Modulation of bone marrow microenvironment following ruxolitinib therapy in myelofibrosis

Myelofibrosis (MF) is characterized by progressive bone marrow fibrosis, ineffective hematopoiesis and extramedullary hematopoiesis with splenomegaly. The primary symptoms resulting from peripheral cytopenias and massive spleen enlargement are often associated to constitutional symptoms such as fever, weight loss and night sweats.

The constitutive activation of the Janus kinase (JAK)-STAT signaling pathway is a hallmark of MF. The JAK2 V617F mutation is one of the factors responsible for up-regulation of this pathway and is present in ∼ 50% of MF patients [1]. Calreticulin (CALR) mutations are also involved and so far have been ascertained in ∼ 25% of patients with JAK2 and MPL unmutated primary MF [2].

Despite the availability of a wide range of conventional treatment strategies, none has so far been able to change the natural course of the disease.

The JAK1/2 inhibitor ruxolitinib has recently been approved for the treatment of intermediate-2 and high risk MF, either primary or post-polycythemia vera (PV) or essential thrombocytemia (ET). Data emerging from two large randomized studies known as COMFORT (Controlled myelofibrosis study with oral JAK inhibitor treatment) I and II provide evidence of significant improvement in splenomegaly and symptoms [3,4]. Further study suggests that ruxolitinib therapy may offer an advantage in survival over traditional therapy [5]. Interestingly, some single case reports and abstracts indicate that the drug is capable of achieving an improvement in bone marrow (BM) fibrosis [6–10]. Newly emerging data suggest that ruxolitinib therapy may lead to a modulation of BM microenvironment. Kvasnicka et al. [11–13] observed, alongside a significant decrease in fibrosis, macrophages and mast cells, a significant reduction of microvascular density (MVD) and microvessel area (MVA) as well as several changes in the shape and tortuosity of BM vessel structure.

These results encouraged us to investigate the BM microenvironment in 10 MF patients treated with ruxolitinib. Eight patients with primary MF were enrolled in the open-label, multi-center, expanded access study JUMP CINC424A2401 and another two patients with post-ET MF were enrolled within the compassionate use program INCB018424. Both studies were approved by the local institutional ethical committee and conducted in accordance with the principles of the Declaration of Helsinki. The primary objective of these studies was to collect additional data on the safety of ruxolitinib treatment, irrespective of JAK2 mutation status, in patients with intermediate-high risk MF who had either been previously treated with commercially available agents or had never been treated. All patients were given 15 or 20 mg of oral ruxolitinib twice daily (BID) depending on baseline platelet count (100 000/μL to 200 000/μL or >200 000/μL, respectively). The protocol permitted dose escalation to 25 mg BID in patients with an inadequate response; dose reduction was mandatory when platelet counts dropped to < 100 000/μL and the drug was stopped altogether with levels below 50 000/μL. Specimens were collected and screened for the JAK2 V617F, CALR and MPL mutations. All patients underwent BM trephine biopsy at baseline and were re-evaluated after a median of 31 months (range = 12–40) of starting treatment with ruxolitinib. Immunohistochemical staining was performed with hematoxylin-eosin, factor VIII, glycoforin A (JC 159), reticulin (Dako stainer) and anti-CD34 monoclonal antibody (QBEnd 10). Stained biopsy specimens were assessed by three independent pathologists for cellularity (%); megakaryocyte proliferation [3-point scoring system based on the number of megakaryocytes: <10 (+); between 10–20 (++) and >20 (+++) at 200× magnification]; clustering and atypia [3-point scoring system ranging from normal (–), through rare dense clustering and moderate atypia (+) to dense clustering of megakaryocytes with aberrant nuclear/cytoplasmic ratio and hyperchromatic and irregularly folded nuclei (++)] [14]; sinusoidal density [3-point scoring system based on the number of sinusoids (<20 (+); between 20–30 (++) and >30 (+++) at 200× magnification) [15]] and dilatation [2-point scoring system based on absence (–) or presence (+)]; marrow fibrosis according to WHO criteria [16]. The average agreement for the six assessed variables among the three pathologists (AM, MP, RM) was 80%, ranging from 70% for sinusoidal density, through 80% for fibrosis, cellularity, megakaryocyte proliferation, clustering and atypia, to 90% for dilatation. In 15% of the evaluations, at least two of the three pathologists agreed. In the remaining cases, where interpretation was complicated by uneven sinusoidal density and fibrosis, the consensus was reached following collegial discussion. Overall, the strength of agreement was substantial (unweighted kappa coefficient equal or greater than 0.61) [17].

After starting treatment, patients were initially evaluated once a week and then on a monthly basis. Complete clinical check-up included body weight measurements, splenic size by palpation and blood analysis.

The characteristics of our patients are summarized in Table I. Median age at diagnosis was 62 years (range = 44–80). The JAK2 V617F mutation was detected in six patients, while CALR and MPL were found in two and one patient, respectively. One patient was triple negative. At diagnosis, median white blood cell (WBC) and platelet (PLT) counts were 20.1 × 10³/μL (range = 3–37) and 357 × 10³/μL (range = 134–480), respectively; median hemoglobin concentration was 11.6 gr/dl (range = 7.3–15.9). Splenomegaly was detected in all patients with a median enlargement of 17.5 cm below the costal margin (range = 6–29). Half of the patients complained of non-specific constitutional symptoms. Based on IPSS scores, 30% of the patients were assigned to the intermediate-1 risk category, 50% to the intermediate-2 risk category and 20% to the high risk category. Six patients had previously been treated with hydroxyurea, one patient with pipobroman and danazol, three patients had not received any specific treatment. No relationship was found between pre-treatment therapy with hydroxyurea and baseline bone marrow features before starting ruxolitinib.

Table I. Characteristics of 10 patients with MF treated with ruxolitinib.

Pts	Treatment before rxt	Months of rxt therapy	Mutation	Spleen Reduction (%)	Weight gain (Kg)	Marrow cellularity (%)		Marrow fibrosis (WHO grade)		Megakaryocyte atypia		Megakaryocyte proliferation		Sinusoidal dilatation		Sinusoidal density
						BL	Post rxt	BL	Post rxt	BL	Post rxt	BL	Post rxt	BL	Post rxt	BL	Post rxt
						1	hydroxyurea	36	JAK2^V617F	20	7	100	40	3	3	++	–	+++	+	–	+	+	++
2	hydroxyurea	36	JAK2^V617F	30	8	80	50	2	2	++	+	+++	+	+	+	++	++
3	hydroxyurea	35	JAK2^V617F	65	11	70	40	2	2	+	–	++	+	–	+	+	++
4	—	27	MPL	100	5	50	70	3	2	+	+	++	+	–	+	+	+
5	—	24	JAK2^V617F	NA	12	90	70	2	2	–	+	+++	+	–	+	+	++
6	—	14	JAK2^V617F	100	14	90	70	3	2	+	–	+++	+	–	+	+++	+++
7	hydroxyurea	12	CALR	35	4	10	10	3	3	+	+	+	+	–	+	+	++
8	danazol  pipobroman	12	Triple neg	100	11	90	70	2	1	+	–	++	++	–	+	+	+++
9	hydroxyurea	40	CALR	20	5	10	10	3	3	–	–	+	+	+	+	+	++
10	hydroxyurea	40	JAK2^V617F	100	3	90	70	2	2	++	–	+++	+	–	+	+	++

Marrow trephines were evaluated for cellularity (%); megakaryocyte proliferation [3-point scoring system based on the number of megakaryocytes: < 10 (+); between 10–20 (++) and >20 (+++) at 200× magnification]; clustering and atypia [3-point scoring system ranging from normal (–), through rare dense clustering and moderate atypia (+) to dense clustering of megakaryocytes with aberrant nuclear/cytoplasmic ratio and hyperchromatic and irregularly folded nuclei (++)] [14]; sinusoidal density [3-point scoring system based on the number of sinusoids (< 20 (+); between 20–30 (++) and >30 (+++) at 200× magnification) [15]] and luminal dilatation [2-point scoring system based on absence (–) or presence (+)]; marrow fibrosis according to WHO criteria [16].

BL, baseline; Post rxt, post-ruxolitinib treatment.

The median age at starting ruxolitinib treatment was 66 years (range = 51–82) and the median administered dose was 15 mg BID (range = 5–25 mg). After a median period of 31 months of therapy, median white blood cell (WBC) and platelet (PLT) counts dropped to 11.4 × 10³/μL (range = 3–28) and 231 × 10³/μL (range = 90–550), respectively; median hemoglobin concentration decreased to 9.8 gr/dl (range = 8.3–12.9); two patients remained transfusion-dependent and one patient required transfusions after starting treatment. All patients showed improvement in constitutional symptoms and quality-of-life; the median weight gain was 7.5 kg (range = 4–14 kg). Splenomegaly decreased by 65% (range = 20–100%). Except for anemia and a single case of reactivation of hepatitis B virus infection [18], no significant side-effects were reported during treatment. Morphological evaluation of trephine biopsies showed significant changes in BM microenvironment compared to baseline. Overall, there was a 20% decrease in median cellularity (range = 20–60%). Interestingly, the two CALR positive patients had very low baseline values for cellularity compared to patients with other mutations.

Fibrosis was improved in 30% of the cases and remained stable in 70%. Reduced proliferation of megakaryocytes and improvement of dysplasia were observed in 70% of the cases. Notably, the majority of patients displayed significant changes in bone marrow angiogenesis and micro-vessel architecture. A particularly interesting and novel finding was the increment in sinusoidal density compared to baseline observed in 70% of our patients. Additionally, 80% of the trephine biopsies revealed a marked increase in the dilatation of sinusoids ([Fig. 1]. Kvasnicka et al. [13] instead observed a reduction of MVD and MVA, independent of BM fibrosis response. These authors also found a correlation between normalization of vessel structure and significant reduction of vascular endothelial growth factor serum levels.

Figure 1. Increased sinusoidal dilatation after ruxolitinib treatment. Anti CD-34 immunohistochemical staining, 200× magnification. Patients 5 and 7 at baseline (a, c) and after 24 and 12 months of treatment, respectively (b, d). In these patients, fibrosis and cellularity remained stable [Table I].

Although we did not perform a study on cytokine expression, our data provide additional support that ruxolitinib therapy may possibly be capable of re-modeling marrow microenvironment and vascular niches. The increase found for sinusoidal density and dilatation is reason for debate. Marrow blood vessel structure is characterized by medullary arteries that enter the trabeculae, divide into arterioles and then into capillaries, which lead into the sinusoids, forming a system of channels of variable width and length with walls consisting of a single layer of endothelial cells. Physiologically, portions of the sinusoidal channels may be collapsed; expansion or contraction of the microvessels contributes to the extreme fluctuation in peripheral blood cell levels. It can be postulated that the reduction in marrow cellularity and improvement in fibrosis creates additional space for the sinusoids, thereby increasing their number and enlarging their shape and size. Unfortunately, a direct effect of anti-JAK2 drugs on CD34 positive hematopoietic progenitor cells is still unknown and needs to be clarified in future studies.

Given the paucity of reports [7–10] and despite the relatively small cohort of patients evaluated in the present study, it would seem possible to confirm that ruxolitinib is able to modify vascular niches of the BM micro-environment and reduce or stabilize fibrosis. Ruxolitinib should, therefore, be considered for use in early stage MF with the aim to challenge the BM alterations observed in the course of the disease [6].

Acknowledgments

We are deeply grateful to all the patients who participated in this study. We also wish to thank Anna Maria Koopmans for professional writing assistance.

Potential conflict of interest:

Disclosure forms provided by the authors are available with the full text of this article at www.informahealthcare.com/lal