Rodent blood-stage Plasmodium survive in dendritic
cells that infect naive mice
Michelle N. Wykesa,1, Jason G. Kayb,2,3, Anthony Mandersonb,2,4, Xue Q. Liua,2,5, Darren L. Brownb,2,
Derek J. Richarda,2,6, Jiraprapa Wipasac, Suhua H. Jianga,d,e, Malcolm K. Jonesa,f, Chris J. Janseg, Andrew P. Watersh,
Susan K. Piercei, Louis H. Millerj,1, Jennifer L. Stowb, and Michael F. Gooda,1,5
aQueensland Institute of Medical Research, Brisbane, Queensland, Australia 4029;bInstitute for Molecular Bioscience, andfSchool of Veterinary Science,
University of Queensland, Brisbane, Queensland, Australia 4072;cUniversity of Chiang Mai Research Institute for Health Sciences, Chiang Mai 50200, Thailand;
dDepartment of Immunology, Tongji Medical College of Huazhong University of Science and Technology, Wuhan City, Hubei, People’s Republic of China
430030;eDepartment of Pathogenic Biology and Immunology, Shihezi University School of Medicine, Shihezi City, Xinjiang, People’s Republic of China
832002;gDepartment of Parasitology, Center of Infectious Diseases, Leiden University Medical Center, 2333 ZA, Leiden, The Netherlands;hWellcome Centre
for Molecular Parasitology and Faculty of Biomedical Life Sciences, Glasgow Biomedical Research Centre, University of Glasgow, Glasgow G12 8TA, Scotland,
United Kingdom; andiLaboratory of Immunogenetics, andjHead, Malaria Cell Biology Section, National Institute of Allergy and Infectious Diseases, National
Institutes of Health, Rockville, MD 20852
Contributed by Louis H. Miller, June 2, 2011 (sent for review January 28, 2011)
Plasmodium spp. parasites cause malaria in 300 to 500 million indi-
viduals each year. Disease occurs during the blood-stage of the
parasite’s life cycle, where the parasite is thought to replicate ex-
clusively within erythrocytes. Infected individuals can also suffer
relapses after several years, from Plasmodium vivax and Plasmo-
dium ovale surviving in hepatocytes. Plasmodium falciparum and
Plasmodium malariae can also persist after the original bout of in-
fection has apparently cleared in the blood, suggesting that host
cells other than erythrocytes (but not hepatocytes) may harbor
these blood-stage parasites, thereby assisting their escape from
host immunity. Using blood stage transgenic Plasmodium berghei-
the parasite had a tropism for CD317+dendritic cells. Other studies
using confocal microscopy, in vitro cultures, and cell transfer studies
showed that blood-stage parasites could infect, survive, and repli-
cate within CD317+dendritic cells, and that small numbers of these
cells released parasites infectious for erythrocytes in vivo. These
data have identifieda uniquesurvival strategy for blood-stage Plas-
modium, which has significant implications for understanding the
escape of Plasmodium spp. from immune-surveillance and for vac-
immune evasion|rodent malaria
ozoites under the skin of the host (1). Studies using Plasmodium
berghei sporozoites, showed that a proportion will remain in the
skin and infect keratinocytes (2), others are drained by the lym-
phatic system and are trapped in lymph nodes, and a fraction of
the deposited sporozoites enter blood vessels to migrate to the
liver (3). In the liver, typically between 1 and 10 sporozoites in-
vade hepatocytes. Other studies have shown that sporozoites can
invade and migrate through other cell types, including macro-
phages (4), Kupffer cells (5, 6), epithelial cells, and fibroblasts (7).
The sporozoites within hepatocytes develop by a process of
schizogony into merozoite forms, which escape from an infected
liver cell into the sinusoid lumen (8) to invade RBC. Within RBC,
the merozoites then develop into “ring” trophozoites, then ma-
ture trophozoites, and finally a schizont containing up to 32 new
merozoites. These schizont-infected RBC then rupture to release
merozoites that are able to invade new RBCs, resulting in an in-
crease of parasite biomass. The Plasmodium life cycle continues
whensomemerozoites develop into thesexualparasite stages,the
male and female gametocytes, which can be taken up by mos-
quitoes during blood meals (9).
Some Plasmodium infections, such as Plasmodium malariae
(10) and Plasmodium inui (11), persist for years and sometimes
for the life of the host (12), and the reasons for this are not un-
derstood. To date, the blood-stage parasites of mammalian
Plasmodium spp. are thought to survive and replicate only within
alaria commences when an infected female anopheline
mosquito bites and deposits, up to 125 Plasmodium spor-
RBC, but it has long been suspected that they may also have an-
other survival strategy. The blood-stage-parasite of rodent Plas-
modium spp. have been observed as sacks of merozoites within
macrophages and polymorphonuclear leukocytes (13). The mer-
ozoites were associated with pigment that suggests that cells
ingested parasitized RBC (pRBC) as opposed to invasion of the
cells by merozoites. These merozoite-associated cells were prin-
cipally in the spleen, thought to be formed by either pitting or
phagocytosis of parasites by macrophages (13). It is thus possible
that this is a site of latency, but proof that the transfer of infection
with these cells was not contaminated with infected RBC is
lacking. Furthermore, blood-stage parasites have also been found
inside platelets (14) but the intrathrombocytic environment did
not support parasite growth or replication (15) and infectious
studies were not undertaken. Finally, avian Plasmodium also has
two stages in which the parasite invades nonerythrocytes: during
the initial pre-erythrocytic stages of infection and later in the
blood-stage (16). The infection starts when the sporozoites from
the mosquito invade skin mononuclear cells where schizonts de-
velop and release merozoites, which have three phases. These
merozoites first invade mononuclear cells throughout the body,
develop through schizogony, and are released to invade eryth-
rocytes as part of the second phase. In the final phase, merozoites
from erythrocytes invade many different endothelia where the
parasite grows (second exo-erythrocytic stage) (16).
We originally suspected that blood stage Plasmodium spp.
might reside within an unidentified cell type and be protected
from immunity, when passive transfer of Plasmodia-specific
antibodies into immunocompromised mice could significantly
reduce but not eliminate parasites in the blood (17). We then
found that CD11c+dendritic cells (DCs) from mice infected with
Author contributions: M.N.W., S.K.P., L.H.M., J.L.S., and M.F.G. designed research; M.N.W.,
J.G.K., A.M., X.Q.L., D.L.B., D.J.R., J.W., S.H.J., and M.K.J. performed research; C.J.J. and
A.P.W. contributed new reagents/analytic tools; M.N.W., S.K.P., L.H.M., J.L.S., and M.F.G.
analyzed data; and M.N.W., C.J.J., L.H.M., J.L.S., and M.F.G. wrote the paper.
The authors declare no conflict of interest.
Freely available online through the PNAS open access option.
1To whom correspondence may be addressed. E-mail: email@example.com, michael.
firstname.lastname@example.org, or email@example.com.
2J.G.K., A.M., X.Q.L., and D.L.B. contributed equally to this work.
3Present address: Division of Cell Biology, Hospital for Sick Children, Toronto, ON, Canada
4Present address: Biological Sciences Section, Therapeutic Goods Administration, PO Box
100, Woden Australia Capital Territory 2606, Australia.
5Present Address: Institute for Glycomics, Griffith University, Gold Coast Campus, Gold
Coast, Queensland, Australia 4222.
6Present address: Faculty of Science and Technology, Queensland University of Technol-
ogy, Brisbane, Queensland 4059, Australia.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
| July 5, 2011
| vol. 108
| no. 27
Plasmodium yoelii transferred an infection to naive mice unless
the donor mice were treated with antimalarial drugs, suggesting
that some parasites might reside within the DCs. Here we dem-
onstrate, predominantly using a transgenic blood-stage P. berghei-
expressing GFP (PbGFP) (18), that the rodent malaria parasite
P. berghei can survive and replicate within CD317+DCs and that
a small percentage of these DCs release parasites that are in-
fectious for erythrocytes on transfer to naive mice.
Blood-Stage Plasmodium Has a Tropism for CD317+DCs. To de-
termine if DCs could harbor P. berghei, we infected C57/B6 mice
with 104blood-stage, transgenic PbGFP (18, 19). Patent infection
(first day of parasites visible on a blood smear) occurs around days
4 to 6 after inoculation (Fig. 1A), and the infection is usually lethal
by expression of CD11c (20), total CD11c+DCs were isolated
of most major cell types, for studies by flow cytometry. After 7 d
in multiple experiments, >20% of spleen CD11c+DCs contained
GFP (Fig. 1C) and >85% of CD11c+GFP+DCs also expressed
CD317 (Fig. 1D), labeled by monoclonal antibodies PDCA-1
and 120G8 (21). Further characterization of the GFP+CD317+
CD11c+DCs found variable expression of other DC-associated
markers (B220, Ly6G, Siglec-H, F4/80, Dec205, and CD8) (Fig.
S1A). Data for DCs from naive mice are provided for contrast (Fig.
spleen contained GFP compared with <10% of CD317−DC sub-
populations by day 7 postinfection (Fig. S1C). Absolute numbers of
CD317+DCs were also significantly increased by day 7 (Fig. S1D).
All GFP+CD11c+DCs were “hypermature,” as defined by high-
level expression of MHC Class II, CD80, and CD86 (Fig. S1E). In
contrast with the >20% of CD11c+splenic cells, only ∼1% of
CD11c+DC in the peripheral blood contained GFP (Fig. S1F).
Blood-Stage Plasmodium Replicates Within CD317+DCs. To de-
termine if the GFP seen in DCs represented pRBC, CD317+DCs
were isolated from PbGFP-infected mice, stained with DAPI
(4′,6-diamidino-2-phenylindole, which binds DNA), and analyzed
by microscopy (Fig. 2). Multinucleated, DAPI+GFP+parasites
were seen within CD317+DCs (Fig. 2A, arrow), and parasites of
similar appearance were also visible within Ter119+pRBC (Fig.
multinucleated parasites within DCs (Fig. 2 C–E). Fig. 2D is a
magnified image ofthe inset boxshown on Fig. 2C tohighlight the
parasite has as a parasitophorous vacuole membrane (PVM). Fig.
2E is another higher magnification example of a parasite with
To determine if the multinucleated GFP+parasites seen within
DCs were derived from replication within CD317+DC, these
DCs were isolated from spleens of PbGFP-infected mice and
cultured for 15 to 18 h in the presence of 5 ethynyl-2′deoxyuridine
(EdU) that is incorporated into replicating DNA. The cells were
then processed to detect EdU, labeled for CD11c expression and
DAPI, and then CD11c+EdU+DAPI+DCs were cell-sorted for
(arrows) were found inside these DCs (Fig. 2F), indicating that
some parasites were proliferating within the DCs as they had in-
corporated EdU into their DNA during replication within cul-
tured DCs. These EdU+DAPI+small parasite nuclei did not,
however, colocalize with the autophagosome marker LC3 in the
same cell (Fig. 2F, arrowhead). Dividing EdU+DAPI+parasites
were never associated with lysosomal or autophagocytic mem-
branes. Other studies found that ∼25% (range of 14–40%) of
GFP+CD317+DCs had replicating parasites at the time of cul-
ture, seen by incorporated EdU (Fig. 2G). This finding was sup-
ported by microscopic analysis of ex vivo CD317+DCs containing
GFP+parasites, which found equivalent numbers of DCs had
single or multiple green parasites, suggesting many of the infected
DCs had probably supported proliferation in vivo (Fig. 2H).
Additional labeling ofCD317+DCswasundertaken asDCs,by
nature, degrade all proteins taken up by endocytosis/pinocytosis
except those maintained for transfer to naive B cells (22). These
studies found that when GFP+parasites were adjacent to or
colocalized with the late endosome/lysosome markers, LAMP1
(early lysosome associated membrane protein-1) (Fig. 3A, ar-
rowhead) and cathepsin D (Fig. 3B, arrowhead), GFP labeling
was reduced, suggestive of parasite degradation. However, there
were also GFP+parasites that did not colocalize with these
markers (Fig. 3 A and B, arrows; and Movie S1, which is a 3D
reconstruction of a PbGFP-infected CD317+DCs).
CD317+DCs Are Infected in Vitro by P. berghei Schizont-Infected RBC
in Culture. In vitro culture followed by adoptive-transfer studies
were undertaken to determine if CD317+DC could be infected
in vitro and transfer infection. Total CD11c+DC or CD317+DC
were isolated from naive mice and multiple samples cultured with
PbGFP schizont-infected RBC. In parallel, similar numbers of
schizont-infected RBC were cultured alone without DCs under
similar conditions for 42 to 48 h.
After 24 h, a sample of the CD11c+DCs cultured with para-
sites was examined by flow cytometry and GFP was found to be
predominantly associated with the CD11clopopulation of DCs
(Fig. 4A). After 42 to 48 h, aliquots of all parasite cultures with or
without DCs were injected intravenously into naive mice (106DC
per mouse), which were monitored for infection (Fig. 4B). All
mice given CD11c+DC or CD317+DC cultured with schizont-
infected RBC developed an infection. In contrast, after the 42 to
48 h in culture, no infection was detected in mice given schizont-
infected RBC cultured alone. This series of studies showed that
the culture conditions used did not support survival of PbGFP
schizont-infected RBC, but that CD317+DCs could support the
survival of PbGFP schizont-infected RBC in vitro and infect naive
mice on transfer of the DCs. Furthermore, despite extensive
attempts, we have never seen red cell membranes surrounding
parasites within DCs, suggesting that merozoites may be released
by PbGFP schizont-infected RBC in culture and invade DCs.
Demonstration That Infected RBC Were Not the Source of Infection in
DC Preparations. Although we could not detect contaminating
pRBC by Ter119 labeling of our DC preparations, it could be ar-
infected with P. berghei that expresses the reporter protein GFP (PbGFP) and
monitored for (A) parasitemia (Mean ± SEM, n = 11) and (B) survival (n = 11)
in multiple experiments. DCs were isolated from the spleens of infected mice
after 7 d and labeled to show expression of (C) CD11c and (D) CD317 by flow
cytometry. The density plot shows uptake of PbGFP by CD11c+DCs and the
histogram shows CD317 (thick line) expression on GFP+CD11c+DCs compared
with labeling by isotype control antibody (dotted line). Data are represen-
tative of multiple independent experiments that had similar results.
Plasmodium has a tropism for CD317+DCs. Groups of mice were
| www.pnas.org/cgi/doi/10.1073/pnas.1108579108Wykes et al.
gued that the infection was transferred as a result of a small
number of contaminating infected RBC in the DC preparations.
Toanalyzethis further,weisolatedCD317+DCsfrommice atday
7 after infection with PbGFP. These DCs were then labeled with
CD11c and Ter119 for FACS sorting of Ter119−CD11c+
DCs or Ter119−CD11c+CD317+GPF−
which were transferred to groups of naive mice. Although GFP+
DCs transferred the infection, no mouse given GFP−DCs de-
veloped infection (Table 1). This result demonstrated that our
sorting method was highly sensitive in discriminating between in-
fectious and noninfectious cells.
To further demonstrate that contaminating pRBC were not
responsible for transferring the infection, CD317+DCs were iso-
lated from naive mice and mixed with infected RBC in a DC:
pRBC ratio of 1:20 or 1:80. One-half of the sample was immedi-
ately treated with Gey’s solution to lyse RBC, CD317+DCs rei-
solated by MACS (magnetic-activated cell sorting) columns and
then labeled with antibody specific for CD11c. The CD11c+DCs
were cell-sorted and 106DCsper mouse transferred tonaive mice.
The other half of the sample was not treated with Gey’s solution
and adjusted to transfer an equal number of DCs (106per mouse)
to naive mice. In duplicate experiments, untreated DC transferred
the infection but all DC preparations treated with Gey’s solution
and isolated failed to transfer infection, confirming our standard
DC isolation procedure removed infected RBC.
To demonstrate that cell surface-associated merozoites, re-
leased following Gey’s lysis of pRBC, were not responsible for the
transfer of infection by GFP+CD11c+DCs, blocking studies were
undertaken (Fig. 4C). For these studies, Ter119+GFP+RBC and
GFP+CD11c+Ter119−DCs were isolated from PbGFP-infected
mice, treated with Gey’s solution (to free merozoites), followed by
incubation with an equal volume of either hyperimmune serum
specific for PbGFP (HIS) or naive mouse serum (NMS) before
transfusion into naive mice. The treatment of lysed pRBC with
HIS delayed the detection of parasitemia in recipient mice by 1 d,
with nearly 10-fold lower parasitemia, compared with freed mer-
ozoites treated with NMS (Fig. 4C; Left). In contrast, HIS did not
affect the infectivity of DCs (Fig. 4C; Right). These studies showed
that the parasites were internalized into DCs and were protected
from antibodies capable of neutralizing free merozoites.
Plasmacytoid DCs from PbGFP, P. yoelii 17XNL, and Pladmodium
chabaudi-infected Mice Are Infectious to Naive Mice. To demon-
strate whether CD317+DCs also supported parasite survival
in vivo, GFP+CD317+CD11c+Ter119−DCs were isolated from
spleens of mice infected with PbGFP. In particular, RBC in the
spleen cell preparation were lysed by Gey’s solution and cell
preparations were passed through a MACS magnet to deplete
hemozoin, residual schizonts, or any other cell containing hemo-
zoin, before positive selection of GFP+CD317+CD11c+Ter119−
DCs. Titrating numbers of these DCs (10–106) or infected RBC
(10–104) were transferred to groups of naive mice (n = 10) and
monitored for the development of blood-stage infections. Using
a limiting dilution analysis, it was estimated that ∼1 in 8,631
DC. Cohorts of naive mice were infected with
PbGFP and DC isolated after 7 d for assess-
ment by confocal microscopy. Representative
confocal z-stack examples show (A) GFP+
(green) parasites with three small DAPI+
(blue) nuclei (arrow on gray-scale DAPI im-
age) adjacent to a large DC nucleus (N).
(Scale bar, 5 μm.) (B) GFP+parasites in Ter-
119+RBC (marker for RBC). (C–E) Trans-
mission electron micrographs. (C) DC with
multinucleated (pN) parasite within DC, ad-
jacent to the DC nucleus (N). (Scale bar, 1
μm.) (D) Magnified image of parasite in C
Inset, to show PVM and Parasitophorous
vacuole (PVac). (E) Another magnified ex-
ample of a parasite in DC with two pN and
rough endoplasmic reticulum (RER) high-
lighted. (F) Confocal z-stack of DCs with EdU+
parasites (green), immunolabeled for the
autophagosomal marker LC3 (red) and DAPI
(blue). The DC shown has its own DAPI
stained nucleus (N) as well as multiple EdU+-
DAPI+PbGFP nuclei (arrows). Of note, ac-
tively dividing, EdU+parasites did not usually
colocalize with LC3 (arrowhead). (Scale bars,
5 μm.) (G) Flow cytometry profile of CD317+
DCs with parasite-associated GFP and uptake
of EdU by replicating parasites. (H) Percent-
age of total DCs with visible whole GFP+
trophozoites/schizonts with their own DAPI+
nuclei not associated with LAMP1.
Microscopy of Plasmodium within
Wykes et al. PNAS
| July 5, 2011
| vol. 108
| no. 27
GFP+CD317+CD11c+Ter119−DCs taken from infected mice
were infectious (Fig. 4D and Table 1), which equates to ∼360 to
580 infectiousDCs per spleen, asthe spleens ofinfectedmicehave
3 to 5 × 106CD317+DCs during a PbGFP infection (Fig S1D).
GPF+CD317−CD11c+, or GFP+CD8+DCs from infected mice
did not transfer the infection (Table 1). In comparison, 1 in 22
PbGFP-infected RBC were infectious (Table 1). Similarly, when
CD11c+CD317+DCs isolated from mice infected with P. yoelii
17XNL and transferred to naive mice, 1 in 2,193 DCs was in-
fectious(Table1 and Fig.S2) which,basedonnumbersofDCs per
spleen, equates to 4,559 infectious DC per spleen (Table 1).
To determine if CD317+DCs harbored Plasmodium after the
apparent clearance of infection, cohorts of mice were infected
20 d followed by recurrent bouts of low-level recrudescent para-
sitemia (Fig. 5A). Purified CD317+DCs and infected RBC were
taken after 8 d during peak parasitemia and after 20 d, when no
parasitized RBC were observed by microscopy, and equal cell
numbers (105on day 8 and 106on day 20) transferred to naive
recipients. Recipient mice were smeared daily to detect the first
appearance of parasites. All mice given RBC or CD317+DCs
taken from mice during peak parasitemia (day 8) developed
infections, although the mice given RBC generated higher para-
sitemia on day 6 posttransfer compared with mice given DC (Fig.
5B). However, when RBC or CD317+DCs were taken from mice
after the apparent clearance of blood-stage infection (day 20),
eight of eight mice given 106CD317+DCs developed ∼0.67%
parasitemia by day 6, compared with five of six mice that received
RBC and only developed ∼0.05% (P < 0.0007) parasitemia by the
same day (Fig. 5B). These data show that during the subpatent
stage of infection as measured by blood smears, CD317+DC can
harbor parasites and transfer infection more efficiently than the
lower level of infection by RBC.
This study shows that rodent-infecting blood-stage Plasmodium
spp. have a tropism for splenic CD317+DCs, which can promote
their survival and replication and that ∼360 to 4,559 DCs per
spleen will sustain infectious parasites. Furthermore, in studies of
nonlethal infections with P. chabaudi, these parasitized DCs can
release infectious parasites even during subpatent infections.
Confocal microscopy studies established that multinucleated par-
asites were within DCs and not associated with the extracellular
membrane folds, as described for the Toxoplasma tachyzoites (23)
or associated with extracellular infected RBC. Additional experi-
ments also excluded extraneous RBC or merozoites as being the
source of the infection in transfer studies. Furthermore, the ob-
servation that only GFP+CD317+CD11c+Ter119−DCs but not
GFP+CD317−CD11c+Ter119−DCs or GFP+CD8+DCs could
transfer infections to naive mice, highlighted that only CD317+
DC supported Plasmodium and transferred infectious parasites.
unique DC populations, termed “inflammatory DCs,” which are
micewere infected with PbGFP and CD317+DC isolated after 7 d for assessment
by confocal microscopy. Representative confocal z-stack examples show (A)
CD137+DCs with whole (arrows) and degrading (arrowhead) PbGFP, immunos-
tained with LAMP1 (red) and DAPI (blue), reconstructed to give a 3D view. Most
Plasmodia are GFP+but one, adjacent to a late endosome (red), has lost its GFP
(arrowhead). (Scale bar, 5 μm.) (B) Cathepsin D-labeled sections show an inverse
with strong GFP that is not in a labeled lysosome (arrow). (Scale bar, 5 μm.)
P. berghei and lysosomal markers within CD317+DCs. Cohorts of naive
CD11c+DCs or CD317+DCs were isolated from naive mice and cultured with
PbGFP schizonts for 42 to 48 h. Schizonts were cultured alone in parallel. All
cultures were then transferred to multiple cohorts of naive mice (n = 4–5 per
group), which were then smeared every 2 to 3 d to detect infection. (A)
CD11cloDC show uptake of PbGFP after 24-h culture compared with cells
labeled with isotype control antibody. Parallel samples of cultures with
schizonts only were used to define infected red cells based on GFP, size, and
granularity. Gate highlights that DCs with low CD11c expression take up the
parasite. (B) Total CD11c+DCs or CD317+DCs from naive mice support PbGFP
schizont survival in vitro as seen by the transfer of infectious Plasmodium to
naive mice. Data represent four independent experiments with similar
results. (C) Extracellular merozoites are not the source infection in DC
preparations. Ter119+GFP+RBC (pRBC) or GFP+CD11c+Ter119−DCs were
from blood or spleen of PbGFP-infected mice were treated with Gey’s so-
lution (to free merozoites) and an equal volume of either hyperimmune
serum specific for P. berghei (HIS) or serum from naive mice (NMS) was
added immediately. Cohorts of naive mice were transfused with either 104
intact pRBC per mouse, lysed-equivalent per mouse (ly pRBC; Left) or 104DCs
and all mice monitored daily for infection (Right). The data represent one of
two experiments with groups of four mice, which had similar results. Error
bars shown are Mean parasitemia ± SEM. (D) To calculate the frequency of
CD317+DC that support parasites in vivo, cohorts of mice were infected with
PbGFP and GFP+CD317+DC isolated after 8 d for transfer to naive recipients.
Plots of the percentage of uninfected mice versus numbers of GFP+CD317+
DCs or pRBC from PbGFP-infected mice were used to calculate the frequency
of infected DC or pRBC. (y = mx + c was used to calculate the frequency of
infectious DCs). The arrows indicate that all mice were infected.
CD317+DCs harbor infectious Plasmodia. To infect DC in vitro, total
| www.pnas.org/cgi/doi/10.1073/pnas.1108579108Wykes et al.
not found in the steady state can appear as a consequence of in-
fection or inflammation (24) and may express this molecule fol-
lowing IFN stimulation (21). Our data show that the GFP+DCs
are heterogeneous in surface expression, but consistent with the
infected CD317+DC being predominantly PDCs because GFP+
DCs secreted more IFN-α than GFP−DCs in infected mice (Fig.
S3); >50% consistently expressed B220 and Ly6G which are
markers of PDC, and <30% were DEC205+(25). Furthermore,
PDC are long-lived (14-d limit of testing) compared with con-
ventional DCs (<3 d) (25), and thus we hypothesize would be
better able to harbor infectious parasites even after peak para-
sitemia has cleared. Finally, we were able to show CD317+DC
from naive mice could take up PbGFP from schizont-infected
RBC,possibly throughinvasionasopposedtophagocytosis. These
infection to naive mice. Unfortunately, because of the very low
numbers of infectious DCs in the spleen and following in vitro
infection, we were unable to reliable quantify cell cycle charac-
teristics and determine if the parasites had: (i) a very prolonged
cell cycle with a slow development of trophozoites/schizonts; (ii)
“arrested forms” that can survive for prolonged periods after
in host RBC which constantly reinfect DC. Based on the low in-
fectivity of DCs on transfer to naive mice and the observation that
CD317+DC can harbor parasites and transfer infection at similar
levels during peak and subpatent infections, we hypothesize that
these DCs may hold arrested forms of parasites that survive for
extended periods. If DCs were constantly reinfected, then like
RBC, their infectivity would have been higher during peak para-
sitemia. Finally, we also suggest that arrested forms would have a
survival advantage over just having a prolonged cycle.
Based on our data, we hypothesize that the survival of Plas-
modium within immuno-privileged DCs may facilitate evasion
from immune surveillance. The observation of blood-stage para-
sites surviving and replicating only within spleen DCs is consistent
with observations of malaria in primates, where it has been shown
that primate malaria P. inui persists for the life of the host and
infection is only cleared if the spleen is removed after infection
(12). Although some monkeys died from high parasitemia after
splenectomy, those that survived were able to clear the infection
radical cure of the infection is the possible survival of parasites
human malaria. It is known that in patients with a previous history
of malaria, often with no new infections for many years, following
splenectomy suffer an abrupt onset of Pladmodium falciparum
(26), P. malariae (27), and Plasmodium vivax (28) malarias. These
splenectomies usually follow splenic trauma. We thus hypothesize
that trauma to the spleen may disturb CD317+DCs to release
parasites and trigger these relapses with malaria.
survival of the blood-stage of rodent malaria parasites, P. berghei,
P.chabaudi,andP.yoelii17XNL.Whether DCsareareservoir for
human Plasmodium spp. remains to be determined. Such findings
could have important implications for the development ofvaccine
strategies and potentially new antimalarial therapeutics.
Materials and Methods
Please refer to the SI Materials and Methods for details, materials, and
Animals. Specific pathogen-free, 6- to 8-wk-old female C57BL/6J mice were
obtained from the Animal Resources Centre. All animal procedures were ap-
Ethics Committee. This work was conducted under Queensland Institute of
Medical Research animal ethics approval number A0209-623M, in strict accor-
dance withthe“Australian code of practice for thecare and useof animals for
Preparation of Spleen Cells Depleted of RBC. Cohorts of 3 to 20 mice were
infected intravenously with 104PbGFP or 105P. yoelii 17XNL, or P.chabaudi
AS-infected RBC. The PbGFP line, as described by Franke-Fayard et al., was
shown to express GFP at all previously known blood stages and successfully
used to identify infected RBC by FACS (18). Spleens from naive or infected
C57BL/6J mice were digested with Collagenase D (Roche Diagnostics) and
DNase (Boehringer), as previously described (22). Approximately 1 × 108
pelleted spleen cells were resuspended in 1 to 2 mL of Gey’s solution and
incubated on ice for 1 to 3 min, shaking occasionally. The lysis was stopped
with Iscove’s Modified Dulbecco’s Media (IMDM; Invitrogen) containing 5%
FCS. The digested spleen cell suspension was then incubated with 100 μg/mL
Rat IgG, to block FcR binding of subsequently used labeling antibodies.
Plasmodium to naive mice
Frequency of CD317+DC that transfer infectious
Cell transferred Frequency of infectious cells
P. yoelii pRBC
P. yoelii CD317+DC
1/8,631 Infectious (360 DC/spleen)
No transfer of infection
No transfer of infection
No transfer of infection
1/2,193 Infectious (4,559 DC/spleen)
Naive mice were infected with PbGFP or P. yoelii 17XNL and CD317+DCs
isolated after 7 d. Titrating numbers (106,105, 104, 103, 102, 101) of sorted
GFP+CD317+DCs, GFP−CD317+DCs, GFP+CD317−DCs, GFP+CD8+DCs, GFP+
red cells (pRBC) from PbGFP infections or P. yoelii 17XNL pRBC and CD317+
DCs from P. yoelii 17XNL infections were transferred to multiple cohorts of
naive mice (n = 5–10 per group), which were then smeared every 2 to 5 d to
detect infection for up to 30 d. A plot of the log-percent uninfected mice
versus number of transferred cells (Fig. 4D and Fig. S2) was used to calculate
the frequency CD317+DCs required to transfer infection. Data are represen-
tative of two to three independent experiments with similar results.
determine if CD317+DC could harbor parasite during subpatent parasitemia,
multiple cohorts of mice were infected with nonlethal P. chabaudi AS. (A)
The percentage of parasitemia monitored for one round of infection (Mean
± SEM; n = 5). (B) DCs or RBC were taken after 8 d during peak parasitemia
and after 20 d when no patent parasite was observed by blood-smear mi-
croscopy (<0.001%). On day 8, 105CD317+DCs or pRBC and on day 20, 106
CD317+DCs or RBC were transferred to naive mice and blood smears made
daily for first sign of infection. Data represents two independent experi-
ments with similar results.
CD317+DCs harbor P. chabaudi during subpatent parasitemia. To
Wykes et al.PNAS
| July 5, 2011
| vol. 108
| no. 27
Isolation of Spleen DCs. Total DC populations were enriched from digested Download full-text
spleen cells with Dynabeads Mouse DC enrichment kit (Dynal) and additional
schizonts, or any other cells containing hemozoin. The DC-enriched cells were
then labeled with anti-CD11c or anti-PDCA (CD317) MACS microbeads to
isolate total CD11c+DC or CD317+DC, respectively (Miltenyi Biotec Gmb).
To isolate GFP+CD317+DCs, the CD317+DCs were labeled with anti–CD11c-
APC and Ter119-PE, and using a narrow gate on the pulse-width settings (to
exclude doublets), large GFP+CD317+CD11c+Ter119−DCs were sorted on a
MoFlo (Beckman Coulter) or FACSAria III cell sorter (BD Bioscience).
PbGFP Cell Transfer Studies. Titrating numbers of sorted GFP+Ter119 RBC (10–
104) or GFP+CD317+CD11c+Ter119−DCs (10–106) per mouse were injected
intravenously into cohorts of 5 or 10 naive mice per group. The development
of blood-stage infections in these mice was monitored by analysis of Giemsa
stained blood films made from tail blood every 1 to 2 d until infection was
patent in control groups or up to 3 wk.
P. yoelii 17XNL and P. chabaudi Cell Transfer Studies. To isolate CD317+CD11c+
Ter119−DCs, CD317+DCs were isolated from spleens of infected mice using
the method described in the previous sections and CD317+CD11c+Ter119−DC
were then cell-sorted.
Confocal Microscopy. GFP+CD11c+CD317+DCs (day 7) from PbGFP-infected
mice were immobilized on poly-L-lysine–coated slides (Sigma) and labeled to
detect LAMP1 (BD Bioscience), LC3 (ubiquitin-like protein Atg8, which is
a marker of autophagy; clone 5F10; Alexis Biochemicals), Cathepsin D, and
nuclear DNA with DAPI (4’,6-diamidino-2-phenylindole; Invitrogen), as in-
dicated in the text using immuno-fluorescent–labeling techniques described
previously (29). Microscopy was performed on a Zeiss LSM 510 META con-
focal microscope (Carl Zeiss).
Electron Microscopy. DCs were isolated as for light microscopy and then
processed for standard EM by fixation in glutarahdehyde, postfixation in
potassium ferricyanide-reduced osmium tetroxide and uranyl acetate, and
embedding in Epon resin following standard protocols.
Preparation of Schizonts. Mature schizonts were isolated from in vitro culture
of synchronized ring forms as described by Janse et al. (19) using Nyodenz
density-gradient centrifugation (Nycodenz 1.077A; NycoMed) or Optiprep
In Vitro Infections of DCs. Total CD11c+DCs or CD317+DCs were isolated from
naive mice (PDC) using MACS and cultured (1–2 × 106per tube) with isolated
schizonts in IMDM with 10% FCS in poly styrene tubes, at an approximate
ratio of one-to-two schizonts per DC. As controls, equivalent numbers of
schizonts were cultured in parallel. After 40 to 48 h, PBS was added to all
cultures, cell spun down at 200 × g, resuspended in PBS, transferred to naive
mice (5 × 105PDC per mouse) by intravenous injection, and all mice moni-
tored daily for infection.
Experiment to Exclude Red Cell Contamination. For these studies, ∼2 × 106
CD317+DCs were isolated from naive mice and mixed with 1.6 × 108infected
RBC (80-fold; 1.15 × 109total RBC). The sample was divided into two halves
and one stored on ice. The other half was treated with Gey’s solution, DCs
reisolated by MACS, and cell-sorted for CD11c cells. Five naive mice per
group were each transfused with 1 × 105sorted CD11c+CD317+DCs or un-
sorted 1 × 105CD11c+CD317+DCs with 8 × 106infected red cells from the
untreated preparations. In the duplicate experiment, 4 × 106DCs was mixed
with 8 × 107infected red cells (20-fold) and treated as described above.
Experiment to Exclude Extracellular Merozoites as a Source Infection in DC
Preparations. Ter119+GFP+RBC and GFP+CD11c+Ter119−DCs were isolated by
cell sorting from the blood and spleens of PbGFP-infected mice, respectively.
All cell preparations were treated with Gey’s solution (to free merozoites) for
1 to 2 min. One-half of the preparation was then incubated with hyperim-
mune serum specific for P. berghei (diluted 1/2 in culture media), produced by
repeated infection of mice followed by drug cure with pyrimethamine. The
other half was incubated with mouse serum from naive mice (diluted 1/2 in
culture media). After 30 min, cohorts of four naive mice were transfused with
monitored daily for infection.
ACKNOWLEDGMENTS. We thank Drs. Denise Doolan, Christian Engwerda,
and James McCarthy for their critical reading of the manuscript; Drs. Alan
Sher and Leanne Tilley for their discussion of the data; Ms. Grace Chojnowski
and Paula Hall for their excellent cell sorting skills; and Ms. Virginia McPhun
for her technical assistance with some of the earlier developmental assays.
This study was supported by the National Health and Medical Research
1. Medica DL, Sinnis P (2005) Quantitative dynamics of Plasmodium yoelii sporozoite
transmission by infected anopheline mosquitoes. Infect Immun 73:4363–4369.
2. Gueirard P, et al. (2010) Development of the malaria parasite in the skin of the
mammalian host. Proc Natl Acad Sci USA 107:18640–18645.
3. Amino R, Thiberge S, Shorte S, Frischknecht F, Ménard R (2006) Quantitative imaging
of Plasmodium sporozoites in the mammalian host. C R Biol 329:858–862.
4. Vanderberg JP, Chew S, Stewart MJ (1990) Plasmodium sporozoite interactions with
macrophages in vitro: A videomicroscopic analysis. J Protozool 37:528–536.
5. Pradel G, Frevert U (2001) Malaria sporozoites actively enter and pass through rat
Kupffer cells prior to hepatocyte invasion. Hepatology 33:1154–1165.
6. Ishino T, Yano K, Chinzei Y, Yuda M (2004) Cell-passage activity is required for the
malarial parasite to cross the liver sinusoidal cell layer. PLoS Biol 2(1):E4.
7. Mota MM, et al. (2001) Migration of Plasmodium sporozoites through cells before
infection. Science 291(5509):141–144.
8. Sturm A, et al. (2006) Manipulation of host hepatocytes by the malaria parasite for
delivery into liver sinusoids. Science 313:1287–1290.
9. Prudêncio M, Rodriguez A, Mota MM (2006) The silent path to thousands of
merozoites: The Plasmodium liver stage. Nat Rev Microbiol 4:849–856.
10. Tiburskaja NA, Vrublevskaja OS (1965) Clinical and experimental studies on quartan
malaria following blood transfusion and methods for preventing its occurrence. Bull
World Health Organ 33:843–851.
11. Schmidt LH, Fradkin R, Harrison J, Rossan RN, Squires W (1980) The course of
untreated Plasmodium inui infections in the rhesus monkey (Macaca mulatta). Am J
Trop Med Hyg 29(2):158–169.
12. Wyler DJ, Miller LH, Schmidt LH (1977) Spleen function in quartan malaria (due to
Plasmodium inui): Evidence for both protective and suppressive roles in host defense.
J Infect Dis 135(1):86–93.
13. Landau I, et al. (1999) Survival of rodent malaria merozoites in the lymphatic
network: Potential role in chronicity of the infection. Parasite 6:311–322.
14. Fajardo LF (1973) Letter: Malarial parasites in mammalian platelets. Nature 243:
15. Perkash A, Kelly NI, Fajardo LF (1984) Enhanced parasitization of platelets by
Plasmodium berghei yoelii. Trans R Soc Trop Med Hyg 78:451–455.
16. Huff CG (1957) Organ and tissue distribution of the exoerythrocytic stages of various
avian malarial parasites. Exp Parasitol 6(2):143–162.
17. Hirunpetcharat C, et al. (1999) Absolute requirement for an active immune response
involving B cells and Th cells in immunity to Plasmodium yoelii passively acquired with
antibodies to the 19-kDa carboxyl-terminal fragment of merozoite surface protein-1.
J Immunol 162:7309–7314.
18. Franke-Fayard B, et al. (2004) A Plasmodium berghei reference line that constitutively
expresses GFP at a high level throughout the complete life cycle. Mol Biochem
19. Janse CJ, Ramesar J, Waters AP (2006) High-efficiency transfection and drug selection
of genetically transformed blood stages of the rodent malaria parasite Plasmodium
berghei. Nat Protoc 1:346–356.
20. Inaba K, et al. (1997) High levels of a major histocompatibility complex II-self peptide
complex on dendritic cells from the T cell areas of lymph nodes. J Exp Med 186:
21. Blasius AL, et al. (2006) Bone marrow stromal cell antigen 2 is a specific marker of type
I IFN-producing cells in the naive mouse, but a promiscuous cell surface antigen
following IFN stimulation. J Immunol 177:3260–3265.
22. Wykes M, Pombo A, Jenkins C, MacPherson GG (1998a) Dendritic cells interact directly
with naive B lymphocytes to transfer antigen and initiate class switching in a primary
T-dependent response. J Immunol 161:1313–1319.
23. Courret N, et al. (2006) CD11c- and CD11b-expressing mouse leukocytes transport
single Toxoplasma gondii tachyzoites to the brain. Blood 107:309–316.
24. Shortman K, Naik SH (2007) Steady-state and inflammatory dendritic-cell development.
Nat Rev Immunol 7(1):19–30.
25. O’Keeffe M, et al. (2002) Mouse plasmacytoid cells: Long-lived cells, heterogeneous in
surface phenotype and function, that differentiate into CD8(+) dendritic cells only
after microbial stimulus. J Exp Med 196:1307–1319.
26. Looareesuwan S, Suntharasamai P, Webster HK, Ho M
splenectomized patients: report of four cases and review. Clin Infect Dis 16:361–366.
27. Tsuchida H, Yamaguchi K, Yamamoto S, Ebisawa I (1982) Quartan malaria following
splenectomy 36 years after infection. Am J Trop Med Hyg 31(1):163–165.
28. Garnham PC (1970) The role of the spleen in protozoal infections with special
reference to splenectomy. Acta Trop 27(1):1–14.
29. Manderson AP, Kay JG, Hammond LA, Brown DL, Stow JL (2007) Subcompartments of
the macrophage recycling endosome direct the differential secretion of IL-6 and
TNFalpha. J Cell Biol 178(1):57–69.
| www.pnas.org/cgi/doi/10.1073/pnas.1108579108 Wykes et al.