Raman spectra were obtained from the following commercially available chemicals (Wuhan Boster Biological Technology Ltd., Wuhan, China) for use in spectral fitting and Raman criterion development: albumin, monopotassium phosphate, cysteine, histidine, phenylalanine, methionine, serine, tyrosine, tryptophan sodium, pyruvate, sodium citrate and taurine. The fitting spectra was calculated using a self-produced program in the Matlab software platform (MathWorks, Natick, MA, USA). Figure 1
shows that the fitting spectra shared great similarity with the measured Raman spectra of the samples. The average spectrum of each sample was calculated using the approach above, obtaining the fitting coefficient of the corresponding components involved in the medium as the relative concentration (C′).
2.1. A Preliminary Analysis of the 45 Samples
Based on the theory above, the Raman spectra of 45 samples from six patients were detected and analyzed with the Matlab software. In comparison, we found that the relative concentrations (C′) of sodium pyruvate and phenylalanine showed a certain degree of regularity.
The currently used embryo grading systems were developed soon after the report of the first successful pregnancy after IVF [10
]. Although their accuracy remained insufficient, these grading systems led to significant improvements in the implantation rate and pregnancy rate [11
], such that morphological assessment is the first-line approach for embryo selection. The establishment of Raman criterion to screen the higher quality of embryos should obey the results of morphological assessment. Based on the morphological assessment, we obtain the threshold of pyruvate/albumin and phenylalanine/albumin to be lower than 0.012 and higher than −0.00085, respectively. The results are shown in Figure 2
and Table 1
By analyzing the Raman information of the samples, combined with the embryo morphology scores, we found that the threshold for the relative concentration of sodium pyruvate was 0.012; that is, the embryos in culture media with less than 0.012 had high reproductive potential, whereas those with more than 0.012 had low developmental potential. For phenylalanine, the threshold is −0.00085; that is, embryos in media with phenylalanine levels above −0.00085 had high developmental capacity, whereas those in media with levels below −0.00085 had low developmental capacity. These results are expressed in detail in Figures 3
A total of 16 labels met the two conditions above simultaneously and were considered to have high developmental potential: 12, 14, 21, 23, 27, 28, 29, 31, 33, 34, 35, 36, 37, 38, 41 and 44. The labels of the corresponding embryos transferred were 12, 21, 23, 27, 29, 35, 36, 38, 41 and 44; the remainder of the embryos were cryopreserved, with the patients’ agreement. The labels of the embryos that resulted in clinical pregnancy were 35, 36, 38, 41 and 44, all of which were in the range of the embryos with high developmental potential by the previous test (Figure 5
2.2. The Verification of Preliminary Results
Of the second batch of 12 samples from four patients numbered 46–57, after removing the lower scoring embryos based on morphological evaluation, the remaining samples were detected by Raman spectroscopy in accordance with the methods above and the metabolism of the spent culture media was analyzed. We found that embryos with a better ability to develop corresponded to the labels 46, 47, 49, 50, 52, 53 and 56. The labels of the embryos that achieved pregnancy in four patients were 46, 47, 49, 50, 52 and 53. Thus, Raman spectroscopy combined with morphological methods to assess early embryo quality could achieve an accuracy rate up to 85.7% (6/7).
In this study, we analyzed the spent culture media of embryos with known outcomes after embryo transfer (ET) on day three. First, we assessed the embryos using the current embryo selection methodology based on cleavage rate and morphology; then, we excluded the spent culture media of the embryos with lower scores. The remaining samples were detected using Raman spectroscopy under certain conditions, and high-quality embryos were selected based on the relative concentrations of pyruvate sodium and phenylalanine. Our results suggest that the metabolomic profiles of embryonic development in IVF may associate with the ability to implant. Additionally, compared with embryos that failed in a pregnancy, in vitro cultured embryos with a high reproductive potential displayed a difference in the alteration of the culture environment that was detectable using spectroscopy.
During early cleavage, pyruvate is the preferred energy substrate for embryos [12
], whereas glucose becomes the primary energy source at the blastocyst stage [13
]. We also observed a trend toward lower pyruvate levels in the day three culture media of embryos that resulted in pregnancy compared with those that failed to implant (Table 1
). As the energy in the cleavage stage is provided mainly by the citric acid cycle, pyruvate consumption in embryo culture media can be regarded as an indicator of embryonic vitality and developmental potential. Our findings are consistent with previous reports by Hardy [14
] and Conaghan [15
There are large quantities of free amino acids in oviductal and uterine fluid [16
], which are important regulatory factors in the process of embryo culture [18
], including acting as prerequisites for the biosynthesis of cells, being involved in carbohydrate metabolism and regulating the cell osmotic pressure and pH. Amino acids can maintain the normal function of cells and improve the implantation. Gardner proved that the ideal environment for embryonic development included the presence of amino acids. Early embryonic development requires several specific amino acids, and after embryonic genome activation, all 20 amino acids are needed. Based on these results, Gardner and coworkers [19
] produced two sequential media, called G1 and G2. Non-essential amino acids and essential amino acids play different roles in the different developmental stages of embryos. Non-essential amino acids can promote the embryonic development from the cleavage stage to the blastocyst, increase the number of trophoblast cells and enhance the ability of blastocyst hatching. Essential amino acids, however, mainly increase the division rate of the inner cell mass and improve the ability of fetal development after implantation. Houghton [20
] and his colleagues hold that amino acid turnover can predict human embryo developmental capacity using high-performance liquid chromatography (HPLC). Brison [9
] used the same approach and found that the combination of decreased glycine and leucine and increased asparagine levels in the culture medium correlates with increased clinical pregnancy rate, and Sturmey [21
] reported similar findings in cryopreserved embryos. Seli [22
] found an association between higher glutamate levels in the culture medium and clinical pregnancy rates using proton nuclear magnetic resonance (1
H NMR). In the present study, we found that phenylalanine levels in the spent culture media samples of embryos that resulted in pregnancy were higher than those of embryos that fail to implant (Table 1
The development of embryo grading systems based on cleavage rate and morphology [23
] led to significant improvements in implantation and pregnancy rate and reductions in multiple gestation rate [11
]. The morphological evaluation method is rapid and non-invasive, but still subjective, and its precision is insufficient to enable most patients and clinicians to reduce the number of embryos transferred. These limitations have led many investigators to pursue adjunctive technologies for determining an individual embryo’s reproductive potential.
Other methods for the improvement of IVF outcomes are currently being studied. Preimplantation genetic screening has been proposed as a technique to improve embryo selection in specific populations, such as patients of older age or with repeated miscarriages. However, the effectiveness of this technique is still questionable [27
]. In addition, routine preimplantation genetic screening is time-consuming, requires considerable resources and is not suitable for all patients. Although there are many other methods to analyze the metabolism of embryos, there are limitations in their clinical application to various extents. For example, many of these technologies require complex equipment and dedicated technical staff that are not available at most embryology laboratories, and several technologies are so time-consuming as to miss the limited window of time available for embryo transfer.
The commercial culture media used in ART laboratories all share the same components, as they are all based on the metabolic requirements of human embryos [28
]. We chose Single Step Medium™ (SSM) culture medium in our study. The culture medium is the direct environment for embryo development in vitro
, which is very important for the physiology and viability of the pre-implanted embryos [29
]. Thus, the analysis of embryo culture medium can predict developmental potential noninvasively and objectively, and this draws public attention. The current detection methods include HPLC [20
], near-infrared (NIR) [30
] and 1
H NMR [22
]. Ahlstrom [31
] in Reproductive Biomedicine 2011 demonstrated that metabolomics profiling by NIR spectroscopy analysis can predict the implantation potential of blastocysts. Nadal-Desbarats [32
] and coworkers in MAGMA 2012 used 1
H NMR metabolomics profiling on 15 micro-L of embryo culture medium with multivariate data analysis to predict embryo viability. However, these technologies require complex equipment and preparation of samples, professional inspection and a long testing time that exceeds the time limit for embryo selection, which have been the limiting factors in routine viability assessment. As a result, the search for an accurate, rapid and noninvasive assessment is becoming an urgent problem to be solved in reproductive medicine.
Most recently, it was discovered that the oxidative status of the early embryo in IVF, as assessed by a thermo-chemiluminescence (TCL) analyzer, was associated with the chances of implantation; this approach was also applicable for centers that routinely perform elective transfer of two embryos [33
]. The ultimate aim is to find a suitable technology to assess embryo quality quickly for clinical practice.
Raman spectroscopy is based on molecular vibration information and can provide a wealth of molecular structure and composition information without causing any damage. Moreover, Raman spectroscopy is especially applicable for aqueous samples, because water only produces a very weak signal. Thus, this technique has become an important means of disease detection and diagnosis. In our study, it only took several minutes and 5 μL of sample volume to assess a sample using the chosen parameters. Raman is a simple, objective, real-time, noninvasive method.
] and Scott [35
] found that the viability index calculated by Raman spectroscopy was higher for embryos that succeed in implantation and delivery, compared with those that failed to implant. They analyzed individual samples using 15 μL of media, and spectra were recorded from 50 to 3450 cm−1
. In contrast, our study sample size was as less as 5 μL. In addition, the spectral range was narrower, from 600 to 1800 cm−1
. In particular, viability models used contributions from the wavelength regions of –SH, –CH and –NH, while our final analysis concluded that the specific substances in the culture medium, phenylalanine and pyruvate, are obviously related to the potential of embryos. Grading systems of embryo were developed by several investigators, leading to significant improvements in implantation and pregnancy rate, so they are always first-line approaches for embryo selection. The establishment of Raman criterion to screen the higher quality of embryos should be in accordance with the results of morphological assessment. In this study, to detect embryo culture medium using Raman spectroscopy combined with embryo morphology score, our method not only obviously reduces the scope in comparison with morphological assessment results, but it also accurately predicts clinical pregnancy.