The transmembrane proteins contribute to immunodeficiencies induced by HIV-1 and other retroviruses
Download the PDF here
AIDS 15 May 2014
Robert Koch Institute, Berlin, Germany.
"To summarize, in-vitro and in-vivo evidence showed that the isu domain of retroviral transmembrane envelope proteins, including that of HIV-1, influences cytokine release, gene expression and immune response, and may therefore contribute to immunosuppression in the infected host. The possible functions of the transmembrane envelope protein at different time points after infection are shown in Fig. 4. As regards HIV-1, the contribution of the transmembrane envelope protein is difficult to evaluate due to the T-cell depletion, the activity of the accessory proteins and hyperactivation. Although the immune system of the host develops a strong response against the virus and the translocated microorganisms, the permanent suppression by the transmembrane envelope protein leads to retardation of the immune response and to immunodeficiency."
Many microorganisms including retroviruses suppress the immune system of the infected host in order to maintain infection. Unfortunately, it is still unclear how retroviruses induce immunosuppression. There is increasing evidence of a common mechanism based on their transmembrane envelope proteins. This review therefore summarizes evidence of the involvement of the transmembrane envelope proteins in the immunopathogenesis of different retroviruses including HIV-1. Mutations in the immunosuppressive (isu) domain of the transmembrane envelope protein of several retroviruses abrogate the immunosuppressive activities in vitro and in vivo. Most importantly, virus sequences with such abrogating mutations were never found in HIV-1-infected individuals despite the fact that the mutated viruses are replication-competent. However, there is also evidence for additional, perhaps even divergent, strategies for each retrovirus. For example, in contrast to many other retroviruses, the HIV directly interacts with immune cells and infects them. In addition, HIV uses several accessory proteins to evade the immune response. Furthermore, the possible contribution of the transmembrane envelope proteins of endogenous retroviruses to immunosuppression when expressed on tumor cells or in the placenta is analyzed.
Viral infections and immunosuppression
Inhibition of the immune system is common for infections with different microorganisms including viruses (for review see [1,2]). Nonretroviruses such as herpes viruses [3,4], measles virus  and other paramyxo viruses , as well as human papilloma viruses , modulate innate and acquired immunity using different strategies. Some viruses even use multiple strategies, including viral inhibition of major histocompatibility complex class I [8,9] and class II-restricted antigen presentation [10-12], viral inhibition of natural killer (NK) cell lysis , viral interference with apoptosis [14,15], as well as viral evasion of humoral immunity . Targeting cytokine function is a common strategy [1,17-24]. Some viruses interrupt the cytokine production ; others encode proteins acting as decoys to inhibit interaction of the cellular cytokine with its receptor . Last but not the least, some viruses express proteins acting like a cellular cytokine: Eppstein-Barr virus (EBV) and the equine herpes virus-2 encode viral analogs to the cellular interleukin (IL)-10 [18,19], human herpes virus (HHV)-8 encodes viral macrophage inflammatory proteins (MIP)-I, MIP-II and MIP-III [86-88], HHV-6 encodes IL-6 and a chemokine , and herpes simplex virus encodes IL-17 . Most interestingly, viral structural proteins such as the glycoproteins of the measles virus interact with the toll-like receptor 2 (TLR2) and other cellular receptors to induce a severe immunosuppression in the infected individuals .
Many retroviruses induce immunodeficiencies
Infections with retroviruses are commonly associated with immunodeficiencies . This was shown for most of the genera of the family Retroviridae, including alpharetroviruses such as avian leucosis viruses ; betaretroviruses such as simian retroviruses (SRV)-1, SRV-2 and SRV-3 ; gammaretroviruses such as the murine leukemia virus (MuLV) ; and the feline leukemia virus (FeLV) , deltaretroviruses such as the human T-lymphotropic virus HTLV-1 [30,31], as well as lentiviruses such as HIV  and the simian immunodeficiency viruses (SIV) . First reports on the immunosuppressive property of the MuLV were published in 1960  and since then, the immunosuppressive properties of gammaretroviruses have been carefully analyzed (for review see ). These viruses inhibit humoral and cellular responses in the infected animals, and immunosuppression was generalized, that is, not limited to the response against the virus. Although named leukemia virus, FeLV induces fatal opportunistic infections in approximately 60% of the infected cats as a result of the virus-induced immunosuppression, whereas only 5-10% of the animals suffer from leukemia . Also, the recently detected koala retrovirus (KoRV) induces lymphomas and an immunodeficiency associated with severe opportunistic infections, mainly chlamydia infection [35,36]. The ability to induce immunodeficiencies by nearly all retroviruses raises the question whether there is a common mechanism behind it. A comparison of the immunosuppression induced by gammaretroviruses and by lentiviruses, including HIV-1, may help to identify a general mechanism, if any, and individual strategies.
The role of the retroviral transmembrane envelope proteins
The transmembrane envelope proteins of retroviruses play an important role during infection facilitating the fusion of the viral and cellular membrane due to an interaction of the C-terminal and N-terminal helical regions (Fig. 1).
It was shown before that inactivated retrovirus particles are able to inhibit proliferation of immune competent cells in vitro and in vivo, indicating that a viral protein is directly involved. For example, cats immunized with ultraviolet (UV)-inactivated FeLV showed abrogation of tumor immunity and an increased tumor incidence compared with nonimmunized animals after challenge with active virus [29,34]. An inhibition of mitogen-triggered proliferation of peripheral blood mononuclear cells (PBMCs) in vitro was observed when preparations of the following viruses were added: HIV-1 [40-42], FeLV , baboon endogenous virus (BaEV) [44,45], porcine endogenous retroviruses (PERVs) , a type D retrovirus [44,47,48], the human endogenous retrovirus (HERV)-K , as well as the KoRV . The virus preparations were all inactivated by UV, ether or freeze-thawing. When the structural proteins of FeLV were fractionated and analyzed for inhibition of lymphocyte activation, only the transmembrane envelope protein p15E, but no other viral protein, acted inhibitory . Similar results were obtained with the transmembrane envelope proteins of a type D retrovirus , and MuLV , purified from virus particles. The transmembrane envelope protein of FeLV was shown to inhibit mitogen-triggered stimulation of T cells, mixed lymphocyte reactions, IL-2-stimulated proliferation of T cells, mitogen-triggered proliferation of B cells, neutrophilic and erythroid cell function, as well as receptor motility on the cell surface (for review see [54,55]). It is important to note that the transmembrane envelope proteins are interspecies-reactive: FeLV-derived p15E inhibited not only feline PBMCs, but also cells from other species including humans . The retroviral transmembrane envelope proteins are not only reactive in in-vitro assays, but also in vivo. They inhibit macrophage accumulation in mice , antibody production against a tumor antigen in cats, leading to an increase in tumor incidence  and inhibited antiviral responses .
Experiments using a tumor rejection assay have given further evidence of immunosuppressive activity in vivo. Transfection and expression of the transmembrane envelope proteins of the MuLV , of the deltaretrovirus Mason-Pfizer monkey virus,  or of the human endogenous retrovirus HERV-H in mouse tumor cells, normally growing in immunocompromized mice only, prevented rejection of the tumors and allowed tumor growth when injected in immunocompetent animals. This indicates that the transmembrane envelope proteins of different retroviruses act immunosuppressively in vivo. In the same tumor rejection model, the immunosuppressive properties of syncytin 1 and syncytin 2 expressed in the human placenta were investigated . Both molecules are important for the successful accomplishment of pregnancy, contributing to the generation of the syncytiotrophoblast. Syncytin 1 corresponds to the envelope protein of HERV-W, and syncytin 2 to that of HERV-FRD (FRD indicates a specific amino acid motif in the polymerase sequence of this virus). Also in the mouse placenta, two syncytins are expressed, A and B. The human and the murine syncytins are not related. Whereas syncytin 2 (human) and B (mouse) have immunosuppressive properties in vivo, syncytin 1 and A are inactive in the tumor rejection model .
Immunomodulatory properties were also shown for the transmembrane envelope protein gp41 of HIV-1. It was investigated either as a recombinant protein produced in bacteria [62-65] or produced in human cells , or in the context of replication-competent viruses . It is important to note that nanogram amounts of gp41 are active . When it was added to PBMCs from healthy donors, a modulation of cytokine release, for example, an enhancement of the release of IL-10 and IL-6, and a modulation of the expression of more than hundred genes were observed .
It is important to indicate that the retroviral transmembrane envelope proteins may be found in different forms in the infected organism, not only on intact virus particles hidden beneath the surface envelope protein (Fig. 2). Transmembrane envelope proteins may be found on the surface of cells after shedding of the surface envelope protein, and in immune complexes (antibody-gp41 complexes in the case of HIV-1). Recent studies showed that gp41 may be present on the surface of virus particles without gp120 as monomeric or trimeric stumps .
The immunosuppressive domain of retroviral transmembrane envelope proteins
The immunosuppressive (isu) domain was first described for p15E of FeLV in 1985 and the consensus sequence which is highly conserved among all gammaretroviruses was designated CKS-17 . A synthetic peptide corresponding to this domain inhibited cellular immunity in vivo and in vitro, mitogen-driven proliferation of PBMCs, B-cell activation, NK cell activity and monocyte functions (given below) .
Effects of the immunosuppressive domain of gamma-retroviruses:
1. Inhibition of NK cells, cytotoxic T lymphocytes, macrophage-mediated tumor lysis, respiratory burst of monocytes, immunoglobulin production, proliferation of IL-2-dependent T cells, alloantigen-stimulated proliferation of lymphocytes, interferon (IFN)-γ-induced IFN regulatory factor (IRF) 1 expression, delayed-type hypersensitivity in vivo.
2. Down-regulation of IFN-γ (dependent on IL-10), IL-2, tumor necrosis factor (TNF)-α, IL-12 (independent of IL-10)
3. Up-regulation of IL-4, IL-5, IL-6, IL-10
4. Activation of intracellular cyclic adenosine monophosphate (cAMP), extracellular signal-regulated kinase 1 and 2 (ERK1/2) via the phospholipase C (PLC 1)-protein kinase C (PKC)- rapidly accelerated fibrosarcoma 1 (Raf1)- mitogen-activated protein kinase/ERK kinase (MEK) pathway, MEK, PKCmu, Raf1, PLC
After the identification of HIV-1 as the cause of AIDS, a synthetic peptide corresponding to the isu domain of the transmembrane envelope protein gp41 was also shown to be inhibitory in different in-vitro assays [69-72].
The isu domain is located N-terminally to the immunodominant Cys-Cys loop (Fig. 1). This domain is highly conserved among retroviruses; however, three groups can be distinguished: gammaretroviruses, lentiviruses and betaretroviruses (Fig. 3). During infection, the transmembrane envelope protein undergoes different conformational changes ending up in a six-helix bundle with the N-terminal helix and the C-terminal helix in an antiparallel position. In this stage, the isu domain is located opposite the S3 domain, which binds to the gC1qR (globular C1q receptor) and induces expression of NKp44L, the ligand of the natural cytotoxicity receptor NKp44 . Antibodies against the S3 domain correlated with a better disease progression , and immunizing macaques infected with a SIV/HIV hybrid with S3 peptide-carrier molecules induced S3-specific antibodies which prevented NKp44L expression on CD4+ T cells in vivo and subsequently preserved the CD4+ central memory T cells . Whether such antibodies also interfere sterically with the isu domain is unclear.
It is important to note that peptides corresponding to the isu domain were biologically active only when bound to a carrier protein or applied as homopolymers [42,49,70-72]. It is thought that conjugation or polymerization is required to achieve a biologically active conformation or a multiplicity that may be required for interactions with the target cell .
Deletion of the isu domain of the Env protein of MuLV abrogated the immunosuppressive activity in the tumor rejection model , supporting the fact that the isu domain is the biologically functional domain in the transmembrane envelope proteins. Furthermore, mutations in the isu domain of the nonimmunosuppressive syncytin 1 converted this protein into an immunosuppressive one and vice versa. Immunization of mice with the nonimmunosuppressive syncytin 1 induced a much better antibody response compared with syncytin 2 . Both results clearly indicate that the isu domain is immunosuppressive in vivo.
Recently, it was confirmed that the FeLV envelope protein is immunosuppressive in vivo, using the tumor rejection assay, and it was shown that this immunosuppressive activity can be abrogated by targeted mutation of a specific amino acid . With the introduction of the mutated envelope sequence into a previously well characterized canarypoxvirus-vectored vaccine, the frequency of vaccine-induced FeLV-specific IFN-γ-producing cells was increased, whereas, conversely, the frequency of vaccine-induced FeLV-specific IL-10 producing cells was reduced. This shift in the IFN-γ/IL-10 response was associated with a higher efficacy of the vaccine .
Single mutations in the isu domain of gp41 of HIV-1 also abrogated its property to induce an elevated production of IL-10 and IL-6 . Interestingly, some mutations in the isu domain of HIV-1 totally abrogated the immunomodulatory property. Other mutations abrogated the immunomodulatory activity in one donor, but not in the other. Mutations of the amino acids 1-4, 9 and 14 totally abrogated the immunomodulatory properties (LQARILAVERYLKDQQL, abrogating mutations typed in bold). Similarly, immunization of rats with a mutated gp41 induced higher antibody titers than the wild type , indicating that gp41 had an immunosuppressive effect in vivo.
Mechanisms of action
Retrovirus particles, their transmembrane envelope proteins and isu peptides have been shown to modulate the cytokine production of normal PBMCs (given as a list under the 'The immunosuppressive domain of retroviral transmembrane envelope proteins' section) [42,55,62-65]. Using cytokine arrays it was shown that the transmembrane envelope proteins of HIV-1, KoRV, PERV, HERV-K and the corresponding isu peptides increased the release of the following cytokines: IL1-ß, IL-10, IL-6, IL-8, monocyte chemoattractant protein (MCP)-1, MCP-2, tumor necrosis factor (TNF)-α, macrophage inflammatory protein (MIP)-1α and MIP-3 [42,49,50,78]. In contrast, the expression of IL-2 and chemokine (C-X-C motif) ligand (CXCL-9, also called monokine induced by gamma interferon, MIG) decreased. Microarray analysis of the expression of more than 25 000 genes in human PBMCs treated with the homopolymer of the HIV-1 isu peptide or with the recombinant transmembrane envelope protein of HERV-K confirmed the cytokine data and showed up-regulation and down-regulation of more than 300 genes [49,78]. Among the genes with the highest up-regulation were IL-6, matrix metalloproteinase 1 (MMP-1), triggering receptor expressed on myeloid cells 1 (TREM-1) and others. Among the down-regulated genes were ficolin-1 (FCN1), selenoprotein P, plasma, 1 (SEPP1), TREM-2 and CXCL-10 (also called interferon gamma-induced protein 10, IP-10), all involved in innate immunity. Using real-time PCR, the data were confirmed and kinetic investigations were performed. It is of interest that the changes in gene expression were similar when the homopolymer of the HIV-1 isu peptide, or the recombinant transmembrane envelope protein of HERV-K were added to PBMCs from the same healthy donors.
It is still unclear how retroviruses, their transmembrane envelope protein and the immunosuppressive peptide act on the target cell to induce the described changes in cytokine release and gene expression. Cell surface binding proteins, which possibly may represent receptors of the isu domain of gp41 and of p15E have been described [79-85]; however, it is unclear whether they function as such and whether they are involved in the signal transduction.
Special features of the immunodeficiency induced by HIV-1
Whereas the gammaretroviruses have no accessory genes, complex retroviruses such as HIV-1 may use numerous accessory proteins such as Vif, Vpu, Vpr, Tat, Rev and Nef to evade the host immune response and to counteract restriction factors (reviewed in ). Nonetheless, retroviruses without functional nef and other accessory genes are also able to induce fatal immunodeficiencies, for example, gammaretroviruses and the feline immunodeficiency virus (FIV).
In contrast to many other retroviruses, HIV infects cells of the immune system, mainly CD4+ cells, for example, lymphocytes, macrophages and dendritic cells. The virus directly interacts with these cells via the surface envelope protein gp120. As a result, a depletion of the CD4+ cells, in the beginning mainly in the gastrointestinal tract, was observed . On the contrary, activation of the target cells is required for an effective virus replication. This demonstrates the dilemma for the virus: on the one hand, the virus has to inhibit the immune system to maintain infection; on the other hand, it indirectly forces activation of the target cells.
In addition to the structural transmembrane envelope protein and the accessory proteins listed above, HIV uses some more strategies to evade the immune system. These strategies include the high variability which allows HIV-1 to escape the immune responses, cell-to-cell spread of the virus, hiding in dormant cells undetected by the immune system and masking of epitopes by carbohydrates (the surface envelope protein gp120 is one of the most glycosylated proteins) (for review see ).
One important hallmark of HIV-1 infection is the hyperactivation of the immune system (however, this topic is not well studied in other retroviral infections). The reason why the immune system is drastically activated is still unclear. One effect of hyperactivation is the generation of activated target cells for infection. Certainly, the immune response against the virus itself is one possibility to explain hyperactivation. Another is the so-called microbial translocation. This term characterizes a massive invasion of microorganisms from the gut into the bloodstream. In a healthy individual, several immune mechanisms exist to kill or transport bacteria back to the lumen of the intestine; however, the massive dysfunction of the mucosal immune system during HIV and SIV infection fails to prohibit systemic microbial translocation . It is still unclear how the virus induces the changes in tight junction gene expression in the epithelial cells leading to an increased permeability. It is also unclear how it inhibits the gut immune system, thus allowing microorganisms to translocate. Interestingly, translocation is still observed after treatment of the patients with combination antiretroviral therapy (cART) .
Exhaustion of the immune cells was reported to contribute to the immunodeficiency by HIV-1. However, it has recently been shown that antibodies binding to programmed cell death protein 1 (PD-1) restore the function of so-called exhausted cells. In vivo, in SIV-infected rhesus monkeys, antibodies against PD-1 markedly reduced the microbial translocation  and increased the animals' life expectancy [88,89]. PD-1 is a cell surface membrane protein of the immunoglobulin superfamily. PD-1 has two ligands, PD-L1 and PD-L2. Several lines of evidence suggest that PD-1 and its ligands negatively regulate immune responses. Meanwhile, it is clear that PD-1+ T cells mark out a particular differentiation stage or trafficking ability rather than exhaustion . In chronic SIV infection, the level of PD-1 expression by T cells does not by itself serve as a reliable marker for immune exhaustion [90,91]. The antibody directing PD-1 that restores the function of the so-called exhausted cells obviously prevents binding of the ligand to the PD-1 receptor. To rephrase it, the antibody prevents the immunoregulatory/immunosuppressive activity of the ligand. This indicates that the so-called exhausted cells are actively suppressed.
It is still unclear why simian 'immunosuppressive' viruses are apothogenic in their natural hosts. Although HIV-1, HIV-2 and SIVmac induce fatal immunodeficiencies, more than 40 primate species are infected with related viruses without inducing a disease. Despite the fact that the virus load (and the amount of the transmembrane envelope protein) in the natural hosts is as high as in late-stage HIV-1 infection, the animals stay healthy . Infection of natural hosts is not associated with high levels of neutralizing antibodies  and suppression of the innate immunity does not induce disease . The sequence of the isu domain is not mutated and when transmitted to another species the virus may induce AIDS as shown in the case of SIVsm/HIV-2. So far, estimates of why these viruses are apathogenic in their natural host are purely speculative. One speculation is that the putative receptor for the transmembrane envelope protein is mutated. Another speculation is that the corresponding signal transduction pathway changed in these animals during co-evolution with the virus. On the contrary, some trans-species transmitted viruses induce immunodeficiency even in the absence of a high virus load .
Also, it is still unclear why - despite cART and absence of virus replication - the abnormalities of the immune cells are not restored completely . This may be due to the fact that the immunosuppressive gp41 is expressed on the surface of infected cells in lymphoid organs. Interestingly, HIV structural proteins may persist in the follicular dendritic network of the germinal centers for months to even years after cART .
Exogenous retroviruses: summary and outlook
To summarize, in-vitro and in-vivo evidence showed that the isu domain of retroviral transmembrane envelope proteins, including that of HIV-1, influences cytokine release, gene expression and immune response, and may therefore contribute to immunosuppression in the infected host. The possible functions of the transmembrane envelope protein at different time points after infection are shown in Fig. 4. As regards HIV-1, the contribution of the transmembrane envelope protein is difficult to evaluate due to the T-cell depletion, the activity of the accessory proteins and hyperactivation. Although the immune system of the host develops a strong response against the virus and the translocated microorganisms, the permanent suppression by the transmembrane envelope protein leads to retardation of the immune response and to immunodeficiency.
HIV-1 infection starts with the entry of a limited number of virus particles. At this stage, cellular restriction factors and the local innate immunity may prevent infection and replication. As shown by us , the transmembrane envelope protein down-regulates genes involved in innate immunity such as ficolin 1, SEPP-1, TREM-2 and CXCL-10, and up-regulates IL-10 and other cytokines. At this stage, all accessory viral proteins, including Nef, are absent and viruses with mutations in the isu domain that abrogate the immunosuppressive properties are unable to replicate to a detectable degree in the infected individual despite being replication-competent .
As shown previously, antibodies against the isu domain reduced the activity of the isu peptide polymer . It would be interesting to study whether antibodies against the isu domain and/or antibodies and drugs against the still unknown receptor may reduce immunosuppression.
Endogenous retroviruses are the result of infection of germ cells and integration of the provirus in all cells of the organism followed by transmission to the progeny. Whereas in mice and pigs, endogenous retroviruses may be expressed in normal tissues, in primates including humans, the expression is more limited . HERV-K, one of the human endogenous retroviruses with open reading frames, is expressed in germ cell tumors, melanomas and in the placenta [98-101]. In the placenta, three other endogenous retroviruses are expressed, HERV-W (also called syncytin 1), HERV-FRD (called syncytin 2) and HERV-3 (for review see ). The envelope proteins of these viruses participate in the generation of the syncytiotrophoblast, the outer layer of the placenta, by cell fusion . Furthermore, immunosuppressive properties were observed for HERV-K, syncytin 2, HERV-H, but not for syncytin 1 [49,61,103]. These data allow speculating that the expression of the endogenous retroviral transmembrane envelope proteins may, among numerous other factors, contribute to the tumor progression as well as to the protection of the embryo from rejection. When human PBMCs were incubated with the recombinant transmembrane envelope of HERV-K, or with a homopolymer of its isu peptide, or with HERV-K particles released from a human teratocarcinoma cell line, the modulation of the cytokine release and gene expression was similar to that induced by HIV-1 .