Background
Tumorigenesis is a multiple step process in which genetic alterations drive the progressive transformation of normal cells into highly malignant derivatives with well-known hallmarks [
1,
2]. In addition, two “emerging hallmarks” of cancer have recently been proposed, namely deregulation of cellular energetics, and avoiding immune destruction [
3]. Furthermore, neoplastic transformation drives genome instability and mutation, and tumor-induced inflammation [
3,
4].
The idea that tumors must escape from immune recognition implies that tumors can be destroyed by the immune response [
5]. However, some tumors generate an immune suppressive environment, thus evading immune destruction [
6]. Gliomas are the most common primary tumors in the brain and are divided into four clinical grades on histopathological and prognosis basis [
7]. Several gene expression alterations and chromosomal abnormalities are commonly found in gliomas and, in some instances, these mutations correlate with the clinical grade [
8].
In most cancers, the oncogenic Ras is activated, and 20-30 % of all tumors harbor oncogenic point mutations in Ras. Moreover, if Ras is not mutated, such as in gliomas, it is frequently found that the Ras signaling pathway is disrupted [
2].
On the other hand, the tumor suppressor Rb regulates cell cycle, inhibiting progression into the S phase, by inactivating the E2F transcription factor, which is critical for DNA replication. The Cancer Genome Atlas (TCGA) project has shown that CDKN2A/p16-CDK4/6-RB pathway is altered in nearly 80 % of primary GBMs with the most frequent genetic alterations being CDKN2A gene deletion or mutation, CDK4 amplification, and RB1 mutation or deletion [
9,
10].
Natural killer cells (NK) are regarded as the first line of defense against tumors [
11]. Therefore, taking advantage of an oncogenic Ras expression and
Rb inactivation-based
in vitro model of gliomagenesis, as previously reported [
12], we explored whether these specific genetic alterations induce a cell phenotype compatible with glioma cell evasion from NK cell-mediated cytotoxicity. In addition,
in vitro transformed glioma cells were injected into SCID mice and after tumor growth, two cell lines that survived the cytotoxic effect of mice NK cells were also analyzed and showed increased resistance to NK cell-mediated cytotoxicity. Together, our results suggest that overexpression of mutated Ras, down-regulation of
Rb, or both genetic traits, confer
in vitro resistance to NK cells and that
in vivo NK cell-based selective pressure, selected cells with an increased
in vitro resistance to NK cells.
Discussion
The immune system is thought to be constantly surveying for the arising of malignant cells that would in turn be eliminated by the immune response [
13]. An “emerging hallmark” of cancer states that tumor cells are capable of avoiding immune destruction [
3].
In gliomas from human origin, the formed tumor is immunosuppressive [
14,
15]. How these tumors reach that immunosuppressive characteristic at the early stages of malignant transformation, and whether specific mutations are associated with the ability to escape from the immune response is poorly understood. Here, by using a previously reported
in vitro model of gliomagenesis [
12], we tested the hypothesis that defined changes in the expression of a proto-oncogene (Ras) and a tumor suppressor (
Rb) confer tumor cells the ability to avoid immune destruction. In particular, we addressed the resistance to NK cell-mediated cytotoxicity. Since NK-cell mediated cytotoxicity is dependent on the tumor cell membrane expression of several ligands that upon engagement with specific NK cell receptors either inhibit or activate NK cell cytotoxic function, the expression of MHC class I (ligand for inhibiting NK receptor Ly49D), Rae1δ, Rae1αβδγε, mult1, H60a, H60b, and H60c (ligands for NKG2D, an activating NK cell receptor) [
16‐
19], as well as of Fas and FasL (molecules involved in programed cell death) was assessed
in vitro transformed astrocytes. In gliomas from human origin, deregulated expression of MHC-I has been associated with NK cell-mediated cytotoxicity [
20], and stem cells from glioma patients do not express protective levels of MHC-I molecules, but they express several ligands that activate NK cells [
21].
These studies assessed the susceptibility of both stem cells and tumor cells to NK cell-mediated cytotoxicity, after the onset of malignant transformation and therefore after immunoediting had likely already taken place [
5,
22]. Although, down-regulation of MHC-I makes tumor cells susceptible to NK cell-mediated cytotoxicity [
22], high expression of MHC-I does not necessarily guarantee resistance to NK cells [
23]. Here, we assessed the expression of MHC-I on various
in vitro transformed astrocytes that therefore were not subjected to any immune-based selective pressure. Results showed a significant increase in the expression of MHC-I in the cRb
−/− cells and a significant decrease in the cRb
−/−/Ras
V12 cells (Fig.
1a). When Rb
−/−/Ras
V12 cells were inoculated into SCID mice, and the T731 tumor cell line was derived, the expression of MHC-I was similar to that of cRb
loxP/loxP cells (Fig.
1a). This MHC-I expression recovery may be the result of selective pressure exerted by NK cells in the SCID mice that selected cells with the highest levels of MHC-I, conferring them an advantage to evade the NK cell anti-tumor response. In this regard, the T731 cell line also showed the lowest percentage of NK cell-mediated cytotoxicity (Fig.
2). However, the lower expression of MHC-1 observed in
Rb
−/−
/Ras
V12 cells, and the higher expression of Rae1α and Rae 1αβγδε observed in c
Rb
−/−
cells seem to be in contradiction with their increased resistance to NK cells. Rae1 expression is low or absent in normal tissues and it is constitutively expressed on some tumor cells. Upon engagement with the NKG2D receptor on NK cells, it activates their cytotoxic activity [
24,
25]. Rae1 expression has been associated with cell proliferation [
26], and Rae1 gene family members are activated by the E2F transcription factor, which plays a central role in regulating cell cycle entry [
27]. A significant increase in the expression of Rae1δ and Rae1αβγδε was found in
Rb
−/−
astrocytes, as compared with that of parental c
Rb
loxP/loxP
astrocytes (Fig.
2b and
c). Ras induces the expression of Rae1 [
28]. However, in this case, overexpression of Ras (Ras
V12 astrocytes) did not correlate with an increase in Rae1δ or Rae1 αβγδε expression. Moreover, in c
Rb
−/−
/Ras
V12 astrocytes a significant reduction in the expression of both Rae1δ, and Rae1αβγδε was observed (Fig.
1). The T653 and T731 astrocyte lines did not show significant changes in Rae1δ, or Rae1αβγδε expression, as compared with parental c
Rb
loxP/loxP
astrocytes. This would suggest that loss of Rb promotes the expression of Rae1δ, and Rae1αβγδε. However, since these two ligands would engage with the NK cell activating NKG2D receptor, increased NK cell-mediated cytotoxicity against cRb
−/− astrocytes, as compared to that of c
Rb
loxP/loxP
astrocytes would be expected. This was clearly not the case, since the percentage of NK cell-mediated cytotoxicity was in fact lower than that for c
Rb
loxP/loxP
astrocytes (Fig.
3). The mRNA expression of four other ligands for NKG2D cell receptors was also assessed (mult1, H60a, H60b, and H60c). These ligands are poorly expressed on most normal cells, but are upregulated on tumor cells, which is keeping with the finding that mult1 mRNA expression was comparatively higher in RasV
12 and cRb
−/−/ RasV
12 cells than in cRb
loxP/loxP cells. On the other hand, the mRNA expression of mult1 was similar between cRb
loxP/loxP and cRb
−/− cells, thus suggesting that transformation due to RasV
12 over expression but not to Rb deletion would render transformed astrocytes more susceptible to NK cytotoxicity. Again, this was not the case, since both RasV
12 and cRb
−/−/ RasV
12 astrocytes were more resistant to NK cytotoxicity. However, in analyzing the mRNA expression of H60a and H60b, results evident that their expression is lower in RasV
12 and cRb
−/−, as compared to cRb
loxP/loxP astrocytes. In addition, mRNA expression of H60a and H60b is similar in cRb
−/−/ RasV
12 and cRb
loxP/loxP astrocytes. Therefore H60a and H60b expression would imply resistance to NK cell cytotoxicity, as it actually happens.
The molecule Fas induces cell death upon engagement with FasL (Fas ligand) [
29]. FasL is expressed on activated T and NK cells [
30] and thus FasL induces apoptotic cell death on Fas-expressing cells [
28]. Accordingly, a decrease in Fas expression would protect cells from cell-mediated cytotoxicity. The loss of Fas expression has been observed in melanoma, breast cancer, leukemia, and lymphoma cells [
31‐
33]. A variety of malignant tumors show increased expression of FasL, thus allowing tumor cells to induce apoptosis on cytotoxic cells, in a process known as “tumor counterattack” [
34‐
36]. Here we showed that overexpression of Ras, deletion of
Rb, or both, are sufficient to decrease the expression of Fas (Fig.
1d). Lower expression of Fas was concomitant to increased resistance to NK cell-mediated cytotoxicity (Fig.
2). No significant increase in the expression of FasL was observed in any of the transformed astrocytes. The FasL expression data suggest that genetic alterations other than just Ras overexpression and
Rb deletion are required for the “tumor counterattack” phenotype acquisition. It is tempting to speculate that this phenotype only takes place after acquisition of resistance to NK cells and further genome instability, a property of neoplastic transformation [
3].
Additional experiments were designed to evaluate the tumor growth in a syngeneic model (FVB mice) and the possible contribution of other immune cells in the tumor implantation outcome. cRb
−/−/Ras
V12, followed by cRb
−/− astrocytes were successfully implanted. This could be explained in part by the mRNA expression of H60a and H60b, as mentioned before, and the consequent resistance to NK cell-mediated cytotoxicity (Fig.
3). However, the expression of other ligands for NKG2D is contradictory, mult1 for instance. If tumor implantation can be attributed to resistance to NK cell cytotoxicity, tumor remission could then be attributed to the increase in the percentages of CD8
+ and granzyme
+ cells in the case of cRb
−/−/Ras
V12 tumor, and to the increase in the percentages of CD8
+CD69
+, CD4
+CD69
+ and granzyme
+ cells, and to the decrease in the percentage of regulatory CD4
+CD25
+ cells in the case of Ras
V12 tumor. It is worth noting that the most resistant tumor is the one with the two mutations and also the one that after 28 days post implantation only induced an increase in the percentage of CD8
+ and granzyme
+ cells as compared with the Ras
V12 that induces phenotype changes more consistent with anti-tumor immunity.
Taken together, our results suggest that Ras overexpression and Rb deletion are sufficient for the malignant transformation of astrocytes and that these genetic alterations confer transformed cells resistance to NK cell-mediated cytotoxicity, by altering the expression of NK cell receptor ligands, such as the higher expression of MHC-I observed in cRb
−/−
cells, or the lower expression of Rae1α and Rae1αβγδε observed in Rb
−/−
/RasV12 cells, and also by altering the expression of cell death associated molecules such as Fas, as observed in RasV12 cells. Its seems that the innate immune system deals with small antigenic differences between the normal and transformed tumor cells that allow tumor cells to initiate proliferation and to develop a number of oncogenic stages, however further analyses will be required.
Methods
Cell culture and cell proliferation assays
Astrocytes were isolated from 3 days old
Rb floxed mice, as previously described [
12]. Animal care and use of all experimental animals were performed in accordance with institutional ethical guidelines. In order to introduce an active
Ras allele or to promote the
Rb loss in astrocytes,
Phoenix-Eco packaging cells (a kind gift from G.P. Nolan) were transfected with empty pBABE, pBABE-HRas
V12, empty PIG, and PIG-CRE retroviral plasmids (a kind gift from P.P. Pandolfi), and
in vitro transformation was achieved by retroviral infection. The resulting cells were denominated; c
Rb
loxP/loxP
, Ras
V12,
cRb
−/−
, and
cRb
−/−
/Ras
V12. The T653 and T731 cell lines were derived from tumors formed by inoculation of Ras
v12 and
cRb
−/−
/Ras
V12 cells in SCID mice, as previously shown [
12]. A feature of the SCID mice is a deficiency in the recombination of genes needed for full maturation of T and lymphocytes. However, SCID mice harbor functional NK cells. Cells were maintained in Dulbecco Modified Eagle Medium (DMEM) (Sigma-Aldrich, St Louis, MO) supplemented with 10 % FBS and antibiotic-antimycotic solution (Bio West, Nuaillé, France). For cell growth analysis, 5x10
3 cells were plated into 24-well culture plates (Corning, NY, USA) and then fixed with methanol/acetic acid (3:1) on days 1, 3, 5, 7 of culture, for subsequent staining with crystal violet (0.1 % in PBS) and distaining with 10 % acetic acid. The relative cell number was assessed by spectrophotometry.
Syngeneic model
The performance of tumor growth in the syngeneic model was evaluated by subcutaneously injecting 1x106 cRbloxP/loxP, RasV12, cRb−/−, or cRb−/−/RasV12 astrocytes previously obtained from FVB mice (n = 10). Tumours from all animals (10 per group) were measured weekly, and their volumes (in cubic millimeters) were determined with the formula 6/π × L × W × H. After 28 days, animals from all groups were anaesthetized and sacrificed by exsanguination.
Flow Cytometry of T lymphocytes, macrophages and granzyme
Immunofluorescence using monoclonal antibodies was used to determine the percentages of CD68+, CD4+, CD8+, CD4+/CD25+, CD8/CD25+, CD8+/CD69+, CD4+/CD69+ and granzyme+ cells in the peripheral blood samples, (Biolegend, USA). Briefly, 30 μl of blood were incubated for 30 min with 5 μl of the corresponding monoclonal antibody (1:100 dilution) afterwards, 200 μl of lysis solution were added (Becton Dickinson, California), incubated in darkness for 10 min and washed twice with 0.1 M PBS (pH 7.2), 0.1 % BSA and 0.1 % NaN3. The cells were then fixed in 1 % paraformaldehyde solution and stored at 4 °C until examination by flow cytometry (FACSCalibur, Becton Dickinson) using the Cell Quest software. 10,000 events in the region corresponding to lymphocytes were analyzed. From this region, the percentage of positive cells from each sample was determined. Results were expressed as means (±SD) for each experimental group.
Senescence assay
Cell senescence was assessed by the expression of β-galactosidase, by using a β-galactosidase staining kit (Cell Signaling, Danvers, MA). Cells (5x103) were plated in triplicate into 24-well culture plates (Corning, NY, USA) and then fixed on day 6 of culture for subsequent β-galactosidase staining.
Immunoblot
Cell proteins were extracted in RIPA buffer (1 % Nonidet P-40, 0.5 % sodium deoxycholate, 0.1 % SDS in PBS) in the presence of 40 μg/ml of aprotinin, 10 μg/ml PMSF and 100 mM orthovanadate. 40 μg of total protein were separated by 8 % or 12 % SDS-PAGE and transferred to nitrocellulose membranes. Western blot were developed with antibodies against p-p53 (Cell signaling, Danvers, MA), p-H2AX (Millipore, MA, USA), p53 (cell signaling, Danvers, MA), pan-Ras-V12 (Calbiochem, MA, USA), p-Rb (BD Biosciences, San Jose, CA, USA), or α-tubulin (Sigma-Aldrich, St Louis, MO).
Phenotypic analysis (by Immunofluorescence)
Cells were growth in 8-well polystyrene chambers (BD Falcon, San Jose, CA, USA) until 80 % confluence, fixed with 4 % paraformaldehyde in PBS, blocked with 1 % BSA in PBS, and labeled overnight at 4 °C with anti-GFAP antibody (Millipore, MA, USA) followed by Alexa fluor-594-conjugated anti-rabbit IgG (Invitrogen). Cells were mounted in Vectashield (Vector, CA, USA) and analyzed by confocal microscopy (LSMS Pascal, Zeiss).
Phenotypic analysis (by Flow cytometry)
Cells were suspended in blocking buffer (0.5 % BSA/ 2 mM EDTA, in PBS) and then labeled with fluorochrome-conjugated antibodies: PE-conjugated anti-Fas (BD Biosciences, San Jose, CA, USA), PE-conjugated anti-FasL (Biolegend, San Diego, CA) (1 μg/ml), or anti-MHC-I (Biolegend, San Diego, CA), anti-RAE-1δ (Biolegend, San Diego, CA), anti-Rae-1αβγδε (scbt, CA, USA) (1:50), followed by APC-conjugated anti-mouse IgG (scbt, CA, USA) or APC-conjugated anti-rabbit IgG (scbt, CA, USA) (1:50) secondary antibodies, as appropriate. Cell membrane expression of these molecules was assessed by Flow Cytometry (FACS Aria III, BD Biosciences). Raw data was further analyzed by using Flow Jo software (Tree Star, Inc. Ashland OR).
NK cytotoxicity assays
NK cells were purified from mice spleens (C57 strain), hosted in the INNN animal house, in accordance with institutional guidelines, by using the NK cell isolation kit II (Myltenyi Biotech, Germany), following the manufacturer”s instructions. NK cell cytotoxicity against tumor cells was evaluated by using Lecoeur et al. method [
37]. Since the tumor cells here used express green fluorescent protein (GFP) due to the transformation procedure, there was no need to label them. Isolated NK cells and tumor cells were co-cultured in 10 % FBS/DMEM at 1:10 target/effector cell ratio, for 4 h. After which, cells were labeled for 15 min with 7-aminoactinomycin D (BD Pharmingen, San Jose, CA, USA) at a final concentration of 20 μl/ml. Flow cytometry analysis was used to calculate the percentage of green fluorescent cells (tumor cells) that were stained by 7-AAD (dead cells). Results are expressed as the percentage of specific lysis calculated by the following formula:
$$ \%\ \mathrm{specific}\ \mathrm{lysis} = 100 \times \left(\%\kern0.5em \mathrm{sample}\ \mathrm{lysis} - \%\ \mathrm{basal}\ \mathrm{lysis}\right)/100 - \%\ \mathrm{basal}\ \mathrm{lysis} $$
Quantitative polymerase chain reaction
Total RNA was extracted via the phenol/chloroform method using TRIzol reagent (Invitrogen). Quantitative polymerase chain reaction (PCR) was performed using EXPRESS One-Step SYBR® GreenER™ Kit (Invitrogen, USA). Emissions from the SYBR Green reporter dye were monitored with an ABI Prism 7500 Real Time PCR (Applied Biosystems). The primer sequences used were as follows:
Mult1, 5′-CAATGTCTCTGTCCTCGGAA-3′ (sense), Mult1, 5′-CTGAACACGTCTCAGGCACT-3′ (antisense); H60a, 5′-TGCCTGATTCTGAGCCTTTTCA-3′ (sense), H60a, 5′-ATTCACTGAGCACTGTCCATGTAGAT-3′ (antisense); H60b, 5′-AGCCTTTTGGTCCTGCTGAAT-3′ (sense), H60b, 5′-ATGTTTTTTATCACCAAAATCAAGGAGT-3′ (antisense); H60c, 5′-CTTCTCTTGATCCTGGAGTCCTGTAGT-3′ (sense), H60c, 5′-GAGAGTCTTTCCATTCACTGAGCAC-3′ (antisense); β-actin, 5′-TTCTACAATGAGCTGCGT-3′ (sense), β-actin, 5′-ATCACAATGCCTGTGGTA-3′ (antisense). All expression levels of interested genes were normalized to the housekeeping gene β-actin. Gene expression values were then calculated based on the ΔΔCt method.
Statistical analyses
Data was summarized as arithmetic means and standard deviations (SD). One-way analysis of variance (ANOVA) and post-hoc (Tukey) test were conducted. Statistical significance was set at p < 0.05 in a two-sided test. SPSS software package V 18.0 for Windows; (SPSS Inc., Chicago, IL) was employed for data analysis.
Competing interests
The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
Authors’ contributions
Conception and design: MOM; JACP; FJSG; BP. Collection and assembly of data: MOM, IGC; NHP, VPdlC, SMJ Data analysis and interpretation: MOM; JACP; FJSG; JS; BP. Manuscript writing: MOM; FJSG and BP. All authors read and approved the final manuscript.