Natural history and biomarkers of KCNV2‐associated retinopathy

KCNV2‐associated retinopathy is an autosomal recessive inherited retinal disease classically named cone dystrophy with supernormal rod response (CDSRR). This study aims to identify the best biomarker for evaluating the condition.


| INTRODUCTION
KCNV2 (potassium voltage-gated channel modifier subfamily V member 2)-associated retinopathy is a particular type of cone dystrophy with a pathognomonic electroretinography (ERG) response including a supernormal rod response; hence it is frequently referred to as cone dystrophy with supernormal rod response (CDSRR). 1 The KCNV2 pathogenic variants are found in 2.2%-4.3% of cone dysfunction cases, higher than previously thought. 2 KCNV2 is a 2-exon gene encoding a 545 amino acid protein K v 8.2 (Figure 1), a voltage-gated potassium channel transmembrane protein expressed predominantly in the inner segment of photoreceptor cells.4][5] Voltage-gated potassium channels (K v channels) are encoded by 40 genes and are divided into 12 subfamilies, K v 1 to K v 12. 6,7 The K v 8.2 subunit is a member of a 'silent' K v S family that needs to co-assemble with K v 2 subunits to be functional. 7,8In human photoreceptors, K v 8.2 needs to interact with K v 2.1 or K v 2.2 to form functional K v heteromeric channels. 5In the human retina, K v 8.2 is situated with K v 2.1 and K v 2.2 in cone inner segments and with only K v 2.1 in rod inner segments. 5K v 8.2-K v 2.1 heteromeric binding is responsible for the assembling of I kx , 5,9 and it is hypothesised that K v 8.2-K v 2.2 also has a similar function (Figure 2).
To understand the impact of KCNV2 mutations, KCNV2's role in phototransduction and synaptic signalling needs to be understood.In healthy photoreceptors, the K v 8.2-K v 2 heteromeric channel is thought to be responsible for I kx , the 'dark current' outward flow which balances the inward 'dark current' that occurs through the cyclic nucleotide-gated (CNG) channels to maintain the depolarisation of the photoreceptor at a stable resting potential (near-40 mV) (Figure 2). 5,9,10In the dark, cyclic guanosine monophosphate (cGMP) binds to the CNG channels enabling the Ca 2+ and 'dark current' influx. 11Light stimulus via the phototransduction cascade activates phosphodiesterase reducing cytosolic cGMP, closing the CNG channels, and terminating the Ca 2+ influx.The cessation of the Ca 2+ influx and the continuous Ca 2+ extrusion by the NCKX (Na + /Ca 2+ , K + exchanger) channel results in the Ca 2+ concentration falling and hyperpolarisation of the photoreceptor. 11 I G U R E 1 Protein structure of K v 8.2 encoded by KCNV2 and our pathogenic variants location.P1 had compound heterozygous variants Ala259Thr and Ala261Asp( a ).P2 had compound heterozygous variants Gln109X and Arg320Cys( b ).P3 and P7 had a homozygous variant of Gly461X( c ). P8 had a homozygous variant of Trp188X( d ).P4, P5, and P6 had a homozygous whole gene deletion variant.P5 and P6 are siblings, while P3 and P7 are first cousins.Gln109X is located in the NAB domain, while Gly461X is located in the P loop domain which are both mutational hotspots. 3NAB facilitates interaction between compatible alpha subunits (K v 8.2-K v 2.1; K v 8.2-K v 2.2).NAB, N-terminal A and B box; P loop, pore forming loop; VSD, voltage sensor domain.
This hyperpolarisation also causes the voltage-gated calcium channel (Ca v 1.4) to close, leading to the termination of photoreceptor glutamate release (Figure 2). 12,13In the normal full-field electroretinogram (ffERG), this is recorded as the electrically negative a-wave.The cellular origins of the ffERG b-wave are a measure of the sum of the ON-bipolar and OFF-bipolar pathways. 14,15In the normal state, the ON bipolar cells have inhibitory glutamate receptors while OFF bipolar cells have excitatory glutamate receptors. 12Humans use both ON and OFF systems simultaneously to aid visual contrast. 16In the dark where there is constant photoreceptor glutamate release, ON bipolar cells do not release glutamate to the next neuron (sign inverting) while OFF bipolar cells release it (sign conserving).Conversely, in the event of light exposure and the cessation of glutamate release, the process is switched with ON bipolar cells releasing glutamate while OFF bipolar cells stop releasing it (Figure 2). 12,17he hypothesis that has been proposed for the KCNV2 ffERG findings is related to the inability of K v 8.2 to make heteromeric binding with K v 2.1 and K v 2.2.This inability disturbs the I kx channel which in turn depolarises the photoreceptor resting potential to the extent that the voltage-gated calcium current (Ca v 1.4, Figure 2) no longer varies linearly with fluctuations in light intensity.The result is that brighter stimuli are required to shut off transmitter release (glutamate).Hence, when this threshold is reached, a 'supernormal' postreceptoral response (b-wave) is displayed in the ffERG. 5,18n understanding of the KCNV2-associated retinopathy pathophysiology and how this impacts the ffERG provides an opportunity to identify specific electrophysiology biomarkers that could have the potential to monitor disease progression.With the development of two animal models for KCNV2-associated retinopathy, 19,20 the prospect of therapies is promising.Our study has examined KCNV2 patients' ffERG and ISCEV-extended protocol ERG responses 3,21 in detail and observed structural macula change to identify optimal biomarkers for monitoring.Additional natural history of structural and functional data is analysed to understand the patient's visual decline.

| METHODS
This study is a retrospective review of genetically confirmed KCNV2 associated retinopathy patients.The biomarkers used are: Best corrected visual acuity (BCVA), fundus imaging, spectral domain-optical coherence tomography (SD-OCT), visual electrophysiology, and visual field assessment were evaluated.

| Best corrected visual acuity
BCVA was assessed using a logarithm of the minimum angle of resolution (logMAR) at each visit.Patients with BCVA (logMAR) worse than 1.0 (Snellen equivalent of 6/60) were examined using Sheridan Gardener single letters.If this failed, then finger counting, hand movements, and perception of light were used and converted to logMAR equivalent values as described by Lange et al. 22 Baseline (BL) data and follow-up (FU) are compared.

| Fundus imaging
Ultra-wide-field (UWF) fundus pseudocolour images and UWF-fundus autofluorescence (UWF-FAF) were obtained using Optos 200TX and subsequently Optos California (Dunfermline, UK).FAF imaging was classified according to Georgiou et al. 23 2.3 | Spectral domain-optical coherence tomography SD-OCT was obtained using the Heidelberg Spectralis (Heidelberg Engineering, Heidelberg, Germany) and Zeiss Cirrus (Carl-Zeiss Meditec, Dublin, CA, USA).SD-OCT imaging was graded according to the previous study to quantify severity wherever possible. 23

| Visual field testing
This study used the GATE (German Adaptive Thresholding Estimation) threshold-estimating static visual field examination to assess visual field data.The GATE strategy algorithm enables threshold estimation with a shorter test duration than full threshold (FT) testing. 24he GATE strategy includes 109 test points covering almost 90 of the visual field 25 and is available as a research option on the Octopus 900 (Haag-Streit Diagnostics, Switzerland).To further assess the central vision, we added 10 additional static points to the central 20 .

| Electrophysiology
Full-field electroretinography (ffERG) and pattern ERG (pERG) were performed according to the International Society for Clinical Electrophysiology of Vision (ISCEV). 26,27We also used the ISCEV-extended protocol for the stimulus-response series for the dark-adapted full-field ERG b-wave. 21This stimulus series incorporates the protocols described in previous KCNV2 studies. 3,19his intensity series was performed to identify the impact on ffERG amplitude and a-wave and b-wave peak time.The change in amplitude was calculated by dividing the amplitude value in one stimulus with the amplitude of exactly one dimmer stimulus before and measured as 'fold'.The mean amplitude increases from all KCNV2 patients was compared with the mean from all unaffected subjects.Visual electrophysiology was performed using the Diagnosys LLC Espion device (Lowell, MA, USA).
The study adhered to the Declaration of Helsinki and was approved by the ethical board of the Sydney Southeast Health District committee (2022/ETH01713).

| Statistical analysis
Statistical analyses were performed in SPSS ® by IBM ® .The distribution normality of the data was analysed using Kolmogorov-Smirnov test.Non-parametric correlation and comparison were done with Spearman's rank test and Mann-Whitney U test respectively.Parametric comparison was done with independent samples t test.

| RESULTS
We report eight patients (P1-P8) with confirmed KCNV2-associated retinopathy from seven families, including three cases that we have previously reported. 28,29P5 and P6 are siblings while P3 and P7 are first cousins.The average FU duration was 12.3 (SD 4.3) years.The summary of the patients' characteristics is displayed in Table 1.We found seven distinct allelic changes that have been published in previous studies: Ala259Thr, Ala261Asp, Gln109X, Arg320Cys, Gly461X, complete gene deletion, 30 and Trp188X (Figure 1). 31

| Best corrected visual acuity
The average BL BCVA was 0.64 (SD 0.26) logMAR with an average age at BL of 7.7 (SD 9.4) years.The average BCVA was 0.85 (SD 0.28) logMAR and the average patient age was 20 (SD 7.5) years at FU.The average BCVA decline rate was 0.02 (SD 0.04) logMAR per year (Figure 3).The Spearman's rank test measuring the age and BCVA found no statistically significant correlation (ρ = 0.051, p = 0.782).

| Spectral domain optical coherence tomography
In all patients, SD-OCT showed a consistent photoreceptor ellipsoid zone (EZ) absence or disruption in the fovea or perifovea region (Table 1, Figures 4 and 5).

| Visual field
Visual field (VF) results were available for six of the eight patients (Figure 6).All tested patients had some degree of reduced sensitivity on central VF.P2 had the worst VF with severe constriction and reduced sensitivity.P2, P3, and P7 had some area of absolute scotoma in the central 40 area.

| Electrophysiology
Under dark adapted conditions, the ffERG b-wave was observed to have an increasing amplitude in all patients with increasing light stimulus strength.Light adapted (LA) ffERG showed reduced amplitudes for LA 30 Hz and LA 3.0 stimuli in all patients (Figures 4 and 5).Previous ffERGs were available for P1-P4, P7, and P8 and were displayed as grey and blue waves (Figures 4 and 5).The DA ffERGs at younger ages showed smaller amplitudes in both a-and b-waves compared with the most recent follow-up ffERG's.The baseline LA ffERGs and the most recent follow-up were similar.The pERG was available for subjects P2, P3, P5, P7, and P8 and consistently showed macular dysfunction (Figures 4 and 5).
The ffERG parameters were compared to reference values (grey and light grey bar in Figure 7) and to three unaffected subjects (green solid and dashed lines in female; N3, 44 years female which specifically underwent the same ffERG protocol as the KCNV2 patients.In all patients, the a-wave amplitude was normal in all the intensity series but mainly in the lower range of reference values (Figure 7A).The a-wave peak time was consistently delayed through all intensities compared to unaffected subjects and reference values in all patients (Figure 7B).
The DA b-wave amplitude increased disproportionately with increasing stimulus strength (log cd.s/m 2 ) (Figure 7C), most marked from DA 0.002 to 0.01 (À2.7 to À2.0 log cd.s/m 2 ).A statistically significant difference of this increase was found between the KCNV2 patients versus the unaffected subjects.The mean [standard deviation (SD)] and median (range) were 10.2 (9.5) and 6.4 (2-37.2)-fold in the KCNV2 patients versus 1.5 (0.1) and 1.5 (1.2-1.6)-fold in the unaffected subjects (p = <0.001)(Table 2).The amplitude changes became more similar between patients and the unaffected subjects with brighter light intensity stimuli (Table 2).For the patients P3 and P6, the b-wave amplitude increased above the reference values in response to DA 3.0 and 12.0 (0.5 and 1.1 log cd.s/m 2 ) (Figure 7C).The b-wave peak time was markedly delayed for dimmer intensity stimuli of DA 0.002, DA 0.01, and DA 0.14 (À2.7,À2.0, and À0.9 log cd.s/m 2 ) (Figure 7D) compared with both reference values and unaffected subjects.However, as the stimulus intensity increased through DA 3.0, DA 12.0, and DA 30 (0.5, 1.1, and 1.5 log cd.s/m 2 ), the b-wave peak time delay lessened and was found at the upper range of reference values but still above the unaffected subjects (Figure 7D).Interestingly, in the two oldest patients (P7 and P8), the b-wave peak time for these three brighter intensities remained delayed compared with the reference values and the unaffected subjects (Figure 7D).
The b:a wave ratio was above or just at the upper limit of the reference values with a trend of declining ratio from dimmer to brighter stimuli intensity (Table 1 and Figure 7E).
Further analysis of ffERG parameter change through the increasing age were available in six patients (P1-P4, P7, and P8) (Supplement S1).The a-and b-wave amplitudes were smaller in the younger age compared with the most recent examination.The b:a wave ratios were above normal or on the upper range of reference values without any observable age-related pattern.The a-wave peak times were consistently delayed without any observable age-related pattern.The b-wave peak times of the DA 0.01 were delayed or in the upper range of reference values while b-wave peak times of DA 3 and DA 12 were in the upper range of reference values for older age (>8 years) and in the lower range of reference values for younger age (<8 years) (Supplement S1).As the previous ffERG did not include the ISCEV-extended protocol, observation of the amplitude jump between DA 0.002 and DA 0.01 could not be completed.

| DISCUSSION
KCNV2 encodes Kv8.2 which consists of N-terminal A and B box (NAB) that mediates interactions between a compatible alpha subunit, six transmembrane domains (S1-S6) with a positively-charged voltage sensor domain (VSD) in S4, extracellular and intracellular loop segments, and the pore-forming loop between S5 and S6 (P loop) 3 (Figure 1).Most of the reported pathogenic variants in KCNV2 are located in the NAB and P loop domains, making these domains mutational hotspots. 3he pathogenic variants in these two locations either prevent K v 8.2-K v 2.1 heteromer formation (NAB) or impair F I G U R E 4 Electrophysiology and Multimodal Imaging of KCNV2-associated retinopathy patients 1-4.This figure shows the ffERG intensity series and ISCEV-extended protocol, pERG, fundus photograph, FAF, and SD-OCT of the right eye (P1-P4).Previous ERGs are shown with grey and blue wave lines.In general, DA ERG waveforms had lower amplitudes in the younger age.(A) ffERG of P1 showed an almost undetectable wave in DA 0.002.Delayed peak time and disproportionate amplitude increase of b-wave can be observed in DA 0.01, DA 0.14, DA 3.0, and DA12.0The LA 30 Hz and LA 3.0 was delayed and reduced.The photopic on-off showed delayed and reduced response.Blue flash on amber was subnormal.The pERG was unavailable in P1.The macular reflex was reduced on the fundus photograph.FAF showed bull's eye maculopathy, perimacular hyperAF ring, and increased foveal AF.SD-OCT showed EZ disruption or absence at the fovea.(B) ffERG in subject P2 was similar to P1.The pERG was reduced and noisy reflecting macular dysfunction.The macular reflex was reduced on the fundus photograph.FAF showed bull's eye maculopathy, perimacular hyperAF ring, and increased foveal AF.SD-OCT showed EZ absence at the fovea.(C) The ffERG of P3 was similar to P1. pERG was reduced and noisy, reflecting macular dysfunction.The macular reflex was reduced on the fundus photograph.FAF showed bull's eye maculopathy and increased foveal AF.SD-OCT showed EZ absence on the fovea.(D) The ffERG of P4 was similar to P1. Photopic on-off and pERG were unavailable.The macular reflex was reduced on the fundus photograph.FAF showed bull's eye maculopathy, perimacular hyperAF ring, and increased foveal AF.SD-OCT showed outer retina and RPE complex loss.(E) An unaffected subject is shown for comparison.DA, dark adapted; EZ, ellipsoid zone; FAF, fundus autofluorescence; ffERG, full field electroretinogram; LA, light adapted; pERG, pattern electroretinogram; SD-OCT, spectral domain optical coherence tomography.K v 8.2-K v 2.1 heteromer conduction (P loop) and have been compared in a functional study and found to have a similar effect and therefore a similar phenotype. 9The KCNV2 variant does not seem to be associated with a specific phenotype.Friedburg et al. study showed three siblings harbouring the same variant Gly461Arg, with only the middle sibling having a 'supernormal' wave. 32nother study comparing KCNV2 patients with no protein (NOP) group versus altered protein (ALP) group did not find any clear functional difference, although the NOP group had more advanced structural damage. 33he characteristic 'supernormal' rod response in KCNV2-associated retinopathy is still poorly understood.Understanding phototransduction physiology and how it relates to the electroretinogram response can assist in deciphering this phenomenon and identifying electrophysiology components that could be used as biomarkers.The KCNV2-associated retinopathy ffERG findings have been ascribed to the abnormal I kx channel (K v 8.2-K v 2.1/ K v 2.2) altering the photoreceptor resting potential (greater depolarisation) resulting in brighter stimuli being required to shut off transmitter release (glutamate) contributing to a supernormal post-receptoral response b-wave (Figure 2). 5 Another possible explanation is related to the production of the b-wave itself.The b-wave in the ffERG originates from both ON (producer) and OFF (limiter) bipolar cells (Figure 2B). 15The OFF bipolar pathway is only connected to cone photoreceptors, the main location of the KCNV2-associated retinopathy defect.The disturbance in the OFF bipolar pathway consequently creates a disturbed limiter in the b-wave and hence the 'supernormal' wave is formed.It is also possible that these two explanations coexist.
In our cohort, there is a suggestion that the younger patients had lower amplitudes and faster peak times (Supplement S1) which may reflect earlier disease state.Further, our KCNV2 patients had noticeably lower starting b-wave amplitude than unaffected subjects for the dimmest stimuli (DA 0.002) (Figure 7C).This contributes to the more significant increase in b-wave amplitude compared to unaffected individuals, especially between the dimmest stimuli (DA 0.002 and DA 0.01; Table 2 and Figure 7C).This may also indicate co-existing rod dysfunction. 3With relatively small increases in stimulus intensity, a disproportionately large increase in b-wave amplitude is formed (Figures 4 and 5 and 7C). 18This increased jump in b-wave amplitude reflects the described pathognomonic 'supernormal rod response'. 34e found a significant amplitude increase between the dimmest stimuli (DA 0.002 and DA 0.01) in the patients (Table 2).Estimating the amplitude change between the two stimuli (DA 0.002 and DA 0.01) can be utilised as a biomarker for KCNV2-associated retinopathy by quantifying the increase.Successful therapy would see a normalisation of this jump.
The delayed a-wave and b-wave peak times for the DA ffERG is an important biomarker.We showed similar to other studies that these delayed peak times should serve as a stimulus for KCNV2 genetic testing even if no supernormal DA ffERG responses have been recorded (Figure 7B,D). 34,35Our KCNV2-associated retinopathy patients had a delayed a-wave peak time compared with the reference values and unaffected subjects in all intensities (Figure 7B).Similarly, the overall b-wave peak time was distinctively prolonged compared with the unaffected subjects in all intensities except it went past the upper range of the reference values in DA 3 and DA 12 (Figure 7D).This contrasts with the b-wave amplitude, where there was some overlap with the reference range and unaffected individuals (Figure 7A).We suggest that this delayed peak time results from the longer time to reach the brighter stimulus needed to shut off the transmitter (glutamate) release. 5This peak time delay may serve as an important biomarker when evaluating natural history or in response to therapeutic intervention.Furthermore, the b:a wave ratio of the patients was above F I G U R E 5 Electrophysiology and Multimodal Imaging of KCNV2-associated retinopathy patients 5-8.This figure shows the ffERG intensity series and ISCEV-extended protocol, pERG, fundus photograph, FAF, and SD-OCT of the right eye (P5-8).Previous ERGs are shown with grey and blue wave line.In general DA ERG waves were smaller in the younger age.(A) ffERG of P5 showed an almost undetectable wave in DA 0.002.Delayed peak time and disproportionate amplitude increase of b-wave can be observed in DA 0.01, DA 0.14, DA 3.0, and DA12.0The LA 30 Hz and LA 3.0 was delayed and reduced.The photopic on-off showed delayed and reduced response.Blue flash on amber background was subnormal.The pERG was reduced and noisy, reflecting macular dysfunction.The macular reflex was reduced on the the fundus photograph.FAF showed an increased foveal AF.SD-OCT showed EZ disruption on the fovea.(B) The ffERG of P6 was similar to P5. Photopic on-off, blue flash on amber, and pERG were unavailable.The macular reflex was reduced on the fundus photograph.FAF showed bull's eye maculopathy, perimacular hyperAF ring, and increased foveal AF.SD-OCT showed EZ disruption on the fovea.(C) ffERG of P7 was similar to P5.The pERG was reduced and noisy, reflecting macular dysfunction.The macular reflex was reduced on the fundus photograph.FAF showed bull's eye maculopathy, perimacular hyperAF ring and increased foveal AF.SD-OCT showed EZ disruption on the fovea.(D) The ffERG of P8 was similar to P5.The pERG is reduced and noisy, reflecting macular dysfunction.The macular reflex was reduced on the fundus photograph.FAF showed bull's eye maculopathy, perimacular hyperAF ring and increased foveal AF.SD-OCT showed EZ disruption on the fovea.(E) An unaffected subject is shown for comparison.normal or at the upper limit of the reference interval (Figure 7E) and could also be used as an additional biomarker to monitor change in KCNV2-associated retinopathy.When assessing the biomarkers of the oldest patient (P8) and patient with the worst BCVA and VF (P2), the biomarkers continue to be distinct from the unaffected control (Supplement S1).This highlights the potential usefulness of these biomarkers in the more advanced cases.
Overall, we suggest three biomarkers for monitoring KCNV2 retinopathy natural history.First, the disproportionate b-wave amplitude jump between the two dimmest stimuli.Second, the a-wave and b-wave peak time delay.
Third, the relatively higher than normal b:a wave ratio.The normalisation of these three biomarkers could indicate therapeutic efficiency.Correlating the normalisation of these biomarkers with improvement in visual function will require additional correlation study with patient reported outcomes and mobility endpoints.
The data of this study and of Guimaraes et al. 3 suggests that in suspected KCNV2 cases, the ISCEV-extended protocol 21 should be incorporated into the testing strategy when possible.However, skin electrodes are frequently utilised in young paediatric patients and the b-wave response may not be detectable for the very dim flashes used in this protocol.In evaluating BCVA over time our cohort showed a decline of 0.02 (SD 0.04) logMAR per year over 12.3 (SD 4.3) years of FU.This number is higher than Georgiou et al. study with 0.01 logMAR/year over 8.4 years of FU (n = 75) 23 but lower than Vincent et al. with 0.1 log-MAR/year over 7.5 years of FU (n = 6). 36Legal blindness is projected to be reached before the age of 25 years (Figure 3), providing evidence for the best time for intervention.
Visual field analysis assessed both the central and peripheral fields using the GATE strategy.During testing, the gaze was monitored to manage the impact of poor central acuity on fixation.The GATE VF was performed under mesopic condition, which allows rod and cone responses to be interchangeable in visual processing.This assists in explaining why the peripheral VF remains essentially normal in several patients (Figure 6). 37,38Nasal and superior field defects (Figure 6) were contributed to by eyelid squeezing.We identified absolute central scotomas in three patients (P2, P3, and P7) similar to previous KCNV2 studies (Figure 6). 39,40Thiagalingam et al. could not document central scotomas in some of their patients due to nystagmus. 41he worst VF with severe constriction was found in P2, the second youngest with biallelic missense variants.P3-P8 with biallelic nonsense or whole gene deletions and older than P2 had better VF.These findings further re-enforce that there is no consistent relationship between visual function, age, or genetic variant in KCNV2-associated retinopathy. 9,33tructural assessment of our patients included FAF and SD-OCT.A perifoveal hyperAF ring was typical in our cohort similar to previous studies. 18,23   KCNV2 mutation. 18,42SD-OCT shows morphological defects localised in the fovea-perifovea region with disruption of the ellipsoid zone and outer retinal structures consistent with previous studies (Figures 4 and 5). 23,34owever, these FAF and SD-OCT findings are also seen in other IRDs such as achromatopsia and, therefore, not specific for KCNV2-associated retinopathy. 43The FAF and SD-OCT are markers of photoreceptor and retinal structural degeneration.The pathophysiology of photoreceptor degeneration in KCNV2-associated retinopathy is poorly understood.An inflammatory contribution is suggested from animal model studies using K v 8.2 knock-out (KO) mice which showed infiltration of natural killer (NK) cells and granulocytes in the retina. 19NK cells have cytolytic activity stimulating neutrophil infiltration and produce cytokines recruiting immune cells into the retina, while granulocytes generate various toxic reactive oxygen species and proteases which can damage photoreceptor cells. 44nother contributory mechanism to photoreceptor damage comes from the disturbance of I kx flow that may disrupt K + (potassium) homeostasis, affecting the Ca 2+ balance in the photoreceptor.The influx of Ca 2+ cannot be compensated by the K + -dependent outflux of Ca 2+ . 45he high intracellular calcium concentration will induce apoptosis and thus photoreceptor degeneration. 46,47On the contrary, dysfunction of NCKX (Na + /Ca 2+ , K + exchanger), which disrupts the extrusion of Ca 2+ from photoreceptors, only causes congenital stationary night blindness (CSNB) with no photoreceptor degeneration. 46,48These mechanisms contribute to our understanding of KCNV2-associated retinal degeneration.
In this study, the KCNV2 retinopathy patients showed peripheral visual function and structural stability over 8-15 years of follow-up.The electrophysiology data showed generalised rod and cone dysfunction.Interestingly the structural changes are concentrated at the macula in agreement with previous studies. 30,40actors that make the cone photoreceptors more vulnerable include: greater susceptibility to damage if they are not protected by surrounding rod photoreceptors, 20 the lack of Müller cells in the fovea which have a regulatory and buffering effect on extracellular K +33 and thirdly, K v 2.2 is expressed more in L/M cones (central fovea) than in the S cones (foveal slope). 5,49Hence, the disrupted K v 8.2 interacts with more K v 2.2 subunits in the central L/M cones in the central fovea.
KCNV2 retinopathy results from loss of function genetic variants.Gene replacement therapies have been developed and are being assessed in preclinical models giving hope to patients. 50,51s the electrophysiological 'supernormal rod response' is not always found in KCNV2 patients, a concurrent finding of DA ffERG peak time delay and localised EZ foveal-perifoveal disruption should raise the possibility of KCNV2 retinopathy.The visual acuity rate of change suggests that the therapeutic window is greatest in the first three decades of life.
In conclusion, we have identified three potential biomarkers that assist in differentiating affected from unaffected individuals: the disproportionate b-wave amplitude jump between the two dimmest dark adapted stimuli (DA 0.002-0.01),a-and b-wave peak time delay, and a higher than normal b:a wave ratio.Restoration of these biomarkers could assist in monitoring responses to future therapies.

2
Photoreceptor physiology.(A) Sodium (Na + ), potassium (K + ), and calcium (Ca 2+ ) exchange in the photoreceptor.KCNV2 encodes K v 8.2 which forms heteromeric bonds with K v 2.1 or K v 2.2 to create I kx .I kx is the outward flow to balance the inward 'dark current' of the CNG channel.(B) The ffERG wave development in the event of light exposure.Photoreceptors continuously release glutamate neurotransmitters in dark conditions, thus exciting the OFF bipolar to release glutamate.When stimulated by light, photoreceptors will stop releasing glutamate and in turn excite ON bipolar to release glutamate.The b-wave in ffERG is formed (push) by ON bipolar cells while it is reduced (pull) by OFF bipolar cells.In KCNV2-associated retinopathy, the dominant disruption of cone photoreceptors interferes with the OFF bipolar pathway and in turn causes an unrestrained b-wave.CNG, cyclic nucleotide-gated; ffERG, full field electroretinogram.

Figure 7 ) 3
, facilitating comparison to the ISCEV-extended protocol.Reference values were derived from unpublished data of normal subjects who underwent ffERG's in our centre.Grey bar reference values came from 30 normal subjects aged 7-20 years (4 male, 26 female).Light grey bar reference values came from 58 normal subjects aged 21-40 years (23 male, 35 female).The three unaffected subjects are N1, 35 years male; N2, 27 years First and last best corrected visual acuity (BCVA) of KCNV2-associated retinopathy patients.The BCVAs are plotted against the age.BL BCVAs are represented by hollow circles while FU BCVAs are presented by filled circles.Average BCVA is represented by the X symbol and shows declining BCVA (greater logMAR score) with a trend of reaching legal blindness (1 logMAR) before the age of 25 years.The Spearman's rank test measuring the age and BCVA found no statistically significant correlation (ρ = 0.051, p = 0.782).BCVA, best corrected visual acuity; BL, baseline; FU, follow up.F I G U R E 4 Legend on next page.SAKTI ET AL.
U R E 6 Visual field (VF) of KCNV2-associated retinopathy patients.Left eye VF is shown using the German Adaptive Threshold Estimation (GATE) test.Superior-nasal and superior-temporal constriction were common in all patients.P2 had the worst VF with severe constriction and reduced sensitivity.All patients had some degree of reduced sensitivity in the central VF.P2, P3, and P7 had an absolute central scotoma.VF, visual field.

F I G U R E 7
Full-field electroretinogram (FFERG) parameters plot.This figure shows the intensity of DA stimuli presented in log cd.s/m 2 plotted against electroretinography (ERG) parameters (amplitude, peak time or b:a ratio) of 16 eyes of eight patients.(A) The a-wave amplitude was at the lower end of the reference values.(B) The a-wave peak time was delayed in all patients.(C) The b-wave amplitudes were above reference values on P3 and P6 while on the upper limit of reference values in other patients.DA 0.002 reflected low amplitude and then increased disproportionately towards the brighter stimuli.(D) The b-wave peak time was delayed on the dimmer stimulus (DA 0.01) and then passed through the upper range of the reference values on the brighter stimuli (DA 3.0 and DA 12.0) except for the two oldest patients (P7 and P8) which were still delayed.The b-wave peak time was consistently prolonged in all intensities compared with the unaffected subjects (green line) except for left eye of P2 which fell to the unaffected subjects' level in DA 30.(E) The b:a wave ratio was above or at the upper end of the reference values with a decreasing trend from dimmer to brighter stimuli.The references values are indicated by a dark grey bar for the younger group (7-20 years) with a light grey bar for the older group (21-40 years).The vertical bars are superimposed on the graphs.The green solid (RE) and dashed (LE) lines represent data from three unaffected subjects to enable comparison with the ISCEV-extended protocol stimuli (DA 0.002, DA 0.
However, this ring is a non-specific sign manifested in various IRD types and thus electrophysiology remains indispensable for precise diagnosis and directing genetic testing for T A B L E 2 The b-wave amplitude changes between stimuli strength with increasing light intensity.
*Significant with Mann-Whitney U test.**Significant with independent samples t test.Abbreviations: DA, dark adapted; SD, standard deviation.