Non-destructive, label free identification of cell cycle phase in cancer cells by multispectral microscopy of autofluorescence

Stan Gronthos,Sharon Paton,Abbas Habibalahi,Saabah Mahbub,Jared M Campbell,Ayad G Anwer,Ewa Goldys,Martin Gosnell

doi:10.1186/s12885-019-6463-x

Abstract

BackgroundCell cycle analysis is important for cancer research. However, available methodologies have drawbacks including limited categorisation and reliance on fixation, staining or transformation. Multispectral analysis of endogenous cell autofluorescence has been shown to be sensitive to changes in cell status and could be applied to the discrimination of cell cycle without these steps.MethodsCells from the MIA-PaCa-2, PANC-1, and HeLa cell lines were plated on gridded dishes and imaged using a multispectral fluorescence microscope. They were then stained for proliferating cell nuclear antigen (PCNA) and DNA intensity as a reference standard for their cell cycle position (G1, S, G2, M). The multispectral data was split into training and testing datasets and models were generated to discriminate between G1, S, and G2 + M phase cells. A standard decision tree classification approach was taken, and a two-step system was generated for each line.ResultsAcross cancer cell lines accuracy ranged from 68.3% (MIA-PaCa-2) to 73.3% (HeLa) for distinguishing G1 from S and G2 + M, and 69.0% (MIA-PaCa-2) to 78.0% (PANC1) for distinguishing S from G2 + M. Unmixing the multispectral data showed that the autofluorophores NADH, FAD, and PPIX had significant differences between phases. Similarly, the redox ratio and the ratio of protein bound to free NADH were significantly affected.ConclusionsThese results demonstrate that multispectral microscopy could be used for the non-destructive, label free discrimination of cell cycle phase in cancer cells. They provide novel information on the mechanisms of cell-cycle progression and control, and have practical implications for oncology research.

Highlights

Cell cycle analysis is important for cancer research
Differentiation of cell-cycle phases Analyses were conducted at a single cell level where, based on proliferating cell nuclear antigen (PCNA) patterning and DNA intensity, each cell was labelled as gap 1 (G1), S, or gap 2 (G2) + M
The classifications were done in two stages: first cells not in the process of replication were distinguished from those which were (i.e. G1 phase cells vs S, G2 and M phase cells) as they formed the most distinct initial cluster (area under the curve (AUC) > 75%), S-phase and G2 + M-phase cells were differentiated from one another

Summary

Introduction

Cell cycle analysis is important for cancer research. Dividing cells must pass through the four phases of the cell cycle to duplicate their DNA and separate into two daughter cells. These phases are gap 1 (G1) during which the cell grows, increasing protein content and organelles; synthesis (S) during which nuclear DNA is replicated; gap 2 (G2) a second growth phase; and mitosis (M) where cell division occurs. Cell-cycle phase identification is important for the basic investigation of the growth characteristics of cell lines, especially in cancer research where the cellular mechanisms of cell growth and division may offer therapeutic opportunities [1]. The assessment of cell cycle distribution in tumours may help enable personalised therapy by informing the selection of therapeutic strategies which they are optimally vulnerable to

Objectives

Methods

Results

Discussion

Conclusion