The short peptide P9 exhibited the highest antiviral activity among mBD4-dervied peptides, smBD4, and rmBD4

Eleven mBD4-derived peptides were designed and synthesized (Table 1). Their antiviral activity against influenza virus H1N1 was evaluated in MDCK cells. Among all short peptides, P9 showed the strongest antiviral activity (Fig. 1a). Interestingly, the dose-dependent antiviral effect of P9 was stronger than those of synthetic mBD4 (smBD4) and recombinant mBD4 (rmBD4). In particular, the 50% inhibitory concentration (IC 50 ) of P9 was 1.2 μg/ml, which was lower than those of smBD4 (3.2 μg/ml) and rmBD4 (1.5 μg/ml) (Fig. 1b). Furthermore, the 50% toxic concentrations (TC 50 ) of P9, smBD4 and rmBD4 were 380, 280 and 360 μg/ml, respectively (Fig. 1c). The selectivity index of P9 was 316, which was higher than those of smBD4 (87) and rmBD4 (240). Notably, both of the liquid form and powder form of P9 demonstrated stable shelf lives as they exhibited similar antiviral activity after storing at −20 °C for more than 1 year (Supplementary Fig. 1). Taken together, among smBD4, rmBD4 and the 11 peptides derived from mBD4, P9 exhibited the highest antiviral activity and selectivity index. Thus, P9 was selected for further studies.

Table 1: Sequences of peptides derived from mBD4. Full size table

Figure 1: Evaluation of antiviral activity and cytotoxicity of peptides in vitro. (a) Antiviral activities of smBD4, rmBD4 and 11 short peptides (25 μg/ml) against H1N1 viral infection were tested in MDCK cell cultures. White and black colors indicate that the virus was pretreated with peptides in PB and MEM, respectively. Zanamivir (Zana) was included in the experiment as positive control. (b) Anti-H1N1 efficacies of P9, smBD4 and rmBD4 at different concentrations were detected in cultured MDCK cells. The inhibitory activity was directly determined by plaque assay and the infection ratio was calculated as plaque numbers in treated samples/plaque numbers in untreated samples. The IC 50 s are indicated by dotted line, indicating that the mean IC 50 s of P9, smBD4 and rmBD4 are 1.2 μg/ml, 3.2 μg/ml and 1.5 μg/ml, respectively. (c) The cytotoxicity of P9, smBD4 and rmBD4 was determined by MTT assay. TC 50 s are indicated by dotted line. All data are presented as means ± SD from three independent experiments. Full size image

P9 protected mice against lethal challenge of H1N1 virus

We further evaluated the prophylactic and therapeutic effects of P9 in a lethal mouse model of H1N1 influenza (Fig. 2). The mice were intranasally (i.n.) inoculated with P9 (50 μg/mouse) or rmBD4 and then i.n. challenged with lethal dose of H1N1 virus. The dose of 50 μg/mouse was applied for animal experiments because it was shown to protect all mice from lethal challenge of H1N1 virus. The survival rate of P9-pretreated mice (100%) was significantly higher than those of rmBD4-pretreated mice (20%, P < 0.01) and untreated mice (0%, P < 0.0001). Notably, pretreatment with P9 and Zanamivir achieved a similar protection effect (Fig. 2a). In parallel, when the mice were i.n. inoculated with 3 doses (50 μg/dose/day) of P9, rmBD4 or Zanamivir at six hours after the lethal challenge with the virus (Fig. 2b), the survival rate in P9-treated mice (60%) were significantly higher than those of untreated control and rmBD4-treated mice (P < 0.05). In this setting, treatment of zanamivir yielded a similar level of protection (80%) in comparison to P9 (P > 0.05). All infected mice suffered from body weight lost regardless of treatment types. However, the survived mice began to regain body weight at ~10–12 days post-infection (Fig. 2c,d).

Figure 2: Evaluation of prophylactic and therapeutic effects of P9 in a mouse model. (a) Survival rates of mice i.n. inoculated with 25 μl of PB (VC), P9 or rmBD4 (2 mg/ml) before H1N1 virus challenge. (b) Survival rates of mice i.n. treated with 25 μl of PB (VC), P9 or rmBD4 (2 mg/ml) after the virus challenge. Zanamivir (Zana, 2 mg/ml) was included as positive control. (c,d) The body weight of the mice corresponding to (a,b). Ten animals per group were used for this experiment. P values are indicated. Full size image

The viral RNA copies and viral titers in lung tissues of infected mice were detected by real-time RT-PCR (Fig. 3a) and plaque assay (Fig. 3b), respectively. The viral loads in lung tissues of P9-pretreated and P9-treated mice were significantly lower than that of the untreated mice (P < 0.05). Histopathological examinations revealed that the alveolar damage and interstitial inflammatory infiltration in untreated mice or mice pretreated or treated with rmBD4 were substantially more severe than those in mice pretreated or treated with P9 (Fig. 3c).

Figure 3: P9 decreased viral loads and pathological changes in lung tissues of infected mice. (a) Viral RNA copies in lung tissues of infected mice receiving prophylactic treatment (Prevention) and therapy by i.n. inoculation (Therapy). (b) Viral titers in lung tissues of the mice were detected by plaque assay. The results are presented as means + SD of five mice and * indicates P < 0.05. (c) Histopathological changes in the mouse lung tissues were tested by H&E staining. Representative histological sections of the lung tissues taken from the P9 or rmBD4 treated mice, untreated mice (Untreated) and uninfected mice (Normal) are shown (original magnification 100×). Full size image

The stability of P9 in vivo was evaluated by P9 biodistribution in mouse lungs and its antiviral activity at different time-points after P9 administration. As shown in Supplementary Fig. 2a, P9 could be detected on the surface of mouse bronchial tubs at 10 min, 2 h and 4 h after P9 administration, but the signal decreased to almost undetectable level at 8 h after P9 administration. In the in vivo protective experiment, when P9 was administrated to mice at 2 h, 4 h and 8 h before the challenge, it protected 60%, 30% and 20% mice from lethal challenge of H1N1 virus, respectively (Supplementary Fig. 2b). These results indicated that P9 could maintain more than half of its antiviral activity at 2 hours and about 1/5 antiviral activity at 8 hours after administration.

The in vivo toxicity of P9 was also assessed in mice. Each mouse was i.n. inoculated with P9 (50 mg/kg) or intra-peritoneally (i.p.) injected with P9 (500 mg/kg) per day for 3 days. As shown in Supplementary Fig. 3, less than 10% of body weight loss was observed in the first 3 days and the body weight began to recover when the treatment was stopped at day 3. There was no obvious reduction of food consumption or sickness during the 10-days observation period. Collectively, our data demonstrated that P9 exhibited prophylactic and therapeutic effects against lethal challenge of H1N1 virus in mice, accompanied with a low toxicity in vivo.

P9 inhibited influenza virus infection through binding to viral surface gylcoproteins

To investigate the antiviral mechanism of P9, we first seek to determine whether P9 inhibited virus entry, replication, or release. When P9 was supplemented to the culture medium at 1 hour after MDCK cells were infected with H1N1 virus (2 MOI), viral RNA copies inside the cells (Fig. 4a) and in culture supernatants (Fig. 4b) harvested at the indicated hours post-infection were comparable with those in the untreated control (P > 0.05). The results suggested that P9 did not inhibit viral replication or release. When the virus was pretreated with P9 and then added to the cells (0.3 MOI), viral loads both inside the cells (Fig. 4c) and in culture supernatants were significantly reduced (Supplementary Fig. 4a,b). In contrast, the viral loads inside the cells (Fig. 4d) and in supernatants (Supplementary Fig. 4c,d) were similar to those of the untreated control when the cells were pretreated with P9 followed by infection. These results indicated that P9 inhibited virus infection by interacting directly with the virus. We further demonstrated that P9 could bind to surface glycoprotein HA of influenza virus (Fig. 4e,f) and S2 of MERS-CoV (Supplementary Fig. 5). Together, our results indicated that P9 inhibited viral infections through binding to viral surface glycoproteins.

Figure 4: P9 inhibited virus infection by binding to viral glycoprotein. (a,b) P9 did not inhibit virus replication or release. Viral RNA copies inside cells (a) and in supernatants (b) of the cultures when P9 was maintained in the culture media after viral infection (P9-maint). (c) P9 inhibited the virus infection when the virus was pretreated with P9 (P9-virus-pre). (d) P9 did not inhibit viral infection when the cells were pretreated with P9 (P9-cell-pre). Untreated virus controls (VC) were included. All data are presented as means + SD of three independent experiments. * indicates P < 0.05. (e) P9 bound to viral surface glycoprotein HA of influenza virus. The binding of P9 to H1N1 viral proteins was detected by anti-mBD4 antibody. Scrambled peptide (SP), Anti-HA1 and Anti-NA antibodies were included as controls. (f) P9 bound to HA but not NA as determined by ELISA. Data are presented as means + SD of three independent experiments. ** indicates P < 0.01. Full size image

P9 blocked viral RNA release from late endosomes

To further dissect how P9 interfered with influenza virus infection, we asked whether virus-receptor binding, endocytosis or viral RNA release was affected by P9. H1N1 virus carrying a green fluorescent label was pretreated with P9 and then incubated with the cells at 4 °C for 3 hours. We found that P9-pretreated virus bound to the cellular membrane in a similar manner compared to the untreated virus (Fig. 5a, left). P9-pretreated virus entered the cells through endocytosis after the virus was incubated at 37 °C for 1 hour (Fig. 5a, middle). Furthermore, the transport of the virus in endosomes from the cell periphery to the proximity of the nucleus was not affected when the virus was pretreated by P9 (Fig. 5a, right). The colocalization of the virus with P9 further confirmed the binding of P9 to the virus, its internalization into cells and its subsequent transport to the perinuclear region together with the virus (Fig. 5b). On the other hand, the levels of viral RNA in infected cells remained essentially unchanged during the first 2 hours post-infection regardless of P9 treatment (Fig. 5c), suggesting that P9 did not inhibit viral binding and entry. The viral RNA levels in cells infected with untreated virus increased at 3.5 hours post-infection, indicative of the initiation of nascent viral RNA replication. In contrast, the viral RNA levels in the cells infected with P9-pretreated virus did not increase until 6.5 hours post-infection and was significantly lower than that of untreated virus control (P < 0.05) (Fig. 5c). These results implicated that P9 perturbed viral uncoating and viral RNA release from late endosomes for subsequent replication.

Figure 5: P9 did not inhibit virus-receptor binding or endocytosis but blocked viral RNA release. (a) P9 did not inhibit virus-receptor binding or endocytosis. DiO dye labeled virus pretreated with P9 or PB (VC) bound to the cell surface at 4 °C (left side) and entered the cells after incubation at 37 °C for 1 h or 2 h. The cell membrane was stained by Alexa-594 dye and representative images were taken by confocal microscope (original magnification 400×). (b) P9 was delivered into the cells by binding to the virus. Colocalization (orange) of virus (green) and P9 (red) was shown after incubation at 37 °C for 2 h. (c) P9 blocked the viral RNA replication. Viral RNA copies were detected in the cells infected with the virus pretreated with P9 or PB (VC) at the indicated time points. The results are presented as means ± SD of three independent experiments. Full size image

P9 prevented endosomal acidification

Endosomal acidification is a prerequisite for influenza virus uncoating and the subsequent genome release. Bearing in mind the basic amino acid-rich composition of P9, we examined whether P9 could block viral RNA release by counteracting endosomal acidification. As shown in Fig. 6a, P9 showed similar antiviral effects in comparison to ammonium chloride (NH 4 Cl) and Bafilomycin A1, which are well-defined inhibitors of endosomal acidification21,22,23. In contrast, Zanamivir, which inhibits viral release, did not show a similar antiviral kinetics in comparison to the endosomal acidification inhibitors. Therefore, P9, NH 4 Cl and Bafilomycin A1 might share a similar antiviral mechanism of inhibiting pH decrease in endosomes. To confirm this hypothesis, we detected the pH change in endosomes using a pH-sensitive dye, the intensity of which mirrored the acidity of the cellular milieu. Red signals in the cells infected with untreated virus (VC) indicated the acidification process in endosomes (Fig. 6b). In contrast to the control, when the cells were infected with virus pretreated with P9, NH 4 Cl or Bafilomycin A1, little to no any red signals were observed, indicating that acidification was abolished in the endosomes. To verify this, we designed a P9 analog named P9-aci-1 containing three additional acidic amino acids at the C-terminus (Table 1). If P9 acts through perturbation of pH lowering, the addition of acidic residues is expected to relieve the perturbation and abrogate the consequent antiviral effect. Indeed, confocal microscopic images indicated that pH lowering in endosomes was unaffected by P9-aci-1 (Fig. 6b). In line with the imaging result, P9-aci-1 lost its antiviral activity (IC 50 > 100 μg/ml). In addition, the polykaryon assay further showed that P9 did not directly inhibit HA-cell membrane fusion (Supplementary Fig. 6). Hence, when delivered into the endosomes, virus-bound P9 prevented pH decrease, resulting in the inhibition of viral RNA release and subsequent replication.

Figure 6: P9 suppressed pH lowering in endosomes. (a) The inhibitory effect of P9 was similar to those of inhibitors of endosomal acidification. Viral RNA copies in the cells infected with the virus pretreated with PB (VC), P9, NH 4 Cl or Bafilomycin were measured at the indicated time-points post-infection. NH 4 Cl and Bafilomycin were maintained in the culture medium after viral infection. The results are presented as means + SD of three independent experiments. (b) P9 inhibited pH decrease in endosomes. Cells were infected with DiO dye labeled virus (green) pretreated with PB (VC), P9, NH 4 Cl, Bafilomycin or P9-aci-1. The low pH in endosomes of infected cells was stained by pH-sensitive dye (pHrodo, red). Representative images were taken by confocal microscope (original magnification 400×). Full size image

P9 showed broad-spectrum antiviral effects against respiratory viruses in vitro and in vivo

We next evaluated whether P9 could inhibit other subtypes of influenza virus and other respiratory viruses such as SARS-CoV and MERS-CoV, because they all depend on endosomal acidification for viral infection. Our data demonstrated that P9 exhibited strong antiviral effects against other subtypes of influenza A virus including H3N2, H5N1, H7N7 and H7N9 with IC 50 s ranging from 1.5 to 4.8 μg/ml (Fig. 7a). The IC 50 s of P9 against SARS-CoV and MERS-CoV were about 5 μg/ml. Notably, at concentrations higher than 25 μg/ml, P9 could inhibit SARS-CoV and MERS-CoV infections to more than 95% (Fig. 7b). These results indicated that P9 has broad-spectrum antiviral activities in vitro against multiple respiratory viruses.

Figure 7: Detection of antiviral effects of P9 against infections of multiple respiratory viruses. (a) P9 inhibited infections of influenza virus subtypes H3N2, H5N1, H7N7 and H7N9 in cells. (b) P9 inhibited infections of SARS-CoV and MERS-CoV in cells. IC 50 s are indicated by dotted lines. The results are presented as means ± SD of three independent experiments. (c) P9 protected mice from lethal challenge of H5N1 virus. (d) P9 protected mice from lethal challenge of H7N9 virus. (e and f) Body weight of the mice corresponding to (c,d). (g,h) P9 inhibited the infection of SARS-CoV in mice. Lung tissues of infected mice were collected at day 3 post-infection. Viral titers in lung tissues were detected by plaque assay (g) and real-time RT-PCR (h). To evaluate prophylactic effect of P9, mice were intratracheally (i.t.) inoculated with 50 μl of PB (VC-P), Zanamivir (Zana-P) or P9 (P9-P). To evaluate therapeutic effect of P9, mice were i.n. treated by PB (VC-T), Zanamivir (Zana-T) or P9 (P9-T) after viral challenge. P values are indicated. Full size image

We further evaluated the protective effect of P9 against infections of H5N1, H7N9 and SARS viruses in mice. As shown in Fig. 7c, one dose (100 μg/mouse) of P9 for prophylaxis (P9-P) and 5 doses (100 μg/mouse) of P9 for therapy (P9-T) could protect 44% and 50% of mice from lethal challenge of H5N1 virus, respectively. These protection rates were significantly higher than that of untreated mice (P < 0.05) and comparable with the mice treated with prophylactic and therapeutic Zanamivir (Zana; 50% and 60%, P > 0.05). One dose of P9 for prophylaxis and 5 doses of P9 for therapy protected 50% and 50% of mice from lethal challenge of H7N9 virus, respectively. The treatment groups provided significantly more protection than the control group (P < 0.01). The survival rates of mice treated with prophylactic and therapeutic Zanamivir were 60% and 70%, respectively, which did not show significant difference compared to those of P9-treated mice (P > 0.05) (Fig. 7d). The body weight of all treated and untreated mice began to decrease at day 2 post-challenge and then gradually recovered from day 12 post-infection in the survival mice (Fig. 7e,f). Since we have not established animal models for lethal infection of MERS-CoV and SARS-CoV, we compared the viral load in mouse lungs at day 3 post-infection in a mouse model of non-lethal SARS-CoV infection. As shown in Fig. 7g,h, one dose of P9 for prophylaxis and 5 doses of P9 for therapy significantly inhibited virus infection in mouse lungs compared to that in the untreated mouse lungs (P < 0.05). These results demonstrated that P9 could provide broad-spectrum protection in vivo against infections of multiple respiratory viruses.