The cure is in the data
New therapeutic targets for pediatric AML
Much gratitude to Dr. Soheil Meshinchi. Any errors here are mine alone.
This article is a positive case study for the kind of breakthrough that becomes possible when the right data is collected and analyzed. I am interested in how genomics data can be used to find cures for cancer. This is the third in a series of articles all focused on one kind of cancer — pediatric AML.
In two earlier articles:
*) I wrote about the science of pediatric AML, how different it is from adult AML, the heterogeneity of the disease and the challenges thus far in incenting the commercial development of new therapies
*) I discussed how we’re not routinely collecting the data we need to fuel new discoveries for cancer.
The outcomes described here stem from the work of a passionate team of pragmatic researchers (led by Drs. Meshinchi, Molina, Loeb and others) with a deep understanding of pediatric AML. Their research has been supported by a team of amazing patient advocates. Together, they have furthered not just the science in this field, but have also discovered a potential targeted therapy for one sub-type of pediatric AML.
In a nutshell:
Over years, the team studied a rare pediatric AML caused by a fusion of two genes. Referred to as the ‘CBF-GLIS’ sub-type, this AML is refractory (hard to cure). With the analysis of deep genomics data, they discovered a receptor called FOLR-1 on the surface of the cancerous cells and showed a causal link between the CBF-GLIS fusion and the expression of FOLR-1. They also found a FOLR-1 targeted immunotherapy drug, STRO-002 (from Sutro Biopharma), that they were able to demonstrate was effective in treating CBF-GLIS in-vivo (in experiments on diseased mice). Based on these positive results, they were able to acquire STRO-002 for compassionate use for 17 patients who had exhausted other therapeutic options. The results were extremely encouraging, with 7/17 patients achieving complete remission of the disease.
This significant advance in the scientific state of the art was possible because:
- Data was collected from 1000s of patients over the years. Biobanking of samples, deep genomic sequencing, and linked clinical and outcomes data enabled researchers to test many hypotheses.
- Funding from the government, cancer non-profits, parents, and foundations enabled research in academic institutions. Lack of traditional big-pharma funding led to pragmatic approaches being investigated and tested.
- Parent-led foundations were instrumental in petitioning pharma/biotech companies for use of their investigational drugs for compassionate use in patients.
The CBF-GLIS sub-type of pediatric AML:
Physicians have known that pediatric AML is genetically very different from adult AML. Genomic fusions are more prevalent than mutations in pediatric AML.
Note in the figure below that the number of genetic fusions is predominant at ages < 5. In other words, for pediatric AML patients, it is important to focus on the genetic fusions, whereas for older adult patients, genetic mutations are predominant. So the researchers analyzed their extensive data on the various fusions detected in pediatric AML patients.
Analyzing the transcriptome of children with AML showed clusters of patients with various fusions, known to be potentially causative to the disease. The CBF-GLIS sub-type is highlighted.
The CBF-GLIS sub-type of AML is considered the most refractory sub-type of pediatric AML with a poor prognosis. 80% of babies and young children with this cancer do not respond to the initial course of treatment, called ‘induction therapy’.
The methodology for data-driven target discovery:
One of the challenges in targeted therapies for blood cancers is that if a target for a drug is present on both normal blood cells as well as cancerous blood cells, any drug used in chemotherapy will kill both types of cells. The Meshinchi lab decided to focus on cell-surface targets that were present in cancerous cells only. Not in healthy blood cells or in normal bone marrow.
I’ve tried to summarize what was many years of work by dozens of researchers into 4 high-level steps.
Step 1: Find a suitable target
Analyzing the data showed that patients with the CBF-GLIS AML also had FOLR-1 receptors on the surface of their cancerous cells. These receptors were not present in normal red blood cells or in normal bone marrow.
In the figure below, we see that CD34+ cells (human bone marrow stem cells) as well as normal marrow, do not express FOLR-1.
Step 2: Verify target causality
The next step was to show a causal link between this fusion and the expression of the FOLR-1 receptor.
The team started with transducing cord blood stem cells (CD34+ in the figure below), with the CBF-GLIS fusion transcript and cultured these in-vitro. This led to these transduced stem cells becoming AML cells, and in 12 weeks, all the cancerous cells expressed the FOLR-1 receptor on their cell surface.
Step 3: Identify target-specific therapeutics and test on mouse models
The team acquired Sutro Biopharma’s drug STRO-002, which is a targeted FOLR-1 immunotherapy. Sutro has an ongoing phase-1 clinical trail where they’re testing the drug on a set of adult cancers (including ovarian cancer).Note that STRO-002 was not developed for pediatric AML. However this kind of “tissue-agnostic” drug repurposing based on genomic data is exactly the kind of therapeutic breakthrough that is possible with a data-driven approach.
The team showed once they induced CBF-GLIS AML in mice, STRO-002 could clear the disease in mice.
In the figure below — on the left are untreated mice and on the right, those treated with STRO-002 (at 2 different doses). Note that as time goes on, the mice on STRO-002 have few if any blue splotches on imaging, which means the CBF-GLIS AML is no longer detectable. Whereas, all the untreated mice die.
Step 4: Compassionate use on patients
Through Sutro Biopharma’s compassionate use program, the team was able to acquire the drug for 17 children with the CBF-GLIS sub-type of AML, who were relapsed and refractory and exhausted other therapeutic options. 7 patients obtained complete remission of the disease (aka ‘MRD negative’). 9 had partial response or stable disease and 1 had some residual disease.
At the time of seeing the data (a few months back), the patients tolerated the treatment well and clinical response was seen in patients even if their initial status was that they’d relapsed after a prior stem-cell transplant.
So, what’s are potential next steps?
We know that monotherapies (i.e. giving a patient just one drug) does not work with pediatric AML. So STRO-002 would need to be incorporated as a modification to the existing treatment regimen. Hopefully, a randomized control clinical trial where patients with the CBF-GLIS fusion are randomized to either the standard of care or one where STRO-002 is added to the regimen. That’s not happened yet — but there is hope that it can happen in the future.
Another option is to construct target specific immunotherapies, via technologies like CAR-T.
Scientific advances in cancer research have always depended on the long and hard work of researchers and patient advocates, as they incrementally further the understanding of the disease and potential cures. I am always reminded that my child is alive today because of the work of prior decades of research by thousands of people.
The next wave of advances will all derive from data.
Deep biological data is key to new cures — whether it is toward new drug discovery, use of drugs in the clinical development pipeline, or drug repurposing (i.e. drugs approved for other indications). And it is never a single magic bullet for a disease. Progress is made in steps and compounds with each discovery. And all progress is driven by mission-oriented research teams and armies of dedicated patients and advocates.
Thanks for reading!
References:
- Distinct age-associated molecular profiles in acute myeloid leykemia defined by comprehensive clinical genomic profiling
- Targeting FOLR1 in high-risk CBF2AT3-GLIS2 pediatric AML with STRO-002 FOLR1-antibody-drug conjugate
- CBFA2T3-GLIS2 oncogenic fusion is sufficient for leukemic transformation
- CBFA2T3-GLIS2 model of pediatric acute megakaryoblastic leukemia identifies FOLR1 as a CAR-T target
- Anti-Leukemic Activity of STRO-002 a Novel Folate Receptor-alpha- Targeting ADC in Relapsed/Refractory CBF2AT3-GLIS2 AML
- Project Stella
