Cumulus cell gene expression as a potential biomarker for oocyte quality

Cumulus cell gene expression as a potential biomarker for oocyte quality

Volume 109, Issue 3, Pages 438–439


Catherine Racowsky, Ph.D., Daniel J. Needleman, Ph.D.


Reflections on "Cumulus cell transcriptome profiling is not predictive of live birth after in vitro fertilization: a paired analysis of euploid sibling blastocysts" by Green et al.

Read the full text here.


Go to the profile of Katherine Green
over 4 years ago

We thank Drs. Racowsky and Needleman for their thoughtful comments on our paper regarding the cumulus cell transcriptome and IVF outcomes.  While a sample size of 34 paired samples (17 patients) may seem small, this is one of the largest human studies on this topic to date.  In addition to being one of the largest studies, it is the only paired study design.  This is recognized by Drs. Racowsky and Needleman as the optimal study design when investigating a biomarker for embryo quality and is arguably a much stronger study design than an unpaired study with a larger sample size.  That being said, certainly, a larger, paired study would increase the likelihood of identifying significantly differentially expressed genes.  The age range of included patients is wide (ages 18-42 years) and we recognize the limitations of this heterogeneity; however, restricting the age criteria for inclusion would further limit the sample size and may yield results only applicable to the narrow range of patients studied.

Patients in our analysis were not more likely to be poor prognosis solely because they underwent double embryo transfer (DET).  As described in the methods, patients underwent DET as part of the study design to investigate transcriptome differences between cumulus cells from sibling implanted versus non-implanted embryos.  All included patients had normal ovarian reserve testing and were, in fact, moderate to good prognosis patients.  It is notable that our analysis was conducted prior to the publication of the 2017 ASRM transfer order guidelines, and the unique study design involving a sibling euploid DET would no longer be possible.

By evaluating only good quality blastocysts suitable for trophectoderm biopsy and determined to be euploid, we eliminated the confounder of an aneuploid-related effect on the transcriptome.  We sought to analyze blastocysts specifically (vs. cleavage stage embryos) because of the increasing use of extended culture to aid in embryo selection and the goal of identifying a biomarker of blastocyst quality.  Several Class 1 studies have shown that extended culture and PGT-A improve outcomes.  We aimed to improve on the most rigorous selection paradigm that exists. Therefore, we do not feel that studying only high quality, euploid blastocysts was a limitation.

At the time of the study, the majority of patients undergoing a freeze-only protocol for preimplantation genetic testing at our center underwent trophectoderm biopsy on day 6.  We recognize that by analyzing only day 6 blastocysts the results may only apply to this group of embryos.  However, the analysis of only day 6 blastocysts also eliminates the possibility of detecting transcriptome differences related to the day of blastulation.

Lastly, we acknowledge the analytical hurdles related to multiple hypothesis testing and the study of gene expression.  Future investigations of candidate genes based on biologic mechanisms and studies of gene networks, rather than individual genes, may indeed provide valuable insight into the relationship between cumulus cell gene expression and oocyte or embryo quality.  However, the importance of proper study design to eliminate differential gene expression related to patient or cycle-specific criteria cannot be overstated.