Title: Anti-proliferative effect of extremely low frequency electromagnetic field on preneoplastic lesions formation in the rat liver

Author's response to reviews Title: Anti-proliferative effect of extremely low frequency electromagnetic field on preneoplastic lesions formation in the rat liver Authors: Mónica N Jiménez-García (mony.piensa@gmail.com) Jaime Arellanes-Robledo (jarellanes@gmail.com) Diana I Aparicio-Bautista (adbi2401@hotmail.com) Miguel A Rodríguez-Segura (mars@fis.cinvestav.mx) Saúl Villa-Treviño (svilla@cell.cinvestav.mx) Juan J Godina-Nava (jj@fis.cinvestav.mx) Version: 2 Date: 1 February 2010 Author's response to reviews: see over

1 Dear editors of BMC Cancer journal, We considered all suggestions of reviewers. Thank to those, the manuscript was substantially improved. All new or modified paragraphs are showed highlighted in blue throughout the manuscript. Reviewer Comments: Referee 1 Reviewer: Santi Tofani 1. Describe how animals were held during the exposure, how many of them were contemporarily exposed or sham exposed, provide also info on the timing of the entire experimental section (s). A scheme of the experimental setup should be given with all the necessary information for in theory allowing a replica of the experiment. Results of all physical parameters monitored during the animal exposure as well as during sham exposure sections should also be given. In the Methods section we included a new subsection entitle: Animal exposure which explains part of these questions: Animal exposure During the exposure, CTF group was divided in two subgroups of three rats each one and placed in a Plexiglas cage (16 x 16 x 25 cm) inside of solenoid. Each subgroup was exposed to 4.5 mt - 120 Hz ELF-EMF for 50 min daily as shown Figure 1; the first group was exposed from 10:00 10:50 h and the second was from 12:00 12:50 h, alternating them on this schedule, during 32 consecutive days (from 7 days before carcinogenic treatment until 25 days after). After exposure all animals were returned to their home cages. Moreover, we also included a new figure (Figure 2) for a better explanation and understanding of entire experimental section. Additionally, in the legend of the Figure 1 we improved the explanation about our experimental setup: Figure 1. Schematic representation of the carcinogenic treatment and ELF-EMF exposure. The sham-exposure or exposure to ELF-EMF was applied before and during carcinogenic treatment, as indicated in schemes of CT and CTF groups, respectively. NC, normal control; CT, carcinogenic treatment and sham-exposure; CTF, carcinogenic treatment plus ELF-EMF exposure. S, sacrifice; DEN, N-diethylnitrosamine; 2AAF, 2- acetylaminofluorene; PH, partial hepatectomy; n = 6 for each group. In the subtitle Experimental design of the Methods section, additional information was included, like light/dark cycle, temperature and humidity in which the animals were held throughout the experiment: Male Fischer-344 rats weighing 160 to 200 g, obtained from the Production Unit of Experimental Laboratory Animals (UPEAL-CINVESTAV, México DF, Mexico), were fed ad libitum and housed in a controlled environment (12 h light/12 h dark cycle, room temperature of 19-22 ºC with relative humidity at 40-70 %).

2 To clarify the moment of the exposure during the days when the carcinogenic treatments were given, in the subtitle Experimental design of the Methods section we included the next paragraph: CTF group. For this group, the administrations of DEN, 2AAF and PH were carried out before ELF-EMF exposure schedule. 2. How and where temperature was recorded inside the coil (in respect to the animal position) and corresponding results obtained during the animal exposure as well as during sham exposure section (s) should be given. For answering this question, in the subtitle System of ELF-EMF exposure of Methods section was included the next paragraph and the figure 2: The temperature sensor location and the shape of signal wave are shown in Figure 2. In the subtitle General observations of the Results section, we also included the next explanation: We also monitored the temperature of sham-exposure (CT) or exposure (CTF) groups during the 32 days and their values were recorded at specific times shown in Table 1. Table 1. Average values of the temperature recorded during the exposure time Treatment group CT CTF Time/min T/ C T/ C 0 23.7±0.4 23.7±0.4 25 23.9±0.6 25.8±0.6 35 24.2±0.6 26.7±0.7 45 24.0±0.5 27.1±0.6 50 24.1±0.6 27.0±0.7 All values are expressed as the mean ± SEM 3. About the histological and immunoistological exams please report the number sections used for each animal and how sections were selected. Please also report results for every single animal in a table where also the number of liver sections is shown. Please provide also the timing of when these exams were performed (i.e. in different days?) for the different experimental groups. Please also report on the knowledge by the examiner of the exposure group. In the subtitles GGT histochemical staining and TUNEL assay and immunohistochemical analysis of the Methods section, we included the number sections used and how were selected them, moreover we included when the exams were performed: GGT histochemical staining For each animal 3 tissue sections were analyzed randomly. The experimental procedure was carried out at the same time for all comparative groups. TUNEL assay and immunohistochemical analysis 4 sequential sections per rat were analyzed; For detection of each protein, the immunostaining protocol was carried out at the same time for all comparative groups.

3 Note: The figures that include tissue sections only show representative tissues of one animal per treatment group. A table (Table 2) that shows the number of liver sections reported as the mean of foci number per cm 2 and the percent of area GGT-positive for every single animal was included. Table 2. Effect of ELF-EMF exposure on foci number per cm 2 and the percent of area GGT-positive Group # animal Foci number/cm 2 % of GGT-positive area CT 1 38.71 ± 3.3 6.93 ± 0.4 2 16.31 ± 2.6 2.27 ± 0.5 3 39.22 ± 11.4 5.63 ± 1.2 4 31.42 ± 1.5 12.76 ± 0.8 5 40.51 ± 9.0 8.77 ± 1.7 6 28.66 ± 6.2 4.98 ± 1.1 CTF 1 0.65 ± 0.6 0.07 ± 0.1 2 8.13 ± 3.5 1.18 ± 0.5 3 27.45 ± 5.7 4.59 ± 0.9 4 14.85 ± 2.6 2.38 ± 0.6 5 24.40 ± 2.2 3.80 ± 0.2 6 14.50 ± 2.5 4.52 ± 0.7 Values represent the mean ± SEM of three liver sections evaluated per animal. In the subtitle of Experimental design of the Methods section we also report on the knowledge by the examiner of the exposure group: Biochemical and molecular evaluations were performed under blind conditions by expert people not involved in the animal exposure. Referee 3 Reviewer: Arthur A Pilla 1. The authors indicate temperature in the coil was monitored, but provide no values for temperature for active and inactive coil states. These should be included. In the subtitle General observations of the Results section, we included the next explanation for this question and we also included the Table 1: We also monitored the temperature of sham-exposure (CT) or exposure (CTF) groups during the 32 days and their values were recorded at specific times shown in Table 1. Table 1. Average values of the temperature recorded during the exposure time Treatment group CT CTF Time/min T/ C T/ C 0 23.7±0.4 23.7±0.4 25 23.9±0.6 25.8±0.6 35 24.2±0.6 26.7±0.7 45 24.0±0.5 27.1±0.6

4 50 24.1±0.6 27.0±0.7 All values are expressed as the mean ± SEM 2. The authors should look at: Strauch et al. Evidence-based use of pulsed electromagnetic field therapy in clinical plastic surgery. Aesthet Surg J. 2009;29:135-43 for further insight into possible mechanisms of PEMF therapeutic effects. In the Discussion section, we included a paragraph that gives an additional possible mechanism of electromagnetic fields effects: Furthermore, evaluations in experimental models have been established that the electromagnetic fields are able to modulate the intracellular calcium (Ca 2+ ) when cellular homeostasis is disrupted [32]. Calcium is a highly versatile intracellular signal that can regulate many different cellular functions, whether normal or pathological; thus, the consequences of Ca 2+ signaling depend of steady state between Ca 2+ influx, efflux, and storage [33]. Cancer development takes place through rapid proliferation and the continuous increase of altered cells that modify the cellular environment [19], including the flow of ionic charges across the cell membrane, such as Ca 2+ flow. Given that several blockers of Ca 2+ entry inhibit tumor growth [34], we cannot discount that ELF-EMF could be regulating Ca 2+ flow in the cells. Therefore, Referee 2 Reviewer: Ronald J Midura Commentary 1 The manuscript should provide a detailed diagram of db/dt versus the time domain for the chosen electro-magnetic field (EMF) treatment. For covering this suggestion, in the manuscript was included the Figure 2 which is mentioned in the subtitle System of ELF-EMF exposure of Methods section. The Figure 2 shows the waveform measured inside the coil. 1. Figure 1 shows that EMF treatments began a week before the first drug treatment to induce the modified resistant hepatocyte model (MRHM) of pre-neoplastic lesions. A rationale for starting EMF treatments PRIOR to induction of MRHM is not described. Furthermore, from a relevance view point, it is hard to justify a biophysical treatment prior to diagnosis of a disease state from ethical and clinical standpoints. What would have happened if EMF treatments started after DEN treatment, or after partial hepatectomy? Without an adequate rationale for this research design, the biological and clinical significance of this work is jeopardized. For example, would the EMF treatments prior and during subsequent drug treatments have had an effect on the pharmacokinetics of these drugs and thus alter their bioavailability in liver tissue? If so, then one would conclude that EMF reduced pre-neoplastic lesions in the liver because it reduced the negative influences of these drugs via altering their accessibility to the liver. With regard to this suggestion, the rationale by which the EMF treatment was given prior to MRHM is that although HCC takes several years to reach invasiveness, the inability to diagnose HCC patients at an early development stage causes it to be a high mortality malignancy (Jia HL et al, Clin Cancer Res 2007;13:1133-1139). Thus, prevention strategies administered during initial stages represent a promising strategy for the treatment of HCC. In this way, by identifying strategies that modulate molecular

5 targets in preclinical models, these could be applied in populations at high risk of developing the disease. In the subtitle Animal exposure of the Methods section and in the Conclusion section, we included information to cover this suggestion: Animal exposure We decided to apply the ELF-EMF from 7 days before starting the carcinogenic treatment, because our initial approach was to assess its effect on the inhibition of preneoplastic lesions development in rat liver, as a strategy to prevent the disease in the early stages of its development. Conclusion Finally, considering that hepatocellular carcinoma is a common form of cancer and that its incidence around the world remains high [35], this finding could be the basis for the design of strategies and clinical applications of ELF-EMF to treat this disease, aimed primarily at high-risk populations. 2. The immunohistochemistry data in Table 1 need to be broken down with regard to signal detection within the boundaries of the pre-neoplastic lesions versus signal detection outside lesion boundaries. Only with this kind of assessment can one determine whether the EMF effects are specific to the cells within the lesion as opposed to having generalized effects on all cells within the liver tissue. Although the carcinogenic treatment induces the formation of preneoplastic lesions, which can be detected by specific markers as the expression of GST-p, the expression of other proteins, such as proliferation and cell cycle proteins, are not limited to preneoplastic lesions, they also are found outside the lesions; indicating that the carcinogenic treatment alters the whole tissue at other levels. Previously, altered cells were already quantified both within and outside the preneoplastic lesions and it was reported that more than 70% of them are located within lesions after carcinogenic treatment (Arellanes-Robledo J, et al, Anticancer Res 2006;26:1271-1280). Furthermore, in the present work, the effects of electromagnetic fields on expression of these proteins were confirmed quantitatively in total tissue by Western blot analysis. That showed a clear effect on the general phenomenon. For this reason, we quantified the changes in proteins expression of the whole tissue. 3. Figure 3 TUNEL results are of good quality, but the authors need to realize that TUNEL detects free 3 -OH in genomic DNA via DNA strand breaks resulting from damaging insults. Thus, TUNEL is best interpreted as an assay of DNA damage, and can only be interpreted regarding apoptosis when additional corroboration is made. The study shows activated (cleaved) caspase 3 Western blots for NC, CT, and CTF liver samples. Close inspection of these Western blots demands that more sampling of liver tissues from additional animals needs to be obtained in order to provide a truly representative sampling. Specifically, in the CTF group two of the sample signals were noticeably lower in intensity than even the NC signals; only one CTF sample exhibited relatively high signal intensity. Does this one specimen skew the data erroneously, and if so, then the CTF would actually be shown to reduce overall activated caspase 3 levels. This would lend an interpretation that EMF treatment might have reduced overall apoptosis levels in the liver tissue. Closer inspection of the TUNEL results support this claim as the intensity of the TUNEL signal per nucleus seems greater in the CT group versus NC or CTF groups. In fact, the TUNEL signal of he CT group seems closer to

6 the DNaseI digestion positive control for the assay. The authors are suggested to reanalyze their activated caspase 3 and TUNEL results and seek a higher level of confidence regarding the quantifiable attributes of their data. Given a lack of a statistical power analysis in the Methods section, it is unclear at present whether a sufficient sample population for each outcome was definitively established. We are agree with this suggestion, TUNEL assay only can tell us if an agent breaks the DNA strands but no if it is able to induce apoptosis; in this way, in the subtitle TUNEL assay and immunohistochemical analysis of the Methods section, we change the aim by which we used the TUNEL assay, remain as follow: Liver tissue sections, 4 µm thick, were deparaffinized and hydrated gradually. DNA fragmentation was determined by a Colorimetric TUNEL System kit, Similarly, in the subtitle ELF-EMF exposure did not induce apoptosis of the Results section, as well as in Discussion section, we modified our insight. ELF-EMF exposure did not induce apoptosis To evaluate the effect of ELF-EMF exposure on the apoptosis induction of altered hepatocytes, we used two different procedures applied to the three groups. First, cells with DNA fragmented were identified in tissues by a TUNEL assay. Representative tissue sections of each treatment are depicted in Figure 4. Although a slight increase of TUNEL-positive cells is observed in the rat tissues of CT group (Figure 4C), these were not different from those of rats in the NC group or those in the CTF group (Figures 4B and 4D). According to this result, the levels of cleaved caspase 3 were not affected by either the CT or the CTF protocols as compared to the NC protocol (Figure 4E). These results indicate that the application of 4.5 mt - 120 Hz ELF-EMF does not induce either DNA fragmentation or the caspase-3 activation, indicating that it does not promote apoptosis of altered hepatocytes. Discussion These results indicate that the application of 4.5 mt - 120 Hz ELF-EMF does not induce either DNA fragmentation or the caspase-3 activation, indicating that it does not promote apoptosis of altered hepatocytes. Respect to detection of cleaved caspase-3 by Western blot analysis, in the Figure 3 (now Figure 4); we only show representative images of six independent animals per group. We consider that six animals per group are enough. Moreover, in Figure 3 (now Figure 4) you can see that two of three animals that belonging to the CTF group presented less intensity in the cleaved caspase-3 band, and not vice versa, this support our results respect to the ELF-EMF dose not induce caspase-3 activation. Initially, we designed our experiments including only the TUNEL assay to see DNA integrity; however, for us also the results of this assay were not clear, and then we decided to include the cleaved caspase-3 by Western blot analysis to confirm the first. Western blot was selected because is a quantitative analysis and we could associate it to apoptosis phenomenon. Respect to a statistical power analysis, we re-analyzed our casapse-3 values but now using one-way analysis of variance (ANOVA) test; and our results were the same. We did not find statistical differences, you can see it below: One Way Analysis of Variance Data source: Data 1 in Notebook 1

7 Normality Test: Passed (P = 0.129) Equal Variance Test: Passed (P = 0.558) Group Name N Missing Mean Std Dev SEM CN 6 0 1.000 0.249 0.101 CT 6 0 1.000 0.352 0.144 CTF 6 0 0.917 0.638 0.261 Source of Variation DF SS MS F P Between Groups 2 0.0278 0.0139 0.0702 0.932 Residual 15 2.966 0.198 Total 17 2.994 The differences in the mean values among the treatment groups are not great enough to exclude the possibility that the difference is due to random sampling variability; there is not a statistically significant difference (P = 0.932). Power of performed test with alpha = 0.050: 0.049 The power of the performed test (0.049) is below the desired power of 0.800. Less than desired power indicates you are more likely to not detect a difference when one actually exists. Be cautious in over-interpreting the lack of difference found here. 4. The above mentioned limitations in the TUNEL and activated caspase 3 analyses reported in this manuscript are also impacted by a re-analysis of the PCNA versus Ki-67 immunostaining results for nuclei in liver tissue sections. Ki-67 is strictly a marker of replication fork presence due to proliferation, while PCNA has been demonstrated to positively stain both proliferating nuclei and those actively undergoing DNA repair [Nucleic Acids Res 27:4476, 1999; Circulation 99:2757, 1999]. Given that serial sections were used to immunostain the liver tissue sections, this study should be able to assess how many nuclei stained positive for both Ki-67 & PCNA, as opposed to those nuclei that stained positive only for PCNA. This would provide an assessment of cell nuclei that are not actively engaged in replication, but rather DNA damage repair. This type of assessment would be able to be cross referenced against the TUNEL/caspase3 data in Figure 3 in order to provide a more definitive assessment of whether EMF was altering apoptosis outcomes. For example in Table 1, if one assumes that all nuclei staining positive for Ki-67 were also staining positive for PCNA, then one could perform a simple subtraction [PCNA results Ki-67 results = # nuclei undergoing DNA damage repair] to provide an assessment of the number of liver cells attempting to repair DNA damage. This type of calculation would yield the following data: Group PCNA Ki67 # DNA damage repair cells # cyclin D1 positive NC 22-7 cells/mm2 = 15 cells/mm2 49 cells/mm2 CT 492-192 cells/mm2 = 300 cells/mm2 265 cells/mm2 CTF 59-25 cells/mm2 = 34 cells/mm2 45 cells/mm2 There is a remarkable similarity in the proportions of cells exhibiting DNA damage repair and cyclin D1 detection. I do not believe this similarity is a coincidence, and this possibility needs to be more rigorously investigated. If this holds true, then EMF altered the amounts of DNA damage to liver cells in the MRHM model. To reveal the serial tissues by immunohistochemistry, the sequence of the analysis was as follows: tissue 1, Cyclin D1, tissue 2, Ki-67; tissue 3, GSTP and tissue 4, PCNA. As you can see the tissue section used to reveal Ki-67 is distant from the tissue section to reveal PCNA, considering that the nucleus of each hepatocyte has an average of 8 µm and each serial tissue has a thickness of 4 µm, it is not possible to observe accurately the marks of each protein overlapped. In addition, the quantification of positive nuclei for

8 PCNA, Ki-67 and cyclin D1 was performed by taking 10 random areas from each tissue section independently per animal, so that no overlaps between them. To make the suggested analysis, it would take us much longer. However, we consider this important suggestion and we decided to discuss it only as a possibility, which was included in the "Discussion" section. Moreover, we included additional information in the subtitled "The application of 4.5 mt - 120 Hz ELF-EMF inhibits proliferation during in vivo hepatocarcinogenesis" of the Results section. Dear Dr. Ronald J Midura, we are grateful for this important suggestion, because it given us insights to investigate new aspects related to our project. "Discussion" The correlation between the presence of Ki-67, PCNA and cyclin D1 has already been studied; while PCNA participates in replication and DNA repair and is also closely associated with the cell cycle machinery, Ki-67 is a specific replication marker associated with cell cycle entry given that participates in all active phases of the cell cycle except for the G0 phase [15, 16, 27]. We evaluated their expression by immunohistochemestry and found that ELF-EMF altered the amounts of positive cell expressing these proteins and we concluded that this effect is associated with the diminishing of cell proliferation. However, given that PCNA takes part else in DNA repair and that the amount of PCNA-positive cells were larger than Ki-67 (see Table 1); if we assume that all Ki-67-positive cells were also PCNA-positive, it is striking to observe that from the subtraction between them ([PCNA Ki67]), the number of remaining PCNA-positive cells is similar to the number of cyclin D1-positive cells. Thus, the remaining PCNA-positive cells would be involved in DNA repair, but not in proliferation processes. A similar analysis was reported in human myocytes, where was observed that TUNEL-positive cells can simultaneously express PCNA, but not Ki-67 [28]. This also could explain why the slight increase of TUNEL-positive cells observed in CT group, was not confirmed by cleaved caspase-3 detection. Together this information, suggests that the ELF-EMF could be altering the amount of cells bearing DNA damage in the MRHM model. Further studies are required to determine the nature of this possible association. The application of 4.5 mt - 120 Hz ELF-EMF inhibits proliferation during in vivo hepatocarcinogenesis" To determine whether the application of 4.5 mt - 120 Hz ELF-EMF had an effect on cell proliferation, which is a characteristic alteration from the induction of experimental hepatocarcinogenesis, we analyzed the expression of PCNA, which participates in replication and DNA repair, and the expression of Ki-67, a specific replication marker, Referee 4 Reviewer: Luciana Dini - Major Compulsory Revisions 1. The last part of the discussion section is too speculative and only one reference, but no data, is given in support of the hypothesis.

9 In the Discussion section, we included additional information the supports our hypothesis about electromagnetic fields effects on hepatocarcinogesis model used: ELF-EMF are able to interact with moving electrons and to increase electron transfer rates in chemical reactions [30]. However, the interaction of ELF-EMF exposure with biological systems, from a physical point of view, remains unclear. Nevertheless, a biophysical model has recently been hypothesized. In this model, the action mechanism of the electromagnetic fields in cells occurs through the forced vibration of each of the free ions that exist on both sides of all plasma membranes and that can move across of them using transmembrane proteins, which disrupt the electrochemical balance of the plasma membrane and, therefore, the whole function of the cell [31]. Furthermore, evaluations in experimental models have been established that the electromagnetic fields are able to modulate the intracellular calcium (Ca 2+ ) when cellular homeostasis is disrupted [32]. Calcium is a highly versatile intracellular signal that can regulate many different cellular functions, whether normal or pathological; thus, the consequences of Ca 2+ signaling depend of steady state between Ca 2+ influx, efflux, and storage [33]. Cancer development takes place through rapid proliferation and the continuous increase of altered cells that modify the cellular environment [19], including the flow of ionic charges across the cell membrane, such as Ca 2+ flow. Given that several blockers of Ca 2+ entry inhibit tumor growth [34], we cannot discount that ELF-EMF could be regulating Ca 2+ flow in the cells. Therefore, 2. Magnification of figs 3 and 4 must be the final magnification and not reported whit the magnification of the objective. The magnifications of figures 3 and 4 (now 4 and 5, respectively) in the Figure legends section were changed; moreover, on their down-right part, the figures show the scale bar corresponding to their magnifications. Figure 4. Effect of ELF-EMF exposure on apoptosis. Representative liver sections of each treatment are shown. (A) Positive control tissue that was treated with DNase I. (B) Normal control tissue. (C) CT protocol. (D) CTF protocol; images magnification 200x. (E) Western blot analysis for cleaved caspase 3 levels. Caspase 3 was normalized with actin expression used as the loading control. The expression of NC was adjusted to one in the densitometric units scale; n = 6 for each group. Figure 5. Immunohistochemical analysis of GST-p, PCNA, Ki-67 and cyclin D1 expression. Four serial liver sections from each treatment are shown in columns. Immunostaining for each protein is displayed in rows. GST-p detection shows the preneoplastic lesions for localization of PCNA, Ki-67 and cyclin D1 (CD1) expression. NC, normal control; CT, carcinogenic treatment; CTF, carcinogenic treatment plus ELF-EMF exposure; images magnification 200x; n = 6 for each group. - Discretionary Revisions 1. It should be also taken into account the effect of ELF-EMF on the normal liver tissue, since the cells response to exposure could also depend on their physiological conditions. Effectively, our normal control group (NC) was not subjected to ELF-EMF exposure, however, there are investigations where normal rats Fischer 344 were exposed to electromagnetic fields by a longer period to demonstrate the effects of electromagnetic

10 fields on the induction of general carcinogenesis and at level of hematological modifications. The results showed that electromagnetic fields were not able to induce the carcinogenesis and to modify the hematological variations in the rats. (R Mandeville, et al. FASEB journal 1997,11:1127-1136; DU Cakir, et al. Archives of Medical Research 2009, 40:352-356). 2. Why the Authors have not investigated on the cytoskeleton, whose modifications exerted by exposure via the action on the membrane ions, could block the formation of the mitotic spindle? We realize that we did not investigate the effect of field exposure on the cell cytoskeleton, and we did not due to different reasons: First, our experiments were conducted in an in vivo model, which has the advantage that it includes all the variables of an organism and not just those contained in the isolated cells. Second, we are also aware that during the initial investigation of a specific area, from the results emerge many questions which must be resolved. However, this work represents our first approach about the effect of electromagnetic field on experimental hepatocarcinogenesis. Undoubtedly there are many unanswered questions, for these reasons we consider to perform further analysis to answer that emerged in this work. Finally, this study clearly demonstrate a phenomenon induced by the electromagnetic field on development of preneoplastic lesions in rat liver, which is supported by the analyses performed at the level of preneoplastic lesions quantification, proliferation and apoptosis.