Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Summary:

The research reported a new lentiviral vector packaging strategy using yellow fever virus 17D-derived RNA replicons to create replicon cell pools in HEK293FT cells that persistently express HIV-1 Gag-Pol. They show that wild-type HIV-1 protease causes cytotoxicity that prevents the establishment of stable replicon-bearing cells, while a T26S protease mutation eliminates this effect. This allows them to generate puromycin-resistant cell pools that maintain Gag Pol expression over multiple passages. Using these pools as packaging cells simplifies lentiviral vector production by requiring only transient transfection of the transfer vector, Tat/Rev, and envelope plasmids. They also include production optimization, a multilayer stack scale-up yielding roughly 10⁹ transducing units, functional transduction of target cells, and replication-competent lentivirus testing that showed no detectable replication-competent lentivirus under their assay conditions. Conceptually, their approach is novel and interesting and could offer real advantages for scalable manufacturing. However, they overinterpret several of their central claims. Specifically, their conclusions about how their method compares to standard production systems, the long-term stability of their replicon-based platform, and the biosafety advantages all need much stronger experimental backing and more careful interpretation before this manuscript can be published.

 

Major Concerns:

1.     Absence of direct benchmarking against standard lentiviral production methods

The study repeatedly claims that their yields are comparable to current methods and standard transient transfection systems. However, they did not actually perform a head-to-head comparison within this study using a conventional second-generation packaging system under the same experimental conditions. Concluding that the yields are comparable without the same study control, using standard transient four-plasmid transfection, is just an assertion rather than a demonstrated result. I recommend that they either conduct a direct comparison experiment under matched harvest conditions, cell density, and titration methods, or significantly tone down their language to state that preliminary yields are within the range reported for transient transfection, with clear caveats about the need for formal benchmarking. Cross-checking with published stable producer platforms like LentiPro26 (which also uses the T26S protease mutation and reports sustained titers >10⁶ TU/mL/day from constitutive producer clones) will highlight how important systematic productivity comparisons are when making claims about manufacturing value.

 

2.     Insufficient evidence for stable replicon maintenance

The manuscript uses the term stable replicon containing cell pools and claims that the replicon itself remains genetically stable. However, their evidence for this stability relies mostly on semi-quantitative VLP output measured by GoStix p24 detection across passages. This assay only reports the phenotypic output, meaning Gag protein release. Still, it does not directly prove replicon RNA persistence, the genetic integrity of the replicon sequence after extended passaging, or the uniformity of replicon retention across the polyclonal cell pool. Relying on this to claim stable replicon maintenance is an overinterpretation of the current data. If possible, I recommend they add direct measurements of replicon RNA levels across passages using qRT-PCR with replicon specific primers. They should also provide sequencing verification of the Gag-Pol and critical regulatory regions from the pool at a late passage to confirm there are no adaptive mutations or deletions. Without these data, they should dial back the claim of stability to sustained VLP production.

 

3.     RCL detection assay requires validation and transparency

They conclude that no replication-competent lentivirus was detected and frame this as a biosafety advantage of their flavivirus replicon platform. However, their assay description is missing critical validation parameters. They state that the VSV G qPCR results are not shown, but these should be presented with actual Ct values for all time points and replicates. The limit of detection for the qPCR assay is also not defined. Additionally, their positive control only verifies that the qPCR primers and probe work; it does not serve as a full-process positive control that mimics amplification throughout the workflow. I recommend presenting all qPCR data transparently and, if possible, including a positive process control by spiking a known low amount of infectious particles into naive cultures at the start. Furthermore, the mechanistic claim that flavivirus compartmentalization reduces the risk of replication-competent lentivirus is a plausible hypothesis, but this was not experimentally demonstrated in this specific system. They should reframe this as a theoretical rationale requiring future validation, rather than presenting it as a proven safety advantage.

 

4.     Statistical analysis and experimental design clarity

The authors state that experiments were performed in triplicate, with log-transformed titers analyzed using statistical tests. However, several methodological details are unclear. For example, in Figure 4C, they report a mean of 12 wells but do not specify whether these are technical or independent biological replicates. While statistical testing is indicated, the failure to define what the error bars represent (e.g., SD vs. SEM) in the MTT figure complicates interpretation. For the passage stability experiments in Figure 5, treating each passage comparison as an independent test risks inflating Type I error. They should probably use a repeated-measures ANOVA that accounts for within-replicate correlation across time. Also, the western blot densitometry in Figure 4D appears to be from a single representative blot, making it unclear whether the quantification reflects multiple independent biological replicates. They need to clarify their experimental unit and biological replicate structure for all quantitative assays, present clear error bars, and use appropriate statistical models for longitudinal data.

 

5.     Limited functional characterization and quality assessment of packaged vectors

The research demonstrate that their packaged lentiviral vectors can transduce HEK293FT target cells. While this confirms functional vector production, it gives very limited information about the actual quality and performance of the vector. They did not assess the physical-to-functional particle ratio, nor did they test the vector in a biologically relevant target cell type, such as primary T cells. They also failed to compare transduction efficiency or potency relative to vectors produced by standard methods. Since this manuscript is positioned around manufacturing value, I recommend they test at least one additional target cell type relevant to their intended application. They should also include a physical titer measurement to calculate the particle to infectivity ratio and provide a side-by-side functional comparison of their vectors versus standard derived vectors in matched target cells.

 

6.     Methodological gaps and inconsistencies

Several details need correction for reproducibility. There is inconsistent plasmid nomenclature, referring to pTat/Rev in one section but pTet/Rev in another (line 237, 349). In Figure 4, the microscopy panels use different objective magnifications and GFP exposure times, which weakens direct visual comparison. They should standardize their imaging conditions or add quantitative image analysis metrics. Also, the Data Availability statement mentions supplementary material, but supplementary figures or methods are not clearly provided in the manuscript. Finally, the selection strategy, clonality, and passage history of the established cell pool are not systematically described.

 

Minor Concerns

• Introduction and Discussion: The Introduction motivates the manufacturing problem well but could explicitly position this approach relative to other intermediate strategies. The discussion rightly acknowledges this is a first-generation platform requiring further development, but it should be much more cautious about biosafety and comparability claims until additional validation is provided.
• Data transparency: They should consider depositing full plasmid sequences, raw flow cytometry files, and complete western blot images in a public repository and providing the accession numbers in the revised manuscript.

Author Response

Response to Reviewer 1

We thank Reviewer 1 for the thorough and constructive comments. We have carefully considered all points and have revised the manuscript accordingly. Below we provide a point‑by‑point response.

 

Major Concern 1. Absence of direct benchmarking against standard lentiviral production methods

The study repeatedly claims that yields are comparable to current methods, but a head‑to‑head comparison was not performed within this study.

Response:

We agree that the original wording was too strong. We have since continued our work and performed additional packaging experiments using RCP cells, as well as conventional transient transfection of HEK293FT cells, to enable a direct side‑by‑side comparison under identical conditions (5‑layer stacks, same transfection method, same titration protocol). These experiments are now described in Section 2.10 and Section 3.6 and Table 1. Benchmarking was therefore conducted, and the results are presented in the revised manuscript. The wording has also been adjusted to tone down claims.

 

Major Concern 2. Insufficient evidence for stable replicon maintenance

Stability is shown only by p24 output; no direct proof of replicon RNA persistence or genetic integrity.

Response:

We have performed RT‑PCR and Sanger sequencing of PCR products covering the entire Gag‑Pol insert after ten passages of RCP cells. The sequencing results are now presented in Section 2.14 (Materials and Methods) and Section 3.5 (Results). Sequencing revealed no mutations in the Gag‑Pol insert, underscoring the accuracy of the YFV replicon used.

In addition, we have extended the passaging of RCP cells under puromycin selection and show that under these conditions, extracellular p24 production levels are maintained (data shown in Section 3.4).

In the Discussion, we now state that Sanger sequencing confirmed the absence of mutations, while also noting that low‑frequency variants may escape detection. All phrases mentioning “stability” have been toned down to reflect the actual findings.

 

Major Concern 3. RCL detection assay requires validation and transparency

Missing critical validation parameters (Ct values, LOD, full‑process positive control); the mechanistic claim about flavivirus compartmentalisation is not experimentally proven.

 

Response:

The original manuscript described a qPCR‑based approach. However, to improve reproducibility, we have switched to a p24 ELISA‑based two‑phase RCL detection assay. This assay does not rely on qPCR and therefore does not require Ct values. Instead, it measures p24 antigen production during an extended culture period (21 days RCL amplification phase and 7 days indicator phase), which is the standard method for RCL detection in research settings. This assay has been published by multiple independent groups, e.g. 1. Escarpe et al., 2003; 2. Cornetta et al., 2011; 3. Sastry et al., 2005 – these sources are now referenced in the revised manuscript as. Refs.39, 40, 41. These sources demonstrate that the p24 ELISA‑based two‑phase assay is widely accepted and appropriate for the detection of RCL in research‑scale lentiviral vector production.

We explicitly state in the revised Discussion that: “The p24 ELISA‑based two‑phase assay used here is appropriate for a proof‑of‑concept study and has been widely employed for RCL screening in research settings [39,40,41]. Nevertheless, future development of this platform for clinical applications will require RCL testing with more stringent assays (e.g., spike‑in controls, and more sensitive technologies).”

Thus, we fully acknowledge the limitations while justifying the suitability of the method for the presented study.

Also, the LOD for the p24 ELISA (8 pg/mL) was determined using the method described in Section 2.15 (Materials and Methods). Raw data and calculations shown in Supplementary Table S6.

 

3.1. Removal of unsupported mechanistic claim “the mechanistic claim that flavivirus compartmentalization reduces the risk of replication-competent lentivirus is a plausible hypothesis”

Response: We agree with Reviewer 1 that the claim about flavivirus compartmentalisation reducing RCL risk was not experimentally proven. Therefore, we have removed that statement from the Discussion. Instead, we now simply cite literature on the intrinsically low recombination rates of flaviviruses (references 30, 31) and state that the absence of detectable RCL is consistent with this property, without claiming causation.

 

Major Concern 4. Statistical analysis and experimental design clarity

Unclear whether replicates are technical or biological; error bars not defined; repeated‑measures ANOVA should be used for passage data; densitometry from a single blot.

Response:

We have clarified the experimental design and statistics throughout.

• Changes made in the manuscript: Section 2.16 now states: “Three biological replicates were done unless otherwise stated. Data are presented as mean ± SD.” For the MTT assay, we specify that each biological replicate had four technical replicates (total 12 wells per condition).
• Figure 2d (describing MTT test results) now states: “Data points represent the mean of three independent biological replicates (each with four technical replicates, total 12 wells per condition). Error bars are SD.”
• For the passage stability experiments (Figures 5 and 6), we used one‑way repeated measures ANOVA with Geisser‑Greenhouse correction; this is stated in Section 2.16 and in the figure legends.
• Regarding densitometry: we removed any quantitative claim based on a single blot. Instead, in Section 3.2 (Figure 2 legend) we state: “This experiment was repeated with qualitatively consistent results. One photograph of a full membrane representative to several blotting experiments is shown in Supplementary Material Figure S7.

 

Major Concern 5. Limited functional characterization of packaged vectors

Vector tested only on HEK293FT, not on primary T cells; no physical‑to‑functional particle ratio; no comparison with standard vectors on relevant cells.

Response:

We have produced these significant new data and added to the revised manuscript.

• Added Section 3.7 (Results) shows production of CAR‑T cell from primary human T lymphocytes transduced with the RCP‑produced LV/CAR vector. The result is obtaining of a cell product with a high transduction rate: 77.4 % CAR+ cells (Figure 8 and Supplementary Figure S8).
• We compared functional titers (TU/mL) and p24 concentrations for lentiviral vector batches produced using the RCP system or the conventional transient transfection system. We have added Table 1 (Section 3.6), which presents the results of these production runs, and Table 2 (Section 3.6), which provides the calculated TU/pg p24 ratios. The values are higher for the conventional system under the conditions tested; however, the limited number of production runs precludes statistical inference.
• In the Discussion, we now highlight that the RCP‑produced vector transduces primary human T cells (lines 936-936 and 953-954).

 

Major Concern 6. Methodological gaps and inconsistencies

Inconsistent plasmid nomenclature (pTat/Rev vs pTet/Rev); different magnifications in microscopy panels; missing supplementary materials; incomplete description of selection strategy and passage history.

Response: All identified inconsistencies have been corrected. Plasmid name is uniformly pTat/Rev. All microscopy images now have consistent magnification 10X. Supplementary Material has been added.

Section 2.13 now describes the selection strategy in detail: RCP obtained after a single electroporation and puromycin selection. The passage history (P0 from cryopreservation, P1–P10 under continuous selection) is clearly stated.

 

Minor Concerns

1. Introduction and Discussion: position the RCP approach among other intermediate packaging strategies

Response: We have expanded the Introduction to explicitly position RCPs as an intermediate strategy between transient transfection and fully stable producer cell lines. In the Discussion, we have toned down the biosafety claims as requested (see response to Major Concern 3).

2. Data transparency:

Response: The complete nucleotide sequence of the key construct in this study – the molecular infectious clone used to produce RCP cells (pYFrep/GAG‑POL*/Pac) - is shown in Supplementary Material Figure S3. We explicitly state: “Other sequences are available from the authors upon request” (lines 193-194). The flow cytometry diagram and uncropped Western blot images are now presented in the Supplementary Material.

 

Sincerely,

Alexandr Shustov, corresponding author

On behalf of all coauthors

Reviewer 2 Report

Comments and Suggestions for Authors

The authors are trying to solve a novel problem. Making CAR-T therapy requires lentiviral vectors (LVs) which are expensive to produce. In terms of scalability, it is difficult as it requires transfecting HIV-derived Gag-Pol carrying plasmids along with multiple plasmids. They are trying to solve this by creating cells that permanently carry gag-pol gene inside a self-replicating RNA molecule derived from Yellow fever virus(YFV). This method is new in terms of ideas and could be useful. However, there are several areas where additional experimental evidence and clarification would considerably strengthen the conclusions drawn.

Major comments:

1. The manuscript lacks a direct comparison with standard lentiviral production systems. Without a side by side comparison, it is difficult to assess how this system performs relative to existing approaches. The Western blot in Figure 4d shows that cells expressing T26S-mutant protease accumulate substantially more unprocessed pr55gag precursor (46%) compared to wild-type (14%). While the authors cite Konvalinka et al. 1995 to argue this does not affect infectivity, it is unclear whether this holds true in the context of a replicon-based packaging system. The authors should provide data directly comparing the particle to infectivity ratio of vectors produced from RCP versus standard transient transfection.
2. GoStix is a semi-quantitative lateral flow test. For a stability study of this nature, a quantitative p24 ELISA would be more appropriate and rigorous. However, the data in Figure 5d are presented as a quantitative graph with numerical values on the Y-axis (VLPs/mL). The authors should clarify how quantitative values were extracted from the GoStix assay.
3. The RDRP lacks proofreading activity. To confirm the genetic stability over the passages, sequence verification of the replicon at key sites particularly the T26S mutation and the EMCV IRES, across passages is essential. Also, five passages are insufficient to support claims of long-term stability relevant to a manufacturing context. Consistent production should be demonstrated over a significantly greater number of passages.
4. The biosafety rationale for using YFV over alphavirus replicons is theoretically sound but not experimentally demonstrated. The RCL assay relies solely on VSV-G qPCR detection, which would not capture recombination events occurring independently of VSV-G incorporation. Also, while classical ADE is not applicable here given the absence of structural proteins, the manuscript provides no discussion of biosafety classification or regulatory handling requirements for cells chronically harboring replicating flavivirus-derived RNA. These details could be important information for groups considering adopting this platform.
5. Given that the stated application is therapeutic CAR-T manufacturing, transduction efficiency should be demonstrated on primary T cells, CD34+ HSCs, or at minimum Jurkat cells. Data from HEK293FT cells alone is insufficient to support therapeutic claims.
6. The authors employ a second-generation lentiviral system requiring Tat for LTR activation. Given that third-generation self-inactivating systems are now standard in the field and offer improved biosafety, the authors should justify this design choice.

Minor comments:

1. There is a typographical inconsistency in the transactivator plasmid name. For example, It is referred to as both "pTat/Rev" (lines 236, 272, 495) and "pTet/Rev" (line 402) in the large-scale production section. This should be corrected throughout.
2. The phrase "particularly in low- and middle-income countries such as the authors' home country, Kazakhstan" (line 712) reads as self-referential in a scientific context. The authors should reframe this more broadly, for example, "in developing countries" which would actually strengthen their argument.

Author Response

Response to Reviewer 2

We thank reviewer 2 for the thorough evaluation of our manuscript and for the constructive comments. We have carefully addressed all major and minor concerns, and the manuscript has been substantially revised. Below we provide a point-by-point response.

 

Major comment 1 – Lack of direct comparison with standard lentiviral production systems

Response:

We have now performed a direct side‑by‑side comparison between the RCP‑based system and a conventional three‑plasmid transient transfection of HEK293FT cells under identical conditions (5‑layer stacks, same transfection method, same titration protocol). The results are presented in Table 1 (Section 3.6, lines 807–809) and Table 2 (lines 822–824).

Regarding the Gag processing: We agree that the T26S mutation leads to accumulation of pr55gag precursor (as shown in Figure 2e). However, our functional assays (CAR‑T cell production, Figure 8) clearly demonstrate that the RCP‑produced vector is fully capable of transducing primary human T cells with high efficiency (77.4% CAR+ cells). This confirms that the residual proteolytic activity of the T26S mutant is sufficient for particle infectivity, consistent with the original report by Konvalinka et al. We have added a sentence in the Discussion (lines 936-937 and 952-954) to explicitly state that the RCP-produced vector is functionally capable.

 

Major comment 2 – GoStix assay for stability study

GoStix is a semi-quantitative lateral flow test. For a stability study of this nature, a quantitative p24 ELISA would be more appropriate and rigorous.

Response:

We agree that a quantitative p24 ELISA is more rigorous. Therefore, we have replaced the GoStix data entirely with quantitative p24 ELISA measurements. The revised Figure 5 (ten RCP passages under selection) now presents p24 concentrations (ng/mL) measured by p24 ELISA (Section 3.4, line 687). The revised Figure 6 (RCP passages without selection and after reintroducing puromycin) also presents data of p24 ELISA (Section 3.4, line 726).

The experimental scheme and statistical analysis (repeated measures ANOVA) are described in Sections 2.13 and 2.16. We thank the reviewer for pointing out this weakness.

 

Major comment 3 – Genetic stability over passages

To confirm the genetic stability over the passages, sequence verification of the replicon at key sites particularly the T26S mutation and the EMCV IRES, across passages is essential.

Response:

We have addressed this concern in two ways:

Extended passaging: We have performed passaging up to P10 under continuous puromycin selection (Section 3.4, lines 677–684). p24 production remained stable across ten passages (Figure 5).

Sanger sequencing: We carried out RT‑PCR and bidirectional Sanger sequencing of the entire Gag‑Pol insert after passage P10. The results had shown no nucleotide substitutions compared to the parental MIC, and the T26S mutation was present (Section 3.5, lines 739–750).

 

Major comment 4 – Biosafety rationale and RCL assay

The RCL assay relies solely on VSV-G qPCR detection, which would not capture recombination events occurring independently of VSV-G incorporation. Also, while classical ADE is not applicable here given the absence of structural proteins, the manuscript provides no discussion of biosafety classification or regulatory handling requirements for cells chronically harboring replicating flavivirus-derived RNA. These details could be important information for groups considering adopting this platform.

Response:

RCL assay: The original manuscript described a qPCR‑based approach. However, to improve reproducibility, we have switched to a p24 ELISA‑based two‑phase RCL detection assay. This assay does not rely on qPCR and therefore does not require Ct values. Instead, it measures p24 antigen production during an extended culture period (21 days RCL amplification phase and 7 days indicator phase), which is the standard method for RCL detection in research settings. This assay has been published by multiple independent groups, e.g. 1. Escarpe et al., 2003; 2. Cornetta et al., 2011; 3. Sastry et al., 2005 – these sources are now referenced in the revised manuscript as. Refs.39, 40, 41. These sources demonstrate that the p24 ELISA‑based two‑phase assay is widely accepted and appropriate for the detection of RCL in research‑scale lentiviral vector production.

We explicitly state in the revised Discussion that: “The p24 ELISA‑based two‑phase assay used here is appropriate for a proof‑of‑concept study and has been widely employed for RCL screening in research settings [39,40,41]. Nevertheless, future development of this platform for clinical applications will require RCL testing with more stringent assays (e.g., spike‑in controls, and more sensitive technologies).”

Thus, we fully acknowledge the limitations while justifying the suitability of the method for the presented study.

Biosafety classification: We have added a paragraph in the Discussion (lines 965–968) addressing the biosafety considerations: The replicon in this study is derived from the live‑attenuated yellow fever virus vaccine strain YFV‑17D, which is classified as a Risk Group 2 agent by the American Biological Safety Association [61]. Because the replicon lacks the structural genes of YFV, it cannot generate infectious viral particles, which supports the biosafety of this platform.

 

Major comment 5 – Transduction of primary T cells

Given that the stated application is therapeutic CAR-T manufacturing, transduction efficiency should be demonstrated on primary T cells…

Response:

We have completely addressed this comment. Section 3.7 (lines 827–839) and Figure 8 now present data on CAR‑T cell manufacturing using primary human T lymphocytes as a starting material. The vector was produced in the RCP system and used to transduce primary T cells. The transduction efficiency was 77.4%. The complete dataset is provided in Supplementary Figure S8. This directly supports the therapeutic applicability of the RCP‑produced vector.

 

Major comment 6 – Justification for second‑generation lentiviral system

Response:

We use the second‑generation LV packaging system because second‑generation systems generally yield higher titers than third‑generation systems. Moreover, second‑generation packaging systems remain in widespread use in academic research and are well‑suited for exploratory studies such as the present investigation.

 

Minor comments

Minor comment 1 – Typographical inconsistency in transactivator plasmid name (pTat/Rev vs pTet/Rev)

Response: Corrected.

Minor comment 2 – Self-referential phrase about low‑ and middle‑income countries

Response: Corrected. We have completely removed the phrase “particularly in low- and middle-income countries such as the authors‘ home country, Kazakhstan”.

 

Sincerely,

Alexandr Shustov, corresponding author

On behalf of all coauthors

Reviewer 3 Report

Comments and Suggestions for Authors

The authors reported a vector packaging system using the Yellow Fever virus that can stably produce replicons for the HIV-1 Gag-Pol protein (compared to wild-type vs T26S protease), allowing transient transfection of transfer vectors and Tat/Rev, and enveloped plasmids, which shall finally yield lentiviruses. 

The manuscript can be enhanced. 

1- Lines 80, 86, 232, 308, 539, 547: Please insert references to support the statement or hypothesis ot technique.

2- In Figures 4 and 5: please insert scale bars. In Figure 4, Please try to represent your MTT results as a percentage in the supplementary section. 

3- Use an identical Y-axis scale for Fig. 5d. How did you do the statistical analysis? Also, how did you quantify using Lenti-X?

4- Reorganize Figure 6. 

5- Discussion is quite redundant, particularly lines 702-781. Try to shorten. 

6- In Line 700, where is Figure 8d?

Author Response

Response to Reviewer 3

We thank reviewer 3 for the constructive comments. We have carefully revised the manuscript accordingly. Below is a point‑by‑point response.

 

Comment 1 Lines 80, 86, 232, 308, 539, 547: Please insert references to support the statement or hypothesis or technique.

Response:

We have added the missing references at the indicated locations. The old line numbers are no longer apply because we have substantially revised the manuscript. Specifically, we added references to describe the following:

Line 80 (previous manuscript) – we added:

“… and RCPs are easier to generate than classical PCLs because they eliminate clonal selection, shortening the development timeline [19].”

Corresponding reference: [19] Nishimura, K.; Segawa, H.; Goto, T.; et al. Persistent and stable gene expression by a cytoplasmic RNA replicon based on a noncytopathic variant Sendai virus. J Biol Chem. 2007, *282*(37), 27383‑27391. https://doi.org/10.1074/jbc.M703041200

Line 86 (previous manuscript) – we added:

“Such recombination could generate replication‑competent lentivirus (RCL), posing a hypothetical safety concern for therapeutic applications [28,29].”

Supporting references: [28] Farley, D.; Stockdale, S.; Moore‑Kelly, C.; Miskin, J.; Reiser, J.; Mitrophanous, K. Risks of replication‑competent retro/lentivirus from associated vector systems: Is it time for a roadmap toward reduced testing? Mol Ther Methods Clin Dev. 2025, *33*(4), 101601.

[29] Sena‑Esteves, M.; Gao, G. Monitoring Lentivirus Vector Stocks for Replication‑Competent Viruses. Cold Spring Harb Protoc. 2018, *2018*(4).

Line 232 (previous manuscript) – we added:

“Plasmids were isolated by alkaline lysis and purified for transfection by banding in cesium chloride density gradients as described in [33].”

Corresponding reference: [33] Syzdykova, L.; Zauatbayeva, G.; Keyer, V.; Ramanculov, Y.; Arsienko, R.; Shustov, A.V. Process for production of chimeric antigen receptor–transducing lentivirus particles using infection with replicon particles containing self‑replicating RNAs. Biochem Eng J. 2023, *191*, 108814. https://doi.org/10.1016/j.bej.2023.108814

Line 308 (previous manuscript) – we added:

“At 24 hours post‑transfection, one plate was removed from the incubator, and cell viability was measured using a 3‑(4,5‑dimethylthiazol‑2‑yl)‑2,5‑diphenyltetrazolium bromide (MTT) assay as described in [36].”

Supporting reference: [36] Zhurinov, M.Z.; Miftakhova, A.F.; Keyer, V.; et al. Glycyrrhiza glabra L. Extracts and Other Therapeutics against SARS‑CoV‑2 in Central Eurasia: Available but Overlooked. Molecules 2023, *28*(16), 6142. https://doi.org/10.3390/molecules28166142

Line 539 (previous manuscript) – we added a reference to the original study by Konvalinka. The text states: “These findings are consistent with the previously published observations of Konvalinka et al. [42].”

Line 547 (previous manuscript) – we added: “These findings are consistent with the previously published observations of Konvalinka et al. [42].”

 

Comment 2

In Figures 4 and 5: please insert scale bars. In Figure 4, Please try to represent your MTT results as a percentage.

Response:

Figures 4 and 5 have been completely revised. In the revised manuscript, all micrographs of cell cultures now include scale bars.

The MTT results are now presented as a percentage of the control (mock‑transfected cells).

 

Comment 3

Use an identical Y-axis scale for Fig. 5d. How did you do the statistical analysis? Also, how did you quantify using Lenti-X?

Response:

In the revised manuscript, uniform Y‑axis scales are used for all graphs representing yields. We performed new experiments to measure the production of extracellular p24 (a surrogate for VLP particles) using a p24 ELISA. Accordingly, the new Figure 5 and  Figure 6 in the revised manuscript now have Y‑axes expressed in terms of p24 concentration (ng/mL).

Regarding Lenti‑X GoStix quantification: we have replaced the semi‑quantitative GoStix Plus data with a fully quantitative p24 ELISA.

 

Comment 4

Reorganize Figure 6.

Response:

We have completely reorganized Figure 6 (from the previous manuscript). The revised manuscript now presents the relevant information on LV production and titration in the newly revised Figure 7. The layout has been simplified, and the axis labels and legend have been improved for clarity.

 

Comment 5

Discussion is quite redundant, particularly lines 702-781. Try to shorten.

Response:

We have significantly reworked and shortened the Discussion. Redundant phrases (lines 702–781 in the previous manuscript) and repetitive statements have been removed.

 

Comment 6

In Line 700, where is Figure 8d?

Response:

We apologise for the erroneous reference to a non‑existent “Figure 8d”. This has been corrected. Figure 8 has now been included and illustrates the results of CAR‑T cell production using a vector produced in RCP cells, thereby demonstrating functionality.

 

Sincerely,

Alexandr Shustov, corresponding author

On behalf of all coauthors

Reviewer 4 Report

Comments and Suggestions for Authors

This manuscript presents a flavivirus-replicon-based lentiviral packaging strategy in which a yellow fever virus replicon is used to stably express HIV-1 Gag-Pol in HEK293FT-derived cell pools. The study shows that the T26S protease mutation largely removes the cytotoxicity associated with wild-type HIV-1 protease, enables establishment of replicon cell pools, and supports lentiviral vector production at titers comparable to conventional transient transfection workflows. The work is interesting because it explores an intermediate platform between transient transfection and fully stable producer cell lines, with possible relevance to gene therapy manufacturing and academic-scale CAR-T workflows. The concept is potentially useful, but the manuscript still needs clearer positioning of novelty, tighter validation of stability and biosafety claims, and more rigorous discussion of the system’s current boundaries.

specific comments

1. The study design is interesting, but the current comparison framework is still a bit narrow for supporting the broader manufacturing claim. Since the manuscript states that this platform may provide “a new direction in packaging platform development,” the authors should explain more clearly whether the key benchmark is against standard transient transfection, pseudostable pools, or published producer cell lines such as LentiPro26, and why this comparison set is sufficient. Please also clarify whether similar replicon-assisted LV packaging studies have already been reported, how this system is technically distinct from them, and which part should be viewed as the true innovation here: the flaviviral backbone, the T26S-enabled long-term maintenance, or the packaging workflow itself.
2. The current writing structure can be made easier to follow, especially for readers outside the immediate vector-manufacturing field. The manuscript would read more clearly if the Results section first separated “proof of concept,” “cell-pool establishment,” “production optimization,” and “scale-up/safety validation” into a more explicitly progressive logic, because several design details are currently introduced before the biological rationale is fully settled. Please also tighten the Discussion around one main message—what problem this system solves now, what it does not solve yet, and where it realistically sits between transient transfection and true stable producer lines.
3. The analytical rigor needs a bit more detail in several places. The manuscript should explain more fully how replicate numbers were defined across viability, Western blot densitometry, VLP readout, flow-cytometric titration, and scale-up experiments, and whether the statistical assumptions behind the selected tests were checked before using t-tests or one-way ANOVA. Please also clarify the gating strategy, linear range for titer calculation, criteria for choosing 1–20% positive cells for back-calculation, and whether assay variability was evaluated across independent transfections, operators, or frozen-thawed batches.
4. “The T26S mutation effectively abolishes replicon-induced cytotoxicity” feels a little too strong as written; could the authors show more direct evidence that the reduced toxicity is specifically due to attenuated protease activity rather than altered replicon fitness, reduced expression burden, or different intracellular RNA levels?
5. “Stable replicon-containing cell pools” needs tighter support; could the authors define what was actually measured as stability, report how many passages were tested under and without selection, and clarify whether replicon RNA integrity, copy number, or sequence preservation was directly examined rather than inferred only from VLP output?
6. The biosafety section is important but still somewhat limited; could the authors explain the detection sensitivity, positive-control performance, assay limit of detection, and whether a 15-day amplification assay plus VSV-G/p24 readout is sufficient to support the broader statement that the system may reduce RCL risk?
7. The scale-up result is promising, but the manufacturing relevance would be stronger if the authors added more process detail; could they report batch-to-batch consistency, recovery losses at each downstream step, cell-specific productivity, and how this workflow compares practically with their own transient-transfection baseline using the same transfer vector and assay system?

Author Response

Response to Reviewer 4

We thank the reviewer for the thorough and constructive evaluation of our manuscript. We have carefully considered all comments and have substantially revised the manuscript to address each point. Below we provide a point‑by‑point response.

 

Reviewer’s comment:

The study is interesting, but the comparison framework is narrow.

Response:

In the revised manuscript, we have now performed a direct side‑by‑side comparison between the RCP‑based system and conventional three‑plasmid transient transfection of HEK293FT cells under identical conditions (5‑layer stacks, same transfection method, same titration protocol). The results are presented in Table 1 (Section 3.6, lines 807–809) and Table 2 (lines 822–824).

Regarding the reviewer’s comment on the comparison framework, we kindly note that our work is an original research study, not a review. We present to the readers a novel, original platform for vector packaging and compare its properties with those of the traditional platform. Furthermore, we demonstrate the functionality of the obtained packaged vector by producing a cell product under conditions closely resembling a real‑world application scenario.

 

Reviewer’s comment:

Needs clearer positioning of novelty, tighter validation of stability and biosafety claims, and more rigorous discussion of the system’s current boundaries.

Response:

1. Positioning of novelty

In the Introduction we now explicitly state that this is the first use of a flavivirus‑derived replicon for stable lentiviral packaging protein expression. The text reads: “In this study, we describe a novel LV packaging strategy based on a replicon derived from the model flavivirus yellow fever virus (YFV)”.

2. Tighter validation of stability

Section 3.4 (lines 676‑696; Figure 5) presents p24 ELISA data showing sustained Gag‑Pol expression over ten passages under puromycin selection. A one‑way repeated measures ANOVA confirms no significant effect of passage number.

Section 3.5 (lines 739‑750) reports Sanger sequencing of the entire Gag‑Pol insert after ten passages: “No nucleotide substitutions were detected … and the attenuating T26S mutation remained present.”

3. Biosafety claims

Section 3.8 (lines 849‑873; Figure 9) describes the two‑phase p24 ELISA‑based RCL assay. No RCL was detected in 5 × 10⁷ TU tested across two independent batches, demonstrating that flavivirus‑based replicons are not prone to productive recombination under the tested conditions.

In the Discussion (lines 965‑968), we state that YFV 17D is BSL-2 agent and “the replicon lacks the structural genes of YFV, it cannot generate infectious viral particles, which supports the biosafety of this platform”.

4. More rigorous discussion of the system’s current boundaries

In the Discussion, we have added a paragraph describing the limitations of our system (lines 969‑979).

 

Specific comment 1

The authors should explain more clearly… Clarify the benchmark (transient transfection, pseudostable pools, or LentiPro26), distinguish this system from any prior replicon‑assisted LV packaging studies.

Response: The benchmark is conventional transient transfection (direct side‑by‑side data in Section 2.10, Table 1 and Table 2). Regarding LentiPro26 (we mention this original development in Discussion in line 921), our platform is distinct by using a flavivirus replicon to express Gag-Pol. To our best knowledge, our study is the first in the field to use a flavivirus replicon for LV packaging.

 

Specific comment 2 – Writing structure and readability

The manuscript would read more clearly if the Results section first separated “proof of concept,” “cell‑pool establishment,” “production optimization,” and “scale‑up/safety” into a more explicitly progressive logic. Please also tighten the Discussion around one main message.

Response:

As suggested by the reviewer, we have restructured the entire manuscript to present a clear logical flow: “proof of concept,” “cell‑pool establishment,” “production optimization,” and “scale‑up/safety.” We have also shortened the Discussion.

The Results section is now organised as follows:

3.1 Engineering YFV Replicons for Transient Gag‑Pol Expression (proof of concept)

3.2 T26S Protease Mutation Reduces Replicon Cytotoxicity and Affects Gag Processing (proof of concept)

3.3 Construction of Selectable Replicons and Establishment of Replicon Cell Pools (cell‑pool establishment)

3.4 Gag‑Pol Expression in RCP Cells During Serial Passaging With and Without Selection (production optimisation)

3.5 Sequencing Reveals No Mutations in the Gag‑Pol Insert (stability)

3.6 Production of Lentiviral Vectors Using RCPs or Conventional Transfection (scale‑up and benchmarking)

3.7 Production of CAR‑T Cells Using the Vector Produced With RCP (functional validation)

3.8 Assessment of RCL (safety)

 

Specific comment 3 – Analytical rigor and statistical details

Explain more fully how replicate numbers were defined across viability, Western blot densitometry, VLP readout, flow‑cytometric titration, and scale‑up experiments, and whether statistical assumptions were checked. Please clarify the gating strategy, linear range for titer calculation, criteria for choosing 1–20% positive cells.

Response:

We have added the requested details:

Replicate definitions: In Section 2.16 (lines 472–479), we specify that three biological replicates (independently seeded cultures) were performed for all experiments unless otherwise indicated. For the MTT assay, each biological replicate contained four technical replicates (total 12 wells per condition). For Western blot densitometry, one representative blot is shown from a series of experiments that gave qualitatively consistent results.

Statistical assumptions: For time‑series data (sequential passages), we used repeated‑measures ANOVA (RM ANOVA). Two‑tailed Student’s t‑test was used for comparisons of independent groups. These details are now included in Section 2.16.

Gating strategy and titer calculation: The gating strategy for flow cytometry is shown in Supplementary Figure S4 and described in Section 2.9 (lines 304–309). The linear range for titer calculation (dilution that yielded between 1% and 20% GFP‑positive cells) was used as recommended in “Sena-Esteves et al. Titration of lentiviral vectors”. (This reference is included in the Reference list #38).

 

Specific comment 4 – Strong wording on T26S abolishing cytotoxicity

“The T26S mutation effectively abolishes replicon‑induced cytotoxicity” feels too strong. Could the authors show more direct evidence that the reduced toxicity is specifically due to attenuated protease activity rather than altered replicon fitness, reduced expression burden, or different intracellular RNA levels?

Response:

We agree that the original wording was too strong. We have modified the statement in the Results (lines 572–573) to read: “These results demonstrate that the T26S mutation reduces replicon‑induced CPE to levels undetectable by the MTT assay.”

Direct evidence linking the T26S substitution to attenuated protease activity is provided by the Western blot (Figure 2e, line 586) which shows reduced Gag processing in the T26S mutant compared to wild‑type. Importantly, under puromycin selection, p24 production remains stable over ten passages (Section 3.4, Figure 5, line 687), indicating that replication of the T26S replicon is not detrimental to RCP cells. These results are in good concordance with the original study by Konvalinka et al. [#42 in the References list], who already extensively characterised the T26S mutation. We discuss our results in comparison to those of Konvalinka et al. (Discussion, lines 918 and following).

 

Specific comment 5 – Definition of “stable replicon‑containing cell pools”

“Stable replicon‑containing cell pools” needs tighter support. Define what was measured as stability, report how many passages were tested under and without selection, and clarify whether replicon RNA integrity, copy number, or sequence preservation was directly examined.

Response:

Definition of stability: stability has been demonstrated as (i) sustained p24 production in the culture supernatant for at least ten passages under selection (Section 3.4, lines 676–696), and (ii) genetic integrity of the Gag‑Pol insert confirmed by sequencing (Section 3.5, lines 739–750).

Passage numbers: Under continuous selection, ten passages (P1–P10) were tested (Figure 5). In the withdrawal experiment, cells were passaged five times without selection (P1–P5) and then re‑selected for five additional passages (P8–P12) (Figure 6).

Direct examination of replicon RNA: We have now performed RT‑PCR to sequence the entire Gag‑Pol insert after passage P10 (Section 3.5).

Thus, the stability claim is now supported by both phenotypic (p24) and molecular (sequencing) evidence.

 

Specific comment 6 – Biosafety section and RCL assay sensitivity

Explain the detection sensitivity, positive‑control performance, assay limit of detection, and whether a 15‑day amplification assay plus VSV‑G/p24 readout is sufficient to support the broader statement that the system may reduce RCL risk.

Response:

We have clarified these points:

Detection sensitivity and LOD: The RCL assay is a p24 ELISA‑based two‑phase protocol (Section 2.15, lines 447–471). The LOD was defined as the p24 concentration (pg/mL) corresponding to the mean OD₄₅₀ of naïve C8166 medium plus three standard deviations. The calculated LOD was 8 pg/mL (Supplementary Table S6).

Positive control: We were unable to include a replication‑competent lentivirus (RCL) positive control because current regulations at our institution forbid working with RCL. Nevertheless, the two‑phase assay with p24 detection is widely used for proof‑of‑concept studies, and the relevant references are included in the text (refs. 39, 40, 41). In the Discussion, we now explicitly state that future validation for clinical applications will require more stringent RCL testing with spike‑in controls (lines 975–976).

Sufficiency of the assay: During preparation of the revised manuscript, we performed new RCL assay in which the amplification phase was extended to 21 days and the indicator phase to 7 days. These durations are recommended for RCL assays (refs. 39, 40, 41).

 

Specific comment 7 – Scale‑up and process details

Report batch‑to‑batch consistency, recovery losses at each downstream step, cell‑specific productivity, and how this workflow compares practically with your own transient‑transfection baseline using the same transfer vector and assay system.

Response:

We have added the requested details in Section 3.6 and Table 1:

Batch‑to‑batch consistency: Two independent production runs (RCP repeats 1 and 2) are shown in Table 1. The functional titers of unconcentrated supernatants were 4.20×10⁶ and 6.06×10⁶ TU/mL.

Recovery losses: Table 1 includes stage yields (%) for each process step (supernatant collection → TFF → centrifugal pelleting → final yield). For RCP, the overall recovery was ∼70–72%.

Comparison with own transient baseline: The conventional transient transfection runs were performed in parallel using HEK293FT cells, same transfection method, same culture vessels, and the same titration protocol (Section 2.10). The yields are directly comparable (Table 1). We also present the specific infectivity values (TU/pg p24, Table 2).

 

Sincerely,

Alexandr Shustov, corresponding author

On behalf of all coauthors

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

I want to commend you on making a very thorough and scientifically rigorous revision. I must say, it is rare to see a response to reviewers that involves returning to the bench to generate such substantial new experimental data. Specifically, your addition of the direct, side-by-side 5-layer stack benchmarking experiments against the conventional transient transfection system adequately addresses my primary concern. Transparency in these yields adds significant value to your manuscript. Furthermore, abandoning the unvalidated qPCR assay in favor of the industry-standard, p24 ELISA-based two-phase RCL detection assay resolves my previous concerns regarding analytical sensitivity and limit of detection (LOD).

The inclusion of the functional characterization data, specifically the one utilizing the RCP-produced vector to transduce primary human CD4+ and CD8+ T lymphocytes and achieving a 77.41% CAR+ cell population, significantly elevates the impact of this study and proves its real-world relevance.

 

Finally, your willingness to responsibly temper your claims regarding manufacturing comparability, replicon stability, and flavivirus compartmentalization shows a high level of scientific integrity.

This is a solid, well-supported research that makes a valuable contribution to lentiviral vector manufacturing. Congratulations on a job well done.

Author Response

Dear Reviewer,

Thank you for your time and thorough analysis. We truly appreciate your recognition of our efforts to develop novel lentiviral packaging systems. This work may one day yield viral‑derived systems that combine genomic integration of transgene with robust proliferative activity in packaging cell cultures, ultimately providing therapeutic vectors to laboratories and clinics in need. We continue working toward that goal.

 

Sincerely,

Alexandr Shustov

On behalf of all co‑authors

 

Reviewer 2 Report

Comments and Suggestions for Authors

I would like to thank the authors for their thorough and constructive responses to my comments. The major concerns regarding direct comparison with conventional systems, stability assessment, genetic sequencing, biosafety classification, and primary T cell transduction have all been adequately addressed in the revised manuscript. The addition of CAR-T manufacturing data using primary human T cells, achieving 77.4% CAR+ efficiency, is particularly valuable and directly supports the therapeutic applicability of the platform.
One observation remains. Table 2 shows that RCP-produced vectors have lower specific infectivity compared to the conventional system, approximately 379-472 TU/pg p24 versus 710-907 TU/pg p24. The authors acknowledge this difference but attribute it to the limited number of experiments. While I appreciate the honest acknowledgment and agree the work is a first-generation platform with considerable room for optimization, a brief discussion of what this difference might mean practically in the context of CAR-T manufacturing dose requirements would strengthen the manuscript.

Author Response

Dear Reviewer,

Thank you very much for your thorough work on review and positive assessment of our manuscript.

 

Reviewer’s comment: One observation remains. Table 2 shows that RCP-produced vectors have lower specific infectivity compared to the conventional system, approximately 379-472 TU/pg p24 versus 710-907 TU/pg p24. The authors acknowledge this difference but attribute it to the limited number of experiments. While I appreciate the honest acknowledgment and agree the work is a first-generation platform with considerable room for optimization, a brief discussion of what this difference might mean practically in the context of CAR-T manufacturing.

Response: We have added a brief discussion on the topic. The relevant text has been inserted into the Discussion (lines 955–964) in the second revision of our manuscript.

 

We much appreciate your recognition of our efforts to develop novel lentiviral packaging systems. This work may one day yield viral‑derived systems that combine genomic integration of a transgene with robust proliferative activity in packaging cell cultures, ultimately providing therapeutic vectors to laboratories and clinics in need. We continue working toward that goal.

 

Sincerely,

Alexandr Shustov

On behalf of all co‑authors