Dual PI3K/mTOR Inhibitor NVP-BEZ235 Enhances Radiosensitivity of Head and Neck Squamous Cell Carcinoma (HNSCC) Cell Lines Due to Suppressed Double-Strand Break (DSB) Repair by Non-Homologous End Joining

Round 1

Reviewer 1 Report

This manuscript investigated the advantage of a dual PI3K/mTor inhibitor in HNSCC tumour by experimenting NPV-BEZ235 in in vitro cell lines both negative and HPV positive. The main goal was to study whether the drug could be an efficient radiosensitizer. The data indicated that targeting PI3K/AKT/mTor by BEZ235 and concurrent radiotherapy is effective in treating HNSCC cells in vitro regardless of HPV and AKT status.

NPV-BEZ235 –mediated suppression of DSB repair recovered radiosensitivity in all cell lines studied particularly in HPV positive.

The effect of inhibition of phospho-ser473-AKT has been already studied in HNSCC and the impact on radiosensitivity (Int. J. Cancer: 136, 2775–2785 (2015)) of high Ser473AKT phosphorylation. This paper has already demonstrated the role of AKT phosphorylation and provides experimental evidence that irradiation augments AKT (Ser473) phosphorylation and that pharmacological interference increases persistent DNA damage and reduces tumour cell proliferation. 

 

 The novelty of this manuscript is in the discovery that BEZ235 suppresses NHEJ due to an impaired phosphorylation of DNA-PKcs which can be activated by AKT.

Thereby, the main lack of this work lies in the fact that the authors used HNSCC-derived cell lines exclusively. Since the goal of authors was to test for the first time the dual PI3K/mTor inhibitor, NPV-BEZ235, in HNSCC system it would have been appreciated the establishment of ex vivo tissue cultures for HNSCC. This would allow demonstrating in a system closer to patient context the radiation induced DSB mechanism and whether NPV-BEZ235 suppresses repair mechanism in tumour patient-derived cells depending or not on AKT phosphorylation. Thereby, this body of data will greatly increase novelty and the impact of the manuscript for a publication in Cancers.

Author Response

Reviewer 1

As mentioned by the reviewer it was already demonstrated that irradiation augments AKT (Ser473) phosphorylation and that pharmacological interference increases persistent DNA damage and reduces tumour cell proliferation (f.i. Freudelsberger et al. Int. J. Cancer: 136, 2775–2785). This is also mentioned in our paper. 

It was however stated that ‘The novelty of this manuscript is in the discovery that BEZ235 suppresses NHEJ due to an impaired phosphorylation of DNA-PKcs which can be activated by AKT’.

Due to the reviewer the main lack of this work lies in the fact that the authors used HNSCC-derived cell lines exclusively.

It is suggested by the reviewer to use the ex vivo tissue cultures for HNSCC to confirm our data in tumour patient-derived samples.

We completely agree with the reviewer and in fact the respective experiments are on the way. However, it should be noted that these experiments including the establishment of the technique as well as the collection of all samples and the completion of all experiments will take at least 12 months. Therefore, this part was considered as a new project.

We feel that the present paper already includes a substantial amount of new and interesting data allowing its separate publication. In addition, we now added more and new data concerning the impact of BEZ235 on cellular radiosensitivity and DSB repair. See Figures 3C, 4B, 5C and supplementary data S5.

Reviewer 2 Report

The radio-resistance is an insuperable impediment in cancer radiotherapy. Many investigators found new and effective drugs to enhance radiosensitivity of tumor cell. Previous studies verified that NVP-BEZ235 is a potent dual function, which is PI3K/mTOR inhibitor and radiosensitizer, candidate in chemoradiotherapy strategies. The authors applied NVP-BEZ235 combined with radiotherapy to HNSCC in this study and furtherly explored the radio-resistant mechanism by clarifying the role of NVP-BEZ235 in NHEJ and HR.

It is laudable that the authors investigated proactively the radio-resistance mechanism in HNSCC cells and designed good experimental control group. However, there are many things that make me doubt in this manuscript. First, the incubation time with BEZ235 (1 h and 2 h), radiation dose (2, 6, and 10 Gy), and time points of the harvested protein extracts were very confusing. Second, the authors stated that Akt is associated with radioresistance in part of Abstract and Introduction. But the authors emphasized from Figure S3B and S4B that the radiosensitivity after irradiation alone or in combination with BEZ235 was also not associated with the Akt status. Why do you want to conduct this analysis about activity of Akt during BEZ235 treatment? This is very contradictory. Third, this study lacked animal experiments to prove the author’s argument. In addition, there are still some questions to be resolved and corrected.

 

Results

I don’t understand the meaning of the sentence “Activity of Akt depends also on the total protein concentration [29,30]”. Please explain it. UD-SCC-1 (HPV neg.) and SCC152 (HPV pos.) both expressed more pAkt than other cell lines in Figure 1A. Why do not adopted UD-SCC-1 and SCC152 to assess the ability of BEZ235 to decrease the level of pAkt? (Figure 1C) In Figure 1. Why do you change the incubation time with BEZ235 (1h to 2h)? Please provide a reasonable explanation. In addition, UM-SCC-11b cell treated with 50 nM BEZ235 alone for 1 hour (0.23) showed lower expression level of pAkt than it treated for 2 hours (0.35). I think that incubation with BEZ235 for 1 hour has better effect to inhibit pAkt compared with 2 hours. In Figure 1C and 1D. UT-SCC-33 cell and UD-SCC-2 cell both showed lower pAkt expression in IR alone group than control group. This is confusing. In Figure 2A and 2B. There is no statistical significance. In the figure legend of Figure 2. What is “from 5 h 48 h”? Please correct it. It is proper to move Figure 3A and 3B into supplementary. In Figure 3C. Why only use four kinds of cells in HPV neg. and HPV pos.? Where is the data of UD-SCC-1 and UPCI: SCC152 ? In Figure 4A and 4B. H1299 is cell line of lung cancer, not HNSCC. The two repair plasmids must be integrated into HNSCC to perform subsequent experiments. Please explain why to use H1299. In Figure 4C and 4E. The western blotting data of UM-SCC-104 cells is deficient. Please provide it. The sentence reads “This strong and significant reduction of cellular survival was confirmed in all HPV pos. and HPV neg. cell lines used in this experiment, indicating an independence of the radiosensitizing effect of BEZ235 from HPV status (Figure 5C)”. However, the data of UD-SCC-1 and UPCI: SCC152 is deficient. In Figure 5E and 5F. What is the definition of DEF? Please describe it in Materials and Methods. In addition, where is the data of HPV pos. cell lines? This study lacked animal experiments.

 

Discussion

The sentence reads “When combined with radiation, an overlay of the BEZ235-induced G1-arrest and the radiation-induced G2-arrest was found”. But I think that the data of G2 from Figure 2D has no significant difference in the group of BEZ235 combined with radiation compared with other groups. Please provide more data to prove. The sentence reads “This group also observed that BEZ235 inhibits HR, probably due to the higher concentrations of 100 nM applied.” Where is the data? 

Author Response

Reviewer 2

It is mentioned by the reviewer that it is laudable by the authors to investigated proactively the radio-resistance mechanism in HNSCC cells and designed good experimental control group. However, there are many things that make him doubt in this manuscript.

First, the incubation time with BEZ235 (1 h and 2 h), radiation dose (2, 6, and 10 Gy), and time points of the harvested protein extracts were very confusing.

Unfortunately there was a typing error in Fig. 1D stating 2 Gy. We corrected that to 6 Gy.
The high doses of 6 Gy or 10 Gy were chosen for Western Blot analysis, since higher doses will create a stronger effect. This will demonstrate sufficiently, that BEZ235 is able to inhibit irradiation-induced phosphorylation of proteins.

Second, the authors stated that Akt is associated with radioresistance in part of Abstract and Introduction. But the authors emphasized from Figure S3B and S4B that the radiosensitivity after irradiation alone or in combination with BEZ235 was also not associated with the Akt status. Why do you want to conduct this analysis about activity of Akt during BEZ235 treatment? This is very contradictory.

There is abundant data in the literature suggesting that radioresistance is associated with pAkt especially when using huge cohorts. Such a significant association but only a trend was seen as now mentioned in the text on page 6 1^st line. This is probably due to the low numbers of cell lines used in our analysis. Therefore, in the last sentence we omitted a statement to the Akt status. 

Third, this study lacked animal experiments to prove the author’s argument. In addition, there are still some questions to be resolved and corrected.

We agree with the reviewer that we have to confirm our data with in-vivo experiments. This is now planed in our next study using the ex-vivo assay. The respective experiments will take us at least 12 months including the collection of patient samples etc. and is therefore considered as a separate project.

 

Results

I don’t understand the meaning of the sentence “Activity of Akt depends also on the total protein concentration [29,30]”. Please explain it.

This statement is not substantial in the written context and is now deleted.

UD-SCC-1 (HPV neg.) and SCC152 (HPV pos.) both expressed more pAkt than other cell lines in Figure 1A. Why do not adopted UD-SCC-1 and SCC152 to assess the ability of BEZ235 to decrease the level of pAkt? (Figure 1C) In Figure 1.

Data in Figure 1C were used to demonstrate that the inhibitory effect is very long lasting and still present after incubation of 24h. That this inhibitory effect is present in all cell lines including UD-SCC-1 and SCC-152, as requested by the reviewer, is shown in Fig. 1D.

Why do you change the incubation time with BEZ235 (1h to 2h)? Please provide a reasonable explanation.
In addition, UM-SCC-11b cell treated with 50 nM BEZ235 alone for 1 hour (0.23) showed lower expression level of pAkt than if treated for 2 hours (0.35). I think that incubation with BEZ235 for 1 hour has better effect to inhibit pAkt compared with 2 hours

It is now clearly mentioned on page 4, line 122, that BEZ235 was given 2 h before irradiation.

We agree with the reviewer that there are slight differences for the cell line UM-SCC-11b between Fig. 1C and 1D (0.23 vs. 0.35). However, we feel that this is mostly due to the biological variation from one experiment to the other. 

In Figure 1C and 1D. UT-SCC-33 cell and UD-SCC-2 cell both showed lower pAkt expression in IR alone group than control group. This is confusing.

In fact, in some experiments we saw no increase in pAkt after IR, as also mentioned in the text. This is also seen in other reports (Toulany et al. Radiother Oncol 2011, 101:140-6; Fokas et al. Radiat Oncol 2012, 7:48). Sometimes, we even observed a lower level as mentioned by the referee. The latter is ascribed to the biological variation often seen in Western Blots. However, the overall message that BEZ235 is able to strongly decline pAkt is clearly seen for all cell lines. 

In Figure 2A and 2B. There is no statistical significance.

It is now shown in this figure, that there is a significant difference between controls and irradiated samples.

In the figure legend of Figure 2: What is “from 5 h 48 h”? Please correct it.

Line141: The sentence was corrected and is now “from 5h to 48h”

It is proper to move Figure 3A and 3B into supplementary.

Unfortunately, we do not agree with the referee. Due to our opinion it is important to show these figures in order to illustrate the high quality of this detection.  

In Figure 3C. Why only use four kinds of cells in HPV neg. and HPV pos.? Where is the data of UD-SCC-1 and UPCI: SCC152 ?

These data were previously skipped from the analysis due to the insecure analysis of the foci data. However, for UD-SCC-1 we are now able to include these, whereas it is still not possible for UPCI:SCC152, because this cell line does not allow a proper analysis of γH2AX foci. This is now mentioned on page 5, line 149-151.

In Figure 4A and 4B. H1299 is cell line of lung cancer, not HNSCC. The two repair plasmids must be integrated into HNSCC to perform subsequent experiments. Please explain why to use H1299.

For this analysis it is important to use a fully repair competent cell line with optimally integrated repair plasmid. This is guaranteed by the cell line H1299. However, we now also show data for the HNSCC cell line UD-SCC-2 (Figure 4B). It should be noted that we also repeated the experiments for H1299. Therefore the data shown now are slightly different from previous data.

We also included new data concerning the impact of BEZ235 on the Rad51 formation after X-irradiation. Now shown in supplementary Figure S5. These data indicates that inhibition of NHEJ by BEZ235 may lead to a small shift of DSB repair to HR. Now mentioned in the manuscript on page 6, line 197-199. 

In Figure 4C and 4E. The western blotting data of UM-SCC-104 cells is deficient. Please provide it.

These data are now also shown in Figure 4C and E.

The sentence reads “This strong and significant reduction of cellular survival was confirmed in all HPV pos. and HPV neg. cell lines used in this experiment, indicating an independence of the radiosensitizing effect of BEZ235 from HPV status (Figure 5C)”. However, the data of UD-SCC-1 and UPCI: SCC152 is deficient.

The data for UD-SCC-1 and UPCI:SCC152 were previously omitted from the analysis because of  unsecure analysis of γH2AX foci as already described above. These data are now included in the analysis.

In Figure 5E and 5F. What is the definition of DEF? Please describe it in Materials and Methods. In addition, where is the data of HPV pos. cell lines?

It was mentioned on page 7, line 248, that the dose enhancement factor (DEF) was calculated at 10% survival.

As mentioned in the manuscript and also above (14), the radiosensitizing effect of BEZ235 was found to be independent of the HPV status and shown to result from a depressed NHEJ. To verify these data, radiosensitization has to be studied in dependence of cell cycle in a HPV neg. cell line known to be proficient in both NHEJ and HR. We don’t see an argument why this needs to be confirmed by a HPV pos. cell line.  

This study lacked animal experiments.

As already outlined above, we agree with the reviewer that we have to confirm our data with in-vivo data.

These experiments will be performed using the ex-vivo assay with samples obtained from patients with HPV neg or pos. HNSCC. This huge project including the pathology etc. will take us at least 12 months and is considered to be a separate project.

 

Discussion

The sentence reads “When combined with radiation, an overlay of the BEZ235-induced G1-arrest and the radiation-induced G2-arrest was found”. But I think that the data of G2 from Figure 2D has no significant difference in the group of BEZ235 combined with radiation compared with other groups. Please provide more data to prove.

As already outlined above (8), it is now shown in this figure, that there is a significant difference between controls and irradiated samples.

The sentence reads “This group also observed that BEZ235 inhibits HR, probably due to the higher concentrations of 100 nM applied.” Where is the data? 

This is now changed to “This group also observed that BEZ235 inhibits HR, which is however not seen here.”

In order to strengthen our data, we now also present data about the impact of BEZ235 on Rad51 foci formation. It is shown in supplementary figure S5 that in our hand BEZ235 slightly enhances this formation rather than a depressed formation.

Round 2

Reviewer 1 Report

Head and neck squamous cell carcinomas (HNSCCs) are an aggressive, genetically complex group of cancers whose heterogeneity makes the treatment difficult. Surgery and radiotherapy are the primary treatment although radioresistance and relapse are very frequent.

In this paper the dual PI3K/mTOR inhibitor, BEZ235, was tested for its ability to increase the radiosensitization of a number of HPV positive and HPV negative HNSCC cell lines.

On the basis of data obtained in vitro using HNSCC derived cell lines authors suggest that BEZ235 affects AKT-DNA-PKcs axis impairing DNA repair by NHEJ.

In the previous review the reviewer suggested to use ex vivo culture of HNSCC in order to have results closer to patient contest.

Authors replied that the data of the present work are interesting and deserve to be published while all the other approaches are on the way and will be considered in a new project.

In this revised form new data were added without changing the impact of this manuscript.

If the ex vivo results will take 12 months of work, the in vivo experiments aimed at demonstrating that BEZ235 works through same pathway contrasting tumor growth and  suppressing NHEJ via AKT-DNA-PKCs In vivo tumour can be done by xenografting  in nude mice H1299 or UD-SCC-2 cells with the repair plasmid integrated. Xenograft experiments and treatments often can take few months of work. The evidences that BEZ235 characterized in HNSCC cell lines in vitro acts by the same mechanism in in vivo tumour will increase the impact and novelty of this manuscript.

The manuscript as a whole is potentially interesting but require intense revision of data presentation in figures and text; irradiation condition are an important phase of this study and the condition should be homogenous unless changes are needed and in this case supported by explanation otherwise the reader is confused; WB analysis support many important results and for this reasons should be presented and evaluated correctly.

 

Questions

Fig1A. Authors should specify how the values of AKT1 and pAKT in WB were obtained: are they arbitrary units or are referred to some sample taken as 1? Authors are invited to specify. As a matter of fact, in Fig 1 B, on the left the relative expression (arbitrary units?) and right Ratio pAKT/AKT1 depends on the values of Fig A. For example in Fig 1 C and D is clear that untreated sample is taken as 1.

 

Fig 1 D and E. The irradiation is here reported as 2 Gy whereas in the text, line 121, and legend, is reported as 6 Gy. The difference between 2 and 6 is very high, see Fig5 B and F. What is the right IR applied to see the effects reported in this experiment?

 

 

 

Upon irradiation an increase in phosphor-Akt was seen…….

A reliable increase in AKT phosphorylation after IR, as showed in Fig 1D and E , is not present in the majority of cell lines but in some cell lines: in two HPV neg and two HPV pos. As matter of fact, 1.17, 1.24 and 1.3 are not reliable increases but rather demonstrated the tendency of phospho-AKT to increase under irradiation unless these values are supported by statistical analysis.  Fig 1 C, D and E.  The levels of AKT1 and phospho AKT were calculated on different gels (hopefully same amounts of same samples) and each with its GAPDH normalization; AKT1 level appears to decrease following Irradiation, BEZ235 treatment alone or in combination with IR in UDSCC2 for example. Authors are invited to explain (in methods or legend) how they calculate the variations of levels in phospho-AKT. As a matter of fact, in WB analysis the level of kinase expression and its phosphorylation status are usually detected in the same membrane exactly  as authors did for experiment of Fig 4C and E. 

BEZ235 Leads to a Moderate G1 Arrest Without Affecting Apoptosis   

There is any reason to have this title of paragraph where is reported the effect of BEZ235 on apoptosis? See lines 155-156.

 

Fig.4 C and E. Same problem as in Fig 1A, that is authors should specify how the expression levels of DNA-PKcs , phospho-DNA-PKcs, ATM and phospho-ATM were calculated.

In these experiments authors used 10Gy of irradiation while all previous results were obtained using 2Gy. The maximum dose authors used was 4Gy as in Fig 5. Authors are invited to explain this different radiation treatment performed only in this experiment.

 

Author Response

Reviewer 1

It was asked by the reviewer to specify how the expression values were obtained.

It is now stated in Mat and Meth on page 12, line 403-404 that “Generally, specific expression was normalized to the mean value of the total blot, or, when specified, to the untreated control”.

The reviewer claims that for Fig.1D and E in contrast to the Fig. in the text, line 121, and legend, the radiation doses is reported to be 6 Gy.

We apologize, this was a typing error and is now corrected to 6 Gy.

The reviewer states that the data obtained for phosphor-Akt after exposure to 6 Gy is not adequately described by saying “Upon irradiation an increase in pAkt was seen for the majority of cell lines with (Fig. 1D and E).”

We partly agree with the reviewer. This sentence is now changed to: Upon irradiation a very heterogeneous pattern was seen for pAkt, with an increase in only few cell lines.
It is now stated in Mat and Meth how the expression was quantified (see above).

The reviewers criticizes the title ‘BEZ235 Leads to a Moderate G1 Arrest Without Affecting Apoptosis’.

We have to agree with the reviewer that this title needs to be corrected. This is now changed to ‘BEZ235 Leads to a Moderate G1 Arrest Without Affecting Radiation-Induced Apoptosis’.

The reviewer asked for a detailed description of the expression analysis

This is now given in Mat and Meth (see above).

The reviewer wants to know why phosphorylation of ATM and DNA PKcs was measured after10Gy and not after 6Gy as used for the other experiments.

It is now mentioned on page 7, line 220-221 that a dose of 10 Gy was used in order to achieve a significant increase especially for DNA PKcs at T2609.

 

Concerning the statement about animal experiments, please refer to response to Reviewer 2

Reviewer 2 Report

Although the authors have answered most of the questions positively, some issues remain unresolved.

 

The irradiation dose in Figure 1D and Figure 1E still does not accord with the content.

 

But I still want to know the reason about the change of incubation time. Maybe the incubation time of 1h had better effect than 2h in combination treatment in Figure 1D.

 

The animal experiments don't take much time (at most 3 months). At least 12 months is too exaggerated. Therefore, this is not a reason to refuse to perform animal experiments.

 

In the Discussion section: The sentence reads “This group also observed that BEZ235 inhibits HR, which is however not seen here’’. This description is too arbitrary and must deleted because the data cannot be presented.  

Author Response

Reviewer 2

 

It is mentioned by the reviewer that the irradiation dose in Figure 1D and Figure 1E still does not accord with the content.

It is now corrected in the figure that the dose used was always 6 Gy.

 

The reviewer wants to know the reason about the change of incubation time.

Sorry, that we forgot to answer in the first reply.

It is demonstrated in Fig. 1D and E that treatment with 50 nM BEZ235 for 2h will reduce  pAkt in an extent similar to that seen after an incubation for 1h. This is now mentioned in the text on page 4, line 125 by “BEZ235 treatment, which was applied 2 h before irradiation, could not only reduce basal phosphorylation activity similar to that shown in Figure 1C, but also to abolish radiation-induced activation of Akt.”

 

The reviewer suggests that the animal experiments don't take much time (at most 3 months). At least 12 months is too exaggerated. Therefore, this is not a reason to refuse to perform animal experiments.

Unfortunately, we cannot agree with the referee. We already submitted a respective ethical proposal to conduct the ex-vivo assay with tumor samples. Following high ethical standards, already this process will take us more than three months. The same holds true for animal experiments.

Most of all, we feel that adding further ex vivo or in vivo studies will not increase the impact of our study. With our study we present already strong data. Examining ten different cell lines, plus the various different techniques used including plasmid repair assay, gH2AX- and Rad51 foci technique as well as synchronized cells is already a rather substantial report.

The main focus of our work was to examine the effect of BEZ235 on HNSCC cells, since this is a tumor entity with a very heterogenous molecular profile, due to the tumor pathogenesis. We found that BEZ235 is a potent radiosensitizer of HNSCC cells, independently from the molecular status of the cells.

 

The reviewer criticizes the sentence in the discussion section: The sentence reads “This group also observed that BEZ235 inhibits HR, which is however not seen here’’. This description is too arbitrary and must deleted, because the data cannot be presented.

Unfortunately, we do not understand this statement. It was shown by Mukherjee et al. (Neoplasia 14:34-43, 2012) for MCF7-DRGFP that BEZ 235 down-regulates HR and for 1BR3 skin fibroblast cells that BEZ235 blocks Rad51 foci formation, which is in clear contrast to the results seen by us for HNSCC cells. Therefore this sentence was changed to “However, this group also observed that BEZ235 inhibits HR in 1BR3 fibroblasts by studying Rad51 foci formation or in MCF7 cells by using a respective DSB repair construct. These results are in clear contrast to the data shown here for HNSCC cells as outlined above indicating that the effect of BEZ235 might depend on the cell line used.”  

Round 3

Reviewer 2 Report

The authors have responded to the comments of reviewer although the animal experiemental has not performed in this revised text.