Every night I read to my daughters at bedtime, then head back to the lab for a second shift. One night, my youngest daughter, then 4 and a half years old, asked, “Daddy, you leave for your research every night...but have you ever saved anyone’s life from the research you do in your lab, ever?” My answer was, unfortunately, “No. not from my research.” I could have said that nobody really ever has - the path from basic science to clinical applicability seemed too complicated to explain. But she had a valid question, and it got me thinking: wouldn’t children with cancer ask the same question, too?
Fast forward to Spring 2014, at the annual Children’s Oncology Group meeting. The Chair announced dramatic cuts in the National Institute of Health’s budget for the COG... suggesting that not every childhood cancer could have an open clinical trial: only the ones with preclinical justification. The standard approach (dropping in adult cancer drugs into trials for children) simply hadn’t worked. At the same meeting, the NCI announced that as a result of the sequester, the budget of the Pediatric Preclinical Testing Program was cut earliest, and deepest, of any NCI program (by 40 percent)!
Among colleagues who were leaders in pediatric oncology at academic centers, the feeling was that preclinical testing of basic science findings, to move exciting discoveries to clinical trials, was too tedious and narrow-scoped for university laboratories and government programs. As a result, the best and brightest scientific discoveries for childhood cancers never actually make it into the clinic. Rare cancers are the hardest hit, with survival rates remaining stagnant for decades. Knowledge that could save kids’ lives simply languishes in this black hole: the preclinical gap.
By chance, my reading material for the airplane ride home was A Life Decoded, the book by J. Craig Venter. In this story of the first group to sequence the human genome, Dr. Venter achieved remarkable speed and cost efficiency by “going outside the box” of academia. Curious, I drove straight from the airport to a biotech incubator. Renting a 250 sq ft lab space for per year: a mere $10,500. The “what if’s” began: what if we could change the existing paradigm of research grants leading to publications (leading to more grants and papers, but never tangible results)?
What if we could bridge the preclinical gap as a mission...with scientists partnering with families to achieve the cures they so desperately desired? What if science driving drugs into the clinic existed as a singular mission? My research team and I simply wanted to know how a non-profit biotech could answer my daughter’s question. The result is the Children’s Cancer Therapy Development Institute. We value our pharmaceutical partners, give parents a seat at the table, and really listen to the clinical trialists. All in the name of converting scientific discovery into clinical trials for children with rare and underserved cancers.
WOW - our thanks to the anonymous donor who is offering a 1:1 match up to $250,000 for this project!
A question families often ask is, how does lab research on embryonal rhabdomyosarcoma apply to me or my child? Our lab doesn't offer medical advice, but the question is still valid. This project's goal is to determine the major types of embryonal rhabdomyosarcoma - and determine the best treatments for each "archetype". In a children's block letter toy set, There are 26 letters. If we divide those into 5 major groups (A-E, F-J, K-O, P-T and U-Z), we can look for treatments with available and emerging new drugs that treat each group in a more "personalized" way.
Our approach is as follows:
We will use publicly available RNA (and DNA) sequencing data to divide embryonal rhabdomyosarcoma into 5 groups. We will use Pax7:Foxo1 rhabdomyosarcoma as a ERMS-similar outgroup. We’ll know that these groups cover the full breadth of embryonal rhabdomyosarcoma (eRMS) because these biopsy samples represent all children across North America who were enrolled in Children’s Oncology Group (COG) clinical trials.
What’s next? We need experimental model systems to test drugs in. One kind of a model system is a cell line. The RD cell line is very popular for studying eRMS, but it was made in 1968. That’s a long time ago! We need a broader range of cell lines to match up to the COG biopsy samples. We’ve cataloged and have in our lab a great number of other eRMS cell lines (for details, see https://www.ncbi.nlm.nih.gov/pubmed/23665679).
We also develop new primary tumor cell cultures from families willing to share tissue from surgery (and sometimes from autopsies). Since these tumor cell cultures are made in recent months or years, they are perhaps especially reflective of contemporary eRMS patients.
Can we do more? Absolutely. We want to be able to go beyond cell lines in petri dishes for testing drugs: we aim to test drugs against the different types of eRMS using mice. We can implant mice with tumors from children, adolescents and young adults. Once tumors have grown, we can treat these mice with the drugs (or drug combinations) that showed the most promise in petri dish (cell line/cell culture) studies. Many models exist that we have already generated with our partners at The Jackson Laboratory. Our collaborators at Champions Oncology have also created a number of these models. We continue to make new models as families donate tissue to this effort.
THE RESULT: Generating data as efficiently and accurately as possible, we plan to have drug – genetic matches for each type of embryonal rhabdomyosarcoma … and to publish this data, but also to share it in real time with pediatric oncologists of eRMS patients who may have scientific interest.
Why is this important?
This project is important because many children, teens and young adults with embryonal rhabdomyosarcoma need new options. When metastatic, the five-year survival rate is only 43%. We want eRMS to be universally survivable --- with the least side effects and highest quality of life after treatment.
As a grandparent recently put it,
"Childhood cancer is vastly underfunded (for a variety of reasons), even though it is a particularly tragic variant of cancer. To watch a two-year-old undergo a drastic surgery and chemotherapy is difficult. But what is really heartbreaking is when the best and most aggressive treatments are unable to save a child's life. Each child saved may have 70 years added to their lifespan, including many of the best years. Community support is urgently needed to work towards breakthroughs."
As a mother conveys,
"My son was diagnosed with an aggressive and deadly form of cancer called rhabdomyosarcoma when he was just 19 months old. When he was first diagnosed, and nearly every week since then, I have taken to reading the medical literature about rhabdomyosarcoma, looking for new treatments that could possibly help improve the odds of my son beating this cancer. Sadly, very little research is being done on rhabdomyosarcoma. First off all, there is not enough government funding of cancer research in general (and my understanding is the funding situation is getting worse and worse). Second, very little--we're talking around 4%--of funding for cancer research in the US goes to pediatric cancer. Of that limited pediatric cancer research funding, only a tiny amount goes to rhabdomyosarcoma because it is considered a "rare" cancer. (However, there are 400-500 kids diagnosed with rhabdomyosarcoma each year... they do matter.... and research on "rare" cancers can help our understanding of other cancers, too.) So, the treatments that are currently used for rhabdo are very old treatments, with high toxicity, and they only cure rhabdomyosarcoma about 60% of the time. This is terrifying and unacceptable."
Who will benefit?
Children, teens and young adults with embryonal rhabdomyosarcoma will benefit from new drug treatments, matched to a child’s tumor’s genetics, being introduced into clinical trials. Our longstanding work on eRMS also tells us that eRMS is related to certain other soft tissue sarcomas (https://linkinghub.elsevier.com/retrieve/pii/S1535-6108(10)00532-5), making it possible that both children & adults affected by eRMS and undifferentiated sarcomas will benefit from this two-year project.
BudgetThis project in its basic form runs two years at $125,000 per year ($250,000 total). The expenditure category breakdown is given below. It is supervised by Dr. Keller, and led first-hand by a biomedical engineer. The project has both bioinformatics/computational biology and wetlab experiment studies. If this crowdfunding project reaches its goal, our anonymous donor will match each dollar 1:1, doubling the funds available allowing us to hire a biologist to work alongside our biomedical engineer ... doubling the speed and throughput towards our mission find the best treatment option for each group of children with embryonal rhabdomyosarcoma.
Our engineer Noah and bioinformatics intern Chandler are busy building archetype groups across child, teen and young adult embryonal rhabdomyosarcoma. Above is a pictograph representation of their preliminary results.
Pictured is team leader, Noah, and bioinformatics interns Nick (front) and Chandler (back) working on the integration of ERMS cell lines into the Archetypes above. more soon!
Work over the past few weeks has been on "data aggregation". Specifically, we're trying to fill in the gap in the last archetype/tumor family tree diagram by adding in typical ERMS cell lines. First we get permission for the data, then we download it, then we run it through an analysis pipeline for each cell line, then we re-analyze all sequencing data (from biopsies, cell lines, GEM mouse models and PDX mouse models). We are constantly being vigilante to use the right "gene list" so that mouse and human ERMS can be compared directly. We're just about there - stay tuned!
For the family tree studies, there are a number of algorithms that can be used, each with slightly different (but important) subgrouping effects that we observe. We are currently using DNA mutations for known outgroups to pick the best dendogram (subgrouping) algorithm. Another next step, by our new engineer Cora, is to integrate Pax7:Foxo1 rhabdomyosarcomas into this study as an ERMS-related outgroup. This is an important "control group" to getting the family trees right.
from Cora (engineer project lead):
Over the past week we have worked to finalize the "database overview", "get analyzed data", and "data aggregation" portions of the project. We are ensuring that users will have the ability to download analyzed data for each sample and obtain a statistical summary of the data without any problems. Next, we will be focusing on integrating Pax7:Foxo1 rhabdomyosarcoma into the study.
Pax7:Foxo1 rhabdomyosarcomas are now integrated into the archetype! (they will serve as an 'outgroup'). As mentioned previously, there are many types of subgrouping algorithms. We are currently deciding the best method of clustering with the Pax7:Foxo1 rhabdomyosarcomas now integrated into the archetype. Once that is completed we can begin to utilize the dendogram for drug discovery.
This week has been dedicated to figuring the best subgrouping algorithm. We have added a table matching specific mutations to each sample as well as the diagnosis of each sample. By using the table, we can decipher which method of clustering groups together specific mutations of interest (Pax7:Foxo1 being one of them). We still have some work to do, but we are well on our way to finalizing the dendogram!
We have also been working on some preliminary experiments. We will be using cell lines derived from eRMS tumors to look at alterations in the cells' phenotype once applied with a small molecule. In a process known as high content analysis (HCA) staining we can label proteins with fluorescent tags and detect changes at the subcellular level using imaging. This will be an especially useful tool for drug discovery in the later stages of this project.
This week has been focused on making small changes to the dendrogram so that we can ultimately decide the best subgrouping algorithm. Here's a closer look at the program we've been working with (above). We also have been continuing our preliminary experiments on the staining protocol. We hope to identify a variety of cellular alterations, such as proliferation, apoptosis, and myo-differentiation. More to come soon!
Dr. Keller and I were lucky enough to attend a wonderful event called Golf Fights Cancer (GFC). This organization gets golfers together and raises money for cancer, and cc-TDI was their chosen mission to support for this event! Moreover, this event was in support of this very project. I was also able to see my initial inspiration for joining cc-TDI's mission- my cousin, Kiley Sullivan, who was diagnosed with stage 4 alveolar rhabdomyosarcoma in 2014. Pictured is Brian, co-founder of GFC, the Sullivan family, Dr. Keller, and myself at GFC. It was amazing and heart-warming to see so much support from so many people at this event.
In terms of the dendrogram, we are close to completing it. We are still working on choosing the best subgrouping algorithm and ensuring that we have the right set of genes that fully recapitulates mouse and human eRMS tumors.
This week has been focused on finalizing the dendrogram. We have incorporated a binary heatmap below the dendrogram (Charles and I are working on the best way to show this in a layman's version) to identify the mutations and diagnosis of each sample listed within the dendrogram. Currently, we are working on adding more information to the binary heatmap, such as demographic data (age and gender), mouse genotypes, mouse cell of origin, and human histology of each sample. Once this has been completed, we can directly compare each of the subgrouping methods and choose the best one!
We have just a few more finishing touches before the dendrogram is completed. We have added demographic data (age and gender), mouse genotypes, mouse cell of origin, and human histology of each sample to the binary heatmap. All we need to do is add a few more mutations to our gene list before we can begin using the dendrogram for drug screening on eRMS samples. Please bear with us as we move this rough draft into a polished result!
ps. a check for $100,000 arrived from Golf Fights Cancer last week ... wow, we're speechless! this is so impactful! The new drug printer is now ordered...
We are ensuring that we get the best possible dendrogram by aggregating even more data, and attempting to expand our sample list. We are working to accrue samples from all over the world and are even looking to add a collection of dog samples into our dendrogram. Our goal is to gather every last available ERMS sequence to capture the full breadth of ERMS.
In addition, we have been working on some preliminary experiments, such as HCA staining, mentioned earlier. We are employing different techniques to get the best, most accurate images. Above is a sneak peak of the program we have been working with to get these images. With the windows on the left and right-hand side, we can control several settings, such as the specific wells of a 96 well plate we are interested in imaging. The middle window displays cells expressing proteins labeled with fluorescent tags captured by the microscope.
We have added several genes to the ERMS legend, including genes that act as a marker of macrophages, genes that target new immunotherapies, and the expression of 4 genes (LAG3 TIM3 PD-L1 CTLA4) that are associated with being able to drug soft tissue sarcomas with immune therapies. All of these additions will become especially important when we begin screening drugs on samples.
I'm still working on a rough draft for this, so I don't have a picture at the moment 🙁
To compare mouse and human ERMS directly we had to be vigilant to use the right "gene list". Similar to the mouse, we have to be careful in using the right "gene list" for comparison of dog and human ERMS. Over the past week we have been working on this "gene list", and were able to add dog samples to the dendrogram. In addition, more mutations were added to the dendrogram's legend. Knowing the common variants among the samples will allow us to predict the best treatments later on.
Within the next week we will work on adding the expression of the four genes mentioned in the previous post. We will also work on aggregating even more data and analyzing more samples.
Happy Thanksgiving everyone! We are truly thankful for your support through this process.
mRNA expression of the four genes (LAG3 TIM3 PD-L1 CTLA4) associated with being able to drug soft-tissue sarcomas are now added to the dendrogram! Next, we plan on adding a secondary dendrogram to correlate the demographics at the bottom of the legend with the DNA mutations that are clustered within the first dendrogram (don't worry a picture will be up soon).
Also, a huge thank you to everyone that contributed to Golf Fights Cancer. The drug printer is up, running, and being used by many in our lab! We all appreciate your support and generosity.
This weekend I was able to attend the West Coast Sarcoma conference in Vancouver, BC, and listened in on presentations from several notable oncologists, radiologists, pathologists, research investigators, etc. from Canada as well as the US. A few of the major topics were advances in sarcoma diagnosis, treatments for fibromatosis, and current therapies and challenges of Ewing's sarcoma. Needless to say, I learned a lot on my trip and am looking forward to applying some of this new knowledge to my work at cc-TDI!
During my time in Vancouver, I was also lucky enough to be able to visit the Fishermen Helping Kids with Cancer during their annual herring sale. All of the funds go directly to the BC Children's hospital to ensure the best quality of life for children fighting cancer. The dedication and hard work that all these fishermen and volunteers put into this event was incredible to see! We're hoping we can convince them to have another sale closer to us in Portland. The first image is of Phil, his daughter (Andrea), and I at the sale. The second image shows the tons of fish that were moved by a conveyor belt and bagged by volunteers waiting close-by.
We are still making progress with the dendrogram-stay with us!
I was so honored to present my progress on this project at Shriners Hospitals for Children last week. Lately, I have been working to collect as many eRMS cell lines as possible. Once we acquire them, we can do RNA and DNA sequencing to determine which group within the dendrogram each cell line falls into. After, we can run drug screens on each of these cell lines. The response of eRMS cell lines that fall into the same group will tell us whether our dendrogram has biological significance.
We are also still working hard to optimize the high content analysis protocol. This week we tried out fish skin gelatin as our blocking reagent and were able to get cleaner results as compared to our previous protocol using 3% FBS. We're almost there!
It's the most wonderful time to blog... This week I am still working on optimizing the immunocytochemistry protocol, as well as adding that second dendrogram! [pictured: blue = DNA (dapi), green = GFP (a surrogate for cleaved caspase 3), red = Ki67 (for proliferating cells). Missing is channel for myosin heavy chain (MHC), a marker of muscle differentiation. We were working with 3 channels at a time for this optimization]
We have officially started our new year by creating our second dendrogram- check it out! Our second dendrogram will be helpful to determine whether some of our demographic data correlates with the DNA (gene) mutations.
We now have our second dendrogram axis! The first axis (horizontal) reflects biopsy samples, cell lines and mouse models. The second axis (vertical) represents demographic (clinical) features and genes.
We also have cell lines currently in culture. Once we have grown enough cells, we'll get them sequenced and add them to our dendrogram. We are still working on adding as many samples as possible to cover the full scope of ERMS.
This week I worked on wet lab experiments. Following myoblast proliferation, skeletal muscle cells (SkMc's) fused to form multi-nucleated myotubes within 72 hours. Cells were then stained with myosin 4 antibody (MF20) as a marker for differentiation. Despite achieving 50-60% differentiation (shown in the accompanying image), no clear pattern was detected with staining (data not shown). We will continue optimizing our protocol until a clear indication of myogenic differentiation is seen with staining (fresh reagents are being ordered).
Also, RMS cells are still growing in culture and will be sent for sequencing soon. We are also hoping to sequence more RMS samples from dogs, so our dendrogram will soon be expanding!
This week we have been doing a lot of culturing. We have some cell lines that are ready for sequencing and drug screening. We will put those in our dendrogram and then analyze their response to drugs to ensure that our groups have biological significance.
We received the fresh reagents for staining, so I will be staining cells with those today. Pics coming soon!
Along with preparing more cells for sequencing, we've been working on our high content analysis. After 5 days in differentiation media, skeletal muscle cells (SkMcs) were fixed, permeabilized, washed, and incubated with myosin 4 monoclonal antibody (MF20) as a marker for muscle differentiation. Some cells were incubated with conjugated MF20, and others were incubated with unconjugated MF20 overnight. Cells incubated with the latter were washed and incubated with a secondary antibody. Pictured is our nuclear counterstain in blue (DAPI), myosin heavy chain (MHC) stained with the pre-conjugated antibody in green, and MHC stained with the primary and secondary antibodies in red. The next step will be to include Ki67 as a marker for cell proliferation and cleaved caspase 3 as a marker for apoptosis.
Right now we're adding filters to view cleaved caspase 3 (CC3), Ki67, and MF20 most optimally (new filters are being ordered). Thanks to the Golf Fights Cancer we also have ordered new software for our CellInsight high content screening. This will greatly improve our high content analysis by quickly and efficiently automating quantitative cellular images. Thank you again to all of those who participated in the event! Our lab is very appreciative.
Also, thanks to our collaborators at St. Jude's and NIH we should have 9 more cell lines to add to our dendrogram very soon. Pictured is one of the cell lines, Rh6, that I am currently growing in culture until we have enough cells for both RNA and DNA sequencing, as well as enough cells to freeze down for later experiments (approx. 20 million).
For a short time, we're also doubling our efforts by working together on the GEARS project, a project similar to the building blocks project that aims to define the subtypes of *alveolar* rhabdomyosarcoma (Building Blocks is embryonal rhabdomyosarcoma). We'll be comparing data and dendrograms to ensure the best results. These are independent projects, but they start with the same data collection process and similar computational approaches. We want to make sure that patients who fall in-between embryonal and alveolar can benefit from both projects. Hence the computational synchronization at the start.
The software for CellInsight came in this week, so we spent some time setting that up. I'm also still culturing cells that will eventually be added to the dendrogram. Some cells grow a little slower than others, but with tender loving care (and patience) they will grow to the number we need for sequencing.
I also presented a paper that was published in Cell where The Cancer Genome Atlas (TCGA) did a genomic analysis of 206 adult sarcomas with a focus on 6 specific subtypes. Similar to our approach, they used cluster analysis and found distinct subsets within certain sarcomas. Proof that our concept works! In contrast they found some sarcomas to be molecularly indistinct (myxofibrosarcoma and undifferentiated pleomorphic sarcoma). There has been evidence that suggests a relationship between UPS and RMS, so there may also be a relationship between RMS and MFS, according to their analysis.
In light of Valentine's Day, as well as our love for genomic analysis, I've included a picture of adenine pairing with uracil in DNA. Happy Valentine's Day everyone!
We're still growing up two RMS cell lines, Rh7 and CCA (pictured). We also received more samples this week, including 7 more cell lines and 4 xenografts from the Children's Oncology Group (COG), 2 cell lines from ATCC, and FFPE slides from SCA in Canada. I will be putting the cell lines into culture shortly, so all of these samples can be added into our dendrogram as soon as possible.
We're almost done with the DNA analysis of some key patient sample as well. Once we upload the raw sequencing data to the server, we have to download the data and then run it through our analysis software. Each step of this process takes several hours.
All of the cells from St. Jude's are now ready for sequencing, but we still have more! I put two more cell lines in culture, Hs 926.T, an RMS cell line, and CW9019. CW9019 is one of our important Pax7:Foxo1 RMS cell lines that will serve as our ERMS-related outgroup. We'll be sequencing both of these cell lines to put them into the dendrogram, as well as doing drug screens to test their biological significance.
We've also made a lot of important changes to the dendrogram. With some edits to our code, we now have all of the Pax7:Foxo1's grouped together, as we'd expect!
We began our drug screening on some of our rhabdomyosarcoma cell lines included within the dendrogram. After preparing 3 million cell lines, such as CW9019 (pictured in culture), our WellMate adds them onto drug plates. Using our panel of 60 drugs with known activity in RMS we will determine if cells within the subgroups react in the same manner to a variety of drugs. After 72 hours of treatment, CellTiter-Glo will determine the cytoxicity and proliferation of cells by quantifying the amount of metabolically active cells in culture with a luminescent signal.
Additionally, I learned how to make a circos plot using RNA and DNA sequencing data from tumor material as well as matched normal tissue used as a comparison. Circos plots are not quite as easy to make as I had thought! We must develop a list of important, high impact mutations, fusion genes, and mutations with high copy number values. This requires sifting through a lot of data (think: 1,000's of genes).
The filters for the ArrayScan have also arrived, and we will be having training early next week on CellInsight. Soon we'll be able to complete our in depth imaging of cells using all four channels that will each identify differentiation, apoptosis, and proliferation.
Finally, we’ve had quite a few visitors and events these past few weeks. We had a very successful fundraiser playing trivia at Gigantic Brewing. Thank you to everyone who supported us! Also, we had one visitor with lots of great questions for us! She was doing her science fair project on sarcomas - pretty cool!
After preparing one last cell line, CCA, for RNA and DNA sequencing, we sent a total of 10 human cell lines for whole exome and RNA sequencing. The cell lines from COG are currently in culture, and will be prepared as our next batch to be sent for sequencing. In addition, I completed my first drug screen for CW9019 by adding CellTiter Glo to the treated cells, and used our plate reader (pictured) to quantify the amount of metabolically active cells via luminescence. The results will allow us to calculate IC50 values and determine the efficacy of each drug.
Two cell lines from COG, SMS-CTR and CB-NJR (pictured), are now ready for RNA and DNA sequencing(Patrick Reynolds at Texas Tech - you are our hero!) I also ran drug screens on SMS-CTR, CB-NJR, and CCA. I'll be applying CellTiter Glo to each plate on Thursday, and then will analyze their results. I'm also currently working on extracting RNA and DNA from the dog samples. The first part of the extraction process (deparaffinization) will be done today. After that, I will use an RNA and DNA extraction kit to isolate the RNA and DNA and preserve it for later RNA and DNA sequencing. So much more data to come!
Yet another week has passed! I did the usual- ran more drug screens on cell lines and prepared some cells for sequencing. I also continued my work from last week on isolating DNA and RNA from dog tumor samples. Pictured is an Eppendorf tube containing one of the samples preserved within an FFPE curl. We hope to get the optimal yield and concentration of DNA and RNA for downstream sequencing. This will be important for ensuring that we have reliable data for our dendrogram. We also received a QIAcube, which will automate the process of prepping DNA and RNA purification with up to 12 samples at a time.
In addition to preparing the dog tumor samples for sequencing, I started working on a manuscript that involves this very project! Currently, we have a general outline with the abstract, introduction, and a few other parts written. More progress will be made in the coming weeks!
Recently, a paper was published in Cell that found the VANGL2/RHOA signaling pathway to be an important pathway for the growth and maintenance of tumor propagating cells (TPCs). Due to their finding, we'll be adding the expression of VANGL2 and RHOA to the list of genes in our figure legend. Perhaps, this signaling pathway could be one of our endotypes, and a potential therapeutic target.
I'm still making progress on improving the dendrogram with the addition of new samples. DNA was isolated from four more dog rhabdomyosarcoma samples this week. Unfortunately, three of the four were not at the optimal yield and concentration for downstream sequencing. Next week, I'll be re-running the DNA extraction protocol to see if one of the other curls included with the sample set will give us the optimal yield and concentration. Additionally, more cell lines are growing in culture that will be sequenced and added to the dendrogram as well. One of those cell lines, Hs.926, is a kidney RMS cell line shown in the image above with a characteristic spindle-like morphology.