Retroviral Restriction Factors and Infectious Risk in Xenotransplantation

Abstract

The clinical application of xenotransplantation poses immunologic, ethical, and microbiologic challenges. Significant progress has been made in the investigation of each of these areas. Among concerns regarding infectious risks for human xenograft recipients is the identification in swine of infectious agents including porcine endogenous retroviruses (PERV) that are capable of replication in some human cell lines. PERV replication has, however, been difficult to demonstrate in primate-derived cell lines and in preclinical studies of non-human primates receiving porcine xenografts. Endogenous “retroviral restriction factors” are intracellular proteins and components of the innate immune system that act at various steps in retroviral replication. Recent studies suggest that some of these factors may have applications in the management of endogenous retroviruses in xenotransplantation. The risks of PERV infection and the potential role of retroviral restriction factors in xenotransplantation are reviewed in detail.

Introduction: Infection in Xenotransplantation

Xenotransplantation, the transplantation of viable cells, organs, or tissues between species, has been proposed as a solution to the shortage of human organs for the treatment of organ failure. Recent advances in experimental xenotransplantation have increased the possibility that clinical trials of xenotransplantation will be feasible. As with human allotransplantation, the prevention of infection associated with xenotransplantation is central to the ultimate success and acceptance of this technology (1, 2). The terms “xenosis”, “direct zoonosis” and “xenozoonosis” were coined to reflect the unique epidemiology of infection due to organisms potentially carried by xenogeneic tissues and transplanted into immunosuppressed human hosts.

Swine are generally considered as providing the greatest advantages as a source species for clinical xenotransplantation. While non-human primates are closer immunologically and metabolically to humans as a potential source of organs for transplantation, these have been excluded by many regulatory agencies due to ethical issues, the presence of human-tropic viruses in many primates, poor size matches, and the expense and difficulty of breeding. Swine, while posing greater immunologic barriers to transplantation into primates, are easier to breed, are good size matches for humans, and have been genetically engineered to express or suppress a variety of specific gene products relevant to transplantation. Swine can also be bred to exclude many common potential human pathogens (“designated-pathogen-free”) with barrier maintenance. However, based on experience with allotransplantion and in xenotransplantation, a major concern is that the immunosuppression required to prevent graft rejection coupled with inflammation associated with graft ischemia and reperfusion and with immune activation following transplantation, will result in the activation of latent viral infections (1). For example, the activation of porcine cytomegalovirus (PCMV) following renal pig-to-baboon xenotransplantation contributes to consumptive coagulopathy via endothelial activation (3). This observation lead to successful efforts to exclude PCMV from breeding herds by early weaning (3, 4). Sensitive, quantitative microbiologic assays are now available for PCMV, porcine lymphotropic herpesvirus, and porcine circovirus (5). Other infections occasionally observed in immunosuppressed swine include adenovirus, hepatitis E virus, and porcine respiratory and reproductive virus. Despite significant advances in knowledge regarding these pathogens, the absolute risk for transmission of such infections remains unknown without human studies.

Retroviral Infection in Xenotransplantation

The clinical course of retroviral infection transmitted with human allografts (e.g., HTLV-1 and HIV) is accelerated in solid organ transplant recipients (1). Concerns regarding retroviral transmission in xenotransplantation relates to the potential for “silent” or asymptomatic transmission, genomic insertion, and the potential for subsequent alterations in gene regulation, oncogenesis, or viral recombination in the recipient (2) The identification of a family of porcine endogenous retroviruses (PERV) in the pig genome and the demonstration of infectivity of some of these viruses for human cells in vitro has raised concerns regarding the safety of clinical xenotransplantation (6–12). Swine carry three replication-competent subtypes of endogenous PERV termed PERV-A, PERV-B and PERV-C (6, 7, 13, 14). Whereas PERV-A and PERV-B can infect human cells in vitro, PERV-C can infect only pig cells (8, 10, 11). The capacity of PERV to infect non-human primate cells is limited, at least partially due to differential display and efficiency of PERV receptors on these cells (15–17). Despite evidence of infection of human cells by PERV in vitro, no evidence of infection has been demonstrated of human cells in vivo and no disease resulting from this family of viruses has been described in swine or humans to date (11, 18, 19).

PERV mRNAs are expressed in all pig tissues; thus, any transplanted organ is a potential source of virus. Swine are classified according to whether or not peripheral blood mononuclear cells (PBMCs) transmit PERV to human cells in vitro (‘transmitters’ or ‘nontransmitters’) (11). Some animals do not transmit PERV to either swine or human cells (“null” animals) (11). It is unclear whether the “transmitting” phenotypes are stable traits (12). Studies suggest that expression and infection may be amplified by stimulation of swine peripheral blood lymphocytes in vitro (9). “Transmitting” animals generally transmit either PERV A or a recombinant PERV AC that contains the envelope region (and receptor binding domain) of PERV-A fused with PERV C (11, 19). Recombinant virus is produced via exogenous recombination between PERV A and C mRNAs or with expression of genomic PERV AC loci (12, 19) The efficiency of infection by PERV AC improves with serial passage in vitro, a trait that has been associated with increased pathogenicity for other retroviruses (9, 20–24).

The risk for viral activation in allotransplantation is increased by graft rejection and intensified immunosuppressive anti-rejection therapy (1, 2). One of the major hurdles to successful xenotransplantation is hyperacute graft rejection or acute vascular rejection (HAR) of xenografts due to the binding of preformed or “natural antibodies” to antigens on the surface of the donor vascular endothelium with subsequent complement activation, endothelial injury, thrombosis and acute graft injury. Most of these antibodies are directed against the carbohydrate determinant Galα1-3Galβ1-4GlcNAc (Gal), a terminal sugar of cell-surface glycoproteins in all mammalian species except Old World primates and humans. These natural antibodies appear to provide some defense against human infection by retroviruses, parasites and other common organisms carrying the Gal epitope. Thus, strategies used to reduce hyperacute rejection, including depletion of anti-Gal antibodies and genetic engineering of swine to express human complement regulatory proteins to decrease complement deposition, might impact host defenses against viral infection. Miniature swine have also been produced with disrupted genes for alpha-1,3-galactosyltransferase (GalT) and which lack the target sugar for HAR – “Gal knockout (GalT-KO) swine” (25). Xenografts from these pigs have improved survival without HAR (26). Increased levels of circulating PERV virus have not been detected in the GalT-KO swine or in immunosuppressed baboon recipients of GalT-KO grafts (27).

A number of additional strategies have been studied to reduce the risk of PERV infection in xenograft recipients. These include use of non-transmitting swine or swine without active PERV loci as source animals, use of antiretroviral agents in recipients, viral vaccines, or the reduction of viral replication in vitro using RNA interference, various antibody therapies, and amplification of antiviral restriction factors (28).

Restriction Factors for Retroviral Replication

Various animal species have evolved protective mechanisms to avoid retroviral infection. This system of antiviral activities is mediated by a group of intracellular proteins referred to as “restriction factors”. Restriction factors are generally species-specific, so that each species can restrict infection due to a subset of viruses with exquisite sensitivity (29). These factors are components of the innate immune system and act at diverse steps in the retroviral replication (30) (Figure 1). A number of these factors have been partially characterized including: APOBEC (apolipoprotein B mRNA-editing catalytic polypeptides) which are cytidine deaminases that disrupt viral DNA during synthesis; TRIM5α which disrupts the viral capsid (CA) after cell entry; TRIM28 which blocks viral transcription; ZAP (zinc-finger antiviral protein) which directs degradation of viral RNAs; and tetherin, which traps virions on the surface of infected cells (30, 31) (Figure 1). Antiviral restriction systems are being studied with the prospect of the development of therapeutic agents to regulate expression of these factors and to enhance antiviral activities (31).

Figure 1.

The retroviral life cycle and sites of activity of the major antiviral restriction factors. Envelope glycoproteins of the retrovirus interact with specific host-cell membrane protein receptors. (1) The retroviral envelope fuses with the plasma membrane and enters the host cell. (2) Following fusion, the nucleocapsid enters the cytoplasm and (3) uncoating of viral core occurs. (4) Viral reverse transcriptase copies single strand viral RNA into double-stranded DNA. (5) Viral DNA is transported into the nucleus and integrated into host-cell chromosomal DNA. (6) Integrated viral DNA is transcribed by the host-cell RNA polymerase, generating mRNA molecules and new viral genomic RNA molecules [TRIM28 blocks viral transcription]. Viral mRNAs are translated into viral proteins (envelope, capsid, and reverse transcriptase). (7) Newly synthesized viral proteins and genomic RNA gather to form immature viral particles [ZAP degrades viral RNAs]. (8) New virions bud from the cell surface, acquiring an envelope including host-cellular and viral proteins from the cell membrane [Tetherin traps virions on the cell surface].

APOBEC

In studies of HIV biology, a restriction factor was identified as the basis of antiviral activity that inhibited production of infectious virus in non-permissive cell lines such as H9 and CEM. The gene encoding this activity was initially identified as CEM15 and renamed APOBEC3G. APOBEC3G causes G-to-A hypermutation in nascent retroviral DNA strands during reverse transcription. The C to U change in the minus strand produces a G to A mutation in the sense strand of the DNA product (31). These changes cause post-transcriptional silencing by mutational inactivation of the retroviruses. The family of APOBEC-related proteins also causes site-selective mutation of DNA through deoxycytidine deamination to deoxyuridine (DNA editing).

APOBEC3G protein converts HIV-permissive cells into a nonpermissive state. In HIV infection, however, Vif protein mediates viral evasion of host antiviral factor CM15/APOBEC3G. APOBEC3G also acts on a broad range of retroviruses other than HIV, suggesting that hypermutation by editing is a general innate defense mechanism (32). Thus, APOBEC3G can restrict hepatitis B virus and adenovirus type 2. Accumulating evidence suggests that some viruses normally inhibited by APOBEC expression have, like HIV, developed mechanisms to avoid deamination.

Ancestral APOBECs originated from a branch of the zinc-dependent deaminase superfamily as a defense against retroviruses concurrent with the evolutionary appearance of vertebrates and of adaptive immunity. APOBEC3G has been demonstrated in spleen, testes, ovary, and peripheral blood leukocytes and T-lymphocytes. APOBECs are also expressed in cancer tissues and cell lines, possibly due to altered control of expression in these tissues or as an on-going defense mechanism. The APOBEC proteins in humans include: AID and APOBEC1 (located on chromosome 12); APOBEC2 (chromosome 6); and a series of seven APOBEC3 genes, which are on human chromosome 22. These are APOBEC3A, APOBEC3B, APOBEC3C, APOBEC3DE, APOBEC3F, APOBEC3G, and APOBEC3H. A new APOBEC subfamily, APOBEC4, was recently identified. APOBEC3G, APOBEC3F, APOBEC3B and APOBEC3DE block replication of human immunodeficiency virus type 1 (HIV-1).

In vitro studies indicate that an APOBEC3 protein that is derived from a mammal to which the virus has not yet adapted may provide an effective strategy to thwart species-specific viral counter-defences (30, 33). Thus, murine APOBEC3 can block HIV-1 replication (regardless of Vif). It is reasonable to hypothesize that cross-species expression of an APOBEC3 protein may be used to create a barrier to retroviral infection.

PERV has been studied largely using the permissive human embryonic kidney (HEK) 293 cell line. HEK293 is an adenovirus-transformed human cell line that does not express APOBEC3G and may not reflect the natural “infectability” of human cells. Given the difficulty in demonstrating PERV replication in human and other primate cells in vitro or in vivo, it has been suggested that endogenous primate APOBEC3G expression inhibits the replication of PERV in normal human cells. Jónsson and colleagues demonstrated that PERV transmission from a high level PERV-producing porcine kidney cell line (PK15) to human cells is markedly reduced by the expression of human APOBEC3G in virus-producing pig kidney cells (33). Notably, over-expression of the porcine APOBEC3 protein (APOBEC3F) failed to interfere significantly with PERV transmission while effectively inhibiting the replication of HIV. Human APOBEC3G and pig APOBEC3F were shown to reside in a largely cytoplasmic location in human and pig cell lines; human APOBEC3G was normally regulated in pig cells (33). Other studies have suggested that both human and porcine APOBEC3 (APOBEC-Ss3F) are potent inhibitors of PERV (34). These studies differ in the cellular source of PERV (porcine versus human) and in the apparent ability of porcine proteins to inhibit replication in human cells. However, both suggest that endogenous APOBEC may limit PERV infectivity for human cells and that enhanced production of restriction proteins might offer an opportunity to interfere with possible retroviral infection (33). The antiviral mechanisms of the APOBEC3 proteins remain under study (33). Clarification is needed of the role of APOBEC and RNA viruses that do not replicate through DNA intermediates.

TRIM Family

TRIM5α is a member of the tripartite motif (TRIM) protein family with over 70 members involved in diverse cellular processes, including proliferation, differentiation, development, oncogenesis and apoptosis (35). Like APOBEC, TRIM factors are zinc finger domains. Some primate TRIM proteins display anti-retroviral properties. The antiviral mechanism of TRIM5α remains incompletely understood but appears to involve virus-specific binding to capsid proteins via a C-terminal domain (PRY/SPRY or B30.2) and interference with viral uncoating, thus preventing reverse transcription of the viral genome and transport to the nucleus. Capsid protein from restricted viruses is removed by a proteosome-dependent degradation pathway (36). Other cellular proteins appear to be involved in TRIM5α inhibition. A variety of TRIM5-spliced isoforms exist but only TRIM5α carries the PRY/SPRY domain with antiviral activity (29).

TRIM5α has generated great interest in the HIV/AIDS field as retroviral restriction has been observed in Old World monkey-derived cell lines, in some New World monkeys and in Rhesus macaque. In susceptible cells, gene silencing of TRIM proteins enhances HIV and MLV viral entry or release (35). Of interest to transplantation, the insertion of a cDNA encoding the enzyme cyclophilin A (CypA) into the TRIM5 locus of New World owl monkey cells renders HIV-1 sensitive to restriction by TRIM5α (29). By contrast in humans, CypA acts as a positive factor for HIV-1 replication (37). Human TRIM5α is a broadly effective restriction factor for viruses including equine infectious anemia virus (EIAV) and N-tropic murine leukemia virus (N-MLV) (38). TRIM factors also have activity against a variety of lentivirus as well as Influenza A, Ebola virus, Lassa fever virus, herpes simplex virus, rabies virus, vesicular stomatitis virus, human foamy virus, HIV, avian leukosis virus, and lymphocytic choriomeningitis virus. PERV A and PERV AC are insensitive to restriction by TRIM5a molecules in permissive feline Crandall-Reese feline kidney (CRFK) cells expressing TRIM5a proteins from human, African green monkey, rhesus macaque, squirrel monkey, rabbit or cattle (36).

Tetherin

Tetherin is a type I interferon-inducible molecule that blocks release of retroviruses and filoviruses from infected cells and is antagonized by viral accessory proteins including HIV-1-Vpu, Ebola virus glycoprotein, and SIVtan Env. Tetherin is a group of protein-based cell surface tethers recently identified as CD317 and also known as BST-2/HM1.24/CD317. Tetherin prevents retroviral particle release from the surface of producer cells. Tetherin was discovered through characterization of HIV-Vpu (30). When Vpu is absent, HIV virions are bound to tetherin at the cell surface and in intracellular compartments. In the presence of Vpu, tetherin-induced particle retention is antagonized and viral spread is facilitated. Vpu activity appears to be species-specific (30). Tetherin is upregulated in some HIV-permissive cell lines upon stimulation by interferon alpha (IFNα); these cells became resistant to infection with Vpu-minus HIV-1 without significantly affecting wild-type HIV-1 replication (30). Tetherin is known to be expressed on terminally differentiated B cells, bone marrow stromal cells and plasmacytoid dendritic cells. Tetherin may restrict a wide range of enveloped viruses. Vpu can enhance the release of Ebola virus in cells treated with IFNα, suggesting that tetherin acts nonspecifically to inhibit enveloped virion release. Tetherin has also been shown to inhibit the release of Lassa and Marburg virions. This prediction is borne out by the observation that tetherin can inhibit the release of MLV as well as HIV-1. It is unclear whether all forms of tetherin carry this antiviral function. Recent studies by Mattiuzzo et al demonstrated that over-expression of either human or porcine tetherin in pig cells significantly reduces PERV production (39).

Restriction factors and PERV in Clinical Xenotransplantation?

Progress in preclinical studies of xenotransplantation has refocused attention on potential infectious hazards posed by this technology. It has been demonstrated that available antiretroviral agents have the capacity to inhibit PERV replication in vitro (40). Given the application of genetic engineering of swine for studies of xenotransplantation, it is reasonable to consider whether expression of restriction factors for PERV might further enhance the safety of pig-to-human xenotransplantation. Given the specificity of each restriction factor, the activity of specific factors for each strain of PERV merits assessment in the appropriate host species. For example, the proviral clone (HERV-KCON) similar to the HERV-K progenitor (human endogenous retrovirus) is resistant to inhibition by (TRIM)5α and APOBEC 3G, but inhibited by APOBEC 3F. Most of these factors can be studied in vitro. It is notable that the major human target cell line used in investigations of PERV in xenotransplantation lacks APOBEC, likely due to adenovirus transformation. This may provide optimism that normal human cells will be relatively resistant to PERV replication in vivo. However, further studies of the infectivity of porcine viruses for human cells are needed to assure the safety of clinical trials of xenotransplantation.