FT-IR spectroscopic metabolic fingerprintiNg of human embryo culture media combined with multivariate statistical analysis to discriminate embryos with different reproductive potential

Abstract

OBJECTIVE: The aim of this work was to study metabolic fingerprinting of spent culture media of human embryos and to develop a multivariate statistical model to discriminate embryonic reproductive potential.DESIGN: Descriptive, Experimental.MATERIALS AND METHODS: We analyzed 30 spent day 3-embryo culture media from 15 patients with known outcome after IVF-ET, using FT-IR spectroscopy. 30 mL of embryo culture media were transferred to ZnSe optical plates and dried to obtain transparent films. FT-IR spectra were acquired between 4000 and 650 cm−1 with 6 cm−1 spectral resolution using a Spectrum One FT-IR spectrometer. To increase the number of discriminative features present in spectra, the first and second derivatives were calculated using OPUS software (Bruker Optics). Dendrograms were constructed by executing Ward's algoritm of OPUS software using first derivate spectra as input data for multivariate hieralchical cluttering analysis (HCA).RESULTS: We found that in vitro cultured embryos that resulted in pregnancy produce a different spectroscopic pattern compared with embryos that do not result in pregnancy. We could determine that these differences could be detectable in first derivate spectra in the 3000-2800 cm−1 range (lipid region), the 1500-1300 cm−1 interval (mixed region) and the 1200-900 cm−1 region (carbohydrate region).When HCA was applied to first derivate spectra in those spectral ranges, the dendrogram produced showed two clusters which clearly distinguish embryos with proven pregnancy from those which failed to implant.CONCLUSION: Our preliminary study shows that changes detected in spectroscopic metabolic fingerprinting of spent human culture media from embryos with different reproductive potential would allow the development of a robust, fast and noninvasive methodology based on a multivariate statistical analysis able to predict pregnancy outcome before embryo transfer. OBJECTIVE: The aim of this work was to study metabolic fingerprinting of spent culture media of human embryos and to develop a multivariate statistical model to discriminate embryonic reproductive potential. DESIGN: Descriptive, Experimental. MATERIALS AND METHODS: We analyzed 30 spent day 3-embryo culture media from 15 patients with known outcome after IVF-ET, using FT-IR spectroscopy. 30 mL of embryo culture media were transferred to ZnSe optical plates and dried to obtain transparent films. FT-IR spectra were acquired between 4000 and 650 cm−1 with 6 cm−1 spectral resolution using a Spectrum One FT-IR spectrometer. To increase the number of discriminative features present in spectra, the first and second derivatives were calculated using OPUS software (Bruker Optics). Dendrograms were constructed by executing Ward's algoritm of OPUS software using first derivate spectra as input data for multivariate hieralchical cluttering analysis (HCA). RESULTS: We found that in vitro cultured embryos that resulted in pregnancy produce a different spectroscopic pattern compared with embryos that do not result in pregnancy. We could determine that these differences could be detectable in first derivate spectra in the 3000-2800 cm−1 range (lipid region), the 1500-1300 cm−1 interval (mixed region) and the 1200-900 cm−1 region (carbohydrate region).When HCA was applied to first derivate spectra in those spectral ranges, the dendrogram produced showed two clusters which clearly distinguish embryos with proven pregnancy from those which failed to implant. CONCLUSION: Our preliminary study shows that changes detected in spectroscopic metabolic fingerprinting of spent human culture media from embryos with different reproductive potential would allow the development of a robust, fast and noninvasive methodology based on a multivariate statistical analysis able to predict pregnancy outcome before embryo transfer.