Ocular biomarkers for Alzheimer’s disease
van de Kreeke, J.A.
2020
document version
Publisher's PDF, also known as Version of record
Link to publication in VU Research Portal
citation for published version (APA)
van de Kreeke, J. A. (2020). Ocular biomarkers for Alzheimer’s disease.
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal ? Take down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
E-mail address:
Summary and
general discussion
In this study, we investigated the diagnostic potential of ocular biomarkers in identification of AD pathology in the preclinical stage and in mild Alzheimer’s type dementia (AD). Additionally, we investigated the role of aging and genes on these ocular biomarkers. The main findings of this thesis are:
‐ Diagnostic potential of ocular biomarkers:
o Several retinal vascular parameters (RVPs) are associated with cerebral white matter hyperintensities (WMH) on MRI.
o RVPs are unable to discriminate cases of (preclinical) AD, mild AD dementia or cognitive impairment.
o Retinal layer thickness on optical coherence tomography (OCT) cannot discriminate (preclinical) AD, mild AD dementia or cognitive impairment,
o Inner plexiform layer (IPL) thickness on OCT decreases less over time in individuals with higher amyloid-beta (Aβ)-load on PET.
o Vessel density on OCT angiography (OCTA) is significantly higher in individuals with preclinical AD.
‐ Role of aging on ocular biomarkers:
o Almost all ocular biomarkers show a strong association with age, in both cross-sectional and longitudinal studies.
‐ Role of genes on ocular biomarkers:
o Almost all ocular biomarkers show a moderate to high correlation between identical twins, suggesting an influence of genes.
o Twin specific analyses support most of the described associations between ocular biomarkers and cerebral pathology, suggesting them to be partly caused by a common, most likely genetic, etiological factor. In this section, we combine, analyze and discuss these main findings from several studies, and place them in the context of what is already known in literature. We also elaborate on some challenges encountered in performing these studies and possible limitations of the use of ocular biomarkers for screening. Finally, we end with general recommendations and future perspectives on the use of ocular biomarkers for (preclinical) AD.
Retinal vascular parameters as biomarker
The first chapter of this thesis illustrated that several retinal vascular parameters (RVPs) measured using Singapore I Vessel Assessment (SIVA) were linked to cerebral small vessel disease (SVD) as measured by white matter hyperintensities (WMH) on MRI. Retinal venous width, fractal dimension of both arteries and veins and retinal venous tortuosity were all positively associated with WMH. This suggests that retinal vasculature may indeed reflect cerebral vascular pathology. This is supported by several other studies (1-4). However, it has to be noted that not all these associations run in the same direction. For example, fractal dimension is often described to be negatively associated with SVD (1, 2). An explanation for this discrepancy may lie in the different populations studied: only cognitively healthy participants were included for our study, whereas most other studies on this subject included participants already predisposed to SVD.
As SVD is part of the pathological spectrum in AD, it could be expected that RVPs may also be associated with AD. Chapter 2 illustrated that this was not the case. In a cohort of well described and biomarker proven AD patients, RVPs could not discriminate the AD cases from healthy controls. This contrasts with other studies. Both Frost et al. and Cheung et al. reported an arteriolar narrowing in AD, while Williams et al. found an arteriolar widening in patients with AD (5-7). Considering the sparsity of the studies on this subject, as well as their conflicting results in addition to our own negative results, the use of RVPs as a potential screening method for AD is still questionable at best.
In chapters 4 and 7, we explored the use of RVPs in discriminating individuals in the preclinical phase of AD from controls. Both cross-sectional as well as longitudinal, RVPs did not differ between participants with and without preclinical AD. Additionally, no associations with cerebral amyloid-beta (Aβ) load were found. The longitudinal study in particular should have been able to find any potential differences in preclinical AD cases, as this eliminates the high inter-person variability of RVPs. To our knowledge, no other studies have been performed on the subject of RVPs in preclinical AD, but the results of our studies imply that the role for RVPs as a biomarker for preclinical AD will be limited. Lastly, chapter 9 looks at the potential of RVPs to discriminate cognitive impairment in a population where this is most prevalent: nonagenarians. RVPs did not show any differences between cognitively impaired nonagenarians and their healthy peers, illustrating yet again that RVPs seem unable to aid in a diagnosis of cognitive impairment.
OCT as biomarker
There is currently a lot of attention on the subject of OCT as a potential biomarker of AD. Quite some studies have been performed to assess the potential differences in retinal thickness in AD patients, most of which report a thinner retina in AD patients (8, 9). Especially thinning of the peripapillary retinal nerve fiber layer (RNFL) and inner macular layers (RNFL, ganglion cell layer and inner plexiform layer) are reportedly associated with a diagnosis of AD-type dementia. Chapter 3 shows that we were unable to reproduce these findings. Total retinal thickness and peripapillary RNFL did not differ between a well characterized, biomarker proven cohort of AD patients and controls, nor did any of the inner retinal layers. Chapter 9 illustrates that OCT was also unable to discriminate between nonagenarians with and without cognitive impairment. We are not the only group to report negative findings on this subject. For example, Sanchez et al. also could not illustrate differences in retinal thickness in their AD participants, notwithstanding their sizeable included population (10).
The possible value of OCT to make a diagnosis of AD as is suggested by many other groups, would be especially valuable in the earliest disease stages of AD. Ideally, a biomarker for AD is already applicable in the preclinical stages, so cognitive decline may still be prevented. Unfortunately, chapters 5 and 8 illustrated that OCT was unable to discriminate cases of preclinical AD, even when using longitudinal data (i.e., look at changes over time). The studies that have been performed on this subject are a lot scarcer. Currently, 2 other groups besides our own report on the use of OCT in preclinical AD. One found a higher IPL volume and significant decrease of macular RNFL volume in participants with preclinical AD, whereas the other found a thinner total foveal thickness (11-13). Yet, neither group described the typical differences often reported in AD-type dementia: a thinner (peripapillary) RNFL, GCL and/or IPL. It seems therefore likely that the typical changes found in retinal thickness in AD patients can be measured in the dementia stage only.
One finding that may, however, be of interest is that of an association between change in IPL thickness and Aβ load, as reported in chapter 8. Participants with a higher Aβ load at baseline decreased less in their IPL thickness over the 22 months follow-up time. This might suggest that cerebral Aβ deposition is in a way linked to changes in retinal IPL. This can be a direct link, i.e.: Aβ deposition may also be occurring in the retina, and more particularly the IPL, explaining why IPL does not follow the regular age-related thinning that is observed in individuals with low cortical Aβ. This theory has been postulated by Snyder et al, who also observed a thicker IPL in cases of preclinical AD (13). They noticed so-called ‘inclusion bodies’ in the IPL, assumed to be consisting of Aβ and observable on autofluorescence. However, the presence of retinal Aβ is still a very controversial subject. While one lab was able to illustrate this presence in post-mortem retinas of AD patients,
many other were not (14-17). The link between IPL thickness and cortical Aβ may therefore also be indirect: a pathological process occurring in both the retina and the brain may be causing cortical Aβ deposition as well as retinal IPL thickening due to some other process. This could be microscopic fluid leakage in the IPL due to inflammation. Regardless of its cause, a possible relation between changes in the IPL and cortical AB may offer leads for further research on the topic of OCT as a (longitudinal) biomarker for preclinical AD.
OCTA as biomarker
OCTA is a very recent technique, making the studies on the subject of OCTA as a biomarker for AD still limited. However, the studies that have been published show a promising role for OCTA. They demonstrated an enlargement of the foveal avascular zone (FAZ) and a decrease in vessel density in AD-type dementia (18-20). These changes are ascribed to microvascular damage, resulting in fewer functional capillaries which in turn can lead to said enlargement of the FAZ and a decrease in vessel density. Chapter 2 illustrated that we could not reproduce these findings. In our cohort of well-defined AD patients in the dementia stage, we found no significant differences in neither vessel density nor FAZ size. A reason for the discrepancy between existing literature and our own findings may be that we corrected for scan quality, whereas none of the other studies report this. We noticed scan quality had a very high impact on especially vessel density. As it is more difficult to scan AD patients suffering from cognitive impairment than healthy controls, this might result in scans of a generally lower quality for AD patients, thereby introducing a bias that should be corrected for.
Interestingly, chapter 6 illustrated that OCTA was able to discriminate cases of preclinical AD using vessel density on OCTA, but vessel density was increased rather than decreased in preclinical AD patients. We found this increased vessel density in all the retinal regions we examined, with the ONH region even achieving a sensitivity and specificity of around 0.8 in diagnosing cases of preclinical AD. A possible explanation for this finding may again lie in inflammation: a very early inflammatory process often leads to an increase in blood flow. This increased blood flow may lead to more capillaries reaching the detection threshold for OCTA and result in an increased vessel density. As inflammation progresses, capillaries may become damaged and perish, ultimately leading to a decrease in vessel density, as was reported in cases of clinical AD. Only one other study tested OCTA as a possible biomarker for preclinical AD (11). Unfortunately, although vessel density was one of the parameters examined, this study reported no results, leading us to believe they found no significant differences in vessel density in their preclinical cases. They did report an increased FAZ size in their preclinical cases of AD, which is something we in turn did
not find. Therefore, whilst OCTA may show promise as a potential biomarker for preclinical AD, more studies are certainly needed to ascertain the somewhat contradicting findings to date.
Aging and ocular biomarkers
Aging has a profound effect on all parts of the body, and ocular biomarkers are not excluded from these effects. Studies have illustrated a negative association between retinal thickness and age (21, 22). Because of this, the referential retinal (layer) thickness in OCT databases is adjusted for age. Both chapter 3 and chapter 9 confirm this negative association between retinal thickness and age in cognitively impaired as well as healthy individuals. Furthermore, in chapter 8 we were able to evaluate the longitudinal effects of age and illustrated that even in a relatively short period of 22 months, almost all retinal layers showed a significant thinning in a healthy elderly population.
The effects of age on RVPs are somewhat less known, but several groups also reported associations with RVPs, in particular a decrease in arterial and venous calibers (23-25).
Chapters 2 and 9 confirmed this negative association between age and vascular calibers,
and even illustrated some other associations (most markedly a negative association with retinal fractal dimension). In chapter 7 we evaluated the effect of age in a longitudinal set-up and confirm that several RVPs changed significantly over the 22 month follow-up time, with decreasing vascular calibers being the most prominent change.
Almost all of these changes (thinning of retinal layers, decreasing vascular calibers) are typically those described to occur in AD, as has been discussed in the previous paragraphs. Hence, age is an extremely important confounder in any study on the subject of ocular biomarkers for neurodegenerative disease. This should always be taken into account, either by matching age groups or by correcting for age in statistical models (or both).
Genes and ocular biomarkers
A substantial part of this thesis is based on a cohort study performed using a very unique population: monozygotic twins. Twins are ideal subjects to investigate the effect of nature versus nurture, as they share either half (dizygotic) or all (monozygotic) of their genes. By analyzing similarities and differences between twins of the same pair, and comparing this between dizygotic and monozygotic pairs, information on the estimated contribution of genes, shared environment and unique environment on biological parameters may be
made (26). Furthermore, twin specific analyses, can offer further insight on the role of genes and environment on associations between biological parameters (27, 28). This enables a more profound understanding on the etiology of such biological parameters, as well as their associations.
As only monozygotic twin pairs were included as the basis for this thesis, the true contribution of genes cannot be calculated exactly, as any similarities between twins from the same pair may be due to either their shared genes or their shared environment. However, as this population consists of elderly twins (average age of 70), it is very likely that most similarities will be largely due to genes rather than shared environment, as most of their lives have by now been spent apart and thus in an unique environment.
Chapters 1 and 3 illustrated that both RVPs as well as retinal layer thickness are biological
measures which are highly determined by genes. Intra-twin pair correlations were moderate to high: around 0.6 for most RVPs and around 0.8-0.9 for most retinal (layer) thicknesses. This is confirmed by several other studies, which also demonstrated a high heritability for both RVPs and retinal thickness (29-32). Chapters 7 and 8 illustrated that even the rate of change in these biological measures is partly shared by twins from the same pair, suggesting that effects of aging on these parameters is also, to a certain degree, genetically determined. Chapter 6 also illustrated a high intra-twin pair correlation for foveal avascular zone size calculated using OCTA (0.85), but slightly lower correlations for vessel density (around 0.5), suggesting genes play a smaller but nonetheless substantial role in these biological parameters. No other studies on the subject of heritability of FAZ and vessel density have been performed, so the suggested contribution of genes has yet to be confirmed by other groups.
Chapter 1 showed, using a cross-twin cross-trait analysis, that retinal venous tortuosity
of twin was associated with deep WMH volume of twin 2, suggesting this relation to have a partly genetic basis. Furthermore, a significant twin difference analysis on this same association (meaning that within twin-pair differences in retinal venous tortuosity were associated to within twin-pair differences in deep WMH volume) also implied the existence of a common unique environmental factor influencing both parameters. This suggests that retinal venous tortuosity and WMH have the same underlying pathology and that RVPs indeed reflect brain vascular pathology. In chapter 6, a twin post-hoc staging model for Aβ load supported the finding that vessel density was increased in preclinical AD cases, as the staging model showed a linear increase along increasing stages of Aβ load.
Limitations of ocular biomarkers
Despite the obvious advantages in using ocular biomarkers, there are some drawbacks. These biomarkers are altered in many ophthalmological conditions. Such ophthalmological diseases, like age-related macular degeneration or glaucoma, are quite common, especially in the elderly age group when (preclinical) AD often manifests. The presence of ophthalmological disease therefore disables the use of ocular biomarkers for neurological disease. Furthermore, systemic conditions such as diabetes mellitus are also known to alter these ocular biomarkers, enlarging this problem to other conditions as well. Currently it seems that only systemically and ophthalmologically healthy people would be suitable for (preclinical) AD screening using ocular biomarkers, thereby limiting their use considerably.
Measures such as retinal thickness and retinal vascularity differ considerably between individuals, simply due physiological anatomical variation. Any difference in these ocular biomarkers therefore needs to be substantial in other to cross a detection threshold when using a normative database. The differences described by other groups, as well as some of the differences found in this thesis, almost never reach such levels, many times encompassing only a few percent of the total average measure. Clinical applicability of reported differences, even when statistically significant, are thus often limited.
Detection of vessels on fundus photography using automated software proved challenging. SIVA was often very inaccurate in the correct detection of veins and arteries, or their anatomy. This required a lot of manual correction by a grader, resulting in a subjective bias. Furthermore, image conditions and quality (e.g. lighting, focus) strongly influenced the grading process by SIVA, introducing even more bias. This severely limited reproducibility of RVPs, making them less suitable as reliable ocular biomarkers.
Methodological considerations
It is estimated cortical Aβ is present in around 20 percent of individuals at around age 70. The largest part of those are cognitively healthy, and therefore considered to be preclinical AD cases. It is very hard to identify a population with preclinical AD of a decent size, as this would implicate that many cognitively healthy individuals need to be scanned with an expensive PET scan. This is a problem which also applies to our preclinical cohort; in the chapters on preclinical AD only around 20 participants or less were positive for Aβ. This limits statistical power, making it harder to detect subtle differences. The use of a population of monozygotic twins also has a drawback. As two individuals of a twin pair are very much alike, this creates an inter-twin dependency that needs to be statistically
accounted for and limits statistical power (i.e. twins do not count as 2 completely separate individuals).
Implications
Based on the findings in this thesis, ocular biomarkers have disappointingly little value in discriminating cases of preclinical AD or early AD. From the biomarkers investigated, it seems that OCTA may have the most potential, as vessel density measured using this technique was consequently higher in preclinical AD cases. Furthermore, a longitudinal measurement of IPL thickness using OCT may also provide leads in gaining more information on cerebral Aβ load. Based on current knowledge however, there is as yet no role for ocular biomarkers in the detection of clinical or preclinical AD.
Future perspectives
As the ocular biomarkers discussed in this thesis all suffer from a high inter person variability, longitudinal studies will probably be more promising as this set-up may eliminate this variance. Focus may also be shifted on more pathognomonic ocular biomarkers, such as pathological protein deposition in the retina. Aβ and Tau are quite specific proteins for AD, and their detection in the retina can offer valuable leads for a highly sensitive and specific ocular biomarker. This does, however, assume the presence of such proteins in the retina, which is as yet not unequivocally established. More post-mortem studies are therefore the first step, to establish whether these proteins are present. In vivo detection of said, or other AD-specific, proteins would be the next step. There are several promising techniques of doing so, such as ultra-wide field imaging, fluorescence lifetime imaging, multispectral imaging or tear protein analyses.
References
1. Doubal FN, de Haan R, MacGillivray TJ, Cohn-Hokke PE, Dhillon B, Dennis MS, et al. Retinal arteriolar geometry is associated with cerebral white matter hyperintensities on magnetic resonance imaging. Int J Stroke. 2010;5(6):434-9.
2. Hilal S, Ong YT, Cheung CY, Tan CS, Venketasubramanian N, Niessen WJ, et al.
Microvascular network alterations in retina of subjects with cerebral small vessel disease. Neurosci Lett. 2014;577:95-100.
3. Hughes AD, Falaschetti E, Witt N, Wijetunge S, Thom SA, Tillin T, et al. Association of Retinopathy and Retinal Microvascular Abnormalities With Stroke and Cerebrovascular Disease. Stroke. 2016;47(11):2862-4.
4. Ikram MK, De Jong FJ, Van Dijk EJ, Prins ND, Hofman A, Breteler MM, et al. Retinal vessel diameters and cerebral small vessel disease: the Rotterdam Scan Study. Brain. 2006;129(Pt 1):182-8.
5. Cheung CY, Ong YT, Ikram MK, Ong SY, Li X, Hilal S, et al. Microvascular network alterations in the retina of patients with Alzheimer's disease. Alzheimers Dement. 2014;10(2):135-42.
6. Frost S, Kanagasingam Y, Sohrabi H, Vignarajan J, Bourgeat P, Salvado O, et al. Retinal vascular biomarkers for early detection and monitoring of Alzheimer's disease. Transl Psychiatry. 2013;3:e233.
7. Williams MA, McGowan AJ, Cardwell CR, Cheung CY, Craig D, Passmore P, et al. Retinal microvascular network attenuation in Alzheimer's disease. Alzheimers Dement (Amst).
2015;1(2):229-35.
8. Chan VTT, Sun Z, Tang S, Chen LJ, Wong A, Tham CC, et al. Spectral-Domain OCT Measurements in Alzheimer's Disease: A Systematic Review and Meta-analysis. Ophthalmology. 2018.
9. den Haan J, Verbraak FD, Visser PJ, Bouwman FH. Retinal thickness in Alzheimer's disease: A systematic review and meta-analysis. Alzheimers Dement (Amst). 2017;6:162-70. 10. Sanchez D, Castilla-Marti M, Rodriguez-Gomez O, Valero S, Piferrer A, Martinez G, et al. Usefulness of peripapillary nerve fiber layer thickness assessed by optical coherence tomography as a biomarker for Alzheimer's disease. Sci Rep. 2018;8(1):16345.
11. O'Bryhim BE, Apte RS, Kung N, Coble D, Van Stavern GP. Association of Preclinical Alzheimer Disease With Optical Coherence Tomographic Angiography Findings. JAMA Ophthalmol. 2018;136(11):1242-8.
12. Santos CY, Johnson LN, Sinoff SE, Festa EK, Heindel WC, Snyder PJ. Change in retinal structural anatomy during the preclinical stage of Alzheimer's disease. Alzheimers Dement (Amst). 2018;10:196-209.
13. Snyder PJ, Johnson LN, Lim YY, Santos CY, Alber J, Maruff P, et al. Nonvascular retinal imaging markers of preclinical Alzheimer's disease. Alzheimers Dement (Amst). 2016;4:169-78. 14. Ho CY, Troncoso JC, Knox D, Stark W, Eberhart CG. Beta-amyloid, phospho-tau and alpha-synuclein deposits similar to those in the brain are not identified in the eyes of Alzheimer's and Parkinson's disease patients. Brain Pathol. 2014;24(1):25-32.
15. Williams EA, McGuone D, Frosch MP, Hyman BT, Laver N, Stemmer-Rachamimov A. Absence of Alzheimer Disease Neuropathologic Changes in Eyes of Subjects With Alzheimer Disease. J Neuropathol Exp Neurol. 2017;76(5):376-83.
16. den Haan J, Morrema THJ, Verbraak FD, de Boer JF, Scheltens P, Rozemuller AJ, et al. Amyloid-beta and phosphorylated tau in post-mortem Alzheimer's disease retinas. Acta Neuropathol Commun. 2018;6(1):147.
17. Koronyo Y, Biggs D, Barron E, Boyer DS, Pearlman JA, Au WJ, et al. Retinal amyloid pathology and proof-of-concept imaging trial in Alzheimer's disease. JCI Insight. 2017;2(16).
18. Bulut M, Kurtulus F, Gozkaya O, Erol MK, Cengiz A, Akidan M, et al. Evaluation of optical coherence tomography angiographic findings in Alzheimer's type dementia. Br J Ophthalmol. 2018;102(2):233-7.
19. Jiang H, Wei Y, Shi Y, Wright CB, Sun X, Gregori G, et al. Altered Macular
Microvasculature in Mild Cognitive Impairment and Alzheimer Disease. J Neuroophthalmol. 2018;38(3):292-8.
20. Grewal DS, Polascik BW, Hoffmeyer GC, Fekrat S. Assessment of Differences in Retinal Microvasculature Using OCT Angiography in Alzheimer's Disease: A Twin Discordance Report. Ophthalmic Surg Lasers Imaging Retina. 2018;49(6):440-4.
21. Demirkaya N, van Dijk HW, van Schuppen SM, Abramoff MD, Garvin MK, Sonka M, et al. Effect of age on individual retinal layer thickness in normal eyes as measured with spectral-domain optical coherence tomography. Invest Ophthalmol Vis Sci. 2013;54(7):4934-40.
22. Won JY, Kim SE, Park YH. Effect of age and sex on retinal layer thickness and volume in normal eyes. Medicine (Baltimore). 2016;95(46):e5441.
23. Leung H, Wang JJ, Rochtchina E, Tan AG, Wong TY, Klein R, et al. Relationships between age, blood pressure, and retinal vessel diameters in an older population. Invest Ophthalmol Vis Sci. 2003;44(7):2900-4.
24. Sun C, Wang JJ, Mackey DA, Wong TY. Retinal vascular caliber: systemic, environmental, and genetic associations. Surv Ophthalmol. 2009;54(1):74-95.
25. Wong TY, Klein R, Klein BE, Meuer SM, Hubbard LD. Retinal vessel diameters and their associations with age and blood pressure. Invest Ophthalmol Vis Sci. 2003;44(11):4644-50. 26. Grasby KL, Verweij KJH, Mosing MA, Zietsch BP, Medland SE. Estimating Heritability from Twin Studies. Methods Mol Biol. 2017;1666:171-94.
27. De Moor MH, Boomsma DI, Stubbe JH, Willemsen G, de Geus EJ. Testing causality in the association between regular exercise and symptoms of anxiety and depression. Arch Gen Psychiatry. 2008;65(8):897-905.
28. Vitaro F, Brendgen M, Arseneault L. The discordant MZ-twin method: One step closer to the holy grail of causality. International Journal of Behavioral Development. 2009;33(4):376-82. 29. Chamberlain MD, Guymer RH, Dirani M, Hopper JL, Baird PN. Heritability of macular thickness determined by optical coherence tomography. Invest Ophthalmol Vis Sci.
2006;47(1):336-40.
30. Hougaard JL, Kessel L, Sander B, Kyvik KO, Sorensen TI, Larsen M. Evaluation of heredity as a determinant of retinal nerve fiber layer thickness as measured by optical coherence
tomography. Invest Ophthalmol Vis Sci. 2003;44(7):3011-6.
31. Sun C, Zhu G, Wong TY, Hewitt AW, Ruddle JB, Hodgson L, et al. Quantitative genetic analysis of the retinal vascular caliber: the Australian Twins Eye Study. Hypertension.
2009;54(4):788-95.
32. Taarnhoj NC, Larsen M, Sander B, Kyvik KO, Kessel L, Hougaard JL, et al. Heritability of retinal vessel diameters and blood pressure: a twin study. Invest Ophthalmol Vis Sci.
