integrin subunit

Receptors of the β1 integrin family are involved in many tumor-promoting activities. There are several approaches currently used to control integrin activity, and thus to potentially restrain tumor metastasis and angiogenesis. In this study, we compared inhibitory efficiencies of siRNA and DNAzymes against the β1 integrin subunit (DEβ1), in a mouse xenograft model. Both inhibitors were used under their most favorable conditions, in terms of concentrations, incubation time and lack of cytotoxic effects. Transfection of siRNAβ1 or DEβ1 remarkably inhibited the growth of both PC3 and HT29 colon cancer cells in vitro, and decreased their capability of initiating tumor formation in the mouse xenograft model. siRNAβ1 appeared to be slightly more efficient than DEβ1 when tested in vitro, however it was comparably less proficient in blocking the tumor growth in vivo. We conclude the DNAzyme, due to its greater resistance to degradation in extra- and intracellular compartments, to be a superior inhibitor of tumor growth in long lasting experiments in vivo when compared to siRNA, while the latter seems to be more efficient in blocking β1 expression during in vitro experiments using cell cultures.


INTRODUCTION
Adhesion molecules such as integrins, mediate direct cell-cell recognition and cell-matrix interactions (Hynes, 1992), which are essential for tumor cell migration (Holly et al., 2000) and basement membrane penetration (Uhm et al., 1999).Although integrins have become attractive therapeutic targets (Albelda, 1993), there are contrasting results on integrin expression patterns in different tumor types, making it difficult to draw general conclusions on their role in metastasis.Transformed cancer cells are often characterized either by the loss/reduction or increase of integrin expression (Pignatelli et al., 1991;Zutter et al., 1990).Furthermore, tumor progression and metastasis are associated with changes in a numerous integrin signaling cascades eliciting various cell functions, such as morphological changes, adhesion, migration and gene activation, which are all relevant to the metastatic cascade.
There are several approaches currently used to downregulate the integrin activity, and thus to restrain tumor metastasis and angiogenesis.Up until now, the most advanced studies researched the blockade of integrin interaction with an extracellular matrix.They chiefly focus on the applications of monoclonal antibodies, small-molecule peptides, and peptidemimetics (Ma & Adjei, 2009;Yazji et al., 2007).Another group of approaches is based on gene-silencing methods, in which compounds that function with sequence-specificity at a post-transcriptional level are used.Among them, the most intensively studied compound is the small interfering RNA (siR-NA), which has recently been developed as a powerful tool to suppress the expression of specific gene products (Schiffelers et al., 2004;Kohlgraf et al., 2003;Mukherjee et al., 2005;Ren et al., 2004).This technique was successfully used to explore its potential therapeutic values (Hannon, 2002).A different approach is represented by DNAzymes, which recently have emerged as a new class of nucleic acid-based gene-silencing agents (Isaka, 2007).DNAzymes are single-stranded DNA catalysts, which cleave the target at predetermined phosphodiester linkages (Schubert et al., 2003).Due to the low cost of synthesis, high stability and flexible rational design features, DNAzymes have been demonstrated to be a potential new class of drugs inhibiting gene expression (Schubert & Kurreck, 2004;Cieslak, 2002;Cieslak, 2003;Manes et al., 2003).
In this report, we compared the capability of siRNA and DNAzymes to inhibit the expression of β1 integrins in colon adenocarcinoma (HT29) and prostate (PC3) cancer cells.The reason for which we targeted the β1 subunit is because this integrin family includes twelve members, thus this method has a broad-spectrum antiintegrin effect (Goel et al., 2005;Niewiarowska et al., 2009).Although both inhibitors were aimed at the same target, the mechanisms by which they caused the degradation of mRNA were different.
Carcinoma cell lines and culture conditions.The human colon adenocarcinoma cell line HT29, was obtained from Ludwik Hirszfeld Institute of Immunology and Experimental Therapy (Polish Academy of Sciences, Wroclaw, Poland).The human prostate carcinoma cell line PC3, was obtained from ATCC company (American Type Culture Collection, Manassas, VA, USA).HT29 and PC3 cells were cultured as described previously (Wiktorska et al., 2010).For the experiments with siRNA, cells were transferred to 6-well dishes and used at 70% confluence in the MEM-α (HT29) or F-12K (PC3) full medium without antibiotic.LipofectAMINE TM Reagent (5 μg/ml) was diluted in Opti-MEM reduced medium (GIBCO BRL, Invitrogen, Carlsbad, CA, USA) containing 5 mM MgCl 2 and joined with a mixture of all four siRNAs diluted with the same medium to obtain the final concentration of 60 nM.After a 24 h incubation, the medium was exchanged for one containing antibiotics.48 h post-transfection cells were detached with trypsin/EDTA.Subsequently, DNAzymes were mixed with LipofectAMINE TM Reagent (5 µg/ml) and suspended in Opti-MEM-reduced medium containing 5 mM MgCl 2 to obtain a final concentration of 1 µM.Transfection was performed for 6 h according to the manufacturer's protocol.After a 12 h incubation in a corresponding medium supplemented with 10% FBS, cells were detached with trypsin/EDTA and used for experiments.Cell viability was determined microscopically by trypan blue exclusion and only cell cultures having less than 1% of dead cells were included in the study.
Animals.Six-to eight-weeks-old BALB/cA nude (nu-/-)-B6.Cg-Foxn1 nu mice (Mus musculus) were purchased from Taconic Europe, Ejby, Denmark.Mice were housed under pathogen-free conditions in microisolator cages with laboratory chow and water available ad libitum.For the experiments animals were divided into three groups (each group n = 5), anesthetized before any invasive procedures, and placed under observation until fully recovered.All experiments and procedures were reviewed by the Local Ethical Committee and performed in accordance with the EU regulations regarding the humane care and use of laboratory animals.
Western blotting.Subconfluent cells were washed with PBS and lysed in Mammalian Protein Extraction Reagent (M-PER, PIERCE, Rockford, IL, USA) supplemented with protein inhibitor cocktail (Roche, Basel, Switzerland).In all, 30 µg samples of total protein from cells mock-transfected or transfected with siRNA or DNAzymes were treated as previously (Wiktorska et al., 2010).The membrane was incubated with rabbit polyclonal anti-human β1 integrin subunit antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and then the level of β-actin was detected with rabbit polyclonal antibody (Abcam, Cambridge, MA, USA).
Flow cytometry.Harvested cells were washed with serum-free appropriate medium.Cells (1 × 10 5 ) suspended in a medium containing 1% bovine serum albumin were incubated in the dark at 4°C for 30 min with fluorescein isothiocyanate-conjugated monoclonal antibodies against the β1 subunit (0.1 μg/ml) (DAKO A/S, Denmark).After being washed and fixed with 1% paraformaldehyde/PBS, cell fluorescence was measured with a FACScan flow cytometer (Becton Dickinson, Mountain View, CA, USA).The results were analyzed with PC Lysis II software.
Adhesion assay.Maxisorp loose Nunc-Immuno TM modules (Pittsburgh, PA, USA) were coated for 2 h with 100 μl of fibronectin or collagen type I at 10 μg/ml/ TBS.Next wells were washed and blocked for 1.5 h at 37°C in a humidified 5% CO 2 atmosphere with 200 μl of 1% BSA/TBS (0.1 mM CaCl 2 , 1 mM MnCl 2 ).Cells were then added at 1 × 10 5 cells/0.1 ml of appropriate medium for 1.5 h.The total cell-associated protein was determined by dissolving the attached cells in 200 μl of BCA protein assay reagent directly in the microtiter wells (PIERCE, Waltham, MA, USA).The absorbance was determined at 562 nm (BioKinetics Reader EL340, Bio-Tek Instruments, Winooski, VT, USA).
Mouse tumor model.Tumor xenograft model was performed as described before with a modification (Niewiarowska et al., 2009).HT29 or PC3 cells were established by s.c.injection of 2 × 10 6 cells mixed with Matrigel™ at the ratio of 1 : 1 into female (n = 20) or male (n = 20) BALB/cA nude (nu-/-)-B6.Cg-Foxn1 nu mice, respectively.When tumors reached the volume of 80-150 mm 3 , mice were divided into two groups.Specimens of the first group received an injection of 0.5 nmol mixed siRNAs (1.67 µg each one per tumor), while the specimens of the second one were injected with 6.67 mg of siRNA C per tumor 8 times every second day.In parallel, when DNAzymes were used, mice were also divided into two groups: the first one received an injection of 1.25 µg mDEβ1, and the second one of 1.25 µg mDE C per tumor 8 times every second day.Tumors were measured three times a week and their volumes were calculated by the formula [π/6 (w 1 × w 2 × w 3 )].Tumor specimens were fixed in 4% buffered formaldehyde and were routinely processed for paraffin embedding.
Immunohistochemistry. Paraffin sections were incubated overnight at 4°C with rat monoclonal anti CD34 (MEC 14.7; Abcam, Cambridge, MA) at a dilution of 1 : 100.Afterwards, the polyclonal rabbit anti-rat immunoglobulins/HRP (P0450; DakoCytomation, Glostrup, Denmark) were used.Visualization was performed by incubating the sections in a 3.3′-diaminobenzidine solution (DakoCytomation, Glostrup, Denmark).For each sample a negative control was processed.The microvessels were Blocking tumor growth by siRNAβ1 and β1DE determined by counting all CD34-positive structures (MultiScan 8.08 software, Computer Scanning Systems, Poland) in a sequence of 10-15 consecutive computer images of 250 × high power fields of 0.047914 mm 2 each.The mean values of microvessels with or without lumen were calculated per mm 2 .Data analysis.All values are expressed as mean ± S.D. and were compared with controls.Significant difference was taken for P values less than 0.05.

Inhibition of cancer cell adhesion and invasiveness
Before comparing the efficiency of siRNAβ1 and hDEβ1 to inhibit adhesive and tumorogenic properties of cancer cells, we evaluated their effect on the β1 expression in PC3 and HT29 cells in preliminary experiments.Incubation of cells with siRNAβ1 and hDEβ1 for 48 and 18 h respectively, specifically inhibited synthesis of the β1 in both types of cells.A quantitative analysis of the β1 by densitometry revealed a significant (P < 0.001) decrease in the β1 protein in both types of cells transfected either with siRNAβ1 or hDEβ1 when compared to controls (Figs.1A1, B1).The β-actin expression was unaffected neither by the controls nor the siRNAβ1 or hβ1DE, indicating that non-specific downregulation of protein expression did not occur and equal quantities of protein were loaded.Under these conditions, both agents significantly reduced the β1 integrin subunit expression on the surface of PC3 and HT29 cells when compared to controls (P < 0.05), as detected by flow cytometry (Figs.1A2,  B2).Both agents were used under their most favorable conditions, in terms of concentrations and incubation times.Such settings, although different, were optimal to bring about the most advanced inhibition of the β1 expression with no cytotoxic effects observed.
Treatment of PC3 and HT29 cells with siRNAβ1 and hDEβ1 resulted in a significant inhibition of adhesion to fibronectin and collagen type I (Fig. 2).In this assay cell adhesion to either fibronectin or col-  lagen was analyzed for 1.5 h at 37°C (Figs.2A, B, D,  E).The extent of adhesion was evaluated based on the amounts of cellular protein detected as associated with wells.siRNAβ1 consistently inhibited the adhesion of both cell types to either fibronectin or collagen by 50-55%.hDEβ1 was less efficient in blocking adhesion to fibronectin.It produced similar inhibition to siRNAβ1 when adhesion of cells was tested to collagen.
To evaluate the effect of both inhibitors on cell invasion, transwell invasion assays were carried out.Transfection drastically reduced the number of PC3 or HT29 cells that invaded through the Matrigel™-coated membrane when compared to untreated cells.(Fig. 2C, F).Downregulation of β1 integrins in these cells led to similar decrease in the number of invading cells, namely by 65 to 80%, pointing out no significant difference in the inhibitory efficiency between both used inhibitors.

Inhibition of tumor growth
To compare the anti-tumor activity of siRNAβ1 and mDEβ1 in vivo we established human PC3 cell xenografts in male BALB/cA nude (nu -/-)-B6.Cg-Foxn1nu mice.Then, 6.67 µg siRNAβ1 or mDEβ1 at 1.25 µg per injection, either active or control, were administered into solid PC3 tumors every second day after the tumor volume was assessed to be 80-150 mm 3 .Fig. 3A shows that solid prostate carcinoma growth is considerably inhibited by siRNA when compared with control.When mDEβ1 was used in the same model, the tumor growth was almost entirely congested (Fig. 3B).To quantify blood vessels in PC3 tumors from control and siRNAβ1-or mDEβ1-animals, tissue sections were stained immunochemically with monoclonal antibody to CD34.Immunostaining demonstrated both blood vessels with wide lumen and with markedly narrowed lumen located within tumor stroma.Treatment caused a statistically significant decrease (P < 0.01) in the number of CD34-positive tumor blood vessels when compared with relevant controls (Fig. 3C).Microscopic evaluation revealed that in siRNAβ1-or mDEβ1-treated tumors the vascular stroma was scant, and that large areas of tumor cells underwent ischemic necrosis (not shown).
Figure 4 shows that mDEβ1 also blocked the tumor growth of human colon adenocarcinoma more efficiently than siRNAβ1.In these experiments PC3 cells were substituted by HT29 cells to develop solid tumors in female BALB/cA nude mice.All experiments were performed exactly as described above.When administered intratumorally, the siRNA or DE showed a direct inhibition of colon solid tumor growth by targeting the β1 integrins.Also, in this system, there was a significant reduction in the number of blood vessels in the HT29 tumors treated with siRNAβ1 or mβ1DE compared with the control tumors (P < 0.01).

DISCUSSION
Several reports showed that β1 integrins contribute to tumor progression through their participation in signaling events that control such functions as migration, proliferation, survival, invasion, and angiogenesis (Rathinam & Alahari, 2010).The β1 integrins, particularly α2β1 and a5β1, promote tumor growth in vivo and are uniquely required in cancer cells for localization, expression, and functioning of the insulin-like growth factor type 1 receptor (IGF-IR), which is known to support cancer cell proliferation and survival (Goel et al., 2005).The mechanism proposed for the β1 integrin controlling the IGF-IR activity involves the recruitment of specific adaptors to the plasma membrane by the β1 and increasing their concentration proximal to the growth factor receptor (Goel et al., 2004).
In our experimental setting, in order to reduce the expression of integrins, we used inhibitory nucleic acids, which cannot work as agonists.To knock down the β1 integrin synthesis two approaches were used and their inhibitory efficiency in anti-tumor activity was compared.The first one utilized DNAzymes, a novel class of antisense molecules.The 10-23 DNAzymes belong to a group of RNA-cleaving DNA molecules that contain a catalytic domain and cleave the RNA sequence at a phosphodiester bond between an unpaired purine and a paired pyrimidine residue (Santoro & Joyce;1998Silverman, 2005;;Tritz et al., 2005).In our previous studies we observed that unmodified DNAzymes were rapidly degraded in cells.Therefore, to avoid the degradation in present studies we used their O-Metyl analogs (Cieslak et al., 2003).
The 10-23 DNAzymes targeting GATA-3 mRNA have recently been developed and their anti-asthmatic effect in mouse models has been successfully demonstrated (Sel et al. 2008).Since GATA-3 plays a central role in Th2 cell differentiation (Barnes, 2008) and in promoting Th2 responses (Zhu et al., 2006), the relevant DNAzymes are expected to be useful for the treatment of inflammatory skin diseases.
In the second approach we used the siRNA to β1 integrin.Previous reports have indicated that siRNA has advantages over antisense oligonucleotides due to its greater resistance to nuclease degradation (Bertrand et al., 2002).Although it has a high specificity in gene silencing, there is a frequent off-target suppression of other genes resulting from partial complementarity (Jackson et a.l, 2003), immunostimulation of adverse effects (Sioud, 2006), and toxicity from interfering with endogenous mi-croRNA pathway (Petrocca & Lieberman, 2011).Hence, there is a need for further investigation and search for more efficient tools to control integrin-dependent cellular processes.
Comparing the levels of knockdown achieved by siR-NA and DNAzymes is not an easy task, since the design rules for sequence and site selection, as well as optimal transfection conditions for both of them are different.Therefore, in our studies both inhibitors used under their optimal conditions, at which they displayed highest efficiency in blocking the β1 expression in cancer cells without producing cytotoxic effects.Interestingly, siRNAβ1 appeared to be a slightly more capable inhibitor than β1DE when tested in vitro, however it was less effective in blocking the growth of tumors produced by PC3 cells.In vitro, even when used at much lower concentration than β1DE, it produced the same or higher level of inhibition of the total β1 synthesis, and more extended downregulation of the β1 expression on cell surface expression in both types of cancer cells used.In contrast, intra-tumor administration of β1DE virtually terminated the tumor growth when both PC3 and HT29 cells were used to produce xenografts.Under the same conditions, siRNAβ1 hindered the tumor growth to a lesser extent when compared to β1DE.Different efficiency in vivo may result rather from the lower stability of siRNA in extra-and intracellular compartments than from the distinctive duration of gene-silencing by both inhibitors.For proliferating tumor cells, gene-silencing produced by siRNA lasts for less than a week because of the dilution of siRNAs that occurs with each cell division as the RISC and the siRNAs bound to it get divided between daughter cells.In both xenograft models, the inhibition produced after a week since the administration of siRNAβ1 and β1DE equaled to 45.2% or 64.5% and 92.2% or 83.8% respectively, when the tumor growth of colon cancer or prostate cancer xenografts was tested.It suggests a significantly higher inhibitory efficiency of β1DE when compared to that of siRNAβ1 produced in xenografted mice.
Taking into consideration all collected data, our results confirm that siRNA and DNAzymes can effectively downregulate the β1 integrin expression with great specificity at the protein synthesis level.The siRNA appears to be quantitatively more efficient with more durable effects in the cell culture, however, the DNAzyme produces more extensive inhibition of tumor growth during in vivo experiments.

Figure 1 .
Figure 1.Inhibition of β1 integrin synthesis in PC3 and HT29 cells.PC3 and HT29 cells were incubated with siRNAβ1 or siRNA C (1.67 µg each) for 48 h, and with hDEβ1 or hDE C (1.25 µg each) for 18 h.Then, the β1 level was measured by Western immunoblotting.Protein bands were scanned, and data presented as the mean ± S.D. was calculated from three separate experiments (A1, B1).Surface expression of β1 integrin subunit in PC3 (A2) and HT29 cells (B2) was analyzed by FACS.It was influenced and decreased by both inhibitors compared to control.(*P < 0.05; **P < 0.001).Data are expressed as % of the control untreated PC3 or HT29 cells.

Figure 2 .
Figure 2. Effects of siRNAβ1 and DEβ1 on cell adhesion and invasiveness.The adhesion of PC3 (A, B) or HT29 (D, E) after transfection with siRNAβ1 (6.67 µg) and hDEβ1 (1.25 µg) was evaluated using plastic wells coated with fibronectin (A, D) or collagen (B, E).The level of adhesion was determined and compared with that of control cells.C and F show the effects of siRNAβ1 and hDEβ1 on invasive properties of PC3 and HT29.Treated cancer cells were allowed to invade Matrigel™ and migrate into the lower part of the filter.The number of invasive cells was expressed in relation to control cells treated with lipofectamine only.

Figure 4 .
Figure 4. Effects of siRNAβ1 and DEβ1on tumor growth of human colon cancer xenografts.HT29 cells were injected into female BALB/cA nude mice to develop solid tumors.The experiment was performed as described in Fig. 3.Both siRNAβ1 and mDEβ1 significantly reduced the volume of the adenocarcinoma tumor (A, B).In both panels inserts show tumors after administering siRNAc or mDEc, and siRNAβ1 or mDEβ1.Panel C shows the quantification of CD34 positive microvessels with lumen in HT29 tumor cross sections.**P < 0.01.