i have attached the paper
i need powerpoint slide presentation of atleat 20 slides to be prepared
Talanta 252 (2023) 123842
Contents lists available at ScienceDirect
Talanta
journal homepage: www.elsevier.com/locate/talanta
Microporous affinity membranes and their incorporation into microfluidic
devices for monitoring of therapeutic antibodies
Joshua D. Berwanger a, 1, Melinda A. Lake b, 1, Sanniv Ganguly c, Junyan Yang d,
Christopher J. Welch e, Jacqueline C. Linnes c, Merlin Bruening a, d, *
a
Department of Chemistry and Biochemistry, University of Notre Dame, Notre Dame, IN, 46556, United States
School of Mechanical Engineering, Purdue University, West Lafayette, IN, 47907, United States
Weldon School of Biomedical Engineering, Purdue University, West Lafayette, IN, 47907, United States
d
Department of Chemical and Biomolecular Engineering, University of Notre Dame, Notre Dame, IN, 46556, United States
e
Indiana Consortium for Analytical Science and Engineering (ICASE), Indianapolis, IN, 46202, United States
b
c
A R T I C L E I N F O
A B S T R A C T
Keywords:
Mimotopes
Monoclonal antibodies (mAbs)
Therapeutic drug monitoring
Vertical flow assays
Point-of-care
Microfluidics
Control of monoclonal antibody (mAb) concentrations in serum is important for maintaining the safety and
efficacy of these lifesaving therapeutics. Point-of-care (POC) quantification of therapeutic mAbs could ensure
that patients have effective mAb levels without compromising safety. This work uses mimotope-functionalized
microporous alumina affinity membranes in vertical flow assays for detection and quantitation of therapeutic
mAbs. Selective capture of bevacizumab from 1000:1 diluted serum or plasma and binding of a fluorescently
labelled anti-human IgG secondary antibody enable fluorescence-based analysis of bevacizumab at its thera
peutically relevant concentration range of ~50–300 μg/mL. The assay results in a linear relationship between the
fluorescence intensity of the antibody capture spot and the bevacizumab concentration. A simple prototype
microfluidic device containing these membranes allows washing, reagent additions and visualization of signal
within 15 min using a total of 5 mL of fluid. The prototype devices can monitor physiologically relevant bev
acizumab levels in diluted serum, and future refinements might lead to a POC device for therapeutic drug
monitoring.
1. Introduction
Pharmaceutical applications of monoclonal antibodies (mAbs) have
grown at an astonishing rate since the introduction of the first thera
peutic mAb in 1986 [1]. As of April 2021, 100 therapeutic monoclonal
antibodies were approved as treatments, and that number has grown
since then, with mAbs accounting for almost a fifth of new approved
drugs in the US [1–4]. mAb therapies are highly specific in treating in
flammatory diseases, autoimmune diseases, pain, and cancers, but their
efficacy often depends on maintaining the proper mAb concentration in
blood [5–10]. For example, clinical pharmacokinetic studies showed a
large patient-to-patient variation in the serum concentrations of two
cancer therapeutics, bevacizumab and trastuzumab, at the same time
point after mAb administration [11–14]. For both mAbs, the serum
concentration affected the outcome of the treatment. In the case of
trastuzumab and ado-trastuzumab emtansine (a drug conjugate), higher
exposure to the drug correlated with higher drug efficacy while patients
with lower exposure had shorter overall survival times [10,15]. With
bevacizumab, higher survival chances correlated with higher mAb
concentrations in both metastatic colon cancer and glioma, but in the
case of glioma, side effects began to arise as the concentration of bev
acizumab increased beyond 250 mg/L [8,11]. The inter-patient vari
ability of pharmacokinetics is an unmet problem of the current standard
dosage regimens that are based on body weight (mg/kg) or set dosages.
Therapeutic drug monitoring could address the challenge of interpatient
pharmacokinetic variability and inform personalized dosage regimens to
increase the clinical effectiveness and potentially lower the cost of these
treatments [8,16]. This approach has shown effectiveness for several
drug classes such as antibiotics, antiepileptics, and immunosuppres
sants, but its usage in therapeutic mAb treatment is still nascent
[16–23].
Ideally, therapeutic monitoring of mAbs will employ a fast and
* Corresponding author. Department of Chemical and Biomolecular Engineering, University of Notre Dame, Notre Dame, IN, 46556, United States.
E-mail address: mbruenin@nd.edu (M. Bruening).
1
equal contribution.
https://doi.org/10.1016/j.talanta.2022.123842
Received 14 May 2022; Received in revised form 8 August 2022; Accepted 11 August 2022
Available online 17 August 2022
0039-9140/© 2022 Elsevier B.V. All rights reserved.
J.D. Berwanger et al.
Talanta 252 (2023) 123842
inexpensive method to determine their concentrations in patient blood.
Current methods for mAb quantitation include enzyme-linked immu
nosorbent assays (ELISAs), liquid chromatography-mass spectrometry,
and Single Molecular Array (SIMOA)/Luminex assays [24–28].
Although these techniques are sensitive and accurate, they require
several time-consuming steps as well as expensive equipment in some
cases. Thus, they are not applicable to point-of-care (POC) therapeutic
drug monitoring [29,30].
Vertical flow through membranes in microfluidic devices may afford
rapid capture of targets to enable POC analyses. Many studies demon
strated the utility of flow through membranes in immunoassays
[31–33]. Examples of such devices include a colorimetric method to
detect B. pseudomallei down to 0.02 ng/mL in a nitrocellulose membrane
[34], and a fluorescence-based assay in a paper fluidic device to detect
cancer biomarkers from 0.1 to 1000 ng/mL [35]. Additionally, re
searchers developed a POC device to monitor flucytosine uptake in a
nitrocellulose membrane using gold nanoparticles to detect 10 μg/mL
levels using surface enhanced Raman spectroscopy [36]. We demon
strated a microplate-based vertical flow assay that employs an
epitope-mimicking peptide (mimotope) to quantify trastuzumab and a
mAb for SARS-Cov-2 in less than 5 min [37].
This study explores vertical-flow immunoassays to rapidly determine
bevacizumab concentrations in patient blood plasma or serum using a
macroscale fluidic device as well as a proof-of-concept microfluidic
device designed to reduce both assay time and reagent volume. Building
on recent work on the capture and analysis of therapeutic mAbs using
microporous membranes [38], we employ a microfluidic workflow that
utilizes membranes covalently modified with a mAb-binding mimotope
[39] to capture and quantify target therapeutic mAbs in patient serum.
Such membranes are attractive because they may efficiently capture
mAbs in minutes (including rinsing steps). Importantly, quantitation of
the mAbs via a fluorescently labelled secondary antibody allows
detection of bevacizumab at clinically relevant levels between ~50 and
300 ng/mL in 1000-fold diluted serum [8,11]. Our microfluidic chip that
exploits vertical-flow capture consists of layers of plastic and pressure
sensitive adhesive (PSA), making it low-cost and scalable through
roll-to-roll manufacturing [40] or adaptable into an injection molded
platform [41]. We integrate a functionalized porous alumina membrane
within the adhesive and plastic layers. Thin, optically clear layers enable
imaging of the membrane within the microfluidic chip, rendering the
device adaptable to point-of-care smartphone-based imaging platforms
[42].
assays were prepared. Wash Buffer 1 is a solution containing 20 mM
PBS, 500 mM NaCl, and 0.1% (v/v) Tween 20 (pH 7.2). Wash Buffer 2 is
a solution containing 20 mM PBS, 500 mM NaCl, 0.1% Tween 20, and
0.1% PVA at (pH 7.2). PVA was added to remove non-specifically bound
proteins.
2.2. Modification of bare alumina membranes with peptide mimotopes
Layer-by-layer adsorption was employed to modify alumina mem
branes (Fig. 1A) [43]. UV/O3-cleaned membranes were washed with
water prior to immersion in a PAA solution (10 mM of the PAA repeating
unit in 500 mM aqueous NaCl, pH 4) for 5 min. The membranes were
then immersed sequentially in deionized water for 1 min, PAH solution
for 5 min (10 mM of the PAH repeating unit in aqueous 500 mM NaCl,
pH 4), and deionized water for 1 min. This process was repeated until
the membrane was modified with the desired number of polyelectrolyte
bilayers (1.5, 2.5, or 3.5; the extra 0.5 bilayer indicates that the film ends
in PAA). The membranes were then dried gently with N2 gas. After
drying, the (PAA/PAH)xPAA-modified alumina membranes were rinsed
with water for 10 s on both sides and immersed in a 0.1 M NHS, 0.1 M
EDC aqueous solution for 30 min. The activated membranes were again
rinsed with water for 10 s on either side and dried with N2 gas. Next,
0.75 μL of Bev17 mimotope peptide (1 mg/mL in 0.1 M NaHCO3, pH 9)
was pipetted onto the middle of the membrane. The membranes were
then placed in covered polystyrene petri dishes saturated with water
vapor by a cotton ball swab to allow covalent immobilization of Bev17
overnight. Note that the poly(acrylic acid) modification occurs over the
entire membrane, so only Bev17 binding and subsequent mAb capture
should affect flow through the spot relative to the rest of the porous
alumina. Membranes used in the microfluidics assay were shipped
overnight on ice and stored in a 4 ◦ C refrigerator upon arrival.
2.3. Capture and analysis of bevacizumab with mimotope-modified
alumina membranes
After letting Bev17-modified alumina membranes sit overnight at
room temperature, the membranes were used in an antibody capture
assay (Fig. 1B) in a custom-built Teflon vertical flow device that holds
the membrane, as detailed elsewhere [44–46]. The membranes were
rinsed for 10 s on each side with water. Membrane testing employed a
peristaltic pump (HV-77120-62 Masterflex, Gelsenkirchen, GER) to pull
fluid from an inlet reservoir through the membranes. The pump is
connected via tubing to the outlet of the membrane holder. After placing
a membrane in a Teflon holder, it was rinsed by flowing 30 mL of Wash
Buffer 1 through the membrane at 1 mL/min using the peristaltic pump.
Next, 1 mL of bevacizumab at concentrations ranging from 0 to 500
ng/mL in 20 mM PBS, 150 mM NaCl, pH 7.4 was circulated through the
membrane for varying amounts of time from 1 to 10 min at 1 mL/min.
The membranes were then washed with varying volumes between 5 and
30 mL of Wash Buffer 1. After washing, 1 mL of Cy5-labelled Anti-
Human IgG (10 μg/mL in 20 mM PBS, 500 mM NaCl, pH 7.4) was
circulated through the membrane at 1 mL/min for varying amounts of
time between 10 min and 50 min. The membranes were then washed
again with Wash Buffer 1, removed from the pump setup, rinsed with
water for 10 s on each side, dried with N2 gas, and analyzed using an
Azure C400 Bioanalytical Imaging System in the Cy5 imaging mode
using a 50 ms exposure time. Quantitation of the images was carried out
using the ImageJ intensity measurement function by integrating the
intensity over the area of a circle of 4.9 mm2 that fit within the edges of
the mimotope spots. The circle size was determined by finding the area
that fit within the spots on all the membranes analyzed in that data set.
The fluorescence intensity was reasonably uniform across the spots (see
Fig. S4 in the supporting information).
After developing conditions for capture and analysis of bevacizumab
in buffer, the process was repeated using mAb spiked in human fluids.
Experiments were repeated with both human serum and human plasma.
2. Experimental section
2.1. Materials
Alumina membranes (Whatman Anodisc inorganic filter membranes,
25 mm diameter, 0.2 μm pore size) were cleaned in a UV/O3 chamber
(Jelight, model 18) for 15 min prior to use. AcetylWLEMHWPAHSGSGSGSK (Bev17, the mimotope that binds to bev
acizumab) was synthesized by Genscript with a purity greater than 95%.
Polyallylamine hydrochloride (PAH, Mw = 50,000), poly(acrylic acid)
(PAA, average molecular weight ~100,000 Da, 35% aqueous solution),
Tween-20 surfactant, N-hydroxysuccinimide (NHS), N-(3-dimethylami
nopropyl)-N′ -ethylcarbodiimide hydrochloride (EDC), and human
serum were used as received from Sigma Aldrich. BioChemEd Services
provided deidentified patient serum samples. Human blood was
received from Innovative Research in a sodium citrate anti-coagulant
tube and centrifuged at 2000 ×g for 10 min before extracting plasma
using a pipette. Poly(vinyl) alcohol (PVA, 99–100% hydrolyzed,
approximate molecular weight 8600 Da) was obtained from Acros.
Bevacizumab (Genentech) was used from its therapeutic formulations.
Buffers were prepared using analytical grade chemicals from various
chemical providers, and Milli-Q, 18.2 MΩ-cm deionized water was used
to prepare all aqueous solutions. Two Wash buffers used throughout the
2
J.D. Berwanger et al.
Talanta 252 (2023) 123842
Fig. 1. Alumina membrane modification and steps of a flow assay. A) Modification of a polyelectrolyte-coated alumina membrane with peptide mimotopes.
Polyelectrolyte deposition, mimotope spotting, and mAb capture occur throughout the depth of the alumina substrate. B) Capture and analysis of bevacizumab using
mimotope-modified alumina membranes and binding of a fluorescently labelled secondary antibody.
The human fluids were first diluted 1000:1 with a solution containing
20 mM PBS and 500 mM NaCl at pH 7.4. These diluted samples were
then spiked with a known amount of the target bevacizumab at con
centrations ranging from 0 to 500 ng/mL. The Bev17-spotted mem
branes were pretreated with Wash Buffer 1 as described above. After
pretreatment, 1 mL of the bevacizumab-spiked, diluted human fluid was
passed once through the membrane at 1 mL/min to capture the target.
After capture, the membranes were washed with varying volumes be
tween 5 and 30 mL of Wash Buffer 2. After washing, 1 mL of Cy5labelled Anti-Human IgG (1–10 μg/mL in 20 mM PBS, 500 mM NaCl,
pH 7.4) was circulated through the membrane at 1 mL/min for 10 min.
Membranes were washed again with Wash Buffer 2, rinsed, dried, and
analyzed as described above using fluorescence imaging.
assembled manually. COP was cut using a laser power of 80% and a
speed of 100%, and the PSA was cut using a laser power of 90% with a
speed of 90%. The membranes have an initial diameter of 25 mm prior
to laser cutting and have an outer ring of plastic that provides a pro
tective border for gripping the membrane. The 25-mm membrane with
an antibody spot was laser cut to 10 mm using a laser power of 56% and
a speed of 80%. Laser cutting removes the protective plastic ring, so
afterward the membrane is directly handled with tweezers. Chip as
sembly could be scaled to a roll-to-roll manufacturing setup or adapted
into an injection moldable design. The PDMS layer is added to secure the
PEEK tubing (Part 1569, Idex Health & Science, Oak Harbor, WA, USA)
to the inlet and outlet. PDMS was fabricated using a 10:1 base:curing
agent ratio, baked in a Petri dish for 2 h at 65 ◦ C, and hole punched and
diced to size after curing. The COP pieces were rinsed with 70% v/v
ethanol in water and dried using a Kimwipe prior to assembly. Layer 0 to
Layer 4 were assembled and stored at room temperature for up to two
weeks, while Layers 5 and 6 were assembled on the day of the experi
ment following placement of the alumina membrane in the pre-designed
slot in layer 4.
2.4. Microfluidic device fabrication
Microfluidic devices were designed in AutoCAD, and the files were
transferred to Adobe Illustrator for production. Each device contained
six layers: 2 layers of 0.19 mm-thick cyclic olefin polymer (COP) (Zeo
nor, Tokyo, Japan), 1 layer of 0.05 mm-thick COP, 3 layers of PSA
(93020LE, 3 M, St. Paul, MN, USA), and 1 layer of polydimethylsiloxane
(PDMS) (Sylgard 184, Dow Corning, Midland, MI, USA) (Fig. 2). Each
plastic and adhesive layer of the chip was cut using a laser cutter
(VLS3.50, Universal Laser Systems Inc., Scottsdale, AZ, USA) and
2.5. Microfluidic device assay
In binding studies for the microfluidic device assay, all solutions
were passed over the membrane through the chip with a syringe pump
Fig. 2. A) Exploded view of the microfluidic device assembly. B) Photograph of an assembled chip filled with blue food coloring. COP = cyclic olefin polymer. PSA =
pressure sensitive adhesive. PDMS = polydimethylsiloxane. (For interpretation of the references to color in this figure legend, the reader is referred to the Web
version of this article.)
3
J.D. Berwanger et al.
Talanta 252 (2023) 123842
(Part 788,212, KD Scientific, Holliston, MA) connected to the inlet to
maintain a constant flowrate and a measurable residence time for
interaction. The outlet side had a tube connected to a waste container.
The binding was performed in 3 separate sets of trials evaluating the
dose-dependent fluorescence in response to various concentrations of
bevacizumab spiked into serum. Trial 1 and Trial 2 were tested out of the
same prepared membrane batch: Trial 1 occurred within 2 days of
membrane delivery and Trial 2 occurred 1 week later. Trial 3 was tested
within 2 days of delivery of a new batch of membranes. After sealing the
laser-cut alumina into the device, the modified alumina membrane
within the chip was first washed with 0.8 mL of Wash buffer 1 at a
constant flow rate of 1 mL/min, followed by running the 1000:1 diluted
serum with 20 mM PBS and 500 mM NaCl at pH 7.2 spiked with varying
concentrations of bevacizumab from 0 to 300 ng/mL at a flow rate of 0.5
mL/min with a volume of 0.16 mL. This was followed by washing the
membrane with 0.8 mL of Wash Buffer 2. Goat Anti-Human IgG H&L
(Cy5 ®) (Abcam) was then diluted to 1 μg/mL in 20 mM PBS, 500 mM
NaCl, pH 7.4, and 0.6 mL of this dilute secondary antibody solution was
flowed through the membrane for 10 min at 0.06 mL/min. This was
followed by washing the membrane with 0.8 mL of Wash Buffer 2 and
rinsing with 0.8 mL of deionized water, both at 1 mL/min. The tubing
and pump were disconnected, then the PDMS layer was peeled off and
the membranes were dried with N2, but not removed from within the
device. The membranes were then imaged from the top in the Azure
c600 fluorescence imager (Azure Biosystems, Dublin, CA, USA) in the
Cy5 mode with a 50 ms exposure time. The contrast settings of the Azure
scanner were adjusted to the same black/white values of 0/6242 ± 23 a.
u. for the 3 trials to make the spots bright enough to see and yet avoid
saturation of the fluorescence. The small deviation in the setting is due
to the machine options. Quantitation of the images was carried out using
the ImageJ intensity measurement function by integrating the intensity
over the area of a 1.5 mm2 circle that fit within the edges of the mim
otope spot. The circle size was determined by finding the area that fit
within the spots on all the membranes analyzed in that data set. The
membranes were also imaged on an inverted microscope (Zeiss Axio
Observer Z1, Carl Zeiss Microscopy, Jena, Germany) in brightfield mode
using the 5X-20X objectives to inspect for cracks.
After selecting parameters for bevacizumab assays in buffer, we
examined bevacizumab-spiked samples in diluted human serum or
plasma using the same parameters. With a 1000-fold dilution, the serum
or plasma components had minimal effect on bevacizumab binding to
the mimotope on the alumina membrane. Removal of non-specifically
bound proteins occurs during a wash step prior to capture of a second
ary antibody as well as a final rinse; the PVA added to Wash Buffer 2 aids
in removing excess proteins. Fig. 3A shows results from the analysis of
bevacizumab in 1000-fold diluted serum. This 0–300 ng/mL assay range
corresponds to the therapeutically relevant bevacizumab concentrations
that would be present in 1000:1 diluted serum from patients [8,11]. The
figure shows a linear relationship between the concentration of bev
acizumab flowed through the membrane and the fluorescence intensity.
Equally important, the background signal for 0 ng/mL of bevacizumab is
low, indicating that under these conditions the 5-mL wash removes
essentially all non-specifically bound secondary antibody. The low
signal with no added bevacizumab also confirms that the signals
observed on the other spots arise from the secondary antibody binding
to bevacizumab and not from other proteins. The relatively large stan
dard deviations in Fig. 3B likely stem from the challenge of exactly
reproducing the membrane preparation, particularly the spotting pro
cess, which was performed by hand. Nevertheless, the fluorescence
values over the range from 100 to 300 ng/mL have a coefficient of
variation (CV) less than 22%, with 80% having a CV of 20% or less.
These uncertainties would likely be reduced further with standardiza
tion of membrane preparation to inform physicians about patient mAb
levels in diluted serum. In particular, the spot size varies as much as 13%
from the average spot size in 3 replicate measurements with the different
concentrations for macroscale tests (Table S2) and 14% for the micro
fluidic tests described below (Table S3).
The results in Fig. 3A and B employed (PAA/PAH)3PAA films for
membrane modification, following a literature procedure [43]. We also
tested the assay with (PAA/PAH)2PAA and (PAA/PAH)PAA films to
decrease the transmembrane pressure, increase porosity, and facilitate
incorporation of the membranes into a microfluidics device. Preliminary
testing in the microfluidic devices with (PAA/PAH)3PAA-modified
alumina resulted in cracked membranes and burst bonds between device
layers due to the high pressures. Fig. S3A shows that the background
points for 0 ng/mL bevacizumab for membranes modified with (PAA/
PAH)2PAA and (PAA/PAH)PAA films also showed minimal signals.
Moreover, as Figs. S3B–C show, the relationship between fluorescence
intensity and the bevacizumab concentration in 1000:1 diluted serum is
linear for membranes modified using (PAA/PAH)2PAA and (PAA/PAH)
PAA films. Thus, we proceeded with membranes coated with (PAA/
PAH)PAA films because they show a linear response, low background,
and relatively low back pressure.
Similar assays with spiked, diluted human plasma also show a linear
relationship between the measured fluorescence intensity and the con
centration of bevacizumab in solutions passed through the membrane
(Fig. 3C). There are three major differences in this assay compared to the
procedure for the diluted serum. First, the assay with plasma required
double the washing volume (10 mL vs 5 mL) to sufficiently reduce nonspecific binding. Second, to achieve sufficient sensitivity the assay
required 10 μg/mL of the Cy5 labelled anti-human IgG, whereas the
serum assay only requires 1 μg/mL. The plasma has an anticoagulant
and thus a different composition than the serum including proteins,
calcium, and magnesium levels that may cause higher background or
interactions with the bevacizumab/Bev17 that interfere with binding
[48]. Nonetheless, the assay still exhibits the needed sensitivity while
taking less than 35 min to complete. Based on the current standard
deviations, the assay affords a coefficient of variation below 23%, except
the 200 ng/mL point which has a much larger standard deviation
(Fig. 3D). The standard deviations are relatively large, but automated
production of the membranes with precise spot placement and calibra
tion of fluorescence using standards can likely overcome this challenge
to reduce the errors to less than 20%.
3. Results and discussion
3.1. Mimotope-modified alumina membranes for analysis of bevacizumab
in human samples
This study uses fluorescence-based detection from porous alumina
membranes with a total analysis time of less than 35 min in a membrane
holder and only 15 min in a microfluidic device. The long optical
pathlength in alumina membranes provides up to two orders of
magnitude greater sensitivity relative to assays using flat surfaces,
allowing for shortened assay times [43]. Our prior work demonstrated
mAb capture in modified nylon membranes, quantitation based on
native mAb fluorescence, and preliminary use of a fluorescent secondary
antibody to increase sensitivity with an analysis time of approximately
2 h [47]. Decreasing the assay time to 35 min requires optimization of all
protocol steps including target antibody capture, washing, and second
ary antibody capture. Initially, we studied analysis of mAbs in buffer to
select the parameters for subsequent assays in serum and plasma. The
Supporting Information describes our choices of times for each step of
the analysis (Figs. S1 and S2, and Table S1) in large membrane holders.
The final assay configuration includes a single pass of 1 mL of the pri
mary mAb bevacizumab through the 4.9-cm2 membrane, 5 mL washing
steps that occur after capture of the primary and secondary antibodies,
and 10 min of secondary antibody circulation. These parameters
represent a compromise between minimal analysis times and achieving
high fluorescence signals with low background. All membranes are
designed for single use due to their fragility and a desire to avoid
contamination.
4
J.D. Berwanger et al.
Talanta 252 (2023) 123842
Fig. 3. Analysis of bevacizumab in diluted human
serum and diluted human plasma. A) Contrastenhanced fluorescence image of Bev17-spotted
(PAA/PAH)3PAA-modified membranes after cap
ture of bevacizumab from the diluted serum con
taining the indicated mAb concentration, rinsing,
passage of a fluorescent secondary antibody, and
further rinsing. B) Average intensity of spotted cap
ture membranes as a function of bevacizumab con
centration for the treatment described in A. The
points on the line in panel B show the average
measured intensity from 4 different membranes
prepared on different days for each different con
centration. C) Contrast-enhanced fluorescence image
of Bev17-spotted (PAA/PAH)PAA-modified mem
branes after capture of Bevacizumab from diluted
plasma containing the indicated mAb concentration,
rinsing, passage of a fluorescent secondary antibody,
and further rinsing. D) Average intensity of spotted
capture membranes as a function of bevacizumab
concentration for the treatment described in C. The
points on the line in panel B show the average
measured intensities from 3 different membranes
prepared on different days at each different concentration. The error bars correspond to the standard deviations.
As a comparison, we performed trastuzumab ELISAs in buffer using
commercial kits and the manufacturer’s protocols. With a commercial
human IgG assay, the coefficient of variation was 10–20% when using
0.5–10 ng/mL and an assay time of 90 min. An anti-Her2 ELISA kit
showed a coefficient of variation of 5% with concentrations from 10 to
100 ng/mL and an assay time of 120 min. Fig. S5 in the supporting
material shows calibration curves. Detection limits were 0.2 ng/mL for
the human IgG assay and 1.4 ng/mL for the anti-Her2 assay. Bev
acizumab ELISA kits have a limit of quantitation around 30 ng/mL [49].
The ELISAs take at least an hour longer than the membrane-based
method, and the human IgG assay shows similar uncertainty as the
membrane-based system. ELISA detection limits are 2 orders of magni
tude lower than those with the current membrane system. However,
typical bevacizumab therapeutic concentrations are in the range of
50–300 μg/mL [11], so low detection limits are not needed for this mAb.
The main advantage of the membrane-based assays is a reduction in
time, which we aim to decrease further in the future. Changes in the
fluorophore on the secondary antibody should also decrease detection
limits in the membrane-based assays.
3.2
Incorporation
ofmimotope-modifiedmembranes
in
amicrofluidicchip
After developing analyses of bevacizumab in serum and plasma, the
assay was miniaturized into a microfluidic workflow. The advantages of
microfluidics include a decreased assay time, smaller reagent volumes,
further method simplification, and the potential for automation, which
could be valuable for future clinical use. Importantly, in a microfluidic
device format this assay occurred in under 15 min and used a total of 4
mL of fluid, including the wash buffer and rinsing water. Thus, the
prototype microfluidic assay demonstrates greater than 5X reduction in
the reagent volumes compared to the standard assay. Other refinements
may further decrease the required volumes to make this device more
compatible with point of care diagnostics. Techniques such as ELISA
employ small volumes, but they typically require >1 h for analysis and a
plate reader for quantitation.
Fig. 4 shows that for a given set of replicate experiments the fluo
rescence generally increases linearly with an increasing concentration of
bevacizumab. Also, the background fluorescence in the membrane and
the surrounding microfluidic chip is low compared to the fluorescent
spot at the center of the membrane (Fig. 4A). The control without bev
acizumab confirms the specificity of binding of the secondary antibody
to bevacizumab.
The fairly large differences in fluorescence intensities between
different experimental replicates at the same bevacizumab concentra
tions may stem from variations in membrane preparation, or shipping
conditions. For example, the alumina material is brittle and despite
careful handling with tweezers, may develop microscopic or difficult to
see cracks during device assembly. Such cracks will affect the assay by
altering the flow-through properties of the membrane. After laser cut
ting the membranes down to 10 mm, the protective plastic ring is
removed and thus the edge of the brittle membrane is directly handled
with tweezers. When cracks were detected on a microscope, data points
were dropped for Trial 2 at 150 and 250 ng/mL and Trial 1 at 300 ng/
mL. Membranes used in the microfluidic assay were shipped on ice from
Notre Dame to Purdue, so they had different storage conditions than the
membranes fabricated and directly tested at Notre Dame. However, the
microfluidic assay still demonstrates an approximately linear trend for
each trial run with increasing bevacizumab concentration. Further, the
manual spotting process for placing the spot at the center of the mem
brane in addition to different membrane storage times may add to
variability in the microfluidic assay results. This could be ameliorated
Fig. 4. Analysis of bevacizumab in diluted (1000:1)
human serum using a microfluidic assay. A) Contrastenhanced fluorescence image of Bev17-spotted mem
branes after capture of bevacizumab from the diluted
serum containing the indicated mAb concentration,
rinsing, passage of a fluorescent secondary antibody,
and further rinsing. B) Raw data collected for each
membrane tested in the microfluidic system, excluding
cases where membrane cracks were discovered. Each
point represents a different membrane. The image in
panel A corresponds to Trial 3 and was contrast-
enhanced to increase visibility of the spots.
5
J.D. Berwanger et al.
Talanta 252 (2023) 123842
with automated membrane production including precise spot place
ment. However, the signal areal intensity did not show a correlation
with spot size (see Fig. S6). Calibration of the fluorescence using stan
dards integrated into the assay may also help future quantification.
Further work is needed to increase reproducibility. With significant
refinement, the microfluidic assay could allow simple point-of-care an
alyses that do not require the instrumentation currently used in assays
that employ 96-well plates.
org/10.1016/j.talanta.2022.123842.
References
[1] R.-M. Lu, Y.-C. Hwang, I.-J. Liu, C.-C. Lee, H.-Z. Tsai, H.-J. Li, H.-C. Wu,
Development of therapeutic antibodies for the treatment of diseases, J. Biomed.
Sci. 27 (2020) 1, https://doi.org/10.1186/s12929-019-0592-z.
[2] Homepage of The Antibody Society, Antib. Soc. https://www.antibodysociety.org/
home/(accessed May 13, 2022).
[3] A. Mullard, FDA approves 100th monoclonal antibody product, Nat. Rev. Drug
Discov. 20 (2021) 491–495, https://doi.org/10.1038/d41573-021-00079-7.
[4] S.S. Wang, Y. (Susie) Yan, K. Ho Yan, US FDA-approved therapeutic antibodies
with high-concentration formulation: summaries and perspectives, Antib. Ther. 4
(2021) 262–272, https://doi.org/10.1093/abt/tbab027.
[5] S. Singh, N.K. Tank, P. Dwiwedi, J. Charan, R. Kaur, P. Sidhu, V.K. Chugh,
Monoclonal antibodies: a review, Curr. Clin. Pharmacol. 13 (2018) 85–99, https://
doi.org/10.2174/1574884712666170809124728.
[6] M. Berger, V. Shankar, A. Vafai, Therapeutic applications of monoclonal
antibodies, Am. J. Med. Sci. 324 (2002) 14–30, https://doi.org/10.1097/
00000441-200207000-00004.
[7] C.L.M. Krieckaert, S.C. Nair, M.T. Nurmohamed, C.J.J. van Dongen, W.F. Lems, F.
P.J.G. Lafeber, J.W.J. Bijlsma, H. Koffijberg, G. Wolbink, P.M.J. Welsing,
Personalised treatment using serum drug levels of adalimumab in patients with
rheumatoid arthritis: an evaluation of costs and effects, Ann. Rheum. Dis. 74
(2015) 361–368, https://doi.org/10.1136/annrheumdis-2013-204101.
[8] A. Papachristos, P. Kemos, H. Kalofonos, G. Sivolapenko, Correlation between
bevacizumab exposure and survival in patients with metastatic colorectal cancer,
Oncol. 25 (2020) 853–858, https://doi.org/10.1634/theoncologist.2019-0835.
[9] A.L. Quartino, C. Hillenbach, J. Li, H. Li, R.D. Wada, J. Visich, C. Li, D. Heinzmann,
J.Y. Jin, B.L. Lum, Population pharmacokinetic and exposure–response analysis for
trastuzumab administered using a subcutaneous “manual syringe” injection or
intravenously in women with HER2-positive early breast cancer, Cancer
Chemother, Pharma 77 (2016) 77–88, https://doi.org/10.1007/s00280-015-29225.
[10] J. Yang, H. Zhao, C. Garnett, A. Rahman, J.V. Gobburu, W. Pierce, G. Schechter,
J. Summers, P. Keegan, B. Booth, Y. Wang, The combination of exposure-response
and case-control analyses in regulatory decision making, J. Clin. Pharmacol. 53
(2013) 160–166, https://doi.org/10.1177/0091270012445206.
[11] G. Nugue, M. Bidart, M. Arlotto, M. Mousseau, F. Berger, L. Pelletier, Monitoring
monoclonal antibody delivery in oncology: the example of bevacizumab, PLoS One
8 (2013), e72021, https://doi.org/10.1371/journal.pone.0072021.
[12] J. Baselga, X. Carbonell, N.-J. Castañeda-Soto, M. Clemens, M. Green, V. Harvey,
S. Morales, C. Barton, P. Ghahramani, Phase II study of efficacy, safety, and
pharmacokinetics of trastuzumab monotherapy administered on a 3-weekly
schedule, J. Clin. Oncol. 23 (2005) 2162–2171, https://doi.org/10.1200/
JCO.2005.01.014.
[13] R. Bruno, C.B. Washington, J.-F. Lu, G. Lieberman, L. Banken, P. Klein, Population
pharmacokinetics of trastuzumab in patients with HER2+ metastatic breast cancer,
Cancer Chemother. Pharmacol. 56 (2005) 361–369, https://doi.org/10.1007/
s00280-005-1026-z.
[14] D. Ternant, N. Cézé, T. Lecomte, D. Degenne, A.-C. Duveau, H. Watier, E. Dorval,
G. Paintaud, An enzyme-linked immunosorbent assay to study bevacizumab
pharmacokinetics, Ther. Drug Monit. 32 (2010) 647–652, https://doi.org/
10.1097/FTD.0b013e3181ef582a.
[15] J. Wang, P. Song, S. Schrieber, Q. Liu, Q. Xu, G. Blumenthal, L. Amiri Kordestani,
P. Cortazar, A. Ibrahim, R. Justice, Y. Wang, S. Tang, B. Booth, N. Mehrotra,
A. Rahman, Exposure-response relationship of T-DM1: insight into dose
optimization for patients with HER2-positive metastatic breast cancer, Clin.
Pharmacol. Ther. 95 (2014) 558–564, https://doi.org/10.1038/clpt.2014.24.
[16] D.J. Touw, C. Neef, A.H. Thomson, A.A. Vinks, Cost-effectiveness of therapeutic
drug monitoring, Ther. Drug Monit. 27 (2005) 8, https://doi.org/10.1097/
00007691-200502000-00004.
[17] W. Darko, J.J. Medicis, A. Smith, R. Guharoy, D.F. Lehmann, Mississippi mud no
more: cost-effectiveness of pharmacokinetic dosage adjustment of vancomycin to
prevent nephrotoxicity, Pharmacotherapy 23 (2003) 643–650, https://doi.org/
10.1592/phco.23.5.643.32199.
[18] T.E. Welty, A.K. Copa, Impact of vancomycin therapeutic drug monitoring on
patient care, Ann. Pharmacother. 28 (2016) 1335–1339, https://doi.org/10.1177/
106002809402801201.
[19] M.J. Eadie, The Role of therapeutic drug monitoring in improving the cost
effectiveness of anticonvulsant therapy, Clin. Pharmacokinet. 29 (1995) 29–35,
https://doi.org/10.2165/00003088-199529010-00004.
[20] S.I. Johannessen, D. Battino, D.J. Berry, M. Bialer, G. Krämer, T. Tomson, P.
N. Patsalos, Therapeutic drug monitoring of the newer antiepileptic drugs, Ther.
Drug Monit. 25 (2003) 347–363, https://doi.org/10.1097/00007691-20030600000016.
[21] L.M. Shaw, B. Kaplan, K.L. Brayman, Prospective investigations of
concentration–clinical response for immunosuppressive drugs provide the
scientific basis for therapeutic drug monitoring, Clin. Chem. 44 (1998) 381–387,
https://doi.org/10.1093/clinchem/44.2.381.
[22] A. Johnston, D.W. Holt, Therapeutic drug monitoring of immunosuppressant
drugs, Br. J. Clin. Pharmacol. 47 (1999) 339–350, https://doi.org/10.1046/j.13652125.1999.00911.x.
4. Conclusions
This study demonstrated fluorescence-based bevacizumab quantita
tion at therapeutically relevant ranges in diluted serum and plasma with
porous alumina membranes. Preliminary results indicate that the pro
cess can occur in less than 15 min in a microfluidic device. In the
macroscale and microfluidic vertical flow assays, the fluorescence signal
varies approximately linearly with the concentration of bevacizumab.
Moreover, the low signal with no added mAb confirms the high speci
ficity of the assay. The study shows promise for a microfluidic antibody
analysis platform. A scalable method to prepare the membranes should
further improve reproducibility with lower errors. Additionally, cali
bration of the fluorescence using a standard could improve measure
ment consistency. Scaling the microfluidic platform through roll-to-roll
manufacturing or injection molding would enable implementation of a
low-cost POC device. Such a device may enable healthcare workers to
rapidly measure therapeutic concentrations to ensure patients have
effective mAb levels.
Author contributions
Joshua Berwanger: Conceptualization, Methodology, Validation,
Formal analysis, Investigation, and Writing – original draft and Writing
– review & editing. Melinda Lake: Conceptualization, Methodology,
Validation, Formal analysis, Investigation, and Writing – original draft
and Writing – review & editing. Sanniv Ganguly: Formal analysis,
Investigation, Writing – original draft and Writing – review & editing.
Junyan Yang: Methodology and validation; Christopher Welch: Super
vision, Project administration, Writing – original draft and Writing –
review & editing. Jacqueline Linnes: Conceptualization, Supervision,
Project administration, Writing – review & editing, and Funding
acquisition. Merlin Bruening: Conceptualization, Supervision, Project
administration, Writing – review & editing, and Funding acquisition.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Data availability
Data will be made available on request.
Acknowledgements
We acknowledge funding from the NSF IUCRC Center for Bio
analytical Metrology (IIP 1916601 and IIP 1916691) and the Purdue
University Lillian Gilbreth Postdoctoral Fellowship. The authors thank
Ivan Budyak (Lilly) and Brandy Verhalen (Corteva Agriscience) for their
technical expertise throughout the project. The authors also thank
Muthu Meiyappan, and Christoph Zlabinger (Takeda) and Samin Akbari
(Sartorius) for their technical expertise.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
6
J.D. Berwanger et al.
Talanta 252 (2023) 123842
[36] A.G. Berger, S.M. Restaino, I.M. White, Vertical-flow paper SERS system for
therapeutic drug monitoring of flucytosine in serum, Anal. Chim. Acta 949 (2017)
59–66, https://doi.org/10.1016/j.aca.2016.10.035.
[37] H.Y. Tan, J. Yang, J.C. Linnes, C.J. Welch, M.L. Bruening, Quantitation of
trastuzumab and an antibody to SARS-CoV-2 in minutes using affinity membranes
in 96-well plates, Anal. Chem. 94 (2022) 884–891, https://doi.org/10.1021/acs.
analchem.1c03654.
[38] W. Liu, A.L. Bennett, W. Ning, H.-Y. Tan, J.D. Berwanger, X. Zeng, M.L. Bruening,
Monoclonal antibody capture and analysis using porous membranes containing
immobilized peptide mimotopes, Anal. Chem. 90 (2018) 12161–12167, https://
doi.org/10.1021/acs.analchem.8b03183.
[39] H.M. Geysen, S.J. Rodda, T.J. Mason, A priori delineation of a peptide which
mimics a discontinuous antigenic determinant, Mol. Immunol. 23 (1986) 709–715,
https://doi.org/10.1016/0161-5890(86)90081-7.
[40] C. Liedert, L. Rannaste, A. Kokkonen, O.-H. Huttunen, R. Liedert, J. Hiltunen,
L. Hakalahti, Roll-to-roll manufacturing of integrated immunodetection sensors,
ACS Sens. 5 (2020) 2010–2017, https://doi.org/10.1021/acssensors.0c00404.
[41] H. Becker, C. Gärtner, Polymer microfabrication technologies for microfluidic
systems, Anal. Bioanal. Chem. 390 (2008) 89–111, https://doi.org/10.1007/
s00216-007-1692-2.
[42] T.J. Moehling, D.H. Lee, M.E. Henderson, M.K. McDonald, P.H. Tsang, S. Kaakeh,
E.S. Kim, S.T. Wereley, T.L. Kinzer-Ursem, K.N. Clayton, J.C. Linnes,
A smartphone-based particle diffusometry platform for sub-attomolar detection of
Vibrio cholerae in environmental water, Biosens. Bioelectron. 167 (2020), 112497,
https://doi.org/10.1016/j.bios.2020.112497.
[43] J. Dai, G.L. Baker, M.L. Bruening, Use of porous membranes modified with
polyelectrolyte multilayers as substrates for protein arrays with low nonspecific
adsorption, Anal. Chem. 78 (2006) 135–140, https://doi.org/10.1021/ac0513966.
[44] S. Bhattacharjee, J. Dong, Y. Ma, S. Hovde, J.H. Geiger, G.L. Baker, M.L. Bruening,
Formation of high-capacity protein-adsorbing membranes through simple
adsorption of poly(acrylic acid)-containing films at Low pH, Langmuir 28 (2012)
6885–6892, https://doi.org/10.1021/la300481e.
[45] S. Wijeratne, W. Liu, J. Dong, W. Ning, N.D. Ratnayake, K.D. Walker, M.
L. Bruening, Layer-by-layer deposition with polymers containing nitrilotriacetate, a
convenient route to fabricate metal- and protein-binding films, ACS Appl. Mater.
Interfaces 8 (2016) 10164–10173, https://doi.org/10.1021/acsami.6b00896.
[46] N. Anuraj, S. Bhattacharjee, J.H. Geiger, G.L. Baker, M.L. Bruening, An all-aqueous
route to polymer brush-modified membranes with remarkable permeabilites and
protein capture rates, J. Membr. Sci. 389 (2012) 117–125, https://doi.org/
10.1016/j.memsci.2011.10.022.
[47] J.D. Berwanger, H.Y. Tan, G. Jokhadze, M.L. Bruening, Determination of the serum
concentrations of the monoclonal antibodies bevacizumab, rituximab, and
panitumumab using porous membranes containing immobilized peptide
mimotopes, Anal. Chem. 93 (2021) 7562–7570, https://doi.org/10.1021/acs.
analchem.0c04903.
[48] B.L. Boyanton Jr., K.E. Blick, Stability studies of twenty-four analytes in human
plasma and serum, Gen. Clin. Chem. 48 (2002) 2242–2247.
[49] Bevacizumab ELISA Kit (ab237642) | Abcam, https://www.abcam.com/bevaci
zumab-elisa-kit-ab237642.html (accessed August 8, 2022).
[23] M. Regazzi, J. Golay, M. Molinaro, Monoclonal antibody monitoring: clinically
relevant aspects, a systematic critical review, Ther. Drug Monit. 42 (2020) 45–56,
https://doi.org/10.1097/FTD.0000000000000681.
[24] M. Ovacik, K. Lin, Tutorial on monoclonal antibody pharmacokinetics and its
considerations in early development, Clin. Transl. Sci. 11 (2018) 540–552, https://
doi.org/10.1111/cts.12567.
[25] A. Patton, M.C. Mullenix, S.J. Swanson, E. Koren, An acid dissociation bridging
ELISA for detection of antibodies directed against therapeutic proteins in the
presence of antigen, J. Immunol. Methods 304 (2005) 189–195, https://doi.org/
10.1016/j.jim.2005.06.014.
[26] R. Pai, N. Ma, A.V. Connor, D.M. Danilenko, J.M. Tarrant, D. Salvail, L. Wong, D.
P. Hartley, D. Misner, E. Stefanich, Y. Wu, Y. Chen, H. Wang, D.M. Dambach,
Therapeutic antibody-induced vascular toxicity due to off-target activation of nitric
oxide in cynomolgus monkeys, Toxicol. Sci. 151 (2016) 245–260, https://doi.org/
10.1093/toxsci/kfw037.
[27] How to Run an R&D Systems Luminex® Assay, R&D Systems https://www.rndsys
tems.com/resources/protocols/how-run-rd-systems-luminex-assay (accessed
January 25, 2021).
[28] D.H. Wilson, D.M. Rissin, C.W. Kan, D.R. Fournier, T. Piech, T.G. Campbell, R.
E. Meyer, M.W. Fishburn, C. Cabrera, P.P. Patel, E. Frew, Y. Chen, L. Chang, E.
P. Ferrell, V. von Einem, W. McGuigan, M. Reinhardt, H. Sayer, C. Vielsack, D.
C. Duffy, The Simoa HD-1 Analyzer: a novel fully automated digital immunoassay
analyzer with single-molecule sensitivity and multiplexing, J. Lab. Autom. 21
(2016) 533–547, https://doi.org/10.1177/2211068215589580.
[29] K.N. Han, C.A. Li, G.H. Seong, Microfluidic chips for immunoassays, Annu. Rev.
Anal. Chem. 6 (2013) 119–141, https://doi.org/10.1146/annurev-anchem062012-092616.
[30] A.H.C. Ng, U. Uddayasankar, A.R. Wheeler, Immunoassays in microfluidic systems,
Anal. Bioanal. Chem. 397 (2010) 991–1007, https://doi.org/10.1007/s00216-0103678-8.
[31] M. Zalis, C.L. Jaffe, Routine dot-blot assay of multiple serum samples using a
simple apparatus, J. Immunol. Methods 101 (1987) 261–264, https://doi.org/
10.1016/0022-1759(87)90158-X.
[32] N. Cardona-Castro, P. Agudelo-Flórez, Immunoenzymatic dot-blot test for the
diagnosis of enteric fever caused by Salmonella typhi in an endemic area, Clin.
Microbiol. Infect. 4 (1998) 64–69, https://doi.org/10.1111/j.1469-0691.1998.
tb00357.x.
[33] S. Ramachandran, M. Singhal, K.G. McKenzie, J.L. Osborn, A. Arjyal, S. Dongol, S.
G. Baker, B. Basnyat, J. Farrar, C. Dolecek, G.J. Domingo, P. Yager, B. Lutz, A rapid,
multiplexed, high-throughput flow-through membrane immunoassay: a convenient
alternative to ELISA, Diagnostics 3 (2013) 244–260, https://doi.org/10.3390/
diagnostics3020244.
[34] P. Chen, M. Gates-Hollingsworth, S. Pandit, A. Park, D. Montgomery, D. AuCoin,
J. Gu, F. Zenhausern, Paper-based vertical flow immunoassay (VFI) for detection of
bio-threat pathogens, Talanta 191 (2019) 81–88, https://doi.org/10.1016/j.
talanta.2018.08.043.
[35] Y. Jiao, C. Du, L. Zong, X. Guo, Y. Han, X. Zhang, L. Li, C. Zhang, Q. Ju, J. Liu, H.D. Yu, W. Huang, 3D vertical-flow paper-based device for simultaneous detection
of multiple cancer biomarkers by fluorescent immunoassay, Sensor. Actuator. B
Chem. 306 (2020), 127239, https://doi.org/10.1016/j.snb.2019.127239.
7