A practical guide to preclinical Alzheimer’s disease biomarkers panel selection when you’re considering outsourcing to a specialist neurology CRO

Biomarkers have transformed Alzheimer’s disease clinical trials, enabling diagnosis confirmation for trial eligibility, patient stratification based on likelihood of progression, longitudinal monitoring of therapeutic effects, and pharmacodynamic verification of candidate therapies as study outcomes.  

The 2026 AD clinical trial drug pipeline report published by Cummings et al., 2026 reflects this shift: 158 drugs are now being assessed across 192 active trials, with 83% incorporating at least one biomarker. Fifty percent of current trials use a biomarker for eligibility, and 27% have a biomarker as a primary outcome measure.  

Fluid biomarkers, in particular, are playing an increasingly central role: used as inclusion criteria in 60% of trials that incorporate an eligibility biomarker, with 44 trials measuring CSF analytes and 26 measuring blood-based readouts. Aβ and tau biomarkers feature as primary outcome measures in 16 and 14 trials, respectively. 

However, for any program entering the clinic today, biomarker-supported evidence of mechanism is no longer optional; it is an expected part of your clinical trial design. 

For preclinical sponsors, this means the biomarkers your preclinical efficacy study generates need to be interpretable in a clinical context, ideally using the same readouts your clinical team will be tracking in Phase 1 and 2.  In this regard, choosing a preclinical CRO that can run validated biomarker panels alongside your efficacy endpoints, rather than outsourcing biomarker analysis separately, keeps your data consistent and your timeline intact.  

Below, we outline the main fluid biomarker readouts available at InnoSer across our Alzheimer’s disease mouse model portfolio, along with representative datasets illustrating their utility as efficacy markers. 

Plasma NfL: does your compound reduce neurodegeneration markers, or just pathological burden?  

As the field moves toward multi biomarker evidence studies demonstrating efficacy beyond pathology clearance, neurofilament light chain (NfL) arises as a relevant biomarker to include in your preclinical efficacy studies.  

NfL is released from damaged axons into CSF and blood, making it a sensitive marker of ongoing neuroaxonal injury. While not specific to Alzheimer’s disease, NfL may be a particularly useful downstream confirmatory readout, indicating whether observed reductions in amyloid or tau pathology are accompanied by measurable neuroprotection.  

Homozygous Tau[P301S] mice show progressive age-related increases in plasma levels of NfL  

FIGURE 1. Elevated plasma Neurofilament Light (NF-L) levels in Tau[P301S] transgenic mice. Longitudinal analysis shows a progressive increase in plasma NF-L levels in homozygous Tau[P301S] transgenic mice over time compared to age-matched wild-type controls. NF-L concentrations were measured using ELISA; data represent mean ± SEM.

Increased plasma NfL levels correlated with plasma pTau181 levels in the APP[V717I]xTau[P301S] mouse model 

FIGURE 2. Increased plasma Neurofilament Light correlates with plasma pTau181, supporting the utility of plasma NF-L and pTau181 as complementary plasma biomarkers for disease progression. (A) Plasma concentrations of Neurofilament Light (NF-L) were measured using MSD technology (kit K1517XR). (B) A significant positive correlation was observed between plasma NF-L and pTau181 levels, suggesting a link between axonal damage and Tau-related neurodegeneration in APP[V717I] × Tau[P301S] mice.

Plasma and CSF NfL are elevated in APPV[717I] × PS1[a246E] mice at 9 months, indicating neuroaxonal damage 

FIGURE 3. Plasma and CSF neurofilament light chain (NfL) levels are significantly elevated in APP[V717I] × PS1[A246E] transgenic mice compared to wild-type controls at 9 months of age, indicating progressive neuroaxonal damage concurrent with established amyloid pathology. NfL was quantified by MSD technology (kit K1517XR).

Aβ42/40 ratio: tracking amyloid burden as a pharmacodynamic biomarker in your efficacy study 

The plasma Aβ42/40 ratio reflects the sequestration of Aβ42 into insoluble amyloid aggregates in the brain. As plaque burden increases, soluble Aβ42 concentrations in plasma decline relative to Aβ40, resulting in a falling ratio. Because it is non-terminal and repeatable across timepoints, the plasma Aβ42/40 ratio may reduce the number of terminal cohorts needed in a longitudinal efficacy study, and it mirrors the amyloid confirmation biomarkers increasingly used for clinical trial eligibility screening. 

At the tissue level, cortical soluble and insoluble Aβ40 and Aβ42 fractions can be quantified separately following sequential extraction, providing a more direct readout of plaque-associated versus soluble amyloid species and their independent responses to treatment. 

APP[V717I]xPS1[A246E] mice show progressive increase of insoluble Aβ 40/42 fibrils in cortex

The APP [V717I]xPS1[A246E] mouse model shows an age-dependent increase in both soluble and insoluble Aβ species, demonstrating progressive amyloid pathology, with the two fractions reflecting distinct temporal aspects of disease progression.

This ultimately supports the use of insoluble cortical Aβ 40/42 fibrils as a pharmacodynamic readout in preclinical Alzheimer’s disease mouse model studies.  

FIGURE 4. Progressive accumulation of soluble and insoluble Aβ40 and Aβ42 in the cortex of APP[V717I] × PS1[A246E] transgenic mice. Cortical levels of soluble (A) Aβ40 and (B) Aβ42, and insoluble (C) Aβ40 and (D) Aβ42, were quantified using ELISAs specific for human Aβ peptides (Thermo Fisher, KHB3481 & KHB3544).

Dose-dependent reduction in soluble Aβ42 correlates with cognitive improvement following donepezil treatment in APP[V717I]×PS1[A246E] mouse model 

In the APP[V717I × PS1[A246E] transgenic mouse model, sub-chronic donepezil treatment produced a dose-dependent reduction in cortical soluble Aβ42 levels that correlated with dose-dependent improvement in spatial reference memory in the Morris Water Maze probe test (Easton et al.,2013).

These results suggest donepezil may alleviate cognitive impairment in part by reducing brain Aβ, and further illustrate the sensitivity of soluble Aβ42 as a pharmacodynamic endpoint in this model. 

Plasma Aβ42/40 ratio correlates negatively with cortical Aβ plaque burden in the APP[V717I] × Tau[P301S] mouse model 

FIGURE 5. In APP[V717I] × Tau[P301S] transgenic mice, plasma concentrations of Aβ42 and Aβ40 were measured using MSD technology (kit K15200E). (A) The plasma Aβ42/40 ratio declines progressively with age in APP[V717I] × Tau[P301S] transgenic mice, mirroring the reduction in Aβ42/40 ratio observed in Alzheimer’s disease patients as amyloid burden increases, although absolute ratio values differ between species. (B) A significant negative correlation was observed between the plasma Aβ42/40 ratio and cortical Aβ plaque burden assessed by immunohistochemistry, supporting the potential use of this ratio as a peripheral pharmacodynamic marker of cerebral Aβ pathology progression in this model.

CSF Aβ1-42 and pan-Tau: biomarker monitoring in chronic Alzheimer’s disease preclinical efficacy studies 

CSF biomarkers provide a more proximal readout of Alzheimer’s disease brain pathology than plasma biomarkers and may be particularly informative in chronic dosing studies where the aim is to track compound effects in reducing pathology over time.  

Accordingly, simultaneous monitoring of CSF Aβ1-42 and pan-Tau levels in a transgenic combined amyloid and tau mouse model may support programs evaluating compounds targeting multiple pathological pathways in a single preclinical study. 

Chronic compound administration significantly reduces CSF Aβ1-42 and pan-Tau in APPV[717I] × TauP[301S] transgenic mice 

FIGURE 6. In APP[V717I] × Tau[P301S] transgenic mice, chronic daily oral administration of an experimental compound from 5 to 13 months of age significantly reduced both CSF Aβ1-42 and pan-Tau secretion compared to vehicle-treated animals. Aβ1-42 levels were quantified by ELISA (vehicle: N=22; compound: N=17) and pan-Tau by separate ELISA (vehicle: N=21; compound: N=15), with statistically significant reductions at p<0.05 to ***p<0.001. These findings demonstrate that simultaneous pharmacodynamic monitoring of amyloid and tau species in CSF is feasible in this model following chronic dosing.

A starting framework for fluid biomarker inclusion in your next preclinical Alzheimer’s disease efficacy study 

The appropriate panel will depend on your compound’s mechanism and downstream clinical endpoints.  

Whether you are designing your first outsourced preclinical Alzheimer’s disease study or building on an existing dataset, the appropriate biomarker panel will depend on your compound’s mechanism and downstream clinical endpoints.

The table below can be used as a general starting point:

Not sure which biomarkers fit your study design? Our neurology team can advise on biomarker panel selection and matrix choice at the protocol stage, before you commit to a design.

ELISA, MSD or Western Blot: how do we reliably quantify preclinical Alzheimer’s disease biomarker responses to your therapy? 

Reliable biomarker quantification in preclinical Alzheimer’s disease studies depends on matching the right analytical platform to the right analyte and matrix.  

When outsourcing your preclinical Alzheimer’s disease program, biomarker platform capability is one of the most practical differentiators between CROs, as not all preclinical research organizations offer validated MSD, ELISA, and Western blot panels in-house. 

Western blot is applied for detection and semi-quantitative assessment of total and phosphorylated protein species in brain tissue, including p-Tau, and APP. Importantly, at InnoSer, Western blot is used following sarkosyl fractionation to separately assess soluble and sarkosyl insoluble tau species (normalised to the total homogenate), distinguishing the aggregation-prone, insoluble tau fraction from soluble tau, which is relevant for programs targeting tau aggregation or clearance mechanisms specifically.  

ELISA and MSD immunoassay platforms are used for quantitative measurement of soluble biomarkers across brain lysates, plasma, and CSF. MSD’s electrochemiluminescence technology offers particular advantages for low-abundance plasma analytes such as Aβ40/42 and NfL, where sensitivity requirements exceed what conventional ELISA can reliably deliver. 

CSF, plasma or brain tissue fractions: which matrix should you collect for your AD biomarker endpoints? 

CSF provides a higher-concentration, lower-background readout of brain-derived biomarkers but requires terminal collection in mice, with limited volumes, meaning careful prioritisation of Alzheimer’s disease biomarkers is necessary at the study design stage. 

Plasma is accessible longitudinally and non-terminally, making it preferable for repeated-measure study designs, though plasma Aβ species require high-sensitivity MSD platforms due to the lower signal relative to CSF. 

Brain tissue lysates, whether TBS-soluble fractions or sarkosyl-insoluble pellets following sequential extraction, provide the most direct readout of pathological burden and protein aggregation state, but are necessarily terminal. Soluble Aβ42 and soluble tau species in brain lysate may be particularly sensitive pharmacodynamic endpoints in treatment studies, as they can change prior to detectable shifts in insoluble plaque or tangle burden. 

The appropriate biological matrix, or alternatively a combination of biological matrices, will depend on your study design, treatment window, and whether longitudinal or terminal endpoints are prioritized. Our scientific team can advise on collection protocols and platform selection at the study design stage. 

Reach out to our neurology team to discuss which Alzheimer’s disease biomarkers and analytical platforms are most appropriate for your study design and compound mechanism.

From preclinical biomarkers to clinical endpoints: partnering with InnoSer for your Alzheimer’s disease efficacy study 

Biomarker strategy is no longer something you figure out after the model is running, it is a study design decision that shapes what your data can say, and to whom.  

At InnoSer, biomarker panel design is part of the conversation from the first protocol discussion, not an afterthought.  

Whether you are building your first preclinical AD biomarker dataset or extending an existing program, our neurology team can help you generate data that speaks the same language as your clinical team.  

Designing your preclinical Alzheimer’s disease biomarker strategy? Contact our neurology team to discuss panel selection, matrix choice, and how InnoSer’s validated MSD, ELISA, and Western blot platforms can generate translationally relevant biomarker data alongside your efficacy endpoints.