Cerebrospinal uid levels of amyloid-beta oligomers in patients with idiopathic normal pressure hydrocephalus are elevated before shunt surgery but decrease afterward

Background The amyloid-beta (Aβ) oligomer has strong neurotoxicity and is associated with cognitive impairment in Alzheimer’s disease (AD). However, its role in patients with idiopathic normal pressure hydrocephalus (iNPH) is poorly understood. We hypothesised that cerebrospinal uid (CSF) stagnation leads to Aβ oligomer accumulation in patients with iNPH. We measured CSF Aβ oligomer levels before and after CSF shunting in patients with iNPH. Methods We evaluated two iNPH cohorts: an analysis cohort (cohort-1) with 52 patients and a validation cohort (cohort-2) with 13 patients. For comparison cohorts, we recruited 27 neurologically normal controls (NCs), 16 patients with AD, 15 patients with Parkinson’s disease (PD), and 14 patients with progressive supranuclear palsy (PSP). We measured CSF Aβ oligomer levels and assessed participants’ neurological statuses. We then compared the iNPH cohorts’ pre-shunting measurements with the comparison groups’ measurements and compared cohort-1’s measurements recorded before and after CSF shunting. to evaluate differences between biomarkers measured before and after CSF shunting. Spearman’s rank correlation coecient was used to evaluate correlations. All statistical analyses were performed with statistical software EZR version 1.41 based on R and R commander.(17) Results are shown as means ± standard deviations for ages and CSF biomarker values and as medians and interquartile ranges for clinical evaluation battery scores. A p-value < 0.05 was considered statistically signicant. KK, MM, MN, CA, SY and HA are contributed to the study concept, design and collection of CSF samples in iNPH cohort-1 and NCs.. IO and HX carried out CSF assay. CH carried out neuropsychologic assessment in iNPH cohort-1. SN contributed to the statistical analysis and interpretation. MK contributed to the collection of CSF samples and geographic information of individuals in iNPH cohort-2. YM contributed to the collection of CSF samples and geographic information of individuals in AD, PD and PSP. KK contributed to the correction and ensuring of English in the manuscript. KK carried out all analysis for the study and manuscript was written by all authors. All authors read and approved the nal manuscript.


Introduction
Idiopathic normal pressure hydrocephalus (iNPH) is a neurodegenerative disease involving gait disturbances, cognitive impairment, and urinary incontinence that was rst described by Hakim et al. in 1965.(1) iNPH is a treatable neurodegenerative condition, and 80% of patients normally improve with cerebrospinal uid (CSF) shunt surgery. (2) However, its pathophysiology remains unclear.
According to the recently articulated "oligomer hypothesis", the amyloid-beta (Aβ) oligomer contributes to cognitive impairments in patients with Alzheimer's disease (AD). (3) The Aβ peptide has a strong tendency to aggregate and the oligomer is considered more toxic than the monomer.(4) It has been reported that a nonamer (called Aβ*52) has strong neurotoxicity, (5) and CSF concentrations of the Aβ oligomer, which usually has 10-20 monomers, are higher in patients with AD than in healthy controls and are correlated with cognitive impairment severities. (6) We have previously published reports regarding the CSF concentrations of the Aβ 42 toxic conformer, which usually contains 2-3 monomers, in patients with iNPH. In patients with iNPH the toxic conformer ratio (i.e. the ratio of toxic conformer levels to total Aβ 42 levels) is higher than in cognitively normal individuals and lower than in patients with AD. Patients with iNPH who have decreased toxic conformer ratios following CSF shunting display better outcomes than patients without such decreases.(7) Impairment of CSF absorption followed by CSF accumulation, impaired CSF secretion, and a brief period of CSF stagnation may play a key role in iNPH pathophysiology, (8) as it may disrupt clearance of waste molecules, such as Aβ peptides, and lead to cognitive impairment.
On the basis of the aforementioned ndings, we hypothesised that CSF stagnation in patients with iNPH may result in the accumulation of the Aβ oligomer and that CSF shunting can restore normal clearance of CSF waste products and thus avert cognitive disturbances. To the best of our knowledge, the CSF levels of Aβ oligomers in patients with iNPH have not been systematically investigated. We therefore aimed to measure the CSF Aβ oligomer concentrations of patients with iNPH and compare them with those of neurologically normal controls and those of patients with other neurodegenerative diseases. We also aimed to determine whether CSF Aβ oligomer concentrations change following CSF shunting and to test for correlations between these biomarkers and cognitive functions.

Study populations
Our evaluations included two iNPH cohorts: an analysis cohort and a validation cohort. We diagnosed patients with iNPH according to the Japanese iNPH guidelines.(9) We enrolled 173 consecutive patients with iNPH who had undergone lumbo-peritoneal shunting at the Department of Neurosurgery of Juntendo University (Tokyo, Japan) between April 2011 and December 2017. The following patients were excluded from the study: 32 patients aged < 70 years or > 85 years, 9 patients without at least 2 years of follow-up after surgery, 9 patients for whom no CSF had been obtained preoperatively or who had a previous tap test history at another institute, and 71 patients who had not been studied with [ 123 I]-io upane singlephoton emission computed tomography (i.e. dopamine transporter scintigraphy [DaTSCAN]). Ultimately, we included 52 patients in the analysis cohort (iNPH cohort-1). For the validation cohort (iNPH cohort-2), we used data from 13 patients who were diagnosed through the neurology department of the National Hospital Organization Takasaki General Medical Centre (Takasaki, Japan) according to the same criteria used for iNPH cohort-1.
Because patients with iNPH may have additional comorbid neurodegenerative conditions, we subclassi ed the iNPH cohort-1 patients according to CSF concentrations of phosphorylated tau (pTau) and DaTSCAN ndings. On the basis of our previous ndings that CSF pTau levels ≥ 30 pg/mL are associated with worsened cognitive function outcomes 3 years after shunting in patients with iNPH, (10) we assumed patients with pTau levels > 30 pg/mL had AD comorbidity.
We used the speci c binding ratios (SBRs) for [ 123 I]-io upane determined with DaTSCAN to detect Parkinson's spectrum (PS) comorbidities. Given that SBRs vary depending on clinical settings, it is di cult to directly compare SBRs measured at different institutions. However, a previous report from our institution indicated that a cut-off SBR value of 2.83 is a useful diagnostic threshold for diffuse Lewy body disease. (11) We therefore, assumed that patients with SBR values < 3 had PS comorbidity. The patients in iNPH cohort-1 were thus subclassi ed into four subgroups: 15 patients in an "iNPH without comorbidities" subgroup, 18 patients in an "iNPH with AD comorbidity" subgroup, 14 patients in an "iNPH with PS comorbidity" subgroup, and 5 patients in an "iNPH with AD/PS comorbidity" subgroup ( Fig. 1).
To evaluate the Aβ oligomer levels in patients with iNPH, we also enrolled neurologically normal controls Fourteen patients with PSP were clinically diagnosed by a neurologist according to the Movement Disorder Society criteria. (14) Twenty-seven NCs who were > 60 years of age and had Mini-Mental State Examination (MMSE) scores of 26 points or higher were enrolled. All NCs had previously undergone lumbar puncture procedures for various reasons, such as investigations into headache aetiologies, preoperative insertion of a lumbar drainage catheter for aortic surgery, undergoing planned myelography for degenerative lumbar spine disease, or undergoing planned surgery with spinal anaesthesia.

Study design
The iNPH cohort-1 patients were followed for at least 2 years after CSF shunting. CSF samples were obtained preoperatively and 1 year after shunt surgery, as described below. Patients were assessed with the MMSE, the Frontal Assessment Battery (FAB), and the modi ed Rankin Scale (mRS) by a neuropsychologist and with the Japanese iNPH Grading Scale (iNPHGS) by a neurosurgeon before surgery and 1 and 2 years after surgery. The patients with iNPH were assessed for radiographical features such as the presence or absence of disproportionately enlarged subarachnoid space hydrocephalus (DESH) according to clinical management guidelines for iNPH.(9) We also assessed vascular burdens according to a classi cation system for deep subcortical white matter hyperintensities (DSWMH) (15) and calculated Evans Index values. CSF biomarkers (i.e. pTau, Aβ 38 , Aβ 42 , the toxic Aβ 42 conformer, and Aβ oligomers) were quanti ed with an enzyme-linked immunosorbent assay (ELISA) kit as described below.
First, we compared iNPH cohort-1 with the comparison cohorts in terms of baseline characteristics and CSF biomarker levels (Study 1). Second, we validated the results of those analyses with data from iNPH cohort-2 (Study 2). Third, we performed comparisons of the following iNPH cohort-1 subgroups: those without comorbidities, with AD comorbidity, and with PS comorbidity. We compared them in terms of baseline characteristics and CSF biomarker levels (Study 3 − 1), and before and after CSF shunting in each subgroup (Study 3 − 2). We compared CSF biomarker levels measured before and 1 year after shunt surgery as well as neurological statuses recorded before and at 1-and 2-year timepoints after shunting.
Using data from iNPH without comorbidity subgroup, we analysed the correlations between Aβ oligomer levels and MMSE scores before surgery, and between Aβ oligomer levels 1 year after surgery and MMSE scores 2 years after surgery (Fig. 2). CSF analysis CSF samples were obtained via lumbar puncture performed in the L3-L4 or L4-L5 interspace with an 18gauge spinal needle. Lumbo-peritoneal shunting was performed with our previously reported method. (16) We used an adjustable valve in all patients (Medtronic Neurosurgery, Goleta, CA, or Integra Codman, Raynham, MA). For postoperative follow-up, we adjusted the valve pressure and monitored proper shunt function with CSF withdrawal via valve reservoir puncture as previously reported. (7,16) If shunt malfunction was detected, the patient underwent shunt revision immediately.

Statistics
Data distribution was evaluated graphically using histograms and Q-Q plots. In study 1, 2 and 3 − 1, the Kruskal-Wallis H test was performed for factorial analysis of variance and the Steel-Dwass test for multiple comparisons. Fisher's exact test was conducted to compare proportions. Areas under receiver operating characteristic curves (AUROCs) were calculated as goodness-of-t descriptors for predictors of iNPH status versus AD or PD and PSP status. In study 3 − 2, the Friedman's test followed by Wilcoxon signed-rank test with Holm's correction was conducted to evaluate differences between neurological statuses before and 1 and 2 years after CSF shunting, and the Wilcoxon signed-rank test was conducted to evaluate differences between biomarkers measured before and after CSF shunting. Spearman's rank correlation coe cient was used to evaluate correlations. All statistical analyses were performed with statistical software EZR version 1.41 based on R and R commander. (17) Results are shown as means ± standard deviations for ages and CSF biomarker values and as medians and interquartile ranges for clinical evaluation battery scores. A p-value < 0.05 was considered statistically signi cant.

Results
Comparisons of the iNPH cohorts with the NC, AD, PD, and PSP groups (Studies 1 and 2) Comparisons of the iNPH cohorts and the NC, AD, PD, and PSP groups revealed significant differences in the sex ratios and the proportions of patients with hypertension and dyslipidaemia. The Aβ oligomer levels in iNPH cohort-1 and the AD group were significantly higher than in the NC, PD, and PSP groups ( Figure 3). The Aβ 42 levels in the AD group were significantly lower than those in iNPH cohort-1 and the NCs (p = 0.014, < 0.001 significantly). The pTau levels in the AD group were significantly higher than those in any other group (Table 1). An Aβ oligomer cut-off level of 5.65 pM distinguished iNPH cohort-1 members from AD group members with 50.0% sensitivity, 76.9% specificity, and an AUROC curve of 0.61. An Aβ oligomer cut-off level of 5.48 pM distinguished iNPH cohort-1 members from PD and PSP group members with 93.1% sensitivity, 80.8% specificity, and an AUROC curve of 0.90 ( Figure 4).
In the validation analysis, the iNPH cohort-2 displayed significantly higher levels of Aβ oligomer than the NC, PD, and PSP subgroups did. This result is consistent with that observed for the iNPH cohort-1.   Comparisons of iNPH subgroups (Study 3-1) In comparisons of the iNPH cohort-1 without comorbidities, iNPH with AD comorbidity, and iNPH with PS comorbidity subgroups, there were no significant differences in terms of Page 9/25 sex ratios or the proportions of patients with DESH, DSWMH scores of 0-1, or comorbidities. The iNPH with PS comorbidity subgroup members were significantly younger than iNPH with AD comorbidity subgroup members. The iNPH without comorbidities subgroup members had significantly higher Evans Index values than the iNPH with PS comorbidity subgroup members did. The iNPH without comorbidity subgroup had significantly higher Aβ oligomer levels than the iNPH with PS comorbidity subgroup and had significantly lower Aβ 38 levels than the iNPH with AD comorbidity subgroup. The iNPH with AD comorbidity subgroup had significantly higher levels of the toxic Aβ 42 conformer than the iNPH with PS comorbidity subgroup. There were no significant differences between the subgroups in terms of Aβ 42 levels ( Table 2).  In the iNPH without comorbidities and iNPH with PS comorbidity subgroups, Aβ 38 levels were significantly increased and Aβ oligomer levels were significantly decreased relative to their pre-CSF shunting levels. Conversely, in the iNPH with AD comorbidity subgroup, the Aβ 38 levels were significantly increased, but Aβ oligomer levels were unchanged (Table 3).
In the iNPH without comorbidities subgroup, two patients showed increases in Aβ oligomer levels ( Figure 5). One of those two patients experienced a shunt malfunction during followup and underwent shunt revision. In the iNPH without comorbidity subgroup, there were no significant correlations between Aβ oligomer levels before surgery and MMSE scores before surgery (p = 0.949) or between Aβ oligomer levels after surgery and MMSE scores 2 years after surgery (p = 0.469).  Before versus 2 years significantly). Differences between biomarkers recorded before and after shunt surgery were evaluated with the Wilcoxon signed-rank test, and significance is shown in the table. Significance is shown as *p < 0.05, **p < 0.01 or *** p < 0.001.

Discussion
Aβ oligomer levels are elevated in patients with iNPH In iNPH, CSF stagnation is one of the causes of Aβ deposition followed by cognitive impairment. The Aβ oligomer levels in the iNPH cohort-1 were signi cantly higher than those in the NCs and the patients with PD or PSP. This result was validated with data from iNPH cohort-2. Consistent with the ndings of previous research, the AD group had higher Aβ oligomer levels than NCs. (6) This result may imply that Aβ aggregation is a disease-speci c phenomenon in iNPH pathophysiology as in AD, and that Aβ aggregation may be as strong in the iNPH group as in the AD group. However, we do not think that these results merely re ect the severity in Aβ aggregation. Although the antibody we used in the present study detects mainly 10-20-monomer oligomers,(6) we were unable to evaluate their molecular sizes.
Therefore, we cannot exclude the possibility that stronger Aβ aggregation produces larger oligomers and apparently decreases the total oligomer numbers in patients.
Patients with iNPH sometimes show clinical symptoms that are similar to those of other neurodegenerative diseases, which complicates differential diagnosis. According to past reports, ADrelated pathology can be observed in 18-75% of patients with iNPH,(23, 24) and 11-86% of patients with iNPH exhibit parkinsonism. (25)(26)(27)(28) Our results indicate that Aβ oligomers are not good diagnostic biomarkers for differentiating iNPH from AD but are useful for differentiating iNPH from PD and PSP.
However, as has been previously reported, AD can be differentiated from iNPH using CSF levels of pTau.
Additionally, Aβ oligomer levels were signi cantly higher in the iNPH without comorbidities subgroup than in the iNPH with PS comorbidity subgroup. This implies that Aβ oligomer levels are useful for distinguishing PS from other comorbidities in patients with iNPH. Although there were no signi cant differences in baseline clinical severities between iNPH subgroups, patients in the iNPH with PS comorbidity subgroup might have had PS-related pathophysiological features at baseline. Therefore, the effects of iNPH pathophysiology might have been less severe in the iNPH with PS comorbidity subgroup than in the iNPH without comorbidities subgroup. The fact that the Evans Index values in the iNPH without comorbidity subgroup was signi cantly higher than iNPH with PS comorbidity subgroup and that the patients in iNPH with PS comorbidity subgroup displayed worse outcomes in gait disturbance after CSF shunting are consistent with this speculation. These results also imply that Aβ oligomer levels in patients with iNPH may become elevated as the disease progresses.

Aβ oligomer decreases after CSF shunt placement
In the iNPH without comorbidities and iNPH with PS comorbidity subgroups, Aβ 38 levels increased and Aβ oligomer levels decreased after CSF shunting. In the iNPH without comorbidities subgroup, only two patients displayed elevated Aβ oligomer levels, and one of those patients experienced a shunt malfunction and received shunt revision. CSF production is disturbed in iNPH, and CSF concentrations of APP-derived peptides increase after shunt surgery. (29,30) It is possible that CSF elimination following shunt surgery improves CSF production and APP metabolism in the periventricular space. (31)(32)(33) In the present study, elevated CSF Aβ 38 levels may re ect an improvement in CSF turnover that resulted in decreased Aβ oligomer levels. On the basis of these results, we can speculate that Aβ oligomer accumulation may be caused by CSF stagnation and may be alleviated by CSF shunting.
Alternatively, in the iNPH with AD comorbidity subgroup, Aβ oligomer levels did not decrease despite the elevated Aβ 38 levels. This result may imply the difference in Aβ metabolism between iNPH and AD. Aβ is derived from the proteolytic cleavage of APP by β-and γ-secretases. It has been reported that γ-secretase is more strongly activated than β-secretase in iNPH, although β-secretase is more strongly activated than γ-secretase is in AD. (34) From these facts, we hypothesise that amyloid metabolism in AD is different from that in iNPH, which may be poorly affected by CSF stagnation.
As in AD, cognitive impairment in iNPH may also be affected by Aβ oligomer levels. However, in the present study, we did not observe any correlation between Aβ oligomer levels and MMSE scores. The lack of such correlation might be due to the fact, that iNPH without comorbidity subgroup had better cognitive function at the baseline. Although there was no signi cant difference in baseline characteristics, the iNPH without comorbidity subgroup manifested higher scores in MMSE at the baseline. These higher scores might affect the correlation between Aβ oligomer levels and MMSE scores at the baseline. Actually, in the present study, iNPH without comorbidity subgroup did not show any signi cant improvement in cognitive function. We considered it to be a result of a "ceiling" effect.

Study limitations
The main limitations of the current study were that we had only two cohorts of patients with iNPH and that the number of recruited participants was low. AD comorbidity in patients with iNPH was de ned only by pTau levels in CSF, and no pathological features were evaluated. PS comorbidity in patients with iNPH was de ned exclusively by DaTSCAN results, which might not have re ected true PS pathology. (33) We excluded patients with iNPH who had past ischemic events, and there were no signi cant differences between the iNPH cohort-1 subgroups in the proportions of patients with DSWMH scores of 0-1.
However, we did not do further investigations for vascular pathologies, such as performing cerebral blood ow studies. Vascular lesions may affect biomarker measurements and DaTSCAN scan ndings.
Although we hypothesised that amelioration of CSF stagnation may improve Aβ oligomer accumulation, we could not quantify CSF ow or the volume of CSF eliminated through the shunt device. Future research should attempt to evaluate the amount of CSF produced and eliminated to test this hypothesis.

Conclusions
We have successfully quanti ed CSF Aβ oligomer levels in patients with iNPH for the rst time. Aβ oligomer levels were higher in the patients with iNPH who had CSF stagnation than in the comparison groups. Aβ oligomer levels in patients with iNPH decreased after CSF shunting.

Availability Of Data And Materials
The datasets used and/or analysed during the current study are available from the corresponding author on a reasonable request. Study designs We evaluated two iNPH cohorts: the analysis cohort (iNPH cohort-1) with 52 patients and the validation cohort (iNPH cohort-2) with 13 patients. For comparison groups, we also enrolled 27 NCs, 16 patients with AD, 15 patients with PD, and 14 patients with PSP. In Study 1, we compared iNPH cohort-1 with the comparison groups (i.e. the AD, PD, PSP, and NC groups). In Study 2, we compared iNPH cohort-2 with the comparison groups to validate the results of Study 1. In Study 3, we compared three iNPH cohort-1 subgroups (i.e. the iNPH without comorbidities, iNPH with AD comorbidity, and iNPH with PS comorbidity subgroups) in terms of baseline characteristics and CSF biomarker levels and in terms of measurements recorded before and after CSF shunting.

Figure 3
Aβ oligomer levels in each group Aβ oligomer levels for each individual group are represented in a scatter plot. The vertical axis shows the CSF concentrations of Aβ oligomers. Signi cance testing was done with the Steel-Dwass test and is indicated as **p < 0.01 or ***p < 0.001 (Versus iNPH cohort-1) and ##p < 0.01 or ###p < 0.001 (Versus iNPH cohort-2).  Signi cant decreases in Aβ oligomer levels 1 year after CSF shunting were observed in the iNPH without comorbidities subgroup (p = 0.002) and the iNPH with PS comorbidity subgroup (p = 0.036, respectively).

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. AdditionalTableJARTFinal.docx