Science - USA (2022-01-14)

protein with a half-life of less than 30 min.
At the end of G 1 phase, cyclin D1 is phosphoryl-
ated at Thr^286 by GSK3b( 15 ). This facilitates
association of cyclin D1 with the nuclear exportin
CRM1 and promotes export of cyclin D1 from the
nucleus to the cytoplasm ( 16 ). Subsequently,
phosphorylated cyclin D1 becomes polyubi-
quitinated by E3 ubiquitin ligases, thereby
targeting it for proteasomal degradation.
Several substrate receptors of E3 ubiquitin
ligases have been implicated in recognizing
phosphorylated cyclin D1, including F-box
proteins FBXO4 (along withaB crystallin),
FBXO31, FBXW8,b-TrCP1/2, and SKP2 ( 17 ).
The anaphase-promoting complex/cyclosome
(APC/C) was also proposed to target cyclin D1
while F-box proteins FBXL2 and FBXL8 target
cyclins D2 and D3 ( 17 , 18 ). Surprisingly, the
level and stability of cyclin D1 was unaffected
by depletion of several of these proteins, indi-
cating that some other E3 plays a rate-limiting
role in cyclin D1 degradation ( 19 ). Indeed,
recent studies reported that D-cyclins are
ubiquitinated and targeted for proteasomal
degradation by the E3 ubiquitin ligase CRL4,
which uses AMBRA1 protein as its substrate
receptor ( 20 – 22 ).

Cyclin DÐCDK4/6 in cancer

Genomic aberrations of the cyclin D1 gene
(CCND1) represent frequent events in different
tumor types. The t(11;14)(q13;q32) translocation
juxtaposingCCND1with the immunoglobulin
heavy-chain (IGH) locus represents the charac-
teristic feature of mantle-cell lymphoma and is
frequently observed in multiple myeloma or
plasma cell leukemia ( 23 , 24 ). Amplification of
CCND1is seen in many other malignancies—
for example, in 13 to 20% of breast cancers
( 23 , 24 ), more than 40% of head and neck
squamous cell carcinomas, and more than
30% of esophageal squamous cell carcinomas
( 23 ). A higher proportion of cancers (e.g., up
to 50% of mammary carcinomas) overexpress
cyclin D1 protein ( 24 ). Also, cyclins D2 and D3,
CDK4, and CDK6 are overexpressed in various
tumor types ( 5 , 9 ). Cyclin D–CDK4/6 can also
be hyperactivated through other mechanisms
such as deletion or inactivation of INK inhibi-
tors, most frequently p16INK4A( 5 , 9 , 23 ). Alto-
gether, a very large number of human tumors
contain lesions that hyperactivate cyclin D–
CDK4/6 ( 5 ).
An oncogenic role for cyclin D–CDK4/6 has
been supported by mouse cancer models. For
example, targeted overexpression of cyclin D1
in mammary glands of transgenic mice led to
the development of mammary carcinomas
( 25 ). Also, overexpression of cyclin D2, D3, or
CDK4, or loss of p16INK4aresulted in tumor
formation ( 9 ).
Conversely, genetic ablation of D-cyclins,
CDK4, or CDK6 decreased tumor sensitivity
( 9 ). For instance,Ccnd1- orCdk4-null mice, or

knock-in mice expressing kinase-inactive

cyclin D1–CDK4/6, were resistant to develop

human epidermal growth factor receptor 2

(HER2)–driven mammary carcinomas ( 26 – 29 ).

An acute, global shutdown of cyclin D1 in mice

bearing HER2-driven tumors arrested tumor

growth and triggered tumor-specific senes-

cence while having no obvious impact on

normal tissues ( 30 ). Likewise, an acute abla-

tion of CDK4 arrested tumor cell prolifera-

tion and triggered tumor cell senescence in

a KRAS-driven non–small-cell lung cancer

(NSCLC) mouse model ( 31 ). These observa-

tions indicated that CDK4 and CDK6 might

represent excellent therapeutic targets in can-

cer treatment.

CDK4/6 functions in cell proliferation

and oncogenesis

The best-documented function of cyclin D–

CDK4/6 in driving cell proliferation is phos-

phorylation of the retinoblastoma protein,

RB1, and RB-like proteins, RBL1 and RBL2

( 5 , 6 ) (Fig. 1). Unphosphorylated RB1 binds

and inactivates or represses E2F transcription

factors. According to the prevailing model,

phosphorylation of RB1 by cyclin D–CDK4/6

partially inactivates RB1, leading to release of

E2Fs and up-regulation of E2F-transcriptional

targets, including cyclin E. Cyclin E forms a

complex with its kinase partner, CDK2, and

completes full RB1 phosphorylation, leading

to activation of the E2F transcriptional program

and facilitating S-phase entry ( 5 , 6 ). In normal,

nontransformed cells, the activity of cyclin D–

CDK4/6 is tightly regulated by the extracellular

mitogenic milieu. This links inactivation of RB1

with mitogenic signals. In cancer cells carry-

ing activating lesions in cyclin D–CDK4/6, the

kinase is constitutively active, thereby decou-

pling cell division from proliferative and in-

hibitory signals ( 5 ).

This model has been questioned by the

demonstration that RB1 exists in a mono-

phosphorylated state throughout G 1 phase

and becomes inactivated in late G 1 by cyclin E–

CDK2, which“hyperphosphorylates”RB1 on

multiple residues ( 32 ). However, recent single-

cell analyses revealed that cyclin D–CDK4/6

activity is required for the hyperphosphoryla-

tion of RB1 throughout G 1 , whereas cyclin E/A–

CDK maintains RB1 hyperphosphorylation in

S phase ( 33 ). Moreover, phosphorylation of RB1

by cyclin D–CDK4/6 was shown to be required

fornormalcellcycleprogression( 34 ).

In addition to this kinase-dependent mech-

anism, up-regulation of D-cyclin expression

and formation of cyclin D–CDK4/6 complexes

lead to redistribution of KIP/CIP inhibitors

from cyclin E–CDK2 complexes (which are

inhibited by these proteins) to cyclin D–CDK4/6

(which use them as assembly factors), thereby

activating the kinase activity of cyclin E–CDK2

( 6 ). Cyclin E–CDK2 in turn phosphorylates RB1

Fasslet al.,Science 375 , eabc1495 (2022) 14 January 2022 2 of 19

G 1

G 2 S

M

CycD

CDK4/6

p19 p57

p27 /

p21

CycE

CDK2

E2F

RB1

RB1

P

P

E2F

p21

p27

CycE

p27 /

p21

p18

p15

p16

X

Fig. 1. Molecular events governing progression through the G 1 phase of the cell cycle.The mammalian

cell cycle can be divided into G 1 , S (DNA synthesis), G 2 , and M (mitosis) phases. During G 1 phase,

cyclin D (CycD)–CDK4/6 kinases together with cyclin E (CycE)–CDK2 phosphorylate the retinoblastoma

protein RB1. This activates the E2F transcriptional program and allows entry of cells into S phase. Members

of the INK family of inhibitors (p16INK4A, p15INK4B, p18INK4C, and p19INK4D) inhibit cyclin D–CDK4/6;

KIP/CIP proteins (p21CIP1, p27KIP1, and p57KIP2) inhibit cyclin E–CDK2. Cyclin D–CDK4/6 complexes use

p27KIP1and p21CIP1as“assembly factors”and sequester them away from cyclin E–CDK2, thereby activating

CDK2. Proteins that are frequently lost or down-regulated in cancers are marked with green arrows,

overexpressed proteins with red arrows.

RESEARCH | REVIEW