Immunity

Perspective

Dying to Replicate: The Orchestrationof the Viral Life Cycle, Cell DeathPathways, and Immunity

Nader Yatim1 and Matthew L. Albert1,*1Unite d’Immunologie des Cellules Dendritiques, Departement d’Immunologie, Institut Pasteur and Inserm U818, Paris France*Correspondence: albertm@pasteur.frDOI 10.1016/j.immuni.2011.10.010

Manipulation of cell death pathways has been identified as a common feature of host-microbe interactions.We examine two examples: influenza A as a representative acute infection and cytomegalovirus as anexample of chronic infection. From the perspective of viral entry, replication, and transmission, we identifypoints of interconnection with the host response to infection, namely the induction of host cell death, inflam-mation, and immunity. Following from this analysis, we argue that the evolution and fine-tuned regulation ofdeath-associated genes may result from constant microbial pressure—past and present—that helped tosupport and coordinate cell death programs within the host. Interestingly, the delay in host cell death allowstime for the virus to replicate while perturbations in cell death allow the host cell to initiate an immuneresponse. This may represent a genetically encoded trade-off ensuring survival of both host and virus, orit may be a part of the complex agenda of infectious microbes.

PrecisThe interconnected pathways of programmed cell death are inti-

mately linked to the emergence of eukaryotic cells as well as the

evolution of multicellular organisms. Mitochondria, which play

a key role in programmed cell death in higher-order organisms,

derived from an endosymbiotic event, and the apparent ‘‘altru-

istic’’ suicide of cells for the good of the colony has been

observed in social unicellular organisms (e.g., Dictyostelium

discoideum). These events are punctuated by the emergence

of programmed cell death pathways and provide insight into

a cell-cell detente that predicates the intrinsic competition for

resources in a multicellular organism (Ameisen, 2004).

Similarly, we argue that host cell-microbe interactions may be

defined through their interdependent regulation of cell death

pathways. This may occur at the level of the infected cell, in

which for a finite period of time the fate of the microbe and the

host cell are joined. Microbe-induced host cell survival may

help to enhance its replication and dissemination within the

organism. Alternatively, the microbe may require the initiation

of cell death pathways in order to escape from the infected

cell. As a counterpoint, the host cell, although inherently inter-

ested in living, may respond to infection by activating pro-

grammed cell death, presumably aimed at clearing the patho-

genic microorganism. Dying cells may also impact host

immunologic responses. As such, understanding the regulation

of cell death pathways, induced either by host cell or microbe,

may require a deeper understanding of how noninfected cells

behave in response to their exposure to, and/or engulfment of,

infection-induced dying cells.

Herein, we examine the orchestration of cell death pathways

from the perspective of viral pathogenesis, which we identify

as the guiding force for host- microbe evolution: inducing and

modifying stress responses at the host cellular level, as well as

influencing downstream effects on the organism. Two detailed

paradigms are offered in order to examine different niches that

478 Immunity 35, October 28, 2011 ª2011 Elsevier Inc.

have been occupied by viruses as their mechanisms of infection,

replication, and transmission intersect with cell death pathways:

(1) Acute viral infection, as represented by influenza A, will illus-

trate how different phases of infection demand opposing pres-

sures on apoptotic pathways in order to achieve efficient replica-

tion. Additionally, the modulation of autophagy and the crosstalk

between autophagic and apoptotic pathways are discussed in

light of their influence on host immune responses. (2) Chronic

viral infection requires distinct relationships with cell death

programs and we utilize cytomegalovirus (CMV) infection to

highlight the mechanisms by which some viruses have evolved

inhibitors of intersecting caspase-dependent apoptosis and

caspase-independent programmed cell death pathways. We

examine the emerging understanding of how cell death and

innate responses are integrated and review the balance that is

established between the virus and adaptive immunity during

latency. Although brief introductions are provided, we assume

prior knowledge of the molecular pathways of cell death (Green,

2011, in this issue of Immunity).

Starting Assumptions

1. Host responses drive the selection of microbial variants.

By this we mean that microbes, intrinsic to their replicative

capacity and high rate of mutagenesis, possess the

genetic diversity to establish themselves within their

respective host. Achieving this includes securing their

niche through the avoidance of host defense; alternatively,

many organisms have adopted the strategy of co-opting

host stress pathways in order to support their replication

and transmission.

2. Our microbial adversaries have a shared interest in our

survival. Aware of the differential rate of mutagenesis

and evolution, Lederberg (2000) concluded that it is not

a Manichaean struggle between pathogen and host,

but instead a process of domestication to achieve

Immunity

Perspective

ametastable equilibrium. From themicrobe’s perspective,

‘‘a dead host is a dead end.’’ Lederberg continued by

arguing that ‘‘most successful parasites travel a middle

path. It helps for them to have aggressive means of

entering the body. but once established they also do

themselves (and their hosts) well bymoderating virulence’’

(Lederberg, 2000).

3. The evolutionary events in microbial genomes intersect

with programmed cell death pathways. By programmed

we mean active cell death as opposed to passive cell

death (Green, 2011). In positing this assumption, we aim

to integrate cell death into an evolutionary genetics

perspective (Barreiro and Quintana-Murci, 2010). In sharp

contrast to microbes, evolution of the host is slow. None-

theless, there exist evidence for microbes selecting for

genetically determined phenotypes—as first suggested

by Haldane, malaria-causing plasmodium species are

responsible for the prevalence of hemoglobinopathies in

African populations (Lederberg, 1999). We argue that in

formulating an association between cell death and immu-

nity, it will become evident that microbes have exerted

selective pressure on genetically encoded mechanisms

of host cell death.

A Viral Perspective to Cell DeathThemicrobes’ agenda is assumed: transmission, infection, repli-

cation, and the readiness for future transmission. From an evolu-

tionary perspective, our agenda (as host) is the same. We

discuss two relationships, using representative microbial agents

responsible for acute and chronic infection in humans.

Acute Virus Infection—Influenza AInfluenza A virus (IVA) is a member of the Orthomyxoviridae

family; it is an enveloped virus with eight minus-sense RNA

segments that encode 11 viral proteins. It causes an acute infec-

tion in humans and is primarily transmitted fromperson to person

via exposure to micron-sized, virus-laden droplets that emerge

from the nose or throat of an infected person. Tropism is princi-

pally the bronchial epithelium (and for more virulent strains may

include the lower respiratory tract) and also includes immune

cells, with the latter having important implications for the host

response. Several phases of the viral life cycle can be delineated:

entry, replication, release, and transmission. Relevant to this

discussion is the fact that human influenza results in a complex

cytopathic effect (CPE) that is intertwined with the viral life cycle

and impacts innate and adaptive immune responses. Strikingly,

severe clinical disease is associated with an exaggerated host

response, including greater cellular infiltration into the infected

lung parenchyma (Kamps et al., 2006).

Inhibition of Apoptosis by IVA

As a result of viral entry and exposure to the viral genome,

several host pathways may be triggered. At the structural level,

this includes perturbation of the actin cytoskeleton, as well as

modulation of endocytic pathways by the IVA matrix protein 2

(M2) ion channel. Host sensors may also signal ‘‘danger,’’ espe-

cially once viral genomes begin to accumulate within the in-

fected cell. Depletion of nutrients and amino acids from the

intracellular environs and overproduction of viral proteins

provide additional triggers for cell stress. All of these host

responses may suffice as inducers of programmed cell death.

Specific examples of entry and viral replication-induced death

include the activation of latent TGF-b, which results in the induc-

tion of endoplasmic reticulum (ER) stress and intrinsic apoptotic

cell death (Carlson et al., 2010; Schultz-Cherry and Hinshaw,

1996), and engagement of the intracellular host sensor RIG-I

by newly synthesized viral genomes, triggering type I interferon

(IFN) and in turn proapoptotic IFN-stimulated genes (ISGs) such

as protein kinase R (PKR) or tumor necrosis factor-related

apoptosis-inducing ligand (TRAIL) (Herold et al., 2008). Although

the first newly produced viral particles may be detected within

5 hr after infection, the virus requires 10–12 hr in order to achieve

stable infection of the host. For these reasons, it is argued that

influenza actively interferes with apoptosis during the early

phase of infection.

Several lines of study have helped to establish a key role for

IVA nonstructural protein 1 (NS1) as an inhibitor of early

apoptosis. NS1 directly activates the phosphatidyl innositol

30OH kinase-(PI3K)-Akt signaling pathway (Ehrhardt et al.,

2007a; Ehrhardt et al., 2007b) (Figure 1), thereby inhibiting pro-

apoptotic factors such as caspase-3, caspase-9, and glycogen

synthase kinase-3B (GSK-3B), as well as stimulating antiapop-

totic Mdm2, X-linked inhibitor of apoptosis (XIAP), and NF-kB

signaling (Downward, 2004; Ehrhardt and Ludwig, 2009). NS1

is also capable of silencing RIG-I (Mibayashi et al., 2007) and

directly inhibiting PKR activity (Bergmann et al., 2000). By dimin-

ishing the induction of IFN production, NS1 may also impact the

kinetics of TRAIL-mediated apoptosis.

Induction of Apoptotis by IVA

Although cell survival may favor infection by providing additional

time for replication and assembly, it has been demonstrated that

caspase-3 activity is necessary for the release of viral ribonu-

cleoprotein (RNP) complexes from the nucleus and maturation

of haemagglutinin (HA) (Wurzer et al., 2003). Strikingly, it has

been demonstrated that inhibition of caspase-3 impairs IVA

propagation (Wurzer et al., 2003). Similarly, it has been reported

that inhibition of the proapoptotic effector Bax or overexpression

of the antiapoptotic Bcl-2 protein results in decreased viral yields

(McLean et al., 2009; Nencioni et al., 2009). In fact, IVA seems to

have evolved to delay apoptosis by inhibiting premature cell

death mediated by host cell stress responses, but actively

contributes to the induction of late-phase cell death (Zhirnov

and Klenk, 2007). This dichotomous effect of IVA on apoptosis

suggests that early induction of cell death may be a protective

response favoring viral clearance, whereas during the later

stages of the viral life cycle, cell death favors virus replication.

Regarding IVA proteins that directly articulate with cell death

effector proteins, one important example is PB1-F2, a recently

identified gene product that is encoded within an alternative

reading frame of the IVA polymerase gene, PB1(Chen et al.,

2001a). PB1-F2 contains a C-terminal mitochondrial targeting

sequence (MTS) and creates pores in planar lipid membranes,

which suggested a role in regulating intrinsic apoptotic cell death

(Chanturiya et al., 2004). Subsequent work has shown that

expression of PB1-F2 in the late phase of influenza A infection

(5–12 hr postinfection) contributes to IVA virulence in various

ways (McAuley et al., 2010). Supporting this argument, it has

been demonstrated that genetically ablating PB1-F2 expression

Immunity 35, October 28, 2011 ª2011 Elsevier Inc. 479

Cell survival

Early phase

Late phase

NS1Intrinsic sensors

Viral cycle

PI3K-Akt

Autolysosome

PB1-F2M2

Viral cycle

Inhibition of autophagosomematuration

Cell death

Perturbation of cytoskeletonER stressNutrient depletion

Viral entry

Autophagicflux

Autophagosome

Activation of Bak and BaxInduction of antiviral responseUpregulation of death receptors

Intrinsic & extrinsic sensors

Time

BaxVDACANT1

MOMP

?

Maturation and release of virus

IVA

Figure 1. Virus Entry and Replication IsInterconnected with Cell survival and CellDeath PathwaysAfter influenza A (IVA) entry, intrinsic cell stresstriggers host cell death pathways. An early phaseof viral protein expression results in a counterbal-ance to the activation of apoptosis. NS1 expres-sion activates PI3kAkt pathway, leading to survivalsignals and the coincident downregulation ofproapoptotic factors (e.g., Caspase-3). Addition-ally, autophagy enhances cell survival through theremoval of damaged cellular components andthrough the direct inhibition of cell death. Duringlate phase of IVA infection, M2-mediated inhibitionof autophagosome maturation results in abortiveautophagy and increased cellular stress dueto accumulation of damaged organelles andunfolded proteins. Moreover, PB1-F2 expressiontriggers mitochondrial outer membrane per-meabilization (MOMP) via a still disputed mecha-nism that may involve Bax, VDAC, and/or ANT1.This combination of events results in caspaseactivation, which appears to facilitate maturationand release of mature virions.

Immunity

Perspective

(DF2) attenuated the virus and limited disease severity in exper-

imental models (Zamarin et al., 2006). Interpretation of in vivo

data, however, requires evaluation of many variables as different

cell types (e.g., epithelial cells and immune cells) may be differ-

entially susceptible to PB1-F2-induced cell death. It is interesting

to note that DF2 viruses did not show an altered viral replication

in Madine Darby Canine Kidney (MDCK) cells (Zamarin et al.,

2006), whereas a 50% reduction in monocyte apoptosis was

observed (Coleman, 2007).

Further characterization of the mechanism of cell death induc-

tion by PB1-F2 suggested a role for adenine nucleotide translo-

case 3 (ANT3) and voltage-dependent anion channel 1 (VDAC1)

interactions that resulted in release of cytochrome c (Cyt c) (Za-

marin et al., 2005). More recent results, however, indicate that

PB1-F2 activation of the intrinsic apoptotic pathway is depen-

dant on Bak- and/or Bax-mediated mitochondrial outer mem-

brane permeabilization (MOMP) (McAuley et al., 2010) (Figure 1).

One caveat of the latter study is that cell death assays were per-

formed with 27-mer peptides corresponding to the C-terminal

sequence of PB1-F2 as opposed to whole protein. An additional

confusion in the literature is the diversity of PB1-F2 genes that

exists—some variants possess proapoptotic functions whereas

others do not seem capable of mediating MOMP (Zell et al.,

2007).

480 Immunity 35, October 28, 2011 ª2011 Elsevier Inc.

IVA and the Modulation of

Autophagy Is at the Crossroads of

Cell Survival and Cell Death

Cell stress and cell death pathways over-

lap as they share common triggers. As

a consequence, there has been in-

creasing attention to the crosstalk

between macroautophagic (a cell stress

and cell survival pathway) and apoptotic

pathways; for example, recent evidence

supports the existence of a complex

between the antiapoptotic protein Bcl-2

and the proautophagy regulator Beclin-1,

thus coordinating reciprocal regulation of these two stress

pathways (Kang et al., 2011). We propose that during the early

phase of the viral life cycle, autophagic flux serves to promote

cell survival and inhibit apoptosis. This is supported by the simul-

taneous single-cell analysis of autophagy and apoptosis: func-

tional autophagic pathways delayed IVA-induced caspase-3

activation (de la Calle et al., 2011).

With respect to the late phase of replication, it has been

recently demonstrated that IVA infection results in autophago-

some accumulation as a result of M2-mediated inhibition of

autophagosome-lysosome fusion (Gannage et al., 2009) (Fig-

ure 1). Surprisingly, the mechanism responsible for the M2

blockade of autophagosome maturation is independent of its

proton channel activity and instead results from its modulation

of Beclin-1 (Gannage et al., 2009). Notably, the prolonged

expression of M2 results in abortive autophagy that leads to

the induction of apoptosis and eventually cell shrinkage and

engulfment by phagocytes (de la Calle et al., 2011; Uhl et al.,

2009; and unpublished data, M.L.A. and C. de la Calle). Interest-

ingly, M2 regulation of autophagy does not impact viral propaga-

tion (Gannage et al., 2009), yet additional experimentation may

be required to assess whether this holds for distinct cell types

(e.g., infected hematopoetic cells). In the context of the dichoto-

mous relationship between IVA and apoptotic programs, we

IVA

Autophagy Apoptosis

Lung- Infection of the epithelium

Lymph node- Infection of the dendritic cellsProtective immunity (e.g. NP-specific CD8+ T cell response)Counter-protective immunity (e.g. PA-specific CD8+ T cell response)

IVA

Abortive autophagy Apoptosis

Lung- Infection of the epithelium

Cough, sneeze, saliva - spread of IVA

Lymph node- Dendritic cell capture and cross-presentation of dying cellsEnhanced cross-priming (e.g. NP-specific CD8+ T cell response)

Early phase

Late phase

Transmission phase

Figure 2. Schematic Representation of thePhases of Influenza InfectionEarly phase influenza (IVA) infection is marked byautophagy and inhibition of apoptosis. As a resultof dendritic cell (DC) infection, there is rapid acti-vation of IVA-specific CD8+ T cells, but a subset ofthis response (at least in experimental models)appears to be counterprotective. During late-phase infection, abortive autophagy and directaction of IVA gene products coordinate theinduction of apoptotis. This contributes toenhanced cross-priming, which favors theexpansion of protective CD8+ T cells. We proposethat the enhanced cellular immune response mayparticipate in the transmission phase by provokingprogrammed reflexes such as coughing andsneezing, thus spreading IVA.

Immunity

Perspective

suggest a reciprocal relationship with respect to autophagy

pathways; macro-autophagy, possibly induced by IVA infection,

facilitates cross-inhibition of premature apoptosis, yet at a later

time in the course of infection, M2-mediated inhibition of auto-

phagic flux (i.e., abortive autophagy) facilitates induction of

late-phase apoptosis (de la Calle et al., 2011; Rossman and

Lamb, 2009).

Taken together, these results suggest that influenza virus has

evolved an interconnected and intimate relationship with host

cell death pathways by specifically modulating the timing,

extent, and progression of apoptosis (Figure 2). As such, evalu-

ation of sequence polymorphisms in IVA NS1, PB1-F2, and M2

may be key to further understanding of strain-specific virulence

and disease pathogenesis. Furthermore, there appears to be

cell-type specificities in the balance between autophagy and

apoptosis induction underlying the importance to elucidate this

crosstalk in several cell types, but also in the context of the

host immune system.

Integration of Innate Immunity and Cell Death

Thus far, we have discussed themanner by which the host stress

response engages cell death pathways as a possible constraint

on viral replication; the other major barrier to transmission and

infection is the host immune response. But ‘‘barrier’’ is perhaps

not an accurate metaphor as microbes have identified numerous

strategies to go over, around and under these so-called immuno-

logic barriers. Instead, we consider it a node of integration and

emerging evidence indicates that the viral life cycle, cell stress,

and/or cell death pathways and host immunity are all intercon-

nected. For IVA infection, host sensors include pattern recogni-

tion receptors for single-stranded RNA (ssRNA): Toll-like

receptor (TLR)-7 and the retinoic acid inducible gene-I (RIG-I)

(Diebold et al., 2004; Kato et al., 2006). Both TLR7 and RIG-I

induce the production of proinflammatory cytokines, including

Immunity 35,

type I IFNs. Although not reported for an

IVA model, we highlight that the TLR7

and RIG-I pathways also have in common

their being regulated by autophagic flux.

In the case of TLR7, it was demonstrated

by Iwasaki and colleagues that the

protein encoded by autophagy gene-5

(Atg5) is responsible for cytosolic ssRNA

being externalized, in a topological

sense, exposing ligand to the luminal domain of TLR7 (Lee et al.,

2007) (Figure 3). In contrast, Atg5-Atg12 has been reported to

blunt RIG-I signaling by interfering with its recruitment of the

adaptor protein IPS-1 (Jounai et al., 2007). More direct evidence

has been suggested for RIG-I intersecting with cell death path-

ways: RIG-I recruits, via the adaptor protein IPS-1, the TNFR-

associated death domain (TRADD) protein, thus establishing

a molecular link with apoptotic effectors (Michallet et al., 2008)

(Figure 3). It remains to be determined how IVA-induced auto-

phagy regulates the activation of TLR7 or RIG-I pathways, and

furthermore, it should be explored whether the NS1 protein, or

for that matter the polymerase PB2, differentially modulates

RIG-I-induced IFN versus induction of cell death (Varga et al.,

2011).

In addition to the established host sensors for viral nucleic

acids, three independent studies have shown that IVA activates

the inflammasome, a distinct intracellular pathway that is linked

to inflammation and host immune responses (Thomas et al.,

2009; Allen et al., 2009; Ichinohe et al., 2009). In all studies, IVA

infection of caspase-1-deficient animals failed to induce IL-1b

or IL-18 secretion, resulting in enhanced susceptibility to infec-

tion. Notably, these initial studies do differ in their implication

of NLRP3 as the upstream trigger of inflammasome activation,

possibly a result of differences in the microbial flora present in

the different mouse colonies (Ichinohe et al., 2011).

Although distinct innate sensors seem to be independently

activated, they are in some cases integrated at the cellular and

transcriptional level. In the case of the inflammasome, activation

requires two signals with ‘‘signal 1’’ being mediated by TLR7,

which is both necessary and sufficient to induce transcription of

pro-IL1-b (Ichinohe et al., 2010). Remarkably, it is the IVA M2

protein that provides signal 2, acting this time via a mechanism

dependant on its pH-gated H+ channel activity. Specifically,

October 28, 2011 ª2011 Elsevier Inc. 481

IVA

Autophagosome

Endosome

ssRNA

TLR7

MyD88

IRF7 NF-κB

Antiviral response(e.g. Type I IFN)

Antiviral response

Proinflammatory response(e.g. cytokines, chemokines)

Viral replication

RIG-I

IPS-1

IRF3

TBK-1

IKKε

RIPK1-FADD

RIPK3Caspase 8

Necroptosis

Apoptosis

M2 M2

NLRP3Inflammasome

Caspase 1

IL-1β, IL-18

Pyroptosis

NS1

H+

K+

ROS

?? Bcl2

Pyropto

TLR NLRRLR

Figure 3. Integrating the Effect of Influenza Viral Proteins on Endosomal and Cytosolic Innate Immune SensorsDuring viral replication, single-stranded RNAmay be delivered to the endosome via the autophagy-mediated sequestration of cytosolic material, thus permittingthe engagement of Toll-like receptor 7 (TLR7). TLR7 in turn recruits Myd88, which results in the nuclear translocation of IRF7 and NFkB and the induction of type Iinterferons and proinflammatory cytokines and/or chemokines, respectively. In the case of the latter, this response includes the expression of pro-IL1b and pro-IL18 (signal 1 for inflammasome-mediated inflammation). Viral replication generates 50-ppp-genome that is recognized by RIG-I, an upstream sensor that recruitsIPS-1 to the mitochondrial outer membrane via a CARD-like domain and may inhibit Bcl2 antiapoptotic effects. Activation of IRF3 via TBK1- and IKK 3-dependentphosphorylation results in expression of additional proinflammatory and antiviral host factors. IPS-1 also activates NFkB via FADD- and RIPK1-dependentpathways with possible crosstalk to cell death pathways via caspase-8 and/or RIPK3 possibly leading to apoptosis or necroptosis. Production of M2 viral proteininhibits autophagosomal maturation, thus having a potential effect on TLR7 and/or RIG-I. Additionally, M2 activates the NLRP3 inflammasome via a mechanismdependent on its proton channel activity (signal 2). NLRP3 oligomerization activates caspase-1 via the adaptor protein ASC, resulting in the maturation andsecretion of Il1b and Il18. Caspase-1 may also induce pyroptosis, which has been reported to be inhibited by NS-1 via still unknown mechanisms.

Immunity

Perspective

M2 destabilizes the Golgi apparatus and creates an imbalance in

cation concentration (Ichinohe et al., 2010) (Figure 3). An impor-

tant and still open question is how the inflammasome is being

activated without inducing pyroptosis, a caspase-1 mediated

inflammatory cell death (Lamkanfi and Dixit, 2010). One possi-

bility is that the induction of prosurvival factors, perhaps induced

by the same signaling pathways, safeguard the cell from cas-

pase-1-mediated cell death. Alternatively, there may be new

secrets to IVA NS1 that await discovery. In fact, NS1 has been

reported to modulate inflammasome activation and caspase-1

maturation (Stasakova et al., 2005), however the dissection of

proinflammatory and pyroptotic effectors have yet to be clarified.

Death Is Not an Endpoint but the Beginning

of an Immune Response

As we move from interplay at the cellular level to integration at

the organism level, it has become evident that dying cells are

482 Immunity 35, October 28, 2011 ª2011 Elsevier Inc.

important regulators of proimmune responses (Albert, 2004).

This is especially true when the dying cell harbors an infectious

microbe and/or has died a proinflammatory death (Green et al.,

2009). Prior work revealed a central role for apoptotic cells in

the delivery of antigen to conventional dendritic cells (cDCs) for

cross-presentation and activation of CD8+ T cells. Notably, the

initial studies employed an IVA infection model, after the transfer

of matrix 1 (M1) from infected donor cells to uninfected human

cDCs competent in presenting peptide epitopes derived from

M1 on major histocompatibility class I (MHC I) molecules (Albert

et al., 1998b; Albert et al., 1998a). More recently, it was sug-

gested that autophagic processes within the dying IVA-infected

cells are critical for achieving efficient antigen transfer and cross-

priming (Giodini and Albert, 2010; Uhl et al., 2009). Intriguingly,

M2-mediated inhibition may be central to these biologic path-

ways as inferred from the seminal observation that chloroquine,

Immunity

Perspective

an inhibitor of autophagosome-lysosome fusion, serves to

enhance cross-priming (Accapezzato et al., 2005). Accordingly,

autophagy is not only integrating cell survival and cell death

pathways—it is also central to the coordination of immune

responses. Through the crosstalk between host stress and cell

death pathways, as well as innate programs triggered by IVA,

there is ample antigen and inflammation to support the establish-

ment of adaptive immune responses.

For a more complete perspective of the immune pressures ex-

erted on IVA during its life cycle, a brief interlude is required in

order to consider the cells that are insensitive to IVA-induced

cell death. Notably, mature cDCs, different from monocytes,

macrophages, or immature cDCs, resist the engagement of

apoptosis (Bender et al., 1998). As a consequence, IVA directly

infected mature cDCs provide a second source of proinflamma-

tory cytokines and presented antigen for the priming of T cell

responses. These two modalities for T cell priming (cross-

priming and direct priming) have been noted by Bennink and

colleagues, following the initial observations of the Doherty labo-

ratory. In C57BL/6 mice, infection with IVA/PR8 results in the

initial priming of two codominant populations of antigen-specific

CD8+ T cells: nucleoprotein (NP) and acid phosphatase (PA)-

specific responses (Andreansky et al., 2005). Although the NP

epitope (NP366–374) is broadly presented, due to its generation

being mediated by the conventional proteasome, PA epitope

(PA224–233) presentation is restricted to cells expressing the im-

munoproteasome (Chen et al., 2001b; Rivett and Hearn, 2004).

At first, many would conclude that regardless of the mode for

T cell activation, a robust CD8+ response should support viral

clearance. Surprisingly, this is not always the case. In a vaccine

study, Woodland and colleagues provided clear evidence that

a pre-existing NP366–374-specific CD8+ T cell response favored

IVA clearance by limiting viral proliferation. This was in contrast

to pre-existing PA224–233-specific CD8+ T cell responses, as

well as other subdominant T cell populations, which favored viral

replication via still undefined mechanisms (Crowe et al., 2006).

Integrating this information with knowledge about antigen

presentation pathways, we suggest that mouse adaptation of

the IVA/PR8 strain may have resulted in a ‘‘private conversation’’

between directly infected cDCs and PA224–233-specific CD8+

T cells. Because epithelial cells do not express the immunopro-

teasome during early phases of cellular infection, they do not

process and present the PA224–233 epitope. In contrast, cross-

presentation favors the priming of NP366–374-specific CD8+

T cells (Blachere et al., 2005), which possess the ability to directly

engage an IVA-infected epithelial cell. Again, we find a temporal

relationship that follows the agenda of the microbe. Although we

recognize that peak viral production occurs prior to peak T cell

priming, it is possible that early adaptive immune responses

favor replication and extend the duration of viral shedding by al-

lowing the survival of mature cDCs. In contrast, late responses,

through the delivery of dying cell-associated antigen, help to

establish a so-called ‘‘protective’’ immune response (Figure 2).

Completing the Life Cycle

As discussed, the virus has evolved complex mechanisms to

modulate different forms of cell stress and cell death. Interest-

ingly, we identify a temporal control of apopotosis and auto-

phagy that maps to the different phases of the viral life cycle

and immune response. Likely, this is a pattern that can be found

in other acute infections. This pathophysiologic analysis of IVA

and cell death stimulates speculation that perhaps the establish-

ment of host immunity is also in line with the virus’ agenda for the

population (i.e., transmission). This would extend far beyond the

initial reflections of Lederberg (Lederberg, 2000). IVA may have

evolved to appropriate the adaptive immune responses initiated

by DCs dying cells—T cell influx into the lung may support sus-

tained inflammation that facilitates transmission (Figure 2). In this

view of viral infection, we can perhaps speculate that CD8+

T cells exert a host-induced apoptotic program (i.e., cytolytic

killing of target epithelium), which is initiated by viral antigen

cross-priming—all tuned to support the release of viral particles

into the airway during the transmission phase. It is exactly these

reflexes (cough, cough) that help to ensure exposure and infec-

tion of the next ill-fated host.

Chronic Virus Infection—CytomegalovirusThe Herpesviridae family are a large group of double stranded

DNA (dsDNA) viruses that are divided in three subfamilies: alpha-

herpesviruses (e.g., HSV, VZV), betaherpesviruses (e.g., CMV,

HHV-6,7) and gammaherpesviruses (e.g., EBV, KSHV)(Arvin

et al., 2007). Most vertebrate hosts have been reported to harbor

at least one herpesvirus infection and interestingly, there is strict

species specificity with most transmissions being a result of

either vertical (less common, with the exception of CMV) or hori-

zontal spread. This suggests an extensive history of coevolution

between virus and host, arguably with the virus having evolved to

exploit a unique niche within their susceptible host. With respect

to human herpesviruses, all have in comment that release of

infectious progeny is accompanied by destruction of the infected

cell, referred to as CPE, and all possess the potential to enter

a state of latency within their natural host. Human herpesviruses

are widespread in the population and cause mostly benign acute

infection followed by an imperceptible chronic, latent-phase

infection. Nonetheless, at-risk individuals (e.g., persons with

acquired or congenital immunodeficiencies) may develop severe

disease (Arvin et al., 2007). Among the herpesviruses, cytomeg-

aloviruses (CMVs) have been highly studied as a model of viral

inhibition of cell death and immunity (Brune, 2011; Hengel

et al., 2005). CMV replicates in many differentiated cell types in

the host and its cell cycle is divided into three phases: immediate

early (IE), delayed early (DE), and late (L). The replication cycle is

relatively slow, with a switch from the IE to L stages occurring

after 24 hr. An additional 24 hr is required for the production of

infectious virions and peak viral titers are not reached until

several days postinfection (Arvin et al., 2007). Thus CMV must

regulate cell death, even during the lytic state of infection.

Although we have included data from mouse and human CMV

infections (MCMV and HCMV, respectively), we are aware and

offer as a caveat that because of species specificity, some of

the mechanisms are not shared between these two viruses.

Regulation of Cell Death during Lytic CMV Infection

Cytomegaly, or cellular swelling, is a hallmark feature of CMV

infection. This unique CPE can be first observed as early as

a few hours postinfection and has been attributed to its rapid

activation by CMV of host pathways. NF-kB activation has

been observed within 5min after exposure of cells to virus (Poole

et al., 2006); CMV glycoprotein B (gB) binding can trigger the

activation of phospholipases C and A2 as well as an increase

Immunity 35, October 28, 2011 ª2011 Elsevier Inc. 483

CMV

Intrinsic stressors-DNA damage-ER stress-mitochondrial alterations

BH3-only proteinactivation

Bax Bak

Expression of IE genes

vIBO

vICA

vIRA

vMIA

Death receptor

RIPK1-FADD

RIPK1-RIPK3Caspase 8

Caspase 3

Bid

Apoptosis

Cytochrome C

Necroptosis

Figure 4. Cytomegalovirus Entry Results inthe Reprogramming of Multiple Cell DeathProgramsCellular infection by cytomegalovirus (CMV)results in a rapid activation of inflammation andstress pathways. The latter engages BH3-onlyproteins and establishes a pro-apoptotic pressureon the host cell. Counterbalancing these stimuli,CMV encodes for specific inhibitors for Bax acti-vation, referred to as viral mitochondrial localizedinhibitor of apoptosis (vMIA), and, in the case ofmouse CMV, an viral inhibitor of Bak oligomeriza-tion (vIBO). Inflammation also engages extrinsicpathways for apoptosis. The viral inhibitor of cas-pase activation (vICA) serves to inhibit caspase-8protease activity, but has the consequence ofsimultaneously sensitizing the cell to necroptosis.Hence, a fourth regulator of cell death (to dateidentified in mouse but not human CMV), the viralinhibitor of RIPK activation (vIRA), perturbs RIPK1and RIPK3 interaction via its RHIM domain.

Immunity

Perspective

in arachadonic acid metabolites (Arvin et al., 2007), and in the

context of an infectious process, engagement of TLR9 in nonin-

fected plasmacytoid dendritic cells (pDCs) results in the produc-

tion of type I IFNs (Dalod et al., 2003). To blunt such activation,

CMV has evolved to express immediate early (IE) gene products

that reprogram the host response, including the regulation of

NF-kB, interferon response factor 3 (IRF3) complex formation,

and IFNa/b (IFNAR) receptor signaling (Arvin et al., 2007).

Although successful at rapidly regulating these host programs,

proapoptotic signals are abundant as CMV enters the nucleus,

triggers host cell DNA damage, initiates an abortive cell cycle,

modifies the ER, alters mitochondrial function and activates

PKR (to name a few early events) (reviewed elsewhere, Tait

and Green, 2010). Here, we focus on the viral encoded mecha-

nism that perturbs different death pathways during the lytic

phase of infection.

HCMV and MCMV both encode viral inhibitors of intrinsic and

extrinsic apoptosis during the early phase of infection. In the

case of HCMV, there exist two genes, ul37x1 and ul36: the

former encodes for and will be referred to as the viral mitochon-

drial localized inhibitor of apoptosis (vMIA), and the latter

encodes the viral inhibitor of caspase 8 activation (vICA). In addi-

tion, the major IE gene products IE1p72 and IE2p86 participate in

the inhibition of death receptor-induced apoptosis; however, the

precise mechanisms remain unknown (Brune, 2011). MCMV

encodes several known inhibitors of apoptosis. These include:

(1) a homolog of vICA, the gene product of m36; (2) a positional

and functional analog of vMIA, encoded by the gene product of

m38.5; and (3) a unique regulator of apoptotic cell death en-

coded bym41.1, generating a gene product called viral inhibitor

of Bak oligimerization (vIBO).

Functionally, vMIA expression results in the translocation of

Bax from the cytosol to the mitochondrial membrane, suggest-

ing that vMIA sequesters oligomerized high molecular weight

484 Immunity 35, October 28, 2011 ª2011 Elsevier Inc.

Bax in an inactive conformation, thereby

inhibiting MOMP and subsequent

intrinsic apoptotic cell death (Arnoult

et al., 2004) (Figure 4). As indicated

above, MCMVm38.5 also sequesters oli-

gomerized Bax at the mithochondria, but is not homologous to

the HCMV ul37x1 gene product. Importantly, D38.5 MCMV

shows enhanced viral induced apoptosis (Jurak et al., 2008).

This observation illustrates the nonredundant role for Bax oligo-

merization in the CMV life cycle.

In addition to its role in inhibtion of host-induced apoptosis,

vMIA might have a role in inhibiting CMV-mediated programmed

cell death (cmvPCD), an hCMV-specific caspase-independent

cell death pathway linked to intramitochondrial accumulation

of the pro-death effector protein Omi (also known as HtrA2)

during the late phases of infection (McCormick et al., 2008).

cmvPCD timing seems to be tightly regulated by vMIA expres-

sion, independent of its function in regulating apoptosis. Via

a still unknown mechanism, vMIA delays cmvPCD, which is

believed to extend the period of viral production by infected

cells, potentially increasing transmission. Additionally, the broad

expression of Omi in mammalian cells and the fact that Omi-

overexpressing cells only die when infected suggested that

this pathway may represent a process of viral detente, defining

the timing of cell death in a manner that is in accord with the viral

agenda. This interplay between host and direct viral-induced

cell death pathways is reflective of the complex relationship

described for IVA.

It has recently been characterized that MCMV also encodes

a vIBO. Specifically, m41.1 is a mitochondrial targeted protein

capable of sequestering oligomerized Bak (Cam et al., 2010)

(Figure 4). While Dm41 MCMV has a striking death phenotype,

its effect on viral replication is restricted to macrophage infection

(Brune et al., 2003). This provides an interesting parallel to the

differential effects on cell death by IVA gene products. As an

additional point, it is interesting to note the distinct strategy

that has evolved in betaherpesviruses, which (at least for rodent

CMVs) employ two distinct proteins in order to influence intrinsic

cell death. This is in contrast to gammaherpesviruses as well as

Immunity

Perspective

other viral families that encode viral homologs of Bcl-2, which

double as inhibitors of both Bak and Bax (Lagunoff and Carroll,

2003). Perhaps the distinct mechanisms of regulation in CMVs

permit the virus’s ability to evolutionarily tune its modulation of

intrinsic apoptosis across several host cell types.

In order to modulate extrinsic cell death, betaherpesviruses

encode a conserved vICA. Although its evolutionary conserva-

tion suggests an important role in viral infection and dissemina-

tion, vICA expression is considered dispensable for the HCMV

life cycle and in MCMV shows a growth phenotype restricted

to macrophages (Cicin-Sain et al., 2008; McCormick et al.,

2005; McCormick et al., 2010; Menard et al., 2003; Skaletskaya

et al., 2001). This may indicate that CMV-infected macrophages

are particularly sensitive to death receptor-mediated cell death.

As a result, the survival advantage conferred by vICA expression

seems to act by inhibiting the death of macrophages, a poten-

tially important cell type for dissemination within the infected

host (Ibanez et al., 1991). vICA binds to the prodomain of cas-

pase-8 and prevents its oligomerization and cleavage (Skalet-

skaya et al., 2001) (Figure 4). Although vICA may be considered

mechanistically similar to viral caspase-8 inhibitors (v-FLIP),

CMV has again evolved a distinct molecular solution—vICA

shares no evident homology to death effector domains typical

of this class of apoptosis inhibitors (Skaletskaya et al., 2001).

Recently, it has become evident that in the absence of cas-

pase-8, cells become susceptible to programmed necrosis

(also referred to as necroptosis) (Kaiser et al., 2011; Oberst

et al., 2011). This pathway is mediated by the receptor interact-

ing protein kinase (RIP) homotypic interaction motif (RHIM)-con-

taining signaling adaptors RIP1 and RIP3. Strikingly, Mocarski

and colleagues demonstrated that MCMV, through its expres-

sion of vICA, sensitizes cells to RIP3-mediated necroptosis (Iba-

nez et al., 1991; Upton et al., 2010); however, the virus (high-

lighting its prior knowledge of this pathway) has evolved to

express a viral inhibitor of RIP activation (vIRA), encoded by

m45. MCMVm45 was first described as being required for repli-

cation and cell death inhibition in endothelial cells, with a notable

impact on in vivo viral propagation (Brune et al., 2001). It has now

been demonstrated that vIRA inhibits RIP1 and RIP3 interaction

through its viral RHIM domain (Rebsamen et al., 2009) (Figure 4).

Remarkably, in vivo infection of T and NK cell-deficient mice

with MCMV engineered to express a RHIM-domain mutant

m45 (m45mutRHIM) resulted in a 7-log reduction in salivary

gland viral titers (Upton et al., 2010). Although RIP3 expression

is required in order for MCMV to sensitize cells to RIP3-mediated

necroptosis, it is unclear what cellular infections (e.g., fibro-

blasts, endothelium or macrophages) account for this dramatic

phenotype. Additionally, the processes leading to RIP3 activa-

tion remain unknown. One possibility is via TLR3 or the DNA

sensor AIM2 (discussed below). Notably, these pathways

engage RIP3 and they too are abrogated by vIRA expression

(Mack et al., 2008). What is clear is that the crosstalk between

cell death and proinflammatory pathways is again an area

where we have much to learn from the study of the CMV life

cycle. The observation that macrophages (and other RIP3-

expressing cell types) are particularly sensitive to infection by

several genetically ablated strains (m41.1,m45) puts pronecrotic

adaptors as central regulators of cell fate (Brune et al., 2001;

Mack et al., 2008).

Integration of Innate Immunity and Cell Death

CMV has a long-term interest in the occupation of secure niches

within its host. Given the initial impact on host cell signaling path-

ways and the profound effect on cellular metabolic function, it is

almost self-evident that ‘‘danger’’ signals will be triggered, both

by microbial-associated and damage-associated molecular

patterns (MAMPs and DAMPs, respectively) (Mackey and

McFall, 2006; Seong and Matzinger, 2004). To counterbalance

host responses, CMVs have evolved various strategies,

including the expression of viral cytokines and chemokines,

decoy receptors, enzymatic proteins capable of altering the

function of host immune proteins, modulators of bioactive lipid

and carbohydrate effectors, and inhibitors of innate and adaptive

signaling pathways. Thorough analysis of how CMVs regulate

immune responses can be found elsewhere (Miller-Kittrell and

Sparer, 2009); here, we focus on those pathways for which there

exists evidence of crosstalk with cell death.

TLR1, TLR2, TLR3, TLR9, and DAI (also known as ZBP1) have

been reported as host sensors of CMV infection and replication.

Interestingly, these four sensor pathways offer examples of

direct and indirect crosstalk with cell death pathways. TLR1

and TLR2 heterodimers are engaged by and has been shown

to coimmunoprecipitate with HCMV gB and glycoprotein H

(gH) (Boehme et al., 2006). It has also been reported that

R753N TLR2 polymorphisms confer susceptibility to higher

HCMV load, as reported in a small cohort of liver transplant

recipients (Kijpittayarit et al., 2007). Moreover, experimental

evidence in mice indicate that TLR2-deficient strains have higher

liver and splenic MCMV titers, which could be explained by

diminished natural killer (NK) cell activation (Szomolanyi-Tsuda

et al., 2006). Although it has yet to be firmly established in amodel

of CMV infection, it has been suggested that apoptotic pathways

may be directly induced as a result of TLR2 activation. For

example, there exists evidence for extrinsic apoptosis, demon-

strated by the in vitro exposure of HCMV to syncytiotropho-

blasts, which produce TNF-a as a result of TLR2 activation,

inducing apoptosis in neighboring cells (Tomkinson et al., 2010).

Less is known about the mechanisms by which CMV compo-

nents (presumably nucleic acids) engage TLR3 and TLR9. TLR3

is engaged by dsRNA, and it has been suggested that such

ligands might be generated as a consequence of bidirectional

transcription during the CMV life cycle (Tabeta et al., 2004). Still

unknown is how such dsRNA would gain access to the ligand

binding domain of TLR3. It has been shown that Tlr3�/� mice

are hypersusceptible to MCMV infection (Tabeta et al., 2004).

This mechanism requires additional study, however, given that

this phenotype seems dependant of the titer of viral inoculation

(Edelmann et al., 2004). Regardless of the impact on host pheno-

type, the engagement of the adaptor protein TRIF by TLR3

provides an exciting link to necroptosis. It has been known for

some time that in contrast to other TLRs, which use interleukin

1 receptor-associated kinase (IRAK) proteins to activate

NF-kB, TLR3 activation of NF-kB is dependant on RIP kinases

(Feoktistova et al., 2011; Meylan et al., 2004). More recently,

evidence has been provided that TLR3 activation can result in

assembly of the so-called ‘‘ripoptosome,’’ which contains

RIP1, Fas-Associated protein with Death Domain (FADD), cas-

pase-8, and caspase-10 (Feoktistova et al., 2011), and under

specific conditions may result in necroptosis. Through an

Immunity 35, October 28, 2011 ª2011 Elsevier Inc. 485

Immunity

Perspective

independent mechanism, TLR3 may trigger intrinsic apoptosis

via an IRF-3-mediated signaling pathway. It has been revealed

that IRF-3 contains a BH3 domain that enables interaction with

Bax, that mediates their colocalization to the mitochondria,

and that is essential for the induction of MOMP (Chattopadhyay

et al., 2010). Future studies will be required to define the parti-

tioning of IRF-3-induced cell death versus activation of type I

IFN pathways. For TLR9 activation, CMVmay engage an indirect

cell death mechanism because the dominant cell type express-

ing this host sensor is not directly infected (Dalod et al., 2003;

Krug et al., 2004). As discussed above, this includes the rapid

induction of TNFR superfamily members that engage death

receptors and activate caspase-8. Presumably, these rapidly

induced innate pathways explain the evolutionary conservation

of the IE gene product vICA across betaherpesviruses.

Considerable advances have been made toward the charac-

terization of DNA sensors and two molecules have been thus

far implicated in CMV pathogenesis. Interestingly, here too,

crosstalk with cell death pathways has been elucidated as well

as the demonstration of CMV-encoded regulators. HCMV

genome engages the DNA-dependent activator of IRFs (DAI,

also known as Z-DNA binding protein 1 or ZBP1) and on the

basis of studies using UV-inactivated virus, it has been sug-

gested that genomic dsDNA, during its transport to the nucleus,

is recognized by host sensors (DeFilippis et al., 2010). In addition

to IRF-3 activation, DAI may also recruit RIP1 and RIP3 through

its two RHIM domains (Rebsamen et al., 2009). This provides

a mechanistic link to NF-kB activation, but may also engage

the ripoptosome. Importantly, MCMV m45 has been reported

to inhibit DAI signaling. One study has implicated a second

DNA sensor, absent in melanoma 2 (AIM2), that creates a link

to the inflammasome. Mice deficient for AIM2 showed a mild,

but significant enhancement in viral propagation and diminished

IFN-g production in NK cells (Rathinam et al., 2010). Although

a clear link between AIM2 activation and pyroptosis has been

established in other infectious models (e.g., Listeria monocyto-

genes), it remains unclear whether caspase-1-mediated cell

death contributes to CMV pathogenesis. What is clear is that

during the early phases of infection, CMV has evolved a mecha-

nism for modulating multiple cell death and innate immune path-

ways. Studies on single cells with kinetic resolution will be

required to precisely establish the temporal regulation of these

pathways as it maps to viral antigen expression.

A ‘‘Death-Defying’’ Balancing Act between Viral

Propagation and Cellular Immunity

Given the mechanisms discussed thus far, it is no surprise that

CMVs have evolved strategies for redirecting normal cellular

pathways implicated in cell-mediated immunity. What is sur-

prising, however, is that in spite of redundant mechanisms of

escaping detection, there exist long-lasting and robust T cell

responses. Many in-depth reviews have been devoted to

CMV-mediated inhibition of NK and T cell activation and is there-

fore not discussed here (Loureiro and Ploegh, 2006; Mocarski,

2004). Instead, we focus on the role of direct versus cross-

priming of T cells in the context of CMV-specific immunity.

Because of the obvious challenges of studying functional

immune activation in humans, most studies of T cell priming

have been performed with MCMV. Given that DCs may be

directly infected by CMV, viral protein destabilization of key

486 Immunity 35, October 28, 2011 ª2011 Elsevier Inc.

antigen-processing machinery might impact priming efficiency.

In one important example, MCMV m152 (gp40) was shown to

be capable of retaining peptide-loaded MHC I complexes in

the cis-Golgi, and in vivo data suggest that Dm152 MCMV was

more rapidly cleared (Holtappels et al., 2004; Ziegler et al.,

2000). Strikingly, clearancewasmediated bym45985-993-specific

Db-restricted CD8+ T cells, which was considered the immuno-

dominant response during acute infection, and presumably

primed via cross-presentation of antigen acquired from dying

cells. This observation led to the conclusion that in some

instances MCMV ‘‘misleads’’ its host by facilitating the priming

of CD8+ T cells specific for an epitope that is not efficiently pre-

sented by directly wild-type-infected cells (Holtappels et al.,

2004). These data highlight an interesting parallel to IVA adaptive

responses. Another intriguing point is that the population of im-

munodominant T cells in mice was reactive to a peptide present

in vIRA. This illustrates a potential trade-off between inhibition of

necroptosis and exposure of high-affinity T cell epitopes, which

in turn must be retained within infected cells by other MCMV

regulatory proteins (e.g., m152). Future studies will be required

in order to clarify the balance between viral inhibitors of cell

death, antigen presentation, and modulators of CMV immunity.

Inhibiting Cell Death during Latent Infection

Herpesviridae share the ability to enter an altered transcriptional

and translational state in the course of their viral life cycle. This

state of latency is defined as a nonproductive, persistent infec-

tion, in which the viral genome is present, but gene expression

is restricted to a smaller subset of genes, and importantly, infec-

tious virus cannot be detected. Studies in humans andmice have

identified endothelial cells and selected cell types of the myeloid

lineage as critical sites of CMV persistence and latency.

Although latent HCMV has the potential to become reactivated

and has been implicated in several human diseases, we are

only beginning to understand the virally encoded determinants

and their role in establishing and maintaining CMV latency.

This is in sharp contrast to other herpes viruses, such as EBV,

in which clearly established latency programs as well as links

to cell death pathways have been defined (e.g., viral Bcl-2) (Virgin

et al., 2009). Investigation into CMV latency has been con-

founded by the species specificity of HCMV and the low fre-

quency of latently infected cells. The establishment of latent

infection is by definition closely linked to host immune responses

and implicates an interconnection with cell survival and cell

death pathways. Therefore, greater attention to CMV latency is

warranted.

Microarray studies of gene expression during HCMV latency

phase have provided some initial insight into the restricted

expression of viral proteins. Human cytomegalovirus latency

associated transcripts (CLTs) are thought to include gene prod-

ucts encoded in the major immediate early (MIE) and UL111A

regions. One interesting latency gene is US28, a functional che-

mokine receptor that binds to a subset of CC chemokines (Slo-

bedman et al., 2010). Notably, US28 induces caspase-8 and

caspase-10 activation, resulting in apoptotic cell death.

Although the precise mechanism of oligomerizing these initiator

caspases is unknown, it does not seem to involve classical death

receptor engagement (Casarosa et al., 2001).

Forging a link between CLT-induced cell death and host

immunity may establish a deeper understanding of latency itself.

Immunity

Perspective

Emerging data in the herpesvirus field suggest that latency, even

in immune competent hosts, is punctuated by intermittent viral

reactivation. During MCMV latency, stochastic episodes of desi-

lencing (or reactivation) have been linked to the transient gener-

ation and presentation of viral peptide epitopes byMHC I (Simon

et al., 2006). Such transient events may also serve to maintain

the robust T cell activation, at least for a subset of CMV-specific

T cells (Munks et al., 2006). Of great interest will be the mapping

of which cells experience transient reactivation and how these

events are linked to antigen production, cell death, and cross-

priming. Characterization of these re-activation episodes will

alter our concept of latency, which may be erroneously associ-

ated with the absence of detectable virus because of the insen-

sitivity of our assay systems. In this regard, the immune system

and the engagement of cell death programsmay provide a better

definition of what constitutes a latent infection. Finally, this anal-

ysis offers an opportunity to consider the relationship between

viral transmission and orchestrated bouts of CMV reactivation.

Indeed, the epithelium within the salivary gland has provided

an interesting example of distinct cell-specific regulation.

HCMV DNA may be detected in the saliva of 2%–10% of the

human population at any interval time (Britt, 2007). A link to

immune regulation is evident, as suggested by studies in iatro-

genic-induced immunosuppression in which shedding rates

can reach 45% in patients after stem cell transplantation (Cor-

reia-Silva et al., 2007). Other target tissues in which shedding

and transmission have been described are the ductal epithelium

of mammary glands (accounting for mother to child infection);

the ductal epithelium of the kidney tubules (accounting for shed-

ding in urine); and the syncytiotrophoblasts (accounting for

transplacental transmission). As such, the intermittent relaxation

of antigen presentation and T cell priming may all be part of the

plan—controlling host immunity and providing ample opportu-

nity for induction of cell death and widespreadmicrobial dissem-

ination (Figure 2).

Although hardly exhaustive, our examination of viral pathogen-

eses has illuminated interesting differences between IVA and

CMV with respect to their molecular articulation among viral

entry, replication, transmission, and the pathways of cell death

and immunity. Even in the comparison between HCMV and

MCMV, we observe striking distinction in how the viruses have

coevolved with their respective hosts. Nonetheless, in the exam-

ples provided, we see an important commonality in the delayed

host cell death, which may favor the virus, whereas the conse-

quent perturbations in cell death might serve as a sensor for

the host cell to initiate an immune response. This may represent

a genetically encoded trade-off ensuring mutual survival, or we

may discover that it is all part of the complex agenda of immuno-

logically niche-filling infectious microbes (Hedrick, 2004).

Cell Death as a Window into CoevolutionThe study of disease pathogenesis from the perspective of cell

death provides interesting insight into host immunity, and it

may provide a mechanism to chart a historical record of our

coevolution with a particular microbe or class of microbes (de-

pending on the rate of evolution). Recent studies have applied

such approaches for the examination of host sensors, discov-

ering that in the human population, there is evidence of genetic

constraint on TLR3, TLR7, TLR8, and TLR9 that is distinct from

the observation of genetic drift within TLR1, TLR2, TLR4,

TLR6, and TLR10. This observation suggest the possibility that

though their interactions with the class of endosomal TLRs,

viruses are exerting selective pressure on their host; in contrast,

there appears to be a redundancy in the latter class of TLRs with

other host sensors playing perhaps a more critical role. Defining

the cell survival and cell death genes under evolutionary

constraint will provide a similarly rich account of past infections

and may provide further insight into the complex crosstalk

among microbe, cell death, and immunity. Influenza in this

context would provide an interesting evolutionary model as it is

possible to track the emergence of new IVA genomes, with

human data regarding the respective virulence of seasonal,

epidemic, or pandemic viral strains. Spanning a different evolu-

tionary timeline and involving multiple microbe-vertebrate host

relationships, the family of herpesviruses will provide a unique

opportunity.

One striking observation is the temporal and apparently

dichotomous relationship that governs themicrobe’s connection

to host cell death. The counterbalanced induction and inhibition

of cell death provides a striking example of ‘‘genetic addition’’—

a selfish gene’s strategy that creates the foundation for symbi-

otic relationships and speciation (Ameisen, 2004; Mochizuki

et al., 2006). Classically stated, a gene or gene complex capable

of achieving genetic addition encodes a toxin (e.g., inducer of

cell death) and an antitoxin (e.g., inhibitor of cell death). The anti-

toxin may act directly on the toxin, with temporal regulation or

differential kinetics of decay ensuring an addicted state. Alterna-

tively, the antitoxin can act via host pathways that crosstalk with

the effector mechanism of toxin-mediated cell death. Moreover,

as suggested herein, the relationshipmay bemulticellular, acting

across the host immune response to counterbalance the activi-

ties of antitoxin and toxin on host physiology.We have previously

speculated that the answer to the question of ‘‘Why cell death?’’

might be phagocytosis. This was first suggested by Charles

David and was based on the discovery that the metazoan Hydra

vulgaris survives in nutrient-deprived conditions by capturing

energy from internalized apoptotic epithelial cells, a survival

mechanism that was recently shown to also implicate autophagy

(Chera et al., 2009). These intriguing studies suggest that phago-

cytosis might have preceded the evolution of apoptosis, with

energy conservation having offered the selective advantage for

communal single-celled organisms to evolve mechanisms of

programmed cell death (Albert, 2004). Recent evolutionary

studies of ancestoral eukaryotes (e.g., Thermoplasma) indicate

that it might have been a trifecta—phagocytosis, eukaryogene-

sis and perhaps intrinsic apoptosis—all emerging as part of the

evolutionary continuum associated with the acquisition of

a bacterial endosymbiont (Yutin et al., 2009). Coming back to

our interconnectedness with viruses (and transposable ele-

ments), we acknowledge the contributions made in the form of

telomeric ends to our DNA (Nosek et al., 2006), the emergence

of genes mediating somatic recombination in lymphocytes (Fug-

mann, 2010), and the possible role of transposable elements in

miotic segregation (Arkhipova and Meselson, 2005). When all

is said and done, we as academics may not only be learning

about cell death pathways from our microbial counterparts, we

may have actually evolved the mechanisms themselves as

a direct response to viruses’ dying need to replicate.

Immunity 35, October 28, 2011 ª2011 Elsevier Inc. 487

Immunity

Perspective

ACKNOWLEDGMENTS

We would like to dedicate this review to the memory of R. Steinman, theconsummate mentor and a leader in the field of dendritic cell biology andhost response to infection. Far ahead of his time, he informed the dendriticcell community in 2000 that we must always consider the microbes agendaif we are going to understand how to vaccinate the host. His insights andsupport will be sorely missed. We would like to thank A. Oberst, O. Schwartz,and M. Laird for their careful read of the manuscript. We also acknowledgehelpful discussions with L. Quintana-Murci, D. Green and members of hislab, and the Laboratory of Dendritic Cell Biology. Funding support is fromthe Agence National de Recherche (ANR) and the Immuno-Oncology LabExProject.

REFERENCES

Accapezzato, D., Visco, V., Francavilla, V., Molette, C., Donato, T., Paroli, M.,Mondelli, M.U., Doria, M., Torrisi, M.R., and Barnaba, V. (2005). Chloroquineenhances human CD8+ T cell responses against soluble antigens in vivo. J.Exp. Med. 202, 817–828.

Albert, M.L. (2004). Death-defying immunity: Do apoptotic cells influenceantigen processing and presentation? Nat. Rev. Immunol. 4, 223–231.

Albert, M.L., Pearce, S.F., Francisco, L.M., Sauter, B., Roy, P., Silverstein,R.L., and Bhardwaj, N. (1998a). Immature dendritic cells phagocytoseapoptotic cells via alphavbeta5 andCD36, and cross-present antigens to cyto-toxic T lymphocytes. J. Exp. Med. 188, 1359–1368.

Albert, M.L., Sauter, B., and Bhardwaj, N. (1998b). Dendritic cells acquireantigen from apoptotic cells and induce class I-restricted CTLs. Nature 392,86–89.

Allen, I.C., Scull, M.A., Moore, C.B., Holl, E.K., McElvania-TeKippe, E.,Taxman, D.J., Guthrie, E.H., Pickles, R.J., and Ting, J.P. (2009). The NLRP3 in-flammasome mediates in vivo innate immunity to influenza A virus throughrecognition of viral RNA. Immunity 30, 556–565.

Ameisen, J.C. (2004). Looking for death at the core of life in the light of evolu-tion. Cell Death Differ. 11, 4–10.

Andreansky, S.S., Stambas, J., Thomas, P.G., Xie, W., Webby, R.J., andDoherty, P.C. (2005). Consequences of immunodominant epitope deletionfor minor influenza virus-specific CD8+-T-cell responses. J. Virol. 79, 4329–4339.

Arkhipova, I., andMeselson,M. (2005). Deleterious transposable elements andthe extinction of asexuals. Bioessays 27, 76–85.

Arnoult, D., Bartle, L.M., Skaletskaya, A., Poncet, D., Zamzami, N., Park, P.U.,Sharpe, J., Youle, R.J., and Goldmacher, V.S. (2004). Cytomegalovirus celldeath suppressor vMIA blocks Bax- but not Bak-mediated apoptosis bybinding and sequestering Bax at mitochondria. Proc. Natl. Acad. Sci. USA101, 7988–7993.

Arvin A., Campadelli-Fiume G., Mocarski E., Moore P.S., Roizman B., WhitleyR., and Yamanishi K., eds. (2007). Human Herpesviruses: Biology, Therapy,and Immunoprophylaxis (Cambridge: Cambridge University Press).

Barreiro, L.B., and Quintana-Murci, L. (2010). From evolutionary genetics tohuman immunology: how selection shapes host defence genes. Nat. Rev.Genet. 11, 17–30.

Bender, A., Albert, M., Reddy, A., Feldman, M., Sauter, B., Kaplan, G., Hell-man, W., and Bhardwaj, N. (1998). The distinctive features of influenza virusinfection of dendritic cells. Immunobiology 198, 552–567.

Bergmann, M., Garcia-Sastre, A., Carnero, E., Pehamberger, H., Wolff, K., Pal-ese, P., and Muster, T. (2000). Influenza virus NS1 protein counteracts PKR-mediated inhibition of replication. J. Virol. 74, 6203–6206.

Kamps B.S., Hoffman C., and Presier W., eds. (2006). Influenza Report(Cologne: Druckhaus Sud GmbH).

Blachere, N.E., Darnell, R.B., and Albert, M.L. (2005). Apoptotic cells deliverprocessed antigen to dendritic cells for cross-presentation. PLoSBiol. 3, e185.

Boehme, K.W., Guerrero, M., and Compton, T. (2006). Human cytomegalo-virus envelope glycoproteins B and H are necessary for TLR2 activation inpermissive cells. J. Immunol. 177, 7094–7102.

488 Immunity 35, October 28, 2011 ª2011 Elsevier Inc.

Britt, W. (2007). Virus entry into host, establishment of infection, spread in host.In Human Herpesviruses, T. Biology, A.A. Immunoprophylaxis, G. Campadelli-Fiume, E. Mocarski, P.S. Moore, B. Roizman, R. Whitley, and K. Yamanishi,eds. (Cambridge: Cambridge University Press).

Brune, W. (2011). Inhibition of programmed cell death by cytomegaloviruses.Virus Res. 157, 144–150.

Brune, W., Menard, C., Heesemann, J., and Koszinowski, U.H. (2001). A ribo-nucleotide reductase homolog of cytomegalovirus and endothelial celltropism. Science 291, 303–305.

Brune, W., Nevels, M., and Shenk, T. (2003). Murine cytomegalovirus m41open reading frame encodes a Golgi-localized antiapoptotic protein. J. Virol.77, 11633–11643.

Cam, M., Handke, W., Picard-Maureau, M., and Brune, W. (2010). Cytomega-loviruses inhibit Bak- and Bax-mediated apoptosis with two separate viralproteins. Cell Death Differ. 17, 655–665.

Carlson, C.M., Turpin, E.A., Moser, L.A., O’Brien, K.B., Cline, T.D., Jones, J.C.,Tumpey, T.M., Katz, J.M., Kelley, L.A., Gauldie, J., and Schultz-Cherry, S.(2010). Transforming growth factor-b: Activation by neuraminidase and rolein highly pathogenic H5N1 influenza pathogenesis. PLoS Pathog. 6, e1001136.

Casarosa, P., Bakker, R.A., Verzijl, D., Navis, M., Timmerman, H., Leurs, R.,and Smit, M.J. (2001). Constitutive signaling of the human cytomegalovirus-encoded chemokine receptor US28. J. Biol. Chem. 276, 1133–1137.

Chanturiya, A.N., Basanez, G., Schubert, U., Henklein, P., Yewdell, J.W., andZimmerberg, J. (2004). PB1-F2, an influenza A virus-encoded proapoptoticmitochondrial protein, creates variably sized pores in planar lipid membranes.J. Virol. 78, 6304–6312.

Chattopadhyay, S., Marques, J.T., Yamashita, M., Peters, K.L., Smith, K.,Desai, A., Williams, B.R., and Sen, G.C. (2010). Viral apoptosis is induced byIRF-3-mediated activation of Bax. EMBO J. 29, 1762–1773.

Chen, W., Calvo, P.A., Malide, D., Gibbs, J., Schubert, U., Bacik, I., Basta, S.,O’Neill, R., Schickli, J., Palese, P., et al. (2001a). A novel influenza A virus mito-chondrial protein that induces cell death. Nat. Med. 7, 1306–1312.

Chen, W., Norbury, C.C., Cho, Y., Yewdell, J.W., and Bennink, J.R. (2001b).Immunoproteasomes shape immunodominance hierarchies of antiviralCD8(+) T cells at the levels of T cell repertoire and presentation of viral anti-gens. J. Exp. Med. 193, 1319–1326.

Chera, S., Buzgariu, W., Ghila, L., and Galliot, B. (2009). Autophagy in Hydra: Aresponse to starvation and stress in early animal evolution. Biochim. Biophys.Acta 1793, 1432–1443.

Cicin-Sain, L., Ruzsics, Z., Podlech, J., Bubi�c, I., Menard, C., Jonji�c, S., Red-dehase, M.J., and Koszinowski, U.H. (2008). Dominant-negative FADDrescues the in vivo fitness of a cytomegalovirus lacking an antiapoptotic viralgene. J. Virol. 82, 2056–2064.

Coleman, J.R. (2007). The PB1-F2 protein of Influenza A virus: Increasing path-ogenicity by disrupting alveolar macrophages. Virol. J. 4, 9.

Correia-Silva, Jde.F., Victoria, J.M., Guimaraes, A.L., Salomao, U.E., deAbreu, M.H., Bittencourt, H., and Gomez, R.S. (2007). Cytomegalovirus shed-ding in the oral cavity of allogeneic haematopoietic stem cell transplantpatients. Oral Dis. 13, 163–169.

Crowe, S.R., Miller, S.C., and Woodland, D.L. (2006). Identification of protec-tive and non-protective T cell epitopes in influenza. Vaccine 24, 452–456.

Dalod, M., Hamilton, T., Salomon, R., Salazar-Mather, T.P., Henry, S.C.,Hamilton, J.D., and Biron, C.A. (2003). Dendritic cell responses to early murinecytomegalovirus infection: Subset functional specialization and differentialregulation by interferon alpha/beta. J. Exp. Med. 197, 885–898.

DeFilippis, V.R., Alvarado, D., Sali, T., Rothenburg, S., and Fruh, K. (2010).Human cytomegalovirus induces the interferon response via the DNA sensorZBP1. J. Virol. 84, 585–598.

de la Calle, C., Joubert, P.E., Law, H.K., Hasan, M., and Albert, M.L. (2011).Simultaneous assessment of autophagy and apoptosis using multispectralimaging cytometry. Autophagy 7, 1045–1051.

Diebold, S.S., Kaisho, T., Hemmi, H., Akira, S., and Reis e Sousa, C. (2004).Innate antiviral responses by means of TLR7-mediated recognition of single-stranded RNA. Science 303, 1529–1531.

Immunity

Perspective

Downward, J. (2004). PI 3-kinase, Akt and cell survival. Semin. Cell Dev. Biol.15, 177–182.

Edelmann, K.H., Richardson-Burns, S., Alexopoulou, L., Tyler, K.L., Flavell,R.A., and Oldstone, M.B. (2004). Does Toll-like receptor 3 play a biologicalrole in virus infections? Virology 322, 231–238.

Ehrhardt, C., and Ludwig, S. (2009). A new player in a deadly game: Influenzaviruses and the PI3K/Akt signalling pathway. Cell. Microbiol. 11, 863–871.

Ehrhardt, C., Wolff, T., and Ludwig, S. (2007a). Activation of phosphatidylino-sitol 3-kinase signaling by the nonstructural NS1 protein is not conservedamong type A and B influenza viruses. J. Virol. 81, 12097–12100.

Ehrhardt, C., Wolff, T., Pleschka, S., Planz, O., Beermann, W., Bode, J.G.,Schmolke, M., and Ludwig, S. (2007b). Influenza A virus NS1 protein activatesthe PI3K/Akt pathway to mediate antiapoptotic signaling responses. J. Virol.81, 3058–3067.

Feoktistova, M., Geserick, P., Kellert, B., Dimitrova, D.P., Langlais, C., Hupe,M., Cain, K., MacFarlane, M., Hacker, G., and Leverkus, M. (2011). cIAPs blockRipoptosome formation, a RIP1/caspase-8 containing intracellular cell deathcomplex differentially regulated by cFLIP isoforms. Mol. Cell 43, 449–463.

Fugmann, S.D. (2010). The origins of the Rag genes—from transposition toV(D)J recombination. Semin. Immunol. 22, 10–16.

Gannage, M., Dormann, D., Albrecht, R., Dengjel, J., Torossi, T., Ramer, P.C.,Lee, M., Strowig, T., Arrey, F., Conenello, G., et al. (2009). Matrix protein 2 ofinfluenza A virus blocks autophagosome fusion with lysosomes. Cell HostMicrobe 6, 367–380.

Giodini, A., and Albert, M.L. (2010). A whodunit: An appointment with death.Curr. Opin. Immunol. 22, 94–108.

Green, D.R. (2011). The end and after: How dying cells impact the livingorganism. Immunity 35, this issue, 441–444.

Green, D.R., Ferguson, T., Zitvogel, L., and Kroemer, G. (2009). Immunogenicand tolerogenic cell death. Nat. Rev. Immunol. 9, 353–363.

Hedrick, S.M. (2004). The acquired immune system: A vantage from beneath.Immunity 21, 607–615.

Hengel, H., Koszinowski, U.H., and Conzelmann, K.K. (2005). Viruses know itall: New insights into IFN networks. Trends Immunol. 26, 396–401.

Herold, S., Steinmueller, M., von Wulffen, W., Cakarova, L., Pinto, R.,Pleschka, S., Mack, M., Kuziel, W.A., Corazza, N., Brunner, T., et al. (2008).Lung epithelial apoptosis in influenza virus pneumonia: The role of macro-phage-expressed TNF-related apoptosis-inducing ligand. J. Exp. Med. 205,3065–3077.

Holtappels, R., Podlech, J., Pahl-Seibert, M.F., Julch, M., Thomas, D., Simon,C.O., Wagner, M., and Reddehase, M.J. (2004). Cytomegalovirus misleads itshost by priming of CD8 T cells specific for an epitope not presented in infectedtissues. J. Exp. Med. 199, 131–136.

Ibanez, C.E., Schrier, R., Ghazal, P., Wiley, C., and Nelson, J.A. (1991). Humancytomegalovirus productively infects primary differentiated macrophages. J.Virol. 65, 6581–6588.

Ichinohe, T., Lee, H.K., Ogura, Y., Flavell, R., and Iwasaki, A. (2009). Inflamma-some recognition of influenza virus is essential for adaptive immuneresponses. J. Exp. Med. 206, 79–87.

Ichinohe, T., Pang, I.K., and Iwasaki, A. (2010). Influenza virus activates inflam-masomes via its intracellular M2 ion channel. Nat. Immunol. 11, 404–410.

Ichinohe, T., Pang, I.K., Kumamoto, Y., Peaper, D.R., Ho, J.H., Murray, T.S.,and Iwasaki, A. (2011). Microbiota regulates immune defense against respira-tory tract influenza A virus infection. Proc. Natl. Acad. Sci. USA 108, 5354–5359.

Jounai, N., Takeshita, F., Kobiyama, K., Sawano, A., Miyawaki, A., Xin, K.Q.,Ishii, K.J., Kawai, T., Akira, S., Suzuki, K., and Okuda, K. (2007). The Atg5Atg12 conjugate associates with innate antiviral immune responses. Proc.Natl. Acad. Sci. USA 104, 14050–14055.

Jurak, I., Schumacher, U., Simic, H., Voigt, S., and Brune, W. (2008). Murinecytomegalovirus m38.5 protein inhibits Bax-mediated cell death. J. Virol. 82,4812–4822.

Kaiser, W.J., Upton, J.W., Long, A.B., Livingston-Rosanoff, D., Daley-Bauer,L.P., Hakem, R., Caspary, T., and Mocarski, E.S. (2011). RIP3 mediates theembryonic lethality of caspase-8-deficient mice. Nature 471, 368–372.

Kang, R., Zeh, H.J., Lotze, M.T., and Tang, D. (2011). The Beclin 1 networkregulates autophagy and apoptosis. Cell Death Differ. 18, 571–580.

Kato, H., Takeuchi, O., Sato, S., Yoneyama, M., Yamamoto, M., Matsui, K.,Uematsu, S., Jung, A., Kawai, T., Ishii, K.J., et al. (2006). Differential roles ofMDA5 and RIG-I helicases in the recognition of RNA viruses. Nature 441,101–105.

Kijpittayarit, S., Eid, A.J., Brown, R.A., Paya, C.V., and Razonable, R.R. (2007).Relationship between Toll-like receptor 2 polymorphism and cytomegalovirusdisease after liver transplantation. Clin. Infect. Dis. 44, 1315–1320.

Krug, A., French, A.R., Barchet, W., Fischer, J.A., Dzionek, A., Pingel, J.T.,Orihuela, M.M., Akira, S., Yokoyama, W.M., and Colonna, M. (2004). TLR9-dependent recognition of MCMV by IPC and DC generates coordinated cyto-kine responses that activate antiviral NK cell function. Immunity 21, 107–119.

Lagunoff, M., and Carroll, P.A. (2003). Inhibition of apoptosis by the gamma-herpesviruses. Int. Rev. Immunol. 22, 373–399.

Lamkanfi, M., and Dixit, V.M. (2010). Manipulation of host cell death pathwaysduring microbial infections. Cell Host Microbe 8, 44–54.

Lederberg, J. (1999). J. B. S. Haldane (1949) on infectious disease and evolu-tion. Genetics 153, 1–3.

Lederberg, J. (2000). Infectious history. Science 288, 287–293.

Lee, H.K., Lund, J.M., Ramanathan, B., Mizushima, N., and Iwasaki, A. (2007).Autophagy-dependent viral recognition by plasmacytoid dendritic cells.Science 315, 1398–1401.

Loureiro, J., and Ploegh, H.L. (2006). Antigen presentation and the ubiquitin-proteasome system in host-pathogen interactions. Adv. Immunol. 92,225–305.

Mack, C., Sickmann, A., Lembo, D., and Brune, W. (2008). Inhibition of proin-flammatory and innate immune signaling pathways by a cytomegalovirusRIP1-interacting protein. Proc. Natl. Acad. Sci. USA 105, 3094–3099.

Mackey, D., andMcFall, A.J. (2006). MAMPs andMIMPs: Proposed classifica-tions for inducers of innate immunity. Mol. Microbiol. 61, 1365–1371.

McAuley, J.L., Chipuk, J.E., Boyd, K.L., Van De Velde, N., Green, D.R., andMcCullers, J.A. (2010). PB1-F2 proteins from H5N1 and 20 century pandemicinfluenza viruses cause immunopathology. PLoS Pathog. 6, e1001014.

McCormick, A.L., Meiering, C.D., Smith, G.B., andMocarski, E.S. (2005). Mito-chondrial cell death suppressors carried by human and murine cytomegalo-virus confer resistance to proteasome inhibitor-induced apoptosis. J. Virol.79, 12205–12217.

McCormick, A.L., Roback, L., and Mocarski, E.S. (2008). HtrA2/Omi termi-nates cytomegalovirus infection and is controlled by the viral mitochondrialinhibitor of apoptosis (vMIA). PLoS Pathog. 4, e1000063.

McCormick, A.L., Roback, L., Livingston-Rosanoff, D., and St Clair, C. (2010).The human cytomegalovirus UL36 gene controls caspase-dependent and-independent cell death programs activated by infection of monocytes differ-entiating to macrophages. J. Virol. 84, 5108–5123.

McLean, J.E., Datan, E., Matassov, D., and Zakeri, Z.F. (2009). Lack of Baxprevents influenza A virus-induced apoptosis and causes diminished viralreplication. J. Virol. 83, 8233–8246.

Menard, C.,Wagner, M., Ruzsics, Z., Holak, K., Brune,W., Campbell, A.E., andKoszinowski, U.H. (2003). Role of murine cytomegalovirus US22 gene familymembers in replication in macrophages. J. Virol. 77, 5557–5570.

Meylan, E., Burns, K., Hofmann, K., Blancheteau, V., Martinon, F., Kelliher, M.,and Tschopp, J. (2004). RIP1 is an essential mediator of Toll-like receptor3-induced NF-kappa B activation. Nat. Immunol. 5, 503–507.

Mibayashi, M., Martınez-Sobrido, L., Loo, Y.M., Cardenas, W.B., Gale, M., Jr.,and Garcıa-Sastre, A. (2007). Inhibition of retinoic acid-inducible gene I-medi-ated induction of beta interferon by the NS1 protein of influenza A virus. J. Virol.81, 514–524.

Immunity 35, October 28, 2011 ª2011 Elsevier Inc. 489

Immunity

Perspective

Michallet, M.C., Meylan, E., Ermolaeva, M.A., Vazquez, J., Rebsamen, M.,Curran, J., Poeck, H., Bscheider, M., Hartmann, G., Konig, M., et al. (2008).TRADD protein is an essential component of the RIG-like helicase antiviralpathway. Immunity 28, 651–661.

Miller-Kittrell, M., and Sparer, T.E. (2009). Feeling manipulated: Cytomegalo-virus immune manipulation. Virol. J. 6, 4.

Mocarski, E.S., Jr. (2004). Immune escape and exploitation strategies of cyto-megaloviruses: Impact on and imitation of themajor histocompatibility system.Cell. Microbiol. 6, 707–717.

Mochizuki, A., Yahara, K., Kobayashi, I., and Iwasa, Y. (2006). Genetic addic-tion: Selfish gene’s strategy for symbiosis in the genome. Genetics 172, 1309–1323.

Munks, M.W., Cho, K.S., Pinto, A.K., Sierro, S., Klenerman, P., and Hill, A.B.(2006). Four distinct patterns of memory CD8 T cell responses to chronicmurine cytomegalovirus infection. J. Immunol. 177, 450–458.

Nencioni, L., De Chiara, G., Sgarbanti, R., Amatore, D., Aquilano, K., Marcocci,M.E., Serafino, A., Torcia, M., Cozzolino, F., Ciriolo, M.R., et al. (2009). Bcl-2expression and p38MAPK activity in cells infected with influenza A virus:Impact on virally induced apoptosis and viral replication. J. Biol. Chem. 284,16004–16015.

Nosek, J., Kosa, P., and Tomaska, L. (2006). On the origin of telomeres: Aglimpse at the pre-telomerase world. Bioessays 28, 182–190.

Oberst, A., Dillon, C.P., Weinlich, R., McCormick, L.L., Fitzgerald, P., Pop, C.,Hakem, R., Salvesen, G.S., and Green, D.R. (2011). Catalytic activity of thecaspase-8-FLIP(L) complex inhibits RIPK3-dependent necrosis. Nature 471,363–367.

Poole, E., King, C.A., Sinclair, J.H., and Alcami, A. (2006). The UL144 geneproduct of human cytomegalovirus activates NFkappaB via a TRAF6-depen-dent mechanism. EMBO J. 25, 4390–4399.

Rathinam, V.A., Jiang, Z., Waggoner, S.N., Sharma, S., Cole, L.E., Waggoner,L., Vanaja, S.K., Monks, B.G., Ganesan, S., Latz, E., et al. (2010). The AIM2 in-flammasome is essential for host defense against cytosolic bacteria and DNAviruses. Nat. Immunol. 11, 395–402.

Rebsamen, M., Heinz, L.X., Meylan, E., Michallet, M.C., Schroder, K., Hof-mann, K., Vazquez, J., Benedict, C.A., and Tschopp, J. (2009). DAI/ZBP1recruits RIP1 and RIP3 through RIP homotypic interaction motifs to activateNF-kappaB. EMBO Rep. 10, 916–922.

Rivett, A.J., and Hearn, A.R. (2004). Proteasome function in antigen presenta-tion: Immunoproteasome complexes, Peptide production, and interactionswith viral proteins. Curr. Protein Pept. Sci. 5, 153–161.

Rossman, J.S., and Lamb, R.A. (2009). Autophagy, apoptosis, and the influ-enza virus M2 protein. Cell Host Microbe 6, 299–300.

Schultz-Cherry, S., and Hinshaw, V.S. (1996). Influenza virus neuraminidaseactivates latent transforming growth factor beta. J. Virol. 70, 8624–8629.

Seong, S.Y., and Matzinger, P. (2004). Hydrophobicity: An ancient damage-associated molecular pattern that initiates innate immune responses. Nat.Rev. Immunol. 4, 469–478.

Simon, C.O., Holtappels, R., Tervo, H.M., Bohm, V., Daubner, T., Oehrlein-Karpi, S.A., Kuhnapfel, B., Renzaho, A., Strand, D., Podlech, J., et al. (2006).CD8 T cells control cytomegalovirus latency by epitope-specific sensing oftranscriptional reactivation. J. Virol. 80, 10436–10456.

Skaletskaya, A., Bartle, L.M., Chittenden, T., McCormick, A.L., Mocarski, E.S.,and Goldmacher, V.S. (2001). A cytomegalovirus-encoded inhibitor ofapoptosis that suppresses caspase-8 activation. Proc. Natl. Acad. Sci. USA98, 7829–7834.

490 Immunity 35, October 28, 2011 ª2011 Elsevier Inc.

Slobedman, B., Cao, J.Z., Avdic, S., Webster, B., McAllery, S., Cheung, A.K.,Tan, J.C., and Abendroth, A. (2010). Human cytomegalovirus latent infectionand associated viral gene expression. Future Microbiol. 5, 883–900.

Stasakova, J., Ferko, B., Kittel, C., Sereinig, S., Romanova, J., Katinger, H.,and Egorov, A. (2005). Influenza A mutant viruses with altered NS1 proteinfunction provoke caspase-1 activation in primary humanmacrophages, result-ing in fast apoptosis and release of high levels of interleukins 1beta and 18. J.Gen. Virol. 86, 185–195.

Szomolanyi-Tsuda, E., Liang, X., Welsh, R.M., Kurt-Jones, E.A., and Finberg,R.W. (2006). Role for TLR2 in NK cell-mediated control of murine cytomegalo-virus in vivo. J. Virol. 80, 4286–4291.

Tabeta, K., Georgel, P., Janssen, E., Du, X., Hoebe, K., Crozat, K., Mudd, S.,Shamel, L., Sovath, S., Goode, J., et al. (2004). Toll-like receptors 9 and 3 asessential components of innate immune defense against mouse cytomegalo-virus infection. Proc. Natl. Acad. Sci. USA 101, 3516–3521.

Tait, S.W., and Green, D.R. (2010). Mitochondria and cell death: Outermembrane permeabilization and beyond. Nat. Rev. Mol. Cell Biol. 11,621–632.

Thomas, P.G., Dash, P., Aldridge, J.R., Jr., Ellebedy, A.H., Reynolds, C., Funk,A.J., Martin, W.J., Lamkanfi, M., Webby, R.J., Boyd, K.L., et al. (2009). Theintracellular sensor NLRP3mediates key innate and healing responses to influ-enza A virus via the regulation of caspase-1. Immunity 30, 566–575.

Tomkinson, A., Tepper, J., Morton, M., Bowden, A., Stevens, L., Harris, P.,Lindell, D., Fitch, N., Gundel, R., and Getz, E.B. (2010). Inhaled vs subcuta-neous effects of a dual IL-4/IL-13 antagonist in a monkey model of asthma.Allergy 65, 69–77.

Uhl, M., Kepp, O., Jusforgues-Saklani, H., Vicencio, J.M., Kroemer, G., andAlbert, M.L. (2009). Autophagy within the antigen donor cell facilitates efficientantigen cross-priming of virus-specific CD8+ T cells. Cell Death Differ. 16,991–1005.

Upton, J.W., Kaiser, W.J., and Mocarski, E.S. (2010). Virus inhibition of RIP3-dependent necrosis. Cell Host Microbe 7, 302–313.

Varga, Z.T., Ramos, I., Hai, R., Schmolke, M., Garcıa-Sastre, A., Fernandez-Sesma, A., and Palese, P. (2011). The influenza virus protein PB1-F2 inhibitsthe induction of type i interferon at the level of the MAVS adaptor protein.PLoS Pathog. 7, e1002067.

Virgin, H.W., Wherry, E.J., and Ahmed, R. (2009). Redefining chronic viralinfection. Cell 138, 30–50.

Wurzer, W.J., Planz, O., Ehrhardt, C., Giner, M., Silberzahn, T., Pleschka, S.,and Ludwig, S. (2003). Caspase 3 activation is essential for efficient influenzavirus propagation. EMBO J. 22, 2717–2728.

Yutin, N., Wolf, M.Y., Wolf, Y.I., and Koonin, E.V. (2009). The origins of phago-cytosis and eukaryogenesis. Biol. Direct 4, 9.

Zamarin, D., Garcıa-Sastre, A., Xiao, X., Wang, R., and Palese, P. (2005). Influ-enza virus PB1-F2 protein induces cell death through mitochondrial ANT3 andVDAC1. PLoS Pathog. 1, e4.

Zamarin, D., Ortigoza, M.B., and Palese, P. (2006). Influenza A virus PB1-F2protein contributes to viral pathogenesis in mice. J. Virol. 80, 7976–7983.

Zell, R., Krumbholz, A., Eitner, A., Krieg, R., Halbhuber, K.J., and Wutzler, P.(2007). Prevalence of PB1-F2 of influenza A viruses. J. Gen. Virol. 88, 536–546.

Zhirnov, O.P., and Klenk, H.D. (2007). Control of apoptosis in influenza virus-infected cells by up-regulation of Akt and p53 signaling. Apoptosis 12,1419–1432.

Ziegler, H., Muranyi, W., Burgert, H.G., Kremmer, E., and Koszinowski, U.H.(2000). The luminal part of themurine cytomegalovirus glycoprotein gp40 cata-lyzes the retention of MHC class I molecules. EMBO J. 19, 870–881.