A point-of-care test to measure tacrolimus directly from unprocessed whole blood

Abstract

Tacrolimus is a cornerstone immunosuppressant used to prevent graft rejection in solid organ transplant recipients. While effective, tacrolimus therapy must be carefully monitored and managed due to the drug’s narrow therapeutic window and variable patient responses to dosing. Overdosing of tacrolimus can lead to opportunistic infections, nephrotoxicity, neurotoxicity, and cancer while under-dosing risks organ rejection. As a result, therapeutic drug monitoring (TDM) of tacrolimus is routinely performed. Current monitoring approaches face significant limitations including slow (4–8 hours or days) turnaround times, complex sample preparation requirements, and reliance on centralized infrastructure that delays critical dosing decisions. Furthermore, emerging evidence shows that the pharmacokinetic exposure parameter best associated with clinical outcomes is the area under the concentration-time curve (AUC). AUC-based monitoring requires sampling patients at multiple time points, which increases the logistical burdens to both patients and providers. Hence, there is an urgent need for a point-of-care test for tacrolimus that can be run at the patient bedside, in outpatient clinics, and eventually in patient homes. Here, we describe a novel point-of-care test, the TAC-D4 POCT, capable of measuring tacrolimus levels directly from whole blood samples without sample preparation steps or expensive analyzers with automated results available in 1-hour. Our test helps overcome the current logistical barriers in the clinical management of patients receiving tacrolimus therapy ultimately leading to improved patient adherence, outcomes, and the ability to further personalize dosing regimens. Moreover, the Tac-D4 is highly modular, and can be used to monitor clinically significantly biomarkers, such as interleukin-6 and matrix metalloproteinase 3, that complement tacrolimus TDM.

INTRODUCTION

Organ transplantation is the treatment of choice for patients with end organ failure offering curative therapy globally to over 100,000 people per year [1]. Its successful outcome relies on a lifetime requirement for immunosuppressive drug therapy to prevent rejection [2]. Among these, tacrolimus is one of the most effective agents and is prescribed in nearly 90% of renal transplant cases [3, 4]. Despite its effectiveness, tacrolimus dosing is complex due to significant inter- and intra-patient variability in both pharmacokinetic and pharmacodynamics. [2, 5] The risks of improper dosing are severe: underexposure can lead to rejection, while overexposure can cause life-threatening toxicities such as nephrotoxicity, neurotoxicity, and increased infection risk [6–9]. To manage this delicate balance, tacrolimus dosing relies on frequent, accurate measurement of blood concentrations, with target trough levels typically ranging from 5 – 20 ng/mL depending on the organ type and patient characteristics and the time since transplant.

Precise tacrolimus dosing is managed through TDM, where trough blood concentrations guide subsequent dose adjustments [2, 9]. Currently, this requires the patient to travel to a facility for venous blood collection followed by transport to a central laboratory. Samples then undergo complex, multi-step extraction procedures before being analyzed using liquid chromatography-tandem mass spectrometry (LC-MS/MS) or immunoassays, with turnaround times typically ranging from 4–8 hours or longer [8, 9]. While TDM is essential for effective tacrolimus therapy, the centralized testing model places a significant burden on both patients and providers [2, 8], samples being taken at the incorrect time leading to delayed dose adjustments, higher healthcare costs, limited access in resource-constrained settings, and restricted monitoring frequency [6, 9, 10]. This constraint is particularly important, given emerging evidence that AUC-based dosing leads to better patient outcomes compared to trough-based monitoring, as AUC drug exposure correlates more strongly with clinical outcomes. Unfortunately, AUC-based dosing is largely impractical with current testing paradigms [11].

The limitations of current TDM highlight an urgent need for new testing methods that can deliver rapid and timely results both in hospital settings and remotely. Point-of-care tests (POCTs) have proven highly effective in other areas, such as glucose monitoring, but have not yet been widely adopted for TDM.

For over a decade, our group has been developing an innovative immunoassay platform, called the D4 for testing of proteins analytes. We reasoned that a redesign of the D4 assay for small molecules could address many of the fundamental limitations of current tests for tacrolimus. Motivated by this rationale, we describe herein the development and validation of a D4 assay for tacrolimus (the “TAC-D4”) directly from whole blood. The TAC-D4 uses a competitive immunoassay format that we optimize for whole blood analysis without sample preparation. The assay delivers clinical-grade results within one hour and, when integrated with our microfluidic platform and a low-cost and portable detector we have developed for the D4 assay—named the D4Scope—that has a form factor that allows it to be used anywhere to read out the test and provide quantitative results. Moreover, the D4 platform is easily adaptable to detect other protein biomarkers, such as interleukin 6 (IL-6) and matrix metalloproteinase 3 (MMP-3), offering the potential for deeper insight into graft rejection risk.

MATERIALS AND METHODS

Study Design

The TAC-D4 employs a competitive immunoassay format in which fluorescently labeled anti-tacrolimus detection antibodies (dAbs) compete for binding between free tacrolimus in the blood sample and surface-immobilized tacrolimus-BSA conjugate (TAC-BSA). The TAC-BSA capture spots are printed on a non-fouling POEGMA-co-GMA polymer brush coating that minimizes non-specific binding, while trehalose serves as a stabilizing excipient for the dried dAbs on the chip surface.

In the absence of tacrolimus, dAbs binds to the immobilized TAC-BSA, producing a high fluorescence signal. Because the running buffer used in the TAC-D4 is non-lytic, the assay senses only the extracellular (free plus plasma protein-bound) tacrolimus in the whole blood sample during the 60-minute incubation; intra-erythrocyte tacrolimus is not interrogated. As tacrolimus concentration in the sample increases, more dAbs are sequestered in solution, resulting in a decrease in surface-bound fluorescence. Thus, the assay produces an inverse fluorescence intensity response that is proportional to the tacrolimus concentration in the sample.

The objective of this investigation was to develop a novel POCT for tacrolimus that addresses the critical need for rapid and frequent TDM in transplant recipients. The development pathway had three phases: (1) demonstrating tacrolimus detection that was sensitive enough to quantify tacrolimus concentrations below and across its therapeutic range, (2) integrating the tacrolimus assay for bedside testing applications using a POCT, and (3) establishing a proof-of-concept for the multiplexed detection of complimentary biomarkers that can be used with tacrolimus TDM to better manage complicated VCA cases.

All commercial materials (e.g., immunosuppressive drugs, protein conjugates, monoclonal antibodies) and human whole blood specimens were procured from established vendors. Preliminary assay development and parameter optimization employed tacrolimus-spiked whole blood from diverse donors. Clinical performance evaluation was conducted using anonymized residual specimens from tacrolimus-monitored patients acquired through the Cooperative Human Tissue Network (CHTN) at Duke University Medical Center. Research personnel were initially blinded to sample concentrations throughout the study except during specified pilot experiments using tacrolimus-spiked samples.

TAC-D4 Assay Development

The TAC-D4 assay is based on established protocols for conventional D4 assays [12–19] with two key modifications: the implementation of POEGMA-co-GMA brushes instead of pure POEGMA brushes, and the substitution of capture antibodies with TAC-BSA (Creative Diagnostics Catalog Number: DAGA-190B). The formation of POEGMA-co-GMA brushes containing 2% GMA was achieved through the following procedure: silicon oxide glass microscope slides underwent initial amine group functionalization via overnight immersion in (3-aminopropyl)triethoxysilane (IUPAC: 1-propanamine,3-(triethoxysilyl), Sigma-Aldrich Catalog Number: 440140, CAS Number: 919-30-2), followed by treatment with triethylamine (IUPAC: N,N-diethylethanamine, Sigma-Aldrich, Catalog Number: 471283, CAS Number: 121-44-8) and bromoisobutyryl bromide (IUPAC: 2-bromo-2-methylpropanoyl bromide, Sigma-Aldrich, Catalog Number: 252271, CAS Number: 20769-85-1). The slides were then placed in a polymerization mixture containing 330 mL deionized water, 500 mL methanol, 150 mL poly(ethylene glycol) methyl ether methacrylate (PEGMEM) monomer (IUPAC: 2-methoxyethyl methacrylate, Mn 300 average ethylene glycol repeat length 5, Sigma-Aldrich Catalog Number: 447935, CAS Number: 26915-72-0), 20 mL glycidyl methacrylate (GMA) monomer (IUPAC: (oxiran-2-yl)methyl 2-methylprop-2-enoate, Sigma-Aldrich Catalog Number: 779342, CAS Number: 106-91-2), 130 μL hexamethyltriethylenetetramine (HMTETA) (IUPAC: N’-[2-[2-(dimethylamino)ethyl-methylamino]ethyl]-N,N,N’-trimethylethane-1,2-diamine, Sigma-Aldrich Catalog Number: 366404, CAS Number: 3083-10-1), and 77 mg Cu(II)Br (ThermoFisher Scientific, Catalog Number: 206331000, CAS Number: 7789-45-9). Surface-initiated atom transfer radical polymerization (SI-ATRP) was carried out within a mBRAUN MB10 LABstar glove box (Serial Number 13–210) for 2 hours. Verification of successful POEGMA-co-GMA brush polymerization was accomplished through polymer brush thickness measurements in air using a J.A. Woollam alpha 2.0 spectroscopic ellipsometer and X-ray photoelectron spectroscopy (XPS) analysis on a ThermoFisher Scientific Nexsa G2 surface analyzer (Catalog Number: 9961253).

After polymerization, fabrication of the TAC-D4 assay utilized non-contact inkjet printing techniques as previously described [12–19]. TAC-BSA printing was performed using a non-contact inkjet microarray printer (Scienion, sciFLEXARRAYER Type S12 from a 1 mg/mL stock concentration. For the open format TAC-D4 assay, anti-tacrolimus antibodies (Creative Diagnostics Catalog Number: DPABH-29477) were labeled with a Alexa Fluor 647 antibody labeling kit (ThermoFisher Scientific Catalog Number: A20173) according to the manufacturer’s instructions, and were then printed at a concentration of 0.02 mg/mL using a BioDot AD1520 Aspirate Dispense System. For dAb-spike experiments dAbs for the TAC-D4 POCT were introduced at 1.5 μg/mL concentration. For all assays using tacrolimus-spiked samples, Molecular Biology Grade tacrolimus (R&D Systems^®, Catalog Number: 3631) was serially diluted into pooled fetal bovine serum (FBS) (Sigma-Adrich, Catalog Number: F4135) or human whole blood obtained from Innovative Research (catalog # IWB1K2E10ML).

Punch Chip Fabrication, Assembly, and Function

Fabrication of the microfluidic punch cassette employed a multilayer methodology combining CLAREX acrylic sheets of different thicknesses (0.2, 0.5, 1, and 2 mm; Astra Products), pressure sensitive adhesive tapes (7955MP and 468MP; 3M Company), and cotton absorbent liner (Whatman CF7; Cytiva Life Sciences). Component geometries were created in AutoCAD 2024 (Autodesk, Inc.) and precision-cut using a Gravograph LS900 CO₂ laser system with Gravostyle 8.0 software. 3D printed components were designed in Fusion 360. Fluid reservoirs and button covers were produced using a Form 4 stereolithography printer (Formlabs, Inc.) utilizing Clear V5 and Elastic 50A V2 resins. The seal-breaking pin was manufactured using a Bambu X1C with standard PLA filament.

Reagent deposition followed the same protocol described in fabrication of the open format TAC-D4 assay except that dAbs for the TAC-D4 POCT were mixed with the sample, rather than being printed on the chip. The TAC-D4 POCT workflow comprises five simple steps. Initially, the user combines 90 μL of whole blood, collected in EDTA-anticoagulated tubes, with 10 μL of 10% Poloxamer-188 running buffer containing anti-tacrolimus antibodies at 1 μg/mL, unless specifically noted otherwise. EDTA was selected as the anticoagulant because it is the standard collection method for tacrolimus TDM and does not interfere with the immunoassay chemistry. While the TAC-D4 has been primarily validated using EDTA-anticoagulated blood, the effect of alternative anticoagulants (e.g., heparin, citrate) on assay performance has not yet been systematically evaluated and will be addressed in future studies. Subsequently, after gentle mixing, 80 μL of this mixture is introduced into the sample inlet of the microfluidic chip. Next, the user adds 600 μL of wash buffer to the designated wash reservoir and allows the test to incubate for 60 minutes.

All user-dispensed volumes in the current study were delivered using calibrated, volume-controlled air-displacement pipettes; for non-laboratory deployment, the same volumes are intended to be delivered by fixed-volume disposable pipettes or pre-metered single-use reagent ampoules so that operator-induced volumetric variability is removed from the workflow without modifying assay chemistry. Reagent stabilization on the printed POEGMA-co-GMA brush is provided by trehalose pads, and environmental resilience of the microfluidic D4 format under temperature- and humidity-stressed conditions has been previously demonstrated for the same platform [19].

Following incubation, pressing the chip’s button triggers an automated wash and dry sequence. Finally, the chip is placed into the D4Scope for automated image capture and analysis. It is worth noting that Poloxamer-188 has a known lower critical solution temperature (LCST) of approximately 50°C in aqueous media. While this is well above typical ambient and clinical operating temperatures (20–37°C), users should ensure that the test is performed and reagents are stored below this threshold to avoid phase separation that could compromise assay performance.

TAC-D4 Assay Performance Assessment

Initial assay performance evaluation involved LOD calculation using established methodologies where LOD = limit of blank (LOB) + 1.645(_{SDlow concentration sample}) and LOB = mean_blank + 1.645(SD_blank) as described in Armbruster and Pry [20]. Assay performance underwent further evaluation through systematic precision characterization across tacrolimus’s clinical range (1–45 ng/mL for our test). Precision measurement utilized relative error (RE) calculated using the following formula:

Equation 1:

Relative error calculation used for all non-anchor points

$$\mathit{\text{RE}}=\mathit{\text{absolute}}\mathit{\text{value}}(\left(\frac{\mathit{\text{Measured}}\mathit{\text{TAC}}-\mathit{\text{True}}\mathit{\text{TAC}}\mathit{\text{value}}}{\mathit{\text{True}}\mathit{\text{TAC}}\mathit{\text{value}}}\right)\times 100\%)$$

where “True TAC” represents the known concentration of tacrolimus-spiked samples prepared by precise weight-to-volume dilution of tacrolimus powder in ethanol (as recommended by the supplier) and then diluted into whole blood, and “Measured TAC” represents the concentration determined from the calibration curve.

Quality control and assay standardization were accomplished through a comprehensive batch-matching strategy designed to compensate for inter-run signal variability. Each test batch was characterized by unique combinations of POEGMA-co-GMA polymerization, device fabrication runs, and final assembly procedures. Calibration curves for each batch were developed using tacrolimus-spiked whole blood, with duplicate measurements spanning a concentration range of 0–50 ng/mL in two-fold serial dilutions. Tacrolimus measurements were determined using a 5-parameter asymmetric sigmoidal curve fit (GraphPad Prism, version 10.2.3) referencing batch-matched calibration curves. This calibration approach facilitated accurate concentration determination while accounting for batch-specific response characteristics.

Assessment of TAC-D4 Tacrolimus Concentrations to DCAL Tacrolimus Concentrations Using Scavenged Samples from CHTN

Determination of free tacrolimus and total concentrations was conducted through collaboration with the Southern Division of the Cooperative Human Tissue Network at Duke University (CHTN). CHTN, founded in 1987 by the National Cancer Institute’s Cancer Diagnosis Program, provides researcher access to human biological specimens through a network of six regional centers [21]. This infrastructure allowed rapid and cost-effective acquisition of well-characterized tacrolimus-positive blood samples while maintaining compliance with Duke University Health Systems Institutional Review Board approved requirements (Pro00109699, PI Ashutosh Chilkoti).

Through CHTN, we obtained access to residual whole blood samples from patients undergoing routine tacrolimus monitoring at DCAL. This arrangement facilitated parallel testing of samples processed through both standard laboratory protocols and the TAC-D4 assay. Specifically, a portion of each blood sample underwent processing into serum for standard DCAL total tacrolimus testing, while the remaining unprocessed whole blood was collected in EDTA-coated tubes for the TAC-D4 assay. CHTN staff-maintained chain of custody and provided essential demographic data and corresponding DCAL test results (including both total tacrolimus and hematocrit) while ensuring patient privacy. This approach permitted direct comparison between the TAC-D4 POCT from blood and the laboratory assay from serum.

DCAL’s process for testing tacrolimus concentrations are as follows: first, water is added to patient blood collected in ethylenediaminetetraacetic acid (EDTA) coated tubes. Next, an internal standard solution and denaturing reagent are added to the blood-water mixture. Then, samples are vortexed, and centrifuged. The supernatant is collected in vials and placed in trays on the instrument for analysis. Reverse-phase high performance liquid chromatography on a C18 column is then used to elute tacrolimus so that its internal standard has predictable retention times. The mobile phases include ionization that aids to apply charges to tacrolimus and its internal standards at the ion source. Analysis is performed using a Acquity-LC Xevo TQD systems liquid chromatograph-tandem mass spectrometer with triple quadrupoles via multiple reaction monitoring (Waters Corporation). Finally, results are quantified by using a 6-point calibration curve.

VCA-D4 Assay Fabrication and ELISA comparison

VCA-D4 fabrication followed the same steps as TAC-D4 fabrication with the additional inkjet printing of IL-6 and MMP3 reagents. Specifically, a dAb cocktail of Alexa Fluor 647-labeled anti-tacrolimus (Creative Diagnostics Catalog Number: DPABH-29477), anti-IL-6 (R&D Systems Catalog Number: AF-206), and anti-MMP3 (R&D Systems Catalog Number: AF513) were printed at 0.02, 0.05, and 0.05 mg/mL respectively with the same BioDot AD1520 Aspirate Dispense System on trehalose pads on slides coated with POEGMA-co-GMA brushes. Following dAb printing, TAC-BSA (Creative Diagnostics Catalog Number: DAGA-190B), anti-IL-6 cAbs (R&D Systems Catalog Number: MAB206), and anti-MMP3 cAbs (R&D Systems Catalog Number: AF513) were printed using the Scienion non-contact inkjet microarray printer (Scienion, sciFLEXARRAYER Type S12). Purified recombinant human IL-6 (R&D Systems Catalog Number: 206-IL) and MMP3 (R&D Systems Catalog Number: 513-MP) were spiked into fetal bovine serum (Sigma Aldrich Catalog Number: F4135). For ELISA comparison studies, R&D Systems Quantikine ELISA kits were used for both IL-6 (R&D Systems Catalog Number: D6050) and MMP3 (R&D Systems Catalog Number: DMP300).

STATISTICAL ANALYSIS

All statistical analyses were conducted using GraphPad Prism software version 10.4.1 (GraphPad Software Inc.). Dose-response curves were generated using a 5-parameter asymmetric sigmoidal curve fit. Relative error (RE) was calculated for each non-anchor concentration point using the formula described in the TAC-D4 Assay Performance Assessment section. All spiked-sample experiments were performed with a minimum of n=3 replicates per concentration, and data are presented as mean ± standard deviation (SD), unless otherwise noted. Correlation between TAC-D4 and DCAL measurements was assessed by linear regression analysis, with the coefficient of determination (R²) used to evaluate goodness of fit. For ELISA comparison studies, Pearson correlation coefficients were computed between VCA-D4 fluorescence intensity and ELISA absorbance values. No formal hypothesis testing (e.g., t-tests or ANOVA) was performed in this proof-of-concept study, as the primary objective was to characterize assay analytical performance rather than to compare group means.

RESULTS

Development of the TAC-D4 Assay

To detect tacrolimus, we built a competition assay, termed the TAC-D4, that utilizes the D4 platform [12–19]. The TAC-D4 uses commercial anti-tacrolimus antibodies and tacrolimus conjugated bovine serum albumin (TAC-BSA) reagents that are inkjet printed onto glass slides coated with poly(oligo(ethylene glycol) methyl ether methacrylate-co-glycidyl methacrylate) (POEGMA-co-GMA) to detect tacrolimus in blood and/or serum samples. We selected these reagents and the assay format shown in Figure 1 based on results from pilot studies (Supplementary Figures 1–2) that compared assay formats (e.g., sandwich immunoassays and competition assays) and assay optimization parameters (e.g., assay reagent amounts and running buffer compositions).

Figure 1.

The TAC-D4. A) The four sequential steps of the TAC-D4. I) A blood sample is dispensed. II) Trehalose pad dissolves, liberating dAbs printed on the trehalose pad. III) dAbs diffuse throughout the sample volume. IV) Fluorescence signal is detected. B) Optimized TAC-D4 dose response curve using tacrolimus-spiked whole blood with relative error in Table 1 below. Error bars for n=3 replicates at each concentration point are included but are too small to be visible.

Figure 1A shows a schema of the TAC-D4. First, a blood sample is dispensed on the chip (Fig. 1A–I) that dissolves the underlying trehalose pads (Fig. 1A–II) and liberates printed fluorescently labeled anti-tacrolimus detection antibodies (dAbs), that diffuse (Fig. 1A–III) throughout the sample volume. dAbs bind to either tacrolimus in solution or immobilized TAC-BSA. Labeled antibodies that remain bound to immobilized TAC-BSA spots generate a fluorescence signal that can be detected (Fig. 1A–IV), by the D4Scope [12–14, 22].

The TAC-D4 can measure tacrolimus with a limit-of-detection (LOD) lower than its clinical range of 5–20 ng/mL for most patients [23]. The TAC-D4 can also quantify tacrolimus concentration across its narrow therapeutic window with a LOD of less than 1 ng/mL and relative error (RE) (as measured by the equation described in the Materials and Methods section of this report) of less than 21%, as shown in Figure 1B above and described in detail in Table 1 below.

Table 1.

TAC-D4 relative error (RE) as a function of concentration. Tacrolimus concentrations in the left column reflect the nominal values of the two-fold serial dilution series prepared by precise weight-to-volume dilution. The number of significant figures shown (up to 4–5 digits) reflects the mathematical precision of the serial dilution rather than the analytical precision of the assay. Interpolated concentrations (columns 2–4) are reported to two decimal places as output by the curve-fitting algorithm.

TACROLIMUS CONC. (NG/ML)	INTERPOLATED TACROLIMUS CONC. (NG/ML)			RE %			AVERAGE RE %
50	**	**	**	**	**	**	**
25	28.83	26.24	24.92	15%	5%	0%	7%
12.5	10.73	11.75	11.96	14%	6%	4%	8%
6.25	6.57	7.52	6.60	5%	20%	6%	10%
3.125	2.38	3.05	2.95	24%	2%	6%	11%
1.5625	1.41	2.18	1.78	10%	40%	14%	21%
0.78125	**	**	**	**	**	**	**
0	***	***	***	***	***	***	***

^**Anchor point – RE not calculated

^***Blank point – RE not calculated, SD for n=3 blank points is 9290.4 (a.u.). Signal intensity in a.u. for all points can be found in the Supplemental Data file.

Performance Assessment of the TAC-D4 Assay Using Clinical Samples

Next, we sought to evaluate how tacrolimus concentrations measured by the TAC-D4 compare with the clinical gold standard test performed at Duke University Hospital. To accomplish this, we obtained blood samples from the Cooperative Human Tissue Network (CHTN), whose tacrolimus levels had already been measured by the Duke Clinical Automated Lab (DCAL) using LC-MS/MS.

Figure 2A shows the fluorescence intensity plotted for eight blood samples with their tacrolimus concentration, as measured by DCAL. The TAC-D4 shows the expected inverse fluorescence response with increasing tacrolimus concentrations and a non-linear curve fit. Figure 2B shows the correlation (R² = 0.70) between TAC-D4 (y-axis) and DCAL (x-axis) concentrations using the same eight samples shown in Figure 2A.

Figure 2.

TAC-D4 clinical test. A) Fluorescence intensity (y-axis) on the TAC-D4 for DCAL measured tacrolimus concentrations (x-axis). B) Correlation plot for TAC-D4 measured tacrolimus concentrations (y-axis) and DCAL measured concentrations (x-axis).

Adaptation of the TAC-D4 to a Point-of-Care Test

Thus far in this report, we have described a TAC-D4 assay capable of quantifying tacrolimus concentrations directly from whole blood samples in an “open format” (shown in Supplementary Figure 3) that is optimized for high throughput testing. This design is well-suited for experimental studies but is not amenable to point-of-care settings. To adapt the TAC-D4 assay into a POCT, we integrated it with a microfluidic cartridge (Figure 3A–B) and the D4Scope detector which have been developed by our group (Figure 3C) [13, 15, 19].

Figure 3.

The TAC-D4 POCT. A) Image of the TAC-D4 “punch” chip before sample loading. B) Image of tacrolimus-spiked whole blood with Poloxamer-188 dispensed into the punch chip. C) Image of the TAC-D4 POCT in the imaging chamber of the D4Scope. D) Dose-response curve using the Poloxamer-188 wetting agent on the TAC-D4 POCT.

This cartridge —dubbed the “punch chip— is a single use microfluidic device that automates the D4 assay after adding sample and wash buffer to the device inlets at the start of testing, which significantly reduces the required user intervention to run the test when compared to the “open format”. This simplification is made possible by enclosing the inkjet printed POEGMA substrate in a reaction chamber, and accessorizing with other fluid handling elements (e.g., inlets, flow resistor, waste reservoir). During incubation, this is a completely closed system, where the sample and reagents will react for any user-specified period regardless of sample type. A user-accessible button is situated on top of a depressible pin, that when depressed breaks the barrier that separates the reaction chamber from the downstream waste collection chamber, to facilitate the washing and drying of the POEGMA substrate within ~2 minutes.

To facilitate the operation of the punch chip, a small amount of minimally dilutive (1:10 v:v) surfactant is added to the sample prior to loading onto the cassette. This helps reduce the surface tension and improve wettability of the sample to enable ventless sample loading. We tested several non-ionic surfactants and found that Poloxamer-188 [24] enables the flow of blood into the punch chip without compromising the performance of the assay (Supplementary Figure 6).

Figure 3D shows a dose-response curve for the TAC-D4 POCT using tacrolimus-spiked whole blood samples with Poloxamer-188, demonstrating that the point-of-care version of the TAC-D4 on the punch chip has a comparable performance to the data in Figure 1B acquired on the open-format chip.

Multiplexed Testing of Tacrolimus, IL-6, and MMP3 for Improved Vascular Composite Allotransplantation Management

To highlight the versatility of the TAC-D4, we also developed a multiplexed version of the test to meet future clinical objectives for the management of vascular composite allotransplantation (VCA) patients. VCA is a form of advanced organ transplant surgery that involves the transfer of a vascularized body part containing multiple tissue types (e.g., skin, muscle, bone, nerves, and blood vessels) as a functional unit (e.g., hand) [25–29]. Tacrolimus is the most common medication used in VCA, which, in addition to the routine TDM, requires careful monitoring of the allograft [30, 31].

Current methods for preventing rejection require in-patient admission and invasive biopsies [26, 28–30]. Our objective was to build a non-invasive bioassay that could be used to monitor patients following VCA procedures to predict rejection before and as it occurs. Studies suggest that by closely monitoring the level of circulating biomarkers such as IL-6 and MMP3, graft rejection can be anticipated and potentially avoided [32, 33]. IL-6 is a key pro-inflammatory cytokine that is rapidly upregulated during acute rejection episodes, often preceding clinical manifestations by several days [33]. In VCA, elevated IL-6 levels have been associated with both acute cellular rejection and chronic allograft vasculopathy, making it an ideal early warning biomarker for rejection surveillance [34]. MMP3 plays a critical role in extracellular matrix remodeling during tissue inflammation and has emerged as a sensitive marker of vascular inflammation in solid organ transplantation [32]. For VCA, where vascular integrity is paramount for graft survival and function, early reports indicate that MMP3 elevation signals early vascular compromise that could progress to chronic rejection if left unaddressed [32]. While the full clinical utility of these biomarkers for VCA is still under active investigation, we aimed to create a test––called the VCA-D4––that could be used to simultaneously quantify these biomarkers while providing necessarily tacrolimus TDM.

Figure 4A shows a schematic of the VCA-D4 where spatially distinct zones, all within a single reaction chamber, are utilized to accommodate the different assay formats and capture specific fluorescence intensities that correspond to each biomarker. This multiplex design uses the same non-contact inkjet printing fabrication approach as the TAC-D4 [15, 16]. Unique to the VCA-D4 is its integration of small-molecule detection via the TAC-D4 competitive format with protein detection using the standard antibody sandwich approach in the same chip. Each zone is visually mapped by the software onboard the D4Scope, enabling quantification of all biomarkers without the need for multiple fluorophores.

Figure 4.

The VCA-D4. A) Schematic representation of the VCA-D4 with its three-zone architecture. Zone 1 contains the TAC-D4 competitive assay format while Zones 2 and 3 have a traditional sandwich immunoassay format for MMP3 and IL-6 respectfully. B) VCA-D4 dose response curves for tacrolimus (gray), MMP3 (yellow) and IL-6 (blue) using fetal bovine serum.

Figure 4B shows the dose-response behavior of the VCA-D4 in the presence of tacrolimus, IL-6, and MMP3-spiked into undiluted fetal bovine calf serum. Fetal bovine calf serum was used to define the initial performance characteristics of the VCA-D4, as it mimics human blood while lacking endogenous IL-6 and MMP3 [32]. Additionally, we performed a simple cross-reactivity study where each biomarker was spiked into fetal bovine calf serum independently (i.e., only tacrolimus, only IL-6, and only MMP3) to characterize the dose response behavior of the VCA-D4 in the presence of one or multiple biomarkers. These results can be found in Supplementary Figure 7.

To further assess the performance of the multiplexed D4 assay, we compared results to a commercially available enzyme linked immunosorbent assay (ELISA) for each biomarker. Figure 5 shows preliminary comparison between the multiplexed VCA-D4 and the IL-6 ELISA (Figure 5A–B) and MMP3 ELISA (Figure 5C–D) using both the protein standards provided in the ELISA kits (Figure 5A and 5C) and the protein standard selected for initial tests on the VCA-D4 (Figure 5B and D). Notably, the VCA-D4 correlates very strongly (R² > 0.9) with the R&D IL-6 ELISA and strongly (R² = 0.8 for points above 1 ng/mL) with the R&D MMP3 ELISA. Below 1 ng/mL the correlation between the VCA-D4 and R&D MMP3 ELISA deteriorates because the VCA-D4 has been tuned to quantify MMP3 levels >1 ng/mL based on our review of the reported concentrations of MMP3 during periods of graft rejection across the literature. If measuring MMP3 concentrations lower than 1 ng/mL is required, the sensitivity of the VCA-D4 can increased by further optimization of the assay.

Figure 5.

Comparison of the performance of the VCA-D4 to commercial ELISAs. A) Correlation plot between the fluorescence response of the VCA-D4 and the measured absorbance from the R&D Systems^® IL-6 ELISA using the ELISA IL-6 standard (R²=0.95). B) Correlation plot between the fluorescence response of the VCA-D4 and the measured absorbance from the R&D Systems^® IL-6 ELISA (R²=0.97) using the selected IL-6 protein used for VCA-D4 development. C) Correlation plot between the fluorescence response of the VCA-D4 and the measured absorbance from the R&D Systems^® MMP3 ELISA using the ELISA MMP3 standard (R²=0.71 when using all points and 0.82 when using points above 0.1 ng/mL–the LOD of the VCA-D4). D) Correlation plot between the fluorescence response of the VCA-D4 and the measured absorbance from the R&D Systems^® MMP3 ELISA (R²=0.7 for all points, 0.8 for points above 1 ng/mL) using the selected MMP3 protein used for VCA-D4 development.

DISCUSSION

Tacrolimus is the most widely used immunosuppressant for solid organ transplant [6, 9]. Regular TDM is critical to ensure patient safety and efficacious treatment [6, 8, 9]. Despite best efforts, and decades of experience with tacrolimus TDM, the relationship between whole blood trough and peak concentrations and clinical outcomes is not always clear and generalizable [35]. Rejection episodes and toxicities have been documented in patients whose tacrolimus levels were within the therapeutic range by conventional measurement [36]. Conversely, some patients maintain good graft function despite lower-than-target concentrations [3]. As a result, novel, timely, and potentially personalized testing methods are required that enable AUC-guided dosing [6, 11, 37, 38] and other emerging methods such as the exploration of free versus total tacrolimus in blood [36, 39–41].

POCTs such as the TAC-D4 described herein represent a potential breakthrough for post-transplant care management. For example, an ongoing clinical trial (NCT04657562 [40]) is specifically examining correlations between free tacrolimus concentrations and transplant outcomes (e.g., acute rejection episodes and toxicity events). In addition to tacrolimus, the simultaneous monitoring of therapeutic drug levels and rejection biomarkers, as in the case of VCA, provides clinicians with actionable information that can be used for immediate prevention of allograft dysfunction and decline and potentially lifesaving therapeutic adjustments [32].

While the results of this study are promising, this study has some limitations. First, the TAC-D4 measure free and plasma protein-bound tacrolimus levels in blood, whereas the LC-MS/MS method used by DCAL measures the total —free, plasma protein-bound and intracellularly bound— tacrolimus concentration. Clinical relevance of free versus total TAC levels is still being defined; because tacrolimus binds extensively to erythrocytes and plasma proteins, whole-blood (“total”) levels are strongly influenced by hematocrit and red blood cell binding, whereas the unbound (“free”) fraction—the pharmacologically active moiety—may better reflect exposure and toxicity risk [42].

As a result, monitoring the free (unbound) tacrolimus fraction could offer several advantages. First, it aligns more closely with the pharmacologically active exposure and is less confounded by hematocrit-driven erythrocyte binding that distorts whole-blood total levels [41]. Second, because tacrolimus extensively binds erythrocytes and other plasma proteins (e.g., albumin), shifts in these binding compartments (e.g., anemia, hypoalbuminemia, inflammation) can change total levels without proportionate changes in effect—whereas free levels better track toxicologically relevant exposure and may correlate more directly with adverse events [35]. Third, recent matrix-comparison work underscores clinically meaningful discrepancies between whole blood, total plasma, and free plasma, supporting the rationale to follow the unbound fraction in susceptible patients [43].

Second, the slope in Figure 2B is 0.7316, which deviates from the expected slope of 1.0 for perfect agreement between the two assays. Despite the discordance in the slope, with a large enough number of clinical samples analyzed by both approaches, a calibration can be devised that provides a correction factor for the difference in the tacrolimus concentration reported by LC/MS-MS and the TAC-D4 with high accuracy and precision, however, sample number 115 with a DCAL measured tacrolimus concentration of 14.5 ng/mL, but an average TAC-D4 concentration of 3.86 ng/mL was identified as a potential outlier that when removed, increases the correlation between our test and DCAL measurements to an R² of 0.87. There are several possible reasons for this discrepancy. The TAC-D4 or DCAL measurement may be erroneous. Alternatively, free and total tacrolimus concentrations may vary from patient to patient, and this patient may have a greater than average difference between their free and total tacrolimus concentrations. As noted above, varying patient hematocrit, for example, has been linked to influencing the amount of free:total tacrolimus in patient blood [41, 42, 44, 45]. Efforts are ongoing to assess the accuracy of the TAC-D4 compared to DCAL measured tacrolimus concentrations when normalized for hematocrit.

Additionally, we acknowledge that the clinical comparison presented here involved a limited number of patient samples (n=8). Our primary objective in this study was to establish the analytical foundation of the TAC-D4 platform—demonstrating its feasibility for measuring tacrolimus directly from unprocessed whole blood and benchmarking its performance against DCAL. The clinical comparison was intended as an initial proof-of-concept rather than a comprehensive clinical validation. A larger, adequately powered clinical validation study with an expanded patient cohort is actively being planned and will be reported in a future manuscript.

Third, while our 1-hour turnaround time represents a significant improvement over centralized laboratory testing, further optimizing the test to reduce the time-to-result would improve the value of our tests in point-of-care settings. In the current study, the 60-minute incubation time was empirically optimized to balance assay sensitivity and practical turnaround requirements for point-of-care use. Preservation of trehalose pad dissolution and dAb binding activity over storage was assessed by running the same printed chip lot at Day 1 and Day 30 post-fabrication on tacrolimus-spiked whole blood serial dilutions; both timepoints delivered an LOD below 0.78 ng/mL and produced dose-response curves with preserved dynamic range (Supplementary Figure 11). Single-point depressions were observed at the Day 1 1.5625 ng/mL concentration and at the Day 30 blank, which lowered the corresponding replicate means and elevated variance. These anomalies did not significantly affect the overall LOD at either timepoint, but they identify the blank and low-end clinical concentrations as the conditions most sensitive to between-run variability and motivate larger stability and reproducibility characterization. A comprehensive temporal kinetics study—including determination of the minimum incubation time for reliable results, the effects of varying temperature and humidity conditions, and a longer-duration stability study, building on the environmental resilience previously demonstrated for the broader D4 platform [19]—will be reported on in a future study.

Finally, for the VCA-D4 a correlation study using patient samples with robust IL-6 and MMP3 data from accepted laboratory tests would further validate the accuracy and reliability of our multiplexed approach. IL-6 and MMP3 are compelling initial targets for a combinatorial TDM, and prognostic tool based on their emerging roles in transplant rejection pathophysiology but their potential for early rejection detection in VCA recipients need to be further defined. Importantly, the modular design of the TAC-D4, as evidenced by the multiplexed VCA-D4, enables rapid adaptation for emerging needs in TDM and beyond. The platform architecture can accommodate future biomarker discoveries, creating a scalable approach to personalized transplant monitoring with an increasing density of information.

To contextualize the TAC-D4 within the broader landscape of tacrolimus monitoring technologies, we provide a brief comparison of key performance metrics. Current LC-MS/MS methods, while considered the gold standard for accuracy, require extensive sample preparation (cell lysis, protein precipitation, and liquid-liquid or solid-phase extraction), specialized instrumentation costing $300,000–$500,000 or more, and trained laboratory personnel, with typical turnaround times of 4–8 hours or longer [8, 9]. Emerging POCT approaches for tacrolimus, such as lateral flow immunoassays [46] and electrochemical biosensors [47], have been reported in the literature but have generally not achieved the combination of direct whole-blood analysis, quantitative measurement across the full clinical range, and integration into a self-contained microfluidic device. In contrast, the TAC-D4 POCT requires no sample preparation beyond a simple 1:10 dilution with surfactant, uses a low-cost portable reader (the D4Scope), delivers quantitative results in approximately 1 hour, and can operate at the point of care without centralized laboratory infrastructure. Furthermore, the multiplexed VCA-D4 format offers the unique capability to simultaneously monitor tacrolimus alongside protein biomarkers (IL-6 and MMP3) on a single chip, a feature not available in any current commercial tacrolimus monitoring platform. These attributes position the TAC-D4 as a complementary tool to existing laboratory methods, particularly for settings where rapid turnaround, decentralized testing, or multi-analyte monitoring are advantageous.

Supplementary Material

The following files are available free of charge:

Supplementary Figures (file type, PDF)

TAC-D4 Raw Data File (file type, XLSX)

ABBREVIATIONS

POCT

point-of-care test

TDM

therapeutic drug monitoring

TAC-BSA

tacrolimus-bovine serum albumin

LC-MS/MS

liquid chromatography-tandem mass spectrometry

DCAL

Duke Clinical Automated Laboratory

CHTN

Cooperative Human Tissue Network

POEGMA-co-GMA

poly[oligo(ethylene glycol) methyl ether methacrylate-co-glycidyl methacrylate]; poly[oligo(ethylene glycol) methyl ether methacrylate

LOD

limit of detection

RE

relative error

SI-ATRP

surface initiated atomic transfer radical polymerization

dAb

detection antibody

cAb

capture antibody

dAg

detection competitive antigen

cAg

capture competitive antigen

VCA

vascular composite allotransplantation

IL-6

interleukin-6

MMP3

matrix metalloproteinase 3