Targeted Hyaluronan Degradation Enhanced Tumor Growth Inhibition in Gastrointestinal Cancer Models

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript entitled “Targeted Hyaluronan Degradation enhanced Tumor Growth Inhibition in Colorectal Cancer Models” was well designed in experimental approaches. The result expressed in a well order and clarity. The main focus of the manuscript is targeting and degradation of stromal hyaluronic acid by a diverse AbEn molecules and the comparison of them with each other. Also they studied in vivo tumor growth inhibition of trispecific TAVO423 AbEn in combination with different drugs ((5-FU, anti-PD-L1 monotherapy, PD-L1/CD3 bispecific antibody, and CD318-MMAE 545 ADC) in colorectal cancer animal model as tumor size and volume. However it needs some corrections to improve quality of the paper.

1. In p-age 2, line 95, the phrase “TAVO423 is a trispecific FAP × LRRC15 × HYAL AbEn that”, should be written as “TAVO423 is a FAP × LRRC15 × HYAL trispecific AbEn that “.
2. The abstract is very general. In the abstract main treatment modal (trispecific AbEn) was highlighted should be revised completely and written in more detailed. The authors should be explain methods in specific and detail fashion. Also the results should be explained and compared for modalities used in the manuscript with numbers and digits.
3. What is the aim of using TAVO412 in this study? Why was the author’s interest to study combination of TAVO412 with PD-L1 × HYAL?
4. It should be better if the authors include a table for all drugs, agent and mAbs constructs used in the study.
5. To improve and to more clarification for the first readers some articles may be considered in the discussion (DOI: 1016/j.ijbiomac.2022.03.057 and DOI: 10.1038/s41392-021-00484-9).

Author Response

The manuscript entitled “Targeted Hyaluronan Degradation enhanced Tumor Growth Inhibition in Colorectal Cancer Models” was well designed in experimental approaches. The result expressed in a well order and clarity. The main focus of the manuscript is targeting and degradation of stromal hyaluronic acid by a diverse AbEn molecules and the comparison of them with each other. Also they studied in vivo tumor growth inhibition of trispecific TAVO423 AbEn in combination with different drugs ((5-FU, anti-PD-L1 monotherapy, PD-L1/CD3 bispecific antibody, and CD318-MMAE 545 ADC) in colorectal cancer animal model as tumor size and volume. However it needs some corrections to improve quality of the paper.

Author’s response: We thank the reviewer for the positive comments.

Suggestions:

1. In page 2, line 95, the phrase “TAVO423 is a trispecific FAP × LRRC15 × HYAL AbEn that”, should be written as “TAVO423 is a FAP × LRRC15 × HYAL trispecific AbEn that“.

Author’s response: We thank the reviewer for the advice. The sentence has been revised as recommended (Page 3, Line 100).

2. The abstract is very general. In the abstract main treatment modal (trispecific AbEn) was highlighted should be revised completely and written in more detailed. The authors should be explain methods in specific and detail fashion. Also the results should be explained and compared for modalities used in the manuscript with numbers and digits.

Author’s response: We thank the reviewer for the comment. As suggested, we have thoroughly revised the Abstract to include more specific methodological details and quantitative results (Page 1, Line 17-37). Key values such as TGI percentages and fold changes in immune cell infiltration have been added to provide a clearer and more concrete summary of our findings. The revised the abstract was also provided as below.

“The dense hyaluronan (HA)-rich stroma in solid tumors can prevent effective tumor growth inhibition by hindering drug delivery and immune cell infiltration. However, the degradation of HA alone by systemic delivery of hyaluronidase has not shown significant improvement of tumor growth inhibition. In this study, we targeted hyaluronan (HA) degradation by using antibody-enzyme (AbEn) molecules by fusing antibodies to a recombinant human hyaluronidase (HYAL). The AbEn molecules were stable, retained both antigen-binding and enzymatic activities, and demonstrated a prolonged serum half-life of 132 hours in rodent models. In the HA-rich colorectal cancer model, the cancer-associated fibroblast (CAF)-directed AbEn, TAVO423 (FAP x LRRC15 x HYAL trispecific antibody) achieved greater intratumoral HA depletion resulting in superior tumor growth inhibition compared to untargeted HYAL. Furthermore, the combination of TAVO423 in combination with other solid tumor cell targeting modalities such as 5-fluorouracil (5-FU), anti-PD-L1 monoclonal antibody, a PD-L1 × CD3 bispecific T-cell engagers (TCEs), and a CD318-targeting antibody-drug conjugates (ADCs) all demonstrated enhanced tumor growth inhibition (TGI) values of 49%–67% as compared to the respective monotherapy TGI values of 1%–28%. In addition, TAVO423 improved the antitumor response of a 5T4 × CD3 TCE with an increase in TGI from 73% to 92% in an in vivo HA-rich pancreatic cancer model. The CAF-targeted HA degradation mediated by TAVO423 also reversed immune exclusion by increasing the density of CD8+ tumor-infiltrating lymphocytes (TILs) by 6-9-fold and synergized with PD-1 blockade to enhance TGI from 33% to 51% in an in vivo immunocompetent EMT-6 breast cancer model. These findings demonstrated the broad potential of the modular AbEn platform for targeted HA degradation to overcome barrier entry in stromal HA-rich solid tumors.”

3. What is the aim of using TAVO412 in this study? Why was the author’s interest to study combination of TAVO412 with PD-L1 × HYAL?

Author’s response: We thank the reviewer for the comment. TAVO412 is a trispecific antibody (EGFR × cMet × VEGF) that has demonstrated broad antitumor activity across in vivo solid tumor models. We selected this molecule as a representative multi‑targeted therapeutic to evaluate whether targeted HA degradation could further enhance the efficacy of an already potent antitumor agent. For targeted enzyme delivery, PD‑L1 was chosen because it is highly expressed on RKO cells. The combination of TAVO412 with a PD‑L1 × HYAL was designed to determine whether stroma remodelling, mediated via a tumor‑associated antigen (TAA)-targeting strategy, could synergize with this potent antitumor drug. We have clarified this rationale in the Results section 3.3.

4. It should be better if the authors include a table for all drugs, agent and mAbs constructs used in the study.

Author’s response: We thank the reviewer for the comment. As suggested, we have added a new Supplementary Table S7 listing all constructs, drugs, and antibodies used in this study, along with their targets, formats, and applications.

 

5. To improve and to more clarification for the first readers some articles may be considered in the discussion (DOI: 1016/j.ijbiomac.2022.03.057 and DOI: 10.1038/s41392-021-00484-9).

Author’s response: We thank the reviewer for these relevant references. We have now cited both suggested articles in the Discussion section to provide broader context regarding HA-based therapeutic strategies and stromal targeting (Page 18, Lines 700-703).

Reviewer 2 Report

Comments and Suggestions for Authors

Figure 6E. X axis should use consistent scale, while a gap would be acceptable, change of scale is not.
FAP/LRRC15/HA colocalization in fig. 4 - it is hard to tell where author see this colocalisation - especially since they show pictures of different tumours or different parts of the same tumour. Moreover, there is a huge discrepancy in staining between those three. Nuclear counterstain is only visible in HA. 
Overall the size and quality of IHC images does not warrant any real assessment of staining quality and thus one cannot really validate and review the findings.
All major findings are reported on a single cell line RKO, those should be independently validate on at least another cell lines.
Authors mention flow cytometry based assessment, but those cannot be found within results
Table S5 should be a figure 
CD11b/myeloid cells require further analysis as this subset contains both pro and anti-inflammatory cells. Similarly to T cells, while an increase in Cd3 is partially due to Tc rise, the rest could be due to accumulation of Tregs which was not assessed.
Authors should work on language quality to improve the readability and general flow of the text. Current version is relatively hard to follow.

Author Response

Suggestions:

1. Figure 6E. X axis should use consistent scale, while a gap would be acceptable, change of scale is not.

Author’s response: We thank the reviewer for this comment. We have revised Figure 6E to maintain a consistent linear scale by implementing a broken axis between 15% and 30%. The X-axis now displays two segments: 0-15% (with ticks at 0, 5, 10, 15) and 30-90% (with ticks at 30, 60, 90), ensuring uniform scaling while optimally accommodating the actual data range. We have also updated the figure legend (Page 18, Lines 693-695) to clearly indicate the use of a broken axis and emphasize that a consistent linear scale is maintained throughout the figure.

 

2. FAP/LRRC15/HA colocalization in fig. 4 - it is hard to tell where author see this colocalisation - especially since they show pictures of different tumours or different parts of the same tumour. Moreover, there is a huge discrepancy in staining between those three. Nuclear counterstain is only visible in HA.

Author’s response: We thank the reviewer for this insightful comment. The images presented indeed originate from adjacent sections of the same tumor sample, which is a technical limitation for demonstrating precise spatial colocalization. Our conclusion of constitutive expression of FAP, LRRC15, and HA within the tumor stroma was based on their consistent presence across multiple independent samples. To avoid overinterpretation, we have revised the manuscript text from "colocalization" to "co-expression" to reflect our findings more accurately from serial sections (e.g., in the Results section describing Figure 4A, Page 12, Lines 504, 510-511). We agree that multiplex immunofluorescence would be required for definitive colocalization analysis, and this is a key objective for our future studies.

 

3. Overall the size and quality of IHC images does not warrant any real assessment of staining quality and thus one cannot really validate and review the findings.

Author’s response: We thank the reviewer for the comment. IHC staining was performed and evaluated by experienced pathologists. Although the image size may be limited by publication format, high-resolution original images are available upon request. All staining results were reproducible across multiple tumor samples and are representative of consistent biological observations.

 

4. All major findings are reported on a single cell line RKO, those should be independently validate on at least another cell lines.

Author’s response: We thank the reviewer for the suggestion. While the RKO model served as our primary HA-rich colorectal cancer model, key findings were corroborated in additional model including an immunocompetent breast cancer model (EMT-6), supporting the broader applicability of our approach. We have included additional data from a HA-high pancreatic cancer model (BxPC-3-HAS3) to validate the efficacy of the TAVO423 AbEn. These results are included in the revised manuscript as Supplementary Figure S9 and described in the Results section (Page 15, Lines 624-638).

 

5. Authors mention flow cytometry based assessment, but those cannot be found within results Table S5 should be a figure.

Author’s response: We thank the reviewer for this suggestion. We believe there may have been a misunderstanding regarding the tables. The flow cytometry-based binding affinity data (EC₅₀ values) for the anti-FAP and anti-LRRC15 antibodies are summarized in Supplementary Table S4, and the corresponding binding curves are provided in Supplementary Figure S7E (new revision version). Supplementary Table S5, conversely, summarizes the in vivo tumor growth inhibition (TGI) rates for combination therapies. We apologize for any confusion and hope this clarification is helpful.

 

6. CD11b/myeloid cells require further analysis as this subset contains both pro and anti-inflammatory cells. Similarly to T cells, while an increase in Cd3 is partially due to Tc rise, the rest could be due to accumulation of Tregs which was not assessed.

Author’s response: We thank the reviewer for this comment. We agree that further analysis of additional immune subsets (e.g., Tregs, myeloid subsets) would provide deeper mechanistic insight. This limitation has been explicitly added to the Discussion section (Page 20, Lines 773-779), where we also indicate that future studies will utilize advanced immune profiling to address this gap.

 

7. Authors should work on language quality to improve the readability and general flow of the text. Current version is relatively hard to follow.

Author’s response: We appreciate the reviewer's comment on language quality. Please note that the manuscript was professionally polished by a native English editor prior to submission. However, in direct response to this comment, we have meticulously re-examined the text to enhance readability and ensure the scientific narrative is as clear and compelling as possible. We are confident that the current version presents our findings with great clarity.

Reviewer 3 Report

Comments and Suggestions for Authors

This is a very nice contribution. 

The authors developed antibody-enzyme (AbEn) molecules with antibodies fused to a recombinant hyaluronidase to target the tumor extracellular stroma resulting in degradation of hyaluronan. The overall purpose being to make the tumor cells more accessible to the immune system and/or anti-cancer drug therapy.

The work is well-justified and the approach seems appropriate. The methods are well-described and the results justify the conclusions. Antibody targeting and hyaluronan degradation were demonstrated in models both in vitro and in vivo.

Author Response

Comments:

This is a very nice contribution.

The authors developed antibody-enzyme (AbEn) molecules with antibodies fused to a recombinant hyaluronidase to target the tumor extracellular stroma resulting in degradation of hyaluronan. The overall purpose being to make the tumor cells more accessible to the immune system and/or anti-cancer drug therapy.

The work is well-justified and the approach seems appropriate. The methods are well-described and the results justify the conclusions. Antibody targeting and hyaluronan degradation were demonstrated in models both in vitro and in vivo.

Author’s response: We are grateful to the reviewer for these positive comments.

Reviewer 4 Report

Comments and Suggestions for Authors

In the manuscript, Zhou et al., developed an “antibody-enzyme conjugate” (AbEn) to target HA in the tumor microenvironment. They demonstrated the therapeutic potential of the AbEn, especially TAVO423, in inhibiting tumor growth and enhancing the efficacy of other therapeutic approaches. Their study is comprehensive and provides a potent platform for treating HA-rich solid tumors. Here, I have a few suggestions to help further improve the manuscript.

1. The authors used EGFR x HYAL to evaluate the HA degradation efficacy and tumor growth inhibition. However, they used RKO model which has low EGFR expression to test the effects, additional in vivo model with high EGFR expression such as HCC827 mentioned in the manuscript should be considered.
2. To assess the targeted HA depletion in the tumor, the authors should provide data that quantifies the HA levels in tumors following systemic administration of each AbEn (including EGFR, PDL1, and TAVO423) and HYAL-Fc. They might use an approach such as ELISA on tumor lysates to quantify the HA levels.
3. Please state clearly if TAVO423 was also constructed with LALA IgG mutant.
4. Please include statistical analysis for all the in vivo tumor studies, and indicate the significance of differences between key groups.

Author Response

Comments:

In the manuscript, Zhou et al., developed an “antibody-enzyme conjugate” (AbEn) to target HA in the tumor microenvironment. They demonstrated the therapeutic potential of the AbEn, especially TAVO423, in inhibiting tumor growth and enhancing the efficacy of other therapeutic approaches. Their study is comprehensive and provides a potent platform for treating HA-rich solid tumors.

Author’s response: We thank the reviewer for the positive comments.

 

Suggestions:

1. The authors used EGFR x HYAL to evaluate the HA degradation efficacy and tumor growth inhibition. However, they used RKO model which has low EGFR expression to test the effects, additional in vivo model with high EGFR expression such as HCC827 mentioned in the manuscript should be considered.

Author’s response: We thank the reviewer for the suggestion. Although the HCC827 CDX model exhibited positive HA staining by IHC (Supplementary Table S3), its HA content appears lower than that of RKO, which was an important determinant for assessing the antitumor activity of targeted HA degradation (data not shown). As an alternative, we included additional in vivo data from the HA‑high BxPC‑3‑HAS3 CDX model to further validate the efficacy of TAVO423. These results are included in the revised manuscript as Supplementary Figure S9 and described in the Results section (Page 15, Lines 624-638).

2. To assess the targeted HA depletion in the tumor, the authors should provide data that quantifies the HA levels in tumors following systemic administration of each AbEn (including EGFR, PDL1, and TAVO423) and HYAL-Fc. They might use an approach such as ELISA on tumor lysates to quantify the HA levels.

Author’s response: We thank the reviewer for the suggestion. We have performed HA quantification via ELISA on tumor lysates from RKO xenografts treated with PD-L1 × HYAL (Supplementary Figure S6A), and TAVO423 (Supplementary Figure S6B). These data are now also described in the Methods and Results section (Page 11, Lines 495–500; Page 13, Lines 558–560).

 

3. Please state clearly if TAVO423 was also constructed with LALA IgG mutant.

Author’s response: Yes, TAVO423 was constructed with LALA mutations to minimize Fc-mediated effector functions. This is now explicitly stated in the Results section (Page 13, Lines 533-534) and Supplementary Figure S8, respectively.

 

4. Please include statistical analysis for all the in vivo tumor studies, and indicate the significance of differences between key groups.

Author’s response: We thank the reviewer for this important comment. We acknowledge that the combination therapy did not achieve statistical significance in some of our in vivo studies, which was likely attributable to the biological variability inherent in tumor models and the limited group size (n=5-6), reducing the statistical power. We assumed that optimizing the dosing regimen and further stratifying models based on HA levels could potentially expand the efficacy window. However, we would like to highlight that the therapeutic trend, demonstrated by the consistent improvement in tumor growth inhibition (TGI), was observed across multiple combination cohorts. This biological effect, coupled with the direct evidence of stromal HA depletion and increased T-cell infiltration, provided strong supportive evidence for our conclusion that targeted HA degradation could potentiate various anti-tumor agents. Therefore, we believed that the TGI data, in the context of these complementary pharmacological and immunohistochemical findings, robustly supported our central hypothesis.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

I am satisfied with authors response. Authors implemented some crucial additional data. 

Reviewer 4 Report

Comments and Suggestions for Authors

The authors addressed all my concerns. Thanks.