Ávila-Pérez et al. 2020
Ávila-Pérez et al. (2020) reports the recovery of a replicating Zika agent directly in live mice from a DNA copy of its genome, without the tissue-culture step that reverse genetics normally requires. It is assessed here against its own claims, against the rationale it gives for its own method, and against the published submission criteria of the journal that reviewed it.
The paper's stated purpose is the removal of a confound. Recovering an agent in cultured cells selects for culture-adapted mutations, so recovering it in the animal directly should yield material "devoid of confounding factors typical associated with in vitro systems" (p. 1). The experiments reported do establish that something replicating and lethal was obtained in mice from the clone, and the vaccine arm carries a negative control whose animals were separately shown to be susceptible — a stronger design element than most papers assessed here possess. What the paper does not report is any determination of the sequence of what it recovered, which is the measurement its own rationale calls for. The assessment is on the first of the two grounds at Which Documents Get An Entry — several of its claims assert more than the reported evidence shows — and it belongs to the reverse-genetics track described at Two Tracks. It is not an instance of the conditional claim stated there for the isolation track: the causal relation this paper asserts is controlled, and better controlled than most, and the flags below concern what the paper asserts beyond what it measured.
Citation and Access
Ávila-Pérez G, Nogales A, Park J-G, Morales Vasquez D, Dean DA, Barravecchia M, Perez DR, Almazán F, Martínez-Sobrido L. "In vivo rescue of recombinant Zika virus from an infectious cDNA clone and its implications in vaccine development." Scientific Reports 10:512 (2020). doi:10.1038/s41598-020-57545-2. Received 19 September 2019; accepted 3 January 2020; published 16 January 2020.
Open access under CC BY 4.0, and retrievable by automated request as well as in a browser, which is not true of every paper assessed on this wiki. Two independent renderings were read: the publisher PDF (fourteen pages, clean text layer) and the Europe PMC full text (PMC6965646). The journal is electronic and carries no separate journal pagination, so page references below are to the numbers printed at the foot of the PDF, 1 to 14.
The journal criteria used below are cited from captures made during this paper's review period rather than from the present versions of those pages, since both have since changed: submission guidelines, captured 5 December 2019 and editorial policies, captured 26 September 2019. The live pages refuse automated retrieval behind an identity-provider redirect; they open normally in a browser.
Reference 30 of the paper carries the clone, the attenuating substitution and the lethal challenge dose, and is cited only as "Sci Rep. In press (2019)". It is Ávila-Pérez G et al., "A natural polymorphism in Zika virus NS2A protein responsible of virulence in mice", Scientific Reports 9 (2019), doi:10.1038/s41598-019-56291-4, published 27 December 2019 — eight days before this paper was accepted. The citation is unresolved in the publisher PDF and in the Europe PMC text alike, and the PDF of record still reads "Published: xx xx xxxx" (p. 13). Both are processing failures at the journal rather than defects in the work, and no flag below rests on either; they are recorded so that a reader can reach the reference.
The Standard Applied
Each flag is measured against one of three things, and each flag states which:
- The paper's own claims — whether the reported evidence supports what the paper asserts.
- The paper's own stated rationale — the confound the method exists to remove, as the paper itself describes it.
- The published submission criteria of Scientific Reports — identifiable, dated, and in force during this paper's review.
No flag is measured against an unnamed standard of competence, and none against an expectation the authors did not set for themselves. In particular, nothing below faults the paper for using an interferon-receptor-deficient mouse or an injected route. The authors name that limitation themselves, and it is credited rather than flagged.
How this paper was chosen. It was read because it is the strongest case known to this wiki against a self-referential reading of virological method, and it is already analysed on those terms at Reverse Genetics. That is selection by relevance to an argument this wiki already makes, and not by any criterion independent of it. Nothing here supports a generalisation about reverse-genetics papers, about this laboratory, or about this journal.
What the Paper Establishes
Stated fairly, before any flag, and against this assessment's own interest.
- Recovery in the animal. A replicating and lethal agent was obtained in mice from the bacterial artificial chromosome clone by three routes — subcutaneous footpad, intramuscular quadriceps and intraperitoneal — both from transfected Vero cells inoculated into mice (p. 2, Fig. 1 p. 3) and from the clone complexed with transfection reagent and inoculated directly (pp. 3–5, Fig. 2 p. 4). Intraperitoneal delivery was the most efficient, affecting 6 of 6 animals (p. 5).
- Dose-dependence. Complexed clone at 100, 50, 25 and 12.5 µg was uniformly lethal; 6.25 µg affected 3 of 6 animals (p. 5, Fig. 3 p. 6).
- Recovery without transfection reagent. Naked clone DNA at 100 µg produced disease in all animals, with a dose-response below that (p. 5, Fig. 5 p. 8, Fig. 6 p. 9). The paper reports this as the first intraperitoneal delivery and expression of naked DNA (p. 8).
- The recovered agent was biologically competent. Serum from affected mice, given to naive mice at 100 focus-forming units by the subcutaneous route, produced disease indistinguishable from that of the cultured agent and death by 7–8 days (p. 5, Fig. 4 p. 7). That material carries neither plasmid nor transfection reagent.
- A negative control whose animals were demonstrably susceptible. In the vaccine experiments, mice given empty vector with the same transfection reagent neither lost weight nor died and yielded no detectable agent (Fig. 7 p. 10), raised no antibodies against viral antigens where clone-inoculated mice did (Fig. 8A p. 11), and then died 8–9 days after challenge with the virulent agent (p. 7, Fig. 9B p. 12). The last of these is the strongest element of the design. A negative control is informative only if the control animals could have shown the effect, and here that was demonstrated rather than assumed.
- Immunogenicity and protection. The attenuated clone delivered intraperitoneally raised antibodies by ELISA and neutralising antibodies by focus-reduction neutralisation test at levels comparable to sublethal infection with the attenuated agent itself, and the animals survived challenge at 50 × MLD50 without weight loss (p. 7, Figs 8–9 pp. 11–12).
- Results reported against the authors' own case. Subcutaneous delivery of the complexed clone affected only 2 of 6 animals and intramuscular 4 of 6 (pp. 3–4), and naked DNA required 100 µg where the complexed clone required 12.5 µg (p. 5). None of this is softened.
Flags
1. The recovered agent was never sequenced
Flag. The paper's rationale is that culture systems "impose bottlenecks on RNA viruses that eventually leads to selection of a virus population more adapted to grow in vitro" (p. 8), and it offers the method as useful "to minimize potential mutations associated with passaging the virus in culture cells" (p. 9). Establishing that requires knowing the sequence of what was recovered. No sequence determination of any recovered agent is reported. The Material and Methods describe cells, the clones, in-vitro recovery, mouse studies, viremia by focus-forming assay, ELISA and focus-reduction neutralisation, and nothing further (pp. 11–13).
The identity of the recovered agent therefore rests on lethality, on serum titre measured with a pan-flavivirus envelope antibody that does not distinguish one flavivirus sequence from another (p. 11), and on the attenuated phenotype. The claim that the method avoids the mutations a culture step would introduce is not measured by the paper that makes it.
Where checked. Material and Methods in full, pp. 11–13; abstract p. 1; discussion pp. 8–9. The whole paper was read before this flag was raised, and both renderings were then searched for "sequencing", "PCR", "RT-PCR", "Sanger" and related terms in an attempt to refute it: there are no occurrences, and the only appearances of "sequence" describe the construct.
Status. Absent from the described design.
What would refute this flag. Any reported sequence determination of a recovered agent, in this paper or in material published with it.
Credit. The second round of infection (p. 5, Fig. 4 p. 7) is a biological check in place of a sequence check, and it does establish that the recovered material retained the virulent phenotype through a passage in which no plasmid was present.
2. "Sterilizing immunity" is claimed from serum viremia beneath a detection limit
Flag. The paper concludes that a single intraperitoneal inoculation of the attenuated clone "is safe, immunogenic and provides sterilizing immunity against lethal challenge with a virulent rZIKV strain" (p. 7). The evidence offered for the sterilising part is that challenged animals showed "viremia below levels of detection" (p. 7). That limit is 200 focus-forming units per millilitre by immunofocus assay, stated in the figure legends throughout; serum was sampled at two time points, days 3 and 6 after challenge (p. 12, Fig. 9 legend); and no tissue, organ or molecular assay appears anywhere in the paper. Sterilising immunity is the absence of replication, which undetectable serum titre at two time points beneath an insensitive threshold does not establish.
The accompanying claim of safety rests on six animals observed for 21 days, in a strain that cannot mount a type-I interferon response. The attenuated group did lose 5–10% of body weight between 8 and 10 days and did show detectable viremia at 6 days (pp. 6–7).
Where checked. Results p. 7; Fig. 9 legend p. 12; Determination of viremia in sera samples p. 12; Material and Methods pp. 11–13.
Status. Observed. The claim is present, and the supporting measurement is narrower than the claim.
What would refute this flag. Tissue or organ virology, a molecular assay, additional sampling time points, or a definition of sterilising immunity given in the paper that undetectable serum viremia would satisfy.
3. Correspondence to human disease is asserted without citation and without measurement
Flag. The discussion states that in this model "infection with ZIKV using different routes (e.g. subcutaneous, intraperitoneal or intramuscular) results in neurological disease and the animals succumb to viral infection with high viral load in the blood and organs (liver, spleen, kidney, brain, and spin cord), consistent with ZIKV infection and manifestation in humans" (p. 9).
That sentence carries no citation. References 33 and 45 attach to the sentence that follows it, about wild-type mice. The organ viral loads it describes were not measured in this paper, whose only virological measurement is serum viremia. And the correspondence it asserts runs against the paper's own introduction, which states that "the majority of ZIKV infections are asymptomatic or associated with a mild febrile illness", with congenital malformation in pregnancy and Guillain-Barré syndrome in adults as the serious outcomes (quotation p. 1; the serious outcomes named p. 2). What was measured is weight loss and hind limb paralysis, with a humane endpoint of more than 20% weight loss or hind limb paralysis (p. 12), so the survival curves record animals meeting euthanasia criteria. Neither of the two serious human outcomes the paper names was measured.
Where checked. Discussion p. 9, with citation placement confirmed in the Europe PMC markup; introduction pp. 1–2; Mouse studies p. 12; all nine figure legends.
Status. Observed.
What would refute this flag. A citation supporting the human-correspondence sentence, organ titres reported in this paper, or a measurement of either serious human outcome.
Note. The same contrast between mouse and human outcome is used for a different purpose at Reverse Genetics, where it bears on whether the experiment meets a specification. Here it bears only on whether a sentence of the paper is supported by the paper.
4. No reversion analysis for a single-residue attenuation
Flag. The introduction names reversion as the hazard the approach must address: "In the case of LAVs, these mutations could lead to the reversion to a virulent phenotype" (p. 2). The attenuated clone differs from the virulent clone by one substitution, valine to alanine at residue 117 of the NS2A protein (pp. 5, 11). That attenuated agent was recovered in mice, replicated to detectable viremia, and was neither sequenced nor passaged to test for virulent revertants. A single-residue attenuation recovered and replicated in a living animal is the case in which reversion is most consequential and cheapest to look for, and the paper proposes the system as a vaccine platform on the strength of it.
Where checked. Introduction p. 2; description of the attenuated clone pp. 5 and 11; results pp. 6–7; Material and Methods pp. 11–13.
Status. Absent from the described design.
What would refute this flag. A reported sequence or phenotype check on the recovered attenuated agent, or a passage experiment testing for revertants.
5. The required data availability statement is absent, and nine results are unshown
Flag. The journal's submission guidelines in force during this paper's review state that "Scientific Reports requires a Data Availability Statement to be included in all submitted manuscripts (at the end of the main text, before the References section)". The published paper carries no such statement. The same page states "We request that authors avoid 'data not shown' statements and instead make their data available via deposition in a public repository", and that "Any data necessary to evaluation of the claims of the paper that are not available via a public depository should be provided as Supplementary Information". Nine results carry the "data not shown" label, there is no Supplementary Information, and no repository deposit is referenced.
The gradient in those three requirements is worth preserving: the statement is required, avoiding "data not shown" is requested, and the supplementary provision is expressed as should. Only the first is breached outright.
The nine fall into three groups — three in-vitro transfection-quality checks, two in-vivo empty-vector control arms, and four observations of clinical signs. The last group is the consequential one. Clinical signs are a declared endpoint: "Morbidity (body weight loss and paralysis) and mortality were monitored for 18 days" (p. 12), and hind limb paralysis is half of the euthanasia criterion that generates every survival curve in the paper. Every clinical-sign observation in the paper is unshown, including the signs recorded as "tremors, disorientation, hind limb weakness and severe paralysis" (p. 2) on which the description of the syndrome rests.
This is not a claim that the data do not exist. The editorial policies in force state that "Supporting data must be made available to Editorial Board Members and referees at the time of submission for the purposes of evaluating the manuscript", so the referees may well have seen them. The flag is that the published record does not contain them, and that the statement required to tell a reader how to obtain them is missing.
Where checked. The whole paper, pp. 1–14, for the absence of the statement and of supplementary material, in both renderings independently; submission guidelines as captured 5 December 2019; editorial policies as captured 26 September 2019.
Status. Observed. The statement is absent from the published paper, not merely not located: the guidelines specify where it belongs, and nothing stands at the end of the main text before the references.
What would refute this flag. A data availability statement or supplementary file in the version of record, a repository accession for the unshown results, or a dated version of the journal's criteria showing the requirement was introduced after this paper was submitted.
6. A figure legend compares an arm the figure does not show
Flag. Figure 5 reports three arms — the clone complexed with Lipofectamine 2000, the clone complexed with GenJet, and the clone without transfection reagent — and its legend describes those three (p. 8). The legend then closes: "No statistically significant differences were observed among mice inoculated using LPF2000, GenJet or empty BAC." No empty-vector arm is reported for that experiment, in the figure or in the text, which gives the third group as naked clone DNA. Either the comparison names the wrong group or it reports an arm that was run and not shown, and the paper does not let a reader determine which.
Where checked. Figure 5 and its legend, p. 8; the corresponding text, p. 5.
Status. Observed.
What would refute this flag. A panel or supplementary data showing an empty-vector arm in that experiment, or a published correction identifying the group.
Claims Not Supported by the Paper
Flags this assessment declined to make, and readings that the paper answers itself.
- That the empty-vector control is missing. It is not. The vaccine experiments document it in three figures (Fig. 7 p. 10, Fig. 8A p. 11, Fig. 9 p. 12), including the survival-then-challenge sequence that proves the control animals were susceptible. A reading of this paper as lacking a negative control arm misdescribes it, and the inference that injection trauma explains the outcome is the arm that was run with the sequence removed.
- That the attenuated construct is an inadequate sequence control. A single-residue change that attenuates the outcome is finer-grained evidence of sequence dependence than an empty vector alone, and the paper is entitled to it. The narrower observation — that no unrelated replicating construct was tested, so the headline contrast is between a replicating and a non-replicating construct — concerns the scope of what the design can answer rather than the support for what is claimed, and it is treated at Reverse Genetics.
- That "data not shown" means the results were not obtained. Nothing supports that reading, and the journal's policy in force required supporting data to be with the referees at submission. Flag 5 concerns the published record, not the conduct of the work.
- That the unresolved reference or the unfilled publication date bears on the science. Neither does. Both are recorded under Citation and Access so that a reader can reach the reference, and no flag rests on them.
What the Authors Stated as Limitations
The authors identify the host limitation plainly, and name it as a constraint rather than a feature. The paragraph opens: "It is important to indicate that our studies are based on the use of IFNAR −/− A129 mice that are deficient in type I IFN responses" (p. 9). It explains the choice by the unsuitability of the alternative — wild-type mice develop no clinical signs, no mortality and no detectable agent — concludes that the deficient strain "represented the best suitable model to demonstrate the successful rescue of recombinant ZIKV using in vivo reverse genetics", and states that "Further studies using a more relevant animal model of ZIKV infection, currently outside the scope of this manuscript, will demonstrate the feasibility of using similar in vivo reverse genetics" (p. 10).
The introduction also identifies, accurately, the two confounds that the method exists to address: culture-adaptive mutations that "might restrict and/or change the phenotype of the virus in vivo", and reversion to a virulent phenotype in live-attenuated vaccines (p. 2). The authors name the problem precisely. Flags 1 and 4 concern the absence of the measurements that would show it had been solved.
What Would Weaken This Assessment
- Any flag refuted by a passage not located during this reading. Each flag names its page, so each is testable at the cost of one fetch.
- A supplementary file, data deposit or data availability statement associated with the article and not served with either of the two renderings read.
- Evidence that the sequence of recovered agents is reported in reference 30 or in another companion paper intended to be read with this one. Flag 1 concerns what this paper reports.
- A dated version of the journal's submission criteria showing that the data availability requirement post-dates this paper's submission.
- For flag 2, an established definition of sterilising immunity that undetectable serum viremia at two time points would satisfy.
- For flag 1, a demonstration that lethality, reactivity to a pan-flavivirus antibody and the attenuated phenotype together constrain the recovered genome tightly enough to support the claim that culture-associated mutation was avoided.
Epistemic Status
The paper is observed. The full fourteen pages were read in two independent renderings — the publisher PDF and the Europe PMC full text — including all nine figures and their legends, the complete Material and Methods, the discussion and the reference list. There is no supplementary material to read.
Each flag is observed as to what the paper contains and does not contain. The count of nine unshown results is observed and was taken from both renderings independently; a whitespace-tolerant match is required, because one instance breaks across lines in the PDF text layer. The placement of references 33 and 45 in flag 3 is observed from the Europe PMC markup rather than from the PDF, where a superscript cannot be attributed to a sentence with certainty. The journal criteria are observed from dated archival captures, not from the live pages. The resolution of reference 30 is observed from the publisher's registered metadata.
That the absent sequencing bears on the method's stated advantage is inferred: it follows from the paper's own rationale rather than from any statement in the paper.
Whether the practices flagged here are characteristic of this journal, this field or this method is unresolved, and this entry supports no such claim. One paper was read, and it was chosen for its relevance to an argument already on this wiki.
See Also
- Source assessments — the standard these flags are measured against, stated for the genre
- Reverse Genetics — the same paper read as the strongest objection to a self-referential account of the method
- Enders and Peebles 1954 — the other paper assessed against its authors' own stated requirements
- Peer Review — the system this assessment tests by demonstration