The Economist USA - 22.02.2020

78 Science & technology The EconomistFebruary 22nd 2020

B

odies are made of cells. Lots of them.

An average adult human contains about

37.2trn cells, 100 times as many as the num-

ber of stars in the Milky Way. Clearly, trying

to map the location of every one of these

cells would be a futile endeavour. But cells

are not identical. They are divided into

many types, each specialised for different

tasks. Mapping the location of each of

these types is a more tractable problem.

And that is the objective of the Human Cell

Atlas project, a collaboration of researchers

from 1,029 institutes in 71 countries around

the world. Work on the atlas began in 2016,

and its organisers hope to complete the ef-

fort by the end of this decade. Several of

those involved gave a progress report to the

aaasmeeting in Seattle.

As Aviv Regev of the Massachusetts In-

stitute of Technology explained, compiling

the atlas has been made possible by a tech-

nique called single-cell transcriptome se-

quencing. This looks, a cell at a time, at the

messenger molecules which carry instruc-

tions from a cell’s nucleus to the protein-

making machinery in its cytoplasm. These

messengers are made by transcribing

genes into a dna-like chemical called rna.

All an organism’s cells have more or less

the same dna. What makes them different

from each other is which bits of that dna

are being actively transcribed, and thus

what proteins are being made. This means

identifying and counting the rnamessen-

gers shows what sort of cell it is.

One early discovery the project’s re-

searchers have made is that there are many

more types of cell than the text books sug-

gest. Human cells have about 20,000 pro-

tein-coding genes and so possess 20,000

different possible messengers. One way to

classify cells, Dr Regev explained, is to de-

vise a graph that has 20,000 axes, each cor-

responding to a gene’s activity, and then

plot particular cells as points on this graph.

Clearly, such a graph would in a physical

form be impossible, but it can exist and be

manipulated in a computer. Plotting cells

out in this way reveals how they cluster to-

gether. Those clusters are cell types.

Based on cells’ appearance under a mi-

croscope, and their reactions with chemi-

cal stains used to make them visible to mi-

croscopy, early histologists came up with

about 300 cell types in a body. Single-cell

rnasequencing is multiplying that num-

ber by showing that cells which look simi-

lar under a microscope often turn out to be

chemically different from one another. It

has also found previously unknown cell

types so rare that microscope studies have

missed them.

One example Dr Regev proffered con-

cerns the lining of the airways of the lungs.

This tissue, pulmonary epithelium, is

more easily sampled pre mortem than

many others, so was an early candidate for

investigation. The text books suggest it has

six types of cell in it. rnaanalysis showed

that three of these six can themselves be di-

vided into three, and that there are also two

minor cell types previous investigators had

overlooked. This turned out to be an im-

portant discovery, because one of the mi-

nor types proved to be where the gene in-

volved in cystic fibrosis, a fatal hereditary

disease of the lungs, is active.

Having identified individual cell types

in this way, it is then possible to locate

them within tissue samples. Fluorescent

chemicals attached to molecular tags that

will stick only to particular rnamessen-

gers show up the cells containing those

messengers. This permits construction of

three-dimensional maps into which an ob-

server may zoom to reveal ever finer levels

of detail, in the same way that the zoom

control works on an internet map of Earth.

That makes understanding the micro-

scopic details of anatomy much easier.

Shannon Hughes, of America’s National

Cancer Institute, illustrated this with the

example of skin cancer. Tumours are par-

ticularly good targets for transcriptome se-

quencing because they are caused by ge-

netic mutations that show up in the rna

messengers. This has led to a parallel effort

to the Human Cell Atlas, the Human Tu-

mour Atlas, a network devoted specifically

to studying cancer. And, like the lining of

the lungs, the skin is easily sampled. Dr

Hughes and her colleagues have been able

to detect pre-cancerous skin cells (those

without the full complement of mutations

needed to make them cancerous), and ob-

serve how these are already attracting the

attention of immune-system cells called t-

lymphocytes, which burrow through the

skin tissue to attack their targets.

Kerstin Meyer of the Wellcome Sanger

Institute in Britain, meanwhile, trailed the

publication of a complete map of an organ

called the thymus. This has subsequently

appeared on February 20th as a paper in

Science, the aaas’s house journal,

The thymus, located just above the

lungs, is where t-lymphocytes develop. By

looking at 25 thymuses from embryos, fe-

tuses, children and adults, Dr Meyer and

her colleagues have constructed a map that

stretches through time, as well as space.

Their transcriptome sequencing distin-

guishes more than 40 cell types, and they

can follow the ebb and flow of these at dif-

ferent stages of life.

The thymus is, admittedly, but a small

continent in the world that is a human

body. Mapping it is, however, an important

advance for the cellular cartographers. Big-

ger organs will follow soon. The result will

be the most granular view yet obtained of

human anatomy. 7

SEATTLE

Building the most detailed map yet of human anatomy

Cell biology

An atlas of the innerverse

Lung cells put on a colourful display