The influence of macrophage inflammatory protein-1α on protective immunity mediated by antiviral cytotoxic T cells

Abstract

Macrophage inflammatory protein 1α (MIP-1α), a member of the CC-chemokine subfamily, is known to induce chemotaxis of a variety of cell types in vivo. Although the role of MIP-1α in inflammatory responses generated following primary infection of mice with many different pathogens has been characterized, the influence of this chemokine on the generation of antigen-specific T-cell responses in vivo is less well understood. This is important, as virus-specific CD8⁺ T lymphocytes (CTL) play a crucial role in defence against viral infections, both acutely and in the long term. In this study, we compared the ability of wild-type and MIP-1α-deficient (MIP-1α−/−) mice to mount CTL responses specific for the immunodominant epitope derived from influenza nucleoprotein (NP366–374). Influenza-specific CTL responses were compared with respect to frequency, cytotoxic activity and ability to clear subsequent infections with recombinant vaccinia viruses expressing the influenza NP. The results indicate that antiviral CTL generated in MIP-1α−/− mice are slightly impaired in their ability to protect against a subsequent infection. However, impaired in vivo CTL-mediated antiviral protection was found to be associated with reduced cytotoxicity rather than with a failure of the CTL to migrate to peripheral sites of infection.

Introduction

The immune responses to pathogens, particularly viruses, are becoming increasingly well understood in terms of the cell populations involved in antiviral protection. Molecular characterization of the way in which such subsets collaborate to mediate antiviral protection is, however, currently lacking. It is becoming clear that co-ordination of effective immune responses requires a complex network of interactions involving soluble mediators (such as chemokines and cytokines) that control interactions between different cell types, including activation, migration to sites of infection and effector function.¹

Cytotoxic T lymphocytes (CTL) have been shown to be critical for effective resolution of primary viral infections.^2–4 They are able to migrate out into peripheral tissues directly to the site of infection, and recognize and lyse infected cells before virus particles are released.⁵ Although the mechanisms of recognition of infected cells have been extensively studied in vitro, the rules governing how CD8⁺ antiviral T cells reach sites of infection and mediate viral clearance are not fully understood.

Macrophage inflammatory protein 1α (MIP-1α), a small protein that binds to the receptors CCR5 and CCR1, is a member of the CC-chemokine subfamily and has been shown both in vitro and in vivo to act as a chemoattractant for monocytes, neutrophils and natural killer (NK) cells, as well as antigen-specific T and B cells.^6–8 Studies performed in mice show that the absence of MIP-1α results in delayed clearance of some pathogens, such as the bacterium Listeria monocytogenes⁹ and influenza virus,¹⁰ but not of others, e.g. herpes simplex virus-1 (HSV-1)¹¹ and respiratory syncitial virus (RSV).¹² In most cases, however, the absence of MIP-1α was accompanied by a reduction in the immunopathological consequences of virus infection. Unlike wild-type mice, MIP-1α-deficient mice did not develop myocarditis, pneumonitis or blinding ocular inflammation as a consequence of Coxsackie virus, influenza virus and HSV-1 infection, respectively.^10,11,13 Similarly, following infection of mice with murine cytomegalovirus (MCMV), the absence of MIP-1α was associated with a decrease in liver inflammation owing to a paucity of infiltrating interferon-γ (IFN-γ)-producing NK cells.^14,15 The implication that MIP-1α increases the risk of immune-mediated damage to the host is supported by the observations that MIP-1α deficiency can protect against the development of experimental autoimmune encephalitis (EAE)¹⁶ type II collagen-induced arthritis and graft versus host disease (GVHD) in mice.^17,18 The relative influence of MIP-1α on antigen-specific versus non-specific immune responses was not addressed in the studies described above.

The effect of MIP-1α on antigen-specific CTL was, however, addressed in a report by Cook et al.⁹ This study showed that MIP-1α-deficient CTL, specific for L. monocytogenes, were impaired in their ability to clear the bacterium following their adoptive transfer into infected wild-type C57BL/6 mice. Recruitment of MIP-1α-deficient T cells to the spleens of recipient mice was reduced compared with wild-type CTL, as was their capacity to lyse listeria-infected cells in vitro. These results indicate that production of MIP-1α by CTL is important for their effectiveness in vivo.

During acute infections, multiple inflammatory processes, involving chemokines and cytokines, may contribute to the activation of CTL.¹⁹ During chronic low-level infection, the redistribution of CTL into the periphery is maintained after virus clearance, even when inflammation has waned.²⁰ During rechallenge, CTL may once again require inflammatory signals for acquisition of full effector function in order to effectively mediate virus clearance. Differential expression of chemokine receptors has been used to distinguish subsets of antiviral CTL during chronic viral infection.²¹ These findings imply that memory CTL, which are reactivated upon rechallenge with virus, differ in their chemokine sensitivity compared with CTL activated in response to primary virus infections. The influence of MIP-1α on virus-specific CTL in primary and secondary immune responses has not been assessed in detail.

MIP-1α clearly has multiple effects on the generation of immune responses in vivo. In the present study we used fluorescently labelled major histocompatibility complex (MHC) class I tetramers²² to examine the influence of MIP-1α on the generation of virus-specific CTL responses after primary infection of C57BL/6 mice with influenza and after secondary infections of these mice with recombinant vaccinia virus (rVVs) expressing influenza nucleoprotein (NP). Following primary infection with influenza virus, NP-specific CTL from both wild-type and MIP-1α-deficient (MIP-1α−/−) mice were compared with respect to frequency, cytotoxicity, IFN-γ expression and ability to mediate clearance of rVVs expressing influenza NP. A role for MIP-1α in reactivation and migration of memory NP-specific CTL was also investigated.

Materials and methods

Mice

C57BL/6 (B6) and C57BL/10 (B10) mice were bred under specific-pathogen free conditions at Biomedical Services (Oxford, UK). Rag-deficient mice transgenic for the αβ T-cell receptor (TCR) from the F5 CTL clone (F5 mice) and MIP-1α−/− mice have been described previously.¹⁰ Female mice, 5–8 weeks of age, were used in all experiments. During experimental procedures, mice were housed in conventional facilities. All experiments were carried out in accordance with local ethical guidelines.

Viruses and virus infection

Influenza virus strain E61-13-H17,²³ a gift from Dr Keith Gould, was used in all experiments. Mice were infected intranasally with 20 haemagglutination units (HAU) of influenza virus. rVVs expressing the glycoprotein of lymphocytic choriomeningitis virus (LCMV) (rVVGP)²⁴ and ubiquitin-linked influenza nucleoprotein (rVVFLNP)²⁵ have been described previously. An rVV expressing a fragment of NP corresponding to the immunodominant CTL epitope NP366–374 (rVVNPP) has also been described previously.²⁶ For detection of protective immunity to rVVs, mice were infected intraperitoneally with 2 × 10⁶ plaque-forming units (PFU) of rVVs. Four days, later ovaries were removed and analysed for rVV titres, as described previously.^27,28 Statistical analyses were carried out using Fisher's Exact test.

Generation of MHC class I tetramers

Generation of MHC class I tetramers has been described previously.²² Briefly, monomeric complexes of the murine MHC class I molecule D^b and human β₂ microglobulin were refolded using the NP peptide derived from influenza virus (NP366–374). In order to produce NP tetramers, purified monomers were mixed at a 4 : 1 molar ratio with phycoerythrin-conjugated extravidin (Sigma, St Louis, MO). Approximately 0·5 µg of NP tetramers were used to stain 10⁶ cells.

Cell staining and flow cytometry

In order to identify tetramer-positive cells in the organs of mice infected with influenza virus and, in some cases, rVVs, single-cell suspensions were prepared from the spleens and ovaries of each mouse and 2 × 10⁶ cells were incubated with 0·5 µg of tetramer and 0·5 µg of Tricolor-conjugated, anti-CD8β antibodies (Caltag, Towcester, UK) for 30 min at 37°. The cells were washed twice and resuspended in fluorescence-activated cell sorter (FACS) Fix (wash buffer containing 2% formalin). Tetramer-positive T cells were subsequently identified by flow cytometry using Cellquest software (Becton-Dickinson, San Diego, CA).

CTL assays

Spleen-cell suspensions, prepared from the experimental mice, were plated at 4 × 10⁶ cells/well (24-well plates) in 1 ml of RPMI (Gibco, Paisley, UK) supplemented with 10% fetal calf serum (FCS) (GlobePharm, Surrey, UK), penicillin–streptomycin (Gibco) and β-mercaptoethanol (Sigma) (R10). The cultures were supplemented with 1 ml of peptide-pulsed, irradiated spleen cells at a concentration of 10⁶ cells/ml. Before irradiation, spleen cells were incubated (for 1 hr at 37°) with 100 µl of peptide at a concentration of 10 ng/ml before extensive washing to remove any unbound peptide. After 2 days of culture, cultures were supplemented with interleukin-2 (IL-2) (Peprotech, London, UK) at a final concentration of 10 U/ml. Following a 5-day culture period, the cells were resuspended in 0·7 ml of medium per culture well and threefold dilutions of the effector cells were performed (referred to as dilution of culture). Cytotoxicity assays were carried out as described previously.²⁹ In all assays, ⁵¹Chromium-labelled RMA-S cells, pulsed for 1 hr with 100 µl of 10⁻⁶m peptide before extensive washing, were used as target cells. Specific lysis was determined, as described previously.

Intracellular cytokine staining

Intracellular cytokine staining was carried out by incubating spleen cells for 6 hr in a 24-well plate at a concentration of 10⁶ cells/ml of R10 supplemented with 10 µg/ml brefeldin-A (Sigma) and the influenza peptide, NP366–374, at a final concentration of 10⁻⁶m. Subsequently, the cells were washed and stained with Tricolor-conjugated anti-CD8β antibodies, as described above. After fixing, the cells were permeabilized in phosphate-buffered saline (PBS) containing 0·1% saponin (Sigma) and 1% FCS (Sigma) for 10 min on ice. Permeabilized cells were stained with fluorescein isothiocyanate (FITC)-conjugated anti-IFN-γ antibodies (Pharmingen, UK) for 30 min on ice. The stained cells were washed twice in permeabilization buffer, resuspended in FACS Fix, and analysed by flow cytometry using Cellquest software (Becton-Dickinson).

Results

CD8⁺ T-cell responses following primary infection with influenza are not impaired in MIP-1α-deficient mice

The immunodominant epitope that is recognized by CTL from B6 mice following infection with influenza virus is a D^b-restricted peptide (NP366-374, ASNENMDAM) derived from the nucleoprotein.³⁰ Fluorescently labelled tetramers comprising D^b and the NP peptide were constructed (NP-Tetramers) and tested for their ability to specifically stain NP-specific T cells. The staining pattern of CD8⁺ T cells recovered from the spleens of F5 mice,^31,32 which are transgenic for a TCR that recognizes the NP peptide, was compared with the staining of CD8⁺ T cells recovered from the spleens of non-TCR transgenic mice of the same genetic background (C57BL/10). As shown in Fig. 1, NP-Tetramers stained only CD8⁺ T cells recovered from F5 mice, indicating that the tetramers were indeed NP specific. We next compared the ability of wild-type B6 and MIP-1α−/− mice to respond to infection with influenza virus. Spleen cells recovered from mice infected 8 days previously with influenza virus were stained with NP-Tetramers and tested for their ability to lyse target cells pulsed with the NP-peptide epitope. As shown in Fig. 2(a), the percentage of NP-tetramer-positive cells was not considerably different in wild-type compared with MIP-1α−/− mice. No appreciable differences were observed in the cellularity of the spleens, indicating that the overall numbers of NP-specific CD8⁺ T cells were also similar in both groups of mice (data not shown). As no cytotoxicity was measurable ex vivo, spleen cells from both groups of mice were restimulated in vitro. After one round of restimulation, CTL activity from both groups of mice was found to be comparable (Fig. 2b). Therefore, compared with wild-type mice, MIP-1α deficiency does not impair the ability of mice to mount an antigen-specific CD8⁺ T-cell response to influenza.

Figure 1.

Staining of nucleoprotein (NP)-specific H-2D^b-restricted T-cell receptor (TCR) transgenic T cells. Spleen cells from a naïve B10 mouse (a) and an F5 TCR transgenic mouse (b) were stained with antibodies specific for CD8 and fluorescently labelled NP-Tetramers. In both cases, the percentage of CD8⁺, NP-Tetramer-positive cells is indicated. PE, phycoerythrin conjugate.

Figure 2.

(a) Nucleoprotein (NP)-Tetramer-positive CD8⁺ T cells in the spleens of influenza virus-infected mice. Cells isolated from the spleens of mice infected 8 days previously with influenza virus were stained with antibodies specific for CD8 and with NP-Tetramers. The graphs describe data collected from dot-plots generated as shown in Fig. 1. Black symbols represent individual wild-type (wt) mice whilst open symbols represent individual macrophage inflammatory protein 1α-deficient (MIP-1α−/−) mice. The data are representative of two independent experiments using groups of four mice each time. Solid lines represent the mean within each group. (b) Cytotoxic activity in the spleens of influenza-infected mice. NP-specific cytotoxic T lymphocyte (CTL) activity was measured using spleen cells from the mice (n = 4) in (a). The cells, which had been restimulated in vitro using peptide-pulsed spleen cells, were tested for their ability to lyse RMA-S cells pulsed with the NP peptide. Lysis of RMA-S cells pulsed with GP33-41, an irrelevant H-2D^b-restricted peptide derived from lymphocytic choriomeningitis virus (LCMV), was less than 8% in all cases (data not shown).

NP-specific protective immunity is slightly impaired in mice deficient for MIP-1α

Approximately 8 weeks after infection with influenza, wild-type and MIP-1α−/− mice were challenged with rVVs expressing an irrelevant antigen (rVVG2), a full-length ubiquitin-linked NP (rVVFLNP), or the NP peptide epitope (rVVNPP). Four days after infection with rVV, ovaries of each mouse were harvested and analysed for virus titres. As shown in Fig. 3, mice deficient for MIP-1α were impaired in their ability to clear infection with rVVFLNP and, to a lesser degree, rVVNPP, compared with control wild-type mice. Neither wild-type nor MIP-1α−/− mice were able to clear infection with the control vaccinia virus, rVVG2. Using Fisher's exact test we found that the difference between the ability of MIP-1α−/− mice and wild-type mice to clear rVVNP was statistically significant (P = 0·0195). In the case of rVVNPP, the peptide epitope is generated from a minigene and requires no further intracellular processing for presentation by MHC class I. This may result in more efficient presentation of the NP peptide in cells infected with rVVNPP than rVVFLNP, thereby rendering infected target cells more sensitive to lysis by NP-specific CTL. All MIP-1α−/− mice cleared rVVNPP and rVVFLNP infections by day 7 after virus challenge (data not shown), thereby further indicating that MIP-1α plays only a minor role in clearing secondary virus infections.

Figure 3.

Protection against recombinant vaccinia virus (rVV) in influenza-primed mice. Ovaries were removed from mice that had been infected approximately 8 weeks previously with influenza virus and challenged 4 days previously with rVVs expressing either an irrelevant antigen (rVVG2), full-length ubiquitin-linked nucleoprotein (NP) (rVVFLNP) or the NP peptide epitope (rVVNPP). Black symbols represent the number of plaque-forming units (PFU) of rVV in the ovaries of individual wild-type mice, whilst open symbols represent the individual number of PFUs in the ovaries of individual macrophage inflammatory protein 1α-deficient (MIP-1α−/−) mice. The dashed line represents the limit of detection of the assay. The data are representative of three independent experiments using groups of four experimental mice each time.

Impaired protective immunity in MIP-1α-deficient mice is not associated with a failure of antigen-specific CD8⁺ cells to migrate to peripheral sites of infection

We hypothesized that the reduced ability of influenza-primed MIP-1α−/− mice to clear infections with rVVFLNP and rVVNPP was the result of impaired migration of NP-specific CTL to the ovaries of the infected mice. We first found that NP-Tetramers could be used to detect NP-specific CD8⁺ T cells in the ovaries of influenza-primed wild-type mice that were subsequently infected with rVVNPP, but not rVVG2 (Fig. 4a). NP-Tetramers were subsequently used to stain ovary cell suspensions recovered from influenza-primed wild-type and MIP-1α−/− mice that had been challenged with rVVFLNP or rVVNPP. The results of this experiment indicated that the percentage and numbers (data not shown) of NP-specific CTL were similar in both wild-type and MIP-1α−/− mice (Fig. 4b). Lungs of the mice described above were also analysed for the presence of NP-specific CD8⁺ T cells. No appreciable differences were observed in the percentages of NP-Tetramer positive cells isolated from wild-type and MIP-1α−/− mice (data not shown). Therefore, a failure to migrate to peripheral organs does not appear to account for the inability of MIP-1α−/− mice to clear infection with rVV.

Figure 4.

Staining of nucleoprotein (NP)-specific CD8⁺ T cells in the ovaries and spleens of influenza-primed, recombinant vaccinia virus (rVV)-challenged mice. (a) Single-cell suspensions of ovaries prepared from wild-type B6 mice infected 4 days previously with rVVs expressing either an irrelevant antigen (rVVG2) or the NP peptide epitope (rVVNPP), approximately 8 weeks following infection with influenza virus, were stained with CD8-specific antibodies and NP-Tetramers. In both cases, the percentage of NP-Tetramer-positive, CD8⁺ T cells is shown. (b) The percentage of NP-Tetramer-positive CD8⁺ cells in the ovaries of mice described in Fig. 3. (c) The percentage of NP-Tetramer-positive CD8⁺ cells in the spleens of mice described in Fig. 3. In both (b) and (c), the graphs describe data collected from dot-plots generated as shown in Fig. 1. Closed symbols represent individual wild-type mice, whilst open symbols represent indivdiual macrophage inflammatory protein 1α-deficient (MIP-1α−/−) mice. Solid lines represent the mean within each group. The data are representative of three independent experiments using groups of four experimental mice each time. rVVFLNP, full-length ubiquitin-linked NP. PE, phycoerythrin conjugate.

Failure of MIP-1α-deficient mice to clear a secondary infection is associated with impaired cytotoxic activity of antigen-specific CTL

We next compared the functional activity of NP-specific CD8⁺ T cells in the spleens of influenza-primed wild-type and MIP-1α−/− mice 4 days after secondary infection with rVVFLNP or rVVNPP. As IFN-γ, but not CTL-mediated cytotoxicity, has previously been shown to be critical for clearing VV infections,³³ we first compared IFN-γ expression by NP-specific CD8⁺ T cells 4 days after secondary infection of either wild-type or MIP-1α−/− mice with rVVNPP or rVVFLNP. As shown in Fig. 5, the ability to express IFN-γex vivo, measured by intracellular staining following stimulation of spleen cells with the NP peptide, was comparable in CD8⁺ T cells recovered from wild-type and MIP-1α−/− mice. Unlike IFN-γ, which has a profound effect on the control of VV infections, perforin is not essential for clearing VV infections. Perforin-deficient splenocytes have, however, been found to be slightly less efficient than wild-type cells in clearing virus from the lungs of VV-infected mice.³⁴ CTL activity was therefore compared in both wild-type and MIP-1α−/− mice. As no ex vivo CTL activity could be observed in either group of mice, spleen cells were restimulated in vitro using peptide-pulsed spleen cells. Five days after restimulation, NP-specific cytotoxicity, as well as the percentages of NP-Tetramer-positive cells in spleen-cell cultures from each mouse, was measured. Figure 6 shows a decrease in the capacity of cells from MIP-1α−/− mice to lyse peptide-pulsed target cells compared to cells recovered from wild-type mice. The percentages of NP-Tetramer-positive cells in the cultures used in the cytotoxicity assays were similar in wild-type (rVVFLNP, 30% ± 11%; and rVVNPP, 27·2% ± 16%) and MIP-1α−/− (rVVFLNP, 26·1% ± 11·4%; and rVVNPP, 25·3% ± 11·4%) mice, indicating that the differences observed in lysis could not be attributed to different numbers of NP-Tetramer-positive cells. These results indicate that virus-specific memory CD8⁺ T cells in MIP-1α−/− mice exhibit impaired cytotoxic activity upon reactivation, which impinges slightly on their ability to clear secondary virus infections.

Figure 5.

Intracellular cytokine staining of CD8⁺ cells isolated from the spleens of influenza-primed, recombinant vaccinia virus (rVV)-infected mice. Spleen cells recovered from the mice described in Figs 3–5 were stimulated in vitro with the nucleoprotein (NP) peptide, NP366–374, and the CD8⁺ cells were subsequently analysed for inteferon-γ (IFN-γ) expression, as described in the Materials and methods. Each symbol represents the percentage of IFN-γ-positive CD8⁺ cells in the spleen of an individual mouse. Closed symbols represent individual wild-type mice, whilst open symbols represent individual MIP-1α−/− mice. Solid lines represent the mean within each group.

Figure 6.

Cytotoxic activity in the spleens of influenza-primed, recombinant vaccinia virus (rVV)-infected mice. Nucleoprotein (NP)-specific cytotoxic T lymphocyte (CTL) activity was measured using spleen cells from the mice described in Figs 3 and 4: (a) influenza-memory wild-type mice rechallenged with rVVs expressing the full-length ubiquitin-linked NP (rVVFLNP); (b) influenza-memory macrophage inflammatory protein 1α-deficient (MIP-1α−/−) mice rechallenged with rVVFLNP; (c) influenza-memory wild-type mice re-challenged with rVVs expressing the NP peptide epitope (rVVNPP); and (d) influenza-memory MIP-1α−/− mice re-challenged with rVVNPP. The cells that had been restimulated in vitro using peptide-pulsed spleen cells, were tested for their ability to lyse RMA-S cells pulsed with the NP peptide using a standard chromium-release assay. Lysis of target cells pulsed with an irrelevant H-2D^b-restricted peptide derived from lymphocytic choriomeningitis virus (LCMV) was less than 10% in all CTL lines tested (data not shown). The data are representative of three independent experiments using groups of four experimental mice each time.

Discussion

The central role of chemokines in directing the migration of cells involved in innate and adaptive immunity is becoming increasingly well understood.³⁵ As a large range of chemokines are secreted and a variety of receptors are shared, it might have been assumed that impairment or inhibition of a single mediator would only have a modest effect on the in vivo function of lymphocytes. However, previous experiments using mice deficient in MIP-1α have revealed important effects of this chemokine on pathogen control and the development of immunopathology.¹⁰ Many experiments have shown that MIP-1α-deficient mice exhibit delayed clearance of individual pathogens and a reduction in immune-mediated tissue damage compared with wild-type mice. Some investigators have suggested that, in addition to defects in migration, impairment of effector function may contribute to the in vivo effects of MIP-1α deficiency.^9,36

We examined the effect of MIP-1α deficiency on CD8⁺ T cells in a model infection system where the frequency, function and distribution of peptide-specific CTL populations could be readily measured. In particular, we were able to examine the effect of MIP-1α deficiency on secondary antigen-specific CTL responses. We identified a slight impairment of memory, influenza NP-specific CTL to control a subsequent infection with rVVs encoding either full-length NP (rVVFLNP) or a peptide epitope derived from NP (rVVNPP). Unexpectedly, we detected no defect in the ability of MIP-1α-deficient NP-specific CTL to migrate to the main site of vaccinia infection (the ovary) after rechallenge. Numbers of NP-specific CTL in both the central compartment (spleen) and peripheral compartment (ovary) did not differ between wild-type and MIP-1α−/− mice. No differences were observed in the ability of NP-specific CD8⁺ T cells, recovered from wild-type or MIP-1α−/− mice, to express IFN-γ. This result was not surprising given that the absence of IFN-γ, unlike the absence of MIP-1α, drastically impairs the control of VV infection in mice.³⁴ In our experiments, which were controlled for cell number, a reproducible deficit in cytotoxic activity was observed in spleen cultures from MIP-1α-deficient mice. We do not know whether the impact of MIP-1α deficiency on the cytotoxic activity of the NP-specific CD8⁺ T cells is attributable to a lack of MIP-1α production by the T cells themselves, the ovaries or other sites of rVV infection, such as the lungs or lymphoid tissue. Interestingly, no difference in cytotoxicity was observed in CTL cultures generated 8 days following primary infection of wild-type and MIP-1α−/− mice with influenza virus. In this situation, a spectrum of proinflammatory mediators, induced in response to primary influenza infection, may contribute to activation of NP-specific CTL and compensate for MIP-1α deficiency. Such compensation may not be offered in the case of secondary infections when rapid control of virus by memory cells limits virus-induced proinflammatory stimuli. It is therefore possible that the effect of MIP-1α on antigen-specific CTL activity is more apparent during memory responses when non-specific inflammatory activity is reduced. Overall, these results imply that fully activated MIP-1α−/− CTLs are not impaired in their ability to lyse target cells and that the deficiency observed in cytotoxic activity is not absolute, but rather relative and dependent upon the maturation or activation status of the cell population.

CTL are a major source of the production of CC-chemokines following encounter with antigen.⁹ CC-chemokines have been shown to regulate adherence of T cells to vascular endothelium and to target immune responses to sites of antigen density through chemoattraction of multiple immune-effector cells.^7,37 It is also known that localization of antigen-specific T cells to the site of infection is a key determinant of effective immunity against intracellular pathogens.^38,39 Several studies indicate that the migration of CTL to sites of infection is indeed impaired in MIP-1α−/− mice. In this study, however, the reduced ability of influenza-primed MIP-1α−/− mice to clear a second infection with rVV was not associated with fewer CTL at the main site of virus infection (ovaries) compared with wild-type mice.

CC-chemokines, including MIP-1α, have also been shown to participate in T-cell activation. MIP-1α has been shown to enhance T-cell proliferation and IL-2 production in a donor-dependent manner and to enhance the cytotoxic activity of CTL and NK cells.^40,41 A separate study indicated a role for the CC-chemokine, regulated on activation, normal, T-cell expressed, and secreted (RANTES), but not MIP-1α, on the cytotoxic activity of human immunodeficiency virus (HIV)-specific CTL clones.⁴² These findings, together with the results presented in this study, imply that the effect of MIP-1α on cytotoxic activity is relative rather than absolute and may depend upon the presence or absence of other chemokines and inflammatory mediators. In this in vivo study, the reduced ability of influenza-primed MIP-1α−/− mice to clear a second infection with rVV was associated with impaired cytotoxicity of antigen-specific CTL compared with wild-type mice. This effect of MIP-1α was not absolute because cytotoxic activity, although reduced, was not abolished.

In conclusion, we have demonstrated that MIP-1α plays a small role in the ability of CTL to protect against a secondary viral challenge. This role appears to relate less to the ability of protective CTL to migrate to the site of infection and more to their functional capacity. Sallusto and colleagues have previously shown that memory cells with distinct phenotypic and functional characteristics can be distinguished by expression of the chemokine receptor CCR7.²¹ Memory cells that are CCR7⁺ (and known as central memory cells) lack effector function but differentiate into CCR7⁻ cells upon reactivation with antigen; CCR7⁻ cells, known as effector memory cells, rapidly express effector cytokines and, in the case of CTL, perforin. These studies suggest that chemokines play an important role in defining subsets of memory T cells. As MIP-1α−/− mice clear secondary virus infections less rapidly than wild-type mice, it is possible that MIP-1α, which binds to the chemokine receptors CCR1 and CCR5, itself impinges upon the development of effector memory versus central memory cells.

Abbreviations

B6

C57BL/6

CTL

cytotoxic T lymphocyte

MIP

macrophage inflammatory protein

IFN

interferon

IL

interleukin

NK

natural killer

NP

nucleoprotein

TCR

T-cell receptor

rVV

recombinant vaccinia virus