Effects of formulation and operating variables on Zanamivir dry powder inhalation characteristics and aerosolization performance

Abstract

The objective of this study was to investigate the influence of formulation and operating variables on the physical characteristics and aerosolization performance of zanamivir spary-dried powders for inhalation. Spray-dried samples of zanamivir, zanamivir/mannitol and zanamivir/mannitol/leucine were prepared from their corresponding aqueous solutions under the same conditions to study the influence of the composition, and zanamivir/mannitol/leucine (1/1/3 by weight) formulation was used for investigation of the effect of the preparation process. Dry powders were characterized afterwards for different physical properties, including morphology, particle size, flowability, density and moisture absorption. The in vitro deposition was also evaluated after the aerosolization of powders at 100 L min⁻¹ via the Aerolizer® into a Next Generation Impactor (NGI). The highest FPF of 41.40 ± 1.1% was obtained with a zanamivir/mannitol/leucine ratio of 1/1/3, which had an average D_g of 3.11 ± 0.13 μm and an angle of repose of 36°± 1. It was found that the influence of the preparation process on zanamivir spary-dried powders characteristics and aerosolization properties was relatively small, but the influence of the composition was relatively large. Optimization of DPI can be achieved by selecting the most appropriate formulation and preparation process.

• Dry powder inhalations
• excipient
• in vitro powder deposition
• spray-drying
• Zanamivir

Introduction

The development of dry powder inhalations (DPIs) formulations continues to be a scientific challenge. Usually a DPI formulation consists of API and excipients. Ideally it would be API alone but because particle/particle interactions increase with decreasing size it is often not feasible to process, de-aggregate and aerosolize the typically fine API powder. To get around this, formulators use excipient particles as carriers. These excipients have an effect on the physical properties of powders, including their morphology, density, hygroscopic nature, particle size and distribution. These properties can directly influence the degree of pulmonary deposition of drug (Maa et al., 1997). So, for optimal, functional and therapeutic performance of dry powders for inhalation, the selection of appropriate excipients and the control of formulation properties are of critical importance.

Unfortunately, the choice of excipients is very limited for pulmonary applications. Currently, nearly all DPI products (Relenza®, Seretide®, Spiriva®, Symbicort®, Beclophar®, Flixotide® and so on) already on the market rely on lactose as a carrier material. The advantages of lactose are its well-investigated toxicity profile, its broad availability and its relatively low price (Steckel et al., 2004). However, the use of lactose has some disadvantages, such as its sugar-associated reducing function that may interact with functional groups of drugs such as budesonide or peptides and proteins (Hickey et al., 2007). Furthermore, lactose can be produced with traces of their bovine source (proteins), which carries a theoretical risk of transmissible spongiform encephalopathy (TSE) (European Commission, 2002). Also, discussions regarding the endotoxin content and the necessity for a specification limit are ongoing (Food and Drug Administration, 1998). Moreover, lactose rendered the powder more cohesive and, hence, reduced the flowability. It is also prone to absorb moisture. It is therefore reasonable to look for alternative carriers that still possess the beneficial properties but overcome the aforementioned drawbacks of lactose.

Several alternative excipients have been tested, such as mannitol and leucine. Mannitol is widely used in pharmaceutical formulations and food products. It may be an attractive alternative carrier to lactose because it does not have the reducing effect and is less hygroscopic. Moreover, mannitol gives a highly sweet after-taste which could be used to confirm to the patient that a dose has been successfully administered, and it has the capacity to provide a high fine particle dose of incorporated drug upon powder aerosolization (Kaialy et al., 2010). In addition, mannitol was shown to stimulate the reserve capacity of the mucociliary system (Daviskas et al., 2008) and to enhance clearance of mucus (Jaques et al., 2008). Therefore, mannitol, which is a sugar alcohol previously inhaled for diagnosis of bronchial hyperresponsiveness, has emerged as a promising excipient (Pilcer et al., 2010). As leucine is an endogenous substance, it might not present major risk of toxicity to the lungs. The addition of leucine to formulations for inhalation obtained by spray-drying has been demonstrated to significantly improve the in vitro deposition profile of a dry powder. Moreover, leucine has been shown to decrease hygroscopicity and improve surface activity and charge density of particles (Li et al., 2005; Arakawa et al., 2007). The influence of the amount of leucine in the powder has been demonstrated in various papers. The consensus is that 10–20% (w/w) of leucine in spray-dried solutions of water gives optimal aerosolization characteristics of powders containing peptides or sodium cromoglycate (Chew et al., 2005).

Three methods, that is, air-jet milling, supercritical fluid precipitation (SCFP), and spray-drying, are usually used to produce dry powders for inhalation. However, the micronization process using air-jet milling is extremely inefficient (Donovan et al., 2011). Although SCFP is attractive, many compounds are not soluble in the available supercritical fluids, and therefore, SCFP is restricted to marketed DPI products (Okamoto et al., 2008). At the start of the 1990s, spray-drying was developed as an alternative method to micronization. Spray-drying, as a single-step processing technique with a great control over the particle characteristics (Stahl et al., 2002; Asada et al., 2004; Chan et al., 2004; Cook et al., 2005) was applied for powder preparation.

The drug used in this study was zanamivir which is marketed by GlaxoSmithkline. It is a potent and highly selective inhibitor of the influenza A and B virus neuraminidase (Cox et al., 2001). When administered directly to the human respiratory tract, zanamivir was shown to have potent antiviral effects. Relenza® (zanamivir for inhalation) is comprised of four regularly spaced double-foil blisters, each of which contains a powder mixture of 5 mg of zanamivir and 20 mg of lactose. Although the dry powder (Relenza®) presented good aerosolization properties, drawbacks are linked to lactose. Therefore, the development of a new inhalable zanamivir formulation would potentially overcome some of the drawbacks of the Relenza®.

So far there have been few reports concerning the effects of excipients on the properties of inhalation dry powders, especially as far as zanamivir is concerned. When formulated with an appropriate composition to produce adequate physical characteristics, the powders exhibited excellent aerosolization properties. The present investigation focused on the physical characteristics and aerosolization performance of zanamivir inhalation spray-dried powders for inhalation with different formulation composition. Powders were prepared by spray-drying using the excipients mannitol and leucine, which were selected as they are recognized as safe and widely used excipients in previous DPIs studies (Bosquillon et al., 2004). The physical characteristics of the powders (i.e. particle size, tapped density, morphology and moisture) were determined. In vitro aerosolization properties were assessed with an Aerolizer® inhaler device in a Next Generation Impactor (NGI).

Experimental materials

Zanamivir (purity: 99.11%) (Haoxin pharmaceutical Company), mannitol (Tianjin Bodi Chemical Reagent Company) and leucine (Shanghai Kangda Amino acids Factory) were obtained from the suppliers indicated. Relenza® (zanamivir) with Diskhaler® inhalation device (GlaxoSmithKline) was purchased commercially. All other materials or solvents were of analytical or chromatographic grade and were used as obtained commercially.

Methods

Preparation of spray-dried powders for inhalation

Dry powders were made with zanamivir, zanamivir/mannitol and zanamivir/mannitol/leucine (Table 1) by spray-drying in order to evaluate the influence of the formulation components on the powder characteristics and in vitro powder deposition. They were weighed and added to a suitable volume of distilled water. Then obtained solution was diluted to volume and passed through a 0.45 μm cellulose acetate filter. The final solution was spray-dried using an SD-06AG spray dryer (Lab-Plant, UK), under the following conditions: inlet temperature, 130 °C; atomization pressure, 18 kPa; feed flow rate, 2 mL min⁻¹; air flow rate, 0.6 m³min⁻¹. The solution was constantly stirred at 25 °C throughout the spray-drying process.

Table 1. Formulation design.

Batch	Composition	Concentration of spray-drying solution (g/100 mL)
1	Zanamivir	1
2	Zanamivir/mannitol	1/4
3	Zanamivir/mannitol/leucine	1/3/1
4	Zanamivir/mannitol/leucine	1/2/2
5	Zanamivir/mannitol/leucine	1/1/3

To estimate the effect of spray-drying conditions on the powder characteristics and in vitro powder deposition, Batch 5 (zanamivir/mannitol/leucine, 1/1/3) was spray-dried under the following conditions (Table 3).

All spray-dried powders were collected and stored in a dessicator (at 30% relative humidity and at room temperature).

Powder characterization

Scanning electron microscopy

Samples were prepared by sputter-coating with gold to a thickness of 10 nm (Quorum model Q150). Images were collected using a scanning election microscopy (S-450, Hitachi, Japan).

Moisture absorption

An experiment was conducted at varying relative humidity (45, 55, 65, 75 and 85% RH) for 24 h at 25 °C in the climatic cabinet (T021, Hengxing, China). After 24 h, the samples that had been spread uniformly in the samples that had been spread uniformly in open Petri dishes and weighed in advance were taken out and weighed again. The moisture absorption profile was expressed as a percentage by the increasing weight at different RH.

Particle size analysis

The geometrical volume mean diameter (D_g) was measured by a light-scattering particle size analyzer (LS800, OMC Corp., China). The spray-dried powders samples were suspended in 0.5% (w/v) Span 80 in cyclohexane solutions and stirred with a magnetic bar at 1000 rpm. The particle size analysis was performed by the WINDOX 3.4 software (Redmond, WA, USA).

Powder tapped density

The powder tapped density (ρ) was obtained by mechanically tapping a measuring cylinder containing a powder sample. After observing the initial volume, the cylinder is mechanically tapped and the volume reading was taken until little further volume change was observed (British Pharmacopoeia Commission, 2005).

Powder flowability

The angle of repose (α) as the most frequently used index for evaluating powder flowability were determined by the method reported (Hickey & Concession, 1997). Powders were poured through a funnel to form a cone-shaped pile which had an angle, α, to the horizontal. The value of α was calculated by measuring the height and radius of the pile. A large angle of repose is indicative of poor flow properties while a small angle of repose indicates a free-flowing powder.

HPLC analysis of zanamivir

The zanamivir concentration was determined with a HPLC (1211, Agilent Technologies Inc., Avondale, PA, USA). Luna SCX HPLC column (250 mm × 4.6 mm, 5 μm, Phenomenex Inc., Torrance, CA, USA) was operated at room temperature. The mobile phase was acetonitrile −0.02 mol L⁻¹ phosphate (60:40 v/v) buffer containing 0.2% triethylamine (pH = 3.5) at a flow rate of 1 mL min⁻¹ and the detection was performed at 233 nm. Calibration curve was generated for each formulation using linear regression over the range of 0.05–0.90 mg mL⁻¹ using five concentrations. The regression coefficient (r²) values were greater than 0.999 for all formulations demonstrating satisfactory linearity.

In vitro powder deposition

In vitro powder deposition of powder was tested using the NGI (Copley Scientific, Nottingham, UK) . Prior to use, all stages were coated with 1% w/v silicone oil in hexane to minimize bouncing. A size 3 gelatin capsule, containing 25 mg of zanamivir spray-dried powders was placed into the sample compartment of an Aerolizer® inhaler, the inhaler was activated, inserted into a United States Pharmacopoeia (USP) throat. The airflow rate was adjusted at 100 L min⁻¹ to produce a pressure drop of 4 kPa over Aerolizer® with an emptied capsule, as recommended by the USP. Accordingly, the test flow duration was set at 2.5 s by the critical flow controller to draw a 4 L simulated inhalation volume from Aerolizer®. In vitro powder deposition of Relenza® was also investigated as a control. Identical to those employed in the NGI experiments, the airflow rate and its duration were set at 90 L min⁻¹ and 2.7 s, respectively, for a pressure drop of 4 kPa over Diskhaler®.

After the run, the capsule, inhaler, adaptor mouthpiece, throat, and NGI stages were washed with 50 mL of mobile phase which was assayed for solute concentration by the HPLC as described above. The fine particle fraction (FPF) was calculated as the ratio of fine particle dose to total loaded dose, expressed as a percentage and corrected for actual zanamivir content in each powder. Each powder was tested in triplicate.

Results and discussion

Influence of formulation composition on powder characteristics

The zanamivir spray-dried powders with different formulations were prepared by a spray-drying method. As shown in Figure 1, the yield of recovered powder varied considerably. The yields of zanamivir/mannitol (Batch 2) and zanamivir/mannitol/leucine preparations (Batch 3, 4, 5) were high, whereas zanamivir (Batch 1) alone exhibited the lowest yield. The increased spray-drying yields seen with the addition of leucine suggest some changes in the powders, such as particle size and cohesive force of the particles, which might contribute to the aerosolization behavior of the powders (You et al., 2007).

Figure 1. The zanamivir spray-drying powders yield of the different formulation.

The SEM micrographs (Figure 2) showed that the formulation components significantly influenced the particle shape and morphology. The addition of leucine resulted in irregularly wrinkled particles (Figure 2c–e). Morphology studies showed an increase in particle corrugation as an effect of leucine presence in spray-dried powders. According to previous report (Lechuga-Ballesteros et al., 2008), during the spray-drying process, the saturation of the lower-soluble component (leucine) may increase faster than that of hydrophilic one (mannitol). This led to the formation of a primary solid shell which collapsed, hence corrugated microparticles were formed. Unlike leucine, mannitol produced more globular and smooth particles (Figure 2b). However, the increasing surface smoothness usually increases adhesion forces between particles as a result of an increased contact area between the interacting species (Boshhiha et al., 2009). Thus, a smooth surface of powders may not be necessary for inhalation. The advantage of wrinkled, hollow particles is that there is less contact between particles, thereby resulting in a better in vitro deposition (Wang et al., 2009).

Figure 2. Scanning electron micrographs showing the particle morphology of zanamivir spray-drying powders at various formulations. (a) The powders prepared with zanamivir, Batch 1; (b) the powders prepared with zanamivir/mannitol (1/4), Batch 2; (c) the powders prepared with zanamivir/mannitol/leucine (1/3/1), Batch 3; (d) the powders prepared with zanamivir/mannitol/leucine (1/2/2), Batch 4; (e) the powders prepared with zanamivir/mannitol/leucine (1/1/3), Batch 5.

The tapped density is an important physical property of dry powder. As demonstrated previously, the tapped density of a powder influences the aerosolization performance of the powder, the latter being improved as density is lowered (Bosquillon et al., 2001). As listed in column 2 of Table 2, the increasing order of powder tapped density is Batch 5 < Batch 4 < Batch 3 < Batch 2 < Relenza®< Batch 1. In a general sense, it could be concluded from this ranking that the proportion of leucine in excipients greatly affects the tapped density of powder. The more leucine incorporated in the formulation, the lighter the particles produced. It was indicated that leucine had significant effects on decreasing the tapped density of spray-dried powders (Lucas et al., 1999). So, the lightest powder (0.32 ± 0.01 g cm⁻³) was Batch 5 (zanamivir/mannitol/leucine, 1/1/3). By contrast, Batch 1 (the excipient-free formulation) had the greatest tapped density of 0.62 ± 0.03 g cm⁻³ followed by Relenza® (0.50 ± 0.01 g cm⁻³).

Table 2. Characteristics of zanamivir spray-drying powders (n = 3).

Batch	ρ (g.cm^−3)	α	Dg (μm)	FPF (%)
1	0.62 ± 0.03	59° ± 1	9.28 ± 0.09	5.43 ± 1.7
2	0.47 ± 0.02	43° ± 1	5.88 ± 0.15	16.33 ± 1.2
3	0.42 ± 0.01	39° ± 2	4.48 ± 0.12	25.62 ± 1.3
4	0.38 ± 0.02	33° ± 2	4.07 ± 0.07	30.62 ± 1.5
5	0.32 ± 0.01	36° ± 1	3.11 ± 0.13	41.40 ± 1.1
Relenza®	0.50 ± 0.01	46° ± 1	3.67 ± 0.17	34.21 ± 0.9

The smaller the angle of repose (α) is, the better the flow property is. As shown in column 3 of Table 2, the increasing order of angle of repose is Batch 4 < Batch 5 < Batch 3 < Batch 2 < Relenza®< Batch 1. The best flowability was observed when the carrier consisted of mannitol/leucine with a ratio of 2: 2 (Batch 4). The angle of repose of the zanamivir spray-dried powders decreased upon addition of leucine and then increased slightly on addition of more leucine. Previous research has shown that leucine is able to act as an anti-adherent in the case of dry powders. It is assumed that leucine interferes with the weak bonding forces, such as Van der Waal’s and Coulomb forces, between the small particles which helps to keep the particles separated and may be thought of as weak links (Chougule et al., 2007). These particular properties of leucine give the powders a better flowability. Nevertheless, higher concentrations of leucine might lead to highly charged particles, thereby adversely affecting the powder flow properties. In contrast to mannitol, lactose rendered the powder more cohesive and, hence, reduced the flowability.

The geometrical volume mean diameter (D_g) of the zanamivir spray-dried powders is shown in column 4 of Table 2. The D_g seen from the scanning electron micrographs appear to be in good agreement with the results of the particle size determination. The D_g of all powders, in descending order, was shown to be Batch 1 > Batch 2 > Batch 3 > Batch 4 > Relenza®> Batch 5. It can be seen that the D_g of powder from Batch 1 to 5 fell as the leucine increased. So it is supposed that the more leucine incorporated in the formulation, the smaller the mean particle size of the spray-dried powders is. As demonstrated previously, leucine is a particularly hydrophobic amino acid and its surfactant-like properties may result in the ability of leucine to migrate to the droplet surface during the rapid drying phase of the spray-drying and, hence, influence the size of the resultant particles (Glińiski et al., 2000).

Moisture will increase powder agglomerates via interparticulate capillary forces, which will reduce the dispersibility of the dry powder (Todo et al., 2003). That is why the characteristics of the excipients should be considered in the design of formulations. The corresponding moisture absorption profiles for the dry powders at 45, 55, 65, 75 and 85% RH are presented in Figure 3. It is clear that, as the RH increases to 85%, there is a marked increase in the weight change of all formulations. Batch 1 (the excipient-free formulation) was found to adsorb a significant amount of water at all of RH levels. While the RH increased to about 85%, the Relenza® showed the highest water affinity (55.2 ± 2.6%). In contrast, the lowest water binding capacity of all formulations was Batch 2 (zanamivir/mannitol, 1/4) followed by Batch 3 (zanamivir/mannitol/leucine, 1/3/1), Batch 4 (zanamivir/mannitol/leucine, 1/2/2) and Batch 5 (zanamivir/mannitol/leucine, 1/1/3). These results show that as the amount of mannitol decreased in formulation, the water binding capacity of the dry powders increased. It can be concluded that mannitol has good antihygroscopic properties for zanamivir for inhalation, because powders with mannitol are less sensitive to moisture at high RH.

Figure 3. Moisture adsorption profile of zanamivir spray-drying powders at 45–85% relative humidity at 25 °C.

Influence of the formulation composition on in vitro powder deposition

Obviously, as shown in column 5 of Table 2, different formulation composition had different effects on physical properties of zanamivir spray-dried powders for inhalation as discussed in earlier sections, which in turn resulted in different effects on the in vitro powder deposition as the same spray-drying operating conditions.

Dealing with the NGI data, it was especially observed that Batch 5 (zanamivir/mannitol/leucine, 1/1/3) gave the greatest FPF (41.40 ± 1.1%) followed by the Relenza® (34.21 ± 0.9%). Batch 1 (the excipient-free formulation) exhibited the poorest FPF of 5.43 ± 1.7% due to the largest particle size and poorest flowability. Indications were that a higher loading of leucine in the formulation contributed to a better aerosolization performance. The influence of leucine on powder in vitro powder deposition has been evaluated previously (Najafabadi et al., 2004; Seville et al., 2007). These researchers reported that the addition of leucine into powder resulted in a more dispersible powder. The leucine can improve the aerosolization performance of powder and was more effective at higher concentrations. Compared with the previous investigations mentioned earlier, in this study, the more leucine incorporated in the formulation, the higher the FPF produced. The contribution of leucine to the improved aerosolization could be due to its surfactant-like and anti-adherent properties. When agglomerates of particles are formed, the addition of leucine reduces the stability of those agglomerates so that they are more likely to break up in the turbulent air stream created on inhalation to form small individual particles which are likely to reach the lower lung (Rabbani et al., 2005).

Influence of the spray-drying conditions on powder characteristics and in vitro powder deposition

The spray-drying conditions showed no significant influence on the powder morphology (figures not shown) and flowability (column 8 of Table 3).

Table 3. Spray-drying variables and characteristics (Batch 5, zanamivir/mannitol/leucine, 1/1/3, n = 3).

No.	Feed flow rate (mL min^−1)	Atomizing pressure (kPa)	Air flow rate (m^3 min^−1)	Inlet temperature (°C)	Yield (%)	ρ (g cm^−3)	α	Dg (μm)	FPF (%)
1	1.5	18	0.6	130	48	0.38 ± 0.01	33° ± 2	3.18 ± 0.20	35.56 ± 2.2
2	2	18	0.6	130	41	0.32 ± 0.01	36° ± 1	3.11 ± 0.13	41.40 ± 1.1
3	2.5	18	0.6	130	23	0.29 ± 0.05	34° ± 1	3.46 ± 0.08	30.83 ± 1.4
4	2	16	0.6	130	51	0.34 ± 0.03	38° ± 1	3.86 ± 0.10	33.12 ± 1.6
5	2	20	0.6	130	18	0.33 ± 0.01	38° ± 1	2.97 ± 0.21	39.11 ± 2.0
6	2	18	0.4	130	39	0.35 ± 0.01	36° ± 1	3.42 ± 0.12	32.56 ± 1.6
7	2	18	0.7	130	40	0.34 ± 0.01	34° ± 1	3.05 ± 0.11	38.56 ± 1.2
8	2	18	0.6	110	20	0.39 ± 0.01	36° ± 2	3.11 ± 0.17	35.16 ± 1.5
9	2	18	0.6	150	47	0.25 ± 0.06	37° ± 1	3.57 ± 0.09	34.03 ± 1.1

As shown in column 6 of Table 3, the yields varied between 18 and 51%. The higher the atomizing pressure, the lower the yield. Decreased atomizing pressure decreases the atomization energy and thus producing enlarged droplets. These droplets dry to larger particles, which are more easily captured through the centrifugal force in the cyclone (Amaro et al., 2011). Moreover, the yield was also increased with increased inlet temperature and decreased feed flow rate. With the rise of inlet temperature, the rate of evaporation increases and spray droplets become rigid enough before contacting the wall of the spray dryer, thus the loss of product decreases and the yield increases. The higher the feed flow rate, the lower the outlet temperature, which leads to an increase of the moisture content. The wet powder easily sticks to the wall of the spray dryer thus decreases the yield (Shoyele & Cawthorne, 2006).

As shown in column 7 of Table 3, the density of the powder decreased with an increase in the inlet temperature and feed flow rate. At high inlet temperature and fast feed flow, as evaporation rates are faster, products dry to a more porous or fragmented structure and implying a lower shrinkage of the droplets, and so a lower density of the powder.

The particle size to the process variables are given in column 9 of Table 3. Two of the main effects are positive, i.e. inlet temperature and feed flow rate which means that when any of these factors increases, larger particles are produced and two are negative, i.e. air flow rate and atomizing pressure giving smaller particles when at higher levels. The higher atomizing pressure, the more energy is supplied for breaking up the liquid into droplets during the atomization step, resulting in smaller droplets (Al-Asheh et al., 2003). However, at high inlet temperature a skin is formed on the outer surface of the spray droplets. When the inner water phase is evaporated the skin is destroyed and the outer surface collapse. Therefore, the increase in particle size might be an effect of increased agglomeration at the higher inlet temperature (Broadhead et al., 1994).

From the above results, the influence of spray-drying conditions on powder properties is relatively small. It is supposed that the spray-dried powder characteristics and aerosolization properties mainly relate to the properties of the spray-drying solution, that is, viscosity and surface tension of the spray-drying solution, which are related to the composition of solution. So, the spray-dried powder characteristics and aerosolization properties are principally affected by formulation composition.

Conclusions

The objective of this research was to evaluate the influence of formulation composition and preparation processes on characteristics and in vitro aerosolization properties of zanamivir spray-dried powders for inhalation. In this study, mannitol and leucine were selected as candidate excipients for zanamivir inhalation. The results revealed that formulation components had significant effects on the physical characteristics of the spray-dried powders, such as morphology, tapped density, particle size, moisture absorption as well as flowability and, thus, influenced their in vitro deposition. The effect of the spray-drying conditions on powder properties is relatively small. It should be stressed again that the difference in aerosolization performance should be caused by a number of complicated factors and a reasonable and accurate analysis is still a challenging area. Encouragingly, among the developed formulations, Batch 5 (zanamivir/mannitol/leucine, 1/1/3) was preferred; the FPF was as high as 41.40 ± 1.1% and, so, it would be anticipated to deposit in the lower regions of the respiratory tract, thereby facilitating systemic delivery of zanamivir.

Declaration of interest

This study was supported by National Natural Science Foundation of China (Grant No. 81202466) and the Important National Science & Technology Specific Projects (Grant No. 2012ZX09301003-001-009) of China.