Proc. Nail. Acad. Sci. USA Vol. 89, pp. 2223-2226, March 1992 Medical Sciences Replication of patient isolates of human immunodeficiency type 1 in T cells: A spectrum of rates and efficiencies of entry (slow/low patient isolate/rapid/high patient isolate/single cycle infections/leu3a entry block/dideoxycytidine reverse block) R. FERNANDEZ-LARSSON, K. K. SRIVASTAVA, S. Lu, AND H. L. ROBINSON* Department of Pathology, University of Massachusetts Medical Center, 55 Lake Avenue North, Worcester, MA 01655 Communicated by Sheldon Penman, December 16, 1991

ABSTRACT Isolates of human immunodeficiency virus entry depended on the presence of CD4, differences in entry type 1 (HIV-1) undergo many different rates of replication, did not correlate with surface levels of CD4 (6). with the time course ofreplication being determined by the host The current study was undertaken to determine whether cell and the virus. Recently, we demonstrated that the permis- virus as well as T-cell factors could affect HIV-1 entry. To siveness of four CD4' T-cell lines for the laboratory strain optimize our chance of detecting viral factors that might NL4-3 correlated with the rate and efficiency of virus entry. In affect entry, the life cycles of three patient isolates were this study, we have analyzed the replication of a "slow/low" compared with that of NL4-3. Two of the patient isolates, isolate from the pre-AIDS period of infection and two "rap- SF94 and SF216, represent serial isolates from a single id/high" isolates from the AIDS period of infection to deter- patient (2). SF94, from the pre-AIDS period of disease, has mine which steps in the virus life cycle determine differences in a limited ability to replicate on H9 cells (slow/low, class 3 the growth characteristics ofpatient isolates. Differences in the phenotype) (3). In contrast, SF216, from the AIDS period of growth of the patient isolates correlated with differences in disease, establishes productive infections on various cell entry but not postentry steps of the virus life cycle. The two lines (rapid/high phenotype). The third patient isolate, SF33, rapid/high patient isolates (SF33 and SF216) underwent 50% is a rapid/high virus with an unusually broad tissue tropism entry in s0.5 hr in C8166 cells, in '1 hr in mitogen-stimulated that includes an osteosarcoma cell line (8). peripheral blood mononuclear cells, and in .2.5 hr in H9 cells. In contrast, a class 3 slow/low patient isolate required 1 hr for MATERIALS AND 50% entry into C8166 cells, 3 hr for 50% entry into peripheral METHODS blood mononuclear cells, and 5-6 hr for 50% entry into H9 Cells. C8166, a human T-lymphotrophic virus type 1 im- cells. Entry efficiency correlated with entry rate, with the mortalized but nonexpressing line ofcord blood lymphocytes rapid/high having a 2-fold higher titer and the (9), and H9 cells, a subline ofHU178 cells (10), were obtained slow/low virus having a 5-fold higher titer on C8166 than H9 from the AIDS Repository, Rockville, MD. Peripheral blood cells. The laboratory strain NL4-3 displayed intermediate rates mononuclear (PBM) cells were cultured from seronegative and efficiencies of entry. These data demonstrate that entry donors according to the consensus protocol used by the characteristics are major determinants of the pathogenic po- National Institutes of Health-sponsored AIDS Clinical Trial tential of patient isolates. Units. Cells were cultured at densities of 1-2 x 106 per ml in RPMI 1640 medium supplemented with 10-15% fetal bovine serum. Human immunodeficiency virus type 1 (HIV-1) isolated from Viruses. NL4-3, a laboratory strain that has been con- patients during the long seropositive pre-AIDS period of structed to encode all known HIV-1 gene products (11), was disease has been reported to be slowly replicating, low- recovered from pNL4-3 DNA by DEAE-dextran transfection yielding viruses (slow/low viruses) that do not cause syncytia of H9 cells. Three patient isolates, SF33, SF94, and SF216, and do not grow well on cell lines. In contrast, isolates from were obtained from C. Cheng-Mayer and J. Levy (Uni ersity patients with fully developed AIDS tend to be rapidly repli- of California, San Francisco, CA) (2, 8). Stocks of tie SF cating, high-yielding viruses (rapid/high viruses) that cause viruses were produced during a single passage on H9 ells. syncytia and readily establish productive infections in T-cell Serial passage ofNL4-3 or SF94 on H9 cells resulted in sto as well as monocyte/macrophage cell lines (1-5). Laboratory with altered growth potentials (R.F.-L. and H.L.R., unpub- strains of HIV-1 also exhibit virus- and cell-line-specific lished observations). Viruses were titered for infectious units differences in growth potential and syncytium forming activ- on C8166 cells (12). ity. Single Cycle Infections. Infections were carried out at 370C Recently we demonstrated that differences in the ability of in the presence of 2 ,ug of Polybrene per ml at multiplicities four cell lines to support the replication of a laboratory strain of 0.2-0.8 C8166 infectious unit per cell (6). At the time of of HIV-1 (NL4-3) correlated with the rate and efficiency of virus addition, cultures were aliquoted in prewarmed 24-well virus entry (6). Virus entry was measured as the time required plates. The time course for entry was followed by timed for 50o ofthe infectious units to escape being blocked by the additions of leu3a, an anti-receptor monoclonal antibody anti-receptor antibody leu3a (7). Fifty percent entry occurred (Becton Dickinson) (7). The time course for reverse tran- within 30 min in infections of the highly permissive C8166 scription was followed using timed additions of dideoxycy- cells. In contrast, 50% entry required an average of 4 hr in tidine (ddC), a reverse transcriptase inhibitor (Sigma) (13). infections of three less permissive cell lines (H9, Jurkat, and Addition of leu3a (250 ng/ml for NL4-3, SF94, and SF216; A3.01). The rate of entry correlated with the efficiency of 750 ng/ml for SF33) or ddC (2 ,M) 10 min prior to the addition entry. Stocks ofNL4-3 had three times higher titers on C8166 of virus completely blocked each of the infections. cells than on the less permissive cells. Interestingly, although Abbreviations: HIV-1, human immunodeficiency virus type 1; PBM, The publication costs of this article were defrayed in part by page charge peripheral blood mononuclear; ddC, dideoxycytidine; IFA+, immu- payment. This article must therefore be hereby marked "advertisement" nofluorescent antibody positive. in accordance with 18 U.S.C. §1734 solely to indicate this fact. *To whom reprint requests should be addressed.

2223 Downloaded by guest on September 28, 2021 2224 Medical Sciences: Fernandez-Larsson et al. Proc. Natl. Acad. Sci. USA 89 (1992)

Escape from the leu3a and ddC blocks was scored by monitoring cultures for plateau levels of virus-expressing cells using indirect immunofluorescence (6, 12). The time course for virus expression was determined by washing an untreated aliquot at 12-15 hr after infection and growing the washed cells in the presence of leu3a and ddC to prevent the LJ spread of progeny virus. Aliquots of these cultures were analyzed at sequential times for virus-expressing cells using indirect immunofluorescence (6, 12) and for levels of viral using an antigen capture ELISA. The ELISA used c 99% pure immunoglobulins obtained from pooled patient 0 plasma for the capture and for detecting antibodies. These 6- . _ 5- immunoglobulins recognized several HIV-1 proteins, includ- L) ing p24, gp4l, and gpl20 (14). U) 4

en 3 a)L. 0 cr 2- RESULTS a) aj) To identify steps in the HIV-1 life cycle that determine 1 IR1E IRM0 IRE differences in the replication of patient isolates, single cycle analyses were conducted for two rapid/high patient isolates (SF33 and SF216), one slow/low patient isolate (SF94), and a laboratory strain (NL4-3). Growth studies were done on the 30: highly permissive C8166 cells, the less permissive H9 cells, 0 and mitogen-stimulated PBM cells. In these analyses, timed additions of an anti-receptor antibody (leu3a) and of an inhibitor of reverse transcriptase (ddC) were used to define ,°20 the times required for entry, reverse transcription, and ex- pression. An example of such a single cycle analysis is given for the H9 PBMC C8166 replication ofthe slow/low virus (SF94) in PBM cells (Fig. 1). The time course for escape from the leu3a block (0), the time FIG. 2. Hours required for 50% ofthe infectious units to undergo course for escape from the dideoxycytidine block (e), and the entry, reverse transcription, and expression. Data for each of the time course for virus expression (o) are indicated in Fig. 1. T-cell types were collected from parallel, simultaneous single cycle These three time courses allowed estimation of the times infections with the four tested viruses. For an example of a single required for 50%o of the infectious units to undergo entry cycle infection see Fig. 1. PBMC, PBM cells. (Top) Entry. Values represent the hours required for 50o ofthe infectious units to escape (escape the leu3a block), undergo reverse transcription (pass the leu3a block. (Middle) Reverse transcription. Values represent the from the leu3a block past the ddC and block), undergo hours required for 50% of the infectious units to pass from the leu3a expression (pass from the ddC block to virus-expressing block past the ddC block. (Bottom) Expression. Values represent the cells). hours required for 50% of the infectious units to pass from the ddC Data from 12 such single cycle infections are summarized block to the expression of viral antigens. in Fig. 2. These 12 infections followed the life cycles of the two rapid/high viruses, the slow/low virus, and NL4-3 in entry. In sharp contrast, little or no differences were ob- C8166 cells, H9 cells, and mitogen-stimulated PBM cells. The served in the times required for reverse transcription and data reveal striking differences in the times required for virus expression. Entry rate depended on the virus and the T cell (Fig. 2). For most in SF94 each of the viruses, the rapid entry occurred C8166 U) 115- cells, the intermediate rates of entry occurred in mitogen- stimulated PBM cells, and the slowest rates of entry took place in H9 cells. Entry rates within a cell type were < 1 10 UL- consistently 2- to 3-fold faster for the rapid/high viruses than the slow/low virus. The laboratory strain NL4-3 had an C-)- intermediate entry profile. To test whether virus concentra- 5 Q) a 0L tion would affect entry rates, two dilutions of SF216 stock were analyzed for entry. The 5-fold dilution in these samples o did not affect the "rapid/high" entry characteristics ofSF216 0 10 20 30 40 50 60 (R.F.-L. and H.L.R., unpublished observations). Entry rates did show some variation from experiment to experiment (6). Hours postinfection However, within each experiment, the relative differences FIG. 1. Single cycle infection of SF94 in PBM cells. Entry, for the different HIV-1 isolates were always the same. reverse transcription, and expression times are defined as the hours The efficiency ofthe entry of each virus correlated with its required for 50o of the infectious units to escape the leu3a block rate of entry (Fig. 3). When titers of virus were determined (entry), the hours for 50%1 ofthe infectious units to traverse from the from data obtained in independent single cycle infections, the leu3a block past the ddC block (reverse transcription), and the hours titers in H9 infections were consistently lower than those in for 50% of the infectious units to pass from the ddC block to scoring the highly permissive C8166 infections. To directly test for for viral in an indirect immunofluorescence assay antigens antibody relative titers on C8166 and virus were (expression). IFA+, immunofluorescent antibody positive. o, Time the H9 cells, stocks of leu3a addition; *, time of ddC addition; o, time for expression of titered on C8166 cells and then diluted to initiate H9 infec- viral antigens in 15-hr leu3a plus ddC point. The vertical dotted lines tions with 0.5 C8166 infectious unit per cell (a multiplicity of indicate how the times for entry, reverse transcription, and expres- infection that gives 50% infected C8166 cells). The leu3a sion were determined. entry block was added to cultures at various times after Downloaded by guest on September 28, 2021 Medical Sciences: Femandez-Larsson et al. Proc. Natl. Acad. Sci. USA 89 (1992) 2225

Under our conditions of single cycle infections, between 10% and 50%6 ofthe cells are infected. This means that the majority of virus-expressing cells are infected by a single infectious

CI, unit. Assuming a Poisson distribution, two active proviruses would be expected in 10%o of the cells in which 50% of a o20 culture is infected and in <1% ofthe cells in which 10% ofthe +A culture is infected. Thus, normalization of the level of virus proteins per infected cell provides an estimate of the relative -4-, levels of expression per infectious unit. When this was done for H9 infections, the two rapid/high viruses and NL4-3 were found to have similar levels of expression, which were ~3-fold higher than that of the slow/low virus (Fig. 4). In contrast, on PBM cells, the slow/low virus was expressed at levels similar to one of the rapid high viruses (SF33) and NL4-3. This level of expression was about 2-fold lower than

0 3 6 9 12 15 18 that of the other rapid/high virus (SF216) (Fig. 4). These results indicate that differences in the levels of virus expres- Time of Leu3a addition (hours) sion per infected cell do not necessarily correlate with the slow/low and rapid/high phenotype. FIG. 3. Rate and efficiency of entry of a constant number of C8166 infectious units into H9 cells. Each infection was initiated by incubating 5 x 105 C8166 infectious units of SF94, SF216, SF33, or DISCUSSION NL4-3 with 1 x 106 H9 cells. A, SF33; o, SF216; *, NL4-3; e, SF94; Spectrum of Rates and Efficiencies of HIV-1 Entry. Our -----, estimates of hours required for 50o entry. results clearly demonstrate that HIV-1 isolates have a whole infection and the cultures were scored for the percent in- spectrum of rates and efficiencies of entry. In our tests, 50%6 fected cells at 60 hr after infection. In each of the infections, entry required from <20 min to >5 hr, depending on the virus the level of that isolate and the host T cell (Fig. 2). Entry rates also affected plateau virus-expressing cells occurred when entry efficiencies (Fig. 3). For our slowest entry virus (SF94), all of the infectious units had passed the leu3a block was the titer of infectious units measured in the most permissive <50%o. Only 10% of the H9 cells became infected by the cell was five times higher than the titer of infectious units slow/low virus. Approximately 25% of the H9 cells became determined in the least permissive cell. infected by the two rapid/high viruses. These results indicate Steps that take place during HIV-1 entry require the viral that the slow/low virus underwent a 5-fold reduction of its envelope glycoproteins gpl20 and gp4l and the host cell C8166 titer on H9 cells, whereas the two rapid/high viruses receptor CD4. The first step in entry is the binding of virus underwent 2-fold reductions in their C8166 titers on H9 cells. to cells. This step is facilitated by gpl20 binding to CD4 (for Consistent with its intermediate entry rate, NL4-3 exhibited a recent article, see ref. 15). gpl20-gp41 oligomers are then an intermediate (3-fold) reduction of its C8166 titer on H9 thought to undergo post-binding conformational changes that cells (6). expose the fusion domain of gp4l (16, 17). These conforma- The levels of virus expression per infected cell were cell tional changes are pH independent (18), may involve shed- type specific and not a general predictor ofwhether an isolate ding of gpl20 (19-21), and may require proteolytic cleavage had slow/low or rapid/high growth characteristics (Fig. 4). of the third variable region of gpl20 (V3 loop) (22, 23). Once the conformational changes have taken place, virus enters the 50 - SF94 cytoplasm by fusion through the cytoplasmic or an endoso- TEC- SF216 mal membrane (24, 25). Our results indicate that virus and U) 40 1 SF33 (I host cell factors affect the rate and efficiency of these entry EM NL4-3 phenomena (Fig. 2). <-a In terms of viral factors, we think it likely that differences in the envelope glycoproteins of patient isolates will deter- LLJ + mine differences in entry characteristics. During HIV-1 in- at LL fections, envelope sequences undergo selection for neutral- _-* ization-resistant variants by the host's immune response (26, 27). Concomitant with this selection and the establishment of the long seropositive period of infection, viruses with slow/ low growth characteristics emerge as the predominant iso- Cell type: PBM PBM lates (1-5, 28, 29). This would suggest that selection of sequences for resistance may also select Sample: cells cells media envelope antibody for slow entry. Such would be consistent with the coselection FIG. 4. Production of virus-specific proteins per infected cell. of neutralization resistance and poor replication in a series of Ratios of virus proteins (ELISA units) per infected cell were obtained molecularly characterized envelope genes from an experi- for infected H9 cells, infected PBM cells, and the culture medium of mentally infected chimpanzee (27). infected PBM cells. The H9 data are for a single point taken at 60 hr In terms ofhost cell factors that might affect entry rates and after infection. Data for PBM cells are from three points taken at 46, efficiencies, differences in the permissiveness ofthe test cells 51, and 61 hr after infection. Data for these three points were did not correlate with differences in the surface densities of normalized by adding the values for all viruses at each point and the CD4 receptor (6). However, differences in susceptibility calculating the percentage of the total for each virus. Normalized to HIV-1-induced syncytium formation did correlate with data from the three points were then averaged to give the presented ratios. Error bars for these ratios are standard deviations. The overall rate and efficiency of virus entry. The entry-permissive levels of virus expression were similar in H9 and PBM cells, ranging C8166 cells are much more susceptible to syncytium forma- from 75 to 225 ELISA units per IFA' cell in the 60-hr H9 samples tion than either mitogen-stimulated PBM cells or H9 cells. and from 100 to 300 ELISA units per IFA' cell in the 51-hr samples The entry times given in Fig. 2 are distinctly longer than from PBM cells and culture medium. those that have been previously reported (24, 25). Our study Downloaded by guest on September 28, 2021 2226 Medical Sciences: Fernandez-Larsson et al. Proc. Natl. Acad. Sci. USA 89 (1992) differs from prior studies on entry in that (i) we did not chini, G., Kalyanaraman, V. S. & Gallo, R. C. (1983) Virology prebind virus and (ii) we followed the entry ofinfectious units 129, 51-64. (rather than physical particles) of virus. 10. Popovic, M., Sarngadharan, M. G., Read, E. & Gallo, R. C. Entry Characteristics and Pathogenic Potential. This study (1984) Science 224, 497-500. represents our first comparative study on the life cycles of 11. Adachi, A., Gendelman, H. E., Koenig, S., Folks, T., Willey, R., Rabson, A. & Martin, M. H. (1986) J. Virol. 59, 284-291. slow/low and rapid/high patient isolates. Unexpectedly, the 12. Robinson, H. L. & Zinkus, D. M. (1990) J. Virol. 64, 4836- data demonstrate that a class 3 slow/low virus differs from 4841. two rapid/high viruses in entry characteristics (Fig. 2). The 13. Mitsuya, H. & Broder, S. (1986) Proc. Natl. Acad. Sci. USA 83, data reveal no differences in the time courses for reverse 1911-1915. transcription and expression (Fig. 2). Data on the levels of 14. Cummins, L. M., Weinhold, K. J., Matthews, T. J., Langlois, expression per infected cell do reveal differences among the A. J., Perno, C. F., Condie, R. M. & Allain, J.-P. (1991) Blood viruses. However, these were cell type dependent and not a 77, 1111-1117. general correlate with the slow/low or rapid/high phenotype 15. Ryu, S.-E., Kwong, P. D., Truneh, A., Porter, T. G., Arthos, (Fig. 4). J., Rosenberg, M., Dai, X., Xuong, N.-H., Axel, R., Sweet, The discovery of regulatory genes in HIV-1 raised high R. W. & Hendrickson, W. A. (1990) Nature (London) 348, interest in the possibility that such would determine differ- 419-425. ences in the growth potentials and tissue tropism of patient 16. Gallagher, W. R. (1987) Cell 50, 327-328. 17. Kowalski, M. L., Potz, J., Basiripour, L., Dorfman, T., Goh, isolates. Interestingly, our analyses on the growth of patient W. C., Terwilliger, E., Dayton, A., Rosen, C., Haseltine, W. isolates in T cells and the work of others on tissue tropism & Sodroski, J. (1987) Science 237, 1351-1355. have demonstrated viral-encoded entry, rather than viral- 18. McClure, M. O., Marsh, M. & Weiss, R. A. (1988) EMBO J. 7, encoded regulatory phenomena, determining differences in 513-518. pathogenic potential (Fig. 2; refs. 30-34). However, our 19. Kirsh, R., Hart, T. K., Ellens, H., Miller, J., Petteway, S. A., studies, as well as those on tropism, have involved only a few Jr., Lambert, D. M., Leary, J. & Bugelski, P. J. (1990) AIDS patient isolates, with the isolates chosen for study represent- Res. Hum. Retroviruses 6, 1209-1212. ing those that could be grown in mitogen-stimulated PBM 20. McKeating, J. A., McKnight, A. & Moore, J. P. (1991)J. Virol. cells or monocyte/macrophages. Thus, these studies do not 65, 852-860. and in 21. Hart, T. K., Kirsh, R., Ellens, H., Sweet, R. W., Lambert, preclude a role for regulatory genes proviral latency D. M., Petteway, S. R., Jr., Leary, J. & Bugelski, P. J. (1991) pathogenic potential. Rather, they highlight the importance Proc. Natl. Acad. Sci. USA 88, 2189-2193. of the entry step of the HIV-1 life cycle in the determination 22. Hattori, T., Koito, A., Takatsuki, K., Kido, H. & Katunuma, of pathogenic potential. N. (1989) FEBS Lett. 248, 48-52. 23. Clements, G. J., Price-Jones, M. J., Stephens, P. E., Sutton, We are indebted to Drs. C. Cheng-Mayer and J. A. Levy for C., Schulz, T. F., Clapham, P. R., McKeating, J. A., McClure, providing the SF viruses. This research was supported by U.S. M. O., Thomson, S., Marsh, M., Kay, J., Weiss, R. A. & Public Health Service Research Grant RO1 Al 24474. R.F.-L. was Moore, J. P. (1991) AIDS Res. Hum. Retroviruses 7, 3-16. supported by U.S. Public Health Service Training Grant T32 AI 24. Pauza, C. D. & Price, T. M. (1988) J. Cell Biol. 107, 959-968. 07349. 25. Grewe, C., Beck, A. & Gelderblom, H. R. (1990) J. Acquired Immune Defic. Synd. 10, 965-974. 1. Asjd, B., Morfeldt-Minson, L., Albert, J., Biberfeld, G., 26. Albert, J., Abrahamsson, B., Nagy, K., Aurelius, E., Gaines, Karlsson, A., Lidman, K. & Fenyo, E. M. (1986) Lancet n, H., Nystrom, G. & Fenyo, E. M. (1990) AIDS 4, 107-112. 660-662. 27. Nara, P. L., Smit, L., Hatch, D. W., Merges, M., Waters, D., 2. Cheng-Mayer, C., Seto, D., Tateno, M. & Levy, J. A. (1988) Kelliher, J., Gallo, R. C., Fischinger, P. J. & Goudsmit, J. Science 240, 80-82. (1990) J. Virol. 64, 3779-3791. 3. Fenyo, E. M., Morfeldt-Minson, L., Chiodi, F., Lind, B., von 28. Clark, S. J., Saag, M. S., Decker, W. D., Campbell-Hill, S., Gegerfelt, A., Albert, J., Olausson, E. & Asjo, B. (1988) J. Roberson, J. L., Veldkamp, P. J., Kappes, J. C., Hahn, B. H. Virol. 62, 4414-4419. & Shaw, G. M. (1990) N. Engl. J. Med. 324, 954-960. 4. Tersmette, M., Gruters, R. A., de Wolf, F., de Goede, 29. Daar, E. S., Moudgil, T., Meyer, R. D. & Ho, D. D. (1990) N. R. E. Y., Lange, J. M. A., Schellekens, P. T. A., Goudsmit, Engl. J. Med. 324, 961-964. J., Huisman, H. G. & Miedema, F. (1989) J. Virol. 63, 2118- 30. O'Brien, W. A., Koyanagi, Y., Namazie, A., Zhao, J.-Q., 2125. Diagne, A., Idler, K., Zack, J. A. & Chen, I. S. Y. (1990) 5. Fenyo, E. M., Albert, J. & Asjo, B. (1989) AIDS 3, Suppl. 1, Nature (London) 348, 69-73. S5-S12. 31. Liu, Z.-Q., Wood, C., Levy, J. A. & Cheng-Mayer, C. (1990) 6. Srivastava, K. K., Fernandez-Larsson, R., Zinkus, D. M. & J. Virol. 64, 6148-6153. Robinson, H. L. (1991) J. Virol. 65, 3900-3902. 32. Shioda, T., Levy, J. A. & Cheng-Mayer, C. (1991) Nature 7. Sattentau, Q. J., Dalgleish, A. G., Weiss, R. A. & Beverley, (London) 349, 167-169. P. C. L. (1986) Science 234, 1120-1123. 33. Westervelt, P., Gendelman, H. E. & Ratner, L. (1991) Proc. 8. York-Higgins, D., Cheng-Mayer, C., Bauer, D., Levy, J. A. & Natl. Acad. Sci. USA 88, 3097-3101. Dina, D. (1990) J. Virol. 64, 4016-4020. 34. Hwang, S. S., Boyle, T. J., Lyerly, H. K. & Cullen, B. R. 9. Salahuddin, S. Z., Markham, P. D., Wong-Staal, F., Fran- (1991) Science 253, 71-74. Downloaded by guest on September 28, 2021