4-Chloropropofol enhances chloride currents in human hyperekplexic and artificial mutated glycine receptors

Background The mammalian neurological disorder hereditary hyperekplexia can be attributed to various mutations of strychnine sensitive glycine receptors. The clinical symptoms of “startle disease” predominantly occur in the newborn leading to convulsive hypertonia and an exaggerated startle response to unexpected mild stimuli. Amongst others, point mutations R271Q and R271L in the α1-subunit of strychnine sensitive glycine receptors show reduced glycine sensitivity and cause the clinical symptoms of hyperekplexia. Halogenation has been shown to be a crucial structural determinant for the potency of a phenolic compound to positively modulate glycine receptor function. The aim of this in vitro study was to characterize the effects of 4-chloropropofol (4-chloro-2,6-dimethylphenol) at four glycine receptor mutations. Methods Glycine receptor subunits were expressed in HEK 293 cells and experiments were performed using the whole-cell patch-clamp technique. Results 4-chloropropofol exerted a positive allosteric modulatory effect in a low sub-nanomolar concentration range at the wild type receptor (EC50 value of 0.08 ± 0.02 nM) and in a micromolar concentration range at the mutations (1.3 ± 0.6 μM, 0.1 ± 0.2 μM, 6.0 ± 2.3 μM and 55 ± 28 μM for R271Q, L, K and S267I, respectively). Conclusions 4-chloropropofol might be an effective compound for the activation of mutated glycine receptors in experimental models of startle disease.


Background
Hereditary hyperekplexia also known as 'startle disease' , 'Kok disease' or 'stiff baby syndrome' is a rare hereditary neurological disorder which is caused by mutations in genes encoding proteins involved in glycinergic neurotransmission, including the α 1 -subunit of the strychnine sensitive glycine receptor (GlyR) [1][2][3]. It predominantly manifests in the newborn with an extreme exaggerated hyperexcitability in terms of an abnormal startle response to sudden, unexpected acoustic, visual or somatosensory stimuli. Patients exhibit an intense tremor of arms and legs. Frequent falling attacks with episodes of convulsive hypertonia occur in adult patients [4,5]. In addition to mutations in the genes GLRB (encodes glycine receptor β-subunit), SLC6A5 (encodes glycine transporter 2) and GPHN (encodes the integral membrane protein gephyrin), mutations in the gene GLRA1 (encodes the α 1 -subunit of the GlyR) account for 40 -80% of hyperekplexia [6,7]. The most common mutations reported are R271L or R271Q [8].
Fast inhibitory postsynaptic transmission in the central nervous system (CNS) is mainly mediated by γ-aminobutyric acid A (GABA A ) receptors, whereas glycine receptors play a major role in the spinal cord, brainstem and retina [9]. The GlyR mutations α 1 R271Q-and α 1 R271L are commonly associated with clinically relevant symptoms in patients with autosomal dominant hyperekplexia, characterized by an impaired GlyRfunction due to reduced glycine sensitivity [7,[10][11][12]. The artificial mutation R271K shows the same startle GlyR features [13,14]. S267 mutations also display startle symptoms resulting from structural alterations of the GlyR. The corresponding mutated genes have been found in hyperekplexic patients and animals [15,16]. Patients suffering from startle disease are commonly treated with GABA A -activating drugs like clonazepam [12]. Clonazepam relieves the symptoms of hyperekplexia indirectly, but may be accompanied by sedative side effects [17].
As there is a compelling connection between startle disease and a distinct GlyR-malfunction, it would be of benefit to discover selective positive allosteric GlyRmodulators which might attenuate the hyperekplexic symptoms by restoring the function of the GlyR. Thus, it is of interest to modify the molecular structure of propofol in order to optimize all its various (aesthetic, sedative, anticonvulsant) activities or to yield drugs with more selective actions. The intravenous aesthetic propofol, well known for its positive allosteric modulatory effects at GABA A -receptors, has been shown to modulate glycine receptors in rat cortical and murine spinal neurons as well as in recombinant expression systems [18][19][20][21]. The effects exhibit a non-selective manner, i.e. the effects at GlyRs require higher concentrations than the effect at GABA A -receptors [19]. It has previously been shown that halogenation of a propofol analogue did not increase GABA-ergic activity [22,23]. A study of our group on heterologously expressed α 1 β glycine receptors found that 4-chloropropofol is almost 1000-fold more potent than propofol in enhancing glycine induced currents at wild-type (WT) glycine receptors [24]. The chemical structures of propofol and its analogue 4-chloropropofol are illustrated in Figure 1.
The aim of this study was to investigate whether 4-chloropropofol improves the function of glycine receptor mutations relevant for the generation of hyperekplexia. Consequently, we investigated the effects of 4-chloropropofol at WT glycine receptors and at the glycine receptor mutations α 1 R271Q-, α 1 R271L-, α 1 R271K and α 1 S267I.

Solutions
The phenol derivative 4-chloropropofol (2,6-dimethyl-4chlorophenol) was provided as pure substance by Prof. It has previously been shown that osmotic controls up to 500mM sucrose produced no currents [13,27]. 300 mM glycine adjusted to pH 7.4 by Na-OH revealed 1096mOsm. High glycine solutions of 300 mM glycine osmolarity subtracted from the osmolarity of the buffer solution itself, resulted in a total of Δ 800 mOsmol. We excluded possible osmotic effects of high concentrations of glycine (up to 300 mM in our experiments) by performing experiments with 1100mM sucrose. These experiments showed a lack of osmotic effects on HEK 293 cells transfected with WT glycine receptors [see Additional file 1]. Bath solution itself did not induce any current amplitude. Thus, any direct effects of 4-chloropropofol in the absence of glycine can be directly attributed to the applied substance.

Experimental set-up
Whole-cell experiments [28] were performed at a holding potential of −30 mV with a mean seal resistance of 1 GΩ. Chloride inward currents, due to agonist-induced channel activation, were resolved in the pA range. A fast liquid filament switch technique was used for the application of the agonist, presented in pulses of 2 s duration every 20 s. The liquid filament switch technique is able to exchange the solution passing an outside-out patch or small whole-cells within 1-2 ms [29,30]. We calculated flow rate of background solution through the chamber with 3,4 ml/min corresponding to 4% of maximal pump rate. With regard to the chamber volume of 100 μl the exchange solution time within the chamber was assessed to approximately 1.8 ms, dependent on cell size accordingly. Piezo-switch and small diameter capillary (ID 0.15 mm) within the measurement chamber promoted the millisecond solution exchange. The correct positioning of the cell, in respect to the liquid filament (ID 0.15 mm), was ensured by applying a saturating glycine pulse (1 mM for WT and S267I, 300 mM for R271Q and L, 10 mM for R271K) before and after each test experiment [see Additional file 2]. The induced current (I) by this saturating control solution was defined as I control .
Care was taken that the amplitude and shape of the glycine induced control currents had stabilized before proceeding with the experiment. The stability of the seal was controlled during the complete experiment via control of the seal resistance. A variation of 10-20% of the basic value was regarded as tolerable. Test solution and the saturating glycine solution were applied via the same glass-polytetrafluoroethylene perfusion system, but from separate reservoirs.
4-chloropropofol was applied either alone, in order to determine its direct agonistic effects at the GlyR mutations or in combination with a sub-saturating (EC 20 ) glycine concentration (20 μM for WT, 10 mM for R271Q, 30 mM for R271L, 100 μM for R271K and 30 μM for S267I), in order to determine its glycine modulatory effects. A new cell was used for each protocol and at least four different experiments were performed for each condition. The concentration of the diluent EtOH corresponding to the highest drug concentration used was 17150 μM. We have performed experiments demonstrating the lack of effect of ethanol on the potentiation of glycine induced currents in this concentration [see Additional file 3].

Current recording and analysis
For data acquisition we used an EPC 10 digitallycontrolled amplifier in combination with Patch Master Software (HEKA Electronics, Lambrecht, Germany). Currents were filtered at 2 kHz. Analysis was performed using Fit Master (HEKA Electronics, Lambrecht, Germany) and Graph Prism 5.0 software (GraphPad, La Jolla, USA). Fitting procedures were performed using a non-linear least-squares Marquardt-Levenberg algorithm.
The concentration-response-curves for receptor activation by the natural agonist and for positive allosteric modulation by 4-chloropropofol were fitted according to the Hill function (I norm = [1 + (EC 50 /[C]) nH ] -1 ). I norm is the current induced by the respective concentration [C] of 4-chloropropofol in the chloropropofol-glycine mixture. EC 50 is the concentration required to evoke a response amounting to 50% of their own maximal response and n H is the Hill coefficient. For 4-chloropropofol, the dose-response curves did not always reach a plateau response, because 4-chloropropofol in high concentrations leads to a decline in seal resistance and thus, did not yield reliable results. Therefore, further curve process could not be described by the Hill function. In these cases, the maximum response was the response at the highest concentration of the test compound for which a reliable response could be recorded.
KS normality test showed no normal distribution for the calculated EC 50 values; consequently two-tailed Mann-Whitney U test was performed to determine significance. All columns depicted in the diagram are means ± SEM and the levels of significance are indicated as *p < 0.05 and **p < 0.01.
Positive allosteric modulation of 4-chloropropofol was expressed as percentage of the current elicited by the sub-saturating glycine solution according to E (%) = 100 [(I-I 0 )/I 0 ], where I 0 is the current response to the subsaturating glycine solution. Currents were normalized to their own maximum response. The rise time τ (ms) of the initial glycine current traces (10-90% of the maximal amplitude I max ) were determined with a monoexponential fit (Fitmaster, HEKA Electronics) [31]. Data in tables and figures as well in the following results section are shown as mean values ± SEM.
A total of 90 cells was included in the study for the investigation of glycine sensitivity (n = 42) and modulating effects (n = 48) of mutant and WT glycine receptors. Cells that did not show a stable seal till the end of the experiment were excluded for EC 50 calculation. Thus, the cells included in the calculation of I max and rise time values of the peak amplitude may differ in the total number. Details, e.g. about the number of cells used for each test experiment, are provided in the appropriate figures and tables.

Results
In our study we were able to characterize glycine sensitivity and modulating effects (positive allosteric modulation and direct activation) of α 1 glycine receptors and its mutations R271Q, L, K and S267I by 4-chloropropofol.

Glycine sensitivity
Glycine induced larger inward currents in HEK293-cells expressing WT glycine receptors than the startle disease mutations α 1 R271Q, α 1 R271L and α 1 R271K, following application of a saturating concentration of the natural agonist. Consequently, the glycine sensitivity of the mutations to the natural agonist glycine was considerably reduced [see Additional file 4]. Glycine induced inward currents at S267I are only marginally reduced in amplitude compared to the WT. The current transient at the WT showed a biphasic time course with a fast increase, followed by a monophasic decay. These characteristics were considerably changed at the startle receptors, whereas R271K reflected to a greater extent the WT. S267I shows nearly similar rise time values compared to the WT. Concentration-response curves are depicted in Figure 2. EC 50 values and corresponding Hill coefficients (n H ), as well as rise time values and mean maximal current amplitudes are shown in Table 1.   marginal direct activation, while R271L and S267I were completely insensitive to the direct effects of 4chloropropofol ( Figure 5). The significant differences of calculated EC 50 values for positive allosteric modulation of 4-chloropropofol at the four startle mutations and the WT are shown in a mean-standard deviation-plot in Figure 6.

Discussion
This study provides evidence that the para-substituted propofol derivative 4-chloropropofol effectively modulates the WT GlyR and the α 1 R271Q, α 1 R271Lα 1 R271K-and S267I-GlyR startle mutations in vitro. 4-chloropropofol appears to be much more potent than propofol at the WT GlyR and at the mutation S267I [21].
These findings show a potential role for halogenated propofol derivatives in restoring the reduced function of mutated glycine receptors, which might alleviate the symptoms of the reduced glycinergic inhibition in experimental models of hyperekplexia. In a previous study we have shown that the halogenation of phenol derivatives is crucial for effective, positive allosteric modulation of the GlyR [32]. Thus, it was our aim to find a substance able to enhance the reduced glycinergic inhibition by pharmacological modulation of mutated receptors.

Positive allosteric modulation of WT and mutated GlyR with impaired glycine sensitivity by 4-chloropropofol
We were able to demonstrate the strongly reduced glycine sensitivity of the mutated receptors associated with lower maximal inducible current amplitudes and higher EC 50 values for glycine compared to the WT [8,13,27]. Hill slope values for alpha1 glycine receptors in HEK293 are known to show cooperativity and described to amount nH 2.4 for human [33]and nH 3.3 for rat alpha 1 GlyR [34]. The remarkably reduced nH values of our study (nH 0.8 for WT) may be due to differences in drug application speed and set up mode. The prolonged rise time of 3.7 ms for WT (vs. 0.4 ms on cell mode, 4 s drug application [35] also corroborates this hypothesis. Nevertheless, the elongated rise time values of hyperekplexic glycine receptors compared to the WT are appreciable. Our results clarify the importance of amino acid position R271 in the transmembrane domain 2 (TM2) of the α 1 subunit for the allosteric modulation of the GlyR [36]. R271 determines the channel properties for chloride binding and entry [11,12] and is thought to be involved in channel gating [37] and ion permeation [38]. The mutations at R271 lead to a reduced chloride conductivity at mutated startle receptors whereas the influence on assembly and oligomerisation of the GlyR is reported to be negligible [13]. Charge reversal mutations of positive charged amino acids in the TM2 of the α1 subunit have been published to reduce single channel conductance by 41% [39].
4-chloropropofol effectively potentiated glycineinduced currents at the mutations R271Q, R271L and R271K. Despite the strongly reduced glycine sensitivity  of the hyperekplexic receptors, modulation of receptor function was detected in the low-micromolar and submicromolar concentration range. Furthermore, our study shows that halogenation of the anesthetic propofol yields a compound with a highly increased potency for activation of chloride currents via WT and mutated α 1 glycine receptors. 4-chloropropofol shows a more than 1000fold increased potency to positively modulate glycine induced chloride currents at WT glycine receptors in comparison to the effect of propofol [21]. A recent study provided evidence that hyperekplexic R271L glycine receptors did not significantly change propofol binding and its strychnine cooperativity in comparison to the WT [40]. Our experiments revealed highly increased EC 50 values for modulation of the mutated receptors R271Q, R271L and R271K-GlyR by 4chloropropofol compared to the effect at the WT.

S267 seems to be crucial for 4-chloropropofol binding pocket
Point mutations in the α 1 subunit at the amino acid position Arg271, mainly R271Q-and R271L-GlyR, are mutations leading to clinical symptoms of the impaired GlyR function in startle disease [7]. The arginine (R) at position 271 of the α 1 GlyR is substituted by glutamine (Q) or leucine (L), respectively. The substitution of arginine to lysine (K) leads to the artificial mutation R271K, which also shows the typical startle GlyR features [13,14,27]. Mutation of the serine residue at position 267 is relevant for hyperekplexia as well and an exchange to isoleucine (S267I) diminishes binding of the anesthetic propofol [10,39]. S267 has previously been shown to be crucial for binding of propofol, ethanol and halogenated anesthetics [21,41,42]. Our results show, that S267 may be a part of the binding pocket cavity for 4-chloropropofol at the glycine receptor.
Compared to propofol [27], the increase in efficacy of 4-chloropropofol at hyperekplexic receptors in this study may be explained due to an interaction at R271 with the halogen (via a possible ion dipole interaction), additive to the already supposed binding site of propofol through the intact S267 (hydrogen bond). Further, more detailed structure-binding analysis will be necessary to substantiate these interactions. Evidence for the importance of  S267 mutations in generation of hyperekplexia comes from a study showing that α 1 S267Q-GlyR knock-in mice displayed a hyperekplexic phenotype and were shown to disrupt normal GlyR function [16]. Recently, it has been found in a hyperekplexic family that a S267N point mutation influences agonist responses and ethanol modulation of the mutated receptor [15]. The nature of the TM2 residue (267) of the glycine α 1 subunit influences the glycine modulatory effect of propofol and direct activation of the receptor by this anesthetic [21]. A comparison of the impact of such mutations on the interaction of 4-chloropropofol with glycine and GABA A receptors should permit a better understanding of the molecular determinants of action of 4-chloropropofol on these structurally related receptors and may aid the development of selective modulators of mutated startle disease receptors.
These in vitro findings are based on homomeric human α 1 -glycine receptors. Glycinergic-neurotransmission in adults is mainly based on heteromeric α 1 β-receptors [43]. The GlyR is a ligand gated ion-channel composed of five subunits [44], comprising 2α and 3β subunits [45]. Four different α isoforms exist (α1-α4) which show a developmental and regional dependant distribution [46]. Coexpression of the α subunit has an impact on the effect of various agonists (among them the natural agonist glycine), whereas its influence on the effects of strychnine seems to be negligible [45]. Other studies revealed that the expression of homomeric α 1 subunits in mammalian cells or Xenopus laevis oocytes is sufficient to generate functional receptors with pharmacological properties that are typical for the native GlyR in the spinal cord [47,48]. We have previously shown that co-expression of the glycine β subunit does not affect the response of heterologously expressed WT α 1 subunits to different halogenated phenol derivatives and propofol [21,32]. Additionally, studies with startle disease transgenic mice showed only a small level (25% of normal) of β subunits to be necessary for a proper GlyR-function [49].
Recessive GLRA1 mutations as being caused by nonsense or frameshift mutations in the α1 gene have been increasingly associated with case reports elucidating the molecular genetics of hyperekplexia [6] As autosomal dominant (AD) startle mutations of GLRA1 such as missense mutations R271Q and R271L define case reports with high frequency in hyperekplexic patients, it is important to mention that these AD cases represent only 5 of 30 index cases as recently published [6]. Cell surface expression and Imax of recessive GLRA1 mutations are markedly reduced compared to AD mutations. It would be an interesting attempt in functional analysis of 4chloropropofol to investigate the effects of 4-chloro-PRO at recessive startle mutations.
In vivo effects of 4-chloropropofol in experimental models of startle disease The impact of glycine receptor modulation by propofol on the clinical effects remains an interesting, yet unresolved issue. Evidence that an effect on glycine receptors may occur in vivo comes from other studies of the startle disease. The intravenous anesthetic propofol is known to activate glycine-induced currents in startle mutations in vitro and emerged as a transient treatment of startle symptoms in a transgenic mouse line carrying the R217Q mutation (tg271Q-300). Patch-clamp experiments at different startle mutations expressed in Xenopus laevis oocytes revealed that propofol in the concentration range of 1-500 μM increases the glycine induced maximal response. Transgenic mice (tg271Q-300) exhibited a decreased righting time and abatements in tremor without sedative side effects following injections of low doses of propofol (15 mg/kg i.p.) [27]. We have detected EC 50 values for modulation of the startle Figure 6 Mean-standard deviation plot of the calculated EC 50 values (mean ± SD) at either α1-(WT) or mutations -R271Q, -R271L, -R271K and -S267I expressed in HEK293 cells following application of 4-chloropropofol. P-values of Mann-Whitney U test result in significant differences between EC 50 of WT and the corresponding GlyR-mutation, indicated as significance levels of *p<0.05 and **p<0.01 (WT vs. R271Q, p=0.0159; WT vs. R271L as well as vs. R271K, p = 0.0095 and WT vs. S267I, p = 0.0286; n = 4-6 per group). receptors R271Q and R271L in a tenfold to hundredfold lower concentration range. Thus, it is conceivable that 4-chloropropofol might act in sub-anesthetic doses at startle disease glycine receptors.

Conclusions
In summary, our results broaden the body of knowledge about approaches to restore the reduced glycinergic inhibition in startle disease. In particular, 4-chloropropofol might lead to an effective enhancement of the function of startle disease glycine receptors in vitro. Based on our results we hypothesize that 4-chloropropofol might alleviate hyperekplexic symptoms in animal models of startle disease.

Additional files
Additional file 1: Osmolarity controls. Whole cell experiments at α1R271Qglycine receptors lack activation following 1100 mM glucose application. High glycine solutions of 300 mM glycine adjusted to pH 7.4 by Na-OH revealed 1096 mOsm. 300 mM glycine osmolarity subtracted from the osmolarity of the buffer solution itself, resulted in a total of Δ 800 mOsmol. Switch to 1100 mM glucose rather reduces baseline leak currents following 2 s application. In addition wild type glycine receptors didn't show sensitivity for osmolarity controls (600 mM sucrose). Thus osmolarity effects resulting from high glycine concentrations up to 300 mM at startle glycine receptor mutations can be excluded.
Additional file 2: Control of rundown effects of glycine receptors. Current amplitude of glycine control (10 mM) traces applied before and after application of 4-chloropropofol at α1R271K-mutation remains almost unchanged with negligible run down in liquid filament switch technique. Same picture for subsequent application of subsaturating glycine solution (10 μM) before and after co-application of 17.15 mM ethanol in custom-designed gravity driven perfusion system. There is almost no change in amplitude size. In any case the addition of ATP to the pipette solution should help to reduce receptor desensibilisation to exclude rundown effects for future studies.
Additional file 3: Ethanol control experiments. Ethanol (17.15 mM) in a sub-saturating glycine solution (10 μM) doesn't lead to an additional activation of wild type glycine receptors, by contrast the initial subsaturating glycine response is reduced when ethanol is added. Consequently the ethanol effect in particular at high 4-chloropropofol doses (where the concentration of the diluent EtOH corresponding to the highest drug concentration is 17.15 mM) has no influence on 4-chloropropofol effect.
Additional file 4: Glycine sensitivity at R271Q. α1R271Q-mutation lacks activation by low glycine concentrations (10-100 μM). Dose response curve starts with current peaks of less than 50 pA at 1 mM glycine. These current traces are depicted as example to illustrate repressed glycine sensitivity of mutated startle receptors.