Dynamic imaging of the effector immune response to listeria infection in vivo.

Abstract

Host defense against the intracellular pathogen Listeria monocytogenes (Lm) requires innate and adaptive immunity. Here, we directly imaged immune cell dynamics at Lm foci established by dendritic cells in the subcapsular red pulp (scDC) using intravital microscopy. Blood borne Lm rapidly associated with scDC. Myelomonocytic cells (MMC) swarmed around non-motile scDC forming foci from which blood flow was excluded. The depletion of scDC after foci were established resulted in a 10-fold reduction in viable Lm, while graded depletion of MMC resulted in 30-1000 fold increase in viable Lm in foci with enhanced blood flow. Effector CD8+ T cells at sites of infection displayed a two-tiered reduction in motility with antigen independent and antigen dependent components, including stable interactions with infected and non-infected scDC. Thus, swarming MMC contribute to control of Lm prior to development of T cell immunity by direct killing and sequestration from blood flow, while scDC appear to promote Lm survival while preferentially interacting with CD8+ T cells in effector sites.

Full text

Page 1

Dynamic Imaging of the Effector Immune Response to
Listeria Infection In Vivo
Janelle C. Waite1, Ingrid Leiner2, Peter Lauer3, Chris S. Rae4, Gaetan Barbet1, Huan Zheng5, Daniel A.
Portnoy4,6, Eric G. Pamer2, Michael L. Dustin1*
1Program in Molecular Pathogenesis, Helen L. and Martin S. Kimmel Center for Biology and Medicine of the Skirball Institute of Biomolecular Medicine, New York
University School of Medicine, New York City, New York, United States of America, 2Infectious Disease Service, Department of Medicine, Memorial Sloan-Kettering Cancer
Center, Immunology Program, Sloan-Kettering Institute, New York City, New York, United States of America, 3Aduro BioTech, Berkeley, California, United States of
America, 4Department of Molecular and Cell Biology, University of California, Berkeley, Berkeley, California, United States of America, 5Department of Chemical
Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts, United States of America, 6School of Public Health, University of California, Berkeley,
Berkeley, California, United State of America
Abstract
Host defense against the intracellular pathogen Listeria monocytogenes (Lm) requires innate and adaptive immunity. Here,
we directly imaged immune cell dynamics at Lm foci established by dendritic cells in the subcapsular red pulp (scDC) using
intravital microscopy. Blood borne Lm rapidly associated with scDC. Myelomonocytic cells (MMC) swarmed around non-
motile scDC forming foci from which blood flow was excluded. The depletion of scDC after foci were established resulted in
a 10-fold reduction in viable Lm, while graded depletion of MMC resulted in 30–1000 fold increase in viable Lm in foci with
enhanced blood flow. Effector CD8+T cells at sites of infection displayed a two-tiered reduction in motility with antigen
independent and antigen dependent components, including stable interactions with infected and non-infected scDC. Thus,
swarming MMC contribute to control of Lm prior to development of T cell immunity by direct killing and sequestration from
blood flow, while scDC appear to promote Lm survival while preferentially interacting with CD8+T cells in effector sites.
Citation: Waite JC, Leiner I, Lauer P, Rae CS, Barbet G, et al. (2011) Dynamic Imaging of the Effector Immune Response to Listeria Infection In Vivo. PLoS
Pathog 7(3): e1001326. doi:10.1371/journal.ppat.1001326
Editor: David S. Schneider, Stanford University, United States of America
Received August 4, 2010; Accepted February 25, 2011; Published March 24, 2011
Copyright: ? 2011 Waite et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the National Institute of Health (Grant P01AI-071195). The funders had no role in study design, data collection and analysis,
decision to publish, or preparation of the manuscript.
Competing Interests: The authors have declared that no competing interests exist.
* E-mail: michael.dustin@med.nyu.edu
Introduction
Appropriate host immune response to pathogenic invasion is
critical for survival. Secondary lymphoid tissues provide a
structural context for multiple layers of the innate and adaptive
immune response to infection. The spleen is a physiologically
relevant tissue for immune responses to blood borne pathogens.
The immune response to systemic Listeria monocytogenes (Lm)
infection has been studied extensively in mice and is focused in
the spleen and liver [1]. In this model, the innate immune system is
responsible for detecting and containing infection while adaptive
immunity is required for clearance of Lm and enhanced protection
against future infections (memory) [2]. Lm is a gram positive
intracellular bacterium and can cause severe infection in immune
compromised individuals [3]. Lm expresses several virulence
factors that enable invasion of the cytoplasm and movement from
cell to cell through a contact dependent mechanism and thus
grows in foci established by infection of one cell [4].
The spleen acts as a blood filter with a network of phagocytic
cells in the marginal zone (MZ) and red pulp (RP) that are in direct
contact with 5% of the the cardiac output. CD11c+dendritic cells
(DC) in the MZ and RP of the spleen sequester Lm from the blood
and are required to initiate infection in the spleen [5,6,7]. At the
site of infection DC orchestrate both recruitment and activation of
innate effectors by secretion of inflammatory cytokines [8].
Aggregations of immune cells including myelomomonocytic cells
(MMC) consisting of neutrophils and inflammatory monocyte
subsets at foci of Lm growth are required to restrict bacterial
growth and contain infected cells [9,10,11]. The process through
which MMC converge on foci in the spleen is not understood. For
example, expression of chemokine receptor CCR2 is required for
monocyte egress from the bone marrow, but is dispensable for
homing to infection sites once in the blood [12].
Although the innate immune system can restrict Lm infection,
cells of the adaptive system, particularly CD8+T cells, are
required for sterilizing immunity [13]. DC bearing Lm antigens
migrate from the MZ to the white pulp (WP) in a pertussis toxin
sensitive process where they present antigen to T cells, which is
required to prime adaptive immune responses [14,15,16]. DC that
prime CD8+T cells may be directly infected or acquire Lm
antigens from infected apoptotic cells such as neutrophils
[17,18,19]. After activation, CD8+T cells proliferate extensively
and exit the WP for bacterial clearance in the RP and to gain
access to other sites of infection through the blood [20]. The
mechanism of clearance is not completely understood, but requires
perforin, IFNc, TNFa and CCL3 [21,22]. Direct visualization of
this process may provide insight as in vivo cytotoxicity has been
associated with stable and prolonged interactions of antigen
specific cytotoxic T lymphocytes (CTL) and target cells [23,24].
Tissue DC also play a role in the effector phase of T cell responses
PLoS Pathogens | www.plospathogens.org1 March 2011 | Volume 7 | Issue 3 | e1001326

Page 2

providing local signals for cytokine production [25,26,27,28].
However, the role of tissue DC in clearance of Lm and nature of T-
DC interactions in RP foci is unknown.
Live imaging of lymphoid structures by intravital microscopy
has provided high-resolution information on the dynamic
interactions that take place during immune responses in situ.
Tracking of cells in real time reveals kinetic information that is
lacking from static images. Intravital microscopy has been useful in
understanding tissue specific responses to pathogens [29]. These
studies have revealed active environmental sampling by resident
DC networks that act as sentinals in peripheral organs such as skin
and intestines during infection by Salmonella in the gut and
protozoan parasites in the skin [30,31,32,33]. In addition,
patrolling neutrophils and monocytes rapidly respond to invading
Leishmania major or tissue injury [34,35,36]. The dynamics of these
cells at inflammatory lesions in zebrafish models of tuberculosis
have yielded unexpected role of macrophages in dissemination of
infection [37]. In mouse models of Leishmania donovani and Bacillus
Calmette-Gue ´rin (BCG) infection, T cells are rapidly recruited to
liver granulomas by inflammatory signals and antigen specific cells
are retained [38,39]. Interactions of effector T cells with parasites
and resident antigen presenting cells (APC) revealed zones of
antigen specific contacts with pathogen associated APC, while
some infected cells were not contacted by T cells in the brain and
in the skin [40,41,42]. The mode of interaction of Lm specific
effector T cells in infectious foci is unknown and cannot be
predicted based on existing data.
Multiphoton intravital imaging in the spleen is hindered by light
absorption and scattering by red blood cells and auto-fluorescent
metabolites concentrated in the RP. This interference is avoided
by directly imaging WP of spleen fragments ex vivo [43,44].
However with these preparations physiological circulation and
recruitment of cells from hematopoietic tissues such as the bone
marrow are lost. We have found that intravital imaging in the RP
with confocal microscopy yields high-resolution images suitable for
analyzing cellular dynamics. We have previously reported the
presence of an extensive DC network in the subcapsular RP
(scDC) that can be directly visualized by intravital microscopy of
the spleen [45]. Early after infection, macrophages and DC
capture the bulk of Lm in the MZ and then migrate to the WP to
present Lm antigens to activate T cells [5,6,16,46]. However, a
minority of Lm does access the RP and establish foci of bacterial
growth there. In addition, T cells exit the WP after activation to
clear infected cells in the RP [20]. The role of the scDC network
and immune cell dynamics in the response to Lm infection in the
RP is unknown. Here we set out to observe the early events in the
innate immune response and later clearance of Lm infection by
adaptive T cells using in vivo imaging of the RP. We show that
scDC interacted with Lm within 2 minutes after introduction into
the blood stream, but at 48–72 hours post infection (p.i.) did not
directly contribute to the innate control of bacterial growth once
Lm foci are established. At 48 hours p.i., MMC converged on foci
by directed migration and restricted blood flow around infected
cells. On day 5 p.i., the peak of Lm growth, CD8+T cells are
recruited to these foci and displayed a two-tiered deceleration with
antigen independent and antigen dependent components.
Results
Systemic Lm is rapidly associated with subcapsular DC in
the spleen red pulp
To test if scDC take up Lm, we used intravital microscopy in the
spleen subcapsular RP to image the arrival of fluorescently labeled
Lm in circulation of CD11c-EYFP mice [47]. Frozen spleen
sections from CD11c-EYFP mice show that YFP+cells are present
in the RP and these cells are heterogeneous for surface staining of
CD11c, MHC class II, F4/80 and CD11b (Figure S1 in Text S1).
In the subcapsular RP, YFP+cells stained strongly for F4/80, a
marker for RP macrophages but also expressed by a subset of DC
isolated from spleen and DC subsets found in peripheral organs
such as skin [25,48]. Most of these subcapsular YFP+cells also
displayed dendritic morphology and time-lapse images of these
cells show they are largely non-motile but are actively extending
and retracting their dendrites (Figure 1A, Video S1). This activity
is similar to other reports of environmental sampling by DC
networks in the lymph node [47]. Thus, it seems that a majority of
these YFP+cells in the subcapsular RP represent a subset of
peripheral tissue DC.
During image acquisition, 107Bodipy-630 labeled Lm were
injected into the retro-orbital plexus (Video S1). In some
experiments, Lm were injected together with a 10 kDa rhodamine
dextran, to mark the time of injection. Lm arrived in the RP within
30 seconds of the rhodamine dextran arrival which was about
60 seconds post injection (data not shown). Of the Lm that came
through the imaging field, 70+7.7% (n=6 mice, 135 bacteria,)
were associated with non-motile scDC 10 minutes after injection
(Figure 1B.). Lm-associated scDC remained non-motile for the
duration of image acquisition (up to 20 minutes) and continued to
actively probe the environment (Video S1). High-resolution z-stack
images show that Lm are primarily located within or at the
periphery of scDC (Figure 1C). Lm were also co-localized with
scDC in images of fixed spleen sections (Figure S1 in Text S1).
The peripheral association may be due to the location of early
phagocytic compartments, Lm attaching to or within dendrites
below the limit of detection or via interactions with non-
fluorescent cells containing Lm. By tracking Lm in the spleen in
vivo, we found Lm speeds were initially high in the first few minutes
post injection representing bacteria in flow however decelerated as
they came in contact with DC (Figure 1D–E). The number of
Author Summary
The pathogenic bacteria Listeria monocytogenes (Lm) can
access the intracellular space and establish infection in the
host spleen and liver. Innate immune cells are required to
detect and contain infection while specific adaptive T cells
are activated and recruited to infection sites. Adaptive
immunity enhances the function of innate cells and also
directly kills infected cells to prevent further bacterial
propagation. These cells respond to local environmental
cues and often communicate via direct cell-to-cell contact.
Here, we observed cellular dynamics in situ by imaging live
tissue using intravital microscopy. We reconstituted the Lm
infection model with bacteria and mouse strains engi-
neered to express fluorescent proteins to label specfic
cells. We find that dendritic cells (DC) in the spleen rapidly
capture Lm from circulation, demonstrating a direct role of
these cells in sequestering bacteria. Neutrophils converge
on infected cells by directed migration to generate
exanguinous foci within the otherwise blood rich red
pulp. Finally, antigen specific CD8+
extensively with infected cells and DC in the foci and are
selectively retained compared to non-specific polyclonal T
cells. These findings significantly advance our understand-
ing of immune cell dynamics in the spleen in both the
steady state and in response to Listeria infection. These
studies could aid in development of disease therapies and
vaccine strategies targeted towards enhancing DC, mye-
lomonocytic and CD8+T cell effector responses.
T cell interact
Intravital Imaging of Listeria Infection
PLoS Pathogens | www.plospathogens.org2 March 2011 | Volume 7 | Issue 3 | e1001326

Page 3

Figure 1. Systemic Lm rapidly associates with sinusoidal DC in the spleen RP. 107Bodipy-630 labeled Lm were injected i.v. while acquiring
images in the spleen of CD11c-EYFP mice. A. Intravital snapshots of labeled Lm (red) and YFP+DC (green) just before and 10 minutes after Lm
injection. White boxes indicate Lm associated with DC. Scale bar=50 mm. B. Frequency of Lm associated with YFP+cells, YFP2cells or in blood flow
(in flow) within the imaging field. C. Intravital Z-stack images taken at 1 mm steps through Lm associated with DC shown in A (white boxes). Scale
bar=17 mm. D. Intravital images of CD11c-EYFP and Lm at the indicated times post injection (minutes). Scale bar=10 mm. E. Mean speed (mm/min) of
bacteria entering the field over time (minutes). F. Number of bacteria arriving in the imaging field over time (minutes). G. Instantaneous speeds of
bacteria over time (minutes). Time is relative to the Lm injection. Data represents the mean from 135 bacteria from 6 mice. Error bars represent SEM.
doi:10.1371/journal.ppat.1001326.g001
Intravital Imaging of Listeria Infection
PLoS Pathogens | www.plospathogens.org3March 2011 | Volume 7 | Issue 3 | e1001326

Page 4

bacteria arriving in the spleen decreased to less than 1 per field
after 2 minutes suggesting they are rapidly taken out of circulation
(Figure 1F). The speeds of Lm that arrived at later time points were
slower than those in flow suggesting they were within a slow
moving non-fluorescent cell (Figure 1G).
Patrolling neutrophils scavenge extracellular Lm
To test if these slow moving Lm were within neutrophils, we
repeated the above experiments with LysM-EGFP knock-in mice
[49]. In uninfected mice, neutrophils and monocytes express high
and intermediate levels of EGFP, respectively (Figure S2 in Text
S1). In the subcapsular RP of uninfected spleens, EGFPhigh
neutrophils were motile in patrolling fashion (Figure 2A–B, Video
S2). Upon injection of 107Lm, there was no acute change in
crawling speed (15 minutes) or at 2 hours post injection (Figure 2C,
Video S2). At 2 hours, neutrophils were significantly more
confined as indicated by a lower straightness ratio and displace-
ment rate (p,0.0001, Figure 2D–E). A very small fraction of
neutrophils (for example 4 out of over 200 in the imaging field)
took up Lm (Figure 2, LM+, Video S2). The Lm containing
neutrophils crawled with an average speed similar to Lm negative
neutrophils in the same field (Figure 2C). The number of
neutrophils in the field increased acutely after Lm injection
(Figure 2F) and increased in frequency of total splenocytes
proportionately to the number of Lm injected suggesting
recruitment from the bone marrow (Figure S3 in Text S1).
During the course of Lm infection the expression level of EGFP
in LysM-EGFP mice shifts such that neutrophils and monocytes
express similar levels (Figure S2 in Text S1). Thus, we are not able
to distinguish neutrophils and inflammatory monocytes in infected
mice and will refer to all EGFP positive cells as myelomonocytes
(MMCs).
DC and MMC dynamics at the site of foci of Lm growth
For long-term visualization of bacteria, we generated Lm that
express TagRFP from the actA promoter (Lm-RFP), which is only
expressed upon entry into the cytosol (Materials and Methods). To
visualize DC and MMC dynamics at the site of Lm growth (Lm
foci), we crossed CD11c-EYFP to LysM-EGFP transgenic mice.
Prior to Lm injection LysM-EGFP+cells crawled through the RP
and displayed transient interactions with scDC (Video S3). Upon
Lm-RFP infection, MMC accumulated at sites of Lm-RFP+foci and
accumulation increased from 24 to 48 hours (Figure 3A–B).
Interestingly, Lm-RFP were detected primarily in DC at 24 hours
Figure 2. A small fraction of patrolling neutrophils take up Lm. 107Bodipy-630 labeled Lm were injected i.v. while acquiring images in the
spleen of LysM-EGFP mice. A. Intravital snapshots of LysM-EGFP (green) and labeled Lm (red). Scale bar=25 mm. B. Snapshots overlaid with neutrophil
tracks (white) and Lm+ neutrophil tracks (blue). Scale bar=100 mm. C–E. Motility parameters before (Pre-LM), acutely after (0–15 min), and 2 hours
after (2–2.5 H) Lm injection separating out neutrophils that took up Lm at 2 hours post injection (LM+). Each dot represents a single cell in the
imaging field. Data are from 1 mouse representative of 3 separate experiments. C. Mean Speed (mm/min). D. Straightness Ratio (maximum
displacement/track length). E. Displacement Rate (Disp. Rate, mm/min). F. Number of neutrophils per field before (pre LM) and at 15 minutes after Lm
injection (15 min). Data represents the mean from 3 mice (n=3). Error bars represent SEM.
doi:10.1371/journal.ppat.1001326.g002
Intravital Imaging of Listeria Infection
PLoS Pathogens | www.plospathogens.org4 March 2011 | Volume 7 | Issue 3 | e1001326

Page 5

but spread to neighboring cells, including MMC, at 48 hours p.i.
(Figure 3C–D).
Although we did not directly observe cell to cell spread, Lm-RFP
were detectable at the tips of EYFP+
‘‘listeriopods’’, a mechanism by which Lm spread to neighboring
cells [50] (Figure S4 in Text S1, Video S4). We also noted fine
EYFP+tubules between well-separated scDC (Figure S4 in Text
S1, Video S4) that may represent the in vivo counterparts of
membrane nanotubes [51]. Infected scDC (and non-DC) con-
tained multiple Lm (Figure S4 in Text S1, Video S4). Most scDC
in the field were enlarged and contained large vacuoles consistent
with activation. Several scDC were clustered together with
trapped or internalized MMCs (Figure S4 in Text S1).
DC extensions or
MMC aggregation at Lm foci is mediated by directed
motility
LysM-EGFP+MMCs that accumulate on Lm-RFP foci were
CD11b+and Gr-1+(Figure S5 in Text S1). To observe the
mechanism by which MMC accumulate at the site of infection,
intravital microscopy was initiated in LysM-EGFP mice infected
with Lm-RFP 48 hours prior. MMCs crawled from the periphery
(up to 300 mm from the foci center) toward established foci by
directed migration, as illustrated by tracks of moving cells
(Figure 4A, Video S5). The biased trajectories of MMCs are
shown by plotting the change in distance to the foci center against
the mean speed for each cell (Figure 4B). Angles (Y) between
MMC trajectories and the foci center [52] were more frequently
less than 90u indicating directed migration towards the focus while
no bias in migration within the RP is observed in the absence of
infection (p,0.0001, Figure 4C). Overall turning angles were not
different compared to cells crawling in uninfected mice however
cells moved slightly faster (Figure 4D–E). The signal to join an
established focus seemed dominant over signals from some
infected cells (Video S5). MMCs within the foci were quite
dynamic, but confined within the region of the foci (Video S5).
MMC protect the host from bacterial spread and tissue
damage in the spleen
The Lm foci observed here are in the blood filled space of the RP.
To test if MMC accumulation affects blood flow around infected
cells, we selectively depleted Gr-1hineutrophils alone or neutrophils
and inflammatory monocytes with 125 or 250 mg of anti-Gr-1
Figure 3. Foci of Lm initiate at DC and spread to aggregated MMC. A. Snapshots of intravital images from CD11c-EYFP and LysM-EGFP
double transgenic mice at the indicated times p.i. with 2.56104Lm-RFP. A. Merged images from Lm-RFP (red), CD11c-EYFP (green) and LysM-EGFP
(blue). Scale bar=50 mm. B. Fraction of MMC or DC per field (number of positive pixels/total pixels) at the indicated time p.i. C. 26 higher
magnification of snapshots in A with single color panels. Scale bar=25 mm. D. Frequency of co-localized fluorescence of Lm-RFP and the indicated
cell type. Data is from 3 mice (n=3). Error bars represent SEM.
doi:10.1371/journal.ppat.1001326.g003
Intravital Imaging of Listeria Infection
PLoS Pathogens | www.plospathogens.org5March 2011 | Volume 7 | Issue 3 | e1001326