Eur J Cardiothorac Surg 1998;13:424-430
© 1998 Elsevier Science NL
Transforming growth factor-ß1 and lung allograft fibrosis1
Ahmed El-Gamela,
Mohammed Awadc,
Ewan Sima,
Philip Hasletona,
Nizar Yonana,
Jim Eganb,
Abdul Deiraniyaa,
Ian V. Hutchinsond
a Cardiothoracic Transplant Unit, Wythenshawe Hospital, Manchester M23 9LT, UK
b North West Lung Centre, Wythenshawe Hospital, Manchester, UK
c Tissue Typing Laboratory, Central Manchester Health Trust, Manchester, UK
d School of Biological Sciences, Manchester University, Manchester, UK
Received 30 September 1997;
received in revised form 29 December 1997;
accepted 10 February 1998.
Corresponding author. Tel.: +44 161 9987070; fax: +44 161 9212091; e-mail: aelgamel@ aol.com
 |
Abstract
|
|---|
Objectives: Transforming growth factor ß1 (TGF-ß1) is a potent immunosuppressive cytokine that promotes fibrosis by enhancing the synthesis of extracellular matrix components. The repair process following lung allograft injury is due to rejection or infection replaces lung parenchyma by fibrotic tissue, leading to pulmonary dysfunction. The role of TGF-ß1 in this excessive healing process and increasing the risk of infection is unknown. Methods: We analysed our patient data to investigate the relevance of different factors on allograft fibrosis and its correlation with TGF-ß1. Fibrosis was graded in H and E stained sections. TGF-ß1 genotype was determined in all patients. Results: Patients were aged between 16 and 62 years (mean age of 39.6 years). Procedures were heart/lung (n=32), double lung (n=18), and SLT (n=41). A total of 46 patients had lung allograft fibrosis diagnosed in transbronchial biopsies sections. Patients who had developed interstitial fibrosis had significantly more acute rejection episodes (mean 3.4±2.8) compared with patients without fibrosis (mean 2.1±2.2) (P=0.024). The presence of eosinophils in the interstitium preceded and were associated with the development of fibrosis regardless of the rejection grade (P=0.0001). TGF-ß1 was heavily expressed in sections with fibrosis with a mean score of 6.8±2.9 compared with 2.4±0.6 in sections with no fibrosis (P<0.0001). TGF-ß1 expression correlated positively with fibrosis grades (P<0.0001). The mean survival for patients with a fibrosis score >6 is 892.4±73 days compared with mean survival 427±78 in patients with scores <6 (P=0.0001). Patients who developed fibrosis had homozygous TGF-ß1 genotype that correlates with excessive TGF-ß1 expression (P=0.01). The use of cardiopulmonary bypass was associated with the development of excessive fibrosis (P=0.02), and 7 patients who had severe fibrosis died of septicaemia (17.5%). FEV1 (forced expiratory volume) was significantly higher in patients without fibrosis (1870±111 ml versus 1590±160; P=0.02). Conclusions: The risks of lung allograft fibrosis increases with recurrent rejection, tissue eosinophilia, homozygous TGF-ß1 genotype and the use of bypass machine. Fibrosis was associated with higher mortality and morbidity might be explained by the TGF-ß1 immunosuppressive and fibrotic properties. Immunological strategies to down-regulate TGF-ß1 production might improve survival and function of lung allografts.
Key Words: Transforming growth factor-ß1 • Lung transplantation • Allograft
|
Introduction
|
|---|
In ongoing chronic rejection after lung transplantation, interstitial fibrosis develops. However, little is known about the mechanisms involved
[1]. The mesenchymal and extracellular matrix alterations that occur in acute and chronic rejection of the lung allograft showed that in early rejection, perivascular and peribronchiolar mononuclear infiltrates were associated with basement membrane disruption of the vessels and airways and an ingrowth of muscle-specific actin-, vimentin-positive, desmin-negative spindle cells accompanied by type IV collagenase-positive histiocytes. Subsequent fibrous scarring was manifested by perforation and reduplication of the basement membrane of airways and vessels and dense collagen deposition, primarily type III
[2]. As has been suggested in idiopathic pulmonary fibrosis, the fragmentation of basement membranes and the deposition of collagen IV and laminin by mesenchymal cells in vessels and airways may reflect the irreversible fibrosis responsible for allograft dysfunction
[2]. The biologic basis of pulmonary fibrosis is akin to the process of normal wound-healing, in which injury to normal tissues is followed by inflammation and then repair by the mechanism of scar formation
[3]. However, whereas wound-healing is normally localised in space and confined in time, the fibrosis in lung diseases generally involves the entire organ and is a chronic, ongoing process. Lung allografts are usually exposed to multiple injuries as a result of ischaemia reperfusion, acute rejection, infection or damage of either pulmonary epithelium or capillary endothelium. The lesion could be focal and inhomogeneous or generalised in distribution. Regardless of the initial site of injury, alterations in capillary permeability rapidly ensue
[4]. Leakage of serum proteins, cells, and platelets into the interstitial and alveolar space is apparent several hours after injury. Marked increases in total DNA content soon after injury reflect the dramatic increases in inflammatory cells that are seen in the lungs or noted by bronchoalveolar lavage (BAL)
[5]. The coagulation system is activated early, and fibrinogen in the intraalveolar space is converted to fibrin
[6]. Fibroblasts contain receptors for fibrinogen–fibrin
[7]. These proteins could function as a nascent extracellular substrate for adhesion, since fibroblasts adhering to a fibrin meshwork in vitro synthesise collagen
[7]. Several studies have shown that lavage fluid obtained soon after acute injury contains a significant increase in procoagulant activity and a significant decrease in fibrinolytic activity
[6]. With resolution of injury, these abnormalities resolve
[8] in a majority of subjects. In contrast, in a small proportion of subjects fibrosis ensues and a persistent predisposition to fibrin formation can be documented
[8]. Upregulated expression of extracellular matrix in the alveolar interstitial space leads to alveolar malfunction by thickening of the wall and, thus, is one of the causative factors of respiratory dysfunction in chronic lung graft rejection
[2]. A number of well-characterised cytokines, including transforming growth factor-ß1 (TGF-ß1) have been either found in the injured lung or produced by inflammatory cells removed from the lung
[9]. Inflammatory cells (mainly mononuclear phagocytes), platelets, endothelial cells, and type II pneumocytes play a direct and indirect role in tissue injury and repair
[10].
Because TGF-ß1 may be the ‘master switch’ for a fibrotic program in lung, we studied the correlation of tissue expression of TGF-ß1, its genotype and fibrosis in lung allografts as well as considering other histological and clinical features.
|
Patients and methods
|
|---|
Patients
Specimens from 91 consecutive lung recipients carried out at Wythenshawe Lung Transplant Programme between January 1990 and March 1996 were analysed. We reviewed 780 transbronchial biopsies from all the recipients to classify and score rejection, infection and fibrosis of our patients. We also collected blood samples from all the recipients and 96 healthy volunteers.
Technical aspects
Lung transplantation operative procedures
[11]
[12]
[13] and posttransplant management
[14], have been reported in detail elsewhere. Follow-up was completed in March 1996, or at the time of the recipient's death. Follow-up was complete in all recipients and ranged from 4 to 64 months.
Immunosuppression
Recipients were treated with triple-agent immunosuppression, consisting of cyclosporine, azathioprine, and corticosteroid. Azathioprine 4 mg/kg IV was administered immediately preoperatively. Cyclosporine 2–3 mg/kg was begun during the first few postoperative hours after satisfactory renal function was ensured. The dose was adjusted to maintain whole blood levels in the range of 200–300 ng/ml. After several days, when an oral diet was initiated, oral cyclosporine was started with twice-daily doses, as determined by the whole blood level. Azathioprine was continued at a dose of 1–2 mg/kg IV, and when an oral diet was initiated, the same dosing level can be administered by the oral route. The dose is adjusted to maintain a white blood cell count in excess of 3500/dl. Methylprednisolone 10–15 mg/kg IV was given intraoperatively just prior to graft perfusion, followed by methylprednisolone 100 mg IV twice a day for 3 days before initiating an oral dose of prednisone of 0.5 mg/kg per day which we tapered to 0.2 mg/kg per day. Antilymphocyte antibodies were used as induction therapy for 3 days between 1990 and 1992 (n=17). We discontinued its use for single lung and bilateral lungs in our lung transplant program after 1992. Acute rejection episodes were treated by pulse steroid therapy (10 mg/kg per day for 3 days).
Fibrosis assessment
A total of 780 transbronchial biopsies from all patients were studied. A paraffin section of lung, stained by haematoxylin and eosin, was systematically scanned in a microscope using a x10 objective. Each successive field was individually assessed for severity of interstitial fibrosis and allotted a score between 0 and 8 using a predetermined scale (Table 1)
[15]. After examining the whole section the mean score of all the fields was taken as the fibrosis score for the section and was expressed correct to two decimal places (Table 1).
In every field the predominant degree of fibrosis was recorded that occupied more than half of the field area. The whole area of the microscope image was considered and the observer first decided whether the parenchyma in the field was normal or fibrotic. If normal tissue predominated, the field was allotted a score of 0. If fibrotic tissue predominated, the observer decided on the predominant level of fibrosis in the field, trying to place the field in one of the odd-numbered categories (Table 1). If there was any difficulty in deciding between two odd-numbered categories, the field would be given the intervening even-numbered score. Fields predominantly occupied by portions of large bronchi or vessels, were not counted, and fields partly occupied by such structures were assessed on the predominant fibrosis in the remaining field area. Fibrosis alone was considered when arriving at a score and other changes which might be present, were ignored. This scoring system seems to be reproducible and easy to apply in comparison to more complicated ones
[16].
Slides were examined by two independent observers (P. Hasleton, A. El-Gamel), slides with scoring differences of more than 1 were reviewed to reach an agreement regarding the final score of the slides
Immunohistochemistry
Paraffin-embedded sections were prepared from the biopsy sample fixed in 4% paraformaldehyde. Immunochemistry was carried out on BAL cytospins and tissue sections.
Each slide was placed in citroclear for 5 min and then 100, 100, 90, 70 and 50% alcohol, respectively, and finally distilled water. The slides were transferred to Tris buffered saline (TBS) bath and was washed for 3 min and slides were then placed in a fresh 3% hydrogen peroxide in TBS solution for 10 min to neutralise natural peroxidases. They were then washed in TBS for 3 min. Each section on the slide was marked with a paraffin pen to prevent the enzymes or antibodies from running off the slide and to reduce the chance of cross contamination. Individually the slides were removed from the TBS and placed into a humidified perspex incubator at room temperature. Each section on the slide was entirely covered with a solution of prokinase K (10%; Dako, UK), 0.1% bovine serum albumen (BSA; Chemicon International, Harrow, England), and TBS to break the formaldehyde bonds for 10 min. Slides were then transferred to a rack in TBS bath and washed for 3 min whilst stirring on a magnetic plate. To block non-specific binding further, the slides were returned to the incubator box and 10% normal swine serum (NSS; Chemicon International, Harrow, England) was left on each section for 20 min. The excess fluid was drained by tapping on the bench top. TGF-ß1 antibodies in TBS (1/10 concentration chicken anti-human TGF-ß1 provided by Dr Paul Bernchely, Department of Immunology, St Mary's Hospital, Manchester University) was added to three of the four sections on each slides whilst the remaining section (the negative control
Fig. 1
) received TBS only without antibodies. This was repeated for all the slides. The slides were left for 1 h and then rinsed in the first, second and third TBS baths whilst stirring for 3 min each time. The secondary antibody (anti chicken IgG, peroxidase conjugated 1/1000 dilution, Sigma Immuno Chemical, USA) was added and allowed to incubate for 2 h. The slides were rinsed in fourth, fifth and sixth TBS baths for 3 min each time whilst stirring. Immediately, the slides were transferred to a fresh diaminobenzedene developer bath (DAB), consisting of 280 ml TBS, 20 ml filtered DAB, and 60 µl of 30% hydrogen peroxide for approximately 45 s, then rinsed in running tap water for 10 min. The slides were placed into filtered Mayer's Haemalum for 20 s and then washed in the sink with a constant flow of cold water. The hydration process was repeated in the reverse order i.e. 50–100% alcohol, then into citroclear for 5 min each time. A cover slip was placed over each slide with the adhesive DPX (dibutyl polystyrene xylene).
Immunohistochemistry quantification
All slides were coded before evaluation. Assessment of staining was carried out with a semiquantitative method, previously described for evaluating staining
[17].
Two observers graded the presence or absence, the extent and intensity of positive staining for each biopsy specimen on a scale from 0 (negative) to 3, for intra- and extracellular staining Table 2. Positive control human tonsillar tissue slide scored maximum of 7 points (
Fig. 2
), and every batch of slides had a positive control slide stained at the same session. All slides were compared with the positive control.
Interclass correlation coefficients between observers were 0.91 for intracellular TGF-ß1 and 0.88 for extracellular TGF-ß1.
DNA samples
Anticoagulated venous blood samples (5 ml) were collected from all recipients and from healthy individuals (n=96) with no recent infection or recent history of treatment with immunosuppressive drugs.
DNA extraction and amplification
DNA was recovered using a phenol precipitation method following digestion with prokinase K (Boehringer Mannheim). Specific oligonucleotide primers were designed based on the published TGF-ß1 sequence (acc. number J04431). Using specifically designed pairs of primers (Table 3), overlapping fragments which covered a segment of sequence -1321 upstream first transcription starting point to 126 base pair of the first exon were amplified using polymerase chain reaction (PCR) amplification.
Sequence-specific oligonucleotide probing (SSOP)
Two biotinylated oligonucleotide probes (Gibco BRL) were designed to detect each polymorphism by a dot blot hybridisation technique. Specifically amplified PCR product was blotted onto Hybond N+ nylon membranes (Amersham).
The membranes were placed in, denatured and neutralised. The membranes were incubated with hybridisation buffer for 30 min at 42.5°C, then 200 ng of the specific probe (Table 4) was added and allowed to hybridise for 2 h at 42.5°C. The membranes were then washed twice at room temperature in 5xSSC with 0.1% SDS for 5 min. The membranes were then placed adjacent to X-ray films (Fuji) which were subsequently developed and the resulting blots were analysed.
The analyses
Clinical variable analysis included age, sex, diagnosis (indication for transplantation), and type of transplantation carried out (SLT versus bilateral lung transplantation (BLT) and heart lung transplantation (HLT), number of biopsies, acute rejection episodes/year, acute infection episodes/year). Distributions were determined for groups a and b to look for a possible association between any of these variables and the development of broncholilis obliterans syndrome (BOS). Mortality was compared between recipients with and without fibrosis.
Statistical analysis
Continuous variables were analysed using Mann Whitney test. Nominal variables were analysed using x2 analysis or Fisher's exact test as appropriate. Life tables survival analysis was used to compare between the two groups.
|
Results
|
|---|
Demographics
Early and late mortality summarised in Table 5 in all patients with interstitial fibrosis TGF-ß1 was immunolocalised in the airways and lung parenchyma (
Fig. 3
). Recipients who developed fibrosis had more acute rejection episodes/year, higher TGF-ß1 tissue scores, tissue eosinophilia, are more likely to have required bypass, and have homozygous TGF-ß1 genotype at codon 25 (Table 6). Recipients with fibrosis score <3 (n=33) after 1 year of transplantation had better 5-years survival than recipients with fibrosis score >3 (n=21) (
Fig. 4 ).
|
Discussion
|
|---|
We have shown an association between tissue expression of TGF-ß1 and lung allograft fibrosis. We studied TGF-ß1 genotype in the normal and transplant population to identify any differences that might explain the different biological responses to lung injury. In previous studies we have elucidated and analysed the polymorphism in the promoter regions of other cytokine genes and have shown or confirmed the correlation between differences in gene sequence and phenotypic expression. The polymorphism at codon 25 was significantly correlated with the development of allograft fibrosis in our study. We have also found that this is significantly correlated with high TGF-ß production in vitro (unpublished results). The homozygous status at codon 25 was also associated with pretansplant conditions that fibrosis is a component of the lung pathology (fibrosing alveolitis, cystic fibrosis, bronchiectesis). These results may suggest that a genotype of high TGF-ß production is a risk factor for the development of excessive lung damage that might lead to transplantation and also might predict posttransplantation pathology. Recurrent acute rejections and tissue eosinophilia are also risk factors for the development of allograft fibrosis. Cytokines and growth factors have many and complex effects on all the processes that constitute collagen metabolism. Interleukin-1 and tumour necrosis factor-
can either stimulate or inhibit collagen accumulation, presumably depending on the immediate environment
[18]. Interleukin-6 inhibits collagen degradation, and 
-interferon inhibits collagen production
[18]. Platelet-derived growth factor and fibroblast growth factor have powerful mitogenic effects on connective tissue cells in lung, and can also affect collagen production directly. The role of TGF-ß1 in the development of lung allograft fibrosis needs further studies. New strategies to down-regulate TGF-ß1 may reduce the injurious effects of ischaemia, reperfusion, allosensitivity and infections on lung allografts. Therapeutic approaches to controlling lung fibrosis by manipulating cytokine levels are promising.
|
Footnotes
|
|---|
Presented at the 11th Annual Meeting of the European Association for Cardio-thoracic Surgery, Copenhagen, Denmark, 28 September–October 1, 1997. 
|
Appendix A. Conference discussion
|
|---|
Dr G. Pettersson (Copenhagen, Denmark): An interesting paper concerning the pathophysiology of obliterative bronchiolitis and the role of TGFß. What you call fibrosis, is that very similar to obliterative bronchiolitis?
Dr A. El-Gamel: We tried to study the fibrosis in the airway as well as in the interstitium and we found that fibrosis is a generalised process in the lung allografts. Although it might start in the small airways, it eventually is a generalised process including the whole lung. And I’m sure everybody has had the experience when you take a lung out after OB it feels like a solid organ more than a lung. We have also presented in the ISHLT meeting our feelings about TGF-ß and obliterative bronchiolitis specifically. So that was a different issue.
Dr A. Haverich (Hannover, Germany): You gave a number of hypotheses now as to what is the underlying factor for the increased expression of TGF-ß in these lungs. Among those there was genetic predisposition, the heart–lung machine, eosinophilia, and also immunosuppressants. So what is your true hypothesis behind it? What are you looking for and what do you think the ultimate answer will be? Where is the TGF-ß coming from and what is the reason for the increasing expression?
Dr A. El-Gamel: That’s a good question. The true hypothesis is that the TGF-ß is the final pathway of the multiple lung injuries, and if you look in the clinically published series, there is no one factor that leads to OB; there is no one factor that leads to graft loss. Our feeling is that if you studied the genotype of a particular person and you know that they are high producers of TGF-ß, you might have to manipulate your immunosuppressant drugs; you might have to manipulate your approach to reduce TGF-ß. Also in the future we might be able to monitor TGF-ß during different episodes of infection or rejection and either give antibodies during the episode to stop it or try to do something to down-regulate it. So I think we believe it is the final pathway that all the other injurious mechanisms utilise and use.
Dr G. Pettersson: What about the timing of the process and its progress? We will present information that says that it seems likely the process starts very early, possibly immediately in relation to the transplantation, and that these patients that later develop obliterative bronchiolitis, never reach a good function.
Dr A. El-Gamel: Actually I do agree with you. We found that you can see the TGF-ß in the tissue about 18–20 months before the development of any problems with the lung allograft in terms of deterioration of function. But also what we found is that there are different types of collagen, and earlier on the collagen that deposits in the lung is collagen you can’t see in normal histology, and it’s immature collagen. Later on it’s solidified and mature, and that explains exactly what you said.
|
References
|
|---|
- Hirabayashi T, Demertzis S, Schafers J, Hoshino K, Nashan B Chronic rejection in lung allografts: immunohistological analysis of fibrogenesis. Transplant Int 1996;9(Suppl 1):S293-S295.
- Yousem SA, Suncan SR, Ohori NP, Sonmez-Alpan E Architectural remodeling of lung allografts in acute and chronic rejection (see comments). Arch Pathol Lab Med 1992;116(11):1175-1180.[Medline]
- Crystal R Interstitial lung disease. In: Wyngaarden JB, ed. Cecil Text Book of Medicine. Philadelphia, PA: Saunders, 1980:421-435.
- Brigham K, Meyrick B Interactions of granulocytes with the lungs. Circ Res 1985;54:623-635.[Abstract/Free Full Text]
- McCullough B, Collins J Bleomycin-induced diffuse interstitial pulmonary fibrosis in baboons. J Clin Invest 1978;61:79-88.
- Sitrin RG, Brubaker PG, Fantone JC Tissue fibrin deposition during acute lung injury in rabbits and its relationship to pro coagulant and fibrinolytic activities. Am Rev Respir Dis 1987;135:930-936.[Medline]
- Colvin RB, Robbin RO, Verderber EL, Lanigan JM Cell surface fibrinogen–fibrin receptors on cultured human fibroblasts. Lab Invest 1979;41:464-472.[Medline]
- Idell S, Giles C, Fair DS, Thrall RS Abnormalities of pathways of fibrin turnover in lung lavage of rats with oleic acid and bleomycin-induced lung injury support alveolar fibrin deposition. Am J Pathol 1989;135:387-399.[Abstract]
- Khalil N, Greenberg AH The role of TGF-beta in pulmonary fibrosis. Ciba Found Symp 1991;157:194-211.[Medline]
- Martinet Y, Menard O, Vaillant P, Vignaud JM, Martinet N Cytokines in human lung fibrosis. Arch Toxicol 1996;18(Supplement):127-139.
- Cooper JD The evolution of techniques and indications for lung transplantation. Ann Surg 1990;212(3):249-256.[Medline]
- Patterson GA Double lung transplantation. Clin Chest Med 1990;11(2):227-233.[Medline]
- Sundaresan S, Trachiotis GD, Aoe M, Patterson GA, Cooper JD Donor lung procurement: assessment and operative technique (see comments). Ann Thorac Surg 1993;56(6):1409-1413.[Abstract]
- Trulock EP Management of lung transplant rejection. Chest 1993;103(5):1566-1576.[Abstract/Free Full Text]
- Ashchroft T, Simpson J, Timbrell V Simple method of estimating severity of pulmonary fibrosis on a numerical scale. J Clin Pathol 1988;41:467-470.[Abstract/Free Full Text]
- Cherniack M, Colby TV, Flint A, BAL Cooperative Group Steering Committee. Quantitative assessment of lung pathology in idiopathic pulmonary fibrosis. Am Rev Respir Dis 1991;144:892–900.
- Lendrum A. The staining of eosinophil, polymorph1 and enterochromassin cells in histology sections. J Pathol Bacteriol 1944;C56:441.
- Bienkowski RS, Gotkin MG Control of collagen deposition in mammalian lung. Proc Soc Exp Biol Med 1995;209(2):118-140.[Abstract]
This article has been cited by other articles:
|
|
|
|
 
M. LIU, M. SUGA, A. A. MACLEAN, J. A. ST. GEORGE, D. W. SOUZA, and S. KESHAVJEE
Soluble Transforming Growth Factor-beta Type III Receptor Gene Transfection Inhibits Fibrous Airway Obliteration in a Rat Model of Bronchiolitis Obliterans
Am. J. Respir. Crit. Care Med.,
February 1, 2002;
165(3):
419 - 423.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. M. Mason, H.-P. Xu, S. K. Rao, A. Leask, M. Barcia, J. Shan, R. Stephenson, and S. Tabibzadeh
Lefty Contributes to the Remodeling of Extracellular Matrix by Inhibition of Connective Tissue Growth Factor and Collagen mRNA Expression and Increased Proteolytic Activity in a Fibrosarcoma Model
J. Biol. Chem.,
January 4, 2002;
277(1):
407 - 415.
[Abstract]
[Full Text]
|
|
|
|
|
|
|
R K Coker, G J Laurent, P K Jeffery, R M du Bois, C M Black, and R J McAnulty
Localisation of transforming growth factor {beta}1 and {beta}3 mRNA transcripts in normal and fibrotic human lung
Thorax,
July 1, 2001;
56(7):
549 - 556.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
A. A. MACLEAN, M. LIU, S. FISCHER, M. SUGA, and S. KESHAVJEE
Targeting the Angiotensin System in Posttransplant Airway Obliteration . The Antifibrotic Effect of Angiotensin Converting Enzyme Inhibition
Am. J. Respir. Crit. Care Med.,
July 1, 2000;
162(1):
310 - 315.
[Abstract]
[Full Text]
|
|
|
|
|
|
|
P. D Arkwright, S. Laurie, M. Super, V. Pravica, M. J Schwarz, A K. Webb, and I. V Hutchinson
TGF-beta 1 genotype and accelerated decline in lung function of patients with cystic fibrosis
Thorax,
June 1, 2000;
55(6):
459 - 462.
[Abstract]
[Full Text]
|
|
|
Top
Abstract
Introduction
