High Permissivity of Human HepG2 Hepatoma Cells for Influenza Viruses

Abstract

Human HepG2 hepatoma cells are highly permissive for influenza virus type A and type B, even without the addition of trypsin, and they exhibit a marked cytopathic effect. This property greatly facilitates the primary isolation of influenza viruses. Virus replication was significantly reduced by the plasmin(ogen)-specific inhibitor tranexamic acid, and this suggests a potential role played by the plasminogen/tissue plasminogen activator complex at the surface of HepG2 cells. This might represent a new approach for study of the interrelations of this complex with influenza viruses.

Various techniques are available for rapid detection of influenza viruses (19, 20). However, continuous isolation of new strains remains indispensable for epidemiologic studies and for the production of up-to-date vaccines of appropriate composition. Cultivation of influenza viruses in embryonated hens' eggs is the method of choice for the production of large amounts of antigens (21). Propagation of influenza viruses in cell lines requires the addition of trypsin to the culture medium to ensure cleavage of the viral hemagglutinin precursor (HA0), a prerequisite for the virus to be infectious (11). The MDCK cell line is generally used for primary isolation from humans (10, 17). Depending on the positive sample tested, this system shows a very weak cytopathic effect (CPE), if any, and repeated blind detections of virus production are thus necessary (17). HepG2 cells are known to express tissue plasminogen activator (tPA) (15). Plasminogen, which is also expressed by HepG2 cells (14), is present in fetal calf serum (FCS), which usually supplements the cell culture medium. The probability that plasmin, a trypsin-like protease known to activate HA0 (12), is thus potentially present in the supernatant of HepG2 cell cultures prompted us to examine the ability of these cells to support the replication of influenza viruses. Results revealed that HepG2 cells are highly permissive and show a marked CPE. This should facilitate the screening of positive clinical samples during attempts at primary isolation.

HepG2 (ATCC HB-8065) and MDCK (ATCC CCL-34) cells were grown in Dulbecco's modified Eagle's medium (Gibco, Invitrogen, Cergy-Pontoise, France) supplemented with 10% FCS (Gibco), penicillin (100 U/ml), and streptomycin (50 μg/ml) and distributed in culture tubes (5 × 10⁴ cells/tube). Nasal or throat swabs were diluted in 2 ml of Hanks' balanced salt solution and spun at 4,500 × g for 15 min. Subconfluent HepG2 and MDCK cells were inoculated with 100 μl of sample and incubated for 2 h at 35°C. Samples were then removed, and each tube was mixed with 1 ml of Dulbecco's modified Eagle's medium and 20 μg of ciprofloxacin (Bayer Pharma, Puteaux, France). HepG2 cells were supplemented with 2% FCS, and MDCK cells were supplemented with 10 μg of TPCK-treated trypsin [trypsin treated with l-(tosylamido-2-phenyl)ethylchloromethylketone (TPCK) to inhibit contaminating chymotryptic activity (9); Worthington, Serlabo, Bonneuil sur Marne, France]/ml and 2% Ultroser G (Biosepra, Cergy Saint Christophe, France) as serum substitute (2, 22). Both cell types were successfully maintained strongly attached to the support for more than 14 days.

The sensitivities of HepG2 and MDCK cells for influenza viruses were compared during two periods, January to March 2001 and January to March 2002, in a total of 82 samples (52 pediatric patients and 30 adults) received in our department for primary isolation. Tubes of subconfluent cells of each type were inoculated with 100 μl of each sample. In HepG2 cells, a CPE became obvious within 2 to 6 days for type A influenza virus replication and 3 to 13 days for type B. The CPE caused the cells to become granular and then fragmented. This was followed by detachment from the support and complete destruction (Fig. 1A). In contrast, MDCK cells showed only a very weak CPE, if any. Influenza virus replication in these cells was determined using the Monofluo influenza virus kit (Bio-Rad, Marnes-la-Coquette, France) and was detected several days after the CPE exhibited in HepG2 cells. To correlate CPE with the growth of influenza viruses in HepG2 cells, various viral concentrations, titrated by a plaque assay, were inoculated. At day 3 postinfection, CPEs were photographed. As shown in Fig. 1B, the degrees of CPE and respective virion production were correlated with the inoculum doses. The growth curves of viral production by HepG2 and MDCK cells were determined after inoculation with an H3N2 strain (A/Panama/379/99), by using a plaque assay. As shown in Fig. 1C, higher titers were observed in the supernatants of infected HepG2 cells. All influenza virus strains were isolated with both cell types, but only HepG2 cells exhibited a clear CPE with all of the positive samples (Table 1). Furthermore, the time necessary to detect influenza virus strains was shorter in HepG2 cells than in MDCK cells: 4 days versus 7 days for influenza virus A strains and 8 days versus 10 days for influenza virus B strains, respectively (Table 1). The results were statistically found to be quite significant (Fig. 2). As the HepG2 system exhibits a clear CPE, it could be used to confirm positive isolation and thereby obviate the need for repeated blind detections of influenza virus antigens. After primary isolation, all of the influenza virus strains that we obtained were successfully serially propagated in HepG2 cells. The period of latency to CPE appearance was progressively shortened to just 1 day. This allowed ready preparation of virus stocks.

FIG. 1.

CPE caused by the replication of H3N2 influenza A virus strain Panama/379/99 and influenza B virus strain Sichuan/379/99 in human HepG2 hepatoma cells. (A) CPEs were photographed on days 2, 5, and 8 postinfection. Infected MDCK and mock-infected cells are shown. (B) Correlation between degrees of CPE and production of H3N2 influenza virus virions in HepG2 cells. (C) Growth curves (means of triplicates) of virion production by HepG2 and MDCK cells (5 × 10⁴) inoculated with 10² PFU. All titrations were performed by a plaque assay.

TABLE 1.

Primary isolation of influenza virus strains A/H1N1, A/H3N2, and B by using HepG2 and MDCK cells^a

Strainb	No. of positive samples (n = 82)		Mean time (days) to detection of CPE (HepG2) or antigen (MDCK)		CPE
	HepG2	MDCK	HepG2	MDCK	HepG2	MDCK
A/H1N1	8	8	4 ± 2	7 ± 3	++++	Weak
A/H3N2	12	12	4 ± 2	7 ± 3	++++	Weak
B	15	15	8 ± 5	10 ± 5	+++	Negative

^aSample positivity was judged by the appearance of CPE in HepG2 cells and thereafter confirmed by the detection of viral antigens. This last method, routinely repeated, was the only technique valuable for MDCK cells.

^bA/H1N1, strain A/New Caledonia/20/99; A/H3N2, strain A/Panama/2007/99; B, strain, B/Sichuan/379/99.

FIG. 2.

Statistical comparison of mean values summarized in Table 1, corresponding to the time (days) necessary for the appearance of CPE in influenza virus-infected HepG2 cells and the detection of viral antigens in infected MDCK cells.

To compare the replication rates of influenza viruses in HepG2 and MDCK cells, a real-time reverse transcription-PCR (RT-PCR) was carried out using a LightCycler instrument (Roche Diagnostics, Meylan, France) with two hybridization probes, according to the principle of fluorescence resonance energy transfer (FRET). An identical number of cells (5 × 10⁴) of both types was inoculated with 10³ PFU of an H3N2 strain (A/Panama/379/99). Cells were harvested on day 3 postinfection, and RNAs were extracted using the QIAamp Blood DNA/RNA minikit (Qiagen, Courtaboeuf, France). A 447-bp-long fragment was amplified using the LightCycler RNA Master Hybridization Probes kit (Roche Diagnostics) with 30 ng of RNAs, carefully measured, and the primer pair H3-S (5′-GAGCATCTATCAGTGTGCAG-3′) and H3-AS (5′-CCAAGAGAAGCCAGCAAACT-3′). Following amplifications, we measured the FRET signal generated by hybridization of the two probes H3-P5 [5′-AATGACAAATTTGACAAATTGTACATTTGGGG(Fluo)-3′] and H3-P3 [5′-(Red640)GTTCACCACACGAGTACGGACAGTGAC(phosphate)-3′]. Three independent influenza virus productions were tested, as described above. As shown in Fig. 3A, the cycle in which each reaction with influenza virus RNAs from both cell types first arose above the background was earlier with HepG2 than with MDCK cells. Moreover, the FRET signal was approximately 10 times greater with influenza virus RNAs from HepG2 cells than with those from MDCK cells, thus confirming the higher permissivity of these cells for influenza viruses. Identical experiments were performed to compare the production of a type B influenza strain (Sichuan/379/99) in HepG2 and MDCK cells, by using the primer set B-S (5′-ATAGTTTCCCTCTGGTCTTTGTT-3′) and B-AS (5′-AATCCACTAACAGTAGAAGTACC-3′). To take into account the longer period of time (8 to 10 days) required to detect B strain replication, RNAs were extracted on day 7 postinfection. Amplifications were followed by measurement of the signal generated by hybridization of the two probes B-P5 [5′-CTTGCCCTAACTCTACCAGTAAAAGCGGATTT(Fluo)-3′] and B-P3 [5′-(Red640)TCGCAACAATGGCTTGGGCTGTCCCAA(phosphate)-3′]. As for type A influenza virus, HepG2 cells showed a markedly higher permissivity for type B influenza virus than MDCK cells did (Fig. 3B).

FIG. 3.

Quantitative and comparative evaluation of influenza virus production in HepG2 and MDCK cells. Three independent influenza virus productions were evaluated using a real-time RT-PCR method (FRET principle) with 30 ng of total RNAs, carefully measured, from each type of cells infected with 10³ PFU. (A) Cells infected with H3N2 influenza A virus (strain Panama/379/99). (B) Cells infected with influenza B virus (strain Sichuan/379/99). (C) Inhibitory activity of 1 and 10 mM tranexamic acid on influenza A virus replication in HepG2 cells, evaluated on day 3 postinfection.

Keskinen et al. (8) observed that HepG2 cells do not show any detectable interferon expression in response to infection with influenza A virus, even though the latter is known to be a very good inducer of alpha interferon (18). These authors analyzed this property over a period of 48 h to monitor viral protein production but did not attempt to investigate the subsequent appearance of CPE. The absence of interferon response might partially explain the greater permissivity of HepG2 cells.

The first likely hypothesis to explain the constitutive ability of HepG2 cells to support the productive replication of influenza viruses without the need for the addition of trypsin is based on the fact that these cells express tPA, which is capable of activating endogenous and/or exogenous plasminogen. Using RT-PCR, we verified that the HepG2 cell line used in our laboratory expressed tPA (results not shown). The lysine analog ɛ-aminocaproic acid, which inhibits the activity of plasmin, was able to decrease the replication of influenza viruses by HepG2 cells. To obtain more accurate results, we used tranexamic acid, a specific inhibitor of plasmin (7) that has also been shown to block the lysine-binding sites of plasminogen and can thus prevent the activation of plasminogen by tPA (1, 3). Three independent influenza virus productions were tested in the presence of 1 and 10 mM tranexamic acid (Sigma, St. Louis, Mo.) by the real-time RT-PCR technique described above. As shown in Fig. 3C, at day 3 postinfection with 10³ PFU of H3N2 (strain A/Panama/379/99), influenza virus replication was significantly reduced by 1 and 10 mM tranexamic acid. In comparison with control HepG2 cells, the first PCR cycle in which significant RNA amplification appeared was delayed by five and eight additional cycles when the tranexamic acid inhibitor was added at 1 and 10 mM concentrations, respectively (Fig. 3C). We therefore concluded that the plasminogen-plasmin system was implicated in the activation of influenza viruses for HepG2 infection. This indirectly demonstrated the role of tPA, which is indispensable in the tPA→plasminogen→HA0 activation cascade.

Nevertheless, the precise mechanism of activation of HA0 by HepG2 cells remains unclear. No plasmin activity was found in culture supernatants of mock-infected HepG2 cells by using N-p-tosyl-Gly-Pro-Lys 4-nitroanilide acetate (Sigma), a specific chromogenic substrate of plasmin (13). Moreover, the culture supernatant of HepG2 cells, which was expected to contain a protease activity able to cleave HA0, did not support the replication of influenza viruses in MDCK cells when used as conditioned medium. These apparently paradoxical results can probably be explained by the mechanism of plasminogen activation in influenza virus-infected HepG2 cells. Physiological tPA activation of plasminogen is known to occur at the level of endothelial cell membranes (5), and plasminogen and tPA are associated with the plasma membrane of HepG2 cells (6, 16). Hence, HA0 could be cleaved during the viral particle attachment-internalization process by simultaneously activated plasminogen. To determine the potentially complementary roles played by the plasminogen/tPA complex and endocytosis, we cloned the tPA cDNA on HepG2 RNAs and transfected it into hTERT-BJ1 cells, which are telomerase-immortalized human fibroblasts (Clontech, BD Biosciences, Le Pont de Claix, France). Expression of the transgene was verified. Neither parental nor transfected hTERT-BJ1 cells were found to be permissive for influenza viruses (unpublished data). This demonstrates that tPA alone cannot render cells permissive for influenza viruses. The attachment of influenza viruses to the plasma membrane of HepG2 cells is probably facilitated by the binding of the envelope-anchored neuraminidase of influenza viruses to plasminogen. Goto and Kawaoka (4) demonstrated that neuraminidase at the virion surface both binds to and sequesters plasminogen and thus contributes to the in situ increase in the concentration of this protease. Similarly, inhibitors of plasmin activity such as tranexamic acid or ɛ-aminocaproic acid that are known to bind to plasmin(ogen) could be internalized on contact with the plaminogen/tPA/viral particle complex. This would explain the inhibition that we observed with these compounds. Finally, plasminogen activator inhibitor is also present at the surface of HepG2 cells (16). The potential activation of plasminogen at this level might be the result of regulated competition between activators and inhibitors.

In addition to proving useful for primary isolation and propagation of influenza viruses, the HepG2 system might represent an interesting model for study of the interrelations between plasminogen-tPA and influenza viruses and investigation of the virulence of influenza virus strains.