THz Spectroscopy for a Rapid and Label-Free Cell Viability Assay in a

Oct 18, 2018 - ... Guidelines · Copyright & Permissions/RightsLink · ACS Publishing Center ... This liquid serves as the essential environment for cel...
0 downloads 0 Views 2MB Size
Subscriber access provided by UNIV OF NEW ENGLAND

Technical Note

THz spectroscopy for rapid and label-free cell viability assay in microfluidic chip based on optical clearing agent Ke Yang, Xiang Yang, Xiang Zhao, Marc Lamy de la Chapelle, and Weiling Fu Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b03665 • Publication Date (Web): 18 Oct 2018 Downloaded from http://pubs.acs.org on October 24, 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 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 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.

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 19 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

THz spectroscopy for rapid and label-free cell viability assay in microfluidic chip based on optical clearing agent Ke Yang,1,# Xiang Yang,1,# Xiang Zhao,1 Marc Lamy de la Chapelle,2 Weiling Fu1,* 1Department

of Laboratory Medicine, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing 400038, China 2Institut des Molécules et Matériaux du Mans (IMMM-UMR CNRS 6283), Université du Mans, Avenue Olivier Messiaen, 72085 Le Mans, France Corresponding Author: *(W.-L. F); E-mail: [email protected] Author Contributions: # K.Y. and X.Y. contributed equally to this work.

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 19

ABSTRACT: Simple, rapid, and efficient cell viability assay plays a fundamental role in many biomedical researches, including the cell toxicology investigation and antitumor drug screening. Here, we demonstrate for the first time a rapid and label-free cell viability assay using THz spectroscopy in combination with a new optical clearing agent (OCA) and microfluidic technology. This strategy uses a considerably less absorptive OCA to replace the highly absorptive water molecules around the living cells, and thus to decrease the background signal interference. Three low-viscosity oils were screened as potential OCA candidates, among which fluorinated oil was selected because of its lower absorption and lowest cytotoxicity. After replacing the liquid medium with fluorinated oil in a microfluidic chip, an obvious THz spectral difference was observed between the fluorinated oil with and without living cells. This change in THz response was preliminary attributed to the distinguishable signals between the cells and the fluorinated oil. In addition, we applied this method to cell viability assays of human breast cancer cells (MDA-MB-231) after treatment with different antitumor drugs. The results indicated that THz spectroscopy with the aid of the proposed water-replacement strategy presented excellent quantification of cell viability with the advantages of a rapid, label-free, nondestructive and micro assay, which offers significant potential to developing a convenient and practical cell analysis platform. Keywords: Terahertz spectroscopy, Optical clearing agent, Microfluidic chip, Cell viability assay

ACS Paragon Plus Environment

Page 3 of 19 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 Rapid and accurate cell viability assay is of great importance in both fundamental science and clinical medicine. Conventional methods to cell viability assay, such as flow cytometry, microplate reader, and optical microscopy, are encumbered by either relatively expensive instrumentation or poor quantitative capabilities 1,2. Moreover, these approaches required to mark the cells with chemiluminescent, fluorescent, or radioactive labels that may influence the physiological functions of cells and created potential biological hazards for researchers 3. Thus, exploring a label-free, inexpensive and efficient approach for cell viability assay is urgently needed. Recently, terahertz (THz) spectroscopy has emerged as a promising label-free detection method for cellular analysis. THz radiation exhibits excellent sensitivity to water molecules, and its photon energy conveniently corresponds to that of low-frequency bond vibrations, such as hydrogen-bond stretches and distortion 4,5. This technique has been applied in label-free investigations of bacterial cells with different water contents 6. However, because of this notoriously high sensitivity to water contents, the cell viability assay by THz spectroscopy has encountered strong constraints. The medium surrounding the living cells is mainly composed of water molecules. This liquid serves as the essential environment for cell survival, but inevitably causes strong THz energy attenuation 7. In particular, a 1-mm-thick layer of water is sufficient to attenuate the radiation energy at 1 THz by approximately 109 times 8. In addition, THz spectra of both the cells and the medium are dominated by water absorption 6. The THz responses of the cells can be largely masked by the background signal interference of the medium. Therefore, the detection of living cells and their viability assay using THz spectroscopy have been markedly restricted. Several strategies for detecting liquid-phase biological samples using THz spectroscopy have been developed, including the utilization of strong THz radiation or THz attenuated total reflection (THz-ATR) (More details were shown in SI). However, these methods are not extensively used in practical applications because of the relatively poor accessibility, integration limitations and high in cost. Current researches on liquid-phase detection generally utilize tailored sample cells, such as microfluidic chips, to minimize the liquid-sample thickness. These devices are attractive tools for this purpose because of their controlled sample thickness, ease of integration and universality 7. In particular, precise confinement of the thicknesses of liquid samples within tens to hundreds of micrometers considerably reduces THz energy attenuation 7,9. Moreover, microfluidic chips can be easily integrated with other photonic and electronic components, such as waveguides 10, birefringence silicon gratings 11, and various electrodes 12,13. This strategy has clearly improved the THz measurement sensitivity and accuracy, as observed in studies on nucleic acids 11, carbohydrates 11, proteins 9, and chemical mixtures 7,14 in liquid solutions. Compared to the utilization of strong THz radiation or THz-ATR, microfluidic chips offer the additional advantages of considerably lower costs and easier implementation. However, background signal interference remains an unresolved problem in some microfluidic chip-based biomedical studies using THz spectroscopy. This technical hurdle is attributed to the inherent similarities in the THz properties of the living cells and medium, which is difficult to eliminate by decreasing the sample thickness alone 12. Notably, some pioneer studies have demonstrated the use of glycerol as a THz penetration-enhancing agent in the THz tissue detection 15,16, which provided great inspiration. As a typical optical clearing agent (OCA) widely applied for different spectral ranges, glycerol is nearly transparent in the THz range because of its very low THz absorption 15. When glycerol penetrates into the tissues, this less absorptive agent replaces the highly absorptive interstitial water molecules, because it can create the osmotic pressure over the tissues 16. As a result, the background interference from the interstitial water in tissues is considerably weakened. 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 4 of 19

Inspired by these previous works, we here to introduce a strategy for coupling a new OCA with a microfluidic chip for the rapid and label-free cell viability assay using THz spectroscopy. Because the viscosity of glycerol is too high to be combined with the microfluidic chip, we screened three low-viscosity oils as potential OCA candidates. Among them, fluorinated oil was selected as the optimal OCA, and was then injected into a microfluidic chip to expel the culture medium around the attached human breast cancer cells (MDA-MB-231). This strategy generated an obvious THz spectral difference between the fluorinated oil with and without living cells. We also applied this method in cell viability assays after the treatment of two antitumor drugs. With the PIstaining based flow cytometry analysis, the validation of THz quantification of cell viability was accomplished. Our findings suggested that THz spectroscopy with the aid of the proposed water-replacement strategy is a convenient and practical cell analysis tool for rapid, label-free and nondestructive cell toxicology evaluation and antitumor drug screening.

Experimental Section Cytotoxicity and THz absorption evaluation of different OCAs The cytotoxicity evaluation of the three potential OCA candidates, silicon oil (S104742, CAS number: 6314862-9, Aladdin, Shanghai, China), fluorinated oil (FluorinertTM Electronic Liquid FC-40, also called fluorocarbon oil, CAS number: 86508-42-1, 3M, Minnesota, USA), and mineral oil (M3516, CAS number: 8042-47-5, SigmaAldrich, Shanghai, China), was conducted using a cell counting kit-8 (CCK-8; Dojindo, Kumamoto, Japan), as per the manufacturer’s protocol. More details of the CCK-8 assay and THz absorption evaluation can be found in the SI. THz measurement of the cells with the OCA A microfluidic chip with a sample thickness of 0.1 mm was used as the sample cell (Figure S1). Cells were firstly injected into the chip and cultured for 12 h to form a fully-covered cell layer. The selected OCA was then pumped into the chip to replace the Dulbecco’s Modified Eagle’s Medium (DMEM). After replacement, the chip was directly inserted into the sample chamber of the THz spectrometer for measurement. The detailed descriptions of instrumentation and data analysis are provided in SI. Atomic force microscope (AFM) imaging Cell suspension was cultured in a 0.01% poly-D-lysine-coated Petri dish for 24 h. The cells were washed twice with PBS solution and added to 4 mL of DMEM or fluorinated oil for 60 min. The cell appearance and height were then measured using a Dimension Edge Instrument (Bruker Nano Surfaces, Santa Barbara, USA) and AFM cantilevers (MLCT, Bruker Nano Inc., Camarillo, USA). Flow cytometry assay Cell suspension was cultured in a 6-well plate for 24 h and treated with two antitumor drugs, cisplatin and paclitaxel (both from Solarbio, Beijing, China), at doses of 0.5, 5, 50, and 500 μg/mL for 48 h. A solvent control with dimethyl sulfoxide (DMSO), a negative control with streptomycin (500 μg/mL, from Solarbio, Beijing, China) were included in each experiment. The cells were then collected and suspended in propidium iodide (PI) solution. After incubation for 15 min, the viabilities of cells were analyzed using flow cytometry (BD AccuriTM C6, San Diego, CA).

ACS Paragon Plus Environment

Page 5 of 19 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 and discussion THz absorption and cytotoxicity of different OCAs Silicon, fluorinated, and mineral oils were initially selected as the three low-viscosity candidates because they have been extensively applied in microfluidic-based biomedical researches 17. As shown in Figure 1A and 1B, DMEM, which mainly consists of water molecules, predictably exhibited the highest THz absorption and refractive index. Three oils have significantly lower absorption and refractive index than that of DMEM. Therefore, the attenuation of THz energies can be neglected when THz radiation propagates through the less absorptive nonpolar oil layer. Because OCA cytotoxicity can significantly influence the THz measurement, we further calculated the viabilities of MDA-MB-231 breast cancer cells subjected to these oils for 5 min, 30 min and 60 min, respectively. As shown in Figure 1C, the cell viabilities after treatment with fluorinated oil for varying durations (96.90 ± 2.00% for 5 min, 88.16 ± 1.42% for 30 min, and 85.44 ± 1.46% for 60 min) were higher than those treated with the other two oils. Although sample preparation and measurement consumed less than 10 min using our method, maintaining the cell activity as long as possible is preferable in practical applications. Therefore, we extended the observed duration to 60 min for a strict cytotoxicity evaluation. Fluorinated oil retained a cell viability of more than 85% over this longer duration, whereas the cell viability when subjected to silicon oil and mineral oil decreased to 79.07 ± 2.42% and 44.41 ± 1.54%, respectively. This result agrees with the previous studies on cell viability in fluorinated oil, which retained between 70% and 94% after 1h incubation 18-20. The lowest cytotoxicity of fluorinated oil may be attributed to its considerable capability for solubilizing oxygen 21. Because fluorinated oil can dissolve approximately 20 times more oxygen than water, it is particularly well-suited for cell-based assays. Moreover, among the candidates, only fluorinated oil has a higher density (1.86 g/cm3) than water (1.00 g/cm3). This property may favor the replacement process because the lighter water molecules can float on the fluorinated oil and then be completely washed away. Therefore, fluorinated oil was finally selected as the OCA for THz measurement because of its lower absorption and cytotoxicity, and higher density.

Figure 1. THz absorption and cytotoxicity evaluation of different candidate OCAs. (A) Absorption coefficients of DMEM and three candidate OCAs, respectively. Inset graph magnifies the signal from 0.9–1.2 THz. (B) 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 6 of 19

Refractive indexes of DMEM and three candidate OCAs, respectively. Inset graph magnifies the signal from 0.9–1.2 THz. (C) Cell viabilities of MDA-MB-231 cells after treatment with three candidate OCAs for different durations. Water replacement process in the microfluidic chip We first validate the effectiveness of the water replacement process by the fluorinated oil. As illustrated in Figure 2A, the detection protocol included the following steps. MDA-MB-231 cells in DMEM solution were first attached on the surface of the sample area in the microfluidic chip. Fluorinated oil was then injected into the microfluidic chip. The DMEM on the cell monolayer could be easily washed away because of the insolubility of water in oil and the higher density of fluorinated oil than that of water. Figure S2 shows the water displacement (red area) by the fluorinated oil (grey area) inside the microfluidic chip during 168 s. The oil-water interface was observed to move from the inlet to the outlet over time, verifying that the water molecules around the cells were progressively expelled after the injection of fluorinated oil. Furthermore, the optimal replacement time was determined by monitoring the variations of the transmitted power spectra at 1 THz during the replacement at 12 s intervals. As shown in Figure 2B, the THz power remained at a relatively constant level (approximately -63 dB) after 156 s, indicating that the DMEM within the chip has been removed as much as possible. Thus, 156 s was selected as the optimal replacement time. We selected a very low flow rate (2 μL/min) for this replacement process to guarantee that the attached cells were not washed away. Thus, the replacement process had no obvious influence on the population of attached cells, which was vital for the subsequent experiments.

Figure 2. Water replacement process in the microfluidic chip. (A) Schematic of water replacement process. (B) Monitoring of power of 1-THz pulses at 12-s intervals during the replacement. THz spectroscopy of the cells with the OCA After the validation of the water replacement process, we evaluated the performance of this strategy for THz measurements on living cells. Figure 3A and 3B displayed the THz time-domain waveforms and normalized THz transmission amplitude (𝐴 ∗ ) of MDA-MB-231 cells in the fluorinated oil and in the DMEM environments. There was nearly no difference in the THz time-domain waveforms of DMEM with and without cells. In contrast, the peak-to-peak values of the THz time-domain waveforms through fluorinated oil with cells were significantly lower than those of pure fluorinated oil. Similarly, the presence of cells induced an obvious increase 𝐴 ∗ only in the fluorinated oil environment. We compared their values of 𝐴 ∗ at three representing frequency points (0.5, 1.0 and 1.5 THz), as shown in the Figure 3C. By conducting an independent t-test, statistically significant differences were observed between the fluorinated oil with and without cells (P < 0.05 at 0.5, 1.0 and 1.5 THz), whereas no statistically significant differences were observed between the DMEM with and without cells. The undetectable differences in DMEM with and without living cells were mainly attributed to background signal interference. ACS Paragon Plus Environment

Page 7 of 19 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

Water absorption dominates the THz responses of both the cells and background medium (DMEM). As a result, the THz signals of cells are substantially masked. This result agrees with a similar microfluidic chip-based study, which attributed the indistinguishable differences in the THz spectra between T-cell solutions in different states to the similarity in the THz absorption properties of the cell and medium 12. Although some cell properties, such as their complex dielectric constants, result in slight differences in the THz response compared to that of the medium 22, those differences are negligible compared to that between fluorinated oil and cells. As previously indicated, the THz absorption of fluorinated oil is considerably lower than that of DMEM. Thus, the background absorption of the surrounding fluorinated oil was unable to largely mask the THz responses of the cells. As a result, the signal from the living cells dominated the overall THz signal of fluorinated oil with cell. We further analyzed the influence of fluorinated oil on the cell morphology using AFM imaging technique. As shown in Figure S3A and S3C, cells in the fluorinated oil for 60 min still have the same spindle shape and the surrounding lamellipodia with those in the DMEM. In addition, the average height of cells in the fluorinated oil was determined to be 6.78 ± 0.36 μm, similar to the cells in the DMEM with an average height of 6.47 ± 0.23 μm, as shown in Figure S3B and S3D. These results suggested that the fluorinated oil did not markedly damage the morphology of samples 23,24. It gives another evidence of the advantage of proposed strategy for THz measurements on living cells from the aspect of morphology. Compared to previous cell detection researches that utilized air drying 25 or paper 3,26 to remove the DMEM layer before THz measurement, replacing the water molecules with biocompatible fluorinated oil can eliminate the background signal interference and simultaneously maintain the cell activity and morphological characteristics. Such conditions are all the more favorable in obtaining the intrinsic THz spectra of the cells under natural states and very preferable for practical applications.

Figure 3. THz spectra of MDA-MB-231 cells in DMEM and fluorinated oil. (A) THz time-domain waveforms of fluorinated oil without cells, fluorinated oil with cells, DMEM without cells, and DMEM with cells, respectively. The grey boxes highlight the peak-to-peak values of the THz pulses. Spectra of DMEM with cells and DMEM without cells are horizontally offset 3 ps for clarity. (B) Normalized THz transmission amplitudes (𝐴 ∗ ) of the four aforementioned samples. (C) 𝐴 ∗ at 0.5, 1.0 and 1.5 THz for the four aforementioned samples. An independent t-test is used to make statistically comparison within the fluorinated oil with and without cells as well as the DMEM with and without cells, respectively. * refers to (P < 0.05) and N.S. refers to no significance. 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 8 of 19

Cell viability assay by THz spectroscopy After elucidating the underlying mechanism, we next investigated the application of this strategy to the study of viability assay of breast cancer cells subjected to different antitumor drugs. MDA-MB-231 cells were treated with cisplatin and paclitaxel at doses of 0.5, 5, 50, and 500 μg/mL for 48 h. Cells treated with DMSO and streptomycin (500 μg/mL) served as the solvent control and negative control, respectively. The viabilities of the cell samples were then measured by both our proposed method and the PI-staining combined with flow cytometry. In flow cytometry assays, the PI-negative and PI-positive cells are regarded as living and dead cells, respectively, as shown in Figure 4A and 4B. In the solvent control and negative control, majority of cells (approximately 95%) were viable. In contrast, when the cells were treated with two antitumor drugs, the average viabilities both gradually decrease with increasing drug concentrations. Correspondingly, samples in the solvent control and negative control also exhibited the highest 𝐴 ∗ in Figures 5A and 5B, whereas the 𝐴 ∗ decreased with increasing drug concentrations.

Figure 4. Cell viabilities measured by PI-staining combined with flow cytometry for MDA-MB-231 cells treated with cisplatin (A) and paclitaxel (B) at doses of 0.5, 5, 50, and 500 μg/mL for 48 h. Cells treated with DMSO and streptomycin (500 μg/mL) served as the solvent control and negative control, respectively.

Figure 5. Normalized THz transmission amplitudes (𝐴 ∗ ) of MDA-MB-231 cells after treatment by cisplatin (A) and paclitaxel (B) at doses of 0.5, 5, 50, and 500 μg/mL for 48 h. Pure fluorinated oil, cells treated with DMSO and cells treated with streptomycin (500 μg/mL) served as the blank control, solvent control and negative control, respectively.

ACS Paragon Plus Environment

Page 9 of 19 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

These results can be mainly attributed to the different THz responses of the living and dead cells. Cisplatin and paclitaxel kill cells by inducing cell apoptosis or even necrosis 27,28, leading to changes in cell membrane permeability and loss of cell membrane integrity 2. As a result, the cytoplasm may leak outside the cell and can be washed away by the injected fluorinated oil. Compared to the living cells, the 𝐴 ∗ associated with dead cells may decrease because of the pronounced loss of the high-THz-absorbing substances, particularly the intracellular water molecules. Based on this principle, a recent study successfully used THz-ATR imaging to measure the permeabilizations of living cells quantitatively 29. In addition, dead cells attached in the detection area may become fewer in number after treatment by a high-concentration drug. Some dead cells have been observed to break away from the substrate and even float in the medium 26. As a result, small amounts of floating cells were considered to have washed away after replacement with fluorinated oil, further decreasing the 𝐴 ∗ .Therefore, the 𝐴 ∗ of cell sample is proportional to the quantity of living cells present on the detection area. The differences between samples and blank control in Figure 5A and 5B are nearly constant with the increasing frequency. Thus, we further compared their value differences at 1 THz for the quantitative analysis of the cell viabilities, as shown in figure 6A. The relative changes in the normalized THz transmission amplitude at 1 THz (∆𝐴



∆𝐴0∗ ) of cell samples were calculated using the following equation: ∆𝐴

*

* * 𝐴sample ― 𝐴blank control

∆𝐴0* (%) = 𝐴 *

solvent control

* ― 𝐴blank control

× 100 %

.

∆𝐴 ∗ ∆𝐴0∗ gradually decreased with increases in both drug concentrations. As the 𝐴 ∗ is proportional to the ∗ quantity of living cells, we can assume that the ∆𝐴 ∆𝐴0∗ corresponds to the cell viabilities. To confirm this, ∗ the Figure 6B shows ∆𝐴 ∆𝐴0∗ as a function of the average cell viabilities measured by flow cytometry. Linear

relationships were found for both drugs with a slope close to one. This result indicates that THz spectroscopy with the aid of the proposed water-replacement strategy can quantify cell viability.

Figure 6. Quantitative analysis of the cell viabilities using THz spectroscopy. (A) Relative changes in the ∗ normalized THz transmission amplitude at 1 THz (∆𝐴 ∆𝐴0∗ ) of MDA-MB-231 cells after treatment by two

drugs with different concentrations. (B) Linear fitting between the relative changes in normalized THz transmission amplitude at 1 THz (∆𝐴



∆𝐴0∗ ) and the cell viabilities measured by flow cytometry.

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 19

Compared to other common cell viability assay methods, such as flow cytometry, microplate reader, and optical microscopy, THz spectroscopy offers several attractive properties, as listed in Table S1. In general, some reagents used in conventional methods are suspected to be carcinogens 1. In contrast, THz spectroscopy enables nondestructive and noncontact cell viability assays, which avoid both any damages to the cell samples and any potential biological hazards for researchers. The cell numbers required for each analysis of flow cytometry are generally more than 10,000 cells 26. By comparison, the numbers of THz spectroscopy in this study were approximately 3,600 cells. This value is higher than that required for the microplate reader (at least 1000 cells/well for the CCK-8 test), but is nonetheless acceptable for most of cell biology experiments. However, the proposed strategy now also has certain limitations. Because of the insolubility of cells in fluorinated oil 21, it is difficult to disperse the cells directly in the OCA. Hence, this method is currently limited to adherent cells. This issue also hinders investigations using THz metamaterials 3,26 and THz-ATR 30. Cell trapping approaches, such as antibody capture 31, are worth trying to realize the specific isolation of suspended cells. In this direction, an integrated assay platform will be established in future investigations, whereby the cell trapping and THz measurement can be successively conducted.

Conclusion In summary, we demonstrated a new strategy involving the replacement of water molecules around living cells with an OCA, for quantification of cell viability under drug treatment using THz spectroscopy. Fluorinated oil was reported for the first time to function as an OCA in THz frequency range. After introducing fluorinated oil in a microfluidic chip, we clearly observed the THz responses of living cells in the fluorinated oil. This THz response change can be attributed to the distinguishable signals between the cells and the fluorinated oil. Using this strategy, we have been able to observe variation of the THz signal of MDA-MB-231 cells treated with antitumor drugs and to quantify the cell viabilities in a rapid, label-free, and non-destructive manner. Our work provided a proof-of-concept of a convenient and efficient cell analysis platform for the detection of living cells and their viability assay using THz spectroscopy. Compared to the previous methods of liquid-phase biological sample detection, such as strong THz radiation and THz-ATR, the proposed strategy is low-cost, simple, fast and easy to operate, all of which are advantages in practical applications, such as the cell toxicology evaluation and antitumor drug screening.

Acknowledgements This work was supported by the National Basic Research Program of China (2015CB755400), the National Natural Science Foundation of China (81430054, 81802118), Military logistics scientific research project (AWS17J010) and Joint medical scientific research project of Chongqing (2018QNXM007).

Supporting information Details of introduction and experimental section. Image of the microfluidic chip (Figure S1). Images of the microfluidic chip at different stages of the replacement process (Figure S2). Analysis of the MDA-MB-231 cell morphology using AFM imaging technique (Figure S3). Comparison of THz spectroscopy with common cell ACS Paragon Plus Environment

Page 11 of 19 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

viability assay methods (Table S1).

Conflicts of interest
 The authors declare no competing financial interest.

References (1) Johnson, S.; Nguyen, V.; Coder, D.; John Wiley & Sons, Inc.: Hoboken, New Jersey, USA, 2013, pp 9.2.19.2.26. (2) Wlodkowic, D.; Telford, W.; Skommer, J.; Darzynkiewicz, Z. Apoptosis and Beyond: Cytometry in Studies of Programmed Cell Death. Method. Cell. Biol. 2011, 103, 55-98. (3) Liu, Y.; Tang, M.; Xia, L.; Yu, W.; Peng, J.; Zhang, Y.; Chapelle, M. L. D. L.; Yang, K.; Cui, H. L.; Fu, W. Cell viability and hydration assay based on metamaterial-enhanced terahertz spectroscopy. Rsc. Adv. 2017, 7, 53963-53969. (4) Xiang, Y.; Xiang, Z.; Ke, Y.; Liu, Y.; Yu, L.; Fu, W.; Yang, L. Biomedical Applications of Terahertz Spectroscopy and Imaging. Trends Biotechnol 2016, 34, 810-824. (5) Sun, Q.; He, Y.; Liu, K.; Fan, S.; Parrott, E. P. J.; Pickwell-MacPherson, E. Recent advances in terahertz technology for biomedical applications. Quant. Imaging Med. Surg. 2017, 7, 345-355. (6) Xiang, Y.; Dongshan, W.; Shihan, Y.; Yueping, L.; Shu, Y.; Mingkun, Z.; Zhongbo, Y.; Xiaoyan, Z.; Qing, H.; Hong‐Liang, C.; Weiling, F. Rapid and label‐free detection and assessment of bacteria by terahertz time‐domain spectroscopy. J. Biophotonics 2016, 9, 1050-1058. (7) Liu, L.; Pathak, R.; Cheng, L. J.; Wang, T. Real-time frequency-domain terahertz sensing and imaging of isopropyl alcohol–water mixtures on a microfluidic chip. Sensor Actuat B-Chem 2013, 184, 228-234. (8) Duponchel, L.; Laurette, S.; Hatirnaz, B.; Treizebre, A.; Affouard, F.; Bocquet, B. Terahertz microfluidic sensor for in situ exploration of hydration shell of molecules. Chemometr. Intell. Lab. 2013, 123, 28-35. (9) Smith, A. E.; Hawkins, B. G.; Kirby, B. J.; Rana, F.; George, P. A.; Hui, W. Microfluidic devices for terahertz spectroscopy of biomolecules. Opt. Express 2008, 16, 1577-1582. (10) Masini, L.; Meucci, S.; Xu, J.; Degl'Innocenti, R.; Castellano, F.; Beere, H. E.; Ritchie, D.; Balduzzi, D.; Puglisi, R.; Galli, A. Terahertz probe of individual subwavelength objects in a water environment. Laser. Photonics. Rev. 2015, 8, 734-742. (11) Saha, S. C.; Grant, J. P.; Ma, Y.; Khalid, A.; Hong, F.; Cumming, D. R. Application of terahertz spectroscopy to the characterization of biological samples using birefringence silicon grating. J. Biomed. Opt. 2012, 17, 067006. (12) Tang, Q.; Liang, M.; Lu, Y.; Wong, P. K.; Wilmink, G. J.; Zhang, D. D.; Xin, H. Microfluidic Devices for Terahertz Spectroscopy of Live Cells Toward Lab-on-a-Chip Applications. Sensors 2016, 16, 476-486. (13) Nemes, C. T.; Swierk, J. R.; Schmuttenmaer, C. A. A Terahertz-Transparent Electrochemical Cell for In Situ Terahertz Spectroelectrochemistry. Anal. Chem. 2018, 90, 4389-4396. (14) Swithenbank, M.; Burnett, A. D.; Russell, C.; Li, L. H.; Davies, A. G.; Linfield, E. H.; Cunningham, J. E.; Wood, C. D. On-Chip Terahertz-Frequency Measurements of Liquids. Anal. Chem. 2017, 89, 7981-7987. (15) Oh, S. J.; Kim, S. H.; Jeong, K.; Park, Y.; Huh, Y. M.; Son, J. H.; Suh, J. S. Measurement depth enhancement in terahertz imaging of biological tissues. Opt. Express 2013, 21, 21299-21305. (16) Smolyanskaya, O. A.; Schelkanova, I. J.; Kulya, M. S.; Odlyanitskiy, E. L.; Goryachev, I. S.; Tcypkin, A. 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 12 of 19

N.; Grachev, Y. V.; Toropova, Y. G.; Tuchin, V. V. Glycerol dehydration of native and diabetic animal tissues studied by THz-TDS and NMR methods. Biomed. Opt. Express 2018, 9, 1198-1215. (17) JC, B. Surfactants in droplet-based microfluidics. Lab. Chip. 2012, 12, 422-433. (18) ClausellTormos; Jenifer; Lieber; Diana; Baret; JeanChristophe; ElHarrak; Abdeslam; Miller; Oliver, J. Droplet-Based Microfluidic Platforms for the Encapsulation and Screening of Mammalian Cells and Multicellular Organisms. Chemistry & Biology 2008, 15, 427-437. (19) Koster, S.; Angile, F. E.; Duan, H.; Agresti, J. J.; Wintner, A.; Schmitz, C.; Rowat, A. C.; Merten, C. A.; Pisignano, D.; Griffiths, A. D.; Weitz, D. A. Drop-based microfluidic devices for encapsulation of single cells. Lab. Chip. 2008, 8, 1110-1115. (20) Chen, F.; Zhan, Y.; Geng, T.; Lian, H.; Xu, P.; Lu, C. Chemical transfection of cells in picoliter aqueous droplets in fluorocarbon oil. Anal. Chem. 2011, 83, 8816-8820. (21) Mazutis, L.; Gilbert, J.; Ung, W. L.; Weitz, D. A.; Griffiths, A. D.; Heyman, J. A. Single-cell analysis and sorting using droplet-based microfluidics. Nat. Protoc. 2013, 8, 870-891. (22) Shiraga, K.; Suzuki, T.; Kondo, N.; Tanaka, K.; Ogawa, Y. Hydration state inside HeLa cell monolayer investigated with terahertz spectroscopy. Appl. Phys. Lett. 2015, 106, 253701-253705. (23) Fletcher, D. A.; Mullins, R. D. Cell mechanics and the cytoskeleton. Nature 2010, 463, 485. (24) Yun, X.; Tang, M.; Yang, Z.; Wilksch, J. J.; Xiu, P.; Gao, H.; Zhang, F.; Wang, H. Interrogation of drug effects on HeLa cells by exploiting new AFM mechanical biomarkers. Rsc. Adv. 2017, 7, 43764-43771. (25) Liu, H. B.; Plopper, G.; Earley, S.; Chen, Y.; Ferguson, B.; Zhang, X. C. Sensing minute changes in biological cell monolayers with THz differential time-domain spectroscopy. Biosens. Bioelectron. 2007, 22, 1075-1080. (26) Zhang, C.; Liang, L.; Ding, L.; Jin, B.; Hou, Y.; Li, C.; Jiang, L.; Liu, W.; Hu, W.; Lu, Y. Label-free measurements on cell apoptosis using a terahertz metamaterial-based biosensor. Appl. Phys. Lett. 2016, 108, 241105-241109. (27) Rebillard, A.; Lagadic-Gossmann, D.; Dimanche-Boitrel, M. T. Cisplatin cytotoxicity: DNA and plasma membrane targets. Curr. Med. Chem. 2008, 15, 2656-2663. (28) Wang, T. H.; Wang, H. S.; Soong, Y. K. Paclitaxel-induced cell death: where the cell cycle and apoptosis come together. Cancer 2000, 88, 2619-2628. (29) Grognot, M.; Gallot, G. Quantitative measurement of permeabilization of living cells by terahertz attenuated total reflection. Appl. Phys. Lett. 2015, 107, 103702-103705. (30) Zou, Y.; Liu, Q.; Yang, X.; Huang, H.-C.; Li, J.; Du, L.-H.; Li, Z.-R.; Zhao, J.-H.; Zhu, L.-G. Label-free monitoring of cell death induced by oxidative stress in living human cells using terahertz ATR spectroscopy. Biomed. Opt. Express 2018, 9, 14-24. (31) Park, S. J.; Hong, J. T.; Choi, S. J.; Kim, H. S.; Park, W. K.; Han, S. T.; Park, J. Y.; Lee, S.; Kim, D. S.; Ahn, Y. H. Detection of microorganisms using terahertz metamaterials. Sci. Rep. 2014, 4, 4988.

ACS Paragon Plus Environment

Page 13 of 19 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. THz absorption and cytotoxicity evaluation of different candidate OCAs. (A) Absorption coefficients of DMEM and three candidate OCAs, respectively. Inset graph magnifies the signal from 0.9–1.2 THz. (B) Refractive indexes of DMEM and three candidate OCAs, respectively. Inset graph magnifies the signal from 0.9–1.2 THz. (C) Cell viabilities of MDA-MB-231 cells after treatment with three candidate OCAs for different durations. 59x43mm (600 x 600 DPI)

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

Figure 2. Water replacement process in the microfluidic chip. (A) Schematic of water replacement process. (B) Monitoring of power of 1-THz pulses at 12-s intervals during the replacement. 32x13mm (600 x 600 DPI)

ACS Paragon Plus Environment

Page 14 of 19

Page 15 of 19 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. THz spectra of MDA-MB-231 cells in DMEM and fluorinated oil. (A) THz time-domain waveforms of fluorinated oil without cells, fluorinated oil with cells, DMEM without cells, and DMEM with cells, respectively. The grey boxes highlight the peak-to-peak values of the THz pulses. Spectra of DMEM with cells and DMEM without cells are horizontally offset 3 ps for clarity. (B) Normalized THz transmission amplitudes (A*) of the four aforementioned samples. (C) A* at 0.5, 1.0 and 1.5 THz for the four aforementioned samples. An independent t-test is used to make statistically comparison within the fluorinated oil with and without cells as well as the DMEM with and without cells, respectively. * refers to (P < 0.05) and N.S. refers to no significance. 61x47mm (600 x 600 DPI)

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

Figure 4. Cell viabilities measured by PI-staining combined with flow cytometry for MDA-MB-231 cells treated with cisplatin (A) and paclitaxel (B) at doses of 0.5, 5, 50, and 500 μg/mL for 48 h. Cells treated with DMSO and streptomycin (500 μg/mL) served as the solvent control and negative control, respectively. 82x39mm (300 x 300 DPI)

ACS Paragon Plus Environment

Page 16 of 19

Page 17 of 19 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 5. Normalized THz transmission amplitudes (A*) of MDA-MB-231 cells after treatment by cisplatin (A) and paclitaxel (B) at doses of 0.5, 5, 50, and 500 μg/mL for 48 h. Pure fluorinated oil, cells treated with DMSO and cells treated with streptomycin (500 μg/mL) served as the blank control, solvent control and negative control, respectively. 30x11mm (600 x 600 DPI)

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

Figure 6. Quantitative analysis of the cell viabilities using THz spectroscopy. (A) Relative changes in the normalized THz transmission amplitude at 1 THz (∆A*⁄∆A0*) of MDA-MB-231 cells after treatment by two drugs with different concentrations. (B) Linear fitting between the relative changes in normalized THz transmission amplitude at 1 THz (∆A*⁄∆A0*) and the cell viabilities measured by flow cytometry. 31x12mm (600 x 600 DPI)

ACS Paragon Plus Environment

Page 18 of 19

Page 19 of 19 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

TOC for web publication 21x8mm (600 x 600 DPI)

ACS Paragon Plus Environment