SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Anti-apoptotic proteins that block death receptor-mediated apoptosis favour tumour evasion of the immune system, leading to enhanced tumour progression. However, it is unclear whether blocking the mitochondrial pathway of apoptosis will protect tumours from immune cell attack. Here, we report that the anti-apoptotic protein Bcl-xL, known for its ability to block the mitochondrial pathway of apoptosis, exerted tumour-progressive activity in a murine lymphoma model. Bcl-xL overexpressing tumours exhibited a more aggressive development than control tumours. Surprisingly, Bcl-xL protection of tumours from NK cell-mediated attack did not involve protection from NK cell-mediated cytotoxicity. Instead, Bcl-xL-blocked apoptosis resulting from hypoxia and/or nutrient loss associated with the inhibition of angiogenesis caused by NK cell-secreted IFN-γ. These results support the notion that NK cells may inhibit tumour growth also by mechanisms other than direct cytotoxicity. Hence, the present results unravel a pathway by which tumours with a block in the mitochondrial pathway of apoptosis can evade the immune system.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Inhibition of apoptosis and immune evasion are two important characteristics of the multistep process that takes place during tumorigenesis. We have previously demonstrated that the inhibition of death receptor-mediated apoptosis, induced by both T cells and NK cells against tumours, lead to tumour progression [1, 2]. Whether inhibition of the mitochondrial apoptotic machinery plays a role in the escape of tumours from immune attack has not been characterized. There are many examples of human cancers in which the mitochondrial apoptotic machinery is disrupted. Examples include the loss of apoptosis-activating factor 1 (Apaf-1), a key factor in mitochondria-mediated apoptosis [3], or the upregulation of survivin, a member of the inducer of apoptosis family, that blocks downstream caspases [4]. Another inhibitor of apoptosis that strongly correlates with malignancy is the B cell lymphoma-2 protein (Bcl-2), a protein involved in approximately 80% of patients with human follicular B cell lymphoma [5]. It is acknowledged that Bcl-2 acts as an oncogene, but how it mediates tumour progression is not completely understood [6]. The B cell lymphoma-extra-large protein (Bcl-xL), the closest relative to the Bcl-2 protein, also interferes with the apoptotic machinery at the mitochondrial level [7-9]. One indication that Bcl family members may interfere with immune cell activity towards tumour cells is their ability to block Granzyme B-mediated apoptosis [10]. Bcl-xL and Bcl-2 also inhibit Fas-mediated apoptosis although this inhibition seems to be cell-type dependent [11]. Bcl-xL also has the capacity to inhibit hypoxia-induced apoptosis [12].

NK cells have the potential to eliminate tumour cells that fail to express normal levels of MHC class I molecules [13-15] or that express stress-induced activation receptor ligands such as MIC or Rae1 [16]. NK cells are well known for their strong, granule-mediated, cytotoxic function and may prevent progression of tumours. In addition to perforin-mediated cytotoxicity, they also may induce apoptosis in target cells using death receptor ligands [1, 17, 18]. NK cells are also potent producers of pro-inflammatory cytokines such as IFN-γ [19].

Here, we set out to test whether blocking the mitochondrial apoptotic machinery by overexpression of Bcl-xL in a NK cell-sensitive tumour, RMA-S [14], would lead to immune evasion and affect its tumorigenicity in vivo. We investigated the mechanism behind the observed tumour-promoting action of Bcl-xL and concluded that it is linked to angiostatic effects mediated by NK cell-derived IFN-γ.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References
Ethics statement

Animal experiments were approved by the Stockholm animal ethical committee north (Stockholms Norra Djurförsöksetiska Nämnd). Approval ID number: N220/01.

Cell lines and mice

The mouse T lymphoma cell lines RMA and RMA-S and the human retroviral packaging cell line Phœnix-Ampho (www.stanford.edu/group/nolan/retroviral_systems/phx.html) were grown as described [2, 20, 21]. RMA-S cells have a mutation in the TAP-2 gene that impairs normal peptide trafficking to the ER, resulting in the expression of low levels of unstable MHC class I molecules [14, 15]. Sex- and age-matched (4–6 weeks old) inbred C57BL/6 mice were obtained from Charles River (Uppsala, Sweden). Fas ligand-mutant mice FasL−/− (gld) bred in the C57BL/6 background were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). Perforin-deficient C57BL/6 mice [22], RAG-2, common γ-chain-deficient C57BL/6 mice (Rag-2−/−γc−/−) [23] and IFN-γ-deficient C57BL/6 mice [24] were obtained from the Karolinska Institutet. Mice were maintained in the animal facility at Stockholm University and at the Karolinska Institutet.

Expression vectors and cell transduction

Human Bcl-xL was excised from the pSFVNeo-Bcl-xL expression vector (kind gift of C. Thompson, Abramson Family Cancer Research Institute and Department of Cancer Biology, University of Pennsylvania, Philadelphia, PA, USA) and inserted into the EcoRI site of the retroviral expression vector pLXIN (CLONTECH, Palo Alto, CA, USA). Vectors were then separately used to transiently transfect the Phœnix-Ampho packaging cell line (kindly provided by G.P. Nolan, Stanford University, Stanford, CA, USA). Supernatants containing recombinant viral particles were used to transduce RMA-S cells, and stable G418-resistant cell lines were obtained. mRNA expression was verified by RT-PCR, and the presence of helper virus was excluded by PCR amplification of viral env using the primers 5′-ACCTGGAGAGTCACCAACC-3′ and 5′-TACTTTGGAGAG GTCGTAGC-3′.

Flow cytometry

Cells were washed twice in PBS containing 2% FCS and 0.1% NaN3, and incubated with specific mAbs in the presence of anti-CD32/CD16 (Fcblock™; BD Biosciences, San Jose, CA, USA). PI was added to discriminate dead from living cells. Cells were analysed in a FACSCalibur (Becton Dickinson, San Jose, CA, USA), gating on the living cells.

Apoptosis assay

Sensitivity to mitochondria-mediated apoptosis was assessed by treating 5 × 105 cells with 1–100 nm of thapsigargin (Sigma, St. Louis, MO, USA) for 24 h at 37 °C. Sensitivity to Fas-induced apoptosis was assessed by treating 5 × 105 cells with mouse soluble FasL (sFasL) at a 1:2 dilution (FasL-hCD8α/pSG5 vector, kind gift of H. Yagita, Juntendo University, Tokyo, Japan) for 24 h at 37 °C. Apoptosis was monitored by flow cytometry analysing cell permeability to propidium iodide and the expression of phosphatidylserine on the cell surface.

Generation of NK cells

IL-2-activated NK cells were generated by isolating NK cells from splenocytes using anti-DX5 antibody linked to MACS beads (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer's recommendations. DX5+ cells were then cultured in LAK medium: αMEM containing 10 mm HEPES, 5 × 10−5 m 2-ME and 10% FCS (all reagents from Life Technologies, Paisley, UK) supplemented with 1 × 103 U/ml human rIL-2 (PeproTech Inc., Rocky Hill, NJ, USA) and cultured at a concentration of 5 × 105 cells/ml in 25-cm2 tissue culture flasks in 10% CO2 at 37 °C. After 5 or 6 days, IL-2-activated NK cells were resuspended by pipetting and scraping and used for cytotoxicity assays.

Cytotoxicity assay

NK cells were washed once, resuspended and used as effectors in standard 4- or 18 h 51Cr-release assays. Briefly, target tumour cells were labelled with 10–20 μl 10 mCi/ml Na[51Cr]O4 for 1 h at 37 °C and then washed. Cells (5 × 103) were incubated with titrated numbers of effector cells in round-bottom 96-well plates for 4 or 18 h at 37 °C in 10% CO2. In some experiments, anti-FasL mAb was added to the culture at 10 μg/ml. After incubation, released radioactivity was measured, and specific lysis was calculated according to the following formula: % specific release = [(experimental release − spontaneous release)/(maximum release − spontaneous release)] × 100.

In vivo rapid elimination assay

Tumour cells were labelled with 10–20 μl 10 mCi/ml Na[51Cr]O4 in 500 μl of LAK medium for 1 h at 37 °C. After washing, the cells were resuspended in PBS. Cells (1 × 106) in 200 μl were injected i.v. to mice. Twelve hours after injection, the mice were sacrificed and lungs were excised. The radioactivity in the lungs was measured, and the remaining radioactivity was calculated as follows: [(cpm lung − cpm background)/(cpm of cells injected − cpm background)].

Tumour engraftment assays

Groups of five to eight mice were injected s.c. in the interscapular region with different amounts of RMA-S transduced cells. As previously shown, the percentage of tumour take is dose-dependent, indicating that greater numbers of RMA-S cells are more difficult for NK cells to reject [1]. Mice were monitored every second day by palpation for 8 weeks. The tumour mass was measured with calipers, and tumour volume was calculated according to the formula length × width2 × 0.52 [25]. Mice were killed with CO2 when tumours reached a size of approximately 1 cm3, as recommended by the Stockholm ethical committee for animal experiments, or when the experiment was terminated. Tumour samples were obtained by surgical excision and used for ex vivo assays.

NK cell depletion

In experiments performed with NK cell-depleted mice, animals were injected i.p. with 100 μg purified anti-NK1.1 monoclonal antibody (PK136) 2 days prior to tumour grafting and then, 2, 7, and 12 days after tumour challenge.

Adoptive transfer assay

Freshly isolated splenocytes were enriched for NK cells using anti-DX5 mAb linked to MACS beads (Miltenyi Biotec) according to the manufacturer's recommendations. DX5+ cells (1.5 × 106) were then injected i.v. to groups of three age- and sex-matched Rag-2−/−γc−/− mice. Thirteen days after cell engraftment, 100 μg of poly I:C (Sigma) was administered by i.p. injection to all animals, including those that had not received any NK cells. The following day, the animals were injected s.c. in the interscapular region with 1 × 105 Bcl-xL-expressing RMA-S cells. Tumour development was measured by palpation. After 14 days, tumour samples were obtained by surgical excision and used for ex vivo assays. Concurrently with tumour excision, the spleens of the animals were removed and FACS analysis was performed to assess the number of NK1.1 (PK136)-positive NK cells.

Angiogenesis inhibition assay

Mice were injected s.c. in the interscapular region with 1 × 106 RMA-S-transduced cells. Shortly after tumour grafting, 20 mg/kg of TNP-470 (a kind gift of Takeda Chemicals, Tokyo, Japan) was injected s.c. in the flank. Treatment was repeated every second day, followed by tumour palpation and measurement.

Immunohistochemistry

Tumours of similar size were fixed with 4% formaldehyde in PBS for 24 h at 4 °C and subsequently embedded in paraffin. Tumour sections (6 μm thick) were processed and stained with biotinylated anti-CD31 mAb (MEC 13.3; BD Biosciences). Peroxidase activity was detected as described [26]. CD31-positive cells were counted at 40× magnification. For each tumour, six to eight areas were counted in perinecrotic areas of the tumour.

Statistic evaluation

P-values were calculated by the use of a two-tailed Student's t-test.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Bcl-xL protects RMA-S tumour cells from NK cell-mediated tumour rejection in vivo

To investigate whether Bcl-xL overexpression could confer an advantage for tumour growth in vivo, we overexpressed Bcl-xL in the murine lymphoma cell line RMA-S. We then transplanted these transfected RMA-S cells and RMA-S control cells, transduced with an empty vector (control), s.c. to mice. RMA-S lacks expression of MHC class I, rendering them sensitive to NK cell activity [1, 14, 15]. As shown previously, when a small number (102) of RMA-S cells were inoculated s.c., they were rejected in an NK cell-dependent manner (Fig. 1A,B). However, when RMA-S cells overexpressing Bcl-xL were inoculated, 27% of the mice developed tumours (Fig. 1A). The few RMA-S control tumours that started to grow developed with delayed kinetics and did not grow as aggressively as the Bcl-xL overexpressing tumours (Fig. 1A and data not shown). When NK cells were removed by injections of depleting monoclonal antibodies, 80–100% of the mice developed tumours regardless of the Bcl-xL overexpression (Fig. 1B). When a greater number of tumour cells (105) were injected, both control RMA-S cells and Bcl-xL overexpressing RMA-S cells gave rise to tumours. Under these conditions, Bcl-xL overexpressing RMA-S cells, however, formed tumours more rapidly than control RMA-S cells (Fig. 1C). Again, in the absence of NK cells, both control and Bcl-xL overexpressing RMA-S tumours grew at nearly identical rates, indicating that Bcl-xL actively protected the tumour cells from NK cell-mediated rejection rather than providing an NK cell-independent survival advantage (Fig. 1D).

image

Figure 1. Effect of Bcl-xL overexpression on RMA-S tumor growth in vivo. (A) Untreated C57BL/6 mice and (B) NK cell- depleted C57BL/6 mice were injected s.c. in the interscapular region with 102 RMA-S tumour cells. Tumour occurrence was monitored for 8 week. The percentage of mice developing tumours when injected with control-transduced control RMA-S cells (thin line) is compared with that of mice injected with Bcl-xL expressing RMA-S cells (thick line). Number of mice is shown in parenthesis in the figures. (C) Untreated C57BL/6 mice (n = 10) and (D) NK cell-depleted C57BL/6 mice (n = 10) were injected with 105 tumour cells and tumour size was monitored and calculated as described in 'Materials and methods'. Control RMA-S cells are represented by open squares and Bcl-xL overexpressing RMA-S by solid circles.

Download figure to PowerPoint

Bcl-xL does not protect against NK cell-mediated cytotoxicity in vitro or in vivo

We have previously shown that RMA-S cells are sensitive to NK cell cytotoxicity-mediated through both FasL and perforin and that overexpression of FLIPL protects RMA-S cells against Fas-induced cell death in the absence of perforin [1, 2, 18]. We thus investigated whether Bcl-xL mediated its tumour-promoting effects by protecting the tumour cells against the direct cytotoxic function of NK cells. To test whether Bcl-xL overexpression could protect against perforin-dependent or FasL-dependent NK cell-mediated cytotoxicity in vitro, we performed cytotoxic assays using IL-2-activated NK cells lacking either of these effector pathways. Bcl-xL was unable to protect RMA-S cells against these direct cytotoxic functions of the NK cells (Fig. 2A,B). As we have shown previously, perforin-independent cytotoxicity requires longer incubation time in vitro, which is why we, in Fig. 2A, incubated the cells for 18 h instead of the usual 4–6 h in the 51Cr release assay. To test the ability of Bcl-xL to protect against in vivo cytotoxicity and rapid rejection of the RMA-S cells, we injected 51Cr-labelled RMA-S cells i.v. and measured the amount of remaining radioactivity in the lungs of the mice after 12 h. In this assay, a high proportion of the injected tumour cells are trapped in the capillary beds of the lung, and the amount of remaining radioactivity reflects killing of RMA-S cells by lung-resident NK cell in an inversely proportional manner [14, 27]. We have previously shown that the rejection of RMA-S cells in the lung can occur by both perforin-dependent and perforin-independent pathways [1]. As a control tumour cell line, which is rejected to a lesser degree in the lung, we used the NK cell-insensitive RMA tumour cell line (from which RMA-S was derived). Again, no protective effect of Bcl-xL overexpression was seen against the rejection of the RMA-S cells in the lungs of either FasL- or perforin-deficient mice (Fig. 2C). Taken together, these data indicate that it is unlikely that the observed tumour-promoting effect of Bcl-xL was due to the protection from NK cell cytotoxicity.

image

Figure 2. Effect of Bcl-xL overexpression on FasL or perforin-mediated NK cell cytotoxicity in vivo and in vitro. (A) In vitro cytotoxicity was assessed by means of specific lysis in a 51Cr-release assay, as described in 'Materials and methods'. NK cells from perforin-deficient mice were mixed with 51Cr-labelled control (open squares)- or Bcl-xL (solid circles)-transduced RMA-S cells and 51Cr-release was measured after 18 h. (B) NK cells from FasL-deficient mice were mixed with 51Cr-labelled control (open squares)- or Bcl-xL (solid circles)-transduced RMA-S cells, and lysis was measured after 4 h. (C) For in vivo assessment of NK activity, 5 × 105 51Cr-labelled RMA, control- and Bcl-xL-transduced RMA-S cells were injected i.v. to mice (n = 5/group), and cytotoxicity was measured as described in 'Materials and methods'. Results shown are representative of three experiments. Error bars denote standard deviation.

Download figure to PowerPoint

NK cells elicit IFN-γ-dependent angiostatic effects in RMA-S tumours

The observation that Bcl-xL could protect tumours from NK cell-mediated rejection despite its inability to block their cytotoxic functions led us to hypothesize that NK cells, in the early stage of tumor growth, may also mediate indirect effects related to the tumour microenvironment. One factor that is crucial for tumour growth is angiogenesis. It has been shown that IFN-γ can inhibit angiogenesis [28] and that NK cells are potent producers of IFN-γ when activated [19]. We thus investigated the role of IFN-γ in the rejection of RMA-S tumours. Interestingly, Bcl-xL overexpressing RMA-S tumours grew even more aggressively in IFN-γ-deficient mice than in wild-type mice. When injecting the mice with a high dose of 104 Bcl-xL overexpressing RMA-S cells, the tumour penetrance in IFN-γ-deficient or NK-depleted wild-type mice was 100% in both cases, whereas tumour penetrance in wild-type mice was 33% (Fig. 3A).

image

Figure 3. Effect of NK cell produced IFN-γ on RMA-S tumour growth and vascularization. (A) 104 Bcl-xL-transduced cells were injected to untreated wild-type (solid thick line), NK-depleted (solid thin line) and IFN-γ-deficient (dashed line) animals, and tumour occurrence was monitored for 8 week. The percentage of mice developing tumours is shown, and the number of mice is indicated within brackets in the figure. (B) 104 control-transduced control RMA-S cells were injected s.c. in the interscapular region of untreated C57BL/6 mice (filled bar) or NK cell-depleted C57BL/6 (open bar). Neovascularization was measured as means of CD31+ cell count in tumour sections, as described in 'Materials and methods'. (C) 104 Bcl-xL-transduced RMA-S cells were injected to Rag-2−/−γc−/− animals reconstituted with IFN-γ K.O. (filled bar), wild-type (open bar) or non-reconstituted (grey bar). Neovascularization was measured as means of CD31+ cell count in tumour sections, as described in 'Materials and methods'. (D), The spleen of the tumour-bearing mice was excised concurrently with tumours analysed in (C), and the percentage of NK1.1+ CD3 cells was assessed by flow cytometry. Error bars denote standard deviation. **P > 0.01.

Download figure to PowerPoint

Next, we studied the degree of vascularization in control RMA-S tumours by determining the CD31+ cell count in tumour sections. The tumours growing in NK cell-depleted mice showed a significantly higher degree of tumour neo-vascularization compared with those growing in non-depleted mice, indicating that NK cells impaired vessel growth in the tumour environment (Fig. 3B). To determine whether NK cell-derived IFN-γ was involved in the reduction in angiogenesis, Rag-2−/−γc−/− mice (lacking NK cells) were reconstituted with wild-type or IFN-γ-deficient NK cells. Fourteen days later, 105 Bcl-xL-expressing RMA-S cells were injected to the NK cell-reconstituted mice or, as a control, to non-reconstituted Rag-2−/−γc−/− mice. This higher dose of tumour cells was used to increase the number of mice with wild-type NK cells that would develop control RMA-S tumours. After an additional 14 days, the tumours were excised and tumour neo-vascularization was examined. Tumours obtained from Rag-2−/−γc−/− mice reconstituted with wild-type NK cells showed significantly fewer CD31+ cells than tumours from non-reconstituted mice. Rag-2−/−γc−/− mice reconstituted with IFN-γ-deficient NK cells had a similar degree of vascularization as non-reconstituted mice (Fig. 3C). The degree of NK cell reconstitution was also analysed, and the number of NK cells in spleens was similar when comparing mice reconstituted with IFN-γ-deficient and wild-type NK cells (Fig. 3D). These results indicate that IFN-γ released by NK cells was responsible for the angiostatic effect observed in growing RMA-S tumours in the presence of NK cells.

Bcl-xL protects tumours from the angiostatic effects of TNP-470

To further strengthen the notion that Bcl-xL enabled the RMA-S cells to form tumours under conditions of limited neo-vascularization, we investigated the tumour-protective role of Bcl-xL on angiostatic treatment of the RMA-S tumours. The synthetic analogue to fumagillin, TNP-470, a potent angiostatic agent, was used for the treatment [29, 30]. TNP-470 binds to and irreversibly inactivates -methionine aminopeptidase-2 (MetAP2), resulting in the endothelial cell cycle arrest late in the G1 phase and inhibition of tumour angiogenesis. TNP-470 may also induce the p53 pathway, thereby stimulating the production of cyclin-dependent kinase inhibitor p21 and inhibiting angiogenesis. Tumours were established by injecting a large number (106) of RMA-S cells. TNP-470 was then injected i.p., every second day, to inhibit angiogenesis in the growing tumours. TNP-470 was able to delay the growth of the control RMA-S tumours (Fig. 4A), but did not significantly delay the growth of the Bcl-xL-expressing tumours (Fig. 4B). Another representation of these data is shown in Fig. 4C, where the effects of TNP-470 on tumour growth of Bcl-xL-expressing RMA-S are compared with control RMA-S tumours, showing a strong protective effect of Bcl-xL against the effects of the drug. These results indicate that Bcl-xL protects the tumour cells in a tumour environment that is hypoxic and nutrient-deprived due to the lack of tumour neo-vascularization.

image

Figure 4. The effect of Bcl-xL overexpression on RMA-S tumour growth in vivo in the presence of the angiostatic compound TNP-470. C57BL/6 mice were injected s.c. in the interscapular region with 1 × 106 cells. TNP-470 was administered every second day by i.p. injection at a concentration of 20 mg/kg. Tumour occurrence was monitored for 4 week by palpation, and tumour size was measured and calculated as described in 'Materials and methods'. (A) Control-transduced control RMA-S cells were injected to untreated (open squares, n = 5) or TNP-470-treated mice (solid squares, n = 5), and mean tumour volume is presented in the figure. (B) Bcl-xL-expressing RMA-S cells were injected to untreated (open squares, n = 5) and TNP-470-treated mice (solid squares, n = 5), and mean tumour volume is presented in the figure. (C) Comparison of the mean tumour size (data from A and B) in mice treated with TNP-470 and injected with control RMA-S (open bars) or Bcl-xL -expressing RMA-S cells (filled bars). Error bars denote standard deviation. *P < 0.05, **P < 0.01, ***P < 0.001.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

Our results indicate that Bcl-xL, introduced into RMA-S cells, promote tumour progression by protecting the tumour cells from NK cell-mediated tumour suppression. We first addressed if Bcl-xL could protect the tumour cells from NK cell-mediated cytotoxicity. NK cells use both perforin-dependent and death receptor-dependent mechanisms for cytotoxicity [18]. Perforin enables the entry of several granzymes into target cells, which lead to the induction of both caspase-dependent and -independent apoptosis. For example, Granzyme B can directly cleave several death substrates, contributing to nuclear damage and cell death [31]. It has been previously observed that Bcl-2 prevents apoptosis induced by cell exposure to purified Granzyme B, but not that mediated by whole cytotoxic lymphocytes [10, 32]. With regard to death receptor-mediated cytotoxicity, the capacity of Bcl-2 or Bcl-xL to inhibit Fas-mediated apoptosis has been shown to be cell-type dependent [11]. In the present study, we were unable to observe any protective effects of Bcl-xL on perforin-mediated or FasL-mediated cytotoxicity in vivo or in vitro.

The observation that Bcl-xL could not protect RMA-S against NK cell-mediated cytotoxicity but still protected the tumour from NK cell inhibition of tumour formation suggested that NK cells affected tumour growth in vivo by mechanisms other than cytotoxicity. Neo-vascularization is required for tumour formation in vivo, and the observed effect of overexpressing Bcl-xL in the RMA-S is only evident in long-term tumour growth experiments in vivo. When investigating the effect of NK cells on the number of CD31+ cells in the RMA-S tumours, we found that NK cells reduced the number of blood vessels in the RMA-S tumours and that this was due to their production of IFN-γ. This is in agreement with previous reports showing an anti-angiogenic effect of IFN-γ [28, 33, 34]. Some of these studies suggest a direct effect of IFN-γ on endothelial cells [28, 34], whereas other studies suggest that the tumour cells need to be responsive to IFN-γ [35]. Also, IFN-γ-inducing cytokines exhibit anti-angiogenic properties [36, 37]. IFN-γ may not be the last player in the chain because an IFN-γ-induced protein, IP-10, also inhibits angiogenesis [38]. One mechanism through which IFN-γ inhibits angiogenesis is STAT-1 regulation of genes involved in angiogenesis in endothelial cells [39]. Published data suggest that certain immune cells can also stimulate angiogenesis under certain circumstances [40, 41]. Our data suggest that, although some cytokines produced by NK cells may have stimulatory effects on angiogenesis, the inhibitory effects of IFN-γ overrides these and the net-result is inhibition of neo-vascularization.

Bcl-xL protects cells from hypoxia-induced cell death [12, 42, 43]. Hypoxia and hypoxia-induced vascular endothelial growth factor also mediate upregulation of Bcl-2 [44, 45]. In our tumour model, Bcl-xL protected the tumour from the hypoxic and nutrient-poor condition generated by angiostatic TNP-470, similar to what was previously reported for Bcl-2 by Fernandez et al. [46]. Contrary to the work of Fernandez et al., we used immunocompetent animals and not the SCID model with transplanted human cells. In our model, we were able to study the role of Bcl-xL in the interaction between tumours and the immune system. In the light of the present data, and previously published data presented above, we suggest that the tumour-progressive activity of Bcl-xL in the present system was mediated through the inhibition of tumour cell apoptosis, induced by a lack of nutrients and/or oxygen due to lack of vascularization. This indicates that NK cells mediate tumour-suppressing activity not only through cytotoxicity, but also through preventing the establishment of a microenvironment favouring tumour growth. Tumours may then be selected to escape this by upregulating anti-apoptotic proteins of the Bcl-2 family to continue to grow, in spite of poor blood supply, and ultimately overwhelming the NK cell response and the host.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References

We thank M.-L. Solberg and M. Hagelin for their excellent technical assistance. We also thank Dr. C. Fluur for helpful advice.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgment
  8. References