Et

CMJBa

b

c

d

e

f

g

a

ARRAA

KAHT

1

tacimeHt

Cf

0d

Vaccine 29 (2011) 5711– 5724

Contents lists available at ScienceDirect

Vaccine

jou rn al h om epa ge: www.elsev ier .com/ locate /vacc ine

x vivo production of autologous whole inactivated HIV-1 for clinical use inherapeutic vaccines

ristina Gila,∗, Núria Climenta, Felipe Garcíaa,b, Carmen Hurtadoa, Sara Nieto-Márqueza, Agathe Leóna, Teresa Garcíaa, Cristina Roviraa, Laia Mirallesa, Judith Dalmaue, Tomás Pumarolac, Manel Almelac,

avier Martinez-Picadoe,f, Jeffrey D. Lifsong, Laura Zamoraa, José M. Miróa,b, Christian Brandere,f,onaventura Clotete, Teresa Gallarta,d, José M. Gatell a,b

AIDS Research Group-IDIBAPS-HIVACAT, SpainServices of Infectious Diseases & AIDS Unit, SpainMicrobiology of the Hospital Clinic Barcelona University of Barcelona, SpainImmunology of the Hospital Clinic Barcelona University of Barcelona, SpainInstitut de Recerca de la SIDA IrsiCaixa-HIVACAT, Hospital Germans Trias i Pujol, Badalona, Barcelona, SpainInstitució Catalana de Recerca i Estudis Avanc ats (ICREA), Barcelona, SpainAIDS and Cancer Virus Program, SAIC-Frederick, Inc., NCI-Frederick, Frederick, MD 21702, USA

r t i c l e i n f o

rticle history:eceived 13 December 2010eceived in revised form 29 April 2011ccepted 31 May 2011vailable online 14 June 2011

eywords:utologous HIV-1 immunogeneat inactivated HIV-1

a b s t r a c t

This study provides a detailed description and characterization of the preparation of individualized lotsof autologous heat inactivated HIV-1 virions used as immunogen in a clinical trial designed to testan autologous dendritic-cell-based therapeutic HIV-1 vaccine (Clinical Trial DCV-2, NCT00402142). Foreach participant, ex vivo isolation and expansion of primary virus were performed by co-culturing CD4-enriched PBMCs from the HIV-1-infected patient with PBMC from HIV-seronegative unrelated healthyvolunteer donors. The viral supernatants were heat-inactivated and concentrated to obtain 1 mL of autol-ogous immunogen, which was used to load autologous dendritic cells of each patient. High sequencehomology was found between the inactivated virus immunogen and the HIV-1 circulating in plasma at

9

herapeutic vaccine the time of HIV-1 isolation. Immunogens contained up to 10 HIV-1 RNA copies/mL showed considerably

reduced infectivity after heat inactivation (median of 5.6 log10), and were free of specified adventitiousagents. The production of individualized lots of immunogen based on autologous inactivated HIV-1 virusfulfilling clinical-grade good manufacturing practice proved to be feasible, consistent with predeter-mined specifications, and safe for use in a clinical trial designed to test autologous dendritic cell-basedtherapeutic HIV-1 vaccine.

. Introduction

Combined antiretroviral therapy (cART) inhibits HIV-1 replica-ion to levels below the limit of detection of standard clinical assays,llowing the restoration of normal or nearly normal CD4 T cellounts and protective T cell immunity to opportunistic pathogensn most patients, and has dramatically reduced the morbidity and

ortality of HIV-1 infection. However, in spite of these clinical ben-

fits, cART is incapable of eradicating HIV-1, and also fails to restoreIV-specific T cell responses able to durably and effectively to con-

rol HIV-1 replication when antiretroviral drugs are discontinued

∗ Corresponding author at: Service of Infectious Diseases & AIDS Unit, Hospitallínic Barcelona, Villarroel, 170, 08036 Barcelona, Spain. Tel.: +34 93 2275400x2322;

ax: +34 93 4514438.E-mail address: cgil@clinic.ub.es (C. Gil).

264-410X/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.oi:10.1016/j.vaccine.2011.05.096

© 2011 Elsevier Ltd. All rights reserved.

[1–5]. Thus, cART must be administered for life, increasing the riskof drug-related adverse effects, the potential for emergence of drugresistant mutant viral variants, along with some inconvenience forthe patients and economic burden for the patient and society [6,7].

Over the past several years, a variety of immune-based strate-gies including therapeutic vaccination and designed treatmentinterruptions have been explored as approaches that might alloweffective host control of viral replication in the absence of con-tinuous cART, but most of therapeutic vaccination studies to datehave been disappointing, and some treatment interruption strate-gies have been associated with increased adverse events [8,9].Nevertheless, the potential of an effective therapeutic vaccinationapproach is sufficiently attractive that such strategies continue to

be evaluated, with dendritic cell-based anti-HIV-1 therapeutic vac-cine strategies being of particular interest [10,11].

Myeloid DC are key cells for the generation and regulation ofthe adaptive immune responses by T cells. They are the most

5 e 29 (

piatdbpgnawa

mdc[brgHsasc

giptofusicmavttwaawfitudmtt

issltpbiwcpfi

712 C. Gil et al. / Vaccin

otent professional antigen-presenting cells, unique in their capac-ty to induce the de novo antigen-specific activation of naïve CD4nd CD8 T cells in vivo and in vitro (priming), as well as inheir ability to present exogenous antigens, including pathogenerived peptides via MHC-class I molecules to stimulate responsesy CD8 T cells, through processes called crosspriming and cross-resentation [12,13]. These extraordinary abilities make ex vivoenerated antigen-loaded myeloid DCs an attractive and potentatural “cellular” adjuvant to induce desired T cell responses tontigens. This approach has been used in clinical trials in patientsith cancer for more than a decade [14], and more recently it is

lso being explored in HIV-1/SIV infection.Several studies performed in vitro and in vivo using animal

odels have demonstrated that autologous monocyte derived-endritic cells (MDDCs) pulsed with HIV-1/SIV based immunogensan induce priming and boosting of virus specific immune response15–19]. This approach has been also applied in infected patientsy pulsing autologous MDDC with different kinds of immunogens:gp160, HIV-1 peptides, canary pox vectors carrying HIV relevantenes, autologous HIV-1 RNA and autologous whole inactivatedIV-1, and, despite the different clinical study designs, the results

uggest that the pulsed autologous MDDC immunizations were safend well tolerated, could induce lymphoproliferation and virus-pecific CD8 cell responses and, in some cases, transient virologicontrol [20–26].

Theoretically the use of whole inactivated HIV-1 as immuno-en could have potential advantages over other HIV immunogens,ncluding provision of the full complement of virion associated viralroteins, and depending on the method of inactivation, providinghese virion proteins in their native conformation, which may bef particular importance for inducing desirable responses to con-ormational determinants on the viral envelope proteins. However,se of whole inactivated virions as a vaccine immunogen also hasome potential drawbacks. Some of the viral proteins are not foundn virions, while others are present in low amounts that may notonfer optimal immunogenicity. In addition, the viral inactivationethod applied must guarantee a high reduction of infectivity to

ssure safety, although this is less critical when using autologousirus for therapeutic vaccination compared to potential prophylac-ic immunization of seronegative subjects. Phase I/II clinical trials ofherapeutic anti-HIV-1 vaccines based on autologous MDDC pulsedith autologous whole inactivated HIV-1 performed by Lu et al. [25]

nd our group [26], have demonstrated the safety of this approachnd its ability to induce HIV-1 specific cellular immune responses,ith a suggestion of antiviral activity in the Lu et al. study. Dif-

erences between the results of Lu et al. and our own prior studyncluded both the amount of virus used for the immunizations andhe method used for inactivation. In our follow study, rather thantilize the limiting amounts of autologous virus we could obtainirectly from patient plasma [26], we developed procedures to opti-ize viral production by primary cell culture in order to increase

he amount of autologous inactivated HIV-1 available for pulsinghe autologous MDDM used at each vaccination.

The production of autologous inactivated HIV-1 to be used asmmunogen in autologous therapeutic vaccines in humans entailsignificantly increased procedural complexity compared to thetandard methods used to isolate and inactivate HIV-1 for basicaboratory research. Requirements include the use of: (i) cell cul-ure methods that permit isolation and high level production ofrimary HIV isolates, in which standard reagents must be replacedy materials compatible with clinical grade GMP conditions, (ii) an

nactivation procedure that provides the necessary level of safety

ith retention of antigenic integrity, (iii) a suitable method for

oncentration and purification of inactivated virions, (iv) suitablerocedures to exclude the presence of adventitious agents, andnally (v) validation of a reproducible immunogen manufacturing

2011) 5711– 5724

process capable of meeting protocol defined product release speci-fications. All these steps and procedures must be in agreement withrelevant regulatory guidelines, which are intended to minimize therisk of adverse effects [27,28].

We describe the detailed methodology for preparation and char-acterization of autologous whole inactivated HIV-1 for use in a pilotclinical trial to assess this immunization approach (Clinical TrialDCV-2, NCT00402142; Phase II Study of Autologous Myeloid Den-dritic Cells as a “Cellular Adjuvant” for a Therapeutic HIV-1 Vaccinein Early Stage HIV-1+ Patients).

2. Material and methods

2.1. Patients

Sixty chronically HIV-1-infected patients were included in theClinical Trial DCV-2 (NCT00402142); a randomized trial withplacebo control and double blind masking to assess the safetyand activity of a therapeutic HIV vaccine consisting of autologousmyeloid dendritic cells pulsed ex vivo with high doses of inactivatedautologous HIV-1 in infected patients in a very early stages of thedisease (CD4 > 450 × 106/L) and plasma viral load (pVL) before anyantiretroviral therapy greater than 10,000 HIV-1 RNA copies/mL.The trial is being performed in two parts: Part I involves patientsoff cART for a minimum period of 2 years (n = 24 patients: 12 casesand 12 placebo control); Part II involves patients on cART withpVL lower than 50 HIV-1 RNA copies/mL for a minimum periodof six months (n = 36 patients: 24 cases and 12 placebo control).HIV-1 was isolated from the cases. A brief treatment interruptionwas required for the participants in Part II to allow virus isolation.The study was explained in detail to all patients, and all signedinformed written consent. The clinical trial was approved by boththe institutional Committee of Ethics and Clinical Investigation andthe Spanish Agency of Drug and Health Products.

2.2. Culture medium and reagents

All reagents used to manufacture the heat inactivated autolo-gous HIV-1 immunogen as a final product were sterile, endotoxinfree and, either were themselves pharmaceutical products or weremanufactured under GMP conditions. The manipulation of allopened tubes and reagents was always done under sterile con-ditions in bio II/A biosafety cabinets following the correspondingstandard operational procedures (SOPs) validated during the pre-clinical evaluation of this project.

Cell cultures were established in T75 flasks (Corning Inc.430641) using X-VIVO 20 media (Serum-free medium, Lonza-Biowhitaker) supplemented with 10% of AB human serum (ABHuS).Buffy coats and AB human plasma from anonymous numberedhealthy HIV-negative blood bank donors were obtained (Banc deSang i Teixits, Barcelona, Spain). They were tested negative forHBsAg, anti-HCV, anti-HIV-1+2, HCV-RNA, HIV-1-RNA, HBV-DNAand syphilis (TPHA) serology. Plasmas were converted into ABHuSby treatment with 10 U of human thrombin (Tissucol Duo®, Baxter)per mL of AB plasma, for 30 min at 37 ◦C, and centrifuged at 6000 × gfor 30 min. Afterwards, ABHuS was heat-inactivated at 56 ◦C for30 min, centrifuged at 6000 × g for 30 min and stored in aliquots at−20 ◦C until use. CD8+ lymphocytes were depleted from peripheralblood mononuclear cells (PBMCs) using CliniMACS Starting Kits,CD8 MicroBeads, (Miltenyi Biotec, Inc.) and sterile LS columns (Mil-tenyi Biotec, Inc.), which were certified endotoxin free (LAL Lonza).

CD8 MicroBeads 7.5 mL vials were aliquoted into 0.1 mL sterilevials and frozen at −20 ◦C until usage before the manufacturer’sexpiry date. Pharmaceutical anti-CD3 (Orthoclone OKT3®) wasprepared as a stock solution at a concentration of 10 �g/mL in X-

e 29 (

VttdPwie

HBBrRBe

2

HHp

SpCtfCcoHwtf

xF3c

2

1Pei1ac1aOtfaaer2acit

C. Gil et al. / Vaccin

IVO 10 media (Serum free medium, Lonza-Biowhitaker), filteredhrough 0.22 �m filters (Millipore) and aliquoted into sterile vials,hen frozen at −20 ◦C until use before the manufacturer’s expiryate. Pharmaceutical grade human recombinant IL-2 (18 × 106 UIroleukin® freeze-dried, Chiron Corporation, Emeryville, CA, U.S.A.)as prepared at 10,000 UI/mL using X-VIVO 10, aliquoted into ster-

le vials and frozen at −20 ◦C until use before the manufacturer’sxpiry date.

Other reagents employed for the standard cell cultures used asIV-1 isolation controls were: RPMI 1640 culture media (Lonza-ioWithaker), heat inactivated fetal bovine serum (FBS, Lonza-ioWithaker), phytohemagglutinin (PHA, Sigma–Aldrich) andecombinant interleukin-2 (rIL-2 10,000 U, Ref. 11011456001G003oche Diagnostic Systems), and the DMEM media (Lonza-ioWithaker Ref. 23718) used for the analysis of the tropism ofach autologous HIV-1 isolated.

.3. Cells

CD4-enriched PBMCs obtained from both blood buffy coats ofIV-negative unrelated healthy volunteer donors and from 150 mLIV-1 infected patients’ whole blood were used to isolate and toropagate autologous HIV from each patient.

PBMCs were obtained by ficoll centrifugation (ACCUSPINTM

ystem-Histopaque®-1077, Sigma Diagnostics) and the CD8+ lym-hocytes were depleted from PBMC by magnetic cell sorting withlini CD8 MicroBeads according to the manufacturer’s instruc-ions to obtain CD4-enriched PBMCs. The CD4-enriched PBMCsrom healthy donors were stimulated with pharmaceutical anti-D3 (10 ng/mL) for 48–72 h, at 2 × 106 cells/mL in culture mediaomposed of serum free X-VIVO 20 media supplemented with 10%f ABHuS. Twenty-four hours before initiating co-cultures withIV-1 infected cells, these CD3-stimulated cells were co-stimulatedith pharmaceutical IL-2 (10 U/mL). Activated cells were washed

hree times with X-VIVO-10 before being used in cell co-culturesor HIV-1 isolation.

Cells were analyzed by flow cytometry by using: mouse �1 (G140) PerCp, mouse �1 (G1 x40) FITC, CD3 (SK7) PerCp, CD4 (SK3)ITC and CD8 (SK1) FITC (references: 345817, 345815, 345766,45768 and 345772 respectively, all from BD Biosciences Immuno-ytometry Systems).

.4. Primary HIV-1 isolation

For each HIV-1 infected patient, the primary autologous HIV- was isolated by means of co-culture of 25 × 106 CD4-enrichedBMCs from the infected patient with 25 × 106 pre-activated CD4-nriched PBMCs obtained from a healthy donor, at 2 × 106 cells/mLn X-VIVO 20 media supplemented with the 10% of ABHuS and0 U/mL of Proleukine®, in upright T75 flasks, for 21 days (here-fter referred to as Pt–HD co-cultures). Half of the volume of eachell culture supernatant was replaced with fresh medium at days 5,1 and 17 and the cell-culture was fed with 25 × 106 of fresh pre-ctivated CD4-enriched PBMCs from a new healthy donor on day 7.n day 14, the entire cell-culture supernatant was recovered from

he flask and the infected cell pellet was mixed with 50 × 106 ofresh pre-activated CD4-enriched PBMC from a new healthy donornd resuspended, at 2 × 106 cells/mL in fresh medium. Thus for eachutologous virus isolation, CD4-enriched PBMC from three differ-nt healthy donors were employed. The cell-culture supernatantsecovered at days 14, 17 and 21, were centrifuged at 6000 × g for0 min at 21 ◦C and stored at −80 ◦C before being pooled. In parallel,

negative control for the procedure was established by separatelyulturing alone the healthy donor CD4-enriched PBMCs employedn the HIV isolation co-culture for each patient (hereafter referredo as HD–HD co-cultures). The supernatants of these HD–HD cell

2011) 5711– 5724 5713

co-cultures were recovered and subjected to the same subsequentsteps as supernatants from Pt–HD co-cultures. A flow diagram ofthe production process is shown in Fig. 1.

An additional HIV-1 co-culture utilizing standard laboratoryprocedures was performed in parallel in order to provide: (1) acontrol for infection in all cell sources (patients’ cells and healthydonors’ cells) used to produce the HIV-1, and (2) to evaluate cellsfrom the healthy donor population that could potentially representseronegative samples obtained from recently infected individu-als during the window period for HIV-1 testing. These standardco-cultures were carried out using 24-well-tissue culture plates(Cultek) with PHA pre-activated healthy donor cells in 2 mL RPMI1640 culture media, 20% FBS, 10 U/mL rIL-2, at a final density of106 cells/mL (according to the DAIDS virology manual for HIV lab-oratories [29]). The susceptibility of all healthy donor cells used inthis project to HIV-1 infection was simultaneously confirmed bystandard procedures, infecting the healthy donor cells with HIV-1BaL strain at moi < 0.008.

2.5. Viral production analyses

Viral production was analyzed by testing supernatants for bothHIV-1 p24 Antigen-ELISA (Ag HIV, Innogenetics # K1048) and forHIV-1 RNA by real time RT PCR (PCR Real time COBAS TAQMANHIV-1 Test, v1.5, Roche Diagnostic Systems), at days 7, 14, 17 and21 of the cell culture, during viral isolation in GMP co-cultures andin standard control co-cultures, as well as during subsequent stepsin the immunogen preparation process (pooled supernatants, con-centrated autologous HIV-1 final products and their correspondingcontrols). Both procedures were previously validated according tothe ICH Topic Q 2 (R1) guidelines [30] to assure that there was noassay interference (data not shown).

2.6. Heat inactivation and concentration of autologous HIV-1

The procedure for heat inactivation of HIV-1 was previouslyestablished to assess the viral infectivity reduction and viral denat-uration (shown in part in Fig. 2) [31]. For each autologous HIV-1isolated, the pool of harvested cell co-culture supernatant wassplit into 10 mL aliquots and inactivated by heat-treatment at 56 ◦Cwith 750 rpm agitation for 30 min using a thermomixer (model AG22331 HAMBURG, Eppendorf). The heat inactivated viral super-natants from the patients who participated in Part-I of the studywere concentrated by ultrafiltration using sterile VivaSpin 20,300 kDa, Centrifugal Filter Units (Vivasience Sartorius Group #VS2051) at 6000 × g for 60 min at 21 ◦C. For each patient multi-ple VivaSpin 20 filters were needed to concentrate the total pooledsupernatant volume of approximately 80 mL. Centrifugal filtrationconcentrates were washed three times using physiological saline(three times 6000 × g for 60 min at 21 ◦C). The 0.5 mL final volumerecovered from each centrifugal filter was pooled and centrifugedat 15,000 × g (CH 007466 rotor and Multifuge 1LR, Heraeus) for 2 h,at 4 ◦C. The pellets were resuspended and pooled into a 1 mL ofphysiological saline and divided into 5 aliquots of immunogen of0.2 mL each, which were stored frozen at −80 ◦C until their usageas immunogen dose to pulse the autologous MDDCs.

On the other hand, the heat inactivated viral supernatantpools from the patients who participated in Part II of the studywere concentrated by ultracentrifugation instead of ultrafiltrationbecause during the validation of the ultracentrifugation procedurewe demonstrated that viral recovery by ultracentrifugation washigher than by ultrafiltration (data not shown). Heat inactivated

viral supernatant pools from participants in Part II were concen-trated by ultracentrifugation at 100,000 × g at 4 ◦C for 32 min insterile polyallomer bottles (S5083 Seton Scientific, USA) using aT1250 fiberlite rotor followed by another ultracentrifugation at

5714 C. Gil et al. / Vaccine 29 (2011) 5711– 5724

Fig. 1. HIV-1 production process flow-chart for the preparation of the autologous HIV-1 immunogen. Bold rectangles show the HIV-1 production process steps (Pt–HD),octagons describe the production control procedures (HD–HD), grey rectangles show the quality control steps during the intermediate and final stages of the productionprocess, and the arrow text boxes describe the production process specifications for each stage. Individual lots of final product were divided into 5 aliquot/doses (3 fora n in qT MDD> negati

1Bmwsswo

2

paocd

2s

ng

utologous MDDC pulsing, 1 for quality analyses, and 1 as a retention sample), frozehe specifications of quality control for the pulsed autologous MDDC were: >8 × 106

40%, mycoplasma negative, Gram stain test negative, and microbiological culture

92,000 × g at 4 ◦C for 10 min in sterile 1.5 mL tubes (357448eckman-coulter) using a F45L-24X1,5 fiberlite rotor and a Ther-ofisher, Sorvall WX Ultra 80 centrifuge. Final pellets of both, thehole inactivated HIV-1 isolated from each patient and its corre-

ponding control, were separately pooled into 1 mL of physiologicalaline and divided into 5 aliquots of immunogen of 0.2 mL each,hich were stored frozen at −80 ◦C until their usage to pulse autol-

gous MDDCs ex vivo to prepare immunogens for vaccination.

.7. Immunogen safety and quality control

Most analyses for immunogen safety and safety controls wereerformed using frozen pooled harvested supernatants and anliquot of the frozen final heat-treated concentrated productsbtained from each patient, as well as from each healthy donorontrol, which were diluted 1/15 in physiological saline. The 1/15ilution factor was normalized for the final results.

.7.1. Analyses for potential cross contamination and viral

equence conservation

Monitoring of the viral cultures for production of the final vacci-ation product was carried out by sequencing of the HIV-1 proteaseene of: (a) the virus present in the patient plasma (extracted at

uarantine until confirmation of compliance with release specifications compliance.Cs, purity >90%, viability >85%, expression of maturation markers (CD80 and CD83)ve (result obtained after MDDC injection); described by García et al. [40].

the time of autologous HIV-I isolation for the virus stock produc-tion), and (b) the pooled supernatants of viral cultures harvestedat days 14, 17 and 21 before heat-treatment, in order to confirmthe identity of each patient’s virus and ensure that no viral cross-contamination or viral population genetic drift had occurred duringthe immunogen manufacturing.

For each sample, RNA was extracted using the QiaAmpViral RNA extraction kit (Qiagen) and a 2829 bp fragment con-tained in the pol gene was amplified using RT-PCR (SuperScriptone-Step RT-PCR Kit, Invitrogen) with the following primers:5′-ATTCTGGACATAARACARGGACC-3′ (1633U23, upper primer)and 5′-CTTCTATATATCCACTGGCTACATG-3′ (4461L25, lowerprimer). Afterwards, amplification of a 2525 bp inner frag-ment was carried out by nested PCR (Platinum®, Tag DNAPolymerase High Fidelity, Invitrogen) using the primers: 5′-CTACACTAGAAGAAATGATGACAG-3′ (1811U24, upper primer)and 5′-GACATTTATCACAGCTGGCTACTAT-3′ (4335L25, lowerprimer). Finally, a 515 bp fragment containing the proteasegene (HXB2: nt2136–nt2650) was sequenced using the Big-

Dye Terminator Cycle Sequencing Kit and the ABI 3100sequence analyzer (Applied BioSystems), with the primers:5′-TCAGAGCAGACCAGAGCCAACAGCCCCA-3′ (5′Prot2, upperprimer) and 5′-AATGCTTTTATTTTTTCTTCTGTCAATGGC-3′ (3′Prot2,

C. Gil et al. / Vaccine 29 (2011) 5711– 5724 5715

Fig. 2. Representative data produced during the validation of heat-inactivation procedure [31]. (A) HIV-1 infectivity reduction was temperature and time-dependent whenprimary HIV-1 clinical isolate (named 69/7) or laboratory HIV-1 strains (MN, HXB2, SF2 and BaL) were analyzed, reaching >4 log (99.99% virus inactivated). Time-dependentHIV-1 infectivity reduction was also observed when the supernatants were stored at room temperature. (B) We assessed denaturation of HIV-1 p24 (CA) as a model of viralprotein damage assessment by analyzing antibody reactivity to HIV-1 inactivated by different heat-treatments, using p24 capture immunoassay. Diminution of antibodyreactivity to HIV-1 p24 Ag was observed in a temperature and time-dependent manner. Heating at 56 ◦C for 30 min reached >4 log of viral infectivity reduction and resultedin less loss of antibody reactivity compared to the other treatments. (C) The potential hazard of infective HIV-1 transmission from MDDC to uninfected T-cells was analyzed.Immature MDDC from an uninfected individual were prepared by using same procedure validated for the anti-HIV1 therapeutic vaccine [40] and pulsed by different dosesof HIV-1BaL + IL-4 and GS-CSF, in presence or absence of 1 �M ZDV + 2.5 �M of amphotericin B, for 48 h. In parallel, PBMCs from the same individual were infected with thesame HIV-1 infective doses (10-fold doses, from moi = 1 to moi = 10−6 and moi = 0 (mock)). After 48 h, pulsed cells were washed three times with PBS and co-culture withautologous PHA pre activated PBMCs (1:1 ratio) and viral production was followed for 21 days. Harvest and feeding was performed at days 7 and 14 with fresh allogenic PHApre-activated PBMCs obtained from healthy donors. Supernatant HIV-1 p24 Ag content was analyzed at days: 0 (last DC-HIV pulsed washing supernatant), 7, 14 and 21 ofDC-PBMC co-cultures. (C1): Infective HIV-1 transmission was only observed at moi doses ≥0.01 when MDDCs were pulsed in the absence of ZDV and amphotericin B. (C2):H and aH 001.

lutam

2

fR

IV-1 transmission was reduced when MDDCs were pulsed in the presence of ZDVIV-1 transmission in common cell culture of PBMCs (C3) was observed at moi ≥ 0.

ower primer). Sequencher 4.1 (Gene Codes Corporation) wassed to edit and align sequences, neighbor-joining phylogeneticrees (1000 replicates) were constructed using genetic distancesnd evolutionary rates were computed using Kimura 2-parameterodel in MEGA 4.1.

.7.2. Quantification of total protein in the immunogenThe quantification of total protein in the immunogen was per-

ormed by the Bradford technique (Quick Start Bradford 1× Dyeeagent, BioRad). The concentration of HIV-1 p24 Ag in the final

mphotericin B and p24 production was observed at moi ≥ 0.1; while the expected

product was not quantified because the heat-treatment can causeprotein denaturation resulting in potential under quantification ofHIV-1 p24 Ag by ELISA.

2.7.3. Residual HIV infectivity analysisInfectious titers (TCID50/mL) of HIV-1 were analyzed from

frozen samples (pooled supernatants and the frozen final heat-treated and concentrated products or immunogen). The procedureused was as previously described with minor modifications [32,33].Briefly, titration was performed by infecting PHA pre-activated

5 e 29 (

PdpTIifViT

2

tiwfTcraacaccSecwtuIoetVm

2

aoae(tAhfp#wEtudp

2

hMtai

716 C. Gil et al. / Vaccin

BMCs from healthy donors with inputs of 40 �L of 10-foldilutions (triplicate wells) for 11 days, in 96-well cell culturelates (Cultek). Infection was determined by HIV p24 Ag ELISA.he lower quantification limit of this assay was 79 TCID50/mL.nfectious titer was calculated by the Kärber formula and thenfectivity reduction factor (Ri) was calculated using the followingormula: Ri = log10(V1 × T1/V2 × T2) where R = the reduction factor;1 = volume of starting material, T1 = concentration of virus in start-

ng material, V2 = volume of material after the inactivation, and2 = concentration of virus after the inactivation [34].

.7.4. In vitro assay for adventitious virus detectionTo ensure that the final products were free of specified adventi-

ious viruses, several samples obtained at different stages of themmunogen preparation process were subjected to cell culture

ith human (MRC-5) and non-human primate (Vero) cell lines,ollowing the recommendations of the ICH guidelines [34–37].hese cell lines are permissive to a wide range of viruses includingytomegalovirus, varicela zoster virus, herpes simple virus, respi-atory syncitial virus, adenovirus, rhinovirus, Coxakie B, Coxakie And poliovirus among others. For each patient, three samples werenalyzed: (a) the supernatant pools of Pt–HD and HD–HD cell co-ulture prepared before heat inactivation; (b) the final heat-treatednd concentrated products; and (c) all ABHuS batches used in theell co-cultures. Briefly, 300 �L of sample were inoculated in cellulture tubes containing monolayer MRC-5 or VERO (both Vircell,L, Spain), respectively. After 2 h of incubation for virus adsorption,ach cell monolayer was covered with fresh cell culture mediumonsisting of EMEM containing 10% of FBS. Inoculated cell-culturesere maintained for 4 weeks and harvested every 7 days. In addi-

ion, EMEM medium and a HSV-1 clinical isolate in EMEM weresed as negative and positive controls, respectively, in each assay.

mmunogens were discarded for its use in the vaccine when anyf the inoculated cell-cultures or subcultures in fresh cells showedither virus-induced cytopathic effect (CPE) or non-specific CPE oroxic effect. This procedure was previously validated in MRC-5 andero cell line, respectively, spiking the samples with HSV-1 as virusodel, according to the ICH guidelines [34] (data not shown).

.7.5. Microbiological sterility analysisThe procedures described here were previously validated

ccording to monograph 2.6.1 de la Ph. Eur. “Sterility” using aer-bic, anaerobic and fungal cultures with spiked positive controls tossure that there was no sample substrate related assay interfer-nce that could underestimate or overestimate the results obtaineddata not shown) [37]. Briefly, 100 �L of supernatant pool, ofhe frozen final heat-treated and concentrated products and allBHuS used in this project were inoculated onto 8 mL of brain-eart medium (BD Diagnostic Systems # 220837) and cultured

or 5 days at 35 ◦C. Later, they were sub-cultured on blood-agarlates (Columbia; agar 5% of calf blood BD Diagnostic Systems

254005) for three additional days. Macroscopic observationsere performed at 24, 48 and 72 h. Staphylococcus aureus and the

scherichia coli (obtained from the American Type Culture Collec-ion (ATCC) references # 25913 and # 25922, respectively) weresed as positive controls. This microbiological test was sensitive toetect a wide spectrum of microorganism including fungus (>100laque-forming cell).

.7.6. Mycoplasma testMycoplasma enzymatic activity was analyzed on frozen final

eat-treated and concentrated products by using MycoAlert®

ycoplasma Detection Assay (Lonza Rockland, Inc.). According tohe assay, any viable mycoplasma in the specimen were lysednd the enzymes reacted with the MycoAlert substrate catalyz-ng the conversion of ADP to ATP with a readout measured as

2011) 5711– 5724

luciferase units (URL). This procedure was previously validated toassure that there was not sample substrate assay interference thatcould underestimate or overestimate the results obtained (data notshown).

2.8. PBMC viral load

To investigate if viral production was correlated with active viralreplication in PBMCs from the patients at the time of blood extrac-tion for HIV-1 isolation, cell associated HIV-1 RNA was quantifiedin PBMCs by Real time RT-PCR (COBAS TAQMAN HIV-1 Test, v1.5,Roche Diagnostic Systems). COBAS TAQMAN is a clinical standardassay to quantify HIV-1 VL in plasma. The procedure had to bemodified and validated previously to quantify cell associated HIV-1RNA in PBMCs [38]. Briefly, total RNA was extracted from 2 × 106

PBMCs pellets using a RNeasy Kit (Qiagen) and eluted in 50 �L ofwater, then quantified. Two micrograms of total RNA diluted in50 �L of water was treated with 2 �L rDNase I (2 U/�L DNA-freeTM

Kit Ambion, Applied Biosystems) at 37 ◦C for 60 min following thekit instructions. Next, a final sample of 50 �L containing an inputof 1 �g of total RNA and 4 �L of added quantitation standard (QS)were mixed with 50 �L of HIV-1 master mix + Mn2+ solution for theamplification reaction by COBAS TaqMan HIV-1. For cellular RNA,the output results were considered as HIV-1 RNA copies/PCR reac-tion and they were adjusted to the sample input (1 �g of total RNA)by dividing the output result by three. The adjusted results wereexpressed as HIV-1 RNA copies/�g of total RNA.

2.9. Tropism of autologous HIV-1

The tropism of the HIV-1 isolated from each patient was ana-lyzed by cell culture in both U87 CD4/CXCR4 and U87 CD4/CCR5cell line. Cells were prepared in 96-well cell culture plates (Cul-tek) at 20 × 103 cells/well in DMEM with 10% of inactivated FBS.When cultures reached semiconfluent monolayers, 40 �L of pooledsupernatants prepared before heat inactivation was inoculated induplicate onto both U87 CD4/CXCR4 and U87 CD4/CCR5 and incu-bated for 12 days at 37 ◦C and 5% CO2 and observed by microscope atdays 3, 6, 9 and 12, until CPE appeared. The reference HIV-1 strainsBaL, HXB2, SF2 (obtained through the AIDS Research and ReferenceReagent Program, Division of AIDS, NIAID, NIH) were used as posi-tive controls and an HIV-1 negative sample obtained from healthydonor cell culture produced in GMP conditions was used as negativecontrol. This procedure provided a qualitative assessment of theviral tropism (R5, X4 and dual) in the viral supernatants obtainedto prepare the immunogens, but did not distinguish between dualtropic R5X4 or dual mix population.

2.10. Electron microscopy

To assess the appearance of the immunogen prepared to pulseMDDC, the final product obtained from a representative HIV-1infected patient and prepared per the protocol for patients whoparticipated in Part II of the study was evaluated. One mL ofimmunogen was pelleted at 50,000 × g for 1 h and fixed in 2% glu-taraldehyde (Merck, 1.04239.0250) in a 0.1 M sodium cacodylatebuffer (Merck 10356.0100) at pH 7.4 for at least 2 h, rinsed in a0.1 M sodium cacodylate buffer at pH 7.4 and postfixed in cold(4 ◦C) 1% osmium tetroxide in the same buffer for 1 h. The samplewas rinsed in a 0.1 M sodium cacodylate buffer at pH 7.4, dehy-

drated in an acetone series and finally embedded in Spurr’s resin.Ultrathin sections were obtained using a Reichert-Jung Ultracut Eultramicrotome, placed on copper grids and double-stained withuranyl acetate and lead citrate [39]. Ultrathin sections were exam-

C. Gil et al. / Vaccine 29 (2011) 5711– 5724 5717

P t I ti t ff ARV f t l t 6 thPart I, patients off ARV for at least 6 months

1010(a)

1010(c)

106

107

108

109

p24 a

nti

gen

su

pern

ata

nt

106

107

108

109

p24 a

nti

gen

su

pern

ata

nt

14 17 20 23103

104

105

HIV

-1 p

pg

/mL

14 17 20 23103

104

105

HIV

-1

pg

/mL

Cell co-culture day s

1011

1012(b)

nt

Cell co -culture da ys

1011

1012(d)

106

107

108

109

1010

HIV

-1 R

NA

mL

su

pern

ata

n

106

107

108

109

1010

14 17 20 23103

104

105

Cell co-culture day s

cp

/m

14 17 20 23103

104

105

Cell co-culture days

t

HIV

-1 R

NA

L s

up

ern

ata

ncp

/m

at

least 6 months,Part II Patien ts on ARV wit h PV L < 20 0 copies/m l for at

F co-cua ctivityr

i8

2

PtlwAtt(o

3

aismtf

3

v

ig. 3. Monitoring of autologous HIV-1 production by CD4-enriched PBMCs (Pt–HD)nd (B) HIV-1 RNA copies/mL supernatant and Part II (C) and (D). The HIV-1 produecovered from HD–HD cell co-culture controls were HIV-1 negative.

ned using a JEOL 1010 TEM operated at an accelerating voltage of0 kV and photographed with a Bioscan 792 camera (Gatan).

.11. Statistical analysis

All data were collected in MS Access and analyzed by usingrism v.5 software. For descriptive analysis of data, the median andhe inter-quartile range (IQR) of quantitative variables were calcu-ated. The median comparisons of variables between two groups

ere achieved by non-parametric Mann–Whitney U-test analysis.nalysis of paired data was performed by Wilcoxon signed rank

est. Spearman rank correlation was used to analyze the associa-ion between the VL in PBMCs and viral production by cell cultureeither VL or HIV-1 p24 Ag). All tests were two-tailed, and p valuesf p < 0.05 were considered significant.

. Results

Thirty-eight HIV-1 infected patients participated in isolation ofutologous HIV-1 for the preparation of heat inactivated HIV-1mmunogen. Fourteen of these 38 patients belonged to Part I of thetudy of therapeutic immunization (patients off cART for a mini-um period of 2 years), and the remaining 24 patients belonged to

he Part II of the study (patients on cART with pVL < 50 copies/mLor a minimum period of 6 months.

.1. Part I, patients off cART for a minimum period of 2 years

To produce the 14 preparation of individualized heat inacti-ated autologous HIV-1 immunogen for this group of patients it

ltures analyzed at days 14, 17 and 21st. Part I, (A) HIV-1 p24 Ag pg/mL of supernatant was highly variable among the different patients of each group. All supernatants

was necessary to perform 21 cell co-culture procedures, since inseven cases the initial process failed to meet defined predeterminedspecifications, such as insufficient HIV-1 production (4 cases), fail-ure in the standard culture for positive control (2 cases), and, inone case, because a corresponding negative control HD–HD finalproduct showed CPE in Vero cell line during adventitious test con-trol. Therefore in these cases it was necessary to repeat the cellco-culture.

At the time of blood sampling for HIV-1 isolation, the median(IQR) of plasma viral load of these 14 HIV-1 infected patientswas 4.4 (4.1–4.8) log10 copies/mL and, the median (IQR) of cellassociated PBMC viral load was 2.1 (1.6–2.8) log10 HIV-1 RNAcopies/�g of total RNA. A total of 73 buffy coats from healthydonors were required for all aspects of immunogen preparation.For each infected patient, autologous HIV-1 isolation and virus pro-duction was performed using CD4-enriched PBMC cell co-culturesfrom both the HIV-1 infected patient and healthy donors. A sig-nificant CD4-enrichment of PBMCs was obtained after negativeselection with CD8-CliniMacs magnetic microbeads from bothinfected patient cells and healthy donor cells (Wilcoxon’s test,p = 0.0001, both). From the HIV-1 infected patients the median per-centage of CD3+CD4+ in PBMC was 24.5 (15.4–39.1) and the medianpercentage of CD3+CD4+ in PBMC-CD8 depleted cells was 38.6(22.2–59.6), while for healthy donor cells the median percentage ofCD3+CD4+ in PBMC was 39.6 (29.7–46.0) and the median percent-age of CD3+CD4+ in PBMC-CD8 depleted cells was 50.8 (37.1–63.3).

Viral production in Pt–HD co-culture supernatants recoveredat days 14, 17 and 21 was highly variable among the differentsubjects. The mean ± SD of HIV-1 p24 Ag was 0.68 ± 0.6 �g/mL atday 14, 0.9 ± 1.9 �g/mL at day 17, and 0.8 ± 0.7 �g/mL at day 21.

5718 C. Gil et al. / Vaccine 29 (2011) 5711– 5724

Part I, patien ts off ARV for at least 6 months Part II, Patients on ARV with PVL < 50 copies/m L for at

least 6 months

(A) (B)

3

4rs = 0.75 p = 0.002

ntigen

ata

nt

pool

3

4rs = 0.07 p = 0.75

ntigen

ata

nt

pool

1

2

HIV

-1 p

24 a

µg/m

L s

upern

a

1

2

HIV

-1 p

24 a

µg/m

L s

upern

a

0 1 2 3 4 50

PBMCs Log HIV RNA cp/µg RNA

µ

0 1 2 3 4 50

PBMCs Log HIV RNA cp/µg RNA

µ

Fig. 4. Analysis of the association between the cell associated viral load in the infected PBMCs (log10 HIV-1 RNA copies/�g of total RNA) used to initiate the co-cultures and theamount of HIV-1 p24 Ag obtained in the supernatant pools from the HIV-1 infected patients. (A) The production of HIV-1 p24 Ag in cell-culture supernatants obtained fromt ctive H 2

p tion ot d to pg

Td8Ac

fynTA(fc(

cii

ccwfiftc

noodtgb

hiRfirm

he group of patients included in Part I of the Clinical Trial was correlated with the a = 0.002). (B) The group of patients included in Part II, which required the interruphe PBMCs used to initiate the cell-cultures and a higher viral production compareroup of patients (Spearman r2 = 0.07, p = 0.75).

he mean ± SD of log10 of HIV-1 RNA copies/mL was 8.6 ± 0.55 atay 14, 8.5 ± 0.75 log10 of HIV-1 RNA copies/mL at day 17, and.9 ± 0.7 log10 of HIV-1 RNA copies/mL at day 21 (Fig. 3a and b).ll supernatants recovered from HD–HD cell co-culture used asontrols were HIV-1 negative.

The supernatants recovered at days 14, 17, and 21 obtainedrom co-cultures (Pt–HD) for each individual patient were pooled,ielding a median volume (IQR) of 80 (77.5–80) mL; correspondingegative control co-cultures (HD–HD) were processed similarly.he supernatant pools contained a median (IQR) of HIV-1 p24g of 0.322 (0.54–0.77) �g/mL and a median (IQR) VL of 8.9

8.3–9.3) log10 copies/mL (Table 1). A positive association wasound between the VL in the infected PBMCs used to initiate theo-culture and the HIV-1 p24 Ag obtained in the supernatant poolsSpearman r2 = 0.75, p = 0.002) (Fig. 4a).

Supernatant pools were heat inactivated at 56 ◦C for 30 min andoncentrated to a final volume of 1 mL in physiological saline (here-nafter referred to as final product or immunogen), which was splitnto 5 aliquots of 0.2 mL each.

The autologous inactivated HIV-1 concentrated final productsontained a median (IQR) VL of 9.8 (9.1–10.5) log10 HIV-1 RNAopies/mL, a median (IQR) of total protein of 5.0 (3.7–9.2) mg/mL. Aide range in the amount of total protein was found among the 14nal products containing autologous HIV-1, but no significant dif-

erences were found when they were compared to the amount ofotal protein in the final product obtained from their own HD–HDell co-culture control (Wilcoxon signed rank test, p = 0.71).

HIV-1 co-receptor usage was analyzed for 10 out 14 super-atant pools. Nine out of the 10, showed R5 tropism and, 1 outf the10 showed R5X4 dual or mixed tropism. Sequence analysisf the HIV-1 protease coding region in the HIV-1 supernatant poolsemonstrated that, there was no inter-patient cross contamina-ion and that the virus isolated by cell-culture for each patient wasenetically related to the virus found circulating at the time thatlood samples were obtained for virus isolation (Fig. 5a).

Viral infectivity was analyzed in the supernatant pool beforeeat inactivation and in the heat inactivated final product. Heat

nactivation resulted in a median reduction of infectivity (IQR)

i of 5.0 (3.8–6) log10. However, in 10 out of 14 (71.4%) of thenal products containing autologous HIV-1 low residual infectivityemained detectable. Our preliminary studies suggested that weight observe low levels of residual infectivity with our standard

IV-1 replication in the PBMCs used to initiate the cell-cultures (Spearman r = 0.75,f treatment for viral isolation, showed a higher level of active HIV-1 replication in

atients from Part I, but no association was found between the two variables in this

heat inactivation treatment, especially for samples with very highpre-treatment infectious titers. However, as pulsing of the autolo-gous MDDCs occurred in the presence of ZDV and amphotericin B,this represented a very limited potential for transmissible infectiv-ity (Fig. 2). Nevertheless, in view of detectable residual infectivityin the concentrated final product we applied an additional heattreatment to reduce the possibility of HIV-1 transmission. Prepara-tions with measurable residual infectivity were thus subjected toan additional round of heat-treatment at 56 ◦C for 30 min, result-ing in elimination of detectable residual infectivity to levels belowthe detection limit of 79 TCID50/mL in all but one case. In thisinstance, despite two rounds of heat treatment and an infectivityreduction of 4 log10, residual infectivity still remained in the sample(5.5 × 103 TCID50/mL in the autologous HIV-1 final product). In theview of the fact that the virus used with this individualized autolo-gous MDDC based vaccine was the patient’s own virus, the potentialharm from residual infectivity was judged to be low. Obviously, ina prophylactic vaccine model for uninfected individuals this risk ofHIV-1 transmission would had been unacceptable.

All final products (autologous HIV-1 isolated and their cor-responding HD–HD control), as well as all ABHuS used in cellco-cultures during the manufacturing process, were tested foradventitious agents. All tested negative after culture on Vero andMRC-5 cell lines, in microbiological cultures and in the mycoplasmatest.

All HD–HD co-culture supernatants, which acted as cell culturecontrols, were HIV-1-negative at all steps of immunogen produc-tion, except one concentrated cell culture which had a borderlinepositive reaction by real time RT-PCR Cobas Taqman HIV-1 Test.Although it probably reflected an nonspecific reaction, the corre-sponding Pt–HD immunogen produced by cell co-culture was thenconsidered not compliant with protocol specifications and, there-fore, rejected for clinical trial use. Unfortunately, the establishmentof a new cell co-culture to produce the immunogen was not possiblefor this patient.

3.2. Part II, patients on cART with pVL <50 copies/mL for aminimum period of 6 months

To isolate autologous HIV-1 and propagate the virus by cellculture from the 24 participants of this group of patients withundetectable level of plasma VL on treatment, it was necessary

C. Gil et al. / Vaccine 29 (2011) 5711– 5724 5719

Table 1Summary of the characteristics of the heat inactivated HIV-1 immunogens obtained from 14 ARV naïve HIV-1 infected subjects and from 24 cART patients withpVL < 200 copies/mL at least for 6 months and subjected to treatment interruption for 3 month. A total of 80 mL of heat inactivated HIV-1 cell co-culture supernatantobtained from each patient was concentrated to 1 mL of final product (immunogen) composed of HIV-1 virions containing some non-autologous cell-proteins inserted andby non-HIV-1 microvesicles, due to the method of viral production. These products had a median of 5.8 × 109 HIV RNA copies/1 mL from Part I patients and a median of 1011

HIV RNA copies/1 mL from Part II patients, and no adventitious agents (viruses, bacteria, fungus and mycoplasma) were detected. One out of thirty-eight final products wasdiscarded at last step because of a low level, likely nonspecific borderline result for HIV RNA was detected in the corresponding HD–HD control by COBAS assay. Eleven of 24final concentrated products, obtained from patients from Part II, still contained residual HIV-1 infectivity after two heating cycles; these preparations were subjected to twoadditional freeze/thaw cycles to reduce the residual infectivity to below the quantitative limit of infectivity meeting the specification for use as immunogen. Final productsmeeting specifications were divided into five aliquots of 0.2 mL for use as MDDC pulsing dose.

Part I of the study,patients off HAART atleast for 6 months

Part II of the study, patientson HAART withpVL < 200 copies/mL at least for 6 months

Comparative analysisby non-parametricMann–Whitney test

Number of final HIV-1 immunogens prepared 14 24Median (IQR) of plasma VL of HIV-1 at the moment of HIV isolation(log10 HIV-1 RNA copies/mL)

4.4 (4.1–4.5) 4.7 (4.3–5.4) p = 0.0546

Median (IQR) of PBMCs VL of HIV-1 at the moment of HIV isolation(log10 HIV-1 RNA copies/�g of total RNA)

2.1 (1.6–2.8) 2.7 (2.2-3.1) p = 0.04

Median (IQR) of percentage of CD3+CD4+ cells from infectedpatients in the established cell co-cultures (% TCD4 cells)

38.6 (22.2–59.6) 48.2 (36.7–56.8) p = 0.5153

Median (IQR) of percentage of CD3+CD8+ cells from infectedpatients in the established cell co-cultures (% TCD8 cells)

1.5 (1–2.7) 1.9 (1.4–3.4) p = 0.2203

Median (IQR) of volume of pool of cell co-culture supernatantcollected (mL)

80 80

Median (IQR) of concentration of p24 HIV antigen in the pool(microgram p24 HIV-1 Ag/mL)

0.322 (0.14–0.77) 1.5 (1.1–2.0) p < 0.0001

Median (IQR) of HIV-1 VL in the supernatant pool (log10 HIV-1 RNAcopies/mL)

8.9 (8.3–9.3) 9.7 (9.4–9.9) p < 0.0006

Median (IQR) HIV-1 VL concentrated in final products (log10 HIV-1RNA copies/mL)

9.8 (9.1–10.5) 11.0 (10.9–11.0) p < 0.001

Median (IQR) of total protein concentrated in final products(mg/mL)

5.0 (3.5–9.2) 0.6 (0.4–0.9) p < 0.0001

Median (IQR) of infectious dose TCID50/mL × concentration factorin mL

2.9 × 107 (1.6 × 106 to 3.68) 2.0 × 109 (1.8 × 108 to 6.3 × 109) p = 0.0003

Median (IQR) of infectivity reduction dose (Ri) after inactivationtreatment (log10 Ri)

5.1 (3.8–6) 7.1 (6.2–7.9) p = 0.0004

Microbiological analysis Negative NegativeAdventitious viral analysis Negative Negative

e

tmpit(t

entraftsdp(dtac

dd1w1

Mycoplasma test Negativ

o interrupt the cART for a variable period of time (median of 3onths), until pVL reached > 4 log10. At the time of blood sam-

ling for HIV-1 isolation, the median (IQR) of pVL of these 24nfected patients was 4.7 (4.3–5.4) log10 HIV-1 RNA copies/mL and,he median (IQR) of cell associated viral load in PBMCs was 2.72.2–3.1) log10 HIV-1 RNA copies/�g of total RNA. After virus isola-ion, patients resumed cART.

A total of 26 samples from these 24 patients were required tostablish 24 cell co-cultures: 2 were discarded because they didot complete the standard culture control. For this part of the pro-ocol, 94 buffy coats from healthy HIV seronegative donors wereequired. A significant CD4-enrichment of PBMCs was obtainedfter negative selection with magnetic microbeads CD8-CliniMacsrom both infected patient cells and healthy donor cells (Wilcoxonest, p < 0.05). A significant decrease in CD8+ was observed withelection for both patient cells, as well cells recovered from healthyonors cells (Wilcoxon test, p < 0.05). From the HIV-1 infectedatients the median percentage of CD3+CD4+ in PBMC was 25.819.9–31.5) and the median percentage of CD3+CD4+ in PBMC-CD8epleted cells was 48.2 (36.7–56.8), while for healthy donor cellshe median percentage of CD3+CD4+ in PBMC was 35.7 (27.6–2.9)nd the median percentage of CD3+CD4+ in PBMC-CD8 depletedells was 43.3 (28.7–54.4).

The co-cultures were maintained for 21 days and fed every 7ays with fresh CD4-enriched PBMC obtained from new healthy

onors. Infected cell culture supernatants were recovered at days4, 17 and 21 and pooled (Fig. 3). The mean ± SD of HIV-1 p24 Agas 0.66 ± 0.5 �g/mL at day 14, a median 2.0 (IQR: ±0.8–2.8) at day

7, and 2.0 ± 0.8 at day 21. The mean ± SD of log10 of HIV-1 RNA

Negative

copies/mL was 9 ± 0.6 at day 14, 9.5 ± 0.7 at day 17, and 9.7 ± 0.3 atday 21 (Fig. 3c and d). All supernatants recovered from HD–HD cellco-culture controls were HIV-1 negative.

A median of 80 mL of supernatant was pooled from each cell co-culture with a median (IQR) of 1.5 (1.1–2.0) �g/mL of HIV-1 p24 Ag,and a median (IQR) VL of 9.7 (9.4–9.9) log10 HIV-1 RNA copies/mL.In contrast to the cultures for patients from Part I, no correlationwas found between the cell associated VL in the infected PBMCsused to initiate the co-culture and the HIV-1 p24 Ag obtained inthe supernatant pools (Spearman r2 = 0.07, p = 0.75) (Fig. 4b). Fiveout of 24 HIV-1 primary isolates (21%) showed R5X4 dual or mixedtropism, the remaining 19 out of 24 (79%) showed R5 tropism.

Supernatant pools were heat inactivated at 56 ◦C for 30 min,then frozen, and afterwards thawed and heat inactivated at 56 ◦C foranother 30 min. Heat-inactivated pools were concentrated by ultra-centrifugation into a final volume of 1 mL in physiological salinethat contained a median (IQR) VL of 11.0 (10.9–11.0) log10 HIV-1RNA total copies. All HD–HD cell co-culture supernatants, whichacted as cell culture controls, were HIV-1 negative in all steps ofimmunogen production.

The median (IQR) of total protein in the final products contain-ing the autologous HIV-1 was 0.63 (0.4–0.9) mg/mL and in the 24final products of the corresponding HD–HD cell co-culture controlstotal protein was 0.54 (0.4–0.7) mg/mL. The amount of total pro-tein obtained in the immunogen final product was slightly greater

than the HD–HD cell co-culture control (Wilcoxon signed rank test,p < 0.001).

Sequence analysis of the HIV-1 protease coding region in theHIV-1 supernatant pools demonstrated no evidence of inter-patient

5720 C. Gil et al. / Vaccine 29 (2011) 5711– 5724

(a) Par t I, patients off cA RV for at leas t 6 mont hs (b) Par t II , Patient s on cARV wit h PVL < 200 copies/ml

for at least 6 month s

0.00 5

● Plasma

▲ Culture’ supernatant

0.005

Fig. 5. The protease coding region was sequenced from the virus present in the patient plasma at the time of blood sampling for the virus stock production and from thepooled supernatants of the viral cell-cultures, in order to ensure that no significant sequence changes or contaminations had occurred. No evidence of inter-patient cross-c changt atient

cec

ht6twottu(soEia

alm

4

siagucvp

ontamination was detected. Phylogenetic trees show that despite some sequencehe virus circulating in plasma (circles), the sequences were close related for each p

ross contamination and that the virus isolated by cell-culture forach patient was not genetically distant from the virus found cir-ulating in their plasma at the moment of blood sampling (Fig. 5b).

Viral infectivity was analyzed in the supernatant pool beforeeat inactivation and in the heat inactivated final product. Heatreatment reduced the initial HIV-1 infectivity a median (IQR) of.2 (5.4–6.7) log10. However, 11 out of 24 of the final concen-rated products still retained measurable residual HIV-1 infectivityith a median (IQR) of 4 (3.4–4.4) log10 TCID50 in a final volume

f 1 mL. These eleven final products were subjected to an addi-ional freeze–thaw cycle to reduce the residual infectivity bellowhe quantitative limit of the assay. Only, 2 out of 11 of the final prod-cts of autologous HIV-1 still retained measurable HIV-1 infectivity2.5 × 103 TCID50 both in a final volume of 1 mL) after 11 days oftandard cell culture in PHA activated PBMCs, despite a reductionf 5.4 log10 and 4.7 log10 from the initial infectivity, respectively.lectron microscopy of final products (representative image shownn Fig. 6) showed viral particles, microvesicles of cellular origin andpparent protein aggregates.

All autologous HIV-1 final products and their controls gave neg-tive results for adventitious agents based on Vero and MRC-5 cellines cultures, microbiological cultures for bacteria and fungi and

ycoplasma testing.

. Discussion

Autologous dendritic cells pulsed with autologous virus repre-ent an attractive approach to therapeutic immunization in HIVnfected patients. Among the challenges for the development of anutologous whole inactivated HIV-1 to be used as the immuno-en in such protocols is the requirement to produce sufficient virus

sing methods and under conditions that satisfy clinical grade GMPonditions. We have recently reported such a study; here we pro-ide detailed methods employed to produce the individualizedreparations of inactivated autologous HIV-1 for use in this clin-

es in the protease coding-region found between the virus isolated (triangles) and, shown in matching colors.

ical trial (Clinical Trial DCV-2, NCT00402142; Phase II Study ofAutologous Myeloid Dendritic Cells as a “Cellular Adjuvant” for aTherapeutic HIV-1 Vaccine in Early Stage HIV-1+ Patients). Autolo-gous HIV-1 was isolated from two groups of patients: (a) patients offcART for a minimum period of 2 years (Part I; García et al. [40]); and(b) patients on cART with undetectable level of pVL (<50 copies/mL)for a minimum period of 6 months (Part II; on going). A treatmentinterruption was required for the participants included in Part II toallow virus isolation.

For timely optimal processing of specimens, physical proxim-ity of the laboratory/manufacturing facilities to the clinics wherethe trial participants were seen was very important. Three yearsbefore approval of the trial, background basic and preclinical stud-ies were initiated for optimizing and validating all methods to beemployed for the viral production by cell culture, identifying andminimizing potential risk factors associated with the quality andsafety of the final product. Finally, the ex vivo production of autol-ogous inactivated HIV-1, manufacturing process controls and theregulatory quality requirements were included in the final Inves-tigational Medicinal Product Dossier (IMPD, European application)for its approval in 2006 [27,28].

Primary HIV-1 isolates from each infected patient included inthe clinical trial were obtained from co-culture of CD4-enrichedPBMCs from the HIV-1 infected subject and from HIV-seronegativeunrelated healthy volunteer donors. As a consequence of the useof allogenic cells from HIV-1 negative blood donors as a substratefor virus production, an exhaustive analysis for potential adventi-tious agents in samples from different stages of the immunogenmanufacturing process was mandatory [36,37,41,42]. For each lotof immunogen prepared this analysis was performed by inoculat-ing both Vero and MRC-5 cell lines with the following samples: (a)all the pooled supernatants prepared before heat inactivation, (b)

the final heat-treated and concentrated products from the patientas well as from the donor, and (c) all the ABHuS’s batches used.All autologous HIV-1 immunogens prepared tested negative foradventitious viruses at the different stages of the manufactur-

C. Gil et al. / Vaccine 29 (2011) 5711– 5724 5721

Fig. 6. Appearance of representative pelleted HIV-1 final products concentrated by ultracentrifugation and observed by electronic microscope. (A) Whole killed HIV-1 virionspresent in a representative final immunogen preparation (white arrowheads), along with microvesicles from cellular origin and variable size (black arrowheads), bothp blackp d by b

iHcptpamtdofbpeait

mpe

roduced in the Pt–HD cell co-culture; (a) higher power view of area indicated byroduced by HD–HD cell co-culture control. (b) Higher power view of area indicate

ng process. However, one was rejected because its correspondingD–HD cell co-culture control produced cytotoxicity effect in Veroells, and a new HIV-1 isolation from this infected patient waserformed, which tested negative. An additional consequence ofhe viral production by allogenic co-culture is that allogenic cellroteins such as HLA-Class I and II antigens, beta-2-microglobulinnd ICAM-1, among others, are expected to be inserted in the viralembrane during virus budding; host cell derived membrane pro-

eins are also expected to be present in microvesicles produceduring the virus production cultures [43–46]. The developmentf anti-HLA specific antibodies against specific HLA allo-antigensrom the cell substrate employed during viral propagation haseen reported in experimental immunotherapies in non-humanrimates and in immunotherapy in humans with Remune®, appar-ntly without the raising of auto-antibodies [47–50]. The potentialppearance of auto-antibodies and anti-HLA specific antibod-es in the vaccinated patients will be evaluated in the clinicalrial.

The cell cultures performed from patients of Part II producedore virus in the pooled supernatant than those performed from

atients Part I. This median difference of HIV-1 production could bexplained mainly because the active HIV-1 replication in the PBMCs

box in panel A; (B) microvesicles (black arrowheads) present at the final productlack box in panel B.

employed to initiate the cell cultures from Part II group was signif-icantly higher than in the PBMCs employed from Part I group as aconsequence of the antiretroviral treatment interruption required.In addition, the fact that the CD8-depleted cells obtained frompatients from Part II had a higher percentage CD4-T cells comparedto those obtained in patients from Part I may also have contributedin the differential production.

At the end of the viral concentration process, the median of HIV-1 concentration in immunogens prepared in patients from PartII was higher than the median obtained in patients from Part I.This was partly due to the concentration procedure used. Accord-ing to the validation of the concentration method performed usingultracentrifugation in comparison to the ultrafiltration procedure,a median difference of 1 log10 HIV-1 RNA copies/mL higher at thefinal product was expected by ultracentrifugation. It is worth men-tioning that the percentage of HIV-1 p24 Ag in total protein inthe final product could not be measured reliably by ELISA in theheat treated final product because heating can alter the conforma-

tional structure of the viral proteins, causing an underestimationof HIV-1 p24 Ag by most of commercial ELISA. Consequently, HIV-1 RNA content was used to quantify the amount of HIV-1 in theimmunogen.

5 e 29 (

pecdpvmfrsinMspot[

oapctntacsfprsip1vgtRsoii

aitntntMr

aNr2mootot

722 C. Gil et al. / Vaccin

On the other hand, an important aspect observed in the finalroduct as a result of the difference between both methodsmployed to concentrate the virus was the decrease of total proteinoncentration by ultraconcentration (mostly cell microvesicles,etritus, ABHuS proteins such as human albumin, and proteinsresent in the X-vivo cell culture medium). The presence of non-iral residual proteins in the final product was unavoidable in theanufacturing process. Nevertheless, our previous studies per-

ormed in vitro during the preclinical evaluation of the T-cellesponses induced by HIV-1 pulsed MDDC, obtained from infectedubjects, showed that: (1) the non-viral residual stock of proteinn the final product obtained after the concentration of the super-atant pools by ultrafiltration did not interfere in the ability ofDDC pulsed with these materials to present derived antigens and

timulate HIV-1 specific immune response, and (2) the proteinsresent in the final product of the concentrated supernatant poolsbtained from uninfected HD–HD cell co-cultures did not inducehe production of IL-2 and INF-gamma as an allospecific response51].

Heating is a well established general method for inactivationf viruses. Thermal treatment at 56 ◦C for 30 min generally givesn infectivity reduction >4 log10 and the manipulation of the sam-les required for this type of the viral inactivation is minimal inomparison to most chemical inactivation procedures. However,he main drawback is that due to potential heat denaturation theative structure of viral proteins can be considerably affected. Inhis study, two rounds of heat treatment were required to assure

high infectivity reduction. It was particularly difficult to achieveomplete elimination of residual infectivity in those samples withtarting levels of more than 8 log10 of infectivity, mainly obtainedrom patients included in Part II of the clinical trial. These sam-les required two additional freeze/thaw cycles to reduce theesidual infectivity bellow the quantitative limit for the vaccinepecification. Of note, we utilized a highly sensitive multiroundnfectivity assay to quantify infectivity. While some immunogenreparations contained low amounts of residual infectious HIV-, the input of immunogen used to load MDDC at each dose ofaccination was not considered a significant hazard (moi < 0.001),iven that the virus was autologous and taking into considera-ion that the pulsing is performed in the presence of 1 �M ofetrovir® and amphotericin B, a fungicidal antibiotic with demon-trated anti-HIV-1 activity (Fig. 2) [52,53]. Nevertheless, any levelf measurable residual infectivity would be unacceptable for anmmunogen to be used for a prophylactic vaccine for uninfectedndividuals.

Heat treatment can denature proteins and thus runs the risk ofltering the conformational structure of the viral proteins, includ-ng Gp120, potentially impairing immune recognition. While heatreatment can indeed denature proteins and alter immune recog-ition, there is at least some published evidence that milder heatreatments do not necessarily completely abrogate immune recog-ition for HIV-1 [54,55], and our own preliminary data suggesthat heat inactivated virions can still be effectively taking up by

DDCs, processed and presented to stimulate HIV-1 specific T cellesponses [26,51].

Several other chemical procedures to inactivate the HIV-1 suchs beta-propiolactone, formaldehyde, psoralen/UV, aldhithiol-2,-ethylmaleimide and recently taurine chloramine have been

eported [44,54–60]. At the time of designing the Clinical Trial DCV- and preparing the documentation for its approval, some of theseethods were considered and tested. However, due to the difficulty

f assuring the complete elimination from the final product of some

f the chemical products available to inactivate HIV-1, along withhe lack of available full and formal toxicological evaluation of somef these products, we took the conservative decision of inactivatinghe HIV-1 by heat-treatment [51].

2011) 5711– 5724

The analysis of the protease coding region of plasma virus andcultured virus derived from contemporaneously isolated patientvirus allowed us to confirm identity of expanded virus as beingfrom the correct patient. There is theoretical concern about thesequence differences that may arise through the expansion of virusin cell cultures in comparison to the viruses circulating in plasma.Analysis of other target genes such as env might be expected to bemore informative in this regard, in conjunction with the ongoingclinical study additional sequence analysis will be performed of theculture expanded virus, in parallel with evaluation of the geneticdivergence among the primary HIV-1 variants isolated, the virusescirculating in plasma at the time of immunization and the virusescirculating in plasma at the clinical trial end point.

In summary, the ex vivo isolation and production of autolo-gous HIV-1 by cell culture, fulfilling clinical grade requirementsfor preparation of a therapeutic vaccine was feasible, consistentand could be performed according to the pre-determined specifica-tions. High sequence homology was found between the inactivatedvirus immunogen and the HIV-1 circulating in plasma at thetime of HIV-1 isolation. Immunogens contained up to 109 HIV-1 RNA copies/mL showed considerably reduced infectivity afterheat inactivation (median of 5.6 log10), and were free of specifiedadventitious agents. The immunogenic activity of this therapeuticvaccine is being assessed in a limited pilot clinical trial (ClinicalTrial DCV-2, NCT00402142; Phase II Study of Autologous MyeloidDendritic Cells as a “Cellular Adjuvant” for a Therapeutic HIV-1Vaccine in Early Stage HIV-1+ Patients), which includes two differ-ent groups of patients: viremic patients [40] and aviremic patients(pVL < 50 copies/mL, currently in progress). If activity is demon-strated, we will want to characterize the basis of this activity, and ifpossible, to define a more broadly applicable, feasible mechanismto induce these activities.

Acknowledgments

The authors are grateful to all blood donors and to techniciansperforming blood draws. We would also thank to Dr Silvia Sauledafrom Banc de Sang i Teixits, Hospital Universitari de la Vall d’Hebronde Barcelona, Spain, to the AIDS Research and Reference ReagentProgram, Division of AIDS, NIAID, NIH for their kind gift of HIV-1 BaL and SF2 strains (Dr. Suzanne Gartner, Dr. Mikulas Popovic,Dr. Robert Gallo and Dr Jay Levy), to Julian Bess, Charles M Trubeyfrom AIDS and Cancer Virus Program, SAIC Frederick Inc., NationalCancer Institute-Frederick, Frederick Maryland, USA for their helpin the development of procedures to concentrate the virus byultracentrifugation and the preparation of viral pellets for elec-tron microscopy; Nuria Cortadellas and Almudena Garcia, Facultatde Medicina. Unitat Microscòpia Electrònica, Serveis Cientifico-tècnics,University of Barcelona, Spain, for electron microscopy and to NuriaPiqué, Department of Microbiology and Parasitology, PharmacyFaculty, Universitat de Barcelona, Spain, for her help to prepare theInvestigational Medicinal Product Documentation (IMPD) requiredby the Spanish Agency of Drug and Health Products for its approval.

This study was supported in part by FIPSE (2005-VA36536),Fondo de Investigaciones Sanitarias (PI061259), ORVACS(Manon07), Instituto de Salud Carlos III- Red de Investigaciónen sida (RD06/0006) and HIVACAT (Projecte de recerca de lavacuna de la SIDA) and in part with federal funds from the NationalCancer Institute, National Institutes of Health (USA) under contractHHSN261200800001E. Dr Felipe García received a Research Grantfrom the “Institut d’Investigacions Biomèdiques August Pi i Sunyer

(IDIBAPS)”. Dr. José M. Miró holds an INT10/219 IntensificationResearch Grant (I3SNS & PRICS programs) from the “Instituto deSalud Carlos III, Madrid (Spain)” and the “Departament de Salut dela Generalitat de Catalunya, Barcelona (Spain)”.

e 29 (

o

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[[

[

[

[

[[

[

[

[

[

[

[

[

[

[

[

[

C. Gil et al. / Vaccin

Conflict of interest: The authors do not have a commercial orther association that might pose a conflict of interest.

eferences

[1] Pantaleo G, Koup RA. Correlates of immune protection in HIV-1 infection:what we know, what we don’t know, what we should know. Nat Med2004;10(8):806–10.

[2] Plana M, García F, Gallart MT, Miro JM, Gatell JM. Lack of T-cell proliferativeresponse to HIV-1 antigens after one year of HAART in very early HIV-1 disease.Lancet 1998;352:1194–2119.

[3] Chomont N, El-Far M, Ancuta P, Trautmann L, Procopio FA, Yassine-Diab B, et al.HIV reservoir size and persistence are driven by T cell survival and homeostaticproliferation. Nat Med 2009;15(August (8)):893–900.

[4] Lozano JM, De la Rosa O, García-Jurado G, Luque J, Solana R, Kindelán JM,et al. Impaired response of HIV type 1-specific CD8(+) cells from antiretroviral-treated patients. AIDS Res Hum Retrov 2007;23(10):1279–82.

[5] Richman DD, Margolis DM, Delaney M, Greene WC, Hazuda D, Pomer-antz RJ. The challenge of finding a cure for HIV infection. Science2009;323(5919):1304–7.

[6] Perno CF, Moyle G, Tsoukas C, Ratanasuwan W, Gatell J, Schechter M. Over-coming resistance to existing therapies in HIV-infected patients: the role ofnew antiretroviral drugs. J Med Virol 2008;80(4):565–76.

[7] Mallewa JE, Wilkins E, Vilar J, Mallewa M, Doran D, Back D, et al. HIV-associatedlipodystrophy: a review of underlying mechanisms and therapeutic options. JAntimicrob Chemother 2008;62(4):648–60.

[8] Letvin NL, Walker BD. Immunopathogenesis and immunotherapy in AIDS virusinfections. Nat Med 2003;9(July (7)):861–6.

[9] Autran B, Carcelain G, Combadiere B, Debre P. Therapeutic vaccines for chronicinfections. Science 2004;305(5681):205–8.

10] Strategies for Management of Antiretroviral Therapy (SMART) Study GroupEl-Sadr WM, Lundgren JD, Neaton JD, Gordin F, Abrams D, et al. CD4+ count-guidedinterruption of antiretroviral treatment. N Engl J Med 2006;355(22):2283–96.

11] Rinaldo CR. Dendritic cell-based human immunodeficiency virus vaccine. JIntern Med 2009;265(1):138–58.

12] Walsh SR, Bhardwaj N, Gandhil RT. Dendritic cells and the promise of ther-apeutic vaccines for human immunodeficiency virus (HIV)-1. Curr HIV Res2003;1(2):205–16.

13] Banchereau J, Steinman RM. Dendritic cells and the control of immunity. Nature1998;392(6673):245–52.

14] Palucka K, Ueno H, Zurawski G, Fay J, Banchereau J. Building on dendritic cellsubsets to improve cancer vaccines. Curr Opin Immunol 2010;22(2):258–63.

15] Climent N, Martinez-Navio JM, Gil C, García F, Rovira C, Hurtado C, et al. Adeno-sine deaminase enhances T-cell response elicited by dendritic cells loaded withinactivated HIV. Immunol Cell Biol 2009;87(8):634–9.

16] Lu W, Wu X, Lu Y, Guo W, Andrieu JM. Therapeutic dendritic-cell vaccine forsimian AIDS. Nat Med 2003;9(1):27–32.

17] Frank I, Santos JJ, Mehlhop E, Villamide-Herrera L, Santisteban C, Gettie A, et al.Presentation of exogenous whole inactivated simian immunodeficiency virusby mature dendritic cells induces CD4+ and CD8+ T-cell responses. J AcquirImmune Defic Syndr 2003;34(1):7–19.

18] Lifson JD, Rossio JL, Piatak Jr M, Bess Jr J, Chertova E, Schneider DK, et al.Evaluation of the safety, immunogenicity, and protective efficacy of wholeinactivated simian immunodeficiency virus (SIV) vaccines with conformation-ally and functionally intact envelope glycoproteins. AIDS Res Hum Retrov2004;20(7):772–87.

19] Kundu SK, Engleman E, Benike C, Shapero MH, Dupuis M, van Schooten WC, et al.A pilot clinical trial of HIV antigen-pulsed allogeneic and autologous dendriticcell therapy in HIV-infected patients. AIDS Res Hum Retrov 1998;14(7):551–60.

20] Connolly NC, Whiteside TL, Wilson C, Kondragunta V, Rinaldo CR, Riddler SA.Therapeutic immunization with human immunodeficiency virus type 1 (HIV-1)peptide-loaded dendritic cells is safe and induces immunogenicity in HIV-1-infected Individuals. Clin Vaccine Immunol 2008;15(2):284–92.

21] Ide F, Nakamura T, Tomizawa M, Kawana-Tachikawa A, Odawara T, Hosoya N,et al. Peptide-loaded dendritic-cell vaccination followed by treatment interrup-tion for chronic HIV-1 infection: a phase 1 trial. J Med Virol 2006;78(6):711–8.

22] Kloverpris H, Karlsson I, Bonde J, Thorn M, Vinner L, Pedersen AE, et al. Induc-tion of novel CD8+ T-cell responses during chronic untreated HIV-1 infectionby immunization with subdominant cytotoxic T-lymphocyte epitopes. AIDS2009;23(11):1329–40.

23] Gandhi RT, O’Neill D, Bosch RJ, Chan ES, Bucy RP, Shopis J, et al. A ran-domized therapeutic vaccine trial of canarypox-HIV-pulsed dendritic cells vscanarypox-HIV alone in HIV-1-infected patients on antiretroviral therapy. Vac-cine 2009;27(43):6088–94.

24] Routy JP, Boulassel MR, Yassine-Diab B, Nicolette C, Healey D, Jain R, et al.Immunologic activity and safety of autologous HIV RNA-electroporated den-dritic cells in HIV-1 infected patients receiving antiretroviral therapy. ClinImmunol 2009.

25] Lu W, Arraes LC, Ferreira WT, Andrieu JM. Therapeutic dendritic-cell vaccinefor chronic HIV-1 infection. Nat Med 2004;10:1359–65.

26] García F, Lejeune M, Climent N, Gil C, Alcamí J, Morente V, et al. Therapeuticimmunization with dendritic cells loaded with inactivated autologous HIV-1in chronic HIV-1 infected patients. J Infect Dis 2005;191:1680–5.

27] Attachment 2: Common Technical Document Headings for InvestigationalMedicinal Product Quality Data. Detailed guidance for the request for authori-

[

2011) 5711– 5724 5723

sation of a clinical trial on a medicinal product for human use to the competentauthorities, notification of substantial amendments and declaration of the endof the trial. Pharmaceuticals: regulatory framework and market authorizations.EMEA. Brussels, ENTR/CT 1; April 2004.

28] Lebron JA, Wolf JJ, Kaplanski CV, Ledwith BJ. Ensuring the quality, potencyand safety of vaccines during preclinical development. Expert Rev Vaccines2005;4(6):855–66.

29] National Institute of Allergy and Infectious Diseases, Division of AIDS. DAIDSvirology manual for HIV laboratories, publication no. NIH-97-3828. Washing-ton, D.C.: U.S. Health and Human Services; 1997.

30] ICH Topic Q 2 (R1) Validation of Analytical Procedures: Text and Methodology[CPMP/ICH/381/95].

31] Gil C, Climent N, Hurtado C, Castro P, García M, Sánchez-Palomino S, et al.Effectiveness of different chemical and physical inactivation methods to reducethe infectivity of HIV-1 Strains Aids Vaccine 2006 Conference Amsterdam, TheNetherlands, Abstract 132.00; 2006.

32] Japour AJ, Mayers DL, Johnson VA, Kuritzkes DR, Beckett LA, Arduino JA, et al.A standardized peripheral mononuclear cell culture assay for the determina-tion of drug susceptibilities of clinical human immunodeficiency virus type 1isolates. Antimicrob Agents Chemother 1993;37:1095–101.

33] Guidance on viral validation studies: the design, contribution and interpre-tation of studies validating the inactivation and removal of viruses; 1996[CPMP/BWP/268/95].

34] ICH Topic Q 5 A (R1). Quality of biotechnological products: viral safety eval-uation of biotechnology products derived from cell lines of human or animalorigin; October 1997 [CPMP/ICH/295/95].

35] ICH Topic Q 5 D. Quality of biotechnological products: derivation and charac-terisation of cell substrates used for production of biotechnological/biologicalproducts; March 1998 [CPMP/ICH/294/95].

36] Schiff LJ. Review: production, characterization, and testing of banked mam-malian cell substrates used to produce biological products. In Vitro Cell DevBiol 2005;41:65–70.

37] European Pharmacopoeia, 5th edition. Strasbourg 2.6.1 Sterility; 2005.38] Gil C, Garcia MT, García F, Miró JM, Agüero F, Alós L, et al. Evaluation of the Roche

COBAS(®) TaqMan(®) HIV-1 test for quantifying HIV-1 RNA in infected cells andlymphoid tissue. J Virol Methods 2011, doi:10.1016/j.jviromet.2011.03.026.

39] Reynolds ES. The use of lead citrate at high pH as an electron opaque stain inelectron microscopy. J Cell Biol 1963;17:208–12.

40] García F, Climent N, Assoumou L, Gil C, González N, Alcamí J, et al. A therapeuticdendritic cell-based vaccine for HIV-1 infection. J Infect Dis 2011;203:473–8.

41] Sheets RL, Goldenthal KL. Traditional approach preventive HIV vaccines:what are the cell substrate and inactivation issues? AIDS Res Hum Retrov1998;14(7):627–33.

42] Sheets RL. Adventitious agent test methods. Dev Biol (Basel) 2006;123:135–45.43] Bess Jr JW, Gorelick RJ, Bosche WJ, Henderson LE, Arthur LO. Microvesicles are a

source of contaminating cellular proteins found in purified HIV-1 preparations.Virology 1997;230(1):134–44.

44] Richieri SP, Bartholomew R, Aloia RC, Savary J, Gore R, Holt J, et al. Characteriza-tion of highly purified, inactivated HIV-1 particles isolated by anion exchangechromatography. Vaccine 1998;16(2–3):119–29.

45] Cantin R, Methot S, Tremblay MJ. Plunder and stowaways: incorporation ofcellular proteins by enveloped viruses. J Virol 2005;79:6577–87.

46] Chertova E, Chertov O, Coren LV, Roser JD, Trubey CM, Bess Jr JW, et al.Proteomic and biochemical analysis of purified human immunodeficiencyvirus type 1 produced from infected monocyte-derived macrophages. J Virol2006;80:9039–52.

47] Bergmeier LA, Walker J, Tao L, Cranage M, Lehner T. Antibodies to humanand non-human primate cellular and culture medium components inmacaques vaccinated with the simian immunodeficiency virus. Immunology1994;83(2):213–20.

48] Leith JG, Clark DA, Matthews TJ, Rosenthal KL, Luscher MA, Barber BH, et al.Assessing human alloimmunization as a strategy for inducing HIV type 1neutralizing anti-HLA responses. AIDS Res Hum Retrov 2003;19(11):957–65.

49] Page M, Ojugo A, Imami N, Hardy G, Gotch F, Almond N. Specificity ofanti-human leukocyte antigen antibody responses after immunization withRemune, an inactivated HIV-1 vaccine. AIDS 2007;21(3):375–7.

50] Lakhashe SK, Thakar MR, Bharucha KE, Paranjape RS. Quantitation of HLA pro-teins incorporated by human immunodeficiency virus type 1 and assessmentof neutralizing activity of anti-HLA antibodies. J Virol 2008;82(1):428–34.

51] Climent N, Gil C, Nomdedeu M, García F, Rovira C, Oliva H, et al. Influence ofdoses and method of inactivation of antigen on in vitro T cell responses inducedby HIV-pulsed dendritic cells. In: AIDS Vaccine 2005. 2005 [Abstract].

52] Hansen JE, Nielsen C, Svenningsen A, Witzke N, Mathiesen LR. Synergisticantiviral effect in vitro of azidothymidine and amphotericin B methyl esterin combination on HIV infection. Scand J Infect Dis 1992;24(1):35–9.

53] Waheed AA, Ablan SD, Mankowski MK, Cummins JE, Ptak RG, Schaffner CP, et al.Inhibition of HIV-1 replication by amphotericin B methyl ester: selection forresistant variants. J Biol Chem 2006;281(39):28699–711.

54] Grovit-Ferbas K, Hsu JF, Ferbas J, Gudeman V, Chen IS. Enhanced binding of anti-bodies to neutralization epitopes following thermal and chemical inactivation

of human immunodeficiency virus type 1. J Virol 2000;74(13):5802–9.

55] Poon B, Safrit JT, McClure H, Kitchen C, Hsu JF, Gudeman V, et al. Induction ofhumoral immune responses following vaccination with envelope-containing,formaldehyde-treated, thermally inactivated human immunodeficiency virustype 1. J Virol 2005;79(8):4927–35.

5 e 29 (

[

[

[

[

724 C. Gil et al. / Vaccin

56] Race E, Stein CA, Wigg MD, Baksh A, Addawe M, Frezza P, et al. A multistepprocedure for the chemical inactivation of human immunodeficiency virus foruse as an experimental vaccine. Vaccine 1995;13(16):1567–75.

57] Whiteside TL, Piazza P, Reiter A, Stanson J, Connolly NC, Rinaldo Jr CR, et al.

Production of a dendritic cell-based vaccine containing inactivated autologousvirus for therapy of patients with chronic human immunodeficiency virus type1 infection. Clin Vaccine Immunol 2009;16(2):233–40.

58] Rossio JL, Esser MT, Suryanarayana K, Schneider DK, Bess Jr JW, Vasquez GM,et al. Inactivation of human immunodeficiency virus type 1 infectivity with

[

2011) 5711– 5724

preservation of conformational and functional integrity of virion surface pro-teins. J Virol 1998;72(10):7992–8001.

59] Morcock DR, Thomas JA, Gagliardi TD, Gorelick RJ, Roser JD, ChertovaEN, et al. Elimination of retroviral infectivity by N-ethylmaleimide with

preservation of functional envelope glycoproteins. J Virol 2005;79(3):1533–42.

60] Dudani AK, Martyres A, Fliss H. Short communication: rapid preparation of pre-ventive and therapeutic whole-killed retroviral vaccines using the microbicidetaurine chloramine. AIDS Res Hum Retrov 2008;24(4):635–42.