A streamlined, resource-efficient immunoprecipitation-mass spectrometry method for quantifying plasma amyloid-β biomarkers in Alzheimer’s disease

High-performance, resource-efficient methods for plasma amyloid-β (Aβ) quantification in Alzheimer’s disease are lacking; existing mass spectrometry-based assays are resource- and time-intensive. We developed a streamlined mass spectrometry method with a single immunoprecipitation step, an optimized buffer system, and ≤75% less antibody requirement. Analytical and clinical performances were compared with an in-house reproduced version of a well-known two-step assay. The streamlined assay showed high dilution linearity (r2>0.99) and precision (< 10% coefficient of variation), low quantification limits (Aβ1–40: 12.5 pg/ml; Aβ1–42: 3.125 pg/ml), and high signal correlation (r2~0.7) with the two-step immunoprecipitation assay. The novel single-step assay showed more efficient recovery of Aβ peptides via fewer immunoprecipitation steps, with significantly higher signal-to-noise ratios, even at plasma sample volumes down to 50 pl. Both assays had equivalent performances in distinguishing non-elevated vs. elevated brain Aβ-PET individuals. The new method enables simplified yet robust evaluation of plasma Aβ biomarkers in Alzheimer’s disease.


Introduction
Brain amyloid β (Aβ) deposition is a pathological hallmark and diagnostic criterion of Alzheimer's disease (AD) [1,2].Following the recent approval of anti-Aβ monoclonal antibody therapies by the United States Food and Drugs Administration (FDA), the importance of reliable yet accessible biomarkers in clinical settings has become increasingly essential [3][4][5].Currently, the most widely used biomarkers for assessing Aβ deposition are positron emission tomography (PET) imaging of Aβ plaques, and cerebrospinal uid (CSF) measurements of Aβ42/40 peptide ratio via immunoassays [6,7].However, these methods are limited by their high costs, invasiveness, and lack of widespread availability, which restrict their use in routine clinical assessments [8].
Blood-based biomarkers, such as the plasma Aβ1-42/1-40 ratio, offers a minimally invasive and affordable option that may guide, complement, or serve as a substitute for PET and CSF testing [9][10][11].
In 2018, Nakamura et al. presented a matrix-assisted laser desorption/ionization (MALDI) time-of-ight (TOF) MS method to measure Aβ peptides from 250 µl of plasma.This assay demonstrated robust predictive capabilities of both the Aβ1-40/Aβ1-42 and APP669-711/Aβ1-42 ratios for assessing brain Aβ burden [22].Subsequent studies revealed potential clinical utility of the assay in predicting individual brain Aβ burden and discriminating AD from non-AD neurodegenerative conditions [31].
Recognizing the importance of plasma Aβ measurement for AD detection, we previously adapted the IP-MS assay originally developed by Nakamura et al., referred to here as the Pittsburgh plasma Aβ assay version 1.0 (PAβ V1.0) [32].The PAβ V1.0 assay was implemented on a cost-effective benchtop MALDI-TOF instrument, the Bruker Micro ex LT.However, considering the limited dynamic range of MS, the presence of strong background interference including albumin and immunoglobulins could impede the accurate detection of plasma Aβ peptides [14,33].To address this limitation, the V1.0 assay uses two rounds of immunoprecipitation (IP).Regrettably, this approach increases per-sample costs, reagent usage, and sample preparation time.
Here, we describe a novel IP-MS assay, referred to as the Pittsburgh Plasma Aβ assay version 2.0 (PAβ V2.0).This new version streamlines the two steps of IP into one, using a commercially available supplement buffer, which was identi ed in a screen of buffers and detergents that can effectively improve the signal-to-noise ratio (S/N).We evaluate the analytical and clinical performance compared with the PAβ V1.0 assay, focusing on the biomarkers Aβ1-42, Aβ1-40, and the Aβ1-42/Aβ1-40 ratio.

Participants recruitment
This study included plasma samples from two cohorts.For the rst cohort, we enrolled participants from the University of Pittsburgh Alzheimer's Disease Research Center (ADRC) in Pittsburgh, Pennsylvania, USA.The participants in this ongoing study undergo annual clinical evaluation to assess their longitudinal brain health and potential development of cognitive impairment and dementia.Annual evaluations include neuroimaging, cognitive testing, and blood collection for use in plasma biomarker analysis outside of the clinical assessment.Neuropsychological evaluation and diagnoses were established through clinical assessments [28,29].The battery of cognitive tests included the Montreal Cognitive Assessment (MoCA), Mini-Mental State Examination (MMSE), and the Clinical Dementia Rating (CDR) scale.The current investigation was a prospective, blinded substudy where participants were enrolled based on their order of clinical attendance and their informed consent to participate.This involved agreeing to provide an additional tube of blood for the project.The ADRC study was approved by the University of Pittsburgh Institutional Review Board (MOD19110245-023).
The second cohort was sourced from the Active Gains in Brain Using Exercise During Aging (AGUEDA) project (NCT05186090).
Participants were recruited from Granada, Spain, based on their classi cation as physically inactive and cognitively normal, assessed by the Spanish Telephone Interview for Cognitive Status modi ed (STICS-M), MMSE, and MoCA.As an outcome, Aβ PET was performed using the [ 18 F] Florbetaben tracer, quanti ed using standardized uptake value ratio (SUVR) values and the Centiloid (CL) scale.Detailed information on eligibility criteria, participant selection methods, and recruitment procedures, as well as details about the study setting, locations, and data collection, can be found in a comprehensive description provided in the AGUEDA protocol [34].Prior to enrollment in the AGUEDA trial, participants provided informed consent, and the trial was conducted in accordance with the approval of the Research Ethics Board of the Andalusian Health Service (CEIM/CEI Provincial de Granada; #2317-N-19).In this cross-sectional analysis, we focused on the baseline data.
Researchers were blinded to all participant information until the completion of data acquisition.

Blood collection and processing procedures
At the University of Pittsburgh ADRC, blood samples were collected via venipuncture by nurses with extensive clinical experience and trained in ADRC procedures [35].Blood collection was performed between 9:00 am and 2:00 pm, with the time of last meal recorded.For the AGUEDA cohort, blood samples were collected at 08:00-10:00 am following longer than 8 hours of fasting, at the Virgen de las Nieves University Hospital, Spain.Brie y, a 10 and 4 ml Lavender top ethylenediaminetetraacetic acid (EDTA) tube was used to collect whole blood from each participant in the ADRC and AGUEDA cohort, respectively.Following each blood draw, the tubes were promptly inverted 8 to 10 times and subsequently centrifuged at 2000 xg for 10 minutes for the AGUEDA cohort and 15 minutes for the ADRC cohort at 4°C to effectively separate the plasma.The resulting plasma samples were aliquoted into cryovials and frozen at -80°C until use, following standard guidelines [35].
The samples were immunoprecipitated with 10 µl of 50 mg/ml Dynabeads (M-270 Epoxy; ThermoFisher #14301) coupled with 5 µg 6E10 anti-Aβ antibody (BioLegend #803003) for 1 hour at 4°C with rotation.The beads were coupled with the antibody following the protocol recommended by the manufacturer.After the IP, the supernatant was discarded, and the beads washed once with 0.5 ml of cold phosphate-buffered saline (PBS, Gibco #2537136).The washed beads were then transferred to a fresh Eppendorf tube using 0.5 ml of cold PBS and eluted with 25 µl of glycine elution buffer (50 mM glycine [pH 2.8, Sigma #G2879-100G], 0.1% DDM) after removing all liquid.The eluates were collected and transferred to fresh tubes containing 0.5 ml of the binding buffer (without any Aβ ISs) for a second round of IP.Following one hour of rotation at 4°C, the beads were washed twice with 0.5 ml of cold HPLC-grade H 2 O (Fisher #7732-18-5) and transferred to a fresh Eppendorf tube by resuspending in 0.2 ml H 2 O.After complete removal of all liquid through vacuum aspiration, the beads were eluted using 6 µl of 3 mg/ml α-cyano-4hydroxycinnamic acid matrix (Bruker #8201344) dissolved in TA50 (50% Acetonitrile [Fisher #75-05-8], 0.1% Tri uoroacetic acid [Alfa Aesar #UN2699], 1 mM ammonium dihydrogen phosphate [Sigma #204005]).The eluate was spotted four times with 1 µl each onto the MALDI target plate (Bruker #8280823) for MS analysis.A schematic illustration of the work ow for this assay is shown in Fig. 1A.
Single IP procedure for detergents and blocking buffer tests Similar to the rst IP step of the PAβ V1.0 assay, we prepared 250 µl of the same assay binding buffer, either used as is or supplemented with one of the following detergents or blocking buffers: 10% v/v SuperBlock (Thermo #37535), 10 µg/ml TruBlock (Meridian #A66803H), 0.5% v/v Triton100 (Millipore #648462), 0.5% v/v Tween20 (BioRad #1610781), or 10% Quanterix This mixture was transferred to a 1.5 ml Eppendorf Protein LoBind tube with 62.4 pg/ml of Aβ1-38 IS, 100 pg/ml of Aβ1-40 IS, and 30 pg/ml of Aβ1-42.Subsequently, 250 µl of human plasma sample was added to the mixture.The sample was immunoprecipitated with 5 µl of 50 mg/ml Dynabeads coupled with 1.25 µg 6E10 Aβ antibody (BioLegend #803003) for 1 hour at 4°C with rotation.After IP, the supernatant was discarded, and the beads resuspended in 0.5 ml of the assay binding buffer with the corresponding supplement added as appropriate and transferred to a new tube.The beads underwent an additional wash with 0.5 ml of the binding buffer with corresponding supplement, two washes with 0.5 ml of PBS and one wash with 0.5 ml of HPLC-grade H 2 O. Finally, the beads were transferred to a fresh Eppendorf tube using 0.2 ml of H 2 O.After removal of all liquid through vacuum aspiration, the beads were eluted using 6 µl of 3 mg/ml α-cyano-4-hydroxycinnamic acid matrix dissolved in TA50.The eluate was spotted four times with 1 µl each onto the MALDI target plate for analysis.

Screening of buffers and blockers for the PAβ V2.0 assay
We evaluated the effects of several buffer systems and heterophilic blocking agents for the PAβ V2.0.These included the 10% N4PE CSF diluent from Quanterix, the 10% v/v SuperBlock, 10 µg/ml TruBlock, 0.5% v/v Triton100 and 0.5% v/v Tween20.The results from the PAβ V2.0 assay were compared to those obtained using the PAβ V1.0 assay.

MALDI-TOF MS
After sample spotting, the MALDI target plate was air dried and then loaded into a benchtop MALDI-TOF mass spectrometer, Micro ex LT (Bruker Daltonics), equipped with a 337 nm nitrogen laser to acquire mass spectra.The Micro ex LT operated in linear mode with a pulsed positive ion extraction setting, utilizing an attenuator offset of 12%, an attenuator range of 30%, and 63% laser power.An external mass calibration was performed using a peptide calibration mixture consisting of two calibration standards (Bruker #8222570, #8206355).The auto scan function was utilized, acquiring one spectrum for each spot through the combination of ion signals from 2,500 laser shots, resulting in four spectra per sample.Aβ1-38 IS was employed to ensure spectrum quality in the auto scan function.Only spectrum, generated from every 50 shots, with Aβ1-38 IS S/N ratios greater than three were collected.After acquisition, the spectra underwent smoothing using the SavitzkyGolay algorithm with a width of 0.1 mass-to-charge (m/z) and baseline subtraction using the TopHat algorithm.The peak intensity and S/N ratios were measured using FlexControl (v3.4,Bruker Daltonics).Subsequently, ClinPro Tools Software (v2.1, Bruker Daltonics) was employed for m/z alignment, peak detection, and peak area calculation.

Analytical assessment
Linearity analysis was conducted using a two-fold serial dilution of an Aβ peptide mixture, starting with concentrations of 400 pg/ml for Aβ1-40 (Anaspec, #AS-24235) and 10 pg/ml for Aβ1-42 (Anaspec, #AS-20276), in 6% bovine serum albumin (BSA)/PBS, diluting up to 64x.The analysis involved six replicates for each dilution, totaling 36 samples, which were evenly processed across two batches.The lower limit of quanti cation (LLOQ) was established as the lowest concentration measurable with a coe cient of variation (CV) under 20% [25].The working range was de ned as the range from the LLOQ to the highest concentration tested.To evaluate the plasma matrix effect, we assessed the recovery by comparing the results in plasma to those in 6% BSA/PBS at three different concentration levels.Both media were spiked with equal amounts of Aβ1-40 and Aβ1-42 prior to the IP procedures.Recovery was calculated using the formula: where P represents the normalized peak area.Intra-and inter-assay variability were determined by analyzing samples at three Aβ concentrations levels across ve batches, each containing six replicates per concentration.

Plasma dilution linearity
The effect of plasma dilution on normalized intensity for both the PAβ V1.0 and PAβ V2.0 assay formats were investigated by testing ve separate amounts of a pooled plasma sample (50 µl to 250 µl), with three replicates each.All samples in this test were diluted to 250 µl prior to processing, and Aβ1-40 and Aβ1-42 levels were normalized using the Aβ1-38 IS only.

Simoa assay for IP recovery assessment
To quantify the proportion of Aβ peptides retained after the IP procedures, Single Molecule Array (Simoa) assays were utilized.These assays were performed using the Simoa Human Neurology 4-Plex E assay (N4PE) kit from Quanterix (#103670) on an HD-X analyzer (Quanterix, Billerica, MA, USA).To monitor assay performance, quality control samples at three different concentrations were analyzed at the beginning and end of each assay run.The average %CV for the quality controls was below 5%.
Mass spectrometric and immunoassay experiments were performed in the Mass Spectrometry facility at the Bio uid Biomarker Laboratory, Department of Psychiatry, School of Medicine, University of Pittsburgh, Pittsburgh, PA, USA.
The assay performance over multiple batches was evaluated using pooled quality control plasma samples at two concentration levels by measuring the Aβ1-40 and Aβ1-42.In both assays, normalization of Aβ1-40 and Aβ1-42 was conducted using the Aβ1-38 IS.The intra-and inter-assay %CV were determined to be less than 15% for both cohorts.

Statistical Analysis
For participant demographic characteristics, continuous variables were summarized using means and standard deviations, while categorical variables were reported as numbers and percentages.Differences across cohorts for continuous variables were examined using the Wilcoxon Rank Sum test or Kruskal-Wallis test, depending on the number of groups involved.Categorical variables were analyzed using Fisher's exact tests.For S/N ratio comparison between different assays, Wilcoxon Rank Sum test was used.For clinical assessments, box and whisker plots were generated using clinical assessments, Aβ PET imaging results, and CL scales over the cohorts.Wilcoxon Rank Sum test was used to assess the disease discriminating performance of biomarkers across cohorts based on the clinical assessments or the Aβ PET imaging results.The Kruskal-Wallis test was used to evaluate the difference among the CL scale groups.The Cohen's d was calculated for multiple assay biomarkers to evaluate the standardized difference between different diagnostic groups.For correlation study, Spearman correlation analysis was conducted to evaluate the strength of the association between Aβ peptide measurements from the two different assays.For all the tests, a p-value less than 0.05 was considered statistically signi cant.All analyses were performed using R statistical software (version 4.2.1, R Foundation for Statistical Computing, Vienna, Austria), available at [http://www.r-project.org/].

Effectiveness of detergents and blocking buffers in reducing IP-MS background
To streamline the PAβ V1.0 assay into a single IP step, we experimented with various supplements in the IP binding buffer to reduce background interference.These included 10% N4PE CSF diluent, 10% SuperBlock, 10 µg/ml TruBlock, 0.5% Triton100, and 0.5% Tween20, were all tested following the Single IP protocol (see Methods section).
Among the supplements tested, the N4PE CSF diluent demonstrated the best performance, effectively eliminating interference peaks while maintaining the highest S/N ratio.Consistent with the PAβ V1.0 assay, the interference peak at 4450 m/z, which often obscures the Aβ1-38 and Aβ3-40 signals, was signi cantly eliminated with the use of the N4PE CSF diluent but not with the other supplements.Notably, SuperBlock and TruBlock resulted in signi cantly lower S/N ratios when compared with the PAβ V1.0 assay.The detergents, on the other hand, showed the lowest S/N ratios for all Aβ peptides.(Fig. 2 and Supplementary Fig. 1).
As a con rmation comparison, the original PAβ V1.0 assay and the single IP assay that used the N4PE CSF diluent were compared with the original PAβ V1.0 assay con guration with one IP step and no binding buffer supplementation.As shown in the representative spectra (Fig. 3A), supplementing the IP binding buffer with the N4PE CSF diluent resulted in the cleanest spectra.Similar to the PAβ V1.0 assay, the interference peaks observed in the PAβ V1.0 assay with a single round of IP at 3200 m/z to 3500 m/z, and at 6400 m/z and 6600 m/z, were reduced by using the N4PE CSF diluent.Furthermore, the single round of IP procedure using N4PE CSF diluent achieved a signi cantly higher S/N ratio, with means of 143.9 for Aβ1-40 and 9.5 for Aβ1-42, compared with 72.4 and 5.5 respectively in the PAβ V1.0 assay, and 23.9 and 1.6 in the PAβ V1.0 assay with one IP (Figs. 3B,  3C).Similar improvements were also observed for other Aβ peptides.In the PAΒ V1.0 assay, the S/N ratios were 16.0, 7.8, 4.9, and 3.6 in PAβ V1.0 assay and 5.5, 0, 1.8, and 2.4 in the PAβ V1.0 assay with a single IP for Aβ1-38, Aβ3-40, Aβ1-39, and APP669-711, respectively.Conversely, in the single IP with N4PE CSF diluent, these ratios improved to 29.8, 13.7, 9.1, and 7.1 for the same peptides (Figs.3B, 3C).
Due to the optimal performance, we selected the single IP with N4PE CSF diluent-supplemented binding buffer as the Pittsburgh assay PAβ V2.0.

Analytical assessment
We proceeded to compare the analytical performance of the PAβ V1.0 assay with the PAβ V2.0 assay.
The inter-assay CV for both Aβ1-40 and Aβ1-42 were below 20% in the sample with the lowest non-zero concentrations.Thus, we set the LLOQs for both assays at 12.5 pg/ml for Aβ1-40 and at 3.125 pg/ml for Aβ1-42.Additionally, since the linearity extended to the sample with the highest concentrations, we set the upper limits of quanti cation (ULOQs) for the assays at 400 pg/ml for Aβ1-40 and 100 pg/ml for Aβ1-42.

Matrix effect assessment
To assess plasma matrix effect, we compared signals of Aβ peptides in plasma samples relative to BSA/PBS at three concentration levels (118.2pg/ml,53.6pg/ml, and 21.4pg/ml for Aβ1-40, 47.2pg/ml, 23.0pg/ml and 10.8pg/ml for Aβ1-42) and calculated the matrix effect recovery following the formula outlined in the "Materials and Methods" section.Both the PAβ V1.0 and PAβ V2.0 assay formats demonstrated similar matrix effects (Table .2, Fig. 4B).The detailed results are listed in Table 2. Interestingly, we observed overall better recovery when using analyte speci c IS to normalize peak area (Table .2, Fig. 5B).These results suggest that different Aβ peptides might exhibit varying matrix effects, and analyte speci c IS might be more robust in normalizing the matrix effect of corresponding analytes.

Assay precision
Assay precision was evaluated at three concentration levels (37.5 pg/ml, 146.4 pg/ml and 382.5 pg/ml for Aβ1-40, 82.8 pg/ml, 49.5 pg/ml and 13.9 pg/ml for Aβ1-42) using normalized peak areas for both intra-and inter-assay assessments.The detailed results are listed in Table .3.

Relationship between plasma dilution and normalized intensity
The relationship between plasma dilution and normalized intensity for both the PAβ V1.0 and PAβ V2.0 assays was linear (r 2 > 0.99), except for the Aβ1-42 of the PAβ V1.0 assay for which r 2 was 0.758.This deviation can be attributed to the inaccuracy introduced by low S/N ratio at low concentration level (Fig. 4D).

Plasma volume requirement
The PAβ V1.0 assay was designed around the use of 250 µl plasma sample for each measurement.To test whether the PAβ V2.0 assay could enable measurement of Aβ peptides at lower plasma volumes, we examined both assays using varying amounts of plasma, ranging from 50 µl to 250 µl in increments of 50 µl, with three replicates for each sample volume.The results showed the PAβ V2.0 assays provided 178.0-22.7%higher S/N of Aβ1-40 and 87.6-26.1% higher S/N of Aβ1-42 from 50 µl to 250 µl (Fig. 4E).

IP recovery
To evaluate the proportion of Aβ peptides that were retained after the IP procedures, we utilized Simoa immunoassay to provide absolute quanti cation of Aβ peptides before and after IP.IP recovery was evaluated at three concentration levels of low, medium, and high (27.4pg/ml, 51.4 pg/ml, and 99.2 pg/ml for Aβ1-40; 7.0 pg/ml, 13.2 pg/ml, and 27.4 pg/ml for Aβ1-42) in triplicates.The result demonstrated that the PAβ V2.0 assay retained a higher proportion of Aβ peptides after IP (Fig. 4C).
Implementing analyte speci c IS in the PAβ V2.0 assay revealed a similar performance, yielding a p-value of 0.0013.The effect sizes for the Aβ-PET-positive versus negative groups were 0.53, 0.56, and 0.73 for the PAβ V1.0, PAβ V2.0, and PAβ V2.0 with analyte speci c IS assays, respectively.
These results indicate that the PAβ V2.0 assay with common IS normalization and analyte speci c IS normalization performed comparably to the PAβ V1.0 assay in the clinical predictive performance.In the ADRC cohort, we observed strong correlations for Aβ1-38, Aβ1-39 and Aβ1-40 when normalized using Aβ1-38 IS (r > 0.8).The remaining Aβ peptides also exhibited strong correlations of 0.8 > r > 0.6 (Fig. 7A).
Similar results were obtained in the AGUEDA cohort, where there were strong correlations (r > 0.8) between Aβ1-40 measures when normalized using Aβ1-38 IS or Aβ1-40 IS.The correlation of Aβ1-42 (normalized using Aβ1-38 IS or Aβ1-42 IS) and Aβ1-38 was strong (0.8 > r > 0.6).Additionally, the correlation of the other Aβ peptides was moderate (0.6 > r > 0.5) (Fig. 7B), all indicating good agreement between the peptide levels measured in the different assay formats

Discussion
Mounting evidence indicate that plasma Aβ ratio has utility to measure brain Aβ pathology [39], and target engagement in therapeutic programs targeting brain Aβ aggregates.It is critical for healthcare systems to utilize cost-effective and minimally invasive methods for clinical diagnosis and patient selection and monitoring for the recently approved immunotherapies.
Additionally, Aβ is an early biomarker showing changes in patients with incipient disease including in cognitively normal older adults compared to non-diseased individuals, highlighting its critical role in the pre-clinical diagnosis of AD [40,41].Early detection provides an opportunity for intervention and potentially altering the disease course.Furthermore, implementing a blood-based biomarker test for patient triaging could potentially reduce the current 50-month wait for treatment access to just 12 months, as projected by specialist referral models for cognitive impairment and dementia patients [42].These factors highlight the necessity of plasma-based IP-MS Aβ assay as a tool for early diagnosis.
Among the various blood-based Aβ assays, IP-MS methods such as the assay from Nakamura et al. stands out for its performance but has limitations needing improvement [22].We adopted and enhanced this assay, resulting in the PAβ V2.0 assay with several enhancements.Firstly, our assay streamlined sample preparation time and preanalytical processing.
Secondly, our new assay demonstrated a substantially stronger signal to noise ratio.Thirdly, the PAβ V2.0 and PAβ V1.0 assays exhibited similar clinical performance and analytical performance across multiple cohorts.To our knowledge, this is the rst time that such signi cant enhancements have been achieved in re ning the landmark Nakamura et al. plasma Aβ method.
The PAβ V2.0 assay successfully streamlined the IP steps using a commercially available buffer -the N4PE CSF diluent.The high detergent, high salt content and the interference blocker mixture in the buffer helped reduce the background noise.This buffer was selected after comparing its performance against several detergents and blocking buffers.While all other tested reagents exhibited lower S/N ratios compared with PAβ V1.0 assay, the N4PE diluent demonstrated higher S/N ratio, supporting its selection for further use as PAβ V2.0 assay.
The PAβ V2.0 assay maintained comparable analytical performance with a higher recovery rate compared to the PAβ V1.0 assay.
This result was veri ed by SIMOA, an immunoassay with a different measurement mechanism than MS.We also tested the S/N of Aβ1-40 and Aβ1-42 in the PAβ V2.0 assay, utilizing a diluted sample volume of pooled plasma.The results demonstrated a higher S/N ratio and suggested the potential feasibility of decreasing the sample volume to 50-100 µl for the PAβ V2.0 assay.
However, further investigation is warranted, including comparisons to Aβ PET imaging and/or CSF analysis, to assess the clinical utility and determine the feasibility of utilizing reduced sample volumes.Additionally, the PAβ V2.0 assay preserved similar clinical performance, with peptide concentrations showing strong correlation with those in the PAβ V1.0 assay.
We further tested the performance of using analyte speci c IS (Aβ1-40 IS and Aβ1-42 IS) comparing to the common IS (Aβ1-38 IS) for normalization.Our results indicated that using analyte speci c IS for normalization can slightly improve the matrix effect recovery of plasma Aβ peptides.However, the use of the analyte speci c IS did not signi cantly change the analytical performance of the Aβ biomarkers.In the clinical performance analysis, regardless of whether the analyte speci c IS was used in the PAβ V2.0 assay or not, the results did not show a signi cant improvement compared to the PAβ V1.0 assay.Our ndings supported Nakamura et al.'s approach, con rming that using a common IS in the MALDI-TOF based IP-MS Aβ assay did not signi cantly alter clinical performance.
The MS instrument we utilized for our assays was a Bruker Micro ex LT MALDI-TOF, widely adopted across numerous clinical facilities.Notably, it has received FDA approval for clinical microbiology diagnosis in humans [43,44], and is widely available in many laboratories.In comparison to other MS instruments utilized in alternative IP-MS plasma Aβ assays, the Micro ex is distinguished by its affordability and simplicity.Furthermore, it offers practical advantages, such as direct compatibility with a standard 110V outlet, without necessitating the use of any special electrical modi cations or voltage converter.Moreover, its user-friendly interface facilitates straightforward operation, enabling general laboratory technicians to operate the instrument pro ciently without requiring specialized training in mass spectrometry.
Our study has several notable strengths.Firstly, we describe in detail the technical development, analytical and clinical validation of an improved plasma Aβ assay by IP-MS.Secondly, our study included two different cohorts.This diverse representation enhances the generalizability and practical relevance of our ndings.Thirdly, the cohorts had been characterized for biological evidence of disease using brain Aβ PET and neuropsychologically using established evaluation instruments such as the MMSE, MoCA, and CDR.Limitation of our study include the fact that the sample size across cohorts was relatively small.Future evaluations utilizing larger-scale studies, ideally conducted in real-world settings, will be necessary to validate and con rm the robustness of our ndings.
In conclusion, we report successful development of a more resource-e cient and cost-effective IP-MS plasma Aβ assay.Supplementary Files

Figures
Figures

Figure 2 Mass
Figure 2

Figure 3 The
Figure 3

Table 2
Matrix effect recovery of the Aβ peptides using both normalization methods of Pittsburgh assays

Table 1
* P-values were calculated using the Wilcoxon Rank Sum or Kruskal-Wallis tests for continuous variables, and Fisher's exact test for a categorical variable.Two participants had missing data for APOE alleles.The percentage was calculated based on a sample size of 75.Abbreviations: APOE, apolipoprotein E; CDR, Clinical Dementia Rating; MMSE, Mini Mental State Examination; MoCA, Montreal Cognitive Assessment; CL, Centiloid.

Table 3
Inter-and Intra assay precision of the Aβ peptides using both normalization methods of Pittsburgh assays Compared with the in-house reproduced Nakamura et al. assay, the new assay demonstrated comparable clinical and analytical performance.The cost, time, and reagent savings, coupled with the utilization of a more affordable and widely available instrument, will enable research laboratories to conduct IP-MS analysis of Aβ in blood more effectively.