A unique resource for studying DCIS and progression to cancer

DCIS is the presence of abnormal but non-invasive cells in the breast milk ducts. In a small number of cases, DCIS can progress to invasive breast cancer, but there is currently no way to distinguish those cases that will progress and those that won’t. The PRECISION team, taking on our Lethal vs Non-lethal challenge, is working to find ways to make this distinction, which could spare thousands of women around the world unnecessary treatment.  

We spoke to Stefan Hutten, first author of the paper published today in Cancer Cell, to find out more about the findings from the study and how these models could help to accelerate research on the mechanisms of DCIS progression. Stefan is a PhD student in the PRECISION team, based at the Netherlands Cancer Institute. 

We wanted to develop models that could be used to track what would happen to DCIS lesions without surgery – would they progress to invasive breast cancer or not?  

We have created a collection of patient-derived xenograft (PDX) models by intraductal injection into immunocompromised mice that we can follow over a 12-month period. Using different ‘omics methods, we compared the two groups to characterise the differences observed between DCIS cases that progressed to invasive cancer and those that did not.  

The main differences we found were that DCIS lesions that did progress to invasive breast cancer overall had a higher copy number aberration burden and, in a lot of cases, were HER2 positive.  

We also looked at 3D growth patterns in our study. We saw two distinct growth patterns – replacement growth, which didn’t change the architecture of the mammary gland when the lesion grew; and expansive growth, where the lesion grew perpendicular to the ductal system and put a lot of pressure on the ducts. The latter growth pattern may eventually lead to the breakage of the duct and correlated with invasive progression.  

I think the most important thing we bring in this paper, which will further accelerate research into DCIS, is that we have developed a collection of 19 DCIS models that can be transplanted over multiple generations that can be shared with the wider research community for further research.  

Our paper provides important direction on which parameters or markers could potentially predict invasive progression in women with DCIS. We have developed a resource that could be used to further test biomarkers that we predict could be important markers in patients before proceeding with clinical trials.  

A difficulty with clinical trials for DCIS is that it can take 10-20 years to follow its progression and see the differences between patient groups that do progress and those that don’t. Using our PDX models, we can monitor these differences in a much shorter timeframe, and the information gained may provide direction on how to create better clinical trials for patients.  

I think that the models we have created vastly improve the modelling space for DCIS. Currently, in vivo DCIS work uses cell lines that don’t always have the common characteristics that most people with DCIS have. Our models still resemble the lesion of the original patient that the sample was taken from. So being able to mimic the patient lesion more directly will really help with making future findings more clinically relevant.   

In 20 years, I hope there will be a paradigm shift in how we think about preinvasive cancer and cancer risk. Currently, a diagnosis of DCIS can lead to severe psychological effects for the patient, as they perceive the risks to be similar to patients with invasive breast cancer and usually undergo the same treatment regardless of whether their DCIS would progress or not. As we and other research groups have shown, most DCIS cases will not become dangerous, and I hope in the future we will be able to better assess prognosis in DCIS patients, dividing patients into three groups. A low-risk group, which receives active surveillance; a high-risk group, which should receive the current standard of care; and a medium-risk group, where the physician makes an informed decision together with the patient about the course of action.  

This is the type of research that could only have been done with the Cancer Grand Challenges approach. Firstly, the size of our study and the researchers required for the study needed the support of an initiative of this scale. On top of this, the techniques and the ‘omics and imaging data that we had access to was only possible because we had such a large international group of people with different specialities. For example, we worked closely with Fariba Behbod (the University of Kansas Medical Centre) and were able to exchange experiences and techniques, which made it possible to create these transplantable models. Without these international collaborations, I don’t think this work would have been possible. 

The patient DCIS samples used in this study were obtained from patients that underwent surgery at the Netherlands Cancer Institute. The researchers would like to express their thanks to the patients who donated their tissue and made this research possible

Lorenzo de la Rica, Research Programme Manager at Cancer Grand Challenges, provides his insights on the importance of this study. 

Patient-derived mouse models provide information that wouldn’t be possible to obtain from humans, where treatments like surgery, radiation and hormone therapy are the current standard of care. These models make it possible to track whether DCIS lesions would progress to invasive breast cancer or not if a patient did not undergo surgery. Additionally, the models shorten the timeframes for those discoveries. 

The mouse models are a significant improvement from the models currently available, and it is a resource that could be used to further test potential biomarkers for progression and treatment outcome, for example. 

The collection also includes 19 distributable DCIS MIND models (accessible through CancerTools.org), available to researchers worldwide to further investigate the molecular subtypes of DCIS. In my opinion, having a DCIS PDX model that closely recapitulates what happens in humans provides a unique resource that will transcend the PRECISION team. It will be a Cancer Grand Challenges legacy that will potentially be used by researchers worldwide, to make discoveries for years to come. 

The research was led by Jos Jonkers (the Netherlands Cancer Institute), Jelle Wesseling (the Netherlands Cancer Institute) and Colinda Scheele (VIB, Belgium). Read the paper in Cancer Cell. 

The Cancer Grand Challenges PRECISION team is funded by Cancer Research UK and the Dutch Cancer Society.  

Image: whole-mount image of DCIS MIND model, with DCIS cells in green and myoepithelial cells in magenta, which represent the ductal tree. Image by Stefan Hutten and Colinda Scheele.