Study of significance of bone marrow microvessel density in myeloproliferative neoplasms in correlation with CD34 blasts, mast cell count and fibrosis

Study setting

The current research is a cross sectional study of cases diagnosed over six years duration spanning from January 2016 to December 2022 (five years two months retrospective and 10 months prospective).

Study design

This study follows ‘The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement guidelines. A completed STROBE checklist can be found in the reporting guidelines.⁹

Study sample

We included 66 cases diagnosed as MPN during a six-year period. This comprised of 32 chronic myeloid leukemia (CML), 31 PMF and three essential thrombocythemia (ET) cases.

Ethical consideration

The Institutional Ethics Committee (Reg. No. ECR/541/Inst/KA/2014/RR20; DHR Reg. No. EC/NEW/INST/2020/742), provided ethical clearance on 17^th February 2022 with reference no. IEC KMC MLR 12-2020/444 after submitting the proposal letter beforehand. The ethics committee also waived the informed consent as the study is performed on archived tissue blocks.

Inclusion and exclusion criteria

All bone marrow biopsies diagnosed as MPNs in our laboratory during the study duration were included. Cases in which the blocks and/or slides were not available or the tissue was inadequate for immunohistochemistry (IHC) were excluded.

Data collection

All previously histomorphologically diagnosed MPNs were retrieved from the Pathology department archives and clinical characteristics, including age, gender, symptoms, signs, laboratory investigations, genetic mutations, treatment and follow up details were gathered from the medical records and patient case record files. A total 66 cases of MPN were studied, out of which 32 cases were CML, 31 cases were PMF and three cases were ET. Bone marrow biopsies were received in 10% neutral buffered formalin fixative. Paraffin embedded tissue blocks were made. Sections of four-micron thickness were prepared using a microtome (Leica RM2255) and stained using haematoxylin and eosin (H&E).

Haematoxylin and eosin staining procedure

Preparation of Mayer’s hematoxylin was done by dissolving one gram of hematoxylin in one litre of water in a bottle. Aluminium sulphate was then dissolved in a small amount of water in a separate beaker, followed by addition of 0.2 g of sodium iodate to this solution and mixed properly. This was then added to the hematoxylin solution bottle and mixed properly. Finally, 0.2 g of citric acid crystals were added and mixed.

Eosin stain was prepared using Eosin Y powder (10 g) dissolved in distilled water (1000 ml) and 1 ml of glacial acetic acid. The mixture was filtered to get a total of 1000 ml of eosin.

Following section cutting, deparaffinization was caried out first, by incubating the slides in an oven for 40 minutes. They were then cooled and dipped in xylene two times, first for 15 minutes and the second time for 10 minutes. The slides were air dried after which they were dipped in absolute alcohol twice for one minute each. They were then air-dried following which they were dipped in water for five minutes. Following deparaffinisation, sections were washed in tap water. They were dipped for 15 minutes in Mayer’s hematoxylin and kept for 5–10 minutes in running tap water to ensure ‘bluing’. Next, they were dipped for 30 seconds in eosin stain and then washed in tap water, dried, and mounted.

Microscopic examination of the cases was based on morphology according to the World Health Organisation classification as per the data proforma⁹ by using an Olympus CX 21i microscope. MPNs were typed according to the WHO 2016 classification. This is comprised of chronic myeloid leukemia (CML), BCR-ABL1 positive, chronic neutrophilic leukemia (CNL), polycythemia vera (PV), primary myelofibrosis (PMF) i) Prefibrotic/early stage ii) Overt fibrotic stage, essential thrombocythemia (ET), chronic eosinophilic leukemia, not otherwise specified (NOS), myeloproliferative neoplasm, unclassifiable.¹⁰

Immunohistochemistry (IHC)

Immunohistochemical evaluation with CD34 and mast cell tryptase were done on the archived blocks. IHC staining was performed with CD34 (RTU mouse monoclonal; clone: QBEND/10; PDM050) (BD Biosciences Cat# 550390; RRID:AB_393656) and mast cell tryptase (RTU mouse monoclonal; clone: JC/70A; PDM020) (IMGENEX Cat# IMG-80250; RRID:AB_1152624).

Immunohistochemistry staining procedure

The tissue taken for IHC was performed on paraffin embedded, formalin fixed samples. A representative block was selected in every case. IHC staining was performed with CD34 (RTU mouse monoclonal; clone: QBEND/10; PDM050) and mast cell tryptase (RTU mouse monoclonal; clone: JC/70A; PDM020) according to the instructions provided by the manufacturer. The secondary antibody used was of Diagnostic Biosystem (Ref: UMR1000PD) against mouse primary antibodies. The staining for IHC was done according to standard procedures.

Preparation of Tris (tris hydroxymethyl aminomethane) EDTA buffer: Tris buffer, 1.21 g; EDTA disodium salt 0.372 g was dissolved in 1L of distilled water. This was used for antigen retrieval.

Preparation of stock solution of wash buffer/tris buffer: Tris buffer and sodium chloride (NaCl) solution were used for preparing the working solution. Tris buffer was prepared by mixing 121.1 g of buffer powder with 1000 ml of distilled water and adjusting the pH to 7.4 by adding hydrochloric acid (HCl) drop by drop. NaCl solution was prepared in a separate beaker by adding 83.33 g of NaCl powder to 1000 ml of distilled water. The working solution was prepared by mixing 30 ml of Tris buffer and 70 ml of NaCl solution in 700 ml of distilled water.

The IHC staining was carried out on 4 μm thick sections of tissue. The tissue sections were incubated on poly L-lysine coated glass slides overnight at 37°C. De-paraffinization was performed by keeping the sections at 60°C for 20 minutes followed by two changes of xylene for 15 minutes each. Slides were dried and alcohol changes for 2 times for 15 minutes each was carried out. The slides were immersed in a solution of 99% methanol and 1% hydrogen peroxide to quench the endogenous peroxidase, for 20 minutes. They were washed in running tap water for 5 minutes.

Antigen retrieval: The EDTA buffer was pre-heated for 5 minutes at high power in the microwave oven. After 5 minutes, washed slides were dipped in pre-heated solution for 5 minutes, 10 minutes, and another 5 minutes and allowed to cool. They were washed under running tap water for 5–10 minutes. Slides were dipped in distilled water, then working solution/wash buffer (pH 7.4) for 10 minutes. Next, the slides were incubated for 10 minutes in buffered casein solution with sodium azide to suppress the non-specific binding (power block/peroxidase blocking solution). The slides were incubated with the primary antibody CD34 (RTU mouse monoclonal; clone: QBEND/10; PDM050) and mast cell tryptase (RTU mouse monoclonal; clone: JC/70A; PDM020) for 30 minutes at room temperature. After incubation, they were washed in the working solution for duration of 10 minutes. Next, the slides were treated with Diagnostic Biosystem rabbit/mouse secondary antibody (catalogue number K8002, monoclonal antibody) and incubation was carried out at room temperature. Slides were washed in working solution for 10 minutes and then treated with a freshly prepared solution of 20 μL of 3,3’- Diaminobenzidine (DAB) chromogen and 1000 μl of substrate buffer for 5 minutes to obtain a brown colour. Next, the slides were washed with working solution for 5 minutes. Counterstaining was performed for 2–3 minutes with Meyer’s hematoxylin and slides were rinsed in tap water thereafter. Finally, the slides were dehydrated, cleared and mounted.

Controls used for IHC: Positive controls used for respective markers were as follows: CD34, Kidney; Mast cell tryptase, Tonsil

Assessment of MVD, mast cells, CD34 blasts and reticulin fibrosis

The average number of thin-walled vessels with or without lumen and sinusoids that were detectable by CD34+ endothelial cells were considered to assess MVD. These were then counted in ten hot spot fields and an average was obtained per high power field (HPF). Average MVD in 10 hotspots was scored as follows; 0–7 mvd/10 hotspots – score 1, 8–15 mvd/10 hotspots – score 2, 16–25 mvd/10 hotspots – score 3.^8,11

CD34 is expressed in the cytoplasm of blast cells. For determination of blast percentage on CD34 stained bone marrow biopsy, five representative fields were selected and the number CD34+ cells were counted in a total of five high power fields (40x).^12,13 The scoring was performed as follows; 1–5 blasts/5HPF – score 1, >5–10 blasts/5HPF – score 2, >10–19 blasts/5HPF – score 3. MCT shows cytoplasmic staining in mast cells. Average mast cell numbers were counted using tryptase in 10 oil immersion fields. The grading of mast cells using tryptase was as follows; 3–5 mast cells – 1+, >5–10 mast cells – 2+, >10 mast cells – 3+.¹⁴

Reticulin fibrosis was graded according to WHO guidelines, where loose network of intersecting reticulin fibres around the perivascular region was considered grade 1, diffuse and dense reticulin fibres with extensive intersections along with focal bundles of thick collagen fibres as grade 2 and diffuse and dense increase in reticulin fibres with extensive intersections and coarse bundles consistent with thick collagen fibres as grade 3 fibrosis.¹⁰

Statistical analysis

For statistical analysis, SPSS software version 25 (IBM SPSS Statistics (RRID:SCR_016479)) was used. Chi-square test, Fishers exact test, mean MVD and standard deviation were calculated. P-value was used to determine significance of the study. Statistical significance was defined as p-value ≤0.05.

Correlation of MVD with bone marrow morphology on biopsy

On morphological assessment of these 66 bone marrow biopsies, the majority of them (n = 59, 89.39%) were hypercellular for age. Increased myelopoiesis was observed in 54 cases (81.81%), out of which left shift was seen in 39 (59%) of them. Erythropoiesis was found to be decreased in 36 (54.54%) cases. Megakaryopoiesis was markedly increased in 62 cases (93.93%), in which 17 (26%) and 8 (12.12%) cases showed MVD scores 2 and 3 respectively. Cases with megakaryocytes in clusters, sheets and diffuse patterns were found to have increased mean MVD (8.85, 8.38 and 11.50 respectively), while the presence of dyspoietic megakaryocytes did not alter the mean MVD significantly. These observations, however, were not statistically significant (Table 1).

Expression of MVD in different subtypes of MPN

In CML, MVD score 2 and 3 were found only in six (18.75%) and five cases (15.62%) respectively. Out of 31 cases of PMF, 12 (38.7%) cases expressed MVD of score 2, while score 3 was seen in three (9.67%) cases only. In ET, two cases (n = 2/3, 66.66%) expressed MVD of score 3 (p = 0.042). Highest mean MVD was observed in ET (Mean MVD -12.67) compared to CML and PMF (Mean MVD -7.34 and 8.29 respectively) as depicted in Table 2.

Correlation of MVD with bone marrow fibrosis in MPN

Out of 66 cases, 9 (13.63%), 26 (39.39%) and 31 (46.96%) cases were showing grade 1, 2 and 3 fibrosis respectively. In cases with MVD score 2, 66.7% cases were found to have grade 2 fibrosis and in cases with score 3, 50% of them were showing grade 2 fibrosis (p = 0.332) (Table 3, Figure 1).

Figure 1. Case of PMF¹ on bone marrow biopsy; A. Showing fibrosis (H&E², 20x); B. Showing hypolobated and micro-megakaryocytes (black arrow) (H&E, 40x); C. Showing WHO³ grade 3 fibrosis (Reticulin stain, 40x); D. CD34 highlights microvessels (black arrow) showing MVD⁴ score 3 (IHC⁵, 40x); E. CD34 stains cytoplasm of blasts (black arrow) showing blast score of 1 (IHC, 40x); F. MCT⁶ highlights cytoplasm of mast cells (black arrow) showing 3+ positivity (IHC, 100x).

¹PMF – Primary Myelofibrosis, ²H&E – Haematoxylin and Eosin stain, ³WHO – World Health Organisation, ⁴MVD – Microvessel Density, ⁵IHC – Immunohistochemistry, ⁶MCT – Mast cell tryptase.

Correlation of MVD with CD34 blasts in MPN

When we correlated CD34 blast count score with mean MVD, there was a positive correlation between CD34 blast count and mean MVD (Table 4). Increased mean MVD was observed with score 2 and 3 of CD34 blasts, which was statistically significant (mean MVD = 8.37 and 10.67 respectively, p = 0.036) (Figure 2).

Figure 2. Case of CML¹ on bone marrow biopsy; A. Showing left shift (H&E², 20x); B. Showing ‘dwarf’ megakaryocytes (black arrow) (H&E, 40x); C. Showing WHO³ grade 2 fibrosis (Reticulin stain, 40x); D. CD34 highlights microvessels (black arrow) showing MVD⁴ score 2 (IHC, 40x); E. CD34 stains cytoplasm of blasts (black arrow) showing blast score of 3 (IHC⁵, 40x); F. MCT⁶ highlights cytoplasm of mast cells (black arrow) showing 2+ positivity (IHC, 100x).

¹CML – Chronic myeloid leukemia, ²H&E – Haematoxylin and Eosin stain, ³WHO – World Health Organisation, ⁴MVD – Microvessel Density, ⁵IHC – Immunohistochemistry, ⁶MCT – Mast cell tryptase.

Correlation of MVD with mast cells in MPN

On analysis, highest mean MVD was seen in cases with mast cell positivity of 3+ and there was a positive correlation between mast cell score and mean MVD. However, we did not get statistical significance with MVD score and mean MVD as seen in Table 5 (p = 0.447).

Correlation of MVD with mutations in MPN

BCR-ABL1 mutation was seen in 32 (48.48%) cases, out of which six (33.3%) and five (50%) cases showed MVD scores 2 and 3 respectively. JAK2 mutation was seen in a total of 32 (48.48%) cases, in which MVD scores of 2 and 3 were found in 11 and five cases respectively. Only five cases showed CALR mutation, out of which only one case showed score 2 MVD as depicted in Table 6. Mean MVD of 8.72 and 6.60 was seen in JAK2 and CALR positive cases respectively (p = 0.053 and 0.059 respectively).

Correlation of MVD with prognosis in MPN

In the present study, all 66 patients were followed up for one year with regular checkups at three-month intervals. Among which, 37 (68.18%) and 17 (15.15%) subjects were rendered disease-free with MVD score of 1 and 2 respectively. Disease relapse was seen in eight cases (12.12%), with the majority (n = 6/8, 75%) expressing MVD score of 3 (p = 0.000) as depicted in Table 7. Thus, mean MVD was higher in those cases with relapse/deceased as compared to disease-free patients.

Limitations

Out of 66 MPN cases, only three cases of ET and no case of polycythemia vera were evaluated. Hence, further studies with larger cohorts comprising a greater number of each subtype of MPN would be recommended to establish the prognostic/theranostic role of MVD in MPNs.