Department of Translational Hematology and Oncology Research

The disseminated malignancy with the best overall survival today is acute promyelocytic leukemia (APL), a rare genetic subtype of de novo acute myeloid leukemia (AML), which accounts for less than 10% of AML incidence. It is treated not with conventional cytotoxic chemotherapy but with just 2 drugs: all-trans retinoic acid and arsenic trioxide. These unblock the myeloid differentiation arrest that defines all AMLs and release the APL cells to terminal granulocyte lineage fates. Given this inspiring example of what is clinically possible, an enduring question has been whether noncytotoxic differentiation-restoring therapy can be extended to the most common genetic subtypes of AML—eg, those containing mutated nucleophosmin (

), which amount to approximately 30% of AML cancers.

The master transcription factor expression pattern of AML cells suggests why this should be achievable. A few of the roughly 100 transcription factors expressed in cells are master regulators, collaborating in couplets or triplets to powerfully determine cell fates and functions, as illustrated by their remarkable capacity to convert cells of one lineage into another—even into embryonic stem cells.

The master regulators highly expressed in all AMLs are PU.1, CEBPA, and RUNX1, which usually collaborate to command terminal granulocyte or monocyte lineage fates (

 Clearly, AML cells evade such terminal lineage fates, but how do they defy these master regulators, and can compliance be restored?

 Historically, investigators have known that translocations of 

 cause cytoplasmic dislocation of the NPM1 protein, but why or how this should transform myeloid cells was unknown. The Cleveland Clinic group performed the first mass-spectrometric analyses of protein—protein interactions of endogenous NPM1 affinity purified from wild-type and 

-mutated AML cell nuclear and cytoplasmic fractions. They found that both wild-type and mutant NPM1 interact with PU.1 and that mutant NPM1 dislocates PU.1 into the cytoplasm.

CEBPA and RUNX1—master transcription factors that collaborate with PU.1 to activate granulomonocytic lineage fates&mdash;remained nuclear; but without PU.1, their coregulator interactions were toggled from coactivators to corepressors, suppressing instead of activating more than 500 granulocyte and monocyte terminal-differentiation genes. The PU.1 dislocation also explained why the AML cells expressed high levels of key precursor genes&mdash;eg, 

—since these are usually repressed by PU.1-helmed forward myeloid differentiation.

Protein macromolecules such as NPM1 require transport factors to enter (importins) and exit (exportins) the nucleus. Use of selinexor, an inhibitor of XPO1-mediated nuclear export, locked mutant NPM1/PU.1 in the nucleus and activated terminal monocytic fates. The same treatment did not induce differentiation of 

Investigators then queried whether the master transcription factors CEBPA and RUNX1, which remained nuclear in 

-mutated AML cells, could be used for a complementary approach to differentiation restoration. CEBPA and RUNX1 are supposed to activate granulocytic fates; however, the granulocyte differentiation program, like the monocyte differentiation program, is suppressed in AML. To investigate how, investigators examined the coregulator interactions of nuclear CEBPA and RUNX1 in 

-mutated AML cells using affinity purification—mass spectrometry/Western blotting. The CEBPA and RUNX1 protein interactomes were enriched for coregulators that repress transcription (eg, DNMT1, NURD) over coactivators that activate genes (eg, SWI/SNF, NUA4). PU.1 introduction into the CEBPA/RUNX1 interactomes by selinexor switched CEBPA and RUNX1 interaction in 

- mutated AML cells from corepressors to coactivators. This outcome reinforced previous findings from the group regarding how PU.1 and RUNX1 collaboratively exchange corepressors for coactivators. This occurred through interactions between their transcriptionregulating domains.

Moreover, use of the clinical small molecule decitabine (Dacogen) to directly deplete the corepressor DNMT1 from the CEBPA/RUNX1 interactomes also reconfigured CEBPA/RUNX1 interaction from corepressors to coactivators, which activated terminal granulocytic fates. Interestingly, as per normal granulocytic differentiation, this led to a natural downregulation of NPM1 (including mutant 

In a patient-derived xenotransplant model of poor-prognosis dual 

-mutated AML, with treatment initiated after confirmation of bone marrow engraftment to greater than 20% AML cells, combination noncytotoxic therapy with pharmacodynamically directed dosing of selinexor and DNMT1-depleting drugs extended time-to-distress by more than 6 months over vehicle treatment. The noncytotoxic mechanism of action was highlighted by the preservation of normal blood counts during more than 8 months of active therapy with the regimen. The 

-mutated AML cells that resisted and progressed through several months of in vivo therapy to cause distress in the mice did so by avoiding both nuclear retention of mutant NPM1/PU.1 by selinexor and DNMT1-depletion by decitabine/ 5-azacytidine. That is, resistance both in vitro and in vivo was by prevention of intended molecular pharmacodynamic effects, underscoring the importance of the targeted pathways to the malignant phenotype. A summary of the data is provided (

 PU.1, CEBPA, and RUNX1 have been shown to promote exponential replication kinetics by binding to 

 enhancers to produce high-grade activation of 

 (the master transcription factor coordinator of cell proliferation) and cobinding with MYC at its target genes. This contrasts with the quiescence imposed by stem cell master transcription factors such as HLF in hematopoietic stem cells. In computational analyses, such skewing of intrinsic replication rates logarithmically favors decoupling replication from forward differentiation in lineage progenitors as the most efficient strategy for malignant transformation. The results demonstrated how mutant 

 executes such decoupling in lineage progenitors and showed how the machinery could be recoupled.

The investigators do not know for sure why neoplastic evolution selects to mutate 

 directly, but possible reasons include that mutant NPM1 creates the graded PU.1 loss of function (more than haploinsufficiency, less than total) that has been shown to be necessary for leukemogenesis. It’s also possible that dislocation of proteins other than PU.1 contributes to transformation. The leukemogenic suspension of cells at an intermediate, inherently replicative stage of their advance along lineage-differentiation axes seems to hinge on partial but not complete loss of function in the PU.1/ CEBPA/RUNX1 circuit. This is because, experimentally, complete inactivation of 

 kills AML cells, even as partial loss of function of any 1 of these is leukemogenic; accordingly, 

mutations, though highly recurrent in AML, are mutually exclusive.

This delicate balancing act of all AMLs, relying on commanders of lineage differentiation and being poised close to the edge of terminal myeloid differentiation, is what the investigators fundamentally exploit with their evaluated treatments. These results add to a body of data suggesting that corepressor-coactivator imbalance in the PU.1/CEBPA/RUNX1 master transcription factor hub is a common final pathway that represses terminal differentiation. This explains the meaningful clinical activity of noncytotoxic DNMT1 depletion by decitabine (or its prodrug 5-azacytidine) in patients with myeloid malignancies containing sundry mutations and translocations.4

 disrupts the PU.1/CEBPA/ RUNX1 master transcription factor hub to stall myeloid differentiation at an inherently proliferative stage of myeloid lineage differentiation. By dissecting this final common pathway to myeloid differentiation arrest and leukemic hematopoiesis, the investigators identified and then validated specific molecular targets for noncytotoxic (p53-independent, normal stem cell sparing), differentiation-restoring therapy of the largest genetic subtypes of AML. Because the treatments have defined and measurable molecular pharmacodynamic effects, dosages of candidate therapeutics can be selected rationally instead of empirically, and measurements for achievement of intended molecular pharmacodynamic effects can facilitate rational investigation of mechanisms of resistance. In sum, the discoveries regarding mutant NPM1 can guide efforts to extend noncyotoxic (nontoxic) differentiation-restoring treatments to the most common genetic subset of AML.

Gu X, Ebrahem Q, Mahfouz RZ, et al. Leukemogenic nucleophosmin mutation disrupts the transcription factor hub regulating granulomonocytic fates. 

. 2018;128(10):4260-4279. doi: 10.1172/JCI97117.

Rapin N, Bagger FO, Jendholm J, et al. Comparing cancer vs normal gene expression profiles identifies new disease entities and common transcriptional programs in AML patients. 

. 2014;123(6):894-904. doi: 10.1182/blood-2013-02-485771.

Bagger FO, Sasivarevic D, Sohi SH, et al. BloodSpot: a database of gene expression profiles and transcriptional programs for healthy and malignant haematopoiesis. 

. 2016;44(D1):D917-924. doi: 10.1093/nar/gkv1101.

Velcheti V, Schrump D, Saunthararajah Y. Ultimate precision: targeting cancer but not normal self-replication. 

. 2018;(38):950-963. doi: 10.1200/EDBK_199753.

Articles

Acute promyelocytic leukemia, a rare genetic subtype of de novo acute myeloid leukemia is treated not with conventional cytotoxic chemotherapy but with just 2 drugs: all-trans retinoic acid and arsenic trioxide. Given this, an enduring question has been whether noncytotoxic differentiation-restoring therapy can be extended to the most common genetic subtypes of AML.

Novel AML Strategies Aim to Restore Regulatory Gene Functions

	Co-Leader, Developmental Therapeutics Program

Today, only three types of nonresectable adult cancers are routinely cured: some types of lymphoma and myeloid leukemia, and testicular cancer 

. This overall picture has not changed substantially in the past 20 years, an indication of the extraordinarily high failure rate of drug development in oncology. The costs of this failure rate shifts to patients and their families, contributing to the financial complications of a cancer diagnosis.

One of the central reasons that curing cancer has been so problematic is the dysfunction of 

, the single most common genetic alteration in cancer. At the Cleveland Clinic, we have been studying ways to overcome the problem of aberrant 

 activity safely and effectively in several hard-to-treat solid tumors, including through the development of a promising experimental new drug.

Through this research focus, we can answer some of the most basic questions in cancer: Why can we cure some cancers but not others? Why are treatments so arduous? Why does oncology drug development have a 95% failure rate?

These are related questions, with related answers. Most importantly, the conceptual leap needed to render these questions obsolete is laid out clearly before us—for all the scientific complexity appended to cancer, its essence is simple: cells that grow and divide and do not stop.

For cells to grow and divide is completely normal; this defines life. But, abnormal (cancer) is when such growth and division does not cease. Why do the billions of healthy cells dividing in our body cease and desist? The main, natural reason is that the dividing cells arrive at their intended lineage-differentiation fate. Another reason it ceases is that an emergency brake—apoptosis&mdash;activates when damage or stress of some sort is sensed, requiring a delay for repair or orderly self destruction.

		Figure. Global Cancer Incidence and Mortality Rates

		World Health Organization. Global cancer incidence and mortality rates/100,000. GLOBOCAN 2012: estimated cancer incidence, mortality and prevalence worldwide in 2012. http://globocan.iarc.fr. Published December 12, 2013. Accessed Octover 12, 2015.

Cancers, to be cancers, must avoid these natural brakes... but, how? Apoptosis is a complex cellular program executed by the master transcription factor p53. Cancer cells, which by their very nature are damaged and/or stressed, frequently physically remove or inactivate the gene for p53 (

Radiation and the vast majority of current cancer drugs have the intent of inducing apoptosis (cytotoxicity). This is hard to do if p53 is missing.

Meanwhile, normal dividing cells with intact p53 readily undergo apoptosis in response to our severe stressful treatments. In fact, the absence of p53-system alterations distinguishes the disseminated cancers we routinely cure from the ones we do not. Thus, it is a fundamental therapeutic deficiency that we have hundreds of drugs that intend apoptosis, but no drugs that broadly intend differentiation. Why? It is relatively easy to develop treatments that stress or poison cells and turn on the emergency brake in the test tube; this concept for treating cancer was established decades before we knew about p53.

Moreover, at that time we did not understand the complex mechanisms of cell differentiation, and how cancers avoided this routine mode of cell cycle exit.

The good news is that now we understand how cancer cells avoid everyday differentiation-mediated cell cycle exits: cancer “stem” cells are actually committed progenitors, expressing supra-physiologic levels of master transcription factors of lineage-differentiation, most likely because this is needed for high-grade activation and stabilization of MYC, the ancient master transcription factor coordinator of cell growth and division.

To avoid terminal differentiation despite lineage-commitment, cancers genetically inactivate genes for selected coactivators or “On” enzymes (eg, SMARCA4, PBRM1, KMT2A), resulting in aberrant epigenetic repression instead of the activation of hundreds of late-differentiation and proliferation-terminating genes. This epigenetic repression is mediated by the unbalanced action of corepressors or “Off”’ enzymes.

Pharmacologic inhibition of these “Off” enzymes (eg, DNMT1) rebalances toward residual “On” enzymes and restores the intended lineage-differentiation fate of cancer cells, terminating cell growth and division physiologically, without need for p53, and without killing any normal cells. Inhibition of “Off” enzymes triggers differentiation-mediated cell cycle exits across the cancer histological and genetic spectrum, indicating the general nature of this cancer strategy, just as p53-system attenuation is cancer histology agnostic as a stratagem to avoid apoptosis. Crucially, this science is translatable in the near-term (only lack of funding stands in the way!), since enzymes such as DNMT1 can, in principle, be drugged by the generic agents decitabine and 5-azacytidine. Decitabine and 5-azacytidine, however, have terrible pharmacology, with trivial penetration into solid tissues and very brief plasma half-lives. Also, these drugs have not been explicitly applied in a way to avoid apoptosis (to avoid toxicity) and maximize DNMT1-depletion (to increase p53-independent cell cycle exits of cancer cells). Cleveland Clinic is developing a pharmacologically rational, orally bioavailable version of decitabine (oral THU-decitabine) with a multihour half-life and pan-tissue distribution, that does penetrate solid tissues. This agent has completed phase I clinical trials, which proved the pharmacologic premise, and we hope to begin clinical trials in refractory/ relapsed metastatic pancreatic and lung cancers before the end of the year.

At this time, there is only one cancer, a rare type of acute myeloid leukemia called acute promyelocytic leukemia, that is routinely treated with differentiation- intending treatment. In fact, switching from apoptosis to differentiation as the method of treatment converted this leukemia from the worst to the best in terms of overall survival.

Scientifically, there is no reason this route to durable, nontoxic cancer control should be restricted to only this rare form of leukemia.

One of the central reasons that curing cancer has
been so problematic is the dysfunction of TP53, the
single most common genetic alteration in cancer.

Yogen Saunthararajah, MD

Articles

Novel AML Strategies Aim to Restore Regulatory Gene Functions

TP53: The Elephant in the Precision Oncology Room