Enrichment and Recovery of Mammalian Cells From Contaminated

5 days ago - This paper describes a density-based method for removing contaminants, including microorganisms and nonviable cells, from mammalian cell ...
2 downloads 10 Views 785KB Size
Subscriber access provided by RMIT University Library

Article

Enrichment and Recovery of Mammalian Cells From Contaminated Cultures Using Aqueous Two-phase Systems Christopher J Luby, Benjamin P. Coughlin, and Charles R. Mace Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b04352 • Publication Date (Web): 29 Dec 2017 Downloaded from http://pubs.acs.org on January 1, 2018

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 29 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

Enrichment and Recovery of Mammalian Cells From Contaminated Cultures Using Aqueous Two-phase Systems

Christopher J. Luby, Benjamin P. Coughlin, and Charles R. Mace*

Department of Chemistry, Tufts University, 62 Talbot Avenue, Medford, MA 02155 United States

*Corresponding author: [email protected]

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 2 of 29

Abstract This paper describes a density-based method for removing contaminants, including microorganisms and nonviable cells, from mammalian cell cultures using an aqueous two-phase system (ATPS). The properties of a 7% w/w polyethylene glycol (PEG) – 11% w/w Ficoll ATPS can be tuned to prepare a biocompatible system that removes contaminants with little to no adverse affects on the viability or growth of the cultured cells after treatment. This system can be used to enrich cell culture populations for viable cells and to reduce the number of microorganism contaminants in a culture, which increases the chances of subsequent antibiotic treatments being successful. We test the effectiveness of our method in model contaminated cultures of both adherent (HeLa) and suspension (HL-60 II) mammalian cells contaminated with bacteria (E. coli) and yeast (S. cerevisiae). An average of 70.2 ± 4.6% of HeLa cells added to the system are subsequently recovered and 55.9 ± 2.1% of HL-60 II cells are recovered. After sedimenting to the interface of the ATPS these cells have an average viability of 98.0 ± 0.2% and 95.3 ± 2.2% respectively. By removing unwanted cells, desired cell populations can be recovered and cultures that would otherwise need to be discarded can continue to be used.

ACS Paragon Plus Environment

2

Page 3 of 29 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 The in vitro culture of mammalian cells is a crucial part of biomedical research. Aseptic technique is necessary to ensure that only the cells of interest are growing in the culture. Contamination, however, is still a relatively common occurrence in cell culture laboratories. Typical contaminants include bacteria, yeast, and mold.1 These microorganisms can affect the way cells behave in culture (e.g., change morphology, growth, and viability)2 and alter the culture conditions (e.g., through changing pH or by competing for resources in the medium).1,3 Nonviable cells (i.e., cells present in culture that lack membrane integrity as demonstrated by the inability to exclude dyes), cell fragments, and debris can also be considered contaminants because their presence can bias seeding densities or skew the results of functional assays.4,5 Cultures that contain a large fraction of nonviable cells (e.g., cultures that are overgrown) may not be usable unless the viable cells can be isolated from the dead cells.5 Antibiotics, such as penicillin and streptomycin, are used to treat contamination or prevent it prophylactically. However, these compounds may adversely interfere with normal processes in cultured cells,6 exacerbate contamination issues by masking poor aseptic technique, or promote antibiotic resistance and the development of resistant microorganisms.7 Ideally the use of antibiotics in cell culture should be eliminated or minimized. In order to prevent the spread of microorganisms to other cultures and avoid the negative effects of antibiotic use, contaminated cultures are typically destroyed with bleach. Although this approach is an effective way to contain and eliminate contaminants, it also results in the disposal of the cultured cells. This is, at a minimum, a waste of time and effort for the researchers and money for the lab.8 Moreover, if the cells are rare or valuable (e.g., primary or engineered), their disposal may not be acceptable. The capability to decontaminate and recover the desired population of cells would be far more appropriate in these cases. We developed a method to separate cultured mammalian cells from contaminants based on density using aqueous two-phase systems (ATPS). This method allows recovered cells to be

ACS Paragon Plus Environment

3

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 4 of 29

cultured after enrichment and can eliminate or minimize the need for antibiotic treatments. ATPS form when solutions of certain water-soluble polymers are mixed above a threshold concentration and molecular weight. These systems form two thermodynamically stable phases with different physical properties (e.g., densities) separated by a molecularly-sharp liquid-liquid interface.9 More generally, aqueous systems composed of multiple phases are known as aqueous multiphase systems (either termed MuPS or AMPS).10–12 Because water is the common solvent, ATPS are particularly well suited for separations of biomolecules and cells when compared to immiscible systems prepared using organic solvents. Cells and small molecules can be partitioned into either phase due to properties of both the species themselves or the system including hydrophobicity, molecule/particle size, surface charge, binding affinities for functionalized polymers, or the interfacial tension of the system.13–16 Cells can also be separated based on differences in buoyant density between cells and the phases of an AMPS. Unlike approaches that use partitioning, in which cells remain suspended within a bulk phase (i.e., a volume), cells will become isolated and concentrated at the liquid-liquid and liquidcontainer interfaces (i.e., an area) by sedimentation upon application of a relative centrifugal force when performing separations using ATPS or AMPS as step-gradients in density.17,18 Other applications that capitalize on the biocompatibility of ATPS include patterning cells or biomolecules in order to perform functional assays.19,20 Microorganisms that may contaminate cell cultures such as bacteria (e.g., E. coli)21 and yeast22 have higher buoyant densities than cultured mammalian epithelial cells (e.g., HeLa, human lung epithelial, and mouse lymphocytic leukemia.)23,24 Under these conditions, contaminated cultures added to an ATPS will separate by density upon centrifugation with the mammalian cells sedimenting to the liquid-liquid interface and the denser contaminates forming a pellet at the bottom of the container. After separation, the desired cells can be recovered from the interface by pipette, washed, and transferred to fresh medium for further treatment or culture (Figure 1). We prepared an ATPS from polyethylene glycol (PEG) and Ficoll (polysucrose)

ACS Paragon Plus Environment

4

Page 5 of 29 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

whose phases formed a step in density that was able to separate lower density mammalian cells from higher density microorganisms in a test system of HeLa cells and E. coli. We also tested the system for broader applicability using HL-60 II cells25 (a human promyelocytic leukemia suspension cell line) as a second mammalian cell type and yeast (S. cerevisiae) as a second microorganism contaminant. This system of biocompatible polymers allowed for the cells to be separated and recovered without a loss of viability. Furthermore, recovered cells had comparable viability and growth rates to control populations that were not introduced to ATPS. By exploiting an intrinsic physical difference between mammalian cells and contaminant microorganisms, the desired cell population can be enriched for, thereby minimizing the need for treatments with antimicrobial compounds. Density-based separation has previously been demonstrated as an effective method for removing contaminants from cell cultures (e.g., removal of contamination by mold,26 cell debris, and microorganisms by centrifugation through Percoll solution).27 Separating cells from contaminants in ATPS based on density is a simple, gentle, and nondestructive process.17,18,28,29 Using ATPS is preferable to continuous density gradients (e.g., Percoll): (i) The sharp step in density between phases in ATPS allow cells with slightly different densities to become concentrated at an interface rather than in a disperse band.17,27 (ii) The interface allows for desired cells and contaminants to be separated from each other by significant phase volume. This separation in space makes it easier to remove cells by pipette without also recovering the contaminants. Density-based separation of cell populations allows non-viable and lysed cells to be similarly separated and removed, enriching the cell population for viable cells with similar densities. Gonzalez-Gonzalez et al. demonstrated that nonviable cells and cell debris can be separated from viable cells by partitioning to separate phases of an ATPS.4 This separation can be achieved in a more controllable manner by characterizing and tuning the physical properties of the phases and considering the interface as region of the system where cells can be located, distinct from the bulk phases. Our ATPS-based

ACS Paragon Plus Environment

5

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 6 of 29

cell recovery method allows for the preservation of potentially valuable cells that would otherwise be discarded.

Experimental Section Model Systems for Contaminated Cultures In order to confirm separation and determine the efficiency of our method, we developed model contaminated cultures using GFP-expressing HeLa cells30,31,32 and mPlum-expressing E. coli. The fluorophores were chosen to avoid overlap in emission, which allowed us to differentiate between the two types of cells by fluorescence. We created model contaminated cultures by adding fluorescent microorganisms to cell cultures. After treating these cultures with our ATPS, we qualitatively inspected the cultures for the presence of red E. coli using confocal microscopy. For a quantitative measure of separation efficiency and number of bacteria present after separation we used flow cytometry and colony formation assays. Through confocal microscopy, we observed the morphology of the HeLa cells to ensure that it was unchanged by the separation process. These images also provided information regarding the interaction of the microorganisms with the cells. To determine if our specific ATPS formulation could also be applied more broadly to cells and contaminants other than our HeLa/E. coli model system, we used a suspension cell line (HL-60 II) with the mPlum E. coli. We also tested separations of mCitrine-expressing yeast in suspensions of HeLa cells and HL-60 cells. The separation efficiency of these systems was quantified by flow cytometry. Design of Aqueous Two-Phase Systems The inherent difference in density between mammalian cells and microorganisms allow them to be separated by centrifugation when placed in a density gradient or ATPS. ATPS provide a sharp step in density at the liquid-liquid interface, which can be used to isolate cells of interest.10 HeLa cells have a buoyant density of 1.04–1.07 g/mL,23,27 while yeast and E. coli have reported buoyant densities of 1.08–1.11 g/mL and 1.12–1.18 g/mL, respectively.22,33 We confirmed these

ACS Paragon Plus Environment

6

Page 7 of 29 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

density ranges experimentally using sink/float assays (Figure S1). Based on these values, we desired an ATPS with a bottom phase that is denser than mammalian cells but less dense than microorganisms. This means that, upon centrifugation, the microorganisms can sediment through the entire system and form a pellet at the bottom of the container. The desired mammalian cells, however, localize to the liquid-liquid interface allowing them to be subsequently recovered and cultured. PEG-Ficoll is among the more than 100 known combinations of polymers that form ATPS.10 Both polymers are biocompatible and are widely used in biological applications including twophase systems34 and gradient-based separations.35 The properties of this system and the polymers comprising it have been previously studied and well-characterized.36–38 The densities, osmolalities, and pH of the phases, as well as the interfacial tension between phases can be tuned through control of polymer and buffer concentrations to provide the desired separation while maintaining biocompatibility. We developed an ATPS with a top phase density of 1.0433 ± 0.0002 g/mL and a bottom phase density of 1.0755 ± 0.0004 g/mL (N = 5, one measurement from each of 5 separate stock solution preparations), which would allow for the desired separation of HeLa and E. coli cells. We evaluated several relative concentrations of PEG and Ficoll and selected final concentrations of 7% w/w PEG (8,000 g/mol) and 11% w/w Ficoll (400,000 g/mol) because they produced the desired densities for the cell separation systems. These concentrations produce a system that is 70% top phase 30% bottom phase by volume. We chose PEG 8K because it allows for sterilization by filtration using 0.22 µm filters, while larger molecular weight PEG does not. In order to ensure biocompatibility, we tested a range of phosphate buffered saline (PBS) concentrations and found that systems prepared in 0.6X PBS yielded a top phase osmolality of 292 ± 3 mOsm/kg, pH 7.38 ± 0.02, and a bottom phase osmolality of 294 ± 4 mOsm/kg, pH 7.38 ± 0.03 (Table 1). This ATPS formulation allows the cells to be separated without adversely affecting their morphology and viability. We also performed a focused study on the

ACS Paragon Plus Environment

7

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 8 of 29

reproducibility of ATPS preparation comparing systems produced from single stock polymer solutions (Table S1) to systems from 5 separate stock polymer solutions (Table 1).

Materials and Methods Preparation of Aqueous Two-Phase Systems Stock solutions of 0.6X PBS were prepared by diluting concentrated 10X PBS stock solutions (Fisher Scientific) to “0.6X” rather than the typical 1X working concentration. The pH of stock solutions of 0.6X PBS was adjusted using 5 M HCl to generate systems with physiological pH. Stock polymer solutions of 14% w/w PEG (Amresco, average molecular weight 8,000 g/mol) and 22% w/w Ficoll (Corning, average molecular weight 400,000 g/mol) were prepared by dissolving the polymers in the 0.6X PBS solvent. These solutions were sterilized by vacuum filtration through “Steriflip” 0.22 µm filters (EMD Millipore). Three grams of each stock solution were added to a 15-mL conical tube for a final overall concentration of 7% w/w PEG and 11% w/w Ficoll. The systems were vortexed for ~20 seconds to ensure complete mixing then centrifuged at 1500g for 15 minutes to achieve phase separation. The phases were divided for analysis by pipetting the top phase and draining the bottom phase through a hole pierced in the bottom of the conical tube in order to avoid remixing. The properties of each phase (i.e., density, osmolality, and pH) were characterized. The densities were measured using an Anton Paar DMA 4100M density meter, the osmolalities were measured using a Wescor Vapro Model 5600 vapor pressure osmometer, and pH was measured using a VWR Symphony B10P pH meter. Culture of Cells and Microorganisms LC3-GFP-expressing HeLa cells (GFP-HeLa)30–32 were generously provided by Prof. Beth Levine (UT Southwestern Medical Center) and unmodified HeLa cells were purchased from the ATCC (CCL-2). Both variants of HeLa cells were cultured in DMEM (Corning), 10% fetal bovine serum (FBS; Gibco) at 37 °C, and 5% CO2. HL-60 II cells were a generous gift provided by Prof. Michael Pawlita (DFKZ German Cancer Research Center). These suspension cells were

ACS Paragon Plus Environment

8

Page 9 of 29 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

cultured in RPMI 1640 (Corning), 10% FBS, and 1% Penn/Strep at 37 °C, and 5% CO2. mPlumexpressing E. coli was a gift from Michael Davidson & Roger Tsien (Addgene plasmid # 54564). These DH5α E. coli are ampicillin resistant and mPlum expression is under a pBad promoter. E. coli were grown in Lysogeny Broth (LB; Amresco) to an OD600 of 0.6–0.7 then induced with 0.2% w/v L-arabinose (Acros Organics) and allowed to grow 15 hours. After induction, red fluorescence was confirmed by fluorescence microscopy. mCitrine-expressing yeast were generously provided by Prof. Stephen Fuchs (Tufts University). Yeast were cultured in yeastextract peptone dextrose (YPD; Fisher) overnight at 30 °C. Separation of Cell Contaminants For our “model contamination” culture, GFP-HeLa cells were grown to confluency. The DMEM was removed from the cultures and replaced with 3 mL of L-15 to allow for incubation in atmospheric CO2. 1 mL of mPlum-E. coli suspension (OD600 ≈ 1.0) was added to the cultures and allowed to incubate for 30 minutes. The medium in this model contaminated cell culture was removed by pipette and the adherent cells were washed with 1X PBS. The cells were then trypsinized (0.25% + EDTA; Hyclone, GE Life Sciences) for 5 min at 37 °C to facilitate removal from a 60-mm plate. Next, the cells were pelleted by centrifugation (200g for 5 min) and resuspended in 1 mL of complete culture medium. This suspension was layered on top of the ATPS by pipette. The system was centrifuged for 15 min at 2000g to allow the contaminants to sediment by density to the bottom of the tube and the mammalian cells to sediment by density to the liquid-liquid interface. This method used to separate cells in ATPS by density is therefore in stark contrast to methods that rely on partitioning: cells encounter each phase sequentially in a process driven by sedimentation under an applied gravitational field, rather than interacting with all phase components simultaneously in a homogenized system. If the densities of the cells are known, then ATPS and their phase densities can be designed a priori to enrich for cells of interest at predetermined locations within the system.

ACS Paragon Plus Environment

9

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 29

The top phase was aspirated by pipette leaving approximately 0.5 mL above the interface. The liquid close to and including the interface (~2 mL) was transferred by pipette into a 15-mL conical tube containing 3 mL of fresh Leibovitz Medium (L-15; Hyclone, GE Life Sciences) in order to dilute the polymer solutions and prevent further phase-separation. L-15 was used for this step because the pH of DMEM is above physiological pH in atmospheric CO2 concentrations. This solution was centrifuged for 10 min at 200g to pellet the recovered cells. The cells were then washed twice with 1 mL L-15. Finally, they were diluted to the desired density to seed new plates or for analysis (Figure 1). Viability Enrichment In order to determine our method’s ability to enrich cell populations for viable cells, 1x106 cells were seeded in a 60-mm plate and left in the incubator for seven days without exchanging media, which led to severe overgrowth, depletion of nutrients, and acidification of the cell culture medium. The medium was removed along with cells in suspension and reserved for analysis while the adherent cells were trypsinized to remove them from the plate. These cells were passed through the PEG-Ficoll ATPS and the cells from the interface were recovered. Viability was analyzed using propidium iodide exclusion and quantified with flow cytometry (EMDMillipore Guava easyCyte Flow Cytometer 6HT-2L). Imaging and Analysis of Model Contaminated Systems To measure the efficiency of our separation method, model contaminated cultures were prepared and treated using the ATPS method described above. The cultures were analyzed for the presence of contaminants before and after treatment by microscopy, flow cytometry, and colony formation assays. Contaminated and recovered cells were imaged by confocal microscopy (Leica DMi8 with Andor Revolution DSD2 confocal imaging system). GFP-HeLa cells were imaged using a GFP filter and E. coli using an RFP filter with a 20X lens (NA = 0.4, WD = 0–2 mm). Cultures were analyzed for the presence of GFP- and mPlum-expressing cells.

ACS Paragon Plus Environment

10

Page 11 of 29 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

Cell counts for percent recovery calculations were performed using a Countess II Automated Cell Counter (ThermoFisher). Flow Cytometry Cells were analyzed using two different methods. The first was designed to identify GFPHeLa cells that had mPlum E. coli bound to them both before and after separation. We identified this population by first gating for single cells, then for cells that were positive for green fluorescence, and then for cells that were positive for red fluorescence (Figure S3). The second was designed to identify all red-fluorescing species in the suspension, which would include free E. coli but would differentiate it from debris. We identified these cells by simply gating for all events that were positive for red fluorescence (Figure S4). Colony Formation Assay Contaminated GFP-HeLa cell suspensions were serially diluted by a factor of 10 with 1X PBS. 10 µL of undiluted, 1:10, 1:100, and 1:1000 suspensions were added to the top of an agar plate in a line. The plate was tilted to allow the droplets to run vertically down the plate until they reached the bottom. These plates were incubated at 37 °C for ~15 hours and imaged with a digital camera. Single colonies were identified by eye and counted using ImageJ to record the analysis (Figure S5). Treatment Effectiveness To determine if the ATPS treatment sufficiently reduced the bacteria count low enough to prevent recurrence of contamination, the recovered HeLa cells were cultured in DMEM containing antibiotic for one passage cycle and compared to cells cultured without antibiotic and without ATPS treatment. Colony formation assays were performed prior to treatment with ATPS method and again after 3 days in culture. Applicability to Other Cells and Contaminants We cultured HeLa cells and HL-60 II cells as described above then suspended them at a density of 1x106 cells/mL in warm 1X PBS. We then added 5 µM DiO (green) or DiI (red) and

ACS Paragon Plus Environment

11

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 12 of 29

incubated for 30 min at 37 °C in order to stain the membranes of the cells and provide fluorescent contrast with the contaminants. The cell suspensions were then washed twice with 1X PBS and resuspended in 5 mL L-15. One HL-60 II suspension was stained green (DiO) and contaminated with 1 mL OD600 mPlum E. coli and another was stained red (DiI) and contaminated with 1 mL OD600 mCitrine yeast. Finally, a HeLa culture was also stained red (DiO) and contaminated with 1 mL OD600 mCitrine yeast. These suspensions were incubated for 30 min to allow microorganisms to interact with the cells and then they were treated with our ATPS and analyzed by flow cytometry to determine the separation efficiency. Gating strategies were established to differentiate fluorescent contaminants from fluorescent mammalian cells. Statistical Analysis Prism 7 (GraphPad) was used for all statistical analyses including generating plots, fitting curves to data, performing student’s t-tests, and comparing linear regression slope comparison.

Results and Discussion Recovery of Mammalian Cells The recovery percentage is important for continuing a cell culture after decontamination treatment because higher cell counts are often required for seeding or other downstream applications. Significant loss of desired cells during separation could inhibit the culture’s ability to grow after recovery particularly for slow growing or fragile cells.39 The density-based method we developed relies on the fact that most of the desired cells will sediment to the liquid-liquid interface where they can easily be recovered. The biocompatibility of the process is also essential because if the recovered cells are damaged or too few cells are recovered, it will not be possible to start a new culture from them. The majority of the desired cells can be isolated according to their density and recovered without substantial loss of cells. We added 1x106 cells to the ATPS by layering 1 mL of cell suspension on top. We recovered an average of 70.2 ± 4.6% of the HeLa cells and 55.9 ± 2.1% of HL-60 II cells after removing them from the interface

ACS Paragon Plus Environment

12

Page 13 of 29 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

by pipette and washing them (Table 2). Loss of cells was likely due to some fraction of the cell population having a different density from the recovered cells as well as adsorption of cells to the pipette tips and conical tubes during separation and wash steps. We observed a positive correlation (r = 0.81) between the number of cells added to the system and the percent of cells recovered (Figure S6). The difference between the recovery percentages between HeLa and HL-60 II cells in this common ATPS can likely be explained by differences in density between these two cell lines (Figure S1). We would expect that a different ATPS comprising a bottom phase with a higher density could be used to recover a greater percentage of HL-60 II cells. Maximizing viability throughout the separation procedure is essential for maintaining these cells in culture after the contaminants are removed. Recovered GFP-HeLa cells had an average viability of 98.0 ± 0.2% compared to control with a viability of 98.4 ± 0.2%. Recovered HL-60 II cells had an average viability of 95.3 ± 2.2% compared to control cells with a viability of 97.2 ±1.1 (Table 2). The control population was aliquoted from the same suspension of cells but was not passed through an ATPS. These data indicate that subjecting the cells to this decontamination procedure, which includes exposure to polymer solutions, relatively high-speed centrifugation, and several pipetting steps, does not result in a loss of viability relative to control. The average viability of cells passed through our ATPS was not significantly different from cells passaged using standard cell culture technique (GFP-HeLa p value = 0.3, HL-60 II p value = 0.4). Because calculations of seeding densities often rely on cell counts, maximizing viability while eliminating nonviable cells allows for more accurate seeding of subsequent cultures. We generated growth curves by plotting cell count vs. time for 3 days after initial seeding and compared the growth rates of the control population to the recovered cells for both GFPHeLa and HL-60 II (Figure 2A). The control curve represents the average from 4 biological replicates, while the ATPS curve represents an average of 4 biological replicates, which are each the average of 3 technical replicates (N = 12 total measurements). The population doubling times for the control and recovered populations of GFP-HeLa cells were 1.08 days and

ACS Paragon Plus Environment

13

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 29

1.55 days, respectively. For HL-60 II cells, the doubling times were 0.70 and 0.72 days for control and recovered populations respectively. The suspension cell lines show almost no difference after being treated while the adherent GFP-HeLa cells show a slight lag in their growth rate. These rates were calculated from the exponential growth curve fit to the data. To determine if the growth rates were significantly different, the natural log (ln) of cell count was plotted vs. days of growth. The slope of this curve is the rate constant. The control and ATPStreated populations were plotted (Figure 2B) and their slopes, and therefore their growth rates, were not found to be significantly different for HL-60 II cells (p value = 0.6) and slightly, but significantly different for the HeLa cells (p value = 0.04). We did not investigate whether there is still polymer adsorbed to the cells after the separation process and wash steps, but growth rates indicate that the polymers did not significantly alter the suspension cells’ behavior in culture and generated only a slight lag in growth for adherent cells. This result highlights the biocompatibility of this procedure. Enrichment of Viability Cell cultures may become contaminated with nonviable cells either as the result of an experiment (e.g., in vitro cytotoxicity assay)40 or due to neglect and overgrowth. These cells will typically have a different density than the general population due to loss of osmotic control or total lysis.41 Because of this difference in density, our method can be used to enrich the population for viable cells. In order to demonstrate this application of our ATPS, we used overgrown cultures of HeLa cells with high numbers of nonviable cells. Cells suspended in the medium of the overgrown culture had an average viability of 7.2%, while overgrown cells adhered to the dish had an average viability of 47.4%. After treatment with our ATPS, cells recovered from the interface had an average viability of 66.8% (N=3 technical replicates containing 3 biological replicates each) indicating that our ATPS enriches the viability of overgrown cells by an average of 40.9% (Figure 3). These data show that our ATPS can be used to enrich the viability of neglected cells, and that treatment with our ATPS significantly

ACS Paragon Plus Environment

14

Page 15 of 29 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

increases viability of cell populations. Student’s t tests for each of the 3 trials indicate significant difference between ATPS-enriched and control populations with p values of 0.002, < 0.001, and < 0.001 in trials 1, 2, and 3, respectively. Enrichment of Mammalian Cells from Model Contaminated Cultures We tracked the effectiveness of the separation and the ability of the system to remove contamination using fluorescence microscopy (Figure 4), flow cytometry, and CFU counts (Table 3). We analyzed suspensions of cells after three separation trials for the presence of total bacteria as well as GFP-HeLa cell-bacteria aggregates by flow cytometry. Cell-bacteria aggregates were identified as cell-sized particles that were positive for both red and green fluorescence. Total bacteria were identified as any particle that was positive for red fluorescence. Additionally we performed colony formation assays to quantify the amount of active bacteria remaining in the culture following treatment. We found that ATPS treatment reduced CFU by 43 ± 28%, bacteria-cell aggregates by 33 ± 15%, and free bacteria were reduced by 26 ± 9%. For trials 1 and 2, our ATPS considerably reduced the bacteria present in the suspension, decreasing the CFU by 60% and 80%, respectively. In these two trials, the ATPS also reduced the cell-bacteria aggregates by 56% and 38% and the total number of bacteria by 40% and 31%, as determined by flow cytometry. In trial 3, however, there were 12% more bacteria in the treated suspension than the control, while the system reduced the cellbacteria aggregates by only 5% and the total number of bacteria by 8%. The three technical replicates for CFU assays were reproducible within each of the three biological replicate trials. The inconsistency in the outcomes among these three trials highlights inherent differences expected of biological replicates. Ideally the method would remove all of the detectable bacteria from the culture, but it may not be necessary. Significantly reducing the bacteria count in the culture can increase the effectiveness and reduce the required duration of antibiotic treatment. We attempted to remove the adhered bacteria from the cells prior to adding them to the ATPS through multiple methods (e.g., wash buffer containing surfactant), but none were successful

ACS Paragon Plus Environment

15

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 29

(Supporting Information). To test the success of post-separation antibiotic treatment, we generated a fourth contaminated culture and treated it with our ATPS separation method. We used this suspension of cells, which had 84% less CFU after treatment (Table 4), to seed a new culture with and without penicillin/streptomycin antibiotics. We grew these cultures for 3 days and then re-evaluated them by colony formation assay. No bacteria were detectable after this brief antibiotic treatment. This result indicates that absent 100% separation efficiency, the ATPS can still be used in conjunction with other treatments to recover the desired cells. Applicability to Other Cultures and Contaminants In order to determine if our system, which was specially tuned to enrich HeLa cells from a culture contaminated with E. coli, could be more broadly applied to other cell cultures and other contaminants, we tested a suspension cell line and yeast contaminants. We found that in an HL60 II culture contaminated with E. coli, the system was able to remove an average of 56 ± 2% of the contaminants. In an HL-60 II culture contaminated with yeast, the ATPS treatment removed 50 ± 1% of the contaminants. Finally in a culture of HeLa cells contaminated with yeast, ATPS treatment removed 32 ± 1% of the contaminant (Figures S7–S9). These results indicate that our ATPS, while unable to fully decontaminate the culture, is currently able to substantially enrich the contaminated cell suspensions for the desired mammalian cells.

Conclusions This work describes a method for density-based separation of desired mammalian cells from contaminants in culture including microorganisms and nonviable cells. Separation was achieved using a PEG-Ficoll ATPS whose physical properties were characterized and tuned to ensure biocompatibility. This method could be useful in laboratory settings where cell cultures have become contaminated or overpopulated with nonviable cells and researchers wish to attempt to save the cultures rather than simply dispose of them. The ATPS separation can be used to reduce the amount of contaminants considerably in most cases. In the event that the separation

ACS Paragon Plus Environment

16

Page 17 of 29 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

is unsuccessful or incomplete, as was the case in some of our trials, the cells can still be disposed of or treated with antibiotics. This makes our inexpensive method a low risk, high reward option in instances where the contaminated cells are of particularly high value. Our ATPS, as currently designed, is unable to achieve 100% efficient separations. This is likely due to some fraction of microorganisms being adhered to the mammalian cells. Future ATPS may be optimized to remove these bacteria and improve on separation efficiency through further manipulation of the physical (e.g., density, osmolality, interfacial tension) and chemical properties (e.g., polymer identities, concentrations, contaminant binding affinities) of the system, or by altering the protocol (e.g., number of wash steps, centrifugation speeds, and number of passes through the ATPS). The sharp, thermodynamically stable, and potentially biocompatible density step at the liquid-liquid interface of ATPS allows us to exploit inherent differences in densities between desired cell populations and a wide array of potential contaminants. In addition to E. coli and nonviable cells, ATPS can be potentially be developed to eliminate any contaminant whose density is different from the cultured cells including other bacteria, mycoplasma, yeast, and mold.

Acknowledgments This work was supported by Tufts University and a Research Corporation for Science Advancement SEED Award granted to Prof. Charles Sykes (Tufts University).

Supporting Information Experimental

details

regarding

bacteria-cell

aggregation,

image

processing,

ATPS

reproducibility, empirical mammalian cell and microorganism density, flow cytometry gating, colony formation assays, cell recovery, enrichment of cells from yeast, and ATPS set up; Figures S1–S10; Table S1.

ACS Paragon Plus Environment

17

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 29

Figure 1. Schematic representation of the cell culture decontamination procedure using aqueous two-phase systems (ATPS). (1) A contaminated culture of adherent cells is identified through discolored, cloudy culture medium. (2) The contaminated medium is removed and the plate is washed with buffer, which reduces the number of bacteria. (3) The cells are trypsinized and resuspended in fresh medium. (4) 1 mL of this contaminated cell suspension is carefully layered on the top phase of a prepared ATPS. (5) The ATPS is centrifuged for 15 minutes at 2000g causing the contaminants to sediment to the bottom of the tube and the cells to collect at the liquid interface. (6) The top phase is removed and the 2 mL surrounding and including the interface is transferred by pipette into 3 mL of fresh medium (7) leaving the bottom phase and contaminants to be discarded (8). (9) The cell suspension is washed with medium to remove phase-forming components, and resuspended in fresh medium to be cultured (10). (11) Finally, the decontaminated cell suspension is seeded in a new culture and monitored for the recurrence of contamination.

ACS Paragon Plus Environment

18

Page 19 of 29 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 2. Growth curves for GFP-HeLa and HL-60 II cells after being recovered from the interface of our ATPS. Cell count data for GFP-HeLa (A) and HL-60 II (C) were plotted as a function of days of growth. The control curve is the average of 4 biological replicates. The ATPS curve is an average of 4 biological replicates, which are each the average of 3 technical replicates (N = 12 total). Error bars represent the standard error of the mean. The data were fit to exponential growth curves for quantitative comparison of growth rate. The natural log (ln) of cell counts vs. days after separation for GFP-HeLa (B) and HL-60 II (D) were plotted and fit to lines. The slopes of these lines represent the growth rate of the cultures, and were not found to differ significantly for HL-60 II (p value = 0.6). For GFP-HeLa cells, the growth rate had a slight difference (p value = 0.04).

ACS Paragon Plus Environment

19

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 29

Figure 3. Viability enrichment of overgrown cells upon separation in ATPS. The viability of overgrown cells was analyzed using propidium iodide dye exclusion and flow cytometry (N = 10,000 events). Three biological replicates each consisting of three technical replicates were performed. Averages are plotted with error bars representing the standard error of the mean. Student’s t-test results indicate significant differences between the viability of control and ATPStreated populations with p values of = 0.002 for trial 1 (*) and < 0.001 for trials 2 and 3 (**).

ACS Paragon Plus Environment

20

Page 21 of 29 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 4. Selected representative images of major steps in the decontamination procedure. A. The initial GFP-HeLa culture after contamination with mPlum E. coli and 30 minutes of incubation in L-15 medium. B. The contaminated medium is aspirated and the plate is washed with 1X PBS to remove most of the free bacteria. C. The positive control (contaminated) population after trypsinization, which contains several bacteria. D. The contaminated population after being recovered from the interface of the ATPS reducing the number of bacteria present. White arrows in C and D mark bacteria. Scale bars in each image represent 50 µm.

ACS Paragon Plus Environment

21

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 29

Table 1. Density, osmolality, and pH of the top and bottom phases of our ATPS. Uncertainty reported is standard deviation and standard error of the mean (N = 5 replicates).

Stock 1 2 3 4 5 Average Stdev SEM

Stock 1 2 3 4 5 Average Stdev SEM

Top Phase Density Osmolality (g/mL)

(mOsm/kg)

1.0431 1.0432 1.0434 1.0435 1.0434 1.0433 0.0002 0.0001

296 289 292 289 294 292 3 1

pH 7.37 7.36 7.37 7.39 7.40 7.38 0.02 0.01

Bottom Phase Density Osmolality (g/mL)

(mOsm/kg)

1.0762 1.0755 1.0755 1.0751 1.0752 1.0755 0.0004 0.0002

296 289 292 296 299 294 4 2

pH 7.42 7.35 7.35 7.41 7.37 7.38 0.03 0.01

ACS Paragon Plus Environment

22

Page 23 of 29 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 2. Percent of cells added to the ATPS recovered from the interface with their respective viabilities. Each trial represents a biological replicate, which is the average of 3 technical replicates. A single control was analyzed for each trial. Final averages are of all replicates (N = 12 for ATPS, N = 4 for controls). Uncertainty shown is standard error of the mean.

Trial 1 Trial 2 Trial 3 Trial 4 Average

GFP-HeLa Percent Control Recovery Viability 72.4 ± 0.7 98.0 76.2 ± 1.5 99.1 75.6 ± 2.0 98.2 56.6 ± 2.0 98.4 70.2 ± 4.6 98.4 ± 0.2

Recovered Viability 97.8 ± 0.2 98.7 ± 0.1 97.8 ± 0.3 97.6 ± 0.2 98.0 ± 0.2

Trial 1 Trial 2 Trial 3 Trial 4 Average

HL-60 II Percent Control Recovery Viability 57.2 ± 7.1 99.8 53.4 ± 2.5 98.1 61.1 ± 3.6 96.3 51.7 ± 3.1 94.5 55.9 ± 2.1 97.2 ± 1.1

Recovered Viability 99.4 ± 0.1 89.4 ± 8.7 97.5 ± 0.3 95.0 ± 0.1 95.3 ± 2.2

ACS Paragon Plus Environment

23

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 29

Table 3. Quantitative measurements of the separation efficiency of the PEG–Ficoll ATPS by counts of colony forming units (CFU) and two flow cytometry gating strategies. Flow cytometry results are reported as the percent of total cells that are within the gating parameters for that experiment. Results from CFU assays are the average of 3 ATPS replicates shown with standard error of the mean.

CFU Counts Contaminated ATPS Treated % Difference

Trial 1 9700 3800 ± 300 60 ± 3

Trial 2 39000 7700 ± 600 80 ± 3

Trial 3 28000 31000 ± 2000 -12 ± 7

Flow Cytometry: Cell-Bacteria Aggregates (% of Events) Trial 1 Trial 2 Trial 3 Contaminated 0.9 4.0 21.8 ATPS Treated 0.4 2.5 20.7 % Difference 55.6 37.5 5.0 Flow Cytometry: Total Bacteria (% of Events) Trial 1 Trial 2 Trial 3 Contaminated 1.7 4.7 15.6 ATPS Treated 1.0 3.2 14.3 % Difference 42.0 31.2 8.4

ACS Paragon Plus Environment

24

Page 25 of 29 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 4. CFU/mL counts for contaminated cell suspensions. ATPS separation reduced CFU count by 84%. Day 3 indicates CFU present after 3 days in culture. Cells cultured with 1% penicillin/streptomycin showed zero CFU.

Initial

Recovered

Day 3

Control

25000

--

6000

ATPS

31000

5100

1600

ACS Paragon Plus Environment

25

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 26 of 29

References (1)

Langdon, S. P. In Methods in Molecular Medicine; Langdon, S. P., Ed.; Humana Press Inc., 2004; Vol. 88: Cancer, pp 309–317.

(2)

Drexler, H. G.; Uphoff, C. C. Cytotechnology 2002, 39, 75–90.

(3)

Olarerin-George, A. O.; Hogenesch, J. B. Nucleic Acids Res. 2015, 43, 2535–2542.

(4)

González-González, M.; Willson, R. C.; Rito-Palomares, M. Sep. Purif. Technol. 2016, 158, 103–107.

(5)

Clynes, M. M.; Smyth, H.; O’Kennedy, R.; Corrigan, A. In Vitro 1980, 16, 809–812.

(6)

Kuhlmann, I. Cytotechnology 1995, 19, 95–105.

(7)

Cohen, S.; Samadikuchaksaraei, A.; Polak, J. M.; Bishop, A. E. Tissue Eng. 2006, 12, 2025–2030.

(8)

Ryan, J. Corning Tech. Bull. 2008, No. Table 1, 1–24.

(9)

Albertsson, P.-Å. Biochim. Biophys. Acta 1958, 27, 378–395.

(10)

Mace, C. R.; Akbulut, O.; Kumar, A. A.; Shapiro, N. D.; Derda, R.; Patton, M. R.; Whitesides, G. M. J. Am. Chem. Soc. 2012, 134, 9094–9097.

(11)

Hennek, J. W.; Kumar, A. A.; Wiltschko, A. B.; Patton, M. R.; Lee, S. Y. R.; Brugnara, C.; Adams, R. P.; Whitesides, G. M. Lab Chip. 2016, 16, 3929–3939.

(12)

Passos, H.; Costa, S. H.; Fernandes, A. M,; Freire, M. G.; Rogers, R. D.; Coutinho, J. A. P. Angew. Chem. Int. Ed. 2017, 56, 1–6.

(13)

Walter, H. Trends Biochem. Sci. 1978, 3, 97–100.

(14)

Sousa, A. F.; Andrade, P. Z.; Pirzgalska, R. M.; Galhoz, T. M.; Azevedo, A. M.; da Silva, C. L.; Aires-Barros, M. R.; Cabral, J. M. S. Biotechnol. Lett. 2011, 33, 2373–2377.

(15)

Fisher, D. Biochem. J. 1981, 196, 1–10.

(16)

Atefi, E.; Joshi, R.; Mann, J. A.; Tavana, H. ACS Appl. Mater. Interfaces 2015, 7, 21305– 21314.

(17)

Kumar, A. A.; Lim, C.; Moreno, Y.; Mace, C. R.; Syed, A.; Van Tyne, D.; Wirth, D. F.;

ACS Paragon Plus Environment

26

Page 27 of 29 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

Duraisingh, M. T.; Whitesides, G. M. Am. J. Hematol. 2015, 90, 31–36. (18)

Kumar, A. A.; Patton, M. R.; Hennek, J. W.; Lee, S. Y. R.; D’Alesio-Spina, G.; Wang, X.; Shevkoplyas, S. S.; Kanter, J.; Brugnara, C.; Whitesides, G. M. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 14864–14869.

(19)

Tavana, H.; Kaylan, K.; Bersano-Begey, T.; Luker, K. E.; Luker, G. D.; Takayama, S. Adv. Funct. Mater. 2011, 21, 2920–2926.

(20)

Frampton, J. P.; Tsuei, M.; White, J. B.; Abraham, A. T.; Takayama, S. Biotechnol. J. 2015, 10, 121–125.

(21)

Kubitschek, H. E.; Baldwin, W. W.; Graetzer, R. J. Bacteriol. 1983, 155, 1027–1032.

(22)

Bryan, A. K.; Goranov, A.; Amon, A.; Manalis, S. R. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 999–1004.

(23)

Wolff, D. A.; Pertoft, H. J. Cell Biol. 1972, 55, 579–585.

(24)

Bryan, A. K.; Hecht, V. C.; Shen, W.; Payer, K.; Grover, W. H.; Manalis, S. R. Lab Chip 2014, 14, 569–576.

(25)

Keppler, O. T.; Hinderlich, S.; Langner, J.; Schwartz-Albiez, R.; Reutter, W.; Pawlita, M. Science. 1999, 284, 1372–1376.

(26)

Kruk, P. A.; Auersperg, N. Vitr. Cell. Dev. Biol. 1991, 27A, 273–276.

(27)

Pertoft, H. J. Biochem. Biophys. Methods 2000, 44, 1–30.

(28)

SooHoo, J. R.; Walker, G. M. Biomed. Microdevices 2009, 11, 323–329.

(29)

Hatti-Kaul, R. Mol. Biotechnol. 2001, 19, 269–277.

(30)

Kabeya, Y.; Mizushima, N.; Ueno, T.; Yamamoto, A.; Kirisako, T.; Noda, T.; Kominami, E.; Ohsumi, Y.; Yoshimori, T. EMBO J. 2000, 19, 5720–5728.

(31)

Orvedahl, A.; Macpherson, S.; Sumpter Jr., R.; Tallóczy, Z.; Zou, Z.; Levine, B. Cell Host Microbe 2010, 7, 115–127.

(32)

Mizushima, N.; Yoshimori, T.; Levine, B. Cell 2010, 140, 313–326.

(33)

Lewis, C. L.; Craig, C. C.; Senecal, A. G. Appl. Environ. Microbiol. 2014, 80, 3622–3631.

ACS Paragon Plus Environment

27

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

(34)

Albertsson, P.-Å. Biochem. Pharmacol. 1961, 5, 351–358.

(35)

Pretlow, T. G.; Boone, C. W.; Shrager, R. I.; Weiss, G. H. Anal. Biochem. 1969, 237,

Page 28 of 29

230–237. (36)

González-González, M.; Rito-Palomares, M. Crit. Rev. Biotechnol. 2014, 34, 318–327.

(37)

Ge, X.; Luo, D.; Xu, J. PLoS One 2011, 6, e28707.

(38)

Croll, T.; Munro, P. D.; Winzor, D. J.; Trau, M.; Nielsen, L. K. Biotechnol. Prog. 2008, 19, 1269–1273.

(39)

Freshney, R. I. Culture of Animal Cells, 6th ed.; John Wiley & Sons, Inc.: Hoboken, NJ, 2010.

(40)

Fischer, D.; Li, Y.; Ahlemeyer, B.; Krieglstein, J.; Kissel, T. Biomaterials 2003, 24, 1121– 1131.

(41)

OptiPrep TM The ideal density gradient medium for isolation of macromolecules and lipoproteins http://www.axis-shield-density-gradient-media.com/Purification and analysis of viruses.pdf. (Accessed May 30, 2017)

ACS Paragon Plus Environment

28

Page 29 of 29 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

For TOC Only

ACS Paragon Plus Environment

29