IMMUNOBIOLOGY

HIV-1 Nef-induced FasL induction and bystander killingrequires p38 MAPK activationKaruppiah Muthumani, Andrew Y. Choo, Daniel S. Hwang, Arumugam Premkumar, Nathanael S. Dayes, Crafford Harris, Douglas R. Green,Scott A. Wadsworth, John J. Siekierka, and David B. Weiner

The human immunodeficiency virus (HIV)has been reported to target noninfectedCD4 and CD8 cells for destruction. Thiseffect is manifested in part through up-regulation of the death receptor Fas li-gand (FasL) by HIV-1 negative factor (Nef),leading to bystander damage. However,the signal transduction and transcrip-tional regulation of this process remainselusive. Here, we provide evidence thatp38 mitogen–activated protein kinase

(MAPK) is required for this process. Loss-of-function experiments through domi-nant-negative p38 isoform, p38 siRNA,and chemical inhibitors of p38 activationsuggest that p38 is necessary for Nef-induced activator protein-1 (AP-1) activa-tion, as inhibition leads to an attenuationof AP-1–dependent transcription. Further-more, mutagenesis of the FasL promoterreveals that its AP-1 enhancer element isrequired for Nef-mediated transcriptional

activation. Therefore, a linear pathway forNef-induced FasL expression that encom-passes p38 and AP-1 has been eluci-dated. Furthermore, chemical inhibitionof the p38 pathway attenuates HIV-1–mediated bystander killing of CD8 cells invitro. (Blood. 2005;106:2059-2068)

© 2005 by The American Society of Hematology

Introduction

Human immunodeficiency virus (HIV) infection typically results inthe eradication of the host’s immune system through eventualdepletion of its CD4 cells.1-3 Although initially controlled, the viruspersists and invariably replicates to high titers. In this regard,disregulated apoptosis is considered a major pathogenesis eventleading to severe CD4 lymphopenia during HIV-1 infection.3

HIV-induced apoptosis of host cells has been reported to both involveand not involve the Fas/Fas ligand (FasL) apoptotic pathway. FasL is notpresent on resting T cells, but activated T cells may undergo apoptosisthrough the CD95/CD95 ligand (CD95L) pathway.4-7

Specifically, negative factor (Nef) has been reported to induceapoptosis of bystander cells while protecting infected cells. Nefexpression in T cells induces FasL expression on the infectedcell.8,9 This proposed mechanism is also detected in lymph nodes ofsimian immunodeficiency virus (SIV)–infected monkeys and HIV-infected patients, further suggesting its importance.5 Concomi-tantly, Nef expression in T cells induces signals that protect theinfected cell from the same Fas-mediated cell death via B-celllymphoma 2 (Bcl2)–antagonist of cell death (Bad) phosphoryla-tion and apoptosis signal-regulating kinase 1 (ASK1) inhibi-tion.10,11 Ideally, this would allow the virally infected cell topersist in the face of the host immune response by becomingresistant to apoptosis while provoking localized destruction ofneighboring cells and effector T cells attempting to mediateviral factory clearance.

Previously, stimulation by Nef of FasL has been directly linkedto Nef’s ability to bind to the CD3 � chain of the T-cell receptor(TCR).8 This interaction is mediated by the proline-rich domain ofNef, which potentiates its interaction with Nef-associated kinase(NAK/p62).9,10 However, the downstream signals that regulate thisNef-mediated FasL transcription remain undetermined. Here, wereport that mitogen-activated kinase p38 is necessary for Nef-mediated FasL transcription. Further, disruption of p38 leads to theinhibition of activator protein 1 (AP-1)–dependent transcription,which we show is necessary for FasL up-regulation.

Materials and methods

Cell culture

Jurkat, 293T, and the monocyte line U937 were obtained from the AmericanType Culture Collection (ATCC, Rockville, MD). Cells were passaged inRPMI-1640 or Dulbecco modified Eagle medium (DMEM; Gibco, Carls-bad, CA) supplemented with 10% fetal bovine serum (FBS), 2 mML-glutamine, 100 U/mL penicillin G, and 100 �g/mL streptomycin. Cellswere maintained at 37°C and 5% CO2. Leukopacks from individual donorswere obtained from the CFAR clinical core facilities at the University ofPennsylvania (UPENN) School of Medicine to isolate T cells as well asmonocytes. Peripheral blood mononuclear cells (PBMCs) were isolated byFicoll-Hypaque (Pharmacia, Piscataway, NJ) density centrifugation andcultured as described previously.4,12 CD14 monocytes were purified from

From the Department of Pathology and Laboratory Medicine, University ofPennsylvania School of Medicine, Philadelphia, PA; the MemorialSloan-Kettering Cancer Center, New York, NY; Johnson & JohnsonPharmaceutical Research and Development, Raritan, NJ; La Jolla Institute forAllergy and Immunology, San Diego, CA; and the Department of Cell Biology,Harvard Medical School, Boston, MA.

Submitted March 9, 2005; accepted May 19, 2005. Prepublished online asBlood First Edition Paper, May 31, 2005; DOI 10.1182/blood-2005-03-0932.

Supported by a Johnson & Johnson Pharmaceutical Research and Devel-opment (PRD) grant to D.B.W. and K.M. Supported also by the NationalInstitutes of Health (NIH) AIDS Research and Reference Reagents program

and the Center for AIDS Research (CFAR), University of Pennsylvania.

An Inside Blood analysis of this article appears in the front of this issue.

Reprints: David B. Weiner, University of Pennsylvania School of Medicine,Department of Pathology & Laboratory Medicine 422 Curie Blvd, Philadelphia,PA 19104; e-mail: dbweiner@mail.med.upenn.edu.

The publication costs of this article were defrayed in part by page chargepayment. Therefore, and solely to indicate this fact, this article is herebymarked ‘‘advertisement’’ in accordance with 18 U.S.C. section 1734.

© 2005 by The American Society of Hematology

2059BLOOD, 15 SEPTEMBER 2005 � VOLUME 106, NUMBER 6

PBMCs by positive enrichment using autoMACS (Miltenyi Biotec, Au-burn, CA) according to the manufacturer’s instructions. Monocytes werecultured in complete RPMI-1640 medium with 500 U/mL recombinanthuman interleukin-4 (rhIL-4) and 1000 U/mL recombinant human granulo-cyte-macrophage colony-stimulating factor (rhGM-CSF; both from R&DSystems, Minneapolis, MN) per 106 cells.4,12

Mitogen-activated protein kinase (MAPK) inhibitorand plasmids

The p38 inhibitors SB203580 and RWJ67657 have been previouslydescribed.13,14 The wild-type Nef (pNef) or pNef (PxxPxR) alaninesubstitution plasmids were amplified by single-round polymerase chainreaction (PCR) using Nef-specific primers or overlap extension PCR andsubcloned into the pVax vector (Invitrogen, Frederick, MD). The wild-typep38 MAP kinase, dominant-negative p38 (KM mutant; K to M mutation atthe adenosine triphosphate [ATP]–binding site), was constructed.15 HumanFasL (hFasL) reporter expression vector hFasL-Luc (1.2-kb FasL promoterreporter construct) and hFasLmut (mutations in the AP-1 site were intro-duced into the 1.2-kb FasL promoter construct) were previously described.16

Four sets of p38 shRNA oligonucleotides (oligos) were designed forp38 sequence using siRNA target finder with appropriate controls fromAmbion (Austin, TX) web resources. For each set, top- and bottom-strandoligos were synthesized separately and annealed together. The primers forsiRNA corresponding to the coding region were made corresponding to thesequence as follows: clone p38-61: top, 5�-gatccACATGAGAAGCTAGGC-GAGTTcaagagaCTCGCCTAGCTTCTCATGTTTTTTTGGAAA-3� andbottom, 5�-AGCTTTTCCAAAAAA CTCCTGGTAGAGAAGAAAAtctct-tgAATTTTCTTCTCTACCAGGAGg-3�; clone p38-352: top, 5�-gatccGCT-CATGAAACATGAGAAGTTcaagagaCTTCTCATGTTTCATGAGCTT-TTTTGGAAA-3� and bottom, 5�-AGCTTTTCCAAAAAAGCTCATGAAACATGAGAAGtctcttgAACTTCTCATGTTTCATGAGCg-3�; andclone p38-1016: top, 5�-gatccTGGAAGCGTGTTACTTACATTcaagaga-TGTAAGTAACACGCTTCCATTTTTTGGAAA-3� and bottom, 5�-AGCT-TTTCCAAAAAATGGAAGCGTGTTACTTACAtctcttgAATGTAAGTAAC-ACGCTTCCAg-3�. The selection of siRNA sequences followed the criteriaprovided byAmbion. The BamHI restriction site was included on each top strandat its 5� end; similarly, on the bottom strand, the HindIII site was included at the 3�end to facilitate cloning into the pSilencer 2.0-U6 neo vector (Ambion). In the topstrand, the uppercase nucleotides correspond to the nucleotide open readingframe (ORF) of the mRNA sequence. Next to the mRNA sequence, a 7-base(caagaga) spacer sequences was included to facilitate to form a small hairpinloop. The second stretch of uppercase nucleotides is the reverse complement ofthe first. The 2 oligos were annealed and inserted into a pSilencer 2.0-U6 neovector (Ambion) that had been linearized with BamHI and HindIII enzymes.Positive clones were identified and sequences were conformed by automatedDNA sequencing. The pSilencer 2.0-U6 neo vector, which contains a 21-basescrambled sequence and has no significant homology to mouse, rat, or humangene sequences, was used as a negative control to transfect cells.17

Transfection and luciferase reporter assay

Jurkat or U937 (2 � 106) cells were washed twice with phosphate-bufferedsaline (PBS), resuspended in 500 �L Opti-MEM (Gibco) culture medium,and transfected with LipofectAMINE Plus (Gibco). The plasmid pNef(5 �g), with a mixture of 2 �g cytomegalovirus (CMV) vector expressinggreen fluorescence protein (GFP), was transfected with the FuGENE 6Reagent (Roche Applied Science, Indianapolis, IN). Cells were washedtwice with complete RPMI-1640 and incubated for 48 hours. Transfectionefficiencies were assessed by fluorescence-activated cell sorter (FACS)analyses of GFP expression. For determination of AP-1 and FasL reporterassay, 2.5 �g reporter plasmid (pAP1-Luc or hFasL-Luc) alone or withpNef (5 �g), mutated pNef with or without inhibitors (1 �M), pNef orpNef plus p38Wt, or pNef plus p38 dominant-negative construct (5 �g)were added to the cells and mixed well as indicated. Electroporation wascarried out at 250 V and 960 �F in a Bio-Rad Gene Pulse II (Bio-Rad,Hercules, CA).16 Cells were grown in complete RPMI 1640 medium for48 hours at 37°C.

Total amounts of DNA and equal molar ratios of promoters were keptconstant in all setups by using empty vectors.16 Because of differences intransfection efficiencies, an expression plasmid pCMV �-galactosidase(�-gal) was cotransfected as a transfection efficiency control, and luciferaseactivities were normalized based on �-gal activity with the �-gal reportergene assay.16 Cells were harvested, washed 3 times with PBS, and lysed in100 �L reporter lysis buffer (RLB) according to the manufacturer’sinstructions (Roche Applied Science). Cell debris was removed by centrifu-gation, and the supernatant was used in the luciferase assay usingLUMAT-LB9501 (Berthold, Bad Wildbad, Germany).16

HIV-1 pseudotype viruses and infection

The HIV-1 proviral infectious constructs pNL4-3/HSA, pNL4-3/HSA/�Env, or pNL4-3/HSA/�Env/�Nef and primary clade specific isolates(subtype A-/94UG103, subtype-B/92US723, subtype-C/96USNG31, andsubtype-D/92UG001) were obtained through the AIDS Research andReference Reagent (RRR)-Program, National Institute of Allergy andInfectious Diseases (NIAID), NIH.18 Constructs containing �Nef weregenerated by recombinant PCR mutagenesis.19 HIV-1 viral particles weregenerated by transfection with pNL4-3/HSA alone or cotransfection ofpNL4-3/HSA/�Env or pNL4-3/HSA/�Env/�Nef with vesicular stomatitisvirus G (VSV-G) envelope18 by FuGENE 6 transfection in 293T cells.Infection was carried out by incubating the cells with HIV-1 virus at aconcentration of 100 tissue-culture infective dose (TCID50)/106 cells permilliliter. These retroviruses encode as a specific cell-surface marker, themurine CD24 antigen (heat-stable antigen [HSA]), which allows foridentification of infected target cells by FACS analysis.12,18 Culturesupernatant was collected at 6-, 12-, and 24-hour intervals and assayed forvirus production by measuring p24 antigen by enzyme-linked immunosor-bent assay (ELISA) kit (Beckman Coulter, Fullerton, CA). Data arepresented as mean plus/minus SEM.

Flow cytometry

Cell suspensions (106) were washed in PBS (pH 7.2) containing 0.2%bovine serum albumin and 0.1% NaN3. Cells were stained with 1 to 2 �g ofthe following monoclonal antibodies (mAb): CD3 fluorescein isothiocya-nate (FITC) mouse immunoglobulin G2a (IgG2a), CD4 phycoerythrin (PE)mouse IgG1, CD8 FITC mouse IgG1, CD14 FITC and PE mouse IgG2a(PharMingen, San Diego, CA), and FasL (NOK-1) mouse IgG1 (eBio-science, San Diego, CA). For intracellular p24gag, staining was performedby fixing 106 cells in 1% paraformaldehyde containing 20 �g/mL lysoleci-thin (Sigma, St Louis, MO) for 5 minutes at 4°C. Samples were thenwashed, permeabilized in 2 mL ice-cold methanol while vortexing, placedon ice for 15 minutes, washed again, and resuspended in 1 mL PBS/0.1%nonidet P-40 (NP-40; Sigma) for 5 minutes at 4°C. Cells were washed andincubated with 1 �g PE-conjugated KC57-RD1 (Coulter) for 15 minutes at4°C. Cells were washed with PBS and resuspended in 200 �L PBS andanalyzed directly on a Coulter EPICS Flow Cytometer (Coulter) usingFlowJo software (TreeStar, San Carlos, CA). All samples were comparedwith their isotype-matched controls. In the case of dual flow cytometry,individual samples treated with each isotype alone were used to determinethe background levels of autofluorescence.12

Apoptosis, caspase, and FasL ELISA assays

FACS analysis was performed to identify cells undergoing apoptosis.4

Apoptosis was evaluated by using an annexin-V assay kit (PharMingen)and analyzed directly on a Coulter EPICS Flow Cytometry using FlowJosoftware. Caspase-3 activity was determined using Caspase-3/CPP32colorimetric protease assay kit according to the manufacturer’s instructions(MBL, Watertown, MA). FasL in the supernatant or lysates was quantifiedby sandwich ELISA with commercial ELISA (R&D Systems) according tothe manufacturer’s instructions.

Protein extraction and Western blotting

Cells were washed with ice-cold PBS and lysed with protein lysis buffer(20 mM Tris [tris(hydroxymethyl)aminomethane, pH 7.4], 150 mM NaCl,

2060 MUTHUMANI et al BLOOD, 15 SEPTEMBER 2005 � VOLUME 106, NUMBER 6

1 mM EDTA [ethylenediaminetetraacetic acid], 1 mM EGTA [ethyleneglycol tetraacetic acid], 1% triton, 2.5 mM sodium pyrophosphate, 1 mM�-glycerolphosphate, 1 mM Na3PO4, 1 �M/mL leupeptin, and 1 mMphenyl methyl-sulfonyl fluoride). After brief sonication, the lysates wereclarified by centrifugation at 12 100g/10 minutes and protein content wasmeasured by the Bradford method (Bio-Rad). Protein (50 �g per lane) wasloaded into acrylamide gels, separated by 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), and blotted to polyvinyli-denefluoride (PVDF) transfer membrane. Primary antibodies to p38 andphospho-p38 (Thr180/Tyr182; Cell Signaling, Beverly, MA), hFasL (USBio-logical, Swampscott, MA), and HIV-1 p24gag (NIH-AIDS reagent; NIH,Rockville, MD) were used. Immunoblot was carried out using the specificprimary antibody and detected with secondary horseradish peroxidase(HRP)–conjugated anti–rabbit or anti–mouse IgG using Amersham (Arling-ton Heights, IL) enhanced chemiluminescence (ECL) system and wasvisualized by autoradiography.12

CD8 T-cell apoptosis assay

Monocytes were purified as described in cell culture techniques and purifiedCD8 T cells were isolated from PBMCs by negative immunoselection usingDynal magnetic beads (Dynal, Lake Success, NY)4,12 for depletion of CD4T cells. Monocytes were infected with HIV-1 NL4-3 wild-type (Wt)– orNL4-3 �Nef–deleted virus at a concentration of 100 TCID50 /106 cells permilliliter/well. Different experimental groups were set up as follows: groupI, as a mock (untreated); group II, incubated with latex beads only; groupIII, incubated with latex beads and treated with 20 �g/mL neutralizinganti-FasL mAb (NOK-2; PharMingen); and group IV, incubated with latexbeads and treated with 1 �M p38 inhibitor (RWJ67657). On day 7 afterinfection, 0.5 � 106 polystyrene latex beads (IDC Spheres; IDC, Portland,OR) was added to groups II, III, and IV for 3 hours at 37°C. Anti-FasL mAbwas added for 90 minutes, before 0.5 � 106 autologous purified CD8 T cellswere added as described.4 After 14 hours of incubation of monocytes plus Tcells, cells were harvested and stained with mAbs including FITC-conjugated anti-CD8 and further stained with annexin-V–PE. The analysis

was performed on gated low forward scatter and CD8 T cells and furtheranalyzed for the CD8/annexin-V–positive cells (CD8 population).

Results

HIV infection stimulates p38 phosphorylation

Activation of p38 kinase following HIV infection in some targetcells has been previously demonstrated.14,20 Activation was not aresult of virus binding to the cell surface but apparently occurs aftervirus internalization. We sought to gain insight into the molecularinteractions responsible for this activation; therefore, we examinedthe effect of HIV-1 infection on p38 phosphorylation, a majorindicator of p38 kinase activity. HIV-1–infected Jurkat T cells orPBMCs were harvested 24 hours after infection for protein analysisof p38 activation. Western blot analysis was carried out usingantibodies specific for either the Thr180 and Thr182 phosphory-lated forms of p38 or controls (Figure 1A). We observed that p38was rapidly phosphorylated following HIV infection in either thecell line or in primary PBMCs, confirming the activation of p38signaling during HIV infection.

HIV-1–induced apoptosis is blocked by p38 MAPK inhibitors

Studies support that active host cell signaling is involved in theinduction of apoptosis.21 A host cell–signaling pathway that mayhave relevance for bystander apoptosis is the p38 MAP kinasepathway. Genetic studies conducted in mice suggest that MAPkinase kinase 3 (MKK3), an upstream kinase of p38, is required foractivation-induced cell death of T cells.22 Therefore, we examinedthe role of the p38 pathway in HIV-1–induced apoptosis. Human

Figure 1. p38 MAPK activation is required for HIV-1–related apoptosis. (A) p38 MAPK is activated by HIV-1 infection. Immunoblot analysis of protein extracted fromPBMCs or Jurkat T cells infected with mock or NL4-3 virus. Total cellular protein extract (50 �g) was extracted 24 hours after infection and analyzed by 12% SDS-PAGE.p38 and phospho-p38 MAPK activation was detected by using specific antibodies that recognize phospho-p38 MAPK (top) only when phosphorylated or the p38 MAPK(middle). HIV infection strongly activates phosphorylation in both cell types. The same lysates were blotted with anti-p24gag (bottom) to monitor viral infection. (B-C)HIV-1–induced apoptosis is blocked by p38 MAPK inhibitor. Human PBMCs were infected with (B) NL4-3 Wt virus or (C) different HIV-clade specific primary viral isolates. Cellswere uninfected (mock) or infected with HIV-1 virus with or without addition of 1 �M p38 inhibitor-I or -II (SB203580 or RWJ67657, respectively). Cells were collected 2 daysafter infection and stained with annexin-V–FITC; p24gag-PE and HIV-induced apoptosis analysis was performed. The values indicated frequencies of positive cells of eachquadrangle. Data were representative of 3 (B) or 2 (C) independent experiments. Both p38 inhibitors block HIV-driven apoptosis. (D) p38 MAP kinase inhibitors blockHIV-induced caspase-3 activity. Cell lysates were prepared from the groups described in panel B. Total protein from each cell lysate (100 �g) was used for the colorimetriccaspase assay as described in “Materials and methods.” Each column represents the mean � standard deviation from triplicate samples derived from 1 of 3 independentexperiments. All these experiments gave similar results.

ROLE OF p38 MAPK IN Nef-INDUCED FasL INDUCTION 2061BLOOD, 15 SEPTEMBER 2005 � VOLUME 106, NUMBER 6

PBMCs were infected with NL4-3 virus at a concentration of 100TCID50 /106 cells per milliliter and tested for apoptosis undervarious conditions, including in the presence or absence of knownspecific p38 inhibitors designated as inhibitor-I or inhibitor-II(SB203580 or RWJ67657, respectively) at 1 �M as described.14

Cells were collected 2 days after infection from each treatmentgroup and analyzed for apoptosis induction in HIV-infected cells.HIV-1 infection promoted significant apoptosis in target cells(Figure 1B). Of importance, both p38 inhibitors could blockHIV-driven apoptosis. Both inhibitors inhibited apoptosis driven byvirus more than 70% at 1-�M concentration. However, p38inhibitor-II exhibited a greater antiapoptotic potency. To confirmthis observation on primary viral isolates, we analyzed the induc-tion of apoptosis by 4 different subtype-divergent HIV viralisolates (Figure 1C). All 4 primary subtype-divergent virusesinduced strong and clearly detectable levels of target cell apoptosisin this infection system. Again, both inhibitors were effective atinhibiting apoptosis. We wanted to ensure that the observedapoptosis was also activating caspases. Accordingly, we chose thecaspase-3 assay as a downstream marker for all caspase-inducedapoptosis,21 and caspase-3 activation was attenuated by pharmaco-logic inhibition of p38 with 70% potency with a 1-�M concentra-tion of inhibitor-II (Figure 1D). These results support an essentialrole for the p38 MAPK pathway in apoptosis induction by diverseisolates of HIV-1 in PBMCs.

HIV-1–mediated FasL up-regulation requires p38 signaling

Previous studies indicate that HIV-infected cells can up-regulateFasL expression and selectively induce the apoptosis of FasL-susceptible T cells from HIV-positive individuals.4 Recent reportshave identified that up-regulation of FasL on HIV-infected cells canbe induced by Nef.23 However, the signaling pathways that arerequired for this activity remain unclear. Based on the inhibition ofapoptosis by p38 inhibitors observed and previous studies implicat-ing Fas/FasL signaling in HIV-induced apoptosis, we decided toexamine the relevance of the p38 pathway to HIV-induced FasLexpression. We first sought to confirm the induction of FasL byHIV infection. Strong induction of FasL was observed in HIVinfection in PBMCs (Figure 2A). Further FasL expression induc-tion was also confirmed using FasL-specific antibody by immuno-blot analysis using the infected samples (Figure 2B). This resultconfirms a relationship between the FasL induction and HIVinfection. Next, PBMCs or Jurkat cells were infected with mock orNL4-3 virus in the presence or absence of p38 inhibitor-II(RWJ67657), and FasL expression was analyzed by FACS usingFasL-specific antibody. We observed strong inhibition of FasLexpression induced by the pharmacologic blockade of p38 signal-ing (Figure 2C-D). This inhibition appears independent of the celltype used in the experiment. Further, the up-regulation of FasL byindividual HIV genes suggests that Nef but not Tat, Vpr, Vpu, orEnv is sufficient to activate high-level FasL expression in Jurkatand the monocytic U937 cells (data not shown). This effect wasdosage dependent with saturating effects observed at 1 �M.

p38 MAPK is necessary for FasL expression by HIV-1 Nef

To investigate the requirement of p38 activity in Nef-induced FasLexpression, we first sought to demonstrate that Nef transfection issufficient to activate p38 and that this activation can be reversed bya p38 inhibitor. Nef induced a strong increase in phosphorylation ofp38 above basal levels (Figure 3A). In contrast, there was no

detectable increase in the phosphorylation of p38 in Nef-treatedcells in the presence of p38 inhibitor. These data indicate that Nefcan induce phosphorylation of the p38 MAP kinase and also thatNef is an activator of this pathway. Next, we studied the p38requirement for FasL transcription in a promoter-specific fashion.Jurkat cells were transfected with pNef in the presence ofcotransfected wild-type p38 (Figure 3B) or dominant-negative p38constructs (Figure 3C), and activation of a FasL promoter–drivenluciferase construct was assayed. The dominant-negative p38contains a K-to-M mutation at the ATP-binding site, and thismolecule has been previously reported to block p38 activity in vitroby protein interference.15 Nef strongly induced FasL promoteractivation in the presence of native p38 (Figure 3B). In contrast,induction of FasL transcription was reduced in cells cotransfectedwith this dominant-negative mutant p38 and the Nef construct(Figure 3C). To test directly if p38 is required to activateNef-mediated FasL expression, we sought to develop specificsiRNAs that target p38. We targeted the p38-� isoform, which isthe dominant isoform expressed in immune cells. These siRNAs

Figure 2. p38 MAPK activation is required for HIV-induced FasL expression. (A)HIV-1–mediated FasL induction in HIV-1–infected cells. Quantification of FasLexpression in human PBMCs infected with HIV-1 NL4-3 virus. Cell lysates orsupernatants were prepared at the indicated times after infections. As a control, mockcells were stimulated with phorbol myristate acetate (PMA, 10 ng/mL) plus ionomycin(500 ng/mL) to induce FasL expression as described in “Materials and methods.”Results are expressed in concentration to demonstrate that the assay detects bothforms of FasL. Data represent the mean � standard deviation from triplicate samplesderived from 1 of 3 independent experiments. (B) Immunoblot analysis of FasLexpression. Total cellular proteins were prepared from the groups described in panelA. Protein lysates were blotted with anti-FasL–specific antibody or anti-p24gag

(bottom) to confirm the FasL expression and monitor viral infection status. (C-D) FasLinduction is blocked by p38 inhibitors. PBMCs (C) or Jurkat T cells (D) were infectedwith HIV-1 NL4-3 virus in the presence or absence of 1 �M p38 inhibitor RWJ67657.At 48 hours after infection, FasL expression was measured by flow cytometry, andanalysis was performed on a gated low forward scatter and CD24(HSA)–positivecells using FasL-specific antibody. It is noteworthy that the p38 inhibitor stronglyblocks FasL induction in both cell types tested. Data shown are representative of 1 of3 independent experiments. Histograms show the indicated surface marker, and filledhistograms represent the isotype-matched control antibodies.

2062 MUTHUMANI et al BLOOD, 15 SEPTEMBER 2005 � VOLUME 106, NUMBER 6

were studied for their ability to suppress the expression of p38.Four different sets of p38 siRNA oligos were examined. Jurkat cellsthat were transfected with siRNA vector clone p38-61 demon-strated a dramatic reduction of p38 expression, while the otherclones, p38-352, p38-775, and p38-1016, showed only moderatedecreases in p38 expression (Figure 3D). Next, Jurkat cells werecotransfected with pNef and siRNA (clone p38-61) construct, andactivation of FasL expression was assayed in equal numbers ofcells by flow cytometry. Nef-mediated FasL expression wassignificantly reduced by clone p38-61 (Figure 3E), and the effect ofp38 siRNA appears independent of the expression level of Nef(Figure 3F). Therefore, these loss-of-function experiments (chemi-cal inhibitor, dominant negative, and siRNA) suggest that p38MAPK is necessary for FasL induction by HIV-1 Nef.

HIV-1 Nef-deleted virus fails to induce FasL on human PBMCs

Next, we wanted to ascertain if p38 is also necessary for FasLinduction in HIV-1 infection. Accordingly, we constructed a virusthat was devoid of Nef expression to ascertain if there was an effecton FasL in the absence of Nef. Nef-deleted virus was constructedby introducing a frameshift mutation in the Nef ORF, thus

eliminating Nef from this virus, and confirmed via proviraltransfection into 293T cells via ELISA and Western blotting(Figure 4A-B). To analyze the effect of viral delivery of the nefgene on FasL expression, we used a pseudoviral infection assay forinfecting PBMCs with the NL4-3 Wt or NL4-3 � Nef–deleted virusin the presence or absence of p38 inhibitor at 1 �M. Flowcytometry analysis of intracytoplasmic p24gag and FasL levels wasdetermined 4 days after infection, allowing us to follow bothmarkers in infected cells (Figure 4C). A direct correlation wasobserved between the presence of the Nef gene and the ability of thevirus to induce FasL expression. The percentage of positive cells isdepicted in the upper-right quadrant for each group. Approximatelyhalf of the infected (p24gag positive) cells were FasL positive onday 4. The FasL-positive cells constitute a unique separatepopulation of cells that was observed only in the Nef-containingvirus infection group. Of interest, viral Nef–mediated FasL expres-sion was suppressed by treatment of the culture with 1 �M p38inhibitor. These results in conjunction with Figure 3 suggest thatNef is critical for HIV-1–mediated FasL up-regulation. Further-more, Nef-deleted viruses exhibit strong attenuation of p38 activa-tion (Figure 4D). A modest activation of p38 was also observed in

Figure 3. Nef-induced FasL induction requires p38 MAPK activation. (A) Nef induced phosphorylation of p38. Western blot analysis of protein extracted from Jurkat cellstransfected with 5 �g vector control or pNef in the presence or absence of p38 inhibitor-II. Samples were prepared as described in “Materials and methods” and resolved on12% SDS-PAGE gels. Gels were transferred to PVDF membranes and immunoblotted with p38MAPK- and phospho-p38MAPK–specific antibodies as indicated. Nef inducedphosphorylation of p38, and this activation can be blocked by the p38 inhibitors (p38 inhi). (B-C) Inhibition of Nef-induced FasL promoter activity by a dominant-negative p38.Jurkat cells were transfected with plasmid containing pVector, the hFasL-Luc promoter, the pNef plus hFasL-Luc promoter, or the pNef plus hFasL-Luc promoter together withwild-type p38 (B) or p38-DN (C) plasmid constructs as indicated. Luciferase activity in whole-cell lysates was assayed after 12 to 18 hours and is shown as the mean value �SEM. Transfection efficiency was monitored by cotransfection of a plasmid encoding �-gal and results were normalized to �-gal levels. Similar results were obtained in 3independent experiments and were reproducible. Nef strongly induced the FasL promoter activation (B). In contrast, induction of FasL transcriptions was reduced in cellscotransfected with dominant-negative mutant p38 (C). (D) Construction and expression of p38siRNAs. Four different target siRNA expression vectors were constructed asdescribed in “Materials and methods.” Jurkat T cells were transiently transfected with 5 �g siRNA control vector or p38siRNA expression constructs as indicated clones. Celllysates were extracted 48 hours after transfection, and immunoblotting was performed by using an anti-p38 antibody and an actin antibody as a control for equal loading. p38expression was suppressed significantly in cells transfected with clone p38-61 and moderately in cells transfected with the other clones compared with control vectortransfected. (E) FasL expression is significantly reduced by p38siRNA. Jurkat T cells were electroporated with 5 �g plasmid-expressing vector, pNef, or p38siRNA (clonep38-61) and pNef plasmids as indicated. At 48 hours after transfection, the surface levels of FasL expression were determined by flow cytometry using a FasL-specific antibody.Filled histograms show the FasL expression, and open histograms represent isotype-matched control antibodies. p38 siRNA inhibited Nef-induced FasL induction. Similarresults were obtained in 3 independent experiments. Transfection efficiency was monitored by cotransfection of a pCMV plasmid encoding GFP, which also served as a markerfor gating on transfected cells. (F) p38 siRNA does not affect Nef expression. Total protein extracts were prepared from the groups transfected with vector, pNef, or p38siRNAplus pNef. Of each protein sample, 50 �g was separated on 12% SDS-PAGE and analyzed by Western blot with a polyclonal antibody against Nef antibody. Immunoblottingwas also performed using an antiactin antibody as an internal control.

ROLE OF p38 MAPK IN Nef-INDUCED FasL INDUCTION 2063BLOOD, 15 SEPTEMBER 2005 � VOLUME 106, NUMBER 6

Nef-deleted viral infections, which could be due to influence fromother HIV antigens, including the envelope antigen.24,25 Theseresults extend the association between Nef, FasL induction, and thep38 pathway within an infection setting.

p38-mediated AP-1 activation is required for Nef-inducedFasL transcription

AP-1 is an important transcription factor in immune activation andis induced by many stimuli, including growth factors, cytokines,T-cell activators, neurotransmitters, and ultraviolet (UV) irradia-tion.16,26,27 Eukaryotic cells respond to external stresses andinflammatory factors through the activation of MAPKs, leading toaltered transcriptional activity. Specifically, c-Jun N-terminal ki-nases (JNKs) and p38 MAPK have been implicated in theseresponses. In most cases, p38 is activated by MKKs 3 and 6through their phosphorylation.28,29 Of interest, MAPK-activatedtranscription has been implicated in driving the transcriptionalactivation of FasL via AP-1.16,30 Accordingly, we hypothesizedthat Nef may exploit this host pathway for inducing FasL duringHIV infection.

As discussed in the section entitled “p38 MAPK is necessary forFasL expression by HIV-1 Nef,” Nef-induced p38 activation isnecessary for the up-regulation of FasL induction in both aninfection and a transfection setting. However, the requirement ofthe AP-1 binding enhancer element for Nef-induced FasL transcrip-tional activation was unclear, particularly since multiple elementshave been proposed to drive FasL induction, including nuclearfactor of activated T cell (NF-AT), nuclear factor B (NF-B),AP-1, and c-Myc.16,28-30

Accordingly, transfection of Jurkat cells with an AP-1 promoterreporter plasmid in conjunction with Nef induced strong AP-1promoter activation in T cells and was sensitive to p38 inhibitors at1-�M concentration. This suggests that Nef requires p38 to activatethe transcription induced by the AP-1 enhancer elements (Fig-ure 5A). To test whether these observations were related, we

mutated the AP-1 enhancer element from the hFasL promoterplasmid by site-directed mutagenesis.16 Cotransfection with Neffailed to activate the AP-1mut promoter, indicating that the activa-tion of AP-1 transcriptional factors by Nef is required for FasLtranscription (Figure 5B). However, transfection of Jurkat cellswith the hFasL AP-1mut expression vector in conjunction withpRelA (a subunit of NF-B) did induce FasL transcription (Figure5C), indicating that mutation of AP-1 sites does not affect theability of other transcriptional factors to induce FasL transcription.Transfection of the AP-1 reporter plasmid into Jurkat cells andsubsequent infection with NL4-3 Wt or NL4-3 �Nef virus suggestthat Nef is necessary for optimal AP-1–dependent transcriptioninduced by HIV-1 (Figure 5D). Collectively, these data show thatNef-induced FasL expression requires p38, which drives the FasLpromoter through AP-1 induction.

The PxxP domain of Nef is required for FasL transcriptionand p38 activation

Previous experiments suggest that Nef possesses various protein-binding domains responsible for its activity.31 Of particular interestis the PxxP motif in Nef, which is a binding site for Src homology 3(SH3)–mediated protein-protein interactions. We next examined ifthe PxxP region of Nef is required for phosphorylation of p38 inT-cell or in monocyte cell lines by mutating prolines to alanines,which retain Nef structure but block PxxP-associated activ-ity.10,31-34 The PxxP mutation did not stimulate phosphorylation ofp21-related kinase family (Pak) in T cells or hematopoietic cellkinase (Hck) phosphorylation in monocytic cell lines (data notshown). Next, Jurkat T cells and monocytic U937 cell lines weretransfected with expression vectors for pNef or pNef (PxxP). The Nefcontaining the PxxP deletion did not stimulate phosphorylation ofp38 in either T cells or monocytic cells (Figure 6A). These datasuggest that Nef activation of p38 requires the PxxP domain of Nef.

We next examined FasL induction in T cells or monocytes

Figure 4. Nef is required for HIV-1–mediated FasL induction and efficient p38 activation. (A) Expression and analysis of the proviral expression constructs. A frameshiftmutation (5� Nef) was introduced in the nef gene in the proviral constructs as described in “Materials and methods.” Cell lysates derived from 293T cells transfected with pNef,pNL4-3 Wt, or pNL4-3 �Nef HIV-1 proviral constructs were separated by 12% SDS-PAGE and then transferred to nitrocellulose membranes. Samples were probed with anHIV-1 Nef-specific antiserum. Nef was not expressed in the pNL4-3 �Nef constructs. WB indicates Western blot. (B) HIV-1 infection was monitored by measuring p24 levels inculture supernatants derived from PBMCs infected with NL4-3 Wt or NL4-3 �Nef virus. Data were obtained 96 hours after infection. These experiments were repeated 3 timesand similar results were obtained. Error bars represent the mean � standard deviation from triplicate samples derived from 1 of 3 independent experiments. (C) Induction ofFasL expression by NL4-3 Wt or NL4-3 �Nef viruses in human PBMCs. Cells (1 � 106) were infected with NL4-3 Wt or NL4-3 �Nef virions and then treated with or without 1 �Mp38 inhibitor RWJ67657 as indicated. Two days after infection, an equal number of cells (analysis was performed on a gated low forward scatter and CD24[HAS]–positive cells)were assayed for intracellular p24gag and surface FasL expression by flow cytometry as described in “Materials and methods.” The values indicated frequencies of positive cellsof each quadrangle. Data were representative of 3 independent experiments and observations of similar suppression of FasL were obtained. (D) Nef-deleted viruses failed toinduce phosphorylation of p38. Western blot analysis of protein extracted from Jurkat cells infected with mock, NL4-3 Wt, or NL4-3 �Nef virus. Samples were prepared 24 hoursafter infection as described in “Materials and methods” and immunoblotted with p38MAPK and phospho-p38MAPK–specific antibodies as indicated.

2064 MUTHUMANI et al BLOOD, 15 SEPTEMBER 2005 � VOLUME 106, NUMBER 6

transfected with pNef in the presence or absence of p38 inhibitor orPxxP-mutated Nef. While Nef strongly induced FasL on both Tcells as well as monocytes (Figure 6B), mutation of the PxxPdomain severely diminished FasL induction, indicating that thePxxP domain of Nef plays an essential role in Nef-driven FasLexpression. These results were verified by FasL promoter reporterassays (Figure 6C-D). Collectively, these results illustrate that Nefactivates FasL through its PxxP domain in monocytes and T cells.Furthermore, this activation appears to be upstream of Pakactivation in T cells and of Hck activation in monocytes (data notshown). This activation in either cell phenotype eventually con-verges at p38, resulting in its activation, which is required forNef-mediated AP-1 induction of FasL expression.

Inhibition of p38 is sufficient to prevent HIV-1–inducedbystander killing of CD8 T cells

FasL is up-regulated in HIV-infected macrophages and T cells.Such FasL-positive cells have been hypothesized to be able toinduce Fas-mediated apoptosis of CD8 effector T cells.3-5 Accumu-lating evidence indicates that antigen-presenting cells such asmacrophages play a key role in the elimination of activated effectorT cells.3,4,35 Accordingly, we next investigated the role of p38activation of HIV-infected macrophages to induce bystander apo-ptosis of CD8 T cells.

CD14 and CD8 cell populations were isolated from HIV-1–negative PBMCs (Figure 7Ai) by cell-negative immune selection.

Figure 5. AP-1 is required for Nef-induced FasL promoter activation through p38. (A) Jurkat T cells were transfected with plasmid containing either the pVector or theAP1-Luc promoter plasmid, and the pNef plus AP1-Luc promoter plasmid or the pNef plus AP1-Luc promoter plasmid and treated with p38 inhibitors (1 �M) as indicated in“Materials and methods.” (B) Jurkat T cells were transfected with the hFasL-Luc promoter plasmid, hFasL-Luc promoter plasmid plus pNef, or pNef plus hFasLmut-Luc (mutationin the AP-1 binding site) promoter plasmid as indicated. (C) As a control, Jurkat cells were transfected with hFasLmut-Luc or hFasLmut-Luc plus an expression construct for pRelA(p65) as indicated. Luciferase activity in whole-cell lysates was assayed after 12 to 18 hours and is shown as the mean value � SEM. Transfection efficiency was monitored bycotransfection of a plasmid encoding �-gal, and results were normalized to �-gal levels. Similar results were obtained in 3 independent experiments. (D) Nef-deleted virus failedto induce AP-1 transcription. Jurkat cells were transiently transfected with pAP1-Luc vector as indicated. At 48 hours after transfection, cells were infected with NL4-3 Wt with orwithout p38 inhibitor or NL4-3 �Nef viruses as indicated. Luciferase activity in whole-cell lysates was monitored after 24 hours and is shown as the mean value � SEM.Transfection efficiency was monitored by cotransfection of a plasmid encoding �-gal. These data are representative of 2 independent experiments. The asterisks indicatesignificant differences between cells infected with Wt virus (P .02).

Figure 6. PxxP domain of Nef is required for the FasL induction. (A) Loss of the PxxP domain of Nef significantly diminishes the phosphorylation of p38. Western blotanalysis of extracts derived from Jurkat or U937 cells transfected with either the expression vector for wild-type Nef or the amino acid–mutated Nef. Cell extracts were prepared48 hours after transfection as described in “Materials and methods” and subjected to 12% SDS-PAGE followed by PVDF membrane transfer and analyzed by Western blottingusing specific p38 and phospho-p38 antibodies as indicated. Note, the amino acid–mutated Nef (pNef(PxxP)) failed to induce phosphorylation of p38. (B) Comparison of FasLinduction by wild-type Nef versus amino acid–mutated Nef. Jurkat T cells or U937 cells were electroporated with 5 �g wild-type Nef, amino acid–mutated Nef, or wild-type Nefplus p38 inhibitor (1 �M). At 48 hours after transfection, the surface levels of FasL were determined by flow cytometry by staining with a FasL-specific antibody. Transfectionefficiency was monitored by cotransfection of a pCMV plasmid encoding GFP. Thick line histograms show the indicated surface markers, and filled histograms represent theisotype-matched control antibodies. Similar results were obtained in 3 independent experiments. (C-D) The PxxP domain of Nef (amino-acid mutated) is essential for inductionof FasL promoter activity in T cell or monocytic cells. Jurkat T cells or U937 cells were transfected with 5 �g hFasL-Luc promoter plasmid, wild-type Nef plus hFasL-Lucpromoter plasmid, or amino acid–mutated Nef plus hFasL-Luc promoter plasmid as indicated. Luciferase activity in whole-cell lysates was assayed after 12 to 18 hours and isshown as the mean value � SEM. Transfection efficiency was monitored by cotransfection of a plasmid encoding �-gal and results were normalized to �-gal levels. Similarresults were obtained in 3 independent experiments.

ROLE OF p38 MAPK IN Nef-INDUCED FasL INDUCTION 2065BLOOD, 15 SEPTEMBER 2005 � VOLUME 106, NUMBER 6

We then examined CD8 T-cell apoptosis after 12 hours ofincubation with CD14 macrophages that were infected with theNL4-3 pseudovirus, which encoded VSV-G envelope, which hadbeen activated through treatment with latex beads to stimulatephagocytosis. Annexin-V staining on the gated CD8 cells showsthat 22.1% of the cells were induced to undergo apoptosis (Figure7Aii). Uninfected macrophages, even when activated by polysty-rene beads, induced a negligible level of apoptosis in the CD8T-cell population under any of the conditions tested (Figure 7B, toppanels). Also, the �Nef virus did not induce bystander killing underthese conditions (Figure 7B, middle panels). In contrast, HIV-infected macrophages stimulated with latex beads drove apoptosisin the CD8 T cells (Figure 7B, bottom panels). The addition of aneutralizing anti-FasL mAb to the cell culture completely blockedthe induction of apoptosis, suggesting its required role. Of interest,p38 blockade was almost as effective as anti-FasL antibody atpreventing bystander apoptosis, but there was clearly a residuallevel of apoptosis that was resistant to the inhibitor. This isconsistent with prior published reports suggesting that other HIVantigens through p38-independent mechanisms can play a role inbystander apoptosis. These findings clearly illustrate that HIV-infected macrophages can induce Fas/FasL-mediated apoptosis ofCD8 T cells during HIV infection, and inhibition of p38 is sufficient toprevent much of the resulting bystander apoptosis.

Discussion

Hallmarks of HIV-1 infection include the destruction of T cellsand the suppression of cellular immune responses in vivo.3-6,35,36

A mechanism that has been previously reported is through thebystander killing of CD8 effector T cells, the cell populationdirectly responsible for immune clearance and controlling viralload.3,5,35,36 Indirect cell killing has been proposed to involve theup-regulation of FasL-inducing apoptosis of effector cytotoxic Tlymphocytes (CTLs) as they approach viral-harboring CD4 Tcells and macrophages. Furthermore, lamina propria of SIV-infected rhesus macaques has been shown to harbor massiveapoptosis of viral-specific memory T cells through the activationof the Fas-FasL pathway.37 Hence this strategy results in asignificant advantage for HIV in evasion of immune recognitionand viral clearance.

A recent report identified that Nef binds to the CD3� chain ofthe TCR complex, and this interaction is important for theinduction of FasL expression.38 The effect required the PxxPdomain, which consequently stimulated the transcriptional activa-tion of FasL.38 Additional evidence suggests that Nef coprecipitateswith the Nef-associated kinase (NAK), a member of the Pak39 inT cells. NAK is activated via the small guanosine triphosphatases

Figure 7. HIV-specific CD8� T-cell apoptosis is blocked by a p38 inhibitor. (Ai) CD14 macrophages and CD8 T cells were isolated from HIV-1–negative PBMC donors asdescribed in “Materials and methods.” (Aii) Macrophages were infected with HIV-1 wild-type pseudovirus (NL4-3 virus that complemented with VSV-G envelope)(100 TCID50 /1 � 106 cells per milliliter) and incubated with latex beads before being mixed with autologous CD8 T cells. Fourteen hours after CD8 T cells were added toinfected macrophages, cells were harvested and stained with CD8-FITC/annexin-V–PE. Note that the CD8 T cells were induced to undergo apoptosis. (B) Autologousmacrophages were uninfected or infected with pseudovirus HIV-1 NL4-3 �Nef virus or infected with NL4-3 Wt virus (viruses that complemented with VSV-G envelope) atconcentration of 100 TCID50 /1 � 106 cells/mL, were stimulated with latex beads, and were incubated with purified mock CD8� T cells in the presence or absence of neutralizinganti-FasL antibody or with 1 �M p38 inhibitor overnight. Cells were then harvested and stained for CD8 plus annexin-V. Apoptosis analysis was performed by gating the lowforward and CD8 scatter of cells. The percentage values represent the gated CD8 annexin V–positive cells. A representative experiment of 3 performed is shown.

2066 MUTHUMANI et al BLOOD, 15 SEPTEMBER 2005 � VOLUME 106, NUMBER 6

(GTPases) CDC42 and Rac1 through Vav, and moreover, Pak1 andPak2 are implicated in this activation.32-34 Although it is believedthat activation of FasL transcription functions through the TCR-CD3� complex mediated by the CD3�� chain,32,38,40 the down-stream signals and their transcriptional regulation required for thisactivation have not been determined.

The results presented here indicate that p38 MAPK activation isnecessary for T-cell and macrophage activation and eventual FasLtranscription in various viral subtypes. Previous work suggests thatmultiple factors may be sufficient to induce FasL expression.16,26,27

Accordingly, we suggest that the AP-1 enhancer is required for Nefto induce FasL transcription. Further, p38 is also required for AP-1activation, suggesting a linear pathway that involves p38 activationand its subsequent AP-1 activation. Recently, Biggs et al41 identi-fied that AP-1 can be induced by Nef via extracellular signal-related kinase 1/2 (ERK1/2) MAPK signaling events. Our resultssuggest that p38-like ERK1/2 is required for AP-1 activation, andthat this leads to a transcriptional up-regulation of FasL. Ourevidence suggests that p38 MAPK is important for FasL-mediatedkilling and specifically targets AP-1 transcriptional factors for thiseffect. Thus our data extend these recent observations and link theAP-1 pathway with p38 activation and bystander killing.

Previous studies indicate that HIV infection of macrophagescan induce FasL expression, which drives bystander killing ofCD4� T cells.4,11 Additionally, macrophages also drive FasLup-regulation to induce bystander killing of HIV-specific CD8 Tcells.35 The destruction of the CD8 T cells by the FasL pathway islikely a significant damper for the cell-mediated immune response,which ultimately could limit immune clearance.4,6,7 We have alsoobserved blockade of apoptosis using primary viral isolates.

Our results suggest that Nef is critical for efficient p38activation and AP-1–driven transcription induced by HIV-1 infec-tion. However, residual p38 phosphorylation and AP-1–mediatedtranscription can be observed in Nef-deleted viruses (Figure 5D),suggesting that other factors such as Env or other HIV accessorygenes may also be capable of activating p38 and AP-1–driventranscription.24,25 On the other hand, Env is neither required norsufficient to up-regulate FasL (data not shown), suggesting thateither the inputs from Env are not sufficiently robust to drive FasLtranscription or Nef is sufficient to activate another undeterminedpathway that is concomitantly required with p38 to drive FasLtranscription (data not shown), and overexpression fails to augmentNef-induced FasL transcription (Figure 3B).

In conclusion, our work identifies a linear signaling pathwaythat is necessary for HIV-1 Nef-induced bystander killing. Weshow that Nef activates and requires p38 MAPK activation and itsAP-1 enhancer target within the FasL promoter to stimulate itstranscription. Through loss-of-function experiments, we were ableto delineate which factors are activated versus required forregulating FasL transcription. We suggest that this finding has importantapplications for the development of novel HIV therapeutics, as anti-p38compounds could target a vital pathogenic function of HIV-1.

Acknowledgments

We are very grateful to Dr Craig B. Thompson, University ofPennsylvania, and Dr Mario Stevenson, University of Massachu-setts Medical Center, for their continuous support, suggestions, andprovision of reagents.

References

1. Ho DD, Neumann AU, Perelson AS, Chen W,Leonard JM, Markowitz M. Rapid turnover ofplasma virions and CD4 lymphocytes in HIV-1infection. Nature. 1995;373:123-126.

2. Letvin NL. Progress in the development of anHIV-1 vaccine. Science. 1998;280:1875-1880.

3. McCune JM. The dynamics of CD4� T-cell deple-tion in HIV disease. Nature. 2001;410:974-979.

4. Mueller YM, De Rosa SC, Hutton JA, et al. In-creased CD95/Fas-induced apoptosis of HIV-specific CD8(�) T cells. Immunity. 2001;15:871-882.

5. Finkel TH, Tudor-Williams G, Banda NK, et al.Apoptosis occurs predominantly in bystandercells and not in productively infected cells of HIV-and SIV-infected lymph nodes. Nat Med. 1995;1:129-134.

6. Gandhi RT, Chen BK, Straus SE, Dale JK, Le-nardo MJ, Baltimore D. HIV-1 directly kills CD4�T cells by a Fas-independent mechanism. J ExpMed. 1998;187:1113-1122.

7. Ogg GS, Jin X, Bonhoeffer S. Quantitation of HIV-1-specific cytotoxic T lymphocytes and plasmaload of viral RNA. Science. 1998;279:2103-2106.

8. Xu XN, Screaton GR, McMichael AJ. Virus infec-tions: escape, resistance, and counterattack.Immunity. 2001;15:867-870.

9. Fackler OT, Luo W, Geyer M, Alberts AS, PeterlinBM. Activation of Vav by Nef induces cytoskeletalrearrangements and downstream effector func-tions. Mol Cell. 1999;3:729-739.

10. Wolf D, Witte V, Laffert B, et al. HIV-1 Nef associatedPAK and PI3-kinases stimulate Akt-independent Bad-phosphorylation to induce anti-apoptotic signals. NatMed. 2001;7:1217-1224.

11. Geleziunas R, Xu W, Takeda K, Ichijo H, GreeneWC. HIV-1 Nef inhibits ASK1-dependent death

signaling providing a potential mechanism forprotecting the infected host cell. Nature. 2001;410:834-838.

12. Muthumani K, Hwang DS, Choo AY, et al. HIV-1Vpr inhibits the maturation and activation ofmacrophages and dendritic cells in vitro. Int Im-munol. 2005;17:103-116.

13. Wadsworth SA, Cavender DE, Beers SA. RWJ67657, a potent, orally active inhibitor of p38 mi-togen-activated protein kinase. Pharmacol ExpTher. 1999;291:680-687.

14. Muthumani K, Wadsworth SA, Dayes NS, et al.Suppression of HIV-1 viral replication and cellularpathogenesis by a novel p38/JNK kinase inhibi-tor. AIDS. 2004;18:1-10.

15. Eyers PA, van den Ijssel P, Quinlan RA, GoedertM, Cohen P. Use of a drug-resistant mutant ofstress-activated protein kinase 2a/p38 to validatethe in vivo specificity of SB 203580. FEBS Lett.1999;451:191-196.

16. Kasibhatla S, Brunner T, Genestier L, Echeverri F,Mahboubi A, Green DR. DNA damaging agentsinduce expression of Fas ligand and subsequentapoptosis in T lymphocytes via the activation ofNF-kappa B and AP-1. Mol Cell. 1998;1:543-551.

17. Hannon GJ, Conklin DS. RNA interference byshort hairpin RNAs expressed in vertebrate cells.Methods Mol Biol. 2004;257:255-266.

18. He J, Choe S, Walker R, Di Marzio P, MorganDO, Landau NR. Human immunodeficiency virustype 1 viral protein R (Vpr) arrests cells in the G2phase of the cell cycle by inhibiting p34cdc2 ac-tivity. J Virol. 1995;69:6705-6711.

19. Gibbs JS, Regier DA, Desrosiers RC. Construc-tion and in vitro properties of HIV-1 mutants withdeletions in “nonessential” genes. AIDS Res HumRetroviruses. 1994;10:343-350.

20. Shapiro L, Heidenreich KA, Meintzer MK, Din-arello CA. Role of p38 mitogen-activated proteinkinase in HIV type 1 production in vitro. Proc NatlAcad Sci U S A. 1998;95:7422-7426.

21. Green DR, Beere HM. Apoptosis: mostly dead.Nature. 2001;412:133-135.

22. Tanaka N, Kamanaka M, Enslen H, et al. Differ-ential involvement of p38 mitogen-activated pro-tein kinase kinases MKK3 and MKK6 in T-cell ap-optosis. EMBO Rep. 2002;3:785-791.

23. Ameisen JC. Apoptosis subversion: HIV-Nef pro-vides both armor and sword. Nat Med. 2001;7:1181-1182.

24. Perfettini JL, Roumier T, Castedo M, et al. NF-kappaB and p53 are the dominant apoptosis-inducing transcription factors elicited by HIV-1envelope. J Exp Med. 2004;199:629-640.

25. Perfettini JL, Castedo M, Nardacci R, et al. Es-sential role of p53 phosphorylation by p38 MAPKin apoptosis induction by the HIV-1 envelope.J Exp Med. 2005;201:279-289.

26. Le-Niculescu H, Bonfoco E, Kasuya Y, Claret FX,Green DR, Karin M. Withdrawal of survival factorsresults in activation of the JNK pathway in neuro-nal cells leading to Fas ligand induction and celldeath. Mol Cell Biol. 1999;19:751-763.

27. Brunner TS, Kasibhatla M, Pinkoski J, et al. Ex-pression of Fas ligand in activated T cells is regu-lated by c-Myc. J Biol Chem. 2000;27:9767-9772.

28. Chang L, Karin M. Mammalian MAP kinase sig-naling cascades. Nature. 2001;410:37-40.

29. Dong CR, Davis J, Flavell RA. MAP kinases inthe immune response. Annu Rev Immunol. 2002;20:55-72.

30. Faris M, Kokot N, Latinis K, et al. The c-Jun N-terminal kinase cascade plays a role in stress-induced apoptosis in Jurkat cells by up-regulating

ROLE OF p38 MAPK IN Nef-INDUCED FasL INDUCTION 2067BLOOD, 15 SEPTEMBER 2005 � VOLUME 106, NUMBER 6

Fas ligand expression. J Immunol. 1998;160:134-144.

31. Geyer M, Fackler OT, Peterlin BM. Structure-function relationsips in HIV-1 Nef. EMBO Rep.2001;2:580-585.

32. Saksela K, Cheng G, Baltimore D. Proline-rich(PxxP) motifs in HIV-1 Nef bind to SH3 domainsof a subset of Src kinases and are required forthe enhanced growth of Nef� viruses but not fordown-regulation of CD4. EMBO J. 1995;14:484-491.

33. Fackler OT, Lu X, Frost JA, et al. p21-activatedkinase 1 plays a critical role in cellular activationby Nef. Mol Cell Biol. 2000;20:2619-2627.

34. Renkema GH, Manninen A, Mann DA, Harris H,Saksela K. Identification of the Nef-associated

kinase as p21-activated kinase 2. Curr Biol. 1999;9:1407-1410.

35. Swingler S, Brichacek B, Jacque JM, Ulich C,Zhou J, Stevenson M. HIV-1 Nef intersects themacrophage CD40L signalling pathway to pro-mote resting-cell infection. Nature. 2003;424:213-219.

36. Barouch DH, Kunstman J, Kuroda MJ, et al.Eventual AIDS vaccine failure in a rhesus mon-key by viral escape from cytotoxic T lymphocytes.Nature. 2002;415:335-339.

37. Li Q, Duan L, Estes JD, et al. Peak SIV replica-tion in resting memory CD4� T cells depletes gutlamina propria CD4� T cells. Nature. 2005;434:1148-1152.

38. Xu XN, Laffert B, Screaton GR, et al. Induction ofFas ligand expression by HIV involves the inter-

action of Nef with the T cell receptor zeta chain.J Exp Med. 1999;189:1489-1496.

39. Sawai ET, Khan IH, Montbriand PM, Peterlin BM,Cheng-Mayer C, Luciw PA. Activation of PAK byHIV and SIV Nef: importance for AIDS in rhesusmacaques. Curr Biol. 1996;6:1519-1527.

40. Latinis KM, Carr LL, Peterson EJ, Norian LA,Eliason SL, Koretzky GA. Regulation of CD95(Fas) ligand expression by TCR-mediated signal-ing events. J Immunol. 1997;158:4602-4611.

41. Biggs TE, Cooke SJ, Barton CH, Harris MP, Sak-sela K, Mann DA. Induction of activator protein 1(AP-1) in macrophages by human immunodefi-ciency virus type-1 NEF is a cell-type-specific re-sponse that requires both hck and MAPK signal-ing events. J Mol Biol. 1999;290:21-35.

2068 MUTHUMANI et al BLOOD, 15 SEPTEMBER 2005 � VOLUME 106, NUMBER 6