessive, multiple congenital anomalies disorder caused by a defect in the last enzymatic step in the cholesterol (Chol) biosynthetic pathway (Smith et al. 1964) (reviewed in: Yu and Patel 2005; Correa-Cerro and Porter 2005; Kelley and Hermann 2001). The dyslipidemia in patients affected with this disorder has a specific biochemical signature: blood and other tissues contain abnormal (typically grossly elevated) levels of 7-dehydrocholesterol (7DHC) and markedly reduced levels of Chol, relative to the levels found in normal controls (Irons et al. 1993; Tint et al. 1994). This is due to mutations in the DHCR7 gene, which encodes the enzyme 3β-hydroxysterol-Δ7-reductase (EC 126.96.36.199) (Fitzky et al. 1998; Waterham et al. 1998; Correa-Cerro and Porter 2005; Yu and Patel 2005), thereby altering the efficiency of conversion of 7DHC to Chol. The nervous system, in particular, is profoundly affected in this disease, as manifested by cognitive function deficits (e.g. mental retardation and autism) (Nowaczyk et al. 1999; Tierney et al. 2000; Sikora et al. 2006). In addition, retinal dysfunction has been reported in SLOS patients, particularly with regard to rod photoreceptor electrophysiology (Elias et al. 2003). An animal model of this human disease has been developed by treating rats with AY9944, a selective pharmacological inhibitor of the same enzyme that is defective in SLOS (Dvornik et al. 1963; Givner and Dvornik 1965; Kolf-Klauw et al. 1996; Wolf et al. 1996). Using a modification of the originally developed model that allows long-term postnatal survival, a progressive retinal degeneration affecting both rod and cone photoreceptors have been described (Fliesler et al. 2004). Whereas the normal rat retina contains virtually no 7DHC, the mole ratio of 7DHC/Chol in retinas of 1-month-old rats treated with AY9944 as gestation approaches ca. 4 : 1, but without associated histological or functional impairments (Fliesler et al. 1999). In contrast, by three postnatal months, the 7DHC/Chol ratio in retinas of AY9944-treated rats reaches levels > 5 : 1, with marked histological degeneration and associated electrophysiological defects (Fliesler et al. 2004). Recently, it was shown that feeding a high-Chol diet to these rats from weaning through three postnatal months provides a marked improvement primarily in cone photoreceptor function, and also significantly improves (toward normal) the sterol profile of the retina, but does not spare the retina from degeneration (Fliesler et al. 2007).
In the course of our studies of the retinal degeneration in the SLOS rat model, we examined the lipidomic profile of the retina as a function of postnatal age, in comparison with age-matched control retinas, to assess possible changes in lipid species other than sterols. We report herein that the fatty acid composition of the major retinal phospholipid molecular species is dramatically altered in the SLOS rat model, relative to that of controls. Remarkably, the relative docosahexaenoic acid (DHA; 22:6n3) content of these phospholipids was significantly reduced within a 3-month time course, to an extent comparable with or greater than that achieved heretofore by raising rats on a n3 fatty acid-deficient diet (Futterman et al. 1971; Anderson and Maude 1972; Tinoco et al. 1977; Wiegand et al. 1991; Anderson et al. 1992; Bush et al. 1994), except under extraordinary circumstances (Tinoco et al. 1978; Ward et al. 1996; Moriguchi et al. 2004). However, under the conditions employed in this study, there was no evidence of generalized, systemic DHA or n3 fatty acid deficiency. We discuss these findings within the context of the role of lipids in supporting normal visual function, and also with regard to the pathobiology of SLOS. To the extent that the AY9944-induced rat model faithfully mimics the human disease, these findings suggest that the metabolic defect in SLOS may involve other pathways beyond the primary defect in the Chol pathway.
- Top of page
- Experimental procedures
In the present study, we have demonstrated striking alterations in the phospholipid molecular species profile of the retina in rats treated with AY9944, compared with age- and sex-matched control rats, particularly with respect to the steady state levels of DHA. The predominant choline glycerophospholipid molecular species in control retinas were 16:0–22:6 and 18:0–22:6; the levels of these molecular species were lower by ∼30–50% in rats treated for 2–3 months with AY9944 in comparison with age-matched control rats. The dominant DHA-containing retinal PE molecular species in both control and treated rats were 18:0–22:6 and di-22:6. In comparison with controls, AY9944 treatment resulted either in no difference (at 1 month), a 22% and 73% difference (by 2 months), or a 21% and 69% difference (by 3 months) of these ethanolamine glycerophospholipid molecular species, respectively. The predominant retinal PS molecular species were 18:0–22:6 and di-22:6; AY9944 treatment resulted in > 90% less di-22:6 PS, relative to controls. These findings demonstrate that the retinal lipidome is globally altered in the AY9944-treated rat, relative to control rats, over and above the changes in sterol composition (elevated 7DHC and reduced Chol levels) because of the primary inhibition of Chol biosynthesis caused by AY9944. To the extent that the AY9944-treated rat represents as suitable model of SLOS, these results suggest that additional metabolic compromise beyond the primary enzymatic defect in the Chol pathway may be involved in the pathobiology of the human disease.
It should be noted that AY9944 is presumed to be a selective inhibitor of 3β-hydroxysterol-Δ7-reductase (DHCR7; EC 188.8.131.52) (Dvornik et al. 1963; Givner and Dvornik 1965), and has no reported effects on any enzyme involved in fatty acid or phospholipid biosynthesis. This is consistent with our finding that the n3 fatty acid content (especially DHA) of plasma and liver was not reduced in rats treated systemically with AY9944. Also, depletion of retinal DHA levels was achieved over a relatively brief time frame (2–3 months), even though the diet contained adequate levels of 18:3n3 (ca. 6 mol% of total fatty acids, as determined by GLC analysis; data not shown) and both AY9944-treated and control rats were fed the same diet. These findings are all the more remarkable when one considers the well-established fact that the vertebrate retina, particularly the outer segment membranes of retinal rod photoreceptor cells, and in striking contrast to other bodily tissues, strongly resists depletion of DHA and other n3 fatty acids when challenged with an n3-deficient diet (Futterman et al. 1971; Anderson and Maude 1972; Tinoco et al. 1977, 1978; Wiegand et al. 1991; Anderson et al. 1992; Bush et al. 1994). This is thought to be a consequence of an extremely active and efficient recycling mechanism, involving both the retinal pigment epithelium and the neural retina (Scott and Bazan 1989; Bazan et al. 1992, 1994).
The retina requires high levels of DHA for optimal function. Reduction of DHA in the rod outer segments (ROS) of rats (Benolken et al. 1973; Wheeler et al. 1975) and monkeys (Neuringer et al. 1984), by dietary restriction of n3 precursors, leads to reduced amplitudes and other changes in the electroretinogram. More recent studies have shown that ROS membranes with low DHA levels have a slower visual response to light (Niu et al. 2004). At the molecular level, this is seen as a slower transition of meta-rhodopsin-I to meta-rhodopsin-II, an important step in the activation of the G protein, transducin, and subsequent initiation of the visual signal. Studies in term and pre-term human infants have established that the development of visual acuity (Uauy et al. 1990, 1992; Birch et al. 1998) is dependent on a maternal supply of n3 fatty acids. Also, humans and dogs with inherited retinal degenerations have reduced plasma levels of n3 polyunsaturated fatty acids (Anderson et al. 1987, 1991). In addition, the retinas of dogs (Aguirre et al. 1997), rats (Anderson et al. 2002), and mice (Anderson et al. 2001a,b) with inherited retinal degenerations have significant reductions of DHA in their ROS membranes. These results are of further interest because the loss of DHA is not accompanied by an increase in 22:5n6, which typically occurs in n3 fatty acid deficiency (Anderson and Maude 1972; Galli et al. 1972; Wiegand et al. 1995). The reason for this reduction in DHA, even in the presence of adequate levels of dietary n3 fatty acids (which we also observed in the current study) is unclear. The only other example of a reduction of DHA in the retina without a compensatory replacement by 22:5n6 occurs in albino rats (Penn and Anderson 1987, 1992) and mice (Káldi et al. 2003) raised in bright cyclic light. The reasons for the reduction of DHA in these various animal models are not known. However, it is clear that retinas destined for degeneration, whether by heredity or by light-induced stress, respond by reducing the levels of DHA in their retinas, particularly in photoreceptor cell membranes. This is consistent with what we observed in this animal model of SLOS. These changes are all the more intriguing because they reflect metabolic events occurring specifically in the retina, rather than those occurring in other tissues.
The mechanism by which this extraordinary remodeling of the retinal lipidome is achieved in SLOS rats likely involves multiple aspects of lipid transport and homeostasis, both systemic as well as retina-specific. First, it is possible that differential uptake of essential fatty acids may occur in the gut as a consequence of alterations in the composition and content of bile acids (required for solvation of lipids), because of reduced and abnormal sterol biosynthesis caused by AY9944. While not examined specifically in AY9944-treated rats, it has been reported that human SLOS patients have altered urinary bile acid composition, both with respect to decreased levels of normal bile acids and appearance of abnormal bile acids (Natowicz and Evans 1994). This would be expected, as Chol is an obligatory precursor of bile acids. However, this finding is somewhat controversial, since a more recent report by Steiner et al. (2000) has shown that while total sterol synthesis was reduced (by about 40%) in their cohort of SLOS patients, bile acid synthesis was not substantially different from control levels, and both normal primary and secondary bile acids were detected. Second, it is possible that the content of lipoprotein particles carrying DHA-enriched lipid constituents, which are assembled in and secreted by the liver, is different (i.e. DHA-deficient) in AY9944-treated versus control rats. This is an important consideration, as the retina derives its DHA from the liver, not from local de novo synthesis (Scott and Bazan 1989). However, the results of our detailed fatty acid analysis of serum and liver rule out this possibility, as serum DHA levels were equivalent to (at 1 and 2 months) or even slightly higher (at 3 months) in AY9944-treated animals, and liver DHA levels were slightly elevated at 3 months, compared with control values (see Figs 8 and 9). In addition to this prima facia lack of reduced DHA levels in serum and liver, systemic n3 fatty acid deficiency in AY9944-treated rats is also ruled out by the fact that there was no observed elevation of 22:5n6 in these tissues, which typically occurs in n3 fatty acid deficiency (see above). If fatty acid absorption in the gut (the first possibility mentioned above) was a significant factor, one would expect the fatty acid composition of serum and liver to be affected as well. However, given the fact that no DHA or more generalized n3 fatty acid deficiency was apparent in either serum or liver as a function of AY9944 treatment, the first possibility also tends to be ruled out. This leads to a third, and most likely, possibility: disruption in the uptake of blood-borne, DHA-containing lipoproteins by the retina, which also involves the retinal pigment epithelium at the choroidal interface, and subsequent intraretinal redistribution of the lipid constituents among the various retinal cell types and metabolic compartments, may be altered in AY9944-treated rats, compared with controls. This latter process is complex, involving multiple carriers (both intracellular and extracellular), receptors, and enzymes (Tserentsoodol et al. 2006). Fourth, there may be adaptive changes in de novo fatty acid synthesis in photoreceptors and other retinal cell types as a consequence of AY9944 treatment (see below). Resolution of these various possibilities is beyond the scope of the present study, but is the subject of active ongoing investigations in our laboratories.
Although DHA-containing phospholipid molecular species were the most profoundly altered lipids detected in this study, the levels of other fatty acids also were affected by AY9944 treatment. To some extent, this could reflect changes in de novo synthesis occurring within the retina per se. Although we have not yet performed metabolic labeling experiments to examine this further, we have obtained preliminary evidence, using microarray analysis, that indicates the expression levels of several genes involved in fatty acid and phospholipid biosynthesis are significantly altered in retinas of AY9944-treated rats, compared with controls, particularly after 2 months of treatment (Siddiqui et al. 2007). These include, but are not limited to, fatty acid synthase, stearoyl-CoA desaturase, acyl-CoA thioesterase, acyl-CoA synthetase, and CDP-diacylglycerol synthase. As AY9944 has no known direct effect on any of these enzymes, it is conceivable that crosstalk between the Chol biosynthetic pathway and the fatty acid and phospholipid biosynthetic pathways, mediated via sterol response elements and the sterol regulatory element-binding protein/SREBP cleavage-activating protein (SREBP/SCAP) system (reviewed in Shimano 2001; Rawson 2003; McPherson and Gauthier 2004), may explain, at least in part, these observations.
It should be appreciated that by 1 month of AY9944 treatment, there is no retinal degeneration and no electrophysiological dysfunction of the retina; yet, at the same time, retinal sterol composition is grossly deranged, with the 7DHC/Chol mole ratio being nearly 4 : 1 (Fliesler et al. 1999). In the present study, we found that alterations to retinal phospholipid molecular species after 1 month of AY9944 treatment were minimal. However, by 3 months of AY9944 treatment, substantial retinal degeneration and dysfunction accompanies even greater derangement of sterol metabolism and composition (Fliesler et al. 2004). Here, we found that retinal phospholipid molecular species composition is dramatically altered within 2–3 months of AY9944 treatment, compared with age-matched controls. Hence, phospholipid composition seems to correlate more closely with retina histological and electrophysiological integrity than does sterol composition in this animal model. However, it should not be inferred from these findings that reduced levels of DHA somehow caused the observed retinal degeneration. [Similarly, in the cases of hereditary retinal degeneration in animals, as mentioned above, there is no evidence that reduced levels of DHA cause the retinal degeneration.] A correlation between n3 fatty acid content and retinal development and function has been established for many years, both for rats (Benolken et al. 1973; Wheeler et al. 1975) as well as humans and non-human primates (Neuringer et al. 1984; Uauy et al. 1990, 1992; Birch et al. 1998). However, given the fact that we found no evidence for a systemic n3 fatty acid deficiency in this SLOS rat model, it is highly unlikely that feeding a high n3-containing (e.g. 18:3n3- or 22:6n3-rich) diet in addition to Chol supplementation, as opposed to a high-Chol diet alone, would provide any additional benefit with regard to improving retinal structure or function in the SLOS rat model. Indeed, even when AY9944-treated rats are fed a high-Chol diet for 2 months, which tends to nearly normalize their serum and retina sterol compositions, there is no protection against retinal degeneration (Fliesler et al. 2007). Furthermore, in a study by Martin et al. (2004), it was found that although supplementation with a diet enriched in n3 fatty acids was able to alter and partially restore the normal steady stated fatty acid profile of the retina (and, more specifically, of rod photoreceptor outer segment membranes), such dietary manipulation was neither able to protect against photoreceptor cell death nor alter the course of hereditary retinal degeneration in two different rat models carrying rhodopsin mutations linked to human retinal disease. In fact, the retina is known to conserve DHA under conditions of dietary n3 fatty acid deficiency (Anderson and Maude 1972; Wiegand et al. 1991), and even the small amounts of DHA present in standard rat chow are normally sufficient to support the high levels of DHA typically found in retinal photoreceptor membranes (Benolken et al. 1973). Taken together, these findings would suggest that the mechanism underlying the profound decrease in retinal DHA levels in the AY9944-treated rat model of SLOS is fundamentally different from that which underlies retinal DHA decline under conditions of systemic n3 fatty acid deficiency (e.g. as induced by dietary manipulation).
The findings obtained with this AY9944-treated rat model would predict that the retinas of SLOS patients may have altered fatty acid composition, particularly abnormally low DHA levels, compared with unaffected, age-matched normal individuals. Unfortunately, there are no reports in the literature, at present, to either confirm or negate this prediction, and obtaining SLOS as well as normal human retinal tissue specimens is extraordinarily difficult. However, we currently are pursuing a study of the serum fatty acid profiles of SLOS patients, comparing them with normal controls, to ascertain if there is any correlation between disease severity and serum fatty acid profile. Our findings with the AY9944 rat model would predict a lack of such correlation.