In March this year, American electrical engineer and scientist Professor Rashid Bashir published a blow-by-blow account of how to design and make a muscle-powered biological machine. Dubbed ‘bio-bot’, the walking, 6mm-long device can be powered by muscle cells, controlled with electrical and optical pulses and could one day roam your body to deliver drugs, detect disease or remove pieces of tissue.

Crucially for would-be bio-bot builders, the process, published in Nature Protocols, describes every fabrication step from 3D-printing a skeleton to tissue-engineering a muscle actuator, including the part numbers for every component Bashir uses in his laboratory.

“We wanted to provide detailed recipes and protocols so that others can duplicate the work and further permeate the idea of ‘building with biology’,” says Bashir. “[With this], other researchers can have the tools and knowledge to build bio-hybrid systems, and attempt to address challenges in health, medicine and the environment that we face as a society.”

It all started several years ago, when Bashir, head of bioengineering and director of the Micro and Nanotechnology Laboratory at the University of Illinois at Urbana-Champaign, and fellow researchers, wanted to answer one simple question: ‘Can we design with cells?’

As Bashir points out: “We realised that if we knew the design rules and knew how to get the cells to communicate, we could use these cells as building blocks for machines and systems.”

Come 2010, the US National Science Federation had thrust some $33m at a new project – ‘Emergent Behaviors of Integrated Cellular Systems’ – to stimulate the development of living, multi-cellular machines (see ‘Engineering living cells’).

In 2012, Bashir revealed a series of 7mm-long caterpillar-like bio-bots that could ‘walk’ at about 235 micrometres a second within a nutrient-rich fluid.

Fabricated layer by layer using a 3D stereo-lithographic printer, the bio-bots comprised a jelly-like hydrogel strip lined with rat cardiac cells perched on a rigid hydrogel base to form a cantilever. Crucially, when cardiac cells twitched in unison, the cantilever would flex, acting as a lever to propel the bio-bot slowly forward.

“These cardiac cell sheets would self-beat, or self-actuate, and autonomously move the bio-bot in one direction,” explains Bashir. “It could produce force and could do some work but the cardiac cell would spontaneously beat and couldn’t be stopped easily.”

In a bid to make the bio-bot more controllable, Bashir switched from rat cardiac cells to skeletal muscle cells from mice, and developed a new class of bio-bot powered by a skeletal muscle strip and controlled with electrical pulses.

As before, the team used 3D printing to fabricate a hydrogel skeleton structure, this time comprising two ‘legs’ connected by a flexible beam. Then they grew the all-important skeletal muscle cells from immature mouse myoblasts and embedded these within a matrix of fibrin and collagen. The matrix was added to the skeleton legs, with cells compacting and forming a muscle strip from one leg to the other within 14 hours. Electrical stimulation triggered the cells within the muscle strip to contract, pulling the legs toward each other to ‘walk’ at a top speed of 156µm/s.

For Bashir, the switch from heart cells to skeletal muscle cells made perfect sense; skeletal muscles are the primary generator of actuation – or movement/operation – in animals, and he could activate this action by applying electrical pulses. In fact he could even control the bio-bot’s speed by altering the electrical pulse frequency; but he had yet another trick under his bio-bot hat.

In March 2016, Bashir and colleagues unveiled a new muscle-powered two-legged, walking biological machine. In order to gain design flexibility and produce a more agile bio-bot, the researchers had made two crucial changes.

First, they had built the bio-bot from elastic-band-shaped muscle rings, rather than muscle strips. As Bashir explains: “You can take these rings, pick them up and move them from one [skeleton] to another.”

“With the rings, we can connect any two joints or hinges on the 3D-printed skeleton, and we can have multiple legs and multiple rings,” he says. “It is a much more flexible, scalable design.”

In addition to design perks, the rings allowed nutrients from the surrounding cell culture to diffuse into the tissue from all sides. By ‘exercising’ the muscles daily – triggering actuation to increase contractile forces – the researchers could strengthen the rings so that the bio-bot could move further with each contraction. Indeed, this time, the bio-bot’s average speed came in at 310µm/s.

In a second, critical, development, the researchers had also ditched the electrodes and turned to optogenetics, in which cells are genetically endowed with light-responsive molecules.

“You shine blue light at muscle cells, which activates the channels within to polarise and de-polarise the cells, and the cells then actuate so the muscle contracts,” says Bashir. “If you shine the light on different parts of the bio-bot, you can actuate locally, and tilt and move it in different directions.”

“You just have so much more control with light,” he adds.

Bashir had obtained the bio-bot’s light-responsive cells from fellow researcher Roger Kamm, professor of biological and mechanical engineering at Massachusetts Institute of Technology. Here, Kamm and colleagues were working on neuromuscular junctions – the vital connection where nerve meets muscle – and had developed a microfluidic device that contained a muscle strip and small set of motor neurons.

Kamm and his students modified the neurons to respond to light, so the cells would then send signals to excite muscle fibre. He explains: “We made the neurons responsive to light and used optogenetics to power the neuromuscular junction”.

However, while Bashir has successfully incorporated such light-activated muscles cells in the latest bio-bot, Kamm’s neuromuscular junction breakthrough also looks set to have a profound impact on future biological machines. As he explains: “No one had ever generated a neuromuscular junction before, and this is a direction we’re now going towards with bio-bots, so we can have neural activation of the muscle.”

Indeed, as Bashir points out, the latest muscular ring design will ease integration of different cell types, such as muscle cells and neurons, into bio-bots.

“We can make a ring out of muscle and then have neurons in a different ring to develop a neuromuscular junction,” he says. “We have been working on this and want to use many different types of cells such as muscles and neurons, as well as muscles with vasculature, to add more intelligence and functionality to bio-bots.”

Indeed, neurons would enable cell-to-cell communication so a bio-bot could begin to monitor its environment and trigger its movements. At the same time, endothelial cells, which are key in the formation of blood vessels, would promote development of a blood-vessel vascular system in muscles, ensuring bio-bot cells have a steady supply of oxygen and necessary nutrients.

“We really would like to vascularise the muscle: introduce a blood supply to it, so we need these endothelial cells to generate that blood vessel network,” explains Kamm. “By co-culturing muscle cells with endothelial cells and fibroblasts you could, in principle, produce a vascularised muscle, so we’ve been working on ways to produce microvascular networks that could perfuse muscle.”

Yet Bashir, Kamm and colleagues are not alone in their pursuit of more sophisticated bio-bots. From crawling crabs to swimming jellyfish, the last decade has seen myriad biological robots.

In 2007, Professor Metin Sitti and colleagues from Carnegie Mellon University, Pennsylvania, attached hundreds of bacteria onto 10µm polystyrene beads. When placed in a water-glucose solution, the bacteria would rotate their corkscrew-like tails (flagella) to propel the bead at around 15µm/s. Sitti hopes that one day, swarms of bacteria-driven micro-robots could deliver drugs inside the human body.

Around the same time, Sukho Park and colleagues from the Nano-Bio Research Centre, Korea Institute of Science and Technology, powered a six-legged crab-like skeleton with beating rat heart muscle, envisaging that the micro-biobot could clear blocked arteries. A few years later, Professor Keisuke Morishima from Osaka University, Japan, unleashed a polypod microbot driven by the dorsal vessel tissue from an inchworm.

Still on the micro-scale, in 2014, Professor Taher Saif from the Micro and Nanotechnology Laboratory at Illinois developed a flagellar swimming bio-bot by combining live heart cells with a flexible polymer body.

As Saif said at the time: “Our long-term vision is simple. Can we make structures and seed them with stem cells that would differentiate into smart structures to deliver drugs, perform minimally invasive surgery or target cancer?”

Back to the larger scale, bioengineer Professor Kit Parker of Harvard University’s Wyss Institute developed a 6mm artificial jellyfish, dubbed Medusoid, in 2012. Consisting of a jellyfish-shaped silicone sheet coated with rat heart-muscle cells, the bio-bot would happily ‘swim’ through nutrient-rich water for an hour once a jolt of alternating current had been applied to the solution to make the cells contract.

Parker recently introduced a more complex bio-bot; a 16mm-​long stingray, that swims through salt solution, but is powered by some 200,000 light-activated rat heart-muscle cells. Like Bashir’s bio-bots, these live-muscle robots could be used in human bodies to, say, repair tissue, but Parker’s hope is that his muscular pump bio-bots will help him build an artificial heart.

This is where the bio-bot landscape gets even more diverse. In 2013, Kamm and Bashir published a paper in Annals of Biomedical Engineering called ‘Creating living machines’. From bio-bots and organs-on-a-chip to smart plants and multi-functional organs, the researchers detailed the state-of-art in so-called engineered living systems, and also offered a glimpse of the more complex living machines that could come.

As Kamm puts it: “Bio-bots are interesting and fun, but we are also using them as test-beds for different concepts on how to generate engineered living systems. Bio-bots are just one aspect of these living systems, with organs-on-a-chip being another.”

Indeed, while Kamm has been developing his neuromuscular junction on a chip, researchers have also developed microfluidic systems that can replicate certain aspects of organ function, including those of the lung, liver and heart. Kamm notes that the grand challenge is to produce a ‘body-on-a-chip’ to model the response of a drug on several organs.

Looking to the future, Kamm reckons it’s relatively easy to imagine how muscle strips, vascular networks, neuro­muscular junctions as well as bio-bots and organs-on-a-chip could be combined to create higher functioning biological machines.

In a departure from mammal-like living machines, plant cells could be genetically modified or mammalian-derived neuron cells could be introduced into a plant, to create a smarter plant that could process environmental information and grow according to its conditions. As Kamm puts it: “We could create plants with new capabilities, such as corn that converts nitrogen in the atmosphere to fertilise itself.”

Kamm believes researchers could one day be fabricating implantable organ systems that perform functions unheard of today. Blood vessels could pump to relieve load on your heart or cell-based sensors could detect blocked blood vessels and release anti-blood clotting drugs from cell-based factories embedded in your blood vessels.

“We could also produce completely new kinds of organ,” Kamm says. “If you had a patient with a chronic illness that requires delivery of a drug, then you could design an organ to sense drug levels in circulation and respond by synthesising and secreting that drug.”

Clearly, ethics are an issue for such a brave new research area. For example, at what level of complexity does a biological machine become a living organism? How can harmful outcomes of this cutting-​​edge research be avoided?

Kamm and Bashir are adamant that the time for ethics discussions is now, and not once the technologies are developed. “If you have a bio-bot migrating around, say, a stomach cavity, you have to be very careful how you design and implement it, and you must make sure you have control over it,” says Kamm. “We’re constantly thinking about the ethical considerations. But we also need to convince people that this research field is a whole new opportunity and we have only just scratched its surface.”

Engineering living cells

The US NSF project Emergent Behaviors of Integrated Cellular Systems (EBICS) was set up in September 2010 with $33m funding and includes the University of Illinois, MIT, UC Merced and Georgia Tech. The aim is to engineer ‘living machines’ that can sense and process information, in order to perform tasks such as detecting toxins, delivering drugs and sequestering carbon dioxide.

In 2015, funding was renewed to the tune of $25m and the centre is still growing and thriving, headed by Professor Roger Kamm. Work includes growing organoids – miniaturised and simplified versions of organs – from stem cells, as well as developing neural circuits, neuromuscular junction and microvascular networks.

Development of biological machines includes so-called ‘motile bots’ to sense and actuate, while ‘pump bots’ are to transfer, filter or remove fluids.

Crucially, the project also has an open access education plan to draw students to this fledgling field. Ritu Raman, EBICS post- doctoral fellow at Illinois, says: “We want to introduce biological materials into the toolkit of the next generation of makers by giving them hands-on experience with bio-bot design and manufacture.”

Tiny implantable microbots made by 3D printing

The impetus to fabricate muscle- powered biological machines lies in the notion that the bio-bot, with its innate biological response, will be more flexible, agile and functional than a ‘traditional’ robot. However, researchers are also developing biocompatible microbots that could serve similar functions.

Earlier this year, biomedical engineer Professor Sam Sia and colleagues from Columbia University, New York, published details of a manufacturing process for biocompatible hydrogel valves, manifolds, rotors, pumps and drives.

A key device called the Geneva drive is composed of two co-attached gears with an iron oxide nanoparticle-infused interior, rotated by a motor- controlled magnet beneath the gears. Crucially, the drive was loaded with chemotherapy drugs, implanted into live mice with bone cancer and independently released doses on rotation. Tumour growth in the mice was limited with the procedure shown to be less toxic than conventional chemotherapy.

Sia’s colleague, Dr Sau Yin Chin, explains that these micro- machines are manufactured via 3D printing and can each be assembled in around half an hour. “Hydrogels are difficult to work with… but we tuned the mechanical properties and matched the stiffness of structures that come in contact with each other within the device,” she says.

Gears that interlock need to be stiff to allow forces to be transmitted and to withstand repeated actuation. Meanwhile, structures that form locking mechanisms should be soft and flexible so the gears can slip past during actuation.

“We have flexible posts that sit between the spokes of a gear and gently hold the gear in place when there is no actuation,” she says. “Yet during magnetic actuation, the gears move and the soft posts bend so the gears can repeat. In this way we find it quite easy to incorporate locking mechanisms by matching different thicknesses of the gels when making devices.”

Crucially, the latest device brings researchers closer to developing miniaturised, battery- free biocompatible robots that can safely interact with humans and other living systems.

“We’ve been intent on using hydrogels as they are entirely biocompatible. [Our devices] don’t use electronics and don’t require on-board batteries, so we are producing really safe implantable devices,” Yin Chin comments.