Scientist Stories: Ross Hannan, Targeting Ribosomal RNA Synthesis to Treat Cancer

And it's a pleasure for me to introduce our first speaker, Ross Hannan, who is the Deputy Dean of Research at our university's College College of Health and Medicine and also the Centenary Chair in Cancer Research at the John Curtin School of Medical Research. Ross has a PhD from the University of Tasmania. He then did postdoctoral research in the United States and returned to Australia in 2000, rose through the ranks as an academic in different places in Melbourne, and as I said before, he's now a Centenary Chair of Cancer Research here at our university. He's known for his multidisciplinary work on ribosome biogenesis, which has led to new treatment paradigms in cancer centred on drugs that activate nucleolar stress. So with that, I hand over to you, Ross.

So thanks for the introduction. It's a pleasure to be here and to speak to everybody. Good morning or afternoon, wherever you are. And thank you so much to Janelia Research Campus and Howard Hughes for allowing us to talk. I'm going to talk about targeting ribosomal RNA synthesis for the treatment of cancer. And just to remind you that we're doing this work in the Shine-Dalgarno Centre for RNA Innovation. And that's been a really important new initiative led by Thomas Price. So these are my disclosures. I received funding from Pomerath Therapeutics, which is a company I helped start, and I'll talk a bit about that work today. So this is ribosome biogenesis, which I study, and ribosome biogenesis is probably one of the most complex processes the cell undergoes. It uses a considerable amount of the cellular energy, and normally constitutes up to 60% of the total transcription in the cell. Growth factor signalling pathways, or negative regulators of growth, directly regulate RNA polymerase I, which is the enzyme dedicated to the synthesis of this large non-coding RNA, the 47S, which gives rise to the 5.8S, 28S, and 18S RNA. But you also require RNA polymerase III to generate the 5S, and RNA polymerase II to generate the ribosomal proteins, which are then assembled into the large and small subunits in the nucleolus, which are then exported for translation. I've got MYC up here just to remind you that MYC is one of the most potent regulators of this pathway.. And we think it sensitises cells to these drugs shown in red, which I'm going to talk about today, which directly inhibit the transcription and synthesis of this ribosomal RNA. And we've used that to treat cancer. So just to remind you about the rDNA repeats, they're unique in the fact you have up to 400 copies of these ranged in a head-to-tail fashion. This is just a blow-up schematic of one of those rDNA units that has a promoter like a Pol II gene. And enhancer, and then the 18th, 5.8S, and 28S. And this is transcribed exclusively by RNA polymerase I. These are found on the acrocentric chromosomes. And they, through phase condensation, they coalesce, the different 5 different chromosomes coalesce into these nucleoli. And that's better shown on this diagram here. The blue is just DAPI staining, this is a nucleus. The red is RNA FISH to demonstrate the RNA transcription. It's probably worth reminding people that the nucleoli have been evolved to undergo many other processes in addition to ribosome biogenesis. And indeed, there's up to 4,000 different proteins in nucleolus, and only 30% of them are involved in ribosome biogenesis. But I won't have time to talk about those other functions today. RNA synthesis is upregulated during malignant transformation. This is a model of a MYC-driven lymphoma where MYC is driven by the EU enhancer. These mice within 20 weeks get significant disease, as shown here. If you look at the pro-pre-B cells, massive accumulation of these malignant cells. This correlates with a significant increase in RNA transcription of wild-type, pre-malignant, malignant— this is pre-RNA expression on the side there. This is a massive increase, probably 95% of all the RNA in these cells are ribosomal. And this is RNA FISH going from wild-type, premalignant to malignant. And you can see significant increase in ribosomal RNA. So it's significantly upregulated in cancer. This is the reason why we thought it might be a good target. This just shows— excuse me— if you do a Pol I ChIP for wild-type, premalignant, malignant across the transcribed region, you see a significant increase in ribosomal RNA. And that's what's driving this increased transcription. So the question is, what happens if you inhibit RNA synthesis? Do you just get less ribosomes? Well, it turns out that actually during evolution, the eukaryotes have evolved mechanisms to sense fidelity of the ribosomal biogenesis pathway. And when this pathway, because it's so important, is dysregulated, it activates these checkpoints. There's many of these checkpoints. I'm not going to go through all of them. But the basic premise is that you disrupt ribosomal biogenesis, which leads to disassembly of the nucleoli. This allows ribosomal proteins that were normally incorporated into the ribosome to go about to come out of the nucleolus and have many other functions. For example, activating the expression of p21, turning off MYC through various mechanisms, turning off E2F. Probably the most predominant one is this one here, which is the regulation of the tumour suppressor p53. 2 ribosomal proteins, L5 and L11, bind the ubiquitin ligase NDM2, therefore preventing it from degrading p53. You get accumulation of p53 and growth inhibition, apoptosis, or senescence. And that's just shown more clearly here. This is a Western blot for p53. At the top here, you can see when we add dactinomycin D at low doses, which inhibits Pol 1 transcription selectively, you get a massive upregulation of p53. And if you take out L5 or L11 by knocking them down, you can actually completely block that activation, showing it's dependent on L5 or L11. We can do a similar experiment where we take a version of MDM2 knocked in which has a mutation which prevents L5 and L11 binding. Once again, that completely blocks the activation of p53, showing this pathway accumulation of this p53 is completely dependent on this nuclear surveillance pathway. Now what's interesting is that it's not just ribosome biogenesis, inhibition of Pol I transcription directly that causes this pathway. Here we've done a high-throughput screen, and this is p53 intensity shown on the bottom. If we knock down a whole bunch of genes in the genome or ribosomal proteins, you activate p53. If you knock down L5 or L11, you block that. So almost any gene in the genome that you knock down requires nuclear surveillance pathway. And it's not just genotoxic things. If you take various treatments such as actinomycin D, alpha-amanitin, etoposide, 5-fluorouridine, UV treatment, irradiation, all of them require an intact nuclear surveillance pathway to stabilise p53, suggesting that this is a primordial, probably one of the most important pathways for regulating p53. P53. So we've tried to harness this to actually test whether we can actually use this to develop drugs. And this is just mainly for the— some of the graduate students that might be listening. This is sort of the pathway we took. Did high-throughput screens of a compound library of 100,000 compounds looking for things which inhibited Pol I transcription. We did a counter screen to make sure that they didn't inhibit a Pol II gene. We went through various PK oral cassettes, secondary screens, primary development, And this is the pathway. I won't go through it in detail. This just shows you the sort of money that we had to generate to get through these different pathways. And here, shown here is actually the sequence, the timing of this. And it's sort of important for graduate students to be developing drugs. You can see the sort of time it took. Began screening 2008, lead drug 2010, our first paper 2011. So that was like, you know, 3 years later. First funding 4 years after we started. And probably the breakthrough paper is this BioWater Red Owl cancer cell paper, which led to funding for clinical trial. This clinical trial has now been completed. This drug inhibits the initiation of Pol I transcription, and this is a tattoo of the actual CX5461 on one of my very enthusiastic students. We recommend all students get a tattoo while they're doing their PhD. And we've published many, many papers on this, so I'm not going to go through it in detail. And they can be found online. This is just one example from the paper. This is the emu MYC model again, where we adoptively transfer the MYC-driven lymphoma cells into mice and they get diseased very rapidly, within 8 days. This is the tumour cells here shown in green, so fluorescent tag. Here's the tumour cells shown here. If we treat with vehicle, has no effect. If we give a single dose of CX5461, we completely eradicate the tumour cells, and importantly, the normal cells come back. So this shows very surprisingly that you can actually target one of the most basic housekeeping processes in the cell, the synthesis of the non-coding ribosomal RNA, and get selectivity to killing tumour cells. And this was totally unexpected at the time. This leads to nuclear disassembly, as I suggested, and a rapid activation of p53 within 1 hour, and apoptosis starting— this is caspase-3 cleavage— within 3 hours. So this is a rapid response. These cells are dying rapidly.. It's not a slow response to the loss of ribosomes. So we've tried to use this to see what sort of models we could treat in cancer. This also leads to a significant increased survival of these mice. They eventually relapsed through acquired mutations which bypass this process. So this drug now has moved to Phase 1 trial, which we've completed, which we published in Cancer Discovery. It's gone on to a couple of Phase 2 trials and now has been fast-tracked for therapeutic treatment of patients with breast ovarian cancers with BRCA mutation. Now, one of the problems with this drug, as I noticed early on, is that in addition to inhibiting Pol I, it caused DNA damage around the nucleus. This is CX compared to topoisomerase. Gamma H2X is a marker of DNA damage. And it turns out this drug actually has activity to inhibit Toba2α in addition to Pol I. And it turns out that Toba2α is part of the initiation complex So this DNA damage was just occurring around the nucleus where the Pol I transcription takes place. And it was probably the reason for the limiting toxicity due to photosensitivity in the trial. Because of that, although this is a useful drug, we set out to develop second-generation Pol I inhibitors which were cleaner and had no effect on DNA damage. And that drug which we developed is PMR116, developed by Nadine Heim and Kate Hannan in the laboratory, and it has significant advantages over the first-generation inhibitor. Particularly, it penetrates the blood-brain barrier, and we want to use this for brain cancers. And a lot of this work was done at the ANU Centre for Therapeutic Discovery in the John Curtin School of Medical Research. So this drug inhibits Pol I transcription. Where is it inhibited? Well, this is a ChIP assay showing polymerase across the IDNA gene and vehicle of the PMR116. We don't decrease polymerase at the promoter and the enhancer, but we do across the gene, suggesting it's blocking promoter escape. It has about a 1,000-fold increase in sensitivity compared to Pol I compared to Pol II. So it's selective, and it selectively kills tumour cells compared to normal cells. So we think this is a good drug to take forward in our treatments. Most importantly, it doesn't induce DNA damage. It's a vehicle Oops, sorry. This is vehicle treatment of— let's go backwards— vehicle treatment of cells. The— sorry. Vehicle treatment of cells on the left, PMR and then CX5461. Green is DNA damage marked by gamma H2X. And you can see significant DNA damage with CX5461, but none with PMR116. Similarly shown here in the Western blots on the bottom, both Drugs activate p53, but only CX-5461 activates Check1 and Check2, which are markers of DNA damage pathway. So we're quite confident this drug doesn't activate DNA damage. It is a clean Pol1 inhibitor. So it treats mice with lymphoma. This is the same model I showed you before, the immunomic lymphoma, with significant increase in survival. We spent quite a bit of time on other models, haematological models. This is a model of AML. This is an incurable disease in humans. This is driven through a mixed lineage leukemia fusion proteins. These mice also bear activated RAS because about 20% of these mixed lineage leukemia cancers have activated RAS. We can also monitor the cells through GFP and luciferase. This is an adoptive transfer of these cells into the mice which come down with leukemia. As can be shown here on the right, this is 3 different doses of the drug. You can see significant increase in survival, which is better shown here on the bottom left, where you have survival versus days post-transplant. If you look at the standard therapies for patients, cytarabine, doxorubicin, they only get about 20, 22 days increase in survival. So this increase in survival is significant, and we have high hopes for this drug for treatment in AML. We've tried, tried this at a number of different tumor models. This is a colorectal cancer syngeneic model, and we can show that PMR116 is superior to Irinotecan, shown here in the survival. Red is PMR116, but this is the standard drug used for patients, Irinotecan. And you can actually follow individual tumor volume here on the left. The red is the PMR116 and blue is the Irinotecan. One of the other models we've spent quite a bit of time with— there's different models of hepatocellular cancer. This is a DEN-induced, you know, chemically induced model. And we see significant reduction in this massive increase in tumors in this model with the treatment of PMI-116. Better shown in this graph here, we can see this robust decrease in both the numbers of tumors within these livers. And so we're following this up for various other models of liver cancer. Probably, this is a collaboration with Luke Furyk from the Peter MacCallum Cancer Centre and some other people such as Garry Risbridger. We're lucky to get access to warm autopsy of patients, unfortunate patients that have died, but we've managed to get access to their tissue within a couple of hours of death, so we can get tissue from places we wouldn't normally be able to get it from, such as bone or the brain or the lung. In this case, we've got a xenograft from the mice. This is a metastatic castrate-resistant prostate cancer, which is lethal in men. And we can see here that 300 milligrams of PMR116 treatment not only does it block the growth, but actually gets rid of all the tumor cells. So this is a fantastic response given this poor unfortunate patient died of this disease. Not all of them respond as well. This is another prostate cancer metastasised to the dura. But even then, we see a significant decrease in tumour size. So we're very hopeful for this. And the phase 2 part of a second part of a phase 1 trial, we're hoping to test this on prostate cancer patients. So where are we up to with this drug? Well, we've got FDA investigational new drug status granted. We've opened a multicentre first in human study with this at the Pigma Callum Cancer Centre and other places around Australia. The first part is a dose escalation phase shown on the left, and then the second part, which we're up to at the moment, will be an expansion phase where we'll be testing those patients which we think will best respond. In this case, we're confident that patients with MYC aberrations will respond best because we know that MYC drives RNA transcription. And just finally, how else might we— in the last couple of minutes I've got— how else might we stratify patients? So it turns out that there's mutations in chromatin modifiers and actually histones themselves, H3.3, ATRX, and DAX. So ATRX complexes with DAX to deposit H3.3 to form this heterochromatic structure. And it's very important for the stability of repetitive DNA such as telomeric regions and also the ribosomal genes. Now these mutations are found in a range of cancers. We're particularly interested in brain cancers, because there's been no increased treatments or improvement in survival in brain cancer patients for the last 30 years. Now it turns out that if you have these mutations in ATRX, you actually, sorry, you actually end up losing half your RNA repeats. They're really important for RNA repeat stability. Now, this loss of the RNA repeats shown on the right-hand side leads to a significant increase in sensitivity of the drug to CXPMR116, suggesting that patients with these mutations may respond very well. And we're going to use this as a strategy to select patients in Phase 2 trials, particularly brain cancer. So I'm just going to summarise then. I've shown you that targeting Pol I transcription is a promising approach to treat cancer. And the first-generation drugs are in Phase 1 and Phase 2 clinical trials in Australia and overseas. This therapeutic effect is mediated by checkpoint activation and also differentiation. I didn't have time to show you. And then this— I just want to re-emphasise, this is not a slow response to the loss of ribosomes. It's a rapid activation of these checkpoints. And we believe this nuptial surveillance pathway has been activated as a primordial sensor for cellular stress. Something like akin to the canary in the coal mine. Almost all cellular stresses that stabilise p53 require this functional nucleolar surveillance pathway. That's quite surprising to us. So with that, I'll finish. This is the people who have done the work: Kate Hedden, Nadine Hyne, Rita Ferreira, Amy George, and Pilita Poh from my lab, and Tijani Udumana. This is also a collaboration with Professor Luke Furek from the Peter MacCallum Cancer Centre. With that, I'll leave this last slide and I'll finish. Thank you very much.