Flocculent Ti3AlC2-Based Two-Dimensional Material Flexible Microneedle Patch for Screening Alzheimer's Disease Animal Biomarkers and In Situ Revealing Spatial Aggregation of Arachidonic Acid.

Flocculent Ti3AlC2-Based Two-Dimensional Material Flexible Microneedle Patch for Screening Alzheimer’s Disease Animal Biomarkers and In Situ Revealing Spatial Aggregation of Arachidonic Acid

Efficient diagnosis and pathological elucidation of Alzheimer’s disease (AD) are constrained by the lack of technologies for in situ capture of metabolites in living brain tissues. Metabolic disorders in the parietal lobe highly correlate with AD progression, yet existing methods cannot achieve in situ capture and spatial analysis. A petal-like flocculent Ti3AlC2-based flexible microneedle patch was developed for in situ extraction and mass spectrometry of parietal metabolites in AD rabbits. Sodium hydroxide etching and cetyltrimethylammonium bromide (CTAB) self-assembly constructed a flocculent micro-/nanostructure, enhancing enrichment ~4-fold (2.5-fold after etching + 1.6-fold after CTAB). The stretchable substrate enables conformal attachment to the curved parietal lobe. Application to the exposed cortex demonstrates mechanical flexibility, linear glucose response (R2 = 0.9958) and resistance to salt/protein interference. Serum profiling of 15 AD patients and 15 controls with OPLS-DA shows clear discrimination (top-5 markers combined area under the curve (AUC) = 0.84). Spatial analysis reveals arachidonic acid’s unique pattern of centralized aggregation with peripheral gradient diffusion. This novel in situ device enables brain metabolite detection and reveals abnormal arachidonic acid aggregation, providing evidence for lipid-pathway AD interventions.

1. Introduction

Alzheimer’s disease is a clinically prevalent neurodegenerative disorder characterized by progressive cognitive impairment and behavioral deficits (Scheme 1a). Its complex pathogenesis and the difficulty of efficient diagnosis impose substantial burdens on families and society [1,2,3]. The cerebral parietal lobe, as a high-level cognitive integration center, participates in spatial perception, attentional regulation, and language comprehension. Recent studies indicate that AD progression is closely associated with metabolic network dysregulation in the parietal lobe, which exhibits metabolic reprogramming—such as decreased glucose metabolism, abnormal phospholipid metabolism, and neurotransmitter imbalance—at a suitable stage [4,5,6,7]. However, existing methods (e.g., cerebrospinal fluid extraction, positron emission tomography) struggle to achieve in situ efficient capture and precise quantification of metabolites in living brain tissues. Consequently, the dynamic changes and spatial distribution characteristics of local parietal metabolites under pathological states remain unclarified, hindering the understanding of AD pathogenesis and the discovery of efficient biomarkers.

Surface-assisted laser desorption/ionization mass spectrometry (SALDI-MS) offers advantages including rapid detection, simple sample preparation, suitability for complex biofluid analysis, and efficient qualitative/quantitative determination of small-molecule metabolites, rendering it a core technology in metabolomics [8,9,10,11,12]. The analytical performance of SALDI-MS relies heavily on the matrix material. Recently, two-dimensional (2D) materials such as MXenes and graphene have shown great potential as SALDI matrices due to their large specific surface areas, excellent electron transport capacities, and laser desorption/ionization catalytic activities [13,14,15,16,17,18,19,20]. Nevertheless, when directly applied for in situ metabolite collection in living brain tissues, conventional 2D matrices face limitations such as low sampling efficiency, narrow metabolite coverage, and insufficient resistance to biological matrix interference. Moreover, traditional rigid substrates cannot conform to the curved contours of brain tissue, preventing tight attachment and stable sampling [21,22,23,24,25].

To overcome these limitations, we designed a flocculent Ti3AlC2-based 2D material flexible microneedle patch (MXene-MNs-SALDI-Biopatch) (Scheme 1b). The patch integrates three core functional modules: (1) a microneedle array structure that penetrates and adsorbs metabolites from brain interstitial fluid (Scheme 1c,d) [26]; (2) a stretchable flexible substrate that conformally fits the skull contour and matches the rabbit parietal curved surface for stable in vivo sampling; and (3) a flocculent Ti3AlC2-based 2D material with a high specific surface area constructed via NaOH etching and CTAB surfactant-driven self-assembly, which enhances metabolite enrichment and laser desorption/ionization efficiency. We hypothesized that this integrated design would enable in situ capture, highly sensitive detection, and spatial distribution analysis of metabolites in AD-affected brain regions.

Therefore, adopting a structured, hypothesis-driven approach as recommended in recent methodological paradigms [27], this study pursues three specific objectives: (i) to verify the material’s mechanical flexibility (Young’s modulus 0.92 MPa) and anti-interference SALDI-MS performance; (ii) to evaluate the diagnostic value of a multi-marker serum panel (e.g., arachidonic acid) via OPLS-DA; and (iii) most critically, to test the spatial hypothesis—whether arachidonic acid exhibits a unique ‘center-aggregated with peripheral gradient diffusion’ pattern in the AD parietal lobe, providing direct spatial evidence for lipid-targeted interventions.

2. Materials and Methods

2.1. Reagents and Instruments

Ti3AlC2 powder (purity ≥98%), sodium hydroxide (NaOH), cetyltrimethylammonium bromide (CTAB), glucose standards, and arachidonic acid standards were of analytical grade and purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai 200002, China). Key instruments included a field-emission scanning electron microscope (SEM, Hitachi SU8010, Tokyo, Japan), transmission electron microscope (TEM, JEOL JEM2100F), surface-assisted laser desorption/ionization time-of-flight mass spectrometer (SALDI-TOF MS, Bruker Autoflex Speed, Bruker, Billerica, MA, USA), high-speed centrifuge (Eppendorf 5424R), and constant temperature drying oven. Deionized water was used throughout.

2.2. Alkaline Etching of Ti3AlC2 Substrate

The Ti3AlC2 powder was dispersed in 1 mol/L NaOH solution, transferred to a high-pressure autoclave, and reacted at 180 °C for 6 h [28,29,30,31]. The product was collected by centrifugation at 8000 rpm for 10 min, washed repeatedly with deionized water until neutral, and vacuum-dried at 60 °C for 12 h to obtain alkaline-etched layered Ti3AlC2.

2.3. CTAB Surfactant Self-Assembly Modification

The alkaline-etched Ti3AlC2 was dispersed in an aqueous solution containing 1.0 mmol/L CTAB. After ultrasonication for 30 min, the mixture was transferred to a high-pressure autoclave and reacted at 120 °C for 4 h. The product was centrifuged, washed alternately with deionized water and absolute ethanol three times, and vacuum-dried at 60 °C for 8 h to obtain the flocculent Ti3AlC2-based 2D material [32,33,34].

2.4. Preparation of Flexible Microneedle Patch

The synthesized flocculent Ti3AlC2-based 2D material was uniformly coated onto the surface of a pre-fabricated flexible polydimethylsiloxane (PDMS) substrate (thickness of approximately 200 μm) [35,36,37,38,39]. Petal-like microneedle arrays were constructed using a template-assisted method. Specifically, the microneedle height is 80 μm, the spacing is 95 μm, and the base diameter is 35 μm. After curing at 60 °C for 4 h, the flexible petal-like Ti3AlC2-based 2D material microneedle patch was obtained.

2.5. Structural Characterization, Mechanical and Biocompatibility Characterization

The microscopic morphologies of the alkaline-etched Ti3AlC2, the CTAB-modified flocculent material, and the microneedle arrays were characterized by SEM and TEM. Elemental composition and distribution were analyzed by energy-dispersive X-ray spectroscopy (EDS). Surface hydrophilicity was assessed using a contact angle goniometer.

The Young’s modulus of the MXene-MNs-SALDI patch was measured using a universal tensile testing machine. The patch was fixed between grips and subjected to a constant-rate tensile load; the stress–strain curve was recorded, and the modulus value was calculated. Meanwhile, the deformation behavior of the patch under uniaxial stretching was observed by optical microscopy, and the tensile strain was recorded.

To evaluate the tissue penetration capability and biocompatibility of the patch, the patch was pressed onto the surface of excised porcine skin and live rabbit skin for 5 min. After removal, the skin samples were processed for hematoxylin and eosin (H&E) staining to assess structural integrity and inflammatory response.

2.6. Animal Model, In Vivo Sampling, and Human AD Serum Collection

Male New Zealand rabbits (approximately 2.5 kg) were purchased from Shanghai Experimental Animal Center. The AD rabbit model was established by intraperitoneal co-injection of D-galactose (60 mg/kg) and AlCl3 (10 mg/kg) daily for 3 weeks [40,41]. All animal procedures were approved by the ethics committee. For in vivo sampling, rabbits were anesthetized with pentobarbital sodium (40 mg/kg) via the ear vein, and a craniotomy was performed to expose the parietal cortex. The flexible microneedle patch was applied to the parietal region for 5 min. After removal, the patch surface was gently rinsed with deionized water and analyzed directly via SALDI-MS.

For human serum collection, fasting blood samples (5 mL each) were drawn from 15 AD patients and 15 healthy controls via the median cubital vein. The blood was left to clot at room temperature for 30 min, and then centrifuged at 3000 rpm for 15 min at 4 °C. The supernatant serum was aliquoted (200 μL/tube) and stored at −80 °C under strictly controlled temperature conditions. For metabolomic profiling, the serum samples were mixed directly with the composite matrix. The human study protocol was approved under reference number NFEC-202503-JS-NT-3.

2.7. SALDI-MS Detection and Metabolomic Analysis

The microneedle patch containing captured metabolites was placed on the SALDI-MS target plate and analyzed using a nitrogen laser (337 nm). The parameters were: acceleration voltage of 20 kV, pulsed extraction delay of 100 ns, and 500 laser shots accumulated per spectrum. The mass spectra were acquired in the range of m/z 50–1000. Each sample was analyzed in triplicate.

After baseline correction, peak alignment, and normalization, the mass spectrometry data were analyzed using SIMCA software (v14.1) for PCA and OPLS-DA. Differential metabolites were screened using variable importance in projection (VIP > 1.0) and Student’s t-test (p < 0.05). Pathway enrichment was performed using MetaboAnalyst (Version 5.0, McGill University). Receiver operating characteristic (ROC) curves and the area under the curve (AUC) were generated via GraphPad Prism (Version 8.0.2, GraphPad Software, San Diego, CA, USA).

2.8. Spatial Distribution Analysis

To evaluate the spatial distribution of metabolites in the parietal region, the sampling area of the patch was divided into a center zone and a peripheral zone. SALDI-MS detection was performed on each zone, and the relative signal intensity ratio (center/periphery, C/P) was calculated for each metabolite.

3. Results

3.1. Material Preparation and Morphological Characterization

The overall structure of the Ti3AlC2-based petal-like microneedle array is characterized (Figure 1). Specifically, the three-dimensional schematic displays the regularly arranged petal-like microneedle array (Figure 1a). The front view shows the uniform distribution of the microneedles (Figure 1b). The side-view morphology demonstrates that the device can undergo free bending deformation based on its highly elastic flexible substrate to conformally fit the curved contour of the cerebral parietal lobe (Figure 1c). Furthermore, the magnified view of a single petal-like microneedle reveals the fine microstructure of the needle tip constructed from the flocculent 2D material (Figure 1d).

The synthesis process and electron microscopic morphology of the Ti3AlC2-based 2D functional material are evaluated (Figure 2). The complete preparation workflow begins with etching the Ti3AlC2 substrate with NaOH solution, followed by CTAB surfactant-driven self-assembly to yield the flocculent Ti3AlC2-based 2D material (Figure 2a). The baseline morphology of Ti3AlC2 after one-step NaOH etching displays a primary roughened layered structure (Figure 2b). After CTAB self-assembly, a fluffy, flocculent, porous microstructure is formed, significantly increasing the specific surface area and surface roughness to provide abundant active sites for metabolite adsorption (Figure 2c). Integrating this material into a microneedle patch, although the microneedles themselves are at the micron scale (80 μm in height, 95 μm in spacing, and 35 μm in base diameter), their internal architecture consists of petal-like MXene nanosheets (Figure 2), which enable efficient enrichment of negatively charged metabolites and sensitive metabolic detection.

3.2. Mechanical Properties and In Vivo Conformal Verification

The mechanical adaptability of the flexible microneedle patch and in vivo experimental characterization in AD rabbits are presented (Figure 3). The flat planar structure of the patch is schematically shown (Figure 3a). The arched, bent state of the patch reflects its outstanding bending performance (Figure 3b). The photograph captures large-angle bending deformation (Figure 3c). The bidirectional stretching state of the patch verifies its high tensile flexibility (Figure 3d and Figure S1). We performed Young’s modulus analysis on our MXene-MNs-SALDI Biopatch, and the measured modulus value was 0.92 MPa (S2), which is on the same order of magnitude as that of biological tissues, with a slight tensile strain of 15.2%.

This ensures that the patch can effectively extract metabolites while avoiding the significant trauma caused by rigid microneedles. Furthermore, we conducted H&E staining analysis after applying the MXene-MNs-SALDI Biopatch to the skin. The results demonstrated that, compared with the control group, the MXene-MNs-SALDI Biopatch did not induce obvious skin inflammation (Figures S3 and S4), and the use of CTAB as a surfactant did not compromise its biocompatibility, indicating that the patch possesses favorable biocompatibility.

These mechanical characterizations confirm that the patch possesses favorable flexibility to adapt to the irregular curved surface of brain tissue. The photograph of the craniotomy in an AD rabbit model shows the fully exposed parietal brain tissue (Figure 3e). Finally, the schematic of the flexible microneedle patch applied to the rabbit parietal lobe demonstrates conformal contact with the brain surface, establishing the physical foundation for in situ metabolite collection (Figure 3f).

3.3. Mass Spectrometry Detection Performance

The SALDI-MS quantitative performance of the flexible microneedle patch is analyzed (Figure 4). The mass spectrum of glucose obtained using the microneedle patch shows clearly identified characteristic peaks, alongside prominent sodium (m/z 203) and potassium (m/z 219) adduct peaks, confirming the favorable signal recognition capability of the device (Figure 4a). The series of mass spectra corresponding to gradient concentrations of glucose exhibit a clear incremental trend in signal intensity with increasing concentration (Figure 4b). The quantitative standard curve of glucose in the concentration range of 0.08–0.8 mmol/L yields a linear correlation coefficient of R2 = 0.9958, demonstrating a wide and excellent linear detection range (Figure 4c). Considering that CTAB, as a surfactant, can adequately modulate the performance of the MXene-MNs-SALDI Biopatch, it does not interfere with the SALDI-MS performance.

As shown in Figure 4d, the mass spectrometry signal of the precursor substance was only 1453 ± 112, while that of the layered MXene material was enhanced to 3644 ± 342, and that of the final flocculent MXene material was enhanced to 5763 ± 326. The signals increased by approximately 2.5-fold and 1.6-fold after the two modification steps, respectively, demonstrating that the stepwise modification protocol progressively enhances the material’s metabolite enrichment and recognition capabilities.

The salt tolerance and anti-protein interference performance of the device are tested (Figure 4e,f). In the presence of high salt (150 mM NaCl) and high protein (5% BSA), the intensity of the characteristic glucose mass spectrometry peaks does not exhibit significant attenuation, demonstrating that the Ti3AlC2-based material possesses outstanding resistance to salt and protein interference, making it highly suitable for the direct analysis of complex brain interstitial fluid samples (Figure 4e,f).

3.4. Serum Metabolomic Profiling

Metabolic profiling analysis of 15 human AD patient samples and 15 healthy controls is performed (Figure 5). Principal component analysis (PCA) shows that the normal and AD samples exhibit overlapping distributions in the 2D score plot, failing to achieve effective separation (Figure 5a). In contrast, the orthogonal partial least squares discriminant analysis (OPLS-DA) score plot reveals a clear boundary between the two groups without overlap, indicating that AD-associated metabolic reprogramming possesses multidimensional and subtle features that supervised learning methods can effectively extract (Figure 5b). A clustering heatmap based on the OPLS-DA results shows significant differences in the levels of various metabolites between the AD and control groups, displaying consistent intra-group and distinct inter-group clustering patterns (Figure 5c). Furthermore, a volcano plot (using VIP > 1.0 and p < 0.05 as thresholds) highlights a large number of statistically significant differential metabolic markers, with upregulated metabolites dominated by phospholipid intermediates and downregulated metabolites mainly consisting of amino acids (Figure 5d).

3.5. Diagnostic Efficacy and Pathway Enrichment

The diagnostic efficacy and pathway enrichment of key metabolic markers are analyzed (Figure 6). A combined receiver operating characteristic (ROC) curve analysis of the top 5 differential markers—including arachidonic acid, proline, glutamine, phosphocholine, and an unidentified lipid metabolite—yields an area under the curve (AUC) of 0.84, with a sensitivity of 0.89 and specificity of 0.70 (Figure 6a). Individually, the core biomarker arachidonic acid achieves a diagnostic AUC of 0.79 (Figure 6b), while proline yields an AUC of 0.73 (Figure S6), suggesting that single metabolic markers offer limited diagnostic efficacy, whereas multi-marker models significantly enhance diagnostic performance (Figure 6b and Figure S6).

Pathway linkage analysis of all differential metabolites revealed significant enrichment in multiple pathways, including arachidonic acid metabolism, glycerophospholipid metabolism, alanine/aspartate/glutamate metabolism, and tryptophan metabolism (Figure 6c). Among these, the significantly altered metabolites comprised arachidonic acid, glycine, phenylalanine, tryptophan, and methionine (VIP > 1.0 and p < 0.05, as the screening thresholds). The t-test results indicated that arachidonic acid levels were significantly different between the control group and the Alzheimer’s disease group (Figure S5), with p < 0.01. In addition, we performed a recovery test for arachidonic acid (n = 5), which yielded recovery rates ranging from 98.2% to 103.2% (Table S1), with a relative standard deviation (RSD) of 2.01%, indicating satisfactory precision. In addition, we have performed a comparative analysis of our MXene-MNs-SALDI strategy with current state-of-the-art techniques (Table S2) [42,43]. Our results demonstrate that the proposed MXene-MNs-SALDI approach exhibits overall superior performance in certain aspects.

The interaction network of these metabolic pathways demonstrates close regulatory connections among different pathways, suggesting that AD-associated parietal lesions involve the reprogramming of complex, multi-pathway metabolic networks, with lipid metabolism pathways situated at the core node of the network (Figure 6d).

3.6. Spatial Distribution of Parietal Metabolites

The spatial distribution of metabolites within the sampling area of the flexible microneedle patch is analyzed (Figure 7). In situ sampling is conducted on the parietal cortex of New Zealand rabbits (Figure 7a), and the collected metabolites are quantified for signal intensities in the center and peripheral regions. Most metabolic products exhibit a center-to-periphery (C/P) signal intensity ratio close to 1.0, showing no significant global aggregation or dispersion across the parietal region (Figure 7b,c). In contrast, arachidonic acid displays a unique spatial distribution characterized by a gradient decline in signal intensity from the center toward the periphery, representing a localized aggregation pattern (Figure 7d). The corresponding C/P ratio is significantly greater than 1.5 (p < 0.01) (Figure 7d). This specific spatial pattern indicates that arachidonic acid and its metabolic pathway may play a distinct regional regulatory role in AD parietal pathology, and its localized aggregation might be associated with focal activation of AD-related neuroinflammation or lipid peroxidation (Figure 7d).

4. Discussion

4.1. Design Advantages and Detection Mechanism of the Flexible Microneedle Patch

In this study, a petal-like Ti3AlC2-based two-dimensional flexible microneedle patch was fabricated, and its SALDI-MS detection performance was systematically evaluated. Compared with the material modified solely by alkaline etching, the flocculent Ti3AlC2-based material functionalized via CTAB self-assembly increased the intensity of identifiable metabolite, demonstrating significant improvements in both signal intensity and metabolite coverage (Figure 4d). This enhancement can be attributed to two synergistic effects: first, alkaline etching introduces micro-/nano-roughness and oxygen-containing functional groups onto the Ti3AlC2 surface, increasing the density of surface active sites; second, CTAB, as a cationic surfactant, drives electrostatic self-assembly to form a fluffy porous structure, which expands the specific surface area and enhances surface hydrophobicity, thereby promoting metabolite enrichment and energy transfer during laser desorption/ionization.

Furthermore, the flexible microneedle patch retains stable detection signals under high-salt (150 mM NaCl) and high-protein (5% BSA) conditions, demonstrating excellent anti-matrix interference capabilities (Figure 4e,f). This resistance is attributed to the intrinsic chemical stability of the Ti3AlC2-based composite and the physical exclusion of inorganic salts and large proteins by the porous micro-/nanostructure. The micro-/nanopores permit the rapid diffusion of small-molecule metabolites to the active surface while excluding macromolecular proteins from competitive absorption. Compared to existing carbon- or noble-metal-based SALDI matrices, the synthesized flocculent Ti3AlC2 material offers comprehensive advantages in cost, biocompatibility, and anti-interference capacity, providing a reliable platform for the direct analysis of living brain metabolites.

4.2. Metabolomic Profiling and Diagnostic Significance of Differential Metabolites

In this study, conventional unsupervised PCA fails to differentiate the serum metabolic profiles of AD patients from healthy controls (Figure 5a), whereas supervised OPLS-DA achieves distinct separation between the two groups (Figure 5b). This indicates that AD-associated metabolic reprogramming possesses multidimensional and subtle features that supervised learning methods can effectively extract. Multiple significantly differential metabolites are identified via the heatmap and volcano plot, among which arachidonic acid (AUC = 0.79) and proline (AUC = 0.73) show moderate diagnostic potential as single markers (Figure 5c,d). These findings underscore that single metabolic markers offer limited diagnostic efficacy, reflecting the complexity and heterogeneity of metabolic regulation in neurodegenerative diseases.

Pathway enrichment analysis reveals that the differential metabolites are predominantly enriched in arachidonic acid metabolism, glycerophospholipid metabolism, and glutamate metabolism pathways (Figure 6c). Arachidonic acid is a key component of cell membrane phospholipids, and its metabolism involves the synthesis of inflammatory mediators (such as prostaglandins and leukotrienes) mediated by cyclooxygenases and lipoxygenases. During AD pathogenesis, β-amyloid deposition and neuroinflammatory responses can activate phospholipase A2, leading to membrane phospholipid hydrolysis and the release of arachidonic acid, which cascades into inflammatory responses. The localized aggregation of arachidonic acid observed in the parietal region of the AD model is highly consistent with this pathological mechanism (Figure 7d), confirming the central role of lipid metabolic disorders in AD.

4.3. Innovation and Significance of Arachidonic Acid Spatial Aggregation

The core innovation of this study lies in the elucidation of the unique spatial distribution pattern of arachidonic acid at the living parietal lobe level, characterized by centralized aggregation with peripheral diffusion (Figure 8).

This finding is significant in three respects: first, it complements the spatial dimension of metabolomic information. Traditional studies typically report only overall concentration changes, overlooking tissue heterogeneity, whereas our results show that even within the same brain region, metabolite distribution is non-uniform, with arachidonic acid forming localized hotspots that may correspond to initiation sites or active lesion zones of AD pathology, thereby offering a new perspective on focal pathogenesis; second, the focal aggregation pattern implies that lipid metabolic disorders in the parietal lobe do not arise as diffuse global alterations, but rather originate from a local core and spread outward, a “center-to-periphery” propagation that conceptually mirrors the prion-like transmission hypothesis of pathological proteins (Aβ, tau), with the aggregation zone potentially serving as the epicenter of lipid peroxidation and neuroinflammation from which downstream inflammatory mediators (e.g., isoprostanes, leukotrienes) diffuse to drive broader neuronal dysfunction; and third, this spatial pattern provides guidance for targeted therapeutic intervention, as key enzymes in the arachidonic acid pathway (e.g., cytosolic phospholipase A2 and cyclooxygenase-2) are likely highly expressed or activated within the aggregation zone, suggesting that future strategies could exploit local lipid metabolism modulation—through focal delivery of specific inhibitors or nanomedicines—to achieve precision therapy for AD while minimizing systemic side effects.

4.4. Study Limitations

Several limitations in the current study should be acknowledged: (1) the pathological characteristics of the rabbit AD model (Aβ deposition and neurofibrillary tangles) differ to some extent from human AD, requiring further validation in transgenic AD mice or other models closely resembling human pathology; (2) the spatial sampling partition is relatively coarse, warranting the future use of mass spectrometry imaging with higher spatial resolution or array-type multi-channel microneedles to construct finer two-dimensional spatial maps; (3) the current 5 min single-sampling approach lacks dynamic monitoring capabilities, highlighting the need for repeatable or continuous sampling strategies to track the spatiotemporal evolution of metabolites along with AD progression; (4) due to the small sample size (n = 15), the diagnostic efficacy of the combined multi-marker model (AUC = 0.84) must be validated in independent, larger cohorts.

5. Conclusions

A petal-like Ti3AlC2-based 2D material flexible microneedle patch was successfully constructed. Through NaOH etching and CTAB self-assembly, a flocculent, porous structure with a high specific surface area was created, enhancing metabolite adsorption and SALDI-MS detection performance. The patch exhibits excellent mechanical flexibility, salt/protein tolerance, and a wide linear detection range, allowing conformal attachment to the rabbit parietal cortex for in situ metabolite extraction. In parietal metabolic profiling, OPLS-DA achieved clear separation between the AD and control groups, and the multi-marker diagnostic model (AUC = 0.84) outperformed single markers. Crucially, this study first reveals the unique spatial aggregation of arachidonic acid in the AD parietal region, characterized by a centralized gathering and a gradient peripheral diffusion pattern. This finding provides spatial evidence supporting the lipid metabolic reprogramming hypothesis in AD and suggests new avenues for targeted localized interventions.

References

1.Arnold M. Nho K. Kueider-Paisley A. Massaro T. Huynh K. Brauner B. MahmoudianDehkordi S. Louie G. Moseley M.A. Thompson J.W. Sex and APOE Ε4 Genotype Modify the Alzheimer’s Disease Serum Metabolome Nat. Commun. 2020 11 1148 10.1038/s41467-020-14959-w 32123170 PMC7052223

2.Janelidze S. Stomrud E. Smith R. Palmqvist S. Mattsson N. Airey D.C. Proctor N.K. Chai X. Shcherbinin S. Sims J.R. Cerebrospinal Fluid P-Tau217 Performs Better than p-Tau181 as a Biomarker of Alzheimer’s Disease Nat. Commun. 2020 11 1683 10.1038/s41467-020-15436-0 32246036 PMC7125218

3.Karikari T.K. Ashton N.J. Brinkmalm G. Brum W.S. Benedet A.L. Montoliu-Gaya L. Lantero-Rodriguez J. Pascoal T.A. Suárez-Calvet M. Rosa-Neto P. Blood Phospho-Tau in Alzheimer Disease: Analysis, Interpretation, and Clinical Utility Nat. Rev. Neurol. 2022 18 400 418 10.1038/s41582-022-00665-2 35585226

4.Zhang J. Zhang Y. Wang J. Xia Y. Zhang J. Chen L. Recent Advances in Alzheimer’s Disease: Mechanisms, Clinical Trials and New Drug Development Strategies Signal Transduct. Target. Ther. 2024 9 211 10.1038/s41392-024-01911-3 39174535 PMC11344989

5.Peng G. Yang Y. Wang Y. Chandekar S.A. Yu J. Biomarkers in Preclinical and Early Alzheimer’s Disease in China: A Scoping Review J. Prev. Alzheimer’s Dis. 2026 13 100599 10.1016/j.tjpad.2026.100599 42155345 PMC13213292

6.Aureliano M. Ma B. Feature Papers in BioChem BioChem 2025 5 1710.3390/biochem5020017

7.Fan J. Liu S. Luo D. Zhang H. Dai X. Xu H. Shen M. Huang Y. Long Y. Chen M. Advances and Challenges in Point-of-care Testing for Pathogenic Bacteria in Battlefield Injuries Interdiscip. Med. 2026 4 70071 10.1002/inmd.70071

8.Hamdi A. Hosu I.S. Addad A. Hartkoorn R. Drobecq H. Melnyk O. Ezzaouia H. Boukherroub R. Coffinier Y. MoS2/TiO2/SiNW Surface as an Effective Substrate for LDI-MS Detection of Glucose and Glutathione in Real Samples Talanta 2017 171 101 107 10.1016/j.talanta.2017.04.061 28551115

9.Wu Q. Yu J. Zhang M. Xiong Y. Zhu L. Wei B. Wu T. Du Y. Serum Lipidomic Profiling for Liver Cancer Screening Using Surface-Assisted Laser Desorption Ionization MS and Machine Learning Talanta 2024 268 125371 10.1016/j.talanta.2023.125371 37931569

10.Arendowski A. Ossoliński K. Ossolińska A. Ossoliński T. Nizioł J. Ruman T. Serum and Urine Analysis with Gold Nanoparticle-Assisted Laser Desorption/Ionization Mass Spectrometry for Renal Cell Carcinoma Metabolic Biomarkers Discovery Adv. Med. Sci. 2021 66 326 335 10.1016/j.advms.2021.07.003 34273747

11.Chiang C.K. Lin Y.W. Chen W.T. Chang H.T. Accurate Quantitation of Glutathione in Cell Lysates through Surface-Assisted Laser Desorption/Ionization Mass Spectrometry Using Gold Nanoparticles Nanomed. Nanotechnol. Biol. Med. 2010 6 530 537 10.1016/j.nano.2010.01.006 20138245

12.Li Z. Xu E. Zhang Y. Du C. Geng Y. Zhu H. Zhang R. Ma C. Zhang D. Deciphering Spatiotemporal Molecular Pattern of Traumatic Brain Injury by Resveratrol-Engineered Two-Dimensional-Material-Based Field-Effect-Transistor Biopatch Biosens. Bioelectron. 2025 279 117360 10.1016/j.bios.2025.117360 40158492

13.Driscoll N. Erickson B. Murphy B.B. Richardson A.G. Robbins G. Apollo N.V. Mentzelopoulos G. Mathis T. Hantanasirisakul K. Bagga P. MXene-Infused Bioelectronic Interfaces for Multiscale Electrophysiology and Stimulation Sci. Transl. Med. 2021 13 eabf8629 10.1126/scitranslmed.abf8629 34550728 PMC8722432

14.Xie S. Fei X. Wang J. Zhu Y.C. Liu J. Du X. Liu X. Dong L. Zhu Y. Pan J. Engineering the MoS2/MXene Heterostructure for Precise and Noninvasive Diagnosis of Prostate Cancer with Clinical Specimens Adv. Sci. 2023 10 2206494 10.1002/advs.202206494 36988431 PMC10214233

15.Eom W. Shin H. Ambade R.B. Lee S.H. Lee K.H. Kang D.J. Han T.H. Large-Scale Wet-Spinning of Highly Electroconductive MXene Fibers Nat. Commun. 2020 11 2 10.1038/s41467-020-16671-1 32499504 PMC7272396

16.Wang M. Hong Y. He W. Zhang Q. Yu G. Liu J. Chen C. Wen L. Wu D. Cheng Y. Nature-Inspired MXene Electrode with the Highly Interconnected Gradient Nanoconfined Architecture Adv. Mater. 2025 37 202511444 10.1002/adma.202511444 40884169

17.Zhang R. Zhang J. Tan F. Yang D. Wang B. Dai J. Qi Y. Ran L. He W. Lv Y. Multi-Channel AgNWs-Doped Interdigitated Organic Electrochemical Transistors Enable Sputum-Based Device towards Noninvasive and Portable Diagnosis of Lung Cancer Mater. Today Bio 2022 16 100385 10.1016/j.mtbio.2022.100385 35991625 PMC9386496

18.Zhang R. Duan R. Yang D. Wu H. Wang B. Chang Z. Chen J. Dai J. Lv Y. Liu Y. Fluorinated Amino-Acid-Doped Microarray-Based Biochemical Patches Deciphers in-Situ Spatiotemporal Dynamics amidst Intricate Bio-Interfaces Chem. Eng. J. 2025 515 163222 10.1016/j.cej.2025.163222

19.Yang Y. Chen Z. Cheng W. Li F. Wang X. Wang Z. Hu W. MXenes-based Flexible Electronics for Wound Monitoring and Treatment Interdiscip. Med. 2025 3 20250001 10.1002/INMD.20250001

20.Halvaeikhanekahdani P. Zandi S. Ahmad Q. Payravand H. Nanoparticle-integrated Electrostimulation-responsive Biomaterials: Innovations in Drug Delivery and Tissue Repair in Clinical Regenerative Medicine Interdiscip. Med. 2026 4 70077 10.1002/inmd.70077

21.Jiao S. Zhou A. Wu M. Hu H. Kirigami Patterning of MXene/Bacterial Cellulose Composite Paper for All-Solid-State Stretchable Micro-Supercapacitor Arrays Adv. Sci. 2019 6 201900529 10.1002/advs.201900529 31380175 PMC6662096

22.Kosnan M.A. Azam M.A. Safie N.E. Munawar R.F. Takasaki A. Recent Progress of Electrode Architecture for MXene/MoS2 Supercapacitor: Preparation Methods and Characterizations Micromachines 2022 13 183710.3390/mi13111837 36363860 PMC9695226

23.Karimipour M. Paingott Parambil A. Tabah Tanko K. Zhang T. Gao F. Lira-Cantu M. Functionalized MXene/Halide Perovskite Heterojunctions for Perovskite Solar Cells Stable Under Real Outdoor Conditions Adv. Energy Mater. 2023 13 2301959 10.1002/aenm.202301959

24.Liu Z. Zhang Y. Zhang H.B. Dai Y. Liu J. Li X. Yu Z.Z. Electrically Conductive Aluminum Ion-Reinforced MXene Films for Efficient Electromagnetic Interference Shielding J. Mater. Chem. C 2020 8 1673 1678 10.1039/c9tc06304h

25.Wang B. Yang D. Chang Z. Zhang R. Dai J. Fang Y. Wearable Bioelectronic Masks for Wireless Detection of Respiratory Infectious Diseases by Gaseous Media Matter 2022 5 4347 4362 10.1016/j.matt.2022.08.020 36157685 PMC9484046

26.Maciel E.M.A. Silva N.C. Santos L.G.P. Ribeiro F.M. Advances in HiPSC-Derived Brain Organoids as a Model to Study Neuroinflammation in Alzheimer’s Disease J. Neurochem. 2026 170 41906426 10.1111/jnc.70416 41906426 PMC13033922

27.Poejo J. Gumerova N.I. Rompel A. Mata A.M. Aureliano M. Gutierrez-Merino C. Unveiling the Agonistic Properties of Preyssler-Type Polyoxotungstates on Purinergic P2 Receptors J. Inorg. Biochem. 2024 259 112640 10.1016/j.jinorgbio.2024.112640 38968927

28.Wang Q.W. Zhang H.B. Liu J. Zhao S. Xie X. Liu L. Yang R. Koratkar N. Yu Z.Z. Multifunctional and Water-Resistant MXene-Decorated Polyester Textiles with Outstanding Electromagnetic Interference Shielding and Joule Heating Performances Adv. Funct. Mater. 2019 29 1806819 10.1002/adfm.201806819

29.Sun Z. Yuan M. Lin L. Yang H. Nan C. Li H. Sun G. Yang X. Selective Lithiation-Expansion-Microexplosion Synthesis of Two-Dimensional Fluoride-Free Mxene ACS Mater. Lett. 2019 1 628 632 10.1021/acsmaterialslett.9b00390

30.Zhang S. Ying H. Yuan B. Hu R. Han W.Q. Partial Atomic Tin Nanocomplex Pillared Few-Layered Ti3C2Tx MXenes for Superior Lithium-Ion Storage Nano-Micro Lett. 2020 12 78 10.1007/s40820-020-0405-7 34138291 PMC7770861

31.Wang Z. Wang F. Hermawan A. Asakura Y. Hasegawa T. Kumagai H. Kato H. Kakihana M. Zhu J. Yin S. SnO-SnO2 Modified Two-Dimensional MXene Ti3C2Tx for Acetone Gas Sensor Working at Room Temperature J. Mater. Sci. Technol. 2021 73 128 138 10.1016/j.jmst.2020.07.040

32.Driscoll N. Richardson A.G. Maleski K. Anasori B. Adewole O. Lelyukh P. Escobedo L. Cullen D.K. Lucas T.H. Gogotsi Y. Two-Dimensional Ti3C2 MXene for High-Resolution Neural Interfaces ACS Nano 2018 12 10419 10429 10.1021/acsnano.8b06014 30207690 PMC6200593

33.Hui X. Zhao R. Zhang P. Li C. Wang C. Yin L. Low-Temperature Reduction Strategy Synthesized Si/Ti3C2 MXene Composite Anodes for High-Performance Li-Ion Batteries Adv. Energy Mater. 2019 9 201901065 10.1002/aenm.201901065

34.Wu Q. Li N. Wang Y. Liu Y. Xu Y. Wei S. Wu J. Jia G. Fang X. Chen F. A 2D Transition Metal Carbide MXene-Based SPR Biosensor for Ultrasensitive Carcinoembryonic Antigen Detection Biosens. Bioelectron. 2019 144 111697 10.1016/j.bios.2019.111697 31536930

35.Lu N. Gao A. Dai P. Mao H. Zuo X. Fan C. Wang Y. Li T. Ultrasensitive Detection of Dual Cancer Biomarkers with Integrated CMOS-Compatible Nanowire Arrays Anal. Chem. 2015 87 11203 11208 10.1021/acs.analchem.5b01729 26473941

36.Cai L. Zhang S. Miao J. Yu Z. Wang C. Fully Printed Stretchable Thin-Film Transistors and Integrated Logic Circuits ACS Nano 2016 10 11459 11468 10.1021/acsnano.6b07190 28024367

37.Wu M.S. Liu Z. Xu J.J. Chen H.Y. Highly Specific Electrochemiluminescence Detection of Cancer Cells with a Closed Bipolar Electrode ChemElectroChem 2016 3 429 435 10.1002/celc.201500361

38.Lai J.-C. Jia X.-Y. Wang D.-P. Deng Y.-B. Zheng P. Li C.-H. Zuo J.-L. Bao Z. Thermodynamically Stable Whilst Kinetically Labile Coordination Bonds Lead to Strong and Tough Self-Healing Polymers Nat. Commun. 2019 10 1164 10.1038/s41467-019-09130-z 30858371 PMC6411951

39.Meskini I. Capet F. Fraqueza G. Dege N. Tahir M.N. Ayed B. Aureliano M. Dawson- and Lindqvist-Type Hybrid Polyoxometalates: Synthesis, Characterization and Ca2+-ATPase Inhibition Potential Molecules 2025 30 433410.3390/molecules30224334 41302394 PMC12654597

40.Li J.-Q. Ma X.-H. Dai H. Wang C.-C. Zhang J. Meng X.-L. Isoliensinine Ameliorates Cognitive Dysfunction in AlCl3/D-Gal-Induced Alzheimer’s Disease-like Mice by Inhibiting the Calcium Signaling Pathway J. Ethnopharmacol. 2025 344 119567 10.1016/j.jep.2025.119567 40020795

41.Adefisan-Adeoye A.O. Oluwadamilare H. Aisedion M. Shittu T.S. Amangele T. Adeoye T.D. Unuofin J.O. Orimolade B.M. Lebelo S.L. Adaramoye O.A. Neuromodulatory Effects of Berberine Chloride against Aluminum Chloride/D-Galactose Induced Alzheimer-like Neurodegeneration Model in Rats Sci. Rep. 2026 10.1038/s41598-026-61600-9 42449163

42.Onigbinde S. Solomon J. Gutierrez-Reyes C.D. Daramola O. Fowowe M. Adeniyi M. DuBois K.N. Bakulski K.M. Kanaan N.M. Lubman D.M. Serum N-Glycan Profiling Identifies Candidate Glycan Biomarkers for Early Detection and Prediction of Alzheimer’s Disease J. Proteome Res. 2025 24 4417 4436 10.1021/acs.jproteome.5c00018 40792479 PMC12814892

43.Arjun A.M. Deshpande S. Liu G. Miura D. Pawlak K. Takahiko T. Higuchi M. Matsumoto M. Umemura T. Tsukakoshi K. Low-Cost Polyphenol–Polypyrrole Molecularly Imprinted Sensor for Point-of-Care Alzheimer’s Detection ACS Sens. 2025 10 8435 8446 10.1021/acssensors.5c01816 40814138 PMC12671001