Normalization of organ bath contraction data for tissue specimen size: does one approach fit all?


Erdogan B. R. , Karaomerlioglu I., Yesilyurt Z. E. , Ozturk N., Muderrisoglu A. E. , Michel M. C. , ...More

NAUNYN-SCHMIEDEBERGS ARCHIVES OF PHARMACOLOGY, vol.393, no.2, pp.243-251, 2020 (Journal Indexed in SCI) identifier identifier identifier

  • Publication Type: Article / Article
  • Volume: 393 Issue: 2
  • Publication Date: 2020
  • Doi Number: 10.1007/s00210-019-01727-x
  • Title of Journal : NAUNYN-SCHMIEDEBERGS ARCHIVES OF PHARMACOLOGY
  • Page Numbers: pp.243-251
  • Keywords: Normalization, Urinary bladder, Aorta, Weight, Length, Cross-sectional area, RAT URINARY-BLADDER, CARBACHOL-INDUCED CONTRACTION, STREPTOZOTOCIN-DIABETIC RATS, DETRUSOR SMOOTH-MUSCLE, ENDOTHELIAL DYSFUNCTION, BLOOD-PRESSURE, RESPONSES, ARTERIES, EXPRESSION, STRIPS

Abstract

Organ bath experiments are a key technology to assess contractility of smooth muscle. Despite efforts to standardize tissue specimen sizes, they vary to a certain degree. As it appears obvious that a larger piece of tissue should develop greater force, most investigators normalize contraction data for specimen size. However, they lack agreement which parameter should be used as denominator for normalization. A pre-planned analysis of data from a recent study was used to compare denominators used for normalization, i.e., weight, length, and cross-sectional area. To increase robustness, we compared force with denominator in correlation analysis and also coefficient of variation with different denominators. This was done concomitantly with urinary bladder strips and aortic rings and with multiple contractile stimuli. Our urinary bladder data show that normalization for strip weight yielded the tightest but still only moderate correlation (e.g., r(2) = 0.3582 for peak carbachol responses based on 188 strips). In aorta, correlations were even weaker (e.g., r(2) = 0.0511 for plateau phenylephrine responses normalized for weight based on 200 rings). Normalization for strip size is less effective in reducing data variability than previously assumed; the normalization denominator of choice must be identified separately for each preparation.