CCN2/CTGF expression does not correlate with fibrosis in myeloproliferative neoplasms, consistent with noncanonical TGF-β signaling driving myelofibrosis

Previous studies, mainly performed in myelofibrosis models, indicate that the mutant/malignant megakaryocytes contribute to the development of fibrosis by increased expression of fibrotic and pro-inflammatory cytokines and interleukins, growth factors (including transforming growth factor-β (TGF-β)), extracellular matrix components, and other factors [1]. Still, many facets of the development of BM fibrosis and the sequential events that drive stromal activation and fibrosis remain elusive.

Due to its critical involvement in many fibrotic processes, a role of Cellular Communication Network 2 (CCN2) in the pathogenesis of BM fibrosis seems likely. CCN2, also known as CTGF (Connective Tissue Growth Factor), is an extracellular matrix cellular protein belonging to the Cellular Communication Network (CCN)-family. It is considered an important driver and biomarker of organ fibrosis in a wide range of diseases [2]. CCN2 is involved in the proliferation, migration, and differentiation of cells and can promote fibrosis directly or by acting as a factor downstream of TGF-β, which is a powerful and well-known inducer of CCN2 transcription [3‐8], but many other factors also directly or indirectly induce CCN2 mRNA expression. The mode of action of CCN2 is complex and variable as illustrated in Fig. 1. The way(s) by which CCN2 act(s) on BM target cells is, however, still largely unknown.

In previous studies, CCN2 mRNA expression has been detected in BM mesenchymal stem and stromal cells of normal BM [9, 10]. In addition, altered CCN2 mRNA expression levels have been associated with BM malignancies: in B-acute lymphoblastic leukemia, increased CCN2 mRNA levels are present in B-lymphoblasts, whereas in acute myeloid leukemia, CCN2 mRNA overexpression has been detected in the mesenchymal stem/stromal cells [11, 12]. Furthermore, CCN2 mRNA extracted from BM biopsies of patients with myelofibrosis showed a 27-fold increase when compared to healthy controls, decreasing after allogeneic stem cell transplantation [13]. None of these studies, however, investigated a possible relationship between CCN2 levels and BM fibrosis.

Thus far, CCN2 protein expression in the BM has only been investigated in three previous studies [13‐15]. Chica et al. found positive staining of several cell populations of normal bone marrow but not in megakaryocytes, except for weak staining near the cell membranes [15], while Åström et al. [14] reported cytoplasmic positivity for CCN2 in a subpopulation (18%) of megakaryocytes in 1 of 5 patients with X-linked thalassemia in almost all (97%) megakaryocytes of all 6 primary myelofibrosis patients, while other hematopoietic cell lineages and the megakaryocytes in normal control BM biopsies were negative [14]. Shergill et al. published an abstract with intriguing observations suggesting a role for CCN2 as a biomarker and a potential target for therapy in MF [13]. However, the abstract (naturally) contained only limited information and statistical detail, and the data presented suggested considerable scatter of data and limited statistical significance. CCN2 staining was found in a higher percentage of megakaryocytes in biopsies of myelofibrosis patients compared to controls (63% vs 40%). This difference did not reach statistical significance (p = 0.28), but the mean percentage of CCN2-positive megakaryocytes was significantly higher in myelofibrosis patients at diagnosis (63%) compared to post-transplant biopsies 22% [13]. A correlation between CCN2 expression and fibrosis was, however, not made. Unfortunately, no follow-up publication on this 2015 abstract has appeared to date, and no other studies have since then addressed the role of CCN2 in BM fibrosis. Therefore, we set out to investigate CCN2 protein expression by immunohistochemical staining in a large cohort of 75 BM biopsies (55 MPN patients and 20 normal controls) and correlated the results with the amount of BM fibrosis and other disease parameters.

In this study, we investigated the protein expression of the profibrotic factor CCN2 in BM by immunohistochemistry. We included 20 normal BM biopsies and 55 BM biopsies of MPN patients and correlated the CCN2 staining results with the amount of BM fibrosis and clinical parameters. CCN2 protein expression was detected in a variable degree in the cytoplasm of megakaryocytes, while granulopoiesis and erythropoiesis were negative. High levels of CCN2 expression were seen in 7 (13%) MPN cases, but no correlation was observed between CCN2 expression and fibrosis, neither in the total study group nor in the subgroup with CCN2 overexpression. Thus (megakaryocytic) CCN2 expression appears not to be a key feature in the development of fibrosis in MPN, in contrast to the prominent role that CCN2 plays in the development of fibrosis in many organ diseases outside the marrow [2]. A lack of correlation between CCN2 expression and BM fibrosis is remarkable, especially because the prototypical fibrosis inducer TGF-β, known to also contribute to myelofibrosis in MPN [19], is a potent inducer of CCN2, the two forming a positive feedback loop [4‐8].

However, whereas induction of CCN2 gene transcription by TGF-β typically involves the canonical SMAD2/3 pathway [6], a recent study showed that noncanonical c-Jun N-terminal kinase (JNK)-dependent TGF-β signaling in mesenchymal stromal cells is responsible for the development of BM fibrosis in MPN [19]. Our findings may thus be interpreted as further evidence that the non-canonical, rather than the canonical pathway of TGF-β signaling drives BM fibrosis in MPN. This further supports the notion that the exploration of novel modalities for treatment and prevention of BM fibrosis in MPN patients might best focus on the noncanonical pathway of TGF-β signaling.

Interestingly, a subgroup of 7 (13%) MPN cases showed significant CCN2 overexpression. Four of these showed no fibrosis (MF-0), 2 showed moderate fibrosis (MF-2), and 1 severe fibrosis (MF-3), not related to the amount of CCN2 expression. Blood values and clinical signs did not reveal a specific trend. Although CCN2 overexpression was specifically observed in megakaryocytes, no correlation was observed between the level of CCN2 expression and the platelet counts or occurrence of thrombovascular events in this small subgroup, but additional studies on larger cohorts of MPN patients will be needed to better investigate the possible meaning of the megakaryocytic CCN2 overexpression in the development or complications of MPN.

We detected CCN2 protein to be mainly present in megakaryocytes, confirming the findings of previous studies [13, 14]. Chica et al. reported no staining of (normal) megakaryocytes [15], but in the study by Åstrom et al., all their 6 PMF cases showed cytoplasmic CCN2 staining, staining on average 97% of megakaryocytes, while their 6 normal controls were negative [14]. Also, Shergill et al. described variable cytoplasmic staining of megakaryocytes, with a higher percentage of megakaryocytes showing positive CCN2 staining in myelofibrosis patients compared to healthy controls [13]. Discrepancies between the studies might be explained by the use of different types of antibodies. To confirm our results and to rule out non-specific binding of the antibody, we used 2 other CCN2 antibodies binding to different CCN2 epitopes, and the results were in line with our findings with the 10095S antibody.

Shergill et al. found mean CCN2 mRNA levels extracted from BM biopsies in myelofibrosis patients to be 27-fold increase compared to controls, but this difference was not significant. This might suggest that the elevated mean was caused by a small subgroup with very high expression, consistent with our observation of only a subgroup of MPN cases showing marked increased CCN2 protein expression. We have also tried to verify our results at the mRNA level by in situ hybridization on BM sections but did not obtain an interpretable result, probably due to poor mRNA quality after fixation and decalcification. As whole tissue sections were used in the study by Shergill et al., and not isolated megakaryocytes, definite proof that the source of the CCN2 mRNA in BM is the megakaryocytes is still lacking.

Megakaryocytes contain and secrete many factors, by which they contribute to the functioning of the BM microenvironment, including the hematopoietic stem cell niche [1, 20, 21]. They also participate in inflammation and immunity [22]. Also, their role in MPN is well established, in which mutant megakaryocytes play a key role by promoting myeloproliferation and fibrosis [20]. The heterogeneous staining of CCN2 of the megakaryocytes within individual biopsies might reflect different megakaryocyte subtypes, as was previously described [23, 24], or differences between subclones. Morphologically, however, a specific subtype of megakaryocytes showing overexpression could not be detected, and also the degree of CCN2 overexpression did not correlate with the JAK2 mutant allele frequency. There also was no difference in CCN2 expression between JAK2 positive and JAK2 negative MPNs, as opposed to the trend reported in a previous small pilot study [13].

Outside the BM, CCN2 and its fragments have been implicated not only in fibrosis but also in the regulation of cell proliferation, differentiation, adhesion, migration, cell survival, apoptosis, and senescence [25‐27]. At least part of CCN2’s biological activity is mediated through interaction with a host of other proteins and receptors, by which it can modify their signaling activity and cross-talk [28‐30]. Therefore, it will be challenging to fully explore the potential roles of CCN2 in megakaryocytes of normal BM (although the expression is low), of the subgroup of MPN with overexpression, and beyond.

In summary, in BM biopsies, we observed variable CCN2 expression in megakaryocytes, a cell type increasingly recognized for its importance in the regulation of the BM microenvironment, including the established role of mutant megakaryocytes in MPN promoting myeloproliferation and fibrosis. In MPN, remarkable CCN2 overexpression was detected in a subgroup (13%). CCN2 expression, however, did not correlate with fibrosis or other disease parameters, neither in the whole study group nor in the subgroup with CCN2 overexpression in megakaryocytes.

Our data suggest that CCN2 is not a key driver of myelofibrosis in MPN, and that noncanonical, CCN2 independent, rather than canonical TGF-β signaling might be responsible for the development of fibrosis in MPN.