Annals of Hematology

, Volume 89, Issue 8, pp 789–794

Bone marrow angiogenesis in multiple myeloma and its correlation with clinicopathological factors

Authors

    • Department of PathologySanjay Gandhi Postgraduate Institute of Medical Sciences
  • Seema Sharma
    • Department of PathologySanjay Gandhi Postgraduate Institute of Medical Sciences
  • Vinita Agrawal
    • Department of PathologySanjay Gandhi Postgraduate Institute of Medical Sciences
  • Uttam Singh
    • Department of BiostatisticsSanjay Gandhi Postgraduate Institute of Medical Science and Research
Original Article

DOI: 10.1007/s00277-010-0919-z

Cite this article as:
Rana, C., Sharma, S., Agrawal, V. et al. Ann Hematol (2010) 89: 789. doi:10.1007/s00277-010-0919-z

Abstract

Increased angiogenesis has been found to be an adverse prognostic factor in solid tumors but evidences show that angiogenesis also plays an important role in hematological malignancies including multiple myeloma (MM). In this report we studied the various angiogenesis parameters like microvessel density (MVD) and total vascular area (TVA), on bone marrow biopsies in 50 newly diagnosed cases of MM. The aim was to study bone marrow angiogenesis in MM using light microscopy (MVD-A) and computerized image analyzer (MVD-B and TVA) and correlate it with clinical features, laboratory findings, histological features, and response to treatment on follow-up. Bone marrow biopsies of test cases (n = 50) were immunohistochemically stained with CD34 for visualization of microvessels. MVD-A (range 8–80; mean 50.4; SD 17.5), MVD-B (5.2–33.2; mean 16.3; SD 5.1), and TVA in percentage (range 0.42–7.20; mean 2.8; SD 1.5) were measured. Ten age- and sex-matched controls were studied and their parameters were taken as grade I. There was a significant correlation between these angiogenesis parameters (MVD-A vs MVD-B, Pearson’s correlation coefficient (pcc) = 0.724; MVD-A vs TVA, pcc = 0.370; MVD-B vs TVA, pcc = 0.406). The angiogenesis was significantly higher in cases as compared to controls. Patients with residual disease had a higher MVD as compared to the complete responders. High tumor burden and diffuse pattern of infiltration were also associated with grade III MVD and TVA. Hence, it can be concluded that angiogenesis correlates with other histological features associated with prognosis and is also a good predictor for complete response in patients with multiple myeloma.

Keywords

Multiple myelomaAngiogenesisClinicopathological factorsMicrovessel densityComputerized image analysis

Introduction

Angiogenesis is the formation of new blood vessel from pre-existing vasculature which occurs in either pathological or physiological conditions. It occurs physiologically during embryonal growth, wound healing, and in the female genital tract during menstrual cycle as well as in several pathological conditions such as neoplasia. It is obligatory in the enhancement of progression, i.e., growth, invasion, and metastasis of solid tumors. In absence of angiogenesis, tumor cannot grow beyond 1 to 2 mm3 in size [1].

In recent years, bone marrow (BM) angiogenesis is indicated to be involved in the pathogenesis and progression of certain hematological malignancies like acute lymphoid leukemia, acute myeloid leukemia, myelodysplastic syndrome, Hodgkin’s lymphoma, low/high grade non-Hodgkin’s lymphomas, and multiple myeloma [2].

Multiple myeloma (MM) is a B-cell malignancy that accounts for 1% of all cancers and about 10% of all malignant hematological neoplasms. Clinically, it is characterized by a pentad of: (a) anemia, (b) elevated serum and urine monoclonal paraprotein levels, (c) abnormal bone radiographs and bone pains, (d) hypercalcemia, and (e) renal insufficiency [3]. MM was the first hematological malignancy in which a prognostic relevance of angiogenesis was demonstrated. Previous studies have suggested that microvessel density was significantly increased in MM compared to monoclonal gammopathy of unknown significance (MGUS) and moreover active versus non-active myeloma [4]. Hence, possibility of using anti-angiogenic agents in patients with poor outcome after conventional therapy might be a novel strategy for treatment.

Different techniques have been developed for visualizing and estimating the number of microvessels in various tissues. Immunohistochemical staining for substances present in the blood vessels such as CD34, Von willebrand factor, CD31, CD105, and others, allow excellent visualization of microvessels. Quantification of microvessel density (MVD) can be done either in terms of microvessel count per field or by calculating the percentage of the surface area which is covered by microvessels. Different approaches have been adopted for estimating the degree of angiogenesis or MVD, including visual grading [5, 6], vessel count per defined area, use of grids for counting, and computerized image analysis (CIA) [7].

Few studies have been done to study angiogenesis in multiple myeloma and most of them used light microscopic analysis of BM biopsies. There are very few studies in literature that used CIA for calculating microvessel density. The aim of the present study was to evaluate various angiogenesis parameters in BM biopsies of newly diagnosed patients of MM by light microscopy and computerized image analyzer in newly diagnosed patients of MM and correlated them with various clinicopathological factors associated with prognosis.

Patients and methods

The study was a retrospective study where 50 newly diagnosed cases of MM were retrieved between 2000 and 2007. These patients received treatment at SGPGIMS, Lucknow, India, and were diagnosed on the basis of WHO criteria for MM. Adequate pre-treatment and post-treatment BM biopsies as well as aspirates were available in all these 50 patients. Angiogenesis in bone marrow biopsy was assessed by MVD count after immunostaining with CD34 (Fig. 1) and it was correlated with other clinical findings, laboratory factors, histological factors, and response to treatment on follow-up.
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-010-0919-z/MediaObjects/277_2010_919_Fig1_HTML.gif
Fig. 1

CD34 +ve microvessels on immunohistochemistry (bone marrow biopsy) a control, b low microvessel density area, and c high microvessel density area

Ten age- and sex-matched bone marrow biopsies performed for non-malignant conditions were taken as control for grading the angiogenesis parameters (Fig. 1a). Lymph node sections were used as positive controls for immunohistochemical staining.

Bone marrow biopsies processing and histological staining

Bone marrow biopsies were collected in B5 fixative solution followed by decalcification using 5% trichloroacetic acid in normal saline and then processed overnight for paraffin embedding. Sections were cut at 3 to 5 μm and stained with hematoxylin and eosin. Reticulin staining was done when required

Marrow aspirations were performed at posterior superior iliac spine by modified Jamshidi needle and multiple smears were made and stained with leishman stain.

Immunohistochemistry

Immunohistochemical staining for angiogenesis was performed in bone marrow biopsies of all the 50 patients of MM as well as ten controls, using streptavidin biotin method. Antigen retrieval was done by microwave method. Endogenous peroxidase activity was blocked by treating sections with 3% hydrogen peroxide in methanol in the dark for 30 min. Then these sections were incubated overnight at 4°C in humid chamber with the monoclonal antibody QBEND10 to CD34 (Dako, Denmark) at 1:25 dilution.

Histological staging and grading

All slides (n = 50; H&E) were evaluated for confirmation of original diagnosis, staged, and graded histologically as described by Bartl et al. [8].

Estimation of angiogenesis parameters

Microvessel density was estimated using two methods.
  1. 1.

    Light microscopy (manually at ×40; MVD-A)

     
All cases were evaluated on an Olympus microscope by two independent pathologists. Slides were first examined at ×10 magnification and three areas of maximum MVD called “hot spot” were identified. In each of these hot spots, microvessels (capillaries and venules) were counted at ×40. Microvessels were defined as endothelial cells either singly or clustered in nests or tubes, clearly separated from one another, with or without lumen not exceeding 10 µm (i.e., not more than 1.5 times the endothelial cell nucleus in transverse diameter).Grading of test cases was performed taking into account the MVD-A of controls, such that the MVD-A of all the controls included was Grade I MVD-A of the test cases. MVD-A was graded as follows: Grade I, ≤25 microvessels; Grade II, >25 microvessels but ≤50 microvessels; and Grade III, >50 microvessels/high power field (×40).
  1. 2.

    Computerized image analysis

     
The total count of microvessels per 50,000 µm2 of marrow area (MVD-B) and the total area occupied by microvessels (total vascular area (TVA)) expressed in percentage were evaluated on an Olympus Digital Image Analyzer using Image Pro-Plus 6.1 software. The hot spots were examined at ×40 magnification for CIA. Using the computerized color segmentation method, microvessel surface area, i.e., total vascular area was determined on the similar digital images on which MVD-B were determined, by taking into account the cellular area, fatty spaces and excluding the bone lamellae, dense connective tissue, necrosis, hemorrhagic areas, and non-marrow empty spaces.

Grading of MVD-B and TVA for test cases and control was done in a manner similar to that of the MVD-A grading, i.e., taking all control as Grade I. MVD-B was graded as follows: Grade I ≤10 microvessels; Grade II, >10 and ≤20 microvessels; and Grade III, >20 microvessels, per 50,000µmm2 . TVA percentage was graded as follows: Grade I ≤1%; Grade II >1% but ≤3%; and Grade III >3% vascular area per total marrow area.

Treatment and follow-up

The patients were treated by either VAD or combination of prednisolone and thalidomide. Response criteria were defined as (a) remission, <5% plasma cells in the bone marrow with disappearance of M protein; (b) residual, >50% reduction in the M protein levels in serum with >5% plasma cells in bone marrow; and (c) refractory, no response to the initial therapy.

Statistical analysis

Correlation between the manual and computerized image analysis in evaluating the angiogenesis parameters were done by Pearson’s correlation coefficient (pcc). Clinical and laboratory parameters related to prognosis of MM were correlated with the MVD using Spearman and Pearson’s correlation coefficient. Significance of association between the bone marrow findings and response to chemotherapy with the microvessel density and total vascular area was evaluated using the tests of proportion. P value <0.05 was considered as statistically significant. The statistical analysis was carried out with the SPSS Software Version 15.

Results

The study group consisted of 50 newly diagnosed patients of MM with mean age of 55.9 years (range of 29–73 years). There was male predominance with M–F = 3.2:1. Clinical, hematological, biochemical, and immunological details along with adequate BM biopsies as well aspirates were available for all cases (Table 1).
Table 1

Demographic, clinical, hematological, biochemical, immunological, and histopathological details of 50 cases

Demographic

Biochemical parameters

Age range (years)

29–73

S. creatinine (>2 mg/dl)

21 (42%)

Mean age (years)

55.9

S. calcium (>12 mg/dl)

04 (08%)

Median age (years)

57

  

Male–female ratio

3.2:1

  

Presenting complaints

Peripheral blood findings

Pallor/anemia

46 (92%)

Hb (<10 g/dl)

47 (94%)

Backache

35 (70%)

Normal platelet count

36 (72%)

Bone pain

34 (68%)

Platelet count (<1,000,000/µl)

14 (28%)

Weakness/fatigue

25 (50%)

Raised ESR

40 (80%)

Renal symptoms

19 (38%)

  

Bleeding/bruising

08 (16%)

  

Physical examination findings

Plasma cells percentage

Bone tenderness

35 (70%)

>50%

32 (64%)

Collapse/fracture

18 (36%)

<50%

18 (36%)

Hepatomegaly

08 (16%)

  

Splenomegaly

04 (08%)

  

Soft tissue/bony swelling

08 (16%)

  

Lymphadenopathy

01 (02%)

  

Clinical stage

Plasma cell morphology

Stage I 03

03 (06%)

Mature (low grade)

16 (32%)

Stage II

17 (34%)

Immature (intermediate grade)

28 (56%)

Stage III 3

30 (60%)

Plasmablastic (high grade)

06 (12%)

Electrophoresis for M component

Pattern of infiltration

“M” component in serum

45 (90%)

Interstitial

03 (06%)

“M” component in urine

19 (38%)

Nodular

03 (06%)

  

Mixed

16 (32%)

Diffuse (packed)

28 (56%)

Interstitial-sheet

00 (00%)

Immunofixation results (n = )

09 (25.0%)

Response to chemotherapy

 

19 (52.8%)

  

IgGλ

04 (11.1%)

Remission

11 (22%)

IgGκ

04 (11.1%)

Residual disease

29 (58%)

IgA λ

 

Refractory

10 (20%)

Light chain

   
MVD-A (range 8–80; mean 50.4; SD 17.5), MVD-B (range 5.2–33.2; mean 16.3; SD 5.1), and TVA (range 0.42–7.20; mean 2.8; SD 1.5) were measured in bone marrow biopsies after immunohistochemically staining the microvessels with CD34. In controls the range of these angiogenic parameters was: MVD-A, 8–24 vessels/hpf (mean 12.6); MVD-B 2.5–13.8 vessels/hpf (mean 8.9); and TVA 0.2–3.4% (mean1.7%). On comparing the cases with the control it was found that BM angiogenesis was significantly higher in MM patients than in controls (P value < 0.001). A correlation between angiogenesis parameters (MVD-A, MVD-B, and TVA) was based on continuous numerical values of each parameter. On analysis it was found that all these angiogenesis parameters were significantly correlated with each other such that MVD-A vs MVD-B, pcc = 0.724; MVD-A vs TVA, pcc = 0.370; and MVD-B vs TVA, pcc = 0.406 (Table 2).
Table 2

Grade of angiogenesis parameters and correlation between them in pre-treatment biopsies

Grades of angiogenesis parameters

 

Grade I

Grade II

Grade III

MVD-A

05

17

28

TVA

04

24

22

Correlation between angiogenesis parameters (MVD-A, MVD-B, TVA)

 

MVD-A vs MVD-B

MVD-A vs TVA

MVD-B vs TVA

Pearson correlation coefficient

0.724

0.370

0.406

Significance (p value)a

0.000

0.008

0.003

aCorrelation is significant at the 0.01 level (two-tailed)

MVD showed a significant association with plasma cell percentage and morphology. It was seen that the patients with grade III MVD-A had a high tumor burden, i.e., higher plasma cell percentage (>50%; P value < 0.004). Majority of patients with mature plasma cell morphology had MVD <50 vessels/hpf while most of the cases with immature morphology had grade III MVD (>50 vessels/hpf) and this correlation was also statistically significant (P value < 0.05) (Table 3).
Table 3

Factors associated significantly with MVD and TVA

Prognostic factors vs high grade MVD-A

P value

Plasma cell percentage >50%

0.004

Immature plasma cell morphology

0.038

Diffuse pattern of infiltration

<0.001

Residual disease

0.03

Prognostic factors vs TVA

Plasma cell percentage >50%

0.037

Diffuse pattern of infiltration

0.005

When the MVD was correlated with the extent of infiltration, it was noticed that most of the patients with a higher grade of angiogenesis had a diffuse pattern of plasma cell infiltration in bone marrow biopsy (P value < 0.001). It was also found that complete responders had a lower grade of angiogenesis than those who had residual disease (P value < 0.05; Table 3).

Total vascular area was also correlated with other factor like microvessels density. Total vascular area was higher, i.e., TVA >3% in patients with high tumor load as compared to cases with low tumor burden (P < 0.05). Diffuse pattern of infiltration was also more frequently encountered in cases with highest total TVA≥3% (Table 3).

Figure 2 gives a diagrammatic representation of the number of patients in different categories of MVD and factors that significantly correlated with MVD.
https://static-content.springer.com/image/art%3A10.1007%2Fs00277-010-0919-z/MediaObjects/277_2010_919_Fig2_HTML.gif
Fig. 2

Number of patients in various significant prognostic factors in different grades of MVD

Discussion

Immunohistochemistry has provided an objective method for assessing degree of neovascularization. It has been shown to be a significant and independent prognostic indicator of various malignancies including myeloma. MM was the first hematological malignancy in which a significant correlation of angiogenesis with prognosis and survival could be identified [4]. Vacca et al. [9] demonstrated for the first time that BM-MVD was significantly increased in MM compared to MGUS and moreover active versus non-active myeloma. They hypothesized that progression from MGUS to myeloma is accompanied by an increase in BM-MVD.

There are only a few studies on angiogenesis in multiple myeloma. In all these studies angiogenesis was measured as microvessel density in the “hot spots”. These measurements were done mostly based on light microscopic analysis of bone marrow biopsies. There are only two studies by Bhatti et al. [7] and Rajkumar et al. [10] in literature that used computerized image analyzer to study the various angiogenesis parameters in MM. Bhatti et al. [7] performed both MVD as well as TVA using CIA and compared them with the various prognostic factors of MM.

We also used light microscopy as well CIA and found that all angiogenesis parameters were significantly higher in myeloma cases as compared to controls (P value < 0.0001). These results were in accordance with the results of previous studies done by Rajkumar et al. [6] and Bhatti et al. [7]. MVD by light microscopy as well as CIA correlated with each other, hence any of the two can be used for MVD analysis, the only difference being that CIA provides easy visualization of microvessel and helps in more accurate counting by avoiding overlapping. Since CIA is costly method, hence light microscopy can be used where computerised image analyser is not available.

Vacca et al. [11] postulated that higher plasma cell labeling index and plasma cell percentage are associated with active myeloma as compared to non-active myeloma. While Rajkumar et al., Munshi et al. [12], and Singhal et al. [13] found no correlation between plasma cell infiltration and MVD. Due to these conflicting results, we investigated this relationship and demonstrated a significant correlation between high tumor burden (plasma cells >50%) and grade III MVD. In our study, there were five patients with plasma cells <50% but had a grade III MVD. When these patients were followed-up, it was found that all of them had a residual or refractory disease after treatment. These findings indicate that angiogenesis is a better factor for predicting the outcome of the patient after treatment as compared to the plasma cell number.

We also demonstrated that patients with mature plasma cells morphology had a grade I MVD, while most of the patients with plasmablastic morphology had higher angiogenesis, i.e., grade III MVD. Carter et al. [14] had predicted a shorter median survival for plasmablastic myeloma as compared to mature and immature myeloma. Similarly, diffuse pattern of infiltration was also associated with grade III MVD and there was not even a single case of diffuse pattern of infiltration with grade I MVD or TVA.

Only two studies have calculated the total vascular areas in BM biopsies of MM. Vacca et al. [9] used planimetric method while Bhatti et al. [7] used CIA. We also studied the total vascular area using CIA and found that grade III MVD was significantly associated with high tumor burden (plasma cell >50%) and diffuse pattern of infiltration had grade III MVD (P < 0.05).

Our study demonstrated that angiogenesis was higher in patients with residual disease as compared to complete responders. This indicates that angiogenesis can a good indicator of response and hence should be done in all newly diagnosed cases of MM as this could predict the outcome of the patient.

Since angiogenesis is known to play a very important role in progression of myeloma from non-active stage to active stage, it can be a fruitful target for treatment of multiple myeloma. Experimental evidences have been provided that angiogenesis is not negatively affected by drug resistance due to genetic stability of endothelial cells in contrast to tumor cells. On this basis various strategies for anti-angiogenic treatment have been developed which include interference with angiogenic stimulators (e.g., vascular endothelial growth factor (VEGF), HGF, and bFGF), angiogenic factor receptors (e.g., VEGF-receptor signaling), extracellular matrix interactions, inhibiting oncogenes controlling the angiogenic response, and proteolysis [2].

It can be concluded from our study that angiogenesis was increased in patients with MM as compared to controls. Manual (MVD-A) and computerized microvessel density (MVD-B) correlated with each other as well as the total vascular area occupied by the microvessels. Secondly, angiogenesis is higher in patients with residual disease as compared to those with complete response after treatment, hence when performed in newly diagnosed cases of MM degree of angiogenesis can predict the outcome. Thirdly, patients with increased plasma cell burden, immature plasma cell morphology, and diffuse pattern of infiltration had a higher microvessel density. Since the abovementioned factors are known independent prognostic factors and angiogenesis correlate with them, angiogenesis can be included as one of the prognostic factor for multiple myeloma. As stated above, it can even be a better factor; however, a more extensive study is required to prove it. Finally, angiogenesis can be a target for various anti-angiogenic therapies thereby opening gateways towards new treatment approach which could probably be less harmful and more effective thereby improving patient survival from this devastating disease.

Copyright information

© Springer-Verlag 2010