Abstract Study question Can combination of metabolomics of spent embryo culture medium (sECM) and cumulus cell (CC) genes judge live birth rate in unexplained infertility (UI)? Summary answer A combination of day3 metabolic profiling of sECM with CC gene expression yields an algorithm to predict live birth rate in UI. What is known already Optimizing embryo selection is rate-limiting for pregnancy success. The metabolomic profile of sECM can aid in advanced prediction of embryonic developmental potential. Recent transcriptomics studies proposed prokinectin 2 (PROK2), and pregnancy up-regulated nonubiquitous CaM kinase (PNCK) as two novel gene/s responsible for perifollicular vascularization and embryonic development respectively; two important events regulating bidirectional CC-oocyte signalling and implantation ability. Given that UI is contributory to impaired oocyte quality, we delineated if combined metabolic signature of CC gene expression encompassing metabolic, developmental competence, and extracellular matrix facet/s and sECM could address live birth rate in women with UI following single embryo transfer. Study design, size, duration A prospective cohort study was conducted at Institute of Reproductive Medicine between January 2022 and October 2022. 48 women (age: 30-39 years) diagnosed with UI undergoing oocyte-retrieval offered consented participation and were sub-stratified according to pregnancy outcome. Cleavage-stage embryos from intra cytoplasmic sperm injection (ICSI) cycles were scored according to combination of gene expression pattern of CC-samples and integral value of metabolite obtained from nuclear magnetic resonance (NMR) spectroscopy of sECM (range 0-3, 4-6, 7-9). Participants/materials, setting, methods Expression profiles of metabolic (LDHA, PFKP, PKM2), extracellular matrix (HAS2, PTX3, TNFAIP6, VCAN), and developmental competence (PNCK, PROK2) related genes were evaluated by real time-quantitative polymerase-chain-reaction (qRT-PCR) from retrieved CC. Day-3 sECM (30 μL) was collected for NMR from 43 single transferred embryos. Univariate and multivariate analysis was performed after NMR data acquisition. Live birth rate was evaluated by area-under-the-curve (AUC) values using combined score panel. Main results and the role of chance The groups were similar for patient age, BMI, AMH, infertility diagnosis, and stimulation protocol. A total of 43 cycles were included in the analysis [embryo transfer cancelled (n = 2), insufficient cumulus mass (n = 3)]. Clinical pregnancy was confirmed in 16 patients (37.21%). Mean mRNA levels for all genes were similar in CC associated with oocyte that fertilized normally (2PN) compared with those that either failed to fertilize or were abnormal (i.e. 1PN, 3PN). Similar results were observed in terms of speed of embryonic cleavage (≥7 cell on Day 3 vs. ≤6 cell on Day 3). Expression of VCAN, (OR: 1.102, 95% CI: 1.010-1.202, p < 0.02), PNCK (OR: 0.931, 95% CI: 0.916-0.947, p < 0.001) and PROK2 (OR: 1.273, 95% CI: 1.011-1.231, p < 0.001) was significantly higher (p < 0.01) in CC that achieved live birth compared with those that failed to result a successful pregnancy. Multivariate and univariate analysis revealed distinct metabolomic signatures between pregnant and non- pregnant group/s. Lactate, pyruvate and proline were the most significant (p < 0.01) altered metabolite in sECM of non-pregnant group when compared to their pregnant counterpart. The combination demonstrated a “good” (VCAN, pyruvate: AUC: 0.87), “fair” (PROK2, proline: AUC: 0.79) and “modest” (PNCK, glucose: AUC: 0.71) in terms of live-birth rate. Limitations, reasons for caution The study population is heterogeneous in terms of duration of infertility and previous infertility treatment outcome, limited by small sample size and subjective nature of embryo scoring. The presumable involvement of unknown factors in developmental competence of UI and impact of frozen cycle/s should also be addressed. Wider implications of the findings The study provides reliable accuracy measures to support relationship between live birth rate and combined score panel in patients with UI. It is envisioned integration of CC gene expression/s and metabolomics of sECM if combined with patient characteristics may tailor therapy by artificial intelligence to model the outcome in UI. Trial registration number Not applicable
Read full abstract