TD immunologie M1 Biologie-Santé Contrôle in vivo de la domiciliation des LT CD8+ activés et mémoires. En vous aidant des questions posées ci-dessous, analysez l’article suivant : Dynamic programming of CD8+ T cell trafficking after live viral immunization. Liu L, Fuhlbrigge RC, Karibian K, Tian T, Kupper TS. Immunity. 2006 Sep;25(3):511-20. 1/ Les auteurs de cet article utilise un modèle expérimental basé sur le transfert adoptif de lymphocytes T transgéniques pour un TCR spécifique du peptide pOVA257-264 (SIINFEKL) dérivé de l’ovalbumine (OVA) présenté par les molécules H2-Kb. Rappelez comment de tels lymphocytes peuvent être obtenus. Comment ces lymphocytes pourront-ils être activés in vivo ? 2/ Définissez le terme "souris congéniques". 3/ Quelle approche expérimentale les auteurs ont-ils choisie pour suivre le devenir des lymphocytes T in vivo ? 4/ Quel virus les auteurs utilisent-ils pour mener leur étude ? Quelles sont les deux voies d’infection utilisées ? 5/ Par quelle technique les auteurs analysent-ils l’activation des LT ? 6/ Analysez la figure 1. Que pouvez-vous conclure ? 7/ De quelles molécules impliquées dans le « homing » des lymphocytes les auteurs analysent-ils l’expression ? A quelle localisation sont-elles associées ? 8/ Analysez la figure 2. Quelles sont vos conclusions ? 9/ Quelle information les auteurs cherchent-ils à obtenir par l’étude du phénotype des LT activés dans les ganglions inguinaux de souris infectées par voie intrapéritonéale ? 10/ Que concluez-vous suite à l’analyse de la figure suivante ?

Figure S2. At 60 hr after rVV-ova i.p. Infection, OT-I CD8+ T Cells Activated in SkinDraining ILN Expressed Skin-Homing Phenotype. Histograms and dot plots were gated on Thy1.1+ donor populations. 11/ Comment les auteurs vérifient-ils que l’expression des molécules de domiciliation résulte de leur activation ? 12/ Pour quelle raison les auteurs effectuent-ils les expériences présentées par la figure 3 ? Que concluez-vous des résultats présentés? 13/ Analysez la figure 4A. Quelle hypothèse peut-on formuler pour expliquer l’absence de cellules en phase P1-3 dans les BLN, CLN, MLN, rate et lavage péritonéal ?

14/ Rappelez le principe de la RT-PCR quantitative en temps réel. Quelle hypothèse les auteurs testent-ils ici dans la figure 4B ? Que pouvez-vous conclure ? 15/ Quelle hypothèse les auteurs testent-ils dans la figure 4 C? Comment les CPA sont-elles purifiées ? Quelle autre façon de purifier ces cellules pourriez-vous proposer ? Quelle conclusion tirez-vous de cette expérience ? 16/ Analysez la figure 5A. 17/ Quelles sont les deux hypothèses proposées par les auteurs pour expliquer l’apparition de nouvelles molécules de « homing » à la surface des LT activés ? Quelle stratégie adoptent-ils pour tester ces hypothèses ? Analysez la suite de la figure. 18/ Analysez la figure 6 et donnez votre conclusion. 19/ Comment les auteurs suivent-il la présence du virus au site d’inoculation ? Que peut-on conclure de l’expérience 7 ? 20/ Récapitulez les principaux résultats de l’article. Quelles sont les cellules qui, selon certaines études, pourraient contrôler l’acquisition des molécules de homing par les LT ? Quel autre mécanisme les auteurs de cette étude proposent-ils sur la base de leur résultat ? Quel résultat expérimental étaie leur hypothèse ? 21/ Concernant l’expression des molécules de « homing » par les cellules T mémoire, quelle sous-population a été étudiée dans cet article ? Quelle autre population reste à étudier ? Une telle étude vous paraît-elle réalisable ? 22/ En quoi les résultats présentés sont-ils importants en terme de vaccinologie ?

Immunity 25, 511–520, September 2006 ª2006 Elsevier Inc.

DOI 10.1016/j.immuni.2006.06.019

Dynamic Programing of CD8+ T Cell Trafficking after Live Viral Immunization Luzheng Liu,1,2,* Robert C. Fuhlbrigge,1,2 Kara Karibian,1 Tian Tian,1,2 and Thomas S. Kupper1,2,* 1 Harvard Skin Disease Research Center Department of Dermatology Brigham and Women’s Hospital Boston, Massachusetts 02115 2 Department of Dermatology Harvard Medical School Boston, Massachusetts 02115

Summary After viral infection, activated T cells are present in multiple tissues regardless of the infection route. How these cells acquire pleiotropic homing ability is unclear. By using a cutaneous vaccinia virus infection model, we demonstrate that regulation of T cell trafficking is multiphasic. Upon completion of three cell divisions, CD8+ T cells upregulated specific skin-homing molecules within draining lymph nodes (LN). By 60 hr after infection, some activated T cells reached the infected tissue, while others entered distant antigenfree LN. These latter cells continued to divide and acquire additional tissue-homing molecules in this new setting, independent of antigen presentation. After viral clearance, the initial skin-homing imprint became the predominant homing phenotype on memory cells and provided superior protection against secondary cutaneous challenge. These observations demonstrate a mechanism by which T cells provide both immediate tissue-specific immune control at the pathogen entry site and a more flexible systemic protection against pathogen dissemination.

Introduction Like most pathogens, viruses typically invade hosts via epithelial interfaces between body and environment. The immune system must both rapidly and efficiently target the tissue of pathogen entry to control the infection in situ, while at the same time protect distant tissues against dissemination of the pathogen. Accordingly, in mouse models, large numbers of effector and memory T cells have been found in all extralymphoid tissues examined after intravenous infection with vesicular stomatitis virus, demonstrating a highly versatile ability of activated T cells to infiltrate different tertiary tissues (Masopust et al., 2001). The ubiquitous presence of effector and memory T cells in multiple tissues after pathogen infection has also been observed after oral infection with Listeria monocytogenes or rotavirus, as well as intranasal infection with Sendai virus (Masopust et al., 2001, 2004). These observations appear to suggest that the anatomic site of pathogen infection, and thus site of initial T cell activation, does not predict the *Correspondence: [email protected] partners.org (T.S.K.)

(L.L.);

tkupper@

subsequent tissue-specific homing of activated T cells; rather, acute microbial infection invariably leads to systemic dissemination of antigen-experienced T cells. The logic of this event is clear, because these T cells can act as sentinels in all extralymphoid tissues upon secondary infection or dissemination of the primary infection. One clinical correlate of this idea is the well-documented fact that epicutaneous inoculation of vaccinia virus (VV) by skin scarification successfully protects people against variola virus infection, which typically occurs via intranasal and oropharyngeal exposure (Lane, 1977). How T cells acquire this ability to home indiscriminately to multiple peripheral tissues after localized infection remains obscure. For example, it has now become clear that T cell migration between different body compartments is a highly regulated process. One current paradigm states that T cell trafficking is controlled by the sequential interactions of adhesion molecules and chemokine receptors that are differentially expressed on various T cell subsets and their target tissues (Kunkel and Butcher, 2002; von Andrian and Mackay, 2000). Naive and central memory T cells (Tcm) predominantly circulate among secondary lymphoid organs (SLO), where they make extensive contacts with professional antigen-presenting cells (Weninger and von Andrian, 2003). These cells express high levels of L-selectin (CD62L) and chemokine (C-C motif) receptor 7 (CCR7) required for optimal lymph nodes (LN) entry (Springer, 1994; Warnock et al., 1998; Weninger et al., 2001). Effector memory T cells (Tem) isolated from skin express specific chemokine receptors such as CCR4 and E-selectin ligands (E-Lig) such as cutaneous lymphocyte antigen in human, while those from small intestine preferentially express a4b7 integrin and CCR9 (Campbell and Butcher, 2002; Campbell et al., 1999; Erdmann et al., 2002; Fuhlbrigge et al., 1997; Kunkel et al., 2000). Recent studies with cell-associated antigen models demonstrated that T cells activated in regional lymphoid nodes (LN) are programmed to express homing molecules specific to the site of the immunization (Calzascia et al., 2005; Dudda et al., 2004). This appears to be at odds with the observations from the infection models outlined above. Since rotavirus has a narrowly defined tissue tropism and wild-type Sendai virus is exclusively pneumotropic in mice (Bass et al., 1990; Tashiro et al., 1990), it is highly unlikely that the ubiquitous presence of activated T cells after local infections results from disseminated infection and T cell activation in multiple SLO. To understand more clearly how T cells develop pleiotropic tissue-homing ability during immune response to local microbial infection, we studied the tissuehoming phenotype and the trafficking pattern of antigen-specific T cells at a series of time points after VV infection by two distinct routes—cutaneous inoculation by skin scarification and intraperitoneal (i.p.) injection. By using recombinant VV expressing the ovalbumin peptide OVA257-264 (rVV-ova) and adoptive transfer of OVA257-264-specific CD8+ T cells from OT-I TCR transgenic mice, we demonstrated multiphasic regulation of T cell trafficking during primary response. At the early

Immunity 512

stage of immune response, activated T cells were imprinted with tissue-specific homing phenotype upon completion of three cell division cycles within regional LN. By 60 hr after infection, some of these cells migrated specifically into infected tissues, while other activated T cells that have divided at least three times also migrated to distant lymphoid tissues that drain uninfected organs. Unexpectedly, these latter T cells acquired additional and distinct tissue-homing imprinting molecules in distant LN, a process that occurs independent of additional antigen activation. Thus, after localized viral infection and T cell priming in localized draining lymph nodes, multiple homing imprinting programs can occur on the activated T cells circulating in different LN microenvironments independent of subsequent systemic antigen presentation. After viral clearance, the initial homing imprint became the predominant, although not the sole, homing phenotype on memory CD8+ T cells present in all SLO. Importantly, this ‘‘memory’’ of the immunization site provides superior protection against secondary challenge via the same route. Taken together, our observations demonstrate a plausible mechanism by which T cells provide both immediate tissuespecific immune control at pathogen entry site and a more flexible systemic protection against pathogen dissemination.

Figure 1. CD8+ T Cell Activation Was Initiated in ILN after Tail Skin Scarification with rVV-ova CFSE-labeled naive Thy1.1+ OT-1 cells were transferred into Thy1.2+ B6 mice. Recipient mice were then infected with rVV-ova by either (A) skin scarification or (B) i.p. injection. 60 hr after the infection, proliferation of OT-I cells in ILN, BLN, CLN, MLN, and spleen were analyzed by flow cytometry. Histograms were gated on Thy1.1+ donor cells and are representative of six independent experiments.

Results

dividing OT-I cells were also found in the spleen. In contrast, after i.p. infection, the most vigorous proliferation observed in MLN and spleen (Figure 1B), while some activated OT-I cells were detected in all the lymphoid tissues examined. This is consistent with the establishment of VV infection involving both the visceral and parietal peritoneum. Together, these results identified ILN and MLN as the major LN for CD8+ T cell activation after skin scarification and i.p. infection, respectively.

Antigen-Specific CD8+ T Cells Were Activated in LN Draining the Infected Tissues after Cutaneous or Intraperitoneal VV Infection To study the trafficking of activated T cells in a physiological viral infection setting, we infected C57B/l6 (B6) mice with VV epicutaneously by skin scarification, a technique widely used to vaccinate humans against smallpox (Breman and Henderson, 2002; Lane, 1977). After VV inoculation near the base of the tail, the infected mice developed a characteristic skin ‘‘pox’’ lesion at the inoculation site (see Figure S1 in the Supplemental Data available online). The temporal evolution of the lesion from vesicle, to pustule, and then to healing crust, closely resembled that of lesions developed by human vaccinees (http://www. bt.cdc.gov/agent/smallpox/vaccination/facts.asp). To track the in vivo activation, proliferation, and trafficking of antigen-specific CD8+ T cells after VV infection, we adoptively transferred Thy1.1+ naive CD8+ T cells from OT-I T cell receptor (TCR) transgenic mice (OT-I cells) into Thy1.2+ B6 mice 24 hr prior to scarification with VV-ova (Norbury et al., 2002). Before transfer, OT-I cells were labeled with carboxy-fluorescein diacetate succinimidyl ester (CFSE). To compare cutaneous viral inoculation with infection involving a distinct anatomic tissue, additional groups of mice were infected with rVV-ova by i.p. injection. At 60 hr postinfection, various secondary lymphoid organs, including inguinal LN (ILN), mesenteric LN (MLN), cervical LN (CLN), brachial LN (BLN), and spleen, were harvested and examined individually for OT-I cell activation. In skin-scarified mice, OT-I cells that had undergone up to seven rounds of proliferation were easily detected in ILN that drained the infected skin area (Figure 1A). No proliferating OT-I cells were found in LN draining other tissues, although a small number of

Tissue-Specific Homing Phenotype Was Imprinted on Activated T Cells in Regional LN within 60 hr of Infection We next analyzed the expression of homing molecules on the activated OT-I cells in ILN or MLN at 60 hr after skin scarification or i.p. infection with rVV-ova, respectively. OT-I cells displayed a naive phenotype (CD62Lhi, E-Lig2, P-Lig2, a4b72) in ILN (Figure 2A), MLN, and spleen (data not shown) of uninfected control mice. In contrast, CD62L was strongly downregulated on a significant number of proliferating OT-I cells after both cutaneous and i.p. infected groups at this time point (Figures 2B and 2C). Concurrently, there was a strong upregulation of the skin-homing molecules E-Lig and P-Lig on the majority of proliferating OT-I cells in ILN of scarified mice, as detected by the specific binding to E-selectin and P-selectin-Ig chimera protein (Figure 2B). The a4b7 was not upregulated on proliferating OT-I cells in ILN of skin-scarified mice. In contrast, after i.p. infection, a4b7 was strongly upregulated on proliferating OT-1 cells in MLN, while expression of E-Lig and P-Lig were not induced significantly (Figure 2C). Interestingly, the downregulation of CD62L and upregulation of CD11a occurred early on after T cell activation and reached a plateau level by the third cell division, while the upregulation of the tissue-homing molecules a4b7, E-Lig, and P-Lig occurred primarily on T cells that had divided more than three times and continued to increase until the seventh cycle of division (Figures 2B–2D). To test whether it is the route of infection or the site of T cell priming that regulates the early expression of tissue-specific homing molecules, we examined the homing molecule expression of OT-I cells activated in ILN of i.p. infected mice at the same time point (Figure 1B). Similar to those activated in ILN of skin-scarified mice,

CD8+ T Cell Trafficking after Viral Infection 513

Figure 3. In Vivo Tissue-Selective Homing of Activated CD8+ T Cells at 60 hr after VV Infection

Figure 2. The LN of T Cell Activation Imprints Differential Expression of Skin or Gut-Homing Molecules on CD8+ T Cells as Early as 60 hr after Infection CFSE-labeled Thy1.1+ OT-I cells were adoptively transferred into Thy1.2+ B6 mice. (A) At 7 days after transfer, the phenotype of OT-I cells in ILN of naive recipient mice was analyzed by flow cytometry. Dot plots were gated on viable lymphocytes. (B and C) ILN of skin-scarified mice (B) and MLN of i.p. infected mice (C) were harvested at 60 hr after rVV-ova infection. The activation, proliferation, and tissue-homing phenotype of OT-I cells were analyzed by flow cytometry. Dot plots were gated on Thy1.1+ cells. The numbers in quadrant indicate the percentages in Thy1.1+ population. (D) The geometric mean fluorescence intensities (GMFI) of the indicated markers on OT-I cells were plotted with the cell division cycles. Filled circles, ILN of scarified mice; unfilled circles, MLN of i.p. mice. Data are representative of six independent experiments.

these cells upregulated skin-homing molecules E-lig and P-lig, but not the gut-homing molecule a4b7 (Figure S2), suggesting that the homing phenotype of activated T cells is imprinted by LN-specific factors rather than the infection route. Together, these data provide strong evidence that very early after infection, activated CD8+ T cells are programmed to express tissue-specific homing phenotype within draining LN. Imprinting of homing phenotype occurs subsequent to the T cell activation, as shown by the fact that OT-I cells in MLN of either uninfected or skin-scarified mice did not express a4b7 and maintained a naive phenotype (data not shown). Activated T Cells Homed Specifically to Infected Peripheral Tissues within 60 hr Having shown that activated CD8+ T cells acquired tissue-specific homing molecules within draining LN 60 hr after viral infection, we asked whether the in vivo

(A) Peritoneal lavage of mice infected with rVV-ova by scarification or i.p. injection were analyzed for the presence of OT-I cells. (B) Proliferation of OT-I cells in peritoneal cavity (solid line) or MLN (dotted line) of i.p. infected mice was analyzed by CFSE signals. Data are representative of two to four independent experiments. (C and D) Tail skin samples were prepared at 60 hr after (C) skin scarification or (D) i.p. infection with rVV-ova. Slides of frozen tissue were stained with anti-CD3 mAb. Scale bars represent 50 mm.

homing preference of these T cells is consistent with their phenotype. Peritoneal lavages and tail skin samples were examined for the presence of T cells. As shown in Figure 3A, significant numbers of Thy1.1+ donor OT-I cells were present in peritoneal lavage of the i.p. infected, but not skin-scarified, mice at 60 hr after infection. Interestingly, the Thy1.1+ cells in peritoneal cavity of the i.p. infected mice had completed more cell division cycles, with the majority of these cells having divided 7–9 times, as compared to cells in MLN having divided on average 5–7 times at the same time point (Figure 3B). Few if any cells in the first three division cycles were found in peritoneal cavity. Peritoneal OT-I cells expressed high level of a4b7 and were CD62Llow (data not shown). These data suggest that the ability to home to nonlymphoid locations (e.g., peritoneal cavity) increases in parallel with increasing rounds of cell division and acquisition of tissue-specific homing molecules. It is possible, however, that effector T cells were recruited to peritoneal cavity nonspecifically by inflammatory signals rather than the interactions between the tissue-homing molecules. To test this possibility, we infected OT-I recipient mice simultaneously with rVV-ova by skin scarification and with wild-type VV (no ova antigen) by i.p. injection. At 60 hr after the infections, OT-I cells were activated and proliferating in ILN, but few OT-I cells were found in peritoneal cavity even in the presence of viral peritonitis (Figure S3). This result suggests that, at this early time point, T cell trafficking is predominantly directed by the tissue-specific homing receptors rather than inflammatory signals. We also investigated T cell homing into skin after VV infection

Immunity 514

targeted the infected tissues early on during primary response against viral infection.

(A) Lymphocytes were prepared from the indicated tissues 5 days after rVV-ova skin scarification. The proliferation of OT-I cells was analyzed by flow cytometry. (B) RNA samples were prepared from the indicated tissues at day 5 after VV scarification. VV infection was measured by real-time RTPCR. Data represent the averages 6 SD of four independent experiments with three mice per group. (C) Antigen-presenting cells were purified with anti-MHC Class II magnetic beads from ILN, MLN, and spleen of B6 mice 4 days after skin scarification with rVV-ova. The cells were cocultured with CFSE-labeled Thy1.1+ OT-I cells for 60 hr. OT-I cell activation and proliferation was monitored by flow cytometer. The histograms were gated on Thy1.1+ population. Data are representative of two independent experiments.

OT-I Cells Activated in ILN Disseminate through Distant LN by Day 5 of VV Skin Scarification We next examined various LNs and spleen for the presence of activated Thy1.1+ OT-I cells at later time points after skin scarification with rVV-ova. In contrast to the highly polarized tissue-specific migration seen at 60 hr after infection, proliferating OT-I cells were found in all SLO examined at day 5 after VV skin scarification (Figure 4A). Notably, while donor T cells in the first three proliferation cycles (OT-I P1-3) were present in ILN as three distinct populations, these cells were barely detectable in other LN and spleen. The lack of OT-I P1-3 in distant LN was further illustrated by the increased ratios of OT-I cells in more advanced cell cycles (OT-I P>3) to OT-I P1-3 in distant lymphoid tissues than in ILN (data not shown). CFSE2 OT-I cells were also found in peritoneal cavity 5 days after skin scarification, although to a much lesser extent than i.p. infected mice (4.07% 6 0.23% versus 22.1% 6 0.078% of Thy1.1+ cells in total peritoneal lavage cells) (Figure 4A and data not shown). Therefore, in contrast to the tissue-specific trafficking at 60 hr after VV infection, at 5 days after skin scarification, activated CD8+ T cells that have divided more than three times in ILN appeared to have disseminated throughout secondary lymphoid organs and reached peripheral sites that were not initially infected by the viruses. We considered the possibility that cutaneous inoculation of VV may have led to viremia and subsequent systemic antigen expression, leading to antigen-driven T cell proliferation in LNs distant from the site of initial infection. To test this, we measured the expression of an early VV gene in different tissues at day 5 after skin scarification by real-time RT-PCR. As expected, abundant VV mRNA was detected in tail skin samples (Figure 4B). While there was a low and barely detectable level of viral gene transcription in ILN, this highly sensitive real-time RT-PCR approach failed to detect any viral gene expression in intestine, MLN, or spleen. Furthermore, antigen-presenting cells (MHC class II+ cells) purified from MLN and spleen on day 4 after rVV-ova skin scarification failed to induce OT-I cell proliferation in vitro (Figure 4C). These data provide strong evidence against a systemic viral infection and/or viral antigen presentation as an explanation for the presence of dividing T cells in distinct LN. Taken together, our results indicate that subsets of activated OT-I cells can leave draining LN after three rounds of cell division and migrate either to the peripheral tissue for which they express homing receptors or to remote LN where they continue to divide in an antigen-independent fashion.

by immunohistochemical staining, because technical issues precluding us from harvesting sufficient cells to perform flow cytometry on this tissue. While very few T cells were found in tail skin samples at 60 hr after i.p. infection, significant infiltration of CD3 T cells was noted in the epidermal and dermal layers of the inoculated area in skin-scarified mice (Figures 3C and 3D). Collectively, these results demonstrate that activated CD8+ T cells imprinted with tissue-homing phenotype specifically

Disseminated OT-I Cells Acquired Additional Homing Molecules in Various LN Microenvironments while Maintaining Their Initial Tissue-Homing Imprinting To study the tissue-homing phenotype of OT-I cells after their dissemination, cells from various SLO were harvested at 5 days post skin scarification and stained for the expression of selectin ligands and integrin a4b7. At this time point, both E-Lig and P-Lig were highly expressed on OT-I cells present in ILN, MLN, and spleen

Figure 4. At Day 5 after Skin Scarification with rVV-ova, Activated OT-I Cells Had Disseminated throughout SLO and Entered Peritoneal Cavity without Concurrent Systemic Viral Infection

CD8+ T Cell Trafficking after Viral Infection 515

Figure 5. At Day 5 after Skin Scarification with rVV-ova, the GutHoming Molecule a4b7 Was Upregulated on Disseminated OT-1 CD8+ T Cells (A) ILN, MLN, and spleen were harvested at day 5 after scarification with rVV-ova. The expression of the indicated homing markers on Thy1.1+ OT-I cells was determined by flow cytometry. Dot plots were gated on Thy1.1+ cells and are representative of five independent experiments. Numbers in quadrants indicate percentages in Thy1.1+ population. (B–D) B6 mice that had received Thy1.1+ OT-I cells were given daily injections of FTY720 starting 24 hr before scarification with rVV-ova. On day 5 after infection, lymphocytes from the indicated SLO were harvested and analyzed by flow cytometry. (B) The percentages of Thy1.1+ cells in total lymphocytes. Black bars, FTY720-treated mice; unfilled bars, untreated mice. Data represent the average 6 SD of six mice from three independent experiments. (C and D) The expression of E-Lig (C) and a4b7 (D) on OT-I cells in ILN. Histograms were gated on Thy1.1+ populations and are representative of three independent experiments. Solid lines, FTY720treated mice; dotted lines, untreated mice.

(Figure 5A). Therefore, the early skin-homing imprint within ILN is maintained on OT-I cells after they have reached distant lymphoid tissues. Unexpectedly, however, the gut-homing molecule a4b7 was also expressed on a significant portion of OT-I cells in all lymphoid tissues examined, although the level of expression was significantly lower than that in i.p. infected mice (GMFI were 49.1 and 318, respectively, Figure 5A and data not shown). Furthermore, while most of the OT-I cells expressing tissue-specific homing phenotype were either E-lig+ or a4b7+, a significant fraction of these cells expressed both molecules. The expression of a4b7 on OT-I cells in skin-scarified mice suggested one of two possibilities: either activated T cells spontaneously began to acquire homing molecules for multiple tissues as the immune response progressed, or alternatively, they received additional tissue-homing imprinting sig-

nals in the new LN microenvironment after their dissemination. To distinguish between these two possibilities, we attempted to block the egress of lymphocytes from ILN after skin scarification by daily administration of FTY720 (2-amino-2-(2[4-octylphenyl]ethyl)-1,3-propanediol hydrochloride) starting 24 hr before rVV-ova skin scarification. FTY720 is a functional antagonist of sphingosine 1-phosphate (S1P) that regulates lymphocyte egress from lymphoid tissues. Administration of FTY720 induces lymphocytopenia via sequestration of lymphocytes in LN without affecting their activation and function (Matloubian et al., 2004; Pinschewer et al., 2000). As expected, the number of Thy1.1+ cells was significantly increased in ILN of skin-scarified FTY720-treated mice, as compared to the numbers in untreated mice on day 5 after infection (Figure 5B). At the same time, there were significantly fewer Thy1.1+ cells in MLN and spleen in skin-scarified FTY720-treated mice, and these cells were undivided according to the CFSE signals (Figure 5B and data not shown). Subsequent phenotyping of cells trapped in the ILN by FTY720 revealed similar levels of E-Lig expression on OT-I cells in ILN of FTY720 treatment and untreated mice (Figure 5C). However, treatment with FTY720 completely blocked the expression of a4b7 on OT-I cells in the ILN (Figure 5D). Meanwhile, FTY720 treatment of i.p. infected mice did not block the expression of a4b7 on OT-I cells activated in MLN (Figure S4). Therefore, it is highly unlikely that the lack of a4b7 expression on OT-I cells in Figure 5D is due to an inhibitory effect of FTY720 on a4b7 upregulation. These data suggest that the upregulation of gut-homing molecules is a facilitated process that occurred after (and as a result of) activated CD8+ T cells circulated through gut-draining LN, rather than a spontaneous development independent of their LN microenvironment. Recently, it was shown that high-number monospecific T cell transfer may artificially alter both the development of activated T cells and their homing receptor expression (Marzo et al., 2005). Specifically, CD62L is aberrantly retained on T cells activated at high precursor frequency. This could significantly affect their in vivo homing preference. To study whether the observed CD8+ T cell dissemination and subsequent acquisition of additional homing molecules also occur in physiological condition or are rather associated with the CD62Lhi OT-I cells generated in the high number cell-transfer setting, experiments with low input cell numbers (5 3 103 OT-I cells per mouse) were performed. At 60 hr post rVV-ova skin scarification or i.p. injection, OT-I cells were activated in ILN or MLN, respectively (Figure S5). Consistent with our findings with higher input cell numbers, there was a significant upregulation of E-Lig on ILN-activated OT-I cells (Figure S5A), while a4b7 was upregulated on MLN-activated OT-I cells (Figure S5B). However, in contrast to Figure 2, with the lower input number of OT-I cells, virtually all the OT-I cells became CD62Llow. Importantly, 5 days after rVV-ova skin scarification, activated OT-I cells were found in all the lymphoid tissues examined, including MLN and spleen. A subset of OT-I cells had acquired gut-homing molecule a4b7 after disseminating into MLN (Figure S6). This is, again, consistent with the observations with higher input OT-I cell number. Therefore, the findings that locally

Immunity 516

Figure 7. Skin-Scarified Memory Mice Were Better Protected against Secondary Cutaneous Viral Challenge than i.p. Immunized Mice

Figure 6. The Imprint of Tissue-Specific Homing Molecule Expression during Acute Viral Infection Was Maintained during the Memory Phase Lymphocytes were harvested from the indicated SLO at 60 hr (A and C) or day 30 (B and D) after rVV-ova infection. OT-I cells were identified by Thy1.1 staining and analyzed by flow cytometry. (A and B) The percentages of Thy1.1+ OT-I cells in total cell population. (C and D) The percentages of E-Lig+ or a4b7+ cells in Thy1.1+ donor cells. Data represent the average 6 SD of six mice in two independent experiments.

activated CD8+ T cells disseminate into nondraining lymphoid tissues and acquire new homing phenotype are recapitulated at a more physiological setting. After Viral Clearance, the Predominant Homing Phenotype on Memory CD8+ T Cells Reflected the Initial Route of Viral Inoculation We have shown that activated OT-I cells were initially programmed within the draining LN to express tissuehoming molecules specific for the site of infection (Figures 2 and 3) but acquired additional homing molecules as a result of subsequent internodal migration (Figure 5). We then asked which set of tissue-homing molecules would persist on memory CD8+ T cells at a time point well after the infection had been cleared. We therefore determined the expression of tissue-homing molecules on Thy1.1+ memory OT-I cells at day 30 after VV infection, either by skin scarification or i.p. injection. At this time point, the frequencies of Thy1.1+ cells had reached equilibrium among all LN examined in both groups of mice (Figures 6A and 6B). Importantly, in skin-scarified mice, while the frequency of a4b7+ OT-I cells had reduced to the baseline level similar to 60 hr post infection, more than 30% of the memory OT-I cells in ILN expressed E-Lig, and a significant portion of memory OTI cells in MLN and spleen were also E-Lig+ (Figures 6C and 6D). In both ILN and MLN, the frequency of memory

B6 mice were immunized with rVV-ova either by skin scarification or i.p. injection. 4 weeks later, the mice were challenged with 2 3 106 pfu rVV-ova by skin scarification. A group of unimmunized mice were also infected at the same time. 6 days after the viral challenge, skin samples were harvested from the inoculated area and virus loads were determined by quantitative real-time PCR. Data represent the average 6 SD of four to seven mice in one experiment.

cells with skin-homing markers is significantly higher than that of gut-homing memory cells (Figure 6D). In contrast, in i.p. infected mice, significant portions of memory OT-I cells expressed a4b7 in all SLO (Figure 6D). Therefore, it appears that the initial, rather than the secondary, imprinting of tissue-selective homing molecules seen on activated CD8+ T cells has become predominant on memory CD8+ T cells present throughout SLO long after the viral clearance. Skin Scarification Provides Superior Protection against Secondary Cutaneous Viral Challenge than i.p. Immunization We have shown that skin- or gut-homing molecules are preferentially maintained on memory CD8+ T cells in scarified or i.p. immunized mice, respectively. To be meaningful, however, it is important to determine whether this leads to more effective protection of the previously exposed tissues against secondary challenge. To address this question, mice immunized previously with rVV-ova either by skin scarification or i.p. injection were challenged with rVV-ova infection by skin scarification. A group of unimmunized mice were also infected as controls. At day 6 after challenge, the skin samples were harvested from the VV inoculation site and viral DNA copy number was measured by realtime PCR. As shown in Figure 7, viral DNA was undetectable at the skin site of secondary challenge in all the mice previously immunized by skin scarification, while significant viral DNA could be detected at the skin challenge site in 4 out of 6 mice previously immunized via an intraperitoneal route. As expected, samples from unimmunized mice had the highest viral load (more than 2 logs higher than that of i.p. immunized group). These results demonstrate that immunization via skin scarification provides demonstrably superior protection

CD8+ T Cell Trafficking after Viral Infection 517

against secondary cutaneous challenge as compared to i.p. immunization.

Discussion Trafficking of activated T cells among different body compartments is central for the host immune response against microbial infection. Recent studies have demonstrated that virus-specific effector and memory T cells are present ubiquitously in a wide range of host tissues regardless of the route of viral infection (Masopust et al., 2001, 2004). It is unclear when and how these cells acquire the ability to enter multiple tissues. In the current study, by using VV infection established at two distinct sites, we demonstrate a multiphasic regulation of T cell trafficking during the immune response. After VV skin scarification or i.p. infection, CD8+ T cells are activated and begin to proliferate within lymphoid tissues draining the infected tissues. Along with cell division, there is convincing upregulation of E-Lig and P-Lig, but not a4b7, on OT-I cells in skin-draining LN of skinscarified mice, while i.p. infection results in upregulation of a4b7 but not E-Lig or P-Lig on OT-I cells in the visceral peritoneal-draining MLN. T cells activated and educated in these LN appear to specifically migrate into the infected extralymphoid tissues as early as 60 hr after viral inoculation. Activated OT-I cells also migrated into distant antigen-free lymphoid tissues, likely after three cell divisions, and continued to proliferate. This is consistent with the previous report that naive CD8+ T cells are activated by antigenic encounter, and that they will continue to divide and differentiate in the absence of further antigenic stimulation (Kaech and Ahmed, 2001). Remarkably, by day 5 post infection, they had begun to acquire additional homing receptors that appeared to be imprinted in their new LN microenvironment. That is, T cells originally activated in ILN of skin-scarified mice expressed a4b7 after entering MLN; these cells could then traffic back to ILN through blood. This was not an artifactual consequence of the high number of OT-1 cells present, since these findings are held true even when small number of (5 3 103) OT-I cells were adoptively transferred into hosts. Migration of ILN-activated OT-I cells to MLN and the subsequent acquisition of a4b7 could be blocked by FTY720 treatment, which prevented activated T cells from leaving ILN. The observation that CD62Llow OT-I cells entered LN efficiently, as shown by us as well as others (Kassiotis and Stockinger, 2004; Marzo et al., 2005), is surprising since CD62L is thought to be required for T cell LN homing. The mechanism for CD62L-independent LN homing of activated CD8+ T cells remains to be elucidated. To our knowledge, this study represents the first in vivo evidence suggesting that after localized infection, while T cell priming occurs rapidly in LN draining the infected tissues, multiple homing-imprinting programs can be acquired by activated T cells circulating through different LN microenvironments, a process independent of antigen presentation. Therefore, T cell priming and homing receptor acquisition are sequential, but not obligately coupled, independent processes that can occur in remote anatomical locations.

The current paradigm of homing imprinting emphasizes the role of DC in determining the T cell trafficking tropism. When stimulated in vitro with antigen-pulsed DC isolated from MLN or PP, CD8+ T cells upregulated a4b7 and CCR9; in contrast, activated T cells expressed P-Lig and E-Lig when cultured with DC from skindraining LN (Johansson-Lindbom et al., 2003; Mora et al., 2003; Stagg et al., 2002). These studies suggest that homing signals are provided by tissue-specific DC during antigen presentation. However, based on our observation that gut-homing molecules can be induced on activated CD8+ T cells migrating from ILN to antigen-free MLN, we propose that homing signals produced by tissue-specific DC and stromal cells may be present constitutively in regional LN. Thus, they can act in trans on circulating T cells that were recently activated through the TCR elsewhere, allowing multiple homing imprinting programs to be received by the same T cells at different times. For example, it was recently discovered that retinoic acid produced by intestinal DC signals activated T cells for a4b7 and CCR9 expression (Iwata et al., 2004). The presence of higher levels of retinoic acid in MLN may contribute to the secondary gut-homing molecule expression on ILN-stimulated T cells. In support of this, it was demonstrated that after injection of DC via different routes, T cells activated in regional LN expressed homing receptors specific for the injected tissues, irrespective of the DC origin (Dudda et al., 2005). Imprinting of multiple homing pathways on activated T cells has been reported recently, but in a distinct experimental setting (Calzascia et al., 2005). When mice were injected with tumor cells subcutaneously and intraperitoneally at the same time, CD8+ T cells with either skin- or gut-homing phenotype (but not both) were found in skin-draining LN 4 days after tumor implantation. The authors postulated that CD8+ T cells specific for the i.p. implanted tumors were imprinted with guthoming phenotype in skin-draining LN and concluded that the site of antigen entry, rather than the identity of LN where T cell are activated, is responsible for the homing phenotype. Our findings in the present study allow for a different interpretation of these data. While skinhoming OT-I cells were present in MLN at day 5 of VV skin scarification, these cells had clearly been activated originally in ILN draining the scarification site. Therefore, it is likely that the gut-homing CD8+ T cells found in skindraining LN in the tumor model were originally activated in gut-draining LN and had migrated into skin-homing LN by day 4 of tumor injection. These cells maintained their initial gut-homing imprinting in the skin-draining LN, similar to what we have observed. Our observation that at 60 hr after i.p. infection, OT-I cells activated in ILN upregulated E-Lig but not a4b7 further demonstrates that the homing phenotype of activated T cells are regulated by LN-specific factors rather than the route of antigen administration. However, unlike the OT-I cells in our study, the tumor-specific T cells did not appear to acquire secondary homing molecules in the new LN microenvironment. The reasons underlying these different observations are unclear. It is possible that different time points used yielded different results. It is also possible that the innate immune systemic inflammatory reactions associated with VV skin scarification may contribute to the secondary homing imprinting

Immunity 518

in distant LN by increasing the production of homing signaling molecules. Although trafficking of activated OT-I cells at early stage of primary response correlated well with their homing phenotype, we did not observe preferential homing of E-Lig+ or a4b7+ effector OT-I cells into skinor gut-draining LN as reported previously (Mora et al., 2005). Instead, the percentages of E-Lig+ and a4b7+ OT-I cells did not differ in various LN, suggesting that different T cell subsets with heterogeneous homing phenotypes enter all LN equally well. However, the percentages of homing-polarized (E-Lig+ or/and a4b7+) OT-I cells were significantly higher in spleen than in LN, suggesting relative preference of unpolarized T cells for LN homing. By day 30 post infection, the percentages of E-Lig+ or a4b7+ memory OT-I cells became significantly lower than those of primary OT-I effector cells. This may be due to the preferential homing and accumulation of these polarized cells into peripheral tissues as tissuespecific Tem. Recent studies with parabiotic mouse model demonstrated that memory T cells are highly mobile (Klonowski et al., 2004). They migrate from immune mice to their parabiotic naive partners and reach equilibrium in all lymphoid and most peripheral tissues. Similarly, we observed that while the OT-I cells were initially activated and expanded only in regional LN at 60 hr after viral inoculation, their frequencies eventually reached equilibrium among different LN. Therefore, memory T cell migration is likely under homeostatic control to keep the memory pool balanced among various tissues. Interestingly, the predominant homing phenotype on the memory OT-I cells is the one imprinted during T cell priming. The same finding was recapitulated in the low CTL precursor frequency setting. Our data suggest that the cellular immune system not only remembers what antigens it has encountered, but also where the antigens were introduced. Since we have analyzed only memory cells in lymphoid tissues, these results have not taken nonlymphoid tissues Tem cells into account. Whether the initial homing imprinting is maintained on the extralymphoid Tem subset remains to be investigated. Finally, we have shown that compared to the unimmunized mice, both i.p. immunized and skin-scarified memory mice were protected against secondary cutaneous challenge. Importantly, at day 6 post challenge, while more than 60% of the i.p. immunized group still had active viral infection at challenged skin area, all the scarified memory mice had completely cleared the virus from the skin, even though the serum anti-VV antibody levels were comparable between the two groups (data not shown). These data strongly suggest that the ability to generate large number of tissue-specific Tem by peripheral immunization correlates with more effective protection of the immunized tissue against secondary challenge. These findings have important implications for the design of vaccination strategies against pathogens that preferentially invade host via certain target tissues. Experimental Procedures Mice, Viruses, and Viral Infection All animal work was in compliance with the guidelines set out by the Center for Animal Resources and Comparative Medicine at Harvard

Medical School (HMS). Thy1.1+ OT-I mice were bred in a biosafety level 1 (BL-1) facility at HMS. 6- to 8-week-old female Thy1.2+ B6J mice were purchased from Jackson Laboratory and housed in a BL-2 facility. rVV-ova and wild-type VV-WR were kind gifts from Dr. Bernald Moss (National Institutes of Health, Bethesda, MD) (Norbury et al., 2002). The virus stocks were expanded and tittered in Hela cells and CV-1 cells (American Tissue Culture Company) by standard procedures (Earl et al., 1998). Mice were infected with the virus either by i.p. injection (2 3 106 pfu in 100 ml of PBS) or by skin scarification (2 3 106 pfu in 5 ml of PBS). For scarification, mice were anesthetized i.p. with 2,2,2 Tribromoethanol (250 mg/kg, Sigma). 5 ml of diluted virus was applied to tail skin 1 cm from the base of the tail. The skin area was then gently scratched 25 times with a 28 1/2 G needle. Adoptive Transfer of OT-I Cells Spleens and LN were harvested from 3- to 4-week-old Thy1.1+ OT-I mice. Red blood cells (RBC) were lysed and single-cell suspension was prepared. CD8+ T cells were isolated with the mouse CD8a+ T cell isolation kit according to the manufacturer’s protocol (Miltenyl Biotec). The purity of the isolated cells was >95% measured by flow cytometry. Isolated CD8+ T cells displayed a naive T cell phenotype (CD62Lhigh, CD44low, CD11aint, E-Lig2, P-Lig2, a4b72, Figure 2A and data not shown). The cells were then labeled with 5 mM CFSE (Molecule Probes) at 37
C for 10 min, washed three times, and intravenously transferred into Thy1.2+ recipient mice (6 3 106 cells per mouse). Antibodies and Flow Cytometry The following mAbs were purchased from BD Pharmingen: FITCconjugated anti-Thy1.2 (53-2.1), PE-conjugated anti-a4b7 (DATK32), PerCP-conjugated anti-Thy1.1 (OX-7), APC-conjugated anti-CD62L (MEL-14), and biotinylated anti-CD11a (2D7). For E-Lig and P-Lig staining, lymphocytes were incubated with 5 mg/mL of CD62E/Fc chimera or CD62P/Fc chimera (R&D systems) in HBSS supplemented with 2 mM calcium, 5% FCS, and 1 mM HEPES buffer for 30 min at 4
C. Cells were then washed and incubated with biotinylated goat F(ab)’;2 anti-human IgG (BD Bioscience) for 30 min at 4
C followed by streptavidine-fluochrome and other cell-surface marker staining. Since selectin-ligand binding was calcium dependent, HBSS buffer supplemented with 5 mM EDTA instead of calcium was used for specificity control for E-Lig and P-Lig staining. Nonspecific binding to CD62E/Fc chimera or CD62P/Fc chimera was lower than 5% of Thy1.1+ T cells (data not shown). Samples were acquired with a FACSCalibur flow cytometer; data were analyzed with Flowjo software (Tree Star). Immunohistochemistry Staining Immunohistochemistry (IHC) was performed with 5 mm thick formalin-fixed, paraffin-embedded tissue sections. In brief, slides were soaked in xylene, passed through graded alcohols, and put in distilled water. Slides were then pretreated with 10 mM citrate (pH 6.0) (Zymed, South San Francisco, CA) in a steam pressure cooker (Decloaking Chamber, BioCare Medical, Walnut Creek, CA) as per the manufacturer’s instructions followed by washing in distilled water. All further steps were performed at room temperature in a hydrated chamber. Slides were pretreated with peroxidase block (DAKO USA, Carpinteria, CA) for 5 min to quench endogenous peroxidase activity. Primary rabbit anti-CD3 antibody (DAKO) was applied at a dilution of 1:250 in DAKO diluent for 1 hr. Slides were washed in 50 mM Tris-Cl (pH 7.4) and detected with anti-rabbit Envision+ kit (DAKO) as per the manufacturer’s instructions. After further washing, immunoperoxidase staining was developed with a DAB chromogen (DAKO) and counterstained with hematoxylin. Real-Time RT-PCR to Detect VV Infection Tissues harvested from rVV-ova-infected mice were immediately snap-frozen in liquid nitrogen. RNA was extracted from these tissues with Trizol reagent (Invitrogen) according to the manufacturer’s instructions. The RNA samples were treated with DNase (Promega) and reverse-transcribed with Oligo dT (Promega) and Omniscript RT Kit (Qiagen) according to the manufacturer’s instructions. The cDNA products were then mixed with Qiagen SYBR Green Reagent, and the primers specific for a subunit of a VV DNA-directed RNA

CD8+ T Cell Trafficking after Viral Infection 519

polymerase were expressed within 2 hr of viral entry (Howell et al., 2004). The real-time PCR was performed with a Bio-Rad iCycler iQ Real-Time PCR Detection System (Bio-Rad Laboratories, Inc. Hercules, CA). Thermal cycle conditions have been described previously (Howell et al., 2004). All PCR reactions included RT minus RNA samples as the negative controls. The CT values for VV gene were normalized to the housekeeper gene cyclophilin A (forward primer, GTGGTCTTTGGGAAGGTGAA; reverse primer, TTACAGGA CATTGCGAGCAG). The level of viral gene expression was normalized to that in the skin samples of uninfected control mice. FTY720 Treatment FTY720 (provided by Dr. V. Brinkmann at Novartis Pharmaceuticals, Basel, Switzerland) was dissolved in distilled water. Mice were injected i.p. daily with FTY720 at dose of 1 mg/kg. Real-Time PCR to Determine Viral Load VV load was evaluated by quantitative real-time PCR. In brief, 6 days after scarification, inoculated skin samples were harvested and DNA was purified with the DNeasy Mini Kit (Qiagen, Valencia, CA) according to the manufacture’s protocol. Real-time PCR was performed with the Bio-Rad iCycler iQ Real-Time PCR Detection System (BioRad Laboratories). The primers and TagMan probe used in the quantitative PCR assay are specific for the ribonucleotide reductase Vvl4L of VV. The sequences are (forward) 50 -GAC ACT CTG GCA GCC GAA AT-30 ; (reverse) 50 -CTG GCG GCT AGA ATG GCA TA-30 ; (probe) 50 -AGC AGC CAC TTG TAC TAC ACA ACA TCC GGA-30 . The probe was 50 -labeled with FAM and 30 -labeled with TAMRA (Applied Biosystems, Foster City, CA). Amplification reactions were performed in a 96-well PCR plate (Bio-Rad Laboratory) in a 20 ml volume containing 23 TaqMan Master Mix (Applied Biosystems), 500 nM forward primer, 500 nM reverse primer, 150 nM probe, and the template DNA. Thermal cycling conditions were 50
C for 2 min and 95
C for 10 min for one cycle, followed by 45 cycles of amplification (94
C for 15 s and 60
C for 1 min). To calculate the virus load, a standard curve was established from DNA of a VV stock with previously determined titer. Corresponding CT values obtained by the real-time PCR methods were plotted on the standard curve to estimate viral load in the skin samples. Statistical Analysis All statistical analyses were made with Microsoft Excel software (Microsoft). Statistical comparisons between data sets were made with a two-tailed homoscedastic Student’s t test. Supplemental Data Seven Supplemental Figures can be found with this article online at http://www.immunity.com/cgi/content/full/25/3/511/DC1/. Acknowledgments We thank Dr. M. Bai (Department of Dermatology, Brigham and Women’s Hospital, Boston, MA) for her help in real-time RT-PCR experiments. We also thank A. Schlesin (CBR Institute of Biomedical Research, Boston, MA) for her review of the manuscript. The work was supported by grant 1U19AI57330 and R01 AI041707 to T.S.K. from the National Institutes of Health and by a Research Career Development Award to L.L. from the Dermatology Foundation. The authors declare that they have no competing financial interests. Received: March 3, 2006 Revised: June 1, 2006 Accepted: June 27, 2006 Published online: September 14, 2006 References Bass, D.M., Mackow, E.R., and Greenberg, H.B. (1990). NS35 and not vp7 is the soluble rotavirus protein which binds to target cells. J. Virol. 64, 322–330. Breman, J.G., and Henderson, D.A. (2002). Diagnosis and management of smallpox. N. Engl. J. Med. 346, 1300–1308.

Calzascia, T., Masson, F., Di Berardino-Besson, W., Contassot, E., Wilmotte, R., Aurrand-Lions, M., Ruegg, C., Dietrich, P.Y., and Walker, P.R. (2005). Homing phenotypes of tumor-specific CD8+ T cells are predetermined at the tumor site by crosspresenting APCs. Immunity 22, 175–184. Campbell, D.J., and Butcher, E.C. (2002). Rapid acquisition of tissue-specific homing phenotypes by CD4(+) T cells activated in cutaneous or mucosal lymphoid tissues. J. Exp. Med. 195, 135–141. Campbell, J.J., Haraldsen, G., Pan, J., Rottman, J., Qin, S., Ponath, P., Andrew, D.P., Warnke, R., Ruffing, N., Kassam, N., et al. (1999). The chemokine receptor CCR4 in vascular recognition by cutaneous but not intestinal memory T cells. Nature 400, 776–780. Dudda, J.C., Simon, J.C., and Martin, S. (2004). Dendritic cell immunization route determines CD8+ T cell trafficking to inflamed skin: role for tissue microenvironment and dendritic cells in establishment of T cell-homing subsets. J. Immunol. 172, 857–863. Dudda, J.C., Lembo, A., Bachtanian, E., Huehn, J., Siewert, C., Hamann, A., Kremmer, E., Forster, R., and Martin, S.F. (2005). Dendritic cells govern induction and reprogramming of polarized tissue-selective homing receptor patterns of T cells: important roles for soluble factors and tissue microenvironments. Eur. J. Immunol. 35, 1056– 1065. Earl, P.L., Cooper, N., Wyatt, L.S., Moss, B., and Carroll, M.W. (1998). Preparation of cell cultures and vaccinia virus stocks. In Current Protocols in Molecular Biology, F.M. Ausubel, R. Brent, R.E. Kingston, D.D. Moore, J.G. Seidman, J.A. Smith, and K. Struhl, eds. (Mississauga, Ontario: John Wiley & Sons, Inc.), pp. 16.16.14–16.16.16. Erdmann, I., Scheidegger, E.P., Koch, F.K., Heinzerling, L., Odermatt, B., Burg, G., Lowe, J.B., and Kundig, T.M. (2002). Fucosyltransferase VII-deficient mice with defective E-, P-, and L-selectin ligands show impaired CD4+ and CD8+ T cell migration into the skin, but normal extravasation into visceral organs. J. Immunol. 168, 2139–2146. Fuhlbrigge, R.C., Kieffer, J.D., Armerding, D., and Kupper, T.S. (1997). Cutaneous lymphocyte antigen is a specialized form of PSGL-1 expressed on skin-homing T cells. Nature 389, 978–981. Howell, M.D., Jones, J.F., Kisich, K.O., Streib, J.E., Gallo, R.L., and Leung, D.Y. (2004). Selective killing of vaccinia virus by LL-37: implications for eczema vaccinatum. J. Immunol. 172, 1763–1767. Iwata, M., Hirakiyama, A., Eshima, Y., Kagechika, H., Kato, C., and Song, S.Y. (2004). Retinoic acid imprints gut-homing specificity on T cells. Immunity 21, 527–538. Johansson-Lindbom, B., Svensson, M., Wurbel, M.A., Malissen, B., Marquez, G., and Agace, W. (2003). Selective generation of gut tropic T cells in gut-associated lymphoid tissue (GALT): requirement for GALT dendritic cells and adjuvant. J. Exp. Med. 198, 963–969. Kaech, S.M., and Ahmed, R. (2001). Memory CD8+ T cell differentiation: initial antigen encounter triggers a developmental program in naive cells. Nat. Immunol. 2, 415–422. Kassiotis, G., and Stockinger, B. (2004). Anatomical heterogeneity of memory CD4+ T cells due to reversible adaptation to the microenvironment. J. Immunol. 173, 7292–7298. Klonowski, K.D., Williams, K.J., Marzo, A.L., Blair, D.A., Lingenheld, E.G., and Lefrancois, L. (2004). Dynamics of blood-borne CD8+ memory T cell migration in vivo. Immunity 20, 551–562. Kunkel, E.J., and Butcher, E.C. (2002). Chemokines and the tissuespecific migration of lymphocytes. Immunity 16, 1–4. Kunkel, E.J., Campbell, J.J., Haraldsen, G., Pan, J., Boisvert, J., Roberts, A.I., Ebert, E.C., Vierra, M.A., Goodman, S.B., Genovese, M.C., et al. (2000). Lymphocyte CC chemokine receptor 9 and epithelial thymus-expressed chemokine (TECK) expression distinguish the small intestinal immune compartment: epithelial expression of tissue-specific chemokines as an organizing principle in regional immunity. J. Exp. Med. 192, 761–768. Lane, J.M. (1977). Smallpox vaccination. JAMA 237, 2419–2420. Marzo, A.L., Klonowski, K.D., Le Bon, A., Borrow, P., Tough, D.F., and Lefrancois, L. (2005). Initial T cell frequency dictates memory CD8+ T cell lineage commitment. Nat. Immunol. 6, 793–799.

Immunity 520

Masopust, D., Vezys, V., Marzo, A.L., and Lefrancois, L. (2001). Preferential localization of effector memory cells in nonlymphoid tissue. Science 291, 2413–2417. Masopust, D., Vezys, V., Usherwood, E.J., Cauley, L.S., Olson, S., Marzo, A.L., Ward, R.L., Woodland, D.L., and Lefrancois, L. (2004). Activated primary and memory CD8+ T cells migrate to nonlymphoid tissues regardless of site of activation or tissue of origin. J. Immunol. 172, 4875–4882. Matloubian, M., Lo, C.G., Cinamon, G., Lesneski, M.J., Xu, Y., Brinkmann, V., Allende, M.L., Proia, R.L., and Cyster, J.G. (2004). Lymphocyte egress from thymus and peripheral lymphoid organs is dependent on S1P receptor 1. Nature 427, 355–360. Mora, J.R., Bono, M.R., Manjunath, N., Weninger, W., Cavanagh, L.L., Rosemblatt, M., and Von Andrian, U.H. (2003). Selective imprinting of gut-homing T cells by Peyer’s patch dendritic cells. Nature 424, 88–93. Mora, C.G., Jr., Picarella, D., Briskin, M., Buchanan, N., and von Andrian, U.H. (2005). Reciprocal and dynamic control of CD8+ T cell homing by dendritic cells from skin- and gut-associated lymphoid tissues. J. Exp. Med. 201, 303–316. Norbury, C.C., Malide, D., Gibbs, J.S., Bennink, J.R., and Yewdell, J.W. (2002). Visualizing priming of virus-specific CD8+ T cells by infected dendritic cells in vivo. Nat. Immunol. 3, 265–271. Pinschewer, D.D., Ochsenbein, A.F., Odermatt, B., Brinkmann, V., Hengartner, H., and Zinkernagel, R.M. (2000). FTY720 immunosuppression impairs effector T cell peripheral homing without affecting induction, expansion, and memory. J. Immunol. 164, 5761–5770. Springer, T.A. (1994). Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 76, 301–314. Stagg, A.J., Kamm, M.A., and Knight, S.C. (2002). Intestinal dendritic cells increase T cell expression of alpha4beta7 integrin. Eur. J. Immunol. 32, 1445–1454. Tashiro, M., Yamakawa, M., Tobita, K., Klenk, H.D., Rott, R., and Seto, J.T. (1990). Organ tropism of Sendai virus in mice: proteolytic activation of the fusion glycoprotein in mouse organs and budding site at the bronchial epithelium. J. Virol. 64, 3627–3634. von Andrian, U.H., and Mackay, C.R. (2000). T-cell function and migration. Two sides of the same coin. N. Engl. J. Med. 343, 1020–1034. Warnock, R.A., Askari, S., Butcher, E.C., and von Andrian, U.H. (1998). Molecular mechanisms of lymphocyte homing to peripheral lymph nodes. J. Exp. Med. 187, 205–216. Weninger, W., and von Andrian, U.H. (2003). Chemokine regulation of naive T cell traffic in health and disease. Semin. Immunol. 15, 257–270. Weninger, W., Crowley, M.A., Manjunath, N., and von Andrian, U.H. (2001). Migratory properties of naive, effector, and memory CD8 (+) T cells. J. Exp. Med. 194, 953–966.