Nanomechanical Systems for the rapid detection of HIV-1 p24 antigen

Abstract

Early detection of HIV is crucial for reducing transmission and ensuring timely initiation of antiretroviral therapy (ART), significantly improving patient outcomes. Although diagnostic tests have advanced from first-generation antibody detection assays to fourth-generation immunoassays that detect both HIV antibodies and the p24 antigen, these are limited to clinical labs. Their longer processing times, high costs, and the requirement for multiple patient visits highlight the need for rapid, affordable point-of-care (POC) diagnostics. This study introduces a nanomechanical cantilever-based biosensor for the rapid detection of HIV-1 p24 antigen, a key marker for early diagnosis. The platform demonstrated remarkable sensitivity, detecting p24 at concentrations as low as 100 fg/mL in solution and 1 pg/mL in human serum, and was quantitative within several orders of magnitude. After functionalizing the microcantilevers with two broadly cross-reactive monoclonal antibodies (ANT-152 and C65690M), the system was able to detect p24 from a wide range of HIV-1 subtypes. Furthermore, this biosensor was found to be compatible with various blood processing methods and with a direct electronic output. This platform’s high sensitivity, specificity, and applicability across multiple HIV subtypes underscores its potential for future development into a next-generation POC diagnostic tool.

1. INTRODUCTION

Human immunodeficiency virus (HIV) infection remains a major global public health concern, with an estimated 1.3 million new infections occurring in 2022 alone. As there is no vaccine or cure for HIV, treatment and prevention efforts have focused on the further development and deployment of highly active antiretroviral therapy (ART). The Joint United Nations Programme on HIV/AIDS (UNAIDS) reflects an international effort to end the HIV epidemic with the 2025 goal of having 95% of people living with HIV (PLWH) aware of their status, 95% of people diagnosed with HIV on ART, and 95% of people on ART maintaining viral suppression [1]. Unfortunately, most countries are not on track to meet these goals due to a myriad of challenges that include a lack of point-of-care testing and complex diagnostic algorithms; barriers to accessing and retaining healthcare; stigma and discrimination in seeking testing and care; and a lack of resources [2].

The Centers for Disease Control and Prevention (CDC) currently recommends a two-step HIV testing algorithm that starts with a fourth- or fifth-generation HIV-1/2 combined antigen/antibody immunoassay followed by a second confirmatory test that can differentiate between HIV-1 and HIV-2 [3]. HIV p24 antigen can be detected in the blood in as little as two weeks following initial exposure with anti-HIV antibodies detectable 3–12 weeks following exposure [4]. In some circumstances (i.e., in the case of indeterminant results, early in acute infection, for the detection of mother-to-child transmission, etc.), a nucleic acid test (NAT) may be employed for the detection of HIV viral RNA in the blood [5]. Most NATs provide a quantitative readout of viral load, and so while they are not suitable for diagnosis at all stages of infection, they are often used in clinical management of disease to monitor viral suppression on ART [6].

Early diagnosis of HIV infection and subsequent linkage to sustainable care remains a critical component of combating the epidemic. Early diagnosis and initiation of ART is linked to improved patient outcomes and a decreased risk of HIV transmission [7]. Current testing algorithms rely on laboratory-based diagnostic services that can pose a substantial barrier to testing [8]. Community-based testing services that offer point-of-care (POC) testing can improve early diagnosis by enhancing test availability and accessibility. Ideally, POC tests are cost effective to administer, easy to perform, and provide rapid results, which can then be used to inform patient care in real-time while HIV status is validated using standard testing algorithms [3]. Unfortunately, most POC tests are limited to the detection of anti-HIV antibodies, which may not be detectable in the blood or saliva until 3–12 weeks after exposure [9]. POC tests that can also detect HIV p24 antigen at high sensitivity akin to current fourth- or fifth-generation tests would help close this diagnostic gap and further improve early diagnosis efforts [3].

One of the hurdles to developing an accurate HIV p24 test is the diversity of the antigen across subtypes. HIV is highly divergent, and different types, groups, and subtypes are found across the globe at varying frequencies [10]. There are two main types of HIV, HIV-1 and HIV-2, with HIV-1 causing the majority of infections worldwide. HIV-1 is further divided into multiple groups (M, N, O, and P), of which M is the most prevalent. Group M further encompasses additional subtypes (A-D, F-H, J, K) and circulating recombinant forms (CRFs). Changes in p24 amino acid sequence across HIV strains can lead to failure of detection by antigen-based diagnostics [11,12]. This has led to efforts to identify antibody combinations that can detect a diverse array of HIV-1 subtypes. For example, a recent study identified two broadly cross-reactive HIV-1 p24 monoclonal antibodies (mAbs), C65690M and ANT-152, which together were able to detect 166 highly diverse viruses representing a broad range of HIV-1 subtypes and CRFs [13].

Recent advances in microcantilever-based nanomechanical platforms have enabled the high sensitivity detection of viral antigens in clinical specimens [14]. These platforms utilize specific antibodies bound to an immunosensor in which the interaction of antibody with antigen is translated into nanomechanical motion of a cantilever, allowing for antigen detection [15]. These platforms can be readily multiplexed for the rapid and sensitive quantification of multiple analytes. Here, we report the development of a high sensitivity microcantilever device functionalized with broadly reactive HIV-1 specific antibodies for the detection of HIV-1 p24 antigen. We demonstrate that this platform can accurately detect and quantify HIV-1 p24 across a broad range of subtypes with femtomolar sensitivity in a less than a 10-minute reaction time. This technology shows promise for the future development of combined antigen/antibody POC tests for the early diagnosis of HIV.

2. MATERIALS & METHODS

2.1. Cells and Reagents

All chemicals and reagents were analytical grade and sourced from commercial vendors as follows. The following reagents were obtained through the NIH HIV Reagent Program, Division of AIDS, NIAID through the NIH BEI Resources Repository: HIV-1 IIIB recombinant p24 protein (Cat. # ARP-12028); HIV-1 HXB2 recombinant p24 protein (Cat. # ARP-13126); and anti-HIV-1 p24 monoclonal antibody clone 183-H12-5C (Cat. # ARP-3537). Anti-HIV-1 p24 broadly cross-reactive monoclonal antibodies C65690M and ANT-152 were purchased from Meridian Biosciences and ProSci, respectively. Phosphate Buffered Saline (PBS), PBS-Tween 20 sachets, Bovine Serum Albumin (BSA), and 11-Mercaptoundecanoic acid (MUA) were purchased from Millipore-Sigma. Dulbeccco’s Phosphate Buffered Saline (DPBS) was purchased from Corning. 1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) and Sulfo-NHS were purchased from Thermo Fisher Scientific. Tipless, silicone cantilevers were purchased from Nanoworld Incorporation. Human embryonic kidney (HEK) 293T cells were purchased from ATCC (#CRL-3216) and cultured in 1x Dulbecco’s Modified Eagle Medium (DMEM) (Corning) supplemented with 10% heat inactivated fetal bovine serum (Gibco) and 1% penicillin/streptomycin (Cytiva). A panel of 60 HIV-1 isolates representing diverse subtypes from chronically infected individuals was sourced from the NIH BEI Resources Repository (Cat. # HRP-11412), contributed by Dr. Robert Gallo, Dr. Nelson Michael, Dr. Smita Kulkarni, Dr. Victoria Polonis and The Joint United Nations Programme on HIV/AIDS (UNAIDS). Blood products from anonymous, healthy, human donors were sourced through Vitalant. Serum was isolated by Ficoll centrifugation and stored at −80°C.

2.2. Preparation of HIV-1 NL4-3 Virus Stocks

HIV-1 NL4-3 viral particles were produced using an HIV-1 NL4-3 molecular clone with GFP cloned behind an IRES cassette following the viral nef gene (NEF-IRES-GFP) sourced from the NIH BEI Resources Repository (Cat. # ARP-11349). Human embryonic kidney 239T (HEK293T) cells were plated at a density of 5×10⁶ cells per 15 cm tissue culture dish 24 hours prior to transfection. 10 μg of HIV-1 NL4-3 NEF-IRES-GFP plasmid DNA, 250 μL serum-free DMEM, and 30 μL of PolyJet (SignaGen Laboratories) were combined per 15 cm dish of cells and transfected according to the manufacturer’s protocol. 25 mL of viral supernatant was collected from each plate at 48 hours post-transfection and replaced with fresh media. A second set of supernatants were collected at 72 hours post-transfection. Viral supernatants were combined and filtered through 0.2 μM filters, then incubated with 8.5% polyethylene glycol and 0.3 M NaCl for 8 hours at 4°C to precipitate out the virus. Tubes were centrifuged at 3500 rpm for 20 minutes, and the virus-containing pellet was resuspended in 250 μL DPBS for 100x concentration. Virus aliquots were stored at −80°C prior to lysate preparation.

2.3. ELISA for p24 Quantification of Virus Stocks

HIV-1 NL4-3 NEF-IRES-GFP viral stocks underwent heat inactivation at 60°C for 30 minutes and were then quantified using an HIV-1 Gag p24 Quantikine ELISA kit (R&D Systems, #DHP240B) according to the manufacturer’s protocol. Heat-inactivated virus samples were tested in technical duplicate at dilutions of 1:500 and 1:2500. Viral stock concentration was determined using a standard curve.

2.4. HIV-1 Lysate Preparation

A panel of 60 HIV-1 isolates from chronically infected individuals was used to generate viral lysates for experiments shown in Figures 3, 4, and S5 (obtained from BEI resources, HRP-11412). The panel consists of 10 virus samples each of subtypes A, B, C, D and circulating recombinant forms CRF01_AE and CRF02_AG. Lysates from the subtype B lab-adapted HIV-1 NL4-3 strain was used for experiments shown in Figures 2, S3 and S6 Briefly, viral particles were heat inactivated by incubation at 60°C for 30 minutes. 200 μL of each viral sample was lysed through the addition of 20 μL of 10% Triton X-100 in Milli-Q water. Samples were vortexed and incubated at 37°C for 1 hour to complete the lysis. Lysed samples were further diluted as needed in a buffer comprised of sterile-filtered 1% BSA and 0.2% Tween-20 diluted in Roswell Park Memorial Institute (RPMI) 1640 media.

Figure 3 |. Monoclonal HIV-1 p24 antibody ANT-152 enables detection of a subset of HIV-1 isolates.

Response curves for the detection of HIV-1 p24 from lysed viral particles from 10 strains representing subtypes a) A, b) B, c) C, d) D, e) CRF01-AE, and f) CRF02-AG. The curves represent deflection of the microcantilever functionalized with anti-p24 ANT-152 mAb (20μg/mL) as measured by f-AFM over the course of a 16 minutes. The p24 concentration of each isolate as reported by the commercial vendor is provided in the respective panel legends. Graphs show the average of triplicate deflection measurements in nm +/− standard deviation per timepoint and condition.

Figure 4 |. Multiplexed HIV-1 p24 antibodies enable broad detection of diverse HIV-1 isolates.

Response curves for the detection of HIV-1 p24 from lysed viral particles from 10 strains representing subtypes a) A, b) B, c) C, d) D, e) CRF01-AE, and f) CRF02-AG. The curves represent deflection of microcantilevers functionalized with a 1:1 equimolar mixture of ANT-152:C65690M anti-p24 antibodies (20μg/mL) as measured by f-AFM over the course of a 12 minutes. The p24 concentration of each isolate as reported by the commercial vendor is provided in the respective panel legends. Graphs show the average of triplicate deflection measurements in nm +/− standard deviation per timepoint and condition.

Figure 2 |. Dose-dependent detection of recombinant HIV-1 p24 using a microcantilever platform.

Dose-dependent response curves for the detection of recombinant HIV-1 IIIB p24 suspended in a) PBS or b) serum from a healthy human blood donor at the indicated concentrations. The curves represent the deflection of a microcantilever functionalized with anti-p24 mAb 183-H12–5C (20μg/mL) in triplicate as measured by f-AFM over the course of 16 minutes. Graphs show the average of triplicate deflection measurements in nm +/− standard deviation per timepoint and condition. Statistical comparisons were made between the serum-only control and the p24 antigen conditions using an ordinary one-way ANOVA. ** = p-value < 0.01, *** = p-value < 0.001, **** = p-value < 0.0001.

2.5. Processing of Recombinant p24 Antigen in Human Blood

2.5.1. Processing Method A (Lysis prior to clotting)

Donor blood collected without additives and was divided into four fractions and recombinant HIV-1 p24 antigen (either HIV-1 HXB2 or HIV-1 IIIB p24), was added directly to the fresh blood sample at a concentration of either 50 ng/mL or 1 ng/mL per tube. Samples were then mixed with lysis buffer consisting of 10% v/v Triton X-100 in ultrapure water at a ratio of 9:1 (blood product to lysis buffer) to ensure efficient lysis of cells and release of intracellular components. The sample was left at room temperature to clot for 20 minutes, then tubes were centrifuged at 400xg for 10 minutes to pellet the clotted fraction. The supernatant was then removed and used for subsequent analysis.

2.5.2. Processing Method B (Clotting prior to lysis)

Donor blood collected without additives was divided into four fractions and recombinant HIV-1 p24 antigen was added as indicated in Method A. The sample was left at room temperature to clot for 20 minutes, then tubes were centrifuged at 400xg for 10 minutes to pellet the clotted fraction. The supernatant was then removed and mixed with lysis buffer consisting of 10% v/v Triton X-100 in ultrapure water at a ratio of 9:1 (blood product to lysis buffer). Samples were then used for subsequent analysis.

2.5.3. Processing Method C (Lysis with anticoagulant)

Donor blood collected and mixed with sodium heparin prior to being divided into four fractions and recombinant HIV-1 p24 antigen was added as indicated in Method A. Samples were then mixed with lysis buffer consisting of 10% v/v Triton X-100 in ultrapure water at a ratio of 9:1 (blood product to lysis buffer). Samples were immediately used for subsequent analysis.

2.6. Functionalization of Microcantilevers

Gold-coated microcantilevers were washed by immersion in deionized water for 1 hour followed by isopropanol for 1 hour, followed by a final wash in deionized water. Prior to functionalization, the microcantilevers were plasma cleaned to remove any potential contaminants. Monoclonal anti-HIV-1 p24 antibodies (183-H12–5C, ANT-152, C65690M, or a 1:1 ANT-152:C65690M mixture) or anti-albumin IgG polyclonal antibody (ThermoFisher PA5–89332) were covalently attached to the gold-coated cantilever surface through 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) and N-hydroxysuccinimide (NHS) coupling chemistry inside a glass cell culture well plates as below.

200 μL of DPBS solution was added to the wells containing the microcantilevers and incubated for 5 minutes. The DPBS solution was then carefully removed using a micropipette. To form the self-assembled monolayer (SAM) on the cantilever surface via thiol-metal (gold) bonding, the cantilevers were immersed in 200 μL of a 10 mM solution of 11-Mercaptoundecanoic acid (MUA) prepared in ethanol for 30 minutes, followed by rinsing three times with deionized water. The carboxyl anions on the SAM surface were then activated by adding 100 μL of a 1:1 mixture of 5 mM Carbodiimide EDC and 5 mM Sulfo-NHS in deionized water for a 40-minute incubation to attach the amino radicals of other biomolecules via amide linkage. Next, the cantilevers were rinsed three times with deionized water and submerged in a 50 μL solution of 52 mM APBA (3-aminophenylboronic acid monohydrate, 98%) (10 mM PBS at pH 7.4) for 3 hours. The SAM-functionalized microcantilevers were then rinsed three times with DPBS buffer and incubated overnight at 4°C in 100 μL of 20 μg/mL antibody solutions (either anti-p24 183-H12–5C, ANT-152, C65690M, a 1:1 ANT-152:C65690M mixture, or anti-albumin IgG) prepared in DPBS and 0.05% BSA (pH 7.4) to facilitate covalent immobilization. BSA was used to block the remaining sensor surface to minimize non-specific interactions. Before performing real-time immunoassay experiments, the microcantilevers were washed with PBS-Tween-20 solution, dried, and then fixed in the AFM sample holder.

We employed scanning electron microscopy (SEM) to thoroughly characterize tipless, gold-coated silicon cantilevers before and after functionalization (Figure S1a–b). Our SEM data showed major topological differences in the gold-coated surface of our cantilevers before and after antibody immobilization as measured by root mean square (RMS) roughness. All cantilevers used in this study had a consistent and uniform size, measuring 500 μm in length, 95 μm in width, and 1 μm in thickness (Figure S1c). The force constant of these cantilevers was approximately 0.03 N/m.

To further validate functionalization of our cantilevers, we performed Nanoscale Infrared (NanoIR) characterization. Antibodies were immobilized on the gold-coated microcantilevers via EDC-NHS surface chemistry as above. The stable amide bond after EDC-NHS chemistry is characterized by the appearance of absorption peaks in the NanoIR scan, including the amide carbonyl stretching band at 1700–1713 cm⁻¹, a C-N stretching band at 1256–1262 cm⁻¹, and an N-H bending band at 1540–1570 cm⁻¹. Additional characteristic peaks include C=C stretching of the aromatic phenyl group at 1500–1600 cm⁻¹, boron vibration signals at 1300–1450 cm⁻¹. The binding of antibodies to APBA via boronate covalent bonding is further confirmed by the presence of Amide I carbonyl stretching peaks at 1600–1700 cm⁻¹, Amide II peaks at 1500–1600 cm⁻¹, and Amide III peaks at 1200–1350 cm⁻¹ (Figure S1d).

2.7. Experimental Detection Method

The microcantilever platform employed in these experiments is hinged at one end and free at the other, allowing for deflection of the free end. This platform functions through deflection at the free end of the microcantilever caused by compressive surface stress. In our experiments, the deflection of the cantilever is due to antigen-antibody interactions, which causes changes in surface stress, allowing for transformation of biomolecule detection into mechanical motion. This deflection can be monitored in real time, leading to highly sensitive measurement of even small deflections of the microcantilever. The deflection of p24 antibody-coated microcantilevers in response to incubation with p24 antigen was determined by an optical detection method performed using fluidic-atomic force microscopy (f-AFM). The deflection experiments were carried out in a microfluidic microwell reaction chamber containing antigen solution (10 μL) at a constant temperature, where the functionalized microcantilevers were brought into proximity with the help of a stepper motor. The Bruker Bioscope Resolve liquid imaging system was used for deflection measurements.

In static mode, the relationship between the microcantilever deflection data ($\Delta \mathrm{z}$) and differential surface stress ($\Delta \mathrm{\sigma }$) is described using the modified Stoney’s equation, expressed as:

$$\Delta \sigma =\frac{E{T}^2}{4\left(1-v\right)L^2}\Delta z$$ Equation 1

Where $L$ represents the effective length of the cantilever, $t$ denotes the thickness, and $E/1-\nu$ is the ratio of Young’s modulus $E$ (130 GPa for silicon) to Poisson’s ratio $\nu$ (0.28 for silicon). The force constant of these cantilevers was approximately 0.03 N/m. All real-time immunobinding deflection experiments were performed in triplicate using three independent sets of microcantilevers. Optimization experiments were performed to determine ability of the microcantilever system to detect p24 antigen across different concentrations, in different sample diluents (PBS versus human serum), across different HIV-1 subtypes, and using different methods of blood sample processing.

As an alternate method for readout through MOSFET (Metal-Oxide-Semiconductor Field-Effect Transistor) cantilever technology was also used as proof-of-concept in Figure 6. Briefly, MOSFET leverages differential readout mechanisms, where one cantilever is coated with a sensing layer (gold-coated) and the other serves as a reference (silicon nitride). This differential allows for the precise measurement of drain current as an alternate output for cantilever deflection (refer to [[16–18]). It promises prospects for handheld, rapid, and real-time diagnostics sensor system, will likely have positive impact on the measurements of disease burden. It offers such a cost-effective, label-free and widely deployable scheme for parallel and multiplexed detection of varied analytes. Our data demonstrates that the shift in the drain current occurs within the 3 minutes making it an ultra-rapid and sensitive detection method.

Figure 6 |. Detection of HIV-1 p24 using a MOSFET device.

a) A biofunctionalized MOSFET device demonstrates rapid p24 antigen detection within 3 minutes, evidenced by a measurable shift in drain current, b) MOSFET based detection of HIV p24 protein in different concentration (in spiked serum).

2.8. Statistical Analysis

All statistics were calculated using GraphPad Prism version 10.4.1. Statistical comparisons of experimental versus control microcantilever deflection values were calculated using an ordinary one-way ANOVA followed by a Dunnett’s test. Statistical significance was determined by p-value <0.05. All statistics for each figure are reported in Supplemental Tables.

3. RESULTS & DISCUSSION

3.1. A highly sensitive microcantilever platform detects HIV p24 in solution and in human serum from healthy donors.

Platforms that enable the rapid and sensitive detection of HIV p24 antigen are crucial for the development of future HIV testing modalities. Towards this end, we developed a microcantilever-based platform as a novel method of p24 antigen detection. In this platform, engagement of anti-HIV-1 p24 antibodies with their cognate antigen leads to deflection of a microcantilever, which can be detected via fluidic-atomic force microscopy (f-AFM). As the basis for our testing platform, gold-coated microcantilevers were functionalized by covalent binding of an anti-HIV-1 p24 antibody to the cantilever surface through EDC-NHS coupling chemistry (Figure 1a). The deflection measurement corresponds to the compressive stress generated on the microcantilever surface through antigen:antibody binding, allowing for the sensitive, quantitative assessment of analyte binding (Figure 1b).

Figure. 1. Schematic of Microcantilever-Based HIV-1 Testing Platform.

a) Monoclonal p24 antibodies are immobilized on the surface of gold-coated microcantilevers using EDC-NHS chemistry. b) Anti-HIV-1 p24 antibodies immobilized on the surface of microcantilevers bind to p24 antigen, leading to deflection of the microcantilever as detected by fluidic-atomic force microscopy (f-AFM).

To determine if a microcantilever based platform can efficiently detect p24 antigen, we began by functionalizing microcantilevers in triplicate with anti-HIV-1 p24 monoclonal antibody (mAb) clone 183-H12–5C at a concentration of 20 ug/mL. Three sets of microcantilevers were used to conduct each experiment, with each set running three separate assays per condition. The functionalized microcantilevers where then used to detect recombinant HIV-1 IIIB p24 antigen suspended in PBS at a range of concentrations from 100 fg/mL to 1 μg/mL. Deflection of the microcantilevers was measured using f-AFM every minute over the course of 16 minutes (Figure 2a, Table S1). Deflection was detectable within 2–4 minutes following antigen exposure, with dose-dependent differences in deflection apparent by 6 minutes. We observed maximum deflection by 14 minutes, at which point the signal was saturated with an average deflection of 69.8 nm at the highest concentration (1 μg/mL) to 7.96 nm at the lowest concentration (100 fg/mL). The observed signals were generally reproducible within ± 3–4 nm for a given concentration between experimental replicates. Deflection at this lowest concentration was statistically different than the PBS control (p-value < 0.0001), suggesting that the limit of detection (LOD) for our assay under these conditions is at least 100 fg/mL (0.1 picogram/mL).

There was negligible deflection of the microcantilever incubated with PBS buffer alone, indicating that the signal was specific to the antigen. To further evaluate the specificity of p24 detection by our testing platform, we functionalized microcantilevers using a polyclonal anti-albumin IgG antibody. The microcantilevers were then incubated with p24 antigen from lysed HIV-1 NL4-3 viral particles and deflection was measured by f-AFM over the course of 15 minutes. We observed less than 1 nm of microcantilever deflection under these conditions, indicating that the deflection observed in our experiments is due to specific recognition of p24 antigen by anti-p24 antibodies (Figure S2).

We next explored the impact of human serum proteins on assay sensitivity. As before, microcantilevers were functionalized with anti-p24 mAb 183-H12–5C at 20 ug/mL. Recombinant HIV-1 IIIB p24 antigen was then spiked into serum isolated from a seronegative human blood donor at concentrations ranging from 1 pg/mL to 1 ng/mL. Real-time deflection measurements showed that the deflection kinetics p24 samples in serum were similar to what was observed for p24 samples suspended in PBS, with initial antigen binding occurring 2–4 minutes following sample introduction and signal saturation occurring by 14 minutes (Figure 2b, Table S2). While the serum-only control exhibited a greater deflection than what was observed in the PBS control (Figure 2a), it was still statistically different (p-value < 0.0001) compared to the lowest concentration of p24 tested (1 pg/mL). Taken together, these results demonstrate that this microcantilever platform is capable of detecting HIV-1 p24 antigen in human serum at concentrations as low as 1.0 pg/mL.

3.2. Optimization of p24 detection using a broadly cross-reactive anti-p24 antibody

Although HIV-1 p24 is relatively conserved, differences in p24 primary sequences across HIV-1 subtypes have been linked to suboptimal performance of previous antigen tests [11,12]. Thus, proper selection of the antibody used in our microcantilever-based immunoassay is critical to ensure our platform is capable of detecting a broad array of HIV isolates. Previous studies identified a combination of monoclonal broadly cross-reactive p24 antibodies, ANT-152 and C65690M, which together detected all tested HIV subtypes [13]. To test whether our microcantilever platform would be compatible with different antibodies against different HIV-1 subtypes, we first functionalized microcantilevers with ANT-152 mAb at 20 μg/mL as before. However, rather than rely on recombinant antigen, which is only readily available for select few subtypes, we performed deflection assays using lysed HIV-1 NL4-3 viral particles suspended in PBS at concentrations ranging from 100 fg/mL to 1 ng/mL (the stock concentration was determined using a p24 ELISA assay compatible with this strain). Deflection kinetics over the course of 16 minutes were similar to what was observed using the 183-H12–5C mAb and recombinant antigen in Figure 2a, with deflection beginning 2–3 minutes following sample introduction (Figure S3, Table S3). Once again, deflection was dose-dependent with the highest p24 concentration (1 ng/mL) yielding the largest deflection (35.28 nm) and the lowest concentration (100 fg/mL) yielding the lowest deflection (7.13 nm). There was negligible deflection observed in the PBS buffer-only control, which was significantly different from the 100 fg/mL specimen at the 15-minute measurement (p-value < 0.0001) (Table S3).

To determine if the change in antibody meaningfully impacted platform readout, we compared the performance of microcantilevers functionalized with either the 183-H12–5C mAb or the broadly cross-reactive ANT-152 mAb using recombinant HIV-1 IIIB p24 diluted in healthy human serum (Figure S4a, Table S4). The microcantilevers functionalized with the 183-H12–5C mAb yielded a maximum deflection of 21.82 nm at the highest p24 concentration of 1 ng/mL and a minimum deflection of 7.15 nm at the lowest p24 concentration of 1 pg/mL. Microcantilevers functionalized with ANT-152 mAb demonstrated a slightly larger range of deflection measurements, with a maximum measurement of 28.01 nm for the highest p24 concentration and a minimum deflection signal of 6.41 nm observed for the lowest p24 concentration (Figure S4a, Table S4). Plotting the deflection results versus the p24 concentration in a logarithmic scale, all concentrations appear to be in the linear range though with more divergence at higher concentrations (Figure S4b). Altogether these data suggest that the platform is amenable to multiple antibodies, but that the choice of antibody may influence platform sensitivity.

3.3. Microcantilevers functionalized using the broadly cross-reactive antibodies detect p24 from a subset of HIV-1 isolates

To determine whether microcantilevers functionalized with ANT-152 mAbs were capable of detecting p24 from a broad array of clinical isolates, we tested our platform across a panel of 60 clinical HIV-1 isolates representing 10 strains each for subtypes A, B, C, D, and circulating recombinant forms CRF01_AE and CRF02_AG. As before, we conducted real-time immunobinding assays using microcantilevers functionalized with ANT-152 mAbs to detect p24 from lysed viral samples suspended in PBS. Deflection kinetics were largely similar for samples across viral subtypes, with maximal deflection achieved by 14 minutes following sample addition (Figure 3a–f, Table S5). In general, deflection values correlated with the known p24 concentrations of each viral isolate as provided by the commercial vendor with higher p24 concentrations resulting in greater deflection values. However, p24 antigen detection was highly strain dependent. Out of the panel of 60 HIV-1 isolates, 70% of HIV-1 subtype A and recombinant AG isolates, 60% of recombinant AE isolates, and 50% of Subtypes B, C, and D isolates were detected with the ANT-152 mAb. This suggests that multiple antibodies may be required to enable broad detection.

Given that 25 of the 60 isolates were unable to be detected with the ANT-152 mAb, we next tested our platform using an alternate broadly cross-reactive anti-p24 mAb, C65690M. Microcantilevers were functionalized with the C65690M mAb at 20μg/mL as before and real-time immunobinding was measured for p24 antigen from the 25 HIV-1 isolates that were not detected by ANT-152. With this antibody, initial antigen-antibody binding was observed within 2 minutes, and maximal binding was achieved within 10 minutes (Figure S5a–f, Table S6). Critically, p24 antigen was detected from all 25 of the tested viral lysates. Again, maximum deflection values at the assay endpoint largely correlated with reported p24 concentrations of each specimen, though this was not always the case (i.e., one subtype B isolate resulted in significant, but minimal deflection despite a reported concentration of 66.4 ng of p24/mL, Figure S5b). Nevertheless, these results suggest that the use of multiple, broadly cross-reactive anti-p24 mAbs is likely to enable detection of a majority of HIV-1 isolates across an array of circulating subtypes.

3.4. Microcantilevers functionalized with two multiplexed anti-p24 antibodies detect p24 from a broad array of HIV-1 isolates

Our previous results showed that two broadly cross-reactive antibodies, ANT-152 and C65690M, were able to detect a diverse array of HIV-1 isolates from 6 different subtypes when assessed independently. To determine whether these antibodies could be used in combination to improve p24 detection in our microcantilever-based platform, we functionalized microcantilevers using a 1:1 equimolar mixture of the ANT-152 and C65690M mAbs. These microcantilevers were then used to detect p24 in lysed viral samples from the previously described panel of 60 HIV-1 isolates representing 6 HIV-1 subtypes. Deflection kinetics for the 1:1 equimolar mixture were similar to what was observed with the C65690M mAb alone (Figure S5), with initial antigen-antibody binding occuring within 2 minutes, and signal saturation occuring in 10 minutes. The combination of the two antibodies, however, enabled detection of all 60 isolates with overall deflection largely correlating with p24 concentration (Figure 4a–f, Table S7).

To determine the limit of detection for our microcantilever platform optimized with a 1:1 mixture of ANT-152 and C65690M mAbs, we conducted real-time deflection measurements to detect p24 antigen from lysed HIV-1 NL4-3 viral particles suspended in PBS at a range of concentrations (1.0 fg/mL - 10 ng/mL). Deflection was detectable within 2 minutes following sample addition, with maximal defection occurring by 10 minutes (Figure S6, Table S8). Use of multiplexed ANT-152 and C65690M mAbs led to an improvement of sensitivity, allowing for detection of p24 at concetrations as low as 1.0 fg/mL with a deflection of 7.10 nm at 12 minutes, which represented a statistically significant difference from the PBS control (p-value < 0.0001). Overall, these results indicate that our optimized microcantilever-based platform using multiplexed anti-p24 antibodies is capable of rapidly detecting p24 antigen from a broad array of HIV-1 isolates with high sensitivity in less than 15 minutes.

3.5. Antigen-antibody binding kinetics and Hill plot equation

The kinetics of antigen-antibody binding can be influenced by properties that occur at solid-liquid interfaces [20]. Although the forward reaction rate in an antigen-antibody reaction is not normally limited by diffusion, reactions at a solid-liquid interface can be limited by diffusion if reactants are depleted close to the surface, as may be the case with our microcantilever testing platform. Thus, we sought to better understand the kinetics of antigen-antibody binding in our microcantilever platform by fitting a linear Hill plot to microcantilever deflection data. Linear Hill plots are an enzyme kinetic model that allows for calculation of a measure of cooperativity of antigen-antibody binding. It can also be used for calculation of the dissociation (K_d) and association (K_a) binding equilibrium constants for enzyme-based assays or biosensor systems such as our microcantilever platform.

In our previous experiments, we show that real-time deflection measurements of binding between p24 antigen and p24 mAbs in our microcantilever platform display a sigmoidal curve shape, indicating that the initial forward reaction becomes diffusion rate limited at the plane surface of the microcantilever (i.e., Figure S6). The high surface concentration of immobilized p24 mAbs binds with p24 antigen at a slow diffusion rate, followed by an association binding reaction. This reaction reaches a concentration-dependent saturation level that is not due to a dynamic equilibrium. The binding kinetics of p24 antigen binding to p24 mAbs on the microcantilever sensor can be represented by the reaction (equation 2) shown below [21].

$$nAg+mAb\underset{k_d}{\overset{k_a}{\rightleftharpoons }}n[Ag-mAbcomplex]$$ Equation 2

The terms ‘$\text{Ag}$’ and ‘$\text{mAb}$’ denote the antigen and antibody, respectively, and the antigen-antibody complex is denoted by [$\text{Ag-mAb}$ complex]. The dissociation constant K_d is derived from the intercept of the linear Hill plot using a logarithmic transformation of concentration versus deflection response. The Hill coefficient, $n$, is an important parameter in the study of antigen-antibody interactions and reflects the degree of cooperativeness of the substrate with the available binding sites. ‘$\text{n}$’ is the number of antigen moieties bound to the mAbs immobilized on the cantilever. $\text{n}>1$ indicates positively cooperative binding, $\text{n}<1$ indicates negatively cooperative binding, and $\text{n}=1$ indicates noncooperative/independent binding. The values of ‘$\text{n}$’ and ‘Kd’ are calculated by constructing a linear Hill plot, which is generated using following equation (Equation 3) [22].

$$ln(y/1-y)=ln\left(k_a\right)+nln\left(Ag\right)$$ Equation 3

We constructed linear Hill plots (Figure S7) using the previously collected real-time deflection measurements (shown in Figure 4) from microcantilevers functionalized with a 1:1 equimolar mixture of ANT-152:C65690M mAbs tested against viral lysates from 60 diverse HIV-1 isolates suspended in PBS. In our calculations using Equation 1, Y is the quotient of the maximum deflection measurement obtained for the highest concentration of p24 analyte among all 60 isolates minus the blank value and the maximum deflection obtained for each concentration of the respective p24 subtype concentration minus the blank value. Blank readings were recorded during experiments with PBS buffer on the antibody-functionalized microcantilever.

As shown in Table S9, at the tested concentrations the association equilibrium binding constant (K_a) is significantly higher than the dissociation constant (K_d), indicating that the microcantilevers functionalized with a 1:1 equimolar mixture of ANT-152:C65690M mAb have a strong affinity for p24 antigen across all tested HIV-1 subtypes. The lower K_d value suggests that the antigen-antibody complex is less likely to dissociate from the cantilever surface, leading to a complex that is stable over time. Each HIV subtype has a different Hill coefficient (n), indicating that there is variation in the cooperative binding behavior between the mAbs and the different p24 antigens. Some of this variation may be due to differential affinity for specific isolates and one of the two antibodies employed.

3.6. Dosage response study using sigmoidal fit

As shown in Figure 4, real-time deflection measurements for our microcantilever testing follow a sigmoidal curve. Thus, we also examined the dose-response of our assay by fitting a sigmoidal curve (Equation 4) to our deflection data from microcantilevers functionalized with a 1:1 equimolar mixture of ANT-152:C65690M mAbs tested against viral lysates from 60 diverse HIV-1 isolates suspended in PBS. We plotted the saturation deflection point for each tested HIV-1 subtype (maximum deflection) against the logarithm of the p24 concentration (Figure S8). From these curves, we derived the EC50 for our assay, which is the effective concentration of p24 at which the fitted curve reaches 50% of the maximum deflection response and gives a measure of optimal p24 concentration for detection by our platform. There is variation in EC50 values across HIV-1 subtypes, with subtype A displaying the lowest EC50 value, indicating high affinity for the p24 mAbs covalently attached to the microcantilevers (Table S10). Overall, these tests show that the p24 is detectable at low concentrations across HIV-1 subtypes.

$$y=A_{min}+(A_{max}-A_{min})/(1+{(x/x_0)}^{-h}{)}^s$$ Equation 4

3.7. Detection of p24 in human blood specimens

Antigen-based HIV-1 diagnostic platforms typically rely on analysis of blood samples, and some methods of blood specimen processing can interfere with antigen detection and impact viral load estimation [19]. To begin to determine which methods were compatible with our microcantilever platform, we spiked recombinant HIV-1 p24 protein (from strains HXB2 and IIIB) into whole blood isolated from an anonymous donor (Vitalant) at two concentrations (50 ng/mL or 1 ng/mL) prior to processing. The blood was subsequently processed in three ways: Method A) in the absence of anticoagulants, blood was mixed immediately with lysis buffer and clarified by centrifugation; Method B) in the absence of anticoagulants, blood was allowed to clot and was clarified by centrifugation before addition of the serum to lysis buffer; Method C) in the presence of anticoagulants, blood was mixed immediately with lysis buffer.

Using our multiplexed microcantilevers, similar deflection patterns were observed for blood specimens processed without the use of anticoagulants across both viral strains with deflection values between 40 and 50nm for the 50ng/mL specimens and between 10 and 15nm for the 1ng/mL specimens (Figure 5, Table S11). This suggests overall lower sensitivity in blood specimens as similar deflection ranges for the same microcantilevers were observed for lysed viral particles in PBS at 1pg/mL (Figure S6). The addition of sodium heparin as an anticoagulant led to even lower sensitivity with deflection values around 30nm for the 50ng/mL specimens and between 5 and 10nm for the 1ng/mL specimens (Figure 5). We also observed that the processed p24-containing blood specimens remained stable when stored overnight in the refrigerator, as confirmed by repeating the deflection experiment the next day to test reproducibility (Figure S9, Table S12). Taken together, these data show that while this microcantilever platform is compatible with HIV-1 p24 detection in minimally processed blood, the choice of sample processing method is critical for ensuring optimal sensitivity.

Figure 5 |. Dose-dependent detection of HIV-1 p24 antigen in minimally processed human blood.

Dose-dependent response curves for the detection of HIV-1 HXB2 or IIIB p24 spiked into human whole blood at a concentration of 50 ng/mL or 1 ng/mL. Blood specimens were processed one of three ways: A) lysis and then clarification without the use of anticoagulants, B) clarification and then lysis without the use of anticoagulants, and C) lysis and direct measurement in the presence of anticoagulants. The curves represent deflection of the microcantilever functionalized with a 1:1 equimolar mixture of ANT-152:C65690M mAbs (20μg/mL) as measured by f-AFM over the course of a 12 minutes. Graphs show the average of triplicate deflection measurements in nm +/− standard deviation per timepoint and condition.

3.8. Multiplexed cantilever arrays and rapid detection of p24 using biofunctionalized MOSFET device.

While AFM is a powerful technology for measuring microcantilever deflection, it is not suitable for the development of POC tests. Ideally, the readout would be electronic to better enable interpretation and reporting in a POC setting. We had previously reported the use of MOSFET (Metal-Oxide-Semiconductor Field-Effect Transistor) technology for the detection and reporting of microcantilever deflection. MOSFET leverages differential readout mechanisms, where one cantilever is coated with a sensing layer (gold-coated) and the other serves as a reference (silicon nitride). The shift in drain current, caused by changes in threshold voltage due to specific biomolecular interactions (compressive surface stress), enables detection. This configuration minimizes the impact of environmental factors such as humidity and temperature by canceling out non-specific signals. Additionally, MOSFET enables a direct electrical readout that is compatible with use in a handheld device for use in point-of-care test.

To see if this would be compatible with our p24 detection assay, MOSFET devices were biofunctionalized with the anti-p24 monoclonal antibody 183-H12–5C at 20 ug/mL as before. A clear change in drain signal current was observed by 4 minutes after introduction of recombinant HIV-1 IIIB p24 in PBS and signal plateaued by 10 minutes, similar to the kinetics observed by AFM (Figure 6a). We next tested the sensitivity of the platform using dilutions of recombinant HIV-1 IIIB p24 in serum from a seronegative human donor. Plotting the drain current by the drain voltage, we can detect clear differences in signal at different p24 concentrations (Figure 6b). The device retained exceptional sensitivity with statistically significant differences in signal at concentrations as low as 0.1 picogram/mL (100 femtogram/mL) as compared to the serum only control. Taken together, the combination of high sensitivity, rapid response, multiplexing capability, and electronic output makes this approach a promising avenue for the development of future point-of-care diagnostics.

4. SUMMARY AND CONCLUSION

We developed a microcantilever-based platform capable of detecting HIV-1 p24 antigen from diverse HIV subtypes with high sensitivity and speed. Using broadly cross-reactive anti-HIV-1 p24 monoclonal antibodies (mAbs), the platform achieves detection limits as low as 1 fg/mL in human serum without signal amplification and delivers results in just 15 minutes. The technology yields a quantitative signal over several orders of magnitude, raising the potential for downstream use in viral load determination. We furthermore demonstrated that our microcantilever-based platform can be leveraged for use in nano-electromechanical system (NEMS) cantilevers with MOSFET readouts, enabling direct electronic signal transduction for improved ease-of-use. These advancements will enhance the platform’s scalability and adaptability, making it an ideal candidate for next-generation, point-of-care HIV diagnostics.

This innovative technology addresses key limitations of current HIV-1 diagnostics, but there are several barriers yet to be overcome. First, while highly sensitive in serum, sensitivity in whole blood specimens was reduced, especially in the presence of anticoagulants. While this is not disqualifying as many clinical settings and tests have been designed for use with whole blood without anticoagulants, it would narrow the window of possible applications without further optimization. Second, additional efforts to expand the linear detection range and account for additional subtype variations will be needed to ensure robustness across diverse HIV-1 strains. We optimized our microcantilever system using two broadly cross-reactive HIV-1 p24 antibodies, C65690M and ANT-152, which together successfully detected all of the 60 tested international HIV-1 isolates which span subtypes A, B, C, D, CRF01-AE, and CRF02-AG. As of 2021, these subtypes collectively accounted for 72.6% of circulating isolates with an estimated prevalence of each as follows: C (23%), A (16.7%), CRF01-AE (9.5%), B (8.5%), D (7.7%), and CRF02-AG (7.2%)[23]. We cannot rule out the need for additional antibodies to cover additional untested subtypes, but this also highlights multiplexed functionality of the platform, which will allow it to be readily modified as the epidemic continues to evolve. Finally, we have yet to test this platform with clinical specimens, so the best methods for disrupting pre-existing antibody:antigen complexes while retaining platform sensitivity remain to be determined.

In sum, we believe that the sensitivity, adaptability, and scalability of microcantilever technology make it a promising candidate for the development of next-generation diagnostics, potentially even in POC settings. We ultimately hope this technology may enable the simultaneous detection of multiple pathogens (such as HIV, HBV, HCV, tuberculosis, and others) on a single chip in future iterations of this platform, providing comprehensive diagnostics in one test.