Understanding the HIV-1 induced pathogenesis remains a fundamental problem in AIDS research and antiviral therapy. Two principal mechanisms are considered to be involved; first, direct or indirect killing of immune cells and second a general dysregulation of these cells. Based on in vitro and in vivo experiments we provide evidence that the TM protein gp41 of HIV-1 may be involved in dysregulation of the immune cells, partially by the modulation of the expression of cytokines associated with immunosuppression.
For the first time glycosylated wt and mutated gp41 were produced in human cells. The proteins were recovered from the cell supernatant, characterised and analysed for their immunosuppressive properties. Furthermore, two infectious and replication competent variants of HIV-1 with mutations in the isu domain of gp41 were obtained and examined for their influence on cytokine expression. The systematic and comparative analyses of the properties of the isu domain of gp41 of HIV-1 included (i) glycosylated gp41 released from transfected human cells, (ii) gp41 in the envelope of viral particles and (iii) homopolymers of the isu peptide. All three forms (the gp41 in nanogram and the isu peptide in microgram amounts) induced a significant increase in IL-10 and IL-6 expression. Key residues in the isu domain required for the biological (immunosuppressive) activity were identified. Based on this data we speculate that the putative binding site of the isu domain involved in the induction of IL-10 release is discontinuous (the first 4 amino acids as the first part and position 9 to 14 as the second part). It was shown that immunisation with gp41 mutated in the isu domain resulted in a higher antibody response. Finally, we demonstrated that a single mutation (Q2A) in the isu domain of gp41 of replication-competent HIV-1 completely abrogated the immunosuppressive effect. All together these results indicate that the isu domain is the biologically active domain of gp41 of HIV-1 and the protein may directly interfere with the immune system. These results also shed light on the mechanism of HIV-1 pathogenesis and should be considered when designing gp41-directed vaccines.
An increase in IL-10 and IL-6 release from PBMCs as observed in our experiments was found in the blood of HIV-1 infected individuals [25–29]. Although considered as an immunosuppressive molecule, IL-10 has pleiotropic properties and modulates the function of several adaptive immunity-related cells and has a stimulatory effect on B cells . Noteworthy, a modulation of cytokine release was also shown in vitro for the recombinant TM proteins and the isu peptide of gammaretroviruses [17, 30] as well as for the human endogenous retrovirus HERV-K (unpublished). Thus, the immunosuppressive effect of the TM proteins seems to be a common property of retroviruses. In this regard it would be of interest to compare the immunosuppressive potential of the TM protein of other retroviruses with that of the gp41 of HIV-1 and the gp36 of HIV-2.
The mechanism of the cytokine modulation induced by gp41 of HIV-1 remains largely unknown. Although several laboratories reported binding of gp41 to the cell surface [31–34], it remains unclear whether there is a receptor(s); and if there is a receptor, how gp41 triggers signal transduction. The following mechanisms may be proposed: (i) The isu domain may interact specifically with a receptor(s) on the cell surface triggering release of cytokines and other factors, which then induce additional changes in gene expression and cytokine release. (ii) The isu domain of gp41 may bind non-specifically to different receptors, e.g., single-spanning transmembrane receptors, for example Toll-like receptors (TLR), inducing dimerization and triggering expression of numerous cytokines, or (iii) it can be a reverse agonist for the receptor(s). In fact, the observed differences in the modulation of cytokine expression in donor 4 (mutations in position L4 and D14 abrogated IL-10 release, but not IL-6 release) may be explained by some differences in the ligand binding site of the receptor(s). In addition a sequence homology between domains in the type I interferons (IFN) and the isu domain of gammaretroviruses and HIV-1 was reported [8, 35], and all type I IFNs although different in their sequence bind to a common receptor .
Recently a highly conserved region (SWSNKS) in the C-terminal helical region of gp41 was described, which binds to a specific receptor, the globular complement 1q receptor (gC1qR), and induces an increased expression of NKp44L, a cellular ligand for an activating NK receptor on CD4+ cells. This may be one reason for the elimination of CD4+ cells by NK cells and indicates another mechanism of immunosuppression induced directly by the gp41 protein .
The question whether the amount of gp41 in an infected individual is high enough to induce immunosuppression in vivo is related to the question where gp41 can be found in the organism. The gp41 with an accessible isu domain may be found on the surface of virus particles after shedding of gp120, in immune complexes, and in debris from dead cells, but the main amount will be found on the surface of infected and virus producing cells in the blood and in lymphoid organs.
Apparently the amount of gp41 in the body of an infected individual is impossible to calculate. Based on our in vitro
data we attempted to estimate the possible amount of gp41 in the blood of an infected individual. The amount of gp41 in the HIV-1pNL4-3
virus preparations was quantified by SDS-PAGE/Western blot analysis using serial dilutions of T20 as a reference (Figure 5
E). In our experiment 12.5 ng of gp41 were found in a preparation containing 1x106
infectious virus particles. Thus, one infectious particle was “associated” with 12.5 fg of gp41 (coefficient of association). It is important to underline that in addition to gp41 present in the infectious particles, the preparation contained gp41 from a large number of non-infectious particles. Based on the coefficient of association the following equation was developed:
where X is the amount of gp41 in the blood; C is the coefficient of association; 10n is the infectious titre/ml; and V is the volume of the blood in ml. For example, if the infectious titre in an infected individual is 1 × 104/ml, then 5 × 107 infectious virus particles associated with 0.625 μg (0.9 × 1013 molecules) of gp41 are present in the blood. Also assuming that 1 × 1010 lymphocytes are circulating in the blood, then 9 × 102 molecules of gp41 may interfere with one cell. In our in vitro experiments (i) 2 × 106 molecules of wt gp41 produced in human cells, and (ii) 1 × 105 molecules of virus associated gp41 per cell were able to induce changes in the cytokine release. Therefore, the amount of wt gp41 per cell that was used in vitro was nearly 2000 times higher and the amount of wt gp41 per cell that was about 100 times higher compared to estimated amount of virus-associated gp41 in the blood of an individual with an infectious titre of 1 × 104/ml. This calculation did not take into account the substantial amount of gp41 on the surface of blood cells, in cell debris and immune complexes. Thus, counting only the virus-associated gp41, an immunosuppressive effect in the blood of the patient comparable to that seen in vitro could be achieved if the infectious titre is above 1 × 106/ml.
However, the lymphoid organs have been proposed as the major reservoir of HIV-1 . Numerous molecules of gp41 expressed on the membranes of infected cells in lymphoid organs can actively interact with neighbouring uninfected cells and modulate the cytokine release in a long lasting manner. Thus, taking this point into consideration, the gp41 induced immunosuppression might be significant when the load of infectious virus is much lower than 1 × 106/ml. Since the conformation of gp41 on the cell surface and on virus particles might be different (full size functional gp41-gp120, and thermodynamically stable gp41 stumps, both as trimers or monomers ), it still remains to be determined which conformation modulates cytokine release best.
The viruses with mutations in the isu domain were infectious (Figure 5D), and therefore the mutation did not abolish virus replication. In this regard, the fact that certain mutations such as Q2A in the isu domain abrogating the cytokine modulating activity were not found in HIV sequences from infected individuals indicates that non-immunosuppressive viruses might be present only as minor quasispecies and could not represent a majority; otherwise they would be cleared by the immune system. Recently, using a classical laboratory model, a murine retrovirus (MuLV), it was shown that mutations in the isu domain abrogated its ability to suppress the innate and adaptive immune response (NK and CD8+ cells) of the immunocompetent host . This mutation also did not influence the replication capacity.
There are still open questions. For example, the virus load in African green monkeys infected with the simian immunodeficiency virus (SIVagm) and in sooty mangabeys infected with SIVsm is as high as in HIV-1-infected individuals . Despite the presence of an isu domain in their TM protein, these SIV do not induce AIDS in their natural host. However, a trans-species transmission of SIVsm to rhesus monkeys resulted in AIDS in the new h-ost, suggesting that the isu domain is functional and that the natural host adapted to it, either by lack of or structural changes in the putative receptor or by disruption of the signal transduction pathway.
Our studies showed that immunisation with gp41 with point mutations in the isu domain induced a better immune response in 4 of 5 rats compared to immunisation with the wt gp41 (Figure 6). It may be possible that conformational changes in the mutated isu domain promote a better immune response, but this was not supported by protein conformation prediction analysis. Similar improved immune responses were reported in mice immunised with the TM protein of the Friend-MuLV mutated in the isu domain in comparison with the wild-type protein; or with the recombinant human syncytin 1 (not immunosuppressive) in comparison with syncytin 2 (immunosuppressive) [10, 11].
Viruses developed numerous, often multiple mechanisms allowing suppression of the innate and adaptive immunity [3–7, 42]. HIV-1 is not an exception and exploits several strategies to overcome immune responses, for example, down-regulating CD4 and MHC-1 by Nef . In addition, infection of cells by HIV-1 endocytosis may be regarded as a strategy to minimise the contact of the virus with the immune system . Both Tat and gp120 contribute to the deregulation of the immune system, and therefore may be involved in immunosuppression leading to AIDS [45, 46]. In this context, it is of interest that the surface envelope protein gp105 of HIV-2 induced a stronger inhibition of T cell proliferation than HIV-1 gp120, in spite of the lower HIV-2 pathogenicity in vivo, indicating that gp120 is not the key molecule in inducing immunosuppression . In in vitro studies it was shown that gp120 from HIV-1JR-FL and HIVLAI induced IL-10 expression in human monocyte-derived dendritic cells via a mannose C-type lectin receptor, but gp120 from HIVKNH1144 did not . Since gp41 is glycosylated, it may interact with such receptors; however the results with the isu peptide homopolymer and the mutations in the isu domain which do not alter glycosylation indicate that glycosylation of gp41 is not involved in cytokine modulation.
It is important to underline that gp41 and gp120 are the first viral proteins that interfere with the immune system during the initial step of infection. Nef and Tat are produced only after infection, and may contribute to the immunosuppression later. In addition, the envelope proteins (especially gp41 anchored in the cell membrane) are expressed on the surface of infected cells and may permanently interact with the immune system.