In glass: why test-tube results so often fail in living bodies
In vitro is Latin for in glass, a nod to the test tubes, flasks and Petri dishes where biologists pull cells and molecules out of the body to study them. The approach cracked open the immune system and powers modern drug screening, yet most drug candidates that work in a dish fail in living organisms.
The opposite is in vivo, work inside a whole living thing, from animals to human clinical trials. Where one ends and the other begins depends on the field. Toxicologists call anything short of a whole animal in vitro, including organ and tissue cultures. Virologists, whose subjects only multiply inside living cells, call cell-culture work in vitro to set it apart from animal studies. Molecular biologists treat a whole cell as the living unit, so for them only cell-free systems count.
The range of what fits in a dish is wide. Wheat germ extract contains working ribosomes that can translate genetic messages outside any cell. Mitochondria and chloroplasts can be isolated and still function. The polymerase chain reaction copies chosen stretches of DNA using nothing but purified enzymes, and in vitro fertilisation joins sperm and egg in a culture dish before an embryo is placed in the womb.
The appeal is simplicity. A living organism holds at least tens of thousands of genes, proteins and other molecules interacting at once, so isolating a few makes their behaviour visible. Much of what is known about how antibodies recognise foreign invaders came from exactly this kind of work. Human cells can be studied directly rather than inferred from animals, and automated miniature tests let agencies such as the US Environmental Protection Agency screen huge numbers of chemicals while reducing animal testing.
The catch is translation back to the whole body. A candidate antiviral drug, say against HIV-1, might block replication in cell culture, then prove useless in animals and people because it cannot be delivered where it is needed. Researchers try to bridge the gap with more lifelike systems, such as human-on-a-chip designs, and with mathematical models fed by laboratory data.
Source: In vitro
