First fully human ovarian organoids provide insights about female reproductive health and disease

Despite the fact that we all start out as an egg cell in one of our mother's ovaries, these human reproductive organs are surprisingly under-studied. Scientists have been working on creating in vitro models of human ovaries so that we can learn more about them and develop treatments for ovarian conditions, but most existing models use a combination of human and mouse cells, which do not faithfully replicate human ovary functions and take a long time to grow in the lab.

Now, researchers at the Wyss Institute for Biologically Inspired Engineering at Harvard University, Harvard Medical School (HMS) and Duke University in collaboration with Gameto, a biotechnology company developing therapies to improve the female reproductive journey, have created a living, fully human ovarian organoid that supports egg cell maturation, develops follicles, and secretes sex hormones. This "ovaroid" model enables the study of human ovarian biology without the need to take tissue from patients, and could enable the development of new treatments for conditions like infertility, ovarian cancer, and more. Through an agreement with Harvard's Office of Technology Development (OTD), this technology has been licensed to Gameto, which is using it to develop therapeutics for diseases of the female reproductive system. The ovaroids are described in detail in a new paper published today in eLife.

Our new method of fully human ovaroid production is several times faster than existing human/mouse hybrid methods, and replicates many of the critical functions of these organs, marking a significant step forward in our ability to study female reproductive health in the lab. In the future, similar technology could also treat infertility by growing egg cells from people whose own eggs aren't viable."

Merrick Pierson Smela, co-first author, graduate student in the lab of George Church, Ph.D. at the Wyss Institute and HMS

A supportive somatic scaffold from stem cells

The developing ovary contains both germ cells, which grow into egg cells, and somatic cells, which support the germ cells. Current lab models of ovaries use human germ cells and mouse somatic cells, but Smela and his co-authors wanted to see if they could coax human stem cells to grow into functional, fully human ovaries with both major cell types. They decided to focus their efforts on granulosa cells, a type of ovarian somatic cell that supports the development of unfertilized egg cells within follicles and secretes the sex hormones estradiol and progesterone. No method existed at the time for efficiently generating granulosa cells from human induced pluripotent stem cells (iPSCs), so they decided to create their own.

The burgeoning field of iPSC technology is based on the discovery that introducing proteins called transcription factors (TFs) - which bind directly to DNA and control whether certain genes are turned on or off - into human iPSCs can guide them to differentiate into different types of cells like neurons, fibroblasts, and many others. The team chose to pursue this strategy to produce human granulosa cells, and started by combing through datasets to identify TFs that are expressed differently in granulosa cells compared to other cell types. They found 35 candidate TFs, and used a technique called "piggyBac transposition" to insert the genes that coded for those TFs into the genomes of iPSCs.

After inducing the expression of their target TFs in the iPSCs, they screened the cells to see which ones also produced a protein called FOXL2, which is a known hallmark of granulosa cells. They identified six top TFs that were associated with FOXL2 expression: NR5A1, RUNX1/RUNX2, TCF21, GATA4, KLF2, and NR2F2. They then tested different combinations of these top candidates, and found that NR5A1 and either RUNX1 or RUNX2 consistently upregulated FOXL2. These combinations also drove the expression of two proteins called AMHR2 and CD82, which are surface markers found on granulosa cells.

The researchers then looked at the full transcriptome of their new cells, and found that they expressed a number of other genes that are known to be active in granulosa cells. Comparing their data with other studies of human fetal ovarian cells, they found that their cells were most similar in their gene expression to granulosa cells in a human ovary at 12 weeks of gestation – but had taken only five days to generate using their new method.

The team now needed to make sure that these new granulosa-like cells also replicated normal granulosa cell functions. One of those functions is the production of estradiol from the precursor molecule androstenedione, which is stimulated in the ovary by the presence of follicle-stimulating hormone (FSH). The researchers treated their granulosa-like cells with androstenedione, then added FSH. The cells successfully produced estradiol from androstenedione without the addition of FSH, and increased their production when FSH was added. They also produced progesterone, which granulosa cells secrete after ovulation.

100% human in 16% of the time

Now that they were confident that their granulosa-like cells functioned much like the real thing, the researchers co-cultured them with human primordial germ cell-like cells (hPGCLCs) to form ovarian organoids or "ovaroids" that included both germ cells and somatic cells.

"Creating the granulosa cells on their own was a significant accomplishment, but making an ovaroid out of only granulosa cells wouldn't tell us anything about their ability to support the maturation of germ cells, which was what we wanted to be able to study in vitro," said co-first author Christian Kramme, Ph.D., the Vice President of Cell Engineering at Gameto and a former graduate student in Church's group at the Wyss Institute and HMS. "This process had been replicated previously using hPGCLCs and mouse somatic cells, but with this new technology, we now have the ability to do it with a fully human model."

After four days of co-culturing their granulosa-like cells with hPGCLCs, the resulting ovaroids started to produce a protein called DAZL, which is a marker of germ cells that have embarked on their maturation journey. In contrast, ovaroids made with mouse somatic cells did not express DAZL until day 32. The human germ cells did not live long enough to develop further into egg cells, but the human ovaroids started to form empty, follicle-like structures composed of the granulosa-like cells after about 16 days, despite the fact that there were no egg cells present. By day 70, numerous follicles of varying sizes had formed within the ovaroids, some of which had developed multiple layers characteristic of mature follicles that are capable of supporting an egg.

"The efficient production of fully human ovaroids that replicate the hormonal signaling, germ cell maturation, and follicle formation seen in the human ovary is a feat in and of itself, but the fact that this can be done within five days instead of the month required with human/mouse hybrid ovaroids will dramatically speed up the discovery of critical information about women's health and reproduction," said senior author Church, who is a Core Faculty Member at the Wyss Institute as well as a Professor of Genetics at HMS.

The Wyss team is continuing to develop its human ovaroid model and plans to integrate additional ovarian cell types, including hormone-producing theca cells, to more fully replicate the complex functions of the human ovary. They also hope to improve their culture system to allow their germ cells to fully develop into egg cells, and determine the optimal dosage of the different TFs. Gameto, meanwhile, has conducted preclinical studies of a derived co-culture system for egg maturation in humans with leading national fertility clinics.

"Half of the human population is female, and yet historically women's health has not received anywhere near the attention or funding that is given to conditions that affect men. I'm very excited to see this important step forward in being able to study human ovaries in the lab, and look forward to the insights that such a model will provide about female reproductive health and disease," said Wyss Founding Director Don Ingber, M.D., Ph.D. Ingber is also the Judah Folkman Professor of Vascular Biology at HMS and Boston Children's Hospital, and the Hansjörg Wyss Professor of Bioinspired Engineering at the Harvard John A. Paulson School of Engineering and Applied Sciences.

Additional authors of the paper include Patrick Fortuna, Jessica Adams, Alina Su, and Edward Dong from the Wyss Institute; Mutsumi Kobayashi, Toshi Shioda, and Garyk Brixi from HMS; Pranam Chatterjee from Duke University; Emma Tysinger from MIT; and former Wyss Institute member Richie Kohman, who is now CSO of the Wyss Center for Bio and Neuroengineering.

This research was supported by the Wyss Institute at Harvard University, Harvard OTD, sponsored research agreements between the Wyss Institute and industry partners Gameto and Colossal, and the National Science Foundation.