Advertisement

Ultrasound Frequency Sonication Facilitates High-Throughput and Uniform Dissociation of Cellular Aggregates and Tissues

Open AccessPublished:January 11, 2023DOI:https://doi.org/10.1016/j.slast.2023.01.001

      Abstract

      A sample preparation step involving dissociation of tissues into their component cells is often required to conduct analysis of nucleic acids and other constituents from tissue samples. Frequently, the extracellular matrix and cell-cell adhesions are disrupted via treatment with a chemical dissociating reagent or various mechanical forces. In this work, a new, high-throughput, multiplexed method of dissociating tissues and cellular aggregates into single cells using ultrasound frequency bath sonication is explored and characterized. Different operating parameters are evaluated, and a treatment protocol with potential for uniform, high-throughput tissue dissociation is compared to the existing best chemical and orbital plate shaking protocol. Metrics such as percent dissociation, cellular recovery, average aggregate size, proportion of various aggregate sizes, membrane circularity, and cellular viability are subsequently assessed and found to be favorable. In optimized conditions, 53 ± 8% of 1 mm biopsy cores are dissociated within 30 minutes using sonication alone, surpassing leading high-throughput orbital plate shaking techniques five-fold. Chemical digestion is also 2 times more effective when complexed with sonication rather than orbital plate shaking. RNA content, quality, and expression are found to be superior to the standard protocol in terms of transcriptional preservation.

      Keywords

      Introduction

      Cellular dissociation is a ubiquitous step in sample preparation workflows for genetic, transcriptomic, proteomic, or metabolomic profiling of tissue sections or aggregates of cells [
      • Welch E.C.
      • Tripathi A.
      Preparation of Tissues and Heterogeneous Cellular Samples for Single-Cell Analysis.
      ,
      • van den Brink S.C.
      • et al.
      Single-cell sequencing reveals dissociation-induced gene expression in tissue subpopulations.
      ,
      • Denisenko E.
      • et al.
      Systematic assessment of tissue dissociation and storage biases in single-cell and single-nucleus RNA-seq workflows.
      ,
      • Welch E.C.
      • Yu H.
      • Barabino G.
      • Tapinos N.
      • Tripathi A.
      Electric-Field Facilitated Rapid and Efficient Dissociation of Tissues Into Viable Single Cells.
      ,
      • Karasu Benyes Y.
      • Welch E.C.
      • Singhal A.
      • Ou J.
      • Tripathi A.
      A Comparative Analysis of Deep Learning Models for Automated Cross-Preparation Diagnosis of Multi-Cell Liquid Pap Smear Images.
      ]. Converting tissues and aggregates into suspensions of individual cells improves extraction efficiency in bulk preparations [
      • Welch E.C.
      • Tripathi A.
      Preparation of Tissues and Heterogeneous Cellular Samples for Single-Cell Analysis.
      ,
      • Welch E.C.
      • Yu H.
      • Tripathi A.
      Optimization of a Clinically Relevant Chemical-Mechanical Tissue Dissociation Workflow for Single-Cell Analysis.
      ,
      • Donlin L.T.
      • et al.
      Methods for high-dimensional analysis of cells dissociated from cryopreserved synovial tissue.
      ]. Additionally, suspensions of intact single-cells can be analyzed by downstream single-cell analysis techniques such as scRNAseq [
      • van den Brink S.C.
      • et al.
      Single-cell sequencing reveals dissociation-induced gene expression in tissue subpopulations.
      ,
      • Welch E.C.
      • Yu H.
      • Tripathi A.
      Optimization of a Clinically Relevant Chemical-Mechanical Tissue Dissociation Workflow for Single-Cell Analysis.
      ,
      • Achim K.
      • et al.
      High-throughput spatial mapping of single-cell RNA-seq data to tissue of origin.
      ,
      • Chen X.
      • Teichmann S.A.
      • Meyer K.B.
      From Tissues to Cell Types and Back: Single-Cell Gene Expression Analysis of Tissue Architecture.
      ].
      Recent work has been conducted to optimize or streamline the process of tissue dissociation into single cells. Commercialized products including Miltenyi Biotec's GentleMACS Dissociator and S2 Genomics Singulator 100 combine mechanical forces with chemical treatment and thermal regulation. These systems work by automating the decades-old sample preparation approach that has been developed at the benchtop into a single, optimized instrument.
      Other work has been conducted that aims to address the problem from a different avenue. Examples include the use of microfluidic dissociation devices that use different types of mechanical forces, and design of devices that use alternative physical mechanisms entirely. In our previous work, we developed a method for tissue dissociation using electric fields [
      • Welch E.C.
      • Yu H.
      • Barabino G.
      • Tapinos N.
      • Tripathi A.
      Electric-Field Facilitated Rapid and Efficient Dissociation of Tissues Into Viable Single Cells.
      ,

      E. C. Welch and A. Tripathi, “Electrical Dissociation Of Tissue Samples Into Single Cells And/Or Smaller Groups Of Cells,” US Patent App. 17/721,618, 2021.

      ,

      E. C. Welch and A. Tripathi, “Electrical Dissociation Of Tissue Samples Into Single Cells And/Or Smaller Groups Of Cells,” WO Patent App. 22/25,451, 2022.

      ]. We herein investigate a method using ultrasonic cavitation generated shear forces.
      Sonication is a technique wherein sound waves are applied throughout a suspension, resulting in particle agitation [
      • Zhang L.
      • Belova V.
      • Wang H.
      • Dong W.
      • Möhwald H.
      Controlled Cavitation at Nano/Microparticle Surfaces.
      ]. The agitation is caused by the movement of sound waves throughout the medium in the form of pressure peaks and troughs. Ultrasonic baths spread this energy throughout the bath volume, while sonication probes apply it to a focused area. Using a bubble formation and collapse cavitation mechanism, the sound waves are converted into mechanical, kinetic energy that can actuate the suspended particles.
      Sonication is used in many material fabrication applications, particularly in ultrasonic cleaning [
      • Weller R.N.
      • Brady J.M.
      • Bernier W.E.
      Efficacy of ultrasonic cleaning.
      ,
      • Mason T.J.
      Ultrasonic cleaning: An historical perspective.
      ,
      • Lim A.
      Membrane fouling and cleaning in microfiltration of activated sludge wastewater.
      ], de-agglomeration [
      • Marković S.
      • Mitrić M.
      • Starčević G.
      • Uskoković D.
      Ultrasonic de-agglomeration of barium titanate powder.
      ,
      • Sauter C.
      • Emin M.A.
      • Schuchmann H.P.
      • Tavman S.
      Influence of hydrostatic pressure and sound amplitude on the ultrasound induced dispersion and de-agglomeration of nanoparticles.
      ], and nanofabrication [
      • Park C.
      • et al.
      Dispersion of single wall carbon nanotubes by in situ polymerization under sonication.
      ,
      • Huang W.
      • Lin Y.
      • Taylor S.
      • Gaillard J.
      • Rao A.M.
      • Sun Y.-P.
      Sonication-Assisted Functionalization and Solubilization of Carbon Nanotubes.
      ,
      • Dhumal R.S.
      • Biradar S.V.
      • Paradkar A.R.
      • York P.
      Ultrasound Assisted Engineering of Lactose Crystals.
      ,
      • Öztürk S.
      • Akata B.
      Oriented assembly and nanofabrication of zeolite A monolayers.
      ]. Sonication generally uses ultrasonic frequencies of > 20 kHz. Many laboratories have ultrasonic baths that can be used for cleaning, mixing reagents, degassing, resuspending nanoparticles, and other applications. Ultrasound is often used in continuous crystallization installations [
      • Dhumal R.S.
      • Biradar S.V.
      • Paradkar A.R.
      • York P.
      Ultrasound Assisted Engineering of Lactose Crystals.
      ] and to separate bulk materials into individual nanoparticles [
      • Gilca I.A.
      • Popa V.I.
      • Crestini C.
      Obtaining lignin nanoparticles by sonication.
      ]. As many applications of sonication involve mixing aimed at particle dispersion, it follows that this technique could be applied to the problem of tissue dissociation into individual component cells.
      While sonication has been extensively applied to cellular lysis for the extraction of cellular contents, particularly proteins, the dissociation of complex, ex vivo tissues into intact component cells has yet to be thoroughly explored in the literature [
      • Obermaier C.
      • et al.
      Free-flow isoelectric focusing of proteins remaining in cell fragments following sonication of thyroid carcinoma cells.
      ]. However, recent studies examining the acoustophoretic manipulation of cells have shown that such manipulations can be gentler and have lower transcriptional modulation than mechanical or electrokinetic methods [
      • Guex A.G.
      • Di Marzio N.
      • Eglin D.
      • Alini M.
      • Serra T.
      The waves that make the pattern: a review on acoustic manipulation in biomedical research.
      ,
      • Mohanty S.
      • Khalil I.S.M.
      • Misra S.
      Contactless acoustic micro/nano manipulation: a paradigm for next generation applications in life sciences.
      ,
      • Ding X.
      • et al.
      On-chip manipulation of single microparticles, cells, and organisms using surface acoustic waves.
      ,
      • Guo F.
      • et al.
      Three-dimensional manipulation of single cells using surface acoustic waves.
      ]. We herein investigate the utility of sonication in high-throughput tissue dissociation with downstream molecular analysis applications in mind.
      We first investigate the capability of sonication-only dissociation in cellular isolation from bovine liver tissue samples, examining results across numerous parameters to determine an optimum treatment protocol. We then combine sonication-based dissociation with chemical dissociation and compare the results to orbital plate shaking and chemical and orbital plate shaking hybrid dissociation in terms of efficiency of dissociation, viability, ease of use, uniformity, and high-throughput capability. Techniques are assessed regarding their ability to process ex vivo tissue biopsy samples and in vitro MDA-MB-231 cellular aggregates. We characterize ultrasonication bath facilitated tissue dissociation and cellular dispersion as a highly effective method that outperforms competing methods in respect to its high uniformity.

      Materials and Methods

      Physical Theory of Sonication-Based Tissue Dissociation

      The physical theory prompting investigation into sonication-based tissue dissociation is based on extant physical research describing fragmentation in sonication. The following equations are presented to represent an overview of the underlying physical mechanism and how the sonication-based dissociation process can be modeled. However, these equations will need to be further expanded to adequately address more complex systems such as ex vivo tissues that contain multiple cell types and will experience attenuation of the generated shear forces closer to the tissue core.
      The basic fragmentation equation can be described, wherein a(x) is the rate of fragmentation of particles of a given size x that results when a particle of size y breaks up (Equation 1) [
      • Ziff R.M.
      New Solutions To The Fragmentation Equation.
      ,
      • Edwards B.F.
      • Cai M.
      • Han H.
      Rate equation and scaling for fragmentation with mass loss.
      ]. In this case, the particles can be considered the individual single cells, while the agglomerate can be considered the cellular aggregate or tissue.
      c(x,t)t=a(x)n(x,t)+xa(y)b¯(x|y)c(y,t)dy
      (1)


      The particle mass distribution is represented by n(x,t), and the distribution of daughter particle masses is represented by b¯(x|y). The continuous-mass-loss rate can also be added to the equation, as explored by Edwards et al. Fragmentation has also been explored using fragmentation number (Equation 2) [
      • Edwards B.F.
      • Cai M.
      • Han H.
      Rate equation and scaling for fragmentation with mass loss.
      ].
      Ft=τσt
      (2)


      Where the fragmentation number is the ratio of applied shear stress (τ) to the cohesive strength of the tissue or cellular aggregate (σt) [
      • Rwei S.P.
      • Manas-Zloczower I.
      • Feke D.L.
      Observation of carbon black agglomerate dispersion in simple shear flows.
      ,
      • Asanuma Y.
      • Faizal F.
      • Khairunnisa M.P.
      • Lenggoro I.W.
      Deagglomeration of spray-dried submicron particles by low-power aqueous sonication.
      ,
      • Bałdyga J.
      • Makowski Ł.
      • Orciuch W.
      • Sauter C.
      • Schuchmann H.P.
      Deagglomeration processes in high-shear devices.
      ]. The cohesive strength of the tissue or cellular aggregates can be estimated using an additional equation (Equation 3) [
      • Asanuma Y.
      • Faizal F.
      • Khairunnisa M.P.
      • Lenggoro I.W.
      Deagglomeration of spray-dried submicron particles by low-power aqueous sonication.
      ]. In this equation, ε¯ is the average porosity, while Etf represents the energy required to break-up the cell-cell interaction, and Dc is the average cell size.
      σt=1.1×1ε¯ε¯EtfDc3
      (3)


      The Rayleigh-Plesset equation has also been used extensively to characterize the stress generated by cavitation bubble dynamics (Equation 4) [
      • Asanuma Y.
      • Faizal F.
      • Khairunnisa M.P.
      • Lenggoro I.W.
      Deagglomeration of spray-dried submicron particles by low-power aqueous sonication.
      ,
      • Tey W.Y.
      • Alehossein H.
      • Qin Z.
      • Lee K.M.
      • Kang H.S.
      • Lee K.Q.
      On stability of time marching in numerical solutions of rayleigh-plesset equation for ultrasonic cavitation.
      ,
      • Khairunnisa M.P.
      • Faizal F.
      • Miyazawa E.
      • Masuda K.
      • Tsukada M.
      • Lenggoro I.W.
      Detachment of Submicron Particles from Substrates Using the Suspension-Assisted Ultrasonic Method.
      ]. In this equation, R(t) contains dynamic pressure terms and P(t) is the ultrasound driving pressure. ρl represents the density of water, 1,000 kg m−3, and P0 is the atmospheric pressure, 1 atm. The right part of the equation is used to model the effects of surface tension at the bubble-water interface, water viscosity, and sound waves emitted from bubbles [
      • Asanuma Y.
      • Faizal F.
      • Khairunnisa M.P.
      • Lenggoro I.W.
      Deagglomeration of spray-dried submicron particles by low-power aqueous sonication.
      ]. The respective values are σ = 0.073 kg s−2, ηl = 1.00 × 10−3 Pa s, cl = 1481 m s−1. pgas(R,t) represents the gas pressure within generated bubbles [
      • Asanuma Y.
      • Faizal F.
      • Khairunnisa M.P.
      • Lenggoro I.W.
      Deagglomeration of spray-dried submicron particles by low-power aqueous sonication.
      ,
      • German R.M.
      Coordination number changes during powder densification.
      ].
      pl(RR¨+32R2)=pgas(R,t)P(t)P0+Rclddtpgas(R,t)4ηlR˙R2σR
      (4)


      The energy of particle-particle interactions is traditionally expressed using an equation in which A is expressed as the Hamaker constant, 0.63 × 10−20 J (Equation 5). As there are more than van der Waals forces at play in cell-cell interactions, this value cannot be used, however, the adhesion energy between cells of interest can be assessed via other biosensor platforms and equations [
      • Sztilkovics M.
      • et al.
      Single-cell adhesion force kinetics of cell populations from combined label-free optical biosensor and robotic fluidic force microscopy.
      ,
      • Sancho A.
      • Vandersmissen I.
      • Craps S.
      • Luttun A.
      • Groll J.
      A new strategy to measure intercellular adhesion forces in mature cell-cell contacts.
      ].
      Etf=A12π
      (5)


      The total energy balance of the system can then be written (Equation 6). Etot represents the total energy produced by the collapse of a cavitation bubble, Etm represents the motion of the tissue or cellular aggregate and Etf represents the energy requirement for the tissue to dissociate or “fragment” [
      • Asanuma Y.
      • Faizal F.
      • Khairunnisa M.P.
      • Lenggoro I.W.
      Deagglomeration of spray-dried submicron particles by low-power aqueous sonication.
      ].
      Etot=Etm+Etf+Eo
      (6)


      Sonication Bath Setup

      The ultrasonication bath used in this study was the VWR B3500A-DTH ultrasonic cleaner (VWR, Radnor, Pennsylvania). This ultrasonication bath is equipped with a sweep function generator, to ensure even processing. User-controllable operating parameters include bath temperature, degassing time, and emitted ultrasound properties. The input parameters that were used to power the instrument were 110-120 Volts, 50/60 Hz, and 350 Watts. The sonication output parameters that were used were 135 Watts with a frequency of 42 kHz ± 6%.
      A foam holder was used for the sonication bath to study the high-throughput ability of this processing method (Heathrow Scientific, Vernon Hills, Illinois). The holder consisted of 4 rows and 6 columns, capable of holding a total of 24 1.5 or 2 mL tubes, each containing 100 µL of PBS (Sigma-Aldrich, St. Louis, Missouri) as well as either a 1 mm tissue biopsy core, or a suspension of cellular aggregates. The specific tubes used in the study were 1.5 mL Eppendorf DNA LoBind PCR tubes (Eppendorf, North Rhine-Westphalia, Germany).

      Orbital Plate Shaking Setup

      The orbital plate shaker that was used was the PerkinElmer TriNEST incubator shaker instrument (PerkinElmer, Waltham, Massachusetts). This orbital shaker enabled customization of temperature, orbital shaking speed, and processing time. 500 RPM was used for all trials. 25°C temperature was used for room temperature processing in orbital plate shaking-only trials. 37°C temperature was used for processing in chemical and orbital plate shaking trials, to ensure collagenase activation.

      Chemical Dissociation Protocols

      The chemical dissociation protocol in tissue consisted of chemical digestion with a 1% collagenase type I solution (Thermo Fisher Scientific, Waltham, Massachusetts) that was prepared in PBS and processed at 37°C. After processing time had elapsed, the collagenase solution was deactivated with equal parts media, as recommended by Thermo Fisher Scientific. Single-use aliquots were stored at -20°C, brought up to temperature for use, and used or discarded within 2 weeks.
      The chemical dissociation protocol in MDA-MB-231 cellular aggregates was altered to reflect the lack of collagen in in vitro cultured cellular aggregates. The modified chemical dissociation protocol consisted of chemical digestion with 1X Gibco TrypLE Express Enzyme (Fisher Scientific, Pittsburgh, Pennsylvania). This is a trypsin-based reagent that is used for dissociating aggregates amongst numerous cultured cell lines. This reagent is often used for dispersing cellular monolayers during cell culture passage.

      Tissue and Cell Sources

      Bovine Liver Tissue

      Bovine liver tissue was used as an experimental tissue model, as described previously [
      • Welch E.C.
      • Yu H.
      • Tripathi A.
      Optimization of a Clinically Relevant Chemical-Mechanical Tissue Dissociation Workflow for Single-Cell Analysis.
      ]. The tissue was obtained from a local butcher and subsequently cryopreserved for later analysis. 1 mm biopsy cores were taken from the liver and individually characterized in terms of weight and dimension.

      MDA-MB-231 Cell Culture

      MDA-MB-231 triple-negative breast cancer cells were used to examine the effects of sonication and other treatments on cellular viability, as the bovine liver tissue was unsuitable to address this research question due to being cryopreserved. The MDA-MB-231 cells were cultured as described previously [
      • Welch E.C.
      • Yu H.
      • Tripathi A.
      Optimization of a Clinically Relevant Chemical-Mechanical Tissue Dissociation Workflow for Single-Cell Analysis.
      ].
      Cellular aggregates were formed by allowing cells to settle in a 15 mL tube overnight, gently resuspending the pellet, performing cell counting using an INCYTO C-Chip Disposable hemocytometer (VWR, Radnor, Pennsylvania), and then splitting the tube of cells into volume and cell-count matched samples for processing.

      Flow Cytometry Analysis

      Flow cytometry analysis was conducted using a FACSAria III Flow Cytometer (BD, Franklin Lakes, New Jersey) and following an established framework created by the authors for use in liver tissue [
      • Welch E.C.
      • Yu H.
      • Tripathi A.
      Optimization of a Clinically Relevant Chemical-Mechanical Tissue Dissociation Workflow for Single-Cell Analysis.
      ]. Cells were treated briefly with DNAse I solution and red blood cell lysis buffer. Cells were then nonspecifically stained with Hoechst33342 Ready Flow Reagent (Thermo Fisher Scientific, Waltham, Massachusetts).
      A combination of cell-type specific size binning was combined with size gating using size-specific fluorescent microspheres of 6 µm (Thermo Fisher Scientific, Waltham, Massachusetts) and ∼30 µm (Cospheric, Santa Barbara, California). Cell counts were collected with respect to their place in various size and complexity bins, and this information was cross-referenced with fluorescence information from the nonspecific nuclear stain to determine a final number of isolated hepatocytes and other cells from the bovine liver tissue.
      The number of cells was expressed in terms of percent dissociation of the liver tissue using an established model for bovine liver tissue [
      • Welch E.C.
      • Yu H.
      • Tripathi A.
      Optimization of a Clinically Relevant Chemical-Mechanical Tissue Dissociation Workflow for Single-Cell Analysis.
      ] based on work cataloging the number of cells per tissue volume [
      • Junatas K.L.
      • et al.
      Stereological analysis of size and density of hepatocytes in the porcine liver.
      ] and weight [
      • Wilson Z.E.
      • et al.
      Inter-individual variability in levels of human microsomal protein and hepatocellularity per gram of liver.
      ]. This model was previously shown to have good predictive capability of total cellular recovery from tissues and was herein applied to assess relative percent dissociation.

      Position-Based Analysis in Multiplexed Trials

      The reproducibility of sonication as a high-throughput processing method was scrutinized with respect to competing mechanical and chemical/mechanical hybrid methods. The dissociation efficacy of tissue samples in different locations of the 24-well floating tube holder and 96-well plate were assessed in sonication and orbital plate shaking dissociation trials, respectively.
      The aforementioned flow cytometry framework and tissue compositional model were utilized to express results of position-based dissociation in terms of an average percent dissociation for 3 matched trials. Row and column effects were assessed to determine the effect of position in various conditions.

      Hemocytometry Cellular Recovery Assay

      Hemocytometry was used to briefly assess cellular recovery in MDA-MB-231 cells exposed to various treatments. In the hemocytometry process, 10 µL of the cellular sample was mixed with 10 µL of trypan blue solution (Thermo Fisher Scientific, Waltham, Massachusetts). 10 µL of the mixture was loaded onto a hemocytometer and imaged under the microscope. Cell counts per total solution were then calculated following traditional formulae [
      • Crowley L.C.
      • et al.
      Dead Cert: Measuring Cell Death.
      ].
      After performing initial cell counts of samples using hemocytometry, the samples were exposed to various treatments, and hemocytometry was performed again using the same protocol. Cell counts before and after treatment were compared to assess cellular recovery, which was expressed as a percentage of the initial cell count. Recovery is a metric of total cells in suspension after treatment, not just single cells. The recovery metric is used to assess whether cellular lysis has occurred.

      Confocal Fluorescent Microscopy Analysis

      Cellular Aggregate Size and Morphology Assay

      MDA-MB-231 cells were co-stained with Hoechst33342 and Vybrant Dil stains (Thermo Fisher Scientific, Waltham, Massachusetts) for co-visualization of the nucleus and the cell membrane, respectively. Images were captured using an Olympus FV3000 confocal microscope (Olympus, Center Valley, Pennsylvania). The morphology of individual cells was assessed using FIJI ImageJ circularity analysis (National Institutes of Health).
      The size of cellular aggregates was assessed using ImageJ. Average cellular aggregate sizes were recorded by taking an average of the size of every cellular aggregate consisting of 3 or more cells. Multiple interrogation regions were examined per sample. Average size distribution histograms were also created to account for the relative proportion of singlet (∼10-15 µm), doublet (∼15-20 µm), small (20-30 µm), medium (30-100 µm), large (100-200 µm) and extra-large (200-400 µm) cellular clumps per interrogation area.

      Live-Dead Assay

      MDA-MB-231 cells were co-stained with Hoechst33342 and DRAQ7 stains for co-visualization (Thermo Fisher Scientific, Waltham, Massachusetts). While Hoechst33342 nonspecifically stains the nuclei of live or dead cells, DRAQ7 selectively stains the nuclei of dead cells only. Hoechst33342 stained cells were used to obtain an overall cell count to assess the number of cells within a given interrogation area in ImageJ. DRAQ7 stained cells were then counted to express viability as the total number of non-dead cells per total cells. The results were cross-referenced with hemocytometry findings.

      Nucleic Acid Analysis

      RNA Extraction

      RNA extraction was performed using the QIAGEN RNeasy Micro Kit (QIAGEN, Hilden, Germany). Samples of 1 million MDA-MB-231 cells were exposed to various conditions. Conditions included a control where the cells were not treated, a buffer condition where the cells were placed in PBS but not treated, a 30-minute chemical and orbital plate shaking hybrid condition, and sonication at 5-, 15-, and 30-minute timepoints using the previously determined optimized sonication conditions.
      After the cells were treated, they were spun down into a pellet via centrifugation and processed using the RNeasy Micro Kit. DNAse I digestion was performed, but RNA cleanup was not performed to give an idea of the raw sample without any purification steps.

      RNA Quantification and Quality Assessment

      The extracted RNA samples were then quantified and assessed using the NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Waltham, Massachusetts). 1 µL of the extracted RNA sample was placed onto the NanoDrop pedestal and absorbances were measured at 260 nm and compared to absorbances at 230 and 280 nm. Total RNA concentration was confirmed using the Agilent 2100 BioAnalyzer and the Nano chip (Agilent Technologies, Santa Clara, California). The BioAnalyzer was also used to assess the RNA Integrity Number (RIN).

      RT-qPCR Stress Marker Analysis

      After RNA content was assessed, the samples were then adjusted to have the same RNA concentration by adding RNase free water. Dilute samples were concentrated using controlled evaporation centrifugation in the Thermo Scientific Savant SpeedVac (Thermo Fisher Scientific, Waltham, Massachusetts).
      RNA was reverse transcribed into cDNA using the Applied Biosystems High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, Massachusetts). RT-qPCR was then performed using 6 probes that have been characterized to be important in the MDA-MB-231 stress response [
      • Chénais B.
      • Cornec M.
      • Dumont S.
      • Marchand J.
      • Blanckaert V.
      Transcriptomic Response of Breast Cancer Cells MDA-MB-231 to Docosahexaenoic Acid: Downregulation of Lipid and Cholesterol Metabolism Genes and Upregulation of Genes of the Pro-Apoptotic ER-Stress Pathway.
      ]. The probes that were used were SERPINE1, INHBE, FLRT1, HSPA5, ECM2, and PLAT. Thermo Fisher TaqMan probes were used to amplify these specific targets in RT-qPCR. Expression changes were assessed using the control treatment as a baseline and were expressed in terms of ΔCq [
      • Ruiz-Valdepeñas Montiel V.
      • Sempionatto J.R.
      • Campuzano S.
      • Pingarrón J.M.
      • Fernández de Ávila B.Esteban
      • Wang J.
      Direct electrochemical biosensing in gastrointestinal fluids.
      ].

      COMSOL Multiphysics Modeling

      COMSOL Multiphysics software (COMSOL, Burlington, Massachusetts) was used to assess the fluid velocity at different well locations within the 96-well plate during the orbital plate shaking process. The simulation was built on existing work extended to a multi-well model to better understand the effect of position-induced variability in high-throughput processing [
      • Welch E.C.
      • Yu H.
      • Tripathi A.
      Optimization of a Clinically Relevant Chemical-Mechanical Tissue Dissociation Workflow for Single-Cell Analysis.
      ].

      Statistical Analysis

      Results are presented in terms of average value ± standard deviation value. Students t-test, one way analysis of variance (ANOVA) or two-way ANOVA were used where applicable. Multiple comparison analysis and 95% confidence interval Tukey's post-hoc tests were also performed. All statistical analyses and calculations were performed using GraphPad Prism software (GraphPad, La Jolla, California). p-values are represented as follows: * p < 0.05, ** p < 0.01, ***p < 0.001, **** p < 0.0001.

      Results

      Assessment of Sonication Dissociation Efficacy Over Time

      Before examining the effects of position and other high-throughput metrics, the optimal parameters for dissociation using ultrasound frequency sonication within a water bath were first assessed. The effect of time was first considered, looking at 5-, 15-, and 30-minute processing times using 1 mm liver tissue biopsy cores submerged in 100 µL of PBS.
      For this initial experiment, no thermal control was used, and the solution bath remained at 25°C. The degassing function was not utilized. Placement in the 24-plex floating tube holder was randomized. All of the processing throughout this work was performed using the same sonication parameters mentioned above.
      Significant dissociation was obtained within 30 minutes. In particular, 33 ± 19% dissociation was obtained at 5 minutes, 51 ± 9% dissociation was obtained at 15 minutes, and 60 ± 15% dissociation was obtained at 30 minutes with N = 6 (Figure 2A).
      One-way ANOVA showed that the difference between 5 and 30 minutes was significant with a p-value < 0.05. The exact p-value that was obtained was 0.0182. Somewhat large standard deviations in the experiment were attributed to the relatively small sample size and tissue heterogeneity.

      Effect of Degassing

      Many commercially available ultrasound frequency sonication baths incorporate a degassing function. Degassing, in this context, refers to sonicating the solution while under a vacuum to remove the gasses from within liquids. During the sonication process, microbubbles are formed. The degassing process depressurizes the solution to reduce the number of cavitating bubbles.
      To infer information about how degassing affects sonication-induced dissociation, a similar 5-, 15-, and 30-minute time trial was conducted with the degassing function turned on for the full processing time.
      At 5 minutes, 28 ± 8% dissociation was obtained, 27 ± 10% dissociation was obtained at 15 minutes, and 35 ± 4% dissociation was obtained at 30 minutes with N = 6 (Figure 2B). One-way ANOVA did not show any significant differences between all 3 combinations of time intervals.
      When observing the 30-minute timepoint only and comparing degassing to sonication only, an unpaired t-test with Welch's correction yielded a p-value of 0.01 and the treatments were found to be statistically significant from one another with p < 0.01 (Figure 2C).
      When comparing the degassing time course to the sonication only time course, a two-way ANOVA with Šídák's multiple comparisons test was conducted. A p-value of 0.003 was obtained when comparing each test across intervals of time. P-values of 0.9, 0.05 and 0.04 were obtained when comparing differences at 5 minutes, 15 minutes, and 30 minutes respectively. The effect was significant across all time points with p < 0.01 and across 15- and 30-minute timepoints with p < 0.05.
      It is also notable that degassing did not eliminate the dissociative effect of sonication entirely but merely dampened the effect. While degassing was able to reduce gas content, bubbles could still be formed and dissipated. In trials where tissue was left to sit without any sonication, minimal dissociation from below 1% to below 10% was observed, with higher recovery at 30-minute timepoints, and 1% or less at 5 minutes.

      Effect of Temperature

      The effect of temperature on sonication-facilitated dissociation was assessed by processing various tissue samples at room temperature (25°C) and body temperature (37°C), both for 30 minutes (N = 6).
      The group processed at 25°C yielded an average of 62 ± 9% dissociation in the timeframe, while the group processed at 37°C yielded an average of 65 ± 25% dissociation in the timeframe (Figure 2D).
      A significant effect was not observed when assessing with an unpaired t-test with Welch's correction, in which the obtained p-value was 0.8. However, a noticeably higher standard deviation was obtained when processing at 37°C.
      Based on a synthesis of these preliminary findings, it was found that the optimal condition for dissociation of tissue biopsy cores for downstream genetic analysis involved treatment at 42 kHz ± 6% and 25°C for 30 minutes without degassing. This condition was then further studied, combined with chemical treatment, and compared to other leading dissociation methods in terms of its high-throughput potential.

      Comparing Dissociation Efficacy in Sonication, Orbital Plate Shaking, Chemical and Hybrid Trials

      To assess the efficacy of the proposed method of dissociating tissue sections using sonication, it was necessary to compare the optimized sonication technique to other existing methods that are more routinely applied in tissue dissociation. These methods consist of dissociation using chemical reagents and mechanical forces.
      Existing chemical and mechanical dissociation methods vary in their efficacy, processing time, and other parameters. Here, we compared the 24-plex sonication method to what is, to our knowledge, the most high-throughput mechanical processing method for dissociating 1 mm biopsy cores currently available - orbital plate shaking in a 96-well plate (Figure 3A).
      To assess the suitability of these mechanical and acoustic methods in high-throughput processing, the effect of sample placement in the organizers was examined to determine the relative reproducibility of the methods across trials and sample locations.
      The average across the orbital plate shaking only condition was 10 ± 9% (Figure 3B,C). The average across the chemical and sonication condition was 72 ± 10%. The average across the sonication only condition was 53 ± 8% (Figure 3D,E). The average across the chemical and orbital plate shaking condition was 37% ± 20%.
      An ordinary one-way ANOVA was used to compare results from all four treatment groups (Figure 3F) to assess the most effective treatment in terms of percent dissociation. All groups were statistically significantly different from one another with a p-value < 0.0001, as assessed using multiple comparisons analysis.
      Grouping comparisons by sonication and orbital plate shaking processing also resulted in a p-value < 0.0001, showing that these processing methods are distinctly statistically significant when compared to one another, and that orbital plate shaking processing methods are uniformly less effective than sonication processing methods, regardless of the presence of chemical dissociating media.

      Assessment of Cellular Recovery

      Cellular integrity following the sonication, chemical, orbital plate shaking, and combination treatments was assessed using various metrics in MDA-MB-231 cell aggregates. First, cellular recovery was assessed using hemocytometry to determine whether significant loss due to cellular lysis was observed during the treatments (Figure 4A). Before treating the samples, initial cell counts were obtained using the hemocytometer. After treating the samples with the various treatments, another cell count was performed.
      The cellular recovery was expressed in terms of a percentage of the initial cell count. None of the groups experienced significant cellular lysis, although some groups were statistically significant from one another. Sonication trials had a mean cellular recovery of 100 ± 1%, chemical had 103 ± 1%, orbital plate shaking had 91 ± 4%, and chemical and orbital plate shaking had 108 ± 1%.

      Assessment of Average Aggregate Size

      Average aggregate size following the different treatments was also assessed to examine the potential utility of sonication-based dissociation in breaking cellular aggregates apart into their component cells. Treated samples were viewed under the microscope in an imaging dish and interrogation regions were randomly selected. The size of all aggregates containing 3 or more cells was evaluated using ImageJ.
      The untreated control samples were compared to orbital plate shaking and sonication treatments of 5-, 15-, and 30-minute time intervals in terms of average aggregate size. None of the orbital plate shaking treated samples were found to have a statistically significant difference in average aggregate size when compared to the control samples. The 5-, 15-, and 30-minute p-values from one-way ANOVA analysis were 0.5, 1 and 0.6, respectively.
      All the sonication treated samples were found to have statistically significant differences in average aggregate size when compared to the control. The 5-minute timepoint had an adjusted p-value of 0.0001, while the 15 and 30 minute timepoints had p-values < 0.0001.

      Assessment of Aggregate Size Distribution

      MDA-MB-231 aggregate size distribution histograms were created following a similar approach to another published work [
      • Qiu X.
      • et al.
      Microfluidic channel optimization to improve hydrodynamic dissociation of cell aggregates and tissue.
      ]. Representing the data from images in size distribution histograms enabled an understanding of the relative proportions of cells in different size groupings.
      Information about different size groupings in interrogation areas was used to get an idea of the relative proportions of cells exposed to different treatments. Size groupings were designated as singlet (∼10-15 µm), doublet (∼15-20 µm), small (20-30 µm), medium (30-100 µm), large (100-200 µm) and extra-large (200-400 µm) cellular clumps per interrogation area. Multiple images were analyzed and characterized via ImageJ of orbital plate shaking and sonication treatments after 30 minutes of treatment (Figure 4C).
      As the tested cells were in vitro cultured aggregates and not complex tissues, and as MDA-MB-231 cells are generally thought to be a less-adherent cell line, the majority of the cells in images were in singlets or doublets [
      • Evani S.J.
      • Prabhu R.G.
      • Gnanaruban V.
      • Finol E.A.
      • Ramasubramanian A.K.
      Monocytes mediate metastatic breast tumor cell adhesion to endothelium under flow.
      ,
      • Huang Z.
      • Yu P.
      • Tang J.
      Characterization of Triple-Negative Breast Cancer MDA-MB-231 Cell Spheroid Model.
      ]. This also is due to the fact that only a few aggregates can reasonably fit in a given interrogation area due to their larger size.
      It is worth noting that sonication trials had 0 large or extra-large aggregates across all examined images, and an average of ∼4 medium aggregates per image, as compared to an average of ∼39 in orbital plate shaking treatments. Additionally, the number of single-cells per image was statistically significantly larger in the sonication treatment with a p-value of 0.0002, according to two-way ANOVA analysis.

      Assessment of Cellular Integrity

      Cellular integrity was assessed by examining cellular morphology from the aforementioned randomly selected interrogation region images with dual nuclear and membrane stains (Figure 5A). Cellular morphology was analyzed by looking at the roundness of cells in ImageJ. Cells exposed to both orbital plate shaking and sonication treatments were observed to have similar characteristics to untreated cells with essentially no visibly apparent differences. However, some cell membrane fragments were observed in sonication trials, indicating that a low degree of cellular lysis may be occurring during the treatment. However, as seen above, this degree of lysis did not have a significant effect on cellular recovery. Additionally, this was also observed in numerous orbital plate shaking treated samples as well.
      When analyzing the cellular morphology in ImageJ, it was found that no major morphological changes were observed in cells and that none of the groups were found to be statistically significantly different from each other or from the control (Figure 5B). Average values of roundness were 0.74 for the control, 0.71, 0.65, and 0.77 for the orbital plate shaking time course, and 0.73, 0.71, and 0.74 for the sonication time course.
      A viability assay was also conducted using live-dead staining to assess whether different treatments differentially affect viability, using an untreated control as a baseline (Figure 5C). Analyses using confocal microscopy were cross-referenced with results from hemocytometry. It was observed that there is no significant decline in viability across sonication trials regardless of treatment time in comparison to a control.
      Adjusted viability values were higher in sonication treatments than orbital plate shaking treatments across all analyzed timepoints, as assessed using a two-way ANOVA. While sonication viability metrics ranged from 91 - 98% of the control viability, the orbital plate shaking metrics ranged from 64 - 75%. The p-values for the 5-, 15-, and 30-minute comparisons were 0.03, 0.03 and 0.002, respectively.

      Extracted RNA Content and Quality Analysis

      The next objective of this study was to assess the potential of sonication-based tissue and cell aggregate dissociation in downstream analyses. Previous work has shown that RNA expression is quite sensitive to various environmental conditions [
      • van den Brink S.C.
      • et al.
      Single-cell sequencing reveals dissociation-induced gene expression in tissue subpopulations.
      ,
      • Denisenko E.
      • et al.
      Systematic assessment of tissue dissociation and storage biases in single-cell and single-nucleus RNA-seq workflows.
      ]. Chemical treatments, for example, have been shown to change the transcriptome of cells, including MDA-MB-231 [
      • Chénais B.
      • Cornec M.
      • Dumont S.
      • Marchand J.
      • Blanckaert V.
      Transcriptomic Response of Breast Cancer Cells MDA-MB-231 to Docosahexaenoic Acid: Downregulation of Lipid and Cholesterol Metabolism Genes and Upregulation of Genes of the Pro-Apoptotic ER-Stress Pathway.
      ].
      To investigate the effect of various treatments on MDA-MB-231 RNA analyses, RNA content was first extracted from cell-count matched cellular samples exposed to different treatments (Figure 6A). PBS samples tested the effect of media used in the sonication and orbital plate shaking treatments to determine if there was any effect based on media alone. Chemical and orbital plate shaking control samples were processed in 1% collagenase solution at 500 RPM for 30 minutes. Sonication samples were processed for 5-, 15-, and 30-minute timepoints.
      The RNA content extracted from the chemical and orbital plate shaking control was slightly lower than that extracted from the other treatments. Specifically, the average RNA content from these samples was 144.70 ng/µL, while it was 211.16 ng/µL for all other samples combined. This could be due to the back-and-forth transfer of the sample to a 96-well plate and then back into a tube for RNA extraction, chemical degradation of sample, or other loss of sample during processing. While liquid and cell/tissue samples are fully contained within tubes during the processing in sonication, the 96-well plate mechanical processing is not as contained and can more easily result in sample spillage, even when using plate covers or low mixing speeds.
      260/280 ratios were also assessed to determine the ratio of nucleic acids, which have an absorbance maximum of 260 nm, in comparison to residual chemical contaminants, such as phenol, or other extraction reagents. Samples were assessed without any purification step. Ratios for all samples were above ∼2.0, which is the standard of purity for RNA (Figure 6B).
      260/230 ratios were analyzed in conjunction with 260/280 ratios to assess purity of solutions. A pure ratio should again be ∼2.0. However, the unpurified samples all had slightly lower values than this ideal (Figure 6C). This could be due to contaminants such as EDTA and phenols.
      Lastly, the last quality control metric that was assessed was the RNA Integrity Number (RIN). Fully intact RNA has a RIN of ∼10. While all the control samples and sonication samples had RINs of 10, the chemical and orbital plate shaking samples ranged from 8-10 (Figure 6D).

      RNA Stress Marker Analysis

      After analyzing the content and quality of extracted RNA from the samples, a stress marker analysis was performed using a set of previously characterized targets that indicate stress-response in MDA-MB-231 [
      • Chénais B.
      • Cornec M.
      • Dumont S.
      • Marchand J.
      • Blanckaert V.
      Transcriptomic Response of Breast Cancer Cells MDA-MB-231 to Docosahexaenoic Acid: Downregulation of Lipid and Cholesterol Metabolism Genes and Upregulation of Genes of the Pro-Apoptotic ER-Stress Pathway.
      ]. Previous work has determined particular changes that are associated with the stress response, including increases and decreases in various targets (Supplemental Table S1).
      Markers determined to have the highest predictive value for MDA-MB-231 stress response were selected. These factors control for migration and invasion, apoptosis induction, proliferation inhibition, cell adhesion, signal transduction, and ER stress. Assessment was performed using RT-qPCR (Figure 6E) and extracted data was presented in terms of ΔCq (Figure 6F).
      In chemical and orbital plate shaking trials, stress response was observed for SERPINE1 and PLAT. However, across all tested samples, no other transcriptional changes characteristic of stress response were observed for INHBE, FLRT1, HSPA5, or ECM2. Notably, very minimal transcriptional change was observed across all of the sonicated samples.

      Discussion

      This study has explored the use of ultrasound frequency sonication in the dissociation of ex vivo tissues and disaggregation of cellular aggregates. While 30 minutes was determined to be the best tested timepoint, the method can work well at shorter time intervals as well. While bath temperature was shown to have little effect on dissociation efficacy, degassing significantly decreased the dissociation of tissues, suggesting a role for bubble formation and cavitation in the production of mechanical forces that contribute to dissociation efficiency.
      A 24-plex floating tube holder was used to investigate the high-throughput potential of this processing method. The sonication conditions were found to be consistently statistically significantly better at dissociating the tissues than the comparative mechanical methods. The sonication-only condition had over a 43% improvement in dissociation when compared to the orbital plate shaking-only condition.
      The use of 1% collagenase as a chemical dissociant improved the utility of orbital plate shaking processing to an average of 37 ± 20%. However, sonication combined with chemical dissociation was found to have over a 34% improvement, with half of the standard deviation. These results show promise for sonication as a more uniform, high-throughput processing method in the dissociation of tissues.
      A plethora of studies were subsequently conducted on MDA-MB-231 cellular aggregates. First, cellular recovery was assessed using hemocytometry before and after treatments with sonication, a chemical treatment of TryPLE Express, an orbital plate shaking treatment, or a hybrid chemical and orbital plate shaking treatment. Cellular recovery was not found to be significantly affected in any of the groups.
      Aggregate size and distribution were then assessed. Cells stained with membrane and nuclear dyes were analyzed by random sampling of multiple interrogation regions and processing with ImageJ. Average aggregate size was recorded for each treatment, factoring in the size of every aggregate in every ROI of 3 or more cells. While none of the orbital plate shaking treated samples were found to have statistically significant differences in average aggregate size when compared to a control, all the sonication samples were found to have statistically significantly smaller average aggregate sizes.
      To portray size distribution more comprehensively, aggregate size distribution histograms were created to represent the relative proportion of singlet (∼10-15 µm), doublet (∼15-20 µm), small (20-30 µm), medium (30-100 µm), large (100-200 µm) and extra-large (200-400 µm) cellular clumps per interrogation area. While most cells were found in singlets and doublets, sonicated samples had statistically more singlet cells, 0 large or extra-large aggregates, and fewer medium sized aggregates.
      While some membrane fragments were observed in sonication images, these were not uncommon amongst orbital plate shaking images either. The relative circularity of membranes was not found to be significantly different in any of the groups. The relative viability, however, was significantly higher for sonication-treated samples at 5, 15, and 30 minutes.
      RNA content was significantly lower in chemical and orbital plate shaking samples in comparison to control and sonication samples. The RIN was also slightly lower for these samples. More transcriptional changes were observed in chemical and orbital plate shaking samples than in those treated with sonication.
      In conclusion, ultrasound frequency sonication represents a possible new technique that could easily be applied to high-throughput tissue dissociation of micrometer to centimeter size tissue and cellular aggregate samples. Sonication results in significantly greater dissociation efficacy and viability, smaller average aggregate sizes, more singlet cells and less large aggregates when compared to orbital plate shaking. Additionally, the method retains good cellular recovery and membrane integrity. When examining its translational potential in cellular nucleic acid analysis, the sonication method was found to yield high quality RNA and have less transcriptional expression changes with respect to a control than chemical and orbital plate shaking treatments. Further research must be conducted to characterize the ideal parameters for sonication-based dissociation amongst different tissue and cell types and optimize acoustic parameters.

      Uncited Tables

      Table 1
      Table 1Comparison of Various Attributes Across Treatment Methods
      Required TimeDissociation EfficiencyViabilityUniformityHigh-Throughput Capability
      Sonication Only30 minutesHighHighHigh24+
      Orbital Plate Shaking Only> 30 minutesLowIntermediate to HighLow96
      Chemical Treatment Only (1% Collagenase at 37 C)> 1 hourIntermediateIntermediate to HighLowN/A
      Chemical and Orbital Plate Shaking Treatment> 30 minutesIntermediateLow to HighLow to Intermediate96
      Chemical and Sonication Treatment30 minutesHighIntermediate to HighHigh24+

      Funding

      We would like to gratefully acknowledge PerkinElmer Inc. for providing financial support for this research.

      Declaration of Competing Interest

      The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:
      Anubhav Tripathi reports financial support was provided by PerkinElmer Inc. Anubhav Tripathi reports a relationship with PerkinElmer Inc that includes: consulting or advisory and funding grants.

      Acknowledgments

      We would like to acknowledge the Brown University Graduate School and the Brown Center For Biomedical Engineering for providing facilities and financial support for this work. Furthermore, we would like to thank the Flow Cytometry Core and the Leduc Bioimaging Core facilities for providing the flow cytometer and microscope used in this study. We would also like to gratefully acknowledge the laboratory of Vikas Srivastava and Gavin Mays for donating the MDA-MB-231 cancer cells used in this study. Figures 1 and 3 A were created using BioRender.com.
      Figure 1
      Figure 1Schematic Representation of Sonication Protocol.
      Schematic representation of the sonication protocol used in the described experiments. A floating tube holder was used in sample processing in this instance; however, alternative suspension apparatuses may also be used. Tissue or cell samples are immersed in a volume of isotonic solution (PBS recommended) sufficient to submerge the entire tissue section and are then placed in the tube holder. Processing is then completed in the sonication bath, commonly available in many lab settings. With the sonication parameters used in these experiments, 30-minute processing time is recommended for best results. Processing can be completed at room temperature, but a slightly elevated temperature of 37°C is recommended for preserving cellular viability.
      Figure 2
      Figure 2Assessment of Sonication-Facilitated Tissue Dissociation.
      2A) Average percent dissociation for sonicated tissues across 5-, 15-, and 30-minute timepoints. Improved results were obtained at 30 minutes with a significance of * p < 0.05. 2B) represents the same time course experiment conducted with simultaneous degassing, illustrating a less significant increase over time. 2C) illustrates the efficacy of sonication with and without degassing at 30 minutes processing time. Note that results improve with increased processing time. Degassing was observed to attenuate the dissociative capacity of sonication, in accordance with established physical theory. Degassing reduced performance significantly at ** p < 0.01. 2D) demonstrates that the effect of bath temperature on dissociation efficacy. No significant effect on temperature was observed in terms of dissociation, although 37°C is recommended for improved cell viability.
      Figure 3
      Figure 3– Investigation of High-Throughput Sample Processing with Different Methods.
      3A) schematic of orbital plate shaking, chemical, chemical and orbital plate shaking and chemical and sonication protocols. Datapoints in B-E represent average values taken for each placement over numerous trials. 3B) illustrates pure orbital plate shaking and chemical and orbital plate shaking hyrbid dissociation at 30 minutes by row in 96 well plate while 3C) shows the data by column. 3D) shows pure sonication and chemical/sonication dissociation after 30 minutes by row in the tube holder while 3E) shows the data by column. 3F) represents a comparison of all extracted datapoints across all tested treatments at 30 minutes. Sonication based processing is notably more consistent. See Supplemental Figure 1 for visual schematics of averaged results.
      Figure 3
      Figure 3– Investigation of High-Throughput Sample Processing with Different Methods.
      3A) schematic of orbital plate shaking, chemical, chemical and orbital plate shaking and chemical and sonication protocols. Datapoints in B-E represent average values taken for each placement over numerous trials. 3B) illustrates pure orbital plate shaking and chemical and orbital plate shaking hyrbid dissociation at 30 minutes by row in 96 well plate while 3C) shows the data by column. 3D) shows pure sonication and chemical/sonication dissociation after 30 minutes by row in the tube holder while 3E) shows the data by column. 3F) represents a comparison of all extracted datapoints across all tested treatments at 30 minutes. Sonication based processing is notably more consistent. See Supplemental Figure 1 for visual schematics of averaged results.
      Figure 4
      Figure 4MDA-MB-231 Recovery and Aggregate Size Assessments.
      4A) represents MDA-MB-231 average cellular recovery following different treatments. Recovery is a metric of total cell retrieval and would only decline in the event of cellular lysis or sample loss. 4B) represents the average aggregate size of MDA-MB-231 cell aggregates following exposure to different treatments. While none of the orbital plate shaking-only treatments significantly reduced aggregate size when compared to a control, the sonication-only treatments all exhibited significances of *** p < 0.001 or greater. 4C) illustrates size distribution histograms of MDA-MB-231 aggregates following exposure to different treatments. Orbital plate shaking treatments had significantly fewer single cells with *** p < 0.001. This enables visualization of the total number of cells per interrogation region in different aggregate sizes from singlets to extra-large aggregates. Thresholding information for the histogram size bins is available in the methods section.
      Figure 5
      Figure 5Cellular Integrity Assays in MDA-MB-231.
      5A) representative images illustrating cellular morphology for all treatment groups. The top row illustrates orbital plate shaking treated cells after 5, 15, and 30 minutes of treatment, respectively. The second row illustrates sonication treated cells after 5, 15 and 30 minutes of treatment, respectively. Magnification bars represent 100 micrometers. Dissociation efficacy is notably significantly improved when using sonication-only forces as opposed to orbital plate shaking-only forces. 5B) illustrates membrane characteristics across the various treatments and timepoints, with no notable significant difference. 5C) shows viability characteristics across the same treatments and timepoints. Orbital plate shaking treatments exhibited a significant reduction in viability in comparison to sonication treatments with * p < 0.05 at 5 and 15 minutes and ** p < 0.01 at 30 minutes.
      Figure 6
      Figure 6Extracted RNA Content and Quality Analysis. RNA Stress Marker Analysis with RT-qPCR.
      6A) shows a graph of extracted RNA content from cell-count-standardized starting samples. 6B) represents the 260/280 ratios of extracted RNA before sample purification. 6C) shows 260/230 ratios of extracted RNA before sample purification. The dotted blue line in 6B and 6C represents ideal results for purified RNA. 6D) represents the RNA Integrity Number (RIN) for each treatment group. All groups have perfect RNA integrity with the exception of the chemical and orbital plate shaking treatment group. 6E) shows representative amplification trends from a single RT-qPCR run while 6F) illustrates expression changes with respect to a baseline control. The chemical and orbital plate shaking control was processed for a total of 30 minutes. Information on markers and established stress expression patterns can be found in Supplemental Table 1.

      Appendix. Supplementary materials

      References

        • Welch E.C.
        • Tripathi A.
        Preparation of Tissues and Heterogeneous Cellular Samples for Single-Cell Analysis.
        Sample Preparation Techniques for Chemical Analysis. IntechOpen, 2021
        • van den Brink S.C.
        • et al.
        Single-cell sequencing reveals dissociation-induced gene expression in tissue subpopulations.
        Nat. Methods. Oct. 2017; 14https://doi.org/10.1038/nmeth.4437
        • Denisenko E.
        • et al.
        Systematic assessment of tissue dissociation and storage biases in single-cell and single-nucleus RNA-seq workflows.
        Genome Biol. Dec. 2020; 21: 130https://doi.org/10.1186/s13059-020-02048-6
        • Welch E.C.
        • Yu H.
        • Barabino G.
        • Tapinos N.
        • Tripathi A.
        Electric-Field Facilitated Rapid and Efficient Dissociation of Tissues Into Viable Single Cells.
        Sci. Rep. 2022; https://doi.org/10.1038/s41598-022-13068-6
        • Karasu Benyes Y.
        • Welch E.C.
        • Singhal A.
        • Ou J.
        • Tripathi A.
        A Comparative Analysis of Deep Learning Models for Automated Cross-Preparation Diagnosis of Multi-Cell Liquid Pap Smear Images.
        Diagnostics. Jul. 2022; 12: 1838https://doi.org/10.3390/diagnostics12081838
        • Welch E.C.
        • Yu H.
        • Tripathi A.
        Optimization of a Clinically Relevant Chemical-Mechanical Tissue Dissociation Workflow for Single-Cell Analysis.
        Cell. Mol. Bioeng. Mar. 2021; https://doi.org/10.1007/s12195-021-00667-y
        • Donlin L.T.
        • et al.
        Methods for high-dimensional analysis of cells dissociated from cryopreserved synovial tissue.
        Arthritis Res. Ther. Dec. 2018; 20: 139https://doi.org/10.1186/s13075-018-1631-y
        • Achim K.
        • et al.
        High-throughput spatial mapping of single-cell RNA-seq data to tissue of origin.
        Nat. Biotechnol. May 2015; 33https://doi.org/10.1038/nbt.3209
        • Chen X.
        • Teichmann S.A.
        • Meyer K.B.
        From Tissues to Cell Types and Back: Single-Cell Gene Expression Analysis of Tissue Architecture.
        Annu. Rev. Biomed. Data Sci. Jul. 2018; 1: 29-51https://doi.org/10.1146/annurev-biodatasci-080917-013452
      1. E. C. Welch and A. Tripathi, “Electrical Dissociation Of Tissue Samples Into Single Cells And/Or Smaller Groups Of Cells,” US Patent App. 17/721,618, 2021.

      2. E. C. Welch and A. Tripathi, “Electrical Dissociation Of Tissue Samples Into Single Cells And/Or Smaller Groups Of Cells,” WO Patent App. 22/25,451, 2022.

        • Zhang L.
        • Belova V.
        • Wang H.
        • Dong W.
        • Möhwald H.
        Controlled Cavitation at Nano/Microparticle Surfaces.
        Chem. Mater. Apr. 2014; 26: 2244-2248https://doi.org/10.1021/cm404194n
        • Weller R.N.
        • Brady J.M.
        • Bernier W.E.
        Efficacy of ultrasonic cleaning.
        J. Endod. Sep. 1980; 6: 740-743https://doi.org/10.1016/S0099-2399(80)80185-3
        • Mason T.J.
        Ultrasonic cleaning: An historical perspective.
        Ultrason. Sonochem. Mar. 2016; 29: 519-523https://doi.org/10.1016/j.ultsonch.2015.05.004
        • Lim A.
        Membrane fouling and cleaning in microfiltration of activated sludge wastewater.
        J. Memb. Sci. May 2003; 216: 279-290https://doi.org/10.1016/S0376-7388(03)00083-8
        • Marković S.
        • Mitrić M.
        • Starčević G.
        • Uskoković D.
        Ultrasonic de-agglomeration of barium titanate powder.
        Ultrason. Sonochem. Jan. 2008; 15: 16-20https://doi.org/10.1016/j.ultsonch.2007.07.008
        • Sauter C.
        • Emin M.A.
        • Schuchmann H.P.
        • Tavman S.
        Influence of hydrostatic pressure and sound amplitude on the ultrasound induced dispersion and de-agglomeration of nanoparticles.
        Ultrason. Sonochem. Apr. 2008; 15: 517-523https://doi.org/10.1016/j.ultsonch.2007.08.010
        • Park C.
        • et al.
        Dispersion of single wall carbon nanotubes by in situ polymerization under sonication.
        Chem. Phys. Lett. Oct. 2002; 364: 303-308https://doi.org/10.1016/S0009-2614(02)01326-X
        • Huang W.
        • Lin Y.
        • Taylor S.
        • Gaillard J.
        • Rao A.M.
        • Sun Y.-P.
        Sonication-Assisted Functionalization and Solubilization of Carbon Nanotubes.
        Nano Lett. Mar. 2002; 2: 231-234https://doi.org/10.1021/nl010083x
        • Dhumal R.S.
        • Biradar S.V.
        • Paradkar A.R.
        • York P.
        Ultrasound Assisted Engineering of Lactose Crystals.
        Pharm. Res. Dec. 2008; 25: 2835-2844https://doi.org/10.1007/s11095-008-9653-9
        • Öztürk S.
        • Akata B.
        Oriented assembly and nanofabrication of zeolite A monolayers.
        Microporous Mesoporous Mater. Dec. 2009; 126: 228-233https://doi.org/10.1016/j.micromeso.2009.06.010
        • Gilca I.A.
        • Popa V.I.
        • Crestini C.
        Obtaining lignin nanoparticles by sonication.
        Ultrason. Sonochem. Mar. 2015; 23: 369-375https://doi.org/10.1016/j.ultsonch.2014.08.021
        • Obermaier C.
        • et al.
        Free-flow isoelectric focusing of proteins remaining in cell fragments following sonication of thyroid carcinoma cells.
        Electrophoresis. Jun. 2005; 26: 2109-2116https://doi.org/10.1002/elps.200410422
        • Guex A.G.
        • Di Marzio N.
        • Eglin D.
        • Alini M.
        • Serra T.
        The waves that make the pattern: a review on acoustic manipulation in biomedical research.
        Mater. Today Bio. Mar. 2021; 10100110https://doi.org/10.1016/j.mtbio.2021.100110
        • Mohanty S.
        • Khalil I.S.M.
        • Misra S.
        Contactless acoustic micro/nano manipulation: a paradigm for next generation applications in life sciences.
        Proc. R. Soc. A Math. Phys. Eng. Sci. Nov. 2020; 47620200621https://doi.org/10.1098/rspa.2020.0621
        • Ding X.
        • et al.
        On-chip manipulation of single microparticles, cells, and organisms using surface acoustic waves.
        Proc. Natl. Acad. Sci. Jul. 2012; 109https://doi.org/10.1073/pnas.1209288109
        • Guo F.
        • et al.
        Three-dimensional manipulation of single cells using surface acoustic waves.
        Proc. Natl. Acad. Sci. Feb. 2016; 113: 1522-1527https://doi.org/10.1073/pnas.1524813113
        • Ziff R.M.
        New Solutions To The Fragmentation Equation.
        J. Phys. A. Math. Gen. 1991; 24: 2821
        • Edwards B.F.
        • Cai M.
        • Han H.
        Rate equation and scaling for fragmentation with mass loss.
        Phys. Rev. A. May 1990; 41: 5755-5757https://doi.org/10.1103/PhysRevA.41.5755
        • Rwei S.P.
        • Manas-Zloczower I.
        • Feke D.L.
        Observation of carbon black agglomerate dispersion in simple shear flows.
        Polym. Eng. Sci. Jun. 1990; 30: 701-706https://doi.org/10.1002/pen.760301202
        • Asanuma Y.
        • Faizal F.
        • Khairunnisa M.P.
        • Lenggoro I.W.
        Deagglomeration of spray-dried submicron particles by low-power aqueous sonication.
        Adv. Powder Technol. Apr. 2022; 33103543https://doi.org/10.1016/j.apt.2022.103543
        • Bałdyga J.
        • Makowski Ł.
        • Orciuch W.
        • Sauter C.
        • Schuchmann H.P.
        Deagglomeration processes in high-shear devices.
        Chem. Eng. Res. Des. Dec. 2008; 86: 1369-1381https://doi.org/10.1016/j.cherd.2008.08.016
        • Tey W.Y.
        • Alehossein H.
        • Qin Z.
        • Lee K.M.
        • Kang H.S.
        • Lee K.Q.
        On stability of time marching in numerical solutions of rayleigh-plesset equation for ultrasonic cavitation.
        IOP Conf. Ser. Earth Environ. Sci. Mar. 2020; 463012117https://doi.org/10.1088/1755-1315/463/1/012117
        • Khairunnisa M.P.
        • Faizal F.
        • Miyazawa E.
        • Masuda K.
        • Tsukada M.
        • Lenggoro I.W.
        Detachment of Submicron Particles from Substrates Using the Suspension-Assisted Ultrasonic Method.
        J. Chem. Eng. JAPAN. Apr. 2021; 54: 135-143https://doi.org/10.1252/jcej.16we319
        • German R.M.
        Coordination number changes during powder densification.
        Powder Technol. Feb. 2014; 253: 368-376https://doi.org/10.1016/j.powtec.2013.12.006
        • Sztilkovics M.
        • et al.
        Single-cell adhesion force kinetics of cell populations from combined label-free optical biosensor and robotic fluidic force microscopy.
        Sci. Rep. Dec. 2020; 10: 61https://doi.org/10.1038/s41598-019-56898-7
        • Sancho A.
        • Vandersmissen I.
        • Craps S.
        • Luttun A.
        • Groll J.
        A new strategy to measure intercellular adhesion forces in mature cell-cell contacts.
        Sci. Rep. May 2017; 7: 46152https://doi.org/10.1038/srep46152
        • Junatas K.L.
        • et al.
        Stereological analysis of size and density of hepatocytes in the porcine liver.
        J. Anat. Apr. 2017; 230https://doi.org/10.1111/joa.12585
        • Wilson Z.E.
        • et al.
        Inter-individual variability in levels of human microsomal protein and hepatocellularity per gram of liver.
        Br. J. Clin. Pharmacol. Oct. 2003; 56https://doi.org/10.1046/j.1365-2125.2003.01881.x
        • Crowley L.C.
        • et al.
        Dead Cert: Measuring Cell Death.
        Cold Spring Harb. Protoc. Dec. 2016; 2016pdb.top070318https://doi.org/10.1101/pdb.top070318
        • Chénais B.
        • Cornec M.
        • Dumont S.
        • Marchand J.
        • Blanckaert V.
        Transcriptomic Response of Breast Cancer Cells MDA-MB-231 to Docosahexaenoic Acid: Downregulation of Lipid and Cholesterol Metabolism Genes and Upregulation of Genes of the Pro-Apoptotic ER-Stress Pathway.
        Int. J. Environ. Res. Public Health. May 2020; 17: 3746https://doi.org/10.3390/ijerph17103746
        • Ruiz-Valdepeñas Montiel V.
        • Sempionatto J.R.
        • Campuzano S.
        • Pingarrón J.M.
        • Fernández de Ávila B.Esteban
        • Wang J.
        Direct electrochemical biosensing in gastrointestinal fluids.
        Anal. Bioanal. Chem. Jul. 2019; 411https://doi.org/10.1007/s00216-018-1528-2
        • Qiu X.
        • et al.
        Microfluidic channel optimization to improve hydrodynamic dissociation of cell aggregates and tissue.
        Sci. Rep. Dec. 2018; 8https://doi.org/10.1038/s41598-018-20931-y
        • Evani S.J.
        • Prabhu R.G.
        • Gnanaruban V.
        • Finol E.A.
        • Ramasubramanian A.K.
        Monocytes mediate metastatic breast tumor cell adhesion to endothelium under flow.
        FASEB J. Aug. 2013; 27: 3017-3029https://doi.org/10.1096/fj.12-224824
        • Huang Z.
        • Yu P.
        • Tang J.
        Characterization of Triple-Negative Breast Cancer MDA-MB-231 Cell Spheroid Model.
        Onco. Targets. Ther. Jun. 2020; 13: 5395-5405https://doi.org/10.2147/OTT.S249756