Skip to Main Content Menu Search Site

Programming complex tissue organizations in 3D

Simultaneous, programmable generation and organization of cell types with different origins and functions in artificial tissues could invigorate future biomedical 3D organ and tissue engineering

By Benjamin Boettner

(BOSTON) — Tissue engineers have developed a variety of approaches to recreate organs and tissues de novo outside the human body for use in regenerative therapies, drug screening, and disease modeling. As two prominent examples, human “organoids” are being assembled from cells to form tiny artificial constructs in a dish that resemble human organs and larger human tissues are being bioprinted. Both approaches use so-called induced pluripotent stem cells (iPSCs) as a cell source, which are derived from blood, skin and other cells provided by human donors, and then differentiated into cell types typical for an organ of interest.

Roadblocks to tissue engineering

ImageJ=1.52p
unit=micron
This image shows an organoid structure generated from human pluripotent stem cells (iPCSs) that were programmed to create an organized structure with a vascular network (red) and neuronal cells (blue) similar to those found in the brain cortex. Credit: Wyss Institute at Harvard University

A growing number of different cell and tissue types are becoming accessible via these methods, and researchers are even beginning to address the central challenge of integrating a blood support system in the form of vascular networks. However, what they still lack is precise control over the exact composition of cell types and their spatial organization into functional units within the 3D spaces of their constructs. Additionally, to progress towards whole-organ biomanufacturing, these artificial tissue constructs must also be generated with greater efficiency, speed, and scalability than current protocols allow.

Now, a collaborative and highly multidisciplinary research team at the Wyss Institute for Biologically Inspired Engineering at Harvard University, Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS), and Harvard Medical School (HMS) has developed an integrated approach to overcome these challenges. By independently and simultaneously programming distinct populations of iPSCs within organoids and 3D bioprinted tissue constructs with cell fate-specifying transcription factors (TFs), they induce them to differentiate into desired cell types that can be assembled in in vivo-like “organotypic” patterns with unprecedented control. Their findings are published in Nature Biomedical Engineering.

“We are leveraging genetic programming of stem cells and organoids, coupled with bioprinting to push the envelope of programmability and multicellular organization in 3D tissue constructs for disease modeling and regenerative medicine,” said senior author and Wyss Core Faculty member Jennifer Lewis, Sc.D. Lewis also is a Lead on the Wyss Institute’s 3D Organ Engineering Initiative and the Hansjörg Wyss Professor of Biologically Inspired Engineering at SEAS.

We are leveraging genetic programming of stem cells and organoids coupled with bioprinting to push the envelope of programmability and multicellular organization in 3D tissue constructs for disease modeling and regenerative medicine.

Jennifer Lewis

Controlling cell differentiation in 3D artificial tissues

To differentiate iPSCs towards a specific cell type, they need to be exposed to specific cues. These can either be chemical and biological factors added to the media that iPSCs are cultured in, pushing them to adopt a specific cell identity. Alternatively, they equip iPCSs with cell fate-instructive transcription factors (TFs) that turn on the gene expression programs driving iPSCs to differentiate towards a specific cell type, bypassing chemical cues in the media conditions. Co-senior author George Church, Ph.D., and his group compiled the most complete collection of TFs with the potential to generate many differentiated human cell types from iPSCs.

“A key challenge to engineering tissue constructs composed of multiple cell types that are organized in patterned structures was that cell culture media can only specify the generation of one or a few cell type(s),” said co-first and co-corresponding author Mark Skylar-Scott, Ph.D. In 2016, as a Postdoctoral Fellow working with Lewis, he discussed the potential of TFs for expanding the range and organization of cells in engineered tissues with Alex Ng, Ph.D., a former Postdoctoral Fellow on Church’s team Ng who spearheaded the TF discovery effort, and is now commercializing a TF-based technology platform as CSO of GC Therapeutics Inc., which, like Church, he co-founded.

“We hypothesized that expressing cell type-specifying TFs could enforce parallel and independent differentiation programs in cell lines even when they are mixed, and thus override limiting media conditions,” said Skylar-Scott, who now is Assistant Professor of Bioengineering at Stanford University.

ImageJ=1.52p
unit=micron
The OID system for stem cell programming enabled the team to generate 3D organoids with a neuronal organization that included differentiated neuronal cells that are linked into a neuronal network via their neuronal processes (green) and neurospheres resembling brain ventricles (blue). The entire construct is supported by a vascular network (red). Wyss Institute at Harvard University

To validate their idea, the collaborators mixed and aggregated iPSCs that turn on the TF ETV2 upon the addition of a drug, which induces them to differentiate into endothelial cells (iEndo cells), iPSCs that via expression of the TF NGN1 generate neuronal cells (iNeuron cells), and regular iPSCs that in the chosen media, by default, generate neuronal stem cells. The iEndo and iNeuron cells differentiated rapidly and more efficiently into endothelial and neuronal cells than earlier methods allowed, and at tailorable ratios. They then went on to form integrated vascular and neuronal networks, whereas the uninduced iPSCs autonomously differentiated into small neurospheres in the developing organoids that resembled brain ventricles.

“This showed that the method we devised enables the highly programmable co-differentiation of cell types with completely different origins in normal embryogenesis and organogenesis, which cannot be easily done in a single tissue construct otherwise,” said co-first author Jeremy Huang, Ph.D., who was mentored on the project as a graduate student by Lewis and Church. “We therefore chose ‘orthogonally induced differentiation’ (OID) as a name for it.” Huang now is a Postdoctoral Fellow at MIT.

Drawing patterns with programmed cells

Next, the team used the same OID-generated cell lines to create patterned tissue constructs using 3D organoid and bioprinting technologies. Different neuronal cell populations and microvascular cells form distinct layers in regions in and around the brain. Being able to pattern cells similarly in vitro would allow researchers to study their relationships close-up, such as in the case of the brain cortex and its formation of neural networks, migration of cells between its different layers, and layer-specific interactions with the brain microvasculature. The team applied a layering technique to combine the different cells in a “multicore-shell” architecture with an inner core containing normal iPSCs, surrounding layer containing iNeuron cells, and an outer layer containing iEndo cells. When induced to express the different TFs, the multicore-shell organoids differentiated as a unit into a structure with an inner “germinal zone center” in which stem cells reside and multiply, an outer zone composed of differentiated neurons that in normal brain development originate from stem cells in the germinal center that have embarked on their differentiation program and moved out of the germinal center, and a vascular network that encapsulates the neuronal compartment.

“Importantly, these patterned organoids developed very consistently and with differentiated neurons after eight days in culture, whereas brain organoids usually require several weeks of culture to generate neurons. By analyzing single cells in the organoids for their gene expression, we confirmed that they followed independent trajectories without interfering with each other. This level of speed and robustness offers tremendous advantages for future biomedical approaches. By programming cell identity and distribution within brain organoids, we can now control organoid composition in ways that weren’t possible before” said co-first author Aric Lu, a graduate student working with Lewis.

ImageJ=1.52p
unit=micron
3D bioprinted tissue architectures generated from normal, iEndo and iNeuron inks. The specific culture media used by the team induced neuronal differentiation of normal iPSCs after four days (left). After six days, a vascular network was formed by iEndo cells (middle), and a neuronal network was formed by iNeuron cells in the printed patterns. Credit: Wyss Institute at Harvard University

Finally, the team developed biological inks based on their OID system for 3D bioprinting. Since normal iPSCs and their counterparts engineered to express drug-inducible TFs but not having been induced yet with the drug, expand significantly in cell culture, sufficient cell material for larger-scale bioprinting applications can be easily produced. Specifically, the team generated bioinks containing concentrated normal iPSCs, iEndo, and iNeuron cells that they filled into three independent delivery channels of a print head. After laying down a pattern of parallel stripes a continuous filament formed in which the researchers then induced the expression of the TFs. After six days, the iPSCs took on a neuronal stem cell identity. iEndo cells went on to differentiate into endothelial cells that formed a microvascular network and iNeuron cells differentiated into neurons, demonstrating that these cell lines can be used together to create patterned, multicellular structures that contain aspects of the developing brain.

This is an important step towards modeling specific regions of the brain as more TFs are being discovered that can help us further specify neuronal sub-identity within those layers. However, […] our approach offers potential for engineering many more human tissues and organs, just by taking pages from the developmental playbook and harnessing genome engineering and other tools that we are developing.

George Church

“This is an important step towards modeling specific regions of the brain as more TFs are being discovered that can help us further specify neuronal sub-identity within those layers,” said Church. “However, while we started our OID project focusing on neuronal cells because they are relatively easy to access with TFs, our approach offers potential for engineering many more human tissues and organs just by taking pages from the developmental playbook and harnessing genome engineering and other tools that we are developing.” Church is a Lead of the Wyss Institute’s Synthetic Biology platform, and also Professor of Genetics at HMS, as well as Professor of Health Sciences Technology at Harvard and MIT.

Additional authors on the study are Tomoya Duenki, a former graduate student in Lewis’ group who was instrumental for analyzing the OID system the multi-core shell and bioprinting analysis, as well as Songlei Liu, Lucy Nam, and Sarita Damaraju. The study was advanced through grants from the National Institutes of Health (NIH)’s National Human Genome Research Institute (#RM1HG008525), NIH Brain Initiative (#RO1MH123977-01), Vannevar Bush Faculty Fellowship Program sponsored by the Office of Naval Research (#N00014-16-1-2823 and N00014-21-1-2958).

Close menu