Early recognition of CLN3 disease facilitated by visual electrophysiology and multimodal imaging

Background Neuronal ceroid lipofuscinosis is a group of neurodegenerative disorders with varying visual dysfunction. CLN3 is a subtype which commonly presents with visual decline. Visual symptomatology can be indistinct making early diagnosis difficult. This study reports ocular biomarkers of CLN3 patients to assist clinicians in early diagnosis, disease monitoring, and future therapy. Methods Retrospective review of 5 confirmed CLN3 patients in our eye clinic. Best corrected visual acuity (BCVA), electroretinogram (ERG), ultra-widefield (UWF) fundus photography and fundus autofluorescence (FAF), and optical coherence tomography (OCT) studies were undertaken. Results Five unrelated children, 4 females and 1 male, with median age of 6.2 years (4.6–11.7) at first assessment were investigated at the clinic from 2016 to 2021. Four homozygous and one heterozygous pathogenic CLN3 variants were found. Best corrected visual acuities (BCVAs) ranged from 0.18 to 0.88 logMAR at first presentation. Electronegative ERGs were identified in all patients. Bull’s eye maculopathies found in all patients. Hyper-autofluorescence ring surrounding hypo-autofluorescence fovea on FAF was found. Foveal ellipsoid zone (EZ) disruptions were found in all patients with additional inner and outer retinal microcystic changes in one patient. Neurological problems noted included autism, anxiety, motor dyspraxia, behavioural issue, and psychomotor regression. Conclusions CLN3 patients presented at median age 6.2 years with visual decline. Early onset maculopathy with an electronegative ERG and variable cognitive and motor decline should prompt further investigations including neuropaediatric evaluation and genetic assessment for CLN3 disease. The structural parameters such as EZ and FAF will facilitate ocular monitoring.


Introduction
The neuronal ceroid lipofuscinoses (NCLs) are a group of autosomal recessive lysosomal storage disorders (LSD) and together are one of the most frequent causes of neurodegenerative disease in children. The incidence of NCL ranges from 0.1 to 8 per 100,000 live births [1][2][3][4][5][6][7]. Isolated retinal CLN3 disease accounted for 1% of all inherited retinal disease (IRD) in a French cohort [8]. There are a number of recent publications reporting isolated retinal findings in patients with CLN3 mutations [9][10][11]. Analysis of these reports suggests that it is more likely that the 1 kb homozygous deletion is associated with the syndromic CLN3 phenotype, while compound heterozygous mutations are more likely to be found in the isolated retinal degeneration phenotype. NCL patients experience myoclonic seizures, progressive visual deterioration, cognitive dysfunction, motor decline, and premature death [11][12][13]. These clinical features often present asynchronously, making diagnosis difficult and often delayed. Classically, NCL was classified based on age at onset (congenital, infantile, late infantile, juvenile, and adult). To date, 13 causative genes have been identified (CLN1 to 8 and CLN10 to 14) with CLN3 being the most prevalent cause [11,12,14].
CLN3 disease was formerly known as 'juvenile neuronal ceroid lipofuscinosis' (JNCL) and can initially present as with isolated visual symptoms or with progressive neurological dysfunction. Wang et al. reported that the CLN3 associated visual symptoms can exhibit rod-cone or cone-rod dystrophy (RCD or CRD) phenotype [15]. In that study, 5 patients from a total of 123 retinal degeneration patients had a CLN3 mutation with 4 RCD and 1 CRD phenotype [15]. Data from our previous study showed that all CLN3 patients in our study centre had an electronegative ERG, suggesting its importance in this particular diagnosis [16].
CLN3 is a lysosomal membrane protein involved with glycosylation and phosphorylation at several sites, with localization to synaptic compartments in neuronal cells. This localization might suggest a distinctive role of the CLN3 protein within neurons that makes the central nervous system (CNS) susceptible in this disease [17].
Understanding of the ophthalmological findings is crucial to early diagnosis of CLN3-related disease, as these commonly precede the development of neurological signs, with retinal examination using multimodal imaging frequently identifying bull's eye maculopathy, optic disc pallor, and/or bone spicule formation. These structural findings overlap with Stargardt disease or retinitis pigmentosa (RP) [18,19]. Where CLN3 disease is a differential diagnosis, it is critical that a full-field electroretinogram (ffERG) is performed, as this may demonstrate an electronegative ERG (b:a ratio ≤ 1 in dark adapted 3.0 or 12.0 ERG) [16,18,[20][21][22]. Other classical ocular features of CLN3 disease may then be elucidated on ophthalmic examination, alerting the clinician to the possibility of this disorder and the need for neurogenetic review.
Novel therapies for CLN3-related disease are currently emerging into clinical trials. Early diagnosis is therefore vital to increase the possibility of administering a novel CLN3 disease therapy at a time when maximal benefit might be achieved. Ocular biomarkers become challenging to obtain as neurological deterioration progress. The purpose of this study is to report ocular findings of CLN3 disease patients to aid early diagnosis, enable disease monitoring, and assist further trials of novel CLN3 therapies.

Methods
Retrospective evaluation of 5 confirmed CLN3 disease patients in our tertiary referral clinic were included in this study. They were referred for ophthalmic review and subsequently underwent single genetic testing for CLN3 disease. The age when the patients were referred and the age of ocular and neurological onset were recorded. Age of ocular and neurological onset was determined by the earliest time point of reported ocular and neurological symptoms. Best corrected visual acuity (BCVA), retinal imaging, spectral domain-optical coherence tomography (SD-OCT), and full-field electroretinogram (ffERG) data were reviewed at baseline (BL) and follow-up (FU). BCVA was measured using a logarithm of minimum angle of resolution (logMAR). Patients with BCVA worse than 1.0 logMAR (6/60 on Snellen) were examined using Sheridan-Gardiner single letter and if failed this continued to finger counting, hand movement, and perception of light. BCVA values were then converted to logMAR equivalent values as described by Lange et al. [23,24].
The study followed the tenets of the Declaration of Helsinki and was approved by the local ethical committee. Disease severity was calculated using the recently described Hamburg CLN3 Ophthalmic Rating Scale [25]. This scale consists of visual acuity, fundus, and OCT score with maximum points of 14. The scale then is translated into CLN3 grades of grade 0 (unaffected) = 14 points, grade 1 (affected) = 10-13 points, grade 2 (severely affected) = 5-9 points, and grade 3 (end stage) = 0-4 points.
Spectral domain-optical coherence tomography (SD-OCT) SD-OCT imaging was acquired using the Heidelberg Spectralis (Heidelberg Engineering, Germany) and Zeiss Cirrus (Carl Zeiss Meditec, Dublin, CA, USA). Retinal layers and central macular thickness were examined. Bruch membrane and internal limiting membrane markers were manually adjusted to ensure precision in measuring retinal thickness. Central subfield thickness (CST) and central macular thickness (CMT) are both commonly used terms in ophthalmology to describe the thickness of the central retina. Central subfield thickness (CST), also known as foveal thickness, was defined as the average thickness of the central 1 mm subfield centred at the fovea on ETDRS grid [26].

Electrophysiology
Testing strategies included pattern ERG (pERG) and full-field ERG (ffERG) using Espion (Diagnosys, Lowell, Massachusetts USA). Visual electrophysiology was performed according to International Society for Clinical Electrophysiology of Vision (ISCEV) standards [27][28][29]. Gold foil; Dawson, Trick, and Litzkow (DTL); or skin electrodes were used depending on the level of patient's cooperation. Paediatric non-standard abbreviated ERG protocol was done using a modified Great Ormond Street Hospital (GOSH) protocol utilizing handheld Grass (Gr) strobe for the most uncooperative patient [28,30]. Pulse period 2 (2/s) and flicker Gr intensity 1 (Gr1) were used instead of 3/s and Gr4, respectively. The b:a wave ratio was calculated from dark adapted (DA) ffERG 3.0 or 12.0 with a value of ≤ 1.0 defined as an electronegative ERG [16,31].

Results
Five unrelated children with biallelic CLN3 pathogenic variants were included in the study, 4 females and 1 male with median age at referral of 6.2 (4.6-11.7) years (yrs). Median age at ocular onset was 5.1 (2.6-11.6) yrs with P1 who had the earliest ocular onset while P5 the latest. Two patients (P1 and P2) had FU data. Hamburg CLN3 ophthalmic rating scale at BL was ranging from 9 (affected) to 13 (severely affected) ( Table 1). P1 progressed from affected to end stage while P2 from affected to severely affected.

Retinal imaging
Assessment of UWF-fundus pseudocolour appearance and UWF-FAF showed a consistent bull's eye macular appearance in all patients (Fig. 2). Additionally, P5 showed macular striae. Progress of yellow-orange macular appearance, retinal atrophy, and vessel rarefication can be observed in P1 and P2 (Fig. 3). The FAF pattern consisted of hyper-autofluorescence (hyperAF) rings surrounding a hypoautofluorescence (hypoAF) fovea (Fig. 2). This perifoveal hyperAF ring became more apparent at first FAF FU in P1 and P2. On the second FU, the ring of hyperAF had disappeared and hypoAF had developed outside the vascular arcade corresponding to the retinal atrophy seen on fundus image (Fig. 3).
Right and left eyes of the patients showed similar phenotype.
Spectral domain-optical coherence tomography (SD-OCT) Foveal ellipsoid zone (EZ) disruption was found in each patient. Those with the largest central EZ disruption had the poorest eventual BCVA (P1&2) ( Table 1). FU OCT was available in P1 and P2 using the Cirrus device and showed progression of EZ loss and signal hypertransmission into the choroid (Fig. 3). Macular IR appearances again showed a bull's eye maculopathy for P1-P5 and also macular striae for P5. CSTs were ranging from 99 to 145 μm (Table 1). In the compound heterozygous CLN3 patient (P5), we identified macular inner and outer retinal microcystic changes in addition to the macular atrophy (Fig. 2). There is a concordance in SD-OCT result of each patient.  The first one showed presence of rod function and undetectable cone response, while the second one showed worsening of rod function (Table 1) (Fig. 4).

Associated neurobehavioural issues
Neurobehavioural issues were documented in 4 patients (P1-4) at the time of presentation. The oldest patient (P5) did not have any systemic symptoms at presentation or the last ocular follow-up. Associated symptoms included autism spectrum features in P1, significant anxiety and speech delay in P2, behavioural issues in P3, and motor dyspraxia and behavioural issues (biting friends) in P4. In two patients (P2 and P4) the neurological abnormalities appeared after the visual symptoms, while P3 had onset before any eye complaint.

Discussion
CLN3-related disease commonly presents with early onset visual decline and variable neurodegeneration in childhood [12]. The visual decline in children with CLN3 disease is frequently more rapid than other early onset maculopathies such as Stargardt disease [32]. The CLN3 protein has a crucial role within neurons specifically in the synaptic space, with animal models of CLN3 disease showing this condition is primarily a disease of the inner retina, with secondary changes in the outer retina [17,20]. CLN3 has a role in the transfer of the palmitoyl-protein thioesterases-1 (Ppt1) protein, and deficiencies in this protein have been associated with inner nuclear layer damage, particularly cone bipolar cells, and further damaging the cone photoreceptor cells over the rod [33,34]. This pathophysiology assists in the understanding of the generation of the electronegative ERG, the one feature that was consistent across our cohort and similar to previous studies [18,21,22,32], reflecting the inner retinal defects. There was significant but variable reduction in both rod and cone responses as found in other studies [9,18,21,35]. The ffERG of P2 in 2 different time points showed early DA ERG preservation associated with an undetectable LA ERG, further reflecting initial cone involvement of this disease and thus resembling CRD [33,34]. In contrast, other studies in CLN3 studies in cases without neurological phenotype showed that DA ERG is more affected that LA ERG resembling RCD [8,9,32,35]. These contrasting phenotypes have electronegative ERG or at least reduced b:a wave ratio as the consistent common finding reflecting inner retina disturbance.
Batten disease is a rare paediatric degenerative disorder, and diagnosis may be delayed due to variable presenting features [18,21,40]. The application of electrophysiology combined with multimodal imaging in patients with reduced vision provides an opportunity of early recognition of this disease. The findings of an electronegative ERG and biomarkers of a bull's eye maculopathy facilitate directed genetic testing. The increasing availability of genetic testing will supplant the use of peripheral blood film microscopy (vacuolated lymphocytes) and electron microscopy (storage lysosomal inclusions) as previously proposed by other authors [18].
Ophthalmic follow-up is challenging for these patients due to poor cooperation as the degenerative disorder progresses. In our study two patients had reliable measurements to enable comparison with BL. In these two patients (4 eyes) the rate of change was a loss of 0.75 (0.41) logMAR letters/year during 3.9 (2) years of FU. It is a slower rate of deterioration with longer FU compared to Wright et al. study with 2.02 (3.78) logMAR letters/year during 0.9 (0.5) years FU [18]. These results provide further evidence to the variability in disease progression in this disorder. The latest-onset patient (P5) with no documented neurological findings had the best BCVA, while the early onset patients (P1&2) had the worst BCVA at FU. P5 was the only patient with a compound heterozygous mutation. These findings were in concordance to a previous non-syndromic CLN3 study that found absence of visual loss in the late onset patients and mild visual loss in their early onset patients [9]. Later onset of the disease appears to be correlated with better BCVA. A vast majority of CLN3 disease patients (± 80%) present with vision impairment [41,42]. A contribution to the visual decline has been postulated to arise from additional damage to the lateral geniculate nucleus and/or primary visual cortex [43].
Bull's eye maculopathy is the most consistent and prominent macular finding in this patient cohort as also found in previous studies [18,21,44]. Other fundus findings reported in CLN3 disease include optic disc pallor, macular atrophy, macular striae, macular oedema, retinal pigment epithelium (RPE) atrophy, RPE granularity, bone spicule formation, epiretinal membrane, arteriolar attenuation, and even a Coatslike reaction [9,18,40,45]. The fundus variability may lead to misdiagnosis of Stargardt disease or retinitis pigmentosa, demonstrating the importance of electrophysiology investigations.
UWF-FAF findings highlighted the central hypoautofluorescence (hypoAF) surrounded by a ring of hyperAF found in our patients. Through 2.2-3.5 years Fig. 4 Full-field ERGs and pattern ERG recordings. The fullfield ERGs were recorded according to ISCEV protocols for paediatric ERG and in one case the non-standard abbreviated ERG protocol was used. P1 and P2 were examined using skin electrodes, P3 used DTL electrodes, P4 underwent a paediatric non-standard abbreviated ERG protocol using skin electrode, and P5 used gold foil electrodes. P4 ERG was performed using a modified Great Ormond Street Hospital (GOSH) protocol as described in the methods. All patients showed severely reduced or undetectable DA 0.01 response. All patients excluding P4 showed a reduced b:a wave ratio (electronegative) for DA 3.0 and DA 12.0. P4 had a very noisy recordings, and there is a suggestion of a reduced b wave. The LA 30 Hz and LA 3.0 were significantly reduced for P1-P2. P4's responses were noisy but also appeared reduced. Patients P3 and P5 were at the lower of the normal range. In patients P1 to P4, the pERG 30 deg p50 amplitude was almost undetectable. Patient P5 showed an identifiable waveform, but the p50 amplitude was reduced. Despite P4 having poor compliance and cooperation during the testing which resulted in noisy recordings, the combination of the potential electronegative scotopic ERG and a significantly reduced LA 30 Hz and LA 3.0 raised the possibility of Batten disease as a potential diagnosis. DA Dark adapted, DTL Dawson, Trick, and Litzkow electrodes, ERG electroretinogram, ffERG full-field ERG, LA light adapted, pERG pattern ERG ◂ of UWF-FAF follow-up in P1&P2, we found that the perifoveal hyperAF ring as found in previous CLN3 study [18] became more apparent and eventually disappeared. Then hypoAF starts to emerge in the periphery corresponding to retinal atrophy [40]. As disease advances, the whole macular region shows generalised hypoAF [9,18,46]. Therefore, we suggest that this specific change in UWF-FAF can be used as biomarker to monitor natural disease progression. A ring of hyperAF is a common finding in rod-cone dystrophies where the ring divides healthy central retina and disturbed peripheral retina [47,48]. Our CLN3 cases initially show the reverse pattern with an abnormal central fovea region and preserved peripheral retina.
The disrupted foveal EZ on SD-OCTs (P1-P5) is consistent with a previous review [18] and supports the CRD phenotype reflected from ERG and UWF-FAF findings in our cohort. In contrast, CLN3 cases with RCD phenotype had the predictably preserved foveal EZ while disrupted in the parafovea [8,9]. In later stage, there is marked macular EZ disruption with difficulty identifying any remaining outer retinal structures and choroidal signal hypertransmission reflecting RPE disturbance [18,46,49,50]. Inner and outer retinal microcystic changes found in P5 were also found in previous reports of CLN1 and CLN3 patients, indicating the involvement of both retinal layers [8,9,51] The mechanism for retinal degeneration in CLN3 disease is yet to be understood [43]. The bull's eye maculopathy, early DA ERG preservation, pERG disturbance, and foveal EZ disruption in our study support the notion that this disease has centrifugal (central to peripheral) progression as also found by Preising et al. in their study [52]. This condition primarily affects the inner retina with secondary defects in outer retina, as suggested in a mouse model where there were significant bipolar cell survival and preserved retinal function after gene therapy [20].
Four of our patients (P1-4) had neurological problems co-existing with their ocular symptoms, while the oldest patient (P5) did not have any systemic symptoms at presentation or the last ocular follow-up. Reflected by our P5 case, electrophysiology is the primary investigations in the event of a bull's eye maculopathy in a child of this age. An electronegative ERG with bull's eye maculopathy should directly lead to investigation of a genetic referral even in the case without neurological symptoms. Neurological onset is variable and may occur before, after, or concurrent with visual decline. Various neurological signs and symptoms have been reported, including: dementia, seizures, speech delay, mood fluctuations, difficult behaviour, balance, or memory changes, cognitive decline, sleep disturbances, feeding difficulties, clumsiness, and poor concentration [13,18,41] with seizure as the most common [13]. CLN3 has a variable phenotype as illustrated by those presenting with mild or delayed neurological defects ranging from 3-to 18-year interval between ocular and neurological onset [53][54][55][56], or no systemic features [9,10,15]. Ocular and neurological phenotypic variability also is frequently reported in those with the same mutations [57][58][59]. Ocular phenotype variability includes RCD and CRD [15]. CLN3 literature implies that syndromic CLN3 disease (mostly homozygous variant) is characterized by CRD with childhood onset and rapid disease progression, while the isolated retinal degeneration case (mostly compound heterozygous variant) is rather a RCD with later onset and slower progression [32]. However, genotype-phenotype correlation in CLN3 disease is not perfect and caution should be given in establishing the diagnosis [8].
Although there is no current definitive treatment for CLN3 disease, early diagnosis is important to give appropriate family counselling and establish supportive therapies at the earliest opportunity [20,[60][61][62][63][64][65][66][67][68][69][70][71][72][73][74][75][76]. In Australia there is Mackenzie's mission a study investigating preconception for autosomal recessive disorder. CLN3 is one of the gene of 500 genes in the panel for both parents. Secondly, whole genome screening is being investigated as an expansion of the newborn screening programme to identify and enable early management of severe genetic diseases [77,78]. There are currently 3 active CLN3 clinical trials which have ophthalmic parameter measurement as an endpoint. These include intrathecal gene therapy AT-GTX-502 (NCT03770572), oral drug PLX-200/ gemfibrozil (NCT04637282), and oral drug Miglustat 100 mg (NCT05174039) which give hope that disease-modifying therapies are emerging [79]. Those studies emphasize the importance of understanding the ocular biomarkers in CLN3 disease natural history. Multimodal imaging results are similar between the two eyes in each patient, making it viable to use fellow eye as control in the event of intraocular therapeutical trials. Combination of therapies might be needed to treat this condition [73,80,81].
Given the retrospective nature of our study and the natural history of neurodegenerative decline in CLN3 patients, there were limitations of follow-up examinations.

Conclusions
The findings of an electronegative ERG with concurrent bull's eye maculopathy in young age should prompt early neurological assessment for signs of neurodegeneration and referral for genomic investigation for CLN3 gene defects. Some children also experience isolated ocular presentations without neurobehavioral features. It is important that CLN3 disease is considered in electronegative ERG-bull's eye maculopathy patients even without neurological defect. Recognition of these features will assist in establishing an early diagnosis enabling appropriate therapies, family planning, disease monitoring, and potential enrolment in clinical trials for novel therapies.
Monitoring visual function is challenging in this cohort as neurological deterioration progresses. Finding ocular biomarkers that can be consistently recorded in an outpatient setting is important for clinical trial outcome measures. Given their change throughout the natural history of the disease, EZ and FAF are the most promising structural parameters identified in our cohort.
Author contribution JRG and EEC. contributed to the study conception and design. Material preparation, data collection, and analysis were performed by DHS, EEC, CLF, TM. S, MMJ, NAR, AMJ and JRG. BMN and RVJ collected and analysed genetic data. The first draft of the manuscript was written by DHS, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding Open Access funding enabled and organized by CAUL and its Member Institutions. This study was partly funded by National Health and Medical Research Council (NHMRC) Grants APP1116360, APP1099165, APP1109056, and Ophthalmic Research Institute of Australia (ORIA). The funding organization had no role in the design or conduct of this research.

Declarations
Conflict of interest DH Sakti, None; EE Cornish, None; CL Fraser, None; BM Nash, None; TM Sandercoe, None; MM Jones, None; NA Rowe; AM Johnson, None; J R Grigg; and RV Jamieson are consultant to Novartis.

Ethical approval
The study was approved by the South-Eastern Sydney Local Health District Human Research Ethics Committee and certified that the study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments.
Informed consent Due to retrospective nature of this study, informed consent is not required. This study does not contain images that may identify a person. Retinal images included in this study are de-identified.

Statement of human rights
The human rights of participants was recognised through the ethical conduct of the retrospective study.

Statement on the welfare of animals This research did not involve animals.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.