Increased bone marrow angiogenesis in B cell chronic lymphocytic leukemia

Transcription

1 (2000) 14, Macmillan Publishers Ltd All rights reserved /00 $ Increased bone marrow angiogenesis in B cell chronic lymphocytic leukemia AR Kini 1, NE Kay 2,3 and LC Peterson 1 1 Northwestern University Medical School, Chicago, IL; 2 Virginia Piper Cancer Institute, Minneapolis, MN; and 3 Mayo Clinic, Rochester, MN, USA Recent studies have shown that angiogenesis may be involved in the pathogenesis of hematopoietic malignancies, apart from its well-characterized role in the growth and metastasis of solid tumors. In this study, we quantified the degree of angiogenesis in B cell chronic lymphocytic leukemia (B-CLL) by measuring the microvessel density and hotspot density in bone marrow trephine biopsy sections with B-CLL involvement (n 12) and compared it to normal bone marrow sections (n 11). The B- CLL samples had a mean microvessel count/high power field (hpf) of 7.64 while the control samples had a mean microvessel count/hpf of 2.11 (P ). The mean hotspot density in the B-CLL sections (14.83/hotspot) was also significantly higher (P ) than the mean hotspot density in control bone marrow sections (7.09/hotspot). Both the microvessel density and hotspot density correlated positively with the clinical stage of the B-CLL patients. In a separate cohort of B-CLL patients, the median urine level of the angiogenic peptide, basic fibroblast growth factor ( pg/g, n 14), was significantly higher (P ) than the bfgf level in normal controls (1084 pg/g, n 58). These results indicate that angiogenesis may be involved in the pathogenesis of B-CLL. (2000) 14, Keywords: chronic lymphocytic leukemia; angiogenesis; bone marrow; microvessel; basic fibroblast growth factor Introduction The hypothesis that angiogenesis is required for the growth of solid tumors, was first proposed nearly 30 years ago. 1 There is now convincing evidence that growth of solid tumors beyond 1 to 2 mm in diameter 2 requires proliferation of new capillaries. Numerous studies have revealed that an angiogenic switch converts a tumor clone to an angiogenic phenotype through alteration of the balance between angiogenic and anti-angiogenic factors. 3 5 The conversion to an angiogenic phenotype leads to rapid growth and metastasis of solid tumors. 2,6 8 Although angiogenesis has been shown to be essential for the growth of solid tumors, the role of angiogenesis in the pathogenesis of hematological malignancies has not been as extensively studied. However, a recent report 9 showed that bone marrows of children with acute lymphoblastic leukemia (ALL) have higher microvessel density as compared to normal controls. In addition, these children had higher urine levels of the angiogenic peptide basic fibroblast growth factor (bfgf) when compared to normal controls. These results indicate that ALL cells are angiogenic, and that angiogenesis plays a role in the pathogenesis of ALL. There is also increased angiogenesis in the bone marrows of patients with multiple myeloma. 10,11 Myeloma cells secrete bfgf, and stimulate angiogenesis in in vitro and in vivo angiogenesis assays. 12 In addition, there is increased angiogenesis in B cell non-hodgkin s lymphomas. 13 These studies, demonstrating the importance of angiogenesis in lymphoid malignancies, led us to Correspondence: LAC Peterson, 251 E Huron Street, Feinberg 7-344, Chicago, IL 60611, USA; Fax: Received 23 November 1999; accepted 15 March 2000 investigate the possible role of angiogenesis in the pathogenesis of B cell chronic lymphocytic leukemia (B-CLL). Specifically, this was done by examining the degree of bone marrow angiogenesis and by evaluating the most important angiogenic factors, bfgf and vascular endothelial growth factor-a (VEGF-A). Materials and methods Patient population The bone marrows with B-CLL involvement were obtained from patients (n = 12) who had bone marrow trephine biopsies at Northwestern Memorial Hospital for clinical assessment of their disease. Nine patients had received no therapy prior to their biopsies; three patients had received chemotherapy. Four patients were classified as stage 0 by the Rai staging system, two patients were stage 1, one patient was stage 2, three patients were stage 3 and two patients were stage 4. There were seven male and five female B-CLL patients with a mean age of 65.8 ± 10.8 (s.d.) years. All the control bone marrows were negative for infiltrative lesions and were obtained from patients undergoing staging biopsies for lymphoma (n = 2) or breast carcinoma (n = 6), and from patients with benign disorders such as anemia (n = 2) and thrombocytopenia (n = 1). The mean age of the control patients was 56.3 ± 16.4 (s.d.) years. Urine samples for assessment of bfgf and VEGF-A levels were obtained at the Virginia Piper Cancer Institute from patients (n = 14) who were diagnosed as B-CLL using the standard clinical and morphologic evaluation. Nine patients had received no treatment; five patients had received chemotherapy. Eight patients were classified as stage 0 by the Rai staging system, one patient was stage 1, three patients were stage 3 and two patients were stage 4. There were six male patients and eight female patients with a mean age of 64.5 ± 11.9 (s.d.) years. Urine levels of bfgf were also measured in healthy controls (n = 58) from the Boston area and the Minneapolis-St Paul area. These controls included 23 males and 35 females, with a mean age of 35.8 ± 9.2 (s.d.) years. Control urine VEGF-A levels were measured in healthy volunteers (n = 7) from the Minneapolis-St Paul area (three males, four females; mean age: 46 ± 7.1 years). There was no significant difference between the female control subjects of menstruating age and post-menopausal female controls, with regard to urine bfgf and VEGF-A levels. The same analysis in the female B-CLL patients also revealed no significant difference in urine bfgf and VEGF-A levels. There was no significant correlation between age and urine bfgf or VEGF-A levels, in the B-CLL patients or in the controls (data not shown). Measurement of microvessel numbers The degree of angiogenesis in B-CLL was quantified by measuring the microvessel numbers in bone marrow trephine

2 biopsy sections that had nodular or diffuse involvement with B-CLL (n = 12), and comparing it to normal control bone marrow sections (n = 11). The blood vessels in bone marrow trephine sections were highlighted by immunohistochemistry using antibodies to CD31, CD34 and Factor VIII. The immunohistochemistry was performed on B5-fixed paraffin blocks using standard techniques described previously. 14 In our analysis, the microvessel counts obtained by using these different antibodies were similar. In this study we used the data obtained from CD34 staining, since in our experience, CD34 produced the least amount of background staining. We used a modification of the criteria used by Weidner et al 15,16 to count microvessels. Microvessel numbers measured using these criteria, have been shown to be a reliable and reproducible measure of angiogenesis. Microvessels were defined as any endothelial cell or group of endothelial cells, distinct from other endothelial cells, non-endothelial cells, and connective tissue. The presence of a lumen was not considered necessary for defining a microvessel, and the length of the microvessel was not a factor in this definition. Microvessel counts were performed by two observers (ARK and LCP) simultaneously, using a double-headed light microscope. Only structures agreed on by both observers to be microvessels were included in the counts. The microvessel counts were performed at 600 ( 60 objective, 10 ocular) magnification for the entire bone marrow trephine biopsy section. The microvessel density was defined as the total number of microvessels divided by the number of 600 high power fields, and is expressed as microvessels/high power field (hpf). We also compared the mean hotspot counts of the B-CLL bone marrow biopsies to the control marrows. A hotspot was defined as the high power field with the highest microvessel count. To confirm the reproducibility of our results, we performed serial sections on six (three B-CLL, three controls) of our cases. The microvessel counts and hotspot counts in serial sections were found to be consistent, with a coefficient of variation of 5% in all cases. The bone marrow cellularity and extent of bone marrow involvement was estimated by examining the trephine biopsy sections under low power ( 10 objective, 10 ocular). To confirm the accuracy of the cellularity estimates, digitized images were obtained by using a digital camera, or by scanning photographs of the trephine biopsy sections. The digitized images were then processed using the Scion Image program (Scion Corporation, Frederick, MD, USA). Using a threshold pixel level equivalent to the intensity of the fat spaces, the proportion of bone marrow space occupied by cellular elements was obtained. To account for the difference in cellularity between the B-CLL and control bone marrow biopsies, the mean microvessel count for the control marrows was adjusted by multiplying by a cellularity factor. This factor (1.4) was obtained by dividing the mean cellularity of the B-CLL bone marrow sections with the mean cellularity of the control bone marrow sections. Measurement of urine levels of bfgf and VEGF-A Urine levels of the angiogenic peptides bfgf and VEGF-A were quantified using the Quantikine bfgf and VEGF-A assays (R&D Systems, Minneapolis, MN, USA), which are solid-phase enzyme-linked immunosorbent assays. The minimum detectable dose of bfgf using the Quantikine kit is 1 pg/ml. The minimum detectable dose of VEGF-A is 9 pg/ml. The samples were collected as clean-catch specimens using sterile techniques, and were either assayed a b Figure 1 Bone marrow sections with B-CLL involvement have more microvessels as compared to normal sections. (a) There are numerous microvessels in a B-CLL bone marrow core section, highlighted by CD34 immunohistochemical staining (brown). Similar areas with high microvessel counts were found throughout the B-CLL marrow sections. The high power field with the highest microvessel count was designated as a hotspot. Original magnification ( 400). (b) In comparision, normal control bone marrow sections had fewer microvessels. A single microvessel (brown) is evident in this field. Original magnification ( 400). immediately or stored frozen at 70 C. To account for differences in urine concentration, the VEGF-A and bfgf values were normalized to urine creatinine levels. Statistical analysis The Shapiro Wilk test was used to assess distribution of the sample population. The t-test was used to compare means if the samples followed a normal distribution. The Mann Whitney rank sum test was used to compare medians if the samples did not follow a normal distribution. Correlation coefficients (R) were determined by using a linear regression model. The F test was used to test whether the correlation coefficient was significant. In all statistical analyses, a P value 0.05 was considered to be significant. All the analyses were performed using the Prophet 5.0 program (BBN Systems and Technologies, Cambridge, MA, USA). 1415

3 1416 Results Microvessel counts in B-CLL bone marrows The degree of angiogenesis in B-CLL and control bone marrow trephine biopsy sections was quantified by measuring the number of microvessels. Immunohistochemical staining using antibodies to CD34 revealed more microvessels in bone marrow trephine biopsy sections with B-CLL involvement (Figure 1a) compared to normal control marrows (Figure 1b). Heterogeneity of the microvessels in the B-CLL cases was evident with some vessels being long and tortuous, while others were short and lacked lumina. The mean microvessel count/hpf (7.64) in B-CLL bone marrows was significantly higher than the microvessel count/hpf (2.11) in control normal bone marrows (P = ; Figure 2a). We also compared the hotspot density in the B-CLL bone marrows to the control bone marrows. Areas of high microvessel density or hotspots may represent the emergence of a tumor clone with a higher angiogenic potential. 17 The mean hotspot density in the B-CLL sections (14.83/hotspot) was also significantly higher (P = ; Figure 2b) than the mean hotspot density in control bone marrow sections (7.09/hotspot). Since the mean cellularity in B-CLL bone marrows was higher than in the control bone marrow sections, the higher microvessel count in the B-CLL bone marrows could reflect the higher cellularity of the B-CLL bone marrows. To address this possibility, we multiplied the microvessel counts in the control samples by a factor (1.4) equal to the difference in cellularity between B-CLL and control bone marrows. After performing this adjustment, the mean microvessel count/hpf was still significantly higher in the B-CLL bone marrows as compared to controls (P = ; Figure 2c). Correlation of clinical stage and bone marrow involvement with microvessel numbers To assess the potential impact of angiogenesis on disease progression, we examined the relationship between the clinical stage and the corresponding microvessel count/hpf for each case. There was a significant positive correlation between the Rai clinical stage at the time of the biopsy, and microvessel density (R = 0.81, P = ; Figure 3a). The hotspot density also positively correlated with clinical stage (R = 0.66, P = ; Figure 3b). We also correlated the extent of bone marrow involvement by B-CLL, and the microvessel numbers. The extent of bone marrow involvement was the proportion of the trephine biopsy section occupied by the B-CLL lymphocytes. There was a positive correlation between both microvessel density (R = 0.78, P = 0.003, Figure 4a) and hotspot density (R = 0.59, P = 0.043, Figure 4b), with bone marrow involvement. Increased urine levels of bfgf in B-CLL We also measured the urine levels of the angiogenic peptides bfgf and VEGF-A in B-CLL patients and controls. The results are summarized in Table 1. The median bfgf level in B-CLL patients ( pg/g) was higher (P = ) than in normal controls (1084 pg/g). The median urine VEGF-A level was also higher in B-CLL patients than controls, though this difference was not significant. There was no significant correlation between urine bfgf or VEGF-A levels of these patients and their corresponding clinical stage. Discussion Recent studies 9,10,12 indicate that angiogenesis may play a role in the pathogenesis of leukemias, apart from its well-established role in solid tumors. In this study we show that bone marrow trephine biopsy sections with B-CLL involvement have a higher microvessel density compared to control bone marrow biopsies. There was a positive correlation between the microvessel counts and extent of bone marrow involvement by the B-CLL lymphocytes. These results indicate that the B-CLL cells may be angiogenic in the bone marrow. While the B-CLL bone marrows had a higher cellularity than the control bone marrows, the higher cellularity alone did not account for the difference in microvessel counts. This indicates that angiogenesis in leukemia is an inefficient process, Figure 2 B-CLL bone marrow sections have higher mean microvessel counts and hotspot counts compared to normal marrow sections. (a) The mean microvessel count/high power field (hpf) in B-CLL bone marrow core sections (n = 12) was 7.64, while the mean microvessel count/hpf in control bone marrow sections (n = 11) was 2.11 (P = , two-sample unequal-variances t-test). The error bars represent the standard error of mean. (b) The B-CLL bone marrow sections marrows (14.83/hotspot) had a significantly (P = , two-sample equal-variances t-test) higher hotspot density in comparison to control marrows (7.09/hotspot). The error bars represent the standard error of mean. (c) The mean microvessel count/hpf was significantly higher in the B-CLL marrow sections (7.64/hpf, n = 12) compared to the normal marrow sections (2.95/hpf, n = 11), after adjusting the mean microvessel counts in the control marrow sections to account for the difference in cellularity. P = , two-sample unequal-variances t-test. The error bars represent the standard error of mean.

4 1417 Figure 3 Correlation between clinical stage and microvessel numbers. (a) There is a significant positive correlation between the Rai clinical stage and bone marrow microvessel count/high power field (hpf). The bold line represents the line of means. The broken lines represent the 95% confidence limits for the line of means. R (correlation coefficient) = 0.81, P = , F test. (b) There is a significant positive correlation between the Rai clinical stage and hotspot density. The bold line represents the line of means. The broken lines represent the 95% confidence limits for the line of means. R (correlation coefficient) = 0.66, P = , F test. similar to solid tumor angiogenesis, 17 or that the requirement for angiogenesis may not increase in a simple linear fashion with higher cellularity. The hotspot density was also higher in B-CLL bone marrows as compared to normal marrows. The hotspot is defined as the high power field with the highest microvessel count, and may represent emergence of a tumor clone with a higher angiogenic potential. 17 Hotspots may also be responsible for facilitating dissemination of the disease. 16 Both the microvessel density and the hotspot density correlated positively with stage, implying that increased angiogenesis may be involved in progression of the disease. We also found that patients with B-CLL have significantly higher urine levels of the angiogenic peptide bfgf than normal controls. Similar high urine levels of bfgf have been found in most solid tumors 18 and in ALL. 9 The source of the urine bfgf may be the B-CLL lymphocytes, since previous Figure 4 Correlation between extent of marrow involvement and microvessel numbers. (a) There is a positive correlation between the extent of bone marrow involvement by B-CLL cells and bone marrow microvessel count/high power field (hpf). The bold line represents the line of means. The broken lines represent the 95% confidence limits for the line of means. R (correlation coefficient) = 0.78, P = 0.003, F test. (b) There is a positive correlation between the extent of bone marrow involvement and hotspot density. The bold line represents the line of means. The broken lines represent the 95% confidence limits for the line of means. R (correlation coefficient) = 0.59, P = 0.043, F test. Table 1 Urine levels of bfgf and VEGF-A Median bfgf levels Median VEGF-A levels (pg/g creatinine) (ng/g creatinine) B-CLL patients ( ) 159 (13 314) Controls 1084 ( ) 100 (98 294) The median urine bfgf level in B-CLL patients (n = 14) was significantly higher than in normal controls (n = 58) (P = , Mann Whitney rank sum test). The median urine VEGF-A level in B-CLL patients (n = 14) was higher than controls (n = 7), athough this difference was not significant (P = 0.322, Mann Whitney rank sum test). The numbers in parentheses represent the respective ranges.

5 1418 studies have shown high intracellular levels of bfgf in B-CLL cells. 19 In addition to the B-CLL lymphocytes, it is possible that the bfgf is also derived from other sources such as endothelial cells and stromal cells. While the urine bfgf levels were found to be higher in B-CLL patients, there was no significant correlation between stage of the disease and urine bfgf levels. This suggests that other angiogenic factors may also be involved in the progression of the disease. Angiogenesis may play a role in the pathogenesis of B-CLL due to increased perfusion in the bone marrow and extramedullary tissues, helping the accumulation of the leukemic B cells. In addition to the perfusion effect there could also be a paracrine effect similar to that seen in solid tumors and in acute leukemias. For example, in acute myelogenous leukemia, leukemic cells secrete VEGF, and the endothelial cells secrete granulocyte macrophage colony-stimulating factor, which may help the leukemic cells to grow. 20 A similar positive-feedback loop may also exist in B-CLL, though the specific factors involved are likely to be different. It is possible that bfgf produced by the B-CLL lymphocytes may help in the proliferation and migration of endothelial cells, while endothelial cells may produce cytokines that act on the B-CLL cells. bfgf, acting in an autocrine or paracrine fashion, may also be involved in inhibition of apoptosis, thereby aiding in the accumulation of B-CLL lymphocytes. Prior results, 21 showing upregulation of bcl-2 in B-CLL cells after addition of bfgf, indicate that this is a distinct possibility. This pathway may be relevant to B-CLL, since this disease is a low-grade leukemia with slow accumulation of clonal B lymphocytes that are resistant to apoptosis. 22 Elucidation of these pathways would be important in understanding the biology of B-CLL, and could result in the development of novel therapeutic strategies. To our knowledge, this is the first study demonstrating increased angiogenesis in a chronic leukemia. Our data do not yet show that angiogenesis is essential for B-CLL and it is possible that the increased angiogenesis in B-CLL may be an epiphenomenon. Further studies, including a large prospective study in which sequential bone marrow biopsies are performed in concert with increasing clinical stage, and the demonstration of disease regression in response to specific anti-angiogenic manipulations, would be needed to show definitively that angiogenesis is necessary for B-CLL pathogenesis. Acknowledgements We would like to thank Nancy Bone and Susan Connors for technical assistance, and Dr Judah Folkman for helpful suggestions and encouragement. References 1 Folkman J. Tumor angiogenesis: therapeutic implications. New Engl J Med 1971; 285: Folkman J. What is the evidence that tumors are angiogenesis dependent? J Natl Cancer Inst 1990; 82: Bouck N. Tumor angiogenesis: the role of oncogenes and tumor suppressor genes. Cancer Cells 1990; 2: Iruela-Arispe ML, Dvorak HF. Angiogenesis: a dynamic balance of stimulators and inhibitors. Thromb Haemost 1997; 78: Bussolino F, Mantovani A, Persico G. Molecular mechanisms of blood vessel formation. Trends Biochem Sci 1997; 22: Blood CH, Zetter BR. Tumor interactions with the vasculature: angiogenesis and tumor metastasis. Biochim Biophys Acta 1990; 1032: Hanahan D, Folkman J. Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 1996; 86: Zetter BR. Angiogenesis and tumor metastasis. Annu Rev Med 1998; 49: Perez-Atayde AR, Sallan SE, Tedrow U, Connors S, Allred E, Folkman J. Spectrum of tumor angiogenesis in the bone marrow of children with acute lymphoblastic leukemia. Am J Pathol 1997; 150: Vacca A, Ribatti D, Roncali L, Ranieri G, Serio G, Silvestris F, Dammacco F. Bone marrow angiogenesis and progression in multiple myeloma. Br J Haematol 1994; 87: Vacca A, Di Loreto M, Ribatti D, Di Stefano R, Gadaleta-Caldarola G, Iodice G, Caloro D, Dammacco F. Bone marrow of patients with active multiple myeloma: angiogenesis and plasma cell adhesion molecules LFA-1, VLA-4, LAM-1, and CD44. Am J Hematol 1995; 50: Vacca A, Ribatti D, Presta M, Minischetti M, Iurlaro M, Ria R, Albini A, Bussolino F, Dammacco F. Bone marrow neovascularization, plasma cell angiogenic potential, and matrix metalloproteinase-2 secretion parallel progression of human multiple myeloma. Blood 1999; 93: Ribatti D, Vacca A, Nico B, Fanelli M, Roncali L, Dammacco F. Angiogenesis spectrum in the stroma of B cell non-hodgkin s lymphomas. An immunohistochemical and ultrastructural study. Eur J Haematol 1996; 56: Hakimian D, Tallman MS, Kiley C, Peterson L. Detection of minimal residual disease by immunostaining of bone marrow biopsies after 2-chlorodeoxyadenosine for hairy cell leukemia. Blood 1993; 82: Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogenesis and metastasis correlation in invasive breast carcinoma. New Engl J Med 1991; 324: Weidner N. Intratumor microvessel density as a prognostic factor in cancer. Am J Pathol 1995; 147: Rak JW, St Croix BD, Kerbel RS. Consequences of angiogenesis for tumor progression metastasis and cancer therapy. Anticancer Drugs 1995; 6: Nguyen M, Watanabe H, Budson A, Richie JP, Hayes DF, Folkman J. Elevated levels of an angiogenic peptide, basic fibroblast growth factor, in the urine of patients with a wide spectrum of cancers. J Natl Cancer Inst 1994; 86: Menzel T, Rahman Z, Calleja E, White K, Wilson EL, Wieder R, Gabrilove J. Elevated intracellular level of basic fibroblast growth factor correlates with stage of chronic lymphocytic leukemia and is associated with resistance to fludarabine. Blood 1996; 87: Fiedler W, Graeven U, Ergun S, Verago S, Kilic N, Stockschlader M, Hossfeld DK. Vascular endothelial growth factor, a possible paracrine growth factor in human acute myeloid leukemia. Blood 1997; 89: Konig A, Menzel T, Lynen S, Wrazel L, Rosen A, Al Katib A, Raveche E, Gabrilove JL. Basic fibroblast growth factor (bfgf) upregulates the expression of bcl-2 in B cell chronic lymphocytic leukemia cell lines resulting in delaying apoptosis. 1997; 11: Reed JC. Molecular biology of chronic lymphocytic leukemia: implications for therapy. Semin Hematol 1998; 35: 3 13.