Increased choroidal thickness in adults with Down syndrome

Abstract Introduction People with Down syndrome (DS) are particularly susceptible to Alzheimer's disease (AD) due to the triplication of the amyloid precursor protein (APP) gene. In this cross‐sectional study, we hypothesized that choroidal thinning reported in sporadic AD (sAD) is mirrored in adults with DS. Methods The posterior pole of the eye for 24 adults with DS and 16 age‐matched controls (Ctrl) were imaged with optical coherence tomography. Choroidal thickness (ChT) was measured and analyzed in relation to cognitive status and cerebral amyloid beta (Aβ) load. Results ChT was increased in people with DS (pwDS) compared to Ctrl. This increase was associated with gender differences and positively correlated with cerebral Aβ load in a small subset. There was no significant correlation detected between ChT and age or cognitive status. Discussion In contrast to sAD this study found a significantly thicker choroid in pwDS. Whether these changes are related to Aβ pathology in DS needs further investigation.


INTRODUCTION
People with Down syndrome (pwDS) are known to be at increased risk of developing Alzheimer's disease (AD) due to the trisomy of chromosome 21, which results in the overproduction of amyloid beta (Aβ) protein and plaque formation in the brain. 1 Improved support has led to a two-fold increase in life expectancy for pwDS resulting in AD becoming an increasing concern. 2 Because the DS population represents the largest cohort of genetic AD, with about 6 million people worldwide, understanding early AD biomarkers in DS is crucial for clinical studies for DS and sAD. 3,4 An increasing number of studies investigating the brain, cerebrospinal fluid (CSF) and blood biomarkers have shown that the natural history of AD in pwDS is very similar to those with sporadic AD (sAD) 3,[5][6][7][8] This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. As a potential early biomarker for AD, retinal thinning has been investigated in DS. In contrast to what has been observed in sAD or mild cognitive impairment (MCI), the retina was found to be thicker in those with DS compared to age-matched controls 9 despite the AD pathology and the accelerated aging effect in DS. 2,10 In sAD, the choroid also undergoes thinning, both in MCI 11 and later stages, 12,13 as assessed in vivo using enhanced depth imaging optical coherence tomography (EDI-OCT). It is believed that Aβ accumulation in the choroid leads to inflammation, which results in neurodegeneration and vascular attenuation, mirroring the evolution of the amyloid cascade in the brain. 11,13 Despite the choroid having been proposed as a potential early and non-invasive biomarker that may reflect neurodegeneration in the brain, the clinical utility of imaging the choroid remains inconclusive in sAD. 14,15 Given that pwDS overproduce Aβ, studying choroidal thickness (ChT) in adults with DS before or during the onset of clinical dementia, may provide a better insight into early choroidal changes due to developing AD pathology. The only published study in the literature assessed ChT in DS in children and adolescents and found no evidence of choroidal thinning relative to an age-matched Ctrl group. 16 However, ChT has never been assessed in adults with DS. For the first time, this cross-sectional study investigated whether choroidal thinning reported in sAD and MCI is mirrored in adults with DS before clinical evidence of dementia is apparent.

Study recruitment and imaging
The imaging was undertaken by trained examiners in Cambridge, UK.
The pwDS were recruited from an existing cohort. 9 Age-matched con-

Choroidal segmentation
As

HIGHLIGHTS
• This is the first study investigating ChT1 in pwDS2.
• Manual segmentation revealed an increased ChT 1 in pwDS 2 . • There is an increased prevalence of suprachoroidal space in pwDS2.
• Cerebral Aβ3 load correlates with ChT 1 in a small subset of pwDS 2 .
• The choroidal thickening in pwDS 2 is contrary to the thinning reported in sAD 4 .

RESEARCH IN CONTEXT
1. Systematic review: Based on reviewing the available literature, using keywords, optical coherence tomography, choroidal thickness (ChT), and Down syndrome (DS), we identified one study assessing ChT in DS. This study recruited children with DS and found no significant difference in ChT compared to age-matched controls. This is appropriately cited in our manuscript.

Statistical analysis
All analysis was conducted using SPPS (version 26.0; SPSS Inc., Chicago, IL) and data were visualized using GraphPad Prism (version 7

Cohort characteristics
There was no significant difference in age or gender ratio between DS and Ctrl groups ( Table 1). The mean CAMCOG-DS score was 78.43, with a standard deviation of ± 16.08 (Table 1).

There are significant changes in the choroid in pwDS
To assess the reliability of the manual segmentation, inter-rater agreement was calculated using ICC and was found to range from 0.88 to 0.97, falling in the good-excellent correlation range. ChT values were not significantly (p > 0.05) different between the right and left eyes ( Figure S1).
In pwDS, regional differences in ChT across the posterior pole were very similar to the pattern of regional differences measured in the Ctrl group (Figure 1 D). The choroid was thinnest in the nasal outer (NO) and thickest in the superior inner (SI) sectors in both pwDS and Ctrls ( Figure 1 D and Table S1). When ChT values were compared between the two groups, apart from the inferior inner (II) sector, a significantly increased ChT was detected across all the sectors of the ETDRS grid in pwDS compared to the Ctrls (Table S1). In the final adjusted model, the differences remained significant in all sectors except in II and IO ( Figure 1 D and Table S1).  Figure 1E and Table S2). Significant differences were not observed in Ctrl (P > 0.05) ( Figure 1E and Table S2).

TA B L E 2
The relationship between choroidal thickness and age, cognition, and brain Aβ

There is a correlation between ChT and cerebral Aβ load
There was no significant correlation (P > 0.05) detected between ChT and age or CAMCOG-DS (Table 2) In addition, the relationship between ChT and retinal thickness, published earlier, 9 was also assessed, and found no significant correlation in any of the ETDRS grid sectors (P > 0.05) (data not presented).

DISCUSSION
Identifying early, non-invasive and inexpensive biomarkers for proxy outcome measure is crucial for the success of AD trials, especially for those pwDS, as many are likely to benefit from treatment trials aimed at preventing the onset of AD pathology. Choroidal thinning has been proposed as an early biomarker for AD in the typically developing population. [11][12][13] We conducted a detailed choroidal assessment by extracting thickness values for each sector of the ETDRS grid. The thicker choroid detected in our DS cohort with no signs of clinical dementia does not mirror choroidal changes demonstrated in sAD and MCI. [11][12][13] Choroidal thickening due to increased vessel number in the choroidal stroma was found in a post-mortem histological study of sAD. 14 The vascular proliferation could be the response to a metabolic dysfunction of the retina due to retinal Aβ deposition. 21,22 An increasing number of publications on post-mortem tissues show Aβ plaque-like deposits in the retina of donors with sAD. [23][24][25][26][27] However, not all groups can verify these findings [28][29][30] or can distinguish between people with sAD and age-matched, cognitively normal controls. 31 Only one study examined the presence of Aβ plaques in the retina of pwDS using a modified laser scanning ophthalmoscope and curcumin labeling. 32 In the absence of controls, this study could not determine the power of the method. Based on emerging data on sAD, 24,33 the need for further examination of retinal Aβ in vivo is warranted.
We also investigated whether the increased ChT measured in the DS group could result from the SCS that was exclusively visible in pwDS.
We found no significant difference in ChT between those with or without SCS. The SCS appears to be an age-related phenomenon most likely F I G U R E 2 Relationship between choroidal thickness and cerebral Aβ load. The scatter plots show the global choroidal thickness correlation with medial orbitofrontal (A) and rostral middle-frontal (B) cortex Aβ load. Unadjusted values are plotted, all cortical areas were analyzed, but only the above two showed significant correlation. Abbreviations: MOF, medial orbitofrontal; RMF, rostral middle-frontal; Aβ, amyloid beta due to fluid accumulation at the scleral and choroidal interface. 34 It is present in approximately 50% of typically developing individuals above the age of 50 and is associated with hyperopia. 34 The known accelerated aging effect and the higher prevalence of refractive error (in particular, hyperopia) in pwDS may explain the exclusive presence of SCS in the DS group. 10,35 ChT is decreased in patients with systemic arterial hypertension and may result from vascular contraction caused by high intravascular pressure in the choroid. 36,37 A study on healthy volunteers aged 18 and 60 found an inverse relationship between systemic blood pressure and sub-foveal ChT. 38 Although blood pressure data were not available for our cohort, DS is associated with lower systemic blood pressure. 39,40 Therefore, we speculate that the increased ChT observed in our DS group might be, at least partially, due to lower systemic and or choroidal blood pressure.
We have previously shown that a thicker retina was associated with DS, and the thickening occurred in the inner retinal layers. 9 pwDS have a markedly increased number of vessels in the retinal, which is believed to be the consequence of altered angiogenesis in DS. 41 Because retinal vessels are present in the inner retinal layers, the higher vessel number may contribute to the inner retinal layer changes. 41 Therefore, apart from the decreased blood pressure, the altered angiogenesis in pwDS may also contributor to the increased ChT.
A recent study showed no significant increase in ChT in children with DS, 16 suggesting that the significantly thicker choroid detected in our study is a feature of an older DS population. This raises the question of whether the observed choroidal thickening in pwDS could result from developing AD pathology. The positive correlation between ChT and the increased Aβ load in the MOF and RMS cortices could be related to inflammatory changes observed in early AD stages. 42 It is perhaps not surprising that differences are associated with the frontal cortices, as these areas are among the first that are affected by Aβ deposition in pwDS. 4 However, the results should not be overinterpreted, considering the low number of study participants with PET data in this study and the lack of correlation with other primarily affected areas such as the striatum. 4 While information on choroidal changes are sparse, 15 retinal thickening had been reported in early stages of sporadic AD (preclinical AD and MCI), and believed to be the result of inflammatory processes due to AD pathology. 43,44 Hence, we cannot rule out the possibility that the thickened choroid in our cohort is, at least partially, the result of inflammatory processes due to developing AD pathology.
We found thinner choroid in females compared to males in pwDS, a difference that was not detected in the Ctrl group. The same heterogeneity in ChT between females and males has been observed in the general population, 45 and it is believed to be driven by hormonal changes in postmenopausal women. 46 Although the information on the onset of menopause in our study was not available, it is well known that the menopause occurs earlier in pwDS women than the general population and is associated with an increased risk of dementia. 47 Phenotypic variability has been shown in sAD, and gender was a significant driving force. 48 We are not aware of studies reporting gender-related phenotypic differences in choroidal or retinal thickness in pwDS or sAD.
Future studies assessing eye biomarkers for AD should consider gender in study design based on our observation.
Decreased ChT had been associated with age. 49 However, we did not observe a significant age effect on ChT in the Ctrl or pwDS, probably due to the relatively young age and the limited age-range in our study. Previous studies have also demonstrated a positive correlation between ChT and cognitive status in sAD and MCI, 11 but this was probably not detectable in our cohort due to the limited differences in cognitive scores. Therefore, further studies should elucidate further this relationship between ChT and cognitive measures in pwDS. Tan et al.
demonstrated that there are diurnal variations in ChT but this could also be influenced by refractive error, intraocular pressure (IOP), and blood pressure. 50 Although our analysis controlled for diurnal variation, refractive error, IOP and blood pressure data were not available.
We demonstrated a potential relationship between brain Aβ load and ChT in a subset of pwDS (n = 6). However, further studies will need to verify this relationship due to the small sample size and possible gender bias.
One of our study's strengths was the comprehensive analysis undertaken by segmenting all the 25 OCT B-scans per posterior EDI scan, which allowed us to extract data for each sector of the ETDRS grid. To ensure that the manual choroidal segmentation was robust, we used a second grader and found an excellent inter-observer agreement.

CONFLICTS OF INTEREST
The authors declare no conflict of interests.

DATA AVAILABILITY STATEMENT
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

STATEMENT OF SIGNIFICANCE
This is the first study that reports choroidal thickening in the eye in adults with Down syndrome. This finding is in contrast to the thinning observed in sporadic Alzheimer's disease.