Science - USA (2022-02-11)

under the control of p21 promoter sequences that
respond to p53 (fig. S6J) ( 30 ). We found that
scratching induced p53 activity along the
scratched epithelial edge (Fig. 5, D, dotted
whitestrip,andE),relativetocellsawayfrom
the scratch (Fig. 5, D, cells within the yellow
dashed line, and E). p53 elevation was not
accompanied by detectable DNA damage (fig.
S6K), suggesting that p53 was not activated
by mechanically induced DNA damage. Instead,
p53 activation in edge cells was prevented by
inhibition of p38 (Fig. 5, F and G), a stress-
related kinase that can activate p53 upon
mechanical stress ( 24 , 31 ). Thus, scratch-
induced mechanical insult, through p38,
elevates p53, which in turn instructs leader
specification to drive epithelial gap closure.
We reasoned that if the p53-p21 pathway
plays an active role during repair of injured

MDCK monolayers, then inhibiting p53 should

slow down closure of the gap, whereas activat-

ing p53 should accelerate epithelial repair.

Inhibiting p53 function by GSE-22 overexpres-

sion was sufficient to reduce the migration

speed of scratched monolayers (Fig. 5, H to J).

Ablatingp21also resulted in slower migra-

tion of scratched epithelial sheets relative to

wild-type control (Fig. 5K and fig. S6, L and M).

To test whether activating p53 can accelerate

migration, we used laser-induced DNA damage

( 32 ) to elevate p53 specifically in the first row of

cellsattheedgeofthemigratingfront(Fig.5L

and fig. S6N). DNA damage in edge cells was

sufficient to accelerate cell migration, an effect

that could be suppressed by p53 inhibition (Fig. 5,

M and N, and movie S8). Thus, p53 and its target

genep21are the signals activated at the edge of

damaged epithelia that instruct leader cell fate.

Leader cells are eliminated by cell competition

upon epithelial repair

p53 elevation in MDCK cells causes mechanical

cell competition ( 24 ): Cells with high p53 ac-

quire hypersensitivity to cell crowding and be-

have as“mechanical losers,”undergoing cell

extrusion and/or apoptosis when compacted

by cells with low p53 ( 24 ). Because our data

indicate that high p53 is a hallmark of scratch-

induced leaders, we wondered whether these

leaders might be eliminated upon closure of

the epithelial gap, when compacted by follower

cells. Both spontaneous leaders (Fig. 6, A and

B) and scratch-induced leaders (Fig. 6, C and D,

and movie S9) underwent extrusion or death at

high rates upon compaction by followers (75.9

and 40%, respectively).This cell elimination

was not mediated by p21; p21 ablation did not

suppress compaction hypersensitivity in cells

Kozyrskaet al.,Science 375 , eabl8876 (2022) 11 February 2022 4 of 10

A B

0

50

100

150

Nuclear p21

s

ignal

Non-

leaders

n=734

Spont.

leaders

n=11

p<0.0001

C

WT

:

p21KO

MMC

0hrs 12hrs

D

% acting as leaders

p<0.0001

0

20

40

60

80

100

WTMMC

n=251

p21KOMMC

n=187

F G

WTGFP/ WTGFP: WTMMC

WT/ WT:p21KOGFPMMC

-200 200

-400

-200

200

x (μm)

y (μ m)

E

Speed (WTGFP) Time (h)

Distance between WTGFP and WTMMC

Speed (WTMMC) n= 8

-12 -8 -4 0 4 8 12 16

0

5

10

15

0

10

20

30

40 Dis

tanc

e

(μ

m)

(

d

ee

p

S

μ

) h /m

H

Persistence

0.0

0.5

1.0

1.5

WTMMCp21KOMMC

After Contact

p = 0.0101

I

Speed (WTGFP)

Speed (p21KOMMC)

Distance between WTGFP and p21KOMMC

n= 7

Time (h)

Distance (μm)

μ(

d

ee

p

S

m h/

)

-12 -8 -4 0 4 8 12 16

0

5

10

15

0

10

20

30

Displacement (μm)

After Contact

0

100

200

300

400

p = 0.0006

WTMMCp21KOMMC

p21

WT

0hrs 22hrs

DAPI

Fig. 3. The p53 target p21 drives leader cell migration.(A) Movie stills of a
migrating spontaneous leader (arrow), subsequently immunostained for p21
(dashed yellow lines indicate the corresponding field) and (B) corresponding
fluorescence intensity quantification. (C) Movie stills of GFP-positive wild-type
cells cocultured with MMC-treatedp21KOcells. (D) Percentage of MMC-treated
wild-type orp21KOcells acting as leaders. Mean values shown. (Eand
F) Tracking of mean migration speed of and mean respective distance between
pairs of untreated wild-type cells and MMC-treated wild-type cells (E) or
p21KOcells (F). (G) Tracks of wild-type cell migration upon contact with

MMC-treated wild-type (blue) orp21KOcells (gray). (HandI) Persistence of

migration (H) and displacement (I) of wild-type colonies upon contact with

MMC-treated wild-type orp21KOcells. Black bars indicate median in (B), (H),

and (I). Error bars indicate ±SEM in (D) to (F). Thenvalues indicate the

number of cells (B) or the number of contacts [(D) to (F)]. Data from one

representative repeat of three biological replicates (B), pooled from three

biological replicates (D), or from selected movies of three biological replicates

[(E) to (I)].Pvalues from Mann-WhitneyUtest [(B), (H), and (I)] or from

logistic regression (D).

RESEARCH | RESEARCH ARTICLE