Repainting the Pituitary Landscape

Always mystifying and bit mysterious, the pituitary is the gland that sends out the orders to the other glands in the body, from regulating the adrenals and the thyroid to coordinating reproduction. However, thanks to a worldwide collaboration and a new database, this “master gland” is slowly but surely unlocking its secrets.

The pituitary – the master gland that produces hormones controlling the functions of many other organs, ranging from blood pressure to sexual maturation and reproduction – has been left out of major single-cell atlases and consortia, and therefore it has not been characterized at single-cell resolution until very recently. There are a couple of reasons for this: animal models can be elusive, since the pituitary gland, about the size of a chickpea in humans, is obviously much smaller in mouse. Moreover, contrary to other organs included in human databases and atlases, the pituitary can only be studied postmortem in humans, which complicates the access to samples.

However, over the past couple of years, researchers at the Icahn School of Medicine at Mount Sinai in New York, along with other laboratories in the U.S., Canada, and the United Kingdom, have been mapping the landscape of the pituitary gland, looking at the gland’s cell types using single nucleus (sn) multi-omics assays, first in murine pituitaries, then in archived post-mortem human pituitaries, and adding their findings to a growing database. They hope this pituitary database will be a resource for others seeking to get a better understanding of this complex and important organ. The investigators published their mouse model-related findings in Nature Communications in May 2021, and followed on with a Cell Reports article in March 2022 describing the results obtained in human pituitary samples.

“Pituitary cell types have been well-established for years,” says Frederique Ruf-Zamojski, PhD, a biomedical experimental scientist and associate professor in Neurology at the Icahn School of Medicine at Mount Sinai in New York. “We know they consist of hormone-producing cells as well as non-hormone producing cells. However, determination of the exact composition, and cell-type specific characterization of the pituitary at the transcriptomic and epigenetic levels to look into potential cell states or subtypes had yet to be achieved.”

Utilizing Single-Nucleus Atlases

Ruf-Zamojski and her colleagues developed single-nucleus atlases instead of single-cell ones because access to healthy pituitary samples is made feasible from frozen archived samples stored in biobanks, and as such can only be analyzed using nuclei. She says that her laboratory developed and optimized specific protocols for frozen pituitaries, since at the time very few groups were using nuclei, and no commercial protocols for nuclei were published or even supported by commercial single-cell platforms.

“Interestingly we found stem cells at different stages, with many uncommitted stem cells, but also detected committed progenitors at all ages. Using pseudotime trajectory analyses, we further characterized their gene expression changes with age, and sex differences. Our sample size in this study was small, and additional studies will be needed to complete a multi omics characterization of human pituitary stem cells.” – Frederique Ruf-Zamojski, PhD, biomedical experimental scientist and associate professor in neurology, Icahn School of Medicine, Mount Sinai, New York, N.Y.

Ruf-Zamojski says that there were three main reasons to build single-nucleus atlases of murine and human pituitaries. First, the lab of Dr Sealfon with whom she is working focuses on the regulation of gonadotrope cells, which represent only 5% of pituitary cells. She explains that there are cell culture models of gonadotrope cells that, although proven useful to tackle some underlying mechanisms, remain imperfect, and do not represent physiological gonadotropes in their environment. “Having single-cell atlases of pituitaries enables the study of gonadotropes in vivo, revealing molecular aspects we cannot get from cell cultures,” she says.

Next, Ruf-Zamojski describes the need for having normal reference atlases for comparing and getting a better understanding of the molecular pathogenesis of neuroendocrine pituitary tumors/adenomas. These tumors may arise from any pituitary cell types; they represent heterogeneous and complex systems composed of different cell types and subtypes; they often recur, and recurrence may be from a single cell. “Having single-cell/nucleus atlases and single-cell resolution data is critical for elucidating the molecular mechanisms underlying pituitary tumor development,” she says.

Finally, Ruf-Zamojski says, it is important to have cell-type specific transcriptomic and epigenetic information to study any gene in the genome and get basic information on its expression and regulation.

Finding Lost Signals

This past summer, Ruf-Zamojski presented some of her lab’s work at ENDO 2022 in Atlanta. During the session, “Breakthroughs in Understanding Pituitary Networks,” she pointed to the power of multi-omics assays for studying the regulatory network and gene control mechanisms relevant to physiology and disease. “We are at a stage where single-cell resolution studies are finally possible, reliable, and reproducible for a detailed characterization of cell types in complex tissues,” she tells Endocrine News. “Our single nucleus multi-omics assays of murine and archived post-mortem human pituitaries represent references of normal pituitaries. It finally brings insights into the characterization of pituitary cell types and cell states in selected conditions.”

In past investigations, researchers used bulk approaches to examine the pituitary, yet those studies had limitations. Ruf-Zamojski likens those approaches to analyzing a bowl of different types of fruits. The bulk approach would be the equivalent of turning that fruit bowl into a smoothie; all the fruits would still be present, but the smaller or less tasty fruits would not be as discernible in the smoothie. “Imagine for example having a single blueberry in it, you will probably not identify it, although it’s present in your drink,” she says. “In a parallel way, we have been able to analyze pituitary responses in bulk approaches, but some signals are lost and cannot be detected even though they are important.”

In her lab’s analysis of pituitary gonadotropes, Ruf-Zamojski and her colleagues focus on the transcription of Fshb, a gene expressed in gonadotropes only. A requirement for Fshb gene transcription is that chromatin be open at the Fshb promoter for the recruitment of transcription factors and for transcription to initiate. “However, if you use a whole pituitary approach, you cannot detect any opening at the level of the Fshb promoter,” Ruf-Zamojski says. “This is not that the chromatin is not accessible in gonadotropes, but it is simply because the signals coming from around 5% of gonadotrope cells within the whole pituitary are too low to be detected in a bulk experiment.”

Single-cell and single-nucleus assays are again the equivalent of a bowl of fruits, but now more like a fruit salad, in which each ingredient can be easily identified and classified – even a single blueberry could be seen and picked up. “The same is true for pituitary analyses at the single-cell level, where even rarer cell types are identified,” Ruf-Zamojski says. “Now we can characterize specifically our gonadotrope cells among all the pituitary cells, computationally extract them from all the other cells analyzed, and we can look specifically at the accessibility of the Fshb gene within the gonadotrope cluster. Unsurprisingly, when Fshb is expressed, we also find that the Fshb promoter is accessible in the gonadotropes only, something we could not detect in a bulk assay, although the same cells/signals were there.”

“So, using single-cell/nucleus atlases, one can now look specifically at any pituitary cell types or subgroups of cells within the pituitary, and get a better understanding of the cellular functions within the cell type of interest,” she continues. “We generated our atlases in normal healthy pituitaries, and these could now be extended to diverse conditions and samples.”

Potential of Stem Cells

For their work published in Nature Communications, the researchers profiled the transcriptome, chromatin accessibility, and methylation status of more than 70,000 single nuclei from adult murine pituitaries and provided a multi-omics resource to investigate transcriptional regulatory mechanisms. “Our study determined epigenetically-defined cell type composition, identified cell type-specific and sex-specific differences in transcriptional and epigenetic programs, an experimentally supported cis-regulatory domain, and epigenetic mechanisms contributing to cell type-specific and sex-specific regulon composition,” the authors write. “Our work lays the foundation for characterizing the general epigenetic regulatory principles that control cell type-specific animal-specific and sex-specific gene expression programs.”

In the Cell Reports paper, the researchers describe a human single-nucleus RNA-seq and ATAC-seq resource from pediatric, adult, and aged postmortem pituitaries and characterize cell-type-specific gene expression and chromatin accessibility programs for all major pituitary cell lineages. Human pituitary stem cells (PSCs) have been incompletely characterized, the authors write, and they were able to identify uncommitted PSCs, committing progenitor cells, and sex differences as well as distinct deterministic mechanisms that contribute to heterogeneous marker expression within PSCs. “These findings characterize human stem cell lineages and reveal diverse mechanisms regulating key PSC genes and cell type identity,” they write.

“Without high-throughput single-cell approaches, studying these cells remained challenging,” Ruf-Zamojski says. “But there were already studies and interest in PSCs for years, including from our collaborator Dr. Cynthia Andoniadou. Stem cells and their regenerative potential are a very active area of investigation. Studies have started to characterize murine PSCs, but human PSCs have largely not been analyzed. Our recent paper in Cell Reports is part of that effort. We found PSCs at all ages in human postmortem pituitary samples, and they express similar markers as observed in mouse (including SOX2, SOX9, WWTR1, YAP1, etc.).”

“Interestingly we found stem cells at different stages, with many uncommitted stem cells, but also detected committed progenitors at all ages,” she continues “Using pseudotime trajectory analyses, we further characterized their gene expression changes with age, and sex differences. Our sample size in this study was small, and additional studies will be needed to complete a multi-omics characterization of human pituitary stem cells.”

“We are at a stage where single cell resolution studies are finally possible, reliable, and reproducible for a detailed characterization of cell types in complex tissues. Our single nucleus multi-omics assays of murine and archived postmortem human pituitaries represent references of normal pituitaries. It finally gives insights into the characterization of pituitary cell types and cell states in selected conditions.” – Frederique Ruf-Zamojski, PhD, biomedical experimental scientist and associate professor in neurology, Icahn School of Medicine, Mount Sinai, New York, N.Y.

Ruf-Zamojski is careful here, noting that the researchers included only six pituitaries in their human postmortem study – three from male donors and three from female donors, all of very different ages. The researchers detected a large heterogeneity across the different samples, which could be attributed to age, but also to other factors, since humans do not live in controlled environments. The pituitary samples were all considered “healthy,” but Ruf-Zamojski says that just means without neurological or pituitary disorders. “But other factors like diabetes, drug use, medication, obesity, may contribute to the variation observed, and unfortunately with postmortem samples from biobanks, we have to work with de-identified samples, meaning that we have only access to basic metadata and information, not the full medical history,” she says. “Thus, a larger cohort with, if possible, an access to metadata for subgroup classifications, should be studied to draw more definite conclusions.”

Research Resource: Connecting Collaborators Around the World

To help researchers develop additional studies that will allow to reach more definite conclusions, Ruf-Zamojski and her colleagues presented their datasets and analyses as a resource for the research community, which can be accessed at snpituitaryatlas.princeton.edu. Ruf-Zamojski says that the datasets relied on the contribution of many wet lab and dry lab scientists in her group, as well as collaborators from around the world. “These projects were very collaborative and have shown the importance of bringing together teams from different fields and subfields to tackle an important project,” she says. “We are lucky to have collaborators from around the country and around the world join our lab meetings, thus providing a diverse expertise that helps in the design and interpretation of these studies.”

Ruf-Zamojski says that their human study set the stage for analyzing human pituitaries at single-cell resolution, an important first step, since it showed how studying postmortem samples can give important information on pituitary cell types and can be used for additional targeted studies. She thinks studies will now focus on different age groups, probably comparing normal with diseased pituitaries, and trying to understand the molecular mechanisms underlying any pituitary dysfunction, with the objective being to establish better prognosis tools and improve therapeutic options.

“We hope our datasets can be a resource for the community,” Ruf-Zamojski says. “We believe sharing datasets and resources (pipelines, algorithms, reagents, protocols) is critical for research to move forward. With our database, any researcher or clinician without any knowledge in coding or software can go and query any genes of interest, and find where it is expressed (which cell type), what is the chromatin accessibility profile around that gene, and in mouse what is the methylation pattern around that gene in specific cell types.”

Bagley is the senior editor of Endocrine News. In the November issue, he wrote about “rebranding” diabetes insipidus with a new name.