June 2022
Volume 63, Issue 7
Open Access
ARVO Annual Meeting Abstract  |   June 2022
Automatic image skeletonization to characterize ZO-1 distribution in the corneal endothelium following hypothermia and oxidative stress
Author Affiliations & Notes
  • Tapanmitra Ravi
    Computer Science & Engineering, Dayananda Sagar College of Engineering, Bangalore, Karnataka, India
  • M.Y. Thanuja
    Chemical Engineering, Siddaganga Institute of Technology, Tumkur, Karnataka, India
  • P. S. Shilpashree
    Electronics and Communication Engineering, Siddaganga Institute of Technology, Tumkur, Karnataka, India
  • Anupama C
    Biotechnology, Siddaganga Institute of Technology, Tumkur, Karnataka, India
  • Sreesha Srinivasan Kuruvadi
    Data Science, Indiana University Bloomington, Bloomington, Indiana, United States
  • Sudhir H Ranganath
    Chemical Engineering, Siddaganga Institute of Technology, Tumkur, Karnataka, India
  • Sangly P Srinivas
    Optometry, Indiana University Bloomington, Bloomington, Indiana, United States
  • Footnotes
    Commercial Relationships   Tapanmitra Ravi None; M.Y. Thanuja None; P. S. Shilpashree None; Anupama C None; Sreesha Kuruvadi None; Sudhir Ranganath None; Sangly Srinivas None
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2022, Vol.63, 2736 – A0225. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Tapanmitra Ravi, M.Y. Thanuja, P. S. Shilpashree, Anupama C, Sreesha Srinivasan Kuruvadi, Sudhir H Ranganath, Sangly P Srinivas; Automatic image skeletonization to characterize ZO-1 distribution in the corneal endothelium following hypothermia and oxidative stress. Invest. Ophthalmol. Vis. Sci. 2022;63(7):2736 – A0225.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : ZO-1 binds to Claudins at the apical junctional complex as an intracellular scaffold protein. Thus, it manifests contiguously at the cellular periphery in the endothelium with an intact barrier function. However, stressful stimuli, such as hypothermia, oxidative stress, and allograft rejection, break down the perijunctional actomyosin ring (PAMR) concomitant with the destruction of ZO-1 organization (insets in Fig.). This study aims to quantify such disruptions of ZO-1 through combined immunocytochemistry, confocal microscopy, and automatic image skeletonization.

Methods : Porcine eyes were subjected to hypothermia (4 °C; 3-7 days) or oxidative stress (100 µM; H2O2 for 60 min). Next, images of PAMR and ZO-1 were registered by immunocytochemistry combined with confocal microscopy at the focal plane of ZO-1. The images were then subjected to automatic segmentation by (U-Net)-based deep learning workflow to delineate ZO-1 distribution at the cellular periphery. Finally, the segmented images were skeletonized using the scikit-image (python) implementation of the Zha84 method, and the corresponding outputs were analyzed for estimating the distribution of branch lengths.

Results : In untreated cells, ZO-1 was contiguous with tricellular junctions. At moderate hypothermic stress, the linear edges were disrupted along with the loss of tricellular links. The linear edges broke down into smaller pieces and punctate blobs at severe stress levels. The severity of the impact on ZO-1 generally paralled those on PAMR. The U-Net led to reliable segmentation with an F1 score of 82%, IoU of 77%, and an accuracy of 94%. Subsequent skeletonization and determination of branch lengths showed a skewed distribution profile compared to those for untreated cells (Fig.) with the frequency of appearance of small segments increasing significantly with higher stress levels.

Conclusions : We have developed a workflow to quantify the ZO-1 disruption patterns in response to hypothermia and oxidative stress in the corneal endothelium.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.

 

Distribution of ZO-1 branch lengths in the corneal endothelium for control and after hypothermic and oxidative stresses. The damage to ZO-1 is severe in Grade 2 images compared to those in Grade 1. Inset A: Immunocytochemistry images representing different degrees of damages. Inset B: Mean modal branch lengths of ZO-1 in control and stressed endothelium.

Distribution of ZO-1 branch lengths in the corneal endothelium for control and after hypothermic and oxidative stresses. The damage to ZO-1 is severe in Grade 2 images compared to those in Grade 1. Inset A: Immunocytochemistry images representing different degrees of damages. Inset B: Mean modal branch lengths of ZO-1 in control and stressed endothelium.

×
×

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.

×