Discovery of a signaling feedback circuit that defines interferon responses in myeloproliferative neoplasms- Published: 01 April 2022

[PhysAssist]

This gets a little deep, but nonetheless may explain some things:

Abstract

Interferons (IFNs) are key initiators and effectors of the immune response against malignant cells and also directly inhibit tumor growth. IFNα is highly effective in the treatment of myeloproliferative neoplasms (MPNs), but the mechanisms of action are unclear and it remains unknown why some patients respond to IFNα and others do not. Here, we identify and characterize a pathway involving PKCδ-dependent phosphorylation of ULK1 on serine residues 341 and 495, required for subsequent activation of p38 MAPK. We show that this pathway is essential for IFN-suppressive effects on primary malignant erythroid precursors from MPN patients, and that increased levels of ULK1 and p38 MAPK correlate with clinical response to IFNα therapy in these patients. We also demonstrate that IFNα treatment induces cleavage/activation of the ULK1-interacting ROCK1/2 proteins in vitro and in vivo, triggering a negative feedback loop that suppresses IFN responses. Overexpression of ROCK1/2 is seen in MPN patients and their genetic or pharmacological inhibition enhances IFN-anti-neoplastic responses in malignant erythroid precursors from MPN patients. These findings suggest the clinical potential of pharmacological inhibition of ROCK1/2 in combination with IFN-therapy for the treatment of MPNs.

Introduction

Type I interferons (IFNs) are critical frontline effectors of the immune defense against pathogens and cancer1,2. Additionally, IFNs act directly on tumor cells blocking their growth, survival, migration, and other pro-tumorigenic events3. Type I IFNs exert their effects by binding to the Type I IFN receptor (IFNAR) and subsequently activating multiple signaling pathways that lead to the expression of IFN-stimulated genes (ISGs), whose protein products drive several distinct biological functions4,5,6. Extensive clinical studies using different formulations of IFN-alpha (IFNα) have demonstrated anti-tumor activity against several malignancies. IFNα treatment is especially effective in Philadelphia chromosome negative myeloproliferative neoplasms (MPNs), and has been approved by the U.S. Food and Drug Administration for the treatment of hairy cell leukemia, AIDS-related Kaposi’s sarcoma, and chronic myelogenous leukemia (CML)3,7.

MPNs are clonal hematopoietic stem cell (HSC) disorders characterized by an increased risk of bleeding, thrombosis, bone marrow fibrosis and, in some cases, transformation to acute myeloid leukemia (AML)8,9. There are three types of MPNs: polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis (PMF). In 95% of MPN patients, somatic mutations in one of three genes: JAK2, CALR, or MPL occur in a single HSC, giving rise to malignant stem cells with constitutively active MPL-JAK-STAT signaling8. JAK2V617F is the most common mutation and occurs in ~95% of PV patients and 50–60% of ET and PMF patients10,11,12,13. JAK2 inhibitors are approved for the treatment of MPNs, however, overall they have a modest effect and rarely achieve complete molecular or pathologic remission14. In contrast, in recent clinical studies, pegylated IFNα (PEG-IFN) was shown to induce durable hematological and molecular responses in MPN patients, and in some cases complete remission7,15,16,17,18,19. The mechanism(s) of action of IFN in MPNs are unclear and the reason for molecular relapses, which have been reported in some patients20, are unknown.

We have previously demonstrated that Unc-51-like kinase 1 (ULK1) is required for IFN-induced anti-clonogenic effects against malignant erythroid progenitor cells from MPN patients and that ULK1 is required for IFN-mediated activation of p38 MAPK and expression of specific ISGs21. In the present study, we identify protein kinase C-delta (PKCδ) as an upstream regulator of the ULK1-p38 MAPK cascade. We also identify ROCK1/2 as interactor proteins of ULK1 and negative feedback regulators of IFNα-induced anti-neoplastic effects in primary MPN cells. Additionally, we provide evidence that increased expression of elements of the pathway correlate with clinical response to IFNα therapy in MPN patients, underscoring the importance of this signaling circuit in the IFN-system.

Results

Type I IFN treatment activates a PKCδ-ULK1-p38 MAPK signaling cascade

At the outset, we sought out the IFN-regulated signals controlling engagement of ULK1. There are putative PKCδ phosphorylation sites in ULK122 and, identified from our previous work, PKCδ activity is essential for type I IFN-induced phosphorylation of p38 MAPK and the subsequent transcription of ISGs23,24. This prompted us to investigate if PKCδ and ULK1 interact during activation of IFNAR in JAK2V617F-positive cells. In co-immunoprecipitation assays using two JAK2V617F-expressing leukemia cell lines, we found ULK1 associated with PKCδ, both before and after IFNα-treatment (Fig. 1a, b). This constitutive interaction between PKCδ and ULK1, however, was not JAK2V617F-dependent, as it was also detected in JAK2V617F-negative leukemia lines (Fig. 1c, d). These results raised the possibility that IFN could stimulate PKCδ-induced phosphorylation of ULK1. Thus, to determine whether ULK1 is a substrate for PKCδ associated with type I IFN signaling, in vitro kinase assays were conducted on anti-PKCδ immunoprecipitates, using inactive GST-ULK1 as an exogenous substrate. Our results revealed that PKCδ phosphorylates ULK1 in a type I IFN-dependent manner (Fig. 1e).

Link to Figure 1e: nature.com/articles/s41467-...

As previous studies using liquid chromatography (LC) and tandem mass spectrometry (MS/MS) have identified Ser341, Ser467 and Ser495 as the putative PKCδ phosphorylation sites in ULK122, we next generated single serine-to-alanine (S/A) ULK1 mutants (S341A, S467A and S495A) and a triple (SSS/AAA) ULK1 mutant (S341/467/495 A). ULK1 WT and Ser-to-Ala ULK1 mutants were overexpressed in HEK293T cells, immunoprecipitated from cell lysates, inactivated, and used as substrates for purified active GST-PKCδ in the presence of 32P-ATP in kinase assays (Fig. 1f). While replacement of Ser467 by Ala did not result in an obvious alteration of phosphorylation, mutation of Ser341 to Ala (S341A) and Ser495 to Ala (S495A) reduced phosphorylation of ULK1 by PKCδ (Fig. 1f). Additionally, this reduction was more evident for the triple mutant (SSS/AAA) ULK1 (Fig. 1f). These results suggest that PKCδ directly phosphorylates ULK1 on both Ser341 and Ser495 residues (Fig. 1f), which prompted us to generate two phospho-specific ULK1 antibodies, one recognizing ULK1 phosphorylated on Ser341 and the other recognizing ULK1 phosphorylated on Ser495. We confirmed the specificity of the antibodies using Ulk1/2−/− mouse embryonic fibroblasts (MEFs) transfected with myc-tagged empty vector, ULK1 WT, ULK1 S341A and ULK1 S495A mutants (Supplementary Fig. 1a, b). The p-Ser341ULK1 antibody gave a strong signal in the Ulk1/2−/− cells transfected with both ULK1 WT and ULK1 S495A mutant, but not in cells transfected with empty vector or the ULK1 S341A mutant. Likewise, the p-Ser495ULK1 antibody gave a strong signal in the Ulk1/2−/− cells transfected with both ULK1 WT and ULK1 S341A mutant, but not in the cells transfected with empty vector or ULK1 S495A mutant, confirming the specificity of these antibodies (Supplementary Fig. 1a, b). Next, using Ulk1/2+/+ MEFs, we provide evidence that phosphorylation of ULK1 on both Ser341 and Ser495 is induced by IFNα treatment (Supplementary Fig. 2). Moreover, using siRNA-mediated knockdown of PKCδ in U937 cells, we show that PKCδ expression is required for IFNα-mediated phosphorylation of ULK1 on Ser341 and Ser495, but not on Ser757, an AKT/mTOR phosphorylation site21 (Fig. 2a–c). As ULK1 is required for IFN-mediated activation of p38 MAPK, leading to transcription of ISGs21, we next sought to determine whether phosphorylation of ULK1 on Ser341 and Ser495 residues is essential for IFN-induced activation of p38 MAPK. To this end, we overexpressed myc-tagged empty vector, ULK1 WT, ULK1 S341A and ULK1 S495A in Ulk1/2−/− MEFs. These cells were then treated with IFNα and the phosphorylation of p38 MAPK was assessed. ULK1 WT, but not the ULK1 S/A mutants, could rescue IFNα-induced activation of p38 MAPK (Fig. 2d). Additionally, we overexpressed myc-tagged empty vector, ULK1 WT, ULK1 S341A and ULK1 S495A in ULK1 KO KT-1 cells. ULK1 protein expression or its absence was confirmed by anti-ULK1 immunoblotting (Supplementary Fig. 3 and Fig. 2e). The different transfected cells were then either left untreated or were treated with IFNα for 6 h and the mRNA expression of ISGs was evaluated by qRT-PCR analysis. In contrast with ULK1 S341A and ULK1 S495A, expression of the ULK1 WT substantially increased the IFNα-induced expression of IFIT1, OAS1 and IFIT3 compared to empty vector-transfected ULK1 KO KT-1 cells (Fig. 2f–h). Moreover, siRNA-mediated targeted inhibition of PKCδ expression reversed the suppressive effects of IFNα on primitive malignant erythroid precursors (Fig. 3a). Thus, as previously shown for ULK121 and p38 MAPK25, PKCδ engagement by activation of IFNAR is essential for the suppressive effects of IFNα on primary malignant PV progenitors, in vitro. Taken together these results provide evidence of previously unknown signaling events, in which activation of IFNAR controls PKCδ-mediated ULK1 phosphorylation on Ser341 and Ser495 and that ULK1 phosphorylation at these sites is required for downstream activation of p38 MAPK and transcription of ISGs.

Link to Figure 2: nature.com/articles/s41467-...

Link to Figure 3: nature.com/articles/s41467-...

To investigate whether constitutive levels of elements of this pathway may therefore discern responsiveness to IFN treatment, we examined mRNA expression levels of PKCδ, ULK1 and p38 MAPK in peripheral blood or bone marrow mononuclear cells from patients enrolled in the Myeloproliferative Disorders Research Consortium (MPD-RC)-111 study who received PEG-IFN-α2a (Pegasys) (clinical trial #NCT01259817)7 (Supplementary Table 1). Patients expressing higher pre-treatment mRNA levels of ULK1 and p38 MAPK were more likely to respond to PEG-IFN-α2a therapy (Fig. 3b), from which we infer a key role for this pathway in the anti-neoplastic effects of IFN in vivo. Future prospective clinical studies will be required to fully validate these findings and examine the potential of ULK1 and p38 MAPK as biomarkers of IFN-responsiveness.

ROCK1 and ROCK2 are overexpressed in MPN patients and interact with ULK1

Next, to identify binding partners of ULK1 in cells of MPN origin, we performed nano-LC MS/MS analysis of endogenous protein-ULK1 complexes isolated from untreated and IFNα-treated JAK2V617F-expressing HEL cells. The data indicate that ULK1 potentially interacts with 40 proteins in untreated cells, and with 38 proteins after IFNα treatment, 33 of which are identical in untreated and treated cells and 5 exclusively interact with ULK1 following IFN treatment (Supplementary Fig. 4, ProteomeXchange identifier PXD021748.). Among these five, is ROCK1 (Supplementary Fig. 4), a protein involved in promoting survival of some malignant hematopoietic cells26,27. We provide evidence that expression of both ROCK1 and, its related isoform, ROCK228, is increased in peripheral blood neutrophils from ET, PMF, and PV patients compared to age-matched healthy donors (Fig. 4a, b). Moreover, we used the DepMap Portal (depmap.org/portal/) to assess the relative protein expression of ROCK1 and ROCK2 in leukemia and other types of cancer cell lines (Supplementary Fig. 5). Interestingly, ROCK1 expression is higher in leukemia and lymphoma cell lines, compared to other types of cancer cell lines, whereas ROCK2 expression is the highest in colon cancer cells (Supplementary Fig. 5). Using co-immunoprecipitation followed by immunoblotting analyses, we found that ULK1 interacts preferentially with the cleaved/active forms of both ROCK1 and ROCK2 proteins28 in JAK2V617F-expressing cells (Fig. 4c, d).

Link to Figure 4: nature.com/articles/s41467-...

Caspases mediate cleavage and subsequent activation of ROCK1/228 and type I IFNs promote apoptosis through activation of caspases29, prompting us to examine whether IFNα treatment could induce cleavage/activation of ROCK1/2, in vivo. For these studies, we used a conditional Jak2V617F knock-in (KI) MPN mouse model30 (Fig. 5a). As expected, 5 weeks after transplantation of Jak2V617F KI bone marrow cells collected from Jak2V617F/+VavCre+ CD45.2 donor mice, a robust PV-like MPN phenotype developed in CD45.1 recipient mice (Supplementary Fig. 6a–f). Murine PEG-IFNα treatment was then initiated once a week for 4 weeks (Fig. 5a). As seen in previous MPN mouse model studies using IFNα31,32,33, PEG-IFNα treatment significantly reduced hematocrit (HCT) and red blood cell (RBC) levels in these mice (Supplementary Fig. 6g–j). Notably, PEG-IFNα treatment increased the cleavage of both ROCK1 and ROCK2 proteins in bone marrow mononuclear cells isolated from CD45.1 recipient mice (Fig. 5b–e). Additionally, we provide evidence that IFNα treatment induces activation/cleavage of caspase 3, PARP and ROCK proteins in both JAK2V617F-positive HEL cells and in JAK2V617F-negative U937 and KT-1 cells, in vitro, and these effects are repressed by co-treatment with a pan-caspase inhibitor, Z-VAD-FMK (Fig. 5f–h). In further studies, to determine whether PEG-IFNα treatment induces activation/cleavage of caspase 3 and activation of ROCK proteins in both Jak2+/+ (wild-type) and Jak2V617F/+ hematopoietic cells in vivo, we transplanted CD45.1 recipient mice with either Jak2+/+ or Jak2V617F/+ KI bone marrow cells collected from Jak2+/+VavCre- and Jak2V617F/+VavCre+ CD45.2 donor mice, respectively (Fig. 6a). Three weeks after transplantation, engraftment of Jak2+/+ and Jak2V617F/+ CD45.2+ cells in CD45.1 recipient mice was confirmed (Supplementary Fig. 7a) and a PV-like MPN phenotype developed in Jak2V617F/+, but not in Jak2+/+, recipient mice (Supplementary Fig. 7b–e). Murine PEG-IFNα treatment was then initiated once a week for 4 weeks (Fig. 6a). At week 3, 24 h post-IFNα treatment we collected peripheral blood and isolated mononuclear cells from the recipient mice (Fig. 6a, p-MYPT1 in peripheral blood mononuclear cells (PBMCs)). In both Jak2+/+ and Jak2V617F/+ recipient mice, PEG-IFNα treatment increased phosphorylation of the ROCK1/2 downstream target myosin phosphatase target subunit 1 (MYPT1)28, compared to vehicle-control-treated mice (Fig. 6b), consistent with an IFNα-induced increase in ROCK1/2 kinase activity. Importantly, IFNα-induced phosphorylation of p-MYPT1 was greater in Jak2V617F/+ compared to Jak2+/+ recipient mice (p = 0.0382). Additionally, PEG-IFNα treatment decreased hematocrit and red blood cell levels in Jak2V617F/+, but not in Jak2+/+, recipient mice (Fig. 6c, d and Supplementary Fig. 7f, g), consistent with earlier reports that IFNα treatment targets JAK2V617F-mutant over non-malignant hematopoietic cells32. Twenty-four hours after the fourth dose of PEG-IFNα, mononuclear bone marrow cells from recipient mice were isolated and we evaluated cleavage/activation of caspase 3 by immunoblotting and phosphorylation of MYPT1 by flow cytometry analyses. The data show that PEG-IFNα treatment induces cleavage of caspase 3 (Fig. 6e) and phosphorylation of MYPT1 (Fig. 6f, g) in both Jak2+/+ and Jak2V617F/+ bone marrow cells. These results were consistent with the analyses performed using bone marrow cells isolated from PEG-IFNα-treated wild-type C57BL/6 J mice (Supplementary Fig. 8). Together these results suggest that IFNα treatment induces caspase-dependent ROCK1/2 activation in normal and malignant hematopoietic cells, suggesting a potential role for ROCK1/2 in regulation of IFN-mediated biological responses.

Link to Figure 5: nature.com/articles/s41467-...

Link to Figure 6: nature.com/articles/s41467-...

Continues below:

[PhysAssist]

Inhibition of ROCK1/2 enhances IFNα-dependent anti-MPN responses

First, we sought to determine whether ROCK1/2 expression is required for transcription of ISGs. siRNA-mediated knockdown of ROCK1/2 (Supplementary Fig. 9a, b) did not enhance or inhibit IFNα-induced transcription of ISGs in JAK2V617F-positive HEL cells (Supplementary Fig. 9c–f). To determine the effects of ROCK1/2 activation on IFN anti-neoplastic responses, we evaluated the effects of genetic or drug-targeted inhibition of ROCK1/2 in JAK2V617F-positive cells, alone or in combination with IFNα. siRNA-mediated inhibition of ROCK1/2 expression (Fig. 7a) in combination with IFNα treatment significantly decreased cellular viability of SET-2 cells, compared to either ROCK1/2 knockdown or IFNα treatment alone (Fig. 7b). Similar effects were observed in SET-2 (Fig. 7c, left panel) and HEL (Fig. 7c, middle and right panel) cells treated with the ROCK1/2 inhibitors GSK429286A or Fasudil in combination with IFNα treatment, compared to treatment with either inhibitor alone. Moreover, GSK429286A in combination with IFNα treatment significantly reduced the clonogenic capacity of HEL cells, compared to either inhibitor alone (Fig. 7d). Importantly, siRNA-mediated knockdown of ROCK1/2 expression or drug-targeted inhibition of ROCK1/2 activity with Fasudil enhanced the suppressive effects of IFNα on primitive malignant erythroid precursors from PV patients who responded poorly to IFN treatment (Fig. 7e, f).

Link to Figure 7: nature.com/articles/s41467-...

Next, we examined the effects of targeting ROCK1/2 on the inhibitory effects of IFNα in our MPN mouse model. Five weeks after transplantation of CD45.1 recipient mice with Jak2V617F KI bone marrow cells (Supplementary Fig. 10a–e), we treated mice for 4 weeks with the ROCK inhibitor Fasudil34 four times per week, with or without PEG-IFNα treatment given once a week (Fig. 8a). Mice tolerated single agent and combination drug treatments well, as monitored by body weight (Supplementary Fig. 10f). At week 3 after initiation of treatment, we collected peripheral blood and isolated mononuclear cells from the recipient mice 26 h post-IFNα treatment and 2 h post-Fasudil treatment or its vehicles (Fig. 8a, p-MYPT1). PEG-IFNα treatment increased phosphorylation of MYPT1 compared to vehicle-control and Fasudil only treated mice (Fig. 8b), and co-treatment with Fasudil and PEG-IFNα inhibited this increase in peripheral blood mononuclear cells (Fig. 8b). Four weeks post-treatment, further decreases in hematocrit (HCT) and hemoglobin (HB) were noticeable in the IFNα + Fasudil-treated mice, compared to mice that were treated with vehicle or either drug alone (Fig. 8c, d). Red blood cell counts were significantly decreased in both IFNα- and IFNα + Fasudil-treated mice compared to vehicle- and Fasudil-treated mice (Supplementary Fig. 10g). By contrast, no significant differences were observed in platelet levels among the four treatment groups (Supplementary Fig. 10h). Additionally, the reductions in spleen weight, percentage of bone marrow and spleen Ter119highCD71high basophilic erythroblasts (R2), and bone marrow myeloid progenitors, granulocyte-monocyte progenitors (GMPs), and megakaryocyte-erythroid progenitors (MEPs) between vehicle-treated and IFNα + Fasudil-treated mice are greater than between vehicle-treated and IFNα-treated mice (Fig. 8e–j). No significant differences were observed between IFNα and IFNα + Fasudil-treated mice for bone marrow and spleen Ter119medCD71high proerythroblasts (R1), Ter119highCD71med polychromatophilic erythroblasts and Ter119highCD71low orthochromatophilic erythroblasts (R3 + R4) and for bone marrow hematopoietic stem cells (Lin-Sca+Kit+, LSK) and common myeloid progenitors (CMP) (Supplementary Fig. 11). Taken together with the in vitro findings, these results suggest that IFNα treatment activates ROCK1/2 proteins in a negative feedback regulatory manner, triggering mechanisms that suppress its anti-neoplastic effects in MPNs (Fig. 9), and inhibition of activated ROCK1/2 proteins enhances the anti-neoplastic effects of IFNα in MPNs.

Link to Figure 8: nature.com/articles/s41467-...

Link to Figure 9: nature.com/articles/s41467-...

Discussion

Over the past 4 decades, accumulating evidence has revealed that type I IFNs exert important anti-tumor effects and have a critical role in immune surveillance against the development of malignancies3,6. A precise understanding of the mechanisms of action of type I IFNs on both tumor and immune cells is critical, as this will provide the means to identify, in advance, patients who will respond to IFN treatment and thereby predict clinical responses. Moreover, this knowledge will lead to new therapeutic approaches and the development of combination treatments to enhance the action(s) of IFNs and/or overcome resistance mechanisms. It is well established that activation of IFNAR initiates signaling events mediated via the classical JAK/STAT pathway and many other “non-classical” pathways6,35. It is the concerted actions of these signaling pathways and their effectors that drive the biological outcomes of IFN treatment and, at the same time, control excessive responses35. We now identify PKCδ as a key initiator of the p38 MAPK pathway and demonstrate that PKCδ binds to and directly phosphorylates ULK1 on serine residues 341 and 495. We establish that phosphorylation of ULK1 on these sites is required for downstream engagement of p38 MAPK and transcription of ISGs. Activation of PKCδ is critical for the generation of IFN responses, as we provide evidence that PKCδ knockdown reverses the IFNα-mediated anti-clonogenic effects in PV patient primary cells. Using MPN patient samples collected prior to initiation of PEG-IFNα therapy, we observed that higher baseline expression levels of the downstream effectors of PKCδ, namely ULK1 and p38 MAPK, predict a better response to IFNα treatment. These findings suggest that ULK1 and p38 MAPK could potentially be used as part of a pool of biomarkers to predict response to IFNα-based therapies for MPN patients, but future large prospective clinical studies will be required to definitively address this.

[PhysAssist in reply to PhysAssist]

In previous studies, PKCδ has been characterized as both a tumor-promoting and a tumor-suppressing kinase, dependent on the stimulus and cell-type context36,37,38,39. PKCδ promotes cellular apoptosis through both p38 MAPK-dependent and -independent mechanisms, induces cell cycle arrest and inhibits tumor growth and progression39. By contrast, PKCδ enhances the migration, invasion, and metastasis of some tumor cells40,41,42,43. Similarly, ULK1 has been shown to have opposing roles in cancer, again depending on the cellular context and stimulus44. In response to stress, such as starvation, ULK1 is phosphorylated on Ser555 by AMPK to induce autophagy45, shown to play tumor-promoting and tumor-suppressing functions in a context-dependent manner46. Additionally, other stimuli, such as IFNs, induce ULK1 engagement in autophagy-independent signaling events, controlling several biological functions21,44,47,48,49,50,51. Phosphorylation of ULK1 on Ser757 (autophagy-inhibitory site) is associated with shorter overall survival in non-small cell lung cancer patients52. In breast cancer patients, low levels of ULK1 correlate with lymph node metastasis and poor survival53. Hypoxia-induced expression of the kinase-dead ULK1 mutant increases lung metastasis in a breast cancer xenograft mouse model54 and, in another study, ULK1 was shown to phosphorylate Exo70 to inhibit breast cancer metastasis55. In contrast, a kinase-dead ULK1 mutant suppressed growth and metastasis and increased survival in a neuroblastoma xenograft mouse model56. We now report that activation of a PKCδ-ULK1-p38 MAPK signaling pathway is required for IFNα-induced anti-neoplastic effects in MPN patients.

Interestingly, in a recent study, the atypical PKC λ/ι isoform was found to directly phosphorylate ULK2, inhibiting its activity and promoting its degradation57. Inactivation/loss of PKCλ/ι was found to induce ULK2 activity, which directly activates the TBK1-STING-IRF3 pathway, leading to IFN production and anti-tumor responses57. In another study, the PKCδ isozyme, PKCα, was shown to induce phosphorylation of ULK1 on Ser423, reducing autolysosome formation and promoting ULK1 degradation58. Additional studies are required to interrogate the potential role of PKCδ-induced phosphorylation of ULK1 on Ser341 and Ser495 in the regulation of autophagy and IFN production.

Beyond defining the mechanism by which the ULK1/p38 MAPK pathway is engaged in PV cells, our studies have identified an important feedback mechanism that negatively regulates IFN-responses, that may be a therapeutic target. This involves ROCK1/2, shown to interact with ULK1. Importantly, we provide evidence that IFNα treatment induces cleavage/activation of ROCK1/2 in vitro and in vivo in both JAK2+/+ and JAK2V617F/+ hematopoietic cells. Since IFNα treatment does not affect hematocrit, hemoglobin, red blood cells and platelet levels in Jak2+/+ recipient mice, our results suggest that IFN-mediated activation of ROCK1/2 could represent a feedback mechanism to control the IFN effects on normal hematopoiesis. In contrast, we found that both ROCK isoforms are overexpressed in peripheral blood cells from MPN patients, compared to healthy individuals and genetic or pharmacological inhibition of ROCK1/2 enhanced IFNα-induced anti-neoplastic effects in primary MPN cells and in a MPN model in vivo. Our findings support a role for the IFN-mediated activation of ROCK1/2 as pro-tumorigenic in MPN cells. In other studies, ROCK1/2 have been shown to regulate actomyosin contractility and cytoskeleton assembly, important features for tumor progression, migration and invasion59,60,61,62. ROCK1 and ROCK2 have also been shown to have redundant roles in cell cycle progression and have been implicated in tumorigenesis63. Genetic deletion of not one, but both ROCK isoforms, blocked tumor formation in mouse models of non-small cell lung cancer and melanoma63. Recently, engagement of ROCK-myosin II signaling was identified as a mechanism of resistance to MAPK inhibitors in melanoma64. Furthermore, the ROCK inhibitor, Fasudil, has been shown to decrease disease burden in myeloproliferative disease (MPD) mouse models27,65 and combination of imatinib with either Fasudil or the ROCK inhibitor Y-27632 showed synergistic inhibition of CD34+ chronic myeloid progenitor cell growth66. To date, the ROCK inhibitors Fasudil and Ripasudil are approved in Japan for the treatment of cerebral vasospasm and glaucoma, respectively, and several ROCK inhibitors are being tested clinically for the treatment of different diseases, including cancer34.

Viewed altogether, our data identify an important IFNα-regulated negative feedback mechanism that leads to caspase-dependent cleavage/activation of ROCK1/2, which interacts with ULK1, and negatively regulates IFNα anti-neoplastic responses. These findings support future studies to evaluate the clinical potential of combinations of ROCK inhibitors with PEG-IFNα and/or the JAK inhibitor ruxolitinib33 for the treatment of MPN patients.

Intensely esoteric Methods section not copied d/t incomprehensibility....

Full-text Source: nature.com/articles/s41467-...