The burden of MPN disease
Myeloproliferative Neoplasms (MPNs) are a group of rare but incurable blood cancers where the bone marrow produces too many myeloid cells—specifically red blood cells, platelets, and some types of white blood cells. While these cancers are typically slow growing, they result in chronic, debilitating symptoms that can drastically impact a patient’s quality of life and require life-long treatment to manage. Beyond the daily burden of distressing symptoms, MPNs carry the unpredictable and ever-present risk of progression to more aggressive, lethal forms, such as Acute Myeloid Leukaemia (AML), or the development of irreversible bone marrow scarring, known as fibrosis.
The only curative therapy currently available for MPNs is allogeneic stem cell transplantation. However, this is usually only offered to a limited proportion of Myelofibrotic patients with advanced-risk disease and no prohibitive comorbidities. Consequently, the treatment of MPNs remains a field of significant unmet clinical need.
The JAK2 breakthrough and the limits of current therapeutics
MPNs are broadly categorized by the presence or absence of the ‘Philadelphia chromosome’ genomic rearrangement. The ‘Philadelphia-negative’ group comprises three ‘classic’ disorders:
- Polycythemia Vera (PV): An overproduction of red blood cells that thickens the blood.
- Essential Thrombocytopenia (ET): An overproduction of platelets, increasing the risk of blood clots.
- Primary Myelofibrosis (PMF): Primary scarring in the bone marrow, often causing anaemia and a severely enlarged spleen.
The discovery of the JAK2V617F mutation in 2005 was a breakthrough in understanding these diseases. Mutations triggering an overactive JAK-STAT pathway are a unifying feature of MPNs, but the specific JAK2V617F mutation is found in ~95% of patients with PV and 50–60% of those with ET and PMF, making it is the most common disease driving mutation in Philadelphia chromosome negative MPNs. In healthy individuals, the JAK2 protein acts like a ‘switch,’ telling blood cells when to grow based on external signals. The V617F mutation swaps the amino acid valine (V) for phenylalanine (F) at position 617, jamming the switch in the ‘on’ position.
Since 2011, the development of small-molecule JAK inhibitors like Ruxolitinib has revolutionised treatment. However, while Ruxolitinib is an effective tool in symptom management, it rarely eradicates the disease and can carry significant side effects. There is an urgent need for therapeutics that selectively target the aberrant effects of the V617F mutation while preserving canonical JAK2 signalling. Thanks to its recently elucidated role in JAK inhibitor resistance, one of the most promising emerging strategies is the ‘one-two punch’ of JAK2 and PIM1 inhibition. Recent pre-clinical and clinical trials of the potent PIM1 inhibitor TP-3654 show its effectiveness in reducing spleen size and potentially stabilising or reducing fibrosis in the bone marrow, whilst in combinatorial therapy trials it has been shown to be a highly successful therapy in treating relapsed or refractory patients.
Dr Jekyll and Mr Hyde: which HIF are we dealing with?
Our research began by looking at an unexpected player in this axis: Hypoxia Inducible Factor 1 (HIF-1). In a healthy system, HIF-1 is the key protein involved in sensing and responding to low cellular oxygen (hypoxia). It is normally stabilised by hypoxia, allowing it to move to the nucleus and initiate the transcription of genes that help the cell to adapt to and survive periods of low oxygen. As a master regulator of transcription – HIF-1 has over 500 experimentally validated gene targets. In the majority of cancers, both solid and blood malignancies, HIF-1 overexpression is oncogenic. However, in myeloid cancers, HIF-1 can be either oncogenic or tumour suppressive. The mechanisms that direct these opposing functions of HIF-1 are not well understood. It was recently discovered that HIF-1 is activated in JAK2V617F-positive MPNs even under normal oxygen conditions (normoxia).
We were therefore interested in investigating how the JAK2V617F mutation was overriding HIF-1’s dependency on oxygen, what effect this was having on its behaviour, and whether this disease-associated activation was contributing to MPN disease pathology.
A rewired regulon
We started by looking at how the primary subunit, HIF-1α, bound to the chromatin and what cofactors were associating with it. By performing chromatin-immunoprecipitation (ChIP) in paired leukaemic cell lines, we found that the repertoire of HIF-1 regulated genes was significantly reduced in the presence of the JAK2V617F mutation and that the ‘order of priority’ in which they were bound by HIF-1α was significantly altered. Interestingly, when these mutant cells were additionally exposed to hypoxia, the typical stabilisation of HIF-1 was "blunted"—the mutation seemed to have hijacked the pathway, preventing the protein from responding to its canonical triggers.
Using a publicly available ATAC-seq dataset, we confirmed that these changes in binding were not due to alterations in chromatin accessibility but were instead a result of how HIF-1 itself was being directed. Furthermore, RIME mass spectrometry revealed that while more HIF-1 was present at the chromatin in mutant cells, it was actually less functionally able to bind to DNA in the traditional way. Instead, it was associated with a significantly altered panel of cofactors. This included a reduction in markers of degradation and an enrichment of RNA-binding and spliceosomal factors—suggesting that HIF-1 activated by JAK2V617F may play a role in aberrant splicing, another hallmark of MPN disease.
Linking gene signatures to patient survival
To understand the clinical impact, we generated a series of gene signatures from our ChIP-seq data and performed Gene Ontology (GO) analysis. The genes preferentially bound by HIF-1 in the presence of the JAK2 mutation were associated with DNA metabolism, the DNA damage response (DDR), and cell cycle progression, and shifted away from the classical hypoxia-responsive HIF-1 pathways that support survival in hypoxic conditions, for example, glycolysis.
When we applied these signatures to RNA-seq data from MPN patients, the results were striking. While canonical hypoxic HIF-1 signalling had no association with survival, the HIF-1 signature activated by the JAK2V617F mutation showed a significant correlation with reduced patient survival.
We further refined this by performing an integrated analysis of several ‘omics’ datasets across specific cell populations. This revealed that some of these signature genes were specifically differentially expressed in the megakaryocyte progenitor populations (MkP/MEP) responsible for driving MPN disease. Furthermore, only the genes expressed in these specific populations were individually associated with worsened disease severity and survival.
The path to leukaemic transformation
The most dangerous potential transition in MPN disease is the transformation to blast-phase (MPN-BP), or AML. We applied our integrated omics analysis to 11 patients who transitioned from chronic MPN to MPN-BP. We identified a sub-signature of genes significantly associated with this transformation. High enrichment of this "HIF1-MPN-BP" gene signature showed a strong and significant association with disease severity, worsened survival, and key disease parameters like fibrosis. By comparison, canonical hypoxia genes were not associated with transformation at all, suggesting that this "rewired" HIF-1 activity is a primary driver of disease progression.
Identifying the switch: The PIM1 mechanism
Having established HIF-1’s relevance, we sought to elucidate the mechanism. Phosphorylation is a key regulator of HIF-1’s transcription activity. Using phosphoproteomics, we discovered two novel phosphorylation sites (T498 and S500) on HIF-1α that were exclusive to JAK2V617F cells. These sites fall within the domain associated with HIF-1’s normal normoxic degradation, suggesting that these phosphorylation events may physically prevent the protein from being broken down in normoxia. We also identified different co-factors associated with HIF-1a in JAK2V617F cells that may be responsible for mediated the differential gene expression induced by HIF-1 in this disease context.
Through an unbiased, high-throughput phospho-flow cytometry approach, we traced the signalling cascades downstream of the JAK2V617F mutation to try and identify a potential responsible kinase. As opposed to global amplification of JAK-STAT signalling, we identified signalling of specific downstream cascades, notably hyperactivation of STAT1 and STAT5. By interrogating their target genes for known HIF-1α kinases, we identified the serine/threonine kinase PIM1, a known cancer-associated HIF-1 kinase.
When we treated JAK2V617F cells with selective PIM1 inhibitors, we observed a direct reduction in HIF-1α protein levels in normoxia. Whilst some mechanisms by which PIM1 inhibition works are well understood, such as inhibiting the ‘apoptosis-escaping’ effect that PIM1 can confer to cancer cells, inhibiting its activation of NF-kB-mediated cytokine storms and inhibiting PIM1-mediated MYC stabilisation, the full mechanism of PIM1s contribution to MPN pathogenesis has never been studied. Our findings confirmed that PIM1 is responsible for the aberrant stabilisation of HIF-1α, potentially offering an additional explanation of why PIM1 inhibition has shown such promise in MPN clinical trials.
A more selective future
In our final tests, we confirmed that PIM1 inhibition was less disruptive to global signalling than Ruxolitinib and selectively killed JAK2V617F cells while sparing wildtype cells. While Ruxolitinib induced changes in both healthy and mutant cells, PIM1 inhibition specifically rescued the aberrant gene expression driven by the mutation.
If we can target the aberrant phosphorylation of HIF-1α itself, we may find an additional way to treat MPNs that conserves healthy JAK2, PIM1 and canonical HIF-1 signalling. Not only does this research present the potential to selectively target HIF-1 in a disease setting in a way that spares normal hypoxic activity, it also offers a potential roadmap for the first therapeutic strategy aimed at reducing the likelihood of AML transformation and halting the progression of myelofibrosis, providing hope for patients who currently have few curative options.