SHP2 inhibition displays efficacy as a monotherapy and in combination with JAK2 inhibition in preclinical models of myeloproliferative neoplasms

Myeloproliferative neoplasms (MPNs), including polycythemia vera, essential throm-bocytosis, and primary myelofibrosis, are clonal hematopoietic neoplasms driven by mutationally activated signaling by the JAK2 tyrosine kinase. Although JAK2 inhibitors can improve MPN patients' quality of life, they do not induce complete remission as disease-driving cells persistently survive therapy. ERK activation has been highlighted as contributing to JAK2 inhibitor persistent cell survival. As ERK is a component of signaling by activated RAS proteins and by JAK2 activation, we sought to inhibit RAS activation to enhance responses to JAK2 inhibition in preclinical MPN models. We found the SHP2 inhibitor RMC-4550 significantly enhanced growth inhibition of MPN cell lines in combination with the JAK2 inhibitor ruxolitinib, effectively preventing ruxolitinib persistent growth, and the growth and viability of established ruxolitinib persistent cells remained sensitive to SHP2 inhibition. Both SHP2 and JAK2 inhibition diminished cellular RAS-GTP levels, and their concomitant inhibition enhanced ERK inactivation and increased apoptosis. Inhibition of SHP2 inhibited the neoplastic growth of MPN patient hematopoietic progenitor cells and exhibited synergy with ruxolitinib. RMC-4550 antagonized MPN phenotypes and increased survival of an MPN mouse model driven by MPL-W515L. The combination of RMC-4550 and ruxolitinib, which was safe and tolerated in healthy mice, further inhibited disease compared to ruxolitinib monotherapy, including extending survival. Given SHP2 inhibitors are undergoing clinical evaluation in patients with solid tumors


| INTRODUCTION
First described over 70 years ago, myeloproliferative neoplasms (MPNs) are chronic blood neoplasms driven by mutations in hematopoietic stem cells. 1 Classical Philadelphia chromosome negative MPNs include polycythemia vera (PV), essential thrombocythemia (ET), and primary myelofibrosis (PMF).These neoplasms are characterized by marrow myeloproliferation (erythroid, megakaryocytic, and/or granulocytic), disease-related symptoms that reduce patients' quality of life, and an increased risk of thrombotic and hemorrhagic complications. 2 Myelofibrosis (MF) is the most advanced form of MPN with a median survival of 6-7 years. 3MF can occur de novo (PMF) or following a prior diagnosis of PV or ET and can, in turn, transform to acute myeloid leukemia. 4Ns are driven by mutations that activate the JAK2 tyrosine kinase, including mutations in the JAK2, MPL, and CALR genes, the latter two leading to activation of the thrombopoietin receptor (encoded by MPL) and subsequent activation of JAK2. 5,6Such mutations are present in nearly all MPN patients and expression of these mutations in mice induces MPN, confirming their disease driving properties. 3,7,8xolitinib was the first FDA-approved JAK2 inhibitor for MPNs based upon its ability to improve symptoms and quality of life; however, its ability to significantly modify underlying disease biology is limited.0][11][12][13] The survival benefit of ruxolitinib in patients with MF underscores the critical role JAK2 inhibition has in treating MPNs, but even this has been attributed to its anti-inflammatory and spleen reducing impacts.Ultimately, the inability of ruxolitinib and other recently approved JAK2 inhibitors (e.g., fedratinib and pacritinib) to effectively antagonize allele burden in patients indicates alternative therapeutic approaches are needed to overcome the persistent survival of MPN cells during JAK2 inhibitor treatment.
How MPN cells persistently evade the effects of JAK2 inhibitors has been the focus of multiple studies. 14Koppikar et al. 15 identified reactivation of JAK2 signaling (including ERK and STAT) by JAK2 heterodimerization with other JAK family members during inhibitor treatment.Winter et al. 16 used gain-of-function studies and showed activation of the RAS pathway supports JAK2 inhibitor persistence.Studies using mouse models led to the identification of JAK2-independent pathways that lead to sustained MEK/ERK signaling during JAK2 inhibitor therapy. 179][20] Together these studies suggest activation of RAS/-MEK/ERK signaling may negatively impact the efficacy of JAK2 inhibitors. 14Indeed, MEK and ERK inhibition have been shown to antagonize MPN phenotypes in mice and improve responses to ruxolitinib therapy including, importantly, antagonizing mutant allele burden levels. 17,18,21Notably, downstream targets and regulators of MAPK signaling, DUSP1 and DUSP6, have recently been shown to play a role in JAK2 inhibitor persistence and progression of disease. 22,23e SHP2 (Src homology region 2 domain-containing phosphatase-2) tyrosine phosphatase mediates RAS activation and subsequent MEK/ERK signaling downstream of cytokine receptors and receptor tyrosine kinases by relieving negative regulatory phosphorylation events on various proteins that normally antagonize RAS activation. 24,259][50] SHP2 inhibitors are being tested clinically for multiple different cancer types.
Given the evidence suggesting activation of MEK/ERK signaling can antagonize the effects of JAK2 inhibition, 14,[16][17][18]21 we investigated the potential SHP2 inhibition may have as a therapeutic strategy alone or in combination with JAK2 inhibition in MPN. Ourpreclinical results indicate SHP2 inhibition has activity as a monotherapy and can enhance effects of JAK2 inhibition, suggesting SHP2 inhibitors may be candidates for clinical investigation to improve targeted therapeutic approaches for MPN patients.
Cell growth was determined by cell counting using trypan blue exclusion.GI 50 assessments were determined using Cell-Titer Glo 2.0 (Promega) following incubation of cells in quadruplicate in 384-well plates with the indicated concentrations of drug for 48 or 72 h, depending on the cell line, and data were analyzed by Prism 10 (GraphPad Software, LLC) with GI 50 values determined following curve fitting with four parameter nonlinear regression.

| Studies using primary cells
Blood and bone marrow samples were obtained from MPN patients (approximately equal male and female) upon written informed consent through the Institutional Review Board-approved Moffitt Cancer Center Total Cancer Care protocol (MCC 14690) and with approval by the Moffitt Cancer Center Scientific Review Committee.Mononuclear cells from MPN patient samples and healthy controls were isolated by ficoll density separation.

| Ex vivo colony-formation assays
Colony formation studies were performed in the presence of stem cell factor, IL-3, and GM-CSF using MethoCult™ H4534 (Stem Cell Technology) or Human Methylcellulose Complete Media Without Erythropoietin (HSC004, R&D Systems).Erythropoietin (3 U/mL) (Amgen, Inc.) was added to assess BFU-E formation of cells from patients with MPL and CALR mutations as well as healthy controls.Colonies were scored after 12-14 days of incubation using a Zeiss Stemi 2000-C stereo microscope.

| Ex vivo treatment
Bone marrow mononuclear cells were enriched for CD34+ cells (EasySep™ Human CD34 Positive Selection Kit II, Stem Cell Technologies) which were cultured in IMDM supplemented with 2% FBS and DMSO or RMC-4550 for 16 h.

| In vivo therapeutic studies
Studies utilizing mice were performed at the Moffitt Cancer Center following a research protocol approved by the Institutional Animal Care and Use Committee.The MPL-W515L bone marrow transplant model of MPN was performed essentially as described. 6,51Briefly, bone marrow cells retrovirally infected to express MPL-W515L (or MPL-WT) and GFP were transplanted by tail vein injection into lethally irradiated (2 Â 450 cGy, 24 h apart) syngeneic recipient 8-week-old Balb/c or C57BL/6 mice (Jackson Laboratory).Complete blood counts (CBCs) were performed weekly following transplant to assess engraftment and disease development.Randomized cohorts containing equal white blood cell counts (WBCs) were generated once animals displayed leukocytosis (WBCs ≥20 000/μL) and were treated with the dosing strategies detailed in the Results section.RMC-4550 was formulated in 2% hydroxypropyl methylcellulose E-50 and 0.5% Tween-80 in 50 mM Sodium Citrate Buffer, pH 4.0, and ruxolitinib was formulated in 5% N,N-Dimethylacetamide (DMAC) (Sigma-Aldrich, Inc.) and 95% of 0.5% methylcellulose.All administrations were by oral gavage.CBCs were performed weekly following initiation of treatment or until vehicle-treated animals exhibited signs of endpoint and succumbed to disease.

| Histology
After tissue fixation, and decalcification of bones, tissues were paraffin embedded and slices were prepared and stained with H&E and reticulin (Chandler's Precision Reticulum Stain, StatLab).Quantitation of megakaryocytes was assessed based on five independent fields of H&E-stained slides (magnification 200Â) and an average score per mouse was determined.Images were captured using an Olympus BX51 microscope with 20Â or 60Â objectives (200Â and 600Â total magnification) at room temperature using cellSens™ Standard software (Olympus LS).All pathological analyses were performed blindly.

| Flow cytometry
Flow cytometry was performed as previously described. 51

| Immunoblot
Immunoblots of total bone marrow cells and splenocytes were performed following cell isolation from tissues after 6 h of a single oral dose of RMC-4550.

| Statistical analyses
All statistical analyses were performed using Prism 10 (GraphPad Software, LLC).Synergy was assessed using the Bliss independence model, calculated with the formula: E(Bliss) = Ea + Eb -EaÂEb, where Ea/b is the inhibition of each inhibitor assessed, and 0 ≤ E ≤ 1.
Synergy score was then calculated as: Synergy score = 100Â(Eab (observed) -E(Bliss)), where Eab (observed) is the inhibition observed with the combination of inhibitors.A synergy score >0 (i.e., Eab (observed) > E(Bliss)) is considered synergistic. 52 I G U R E 1 Legend on next page.

| SHP2 inhibition suppresses ERK activation in, and proliferation of, JAK2-driven MPN cell lines
7 ERK activity is generally mediated by activated RAS proteins and given the known role of SHP2 in controlling RAS activity in response to signals initiated by kinase activation we investigated if SHP2 contributes to JAK2-V617F-driven cell growth.30 First, we assessed if active SHP2 could be detected in MPN model cell lines.Phosphorylation of SHP2 on tyrosine 542 positively regulates SHP2 phosphatase activity and is used as a marker for activation.30,53,54 We assessed the JAK2-driven MPN model cell lines SET2 and BaF3-JAK2-V617F by immunoblot and detected phosphorylated SHP2-Y542 (pY-SHP2) in these cells (Figure 1A). GivenSHP2 activation is mediated by upstream tyrosine kinase signaling, 30,55 we treated cells with the JAK1/2 inhibitor ruxolitinib which led to an attenuation of pY-SHP2 in cells, suggesting activating tyrosine phosphorylation of SHP2 is dependent on JAK activity (Figure 1A) in these cells.
Next, we treated JAK2-V617F-driven cells with the allosteric SHP2 inhibitor RMC-4550 to assess effects on cell growth. 45RMC-4550 exhibited GI 50 values of 0.31, 0.47, and 2.1 μM in SET2, UKE1, and BaF3-JAK2-V617F cell lines, respectively (Figure 1B).RMC-4550 was over 10 fold more potent than another allosteric SHP2 inhibitor, SHP099, at inhibiting cell growth, likely reflecting the fact that the biochemical potency of RMC-4550 is over 10 fold higher (Figure S1A). 45,49The sensitivity of these cells to ruxolitinib is shown for comparison (Figure 1B).To confirm on-target inhibition of SHP2 by RMC-4550, we immunoblotted for phosphorylated/activated ERK (pERK) and found RMC-4550 diminished pERK levels in a concentration dependent manner, suggesting ERK activity in these JAK2-driven cells is dependent on SHP2 activity (Figure 1C,D).

| Concomitant inhibition of SHP2 and JAK2 synergistically suppresses growth of MPN cell lines, enhances apoptosis, and suppresses RAS and ERK activation
We next combined SHP2 and JAK2 inhibition and assessed longterm cell proliferation.Using concentrations of each inhibitor that diminish but do not prevent proliferation, the combination of RMC-4550 with ruxolitinib strongly antagonized, and effectively prevented, long-term proliferation of SET2 and UKE1 cells (Figure 1E).
Identical results were obtained combining ruxolitinib with SHP099 (Figure S1D).This combination effect was not observed with the two non-JAK2-driven leukemia cell lines tested, SKM-1 and Jurkat (Figure S1E).The combination of RMC-4550 and ruxolitinib enhanced apoptosis of SET2 and UKE1 cells compared to treatment with single agents (Figure 1F).
We next assessed pERK levels following treating cells with both RMC-4550 and ruxolitinib.Treatment of SET2, UKE1, and BaF3-JAK2-V617F cells with a combination of individually suboptimally effective concentrations of inhibitors led to greater suppression of pERK than treatment of either inhibitor alone, and enhanced SHP2 activity mediates RAS activation, the exchange of GDP for GTP bound to RAS proteins via a variety of mechanisms, leading to positive regulation of ERK activity. 25,59Treatment of SET2 and UKE1 cells with RMC-4550 and ruxolitinib each led to a detectable decrease in the amount of cellular RAS-GTP, most evident at higher concentrations, as did the combination of these inhibitors (Figure S2), as measured by a RAS-GTP pulldown assay.These results are consistent with Nichols et al. 45 that demonstrate that higher RMC-4550 concentrations are required to detect changes in wild-type RAS-GTP levels than are required to diminish pERK levels in solid tumor models.Differential dynamic regulation of RAS-GTP levels and the phosphorylation of ERK could also contribute to changes in pERK being easier to capture at lower inhibitor concentrations than changes in RAS-GTP levels.
Finally, SET2 cells derived to persistently proliferate in ruxolitinib (about five times the GI 50 ) remained sensitive to SHP2 inhibition, as 0.1 μM of RMC-4550 prevented continued proliferation and induced loss of cell viability over time (Figure S3A).
Similar results were obtained with ruxolitinib persistent BaF3-JAK2-V617F cells (Figure S3B).The level of pERK in persistent cells remained sensitive to RMC-4550, and levels of pRSK3 were also reduced, indicating SHP2 continued to mediate signaling to ERK activation in the context of JAK2 inhibitor persistence (Figure S3C).

| SHP2 inhibition antagonizes growth of primary MPN patient cells and synergistically enhances growth inhibition with ruxolitinib
To translate our findings to primary MPN patient cells we assessed the effect of SHP2 inhibition on the growth of hematopoietic progenitor cells in colony forming assays.These cells from JAK2-V617F+ MPN patients can form erythroid colonies in the absence of erythropoietin (termed endogenous erythroid colonies, EECs), a pathologic hallmark induced by aberrant JAK2 activation.
RMC-4550 inhibited this neoplastic EEC as well as CFU-GM formation in a concentration dependent manner (Figure 2A) (see Figure S4 for patient details).Colony formation of cells from every patient sample utilized was sensitive to SHP2 inhibition (n = 8) (Figure 2B).Similar results were obtained using SHP099, albeit at higher concentrations, again, likely reflecting its lower biochemical activity compared to RMC-4550 (Figure S5A,B).Treatment of CD34+ selected bone marrow cells from an MPN patient with RMC-4550 led to decreased pERK and pRSK3, indicating expected on-target activity in relevant primary cells (Figure 2C).We next combined SHP2 and JAK2 inhibition in colony forming assays.In our experience, ruxolitinib has an IC 50 of about 50 nM against EEC formation. 60-62RMC-4550 further suppressed EEC formation inhibition observed with ruxolitinib in all patient samples (n = 6) (Figure 2D).This neoplastic erythroid colony formation was nearly completely suppressed in multiple samples with both inhibitors in combination (Figure 2D), which exhibited synergy against EEC formation in all samples (Figure 2F).Similar results were obtained assessing CFU-GMs (Figure 2E), although synergy was less frequent (Figure 2F).Again, results of these experiments were confirmed with SHP099 (Figure S5C-E  We next sought to determine the therapeutic potential of SHP2 inhibition for MPN.We utilized the MPL-W515L transplant model, which aggressively models MPN phenotypes including leukocytosis, thrombocytosis, organomegaly, and bone marrow fibrosis. 6Following transplant with bone marrow cells infected with retrovirus to express MPL-W515L along with GFP, recipient mice were placed into cohorts with equal levels of leukocytosis and treated once daily with RMC-4550, which we chose due to its higher activity against SHP2 than SHP099 and its established PK properties and activity in vivo. 45Treatment for 1 week at 10 mg/kg and 30 mg/kg suppressed the advancing leukocytosis observed in vehicle treated animals but did not influence platelet levels (Figure 3A,B).RMC-4450 significantly antagonized the development of hepatosplenomegaly that was evident in vehicle treated animals (Figure 3C,D).RMC-4550 treatment also led to a decrease in the percent of GFP+ cells in the peripheral blood (Figure 3E), a surrogate for allele burden in this model.MPL-W515L mice treated with 30 mg/kg of RMC-4550 showed improved whole-body weights, similar to non-diseased mice transplanted with MPL-WT, suggestive of improved health of these mice (Figure 3C).We confirmed on target SHP2 inhibition in bone marrow cells and splenocytes of MPL-W515L mice treated with a single dose of RMC-4550, where treatment with the SHP2 inhibitor suppressed pERK in bone marrow cells and pERK and pRSK3 in splenocytes (Figure 3F).

| The SHP2 inhibitor RMC-4550 along with ruxolitinib treatment provides superior suppression of MPN phenotypes in the MPL-W515L mouse model
We next assessed if combining SHP2 and JAK2 inhibition could enhance suppression of MPN.Because of the impressive single agent activity of RMC-4550 in the MPL-W515L model (Figure 3), we modified the RMC-4550 dosing strategy to assess the combination therapy, this time using every other day dosing.To get a better sense of the potential the combination treatment may provide improved therapeutic responses, we also used a lower dosing of ruxolitinib than standardly used in this model, as the effect of ruxolitinib can be dominant in combination therapy studies.Single agent therapy of RMC-4550 for 7 days decreased WBC levels from those present at treatment initiation (Day 0) and prevented the advancing leukocytosis observed in vehicle treated mice (Figure 4A), in line with the daily dosing strategy in our initial mono-therapy experiment (Figure 3A).Mice treated with ruxolitinib for 7 days had WBC levels similar to those present upon treatment initiation, indicating this dosing of ruxolitinib suppressed the increasing leukocytosis observed in vehicle treated mice (Figure 4A).The combination of RMC-4550 and ruxolitinib led to profound inhibition of WBC levels compared to vehicle control and significantly diminished leukocytosis compared to monotherapies (Figure 4A).After 14 days of treatment, this trend continued, with the average WBC levels following ruxolitinib treatment being over fivefold that of the combination treated animals, the latter still having levels lower than when treatment was initiated (Figure 4A).After 7 days of treatment, neither monotherapy suppressed thrombocytosis compared to vehicle treated mice (Figure 4B), but there was a suppression of platelet levels by the combination therapy compared to vehicle treatment (Figure 4B).While no treatment demonstrated suppression of platelet levels compared to vehicle after 14 days of treatment, the combination of RMC-4550 and ruxolitinib provided superior suppression of platelets compared to ruxolitinib treatment alone (Figure 4B).
RMC-4550 and ruxolitinib monotherapy treatment each suppressed hepatomegaly compared to vehicle treatment, and the addition of RMC-4550 to ruxolitinib suppressed liver size further (Figure 4C).Effects on spleen size, however, were not observed with any treatment (Figure S8A).Both RMC-4550 and the combination therapy resulted in suppression of GFP+ cells in the peripheral blood, and RMC-4550 treatment led to significant enhancement of survival compared to vehicle and ruxolitinib monotherapy mice (Figure 4D,E).
There was no difference in survival between RMC-4550 monotherapy and the combination following the termination of treatment after 24 days (Figure 4E).
We next performed pathological assessment of liver, spleen, and bone marrow to determine the effects of SHP2 and JAK2 inhibitor therapy on the disease state of these tissues.H&E staining indicated (Figure 4H).Results were less clear in bone marrow where only the combination led to a trend toward inhibition of megakaryocyte hyperplasia compared to vehicle (Figure S8C).Assessment of fibrosis in the bone marrow and the spleen using reticulin staining indicated neither RMC-4550 nor ruxolitinib improved fibrosis compared to vehicle mice, and there was no improvement of fibrosis detected using the combination of RMC-4550 and ruxolitinib (Figure S8D).
Given MPNs are associated with levels of inflammatory cytokines and targeting MEK/ERK signaling affects cytokine levels in MPN models, 17,21 we utilized serum obtained at endpoint to determine the effect of SHP2 inhibition on inflammatory cytokine levels (Figure S9).
Notably, while RMC-4550 monotherapy treatment led to a trend to lower levels of TNFα, which has been implicated in MPN pathogenesis, [63][64][65] statistically significant inhibition of TNFα levels in combination with ruxolitinib was evident compared to vehicle or ruxolitinib monotherapy mice (Figure 4I).In addition, compared to vehicle treated mice, the combination therapy led to a statistically significant decrease in several cytokines, including CXCL2, IL-2, IL-9, and IL-3, while RMC-4550 monotherapy led to a decrease in these cytokines as well as GM-CSF and M-CSF (Figure S9B).
It should be noted that in these pathological analyses, the vehicle treated mice were analyzed on the ninth day of treatment, at an early endpoint due to their rapidly declining health which was prior to a planned endpoint for pathological analyses after 14 days of treatment of all cohorts.This may have suppressed the differences in vehicle treated mice compared to other cohorts (e.g., possibly lower reticulin staining score, megakaryocytes, and blood cell counts for vehicle mice).Nonetheless, RMC-4550 addition to ruxolitinib displayed superior improvement to multiple disease phenotypes.Finally, MPN mice treated with the combination of RMC-4550 and ruxolitinib exhibited similar body weights as those treated with ruxolitinib monotherapy (Figure S8E) and 4 weeks of treatment of healthy wild-type mice indicated the combination therapy was well tolerated (Figure S10).

| DISCUSSION
JAK2 inhibitor therapy can afford quality of life benefits to MPN patients via amelioration of constitutional symptoms, a result of inhibition of the inflammatory nature of MPNs. 66[13] Still, the inability of JAK2 inhibitors to efficiently alter the natural course of disease (e.g., decrease mutant driver allele burden), however, exemplifies efficacy limitations of this therapy.Thus, there remains a critical need to improve targeted therapies for MPN patients.
While JAK2-V617F induces MPN in mice, genetic removal of this mutation reverts MPN phenotypes, indicating aberrant JAK2 signaling is required to maintain disease. 67These findings suggest improved targeting of aberrant JAK2 signaling has potential to improve molecular responses in patients.While JAK2 mutations are not a mechanism of JAK2 inhibitor inefficacy, potential mechanisms of inhibitor persistence have been described.Persistent survival of cells to type 1 JAK2 inhibitors (e.g., ruxolitinib) is associated with induced heterodimeric formation of JAK2 with other members of the JAK family, allowing for continued JAK-dependent signaling. 15Chronic treatment with type I JAK2 inhibitors stabilizes phosphorylation of JAK2, which may contribute to resistance signaling. 68,69In addition, activation of PDGFR signaling has been identified as a JAK2-independent mechanism by which some signals downstream of JAK2 can be maintained in the presence of JAK2 inhibition. 1718]21 In addition to developing more effective and potentially mutant specific drugs to inhibit JAK2 kinase activity, many combination therapies have been assessed pre-clinically to enhance inhibition of JAK2 activated signaling pathways or to inhibit parallel pathways.For example, combination therapies with MEK or ERK inhibition, among others, with ruxolitinib lead to deeper responses in mouse models than monotherapies. 17,18,21,70Two combination approaches are showing promise in clinical trials for myelofibrosis, using co-targeting of JAK2 with BCL2 family members to target cell survival, 71 or with BET domain containing proteins to further inhibit the inflammatory signals associated with disease phenotypes and support of the malignant clone. 72,73 this study, we identified SHP2 as a potential therapeutic target been obtained with higher ruxolitinib dosing, as we intentionally strategized the dosing to identify a potential combination effect that was not dominated by a single agent.While SHP2 has been proposed to be a therapeutic target in non-classical MPNs such as mastocytosis, 33 our preclinical studies suggest SHP2 inhibition has therapeutic efficacy as a monotherapy and may improve the efficacy of JAK2 inhibitor monotherapy in classical MPNs.
SHP2 activity leads to RAS activation and signaling. 25,30mpared to direct MEK/ERK kinase inhibition, SHP2 inhibition may antagonize activation of other RAS effector pathways that potentially contribute to the effects of activated JAK2 in MPN.In this regard, while MEK and ERK inhibition provided enhanced activity with ruxolitinib in MPN mouse models, direct inhibition of these kinases alone displayed limited activity against splenomegaly and leukocytosis, 17,21 unlike our observations with SHP2 inhibition monotherapy.As SHP2 is a regulator of RAS/ERK activation downstream of cytokine receptors and tyrosine kinases, 24,25,30,74 its inhibition may also antagonize RAS/ERK signaling by potential JAK2-independent mechanisms that contribute to JAK2 inhibitor persistence.SHP2 is a direct effector of PDGFR signaling, 24,75,76 which may contribute to JAK2 inhibitor persistence. 17[77][78][79][80][81][82][83][84][85][86][87][88][89] The addition of RMC-4550 to ruxolitinib in our studies drove the levels of TNFα, which contributes to clonal expansion in MPN models, 63 below those detected after vehicle and ruxolitinib monotherapy.While BET inhibition antagonizes MPN associated inflammatory signals via suppression of NFκB activity, 73 SHP2 has been implicated in positively regulating signaling by IL-1β, [90][91][92] IL-6, 93,94 CXCL8/IL-8, 95,96 as well as IL-13, 97 which has recently been implicated in driving myelofibrosis in MPN models. 86SHP2 can also promote NFκB activation and IL-6 production induced by IL-1 signaling 98 in fibroblasts, but we did not detect modulation of the phosphorylation of p65 in MPN model cell lines or in cells from our mouse model experiments (Figure S11).Interestingly, TGFβ-induced dermal and pulmonary fibrosis can be blocked by SHP2 inhibition. 99Fisher et al. 100 demonstrated that ruxolitinib therapy alone cannot diminish a subset of MPN-associated cytokines (e.g., IL-6 and IL-8, among others), and suggest inhibition of NFκB and ERK may be required for more complete cytokine suppression.Thus, SHP2 inhibition has the potential to antagonize diverse signaling pathways, including growth promoting and inflammatory signals, that may contribute to MPN pathogenesis.5][106] Thus, it is possible SHP2 inhibition may enhance anti-neoplastic cell immunity in MPN.
In summary, our preclinical studies show that SHP2 may be a viable therapeutic target as a monotherapy and to enhance the inhibition of JAK2 signaling in MPN.Inhibiting SHP2 may improve the efficacy of targeted therapeutic approaches for MPN patients, potentially leading to better upfront efficacy and countering the development of JAK2 inhibitor persistence.
inhibition antagonizes the growth of, and signaling to ERK activation in, JAK2-V617F MPN model cells and the combination of SHP2 inhibition with ruxolitinib prevents long-term growth, enhances apoptosis, and further enhances ERK inhibition.(A) SET2 and BaF3-JAK2-V617F cells were left untreated or treated with ruxolitinib (0.25 and 1.0 μM, respectively) for 2 h.Cell lysates were prepared and immunoblotted for pY542-SHP2 (a marker of SHP2 activation, pY-SHP2) and total SHP2.(B) Relative viable SET2, UKE1, and BaF3-JAK2-V617F cells were determined after treatment with a range of the SHP2 inhibitor RMC-4550 (top) and ruxolitinib (bottom) for 72 hours for SET2 and UKE1 cells and 48 hours for BaF3-JAK2-V617F cells.The mean GI 50 (±SD, n = 3-7 independent experiments) for each cell line is shown in the bar graphs (left) along with representative concentration-response curves (right).(C) SET2 cells and (D) UKE1 cells were treated with 0, 0.025, 0.05, 0.1, 0.5, and 1 μM of RMC-4550 for 4 h and lysates were immunoblotted for pERK, pRSK3, pSTAT5, and vinculin or GAPDH.(E) SET2 and UKE1 cells were incubated with DMSO (black lines), ruxolitinib (0.1 μM for SET2 and 1 μM for UKE1, blue lines), RMC-4550 (1 μM, red lines), and both inhibitors together (purple lines).Total viable cell numbers were determined by trypan blue exclusion over time.(F) SET2 cells and UKE1 cells were treated with DMSO, ruxolitinib (0.05 μM for SET2 cells and 0.1 μM for UKE1 cells), RMC-4550 (0.1 μM), and both inhibitors together for 72 h and annexin V binding as a measurement of apoptosis was determined.Statistically significant differences following one-way ANOVA with multiple comparisons are indicated (error bars represent SD, and p-values are indicated by: * = p < .05,** = p < .01,and **** = p < .0001).(G) SET2 cells and (H) UKE1 cells were treated with ruxolitinib (0.025 μM for SET2 cells, 0.01 μM for UKE1 cells) and RMC-4550 (0.025 μM for SET2 cells, 0.05 μM for UKE1 cells) alone and in combination for 4 h and cell lysates were prepared and immunoblotted for pERK, pRSK3, pSTAT5, and vinculin or GAPDH, as indicated.[Color figure can be viewed at wileyonlinelibrary.com] suppression of pRSK3, a downstream target of ERK signaling, was observed (Figures1G,H, and S1F).There was no evidence that treatment with both inhibitors together further suppressed pSTAT5, suggesting SHP2 inhibition in combination with ruxolitinib selectively enhanced suppression of activation of ERK/RSK3 downstream of JAK2 (Figures1G,H, and S1F).
). Erythroid and CFU-GM colony formation of cells from a patient with a MPL mutation were also inhibited by RMC-4550, and as with cells from JAK2-V617F patients, synergy with ruxolitinib was observed (FigureS5F,G).Colony formation of cells from CALR+ MPN patients were also sensitive to RMC-4550, which further enhanced colony suppression seen with ruxolitinib (FigureS5H,I).To determine if inhibitor treatment in colony forming assays presented any selectivity for JAK2-V617F expressing cells, we performed colony assays (using medium containing erythropoietin among other required cytokines) (FigureS6A) and used two assays to determine the mutant allele fraction of pooled colonies obtained from plating cells from each of four MPN patients.Both digital PCR detection (FigureS6B) and deep sequencing of JAK2-V617-encoding amplicons (FigureS6C) indicated there was an evident decrease in the JAK2-V617F allele fraction with the combination of RMC-4550 and ruxolitinib in one patient and with RMC-4550 alone in another.Finally, we also treated cells fromF I G U R E 2 SHP2inhibition suppresses growth of primary MPN cells and exhibits synergy with ruxolitinib.(A) Mononuclear cells were isolated from MPN patients (see Figure S4 for patient and sample details) and plated in colony forming assays in the presence of the indicated concentrations of RMC-4550 (n = 8) and EEC and CFU-GM colony formation were assessed and shown relative to DMSO treated samples (100%, dashed line).The numbers within the bars indicate how many patient samples contributed to the data for that concentration.Statistically significant differences following one-way ANOVA with multiple comparisons comparing each concentration to the control DMSO treatment (100%, dashed line) are indicated (error bars represent SEM, and p-values are indicated by: * = p < .05,** = p < .01,*** = p < .001,and **** = p < .0001).(B) Individual patient sample EEC and CFU-GM formation from (A) are shown.Statistically significant differences following oneway ANOVA with multiple comparisons comparing each concentration to the control DMSO treatment of each individual sample (100%, dashed line) are indicated (error bars represent SD, and p-values are as in (A).(C) CD34+ cells selected from bone marrow mononuclear cells of a MF patient (MPN17) were treated with 1 μM RMC-4550 for 16 h and resulting cell lysates were immunoblotted for pERK, pRSK3, total ERK, and GAPDH.(D) EEC and (E) CFU-GM formation of patient samples were assessed with ruxolitinib (blue bars), RMC-4550 (red bars), and the combination of ruxolitinib and RMC-4550 (purple bars).Statistically significant differences are shown following one-way ANOVA with multiple comparisons comparing each combination treatment to single agent treatments of individual patient samples used (data are normalized to the control DMSO treatment of each individual sample, error bars represent SD) and p-values are indicated as in (A).(F) Colony formation data from (D) and (E) were analyzed for synergy using the Bliss independence model and corresponding synergy scores are presented in a heat map where rows represent individual patient samples.[Color figure can be viewed at wileyonlinelibrary.com] healthy individuals for comparison and observed variable inhibition of colony numbers and erythroid colony size with RMC-4550 which was also evident in combination with ruxolitinib (Figure S7), consistent with the inhibition observed with the combination of ERK inhibition and ruxolitinib previously reported in healthy cell colony assays. 21F I G U R E 3 Legend on next page.

3. 4 |
The SHP2 inhibitor RMC-4550 antagonizes MPN phenotypes in the MPL-W515L mouse model of MPN The SHP2 inhibitor RMC-4550 suppresses MPL-W515L-driven MPN phenotypes.Lethally irradiated C57Bl/6 mice were transplanted with bone marrow cells transduced with retrovirus designed to express MPL-WT or MPL-W515L as well as GFP.Disease was monitored by CBCs, and equal cohorts based on leukocytosis were generated 20 days after transplant.Cohorts transplanted with MPL-W515L transduced bone marrow cells were treated with vehicle (n = 5 mice), RMC-4550 (10 mg/kg, qd-once per day) (n = 4), and RMC-4550 (30 mg/kg, qd-once per day) (n = 5).Mice transplanted with bone marrow transduced to express MPL-WT were treated with vehicle as a control for disease development.(A) WBC and (B) platelet counts of mice before and after 7 days of treatment (Tx).(C) Liver and spleen weights normalized to whole body weights and whole body weights of mice after treatment are shown.(D) Images of representative livers and spleens of each cohort after treatment are shown.(E) The % of CD11b + GFP+ cells present in peripheral blood following treatment are shown.Statistically significant differences were determined following one-way ANOVA with multiple comparisons comparing each treatment to vehicle treatment (** = p < .01,*** = p < .001,and ns = no significant difference).Box and whisker plots: boxes represent the 25th to the 75th percentile, the whiskers represent the minimum and maximum value, and the line designates the median of the data.(F) MPL-W515L mice were treated with a single dose of vehicle (n = 2 mice) or RMC-4550 (20 mg/kg) (n = 2 mice) for 6 h before total bone marrow cells and splenocytes were isolated and immunoblotted for pERK, ERK, pRSK3, and GAPDH, as indicated.[Color figure can be viewed at wileyonlinelibrary.com]F I G U R E 4 Legend on next page.thepresence of extramedullary hematopoiesis in the liver, which was not significantly altered by RMC-4550 or ruxolitinib alone (Figure4F,G).Interestingly, however, the combination significantly antagonized extramedullary hematopoiesis in the liver compared to vehicle or monotherapy treated mice (Figure4F,G).Megakaryocyte hyperplasia with cytologic atypia, syncytial clustering, typically seen in MPN, was evident in all tissues (Figures4G and S8B).While there was no inhibition of megakaryocyte hyperplasia by ruxolitinib, there was less megakaryocyte hyperplasia in liver and a trend toward less in the spleen following RMC-4550 treatment, and the combination of RMC-4550 and ruxolitinib led to reduced megakaryocytes in the spleen and liver compared to vehicle as well as ruxolitinib monotherapy

4
The addition of the SHP2 inhibitor RMC-4550 to ruxolitinib provides superior activity against MPN phenotypes in an MPL-W515L-driven mouse model.The MPL-W515L transplant model in Balb/c mice was used to assess the combination of RMC-4550 (20 mg/kg, qod-once every other day) and ruxolitinib (30 mg/kg-once per day) (n = 9) compared to single agent RMC-4550 (n = 9) and ruxolitinib (n = 10) using the same dosing strategy, and mice treated with vehicle alone (n = 10).Cohorts normalized for elevated WBC levels were created and treatment was initiated 13 days after transplant.(A) WBC counts and (B) platelet counts prior to treatment (Day 0) and after the 7th and 14th days of treatment are shown.The rapid demise of vehicle treated mice led to CBCs being performed on the 9th day of treatment for this cohort, in lieu of the planned endpoint after 14 days of treatment, but data are shown with the Day 14 data for comparison.Five mice were taken from each cohort for pathology after 14 days of treatment and the remaining animals were left for survival analysis with treatment terminated after 24 days of treatment.(C) Liver weights normalized to body weight for each treatment cohort (n = 5 for each cohort) (±SD).(D) The percent of GFP positive cells in peripheral blood for each cohort after 14 days of treatment (vehicle n = 5, RMC-4550 n = 9, ruxolitinib n = 10, and RMC-4550 + ruxolitinib n = 9).Box and whisker plots: the box represents the 25th to the 75th percentile, the whiskers represent the minimum and maximum value, and the lines designate the median of the data.(E) Animal survival analyzed by Kaplan-Meier is shown.(F) The percent of liver architecture replaced by extramedullary hematopoietic cells (±SD) and (G) representative images of H&E staining of liver are shown.(H) Megakaryocytes per field of H&E-stained spleen and liver are shown after 14 days of the indicated treatments, or the demise of the vehicle treated animals (±SD).(I) TNFα levels in serum following 14 days of treatment are shown (±SD).Statistically significant differences following oneway ANOVA with multiple comparisons are indicated (* = p < .05,** = p < .01,*** = p < .001,**** = p < .0001,and ns = no significant difference).[Color figure can be viewed at wileyonlinelibrary.com] for classical MPNs.While JAK2-V617F MPN cell models exhibited sensitivity to SHP2 inhibition by two allosteric SHP2 inhibitors (RMC-4550 and SHP099), the addition of SHP2 inhibition to the JAK2 inhibitor ruxolitinib led to profound suppression of cell growth, due at least in part to enhanced induction of apoptosis.ERK activity is dependent on JAK2 activity in these cells and SHP2 inhibition led to inhibition of ERK activation, which places SHP2 downstream of JAK2 and upstream of ERK.Indeed, ruxolitinib treatment of MPN cells led to a decrease of activating phosphorylation of SHP2, suggesting SHP2 is regulated in a JAK-dependent manner.Stronger inhibition of RAS/ERK signaling could be obtained with concomitant SHP2/JAK2 inhibition.Not only did SHP2 inhibition prevent the development of ruxolitinib persistent growth of MPN model cells, but the growth and viability of established ruxolitinib persistent cell lines remained sensitive to SHP2 inhibition and SHP2 activity was still required to mediate ERK activation in the context of persistence.Notably, SHP2 inhibition antagonized colony formation of cells from MPN patients, including both neoplastic EECs and CFU-GMs, and synergized with ruxolitinib.SHP2 inhibition via RMC-4550 antagonized multiple phenotypes of MPL-W515L-driven MPN including leukocytosis, organomegaly, extramedullary hematopoiesis, and allele burden, and enhanced survival of MPN mice.The therapeutic combination of RMC-4550 with ruxolitinib led to enhanced suppression of numerous MPN phenotypes in mice compared to vehicle and monotherapies.All treatments antagonized the advancing leukocytosis observed in vehicle mice and the combination therapy abrogated pretreatment levels of leukocytosis.In addition, compared to vehicle and monotherapy treatments, the combination of RMC-4550 and ruxolitinib provided superior activity against extramedullary hematopoiesis in the liver, as well as megakaryocytic hyperplasia with atypia in the liver and spleen, and significantly increased survival compared to vehicle and ruxolitinib monotherapy.It is possible deeper anti-MPN responses may have Importantly, RMC-4550 (at 10 μM) did not inhibit any of 468 kinases tested,45 suggesting allosteric SHP2 inhibitors may antagonize cell signaling with limited off target effects compared to MEK/ERK kinase inhibition.This may provide a unique safety and tolerability profile compared to MEK or ERK kinase inhibitors, which could also be reflective of the nature of SHP2 inhibitors acting upstream, instead of downstream, of RAS activation.Of note, treatment of MPL-W515L MPN mice with an ERK inhibitor in combination with ruxolitinib did not enhance survival due to exacerbated thrombocytopenia,21 which we did not observe with SHP2 inhibitor therapy.SHP2 inhibitors (e.g., RMC-4630, JAB-3068, TNO155, BBP-398, and PF-07284892, among others) are being evaluated in clinical trials, mostly in the RAS-driven solid tumor space (e.g., NSCLC, pancreatic cancer).The findings from such studies could assist in rapid translation of clinical studies to test SHP2 inhibitors in MPN patients.Inhibition of SHP2 may antagonize disease in both MPN cell autonomous and non-autonomous manners.Inflammatory factors including IL-1,