Myeloproliferative neoplasms (MPNs) are clonal neoplastic disorders of hematopoiesis, leading to a hypercellular bone marrow and eventually to severe fibrosis, termed myelofibrosis (MF).1 Particularly advanced stage MF patients suffer from splenomegaly, anemia and constitutional symptoms due to elevated levels of proinflammatory cytokines.A gain-of-function V617F mutation in the JAK2 kinase causes the constitutive activation of the JAK/STAT pathway and is supposed to be at least in part responsible for the disease phenotype of MPN.As the mutation is present in 65% of primary MF patients, 95% of polycythemia vera (PV) patients and 55% of essential thrombocythemia (ET) patients, various JAK inhibitors are currently tested as targeted therapeutics in MPN.2.Ruxolitinib is an oral JAK inhibitor recently approved for treatment of primary and secondary MF that shows impressive symptom control by suppression of inflammation irresp. of the JAK mutational status.Moreover, ruxolitinib is also a promising drug for acute and chronic GvHD treatment.3, 4 We recently described various aspects of immune modulation by ruxolitinib5, 6 including impaired DC function.7.DCs are important antigen-presenting cells maturing and migrating via afferent lymphatic vessels from peripheral organs into T-cell areas of draining lymph nodes (LNs) upon antigen contact.8.This process is guided by chemokine gradients and requires cytoskeletal rearrangement promoting cell movement.CCR7 is expressed on DCs upon antigen contact and senses chemokine gradients of CCL19 and CCL21.CCL19 is essential for guiding DCs into draining LNs.9, 10.Notably, recent work suggested that activation of JAKs is involved in lymphocyte migration.11.The aim of this study was a detailed anal. of the impact of ruxolitinib on DC migration with a particular focus on short-term effects of the drug and the identification of potential target mols. mediating these effects.We recently already described that in vivo DC migration under conditions of inflammation is impaired after long-term exposure with ruxolitinib.7 We now extended our previous findings by evaluating the effects of ruxolitinib on DC migration.Using a set-up of an ex vivo crawl-in assay within explanted ear dermis, we could visualize DCs migrating towards and entering afferent lymphatic vessels.Vehicle-treated bmDCs (green) appropriately migrate to and localize around lymphatics (stained for laminin in red), while ruxolitinib-exposed bmDCs cannot be detected around lymphatic vessels (Figure 1a).We then analyzed the effect of ruxolitinib on chemokine-induced migration of LPS-matured human moDCs in vitro using three-dimensional (3D) collagen matrixes.After collagen polymerization, cells are embedded in a complex network of fibrils, mimicking the extracellular matrix.Again, the JAK inhibitor ruxolitinib was added directly to the assay, limiting the maximum exposure of the cells to 4 h.Thus, the observed effects are unlikely due to dedifferentiation or dematuration of DCs, which are usually seen as early as 48 h after drug exposure.7.Ruxolitinib dose-dependently impairs directional migration of moDCs towards a CCL19 gradient, as mirrored by reduced velocity and accumulated distance (Figures 1b and c).Toxicity of the drug was excluded by annexin V/7-AAD staining (data not shown).Importantly, ruxolitinib-exposed moDCs are still able to sense the chemokine gradient as they form lamellipodia at the leading edge of the cell, whereas retraction of the uropod is clearly inhibited, resulting in an elongated cell body of the cells (Supplementary Videos 1 and 2).According to proper chemokine sensing, ruxolitinib does not alter the expression of the CCL19-sensing chemokine receptor CCR7 (data not shown).Anal. of the cell cytoskeleton by fluorescent labeling of the actin and tubulin network also revealed no overt structural alterations after ruxolitinib exposure (Figure 1d).To investigate if the observed defects in DC migration after ruxolitinib treatment are based on an imbalance of adhesion, migration of talin1-deficient moDCs in 3D collagen matrixes, static adhesion analyses on different two-dimensional (2D) substrates and flow cytometry-based characterization of integrin subsets was performed.As talin is of critical importance for integrin activation and the coupling of integrins to the actin cytoskeleton,12 talin1-deficient moDCs were evaluated for their migratory capacity within complex collagen networks.In accordance with the previous reports, the absence of talin had no impact on the migratory ability, as DCs are able to switch between integrin-dependent and -independent migration modes (Figure 1e).Knockdown efficiency of talin1 was proven by western blot anal. (Figure 1f).Of note, the migratory defects seen in ruxolitinib-treated control cells could be induced irresp. of the presence or absence of talin (Figure 1e).In line with these observations, we could not detect any changes of adhesion of mature moDCs to fibronectin or collagen (Figure 1g) as a potential reason for decreased migration due to increased 'stickiness' of the moDCs.As expected, in contrast to the migration within 3D environments, adhesion of talin1-deficient moDCs to both substrates was markedly impaired.Ruxolitinib treatment did not further reduce adhesion in the 2D setting.Similarly, no alteration of integrin expression levels could be detected in ruxolitinib-exposed DC (Figure 1h).Due to the highly conserved structure of the catalytic sites of protein kinases, ATP-competitive inhibitors like ruxolitinib often show side effects due to unspecific binding to different target kinases.It is known that ruxolitinib impairs the function of other kinases in addition to JAKs.13 Therefore, we first tested whether siRNAs targeting either JAK1, JAK2 or JAK1 and 2 affect moDC migration towards CCL19 in the 3D migration assay.Interestingly, CCL19-induced chemotaxis is not significantly altered in the absence of JAK1 and/or JAK2 expression (Figure 2a), highlighting that the inhibitory effects of ruxolitinib are independent of JAK1 and/or JAK2, the main targets of the drug.Knockdown efficiency was verified by western blot anal. of the resp. JAKs (Figure 2b).This data is in contrast to recent publications reporting a functional involvement of JAKs in lymphocyte migration,11 as JAK1/JAK2-deficient lymphocytes do not properly migrate in vitro and their homing to LNs in vivo is also reduced.However, in these reports intranodal localization and motility of lymphocytes were not affected by JAK deficiency.The study performed by Stein et al. also applied relatively high concentrations of AG490, a JAK inhibitor currently not used/tested in clin. praxis.We confirmed our pharmacol. data by using momelotinib, another clin. tested JAK inhibitor, mimicking the reduction of the migratory capacity induced by ruxolitinib (data not shown) and the discrepancy may also at least in part be explained by the different cell types (lymphocytes vs. DCs) examinedOn the basis of literature findings describing 33 kinases inhibited more than 50% in the presence of 1 μM of ruxolitinib, we next aimed to identify the target of the drug mediating the migration defects.Interestingly, the screen revealed two kinases, Rho-associated coiled-coil kinase 1 and 2 (ROCK1 and ROCK2), with a functional impairment of 71 and 65% after ruxolitinib treatment.13 The serine-threonine kinase ROCK controls non-muscle myosin II activity, thereby regulating reorganization and contraction of cellular actin-myosin filaments14 and is of critical importance in DC migration as it mediates the deadhesion of cells from ICAM-1 and propels the rigid nucleus through narrow gaps of the extracellular matrix.12, 15.To evaluate the data from the kinase screen in our cellular setup, we analyzed various target proteins known to be regulated by ROCK.Therapeutic doses of ruxolitinib clearly inhibit ROCK activity as phosphorylation of myosin phosphatase targeting subunit 1 (pMYPT1), a downstream target of ROCK (Figure 2c), was clearly reduced.As a consequence, the phosphorylation of moesin in LPS-matured moDCs was also affected (Figure 2d) and more of the phosphorylated inactive form of cofilin could be detected (Figure 2e).To compare the effect of ROCK inhibition on migration behavior observed in ruxolitinib-treated moDCs, Y-27632 an ATP-competitive ROCK inhibitor was added to the migration assays.Similar to ruxolitinib, Y-27632 potently inhibited moDC migration (Figure 2f) and induced a morphol. phenotype showing conserved chemokine sensing but impaired uropod retraction (Supplementary Video 3).In contrast to ruxolitinib,7 sole inhibition of ROCK by Y-27632 did not induce any phenotypic changes of bona fide DC markers or the in vitro differentiation from monocytes (data not shown).Hence, while inhibition of DC migration is independent of JAK, the effects on DC differentiation and maturation cannot be explained by the here-described off-target inhibition of ROCK.Consequently, the immunosuppressive effects induced by ruxolitinib are most likely a consequence of a combined targeting of several kinases involved in the regulation of immune cell functions.In summary, we here provide the mol. basis for a deeper understanding of ruxolitinib-mediated inhibition of DC migration, as we identified ROCK as ruxolitinib target.Thus via interference with ROCK activation, ruxolitinib profoundly impairs DC migration.Subsequently, the loss of trafficking DCs may lead to reduced activation of T cells in draining LNs and might therefore explain the abundance of these proinflammatory cells from the blood of MPN patients after ruxolitinib treatment.5.Our data may be of importance to better understand the immune-modulatory effects of ruxolitinib in vivo and add further pieces to the puzzle how pleiotropic ruxolitinib-mediated immune-suppressive effects in mice and men may be.The data also complement recent reports on the immune-suppressive role of this compound class for the treatment of immune-mediated diseases, such as GvHD.3, 4.
