In Vitrica

Britain is about to spend £20m on growing miniature human organs from patients' cells, reports The Guardian. The Cambridge programme is intended to reduce the use of animals in drug testing and eventually make it easier to determine which treatments work for particular patients. It will create a library of standardised, validated »organoids« for academics and industry.¹

»We are moving from experimenting on reality towards constructing realities to experiment on«

An organoid is not quite a miniature organ. It’s a three-dimensional collection of cells, usually derived from stem cells, grown in the laboratory to reproduce some of an organ's structure and function. Organoids can model intestines, livers, brains, and tumours; patient-derived ones can preserve characteristics of the person they came from.²

The strange thing isn't that we're replacing mice with miniature human organs, but that we're beginning to construct the objects we want to interrogate.

Mary Shelley's Frankenstein is usually read as a warning about scientific hubris. But science writer Philip Ball has pointed out that its cultural afterlife is more complicated.³ The provocation is not simply that a scientist creates life, but that he constructs something from parts and then has to decide what, exactly, he has made.

We have been asking versions of that question ever since. In 1991, Marc Quinn made Self, a cast of his own head containing ten pints of his frozen blood. It might be the ultimate self-portrait: made from the material of the person it depicts. The sculpture is not Quinn, but also not merely a representation of him.⁴

An organoid is stranger still. It’s not a picture of you, and it’s not you. It’s a biological abstraction: material derived from you, organised to reproduce some of your function.

That matters because science has traditionally experimented on things that stand in for the thing we actually care about. We test a drug on a mouse because the mouse is a model of a human. We test it on a population because the population is a model of the patient.

But what if we could construct a model of the particular person we want to treat? Instead of testing drugs first on mice and then on humans who vaguely resemble us, we could eventually apply thousands of possible treatments to models of a particular person's biology. The system could learn from the results, concentrate on interesting regions of the search space, and compare them with those of biologically related people.

The result would be more like cartography than traditional experimentation. By constructing models in order to interrogate them, we would draw the map as we traversed the landscape. This is already the logic behind computational design. I can generate thousands of possible buildings, subject them to structural, environmental, economic or spatial tests, and use the results to decide which deserve to exist. The building is constructed, in a sense, so that I can interrogate it.

Biology seems to be acquiring the same property. If we can construct useful abstractions of particular people, we could search personalised spaces of nutrition, exercise, fertility, ageing, and disease. A treatment might eventually be the result not of one experiment performed on you, but of a million experiments performed on a model of you.

And those experiments would not need to remain yours. Their results could be shared, allowing each model to become one point in a much larger map. The next experiment could be chosen partly because of what happened in thousands of previous experiments on people whose biological systems resemble yours.

In the news, this is described as a progression towards replacing animal experiments. That’s probably true – and important. But it isn't the most interesting thing happening here. The deeper change is that the experiment itself is becoming an engineered object. We are moving from experimenting on reality towards constructing realities to experiment on.

The future of medicine may be neither in vitro nor in vivo. It may be something else: in vitrica – medicine conducted on manufactured abstractions of specific biological individuals, multiplied, tested, and compared until the system begins to know what might work before we try it.

Once we can construct sufficiently useful versions of things in order to interrogate them, experimentation becomes a form of design. We no longer have to wait for reality to present us with the case we want to study. We can make the case, multiply it, vary it, and see what happens.

Perhaps that’s the strange thing the organoid has inherited from Frankenstein. The monster was a thought experiment about constructing life in order to discover what it was. The organoid is something more practical: a piece of constructed life that lets us discover our own future. Not a miniature human. A machine for asking what a human might become.

References