Science - USA (2022-04-22)

SCIENCE science.org

By Kathleen H. Burns1,2

I

n 1988, physician-scientists studying

the genetic basis of hemophilia dis-

covered pathogenic long interspersed

element 1 (LINE-1) insertions in the co-

agulation factor VIII gene ( 1 ). Therein

was unequivocal evidence of a new mu-

tagenic mechanism in humans: insertions

of mobile genetic elements (transposons)

originating from the mostly silent, repeti-

tive sequences of the human genome. In

subsequent years, insertional mutagen-

esis proved a recurrent, albeit uncommon,

cause of genetic disease. Now, a paradigm

shift is placing new emphasis on transpo-

son control as a vital cellular function that

is compromised by and potentially contrib-

utes to many diseases. Transposon expres-

sion and activity is perhaps most overt in

cancers, although aberrant expression of

LINE-1, endogenous retroviruses (ERVs),

and other repeats have also been invoked

as a dimension of developmental, degen-

erative, and autoimmune diseases. As rig-

orous and accessible tools are developed to

investigate, new discoveries will clarify the

roles of repetitive elements and their regu-

lation in disease.

The human genome, like other complex

eukaryotic genomes, is filled with highly

repetitive DNA attributable to the activ-

ity of self-propagating genetic sequences.

In humans, this landscape is dominated

by retrotransposons that make new copies

of themselves by first being transcribed to

RNA intermediates and then reverse tran-

scribed to cDNA that is ultimately inte-

grated into the genome. Altogether, nearly

half of the human nuclear genome is trans-

poson-derived—a record of the activity of

various types of retrotransposons over mil-

lennia. The overwhelming majority of these

sequences are incomplete or mutated and

no longer active as mobile elements; how-

ever, there are exceptions. Subfamilies of

LINE-1 retain the ability to encode proteins

that make new genomic copies of LINE-1

and other noncoding repeats—namely Alu

short interspersed elements (SINEs) and

composite SINE-variable number tandem

repeat (VNTR)–Alu (SVA) elements. Other

functions can reside in transposable ele-

ments that are now inactive as insertional

mutagens. For example, many ERVs [ long

terminal repeat (LTR) sequences] contain

strong promoters or enhancers that can act

on nearby genes. To varying degrees, these

elements retain residual protein-coding ca-

pacity as well.

LINE-1 sequences make copies of them-

selves by producing a 6-kilobase (kb), bi-

cistronic RNA intermediate transcribed

from an internal RNA polymerase II pro-

moter. The first of its open reading frames

(ORFs) encodes ORF1p, which forms an

RNA-binding homotrimer. The second

ORF encodes ORF2p, which encompasses

endonuclease and reverse transcriptase

domains, which are responsible for mak-

ing cDNA copies of the LINE-1 RNA and

inserting these into the genome. ORF2p

functions can also be co-opted by noncod-

ing Alu and SVA elements to reverse tran-

scribe and integrate their sequences into

the genome instead of the protein-coding

LINE-1 RNA. Collectively, LINE-1, Alu, and

SVA insertions are uncommon but recur-

rent causes of genetic mutation, leading

to various constitutional genetic diseases.

In almost every case, the transposable el-

ement inserts into a critical portion of a

gene and causes a loss-of-function allele.

Disease-causing insertions can occur in

or be transmitted by the germline and thus

establish an allele frequency in a human

population. Another effect of germline

transmission is that each individual inher-

its a different complement of active trans-

posons ( 2 ). It may follow that the expres-

sion of mobile elements and the burden

of retrotransposition consequently varies

from person to person, and whether there

are limits to what is tolerable is unknown.

During development and in adults, trans-

position may contribute to genetic hetero-

geneity in various tissues (somatic mosa-

icism), but to what degree and whether

this is pathogenic are difficult to know.

Baseline activity may be physiologic. It has

been hypothesized that somatic retrotrans-

position is a source of functional genetic

diversity in normal neuronal development.

However, single-cell sequencing experi-

ments indicate that the level of somatically

acquired insertions in neurons is very low,

making a required role for these insertions

in normal development seem unlikely ( 3 ).

Like gene expression, transposable ele-

ment expression patterns are cell type–

specific and can depart from normal in

diseased states. For example, LINE-1 pro-

moter hypomethylation and LINE-1 over-

expression are hallmarks of many cancers,

and ORF1p accumulates in several of the

most common and commonly lethal human

malignancies ( 4 ) in association with loss of

the TP53 tumor suppressor gene ( 5 ). Other

types of retroelements show altered expres-

sion in cancers as well. This inspires prac-

tical questions, such as whether these can

be used as biomarkers of disease—as well

as more fundamental questions, such as

what regulatory mechanisms are responsi-

ble for expression of transposable elements,

and whether the expression of transpos-

able elements is epiphenomenal or causes

malignancy.

Transposable element–encoded regula-

tory sequences can alter chromatin states

and affect nearby gene expression (see the

figure). Thus, transposon activation may

underlie or be coordinated with gene ex-

pression changes that promote transforma-

tion, although demonstrations of this are

limited. What about transposon activation

independent of these regulatory effects?

There are anecdotal cases in which an in-

dividual has inherited a LINE-1 locus that

escaped silencing by promoter methylation

and went on to “drive” tumor development

through insertional mutagenesis, disrupt-

ing expression of the adenomatous pol-

yposis coli (APC) tumor suppressor gene

in colonic epithelium and leading to colon

cancer ( 6 ). Overt cancer-driving mutations

such as these are rare, however.

Somatically acquired LINE-1 insertions

that appear to be “passengers” (nondriver

alterations) in cancer genomes are much

more commonplace ( 7 ). If individual inser-

tions are functionally unimportant, do they

collectively represent a meaningful muta-

tional signature? Recent studies indicate

that LINE-1 insertion intermediates can

cause chromosomal rearrangements ( 7 ) and

pose barriers to DNA replication ( 8 ). These

data suggest that DNA lesions induced dur-

ing transposition may have underrecog-

nized roles in shaping cancer genomes and

that acquiring tolerance for LINE-1 activity

may be a prerequisite for tumorigenesis.

Given the widespread expression of

LINE-1 in cancers, it may be an intuitive

assumption that transposable element

activation promotes malignant growth.

Paradoxically, however, imposing transpo-

son expression in experimental systems

is often cytotoxic. Expression of LINE-1,

Alu, and ERV elements can induce in-

(^1) Dana Farber Cancer Institute, Boston, MA, USA.
(^2) Harvard Medical School, Boston MA, USA.
Email: [email protected]
Transposons become a focus of speculation and
scrutiny in biomedical research
GENETICS
Repetitive DNA in disease
22 APRIL 2022 • VOL 376 ISSUE 6591 353