In the Clinic

One Molecule, Multiple Cancers: The Devil is in the Details

“Inside the beltway,” is a phrase normally reserved for discussions of careers in national politics, referring as it does to the highway that surrounds the Washington D.C. metropolitan area. However, Christina Annunziata, M.D., Ph.D., has also developed her career as a physician- scientist inside the beltway, first as a medical student and resident at Georgetown University and then rising through the ranks of CCR training opportunities to become a tenure-track Investigator. Over the years, Annunziata’s responsibilities have involved her in many NCI protocols, but her own research has remained firmly rooted in the family of transcription factors, NF-kB. As a student with Jeffrey Cossman, M.D., at Georgetown, she studied nuclear factor kappa B (NF-kB) signaling in Hodgkin’s disease. As a Medical Oncology Fellow with Louis Staudt, M.D., Ph.D., in CCR’s Metabolism Branch, she studied the role of NF-kB in multiple myeloma. Now her laboratory investigates the effects of NF-kB signaling in ovarian cancer. Annunziata believes that understanding the nuances of NF-kB function in distinct cell types could lead to effective pharmacological interventions for cancer.

Christina Annunziata, M.D., Ph.D.
Christina Annunziata, M.D., Ph.D. (Photo: E. Branson)

It was definitely the opportunity for strong and dedicated research in a clinical environment that drew me to NIH. I met Lou Staudt, when I was still a doctoral student, and it was a natural fit for me to continue my postdoctoral research in his laboratory. NF-kB is a very interesting family of molecules. It consists of five subunits that form various combinations of dimers capable of coordinating the expression of multiple genes. It is present as an inactive form in the cytoplasm of most cells, where a variety of external cellular signals can prompt its rapid separation from a molecular complex and migration to the nucleus to modulate gene expression. NF-kB signaling is important for many different cell types, however, the pathway functions differently according to cell type.

In multiple myeloma, we found that NF-kB signaling was turned on in most of the tumors we studied. In many cases, that may have been the result of influences from the surrounding tissue—the tumor microenvironment—but in some cases, the tumors had autonomous aberrant NF-kB activity. However, we didn’t identify just one specific mutation. We found that there were multiple points throughout the pathway that were dysregulated in different myeloma subtypes. In one case, a protein might be amplified, in another case a negative regulator might be lost. We did these analyses by first looking at changes in gene expression, but in many cases, validated our findings by looking at protein levels. We believe that one therapeutic action of the protease inhibitor, bortezomib, now approved for the treatment of multiple myeloma, may be to inhibit NF-kB signaling.

NF-kB Signaling in Ovarian Cancer

The NF-kB signaling pathway can be activated by a variety of receptors on the surface of cells to activate transcriptional networks in the nucleus. Several signaling components within the cell regulate NF-kB activity.
The NF-kB signaling pathway can be activated by a variety of receptors on the surface of cells to activate transcriptional networks in the nucleus. Several signaling components within the cell regulate NF-kB activity. (Image: C. Annunziata)

While working in the Staudt laboratory, I was performing clinical duties one day per week with Elise Kohn, M.D., in CCR’s Medical Oncology Branch (See “Ovarian Cancer: A Silent Killer “Speaks” through Proteins,” CCR connections Vol. 2, No. 2). I began seeing patients with ovarian cancer, and I began to wonder whether NF- kB signaling might also play a role in this disease. I could not find much in the scientific literature that addressed this question, and I wanted to move my research more towards helping the patients I was actually seeing. So, when I began my research program as an Assistant Clinical Investigator, I began to study NF-kB signaling in ovarian cancer (See “CCR’s Clinical Investigator Development Program,” Page 31).

My laboratory used a small- molecule drug and small interfering- RNA molecules (siRNA) to inhibit NF-kB signaling in ovarian cancer cell lines. We found that we could define a genetic signature that reflected the upregulation of NF-kB signaling in ovarian cancer and that inhibiting NF-kB signaling affected both this genetic signature and measures of aggressiveness in disease. The signature included genes associated with proliferation, survival, inflammation, adhesion, invasion, and angiogenesis; i.e., all the hallmarks of cancer.

It turns out that NF-kB regulates a very different set of genes in ovarian cancer as compared to multiple myeloma cells. That is hardly surprising, given the very different developmental and functional profiles of the two kinds of cells, but it does mean that the NF-kB activation signature needs to be identified for each cancer type.

We collaborate with Michael Birrer, M.D., Ph.D., who moved from CCR to the Dana Farber/Harvard Cancer Center a few years ago. He has access to a lot of ovarian cancer patient samples, and we have been lucky to work with him to verify that the genetic signature we identified in ovarian cancer cell lines is also found in primary tumors. In fact, we have shown that the NF-kB signature in primary tumors is associated with poor prognosis.

Finding the Right Drug

There aren’t many direct NF-kB inhibitors available, mainly because of issues related to toxicity. So, we’re looking at other points in the pathway that might be more amenable to therapeutic intervention. I am currently working with an investigational drug from Tetralogic Pharmaceuticals, TL- 32711, which mimics the cell signaling molecule, Smac. Under certain conditions, Smac shifts the balance of cell signaling from NF-kB-related proliferation to controlled cell death by apoptosis. (See “Smac, IAP, and Cancer Therapy”). TL-32711 seems to be more potent and specific than other drugs in its class because of reduced cross-reactivity.

Cells taken from primary ovarian tumors show selective expression of NF-kB family proteins. Cells stained in blue show expression of IKK-α (left) and IKK-β (middle) stained in brown, but not IKK-ε (right).
Cells taken from primary ovarian tumors show selective expression of NF-kB family proteins. Cells stained in blue show expression of IKK-α (left) and IKK-β (middle) stained in brown, but not IKK-ε (right). (Image: C. Annunziata)

We are now studying the effect of Smac mimetics on ovarian cancer cell lines. I also plan to test this drug in mouse models. NF-kB is an important part of normal cellular signaling, particularly in the immune system, which is one reason for the toxicities associated with direct inhibitors of NF-kB. Thus, it will be important to study our Smac mimetic in a mouse model with a normal immune system, so that we can observe the effect of this drug on normal and tumor-infiltrating immune cells. Many mouse models of cancer rely on human cancer cells grafted into a mouse with a deliberately dampened immune system, making such studies impossible.

If, as we predict, Smac mimetics alter expression of the genes we identified as a signature of NF-kB activity, we will want to move our work into human trials for recurrent ovarian cancer. Although such a trial would not be restricted in its enrollment, our hypothesis would be that patients with the highest levels of NF-kB activity would be most responsive to the drug. As the trial proceeds, we would look at the treatment response relative to the level of gene expression from an initial biopsy. In that way, we would hope to hone in on the most responsive patient population.

Finding the Right Cancer

I am currently an Investigator on about 10 clinical protocols at CCR, at least half of which are for ovarian cancer. Although my research focus is on NF-kB signaling, there are several other candidate targets for this disease including angiogenesis and poly ADP-ribose polymerases (PARPs). Because some ovarian cancers have dysfunction of BRCA 1 or 2, including germline mutations, these ovarian cancers have compromised DNA repair. Inhibiting PARP-associated mechanisms of DNA repair is thought to overwhelm the cell’s ability to withstand standard chemotherapeutic agents that disrupt DNA. These were the hot topics at the annual meeting of the American Society of Clinical Oncologists (ASCO) last year, but there are many genetic mutations implicated in this disease, albeit at relatively low frequencies.

A major limitation with all our clinical trials is that we have no way to identify which patients are most likely to respond to a given therapy. Ovarian cancers come in four different histological varieties: serous, clear cell, endometrioid, and mucinous. Interestingly, clear cell ovarian cancer shares some similarity in molecular mutations with clear cell renal cancer and may have a higher response to the drugs sorafenib and sunitinib that are used for renal cancers. So, clear cell ovarian cancer may be a molecular subtype that also has a histologic definition. But, that is an exception rather than the rule.

Ovarian cancer is extraordinarily heterogenous at a molecular level.

Perhaps even more than in other cancers, it seems that ovarian cancer is extraordinarily heterogenous at a molecular level. For instance, looking at data on genomic instability from The Cancer Genome Atlas (TCGA), you can see distinct hotspots of genetic abnormalities for glioblastoma and lung cancer patients, but so far, you don’t see that kind of clustering for cases of ovarian cancer. It might be, therefore, that only five percent of ovarian cancers will respond to a particular drug so, without the appropriate molecular testing, clinical trials will continue to see very low response rates.

Our trials are designed so that patients can stay on the therapy as long as there is a response and side effects are manageable. Patients may, in fact, stay on a drug regimen indefinitely. For instance, in the case of the angiogenesis inhibitors bevacizumab and sorafenib, we’ve had several patients treated for two to three years (See “Warrior Drugs,” page 32). Although there are side effects including high blood pressure and rashes, this drug combination seems to have tolerable levels of toxicity over time, which patients deem an acceptable impact on their quality of life.

There will likely come a point when the tumor evolves to evade a particular drug, and we will have to switch to another drug or another combination of drugs. So, options are important, which is one reason I continue to pursue my work in NF-kB. My personal goal, probably like many of my colleagues here at NCI, is to bring my research into the clinic.

To learn more about Dr. Annunziata’s research, please visit her CCR Web site at