C-terminal truncations in human
3¢-5¢ DNA exonuclease TREX1
cause autosomal dominant
retinal vasculopathy with
Anna Richards1,22, Arn M J M van den Maagdenberg2,3,22,
Joanna C Jen4,22, David Kavanagh1,22, Paula Bertram1,
Dirk Spitzer1, M Kathryn Liszewski1, Maria-Louise Barilla-LaBarca5,
Gisela M Terwindt3, Yumi Kasai6, Mike McLellan6,
Mark Gilbert Grand7, Kaate R J Vanmolkot2, Boukje de Vries2,
Jijun Wan4, Michael J Kane4, Hafsa Mamsa4, Ruth Scha ¨fer4,
Anine H Stam3, Joost Haan3, Paulus T V M de Jong8–10,
Caroline W Storimans11, Mary J van Schooneveld12,
Jendo A Oosterhuis13, Andreas Gschwendter14, Martin Dichgans14,
Katya E Kotschet15, Suzanne Hodgkinson16, Todd A Hardy17,
Martin B Delatycki18,19, Rula A Hajj-Ali20, Parul H Kothari1,
Stanley F Nelson21, Rune R Frants2, Robert W Baloh4,
Michel D Ferrari3& John P Atkinson1
Autosomal dominant retinal vasculopathy with cerebral
leukodystrophy is a microvascular endotheliopathy with
middle-age onset. In nine families, we identified heterozygous
C-terminal frameshift mutations in TREX1, which encodes a
3¢-5¢ exonuclease. These truncated proteins retain exonuclease
activity but lose normal perinuclear localization. These data
have implications for the maintenance of vascular integrity in
the degenerative cerebral microangiopathies leading to stroke
We have previously described three families sharing common features
of retinal and cerebral dysfunction. Visual loss, stroke and dementia
begin in middle age, and death occurs in most families 5 to 10 years
later. These diseases map to 3p21.1–p21.3 (ref. 1) and are called
cerebroretinal vasculopathy (CRV)2, hereditary vascular retinopathy
(HVR)3,4and hereditary endotheliopathy, retinopathy, nephropathy
and stroke (HERNS)5. We now designate these illnesses as autosomal
dominant retinal vasculopathy with cerebral leukodystrophy (RVCL)
(OMIM 192315). The neurovascular syndrome features a progressive
loss of visual acuity secondary to retinal vasculopathy, in combination
with a more variable neurological picture1–7. In a subset of affected
individuals, systemic vascular involvement is evidenced by Raynaud’s
phenomenon and mild liver (micronodular cirrhosis)2,5and kidney
This retinal vasculopathy is characterized by telangiectasias, micro-
aneurysms and retinal capillary obliteration starting in the macula.
Diseased cerebral white matter has prominent small infarcts that often
coalesce to pseudotumors. Neuroimaging studies demonstrate con-
trast-enhancing lesions in the white matter of the cerebrum and
cerebellum. Histopathology shows ischemic necrosis with minimal
inflammation and small blood vessels occluded with fibrin5. The white
matter lesions resemble post-radiation vascular damage2. Ultra-
structural studies of capillaries show a distinctive, multilamellar
subendothelial basement membrane5.
By combining haplotypes in the three RVCL families, we narrowed
the disease gene to a 3-cM region between markers D3S1578 and
D3S3564 that encompassed B10 Mb, containing over 120 candidate
genes1. We then sequenced the full coding region and intron-exon
boundaries of 33 candidate genes within this region (Supplementary
Table 1 online).
Here we report the identification of mutations in TREX1
(NM_033627), encoding DNA-specific 3¢ to 5¢ exonuclease DNase III.
In the CRV2and HVR3,4pedigrees, a heterozygous 1-bp insertion
(3688_3689insG) leads to V235fs and a consequent premature stop. In
HERNS5, a heterozygous 4-bp insertion (3727_3730dupGTCA) results
in a frameshift at T249 (Fig. 1a,b).
Next, we examined six families with putative RVCL (Supplemen-
tary Table 2 online)2,6,7. In each, we identified frameshift mutations
affecting the C terminus of TREX1. In three, the alterationwas V235fs,
the same as that in the CRV and HVR pedigrees. Haplotype analysis
suggests that they are not related (data not shown). We did not detect
any of the mutations in panels of chromosomes matched by ancestry
or location (Supplementary Methods online). In the CRV and
Received 22 March; accepted 4 June; published online 29 July 2007; doi:10.1038/ng2082
1Department of Medicine, Division of Rheumatology, Washington University School of Medicine, St. Louis, Missouri 63110, USA.2Department of Human Genetics, and
3Department of Neurology, Leiden University Medical Centre, 2300 RC Leiden, The Netherlands.4Department of Neurology, University of California at Los Angeles,
Los Angeles, California 90095, USA.5Department of Medicine, Division of Rheumatology, North Shore Long-Island Jewish Health System, Lake Success, New York
11030, USA.6Genome Sequencing Center, and7Department of Ophthalmology, Washington University School of Medicine, St. Louis, Missouri 63110, USA.
8Department of Ophthalmogenetics, Netherlands Institute for Neuroscience, Royal Netherlands Academy of Arts and Sciences, 1000 GC Amsterdam, The Netherlands.
9Department of Ophthalmology, Academic Medical Centre, 1100 DD Amsterdam, The Netherlands.10Department of Epidemiology and Biostatistics, Erasmus Medical
Centre, 3000 CA Rotterdam, The Netherlands.11Meander Medical Centre, 3800 BM Amersfoort, The Netherlands.12Department of Ophthalmology, University Medical
Centre, 3508 GA Utrecht, The Netherlands.13Department of Ophthalmology, Leiden University Medical Centre, 2300 RC Leiden, The Netherlands.14Department of
Neurology, Klinikum Grosshadern, Universita ¨t Mu ¨nchen, D-81377 Mu ¨nchen, Germany.15Department of Neurology, Monash Medical Centre, Clayton, Victoria 3168,
Australia.16Department of Neurology, Liverpool Hospital, Liverpool, New South Wales 2170, Australia.17Department of Neurology, Concord Repatriation General
Hospital, Concord, New South Wales 2139, Australia.18Bruce Lefroy Centre for Genetic Health Research, Murdoch Childrens Research Institute, and19Department of
Paediatrics, University of Melbourne, Royal Children’s Hospital, Parkville, Victoria 3052, Australia.20Department of Rheumatic and Immunologic Diseases, Cleveland
Clinic Foundation, Cleveland, Ohio 44195, USA.21Department of Human Genetics, University of California at Los Angeles, Los Angeles, California 90095, USA.
22These authors contributed equally to this work. Correspondence should be addressed to J.P.A. (firstname.lastname@example.org).
1068VOLUME 39 [ NUMBER 9 [ SEPTEMBER 2007 NATURE GENETICS
© 2007 Nature Publishing Group http://www.nature.com/naturegenetics
HERNS families, all affected individuals over the age of 60 (but none
of the unaffected individuals over the age of 60) carried a TREX1
mutation (100% penetrance). In the HVR3,4family, 10 of the 11
mutation carriers over 60 years of age have retinopathy.
TREX1 (DNase III) is a DNA-specific 3¢ to 5¢ exonuclease ubiqui-
tously expressed in mammalian cells8–10. It is thought to function as a
homodimer, with a preference for single-stranded DNA and mispaired
3¢ termini8. TREX1 is a part of the SET complex11that normally
resides in the cytoplasm but translocates to the nucleus in response to
oxidative DNA damage12.
Recently, homozygous mutations in TREX1 have been reported to
cause Aicardi-Goutie `re syndrome (AGS)13. AGS is a rare, familial,
early-onset progressive encephalopathy featuring basal ganglia calcifi-
cations and cerebrospinal fluid lymphocytosis, mimicking congenital
viral encephalitis14. Notably, mutations associated with AGS disrupt
the enzymatic sites in TREX1. This loss of exonuclease function13
(Fig. 1) is hypothesized to cause the accumulation of altered DNA that
triggers a destructive autoimmune response13. No phenotype was
reported for the heterozygous carriers of these mutations; however,
a heterozygous mutation in TREX1 causing familial chilblain lupus
has been reported recently15.
The distinctive clinical course and pathology of RVCL compared
with AGS suggests separate disease mechanisms. The frameshift
mutations observed in RVCL are downstream of the regions encoding
the catalytic domains, whereas in AGS, homozygous mutations occur
that alter exonuclease function. The heterozygous mutations observed
in RVCL did not impair the enzymatic activity of TREX1 (Fig. 2a), in
comparison with the R114H substitution in AGS13.
To investigate how the RVCL TREX1 proteins differ from the wild
type, we performed expression studies using confocal microscopy on
cells transfected with TREX1 tagged with a fluorescent protein
(Fig. 2b and Supplementary Fig. 1 online). The wild-type TREX1
labeled with fluorescent protein (FP-TREX1) localized to the peri-
nuclear region. In contrast, the TREX1 proteins FP-V235fs and
FP-T249fs were diffusely distributed in the cytoplasm and the nucleus,
as was the case for the fluorescent protein alone (Fig. 2b and
Supplementary Videos 1–4 online). Protein blotting confirmed that
the expressed proteins were of the correct size (Fig. 2c). These results
suggest a perinuclear targeting signal within the C terminus of TREX1.
Consequently, we generated a construct containing the C-terminal
106 amino acid residues of TREX1 (FP-C-106). This protein showed a
perinuclear localization pattern identical to that of the wild-type
TREX1 protein (Fig. 2b). The TREX1 protein containing amino
acid change R114H, found in AGS, also had the same pattern as the
wild-type protein. In contrast, the proteinwith the alteration closest to
the C terminus of TREX1, FP-287fs, was diffusely distributed, like the
other two truncated proteins (data not shown).
The TREX1 proteins foundin individualswith RVCL lack part of the
tion with the SET proteins and therefore may prevent formation of the
SET complex. The SET complex is hypothesized to target DNA repair
factors, including TREX1, to damaged DNA under conditions of
oxidative stress11,12. Lack of sufficient TREX1 associated with the SET
complex may result in failure of granzyme A–mediated cell death12.
cytoplasm may have detrimental effects, especially on endothelial cells.
The clinical syndromes in these families and the study of
their mutations should deepen our understanding of exonuclease
function, homeostasis of the endothelium and events leading to
premature vascular aging. RVCL and cerebral autosomal dominant
arteriopathy with subcortical infarcts and leukoencephalopathy
(CADASIL) represent two examples of monogenic disease featuring
a cerebral microangiopathy for which the genetic defects are now
240250 260270 280
290300 310 320
R114H fs E20 R164X
Figure 1 Diagram of TREX1 protein. (a) TREX1 has three exonuclease
domains. Mutations in italics are associated with AGS13, and those in
boldface blue at the C terminus are associated with RVCL. (b) Comparison
of the amino acid sequence of the C terminus of wild-type (WT) TREX1 with
RVCL associated mutations. The abnormal sequence introduced by the
frameshifts is depicted in blue.
FP FP-TREX1 FP-V235 fsFP-T249 fsFP-C-106
2 3 4 5 6
CPM × 103
Figure 2 Functional consequences of RVCL associated TREX1 mutations.
(a) Assessment of 3¢-5¢ exonuclease activity using equivalent amounts of
purified recombinant proteins expressed in E. coli. (b) Confocal microscopy
of HEK293T cells showing transiently expressed fluorescent protein (FP)-
tagged TREX1 proteins (green), TOPRO3 staining of nuclei (red) and overlay
(yellow). Similar expression patterns were obtained for wild-type protein
and for proteins derived from constructs containing mutations associated
with AGS and RVCL in CHO, HL-60 and HeLa cells (data not shown).
(c) Protein blot analysis of untransfected cells (1) and cells transfected with
enhanced yellow fluorescent protein (eYFP) (2), wild-type TREX1 (3), TREX1
mutants (4,5) and the C-terminal 106 amino acids (6), all linked to eYFP.
NATURE GENETICS VOLUME 39 [ NUMBER 9 [ SEPTEMBER 20071069
© 2007 Nature Publishing Group http://www.nature.com/naturegenetics
known and from which we can gain new insights into the origin of Download full-text
strokes and dementia.
We obtained consent from all participants in this study, and the
study was approved by the Office for Protection of Research Subjects
at UCLA and the Human Research Protection Office at Washington
University School of Medicine.
Note: Supplementary information is available on the Nature Genetics website.
We appreciate the cooperation of the participating families. We thank
M. Bogacki, E. van den Boogerd, J. van Vark and S. Keradhmand-Kia. A.R. is a
2005/2006 Fulbright Distinguished Scholar. D.K. is a Kidney Research UK clinical
training fellow. A.R. and D.K. are recipients of Peel Medical Trust Travel
Fellowships. At Washington University in St. Louis, this study has been funded by
the Center for Genome Sciences Pilot-Scale Sequencing Project Program and by
the Danforth Foundation. The Netherlands Organization for Scientific Research
(NWO) (Vici 918.56.602), the European Union ‘‘Eurohead’’ grant (LSHM-
CT-2004-504837) and the Center of Medical System Biology established by the
Netherlands Genomics Initiative/NWO supported the work in the Netherlands.
US National Institutes of Health (NIH)/National Institute on Deafness and Other
Communication Disorders (NIDCD) grant P50 DC02952 (R.W.B.), NIH/National
Eye Institute grant R01 EY15311 and a Stein-Oppenheimer Award (J.C.J.)
supported the work at University of California, Los Angeles. R.S. is the recipient
of a scholarship from the German National Scholarship Foundation.
COMPETING INTERESTS STATEMENT
The authors declare no competing financial interests.
Published online at http://www.nature.com/naturegenetics
Reprints and permissions information is available online at http://npg.nature.com/
1. Ophoff, R.A. et al. Am. J. Hum. Genet. 69, 447–453 (2001).
2. Grand, M.G. et al. Ophthalmology 95, 649–659 (1988).
3. Storimans, C.W. et al. Eur. J. Ophthalmol. 1, 73–78 (1991).
4. Terwindt, G.M. et al. Brain 121, 303–316 (1998).
5. Jen, J. et al. Neurology 49, 1322–1330 (1997).
6. Cohn, A.C. et al. Clin. Experiment. Ophthalmol. 33, 181–183 (2005).
7. Weil, S. et al. Neurology 53, 629–631 (1999).
8. Mazur, D.J. & Perrino, F.W. J. Biol. Chem. 276, 17022–17029 (2001).
9. Mazur, D.J. & Perrino, F.W. J. Biol. Chem. 274, 19655–19660 (1999).
10. Hoss, M. et al. EMBO J. 18, 3868–3875 (1999).
11. Chowdhury, D. et al. Mol. Cell 23, 133–142 (2006).
12. Martinvalet, D., Zhu, P. & Lieberman, J. Immunity 22, 355–370 (2005).
13. Crow, Y.J. et al. Nat. Genet. 38, 917–920 (2006).
14. Goutieres, F. Brain Dev. 27, 201–206 (2005).
15. Rice, G. et al. Am. J. Hum. Genet. 80, 811–815 (2007).
1070VOLUME 39 [ NUMBER 9 [ SEPTEMBER 2007 NATURE GENETICS
© 2007 Nature Publishing Group http://www.nature.com/naturegenetics