dc.description.abstract | This study presents work on the effects of acetaminophen on female mice’s reproductive
hormone levels in blood using Raman spectroscopy together with chemometrics. Here weak
Raman signals are enhanced by a factor of 3.94 when conductive silver paste smeared glass
slides are used as Raman substrates. These Raman substrates were characterized using Raman
spectroscopy upon 785 nm excitation and for the first seven days after preparation, they were
found to possess chemical stability. The substrates were applied in four female hormones
(estradiol, follicle-stimulating hormone (FSH), luteinizing hormone (LH) and progesterone)
characterization and level determination in blood using Raman spectroscopy. The spectral
profiles of respective standard hormones (with no blood) displayed common (480, 1244 and
1454 cm-1) and other bands (540 cm-1 (in LH and progesterone), 1005 cm-1 (estradiol and FSH),
880 cm-1 (FSH), 837 cm-1 (FSH and LH) and 1360 cm-1 (progesterone). The spectral data set
analyses to reveal these bands was done using Principal Component Analysis (PCA) and analysis
of variance (ANOVA). In order to identify biomarker bands of the respective hormones in the
blood, Raman spectroscopy of simulate samples (prepared by mixing each separately with male
mouse’s blood at different concentrations) were done. The Raman experimental parameters
were: Excitation wavelength, 785 nm; spectra accumulation, 10; exposure time, 10 s; center
wavelength, 1000 cm-1; microscope objective, ×10, 0.3 numerical aperture; 600 lines grating; and
≈68.5 μm beam spot size. The biomarker bands were identified to be centered around
wavenumbers 668, 902, 1011 cm-1 for estradiol; 1219 and 1296 cm-1 for FSH; 1440 cm-1 for LH
and 1569 cm-1 for progesterone. These bands exhibited intensity variation with the concentration
of the respective hormones in the blood and were used in determining the levels in the blood of
treated (acetaminophen administered) and normal (untreated/control) mice. The bands
(biomarker bands for each hormone) were used to build Artificial Neural Network (ANN)
models to achieve quantitative analysis. A determination coefficient (R2) and root mean square
error (RMSE) values of greater than 98.05% and less than 0.0602 respectively were estimated.
The high (low) value of R2 (RMSE) indicated a good model for concentration prediction. In this
work, upon applying the calibrated ANN model in hormone concentration level determination in
blood, it was found that mice treated with acetaminophen had on average a level increase from
normal of 29%, 138%, 1.1% and 44% in estradiol, FSH, LH and progesterone hormones
respectively. Such large hormone level modification away from normal has adverse effects on
v
fertility. These results served to prove that application of Raman spectroscopy together with
chemometrics has a great potential in use for hormone level determination in blood. Besides,
the limits of detection were estimated to be 38 pg/ml, 0.45 mIU/ml, 0.69 mIU/ml and 7.14 ng/ml
for estradiol, follicle-stimulating hormone, luteinizing hormone and progesterone respectively.
These values were way lower than reported values of HPLC and ELISA methods thus indicating
the high sensitivity of the Raman spectroscopic method. | en_US |