Vascular Endothelial Growth Factor and Activin-A Serum Levels ...

Biology of Blood and Marrow Transplantation 13:942-947 (2007) 䊚 2007 American Society for Blood and Marrow Transplantation 1083-8791/07/1308-0001$32.00/0 doi:10.1016/j.bbmt.2007.04.007

Vascular Endothelial Growth Factor and Activin-A Serum Levels Following Allogeneic Hematopoietic Stem Cell Transplantation David Nachbaur, Petra Schumacher, Jutta Auberger, Johannes Clausen, Brigitte Kircher Division of Hematology and Oncology, Innsbruck Medical University, Anichstrasse 35, A-6020 Innsbruck, Austria Correspondence and reprint requests: David Nachbaur, MD, Clinical Division of Hematology and Oncology, Department of Internal Medicine, Anichstrasse 35, A-6020 Innsbruck, Austria (e-mail: [email protected]). Received February 5, 2007; accepted April 12, 2007

ABSTRACT Allogeneic hematopoietic stem cell transplantation is frequently complicated by syndromes characterized by a disruption of the endothelial integrity such as graft-versus-host disease or liver toxicity. Vascular endothelial growth factor and activin-A, a member of the transforming growth factor beta (TGF-␤) superfamily, are important for endothelial integrity and tissue repair. We retrospectively measured endogenous vascular endothelial growth factor and activin-A serum levels in 70 patients following allogeneic stem cell transplantation. Vascular endothelial growth factor serum levels were significantly decreased within the first 2 weeks after the transplant and returned to pre transplant levels by day ⴙ15. Activin-A serum levels were significantly elevated from day ⴙ7 with peak levels reached on day ⴙ10. By using the median value as cutoff high vascular endothelial growth factor levels on day ⴙ15 were associated with significantly better overall survival, less liver toxicity, faster neutrophil recovery, and a trend towards less severe acute graft-versus-host disease. No correlation was found between activin-A serum levels and survival, liver toxicity, neutrophil recovery, or graft-versus-host disease. Monitoring of vascular endothelial growth factor levels following allogeneic hematopoietic stem cell transplantation might help to identify patients with a very high risk for early transplantrelated complications. © 2007 American Society for Blood and Marrow Transplantation

KEY WORDS Vascular endothelial growth factor tion

●

Activin-A

INTRODUCTION Allogeneic hematopoietic stem cell transplantation (HSCT) is the treatment of choice for a variety of otherwise life-threatening diseases of the lympho-hematopoietic system. However, allogeneic SCT may be complicated by syndromes such as graft-versus-host disease (GVHD) or chemo/radiotherapy-induced liver toxicity resulting in substantial transplantationrelated morbidity and mortality (TRM) [1]. Various up- or dysregulated proinflammatory cytokines, endothelial markers, and coagulation parameters have been discussed to be involved in the endothelial injury observed in both syndromes [2-4]. Both vascular endothelial growth factor (VEGF) and activin-A, a member of the transforming growth factor beta (TGF-␤) superfamily, are known key players in vasopermeability 942

●

Allogeneic hematopoietic stem cell transplanta-

and integrity, inflammation, repair, neovascularization, and fibrosis [5-7]. To better define the role of VEGF and activin-A in the context of allogeneic HSCT and its complications we conducted a retrospective analysis of endogenous VEGF and activin-A serum concentrations in the early posttransplant period following allogeneic SCT.

PATIENTS AND METHODS Patients

Seventy patients receiving an allogeneic hematopoietic SCT at our institution between May 1998 and September 2004 after giving written informed consent were included in the study, which has been approved

943

VEGF and Activin-A in Hematopoietic Stem Cell Transplantation

Table 1. Patient Characteristics N Total patient number Median patient age, years (range) Median donor age, years (range) Diagnosis Acute leukemia Chronic myelogenous leukemia Lymphoma Intermediate-, high-risk myelodysplastic syndrome Others Donor HLA-identical sibling donor Haploidentical family donor Volunteer unrelated donor Conditioning regimen Total-body irradiation containing Busulfan-based Sex match Male recipient/female donor Others Stem cell source Bone marrow Peripheral blood Stage at transplantation Standard risk High risk

70 40 (17-76) 38 (15-79) 42 8 7 6 7 38 3 29 26 44 15 55 22 48 25 45

jiroveci pneumonia prophylaxis with trimethoprimsulfamethoxazole, 1 double-strength tablet 3 times weekly until day 180 after transplantation. For Herpes simplex and Varicella-zoster virus prophylaxis patients received either valacyclovir 500 mg twice daily orally or low-dose acyclovir 250 mg/m2 intravenously 3 times daily from the beginning of conditioning. Antifungal prophylaxis consisted of fluconazole 400 mg/day from the beginning of conditioning until d ⫹73 [15]. Irradiated (25 Gy) leukocyte-depleted random platelet units from single donors and red cells were transfused when hemoglobin levels dropped to ⬍8.0 g/dL and platelet counts dropped to ⬍20 g/L. Quantitative Determination of VEGF and Activin-A

Serum samples were collected every 4-7 days from day ⫺4 until day ⫹70 and stored at ⫺70°C until use. Human VEGF was detected by quantitative sandwich enzyme immunoassay technique (R&D Systems, Inc., Minneapolis, MN) with a detection limit of 25 pg/mL. Activin-A was measured by solid phase sandwich ELISA (Oxford Bio-Innovation Ltd, Oxfordshire, UK) with a detection limit of ⬍0.076 ng/mL. Statistical Methods

by the local ethical committee. Detailed patient characteristics are listed in Table 1. Conditioning regimen were myeloablative (n ⫽ 48) or (RIC) (n ⫽ 22) in case of contraindication for conventional myeloablative conditioning as defined previously [8]. Donors were either human leukocyte antigen-identical siblings (n ⫽ 38) or haploidentical family members (n ⫽ 3) [9]. Twenty-nine patients received a stem cell graft from a volunteer unrelated donor. Only patients surviving more than 50 days after transplantation were included in the analysis. Prophylaxis, Diagnosis, and Treatment of GVHD

Cyclosporine A (CSa) and methotrexate or mycophenolate mofetil (MMF) were administered according to the Seattle protocols for myeloablative or reducedintensity transplants [10,11]. Patients receiving unrelated stem cell grafts additionally received low-dose rabbit antithymocyte globulin (ATG) (Thymoglobulin®, IMTIX SangStat, Germany ) as described [12]. Acute and chronic GVHD (aGVHD, cGVHD) were diagnosed from clinical symptoms and/or biopsies from skin, oral mucosa, liver, or gut, and classified according to the previously published standard Seattle criteria [13,14]. aGVHD ⬎ grade II was treated with steroids (2 mg/kg/day). Supportive Care

No prophylactic antibiotics were administered during neutropenia. All patients received Pneumocystis

Statistical analyses were completed using the Mann-Whitney U-test for unpaired samples, the chisquare test for independence, and the Spearman rank correlation coefficient using GraphPad Prism 4.0 (GraphPad Software, Inc., San Diego, CA) with a 5% level of significance. Survival data were analyzed as of May 2006. Overall survival was calculated from the date of SCT to the date of death from any cause or date of last follow-up. Probabilities of overall survival (OS) were estimated using the Kaplan-Meier method and compared using the log-rank test [16]. Cumulative incidence estimates were calculated for transplant-related mortality (TRM) and aGVHD using the NCSS statistical software package (Kaysville, UT) [17]. TRM was defined as the probability of dying without previous occurrence of relapse/progression, which is a competing event. Median values of VEGF serum levels, activin-A serum levels, and maximal total serum bilirubin were used as cutoff for categorization into high and low value groups. A Cox regression was performed including recipient age, donor age, time interval between diagnosis and SCT, CD34 cell dose, maximal total serum bilirubin, and number of days with a total serum bilirubin ⬎2 mg/dL as continuous variables, and donor type (HLA-identical sibling versus others), risk category by the underlying disease (standard versus high risk), sex match (male recipient/female donor versus others), VEGF serum level on day ⫹15 (high versus low), and

944

D. Nachbaur et al.

CD34⫹ cell dose and day ⫹10 activin-A serum levels (data not shown). VEGF and Activin-A Serum Levels and Liver Toxicity

Figure 1. Mean VEGF and activin-A serum levels ⫾ SE following allogeneic hematopoietic SCT.

stem cell source (peripheral blood versus bone marrow stem cells) as categoric variables. Standard-risk disease was defined as chronic myelogenous leukemia (CML) in first chronic phase and acute leukemia in first complete remission (CR1). All other indications were classified as high-risk disease. RESULTS

Liver toxicity was assessed by maximal total serum bilirubin levels, number of days with total serum bilirubin ⬎2 mg/dL, and % maximal weight gain within day ⫹20. The median maximum total serum bilirubin before day ⫹20 was 2.89 (range, 0.72-15.45) mg/dL, the median day number with a total serum bilirubin ⱖ 2 mg/dL was 4.5 (range, 0-20) days, and the median weight gain (% increase) was 4.2 (range, 0-17.6) percent. Within the first 2 weeks following transplant no significant correlation was found between VEGF serum levels and maximal total serum bilirubin or days with hyperbilirubinemia within day ⫹20. From day ⫹15 through day ⫹70, however, VEGF serum levels were significantly inverse correlated to maximal total serum bilirubin levels and to the number of days with total serum bilirubin serum levels ⱖ 2 mg/dL within day ⫹20 (Figure 2a and b). In contrast, no correlation was found between activin-A levels and maximal total serum bilirubin or

VEGF and Activin-A Serum Levels

Mean VEGF serum levels were significantly decreased during the first 2 weeks with trough levels on day ⫹7 (61.5 ⫹ 4.5 pg/mL versus 315.5 ⫹ 48.4 pg/mL on day ⫺4, P ⬍ .0001) returning to pretransplant levels by day ⫹15. VEGF peak levels were measured around day ⫹40 (437 ⫹ 41 pg/mL, P ⫽ .02 versus pretransplant VEGF levels). In contrast, mean activin-A serum levels were significantly increased from day ⫹7 through ⫹70 with peak levels on day ⫹10 (0.717 ⫹ 0.103 ng/mL versus 0.422 ⫹ 0.034 ng/mL day ⫺4, P ⫽ .0017) (Figure 1). VEGF and Activin-A Serum Levels, Neutrophil Recovery, and CD34 Cell Count

Neutrophil recovery was defined as the first of 3 consecutive days with a persistent absolute neutrophil count ⬎0.5 G/L. Neutrophil recovery was significantly faster in patients with high VEGF levels (⬎151.2 pg/ mL) on day ⫹15 (median 11 [range, 8-16] days versus 15 [range, 10-35] days, P ⬍ .0001, Mann-Whitney U-test), whereas no such correlation could be demonstrated for patients with high versus low day ⫹10 activin-A serum levels. There was also no correlation between pretransplant VEGF or activin-A serum levels and time to neutrophil recovery (data not shown). Because of patient number there was only a trend for a faster neutrophil recovery in patients receiving a higher number of CD34⫹ stem cells (P ⫽ .08). There was a highly significant correlation between CD34 cell dose and day ⫹15 VEGF serum levels (r ⫽ 0.5, P ⫽ .0003). No such correlation was found between

Figure 2. Correlation between VEGF serum levels on day ⫹15 and a, maximal total serum bilirubin and b, number of days with total serum bilirubin ⱖ 2 mg/dL within the first 20 days following allogeneic hematopoietic SCT.

945


1,0

% survival

0,8

VEGF ≥ 151.2 pg/mL, overall survival 65% (48%-82%, 95% CI)

0,6

VEGF < 151.2 pg/mL , overall survival 29% (12%-45%, 95% CI)

0,4

0,2

p=0.0033, log rank test 0,0 0

360

720

1080 1440 1800 2160 2520 2880 3240

days after SCT Figure 3. Overall survival of patients with high versus low VEGF serum levels on day ⫹15 following allogeneic hematopoietic SCT. The median value of VEGF (151 pg/mL) on day ⫹15 was used to discriminate between patients with high and low VEGF levels.

days with total serum bilirubin levels ⱖ 2 mg/dL. Furthermore, no correlation was found between VEGF or activin-A serum levels measured at any time point (pre- or posttransplant) and % maximal weight gain within day ⫹20, respectively (data not shown). VEGF and Activin-A Serum Levels, OS, TRM, and Relapse Incidence

Overall survival at 4 years for the entire cohort was 49% (37%-61%, 95% confidence interval, [CI]). By univariate analysis patients with a maximal total serum bilirubin ⬎2.89 mg/dL (ie, median) within day ⫹20 had a significantly poorer outcome (OS 38%, 22%55%, 95% CI versus 61%, 44%-78%, 95% CI, for patients with a maximal total serum bilirubin ⬍2.89 mg/dL; P ⫽ .02, log rank test) because of a higher TRM (35% versus 21%, P ⫽ .07, log-rank test). No differences were observed between patients with high or low maximal total serum bilirubin within day ⫹20 with regard to relapse incidence (data not shown). To evaluate whether VEGF serum levels were associated with outcome, day ⫹15 levels were first chosen for the analysis because this time point was the earliest to show a significant correlation with parameters of liver toxicity. The median value at the respective time point was chosen to discriminate between “high” and “low” value groups. Patients with a day ⫹15 VEGF serum level above the median value of 151.2 pg/mL (ie, “high” value group) had a significantly better OS of 65% (48%-82%, 95% CI) versus 29% (12%-45%, 95% CI) for patients with low VEGF serum levels (P ⫽ .0033, log rank test; Figure 3). This effect resulted from a lower TRM (22% versus 35%, P ⫽ .06, log-rank test) and a lower relapse incidence (25% versus 41%, P ⫽ .05, log rank test) in the high-value group. If day ⫹40 VEGF levels were used for the analysis a similar survival benefit was observed for patients with high VEGF serum levels ( 69%, 52%85%, 95% CI, versus 33%, 16%-50%, 95% CI, for

patients with low VEGF serum levels, P ⫽ .0067, logrank test). However, in contrast to earlier time points after the transplant, this survival benefit was only the result of a lower TRM (15% versus 35%), whereas relapse incidence was similar in both cohorts (36% versus 31%). Pretransplant VEGF and activin-A serum levels as well as activin-A serum levels on day ⫹10 had no significant impact on OS, TRM, relapse (data not shown). VEGF and Activin-A Serum Levels and aGVHD

Overall, the cumulative incidence of aGVHD grades II-IV was 64% (54%-77%, 95% CI) with no significant differences in the incidence with respect to donor type (HLA-identical sibling versus others), stem cell source (peripheral blood versus bone marrow), or conditioning regimen (myeloablative versus RIC) (data not shown). There were also no significant differences in the cumulative incidence of aGVHD grades II-IV in patients with high or low VEGF levels either pretransplant or on day ⫹15, or in patients with high or low activin-A serum levels either pretransplant or on day ⫹10 (data not shown). There was a trend towards less severe aGVHD grades III-IV in patients with high day ⫹15 VEGF levels (P ⫽ .07, chi-square test). There was no significant difference in the incidence of aGVHD grades II-IV requiring steroid treatment before day ⫹15 between patients with high or low day ⫹15 VEGF levels (16% versus 28%, P ⫽ .2 chi-square test). Multivariate Analysis

By multivariate Cox regression standard-risk disease was the most powerful predicitve parameter for a lower risk for death (relative risk 0.1126, P ⫽ .02) followed by high day ⫹15 serum VEGF levels (rela-

946

D. Nachbaur et al.

Table 2. Multivariate Cox Regression Analysis for the Risk of Death following Allogeneic Hematopoietic Stem Cell Transplantation Variable Continuous Time interval between diagnosis and HSCT CD34 cell dose Recipient age Donor age Maximal total serum bilirubin Number of days with bilirubin >2 mg/dL Categorical HLA-identical sibling donor Standard-risk disease Peripheral blood stem cells Male recipient/female donor High dⴙ15 VEGF serum level

Relative Risk

P Value

1.0002 0.9852 0.9857 0.9994 1.1677 0.9059

(0.9999-1.0006) (0.8990-1.0796) (0.9439-1.0293) (0.9488-1.0528) (0.8850-1.5408) (0.7840-1.0467)

.22 .75 .51 .98 .27 .18

0.3913 0.1126 1.1527 1.7870 0.1936

(0.1391-1.1009) (0.0174-0.7284) (0.0967-13.7363) (0.3425-7.3855) (0.0503-0.7442)

.08 .02 .91 .42 .02

VEGF indicates vascular endothelial growth factor.

tive risk 0.1936, P ⫽ .02) There was a trend for a lower risk for death in patients with an HLA-identical sibling donor (relative risk 0.3913, P ⫽ .08) (Table 2). DISCUSSION In the present analysis higher VEGF serum levels after hematopoietic regeneration following allogeneic SCT were associated with a significantly better OS. Even by multivariate Cox regression analysis higher VEGF serum level on day ⫹15 remained a significant parameter for a lower risk for death. This survival advantage in patients with high (ie, greater than the median value) VEGF levels on day ⫹15 following allogeneic SCT resulted from a lower TRM and a lower relapse incidence. A lower 1-year nonrelapse mortality (NRM) for patients with low VEGF levels between day ⫹7 and day ⫹14 has already been observed by Min and coworkers [18], although they did not comment on survival in their cohort. The interrelation between VEGF and relapse incidence is not clear, and warrants further investigations. It can at least partly be explained by the fact that significantly more high-risk patients belonged to the cohort of patients with low day ⫹15 VEGF serum levels compared with patients with high day ⫹15 VEGF levels (data not shown). However, when day ⫹40 VEGF levels were used for the analysis instead of day⫹15 values the difference in relapse incidence between high and low value groups disappeared, whereas the effects on survival and TRM remained significant, suggesting that over time VEGF levels become more important for transplant-related complications than for relapse. One of these complications is liver toxicity, which is usually thought to be caused by disturbances of various coagulation parameters and/or inflammatory cytokines, resulting in endothelial injury [3,4]. From an initial report in 6 patients with typical manifestations of veno-occlusive disease it was postulated that

VEGF might be 1 of the cytokines predictive and causative for hepatic veno-occlusive disease [19]. Our results, however, in a substantially larger group of patients do not confirm these initial findings, because high VEGF levels were significantly correlated with less liver toxicity and significantly fewer days with elevated serum bilirubin within the first 20 days. GVHD is another frequent complication following allogeneic SCT, resulting in significant transplant-related morbidity and TRM. Similar to our results, a trend for a lower incidence of severe aGVHD (ie, grades III-IV) in patients with high day ⫹15 VEGF levels has already been described by Min et al. [18]. Together, these findings do not support a role of circulating endogenous VEGF in the cytokine network involved in the pathogenesis of aGVHD [20,21]. Whether steroids affect VEGF serum levels is currently unknown. At least in our cohort there were no significant differences in the percentage of patients receiving steroid treatment because of aGVHD before day ⫹15 between high or low day ⫹15 VEGF value groups. Another important finding of our study is the correlation between VEGF levels, CD34 cell dose, and neutrophil recovery, suggesting that upcoming donor-derived peripheral blood cells are the main source of circulating endogenous VEGF following allogeneic SCT [22,23]. Interestingly, by multivariate Cox regression analysis, high day ⫹15 VEGF levels were significantly more predeictive for a better outcome after transplant than CD34 cell dose in the respective model. The second cytokine that was analyzed in our study was activin-A, a member of the TGF-␤ family. Although increased active-A serum levels after day ⫹10 suggest a possible role in general repair following allogeneic HSCT, we failed to demonstrate any correlation between activin-A serum levels and survival, liver toxicity, hematopoietic recovery, or GVHD in our patient cohort [24]. Whether activin-A might be

947


involved in later stages of liver toxicity when stellate cell proliferation and collagenization by types I, III, and IV collagen take place remains to be shown [25]. In vitro studies have demonstrated a stimulatory effect of activin-A on collagen type I messenger RNA expression, and elevations of procollagen propeptide has been observed in patients with veno-occlusive disease [26,27]. Moreover, elevated plasma TGF-␤ levels prior to autologous SCT have been reported to be strongly correlated with the development of hepatic veno-occlusive disease [28]. In conclusion, although the presented data must be considered, preliminary, patients with high post transplant VEGF levels showed improved survival, a faster neutrophil recovery, less liver toxicity, and less severe aGVHD. Patients with low posttransplant VEGF levels represent a high-risk cohort of patients with a higher relapse rate and/or a higher TRM. Monitoring VEGF serum levels following allogeneic SCT might help to identify patients at a very high risk for early death.

REFERENCES 1. Appelbaum FR. The current status of hematopoietic cell transplantation. Annu Rev Med. 2003;54:491-512. 2. Delve LD, Schulman H, McDonald GB. Toxic injury to hepatic sinusoids: sinusoidal obstruction syndrome (venom-occlusive disease). Semin Liver Dis. 2002;22:27-41. 3. Barman SI. The syndrome of hepatic venom-occlusive disease after marrow transplantation. Blood. 1995;85:3005-3020. 4. Wadleigh M, Ho V, Momtaz P, Richardson P. Hepatic venoocclusive disease: pathogenesis, diagnosis and treatment. Curr Opin Hematol. 2003;10:451-462. 5. Ferrara N, Gerber HP, LeCouter J. The biology of VEGF and its receptors. Nat Med. 2003;9:669-676. 6. Munz B, Hübner G, Tretter Y, Alzheimer C, Werner S. A novel role of activin in inflammation and repair. J Endocrinol. 1999;161:187-193. 7. Jones KL, de Kretser DM, Patella S, Phillips DJ. Activin A and follistatin in systemic inflammation. Mol Cell Endocrinol. 2004; 225:119-125. 8. Nachbaur D, Clausen J, Kircher B. Donor cytomegalovirus seropositivity and the risk of leukemic relapse after reducedintensity transplants. Eur J Haematol. 2006;76:414-419. 9. O’Donnell PV, Luznik L, Jones RJ, et al. Nonmyeloablative bone marrow transplantation from partially HLA-mismatched related donors using posttransplantation cyclophosphamide. Biol. Blood Marrow Transplant. 2002;8:377-386. 10. Storb R, Deeg HJ, Whitehead J, et al. Methotrexate and cyclosporine compared with cyclosporine alone for prophylaxis of acute graft versus host disease after marrow transplantation for leukemia. N Engl J Med. 1986;314:729-735. 11. McSweeney PA, Niederwieser D, Shizuru JA, et al. Hematopoietic stem cell transplantation in older patients with hematologic malignancies: replacing high-dose cytotoxic

12.

13.

14.

15.

16. 17.

18.

19.

20. 21. 22.

23.

24.

25.

26.

27.

28.

therapy with graft-versus-tumor effects. Blood. 2001;97:33904000. Nachbaur D, Eibl B, Kropshofer G, et al. In vivo T cell depletion with low-dose rabbit antithymocyte globulin results in low transplant-related mortality and low relapse incidence following unrelated hematopoietic stem cell transplantation. J Hematother Stem Cell Res. 2002;11:731-737. Glucksberg H, Storb R, Fefer A, et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HLA-matched sibling donors. Transplantation. 1974;18:295-301. Shulman HM, Sullivan KM, Weiden PL, et al. Chronic graftversus-host syndrome in man. A long term clinicopathologic study of 20 Seattle patients. Am J Med. 1980;69:204-211. Marr K, Seidel K, Slavin M, et al. Prolonged fluconazole prophylaxis for fungal infections after marrow transplantation—a prospective, randomized, double-blind study. Blood. 2002;96:2055-2061. Kaplan EL, Meier P. Non parametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457-481. Gooley TA, Leisenring W, Crowley J, Storer BE. Estimation of failure probabilities in the presence of competing risks: new representation of old estimators. Stat Med. 1999;18:695-706. Min C-K, Kim SY, Lee MJ, et al. Vascular endothelial growth factor (VEGF) is associated with a reduced severity of acute graft-versus-host disease and nonrelapse mortality after allogeneic stem cell transplantion. Bone Marrow Transplant. 2006;38: 149-156. Iguchi A, Kobayashi R, Yoshida M, et al. Vascular endothelial growth factor (VEGF) is one of the cytokines causative and predictive of hepatic veno-occlusive disease (VOD) in stem cell transplantation. Bone Marrow Transplant. 2001;27:1173-1180. Ferrara N, Gerber HP, LeCouter J. The biology of VEGF and its receptors. Nat Med. 2003;9: 669-676. Ferrara JL, Reddy P. Pathophysiology of graft-versus-host disease. Semin Hematol. 2006;43:3-10. Gunsilius E, Petzer A, Stockhammer G, et al. Thrombocytes are the major source for soluble vascular endothelial growth factor in peripheral blood. Oncology. 2000;58:169-174. Salven P, Orpana A, Joensuu H. Leukocytes and platelets of patients with cancer contain high levels of vascular endothelial growth factor. Clin Cancer Res. 1999;5:2978-2989. Yu J, Dolter KE. Production of activin A and its roles in inflammation and hematopoiesis. Cytokines Cell Mol Ther. 1997; 3:169-177. Shulman HM, Gown AM, Nugent DJ. Hepatic veno-occlusive disease after bone marrow transplantation. Immunohistochemical identification of the material within occluded central venules. Am J Pathol. 1987;127:549-558. Eltumi M, Trivedi P, Hobbs JR, et al. Monitoring of venoocclusive disease after bone marrow transplantation by serum aminopropeptide of type III procollagen. Lancet. 1993;342:518521. Rio B, Bauduer F, Arrago JP, Zittoun R. N-terminal peptide of type III procollagen: a marker for the development of hepatic veno-occlusive disease of the liver in children undergoing bone marrow transplantation. Blood. 1994;83:3036-3040. Anscher MS, Peters WP, Reisenbichler H, Petros WP, Jirtle RL. Transforming growth factor beta as a predictor of liver and lung fibrosis after autologous bone marrow transplantation for advanced breast cancer. N Engl J Med. 1993;328:1592-1598.