Interventional hydrogel microsphere vaccine as an immune amplifier for activated antitumour immunity after ablation therapy

The response rate of pancreatic cancer to chemotherapy or immunotherapy pancreatic cancer is low. Although minimally invasive irreversible electroporation (IRE) ablation is a promising option for irresectable pancreatic cancers, the immunosuppressive tumour microenvironment that characterizes this tumour type enables tumour recurrence. Thus, strengthening endogenous adaptive antitumour immunity is critical for improving the outcome of ablation therapy and post-ablation immune therapy. Here we present a hydrogel microsphere vaccine that amplifies post-ablation anti-cancer immune response via releasing its cargo of FLT3L and CD40L at the relatively lower pH of the tumour bed. The vaccine facilitates migration of the tumour-resident type 1 conventional dendritic cells (cDC1) to the tumour-draining lymph nodes (TdLN), thus initiating the cDC1-mediated antigen cross-presentation cascade, resulting in enhanced endogenous CD8+ T cell response. We show in an orthotopic pancreatic cancer model in male mice that the hydrogel microsphere vaccine transforms the immunologically cold tumour microenvironment into hot in a safe and efficient manner, thus significantly increasing survival and inhibiting the growth of distant metastases.

In this manuscript Liu et al. develop and study the effects of a hydrogel microsphere vaccine in the treatment of pancreatic cancer. The authors report that treatment of orthotopic KPC pancreatic mouse models with the hydrogel microsphere vaccine increased animal survival and inhibited the growth of distant metastasis. This effect is due to an activation of cDC1 cells and consequent antigen presentation and activation of CD8 Tcells. Overall, the manuscript is well written, well positioned in the state-of-the-art of the field and the authors provide clear assays that support their claims making this manuscript a good addition to the literature. However, the data presented does not fully support the conclusions and some issues need to be addressed to clarify the results: 1) Regarding the data in figure 2p and 2q, the authors should check if the hydrogel microsphere vaccine had any effects on cDC1s' proliferation and differentiation.
2) On figure 3k and 4b the authors show IHCs for CD45 on mice pancreatic tumors to assess whether IRES ablation and the hydrogel microsphere vaccine affected immune cell infiltration in orthotopic pancreatic tumors. Because the study focuses on cDC1s and CD8+ T cells, the authors should also perform staining for these immune cell types. This would allow the assessment of whether IRES ablation and the hydrogel microsphere vaccine being tested are able to promote the recruitment of these specific immune cells into the TME.
3) On line 289, while describing figure 4c, I believe the authors mistakenly wrote FLT3L, as in the figure it is the CD40L's highest concentration that lasted from 48 hours to 96 hours. Figure 4b, the hydrogel microsphere vaccine group is labelled as "IRE+Hydrogel microspheres vaccine", while on figures 4e, 4f, 4k and 4l it is labelled as "Group 5". Please change these labelled to something more uniform. Also, I believe calling it Group 5 might cause confusion as all the other groups are controls. 5) Please explain, in the results section, the treatment groups presented on figures 4i, 4j, 4m and figures 5a-5e. Are these the same as Groups 1 to 4 presented in other figures? This should be clarified and the names given should be simplified in a way that the reader immediately knows to which group are the authors referring to while describing the results. 6) In Materials and Methods the authors say that tumor volume was evaluated every 3 days but the method used is not mentioned. Also, on the experiment depicted on figure 6, did the authors asses tumor volume and weight at the time of euthanasia? Was there any difference in tumor volume between groups? 7) In Materials and Methods it is not mentioned how did the metastasis were evaluated. Also, the authors mention on line 406 that the metastasis that are being evaluated are "distant subcutaneous metastases". Orthotopic pancreatic mouse models don't develop subcutaneous metastasis, but rather liver and lung, and sometimes peritoneal. dissemination can also be seen. Could you please clarify this issue?

4)
8) The authors never mention some important immune cells that are crucial in the TME and inflammation response, like macrophages. Analysis for these inate imune cells and, additionally, if these cells change fate -from pro to antitumorigenic, or vice-versa, especially in the survival studies, should be included.
Minor points: Figure S7 misses legend for panels c and d. Figure 5k can be deleted. I believe it is a repetition of 1d.
co-inhibition coupled with αCD40 invoked immune repolarization and was an attractive therapeutic approach for pancreatic cancer immunotherapy development. In their report, MEK and autophagy co-inhibition triggered the release of inflammatory cytokines in cancer cells, and these signals affected macrophages' polarization to favor an M1-like, antigen-presenting phenotype.
Compared with MEK and autophagy co-inhibition, irreversible electroporation ablation is an emerging local therapy for cancers. This non-thermal ablation therapy directly induces cancer cell apoptosis and triggers a strong inflammatory response. In our study, we highlighted the amplification of the CD8 + T-cell immune cascade induced by the activation of cDC1s. Compared with M1-like macrophages, cDC1 is currently recognized as the strongest antigen-presenting cell targeting CD8 + T cells, and this immune cascade amplification has superior efficiency and specificity [2,3].
In addition, Jiang and colleagues reported that the therapy mostly increased the CD301bsubset of cDC2, whereas the CD301b + subset was decreased in the conditions where αCD40 was used. More importantly, the moderate loss of cDC1s in the tumor microenvironment may impair antigen presentation to CD8 + T cells and attenuate the expansion and activation of CD8 + T cells. In our study, we highlighted that the combination of local and sequential administration of FLT3L and αCD40 could significantly enhance the cDC1-mediated antigen-presentation cascade. As we mentioned in the Introduction and Discussion sections, since myeloid cells are highly plastic, different combinations of cytokines and delivery strategies may lead to differences in the differentiation and activation status of myeloid cells. Here, we have addressed the critical issue of efficient local amplification of tumor-resident cDC1s.
Based on the hydrogel vaccine system, we proposed a novel local therapeutic strategy for pancreatic cancer, and elucidated the immune basis of this anti-tumor immune amplification strategy from three levels: local antitumor immune remodeling, draining lymph node immune response, and systemic adaptive antitumor immunity. Recent evidence has also suggested that the irreversible electroporation may induce immune activation and epitope expansion effects in prostate tumors [4]. The amplification of cDC1-mediated antigen-presentation cascade using hydrogel microsphere vaccine undoubtedly promoted the translational application of irreversible-electroporation based local therapies. In the revised manuscript, we have also reinforced our advantages and innovations with this previous study (Discussion Section, Page 15， line 21-27 ). We sincerely hope sincerely this will meet your criteria. [1]

Reviewer #2
This article reports the potential of hydrogel microspheres for enhancing antitumor immunity after tumor ablation. The authors propose that immune amplification function of hyaluronic acid-based hydrogel microspheres could play a major role in boosting antitumor immunity. I evaluate that the manuscript does not meet the standard of Nature Comm. My comments are as follows. Response：We thank the reviewer for the valuable comment. We understand the concerns of the reviewer. Hydrogel microspheres, liposomes, and calcium carbonate nanoparticles have been widely used in the development of drug delivery systems because of their good bio-affinity and bio-responsiveness. We can understand that the reviewer raised this concern from the perspective of materials science. We accept that developing entirely new materials and drug delivery systems is very innovative work.
Meanwhile, we believe that the utilization and modification of current biomaterial carriers for new biological or medical applications offers equivalently significant innovative and clinical translational value [1][2][3].
In this study, we developed a locally injectable "hydrogel microsphere immunomodulatory chamber" using common and biocompatible materials to achieve programmatic release of FLT3L and CD40 agonist. One of the most significant characteristics of the vaccine is the precise amplification of the antitumor immune effect of cDC1/CD8 + T cells. Within the context of immunotherapy, the strategy of targeted amplification of the immune cascade cDC1/CD8 + T is highly innovative. The amplification and activation of CD8 + T cells are the core of current immunotherapy.  Response：Thanks for the valuable comment. In the revised manuscript, we simulated pH value of the tumor microenvironment and performed the release of CD40L and FLT3L from the hydrogel microspheres in 1 × PBS solution at pH 6.8, which was shown in Figure 2l.
In this revision, the release of CD40L from the hydrogel microspheres system in 1 × PBS at pH 7.4 and pH 6.8 was exhibited. Here, the pH value of 6.8 was chosen for two reasons. Firstly, the pH value of 6.8 have been widely accepted as the pH value for simulating the tumor microenvironment in vitro [1,2]. Secondly, according to  Response：We thank the reviewer for the valuable comment. The size and zeta potential of cationic liposomes before and after Cas9 plasmid entrapment have been presented in Figure 2. The Z-average diameter of cationic liposomes is 140 nm, and the zeta potential is 24.9 mV, and the particle size of liposomes does not change much after loading the Cas9 plasmid. The results showed that the Z-average diameter is 124 nm, and the zeta potential drops to 9.52 mV of cas9@lip.
We also provide more extensive physicochemical characterization data of hydrogel microspheres in Supplementary Fig. 1c Response：Thanks for the comment. In the hydrogel microsphere system, CD40L was first loaded into the calcium carbonate nanoparticles, and the calcium carbonate nanoparticles could be decomposed under acidic pH conditions to release the loaded CD40L. The specific preparation process was to first load CD40L into HOOC-PLGA-PEG-PLGA-COOH micelles, and then form calcium carbonate nanoparticles around the micelles through biomineralization. The release of CD40L required calcium carbonate erosion in the slightly acidic environment of tumors.

Comment 6. The authors should show evidence on the actual acidity values of residual tumors after tumor ablation.
Response：Thanks for the valuable suggestion. In the revised manuscript, we assessed the pH value of the tumor before and after the ablation therapy using a pH microelectrode (CAT#pH-500c 400-600um diameter, UNISENSE company, Denmark).
About 3~4 mm of the pH micro-electrode was punctured into the tumor carefully. The pH values were measured before ablation, 0h (immediately after ablation), 24h, 48h and 72h after the ablation therapy. The mean pH value for each group is 6.84, 6.81, 6.75,6.70, and 6.67, respectively ( Supplementary Fig. 6).
There are no significant differences in pH values before and immediately after the ablation therapy. The intratumour pH values were slightly but significantly reduced after 24, 48h, and 72h after the IRE ablation therapy. The slight decrease in pH value after ablation was compatible to the release of CD40L from the hydrogel vaccine. It was previously reported that the intratumour pH value was slightly decreased after ablation therapy [1,2]. This decrease in pH may be due to the weakening of microvascular drainage, accumulation of metabolites, and the release of waste products caused by cell damage and death. The results showed that the hydrogel vaccine had minimal effect on pancreatic tumor microenvironment without IRE. We think that this misunderstanding may be caused by the ununiformed figure legends. In the revised manuscript, the figure legends labeling control and vaccine group are uniformed according to Comment 4-5, Reviewer 3.
Thanks for the valuable suggestion, and we hope these revisions could clarify the issue.   Fig. 23b). This result demonstrates the efficient knockout of PD-L1 by the Cas9 system. We thanks again for the valuable comment.   immunological development. We have tried our best to clarify this issue, and we sincerely hope these results will meet your criteria. Flow cytometry analysis showed that there was no significant difference in the number of infiltrating CD45 + cells across the IRE, IRE combined with αPD-L1, and IRE combined with hydrogel vaccine group (Figure 7a). Further results suggested that the proportion of CD103 + CD11b -cDC1s in subcutaneous tumors did not change significantly, indicating that cDC1 may not be a direct effector inducing the abscopal effects (Figure 7b). Considering that previous study reported that CD40L may also have a function on tumor-infiltrating macrophages [1], the F4/80 + CD11b + total macrophages, CD206 + CD80 + and CD163 + CD80 + M2-like macrophages were then analyzed. The results showed that in situ administration of hydrogel vaccine had no significant effect on the proportion and phenotype of macrophages in the distant subcutaneous tumors (Figure 7b). Finally, the infiltration and function of CD8 + T cells were analyzed. After the hydrogel vaccination, there is a significant increase in the proportion of infiltrating IFN-γ + CD8 + T cells in the subcutaneous tumors (Figure 7c).  The detailed amounts of plasma and Liposome corresponding to each ratio in the pilot test. Response：Thanks for the kind reminder. We sincerely apologize for our carelessness.
The plasmid-to-lip ratio (w/v) used in the  Fig. 4a).
Considering the potential role of CD40L on macrophages, we also conducted a preliminary analysis of the bone-marrow derived macrophages. Positive expression of CD206 or CD163 is used as independent marker to distinguish M2-like macrophages.
The results showed that the additional CD40L or FLT3L slightly promoted the differentiation of BMDC into M1-like macrophages and cDC1s (Supplementary Fig.   4b, c). The sequential administration of FLT3L and CD40L, including the hydrogel microsphere vaccine, promoted the differentiation of BMDC into cDC1 to a greater extent and kept the higher expression level of Ki67 in these cDC1s (Supplementary

Response:
We thank the reviewer for the valuable suggestion. According to the suggestion, we performed the immunofluorescence staining assays and the results were presented in Supplementary Fig.9  Due to the limitations of antibodies and fluorescence channels, CD103 + cells were identified as potentials cDC1s, and CD8 + cells were considered as CD8 + T cells. The results of immunofluorescence staining were highly consistent with those of flow cytometry analyses in Figure 4. The IRE ablation combined with the hydrogel microsphere vaccine substantially promoted the immune cell infiltration into orthotopic pancreatic tumors (Figure 4, Supplementary Fig. 9b, c). In addition, the results of the immunofluorescence staining also showed that a large number of CD103 + and CD8 + cells were recruited into the peripheral and marginal regions of the tumor after the combination therapy (Supplementary Fig. 9b, c). Since flow cytometry is more accurate for cell qualitative and quantitative analysis, we still use the quantitative analysis results of flow cytometry in Figure 4. We hope these additional experiments will meet your criteria.

Response:
We thank the reviewer for the kind reminder. We sincerely apologize for our carelessness. We have checked the figure panels and the corresponding text and corrected the error (Page 10, line 3-4). Figure 4b, the hydrogel microsphere vaccine group is labelled as "IRE+Hydrogel microspheres vaccine", while in figures 4e, 4f, 4k and 4l it is labelled as "Group 5". Please change these labelled to something more uniform.

Comment 4. In
Also, I believe calling it Group 5 might cause confusion as all the other groups are controls.
Response: Thanks for the kind reminder. We sincerely apologize for our carelessness.
We think this issue is similar to Comment 5, where confusion arises due to the inconsistent labeling of control and treatment names. In the revised manuscript, we have uniformed the name of the hydrogel microsphere vaccine and different controls.
In the revised manuscript, the hydrogel microsphere vaccine group is labeled as "hydrogel microsphere vaccine" or "hydrogel vaccine" in relevant figures. We hope that this revision will minimize the possibility of confusion. All of these changes can be traced in the manuscript with the change track file. Response: Thanks for the kind reminder. We sincerely apologize for our carelessness.
The treatment groups presented on figures 4i, 4j, 4m and figures 5a-5e are the same as Groups 1 to 4 presented in other Figure 5. We think this issue is similar to Comment 4, where confusion arises due to the inconsistent labeling of control and treatment names.
In the revised manuscript, we have uniformed the name of the hydrogel microsphere vaccine and different controls. The hydrogel microsphere vaccine group is labeled as "hydrogel microsphere vaccine" or "hydrogel vaccine" in relevant figures and manuscripts. The names of the hydrogel counterparts and their compositions were stressed in the Methods (Page 18, line 7-10), which also listed as follow: Group1 represented the CaCO3/Cas9@Lip/@HAMA counterpart.
In the revision, these names are uniformly used in the barplot across Response: Thanks for the kind comment. To evaluate the subcutaneous tumor volumes, the maximum and minimum axis of the subcutaneous tumors were evaluated using a caliper after the mice were isoflurane-anesthetized. The tumour volume (V) was calculated using the following formular: tumour volume = 0.52 × L × W2 (L=the major tumour axis; W= the minor tumour axis). The approach was also reported in our previous study [1,2]. In Figure 6, Next, to further clarify the issue, we introduced a set of pancreatic cancer lung metastases and liver metastases models. Briefly, the lung metastases were established by tail-vein injection of tumor cells, while the liver metastases were established by semispleen injection of tumor cells [1]. Panc02 and KPC pancreatic cells were used to mimic the lung and liver metastasis of pancreatic cancers. To further test whether the immunotherapeutic effect of the combination therapy is specific for pancreatic cancers, KL (murine lung squamous cell carcinoma) and Hepa1-6 (murine HCCs) cell lines were also used to establish the lung or liver metastasis models (Reviewer 2, Comment 8).
We compared the abscopal effect across these metastasis models. Figure 6h-l, the combination therapy also effectively attenuated the formation and progression of lung or liver metastases of KPC and Panc02 cells, while no significant attenuation of the distant metastasis of KL and Hepa1-6 tumors was observed (Figure 7a-e,   Supplementary Fig. 18 c-d). It should be noted that the efficiency of the abscopal effect may differ in lung and liver metastatic models. We proposed that the combination therapy enhanced the abscopal effect by promoting an increase in the proportion of activated CD8 + T cells rather than by enhancing the infiltration of cDC1s, or M1-like macrophages in distant tumors (Reviewer 2, Comment 12) [2,3]. The various immunological milieu of the target organs may also involve the formation of abscopal effect.

Consistent with the results of subcutaneous tumors shown in
In conclusion, these results strengthened the conclusion that IRE ablation combined with hydrogel microsphere vaccine could induce a significant abscopal effect in a preclinical pancreatic cancer model and also suggested that the abscopal effect was amplified by adaptive antitumor immunity specifically against pancreatic cancers. We hope you will find these revisions satisfactory. The results showed that the hydrogel microsphere vaccine slightly increased the abundance of total macrophages, which mainly caused a slight increase in the proportion of M1-like proinflammatory macrophages. This may be related to the activation effect of the CD40 agonist on M1-like macrophages [1]. We also observed a trend of M1-like macrophage activation in the groups with CD40 agonist, although it was not statistically significant across some groups (Supplementary Fig. 17a). However, according to current reports, M1-like macrophages commonly do not induce immunosuppressive effects to attenuate the effect of immune therapy. The flow cytometry gating strategy for the phenotype of macrophages were also provided in Supplementary Fig. 17b. The detailed information on the antibodies for macrophage analysis was also provided in the Methods (Page 25, line 6-12). We hope you will find these supplementary experiments and explanations satisfactory. Agonism. [J] .Gastroenterology, 2022, 162: 590-603.e14.

Minor issues
Comment 9. Figure S7 misses legends for panels c and d.
Response: Thanks for the kind reminder. We sincerely apologize for our carelessness.
The figure legends have been added in the revised Figure S7.
Comment 10. Figure 5k can be deleted. I believe it is a repetition of 1d.
Response: Thanks for the kind reminder. Figure 5k has been deleted in the revised manuscript.

REVIEWER COMMENTS
Reviewer #1 (Remarks to the Author): I am satisfied with the responses to my original comments.
Reviewer #3 (Remarks to the Author): Thank you for your comprehensive responses to the reviewers' comments. The authors have made all the necessary changes in the manuscript to address the issues pointed out by the reviewers and to bring more depth into the study. An effort was made to improve the manuscript with all the information requested and it greatly improved the manuscript regarding the use of IRE therapy in combination with hydrogel microsphere vaccine. The improved Figures and Tables provided also further complement the manuscript and help the reader fully understand the information included in the text.
Reviewer #4 (Remarks to the Author): Please see below my feedback on the authors' response to the comments of Review 2. 1. Comment 1. Regarding the innovation of this work, the authors responded with two points. Firstly, they claim that the targeting of cDC1 for priming CD8+ with synthetic materials is innovative "However, there is still no effective strategy to locally activate and augment cDC1mediated anti-tumor immunity".  Supplementary Figure 2a shows that the release rate between pH 6.8 and 7.4 is very minor, considering the relatively high SD values of the data points, the difference seems to be insignificant, which is also not tested by the authors. This small difference can be barely considered useful as a mechanism to trigger tumor-specific payload release.
3. Comment 3. The authors claimed that Ca ions were loaded into the micelles via complexation with carboxylic groups. This raised a fundamental question of the micelle formation: with the hydrophilic carboxylic groups at the end of the hydrophobic chain end of PLGA-PEG-PLGA, this will affect micelles formation as the micelles are self-assembled via hydrophobic interactions. This question is neglected but is fundamental in the design of the materials.
4. Comment 5. The authors claimed "The specific preparation process was to first load CD40L into HOOC-PLGAPEG-PLGA-COOH micelles, and then form calcium carbonate nanoparticles around the micelles through biomineralization." However, this is another fundamental question that needs to be validated. Because it is known that hydrophilic biomolecules such as CD40L do not load in micelles which normally encapsulate payloads via hydrophobic interactions.
5. Comment 6. The pH difference before and after the IRE therapy was really small (~6.8 vs. 6.7) to be used as a trigger for tumor-specific drug release, as mentioned in comment 2. Besides, even the analysis method error can be higher than this small pH change. Therefore, the authors should also show that their method is accurate enough to distinguish the small pH change before and after IRE treatment.
6. Comment 12. Regarding the question "which immune cells in the tumor microenvironment are the major cell types for preventing the growth of distant tumors", The authors addressed this question by quantifying immune cell populations in tumors. This is not a reliable method and the data are not convincing. Instead, the authors should perform knock-out/knock-in experiments in which they deplete the target cells and then adoptively infuse the cells to confirm whether the cDC1 induced the anti-tumor immunity. What was also questionable is that cDC1, as antigen-presenting cells, should play their role in secondary lymphoid organs to activate CD8 T cells. This is another reason why the tumor immune cell counting in this case is not able to fully address this question.

1-Comment 1.
Regarding the innovation of this work, the authors responded with two points. Firstly, they claim that the targeting of cDC1 for priming CD8+ with synthetic materials is innovative "However, there is still no effective strategy to locally activate and augment cDC1-mediated anti-tumor immunity".

Response:
We are grateful to the reviewer for the constructive comments. We fully understand the concerns of the reviewer. As the reviewer noted, there have been previous reports of using biomaterials to treat pancreatic cancer, as well as original studies employing biomaterials to activate cDC1s. It also demonstrates that, as a deadly digestive system tumour, pancreatic cancer has always been the focus of clinical and basic research. In addition to immune checkpoint blockade, the activation of antigen-presenting cells has been a frontier of immunotherapy for the past decade. However, due to our limited understanding of the microenvironment of pancreatic cancer and the regulatory mechanism of antigen-presenting cells in tumour microenvironment, it is challenging to effectively activate antigen-presenting cells to fight against pancreatic tumours.
As we have explained previously, it is because of recent advances in oncology and immunology that we could attempt to efficiently activate the antitumor immunity of cDC1s in pancreatic cancer through new combination therapy and ultimately achieve a significant therapeutic effect in vivo. Considering the current clinical predicament in the treatment of advanced pancreatic cancers, we believe this new strategy gains novelty and has potential for clinical translation from the perspective of oncology. We thank the reviewer again for the valuable comment. We sincerely hope this explanation can clarify this issue.

2-Comment 2.
Regarding the controlled release of CD40L from the hydrogel microspheres, Supplementary Figure 2a shows that the release rate between pH 6.8 and 7.4 is very minor, considering the relatively high SD values of the data points, the difference seems to be insignificant, which is also not tested by the authors. This small difference can be barely considered useful as a mechanism to trigger tumor-specific payload release.

Response:
We thank the reviewer for the valuable comments. We apologize for the confusion caused by the imperfect analysis of the experimental data. We have added statistical comparisons of differences in CD40L release on the same day at different pH values to the newly provided Supplementary Fig. 2a. The raw data is also provided in the Source Data file. The results showed that pH level had minimal effect on drug release on day 0-2 and that there was no significant difference between CD40L release levels at different pH levels.
However, there was a statistically significant difference in drug release at different pH levels over the next 6 days. The release of CD40L from CaCO3 nanoparticles was designed for two primary purposes in our study.
To complete the biological function of sequential regulation of cytokines, it was necessary to ensure that the early release of CD40L was slower than that of FLT3L (relevant results can be found in Fig. 2i)

Response:
We thank the reviewer for the very insightful comment. We understand the concerns of the reviewer. The terminal micelles are one of the critical factors influencing the formation of stable micellar structures in amphiphilic block polymers such as PLGA-PEG-PLGA. The mechanism by which amphiphilic polymers form micelles is that when the polymer reaches a certain concentration in the water system, the molecules self-assemble to form an ordered arrangement of thermodynamically stable structures. Due to the presence of hydrophobic groups, the force of repulsion between water molecules and polymers is greater than the force of attraction. Under the influence of van der Waals force, the hydrophobic portion forms the micellar core, while the hydrophilic group forms the micellar's outer layer, which is stably dispersed in aqueous solution. Although the modification of the terminal carboxylation enhances the hydrophilicity of PLGA-PEG-PLGA, the combination of PLGA with a hydrophobic segment is relatively hydrophobic overall.
In a previous study [1], Prof. Ding and Yu's research group conducted a comprehensive analysis of the self-assembly behaviour of PLGA-PEG-PLGA containing terminal carboxyl groups. In Figure 10 of the report, the authors demonstrated the ability of PLGA-PEG-PLGA modified with a carboxyl group to self-assemble into micelles at various pH values. In our manuscript, the results of Figure 2 (h) also presented that stable nano-micelles could still be formed after carboxyl modification. The carboxyl groups in the micelle core can complex with supersaturated Ca 2+ in solution to provide nucleation sites, while Ca 2+ concentrates around the nucleation sites to increase the local supersaturation, while spontaneously forming aggregates.
Under the action of Ca 2+ bridging and hydrogen bonding, the aggregates rearrange and self-assemble to form calcium carbonate nanoparticles [2,3]. In the process of carboxyl groups forming nanoparticles with Ca 2+ complex, micelles only act as a template for polycarboxyl groups at the beginning to promote Ca 2+ complexation and aggregation. As the reviewer mentioned, this problem is a fundamental problem in material synthesis. However, we regret not having discovered a kind a direct method for determining whether the micelle still exists after complexation. We strongly agree with the reviewer's opinion. At the same time, we also believe that based on existing theories and evidence, it is reasonable to explain the synthesis mechanism of our material.

4-Comment 5.
The authors claimed "The specific preparation process was to first load CD40L into HOOC-PLGA-PEG-PLGA-COOH micelles, and then form calcium carbonate nanoparticles around the micelles through biomineralization." However, this is another fundamental question that needs to be validated. Because it is known that hydrophilic biomolecules such as CD40L do not load in micelles which normally encapsulate payloads via hydrophobic interactions.

Response:
We thank the reviewer for the valuable comment. We apologize for the confusion caused to the reviewers as a result of our inadequate explanation of the loading mechanism of CD40L in the paper. CD40L is loaded into CD40L@CaCO3 nanoparticles via two mechanisms. First, CD40L cytokine is a kind of protein with hydrophilic and hydrophobic regions in its polypeptide sequence. The first mechanism is hydrophobic interaction between the hydrophobic region of CD40L protein and the hydrophobic chain segment of PLGA-PEG-PLGA. Another mechanism is that CD40L can be mineralized to form nanosized mineral solids by means of the reaction between acidic amino acid residues and calcium ions in a supersaturated environment with negligible influence on peptide bioactivity. These are the two primary mechanisms by which CD40L is loaded into CaCO3 nanoparticles. The use of CaCO3 nanoparticles loaded with small molecular proteins to achieve gradual dissolution and release of encapsulated proteins in the acidic and inflammatory microenvironment of tumours has been validated and implemented in several other previous studies [1][2].
Thanks for the valuable suggestion, and we hope these revisions could clarify the issue. 5-Comment 6. The pH difference before and after the IRE therapy was really small (~6.8 vs. ~6.7) to be used as a trigger for tumor-specific drug release, as mentioned in comment 2. Besides, even the analysis method error can be higher than this small pH change. Therefore, the authors should also show that their method is accurate enough to distinguish the small pH change before and after IRE treatment.
Response: Thanks for the valuable suggestion. In the original manuscript, we have explained that the tumour acidic microenvironment is a driving force that triggers the release of CaCO3-loaded drugs. In this study, the acidic microenvironment of pancreatic cancer itself rather than the difference in pH before and after treatment drives the drug release.
In the initial round of peer review, Reviewer 2 believed that there was no evidence indicating the pH change before and after tumour ablation therapy, so he suggested that we confirm that the tumour microenvironment remained acidic after ablation in order to facilitate the sustained drug release. According to the suggestion, we measured the pH value of the tumour microenvironment and found that pH did not elevate after the ablation therapy. This result was consistent with previous reports [1][2]. Thus, the current measurement accuracy was sufficient to indicate that the tumour microenvironment after ablation therapy could still drive the release of drugs in CaCO3. We believe that this problem is mainly due to the interpretation of the manuscript text, rather than a technical problem.
We appreciate the reviewer's suggestion which enables us to further clarify this issue.
References: [1] Nikfarjam Mehrdad, Muralidharan Vijayaragavan, Christophi Christopher, Mechanisms of focal heat destruction of liver tumors. [J]. J Surg Res, 2005, 127: 208-23. immunity. What was also questionable is that cDC1, as antigen-presenting cells, should play their role in secondary lymphoid organs to activate CD8 T cells. This is another reason why the tumor immune cell counting in this case is not able to fully address this question.

Response:
We thank the reviewer for the insightful comment. In the original version of the manuscript, we have recognized the importance of secondary lymphoid organs, such as tumour-draining lymph nodes (TdLNs), in the activation and amplification of antitumor immunity. Thus, we first analyzed the activation and migration of cDC1s in TdLNs after different treatment combinations, and found that the migration of cDC1s into TdLNs was significantly enhanced following a combination of IRE and hydrogel vaccine. In addition, the enrichment of cDC1s in TdLNs was closely related to the activation of CD8 + T cells after the combination therapy.
In addition to preventing local tumour progression, we observed that the combination therapy also induced an abscopal effect. According to the comments of reviewer 2, we evaluated the potential immunological mechanism that may be responsible for the abscopal effect. It has been reported that the activation of various immune cells, such as CD8 + T cells, antitumour macrophages, and NK cells, can lead to abscopal effect in tumour therapies [1][2][3][4]. Considering the local enhancement of cDC1/CD8 + T cell antitumor immunity induced by the combination therapy, we hypothesized that the activated CD8 + T cells might be responsible for the distant effect and evaluated the change in the number of CD8+ T cells in distant metastatic tumours.
Here, according to the suggestions of Reviewer 4 and the Editors, we performed a series of rescue experiments to evaluate the effect of depletion of CD8 or blockade of MHC-I on the antitumour abscopal effect (Fig. 8a). Single dose intraperitoneal administration of antibody drugs could achieve CD8 + T cell depletion in circulation, and MHC-I blockade on cDC1s infiltrating in TdLNs (Fig. 8b). The results showed that after treatment with IRE and Hydrogel Vaccine for subcutaneous tumor, CD8 + T cell depletion or MHC-I blockade significantly promoted the growth of distant subcutaneous metastases (Fig. 8c). The CD8 + T cell depletion or MHC-I blockade also promoted lung (Fig. 8d, e) and liver (Fig. 8f, g) metastases of Panc02 cells after combination therapy for the orthotopic Panc02 tumors. The CD11c-DTR transgenic mice are the classical method for specific depletion of CD11c + cDC1s [5], and there is currently no effective antibody-based method for in vivo cDC1 depletion. Referring to previous studies, we prohibited the antigen-presenting function of cDC1s and its activation of CD8 + T cells using MHC-I blocking antibodies [6].
The results showed that the blockade of MHC-I also significantly promoted the growth of distant metastases.
Due to the complexity of the immune system, it is challenging to exclude the possibility that other immune cells contribute to the abscopal effects induced by the combination therapy. However, combined with all the data we have presented, the results at least demonstrated the important role of cDC1s and CD8 + T cells in inducing the abscopal effect in our study. Relevant revision can be found in Results section (page 14, line   25-29; page 14, line 1-6).
We thank the reviewer again for the valuable comment. We sincerely hope this explanation could clarify this issue.