Persistence of a HIV-1 CTL epitope despite in vitro reactivity

Eur. J. Immunol. 1991. 21: 2637-2640

2637

Short paper Andreas Meyerhans.+*, Gaes DadaglioAv, Jean-PierreVartanian.", Pierre Langlade-DemoyenA, Ronald Frankv, B i i t t a &om, Fernando PlataA and Simon Wain-Hobson. Laboratoire de R6trovirologie MolBculaire., Laboratoire d'Immunopathologie des RBtrovirusA,Institut Pasteur, Paris and GBF', Braunschweig and State Bacteriological Laboratory., Karolinska Institute, Stockholm

In vivo persistence of a HIV-1-encoded HLA-B27-restricted cytotoxic T lymphocyte epitope despite specific in vitro reactivity* A large number of human immunodeficiency virus type 1 (HIV-1) specific HLA-restricted cytotoxicTcel1 (CTL)epitopes have been mapped, including an HLA-B27-restricted immunodominant epitope within p25gag. Accordingly, this segment of the HIV-1 provirus was amplified by the polymerase chain reaction from DNA derived from fresh uncultured peripheral blood mononuclear cells (PBMC) of four HLA-B27 HIV-l-infected individuals. In all cases the majority of infected PBMC bore sequences encoding the HLA-B27-restrictedpeptide. CTL escape mutants had not accumulated in vivo 8 and 14months later despite demonstrable CTL activity in vitro. These data emphasize the importance of silently infected lymphocytes in evading immune surveillance.

from uninfected [7] lymphocytes may restrict viral replication ex vivo.The role of CTL in HIV infection in vivo is less One of the problems associated with HIV, as indeed with clear, although the high frequency of HIV-specific HLARNA viruses, is that replication proceeds in the absence of restricted CTL [8-101 during the asymptomatic phase proof-reading. Consequently, average nucleotide misincor- would suggest that they may play a role in disease poration errors per base per replication cycle of the order of progression. 10-4-2 x lop5 are encountered for retroviral genomes [l, 21. Given that the HIV genome is approximately 104 A large number of HIV-l-specific HLA class I-restricted bases, any description of HIV in vivo or a HIV isolate ex CTL epitopes have been mapped [ll-141 including a vivo must take into account the inherent genetic dispersion. HLA-B27-restricted immunodominant epitope within Populationsof closely related but distinct viral genomes are p25gag [13]. They have all been identified by the use of referred to as quasispecies.Vira1quasispecies are exquisite- syntheticpeptides to appropriate target cells. In an attempt ly sensitive to selection pressures, the immune system being to evaluate the relevance of these findings to the in vivo perhaps the most potent one. Ex vivo culture is another, situation, as well as to circumvent the problems associated invariably resulting in the outgrowth of a minor variant in with ex v i m culture, the DNA segment encoding this HLA-B27-restricted epitope was amplified by PCR from vivo [3]. fresh uncultured PBMC of four HLA-B27 HIV-1 seroposiThe isolation of HIV is somewhat special in that the ex vivo tive individuals. culture of infected CD4 cells takes place in the presence of HIV-antigen-specificCTL. Certainly the CD8+ population can greatly restrict and even abolish HIVreplication ex vivo 2 Materials and methods [4-6].Whether this is due to direct lysis of the target cells or inhibition of HIV replication is not clear. Other data 2.1 Patients suggest that a soluble factor secreted by the CD8+ suppressor cell population from infected individuals [5-71 or even Two (FOU and FOI) were asymptomatic carriers (CDC stage 11), while two others (DEM and 04)had progressed further (CDC stagem). The CD4+ Tlymphocyte cell [I 95341 counts were: patient DEM, 414/pl; patient 04, 457/p1; * This work was supported by grants from Institut Pasteur, patient FOU, 29.03.89 not determined, 7.11.89 79O/pl, 21.05.90 640/pl; patient FOI, 29.03.89 69O/pl, 7.11.89 EAgence National de Recherches sur le SIDA. Present address: Department of Virology, Institut for Medical 90O/pl, 21.05.90 820/pl. The HLA types were: DEM, Microbiology and Hygiene, University of Freiburg, D-7800 A(1,2), B(8,27); patient 04, A(30,31), B(27,40); FOU, Freiburg, FRG. A(30,32), B(14,27); FOI, A(3,25), B(18,27).

1 Introduction

A

Recipient of a fellowship from the Centre International des Etudiants et Stagiaires. Supported by the Ministbre de la Recherche et de la Technologie.

Correspondence: Simon Wain-Hobson, Laboratoire de R6trovirologie Molkculaire, Institut Pasteur, 28 rue du Dr. Roux, F-75724 Paris Cedex 15, France We dedicate this paper to the memory of Fernando Plata who died on May 5, 1991. 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1991

2.2 Polymerase chain reaction (PCR)

PBMC were purified on Ficoll-Hypaque and approximately 5 x lo6cells were taken for PCR analysis; the rest were used for CTL studies. Cellswere lysed and DNA was extracted as usual [3] in a laboratory not working on HI\! Onemicrogram samples of DNA were amplified using the SK38 and SK39 amplificationprimers [W] under standard condi0014-2980/91/1010-2637$3.50+ .25/0

2638

Eur. J. Immunol. 1991. 21: 2637-2640

A. Meyerhans, G. Dadaglio, J.-PVartanian et al.

tions. Thermal cycling parameters were: 55 "C, 30 s, 72°C 30 s and 94"C, 30 s for 45 cycles followed by a 10-min elongation period. High temperature Taq addition was employed to start the reactions. PCR products were blunt end ligated into the Sma I site of M13mp18 RF DNA. After transformation of Escherichia coli strainTG1 recombinants were screened with the SK19 hybridization probe [16]. Twenty recombinants from most samples were sequenced by the dideoxy method resulting in a resolution of 5%. Eighteen clones were sequenced for FOU (21.05.90) and nineteen for FOI (7.11.89) and 04 (15.02.89), respectively. The size of the segment is such that the PCR-associated error is essentially of the order of 0.3 amino acid per 20 sequences [3, 161.

-

2.3 Peptides Peptides were synthesized on cellulose discs using FmodtBu chemistry [17] and all were purified by FPLC on C18-silicagel. The control represented an arbitrarily chosen and unrelated, non HIV-1, peptide (RKRSQMLFRGRRASQ). 2.4 CTL assays

Human HIV-specific CTL lines were generated in autologous MLC by repeated stimulation of 106 PBMC with lo5 X-irradiated (lo4 rad) HIV-1 Lai (which encodes the peptide 30 sequence) infected blasts in 2 ml RPMI 1640culture medium supplemented with 1 m~ sodium pyruvate, 20% FBS and 10 U human IL2 from conditioned medium [18]. CTL were harvested after 3 weeks in culture. Prior to each cytotoxicity assay, peptides were incubated for 2 h with 104 51Cr-labeledP815 (HLA-B27) target cells [19], in a final volume of 200 ml.While peptide 30 was recognized over a 0.01-10 mM range, the final peptide concentration used in the experiment shown was 1mM. Percent specific 51Cr release was determined after 18 h of incubation.

3 Results The populations of proviral sequences, or quasispecies, present in the PBMC taken between February and April 1989 are given in Fig. 1. In each case the major form, appearing at 75% -95% of all sequencespresent, was that of the HLA-B27-restrictedpeptide (identified here as peptide 30). As usual, a number of minor forms accompanied the major form and most were derived from the major form by one or two base substitutions. The structure of the gag quasispecies was established for the two asymptomatic patients (FOU and FOI) 8 and 14 months 1ater.The additional quasispecies are also shown in Fig. 1. As noted previously [3] there were significant fluctuations within the viral populations. Thus for FOU peptide 41, hitherto a minor form, predominated 8 months later. After a further 6 months, a mixture of peptides 37 and 30 was present at approximately equal proportions. A similar situation pertained for patient FOI. In order to rule out the possibility of a PCRcontamination, laboratory plasmids HIV-1 Lai [20,21] and Oyi [22] encode

BEM 11.04.89 Freq Pep YKRWIILGLNKIVRMYS 95% 30 T 5% ns

Freq Pep 15.02.89 YKRWIILGLNKIVRMYS 76% 30 V 12% ns R 6% ns H 6% n s

29.03.89 Freq Pep YKRWIILGLNKIVRMYS 75% 30 * I 1 0 % def VM 5% 41 E K 5% def K* V 5% ns

29.03.89 Freq Pep YKRWIILGLNKIVRMYS 80% 30 M 20% 37

07.11.89 YKRWIVMGLNKIVRMYS 65% I 20% IL 5% I 5% IIR 5%

07.11.89 YKRWIIMGLNKIVRMYS T

41 37

V

30

88% 37 6% ns

6% ns

ns ns

21.05.90 YKRWIIMGLNKIVRMYS 44% 37 L 39% 30 V 11% 41 6% 33 VL

21.05.90 YKRWIIMGLNKIVRMYS *

L I

L I

K

H

65% 10% 10% 10% 5%

37 30 def

ns ns

Figure 1. Structure of HIV-1 gag quasispecies from four patients (DEM, 04,FOU, FOI) all bearing the HLA-B27 antigen and from two (FOU, FOI) over a 1Cmonth period.The dates corresponding to the samples are given. All protein sequences are shown using the one-letter code which is: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser;T,Thr; V,Val; W,Trp and Y,vr.Those of the minor forms have been aligned with respect to the major form. Only amino acid differences are noted. The frequency of each sequence is given on the right along with a peptide identifier code. An additional two residues at the carboxy terminus were sequenced as part of the amplified region but are not shown. The sequences given corresponds to the peptides synthesized. Def identifies a defective viral sequence due to the presence of an in phase stop codon (*). Ns means that the correspondingpeptide was not synthesized.

precisely the sequences shown above, the first hypervariable envelope region was amplified from each of the samples. Fingerprinting [23] using the length variation inherent to this region proved that each sample was distinct from every other and unlike any laboratory plasmid. In addition, part of the amplified product was cloned into M13mp18 as described above and a couple of recombinants were sequenced. All sequences were distinct and did not resemble any so far derived in the laboratory (data not shown). The interpretation of these gag quasispecies data was contingent upon there being detectable CTL response in v i m to these epitopes. Using the appropriate synthetic peptides (i.e. correspondingto peptide sequences 30,33,37 and 41, Fig. 1)in v i m CTL activities were analyzed using a mouse mastocytoma cell line (P815) stably expressing the HLA-B27 cDNA clone [191. CTL assays were undertaken for the initial and final samples from patients FOU and FOI. Representative data are given in Fig. 2 and in all cases specific lysis was observed. In the case of FOU all four peptides (30, 33, 37 and 41) were recognized by CTL despite some small differences at low E/T ratios. Peptides 30 and 37 were both efficiently recognized ex vivo by CTL

Eur. J. Immunol. 1991. 21: 2637-2640 100

-

80

-

60

-

Pept. 37 Pept.41

-

Pept. 33

v)

.4

&

Persistence of a HIV-1 CTL epitope despite in vitro reactivity

4

100

u

5u OI k cc 0

.u

40

i

E:

aJ

;

20-

k

# 0.26

, 0.8

#

T 2.6

I

< 8

26

Pept.30 Control

2639

1FOI

4 “1

Pept. 30 Pept. 37 Control

20 0

E:T ratio

0.013 0.04

0.13

0.4

1.3

4

E:T ratio

Figure 2. T r release of peptide-labeled P815(HLA-B27) targets as a function of the E/T ratio for two patients. The peptides are identified by the code used in Fig. 1. Analyses of FOU and FOI at the other time points gave comparable results, i.e. the CTL always recognized all the peptides present at that time point.

from patient FOI. All gag peptides were tested on similarly constructed P815(HLA-A2) and P815(HLA-A3) target cells which consistently yielded background values (not shown).These controls show that the lysis was due to classic HLA-restricted CTL.

4 Discussion The critical residues within this particular epitope for efficient MHC class I binding and presentation are those KRWIILGNKIVRMYS _ _ _ ([13,24] and underlined, i.e. _ McMichael, personal communication).Almost none of the variant sequences identified above encoded substitutions within these residues. These data help explain the efficient recognition of peptides corresponding to minor forms - for example, peptide 33 which represented a mere 6% of proviral sequences (e.g. FOU, sample 21.05.90; Fig. 2). It must be noted that 2/17 peptide sequences described in this study, encoding either a conservative K+R or nonconservative G + R substitution in the 04 and FOU (7.11.89) quasispecies, respectively, might possibly escape correct HLA-B27 recognition andor presentation. This possibility is strongly suggested by the observation that both CTL fromFOI (29.03.89) andFOU (29.03.89) failed to lyse labeled P815(HLA-B27) targets incubated with a peptide containing a G + E substitution over a 100-fold range of peptide concentration and at an E R ratio of 2/1 (data not shown). However, in the case of FOU the proportion of viral genomes containing this G +R sequence was not found expanded in the subsequent quasispecies, i.e. was not detected at a frequency of > 1/20, the resolution of the analysis.These data would suggest that despite demonstrableCTL activity ex v i v o , time-dependent cytotoxic T cell-driven selection of variant HIV sequences was not occurring in v i v o , at least for this epitope. This is particularly striking given that the structures of the quasispecies fluctuated over the 1Cmonth period, itself evidence of ongoing viral replication and replacement of HIVinfected lymphocytesin peripheral blood.The single report of a viral CTL escape mutant, which concerned lymphochoriomeningitis virus infection of transgenic mice, has recently been described. However, this represents a somewhat special system [25]. In humans CTL apparently aid recovery from influenza virus [26] and control, but fail to eliminate Epstein-Barr

and cytomegalovirus infections [27,28]. The evidence for latent HIV-infected lymphocytes [29] and the sequestering of HIV in privileged sites such as the brain would strongly suggest that HIV-specific CTL cannot completely clear the virus. The turnover observed in the gag quasispecies could be taken as evidence of antigenic drift,the selection of viral variants perhaps being dominated by neutralizing antibodies [30]. The dominant antigen, and hence epitope, involved in HIV-specific CD8+ reactivity is unknown, if indeed a dominant antigen exists. Perhaps the CTL response to this gag epitope could dampen down virus production but not prevent it. It would be interesting to define in an analogous manner the fate of an epitope derived from an early (i.e. tat) as opposed to a late HIV protein (i.e. gag, pol or env) especially as protective immunity to cytomegalovirusinfection in mice is directed against an intermediate early antigen rather than structural proteins [31].While the data presented here caution against the extrapolation of in vitro data to the patient the strategy described above should prove useful in assessing the importance of CTL epitopes thus defined. We are greatly indebted to the four patients (FOU, FOI, DEM and 04) who participated in this stud3 particularly FOU and FOI who came back three times upon request. Drs. Martine Mariotti, Jacqueline Salpetrier and Antoine Toubert represented the essential bridges between the experimental and clinical parts inherent in such a project. We would like to thank Dr.Andrew McMichael for communicating unpublished data and for his comments on the manuscript. Received May 14, 1991; in final revised form July 3, 1991.

5 References 1 Leider, J. M., Palese, l? and Smith, F. I., J. Virol. 1988. 62: 3084. 2 Pathak,V K. and Temin, H., Proc. Natl. Acad. Sci. USA 1990. 87: 6019. 3 Meyerhans, A., Cheynier, R., Albert, J., Seth, M., Kwok, S., Sninsky, J., Morfeldt-Mbnson, L., Asjo, B. and Wain-Hobson, S., Cell 1989. 58: 901. 4 Kannagi, M., Masuda, T., Hattori, T., Kanoh, T., Nasu, K., Yamamoto, N. and Harada, S., J. Virol. 1990. 64: 3399.

2640

A. Meyerhans, G. Dadaglio, J.-PVartanian et al.

5 Walker, C. M., Moody, D. J.,Stites, D. F! and Levy, J. A., Science 1986.234: 1563. 6 Walker, C. M. and Levy, J. A., Immunology 1989. 66: 628. 7 Brinchmann, J. E., Gaudernack, G. and Vartdal, F., J. Irnmunol. 1990.144: 2961. 8 Hoffenbach, A., Langlade-Demoyen,I?, Dadaglio, G. ,Vilmer, E, Michel, F., Mayaud, C, Autran, B. and Plata, F., J. Immunol. 1989. 142: 452. 9 Mills, K. H. G., Nixon, D. F. andMcMichae1,A. J., AIDS 1989. 3: s101. 10 Walker, B. D. and Plata, F., AIDS 1990. 4: 177. 11 Claverie, J. M., Kourilsky, I?, Langlade-Demoyen, F!, Chalufour-Prochnicka, A., Dadaglio, G., Tekaia, F., Plata, F. and Bougueleret, L., Eur. J. Irnrnunol. 1988. 18: 1547. 12 Culmann, B., Gomard, E., Kieny, M. F!, Guy, B., Dreyfus, F., Saimot, A. G., Sereni, D. and Levy, J. F,! Eur. J. Irnmunol. 1989. 19: 2383. 13 Nixon, D. E,Townsend, A. R. M., Elvin, J. G., Rizza, C. R., Gallwey, J. and McMichael, A. J., Nature 1988. 336: 484. 14 Plata, F., Dadaglio, G., Chenciner, N., Hoffenbach, A., Wain-Hobson, S., Michel, F. and Langlade-Demoyen, I?, Irnrnumdef. Rev. 1989. I : 227. 15 Ou, C Y , Kwok, S., Mitchell, S.W., Mack, D. H., Sninsky,J. J., Krebs, J. W., Feorino, F!,Warfield, D. and Schochetman, G., Science 1988.239: 295. 16 Goodenow, M., Huet,T., Saurin,W., Kwok, S., Sninsky, J. and Wain-Hobson, S., AIDS 1989. 2: 344. 17 Frank, R. and Doring, R., Tetrahedron 1988. 44: 6031. 18 Langlade-Demoyen, F!, Michel, F., Hoffenbach, A., Vilmer, E., Dadaglio, G., Garcia-Pons, F., Mayaud, C., Autran, B., Wain-Hobson, S. and Plata, F., J. Immunol. 1988. 141: 1949.

Eur. J. Immunol. 1991. 21: 2637-2640 19 Healy, F.,Toubert, A., Gomard, E., Jordan, B. R. and Levy, J. F!, J. Immunol. 1989. 143: 2357. 20 Wain-Hobson, S., Sonigo, I?, Danos, O., Cole, S. and Alizon, M., Cell 1985. 40: 9. 21 Wain-Hobson, S.,Vartanian, J.-F!, Henry, M., Chenciner, N., Cheynier, R., Delassus, S., Pedroza Martins, L., Sala, M., Nugeyere, M.-T., GuCtard, D., Klatzmann, D., Gluckman, J.-C., Rozenbaum, W., BarrC-Sinoussi, F. and Montagnier, L., Science 1991. 252: 961. 22 Huet, T., Dazza, M. C., Brun-Vtkinet, E, Roelants, G. and Wain-Hobson, S., AIDS 1989. 3: 707. 23 Williams, F!, Simmonds, P.,Yap, F! L., Balfe, F!, Bishop, J., Brettle, R., Hague, R., Hargreaves, D., Inglis, J., Leigh Brown, A., Peutherer, J., Rebus, S. and Mok, J., AIDS 1990.4: 393. 24 Huet, S., Nixon, D. F., Rothbard, J. B.,Townsend, A., Ellis, S. A. and McMichael, A. J., Int. Irnmunol. 1990. 2: 311. 25 Pircher, H., Moskophidis, D., Roher, U, Burki, K., Hengartner, H. and Zinkernagel, R. M., Nature 1990. 346: 629. 26 McMichael, A. J., Gotch, F. M., Noble, G. R. andBeare, I? A. S., N. Engl. J. Med. 1983. 309: 13. 27 Borysiewicz,L. K., Moms, S., Page, J. D. and Sissons, J. G. I?, Eur. J. Irnrnunol. 1983. 13: 804. 28 Rickinson, A. B., Moss, D. J., Allen, D. J., Wallace, L. E., Rowe, M. and Epstein, M. A., Int. J. Cancer 1981. 27: 593. 29 Schnittman, S. M., Psallidopoulos, M. C.,Lane, H. C.,Thompson, L., Baseler, M., Massari, F., Fox, C. H., Salzman,N. F! and Fauci, A. S., Science 1989. 245: 305. 30 Albert, J., Abramhansson, B., Nagy, K., Aurelius, E., Gaines, H., Nystrom, G. and Fenyo, A. M., AIDS 1990. 4: 107. 31 Reddehase, M. J., Mutter, W., Munch, K., Buhring, H. J. and Koszinowski, U. H., J. Virol. 1987. 61: 3102.