Function and Regulation of Natural Killer (NK) Cells during Viral Infections: Characterization of Responsesin Vivo

METHODS: A Companion to Methods in Enzymology 9, 379 –393 (1996) Article No. 0043

Function and Regulation of Natural Killer (NK) Cells during Viral Infections: Characterization of Responses in Vivo1 Christine A. Biron,2 Helen C. Su, and Jordan S. Orange Division of Biology and Medicine, Brown University, Providence, Rhode Island 02912

Although in vitro systems have provided important information about the composition and nature of various immune responses, understanding physiologically relevant function and regulation requires evaluating in vivo conditions. Two different models of acute viral infections have made possible the characterization of a variety of responses to these agents, including natural killer (NK) cell activation and regulation during infection; these are mouse infections with lymphocytic choriomeningitis virus (LCMV) and murine cytomegalovirus (MCMV). The results of our characterization of the NK cell responses elicited in these models and the methods used to dissect the regulation of these responses are reviewed here. Cytotoxicity, proliferation, and cytokine expression assays as well as flow cytometric analyses are used to characterize the in vivo responses. Both of the infections induce early NK cell cytotoxicity and blastogenesis. Infection with MCMV but not LCMV also induces NK cell production of the antiviral cytokine, interferon-g (IFN-g). Antibodies, to mediate in vivo cell subset depletion or cytokine neutralization, and mice, genetically altered to have cell subset or cytokine deficiencies, are utilized to identify the regulatory pathways and mechanisms controlling endogenous NK cell responses to the infections. The major mediators of the regulation of NK cell function during viral infection of normal mice are IFN-a/b and interleukin-12 but not interleukin-2. Furthermore, the induction of later T-cell responses contributes to the negative regulation of NK cells by promoting the production of inhibitory factors including biologically active transforming growth factors-b. Thus, the study of immune responses to viral infections has provided and will continue to provide important insights into the characteristics of endogenous NK cell responses and the cells and factors that regulate these responses in vivo. q 1996 Academic Press, Inc.

1

The work in this laboratory is supported by National Institutes of Health Grants RO1-CA41268 and RO1 CA55726. H.C.S. is a HHMI Predoctoral Fellow. 2 To whom correspondence should be addressed at the Division of Biology and Medicine, Box G-B618, Brown University, Providence, RI 02912. Fax: (401) 863-9045.

A role for natural killer (NK) cells in defense against viral infections was first proposed as a result of the observation that these cells could be activated by virus-induced interferons-a/b (IFN-a/b) to mediate lysis of virus-infected target cells (1–4). It was thought that NK cells might reduce viral replication by killing infected host cells. The characterization of NK cell activation to produce certain antiviral cytokines, i.e., interferon-g (IFNg) and tumor necrosis factor-a (TNF-a) (5, 6), suggested that these cells may also mediate antiviral defense through the production of cytokines inhibiting viral replication. Furthermore, as the NK cell-produced cytokines can facilitate antigen processing and presentation by inducing increased expression of histocompatibility molecules (7, 8) and activation of antigen-presenting cells (8, 9), they might additionally contribute to antiviral defense by using these pathways to promote adaptive immune responses. Thus, NK cells are equipped to access at least three different antiviral defense mechanisms and may use these either independently or synergistically to promote antiviral states in the host. Evidence supporting the actual in vivo function of NK cells in defense against viral infections has been accumulating over time. Although not definitive, the correlative data on natural infections of humans are compelling. Human NK cell deficiencies are associated with increased sensitivity to the herpes group viruses. Individuals with reduced NK cell functions have poor defense against herpes simplex virus (HSV) (10) and Epstein–Barr virus (EBV) (11) infections. Bone marrow transplant recipients slow to reconstitute NK cells demonstrate increased susceptibility to human cytomegalovirus (HCMV) and succumb more readily to this infection (12). Furthermore, a case report of a patient previously identified with a complete lack of NK cells and NK cell functions has documented the consequence of an unusual and life threatening primary HCMV infection in the absence of these cells (13). The extreme sensitivity to acute primary viral infections in the absence of NK cells is also apparent during varicella379

1046-2023/96 $18.00 Copyright q 1996 by Academic Press, Inc. All rights of reproduction in any form reserved.

/ 6706$$357b

04-16-96 00:24:02

metha

AP: Methods

BIRON, SU, AND ORANGE

380

zoster virus and HSV infections (13). Studies in experimental infections of mice have provided definitive data that NK cells mediate resistance to viruses. Sensitivity to murine cytomegalovirus (MCMV) infection correlates with low endogenous NK cell activity in particular strains of mice (14, 15), and depletion of NK cells by in vivo antibody treatment reduces resistance against MCMV (16, 17), HSV (18), and influenza (19). Furthermore, protection against MCMV can be adoptively transferred with purified NK cells or cloned cells mediating NK-like activity (20). Interestingly, although NK cells are activated to mediate elevated levels of lysis during lymphocytic choriomeningitis virus (LCMV) infection (3), it has not been possible to demonstrate an antiviral effect of endogenous NK cells against this agent (16). Thus, NK cells clearly contribute to protection against certain but not all viral infections. Promoting or calling forth NK cell-mediated defenses against infectious challenge might be facilitated by inducing elevated total NK cell function(s) in vivo. However, as NK cells also have the potential to damage the host, curtailing or limiting the magnitude of their responses will most likely be important as well. NK cells can be regulated at the levels of cell localization, total cell proliferation, cytotoxic activity, and cytokine production. There are a number of cells and factors that have been shown to control these parameters in both humans and mice. Work in culture and in vivo has clearly demonstrated that IFN-a, IFN-b, and/or IFN-g all induce NK cell cytolytic activity (1–3, 5, 21, 22), and in vivo studies have shown that IFNs and IFN inducers such as polyinosinicpolycytidylic acid (polyI:C) elicit profound cellular localization and trafficking changes in vivo (23–25). IFNs do not, however, maintain NK cell proliferation in culture. Interleukin-2 (IL-2) is the best growth factor for NK cells characterized to date. IL-2 readily activates NK cell cytolytic activity and, at high concentrations, supports NK cell proliferation in culture (26, 27). Interleukin-12 (IL12) is a potent inducer of IFN-g production by NK cells. This factor can also activate the cytotoxicity mediated by and support the proliferation of NK cells (28, 29). In contrast, transforming growth factors-b (TGF-bs), interleukin-4 (IL-4), and interleukin-10 (IL-10) have been shown to be negative regulators of NK cell responses. Although the precise mechanisms and range of activities mediated by each of the factors have not been thoroughly characterized in both human and mouse, TGF-bs are clearly inhibitors of NK cell proliferation (30, 31) and IFN-g production (31, 32), IL-10 has been shown to inhibit indirectly IFN-g expression by NK cells (33, 34), and IL-4 can inhibit the IL-2-driven proliferation and activation of mature human NK cells (35). Despite the fact that a role for NK cells in mediating resistance to viral infections has been clearly delineated, in vivo NK cell regulation and function are only beginning to be understood. A thorough characteriza-

/ 6706$$357b

04-16-96 00:24:02

metha

tion of the mechanisms by which NK cells may mediate antiviral defense as well as the pathways regulating NK cells is dependent upon examining responses in the context of intact viral infections. Two different mouse infections have been particularly informative in regard to characterizing the generation and regulation of NK cell responses. They are acute primary infections with LCMV and MCMV (Table 1). In this article, we present the methods for characterizing NK cell responses to viral infections and investigating the regulation of NK cell responses in the context of these models. The knowledge resulting from the experiments carried out to date in these models is also reviewed.

DESCRIPTION OF METHODS Animal Protocols Mice The following specific pathogen-free mouse strains are purchased from commercial vendors: C3H/HeNTacBR (H2k) and C57BL/6TacfBR (H-2b) mice from Taconic Laboratory Animals and Services (Germantown, NY); C57BL/6nu/nu mice from Taconic Laboratory Animals and Services or Jackson Laboratory (Bar Harbor, ME); C.B-17, SCID (C.B-17, scid/scid), and Tac:NIHS-nufDF nu/nu mice from Taconic Laboratory Animals and Services; and CD1 mice from Charles River Laboratories (Wilmington, MA). Mice are 4 to 10 weeks old when purchased and 4 to 20 weeks when used in experiments. The following strains are maintained in breeding facilities at Brown University (Providence, RI): b2-microglobulin-deficient mice originally obtained from Drs. B. H. Koller and O. Smithies, University of North Carolina (Chapel Hill, NC); athymic BALB/cAnBOM-nu/nu mice (H-2d); and IL-2-deficient mice originally obtained from Dr. I. Horak, University of Wu ¨ rzburg (Wu¨rzburg, Germany) (48). b2-microglobulin-deficient mice also lacking IL-2 genes are generated at Beth Israel Research (Boston, MA) in collaboration with Drs. Stephen Simpson and Cox Terhorst (49). Mice are handled in accordance with institutional guidelines for animal care and use. Infections Mice are infected ip with 2.5 1 104 plaque-forming units (PFU) of Armstrong strain lymphocytic choriomeningitis virus (LCMV). Either uncloned Armstrong or clone E350 (50) is used. Mice are infected ip with 1 to 4 1 105 PFU of Smith strain murine cytomegalovirus (MCMV). Virulent MCMV stock is passaged at least twice serially in salivary glands of CD1 or C57BL/6 mice and titrated in vitro for infectious center formation on confluent mouse embryo fibroblast (MEFs) monolayers and in vivo for formation of hepatic necrotic foci and induction of interferong (IFN-g) production (51, 52).

AP: Methods

NATURAL KILLER CELLS AND VIRAL INFECTION

In Vivo Treatments

381

Cell Preparation Isolation of Leukocytes Splenic leukocytes are obtained after teasing apart spleens, passing cells through nylon mesh, and osmotically lysing erythrocytes using ammonium chloride or lysing buffer (0.156 M NH4Cl, 0.010 M Na 2HCO3 , 0.116 mM EDTA) followed by stopping buffer (0.120 M NaCl, 0.015 M sodium citrate). In some experiments, cell populations are characterized by in vitro depletion of NK or T cells using anti-NK1.1 mAb SW3A4 or anti-CD3 mAb 29B followed by mouse anti-rat Ig mAb MARS 18.5 to enhance C* fixation (49). Incubations with each mAb are for 30 min at 47C. Rabbit complement (C*) (Pel-Freez Biologicals, Rogers, AR) is added at a sixfold dilution for IgM mAbs or a threefold dilution for IgG mAbs and incubated for 10 min on ice followed by 1 h at 377C. Peripheral blood leukocytes are prepared from whole blood subjected to multiple rounds of ammonium chloride lysis. For isolation of lymph node leukocytes, inguinal lymph nodes are teased apart without ammonium chloride lysis.

Mice are ip administered 100 mg polyI:C (Sigma Chemical Co., St. Louis, MO) in PBS, once at 24 to 36 h prior to cell harvest (53, 54). Cyclosporin A (CsA) (Sandoz Pharmaceuticals, Hanover, NJ) is given ip in olive oil over 3 days at concentrations inhibiting T-cell responses (49, 55, 56). Hydroxyurea is given in PBS ip at doses of 1 mg/g body weight at 9 and 2 h prior to cell harvest (57). For in vivo subset depletions, the following antibodies are injected ip: rabbit antiserum to asialo ganglio-n-tetraosylceramide (anti-AGM1) (Wako Chemicals, Dallas, TX); the anti-NK1.1 monoclonal antibody (mAb) PK136 (58); anti-CD4 mAb GK1.5.6 against L3T4; anti-CD8 mAb 53-6.72 against Lyt-2; and/or anti-CD8 mAb 2.43 against Lyt-2.2. PK136 and GK1.5.6 are prepared as ascities, and 53-6.72 and 2.43 as hybridoma supernatants (49, 52, 59). Antibody treatments commence at least 2 h prior to infection and are repeated every 2 to 4 days during the course of infection. Cytokines are administered ip as described (30, 52, 53, 60). For in vivo neutralization of cytokines, the following neutralizing antibodies are injected ip: 1 mg of the anti-IFN-g mAb XMG1.2 (52); 1 mg of the anti-IL-12 mAb C17.8 (61; 61a); and/or 150 mg of polyclonal sheep anti-IFN-a/b (62; Cousens, Orange, and Biron, unpublished).

Purification of Lymphocyte Populations Populations containing resting as well as activated cells are enriched by negative depletions using antibody and magnetic beads. Cells are incubated with biotinylated rat-anti-mouse Ig F(ab*)2 , followed by streptavidin-conjugated magnetic beads. Surface Ig/ cells are physically removed using a magnet (63). Alternatively, antibody and C* treatments are carried out to eliminate cell subsets, and viable cells are recovered over Percoll (Pharmacia Fine Chemicals, Uppsala, Sweden) density gradients (see below).

Plasma Collection Mice are anesthetized with methoxyflurane, and blood is harvested from the retroorbital sinus into heparin. After centrifugation, plasma is collected and stored at 0807C until use.

TABLE 1 Characteristics of the Acute Experimental Viral Infections in Mice Lymphocytic choriomeningitis virus (LCMV)

Murine cytomegalovirus (MCMV)

Arenavirus

Herpes group virus

Pathology during infection Direct viral-induced Immunopathology

0 /

/ (?)

Cellular immune response NK cell cytotoxic activity CD8/ T-cell expansion CTL activity

/ / /

/ / //0

0 0// (Long-term protection) /

/ / (Salivary gland clearance) /

(3, 16, 17, 36–43)

(14, 16, 17, 44 –47)

Characteristic Virus family

Cellular immune protection NK cell CD4/ T cell CD8/ T cells References

/ 6706$$357b

04-16-96 00:24:02

metha

AP: Methods

BIRON, SU, AND ORANGE

382

Characterization and Purification of Blast Lymphocyte Populations Cells are separated on the basis of size using a centrifugal elutriator (30, 53 ,54, 64). Deoxyribonuclease treatments are carried out to prevent clumping, and cells are loaded in a Beckman JE-6B elutriation system (Beckman Instruments, Inc., Palo Alto, CA) spinning at a constant 3200 rpm at 5–107C. Cell fractions are collected at increasing flow rates by eluting 200–250 ml Hanks’ balanced salt solution containing 1.5% FBS. Fractions 1 through 6 are eluted at 15, 22, 28, 32, 38, and 46 ml/min, respectively. Under these conditions, fractions 4 through 6 are enriched three- to ninefold for [3 H]thymidine incorporating lymphocytes. Blast lymphocytes in fractions 4 through 6 can be further enriched by antibody and complement depletions, using one or a combination of the following mAbs: Jlld, specific for B cells and PMN; B23.1, specific for monocytes and macrophages; and in some cases, 29B, specific for T cells, followed by mouse anti-rat Ig, MARS 18.5, to enhance C* fixation. Following C* treatment, viable cells are recovered by centrifugation through 35% Percoll. Alternatively or additionally, blast-sized lymphocyte populations are isolated on the basis of their low density. Following antibody and C* depletion, dead cells and high-density cells are separated on multi-step Percoll gradients (54, 57). Cell and Tissue Histological Analysis Cells are spun onto microscope slides using a cytocentrifuge (Shandon Southern Instruments, Inc., Sewickley, PA). After air drying, cells are stained with Wright’s– Giemsa, and differential morphology is calculated on the basis of 200–400 counted cells (55). Tissues are fixed in 10% Bouin’s fixative and/or neutral buffered formalin, processed in an automated tissue processor, and paraffin embedded. Sections, 5 mm, are cut for staining with hematoxylin and eosin (H & E) (25, 65). Cytotoxicity Assays NK cell lytic activity is analyzed as cytotoxicity against the NK-sensitive target cell line, YAC-1 (64). CTL activity is determined as virus-specific lysis of the histocompatible cell lines L929 (H-2k), MC57G (H-2b ), and 3T3 (H-2d), where specific lysis of virus-infected cells is calculated as the percentage lysis of virus-infected cells minus the percentage lysis of uninfected cells (59, 66). CD3-dependent lytic activity is determined as percentage lysis of the P815 target cell line in the presence of anti-CD3 145-2C11 mAbs minus percentage lysis of P815 cells in the absence of anti-CD3 mAbs (49). Microcytotoxicity assays are performed in RPMI medium containing 0.01 M Hepes, 10% FBS, antibiotics, and glutamine. Target cells are labeled with sodium 51 Cr (ICN Radiochemicals, Irvine, CA) for 1 h

/ 6706$$357b

04-16-96 00:24:02

metha

at a concentration of 100 mCi/106 cells, incubated in medium for 1 h, and washed 31 before use. Target cells are incubated with splenic effector cells in a total volume of 0.2 ml per well for 4.5 to 6 h at 377C in microtiter plates. Before harvest, plates are centrifuged at 200g for 5 min, and 0.1 ml is pipetted off for radioactivity counting. Spontaneous lysis is determined by incubating medium with targets, whereas maximum 51 Cr release is determined by adding 1% Nonidet P-40 to target cells. Percentage of lysis is calculated as 100 1 (cpm test sample 0 cpm spontaneous lysis)/(cpm maximum release 0 cpm spontaneous lysis). Effector cells are tested at a minimum of three different effector to target cell ratios (E:T). Assays are performed in quadruplicate. Replicate samples have an SD of õ10% of the mean. Samples are counted in an Isoflex automatic gamma counter (ICN Micromedic Systems, Huntsville, AL). Lytic units are calculated as the number of effector cells required to mediate a defined percentage lysis of target cells. Single-cell cytotoxicity assays have been performed by incubating effector cells with K562 target cells at an E:T ratio of 1:1 for 30 min to allow conjugate cell formation. Conjugates are resuspended in 0.5% agarose and spread in a thin layer onto agarose-coated slides. After incubation at 377C for 4 h, slides are stained with 0.1% trypan blue to visualize dead cells and fixed in 0.1% formaldehyde. Labeling effector cells for 1 h with high specific activity [3H]thymidine makes it possible to identify proliferating effector cells with nuclear track emulsion (type NTB-2; Kodak, Rochester, NY) (64). Flow Cytometric Analysis For flow cytometric analysis of cell subsets, splenic leukocytes are blocked in 20% FBS with 25 mg/ml mouse Ig (Sigma) for 30 min before addition of conjugated Abs. Cell subsets are identified by staining with the following Abs: anti-CD8 fluorescein (FITC)-conjugated rat mAb 53-6.7, anti-CD3e FITC-conjugated or biotinylated hamster mAb 145-2C11, anti-CD4 R-phycoerythrin (PE)-conjugated rat mAb RM4-5, anti-NK1.1 R-PE-conjugated mouse mAb PK136, anti-CD25 biotinylated mAb 3C7, and anti-IL2R p75 b-chain mAb FITC-conjugated TM-b1 (all from PharMingen, San Diego, CA). For three-color flow cytometric analysis, biotinylated Abs are followed by streptavidin–PerCP (Becton Dickinson, Mountain View, CA). Nonspecific binding is determined by using control antibodies lacking specificities for murine determinants (PharMingen). Antibody incubations are performed for 30 min on ice, with vortexing every 10–15 min. After staining, cells are fixed in paraformaldehyde. Analyses are performed on a FACScan (Becton Dickinson, San Jose, CA) using the LYSYS I, LYSYS II, or CELLQUEST software packages. Argon laser output is 15 mW at 488 nm. Cell populations are gated and the relative percentages determined after background subtraction.

AP: Methods

NATURAL KILLER CELLS AND VIRAL INFECTION

For cell cycle analysis, DNA is quantitated by staining with 20 mg/ml propidium iodide in the presence of 50 mg/ml ribonuclease A, after fixation with 70% ethanol (60, 67). Alternatively, using two-color flow cytometry, gated cell subsets are examined for forward side scatter (FSC) (68) and percentage of blast-size cells calculated as proportions of cells giving more scatter than the corresponding baseline lymphoid cell populations from uninfected mice. Cell Proliferation Assays To evaluate proliferation of cell populations, cells are incubated in the absence or presence of exogenous factors at 105 per well in microtiter plates at 377C. Tritiated thymidine (ICN Radiochemicals) is added at 1 mCi per well during 6- to 24-h incubations. Cells are harvested using an automated cell harvester, and incorporation of the DNA precursor is quantified by liquid scintillation counting (30, 40, 41, 55, 56). Proliferation at the single-cell level is evaluated by incubating cells for 1 h with high specific activity [3H]thymidine, spinning them onto microscope slides using a cytocentrifuge, fixing in paraformaldehyde, and coating with nuclear track emulsion. After exposure for 24 h at 47C, slides are developed and stained with Wright’s – Giemsa. The proportion of DNA-synthesizing cells is enumerated by microscopic examination (30, 55). Viral Replication Assays LCMV and MCMV viral titers are quantitated by infectious center plaque formation on confluent monolayers of Veros cells (65) or MEFs (52), respectively. For assay of titers in tissues, splenic, renal, and liver homogenates are prepared in ice-cold Teflon pestle homogenizers (Wheaton, Milville, NJ). Cell debris is removed from organ supernatants by refrigerated centrifugation of homogenates for 20 min at 1000g. Duplicate log dilutions of supernatants are made and 0.1 ml is incubated on cell monolayers in medium for 1 h at 377C. After incubation, monolayers are overlaid with 2–4 ml of 11 medium 199 supplemented with 5% FBS, antibiotics, glutamine, and either 0.5% ME agarose for LCMV plaque assays or 0.6% low-gelling agarose for MCMV plaque assays (FMC Bioproducts, Rockland, ME). Agarose is allowed to gel and plates are incubated at 377C for 4 to 4.5 days. LCMV plaques are visualized by overlay with 2 ml of 11 medium 199 supplemented with 5% FBS, antibiotics, glutamine, 0.5% ME agarose, and 0.034% neutral red, and MCMV plaques are revealed by staining of monolayers with 0.1% crystal violet in 1% formaldehyde. Experimental samples are run in parallel with virus standards and negative controls. Cytokine Analysis Northern Blot Hybridization Analysis For the analysis of endogenous mRNA expression of particular genes, including the genes for IFN-b (25),

/ 6706$$357b

04-16-96 00:24:02

metha

383

IFN-g (60), and TNF-a (69), cellular RNA is isolated from total cell populations using RNAzol-B (BIOTECX LAB, Houston, TX) according to company recommendations or lysis with guanidium thiocyanate and centrifugation through a 5 M cesium chloride gradient. Cytoplasmic RNA is prepared by lysing cells with Nonidet P-40 in the presence of RNase inhibitors and proteinase K, followed by phenol –chloroform extraction and ethanol precipitation. RNA is stored at 0807C in ethanol until use. Samples containing up to 40 mg RNA are denatured by heating at 657C in 50% formamide, separated by gel electrophoresis in 1.2% agarose containing 2.2 M formaldehyde, blotted by capillary action onto Hybond-N filters (Amersham Corp., Arlington Heights, IL) (69), nitrocellulose filters (Schleicher & Schuell, Inc., Keene, NH) (25), or Biodyne filters (ICN Biochemicals. Inc., Irvine, CA) (60, 67), and baked dry for several hours. For Hybond-N filters, prehybridization is carried out according to the manufacturer’s recommendations, and hybridization is performed overnight at 657C. For nitrocellulose filters, prehybridization is carried out overnight at 427C in 11 Denhardt’s, 51 SSC, 50 mM sodium phosphate, 50% formamide, and 0.1 mg/ ml Brewer’s yeast tRNA, followed by hybridization overnight at 427C with probe in a mixture of 4 parts prehybridization solution and 1 part 50% (w/v) dextran sulfate. For Biodyne filters, prehybridization is carried out for 30 min at 427C in 51 Denhardt’s, 51 SSC, 50 mM sodium phosphate, 0.1% SDS, 50% formamide, and 250 mg/ml salmon sperm DNA, followed by hybridization overnight at 427C with probe in prehybridization solution. The DNA probes used in hybridization, from plasmid gene inserts or products amplified by the polymerase chain reactions, are labeled with [32P]dATP (New England Nuclear, Boston, MA) or [32 P]dCTP (ICN) by random hexanucleotide priming (BoehringerMannheim Biochemicals, Indianapolis, IN) to a specific activity of ú108 cpm/mg, added to hybridization solution, and denatured at 1007C. After hybridization, filters are washed several times in SSC and SDS at decreasing salt concentrations and increasing temperatures, prior to exposure to film (Kodak X-OMAT) at 0807C with intensifying screen(s). Densitometric tracings are now performed using the Fotodyne image analysis system (Fotodyne, Hartland, WI). In Situ Hybridization In situ hybridization of isolated cell populations is carried out to evaluate the frequency of cells expressing mRNA for particular cytokine genes including IL-2 (40, 41, 55, 56). For this procedure, cells are spun onto microscope slides using a cytocentrifuge, fixed in 4% paraformaldehyde, and stored dehydrated in 70% EtOH until use. After rehydration in PBS, cells are treated with 50% formamide at 707C for 5 min and hybridized for 3 h at 377C. The probes used in hybridization are labeled

AP: Methods

BIRON, SU, AND ORANGE

384

with [35S]dCTP (New England Nuclear) by random hexanucleotide priming (Boehringer-Mannheim Biochemicals, Indianapolis, IN) or nick-translation to a specific activity of ú108 cpm/mg, added to a mixture of salmon sperm DNA and brewer’s yeast tRNA, dehydrated, and denatured in 100% formamide at 907C. After hybridization, cells are washed in formamide and SSC at 47C, dried, dipped in nuclear track emulsion, and after exposure at 47C for 7–10 days, developed and stained with Wright’s –Giemsa (40, 55). To identify the tissue localization of cells expressing mRNA for a variety of cytokines including IFN-b (25), TNF-a (69), and IFN-g (Salazar-Mather, Ishikawa, and Biron, submitted for publication), tissue sections are evaluated by in situ hybridization. For these experiments, freshly harvested organs are snap-frozen in liquid nitrogen and embedded in Tissue-Tek O.C.T. compound (Miles Inc., Elkhart, IN). Frozen sections, 8–10 mm, are cut, fixed in paraformaldehyde for 20 min to 1 h, and stored at 0707C until use (preferably õ2 weeks) (25, 69; Salazar-Mather, Ishikawa, and Biron, submitted for publication). After rehydration in PBS, sections are treated with 0.2 M Tris, 0.1 M glycine, pH 7.4, for 10 min, denatured in 50% formamide at 707C for 5 min, and hybridized overnight at 377C. Alternatively, sections are treated with 0.3% Triton X-100 in 0.1 M Tris base, pH 7.5, for 15 min; 0.01 M Tris base for 5 min; 1 mg/ml proteinase K in Tris –EDTA for 10 min; a rinse of nuclease-free water; a rinse of 0.1 M Triethanolamine, pH 8.0; 0.25% acetic anhydride in 0.1 M Triethanolamine for 10 min; 0.21 SSC for 5 min; and then hybridized overnight at 427C. Prehybridization of sections can be carried out without penetration with Triton X-100, proteinase K treatment, and ethanol dehydration. The probes for hybridization are labeled with digoxigenin –dUTP using the Genius nonradioactive DNA labeling kit (Boehringer-Mannheim). For each section, 10 to 1000 ng labeled probe is denatured in 100% formamide at 1007C, mixed in hybridization solution containing 20% dextran sulfate, 4 mg/ml BSA, and 10% vanadyl ribonucleoside complex. After hybridization, sections are washed in formamide and SSC. Hybridization signals are detected immunologically using alkaline phosphatase-conjugated sheep anti-digoxigenin antibody. Blue–purple alkaline phosphatase precipitates are developed using the substrate nitro blue tetrazolium salt or BM purple precipitating substrate (Boehringer-Mannheim). Tissues are sometimes evaluated without a counterstain (69). Alternatively, they are counterstained with Giemsa (Salazar-Mather, Ishikawa, and Biron, submitted for publication), 0.05% nuclear fast red solution in 5% aluminium sulfate (25), or hematoxylin. Immunohistochemistry For immunohistochemical staining, freshly harvested organs are snap-frozen in liquid nitrogen and

/ 6706$$357b

04-16-96 00:24:02

metha

embedded in Tissue-Tek O.C.T. compound. Frozen sections of 8–10 mm are cut, fixed in acetone for 20 s, and blocked with 1.5% normal goat serum in PBS for 20 min. Tissue sections are incubated for 30 min with polyclonal antiserum specific to the cytokine of interest or equivalent concentrations of control serum. Bound antibody is detected using the Vectastain ABC-AP kit (Vector Laboratories, Inc., Burlingame, CA) as follows: sections are incubated with a 1/200 dilution of biotinylated goat anti-rabbit IgG (0.5 mg/ml) for 30 min, with ABC reagent containing alkaline phosphatase for 30 min, and with Vectastain substrate kit II (Vector Laboratories, Inc.) to develop the brown–black enzyme substrate precipitates. Levamisole at 10 mM is used to inhibit endogenous alkaline phosphatase activity. Sections are counterstained with 0.05% nuclear fast red solution in 5% aluminium sulfate (25, 59). Staining specificity is tested by lack of control serum staining and blocking of specific antibody staining with an excess of added soluble cytokine. ELISAs Cytokines, i.e., IFN-g (52), IL-12 p40 (70), or IL-4 (Su and Biron, unpublished), in serum or conditioned media are quantitated by sandwich ELISA. Conditioned media are prepared with splenic leukocytes at 5 1 106 or 107 cells/ml under media conditions determined to be optimal for individual cytokines. For measurement of IL-4, 0.25 mg/ml of anti-murine IL-4R mAb (Genzyme) is added to block IL-4 consumption during the production phase. After incubation for 24 h at 377C, 5% CO2 , samples are centrifuged to remove cells and stored at 0807C until use. In some cases, it is necessary to concentrate conditioned media using Centriprep-10 or Centricon-10 concentrators (Amicon, Beverly, MA). Immulon 4 microtiter plates (Dynatech Laboratories, Inc., Chantilly, VA) are coated overnight at 47C with primary anti-cytokine antibodies and blocked with 5 to 10% FBS PBS, and biological samples are incubated overnight at 47C. Secondary anti-cytokine mAb or polyclonal sera directed at different epitopes are then added, followed by peroxidase-linked tertiary reagent. Washes between incubations are performed using PBS containing 0.05% Tween 20. Bound cytokine is detected, after incubation with ABTS substrate (Kierkegaard and Perry, Gaithersburg, MD), by colorimetric changes read at 410 using a Dynatech MR 4000 plate reader. Quantitation is performed by comparison to standard murine recombinant samples obtained as follows: IFN-g (PharMingen), IL-4 (Gibco), and IL-12 (Genetics Institute, Cambridge, MA). For detection of IFNg, rat anti-mouse IFN-g XMG1.2 is used as the capture antibody, and polyclonal rabbit anti-IFN-g (a gift from Dr. P. Scott, University of Pennsylvania, PA) is used as the secondary antibody. Peroxidase-conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch, West Grove,

AP: Methods

NATURAL KILLER CELLS AND VIRAL INFECTION

PA) is a tertiary reagent. Sensitivity of IFN-g ELISA is to 39 pg/ml. For detection of IL-12 p40, rat anti-mouse IL-12 mAb C15.1 (61) (a gift of Drs. M. Wysocka and G. Trinchieri) is used as the capture antibody, with polyclonal sheep anti-IL-12 (Genetics Institute) as the secondary antibody and peroxidase-conjugated donkey antisheep IgG (Jackson ImmunoResearch) as the tertiary reagent. Sensitivity of IL-12 p40 ELISA is to 45 pg/ml. For detection of IL-4, rat anti-mouse IL-4 mAb BVD4-1D11 (PharMingen) is used as the capture antibody, with biotinylated-rat anti-mouse IL-4 mAb BVD6-24G2 (PharMingen) as the secondary antibody and peroxidase-conjugated avidin (Sigma) as the tertiary reagent. Sensitivity of IL-4 ELISA is to 10 pg/ml. TGF-b1 is measured with ELISA kit assays purchased from R & D Systems (Minneapolis, MN) (work in progress). Interferon Bioassay IFN, in serum or organ homogenate supernatants, is measured on L929 cells in a standard microtiter protection assay against infection with vesicular stomatitis virus (VSV) as described (3, 25, 57). Samples are serially diluted twofold in 96-well microtiter plates before addition of 104 L929 cells/well. After an 18- to 24-h incubation at 377C, 5% CO2 , VSV is added, and cytopathic effects (CPE) are visually scored 2 to 3 days later. In this assay, a 50% reduction of virus-induced CPE is equal to 1 IRU/ml of IFN. Thus, units of IFN are equal to the reciprocal of the dilution at which 50% CPE is observed. Characterization of the factor as IFNa or IFN-b can be made by neutralization with antibodies to the factors (25).

385

plates. CTLL-2 cells at 5 1 103/well are added in the presence of 8 mg/ml methyl a-D-mannopyranoside (Sigma) and incubated for 48 h at 377C, 5% CO2 . Tritiated thymidine is added during the last 6 h of the 48h incubation, and incorporation of the DNA precursor is quantified by liquid scintillation counting after cell harvesting. The active factor in conditioned media is identified as IL-2 by the ability of a mixture of anti-IL2R p55 a chain mAbs 7D4 and PC61, or the anti-murine IL-2 mAb S4B6, to block IL-2-driven CTLL proliferation (40, 49). Transforming Growth Factor-b (TGF-b) Bioassay For evaluation of TGF-b production, conditioned media are prepared with splenic leukocytes at 107 cells/ml in DMEM (GIBCO) supplemented with 2% Nuserum (Collaborative Research, Bedford, MA) or 200 to 300 mg/ml BSA, 10 mM Hepes, antibiotics, and glutamine. After incubation for 24 h at 377C, 5% CO2 , samples are centrifuged to remove cells, and supernatants are collected under sterile conditions. In some cases, supernatants are supplemented with final concentrations of 0.23 TIU/ml aprotinin, 0.10 mM PMSF, and 10 mg/ml soybean trypsin inhibitor (all from Sigma), dialyzed with a molecular cutoff of 12 to 14 kDa against DMEM, and sterilized through a low protein-binding 0.2-mm filter (Acrodisc Gelman Sciences, Ann Arbor, MI). Samples are stored at 0807C until use. To evaluate total TGF-b, conditioned media are treated with 0.12 N HCl for 5 min to acid activate latent TGF-b and then neu-

IL-12 Bioassay IL-12 biological activity is assessed by the factor’s ability to induce IFN-g production from unstimulated splenocytes. Immulon 4 plates coated with nonneutralizing rat anti-mouse IL-12 mAb C15.1 are used to capture IL-12 from biological samples (61; 61a). One million C57BL/6 splenic leukocytes are added per well without or with 50 U/ml IL-2 to increase sensitivity. Resulting IFN-g production is measured by ELISA and compared to levels obtained using known concentrations of IL-12 (Genetics Institute). Specificity is verified using neutralizing C17.15 rat anti-mouse IL-12 (61). Sensitivity of the IL-12 biological assay ranges from 1 to 10 pg/ml. CTLL Bioassay For evaluation of IL-2 production, conditioned media are prepared with splenic leukocytes at 5 1 106 cells/ ml, with or without the anti-IL-2R a chain mAb, 7D4, to block IL-2 consumption during the production phase. After incubation for 24 h at 377C, 5% CO2 , samples are centrifuged to remove cells, and supernatants are collected and serially diluted in 96-well microtiter

/ 6706$$357b

04-16-96 00:24:02

metha

FIG. 1. Generation of splenic NK cell cytotoxic activity during acute viral infections. C57BL/6 male mice were uninfected (Day 0) or infected for 1, 2, 3, 5, 7, or 9 days with LCMV ( l ) or MCMV (- -h--). LCMV-infected mice received 2 1 104 PFU LCMV Armstrong strain clone E350 ip and MCMV-infected mice received 2 1 105 PFU MCMV Smith strain salivary gland extract ip. Splenic leukocytes were evaluated for NK cell-mediated cytotoxicity by 51 Cr release assay against YAC-1 target cells as described under Description of Methods. All values are means of 3 –6 individual animals or pooled cell populations from 2 –8 animals.

AP: Methods

BIRON, SU, AND ORANGE

386

tralized with a final concentration of 0.08 M Hepes, 0.12 N NaOH prior to use. TGF-b in conditioned media is quantitated by inhibition of [3H]thymidine incorporation in Mv 1 Lu cells (30, 59). Mv 1 Lu cells at 8 1 103 cells/well, and equal volumes of diluted test samples or known concentrations of TGF-b (R & D Systems, Minneapolis, MN), are added and incubated for 48 h at 377C, 5% CO2 . Tritiated thymidine is added during the last 10 h of the 48h incubation, and incorporation of the DNA precursor is quantified by liquid scintillation counting after cell harvesting. The half-maximal sensitivity of the assay is 12.5 to 25 pg/ml of TGF-b.

RESULTS FROM APPLICATION OF METHODOLOGY Cellular Responses Following ip infection of normal C57BL/6 mice with LCMV or MCMV, NK cells are activated to mediate elevated cytotoxic activity. These responses have characteristic kinetics with NK cell-mediated lysis peaking at early times, Days 2 to 5, in the spleens after infection with either virus (Fig. 1) (3, 14, 44, 52, 64). Activation of cytotoxicity occurs without apparent or only modest expansion of NK cell numbers in this compartment (Fig. 2A). T-cell responses to infection are induced at later times, beginning on Day 5 of infection. Profound expansions in CD8/ T-cell numbers (Fig. 2B) but modest changes in CD4/ T-cell numbers (Fig. 2C) per spleen are induced (40, 65; 61a). The measurable mag-

nitudes of T-cell responses to LCMV infection are higher than those to MCMV infection (Figs. 2B and 2C) (61a). The early virus-induced activation of NK cellmediated cytotoxicity can be mimicked with an accelerated kinetics by treatment with the IFN-b-inducer, polyI:C (53, 55). Histologically, the spleens are undergoing extensive changes in both cell morphology and overall architecture. Splenic sections from uninfected C57BL/6 mice, stained with H & E, have characteristic normal architecture with red and white pulp regions (Fig. 3A). Leukocytes in both pulp regions are predominantly small, darkly staining cells with low cytoplasm to nucleus ratios. The red pulp regions contain numerous nucleated leukocytes along with erythrocytes. At early times after LCMV infection, Days 2 and 3 (Fig. 3B), there is a clear margination between white and red pulp areas. The proportions of white pulp to total splenic area are increased, and there are decreases in nucleated cells per red pulp regions. Furthermore, reactive phenotype nucleated cells with a high cytoplasm to nucleus ratio are induced and particularly apparent in white pulp regions. All of these changes occur at times preceding measurable increases in cell expansion (40, 64, 65; Salazar-Mather, Ishikawa, and Biron, submitted for publication). The early changes are also observed at Days 2 and 3 after MCMV infection (Salazar-Mather, Ishikawa, and Biron, submitted for publication) and after treatment with polyI:C (25). By Days 5 (Fig. 3C) and 7 (65) after LCMV infection, defined regions of reactive areas in splenic white pulp regions are developing, and there are increases in reactive leukocytes with high cytoplasm to nucleus ratios in red pulp regions. These

FIG. 2. Effect of acute viral infections on splenic NK- and T-cell yields. C57BL/6 male mice were uninfected (Day 0) or infected with LCMV (solid bars) or MCMV (stippled bars) as described in Fig. 1. Splenic leukocytes from experimental animals were prepared and evaluated by flow cytometry as explained under Description of Methods. NK cell populations (A) were defined by NK1.1/CD30 cell staining. CD8/ T-cell populations (B) were defined as being CD8/CD40. CD4/ T-cell populations (C) were defined as being CD4/CD80. All results are expressed as millions of cells per spleen. Each value represents the mean of 3 –6 individual animals or pooled cell populations from 2– 8 animals. ND, not done.

/ 6706$$357b

04-16-96 00:24:02

metha

AP: Methods

387

/ 6706$$357b

04-16-96 00:24:02

metha

AP: Methods

FIG. 3. Virus-induced changes in the morphology of splenic tissues from normal and immunodeficient mice. Spleens were harvested from immunocompetent C57BL/6 mice (A, B, and C), T-cell-deficient nude mice (D, E, and F), and T- and B-cell-deficient SCID mice (G, H, and I) that had been uninfected (A, D, and G), or ip infected with LCMV for 3 (B, E, and D) or 5 (C, F, and I) days. Paraffin sections were prepared and stained with H & E as described under Description of Methods. Original magnification, 112.

BIRON, SU, AND ORANGE

388

late changes are apparent but of a lower magnitude during MCMV infection (65; Orange and Biron, unpublished). Tissues from T-cell-deficient nude (Figs. 3D, 3E, and 3F) and T- and B-cell-deficient SCID (Figs. 3G, 3H, and 3I) mice show that margination between red pulp and white pulp regions as well as induced nucleated cell reductions in red pulp areas occur independently of T and B cells. These morphological changes are still readily discernible through Day 5 after infection (Figs. 3F and 3I). Because the virus-induced splenic increases in white pulp regions and the development of reactive leukocyte morphology are reduced in nude mice (Figs. 3E and 3F) and not apparent in SCID (Figs. 3H and 3I) mice, these parameters appear to be primarily dependent on the presence of T and B lymphocytes. Trafficking studies, under conditions of polyI:C treatment, indicate that the morphological changes are at least in part a consequence of the induction of T/B-lymphocyte migration into splenic white pulp regions (25; Salazar-Mather, Ishikawa, and Biron, submitted for publication). Characterization of the NK cell phenotype during early viral infections and after polyI:C induction of IFN-b in vivo has shown that the cells elicited under these conditions are CD30 cells (49, 63, 67). Although they are not undergoing extensive expansion in the spleen, these NK cells are undergoing blastogenesis and proliferation. Compared to resting NK cells in unmanipulated mice, the activated NK cell populations elicited during early LCMV infection or after polyI:C treatment are: (1) larger and less dense (53, 54, 64); (2) more sensitive to the cell cycle-specific toxin, hydroxyurea (57); (3) induced to incorporate the DNA precursor, [3 H]thymidine (53, 55, 64); and (4) elevated in proportions of cells with DNA in the S and G2/M phases of the cell cycle (60, 67). Thus, the activated NK cells are proliferating cells. Given this background information, it is possible to routinely quantitate the induction of NK cell blastogenesis using flow cytometry by staining for NK cell surface markers with fluorescent antibody and evaluating size changes within NK cell subsets based on FSC (49). For example (Fig. 4), splenic leukocytes, isolated from uninfected C57BL/6 and enriched in T- and NK-cell populations by depletion of B cells and PMNs with J11d antibody, can be shown to contain approximately 10% NK1.1 /CD30 NK cells, with up to approximately 10% of these being blast-size cells. In contrast, equivalent populations from C57BL/ 6 mice on Day 3 after infection with LCMV represent about the same proportion of the splenic populations but up to approximately 67% of these are blast-size cells (Fig. 4). Interestingly, the blast-size proportion of the small percentage of NK1.1/CD3/ cells is not modified by viral infection. The kinetics of the NK1.1/ CD30 NK cell blastogenic response, during either the LCMV

/ 6706$$357b

04-16-96 00:24:02

metha

or the MCMV infections, coincides with the activation of NK cell cytotoxicity (Fig. 5 compared to Fig. 1). Parallel studies characterizing the Days 7 and 9 Tcell responses during LCMV infections have shown that CD8/ T cells, expanding in total numbers per spleen (Fig. 2) and induced to mediate virus-specific CTL activity (40, 41, 64, 65, 69), are also larger and less dense than CD8/ T cells from uninfected mice (57), have an increased sensitivity to hydroxyurea (57; 70a), and are induced to synthesize DNA (40, 64, 66, 67). Flow cytometric analyses of the splenic T cells from uninfected and Day 7 LCMV-infected C57BL/6 mice demonstrate that, in contrast to the CD4/CD80 T cells, the percentages CD8/ CD40 are increased by ú twofold and the proportions of these that are blast-size cells are increased by up to fourfold (Fig. 6). The kinetics of the T-cell blastogenic and proliferative responses to LCMV infection in normal mice peaks at times after the NK cell response has subsided (Fig. 7). The T-cell changes actively contribute to negative regulation of the NK cell response because activation of NK cell cytotoxicity is extended in T-cell-deficient nude and T- and B-cell-deficient SCID mice (59) as well as in mice rendered CD8/ T-cell-deficient as a consequence of cell depletion either by antibody treatment in vivo or by

FIG. 4. Blastogenesis of NK cells during acute LCMV infection. Male C57BL/6 mice were uninfected (D0, Day 0) or infected for 3 days with LCMV (D3, Day 3). Splenic leukocytes were prepared and B cells and PMNs were depleted using J11d antibody and C* as described under Description of Methods. Flow cytometry was performed using PE-conjugated anti-NK1.1 and FITC-conjugated antiCD3e. Representative contour plots are shown at the top, with percentages of NK1.1/ CD30 and NK1.1/CD3/ cells given. Forward scatter analysis of the gated NK1.1/CD30 (bottom left) and NK1.1/CD3/ cell subsets (bottom right) were acquired. Blast-size cells were defined as cells larger than 90% of resting NK cells. The percentages of the respective cell subsets that are blast size are shown.

AP: Methods

NATURAL KILLER CELLS AND VIRAL INFECTION

genetic mutation of the b2-microglobulin gene, i.e., b2m0/0 mice (49). Interestingly, overall NK cell proliferation in the spleen is dramatically elevated at late times after LCMV infection of b2m0/0 mice (49). This observation suggests that, in the absence of CD8/ T cells, activation of CD4/ T cell responses establishes conditions supporting extended NK cell blastogenesis and expansion. Expression and Function of in Vivo Cytokine Responses Cytokine expression during LCMV and MCMV infections has been and is being characterized using a variety of techniques (Fig. 7). These include mRNA expression by Northern blot analysis and in situ hybridization, protein expression by ELISA and immunohistochemistry, and production of biologically active factor with functional assays. The virus-induced IFNs a/b are elicited at early times after infection with either virus (3, 62, 71). These cytokines are readily detected in serum. Expression coincides with the activation of NK cell cytotoxicity, and in vivo treatments with antibodies neutralizing these factors block the induction of this NK cell function during MCMV infection (62, 71; Orange and Biron, unpublished). IFN-a/b expression is resistant to the inhibitor of IL-2 transcription, CsA. The early NK cell cytotoxicity and blastogenesis, elicited during LCMV or MCMV infection as well as in response to polyI:C treatment, are also resistant to CsA

FIG. 5. Kinetic analysis of acute viral infection induction of blastsize NK cells. C57BL/6 male mice were uninfected (Day 0) or infected with LCMV (solid bars) or MCMV (stippled bars) as described in Fig. 1. Isolated splenic leukocytes were evaluated by flow cytometry as explained under Description of Methods. Blast-size NK cells were determined by performing forward scatter analysis of NK1.1 /CD30 NK cells. Cells with forward scatter ú90% of resting populations were identified as blast-size cells. Results are expressed as percentage of the NK cell population that is of blast size. Each value represents the mean of 3 –6 individual animals or pooled cell populations from 2 –8 animals. ND, not done.

/ 6706$$357b

04-16-96 00:24:02

metha

389

(49, 55; Biron, unpublished). As the late T-cell responses to LCMV infection are sensitive to this drug (56), the signals for early NK cell activation and blastogenesis are fundamentally different from those for Tcell activation. Recent studies from this laboratory have demonstrated that another regulator of NK cell function can be induced in some but not all viral infections. Protein expression and production of biologically active IL-12 is induced in a sharp peak at early times after MCMV but not LCMV infection (70; 61a). The kinetics of this response coincides with induction of detectable NK cell IFN-g protein production; although both infections induce NK cell expression of IFN-g at the level of mRNA (60; Salazar-Mather, Ishikawa, and Biron, submitted for publication), NK cell IFN-g protein production is detectable only during MCMV infection (52, 70; 61a). The NK cell IFN-g protein can be measured both in media conditioned with splenic leukocytes (70) and in serum (61a) from mice infected with this virus. Treatments with antibodies neutralizing IL-12 block the induction of NK cell IFN-g production (61a). Thus, there is definitive evidence that endogenous cytokine expression during MCMV infection controls particular aspects of NK cell activation, IFN-a/b production activates NK cell cytotoxic activity, and IL-12 induces NK cell IFN-g protein expression. IFN-g expression can also be detected as a result of T-cell activation during infection with either MCMV or LCMV. IFN-g production at the late time points is more readily detectable in conditioned media than in serum. The magnitude of the late MCMVinduced IFN-g response is less than the early NK cell-produced IFN-g response observed during this infection. In contrast, the LCMV-induced T-cell IFNg response is significantly larger and is sustained for a longer period of time than either of the MCMVinduced responses (70; 61a). Additional and different cytokines are associated with the later T-cell responses. Of particular relevance to understanding the physiological factors promoting NK cell responses in vivo is the kinetics of IL-2 expression during LCMV infection (40, 41, 56). Although IL2 can activate NK cell killing and support NK cell proliferation, peak production of this factor occurs at times of amplifying T-cell but waning NK cell responses. Both the CD4/ and CD8/ T cells are induced to express IL2 mRNA as evaluated by in situ hybridization (40, 41). The CD4/ T cells are the major producers of detectable IL-2 protein, whereas the CD8/ T cells are the major consumers of the factor (41). As LCMV-specific CTLs are induced in mice rendered IL-2-deficient by genetic mutation (48), this T-cell function is not dependent upon IL-2. The factor is required, however, for maximal CD8/ T-cell expansion and peak induction of IFN-g during LCMV infection (70a). The Day 7 in vivo acti-

AP: Methods

BIRON, SU, AND ORANGE

390

vated T cells but not the Day 3 activated NK cells are induced to express the high-affinity form of the IL-2 receptor (55, 56, 60). As a consequence, they have a competitive advantage for utilization of the factor at times of peak IL-2 expression. This may contribute to the apparent lack of NK cell activation at times of IL2 expression during LCMV infection. In contrast to the NK cell responses in normal mice, the late NK cell response observed during LCMV infection of the CD4/ T-cell-containing but CD8/ T-cell-deficient b2m0/0 mice does occur at times of IL-2 expression and is dependent upon the endogenous availability of this factor (49). Thus, IL-2 can participate in the amplification of NK cell responses, but the activation of CD8/ T cells appears to compete for and interfere with NK cell utilization of this factor during viral infections. In addition to the apparent competition for augmenting factors, factors inhibiting NK cell responses are expressed at later times after viral infection (30, 59). Although certain splenic cell populations constitutively express TGF-b1 mRNA (30), expression of the protein cannot be detected in splenic tissues from unin-

fected mice (59). LCMV infection induces the expression of immunohistochemically detectable TGF-b protein in tissues (59). TGF-b proteins are produced as inactive precursor molecules. Further processing is required for the release of biologically active factor. Splenic leukocytes from LCMV-infected mice are induced to release increased levels of total TGF-b1 protein into conditioned media, as measured in bioassays (30, 59) and ELISAs (Su and Biron, manuscript in preparation) after experimental processing of the factor. Active factor production, measured directly in biological assays, is also induced during infection (30, 59). Low levels of TGF-b protein expression are detected as early as Day 3 and peak on Days 7 to 14 after infection. Thus, production coincides with the activation of the T-cell responses but the inhibition of the NK cell responses. Studies in T-cell-deficient nude and T- and B-cell-deficient SCID mice show that T-cell responses are required for peak production of biologically active TGFb (59). Interestingly, in comparison to the T cells activated at late times, the NK cells elicited at early times after LCMV infection are 100- to 1000-fold more sensi-

FIG. 6. Blastogenesis of CD8/ T cells during acute LCMV infection. Male C57BL/6 mice were uninfected (Day 0) or infected for 7 days with LCMV (Day 7). Splenic leukocytes were prepared and flow cytometry was performed using PE-conjugated anti-CD4 and FITC-conjugated antiCD8. Representative contour plots are shown at the top. Forward scatter analysis of CD4/CD80 T cells (bottom left) and CD8/CD40 T cells (bottom right) was performed and cells larger than approximately 90% of resting T cells were identified as blast-size cells.

/ 6706$$357b

04-16-96 00:24:02

metha

AP: Methods

NATURAL KILLER CELLS AND VIRAL INFECTION

tive to TGF-b1-mediated inhibition of proliferation (30). Thus, T-cell responses actively contribute to the regulation of NK cell responses by promoting the production of inhibitory factors that act preferentially on this subset. Another factor with the potential to regulate NK cells is IL-4. Recent studies show that this cytokine is also produced at later times postinfection (72; Su and Biron, manuscript in preparation). The regulatory effects of this factor on NK cell responses are currently under investigation.

DISCUSSION Although there has been extensive characterization of NK cell responses and of the cells and cytokines that can regulate NK cells in culture, the in vitro approaches only weakly mimic endogenous immune responses. They cannot distinguish the particular in vivo components called forth during, or evaluate the complex interactions occurring in the context of, an intact host response to a viral infection. Thus, in vitro approaches identify mediators that can affect NK cells but do not necessarily define the responses happening during challenge in vivo. The mouse models of LCMV and MCMV infection reviewed here provide a powerful approach to charac-

391

terizing the regulation and function of NK cells during acute viral infections. They present unique opportunities to examine responses in a three-dimensional context. During these viral infections, NK cell responses are taking place with access to the many different immune and tissue compartments of the host, subject to multiple cellular trafficking and activation events taking place within the various mixed in vivo cell populations, and occurring in concert with the dynamic host/ virus interactions. Although characterization of individual components mediating specific effects under these conditions is challenging, understanding the biological relevance of each of these is absolutely dependent upon studying them together as well as apart. The in vivo studies are now being greatly facilitated by the recent and continuing development of antibodies eliminating cell subsets or neutralizing cytokines in vivo and mice rendered deficient for cell subsets or cytokines by specific genetic mutation. These powerful tools have made it possible to specifically dissect the effects requiring particular cells or cytokines and those modified by each of these. In addition to allowing the precise characterization of in vivo NK cell phenotypes, studies of the viral infection models have yielded important information about NK cell regulation in the host. Certain of the experiments have revealed surprising in vivo effects and responses that were not predicted by any of the in vitro studies. For example, they demonstrate that, in addi-

FIG. 7. Kinetics of cellular and cytokine responses to infections with either LCMV or MCMV. NK cell-mediated cytotoxicity and blastogenesis are induced at early times, peaking on Days 3 to 5 after infection. T-cell responses are activated at later times with a peak CD8/ Tcell expansion on Days 7 to 9 after infection. Representation of the kinetics of cytokine expression is based on results presented and/or discussed in the text. Specifically: IFN-a/b (3, 62, 71); IL-12 (70; 61a); IFN-g (52, 70; 61a); IL-2 (40, 41, 49, 56, 73) IL-4 (72; Su and Biron, manuscript in preparation); and TGF-b (30, 59).

/ 6706$$357b

04-16-96 00:24:02

metha

AP: Methods

BIRON, SU, AND ORANGE

392

tion to the activation of NK cell cytotoxicity (Fig. 1), early responses to viral infections include the induction of NK cell blastogenesis (Figs. 4 and 5) and profound changes in cellular localization (Fig. 3). They also show that, although IL-2 can be a potent activator of NK cell cytotoxic activity and proliferation in culture, IL-2 does not appear to be a major factor for activating NK cells during acute viral infections of immunocompetent mice because of the in vivo condition of IL-2 production at times of T-cell expression of the high-affinity IL-2R and T-cell proliferation (40, 55, 60). Finally, they demonstrate the differential sensitivity of NK compared to T cells to particular cytokines induced during viral infections and the role of T cells in negatively regulating NK cell responses (49, 59). Thus, these models have made and will continue to make possible the identification of the cells and factors actually playing a role in controlling in vivo NK cell responses. In conclusion, it is clear that the basic understanding of immunity as well as the development of therapeutic strategies for controlling disease depends upon characterization of in vivo conditions. The well-developed mouse models of acute viral infections with LCMV or MCMV have been used to characterize the regulation and function of NK cells in vivo. There are still many unknowns about in vivo NK cell responses. These models can be further exploited to facilitate the precise and complete characterization of cellular and cytokine effects relevant to promoting endogenous protective responses against viral infections.

13. Biron, C. A., Byron, K. S., and Sullivan, J. L. (1989) N. Engl. J. Med. 320, 1731 –1735. 14. Bancroft, G. J., Shellam, G. R., and Chalmer, J. E. (1981) J. Immunol. 126, 988– 994. 15. Shellam, G. R., Allen, J. E., Papadimitriou, J. M., and Bancroft, G. J. (1981) Proc. Natl. Acad. Sci. USA 78, 5104 – 5108. 16. Bukowski, J. F., Woda, B. A., Habu, S., Okumura, K., and Welsh, R. M. (1983) J. Immunol. 131, 1531 – 1538. 17. Welsh, R. M., Dundon, P. L., Eynon, E. E., Brubaker, J. O., Koo, G. C., and O’Donnell, C. L. (1990) Nat. Immun. Cell. Growth Regul. 9, 112– 120. 18. Habu, S., Akamatsu, K., Tamaoki, N., and Okumura, K. (1984) J. Immunol. 133, 2743 –2747. 19. Stein-Streilein, J., Guffee, J., and Fab, W. (1988) Reg. Immunol. 1, 100–105. 20. Bukowski, J. F., Warner, J. F., Dennert, G., and Welsh R. M. (1985) J. Exp. Med. 161, 40– 52. ¨ rn, A., Wigzell, H., Senik, A., and Gresser, I. 21. Gidlund, M., O (1978) Nature 273, 759– 761. 22. Trinchieri, G. (1989) Adv. Immunol. 47, 187–376. 23. Gresser, I., Guy-Gran, D., Maury, C., and Maunoury, M. (1981) J. Immunol. 127, 1569 –1575. 24. Korngold, R., Blank, K. J., and Murasko, D. M. (1983) J. Immunol. 130, 2236 –2240. 25. Ishikawa, R., and Biron, C. A. (1993) J. Immunol. 150, 3713 – 3727. 26. Suzuki, R., Handa, K., Itoh, K., and Kumagi, K. (1983) J. Immunol. 130, 981–987. 27. Trinchieri, G., Matsumoto-Kobayashi, M., Clasr, S. C., Seehra, J., London, L., and Perussia, B. (1984) J. Exp. Med. 160, 1147 – 1169. 28. Kobayashi, M., Fitz, L., Ryan, M., Hewick, R. M., Clark, S. C., Chan, S., Loudon, R., Sherman, F., Perussia, B., and Trinchieri, G. (1989) J. Exp. Med. 170, 827– 845. 29. Chan, S. H., Perussia, B., Gupta, J. W., Kobayashi, M., Pospisil, M., Young, H. A., Wolf, S. F., Young, D., Clark, S. C., and Trinchieri, G. (1991) J. Exp. Med. 173, 869–879.

REFERENCES 1. Trinchieri, G., Santoli, D., and Koprowski, H. (1978) J. Immunol. 120, 1849 –1855. 2. Trinchieri, G., Santoli, D., Dee, R. R., and Knowles, B. B. (1978) J. Exp. Med. 147, 1299 – 1313. 3. Welsh, R. M. (1978) J. Exp. Med. 148, 163– 181. 4. Santoli, D., Trinchieri, G., and Koprowski, H. (1987) J. Immunol. 121, 532–538.

30. Su, H. C., Leite-Morris, K. A., Braun, L., and Biron, C. A. (1991) J. Immunol. 147, 2717 –2727.

5. Anegon, I., Cuturi, M. C., Trinchieri, G., and Perussia, B. (1988) J. Exp. Med. 167, 569– 583. 6. Perussia, B. (1991) Curr. Opin. Immunol. 3, 49–55. 7. Wong, G. H., Clark-Lewis, I., McKimm-Breschkin, J. L., Harris, A. W., and Schrader, J. W. (1983) J. Immunol. 131, 788–793.

33. Tripp, C. S., Wolf, S. F., and Unanue, E. R. (1993) Proc. Natl. Acad. Sci. USA 90, 3725 –3729.

8. Bancroft, G. J., Sheehan, K. C. F., Schreiber, R. D., and Unanue, E. R. (1983) J. Immunol. 143, 127 –130. 9. Bancroft, G. J., Schreiber, R. D., and Unanue, E. R. (1991) Immunol. Rev. 124, 5 –24. 10. Lopez, C., Kirkpatrick, D., and Read, S. (1983) J. Infect. Dis. 147, 1030 –1035. 11. Sullivan, J. L., Byron, K. S., Brewster, F. E., and Purtilo, D. T. (1980) Science 210, 543 –545. 12. Quinnan, G. V., Jr., Kirmani, N., Esber, E., Saral, R., Manischewitz, J. F., Rogers, J. L., Rook, A. H., Santos, G. W., and Burns, W. H. (1981) J. Immunol. 126, 2036 –2041.

/ 6706$$357b

04-16-96 00:24:02

metha

31. Ortaldo, J. R., Mason, A. T., O’Shea, J. J., Smyth, M. J., Falk, L. A., Kennedy, I. C. S., Longo, D. L., and Ruscetti, F. W. (1991) J. Immunol. 146, 3791 –3798. 32. Bellone, G., Aste-Amezaga, M., Trinchieri, G., and Rodeck, U. (1995) J. Immunol. 155, 1066 – 1071.

34. Howard, M., and O’Garra, A. (1992) Immunol. Today 13, 198– 200. 35. Nagler, A., Lanier, L. L., Cwirla, S., and Phillips, J. H. (1989) J. Immunol. 143, 3183 –3191. 36. Doherty, P. C., and Zinkernagel, R. M. (1975) J. Exp. Med. 141, 502–507. 37. Welsh, R. M., and Zinkernagel, R. M. (1977) Nature 268, 646– 648. 38. Zinkernagel, R. M., and Hengartner, H. (1994) Immunol. Today 15, 262–268. 39. Bachmann, M. F., and Ku¨ndig, T. M. (1994) Curr. Opin. Immunol. 6, 320 –326. 40. Kasaian, M. T., and Biron, C. A. (1989) J. Immunol. 142, 1287 – 1292.

AP: Methods

NATURAL KILLER CELLS AND VIRAL INFECTION 41. Kasaian, M. T., Leite-Morris, K. A., and Biron, C. A. (1991) J. Immunol. 146, 1955 –1963. 42. Fung-Leung, W.-P., Ku¨ndig, T. M., Zinkernagel, R. M., and Mak, T. W. (1991) J. Exp. Med. 174, 1425 – 1429. 43. Christensen, J. P., Marker, O., and Thomsen, A. R. (1995) The 9th International Congress of Immunology Abstracts, 365. 44. Quinnan, G. V., and Manischewitz, J. E. (1979) J. Exp. Med. 150, 1549 –1554. 45. Quinnan, G. V., Manischewitz, J. E., and Ennis, F. A. (1978) Nature 273, 541–543. 46. Jonjic, S., Mutter, W., Wieland, F., Reddehase, M. J., and Koszinowski, U. H. (1989) J. Exp. Med. 169, 1199 – 1212. 47. Reddehase, M. J., Mutter, W., Munch, K., Buhring, H.-J., and Koszinowski, U. H. (1987) J. Virol. 61, 3102 –3108. 48. Ku¨ ndig, T. M., Schorle, H., Bachman, M. F., Hengartner, H., Zinkernagel, R. M., and Horak, I. (1993) Science 262, 1059 – 1061. 49. Su, H. C., Orange, J. S., Fast, L. D., Chan, A. T., Simpson, S. J., Terhorst, C., and Biron, C. (1994) J. Immunol. 153, 5674 – 5681. 50. Oldstone, M. B., Ahmed, R., Buchmeier, M. J., Blount, P., and Tishon, A. (1985) Virology 142, 158– 174. 51. Selgrade, M. J. K., Nedrud, J. G., Collier, A. M., and Gardner, D. E. (1981) Infect. Immunol. 33, 840–847. 52. Orange, J. S., Wang, B., Terhorst, C., and Biron, C. (1995) J. Exp. Med. 182, 1045 –1056. 53. Biron, C. A., Sonnenfeld, G., and Welsh, R. M. (1984) J. Leuk. Biol. 35, 31 –37. 54. Biron, C. A., Pedersen, K. F., and Welsh, R. M. (1986) J. Immunol. 137, 463–471. 55. Kasaian, M. T., and Biron, C. A. (1990) J. Exp. Med. 171, 745– 762. 56. Kasaian, M. T., and Biron, C. A. (1990) J. Immunol. 144, 299– 306. 57. Biron, C. A., Turgiss, L. R., and Welsh, R. M. (1983) J. Immunol. 131, 1539 –1545.

/ 6706$$357b

04-16-96 00:24:02

metha

393

58. Koo, G. C., and Peppard, J. R. (1984) Hybridoma 3, 301 –303. 59. Su, H. C., Ishikawa, R., and Biron, C. A. (1993) J. Immunol. 151, 4874 –4890. 60. Biron, C. A., Young, H. A., and Kasaian, M. T. (1990) J. Exp. Med. 171, 173–188. 61. Kubin, M., Viera, L. Q., Ozmen, L., Garotta, G., Scott, P., and Trinchieri, G. (1995) Eur. J. Immunol. 25, 672– 676. 61a. Orange, J. S., and Biron, C. A. (1996) J. Immunol. 156, 1138 – 1142. 62. Chong, K. T., Gresser, I., and Mims, C. A. (1983) J. Gen. Virol. 64, 461–464. 63. Wang, B., Biron, C., She, J., Higgins, K., Sunshine, M.-J., Lacy, E., Lonberg, N., and Terhorst, C. (1994) Proc. Natl. Acad. Sci. USA 91, 9402 –9406. 64. Biron, C. A., and Welsh, R. M. (1982) J. Immunol. 129, 2788 – 2795. 65. Orange, J. S., Wolf, S. F., and Biron, C. A. (1994) J. Immunol. 152, 1253 –1264. 66. Biron, C. A., Natuk, R. J., and Welsh, R. M. (1986) J. Immunol. 136, 2280 –2286. 67. Biron, C. A., Van Den Elsen, P., Tutt, M., Medveczky, P., Kumar, V., and Terhorst, C. (1987) J. Immunol. 139, 1704 –1710. 68. Parks, D. R., Lanier, L. L., and Herzenberg, L. A. (1986) in Handbook of Experimental Immunology: Immunochemistry (Weir, D. M., Herzenberg, L. A., Blackwell, C., and Herzenberg, L. A., Eds.) Vol. 1, pp. 29.1 –30.11, Blackwell Sci., Boston. 69. Orange, J. S., Salazar-Mather, T. P., Opal, S. M., Spencer, R. L., Miller, A. H., McEwen, B. S., and Biron, C. A. (1995) J. Exp. Med. 181, 901– 914. 70. Biron, C. A., and Orange, J. O. (1995) Res. Immunol., in press. 70a. Cousens, L. P., Orange, J. S., and Biron, C. A. (1995) J. Immunol. 155, 5690 –5699. 71. Grundy (Chalmer), J. E., Trapman, J., Allan, J. E., Shellam, G. R., and Melief, C. J. M. (1982) Infect. Immunol. 37, 143–150. 72. Su, H. C., Kasaian, M. T., Fast, L. D., and Biron, C. A. (1995) J. Cell. Biochem. 19A, 288. 73. Blackett, S., and Mimms, C. A. (1988) Arch. Virol. 99, 1– 8.

AP: Methods