Science - USA (2022-01-28)

INSIGHTS | PERSPECTIVES

GRAPHIC: KELLIE HOLOSKI/

SCIENCE

science.org SCIENCE

By Rebecca A. Ihrie^1 and Elizabeth P. Henske^2

T

uberous sclerosis complex (TSC) is a

devastating disease characterized by

distinctive tumors of the skin (an-

giofibromas), brain [cortical tubers,

subependymal nodules (SENs), and

subependymal giant cell astrocy-

tomas (SEGAs)], heart (rhabdomyomas),

lungs [lymphangioleiomyomatosis (LAM)],

and kidney (angiomyolipomas and renal

cell carcinoma) ( 1 ). Neurologic manifesta-

tions can also include seizures and TSC-as-

sociated neuropsychiatric disorder, which

encompasses aggressive behaviors, autism

spectrum disorders, intellectual disability,

and psychiatric disorders. TSC is caused

by heterozygous germline mutations that

inactivate TSC1 or TSC2, which normally

inhibit mechanistic target of rapamycin

complex 1 (mTORC1). mTORC1 inhibitors

are used for the treatment of brain, lung,

and kidney tumors and seizures but do not

improve all symptoms, and uncertainty

persists about how brain tumorigenesis

begins. On page 401 of this issue, Eichmül-

ler et al. ( 2 ) identify a precursor cell that

generates cortical tuber–like and subep-

endymal tumor–like cells, increasing the

understanding of neurological manifesta-

tions of TSC.

The unmet clinical need for patients

with TSC is enormous, especially for the

complex spectrum of neurologic manifesta-

tions. Although many neurological features

can present soon after birth, whether these

features are due to altered prenatal cellular

development versus additional effects on

postnatal processes such as synaptic prun-

ing, white matter maturation, and brain cir-

cuit connectivity remains unclear. A barrier

to progress in this area has been challenges

in the development of models that fully re-

capitulate the brain malformations and neu-

ropsychiatric manifestations of TSC. Many

animal models diverge from what is seen in

patients—for example, requiring loss of addi-

tional tumor suppressor genes for tumor de-

velopment or exhibiting seizures but without

evident cortical tubers ( 3 – 7 ).

Uncertainty persists about the mecha-

nisms underlying brain hamartomas, in-

cluding SEN, SEGA, and tubers. Second-hit

genetic events, which cause loss of function

in the remaining wild-type copy of TSC1 or

TSC2, occur in virtually all TSC-associated

renal and lung tumors, which tend to pres-

ent later in life (see the figure). The second

hits lead to unrestrained mTORC1 activity

and are believed to be required for tumori-

genesis. Second hits in TSC1 or TSC2 have

been detected in many SENs and SEGAs

but not in most cortical tubers; whether

these events are required for pathogenesis

is unknown ( 8 ).

The cell of origin for several manifesta-

tions of TSC (including angiomyolipomas,

LAM, and brain hamartomas) is not well de-

fined. An additional knowledge gap relates to

how TSC is affected by factors such as sex,

immune infiltrate, modifying genes, diet, and

environmental exposures. Phenotypes and

severity are not strictly genotype dependent;

individuals with TSC in the same family can

have widely different clinical presentation.

These knowledge gaps are interrelated with

respect to the optimization of disease models

for preclinical studies.

Eichmüller et al. generate organoids

(cell cultures that form three-dimensional

masses) using induced pluripotent stem cells

(iPSCs) that are derived from TSC patients

and then differentiated into neural lineages.

These organoids recapitulate two manifes-

tations of TSC: proliferative lesions that re-

semble SENs or SEGAs and giant cells similar

to those in cortical tubers. Using single-cell

RNA-sequencing, the authors identify a pu-

tative shared cell of origin for both types of

lesion, called a c audal late interneuron pro-

genitor (CLIP) cell. CLIP cells express low

amounts of TSC2 and thus are hypothesized

to be more sensitive to the loss of one TSC2

allele. CLIP cells are transcriptionally separa-

ble from nearby progenitor populations that

have been shown to have increased mTORC1

activity and susceptibility to tumor formation

( 7 ). The authors argue that brain lesions ini-

tially consist of CLIP cells, but other cell types

appear during disease progression.

The discovery of CLIP cells, and the hy-

pothesis that they are the precursor of both

cortical tubers and subependymal tumors, is

potentially a breakthrough in understanding

the neurologic manifestations of TSC. These

findings contrast with other recent work in

human embryonic stem cells cultured into

cortical spheroids ( 9 ), which showed that

loss of both alleles of TSC1 or TSC2 was re-

quired for altered differentiation and the

emergence of dysmorphic cells, supporting

a two-hit model of tuber development. This

suggests that more than one mechanism can

lead to tubers, perhaps depending on the

developmental stage. Human iPSC-derived

kidney organoids have been found to develop

angiomyolipoma-like tumors, but only in the

setting of biallelic inactivation of TSC2, con-

firming that two hits are required ( 10 ). As or-

CANCER

Modeling tuberous sclerosis with organoids

S ingle-cell profiling reveals a different path for the development of brain lesions

Angio-

myolipoma

Lymphangio-

leiomyomatosis

X XX

Cortical tuber

Subependymal

nodule

Subependymal

giant cell

astrocytoma

CLIP

cells

CLIP-like

cells?

TSC1/2

(^1) Cell & Developmental Biology and Neurological Surgery,
Vanderbilt University School of Medicine, Nashville, TN,
USA.^2 Pulmonary and Critical Care Medicine, Brigham and
Women’s Hospital, Harvard Medical School, Boston, MA,
USA. Email: [email protected]
Tumorigenesis in tuberous sclerosis complex
Heterozygous germline inactivating mutations in tuberous sclerosis complex 1 (TSC1) and TSC2 result in
distinct tumors developing throughout life. Brain lesions present early and likely arise from caudal late
interneuron progenitor (CLIP) cells, whereas kidney and pulmonary tumors present later in life when a second-
hit inactivating mutation occurs in the wild-type TSC1/2 allele, possibly in CLIP-like cells.
382 28 JANUARY 2022 • VOL 375 ISSUE 6579