TIMP-1-GPI in combination with hyperthermic treatment of melanoma increases sensitivity to FAS-mediated apoptosis

Abstract

Resistance to apoptosis is a prominent feature of malignant melanoma. Hyperthermic therapy can be an effective adjuvant treatment for some tumors including melanoma. We developed a fusion protein based on the tissue inhibitor of matrix metalloproteinase-1 linked to a glycosylphosphatidylinositol anchor (TIMP-1-GPI). The TIMP-1-GPI-fusion protein shows unique properties. Exogenous administration of TIMP-1-GPI can result in transient morphological changes to treated cells including modulation of proliferation and decreased resistance to apoptosis. The effect of TIMP-1-GPI on the biology of melanoma in the context of a defined hyperthermic dose was evaluated in vitro. Clonogenic assays were used to measure cell survival. Gelatinase zymography determined secretion of MMP-2 and MMP-9. Monoclonal antibody against FAS/CD95 was applied to induce apoptosis. The expression of pro- and anti-apoptotic proteins and the secretion of immunoregulatory cytokines were then evaluated using Western blot and ELISA. TIMP-1-GPI combined with a sub-lethal hyperthermic treatment (41.8°C for 2 h) suppressed tumor cell growth capacity as measured by clonogenic assay. The co-treatment also significantly suppressed tumor cell proliferation, enhanced FAS receptor surface expression increased tumor cell susceptibility to FAS-mediated killing. The increased sensitivity to FAS-induced apoptosis was linked to alterations in the apoptotic mediators Bcl-2, Bax, Bcl-XL and Apaf-1. The agent works in concert with sub-lethal hyperthermic treatment to render melanoma cells sensitive to FAS killing. The targeted delivery of TIMP-1-GPI to tumor environments in the context of regional hyperthermic therapy could be optimized through the use of thermosensitive liposomes.

Introduction

The matrix metalloproteinases (MMPs) represent a family of zinc-dependent endopeptidases that collectively degrade components of the extracellular matrix. MMPs have been implicated in tissue remodeling, tumor invasion, resistance to apoptosis and metastasis [1, 2]. MMP activity is regulated at many levels including through association with the four endogenous MMP inhibitors, the tissue inhibitor of matrix metalloproteinases (TIMP)-1, -2, -3 and -4 [3]. In vivo, the balance between MMP and TIMP activities determines whether matrix resorption or deposition occurs [4].

While TIMPs inhibit MMP activity and thus can inhibit tumor growth and metastasis, TIMPs are also multifunctional proteins shown to regulate cell proliferation, apoptosis and angiogenesis. TIMP-1 is a broadly acting MMP inhibitor [3]. It is a soluble protein that can sometimes be detected on the cell surface through its association with surface bound proteins [5, 6]. The overall role of TIMP-1 in cancer biology remains the subject of conflicting reports [7, 8]. TIMP-1 has been linked to angiogenesis, cell migration, proliferation, tumor growth, metastases and tumor prognosis [9, 8].

Proteins anchored by glycosylphosphatidylinositol (GPI) when purified and added to cells in vitro are efficiently incorporated into their surface membranes [10–12]. This protein engineering of cell surfaces using GPI-anchors can be used to study protein function [10–12]. We recently reported the generation of a fusion protein based on human TIMP-1 combined with a GPI-anchor (derived from LFA-3) to generate a recombinant reagent used to focus exogenously added TIMP-1 directly onto cell surfaces independently of protein–protein interactions [11–13]. TIMP-1-GPI treatment lead to a slowing of cancer cell growth in vitro and to an increase in the sensitivity of renal cell carcinoma cells (RCC) to FAS-induced apoptosis [12].

Melanoma is the most aggressive form of skin cancer and is notoriously resistant to current modalities of cancer therapy [14]. A large set of genetic, functional and biochemical studies suggest that melanoma cells become resistant to chemotherapeutic drugs by exploiting their intrinsic resistance to apoptosis and by reprogramming their proliferation and survival pathways during tumor progression [15, 16]. An important challenge today is the application of therapeutic strategies that are potent enough to compensate, bypass or modulate the cell death defects associated with melanoma to improve the prognosis of patients at late stages of the disease [17]. Initial results demonstrated that, unlike what was described for RCC [12], melanoma cell lines remained largely resistant to treatment with TIMP-1-GPI.

The application of hyperthermia has been demonstrated to be an effective modality for cancer therapy [18–20]. Hyperthermia is applied for two purposes: the direct induction of tumor cell death and the sensitization of cells to chemotherapeutics, apoptosis or anticancer treatments [21–23]. For melanoma, hyperthermic treatment has also been shown to affect the antigenicity of human melanoma cells, their recognition by cytotoxic lymphocytes and immune cell induced apoptosis [24].

As hyperthermic therapy is an adjunct treatment for melanoma [18], we sought to explore the effects of exogenously added TIMP-1-GPI protein in combination with thermal shock on the biology of melanoma cells in vitro. Hyperthermic treatment was found to synergize with TIMP-1-GPI to render melanoma cells sensitive to immune mediated apoptosis.

Material and methods

Tumor cell lines and cell culture

The human melanoma cell lines used (624.38-MEL, 93.04A12MEL, SK-MEL23, WM115, WM266-4) represent a spectrum covering diverse melanoma characteristics. 624.38-MEL is a clone of the bulk melanoma cell line 624 expressing high levels of HLA-A2 molecules on the cell surface [25]. WM115 (ESTDAB-066) and WM266-4 (ESTDAB-076) are human melanoma cell lines derived from primary tumor and from metastatic tumor, respectively. The cell line 93.04A12MEL was a gift (Peter Schnier, Leiden, Netherlands). SK-MEL23 was obtained from the ATCC and is melanotic while all other lines studied represent amelanotic tumors. Cells were incubated at 37°C in a humidified atmosphere of 5% CO₂ and 95% humidity, unless otherwise stated, in medium supplemented with 1% streptomycin sulfate and sodium penicillin G and 1% MEM non-essential amino acids and 10% FCS.

Purification of TIMP-1-GPI protein

The TIMP-1-GPI protein was produced and purified as previously described [11]. Briefly, human TIMP-1 was cloned from cDNA using hTIMP-1 specific primers, fused without a translation stop codon to the GPI-signal sequence cloned from human LFA-3 [10, 26] and subcloned into pEF-DHFR. The expression plasmid was stably introduced into DHFR deficient Chinese hamster ovary (CHO) cells and selected as described [27]. TIMP-1-GPI-fusion protein was purified from the CHO cells by Triton X-100H detergent extraction followed by column purification using DEAE, heparin sepharose and size exclusion [11]. Phosphatidylserine was routinely used as a GPI-anchor control (P5660, Sigma Chemical Co.).

Clonogenic assay for survival of heat-treated melanoma cells

The ability of tumor cells to form colonies after hyperthermia exposure was assessed using the clonogenic assay as described [28, 29]. The melanoma cells were exposed to 41.8 or 37°C for 2 h. The cells were then harvested by trypsinization and seeded in duplicate T25 tissue culture flasks at 200 and 500 cells per flask and allowed to form colonies in an undisturbed, humidified, 37°C/5% CO₂ air atmosphere. After 7 days, the flasks were washed with 0.9% NaCl solution and cell colonies were stained with crystal violet solution (20% ethanol, 0.8% ammonium oxalate and 2% crystal violet). Only colonies containing at least 50 cells were considered to be viable survivors. The number of colonies was counted and compared to the input cell numbers to obtain the percent survival.

Fluorescence-activated cell sorting (FACS) analysis

Cells were detached with 1.5 mM EDTA (Biochrom A, Berlin, Germany No. L2113) in 1 × PBS and incubated for 60 min on ice with antibodies specific for human TIMP-1 (IM32L) and IgG_1κ (isotype control is a mixture of murine IgG_κ, Sigma-Aldrich, Taufkirchen, Germany No. M9269). Cells were washed three times with 1 × PBS, incubated with FITC-conjugated donkey anti-mouse mAB (DAKO A/S, Glostrup, Denmark No. F0313) for 45 min on ice, then washed three times with 1 × PBS and analyzed using a flow cytometer (FACSCalibur, Becton Dickinson and Company, San Jose, CA, USA) and CellQuest software. Data are presented as geometric mean.

Incorporation of TIMP-1-GPI into cell membranes

Melanoma cells (5–10 × 10⁶ cells/ml) were incubated with 14 ng/ml of purified hTIMP-1-GPI at 37°C/5% CO₂. The cells were then washed three times with cold 1 × PBS and analyzed by FACS using human TIMP-1 specific monoclonal antibodies (see above).

Proliferation

Cell lines (30 × 10³/100 μl medium) were cultured in 96-well micro-titer plates for 24 h under standard conditions to yield firmly attached and stably growing cells. After discarding supernatants, 50 μl of medium containing TIMP-1-GPI, buffer, denatured TIMP-1-GPI, phosphatidylserine or rhTIMP-1 was added to the cells and incubated for 24–72 h. Additional plates were treated in parallel or in combination with hyperthermic stress for 2 h at 41.8°C. Then 50 μl of a 1 mg/ml solution of [(3,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide] MTT (Sigma-Aldrich, Taufkirchen, Germany No. M2128) was added. After 3-h incubation at 37°C, formazan crystals were dissolved by addition of 100 μl isopropanol and 0.04 N HCl. Absorbance was then measured at 550 nm using GENios plus TECAN ELISA reader. Cell proliferation was calculated as previously described [12]. For each experiment at least 6-wells were analyzed per experimental condition and time point. A minimum of six wells were analyzed per experimental condition, and time point, and are presented as mean values with standard deviation.

Zymography

Cell lines were cultured in 24-well plates (5 × 10⁴ cells/well) to subconfluency. The medium was exchanged for 24 h with serum-free medium containing either rhTIMP-1 or increasing amounts of TIMP-1-GPI and incubated for 24, 48 and 72 h. Cell supernatants were then analyzed by gelatin zymography using 10% SDS-polyacrylamide gels (Invitrogen, Groningen, Netherlands, No. EC61755BOX) as described [11]. Recombinant MMP-9 protein (Amersham Biosciences, Uppsala, Sweden, No. RPN2634) was used as positive signal on the gel.

FAS detection

FAS expression on the surface of melanoma lines was assessed by flow cytometry using a non-activating anti-FAS mAB (L-958 was a gift from H. Engelmann, Munich). To detect FAS expression, untreated cells and cells treated for 72 h with 14 ng/ml TIMP-1-GPI, denatured TIMP-1-GPI or rhTIMP-1 protein either with, or without hyperthermic exposure (2 h at 41.8°C) were -detached with 1.5 mM EDTA and stained with L-958 and analyzed by flow cytometry.

FAS-induced apoptosis and annexin-V-detection

Detection and quantification of apoptotic versus necrotic cells at the single cell level was performed using annexin-V-FLUOS staining Kit (Becton, Dickinson and Company, Heidelberg, Germany, No. 556547). Melanoma cells were seeded at 1 × 10⁶ cells/well into 24-well plates and allowed to attach over night. The wells were then rinsed three times with 1 × PBS and 1 ml of serum-free RPMI 1640 medium was added, followed by 14 ng/ml of TIMP-1-GPI, denatured TIMP-1-GPI or rhTIMP-1. Cells were incubated at 37°C/5% CO₂. After 72 h, 1 μg/ml anti-FAS activating mAB L-957 [30] or isotype control were added and the cells were further incubated for 24 h at 37°C/5% CO₂. The cells were then washed with 1 × PBS, gently pelleted and resuspended in staining solution [annexin-V-fluorescein labeling reagent and propidium iodide (PI) in Hepes buffer] for 15 min at room temperature. The cells were then analyzed by flow cytometry. A time-course study showed that annexin-V binding to melanoma cells precedes PI reactivity.

Western blot analysis

Western blot was used for the detection of Bcl-XL (Santa Cruz Biotechnology, INC, Santa Cruz, CA USA No. SC-7195) Bcl-2, (ALX-804-225), Bax (ANC-357-040), Apaf-1, (Stressgen, Ann Arbor, USA, No. AAP-300) and β-actin (Acris Hiddenhausen, Germany, No. ab8227). The anti-Hsp70 monoclonal rat IgG1 antibody 6B3 was a gift (E. Kremmer, Munich) [31]. The cells were washed three times with cold 1 × PBS and detached with 1.5 mM EDTA. The cells were rotated 1 h at 4°C with hypertonic lysis buffer [5 mM Tris, 2 mM MgCl, 0.1 mM EDTA, 1 mM PMSF (Roche, Basel, Switzerland No. 236608)], 2 μg/ml aprotinin (Roche, Basel, Switzerland No. 1583794) 2 μg/ml leupeptin (Sigma, Taufkirchen, Germany No. L2884), 1 μg/ml pepstatin A (Roche, Basel, Switzerland No.1524488), pH 7.4). Cell membranes were isolated with extraction buffer (100 mM NaCl, 1% Triton X-100, 10 mM Tris, 5 mM EDTA, 1 mM PMSF 2 μg/ml Aprotinin, 2 μg/ml leupeptin, 1 μg/ml pepstatin A, pH 7.4) followed by centrifugation at 11,000g for 20 min at 4°C. The supernatant was used for further Western blot tests.

Whole-cell lysates containing 40 μg protein, as determined by the Bradford method (BioRad, Munich, Germany), were denatured by boiling for 15 min in SDS sample buffer, loaded onto one lane and separated by 4–20% SDS-Page electrophoresis. After electrophoresis, proteins were transferred to PVDF membranes (Invitrogen, Groningen, Netherlands. No. LC2002) and probed with appropriate antibodies. Specific proteins were detected using a commercial Western blot analysis kit, Chemiluminescent Immunodetection System (Invitrogen, Groningen, Netherlands).

ELISA for TGF-β and IL-10 proteins

Melanoma cells were cultured in 24-well plate (5 × 10⁴ cells/well). The medium was exchanged for 24 h with serum-free medium containing either rhTIMP-1 or increasing TIMP-1-GPI. The cells were washed and a portion of the cells were exposed to 41.8°C for 2 h. Then the cells were again washed with 1 × PBS and incubated again in serum-free culture media for an additional 72 h at 37°C in a humidified 5% CO₂ atmosphere. The supernatant was harvested and used for ELISA. The TGF-β and IL-10 concentrations were measured using commercial ELISA kits from BD Biosciences (San Jose, CA USA, No. 559119 for TGF-β and No. 555157 for IL-10).

Results

Incorporation of exogenously added TIMP-1-GPI into the surface membrane of melanoma cell lines

GPI-anchored TIMP-1 protein was generated and isolated as previously described [11, 12]. The incorporation of purified GPI-anchored TIMP-1 protein into the surface membranes of melanoma cells was demonstrated using two exemplary cell lines, SK-MEL23 and WM266-4. The cell lines were incubated with 14 ng/ml of purified TIMP-1-GPI or recombinant human (rh)TIMP-1 control protein for one h at 37°C. Surface associated TIMP-1 protein was then detected using FACS and an anti-human TIMP-1 monoclonal antibody. Addition of control rhTIMP-1 did not lead to detectable TIMP-1 on the cell surface while GPI-anchored TIMP-1 resulted in a surface signal for TIMP-1 (Fig. 1). The GPI-anchored TIMP-1 protein could be efficiently cleaved from the surface of the melanoma cells following treatment with phospholipase C (data not shown) [11, 12].

Fig. 1.

TIMP-1-GPI incorporation into cell membranes of melanoma cells. To demonstrate incorporation of GPI-anchored TIMP-1 protein into cell membranes, purified TIMP-1-GPI or control rhTIMP-1 was added to native SK-MEL23 and WM 226-4 cells. TIMP-1 was detected on the cell surface by FACS analysis. Gray histograms represent isotype control stain, solid-line histograms represent the anti-human TIMP-1 antibody staining

The effect of TIMP-1-GPI treatment in conjunction with thermal stress on the survival and proliferation of melanoma cells

The survival capacity of cells in the context of heat treatment is commonly measured using a clonogenic assay [24]. The cells are exposed to a specific thermal dose represented by a defined period of time and temperature. We previously established the general thermal sensitivity of melanoma cell lines [24] and Milani et al. (unpublished data).

By clonogenic assay, a temperature of 41.8°C delivered for an interval of 2 h was found to represent a thermal dose that was sub-lethal or showed a low level of lethality for melanoma cell lines meaning that this thermal dose does not cause exponential death [24]. This also represents a thermal dose that is usually measured within the tumor during the clinical application of local hyperthermia [19, 20, 32]. For subsequent experiments this thermal dose was used as a basis to study the potential additive effects of TIMP-1-GPI treatment in the context of thermal treatment.

The cumulative effect of the sub-lethal thermal dose (41.8°C for 2 h) with TIMP-1-GPI treatment was then evaluated using SK-MEL23 and WM266-4. The cell lines were treated with 14 ng/ml TIMP-1-GPI for 1 h, washed, and subjected to either hyperthermic stress or physiologic temperature (37°C). This procedure for the combined effect of TIMP-1-GPI with hyperthermia was used in all subsequent experiments. In Fig. 2a and b the resultant survival data at 37 and 41.8°C in the context of TIMP-1-GPI and control rhTIMP-1 are depicted. The figures represent the number of surviving colonies from 200 melanoma cells following treatment. TIMP-1-GPI, but not rhTIMP-1 was found to reduce the clonogenic survival of SK-MEL23 melanoma cells, and to a lesser extent WM226-4. Hyperthermic stress in combination with TIMP-1-GPI further reduced the survival capacity of all melanoma cell lines tested.

Fig. 2.

Clonogenic assay was used to assess the survival and proliferation of hyperthermic/TIMP-1-GPI treated melanoma cells. Survival of melanoma cells after heat and TIMP-1-GPI treatment a SK-MEL23 and b WM 226-4 at 37 and 41.8°C for 2 h as described in Sect. “Materials and methods”. The surviving clones from two independent flasks were counted per experimental condition. The mean values with standard deviation are shown. The effect of increasing levels of TIMP-1-GPI or rhTIMP-1 control protein on the proliferation of c SK-MEL23 and d WM 226-4 was measured using MTT assay [24]. MTT was added after 24 h. While TIMP-1-GPI alone inhibited proliferation of the melanoma cells, hyperthermic treatment amplified the effect. rhTIMP-1-GPI did not affect proliferation of the melanoma cells. In each experiment at least six wells were analyzed per experimental condition and time point. The mean values with standard deviation are shown

We have previously shown that treatment with TIMP-1-GPI can either enhance or suppress cell proliferation depending on the specific cell type studied [11, 12]. MTT assays were performed to assess the effect of TIMP-1 surface engineering on the proliferation of melanoma cells. The exogenously added TIMP-1-GPI protein was found to elicit a dose-dependent decrease in proliferation of SK-MEL23 and WM226-4 at 24 h (Fig. 2c, d). In contrast, heat denatured TIMP-1-GPI and rhTIMP-1 had no effect on proliferation. Additional controls using an equal molar concentration of phosphoinositol did not alter proliferation of the cells (data not shown) [12]. The sublethal hyperthermic dose also reduced cell proliferation while the combined TIMP-1-GPI/hyperthermic treatment amplified the inhibition seen with TIMP-1-GPI alone.

TIMP-1-GPI protein blocks release of proMMP-2 and proMMP-9 from melanoma cells

Secretion of the gelatinases MMP-2 and MMP-9 is thought to play an important role in the general aggressiveness of tumors including melanoma [33, 34]. The refocusing of TIMP-1 functional domains directly onto the cell surface has been shown to influence the secretion of MMPs [11, 12]. Gelatinase zymography was used to measure the release of MMP-2 and MMP-9 into the culture medium from the melanoma cells. SK-MEL23 and WM226-4 cell lines constitutively secrete both proMMP-2 and proMMP-9 (Fig. 3). The effect of TIMP-1-GPI on the release of gelatinases was tested in response to two concentrations of TIMP-1-GPI (7, 14 ng/ml) and in the context of thermal stress (41.8°C for 2 h). As previously reported for renal cell carcinoma [12], treatment with TIMP-1-GPI effectively blocked the secretion of both gelantinases. Control rhTIMP-1 protein at 14 ng/ml had no effect on proMMP-2 or proMMP-9 secretion. Thermal treatment did not significantly influence the release of the gelatinases into the growth media with or without TIMP-1-GPI treatment.

Fig. 3.

TIMP-1-GPI inhibits the release of proMMP-2 and proMMP-9 from melanoma cells. Gelatinase zymography was used to measure the secretion of MMP-2 and MMP-9 from the WM226-4 and SK-MEL23 cell lines. The cells were treated with different concentrations of TIMP-1-GPI, or control rhTIMP-1, and after 48 h the serum-free culture supernatant was removed and analyzed by gelatinase zymography

TIMP-1-GPI and heat shock treatment renders melanoma sensitive to FAS-mediated killing

Melanoma is generally resistant to induced apoptosis [15–17]. The level of FAS surface expression on SK-MEL23, WM226-4, WM115, 624.38 MEL and 93.04A12 MEL cell lines was assessed by flow cytometry using a non-activating anti-FAS mAB (L-958). Untreated cells as well as cells treated with 14 ng/ml TIMP-1-GPI or rhTIMP-1 control protein for 72 h were stained with L-958 and analyzed for surface FAS expression. With the exception of the 624.38 MEL cell line, surface FAS-protein was expressed by the other four melanoma lines. Treatment with TIMP-1-GPI or rhTIMP-1 had little or no effect on cell surface FAS expression (Fig. 4a). Thermal treatment alone showed an effect on surface FAS expression. The combined treatment of TIMP-1-GPI and hyperthermic treatment resulted in an additive increase in FAS surface expression, but similar effects could also be seen with the rhTIMP-1 control protein in SK-MEL23 and 93.04A12-MEL (Fig. 4a).

Fig. 4.

TIMP-1-GPI in combination with heat shock renders melanoma cells sensitive to FAS-induced apoptosis. FAS expression on the melanoma cell lines SK-MEL23, WM-226, WM115, 624.38 MEL and 93.04A12MEL was assessed by flow cytometry using a non-activating anti-FAS mAB (L-958) (a). Untreated cells and cells treated with 14 ng/ml TIMP-1-GPI or 14 ng/ml rhTIMP-1 control protein for 72 h were then stained with L-958 and analyzed for the surface expression of FAS. FAS-induced apoptosis of the melanoma cell lines after treatment with L-957 for an additional 24 h was detected by flow cytometry for cell surface binding of annexin V-fluoroisothiocyanate (FITC) and incorporation of propidium iodide (PI): b MEL 624.38, c SK-MEL23, d 93.04A12MEL e WM266-4 and f WM115 cell lines

We have previously demonstrated that the treatment of renal cell carcinoma (RCC) with TIMP-1-GPI can render RCC more sensitive to FAS/CD95-ligation-mediated apoptosis [12]. The anti-FAS mAB L-957 can mimic CD95 induced apoptosis in FAS-expressing cells [30]. Staining with antibody directed against annexin V-fluoroisothiocyanate (FITC) and incorporation of propidium iodide (PI) into melanoma cells after treatment with L-957 for an additional 24 h was used to detect apoptosis using flow cytometry. The effect of thermal treatment in combination with TIMP-1-GPI on FAS-induced apoptosis was assessed in parallel. As demonstrated in Fig. 4b–f, in keeping with previous reports, the melanoma cells were found to be largely resistant to FAS-mediated apoptosis before TIMP-1-GPI or hyperthermic treatment. The apoptotic response of the various melanoma cell lines to TIMP-1-GPI or hyperthermic treatment alone was variable, and depended on the individual cell line tested. However, when the cell lines were treated by a combination of TIMP-1-GPI with thermal stress, the FAS inducing antibody (L-957) lead to a more than additive increase in sensitivity to FAS-mediated killing in each of the cell lines tested.

TIMP-1-GPI and thermal treatment of melanoma cell moderates the expression of apoptotic proteins

The BCL-2 family of proteins are involved in the regulation of apoptosis (reviewed in [35]). Apoptosis protease-activating factor-1 (Apaf-1) is a key regulator of the mitochondrial apoptotic pathway [36]. We have previously shown for renal cell carcinoma that TIMP-1-GPI can alter the expression of pro- and anti-apoptotic proteins [12]. The effect of TIMP-1-GPI and thermal treatment on the protein expression of Bcl-2, Bcl-XL, Bax and Apaf-1 was evaluated. Following a 72 h preincubation period with 14 ng/ml of TIMP-1-GPI or rhTIMP-1 control, with or without hyperthermic treatment, the two exemplary melanoma cell lines SK-MEL23 and WM226-4 were stimulated with 1 μg/ml L-957 (or control mAB) for an additional 24 h. The level of Bcl-2, Bcl-XL, Bax and Apaf-1 proteins were determined using Western blot and specific antibody reagents. In both cell lines, treatment with TIMP-1-GPI lead to an increase in expression of pro-apoptotic Bax and decreased expression of anti-apoptotic Bcl-2 and Bcl-XL proteins. Levels of Apaf-1 were largely unaffected by TIMP-1-GPI treatment (Fig. 5a, b). The effect of thermal dose augmented the general increase or decrease in Bcl-2 protein expression and increased Apaf-1 expression in combination with TIMP-1-GPI. Antibody directed against human Hsp70 verified an increase in protein in response to hyperthermic stress (Fig. 5a, b).

Fig. 5.

TIMP-1-GPI and hyperthermic treatment of melanoma cells moderates expression of proteins linked to apoptosis. The effect of TIMP-1-GPI treatment on the expression of Hsp70, Bcl-2, Bcl-XL, Bax and Apaf-1 was determined by Western blot using protein specific antibody reagents. Following a 72 h preincubation period with 14 ng/ml of TIMP-1-GPI or 14 ng/ml rhTIMP-1 control protein, with or without hyperthermic treatment, the melanoma cells were stimulated with 1 μg/ml L-957 (or control isotype antibody) for an additional 24 h. The level of Hsp70, Bcl-2, Bcl-XL, Bax and Apaf-1 protein was then determined in a SK-MEL23 and b WM226-4

Downregulation of active TGF-β and induced IL-10 secretion by combined TIMP-1-GPI and hyperthermic treatment

The cytokine TGF-β has been proposed to have a regulatory effect on melanoma biology [37]. TGF-β is processed from a latent to active form by the actions of various proteins including MMPs [38, 39]. To investigate the effect of TIMP-1-GPI and heat shock on the activation of TGF-β from the latent to active form, total TGF-β protein was compared to that of active TGF-β using specific ELISA on the culture supernatants from SK-MEL23 and WM226-4. Hyperthermic stress or TIMP-1-GPI treatment alone reduced the ratio of active TGF-β to total protein to some extent, however, the combined treatment strongly reduced expression suggesting a combined effect of TIMP-1-GPI and hyperthermic treatment on the cleavage of the pro form of the cytokine to the active protein (Fig. 6a, b).

Fig. 6.

Downregulation of TGF-β and induction of IL-10 by a combination of TIMP-1-GPI and hyperthermic treatment of melanoma cells. Specific ELISA was used to compare total TGF-β protein to that of active TGF-β. The active form of TGF-β was effectively reduced by treatment with TIMP-1-GPI and hyperthermia. a SK-MEL23 and b WM226-4. IL-10 specific ELISA assay was used to measure the release of IL-10 from the melanoma cells before or after single or combined TIMP-1-GPI/hyperthermic treatment in c SK-MEL23 and d WM226-4

The immunogenicity of melanoma has been linked to expression of the regulatory cytokine IL-10 [40]. Untreated melanoma cells were found to secrete little IL-10 and the secretion was largely unaffected by either hyperthermic or TIMP-1-GPI treatment. Interestingly, the two modalities appear to act in a more than additive manner to increase the secretion of IL-10 protein by the treated SK-MEL23 and WM226-4 cell lines (Fig. 6c, d).

Discussion

Melanoma is a highly malignant tumor of the skin characterized by resistance to common therapeutic modalities [15, 17, 24]. The relatively weak responsiveness of melanoma to chemotherapy and immunotherapy is based in part on a general resistance of the tumor to apoptosis. The engineering of cells by exogenous addition of GPI-TIMP-1 can elicit enhanced as well as novel TIMP-1 biologic activities [11, 12]. Exogenously administered TIMP-1-GPI efficiently inserted into the cell membranes of melanoma cells and in combination with sub-lethal thermal dose induced a variety of therapeutically relevant effects.

TIMP-1-GPI treatment alone, lead to a dose-dependent reduction in the clonogenic survival capacity of the melanoma cells. This effect was augmented by thermal stress. Hyperthermia increased the expression of FAS receptor in all the melanoma cell lines tested. Neither TIMP-1-GPI, nor sub-lethal levels of heat shock alone, efficiently increased sensitivity of the melanoma cell lines tested to FAS-mediated apoptosis. However, co-treatment with TIMP-1-GPI and hyperthermia rendered all cell lines tested sensitive to FAS-induced apoptosis.

The FAS-apoptosis pathway is regulated through caspase activation [35, 41]. Mitochondria are crucial regulators of apoptosis through several pathways and their function in apoptosis is tightly controlled by the actions of a set of regulatory proteins [35, 42]. The upstream events leading to caspase activation are controlled by the overall balance between pro- and anti-apoptotic BCL-2-family proteins [42]. The general inducibility of apoptosis in the treated melanoma cells was contrasted with the level of expression of pro- and anti-apoptotic BCL-2 proteins. The pro-apoptotic multidomain protein Bax and the anti-apoptotic Bcl-2 and Bcl-XL proteins showed moderate basal expression in each of the melanoma cell lines analyzed. GPI-TIMP-1 treatment in combination with heat shock resulted in reduced expression of Bcl-2 and Bcl-XL proteins and a corresponding increase in the Bax protein. This shift towards a higher concentration of pro-apoptotic proteins may be one reason for the increased sensitivity of the TIMP-1 surface engineered melanoma cells to FAS-mediated apoptosis. This effect was found to be augmented by sub-lethal hyperthermic treatment.

Apoptosis protease-activating factor-1 (Apaf-1) is also a key regulator of the mitochondrial apoptotic pathway, being the central element of the multimeric apoptosome formed by procaspase 9, cytochrome c, and Apaf-1 itself [43]. The role of Apaf-1 is generally described as pro-apoptotic. A reduced expression of Apaf-1 is thought to compromise the apoptotic response in melanoma cells [17]. Apaf-1 expression has been shown to inversely correlate with the chemosensitivity of melanoma and Apaf-1-negative melanomas are unable to execute a typical apoptotic program in response to p53 activation and are chemoresistant [17]. In both melanoma cell lines tested, treatment of the lines with TIMP-1-GPI in combination with thermal dose resulted in an increase in Apaf-1 expression. Treatment with TIMP-1-GPI in combination of heat shock alters the balance of a series of apoptosis regulating proteins to elicit a more “pro-apoptotic” expression profile. This effect appeared to parallel a dose-dependent suppression of proliferation of the melanoma cells.

Growth regulatory proteins of the transforming growth factor-beta family (TGF-β) are one of the few classes of endogenous inhibitors of cell growth [44]. Generally, an increase in the expression of TGF-β is associated with human cancers (in vivo) [45]. TGF-β can have a regulatory effect on melanoma invasion as well as the anti-tumor immune response [45]. TGF-β can suppress the development of effector activity in T cells through an inhibition of the expression of cytotoxins in cytotoxic T cells [40, 44, 46]. TGF-β is produced as a latent inactive form, which is cleaved to gain activity. The cleavage of TGF-β from latent to active forms is mediated at least in part by MMPs [38, 39]. The treatment of cells with the TIMP-1-GPI or hyperthermic stress alone did not appear to significantly influence cleavage to the mature protein, however, the combined effect of TIMP-1-GPI with thermal stress lead to a pronounced reduction in the generation of the active form of the protein.

IL-10 can enhance the effector function of natural killer cells resulting in increased tumor destruction and thus increased availability of antigen [47]. The combination of heat treatment with TIMP-1-GPI also led to a strong increase in the secretion of IL-10. At the same time, IL-10 has been shown to slow the maturation of dendritic cells thus allowing them to more completely sample the antigen environment before undergoing maturation and trafficking to secondary lymph nodes. In addition, IL-10 can act as a co-stimulatory factor on effector T cells to increase Th1 cytokine expression and to promote the maintenance of CD8 + effector function [48, 49]. Taken together, increasing IL-10 and reducing TGF-β may help to positively condition the tumor environment to allow better effector cell activity as well as increased antigen acquisition and presentation at the tumor site.

The results presented here suggest that the membrane-fixed form of TIMP-1 represents an effective version of the protein for use in therapeutic applications directed against melanoma. The data appears to contrast recent publications suggesting a general anti-apoptotic effect of TIMP-1 [7]. It is important to note that in our studies the TIMP-1-GPI is not over-expressed by the target cell, but rather added to engineer the proteolytic environment of the cell surface. In addition, the agent appears to have unique properties not previously associated with TIMP-1 [11, 12]. Based on the chemistry of the GPI-anchor, the TIMP-1 fusion protein is ideally suited to be efficiently incorporated into liposomes. Through the use of thermally sensitive liposomes, the perfusion of TIMP-1-GPI loaded liposomes in the context of regional hyperthermic therapy would allow the targeted delivery and release of the agent into the tumor environment [50–52]. The 41.8°C temperature used in this study is optimal for vesicle “release” and represents a temperature commonly achieved during the clinical application of hyperthermic treatment [19, 20, 32, 51, 53–55]. In the experiments detailed here, 37°C was used for the controlled incorporation of the GPI-anchored TIMP-1 protein, followed by treatment with hyperthermia or control temperatures. Between 50–75% of the input TIMP-1-GPI is incorporated in cell surfaces at this temperature. The efficiency of incorporation is temperature dependent [11, 12]. Released at the tumor site, TIMP-1-GPI may then synergize with the hyperthermic therapy to reduce tumor growth over multiple mechanisms, including direct effects such as inhibition of tumor cell proliferation and increased sensitivity to apoptosis, as well as indirect effects by releasing immune effector cells from inhibition and fostering their cytotoxic activity through reduction of active TGF-β and increase in IL-10, respectively.