Translating genetic and functional data into clinical practice: a series of 223 families with myotonia

Abstract

High-throughput DNA sequencing is increasingly employed to diagnose single gene neurological and neuromuscular disorders. Large volumes of data present new challenges in data interpretation and its useful translation into clinical and genetic counselling for families. Even when a plausible gene is identified with confidence, interpretation of the clinical significance and inheritance pattern of variants can be challenging. We report our approach to evaluating variants in the skeletal muscle chloride channel ClC-1 identified in 223 probands with myotonia congenita as an example of these challenges. Sequencing of CLCN1, the gene that encodes CLC-1, is central to the diagnosis of myotonia congenita. However, interpreting the pathogenicity and inheritance pattern of novel variants is notoriously difficult as both dominant and recessive mutations are reported throughout the channel sequence, ClC-1 structure-function is poorly understood and significant intra- and interfamilial variability in phenotype is reported. Heterologous expression systems to study functional consequences of CIC-1 variants are widely reported to aid the assessment of pathogenicity and inheritance pattern. However, heterogeneity of reported analyses does not allow for the systematic correlation of available functional and genetic data. We report the systematic evaluation of 95 CIC-1 variants in 223 probands, the largest reported patient cohort, in which we apply standardized functional analyses and correlate this with clinical assessment and inheritance pattern. Such correlation is important to determine whether functional data improves the accuracy of variant interpretation and likely mode of inheritance. Our data provide an evidence-based approach that functional characterization of ClC-1 variants improves clinical interpretation of their pathogenicity and inheritance pattern, and serve as reference for 34 previously unreported and 28 previously uncharacterized CLCN1 variants. In addition, we identify novel pathogenic mechanisms and find that variants that alter voltage dependence of activation cluster in the first half of the transmembrane domains and variants that yield no currents cluster in the second half of the transmembrane domain. None of the variants in the intracellular domains were associated with dominant functional features or dominant inheritance pattern of myotonia congenita. Our data help provide an initial estimate of the anticipated inheritance pattern based on the location of a novel variant and shows that systematic functional characterization can significantly refine the assessment of risk of an associated inheritance pattern and consequently the clinical and genetic counselling.

Keywords: CLCN1; ClC-1; chloride channel; myotonia congenita; skeletal muscle channelopathy.

Figure 1

Overview of functional properties of ClC-1 variants. (A) Representative traces of wild-type channel, a variant with wild-type-like functional features (M646T) and a variant with shifted voltage dependence of activation (S289G). From holding voltage of −80 mV, 250 ms pre-pulse step to +60 mV, 250 ms test voltage steps from −150 mV to +190 mV in 10-mV increments (only traces in response to pulses up to +60 mV are shown) and a tail voltage step to −100 mV were applied. Scale bars: = 250 ms (x), 3 µA (y). (B) Voltage dependence of activation of wild-type (black), M646T (blue) and S289G (yellow) channels. Current at the beginning of tail voltage step was normalized to peak amplitude of the Boltzmann fit for each cell and the mean ± SEM normalized current data are shown. Solid lines show fit of the Boltzmann equation to mean data. (C) Mean V_1/2 ± SEM is shown for each variant. If the V_1/2 of the variant was depolarized relative to the cut-off voltage (wild-type ± 1.5× SD red dashed vertical line) the variant was considered pathogenic. If the V_1/2 was to the left of the cut-off it was considered wild-type-like. (D) Representative current traces of a variant (A221E) with minimal ClC-1 currents. Scale bars as in A. (E) Mean ± SEM tail current amplitude of variants with wild-type-like voltage dependence. Red bars show SD of wild-type current amplitude data. While most variants showed wild-type-like current amplitude, for five variants many cells did not express currents and when the currents were detectable the mean amplitude was outside the limits of wild-type ± SD. These variants were considered pathogenic due to reduced expression. Numbers are shown in Supplementary Table 4.

Figure 2

Functional properties of variants where the dysfunction could not be described by analysis of V_1/2 or current amplitude only. (A) Representative current traces of wild-type, L332R and P342L channels. Voltage protocol was as in Fig. 1A except that the holding voltage was −40 mV (V_h = −40 mV) (top row) or as in Fig. 1A but the pre-pulse step was to −140 mV (V_PP = −140 mV) (bottom row). Scale bars = 250 ms (x), 5 µA (y). (B) Representative current traces of R421C and M485K channels. Voltage protocol was as in Fig. 1A except that the pre-pulse step was to −140 mV. (A and B) Traces in response to test pulses up to +60 mV are shown. Scale bars = 250 ms (x), 1 µA (R421C) or 3 µA (M485K) (y). (C) Data for wild-type channels are shown in response to test voltages (grey) or to tail voltage (black) using the V_h = −40 mV protocol (solid symbols) and V_PP = −140 mV protocol (open symbols). Data for each cell are normalized to current in response to (grey) or immediately after (black) a voltage step to +100 mV. Only a subset of wild-type cells was analysed (n = 18). (D and E) Data for L332R (red) and P342L (blue) are shown for V_h = −40 mV (solid) and V_PP = −140 mV (open) protocols. Data are normalized as in C. L332R n = 13 and P342L n = 11. Wild-type data as in C are shown in black. (F) Mean current at the end of the test pulse is plotted against the test voltage for R421C (green) (n = 19) and M485K (n = 9) channels. Wild-type data for cells in C are shown in grey.

Figure 3

Properties of simulated heterozygous mutant channels. (A) Representative current traces of wild-type channels and mutant channels that displayed no currents in the homomeric condition. In simulated heterozygous condition these either showed currents with wild-type-like (A493Ehet) or shifted (P480Hhet) voltage dependence of activation. Inserts show a zoom in on the first 90 ms of the tail current to illustrate two distinct components of activation for P480Hhet channels. Scale bars = 250 ms (x), 3 µA (y). (B) Voltage dependence of activation of wild-type, A493Ehet and P480Hhet channels. The data were normalized as in Fig. 1B except for P480Hhet where tail currents in individual cells were normalized to peak amplitude of a double Boltzmann equation. (C) Representative current traces of a variant with shifted voltage dependence of activation in homomeric condition (S289G) co-expressed with wild-type subunits. Scale bars = 250 ms (x), 3 µA (y). (D) Voltage dependence of activation of wild-type (black), S289G (solid red) and S289Ghet (open red) channels. (E) V_1/2 of simulated heterozygous channels. Mean wild-type data ± 1.5× SD cut-off voltage is shown in red, data in homomeric conditions with solid symbols and in simulated heterozygous conditions with open symbols. For variants with no, or reduced currents, in homomeric conditions only data in simulated heterozygous condition are shown. Variants with V_1/2 left of the cut-off voltage were considered recessive, while variants right of the cut-off voltage were considered dominant. Data for variants for which two-component Boltzmann equation was used to describe the voltage dependence are not included. (F) Current amplitude of simulated heterozygous variants that in homomeric condition showed no or reduced currents and in heterozygous conditions showed wild-type-like voltage dependence. None of the variants suppressed current amplitude to below 50% of wild-type current amplitude, consistent with an absence of dominant negative effects on wild-type subunit function. (G) Voltage dependence of activation of L332Rhet (orange) and P342Lhet (blue) channels using the V_h = −40 mV (solid symbols) or V_PP = −140 mV (open symbols) protocols. While the V_1/2 was wild-type-like using the V_h = −40 mV protocol, this was shifted when using the V_PP = −140 mV protocol [V_1/2(wild-type) = −4.4 ± 0.4 mV, n = 271, V_1/2(L332Rhet) = 27.0 ± 4.1 mV, n = 8, V_1/2(P342L) = 28.4 ± 4.9 mV, n = 6]. (H) Voltage dependence of activation of R421Chet (green) and M485Khet (yellow) channels. (I) Voltage dependence of currents at the end of the test voltage pulse for R421Chet (green) and M485Khet (yellow) measured using the V_PP = −140 mV protocol. Data were normalized to peak tail current amplitude of the same cell measured using the standard protocol. The normalized current data indicate increased currents at hyperpolarized voltages for the simulated heterozygous mutant channels compared to wild-type channels.

Figure 4

Correlation of the functional properties with the inheritance pattern of clinical symptoms. Bar graph of inheritance patterns of all pedigrees for missense variants with wild-type-like (WT-like), recessive and dominant functional features. Dominant inheritance pattern is indicated in red, sporadic inheritance in yellow and recessive inheritance in blue. Variants with uncertain association with clinical symptoms are shown in grey. Variants with unknown inheritance pattern are excluded from the figure.

Figure 5

Alternative pathogenic mechanisms. (A) Representative current traces showing time course of activation of wild-type, A331S, F333L and L587V. Holding voltage was −80 mV, responses steps to voltages between −60 and +60 mV are shown. (A and D) Scale bars = 50 ms (x), 5 µA (y). (B and C) Time constant (B) and voltage dependence (C) of activation for wild-type (black), A331S (blue) and F333L (orange) channel. Solid symbols show data for V_h = −40 mV protocol, open symbols for V_PP = −140 mV protocol. (D) Representative current traces showing time course of deactivation of wild-type and L587V channels. Holding voltage was −80 mV, traces show the time course of closure following pre-pulse to +60 mV for voltage range +50 to −150 mV. (E and F). Time constants of activation and deactivation (E) and voltage dependence of activation (F) for wild-type (black) and L587V (pink) channels. Solid symbols show data for the V_h = −40 mV protocol, open symbols for the V_PP = −140 mV protocol.

Figure 6

Mapping of substituted residues (variants) to ClC-1 structure. (A) Two-dimensional map of variants. Top: ClC-1 intramembrane helices are represented as squares, CBS domains as ovals and the connecting loops as lines. Blue indicates TM1, pink TM2, black intracellular (IC). Helices B, I, J and R are indicated. Bottom: Location of all the variants in this cohort is shown on the top row and location of the variants with distinct functional features is specified below. Two rows of recessive variants are shown depending on if the variant in homomeric condition shifted voltage dependence [Recessive (shift)] or just reduced functional expression [Recessive (no shift)]. Outliers include the variants shown in Figs 2 and 5. (B) Percentage of variants located in intracellular domains (grey), TM1 (blue) or TM2 (pink) is plotted for variants with distinct functional features. (C) Percentage of variants with dominant (yellow), recessive (shift) (orange) or recessive (no shift) (red) and wild-type-like functional features (blue) are plotted for variants located in the intracellular domains (IC), TM1 or TM2. (D) Mapping variants to TM1s (left, light blue) and TM2 (right, light pink) to ClC-1 structure (6COY). TMs of both subunits are shown, variants are plotted on both subunits, Cl⁻ are shown in black. View is on membrane plane. Cl⁻ ions (black) on left and right graphs are aligned to illustrate location of TM2 higher up in membrane plane compared to TM1. Top row shows dominant variants in yellow and Recessive (shift) in orange. Bottom row shows Recessive (no shift) variants in red. Variants with wild-type-like functional features are shown in blue. (E) Top row shows a view from above membrane plane with the two TM1s (left, light blue) or TM2s (right, light pink) shown in a cartoon while the other TM is shown in ribbon. On the bottom row all variants are mapped based on their functional group as in B. Most of subunit interface is formed by the two TM1s (top) and most variants that shift voltage dependence of activation at any condition are located here. (F) Variants with attenuated activation particularly following hyperpolarized pre-pulse (L332 and P342 in red, A331 and F333 in orange) are shown viewed from above the membrane plane. These variants are located on IJ-linker (main chain is shown in light green spheres) that forms an interface with the IJ-linker of the neighbouring subunit and reaches the proximity of variants that showed enhanced currents at hyperpolarized voltage [R421C (magenta) and M485K (cyan)]. (G) Location of L587V (green) variant that accelerated both opening and closing of the channel at the intracellular opening of the selectivity filter pathway. The F297S (orange) variant that displayed a larger shift in voltage dependence of activation when co-expressed with wild-type subunits compared to homomeric F297S channels is also shown. Note that the variants mentioned in F or the L587V variant are not shown in B or C.