Virology tidbits

Sunday, 1 February 2015

HTLV-1 Tax, NF-κB and autophagy

Human T-Cell Leukaemia/Lymphoma Virus type-1 (HTLV-1) is a member of the deltaretrovirus family which also include Bovine Leukaemia Virus (BLV) and Simian T-Lymphotropic Viruses (STLVs) and was the first human retrovirus to be associated with the formation of tumours in humans, being the etiologic agent of Adult T-Cell Leukaemia/Lymphoma (ATL/L; commonly referred to as ATL) as well as a chronic neuroinflammatory disease, HTLV-1-associated myelopathy (HAM)/tropical spastic paraparesis (TSP). Worldwide an estimated number of 10 to 20 million individuals are infected with HTLV-1, with a high prevalence in Japan and the Caribbean, out of which approx. 1-5% develop ATL during their lifetime usually late in life following a period of persistence which can be up to 20 years. HTLV-1 infects a wide variety of cells, including CD4+ T lymphocytes, CD8+ T lymphocytes, B lymphocytes, macrophages, and fibroblasts thus reflecting the ubiquitous expression pattern of its hypothesised receptors (glucose transporter 1, heparan sulfate and proteoglycans and neuropilin-1), although primarily CD4+ T lymphocytes are targeted.

HTLV-1 genome

The key protein involved in the malignant transformation of CD4+ T lymphocytes is the viral Transactivator from the X-gene region (Tax) protein, which activates the expression of viral proteins via the viral cyclic-AMP-responsive promoter located within the HTLV-1 LTR region by binding the cellular transcription factor CREB as well as recruiting the cellular Histone Acetyltransferases p300 and P/CAF via its N-terminus as well as TATA-box-bound TBP protein via the C-terminus, thus enabling the initiation of transcriptional and RNA polymerase elongation. It should be noted that Tax however does not exhibit any DNA binding activity.

As described below, the expression of Tax modulates the cell cycle, induces the DNA damage response, autophagy and other processes, with the activation of Nuclear Factor-kB (NF-κB) pathway being central. The prevailing model involves the binding of Tax to IKKγ/NEMO in the cytoplasm thereby activating IKKα/IKKβ/IKKγ, inducing the phosphorylation (and proteasomal degradation) of IKKα and IKKβ and subsequent nuclear localisation of NF-κB. Following the nuclear translocation of both NF-κB/Rel family members, nuclear Tax associates with c-Rel, RelA, p50 and p52, thereby increasing their dimerisation which is necessary to bind to promoters, both viral and cellular, with p300/CBP and P/CAF being co-activators. Tax mediated nuclear translocation of members of the NF-κB/Rel family therefore facilities the expression of both viral and cellular genes, and indeed hyperactivation of NF-κB is required for Tax induced malignant transformation of cells both in vitro and in vivo.

Tax and Autophagy: inhibition v. activation

Jurkat cells which have been infected with HTLV-1 by co-culturing them with a HTLV-1 transformed cell line, MT-2, exhibit an increase in NF-κB activity as measured by the luciferase activity following the transfection with a pNF-κB-luc reporter plasmid as expected, similar to a Tax-positive Jurkat-TaxP cell line stably expressing Tax. Upon examination of the changes in gene expression in the latter using Microarray analysis, approx. 48 genes exhibited a change in gene expression of at least twofold with 10 genes associated with oncogenesis, differentiation, proliferation, cell development being unregulated and 38 genes down regulated in the presence of Tax. Interestingly one of the genes, Bcl-3, was upregulated not only in Jurkat-TaxP cells but also significantly in MT-2 cells, accompanied by nuclear translocation of Bcl-3 in Jurkat-TaxP. These results indicate that Tax not only induces the translocation of NF-κB but also the expression and nuclear translocation of Bcl-3, which is a negative inhibitor of NF-κB. Furthermore, Bcl-3 also binds Tax and inhibits Tax mediated transcriptional activation of the viral LTR independent of the ability of Tax to activate NF-κB. Also, IKK induced phosphorylation of AMPK and subsequent inactivation of mTOR induces the formation of autophagosomes thus explaining the inability of TaxM22 to induce the formation of autophagosomes since TaxM22 can not bind IKK.

Regarding autophagy, the overexpression of Bcl-3 in HeLa cells co-transfected with Tax induces the formation of autophagosomes following Rapamycin or Pifithrin-α treatment as well as in mock treated cells whereas under starvation the formation of autophagosomes decreased for unknown reasons. The formation of autophagosomes also increases in MT2 and MT4 cells treated with an inhibitor of NF-κB, Bay 11-7082, that inhibits IκBα phosphorylation, suggesting that the inhibition of the NF-κB signalling pathway is sufficient to induce autophagosome formation. Interestingly the expression of a Tax mutant, TaxM22, incapable of inducing the activation of NF-κB also fails to induce the formation of autophagosome, indicating that inducing the expression of Bcl-3 via NF-κB activation is necessary for the induction of autophagy, whereas the expression of TaxM47 not only activates NF-κB but also increases the formation of autophagosomes similar to wt Tax.
In addition to Bcl-3, Tax mediated activation of the IKKβ kinase complex has been postulated to activate autophagy by recruiting a complex containing Beclin-1 and Bif-1 to lipid raft microdomains. In this model, Tax interacts with both complexes directly, thereby promoting the assembly of LC3-II positive autophagosomes.

Increasing the formation of autophagosomes has been shown to increase the replication of HTLV-1 in HEK-293T cells transfected with a HTLV-1 molecular clone (K30) whilst the expression of Tax in HeLa and Jurkat cells as well as in other established T cell lines such as PTX4-1 or PL9-1 not only increases the formation of autophagosomes as measured by the increase in LC3-II but at the same time also inhibiting the fusion with the lysosome as indicated by the accumulation of Neutrophil Cytosolic Factor 4(NCF)/p40phox, which -as discussed for positive strand RNA viruses before - is a common principle in cells expressing viral proteins interfering with the autophagy pathway. Paradoxically, both in HEK-293T cells transfected with the HTLV-1 molecular clone as well as in MT1, MT2, and HUT102 cells the fusion of the autophagosome with the lysosome is not inhibited, suggesting that that in the context of the presence of other viral proteins autophagic flux is functional; it should be noted that in 50% of ATL patients, HTLV-1 infected T lymphocytes do not express Tax as a result of the accumulation of mutations which render Tax inactive.

NF-ϰB dependent and independent signalling pathways

Blocking the fusion step using Bafilomycin-A, Chloroquine or siLAMP2 however increases the production of infectious HTLV-1 by approx. 3-fold as measured by p19 ELISA, suggesting that autophagy is part of the antiviral response. Inhibiting autophagy by the application of PI3K class III inhibitors or transfection of siBeclin-1 as well as shTax indeed resulted in impaired proliferation of HTLV-1 infected SLB1 and MT2 cells as well as a decrease in LC3-II levels. One mechanism of Tax induced blocking of the fusion might be by sequestering Beclin-1 to lipid rafts. The expression of a Tax mutant not capable of binding Beclin-1 or overexpression of Beclin-1 therefore might allow the formation of the autolysosome. Unfortunately, such a mutant would also fail to prevent the induction of autophagosome formation. An alternative might be the overexpression of a Beclin-1 mutant unable to bind Tax providing that such a mutant does not prevent the formation of the phagophore.

Tax induced autophagy has also been implicated in the degradation of phosphorylated IKKβ. Therefore it might be possible that the formation of autophagosomes via direct recruitment of Beclin-1 and Bif-1 is necessary in order to activate NF-κB early infection, whereas later in infection the inactivation of NF-κB mediated signalling is necessary for persistence and that the formation of the autophagosome indeed is beneficial. Unfortunately, so far it has not been shown if the induction of autophagy by Bcl-3 is accompanied by an increase in autophagic flux as measured by the degradation of p62/SQSTM-1. It should be noted however that cytosolic -but not nuclear- HTLV-1 Tax interacts with HDAC-6 and prevents the formation of stress granules. Since the deacetylation of microtubuli by HDAC-6 has been shown to be required for the fusion of autophagosomes with lysosomes, the expression of Tax-1 (at least the cytosolic form) might therefore increase the clearance of stress granules by autophagy induced via the recruitment of Beclin-1 to lipid rafts as well as facilitating the fusion of the autophagosome with the lysosome via interaction with HDAC-6. In the opinion of the author of these lines, the observed decrease in autophagy in cells infected with HTLV-1 might therefore be the result of the decrease of cytosolic Tax-1 -and maybe even nuclear Tax-1 during infection. This again would favour a "biphasic" model in which Tax induces autophagy early in infection whilst inhibiting autophagy at later stages.

Tax and autophagy: promotion of the formation of autophagosomes via
NF-ϰB dependent and independent pathways

HTLV-1 Tax and KSHV vFLIP: similarities and differences

Kaposi’s Sarcoma associated Herpesvirus (KSHV) or Human Herpesvirus 8 is a γ-Herpesvirus associated with two lymphoproliferative diseases, body cavity-based B-cell lymphoma/primary effusion lymphoma and multicentric Castleman’s disease. As discussed previously, the viral vCyclin-D and vFLIP proteins have to the ability to transform infected cells by promoting cell cycle progression whilst inhibition the induction of oncogene induced senescence (OIS) by way of inhibiting the formation of mature autophagosomes, in addition to NF-κB activation, the latter being reminiscent of Tax mediated activation of NF-κB. Both proteins have been shown to bind the regulatory subunits of I-κB kinases (IKK), NF-κB B essential modulator/IKKγ as well as activating IKKα and IKKβ, thus causing the activation of both the classical and alternative NF-κB signalling pathways. Indeed, KSHV vFLIP mediated activation of NF-κB has been demonstrated to be crucial for the survival of B-cell lymphoma/primary effusion lymphoma cells and the expression of Tax in mice and cell lines has been demonstrated to be required for the oncogenic activity of Tax. Similar to vCyclin, the expression of Tax induces senescence, which in this case is induced by (hyperactivated) p65/RelA and mediated by two Cyclin Kinase Inhibitors (CKI), p21^CIP1/WAF1 (p21) and p27^Kip1 (p27). Interestingly, similar to Tax, the expression of KSHV vFLIP in HeLa cells induces an arrest of the cell cycle in G1 phase in the absence of senescence concomitant with a decrease in I-κBα and an increase in p100, suggesting that both the classical and alternative NF-κB signalling pathway are activated, as well as an increase in both p21 and p27 levels. Co-expression of KSHV vCyclin however can abrogate vFLIP induced cell cycle arrest as well as Tax induced senescence probably by sequestering both p21 and p27 without affecting the ability of vFLIP and Tax to activate NF-κB dependent signalling. In contrast to Tax expressing HeLa cells however, cells transduced vFLIP in the presence of vCyclin do not present binucleation nor increased formation of micronuclei, which are commonly present in Tax expressing cells. Since the increase in both p21 and p27 is induced by activation of p65/RelA, the expression of ΔN-I-Kα (which cannot be degraded) as well as the transfection with shRelA prevents vFLIP induced increase of p21 and p27 and thus presumably vFLIP induced cell cycle arrest akin to cells expressing Tax.

In conclusion, both KSHV vFLIP and HTLV-1 Tax increase p21^CIP1/WAF1 and p27^Kip1levels via hyperactivation of RelA/ NF-κB; in the case of HTLV-1 Tax this induces senescence whereas in the case of KSHV vFLIP cells are arrested in G1 phase of the cell cycle. In the case of HTLV-1 infected cells, senescence is inhibited by the accumulation of mutations within the cellular genome leading to a deregulation of cellular G1 Cyclin dependent kinases and CKI, whereas in KSHV infected cells the expression of the viral Cyclin inhibits p21CIP1^/WAF1 and p27^Kip1induced cell cycle arrest by sequestering these proteins. It remains to be seen if the ability of Tax to induce the formation of mature autophagosomes is related to the induction of senescence and if inhibiting autophagy by e.g. by expression of the Atg3 binding domain of vFLIP or specific inhibitors prevents Tax induced senescence. Indeed, Tax induced overexpression of Bcl-3 might not only promote cell survival but also inhibit Tax induced senescence by inhibiting NF-κB. It is therefore interesting that the expression of KSHV vFLIP does not induce senescence - despite the hyperactivation of NF-κB. Future experiments might shed some light of the complex interactions between these oncogenic proteins, senescence, autophagy, and NF-κB.

Sequestering of Beclin-1 and Bif-1 by Tax: inhibition of fusion of the autophagosome
with the lysosome?

Tax not only induces the hyperactivation of NF-κB via facilitating the degradation of the IKK complex but also via attenuating the ATM dependent DNA damage response by sequestering DNA repair factors which can be inhibited by Bcl-3. Constitutive activation of ATM results in an intranuclear activation of NEMO and translocation of the ATM/NEMO complex to the cytoplasm where in subsequent steps the IKK complex is degraded and the p65/p50 complex is activated. It remains to be seen to which extent this pathway is responsible for Tax induced autophagy, in particular early in infection prior to the establishment of persistently infected cells as well as in cell transformation.

The expression of Tax has also been linked to the prevention of TRAIL induced apoptosis in U251 astroglioma cells by inducing the expression of c-FLIP, the cellular equivalent of KSHV vFLIP, as well as inducing the formation of autophagosomes. In this model, the expression of c-FLIP via Tax prevents the cleavage of Beclin-1 by preventing the cleavage of Caspase-9/-3. Tax therefore not only promotes the formation of autophagosomes by sequestering Beclin-1 to lipid rafts and increasing the expression of Bcl-3 but also by preventing the activation of Caspase-3 via inducing c-FLIP expression. Interestingly, the expression of KSHV vFLIP also inhibits Caspase-3 cleavage, suggesting that the expression of KSHV vFLIP in Tax expressing cells depleted of cFLIP might increase the resistance of these cells to TRAIL induced apoptosis – hypothetically.

Tax and autophagy: central role of NF-ϰB in inducing autophagy
and preventing apoptosis

One final question is if the expression of Tax also mediates the secretion of neuroinflammatory factors since autophagy has also been implicated in the secretion of IL1B (Interleukin 1-beta), CXCL8 (Chemokine (C-X-C motif) ligand 8), LIF (Leukemia inhibitory factor), FAM3C (family with sequence similarity 3, member C), and DKK3 (Dickkopf WNT signaling pathway inhibitor 3). In order to answer this question, not only is it necessary to characterise the secretome of Tax expressing cells, but also determine if Tax indeed increases autophagic flux. So far the data obtained from HeLa cells expressing Tax indicate that the fusion of mature autophagosomes with lysosomes is inhibited thus increasing the stability of Tax. Consequently, the induction of senescence in Tax expressing cells is independent of the induction of autophagy and thus different from cells expressing KSHV vCyclin.

Wednesday, 21 January 2015

Porcine Respiratory Syndrome Virus (PRRSV) and autophagy: a link between the ER stress response and p53?

Porcine reproductive and respiratory syndrome virus (PRRSV) is an enveloped virus with a positive strand ssRNA genome of approx. 15kB in length, encoding for ten open reading frames (ORFs).

Similar to the genome of the Coronaviridae, the PRRSV genome ORF1a and 1b genes encode for a RNA-dependent RNA Polymerase as well as for a number of non-structural proteins (nsp; 12 in total) which are generated by autoproteolytic cleavage by a virally encoded cysteine protease (nsp2), 3C-like serine protease (nsp4), and papain-like cysteine protease nsp1α/1β. As it is the case for the Coronaviridae, the remaining nsp’s of ORF1 encode for enzymes required for the replication of the viral genome, including a viral RNA Helicase (nsp-10) and endonuclease (nsp-11).

The remaining ORFs 2-7 encode for ORF2a (GP2a), ORF2b (E), ORF3 (GP3), ORF4 (GP4), ORF5a (GP5a), ORF5b (GP5b), ORF6 (M), and ORF7, the viral Nucleocapsid (N) which akin to the coronaviral N protein localises to the nucleolus and is phosphorylated.

PRRSV virus

PRRSV genome: ORFs and nsp's

In PRRSV infected MARC-145 cells, the viral nsp-2 protein is localised in the perinuclear region resembling a localisation at the ER, akin to the coronaviral proteases, thus suggesting that the expression of of nsp-2 may induce autophagy and a ER Stress response as described for the PLP2 and PLpro proteins derived from CoV. Indeed, PRRSV nsp-2 has been described to contain a DUB domain as well as inhibiting the Interferon response although an interaction with STING has not been demonstrated yet. Although the data for the intracellular localisation of PRRSV structural proteins are incomplete, in the case of the related Equine Arterivirus (EAV), in BHK-21 cells the EAV E protein (encoded by ORF 2a) predominantly localises to the ER and to a lesser degree to the Golgi complex, whereas the viral G_Llocalises to the Golgi.

As described before for other positive strand RNA viruses such as Chikungunya Virus, Measles Virus, Coxsackie B Virus, Coronavirus as well as EMCV, the expression of viral proteins induces the formation of replication transcription centers (RTCs), double membrane vesicles (DMV) which contain the viral RNA (both ssRNA and dsRNA intermediates) as well as the enzymes required for transcription including the viral RNA dependent RNA Polymerase and RNA Helicase. Commonly the DMV derive from the ER in a process subverting the autophagy pathway. As described before autophagy -which involves the formation of mature autophagosomes that fuse with lysosomes, ultimately leading to the degradation of the proteins localised within the autophagosome. Alternatively, the contents of autophagic vesicles might be secreted or in the case of viral proteins be processed to be displayed by MHC-Class I and MHC-Class II molecules. As discussed before, viral proteins -sometimes the same which promote autophagy as for instance the CoV nsp-6 protein- not only promote the formation of DMV but also inhibit the formation of mature autophagosomes and/or the fusion of autophagosomes with lysosomes.

In the case of PRRSV, the viral nsp-2. nsp-3, and nsp-5/6/7 proteins have been demonstrated to localise to the ER and the expression of of the nsp-5/6/7 protein induces the formation of GFP-LC3 positive vesicles, indicating the induction of the formation of autophagosomes. Akin to the CoV nsp-6 protein, the expression of nsp-5/6/7 protein in Vero cells has been postulated to inhibit the fusion of the autophagosome with the lysosome. In the case of both nsp-2 and nsp-3 however the formation of autophagosomes has not been demonstrated (to my knowledge) although in MARC-145 cells and porcine pulmonary alveolar macrophages infected with PRRSV, LC3-II positive autophagosomes accumulate 24 hrs p.i. whilst the fusion with the lysosome is inhibited since the application of Chloroquine does not increase the number of GFP-RFP LC3-II positive punctae nor the levels of LC3-II as measured by western blot. Contrary to these results however, the treatment of PRRSV infected MARC-145 cells with Bafilomycin-A1 suggest that at 120 hrs p.i. PRRSV titers are decreased compared to mock treated cells and that the levels of p62/SQSTM-1 in PRRSV infected cells are lower than in non-infected cells. The difference observed might be due to the experimental conditions since Bafilomycin-A treatment lasted for 48 hrs compared to 6 hrs for Chloroquine treatment as well as different virus strains (PRRSV JXwn06 v. VR-2385), so more experiments are needed to address this issue.

Similar to the RTC induced following the infection of BHK-21 cells with EAV, these vesicles contain the viral nsp-2 and N protein although the presence of dsRNA has not been demonstrated to my knowledge. In contrast to EAV infected MEF, the autophagic machinery however is required for PRRSV replication as viral titers are significantly lower in MARC-145 transfected with shLC3B, siATG7, siBeclin-1 or shATG5, suggesting that PRRSV -in contrast to EAV- does not induce the formation of autophagy-like vesicles via the ERAD pathway but via the induction of the phagophore via the ATG5/ATG7/Beclin-1 pathway; if however EDEMosomes are formed during PRRSV infection remains to be seen. Viral replication can also be induced by treating cells with Rapamycin, thus inhibiting mTORC1 and promoting autophagy, whereas treatment with 3-Methyladenine (3-MA) decreases viral titers.

Interestingly, the infection of MARC-145 with PRRSV strain VR-2385 activates mTORC1 (and thus inhibits autophagy) at early times post infection (6 h p.i.). So far the impact on viral or starvation induced autophagy has not been investigated, but the author these lines suggests that PRRSV inhibits autophagy at early timepoints p.i. whereas at later timepoints the formation of autophagy like vesicles is induced. This hypothesis is supported by results indicating that PIK-K-Akt kinase signalling is modulate by PRRSV in so far as phosphorylated Akt kinase levels increase at earlier timepoints, but decrease at 12 hrs p.i. . It is however crucial to compare proteins derived from highly virulent strains to those derived from attenuated or less pathogenic strains.

Induction of p53 and DRAM-1 dependent autophagy via the ER stress by
PRRSV: hypothetical model

PRRSV and the ER stress response: does nsp-2, nsp-4, or nsp-5/6/7 induce the
ER stress response?

It remains therefore to be seen if the expression of PRRSV proteins increases the formation of autophagosomes and/or autophagy-like vesicles similar to the coronaviral nsp-3/-4/-6 proteins whilst inhibiting the fusion of the lysosome. Also, it remains to be seen if the expression of PRRSV nsp-2 -and other viral proteins including nsp-5/6/7 induces the ER stress response by lipid depletion and subsequent autosis. Interestingly the infection of MARC-145 cells with PRRSV strain CH-1a results in a PERK and IRE1 induced ER stress response whose inhibition is associated with decreased viral replication. Since autophagy is induced fooling the activation of the ER stress response it seems possible that the decrease in viral replication is due to a decrease in autophagy or alternatively to apoptosis (autophagy dependent or independent). In this case, the activation of the ER stress response might induce p53 and thus DRAM-1; indeed, the inhibition of p53 has been demonstrated to decrease viral titers, but so far no link has been established between PRRSV, the ER stress response, p53, and autophagy.

Interplay of the induction of Akt, Akt dependent inhibition of autophagy and induction

of autophagy via the ER stress response: activation of Akt early during the infection, induction of the ER

stress response late in infection?