- Email: [email protected]
E1a is an exogenous in vivo tumour suppressor
Accepted Manuscript E1a is an exogenous in vivo tumour suppressor F.J. Cimas, J.L. Callejas-Valera, D.C. García-Olmo, J. Hernández-Losa, P. MelgarRojas, M.J. Ruiz-Hidalgo, R. Pascual-Serra, M. Ortega-Muelas, O. Roche, P. Marcos, E. Garcia-Gil, D.M. Fernandez-Aroca, S. Ramón y Cajal, J.S. Gutkind, R. SanchezPrieto, Ph.D PII:
S0304-3835(17)30240-9
DOI:
10.1016/j.canlet.2017.04.010
Reference:
CAN 13314
To appear in:
Cancer Letters
Received Date: 14 February 2017 Revised Date:
5 April 2017
Accepted Date: 9 April 2017
Please cite this article as: F. Cimas, J. Callejas-Valera, D. García-Olmo, J Hernández-Losa, P MelgarRojas, M. Ruiz-Hidalgo, R Pascual-Serra, M Ortega-Muelas, O Roche, P Marcos, E Garcia-Gil, D. Fernandez-Aroca, S Ramón y Cajal, J. Gutkind, R Sanchez-Prieto, E1a is an exogenous in vivo tumour suppressor, Cancer Letters (2017), doi: 10.1016/j.canlet.2017.04.010. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT E1a is an exogenous in vivo tumour suppressor Cimas, FJ1, Callejas-Valera1,2, JL, García-Olmo DC3, Hernández-Losa J4, Melgar-Rojas P1, Ruiz-Hidalgo MJ1,5, Pascual-Serra R1, Ortega-Muelas M1, Roche O1, Marcos P6, Garcia-Gil E1, Fernandez-Aroca DM1. Ramón y Cajal S4, Gutkind JS2, Sanchez-Prieto R1*.
RI PT
1. Laboratorio de Oncología, Unidad de Medicina Molecular, PCYTCLM/Centro Regional de Investigaciones Biomédicas Universidad de Castilla-La Mancha. Unidad Asociada de Biomedicina CSIC-UCLM, 02006, Albacete, Spain. 2. Moores Cancer Center, University of California San Diego, San Diego, CA, USA.
SC
3. Unidad de Investigación Experimental, Hospital General Universitario de Albacete, Spain/Institut de Recerca Biomèdica de Lleida, Centre de Recerca Experimental Biomèdica Aplicada, Lleida, Spain.
M AN U
4. Servicio de Anatomía Patológica, Hospital Universitario Vall d'Hebron, Universidad Autónoma de Barcelona. Barcelona, Spain. 5. Área de Bioquímica y Biología Molecular. Facultad de Medicina. Universidad de Castilla-La Mancha, Albacete, Spain. 6. Neurobiología Celular y Química Molecular del Sistema Nervioso Central. Centro Regional de Investigaciones Biomédicas. Universidad de Castilla-La Mancha, Albacete, Spain.
TE D
Short title: Anti-tumour role of E1a in vivo * To whom correspondence should be addressed:
Ricardo Sánchez Prieto Ph.D.
EP
Laboratorio de Oncología Molecular CRIB/PCYTA Universidad de Castilla-La Mancha
AC C
C/Almansa 14
02006 Albacete
[email protected] Tel: +34-967 599200 Ext: 2981 Fax: +34-967-599327
1
ACCEPTED MANUSCRIPT Abstract. The E1a gene from adenovirus has become a major tool in cancer research. Since the discovery of E1a, it has been proposed to be an oncogene, becoming a key element in the model of cooperation between oncogenes. However, E1a’s in vivo behaviour is consistent with a tumour suppressor gene, due to the
RI PT
block/delay observed in different xenograft models. To clarify this interesting controversy, we have evaluated the effect of the E1a 13s isoform from adenovirus 5 in vivo. Initially, a conventional xenograft approach was performed using previously unreported HCT116 and B16-F10 cells, showing a clear antieffect
regardless
of
the
mouse’s
immunological
background
SC
tumour
(immunosuppressed/immunocompetent). Next, we engineered a transgenic
M AN U
mouse model in which inducible E1a 13s expression was under the control of cytokeratin 5 to avoid side effects during embryonic development. Our results show that E1a is able to block chemical skin carcinogenesis, showing an antitumour effect. The present report demonstrates the in vivo anti-tumour effect of E1a, showing that the in vitro oncogenic role of E1a cannot be extrapolated in
Keywords:
TE D
vivo, supporting its future use in gene therapy approaches.
AC C
gene therapy
EP
E1a, tumour suppressor, oncogene, transgenic mouse, skin carcinogenesis,
2
ACCEPTED MANUSCRIPT 1. Introduction The E1a gene from adenovirus has become a potent tool in cancer research. In this regard, during the 1990s, E1a was used as a model to study cooperation with well-established oncogenes as v-H-Ras (1). Indeed, the effect of E1a as a blocker of the tumour suppressor pRB (2) led to its acceptance as
RI PT
an oncogene for most of the scientific community. However, a growing body of experimental evidence showed an unexpected behaviour of E1a as a tumour suppressor gene. On one hand, E1a was able to block tumour growth in xenograft models (3) and showed a surprising ability to promote chemo/radio
SC
sensitivity in different experimental models (4). All these evidence led to the consideration of E1a as a therapeutic gene which could be used in gene
M AN U
therapy approaches. Indeed, several types of tumours have been proposed as potential targets for E1a-based therapy including breast, ovarian, etc. (5). Furthermore, some attempts have been performed to use E1a as a therapeutic agent in clinical trials (6,7).
The mechanisms proposed to explain E1a associated transformation has related to cell cycle alteration, scape from oncogenic-induced senescence,
TE D
blockage of tumour suppressor genes, etc. (for a review see (8). In the case of the anti-tumour behaviour of E1a, several possibilities have been proposed. For example, the effect of E1a on certain oncogenic proteins as Her2/neu or EGFR
EP
have been considered as major mechanisms (9,10). However, this type of mechanism does not seem to be universal (11). Indeed, recent evidence supports the existence of more complex mechanisms that could account for the
AC C
anti-tumour activity of E1a, such as the deregulation of miRNA 520 h (12). Nonetheless, the molecular basis of E1a functions in transformation or in tumour suppression is an intriguing question. For example, both properties seem to share the same region in terms of binding to cellular proteins, mainly through the CR2 domain (13). For instance, the presence of a CR2 domain seems to be mandatory to avoid senescence induced by v-H-Ras in normal cells as a preliminary step for transformation (14), but it has also been shown that this region is strictly required for the anti-tumour effect of E1a in murine carcinoma derived cell lines (15). Nonetheless, most, if not all, of the evidence of the properties of E1a in terms of transformation/tumour suppression are 3
ACCEPTED MANUSCRIPT based on cell culture and xenograft approaches in immunocompromised mice which, although useful tools in cancer research, are quite far from the context of real tumours. In an attempt to fully clarify this interesting discrepancy of E1a as an oncogene
or
a
putative
tumour
suppressor,
we
decided
to
use
RI PT
immunosuppressed and immunocompetent animal models, and to develop an inducible, transgenic mouse model to study the oncogenic or anti-tumour properties of E1a.
Our results show that expression of E1a 13s is related to tumour
SC
suppression in animals. Specifically, E1a expression was related to the blockage of tumour growth of either tumours produced by injection of tumour
M AN U
cells in animals, or those induced by 7,12-Dimethylbenz(a)anthracene/12-O-
AC C
EP
TE D
tetradecanoyl-phorbol-13-acetate (DMBA/TPA) in a skin carcinogenesis assay.
4
ACCEPTED MANUSCRIPT 2. Materials and Methods
2.1 Cell lines. Human colon cancer cell line HCT116 and mouse melanoma B16-F10 cells (ATCC) were maintained in 5% CO2 and 37°C. Cells were cultured in
RI PT
Dulbecco’s modified Eagle’s medium supplemented with 10% foetal bovine serum, 1% glutamine plus 1% antibiotics (Sigma Aldrich, Tres Cantos, Madrid, Spain).
SC
2.2 Transfections and infections.
Lentiviral production and infection was performed as previously described (16). Host cells were infected with lentivirus expressing empty vector (E.V.) or E1a
M AN U
13s and 48 hours later, infected cells were selected using puromycin (SigmaAldrich) at 2 µg/ml for HCT116 and 1.5 µg/ml for B16 cells. Infected cells were routinely maintained at the appropriate concentrations of puromycin.
2.3 Western blotting.
Cell collection, lysis and western blotting were developed as previously
TE D
described (16). Antibodies against E1a (sc-25) and tubulin (sc-32293) were purchased from Santa Cruz Biotechnology (Quimigen, Madrid, Spain). Antibody detection was achieved by enhanced chemoluminescence (Amersham, GE
EP
Health Care, Barcelona, Spain) in a LAS-3000 system (FujiFilm, Tokyo, Japan). The results show a representative blot out of three with nearly identical results.
AC C
Tubulin was used as a loading control.
2.4 Viability assays.
Viability was evaluated by MTT assay (17). Briefly, the MTT assays were performed using 2x104 cells/well plated in 24-well plates up to 96 hours. The absorbance at 570 nm was determined using a Biokinetics plate reader (BioTek Instruments, Inc, Winooski, VT, USA). Data are the average of at least 3 independent experiments performed in triplicate.
5
ACCEPTED MANUSCRIPT 2.5 Animal studies. Animal studies were carried out according to the NIH-Intramural Animal Guide for the Care and Use of Laboratory Animals and approved by the Ethics in
RI PT
Animal Care Committee of the University of Castilla-La Mancha.
2.6 Xenograft assays.
For xenograft assays, 5x106 HCT116 cells and 1x106 B16-F10 cells expressing E1A or E.V. were injected in 5-6-week-old BALB/c-Nude or C57BL/6 female using the formula V=Dxd2/2.
M AN U
2.7 Animal transgenesis and genotyping.
SC
mice, respectively. Tumour size was measured by a caliper and calculated
The cK5-rtTA/tet-E1a transgenic mouse model was generated by crossing cK5rtTA and Tet-E1A transgenic mice derived from the FVB/N mouse strain. Transgenic cK5-rtTA and wild type FVB/N mice have been previously described (18). For the generation of Tet-E1a transgenic FVB/N mice, E1a 13s coding sequences were cloned between BamHI and NotI downstream of the seventh
TE D
Tet-responsive element (Tet-O7) in a modified pBSRV vector by PCR using pLSIP-13s as a template, which contains E1a 13s sequence from adenovirus 5 (16,19). Briefly, E1a 13s was amplified by conventional PCR from pLESIP-13s
sequences:
EP
vector by using the following primers to add BamHI and NotI cloning forward
sequence
GGGGGGATCCACCATGAGACATATTATCTGCCACGGAGG-3´,
5´reverse
AC C
sequence 5´-GGGGGCGGCCGCTTATGGCCTGGGGCGTTTAC-3´. The PCR amplification was performed starting with an initial denaturation step (94°C for 10 min) followed by 35 cycles of 94°C for 40 s, 55° C for 40 s and 72°C for 1 min, and a final elongation step (72°C for 10 min). The amplified E1a 13s was run in a 2% agarose gel and purified by using a DNA Purification Kit (Promega). After purification, E1a 13s PCR product and the modified pBSRV vector were digested using BamHI and NotI restriction enzymes (Fermentas), then cloned with TAKARA ligation kit. Five positive colonies were tested by PCR, BamHI and NotI digestion, and sequenced. The sequences were aligned by using the ClustalW software and all five clones exhibited 100% homology regarding the 6
ACCEPTED MANUSCRIPT E1a 13s sequence (data not shown). Response to doxycycline induction was evaluated by western blot (data not shown) and the most inducible clone was selected for microinjection. The DNA fragment containing the expression cassette was isolated from vector by PmeI digestion, and purified for microinjection into FVB/N mouse fertilised oocytes. Transgenic mice were
RI PT
identified for the presence of transgenes by screening genomic DNA from tail biopsies by PCR using the following primers for E1a: forward sequence 5′reverse
GCAGGAAGGGATTGACTTACTC-3′,
sequence
5′-
CAAACTCCTCACCCTCTTCATC-3′; and for rtTA: forward sequence 5′reverse
sequence
5′-
SC
CCGGATCCACCATGCCTAAGAGCCCACG-3′,
ATCTGAATGTACTTTTGCTCCATTGCGAT-3′. PCR was performed under the following conditions: 95°C for 4 min, followed by 3 5 cycles of 95°C for 30 s,
M AN U
55°C for 30 s, and 72°C for 1 min, and a final cycl e of 5 min at 72°C. The identification protocol for the presence of the transgenes is the same throughout the present work. For E1a induction in transgenic mice, Doxycycline was provided after birth in grain-based food pellets (Test-Diet Ltd.) at 6 g/kg. No randomisation was used and all experiments were conducted using littermate
TE D
controls.
2.8 Tumour induction.
Mice were shaved in the back and tumours were initiated by topical treatment
EP
with a single dose of DMBA (0.5 µg/µl in acetone) and followed 15 days later by the tumour promotion phase in which mice were treated twice a week with TPA
AC C
(0.06 µg/µl in acetone) for 24 weeks.
2.9 RNA isolation, reverse transcription and Real-time quantitative PCR.
Expression of E1a in mice was analysed by SYBR Green quantitative real-time PCR using a 2−∆∆Ct method and referred to the lowest positive expressing cK5rtTA-E1a + Doxy animal. Total RNA was obtained using the RNeasy Fibrous Tissue Mini Kit (Cat. 74704, Qiagen) following the instructions provided by the manufacturer, and reverse transcription was performed from 1 µg of RNA following the instructions for RevertAid First Strand cDNA Synthesis Kit (Cat. K1621, ThermoFisher). The amount of cDNA was quantified by SYBR Green quantitative real-time PCR using an ABIPrism 7500 FAST Sequence Detection 7
ACCEPTED MANUSCRIPT System (Applied Biosystems). cDNA was amplified using SYBR1 Green PCR Master Mix (Applied Biosystems) in the presence of specific oligonucleotides. Primers for all target sequences were designed using the Primer Express software provided with the 7500 Sequence Detection System (Applied Biosystems). Oligonucleotides used for E1a amplification were the following:
AAGGACCGGAGTCACAGCTA-3′.
As
an
RI PT
forward sequence 5′-TACCCGCCGTCCTAAAATGG-3′, reverse sequence 5′endogenous
control,
mouse
ribosomal P0 mRNA levels were evaluated using the following primers: forward sequence
5′-AAGCGCGTCCTGGCATTGTCT-3′,
reverse
sequence
5′-
performed as previously described (17).
SC
CCGCAGGGGCAGCAGTGGT-3′. The PCR conditions and quantification was
M AN U
2.10 Tissue preparation, histology, and immunohistochemistry. Histological images were obtained from fixed sections of skin samples. All tissue samples were fixed in zinc formalin fixative buffer (Sigma-Aldrich) overnight and then transferred to 70% ethanol. Fixed tissues were embedded in paraffin and sectioned to a thickness of 4 µm. For immunohistochemistry, paraffin sections were automatically de-paraffinised and treated with cell
TE D
conditioning 1 solution (pH 8) for antigen retrieval (Ventana Medical Systems, Tucson, AZ, USA). Staining was performed with an automated immunostainer (Beckmarck XT, Ventana Medical Systems) and was visualised using the
EP
ultraViewTM Universal DAB detection Kit (Ventana Medical Systems) using E1a antibody (sc-25). Haematoxylin and eosin staining was developed following the
AC C
Ventana Medical Systems protocols and visualised as previously described.
2.11
Statistical analysis
The data are reported as the mean ± SD. Statistical analysis was performed using the Prism 5.00 software (GraphPad) and Office Excel 2013 (Microsoft). Significance was determined using a t-test. The statistical significance of differences is indicated in figures by asterisks as follows: * ⇒ p<0.05; ** ⇒ p<0.01; and *** ⇒ p<0.001.
8
ACCEPTED MANUSCRIPT
3. Results 3.1 E1a blocks tumour growth in athymic mice and in immunocompetent
RI PT
animals. Initially, we evaluated the effect of E1a 13s in an experimental model in which E1a has been previously expressed, such as the human colon cancer derived cell line HCT116, but in which the in vivo effect remains unknown (20). As expected, E1a expression (Fig. 1A) led to an increase of cell proliferation, as
SC
shown by an MTT assay (Fig. 1B), which suggested an oncogenic role. Then, we examined E1a expression in an in vivo model and, as shown in Figure 1C,
M AN U
the expression of E1a was related to a full blockage of tumour generation in athymic mice. Thus, paradoxically, the in vivo assays suggested an anti-tumour role of E1a. In fact, our result in HCT116, not previously published, increases the number of models in which E1a exerts an anti-tumour effect in vivo. Although the results obtained in the HCT116 cell line supported an anti-
TE D
tumour role for E1a in vivo, the lack of an immune competent environment remained a major issue. Therefore, we decided to challenge the role of E1a as an anti-tumour gene in the murine melanoma B16-F10 cell line, which allowed us to develop xenograft assays in the C57BL/6 WT mouse strain. Indeed, the
already
EP
effect of E1a expression in melanocyte-derived cell lines when grafted in mice is known,
with
differential
histological
behaviour
dependent
on
AC C
immunologic background (21). After achieving a positive E1a 13s expression in B16-F10 cells (Fig. 2A), an increase in proliferation was observed (Fig. 2B). Next, we challenged the in vivo role of E1a. Interestingly, B16-F10 cells expressing E1a showed an almost full blockage of tumour formation (Fig. 2C). Therefore, these experiments allow us to conclude that E1a 13s can
exert its anti-tumour abilities in the presence of a functional immune system.
9
ACCEPTED MANUSCRIPT
3.2 E1a expression is safe and is able to block skin carcinogenesis in vivo. Despite previous evidence supporting the idea of E1a 13s as an antitumour gene in vivo, those approaches presented poor physiological and
RI PT
therapeutic relevance. In addition, other interesting questions such as the effect of E1a in non-transformed cells or the role of E1a in response to carcinogenesis cannot be fully answered using a xenograft-based approach. Therefore, we generated a transgenic mouse model based on a tet-on system, in which E1a
SC
13s expression was controlled by the cytokeratin 5 promoter, allowing its conditional expression in epithelial cells. In this regard, other attempts to prove
M AN U
E1a function in a similar context to our approach have been previously performed. However, in those cases, a refractory background to skin carcinogenesis and a constitutive expression of a mutant form of E1a lacking p300 binding was used, showing gross phenotypic abnormalities and a minimum effect in skin carcinogenesis (22).
TE D
In our mouse model, showing the expected Mendelian ratios (Sup. 1), mice at 7-8 weeks of age, exposed to doxycycline since birth, did not show any apparent phenotypic alterations (Fig. 3 A, B and C). E1a expression was analysed by immunohistochemistry showing the expected pattern, with a
EP
marked nuclear positivity of epithelial cells in the epidermal basal layer and hair follicle (Fig. 3D). Histological structure of skin did not show any difference
AC C
between E1a positive and negative animals (Fig. 3E), showing a normal development of skin in both groups. Therefore, our strategy to overcome side effects associated with E1a expression during development was fully successful.
To prove the putative anti-tumour properties of E1a in our transgenic
model, we challenged the role of E1a 13s in a conventional skin carcinogenesis protocol based on DMBA/TPA application. Therefore, 6-7 week old mice, fed with doxycycline since birth, were analysed for E1a expression by qRT-PCR using a tail biopsy. The 8 animals with the highest expression of E1a from the cK5-rtTA/tet-E1a group were selected for follow-up during the carcinogenesis 10
ACCEPTED MANUSCRIPT protocol as well as 8 random controls selected from the cK5-rtTA group (Fig. 4A). Then, animals were exposed to an initial dose of DMBA and, 2 weeks later, mice were treated twice a week with TPA for 22 weeks. As shown (Fig. 4B), E1a positive animals showed a lower number of lesions from the beginning of the assay, with statistical significance seen from week 14. However, the size
RI PT
distribution of the lesions was similar between both experimental groups (Fig. 4C). This suggests that the anti-tumour effect of E1a was mainly exerted on tumour generation, but not during the development of the tumour. Finally, 6 weeks after the end of the carcinogenesis protocol, lesions were collected and
SC
histologically analysed. Interestingly, no differences in the distribution (Fig. 4D) or histology (Fig. 5) of papillomas and carcinomas were observed between the two groups of mice. Figure 5A shows different spherical or flat papillomas (lower
M AN U
right panel Fig. 5A) with a narrow neck, with epithelial cells growing outward around a connective tissue core with finger-like or warty projections. The papillae resembled normal skin but with an increase of layer number (acanthosis),
hypergranulosis
and
hyperkeratosis.
Figure
5B
shows
representative images of carcinomas obtained from both genotypes. In squamous cell carcinomas, the tumour cells destroy the basement membrane
TE D
and form sheets or compact masses, which invade the subjacent connective tissue (dermis). Upper panels of Fig. 5B show well differentiated carcinomas, and it can be observed that the tumour cells are atypical, resembling
EP
keratinocytes. Tumour cells form round nodules with concentric, laminated layers or keratinous pearls. In lower panels of Fig. 5B, poorly differentiated squamous carcinomas are shown, with the malignant cells more pleomorphic
AC C
and keratinisation is not observed. Immunohistochemistry against E1a was performed in all the obtained
lesions, revealing a clear lack of E1a expression, except for a single papilloma with a marginal expression (Sup. 2). In addition, to prove the presence of E1a transgene in the obtained lesions, three randomly chosen carcinomas from cK5rtTA/tet-E1a+ Doxy were genotyped, showing a clear positivity for the E1a transgene; meanwhile, the sample obtained from a cK5-rtTA+ Doxy carcinoma was negative (Sup. 3). As a control, all the samples were also genotyped for rtTA, showing the expected positivity (Sup.3). Therefore, these data exclude the 11
ACCEPTED MANUSCRIPT loss of the transgenes as a putative mechanism to explain the lack of E1a expression observed in tumours. It is also noteworthy that no spontaneous tumours were observed in any of the mice in the necropsies performed. Furthermore, no differences were observed in the number or histology of metastatic tumours (Sup. 4) which are usual in DMBA/TPA carcinogenesis as in
propose an anti-metastatic role for this gene (25).
RI PT
lung or lymphatic ganglia (23,24), supporting previous observations that
Therefore, all these experiments support the idea that E1a has the ability to block skin carcinogenesis in our experimental model, being unable to
AC C
EP
TE D
M AN U
SC
promote carcinogenesis.
12
ACCEPTED MANUSCRIPT 4. Discussion Several conclusions can be obtained from the present report: First, although E1a promotes proliferation in vitro, consistent with an oncogenic character, it exhibits opposite behaviour in vivo, blocking tumour
RI PT
formation. It is noteworthy that several studies have shown the oncogenic potential of E1a, even at the molecular level, but always in non-transformed fibroblasts (1,14,26). The fact that E1a blocks in vivo tumour growth, without affecting in vitro growth or even potentiating it, supports the paradoxical character of this gene. Indeed, the increase in cell proliferation could be
SC
considered a hallmark of an oncogenic activity. However, this uncontrolled proliferation could render a situation consistent with replicative stress that could
M AN U
be a lethal mechanism for E1a expressing cells, especially under stress conditions (16). Therefore, the first and obvious conclusion might be that the effects of E1a expression in vitro (on cultured cells) cannot be extrapolated to an in vivo setting (mouse models). In fact, certain processes, such as angiogenesis, microenvironment interactions, and hypoxic conditions, cannot be mimicked in a simple cell culture plate, suggesting that the oncogenic or anti-
TE D
tumour abilities of E1a are not due to a simple deregulation of the proliferation rate. Obviously, in addition to the in vitro vs. in vivo differences, there are other big differences. For example, the oncogenic role of E1a is mainly, if not
EP
exclusively, shown in fibroblasts, while our experiments were developed in epithelial cells. Furthermore, at the molecular level, our report is a good example of the paradoxical effects of E1a in vivo vs. in vitro. The oncogenic
AC C
forms of H-Ras have been widely connected with the oncogenic ability of E1a in cell culture (26-29). However, in our transgenic model, E1a diminishes tumour formation in a model in which mutation in H-Ras is a key event (30), suggesting that the proposed model of oncogenic cooperation between H-Ras and E1a is not biologically relevant in vivo. Second, deleterious effects and gross phenotypic abnormalities have been associated with E1a expression in animal models with tissue restricted/preferential expression, such as in the eye, lung or skin (22,31,32). In this regard, our strategy to overcome side effects associated with E1a 13
ACCEPTED MANUSCRIPT expression during development was fully successful due to the lack of phenotype observed. Previous studies have shown that constitutive E1a expression in vivo promotes several alterations, as low weight and size, aberrant hair coat phenotype with crooked, shorter hair shafts and with hair follicles exhibiting a dystrophic or absent inner root sheath (22). However, as
RI PT
previously indicated, we did not detect any gross alteration in our transgenic mice model. To explain this apparent discrepancy, several issues should be considered
between
both
experimental
models.
In
addition
to
the
aforementioned expression during embryonic development that could be the
SC
main factor, the different background of mice, the use of a mutant form of E1a (NTdl646) lacking p300 binding (33) or the different K5 promoter used to control E1a expression, could account for the observed differences between both
M AN U
models.
Third, regarding the mechanism proposed for the anti-tumour effect of E1a, the previous work in a constitutive transgenic mouse model showed that E1a had a negligible effect in blocking the skin carcinogenesis induced by DMBA/TPA treatment, with no histological differences found in the lesions
TE D
obtained (20). The authors concluded that E1a is probably targeting a population of cells in which growth-differentiation behaviour was not influenced by E1a expression, based on the same histology observed in the lesions from positive and negative E1a animals in response to skin carcinogenesis. Our
EP
present study can solve, at least partially, some of these interesting questions. As we have shown, in a background prone to skin carcinogenesis (such as FVB
AC C
mice), the anti-tumour effects of E1a were clearly observed with marked differences in the number of lesions between E1a 13s expressing and control animals. In addition, our report also shows that lesions obtained from both E1atransgenic and control mice are almost identical and that no E1a 13s expression is detected, in spite of the presence of the transgene in these lesions. Our findings could be explained by different mechanisms. One possibility could be that E1a expressing cells become refractory to carcinogenesis, and only those that have lost or never have had E1a expression allow tumour formation. Another possibility could be that long-term E1a expression exerts some kind of toxicity, as in the case of E1a mediated 14
ACCEPTED MANUSCRIPT replicative stress (17), preventing tumour development. Although p300 has been proposed as a key determinant of the in vitro transforming and immortalising activity of E1a (34,35) through its implication in cell cycle deregulation, metabolic alterations and resistance to TGF-β, among other mechanisms (36-38), our in vivo data do not support this role. First, we did not
RI PT
observe spontaneous tumour formation in animals expressing the wild type E1a isoform with a functional p300 binding site. Second, our data showed a potent anti-tumour activity in an E1a isoform that retains the ability to bind p300. Therefore, all our in vivo evidence support previous observations that propose
SC
an additive role for p300-binding in the anti-tumour activity of E1a (15). Nonetheless, further studies are necessary to fully clarify the molecular basis and mechanisms of this anti-tumour ability.
M AN U
Finally, another important issue is how safe E1a expression is. Although previous reports showed that E1a can promote alteration in some organs (e.g., lung), no tumour formation was observed due to the presence of E1a (32). Indeed, it is significant that in a transgenic mouse model carrying the E1a 12s isoform and E1b, with preferential expression in lungs, tumour formation was detected (39), while mice expressing E1a 12s alone only developed hyperplasia
TE D
(40). Furthermore, E1a 13s expression in the same experimental lung model did not even promote hyperplasia (40), in agreement with our observations in skin. Therefore, all existing evidence suggests that E1a 13s is a non-tumourigenic
EP
gene when it is expressed in vivo. This observation could be relevant not only in the field of viral oncogenes, but could also have interesting implications for oncolytic viruses, especially those based on adenovirus, which are becoming a
AC C
new breakthrough in cancer therapy (41), reinforcing the role of E1a as a therapeutic gene.
In summary, independent of its in vitro effects, E1a 13s can promote a
marked anti-tumour effect in vivo in a safe way. The present report clarifies the controversy about the oncogenic role of E1a in vivo. The next step will be the identification of the putative mechanisms that may be affected by E1a to explain its in vivo anti-tumour activity.
15
ACCEPTED MANUSCRIPT Acknowledgements This work was supported by grants from Fundación Leticia Castillejo Castillo and MINECO (SAF 2015 64215-R) and from JCCM (PPII10-0141-0404) to RSP. F. Cimas was supported by Fundación Leticia Castillejo Castillo and Cátedra Enresa. RSP research institution received support from the European
RI PT
Community through the regional development funding programme (FEDER). We appreciate the helpful collaboration of J. García-Cano, L. Serrano-Oviedo, M. Gómez-Juárez and M.J Herreros and Piet de Groot. We are also grateful for
AC C
EP
TE D
M AN U
SC
the excellent technical assistance of M.G. Picazo.
16
ACCEPTED MANUSCRIPT References
Franza BR, Maruyama K, Garrels JI, Ruley HE. In vitro establishment is not a sufficient prerequisite for transformation by activated ras oncogenes. Cell. 1986 Feb 14;44(3):409–18.
2.
Kaelin WG, Pallas DC, DeCaprio JA, Kaye FJ, Livingston DM. Identification of cellular proteins that can interact specifically with the T/E1A-binding region of the retinoblastoma gene product. Cell. 1991 Feb 8;64(3):521–32.
3.
Frisch SM. Antioncogenic effect of adenovirus E1A in human tumour cells. Proc Natl Acad Sci USA. National Academy of Sciences; 1991 Oct 15;88(20):9077–81.
4.
Sanchez-Prieto R, Quintanilla M, Cano A, Leonart ML, Martin P, Anaya A, et al. Carcinoma cell lines become sensitive to DNA-damaging agents by the expression of the adenovirus E1A gene. Oncogene. 1996 Sep 5;13(5):1083–92.
5.
Chang Y-W, Hung M-C, Su J-L. The Anti-Tumour Activity of E1A and its Implications in Cancer Therapy. Arch Immunol Ther Exp. 2014 Feb 7;62(3):195–204.
6.
Hortobagyi GN, Ueno NT, Xia W, Zhang S, Wolf JK, Putnam JB, et al. Cationic liposome-mediated E1A gene transfer to human breast and ovarian cancer cells and its biologic effects: a phase I clinical trial. Journal of Clinical Oncology. 2001 Jul 15;19(14):3422–33.
7.
Yoo GH, Hung M-C, Lopez-Berestein G, LaFollette S, Ensley JF, Carey M, et al. Phase I trial of intratumoural liposome E1A gene therapy in patients with recurrent breast and head and neck cancer. Clinical Cancer Research. 2001 May;7(5):1237–45.
8.
Yousef AF, Fonseca GJ, Cohen MJ, Mymryk JS. The C-terminal region of E1A: a molecular tool for cellular cartography. Biochem Cell Biol. 2012 Apr;90(2):153–63.
SC
M AN U
TE D
EP
AC C
9.
RI PT
1.
Yu DH, Scorsone K, Hung M-C. Adenovirus type 5 E1A gene products act as transformation suppressors of the neu oncogene. Molecular and Cellular Biology. American Society for Microbiology (ASM); 1991 Mar;11(3):1745–50.
10.
Flinterman M, Gäken J, Farzaneh F, Tavassoli M. E1A-mediated suppression of EGFR expression and induction of apoptosis in head and neck squamous carcinoma cell lines. Oncogene. 2003 Apr 3;22(13):1965–77.
11.
Frisch SM, Dolter KE. Adenovirus E1a-mediated tumour suppression by a 17
ACCEPTED MANUSCRIPT c-erbB-2/neu-independent mechanism. Cancer Research. 1995 Dec 1;55(23):5551–5. Su JL, Chen PB, Chen YH, Chen SC, Chang YW, Jan YH, et al. Downregulation of MicroRNA miR-520h by E1A Contributes to Anticancer Activity. Cancer Research. 2010 Jun 14;70(12):5096–108.
13.
Frisch SM, Mymryk JS. Adenovirus-5 E1A: paradox and paradigm. Nat Rev Mol Cell Biol. 2002 Jun;3(6):441–52.
14.
Callejas-Valera JL, Guinea Viniegra J, Ramírez-Castillejo C, Recio JA, Galan-Moya E, Martinez N, et al. E1a gene expression blocks the ERK1/2 signaling pathway by promoting nuclear localization and MKP upregulation: implication in v-H-Ras-induced senescence. Journal of Biological Chemistry. American Society for Biochemistry and Molecular Biology; 2008 May 9;283(19):13450–8.
15.
Sanchez-Prieto R, Quintanilla M, Martin P, Lleonart M, Cano A, Dotto GP, et al. In vivo antitumour effect of retrovirus-mediated gene transfer of the adenovirus E1a gene. Cancer Gene Ther. 1998 Jul;5(4):215–24.
16.
Valero ML, Cimas FJ, Arias L, Melgar-Rojas P, García E, Callejas-Valera JL, et al. E1a promotes c-Myc-dependent replicative stress: implications in glioblastoma radiosensitization. Cell Cycle. Taylor & Francis; 2014;13(1):52–61.
17.
Cimas FJ, Callejas-Valera JL, Pascual-Serra R, García-Cano J, GarcíaGil E, la Cruz-Morcillo de MA, et al. MKP1 mediates chemosensitizer effects of E1a in response to cisplatin in non-small cell lung carcinoma cells. Oncotarget. 2015 Dec 12;6(42):44095–107.
18.
Vitale-Cross L. Conditional Expression of K-ras in an Epithelial Compartment that Includes the Stem Cells Is Sufficient to Promote Squamous Cell Carcinogenesis. Cancer Research. American Association for Cancer Research; 2004 Dec 15;64(24):8804–7.
19.
Callejas-Valera JL, Iglesias-Bartolome R, Amornphimoltham P, PalaciosGarcia J, Martin D, Califano JA, et al. mTOR inhibition prevents rapidonset of carcinogen-induced malignancies in a novel inducible HPV-16 E6/E7 mouse model. Carcinogenesis. Oxford University Press; 2016 Oct;37(10):1014–25.
AC C
EP
TE D
M AN U
SC
RI PT
12.
20.
Chattopadhyay D, Ghosh MK, Mal A, Harter ML. Inactivation of p21 by E1A leads to the induction of apoptosis in DNA-damaged cells. Journal of Virology. American Society for Microbiology; 2001 Oct;75(20):9844–56.
21.
Ramon y Cajal S, Suster S, Halaban R, Filvaroff E, Dotto GP. Induction of different morphologic features of malignant melanoma and pigmented lesions after transformation of murine melanocytes with bFGF-cDNA and H-ras, myc, neu, and E1a oncogenes. AJPA. American Society for Investigative Pathology; 1991 Feb;138(2):349–58. 18
ACCEPTED MANUSCRIPT Missero C, Serra C, Stenn K, Dotto GP. Skin-specific expression of a truncated E1a oncoprotein binding to p105-Rb leads to abnormal hair follicle maturation without increased epidermal proliferation. J Cell Biol. The Rockefeller University Press; 1993 Jun;121(5):1109–20.
23.
Chiba M, Aldaz CM, Conti CJ, Klein-Szanto AJ. Metastatic potential of mouse skin carcinomas produced by different protocols of chemical carcinogenesis. Invasion Metastasis. 1991;11(5):288–96.
24.
Abel EL, Angel JM, Kiguchi K, DiGiovanni J. Multi-stage chemical carcinogenesis in mouse skin: Fundamentals and applications. Nat Protoc. 2009 Aug 27;4(9):1350–62.
25.
Fabra A, Parada C, Vinyals A, Martín Duque P, Fernandez V, SanchezPrieto R, et al. Intravascular injections of a conditional replicative adenovirus (adl118) prevent metastatic disease in human breast carcinoma xenografts. Gene Ther. 2001 Nov;8(21):1627–34.
26.
Serrano M, Lin AW, McCurrach ME, Beach D, Lowe SW. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell. 1997 Mar 7;88(5):593–602.
27.
Callejas-Valera JL, Guinea Viniegra J, Ramírez-Castillejo C, Recio JA, Galan-Moya E, Martinez N, et al. E1a gene expression blocks the ERK1/2 signaling pathway by promoting nuclear localization and MKP upregulation: implication in v-H-Ras-induced senescence. Journal of Biological Chemistry. 2008 May 9;283(19):13450–8.
28.
Jenkins JR, Rudge K, Currie GA. Cellular immortalization by a cDNA clone encoding the transformation-associated phosphoprotein p53. Nature. 1984 Dec;312(5995):651–4.
29.
Zerler B, Moran B, Maruyama K, Moomaw J, Grodzicker T, Ruley HE. Adenovirus E1A coding sequences that enable ras and pmt oncogenes to transform cultured primary cells. Molecular and Cellular Biology. American Society for Microbiology (ASM); 1986 Mar;6(3):887–99.
30.
Quintanilla M, Brown K, Ramsden M, Balmain A. Carcinogen-specific mutation and amplification of Ha-ras during mouse skin carcinogenesis. Nature. 1986 Jul;322(6074):78–80.
31.
Yang Y, McKerlie C, Borenstein SH, Lu Z, Schito M, Chamberlain JW, et al. Transgenic expression in mouse lung reveals distinct biological roles for the adenovirus type 5 E1A 243- and 289-amino-acid proteins. Journal of Virology. American Society for Microbiology (ASM); 2002 Sep;76(17):8910–9.
32.
Wang Y, Albert DM, Shenk T. Adenovirus E1A expression controlled by the red pigment promoter in transgenic mice: a model for developmental abnormalities of the eye. Cell Growth Differ. 1994 Oct;5(10):1061–8.
33.
Drews R, Kolker M, Moran C, Sachar D, Chan V, Schnipper L. Genetic
AC C
EP
TE D
M AN U
SC
RI PT
22.
19
ACCEPTED MANUSCRIPT analysis of adenovirus E1A: induction of genetic instability and altered cell morphologic and growth characteristics are segregatable functions. Mutat Res. 1998 Oct 12;421(1):9–25. Whyte P, Williamson NM, Harlow E. Cellular targets for transformation by the adenovirus E1A proteins. Cell. 1989 Jan 13;56(1):67–75.
35.
Gopalakrishnan S, Douglas JL, Quinlan MP. Immortalization of primary epithelial cells by E1A 12S requires late, second exon-encoded functions in addition to complex formation with pRB and p300. Cell Growth Differ. 1997 May;8(5):541–51.
36.
Somasundaram K, El-Deiry WS. Inhibition of p53-mediated transactivation and cell cycle arrest by E1A through its p300/CBPinteracting region. Oncogene. 1997 Mar 6;14(9):1047–57.
37.
Gopalakrishnan S, Fischer RS, Quinlan MP. Induction of a complex between rasGAP and a novel 110 kD protein is required for immortalization of primary epithelial cells by the E1A 12S oncoprotein of adenovirus. Oncogene. 1996 Dec 19;13(12):2659–69.
38.
Missero C, Filvaroff E, Dotto GP. Induction of transforming growth factor beta 1 resistance by the E1A oncogene requires binding to a specific set of cellular proteins. Proc Natl Acad Sci USA. National Academy of Sciences; 1991 Apr 15;88(8):3489–93.
39.
Koike K, Jay G, Hartley JW, Schrenzel MD, Higgins RJ, Hinrichs SH. Activation of retrovirus in transgenic mice: association with development of olfactory neuroblastoma. Journal of Virology. 1990 Aug;64(8):3988–91.
40.
Yang Y, McKerlie C, Borenstein SH, Lu Z, Schito M, Chamberlain JW, et al. Transgenic Expression in Mouse Lung Reveals Distinct Biological Roles for the Adenovirus Type 5 E1A 243- and 289-Amino-Acid Proteins. Journal of Virology. 2002 Sep 1;76(17):8910–9.
41.
Rosewell Shaw A, Suzuki M. Recent advances in oncolytic adenovirus therapies for cancer. Current Opinion in Virology. 2016 Dec;21:9–15.
AC C
EP
TE D
M AN U
SC
RI PT
34.
20
ACCEPTED MANUSCRIPT Figures Figure 1. E1a expression blocks tumorigenesis in an HCT116 cell lineBALB/c-Nude xenograft model. A) E1a expression was evaluated by western blot in HCT116 cells infected with lentivirus expressing E.V. or E1a 13s. Fifty µg of total cell lysates (TCL) were
RI PT
blotted against E1A, using 50 µg of 293T cell lysate as a positive control of E1a expression. Membranes were reprobed against Tubulin as loading control. The image is representative of three independent experiments. B) E.V. or E1a 13s HCT116 cell viability was assessed up to 96 h by MTT. Bars indicate standard
SC
deviation (SD) of three independent experiments performed in triplicate. C) Mean tumour volume derived from the subcutaneous injection of 5 x 106 of E.V.
M AN U
or E1a 13s HCT116 cells. Groups were composed of six females per group. Bars indicate SD within each experimental group.
Figure 2. E1a expression blocks tumorigenesis in a syngenic B16-F10 cell line-C57BL/6 xenograft model.
TE D
A). E1a expression was evaluated by western blot in B16-F10 cells infected with lentivirus expressing E.V. or E1a 13s. Fifty µg of total cell lysates (TCL) were blotted against E1A, using 50 µg of 293T cell lysate as a positive control of
EP
E1a expression. Membranes were reprobed against Tubulin as loading control. The image is representative of three independent experiments. B) E.V. or E1a 13s B16-F10 cell viability was assessed up to 96 h by MTT. Bars indicate SD of
AC C
three independent experiments performed in triplicate. C) Mean tumour volume derived from the subcutaneous injection of 1 x 106 of E.V. or E1a 13s B16-F10 cells. Groups were composed of six females per group. Bars indicate SD within each experimental group.
Figure 3. Characterisation of a skin-specific, inducible transgenic mouse model expressing E1a. A) Representative picture of cK5-rtTA+ Doxycycline (upper) and cK5-rtTA/tetE1a+ Doxy (lower), 7-week-old mice exposed to doxycycline since birth, 21
ACCEPTED MANUSCRIPT showing no apparent differences between both genotypes. B) Average weight (gr) during weeks 5, 6 and 7 of both cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy mice exposed to Doxy since birth. Eight mice per group were analysed. Bars indicate SD. No differences were observed. C) Average length (mm) during weeks 5, 6 and 7 for both cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy
RI PT
mice exposed to doxy since birth. Eight mice per group were analysed. Bars indicate SD. No differences were observed. D) Representative images of immunohistochemistry against E1a from fixed sections of skin samples acquired from individuals of both groups analysed in A, B and C. Left side images for
SC
both genotypes were acquired at 20X and right side images are magnifications of the selected areas. E1a expression was detected in the expected areas. E) Images show a representative haematoxylin and eosin staining from fixed
M AN U
sections of skin samples acquired from both groups analysed in A, B and C following the protocol previously described. Left side images for both genotypes were acquired at 20X and right side images are magnifications of the selected areas.
Figure 4. Expression of E1a blocks tumorigenesis associated with
TE D
DMBA/TPA skin carcinogenesis.
A) Expression of E1a in eight cK5-rtTA/tet-E1a + Doxy genotype mice analysed by q-RT-PCR. The eight highest-expressing animals for E1a, relative to the
EP
cK5-rtTA/tet-E1a + Doxy lowest-expressing animal of the group, were selected for following-up under DMBA/TPA carcinogenesis assay. B) Average number of lesions per mouse in both cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy
AC C
mouse groups during DMBA/TPA treatment, evaluated once a week. Each experimental model was composed of eight mice per group, being the cK5rtTA/tet-E1a + Doxy group composed by the animals analysed in A. Bars indicate SEM. C) All the lesions obtained in DMBA/TPA assay in both cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy mouse groups (left and right column for each time point, respectively) were classified by their size. Histogram represents data from week 16 to 26 after the DMBA application. No significant differences were observed between both experimental groups. D) Recount of the total lesions recovered at the endpoint of the experiment (36 weeks) in both cK5-rtTA+Doxy and cK5-rtTA/tet-E1a + Doxy mouse groups. No significant 22
ACCEPTED MANUSCRIPT differences were observed in terms of carcinoma conversion between both experimental groups. Figure 5. Histological analysis of DMBA/TPA-derived lesions reveals no differences due to E1a expression. A) Representative haematoxylin and eosin stained fixed sections of papilloma
RI PT
samples acquired from both cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy mouse groups after the DMBA/TPA assay. No histological differences between groups were found. Images were taken at 4X. B) Representative haematoxylin and eosin staining from fixed sections of carcinoma samples acquired from both
SC
cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy mouse groups after the DMBA/TPA assay, showing no differences between groups. Images were taken
AC C
EP
TE D
differentiated carcinoma).
M AN U
at 4X (upper panels, well differentiated carcinoma) or 10X (lower panels, poorly
23
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Highlights. •
E1a blocks xenograft formation independently of the immune system status.
•
Transgenic mice expressing E1a are refractory to DMBA/TPA skin
E1a is not and in vivo oncogene.
•
E1a exerts an anti-tumour effect in vivo.
AC C
EP
TE D
M AN U
SC
•
RI PT
carcinogenesis.
S0304-3835(17)30240-9
DOI:
10.1016/j.canlet.2017.04.010
Reference:
CAN 13314
To appear in:
Cancer Letters
Received Date: 14 February 2017 Revised Date:
5 April 2017
Accepted Date: 9 April 2017
Please cite this article as: F. Cimas, J. Callejas-Valera, D. García-Olmo, J Hernández-Losa, P MelgarRojas, M. Ruiz-Hidalgo, R Pascual-Serra, M Ortega-Muelas, O Roche, P Marcos, E Garcia-Gil, D. Fernandez-Aroca, S Ramón y Cajal, J. Gutkind, R Sanchez-Prieto, E1a is an exogenous in vivo tumour suppressor, Cancer Letters (2017), doi: 10.1016/j.canlet.2017.04.010. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT E1a is an exogenous in vivo tumour suppressor Cimas, FJ1, Callejas-Valera1,2, JL, García-Olmo DC3, Hernández-Losa J4, Melgar-Rojas P1, Ruiz-Hidalgo MJ1,5, Pascual-Serra R1, Ortega-Muelas M1, Roche O1, Marcos P6, Garcia-Gil E1, Fernandez-Aroca DM1. Ramón y Cajal S4, Gutkind JS2, Sanchez-Prieto R1*.
RI PT
1. Laboratorio de Oncología, Unidad de Medicina Molecular, PCYTCLM/Centro Regional de Investigaciones Biomédicas Universidad de Castilla-La Mancha. Unidad Asociada de Biomedicina CSIC-UCLM, 02006, Albacete, Spain. 2. Moores Cancer Center, University of California San Diego, San Diego, CA, USA.
SC
3. Unidad de Investigación Experimental, Hospital General Universitario de Albacete, Spain/Institut de Recerca Biomèdica de Lleida, Centre de Recerca Experimental Biomèdica Aplicada, Lleida, Spain.
M AN U
4. Servicio de Anatomía Patológica, Hospital Universitario Vall d'Hebron, Universidad Autónoma de Barcelona. Barcelona, Spain. 5. Área de Bioquímica y Biología Molecular. Facultad de Medicina. Universidad de Castilla-La Mancha, Albacete, Spain. 6. Neurobiología Celular y Química Molecular del Sistema Nervioso Central. Centro Regional de Investigaciones Biomédicas. Universidad de Castilla-La Mancha, Albacete, Spain.
TE D
Short title: Anti-tumour role of E1a in vivo * To whom correspondence should be addressed:
Ricardo Sánchez Prieto Ph.D.
EP
Laboratorio de Oncología Molecular CRIB/PCYTA Universidad de Castilla-La Mancha
AC C
C/Almansa 14
02006 Albacete
[email protected] Tel: +34-967 599200 Ext: 2981 Fax: +34-967-599327
1
ACCEPTED MANUSCRIPT Abstract. The E1a gene from adenovirus has become a major tool in cancer research. Since the discovery of E1a, it has been proposed to be an oncogene, becoming a key element in the model of cooperation between oncogenes. However, E1a’s in vivo behaviour is consistent with a tumour suppressor gene, due to the
RI PT
block/delay observed in different xenograft models. To clarify this interesting controversy, we have evaluated the effect of the E1a 13s isoform from adenovirus 5 in vivo. Initially, a conventional xenograft approach was performed using previously unreported HCT116 and B16-F10 cells, showing a clear antieffect
regardless
of
the
mouse’s
immunological
background
SC
tumour
(immunosuppressed/immunocompetent). Next, we engineered a transgenic
M AN U
mouse model in which inducible E1a 13s expression was under the control of cytokeratin 5 to avoid side effects during embryonic development. Our results show that E1a is able to block chemical skin carcinogenesis, showing an antitumour effect. The present report demonstrates the in vivo anti-tumour effect of E1a, showing that the in vitro oncogenic role of E1a cannot be extrapolated in
Keywords:
TE D
vivo, supporting its future use in gene therapy approaches.
AC C
gene therapy
EP
E1a, tumour suppressor, oncogene, transgenic mouse, skin carcinogenesis,
2
ACCEPTED MANUSCRIPT 1. Introduction The E1a gene from adenovirus has become a potent tool in cancer research. In this regard, during the 1990s, E1a was used as a model to study cooperation with well-established oncogenes as v-H-Ras (1). Indeed, the effect of E1a as a blocker of the tumour suppressor pRB (2) led to its acceptance as
RI PT
an oncogene for most of the scientific community. However, a growing body of experimental evidence showed an unexpected behaviour of E1a as a tumour suppressor gene. On one hand, E1a was able to block tumour growth in xenograft models (3) and showed a surprising ability to promote chemo/radio
SC
sensitivity in different experimental models (4). All these evidence led to the consideration of E1a as a therapeutic gene which could be used in gene
M AN U
therapy approaches. Indeed, several types of tumours have been proposed as potential targets for E1a-based therapy including breast, ovarian, etc. (5). Furthermore, some attempts have been performed to use E1a as a therapeutic agent in clinical trials (6,7).
The mechanisms proposed to explain E1a associated transformation has related to cell cycle alteration, scape from oncogenic-induced senescence,
TE D
blockage of tumour suppressor genes, etc. (for a review see (8). In the case of the anti-tumour behaviour of E1a, several possibilities have been proposed. For example, the effect of E1a on certain oncogenic proteins as Her2/neu or EGFR
EP
have been considered as major mechanisms (9,10). However, this type of mechanism does not seem to be universal (11). Indeed, recent evidence supports the existence of more complex mechanisms that could account for the
AC C
anti-tumour activity of E1a, such as the deregulation of miRNA 520 h (12). Nonetheless, the molecular basis of E1a functions in transformation or in tumour suppression is an intriguing question. For example, both properties seem to share the same region in terms of binding to cellular proteins, mainly through the CR2 domain (13). For instance, the presence of a CR2 domain seems to be mandatory to avoid senescence induced by v-H-Ras in normal cells as a preliminary step for transformation (14), but it has also been shown that this region is strictly required for the anti-tumour effect of E1a in murine carcinoma derived cell lines (15). Nonetheless, most, if not all, of the evidence of the properties of E1a in terms of transformation/tumour suppression are 3
ACCEPTED MANUSCRIPT based on cell culture and xenograft approaches in immunocompromised mice which, although useful tools in cancer research, are quite far from the context of real tumours. In an attempt to fully clarify this interesting discrepancy of E1a as an oncogene
or
a
putative
tumour
suppressor,
we
decided
to
use
RI PT
immunosuppressed and immunocompetent animal models, and to develop an inducible, transgenic mouse model to study the oncogenic or anti-tumour properties of E1a.
Our results show that expression of E1a 13s is related to tumour
SC
suppression in animals. Specifically, E1a expression was related to the blockage of tumour growth of either tumours produced by injection of tumour
M AN U
cells in animals, or those induced by 7,12-Dimethylbenz(a)anthracene/12-O-
AC C
EP
TE D
tetradecanoyl-phorbol-13-acetate (DMBA/TPA) in a skin carcinogenesis assay.
4
ACCEPTED MANUSCRIPT 2. Materials and Methods
2.1 Cell lines. Human colon cancer cell line HCT116 and mouse melanoma B16-F10 cells (ATCC) were maintained in 5% CO2 and 37°C. Cells were cultured in
RI PT
Dulbecco’s modified Eagle’s medium supplemented with 10% foetal bovine serum, 1% glutamine plus 1% antibiotics (Sigma Aldrich, Tres Cantos, Madrid, Spain).
SC
2.2 Transfections and infections.
Lentiviral production and infection was performed as previously described (16). Host cells were infected with lentivirus expressing empty vector (E.V.) or E1a
M AN U
13s and 48 hours later, infected cells were selected using puromycin (SigmaAldrich) at 2 µg/ml for HCT116 and 1.5 µg/ml for B16 cells. Infected cells were routinely maintained at the appropriate concentrations of puromycin.
2.3 Western blotting.
Cell collection, lysis and western blotting were developed as previously
TE D
described (16). Antibodies against E1a (sc-25) and tubulin (sc-32293) were purchased from Santa Cruz Biotechnology (Quimigen, Madrid, Spain). Antibody detection was achieved by enhanced chemoluminescence (Amersham, GE
EP
Health Care, Barcelona, Spain) in a LAS-3000 system (FujiFilm, Tokyo, Japan). The results show a representative blot out of three with nearly identical results.
AC C
Tubulin was used as a loading control.
2.4 Viability assays.
Viability was evaluated by MTT assay (17). Briefly, the MTT assays were performed using 2x104 cells/well plated in 24-well plates up to 96 hours. The absorbance at 570 nm was determined using a Biokinetics plate reader (BioTek Instruments, Inc, Winooski, VT, USA). Data are the average of at least 3 independent experiments performed in triplicate.
5
ACCEPTED MANUSCRIPT 2.5 Animal studies. Animal studies were carried out according to the NIH-Intramural Animal Guide for the Care and Use of Laboratory Animals and approved by the Ethics in
RI PT
Animal Care Committee of the University of Castilla-La Mancha.
2.6 Xenograft assays.
For xenograft assays, 5x106 HCT116 cells and 1x106 B16-F10 cells expressing E1A or E.V. were injected in 5-6-week-old BALB/c-Nude or C57BL/6 female using the formula V=Dxd2/2.
M AN U
2.7 Animal transgenesis and genotyping.
SC
mice, respectively. Tumour size was measured by a caliper and calculated
The cK5-rtTA/tet-E1a transgenic mouse model was generated by crossing cK5rtTA and Tet-E1A transgenic mice derived from the FVB/N mouse strain. Transgenic cK5-rtTA and wild type FVB/N mice have been previously described (18). For the generation of Tet-E1a transgenic FVB/N mice, E1a 13s coding sequences were cloned between BamHI and NotI downstream of the seventh
TE D
Tet-responsive element (Tet-O7) in a modified pBSRV vector by PCR using pLSIP-13s as a template, which contains E1a 13s sequence from adenovirus 5 (16,19). Briefly, E1a 13s was amplified by conventional PCR from pLESIP-13s
sequences:
EP
vector by using the following primers to add BamHI and NotI cloning forward
sequence
GGGGGGATCCACCATGAGACATATTATCTGCCACGGAGG-3´,
5´reverse
AC C
sequence 5´-GGGGGCGGCCGCTTATGGCCTGGGGCGTTTAC-3´. The PCR amplification was performed starting with an initial denaturation step (94°C for 10 min) followed by 35 cycles of 94°C for 40 s, 55° C for 40 s and 72°C for 1 min, and a final elongation step (72°C for 10 min). The amplified E1a 13s was run in a 2% agarose gel and purified by using a DNA Purification Kit (Promega). After purification, E1a 13s PCR product and the modified pBSRV vector were digested using BamHI and NotI restriction enzymes (Fermentas), then cloned with TAKARA ligation kit. Five positive colonies were tested by PCR, BamHI and NotI digestion, and sequenced. The sequences were aligned by using the ClustalW software and all five clones exhibited 100% homology regarding the 6
ACCEPTED MANUSCRIPT E1a 13s sequence (data not shown). Response to doxycycline induction was evaluated by western blot (data not shown) and the most inducible clone was selected for microinjection. The DNA fragment containing the expression cassette was isolated from vector by PmeI digestion, and purified for microinjection into FVB/N mouse fertilised oocytes. Transgenic mice were
RI PT
identified for the presence of transgenes by screening genomic DNA from tail biopsies by PCR using the following primers for E1a: forward sequence 5′reverse
GCAGGAAGGGATTGACTTACTC-3′,
sequence
5′-
CAAACTCCTCACCCTCTTCATC-3′; and for rtTA: forward sequence 5′reverse
sequence
5′-
SC
CCGGATCCACCATGCCTAAGAGCCCACG-3′,
ATCTGAATGTACTTTTGCTCCATTGCGAT-3′. PCR was performed under the following conditions: 95°C for 4 min, followed by 3 5 cycles of 95°C for 30 s,
M AN U
55°C for 30 s, and 72°C for 1 min, and a final cycl e of 5 min at 72°C. The identification protocol for the presence of the transgenes is the same throughout the present work. For E1a induction in transgenic mice, Doxycycline was provided after birth in grain-based food pellets (Test-Diet Ltd.) at 6 g/kg. No randomisation was used and all experiments were conducted using littermate
TE D
controls.
2.8 Tumour induction.
Mice were shaved in the back and tumours were initiated by topical treatment
EP
with a single dose of DMBA (0.5 µg/µl in acetone) and followed 15 days later by the tumour promotion phase in which mice were treated twice a week with TPA
AC C
(0.06 µg/µl in acetone) for 24 weeks.
2.9 RNA isolation, reverse transcription and Real-time quantitative PCR.
Expression of E1a in mice was analysed by SYBR Green quantitative real-time PCR using a 2−∆∆Ct method and referred to the lowest positive expressing cK5rtTA-E1a + Doxy animal. Total RNA was obtained using the RNeasy Fibrous Tissue Mini Kit (Cat. 74704, Qiagen) following the instructions provided by the manufacturer, and reverse transcription was performed from 1 µg of RNA following the instructions for RevertAid First Strand cDNA Synthesis Kit (Cat. K1621, ThermoFisher). The amount of cDNA was quantified by SYBR Green quantitative real-time PCR using an ABIPrism 7500 FAST Sequence Detection 7
ACCEPTED MANUSCRIPT System (Applied Biosystems). cDNA was amplified using SYBR1 Green PCR Master Mix (Applied Biosystems) in the presence of specific oligonucleotides. Primers for all target sequences were designed using the Primer Express software provided with the 7500 Sequence Detection System (Applied Biosystems). Oligonucleotides used for E1a amplification were the following:
AAGGACCGGAGTCACAGCTA-3′.
As
an
RI PT
forward sequence 5′-TACCCGCCGTCCTAAAATGG-3′, reverse sequence 5′endogenous
control,
mouse
ribosomal P0 mRNA levels were evaluated using the following primers: forward sequence
5′-AAGCGCGTCCTGGCATTGTCT-3′,
reverse
sequence
5′-
performed as previously described (17).
SC
CCGCAGGGGCAGCAGTGGT-3′. The PCR conditions and quantification was
M AN U
2.10 Tissue preparation, histology, and immunohistochemistry. Histological images were obtained from fixed sections of skin samples. All tissue samples were fixed in zinc formalin fixative buffer (Sigma-Aldrich) overnight and then transferred to 70% ethanol. Fixed tissues were embedded in paraffin and sectioned to a thickness of 4 µm. For immunohistochemistry, paraffin sections were automatically de-paraffinised and treated with cell
TE D
conditioning 1 solution (pH 8) for antigen retrieval (Ventana Medical Systems, Tucson, AZ, USA). Staining was performed with an automated immunostainer (Beckmarck XT, Ventana Medical Systems) and was visualised using the
EP
ultraViewTM Universal DAB detection Kit (Ventana Medical Systems) using E1a antibody (sc-25). Haematoxylin and eosin staining was developed following the
AC C
Ventana Medical Systems protocols and visualised as previously described.
2.11
Statistical analysis
The data are reported as the mean ± SD. Statistical analysis was performed using the Prism 5.00 software (GraphPad) and Office Excel 2013 (Microsoft). Significance was determined using a t-test. The statistical significance of differences is indicated in figures by asterisks as follows: * ⇒ p<0.05; ** ⇒ p<0.01; and *** ⇒ p<0.001.
8
ACCEPTED MANUSCRIPT
3. Results 3.1 E1a blocks tumour growth in athymic mice and in immunocompetent
RI PT
animals. Initially, we evaluated the effect of E1a 13s in an experimental model in which E1a has been previously expressed, such as the human colon cancer derived cell line HCT116, but in which the in vivo effect remains unknown (20). As expected, E1a expression (Fig. 1A) led to an increase of cell proliferation, as
SC
shown by an MTT assay (Fig. 1B), which suggested an oncogenic role. Then, we examined E1a expression in an in vivo model and, as shown in Figure 1C,
M AN U
the expression of E1a was related to a full blockage of tumour generation in athymic mice. Thus, paradoxically, the in vivo assays suggested an anti-tumour role of E1a. In fact, our result in HCT116, not previously published, increases the number of models in which E1a exerts an anti-tumour effect in vivo. Although the results obtained in the HCT116 cell line supported an anti-
TE D
tumour role for E1a in vivo, the lack of an immune competent environment remained a major issue. Therefore, we decided to challenge the role of E1a as an anti-tumour gene in the murine melanoma B16-F10 cell line, which allowed us to develop xenograft assays in the C57BL/6 WT mouse strain. Indeed, the
already
EP
effect of E1a expression in melanocyte-derived cell lines when grafted in mice is known,
with
differential
histological
behaviour
dependent
on
AC C
immunologic background (21). After achieving a positive E1a 13s expression in B16-F10 cells (Fig. 2A), an increase in proliferation was observed (Fig. 2B). Next, we challenged the in vivo role of E1a. Interestingly, B16-F10 cells expressing E1a showed an almost full blockage of tumour formation (Fig. 2C). Therefore, these experiments allow us to conclude that E1a 13s can
exert its anti-tumour abilities in the presence of a functional immune system.
9
ACCEPTED MANUSCRIPT
3.2 E1a expression is safe and is able to block skin carcinogenesis in vivo. Despite previous evidence supporting the idea of E1a 13s as an antitumour gene in vivo, those approaches presented poor physiological and
RI PT
therapeutic relevance. In addition, other interesting questions such as the effect of E1a in non-transformed cells or the role of E1a in response to carcinogenesis cannot be fully answered using a xenograft-based approach. Therefore, we generated a transgenic mouse model based on a tet-on system, in which E1a
SC
13s expression was controlled by the cytokeratin 5 promoter, allowing its conditional expression in epithelial cells. In this regard, other attempts to prove
M AN U
E1a function in a similar context to our approach have been previously performed. However, in those cases, a refractory background to skin carcinogenesis and a constitutive expression of a mutant form of E1a lacking p300 binding was used, showing gross phenotypic abnormalities and a minimum effect in skin carcinogenesis (22).
TE D
In our mouse model, showing the expected Mendelian ratios (Sup. 1), mice at 7-8 weeks of age, exposed to doxycycline since birth, did not show any apparent phenotypic alterations (Fig. 3 A, B and C). E1a expression was analysed by immunohistochemistry showing the expected pattern, with a
EP
marked nuclear positivity of epithelial cells in the epidermal basal layer and hair follicle (Fig. 3D). Histological structure of skin did not show any difference
AC C
between E1a positive and negative animals (Fig. 3E), showing a normal development of skin in both groups. Therefore, our strategy to overcome side effects associated with E1a expression during development was fully successful.
To prove the putative anti-tumour properties of E1a in our transgenic
model, we challenged the role of E1a 13s in a conventional skin carcinogenesis protocol based on DMBA/TPA application. Therefore, 6-7 week old mice, fed with doxycycline since birth, were analysed for E1a expression by qRT-PCR using a tail biopsy. The 8 animals with the highest expression of E1a from the cK5-rtTA/tet-E1a group were selected for follow-up during the carcinogenesis 10
ACCEPTED MANUSCRIPT protocol as well as 8 random controls selected from the cK5-rtTA group (Fig. 4A). Then, animals were exposed to an initial dose of DMBA and, 2 weeks later, mice were treated twice a week with TPA for 22 weeks. As shown (Fig. 4B), E1a positive animals showed a lower number of lesions from the beginning of the assay, with statistical significance seen from week 14. However, the size
RI PT
distribution of the lesions was similar between both experimental groups (Fig. 4C). This suggests that the anti-tumour effect of E1a was mainly exerted on tumour generation, but not during the development of the tumour. Finally, 6 weeks after the end of the carcinogenesis protocol, lesions were collected and
SC
histologically analysed. Interestingly, no differences in the distribution (Fig. 4D) or histology (Fig. 5) of papillomas and carcinomas were observed between the two groups of mice. Figure 5A shows different spherical or flat papillomas (lower
M AN U
right panel Fig. 5A) with a narrow neck, with epithelial cells growing outward around a connective tissue core with finger-like or warty projections. The papillae resembled normal skin but with an increase of layer number (acanthosis),
hypergranulosis
and
hyperkeratosis.
Figure
5B
shows
representative images of carcinomas obtained from both genotypes. In squamous cell carcinomas, the tumour cells destroy the basement membrane
TE D
and form sheets or compact masses, which invade the subjacent connective tissue (dermis). Upper panels of Fig. 5B show well differentiated carcinomas, and it can be observed that the tumour cells are atypical, resembling
EP
keratinocytes. Tumour cells form round nodules with concentric, laminated layers or keratinous pearls. In lower panels of Fig. 5B, poorly differentiated squamous carcinomas are shown, with the malignant cells more pleomorphic
AC C
and keratinisation is not observed. Immunohistochemistry against E1a was performed in all the obtained
lesions, revealing a clear lack of E1a expression, except for a single papilloma with a marginal expression (Sup. 2). In addition, to prove the presence of E1a transgene in the obtained lesions, three randomly chosen carcinomas from cK5rtTA/tet-E1a+ Doxy were genotyped, showing a clear positivity for the E1a transgene; meanwhile, the sample obtained from a cK5-rtTA+ Doxy carcinoma was negative (Sup. 3). As a control, all the samples were also genotyped for rtTA, showing the expected positivity (Sup.3). Therefore, these data exclude the 11
ACCEPTED MANUSCRIPT loss of the transgenes as a putative mechanism to explain the lack of E1a expression observed in tumours. It is also noteworthy that no spontaneous tumours were observed in any of the mice in the necropsies performed. Furthermore, no differences were observed in the number or histology of metastatic tumours (Sup. 4) which are usual in DMBA/TPA carcinogenesis as in
propose an anti-metastatic role for this gene (25).
RI PT
lung or lymphatic ganglia (23,24), supporting previous observations that
Therefore, all these experiments support the idea that E1a has the ability to block skin carcinogenesis in our experimental model, being unable to
AC C
EP
TE D
M AN U
SC
promote carcinogenesis.
12
ACCEPTED MANUSCRIPT 4. Discussion Several conclusions can be obtained from the present report: First, although E1a promotes proliferation in vitro, consistent with an oncogenic character, it exhibits opposite behaviour in vivo, blocking tumour
RI PT
formation. It is noteworthy that several studies have shown the oncogenic potential of E1a, even at the molecular level, but always in non-transformed fibroblasts (1,14,26). The fact that E1a blocks in vivo tumour growth, without affecting in vitro growth or even potentiating it, supports the paradoxical character of this gene. Indeed, the increase in cell proliferation could be
SC
considered a hallmark of an oncogenic activity. However, this uncontrolled proliferation could render a situation consistent with replicative stress that could
M AN U
be a lethal mechanism for E1a expressing cells, especially under stress conditions (16). Therefore, the first and obvious conclusion might be that the effects of E1a expression in vitro (on cultured cells) cannot be extrapolated to an in vivo setting (mouse models). In fact, certain processes, such as angiogenesis, microenvironment interactions, and hypoxic conditions, cannot be mimicked in a simple cell culture plate, suggesting that the oncogenic or anti-
TE D
tumour abilities of E1a are not due to a simple deregulation of the proliferation rate. Obviously, in addition to the in vitro vs. in vivo differences, there are other big differences. For example, the oncogenic role of E1a is mainly, if not
EP
exclusively, shown in fibroblasts, while our experiments were developed in epithelial cells. Furthermore, at the molecular level, our report is a good example of the paradoxical effects of E1a in vivo vs. in vitro. The oncogenic
AC C
forms of H-Ras have been widely connected with the oncogenic ability of E1a in cell culture (26-29). However, in our transgenic model, E1a diminishes tumour formation in a model in which mutation in H-Ras is a key event (30), suggesting that the proposed model of oncogenic cooperation between H-Ras and E1a is not biologically relevant in vivo. Second, deleterious effects and gross phenotypic abnormalities have been associated with E1a expression in animal models with tissue restricted/preferential expression, such as in the eye, lung or skin (22,31,32). In this regard, our strategy to overcome side effects associated with E1a 13
ACCEPTED MANUSCRIPT expression during development was fully successful due to the lack of phenotype observed. Previous studies have shown that constitutive E1a expression in vivo promotes several alterations, as low weight and size, aberrant hair coat phenotype with crooked, shorter hair shafts and with hair follicles exhibiting a dystrophic or absent inner root sheath (22). However, as
RI PT
previously indicated, we did not detect any gross alteration in our transgenic mice model. To explain this apparent discrepancy, several issues should be considered
between
both
experimental
models.
In
addition
to
the
aforementioned expression during embryonic development that could be the
SC
main factor, the different background of mice, the use of a mutant form of E1a (NTdl646) lacking p300 binding (33) or the different K5 promoter used to control E1a expression, could account for the observed differences between both
M AN U
models.
Third, regarding the mechanism proposed for the anti-tumour effect of E1a, the previous work in a constitutive transgenic mouse model showed that E1a had a negligible effect in blocking the skin carcinogenesis induced by DMBA/TPA treatment, with no histological differences found in the lesions
TE D
obtained (20). The authors concluded that E1a is probably targeting a population of cells in which growth-differentiation behaviour was not influenced by E1a expression, based on the same histology observed in the lesions from positive and negative E1a animals in response to skin carcinogenesis. Our
EP
present study can solve, at least partially, some of these interesting questions. As we have shown, in a background prone to skin carcinogenesis (such as FVB
AC C
mice), the anti-tumour effects of E1a were clearly observed with marked differences in the number of lesions between E1a 13s expressing and control animals. In addition, our report also shows that lesions obtained from both E1atransgenic and control mice are almost identical and that no E1a 13s expression is detected, in spite of the presence of the transgene in these lesions. Our findings could be explained by different mechanisms. One possibility could be that E1a expressing cells become refractory to carcinogenesis, and only those that have lost or never have had E1a expression allow tumour formation. Another possibility could be that long-term E1a expression exerts some kind of toxicity, as in the case of E1a mediated 14
ACCEPTED MANUSCRIPT replicative stress (17), preventing tumour development. Although p300 has been proposed as a key determinant of the in vitro transforming and immortalising activity of E1a (34,35) through its implication in cell cycle deregulation, metabolic alterations and resistance to TGF-β, among other mechanisms (36-38), our in vivo data do not support this role. First, we did not
RI PT
observe spontaneous tumour formation in animals expressing the wild type E1a isoform with a functional p300 binding site. Second, our data showed a potent anti-tumour activity in an E1a isoform that retains the ability to bind p300. Therefore, all our in vivo evidence support previous observations that propose
SC
an additive role for p300-binding in the anti-tumour activity of E1a (15). Nonetheless, further studies are necessary to fully clarify the molecular basis and mechanisms of this anti-tumour ability.
M AN U
Finally, another important issue is how safe E1a expression is. Although previous reports showed that E1a can promote alteration in some organs (e.g., lung), no tumour formation was observed due to the presence of E1a (32). Indeed, it is significant that in a transgenic mouse model carrying the E1a 12s isoform and E1b, with preferential expression in lungs, tumour formation was detected (39), while mice expressing E1a 12s alone only developed hyperplasia
TE D
(40). Furthermore, E1a 13s expression in the same experimental lung model did not even promote hyperplasia (40), in agreement with our observations in skin. Therefore, all existing evidence suggests that E1a 13s is a non-tumourigenic
EP
gene when it is expressed in vivo. This observation could be relevant not only in the field of viral oncogenes, but could also have interesting implications for oncolytic viruses, especially those based on adenovirus, which are becoming a
AC C
new breakthrough in cancer therapy (41), reinforcing the role of E1a as a therapeutic gene.
In summary, independent of its in vitro effects, E1a 13s can promote a
marked anti-tumour effect in vivo in a safe way. The present report clarifies the controversy about the oncogenic role of E1a in vivo. The next step will be the identification of the putative mechanisms that may be affected by E1a to explain its in vivo anti-tumour activity.
15
ACCEPTED MANUSCRIPT Acknowledgements This work was supported by grants from Fundación Leticia Castillejo Castillo and MINECO (SAF 2015 64215-R) and from JCCM (PPII10-0141-0404) to RSP. F. Cimas was supported by Fundación Leticia Castillejo Castillo and Cátedra Enresa. RSP research institution received support from the European
RI PT
Community through the regional development funding programme (FEDER). We appreciate the helpful collaboration of J. García-Cano, L. Serrano-Oviedo, M. Gómez-Juárez and M.J Herreros and Piet de Groot. We are also grateful for
AC C
EP
TE D
M AN U
SC
the excellent technical assistance of M.G. Picazo.
16
ACCEPTED MANUSCRIPT References
Franza BR, Maruyama K, Garrels JI, Ruley HE. In vitro establishment is not a sufficient prerequisite for transformation by activated ras oncogenes. Cell. 1986 Feb 14;44(3):409–18.
2.
Kaelin WG, Pallas DC, DeCaprio JA, Kaye FJ, Livingston DM. Identification of cellular proteins that can interact specifically with the T/E1A-binding region of the retinoblastoma gene product. Cell. 1991 Feb 8;64(3):521–32.
3.
Frisch SM. Antioncogenic effect of adenovirus E1A in human tumour cells. Proc Natl Acad Sci USA. National Academy of Sciences; 1991 Oct 15;88(20):9077–81.
4.
Sanchez-Prieto R, Quintanilla M, Cano A, Leonart ML, Martin P, Anaya A, et al. Carcinoma cell lines become sensitive to DNA-damaging agents by the expression of the adenovirus E1A gene. Oncogene. 1996 Sep 5;13(5):1083–92.
5.
Chang Y-W, Hung M-C, Su J-L. The Anti-Tumour Activity of E1A and its Implications in Cancer Therapy. Arch Immunol Ther Exp. 2014 Feb 7;62(3):195–204.
6.
Hortobagyi GN, Ueno NT, Xia W, Zhang S, Wolf JK, Putnam JB, et al. Cationic liposome-mediated E1A gene transfer to human breast and ovarian cancer cells and its biologic effects: a phase I clinical trial. Journal of Clinical Oncology. 2001 Jul 15;19(14):3422–33.
7.
Yoo GH, Hung M-C, Lopez-Berestein G, LaFollette S, Ensley JF, Carey M, et al. Phase I trial of intratumoural liposome E1A gene therapy in patients with recurrent breast and head and neck cancer. Clinical Cancer Research. 2001 May;7(5):1237–45.
8.
Yousef AF, Fonseca GJ, Cohen MJ, Mymryk JS. The C-terminal region of E1A: a molecular tool for cellular cartography. Biochem Cell Biol. 2012 Apr;90(2):153–63.
SC
M AN U
TE D
EP
AC C
9.
RI PT
1.
Yu DH, Scorsone K, Hung M-C. Adenovirus type 5 E1A gene products act as transformation suppressors of the neu oncogene. Molecular and Cellular Biology. American Society for Microbiology (ASM); 1991 Mar;11(3):1745–50.
10.
Flinterman M, Gäken J, Farzaneh F, Tavassoli M. E1A-mediated suppression of EGFR expression and induction of apoptosis in head and neck squamous carcinoma cell lines. Oncogene. 2003 Apr 3;22(13):1965–77.
11.
Frisch SM, Dolter KE. Adenovirus E1a-mediated tumour suppression by a 17
ACCEPTED MANUSCRIPT c-erbB-2/neu-independent mechanism. Cancer Research. 1995 Dec 1;55(23):5551–5. Su JL, Chen PB, Chen YH, Chen SC, Chang YW, Jan YH, et al. Downregulation of MicroRNA miR-520h by E1A Contributes to Anticancer Activity. Cancer Research. 2010 Jun 14;70(12):5096–108.
13.
Frisch SM, Mymryk JS. Adenovirus-5 E1A: paradox and paradigm. Nat Rev Mol Cell Biol. 2002 Jun;3(6):441–52.
14.
Callejas-Valera JL, Guinea Viniegra J, Ramírez-Castillejo C, Recio JA, Galan-Moya E, Martinez N, et al. E1a gene expression blocks the ERK1/2 signaling pathway by promoting nuclear localization and MKP upregulation: implication in v-H-Ras-induced senescence. Journal of Biological Chemistry. American Society for Biochemistry and Molecular Biology; 2008 May 9;283(19):13450–8.
15.
Sanchez-Prieto R, Quintanilla M, Martin P, Lleonart M, Cano A, Dotto GP, et al. In vivo antitumour effect of retrovirus-mediated gene transfer of the adenovirus E1a gene. Cancer Gene Ther. 1998 Jul;5(4):215–24.
16.
Valero ML, Cimas FJ, Arias L, Melgar-Rojas P, García E, Callejas-Valera JL, et al. E1a promotes c-Myc-dependent replicative stress: implications in glioblastoma radiosensitization. Cell Cycle. Taylor & Francis; 2014;13(1):52–61.
17.
Cimas FJ, Callejas-Valera JL, Pascual-Serra R, García-Cano J, GarcíaGil E, la Cruz-Morcillo de MA, et al. MKP1 mediates chemosensitizer effects of E1a in response to cisplatin in non-small cell lung carcinoma cells. Oncotarget. 2015 Dec 12;6(42):44095–107.
18.
Vitale-Cross L. Conditional Expression of K-ras in an Epithelial Compartment that Includes the Stem Cells Is Sufficient to Promote Squamous Cell Carcinogenesis. Cancer Research. American Association for Cancer Research; 2004 Dec 15;64(24):8804–7.
19.
Callejas-Valera JL, Iglesias-Bartolome R, Amornphimoltham P, PalaciosGarcia J, Martin D, Califano JA, et al. mTOR inhibition prevents rapidonset of carcinogen-induced malignancies in a novel inducible HPV-16 E6/E7 mouse model. Carcinogenesis. Oxford University Press; 2016 Oct;37(10):1014–25.
AC C
EP
TE D
M AN U
SC
RI PT
12.
20.
Chattopadhyay D, Ghosh MK, Mal A, Harter ML. Inactivation of p21 by E1A leads to the induction of apoptosis in DNA-damaged cells. Journal of Virology. American Society for Microbiology; 2001 Oct;75(20):9844–56.
21.
Ramon y Cajal S, Suster S, Halaban R, Filvaroff E, Dotto GP. Induction of different morphologic features of malignant melanoma and pigmented lesions after transformation of murine melanocytes with bFGF-cDNA and H-ras, myc, neu, and E1a oncogenes. AJPA. American Society for Investigative Pathology; 1991 Feb;138(2):349–58. 18
ACCEPTED MANUSCRIPT Missero C, Serra C, Stenn K, Dotto GP. Skin-specific expression of a truncated E1a oncoprotein binding to p105-Rb leads to abnormal hair follicle maturation without increased epidermal proliferation. J Cell Biol. The Rockefeller University Press; 1993 Jun;121(5):1109–20.
23.
Chiba M, Aldaz CM, Conti CJ, Klein-Szanto AJ. Metastatic potential of mouse skin carcinomas produced by different protocols of chemical carcinogenesis. Invasion Metastasis. 1991;11(5):288–96.
24.
Abel EL, Angel JM, Kiguchi K, DiGiovanni J. Multi-stage chemical carcinogenesis in mouse skin: Fundamentals and applications. Nat Protoc. 2009 Aug 27;4(9):1350–62.
25.
Fabra A, Parada C, Vinyals A, Martín Duque P, Fernandez V, SanchezPrieto R, et al. Intravascular injections of a conditional replicative adenovirus (adl118) prevent metastatic disease in human breast carcinoma xenografts. Gene Ther. 2001 Nov;8(21):1627–34.
26.
Serrano M, Lin AW, McCurrach ME, Beach D, Lowe SW. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell. 1997 Mar 7;88(5):593–602.
27.
Callejas-Valera JL, Guinea Viniegra J, Ramírez-Castillejo C, Recio JA, Galan-Moya E, Martinez N, et al. E1a gene expression blocks the ERK1/2 signaling pathway by promoting nuclear localization and MKP upregulation: implication in v-H-Ras-induced senescence. Journal of Biological Chemistry. 2008 May 9;283(19):13450–8.
28.
Jenkins JR, Rudge K, Currie GA. Cellular immortalization by a cDNA clone encoding the transformation-associated phosphoprotein p53. Nature. 1984 Dec;312(5995):651–4.
29.
Zerler B, Moran B, Maruyama K, Moomaw J, Grodzicker T, Ruley HE. Adenovirus E1A coding sequences that enable ras and pmt oncogenes to transform cultured primary cells. Molecular and Cellular Biology. American Society for Microbiology (ASM); 1986 Mar;6(3):887–99.
30.
Quintanilla M, Brown K, Ramsden M, Balmain A. Carcinogen-specific mutation and amplification of Ha-ras during mouse skin carcinogenesis. Nature. 1986 Jul;322(6074):78–80.
31.
Yang Y, McKerlie C, Borenstein SH, Lu Z, Schito M, Chamberlain JW, et al. Transgenic expression in mouse lung reveals distinct biological roles for the adenovirus type 5 E1A 243- and 289-amino-acid proteins. Journal of Virology. American Society for Microbiology (ASM); 2002 Sep;76(17):8910–9.
32.
Wang Y, Albert DM, Shenk T. Adenovirus E1A expression controlled by the red pigment promoter in transgenic mice: a model for developmental abnormalities of the eye. Cell Growth Differ. 1994 Oct;5(10):1061–8.
33.
Drews R, Kolker M, Moran C, Sachar D, Chan V, Schnipper L. Genetic
AC C
EP
TE D
M AN U
SC
RI PT
22.
19
ACCEPTED MANUSCRIPT analysis of adenovirus E1A: induction of genetic instability and altered cell morphologic and growth characteristics are segregatable functions. Mutat Res. 1998 Oct 12;421(1):9–25. Whyte P, Williamson NM, Harlow E. Cellular targets for transformation by the adenovirus E1A proteins. Cell. 1989 Jan 13;56(1):67–75.
35.
Gopalakrishnan S, Douglas JL, Quinlan MP. Immortalization of primary epithelial cells by E1A 12S requires late, second exon-encoded functions in addition to complex formation with pRB and p300. Cell Growth Differ. 1997 May;8(5):541–51.
36.
Somasundaram K, El-Deiry WS. Inhibition of p53-mediated transactivation and cell cycle arrest by E1A through its p300/CBPinteracting region. Oncogene. 1997 Mar 6;14(9):1047–57.
37.
Gopalakrishnan S, Fischer RS, Quinlan MP. Induction of a complex between rasGAP and a novel 110 kD protein is required for immortalization of primary epithelial cells by the E1A 12S oncoprotein of adenovirus. Oncogene. 1996 Dec 19;13(12):2659–69.
38.
Missero C, Filvaroff E, Dotto GP. Induction of transforming growth factor beta 1 resistance by the E1A oncogene requires binding to a specific set of cellular proteins. Proc Natl Acad Sci USA. National Academy of Sciences; 1991 Apr 15;88(8):3489–93.
39.
Koike K, Jay G, Hartley JW, Schrenzel MD, Higgins RJ, Hinrichs SH. Activation of retrovirus in transgenic mice: association with development of olfactory neuroblastoma. Journal of Virology. 1990 Aug;64(8):3988–91.
40.
Yang Y, McKerlie C, Borenstein SH, Lu Z, Schito M, Chamberlain JW, et al. Transgenic Expression in Mouse Lung Reveals Distinct Biological Roles for the Adenovirus Type 5 E1A 243- and 289-Amino-Acid Proteins. Journal of Virology. 2002 Sep 1;76(17):8910–9.
41.
Rosewell Shaw A, Suzuki M. Recent advances in oncolytic adenovirus therapies for cancer. Current Opinion in Virology. 2016 Dec;21:9–15.
AC C
EP
TE D
M AN U
SC
RI PT
34.
20
ACCEPTED MANUSCRIPT Figures Figure 1. E1a expression blocks tumorigenesis in an HCT116 cell lineBALB/c-Nude xenograft model. A) E1a expression was evaluated by western blot in HCT116 cells infected with lentivirus expressing E.V. or E1a 13s. Fifty µg of total cell lysates (TCL) were
RI PT
blotted against E1A, using 50 µg of 293T cell lysate as a positive control of E1a expression. Membranes were reprobed against Tubulin as loading control. The image is representative of three independent experiments. B) E.V. or E1a 13s HCT116 cell viability was assessed up to 96 h by MTT. Bars indicate standard
SC
deviation (SD) of three independent experiments performed in triplicate. C) Mean tumour volume derived from the subcutaneous injection of 5 x 106 of E.V.
M AN U
or E1a 13s HCT116 cells. Groups were composed of six females per group. Bars indicate SD within each experimental group.
Figure 2. E1a expression blocks tumorigenesis in a syngenic B16-F10 cell line-C57BL/6 xenograft model.
TE D
A). E1a expression was evaluated by western blot in B16-F10 cells infected with lentivirus expressing E.V. or E1a 13s. Fifty µg of total cell lysates (TCL) were blotted against E1A, using 50 µg of 293T cell lysate as a positive control of
EP
E1a expression. Membranes were reprobed against Tubulin as loading control. The image is representative of three independent experiments. B) E.V. or E1a 13s B16-F10 cell viability was assessed up to 96 h by MTT. Bars indicate SD of
AC C
three independent experiments performed in triplicate. C) Mean tumour volume derived from the subcutaneous injection of 1 x 106 of E.V. or E1a 13s B16-F10 cells. Groups were composed of six females per group. Bars indicate SD within each experimental group.
Figure 3. Characterisation of a skin-specific, inducible transgenic mouse model expressing E1a. A) Representative picture of cK5-rtTA+ Doxycycline (upper) and cK5-rtTA/tetE1a+ Doxy (lower), 7-week-old mice exposed to doxycycline since birth, 21
ACCEPTED MANUSCRIPT showing no apparent differences between both genotypes. B) Average weight (gr) during weeks 5, 6 and 7 of both cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy mice exposed to Doxy since birth. Eight mice per group were analysed. Bars indicate SD. No differences were observed. C) Average length (mm) during weeks 5, 6 and 7 for both cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy
RI PT
mice exposed to doxy since birth. Eight mice per group were analysed. Bars indicate SD. No differences were observed. D) Representative images of immunohistochemistry against E1a from fixed sections of skin samples acquired from individuals of both groups analysed in A, B and C. Left side images for
SC
both genotypes were acquired at 20X and right side images are magnifications of the selected areas. E1a expression was detected in the expected areas. E) Images show a representative haematoxylin and eosin staining from fixed
M AN U
sections of skin samples acquired from both groups analysed in A, B and C following the protocol previously described. Left side images for both genotypes were acquired at 20X and right side images are magnifications of the selected areas.
Figure 4. Expression of E1a blocks tumorigenesis associated with
TE D
DMBA/TPA skin carcinogenesis.
A) Expression of E1a in eight cK5-rtTA/tet-E1a + Doxy genotype mice analysed by q-RT-PCR. The eight highest-expressing animals for E1a, relative to the
EP
cK5-rtTA/tet-E1a + Doxy lowest-expressing animal of the group, were selected for following-up under DMBA/TPA carcinogenesis assay. B) Average number of lesions per mouse in both cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy
AC C
mouse groups during DMBA/TPA treatment, evaluated once a week. Each experimental model was composed of eight mice per group, being the cK5rtTA/tet-E1a + Doxy group composed by the animals analysed in A. Bars indicate SEM. C) All the lesions obtained in DMBA/TPA assay in both cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy mouse groups (left and right column for each time point, respectively) were classified by their size. Histogram represents data from week 16 to 26 after the DMBA application. No significant differences were observed between both experimental groups. D) Recount of the total lesions recovered at the endpoint of the experiment (36 weeks) in both cK5-rtTA+Doxy and cK5-rtTA/tet-E1a + Doxy mouse groups. No significant 22
ACCEPTED MANUSCRIPT differences were observed in terms of carcinoma conversion between both experimental groups. Figure 5. Histological analysis of DMBA/TPA-derived lesions reveals no differences due to E1a expression. A) Representative haematoxylin and eosin stained fixed sections of papilloma
RI PT
samples acquired from both cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy mouse groups after the DMBA/TPA assay. No histological differences between groups were found. Images were taken at 4X. B) Representative haematoxylin and eosin staining from fixed sections of carcinoma samples acquired from both
SC
cK5-rtTA + Doxy and cK5-rtTA/tet-E1a + Doxy mouse groups after the DMBA/TPA assay, showing no differences between groups. Images were taken
AC C
EP
TE D
differentiated carcinoma).
M AN U
at 4X (upper panels, well differentiated carcinoma) or 10X (lower panels, poorly
23
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Highlights. •
E1a blocks xenograft formation independently of the immune system status.
•
Transgenic mice expressing E1a are refractory to DMBA/TPA skin
E1a is not and in vivo oncogene.
•
E1a exerts an anti-tumour effect in vivo.
AC C
EP
TE D
M AN U
SC
•
RI PT
carcinogenesis.