Abstract
Cerebral amyloid angiopathy (CAA), limbic-predominant age-related TDP-43 encephalopathy neuropathological change (LATE-NC) and Lewy bodies occur in the absence of clinical and neuropathological Alzheimer’s disease, but their prevalence and severity dramatically increase in Alzheimer’s disease. To investigate how plaques, tangles, age and apolipoprotein E ε4 (APOE ε4) interact with co-pathologies in Alzheimer’s disease, we analysed 522 participants ≥50 years of age with and without dementia from the Center for Neurodegenerative Disease Research (CNDR) autopsy program and 1340 participants in the National Alzheimer's Coordinating Center (NACC) database. Consensus criteria were applied for Alzheimer’s disease using amyloid phase and Braak stage. Co-pathology was staged for CAA (neocortical, allocortical, and subcortical), LATE-NC (amygdala, hippocampal, and cortical), and Lewy bodies (brainstem, limbic, neocortical, and amygdala predominant). APOE genotype was determined for all CNDR participants. Ordinal logistic regression was performed to quantify the effect of independent variables on the odds of having a higher stage after checking the proportional odds assumption. We found that without dementia, increasing age associated with all pathologies including CAA (odds ratio 1.63, 95% confidence interval 1.38–1.94, P < 0.01), LATE-NC (1.48, 1.16–1.88, P < 0.01), and Lewy bodies (1.45, 1.15–1.83, P < 0.01), but APOE ε4 only associated with CAA (4.80, 2.16–10.68, P < 0.01). With dementia, increasing age associated with LATE-NC (1.30, 1.15–1.46, P < 0.01), while Lewy bodies associated with younger ages (0.90, 0.81–1.00, P = 0.04), and APOE ε4 only associated with CAA (2.36, 1.52–3.65, P < 0.01). A longer disease course only associated with LATE-NC (1.06, 1.01–1.11, P = 0.01). Dementia in the NACC cohort associated with the second and third stages of CAA (2.23, 1.50–3.30, P < 0.01), LATE-NC (5.24, 3.11–8.83, P < 0.01), and Lewy bodies (2.41, 1.51–3.84, P < 0.01). Pathologically, increased Braak stage associated with CAA (5.07, 2.77–9.28, P < 0.01), LATE-NC (5.54, 2.33–13.15, P < 0.01), and Lewy bodies (4.76, 2.07–10.95, P < 0.01). Increased amyloid phase associated with CAA (2.27, 1.07–4.80, P = 0.03) and Lewy bodies (6.09, 1.66–22.33, P = 0.01). In summary, we describe widespread distributions of CAA, LATE-NC and Lewy bodies that progressively accumulate alongside plaques and tangles in Alzheimer’s disease dementia. CAA interacted with plaques and tangles especially in APOE ε4 positive individuals; LATE-NC associated with tangles later in the disease course; most Lewy bodies associated with moderate to severe plaques and tangles.
Keywords: Alzheimer’s disease, co-pathology, CAA, TDP-43, a-synuclein
See Tomé and Thall (doi:10.1093/brain/awab027) for a scientific commentary on this article.
Cerebral amyloid angiopathy, TDP-43 inclusions and Lewy bodies frequently, but not universally, arise in Alzheimer’s disease dementia. In a neuropathological analysis of almost 2000 individuals, Robinson et al. examine when these co-pathologies occur dependent on plaques, tangles, APOE ε4 status and age.
See Tomé and Thall (doi:10.1093/brain/awab027) for a scientific commentary on this article.
Introduction
The accumulation of plaques and tangles in the human brain are the pathological hallmarks of Alzheimer’s disease, but it has long been recognized that additional pathologies frequently co-occur including cerebral amyloid angiopathy (CAA), TDP-43 inclusions and Lewy body pathology (Montine et al., 2012). CAA is pathologically defined by the presence of amyloid-β protein deposition within the walls of meningeal and parenchymal blood vessels. TDP-43 pathology is primarily intraneuronal accumulations of phosphorylated TDP-43, while Lewy bodies are aggregates of phosphorylated α-synuclein in neuronal cytoplasmic inclusions.
CAA, TDP-43 and Lewy body pathologies have defined stereotypic patterns of accumulations. CAA is first seen in leptomeningeal and neocortical vessels before extending to medial temporal lobe and cerebellar vessels and finally occurring in the basal ganglia and thalamus (Thal et al., 2003). TDP-43 pathology is commonly found first in the amygdala before extending to regions of the hippocampal formation and finally spreading outward to affect the rest of the brain including the neocortex in a pattern called limbic-predominant age-related TDP-43 encephalopathy neuropathological change (LATE-NC) (Nelson et al., 2019). Lewy bodies are stereotypically observed in the brainstem before they accumulate in a limbic pattern of deposition with the most severe stage showing Lewy body accumulation in the neocortex; a fourth Lewy body pattern is described as amygdala predominant with Lewy bodies concentrated in the medial temporal lobe without significant involvement of the brainstem or neocortex (Leverenz et al., 2008).
At least three theories have been put forth to account for the multiple proteinopathies that occur in Alzheimer’s disease. First, healthy ageing without any cognitive decline is frequently associated with the accumulation of multiple, age-related proteinopathies including neurofibrillary tangles, plaques, CAA, LATE-NC and Lewy bodies (Saito et al., 2004; Keage et al., 2009; Wennberg et al., 2019). Disentangling the increased prevalence in Alzheimer’s disease from the pathologies present in normal ageing remains challenging as increasing age and the presence of an apolipoprotein E ε4 (APOE ε4) allele are the biggest risk factors for Alzheimer’s disease and these same factors influence CAA, LATE-NC and Lewy bodies (Zaccai et al., 2015; Shinohara et al., 2016; Wennberg et al., 2018). According to this theory, co-pathologies observed in the Alzheimer’s disease brain may simply be incidental pathologies that would have been present had Alzheimer’s disease not occurred. Second, co-pathologies may also lower the threshold for, and contribute to the clinical symptoms of Alzheimer’s disease as primary neurodegenerative processes that are subsequently overwhelmed by Alzheimer’s disease pathology (Jellinger, 2002; Kapasi et al., 2017; Nelson et al., 2019). Third, co-pathologies occur after interacting concurrently or downstream of plaque and tangle accumulation (James et al., 2016; Spires-Jones et al., 2017; de Flores et al., 2020).
This retrospective clinical and pathological study uses an Alzheimer’s Disease Center cohort and a larger cohort obtained from the National Alzheimer's Coordinating Center (NACC) to report in patients with dementia that CAA, LATE-NC and Lewy pathology are interacting co-pathologies at moderate or severe stages that function as additional correlates of cognitive decline in Alzheimer’s disease, while, in the absence of dementia, each is best described as an age-associated, incidental co-pathology. To our knowledge, our neuroanatomic study is the largest cohort reporting these correlations using amyloid-β, tau, TDP-43, and α-synuclein immunohistochemistry to stage plaque, CAA, tangle, LATE-NC and Lewy pathology.
Materials and methods
Participants
We analysed a Center for Neurodegenerative Disease Research (CNDR) cohort of 522 participants 50 years of age or older with and without clinical dementia. Of 665 participants referred to the CNDR autopsy program by clinicians in several clinical cores at the University of Pennsylvania, 628 were individuals with a primary neuropathological diagnosis of Alzheimer’s disease or primary ageing-related tauopathy (PART), 535 had a known APOE genotype, and 522 had enough tissue available to stage the burden of Alzheimer’s disease, CAA, LATE-NC and Lewy body pathology as described below. Only dementia participants with a clinical diagnosis of probable (n = 336) or possible (n = 33) Alzheimer’s disease, vascular dementia (n = 6), posterior cortical atrophy (n = 3), or dementia not otherwise specified (n = 9) were included. Many of the patients with no cognitive impairment were recruited at time of death due to other causes. Informed consent for autopsy was obtained in accordance with state laws and protocols approved by the University of Pennsylvania.
Further analysis was done using a similar NACC cohort of dementia participants with clinical probable or possible Alzheimer’s disease, and non-dementia participants, with all participants having a primary pathological diagnosis of PART or Alzheimer’s disease. The NACC datasets used in our study were the Uniform Data Set and the Neuropathology Data Set and included all participants as of the March 2020 data freeze. Of 10 485 NACC participants who came to autopsy at age 50 years or older, 2760 had available immunohistochemically determined amyloid phase and Lewy body measures that included the amygdala predominant pattern. Of these, we excluded 468 because of a primary non-Alzheimer’s disease pathological diagnosis (usually frontotemporal lobar degeneration with tau or TDP-43 inclusions). We also excluded 657 dementia patients without a clinical diagnosis of probable or possible Alzheimer’s disease. Another 295 were excluded because their last clinical assessment occurred ≥2 years prior to death. The final NACC cohort included 1340 participants including 968 with dementia and 372 without dementia. NACC participants are recruited from individual Alzheimer’s disease centres including the one at CNDR, with the result that 57 participants were present in both cohorts.
Neuropathology
Sixteen regions are routinely examined in the CNDR neuropathology evaluations as described previously (Toledo et al., 2014). Each region was assigned a semiquantitative score, i.e. none, rare, mild, moderate, or severe for individual lesions based on immunohistochemistry against tau (mouse antibody PHF1, a gift from Dr Peter Davies), amyloid-β (mouse antibody NAB228, generated in CNDR), α-synuclein (mouse antibody Syn303, generated at the CNDR) and TDP-43 (rat antibody 1D3, a gift from Dr Manuela Neumann). All cases are reviewed by a board-certified neuropathologist (J.Q.T. and/or E.B.L.) for quality assurance and accurate grading.
Tau-positive neurofibrillary tangles and amyloid-β positive plaques were assigned a Braak stage, Amyloid phase and CERAD score to determine the level of Alzheimer’s disease neuropathological change (ADNC) according to consensus criteria (Montine et al., 2012). Cases were diagnosed with a low, intermediate, or high level of ADNC or PART. PART was defined as Braak stage IV or less without plaques (Crary et al., 2014).
CAA stages were assigned after screening middle frontal cortex, superior temporal cortex, angular gyrus and occipital cortex for amyloid-β-positive CAA in meningeal and parenchymal blood vessels. For cases with neocortical CAA, additional regional analysis determined the extent of CAA in the brain, with increasing CAA stages also reflecting increased neocortical CAA severity (Thal et al., 2003; Love et al., 2014). In stage 1, CAA is limited to the neocortex and the mean neocortical scores are rare to mild; in stage 2, CAA extends into the amygdala, hippocampus, anterior cingulate and/or cerebellum and the mean neocortical scores are mild to moderate; in stage 3, CAA additionally involves the caudate, putamen and/or thalamus and the mean neocortical scores are moderate to severe. Consistent with other studies, we found that the occipital cortex is particularly susceptible to CAA (Love et al., 2014). In 27.6% (24/87) of CAA stage 1 cases, CAA was limited to the occipital cortex. In the NACC cohort, CAA stage information was not available, but CAA was measured on a 0–3 severity scale. The presence or absence of CAA in capillaries was not available for all cases in either the CNDR or NACC cohorts and was not analysed (Thal et al., 2002; Love et al., 2014).
LATE-NC was observed in a progression similar to previous studies (Josephs et al., 2016). In stage 1, TDP-43 pathology is limited to the amygdala. In stage 2, it extends into entorhinal cortex and/or subiculum followed by the dentate gyrus of the hippocampus. In stage 3, additional pathology extends into orbital frontal cortex, superior or middle temporal cortex and/or anterior cingulate followed by substantia nigra, inferior olive and midbrain tectum and finally middle frontal cortex. In the NACC cohort, TDP-43 pathology was available for a subset of the cohort (58.5%, 784/1340) as stage 1 (amygdala), stage 2 (hippocampus and/or entorhinal cortex), and stage 3 (neocortex).
Lewy pathology was classified into brainstem (stage 1), limbic (stage 2), neocortical (stage 3), or amygdala predominant distributions for both cohorts as described in the ‘Fourth Consensus Report Of The DLB Consortium’ (McKeith et al., 2017).
Genetics
APOE allele status was defined using two single nucleotide polymorphisms—rs7412 and rs429358—which were genotyped by TaqMan® allelic discrimination assays (Thermo Fisher) in the CNDR cohort. For the NACC cohort, APOE allele status was not obtained.
Clinical
A Clinical Dementia Rating® dementia staging instrument global score (CDR) was available for 201 CNDR participants. 164 additional CDRs were deduced from Mini-Mental State Examination (MMSE) scores according to the formula: MMSE 28–30 = CDR 0; MMSE 25–27 = CDR 0.5; MMSE 19–24 = CDR 1, MMSE 10–18 = CDR 2, MMSE 0–9 = CDR 3 (Perneczky et al., 2006). In the CNDR cohort, CDRs of 3 and CDRs of 0–2 available within 2 years of autopsy were analysed for 65.7% of participants (343/522). For the NACC cohort, a CDR was available for all cases.
Statistical analysis
R 3.4.1 was used for all regression analyses. Unadjusted analysis of demographic features between participants with and without dementia was performed using Pearson’s χ2 tests. Ordinal logistic regression (proportional odds model) was performed to quantify the effect of independent variables on the odds of having a larger pathology score versus a smaller one. The proportional odds assumption was tested by dichotomizing the outcome using each score as a cut-off and performing logistic regression using one independent variable at a time. The independent variable satisfied the proportional odds assumption if the confidence intervals for all outcome values overlapped. Binary logistic regression was used to build a model of dementia to assess each measure’s contribution. All statistical tests were two-sided. Statistical significance was set at <0.05 level.
Results
Participant characteristics
Of the 522 participants in the CNDR cohort, 387 (74.1%) had clinically diagnosed dementia at the time of death and 135 (25.9%) were without dementia, including 21 (4.0%) with cognitive impairment but no dementia and 114 (21.8%) without cognitive impairment (Table 1). The without dementia group were demographically similar to the dementia group with nearly identical ages at death (76.7 and 79.1, respectively) and with same the proportions of females (51.9% and 54.0%, respectively). Genetically and pathologically the two groups were significantly different. Without dementia, the most frequent APOE genotype was ε3/ε3 (54.1%), while with dementia one APOE ε4 allele was present in 45.7% (χ2 = 51.7, df = 3). Without dementia, a low level of ADNC was common (52.6%), but with dementia, a high level of ADNC accounted for the majority (82.9%) (χ2 = 293.8, df = 3). CAA accumulations were found in less than half (43.7%) of the without dementia group, but with dementia, stage 2 was the most frequent distribution (53.7%) (χ2 = 85.8, df = 3). LATE-NC was limited to 13.3% of the without dementia group but was present in 47.8% of the dementia group (χ2 = 51.5, df = 3). Lewy bodies were limited to 13.3% of the without dementia group but were identified in 54.3% of the dementia group (χ2 = 39.7, df = 4).
Table 1.
Participant characteristics
| Characteristic | Total (n = 522) | Without dementia (n = 135) | With dementia (n = 387) |
|---|---|---|---|
| Cognitive status, n (%) | |||
| No impairment | 114 (21.8) | 114 (84.4) | – |
| Cognitive impairment | 21 (4.0) | 21 (15.6) | – |
| Dementia | 387 (74.1) | – | 387 (100) |
| Age at death, years, mean (SD) | 78.5 (10.1) | 76.7 (12.4) | 79.1 (9.1) |
| Disease duration, years, mean (SD) | 9.5 (4.3) | – | 9.5 (4.3) |
| Age at onset, years, mean (SD) | 69.5 (9.5) | – | 69.5 (9.5) |
| Female sex, n (%) | 279 (53.4) | 70 (51.9) | 209 (54.0) |
| APOE allele, n (%) | |||
| ε2/ε2 or ε2/ε3 | 35 (6.7) | 19 (14.1) | 16 (4.1) |
| ε3/ε3 | 194 (37.2) | 73 (54.1) | 121 (31.3) |
| ε4/ε3 or ε4/ε2 | 215 (41.2) | 38 (28.1) | 177 (45.7) |
| ε4/ε4 | 78 (14.9) | 5 (3.7) | 73 (18.9) |
| ADNC, n (%) | |||
| PART | 21 (4.0) | 21 (15.6) | – |
| Low | 82 (15.7) | 72 (52.6) | 11 (2.8) |
| Intermediate | 80 (15.3) | 25 (18.5) | 55 (14.2) |
| High | 339 (64.9) | 18 (13.3) | 321 (82.9) |
| CAA stage, n (%) | |||
| Absent | 144 (27.6) | 76 (56.3) | 68 (17.6) |
| 1 | 87 (16.7) | 25 (18.5) | 62 (16.0) |
| 2 | 233 (44.6) | 25 (18.5) | 208 (53.7) |
| 3 | 58 (11.1) | 9 (6.7) | 49 (12.7) |
| LATE-NC stage, n (%) | |||
| Absent | 319 (61.1) | 117 (86.7) | 202 (52.2) |
| 1 | 53 (10.2) | 8 (5.9) | 45 (11.6) |
| 2 | 107 (20.5) | 7 (5.2) | 100 (25.8) |
| 3 | 43 (8.2) | 3 (2.2) | 40 (10.3) |
| Lewy bodies, n (%) | |||
| Absent | 294 (56.3) | 117 (86.7) | 177 (45.7) |
| 1 | 48 (9.2) | 9 (6.7) | 39 (10.1) |
| 2 | 51 (9.8) | 5 (3.7) | 46 (11.9) |
| 3 | 32 (6.1) | 1 (0.7) | 31 (8.0) |
| Amygdala | 97 (18.6) | 3 (2.2) | 94 (24.3) |
The NACC cohort was demographically similar to the CNDR cohort (Supplementary Table 1). In the CNDR cohort, 74.1% had dementia compared to 72.2% of the NACC cohort, and 80.2% of the CNDR cohort were diagnosed with an intermediate or high level of ADNC compared to 79.8%% of the NACC cohort. Clinicopathological correlations were also similar. In the CNDR cohort, 13.3% without dementia had a high level of ADNC compared to 82.9% with dementia. In the NACC cohort, 11.8% without dementia had a high level of ADNC compared to 78.7% with dementia. The major difference between the two cohorts was the age at death, which was higher in the NACC cohort. The mean (standard deviation, SD) age of death for NACC participants was 87.3 (8.5) years without dementia and 81.2 (10.6) years with dementia compared to 76.7 (12.4) years without dementia and 79.1 (9.1) years with dementia in the CNDR cohort.
Pathological correlations with clinical dementia
As measured by increasing CDR, progressive cognitive decline correlated with the increasing prevalence and severity of ADNC, CAA, LATE-NC, and Lewy body pathology (Fig. 1). In the NACC cohort, neuropathological Alzheimer’s disease, defined by the presence of an intermediate or high level of ADNC, was present in 35.6% of individuals without impairment (CDR 0) and in 83.5% of the mild dementia cases (CDR 1). CAA was the most common co-pathology, increasing from 42.2% with no impairment to 66.9% with mild dementia with the bulk of the increase reflecting severe CAA. Both LATE-NC and Lewy bodies were less common, but still prevalent co-pathologies. LATE-NC prevalence went from 13.5% with no impairment to 34.2% with mild dementia, including substantial increases in hippocampal and neocortical TDP-43 pathology. Lewy body prevalence went from 15.5% with no impairment to 33.7% with mild dementia including large increases in amygdala, limbic and neocortical Lewy bodies. Similar results were observed in the CNDR cohort. These increases in pathology prevalence and severity were even more pronounced with severe dementia (CDR 3), which accounted for the majority of dementia cases (66.7% and 59.6% of the CNDR and NACC cohorts, respectively).
Figure 1.
Clinicopathological correlations along the Alzheimer’s disease spectrum. Increasing cognitive decline as measured by the CDR scale correlates with increasing levels of ADNC, CAA, LATE-NC and Lewy body pathology in two different cohorts. Intermediate and high ADNC and CAA were uncommon in CDR 0 cases but were observed in a majority with mild dementia (CDR 1). Almost half of participants with moderate to severe dementia (CDR 2 and 3) had additional LATE-NC and Lewy body pathologies. CDR scores were obtained within 2 years of death for 343 CNDR participants and 1340 NACC participants.
To assess the contribution of each co-pathology to dementia, we created binary logistic regression models that included neuropathological Alzheimer’s disease and age at death as variables (Table 2). In both cohorts, neuropathological Alzheimer’s disease was the major driver of dementia (P < 0.01). For the co-pathologies, the presence of each at an early stage did not relate to dementia, while more widespread distributions did. In the NACC cohort, CAA severity 2/3, LATE-NC stages 2/3, Lewy body stages 2/3 and the presence of amygdala Lewy bodies associated with dementia (P < 0.01). In the CNDR cohort, CAA stages 2/3 and LATE-NC stages 2/3 associated with dementia (P < 0.01), but Lewy body stages 2/3 and amygdala Lewy bodies were not significant (p-values > 0.1). In summary, intermediate or high levels of ADNC associated with dementia, but intermediate or high levels of co-pathologies also associated with dementia.
Table 2.
Impact of co-pathologies on dementia risk
| Measure | CNDR |
NACC |
||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P-value | OR | 95% CI | P-value | |
| Age at death | 0.83 | 0.71–0.97 | 0.02 | 0.77 | 0.71–0.84 | <0.01 |
| ADNC | ||||||
| Intermediate/high | 34.01 | 15.37–75.26 | <0.01 | 7.53 | 5.37–10.55 | <0.01 |
| CAAa | ||||||
| 1 | 1.13 | 0.48–2.63 | 0.78 | 1.18 | 0.83–1.69 | 0.35 |
| 2/3 | 3.17 | 1.55–6.47 | <0.01 | 2.23 | 1.50–3.30 | <0.01 |
| LATE-NCb | ||||||
| 1 | 1.73 | 0.59–5.05 | 0.31 | 2.45 | 0.68–8.81 | 0.17 |
| 2/3 | 4.60 | 1.91–11.10 | <0.01 | 5.24 | 3.11–8.83 | <0.01 |
| Lewy bodies | ||||||
| 1 | 2.16 | 0.71–6.56 | 0.17 | 0.64 | 0.32–1.28 | 0.21 |
| 2/3 | 1.91 | 0.79–4.59 | 0.15 | 2.41 | 1.51–3.84 | <0.01 |
| Amygdala | 2.63 | 0.62–11.14 | 0.19 | 4.05 | 1.30–12.58 | <0.01 |
CAA is staged in the CNDR cohort and represents cortical severity in the NACC cohort.
Available for 784 participants (58.5%) in the NACC cohort.
Co-pathology associations with plaques and tangles
In the NIA-AA criteria, amyloid phase represents the distribution of plaques, Braak stage the distribution of tangles, and CERAD scores the severity of neocortical neuritic plaques. These pathologies are highly correlated in the Alzheimer’s brain; therefore, increases in ADNC level (Fig. 2) or amyloid phase and Braak stage (Supplementary Fig. 3) show stepwise increases in co-pathology prevalence and severity. Nonetheless, we next asked which Alzheimer’s disease pathology associated with each co-pathology (Table 3). By ordinal logistic regression, both increased amyloid phase and CERAD plaque scores correlated with increased CAA, Lewy body stage, and amygdala-predominant Lewy bodies (P < 0.03) in both cohorts. Neither plaque measure correlated with LATE-NC (P > 0.06) in either cohort. Increased Braak stage correlated with all three co-pathologies in the CNDR cohort (P < 0.01), but no co-pathology associated with increased Braak stages in the NACC cohort (P > 0.06).
Figure 2.
Co-pathology distribution and severity is related to Alzheimer’s disease pathology. In the CNDR and NACC cohorts, CAA, LATE-NC and Lewy body prevalence increase at higher levels of ADNC, primarily driven by increases in moderate to severe levels of deposits. In both cohorts, CAA prevalence approximately doubles from low to high ADNC. LATE-NC prevalence between PART and high ADNC is 4-fold higher in the CNDR cohort and 3-fold higher in the NACC cohort. Lewy body prevalence between low and high ADNC is 5-fold higher in the CNDR cohort and more than doubles in the NACC cohort, reflecting increases in both the limbic/neocortical and amygdala-predominant patterns. In the NACC cohort, both LATE-NC and Lewy body pathology are more prevalent in PART and low ADNC cases, perhaps reflecting age-related incidental levels of each as the mean (SD) age of death for the NACC PART/low ADNC cohort was 86.0 (9.6) years compared to the equivalent CNDR cohort at 74.2 (11.7) years. AD = Alzheimer’s disease.
Table 3.
Co-pathology associations with Alzheimer’s pathology
| Measure | CNDR |
NACC |
||||
|---|---|---|---|---|---|---|
| ORa | 95% CI | P-value | OR | 95% CI | P-value | |
| Amyloid phase | ||||||
| CAA | 2.27 | 1.07–4.80 | 0.03 | 3.95 | 2.78–5.60 | <0.01 |
| LATE-NC | 1.71 | 0.64–4.58 | 0.28 | 1.37 | 0.83–2.26 | 0.21 |
| Lewy body stage | 6.09 | 1.66–22.33 | 0.01 | 2.31 | 1.52–3.49 | <0.01 |
| Amygdala Lewy bodiesb | – | – | – | 3.51 | 1.42–8.71 | 0.01 |
| Braak stage | ||||||
| CAA | 5.07 | 2.77–9.28 | <0.01 | 1.42 | 0.98–2.05 | 0.06 |
| LATE-NC | 5.54 | 2.33–13.15 | <0.01 | 1.13 | 0.68–1.87 | 0.65 |
| Lewy body stage | 4.76 | 2.07–10.95 | <0.01 | 1.03 | 0.68–1.56 | 0.88 |
| Amygdala Lewy bodies | – | – | – | 1.84 | 0.77–4.39 | 0.17 |
| CERAD plaques | ||||||
| CAA | 1.94 | 1.18–3.19 | 0.01 | 2.74 | 2.02–3.71 | <0.01 |
| LATE-NC | 1.14 | 0.64–2.00 | 0.66 | 1.56 | 0.98–2.48 | 0.06 |
| Lewy body stage | 1.51 | 0.87–2.63 | 0.14 | 1.76 | 1.23–2.50 | <0.01 |
| Amygdala Lewy bodies | – | – | – | 2.68 | 1.41–5.08 | <0.01 |
Ordinal logistic regression analysis using ABC scores as predictors, stages 2/3 versus 0/1 (reference group), and the co-pathology as the outcome.
Amygdala Lewy bodies are not analysed in the CNDR cohort because they were not observed in the reference group.
Co-pathology prevalence varies by age, APOE allele and dementia status
As increasing age and APOE allele status are commonly recognized risk factors for neuropathology and clinical outcomes, we next analysed if these characteristics also influence co-pathology prevalence (Table 4 and Supplementary Figs 1 and 2). In the CNDR cohort without dementia, all pathologies were associated with increasing age (P < 0.01). With dementia, LATE-NC associated with increasing age (P < 0.01), while the Lewy body measures, by stage and amygdala predominance, were both more likely at younger ages (P < 0.05). The presence of an APOE ε4 allele associated with CAA in cases with and without dementia (P < 0.01) but did not associate with any other co-pathology (P > 0.20). The presence of an APOE ε2 allele did not associate with any co-pathology, nor was it protective against any co-pathology (P > 0.20), although it trended towards significance for being protective for Lewy bodies (P = 0.05). Similar results were obtained for the APOE alleles when the patients with and without dementia were analysed as a whole cohort (Supplementary Table 2). For cases with dementia, LATE-NC associated with a longer dementia duration (P = 0.01), but CAA and Lewy bodies were not more prevalent with longer disease courses (P > 0.10). This analysis was not performed on the NACC cohort.
Table 4.
Co-pathologies vary with age, APOE ε4 allele, and dementia status
| With dementia, n = 387 |
Without dementia, n = 135 |
|||||
|---|---|---|---|---|---|---|
| OR a | 95% CI | P-value | OR | 95% CI | P-value | |
| Age at deathb | ||||||
| CAA | 1.01 | 0.90–1.14 | 0.86 | 1.63 | 1.38–1.94 | <0.01 |
| LATE-NC | 1.30 | 1.15–1.46 | <0.01 | 1.48 | 1.16–1.88 | <0.01 |
| Lewy body stage | 0.90 | 0.81–1.00 | 0.04 | 1.45 | 1.15–1.83 | <0.01 |
| Amygdala Lewy bodiesc | 0.86 | 0.76–0.98 | 0.03 | – | – | – |
| APOE ε4d | ||||||
| CAA | 2.36 | 1.52–3.65 | <0.01 | 4.80 | 2.16–10.68 | <0.01 |
| LATE-NC | 1.15 | 0.75–1.75 | 0.52 | 0.83 | 0.26–2.59 | 0.74 |
| Lewy body stage | 1.08 | 0.73–1.62 | 0.69 | 1.91 | 0.68–5.39 | 0.22 |
| Amygdala Lewy bodies | 1.02 | 0.62–1.70 | 0.93 | – | – | – |
| APOE ε2e | ||||||
| CAA | 1.53 | 0.54–4.29 | 0.42 | 1.71 | 0.69–4.25 | 0.25 |
| LATE-NC | 0.68 | 0.28–1.69 | 0.41 | 1.30 | 0.38–4.53 | 0.68 |
| Lewy body stage | 0.40 | 0.16–1.01 | 0.05 | 1.41 | 0.40–5.03 | 0.59 |
| Amygdala Lewy bodies | 0.53 | 0.15–1.86 | 0.32 | – | – | – |
| Dementia durationf | ||||||
| CAA | 1.04 | 0.99–1.10 | 0.12 | – | – | – |
| LATE-NC | 1.06 | 1.01–1.11 | 0.01 | – | – | – |
| Lewy body stage | 1.02 | 0.97–1.06 | 0.55 | – | – | – |
| Amygdala Lewy bodies | 1.02 | 0.96–1.07 | 0.60 | – | – | – |
Ordinal logistic regression analysis with co-pathology measures represented as 0, 1, or 2/3.
Odds ratios (OR) represent 5-year increments.
Amygdala Lewy bodies were not analysed because of the limited number of cases with amygdala predominant Lewy bodies.
One or more ε4 alelle.
One or more ε2 alelle.
Odds ratios represent 1-year increments.
Discussion
We attempted to deduce when in both the clinical and pathological course of Alzheimer’s disease do CAA, LATE-NC and Lewy pathologies appear. We found that only rarely can clinical dementia of the Alzheimer’s type be assumed to be plaques and tangles alone. In our two cohorts, dementia associated with more widespread distributions of each co-pathology. CAA correlated with both plaques and tangles and associated with the APOE ε4 allele independent of dementia status. LATE-NC correlated with tangles and longer disease durations. LATE-NC was also more common at older ages independent of dementia status. Without dementia, Lewy bodies were age-associated, but with dementia, Lewy bodies were more prevalent at younger ages. Additionally, both the neocortical and amygdala predominant Lewy body patterns appeared to depend on a moderate accumulation of both plaques and tangles. Clearly, CAA, LATE-NC and Lewy bodies exhibit a complex relationship with plaques and tangles, with each individually influenced by age and genetics. Nonetheless, we review the evidence for each as incidental, primary or interacting Alzheimer’s disease co-pathologies below and note when we are unable to cleanly separate the multiple hypotheses.
We report three main findings regarding the development of CAA. First, CAA can be present in individuals without any plaques (Fig. 2), indicating that CAA may be an incidental pathology. Second, consistent with other studies (Love et al., 2014), CAA is more likely in APOE ε4 individuals in cases with or without dementia (Table 4). Without dementia, CAA prevalence in the CNDR cohort is 35.9% with an APOE ε3/ε3 genotype and 76.0% with an APOE ε4/ε4 genotype (Supplementary Fig. 1). With dementia, CAA is reported in 75.9% of APOE ε3/ε3 and 100.0% of APOE ε4/ε4 participants (Supplementary Fig. 2). As such, CAA may be a primary pathology that leads to a later accumulation of plaques and tangles, especially in APOE ε4 allele carriers, as a recent imaging study concluded (Montagne et al., 2020). The presence of CAA has also been linked to the APOE ε2 allele (Love et al., 2014), but we did not observe a correlation in either individuals with or without dementia (Table 4). Third, the presence of amyloid-β-positive CAA associated with tangles just as well, if not better, than amyloid-β-positive plaques (Table 3). Thus, CAA may develop as an interacting pathology with both plaques and tangles. Alzheimer’s disease is already characterized by an interaction between plaques and tangles, and CAA may be a prevalent, third part of the disease process (Charidimou et al., 2017).
Our data reveal four findings regarding the development of LATE-NC. First, LATE-NC in clinical Alzheimer’s disease associates with longer dementia durations (Table 4) implying that LATE-NC is an interacting pathology that may occur after both plaques and tangles have already accumulated (Jack et al., 2013). Second, since LATE-NC in neuropathological Alzheimer’s disease is the only co-pathology that associates with Braak stage and not Amyloid phase in the CNDR cohort (Table 3) and both tangles and LATE-NC start in the medial temporal lobe, LATE-NC may be a neurofibrillary tangle interacting proteinopathy. Third, LATE-NC may also be described as an incidental pathology independent of plaques and tangles (Fig. 2). In the CNDR cohort, 9.5% of the PART cases and 9.9% of the low ADNC cases had LATE-NC. This prevalence of LATE-NC was even higher in the NACC cohort where 14.7% of the PART cases and 25.8% of the low ADNC cases were positive and is consistent with previous studies (James et al., 2016; Wennberg et al., 2019). Fourth, LATE-NC may also be a primary pathology that correlates with early cognitive decline, especially at older ages (Fig. 1). In the CNDR cohort, 23.8% of the cognitively impaired CDR 0.5 cases were ≥ LATE-NC stage 2 at a mean (SD) age of 87.9 (8.8) years. Similarly, 13.7% of the NACC cognitively impaired cases were also ≥ LATE-NC stage 2 at a mean (SD) age of 84.5 (9.3) years.
We highlight two conclusions regarding the development of Lewy bodies in Alzheimer’s disease. First, Lewy body pathology had significantly different distributions when Alzheimer’s disease pathology was present (Fig. 2). In the CNDR cohort, the neocortical and amygdala predominant patterns were not observed in the cases with PART or low ADNC, but accounted for approximately half of the Lewy body pathology seen in intermediate (40.0%) and high ADNC (61.5%). Second, age as a risk factor for Lewy bodies had opposite effects when dementia was or was not present (Table 4). Without dementia, Lewy bodies occurred more frequently with increasing age, but, with dementia, Lewy bodies were more prevalent at younger ages, as others have reported (Chung et al., 2015). Together with the increased prevalence of Lewy body pathology in the NACC cohort in PART and low ADNC cases (Fig. 2), we interpret these findings to indicate that brainstem and limbic Lewy bodies are age-related incidental pathologies while neocortical and amygdala predominant Lewy bodies are Alzheimer’s disease interacting pathologies. Finally, we note that the neocortical pattern of Lewy bodies may arise from the brainstem to limbic to cortical pattern described for Parkinson’s disease (Braak et al., 2003). but this distribution may also occur as an outward progression from the amygdala to cortical and brainstem regions (Toledo et al., 2016).
The strengths of our study included the use of a large, well-characterized cohort with additional data analyses from a cohort of participants collected from Alzheimer’s Disease Centers across the USA. Nevertheless, our study had some limitations. We attempted to dissect out the progression from low to high ADNC, but most of our cases were high ADNC. As a retrospective, cross-sectional study, the actual plaque, tangle and co-pathology burden during each of the mean (SD) 9.5 (4.3) years of clinical dementia is unknown. Alzheimer’s Disease Center referrals are not representative of the overall US population thereby limiting the generalizability of our conclusions. We attempted to characterize the most prevalent Alzheimer’s disease related co-pathologies, but other common pathologies including ageing-related tau astrogliopathy, ischaemic injury and non-CAA vascular pathologies also occur in the ageing brain. Finally, while we staged CAA and LATE-NC according to published criteria, important morphological characteristics of each co-pathology were not analysed, including distinguishing parenchymal, meningeal and capillary CAA (Love et al., 2014) and distinguishing the histopathologic types of TDP-43 deposits (Lee et al., 2017; Josephs et al., 2019). Future studies should consider these distinctions in a more fine-grained analysis.
Given the link of each co-pathology with the severity of Alzheimer’s disease pathology, our data argues for the value of earlier disease interventions, potentially before more widespread concomitant pathology develops. Thus, recognizing and predicting the amount of each co-pathology has implications for therapeutics. For CAA, the contribution of additional amyloid-β in blood vessels may already have been a factor in the failure of anti-amyloid-β clinical trials, but since CAA is less prevalent in APOE ε3 individuals, future drug studies that account for an individual’s APOE genotype may reduce the risk of failure. For LATE-NC, increasing age is a big risk factor. Alzheimer’s disease trials may miss their clinical targets in older age groups, especially since LATE-NC prevalence may influence early cognitive decline. Nonetheless, if a significant burden of LATE-NC is downstream of tangle accumulation as our data suggest, then anti-tau compounds may be effective treatments for both Alzheimer’s disease and Alzheimer’s disease with LATE-NC. For Lewy body pathology, the majority of Lewy body pathology appears to be driven by plaques and tangles in earlier onset Alzheimer’s disease. Compounds targeting both amyloid-β and tau in younger age groups may be effective at treating both Alzheimer’s disease and Alzheimer’s disease with Lewy bodies. In summary, our findings reflect abundant and significant interactions between pathologies which associate with clinical Alzheimer’s disease progression. Future research should continue to focus on disentangling the multiple pathologies that occur in Alzheimer’s disease, while future clinical trials should consider combination therapies that include compounds targeting TDP-43 and α-synuclein in addition to amyloid-β and tau.
Supplementary Material
Acknowledgements
We wish to thank Terry Shuck for her assistance, the Human Studies Group for their feedback, and, most importantly, the individuals, families, and doctors who made this research possible. The NACC database is funded by NIA/NIH Grant U01 AG016976. NACC data are contributed by the NIA-funded ADCs: P30 AG019610 (PI Eric Reiman, MD), P30 AG013846 (PI Neil Kowall, MD), P30 AG062428-01 (PI James Leverenz, MD) P50 AG008702 (PI Scott Small, MD), P50 AG025688 (PI Allan Levey, MD, PhD), P50 AG047266 (PI Todd Golde, MD, PhD), P30 AG010133 (PI Andrew Saykin, PsyD), P50 AG005146 (PI Marilyn Albert, PhD), P30 AG062421-01 (PI Bradley Hyman, MD, PhD), P30 AG062422-01 (PI Ronald Petersen, MD, PhD), P50 AG005138 (PI Mary Sano, PhD), P30 AG008051 (PI Thomas Wisniewski, MD), P30 AG013854 (PI Robert Vassar, PhD), P30 AG008017 (PI Jeffrey Kaye, MD), P30 AG010161 (PI David Bennett, MD), P50 AG047366 (PI Victor Henderson, MD, MS), P30 AG010129 (PI Charles DeCarli, MD), P50 AG016573 (PI Frank LaFerla, PhD), P30 AG062429-01(PI James Brewer, MD, PhD), P50 AG023501 (PI Bruce Miller, MD), P30 AG035982 (PI Russell Swerdlow, MD), P30 AG028383 (PI Linda Van Eldik, PhD), P30 AG053760 (PI Henry Paulson, MD, PhD), P30 AG010124 (PI John Trojanowski, MD, PhD), P50 AG005133 (PI Oscar Lopez, MD), P50 AG005142 (PI Helena Chui, MD), P30 AG012300 (PI Roger Rosenberg, MD), P30 AG049638 (PI Suzanne Craft, PhD), P50 AG005136 (PI Thomas Grabowski, MD), P30 AG062715-01 (PI Sanjay Asthana, MD, FRCP), P50 AG005681 (PI John Morris, MD), P50 AG047270 (PI Stephen Strittmatter, MD, PhD).
Funding
S.X.X. (AG062418), V.M.Y.L. (AG17586) and J.Q.T. (AG010124, AG062418) report grants from the US National Institutes of Health during this study.
Competing interests
The authors report no competing interests.
Supplementary material
Supplementary material is available at Brain online.
Glossary
- ADNC
Alzheimer’s disease neuropathological change
- CAA
cerebral amyloid angiopathy
- CDR
Clinical Dementia Rating
- CNDR
Center for Neurodegenerative Disease Research
- LATE-NC
limbic-predominant age-related TDP-43 encephalopathy neuropathological change
- MMSE
Mini-Mental State Examination
- NACC
National Alzheimer's Coordinating Center
- PART
primary age-related tauopathy
References
- Braak H, Del Tredici K, Rüb U, de Vos RAI, Jansen Steur ENH, Braak E.. Staging of brain pathology related to sporadic Parkinson’s disease. Neurobiol Aging 2003; 24: 197–211. [DOI] [PubMed] [Google Scholar]
- Charidimou A, Boulouis G, Gurol ME, Ayata C, Bacskai BJ, Frosch MP, et al. Emerging concepts in sporadic cerebral amyloid angiopathy. Brain J Neurol 2017; 140: 1829–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chung EJ, Babulal GM, Monsell SE, Cairns NJ, Roe CM, Morris JC.. Clinical features of alzheimer disease with and without lewy bodies. JAMA Neurol 2015; 72: 789–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Crary JF, Trojanowski JQ, Schneider JA, Abisambra JF, Abner EL, Alafuzoff I, et al. Primary age-related tauopathy (PART): a common pathology associated with human aging. Acta Neuropathol 2014; 128: 755–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- de Flores R, Wisse LEM, Das SR, Xie L, McMillan CT, Trojanowski JQ, et al. Contribution of mixed pathology to medial temporal lobe atrophy in Alzheimer’s disease. Alzheimers Dement 2020; 16: 843–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jack CR Jr, Knopman DS, Jagust WJ, Petersen RC, Weiner MW, Aisen PS, et al. Tracking pathophysiological processes in Alzheimer’s disease: an updated hypothetical model of dynamic biomarkers. Lancet Neurol 2013; 12: 207–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- James BD, Wilson RS, Boyle PA, Trojanowski JQ, Bennett DA, Schneider JA.. TDP-43 stage, mixed pathologies, and clinical Alzheimer’s-type dementia. Brain J Neurol 2016; 139: 2983–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jellinger KA.Vascular-ischemic dementia: an update. J Neural Transm Suppl 2002; 62: 1–23. [DOI] [PubMed] [Google Scholar]
- Josephs KA, Murray ME, Tosakulwong N, Weigand SD, Serie AM, Perkerson RB, et al. Pathological, imaging and genetic characteristics support the existence of distinct TDP-43 types in non-FTLD brains. Acta Neuropathol 2019; 137: 227–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Josephs KA, Murray ME, Whitwell JL, Tosakulwong N, Weigand SD, Petrucelli L, et al. Updated TDP-43 in Alzheimer’s disease staging scheme. Acta Neuropathol 2016; 131: 571–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kapasi A, DeCarli C, Schneider JA.. Impact of multiple pathologies on the threshold for clinically overt dementia. Acta Neuropathol 2017; 134: 171–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Keage HAD, Carare RO, Friedland RP, Ince PG, Love S, Nicoll JA, et al. Population studies of sporadic cerebral amyloid angiopathy and dementia: a systematic review. BMC Neurol 2009; 9: 3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lee EB, Porta S, Michael Baer G, Xu Y, Suh E, Kwong LK, et al. Expansion of the classification of FTLD-TDP: distinct pathology associated with rapidly progressive frontotemporal degeneration. Acta Neuropathol 2017; 134: 65–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Leverenz JB, Hamilton R, Tsuang DW, Schantz A, Vavrek D, Larson EB, et al. Empiric refinement of the pathologic assessment of Lewy-related pathology in the dementia patient. Brain Pathol Zurich Switz 2008; 18: 220–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Love S, Chalmers K, Ince P, Esiri M, Attems J, Jellinger K, et al. Development, appraisal, validation and implementation of a consensus protocol for the assessment of cerebral amyloid angiopathy in post-mortem brain tissue. Am J Neurodegener Dis 2014; 3: 19–32. [PMC free article] [PubMed] [Google Scholar]
- McKeith IG, Boeve BF, Dickson DW, Halliday G, Taylor J-P, Weintraub D, et al. Diagnosis and management of dementia with Lewy bodies: fourth consensus report of the DLB Consortium. Neurology 2017; 89: 88–100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Montagne A, Nation DA, Sagare AP, Barisano G, Sweeney MD, Chakhoyan A, et al. APOE4 leads to blood-brain barrier dysfunction predicting cognitive decline. Nature 2020; 581: 71–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Montine TJ, Phelps CH, Beach TG, Bigio EH, Cairns NJ, Dickson DW, et al. National Institute on Aging-Alzheimer’s Association guidelines for the neuropathologic assessment of Alzheimer’s disease: a practical approach. Acta Neuropathol 2012; 123: 1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Nelson PT, Dickson DW, Trojanowski JQ, Jack CR, Boyle PA, Arfanakis K, et al. Limbic-predominant age-related TDP-43 encephalopathy (LATE): consensus working group report. Brain J Neurol 2019; 142: 1503–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Perneczky R, Wagenpfeil S, Komossa K, Grimmer T, Diehl J, Kurz A.. Mapping scores onto stages: mini-mental state examination and clinical dementia rating. Am J Geriatr Psychiatry 2006; 14: 139–44. [DOI] [PubMed] [Google Scholar]
- Saito Y, Ruberu NN, Sawabe M, Arai T, Kazama H, Hosoi T, et al. Lewy body-related α-synucleinopathy in aging. J Neuropathol Exp Neurol 2004; 63: 742–9. [DOI] [PubMed] [Google Scholar]
- Shinohara M, Murray ME, Frank RD, Shinohara M, DeTure M, Yamazaki Y, et al. Impact of sex and APOE4 on cerebral amyloid angiopathy in Alzheimer’s disease. Acta Neuropathol 2016; 132: 225–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Spires-Jones TL, Attems J, Thal DR.. Interactions of pathological proteins in neurodegenerative diseases. Acta Neuropathol 2017; 134: 187–205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Thal DR, Ghebremedhin E, Orantes M, Wiestler OD.. Vascular pathology in Alzheimer disease: correlation of cerebral amyloid angiopathy and arteriosclerosis/lipohyalinosis with cognitive decline. J Neuropathol Exp Neurol 2003; 62: 1287–301. [DOI] [PubMed] [Google Scholar]
- Thal DR, Ghebremedhin E, Rüb U, Yamaguchi H, Del Tredici K, Braak H.. Two types of sporadic cerebral amyloid angiopathy. J Neuropathol Exp Neurol 2002; 61: 282–93. [DOI] [PubMed] [Google Scholar]
- Toledo JB, Gopal P, Raible K, Irwin DJ, Brettschneider J, Sedor S, et al. Pathological α-synuclein distribution in subjects with coincident Alzheimer’s and Lewy body pathology. Acta Neuropathol 2016; 131: 393–409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Toledo JB, Van Deerlin VM, Lee EB, Suh E, Baek Y, Robinson JL, et al. A platform for discovery: the University of Pennsylvania Integrated Neurodegenerative Disease Biobank. Alzheimers Dement 2014; 10: 477–84.e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wennberg AM, Tosakulwong N, Lesnick TG, Murray ME, Whitwell JL, Liesinger AM, et al. Association of Apolipoprotein E ε4 With Transactive Response DNA-Binding Protein 43. JAMA Neurol 2018; 75: 1347–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wennberg AM, Whitwell JL, Tosakulwong N, Weigand SD, Murray ME, Machulda MM, et al. The influence of tau, amyloid, alpha-synuclein, TDP-43, and vascular pathology in clinically normal elderly individuals. Neurobiol Aging 2019; 77: 26–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zaccai J, Brayne C, Matthews FE, Ince PG; on behalf of the MRC Cognitive Function and Ageing Neuropathology Study. Alpha-synucleinopathy and neuropsychological symptoms in a population-based cohort of the elderly. Alzheimers Res Ther 2015; 7: 19. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.