June 2020
Volume 61, Issue 7
Open Access
ARVO Annual Meeting Abstract  |   June 2020
Super-resolution microscopy reveals mitochondria morphology variations under different cellular metabolic activities
Author Affiliations & Notes
  • Zheyuan Zhang
    BIOMEDICAL ENGINEERING, NORTHWESTERN UNIVERSITY, Evanston, Illinois, United States
  • Yang Zhang
    BIOMEDICAL ENGINEERING, NORTHWESTERN UNIVERSITY, Evanston, Illinois, United States
  • Nader Sheibani
    EPARTMENT OF OPHTHALMOLOGY & VISUAL SCIENCES, University of Wisconsin, Madison, Wisconsin, United States
  • Hao Zhang
    BIOMEDICAL ENGINEERING, NORTHWESTERN UNIVERSITY, Evanston, Illinois, United States
  • Footnotes
    Commercial Relationships   Zheyuan Zhang, None; Yang Zhang, None; Nader Sheibani, None; Hao Zhang, None
  • Footnotes
    Support  CBET-1706642, CBET-1604531, EFRI-1830969, and EEC-1530734 R01EY026078 and R01EY029121
Investigative Ophthalmology & Visual Science June 2020, Vol.61, 912. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Zheyuan Zhang, Yang Zhang, Nader Sheibani, Hao Zhang; Super-resolution microscopy reveals mitochondria morphology variations under different cellular metabolic activities. Invest. Ophthalmol. Vis. Sci. 2020;61(7):912.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : Metabolism are essential for cellular homeostasis and functions. The current understanding of metabolism associated disease, however, is obtained from biochemical assays that requires large population of cells. We propose spectroscopic single-molecule localization microscopy (sSMLM) to visualize complex biomolecular interactions related to metabolic activities. To minimize color contamination in identifying different molecular labels in sSMLM, we aim to develop machine-learning (ML) based spectral classification method to image and further explore morphological variations of mitochondria and protein interactions at different cellular metabolic statuses as model systems to promote the understanding of irregular metabolic activity involved in diabetic retinopathy.

Methods : To train ML network, we prepared three single-dye labeled samples (AF647, CF660, or CF680) as the ground truths. The network consists of four hidden layers with 128, 128, 64, and 32 neurons in each respective layer (Fig 1.a). We further set softmax layer with three neurons as the output layers. We compared drug-induced mitochondria morphological variations between normal and STS-induced apoptosis COS-7 cells as well as intrinsic mitochondria morphological differences in pericyte and endothelial cells.

Results : The neural network converged at 99.10% accuracy on training data and about 99% accuracy on test data. We showed that the morphology of mitochondria exhibited significant change between STS drug-treated and control groups. Specifically, the mitochondria remained elongated with relatively large sizes in healthy cells while they switch to isolated round morphology with smaller size in STS treated group. In addition, different cell types (pericyte vs. endothelial cells) also showed significant morphological changes of mitochondria, which is consistent with their different metabolic rates reported in literature.

Conclusions : ML-based classification method in sSMLM minimizes the misidentification rate by 5-fold with accuracy up to 99%. We found mitochondria fission into smaller round-shaped morphology in STS-induced apoptosis as well as in pericytes. These techniques potentially permit the understanding metabolic pathways associated with mitochondria dysfunctions in diseases, such as diabetic retinopathy, at molecular level.

This is a 2020 ARVO Annual Meeting abstract.

 

Fig. 1: (a) neural network structure, (b) reconstruction result through ML

Fig. 1: (a) neural network structure, (b) reconstruction result through ML

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×