original article
© The American Society of Gene Therapy
Acid Sphingomyelinase Overexpression Enhances the Antineoplastic Effects of Irradiation In Vitro and In Vivo Eric L Smith1 and Edward H Schuchman1 Department of Genetics and Genomic Sciences, Mount Sinai School of Medicine, New York, New York, USA
1
Exposure of cells or animals to stress frequently induces acid sphingomyelinase (ASM)-mediated ceramide production that leads to cell death. Consistent with this, overexpression of ASM in subcutaneous B16-F10 mouse melanomas, in combination with irradiation, resulted in tumors that were up to 12-fold smaller than irradiated control melanomas. Similarly, when irradiated melanomas were pretreated with a single, peritumoral injection of recombinant ASM (rhASM), the tumors were up to threefold smaller. The in vivo effect of ASM was likely due to enhanced cell death of the tumor cells themselves, as well as the surrounding microvascular endothelial cells. In vitro, rhASM had little or no effect on the growth of tumor cells, even in combination with irradiation. However, when the culture media was acidified to mimic the acidic microenvironment of solid tumors, rhASMmediated cell death was markedly enhanced when combined with irradiation. Microscopic analysis suggested that this was associated with an increase in autophagy. rhASM has been produced for the treatment of the lysosomal storage disorder, type B Niemann–Pick disease, and is currently being evaluated in a phase-1 clinical trial. Based on the data presented in this article, we propose that further investigation of this protein and gene as antineoplastic agents also is warranted. Received 30 April 2008; accepted 12 June 2008; published online 15 July 2008. doi:10.1038/mt.2008.145
Introduction Acid sphingomyelinase (ASM; EC 3.1.4.12) hydrolyzes the membrane lipid, sphingomyelin, to phosphorylcholine and the bioactive lipid, ceramide. Mutations in the ASM gene (SMPD1) lead to the type A and B forms of the lysosomal storage disorder, Niemann–Pick disease (NPD). In addition to its housekeeping function in lysosomes, ASM has an important role in ceramidemediated cell death. For example, transformed lymphocytes from NPD patients are resistant to cell death caused by irradiation.1,2 Similarly, angiogenic endothelial cells from ASM knockout mice exhibited enhanced survival after irradiation,3 and ASM knockout animals are resistant to levels of whole-body irradiation that cause
“gastrointestinal syndrome” (endothelial cell death, followed by lethal bleeding into the gastrointestinal track).4 In addition to irradiation, the cellular response to other stress factors, including chemotherapeutic drugs,5–7 has also been shown to involve ASM. ASM participates in cell death by moving from intracellular compartments to the outer leaflet of the cell membrane after stress.8,9 There it converts sphingomyelin to ceramide, which is thought to be a critical process in the cell death response.10 The link between cancer, ASM, and ceramide has been extensively studied. For example, decreased ceramide has been reported in astrocytomas, ovarian, lung, and other cancers.11–14 High-grade gliomas have even been shown to have lower levels of ceramide than low-grade gliomas.12 The gene encoding acid ceramidase (ASAH1), an enzyme that degrades ceramide, is overexpressed in 70% of head and neck squamous cell carcinomas,15 as well as 60% of prostate cancers.16 Its transfection into DU145 prostate cancer cells caused enhanced proliferation, migration, and chemotherapeutic resistance, and resulted in larger subcutaneous tumors.17 It was also shown that shunting ceramide to glucosylceramide via glucosylceramide synthase caused MCF-7 breast cancer cells to become resistant to doxorubicin,18 and inhibiting this pathway recovered doxorubicin sensitivity.19 Furthermore, an analysis of the microarray database, Oncomine (http://www.oncomine.org), revealed that 13/110 matched cancer vs. normal tissue comparisons underexpressed the ASM mRNA (P < 0.0005), suggesting that downregulation of ASM expression may also be responsible for low levels of ceramide in some cancers. Additionally, irradiation of a radiosensitive, but not a matched, radioresistant head and neck squamous cell carcinoma cell line induced ASM externalization and activation followed by lipid raft microdomain formation.20 Thus, increasing ASM at the membrane of cancer cells during irradiation or exposure to other chemotherapy drugs may be a viable mechanism to enhance their sensitivity. Recombinant ASM (rhASM) has been manufactured for clinical use and is currently being evaluated in clinical trials for the treatment of type B NPD. It is nontoxic after administration into normal mice at high doses,21 creating a potentially large therapeutic window. Notably, solid tumors are known to have modified microenvironments where the extracellular pH can be as low as 5.8, and frequently is in the range of 6.5–6.9.22 Melanomas, in particular, have been shown to create a microenvironment with a pH as low as 6.4.23
Correspondence: Edward H. Schuchman, Department of Genetics & Genomic Sciences, Mount Sinai School of Medicine, 1425 Madison Avenue, Room 14-20A, New York, New York 10029, USA. E-mail:
[email protected] Molecular Therapy vol. 16 no. 9, 1565–1571 sep. 2008
1565
© The American Society of Gene Therapy
Antineoplastic Effects of Acid Sphingomyelinase
a
b 2,400
CHO-HT CHO-18
Percent cell viability
100 80 60 * 40 20
* 0
0
2
Gy
5
20
Figure 1 Effect of ASM overexpression on the viability of Chinese hamster ovary (CHO) cells after ionizing irradiation. Parental (CHO-HT) and ASM overexpressing (CHO-18) cells were exposed to 0-, 2-, 5-, or 20-Gy irradiation. Cell viability was measured after 72 hours (*P < 0.01).
1566
ASM overexpressing Control infected
15 CONT B-1
ASM C-1 10
5
Percent tumor growth
1,400
***
1,200 1,000 800 600 400 200
CONT B-1
ASM C-1
c
d 1,000
ASM overexp + 15 Gy Control + 15 Gy ASM overexp Control
900 800 700 600 500 400 300 200
110
2,200
20
× 104 RLU
ASM overexpression or rhASM injection sensitizes subcutaneous mouse melanomas to irradiation in vivo To determine whether overexpression of ASM had an effect on tumors in vivo, B16-F10 mouse melanoma cells were transduced with a retroviral vector to stably overexpress the human ASM complementary DNA (cDNA). ASM activity was enhanced 11-fold in these cells (Supplementary Figure S1), activity did not diminish after 20 passages, and did not affect proliferation in vitro as compared with B16-F10 cells infected with a control vector. Consistent with those results, 4 days after mice were inoculated with B16-F10 cells (transduced with ASM or control vectors), there were no differences in tumor size. However, 2 weeks after irradiation, in vivo molecular imaging showed that ASM overexpression markedly slowed tumor growth, and in some cases decreased tumor volumes (Figure 2a). Quantification of the bioluminescence revealed that, on average, tumors overexpressing ASM were 6.5-fold smaller than control tumors (P < 0.001; Figure 2b). Caliper volume measurements found ASM overexpressing tumors to be 12.4-fold
Day of irradiation
To purify rhASM, an overexpressing cell line, Chinese hamster ovary (CHO)-18 was derived.24 ASM activity from CHO-18 cell lysates was 70-fold higher than that from parental (CHO-HT) lysates, and activity in the media was elevated >700-fold.24 We exposed these two cell lines to irradiation to determine the effect of ASM overexpression on cell viability. CHO-18 cells were 39% more sensitive to 5-Gy irradiation, and 60% more sensitive to 20-Gy irradiation than the CHO-HT parental cells (both P < 0.01; Figure 1). Although the lack of ASM has been shown numerous times to protect cells or mice from irradiation-induced cell death and other stresses, this data is the first to correlate ASM overexpression with sensitization to irradiation, suggesting its possible role as an antineoplastic agent.
2 weeks postirradiation
Results ASM overexpression sensitizes Chinese hamster ovary cells to irradiation
smaller then control tumors (P < 0.05; Figure 2c). The size difference of the ASM overexpressing tumors can be easily appreciated in gross images (Figure 2d). rhASM is currently being evaluated in clinical trials for the treatment of NPD. To determine whether, in principle, this enzyme could also be used as an antineoplastic drug, we administered a single, peritumoral injection of rhASM in the area surrounding B16-F10 subcutaneous melanomas. Within 1–2 hours of rhASM injection, the tumors were exposed to 15-Gy linear irradiation and tumor development was monitored. Despite the fact that ASM exposure to the tumors was transient, and that the dose delivered could not be easily controlled due to release into underlying tissue, a single injection of exogenous rhASM in this manner also sensitized tumors to irradiation as determined by caliper volume measurements or in vivo molecular imaging 2 weeks postirradiation (3.1 and 1.9-fold smaller, respectively; both P < 0.05; Figure 3a). This is the first time that ASM has been shown to sensitize tumors in vivo, as well as the first time that the recombinant protein has been used for this purpose.
Percent tumor growth
Thus, the enzyme’s acidic pH optimum may favor its activity at tumor cells and surrounding tissues. For these reasons, ASM is a logical choice to evaluate as an antineoplastic therapeutic agent.
*
100 0
0
4
*
**
8 12 Days postirradiation
**
* 16
Control
ASM
Figure 2 The effect of ASM overexpression in B16-F10 melanoma cells on subcutaneous tumor growth after linear irradiation in vivo. Mice were injected with 5 × 105 B16-F10 cells either infected with a retroviral vector expressing the ASM complementary DNA or vector alone. Tumors were allowed to grow for 4 days and then given a single dose of 15-Gy linear irradiation. (a) Bioluminescent images showing difference in tumor size from the same mice at the day of irradiation, and 2 weeks after, irradiation of ASM overexpressing (ASM C-1) or control (CONT B-1) subcutaneous B16-F10 melanomas. (b) Scatter plot displaying percent tumor growth 2 weeks after irradiation as quantified by relative bioluminescent light units/second (***P < 0.001; each mark represents one tumor). (c) Mean percent change in tumor volume over time as quantified by caliper volume measurement. Sixteen days after irradiation, ASM overexpressing tumors were 12.4-fold smaller then control tumors (*P < 0.05; **P < 0.01). (d) Representative gross images of ASM overexpressing or control tumors displaying the size of melanomas and visible vasculature 2 weeks after linear irradiation.
www.moleculartherapy.org vol. 16 no. 9 sep. 2008
© The American Society of Gene Therapy
a
2,000 1,800
Antineoplastic Effects of Acid Sphingomyelinase
b
Vehicle INJ + 15 Gy ASM INJ + 15 Gy
Percent cell viability
Percent tumor growth
1,400 1,200 *
1,000 800 *
400
pH = 7.4 pH = 6.5
*
100
1,600
600
120
**
80 60 40 20
200 Caliper
c
IVMI pH = 7.4 pH = 6.5
d
−
+
−
+
−
+
2 Gy
−
+
−
+
4 Gy *
1.6 1.4
20 15
***
10
Red:Green ratio
25 Colony number
+
0 Gy
30
5
1.2 1 0.8 0.6 0.4 0.2
BSA
0
ASM
e
25
Percnet autophagic cells
0
−
0
0
20
4 Gy-BSA- 4 Gy-ASM- 4 Gy-BSA- 4 Gy-ASM- No Irr-BSA- No Irr-ASMpH 7.4 pH 7.4 pH 6.5 pH 6.5 pH 6.5 pH 6.5
**
15 10 5 0 4 Gy-BSApH 7.4
4 Gy-ASMpH 7.4
4 Gy-BSApH 6.5
4 Gy-ASMpH 6.5
No Irr-BSApH 6.5
No Irr-ASMpH 6.5
Figure 3 rhASM sensitizes B16-F10 melanomas to irradiation in vivo and B16-F10 cells in vitro under acidic conditions. (a) Quantification of percent tumor growth in vivo 2 weeks after a single rhASM (20 mg/kg) or vehicle injection preceding 15-Gy linear irradiation and in vivo molecular imaging (IVMI) (*P < 0.05). (b) Media on B16-F10 cells was replaced with new physiologic (pH 7.4) or acidified (pH 6.5) media. Five µmol/l rhASM (+) or BSA (−) was added. Cells were irradiated at 0-, 2-, or 4-Gy, and incubated for 2 hours. The media was then replaced with rhASM-free, physiologic (pH 7.4) media. Forty-eight hours later cell viability was measured (*P < 0.05; **P < 0.01). (c) For colony formation assays, sparsely plated cells were treated as in b and exposed to 2-Gy irradiation (***P < 0.001). (d) Cells were treated as in b. Twenty-four hours after treatment, acridine orange was used to stain DNA green and acidic compartments red. The ratio of red:green fluorescence is shown (*P < 0.05). (e) Cells were transfected with the GFP-LC3 complementary DNA and 16 hours later treated as in b. Twenty-four hours after treatment cells were analyzed for autophagy. (**P < 0.01, calculated by χ2-analysis). Standard error bars are shown.
The effect of pH on rhASM sensitization Next, we sought to investigate the apparent paradox of how a nontoxic, housekeeping enzyme can work synergistically with irradiation to kill tumors. We hypothesized that the acidic pH optimum of ASM may allow it to act preferentially in the unique, acidic microenvironment that exists in vivo at the site of solid tumors. As Molecular Therapy vol. 16 no. 9 sep. 2008
can be seen in Figure 3b, rhASM did not decrease the viability of B16-F10 cells in physiologic media (pH 7.4), even when combined with irradiation. In contrast, B16-F10 cells incubated for 2 hours in acidic media (pH 6.5) containing rhASM had a large decrease in viability following 2- and 4-Gy irradiation, as compared with cells incubated with bovine serum albumin (BSA) (24%, P < 0.05; and 1567
© The American Society of Gene Therapy
Antineoplastic Effects of Acid Sphingomyelinase
1568
Day of irradiation
a
ASM
2 weeks postirradiation
CONT
CONT
ASM
c
40
Percent secreted ASM activity
b
35
Vessel number
30 25 20 15 10
*
5 0
Control
ASM
Control
Day of irradiation
1,600
*
1,400 1,200 1,000 800 600 400 200 0
ASM
B16-F10CONT
2 weeks Postirradiation
d 120
BSA
B16-F10ASM
ASM
100 Percent cell viability
41%, P < 0.01, respectively). We also tested the effect of rhASM on colony formation of B16-F10 cells after 2-Gy irradiation under similar conditions, and found that at pH 6.5, but not 7.4, the addition of rhASM inhibited the number of colonies formed by over 55% (P < 0.001; Figure 3c). This difference in colony number was consistent over time (Supplementary Figure S2). To further test the hypothesis that rhASM has increased activity on B16-F10 cells in acidic vs. physiologic media, we examined changes in the cellular ceramide content after treatment. Cells were treated as in Figure 3b and irradiated at 8 Gy. After a 1-hour incubation at physiologic pH, irradiated cells with rhASM in the media had a 23% increase in ceramide above BSA-treated cells (P < 0.05); notably, however, at acidic pH, rhASM treatment elevated the level of ceramide in irradiated B16-F10 cells by 60% above BSA controls (P < 0.01), a 2.3-fold greater elevation compared with cells treated at physiologic pH (Supplementary Figure S3). The mechanism leading to ceramide-mediated cell death, and, in particular, the role of ASM in the ceramide-mediated irradiation response has been investigated extensively and is the subject of numerous reviews.25,26 Consistent with these prior reports, our data clearly shows that treatment of cancer cells with rhASM induces ceramide and decreases cell viability after irradiation, and that this effect is markedly enhanced in an acidic environment (pH 6.5). ASM’s ability to sensitize cells to irradiation was not specific to the B16-F10 mouse melanoma line. As previously mentioned, we identified 12 microarray studies where the ASM mRNA was significantly downregulated in cancer. Of these, the only tumor type to be replicated by multiple, independent studies was renal cell carcinoma, which was shown to underexpress ASM in three studies.27–29 For this reason, we used the human renal cell carcinoma line, CAKI-1, to examine whether rhASM could effect cell viability in combination with irradiation in a different cell line. We observed that similar to the B16-F10 cells, CAKI-1 cells incubated in acidic media showed a larger decrease in viability following 2- and 4-Gy irradiation (19 and 32% decrease, respectively, P
