Responses to the Reviewers

Reviewer 1

1. Figure 1&2 show the results of HL-60, KG-1 and primary AML cells. But other results just show HL-60 alone. I would like to see all the data regarding to KG-1 and primary AML cells to exclude some specific effect only in HL-60.

In agreement with the reviewer, we performed also other experiments (in particular ATP release, DC- maturation, T-cell proliferation and Treg-induction) with KG-1 cell line and primary AML cells. We added the obtained results and the figures in the text.

2. For figure 3 &4, Facs should be a more precise method to quantitate Fitc positive cells compared to IF which is a semi-quantitative method if there is no technical problem. Do the authors have some FACS data?

We agree with the reviewer that IF is semi-quantitative method, which certainly may represent a limitation. Unfortunately, for HMGB-1 release detection flow cytometry is not capable to detect the HMGB1 released from the nucleus in cytoplasm or extracellular milieu. With these limitations, we used the immunofluorescence method in line with other reports, which tested HMGB1 release in AML cells (Fucikova et al, Int J Cancer, 2014).

3. In vivo study should be a better experiment for immunogenic cell death research to draw a solid conclusion. Has there been any in vivo study for this research even just based on leukemia cell line?

We agree with the reviewer that in vivo studies may provide important evidence. The capacity of DNR and Ara-C to induce ICD in a mouse model of AML has already been investigated in our previous work (Lecciso, Ocadlikova et al 2017; Frontiers in Immunology), focused on ATP release and induction of Tregs. As for other drugs, such as Eto and Flu, we are planning to extend the experiments also to in vivo mouse model. Thanks to the reviewer’s suggestion, we added this point as future plan in the Discussion/Conclusion section.

4. Any possible mechanism for Eto-induced ICD? Shall we have a potential molecular target specific for ICD related therapy in the future?

We thank the reviewer for this very interesting and relevant question. Assuming that immunogenic cell death (ICD) is initiated by the translocation of CRT from the endoplasmic reticulum (ER) lumen to the plasma membrane thus facilitating the transfer of tumor-associated antigens to DCs, we hypothesized that a first molecular target of ICD process initiation after Etoposide treatment could be associated with ER stress induction.

In ICD context, it was well demonstrated that phosphorylation of the Eukaryotic Initiation Factor 2 (eIF2α ) is essential for the ER stress response and is correlated with CRT/ERp57 complex exposure leading to DCs activation in various tumor models (Bezu et al 2018, Cell Death Differ; Panaretakis et al 2009, EMBO J; Bénéteau et al 2012, Proc Natl Acad Sci U S A ; Giglio et al 2019, Genes Immun and others).

Therefore, we think that the well characterized Etoposide activity which is mediated by its interaction with topoisomerase II leading to DNA strand breaks that are lethal to the cell (Montecucco et al 2007, Cancer Letters; Montecucco et al 2015; EXCLI J; Maanen et al 1988, J Natl Cancer Inst and many others), is accompanied also with other type of cell death which can be immunogenic by ER stress induction through the eIF2α phosphorylation. Indeed, some authors demonstrated that Etoposide may possess an uncharacterized cytotoxic function via other channels, such as the induction of reactive oxygen species, the stimulation of Akt/mTOR signaling and the activation of apoptosis through caspase- 3 (Demel et al 2015, Am J Cancer Res; Oh et al 2007, Mol Cancer Ther; Zhang et al 2013, Expert Opin Drug Deliv).

Recently, Chen Wang et al (Wang et al 2016, Sci Rep) have demonstrated that Etoposide triggers apoptosis through the ER stress-mediated activation of the IRE1/ASK1/JNK pathway in activated human hepatic stellate cells. In particular, after Etoposide exposure, protein kinase RNA-like endoplasmic reticulum kinase (PERK) and eIF2a are phosphorylated and the expression of some ER- resident chaperones is also upregulated.

These findings indicate that Etoposide has antiproliferative and proapoptotic effects and induces cell apoptosis through a mechanism involving the ER stress pathway (Wang et al 2016, Sci Rep).

To support our hypothesis, we report the work of (Beneteau et al 2014, Proc Natl Acad Sci U S A).

The authors demonstrated that Etoposide in combination with 2-deoxyglucose (2-DG; inhibitor of glycolysis), but not 2-DG alone, induces ICD in mouse lymphoma model and that this effect was at least partially mediated through Rp57/calreticulin exposure on the plasma membrane. This is a first hint/sign that Etoposide has some immunogenic properties also in other tumor model.

In light of this findings, Etoposide is cited as ICD inducer when used in combination with glycolysis inhibitor in a publication of Kepp, Zitvogel and Kroemer’s group from 2014 (Kepp et al 2014, Oncoimmunology) entitled as “Consensus guidelines for the detection of immunogenic cell death”.

In line also with the suggestion of second reviewer, we changed the title of this manuscript and included a paragraph reporting these findings about Etoposide and its immunogenicity in the Discussion Section.

Reviewer 2

1. Title: should not Eto also be mentioned? The authors focus a lot about its actions, but still it is let out of the title.

We agree with the reviewer, we have accordingly changed the title of the manuscript.

2. Abstract: The results could be a bit clearer. The first sentence contains a lot of information, and Eto is mentioned at the end so that it might escape the reader. Conclusion: please revise the «poorly investigated». It might be interpreted as if this study is meant. Maybe add «until now».

We agree with the reviewer, we have accordingly modified the abstract text.

3. Primary AML cells: from the method section it appears as if cells from several patients were used. More information is needed here: how many patients, what are the cells` morphological features and karyotypes, are the results a mean of these values?

Thank you for this observation. We used samples from 3 patients with AML and we specified their characteristics (in particular Fab classification and karyotype) in the methods section (see 4.1.). The results are represented by the mean values from these 3 patients.

4. Figure 4B: The plot is difficult to interpret as colours for Dnr and Flu are similar. Maybe the different part of the plot can be moved further apart? Alternatively, a baseline level for the drugs without cases can be indicated?

We agree with the reviewer, we moved the groups on X asses in FIGURE 4B to make them more distant.

5. Figure 6: are the colour codings correct? From the plot it appears as if Dnr and Flu induces CCR7 expression, not Dnr, Eto and Ara-C as stated in the text. The same is true for CD80, where only Ara-C seems to not upregulate the surface

We thank the reviewer for this important notice, there was a mistake in the graph (Figure 6 – HL-60 cells). We corrected accordingly.

6. Line 294: «proved» instead of «roved».

We thank the reviewer, we corrected.

7. Line 300f: The authors mention the possibility to use different drugs in personalized medicine. Could they discus more about this issue? Are there specific patients (subgroups) where immunosuppressive effects (Flu) might be an advantage compared to drugs that stimulate the immune system (Dnr and Eto)?

We agree with the reviewer, we have accordingly modified the discussion text.

Round 1

Reviewer 1 Report

The authors addressed all of my questions. 

Reviewer 2 Report

The authors have answered all my requests, and have edited the manuscript and the figures according to my suggestions. 

I therefore recommend the manuscript for publication without further changes.