Dr Peter J Houghton, of Greehey Children’s Cancer Research Institute, explains why collaboration is the key to accelerating drug development for the treatment of childhood cancer.
Childhood cancer is rare, although the most recentreports suggest the incidence has increased by 13% from the 1980s to the most recent analysis in 2010.1 Worldwide, the incidence for children (0-14 years) is 140.6 per million person years and in the age range 15-19 years 185.3 per million person years. For younger patients the most common cancers are leukaemias followed by CNS tumours, whereas for the older age group the most frequent cancers are lymphomas and a group comprising epithelial cancers and melanomas. There are approximately 15,700 new cases of cancer diagnosed in patients under the age of 21 annually in the USA.
It is estimated that current multimodality therapies (surgery, radiation therapy, and intensive chemotherapy) cure around 70% of patients, with five-year event-free survival (EFS) approaching 80%.2 Most notable has been the chemotherapy-driven improvement in five-year survival for patients with acute lymphoblastic leukaemia (ALL), which has increased from 4% in 1962 to over 90%. Less dramatic but important gains have been made in other diagnoses, for example medulloblastoma (10% to 75%), Ewing sarcoma (5% to 65%) and five-year EFS for Wilms tumour; retinoblastoma Hodgkin lymphoma exceeds 85%.
Despite these advances, there is clear evidence that current therapies lead to severe long-term chronic health conditions.3-5 Cardiovascular disease and second malignancies are major contributors to life-threatening morbidity.6 Other morbidities include neurocognitive decline, as well as stroke in patients surviving CNS radiation for brain tumours. Thus, although the five-year EFS and cure rates are 80% and 70%, respectively, these statistics hide the burden that childhood cancer survivors bear as they age.
In the USA there are currently estimated to be 420,000 childhood cancer survivors, and this number is anticipated to increase to exceed half a million by 2020. In Europe it is estimated that over 6,000 childhood cancer patients die annually. Thus, it is clear that to impact survival and quality of life, new, less toxic but equally effective therapeutics have to be identified and developed.
What limits the development of new therapies for childhood cancer?
Cancer in children is multiple diseases with known or unknown genetic or epigenetic events leading to tumorigenesis. The distribution of cancers is shown in Fig. 1. (See image below) However, even for leukaemia, the most frequent cancer in children 0-14 years of age, the genetic heterogeneity is such that each type of leukaemia has relatively few patients diagnosed each year (click here for more information: https://www.ncbi.nlm.nih.gov/pubmed/23233609).7 Similarly, for medulloblastoma, a type of brain tumour that is diagnosed in around 250 patients per year in the USA, there are at least four subgroups, determined by genetic studies that impact outcome and response to current therapy.8 Consequently, because of the low numbers of patients, there is little incentive for pharmaceutical companies to develop novel therapies for these patients, and because of the small numbers of children eligible for experimental therapy, the process of drug development is slow. Collaborations between European and US groups would certainly speed trials. In the USA, the consortium of cancer treatment centres that comprise the Children’s Oncology Group (COG) is responsible for the treatment of most childhood cancers yet can only perform a limited number of phase I/II trials each year. Joint trials between US and European organisations, if carefully designed, could advance drug development more rapidly.
Can preclinical models assist in developing novel therapies?
It is estimated that between 600 and 800 entities are being developed for the treatment of cancer, although very few, if any, are being specifically developed for the treatment of childhood cancers. This raises the question of how one can develop a pre-screen that would assist in prioritising drugs that enter early clinical trials in children. One approach has been to heterograft patient tumours into immune-deficient mice (so-called ‘patient-derived xenografts’). Although the approach using PDX models of childhood cancer is not new,9 it is only more recently that this approach has been adopted as a primary method for identifying new drugs and drug combinations.10 With the development of increasingly immune-deficient mice, virtually all types of paediatric cancers, including many subtypes of childhood leukaemias, can be successfully heterografted and demonstrate growth characteristics consistent with their use in drug screening.10,11
While PDX models have been valuable in identifying novel and effective agents and combinations, the models have significant constraints.12 In terms of their predictive value, our experience has been that the models over-predict anti-tumour activity, most frequently because the drug systemic exposure in the mouse significantly exceeds the exposure that can be achieved in children. However, as most new drugs are tested in children after definition of the recommended phase 2 dose (RP2D) in adults, pharmacokinetic data for the drug and schedule are usually available, so modelling relevant human exposures in mice can be achieved. While these models have value, they are resource-limited (expensive) and, compared to assessment of sensitivity using cell culture, relatively slow. Of greater importance is that using traditional experimental designs, resource-constraints do not allow the models to encompass the clinical, genetic or epigenetic heterogeneity of each disease.
The Pediatric Preclinical Testing Program (PPTP) and subsequent PPT-Consortium has been supported by the National Cancer Institute (NCI) of the National Institutes of Health (NIH) since 2004 and has established, and genetically characterised, approximately 230 PDX and cell line-derived xenograft models. In Europe, the Pediatric Preclinical Proof of Concept Platform (ITCC P4) received funding from the Innovative Medicines Initiative 2 Joint Undertaking through the European Union’s Horizon 2020 research and innovation programme and the European Federation of Pharmaceutical Industries and Associations. Their objectives are to establish and characterise 400 new solid and brain tumours (no leukaemias) and to use these to identify novel agents. One notable difference in the approach taken by ITCC is that the objective is to develop a sustainable programme where commercial drug development will support preclinical testing. ITCC is supported in part by pharmaceutical companies, and this is a positive step, showing commitment to developing therapeutics for childhood cancer. In contrast the PPTP/C approach has been to engage industry as a supplier of drugs without a fee for screening. One incentive for pharma to use the PPTP/C is that data generated can be used in paediatric implementation plans presented to the European Medicines Agency and hence the data generated can be valuable for companies to develop paediatric plans required for drug registration in Europe.
Can the ITCC P4 and PPTP/C programmes complement each other, and what steps need to be accomplished?
Both the European and US programmes have a common goal: to identify novel effective therapeutics and, where possible, develop biomarkers that predict therapy response. The ITCC will focus on brain tumours and solid tumours representing ten cancer types, but not leukaemias. The PPTP/C has panels of brain tumours (medulloblastoma, ependymoma, glioblastoma, atypical teratoid rhabdoid tumours), as well as solid tumour panels representing kidney cancers, sarcomas (Ewing, rhabdomyosarcoma, osteosarcoma), neuroblastoma, and acute leukaemias. Additional panels may include hepatoblastoma and undifferentiated sarcoma. Consequently, there will be significant overlap between the model histologies between the two programmes. However, when one considers the genetic variation associated with a particular histology, it is clear that a large number of models will be required to recapitulate the genetic/epigenetic diversity of a particular cancer type. From the genomics aspect of characterising tumours, standard exome sequencing, RNAseq, and methylseq approaches should facilitate use of raw data files between the two groups, thus expanding the pool of genetic models available for proof-of-principle studies, and discovery and validation of potential biomarkers. Standardisation of methodologies and approaches to the analysis of data will also be issues that will be critical for harmonising how we develop therapeutics. Indeed, defining ‘tumour response’ is not trivial as no preclinical standards have been adopted that would allow cross-interpretation of drug activity.
The PPTP/C approach has been to model tumour responses using an adaptation of RECIST criteria, where partial response (≥ 50% tumour volume regression) is considered the minimum drug activity regarded as clinically meaningful. Standardisation of response criteria would greatly facilitate interpretation of results from the two programmes. Perhaps one of the critical issues, as we move forward, is how we design our experiments. The use of the ‘n of 1’ approach, where each mouse has a different patient-derived tumour, was pioneered by Novartis.13 Retrospective analysis of PPTP/C data showed that use of a single mouse (instead of eight or ten mice per treatment group) was accurate in around 80% of experiments and minimally over- or under-predicted the group response in 95% of experiments.14 Adopting this approach, and using stringent clinically relevant response criteria, will allow testing of agents against an increasing number of PDX models that more accurately simulate disease heterogeneity (e.g. 30 neuroblastomas or 30 medulloblastoma models) and allow identification of exceptional responder subgroups that will be critical in identifying and validating response biomarkers. There will always be issues that make the reproducibility of data problematic (the strain of mouse, gender, age etc.); however, harmonising experimental designs and adopting similar response criteria will make easier collaborations where confirmation of the activity of a drug, or combination, is undertaken between the respective groups.
One of the major deficiencies of the PDX approach is that these models in immune-deficient mice are not suitable for testing immuno-oncology (IO) agents, such as PD1 or CTLA4 antibodies. Although early in development, humanised mice are being developed to allow at least partial reconstitution of the human immune system that will allow evaluation of IO agents. While the current models are clearly sub-optimal, harmonisation of efforts between ITCC and PPTP/C would benefit both programmes. Addition of genetically engineered mouse models of childhood cancers in syngeneic mice, proposed in the ITCC programme, may ultimately provide models more suited to IO discovery and development.
Non-scientific barriers to accelerating drug development for treating childhood cancer
The ITCC and PPTP/C have a fundamentally different structures. The ITCC programme involves pharmaceutical companies as sponsors and ultimately will sustain the programme through providing fee for service. The PPTP/C is federally funded and provides service to pharma without any cost. Both approaches have merit: the PPTP/C is independent of pharma influence, whereas ITCC may have greater access to the pipeline of the sponsoring companies, thus potentially facilitating clinical translation. However, common to both programmes is the issue of combining drugs from different companies. Experience in screening over 30 signalling inhibitors in the PPTP/C indicates that as single agents signalling inhibitors (kinase inhibitors or receptor-blocking antibodies) have relatively modest antitumor activity12 causing relatively few tumour models to regress. Consequently, it will become imperative to evaluate drug combinations to identify effective therapies, and these agents may have been developed by different companies, which often presents almost insurmountable barriers to testing combinations. Similarly, developing a seamless interface between ITCC and PPTP/C programmes that facilitates rapid transfer of drugs between the programmes allowing for expanded testing and validation against additional ‘omically’ appropriate models should be considered in developing materials transfer agreements in the future.
- Steliarova-Foucher E, Colombet M, Ries LAG, et al. International incidence of childhood cancer, 2001-10: a population-based registry study. Lancet Oncol. 2017;18(6):719-731
- Smith MA, Altekruse SF, Adamson PC, Reaman GH, Seibel NL. Declining childhood and adolescent cancer mortality. Cancer. 2014;120(16):2497-2506
- Oeffinger KC, Mertens AC, Sklar CA, et al. Chronic health conditions in adult survivors of childhood cancer. N Engl J Med. 2006;355(15):1572-1582
- Phillips SM, Padgett LS, Leisenring WM, et al. Survivors of childhood cancer in the United States: prevalence and burden of morbidity. Cancer Epidemiol Biomarkers Prev. 2015;24(4):653-663
- Bhakta N, Liu Q, Ness KK, et al. The cumulative burden of surviving childhood cancer: an initial report from the St Jude Lifetime Cohort Study (SJLIFE). Lancet. 2017;390(10112):2569-2582
- Henderson TO, Oeffinger KC. Paediatrics: Addressing the health burden of childhood cancer survivors – improvements are needed. Nat Rev Clin Oncol. 2018;15(3):137-138.
- Mullighan CG. The molecular genetic makeup of acute lymphoblastic leukemia. Hematology Am Soc Hematol Educ Program. 2012;2012:389-396
- Ramaswamy V, Remke M, Bouffet E, et al. Risk stratification of childhood medulloblastoma in the molecular era: the current consensus. Acta Neuropathol. 2016;131(6):821-831
- Houghton JA, Houghton PJ, Webber BL. Growth and characterization of childhood rhabdomyosarcomas as xenografts. J Natl Cancer Inst. 1982;68(3):437-443
- Houghton PJ, Morton CL, Tucker C, et al. The pediatric preclinical testing program: description of models and early testing results. Pediatr Blood Cancer. 2007;49(7):928-940
- Jones L, Carol H, Evans K, et al. A review of new agents evaluated against pediatric acute lymphoblastic leukemia by the Pediatric Preclinical Testing Program. Leukemia. 2016;30(11):2133-2141
- Kurmasheva RT, Houghton PJ. Identifying novel therapeutic agents using xenograft models of pediatric cancer. Cancer Chemother Pharmacol. 2016;78(2):221-232
- Gao H, Korn JM, Ferretti S, et al. High-throughput screening using patient-derived tumor xenografts to predict clinical trial drug response. Nat Med. 2015;21(11):1318-1325
- Murphy B, Yin H, Maris JM, et al. Evaluation of Alternative In Vivo Drug Screening Methodology: A Single Mouse Analysis. Cancer Res. 2016;76(19):5798-5809
This is a commercial article that features in SciTech Europa Quarterly issue 27.