Automated systematic random sampling and Cavalieri stereology of histologic sections demonstrating acute tubular necrosis after cardiac arrest and cardiopulmonary resuscitation in the mouse

• Rumie Wakasaki ^[1] ; Mahaba Eiwaz ^[1] ; Nicholas McClellan ^[1] ; Katsuyuki Matsushita ^[1] ; Kirsti Golgotiu ^[1] ; Kirsti Golgotiu ^[1]
  1. [1] Oregon Health & Science University

     Oregon Health & Science University

     Estados Unidos

     
• Localización: Histology and histopathology: cellular and molecular biology, ISSN-e 1699-5848, ISSN 0213-3911, Vol. 33, Nº. 11, 2018, págs. 1227-1234
• Idioma: inglés
• Títulos paralelos:
  • Automated systematic random sampling and Cavalieri stereology of histologic sections demonstrating acute tubular necrosis after cardiac arrest and cardiopulmonary resuscitation in the mouse
• Enlaces
  • Texto completo
• Resumen

  • A technical challenge in translational models of kidney injury is determination of the extent of cell death. Histologic sections are commonly analyzed by area morphometry or unbiased stereology, but stereology requires specialized equipment. Therefore, a challenge to rigorous quantification would be addressed by an unbiased stereology tool with reduced equipment dependence. We hypothesized that it would be feasible to build a novel software component which would facilitate unbiased stereologic quantification on scanned slides, and that unbiased stereology would demonstrate greater precision and decreased bias compared with 2D morphometry.

    Material and methods. We developed a macro for the widely used image analysis program, Image J, and performed cardiac arrest with cardiopulmonary resuscitation (CA/CPR, a model of acute cardiorenal syndrome) in mice. Fluorojade-B stained kidney sections were analyzed using three methods to quantify cell death: gold standard stereology using a controlled stage and commercially-available software, unbiased stereology using the novel ImageJ macro, and quantitative 2D morphometry also using the novel macro.

    Results. There was strong agreement between both methods of unbiased stereology (bias -0.004±0.006 with 95% limits of agreement -0.015 to 0.007). 2D morphometry demonstrated poor agreement and significant bias compared to either method of unbiased stereology.

    Conclusion. Unbiased stereology is facilitated by a novel macro for ImageJ and results agree with those obtained using gold-standard methods. Automated 2D morphometry overestimated tubular epithelial cell death and correlated modestly with values obtained from unbiased stereology. These results support widespread use of unbiased stereology for analysis of histologic outcomes of injury models.