Expansion of Simian Immunodeficiency Virus (SIV)-Specific CD8 T Cell Lines from SIV-Naive Mauritian Cynomolgus Macaques for Adoptive Transfer

ABSTRACT

CD8 T cells play a crucial role in the control of human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV). However, the specific qualities and characteristics of an effective CD8 T cell response remain unclear. Although targeting breadth, cross-reactivity, polyfunctionality, avidity, and specificity are correlated with HIV control, further investigation is needed to determine the precise contributions of these various attributes to CD8 T cell efficacy. We developed protocols for isolating and expanding SIV-specific CD8 T cells from SIV-naive Mauritian cynomolgus macaques (MCM). These cells exhibited an effector memory phenotype, produced cytokines in response to cognate antigen, and suppressed viral replication in vitro. We further cultured cell lines specific for four SIV-derived epitopes, Nef_103–111 RM9, Gag_389–394 GW9, Env_338–346 RF9, and Nef_254–262 LT9. These cell lines were up to 94.4% pure, as determined by major histocompatibility complex (MHC) tetramer analysis. After autologous transfer into two MCM recipients, expanded CD8 T cells persisted in peripheral blood and lung tissue for at least 24 weeks and trafficked to multiple extralymphoid tissues. However, these cells did not impact the acute-phase SIV load after challenge compared to historic controls. The expansion and autologous transfer of SIV-specific T cells into naive animals provide a unique model for exploring cellular immunity and the control of SIV infection and facilitate a systematic evaluation of therapeutic adoptive transfer strategies for eradication of the latent reservoir.

IMPORTANCE CD8 T cells play a crucial role in the control of human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV). Autologous adoptive transfer studies followed by SIV challenge may help define the critical elements of an effective T cell response to HIV and SIV infection. We developed protocols for isolating and expanding SIV-specific CD8 T cells from SIV-naive Mauritian cynomolgus macaques. This is an important first step toward the development of autologous transfer strategies to explore cellular immunity and potential therapeutic applications in the SIV model.

INTRODUCTION

CD8 T cells are essential for the control of human immunodeficiency virus (HIV) and simian immunodeficiency virus (SIV) replication in the infected host. Experimental depletion of CD8 T cells with monoclonal antibodies leads to increased plasma viremia, and immune control is subsequently reestablished when these cells recrudesce (1–3). Additionally, pressure from CD8 T cells selects for escape variants in the acute and chronic phases of HIV/SIV infection (4–7). However, there are clearly qualitative differences between CD8 T cell populations. Both human and macaque cohorts that control viral replication to low or undetectable levels are enriched for the expression of specific major histocompatibility complex (MHC) class I alleles (8–11). These observations suggest that CD8 T cell specificity for certain viral peptides presented on the target cell surface by particular MHC class I molecules is a key determinant of efficacy. Moreover, epitope targeting breadth and polyfunctionality have been linked with CD8 T cell-mediated control of HIV/SIV (12). Although a composite response directed against a broad array of epitopes may reduce the chances of viral escape at any particular site, the reported associations with immune control are conflicting (13–16). Similarly, the ability of CD8 T cells to deploy multiple effector functions in response to cognate antigen encounter is likely beneficial, but the extent to which the observed correlations reflect causality remains unclear (17, 18). Antigen avidity, variant epitope cross-recognition, and clonotype recruitment also play key roles in this complex scenario (19–23). Further investigation of these various factors is therefore needed to distinguish effective from ineffective CD8 T cell responses in HIV/SIV infection.

Although the development of antiretroviral therapy (ART) has greatly improved the prognosis for HIV-infected individuals, complete eradication of the virus remains an elusive goal. In addition to CD8 T cell-mediated pressure during the acute phase, rapid seeding of the viral reservoir occurs within the first few days of infection (24). Establishment of a latent reservoir requires long-term ART to maintain control of the virus. However, recent studies have demonstrated the potential utility of a “shock-and-kill” strategy designed to force latent virus out of the reservoir and enable CD8 T cell-mediated eradication (25, 26). Further studies of CD8 T cell efficacy may therefore contribute to the “kill” component of this novel therapeutic approach.

Nonhuman primate models provide a unique opportunity to explore immune responses in HIV/SIV infection. We identified a geographically isolated population of Mauritian cynomolgus macaques (MCM) in which seven common haplotypes account for all MHC diversity (27). These simple haplotypes facilitate studies of cellular immunity. MCM are particularly useful for investigations of therapeutic strategies because confounding genetic variances that may influence the immune response are eliminated. Experimental studies with MHC-matched MCM are therefore ideally suited to the development of adoptive transfer protocols as part of an eradication strategy.

Adoptive transfer studies provide a powerful method for studying the cellular immune response (28–32). Previous experiments in which bulk lymphocytes were allogeneically transferred between MHC-matched MCM revealed limited donor cell persistence beyond 14 days (31, 32), although increased donor cell persistence was observed for MHC-identical siblings (30). Autologous adoptive transfer provides distinct advantages in this context (30, 31). Furthermore, this approach prevents complications related to donor matching and graft-versus-host effects. However, previous autologous transfer experiments relied on the use of SIV-infected animals. In this study, we developed a novel protocol to isolate and rapidly expand SIV-specific CD8 T cell lines from SIV-naive MCM. The expansion of bulk lymphocytes required for these experiments represents a significant achievement considering the low precursor frequencies in uninfected animals. Autologous adoptive transfer of defined CD8 T cell populations followed by infectious challenge with SIV may help define the critical elements of an effective T cell response to HIV and SIV infection.

MATERIALS AND METHODS

Animal care and ethics statement.

Members of the Wisconsin National Primate Research Center (WNPRC) cared for all animals according to the regulations and standards set by the International Animal Care and Use Committee (IACUC). Two animals with the M1/M3 MHC haplotype were selected for autologous transfer. Genotyping was based on microsatellite analysis, and the MHC class I alleles on these haplotypes were characterized previously (27).

Autologous transfer and SIV challenge.

Expanded cell lines were processed by Ficoll-Paque Plus (GE Health Sciences, Piscataway, NJ) density centrifugation and resuspended in RPMI 1640 (HyClone, Logan, UT) supplemented with 10% fetal bovine serum (HyClone), 1% antibiotic-antimitotic (HyClone), and 1% l-glutamine (HyClone) (R10 medium). Numbers of cells from individual cell lines were counted, and these cells were then combined into a single homogenate for transfer. Cells were labeled with 80 μM carboxyfluorescein succinimidyl ester (CFSE) prior to infusion for tracking by flow cytometry, as described previously (32). After resuspension in 8 to 9 ml of 1× phosphate-buffered saline (PBS) supplemented with 15 U/ml heparin, cells were transfused into the saphenous vein by WNPRC staff. At 1 day posttransfer, animals were challenged intrarectally with 7,000 50% tissue culture infective doses (TCID₅₀) of SIVmac239 stock virus.

Peptides.

Most of the peptides used in this study were synthesized by GenScript (Piscataway, NJ), ProImmune (Sarasota, FL), or the Biotechnology Center at the University of Wisconsin (Madison, WI), using standard tertiary butyloxycarbonyl or fluorenylmethoxycarbonyl solid-phase methods. The 15-mer peptides used to stimulate CD8 T cell cultures and determine their restriction were provided by the NIH AIDS Research and Reference Reagent Program (Germantown, MD). All peptide sequences were derived from the SIVmac239 sequence.

Generation of MHC class I transferents.

Transferents expressing single MCM MHC class I alleles were generated by using either K562 cells with plasmids synthesized by GenScript or cells of the 721.221 HLA-deficient human B lymphoblastoid cell line (BLCLs) with full-length MHC class I amplicons ligated into pcDNA3.1(+) (Invitrogen, Carlsbad, CA), as described previously (33). Constructs were transfected by using Nucleofector kit C (Amaxa, Gaithersburg, MD). After 3 weeks of culture in R10 medium, K562 cells or 721.221 cells were stained with an anti-MHC class I antibody (W6/32) conjugated to phycoerythrin (PE), provided by David Watkins. Cells expressing MHC class I were then isolated by magnetically activated cell sorting (MACS) using anti-PE microbeads with LS columns (Miltenyi Biotec, Auburn, CA). Transferents were maintained under drug selection conditions in R10 medium with G418 (Mediatech, Manassas, VA).

IFN-γ ELISPOT assays.

Peripheral blood mononuclear cells (PBMCs) were isolated from EDTA-treated whole blood by using Ficoll-Paque Plus (GE Health Sciences) density centrifugation. Enzyme-linked immunosorbent spot (ELISPOT) assays were conducted according to the manufacturer's protocol. Briefly, 10⁵ cells in 100 μl of R10 medium were added to precoated monkey gamma interferon (IFN-γ) ELISpotPLUS plates (Mabtech Inc., Mariemont, OH) with 10 μM peptide. Assays of all samples were repeated in duplicate or triplicate. Full-proteome peptide sets comprised pools of 10 peptides, each at a concentration of 1 μM. Concanavalin A (10 μM) was used as a positive control. Cells alone in the absence of stimulant were used as a negative control. Wells were imaged by using an AID ELISPOT reader, and spots were counted by using an automated program with parameters including size, intensity, and gradient. Experimental responses exceeding the arithmetic mean of results for the negative-control wells plus two standard deviations were considered positive. The limit of detection was set at 50 spot-forming cells per million PBMCs.

Expansion of SIV-specific CD8 T cell lines.

PBMCs were isolated from EDTA-treated whole blood by Ficoll-Paque Plus (GE Health Sciences) density centrifugation. Cells were split into Corning T75 cell culture flasks (Fisher Scientific, Pittsburgh, PA) with 20 ml of RPMI 1640 (HyClone) supplemented with 15% fetal bovine serum (HyClone) and 100 IU/ml interleukin-2 (NIH AIDS Research and Reference Reagent Program, Germantown, MD) (R15-100 medium). Each T75 flask was stimulated with 10 μl of 1 mM peptide for each condition: Nef_103–111 RM9, Gag_389–394 GW9, Env_338–346 RF9, or Nef_254–262 LT9 (ProImmune). Every week thereafter, cell lines were restimulated with irradiated peptide-pulsed BLCLs as described previously (34). After a total of 4 weeks, cells were tetramer sorted by MACS (Miltenyi Biotec) and transferred into G-Rex10 flasks (Wilson Wolf, Saint Paul, MN) for rapid expansion. Resorting occurred every 2 weeks, in addition to restimulation, until at least one cell line achieved >50% specificity. Two weeks prior to infusion, cell lines were transferred into G-Rex100 flasks (Wilson Wolf) for further rapid expansion.

Tetramers.

Biotinylated peptide-MHC class I (pMHCI) monomers were synthesized as described previously (35). Tetramers were produced by mixing pMHCl monomers at a 4:1 molar ratio with purified streptavidin conjugated to either PE or allophycocyanin (APC). The tetramer staining protocol was modified slightly from a protocol described previously (36). Briefly, ∼1 × 10⁶ cells were resuspended in 100 μl of R10 medium with 1 μg/ml of the appropriate tetramer for 30 min at 37°C. Cells were then surface stained with 1 μl of anti-CD3-AlexaFluor700, 2.5 μl of anti-CD8-Pacific Blue, 5 μl of anti-CD14–PE-Texas Red and 5 μl of anti-CD19–PE-Texas Red (BD Biosciences, San Jose, CA) for 30 min at room temperature; washed with 1× PBS supplemented with 10% fetal bovine serum (fluorescence-activated cell sorting [FACS] buffer); and fixed with 2% paraformaldehyde (PFA) (Fisher Scientific). Samples were acquired by using an LSRII flow cytometer (BD Biosciences), and data were analyzed with FlowJo software (TreeStar Inc., Ashland, OR).

Phenotypic analysis.

Approximately 1 × 10⁶ CD8 T cells from cryopreserved homogenates were thawed and resuspended in 100 μl of R10 medium under each test condition. Cells were then stained for 30 min at room temperature with the following conjugates: (i) 1 μl of anti-CD3-AlexaFluor700, 5 μl of anti-CD4-APCH7, and 5 μl of anti-CD8-AmCyan (BD Biosciences); (ii) 5 μl of anti-CD28-Pacific Blue, 5 μl of anti-CCR7-fluorescein isothiocyanate (FITC), and 5 μl of anti-CCR9-peridinin chlorophyll protein (PerCP)-Cy5.5 (BioLegend, San Diego, CA); (iii) 5 μl of anti-CD95-PE-Cy7 (Fisher Scientific); (iv) 5 μl of anti-α4β7-APC (NIH Nonhuman Primate Reagent Resource); and (v) 5 μl of anti-CXCR3-PE (eBioscience, San Diego, CA). After washing with FACS buffer, cells were fixed with 2% PFA (Fisher Scientific). Samples were acquired by using an LSRII flow cytometer (BD Biosciences), and data were analyzed with FlowJo software (TreeStar Inc.).

Intracellular cytokine staining.

Activation of CD8 T cell lines was measured via intracellular staining for IFN-γ and tumor necrosis factor alpha (TNF-α), as described previously (34). Briefly, 5 × 10⁴ peptide-pulsed BLCLs were cocultured with 1 × 10⁵ CD8 T cells from each cell line. The following peptides were tested in each case: Nef_103–111 RM9, Gag_389–394 GW9, Env_338–346 RF9, and Nef_254–262 LT9. Brefeldin A (20 μl of a 1:100 dilution) was added, and the cells were incubated for 4 h at 37°C on an angle. Cells were then surface stained with 1 μl of anti-CD3-AlexaFluor700, 2.5 μl of anti-CD4-APC, and 2.5 μl of anti-CD8-Pacific Blue (BD Biosciences); washed once in FACS buffer; fixed with 2% PFA (Fisher Scientific); and left overnight at 4°C. The following morning, cells were permeabilized in FACS buffer containing 0.1% saponin (saponin buffer), stained with pretitrated concentrations of anti-IFN-γ–FITC and anti-TNF-α–PE (BD Biosciences), washed once in saponin buffer, and fixed with 2% PFA (Fischer Scientific). Samples were acquired by using an LSRII flow cytometer (BD Biosciences), and data were analyzed with FlowJo software (TreeStar Inc.).

Viral suppression assay.

Effector cells and infected targets were prepared as described previously (37). Briefly, targets were CD8 depleted by MACS (Miltenyi Biotec), incubated with concanavalin A (5 μg/ml) overnight, and washed. After 4 days, targets were washed again and plated at 2 × 10⁶ cells/ml in 48-well plates. Virus was prepared by layering SIVmac239 over sucrose and spinning at 14,000 rpm for 60 min. Targets were exposed to virus via magnetofection. Effectors were then plated with infected targets at the appropriate ratio. After 4 days in culture, cells were surface stained with anti-CD3-AlexaFluor700, anti-CD4-APC, and anti-CD8-Pacific Blue as described above and then stained intracellularly with a pretitrated concentration of anti-p27-FITC (NIH AIDS Research and Reference Reagent Program) combined with bulk permeabilization reagent (Invitrogen). After washing, cells were fixed with 2% PFA (Fischer Scientific) and analyzed by flow cytometry as described above. Values were normalized by dividing the average percentage of p27-positive (p27⁺) cells in the experimental wells by the average percentage of p27⁺ cells in the infected-control wells.

Isolation of tissues.

Fresh PBMCs were isolated from EDTA-treated whole blood by Ficoll-Paque Plus (GE Health Sciences) density centrifugation. Bronchoalveolar lavage (BAL) fluid was passed through a 70-μm filter (BD Biosciences) and centrifuged. Tissues were diced with a scalpel, pressed over a 100-μm filter (BD Biosciences), and washed through with R10 medium.

Plasma viral load.

SIVmac239 plasma viral loads were measured as described previously (38, 39). Briefly, viral RNA was reverse transcribed and then quantified by using a LightCycler 2.0 or LightCycler 480 instrument (Roche, Indianapolis, IN). Serial dilutions of SIV Gag in vitro transcripts were used as an internal standard curve for each run. The limit of detection was 100 viral RNA copies/ml plasma.

RESULTS

Characterization of nine novel SIV-specific CD8 T cell responses in MCM.

Previously, we identified optimal SIV-derived epitope-specific CD8 T cell responses restricted by MHC class I molecules encoded universally across all MCM haplotypes and those confined to M1/M3 (27, 34, 40, 41). Although the M1 and M3 haplotypes are extremely common, animals are often heterozygous for one of the other five major MHC haplotypes. We therefore mapped additional responses for the M2, M4, M5, M6, and M7 haplotypes to account for the entire MHC diversity of MCM. Using a full-proteome ELISPOT assay, we identified nine novel responses and established their MHC restriction by testing SIV-specific CD8 T cell lines against BLCLs transfected with a single MHC class I allele and pulsed with the appropriate 15-mer peptides (Fig. 1A). Furthermore, we determined the minimal optimal epitope sequence in each case by pulsing BLCLs with serial dilutions of progressively truncated peptides (Fig. 1B). In this way, we identified responses specific for Nef_254–262 LT9, Pol_1030–1038 RY8, Gag_437–444 LP8, Env_661–669 QL9, Gag_34–41 VL8, VpX_19–27 GR9, Pol_639–648 DT10, Nef_216–223 DY8, and Gag_255–263 NY9 and generated MHC tetramers to confirm epitope specificity (Table 1). The discovery of the dominant MHC-restricted SIV-specific CD8 T cell responses in MCM facilitated our selection of specific epitopes for the naive expansion experiments.

FIG 1.

Identification of SIV-specific CD8 T cell responses. Peptide-pulsed BLCLs were cultured with SIV-specific CD8 T cell lines. Activation was measured by intracellular cytokine staining with anti-IFN-γ–FITC and anti-TNF-α–PE. (A) Representative activation of a CD8 T cell line specific for Env_661–669 QL9. CD8 T cell lines were tested against MHC-matched BLCLs or transferents expressing specific alleles. (B) Representative determination of the minimal optimal epitope sequence for Gag_255–263 NY9. MHC-matched BLCLs were pulsed with serial dilutions of prospective optimal peptides. Responses were normalized to the percentage of the maximum response.

TABLE 1.

Characterization of MHC-restricted SIV-specific CD8 T cell responses

Protein and amino acids positive for restriction	Sequence positive for restriction, as determined by:		Establishment of SIV-specific CD8 T cell line	Allele specificity	Haplotype	Optimal epitope sequence	Optimal epitope	Observed tetramer response
	Full-proteome ELISPOT assay	Peptide scanning
Tat 59–67 (41)		CCYHCQFCFC	Yes	A1*063:02	Universal	CCYHCQFCF	Tat CF9
Env 338–346 (41)	QPINDRPKQAWCWFG	RPKQAWCWF	Yes	A1*063:02	Universal	RPKQAWCWF	Env RF9	Yes
Nef 103–111 (34)	VRPKVPLRTMSYKLA		Yes	A1*063:02	Universal	RPKVPLRTM	Nef RM9	Yes
Gag 386–394 (34)	AAQQRGPRKPIKCWN		Yes	A1*063:02	Universal	GPRKPIKCW	Gag GW9	Yes
Pol 591–598 (34)	IWGQVPKFHLPVEKD		Yes	A4*01:01	Universal	QVPKFHLP	Pol QP8	Yes
Nef 254–262	TARGLLNMADKKETR		Yes	B*104:01	M1	LNMADKKET	Nef LT9	Yes
Pol 1030–1038	KVVPRRKAKIIKDYG		Yes	B*150:01:01	M2	RKAKIIKDY	Pol RY8	Yes
Rev 59–68 (41)	RIYSFPDPPTDTPLD	SFPDPPTDT	Yes	B*075:01	M3	SFPDPPTDTP	Rev SP10	Yes
Gag 459–467 (41)		TAPPEDPAV	Yes	B*075:01	M3	TAPPEDPAV	Gag TV9
Env 620–628 (41)		TVPWPNASL	Yes	B*075:01	M3	TVPWPNASL	Env TL9
Gag 221–229 (41)		PAPQQGQLR	Yes	B*075:01	M3	PAPQQGQLR	Gag PR9
Gag 146–154 (41)	VHLPLSPRTLNAWVK	HLPLSPRTL	Yes	B*075:01	M3	HLPLSPRTL	Gag HL9	Yes
Tat 42–29 (41)	SQLYRPLEACYNTCY		Yes	B*075:01	M3	QLYRPLEA	Tat QA8	Yes
Gag 28–37 (41)	GKKKYMLKHVVWAAN		Yes	B*011:01	M3	KYMLKHVVWA	Gag KA10
Gag 437–444	LGPWGKKPRNFPMAQ		Yes	B*147:01	M4	LGPWGKKP	Gag LP8	Yes
Env 661–669	EAQIQQEKNMYELQK		Yes	A1*031:01	M4	QQEKNMYEL	Env QL9
Gag 34–41	YMLKHVVWAANELDR		Yes	B*050:04	M5	VVWAANEL	Gag VL8	Yes
VpX 19–27	TIGEAFEWLNRTVEE		Yes	B*095:01	M6	GEAFEWLNR	VpX GR9	Yes
Pol 639–648	VKDPIEGEETYYTDG		Yes	A1*060:05	M7	DPIEGEETYT	Pol DT10
Nef 216–223	EVLAWKFDPTLAYTY		Yes	A1*060:05	M7	DPTLAYTY	Nef DY8	Yes
Gag 255–263	WMYRQQNPIPVGNIY		Yes	A1*060:05	M7	NPIPVGNIY	Gag NY9

Rapid expansion can generate highly specific CD8 T cell lines that express common phenotypic homing markers.

Early CD8 T cell responses typically recognize epitopes present in the wild-type virus but do not recognize epitope escape variants in the later stages of infection. We suspected that these initially mobilized T cell populations might have higher precursor frequencies, which would facilitate their expansion from naive animals. Accordingly, we hypothesized that epitopes targeted by acute immunodominant CD8 T cell responses could be used to expand specific cell lines from SIV-naive animals for subsequent in vivo efficacy testing. Using PBMCs isolated from two SIV-naive MCM, we cultured cell lines specific for four SIV-derived CD8 T cell epitopes, Nef_103–111 RM9, Gag_389–394 GW9, Env_338–346 RF9, and Nef_254–262 LT9. We measured the specificity of these expanded lines by MHC tetramer analysis prior to transfer (Fig. 2). Cell lines specific for Gag_389–394 GW9 expanded vigorously in both animals and reached up to 94.4% purity. In contrast, cell lines specific for Env_338–346 RF9 and Nef_254–262 LT9 from animal cy0573 did not expand above 1% purity, whereas the corresponding cultures from animal cy0574 expanded to specificities of >60%.

FIG 2.

Specificity of CD8 T cell lines determined by MHC tetramer analysis. The specificity of each SIV-specific CD8 T cell line was assessed prior to CFSE labeling and transfer. Plots are pregated on CD3⁺ lymphocytes. Frequencies of tetramer-positive CD8⁺ cells are shown in the depicted gates (boxed areas).

To determine whether these rapidly expanded CD8 T cell lines displayed homing markers typically associated with an effector memory phenotype, we conducted flow cytometric analysis to evaluate the surface expression of α4β7, CXCR3, CCR7, and CCR9 (Table 2). The combined pretransfer homogenates from both animals cy0573 and cy0574 were tested and compared to that from a PBMC control (animal cy0391). In both animals cy0573 and cy0574, the expanded CD8 T cell populations expressed CXCR3 at frequencies of >97%. Lower levels of α4β7 expression than those in the control were observed, and neither CCR7 nor CCR9 was detected in significant amounts. These phenotypic characteristics are consistent with effector memory differentiation.

TABLE 2.

Phenotypic homing markers expressed by CD8 T cell lines

Cell line	Frequency (%) of CD8 T cell population
	α4β7	CCR7	CXCR3	CCR9
cy0391a	73.3	26.8	50.4	2.65
cy0573	36.6	0.17	97.2	2.46
cy0574	17.4	0.59	97.2	2.63

^aPBMC control sample from a previous adoptive transfer experiment at 1 year postinfection.

Cytokine production in RM9-specific CD8 T cell lines.

To evaluate functional activation, each SIV-specific CD8 T cell line was cocultured with cognate peptide-pulsed BLCLs, irrelevant peptide-pulsed BLCLs, or unpulsed BLCLs. Intracellular production of IFN-γ and TNF-α was measured in each case by flow cytometry (Fig. 3). The Nef_103–111 RM9-specific CD8 T cell lines responded to cognate peptide stimulation with the highest frequencies of cytokine-producing cells (12.4% IFN-γ⁺ TNF-α⁺ cells for animal cy0573 and 47.7% IFN-γ⁺ TNF-α⁺ cells for animal cy0574). These cell lines were also tested against BLCLs pulsed with Nef_254–262 LT9. The frequencies of detection of IFN-γ⁺ TNF-α⁺ cells were 7.3% for animal cy0573 and 1.2% for animal cy0574, compared to 7.1% and 1.4%, respectively, in the absence of peptide. Thus, there was no cytokine production above baseline levels in the presence of an irrelevant peptide. Strong responses were not observed for the CD8 T cell lines specific for Gag_389–394 GW9, Env_338–346 RF9, and Nef_254–262 LT9 (<5% IFN-γ⁺ TNF-α⁺ cells). These data suggest that Nef_103–111 RM9-specific CD8 T cells are functionally more potent than CD8 T cells specific for the other three SIV-derived epitopes.

FIG 3.

Cytokine production by SIV-specific CD8 T cell lines. CD8 T cell lines specific for Nef_103–111 RM9, Gag_389–394 GW9, Env_338–346 RF9, and Nef_254–262 LT9 were cocultured with peptide-pulsed BLCLs, as indicated. The frequency of dual IFN-γ and TNF-α cytokine-producing cells measured by flow cytometry under each condition is displayed on the y axis for each SIV-specific CD8 T cell line (x axis).

Suppression of viral replication by RM9-specific CD8 T cell lines.

Next, we measured the functionality of these CD8 T cell lines in terms of viral suppression. Each SIV-specific cell line was tested individually and as a combined homogenate of all four specific cell lines, both before and after cell transfer (Fig. 4). In each animal, PBMCs isolated before transfer and after transfer showed increased viral replication compared to that of the baseline control (no effectors). The Nef_103–111 RM9-specific CD8 T cell line from animal cy0574 strongly suppressed viral replication (>90%), while the remaining epitope-specific cell lines were markedly less effective. Suppression by the combined homogenate was comparable to that mediated by the CD8 T cell lines specific for Gag_389–394 GW9, Env_338–346 RF9, and Nef_254–262 LT9 in isolation. Similar results were observed with the corresponding cell lines from animal cy0573. In this case, CD8 T cells specific for Nef_103–111 RM9 suppressed viral replication by >30%, while the other epitope-specific cell lines and the combined homogenate were substantially less potent. Together with the intracellular cytokine staining data, these results suggest that Nef_103–111 RM9-specific cells are the best option for therapeutic transfer in MCM.

FIG 4.

Suppression of viral replication by CD8 T cell lines. Viral suppression assays were conducted with each SIV-specific CD8 T cell line and the combined homogenate, as indicated. Nef_103–111 RM9-specific CD8 T cell lines from both animals suppressed viral replication most effectively compared to the other specific cell lines. PBMCs from recipient blood taken immediately prior to transfer, immediately after transfer, and at 1 day posttransfer appeared to have increased viral replication. Values were normalized by dividing the average percentage of p27⁺ cells in the experimental wells by the average percentage of p27⁺ cells in the control wells (no effectors).

Autologous transfer of CD8 T cells into naive animals.

To prepare for autologous transfer into naive animals, the rapidly expanded cells were first counted to determine the frequency of each SIV-specific cell line in the combined transfusate (Fig. 5). A combined total of 2.82E+09 cells was transferred into animal cy0573, while a combined total of 2.69E+09 cells was transferred into animal cy0574 (Fig. 5A). We discovered contamination in two of the G-Rex100 flasks on the day of transfer, so Nef_103–111 RM9-specific cells from backup G-Rex10 flasks were used as substitutes, which resulted in smaller overall quantities of CD8 T cells with this specificity because G-Rex10 flasks have 1/10 of the surface area of G-Rex100 flasks (Fig. 5B). However, we estimated that the purity of these Nef_103–111 RM9-specific cell lines was sufficient to include the smaller cell populations in the transfusate.

FIG 5.

Cell counts and composition of the transfusate. (A) Cell counts for SIV-specific CD8 T cell lines. Due to contamination of two of the G-Rex100 flasks on the day of transfer, Nef_103–111 RM9-specific cells from backup G-Rex10 flasks were used, resulting in lower absolute numbers for this specific cell line. (B) Percentage of cells from each SIV-specific CD8 T cell line that contributed to the total transfusate.

Prolonged persistence of transfused CD8 T cells in multiple tissues.

Ensuring cell persistence in recipients is a major challenge for adoptive transfer. In the autologous setting, however, this is less problematic because there is no genetic variance. We monitored the persistence of autologously transferred CFSE⁺ donor cells in animals cy0573 and cy0574 for a period of 24 weeks posttransfusion (Fig. 6). Animal cy0573 was necropsied at 37 days postinfection (38 days posttransfusion) due to complications unrelated to the experiment. After transfer, there was a rapid decrease in the frequency of CFSE⁺ CD8⁺ T cells over the first 2 weeks. Nevertheless, transferred cells remained detectable in peripheral blood for at least 24 weeks posttransfer in animal cy0574 (Fig. 6A). Substantially higher frequencies of transferred cells were observed in BAL fluid. Moreover, these cells showed evidence of ongoing proliferation (Fig. 6B). In animal cy0573, we detected donor cells at frequencies of up to 45% in BAL fluid at 14 days postinfection; at the time of necropsy, these cells were still detectable at a frequency of 4.9%. In animal cy0574, donor cells persisted in BAL fluid until at least 8 weeks postinfection (0.45% CFSE⁺ CD8⁺ cells). Notably, these frequencies are higher than those observed for peripheral blood even at 1 week postinfection, suggesting that donor cells may become trapped or traffic predominantly to the lung. For animal cy0573, other tissue sites were examined at necropsy (Fig. 6C). Donor cells were detected in spleen, bone marrow, and internal iliac lymph nodes, with trace amounts being present in other lymph nodes. These data provide evidence that autologously transferred SIV-specific CD8 T cells traffic to multiple tissues and persist for several weeks and may have implications for control of the viral reservoir.

FIG 6.

Cell trafficking and persistence in multiple tissues. Animal cy0573 was necropsied at 37 days postinfection. (A) CFSE⁺ donor cells persisted in peripheral blood for at least 24 weeks postinfection. (B) CFSE⁺ donor cells trafficked to the lung at high frequencies, as measured in BAL fluid, and persisted until at least 8 weeks posttransfer. dpi, days postinfection. (C) CFSE⁺ donor cells trafficked to spleen and bone marrow, as detected postnecropsy. Trafficking to lymph nodes (LN), especially the internal iliac lymph nodes, was also observed.

No observed impact of CD8 T cells on acute-phase viral loads following autologous transfer.

Given that SIV-specific CD8 T cells control viral replication in vitro, we hypothesized that we would detect a change in the acute-phase viral load profile after autologous transfer and subsequent infection. At 1 day posttransfer, animals were challenged intrarectally with 7,000 TCID₅₀ of SIVmac239 stock virus. Historically, the peak viral load is typically 10⁷ viral RNA copies/ml during acute-phase SIV infection (38). In both animals cy0573 and cy0574, plasma viral loads were consistent with those of historic controls, peaking at 10⁷ viral RNA copies/ml (Fig. 7). Plasma viral loads were monitored through 24 weeks postinfection, with no detectable impact during the acute or chronic phase.

FIG 7.

SIVmac239 challenge after autologous CD8 T cell transfer. Animals were challenged intrarectally with a standard high dose of SIVmac239 (7,000 TCID₅₀) 1 day after autologous transfer of SIV-specific CD8 T cells. Quantitative PCR was used to detect viral RNA copies in plasma. Animal cy0573 was necropsied at 37 days postinfection. Historic Control represents median viremia values from a previous study where animals were challenged with 1,000 TCID₅₀ (red).

DISCUSSION

MCM are an ideal experimental model for immunological studies because they display limited genetic diversity. Here, we identified and characterized MHC class I-restricted SIV-specific CD8 T cell responses accounting for all seven major haplotypes. These newly defined immunodominant specificities expand the potential for adoptive transfer studies in SIV-infected MCM (31, 32).

One of the major challenges with adoptive transfers is to ensure cell persistence in recipient animals while maintaining a functional effector phenotype (28–32). Typically, CD8 T cells expanded in vitro have a limited ability to proliferate, making it difficult to achieve the significant numbers required for adoptive transfer (28). In this study, we developed a novel protocol to isolate and expand SIV-specific CD8 T cell lines from naive MCM. This rapid expansion method provides bulk quantities of CD8 T cells suitable for transfer within a relatively short period of time. Although CD8 T cells amplified in vitro typically display a fully differentiated phenotype, we found that rapid expansion induced an effector memory profile with characteristic patterns of homing marker expression. Additionally, we demonstrated that these expanded cells persisted in the recipient animals longer than in previous adoptive transfer studies (28–32). More detailed phenotypic studies in vivo were precluded by CFSE labeling prior to transfer. We also observed important differences among expanded cells with distinct specificities. CD8 T cells specific for Nef_103–111 RM9 were readily amplified and achieved very high levels of purity. Moreover, the resulting Nef_103–111 RM9-specific CD8 T cell lines were most effectively able to produce cytokines and suppress viral replication in vitro. Accordingly, we predict that in vivo viral suppression would have been pronounced had we been able transfer larger quantities of these cells. In contrast, CD8 T cells specific for Gag_389–394 GW9, Env_338–346 RF9, and Nef_254–262 LT9 were poorly functional and therefore unlikely to suppress viral replication in vivo after transfer, although our results do no preclude in vivo efficacy in future experiments. It is notable that we observed an increase in posttransfer viral replication in our suppression assays. This likely reflects the use of fresh PBMCs, which provide supplemental targets for SIV infection. Moving forward, we will continue to optimize these protocols to culture CD8 T cell lines specific for Nef_103–111 RM9 with the intention of performing further autologous transfers to assess in vivo efficacy both during acute SIV infection and in the context of a shock-and-kill strategy to purge viral reservoirs in previously infected animals.

Previously, we showed that adoptively transferred cells persisted for up to 14 days in MHC-matched MCM (31, 32). In this autologous transfer study, CFSE⁺ donor cells persisted through 24 weeks postinfection, a period significantly longer than in previous transfer studies conducted in our laboratory. Although donor cells were detected in multiple tissue sites, the highest frequencies were observed in the lung. This “trapping” of these relatively large cells is a previously described phenomenon in transfer studies in both the MCM model and mice (28, 29, 42). Such anatomical preferences may provide another explanation for the lack of an impact on the SIV load. It is therefore possible that greater efficacy would be observed in vivo if more cells were able to traffic to other tissues. Necropsy of animal cy0573 at 37 days postinfection revealed trafficking to spleen and bone marrow, with a minimal presence in lymph nodes. These trafficking sites may provide additional insights into our understanding of how these potential effector cells interact with the latently infected reservoir.

It may be possible to utilize a conditioning regimen to eliminate recipient immune cells prior to transfer and increase cell trafficking. However, this approach would result in a depletion of lymphocyte targets for viral infection. Alternatively, a “prime-and-pull” vaccine strategy may be effective in establishing memory T cells at peripheral tissue sites affected in the early phases of SIV infection (43). In this scheme, active T cells are “primed” through conventional vaccination strategies and then “pulled” to a localized site via topical chemokine application, resulting in protective immunity via the establishment of a stable memory T cell population. This method prevented the development of disease in herpes simplex virus 2 (HSV-2)-infected mice but demonstrated only a modest effect on antibody responses to HIV gp140. We plan to utilize this strategy to target active T cells to localized sites of initial viral replication and bypass the trapping phenomenon in the lung, potentially improving the impact on SIV.

In the present study, autologous transfer of SIV-specific CD8 T cells ultimately failed to control SIV replication. Peak viral loads were comparable to those of historic controls, and no significant changes were observed during the chronic phase. The lack of cytokine production by CD8 T cells specific for Gag_389–394 GW9, Env_338–346 RF9, and Nef_254–262 LT9 may explain this lack of efficacy. In future studies, we plan to optimize our approach to focus on Nef_103–111 RM9-specific CD8 T cells, which exhibited potent cytokine production and suppression of viral replication. Trafficking cells out of the lung by using a prime-and-pull strategy may help direct T cells to the sites of viral replication, while an activating agent in the context of ART may coax latent virus out of reservoir sites. These supplemental strategies could further serve to enhance virus control. Thus, the protocol developed here represents an important first step toward the optimization of autologous transfer strategies with potential therapeutic utility.