One New Virus—How Many Old Diseases?

[Guest guest]

One New Virus—How Many Old Diseases?

M. Coffin1 and P. Stoye2

1Tufts University, Boston MA and 2National Institute for Medical Research, Mill

Hill,

London, UK

john.coffin@...

Gammaretroviruses are well-known causes of cancer, neurological disease,

immunodeficiency

and other disease in mice, cats, koalas, and some non-human primates. Indeed,

studies of these viruses, especially murine leukemia virus (MLV), in the last

century led

to important discoveries on which much of our current understanding of cancer

rests.

Despite considerable effort in those early studies, no clear evidence was

obtained convincingly

demonstrating human infection with gammaretroviruses or associating these

viruses with any human disease. The sarcasm " human rumor viruses " was coined to

describe

the many false alarms arising from this work (1). Recent studies may finally put

an end to the rumors, because they provide compelling evidence for frequent

human infection

with a mouse-derived gammaretrovirus and provide tantalizing clues pointing to

an association with two very different diseases.

In a study appearing in this issue of Science, collaborators from the Whittemore

Institute, the Cleveland Clinic, and the National Cancer Institute describe the

detection of

a virus called XMRV in about two-thirds of patients diagnosed with chronic

fatigue syndrome

(CFS) (Lombard et al.). As its name implies, CFS is an elusive condition

characterized

by debilitating fatigue persisting for many years, which has been reported to

affect

as much as 1% of the world population. Although abnormal immunological

parameters

consistent with chronic inflammation are often found in CFS patients, no

infectious or

toxic agent has ever been clearly implicated in this disease, and diagnosis is

largely by

exclusion of other conditions that cause similar symptoms (2). There is, in

fact, little

consensus in the medical community on whether CFS is a distinct disease.

Remarkably, CFS is not the first human disease to which XMRV has been linked.

The

virus was first described about 3 years ago in a few prostate cancer patients

(3). Although

the initial report received relatively little attention, a recent paper in PNAS

(4)

reported detection of XMRV in nearly a quarter of all prostate cancer biopsies,

and

higher grade tumors were more likely to contain the virus.

XMRV stands for " xenotropic MLV related virus, " reflecting the similarity of the

virus

isolated from both the prostate cancer and CFS patients to a group of endogenous

MLVs

found in the genomes of inbred and related wild mice. Endogenous viruses arise

when

retroviruses infect germ line cells. The integrated viral DNA, or provirus, is

then passed

on as part of the host genome (see Figure). Endogenous proviruses form a large

part of

the genetic complement of modern mammals – about 8% of the human genome, for

example.

The xenotropic proviruses first entered the mouse germline about a million years

ago and some endogenous proviruses still retain the ability to produce virus

that can infect

a variety of mammals, including humans. However, the viruses cannot infect cells

of

many of the mice that carry them (hence the name " xenotropic " ) because of a

mutation in

the cell surface receptor for the virus that arose in mice subsequent to entry

of the viruses

into the germ line. The propensity of xenotropic MLV to infect rapidly dividing

human

cells has made it a common contaminant in cultured cells, particularly in human

tumor

cell lines established by passage through nude mice (1).

There is greater than 90% sequence identity between XMRV and xenotropic MLV and

their biological properties are virtually indistinguishable (5-8) leaving little

doubt that the

former is derived from the latter, perhaps very recently, by one or more

cross-species

transmission events. There are several lines of evidence that transmission

happened in

the outside world and was not a laboratory contaminant. First, XMRVs from

disparate

locations and from both CFS and prostate patients are nearly identical: The

sequences of

the genomes differ by only a few nucleotides. However, there are hundreds of

differences

in the sequences of the XMRV viruses and the known endogenous murine xenotropic

proviruses of laboratory mice (Figure). Second, XMRV can be directly detected in

freshly isolated tissue from CFS and prostate cancer patients by antibody

staining, in situ

hybridization, electron microscopy, and by isolation of infectious virus. Third,

high levels

of antibodies reactive with XMRV and other MLVs are found in the blood of

affected

individuals. Taken together, these observations provide compelling evidence that

a significant

percentage of the CFS and prostate cancer patients are infected with XMRV.

There is still a great deal we do not understand. First, although there is a

clear association

of XMRV infection with CFS and prostate cancer, it is not clear whether XMRV

plays a causative role in either disease. For example, XMRV infection might, for

coincidental

reasons, be more frequent in the same geographical region as the CFS cluster.

Individuals

with CFS or prostate cancer might be more readily infected due to immune

activation.

In the prostate cancer patients, XMRV might prefer to grow in rapidly-dividing

prostate cancer cells (9). Also, the presence of rapidly dividing cells

associated with

these diseases might make infection more readily detectable. Second, we do not

know

how the virus is transmitted. The suggestion that there is sexual transmission

(7) is premature.

Given that infectious virus is present in plasma and in blood cells, blood-borne

transmission is also a possibility. The striking similarity of disparate

isolates raises the

possibility that there have been multiple independent mouse-to-human

transmission

events from an as-yet unidentified endogenous MLV. Third, we have no idea of the

prevalence or distribution of this virus in either human or wild mouse

populations. Finally,

the full pathogenic potential of XMRV remains to be explored. Animal models for

infection and pathogenesis are badly needed.

Two characteristics of XMRV are particularly noteworthy. First is the near

genetic identity

of isolates from different diseases from individuals in different parts of the

US. The

two most distantly related genomes sequenced to date differ by fewer than 30 out

of 8000

nucleotides. Thus, all of the XMRV isolates are more similar to each other than

are genomes

isolated from any one HIV-infected individual. In this respect, XMRV more

closely resembles human T-cell lymphotropic viruses (HTLVs) isolated from the

same

geographic region (10). As in the case with HTLV, the lack of diversity implies

that the

virus recently descended from a common ancestor, and that the number of

replication cycles

within one infected individual is limited. Second, the frequency of infection in

nondiseased

controls is remarkably high, about 4% in normal individuals from the same part

of Nevada as the CFS patients, and a similar fraction of normal prostate samples

was reported

to be infected in the cancer study. An initial survey of Japanese blood donors

showed a similar frequency of viral antibody-positive individuals (R. Furuta,

pers. com.)

If these figures are borne out in larger studies it would mean that perhaps 10

million people

in the United Sates and hundreds of millions worldwide are infected with a virus

whose pathogenic potential for humans is still unclear. However, it is clear

that closely

related viruses cause a variety of major diseases, including cancer, in many

other mammals.

Both laboratory and epidemiological studies are urgently needed to determine

whether this virus has a causative role, not only in these two diseases, but in

others as

well. However, it is important to remember that there is no evidence of any new

spreading

epidemic associated with this " new " virus. Instead, further study may reveal

XMRV

as a cause of more than one well-known " old " disease, with potentially

significant implications

for diagnosis, prevention, and therapy.

1. R. A. Weiss, in RNA Tumor Viruses R. A. Weiss, N. Teich, H. E. Varmus, J. M.

Coffin, Eds. (Cold Spring Harbor laboratory, Cold Spring Harbor, 1982).

2. L. D. Devanur, J. R. Kerr, J Clin Virol 37, 139 (Nov, 2006).

3. A. Urisman et al., PLoS Pathog 2, e25 (Mar, 2006).

4. R. Schlaberg, D. Choe, K. Brown, H. Thaker, I. Singh, Proc. Nat'l. Acad. Sci.

USA E Pub Sept 8 (2009).

5. B. Dong et al., Proc Natl Acad Sci U S A 104, 1655 (Jan 30, 2007).

6. B. Dong, R. H. Silverman, E. S. Kandel, PLoS One 3, e3144 (2008).

7. S. Hong et al., J Virol 83, 6995 (Jul, 2009).

8. S. Kim et al., J Virol 82, 9964 (Oct, 2008).

9. E. C. Knouf et al., J Virol 83, 7353 (Jul, 2009).

10. S. Van Dooren et al., Mol Biol Evol 21, 603 (Mar, 2004).

Endogenous MLV and XMRV. Xenotropic MLVs were derived from exogenous MLVs that

became established as a proviruses in the mouse germline, followed by loss of

the receptor

from the host. Although the virus can no longer infect mice, it can infect

humans, leading to

one (or more?) cross-species infection events to become XMRV. The inset shows

the

relationship among MLV genomes, including (from top to bottom): 5 XMRVs isolated

from

different patients, 2 CFS, and 3 prostate cancer (PC); XMV13, the most closely

related

provirus in the mouse genome sequence; endogenous xenotropic and ecotropic

viruses from

the indicated inbred strains, and a commonly used laboratory virus.

Endogenization

Receptor Mutation

CFS pt 1010

CFS pt 1042

PC VP 62

PC VP 42

PC VP 35

XMV13 Provirus

NZB Xenotropic MLV

AKR MLV

2%

BALB/c Xenotropic MLV

Moloney MLV

?