Article Text

Download PDFPDF

Comparative toxicity and proliferation testing of aflibercept, bevacizumab and ranibizumab on different ocular cells
  1. Sven Schnichels,
  2. Ulrike Hagemann,
  3. Kai Januschowski,
  4. Johanna Hofmann,
  5. Karl-Ulrich Bartz-Schmidt,
  6. Peter Szurman,
  7. Martin S Spitzer,
  8. Sabine Aisenbrey
  1. Centre of Ophthalmology, University Eye Hospital Tübingen, Tübingen, Germany
  1. Correspondence to Kai Januschowski, Centre of Ophthalmology, University Eye Hospital Tübingen, Schleichstr. 12/1, Tübingen D-72076, Germany; Kai.januschowski{at}med.uni-tuebingen.de

Abstract

Background/aims Vascular endothelial growth factor (VEGF) is a key factor in the pathogenesis of neovascular retinal diseases including age-related macular degeneration. VEGF inhibitors including ranibizumab, pegaptanib or bevacizumab improve retinal morphology and vision in many patients. The recently approved drug aflibercept (VEGF Trap-Eye/Eyelea, Regeneron, Tarrytown, New York, USA) offers a new therapy modality. We therefore tested for toxic and anti-proliferating effects of aflibercept.

Methods The effects of aflibercept (0.125, 0.5, 2 mg), ranibizumab (0.125 mg) and bevacizumab (0.3125 mg) after 1, 24, 48 and 72 h on cell morphology via phase contrast pictures, cell viability via MTS assay, total cell amount via crystal violet staining, apoptosis induction via caspase 3/7 assay and proliferation via BrdU assay were investigated. Three ocular cell lines were chosen for toxicology testing: ARPE19 cells, RGC-5 cells and 661W cells.

Results Aflibercept did not cause changes in cell morphology, induce apoptosis or cause permanent decrease in cell viability, cell density or proliferation in any cell line or concentration investigated. In general, aflibercept had fewer effects (upregulation or downregulation) compared with controls than bevacizumab or ranibizumab.

Conclusions In our experiments, aflibercept did not lead to any negative effects on retinal cell lines and might therefore be used safely in clinical applications.

  • Apotosis
  • Drugs
  • Retina

Statistics from Altmetric.com

Introduction

Clinical trials have shown that inhibition of vascular endothelial growth factor (VEGF) by either intravitreal pegaptanib (Macugen, Eyetech Pharmaceutical, New York, USA), ranibizumab (Lucentis, Genentech, San Francisco, California, USA) or bevacizumab (Avastin, Roche, Basel, Switzerland) can result in stabilisation or improvement of retinal morphology and vision in many patients with neovascular age-related macular degeneration (AMD)1–4 and other retinal diseases.5–9

Aflibercept (VEGF Trap-Eye/Eyelea, Regeneron, Tarrytown, New York, USA) is a new agent available for the treatment of exudative AMD. It is a decoy receptor with a longer half-life in rabbits and a higher affinity to VEGF compared with ranibizumab or bevacizumab.10 ,11 In addition to its effect on VEGF, aflibercept inhibits placental growth factors 1 and 2.11 ,12 Aflibercept has shown benefits in treating wet AMD in phase III trials comparable with those of ranibizumab. Since it was approved by the FDA, the drug seems to represent a safe treatment option for patients suffering from exudative AMD.13

In this study, we compare the proliferative and cytotoxic effects of three different aflibercept concentrations (0.125, 0.5 and 2.0 mg/ml), bevacizumab and ranibizumab on three different ocular cell lines (ARPE19, RGC-5 and 661W) at four different time-points (1, 24, 48, 72 h). We used the aflibercept diluent without aflibercept as an additional control.

Materials and methods

Cell culture

The rat ganglion cell-like/neuronal progenitor cell line—RGC-5—was kindly provided by Professor Neeraj Agarwal (UNT Health Science Center, Fort Worth, Texas, USA). The human retinal pigment epithelium cell line—ARPE19—was purchased from American Type Culture Collection (Manassas, Virginia, USA). The mouse photoreceptor cell line—661W—originated from Professor Dr Muayyad Al-Ubaidi.14 ,15 All cell lines were maintained throughout the experiments in Dulbecco’s modified Eagle’s medium containing 4 mM l-glutamine, 10% fetal bovine serum, 100 U/ml penicillin G and 100 μg/ml streptomycin sulphate at 37°C and 5% CO2. Depending on the specific experiment and cell line, ARPE19 cells were seeded either at a density of 30 000 cells/well in a 24-well plate or at a density of 10 000 cells/well in a 96-well plate; RGC-5 and 661W cells were seeded at a density of 10 000 cells/well in a 24-well plate or at a density of 5000 cells/well in a 96-well plate. Bevacizumab, ranibizumab and aflibercept were diluted with culture medium to obtain bevacizumab at concentrations of 0.3125 mg/ml, ranibizumab at concentrations of 0.125 mg/ml and aflibercept at concentrations of 0.125, 0.5 and 2.0 mg/ml, representing the injection doses, epiretinal doses and taking into account presumed dilution by the vitreous humor. The formula for the aflibercept diluent was prepared by the university pharmacy, Tübingen. In addition, the apoptosis inducer staurosporine (600 nM) was included in this study. Twenty-four hours after seeding, diluent, bevacizumab, ranibizumab, aflibercept or staurosporine at the desired concentration was added to the medium. Twenty-four, 48 and 72 hours after supplementation photographs were taken (100× magnification) from the 24-well plates using a phase contrast/fluorescence microscope (Axiovert 135, Zeiss, Göttingen, Germany) and AxioVision 4.6 software (Zeiss)

MTS viability assay

One, 24, 48 and 72 hours after supplementation, 20 μl of the CellTiter 96 AQueous One Solution Reagent (Promega) was directly added to the culture wells and incubated for 90 min. Then the absorbance was recorded at 490 nm with a Microplate Reader (BioTek, Synergy HT, Bad Friedrichshall, Germany) with the correction of interference at 690 nm.16

Crystal violet staining

After the MTS assay medium was removed and the cells fixed overnight with 4% paraformaldehyde, the cells were washed three times, stained with crystal violet solution (Sigma Aldrich, Steinheim, Germany), washed again, and incubated with 1% SDS for 1 h. Absorbance was determined at 595 nm (BioTek).16

Caspase 3/7 activity assay

One, 24, 48 and 72 hours after supplementation, caspase 3/7 activity was determined using the CaspaseGlo 3/7 activity kit (Promega, Madison, Wisconsin, USA) according to the manufacturer’s protocol. Luminescence was measured with a luminometer (BioTek).16

BrdU cell proliferation assay

Cellular proliferative activity was directly monitored by quantification of 5′-bromo-2′-deoxyuridine (BrdU) incorporation into the genomic DNA during cell growth. DNA synthesis was assessed by BrdU Cell Proliferation Assay (Calbiochem, La Jolla, California, USA). Briefly, 24, 48 and 72 h after supplementation, cell proliferation was monitored by adding 20 µl BrdU label to the media 24 h prior to the desired time-point. All additional steps were performed according to the manufacturer's protocol. Absorbance was determined at 450 and 595 nM (BioTek).

Statistical analysis

Data are represented as mean±SD. With every assay, four to ten different experiments were conducted per cell line and concentration or control, respectively (n=4–10). Statistical analysis was performed using JMP (V.9.0.0, SAS Institute Inc., Cary, North Carolina, USA). Dunnett's analysis was used for comparison between medium versus diluent or drugs and diluent versus drugs. Differences were considered to be significant at p<0.05. Results differing more than 0.2 or 20% were considered to be relevant.

Results

Effect of aflibercept on RGC-5 cells

Untreated RGC-5 cells appeared as a spindle-type shape (figure 1A) as observed by phase contrast pictures. This shape did not change after the treatment of aflibercept, bevacizumab or ranibizumab, independent of the time-point (figure 1A). Only staurosporine changed the morphology at every time-point investigated (figure 1A).

Figure 1

Effects of aflibercept, bevacizumab or ranibizumab on RGC-5 cells. Representative phase contrast pictures of RGC-5 cells 72 h after supplementation with medium, diluent, aflibercept (0.125, 0.5, 2 mg), ranibizumab (0.125 mg) or bevacizumab (0.3125 mg) and staurosporine (600 nM): no morphological changes of the RGC-5 cells were observed by phase contrast pictures at any time-point or concentration or drug, except for staurosporine (A). No drug (except staurosporine) had a relevant negative effect on cell viability, total cell amount, apoptosis and proliferation in RGC-5 cells. Bar graphs represent cell viability (B), total cell amount (C), caspase 3/7 activity (D) and proliferation (E) expressed as arbitrary units with the control set as 1 of RGC-5 cells 1, 24, 48 and 72 h after supplementation with different concentrations of aflibercept (0.125, 0.5, 2 mg), bevacizumab (0.3125 mg) and ranibizumab (0.125 mg) (n=4–6). (B) Although the application of the drugs shows an immediate decrease of cell viability, no drug permanently negatively affects the cell viability of the RGC-5 cells. Even more, the cell viability increased continuously with time. In contrast, staurosporine decreased the cell viability. (C) The application of the drugs does not affect the amount of cells in a relevant way. All cell amounts vary between 86% and 108% of the medium controls. In contrast, staurosporine decreased the cell amount in a time-dependent manner. (D) One hour after treatment, the caspase 3/7 activity was significantly higher in all cells supplemented with a drug. This effect was not observable at any other time-point, at which the caspase 3/7 activity varied between 89% and 115% of the control. However, staurosporine induced apoptosis, with a peak caspase 3/7 activity at the 24-hour time-point. (E) Cell proliferation was significantly changed in 3 of 18 investigated treatments: no relevant increase or decrease was observed. Proliferation varied between 101% and 118% compared with the medium only control. Only staurosporine reduced the proliferation to a minimum (2%). This figure is only reproduced in colour in the online version.

Cell viability of RGC-5 cells treated with varying concentrations of aflibercept (0.125, 0.5 or 2 mg), bevacizumab, ranibizumab and diluent did not show a relevant decrease in cell viability. However, some significant differences were detected (figure 1B). For all substances, cell viability was significantly reduced 1 h after treatment compared with medium only treated cells. The cell viability at all other time-points varied between 98% and 137% of the control. Only 2 mg aflibercept compared with medium and diluent 24 h, bevacizumab compared with medium 48 h and bevacizumab compared with medium and diluent 72 h after supplementation showed a significant difference with only bevacizumab 72 h after supplementation differing relevantly. In contrast, staurosporine significantly reduced cell viability down to 11%. No substance decreased cell viability below the amount of the apoptosis inducer staurosporine at any time-point (figure 1B).

Accordingly, cell density via crystal violet staining did not show any relevant differences. Cell density was significantly lower 1 h after application of aflibercept (2 mg) and bevacizumab compared with diluent. Cell density increased significantly for bevacizumab 24 h after treatment compared with diluent supplemented probes. All other drugs or time-points investigated did not show a significant or even relevant difference (figure 1C).

With regard to cell viability and cell density, a significant effect of the applied drugs could only be observed on caspase 3/7 activity at the 1-hour time-point. All drugs (except for 0.125 mg aflibercept) showed a significant increase of the caspase 3/7 activity. All effects were diminished at the 24-hour and later time-points, except for staurosporine treated probes with a caspase activity up to 13-fold higher compared with the control (figure 1D).

Proliferation was similar in all measurements except for 0.125 mg aflibercept 24 h and bevacizumab 48 h after supplementation. The proliferation of staurosporine treated probes was significantly reduced at all time-points (figure 1E).

Effect of aflibercept on ARPE19 cells

Untreated and treated ARPE19 cells showed no obvious morphological changes as observed by phase contrast pictures (figure 2A). Staurosporine changed the morphology at every time-point investigated (figure 2A).

Figure 2

Effects of aflibercept, bevacizumab or ranibizumab on ARPE19 cells. Representative phase contrast pictures of ARPE19 cells 72 h after supplementation with medium, diluent, aflibercept (0.125, 0.5, 2 mg), ranibizumab (0.125 mg) or bevacizumab (0.3125 mg) and staurosporine (600 nM): no morphological changes of the ARPE19 cells were observed by phase contrast pictures at any time-point or concentration or drug, except for staurosporine (A). No drug (except staurosporine) had a permanent and relevant negative effect on cell viability, total cell amount, apoptosis and proliferation in ARPE19 cells. Bar graphs represent cell viability (B), total cell amount (C) caspase 3/7 activity (D) and proliferation (E) expressed as arbitrary units with the control set as 1 of ARPE19 cells 1, 24, 48 and 72 h after supplementation with different concentrations of aflibercept (0.125, 0.5, 2 mg), bevacizumab (0.3125 mg) and ranibizumab (0.125 mg) (n=5–10). (B) No drug showed a relevant reduction in cell viability at any time-point. Although some variations between the medium were significant, none of these are regarded relevant, as the cell viability varied between 96% and 107% of the medium only vials. However, staurosporine, which served as an additional control for a negative induction, decreased the cell viability in a time-dependent manner. (C) No significant or even relevant negative effect on the cell density was observed at any time-point or condition investigated except for staurosporine, which reduced the amount of cells in a time-dependent manner. In contrast, the cell amount was even higher at many time-points compared with the controls. (D) Although the caspase 3/7 activity was significantly and relevantly higher in some drug treated probes 1 h after application, these effects were not permanent and not found at further time-points. Staurosporine induced apoptosis in a time-dependent manner. (E) At the 24-hour time-point, all drugs caused an increased proliferation. Moreover, ranibizumab even caused a relevant increase. At the 48- and 72-hour time-point, the proliferation rate was lower than that of the controls. Staurosporine reduced proliferation in a time-dependent manner. This figure is only reproduced in colour in the online version.

Cell viability of ARPE19 cells treated with varying concentrations of aflibercept, bevacizumab, ranibizumab and diluent did not show relevant decrease in cell viability. Only ranibizumab at 1 h was significantly different from medium only and diluent supplemented probes. Staurosporine decreased the cell viability in a time-dependent manner (figure 2B).

Accordingly, cell density via crystal violet staining did not show any relevant negative effects: no significant differences were found 1 h after treatment. At all other time-points, an increase in cell density could be observed in treated probes. Only a few were significant to the medium only probes (figure 2C). No significant difference was found for any drug compared with diluent (figure 2C). In contrast, staurosporine caused a time-dependent decrease in cell density (figure 2C).

Few significant differences in caspase 3/7 activity were observable immediately after application of the drugs (1 h) compared with medium or diluent supplemented probes. No differences were observed at 24, 48 and 72 h after treatment. In contrast, staurosporine induced apoptosis in a time-dependent manner (figure 2D).

All drugs had a positive effect on proliferation 24 h after application compared with control. At 48- and 72-hour time-points, the proliferation rate of treated probes ranged below the proliferation rate of control cells. These findings were significant for 2 mg aflibercept, bevacizumab and ranibizumab 48 h after application and all drugs except for ranibizumab at 72 h after application. No change in proliferation rate was defined as relevant. Staurosporine reduced proliferation in a time-dependent manner.

Effect of aflibercept on 661W cells

Untreated 661W cells showed spindle-type shape (figure 3A) as observed by phase contrast pictures. Treatment of the cells with aflibercept, bevacizumab or ranibizumab, independent of the time-point did not result in obvious changes in morphology (figure 3A). In contrast, staurosporine obviously changed morphology at every time-point investigated (figure 3A).

Figure 3

Effects of aflibercept, bevacizumab or ranibizumab on 661W cells. Representative phase contrast pictures of 661W cells 72 h after supplementation with medium, diluent, aflibercept (0.125, 0.5, 2 mg), ranibizumab (0.125 mg) or bevacizumab (0.3125 mg) and staurosporine (600 nM): no morphological changes of the 661W cells were observed by phase contrast pictures at any time-point or concentration or drug, except for staurosporine (A). No drug (except staurosporine) had a distinct negative permanent effect on cell viability, total cell amount and proliferation in 661W cells. However, an increased caspase 3/7 activity was measured in some settings. Bar graphs represent cell viability (B), total cell amount (C), caspase 3/7 activity (D) and proliferation (E) expressed as arbitrary units with the control (medium only) set as 1 of 661W cells 1, 24, 48 and 72 h after supplementation with different concentrations of aflibercept (0.125 mg; 0.5 mg; 2 mg), bevacizumab (0.3125 mg) and ranibizumab (0.125 mg) (n=4–6). (B) One and 24 hours after supplementation no drug showed a negative effect. However, 48 h after treatment, cell viability was reduced (87%–100%) in many probes compared with the medium only treated samples. In contrast, 72 h after treatment cell viability was higher in every drug treated sample than in the controls (1.30–1.59-fold higher). In contrast, staurosporine decreased the cell viability permanently. (C) The amount of cells was always higher than the amount of cells in the medium only treated cells. In contrast, staurosporine reduced the amount of cell in a time-dependent manner. (D) The caspase 3/7 activity was not more than 1.3-fold higher than that of the controls at any time-point investigated after application of the drugs or diluents. However, based on our definition that any increase of more than 20% should be regarded as relevant, the following settings are crucial: aflibercept 0.125 mg 1 h after treatment, aflibercept 0.5 and 2 mg 48 h after treatment and bevacizumab and ranibizumab 48 and 72 h after treatment. Although one should pay attention to these findings, this does not mean that these drugs are not safe because a real induction of apoptosis would be much higher, as observable with staurosporine. Staurosporine induced caspase 3/7 activity with a peak (6.8-fold) at the 24-hour time-point. (E) Cell proliferation was significantly and sometimes relevantly increased (between 1.1- and 1.28-fold) in several drug treated probes 24 and 48 h after application. In contrast, 72 h after treatment proliferation was reduced to 95%–88% relatively compared with the medium only samples (controls). Only staurosporine reduced the proliferation to a minimum (5%) in a time-dependent manner. This figure is only reproduced in colour in the online version.

Application of the drugs or the diluent led to immediate, sometimes significant increase of cell viability 1 h after treatment (103% and 125% of the control) (figure 3B). Cell viability of most treated 661W cells did not show a significant or relevant difference 24 and 48 h after application. Only two samples differed significantly: 0.125 mg aflibercept versus diluent and 2 mg aflibercept versus medium supplemented samples 48 h after stimulation. In contrast, at the 72-hour time-point, all treated probes showed significant and relevant increase in cell viability compared with untreated cells. Compared with the cells treated with diluent only, the 2 mg aflibercept treated cells had significant and relevant higher cell viability. Cell viability of staurosporine treated cells was reduced at every time-point from 24 h onwards.

Except for the 0.125 and 0.5 mg aflibercept, bevacizumab and ranibizumab treated probes 24 h, 0.125 and 2 mg aflibercept 48 h, ranibizumab 72 h after treatment, we did not record a difference in cell density. Only the 0.125 and 0.5 mg aflibercept treated probes 24 h after stimulation showed a relevant difference compared with medium. In staurosporine treated probes, the cell density was significantly and relevantly lower compared with medium only probes from 24 h onwards.

Neither aflibercept (except 0.125 mg 1 h) nor bevacizumab or ranibizumab significantly increased caspase 3/7 activity in the 661W cells 1 or 24 h after treatment compared with the control or the diluent supplemented probes. In contrast, 48 and 72 h after supplementation the caspase 3/7 activity significantly increased in all drug-treated probes compared with the medium only cells. In addition, caspase 3/7 activity increased significantly in 2 mg aflibercept, bevacizumab or ranibizumab treated probes compared with the diluent 72 h after supplementation. Moreover, except for 0.125 mg aflibercept all differences were relevant 48 h after supplementation compared with medium and diluent, whereas only bevacizumab and ranibizumab showed relevant differences 72 h after treatment. However, the known apoptosis inducer staurosporine increased the caspase 3/7 activity at 24 h up to 6.8-fold compared with control (figure 3C).

The cell proliferation rate was higher in several of the treated probes: the proliferation rate was significantly higher in bevacizumab or ranibizumab at 24 h, all drugs or diluent at 48 h, 2 mg aflibercept, bevacizumab and ranibizumab at 72 h after stimulation compared with medium only cells. Compared with diluent, the proliferation rate of bevacizumab and ranibizumab treated cells were significantly higher 24 h after treatment. Staurosporine decreased proliferation rates in a time-dependent manner.

Discussion

The most important finding of this study is the observation that aflibercept, even at the highest concentration tested, did not lead to any obvious change in cell morphology and did not induce apoptosis or a permanent decrease in cell viability, cell density or proliferation in any of the three cell lines investigated. It is important to consider that while testing aflibercept we used the maximum dose for intravitreal injection taking an inhomogeneous distribution of aflibercept into account.

The tests applied in this study (MTS viability assay, caspase 3/7 activity assay, crystal violet staining and BrdU assay) are well evaluated and standardised tests. Although each of these tests examines only a single effect: MTS for cell viability, caspase 3/7 activity for apoptosis, crystal violet staining for the amount of cells, phase contrast pictures for morphology, BrdU assay for proliferation, the combination of these assays would reveal any negative effects that might be hidden if only a single test is performed. Single tests alone could be misleading: apoptosis as an active process is sometimes characterised by a high cell activity leading to a higher reduction of MTS, which can be misinterpreted as a higher or normal cell viability. With the additional assays performed in this set of experiments, such an effect can be excluded since apoptosis would have been detected with higher caspase 3/7 activity, a less amount of cells or a lower proliferation rate.

Other interesting observations concerning the results need to be discussed further. Aflibercept was constructed to bind human VEGF isoforms.11 The RGC-5 and 661W cell lines are of rat or mouse origin, whereas the ARPE19 cell line is of human origin. However, Holash et al stated that despite its affinity for human proteins, aflibercept binds VEGF in all species tested.11 Therefore, we assume that independent of the cell line origin any possible effect should be present in these cells. Moreover, several authors successfully reported the use of aflibercept in vivo and in tissue of mice and rat origin under different conditions.17–21

In a previous study, we investigated proliferation and cytotoxic effects of bevacizumab and ranibizumab on ARPE19 and RGC-5 cells; in these experiments only bevacizumab at 0.3 mg/ml showed a significant but not relevant inhibition of proliferation.22 The results of the present study are in accordance with our previous results. We also investigated the effects of staurosporine on RGC-5 cell viability, cell density and apoptosis induction using the same assays and achieved similar results.23

Since the diluent of the agent might also cause certain effects, we decided to include the diluent in the tests. It contains sucrose, which is known to cause a higher proliferation rate in cells.24 This might explain the higher proliferation rates and higher cell density, cell viability and sometimes apoptosis ratio in our experiments and is additionally supported by the increasing cell density rate seen in figures 1C, 2C and 3C. Our results were not normalised for cell density. Therefore, with increasing cell numbers the amount of the converted agents in the assays also increases; this effect has to be taken into consideration in the interpretation of the data (figures 1–3). The diluent in bevacizumab and ranibizumab solutions, both including trehalose instead of sucrose, differs from the diluent used in the aflibercept formula. Our data indicate that the proliferation rate decreases for all drugs at the 72-hour time-point. This effect could be explained by the fact that in general the proliferation rate decreases with increasing confluence. The probes maintained in medium only proliferated slower due to the missing sucrose or trehalose. However, after 72 h, these cells proliferated relatively faster due to less confluence (figures 1D, 2D and 3D).

In the RGC-5 and ARPE19 cells, we observed changes in cell viability, cell density, a slight induction of apoptosis or changes in proliferation rate in some drug-treated probes (figures 1B–D and 2B–D). Bevacizumab only shows long term effects in cell viability in RGC-5 cells. Recently, 24-hour tests with an MTT assay on ARPE19 cells were performed by another group.25 The results of the two identical concentrations tested in that study (0.125 and 0.5 mg) are in accordance with our results.25

In the 661W cell line, we observed a slight increase in the caspase 3/7 activity. However, this is accompanied by a higher cell density and cell viability. Since we also observed a higher proliferation rate at 72 h, we assume that the sugar in the drugs increases the proliferation rate until confluence is reached. Therefore, a higher number of cells is present in the drug-treated probes and accordingly a higher cell viability and caspase 3/7 activity are observable. In conclusion, the slight induction of apoptosis cannot be considered as a concerning effect.

Our data show that the recently approved drug aflibercept does not induce toxic effects over a 72-hour time period at any of the clinically used and tested concentrations. Based on our findings, administration of aflibercept can be considered a safe treatment option for neovascular retinal diseases with regard to local effects on the retina.

Acknowledgments

The authors thank Regina Hofer for her support as well as Dr Stine Mencl, Dr François Paquet-Durand and Prof. Dr Muayyad Al-Ubaidi for providing the 661W cells. We also thank Jolene Hayes for her proofreading this manuscript.

References

Footnotes

  • Contributors SS planned the study design, performed the statistical analysis, analysed the data, wrote the manuscript and generated the figures. UH performed the experiments, planned the study design, performed the statistical analysis, cleaned and analysed the data and revised the paper. KJ planned the study design and wrote and revised the manuscript. JH performed the experiments. K-UB-S and PS revised the manuscript. MSS planned the study design and revised the manuscript. SA planned the study design and revised the manuscript.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.