A Lethal Case of Natural Infection with the H5N8 Highly Pathogenic Avian Influenza Virus of Clade 2.3.4.4 in a Mandarin Duck

Round 1

Reviewer 1 Report

Khali performed the pathological examination of a mandarin duck infected with H5N8 subtype of highly pathogenic avian influenza virus, which was found in Japan. Moreover, the genetic analysis of the isolate was also performed. As the authors pointed out, mandarin ducks are recognized as a specific avian species with different epidemiology in avian influenza virus infection from others, findings obtained in this study should be important, providing the new insights for avian influenza ecology. This manuscript should be published after minor modifications in below.

 

 

L38-40; though the human infection of H5N8 highly pathogenic avian influenza virus infection was reported, it should not be recognized as a common event (only one case and should not be even recognized as a sporadic case). If the authors should be concern for the H5N8 avian influenza virus infection in human, the authors may argue the potential of these viruses as public health evens using the results of genetic markers already reported, such as 627aa of PB2.

 

L51-52; please indicate the name of District or Province where Izumi locates.

 

L73-77; why do the authors suspect this genetic reassortment occurred in overwintering site? Not in nested lakes and pongs in summer? Please indicate the reasonable bases or considerations for the authors’ conclusion.

 

L78-80; the authors should indicate the approval for euthanizing of the infected birds; putting the ethical clearance number for the animal experiment. Or, if this bird was euthanized under the administrative issue or permission, the authors are highly recommended to put the description of approval from the competent authority in the text.

 

L114-122; this paragraph is not meaningful. The contents written at L115-116, is inconsistent to the one in L107-109. Sentences starting at L114 and L116 are somehow overlapped, and will be combined.     

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Ahmed Magdy Khalil et al. report the case of a mandarin duck infected with an H5N8 reassortant strain who developed neurological signs and was humanely euthanized because of poor conditions. Although oral and cloacal swabs were negative with rapid antigenic testing, viral RNA was identified though RT-PCR in the oral swab and the virus was successfully isolated. Additionally, the virus was also detected in most organs (except the intestines). Gross pathology and Histopathology showed the presence of neurological lesions, necrotic areas with macrophage infiltrations. Additionally, immunohistochemistry methods detected viral antigens in tissues. Finally, genetic characterization showed that this virus was a reassortant. The study is interesting and well performed since it includes several different approaches (molecular characterization, pathology, immunohistochemistry, virus isolation) and the results are relevant since these viruses are usually nonpathogenic in these animals. The manuscript is well written, and most conclusions are supported by the results. However, some procedural details are missing from the manuscript. Here are my specific comments:

1. In the title and often in the text you refer to this as a “lethal case”. However, technically the animal was euthanized so, although the duck would have likely died, I don’t think you can define this a lethal case. Maybe instead of highlighting this aspect, you can stress out the neurological complications you identified.
2. Some additional methodological details should be provided. In the caption of supplementary figure 1 you should add which model was used to build the phylogenetic trees. At line 57 you should include a reference for this PCR or provide (as supplementary material in fine) the used primers. Similarly, at line 63 you should include a reference for this full-genome amplification method or provide the primers.
3. Nanopore sequencing is not an accurate sequencing method, and a high sequence coverage should be obtained. To show that your results are reliable, you should provide details about sequence coverage for each segment. Additionally, have you considered and evaluated the possibility of a co-infection? Maybe instead of a reassortant strain, two different viruses were simultaneously infecting the duck. Can you confirm that only one virus was there?
4. Unlike all other segments, PB2 showed lower identity respect to reference strains. Have you verified that no sequence error or multiple strains affected the quality of the obtained consensus sequence?

Minor:

- Line 41. Species names should be italicized.

- Line 42. Family and order names should not be italicized.

- Line 86. M1 is part of the viral capsid, so technically detecting it detects viral presence, not replication.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

I appreciate the efforts of the authors to answer my comments. However, I still think that more details about the results of the sequencing should be provided to show that the sequences were accurate. 
- Although previous studies have shown that Nanopore sequencing can be used to sequence influenza virus, this does not necessarily mean that it worked for you. The accuracy of this sequencing method, as I explained before, depends on the sequence coverage, and this changes for each sequencing run. The fact that you sequenced almost correctly another virus before doesn’t mean that the sequencing was accurate in this case. The only parameter that can give you an indication of how good the sequencing was is the sequencing coverage, which has to be high for the entirety of the sequence otherwise, the consensus could be doubtful. As in those 2 papers cited at line 70, an idea of the sequence coverage should be provided. This is particularly important for that PB2 divergent sequence. You could simply add the information about sequencing coverage for each segment in an additional column in Table 1. 
- Co-infection. How were consensus sequences obtained? Did you map towards a reference strain, or did you assemble? Di you only obtain 8 sequences or more? Have you tried to assemble all reads of each segment separately to see if you obtained more than one consensus for one segment? I understand that this methodology can’t rule out with 100% certainty the presence of multiple infections, but this is something that should be verified and, at least, mentioned in the manuscript. Maybe you can add a short description of how the consensus sequences were obtained, explaining why you think that there is no evidence for multiple viruses.
- I saw that you used ML to build your trees, but that is the method. The information missing is the model used to calculate genetic distances (was it a K2, HKY, GTR…? Did you use a gamma distribution? How did you choose the models? Was a model-test performed?).
- Since the animal was euthanized, you cannot say for sure whether the infection was sufficient to cause death or not. Therefore, I still don’t think "lethal" can be used. If you write "lethal", the reader will understand that the animal died as a consequence of the infection, which is technically not true. Maybe you can ask the opinion of the Editor on this point. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 3

Reviewer 2 Report

Although the potential for co-infection was not addressed, the reader now has all the elements to evaluate sequencing accuracy.