pubmed.ncbi.nlm.nih.gov

Deformability-based cell selection with downstream immunofluorescence analysis - PubMed

  • ️Fri Jan 01 2016

Deformability-based cell selection with downstream immunofluorescence analysis

Josephine Shaw Bagnall et al. Integr Biol (Camb). 2016.

Abstract

Mechanical properties of single cells have been shown to relate to cell phenotype and malignancy. However, until recently, it has been difficult to directly correlate each cell's biophysical characteristics to its molecular traits. Here, we present a cell sorting technique for use with a suspended microchannel resonator (SMR), which can measure biophysical characteristics of a single cell based on the sensor's record of its buoyant mass as well as its precise position while it traverses through a constricted microfluidic channel. The measurement provides information regarding the amount of time a cell takes to pass through a constriction (passage time), as related to the cell's deformability and surface friction, as well as the particular manner in which it passes through. In the method presented here, cells of interest are determined based on passage time, and are collected off-chip for downstream immunofluorescence imaging. The biophysical single-cell SMR measurement can then be correlated to the molecular expression of the collected cell. This proof-of-principle is demonstrated by sorting and collecting tumor cells from cell line-spiked blood samples as well as a metastatic prostate cancer patient blood sample, identifying them by their surface protein expression and relating them to distinct SMR signal trajectories.

PubMed Disclaimer

Figures

Fig. 1
Fig. 1. Cell sorting and collection method for biophysical-molecular correlation

A) A diagram demonstrating a single cell passing through a suspended microchannel resonator (SMR), squeezing through a constriction. B) An example of the raw resonance frequency signal obtained from a typical cultured tumor cell (an H1650 lung cancer cell in this case) as it transits through the SMR sensor. The height of the frequency shift is proportional to the buoyant mass of the cell, while the width of peak, as it transits from position 3 to position 4, corresponds to the passage time of the cell through the constriction. The numbers in (B) correspond to the position of the cell as labeled in (A). C) A schematic diagram of the cell sorting and collection technique, showing that cells are stained and measured in an SMR. Once the software detects a passage time in the range of interest, it stops the fluid flow in the device, allowing for subsequent collection of the cell off-chip, into a 96-well plate for downstream fluorescence microscopy.

Fig. 2
Fig. 2. Healthy donor versus metastatic prostate cancer patient blood sample

A) Passage time vs. buoyant mass measurements of a healthy donor blood sample, after having been depleted of most of the erythrocytes and leukocytes. An equivalent volume of ~280 µL of blood was measured. B) Passage time vs. buoyant mass measurements of a metastatic prostate cancer patient blood sample, having been depleted of most of the erythrocytes and leukocytes. An equivalent volume of ~360 µL of blood was measured. The difference in the number of cells measured was due to different efficiencies of the preprocessing steps in depleting the erythrocytes and leukocytes from the samples (Materials and Methods). Colors correspond to the density of data points, with red being the highest and blue being the lowest density.

Fig. 3
Fig. 3. Validation of single tumor cell detection and collection in 96-well plate

A) Passage time vs. buoyant mass measurements of H1650 cells spiked into enriched mononuclear blood cells. The cells included in the boxed region (> 50pg buoyant mass & > 0.01 s passage time) were imaged on-chip (B) as well as collected off-chip for additional verification (C). Gray dots outside of the boxed region represent blood cells that were also measured in the SMR during the experiment. B) Images taken on-chip, in the exit channel of the SMR, immediately after custom software detected a long passage time signal. It was confirmed that all detected SMR signals falling into the boxed region of interest (A) corresponded to an H1650 tumor cell by immunofluorescence imaging (EpCAM/CDH11+ and CD45−). C) Each detected H1650 cell was successfully collected off-chip in a 96-well plate. Based on additional immunofluorescence imaging of each well, a tumor cell was identified in each well (EpCAM/CDH11+ and CD45−), along with a few leukocytes (CD45+ and EpCAM/CDH11−).

Fig. 4
Fig. 4. Spiked blood samples and SMR peak shape correlation with immunofluorescence staining

A) A blood sample spiked with LNCaP cells was measured in the SMR. The labeled measurement events were collected off-chip (L1–L7). B) A blood sample spiked with PC3 cells was measured in the SMR. The labeled measurement events were collected off-chip (P1–P9). In both panels, the dotted area corresponds to the region of interest (> 50 pg and > 0.01 s). C) Raw SMR frequency signal versus time acquired for each measurement event, demonstrating different peak shapes for different types of cells or particles being measured. Labels correspond to those in (D) as well as in (A). Note that the axes are different for each peak for better visualization of the details of each peak shape. D) Immunofluorescence images of cells corresponding different peak shapes from (C), with L2 and L3 being LNCaP cells, L7 being an activated neutrophil, and WBC being a typical white blood cell. Note that the intensity of EpCAM for L3 was reduced to avoid saturation, as it had a much higher intensity than all other imaged cells. E) The position of the cell or particle in the SMR as it passes through the constriction (−1 corresponds to the tip, and 0 corresponds to the base of the cantilever) is plotted versus normalized time (0 corresponds to when the cell reaches the tip of the cantilever and 1 corresponds to the time the cell exits from the cantilever) for each of the SMR signals that were collected for visual analysis. F) The median positions of each measured signal within each bin of normalized time are plotted as a heat map. A dendrogram derived from hierarchical clustering analysis is shown to the right of the heat map, demonstrating that the shape of the SMR signals can be categorized into different groups, potentially related to the type of particle measured, as identified by fluorescence imaging.

Fig. 5
Fig. 5. SMR biophysical measurement correlation with quantified immunofluorescence levels

A–C) H1650 cells stained for EpCAM were measured by the SMR and individually collected off-chip for assessment by fluorescence microscopy. A) Passage time versus buoyant mass for H1650 cells stained for EpCAM (N = 46 cells, simple linear regression R2 = 0.63). B) From the same data set as in (A) but showing EpCAM fluorescence intensity versus buoyant mass (simple linear regression R2 = 0.14). C) From the same data set as in (A) but showing EpCAM fluorescence intensity versus passage time (simple linear regression R2 = 0.02). D–F) The fluorescence intensities of H1650 and HCC827 cells stained for surface protein expression were assessed on-chip on an SMR via a PMT. D) Fluorescence intensity versus buoyant mass of H1650 cells stained for EpCAM (N = 400, R2 = 0.13). E) Fluorescence intensity versus buoyant mass of H1650 cells stained for CD45, as a control for nonspecific antibody binding (N = 337, R2 = 0.31). Note that cells having fluorescence intensities below the limit of detection were not included. F) Fluorescence intensity versus buoyant mass of HCC827 cells stained for EpCAM (N = 401, R2 = 0.35).

Fig. 6
Fig. 6. Metastatic prostate cancer patient blood sample

A) Passage time versus buoyant mass measurements of a metastatic prostate cancer patient blood sample, after pre-preprocessing in the CTC-iChip. An equivalent total of 1.8 mL of blood was measured in the SMR. The region of interest (buoyant mass greater than 50pg and passage time greater than 10ms) is shown as a dotted gray box. B) One cell that fell in the region of interest that corresponded to a tumor cell-like frequency signal in the SMR as well was collected off-chip and imaged. Fluorescence images indicate that the SMR measurement correlated to a CTC (EpCAM positive, CD45 negative). Images of a white blood cell (WBC) from the remainder of the sample are shown for comparison. False color overlays were applied to composite fluorescence images. C) Hierarchical clustering analysis was performed for the SMR peak trajectories of cells collected from the patient sample, as demarcated in (A), combined with cells collected in Fig. 4. The patient CTC (Pa1) clusters together with other cells from prostate tumor cell lines.

Similar articles

Cited by

References

    1. Guck J, et al. Optical deformability as an inherent cell marker for testing malignant transformation and metastatic competence. Biophys. J. 2005;88:3689–3698. - PMC - PubMed
    1. Tse HTK, et al. Quantitative diagnosis of malignant pleural effusions by single-cell mechanophenotyping. Sci. Transl. Med. 2013;5:212ra163. - PubMed
    1. Byun S, et al. Characterizing deformability and surface friction of cancer cells. Proc. Natl. Acad. Sci. 2013;110:7580–7585. - PMC - PubMed
    1. Ekpenyong AE, et al. Viscoelastic properties of differentiating blood cells are fate- and function-dependent. PLoS One. 2012;7:e45237. - PMC - PubMed
    1. Remmerbach TW, et al. Oral cancer diagnosis by mechanical phenotyping. Cancer Res. 2009;69:1728–1732. - PubMed

Publication types

MeSH terms