A peptide-based viral inactivator inhibits Zika virus infection in pregnant mice and fetuses

Zika virus (ZIKV), a re-emerging flavivirus associated with neurological disorders, has spread rapidly to more than 70 countries and territories. However, no specific vaccines or antiviral drugs are currently available to prevent or treat ZIKV infection. Here we report that a synthetic peptide derived from the stem region of ZIKV envelope protein, designated Z2, potently inhibits infection of ZIKV and other flaviviruses in vitro. We show that Z2 interacts with ZIKV surface protein and disrupts the integrity of the viral membrane. Z2 can penetrate the placental barrier to enter fetal tissues and is safe for use in pregnant mice. Intraperitoneal administration of Z2 inhibits vertical transmission of ZIKV in pregnant C57BL/6 mice and protects type I or type I/II interferon receptor-deficient mice against lethal ZIKV challenge. Thus, Z2 has potential to be further developed as an antiviral treatment against ZIKV infection in high-risk populations, particularly pregnant women.

2. fig 2. any reason why a standard plaque assay was not used? Fig 2C, where is the control peptide as a control image? Same critique for D. Scrambled peptide derived from Z2 would have been the best control to use throughout. 3. The flow plots do not look that different if the gates were drawn NOT to include half of what appears to be a negative population (minor shift in the population. probably the cells that are fluorsecing around 105 on the X axis is positive, not the ones that are next to the negative population. Given that, there is a lot of background. Also, the authors should demonstrate that the cells that are ZIKV E-positive are the same ones where Z2 is binding. This should be fairly easy given that Z2 is already labeled with Cy5. Fig 3C,

Summary:
The authors present a new peptide drug (Z2 with 33 amino acids) with efficacy against Zika virus (ZIKV), which places a particular threat on pregnant women and, more specifically, on the developing fetus via teratogenic effects. The synthetic peptide Z2 is derived from the stem region of ZIKV envelope protein and is able to penetrate the placental barrier of pregnant, non-infected mice. Conversely, daily injection of Z2 protected non-pregnant mice from a lethal infection challenge with ZIKV. As other peptide drugs were previously used in pregnant women to prevent HIV transmission to the child and based on the presented data, this concept seems promising. As mode of action, the manuscript shows that Z2 disrupts the membrane of the enveloped ZIKV, leads to the release of viral genomes, and thus abrogates infectivity. The used methods are well suited to address the raised research questions, and the experiments are well controlled, including unrelated peptides or viruses. Nevertheless, several key points on the route of virus challenge and safety profile need to be addressed as outlined below before the manuscript may be suited for publication in Nature Communications.
MAJOR COMMENTS 1. The results that Z2 reaches fetuses of non-infected A129 mice (IVIS imaging of fluorescently labeled Z2, Fig. 4) seem disconnected from the data, showing that Z2 can protect adult not pregnant mice from ZIKV infection (survival data, Fig. 6). A major claim of the manuscript is that Z2 can be used during pregnancy to protect fet uses from ZIKV infection, but this claim is actually not addressed by the data. The authors need to provide evidence that Z2 can prevent teratogenic effects in pregnant A129 mice by protecting the developing fetuses from ZIKV infection. These experiments c ould be performed using a lower and sublethal dose of ZIKV in pregnant mice and determine fetal development.
2. Fig. 5: The authors claim that Z2 is a safe peptide drug candidate for use in pregnant women. Safety of Z2 was, however, only tested in pregnant mice the dose that protected non-pregnant mice from ZIKV infection (10 mg/kg) and a 2-fold higher dose (20 mg/kg). Even though not showing in vitro cytotoxicity for cell lines (Fig. S4) or mouse red blood cells (Fig. S4), the performed test in mice is not sufficient to conclude about toxicity in vivo. Along the same lines, the shift in fluorescence of E-protein negative cells in Fig. 3B may indicate a non-specific binding of Z2 to non-infected cells, and Z2 reaches non-infected fetuses (Fig. 4), suggesting binding to non-infected cells and potential off-target effects. Consequently, a dose-escalation study in pregnant mice needs to be performed to determine potential toxicity.
• To test the efficacy of Z2 in vivo, mice were injected intraperitoneally with Z2 one hour after infection with ZIKV with the same route. In addition, in vitro experiments were performed via pre-incubating ZIKV and Z2 for 1.5 hours. These procedures do not reflect the natural events that occur during ZIKV transmission via infected mosquitoes. The manuscript would largely benefit from showing that virus inoculation and treatment is independent. The authors could for instance inoculate ZIKV intravenous, subcutaneous, or intradermal to model natural routes of virus transmission and then inject Z2 intraperitoneal.
• Along the same line as the point above, testing the spread of Z2 to the fetus was performed via intravenous injection (line 677), but testing the efficacy of the drug against ZIKV infection was performed via i.p. injection. For testing the spread of Z2 to the fetus, the same route of inoculation should be performed as for testing the efficacy of the drug.

MINOR COMMENTS
• Line 46-47: The authors should mention the Americas in addition to Africa and Asia -Pacific region as areas at risk for ZIKV transmission.

Reviewers' comments:
Reviewer #1 (Remarks to the Author): Overall, the experiments in this study are performed rigorously and the results are of great importance. The paper is clearly written and will attract interest of researchers working in related areas and stimulate further progress in development of anti-ZIKV therapeutics. However, I have few recommendations that need to be considered in order to strengthen the manuscript and make the results and the claims of this paper more convincing for the scientific community. I had divided them into major and minor points.

Response:
We thank the reviewer for the encouraging comments and helpful suggestions.  "Actually, ZIKV can infect pregnant woman and cause severe congenital brain developmental abnormalities in fetus, as well as can infect testis and cause damage that leads to male infertility 52, 53 . Therefore, we also detected the distribution of Z2 in the genital organs of male mice. Supplementary Figure 12 showes that Z2 could distribute to the genital organs of male mice, suggesting that it may prevent testis damage caused by ZIKV. However, further experiments need to be carried out to confirm this finding."  Response: We thank the reviewer for the helpful suggestion. We have determined the interaction between Z2 and ZIKV surface protein expressed on the ZIKV-infected BHK21 cells. As shown in Supplementary Figure 4, E protein (green, stained by anti-E mAb 4G2) and Z2-Cy5 (red) overlapped and colocated together (Supplementary Figure  4A). In the flow cytometry assay (Supplementary Figure 4B), Z2-Cy5 bound with 63.5% of ZIKV-infected BHK21cells, significantly higher than that of mock-infected cells. The results suggest that Z2 peptide can interact with ZIKV surface protein expressed on ZIKV-infected BHK21 cells.
The result was added to the revised manuscript (lines 144-145): "Similar results were obtained from ZIKV-infected BHK21cells (Supplementary Figure 4)." The related methods were added in the revised manuscript (lines 387 -392): "To test the binding of Z2 to ZIKV-infected BHK21 cells, BHK21 cells seeded in cover slips were ZIKV or mock infected, fixed with 4 % PFA, perforated by 0.2% Triton X-100 and blocked with 3% BSA. The cells were then incubated with anti-E mAb 4G2. After 5 washes, the cells were incubated with Alexa Fluor 488-labeled donkey antimouse antibody and Z2-Cy5 at RT for 1 h. After another 5 washes, the cover slips were sealed for scanning with the Leica SP8 confocal microscope. "

Minor points
1. In Figure 1B, how were the amino acid residues in Z2 peptide numbered?
Response: Z2 peptide was derived from the membrane-proximal stem region (residues 421-453) of ZIKV E protein. Therefore, we numbered the peptide as 421-453 in Figure  1B "ZIKV-117, a human monoclonal antibody, could broadly neutralize infection of divergent ZIKV strains 16 . However, the high cost may limit its application in developing countries, such as Brazil." (lines 58 -60 in the revised manuscript).
"Modification of the anti-ZIKV antibodies to decrease their binding to FcγR may reduce risk of ADE 16 . However, this may further increase the cost of these antibodies." (lines 293 -295 in the revised manuscript).

In line 71, when the abbreviation "E protein" first appears in the manuscript, it should be explained.
Response: The abbreviation of "E protein" was explained upon its first appears in the manuscript.
Response: Thanks, we have done our best to correct the English grammar errors in the manuscript. Figure 4D appears before Figure 4B and 4C. This should be adjusted so that the figures appear in sequence.

Response:
The order of panels in Figure 4 was adjusted accordingly.
6. In line 316, the subtitle in the Method section: "Assays for antiviral activity and cytotoxicity" should be changed to "Assays for antiviral activity". The cytotoxicity experiments were not included in this section.

Response:
We have changed the subtitle according to the reviewer's suggestion.
7. In line 326, both reference 31 and 53 are related with pseudotyped MERS-CoV, but no reference was provided about the preparation of pseudotyped VSV.

Response:
We apologize for this mistake. We have added the following paragraph: "The pseudotyped VSV and MERS-CoV were set as unrelated virus controls and prepared as previously described 60,61 . Briefly, 293T cells were cotransfected with a plasmid encoding MERS-CoV S protein or VSV-G protein and pNL4-3.luc.RE using VigoFect. Supernatants containing pseudotyped MERS-CoV or VSV were harvested 48 h posttransfection. Huh7 cells were infected by these pseudotyped viruses as described above." (lines 364 -366 in the revised manuscript).

Reviewer #2 (Remarks to the Author):
The authors describe the development and characterization of a peptide derived from ZIKV E protein to be used as a fusion inhibitor of ZIKV to be used in pregnant women. The manuscript details Z2's inhibitory properties against ZIKV in vitro, its distribution in vivo, and its ability to interrupt pathogenesis in a mouse model. The paper offers several novelties, including a peptide that could potentially move forward for use in humans, but the manuscript falls short, especially given a title that suggests that it may be beneficial for use in pregnant women. The paper does not even demonstrate this in mice, thus it is a stretch, and misleading, to write this in the title, as it is not a major finding. Response: We thank the reviewer for the insightful comments, which help us to improve our manuscript. We changed the title to "A novel antiviral peptide inactivates Zika virus and prevents its infection in pregnant mice and their fetuses", and responded to the reviewer's critiques one by one as shown below.

Authors show that Z2 has protective activity in vivo, only when used 1 hour post infection. have the authors tried any other time?
Response: We did not try any other time study before. According to the constructive suggestion, we administered Z2 peptide to A129 mice 24 h after inoculation with ZIKV and found that about 33.3 % of these treated mice had survied (Supplementary Figure  11A). Viral load in the Z2-treated A129 mice at 3 days post-infection was about 4-fold lower than that of vehicle-treated mice (p < 0.01) (Supplementary Figure 11B). These results indicate that Z2 peptide still has protective activity in vivo even when administered 24 h after ZIKV infection. We have added the results in the revised manuscript: "Twenty-four hours after inoculation with ZIKV, treatment with Z2 still could protect 33.3 % of A129 mice from death (p < 0.05) (Supplementary Figure 11A). Viral load in the Z2-treated A129 mice at 3 days post-infection was about 4-fold lower than that of the vehicle-treated mice (p < 0.01) (Supplementary Figure 11B). Although Z2 inactivates ZIKV at the early stage of viral replication, consecutive Z2 injection after ZIKV penetrance of cells could still provide some protection of the infected A129 mice, possibly by inactivating newly produced ZIKV virions and prevent their infection of more target cells." (lines 251 -257 in the revised manuscript) During this study, we could not obtain more A129 mice from the animal suppliers in China; therefore, we will test more time points after administration of Z2 peptide in the future once we have obtained enough A129 mice for the experiments.

Is the drug capable of blocking transmission of ZIKV transplacentally?
Response: Yes, Z2 peptide could block transmission of ZIKV transplacentally in mice. We have added experiments to prove it (new Figure 6), and the related results in the revised manuscript (lines 223 -234): "To determine whether Z2 could protect against vertical transmission of ZIKV, pregnant C57BL/6 mice were infected by ZIKV as described previously 39 and were then treated with Z2 at 10 mg/kg of body weight (n=12) or vehicle control (n=12). The results showed that Z2 treatment could reduce viremia in ZIKV-infected pregnant C57BL/6 mice (p < 0.05) (Figure 6A). At the same time, viral RNA load in placentas from the Z2-treated pregnant mice was significantly lower than that from vehicle-treated mice (p < 0.01), and the infection rate decreased from 18/24 to 12/24 ( Figure 6B). Interestingly, Z2 treatment resulted in the decrease of infection rate of fetal head from 14/24 to 2/24 (p < 0.001) ( Figure 6C). These results suggest that Z2 may inactivate ZIKV virions either before or after the virions have penetrated the placenta to fetus, thus reducing the infection rate of fetuses, as well as protecting against vertical transmission of ZIKV in pregnant mice." "Antiviral efficacy of Z2 in pregnant C57BL/6 mice Antiviral efficacy of Z2 in pregnant mice was evaluated using the method as previously reported 39 . Briefly, 24 pregnant C57BL/6 mice (E12-14) were assigned randomly to two groups and infected intraperitoneally (i.p.) with 1 x 10 5 PFU of ZIKV (SZ01). One h later, the infected mice were i.p. administered with Z2 at 10 mg/kg of body weight (n=12) or vehicle control (n=12). At day 1 post-infection, mice were retroorbitally bled to measure viremia by RT-qPCR. Two embryos of each pregnant mouse were randomly collected and the viral RNA load in placenta and fetal head of each collected embryo was determined by RT-qPCR." 2. fig 2. any reason why a standard plaque assay was not used? Fig 2C, where is the control peptide as a control image? Same critique for D. Scrambled peptide derived from Z2 would have been the best control to use throughout.

fig 2. any reason why a standard plaque assay was not used?
Response: Based on our previous experience 1 and others' publication 2 , we quickly developed a colorimetric viral infection assay using CCK8 kit for screening anti-ZIKV compounds. Since we found this method to be convenient, quantitative, and reproducible, we used it to evaluate the anti-ZIKV activity of Z2 peptide. The validation of this colorimetric assay was recently confirmed by Muller et al 3 . To address your concerns, we performed the standard plaque assay to determine the anti-ZIKV activity of Z2 using BHK21 cells. As shown in Supplementary Figure 2, the result derived from the plaque assay (IC 50 : 2.61 ± 0.46 μM) is consistent with that obtained from the CCK8 assay (IC 50 : 1.75 ±0.13 μM).

Supplementary Figure 2: Plaque reduction assay to determine ZIKV infection in BHK21 cells. (A)
Plaque reduction assay for Z2. ZIKV was incubated with Z2 or Z2-scr in different concentrations for 1.5 h. Then the mixture was added to BHK21 cells seeded in 6-well plates. After 1.5 h, the viral inoculum was removed and 2.5 ml DMEM with 2% FBS and 1% low melting-point agarose overlaid the cells. Four to five days later, the cells were fixed with 4 % PFA and stained with 1% crystal violet overnight. (B) Plaque reduction curves. Plaques were counted and percentage of plaque reduction was calculated.
We added the results in the revised manuscript (lines 111 -113): "We also used the plaque reduction assay and BHK 21 cells to test anti-ZIKV activity. As shown in Supplementary Figure 2, the IC 50 is 2.61 ± 0.46 μM, suggesting that the result derived from the plaque reduction assay is consistent with that obtained from colorimetric CCK8 assay."

Fig 2C, where is the control peptide as a control image?
Response: The control image of the control peptide was added to new Figure 2C, as shown below.

Same critique for D. Scrambled peptide derived from Z2 would have been the best control to use throughout.
Response: We agree. We designed the scrambled peptide of Z2 (Z2-scr: LDIIAGLSAGFQGGATFVDAHGMVKASFLGGNW) by using the Pep Controls (scramble pep) Program v.1.2 and synthesized this peptide. Using this scrambled peptide as a control, we have repeated a series of experiments to compare its anti-ZIKV activity with that of Z2 peptide. The new results were added to the following figures: Figure 2A

The flow plots do not look that different if the gates were drawn NOT to include half of what appears to be a negative population (minor shift in the population.
probably the cells that are fluorsecing around 10 5 on the X axis is positive, not the ones that are next to the negative population. Given that, there is a lot of background. Also, the authors should demonstrate that the cells that are ZIKV E-positive are the same ones where Z2 is binding. This should be fairly easy given that Z2 is already labeled with Cy5. Fig 3C, no control peptide control was used.

The flow plots do not look that different if the gates were drawn NOT to include half
of what appears to be a negative population (minor shift in the population. probably the cells that are fluorsecing around 10 5 on the X axis is positive, not the ones that are next to the negative population. Given that, there is a lot of background.

Response:
We have drawn the gate to the cells that are fluorescing around 10 5 on the X axis. Indeed, the difference was reduced. The shedding or the residues of Cy5 from Z2-Cy5 may have contributed to the background.

Also, the authors should demonstrate that the cells that are ZIKV E-positive are the same ones where Z2 is binding. This should be fairly easy given that Z2 is already labeled with Cy5.
Response: As suggested by Reviewer #1, we have carried out immunofluorescence staining experiments using ZIKV-infected BHK21 cells with high expression level of E protein. We have demonstrated that the ZIKV E-positive cells are the same as those binding with Z2. Although we washed the cells for a longer time, some background staining remained. As shown in Supplementary Figure 4, the green (E protein, stained by 4G2) and red (Z2-Cy5) fluorescence overlapped, suggesting that Z2-Cy5 could bind to E protein in ZIKV-infected BHK21 cells.

Supplementary Figure 4: Binding of Z2-Cy5 to ZIKV-infected BHK21 cells. (A)
Immunofluorescence staining assay. Green, ZIKV E protein ; Red, Z2-Cy5; Blue, nuclei. Scale bar = 100 μm. 3C, no control peptide control was used. Response: The scrambled peptide of Z2 (Z2-scr) was used as a control peptide and the new result was added to new Figure 3C in the revised manuscript. Supplementary Figure 13: Pharmacokinetic study of Z2. SD rats (200 ± 10 g, n=3) were administrated 10 mg/kg Z2 intravenously via the tail vein. Blood samples were collected through retro-orbital bleeding at 0.25, 0.5, 1, 2, 3, 4, 8 and 12 h after administration and centrifuged at 6000 rpm for 5 min to obtain the serum samples. Then 150 μl methanol containing 15 μg C24M peptide (MTWEEWDKKIEEYTKKIEELIKKS) as an internal standard was added to 50 μl serum to precipitate the protein. After centrifugation at 17,000 rpm for 10 min, the supernatant was subjected to liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis, using an AB SCIEX QTRAP 6500 instrument (SCIEX, Boston, USA). All concentration data were dose-normalized and plotted as serum drug concentration time curves. PK solutions 2.0 (noncompartmental pharmacokinetics data analysis) was utilized to analyze the pharmacokinetic parameters. Response: We apologize for the confusion. In our revised manuscript, we removed Figure 5A and added the following statement "The same legend was used for Figure 5A-5D." (highlighted in yellow) in the figure legend of the new Figure 5. We converted the original Figure 5F to Supplementary Figure S11.   (Fig. S4) or mouse red blood cells (Fig. S4), the performed test in mice is not sufficient to conclude about toxicity in vivo. Along the same lines, the shift in fluorescence of E-protein negative cells in Fig. 3B may indicate a non-specific binding of Z2 to non-infected cells, and Z2 reaches noninfected fetuses (Fig. 4), suggesting binding to non-infected cells and potential off-target effects. Consequently, a dose-escalation study in pregnant mice needs to be performed to determine potential toxicity.

Response:
We thank the reviewer for the suggestion. We have added experiments to determine the potential in vivo toxicity of Z2 at dose of 40, 80, and 120 mg/kg body of weight, to pregnant mice. As shown in new Figure 5, moms or pups in each group grew normally. Z2 treated groups camparing with PBS group showed no significant differences in ALT and creatinine in the blood, and histological sections of organs were normal without exception. These results suggest that Z2 at the dose as high as 120 mg/kg is safe for pregnant mice, which is 11-fold higher than that providing protection against ZIKV infection in vivo. Indeed, some non-specific binding of Z2 was observed. However, we believe that this very limited non-specific binding will not significantly affect the in vivo efficacy of Z2.
We have added the following statement in the revised manuscript (lines 219-221): "Overall, Z2 is safe for pregnant ICR mice and fetuses, even at the dose as high as 120 mg/kg of body weight, which is 11-fold higher than that providing protection against ZIKV infection in vivo ".
Mouse survival was observed and recorded daily until 21 days post-infection. (B) Viral RNA load in sera of ZIKV-infected A129 mice. At day 2 post-infection, mice were retroorbitally bled to measure viral RNA load in sera by RT-qPCR. (C) Survival of ZIKVinfected AG6 mice. AG6 mice (6 weeks old) were infected with 1×10 3 PFU of ZIKV via a subcutaneous route in the footpad. After 1 h, mice were treated with Z2 (n=6) at 10 mg/kg of body weight, and vehicle (n=6) as control. Mouse survival was observed and recorded daily until 21 days post-infection. (D) Viral RNA load in sera of ZIKV-infected AG6 mice. At day 2 post-infection, mice were retro-orbitally bled to measure viral RNA load in sera by RT-qPCR. Whiskers: 5-95 percentile. **, p < 0.01; ***, p < 0.001.
We have added the results in the revised manuscript (lines 247-250): "Similarly, treatment with Z2 protected 67% of AG6 mice from death caused by subcutaneous administration of ZIKV and significantly prolonged MST from 10 days to the end of the experiment (p < 0.01) ( Figure 7C). The viral load in Z2-treated AG6 mice at 2 days postinfection was about 13-fold lower than that of vehicle-treated mice (p < 0.01) ( Figure  7D)." We have added the method in the revised manuscript (lines 471-478) : "Twelve 6-weekold AG6 mice were assigned randomly to two groups and infected with 1 x 10 3 PFU of ZIKV (SZ01) via a subcutaneous route in the footpad. Then, the infected mice were i.p. administered with Z2 at 10 mg/kg of body weight (n=6) or vehicle control (n=6) every day for 6 consecutive days post-infection. The observation and examination of AG6 mice were then performed as described above." 4. Along the same line as the point above, testing the spread of Z2 to the fetus was performed via intravenous injection (line 677), but testing the efficacy of the drug against ZIKV infection was performed via i.p. injection. For testing the spread of Z2 to the fetus, the same route of inoculation should be performed as for testing the efficacy of the drug.
Response: Thanks for the suggestion. Accordingly, we have assessed the distribution of Z2 peptide after intraperitoneal injection. As shown in Supplementary Figure 9, the Z2-Cy5 fluorescence was seen in the uterus and fetus, suggesting that Z2 could penetrate the placental barrier and enter into the fetus. We have added the results in the revised manuscript.