Cytoplasmic heat shock protein 90 (Hsp90), which is an indispensable component of every eukaryotic cell, is represented in vertebrate cells by two closely related proteins, called Hsp90α and Hsp90β. Hsp90 is involved in the chaperoning of 200–300 clients proteins, and it is among the most abundant proteins in the cell [1
]. Hsp90 clients include many proteins important for the regulation of cellular processes, such as hormone receptors, transcription factors, and protein kinases [3
]. Many of the Hsp90 clients are involved in the development and progression of cancers, which makes this protein an attractive target for pharmacological intervention. The chaperoning activity of Hsp90 depends on the binding and hydrolysis of ATP by the N-terminal domain of this protein. Most of the identified inhibitors of Hsp90 binds to the ATP-binding pocket of ATPase. One of the first ATPase inhibitors of Hsp90 successfully used to inhibit Hsp90 chaperone activity in cultured mammalian cells was 17-N-allylamino-17-demethoxygeldanamycin (17-AAG) [4
]. 18 Hsp90 inhibitors reached different stages of more than 170 clinical trials as a potential drugs in humans [5
]. The dose-limiting toxicity was a major problem reported for the compounds tested so far. No clinical trials of the antiviral activity of the Hsp90 inhibitors have been conducted so far. The Hsp90 inhibitors interfered with the replication of many viruses that were tested in vitro, indicating that it might be possible to fight viral infections through Hsp90 inhibition. Interestingly, viruses tested so far appear to be sensitive to the non-toxic doses of Hsp90 inhibitors [7
]. The replication of a virus requires the production of a large quantity of several types of proteins. These proteins often require help from cellular chaperones in proper folding and protection from aggregation. Hsp90 is involved in the replication of many viruses at different stages of the replication cycle by facilitating virus particle entry in the cell, intracellular transport, expression, and the stabilization of viral proteins, and genome replication [9
]. Hsp90 may also be required for the virus assembly and trafficking [12
]. Interestingly, the requirement for Hsp90 chaperone seems to be universal for the replication of viruses that belong to different taxonomic groups, but the function of Hsp90 seems to be specific for each virus [14
Human adenoviruses (HAdV) belong to the Adenoviridae
family and they are classified in the Mastadenovirus
genus. Non-enveloped icosahedral virions of human adenoviruses are 70 to 90 nm in diameter with over 30 proteins encoded in a 35 kbp long double-stranded DNA. Human adenoviruses are divided into seven species (Human mastadenovirus A-G
) with best-studied HAdV-2 and HAdV-5, both belonging to the species C [17
]. Adenovirus infection is mostly associated with respiratory tract disease and conjunctivitis, while gastrointestinal and urinary tract disease less commonly occurs. Primarily associated with both occasional cases and epidemic infections among children, HAdVs now emerged as opportunistic pathogens causing significant morbidity and mortality in the immunocompromised population [20
]. In immunocompetent hosts, adenovirus infections are usually mild and self-limiting, with the rare need for medical intervention.
HAdV-5 infection begins with the virus binding to the cell membrane, through the interaction of fiber protein with Coxsackievirus–adenovirus receptor (CAR), CD46, or sialic acid [22
]. Subsequently, the penton base protein binds to integrins from αv
β family that serve as an entry receptor and the virus is internalized in the endosomes by receptor-mediated endocytosis [26
]. The transport of the viral DNA is usually completed in less than 1 h [28
E1A RNA is the first to be transcribed after HAdV DNA enters the nucleus [29
]. E1A proteins, which are translated from differentially spliced mRNAs, serve as co-activators of the remaining early promoters (E1B, E2A, E2B, E3, and E4) and regulate the transcription of many cellular genes [30
]. E2 region encodes the proteins necessary for AdV DNA replication. E2A transcript translates to DNA-binding protein (DBP) and E2B transcript encodes polymerase and precursor terminal protein [31
The subsequent transcription of the early AdV regions E3 and E4 results in the production of proteins active in inhibition of apoptosis and suppression of intracellular immune response and activation of the late promoter L4 [32
DNA replication is initiated by the products of the E2 region. DNA replication also activates a transcription of the major late transcript from the late promoter. This transcript is alternatively spliced into several mRNAs that encode hexon, penton, fiber, and other structural proteins of the AdV capsid [38
]. After the replication and capsids assembly is completed, virus is released by cell lysis [39
HAdV infection leads to increased transcription of HSP27
, and HSP90
]. Hsp70 interacts with adenoviral capsid proteins [42
]. However, the specific function of heat shock proteins in HAdV replication was not studied. Therefore, in the present work, we decided to investigate the possible role of Hsp90 in HAdV-5 replication.
Viruses utilize host cell cellular mechanisms to synthesize a large number of proteins that are involved in their replicative cycle. The inhibition of chaperoning activity of Hsp90 during infection prevents or suppresses the replication of many viruses that belong to different groups [46
The aim of this study was to evaluate the effect of 17-AAG, the Hsp90 inhibitor, on human HAdV-5. We demonstrated that 17-AAG exerted a strong, concentration-depending, inhibitory effect on HAdV-5 replication at concentrations that did not affect cell viability. This effect was especially pronounced when the inhibitor was applied at the time of infection, which suggested that Hsp90 is required at the early steps of HAdV-5 replication.
Hsp90 inhibition does not influence the expression of the receptors that are necessary for HAdV-5 entry into the human mesothelioma JMN-1B cells [51
]. We confirmed that this is also true for A549 cells, because, in 17-AAG treated cells, there was no decrease in the expression level of CAR and integrin αv
, the receptors that are necessary for this process. Moreover, the synthesis of the viral proteins was inhibited by 17-AAG, even 9–12 h after infection, when the viral DNA reached the nucleus and transcription of the viral genes begun.
The time-course analysis of transcription revealed that 17-AAG inhibits the expression of the HAdV-5 early genes E1A and DBP at the time of infection, but it seems to be relatively ineffective when applied later. The expression of HAdV genes begins with the E1A transcription [52
]. The E1A protein is necessary for the efficient transcription of other early HAdV-5 mRNAs and stimulates its own transcription [53
]. Immediately after infection with the virus, the E1A transcription is catalyzed by the cellular proteins. Hsp90 inhibition did not affect this early transcription of the E1A gene, but the E1A protein level was decreased, which suggested that Hsp90 chaperones E1A protein. This conclusion was further supported by the Hsp90α-E1A association detected by co-immunoprecipitation. However, the E1A protein constitutively expressed in HEK 293 cells was not affected by the Hsp90 inhibition. Therefore, it seems that Hsp90 stabilized and protected from degradation the newly translated, but not the mature, E1A protein that was present in the cells before they were exposed to the inhibitor. The conclusion that 17-AAG affects only de novo
expressed E1A protein was supported by the observation that 17-AAG did not increase the decay rate of E1A in HEK 293 cells after the protein synthesis was inhibited by cycloheximide.
Recently, the anti-HAdV activity of mifepristone was reported, attributed to the interference with steps preceding an entry of the virus genomic DNA into the nucleus [54
]. The study of three salicylanilide anthelmintic drugs demonstrated that two of these compounds inhibit the HAdV life cycle by restricting access of the viral DNA to the nucleus, whereas the third one inhibited HAdV replication by decreasing E1A transcription [55
]. All of these compounds were effective within 1 h after infection. The data presented here demonstrated that the Hsp90 inhibitor effectively limited HAdV-5 replication much later after infection. The expression of mRNAs for the proteins necessary for the viral genome replication, polymerase DBP, and PTP is activated by E1A. Therefore, the viral DNA replication depends indirectly on E1A. The decreased rate of the genome replication results in decreased production of the late viral proteins, not only due to the lower number of the gene copies, but also because the viral DNA replication activates the transcription of late promoter activator IVa [56
The expression of capsid proteins, controlled by the late promoter, was especially sensitive to Hsp90 inhibition late after infection. These proteins are the last components of the virus to be synthesized. The expression of late mRNAs begins with the activation of the L4 promoter that drives the expression of L4-22K and L4-33K proteins [57
]. These proteins are essential activators of the full set of late mRNAs [58
]. L4 promoter is activated by the viral proteins E1A, E4 Orf3, and IVa2 [37
]. The decreased E1A expression is a limiting factor for the viral capsid protein’s expression, not only directly, but also indirectly, because E1A also stimulates the expression of IVa2 and E4 orf3 [59
A recently published study on HAdV inhibition by ivermectin, a drug preventing E1A protein from entering the nucleus, reports similar effects on the virus mRNA and protein expression and DNA replication, resulting in the decreased production of viral progeny similar to the reported here [61
]. However, ivermectin inhibits E1A expression more effectively 24 h and 36 h after infection, whereas Hsp90 inhibition is most effective up to 9 h after infection.
Although there is a number of investigations focused on the role of Hsp90 in supporting viral replication, intracellular antiviral response, and virus trafficking, there are only limited data available concerning the chaperoning activity of Hsp90 for early viral activators of the replication process. Basha and collaborators presented an impact of geldanamycin (GA) on the expression of immediate early (IE) and major immediate early genes of human cytomegalovirus (CMV) [62
]. There was a delay in IE2 (but not IE1) protein synthesis, and a significantly lower amount of IE2 was produced. The addition of GA at early steps of infection (0–8 hpi) lead to the effective inhibition of immediate early genes, which also led to a decreased synthesis of the second tier of major immediate early genes. Interestingly, following applications of GA doses during changes of cell culture medium led to the complete inhibition of CMV replication. Another work, by Katsuma, revealed the dependency of baculoviral IE protein on Hsp90 chaperone function [63
]. Similar to our observations, treatment with 17-AAG did not affect the initiation of IE gene transcription, but it had a significant negative effect on stable IE protein synthesis. In both cases, conclusions indicated a fundamental role of Hsp90 in supporting viral replication via chaperoning of immediate early genes, while inhibition of this process resulted in disruption of viral gene expression cascade and eventually led to the delay or inhibition of the entire process of virus replication.
Contrary to the above-mentioned drugs that had anti-adenoviral activity, none of the Hsp90 inhibitors was approved for use in humans. This however may change with the new inhibitors being developed and numerous trials conducted. The data reported here demonstrated that 30 nM 17-AAG effectively inhibited HAdV-5 replication in vitro. 17-AAG antiviral activity at the non-toxic concentration was also reported in other studies [64
]. The 17-AAG in plasma of the patients during the clinical trials reached 6–16 μM concentration, depending on the administered dose, with moderate adverse effects [67
]. A lower concentration of the drug necessary to suppress viral infection may limit its toxic side effects. Antiviral drugs tend to lose effectiveness due to drug-resistant mutations. This may not be the case for the Hsp90 inhibitors, because a protein that depends on the Hsp90 chaperone for maturation and stability is not likely to be converted to the chaperone-independent and still functional variant by the simple mutation. There are known Hsp90 mutations that are resistant to ATPase inhibitors, but such mutations might occur in a limited number of cells, and they would not have an impact on the viral infection progress at the whole organism level [69
Modified adenoviruses are widely used as vectors for DNA delivery into mammalian cells and HAdV modified to target tumor cells are studied as a potential means to treat cancers [70
]. The applications of Hsp90 inhibitors in cancer treatment are also studied. The possible interference between the clinical application of Hsp90 inhibitors and HAdV-based agents should be considered. The data presented here suggest that replication-competent oncolytic adenoviruses may be particularly sensitive to the adverse effects of the Hsp90 inhibitors. The HAdV infections tend to be mild, but they can be serious in young and immunocompromised individuals, and specific drugs to treat such infections are lacking [73
]. Our results indicate that Hsp90 inhibitors could be used to suppress the adenoviral infection.
4. Materials and Methods
4.1. Cell Lines and Virus Infection
The human epithelial cell line derived from lung carcinoma (A549) and human embryonic kidney 293 cells (HEK293) were obtained from ATCC and grown in Iscove’s Modified Dulbecco’s Medium (IMDM) that was supplemented with 10% fetal bovine serum, 100 U of penicillin, and 100 µg of streptomycin/mL (Sigma). Transfections were performed using Metafectane (Biontex), as suggested by the manufacturer. Human adenovirus 5 (VR-5) was obtained from ATCC.
In order to determine the effect of 17-AAG on HAdV-5 replication, A549 0.8 × 106 cells/well were seeded in a six-well plate and then infected with the virus at the indicated titer. The inhibitor was added at the specified time concerning infection. Cycloheximide was used at 100 μg/mL concentration and 17-AAG was 0.25 μM, unless specified otherwise.
The virus titer was measured by 50% tissue-culture infectivity endpoint (TCID50
) method of Reed and Muench [76
4.2. Cell Viability Assay
Cell viability was determined using the Cell Counting kit–8 (Sigma). The cells were seeded into a 96-well plate in IMDM medium with different concentration of 17-AAG (0, 0.125, 0.25, 0.5 μM). After 72 h, cell viability was measured according to manufacturer instructions.
4.3. Plasmid Construction
E1A 289R DNA was amplified by RT-PCR from a total RNA isolated from HAdV-5 infected A549 cells using RNAzol and then converted to cDNA with hexamer primers in a reaction with AMV transcriptase. Primers E1A289F and E1A289R were used in this reaction. The resulting DNA fragment was cloned in the plasmid pcDNA-Myc using Kpn I and Xho I restriction sites that were incorporated in the sequence of the primers. The resulting plasmid expresses E1A 289R protein with the C-terminal Myc-tag from CMV promoter.
A plasmid that expresses the Flag-tagged Hsp90α gene in human cells was described earlier [69
]. The Hsp90α E46A mutation was generated by PCR mutagenesis
4.4. Immunofluorescence Microscopy
Prior to infection, A549 cells were seeded on coverslips and cultured overnight to adhere. The medium was removed and replaced with IMDM containing 0.25 μM 17-AAG and with HAdV-5 at 500 TCID50/mL and cultured for 24 h followed by 4% formaldehyde fixation, 0.1% Triton X-100 permeabilization and blocking with PBS containing 3% bovine serum albumin (BSA). The fixed cells were then stained with an anti-HAdV-5 rabbit polyclonal antibody (Abcam). Goat anti-rabbit secondary IgG antibody conjugated with Alexa Fluor Plus 598 (Invitrogen) was then added and DAPI was used to stain the nuclei.
4.5. Co-Immunoprecipitation Assay (Co-IP)
HEK293 cells were transfected with E1A 289R-Myc and Hsp90α E46A plasmids, while the control cells were transfected with pcDNA-Myc and Hsp90α E46A plasmids. After 48 h, cells were harvested and lysed with IP buffer (0,25% Triton X-100, 10 mM Tris, 20 mM NaF, 100 mM, 10 mM β-glycerol phosphate, 2 mM sodium orthovanadate, 5 mM ATP, and protease inhibitors cocktail (Roche)). The lysates were cleared by centrifugation at 12,000× g for 15 min. at 4 °C. The protein concentration was measured using the BCA assay (Sigma), and adjusted with IP buffer to 1 mg/mL. 10 μL anti-Flag agarose beads (Pierce) were added to the supernatant (700 μL), and then incubated for 2 h at 4 °C with mixing. The immunoprecipitates were washed with ice-cold PBS four times and eluted with 40 μL 1× SDS PAGE Loading buffer. The samples were boiled for 10 min. and analyzed by western blot.
4.6. Western Blot Analysis
The proteins were extracted by lysis with RIPA buffer. Rabbit polyclonal antibody for the HAdV-5 (ab6982) and for E1A (ab204123) were obtained from Abcam. Monoclonal antibodies were purchased from: Flag (Sigma, F3165), Myc (Merck, MABE282), CAR (Cell Signaling, 5670S), and integrin Vα (Cell Signaling, 60896S). Secondary antibodies that were conjugated to Alexa488 (Invitrogen, A32723) and Alexa594 (Invitrogen, A32740) were obtained from Abcam. Goat anti-rabbit IgG-HRP and Goat anti-mouse IgG-HRP antibodies were obtained from Bio-Rad (cat. no. 170-6515 and 170-6516). Western blot and immunofluorescence staining were performed according to the standard protocols with the antibodies diluted, as recommended by manufacturers.
2 μg of RNA extracted with RNAzol reagent (Sigma) was used for cDNA synthesis while using hexamer primers and AMV reverse transcriptase. After synthesis, polymerase was inactivated and the reaction mixture was diluted with nine volumes of water. 1 μL of the cDNA was used as a template in a 20 μL PCR reaction with primers specific for E1A, DBP, hexon, and GAPDH (listed in Supplementary Table S1
). The cDNA obtained as described above was used as a template in a qPCR reaction with primers and TaqMan probes listed in Supplementary Table S1
. The results were expressed as the relative copy number of HAdV-5 mRNA or DNA normalized to GAPDH and G6PD, and then for a number of cells used for RNA extraction. QPCR for the viral DNA was performed using primers Hexon-F, Hexon-R, and Hexon probe.
4.8. Statistical Analysis
The one-tailed Student’s t-test was used for data analysis, with the significance set at 0.05.