Identification of RANTES, MIP-1α, and MIP-1β as the Major HIV-Suppressive Factors Produced by CD8+ T Cells

Evidence suggests that CD8+ T lymphocytes are involved in the control of human immunodeficiency virus (HIV) infection in vivo, either by cytolytic mechanisms or by the release of HIV-suppressive factors (HIV-SF). The chemokines RANTES, MIP-1α, and MIP-1β were identified as the major HIV-SF produced by CD8+ T cells. Two active proteins purified from the culture supernatant of an immortalized CD8+ T cell clone revealed sequence identity with human RANTES and MIP-1α. RANTES, MIP-1α, and MIP-1β were released by both immortalized and primary CD8+ T cells. HIV-SF activity produced by these cells was completely blocked by a combination of neutralizing antibodies against RANTES, MIP-1α, and MIP-1β. Recombinant human RANTES, MIP-1α, and MIP-1β induced a dose-dependent inhibition of different strains of HIV-1, HIV-2, and simian immunodeficiency virus (SIV). These data may have relevance for the prevention and therapy of AIDS.

suppressing the response pathways even 19. Total yeast protein was extracted, separated by when wild-type receptors were saturated SDS-polyacrylamide gel electrophoresis, and ETR1 31 July 1995; accepted 30 October 1995 with ethylene. If true, the latter model must be reconciled with the finding that Never-ripe (Nr), a mutation in tomato that Identification of RANTES, MIP-1 x, and MIP-1 , as seems to be attributable to a lesion in a proteins purified from the culture supernatant of an immortalized CD8+ T cell clone Kende, ibid. 241, 1086Kende, ibid. 241, (1988. and therapy of AIDS. 26, 71 (1992); J. S. Parkinson, Cell 73, 857 (1993). 9. ETR1 was expressed in the yeast Saccharomyces cerevisiae (strain LRB520) with the use of the vector pYcDE-2 with a constitutive ADC1 promoter. Constructs for full-length, single site-mutated, and trun-As documented in several viral diseases in parallel with the increasing deterioration cated forms of ETR1 were made as described (7). + For expression of the NH2-terminal region of ETR1 (1),CD8 T lymphocytes are believed to of the immune system (6).
representing amino acids 1 to 165, a 0.65-kb Eco play a critical role in the containment of We established a sensitive test system for RI-Dra fragment of one of our complementary DNA HIV infection, particularly during the phase HIV-SF (7), based on a CD4+ T cell clone clones (cETR1 -9) was cloned into the Eco RI site of pYcDE-2, after addition of a 1 0-mer Eco RI linker to of clinical latency and in long-term nonpro-(PMI) that has a broad susceptibility to the Dra site. gressors (2). Activated CD8+ T cells de-macrophage-tropic and primary HIV-1 iso-10. E. C. Sisler, Plant Physiol. 64, 538 (1979). [14C]ethrived from the peripheral blood of HIVlates (8). The HIV-SF was tested on PM1 ylene (specific activity = 56.9 mCi/mmol) was obinfected individuals (3), as well as from cells acutely infected with HIV-1 tained from American Radiolabeled Chemicals and BaL (9), a trapped as the mercuric perchlorate complex. Yeast HIVor SIV-infected nonhuman primates macrophage-tropic isolate with biological cells were grown to mid-log phase at 30°C, harvest-(4), secrete one or more soluble HIV-supproperties resembling those of non-syncyed by centrifugation at 15OOg for 5 min, washed with pressive factors (HIV-SF) that may contribtia-inducing primary isolates (8,9). To water, and collected by vacuum filtration on glass pre IV th ontibia-nding primar solate (8 9). To fiber disks (1 g of yeast per disk). Samples were ute to the control of HIV infection in vivo identify a reproducible source of HIV-SF, incubated in sealed glass chambers (Ball jelly jars) for (5). The production of HIV-SF by CD8+ T we tested (i) three CD8+ T cell lines im-4 hours in the presence of[14Clethylene with or withcells isolated in vitro correlates with the mortalized in vitro with human T cell leu- peusrcls(1l) ycnrs l he tain an optimized source of HIV-SF, the with three different portions of human HIV-SF activity. Nonimmune goat IgG, two best-producer CD8+ T cell lines (67-I RANTES, an 8-kD polypeptide belonging used as a control, had no effect. When the and CD8-UI) were cloned by the limiting to the C-C or ,B-chemokine subfamily (15). NAb were used alone, only anti-RANTES dilution technique, at 0.5 cells per well. The sequences of tryptic peptides from pro-demonstrated a partial blocking activity. Heterogeneity was observed among the tein 2 (14) matched those of two different However, the combination of NAb against clones (Table 1); one of the two clones with portions of human MIP-loa, another 8-kD all three chemokines totally abrogated the the highest HIV-SF activity (FC36.22) was C-C chemokine that is highly related to HIV-suppressive effect of the FC36.22 culselected for subsequent studies. RANTES (15). ture supematant ( Fig. 2A). Thus, the HIV-Cell-free culture supernatant from High concentrations of RANTES (139 SF activity of clone FC36.22 cannot be as-FC36.22 was filtered through a 0.22-iim to 1624 ng/ml) and MIP-la (112 to 616 cribed to the effect of a single chemokine, membrane and fractionated through a tan-ng/ml) were detected by specific enzyme but rather to their combined action. gential flow filter and centrifugal concen-immunoassay (EIA) in the culture supema-Moderate to high levels of RANTES (6 trators (12). The concentrate displaying tants of all CD8+ clones tested and their to 95 ng/ml), MIP-la (28 to 255 ng/ml), HIV-suppressive activity in the PM1/HIV-parental cell line (Table 1). In contrast, the and MIP-1l1 (37 to 191 ng/ml) were detect-1BaL test was further fractionated by weak only cytokine produced at high levels by ed in culture supematants from activated anion-exchange high-performance liquid the MT-2 cell line was MIP-la (340 ng/ CD8+T cells of HIV-infected patients (Tachromatography (HPLC). Fractions con-ml), as recently reported also for MT-4 ble 1) (18). Lower amounts of all three taining high levels of HIV-SF were pooled (16), another HTLV-I-immortalized CD4+ chemokines were produced by unfractionand further purified by reversed-phase T cell line. These results confirmed a pre-ated peripheral blood mononuclear cells HPLC (12). Potent HIV-SF activity, in the vious report demonstrating CD8+45RO+T (PBMC), stimulated in the same fashion, absence of significant cytotoxic effects, was cells to be the most potent producers of from a normal donor. No significant levels recovered in two separate fractions, each RANTES among blood leukocytes (17). of chemokines were produced by unstimucontaining a single major protein peak (Fig. Because of the close similarity to MIP-loa, 1). The persistence of the HIV-suppressive we also tested the production of MIP-113, a activity after reversed-phase HPLC at pH third member of the C-C chemokine sub-A 0.5 2.0 was consistent with previous reports family (15). Moderate to high levels of that HIV-SF is acid-stable (5). Proteins 1 MIP-lp were detected in all the culture 100l and 2 were each subjected to proteolytic supernatants tested (9 to 112 ng/ml), except 90 -0.4 digestion, followed by sequencing of disin MT-2 ( lated cells from two of the patients. duced in Escherichia coli. A dose-depenin primary PBMC that had been activated HIV-SF derived from CD8+ T cells of dent inhibition of the extracellular release in vitro with phytohemagglutinin (PHA) four patients tested (CD8.Ptl, 2, 3, and 5) of HIV-1 p24 antigen was observed in the (21). Two primary isolates (HIV-1573 and induced a dramatic inhibition of HIV in-PMI/HIV-lB.,L system with rhRANTES, HIV-1BaL), never previously passaged in fection (Fig. 2B). Pretreatment with the rhMIP-loc, or rhMIP-11, but not with continuous cell lines, and two laboratory anti-RANTES NAb alone had only a lim-rhMCP-1, a related C-C chemokine (Fig. isolates (HIV-111IBand HIV-lMN), grown in ited blocking effect (less than 20%) on 3). RANTES was the most effective inhibthe CD4+ T cell line H9, were tested. CD8.Ptl and virtually no effect on itor, with the dose inducing -95% sup-Treatment with rhRANTES, rhMIP-lx, or CD8.Pt3. Similarly, anti-MIP-1co or antipression of HIV p24 release (ED95) be-rhMIP-13 induced a dose-dependent inhi-MIP-13 NAb alone had no effects. Howtween 3.12 and 6.25 ng/ml; MIP-lo and bition of infection by HIV-1573, HIV-lBaLl ever, in three of the four patients tested, MIP-13 showed ED95 values of 12.5 to and HIV-IMN, albeit with different ED95 the combination of NAb to all three che-25.0 and 6.25 to 12.5 ng/ml, respectively. values, whereas HIV-lIIIB was virtually inmokines completely blocked the HIV-SF The cellular viability was markedly higher sensitive (Fig. 4A). Similar results were obactivity; in the fourth case (CD8.Pt2), the in cultures treated with effective doses of tained with culture supernatant of clone activity was blocked by more than 80%. the chemokines, compared to untreated FC36.22 (19). The ED95 of the three che-Thus, the HIV-SF activity produced by in controls (19). Consistent with previous mokines in primary PBMC was higher than vitro-activated CD8+ T cells of HIV-in-observations with crude HIV-SF (5), the in the PM1/HIV-lBaL system. This phefected patients is mostly, if not exclusiveexpression of HIV-1 RNA was totally sup-nomenon was not dependent on the viral ly, a result of the combined effects of pressed by treatment of infected PM1 cells strain used. RANTES, MIP-lo, and MIP-11. with either RANTES (20 ng/ml), MIP-lo The rhRANTES, rhMIP-Io, and rhMIP-The HIV-suppressive effect of C-C (100 ng/ml), or MIP-13 (50 ng/ml) (20). 13 induced a dose-dependent inhibition of chemokines was further investigated with We then analyzed the effects of the cheinfection by two HIV-2 and two SIV isolates recombinant human (rh) proteins, pro-mokines on different HIV-1 strains grown in primary human PBMC (Fig. 4B)

11
[3H]thymidine incorporation between 87 rum (normal GS) was used = and 105% of the control. None of the cheas a control at 200 p.g/ml.°8 0 ! mokines, at the doses used, had stimulatory Subsequently, both treated a effects on resting cells. and untreated superna-= 60 This study demonstrates that chemotants were added to infect- against HIV-lBaL was between 3.12 and the evaluation of inhibitory factors and are infection. Clinical studies on the produc-6.25 ng/ml in PM1 and between 50 and 200 of unlikely physiological relevance. Moretion of RANTES, MIP-lot, and MIP-lf3 in ng/ml in PBMC. However, many of the over, the documented propensity of C-C vivo will be critical to define their role in biological activities of RANTES (in terms chemokines to form large aggregates at the natural history of HIV infection. In of Ca 21 influx, chemotactic responses, ba-physiological pH may significantly reduce particular, it will be important to detersophil and eosinophil activation, and T cell the bio-available levels of these peptides, mine whether high levels of these chemosignaling) have been found between 40 and complicating the interpretation of efficacy kines are associated with a delayed pro-8000 ng/ml (26). Similarly, several physiostudies (15).
gression of HIV disease. Chemokine levels logical effects of MIP-loa and MIP-l 1, in-Chemokines are actively produced at may also provide a reliable correlate of cluding Ca 2+ influx, chemotaxis, enhancsites of inflammatory processes and exert protection in monkeys treated with expering and suppressive effects on bone marrow proinflammatory effects (15). Chemoimental vaccines (28). In addition, it is progenitor cells, and eosinophil activation, kine-mediated control of HIV may occur possible that some of the known difficuloccur in the dose-range between 10 and either directly, through their inherent an-ties with neutralizing antibody assays for 1000 ng/ml (26,27). The experimental ti-lentiretroviral activity, or indirectly, HIV are due to the variable amounts of conditions used in our PBMC tests (syn-through their ability to chemoattract T these chemokines in different human sera. chronous polyclonal activation with a cells and monocytes in proximity of the Finally, the identification of the HIVstrong plant-derived mitogen and acute in-infection foci. However, this latter mechsuppressive chemokines may open new fection with exogenous HIV at relatively anism may also have the opposite effect of perspectives for the development of effechigh multiplicity) are the most difficult for providing new, uninfected targets for HIV tive therapeutic approaches to AIDS.
Note added in proof:-We recently obtained the amino acid sequence analysis of Semin. Immunol. 5, 203 (1993