Skip to Main Content
Text size: SmallMediumLargeExtra-Large

The Impossible Will Take A Little While

October 2010

Defining A Key Developmental Pathway in the Salivary Gland

In the 1940s, lyricist Bob Russell penned the lines, “The difficult I’ll do right now/The impossible will take a little while.” Although Russell reportedly had jazz legend Billie Holiday in mind, his words today capture well the seven-year quest in NIDCR’s Laboratory of Cell and Developmental Biology (LCDB) to define the molecular “how” and “why” of clefting, an early motif in organ development in which growing epithelial tissues bifurcate.

Think of clefting as the letter “O” morphing into the letter “V.” The clefts generate spatially distinct populations of cells that branch tree-like in divergent directions. This continuous cleft-branch-cleft expansion allows the dividing cells to pattern into multiple sacs, ducts, and other structures needed to build a viable secretory organ such as a lung, kidney, salivary gland, and other bodily organs. The tightly packed, highly iterative engineering of these organs is quite remarkable. Just as five fingers work better than one, the multiple sacs and ducts serve to expand the number of epithelial surfaces in the organ to allow greater secretion, air exchange, urine production, and other metabolic processes.

That takes us to the difficult – but first some needed background on this developmental process, known by the umbrella term of branching morphogenesis. Scientists long have studied branching morphogenesis in the submandibular salivary gland, one of the body’s three major salivary glands. In the mouse, the submandibular gland starts as a spherical sheet of epithelial cells attached to a single epithelial stalk, or bud.

Take a look here. You can see the narrow, crack-like clefts in the bud:
Branched salivary gland




Photo of the branched salivary gland.  The purple indicates epithelial buds; the green is the surrounding connective tissue-like mesenchyme.

In 2003, LCDB post doc Takayoshi Sakai, D.D.S., Ph.D., and colleagues discovered that cells in certain areas of the mouse’s embryonic submandibular gland secrete fibronectin, a much-studied adhesive structural protein. Wedges of fibronectin translocate inward and shallow clefts form in the bud. Their discovery was published in the journal Nature.

But, as Sakai and LCDB lab chief Kenneth Yamada, M.D., Ph.D., suspected, there was more to the story than fibronectin. Like a finger flipping a switch, fibronectin had to activate another protein that, in turn, passed the signal to another protein. What they needed next was to find these downstream proteins and define in fine detail the multi-step signaling pathway and physical mechanisms that culminate in a deep cleft and then branching.

After some brainstorming, Sakai decided to perform the difficult task of sifting through the existing gene expression data to see if any promising candidates might fall out for further testing. “We wanted to know which genes are turned on and off after fibronectin is produced,” said Yamada. “Our strategy was to look at the cleft region, particularly the most active region, which is located at the bottom of the epithelial bud.”

Sakai got his candidate. It’s a gene that encodes Btbd7, or the BTB/POZ domain-containing protein 7. That translates to the seventh protein in a family of regulatory proteins that contain a characteristic protein-binding structural motif and which have a wide variety of functions.

The news got even better from there. Btbd7 was highly expressed during stages of active branching in the developing salivary gland, and the expression occurred around the bottom and lower sides of the forming cleft. Better yet, Btbd7-expressing cells surrounded areas of high fibronectin expression. Then best of all, when depleted of Btbd7, developing glands and lungs could not form proper clefts and branches. Sakai knew he now had his hands on an important molecule.

But why was Btbd7 needed to complete the process? Finding the answer belonged to Sakai’s labmate Tomohiro Onodera, M.D., Ph.D. Sakai had received good news from home that he’d been offered a full professorship in the Osaka University Graduate School of Dentistry. It was an offer he couldn’t refuse. Sakai handed over the project, and Onodera worked intensely to piece together new links in the pathway.

As the months and experiments passed, Onodera filled in the blanks. He showed that fibronectin induced Btbd7, which in turn suppressed the cell-adhesion protein E-cadherin and induced Snail2, a protein well known to developmental biologists for its role in tissue formation. This transition causes the cells to disperse and pattern into a cleft.

So, to recap:

  • The submandibular gland starts as a single bud of epithelial cells attached to a stalk;
  • At embryonic day 13, some still unknown factor induces an accumulation of fibronectin;
  • Wedges of activated fibronectin translocate inward;
  • Btbd7 is induced in the cells surrounding sites of high fibronectin concentration at the base and lower sides of the forming clefts;
  • Btbd7 suppresses E-cadherin and induces Snail2, which causes epithelial cells to scatter.

A manuscript was submitted to the journal Science. The peer review was both extremely positive and perplexing. The reviewers requested additional experiments to show how the fibronectin-Btbd7- Snail2 pathway works mechanistically to form the cleft. Specifically, the reviewers wanted to know whether Onodera et al. could induce clefts by overexpressing Btbd7 in their experimental system and then answer the larger question of how and why the resulting scattering of cells contributed to the progression of the cleft?

As Yamada knew, the first request was technologically impossible to obtain. “It’s an extraordinarily tough request because these types of effects are probably sequential,” he said. “In other words, you induce some initial action. Then, to get the cleft to progress and form the branch, you can’t just express Btbd7 everywhere. You would have to express it, move your point of expression, and move it again. There’s just no way with current technology that you can do that.”

The second request forced them to return to the bewildering dynamics of cell scattering in the developing submandibular gland, a discovery made by former LCDB post doc Melinda Larsen, Ph.D. Larsen found that as the gland develops, cells rustle and can scatter like leaves, although much of the movement was random. Or so it seemed. Making sense of cell scattering in cleft formation was like predicting the path of a tornado. It certainly was possible. But again, it remained impossible without the right tools in hand to visualize the process.


Further complicating matters, the hard-working Onadera was headed back to Japan. He had a full-time position waiting for him, and sorting out the impossible now fell to the technologically savvy LCDB postdoc, Jeff Chi-feng Hsu, Ph.D.
Yamada and Hsu brainstormed that their best chance to stimulate their pathway of interest and – if lucky – stimulate cleft formation would be with viral vectors. Hsu could place copies of the Btdb-7 gene into viral vectors and inject them into the gland; the viral vectors, like Trojan horses, introduce the Btbd-7 gene en masse into the epithelium; boost its levels in the developing tissue; and accelerate cleft formation and branching.

They had two things going for them. One, Hsu had experience working with viral vectors. Two, staff in other NIDCR laboratories, particularly in the Adeno-Associated Virus Biology Section, had been developing viral vectors for several years for use in the salivary glands. If Hsu could tap into the wealth of experience present nearby and cobble together a viable protocol to micro-inject the vector into the gland, they had a fighting chance.

But, as lyricist Bob Russell penned so long ago, the impossible would take a little while. “It took me about a year and a half,” said Hsu. “I would try one thing, then try something else until eventually I got the virus where I needed it, could express it everywhere, and induce the clefting and branching.”

This brought Hsu to his next Mount Everest – cell scattering. Through trial and error, he devised a method to watch the movement of epithelial cells in the developing gland labeled with green fluorescent protein. Using a red dye to identify spaces between cells, they could directly follow how the clefts formed. What he saw next – and for the first time ever - was pure magic:

Here is the time-lapse video:

What did Hsu think as watched the process unfold for the first time?

Yamada elaborated, “The Btbd7-induced cell scattering that we showed elsewhere was reflected in a jostling and opening up of tiny gaps by cells that we think are expressing Btbd7 and Snail2. That presumably reduces cell-cell adhesion and opens gaps. Some of the gaps were transient, but others allowed the cleft to spread downwards.”

“To summarize, we originally thought that the steady inward movement of fibronectin cleaved the epithelial bud,” he said. “It turns out that fibronectin is following the key activities at the very base of the cleft, where transient gaps form as cells separate. But only some of these gaps seem to contribute to the cleft as it penetrates deeper into the bud. This discovery provided a functional link between the observation of cell scattering and cleft progression. It also explains the need for highly plastic movements of the cells within glands, which we think is augmented by scattering at the base of the cleft.”

A revised manuscript was submitted, and the paper ultimately was published in the July 30, 2010 issue of Science. It featured three first co-authors - Onodera, Sakai, and Hsu – a reflection of the team approach that produced the paper.

So, where does the Yamada lab go from here? Think in the classical sense of a story with a beginning, middle, and end.

“If you define the end of the story as forming the complex branching structure, we think that we have it,” said Yamada. “We also have the middle of the story with the fibronectin wedging into the bud. What we still lack is the beginning of the story. We need to know what starts the process.”

Intriguingly, Yamada and colleagues have learned the mouse submandibular gland has no preprogrammed map of where the clefting and branching will occur. But when the cleft forms, the cells adjust quickly and get with the developmental program. “By identifying the mechanisms involved in branching morphogenesis, we think it will be possible to recapitulate the steps in the laboratory and employ them to regenerate damaged salivary glands or even engineer artificial ones,” said Yamada. “The process can be difficult to study, but we’re getting there.”

Share This Page

GooglePlusExternal link – please review our disclaimer

LinkedInExternal link – please review our disclaimer


This page last updated: July 31, 2014