Label-free metabolic fingerprinting of motile mammalian spermatozoa with subcellular resolution

Abstract

Background: Sperm metabolic pathways that generate energy for motility are compartmentalized within the flagellum. Dysfunctions in metabolic compartments, namely mitochondrial respiration and glycolysis, can compromise motility and male fertility. Studying these compartments is thus required for fertility treatment. However, it is very challenging to capture images of metabolic compartments in motile spermatozoa because the fast beating of the flagellum introduces motion blur. Therefore, most approaches immobilize spermatozoa prior to imaging. Results: Our findings indicate that immobilizing sperm alters their metabolic profile, highlighting the necessity for measuring metabolism in spermatozoa during movement. We achieved this by encapsulating mouse epididymis in a hydrogel followed by two-photon fluorescence lifetime imaging microscopy for imaging motile sperm in situ. The autofluorescence of endogenous metabolites—FAD, NADH, and NADPH—enabled us to visualize sperm metabolic compartments without staining. We trained machine learning for automated image segmentation and generated metabolic fingerprints using object-based phasor analysis. We show that metabolic fingerprints of spermatozoa and the mitochondrial compartment (1) can distinguish individual males by genetic background, age, or fecundity status, (2) correlate with fertility, and (3) change with age likely due to increased oxidative metabolism. Conclusions: Our approach eliminates the need for sperm immobilization and labeling and captures the native state of sperm metabolism. This technique could be adapted for metabolism-based sperm selection for assisted reproduction.

Lenka Backová

PhD Student

My research focuses on deep learning-based bioimage analysis and biophysical modeling of multicellular systems.