Alpha-1 chain of human haptoglobin as viability marker of in vitro fertilized human embryos: information beyond morphology

ABSTRACT

Only one third of the in vitro fertilization treatments result in successful delivery following morphological viability assessment worldwide. A paper by Montskó et al. (2015) describes the identification of the alpha-1 chain of human haptoglobin as a potential marker of embryo viability. Using mass spectrometry, the concentration of the haptoglobin alpha-1 chain was determined in spent culture media samples of in vitro fertilized embryos and correlation was found with the outcome of the respective transfer. In the present study we investigated, whether the concentration of haptoglobin alpha-1 chain shows any correlation with morphological scores to clarify whether levels of the alpha-1 chain provide additional information on embryo viability unnoticed by the morphological assessment. In the study, pregnancy and live birth rates were examined in 143 transferred samples of 86 patients, retrospectively. Two sample groups were created. The control group contained embryos classified as ‘good’ or ‘fair’ based on the Istanbul Consensus Criteria System, while the double-assay group contained embryos assessed as ‘good’ or ‘fair’ by the morphological evaluation and as ‘viable’ by the haptoglobin assay. Clinical pregnancy rate was 30.2% in the control group, while 47.6% in the group scored parallel with morphological criteria and proteomic analysis (p < 0.05). The increased clinical pregnancy rate observed in the double-assayed group can be attributed to decreased false-positivity of the double assay.

Abbreviations: IVF: in vitro fertilization; SEC: spent embryo culture medium; HSA: human serum albumin; Hpt: haptoglobin; HptA1: haptoglobin alpha-1 chain; ICCS: Istanbul Consensus Criteria System; BMI: body mass index; ICSI: intracytoplasmic sperm injection

KEYWORDS:

• In vitro fertilization
• embryo viability
• biomarker
• mass spectrometry
• success rate

Introduction

Infertility is a major public health issue worldwide that is also recognized by the World Health Organization (Boivin et al. 2007). Since the first reported case of successful in vitro fertilization treatment (IVF) (Steptoe and Edwards 1978), assisted reproduction techniques have advanced significantly. IVF became increasingly widespread resulting in 142,000 IVF cycles a year in the United States in 2007 (Seli et al. 2007). By 2014 this number reached more than 200,000 (Assisted Reproductive Technology 2014). In the European Union a similar growing tendency is observed (de Mouzon et al. 2012; Calhaz-Jorge et al. 2016). Despite evolving assisted reproduction technologies and culturing methods of embryos, successful delivery following IVF is below that expected (Corcoran et al. 2005; Lane and Gardner 2007), around 33% (Calhaz-Jorge et al. 2016).

In the current practice of assisted reproduction, embryos are selected for transfer based on noninvasive morphological evaluation. A three-part grading system was originally reported in 1999 (Gardner and Schoolcraft 1999) that was later modified specifically for blastocysts (Stephenson et al. 2007). Due to diversity of blastocyst morphology, there was a need for a reliable grading system. An international consensus on oocyte and embryo assessment was created by Alpha Scientists in Reproductive Medicine (Alpha Scientists 2011) during the Istanbul Consensus Workshop in 2011. The guidelines that were set on the workshop represent the currently accepted protocol for viability assessment of in vitro fertilized embryos. In parallel with recent advances in the field of morphological evaluation (Paternot et al. 2013; Gardner et al. 2015) noninvasive metabolomic (Gardner and Wale 2013; Rødgaard et al. 2015) and proteomic (Katz-Jaffe et al. 2006a; Nyalwidhe et al. 2013) methods were developed based on molecular markers present in the culture media of embryos. Analysis of spent embryo culture media samples (SEC) offers an exceptional, noninvasive method to examine embryo viability (Nagy et al. 2008; Seli et al. 2010). Initially, catabolic activity of the developing embryo was monitored by the consumption of nutrients, for example, glucose in the culture medium (Devreker et al. 2000; Gardner et al. 2001). Later, with the help of new analytical techniques such as mass spectrometry, analysis of the embryonic secretome (Katz-Jaffe et al. 2006b; Cortezzi et al. 2011) became possible.

Our research group has recently described (Montskó et al. 2015) the level of alpha-1 chain of haptoglobin in the culture medium as a novel indicator of embryo viability. Haptoglobin (Hpt) is present as a known contaminant (Dyrlund et al. 2014) in purified human serum albumin (HSA) products that are also used to supplement culture media. The alpha-1 chain of haptoglobin (HptA1) can be detected in blank unconditioned media samples of HSA supplemented media, during in vitro embryonic development further fragmentation of Hpt was observed. Based on our study, the quantity of HptA1 inversely correlates with embryo viability: increased concentration of HptA1 was detected in SEC of embryos which failed to implant.

The aim of the present, retrospective study was to test whether HptA1 levels correlate with the results of morphological evaluation of in vitro fertilized embryos. An additional goal was to investigate whether measurement of HptA1 levels can contribute to elevation in clinical pregnancies and increased live birth rates.

Results

Pregnancy and live birth rates were divided and compared as two groups of in vitro fertilized embryos (n = 143). The control group contained embryos classified as ‘good’ or ‘fair’ by the Istanbul ICCS, while the double-assay group was constructed using embryos assessed as ‘good’ or ‘fair’ by the ICCS and were also assayed using liquid chromatography coupled mass spectrometry (HptA1 assay). The complete workflow is summarized in Figure 1A, the HptA1 assay was performed retrospective. Figure 1B summarizes the statistical ‘worst and best scenarios’ approach that was used as detailed in the Materials and Methods. To estimate whether the number of samples were sufficient for the study, a power analysis was performed. The power analysis, based on alpha (level of significance), sample size, and effect size (d), showed that in case of clinical pregnancy, as primary outcome, the achieved powers were above 80% (worst scenario: d = 0.643, power = 0.858; best scenario: d = 1.522, power = 0.999). For live birth as a secondary outcome only the best scenario was above than 80% (d = 1.310, power = 0.999) but not in the worst case (worst scenario: d = 0.467, power = 0.591). One-way ANOVA and logistic regression analysis showed that the level of HptA1 is not influenced by any etiological factors described in the materials and methods (df = (5, 67), F = 1.78, p = 0.12). Patient data is summarized in Table 1.

Table 1. Study characteristics, values are expressed as mean ± SD.

	Study group
Number of patients	86
Number of transferred samples	143
Maternal age	33.8 ± 4.7
Maternal BMI	22.7 ± 2.8
FSH treatment (IU)	1715.9 ± 841.9
Number of follicles	8.0 ± 4.8
No. of metaphase II oocytes	6.7 ± 4.2
Fertilization rate (%)	67.9 ± 27.4

Control group

The control group (n = 86) contained embryos assessed as ‘good’ (n = 48) or ‘fair’ (n = 38) by ICCS. Embryos in the control group were only graded using the morphological viability assessment. The 48 embryos graded as ‘good’ resulted in 17 pregnancies while the 38 embryos graded as ‘fair’ resulted in 9 pregnancies meaning clinical pregnancy rates of 35.41% (17 out of 48) and 23.68% (9 out of 38), respectively. Live birth rates were 31.25% (15 out of 48) and 18.42% (7 out of 38), respectively. The rate of miscarriage/abortion was 4.16% (2 out of 48) in the population of embryos graded as ‘good’ while 5.26% (2 out of 38) in the population of embryos graded as ‘fair.’ The total 86 transferred embryos resulted in 26 pregnancies with an overall clinical pregnancy rate of 30.23%. The overall live birth rate was 25.58% (22 out of 86) and the overall miscarriage/abortion rate 4.65% (4 out of 86). Chi-square analysis showed no significant association between the scores given by ICCS and pregnancy as clinical outcome (χ² = 0.883, df = 1, p = 0.347) nor with live birth (χ² = 1.226, df = 1, p = 0.269).

Double-assay group

In the double-assay group morphological viability assessment was complemented by the results of the HptA1 assay. Sample numbers were identical to those of the control group. During the HptA1 assay respective lots of blank G-1TMv5+ HSA culture media incubated under the same circumstances as the embryos were used as controls. Student’s t-test showed significant differences between HptA1 and the clinical outcomes, no-pregnancy versus pregnancy as well as no live birth versus live birth in both scenarios (Table 2). In case of worst scenario of live birth this difference was only marginally significant.

Table 2. Average amount of HptA1 (and standard deviation) based on the clinical outcomes and the results of the related statistical analysis.

	Pregnancy
	No-pregnancy (sd)	Pregnancy (sd)	t	df	p
Worst	65.50 (14.25)	57.51 (10.28)	2.92	64.90	0.004
Best	72.94 (15.01)	54.34 (8.56)	7.25	77.57	2.68e-10
	Live birth
	No live birth (sd)	Live birth (sd)	t	df	p
Worst	64.58 (14.33)	58.72 (10.45)	2.05	49.95	0.045
Best	71.56 (15.56)	54.98 (8.83)	6.12	65.05	5.92e-08

sd: standard deviation; t: value of t-test; df: degrees of freedom; p: significance

The correlation analysis revealed significant correlation (worst scenario: 0.27, p = 0.011; best scenario: 0.54, p = 7.32e-08) between the amount of HptA1 and the clinical outcome of transfer (pregnancy). Comparing the mean HptA1 amounts in ICCS groups ‘good’ and ‘fair’ Student’s t-test did not show any significant differences (worst scenario: t = −0.52, df = 82.98, p = 0.602; best scenario: t = −1.75, df = 78.68, p = 0.243) as the amount of HptA1was found to be almost identical (Figure 2) in both groups.

Figure 1. Study outline of the main assay steps (A) and the data analysis (B). Spent culture medium samples were stored frozen at –80°C until the haptoglobin alpha-1 assay.

Correlation analysis between the results of the HptA1 assay and the ICCS scores did not show any significant correlation (worst scenario: r = 0.127, t = 1.17, df = 84, p = 0.241; best scenario: r = 0.056, t = 0.514, df = 84, p = 0.608).

The combined evaluation of ICCS scores and the HtpA1 categories showed that the highest pregnancy rates were within the ‘good’ graded embryos, where the HptA1 category is viable, in both scenarios (Table 3). Using a Kruskall-Wallis test, it was investigated whether the three categories based on HptA1 concentration (viable, unclear and non-viable) significantly differ in HptA1 concentrations. It was found that in both scenarios there were significant differences between the categories in the concentration of HptA1 (worst scenario: KS = 72.907, df = 2, p < 0.05, best scenario: KS = 74.445, df = 2, p < 0.05). For further multiple comparison, Dunn’s test was applied with Benjamini-Hochberg p values adjusted method. The analysis revealed that significant differences exist in all pairwise comparison in both scenarios (p < 0.05).

Table 3. Number of pregnancies and no-pregnancies based on ICCS scores and the HptA1 categories. In the parentheses the live birth (or no live birth) data are also shown, where there is no parentheses the values were the same.

	ICCS	HptA1 categories	No-pregnancy	Pregnancy	%
Best scenario	Good	Nonviable	14	1	6.66
		Unclear	12 (13)	9 (8)	42.85 (38.09)
		Viable	5 (6)	7 (6)	58.33 (50.00)
	Fair	Nonviable	15	0	0.00
		Unclear	12	2	14.28
		Viable	2 (4)	7 (5)	77.77 (55.55)
Worst scenario
	Good	Nonviable	14	5	26.31
		Unclear	6 (7)	4 (3)	40.00 (30.00)
		Viable	11 (12)	8 (7)	42.10 (36.84)
	Fair	Nonviable	16	2	11.11
		Unclear	5	1	16.66
		Viable	8 (10)	6 (4)	42.85 (28.57)

It was noted that within the ‘fair’ graded embryos the non-viable group measured by HptA1 levels had the lowest rates. In case of live birth rates and the ICCS scores, the results were similar to those mentioned above (Table 2). The total number of embryos that were considered as ‘double-positive’ were 21 (worst scenario) and 33 (best scenario) resulting in 14–14 pregnancies meaning an overall clinical pregnancy rate of 66.66% and 42.42%, respectively.

The overall live birth rates were 52.38% (best) and 33.33% (worst) (11 out of 21, and 11 out of 33). The overall missed abortion rate was 4.65% (4 out of 86) with 29 and 30 (best, worst scenario, respectively) embryos graded as ‘good’ or ‘fair’ by the ICCS, however, as ‘nonviable’ by the HptA1 assay. These cases all proved to be true negative cases: embryos classified as ‘nonviable’ were all corresponding to embryos which failed to implant (no clinical pregnancy was observed by ultrasound examination).

Discussion

Haptoglobin is present in the purified HSA product used to supplement the embryo culture medium as a contaminant (Darcel and Kaldy 1986; Dyrlund et al. 2014). Interestingly, however, there are publications describing the role of haptoglobin in the process of human fertility and implantation (Berkova et al. 2001; Herrler et al. 2003; Weiss et al. 2013). Our research follows the noninvasive search for biomarkers of embryo viability using a proteomic approach. This is a still developing field by analyzing the pooled (Cortezzi et al. 2011) or individual (Katz-Jaffe et al. 2006a) culture medium samples of in vitro fertilized embryos identifying several new candidate biomarkers. In this paper the possible use of HptA1 was investigated as a marker of embryo viability (Montskó et al. 2015). We previously observed that the grade of Hpt fragmentation and thus the concentration of HptA1 in the culture medium can reliably predict the outcome of embryo transfer. Although HptA1 is already present in the G-1TMv5+ HSA medium, its concentration further increases during embryo culturing. In the present work it was clarified in a retrospective analysis that HptA1 concentration used as a quantitative biomarker is truly a new indicator of embryo viability. Live birth rates were also compared in experimental groups of embryos assessed as ‘good’ or ‘fair’ by the ICCS and embryos assessed as ‘good’ or ‘fair’ by the ICCS and additionally graded as ‘viable’ by the HptA1 assay.

Comparing the quantitative results of the HptA1 assay to the ICCS scores it can be concluded that the levels of HptA1 in ICCS groups ‘good’ and ‘fair’ do not significantly differ. This result suggests that the amount of HptA1 in the culture medium is an indicator that is independent from the morphological features of the embryo and adds novel, biochemical information to and above morphological grading of viability assessment. Our results suggest that the amount of HptA1 in the medium is in fact an independent indicator of implantation potential carrying information beyond morphological aspects. The potential of the HptA1 assay is the identification of embryos having low, or no implantation potential due to visually undetectable reasons. In the double-assay group (n = 86) 33 embryos graded as ‘good’ or ‘fair’ and as ‘viable’ by the HptA1 assay resulted in 22 live births. The remaining 50 samples represent true negative cases, embryos having low or no implantation potential that were not noticed by morphological viability assessment. The increased overall clinical pregnancy rate observed in the double-assay group compared to the control group (30.23% vs. 47.61%) is due to the decrease of false positive cases detected by ICCS. The real potential of the HptA1 assay is the identification of low-grade embryos rather than selecting the best ones. The most significant result of the assay is to decrease the number of potentially unviable embryos leading to an increased rate of live birth/transfer. The two best clinical pregnancy rates (77.8% and 58.3%) were achieved in the group of embryos graded as ‘fair’ or ‘good’ by the ICCS and as ‘viable’ according to the HptA1 assay, respectively.

This study has some limitations, however. First is the fact that using HptA1 concentration data, three categories were found (viable, unclear, nonviable). The first and the last categories carry obvious messages in clinical application of the method; however, in the ‘unclear’ category this is not that simple. Under ‘real’ (routine) clinical circumstances ‘viable’ embryos should be preferred for transferring over ‘unclear’ embryos. If this is not possible in the absence of ‘viable’ embryos, ‘unclear’ embryos should and can be transferred (selected) since an acceptable percentage of clinical pregnancy is still achievable. Second is that the published results are still retrospective. Our goal is to implement the HptA1 assay in a prospective clinical trial. At present, the prospective clinical trial is under development. Further the fragmentation of Hpt is an indicator of a yet unknown phenomenon. At the moment, it is not more than a working hypothesis, that HptA1 levels indicate cellular damage of certain embryos and therefore unviable embryos can be detected earlier with such biochemical marker then visualized under the microscope. These are the embryos who appear ‘good’ or ‘fair’ with the morphological scoring and appear non-viable with mass spectrometry. Such a phenomenon, that is, that a biochemical disease marker indicates a pathological state earlier than a morphological marker, is not uncommon in medical diagnostics (e.g., the early elevation of certain tumor marker molecules).

The reaction which leads to the increased amount of HptA1 seems to correlate with implantation potential of the embryo (Montskó et al. 2015). As the purpose of HptA1 and its production by chemical or biochemical reactions are not clear, further studies are necessary in the future. The presented results, however, confirm that the HptA1 assay has the potential to provide additional information on embryo viability. The HptA1 assay can identify embryos amongst low or no implantation potential that remain unnoticed by visual examination.

Materials and methods

Patients

The study was performed between 2016 August and 2016 December in the Assisted Reproduction Unit of the Department of Obstetrics and Gynecology, University of Pécs in an unselected population. In this period, we started 86 cycles, and transferred 143 embryos. The couples were presented with the following main infertility diagnosis: oligozoospermia (n = 23), damaged or blocked Fallopian tubes (n = 24), asthenozoospermia (n = 3), endometriosis (n = 11), male factors (n = 2), unexplained infertility (n = 23). Embryos were fertilized using intracytoplasmic sperm injection, embryos fertilized by conventional IVF were excluded.

The study was approved by the Committee of Human Reproduction, National Science Council of Hungary (5273–2/2012/HER), and by the Public Health Officer, Hungarian Government Office in Baranya County (BAR/006/58–2/2014). The statement of informed consent was part of the ethical permission.

Stimulation and fertilization protocol

The complete and detailed stimulation and fertilization protocol is found in Montskó et al. (2015). Every embryo involved in the study was fertilized using ICSI in order to eliminate any interference caused by cumulus cells. Embryo transfer was done 3 days after oocyte retrieval. Spent culture medium samples were stored frozen at −80°C until the mass spectrometric measurement. The presence of gestational sac was detected by using transvaginal ultrasound examination 21 days after embryo transfer.

Study design

The concentration of HptA1 was compared in samples of embryos assessed by ICCS as ‘good,’ or ‘fair’ and it was examined whether there is any correlation between HptA1 concentration and morphological scores. The primary outcome of the measurement is observing the clinical pregnancy rate in the experimental groups. Live birth and abortion rates were also compared. The sample population was composed of single and double embryo transfers.

Figure 2. Average HptA1 amounts in ICCS groups (1 ‘good’ and 2 ‘fair’) analyzed with Student’s t-test, in case of the worst and the best scenarios.

In the case of DET transfers only those samples were included in the dataset, where the clinical outcomes (pregnancy or live birth) were the same, that is, either both embryos failed to result in clinical pregnancy (failed transfer) or both embryos resulted in clinical pregnancy (gemini). Therefore, a statistical approach called as ‘worst and best scenarios’ was applied to investigate pregnancy and live birth rates. Figure 1B shows the schematic representation of this approach. In the case of the worst scenario, the minimum value of HptA1 (when clinical outcome = 0, no pregnancy) and the maximum value of HptA1 (when clinical outcome = 1, clinical pregnancy) were used. In the best scenario case, the maximum value of HptA1 (when clinical outcome = 0, no pregnancy) and the minimum value of HptA1 (when clinical outcome = 1, clinical pregnancy) were analyzed. Supplementary 1 shows an example of this approach with fictive sample dataset to demonstrate the worst and the best scenarios. We created three categories for pregnancy, based on the average amount of HptA1. If the value was above or equal than the mean of ‘no-pregnancy’ group, it was marked as ‘nonviable’, when it was lower than the mean of ‘pregnancy’ group, it was marked as ‘viable.’ If the value was between these two it was classified as ‘unclear.’

Morphological viability assessment

Embryo photos were recorded on the third day following fertilization using a Nikon Diaphot 300 (Nikon Corp., Tokyo, Japan) inverted phase contrast microscope at 40-fold magnification and resolution of 1024 × 768 pixels. Digital images of embryos were recorded using the Octax CytoScreen^TM EyeWare (Vitrolife GmBH, Bruckberg, Germany) software. EyeWare (Vitrolife GmBH, Bruckberg, Germany) software. For viability scoring the guidelines set in the ICCS were used, embryos involved in the study were graded as ‘good’ (< 10% fragmentation, stage-specific cell size, no multinucleation) or ‘fair’ (10–25% fragmentation, stage-specific cell size for majority of cells, no evidence of multinucleation). One or two embryos/patients were transferred on request.

Viability assessment using liquid chromatography coupled mass spectrometry (HptA1 assay)

SEC samples were thawed, then 5 µl of internal standard solution was added to 25 µl of each sample. Internal standard solution was prepared by dissolving cortisol (Sigma-Aldrich, Budapest, Hungary) in 20% methanol in a 3.86 µM/L concentration. Calibration curve was constructed using HptA1 calibrators prepared in G-1^TMv5 culture medium at concentrations of 0.05 ng/µl, 0.1 ng/µl, 0.2 ng/µl, 0.5 ng/µl, and 1 ng/µl, respectively. The HptA1 synthetic peptide used to construct the calibrators was purchased from NovoPep Limited (Shanghai, China). A total of 10 µl of the samples or calibrators were injected and analyzed. Detailed analytical conditions are found in Montskó et al. (2015).

Statistics

All statistical analysis was performed using R (R Core Team) software. Sample size was determined by a post hoc power analysis due the nature of the data. During the analysis, Chi-square test, Student’s t-test and Pearson’s correlation test were used when it was appropriate, otherwise noted. One-way ANOVA was applied, after checking the normality and heterogeneity assumptions, to examine the relationship between infertility and the amount of HptA1.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website.

Additional information

Funding

The presented work was funded by NKFI-EPR K/115394/2015/HU ‘Early biochemical indicators of embryo viability’ and EDIOP-2.3.2-15-2016-00021. ‘The use of chip-technology in increasing the effectiveness of human in vitro fertilization’ and by the ÚNKP-17-4-III ‘New National Excellence Program of the Ministry of Human Capacities’ grants [ÚNKP-17-4-III]; National Research, Development and Innovation Office, Hungary [EDIOP-2.3.2-15-2016-00021, NKFI-EPR K/115394/2015/HU].

Notes on contributors

Gábor L. Kovács

Performed HptA1 assay, the statistics and prepared the manuscript: GM, RH. Performed the culturing of embryos, the morphological viability assessment and collected the samples: KG. Supervised the clinical protocol and prepared the sections of the manuscript describing oocyte collection, fertilization and embryo transfer procedures: ÁV, BJ. Supervised the preparation of the manuscript the statistical analysis and the corresponding author of the manuscript: GLK.