Scientists Use Stem Cells to Grow Three-Dimensional Mini Lungs.

In research done in several laboratories, lung tissue was derived from flat cell culture systems or by growing cells on scaffolds made from donated organs.

Now in a new study published in the online journal eLife, a multi-institution team has defined a culture system for generating self-organizing human lung organoids, which are three-dimensional structures that mimic the structure and complexity of human lungs.

“These mini lungs can mimic the responses of real tissues and will be a good model to study how organs form, change with disease, and how they might respond to new drugs,” said study senior author Jason R. Spence, Ph.D., an assistant professor of internal medicine and cell and developmental biology at the University of Michigan Medical School.

Spence and his colleagues successfully grew structures that resembled both the large airways or bronchi and small lung sacs, known as alveoli.

These mini lung structures were developed in a cell culture system. Therefore, they lack several components of the human lung, including blood vessels, which are a critical component of gas exchange during breathing.

Despite that, these cultured organoids can serve as a unique research model system for researchers as they grind out basic science ideas that are turned into clinical innovations. These three-dimensional mini-lungs should be an excellent complement to research in liver laboratory animals.

Traditionally, the behavior of cells has been investigated in the laboratory in two-dimensional culture systems where cells are grown in thin layers on cell-culture dishes. Most cells in the body, however, exist in a three-dimensional environment as part of complex tissues and organs. Tissue engineered have been trying to re-create these environments in the laboratory by successfully generating small version of particular organs known as organoids, which serve as models of the stomach, brain, liver and human intestine. The advantage of growing three-dimensional structures of lung tissue, according to Dr. Spence, is that the organization of organoids bears greater similarity to the human lung.

To make these lung organoids, researchers at the U-M’s Spence Lab and colleagues from the University of California, San Francisco; Cincinnati Children’s Hospital Medical Center; Seattle Children’s Hospital and University of Washington, Seattle manipulated several of the cell signaling pathways that control the formation of organs.

First, stem cells were induced to form a type of tissue called endoderm, which is found in early embryos and gives rise to the lung, liver and several other internal organs. Second, the group activated two important development pathways (FGF and WNT signaling ) that are stimulate endoderm to form three-dimensional tissue. By inhibiting two other key development pathways at the same time (BMP and TGFβ signaling), the endoderm became tissue that resembles the early lung found in embryos.

In the laboratory, this early culture-derived lung-like tissue spontaneously formed three-dimensional spherical structures as it developed. Afterwards, they had to expand these structures and develop them into lung tissue. In order to do this, Spence and his colleagues and collaborators exposed the cells to additional proteins involved in lung development (FGF and Hedgehog).

After all this manipulation, the resulting lung organoids survived in the laboratory for over 100 days.

“We expected different cells types to form, but their organization into structures resembling human airways was a very exciting result,” said author Briana Dye, a graduate student in the U-M Department of Cell and Developmental Biology.

While this type of experiment is remarkable, this is only the beginning of lung tissue engineering.  These mini-lungs  will hopefully serve and new model systems for drug testing and researching genetic diseases that affect the lungs, such as cystic fibrosis, sarcoidosis, or inherited forms of emphysema.  It will be a while before scientists can make replacement lungs for human patients, but these experiments by Spence and others are a remarkable start.

Induced Pluripotent Stem Cells Make Lungs

Since my father died of disseminated lung cancer (squamous cell carcinoma), this report has particular meaning to me.

When a person dies, their lungs can be harvested and stripped of their cells. This leaves a so-called “lung scaffold” that can then be used to build new lungs by means of tissue engineering techniques. Lung scaffolds consist of a protein called collagen, and sugar-rich proteins called “proteoglycans” (say that fast five times) and a rubber band-like protein called elastin. Depending on how the lung scaffolds are made more or less of these components can remain in the lung scaffold (see TH Peterson, and others, Cells Tissues Organs. Feb 2012; 195(3): 222–231). The important thing is that the cells are gone and this greatly reduces the tendency for the lung scaffold to be rejected by someone else’s immune system.

Once a lung scaffold is generated from a whole lung, cells can be used to reconstitute the lung. The key is to use the right cell type or mix of cell types and to induce them to form mature lung tissue.

The laboratory of Harald Ott at Harvard University Medical School used a technique called “perfusion decellularization” to make lung scaffolds from the lungs of cadavers. Then he and his co-workers used lung progenitor cells that were derived from induced pluripotent stem cells (iPSCs). This study was published in The Annals of Thoracic Surgery, and it examined the ability of iPSCs to regenerate a functional pulmonary organ

Whole lungs from rat and human cadavers were stripped of their living material by means of constant-pressure perfusion with a strong detergent called sodium dodecyl sulfate (SDS; 0.1% if anyone is interested). Ott and his crew then sectioned some of the resulting lung scaffolds and left others intact, and then applied human iPSCs that had been differentiated into developing lung tissue.

Lung tissue develops from the front part of the developing gut. This tissue is called “endoderm,” since it is in the very innermost layer of the embryo.

Lung Development

Therefore, the iPSCs were differentiated into endoderm with a cocktail of growth factors (FGF, Wnt, Retinoic acid), and then further differentiated in the anterior endoderm (foregut; treated cells with Activin-A, followed by transforming growth factor-β inhibition), and then even further differentiated into anterior, ventral endoderm, which is the precise tissue from which lungs form. In order to be sure that this tissue is lung tissue, they must express a gene called NK2 homeobox 1 (Nkx2.1). If these cells express this gene, then they are certainly lung cells.

Ott and his group showed that their differentiate iPSCs strongly expressed Nkx2.1, and then seeded them on slices and whole lung scaffolds. Then Otts’s group maintained these tissues in a culture system that was meant to mimic physiological conditions.

Those cells cultured on decellularized lung slices divided robustly and committed to the lung lineage after 5 days. Within whole-lung scaffolds and under the physiological mimicking culture, cells upgraded their expression of Nkx2.1. When the culture-grown rat lungs were transplanted into rats, they were perfused and ventilated by host vasculature and airways.

Thus these decellularized lung scaffolds supports the culture and lineage commitment of human iPSC-derived lung progenitor cells. Furthermore, whole-organ scaffolds and a culture system that mimics physiological conditions, allows scientists to enable seeding a combination of iPSC-derived endothelial and epithelial progenitors and enhance early lung fate. Transplantation of these laboratory-grown lungs seem to further maturation of these grafted lung tissues.

Prostaglandin E Switches Endoderm Cells From Pancreas to Liver

The gastrointestinal tract initially forms as a tube inside the embryo. Accessory digestive organs sprout from this tube in response to inductive signals from the surrounding mesoderm. Both the pancreas and the liver form at about the same time (4th week after fertilization) and at about the same place in the embryonic gut (the junction between the foregut and the midgut).

Pancreatic development

The pancreas forms as ventral and dorsal outgrowths that eventually fuse together when the gut rotates. The liver forms from the “hepatic diverticulum” that grows from the gut about 23-26 days after fertilization. These liver bud cells work with surrounding tissues to form the liver.

Liver development

What determines whether an endodermal cell becomes a liver or pancreatic precursor cell?

Wolfram Goessling and Trista North from the Harvard Stem Cell Institute (HSCI) have identified a gradient of the molecule prostaglandin E (PGE) in zebrafish embryos that acts as a liver/pancreas switch.

Postdoctoral researcher Sahar Nissim in the Goessling laboratory has uncovered how PGE toggles endodermal cells between the liver-pancreas fate. Nissim has shown that endodermal cells exposed to more PGE become liver cells and those exposed to less PGE become pancreas. This is the first time that prostaglandins have been reported as the factor that can switch cell identities from one fate to another.

After completing these experiments, HSCI scientists collaborated with colleague Richard Mass to determine if their PGE-mediated cell fate switch also occurred in mammals. Here again, Richard Sherwood from the Mass established that mouse endodermal cells became liver if exposed to PGE and pancreas if exposed to less PGE.  Sherwood also demonstrated that PGE enhanced liver growth and regeneration.

Goessling become interested in PGE in 2005, when a chemical screen identified PGE as an agent that amplified blood stem cell populations in zebrafish embryos. Goessling that transitioned this work to human patients, and a phase 1b clinical trial that uses PGE to increase umbilical cord blood transplants has just been completed.

PGE might be useful for instructing pluripotent human stem cells that have been differentiated into endodermal cells to form completely functional, mature liver cells that can be used to treatment patients with liver disease.

Repopulation of Damaged Livers With Skin-Derived Stem Cells

Patients with severe liver disease must receive a liver transplant. This major procedure requires that the patient survives major surgery and then takes anti-rejection drugs for the rest of their lives. In general, liver transplant patients tend to fair pretty well. The one-year survival rate of liver transplant patients approaches 90% (see O’Mahony and Goss, Texas Heart Institute Journal 2012 39(6): 874-875).

A potentially better way to treat liver failure patients would be to take their own liver cells, convert them into induced pluripotent stem cells (iPSCs), differentiate them into liver cells, and use these liver cells to regenerate the patient’s liver. Such a treatment would contain a patient’s own liver cells and would not require anti-rejection drugs.

Induced pluripotent stem cells or iPSCs are made from genetically-engineered adult cells that have had four specific genes (Oct4, Klf4, Sox2, and c-Myc) introduced into them. As a result of the heightened expression of these genes, some of the adult cells dedifferentiate and are reprogrammed into cells that resemble embryonic stem cells. Normally, this procedure is relatively inefficient, slow, and induces new mutations into the engineered cells. Also, when iPSCs are differentiated into liver cells (hepatocytes), they do not adequately proliferate after differentiation, and they also fail to properly function the way adult hepatocytes do.

New work from laboratories at the University of California, San Francisco (UCSF), has differentiated human hepatocytes by means of a modified technique that bypasses the pluripotency stage. These cells were then used to repopulate mouse livers.

“I really like this paper. It’s a step forward in the field,” said Alejandro Soto-Gutiérrez, assistant professor of pathology at the University of Pittsburgh, who was not involved in the work. “The concept is reprogramming, but with a shortcut, which is really cool.”

Research teams led by Holger Willenbring and Sheng Ding isolated human skin cells called fibroblasts and infected them with engineered viruses that forced the expression of three genes: OCT4, SOX2, and KLF4. These transduced cells were grown in culture in the presence of proteins called growth factors and small molecules in order to induce reprogramming of the cells into the primary embryonic germ layer known as endoderm. In the embryo, the endoderm is the inner-most layer of cells that forms the gastrointestinal tract and its associated structures (liver, pancreas, and so on). Therefore, the differentiation of adult cells into endodermal progenitor cells provides a handy way to form a cell type that readily divides and can differentiate into liver cells.

“We divert the cells on their path to pluripotency,” explained coauthor Holger Willenbring, associate professor of surgery at UCSF. “We still take advantage of what is intrinsic to reprogramming, that the cells are becoming very plastic; they’ve become flexible in what kind of cell type they can be directed towards.”

The authors called these cells induced multipotent progenitor cells (iMPCs). The iMPCs were easily differentiated into endodermal progenitor cells (iMPC-EPCs). These iMPC-EPCs were grown in culture with a cocktail of small molecules and growth factors to increase iMPC-EPC colony size while concomitantly maintain them in an endodermal state. Afterwards, Willenbring and others cultured these cells with factors and small molecules known to promote liver cell differentiation. When these iMPC-Hepatocytes (Heps) were transplanted into mice with damaged livers, the iMPC-Hep cells continued to divide at least nine months after transplantation. Furthermore, the transplanted cells matured and displayed gene expression profiles very similar to that of typical adult hepatocytes. Transplantation of iMPC-Heps also improved the survival of a mouse model of chronic liver failure about as well as did transplantation of adult hepatocytes.

“It is a breakthrough for us because it’s the first time that we’ve seen a cell that can actually repopulate a mouse’s liver,” said Willenbring. Willenbring strongly suspects that iMPCs are better able to repopulate the liver because the derivation of iMPC—rather than an iPSC—eliminates some steps along the path to generating hepatocytes. These iMPCs also possess the ability to proliferate in culture to generate sufficient quantities of cells for therapeutic purposes and, additionally, can functionally mature while retaining that proliferative ability to proliferate. Both of these features are important prerequisites for therapeutic applications, according to Willenbring.

Before this technique can enter clinical trials, more work must be done. For example: “The key to all of this is trying to generate cells that are identical to adult liver cells,” said Stephen Duncan, a professor of cell biology at Medical College of Wisconsin, who was not involved in the study. “You really need these cells to take on all of the functions of a normal liver cell.” Duncan explained that liver cells taken directly from a human adult might be able to repopulate the liver in this same mouse model at levels close to 90 percent.

Willenbring and his colleagues observed repopulation levels of 2 percent by iMPC-Heps, which is substantially better than the 0.05 percent repopulation typically accomplished by hepatocytes derived from iPSCs or embryonic stem cells. However: “As good as this is, the field will need greater levels of expansion,” said Ken Zaret of the Institute for Regenerative Medicine at the University of Pennsylvania, who did not participate in the work. “But the question is: What is limiting the proliferative capacity of the cells?”

Zaret explained that it is not yet clear whether some aspect of how the cells were programmed that differed from how they normally develop could have an impact on how well the population expands after transplantation. “There still is a ways to go [sic],” he said, “but [the authors] were able to show much better long-term repopulation with human cells in the mouse model than other groups have.”

See S. Zhu et al., “Mouse liver repopulation with hepatocytes generated from human fibroblasts,” Nature, doi:10.1038/nature13020, 2014.

Human Fat Contains Multilineage Differentiating Stress Enduring Cells With Great Potential for Regenerative Medicine

A collaboration between American and Japanese scientists has discovered and characterized a new stem cell population from human fat that do not cause tumors and can differentiate into derivatives from ectoderm, mesoderm, and endoderm.

Multilineage Differentiating Stress-Enduring or Muse cells are found in bone marrow and the lower layers of the skin (dermis). Muse cells are a subpopulation of mesenchymal stem cells, and even express a few mesenchymal stem cell-specific genes (e.g., CD105, a cell-surface protein specific to mesenchymal stem cells). However, Muse cells also express cell surface proteins normally found in embryonic stem cells (e.g., stage-specific embryonic antigen-3, SSEA-3). Additionally, Muse cells have the ability to self-renew, and differentiate into cell types from all three embryonic germ layers, ectoderm (which forms skin and brain), mesoderm, (which forms muscle, bone, kidneys, gonads, heart, blood vessels, adrenal glands, and connective tissue), and endoderm (which forms the gastrointestinal tract and its associated tissues). Finally, Muse cells can home to damaged sites and spontaneously differentiate into tissue-specific cells as dictated by the microenvironment in which the cells find themselves.

A new publication by Fumitaka Ogura and others from Tohoku University Graduate School of Medicine in Sendai, Japan and Saleh Heneidi from the Medical College of Georgia (Augusta, Georgia), and Gregorio Chazenbalk from the David Geffen School of Medicine at UCLA has shown that Muse cells also exist in human fat.

The source of cells came from two places: commercially available fat tissue and freshly collected fat from human subjects, collected by means of liposuction. After growing these cells in culture, the mesenchymal stem cells and Muse cells grew steadily over the 3 weeks. Then the Dezawa research group used fluorescence-activated cell sorting (FACS) to isolate from all these cells those cells that express SSEA-3 on their cell surfaces.

FACS uses antibodies conjugated to dyes that can bind to specific cell proteins. Once the antibodies bind to cells, the cells are sluiced through a small orifice while they are illuminated by the laser. The laser activates the dyes if the cell fluoresces, one door opens and the other closes. The cell goes to one test tube. If the cell does not fluoresce, then the door stay shut and another door opens and the cell goes into a different test tube.  In this way, cells with a particular cell-surface protein are isolated from other cells that do not have that cell-surface protein.

Fluorescent-Activated Cell Sorting
Fluorescent-Activated Cell Sorting

In addition to expression SSEA-3, the fat-based Muse cells expressed other mesenchymal stem cell-specific cell-surface proteins (CD29, CD90), but they did not express proteins usually thought to be diagnostic for fat-based mesenchymal stem cells (MSCs) such as CD34 and CD146.  Muse cells also expressed pluripotency genes (Nanog, Oct3/4, PAR4, Sox2, and Tra-1-81).  The Muse cells grew in small clusters and some cell expressed ectodermal-specific genes (neurofilament, MAP2), others expressed mesodermal-specific genes (smooth muscle actin, NKX2) and endodermal-specific genes (alpha-fetoprotein, GATA6).  These data suggested that the cultured Muse cells were poised to form either ectoderm, mesodermal, or endodermal derivatives.

When transplanted into mice with non-functional immune systems, the Muse cells never formed any tumors or disrupted the normal structure of the nearly tissues.  When placed in differentiating media, fat-derived Muse cells differentiated into cells with neuron-like morphology that expressed neuron-specific genes (Tuj-1), liver cells, and fat.  When compared with Muse cells from bone marrow or skin, the fat-derived Muse cells were better at making bone, fat, and muscle, but not as good as bone marrow Muse cells at making neuronal cell types, but not as good at making glial cells.  Many of these assays were based on gene expression experiments and not more rigorous tests.  Therefore, the results of these experiments might be doubtful until they are corroborated by more rigorous experiments.

These cells are expandable and apparently rather safe to use.  More work needs to be done in order to fully understand the full regenerative capacity of these cells and protocols for handling them must also be developed.  However, hopefully pre-clinical experiments in rodents will give way to larger animal experiments.  If these are successful, then maybe human trials come next.  Here’s to hoping.

Forming Induced Pluripotent Stem Cells Inside a Living Organism

A team from the Spanish National Cancer Research Centre (CNIO) has become the first research team to convert adult cells that are still within a living organism into cells that show characteristics of embryonic stem cells.

The CNIO researchers also say that these embryonic stem cells, which were obtained directly from inside an organism, have a broader capacity for differentiation than those obtained by means of an in vitro culture system. Specifically, they have the characteristics of totipotent cells, a primitive state never before obtained in a laboratory, according to the CNIO team.

Manuel Serrano, Ph.D., director of CNIO’s Molecular Oncology Program and head of the Tumor Suppression Laboratory, led this study. It was supported by Manuel Manzanares, Ph.D., and his team from the Spanish National Cardiovascular Research Centre.

The CNIO researchers say their work extends that of Nobel Prize winner Shinya Yamanaka, M.D., Ph.D., one step forward. Yamanaka opened a new horizon in regenerative medicine when, in 2006, he demonstrated that stem cells could be created from adult cells by using a cocktail of genes. But while Yamanaka induced his cells in culture in the lab (in vitro), the CNIO team created theirs directly in mice (in vivo). Generating these cells within an organism brings this technology even closer to regenerative medicine, they say.

In a study published online Sept. 11 in the journal Nature, the CNIO research team details how it used genetic manipulation techniques to create mice in which Dr. Yamanaka’s four genes could be activated at will. When these genes were activated, they observed that the adult cells were able to de-differentiate into embryonic stem cells in multiple tissues and organs.

María Abad, Ph.D., lead author of the article and a researcher in Dr. Serrano’s group, said, “This change of direction in development has never been observed in nature. We have demonstrated that we can also obtain embryonic stem cells in adult organisms and not only in the laboratory.”

Dr. Serrano added, “We can now start to think about methods for inducing regeneration locally and in a transitory manner for a particular damaged tissue.” Stem cells obtained in mice also show totipotent characteristics never generated in a laboratory. Totipotent cells can form all the cell types in a body, including the placental cells. Embryonic cells within the first couple of cell divisions after fertilization are the only cells that are totipotent.

The researchers reported that they were also able to induce the formation of pseudo-embryonic structures in the thoracic and abdominal cavities of the mice. These pseudo-embryos displayed the three layers typical of embryos (ectoderm, mesoderm, and endoderm), and extra-embryonic structures such as the vitelline membrane, which surrounds the egg, and even signs of blood cell formation, which first appears in the primary embryonic vesicle (otherwise known as the “yolk sac”).

“This data tell us that our stem cells are much more versatile than Dr. Yamanaka’s in vitro inducted pluripotent stem cells, whose potency generates the different layers of the embryo but never tissues that sustain the development of a new embryo, like the placenta,” the CNIO researcher said.  Below is a figure from their paper.  The pictures look pretty convincing.

a, Cysts in the abdominal cavity of a reprogrammable mouse. b, Frequency of embryo-like structures after intraperitoneal injection of in vivo iPS cells (3 clones), in vitro iPS cells (2 clones) and ES cells (JM8.F6). Fisher’s exact test: *P < 0.05. c, Cyst generated by intraperitoneal injection. Left panels, germ layer markers: SOX2 (ectoderm), T/BRACHYURY (mesoderm) and GATA4 (endoderm). Right panels, extraembryonic markers: CDX2 (trophectoderm), and AFP and CK8, both specific for visceral endoderm of the yolk sac. d, Cyst generated by intraperitoneal injection presenting TER-119+ nucleated erythrocytes and LYVE-1+ endothelial cells in structures resembling yolk sac blood islands.
a, Cysts in the abdominal cavity of a reprogrammable mouse. b, Frequency of embryo-like structures after intraperitoneal injection of in vivo iPS cells (3 clones), in vitro iPS cells (2 clones) and ES cells (JM8.F6). Fisher’s exact test: *P < 0.05. c, Cyst generated by intraperitoneal injection. Left panels, germ layer markers: SOX2 (ectoderm), T/BRACHYURY (mesoderm) and GATA4 (endoderm). Right panels, extraembryonic markers: CDX2 (trophectoderm), and AFP and CK8, both specific for visceral endoderm of the yolk sac. d, Cyst generated by intraperitoneal injection presenting TER-119+ nucleated erythrocytes and LYVE-1+ endothelial cells in structures resembling yolk sac blood islands.

The researchers emphasize that any possible therapeutic applications of their work are still distant, but they believe that it could mean a change of direction for stem cell research, regenerative medicine and tissue engineering.

“Our stem cells also survive outside of mice in a culture, so we can also manipulate them in a laboratory,” said Dr. Abad. “The next step is studying if these new stem cells are capable of efficiently generating different tissues such as that of the pancreas, liver or kidney.”

This paper is very interesting, but I find it rather unlikely that their approach will take regenerative medicine by storm.  Engineering mice to express these four genes in an inducible manner caused the formation of unusual tumors throughout the mice.  Maybe they can be coaxed to differentiate into kidney or heart muscle or whatever, but learning how to get them to do that will take a fair amount of in vitro work.  This is interesting, but I doubt that it will change the field overnight.

Using Tissue-Specific Mesenchymal Stem Cells to Make Insulin-Producing Beta Cells from Embryonic Stem Cells

Embryonic stem cell lines are made from four-five-day-old human embryos. At this stage of development, the embryo is a sphere of cells with two distinct cell populations; an outside layer of flat trophoblast cells and an inner clump of round inner cell mass (ICM) cells. The embryo consists of ~100 cells four days after fertilization, and ~150 cells five days after fertilization.


Embryonic stem cell (ESC) derivation involves the removal of the trophoblast cells (which are collectively called the trophectoderm) and the isolation of the ICM cells. There are several ways to remove the trophectoderm, but the most commonly-used technique is “immunosurgery,” which uses antibodies that bind to proteins on the surfaces of the trophectoderm, and serum to initiate destruction of the trophoblast cells. The isolated ICM cells are then cultured, and if they grow, they may produce an embryonic stem cell line.

Immunosurgery was first perfected by Davor Solter and Barbara B. Knowles on mouse embryos. They used an antiserum that was raised in rabbits when the rabbits were immunized against mouse spleen tissue. When mouse embryos were incubated with this antiserum plus serum from mice, all the cells of the mouse embryo died. However, if they used the rabbit antiserum and serum from guinea pig, then only the trophoblast cells were destroyed. For human embryonic stem cell derivation, the rabbits are immunized again human red blood cells, and this rabbit antiserum is used with guinea pig serum. The serum contains proteins called “complement,” which bind to cells that have antibodies attached to them and bore holes in those cells, thus destroying them.

When ICM cells are cultured, they are placed on a layer of mouse cells that have been treated with a chemical called mitomycin C to prevent them from dividing. These non-dividing cells act as “feeder cells” that keep the ICM cells from differentiating. Because ICM cells are grown on animal cells, they cannot be used for clinical purposes, since they will possess animal proteins can carbohydrates on their surfaces, which would be attacked by the patient’s immune system. However, several ESC lines have been derived without animal products, and it is possible to make ESC lines that would be safe or human use.

ESC derivation

ESC derivation results in the destruction of human embryos. There is not two ways about it. Even though there are potential ways around this problem, the majority of ESC lines were made literally over the dead bodies of very young human beings. All the rationalization in the world (the embryo is too young, too small, too inchoate, too unwanted, going to die anyway, in the wrong place at the wrong time) do not undo the fact that the embryo is a very young human being, and making an ESC line from it ends his/her life.

Getting the ESC line to differentiate in what you want it to be is another problem. If any undifferentiated cells remain after differentiation, they can cause tumors. Therefore, there is a need to ensure that differentiation is efficient and complete. To this end, Doug Melton’s lab at Harvard University has published a remarkable paper in the journal Nature that uses mesenchymal stem cells from particular organs to direct the differentiation of ESC lines.

Melton’s lab, in particular Julie M. Sneddon and Malgorzata Borowiak (say that fast five times), established 16 lines of tissue-specific mesenchymal stem cells (MSCs) from embryonic, neonatal and adult mouse intestine, liver, spleen, and pancreas and human pancreas too. Then they cultured mouse ESCs on these MSC lines to determine if they could drive the ESCs to differentiate into pancreas cells. In the embryo, pancreatic precursors express several genes in a nested, hierarchical fashion. First, they express Sox17, which is a common endodermal marker, and then pancreatic progenitors all express Pdx1. Of these pancreatic progenitors, some express Ngn3 and these will become endocrine rather than exocrine cells, and othe the Ngn3-expressing cells, a few will become beta cells that make insulin.

Melton and his co-workers tried to determine if any of these genes was up-regulated in their ESC lines if that were co-cultured with their established MSC lines. They discovered that four lines – MSC1, 2, 3, & 4, all affected gene expression when co-cultured with ESCs. MSC 1 and 2 induced and increase in Sox17 expression and MSC 3 and 4 increased the expression of Ngn3 in ESCs.

These changes in gene expression were due to increased cell proliferation of cells actually expressing these genes and not due to differential survival. Also, no combination of growth factors could achieve the same results as the accompanying MSC lines. Thus there is more going on here than the MSCs just secreting the right growth factors. The MSCs must be making contact with the ESCs and inducing them to differentiate into a particular cell type.

Next Melton and his colleagues determined if this interaction with MSCs caused the ESCs to lose their ability to self-renew. The answer was a clear “no.” Even though these ESC lines were expressing genes characteristic of endodermal or pancreatic tissue, they did not lose their ability to differentiate into pancreatic tissue when appropriately induced to do so, and they also id not lose their ability to self-renew and grow competently in culture.

In a more stringent test, these ESCs that had been grown on tissue-specific MSCs were implanted into mice. As Melton points out in the paper, the “most efficient published protocols for in vitro differentiation of pluripotent cells to beta-cells yield only a small percentage (typically 0-15%) of insulin-positive cells, and these do not secrete insulin in a glucose-responsive manner.” Could the MSC-conditioned ESCs do any better?

Before implantation, the ESCs were differentiated into endodermal progenitors (Sox17-expressing cells), and co-cultured with MSCs for at least 3-7 passages. Then they were differentiated into beta cells and transplanted into mice. There were a few important controls that were used; Just saline, implantations of MSCs alone, and ESCs that had been differentiated into beta cells, but had never been passaged on MSCs. Finally, human pancreatic islets were used as a positive control.

The results were interesting to say the least. The saline and MSC alone implantations showed no insulin production with or without glucose. Likewise the human pancreatic islets made insulin in a glucose-dependent manner (no surprise there). The ESC-derived beta cells that had never been passaged on MSCs made insulin, and even showed some ability to respond to glucose and make more insulin after glucose ingestion. However, the beta cells derived from ESCs that had been passaged on MSCs made insulin in a glucose-dependent manner. The experiment produced a wide range of variability since the number of transplanted cells differed between each trial, but the implanted beta cells derived from ESCs passaged on tissue-specific MSCs definitely performed the best, and even did as well or better than the implanted human beta cells in some cases.

a, Schematic depicting implantation of human ESC-derived progenitors. b, Immunofluorescence staining of human ESC-derived endoderm, passaged seven times on mesenchyme and engrafted for 3 months (top panel) or further differentiated to Pdx1+ stage and then engrafted for 2 months (bottom panel). c, Glucose-tolerance test of animals implanted with PBS or mesenchyme only, human islets or Pdx1+ pancreatic progenitors derived from unpassaged (P0), or passaged (P4 or P7) human endoderm. d, Fasting- and glucose-induced (45 min glucose) plasma human C-peptide levels. Pairs of bars represent two time points per animal; data represent mean of two technical replicates ± s.d.
a, Schematic depicting implantation of human ESC-derived progenitors. b, Immunofluorescence staining of human ESC-derived endoderm, passaged seven times on mesenchyme and engrafted for 3 months (top panel) or further differentiated to Pdx1+ stage and then engrafted for 2 months (bottom panel). c, Glucose-tolerance test of animals implanted with PBS or mesenchyme only, human islets or Pdx1+ pancreatic progenitors derived from unpassaged (P0), or passaged (P4 or P7) human endoderm. d, Fasting- and glucose-induced (45 min glucose) plasma human C-peptide levels. Pairs of bars represent two time points per animal; data represent mean of two technical replicates ± s.d.

Melton notes at the end of his paper that this technique worked rather well for coaxing ESCs to form pancreatic derivatives, but it could very well be applicable to other systems as well. Also, it could probably work with induced pluripotent stem cells, which have many (though not all) of the characteristics of ESCs and can be made without killing human embryos. Thus another technique for increasing ESC differentiation seems to be on the table.