METHODS

We studied 18 patients with tuberous sclerosis and a history of epilepsy who were referred to the department of neurology of Moscow University. As shown in table 1, of these 18 patients, six had been seizure-free for at least 2.5 years (patients 1 to 6), four were experiencing monthly seizures (patients 7 to 10), four had weekly seizures (patients 11 to 14), and four had daily seizures (patients 15 to 18). Clinical information on the patients is summarised in table 1. The diagnosis of tuberous sclerosis had been made using the new revised clinical diagnostic criteria.¹⁹

Six healthy subjects matched for age and sex referred to the clinic for a previous infective disease formed the control group (three male, three female; mean (SD) age 10.8 (10.8) years).

Exclusion criteria were as follows: neurological disorders other than tuberous sclerosis; endocrinological and dismetabolic disorders; and other pharmacological treatments in the previous three months.

The study was approved by the local ethics committee, and informed consent was obtained from the patients or their parents before investigations were carried out.

Heparinised blood samples were taken between 9 am and 10 am during the interictal period for plasma determination of 3α_s-THP, 3β_s-THP, and progesterone. In patients 7 to 18, carbamazepine and valproic acid were given 12 to 13 hours before blood sampling. In female subjects with menstrual cycles, the blood samples were taken during the follicular phase.

We used a GC-MS method for the determination of neuroactive steroids.^17,18 This profiles and quantifies very small amounts (fmol) of steroidal compounds with high accuracy and allowed us to identify the stereoisomers. Briefly, approximately 5000 dpm of [H³]-progesterone were added to the plasma to monitor recovery. Ethyl acetate was used for extraction (3 × 2 ml) and thin layer chromatography for separation (2 mm silica gel 60; solvent systems: carbon tetrachloride/methanol 99:1 (vol/vol), followed by a consecutive run in one direction using cyclohexane/ethyl acetate (3:2, vol/vol)). We then added 7 pmol of progesterone to the eluate containing 3α_s-THP or 3β_s-THP as an internal standard, and 10 pmol of 3α,α-THP to the eluate containing progesterone. These eluates were freeze dried and derivatised with heptafluorobutyric acid anhydride. Derivatised steroids were analysed using a Finnigan Trace GC/MS instrument equipped with a capillary column (HP-35MS: length 30 m, internal diameter 0.25 mm, film thickness 0.25 μm). The derivatised steroids were assayed in the negative ion chemical ionisation mode. The recovery of tritiated steroids ranged from 80% to 90%. The detection limit for the steroids studied was approximately 10 fmol.

Statistical analysis

Differences in steroid concentrations and in the ratio between 3α_s-THP and 3β_s-THP within the tuberous sclerosis groups of patients and the control subjects were tested for significance using a one factor analysis of variance (ANOVA). The Tukey–Kramer test was subsequently used for any steroid or 3α_s/3β_s-THP ratio that showed a significant effect at p < 0.05 on ANOVA. Statistical analysis was done using SPSS for Windows 10.8.

RESULTS

No significant age effect was detected by ANOVA between the groups of tuberous sclerosis patients and controls. No significant differences were found between the tuberous sclerosis groups for age of seizure onset and the number or localisation of cortical tubers.

In comparisons of steroid concentrations, ANOVA showed significant differences in the concentrations of 3α_s-THP and 3β_s-THP between seizure-free tuberous sclerosis patients, tuberous sclerosis patients with epilepsy, and the control group (fig 1), but not in the concentrations of progesterone (F = 2.36; df = 159.54; p = 0.08).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Histograms showing: (A) mean and SD (error bars) of the neuroactive steroids 3α_s-THP and 3β_s-THP; (B) Mean and SD of the ratio between 3α_s-THP and 3β_s-THP values, in control subjects and patients with tuberous sclerosis (TS) seizure-free, or with monthly, weekly, and daily seizure frequencies. Analysis of variance probability (p) values are shown above each group of neuroactive steroids and their ratio; asterisks and daggers indicate significant post hoc results. (A) Post hoc results showed that 3α_s-THP levels were decreased in TS patients with daily seizures compared with control subjects (p < 0.05) and seizure-free TS patients (p < 0.05). The 3β_s-THP concentrations were increased in TS patients with monthly, weekly, and daily seizures compared with control subjects (p < 0.0001, p < 0.0001, and p < 0.005, respectively) and seizure-free TS patients (p < 0.0001, p < 0.0001, and p < 0.005, respectively). (B) Post hoc results showed that the ratio between 3α_s-THP and 3β_s-THP values was decreased in TS patients with monthly, weekly, and daily seizures compared with control subjects (p < 0.0001, p < 0.0001, and p < 0.0001, respectively) and seizure-free TS patients (p < 0.0001, p < 0.0001, and p < 0.0001, respectively). *Post hoc significance compared with control subjects; †Post hoc significance compared with seizure-free TSC patients. 3α_s-THP, 3α, 5α-tetrahydroprogesterone and 3α, 5β-tetrahydroprogesterone; 3β_s-THP, 3β, 5α-tetrahydroprogesterone and 3β, 5β-tetrahydroprogesterone.

As shown in fig 1, plasma levels of 3β_s-THP were increased in all tuberous sclerosis patients with epilepsy, independent of seizure frequency, when compared with seizure-free patients and control subjects. The concentrations of 3α_s-THP were unchanged in patients with monthly seizures, were decreased (although not significantly) in patients with weekly seizures, and were significantly reduced in patients with daily seizures compared with controls and with seizure-free tuberous patients.

There was a highly significant decrease in the ratio 3α_s/3β_s-THP in all tuberous sclerosis patients with epilepsy compared with seizure-free patients and controls (fig 1).

DISCUSSION

The 3α_s-THP metabolites are potent positive allosteric modulators of GABA action at GABA_A receptors in the central nervous system, and may serve as endogenous anxiolytic and analgesic agents. In contrast, the 3β_s-THP are devoid of intrinsic activity at GABA_A receptors and behave as functional antagonists of the actions of 3α_s-THP in potentiating GABA_A receptor function.^8–11

Various investigators have shown that naturally occurring 3α reduced neuroactive steroids (such as 3α_s-THP) exert anticonvulsant effects.^12,13 To our knowledge, no similar investigations have examined the 3β reduced compounds, owing to the difficulty in detecting stereoisomers by simple radioimmunoassay methods. Thus this is the first clinical study on epilepsy that has evaluated the entire spectrum of THP enantiomers by means of a GC-MS technique, which is highly sensitive and specific.^18,19

The finding of unchanged THP enantiomer plasma levels in seizure-free tuberous sclerosis patients compared with controls may suggest that alterations in neuroactive steroids are not primary in tuberous sclerosis but represent an epiphenomenon related to seizures or their pharmacological treatment.

The reduced 3α_s-THP concentrations found only in patients with daily seizures, in which the interictal period is very short, may be caused by seizure induced stress. However, in a previous study by Galli et al it was reported that there was an increase in allopregnanolone (3α, 5α-THP) in the postictal phase.¹³ A possible explanation for this discrepancy is that the radioimmunoassay method used in that study did not discriminate between the different enantiomers and thus considered the whole pool of THP stereoisomers. Furthermore, they reported increased allopregnanolone levels up to 15 minutes after seizure occurrence, which is a very short time span compared with our paradigm, where the minimum interval was never less than one hour.

However, the major finding of our study was the decreased ratio between 3α_s and 3β_s-THP in all tuberous sclerosis patients with epilepsy (fig 1B). This ratio should be considered the most relevant index of GABA_A receptor function. A decrease in the modulatory effect of 3α_s-THP, caused mainly by the augmented 3β_s-THP levels in the brain, could result in a reduced sensitivity of GABA_A receptors to GABA, and in a great reduction in the sedative anticonvulsant action of several GABAmimetic drugs.¹⁶ Thus the high GABA levels found in brain biopsies of drug resistant tuberous sclerosis patients² could represent a compensatory mechanism for the downregulation of GABA_A receptor function resulting from a decreased 3α_s/3β_s-THP ratio.

These findings are consistent with the evidence of a relation between neuroactive steroids and seizure susceptibility.¹² A possible hypothesis to explain our data is that there might be an inadequate induction of the enzymes 5α/5β-reductases or the 3α/3β-hydroxysteroid oxidoreductases responsible for the biosynthesis of the 3α_s or 3β_s reduced derivatives of progesterone in tuberous sclerosis patients with seizures.^15,20

On the other hand, treatment with antiepileptic drugs could cause changes in the enzymes that metabolise 3α_s- or 3β_s-THP. It has been shown, for instance, that carbamazepine induces steroidogenesis in rat brain, increasing 3α, 5α-THP concentrations.²¹ However, we did not found significant changes in the 3α_s-THP levels except for a decrement in patients with daily seizures.

Although we need further analyses of samples from a larger population of patients with tuberous sclerosis before and after antiepileptic treatment, and also appropriate pharmacological studies in animal models, the decreased 3α_s/3β_s-THP ratio in the patients with epilepsy may reflect an altered GABA_A receptor mediated excitability which could facilitate the occurrence of seizures.

Conclusions

This preliminary evidence is significant not only for a better understanding of the pathophysiology of seizures but also in the clinical perspective in tuberous sclerosis, as it opens the field to the development of possible therapeutic agents. Recently, a 3α_s-THP analogue, ganaxolone, with anticonvulsant properties in various seizure models such as partial epilepsy, has been synthesised.²² Ganaxolone might also be considered as a potential therapeutic agent in patients with tuberous sclerosis who have epilepsy.