Multiple Myeloma Cell Drug Responses Differ in Thermoplastic vs

Oct 3, 2017 - These results highlight the biases that exist in PDMS devices and the importance of material selection in microfluidic device design, es...
0 downloads 6 Views 1MB Size
Subscriber access provided by UNIV NEW ORLEANS

Article

Multiple Myeloma Cell Drug Responses Differ in Thermoplastic vs. PDMS Microfluidic Devices Thomas Moore, Peter M Brodersen, and Edmond W.K. Young Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b02351 • Publication Date (Web): 03 Oct 2017 Downloaded from http://pubs.acs.org on October 8, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Multiple Myeloma Cell Drug Responses Differ in Thermoplastic vs. PDMS Microfluidic Devices

Thomas A. Moore1, Peter Brodersen2 and Edmond W. K. Young1*

1

Department of Mechanical & Industrial Engineering and the Institute of Biomaterials &

Biomedical Engineering, University of Toronto, Toronto, ON, Canada 2

Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto,

ON, Canada

* Corresponding author: Prof. Edmond W.K. Young Department of Mechanical & Industrial Engineering Institute of Biomaterials & Biomedical Engineering University of Toronto, Toronto, ON, Canada E-mail: [email protected] Tel.: +1 (416) 978-1521 Fax: +1 (416) 978-7753

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract Poly(dimethylsiloxane) (PDMS) is a commonly used elastomer for fabricating microfluidic devices, but has previously been shown to absorb hydrophobic molecules. While this has been demonstrated for molecules such as estrogen and Nile Red, absorption of small hydrophobic molecules in PDMS specifically used to treat cancer, and its subsequent impact on cytotoxicity measurements and assays, have not been investigated. This is critical for the development of microfluidic chemosensitivity and resistance assay (CSRA) platforms that have shown potential to help guide clinical therapy selection, and rely on the accuracy of the readout involving interactions between patient-derived cells and cancer drugs. It is thus important to address the issue of drug absorption into device material. We investigated drug absorption into microfluidic devices by treating multiple myeloma (MM) tumour cells with two MM drugs (bortezomib (BTZ) and carfilzomib (CFZ)), in devices fabricated using three different materials (polystyrene (PS), cyclo-olefin polymer (COP), and PDMS). Half-maximal inhibitory concentrations (IC50) were obtained for each drug-material combination, and an increase in IC50 of ~4.3x was observed in PDMS devices compared to both thermoplastic devices. Additionally, each MM drug was exposed to polymer samples, and samples were analyzed using time-of-flight secondary ion mass spectrometry (ToF-SIMS) to characterize adsorption and absorption of the drugs into each material. ToF-SIMS data showed the bias observed in IC50 values found in PDMS devices was directly related to the absorption of drug during dose-response experiments. Specifically, BTZ and CFZ absorption in both PS and COP were all in the range of ~100-300 nm, whereas BTZ and CFZ absorption in PDMS was ~ 5.0 µm and ~3.5µm, respectively. These results highlight the biases that exist in PDMS devices, and the importance of material selection in microfluidic device design, especially in applications involving drug cytotoxicity and hydrophobic molecules.

ACS Paragon Plus Environment

Page 2 of 27

Page 3 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Introduction Chemosensitivity and resistance assays (CSRAs) are assays designed to identify the optimal drug treatment for an individual by testing cells derived from a patient with various chemotherapy options.1 While a more personalized approach to treatment selection via CSRAs is desirable, CSRAs have yet to be recommended for use in clinical practice due to a lack of evidence supporting improved clinical outcomes.2,3 Two significant factors that may adversely affect CSRA performance (i.e., sensitivity and specificity) are: (i) a lack of consideration of the tumour microenvironment in in vitro CSRAs for predicting in vivo drug response, and (ii) limited patient sample availability.4,5 As an example, multiple myeloma (MM) is a bone marrow malignancy caused by an over-proliferation of abnormal plasma B cells in the bone marrow microenvironment (BMME).6 The BMME is highly complex with multiple cell-cell interactions leading to disease progression and drug resistance.7,8 Traditional tools used for testing drugs and studying MM cells in vitro such as well plates3 and Transwell inserts9 do not sufficiently mimic key features including the 3D architecture and cell-cell interactions present in the in vivo microenvironment.4 Furthermore, these conventional CSRA platforms typically require a large number of patient cells (>105 cells per condition) for a single test,4,10 and this poses a major challenge due to the high variability in cell yield across patients at various disease stages.4,5,10 Ultimately, these limitations must be addressed to enhance the predictive potential of CSRAs, improve clinical outcomes, and advance CSRAs toward approval for clinical practice. Recently, microfluidics and microfabrication techniques have been leveraged to advance the development of various platforms designed to improve CSRA performance. These platforms have been developed to: (i) test differences between 2D vs. 3D cultures in terms of drug sensitivity;11 (ii) improve physiological relevance in 3D cultures by demonstrating hematopoietic

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

function;12 and (iii) recapitulate cell-cell communication between bone marrow cells and osteoblasts in a 3D model of ossified tissue to improve primary cell proliferation ex vivo.13 CSRA platforms with high-throughput potential using small cell populations have also been explored, including a CSRA device designed to culture small cell populations leveraging existing well plates for 3D co-cultures,14 a high-throughput CSRA device designed for culturing and screening spheroids using small cell populations from adenocarcinoma patients,15 and an automated high-throughput tool integrating micropumps and valving designed to screen small quantities of prostate cancer cells across 64 culture wells.16 Recently, Young et al. designed an arrayable microfluidic system to study MM, which addressed several key limitations of conventional CSRA platforms.10 The microscale device allowed co-culture and uniform concentrations of cytokines via cell-cell communication between MM tumour cells and stromal cell types,17 and significantly reduced the number of cells required for each experimental condition. The platform was subsequently used to test cancer drug toxicity on MM patientderived cell samples in both monoculture and co-culture configurations to determine which culture arrangement would better predict patient response to a specific cancer drug. The platform demonstrated improved predictive capabilities in co-cultures of patient cell samples (CD138+ and CD138- cells) compared to MM cell monocultures (CD138+ cells only).5 Collectively, these examples demonstrate that microfluidic systems offer more features and functionality compared to typical well plates or other traditional in vitro platforms, have the ability to address previous limitations of CSRA platforms, and have potential to enhance their throughput and predictive capabilities. An important but often overlooked issue of microfluidic devices, particularly those developed for cell biology applications, is that many of them are fabricated using the elastomeric

ACS Paragon Plus Environment

Page 4 of 27

Page 5 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

material poly(dimethylsiloxane) (PDMS). PDMS is an attractive material for the development of cell-based microfluidic devices because fabrication of PDMS molds via soft lithography is simple, fast (i.e., new PDMS prototype devices can be made in ~2-4 days), and produces molds that can be easily bonded to glass substrates for imaging applications.18–20. However, PDMS is also known to possess several disadvantages,18 with absorption (i.e., accumulation of molecules into the bulk volume of a material)21 and adsorption (i.e., accumulation of molecules on the exposed surface of a material) of small hydrophobic molecules being the most relevant issue to CSRA technology development. The absorption of small hydrophobic molecules into PDMS was shown with hydrophobic fluorophores22 and estrogen23, while PDMS adsorption was demonstrated and found to be significant compared to thermoplastic adsorption.24 In particular, Toepke et al.22 was one of the first to demonstrate absorption of hydrophobic molecules in microchannels, showing how the common fluorophore Nile Red absorbed into PDMS channel walls within minutes, accumulated in the channel walls when more Nile Red was added, and remained stably and irreversibly absorbed even after repeated rinsing with an Alconox solution (detergent) and DI water (~200 channel volumes each). Adsorption and absorption issues are critical to CSRA devices because many cancer drugs are small hydrophobic molecules,25–27 and excessive absorption and adsorption of these molecules could lead to significant bias in the interpretation of drug toxicity results on patient samples. Additionally, the therapeutic index (TI), which is defined as the range of effective doses of medication that do not result in adverse events, can be quite narrow in some drugs.4 In the context of MM, bortezomib (BTZ)25 is a commonly used and effective therapeutic agent for treating MM,28 but is known to have a narrow TI.29 For this reason, it is imperative that results obtained from patient cell-based assays are as accurate and unbiased as possible to provide

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

information that will properly guide therapy selection without affecting patient health and safety. To date, however, cancer drug absorption and adsorption within different microfluidic device materials have not been studied in detail to evaluate their impact and bias on drug toxicity. Here we compare the absorption and adsorption of two MM drugs, BTZ and carfilzomib (CFZ),26 both of which are hydrophobic molecules, in a microfluidic device fabricated using three different polymeric materials: polystyrene (PS), cyclo-olefin polymer (COP) and PDMS. Our objective was to determine whether biases are observed in drug sensitivity experiments performed in devices fabricated from different polymers, and whether these biases correlate with absorption and/or adsorption observed through mass spectrometry measurements. Using an MM cell line (RPMI8226), cells were introduced into a microsystem (Fig. 1) and treated with a range of drug concentrations. Dose-response curves were generated to determine half maximal inhibitory concentrations (IC50) for each drug-material combination. Second, surface characterization experiments were performed to study drug absorption and adsorption at the molecular scale using time-of-flight secondary-ion mass spectrometry (ToF-SIMS) to correlate dose response data with the level of absorption/adsorption observed. These differences have not been previously reported in detail, and will serve as a guide for researchers interested in understanding the bias associated with drug response and material selection.

ACS Paragon Plus Environment

Page 6 of 27

Page 7 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 1. (A) PDMS-based microfluidic device consisting of 6 microfluidic culture systems. Scale bar = 10 mm. (B) Plastic-based microfluidic device, consisting of 8 microfluidic culture systems (COP device shown; PS device is identical). Grooved borders (white arrow) at outlet region enable containment of the outlet droplet (droplet not displayed in photo for convenience). Scale bar = 10 mm. (C) Top view shows a cell culture system containing MM cells, including a culture well, inlet and outlet ports, and an outlet reservoir for collecting media and preventing cross-contamination between separate cell culture systems. The side view shows a microsystem illustrating the culture of suspension cells within a well that traps suspension cells during volume replacements of media or staining reagents. (D) Delivery of bortezomib into a microsystem, and absorption of the drug into the PDMS bulk volume of the microsystem. (E) Illustration of two expected dose-response curves, one curve representing polystyrene (PS) with a lower IC50 value, and one curve shifted to the right representing drug response in a PDMS device with a higher IC50 value, potentially caused by the absorption or adsorption of drug into the PDMS material. Materials and Methods Microsystem design and fabrication Each microsystem consisted of a circular culture well 3 mm in diameter and 800 µm in depth for trapping and culturing suspension MM cells. The culture well is connected to inlet and outlet ports by 1-mm wide and 300-µm deep microchannels. The microsystem used here (Fig. 1) was a simplified version of the original design by Young et al.10 The original design contained more complex features such as side chambers and diffusion ports not required for these experiments, and were thus removed for simplicity.

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The PDMS version of this microsystem design was fabricated from SU-8 master molds generated using soft lithography techniques previously described10 (see Supporting Information for detailed protocols). These molds were used to create two unique PDMS layers, a cell culture channel layer and an inlet/outlet port layer. These two layers, along with a 3” x 2” glass slide, were bonded together via oxygen plasma treatment to form a completed device consisting of two 2 x 3 microsystem arrays (Fig. 1A). COP and PS versions of this microsystem design were fabricated using a micromilling machine to create the feature geometry in plastic sheets approximately 1.2 mm thick.30 COP was obtained as pellets (Zeonor #1020R, Zeon Chemicals, Louisville, KY), and then flattened into sheets using an in-house protocol.31 PS was obtained directly as 1.2-mm thick sheets (ST313120, Delta Scientific, Mississauga, ON, Canada). Thermoplastic devices were fabricated using one cell culture channel layer and one inlet/outlet port layer. These two device layers were bonded together to create a 2 x 4 microsystem array (Fig. 1B) using a previously reported solvent bonding technique for COP,32 which involved the use of a cyclohexane/isopropanol mixture, and for PS bonding a (20:80) water/acetone mixture pressed with 4450 N of force for 2 minutes at 70°C in a hydraulic laboratory press (Model 38891NE1000, Carver Inc., Indiana, USA). Grooves measuring 1-mm wide and 0.5-mm deep were fabricated to surround each microsystem (Fig. 1B) in order to receive and facilitate the transport of solvent across a device.33 Two 1.2 mm diameter holes were added to the corners of the inlet/outlet layer (Fig. 1B) to allow solvent to evaporate and escape during bonding. To create outlet reservoirs in COP/PS devices, the outlet region of each microsystem was scored with a scalpel blade to create a grooved border that served to contain the outlet droplet and prevent cross contamination between neighbouring microsystems (Fig. 1B, white arrows). The grooved

ACS Paragon Plus Environment

Page 8 of 27

Page 9 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

border served the same function as the fabricated outlet reservoirs in PDMS devices, but was easier to create in plastics. Cell preparation RPMI8226 (human B lymphocyte plasmacytoma cells, ATCC CCL-155) cell line was obtained from Cedarlane Labs (Burlington, ON, Canada). RPMI8226 (MM) cells were cultured at 37 ºC and 5% CO2 in high-glucose DMEM (#11995-065, Life Technologies) containing 10% fetal bovine serum (FBS), 1% penicillin-streptomycin (P/S) and 1% HEPES buffer. The cells were seeded in T-25 culture flasks at a density of 7.5 x 105 cells/mL, and were passaged every 2 to 3 days. RPMI8226 cells were used in experiments up to a maximum of 15 passages. Please see Supporting Information for details on other biological reagents used. Device preparation and cell seeding Each microsystem within the device was disinfected with 70% ethanol for 10 min, and flushed with 3 volume replacements (VRs) of 1x Dulbecco's Phosphate-Buffered Saline (DPBS), where one VR is equal to approximately 7.5 µL in the center channel. 5 VRs of DMEM media were flushed through each microsystem, and care was taken to ensure a final DMEM media volume of ~35 µL in the outlet reservoir (same for all devices). After filling, the device was placed in a Nunc OmniTray with sterile Kimwipes pre-soaked with approximately 4 mL of 1x DPBS each. The OmniTray was placed in a bioassay dish with an additional 5 mL of DPBS added to a 50 mL conical tube lid, and this entire assembly was placed in an incubator warmed to 37°C and maintaining a 5% CO2 environment. For cell experiments, 7.5 µL of an RPMI8226 cell suspension was seeded in the center well of each microsystem at a density of ~2.0 x 106

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 27

cells/mL, resulting in ~5000 MM cells settling in the center well, and leaving ~42 µL at the device outlet. Dose response experiments MM cells were seeded and allowed to settle for 24 h before drug treatment. Dose-response experiments were performed using concentrations of CFZ or BTZ ranging from 0.1 nM to 1000 nM. A drug was thawed 5 min before serial dilutions of 1000, 100, 10, 1 and 0.1 nM were completed in 500-µL of cell media within microcentrifuge tubes. After serial dilution, each tube was placed in a 37°C water bath. After warming, drug treatments were applied with a single application of 20 µL per concentration into a microsystem. Each outlet was aspirated, and then 40 µL of each concentration was added to each microsystem. Treatments were quickly added, in order of increasing concentration, and the pipette tip was replaced after the first round of 20 µL doses to avoid contamination between microsystems. Following 24 h of drug treatment, Live/Dead reagents (#L3224, ThermoScientific) were diluted in 37°C DMEM serum-reduced media (#11058-021, Life Technologies), warmed to 37°C, and dispensed into microchambers in 3 sequential VRs followed by aspiration of the outlet port. Devices containing Live/Dead reagents were incubated at 37°C for 45 min, at which time each microsystem was rinsed with 5 VRs of DPBS prior to imaging. Two technical replicates (i.e., two microsystems) were used per condition in all experiments. Fluorescence imaging and analysis All fluorescent images for dose-response experiments were captured at room temperature using an Olympus IX83 inverted fluorescent microscope (Olympus Canada Inc.) with an UPlan 4x objective (NA = 0.13, through air; Olympus Canada Inc.) and an Orca Flash 4.0 CCD camera

ACS Paragon Plus Environment

Page 11 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

(C11440-22CU, Hamamatsu, Japan). One image was captured per microsystem for each fluorescence channel (LIVE = green, DEAD = red) using MetaMorph software (version 7.10.0.119, Molecular Devices LLC, CA, USA). Image analysis was performed using ImageJ and customized software for image database management and analysis (Je’Xperiment, or JeX: https://github.com/jaywarrick/JEX). The software calculated the ratio of live cells to total cells across the range of drug concentrations tested. Statistical Analysis For all dose-response experiments, experimental conditions were tested in duplicate per experiment, and all experiments were repeated three times (n = 3). The mean IC50 for each drugmaterial combination was determined using all IC50 values obtained for a given polymer, and were statistically analyzed by ANOVA followed by Tukey’s post hoc procedure to determine which materials were significantly different. Mass spectrometry A small test component (Fig. S1A) was fabricated to study drug interaction with each material studied. The test component was a 10-mm x 10-mm square chip with 4 square pedestals approximately 2-mm x 2-mm. For PS and COP, each component was micro-milled to create these pedestals. For PDMS, 2-mm x 2-mm squares of 500-µm thickness were cut to size, oxygen-plasma treated, and bonded to 1-mm thick by 10-mm x 10-mm glass substrates. BTZ and CFZ were diluted in DMEM media to 10 µM. 2.5 µL droplets of either the drug-media mixture or the media only (negative control) were placed on each of the four pedestals per chip (Fig. S1B), and then incubated at 37°C and 5% CO2 for 24 h. Droplets were aspirated and stored in vacuum chambers prior to ToF-SIMS experiments. All sample analyses were performed in a ToF-SIMS V instrument, and 500-µm x 500-µm regions of each polymer were analyzed. All data

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

captured using ToF-SIMS were analyzed using SurfaceLab 6 software (IONTOF) to generate 2D profiles of drug and polymer signals, and 3D volume plots of drug signal intensity throughout each polymer. To generate these plots, peak lists of signals were generated using the appropriate representative mass fragments for each drug and polymeric material from baseline analyses, and these peak lists were filtered to eliminate any potential interfering signals and potential contaminants. See Supporting Information for further details on mass spectrometry analyses.

ACS Paragon Plus Environment

Page 12 of 27

Page 13 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Results We studied the absorption and adsorption of two specific MM drugs on three different polymer materials using two different approaches. First, dose-response experiments were carried out to compare IC50 values obtained for each drug-material combination. Second, we exposed each material to drug diluted in cell culture media, and after 24-h exposure, we used ToF-SIMS to detect and observe absorption or adsorption for each drug-material combination. MM-cell dose responses to drugs Based on previous studies, we expected some bias in IC50 values obtained from PDMS devices due to potential absorption of hydrophobic drug molecules.22–24 As mentioned, MM cells were first cultured in PS, COP and PDMS devices and treated with BTZ for 24 h. Results showed that the mean IC50 values obtained for MM cells treated with BTZ in PS, COP and PDMS devices were 8.3 nM, 9.0 nM and 37.0 nM, respectively (Fig. 2A-C). These values were obtained by averaging three independent experiments per material. When BTZ dose-response curves for the three different materials were overlaid on each other, a noticeable shift was evident in the sigmoidal curve for the PDMS material (Fig. 2D). Mean IC50 values obtained for MM cells treated with CFZ in PS, COP and PDMS devices were 27.2 nM, 19.7 nM and 102.2 nM respectively (Fig. 2E-G). Again, when CFZ dose-response curves for the three different materials were overlaid on each other, the sigmoidal curve for the PDMS material noticeably shifted to a higher IC50 (Fig. 2H). Overall, IC50 values obtained from PDMS devices were ~4.3x greater than IC50 values from thermoplastic devices.

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 27

Figure 2. Dose response curves for different drug-material combinations. (A-C) Graphs for each BTZ-material combination illustrating drug response of each independent experiment (grey lines), along with the mean response (points), mean dose-response curve (coloured line curves) and fitted IC50 values (A) PS IC50 = 8.3nM, (B) COP IC50 = 9.0nM, (C) PDMS IC50 = 37.0nM. (D) Overlay of average dose-response curves for all 3 polymer materials, generated based on BTZ treatment of MM cells. (E-G) Graphs for each CFZ-material combination illustrating drug response of each independent experiment (grey lines), along with the mean response (points), mean dose-response curve (coloured line curves) and fitted IC50 values (E) PS IC50 = 27.2nM, (F) COP IC50 = 19.7nM, (G) PDMS IC50 = 102.2nM. (H) Overlay of average dose-response curves for all 3 polymers materials, generated based on CFZ treatment of MM cells. Error bars represent standard error (SE) for three independent experiments (n = 3) for each material and drug tested within it. Standard errors between the two technical replicates per condition were consistently smaller (~50% or less) than the standard errors between independent experiments.

ACS Paragon Plus Environment

Page 15 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 3 summarizes mean IC50 values for all the drug-material combinations. Mean IC50 values determined for thermoplastic devices were found to be statistically equivalent for both drugs. Mean IC50 values determined for PDMS devices were found to be significantly different for both drugs, supporting our hypothesis that devices fabricated from PDMS may lead to a biased IC50 values when compared with equivalent thermoplastic devices.

Figure 3. Mean IC50 values determined for all drug-material combinations. Error bars represent standard error (SE) for three independent experiments (n = 3) for each material and drug tested within it. Statistical significance (*p < 0.0001) was determined by ANOVA and by Tukey’s post-hoc procedure. Mass spectrometry To examine the origin of the bias in IC50 values determined from dose-response experiments, a test component (Fig. S1A) was designed and exposure experiments performed in which each polymeric material was exposed to various concentrations of BTZ-media and CFZ-media mixtures (Fig. S1B), and these test components were then analyzed using ToF-SIMS. Control samples were used to isolate and identify mass fragments for each molecule studied (Fig. 4) that are representative of each drug, cell media component and polymeric material being studied.

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 27

Figure 4. The chemical structure of each molecule analyzed using ToF-SIMS. (A) Bortezomib (BTZ), (B) carfilzomib (CFZ), (C) polystyrene (PS), (D) cyclo-olefin polymer (COP), and (E) poly(dimethylsiloxane) (PDMS). Using the mass fragment lists specified in Tables S1 and S2, 2D profiles were generated demonstrating changing polymer and drug intensity as an ion beam etched into the test component. We expected drug to be present on the surface of all three polymers (i.e., adsorption), but anticipated that at increasing depth (i.e., absorption), drug would only be present in PDMS and not in thermoplastics. Each 2D profile consisted of a black line representing the drug signal, and either a green (PS), red (COP) or blue (PDMS) line representing the polymer signal (Fig. 5A-C and 6A-C). The drug signal decreased substantially for both drugs in PS and COP within several hundred nanometers of the surface, and corresponded with a sharp increase in thermoplastic signal, thereby indicating the location where drug penetration into the material was significantly abrogated. In contrast, BTZ penetrated ~7 µm (Fig. 5C) and CFZ penetrated ~3.5 µm (Fig. 6C) into PDMS, which was at least an order of magnitude higher than in PS and COP. 3D volume plots of the analyzed polymer volumes were also created to represent how the drug signal changed through the polymer, where increased red pixel density indicates regions of increased drug concentration. Data presented in 3D volume plots for both BTZ (Fig. 5D-F) and CFZ (Fig. 6D-F) showed that drug was primarily present on the surface of both PS and COP (~

ACS Paragon Plus Environment

Page 17 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

100-300 nm), while drug had absorbed deeper into PDMS (BTZ penetration ~ 5000 nm, CFZ > 3500 nm). Last we present a 2D profile plot overlaying BTZ signals (Fig. 5G) and CFZ signals (Fig. 6G) for each material, highlighting the increase in drug signal intensity observed in PDMS samples compared with drug signal observed in thermoplastic samples.

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 27

Figure 5. ToF-SIMS data for interactions between BTZ and all polymers studied. 2D profile plots showing how polymer and drug signals change with increasing depth for: (A) COP, (B) PS and (C) PDMS. 3D volume plots of sputter craters, with top plane of each image representing z = 0 nm correlating with the 2D profile plots. Each 3D plot illustrates drug intensity throughout the sputter crater for each polymer (D) COP, (E) PS, and (F) PDMS. (G) A 2D profile plot overlaying BTZ signal intensity measured in each material, showing how BTZ signal intensity changes in each material with increasing depth.

ACS Paragon Plus Environment

Page 19 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 6. ToF-SIMS data for interactions between CFZ and all polymers studied. 2D profile plots showing how polymer and drug signals change with increasing depth for: (A) COP, (B) PS and (C) PDMS. 3D volume plots of sputter crater, with top plane of each image representing z = 0 nm correlating with the 2D profile plots. Each 3D plot illustrates drug intensity throughout the sputter crater for each polymer (D) COP, (E) PS, and (F) PDMS. (G) A 2D profile plot overlaying CFZ signal intensity measured in each material, showing how CFZ signal intensity changes in each material with increasing depth.

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 27

Discussion By combining drug cytotoxicity dose-response experiments and ToF-SIMS, we detected a measurable bias in IC50 concentrations measured in PDMS versus thermoplastic devices, and determined that the nature of this bias stemmed, at least in part, from significant drug absorption observed in PDMS. We designed and fabricated devices in two thermoplastics (PS and COP) and one elastomeric material (PDMS), and performed drug cytotoxicity assays on RPMI8226 MM cells treated with two different MM drugs (BTZ and CFZ), and obtained IC50 values for each drug-material combination. We hypothesized that some drug absorption would take place in PDMS, and lead to biased IC50 values relative to IC50 values obtained in PS and COP devices for both drugs, and we confirmed this by observing ~4.3x increase in IC50 values obtained in PDMS devices relative to COP and PS devices. COP and PS devices produced IC50 values that were relatively similar to one another, with a larger difference observed in CFZ than BTZ. To determine the nature of this bias, ToF-SIMS was performed for each drug-material combination to quantify the level of drug absorption and adsorption. We suspected that the IC50 bias in PDMS was related to drug adsorption and absorption, but was unaware of any previous data showing evidence of the levels of adsorption and absorption in the context of cancer drugs in different microfluidic device materials. Our results provided evidence that significantly more drug was absorbed to greater depths in PDMS-fabricated devices (~5 µm) compared to thermoplasticfabricated devices (~0.2 µm). BTZ was observed to absorb into PDMS at greater depths than CFZ, which may be attributed to the molecular weight of CFZ (720 g/mol) being ~2x that of BTZ (384 g/mol). COP and PS showed similar results, with drug-associated molecules mainly adsorbing on the material surface (with slightly more adsorption in COP than PS), and little to no absorption of drug-associated molecules into the bulk material. Thus, there is a correlation

ACS Paragon Plus Environment

Page 21 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

between absorption of small hydrophobic molecules in PDMS and the 4.3x increase in IC50 values in PDMS devices compared to thermoplastic devices. Regarding stability of this absorption, we note that Toepke22 previously demonstrated highly stable and irreversible absorption of the hydrophobic fluorophore Nile Red, even after repeated washing steps, but showed that reversible absorption is possible in low pH conditions. Since our culture media was neither at a low pH nor repeatedly rinsed away during the assay, we believe the drug was stably absorbed and did not leach back out into the media at a later time. This was further supported by the marked increase in the IC50 values in PDMS devices, as previously stated. These results illustrate that PDMS-based CSRA devices produce significantly biased results compared to their thermoplastic counterparts, and their use may present an increased barrier to adoption in a clinical setting since many cancer drugs are hydrophobic and therefore susceptible to absorption into PDMS.27 For low TI drugs such as BTZ,29 use of PDMS-based CSRA devices to determine appropriate drug dosage and concentration may lead to overestimation of prescribed doses, with potential to dangerously fall in the range of toxic doses. In addition, PDMS-based CSRA devices would be prone to producing false-negative predictions: a patient-derived cell sample may respond to a lower drug concentration in a thermoplastic device (and be deemed “responsive” to the drug), whereas the same sample may not show any response to treatment in a PDMS-based device at that concentration due to drug absorption into PDMS causing a shift toward higher IC50 values. While our mass spectrometry analyses provided strong evidence of significant differences in drug absorption between PDMS and thermoplastic materials, other factors may also be involved in the overall differences in cell response to drug treatment. It is well documented that PDMS leaches uncrosslinked oligomers, while thermoplastics can also leach potentially

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 27

bioactive chemicals, such as plasticizers.18 We mitigated leaching from PDMS by performing Soxhlet extraction of all our PDMS layers (see Supporting Information), while the thermoplastics used in this study were not treated additionally. Furthermore, it is also well known that PDMS has higher oxygen permeability compared to thermoplastics, and this is an advantage for cell culture applications because of the need for proper gas exchange. However, the fact that our live fraction was consistently ~80% in all experiments – independent of device material – suggested that cell viability was not impacted by any differences in gas exchange between device types. Besides small hydrophobic molecules, proteins may also be adsorbed on inner microchannel walls. Most current microfluidic devices are coated with blocking proteins (e.g., bovine serum albumin, BSA) as a method to reduce further adsorption (most often non-specific adsorption).34,35 However, because protein-based drugs (e.g., therapeutic antibodies, recombinant human proteins) have become increasingly prevalent for therapy36 understanding protein adsorption within microfluidic systems may also be important for the ongoing development of microfluidic cell-based assays, particularly if protein-based drugs are studied. The research work presented here was facilitated by recent advances in microfabrication with thermoplastic materials including micromilling30 and solvent bonding.32,33 In our lab, turnaround time for micromilling a device is currently several hours, which is faster than previously reported methods of prototyping using soft lithography for PDMS device fabrication, and hot embossing processes for thermoplastic devices.19 In addition, our lab recently developed solvent bonding with high bond strength and no leakage,32 a method that was used in this study to ensure reliable, leak-free culture of MM cells in the devices over several days, and reliable device performance during drug dosing and volume replacement operations. The thermoplastic

ACS Paragon Plus Environment

Page 23 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

versions of this device can now be fabricated in larger batches by combining micromilling and hot embossing techniques31 to further accelerate lab experimentation. Since biologists have long been using in vitro lab culture-ware made of thermoplastic materials, particularly PS,37 in vitro microfluidic devices made of thermoplastic materials rather than PDMS may be more likely to achieve adoption,18 especially given the results of the current study.

Conclusion We compared the response of MM cells cultured in microsystem arrays fabricated using three different polymers to two small hydrophobic MM cancer drugs. We found an increase of ~4.3x between IC50 values obtained from PDMS-derived devices compared to thermoplastic-derived devices. To explain this bias, we used ToF-SIMS to study drug-plastic interactions, and revealed significant amounts of drug absorption in PDMS, compared to PS and COP, both of which presented some adsorption and negligible absorption. This data provides evidence that performing CSRAs in PDMS-derived devices will lead to biased results, while using thermoplastic-derived devices would mitigate the bias caused by drug absorption into the device material. Recent advances in solvent bonding and micromilling for thermoplastic microfluidic devices facilitated the comparative experiments conducted in this study, and may play an increasingly important role in accelerating adoption of thermoplastic microfluidic devices and the use of microfluidic CSRA devices in biomedical applications.

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 27

Acknowledgments

The authors would like to thank Dr. Craig Simmons for his valuable input regarding polymer testing and ToF-SIMS. We would also like to acknowledge the Ontario Center for the Characterization of Advanced Materials (OCCAM). We acknowledge financial support from the Cancer Research Society and the Natural Sciences and Engineering Research Council of Canada (NSERC) to EY, from the NSERC CREATE Training Program in Organ-on-a-Chip Engineering & Entrepreneurship (TOeP) to TM, and the Canadian Foundation for Innovation, the Ontario Research Fund and the University of Toronto - OCCAM funding partners.

ACS Paragon Plus Environment

Page 25 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

References (1)

Unger, F. T.; Witte, I.; David, K. A. Cell. Mol. Life Sci. 2015, 72, 729–757.

(2)

Burstein, H. J.; Mangu, P. B.; Somerfield, M. R.; Schrag, D.; Samson, D.; Holt, L.; Zelman, D.; Ajani, J. A.; American Society of Clinical Oncology. J Clin Oncol 2011, 29, 3328–3330.

(3)

Samson, D. J.; Seidenfeld, J.; Ziegler, K.; Aronson, N. J Clin Oncol 2004, 22, 3618–3630.

(4)

Morgan, M. M.; Johnson, B. P.; Livingston, M. K.; Schuler, L. A.; Alarid, E. T.; Sung, K. E.; Beebe, D. J. Pharmacol. Ther. 2016, 165, 79–92.

(5)

Pak, C.; Callander, N. S.; Young, E. W. K.; Titz, B.; Kim, K.; Saha, S.; Chng, K.; Asimakopoulos, F.; Beebe, D. J.; Miyamoto, S. Integr. Biol. 2015, 7, 643–654.

(6)

Laubach, J.; Richardson, P.; Anderson, K. Annu. Rev. Med. 2011, 62, 249–264.

(7)

Hideshima, T.; Mitsiades, C.; Tonon, G.; Richardson, P.; Anderson, K. Nat. Rev. Cancer 2007, 7, 585–598.

(8)

Mimura, N.; Hideshima, T.; Anderson, K. C. Exp. Hematol. 2015, 43, 732–741.

(9)

Markovina, S.; Callander, N. S.; O’Connor, S. L.; Xu, G.; Shi, Y.; Leith, C. P.; Kim, K.; Trivedi, P.; Kim, J.; Hematti, P.; Miyamoto, S. Mol. Cancer 2010, 9, 176.

(10)

Young, E. W. K.; Pak, C.; Kahl, B. S.; Yang, D. T.; Callander, N. S.; Miyamoto, S.; Beebe, D. J. Blood 2012, 119, E76–E85.

(11)

Bruce, A.; Evans, R.; Mezan, R.; Shi, L.; Moses, B. S.; Martin, K. H.; Gibson, L. F.; Yang, Y. PLoS One 2015, 10, 1–16.

(12)

Torisawa, Y.; Spina, C. S.; Mammoto, T.; Mammoto, A.; Weaver, J. C.; Tat, T.; Collins, J. J.; Ingber, D. E. Nat. Methods 2014, 11, 663–669.

(13)

Zhang, W.; Lee, W. Y.; Siegel, D. S.; Tolias, P.; Zilberberg, J. Tissue Eng. Part C. Methods 2014, 20, 663–670.

(14)

Silva, A.; Jacobson, T.; Meads, M.; Distler, A.; Shain, K. J. Vis. Exp. 2015, 101, 1–12.

(15)

Ruppen, J.; Wildhaber, F. D.; Strub, C.; Hall, S. R. R.; Schmid, R. A.; De, T. G.; Guenat, O. T. Lab Chip 2015, 15, 3076–3085.

(16)

An, D.; Kim, K.; Kim, J. Biomol. Ther. 2014, 22, 355–362.

(17)

Moore, T. A.; Young, E. W. K. Biomicrofluidics 2016, 10, 044105, 1-17.

(18)

Berthier, E.; Young, E. W. K.; Beebe, D. Lab Chip 2012, 12, 1224–1237.

(19)

Young, E. W. K.; Berthier, E.; Guckenberger, D. J.; Sackmann, E.; Lamers, C.; Meyvantsson, I.; Huttenlocher, A.; Beebe, D. J. Anal Chem 2011, 83, 1408–1417.

(20)

Duffy, D.; McDonald, J.; Schueller, O.; Whitesides, G. Anal. Chem. 1998, 70, 4974–4984.

(21)

Lee, J. N.; Park, C.; Whitesides, G. M. Anal. Chem. 2003, 75, 6544–6554.

ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(22)

Toepke, M. W.; Beebe, D. J. Lab Chip 2006, 6, 1484–1486.

(23)

Regehr, K. J.; Domenech, M.; Koepsel, J. T.; Carver, K. C.; Ellison-Zelski, S. J.; Murphy, W. L.; Schuler, L. A.; Alarid, E. T.; Beebe, D. J. Lab Chip 2009, 9, 2132–2139.

(24)

van Midwoud, P. M.; Janse, A.; Merema, M. T.; Groothuis, G. M. M.; Verpoorte, E. Anal Chem 2012, 84, 3938–3944.

(25)

Pu, G.; Ren, C.; Li, D.; Wang, L.; Sun, J. RSC Adv. 2014, 4, 50145–50147.

(26)

Ao, L.; Reichel, D.; Hu, D.; Jeong, H.; Kim, K. B.; Bae, Y.; Lee, W. J. Pharmacol. Exp. Ther. 2015, 355, 168–173.

(27)

Kalepu, S.; Nekkanti, V. Acta Pharm. Sin. B 2015, 5, 442–453.

(28)

Leung-Hagesteijn, C.; Erdmann, N.; Cheung, G.; Keats, J. J.; Stewart, A. K.; Reece, D. E.; Chung, K. C.; Tiedemann, R. E. Cancer Cell 2013, 24, 289–304.

(29)

Piperdi, B.; Ling, Y.-H.; Liebes, L.; Muggia, F.; Perez-Soler, R. Mol. Cancer Ther. 2011, 10, 2029–2030.

(30)

Guckenberger, D. J.; de Groot, T.; Wan, A. M.-D.; Beebe, D.; Young, E. Lab Chip 2015, 15, 2364–2378.

(31)

Konstantinou, D.; Shirazi, A.; Sadri, A.; Young, E. W. K. Sensors Actuators, B Chem. 2016, 234, 209–221.

(32)

Wan, A. M. D.; Moore, T. A.; Young, E. W. K. J. Vis. Exp. 2017, No. 119, 1–9.

(33)

Wan, A. M. D.; Sadri, A.; Young, E. W. K. Lab Chip 2015, 15, 3785–3792.

(34)

Schrott, W.; Slouka, Z.; Červenka, P.; Ston, J.; Nebyla, M.; Přibyl, M.; Šnita, D. Biomicrofluidics 2009, 3, 44101-1–15.

(35)

Kim, D.; Herr, A. E. Biomicrofluidics 2013, 7, 41501-1–47.

(36)

Tomlinson, I. M. Nat. Biotechnol. 2004, 22, 521–522.

(37)

Sackmann, E. K.; Fulton, A. L.; Beebe, D. J. Nature 2014, 507, 181–189.

ACS Paragon Plus Environment

Page 26 of 27

Page 27 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Table of Contents (TOC) Figure

ACS Paragon Plus Environment