PDF Author's personal copy - NJU

(1)

Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies are encouraged to visit:

http://www.elsevier.com/copyright

(2)

Pathway of programmed cell death in HeLa cells induced by polymeric anti-cancer drugs

Yan-Qing Guan

^a

, Zhibin Li

^a

, Jiamei Chen

^a

, Huimin Tao

^a

, Wenwen Wang

^a

, Zhe Zheng

^a

, Ling Li

^a

, Jun-Ming Liu

^b^,^c^,^d^,^*

aSchool of Life Science, South China Normal University, Guangzhou 510613, China

bSchool of Physics, South China Normal University, Guangzhou 510613, China

cLaboratory of Solid State Microstructures, Nanjing University, Nanjing 210093, China

dInternational Center for Materials Physics, Chinese Academy of Sciences, Shenyang 110016, China

a r t i c l e i n f o

Article history:

Received 13 January 2011 Accepted 20 January 2011 Available online 12 February 2011

Keywords:

Co-immobilized IFN-gplus TNF-a Polystyrene material

HeLa EndoG

a b s t r a c t

Synthesis of anticancer polymeric materials plus their biological applications is one of the most charming and active research areas in biological functional materials. However, the predominant mechanisms for controlling cancer cell viability are not yet clear. In this work, cell culture polymeric materials co-immobilized with death signal proteins interferon-g(IFN-g)/tumor necrosis factor-a(TNF-a) on the surface were prepared by photochemical method to develop an anticancer polymeric drug model. Various characterizations on the microstructures and compositions, including the Fourier transform infrared spectroscopy, UV absorption spectroscopy,fluorescence measurement, atomic force microscopy, and electron spectroscopy for chemical analysis, were performed. For addressing the biological applications, we investigated systematically the death pathways of HeLa cells attached onto the drug model by means of a series of cell- biology techniques. It was demonstrated that the IFN-gplus TNF-aco-immobilized on the polymeric material surface exhibited more notable inhibitive effects than the free IFN-gplus TNF-a, and the induced HeLa cells were mainly along apoptosis-like PCD with the translocation of EndoG from the cytoplasm to the nucleus. Thesefindings indicate that the polymeric drugs with the co-immobilized IFN-gplus TNF-amay offer significant potentials for therapeutic manipulation of human cervical cancer.

1. Introduction

Biomaterials have demonstrated great promise for their extensive application potentials in cell-culture engineering, artificial organs and multifunctional nanoparticle drugs. Approaches to control the interactions at the interface between living tissues or living cells and biomaterials can be classified as chemical modifications and surface biolizations. The chemical modifications include the surface hydrophilization or surface hydrophobilization of polymer films by chemical or plasma treatments, while the surface biolizations address the immobilization of anticoagulant or hybridization of endothelial cells for synthesizing blood-compat- ible biomaterials and realize the immobilization of cell-adhesion peptide for tissue generation. In addition, it has been demonstrated that immobilized biosignal proteins can control cell behaviors such

as growth, differentiation, secretion, mobility, and programmed cell death (PCD, extra for cancer cells)[1e5]. In this regard, synthesis of biomaterials with immobilized biosignal proteins as anti-cancer drugs becomes speciﬁcally attractive.

Along this line, one needs to address two key issues. One is the synthesis technique for biomaterials with effective functionality against cancer cells, and the other is a choice of applicable proteins that can effectively inhibit the growth of cancer cells. For the former issue, biomaterial surface modiﬁcation is a powerful tool to provide surfaces that can control cell behaviors. A number of peptides, enzymes, growth factors, proteins etc, can be integrated onto the surface of polymeric biomedical materials by using various techniques such as photochemical immobilization, in order to produce biomedical materials with a variety of activities and functionalities [1,2]. These photo-immobilized factors were demonstrated to have longer activities than those dissoluble factors[3e5].

For the latter issue, we need to comprehend the mechanism for the PCD of cancer cells induced by biomaterials. The central role of cell mitochondria in mediating PCD has been receiving special attentions, due to their key implications in many pathways essential to both the

*Corresponding author. Laboratory of Solid State Microstructures, Nanjing University, Hankou Road no. 22, Nanjing 210093, China.

E-mail address:[email protected](J.-M. Liu).

Contents lists available atScienceDirect

Biomaterials

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / b i o m a t e r i a l s

doi:10.1016/j.biomaterials.2011.01.060

Biomaterials 32 (2011) 3637e3646

(3)

life and death of cells[6]. Mitochondrial function can be altered through a variety of signals, ultimately leading to the opening of mitochondrial permeability transition pore that is modulated by the members of Bcl-2 gene family [7]. Subsequently, pro-apoptotic proteins, such as apoptosis-inducing factor (AIF) and/or endonuclease G (EndoG), are released from the mitochondria. This phenomenon is another kind of PCD pathway that is called the apoptosis-like PCD pathway where the chromatin condenses at low degree with the DNA fragmenting into the macromolecules of 30w50 kDa[8].

Motivated by the above cell-biology mechanism, one notes that tumor necrosis factor-a(TNF-a) is a cytokine with direct and strong anti-tumor effect on the human body. In parallel, interferon-g(IFN-g) can be also used in the adjuvant therapy of clinical cancer. The synergy mechanism of IFN-gplus TNF-alies in that the TNF-areceptors can inhibit its own expression while IFN-gcan enhance the tumor inhibition of TNF-aand the expression of TNF-areceptors (at 2e3 times), without affecting the afﬁnity constant of IFN-g [9]. TNF-a could enhance the cell cycle arrest induced by IFN-g in HeLa cells, suggesting that TNF-aplays a major role in a short period and IFN-g plays an auxiliary role[10]. For endothelial cell injury induced by IFN- gplus TNF-a[11], the cathepsin B is involved in this cell death process and two cell death pathways are activated, which cannot be inhibited by zVAD.fmk. Simultaneously, IFN-gcan promote the HEK293 cell death rate induced by TNF-a, but this induction can be completely inhibited by zVAD.fmk. These results suggest different cell death pathways upon different cell lines induced by the synergy of IFN-g plus TNF-a.

Keeping in mind the above two key issues, it is of signiﬁcance to synthesize a polymeric material taking IFN-gplus TNF-ain an effort to develop a biomedical anticancer drug. One is challenged with the polymeric drug synthesis in one hand, on the other hand and perhaps more substantially a comprehensive capturing of the effect of the polymeric drug against cancer cells in terms of their growth and PCD etc is required. In fact, a photo-immobilization of IFN-g plus TNF-aonto the polystyrene material surface as a preliminary step toward the polymeric drug was developed in our earlier works [12,13]. While it is necessary to optimize this technique and present evident data on synthesis of the polymeric drug in terms of chemical composition and microstructures, more effort should be devoted to subsequent biomedical characterizations in order to understand the mechanisms with which the prepared drug offers substantial inhibition of tumor growth.

In fact, no work on the PCD upon implication of this polymer drug material was available although there have been many reports on the synergistic inhibition of tumor growth with TNF-aor IFN-g [14e19]. We demonstrated that the co-immobilized TNF-a plus IFN-gdoes cause the inhibition of receptor protein TNFR1, activation of receptor protein IFNR2, and thus the caspase-independent PCD in HeLa cells. The nuclear chromatin of the most cells mildly condenses into theﬂoccular structures, like apoptosis-like PCD[13].

In this paper, we report our substantial effort in synthesizing a polymeric material co-immobilized with TNF-aplus IFN-g and investigating the down-stream biomedical effects in utilizing this drug to attack tumor cell growth (HeLa). We paid particular attention to the apoptosis-like PCD in HeLa cells induced by this drug, and performed extensive experiments on the cell morphology, mortalities, loss ofDJm, expression of p53, Bax, Bcl-2 and caspases in protein level, the expression of AIF and EndoG in mRNA and protein level upon the drug treatment for 24 h. Since the caspase-independent PCD seems to be a new predominant mechanism for controlling cancer cell viability in the absence of apoptosis or in coordination with apoptosis and has less been touched in literature on polymeric anticancer drugs, exploration of EndoG-activation or inhibition in the polymer drugs would be especially critical, which will be performed here.

Actually, due to the problem of drug resistance (some cancer cells can escape apoptosis) remains to be a major obstacle in chemotherapeutic treatment, and also because the underlying caspase-independent PCD mechanism for high anti-cervical cancer functionality of the co-immobilized TNF-aplus IFN-gis the core issue for any success of such polymeric anticancer drugs, the present work which addresses the down-stream PCD mechanism in a comprehensive manner becomes more important.

2. Materials and methods

2.1. Preparation of AzPhIFN-gand AzPhTNF-a

The materials preparation was carried out in the procedure (shown in the Supplement Materials Fig. S1). For identifying the as-generated AzPhIFN-g and AzPhTNF-afrom the pure IFN-gand TNF-a, the Fourier transform infrared spectroscopy (FTIR) (TENSOR27, Bruker, Germany) was performed to probe the characteristic peaks, as shown inFig. 1. In addition, the obtained AzPhIFN-gand AzPhTNF-awere characterized by the UV (UV2450, SHIMADZU, Japan) absorption spectroscopy andﬂuores- cence (Perkin Elmer LS55, USA) measurement, respectively, as shown inFig. 2.

2.2. Immobilization of AzPhIFN-gand/or AzPhTNF-a

The as-prepared AzPhIFN-gand/or AzPhTNF-awere/was added into 24-well cell culture polystyrene plates at a dose of 20 ng/well to prepare the immobilized IFN-g and/or TNF-a. It should be mentioned that we prepared two co-immobilized samples with different IFN-g/TNF-a ratios, e.g. (1:1) and (1:2). The surface morphology and molecular structure of the as-prepared plated samples were investigated by electron spectroscopy for chemical analysis (ESCA LAB 250, Thermo Fisher Scientiﬁc, USA) (Table 1) equipped with AlKaat 1486.6eV and 150W power at the anode, and atomic force microscopy (AFM, SPA-300HV, SIINT, Japan) (Fig. 3). For the ESCA analysis, the survey scan spectra and c-1s core level scan spectra were also obtained and the probed peak height offers the atomic sensitivity.

2.3. Cell morphology, mortalities and mitochondrial membrane depolarization (DJm) analysis

The human cervical carcinoma cells (HeLa cells) (110⁵/well) were seeded in the 24-well polystyrene plates with IFN-gand/or TNF-ain various forms and combinations for 24 h. All the cells were maintained at 37C in the 5% CO2ambient.

First, the morphology and inner structure of the HeLa cells were characterized by transmission electron microscopy (TEM, Philips EM400), as shown inFig. 4A. Second, the HeLa cells were stained according to conventional Annexin V binding protocol provided by Oncogene, followed by an immediate analysis of the cell mortalities by flow cytometry (Becton Dickinson, FACSCalibur, San Jose, CA) using Cell Quest Program, as shown inFig. 4B. Third, the loss ofDJm was monitored with the dye 5, 5, 6, 6-tetrachloro-1, 1, 3, 3-tetraethyl -benzimidazolyl- carbocyanine iodide (JC-1). The ratio between green and redfluorescences provides an estimatedDJm that is independent of the mitochondrial mass. HeLa cells were analyzed using theflow cytometry (Becton Dickinson, FACSCalibur, San Jose, CA), as shown inFig. 4C.

2.4. Immunohistochemistry analysis

The HeLa cells (110⁵/well) were seeded in 24-well TNF-aplus IFN-g-co- immobilized (1:1) cell culture polystyrene plates and 24-well cell culture polystyrene plates treated only with free TNF-aplus IFN-g, for the indicated time periods (12, 24 and 48 h), respectively.

For the immunohistochemistry (IHC) analysis, the HeLa cells were collected into an Eppendorf tube of 1.5 ml and put in 10% neutral buffered formalin for 12 h. Then the cells were stained with eosin for 1 min. Three micrometer-thick sections were cut onto positive-charged slides and used for IHC detection of p53, Bax and Bcl-2.

Briefly, after the deparaffinization in xylene and rehydration in graded alcohols, these sections were microwaved in 0.01Mcitrate buffer (pH 6.0) for 15 min. The Endogenous peroxidase activity was blocked by immersion in 0.03% H2O2in methyl alcohol for 15 min, where 10% normal rabbit (for mouse primary antibodies) or goat serum (for rabbit primary antibodies) was applied to avoid non-specific reaction if any. These sections were then incubated with primary antibodies for 2e3 h. After washing with PBS, the biotinylated anti-muse or rabbit IgG was applied for 20 min at room temperature. The peroxidase-conjugated-streptavidin solution was applied for 30 min and visualized using 0.05% 3⁰-3⁰diaminobenzidine (DAB). The counter- staining was performed with hematoxylin.

In subsequence, the breast cancer cell known to be positive for p53, Bax and Bcl- 2 overexpression was used as a positive control. For a negative control, the primary antibody was replaced by a non-speciﬁc negative control antibody.

Finally, the average optic density of reaction products was tested by image analysis. The immunohistochemistry analysis was shown inFig. 5.

(4)

2.5. Western blotting analysis

The HeLa cells were lysed in extraction buffer (10 mMTris(pH7.4),150 mMNaCl,1%

Triton x-100, 5 mMEDTA(pH8.0)). The protein samples were separated by SDS-PAGE (12%) and electro-transferred onto a NC (nitrocellulose) membrane (Boster Biotech- nology Co., Ltd., China). The expression of proteins was performed using antibodies to

caspase-9, caspase-8, caspase-3, AIF and EndoG (Boster Biotechnology Co., Ltd., China). The blots were incubated with appropriate secondary antibodies conjugated to alkaline phosphatase (AP) peroxidase (Boster Biological Technology Co., Ltd., China) and developed using ALP reagent (Beyotime Institute of Biotechnology, China).

The protein levels were normalized by re-probing the blots with antibody tob-actin (Boster Biological Technology Co., Ltd., China), as shown in Figs.6and7D.

Fig. 1.Measured Fourier transform infrared spectroscopy spectra of IFN-g, AzphIFN-g, and 4-Azidobenzic acid (A); TNF-a, AzphTNF-a, and 4-Azidobenzic acid (B). The concentration of all the samples is 0.01%.

Fig. 2.Measured UV spectra (A) andﬂuorescence spectra (B) of IFN-g, TNF-a, AzphIFN-g, AzphTNF-a, and 4-Azidobenzic acid, respectively. The excitation wavelength is 341 nm. The concentration of all the samples is 0.01% and no apparent absorbance in the UV spectra for IFN-gand TNF-aover the measured wavelength range.

Y.-Q. Guan et al. / Biomaterials 32 (2011) 3637e3646 3639

(5)

2.6. Real-timeﬂuorescent quantitative PCR and PCR analysis

The relative level of apoptosis-inducing factor (AIF) or EndoG mRNA in HeLa was examined by RTPCR (SYBR Green) (ABI 3900, High-Throughput DNA/RNA Synthe- sizer; ABI 9700, PCR instrument; ABI 7500,ﬂuorescent quantitative PCR instrument;

Applied Biosystems, USA), as shown inFig. 7A and B. The HeLa cells were treated by IFN-gand/or TNF-ain various forms and combinations for 24 h. The following primers of AIF, EndoG designed with Primer Express 2.0 Software were used for the PCR step: (1) AIF, EndoG sense, 5⁰-CGGCAGGAAGGTAGAAACTGA-3⁰, 5⁰-CCAGAA TGCCTGGAACAACC-3⁰; (2) antisense, 5⁰-CGGAAGCCACCAAAATCTGA-3⁰, 5⁰-TCAGCC TCTGTCCTGGGC-3⁰; (3)b-actin sense, 5⁰-GCATGGGTCAGAAGGATTCCT-3⁰; and (4) antisense, 5⁰-TCGTCCCAGTTGGTGACGAT-3⁰. The gene levels were normalized by re- probing the blots with antibody tob-actin (Boster Biological Technology Co., Ltd., China), as shown inFig. 7C.

2.7. Immunoﬂuorescence analysis

Finally, the HeLa cells were incubated with appropriately diluted antibodies to AIF, EndoG (Beijing Zhongshan Golden Biotechnology Co; China) overnight, washed and incubated with corresponding secondary antibodies labeled with tetramethylrhodamine isothiocyanate (TMRITC) for 1 h. At last, they were washed with PBS, stained with DAPI (0.5mg/ml) and viewed underﬂuorescence microscopy, as shown inFig. 8. The controls are comprised of cells processed without primary antibody.

2.8. Statistical analysis

Statistical results were obtained using the statistical software SPSS17.0. ANOVA was used to analyze statistical differences between groups, and the Student’st-test was performed. A signiﬁcancep<0.05 is considered to be statistically signiﬁcant.

3. Results

The central issue associated with synthesis of polymeric drugs is to modify the polymeric material surface with death signal proteins. We employed TNF-a, IFN-g and TNF-a plus IFN-g respectively as the death signal proteins and photo-immobilization method toﬁx these proteins onto the cell culture polystyrene (PSt) plate surfaces. The details of the results are given in two sections below.

Table 1

Atomic percentage of C, O, and N on the surfaces of raw PSt and PSt immobilized with IFN-g, TNF-a, and IFN-gplus TNF-a(1:1), respectively. The data are evaluated from the ESCA survey scan spectra.

Substrate Atomic percentage (%)

C O N

Raw PSt 96.14 3.86 0

PSteIFN-g 67.23 15.48 17.29

PSteTNF-a 77.10 11.45 11.45

PSteIFN-gplus TNF-a(1:1) 73.26 12.63 14.11

Fig. 3.AFM surface images of the as-prepared polystyrene plates covered with free (absorbed) TNF-aplus IFN-g(A), immobilized IFN-g(B), immobilized TNF-a(C), and co- immobilized TNF-aplus IFN-g(D), respectively. For each case, the top-view (left-up), 3D-view (right-up), and cross-section (bottom) proﬁles were shown with various geometric data labeled, and the cross-section scan was done along the line shown in the top-view image.

(6)

Fig. 4.HeLa cells morphology as evaluated by transmission electron microscopy (A), cell mortalities as evaluated byﬂow cytometry (B), and the loss of mitochondrial membrane depolarization (DJm) monitored with the dye JC-1 (C). HeLa cells were cultured on the polystyrene surface in the presence of control (a), free IFN-g(b), free TNF-a(c), free IFN-g plus TNF-a(1:1) (d), immobilized IFN-g(e), immobilized TNF-a(f), co-immobilized IFN-gplus TNF-a(1:1) (g), and co-immobilized IFN-gplus TNF-a(1:2) (h). In (D), the rates of mortality and loss ofDJm are presented in the bar plot, with signiﬁcancep<0.05 labeled by symbol*, andp<0.01 labeled by symbol**, in comparison with control group. The bars stand for the standard deviations (n¼3). The length red bar in (A-h) stands for 2.5mm.

Fig. 5.Immunohistochemistry data (green bars: 50mm) (top) and the optical density analysis (bottom, bar plots) of the expression of p53, Bax and Bcl-2 induced by 20 ng/well co- immobilized TNF-aplus IFN-g(1:1). The HeLa cells were treated for 12, 24, and 48 h, respectively, with 20 ng/well free and co-immobilized TNF-aplus IFN-g(1:1). (A): p53; (B):

Bax; (C): Bcl-2; (D): optical density analysis. The optical densities are plotted with the signiﬁcancep<0.05 labeled by symbol*, 0.001<p<0.01 labeled by symbol**andp<0.001 labeled by symbol***, in comparison with the free group. The bars stand for the standard deviations (n¼3).

(7)

3.1. Materials characterizations

A schematic diagram of the PSt preparation and molecular structures of photoactive TNF-aor IFN-g(AzPhIFN-gor AzPhTNF-a) is shown elsewhere (Supplement Materials Fig. S1). The measured Fourier transform infrared spectroscopy (FITC) data on the AzPhTNF-aor AzPhIFN-g, is shown inFig. 1. Because of the presence of azidophenyl group bonded with AzPhTNF-aor AzPhIFN-g, a shift of the characteristic peak associated with this group is demonstrated (2109/2106 cm¹). The carboxyl of contraposition in the azido group becomes combined with the amino group, resulting in the amide bond, featured by the characteristic peak (1105/

1087 cm¹& 1660/1645 cm¹). The UV spectra of the AzPhTNF- aand AzPhIFN-gshown respectively inFig. 2A reveal the diffusive absorption at 206 nm or 212 nm assigned to the azidophenyl group, while the 4-Azidobenzic acid shows an absorption at 224 nm. This difference is due to the electron delocalization of the azidophenyl group caused by the amide bond formation. It is noted that no apparent absorbance for TNF-aor IFN-gitself in the wavelength range of 190e350 nm is observed. The measured ﬂuorescence spectra shown inFig. 2B indicate no substantial difference between AzPhTNF-a/AzPhIFN-g and raw TNF-a/IFN-g, in spite of some difference in magnitude. The remarkable difference of these spectra from those of 4-Azidobenzic acid indicates that the azidophenyl groups are inside the AzPhTNF-aor AzPhIFN-g.

The above spectroscopy evidences show clearly the effective reaction of TNF-aor IFN-gwith 4-Azidobenzic acid, i.e. formation of AzPhTNF-aor AzPhIFN-g.

Fig. 6.Western blotting data of caspase-3, caspase-8, and caspase-9 activities in HeLa cells. Cells were cultured in serum free medium and treated with control, free TNF- aplus IFN-gand co-immobilized TNF-aplus IFN-g(1:1) for 24 h, respectively. Data are the average of three representative experiments.

Fig. 7.Measured mRNA expression of AIF (A) and EndoG (B) by RTPCR analysis, mRNA expression by PCR analysis(C) and protein expression by western blotting analysis (D) of AIF and EndoG, in HeLa cells. In the RTPCR anaylsis, the HeLa cells were treated for 24 h with 20 ng/well free or immobilized IFN-g, TNF-aand IFN-gplus TNF-aon the polystyrene surface. The relative levels of AIF and EndoG are plotted with the signiﬁcancep<0.05 labeled by symbol*andp<0.01 labeled by symbol**, in comparison with the control group.

The bars stand for the standard deviations (n¼3). In the PCR and western blotting analysis, cells were treated on the polystyrene surface with control, free TNF-aplus IFN-gand co- immobilized TNF-aplus IFN-g(1:1) for 24 h, respectively. Data are representative of three experiments.

(8)

Subsequently, we checked the immobilization of the IFN-gand TNF-awith the PSt surfaces and the cross-linking between TNF- aand IFN-gfor the case of TNF-aplus IFN-g. The AFM images in Fig. 3show the peak height, Ave Rough (Ra), and surface coverage of immobilized TNF-a/IFN-gand co-immobilized TNF-aplus IFN-g. For IFN-g, the peak height is w25 nm with surface coverage w72.96% and Ra w3.609 nm (B). For TNF-a, these values arew50 nm, 98.92%, and 2.722 nm (C). For TNF-aplus IFN-g, they arew30 nm, 97.12%, and 4.368 nm (D), in comparison with the case of free TNF-aplus IFN-gabsorbed on the PSt surface for 24 h (A).

The formation of many nano-steps on the surface suggests that the TNF-aor IFN-gis cross-linked respectively with itself and also with the PSt molecules.

The ESCA survey scanning of the raw PSt surface and modiﬁed surface also conﬁrms the immobilization of IFN-g/TNF-aon the PSt.

For the raw surface, the two characteristic peaks respectively correspond to the C-1s (binding energy, 285eV) and O-1s (binding energy, 532eV). For the modified surface, three characteristic peaks corresponding respectively to the C-1s, O-1s and N-1s (binding energy, 400eV) are identified. The evaluated chemical compositions of the modified PSt surface in various cases are shown inTable 1.

The C-1s content of the raw PSt is 96.14%, much higher than 73.26%, the content for the modiﬁed PSt. The N-1s content is 17.29%, 11.45%, and 14.11%, for the PSt surfaces modiﬁed respectively by immobilized IFN-g, TNF-a, and TNF-aplus IFN-g.

3.2. Biomedical effects

The above characterizations demonstrated the successful immobilization of the IFN-g, TNF-a, and IFN-gplus TNF-aon the PSt surfaces. These polymeric substrates with immobilized death signal proteins are the polymeric anticancer materials. Subsequently, we prepared a series of polymeric samples with immobilized TNF- a/IFN-g, and TNF-a plus IFN-g and focused on the biomedical effects of these samples against the HeLa cells. For comparison purpose, we also prepared the control (polymeric sample) and polymeric samples absorbed with free TNF-a/IFN-g, and TNF-aplus IFN-g, respectively.

First, we checked the morphology of the HeLa cells upon treatments by various samples using TEM, as shown inFig. 4A. The chromatin of the control cells is homogeneously distributed in the

nuclei, without chromatin condensation. For the cells treated by the samples with free TNF-a/IFN-gand TNF-aplus IFN-g, the nuclear chromatin apparently condenses to the deep degree, forming chromatin blocks. The cell size is also smaller than the control cells, due to the cytoplasm reduction. For the cells treated by the samples with immobilized TNF-a/IFN-g and co-immobilized TNF-a plus IFN-g, the nuclear chromatin mildly condenses to theﬂoccular structures, just like the apoptosis-like PCD. These differences suggest that the HeLa cells treated by these death signal proteins immobilized polymeric materials may take different PCD pathways.

Then we compared theﬂow cytometry data on the control cell group (16.1% at the cell death state) and the cell groups treated by these proteins immobilized materials. It is shown inFig. 4B and D that a treatment by the samples with IFN-gplus TNF-aresults in more HeLa cells at the death state (33.3% for that with free IFN-gplus TNF-a; 45.4% and 27.6% for those with co-immobilized IFN-gplus TNF-a(1:1) and (1:2)). The mortalities analysis allows one to conclude that the materials with co-immobilized IFN-gplus TNF-a(1:1) have the most signiﬁcant effect in inducing the death of HeLa cells.

Moreover, as shown in Fig. 4C and D, once the HeLa cells are exposed to the sample with 20 ng co-immobilized IFN-gplus TNF- a (1:1) for 24 h, the loss of mitochondrial membrane potential (DJm) turns to be 20.42.5%, in contrast with 16.91.1% after 24 h for the treatment by the sample with free IFN-gplus TNF- a(1:1). These results indicate that the co-immobilized IFN-gplus TNF-abiomaterial may trigger the PCD by altering the integrity of mitochondria.

In addition, we addressed the effect of TNF-aplus IFN-g(1:1) on the expression of p53, Bax and Bcl-2 in HeLa cells by immunohistochemistry for 12, 24 and 48 h. The analysis was performed for the free and co-immobilized TNF-aplus IFN-g(1:1) treated HeLa cells (Fig. 5). The optical density analysis of p53, Bax and Bcl-2 demonstrated the difference in the variation speed of expression for free and co-immobilized TNF-aplus IFN-gtreated cells, respectively. In addition, the Bax/Bcl-2 for the free and co-immobilized TNF-aplus IFN-g(1:1) treated cells was evidently different.

Also, we look at the caspase activity in response to the two different cases. The antibody identiﬁed three caspase-9 bands with molecular weights of 47 kDa, 37 kDa and 35 kDa, three caspase-8 bands with molecular weights of 57 kDa, 45 kDa and 18 kDa, and three caspase-3 bands with molecular weights of 32 kDa, 17 kDa, Fig. 8.Immunoﬂuorescence data (white bar stands for 20mm) of protein localization of AIF (A) and EndoG (B), induced by TNF-aplus IFN-g. The HeLa cells were treated on the polystyrene surface for 24 h with 20 ng/well free and co-immobilized TNF-aplus IFN-g(1:1). Data are representative of three experiments.

(9)

and 12 kDa, respectively (Fig. 6). The results show no up-regulation of cleaved caspase-9 (35 and 37 kDa) expression with respect to the treatment by free and co-immobilized TNF-a plus IFN-g (1:1).

However, the levels of caspase-8 (57 kDa, 45 kDa and 18 kDa) expression for both the free and co-immobilized TNF-aplus IFN-g (1:1) cases show the signiﬁcant up-regulation. Similarly, the levels of caspase-3 (32 kDa) expression for the two cases were up-regulated apparently. From theseﬁndings, it is realized that a minor part of the HeLa cell death induced by the co-immobilized TNF-aplus IFN-g(1:1), similar to the case of free TNF-aplus IFN-g, is caspase-8 and then caspase-3 dependent apoptosis.

However, the morphology of most cells induced by co-immobilized TNF-aplus IFN-g(1:1) seems to be the outcome of apoptosis- like PCD instead of apoptosis. Thus, we subsequently investigated the AIF mRNA and EndoG mRNA levels in various HeLa cell groups, treated respectively by these polymeric samples for 24 h. The obtained results are summarized in Fig. 7A and B. Indeed, the samples with immobilized proteins (IFN-gand 1:1 IFN-gplus TNF-a) at a dose of 20 ng/well cause remarkable enhancement of EndoG mRNA level at 24 h, but the samples with free proteins does not. The amount of AIF mRNA at 24 h shows slight increase upon the treatment by the samples with co-immobilized IFN-gplus TNF-a(1:1, 20 ng/well). These results indicate that the co-immobilized IFN-g plus TNF-a(1:1) may trigger the PCD by up-regulating the mRNA expression of EndoG, but not that of AIF. Furthermore, based on the fact that a more obvious band with 1145 bp of EndoG is potently induced by co-immobilized IFN-gplus TNF-a(1:1) (Fig. 7C), we argue that the co-immobilized IFN-gplus TNF-a(1:1) may trigger the PCD by up-regulating the mRNA expression of EndoG..

In addition, a speciﬁc band with molecular weight of 57 KDa and 30 kDa in the antibody of AIF and EndoG is recognized, as shown in Fig. 7D. Although protein expression levels of AIF do not change obviously for the treatment by samples with free and immobilized proteins, the sample with co-immobilized IFN-gplus TNF-a(1:1) clearly regulates the protein expression levels of EndoG. These results indicate that the co-immobilized IFN-g plus TNF-a (1:1) may trigger the PCD also by up-regulating the protein expression of EndoG.

To further determine whether the EndoG (or AIF) expression leads to the HeLa cell death or not, we examined the EndoG (or AIF) translocation from the mitochondria into the nucleus. This translocation may be responsible for the late apoptotic execution of the HeLa cells upon the treatment by the sample with co-immobilized IFN-g plus TNF-a(1:1). We then employed the immunoﬂuores- cence staining methods to identify this issue. As shown inFig. 8, a trace of AIF is observed in the nuclear but some amount of EndoG is present in the nuclear. Because EndoG is translocated into the nucleus to cleave DNA chromatin into nucleosomal fragments, our results demonstrate that in this case, the HeLa cell death may mainly involve a caspase-independent pathway associated with mitochondrial.

4. Discussion

Right till now, many methods including chemotherapy, radia- tion therapy, and biological therapy, are used for cancer treatment.

The main problem for these treatments is the poor selectivity of their actions against rapidly growing cancer cells, so that normal cells in patients are also damaged[20]. Therefore, developing new therapeutic drugs substituting for or synergizing with the existing ones is substantial.

Recently many researchers[21e26]have turned their interests into cancer treatments with biomaterials. Along this line, polymeric micelles were successfully prepared and optimized, in order to improve tumor targeting and enhance the antitumor activity. In

addition, biodegradable polymersomes which cause two-fold higher cell death rate than that free drug can do in tumors was also described.

Generally, it is of importance to understand the mechanism of PCD in cancer cells, because PCD is believed to be one of the major consequences of anticancer drug treatment against malignancies. In our earlier work, the apoptotic impact of the co-immobilized TNF- aplus IFN-gon HeLa cells was revealed[13]. Given the fact that the co-immobilized TNF-aplus IFN-gis the vital effective cytokine in induction of cell death for the HeLa cells, the systematic characterizations presented here shed further insight into the mechanisms by which PCD in human cervical cancer HeLa cells is induced.

In the present study, we show that the morphological charac- teristics of HeLa cells induced by co-immobilized IFN-gplus TNF- abiomaterial are similar to those of apoptosis, forming extracel- lular apoptotic bodies, so this death pathway should not be necrosis-like PCD[27]. It should not be autophagy[28]because the death pathway is with integral membrane and without obvious intracellular vacuoles. Moreover, due to the chromatin condensation at low degree, no chromatin bulks, indicating that the PCD could not be classic apoptosis. Thus, based on the preliminary deductions above, we argue that it should be apoptosis-like PCD.

On the other hand, it has been shown that the mitochondrial Bcl-2 protein family plays an essential role in apoptosis. This family is a large group of apoptosis-regulating proteins that modulate the mitochondrial pathways. They include both the anti-apoptotic proteins (e.g. Bcl-2, Bcl-XLand Mcl-1) and various pro-apoptotic proteins (e.g. Bax, Bak and Bad). These proteins regulate the mitochondrial membrane permeability, either promoting or suppress- ing the release of apoptogenic proteins from these organelles [29,30]. In this study, the immunohistochemistry data reveal that the treatment of HeLa cells with the co-immobilized TNF-aplus IFN-g remarkably enhanced the cellular expression of activated Bax. The present results suggest that the mitochondrial pathway of cell death, including Bcl-2 family and the release of apoptogenic proteins (including AIF or EndoG), may be signiﬁcant.

At the same time, the tumor suppressor p53 in regulating the expression of genes that mediate cell cycle arrest and/or apoptosis in response to genotoxic insults has been well recognized[31].

Following the environmental insults, p53 is activated by post- translational modiﬁcations such as phosphorylation and acetyla- tion, which enhances its protein stability and DNA binding activity.

The activated p53 up-regulates the expression of down-stream target genes, including Bax, the p53-up-regulated modulator of apoptosis. Our data also conﬁrm that the p53-expression was earlier than Bax (seeFig. 5). In fact, several earlier works indicated that p53 activates the apoptotic machinery through the regulation of the Bax function and mitochondrial integrity, which results in the release of apoptogenic proteins[32,33]. In our study, the co- immobilized TNF-aplus IFN-gwas found to cause the rapid accu- mulation of p53 at 24 h, which is positively correlated with subsequent Bax and the activation, release of EndoG.

Furthermore, the key proteins that modulate the apoptotic response are represented by a family of cysteine proteases called caspases. These caspases appear to be involved in regulating the activation of apoptotic signal transmission[34,35]. Among these, caspase-3 is a key executioner of apoptosis, whose activation is mediated by the initiator caspases such as caspase-8 and caspase-9.

In our experiments, caspase-8 and caspase-3, were activated in the PCD process induced by the co-immobilized TNF-aplus IFN-glike the free cases (Fig. 6). It is reasonable to conclude that the caspase-8 and -3 dependent apoptosis and mainly the caspase-independent apoptosis-like PCD are the reasons for the antiproliferative effect induced by the co-immobilized TNF-aplus IFN-g.

Actually, apoptosis-like PCD, which is caspase-independent, is a new kind of pathway that has been receiving attentions recently.

(10)

In apoptosis-like PCD, the AIF and (or) EndoG release and trans- locate into nucleus, causing large fragment of DNA, and the cells ﬁnally die[20,36e41]. Interestingly, ourﬁndings also suggest that the treatment by the samples with immobilized IFN-g and co-immobilized IFN-g plus TNF-a (1:1, 20 ng/well) causes remarkable enhancement of EndoG mRNA level at 24 h (Fig. 7B), indicating that both the immobilized IFN-gand the co-immobilized IFN-gplus TNF-amight up-regulate the activation of EndoG in HeLa cells. In addition, we could argue that the synergistic effect of co- immobilized IFN-g plus TNF-a might mainly come from the immobilized IFN-g, and the IFN signaling pathway induced by co- immobilized IFN-gplus TNF-a(1:1, 20 ng/well) might be mainly responsible for the HeLa cell death[13].

Moreover, it is also clearly demonstrated that both immobilized IFN-g and co-immobilized IFN-g plus TNF-a (1:2, 20 ng/well) caused remarkable enhancement of AIF mRNA level at 24 h (Fig. 7A). However, according to the results of cell mortalities and the loss ofDJm in mitochondria, AIF might be regulated strongly in protein level so that the effect to induce the cell death is weak.

Therefore, the fact that the co-immobilized IFN-gplus TNF-a(1:2, 20 ng/well) cannot potentiate cell death pathway just as the co- immobilized IFN-gplus TNF-a(1:1, 20 ng/well) do is not surprising.

Finally, we investigate the protein expression and the protein translocation of AIF or EndoG to the nuclear, and the data clearly reveal the enhanced protein expression and nuclear localization of EndoG at 24 h in the cells treated by co-immobilized IFN-g plus TNF-a, while no such convincing effect on the cells treated by free IFN-gplus TNF-ais detected. Thisﬁnding suggests that the EndoG- activation is responsible for cell death upon the stimulus of co- immobilized TNF-a plus IFN-g (1:1, 20 ng/well). Therefore, the results clearly indicate that EndoG-mediated mitrochondrial death pathway might play a critical role in co-immobilized IFN-g plus TNF-a(1:1)-induced programmed cell death of HeLa cells.

We thus present our mechanistic model for the programmed cell death pathway induced by the polymeric drug with co-immobilized TNF-aplus IFN-g, as shown inFig. 9. Although immobilized TNF-a

may also activate the death-receptor pathway[42]and result in a caspase-dependent apoptosis which up-regulates the caspase-8 and then caspase-3, we argue that the synergistic action by the combined IFN-g and TNF-a might mainly up-regulate the IFN signaling components and result in some factors (e.g. p53), which could act in the mitochondria and alter the mitochondrial function through a variety of signals (e.g. Bax and Bcl-2)[13]. This leads to mitochondrial protein EndoG but not AIF released, inducing the apoptosis-like programmed cell death in HeLa. Although we still need further investigations to elucidate the role of EndoG clearly, our data presented here illustrate how the programmed cell death pathway induced by the biomaterial of IFN-gplus TNF-asynergism in HeLa cells works.

5. Conclusion

We have synthesized polymeric materials with co-immobilized IFN-gplus TNF-aand shown that the induced HeLa cell death may be mainly apoptosis-like PCD. It has been demonstrated that the releasing EndoG may be a mechanism in the induction of apoptosis-like PCD. Our study thus makes a new model, with which the subsequent research of apoptosis-like PCD induced by biomaterials can be explored. Furthermore, the merit of these studies might also enable the eventual development of multifunctional nanoparticle drugs in cervical cancer treatment.

Acknowledgments

The expenses of this work were supported by the "Natural Science Foundation of China (30970731, 50832002), the Natural Science Foundation of Guangdong Province (9151063101000015) Appendix. Supplementary material

Supplementary data associated with this article can be found in online version atdoi:10.1016/j.biomaterials.2011.01.060.

Fig. 9.Schematic representation of the proposed mechanism on the cell death pathway in the induction of co-immobilized TNF-aplus IFN-gbiomaterial.

(11)

Appendix

Figure with essential color discrimination.Figs. 1e5,8 and 9in this article is difﬁcult to interpret in black and white. The full color image can be found in the online version, at doi:10.1016/j.

biomaterials.2011.01.060.

References

[1] Matsuda T, Sugawara T. Development of a novel proteinﬁxation method with micron-order precision. Langmuir 1995;11(6):2267e71.

[2] Ohno K, Kayama Y, Ladmiral V, Fukuda T, Tsujii Y. A versatile method of initiatorﬁxation for surface-initiated living radical polymerization on polymeric substrates. Macromolecules 2010;43(13):5569e74.

[3] Ito Y, Hasuda H, Yamauchi T, Yamauchi T, Komatsu N, Ikebuchi K. Immobili- zation of erythropoietin to culture erythropoietin-dependent human leukemia cell line. Biomaterials 2004;25(12):2293e8.

[4] Kang IK, Kim GJ, Kwon OH, Ito Y. Co-culture of hepatocytes andﬁbroblasts by micropatterned immobilization of b-galactose derivatives. Biomaterials 2004;25(18):4225e32.

[5] Chen GP, Kawazoe NK, Fan YJ, Ito Y, Tateish T. Grid pattern of nanothick microgel network. Langmuir 2007;23(11):5864e7.

[6] Leist M, Jäättelä M. Four deaths and a funeral: from caspases to alternative mechanisms. Cell Biol 2001;2(8):589e98.

[7] Jäättelä M, Tschopp J. Caspase-independent cell death in T lymphocytes. Nat Immunol 2003;4(5):416e23.

[8] Dalla VL, Marini AM, Salerno S, La Motta C, Condello M, Arancia G, et al.

Synthesis and biological activity of 1, 4-dihydrobenzothiopyrano [4, 3-c]

pyrazole derivatives, novel pro-apoptotic mitochondrial targeted agents.

Bioorg Med Chem 2009;17(1):326e36.

[9] Aggarwal BB, Eessalu TE, Hass PE. Characterization of receptors for human tumor necrosis factor and their regulation by r-interferon. Nature 1985;318 (6047):665e7.

[10] Yarden A, Kimchi A. Tumor necrosis factor reduces c-myc expression and cooperates with interferon-gamma in HeLa cells. Science 1986;234 (4782):1419e21.

[11] Li JH, Pober JS. The cathepsin B death pathway contributes to TNF plus IFN- {gamma}-mediated human endothelial injury. J Immunol 2005;175(3):1858e66.

[12] Guan YQ, He LM, Cai SM, Zhou TH. Anti-cervix-cancer effect of the co- immobilized tumor necrosis factor-a and interferon-g. J Mater Sci Tech 2006;22(2):200e4.

[13] Guan YQ, Li ZB, Liu JM. Death signal transduction induced by co-immobilized TNF-a plus IFN-g and the development of polymeric anti-cancer drugs.

Biomaterials 2010;31(34):9074e85.

[14] Beyaert R, Vanhaesebroeck B, Heyninck K, Boone E, De VD, Schulze-Osthoff K, et al. Sensitization of tumor cells to tumor necrosis factor action by the protein kinase inhibitor staurosporine. Cancer Res 1993;53(11):2623e30.

[15] Suk K, Chang I, Kim YH, Kim SS, Kim JY, Kim H, et al. Interferon-gamma (IFN gamma) and tumor necrosis factor-alpha synergism in ME-180 cervical cancer cell apoptosis and necrosis. IFN gamma inhibits cytoprotective NF-kappa B through STAT1/IRF-1 pathways. J Biol Chem 2001;276(16):13153e9.

[16] Thomas M, Finnegan CE, Rogers KMA, Purcell JM, Trimble A, Johnston PG, et al.

STAT1: a modulator of chemotherapy-induced apoptosis. Cancer Res 2004;

64(22):8357e64.

[17] Buckley NE, Hosey AM, Gorski JJ, Purcell JW, Mulligan JM, Harkin DP, et al.

BRCA1 regulates IFN-{gamma} signaling through a mechanism involving the type I IFNs. Mol Cancer Res 2007;5(3):261e70.

[18] Hayashi T, Ishida Y, Kimura K, Iwakura Y, Mukaida N, Kondo T. IFN-{gamma}

protects cerulein-induced acute pancreatitis by repressing NF-{kappa}B activation. J Immunol 2007;178(11):7385e94.

[19] Tliba O, Amranl Y. Airway smooth muscle cell as an inﬂammatory cell: lessons learned from interferon signaling pathways. Proc Am Thorac Soc 2008;5 (1):106e12.

[20] Lin ML, Lu YC, Chung JG, Li YC, Wang SG, NG SH, et al. Aloe-emodin induces apoptosis of human nasopharyngeal carcinoma cells via caspase-8-mediated activation of the mitochondrial death pathway. Cancer Lett 2010;291 (1):46e58.

[21]. Nishiyama N, Okazaki S, Cabral H, Miyamoto M, Kato Y, Sugiyama Y, et al.

Novel cisplatin-incorporated polymeric micelles can eradicate solid tumors in mice. Cancer Res 2003;63(24):8977e83.

[22] Bae Y, Nishiyama N, Fukushima S, Koyama H, Yasuhiro M, Kataoka K. Prepa- ration and biological characterization of polymeric micelle drug carriers with intracellular pH-triggered drug release property: tumor permeability, controlled subcellular drug distribution, and enhanced in vivo antitumor efﬁcacy. Bioconjug Chem 2005;16(1):122e30.

[23] Kataoka K, Nishiyama N, Cabral H. Optimization of (1, 2-diamino-cyclo- hexane) platinum (II)-loaded polymeric micelles directed to improved tumor targeting and enhanced antitumor activity. J Control Release 2007;121 (3):146e55.

[24] Lee ES, Na K, Bae YH. Polymeric micelle for tumor pH and folate-mediated targeting. J Control Release 2003;91(1e2):103e13.

[25] Bae Y, Fukushima S, Harada A, Kataoka K. Design of environmentesensitive supramolecular assemblies for intracellular drug delivery: polymeric micelles that are responsive to intracellular pH change. Chem Int Ed Engl 2003;42 (38):4640e3.

[26] Savic R, Luo LB, Eisenberg A, Maysinger D. Micellar nanocontainers distribute to deﬁned cytoplasmic organelles. Science 2003;300(5619):615e8.

[27] Lin CY, Wu HY, Wang PL, Yuan CJ. Mammalian Ste20-like protein kinase 3 induces a caspase-independent apoptotic pathway. Int J Biochem Cell Biol 2010;42(1):98e105.

[28] Zhang WG, Li D, Mehta JL. Role of AIF in human coronary artery endothelial cell apoptosis. Am J Physiol Heart Circ Physiol 2004;286(1):H354e8.

[29] Adams JM, Cory S. The Bcl-2 protein family: arbiters of cell survival. Science 1998;281(5381):1322e6.

[30] Willis SN, Fletcher JI, Kaufmann T, Delft MF, Chen L, Czabotar PE, et al.

Apoptosis initiated when BH3 ligands engage multiple Bcl-2 homologs, not Bax or Bak. Science 2007;315(5813):856e9.

[31] Aylon Y, Oren M. Living with p53, dying of p53. Cell 2007;130(4):597e600.

[32] Cano CE, Gommeaux J, Pietri S, Culcasi M, Garcia S, Seux M. Tumor protein 53- induced nuclear protein 1 is a major mediator of p53 antioxidant function.

Cancer Res 2009;69(1):219e26.

[33] Miyashita T, Reed JC. Tumor suppressor p53 is a direct transcriptional acti- vator of the human bax gene. Cell 1995;80(2):293e9.

[34] Salvesen GS, Dixit VM. Caspases: intracellular signaling by proteolysis. Cell 1997;91(4):443e6.

[35] Fuentes-Prior P, Salvesen GS. The protein structures that shape caspase activity, speciﬁcity, activation and inhibition. Biochem J 2004;384(Pt 2):201e32.

[36] Zong WX, Thompson CB. Necrotic death as a cell fate. Genes Dev 2006;20 (1):1e15.

[37] Yoshimori T. Autophagy: paying charon’s toll. Cell 2007;128(5):833e6.

[38] Candé C, Cecconi F, Dessen P, Kroemer G. Apoptosis-inducing factor (AIF): key to the conserved caspase-independent pathways of cell death? J Cell Sci 2002;115(24):4727e34.

[39] Li LY, Luo X, Wang XD. Endonuclease G (EndoG) is an apoptotic DNase when released from Mitochondria. Nature 2001;412(6842):95e9.

[40] Zhang ZF, Yang XF, Zhang S, Ma XL, Kong JM. BNIP3 upregulation and EndoG translocation in delayed neuronal death in stroke and in hypoxia. Stroke 2007;38(5):1606e13.

[41] Buttner S, Eisenberg T, Carmona-Gutierrez D, Ruli D, Knauer H, Ruckenstuhl C, et al. Endonuclease g regulates budding yeast life and death. Mol Cell 2007;25 (2):233e46.

[42] Russo M, Mupo A, Spagnuolo C, Russo GL. Exploring death receptor pathways as selective targets in cancer therapy. Biochem Pharmacol 2010;80(5):674e82.