ASH 2013 preview: Myeloproliferative Neoplasms, New Targets

November 25, 2013. by P.D. Rennert (I keep forgetting to sign these entries)
Part 6b. Myelofibrosis and related Myeloproliferative Neoplasms: Phase 1 and preclinical updates from ASH 2013.

The American Society of Hematology Meeting will take place in New Orleans, December 7 – 10, 2013. The abstracts are available at
Back to it. In part 6a we looked at the phase 2/3 studies, and these are certainly focused on Jak inhibition, for reasons we discussed in part. Its worth noting that Jaks are a class of signaling proteins that hybridize shamelessly and because of this blocking any given Jak (Jak1, Jak2, Jak3, Tyk2) will necessarily block the activity of others in the context of signaling that requires hybridization. Conversely it has been demonstrated that blocking any given Jak may not be effective in settings where other family members can substitute. As always, it is complicated.
It has been mentioned that Jak inhibitors are not considered disease modifying or curative. They offer symptomatic relief, which relieve the pathological degradation of the bone marrow, and can do so for extended periods, as has been demonstrated with ruxolitinib. In this sense the treatment paradigm is not so different from some other targeted cancer therapies, where the name of the game is to keep the wolf from the door, at least for a time.
Regardless of the limited potential, Jak inhibitor development continues at a brisk pace, as demonstrated in the previous section. In earlier development we find yet more Jak inhibitors. Gilead will present data from the phase 1/2 trial (with extension) of the Jak1/2-selective inhibitor, momelotinib (Abstract #108). Confusingly, this trial used some new (to me) endpoints. One was a reduction in “palpable splenomegaly” which suggests the spleen was not imaged (?). Second (and useful I think) was “transfusion independence”. I like this because it clinically measures the impact of treatment on the most pathologically relevant endpoint, loss of RBCs and platelets in circulation. What is even more interesting about this trial is the extension data, showing a “median spleen response” of 324 days, with pretty huge variance. As noted with ruxolitinib, the variance cannot be traced to the presence or absence of Jak2 V617F mutation - I suspect Gilead will work to sort this out pretty quickly, so that bears watching unless of course the variance is due to the manner of measurement!
Early trial data will be reported for BMS911543, another Jak2-selective inhibitor (Abstract #664). Note that Jak3 and Tyk2 are hit with IC50s less than 100nM, so selectivity will depend on actual exposure, which is data not available yet. Anyway, the drug is given BID and the goal of the trial is to establish an MTD (not reached). Early efficacy data included analysis of responses of patients who had received a prior Jak inhibitor. Symptoms were well controlled and the spleen volume response (> 35% reduction in splenomegaly) was dose responsive – up to 70% of patients responded at the highest doses (note: small sample size). AEs were were the usual diarrhea, nausea and cytopenias. The investigators will update with data from 84 patients at the meeting.
Eli Lilly will present data from the LY2784544 phase 1 dose escalation trial (Abstract #665). Of note, this drug is described as being “selective” for the V617F mutant of Jak2. They are enrolling 38 patients (most MF, some PV) at 30 -300 mg per day, although at 200 mg and above they hit liver tox (grade 3 serum creatinine elevation) and the MTD is established at 120 mg/day. Other AEs are typical of the class and not severe. We are introduced to yet another way of measuring splenomegaly, this time as “spleen length”. There was a >50% improvement in 56% of the evaluable patients, furthermore, >50% of patients reported improved symptoms by TSS. An update is promised for the meeting.
Novel inhibitors are finally moving through preclincal and early clinical evaluation. Geron has a novel telomorase inhibitor, imetelstat, a lipid-conjugated oligonucleotide inhibitor of human telomerase (Abstract #662). Early data are impressive. AEs were modest, and responses (44%) included a fair number that appeared to be CRs or PRs. Here I quote: 
“The four (22%) CR patients experienced reversal of BM fibrosis and recovery of normal megakaryocyte morphology. Two CR patients were transfusion-dependent at baseline and became transfusion-independent … Among 13 patients with leukocytosis, 10 (77%) normalized their count or had >50% reduction. Eleven (61%) patients had complete or partial resolution of leukoerythroblastosis.”
Thats pretty impressive. The downside is that this was a small single center open label trial, so really, we just have to wait and see.
For an example of why such results can be so misleading, let look at a Phase 3 trial failure (Abstract #394). Here we have Celgene’s pomalidomide (approved under the name Pomalyst for the treatment of Multiple Myeloma). Phase 2 trials suggested that pomalidomide imporved severe anemia in MPN patients, including MF, PV and ET patients. The phase 3 trial enrolled 252 patients with MPN-associated MF who were dependent on RBC transfusions. There was no significant improvement in time to transfusion in drug-treated v placebo-treated cohorts.
So these early trials can be very misleading, but lets look even earlier, at some preclinical studies.
Novartis has some nice data showing elevation of Hedgehog signaling pathway targets in MF granulocytes (Abstract #666). Using in vitro and in vivo (mouse) models they demonstrate that Jak inhibition and Hedgehog inhibition have additive activity. The Jak inhibitor was INC424 and the Hedgehog inhibition was provided by Smoothened antagonism (sonidegib). Another novel approach is to target megakaryocytes, that contribute to MF pathogenesis. Using in vitro and in vivo models an academic collaboration (Abstract #109) shows that Aurora A kinase inhibition induces megakaryocyte arrest and this results in decreased bone marrow fibrosis, decreased infiltration of megakaryocytes and granulocytes into the liver and spleen, and decreased plasma TGFbeta, a potent pro-fibrotic growth factor.
Genetic analyses have demonstrated that certain signatures indicate risk for increased risk of leukemic conversion and decreased OS. These include ASXL1, EZH2, IDH1/2 and SRSF2 gene mutations (Abstract #104). EZH2 mutation accelerates the onset of primary MF (Abstract #110). However, patients carrying these mutations are not differentially impacted by ruxolitinib therapy, instead they have similar responses to patients without such mutations (Abstract #105). Novel pathways associated with MF include the arachidonate 5-lipoxygenase pathway (Abstract #111). An exciting new development is the identification of calreticulin mutations in patients that do not have Jak2 mutations, that is, the two appear to be mutually exclusive (Abtract #LBA1). This will spur investigation in novel drivers of MPN pathogenesis.
Obviously there is much more going on then can be covered here. The identification of the Jak2 mutation in MPN opened up a new era of drug development for these rare but nasty and lethal neoplasms. New findings will push us even further, and we can reasonably expect to see important advances over the next few years.