Objectives. To investigate if exogenous cholesterol affects sterol turnover in the cholesterol-synthesis defect Smith–Lemli–Opitz syndrome (SLOS) and if clinical effects justify long-time supplementation. The SLOS is caused by a deficiency of the enzyme 7-dehydrocholesterol-7-reductase with markedly reduced cholesterol levels and greatly increased levels of 7-dehydrocholesterol (7-DHC).
Design. Treatment with dietary cholesterol in patients with SLOS in a case series study.
Setting. All biochemical analyses were performed in one laboratory. The clinical follow-up was carried out by one of the authors (LS), a paediatric neurologist.
Subjects. Seven patients with biochemically verified SLOS have been diagnosed in Sweden and all of them are included in the study.
Interventions. Six patients were treated for 0.5–6 years orally with cholesterol and the bile acid taurocholate and one patient was supplemented with cholesterol only.
Main outcome measures. In addition to cholesterol, 7- and 8-DHC, lathosterol was used as a marker of endogenous cholesterol synthesis and the patients were followed clinically. Nerve conduction velocities (NCV) were measured before treatment in all patients and a UVA-light test was performed in one of them.
Results. Lathosterol was initially increased by cholesterol supply in subjects with very low cholesterol levels with subsequent rise of 7- and 8-DHC. Photosensitivity clinically improved in all, verified by UVA-light testing in one. Progressive polyneuropathy improved, whilst stationary forms did not.
Conclusion. Dietary cholesterol can up-regulate sterol turnover in severely affected patients. Although some specific features are treatable and verifiable by objective methods, data supporting life-long treatment dietary cholesterol in all SLO patients are still lacking.
The Smith–Lemli–Opitz syndrome (SLOS) was first described in 1964 as an autosomal recessive disorder with microcephaly, dysmorphic features, genital abnormalities, minor abnormalities of the limbs and mental retardation . Internal malformations, severe feeding problems, photosensitivity and increased susceptibility to infections are common. The most severely affected individuals die prenatally or in infancy. The clinical presentation is, however, extremely variable and milder forms have been described with only minor mental retardation, behavioural abnormalities and subtle dysmorphic features . Clinical diagnosis of adult patients may also be difficult because of changing phenotype with age . The incidence of the SLOS varies amongst different ethnic groups but seems in some to be at least 1/60 000 . In Canada an incidence of 1/29 000 was recently reported . The true incidence of the syndrome is unknown.
The syndrome is caused by a deficiency of the enzyme 7-dehydrocholesterol reductase (7-DHCR) in the final step of the synthesis leading to hypocholesterolaemia and accumulation of the immediate precursor 7-dehydrocholesterol (7-DHC) and its epimere 8-DHC . It was the first described inborn error of an enzyme in the postsqualene biosynthesis. Mutations in the gene for 7-DHCR were defined in 1998 [7–9] and 80 different mutations have so far been described . Biochemical diagnostic methods, based on gas chromatography-mass spectrometry (gc-ms), which can also be used prenatally have been developed . The diagnostic hallmark is an elevation of 7-DHC.
Major human diseases are caused by excess of cholesterol. Cholesterol is, however, essential for embryogenesis and membrane properties. It is also a substrate for bile acids, steroid hormones and myelin. A complete lack of cholesterol is thus not likely to be compatible with life and in the SLOS a large part of the cholesterol pool is replaced by 7- and 8-DHC.
Besides possibilities of increasing our understanding of pathogenetic mechanisms, the discovery of the defective cholesterol synthesis raised hopes offinding a treatment for the disorder. The aim of therapeutic invention in the SLOS is to increase cholesterol in plasma and tissues and thereby down-regulate the de novo synthesis in order to decrease the accumulated and possibly toxic precursors.
Some benefits of dietary supplementation have been reported [12–15] but they have not been correlated to biochemical changes. The efficacy of different treatment protocols used in these trials has been difficult to evaluate because of small numbers and variable clinical and biochemical severity.
We have treated six children with biochemically verified SLOS with cholesterol and bile acids, one of them for 6 years. An additional patient has been treated with cholesterol only. As the effect on the sterol turnover of a cholesterol diet seems to vary, we have tried to evaluate cholesterol synthesis by analysing serum lathosterol, in addition to cholesterol and dehydrocholesterols. We also wanted to see if any specific features are treatable, if these can be related pathophysiologically to the metabolic defect and if they are possible to verify by objective methods. We have also questioned if the present-day knowledge of therapeutic clinical and biochemical effects justify long-term dietary supplementation with cholesterol in the SLOS.
The patients were three girls and four boys, of whom one (D) was operated upon neonatally and raised as a girl because of ambiguous genitals. One (A) had a patent ductus arteriosus and two (C, D) had neurogenic hearing loss. All were severely mentally retarded and could not speak or walk unsupported. Three (E, F, G) were fed orally, one (A) has a feeding tube and three (B, C, D) had gastrostomies. Some clinical parameters, age at the start of treatment and its duration are listed in Table 1. One patient (A) had a moderate cataract.
Table 1. Some clinical parameters and data for seven patients with the SLOS treated with dietary supplementation
Values without brackets are weight at start, with brackets the maximal deviation of weight before treatment.
Details about patient A have been published before .
Personal examination of the children and management of the dietary protocol was carried out by one of the authors (LS). Laboratory specimens were collected by their ordinary paediatricians.
The treatment started with 7–9 mg per kilo body weight of the bile acid taurocholate divided into three doses with meals for 2 weeks. Cholesterol was then added, initially 20–40 mg, later 80–100 mg per kilo daily. The first treated patient (A) did not tolerate egg yolk and after that all patients were given cholesterol powder with the exception of one (F) whose parents started him on egg yolks only for 1.5 years before taurocholate was added. One patient (G) started on cholesterol powder only 6 months ago.
Before treatment was begun, the children were investigated according to the same protocol, including measurement of the conduction velocities of peripheral nerves, blood count, liver enzymes, immunoglobulins, and vitamins. Endocrinological tests, including corticotrophin stimulation, were also performed.
Serum cholesterol and dehydrocholesterols were followed. Serum lathosterol was analysed at intervals. Liver enzymes and blood count were checked repeatedly. If polyneuropathy was diagnosed by the initial neurography, repeated recordings were made. In one patient (E) in whom photosensitivity was the main clinical indication for treatment, testing with UVA-light was carried out before and during treatment. The children were clinically followed concerning growth, development and photosensitivity.
Analyses of cholesterol, dehydrocholesterols and lathosterol were carried out by gc-ms . Mutation analyses of four of the patients (A, B, F, G) were earlier performed by the Patel laboratory in Charleston, South Carolina .
The present study was approved by the regional ethical committee for research at Karolinska Institute, Stockholm.
During the trial, one patient (B) died of pneumonia after 11 months of treatment at the age of 2.7 years. He was supplemented with 35 mg cholesterol kg−1 and 7.9 mg kg−1 taurocholate. He had a normal cortisol level before corticothrophin stimulation but a poor response (429–541 nmol L−1) after 60′. Circulatory shock could not be excluded and at the end he was given steroids with no effect.
Before treatment, serum lathosterol in the study group was 0.78–3.67 μmol L−1, with a mean of 2.23 μmol L−1. In 10 healthy children it was 0.85–2.25 μmol L−1, mean 1.67 μmol L−1.
All patients, including B, showed increased serum cholesterol (Fig. 1) and its fraction of total sterols (Table 2). The cholesterol supply initially increased dehydrocholesterols in the three patients (B, D, F) with cholesterol below 0.31 mmol L−1 (Fig. 2) as well as lathosterol. These three patients also had the lowest level of sterols. Two of them (D, F) later showed decreased dehydrocholesterols; data from the third is missing because of his death. Of the remaining patients three (A, E, G) showed decreased lathosterol and dehydrocholesterol levels from the beginning and in one patient (C) decrease of lathosterol but unchanged level of dehydrocholesterol.
Table 2. Effect of cholesterol treatment on cholesterol/sterol ratio *
Normal cholesterol/sterol ratio > 0.98.
Figure 3a and b show the sterol changes, including lathosterol, during treatment in one patient (F) with initially low cholesterol compared with one (E) who started on a higher level. The one with the lower initial cholesterol level was given egg yolks. After the addition of taurocholate, his fraction of cholesterol increased from 0.71 to 0.81.
Patient A, treated for 6 years, had a progressive course with deterioration, precocious puberty and polyneuropathy which was stopped during the treatment. She regained some abilities and her nerve conduction velocities improved . Her cholesterol synthesis was initially down-regulated according to decreasing lathosterol and dehydrocholesterols but had later increased and was higher than before treatment. The cholesterol/sterol ratio was gradually falling. The number of infections have increased during the last few years and her liver enzymes were slowly rising. Her polyneuropathy, which has been stationary for the last few years, showed signs of progression at the last neurography.
In the other two patients (E, G) with milder polyneuropathy and stationary decrease of nerve conduction velocities, these did not change substantially during treatment.
Some weight gain occurred in five of seven patients, with an average of 0.7 SD (Table 1).
All seven patients had increased photosensitivity which diminished during treatment. In patient E, who was tested with UVA light, the tolerance increased from 3 to 10 joule cm−2.
The patients' alertness seemed to increase during the treatment, but this as well as progress in development were difficult to evaluate by objective methods. Decreasing susceptibility to infections were observed in most children during the treatment period.
Five patients initially had values for vitamin A moderately below the normal range, and two (A, E) had a slight increase of liver enzymes. These changes were normalized during treatment. No side-effects, clinical or biochemical, related to the treatment were seen.
General effects of the treatment on biochemical and clinical parameters
In agreement with previous treatment trials [12,14], all patients showed increased serum cholesterol levels during dietary supplementation with cholesterol but this was not correlated to clinical improvement. The three with the best response (E, F, G) were all orally fed and their cholesterol doses were increased faster, although the doses per kilo for all patients, except for B and G, were about the same for the last 3 years. Compliance, more infections or biochemical variations can be factors influencing the levels seen in the other four.
In the first published trials, bile acids were usually added to the dietary cholesterol, but they are not included in some of the ongoing protocols. It has been proposed that exogenous bile acids may block the LDL- receptor and cause a rise in serum dehydrocholesterols and cholesterol  but the rise in dehydrocholesterols observed in our study seems mainly be an effect of the increased synthesis. Chenodeoxycholic acid, used in some of the first assays, is known to be hepatotoxic. We have not seen any side-effects of taurocholic acid and in patient F, who received only cholesterol for 1.5 years, the ratio of serum cholesterol to sterols increased when taurocholate was added.
Lathosterol is a suitable marker of cholesterol synthesis  and a precursor of the dehydrocholestrols. In general, the changes in lathosterol were paralleled by similar changes in 7- and 8-DHC. With exogenous supply of cholesterol, the patients with very low cholesterol levels (B, D, F) initially increased their synthesis and it was not down-regulated until after 2 years of treatment. In the other patients, the treatment reduced the rate of synthesis from the beginning. Varying effects on 7-DHC levels have been reported from other studies [15, 19, 20]. Many enzymes involved in cholesterol synthesis, including the rate-limiting 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) reductase, are bound to cholesterol-containing membranes and it seems likely that a certain minimum concentration of cholesterol is needed for optimal activity at the specific site of synthesis.
The normal response to cholesterol deficiency is an increased cholesterol synthesis, but this does not seem to be the case in the SLOS. The mean lathosterol level was within in the range of control patients, even if the patients A, B and G initially had slightly elevated lathosterol levels. According to Honda and coworker, HMG-CoA reductase was not elevated in liver biopsies from SLO patients in spite of severe cholesterol deficiency  and Steiner et al. have reported that the total sterol synthesis is reduced in the SLOS . Recently Fitsky et al. using a knockout mouse model, showed that accumulated 7-DHC suppresses sterol biosynthesis post-translationally by accelerating the proteolysis of (HMG)-CoA reductase .
The severity of the SLO syndrome has mostly been correlated to the degree of hypocholesterolemia [16,24, 25]. Some groups have observed correlations between 7-DHC, severity and mutation classes . However, in a report of 22 mutations found in 32 patients (including patients A, B, F and G from the present study) no such correlation was found . Patients A and G are unrelated but have identical mutations. Both had polyneuropathy but G is orally fed and had better motor function and no signs of deterioration. Both had some up-regulation of the cholesterol synthesis before the treatment as judged by lathosterol but serum cholesterol and the response to exogenous cholesterol supply were different.
Weight gain has been used as an efficacy parameter. In our study, however, six patients gained on average 2.75 SD before treatment started and it therefore seems to be part of the natural course in these patients. As growth velocities for height and weight are retarded during the first years but often normalize later , growth is not a reliable parameter for evaluation of treatment.
Behavioural improvement has frequently been observed in other treatment studies. Tierney et al.  have reported that an autistic trait appeared to be more common if cholesterol supplementation was started after the age of 5. However, autism is not uncommon amongst mentally retarded subjects and may vary with time. So far, behavioural improvement has not been generally accepted to confirm treatment efficacy.
A specific intolerance of UVA light is seen in 60–70% in the SLOS [28, 29] and has been suggested to be caused by cholesterol deficiency in the epidermal cell membranes rather than by accumulation in the skin of 7-DHC, which is a precursor of vitamin D . As in previous reports [4, 12, 14], we found reduced photosensitivity to be the most constant beneficial effect of the treatment.
Polyneuropathy is seen in some SLO patients and it has been suggested that 7-DHC is a less effective substrate for myelin synthesis than cholesterol . This may have influenced the improvement of the patient A with progressive polyneuropathy, reported earlier , during dietary cholesterol supply.
The hyperplasia of the adrenals in patient A disappeared during dietary supplementation. This may have been because of reduced availability of substrate cholesterol for adrenal synthesis of steroid hormones in the untreated state. Most patients with SLOS have been reported to have normal corticosteroids and response to stimulation, although defective adrenal synthesis has been observed in some cases . It may be speculated that such a defect may have contributed to the death of patient B.
Is long-term treatment of SLO patients with cholesterol justified?
It is not yet known if the low levels of cholesterol or accumulation of dehydrocholesterols is most harmful in SLOS or if it differs with age. The natural history is variable but some symptoms, like feeding difficulties, growth deficiency, sleep disturbance and susceptibility to infections, usually decrease with time. The most obvious limitation of the treatment with cholesterol is that very little change can be expected in the sterol composition in the brain, because of the efficient blood–brain barrier. From this point of view, improvement of psychomotor function is doubtful.
The dehydrocholesterols have not been completely eliminated in this or any other reported treatment trial. Their toxicity is still largely unknown. How the increase in plasma cholesterol from exogenous supply affects synthetic pathways or intracellular metabolism is not known. The positive effects of treatment seem at least in some cases to be temporary . One of our patients (A) is now again deteriorating after 6 years on dietary cholesterol.
In one of our patients (E) the only obvious clinical motive for treatment would have been her photosensitivity. A period of cholesterol supply also seems to be indicated in patients with polyneuropathy, malnutrition, a progressive course or serious behavioural problems.
Recently, a new treatment approach with simvastastin, an inhibitor of cholesterol synthesis, was described. This significantly decreased the dehydrocholesterols also in the cerebrospinal fluid in two patients . Whether long-term treatment with statins is beneficial to SLO patients is, however, very difficult to assess at the present state of knowledge.
In conclusion, we have shown that dietary cholesterol can up-regulate the endogenous cholesterol synthesis in severely affected patients with low levels of cholesterol and total sterols. Some specific features of SLO, such as photosensitivity and progressive polyneuropathy, seems to be treatable with cholesterol and the effects possible to verify by objective methods. As we still do not know enough about the consequences of the biochemical effects or efficacy of dietary cholesterol, except for a few specific symptoms, it is questionable if it is justified to treat all patients with the SLOS indefinitely.
This work was supported by a grant from Stiftelsen Samariten, Stockholm.