Nucleic acid-containing amyloid fibrils potently induce
type I interferon and stimulate systemic autoimmunity
Jeremy Di Domizioa, Stephanie Dorta-Estremeraa, Mihai Gageab, Dipyaman Gangulyc, Stephan Mellerc, Ping Lid,
Bihong Zhaoe, Filemon K. Tanf, Liqi Bid, Michel Gillietc, and Wei Caoa,1
Departments ofaImmunology andbVeterinary Medicine and Surgery, University of Texas MD Anderson Cancer Center, Houston, TX 77030;cDepartment of
Dermatology, University Hospital Lausanne, CH-1011 Lausanne, Switzerland;dSection of Rheumatology, University of Jilin University, Changchun 130033,
China;eDepartment of Pathology and Laboratory Medicine andfDivision of Rheumatology, University of Texas Medical School, Houston, TX 77030
Edited* by Michael B. A. Oldstone, The Scripps Research Institute, La Jolla, CA, and approved July 24, 2012 (received for review April 25, 2012)
The immunopathophysiologic development of systemic autoimmu-
I) produced by plasmacytoid dendritic cells (pDCs) critically promotes
the autoimmunity through its pleiotropic effects on immune cells.
However, the host-derived factors that enable abnormal IFN-I pro-
duction and initial immune tolerance breakdown are largely un-
known. Previously, we found that amyloid precursor proteins form
amyloid fibrils in the presence of nucleic acids. Here we report that
nucleic acid-containing amyloid fibrils can potently activate pDCs and
enable IFN-I production in response to self-DNA, self-RNA, and dead
cell debris. pDCs can take up DNA-containing amyloid fibrils, which
are retained in the early endosomes to activate TLR9, leading to high
IFNα/β production. In mice treated with DNA-containing amyloid
fibrils, a rapid IFN response correlated with pDC infiltration and acti-
vation. Immunization of nonautoimmune mice with DNA-containing
amyloid fibrils induced antinuclear serology against a panel of self-
antigens. The mice exhibited positive proteinuria and deposited anti-
bodies in their kidneys. Intriguingly, pDC depletion obstructed IFN-I
response and selectively abolished autoantibody generation. Our
study reveals an innate immune function of nucleic acid-containing
amyloid fibrils and provides a potential link between compromised
protein homeostasis and autoimmunity via a pDC-IFN axis.
autoimmune disease|innate immune response|disease model
antibodies (ANA), including those directed against DNA, ribonu-
cleoprotein complex (RNP), and nucleosomes (1, 2). These auto-
antibodies can form immune complexes (ICs), which are deposited
withinthekidneysandbloodvessels,and contribute critically tothe
pathogenesis of such diseases as lupus nephritis and vasculitis. A
significant number of patients with SLE have inadequate clearing
of apoptotic cell remnants, which include complex antigens con-
taining nucleic acids. The accumulation of these autoantigens
permits the eventual development of ANA.However, because self-
nucleic acids and apoptoticcelldebris are poorlyimmunogenic,the
mechanism behind the initial breakdown of immune tolerance
leading to systemic autoimmunity remains enigmatic.
Patients with SLE show increased levels of IFN-I in the serum
and expression of IFN-inducible genes in both peripheral blood
(3–5). Administration of IFNα to patients with malignant or viral
disease occasionally induces a lupus-like syndrome (5). In auto-
production and glomerulonephritis, whereas IFNAR deficiency
significantly ameliorates the disease (6–8). Functionally, IFN-I
potently differentiates monocytes, matures dendritic cells (DCs),
promotes B-cell differentiation and antibody production, modu-
lates survival, proliferation, and differentiation of T cells, and
primes neutrophils for death by NETosis (9–14). Therefore, IFN-I
acts as a central effector molecule to promote autoimmunity.
Amyloid fibrils are stable insoluble aggregates of misfolded
protein products with extensive β-sheet structures (15). Multiple
he precise etiology of systemic lupus erythematosus (SLE), a
heterogeneous autoimmune disease with multiple organ in-
aberrant polypeptides are implicated in more than 20 human
pathologies (16). Amyloid and related misfolded protein spe-
cies critically affect neuronal functions in the central nervous
system (CNS) and participate in inflammatory responses in both
CNS and peripheral organs (15, 17, 18). Previously, we charac-
terized how misfolded amyloid precursor proteins form amyloid
The fibrous aggregates containing nonproteinaceous cofactors
displayed the biophysical and biochemical features of amyloids
obtained in vitro and from patients, the latter of which are well
known to harbor significant amounts of nucleic acids and/or gly-
cosaminoglycans (20, 21).
Plasmacytoid dendritic cells (pDCs) are a unique innate im-
mune cell population that produces high amounts of IFN-I
(IFNα, -β, -ω, and -τ) upon sensing RNA or DNA by endosomal
TLR7 and TLR9, respectively (22, 23). ICs of autoantibodies to
chromatin and RNPs from SLE patients trigger the production
of IFN-I via activation of pDCs, a process that is mediated by the
Fcɣ receptor (24, 25). DNA-containing neutrophil extracellular
traps (NETs), production of which is accelerated by IFN-I and
autoantibodies in the SLE serum, also induce IFN-I production
by pDCs (13, 14). In another autoimmune condition, psoriasis,
complex of antimicrobial peptide LL-37 and self-nucleic acids
stimulates pDCs to secrete IFN-I (26). Given these findings on
protein–nucleic acid complexes, we examined whether nucleic
acid-containing amyloid fibrils can activate pDCs to induce IFN-
I and its immunological effects in vivo. Our results suggest that
nucleic acid-containing amyloid fibrils can function as a potent
IFN-I inducer both in vitro and in vivo. Intriguingly, a healthy
rodent host, in response to these complexes, developed systemic
autoimmunity with features mimicking SLE.
DNA-Containing Amyloid Fibrils Induce Strong IFNα/β Production by
pDCs. Twoprototypicamyloidogenic peptides, prionfragment and
amyloid β peptide 1–42 (Aβ), bind directly to DNA (19, 21). To
test whether DNA-containing prion or Aβ fibrils can activate
pDCs, we complexed them with oligonucleotide CpG B. CpG B
engages TLR9 in the late endosome and induces pDC to produce
abundant TNFα and IL-6, but little IFN-I (22). Interestingly with
CpG B, both peptides stimulated pDCs to produce elevated levels
of IFNα in a dose-dependent manner, with little effect on other
cytokines (Fig. S1), suggesting that DNA-containing amyloid
fibrils may selectively enhance IFN-I response by pDCs.
The natural amyloidogenic peptides, such as Aβ, undergo
spontaneous intermolecular rearrangement in solution to gen-
erate miscellaneous misfolding species (27). However, in fact,
Author contributions: J.D.D. and W.C. designed research; J.D.D., S.D.-E., and W.C. per-
formed research; D.G., S.M., P.L., and F.T. contributed new reagents/analytic tools;
M. Gagea, B.Z., L.B., and M. Gilliet analyzed data; and J.D.D. and W.C. wrote the paper.
The authors declare no conflict of interest.
*This Direct Submission article had a prearranged editor.
1To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
www.pnas.org/cgi/doi/10.1073/pnas.1206923109PNAS Early Edition
| 1 of 6
most polypeptides can adopt a native structure or form amyloid
fibrils by transitioning through a precursor state (16, 28). We have
characterized stabilized amyloid precursors with misfolded struc-
tures from various native proteins (19). Human serum albumin
(HSA), an abundant protein with native globular structure without
apparent immune stimulatory function, was chosen as a model to
investigate the innate immune functions of defined amyloid fibrils.
For that, four distinct structural variants of HSA were prepared:
native (HSA), amyloid precursor (AP-HSA), amyloid (A-HSA),
and fully denatured (D-HSA). Both native and AP-HSA are sol-
uble, whereas A-HSA and D-HSA are insoluble precipitates. A-
HSA readily bound to amyloid-specific dye Congo Red, indicating
the presence of β-sheet rich amyloid fibrils (Fig. 1A). AP-HSA, the
amyloid precursor species, which interacted mildly with Congo
Red (19), formed an insoluble precipitate upon mixing with CpG
B, which exhibited significantly enhanced Congo Red fluorescence
(Fig. 1A). This result is consistent with our previous observation
that amyloid precursors convert to amyloid fibrils upon binding to
DNA in a sequence-independent manner (19).
After overnight culture, AP-HSA complexed with CpG B in-
duced human primary pDCs to secrete significant levels of IFNα
and slightly increased TNFα and IL-6 (Fig. 1B). This result was
verified by FACS staining on intracellular IFNα associated with
pDCs and detection of increased IFN-I gene products after
stimulation by AP-HSA complexed with CpG B (Fig. 1 C and D).
Despite thestrong IFN-I stimulation, AP-HSA did not affectpDC
maturationbyCpG B (Fig.1D andFig. S2).Incontrast, neitherA-
HSA nor D-HSA, two other misfolded variants, affected IFN-I
production (Fig. 1 B and C). Because pDCs account for <1% of
the mononuclear cells in the blood (23), we investigated whether
CpG-containing HSA amyloid fibrils could selectively activate
pDCs amid other leukocytes. AP-HSA in the presence of CpG B
stimulated prominent and selective IFNα secretion from PBMCs,
which depended on the function of pDCs as IFN production was
abrogated by selective pDC depletion (Fig. 1E).
AP-HSA readily binds to genomic DNA isolated from salmon
sperm, which resulted in the generation of Congo Red positive
complex that displayed apple-green birefringence under polar-
ized light, a definitive indication of amyloid formation (Fig. 1F).
pDCs secreted significant levels of IFNα after stimulation by
HSA amyloid containing salmon sperm DNA (Fig. 1F, Lower)
and similarly by amyloid containing human genomic DNA (Fig.
S3). Therefore, our data indicate that DNA-containing amyloid
fibrils potently induce IFN-I by activating human pDCs.
Amyloid Fibrils Containing DNA Are Required for IFN-I Induction.
and AP-HSA may trigger separate signaling pathways in pDCs that
synergistically heighten the IFN-I response. To examine this pos-
sibility, we cultured pDCs with DNA and AP-HSA sequentially:
pDCs were cultured with DNA for 2 h, the cells were washed, and
then AP-HSA was added; in a second test, the cells were cultured
first with AP-HSA then incubated with DNA. pDCs produced
IFNα only in the presence of both DNA and AP-HSA, a condition
favoring the generation of DNA-containing amyloid fibrils (Fig.
S4A). Furthermore, fluorescent staining with the amyloid-specific
dye thioflavin S revealed the presence of amyloid aggregates inside
not to the native HSA-DNA mixture (Fig. S4B).
In addition to nucleic acids, amyloid precursor proteins bind to
other polyanionic cofactors, such as heparan sulfate glycosamino-
glycan, and form amyloid fibrils (19). In contrast to DNA-con-
taining amyloid fibrils, AP-HSA mixed with heparin failed to
activate pDCs to induce IFN-I (Fig. S4C). Moreover, heparin
inhibited the production of IFNα by PBMCs in response to CpG-
containing HSA amyloid fibrils (Fig. S4D), which is consistent with
its ability to compete with the formation of DNA-containing HSA
amyloid fibrils (19). Therefore, among the different types of amy-
loidfibrils, onlythe nucleicacid-containing aggregatescanpotently
induce IFN-I. A small compound polyphenol(-)-epigallocatechin
6hr 20hr 6hr 20hr 6hr 20hr
IFN 1 IFN 4 IFN
6hr 20hr 6hr 20hr 6hr 20hr
AP-HSA + DNA
No CpG B
absence or presence of CpG B with Congo Red. The intensity of fluorescent emission at 646 nm was plotted (mean ± SD from three independent experiments).
*P < 0.05. (B and C) Cytokine production by human pDCs stimulated with 1 μM CpG B with 5 μg/mL HSA variants measured by ELISA (B), or by intracellular
FACS staining (C). Data shown in B are the mean ± SEM from a representative donor (n > 10). Numbers in C indicate the percentage of population (rep-
resentative of four donors). (D) Gene expression by pDCs after stimulation determined by quantitative PCR analysis (levels are relative to that of a resting
PBMC sample). (E) Cytokine production by human PBMC cultured with HSA variants (10 μg/mL) with or without CpG B. Data shown are the mean ± SEM from
a representative donor (n = 4). (F) Birefringence analysis of AP-HSA in without or with salmon sperm DNA (Upper). IFNα secreted from pDCs cultured with HSA
or AP-HSA (5 μg/mL) and salmon sperm DNA (Lower) (mean ± SD).
DNA-containing amyloid induces prominent IFN-I production by pDCs. (A) Detection of β-sheet–rich structures in the HSA structural variants in the
2 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1206923109Di Domizio et al.
gallate (EGCG) selectively binds to amyloid precursors and
interferes with the amyloid formation (29). Preincubation of AP-
HSA with EGCG significantly reduced levels of IFNα produced by
PBMCs stimulated by CpG-containing HSA amyloid fibrils (Fig.
S4E). These results illustrate the functional importance of fibril
formation for DNA-containing amyloid to induce IFN-I.
DNA-Containing Amyloid Uptake by pDCs Activates TLR9 in Early
Endosomes. To understand how IFN-I production is selectively
enhanced by DNA-containing amyloid fibrils, we examined the
uptake of genomic DNAs by pDCs. In contrast to native HSA,
AP-HSA significantly increased the amount of DNA associated
with pDCs in a dose-dependent manner (Fig. 2A). The uptake of
DNA-containing amyloid seems to be cell type independent,
because effective internalization of DNA-containing amyloid
fibrils was also observed by Jurkat cells (Fig. S5).
Because TLR signaling from different endosomal compart-
ments leads to distinct cytokine response by pDCs (22), we then
examined the subcellular localization of AP-HSA complexed with
DNA in pDCs by costaining the cells with the early endosomal
marker EEA1 (Fig. 2B) and the late endosome/lysosome marker
LAMP1 (Fig. S6). After 4 h in culture, fluorescent AP-HSA and
DNA were detected inside pDCs, where they remained tightly
bound, as illustrated by the complete colocalization of their sig-
nals (Fig. 2B). Interestingly, the DNA–amyloid complex showed
significant colocalization with EEA1 but not with LAMP1, sug-
gesting the exclusive early endosome localization of the amyloid
fibrils. Similarly, both Aβ and prion peptides facilitate CpG B
to early ensosome (Fig. S7), which is likely responsible for the
enhanced IFN-I response (Fig. S1).
TLR9 recognizes the unmethylated CpG motifs present in the
2′ deoxyribose backbone of natural DNA (22, 23). In addition to
TLR9, other cellular DNA sensors can participate in IFN-I re-
sponse by recognizing biochemical features of DNA irrespective
of unmethylated CpGs (30). To investigate the specific role of
TLR9, we enzymatically methylated two DNA species, i.e.,
plasmid DNA and human genomic DNA, and prepared amyloid
fibrils by complexing them with AP-HSA. After methylation,
plasmid DNA obtained from a bacterial source became resistant
to digestion by the restriction enzyme BstU1, which recognizes
unmethylated CGCG sequences, whereas human DNA that
contains low-frequency unmethylated CpGs showed enhanced
BstU1 resistance (Fig. 2C, Left). When added to pDCs, amyloid
containing methylated DNA completely lost its ability to induce
IFN-I, demonstrating the requirement of unmethylated CpGs to
trigger IFN production (Fig. 2C, Right). Therefore, our data
collectively reveal that DNA as part of the complex amyloid is
effectively taken up by pDCs and delivered to early endosomes
to potently trigger TLR9-mediated IFN-I production.
Nucleic Acid-Containing Amyloid Fibrils Mediate IFN-I Response to
Self-RNA and Dead Cell Debris. Similar to their interaction with
DNA, AP-HSA mixed with total cellular RNA produced in-
soluble high molecular aggregate with enhanced Congo Red
emission, fibril formation (19), and retarded migration during
electrophoresis (Fig. 3A). Within this complex, RNA became
resistant to RNase digestion, indicating a protective effect of
the amyloid structure to the complexed nucleic acids. When
added to pDCs, the RNA-containing amyloid fibrils induced
significant levels of IFNα (Fig. 3B).
Upon death, cells release their cellular components and nu-
clear antigens. Preincubation of AP-HSA, but not other forms of
HSA, with the lysates of necrotic cells stimulated purified pDCs
to secrete IFNα (Fig. 3C). Consistently, IFNα production was
detected in PBMCs stimulated by AP-HSA complexed with the
cell debris (Fig. 3D). Such activation was sensitive to the pre-
treatment of the lysates with DNase and RNase, suggesting that
DNA- and RNA-containing amyloid fibrils formed in the mixture
are likely responsible for triggering IFN-I secretion.
Because IFN-inducing ICs implicated in SLE rely on the func-
tionofFcγR tomediate their cellularentry(5, 22),weinvestigated
whether DNA-containing amyloid fibrils also use surface FcγRIIa
(CD32) on pDCs. Although blocking CD32 significantly reduced
the amount of internalized DNA and secreted IFNα induced
by SLE serum, it had no effect on either the uptake of DNA-
containing amyloid fibrils or the amount of IFNα secreted by
pDCs after amyloid stimulation (Fig. 3 E and F). Overall, our
results demonstrate that amyloid fibrils containing self-DNA,
self-RNA, and dead cell debris can directly activate pDCs to
DNA-Containing Amyloid Fibrils Induce Infiltration of pDCs and IFNα/β
Production in Vivo. To examine the function of nucleic acid-con-
taining amyloid fibrils in an in vivo tissue environment, we injected
mixtures of native HSA or AP-HSA together with endotoxin-free
bacterial DNA into the peritoneal cavities of mice. pDCs infil-
trated to the site where DNA-containing amyloid fibrils were in-
oculated and retained locally for days afterward (Fig. 4A). In
contrast, few pDCs were found in the peritoneal cavities of mice
injected with DNA and/or HSA. An elevated number of DCs and
macrophages were also detected, but no significant difference
Hence, pDCs selectively infiltrate in vivo in response to DNA-
containing amyloid fibrils. We next analyzed the gene expression
by the peritoneal exudate cells 24 h after injection. Strikingly, in-
multiple IFN-I subtypes and a group of IFN inducible genes at
levels significantly higher than other treatments (Fig. 4B).
To examine the functional involvement, we depleted pDCs by
injecting mAb 120G8, an antibody recognizing the mouse pDC-
specific receptor BST2 (31), i.p. 24 h before amyloid inoculation.
DNA AP-HSA EEA1
plasmid genomic DNA
TLR9. (A) Uptake of DNA by pDCs after incubation with HSA or AP-HSA and
Alexa 647-labeled salmon sperm DNA. (B) Confocal analysis of pDCs con-
taining biotinylated AP-HSA (green) and A647-labeled DNA (red). Cells were
also stained for the early endosome marker EEA1 (blue). (C Left) Plasmid
DNA and human genomic DNA, unmodified or treated with CpG methyl-
transferase, were digested by BstU1. Shown is the separation of DNA fragments
together with a DNA ladder (center lane). (C Right) IFNα secretion by human
pDCs stimulated by AP-HSA complexed with unmodified or methylated DNA.
Shown are representative results from at least three different donors.
DNA-containing amyloid fibrils are endocytosed into pDCs to activate
Di Domizio et al.PNAS Early Edition
| 3 of 6
Preinjection of 120G8, but not isotype-matched control, reduced
the infiltrating pDCs by more than 90% in the peritoneal cavity
after DNA–AP-HSA complex inoculation (Fig. S8). As a result,
the transcription of not only IFN-I genes but also IFN-inducible
genes induced by amyloid fibrils was drastically decreased (Fig.
4C). Note that although masked in the heat map, the levels of
irf7 and isg15, two prominent IFN-inducible genes, were reduced
after pDC depletion (Fig. S9). Interestingly, the up-regulated
expression of several chemokines, such as CCL5 and CXCL9–11,
was unaffected by pDC depletion, suggesting an inflammatory
reaction that does not rely on the function of pDCs. Further
analysis revealed an elevated transcription of IL-1β triggered by
the DNA-containing amyloid fibrils, which was likewise inde-
pendent of pDCs (Fig. 4C). Overall, our in vivo analysis confirms
thein vitro human studyandsuggests that pDCsacutely senseand
respond to nucleic acid-containing amyloid in tissues.
- + - + -+ - + -+
bufferHSA AP-HSAA-HSA D-HSA
IFNα α (pg/ml)
necrotic cell supernatant
variants in the absence or presence of RNase A. (B and C) Cytokine secreted by pDCs stimulated with HSA proteins in the presence of human total RNA
(B; mean ± SD, four donors) or supernatants of necrotic Jurkat cells (C; mean ± SEM, representative of at least three donors). (D) IFNα secretion by PBMCs in
response to necrotic Jurkat supernatants in the presence of HSA or AP-HSA (n = 11). In some, necrotic supernatants were treated with enzyme before mixing
with HSA. P values were determined by a two-way ANOVA test. **P < 0.01. (E) Blocking of CD32 on uptake of DNA by pDCs in the presence of HSA proteins or
sera from healthy donor or SLE patient. Values of mean fluorescent intensity (MFI) are shown. (F) IFNα produced by pDCs in the presence of anti-CD32
blocking antibody. (E and F) Data are presented as mean ± SD (n = 4).
Amyloid fibrils containing self-nucleic acids trigger IFN-I production by pDCs. (A) Gel shift analysis of human total RNA mixed with HSA structural
Type I IFN
DNA + HSA
DNA + AP-HSA
rat IgG120G8 120G8
DNA + HSA
DNA + AP-HSA
Shown at Top are profiles of pDCs within CD11c+MHC-II+CD11b−population 24 h after i.p. injection. Quantification of kinetic infiltration is shown at Bottom
(mean ± SD, four mice per time point). (B) Peritoneal gene expression presented as a heat map. A PBS-treated animal was used as a reference. (C) Gene
expression by peritoneal cells 24 h after i.p. injection of HSA proteins with DNA in mice received pretreatment of antibodies (B and C) Shown are results from
one experiment of at least two independent experiments with similar results.
DNA-containing amyloid induces pDC-mediated IFN-I response in vivo. (A) Number of infiltrating antigen presenting cells in the peritoneum of mice.
4 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1206923109Di Domizio et al.
and Proteinuria in Healthy Mice. To evaluate the long-term immune
response by a healthy host to nucleic acid-containing amyloid
fibrils, we immunized wild-type female BALB/c mice with such
amyloid in comparison with PBS, DNA, and DNA mixed with
of complete Freund’s adjuvant (CFA), followed at 2-wk intervals
the total IgM and IgG levels between the experimental groups
(Fig. S10). However, the sera from the mice that received DNA-
containing HSA amyloid, but not from the other groups, displayed
nuclei of Hep-2 cells, the staining was also positive in the cyto-
plasm, implying a broad reactivity toward cellular antigens.
To elucidate the specificity of the ANA, we examined sera re-
activity to several well-known autoantigens implicated in SLE. In-
terestingly, mice that received DNA-containing HSA amyloid
developed significant antibody responses against single-stranded
DNA (ssDNA), total RNA, Sm/RNP complex, and histone over
that IgG1 and, to a lesser extent, IgG2a were the major Ig isotypes
within the anti-ssDNA response (Fig. S11). Consistent with the
negative ANA, control animals and mice immunized with DNA
and HSA showed no sign of autoantibody. The autoantibodies in-
duced by DNA-HSA amyloid are independent of the Abs against
HSA, because depletion of HSA-specific Ig had no impact on the
serum reactivity to the autoantigens (Fig. S12). At the time when
the mice were terminated 16 wk after immunization, none of sera
reacted with double-strand DNA. Because lupus nephritis is a ma-
jor organ-specific pathology associated with SLE, we analyzed the
renal function of the immunized mice and detected proteinuria in
the group that received DNA-containing HSA amyloid (Fig. 5C).
Furthermore, the deposition of IgG was found in the glomeruli of
the kidneys from these mice (Fig. 5D).
Because pDC depletion resulted in diminished acute IFN-I
response (Fig. 4C), we next investigated the role of pDCs in
antibody development. Strikingly, 120G8 preinjection largely
abolished the ANA response and severely affected the generation
of specific autoantibodies induced by DNA-containing amyloid
(Fig. 5 E and F). However, it did not affect the titer of anti-HSA
antibody or the proteinuria in the immunized mice (Fig. S13).
This finding suggests that pDC-IFN axis strongly influence the
immune reactions leading to autoantibody development. Overall,
these data collectively demonstrate that exposure of nucleic acid-
containing amyloid fibrils to a nonautoimmune host can result in
the development of lupus-like systemic autoimmunity.
By forming fibrous aggregates with amyloid precursor proteins,
self-nucleic acids are protected from nucleases in the environ-
ment, effectively taken up by pDCs, and then transported to the
endocytic compartment. A unique membrane trafficking pathway
with characteristics of endolysosomes are essential for TLR7/9
signaling and IFN production in pDCs (22, 32, 33). The nucleic
acid-containing amyloid is retained in the early endosomes of
pDCs, where the prolonged TLR9 activation can promote MyD88
signaling and subsequent IRF7 activation, which initiates the
transcription of all IFN-I subtypes. This mechanism is analogous
to other potent IFN-I inducers, i.e., type A CpG oligonucleotide
and LL-37 complexed with nucleic acids (22, 26). It is unclear
whether any specific pDC surface receptor mediates this process.
Amyloid β fibrils effectively attach to cells by interacting with
a wide array of surface receptors and directly with the phospho-
lipid bilayer (34). Interestingly, multiple amyloidgenic peptides
DNA + HSA
DNA + AP-HSA
DNA + HSA
DNA + AP-HSA
DNA + HSA
DNA + AP-HSA
immunization. (B) Levels of autoantibodies in the sera of the immunized mice determined by ELISA analysis (mean ± SD, five mice per group). (C) Levels of
albumin in the urine of immunized mice (mean ± SD, five mice per group). (D) Detection of IgG in the kidneys from immunized mice by staining with A488-
labeled anti-mouse IgG. (A and D) Shown are results with a representative mouse (n = 5). (A–D) Shown are results from one experiment of two independent
experiments with similar results (8–12 mice per group in combination). ANA response (E) and levels of antigen-specific autoantibodies (F) from mice that
received Ab pretreatment. Hep-2 cells were stained with a representative serum 9 wk after immunization (E). The box and whiskers plots of the data dis-
tribution of four mice per group are shown in F. P values were determined by a two-way ANOVA test.
BALB/c mice develop lupus-like autoimmunity after immunization with DNA-containing amyloid. (A) ANA reactivity from sera of mice 13 wk after
Di Domizio et al. PNAS Early Edition
| 5 of 6
enhance HIV attachment and entry into cells (35). Therefore,
nucleic acid-containing amyloid fibrils are unusually effective to
deliver nucleic acids to elicit IFN-I production by pDCs.
Despite the recognized importance of IFN-I in many autoim-
mune diseases, its role in the initiation phase of autoimmunity has
not been fully defined. In fact, only a minor fraction of patients
treated with IFNα develop ANA, and even a smaller fraction
spontaneous lupus do not exhibit significant upregulation of IFN-I.
Excessive IFN-I exposure exacerbates disease only in certain lupus-
that IFN-I requires certain genetic susceptibility or perhaps activa-
tion of additional pathway(s) to break immune tolerance. Here, we
demonstrate the capacity of nucleic acid-containing amyloid to in-
duce early IFN-I production upstream in a cascade of immune
responses, which eventually lead to the autoantibody generation.
Aberrant IFN-I production by pDCs has been implicated in
several human autoimmune disorders (4, 23). In SLE patients, the
numbers of circulating pDCs are reduced, whereas increased pDC
presence has been observed in the inflamed tissues (5, 22). Besides
secreting IFN-I, TLR-activated pDCs promote the generation of
plasma cells and antibody responses via IFN and IL-6 in vitro (10).
However, how pDCs participate in systemic autoimmunity in vivo
remains obscure. Here, we show the infiltration of pDCs shortly
after inoculation of nucleic acid-containing amyloid. Depletion of
pDCs not only abolished the IFN-I induction, but also severely and
selectively diminished the development of autoantibodies against
nuclear antigens. Therefore, IFN-producing pDCs play an essential
role in initiating systemic autoimmunity.
Amyloid fibrils are a product of failed protein homeostasis be-
cause of germ-line mutation, erroneous transcription/translation,
physical damage, or abnormal posttranslational processing (15).
Because human “amylome” constitutes approximately 15% of
all coding polypeptides in the genome, many “self” proteins have
the potential to form amyloid (28). Amyloid depositions are fre-
quently heterogeneous containing nonproteinaceous cofactors
(15, 20, 21). Our results suggest that only the type of amyloid-
containing nucleic acids is capable of inducing IFN-I through
activating nucleic acid-sensing TLRs. Interestingly, protein
misfolding products display another innate immune function: both
fibrillar Aβ and amyloid precursor of islet amyloid polypeptide
potently activate NALP3 inflammasome and induce IL-1β matu-
ration (17, 18). We also observed that DNA-containing amyloid
fibrils induced peritoneal inflammation in a pDC- and IFN-
independent manner, likely due to inflammasome activation
and IL-1β induction. Therefore, the protein misfolding products
can activate multiple innate immune pathways in vivo.
By immunizing nucleic acid-containing amyloid fibrils, we have,
in effect, created an inducible experimental lupus model. Previous
attempts to immunize nonautoimmune mice with self-antigens,
such as DNA, apoptotic cells, or purified nucleosome, only re-
sulted in limited or transient autoantibody generation (36–38).
Tetramethylpentadecane (TMPD) induces an array of auto-
antibodies and glomerulonephritis in BALB/c mice (39). In-
terestingly, prolonged oral administration of TMPD reportedly
leads to amyloidosis (40). Our model of experimental lupus
uniquely centers on the activation of pDC-IFN axis. It would be
important to investigate the critical cellular players and pathways
that lead to systemic autoimmunity.
Materials and Methods
Reagents. HSA structural variant proteins were prepared essentially as de-
scribed (19). Wild-type BALB/cByJ mice were obtained from the Jackson
Laboratory. Additional methods and detailed information can be found
in SI Materials and Methods.
ACKNOWLEDGMENTS. We thank Cametria Thompson, Ming Zhuo, and Ran
Zhang for technical assistance, and Drs. Yong-Jun Liu, Stephanie Watowich,
and Shao-Cong Sun for valuable suggestions. This research is supported by
the University of Texas MD Anderson Cancer Center Institutional Research
Grant (to W.C.), National Institutes of Health (NIH) Grant AI074809 (to W.C.),
and NIH MD Anderson’s Cancer Center Support Grant CA016672.
1. Cho JH, Gregersen PK (2011) Genomics and the multifactorial nature of human au-
toimmune disease. N Engl J Med 365:1612–1623.
2. Hahn BH (1998) Antibodies to DNA. N Engl J Med 338:1359–1368.
3. Theofilopoulos AN, Baccala R, Beutler B, Kono DH (2005) Type I interferons (α/β) in
immunity and autoimmunity. Annu Rev Immunol 23:307–336.
4. Banchereau J, Pascual V (2006) Type I interferon in systemic lupus erythematosus and
other autoimmune diseases. Immunity 25:383–392.
5. Rönnblom L, Pascual V (2008) The innate immune system in SLE: Type I interferons
and dendritic cells. Lupus 17:394–399.
6. Mathian A, Weinberg A, Gallegos M, Banchereau J, Koutouzov S (2005) IFN-α induces
early lethal lupus in preautoimmune but not in BALB/c mice. J Immunol 174:2499–2506.
7. Santiago-Raber M-L, et al. (2003) Type-I interferon receptor deficiency reduces lupus-
like disease in NZB mice. J Exp Med 197:777–788.
8. Agrawal H, et al. (2009) Deficiency of type I IFN receptor in lupus-prone New Zealand
mixed 2328 mice decreases dendritic cell numbers and activation and protects from
disease. J Immunol 183:6021–6029.
9. Blanco P, Palucka AK, Gill M, Pascual V, Banchereau J (2001) Induction of dendritic cell
differentiation by IFN-alpha in systemic lupus erythematosus. Science 294:1540–1543.
10. Jego G, et al. (2003) Plasmacytoid dendritic cells induce plasma cell differentiation
through type I interferon and interleukin 6. Immunity 19:225–234.
11. Le Bon A, et al. (2001) Type i interferons potently enhance humoral immunity and can
promote isotype switching by stimulating dendritic cells in vivo. Immunity 14:461–470.
12. Tough DF (2012) Modulation of T-cell function by type I interferon. Immunol Cell Biol
13. Garcia-Romo GS, et al. (2011) Netting neutrophils are major inducers of type I IFN
production in pediatric systemic lupus erythematosus. Sci Transl Med 3:73ra20.
14. Lande R, et al. (2011) Neutrophils activate plasmacytoid dendritic cells by releasing self-
DNA-peptide complexes in systemic lupus erythematosus. Sci Transl Med 3:73ra19.
15. Sideras K, Gertz MA (2009) Amyloidosis. Adv Clin Chem 47:1–44.
16. Schnabel J (2010) Protein folding: The dark side of proteins. Nature 464:828–829.
17. Halle A, et al. (2008) The NALP3 inflammasome is involved in the innate immune
response to amyloid-β. Nat Immunol 9:857–865.
18. Masters SL,etal.(2010)ActivationoftheNLRP3inflammasomebyisletamyloidpolypeptide
provides a mechanism for enhanced IL-1β in type 2 diabetes. Nat Immunol 11:897–904.
19. Di Domizio J, et al. (2012) Binding with nucleic acids or glycosaminoglycans converts
soluble protein oligomers to amyloid. J Biol Chem 287:736–747.
20. Zhang X, Li J-P, Lijuan Z (2010) Heparan sulfate proteoglycans in amyloidosis. Prog
Mol Biol Transl Sci 93:309–334.
21. Jiménez JS (2010) Protein-DNA interaction at the origin of neurological diseases:
A hypothesis. J Alzheimers Dis 22:375–391.
22. Gilliet M, Cao W, Liu Y-J (2008) Plasmacytoid dendritic cells: Sensing nucleic acids in
viral infection and autoimmune diseases. Nat Rev Immunol 8:594–606.
23. Colonna M, Trinchieri G, Liu YJ (2004) Plasmacytoid dendritic cells in immunity. Nat
24. Vollmer J, et al. (2005) Immune stimulation mediated by autoantigen binding sites
within small nuclear RNAs involves Toll-like receptors 7 and 8. J Exp Med 202:1575–1585.
25. Means TK, et al. (2005) Human lupus autoantibody-DNA complexes activate DCs
through cooperation of CD32 and TLR9. J Clin Invest 115:407–417.
26. Lande R, et al. (2007) Plasmacytoid dendritic cells sense self-DNA coupled with anti-
microbial peptide. Nature 449:564–569.
27. Stefani M (2010) Biochemical and biophysical features of both oligomer/fibril and cell
membrane in amyloid cytotoxicity. FEBS J 277:4602–4613.
28. Goldschmidt L, Teng PK, Riek R, Eisenberg D (2010) Identifying the amylome, proteins
capable of forming amyloid-like fibrils. Proc Natl Acad Sci USA 107:3487–3492.
29. Ehrnhoefer DE, et al. (2008) EGCG redirects amyloidogenic polypeptides into un-
structured, off-pathway oligomers. Nat Struct Mol Biol 15:558–566.
30. Kumar H, Kawai T, Akira S (2011) Pathogen recognition by the innate immune system.
Int Rev Immunol 30:16–34.
31. Asselin-Paturel C, Brizard G, Pin J-J, Brière F, Trinchieri G (2003) Mouse strain differences
in plasmacytoid dendritic cell frequency and function revealed by a novel monoclonal
antibody. J Immunol 171:6466–6477.
32. Blasius AL, et al. (2010) Slc15a4, AP-3, and Hermansky-Pudlak syndrome proteins are
required for Toll-like receptor signaling in plasmacytoid dendritic cells. Proc Natl Acad
Sci USA 107:19973–19978.
33. Sasai M, Linehan MM, Iwasaki A (2010) Bifurcation of Toll-like receptor 9 signaling by
adaptor protein 3. Science 329:1530–1534.
34. Williams TL, Serpell LC (2011) Membrane and surface interactions of Alzheimer’s Aβ
peptide—insights into the mechanism of cytotoxicity. FEBS J 278:3905–3917.
35. Roan NR, et al. (2011) Peptides released by physiological cleavage of semen coagulum
proteins form amyloids that enhance HIV infection. Cell Host Microbe 10:541–550.
36. Pisetsky DS (1996) Immune activation by bacterial DNA: A new genetic code. Immu-
37. Mevorach D, Zhou JL, Song X, Elkon KB (1998) Systemic exposure to irradiated apo-
ptotic cells induces autoantibody production. J Exp Med 188:387–392.
nucleosome complexes: Implications for thepathogenesis of SLE.J Exp Med 205:3007–3018.
39. Reeves WH, Lee PY, Weinstein JS, Satoh M, Lu L (2009) Induction of autoimmunity by
pristane and other naturally occurring hydrocarbons. Trends Immunol 30:455–464.
40. Ho FC, Fu KH (1987) A new model of AA-amyloidosis induced by oral pristane in BALB/c
mice. Br J Exp Pathol 68:413–420.
6 of 6
| www.pnas.org/cgi/doi/10.1073/pnas.1206923109 Di Domizio et al.