| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
by the American Institute of Ultrasound in Medicine J Ultrasound Med 23:1569-1582 • 0278-4297 Bioeffects of Low-Frequency Ultrasonic Gene Delivery and Safety on Cell Membrane Permeability ControlNational Medical Instrument Special Laboratory, Life and Science Technological School, Xi’an Jiaotong University, Xi’an, China (W.W., B.Z.-z., Z.Q.-w., M.Y.-l.); and Medical Experimental Center, Lanzhou Medical College, Lanzhou, Gansu, China (W.W., W.Y.-j.). Address correspondence and reprint requests to Wang Wei, Education Center of Modern Technology, Lanzhou Medical College, Lanzhou, Gansu 730000, China. E-mail: wangwei{at}mail.lazmc.edu.cn.
Objective. To develop a novel method of ultrasonic naked gene delivery (UNGD); to examine the relationship between optimal parameters of ultrasound exposure and cell membrane permeability, enzymes, and free radicals; and to find optimal control parameters that were realizable, reliable, and noncytotoxic for use in gene therapy. Methods. Suspensions of chicken, rabbit, and rat red blood cells and S180 cells were exposed to a calibrated ultrasonic field with different parameters in both the still and flowing states to obtain optimal parameters for UNGD. The optimal parameters then were used to implement UNGD. We examined morphologic characteristics, membrane permeability, enzymes, free radicals, naked gene expression efficiency, cell damage threshold, and cell viability by laser scanning confocal microscopy, fluorescent microscopy, flow cytometry, and spectrophotometry. Results. Green fluorescent protein (GFP) as a reporter was delivered into S180 cells under the optimal parameters without cell damage or cytotoxicity. The transfection rate (mean ± SD) was approximately 35.83% ± 2.53% (n = 6) in viable cells, and cell viability was 90.17% ± 1.47% (n = 6). The intensity of GFP expression with UNGD showed a higher fluorescent peak over both an adeno-associated virus vector–GFP group and a control group (P < .001). Additionally, malondialdehyde, hydroxyl free radicals, alkaline phosphatase, and acid phosphatase displayed an S-shaped growth model (r = 0.98 ± 0.01) in response to permeability and morphologic alteration. Conclusions. Under optimal conditions, low-frequency ultrasound can safely deliver naked genes into cells without causing cell damage. The analytical results indicate that, except for subcavitation, free radical products are responsible for bioeffects in gene delivery. The constant E of energy deposition at 90% cell viability is the optimal control factor, and 80% viability represents the damage threshold. Optimal gene uptake by cells and safety depend on E. Constant E can be applied to control the gene delivery effect in combination with other parameters.
Key Words: drug delivery • ultrasound bioeffects • ultrasound gene delivery Abbreviations: ACP, acid phosphatase • AKP, alkaline phosphatase • AVV, adeno-associated virus vector • CLSM, confocal laser scanning microscopy • FCM, flow cytometry • FM, fluorescence microscopy • GFP, green fluorescent protein • MDA, malondialdehyde • OH•, hydroxyl free radicals • PBS, phosphate-buffered saline • SEM, scanning electron microscopy • SOD, superoxide dismutase • TUET, total ultrasound exposure time • UE, ultrasound exposure • UNGD, ultrasonic naked gene delivery • UP, ultrasonic pressure
Ultrasound is indispensable in diagnostic medicine and has established therapeutic applications. Biological effects of ultrasound have been observed since effects on fish were observed during its early use, and ultrasound can damage cells and tissues. Bioeffects mechanisms can be broadly characterized as resulting from either thermal or nonthermal mechanisms. The nonthermal mechanisms include radiation force effects, bulk streaming of liquids, and cavitation. Heat and cavitation are generally considered the 2 most important mechanisms of damage from ultrasound.1 Ultrasound can further increase the porosity of the cell membrane, incorporate large external molecules into the cell, and then achieve cell healing.2–4 The bioeffects of ultrasound can be used to deliver drugs and genes into cells and also to improve cell uptake, both of which can be developed for gene therapy.5–9 Ultrasound can help deliver DNA to specific areas of the body and can track its progress with contrast agents injected with microbubbles of genetic material.10–12 The microbubbles implode when exposed to ultrasound, fracturing the cell walls. This phenomenon enables the genetic material to enter cells. The cells then heal and express the gene. Researchers are also applying ultrasound to targeted drug delivery via a process known as sonophoresis.13,14 This technique uses ultrasound rather than needles to deliver drugs such as insulin and interferon directly through skin. Ultrasound can open tiny holes in the cell membrane, making the cells temporarily permeable in localized areas and improving drug penetration.15 This phenomenon improves drug effectiveness, reduces dosage requirements and toxicity, and enables clinicians to deliver drugs to specific areas of the body for localized treatment. Ultrasound opened a new approach to implementing gene transfection. Cells could be transfected so that they expressed a protein of an infectious agent, to which the body then developed protective immunity. Some scientists have tried to use specially created DNA codes to treat cancerous tumors via gene therapy. However, progress has been slow because of the lack of an ideal method for delivering therapeutic genes into tumor cells.16,17 Ultrasonic gene delivery thus offers the potential to provide an improvement on existing solutions. This study attempted to apply the bioeffects of cell membrane porosity produced by low-frequency ultrasound to deliver naked genes into cells, to examine the safety of ultrasound exposure (UE), and to determine the threshold and control parameters of cell membrane permeability. Furthermore, changes in enzymes, morphologic characteristics, and other damage factors during low-frequency ultrasound-mediated naked gene delivery were assessed.
Preparation of the Cell Sample Freshly drawn blood from rabbits, rats, and chickens was mixed with an anticoagulant and stored at 4°C before use. Red blood cells were collected by centrifugation (Heraeus Biofuge 15R; Kendro Laboratory Products GmbH, Langenselbold, Germany; 400 x g, 8 minutes, 4°C), washed 3 times with phosphate-buffered saline (PBS, pH 7.4; Sigma-Aldrich Corp, St Louis, MO), and finally suspended in PBS at a concentration of 10% by volume. The cell suspension was stored at 4°C and then gently diluted to 2 x 106 cells/mL before the experiment. In the still experiments, 4 mL of the red cell suspension was added to a 10-mL airtight polypropylene centrifuge tube to prepare UE. In the flowing state experiments, 20 mL of a rat red blood cell suspension was added to a 25-mL airtight polypropylene centrifuge tube to prepare UE in the flowing state; the flowing state was produced by a small extracorporeal circulation pump (WSQ-A wheel infusion pump; Suzhou Medical Instrument Company, Jiangsu, China). After the cell sample was added to the tube, a rubber stopper was carefully inserted into the tube, and the air was pumped out with a medical injector to minimize the air effect. S180 tumor cells were inoculated into the abdomens of small white mice. After 4 days, ascites were collected by abdominocentesis, and cells were harvested by centrifugation (Heraeus Biofuge 15R; 400 x g, 6 minutes, 4°C) and washed 4 times with PBS. The suspension of S180 cells was cultured in a humidified atmosphere comprising 95% air and 5% carbon dioxide at 37°C in RPMI-1640 media and supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum and penicillin-streptomycin at a concentration of 100 µg/mL (Sigma-Aldrich Corp). One day later, cells were collected by centrifugation (Heraeus Biofuge 15R; 200 x g, 5 minutes, 4°C) and then suspended in RPMI-1640 media at a cell concentration of 106 cells/mL in preparation for subsequent use. In the experiments, 4 mL of the cell suspension was added to a 5-mL polypropylene centrifuge tube and sealed by the method described above to prepare for UE. After UE, the sample was added to 5% (vol/vol) heat-inactivated fetal bovine serum before being cultured for approximately 48 hours and observed.
Ultrasound Application and Calibration A double-layer acrylic water tank (inner size, 20 x 15 x 15 cm) for UE was constructed, and the transducer was fixed on its wall. The outer layer of the tank was a cooling chamber containing cooling water, which minimized the thermal effects on cells by maintaining the sample at 24 ± 1°C. The foam material used to absorb ultrasound was plastered on the tank wall to eliminate the ultrasonic reflection and maintain a stable ultrasonic field. The tank base was sealed and filled with filtered, deionized water. The cell sample sealed in a tube or in a rubber tube (inner diameter, 2 mm) from the infusion pump was positioned directly in front of the transducer facing the axis at 4 cm from the transducer for UE. The experimental conditions must be carefully specified to produce valid results; thus, the calibration of acoustic pressure is important. For a traveling wave from a planar transducer at high frequency, ultrasound is primarily described as intensity. Because UNGD needed a stable sound field and low pressure to avoid irreversible damage to cells, the incident pressure was considered the most suitable for describing the ultrasonic field in this study. The incident pressure is defined as the pressure level that exists in subcavitation or without cavitation and supracavitation, which occurs primarily at the driving frequency (35.1 kHz).2 This definition is useful at low acoustic pressure levels, at which the measured pressure equals the incident pressure. This acoustic pressure level was assumed to apply in this study, and the bioeffects of UNGD were assumed to occur in subcavitation because the pressure level in this work was less than 120 kPa.
Because of the cylindrical configuration of the proposed transducer and the use of a low frequency with a wavelength comparable with the apparatus diameter, 5/4.8 cm = 1.1, the ultrasonic wave is not a well-formed traveling wave, and approximates a spherically outgoing wave. The ultrasonic pressure (UP) at which the cells are exposed does not equal the value on the source surface. Thus, the average incident UP to which the cells were exposed was measured with a calibrated hydrophone (Reson, Goleta, CA), at room temperature (24 ± 1°C). Furthermore, the signals of the hydrophone were analyzed with an oscilloscope, and UP was calculated from the calibration curve provided by the manufacturer. The spatial-average temporal-average intensity was calculated to provide a reference for observing energy flux deposition, despite only partially accounting for the UE at a higher sound level in cavitation or supracaviation.2 The intensity of the spherical wave was calculated by the following equation, I = P2/
Evaluation of Delivery Parameters In the still experiment, red blood cell samples of rabbit, rat, and chicken were individually exposed to the ultrasonic field with different TUET and UP values. After exposure, the morphologic characteristics were observed, and cell viability was checked through cell counting by microscopy (DM RXA; Leica, Nussloch, Germany). In the flowing state experiment, 20 mL of the rat red blood cell suspension was added to the pump to circulate through a 2-mm rubber tube located at the exposed point of the tube. The porous, acanthoid cells and cell viability were evaluated. Application of these methods permitted estimation of the degree of damage to the cell membrane, and hemoglobin analysis was used to optimize the permeable parameters. To validate the change of cell structure caused by UE, the value of hemoglobin in the supernatant fluid was assumed to increase when cell membrane permeability increased with optimal cell viability, for example, 90%. The release of hemoglobin in the supernatant had 2 causes: the increase in cell membrane permeability due to reversible instantly recovered porosity on the cell membrane and irreversible breakage of the cell membrane. According to the prestudy state, when the cell viability was kept to greater than 90%, the porosity on the cell membrane could instantly recover with much less cell breakage, and the hemoglobin in the supernatant mostly permeated through the membrane. When the cell viability was less than 85%, the cell breakage increased, and the hemoglobin the in the supernatant was due to the cell breakage and the increase in cell membrane permeability. Combined with the morphologic findings, the value of hemoglobin in the supernatant could be used to indirectly assess the degree of cell membrane permeability. After exposure with different TUET and UP values, the supernatant liquid of red blood cells was collected via centrifugation (200 x g, 5 minutes, 4°C). The optical value of hemoglobin in the supernatant fluid was measured at 575 nm with a spectrophotometer (DU-64; Beckman Coulter, Inc, Fullerton, CA) by a standard method15 to assess the degree of membrane permeability. The control group without UE was assessed by the same approach. To validate the threshold of cell trauma and cell viability, the morphologic characteristics of S180 cells were estimated immediately after exposure by addition of 10 µL of 0.01% trypan blue in samples, with microscopy and scanning electron microscopy (SEM; JSM-5600; JEOL, Tokyo, Japan). The viability and damage of S180 cells were estimated via flow cytometry (FCM; Epxcs XL; Beckman Coulter, Inc) using the propidium iodide dye16 and also by confocal laser scanning microscopy (CLSM) using the acridine orange method. The optimal UP and TUET were determined based on the analysis of morphologic characteristics, enzymes, free radicals, and the value of the hemoglobin in the supernatant fluid. The control group without UE was evaluated by the same method. The cell morphologic characteristics were observed at 1 to 24 minutes of TUET. To assess the cell function, the exposed S180 mouse tumor cells were reinoculated into the abdomens of the small white mice. Four days later, ascites were collected by abdominocentesis and washed 4 times with PBS by centrifugation to observe the results of cell viability and green fluorescent protein (GFP) expression.
Reporter Gene
Gene Delivery and Expression
Enzymes and Free Radicals
Statistical Analysis
Optimal Parameter Assessment of Ultrasound Bioeffects For the purpose of UNGD, the parameters of UE must produce maximal permeability of the cell membrane and the least cell damage and breakage. The point of 90% cell viability was considered the observation point for optimal parameters, which was near the normal physiologic state. When cell viability was less than 80%, cell damage and breakage occurred. Then the UP and TUET at 90% cell viability were determined and used for observation. The porous, acanthoid, and abnormal red blood cells served as the evaluation objects. The values of UP and TUET at the point of 90% cell viability were considered the optimal parameters, which were described as a pair of parameters, UP/TUET: 102.5 kPa/5 min, 94.9 kPa/6 min, 86.6 kPa/7 min, 77.5 kPa/9 min, and 67.1 kPa/12 min in this study.
Figures 1
To validate the optimal parameters above, chicken red blood cells were exposed to different UP values, 102.5, 95.1, 86.6, and 77.5 kPa, with different TUET values. The results showed that 102.5 kPa/5 min, 95.1 kPa/6 min, 86.6 kPa/7 min, 77.5 kpa/9 min, and 67.1 kPa/12 min were the optimal parameters for achieving 90% cell viability and UNGD. Findings from chicken red blood cells closely reflected the above results with the same parameters: 89.8% ± 0.84% (n = 5) cell viability. Figure 4  illustrates that 90% cell viability was achieved with optimal parameters of 102.5 kPa/5 min, 95.1 kPa/6 min, 86.6 kPa/7 min, 77.5 kpa/9 min, and 67.1 kPa/12 min. The results of other validating tests for optimal parameters using S180 cells were similar to those from rabbit and chicken red blood cells at 102.5 kPa/5 min (Figure 5; P > .05,  2 test); the cell viability was 90.8% ± 1.64% (n = 5).
These data showed that cell viability tended to decrease with increasing TUET and UP, whereas the count of viable porous, abnormal, and acanthoid cells simultaneously increased. With the observation point fixed at 90% cell viability, TUET decreased with increasing UP. Conversely, TUET increased with decreasing UP. The energy flux of UE appeared to approach a constant E, which was related to UP, cell membrane permeability, and TUET. At 90% cell viability, the constant E was theorized to be the optimal energy deposition parameter of cell membrane permeability or naked gene delivery: E = TUET x P2/  c. With the optimal energy deposition parameter E at 90% cell viability, the observational results of cell viability and viable porous, abnormal, and acanthoid cells from different cell types were very similar (P > .05, 2 test) and consistent with one another.
Another test was used to assess the safety of the optimal parameters. The hemoglobin in the rat red blood cell supernatant fluid was assayed. The optimal parameters of 102.5 kPa/5 min, 94.9 kPa/6 min, 86.6 kPa/7 min, 77.5 kPa/9 min, and 67.1 kPa/12 min were validated on the basis of the change in hemoglobin permeability of the rat red blood cell membranes. The increase in hemoglobin in the supernatant fluid could indirectly indicate a change in the permeability of the cell membrane. The optical density of hemoglobin in the supernatant fluid began to increase rapidly after 102.5 kPa/5 min in the static condition. The change in hemoglobin in the supernatant fluid of the rabbit red blood cells in the flowing state of the red cells was also observed in a simulating condition of the blood flow speed. The experimental results shown in Figure 6
Gene Delivery and Expression This work used the optimal parameter of 102.5 kPa/5 min to deliver naked plasmid GFP genes into S180 cells. As mentioned previously, the experiment was divided into 4 groups: A through D. Forty-eight hours after UNGD, CLSM and FM were used to obtain the results shown in Figures 7 and 8, respectively. Figure 7 shows that the naked plasmid GFP gene was successfully delivered into S180 cells, which showed strong fluorescence. Moreover, Figure 8 shows that group A also had successful delivery of the naked plasmid GFP gene into S180 cells and additionally showed a higher fluorescent intensity than group B (AVV-GFP). The GFP expression in group C was extremely weak. Histograms of group A in Figure 9 show 3 high-fluorescence peaks with a lowfluorescence background, which indicate extensive naked plasmid GFP uptake within each cell population. Peak A was located in a high-fluorescence area, standing for the highest intensity of GFP expression, and showed more gene uptake in cells; peak B indicates the middle intensity of GFP expression; and peak C illustrates the weakest intensity of GFP expression. Group B had a broad distribution of low-fluorescence background with no peak, which explained the small amount of GFP uptake within each cell population and the higher transfection rate. Although the cell viability was higher with the AVV-GFP vector than with UNGD (P < .05, t test), approximately 98.17% ± 0.75% (n = 6; Figure 8B), the histogram analysis revealed that the intensity of GFP expression yielded by UNGD clearly exceeded the transfection of the AVV-GFP vector at 104 v.g (group B; Figure 10; P < .001, t test; group C; Figure 9; P < .001, t test). Green fluorescent protein–positive and –negative cells on FM were counted to estimate the transfection rate. The transfection rate was 40% ± 2.2% (n = 6) for the AVV-GFP vector and 35.83% ± 2.53% (n = 6) for UNGD, with 90.17% ± 1.47% (n = 6) cell viability. Figure 9 illustrates the low-fluorescence background, which could be considered to represent the transfection rate, and the high-fluorescence peak could explain the expression intensity. The analytical results agreed with the count of positive and negative cells.
Safety Evaluation After Gene Delivery To estimate cell trauma after UNGD, cell trauma on SEM (magnification x5000, 20 kV), CLSM, and FCM at 102.5 kPa was observed. No trace of irreversible cell membrane destruction existed at less than 3 minutes’ TUET, as shown by the absence of damage peaks in FCM histograms. At 5 minutes’ TUET, a trace of recovered cell membrane destruction (Figure 10A  ) was noted, with a 2.1% damage peak in the FCM histogram. The cell membrane was basically slick and had a recovered trace without breakage. At 8 minutes’ TUET, uncovered pores were identified with a 5.3% damage peak in the FCM histogram (Figure 10B). The cell membrane was rougher than in Figure 10B and showed an evident uncovered hole and cell membrane breakage. The findings from CLSM at 8 minutes’ TUET showed that a pyknotic nucleus (Figure 11A) and karyostenosis (Figure 11B) appeared in numerous cells, indicating nucleus trauma. Simultaneously, the exposed S180 cells were reinoculated into the abdomens of the small white mice. After 4 days, appreciable growth occurred, and 24.1% ± 1.2% (n = 5) of S180 cells showed GFP expression.
Assessment of Enzymes and Free Radicals After UNGD with 102.5 kPa/5 min, MDA, SOD, OH•, AKP, and ACP were analyzed. Malondialdehyde began to increase at around 8 minutes’ TUET (2.8 ± 0.2 to 7.38 ± 0.14 nmol/mL; n = 5; Figure 12 ); the increase responded to the damage threshold of the cell membrane. Superoxide dismutase decreased with increasing energy deposition of UE (68.71 ± 3.1 to 10.26 ± 2.3 nU/g; n = 5; Figure 13). The hydroxyl free radicals tended to increase (299.93 ± 7.43 to 422 ± 6.82 U/mL; n = 5; Figure 14), which was consistent with the change in MDA and SOD. Alkaline phosphatase (33.75 ± 0.42 to 34.79 ± 0.63 U/g; n = 5; Figure 15) and ACP (4.97 ± 0.53 to 8.92 ± 0.76 U/g; n = 5; Figure 16) displayed an increase that corresponded to the increased energy deposition. With the exception of SOD, most results showed the tendency of an S-shaped growth model, y = (ab + cxd)/(b + xd), where y was the amount of MDA, AKP, ACP, and OH•, and x was TUET. Malondialdehyde, SOD, AKP, ACP, and OH• are the indicators of cell damage, in which the increase in MDA and decrease in SOD were due directly to the rise of OH•, and AKP and ACP increased after cell damage had occurred. In the sequence of TUET, the increase in OH• occurred earlier, and the changes in MDA, SOD, AKP, and ACP followed. The increases in AKP and ACP were the results of cell damage. These data show that OH• was an important element in causing the increase in cell membrane permeability.
Genetic diseases can be cured through proper gene replacement and manipulation.6 Numerous attempts have already been made to overcome the relatively low transfection efficiency of naked DNA gene transfer, including the use of cationic lipids and polymers18–20 and the application of electric pulses to DNA injection sites.17
Here we describe an ultrasonic method for efficiently delivering naked genes into cells. The proposed method has been used for cancer hyperthermia therapy and gene transfection both in vitro and in vivo. To understand the gene transfection mechanism and explore its potential application for cancer gene therapy, this experiment examined the optimal ultrasonic parameters and morphologic changes for gene transfection efficiency. The findings indicated that maximizing delivery efficiency involved increasing membrane porosity without causing irreversible cell damage. Ultrasound exposure increased the permeability of the cell membrane and thus improved uptake of naked GFP genes with chosen optimal parameters. The analytical results suggested that UE might be more likely to achieve the optimal balance between membrane pore formation and cell resealing. The safety of UNGD was monitored, and the parameters at 90% cell viability were found to be optimal. Naked genes could be delivered into cells without causing cell damage either in vitro or in vivo. Importantly, in this study, we observed that the cell membrane permeability under UE showed a nonlinear energy deposition procedure. At 90% cell viability in different cells, the energy deposition of UE was found to apparently approach an energy deposition constant E (5.97 ± 0.55 [n = 6] in this study); E is useful for controlling bioeffects when combined with measurement of half the applied frequency.2 When (TUET x P2/
This investigation revealed that permeability increased strongly as a function of UP and TUET, which indicated that the selection of appropriate UP and TUET was important for optimizing the cell membrane permeability. The results from Figures 12
Previous studies considered cavitation as the mechanism responsible for increased cell membrane permeability. Cavitation involves the creation and oscillation of gas bubbles in a liquid. During the low-pressure portion of ultrasonic waves, dissolved gas and vaporized liquid can form gas bubbles. These bubbles then shrink or expand, oscillating in response to high- and low-pressure portions of the ultrasonic wave. This phenomenon is known as stable cavitation or subcavitation. Another type of cavitation is termed transient cavitation and occurs under higher acoustic pressure, at which bubbles violently implode after a few cycles.13 Implosion can cause numerous effects, including a transient increase of hundreds of degrees Celsius in the local temperature, an increase of hundreds of atmospheres in the local pressure, light emission owing to sonoluminescence, and the launch of a high-velocity liquid microjet. The bioeffects of UNGD in this work that were below that cell damage threshold resulted from stable cavitation or subcavitation, which yielded optimal gene delivery parameters via washing and eroding behavior in the cell membrane with an appropriate increase in free radicals. In addition to subcavitaion, the increase in OH• is another main reason for increased cell membrane permeability. The cell damage threshold may result from transient cavitation, which sharply increases the cell membrane permeability to cause irreversible cell damage with a rapid increase in OH•. Conversely, the increase in OH• may cause apoptosis of tumor cells (Figures 10 The advantages of UNGD are that the gene delivery efficiency is not strongly influenced by the cell growth stage, and the gene transfection can be performed in complete medium with serum and in the body. These advantages, along with the ability to focus ultrasound on most areas of the body, could provide a novel approach for numerous medical applications such as cancer gene therapy. Although previous studies have achieved enhanced effects in gene transfection,7,19 we think that this study represents an important demonstration of a multiple-cell-type correlation, multiple parameters, multiple factors, and multiple observations for optimal UNGD control parameters. In conclusion, this study shows that low-frequency ultrasound can safely deliver naked genes into cells without damaging cell function under optimal parameters. Optimal gene uptake and expression both depend on the energy deposition parameter E at 90% cell viability. The constant E can be applied along with other parameters to control bioeffects. The analytical results suggest that subcavitation and free radicals are responsible for the bioeffects associated with gene delivery. The amount of local free radicals could be considered another value for monitoring the change in cell membrane permeability. The results of this study can be used to develop novel clinical gene therapy and cancer therapy systems.
Received March 2, 2004, from the National Medical Instrument Special Laboratory, Life and Science Technological School, Xi’an Jiaotong University, Xi’an, China (W.W., B.Z.-z., Z.Q.-w., M.Y.-l.); and Medical Experimental Center, Lanzhou Medical College, Lanzhou, Gansu, China (W.W., W.Y.-j.). Revision requested April 1, 2004. Revised manuscript accepted for publication July 29, 2004. We thank Bai Decheng and the Medical Experimental Center of Lanzhou Medical College and their colleagues for support and contributions. This work was supported by National Nature Science Foundation of China grant 60271022 and the Doctorial Foundations of Xi’an Jiaotong.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Top
Abstract
Introduction
