The guanine nucleotide exchange factor Tiam1 increases ... - Nature

3 downloads 4 Views 296KB Size Report
Feb 14, 2005 - 1Department of Cancer Biology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Box ..... bVersus SW480 with two-sided Fisher's exact test ... with activated Src (v-Src), and cooperates with v-Src to.
Oncogene (2005) 24, 2568–2573

& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00


The guanine nucleotide exchange factor Tiam1 increases colon carcinoma growth at metastatic sites in an orthotopic nude mouse model Meghan E Minard1,2, Matthew H Herynk1,2, John G Collard3 and Gary E Gallick*,1,2 1

Department of Cancer Biology, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Box 173, Houston, TX 77030, USA; 2The Program in Cancer Biology, University of Texas Graduate School of Biomedical Sciences at Houston, Houston, TX 77030, USA; 3The Netherlands Cancer Institute, Division of Cell Biology, 1066 CX Amsterdam, The Netherlands

Alterations in migration and adhesion are critical to invasion and metastasis. To examine signaling pathways important for colon tumor metastasis, cells of increased migratory potential from the low migratory SW480 human colorectal carcinoma parental cell line were biologically selected by serial migration through modified Boyden chambers. Several sublines were obtained with statistically significantly increased migration relative to the parental cell line. One highly migratory population was single-cell cloned and characterized. The migratory clones exhibit a four- to five-fold increase in protein and mRNA expression of T-lymphoma invasion and metastasis gene 1 (Tiam1), a guanine nucleotide exchange factor. To determine directly the role of Tiam1 in the migration of these migratory sublines, the parental SW480 cell line was transfected with a plasmid encoding the Tiam1 protein, and single cell clones were established. Ectopic expression of Tiam1 in these clones led to morphologic changes identical to biologically selected clones and increased migration. Finally, the implantation of clones that overexpress Tiam1 into the cecum of athymic mice resulted in tumor growth in the spleen, liver, and lung, whereas parental cells do not form tumors by this route of injection. These results demonstrate that overexpression of Tiam1 contributes to the metastatic phenotype of colon cancer cells. Oncogene (2005) 24, 2568–2573. doi:10.1038/sj.onc.1208503 Published online 14 February 2005 Keywords: Tiam1; colon carcinoma; migration; metastasis

Colorectal carcinoma ranks third in incidence and cancer deaths in the United States in 2003 (Jemal et al., 2004). Deaths from colorectal carcinoma almost invariably result from metastatic spread of the disease. While a cascade of molecular alterations is required for the development of metastases, one of the critical steps in the process is increased cellular migration. To study signal transduction pathways responsible for increased *Correspondence: GE Gallick; E-mail: [email protected] Received 15 December 2003; revised 25 May 2004; accepted 27 May 2004; published online 14 February 2005

migration and metastasis in colon tumor cells, we established migratory variants of the low migratory SW480 human colon adenocarcinoma cells. We then determined alterations in gene expression responsible for increased migration of these biologically selected cells. A modified Boyden chamber system (Albini et al., 1987) was used to select SW480 cells of increased migratory potential. One hundred thousand SW480 parental cells suspended in complete media were seeded in the top chamber. After 72 h, the cells that had migrated through the pores, detached from the filter, and were growing on the bottom of the dish were trypsinized and expanded. This selection process was repeated twice, each time starting with the line generated from the previous selection, resulting in the establishment of cell populations selected three times for increased migration. As shown in Figure 1a, each selection resulted in a subline that was more migratory than the line from which it was generated. The lines from the first selection were termed SW480mmig1, and demonstrate a 1.7-fold increase in migration with respect to SW480 (P ¼ 0.0074); the lines from the second selection were termed SW480mmig2, with a 3.2-fold increase with respect to SW480 (P ¼ 0.0078); and the third were termed SW480mmig3, and demonstrate a 3.6fold increase in migration with respect to SW480 (Po0.0001). The migration selection process was ended after the third selection because a fourth selection did not result in a cell population further increased in migratory potential (data not shown). To characterize the migratory variants, expression, phosphorylation, and/or activity of FAK, Src, PI3 Kinase, and Cas were examined. No obvious differences in these signaling molecules were observed in the migratory variants versus parental cells (data not shown). We next examined the expression of Tiam1, a guanine nucleotide exchange factor implicated in promoting migration in several cell types. As shown in Figure 1b, while SW480mmig1 expresses approximately equivalent Tiam1 protein to SW480 parental cells, SW480mmig2 has elevated Tiam1 expression with respect to SW480mmig1, and SW480mmig3 is elevated in Tiam1 protein with respect to SW480mmig2. Therefore, cells biologically selected for increased migration overexpress Tiam1 protein, and expression correlates with migratory capability.

Tiam1 increases in vivo growth of colon carcinoma cells ME Minard et al


From the final selection, three sublines were developed independently. Each was increased in migration (Po0.05) relative to the parental line. One of these lines (SW480mmig3) was expanded further and single cell cloned. Three clones, termed SW480mmig3 Clone1, SW480mmig3 Clone6, and SW480mmig3 Clone8, were obtained and are significantly increased in migration (six- to nine-fold, Po0.05), with respect to the parental SW480 line (Figure 1c). Immunoblotting was performed on SW480 and SW480mmig3 Clones 1, 6, and 8 for Tiam1. As shown in Figure 1d, Tiam1 protein expression is approximately four- to five-fold increased in all the migratory clones when compared to SW480

parental. To determine if increased protein expression were due to increased steady-state levels of Tiam1 mRNA, Northern blotting was performed. Each of the clones is increased approximately four- to five-fold in Tiam1 mRNA levels with respect to parental cells (Figure 1e), consistent with the increased protein expression. Selected clones have been stable in Tiam1 overexpression for more than 30 passages. To determine if the Tiam1 overexpression were sufficient to induce migration in SW480 colorectal carcinoma cells, SW480 parental cells were transfected with a plasmid encoding the Tiam1 C1199HA protein (Figure 2b). The transfected plasmid encodes for the Cterminal 1199 amino acids of the Tiam1 protein, while the endogenous protein is 1591 amino acids (Figure 2a). This truncated construct was chosen due to previous results demonstrating that full-length expression of Tiam1 inhibits cell growth, yet the truncated construct does not inhibit growth, and contains the majority of the domains of the full-length construct (van Leeuwen et al., 1995). In addition, the C1199 construct contains all of the necessary domains for protein localization to the membrane (Michiels et al., 1997; Stam et al., 1997). Several clones expressing the HA-tagged Tiam1 construct, as confirmed by immunoblotting, were obtained. Two of these were chosen randomly for further study. Expression of the HA-tagged Tiam1 constructs in these clones is shown in Figure 2c. Clones ectopically expressing this Tiam1 construct were examined for migratory potential in the Boyden Chamber migration assay. As show in Figure 2d, Figure 1 Generation of migratory lines and expression of Tiam1. (a) The average number of cells migrating in 72 h in five 200  fields for SW480, SW480mmig1 (derived from the migration of SW480), SW480mmig2 (derived from the migration of SW480mmig1), and SW480mmig3 (derived from the migration of SW480mmig2) is illustrated. The stained cells on each insert were counted in five microscopic 200  fields and averaged. The means of two cell lines were compared using a level of significance of a ¼ 0.05 with an unpaired one-sided Student’s t-test for two samples. The results represent the average and standard deviation of three independent assays. (b) Immunoblotting for Tiam1 was performed to compare Tiam1 expression for the parental SW480 and for the sublines selected for increased migratory potential, SW480mmig1, SW480mmig2, and SW480mmig3. Confluent dishes of cells were washed twice with phosphate-buffered saline and lysed on ice in radioimmunoprecipitation assay lysis buffer with a Protease Inhibitor Cocktail tablet and 30 mg was resolved by SDS– PAGE and immunoblotted as described previously (Windham et al., 2002). (c) Cells that had migrated after 48 h through the filter and attached to the bottom of the filter were stained and counted. This figure illustrates the average number of cells migrating in five 200  fields for SW480 and SW480mmig3 Clone1, SW480mmig3 Clone 6, and SW480mmig3 Clone 8. The results represent the average and standard deviation of three independent assays. (d) Immunoblotting for Tiam1 was performed on SW480, SW480mmig3 Clone1, SW480mmig3 Clone6, and SW480mmig3 Clone8. Whole-cell lysates were resolved and transferred as described in (b). (e) Northern Blotting was performed on SW480, SW480mmig3 Clone1, SW480mmig3 Clone6, and SW480mmig3 Clone8 with 5 mg of mRNA. The Tiam1 probe was designed by digesting the pcDNA3 C1199 plasmid with the restriction enzymes EcoRI and NcoI to produce a 545 bp fragment. Probes were labeled and Northern Blots were performed as described previously (Ellis et al., 1998) Oncogene

Tiam1 increases in vivo growth of colon carcinoma cells ME Minard et al

2570 Table 1 Incidence of metastasis following orthotopic injection Cell line

Mice with mets/total mice (percent)a

SW480 mmig3 Clone1 2TClone68

0/7 (0%) 4/9 (44%) 5/8 (63%)

P-valueb 0.088 0.026

a Gross, macroscopic tumors in the cecum were not present. Most metastases were to the spleen, though liver and lung metastases were present. bVersus SW480 with two-sided Fisher’s exact test

Figure 2 Transfection of SW480 parental cells with a Tiam1 construct increases migration. (a) Structure of the Tiam1 protein. Various domains are indicated in the figure: P ¼ PEST domain, PHn ¼ N-terminal pleckstrin homology domain, CC ¼ coiled-coil region, Ex ¼ additional flanking region, DHR ¼ discs-large homology domain, DH ¼ Dbl homology domain (the catalytic domain), and PHc ¼ C-terminal pleckstrin homology domain. (b) Structure of the truncated Tiam1 protein, fused to a viral hemagglutinin (HA) protein tag (Habets et al., 1994). (c) Expression of HA-tagged Tiam1 in SW480 cells is shown. The first two lanes contain wholecell lysate. (d) A 72-h migration assay was performed for SW480, 2TClone65, and 2TClone68. The relative increase in migratory cells was determined after 72 h

2TClone 65 and 2TClone68 are 10- to 14-fold increased in migration with respect to vector-transfected cells. These results demonstrate that overexpression of Tiam1 alone is sufficient to increase migration in SW480 colon carcinoma cells, and that the critical domains involved in Tiam1-mediated migration are present in the C1199 construct. To examine the effect of increased Tiam1 on tumor growth and metastasis in vivo, 2.0  106 cells from parental SW480, biologically selected clone mmig3 Clone1, and the Tiam1-transfected clone 2TClone68 were injected orthotopically into the cecal wall of athymic nude mice. The results of these experiments are shown in Table 1. SW480 cells did not produce detectable tumors in any mice, either at the orthotopic injection site or in other organs, consistent with results of others (Hewitt et al., 2000). In contrast, while no mice from the clones with increased Tiam1 expression developed gross, macroscopic tumors in the cecum, Oncogene

four of the nine mice injected with mmig3 Clone1, and five of the eight mice injected with 2TClone68 had tumor growth in their spleens, livers, and/or lungs. An example of a spleen from an SW480mmig3 Clone 1-injected mouse is shown in Figure 3a. Multiple foci of small tumors are observed throughout the spleens of mice that developed tumors. Similarly, by hematoxylin and eosin staining, additional microscopic clusters of tumor cells were observed in the spleen (Figure 3b). To confirm that the tumors arose from the injected cells, tumor sections from spleens of mice injected with 2TClone68, and SW480 as a control, were stained for the HA tag. As shown in Figure 3c, the tumor cells in the spleen from a 2T68-injected mouse express the HA tag in the cytoplasm, while no cells in the SW480-injected mouse spleen stain positive for the HA tag. In addition, a spleen from a 2TClone68-injected mouse incubated without primary antibody showed no staining. These results demonstrate that the splenic tumors resulted from the Tiam-1 transfected clones. Taken together, these results indicate that increased Tiam1 expression is sufficient to increase both migration and metastasis in colon carcinoma cells. As one of Tiam1’s known functions is as a guanine nucleotide exchange factor for Rac, we examined GTPbound (active) Rac in parental SW480, biologically selected SW480mmig3 Clones 1 and 6, and Tiam1transfected clones 65 and 68. GTP-bound Rac was assessed by pull-down assays with the p21-binding domain of PAK-1. As shown in Figure 4a, when cells were serum-starved overnight and then stimulated with complete media for 10 min, Rac activity was minimally increased (to 1.5-fold) in the biologically selected clones compared to SW480, as well as in the Tiam1-transfected clones, relative to vector-transfected cells. Similar results were seen for GTP-bound Rho (Figure 4b), by pull down assays with Rhotekin, and for Rac when cells were treated with 100 nM PMA (data not shown). The modest increases in GTP-bound Rac in Tiam1 overexpressing SW480 cells are likely due to high intrinsic GTP-bound Rac in these cells, the result of the expression of an activated K-Ras (Capon et al., 1983). These results suggest the possibility that Tiam1 may have Rac-dependent and Rac-independent functions, and are consistent with other studies on Tiam1 function. In vitro, Tiam1 acts as an exchange factor for Rac1 and Cdc42, and to a lesser extent, for RhoA. In vivo, Tiam1 acts as an exchange factor primarily for Rac1 (Michiels et al., 1995). A strong direct correlation was found

Tiam1 increases in vivo growth of colon carcinoma cells ME Minard et al


Figure 3 Growth of biologically selected and Tiam1-transfected migratory cells in nude mice. SW480 parental, mmig3 (biologically selected for increased migration) and 2TClone68 (Tiam1-transfected) cells (2.0  106) were inoculated intracecally. Prior to implantation, cells were washed, and enzymatically and physically removed from the tissue culture dishes. Viability of cells was determined by trypan blue exclusion. Mice were anesthetized by Methoxyfluorane, and cells were injected in a volume of 0.05 ml. Incisions were closed with surgical staples, which were removed after 1 week. Mice were killed when moribund or after 22 weeks. Spleens and livers were removed and fixed in formalin, and spleens containing tumors were also quick-frozen in liquid nitrogen. (a) Photograph of a representative spleen from a mmig3 Clone1-injected mouse. Yellow arrows point to tumor nodules. (b) H and E staining of formalin-fixed, paraffin-embedded spleen sections from an uninjected mouse, and mice injected with SW480, mmig3 Clone1, and 2TClone68. Yellow arrows indicate some of the clusters of tumor cells in the mmig3 Clone1 and 2TClone68 sections. (c) Formalin-fixed, paraffin-embedded sections were stained with an antibody to the viral hemagglutinin protein, which was used as a tag on the c-terminus of the Tiam1 construct used for tranfection. Primary antibody was visualized by incubation with a cy5-conjugated secondary antisera (goat anti-rabbit)

between the levels of Tiam1 and Rac mRNA expression and migration potential in four ovarian tumor cell lines (Wu et al., 2003). Activated Ras and Tiam1 can cooperate to induce Rac activation (Lambert et al., 2002). Tiam1-deficient mice are resistant to classic skin tumorigenesis assays in which Ras is activated (Malliri et al., 2002), demonstrating that Tiam1 can be an important mediator of Ras-induced tumorigenesis and/ or tumor progression. Recently, Tiam1 has shown to be tyrosine phosphorylated in NIH 3T3 cells transformed with activated Src (v-Src), and cooperates with v-Src to induce Rac activation, required for transformation in this system (Servitja et al., 2003). Both Ras and Src activation are common in colon cancer, with approximately 50% of colon tumors harboring an activated Ras gene (Bos et al., 1987), and nearly 80% increased in Src activity (Talamonti et al., 1993; Frame, 2002; Summy and Gallick, 2003). Thus, Tiam1 may cooperate with activation of both the Src and Ras signaling pathways in driving tumor progression.

While Rac activation has been implicated in many of the above processes, Tiam1 is likely to have Racindependent functions. Tiam1 is highly conserved, and is expressed in virtually all murine tissues that have been examined (Habets et al., 1995). As shown in Figure 2a, in addition to the Dbl-homologous (DH) domain, which is involved in GDP–GTP exchange, Tiam1 contains a myristoylation site at the N-terminus, two PEST domains, and two pleckstrin homology (PH) domains (Habets et al., 1994). Tiam1 also contains a discs-large homology region (DHR) (van Leeuwen et al., 1995). Two additional regions in the Tiam1 protein, a coiledcoiled (CC) domain and another region termed Ex are immediately adjacent to the PHn domain (Stam et al., 1997). While most GNEFs have only one PH domain (PHc) (Whitehead et al., 1997), Tiam1 has two PH domains. Thus, PHn domains may be involved in Racindependent functions of Tiam1. The complex interactions of Tiam1 likely account for its different effects on migration in different cell types. Oncogene

Tiam1 increases in vivo growth of colon carcinoma cells ME Minard et al


Figure 4 GTP-bound Rac and Rho in SW480 and Tiam1 overexpressing clones. Cells were grown to approximately 80– 85% confluency in media with 10% serum. Cells were then serum starved for 16 h, then stimulated with complete media for 10 min. Cells were lysed and assayed according to manufacturer’s (UBI) instructions (GTP-bound Rac was assessed by pull down with a GST fusion-protein corresponding to the p21-binding domain (PBD) of human PAK-1; GTP-bound Rho by pull down with a GST fusion protein corresponding to residues 7–89 of Rhotekin Rho binding domain). The beads were pelleted, washed, boiled in reducing and denaturing buffer, and proteins were separated by SDS–PAGE (12% gel), and immunoblotted for Rac or Rho. For negative and positive controls, lysates were first incubated with GDP or GTPgS, respectively, and then incubated with the GST fusion protein-agarose beads. (a) GTP-bound Rac, total Rac, and b-actin as a loading control are illustrated for SW480, mmig3 Clone1, mmig3 Clone6, Vector, 2TClone 65, and 2TClone 68. (b) GTP-bound Rho, total Rho, and b-actin as a loading control are illustrated for SW480, mmig3 Clone1, mmig3 Clone6, Vector, 2TClone 65, and 2TClone 68

For example, several studies have demonstrated that Tiam1 promotes cellular adhesion and/or decreased

migration (Hordik et al., 1997; Sander et al., 1999; Engers et al., 2000, 2001). In contrast, other studies have demonstrated that overexpression of Tiam1 increases migration in several cell types (Ehler et al., 1997; van Leeuwen et al., 1997; Bourguignon et al., 2000a, b; Fleming et al., 2000). Recently, upregulation of Tiam1 was associated with increased migration in vitro, and increased invasion in vivo, of B88t oral squamous cell carcinoma cells following downregulation of p27Kip1 (Supriatno et al., 2003). Further, Tiam1 was first isolated by a retroviral insertion strategy to select for increased invasiveness of T lymphoma cells (Habets et al., 1994). Nearly all tumor cell lines examined express Tiam1 (Habets et al., 1995). N-terminal truncation of the Tiam1 protein in NIH 3T3 cells activates its oncogenic potential, similar to other DH-containing proteins. In addition, a plasmid encoding the C-terminal 1199 amino acids of Tiam1 transfected into NIH 3T3 cells produces tumors in nude mice (van Leeuwen et al., 1995). Increased Tiam1 expression correlates with increased grade of human breast tumors (Adam et al., 2001) and breast tumor cell lines selected for increased metastatic potential are increased in Tiam1 expression (Minard et al., 2004). The results presented here not only demonstrate increased migratory potential of Tiam1 overexpressing cells, they are the first to indicate that Tiam1 overexpression contributes to the development of metastases in colon tumor cells. No previous studies examined the potential role(s) for Tiam1 in regulation of colon tumor cell migration and potential roles of Tiam1 in colon tumor progression. The SW480 and migratory, metastatic variants described in this study therefore provide a novel model to further study roles of Tiam1 in promoting the metastatic phenotype. Acknowledgements We thank Nila U Parikh and Charles G Minard for excellent assistance with experimental techniques and statistical analyses, respectively. This research was supported by 2 RO-1 CA25567 and U54 CA 090810 to GEG.

References Adam L, Vadlamudi RK, McCrea P and Kumar R. (2001). J. Biol. Chem., 276, 28443–28450. Albini A, Iwamoto Y, Kleinman HK, Martin GR, Aaronson SA, Kozlowski JM and McEwan RN. (1987). Cancer Res., 47, 3239–3245. Bos JL, Fearon ER, Hamilton SR, Verlaan-de Vries M, van Boom JH, van der Eb AJ and Vogelstein B. (1987). Nature, 327, 293–297. Bourguignon LY, Zhu H, Shao L and Chen YW. (2000a). J. Cell Biol., 150, 177–191. Bourguignon LY, Zhu H, Shao L and Chen YW. (2000b). J. Biol. Chem., 275, 1829–1838. Capon DJ, Seeburg PH, McGrath JP, Hayflick JS, Edman U, Levinson AD and Goeddel DV. (1983). Nature, 304, 507–513. Ehler E, van Leeuwen F, Collard JG and Salinas PC. (1997). Mol. Cell Neurosci., 9, 1–12. Oncogene

Ellis LM, Staley CA, Liu W, Fleming RY, Parikh NU, Bucana CD and Gallick GE. (1998). J. Biol. Chem., 273, 1052–1057. Engers R, Springer E, Michiels F, Collard JG and Gabbert HE. (2001). J. Biol. Chem., 276, 41889–41897. Engers R, Zwaka TP, Gohr L, Weber A, Gerharz CD and Gabbert HE. (2000). Int. J. Cancer, 88, 369–376. Fleming IN, Gray A and Downes CP. (2000). Biochem. J., 351, 173–182. Frame MC. (2002). Biochim. Biophys. Acta, 1602, 114–130. Habets GG, Scholtes EH, Zuydgeest D, van der Kammen RA, Stam JC, Berns A and Collard JG. (1994). Cell, 77, 537–549. Habets GG, van der Kammen RA, Stam JC, Michiels F and Collard JG. (1995). Oncogene, 10, 1371–1376. Hewitt RE, McMarlin A, Kleiner D, Wersto R, Martin P, Tsokos M, Stamp GHW and Stetler-Stevenson WG. (2000). J. Pathol., 192, 446–454.

Tiam1 increases in vivo growth of colon carcinoma cells ME Minard et al

2573 Hordijk PL, ten Klooster JP, van der Kammen RA, Michiels F, Oomen LC and Collard JG. (1997). Science, 278, 1464–1466. Jemal A, Tiwari RC, Murray T, Ghafoor A, Samuels A, Ward E, Feuer EJ and Thun MJ. (2004). CA Cancer J. Clin., 54, 8–29. Lambert JM, Lambert QT, Reuther GW, Malliri A, Siderovski DP, Sondek J, Collard JG and Der CJ. (2002). Natl. Cell Biol., 4, 621–625. Malliri A, van der Kammen RA, Clark K, Van DV, Michiels F and Collard JG. (2002). Nature, 417, 867–871. Michiels F, Habets GG, Stam JC, van der Kammen RA and Collard JG. (1995). Nature, 375, 338–340. Michiels F, Stam JC, Hordijk PL, van der Kammen RA, Ruuls-Van Stalle L, Feltkamp CA and Collard JG. (1997). J. Cell Biol., 137, 387–398. Minard ME, Kim LS, Price JE and Gallick GE. (2004). Breast Cancer Res. Treat., 84, 21–32. Sander EE, ten Klooster JP, van Delft S, van der Kammen RA and Collard JG. (1999). J. Cell Biol., 147, 1009–1022. Servitja JM, Marinissen MJ, Sodhi A, Bustelo XR and Gutkind JS. (2003). J. Biol. Chem., 278, 34339–34346.

Stam JC, Sander EE, Michiels F, van Leeuwen FN, Kain HE, van der Kammen RA and Collard JG. (1997). J. Biol. Chem., 272, 28447–28454. Summy JM and Gallick GE. (2003). Cancer Metastasis Rev., 22, 337–358. Supriatno, Harada K, Kawaguchi S, Yoshida H and Sato M. (2003). Oncol. Rep., 10, 527–532. Talamonti MS, Roh MS, Curley SA and Gallick GE. (1993). J. Clin. Invest., 91, 53–60. van Leeuwen FN, Kain HE, Kammen RA, Michiels F, Kranenburg OW and Collard JG. (1997). J. Cell Biol., 139, 797–807. van Leeuwen FN, van der Kammen RA, Habets GG and Collard JG. (1995). Oncogene, 11, 2215–2221. Whitehead IP, Campbell S, Rossman KL and Der CJ. (1997). Biochim. Biophys. Acta, 1332, F1–F23. Windham TC, Parikh NU, Siwak DR, Summy JM, McConkey DJ, Kraker AJ and Gallick GE. (2002). Oncogene, 21, 7797–7807. Wu MF, Xi L, Chen G, Li J, Xu Q, Li FJ, Lu YP, Wang SX, Liao GN and Ma D. (2003). Zhongguo Yi. Xue. Ke. Xue. Yuan Xue. Bao., 25, 434–437.


Suggest Documents