Overcoming ovarian cancer resistance

Video Transcript

• 00:00 - 00:14:  Welcome, everyone, to the second session of our series on translational and therapeutic advances
• 00:14 - 00:20:  in oncology and epigenetics. Today, we'll be focusing on overcoming ovarian cancer resistance.
• 00:20 - 00:26:  We have three exciting talks, followed by a panel discussion. Please submit your questions
• 00:26 - 00:32:  to the speakers at any time during today's event using the Q&A box found at the bottom of your
• 00:32 - 00:39:  screen. Abcam is approved as a provider of continuing education programs in the clinical
• 00:39 - 00:46:  laboratory sciences by the ASCLS PACE program. PACE credits are available for this live session,
• 00:46 - 00:50:  and a link to request credit will be sent in the chat box at the end of the event.
• 00:51 - 00:54:  You'll also find more details in our post-event email.
• 00:56 - 01:01:  So, my name is Mike Brayter. I'm Abcam's senior commercial strategy manager for oncology.
• 01:01 - 01:06:  I'm going to spend a few minutes talking about how Abcam is supporting oncology research
• 01:06 - 01:11:  and development through new product developments and collaborations,
• 01:11 - 01:14:  and highlight ways in which you can help shape this and get involved.
• 01:15 - 01:20:  So, my goal is to ensure that oncology researchers have the reagents that they need
• 01:20 - 01:27:  to advance cancer research, both right now and in the future. And that involves tracking key
• 01:27 - 01:33:  emerging targets and trends within the oncology field. Importantly, it involves speaking to
• 01:33 - 01:39:  researchers to help understand common needs, and then to feed those insights into our new
• 01:39 - 01:46:  development pipelines. We also collaborate and partner to integrate
• 01:46 - 01:51:  our reagents into novel diagnostic assays and novel technologies.
• 01:53 - 01:57:  So, all of those insights from academia and industry are fed into our
• 01:57 - 02:02:  development pipelines, and the goal is to ensure that we have a comprehensive
• 02:02 - 02:08:  reagent portfolio covering all of the key oncology pathways, hallmarks, and functions.
• 02:10 - 02:15:  Diving a bit deeper into ovarian cancer, you're likely very familiar with our recombinant
• 02:15 - 02:23:  antibodies to key ovarian cancer targets, but we're also investing heavily into in-house
• 02:23 - 02:28:  product developments and into other product lines, such as recombinant proteins,
• 02:29 - 02:38:  specifically native full-length proteins. Also, ELISA kits, highly sensitive assays for circulating
• 02:38 - 02:45:  tumor biomarkers, and also knockout cell lines to key targets in relevant cell line backgrounds.
• 02:48 - 02:55:  So, we're supporting cutting-edge research. I want to highlight a couple of examples here,
• 02:55 - 03:01:  and the first one is we co-developed a reagent, a recombinant antibody, with Dr. Ryan Russell
• 03:01 - 03:11:  at the University of Ottawa to phospho-ATG16L1, which identifies cells at a very early stage of
• 03:11 - 03:20:  autophagy initiation, and this allows studies into autophagy as a driver of cancer resistance
• 03:21 - 03:28:  and resistance to immunotherapies. I'd also like to highlight a webinar given by our chair,
• 03:28 - 03:31:  Professor, who's chairing this session today, Professor Wendy Fandall,
• 03:33 - 03:41:  where she explores a novel NK cell subtype that's associated with a novel resistance mechanism
• 03:41 - 03:48:  in ovarian cancer. As well as supporting fundamental academic research, our reagents
• 03:48 - 03:56:  are increasingly being used by partner companies within novel next-gen diagnostic assays to help
• 03:56 - 04:01:  identify which patients are most likely to respond to targeted therapies.
• 04:04 - 04:08:  I've highlighted here some of the ways in which we collaborate with oncology researchers,
• 04:09 - 04:17:  sharing of insights, both scientific and technical. We support individual projects with
• 04:18 - 04:26:  technical expertise, with product recommendations, and also giving visibility of products that are in
• 04:26 - 04:32:  our late-stage development pipelines, and giving earlier access to those and beta testing
• 04:34 - 04:41:  and data sharing, particularly in novel applications such as multiplex imaging of tumors.
• 04:43 - 04:48:  I'd like to highlight the next session within this series, which is coming up on Friday,
• 04:48 - 04:53:  June the 10th, focusing on earlier detection of pancreatic cancer.
• 04:54 - 04:57:  The link to sign up should also be in the chat.
• 05:00 - 05:06:  Feel free to contact me at the email addresses shown to discuss anything I've presented or to
• 05:06 - 05:11:  discuss collaboration opportunities. So, moving on to today's session, it'll be chaired by
• 05:11 - 05:17:  Professor Wendy Fandall from Stanford University, whose career has spanned academia, biotech,
• 05:17 - 05:23:  diagnostics, and drug development. Her team's translational research explores why ovarian
• 05:23 - 05:28:  cancers frequently become resistant to chemotherapy and often fail to respond to immunotherapy.
• 05:29 - 05:35:  Her lab uses single-cell multiplex proteomic technologies, such as mass cytometry and
• 05:35 - 05:40:  multiplex imaging, to reveal tumor and immune cell heterogeneity and drivers of resistance.
• 05:41 - 05:44:  I'll now hand over to Professor Fandall to start today's session.
• 05:47 - 05:54:  So, good morning from San Francisco, everybody. A very warm welcome, whichever time zone or
• 05:54 - 05:59:  geographical location you're in. First, I'd like to thank Abcam for giving me the opportunity to
• 05:59 - 06:05:  organize this meeting. And, of course, Mike and Sarah could not have done this without you.
• 06:06 - 06:12:  For those members of the audience who may be new to this field, I thought I'd just put some
• 06:12 - 06:15:  introductory slides, share with you a couple of introductory slides
• 06:16 - 06:19:  about the disease and what we're dealing with.
• 06:21 - 06:27:  Okay. So, the major objective of this meeting is to discuss strategies for overcoming platinum
• 06:27 - 06:34:  resistant high-grade serous ovarian carcinoma. It's the most prevalent epithelial ovarian cancer.
• 06:34 - 06:39:  There are about a quarter of a million newly diagnosed cases worldwide, about 22,000 in the
• 06:39 - 06:47:  U.S. and about 7,000 or 8,000 in the U.K. This graphic here shows that even though it's called
• 06:47 - 06:53:  ovarian cancer, it's recognized now that the cell of origin comes from the fallopian tube,
• 06:53 - 06:57:  and you'll be hearing more about that from Professor Ahmed later.
• 06:58 - 07:05:  So, just a brief overview of the disease. It's considered a stealth disease because it's
• 07:05 - 07:10:  asymptomatic in the early stages. There's a lack of effective screening tools to detect
• 07:10 - 07:18:  early disease, and consequently, it's diagnosed at high stage. The treatment has been about the
• 07:18 - 07:22:  same treatment for about the past three decades. It's a combination of surgical debulking with
• 07:22 - 07:29:  platinum-based chemotherapy. A subgroup of patients, about 25 percent, have intrinsic
• 07:29 - 07:34:  resistance to chemotherapy. In other words, they don't respond. Most patients, however,
• 07:34 - 07:41:  do respond, but then they relapse with acquired platinum resistance. This results in a rather
• 07:41 - 07:50:  dismal five-year survival rate of about 30 to 40 percent. And hereditary mutations for women
• 07:50 - 07:56:  have, in BRCA1 and BRCA2 DNA damage repair genes, increased the lifetime risk of high-grade serous
• 07:56 - 08:04:  ovarian cancer. And recently, PARP inhibitors, poly-ADP ribose polymerase inhibitors, have been transformative
• 08:04 - 08:14:  for women with these mutations. And so now, I will introduce today's speakers. Today's roster
• 08:14 - 08:21:  comprises three physician scientists who are all experts in the field of gynecologic oncology.
• 08:21 - 08:26:  They are dedicated clinicians, and their work in the lab has made seminal contributions
• 08:26 - 08:31:  to the field. You'll hear about their different approaches to this problem that hold great
• 08:31 - 08:39:  potential. You will hear Dr. Professor James Brenton's work, his approach to unravel the
• 08:39 - 08:46:  genetic complexity. It's a malignancy with one of the most unstable genomes. Professor
• 08:46 - 08:53:  Catherine Fu will discuss her approaches that include the tumor microenvironment and stromal
• 08:53 - 08:59:  cells. And Professor Ahmed Ahmed will discuss chemoresistance within the context of non-genetic
• 08:59 - 09:06:  heterogeneity. And so, to introduce the first speaker, it's my very great pleasure to introduce
• 09:06 - 09:11:  Professor James Brenton from the University of Cambridge, UK. James is a practicing medical
• 09:11 - 09:17:  oncologist and senior group leader at the Cancer Research UK Cambridge Institute and professor of
• 09:17 - 09:22:  ovarian cancer medicine in the Department of Oncology at the University of Cambridge.
• 09:22 - 09:27:  His contributions have focused on understanding chromosomal instability in high-grade
• 09:27 - 09:33:  serous ovarian carcinoma and include the conclusive demonstration that mutations in the
• 09:33 - 09:41:  p53 or TP53 gene are a ubiquitous hallmark of high-grade serous ovarian carcinoma, thereby permitting that those mutations
• 09:41 - 09:48:  permit the diverse and extreme chromosomal instability. Second major contribution,
• 09:48 - 09:53:  clonal expansions determined by copy number alterations in ovarian cancer correlate with
• 09:53 - 10:01:  increased risk of recurrent disease. And also, he's also been involved in a team developing the
• 10:01 - 10:07:  first targeted and whole genome next-gen sequencing of circulating tumor DNA assays for women with
• 10:07 - 10:16:  ovarian cancer. And recently, his study was the lead senior author using copy number alterations
• 10:16 - 10:22:  discovering new mutational signatures to allow the classification of tumors with the potential
• 10:22 - 10:28:  for predicting response to chemotherapy. We're really lucky to have him. I'm absolutely delighted
• 10:28 - 10:37:  to introduce him. James, the stage is yours. Wendy, thank you very much. Let me just share
• 10:37 - 10:47:  the screen. I hope everyone can see pairs in front of them. So, I'm going to spend the next
• 10:47 - 10:54:  20 minutes or so describing why it's important to understand the building blocks of the diversity
• 10:54 - 11:00:  and heterogeneity in high-grade serous ovarian carcinoma. That is the underlying wiring diagram
• 11:00 - 11:08:  of chromosomal instability. And I use this slide of pairs to try and indicate this requirement to
• 11:08 - 11:12:  understand the heterogeneity in the disease because without molecular classification,
• 11:12 - 11:18:  we are not going to get to the ambition of precision medicine. So, just a quick set of
• 11:18 - 11:24:  aims for today's talk. I want to get over the challenges about molecularly stratifying
• 11:24 - 11:31:  high-grade serous ovarian carcinoma, what are biomarkers from whole genome sequencing that may
• 11:31 - 11:37:  be relevant both in the clinic now and in the future, and potential mechanisms by which
• 11:37 - 11:42:  whole genome sequencing can get into the clinic and start defining intrinsic sensitivity.
• 11:43 - 11:49:  So, I'm conscious that I'm talking in a meeting sponsored by a protein company
• 11:50 - 11:57:  with colleagues and friends who are also using protein-based approaches, but I don't—you know,
• 11:57 - 12:01:  this is one part of the story that I want to concentrate on today. In other words,
• 12:01 - 12:07:  what is the hammer that determines genomic aberrations? What I'm not going to cover
• 12:07 - 12:16:  is how cancer cells adapt and avoid treatment, and I'm very much looking forward to my
• 12:16 - 12:23:  two colleagues' talks to represent those key questions. So, as Wendy has summarized,
• 12:24 - 12:29:  high-grade serous ovarian cancer really hasn't seen a significant change in overall survival
• 12:30 - 12:34:  over the past 20 years. We don't have a useful molecular classification in the clinic,
• 12:34 - 12:41:  apart from identifying women with BRCA1 or BRCA2 mutations. But critically,
• 12:42 - 12:49:  the chromosomal instability, which is the most extreme of all tumors in TCGA and ICGC
• 12:50 - 12:56:  studies, may be a key to identifying treatment options. So, the development of synthetic
• 12:56 - 13:03:  lethality approaches for mutational phenotypes has been extremely important. In fact, I think
• 13:04 - 13:09:  this is probably the most exciting slide in ovarian cancer in the past 20 years. It's the
• 13:10 - 13:17:  current still-preliminary results from the STELAR1 trial, in which patients who have a BRCA1 or BRCA2
• 13:17 - 13:24:  germline anastomotic mutation receive two years of maintenance treatment with a PARP inhibitor
• 13:24 - 13:30:  olaparib. What I think is so extraordinary about this graph is not simply the differences
• 13:30 - 13:34:  in survival between patients who received a PARP inhibitor and those who did not,
• 13:35 - 13:42:  but the fact that there is a strong indication that the proportion of patients who are disease-free
• 13:43 - 13:48:  as time goes on is now plateauing, which is reminiscent of the changes seen in imatinib
• 13:48 - 13:54:  for treatment of chronic myeloid leukemia. So, the survival data is still not mature yet from
• 13:54 - 14:02:  this trial, so we still can't say for sure what the cure rate is, but I think that the huge,
• 14:02 - 14:08:  huge exciting thing is that actually women who would otherwise have died of high-grade
• 14:08 - 14:12:  serous ovarian cancer are cured by maintenance PARP inhibitors, which is an extraordinary
• 14:13 - 14:21:  thing. The title of this session is really about platinum resistance, but this is a phrase we use
• 14:21 - 14:28:  rather imprecisely, and we should think about how we define platinum resistance. That phrase
• 14:28 - 14:32:  encapsulates the fact that upfront treatment, as Wendy summarized a few moments ago,
• 14:32 - 14:36:  has not changed for over 30 years. This is platinum-based chemotherapy.
• 14:36 - 14:42:  Possibly the best way of understanding sensitivity to treatment is to look at
• 14:42 - 14:48:  the results of neoadjuvant treatment from the ICON-H trial, so a large, multi-center trial,
• 14:48 - 14:54:  and this trial really provides the best evidence for what the neoadjuvant response is. Not all
• 14:54 - 15:00:  patients had neoadjuvant treatment, but subset analysis of those that did show that only 6%
• 15:00 - 15:06:  of patients with advanced ovarian cancer do not respond at all to treatment and progress
• 15:06 - 15:10:  through therapy, so that would be platinum-refractory disease or platinum-resistant
• 15:10 - 15:17:  disease. Actually, 61% will achieve an objective response, so complete and partial responses,
• 15:17 - 15:24:  of course, are formal radiological terms using RECIST criteria, so a shrinkage of 30% or more
• 15:24 - 15:28:  of the longitudinal diameters of index lesions, so that's a pretty high rate,
• 15:28 - 15:33:  especially given the bulk of disease at presentation. But more concerning, I think,
• 15:33 - 15:40:  is that 33% of patients have no objective benefit from upfront neoadjuvant chemotherapy,
• 15:40 - 15:47:  and the reasons for that remain unclear. The second—so I would say, just talking about this,
• 15:47 - 15:52:  we should talk about intrinsic sensitivity to treatment and how we define that, and I'm going
• 15:52 - 15:58:  to focus on that for quite a lot of this talk. The other aspect is platinum-resistant. So,
• 15:58 - 16:04:  this is a woman who was under my care who was part of the original whole-genome sequencing
• 16:04 - 16:10:  pilot carried out by the Genomics England project. She presented with very poor outcome disease,
• 16:10 - 16:15:  very large volume ascites, bulky stage 4 disease at time of presentation,
• 16:16 - 16:22:  but by virtue of whole-genome sequencing, we discovered, as her treatment progressed with
• 16:22 - 16:27:  neoadjuvant chemotherapy, that she had a translocation in BRCA1. So, in other words,
• 16:27 - 16:32:  she had a somatic mutation of BRCA1, which would not have been detected by panel gene testing.
• 16:33 - 16:39:  Concomitant with that, she had a very rapid response to initial upfront carboplatin paclitaxel,
• 16:40 - 16:44:  had successful surgery—not perfect surgery, but successful surgery—and
• 16:45 - 16:50:  then had a long disease-free interval before she recurred again, responded again to platinum-based
• 16:50 - 16:56:  chemotherapy, and then 12 months on a PARP inhibitor, but ultimately progressed with
• 16:56 - 17:01:  platinum-resistant disease in which we challenged it. Platinum chemotherapy did not have response.
• 17:01 - 17:05:  And I think for the historical basis of the word platinum-resistant, we're mostly talking about
• 17:05 - 17:10:  acquired platinum resistance. So, this patient clearly was platinum-sensitive, but other
• 17:11 - 17:17:  epigenetic, genetic transcriptional changes have occurred to alter that sensitivity. So,
• 17:17 - 17:23:  this remains an important question, but I'm going to focus mostly on intrinsic sensitivity to
• 17:23 - 17:27:  treatment. And the first problem that we have been—well, the main problem we've been trying
• 17:27 - 17:33:  to solve in the past five-plus years is really how to understand the genome of high-grade
• 17:33 - 17:39:  serous ovarian cancer. And this cartoon is just showing the sort of information we're dealing
• 17:39 - 17:45:  with. The top plot is showing what happens if you look at copy number profiles with whole-genome
• 17:45 - 17:51:  sequencing in normal cells. You have two copies of every segment of the chromosome. The chromosomes
• 17:51 - 17:57:  are aligned along the x-axis. One is obviously largest. Twenty-two is smallest. But in tumors
• 17:57 - 18:04:  like lung cancer, triple-negative breast cancer, glioblastoma, you see this extraordinary complexity
• 18:04 - 18:08:  with different copy number changes and other changes not shown here of structural variance.
• 18:09 - 18:15:  And this movie is just showing you 117 patients from the national—UK national—Britrock study,
• 18:15 - 18:20:  which myself and Ian McNeish set up to study the basis of chromosome instability and evolution
• 18:20 - 18:27:  in high-grade serous ovarian cancer. So, every dot is an abnormal copy number state, and the cases
• 18:27 - 18:32:  are ordered by complexity. But the main purpose of this movie is to show you how very different
• 18:32 - 18:38:  patients are when you look at whole-genome profiles from patients. So, it's very important
• 18:38 - 18:44:  to study copy number aberrations, but it's not a main feature of clinical care at the moment.
• 18:44 - 18:49:  It's important because they're cardinally—a cardinal feature of chromosomal instability.
• 18:49 - 18:57:  The amplicons selected by chromosome instability—say, PIK3CA, MYC—are important drivers
• 18:57 - 19:02:  in the disease. And we know, I think most people in this audience know, that non-synonymous or
• 19:02 - 19:08:  oncogenic mutations are, in fact, very infrequent in high-grade serous ovarian carcinoma. And they
• 19:08 - 19:14:  have predictive value, not only in established cancer. So, my colleague Rebecca Fitzgerald
• 19:14 - 19:21:  recently showed that early signs of chromosomal instability in the esophagus was a very powerful
• 19:21 - 19:26:  precursor to the development of invasive disease. And we've shown in work I'm not going to talk
• 19:26 - 19:30:  about today that chromosomal copy number changes can start to be a basis for choosing
• 19:30 - 19:36:  therapeutic choices. There are problems, though. So, first of all, it's complex to interpret copy
• 19:36 - 19:41:  number changes. They're much less binary and countable than non-synonymous changes.
• 19:42 - 19:47:  And in terms of drug discovery, classical CRISPR approaches have underrepresented discoveries
• 19:47 - 19:51:  around copy number changes, mostly because of off-target effects that lead to false positive
• 19:51 - 19:56:  hits. So, this is a relatively unexplored area when you look across CRISPR technologies.
• 19:56 - 19:59:  And also, though, many people will
• 20:00 - 20:06:  talk a good talk about how important copy number states are, there are still major challenges.
• 20:06 - 20:12:  First of all, you have to deal with ploidy differences between haploid, diploid,
• 20:12 - 20:18:  and whole-genome-duplicated genomes, which can be difficult to tease out using
• 20:18 - 20:23:  next-generation sequencing alone. You've always got the problem of tumor purity for any molecular
• 20:23 - 20:30:  analysis in the clinic. And lastly, you've got tumor subclonality, which tends to be underestimated
• 20:30 - 20:34:  because the tools we have, of course, are mostly based around bulk sequencing.
• 20:35 - 20:39:  But it's still vital, I think, to see how we will bring these analyses into the clinic,
• 20:39 - 20:44:  partly because the understanding of the mutational processes that are patterning the genome,
• 20:44 - 20:51:  those hammers that are inducing specific patterns of chromosomal damage, are probably,
• 20:51 - 20:58:  in large part, determining the diversity and ability of tumors to explore evolutionary space.
• 20:58 - 21:04:  And the selection that occurs under the effect of those hammers may also determine other important
• 21:04 - 21:11:  features, not least of all patterns of metastatic spread, tropism for particular tissues, and, of
• 21:11 - 21:16:  course, the patterning and instruction set of the tumor microenvironment, both in a positive way,
• 21:16 - 21:22:  sequestering, stimulation, and innate immune responses, and in negative ways by immunoediting,
• 21:22 - 21:27:  deletion of 8-PTLA. So, understanding the whole picture here is critical. And for that,
• 21:27 - 21:33:  you really need a whole genome, not a targeted sequencing. And we showed some years ago, in 2015,
• 21:33 - 21:40:  that actually you can measure ongoing evolution through time and space using copy number profiles.
• 21:40 - 21:46:  So, in this case here of a woman who was diagnosed in 2008, relapsed in 2010,
• 21:47 - 21:53:  whole genome sequencing showed the acquisition of an NF1 mutation in the ascites at recurrence.
• 21:53 - 21:58:  And when we used that as a molecular barcode for the clonal population that was predominantly
• 21:58 - 22:04:  relapsed, what we were able to show is variable contributions of those clones over time and space,
• 22:04 - 22:13:  really fulfilling models that had been postulated for some time across many cancer types of selection
• 22:13 - 22:20:  of specific clones. The other thing to say is that copy number aberrations are frequently subclonal
• 22:20 - 22:24:  in patients. This is work building on our plus medicine, which is currently on bioRxiv,
• 22:25 - 22:32:  really showing that key drivers such as MYC, KRAS, cyclin E can exist in subclonal populations. So,
• 22:32 - 22:38:  just one biopsy is not likely to tell you everything about the potential for driver
• 22:38 - 22:45:  genes in this disease. But if I quickly switch to how we make sense of that chaos,
• 22:45 - 22:50:  we have been inspired for several decades by the work of Mike Stratton, which goes back
• 22:50 - 22:58:  to the early 1990s of looking at specific patterns and mutation genes such as TP53,
• 22:58 - 23:06:  and how those reflect specific mutational exposures. So, in this still very useful
• 23:06 - 23:11:  review by Serena Nixon-Arle and Thomas Helleday, this concept really of the
• 23:11 - 23:15:  mutational signatures and the presence of mutations being an archaeological record
• 23:16 - 23:21:  of what mutational processes have done to a genome across the life of that cell.
• 23:22 - 23:29:  And I think that the major advance by Stratton and colleagues was to really understand ways of
• 23:29 - 23:36:  deconvoluting complex patterns as signatures that could then be turned into identification
• 23:36 - 23:41:  mutational processes based on nucleotide substitution. And so, about four or five
• 23:41 - 23:46:  years ago, we started wondering whether instead of looking at substitution patterns, which are
• 23:46 - 23:50:  not necessarily the most informative things in high-grade serous ovarian cancer, we started to look at specific
• 23:50 - 23:56:  copy number features to see if they themselves could be used as the patterns to deconvolve
• 23:56 - 24:01:  ongoing mutational processes. So, this slide really just summarizes our Nature Genetics
• 24:01 - 24:06:  paper from 2018, where we took those same 117 patients you've just seen on the movie
• 24:07 - 24:14:  and used six features to look at what patterns of copy number change were present genome-wide,
• 24:14 - 24:20:  and turn those patterns into statistical distributions of and using mixture modeling
• 24:20 - 24:24:  of what those features were. And then in a similar way to the Stratton approach
• 24:24 - 24:31:  of using non-negative matrix factorization, we were then able to pull out candidate signatures,
• 24:31 - 24:36:  copy number signatures, and map those with statistical associations onto processes such
• 24:36 - 24:42:  as chromothripsis, homologous recombination deficiency, and tandem duplication. Those
• 24:42 - 24:49:  changes, when applied to large public datasets, so there's about 600 patients on this image here,
• 24:49 - 24:55:  reveal several things. So, first of all, bulk signatures, just as they do, just as for nucleotide
• 24:55 - 25:01:  signatures, are present. Most patients have more evidence of more than one ongoing signature
• 25:01 -  25:07:  throughout the lifetime of their cancer. That poor outcomes associated with something called
• 25:07 - 25:12:  copy number signature one, which I'll describe in a minute, which is really related to
• 25:12 - 25:18:  intrinsic platinum insensitivity, whereas increased survival is related to homologous
• 25:18 - 25:23:  recombination deficiency with signatures three and seven. So, if you look at this, you see there's
• 25:23 - 25:30:  an enrichment of the green signature number one, which attenuates over time, whereas the HRD
• 25:30 - 25:37:  signatures of three and seven tend to be present throughout the lifetime, sorry, across survival
• 25:37 - 25:41:  differences. And these straight lines below are just showing the Cox proportional hazards analysis
• 25:42 - 25:47:  for the strength of these associations. But rather like nucleotide signatures, those features that
• 25:47 - 25:52:  we're pulling out in those distributions do actually give us insights into what the underlying
• 25:52 - 25:59:  mutation processes are. So, this is copy number signature one, which is characterized by relatively
• 25:59 - 26:06:  large segment size, shown here in the purple, relatively few breakpoints per chromosome arm,
• 26:06 - 26:12:  and relatively few breakpoints per 10 megabase features. And when you put those things together,
• 26:12 - 26:17:  this is really a feature of breakage-fusion-bridge or fallback inversion mutations,
• 26:17 - 26:26:  in which during mitosis, the ends of chromosomes become, have telomeric dysfunction and
• 26:26 - 26:32:  undergo sticking during mitosis with pull-apart mutations.
• 26:34 - 26:42:  So, these signatures, by their associations, suggest therapies, but I would be the first
• 26:42 - 26:47:  person to say that this, apart from HRD, this approach on the sample set we had for the Nature
• 26:47 - 26:54:  Genetics paper is still taking us toward predictions, not proof. Lastly, just before I move on to
• 26:54 - 26:58:  the second part of the story, the other useful thing about copy number aberrations
• 26:58 - 27:04:  is they're relatively easy to measure using non-invasive measures, particularly blood
• 27:04 - 27:10:  biomarkers such as circulating tumor DNA. So, this is the shallow whole genome profile of a patient
• 27:10 - 27:15:  with high-grade serous ovarian carcinoma, shown on the left-hand side, and this is the profile from a
• 27:15 - 27:23:  single blood spot obtained from that patient with whole genome sequencing. So, by using specific
• 27:23 - 27:30:  features of circulating tumor DNA characteristics and a purification, we can now measure copy number
• 27:30 - 27:37:  profiles on literally 50 micrometers of blood, and we've now developed this further for preclinical
• 27:37 - 27:43:  models in PDXs so that many of the predictions around copy number signatures can be tested
• 27:43 - 27:50:  using the PDX models that we've derived. So, this is a quick timeline of where I think we're going
• 27:50 - 27:54:  with the copy number signature work. This is the work that you've seen already in 2018.
• 27:55 - 28:02:  In 2020, we showed that contributions for copy number profiles and signatures might have a role
• 28:02 - 28:08:  in altering immune states, so this is Martin Mellor's Nature Genetics paper, and in a very
• 28:08 - 28:13:  nice collaboration with the Marsden and Clare Turnbull, we showed that copy number signatures
• 28:13 - 28:19:  could explain patterns of platinum resistance in testicular cancer. And then we were delighted to
• 28:19 - 28:24:  see application of copy number signature approaches in myeloma to detect poor outcome
• 28:24 - 28:28:  myeloma from chromothripsis patterns from copy number aberrations, which have
• 28:28 - 28:33:  been directly used from here. And then there's a paper that's just coming out at the moment,
• 28:33 - 28:37:  which we've worked on with Ian McNeish, showing that whole genome duplication based on copy
• 28:37 - 28:45:  number signatures is a feature of high-stage versus low-stage ovarian cancer. And then in
• 28:45 - 28:51:  work that's just about to come out and is currently in press, my colleagues Friar Markowitz
• 28:51 - 28:57:  and Geoff McIntyre have taken the copy number signature work to a pan-cancer analysis.
• 28:58 - 29:07:  So, this cartoon just summarizes the work done. So, taking TCGA and using the SNP6 data,
• 29:08 - 29:15:  we've identified 6,000 cases across all cancer types that have high degrees of
• 29:17 - 29:21:  chromosomal instability and used a very similar approach to the Nature Genetics approach. And,
• 29:21 - 29:26:  of course, Geoff McIntyre, now a group leader at CNAO in Spain, was the first author of the
• 29:26 - 29:33:  Nature Genetics paper. But from this, we now have 17 signatures of chromosomal instability,
• 29:33 - 29:40:  and we've been able to use the power of TCGA and the ICGC Peacock sets to carry out detailed
• 29:41 - 29:45:  association studies using this. And I just want to touch on a couple of high points
• 29:45 - 29:50:  from this, which is really how this has helped refine the diagnosis of homologous recombination.
• 29:51 - 29:59:  So, 17 signatures now. These are now numbered by prevalence. So, CX1 and CX2 are the most prevalent
• 30:00 - 30:05:  mutational processes defined by copy number signatures across all cancer types. And I think
• 30:05 - 30:10:  most people in this audience will be interested to see that this is both measures of impaired
• 30:10 - 30:16:  homologous recombination, but also the breakage-fusion-bridge type pattern, which is copy number
• 30:16 - 30:23:  signature 1 from the ovarian work. The pan-cancer signatures have also meant that we've seen
• 30:24 - 30:32:  different degrees of homologous recombination repair, really defined by the patterns of changes
• 30:32 - 30:42:  across the genome and features of replication stress. So, CX2, CX5, and CX3 are really the
• 30:42 - 30:48:  signatures that come out. And as you can see here, that CX3 is most strongly associated with
• 30:48 - 30:55:  better outcome using survival as a surrogate for platinum sensitivity, whereas CX2 and CX5
• 30:55 - 31:01:  have different effects. So, CX2 does not have a significant effect on survival, whereas CX5
• 31:01 - 31:07:  may actually be deleterious to survival. But by analysis of these features, we came up with
• 31:07 - 31:16:  a relatively simple metric, which says that if replication stress, larger loss of heterozygosity.
• 31:16 - 31:24: events, has the strongest association in CX3 with outcome, and CX2 has no difference, then simply
• 31:25 - 31:30: looking at the ratio of CX3 to CX2 might be an easy classifier for understanding
• 31:31 - 31:39: whether a patient has HRD. And indeed, this is a remarkably easy approach that functions as well
• 31:39 - 31:44: as other metrics used across whole genome sequencing, particularly as Srin and Nixon
• 31:44 - 31:51: asked HR Detect and others. Similarly, we were able to take the bulk signatures and apply them
• 31:51 - 31:57: across a large panel of public data sets from cell line profiling. I think the exciting thing about
• 31:57 - 32:03: this is we start to identify associations with synthetic lethal approaches that are exciting
• 32:03 - 32:09: around copy number signature 1, CX1, typically targeting telomere maintenance and mitosis.
• 32:10 - 32:15: So, I've got a couple of minutes to finish up. I'm just going to touch on a couple of questions,
• 32:15 - 32:20: then probably skip over the last few slides so I don't keep everybody waiting. But, you know,
• 32:20 - 32:24: I think the key question is, how do we get these discoveries into the clinic? And we've seen in the
• 32:24 - 32:33: past three years, the widespread uptake of obviously routine BRCA1, BRCA2 testing, early access of HRD
• 32:33 - 32:41: testing around 2021, and clear guidelines from ASCO and this. But what is very exciting now is
• 32:41 - 32:45: in the UK, we have now got approval for bringing deep whole genome sequencing into the clinic
• 32:45 - 32:52: as of the end of 2021 into 2022. So, we at Cambridge have just sequenced the first 10
• 32:52 - 32:59: patients as routine patient care to bring these assays into the clinic. And I think that's
• 32:59 - 33:03: important because although we have commercial tests like the Myriad test, which are really
• 33:04 - 33:12: summary measures of how close your genome resembles a BRCA1, BRCA2 genome as shown here.
• 33:12 - 33:19: So, there's a simple additive model that takes these features and compares them against BRCA1,
• 33:19 - 33:24: BRCA2 patients. This is never going to be a very good way of identifying patients like
• 33:24 - 33:27: the copy number signature and breakage fusion bridge patterns.
• 33:28 - 33:33: And we need to understand the diversity of genomes that are present. And as I've shown you already,
• 33:33 - 33:38: there's at least seven patterns within high-grade serous ovarian cancer. So, these are now tests
• 33:38 - 33:44: that are available. The challenge, of course, is getting fresh frozen tissue and that diagnostic
• 33:44 - 33:51: pathway into the NHS, which we did originally on the pilot where we sequenced over 350 women with
• 33:51 - 33:55: ovarian cancer. Now we've got to do it for real together with the tumor boards and everything
• 33:55 - 34:02: else. But our plan for this has really been partially implemented. We've brought open source
• 34:02 - 34:10: algorithms into HRD testing based on retrospective data. But our aim is to bring this forward,
• 34:10 - 34:16: including the signatures I've described earlier today. But ultimately, I think it's not just
• 34:16 - 34:24: about the genome. So, the important predictors are going to be a mixture of both functional
• 34:24 - 34:29: protein-based assays, as well as other features of the tumor microenvironment to predict outcome.
• 34:32 - 34:36: There's just one little unpleasant secret that I want to highlight before I finish talking,
• 34:36 - 34:40: that is that when we're measuring chromosome instability with whole genome sequencing,
• 34:40 - 34:46: we're really looking at bulk profiles. So, that encompasses both ongoing active signatures,
• 34:46 - 34:52: but also signatures that are extinct. That's the archaeological bit, and also episodic changes.
• 34:53 - 34:56: But that means that at the time of a treatment decision, we may not actually be measuring
• 34:58 - 35:01: with the highest sensitivity what's actually going on in the patient's tumor.
• 35:02 - 35:06: And the way to do that is to actually start looking at what is happening in single cells
• 35:07 - 35:12: and looking at recent copy number aberrations, because bulk sequencing is really telling you
• 35:12 - 35:16: about what the history of the tumor is and clonally selected expansions.
• 35:17 - 35:24: So, just very briefly as an exemplar of why this is important, if we look at the PO1, PO4 cell lines,
• 35:24 - 35:34: PO4 has a reversion mutation in BRCA2. PO1 has a loss of function mutation in BRCA2.
• 35:34 - 35:39: So, this is platinum sensitive and PARP sensitive. This is partially platinum resistant,
• 35:39 - 35:44: partially PARP resistant. What you can see from the heat map here of single cell sequencing is
• 35:44 - 35:52: many more copy number aberrations that have recently derived in PO1 and far fewer in PO4.
• 35:52 - 35:57: And when we look at the bar plots here, we see significant differences in the numbers of
• 35:57 - 36:03: aberrations and the dynamic range of aberrations over time. So, what we're driving at now is really
• 36:03 - 36:11: using these patterns to apply those to active signatures as predictors of both HRD and other
• 36:11 - 36:17: tools. And I think the future is really about bulk and single cell sequencing in the clinic if we're
• 36:17 - 36:23: going to use just the genome. But as I've said, that may not be the case. Right, I'm going to
• 36:24 - 36:30: spin over this and just summarize where I think our major gaps exist. I've told you about some
• 36:30 - 36:35: approaches to understanding chromosomal instability, but we need better approaches.
• 36:35 - 36:42: And we need other innovations here. I think some of these things that I've described will improve
• 36:42 - 36:48: clinical testing for BRCA1 and BRCA2. But again, big area of importance because we want to say to
• 36:48 - 36:52: the woman in front of us, you know, before you start treatment, whether it's going to help.
• 36:52 - 36:58: Data science is critical to this. I haven't, I've skipped over the last two slides. And lastly,
• 36:58 - 37:03: we still need much better functional models going forward. And with that, I'd just like to thank the
• 37:03 - 37:09: people who are credited on the slides and our funders, and I'd be very happy to take any
• 37:09 - 37:17: questions. Thank you. That was a fantastic talk, James. Thank you so much. I will start with our
• 37:17 - 37:24: first question. Really amazing that you're going to be able to use deep whole genome sequencing,
• 37:24 - 37:30: actually in the clinic. And I was just wondering, well, actually two questions. Number one,
• 37:32 - 37:36: as everybody knows, ovarian cancer goes, it has different metastases to different anatomical
• 37:36 - 37:42: sites. So how do you take that? How will you take that into account? Are there the heterogeneity
• 37:42 - 37:48: between say a site in the peritoneum versus the omentum? And also, and maybe we can save this
• 37:48 - 37:54: for our discussion, is how exactly are you going to use that for informing the patient's care?
• 37:56 - 38:03: So nice, tough questions. I think the first thing to say is most of my talk today has been really
• 38:03 - 38:09: about trying to understand the active or ongoing mutation process. That's the starting point for
• 38:09 - 38:14: therapy. It doesn't tell us very much about the potential selection or reprogramming that can
• 38:14 - 38:21: occur during therapy. So the tools that I'm describing, how I see these being used, absolutely
• 38:21 - 38:28: at the start of therapy and sequentially onwards. But my caveat is that actually active mutational
• 38:28 - 38:35: processes are pretty constant across multi-site disease. So although there are different selected
• 38:35 - 38:40: copy number aberrations and structural variants, the underlying mutational processes are relatively
• 38:40 - 38:43: constant. And if you look at the recent clinical cancer research paper, there's
• 38:44 - 38:49: methods that we've developed to show that and in a forthcoming paper that we're just preparing
• 38:49 - 38:55: at the moment. So a single biopsy, I think, is going to be highly informative for your class
• 38:55 - 39:01: and your molecular classification. I think you're right to highlight interactions that then occur,
• 39:01 - 39:08: say, between HRD and MYC amplification or HRD and RB1 loss, for instance. And those things may not
• 39:08 - 39:14: be available immediately from the single biopsy you could get before, or the relatively finite
• 39:14 - 39:20: amount of tissue sampling you get before treatment. And those, I think, are where ctDNA and repeat
• 39:20 - 39:28: biopsy are going to become a really important part of further personalizing care as treatment goes on.
• 39:29 - 39:34: I'm going to read a couple of questions, two questions from the chat, and then we'll move on.
• 39:35 - 39:44: From Abir Mukherjee, great talk. Are liquid biopsies ctDNA comparable to fresh tissue
• 39:44 - 39:49: in terms of their predictive diagnostic value using whole genome sequencing?
• 39:50 - 39:57: So that's a great question. So I think at the moment, of course, whole genome sequencing from
• 39:57 - 39:59: ctDNA requires enough allele fraction.
• 40:00 - 40:07: for tumor material, and that's easy in the recurrent setting when patients often have
• 40:07 - 40:11: large volume disease. And if you look back to our PLOS Medicine paper, you can see
• 40:11 - 40:18: how that can be done. In the upfront setting, I think there are more challenges. But I think
• 40:18 - 40:24: the good news is that some of this is a technology question. We are now seeing the advantages
• 40:24 - 40:30: of long read genomic sequencing in being able to pull out much more specific information
• 40:30 - 40:34: about the genome, particularly the structural variants. So I'm really talking about large
• 40:34 - 40:40: solutions and rearrangements that are going to be very, very specific for a patient, and
• 40:40 - 40:44: if detected, are very sensitive markers of recurrence. So if your model is trying to
• 40:44 - 40:49: pick up patients who are recurring, I think that's good news. And also, we're going to
• 40:49 - 40:56: get much more detail about what may be key aberrations affecting targetable therapies.
• 40:56 - 41:02: Okay, I'm just going to do one more question before we move on. I'm from Natalie Sosa-Sunderman.
• 41:02 - 41:07: Thank you for your interesting talk. I have a question regarding the clinical trial with the placebo
• 41:07 - 41:13: group in the first few slides. In my limited experience, the control group in cancer research
• 41:13 - 41:18: often is another already known treatment as a companion for a new treatment. Wouldn't
• 41:18 - 41:26: it be more ethical than a placebo group?
• 41:26 - 41:34: So I'm not quite sure what is more ethical than a placebo group. I think that the difficulty
• 41:34 - 41:39: with all trials, and I think the important thing to remember is that therapies may not
• 41:39 - 41:44: make a difference and may actually make outcomes worse. And to see an example of that, you need
• 41:45 - 41:51: to look at the huge enthusiasm for getting immunotherapy checkpoint inhibitors in the
• 41:51 - 41:56: clinic, which arguably have caused much more harm for women with ovarian cancer than benefit.
• 41:58 - 42:04: It's very difficult not to do a trial without a placebo arm. And in this case, because the
• 42:04 - 42:10: standard of care for maintenance therapy was no treatment, that was the control arm.
• 42:11 - 42:17: As we go forward, and PARP is the standard of care for maintenance treatment, future trials
• 42:17 - 42:21: will look at combinations of PARP and other medicines against PARP. And I think it's always
• 42:21 - 42:28: difficult, but most trials fail, unfortunately. So SOLO1 is a great expense, but I understand
• 42:28 - 42:34: your concerns about that, but I think it really is the difficulty. And you shouldn't also forget
• 42:34 - 42:39: that many of those women will still have gone on to receive a PARP inhibitor post-trial.
• 42:39 - 42:45: So after progression, they still had opportunities to receive a PARP inhibitor on future therapies.
• 42:46 - 42:50: Okay, well, thank you for the answers. There are more questions, but we will
• 42:50 - 42:56: save those for the end of the session. And I'd like to move on. Absolutely delighted to
• 42:56 - 43:00: introduce Professor Katherine Fu, who's an Associate Professor in the Department of
• 43:00 - 43:06: Obstetrics and Gynecology, the Division of Gynecologic Oncology. She's the Director of
• 43:06 - 43:11: Basic and Translational Research at Washington University, St. Louis. Katherine's research
• 43:12 - 43:18: focuses on developing strategies for treatment resistance in ovarian and endometrial cancer
• 43:18 - 43:22: by incorporating the tumor microenvironment, including patient-derived stromal cells
• 43:22 - 43:29: and patient-derived xenograft models. Her lab takes a pathway-driven approach to identify
• 43:29 - 43:34: how to improve the sensitivity to the current treatments such as chemotherapy, the anti-angiogenics,
• 43:35 - 43:40: immunotherapy, and PARP inhibitors. She's published seminal work on two very important
• 43:40 - 43:46: ovarian cancer targets, the Axil tyrosine kinase and the discoidin domain receptor tyrosine kinase.
• 43:47 - 43:52: Katherine moves to her new faculty position at UCSF at the beginning of July. It is a tremendous
• 43:52 - 43:58: pleasure to introduce you. The title of the talk is Targeting Chemoresistant Ovarian Cancer
• 43:58 - 44:01: from Pathway to Clinic. Katherine, the stage is yours.
• 44:02 - 44:08: Thank you so much. I really appreciate being part of this wonderful series. And I hope to share with
• 44:08 - 44:14: you work that we've done in our lab to really help bring it from the laboratory to the clinical
• 44:14 - 44:19: trial setting. And so in the United States, ovarian cancer is a leading cause of death in
• 44:19 - 44:25: reproductive cancers. Although it's the 11th leading cause of new cancer cases, it is the fifth
• 44:25 - 44:30: common cause of death. So it really highlights the idea that once diagnosed with ovarian cancer,
• 44:30 - 44:35: we really need to make greater strides in how to actually improve the survival rate.
• 44:36 - 44:41: And I always like to think about ovarian cancer framing with other most common cancers,
• 44:41 - 44:46: female breast cancer and prostate cancer. You can see here that, you know, the five-year survival
• 44:46 - 44:52: rate for ovarian cancer, you know, overall is much lower than female breast and prostate.
• 44:52 - 44:57: And part of this can be due to that most ovarian cancers are diagnosed at advanced stages,
• 44:57 - 45:00: but really, again, highlights the idea that, you know, we need more targeted therapy.
• 45:01 - 45:05: And when we think about targeted therapy, we must, and Professor Brenton and Professor
• 45:05 - 45:10: Fantel mentioned this too, you know, think about it in terms of homologous recombination deficiency.
• 45:10 - 45:17: And one of the purposes for this is to understand that those patients who have tumors that are HRD,
• 45:17 - 45:21: you know, are more susceptible to PARP inhibitors. But this leads to the other,
• 45:21 - 45:26: you know, 50% that are non-HRD, in which target therapies are quite limited.
• 45:26 - 45:32: And so, our laboratory is really focused on this subset, the non-HRD. And thinking about,
• 45:32 - 45:38: you know, how do we actually target platinum resistance or treatment resistance? You know,
• 45:38 - 45:43: this is a common treatment course in which, in terms of the patient, the woman is diagnosed.
• 45:44 - 45:50: And, you know, about 70 to 80% will relapse and recur with disease. And, you know, as Professor
• 45:50 - 45:54: Brenton mentioned right now, we do classify these as platinum sensitive or platinum resistant.
• 45:54 - 45:59: And so, for those who relapse greater than six months after the completion of platinum,
• 45:59 - 46:03: this is a platinum sensitive setting. It can be retreated with platinum and other,
• 46:03 - 46:08: perhaps, PARP inhibitors and other agents. But for those on platinum resistant settings, so they,
• 46:08 - 46:11: you know, there's relapse within the six months from completion,
• 46:12 - 46:16: the options are a bit more limited. And so, our focus has been on this population of platinum
• 46:16 - 46:23: resistant. And I'll share with you our work on how Axil inhibition can improve sensitivity
• 46:23 - 46:27: to paclitaxel. And this has led to a clinical trial that's currently ongoing.
• 46:28 - 46:35: And so, the idea of receptor tyrosine kinase Axil leading to resistance has been investigated in
• 46:35 - 46:39: other tumor types. We were one of the first to look at it in ovarian cancer and a mutual cancer.
• 46:39 - 46:45: And so, the idea is that the ligand, GAS6, binds to the receptor tyrosine kinase Axil,
• 46:45 - 46:49: the membrane receptor. And this leads to a whole host of mechanisms in terms of cell
• 46:49 - 46:56: survival, cell proliferation. Additionally, Axil does regulate phagocytosis, apoptotic cells,
• 46:56 - 47:02: innate immune response, angiogenesis, and vascular integrity. And as I mentioned before,
• 47:02 - 47:07: Axil has been identified to be highly expressed in other cancer types, as you can see highlighted
• 47:07 - 47:12: there. And, you know, our question really was, you know, particularly for ovarian cancer,
• 47:12 - 47:19: no one had really looked at it in terms of treatment resistance. And so, that was a question
• 47:19 - 47:26: that we certainly had. And so, to introduce GAS6 Axil is the idea that Axil is part of the TAM
• 47:26 - 47:34: family. There are two other receptors, MER-TK and Tyro3. And GAS6 is the main ligand for this
• 47:34 - 47:42: family. And one other concept is that although Axil is a membrane receptor, it actually can be
• 47:42 - 47:49: cleaved by proteases, ADAM10 and ADAM17, and leads to soluble Axil receptor. And this is important
• 47:49 - 47:55: because I'll highlight later on in terms of one of the biomarkers we're looking into is whether
• 47:55 - 48:04: the serum soluble Axil, as well as the GAS6 in the serum, can help perhaps associate or correlate
• 48:04 - 48:10: with response to treatment. And so, this is the biology to sort of highlight how the soluble Axil
• 48:10 - 48:17: receptor can be identified in the serum of our women with ovarian cancer. And I think one main
• 48:17 - 48:23: question has always been, you know, what is better perhaps to target, the Axil receptor itself or
• 48:23 - 48:29: perhaps the ligand? And so, at times we turn to mouse models to really understand, you know,
• 48:29 - 48:35: perhaps the functional consequences and significance for this. So, this table highlights the different
• 48:35 - 48:41: phenotypes that can be affected by perhaps a triple knockout. So, this is the Tyro3, Axil,
• 48:41 - 48:47: and MER. So, this is actually knocking out the three receptors together versus each receptor
• 48:47 - 48:54: alone versus the GAS6 knockout. And I think you can see here that for those with a triple knockout,
• 48:54 - 48:59: the phenotype and the functional consequences seem to be much higher than if we were to actually
• 48:59 - 49:05: knock out GAS6 or to inhibit GAS6 on its own ligand. And by inhibiting the ligand, GAS6 on
• 49:05 - 49:10: its own, you're still contributing to inhibition of the Axil signaling pathway since GAS6 is the
• 49:10 - 49:14: main ligand for the Axil receptor. So, this is sort of, you know, highlighting, you know, in
• 49:14 - 49:20: terms of how the approach of, you know, thinking about how one can actually target a particular
• 49:20 - 49:27: pathway and thinking about, you know, perhaps pairing this with, you know, side effect profiles
• 49:27 - 49:34: and other downstream, perhaps, consequences. And one question has always been, you know,
• 49:34 - 49:39: pretty common, you know, what is the mutational rate of Axil and its other family members?
• 49:39 - 49:47: And so, this is taken from a cohort of ovarian cancer tumors from the TCGA as well as the MSK.
• 49:47 - 49:53: And you can see that in terms of high-grade serous ovarian cancers. And when it's been
• 49:53 - 49:59: sequenced, these tumors, the mutational rate for the TAM family is, you know, very low,
• 49:59 - 50:03: not very common. You can see less than 5%. And I'll show you data, though, that when we look at
• 50:03 - 50:10: the protein level and the functional significance of the inactivation of Axil at the protein level,
• 50:10 - 50:17: you'll see that there is a significant difference. And so, we asked the question, and this is a
• 50:17 - 50:21: question that I asked with Erin Rankin back when I was a PhD and she's a postdoctoral fellow in
• 50:21 - 50:27: Madhav Josh's lab. And, you know, we asked, you know, how often is Axil highly expressed
• 50:27 - 50:32: in ovarian cancer? In particular, high-grade serous ovarian cancers, which we know are, you
• 50:32 - 50:38:  know, need, you know, much more new therapy. And so, here is, you know, when we first initially
• 50:38 - 50:43: looked at this in 2010 or earlier than that, we looked at ovarian surface epithelium. And we
• 50:43 - 50:48: noted that when we looked for Axil expression, there was, you know, very, there was no Axil
• 50:48 - 50:53: expression. And we looked actually at other ovarian cancer primaries, tumor types, as well
• 50:53 - 50:59: as mental metastases. You can see that the intense brown staining of Axil expression was quite high.
• 50:59 - 51:04: And this table on your right highlights the quantitative data. And I actually highlighted
• 51:04 - 51:09: here that when we look at ovarian cancers, high-grade serous ovarian cancers that had either 2
• 51:09 - 51:15: plus or 3 plus IHC expression, so, you know, highly expressed, you can see the majority,
• 51:15 - 51:20: up to 90 percent, do have Axil expression, particularly in the high-grade serous. And
• 51:20 - 51:26: then about 100 percent of the mental metastases also express Axil. So, this led us to ask the
• 51:26 - 51:32: question, you know, how can we actually inhibit Axil and see perhaps functional significance?
• 51:34 - 51:39: And so, when I moved to Washington University in St. Louis, we asked this question in terms
• 51:39 - 51:43: of the treatment resistance. And really identifying advanced stage high-grade
• 51:43 - 51:48: serous ovarian cancers and actually looking at the baseline tumors and asked the question,
• 51:48 - 51:53: you know, if we, you know, can we identify Axil expression that's more highly expressed in those
• 51:53 - 51:58: that are going to become more chemotherapy resistant? And we actually, indeed, did find
• 51:58 - 52:03: this correlation that those women that had treatment resistant or chemotherapy-resistant
• 52:03 - 52:09: ovarian cancer had higher Axil expression by IHC compared to those with chemosensitive and compared
• 52:09 - 52:14: to those with no recurrence. And we actually looked at the correlation for progression-free survival.
• 52:14 - 52:20: We saw that those tumors that had low Axil expression had much better progression-free
• 52:20 - 52:24: survival compared to those with high Axil. And this was the same that we saw with overall survival
• 52:24 - 52:31: here. And down here, you can see the highlighted immunostaining for the chemoresistant tumors
• 52:31 - 52:36: with this intense brown Axil expression staining compared to the, you know, lower expression that
• 52:36 - 52:42: you see in the chemosensitive tumors and then really quite few in the no recurrent tumors.
• 52:43 - 52:47: So this led us to ask the biological functional significance of this. So we
• 52:47 - 52:53: turned to ovarian cancer cell lines that are known to be more high-grade serous and asked, you know,
• 52:53 - 52:58: is Axil highly expressed in these ovarian cancer cell lines and do they correlate actually with
• 52:58 - 53:04: chemotherapy resistance? And you can see here that the OVCAR3s that have really no Axil expression
• 53:04 - 53:09: nor GAS6 expression were indeed more chemosensitive. You can see with the blue
• 53:09 - 53:13: curve here, they're more sensitive to carboplatin as well as paclitaxel.
• 53:14 - 53:19: So then we asked, perhaps if we genetically activate this more highly Axil-expressing cell
• 53:19 - 53:25: line, can we induce more sensitivity to paclitaxel and chemo in carboplatin? And we indeed did find
• 53:25 - 53:31: that we could actually induce more sensitivity to paclitaxel and carboplatin with genetic
• 53:31 - 53:37: inactivation of Axil. And then we looked at PDX (patient-derived xenograft) mouse models, as well as
• 53:37 - 53:44: the human xenograft intraperitoneal mouse model. We did find that the combination of chemotherapy with
• 53:44 - 53:51: an Axil inhibitor, and I'll describe this more, which is AB500, you know, we did see actually
• 53:51 - 53:59: decreased tumor volume when combined with Axil inhibitor. So this then led us to ask, you know,
• 53:59 - 54:03: what could be the reason for this? And so we took two approaches. One was to actually
• 54:05 - 54:12: take tumors of our women who we knew were either no Evans disease or live disease or dead disease.
• 54:12 - 54:17: And we asked if we actually were to perform RNA sequencing on these tumors and then looked at the
• 54:17 - 54:22: Axil RNA expression and see if there's any correlation with perhaps resistance transporters.
• 54:22 - 54:27: We found that there was a correlation. So for those women with live disease or dead disease,
• 54:27 - 54:32: they had higher Axil alterations, and they actually correlated also with drug-resistant
• 54:32 - 54:40: transporters ABCA3 and ABCB9. We then turned to our cells that had Axil expression and those that
• 54:40 - 54:45: did not have Axil expression and asked, you know, perhaps inactivation of Axil might actually lead
• 54:45 - 54:50: to increased paclitaxel accumulation in these tumor cells. And we did find that this was the
• 54:50 - 54:55: case. So you can see in the figure on your right that those tumor cells that had Axil expression,
• 54:55 - 55:01: so the SH scrambled, you know, compared to the SH Axil, the genetic inactivated Axil,
• 55:01 - 55:05: there was an increased accumulation of paclitaxel in these tumor cells, perhaps
• 55:05 - 55:09: suggesting that one of the mechanisms for inactivating Axil and improving response to
• 55:09 - 55:17: paclitaxel could be this increase in accumulation. So then the next question was, well, how do we
• 55:17 - 55:26: develop an Axil or GAS6 inhibitor? And so with Amado Jacha and his team, you know, really the
• 55:26 - 55:32: question was how to actually engineer this exceptional stability and very high affinity,
• 55:32 - 55:38: and this was also collaborated with Jennifer Cochran. And the issue had been actually that
• 55:38 - 55:43: GAS6 and Axil, there's a really strong binding affinity. So it actually had already made
• 55:43 - 55:48: developing inhibitors to this pathway very quite challenging. And so using the yeast immunogenesis
• 55:48 - 55:55: pathway, Jennifer Cochran and her group were able to actually engineer this very high affinity GAS6
• 55:55 - 56:02: binder. So really more of a decoy Axil receptor. So it had much higher affinity for GAS6 compared
• 56:02 - 56:07: to the native Axil receptor. And it was identified in normal healthy volunteer studies, actually,
• 56:07 - 56:12: there was excellent safety, that there really wasn't any other sort of off-target effects.
• 56:13 - 56:19: And the figure on your right here highlights how this actual decoy Axil receptor, in terms of
• 56:19 - 56:25: binding to GAS6 before it can even reach the Axil native receptor on the tumor cells themselves.
• 56:26 - 56:34: And I'll highlight some data later about serum Axil again over GAS6 ratio levels as a way to
• 56:34 - 56:43: perhaps again predict response to this targeted agent. So using this receptor, we did go ahead and
• 56:43 - 56:50: run the phase 1b trial that had recently closed and has been published. And I'll kind of walk you
• 56:50 - 56:55: through this phase 1 trial and then highlight towards the end the current ongoing phase 3 trial.
• 56:56 - 57:01: So in terms of the population, this is a phase 1, so included all women with platinum-resistant
• 57:01 - 57:08: ovarian cancer with measurable disease and one to three prior lines of therapy. And really,
• 57:08 - 57:12: again, focusing on platinum-resistant, which I mentioned before, the definition of less than six
• 57:12 - 57:19: months from the last platinum agent. And this included high-grade serous ovarian cancer,
• 57:19 - 57:25: as well as high-grade endometrioid. And there were two arms in terms of the combination of the
• 57:25 - 57:31: GAS6 receptor decoy with paclitaxel or the GAS6 decoy with pegylated liposomal doxorubicin.
• 57:32 - 57:41: And this table highlights the idea that in terms of prior lines, so women with two to three prior
• 57:41 - 57:48: lines actually interestingly had received paclitaxel. So this is actually very much a patient
• 57:48 - 57:54: and physician decision in terms of which chemotherapy to combine with this inhibitor.
• 57:54 - 57:58: And so this is to highlight that the response rates that I'll show you later
• 57:59 - 58:05: really are in women with more platinum-resistant, more higher lines of treatment.
• 58:08 - 58:13: And so this is a busy table, but really to highlight here that we're actually looking
• 58:13 - 58:18: at the combination of paclitaxel with this inhibitor. You know, a lot of the side effects
• 58:18 - 58:24: that were experienced by the women were very common in terms of paclitaxel, fatigue, anemia.
• 58:25 - 58:30: And when we look at any treatment adverse events leading to death or discontinuation of the study
• 58:30 - 58:36: drug, there were none. So really highlighting that the combination from this phase 1b showed
• 58:36 - 58:42: that there was a safety. And although it's a phase 1b, we should not really be sort of highlighting
• 58:42 - 58:48: the efficacy. We did take a peek at it and did notice that those that with the combination
• 58:48 - 58:53: I'll kind of point to this smaller table here that's a bit easier to digest, that those with
• 58:53 - 58:59: the Axil inhibitor, GAS6 inhibitor plus paclitaxel, there were two women that had a complete response
• 58:59 - 59:06: and there were 26% had a partial response for an objective response rate of 36.8%. And this was
• 59:06 - 59:11: compared to the pegylated liposomal doxorubicin in which the objective response rate was lower at 16%.
• 59:12 - 59:18: And this data also had in combination with other data really led to the idea for the phase 3 trial
• 59:18 - 59:27: to focus on bataricept or the AB500 plus paclitaxel alone and not necessarily pursue
• 59:27 - 59:34: the combination with pegylated liposomal doxorubicin. And this is a spider plot looking at the women
• 59:34 - 59:40: that were treated with bataricept and paclitaxel only. And so you can see that there was a response
• 59:40 - 59:45: where you can see pretty quickly from the beginning initiation of the treatment. Again,
• 59:45 - 59:51: this is a phase 1 trial and just really highlighting a different way to look at the response.
• 59:53 - 59:59: So questions that we've sort of had some analysis wise is to understand the actual, you know,
• 60:00 - 60:04: when women are treated in this platinum-resistant setting, we know bevacizumab is given quite often,
• 60:04 - 60:09: either in the frontline setting or in the resistant or platinum-sensitive setting.
• 60:10 - 60:15: And so one question could be, you know, perhaps, you know, for women who've received bevacizumab,
• 60:15 - 60:23: do they have the same response as those who have not had bevacizumab or they have? And here
• 60:23 - 60:28: actually suggests, and this is very small numbers, but I'm only mentioning this because this actually
• 60:28 - 60:37: has influenced the way, you know, the phase 3 has been designed and thought about in the sense that
• 60:37 - 60:43: for those who had no bevacizumab, there seems to be a high objective response rate of 66%
• 60:43 - 60:50: compared to 10%. You know, the biology behind this is still ongoing, but it really was a point
• 60:50 - 60:54: more to say, gosh, you know, should we really think about this when I'm designing the phase 3 trial?
• 60:55 - 61:00: And one other piece of some analysis, and I mentioned this earlier in terms of perhaps how
• 61:00 - 61:07:  to measure and have a biomarker for response. And so looking at the serum-soluble-axel-to-GAS6
• 61:07 - 61:13:  ratio and asking whether or not this can actually correlate with clinical response to batiricept
• 61:13 - 61:18:  and paclitaxel together. And so this is to highlight that, you know, perhaps there is
• 61:18 - 61:24:  a correlation with response. And here, the soluble-axel-to-GAS6 ratio was defined as
• 61:24 - 61:31:  0.773 or higher. And these were actually suggestive of responders. And when we actually
• 61:31 - 61:38:  look at the population with those that are biomarker high, the response rate is about 47%
• 61:39 - 61:46:  compared to the entire paclitaxel-batiricept cohort, which is 35%. And so with this and the
• 61:46 - 61:52:  last slide with pembrolizumab has helped inform the phase 3 trial to, you know, with the greater
• 61:52 - 61:57:  numbers to really ask if, to really validate these findings from the phase 1 trial.
• 61:59 - 62:04:  And so in terms of the phase 3 trial, it is ongoing right now. It is an international
• 62:04 - 62:08:  randomized phase 3 trial for platinum resistant. This is including high-grade
• 62:08 - 62:14:  serous ovarian cancers with one to four priors. And I mentioned before that based on the phase 1
• 62:14 - 62:23:  data, and the inclusion criteria will, would be for those with platinum resistance,
• 62:23 - 62:31:  and then to receive weekly paclitaxel with the batiricept in Q2 weeks. And it's the day 1 and
• 62:31 - 62:37:  day 15, and kind of in the not really having holiday treatment is really important. And the
• 62:37 - 62:44:  fact that, you know, GAS6 is highly expressed and secreted by these tumors. And so really the
• 62:44 - 62:51:  continuous inhibition of GAS6 and also the Axil expression is really important in combination
• 62:51 - 62:59:  with the paclitaxel. And it's currently accruing and has been open for a little over a year.
• 62:59 - 63:05:  And so just to summarize Axil and platinum resistant ovarian cancer, Axil is highly expressed
• 63:05 - 63:09:  in tumor resistant ovarian cancers that I've highlighted, particularly in high-grade serous.
• 63:09 - 63:16:  Batiricept, you know, was designed and developed to be a novel hyphenated Axil decoy protein
• 63:16 - 63:21:  that binds GAS6 and inhibits the GAS6-Axil signaling. It has been found to be well tolerated
• 63:21 - 63:27:  with no discontinuation due to adverse events. And this was seen in both the combination paclitaxel
• 63:27 - 63:33:  as well as the normal healthy volunteer study in which normal healthy volunteers received
• 63:33 - 63:41:  single-agent batiricept. And there were very few side effects that were seen in this population.
• 63:42 - 63:46:  And there appears to be substantial evidence for the benefit of this combination. Objective
• 63:46 - 63:52:  response rate for the entire cohort was about 37%. And if we actually looked and I didn't
• 63:52 – 64:01:  have a site on the MEC-HI, but there was minimal efficacious concentration, the pharmacokinetics
• 64:01 - 64:08:  really identify, you know, what if the women were receiving appropriate concentration of this novel
• 64:08 - 64:13:  agent. And the objective response rate was higher for those that had met this minimal efficacious
• 64:13 - 64:19:  concentration. And as I mentioned before, for those with no pembrolizumab treatment, the
• 64:19 - 64:24:  objective response rate was higher at 67% compared to 10% in the pembrolizumab group.
• 64:24 - 64:31:  And this finding has really led the idea for the phase three to really make sure that we have
• 64:31 - 64:36:  women enrolled that have not been treated by pembrolizumab just to really understand if
• 64:36 - 64:41:  there really is a difference in the biology of response for those who had had pembrolizumab
• 64:41 - 64:49:  versus those who have not. And excitingly, the serum-soluble-GAS6 ratio, you know,
• 64:49 - 64:54:  whether or not these actually can identify those who would best respond. So the serum-soluble-GAS6
• 64:54 - 65:01:  ratio is obtained right before infusion of soluble of the batiricept. So this really
• 65:01 - 65:06:  might help identify, you know, those who may have an improved response to this combination.
• 65:07 - 65:10:  And as I mentioned, the phase three trial is ongoing with the first patient dosed
• 65:10 - 65:18:  a little over a year ago. And with that, I also wanted to acknowledge those in my laboratory
• 65:18 - 65:22:  that has really helped with the work. But truly, Madha Jatia and Aaron Rankin, who, when I began
• 65:22 - 65:30:  my PhD studies, really helped and really, you know, led the way with the biology of this pathway.
• 65:31 - 65:32:  Thank you.
• 65:34 - 65:42:  Well, that was fantastic. Thank you, Katherine. I have a question from Benjamin Schaefer.
• 65:43 - 65:47:  Thank you for a nice presentation. Have you ever tried using existing
• 65:47 - 65:54:  tyrosine kinase inhibitors like Futibatinib, Acalbrutinib, which were approved for lung cancer
• 65:54 - 66:02:  treatment in the context of APSL? And the little end of his question, does this make no sense?
• 66:04 - 66:12:  Yes. So, you know, APSL and EGFR, you know, certainly in lung cancer has been shown,
• 66:12 - 66:18:  you know, APSL is more highly expressed for those that are EGFR inhibitor resistant.
• 66:18 - 66:27:  In ovarian cancer, EGFR receptors and inhibition hasn't really seemed to have the same response
• 66:27 - 66:31:  as lung cancer. And perhaps it's because lung cancers have EGFR mutations and
• 66:31 - 66:39:  ovarian cancers don't have, it's not as common and prevalent. So, in terms of ovarian cancer,
• 66:39 - 66:47:  you know, we have not, but in terms of looking at it for lung cancer and APSL inhibitor,
• 66:47 - 66:51:  it has not been done with this agent in particular. But, you know, certainly,
• 66:52 - 66:57:  I think would be a great idea. Yeah. And I have a question. So, if you,
• 66:58 - 67:04:  the therapy is APSL in combination with paclitaxel, does that mean you've eliminated
• 67:04 - 67:11:  DNA damaging agents for the treatment? And does APSL, how does APSL communicate with
• 67:11 - 67:18:  the DNA damage response pathway? Sure. So, in separate data that we have recently published,
• 67:18 - 67:24:  and I didn't show just to kind of focus on the paclitaxel, we have identified that an inhibition
• 67:24 - 67:30:  of APSL can improve sensitivity to carboplatin as well. And actually, the combination of carboplatin
• 67:30 - 67:37:  plus batiricept can actually increase DNA damage that we have identified with increased gamma H2AX
• 67:37 - 67:43:  foci. We actually also looked at whether or not this DNA repair might be associated with either
• 67:43 - 67:50:  homologous recombination repair or non-homologous end joining. We do see that with immunofluorescent foci that
• 67:50 - 67:58:  there is an increase in 53BP1 foci and a decrease in RAD51. So, perhaps maybe inducing
• 67:58 - 68:05:  this BRCA-like phenotype. And then we do also see a difference in replication for perturbations. So,
• 68:05 - 68:11:  with the combination of carboplatin with batiricept or Axil inhibitor, we do see increased
• 68:11 - 68:17:  replication for perturbation, perhaps also suggesting another mechanism for this as well.
• 68:18 - 68:23:  But this was really focused on the sort of the platinum-resistant setting. So, we didn't
• 68:23 - 68:28  incorporate carboplatin in here, but I appreciate that question because it's a really exciting
• 68:28 - 68:32:  part as well, thinking about, you know, moving this perhaps even the frontline setting
• 68:32 - 68:35:  where carboplatin or platinum-based therapies are given.
• 68:35 - 68:41:  Yeah. And there was one question here, and then we'll move on. It says, is Axil,
• 68:42 - 68:48:  sort of two questions for the same person from Rohit Nagare. Apologies if I mispronounce your
• 68:48 - 68:54:  name. Is Axil expressed in embryonic stem cells considering high expression in ovarian cancer and
• 68:54 - 69:09:  chemoresistant cells? And then he asks, what is its role in cancer stem cell mediated resistance? Yes. No, these are great questions. And we ourselves have not looked into
• 69:09 - 69:16:  it. I believe others have, and I believe there is an association perhaps with it. But it's certainly
• 69:16 - 69:21:  a field that's in a question that should be further explored, but we ourselves have not done
• 69:21 - 69:29:  that. Great. Well, this was fabulous. And if folks have more questions, there'll be a chance in the
• 69:29 - 69:36:  Q&A at the end of this. So I'm going to now open this up for discussion. And I thought to
• 69:37 - 69:43:  kick this off, let me just share my screen. Just some extremely difficult questions.
• 69:45 - 69:54:  So just, I'm going to ask the panel to take on one of these, one or whatever number of questions of
• 69:54 - 70:01:  these you want to address. We've had, you know, this fantastic orthogonal approaches to this
• 70:03 - 70:08:  platinum resistance that we really, it's like what, it is one of the key unmet medical needs
• 70:08 - 70:15:  for this malignancy. And so I'd ask you, and this may be more into philosophy, but with currently
• 70:15 - 70:22:  available data, and the three of you certainly have just made major contributions. And also in
• 70:22 - 70:29:  the context of the horrendous heterogeneity, how do you see moving forward to this challenge?
• 70:30 - 70:37:  In your opinion, you know, do you, is there, is the information that we have on hand right now,
• 70:37 - 70:44:  not just from you, but from other investigators necessary, but maybe not sufficient? What kind
• 70:44 - 70:49:  of companion diagnostic would you need to determine the most suitable therapeutic for
• 70:49 - 70:55:  overcoming, targeting the carboplatin resistance? And what is it going to take to get there? And
• 70:55 - 71:01:  that's really a question directed that patients really want to know. So as I say, any of you,
• 71:01 - 71:06:  any of the speakers, and then we can also open this up to the audience. And if audience members
• 71:06 - 71:12:  would like to ask their questions live, you're very welcome to, and I'll try and navigate the
• 71:12 - 71:17:  chat as well. Well, maybe if you allow me, Wendy, I can start giving some thoughts.
• 71:18 - 71:24:  We'd love to hear, that would be fantastic. Two of the questions about heterogeneity.
• 71:25 - 71:30:  I am a bit biased because of my interest in minimal residual disease, but I actually
• 71:30 - 71:37:  see that minimal residual disease focus would be, would sort of cut down the heterogeneity
• 71:37 - 71:45:  and the complexity that we see in bulk tumor. So perhaps, you know, a focus on overcoming
• 71:45 - 71:48:  resistance mechanisms in minimal residual disease may be fruitful.
• 71:50 - 71:55:  But also the idea of a companion diagnostic, if I may address this question,
• 71:56 - 71:59:  I'd really highlight sort of, I'd like to highlight this in a kind of a,
• 72:02 - 72:08:  in two different aspects. There's one, as you very rightly say, a companion diagnostic that
• 72:08 - 72:13:  would enable us to predict whether or not a particular therapeutic would be useful,
• 72:13 - 72:20:  such as, for example, BRCA1-2 mutations in PARP inhibitors or homologous recombination repair
• 72:20 - 72:27:  scoring, as James explained earlier, in terms of, you know, in terms of PARP inhibitor sensitivity
• 72:32 - 72:38:  or other forms of markers for particular treatments. But there's also a need, I believe,
• 72:39 - 72:46:  to, for something like the Oxford classic defined EMT score to essentially, from the outset,
• 72:47 - 72:54:  put patients into prognostic groups, like also what James explained with the chromosome
• 72:54 - 73:00:  numeric aberration signatures. And if we knew in advance that these patients are in a poor
• 73:00 - 73:06:  prognosis group, then perhaps we need to think ahead about early recruitment to clinical trials,
• 73:07 - 73:14:  because I, I believe, and, you know, others may disagree with me, that the more we wait for
• 73:14 - 73:20:  second, third, fourth line treatment, the more complex the tumor will become, and the more
• 73:20 - 73:27:  difficult it will be to then, for a, something like an Axil inhibitor to, to work or other,
• 73:27 - 73:33:  or other inhibitors. So I think that early resort for clinical trials when patients still,
• 73:33 - 73:38:  you know, finishing first line or maximum second line treatment, if we can predict that they are
• 73:39 - 73:42:  in that, that traditional treatment is not going to work for them,
• 73:42 - 73:47:  then I think that would be very helpful. Terrific answer, terrific thoughts.
• 73:48 - 73:53:  So, Wendy, if I can jump in. So I just want to start saying something for the audience. I think,
• 73:53 - 74:00:  you know, to hear stories about how someone's taken a target into a phase three trial,
• 74:01 - 74:07:  all the very careful, minimal residual disease work from Ahmed and obviously Catherine in the
• 74:07 - 74:12:  phase three trial, these are amazing stories. These are really, really difficult things to do.
• 74:12 - 74:17:  So I think, you know, what, what you've heard today is a kind of vignette into, you know,
• 74:17 - 74:24:  really critical approaches to it. I guess I, I strongly agree with Ahmed. I think that if we're
• 74:24 - 74:29:  going to make a difference for the disease, we have to make it early. And it's, you know,
• 74:29 - 74:34:  unfortunately, traditional drug development is in the platinum resistant population. And we have
• 74:34 - 74:44:  to find ways of bringing this in, in my case, in neoadjuvant chemotherapy. Because I think we have
• 74:44 - 74:50:  this huge unmet need for women who are having treatment, the same treatment before and after
• 74:50 - 74:56:  surgery. And there's no interpretation the way you've heard from both Catherine and Ahmed today
• 74:56 - 75:01:  in terms of what's actually happening in these tissues and what, and what we should do. The
• 75:01 - 75:05:  problem, the challenge though, is to bring medicines into that, that group. But I think
• 75:05 - 75:11:  that's why I was emphasizing the ICON-8 data and the stable disease data. We can predict that group,
• 75:11 - 75:19:  we can understand molecularly that group, we could make relatively rapid advances. And my last
• 75:19 - 75:24:  question, which is a question for everyone interested in Ahmed's views about that, we've seen
• 75:25 - 75:33:  a lot of stuff coming out in the past year about compelling conserved pathways. So diapause is the,
• 75:33 - 75:38:  you know, flavor of the moment in terms of how some cells may go into this arrest. We've known
• 75:38 - 75:44:  for sort of three or four years about senescence pathways and SASP and what that's happening.
• 75:45 - 75:50:  I guess the question to me is how we address this question in patient studies, because
• 75:51 - 75:57:  the statistical power that we can get, notwithstanding the beautiful stuff that Ahmed
• 75:57 - 76:04:  has done by collecting these samples, means that we can't sample the complete expanse of what's
• 76:04 - 76:08:  happening across patients. And I wondered what approaches everyone else thinks we need to have
• 76:08 - 76:13:  a systematic approach going forward to finding out what those adaptation, particularly these
• 76:13 - 76:20:  non-genetic adaptation systems should be, might be, sorry, might be in patients. Because my sense
• 76:20 - 76:26:  about just taking samples at post neoadjuvant treatment, we need an awful lot of samples
• 76:26 - 76:29:  to try and see what that repertoire is and start designing a trial.
• 76:29 - 76:38:  I agree. I think another challenge is the neoadjuvant, like the post-treatment samples,
• 76:38 - 76:44:  but also being able to take multiple. I mean, I think that's part of the heterogeneity.
• 76:44 - 76:49:  And I know there's logistically some challenges. Of course, it's also with the patients.
• 76:50 - 76:57:  We try to do it laparoscopically to take multiple. But then, yes, I think that's where
• 76:58 - 77:05:  it is complicated. Because I would think ideally, we obviously have some blood marker,
• 77:05 - 77:13:  but also at the time of, let's say, they do relapse, we do take another biopsy just in case.
• 77:13 - 77:20:  And we think it probably will alter either genomically, transcriptomically, or proteomically.
• 77:20 - 77:27:  And using that all at the single-cell level are all the challenges that I think would help
• 77:28 - 77:33:  really be the key study, I would say, and with hundreds and thousands of patients with
• 77:33 - 77:40:  diversity in terms of ethnic groups and locations. But I think that would be amazing
• 77:40 - 77:44:  to be able to really finally drill down. But it would take a huge effort.
• 77:46 - 77:50:  OK. I think there were some questions, but I seem to have lost them.
• 77:52 - 77:55:  I was just wondering if, Sarah, if you can help me find them.
• 77:57 - 78:00:  I thought they were in the Q&A box, but there's nothing in there.
• 78:00 - 78:04:  I've put them in the chat for you. Let me just post them again.
• 78:05 - 78:09:  Sorry. Can you just put them in the Q&A, and then I'll put them on my big screen?
• 78:09 - 78:14:  Sorry. Unfortunately, I can't put them in the Q&A because they've been moved over. But if you
• 78:14 - 78:19:  open up your chat, it's a direct message that's come to you. These are the questions.
• 78:19 - 78:25:  Yeah, I do. They're just very small. OK. Let's see. Who have we got here?
• 78:26 - 78:30:  There's somebody called Georgette had a question for James.
• 78:33 - 78:39:  Oh, I'm not sure which bit you're looking in right now. If you scroll down a little bit further
• 78:39 - 78:45:  down, the first one says questions for James. All right. OK. I can read out the first one.
• 78:45 - 78:53:  Yeah. Well, I think I'll just let's see. OK. There's one here. This is for James.
• 78:53 - 78:56:  Are you seeing certain chromosomal alteration signatures
• 78:56 - 79:03:  correlating with distinct pathways consistently being enriched, especially in the pan-cancer
• 79:03 - 79:08:  analysis? And could this help with treatment options and recurrence?
• 79:10 - 79:16:  So that's a great question. So that we so we don't have a definitive answer yet,
• 79:16 - 79:18:  but that's what we're looking at at the moment. So in other words, do these
• 79:19 - 79:23:  different mutational processes allow tumors to explore
• 79:24 - 79:31:  any space, but obviously in the case of specific DNA selection in different ways? So, for instance,
• 79:31 - 79:39:  HRD patients, a very key feature of most of those tumors is whole genome duplication.
• 79:39 - 79:44:  And that makes sense because of double strand breaks and loss of vital genes. So you double
• 79:44 - 79:51:  your genome to make sure your tumor suppressor and haploinsufficient genes are well protected
• 79:51 - 79:58:  from that. And conversely, tandem duplicator and copy number signature six seem to be driving
• 79:58 - 79:59:  high level copy number aberration.
• 80:00 - 80:06:  So again, I'm a little bit worried about the power of these approaches. In other words,
• 80:06 - 80:11:  how many samples do we need to make sense of that? What we have got, there's an interesting
• 80:11 - 80:18:  manuscript that we have on bioarchive looking at co-occurrence of particular drivers. So
• 80:18 - 80:24:  co-occurrence between MYC and other survival factors. And I think we're still in the phase
• 80:24 - 80:29:  of trying to tease out what's predetermined, possibly limited for a cancer cell, exploring
• 80:29 - 80:36:  evolution and what and what things have to happen together for cells to survive. So great
• 80:36 - 80:41:  question. I don't have a definitive answer, but we are really looking at that at the moment.
• 80:41 - 80:47:  Maybe a segue here is a question for Catherine. Have you looked at Axil expression throughout
• 80:47 - 80:54:  the disease to look at the best and most effective therapeutic window? Very tough.
• 80:54 - 80:59:  Yes, absolutely. You know, we've really focused on the baseline tumors prior to treatment,
• 80:59 - 81:05:  mostly because that's what we have access to. And mostly because we don't tend to, you
• 81:05 - 81:10:  know, sort of standard care biopsy at the time of recurrence. I, you know, for logistical
• 81:10 - 81:16:  reasons in terms of patient comfort and also payment and, you know, just sort of all the
• 81:16 - 81:22:  system issues. So it's difficult to answer that question. But I would, that would be
• 81:22 - 81:28:  very helpful to really study intrinsic versus acquired resistance. You know, I think correlatively
• 81:28 - 81:36:  it does seem to correlate with chemo resistance and likely would be able to, but we do have
• 81:36 - 81:41:  some data, neoadjuvant pre and post neoadjuvant chemotherapy, where we do see Axil expression
• 81:41 - 81:47:  increase in those that are more chemotherapy non-responsive and neoadjuvant chemotherapy.
• 81:47 - 81:52:  So there's a hint there, but we haven't been able to do it, you know, at the time of subsequent
• 81:52 - 81:59:  relapses, but we have, we do have it in terms of neoadjuvant chemotherapy response or lack
• 81:59 - 82:09:  of response. Okay. There is a question for Ahmed. Dear Dr. Ahmed, do MRD cells have more
• 82:09 - 82:21:  lipid droplets? Great question. We have embarrassingly, we haven't done the experiment, but we have
• 82:21 - 82:31:  looked at this by qPCR at lipid genes involved in lipid metabolism by qPCR and by RNA sequencing,
• 82:31 - 82:37:  and we see the increase in those cells, but we have not done that experiment. But thank
• 82:37 - 82:44:  you for the suggestion. Okay. Well, I don't see any more questions in the chat or Q and
• 82:44 - 82:50:  A, but is anybody in the audience would like to reveal yourself? We'd love to hear from
• 82:50 - 83:07:  you and join, have anything to add to our discussion. Maybe, maybe not. I know sometimes
• 83:07 - 83:13:  you think of things after events. So we're all, you know, you have our emails and we're
• 83:13 - 83:20:  all more than happy to answer questions. I will wrap up. And I think the thing that my
• 83:20 - 83:25:  take home message from this has been, we've got these absolutely brilliant orthogonal
• 83:25 - 83:31:  approaches that have been developed, making major contribution from all three speakers.
• 83:31 - 83:38:  And I think maybe what we really need to think about is how these methods and work from other
• 83:38 - 83:45  labs, how are they going to be integrated? And we already have seen a connection Ahmed
• 83:45 - 83:49:  pointed out between Catherine and Ahmed, and I could see a connection between, you know,
• 83:49 - 83:56:  James and there. So we've, so I think this is what it's going to take. People's orthogonal
• 83:56 - 84:03:  approaches being incorporated into that one strategy that is really going to help women
• 84:03 - 84:11:  and hopefully just help us conquer carboplatin resistance. It is, as I pointed out, it is
• 84:11 - 84:15:  a key unmet medical need for this malignancy. And I think the second one, although we did
• 84:15 - 84:25:  not discuss it, is why women with HGSC have such a poor response to immunotherapy. There's
• 84:25 - 84:32:  a really important points because if they either or both of them could be overcome,
• 84:32 - 84:38:  we'd make incredible progress. And there has to be a conversation between the DNA damage
• 84:38 - 84:47:  response and response to immunotherapy. And I think that might be where integrating what
• 84:47 - 84:53:  we've heard today and moving that forward into somehow combination of exploiting the
• 84:53 - 85:00:  DNA damage defects and now reinvigorating the immune system would be the wave of the
• 85:00 - 85:06:  future. And I don't know, and the speakers, if you have anything to add, of course, we'd
• 85:06 - 85:11:  love to hear from you. And if not, this has been, you've all been amazing. It's a terrific
• 85:11 - 85:16:  roster. And I wanted to thank you personally for your time and also Abcam for giving us
• 85:16 - 85:18:  the opportunity to do this.