Chips with everything

Stem cells that have the ability to differentiate into specific tissues and organs are available, but for this to occur, they need the right kind of environment. Depending on the tissue type, they may require chemical gradients, mechanical movement, pressure, or liquid flow as triggers to guide their development in the right direction.

Modern microfabrication techniques derived from computer manufacturing can offer all these environmental factors on a microscopic level, helping researchers to trick cells into behaving as if they really are taking part in the natural development of a living organ. For example, the ‘soft lithography’ approach, pioneered by Harvard’s George Whitesides in the 1990s, uses polydimethylsiloxane (PDMS) to create cell-scale scaffolds complete with channels for flowing liquids, permeable membranes, and pumps. Such devices can be equipped with all kinds of sensors to follow the development. Moreover, the transparency of the material enables researchers to do microscopic measurements from the outside.¹

Microfluidic chip-sized working models have already been produced for a number of human organs and tissues, including heart, lung, liver, kidney, intestine, muscle, bone marrow, fat, skin, blood vessels and nerves. A key feature to reproduce is the interface between the organ and the bloodstream, represented by separate layers of vascular endothelium and organ tissue.

When seeding their devices with stem cells, researchers have often found that applying the characteristic external physical and chemical parameters helps the cells to live up to their destiny. For instance, Sangeeta Bhatia and colleagues at the Wyss Institute seeded primary hepatocytes – liver cell precursors – and found that, after 24 hours of exposure to dynamic flow in the microfluidic chamber, the cells developed into two separate zones as one would expect them to in a developing liver.² Similarly, periodic stretching can aid the development of a heart-on-a-chip, and breathing motions will enable cells to form a small-scale model of a functional human lung.

A challenge overcome only recently was the recreation of bone marrow in a realistic chip model. Bone marrow is of particular medical interest as it generates the precursors for all types of blood cells. Studies of this process have so far been limited to animal models.

To recreate a working bone marrow in vitro, Ingber’s group used a tissue engineering approach. The researchers pulverised bone matrix and filled it into a pre-shaped PDMS compartment, which they implanted into mice. In a first attempt, the compartment was invaded by fat cells, forcing the researchers to add an extra barrier to the system. With this improvement, they produced what they call engineered bone marrow (eBM) and showed that it reproduces normal biological function when maintained under flow for at least one week ex vivo.³ 

This advance in biomedical engineering also opens up new opportunities for stem cell science. Not only do the researchers need stem cells as seeds to produce their model organs, but by reproducing tissues that harbour stem cells and trigger their development, like the bone marrow does for the blood cells, they give stem cell scientists new opportunities to manipulate and investigate the development of such cells.

Paul Fairchid, co-director of the Oxford Stem Cell Institute, UK, welcomed the recent advances and commented: ‘By recreating the complexity of cellular interactions in a temporal and spatial manner that emulates microenvironments in vivo, researchers may finally begin mining the secrets of stem cell niches that have so far proven impervious to investigation.’

Once researchers have established a working model of a human organ in chip format, they can move on to the next step and disrupt its function, either by modelling disease, or by trying drugs or other interventions. In both cases, they have the opportunity to experiment with human tissues and ask questions they would otherwise only be able to address by proxy, in animal models.

In 2012, Ingber’s team developed a model of human pulmonary oedema, or ‘fluid-on-the-lungs’, by using the approved cancer drug interleukin-2. This drug is known to produce pulmonary oedema, which is its dose-limiting side effect.⁴ ‘When delivered into the vascular channel of the lung chip at the same dose it is injected intravenously in humans, it produced fluid accumulation in the air channel of the chip over the same time course as observed in humans,’ Ingber explains. ‘Moreover, our team discovered that breathing motions are critical for this effect, which we then confirmed in a mouse model; collaborating with the pharmaceutical company Glaxo SmithKline, we used the chip to discover a new drug that prevented pulmonary oedema, which also was shown to work in animal models.’ 

The groups of Kevin ‘Kit’ Parker and William Pu at the Wyss Institute have recently used a complex microfluidic system to model and analyse the symptoms of Barth syndrome, a rare disorder of the heart and skeletal muscles, arising from malfunction of lipid metabolism in mitochondria. With this model, the researchers established that a mutation of the TAZ gene is the cause of the disruption of muscle organisation that weakens the muscle’s ability to contract. Using gene editing on their model, the researchers could repair the gene and thereby cure the disease that afflicted their heart-on-a-chip.⁵

While gene therapy is unlikely to become widely available in the near future, the study revealed that the mechanistic link between the gene and the muscle disorder relies on reactive oxygen species (ROS) released by the mitochondria.

This suggests that reactive oxygen scavengers could be used to treat the disease. Furthermore, the authors concluded that their findings ‘may be relevant to more common conditions such as ischemia and ageing, as mature cardiolipin depletion and elevated ROS production also occur in these states’.

Separate studies of disease models in vitro have also been done using ‘organoids’. Not to be confused with organ-on-a-chip technology, these are cell assemblies without the microfabricated support structure. For prostate cancer, for example, Yu Chen and colleagues at the Memorial Sloan Kettering Cancer Center, New York, US, have grown organoids from the cells of a patient’s biopsies.⁶ If doctors could apply this technology when biopsies are made for diagnosis, they would have a quick way of establishing which therapy will work for a given patient, an important step towards personalised cancer treatments.

Personalisation is also a key issue in asthma, another disease targeted by the Wyss Institute. In a recent publication, Parker and colleagues described a chip model for the human airway system. Applying interleukin-13, a transmitter of allergic responses, and acetylcholine, they reproduced in vitro the hypercontraction of smooth muscle, which is the physiological response that characterises asthma attacks.⁷ ‘Our chip offers a simple, reliable and direct way to measure human responses to an asthma trigger,’ said team member Alexander Peyton Nesmith.

Having modelled the disease on a chip, the researchers then went on to try out treatment options. They showed that β-agonists, the drugs used in asthma inhalers, stopped the model asthma attack and allowed the muscle cells to relax. Next, they studied the effects of the drug candidate HA1077, which is thought to work through an association with the RhoA protein. Detailed mechanisms of this protein’s role, and the way the drug candidate interacts with it, remain to be elucidated, but the asthma chip will provide an excellent platform to do this research in a realistic environment.

In asthma, cancer and in other diseases where the existing drugs don’t always work, the progress in cell science and developments like organ-on-a-chip technology could be combined to deliver highly sophisticated therapies tailored to individual patients. ‘By combining chip technology with the routine production of induced pluripotent stem cells from individual patients, it may be possible to investigate not only the pathogenesis of their disease, but to predict, in advance, their response to drugs and any likely adverse reactions,’ explains Fairchild.

Beyond the fundamental research into organ function and disease mechanism, and the promising but cost-intensive prospect of personalised medicine, organ-on-a-chip technology may soon become widely used in routine testing in a range of areas, from drug development to toxicity screening.

Once chips for specific organs are standardised to an extent that they can be mass-produced and commercialised, regulatory authorities will have to confront the question of whether, and under what conditions, they accept chip-generated data as a full or partial replacement of the animal experiments that are currently mandatory. As animal models only have a limited predictive ability, estimated to be around 50–70%, the challenge for chip researchers will be to demonstrate that their predictions are at least as good, if not better.

Ingber’s team at the Wyss Institute has just launched the spin-out company Emulate (http://emulatebio.com/), which will take this approach forward and introduce it to the global markets. Senior staff scientist Geraldine Hamilton from the Wyss Institute will be president and chief scientific officer at the new company, while the institute’s ‘entrepreneur in residence’, James Coon, will be its ceo.

Drug tests on individual organs on chips can already give useful information, but to replace animal models, the approach will have to be broadened to include all organs that may be affected by a given substance or treatment. The more ambitious term of a ‘human body-on-a- chip’ is already becoming a buzzword as researchers begin to establish automated platforms connecting several of the model organs. In 2012, Ingber and Parker received a large grant from DARPA to accomplish precisely this goal – so watch this tiny microfabricated space.

1 S. N. Bhatia and D. E. Ingber, Nature Biotechnol., 2014, 32, 760.
2 J. W. Allen and S. N. Bhatia, Biotchnol. Bioeng., 2003, 82, 253.
3 Y. Torisawa et al, Nature Methods, 2014, 11, 663.
4 D. Huh et al, Sci. Transl. Med., 2012, 4, 159ra147.
5 G. Wang et al, Nature Medicine, 2014, 20, 616.
6 D. Gao et al, Cell, 2014, 159, 176.
7 A. P. Nesmith et al, Lab Chip, 2014, 14, 3925.