Fuse me IFITM can!

No abstract


Ariberto Fassati
Two independent papers show that IFITMs proteins are incorporated into HIV-1 particles and restrict infection at an early stage Five members of the interferon inducible transmembrane (IFITM) gene family are found in humans: IFITM1, 2, 3, 5 and 10. Typically, IFITMs 1, 2 and 3 are induced by both type I and type II interferon [1,2] although they are also constitutively expressed in cells of the upper airway and visceral pleura. IFITM 5 is not interferon-inducible and is involved in bone mineralization, and IFITM10 function is unknown. IFITM1, 2 and 3 are likely to have arisen from lineage and species-specific gene duplications. Indeed they are highly similar and highly conserved genes throughout evolution, with at least one member present in jawless fish (lampreys). The rapid expansion of the IFITM gene family during mammalian evolution was accompanied by positive selection, suggesting a broad role as antivirals. Conversely, IFITM5 and 10, which are not interferon-inducible, show no sign of gene duplications and positive selection [3,4].
IFITMs, as the name suggests, are located at membranes; IFITM1 is found mainly at the plasma membrane, IFITM2 and 3 are found mainly in intracellular compartments and endosomal membranes, although stimulation with interferon can change their distribution [5,6].
IFITMs have been shown to restrict influenza A virus (IAV) infection, in vitro and in a Ifitm3 -/mouse model [7][8][9]. IFITM3 restricts IAV infection more potently than IFITM1 and 2, acting at the early stages of the viral replicative cycle, most likely fusion [7]. Furthermore, Caucasian and Han Chinese individuals bearing the synonymous single nucleotide polymorphism (SNP) rs12252 in the IFITM3 gene were significantly more likely to suffer from severe IAV infection and be hospitalized [8]. This SNP resulted in a truncated IFITM3 protein that is found mainly at the cell membrane instead of endosomal membranes [8]. IAV fusion and release into the cytoplasm occurs in the late endosomal compartment following acidification. Therefore it is likely that IFITM3 needs to be in the right intracellular compartment to exert its anti-IAV activity.
IFITMs proteins have been reported to restrict infection of different viruses, including West Nile and Dengue, Vescicular stomatitis, Rabies, SARS coronavirus, Semliki forest and several Bunyaviridae, with the exception of Crimean-Congo haemorragic fever virus. IFITMs inhibit virus infection by preventing efficient virus-cell fusion and it is quite possible that several IFITMs with different intracellular location have arisen to protect against viruses using different entry pathways [6]. Interestingly, Hepatitis C virus, murine leukaemia virus and Lassa virus are not restricted by IFITMs proteins, which is quite puzzling given that these viruses have overlapping entry pathways with other viruses that are restricted [6].
Whether HIV-1 is restricted by IFITMs has remained somewhat controversial. Early reports suggested that HIV-1 was not restricted, at least in HeLa-CD4 cells [7]. However a later study showed that in CD4+ T cells, replication of HIV-1 was indeed impaired [10]. In this study, IFITM 2 and 3 reduced infection at an early stage, presumably at entry, whereas IFITM1 mainly reduced virus production [10]. Using nice imaging, the two studies demonstrate that, upon stimulation with interferon, IFITMs co-localises with sites of virus budding at the plasma membrane and with sites of Env accumulation at intracellular sites.
Both studies agree that the presence of IFITMs at the virus membrane impairs fusion, which is consistent with studies on other viruses. In addition, Compton at al. found that cell-to-cell virus transmission is also impaired by IFITMs and that this reduction is due to less efficient fusion at the virological synapse.
Several questions arise from these studies: why must IFITMs be incorporated into HIV-1 particles to be restrictive? Consensus has it that amphotropic MLV and HIV-1 fuse at the cell membrane in a pH-independent way, so why is HIV-1 susceptible and amphotropic MLV not susceptible to IFITMs? How is HIV-1 fusion inhibited? Furthermore, it will be interesting to examine the effect of the IFITM3 SNP rs12252 on HIV-1 infection: would aberrant localization of IFITM3 at the plasma membrane make HIV-1 more or less susceptible to restriction?
The two studies by Crompton et al. and Tartour et al. are important because they extend our knowledge on interferon-regulated genes that impact on HIV-1 infection and also because they point to a potentially new mechanism to explain IFITMs antiviral activity. It will be important to investigate if IFITM3 SNPs found in Han Chinese adversely affects progression to AIDS and if HIV-1 evolves differently in individuals carrying this SNP.