ORIGINAL RESEARCH published: 21 July 2021 doi: 10.3389/fnagi.2021.712545 Plasma Proteomic Biomarkers Relating to Alzheimer’s Disease: A Meta-Analysis Based on Our Own Studies Liu Shi 1 *, Noel J. Buckley 1 , Isabelle Bos 2,3 , Sebastiaan Engelborghs 4,5 , Kristel Sleegers 6,7 , Giovanni B. Frisoni 8 , Anders Wallin 9 , Alberto Lléo 10 , Julius Popp 11,12 , Pablo Martinez-Lage 13 , Cristina Legido-Quigley 14,15 , Frederik Barkhof 16,17 , Henrik Zetterberg 9,18,19,20 , Pieter Jelle Visser 2,3 , Lars Bertram 21,22 , Simon Lovestone 1,23 and Alejo J. Nevado-Holgado 1 1 Department of Psychiatry, University of Oxford, Oxford, United Kingdom, 2 Department of Psychiatry and Neuropsychology, School for Mental Health and Neuroscience, Alzheimer Centrum Limburg, Maastricht University, Maastricht, Netherlands, 3 Alzheimer Center, VU University Medical Center, Amsterdam, Netherlands, 4 Reference Center for Biological Markers of Dementia (BIODEM), Institute Born-Bunge, Department of Biomedical Sciences, University of Antwerp, Antwerp, Belgium, 5 Department of Neurology, Universitair Ziekenhuis Brussel and Center for Neurociences (C4N), Vrije Universiteit Brussel, Brussels, Belgium, 6 Complex Genetics Group, VIB Center for Molecular Neurology, VIB, Antwerp, Belgium, 7 Institute Born-Bunge, Department of Biomedical Sciences, University of Antwerp, Antwerp, Belgium, 8 Department of Psychiatry, University of Geneva, Geneva, Switzerland, 9 Institute of Neuroscience and Physiology, Department of Psychiatry and Edited by: Neurochemistry, The Sahlgrenska Academy at the University of Gothenburg, Mölndal, Sweden, 10 Hospital de la Santa Creu i Can Martin Zhang, Sant Pau, Barcelona, Spain, 11 Department of Psychiatry, University Hospital of Lausanne, Lausanne, Switzerland, 12 Geriatric Harvard Medical School, Psychiatry, Department of Mental Health and Psychiatry, Geneva University Hospitals, Geneva, Switzerland, 13 CITA-Alzheimer United States Foundation, San Sebastian, Spain, 14 Kings College London, London, United Kingdom, 15 The Systems Medicine Group, Reviewed by: Steno Diabetes Center, Gentofte, Denmark, 16 Department of Radiology and Nuclear Medicine, VU University Medical Center, Jordi A. Matias-Guiu, Amsterdam, Netherlands, 17 UCL Institutes of Neurology and Healthcare Engineering, London, United Kingdom, 18 Clinical Hospital Clínico San Carlos, Spain Neurochemistry Laboratory, Sahlgrenska University Hospital, Mölndal, Sweden, 19 UK Dementia Research Institute at UCL, David Morgan, London, United Kingdom, 20 Department of Neurodegenerative Disease, UCL Institute of Neurology, London, Michigan State University, United Kingdom, 21 Lübeck Interdisciplinary Platform for Genome Analytics, University of Lübeck, Lübeck, Germany, 22 United States Department of Psychology, University of Oslo, Oslo, Norway, 23 Janssen R&D, High Wycombe, United Kingdom *Correspondence: Liu Shi Background and Objective: Plasma biomarkers for the diagnosis and stratification of liu.shi@psych.ox.ac.uk Alzheimer’s disease (AD) are intensively sought. However, no plasma markers are well Received: 20 May 2021 established so far for AD diagnosis. Our group has identified and validated various blood- Accepted: 21 June 2021 based proteomic biomarkers relating to AD pathology in multiple cohorts. The study Published: 21 July 2021 aims to conduct a meta-analysis based on our own studies to systematically assess the Citation: diagnostic performance of our previously identified blood biomarkers. Shi L, Buckley NJ, Bos I, Engelborghs S, Sleegers K, Methods: To do this, we included seven studies that our group has conducted during Frisoni GB, Wallin A, Lléo A, Popp J, Martinez-Lage P, Legido-Quigley C, the last decade. These studies used either Luminex xMAP or ELISA to measure Barkhof F, Zetterberg H, Visser PJ, proteomic biomarkers. As proteins measured in these studies differed, we selected Bertram L, Lovestone S and protein based on the criteria that it must be measured in at least four studies. We Nevado-Holgado AJ (2021) Plasma Proteomic Biomarkers Relating to then examined biomarker performance using random-effect meta-analyses based on the Alzheimer’s Disease: A Meta-Analysis mean difference between biomarker concentrations in AD and controls (CTL), AD and Based on Our Own Studies. Front. Aging Neurosci. 13:712545. mild cognitive impairment (MCI), MCI, and CTL as well as MCI converted to dementia doi: 10.3389/fnagi.2021.712545 (MCIc) and non-converted (MCInc) individuals. Frontiers in Aging Neuroscience | www.frontiersin.org 1 July 2021 | Volume 13 | Article 712545 Shi et al. Plasma Biomarkers of Alzheimer’s Disease Results: An overall of 2,879 subjects were retrieved for meta-analysis including 1,053 CTL, 895 MCI, 882 AD, and 49 frontotemporal dementia (FTD) patients. Six proteins were measured in at least four studies and were chosen for meta-analyses for AD diagnosis. Of them, three proteins had significant difference between AD and controls, among which alpha-2-macroglobulin (A2M) and ficolin-2 (FCN2) increased in AD while fibrinogen gamma chain (FGG) decreased in AD compared to CTL. Furthermore, FGG significantly increased in FTD compared to AD. None of the proteins passed the significance between AD and MCI, or MCI and CTL, or MCIc and MCInc, although complement component 4 (CC4) tended to increase in MCIc individuals compared to MCInc. Conclusions: The results suggest that A2M, FCN2, and FGG are promising biomarkers to discriminate AD patients from controls, which are worthy of further validation. Keywords: Alzheimer’s disease (AD), diagnosis, blood biomarkers, meta-analysis, proteomic INTRODUCTION et al., 2010b; Kiddle et al., 2012; Westwood et al., 2016), have found that a range of proteins in plasma might act as biomarkers. Currently, the diagnosis of Alzheimer’s disease (AD) clinically In our review published in 2018 (Shi et al., 2018), we summarized is based on clinical examination, patient and carer interview, some of the main findings and approaches taken in the studies and structural or glucose metabolism imaging (McKhann et al., that we have conducted during the last decade. Since then, we 2011). The limitation of this approach is that a significant have further validated these identified biomarkers in two large proportion of AD patients have their diagnosis changed after independent cohorts including over 1,500 individuals. Taking all a measure of amyloid either by positron emission tomography these studies together, this study aims to conduct a meta-analysis (PET) or lumbar puncture (Barthel and Sabri, 2017), or to systematically examine the level of individual biomarkers post-mortem studies (Beach et al., 2012; Selvackadunco et al., in blood as diagnostic tools to discriminate AD patients from 2019). Furthermore, syndrome-based AD diagnosis is too late healthy subjects. given that AD pathology happens 10 or even 20 years before a clinical symptom appears (Jack et al., 2010). Measuring biomarkers through PET or in cerebrospinal fluid (CSF) can MATERIALS AND METHODS close this gap not only for clinical research but also to define AD pathophysiologically (Jack et al., 2018). However, the expense, Study Selection invasiveness, and dependence on relevant infrastructure limit Studies were selected for meta-analysis based on two inclusion their utility in clinical practice (de Almeida et al., 2011; Lista et al., criteria: (1) the study must include a group of AD patients 2013). and CTL (control) or MCIc (MCI converted to AD) and Blood-based biomarkers represent a less invasive and MCInc to (MCI non-converted to AD) conduct pare-wise potentially cost-effective approach for the diagnosis and meta-analysis; and (2) Using a quantitative method to assess classification of AD processes. Numerous studies have sought biomarker concentrations in blood (such as ELISA and Luminex plasma biomarkers relevant to AD and great progress has been xMAP). As a result, seven studies were included in this made during the past several decades (Blennow, 2017). For study including European Medical Information Framework example, recent studies demonstrated that AD hallmarks in (EMIF1000; Westwood et al., 2020), AddNeuroMed (Hye et al., plasma such as Aβ42/40, p-tau 181, and 217 can predict brain 2014), EMIF500 (Westwood et al., 2018), VU University Medical pathology with high accuracy, further adding evidence that Center (VUMC; Westwood et al., 2018), Australian Imaging, they can be used as a non-invasive approach for the diagnosis Biomarkers and Lifestyle Flagship Study of Ageing (AIBL; and prognosis of AD (Nakamura et al., 2018; Karikari et al., Ashton et al., 2015), University of California, San Francisco, 2020, 2021; Mattsson-Carlgren et al., 2020; Thijssen et al., 2020; Memory and Aging Center (UCSF; Ashton et al., 2015) and Janelidze et al., 2021). Despite these advances, great variability GE (Westwood et al., 2018) study. Among these studies, the has been observed in blood biomarker validity in individual diagnosis of AD-type dementia was based on the National studies. A recent meta-analysis (Koychev et al., 2021) showed Institute of Neurological and Communicative Disorders and that analytical assays have played an important role in deciding Stroke–Alzheimer’s Disease and Related Disorders Association the reliability of detection of AD hallmarks in blood and further criteria (McKhann et al., 1984). Of note, the EMIF1000 and research is needed to further validate their use as screening tools. EMIF500 datasets included different subjects. As proteins Apart from hallmarks in blood, an increasing number of measured in these studies differed, a protein was selected for studies (Baird et al., 2015; Zetterberg and Burnham, 2019; meta-analysis based on the criteria that it must be measured in Manzine et al., 2020), including those by ourselves (Thambisetty at least four studies. Frontiers in Aging Neuroscience | www.frontiersin.org 2 July 2021 | Volume 13 | Article 712545 Shi et al. Plasma Biomarkers of Alzheimer’s Disease TABLE 1 | Description of subjects and the included studies in this meta-analysis. Study Sample size Number of proteins CTL (n = 1,053) MCI (n = 895) MCInc (n = 539) MCIc (n = 216) AD (n = 882) FTD (n = 49) EMIF1000 (Westwood et al., 2020) 408 400 237 103 192 / 25 AddNeuroMed (Hye et al., 2014) 452 220 169 51 476 / 30 EMIF500 (Westwood et al., 2018) 97 235 / / 162 / 21 VUMC (Westwood et al., 2018) 43 17 / / 22 / 9 AIBL (Ashton et al., 2015) 49 23 12 10 6 / 20 UCSF (Ashton et al., 2015) 4 / / / 24 49 8 GE (Westwood et al., 2018) / / 121 52 / / 34 AD, Alzheimer’s disease; CTL, control; MCI, mild cognitive impairment; MCIc, MCI converted to AD; MCInc, MCI non-converted to AD; FTD, frontotemporal dementia; EMIF, European Medical Information Framework; VUMC, VU University Medical Center; AIBL, Australian Imaging, Biomarkers and Lifestyle Flagship Study of Ageing; UCSF, University of California, San Francisco, Memory and Aging Center. Meta-Analysis alpha-2-macroglobulin (A2M), complement component 4 All analyses were completed using the R package meta for (CC4), apolipoprotein A-I (ApoA1), clusterin (CLU), ficolin-2 (Viechtbauer, 2010). An effect size was defined by the mean (FCN2), and fibrinogen gamma chain (FGG). Meta-analysis difference of biomarker concentration in the two groups such as showed that three proteins had significant differences between AD vs. the control group. Values below 0 indicate that the mean AD and controls, among which A2M and FCN2 increased in concentration of biomarker was higher in the disease group, AD (Figures 1A,E) while FGG decreased in AD (Figure 1F). In otherwise indicating lower in the disease group. The variance comparison, the other three proteins did not show significant of difference was estimated using the delta method. Given that differences between AD and controls (Figures 1B–D). None these studies are not exactly identical in the characteristics of the proteins passed the significance between AD and MCI of the included samples, we, therefore, chose random-effect or MCI and controls. Furthermore, three proteins (A2M, models to calculate each biomarker separately. Random-effect FCN2, and FGG) were also measured in FTD individuals in the models assume that the true effect size varies across studies UCSF study. Pairwise comparisons showed that no significant based on a normal distribution with mean µ and variance τ2 difference was found for A2M (Figure 2A) or FCN2 (Figure 2B) (heterogeneity), indicating each study has its true effect size θi. between FTD and AD while FGG significantly increased in FTD Estimated effect sizes yi is the study-specific sampling variance compared to AD (Figure 2C). that is caused by measurement error, assuming to be normally distributed with mean θi and variance vi. Confidence intervals Meta-Analyses of Blood Biomarkers were used to assess the significance of the estimated overall for Predicting MCI Conversion effect size. The alpha level was set to 5% and confidence and Overall, five proteins were measured in at least four prediction intervals were 95% for all tests reported below unless studies in MCIc and MCInc and were chosen for it is specified. meta-analysis. The proteins were A2M, ApoA1, CC4, CLU, and complement factor H (CFH). Results showed RESULTS that none of the proteins reached the significance between MCIc and MCInc (Figures 3A,C–E), although Description of Studies CC4 tended to increase in MCIc individuals compared to We included seven studies that our group has conducted MCInc (Figure 3B). during the last decade for meta-analysis. These studies measured candidate biomarkers for AD pathology-related processes using either Luminex xMAP or ELISA. The number of sample size DISCUSSION and proteins measured in each study are shown in Table 1. Overall, the meta-analysis contained 2,879 individuals including In this meta-analysis study, we aimed to evaluate the diagnostic 1,053 healthy individuals, 895 mild cognitive impairment (MCI) value of our previously identified blood-based biomarkers for patients, 882 AD patients, and 49 frontotemporal dementia AD. We found that A2M and FCN2 increased in AD while (FTD). Furthermore, MCI individuals included 216 subjects who FGG decreased in AD compared to CTL. Furthermore, FGG subsequently converted to dementia within 3 years and 539 non- significantly increased in FTD compared to AD, indicating it converted. Details of proteins were shown in Supplementary might be specific for AD diagnosis, while further validation in Table 1. large independent cohorts is needed. In contrast, none of the proteins passed the significance between AD and MCI or MCI Meta-Analyses of Blood Biomarkers for AD and controls from the meta-analysis. This might be caused by the Diagnosis fact that MCI included both MCIc subjects and MCInc subjects. Six proteins were measured in at leastfour studies and were As MCIc and MCInc are different groups, combining them as a chosen for meta-analysis for AD diagnosis. They were single group might lead to insignificant results. Frontiers in Aging Neuroscience | www.frontiersin.org 3 July 2021 | Volume 13 | Article 712545 Shi et al. Plasma Biomarkers of Alzheimer’s Disease FIGURE 1 | Forest plot of six proteins between Alzheimer’s disease (AD) and controls. RE, random effect; (A) A2M, alpha-2-macroglobulin; (B) CC4, complement component 4; (C) ApoA1, apolipoprotein A-I; (D) CLU, clusterin; (E) FCN2, ficolin-2; (F) FGG, fibrinogen gamma chain; EMIF, European Medical Information Framework; VUMC, VU University Medical Center; AIBL, Australian Imaging, Biomarkers and Lifestyle Flagship Study of Ageing; UCSF, University of California, San Francisco, Memory and Aging Center. Our initial discovery-phase studies demonstrated that plasma A2M, FCN2, and FGG were closely associated with AD pathology (Shi et al., 2018). For example, all the three markers were associated with amyloid deposition (Kiddle et al., 2012; Ashton et al., 2015; Westwood et al., 2016, 2017). Furthermore, FCN2 and FGG were related to brain atrophy and rate of cognitive decline (Thambisetty et al., 2010a, 2011; Sattlecker et al., 2014). Apart from our own studies, other studies found that these proteins were biologically relevant to the disease process. For example, it was found that A2M was localized to diffuse amyloid plaques in AD brains (Kovacs, 2000). From the genetic evidence, A2M gene DNA polymorphisms caused increased FIGURE 2 | (A) A2M, alpha-2-macroglobulin; (B) FCN2, ficolin-2; (C) FGG, accumulation of amyloid plaques in the brain of AD patients fibrinogen gamma chain. Box plot of three proteins’ expression between AD and frontotemporal dementia (FTD). NS., not significant. ∗ p < 0.05. (Kovacs, 2000). Ficolins are activators of the lectin complement pathway (Fujita et al., 2004). Ficolin-3 (FCN3) is another member of the ficolin family, sharing approximately 50% amino acid it has been found that fibrinogen binds to amyloid, enhancing sequence homology with FCN2 (Kilpatrick and Chalmers, 2012). amyloid aggregation and fibrillization (Ahn et al., 2010). All the It has been found that FCN3 is related to insulin resistance evidence further confirms the relevance of A2M, FCN2, and FGG and diabetes (Li et al., 2008; Chen et al., 2012; Zhang et al., in AD pathogenesis, indicating they are promising biomarkers 2016). This is very interesting because there is a close relationship for AD diagnosis. between diabetes and AD (Janson et al., 2004; Talbot et al., 2012). Current findings on plasma biomarkers have generated Fibrinogen was found to accumulate along with AD pathology new enthusiasm in the blood biomarker field, particularly progresses (Ryu and McLarnon, 2009) and co-deposits with plasma neurofilament light (NfL), Aβ42/40, p-tau 181 and amyloid plaques in brain tissue (Klohs et al., 2012). Furthermore, 217, and glial fibrillary acidic protein (GFAP; Nakamura Frontiers in Aging Neuroscience | www.frontiersin.org 4 July 2021 | Volume 13 | Article 712545 Shi et al. Plasma Biomarkers of Alzheimer’s Disease FIGURE 3 | Forest plot of five proteins between MCIc and MCInc. MCI, mild cognitive impairment; MCIc, MCI converted to AD; MCInc, MCI non-converted to AD; RE, random effect; (A) A2M, alpha-2-macroglobulin; (B) CC4, complement component 4; (C) ApoA1, apolipoprotein A-I; (D) CLU, clusterin; (E) CFH, complement factor H; EMIF, European Medical Information Framework; AIBL, Australian Imaging, Biomarkers and Lifestyle Flagship Study of Ageing. et al., 2018; Karikari et al., 2020, 2021; Mattsson-Carlgren was 0.64 (Westwood et al., 2020). The combination of FGG et al., 2020; Sugarman et al., 2020; Thijssen et al., 2020; with age achieved an AUC of 0.69 in discriminating amyloid Chatterjee et al., 2021; Cicognola et al., 2021; Clark et al., status (Ashton et al., 2015). Despite lower AUC compared to 2021; Janelidze et al., 2021). However, the samples in these AD core markers, the three proteins can add extra value as studies were clinical trial populations, the performance of these they reflect different aspects of the disease. Therefore, building biomarkers in community-based populations was much worse. algorithms combining AD core markers with additional factors For example, one study in ADNI reported that among people such as demographic information as well as other potential who are cognitively impaired, plasma p-tau181 distinguished blood-based biomarkers are needed to add diagnostic value. amyloid-positives with a moderate area under curve (AUC) Here, our meta-analysis showed that A2M, FCN2, and FGG of 0.67 (Tosun et al., 2021), much lower than the AUCs of are good candidates for AD diagnosis and worthy of further 0.77–0.91 reported in some memory clinic cohorts (Karikari validation. et al., 2020; Thijssen et al., 2020). Likewise, another study Our study has two main limitations. First, the diagnosis using a small cohort found that plasma p-tau181 discriminated of AD in our previous studies was based on clinical diagnosis 20 cognitively normal amyloid-positive people from 31 amyloid- instead of using the ATN framework. Therefore, future studies negative people with an AUC of only 0.67 (Barthélemy et al., are needed to confirm these biomarkers in discriminating 2020). Therefore, these biomarkers cannot stand alone in pathologically confirmed AD-type dementia. Second, this predicting AD diagnosis or AD pathology. study is not a meta-analysis of the published literature Compared to AD core markers, the three markers (A2M, but rather a meta-analysis of the data we generated on FCN2, and FGG) obtained in this study had relatively lower our own. However, because our studies used the same AUC. For example, our previous study showed that plasma analytical platforms to measure these biomarkers, the A2M distinguished AD from controls with an AUC of 0.61 meta-analysis results were not affected by the difference (Hye et al., 2006). The AUC of FCN2 to classify amyloid status in platforms. Furthermore, there is a higher probability Frontiers in Aging Neuroscience | www.frontiersin.org 5 July 2021 | Volume 13 | Article 712545 Shi et al. Plasma Biomarkers of Alzheimer’s Disease of effective utility in practice as the current methods FUNDING (Luminex xMAP and ELISA) can be easily adopted in clinic settings. This research was conducted as part of the EMIF-AD project In conclusion, we demonstrate that A2M, FCN2, and FGG in which has received support from the Innovative Medicines blood have the potential use as screening tools to diagnose AD Initiative Joint Undertaking under EMIF grant agreement along with other promising blood biomarkers. Further validation no. 115372, resources of which are composed of financial in bigger, more epidemiologically sampled populations that contribution from the European Union’s Seventh Framework better represent the community populations are needed. Programme (FP7/2007-2013) and EFPIA companies’ in-kind contribution. The DESCRIPA study was funded by the European DATA AVAILABILITY STATEMENT Commission within the 5th framework program (QLRT- 2001-2455). The EDAR study was funded by the European The datasets will be made available by the authors to qualified Commission within the 5th framework program (contract # researchers upon reasonable request. Requests to access the 37670). The San Sebastian GAP study was partially funded by datasets should be directed to the corresponding author. the Department of Health of the Basque Government (allocation 17.0.1.08.12.0000.2.454.01.41142.001.H). The research at VIB-CMN was funded in part by the University of Antwerp ETHICS STATEMENT Research Fund. The studies involving human participants were reviewed and approved by all 23 medical Ethics Committees. The ACKNOWLEDGMENTS patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained We acknowledge the contribution of the personnel of the from the individual(s) for the publication of any potentially Genomic Service Facility at the VIB-U Antwerp Center for identifiable images or data included in this article. Molecular Neurology. AUTHOR CONTRIBUTIONS SUPPLEMENTARY MATERIAL LS carried out data analysis and interpretation as well as drafted The Supplementary Material for this article can be found the manuscript. All authors contributed to the article and online at: https://www.frontiersin.org/articles/10.3389/fnagi. approved the submitted version. 2021.712545/full#supplementary-material. REFERENCES Cicognola, C., Janelidze, S., Hertze, J., Zetterberg, H., Blennow, K., Mattsson- Carlgren, N., et al. (2021). Plasma glial fibrillary acidic protein detects Ahn, H. J., Zamolodchikov, D., Cortes-Canteli, M., Norris, E. H., Glickman, J. F., Alzheimer pathology and predicts future conversion to Alzheimer dementia and Strickland, S. (2010). Alzheimer’s disease peptide beta-amyloid interacts in patients with mild cognitive impairment. Alzheimers Res. Ther. 13:68. with fibrinogen and induces its oligomerization. Proc. Natl. Acad. Sci. U S A doi: 10.1186/s13195-021-00804-9 107, 21812–21817. doi: 10.1073/pnas.1010373107 Clark, C., Lewczuk, P., Kornhuber, J., Richiardi, J., Maréchal, B., Karikari, T. K., Ashton, N. J., Kiddle, S. J., Graf, J., Ward, M., Baird, A. L., Hye, A., et al. (2015). et al. (2021). Plasma neurofilament light and phosphorylated tau 181 as Blood protein predictors of brain amyloid for enrichment in clinical trials? biomarkers of Alzheimer’s disease pathology and clinical disease progression. Alzheimers Dement. 1, 48–60. doi: 10.1016/j.dadm.2014.11.005 Alzheimers Res. Ther. 13:65. doi: 10.1186/s13195-021-00805-8 Baird, A. L., Westwood, S., and Lovestone, S. (2015). Blood-based proteomic de Almeida, S. M., Shumaker, S. D., LeBlanc, S. K., Delaney, P., Marquie- biomarkers of Alzheimer’s disease pathology. Front. Neurol. 6:236. Beck, J., Ueland, S., et al. (2011). Incidence of post-dural puncture headache in doi: 10.3389/fneur.2015.00236 research volunteers. Headache 51, 1503–1510. doi: 10.1111/j.1526-4610.2011. Barthel, H., and Sabri, O. (2017). Clinical use and utility of amyloid imaging. 01959.x J. Nucl. Med. 58, 1711–1717. doi: 10.2967/jnumed.116.185017 Fujita, T., Matsushita, M., and Endo, Y. (2004). The lectin-complement pathway- Barthélemy, N. R., Horie, K., Sato, C., and Bateman, J. R. (2020). Blood plasma -its role in innate immunity and evolution. Immunol. Rev. 198, 185–202. phosphorylated-tau isoforms track CNS change in Alzheimer’s disease. J. Exp. doi: 10.1111/j.0105-2896.2004.0123.x Med. 217:e20200861. doi: 10.1084/jem.20200861 Hye, A., Lynham, S., Thambisetty, M., Causevic, M., Campbell, J., Byers, H. L., et al. Beach, T. G., Monsell, S. E., Phillips, L. E., and Kukull, W. (2012). Accuracy (2006). Proteome-based plasma biomarkers for Alzheimer’s disease. Brain 129, of the clinical diagnosis of Alzheimer disease at National Institute on Aging 3042–3050. doi: 10.1093/brain/awl279 Alzheimer Disease Centers, 2005–2010. J. Neuropathol. Exp. Neurol. 71, Hye, A., Riddoch-Contreras, J., Baird, A. L., Ashton, N. J., Bazenet, C., Leung, R., 266–273. doi: 10.1097/NEN.0b013e31824b211b et al. (2014). Plasma proteins predict conversion to dementia from prodromal Blennow, K. (2017). A review of fluid biomarkers for Alzheimer’s disease: moving disease. Alzheimers Dement. 10, 799.e2–807.e2. doi: 10.1016/j.jalz.2014. from CSF to blood. Neurol. Ther. 6, 15–24. doi: 10.1007/s40120-017-0073-9 05.1749 Chatterjee, P., Pedrini, S., Stoops, E., Goozee, K., Villemagne, V. L., Asih, P. R., Jack, C. R. Jr., Bennett, D. A., Blennow, K., Carrillo, M. C., Dunn, B., et al. (2021). Plasma glial fibrillary acidic protein is elevated in cognitively Haeberlein, S. B., et al. (2018). NIA-AA research framework: toward a normal older adults at risk of Alzheimer’s disease. Transl. Psychiatry 11:27. biological definition of Alzheimer’s disease. Alzheimers Dement. 14, 535–562. doi: 10.1038/s41398-020-01137-1 doi: 10.1016/j.jalz.2018.02.018 Chen, H., Lu, J., Chen, X., Yu, H., Zhang, L., Bao, Y., et al. (2012). Low serum levels Jack, C. R. Jr., Knopman, D. S., Jagust, W. J., Shaw, L. M., Aisen, P. S., of the innate immune component ficolin-3 is associated with insulin resistance Weiner, M. W., et al. (2010). Hypothetical model of dynamic biomarkers and predicts the development of type 2 diabetes. J. Mol. Cell Biol. 4, 256–257. of the Alzheimer’s pathological cascade. Lancet Neurol. 9, 119–128. doi: 10.1093/jmcb/mjs032 doi: 10.1016/S1474-4422(09)70299-6 Frontiers in Aging Neuroscience | www.frontiersin.org 6 July 2021 | Volume 13 | Article 712545 Shi et al. Plasma Biomarkers of Alzheimer’s Disease Janelidze, S., Berron, D., Smith, R., Strandberg, O., Proctor, N. K., Dage, J. L., diseases in two centres from the brains for dementia research (BDR) et al. (2021). Associations of plasma phospho-Tau217 levels with tau positron cohort. J. Neural Transm. 126, 327–337. doi: 10.1007/s00702-018- emission tomography in early Alzheimer disease. JAMA Neurol. 78, 149–156. 01967-w doi: 10.1001/jamaneurol.2020.4201 Shi, L., Baird, A. L., Westwood, S., Hye, A., Dobson, R., Thambisetty, M., Janson, J., Laedtke, T., Parisi, J. E., O’Brien, P., Petersen, R. C., and Butler, C. P. et al. (2018). A decade of blood biomarkers for Alzheimer’s disease research: (2004). Increased risk of type 2 diabetes in Alzheimer disease. Diabetes 53, an evolving field, improving study designs and the challenge of replication. 474–481. doi: 10.2337/diabetes.53.2.474 J. Alzheimers Dis. 62, 1181–1198. doi: 10.3233/JAD-170531 Karikari, T. K., Benedet, A. L., Ashton, N. J., Lantero Rodriguez, J., Snellman, A., Sugarman, M. A., Zetterberg, H., Blennow, K., Tripodis, Y., McKee, A. C., Suárez-Calvet, M., et al. (2021). Diagnostic performance and prediction of Stein, T. D., et al. (2020). A longitudinal examination of plasma neurofilament clinical progression of plasma phospho-tau181 in the Alzheimer’s Disease light and total tau for the clinical detection and monitoring of Alzheimer’s Neuroimaging Initiative. Mol. Psychiatry 26, 429–442. doi: 10.1038/s41380- disease. Neurobiol. Aging 94, 60–70. doi: 10.1016/j.neurobiolaging.2020. 020-00923-z 05.011 Karikari, T. K., Pascoal, T. A., Ashton, N. J., Janelidze, S., Benedet, A. L., Talbot, K., Wang, H. Y., Kazi, H., Han, L. Y., Bakshi, K. P., Stucky, A., et al. Rodriguez, J. L., et al. (2020). Blood phosphorylated tau 181 as a biomarker (2012). Demonstrated brain insulin resistance in Alzheimer’s disease patients for Alzheimer’s disease: a diagnostic performance and prediction modelling is associated with IGF-1 resistance, IRS-1 dysregulation and cognitive decline. study using data from four prospective cohorts. Lancet Neurol. 19, 422–433. J. Clin. Invest. 122, 1316–1338. doi: 10.1172/JCI59903 doi: 10.1016/S1474-4422(20)30071-5 Thambisetty, M., Simmons, A., Hye, A., Campbell, J., Westman, E., Zhang, Y., et al. Kiddle, S. J., Thambisetty, M., Simmons, A., Riddoch-Contreras, J., Hye, A., (2011). Plasma biomarkers of brain atrophy in Alzheimer’s disease. PLoS One Westman, E., et al. (2012). Plasma based markers of [11C] PiB-PET brain 6:e28527. doi: 10.1371/journal.pone.0028527 amyloid burden. PLoS One 7:e44260. doi: 10.1371/journal.pone.0044260 Thambisetty, M., Simmons, A., Velayudhan, L., Hye, A., Campbell, J., Zhang, Y., Kilpatrick, D. C., and Chalmers, D. J. (2012). Human L-ficolin (ficolin- et al. (2010a). Association of plasma clusterin concentration with severity, 2) and its clinical significance. J. Biomed. Biotechnol. 2012:138797. pathology and progression in Alzheimer disease. Arch. Gen. Psychiatry 67, doi: 10.1155/2012/138797 739–748. doi: 10.1001/archgenpsychiatry.2010.78 Klohs, J., Baltes, C., Princz-Kranz, F., Ratering, D., Nitsch, R. M., Knuesel, I., Thambisetty, M., Tripaldi, R., Riddoch-Contreras, J., Hye, A., An, Y., Campbell, J., et al. (2012). Contrast-enhanced magnetic resonance microangiography reveals et al. (2010b). Proteome-based plasma markers of brain amyloid-beta remodeling of the cerebral microvasculature in transgenic ArcAβ mice. deposition in non-demented older individuals. J. Alzheimers Dis. 22, J. Neurosci. 32, 1705–1713. doi: 10.1523/JNEUROSCI.5626-11.2012 1099–1109. doi: 10.3233/JAD-2010-101350 Kovacs, M. D. (2000). alpha2-macroglobulin in late-onset Alzheimer’s disease. Thijssen, E. H., La Joie, R., Wolf, A., Strom, A., Wang, P., Iaccarino, L., et al. Exp. Gerontol. 35, 473–479. doi: 10.1016/s0531-5565(00)00113-3 (2020). Diagnostic value of plasma phosphorylated tau181 in Alzheimer’s Koychev, I., Jansen, K., Dette, A., Shi, L., and Holling, H. (2021). Blood-based disease and frontotemporal lobar degeneration. Nat. Med. 26, 387–397. ATN biomarkers of Alzheimer’s disease: a meta-analysis. J. Alzheimers Dis. 79, doi: 10.1038/s41591-020-0762-2 177–195. doi: 10.3233/JAD-200900 Tosun, D., Veitch, D., Aisen, P., Jack, C. R. Jr., Jagust, W. J., Petersen, R. C., Li, R. X., Chen, H. B., Tu, K., Zhao, S. L., Zhou, H., Li, S. J., et al. (2008). et al. (2021). Detection of β-amyloid positivity in Alzheimer’s Disease Localized-statistical quantification of human serum proteome associated with Neuroimaging Initiative participants with demographics, cognition, MRI and type 2 diabetes. PLoS One 3:e3224. doi: 10.1371/journal.pone.0003224 plasma biomarkers. Brain Commun. 3:fcab008. doi: 10.1093/braincomms/ Lista, S., Faltraco, F., Prvulovic, D., and Hampel, H. (2013). Blood and plasma- fcab008 based proteomic biomarker research in Alzheimer’s disease. Prog. Neurobiol. Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. 101–102, 1–17. doi: 10.1016/j.pneurobio.2012.06.007 J. Stat. Soft. 36, 1–48.doi: 10.18637/jss.v036.i03 Manzine, P. R., Vatanabe, I. P., Peron, R., Grigoli, M. M., Pedroso, R. V., Westwood, S., Baird, A. L., Anand, S. N., Nevado-Holgado, A. J., Kormilitzin, A., Nascimento, C. M., et al. (2020). Blood-based biomarkers of Alzheimer’s Shi, L., et al. (2020). Validation of plasma proteomic biomarkers relating disease: the long and winding road. Curr. Pharm. Des. 26, 1300–1315. to brain amyloid burden in the EMIF-Alzheimer’s disease multimodal doi: 10.2174/1381612826666200114105515 biomarker discovery cohort. J. Alzheimers Dis. 74, 213–225. doi: 10.3233/JAD- Mattsson-Carlgren, N., Janelidze, S., Palmqvist, S., Cullen, N., Svenningsson, A. L., 190434 Strandberg, O., et al. (2020). Longitudinal plasma p-tau217 is Westwood, S., Baird, A. L., Hye, A., Ashton, N. J., Nevado-Holgado, A. J., increased in early stages of Alzheimer’s disease. Brain 143, 3234–3241. Anand, S. N., et al. (2018). Plasma protein biomarkers for the prediction of doi: 10.1093/brain/awaa286 csf amyloid and tau and [18 F]-flutemetamol PET scan result. Front. Aging McKhann, G., Drachman, D., Folstein, M., Katzman, R., Price, D., and Neurosci. 10:409. doi: 10.3389/fnagi.2018.00409 Stadlan, M. E. (1984). Clinical diagnosis of Alzheimer’s disease: report of Westwood, S., Leoni, E., Hye, A., Lynham, S., Khondoker, M. R., Ashton, N. J., et al. the NINCDS-ADRDA Work Group under the auspices of Department of (2016). Blood-based biomarker candidates of cerebral amyloid using PiB PET Health and Human Services Task Force on Alzheimer’s Disease. Neurology 34, in non-demented elderly. J. Alzheimers Dis. 52, 561–572. doi: 10.3233/JAD- 939–944. doi: 10.1212/wnl.34.7.939 151155 McKhann, G. M., Knopman, D. S., Chertkow, H., Hyman, B. T., Jack, C. R. Jr., Westwood, S., Liu, B., Baird, A. L., Anand, S., Nevado-Holgado, A. J., Newby, D., Kawas, C. H., et al. (2011). The diagnosis of dementia due to Alzheimer’s et al. (2017). The influence of insulin resistance on cerebrospinal fluid and disease: recommendations from the National Institute on Aging-Alzheimer’s plasma biomarkers of Alzheimer’s pathology. Alzheimers Res. Ther. 9:31. association workgroups on diagnostic guidelines for Alzheimer’s disease. doi: 10.1186/s13195-017-0258-6 Alzheimers Dement. 7, 263–269. doi: 10.1016/j.jalz.2011.03.005 Zetterberg, H., and Burnham, C. S. (2019). Blood-based molecular biomarkers Nakamura, A., Kaneko, N., Villemagne, V. L., Kato, T., Doecke, J., Doré, V., for Alzheimer’s disease. Mol. Brain 12:26. doi: 10.1186/s13041-019- et al. (2018). High performance plasma amyloid-β biomarkers for Alzheimer’s 0448-1 disease. Nature 554, 249–254. doi: 10.1038/nature25456 Zhang, X., Hu, Y., Shen, J., Zeng, H., Lu, J., Li, L., et al. (2016). Low levels of Ryu, J. K., and McLarnon, G. J. (2009). A leaky blood-brain barrier, fibrinogen ficolin-3 are associated with diabetic peripheral neuropathy. Acta Diabetol. 53, infiltration and microglial reactivity in inflamed Alzheimer’s disease brain. 295–302. doi: 10.1007/s00592-015-0780-6 J. Cell. Mol. Med. 13, 2911–2925. doi: 10.1111/j.1582-4934.2008.00434.x Sattlecker, M., Kiddle, S. J., Newhouse, S., Proitsi, P., Nelson, S., Williams, S., et al. Conflict of Interest: SL is named as an inventor on biomarker intellectual (2014). Alzheimer’s disease biomarker discovery using SOMAscan multiplexed property protected by Proteome Sciences and Kings College London unrelated protein technology. Alzheimers Dement. 10, 724–734. doi: 10.1016/j.jalz.2013. to the current study and within the past 5 years has advised for Optum labs, 09.016 Merck, SomaLogic and been the recipient of funding from AstraZeneca and other Selvackadunco, S., Langford, K., Shah, Z., Hurley, S., Bodi, I., King, A., et al. (2019). companies via the IMI funding scheme. SL is employed by company Janssen. HZ Comparison of clinical and neuropathological diagnoses of neurodegenerative has served at scientific advisory boards for Alector, Denali, Roche Diagnostics, Frontiers in Aging Neuroscience | www.frontiersin.org 7 July 2021 | Volume 13 | Article 712545 Shi et al. Plasma Biomarkers of Alzheimer’s Disease Wave, Samumed, Siemens Healthineers, Pinteon Therapeutics and CogRx, has The remaining authors declare that the research was conducted in the absence of given lectures in symposia sponsored by Cellectricon, Fujirebio, Alzecure and any commercial or financial relationships that could be construed as a potential Biogen, and is a co-founder of Brain Biomarker Solutions in Gothenburg AB conflict of interest. (BBS), which is a part of the GU Ventures Incubator Program (all unrelated to this study). AL has served at scientific advisory boards of Fujirebio Europe, Eli Copyright © 2021 Shi, Buckley, Bos, Engelborghs, Sleegers, Frisoni, Wallin, Lilly, Novartis, Nutricia and Otsuka and is the inventor of a patent on synaptic Lléo, Popp, Martinez-Lage, Legido-Quigley, Barkhof, Zetterberg, Visser, Bertram, markers in CSF (all unrelated to this study). JP has served at scientific advisory Lovestone and Nevado-Holgado. This is an open-access article distributed under the boards of Fujirebio Europe, Eli Lilly and Nestlé Institute of Health Sciences, all terms of the Creative Commons Attribution License (CC BY). The use, distribution unrelated to this study. SE has received unrestricted research grants from Janssen or reproduction in other forums is permitted, provided the original author(s) and Pharmaceutica and ADx Neurosciences and has served at scientific advisory the copyright owner(s) are credited and that the original publication in this journal boards of Biogen, Eisai, Novartis, Nutricia/Danone and Roche, all unrelated to is cited, in accordance with accepted academic practice. No use, distribution or this study. reproduction is permitted which does not comply with these terms. Frontiers in Aging Neuroscience | www.frontiersin.org 8 July 2021 | Volume 13 | Article 712545