DLL3 Across Neuroendocrine Carcinomas: Why This Biomarker Matters

Dr Anja Roden introduces delta-like ligand 3 (DLL3) biology and its emerging relevance across neuroendocrine carcinomas. This video reviews neuroendocrine carcinoma classification, patterns of DLL3 expression, similarities and differences between thoracic and extrapulmonary disease, and the expanding DLL3-directed therapeutic landscape that is driving interest in biomarker assessment.

Transcript

Anja Roden, MD: Hello, everyone. I'm Anja Roden. I'm a thoracic and surgical pathologist at Mayo Clinic in Rochester, Minnesota. I, until recently, was also the Immunostains Laboratory director, and I'm really interested in biomarker testing. I was involved in validation of quite a few biomarkers, so this is really deep to my heart.

If you think about neuroendocrine neoplasms, start with the lung neuroendocrine neoplasms. We distinguish between neuroendocrine tumors, which are the well-differentiated neuroendocrine neoplasms, and the neuroendocrine carcinomas, which are the poorly-differentiated neuroendocrine neoplasms. Why do we distinguish them? We distinguish them because they do have different morphologies, they have different mitotic activity, and they do or do not have necrosis. Also, their molecular abnormalities differ, and they have very different outcomes and will react to treatment differently.

So let's start with the neuroendocrine tumors. We do use the WHO classification to describe them and to classify them. Amongst the neuroendocrine tumors—so the well-differentiated neuroendocrine neoplasms—we have typical and atypical carcinoids. According to the WHO classification, they are distinguished by their number of mitosis per 2 square millimeter. Typical carcinoids have 1 or less mitosis per 2 square millimeter, while atypical carcinoids have 2–10 mitosis per 2 square millimeter and/or necrosis. Their growth pattern is either nested or trabecular, or maybe you see pseudo-resetting. And importantly, on the cellular level, the nuclei are round to oval, sometimes slightly spindled, and they have distributed what we call the salt and pepper chromatin. And these features are distinct from what we see in the high-grade neuroendocrine carcinomas.

Amongst the neuroendocrine carcinomas, we distinguish between small cell and large cell neuroendocrine carcinomas. And that is foremost based on the morphology again. Small cell carcinomas are characterized by small- to medium-sized cells. They have a high nuclear-to-cytoplasmic ratio. They have a finely dispersed nuclear, dark nuclear chromatin. And because there is almost no cytoplasm, the nuclei are often molding. And there is high mitotic activity. According to the classification, it has to be more than 10 mitosis per 2 square millimeter, but we usually see way over 50 mitosis per 2 square millimeter, and usually also quite a large amount of necrosis. In contrast to small cell carcinomas, the large cell neuroendocrine carcinomas have quite a bit of cytoplasm. They have usually round nuclei and more conspicuous nucleoli. They also have a high mitotic activity, and they also usually have necrosis.

Now, besides these morphological features, there are also, as I said, differences in the molecular alterations. So for instance, small cell lung carcinomas in general have TP53 and RB1 mutations, which we do not see in the carcinoid tumors. In the carcinoid tumors, we sometimes see MEN1 mutations, for instance, but not TP53 and RB1.

If we think about the outcome of these patients, the patients with the neuroendocrine tumors, so the carcinoid tumors, have a much better outcome than patients with neuroendocrine carcinomas. So for instance, patients with small cell carcinomas usually only have an overall survival of a few months, while patients with carcinoid tumors, so with the neuroendocrine tumors, either do not even die of the disease or live very, very long time with the disease, even though specifically the atypical carcinoids may metastasize or recur after they have been resected.

So DLL3 is the delta-like ligand 3. And this is a Notch inhibitor, which is on small cell carcinomas specifically or high-grade neuroendocrine carcinomas. It is expressed on the tumor cells, but it is not expressed on benign cells, which makes it really an ideal target because you want to target the tumor cells with your treatment and you want to leave the benign cells alone. So DLL3 is a protein that is expressed cytoplasmic and membranous in these tumor cells. And so with immunohistochemistry, you can easily detect the DLL3 expression.

DLL3 is a predictive biomarker, foremost. And it is a biomarker to identify patients who may benefit from anti-DLL3 therapy. We don't use DLL3 for diagnostic purposes. We don't need the marker to diagnose neuroendocrine neoplasms or distinguish neuroendocrine carcinomas from neuroendocrine tumors. We have much better diagnostics for this purpose. So DLL3 immunostains is really used as a predictive marker to predict a possible outcome of patients that are treated with anti-DLL3 therapy.

Small cell lung carcinomas, as I said, is a quite deadly disease. And until recently, essentially until DLL3 came out, we had to treat them or manage them with conventional chemotherapy and/or radiation. An oncologist can talk about that much more in detail.

So can we use DLL3 expression as a marker for potential targeted therapy for extrapulmonary neuroendocrine neoplasms? This is an interesting question because even though the morphology and the neuroendocrine immunophenotype of extrapulmonary neuroendocrine neoplasms is very similar to what we see in the lung counterparts, their behavior and their molecular abnormalities and their biology is somewhat different than the neuroendocrine neoplasms of the lung. So therefore, for instance, while all neuroendocrine carcinomas in general have a high number or a high percentage of TP53 mutations, RB1 mutations are much more commonly seen in small cell carcinomas of the lung than in their counterpart in the GI or pancreatobiliary tract, for instance. Those high-grade neuroendocrine carcinomas in the GI and pancreatobiliary tract, they have other molecular alterations such as KRAS. And therefore, you may not be just able to say if small cell carcinomas of the lung express DLL3 and you can actually target the DLL3 on these lung neuroendocrine carcinomas, it will also work equally with extrapulmonary neuroendocrine carcinomas.

So clinical trials need to look at both the lung neuroendocrine carcinomas, as well as extrapulmonary neuroendocrine carcinomas, to actually ensure that the therapy has similar outcomes in those tumors because of differences in biology of these tumors.

So DLL3 was first trialed on small cell lung carcinoma. And one of the reasons is that almost 100% of small cell carcinomas in the lung actually do express DLL3. And we desperately needed some additional therapies for these patients. However, now that we do have a good marker to actually identify the tumor cells that do express DLL3, we can venture out to other tumors to see whether they also express DLL3 and therefore may be a good target for the same treatment.

So, for instance, atypical carcinoids of the lung, even though I said they belong to the neuroendocrine tumors, so the well-differentiated neuroendocrine neoplasms, some of those patients actually do experience widespread metastasis and/or recurrence. And those patients, if the tumor would express DLL3, may be also benefiting from this treatment. And there are already studies that have looked into that, and they found even though not all atypical carcinoids express DLL3, there is a good percentage of these tumors that express DLL3.

Clinical trials would have to look into the possibility whether these patients actually react the same way to the anti-DLL3 targeted therapy as small cell carcinomas of the lung. Since the biology is different of the atypical carcinoids from the small cell lung carcinoma, there needs to be some clinical trials that actually look into that possibility. And the same is true for extrapulmonary neuroendocrine tumors elsewhere, whether or not the question is whether or not they actually express DLL3. And if so, do these patients also react to the treatment similar to the neuroendocrine carcinomas?

I maybe should add that the immunostain is quite easy to interpret. And so it's really a nice and fast tool to assess tumors for expression of DLL3.

So in extrapulmonary neuroendocrine neoplasms, first of all, we should evaluate how many of these express DLL3. And clinical trials, in my opinion, are needed to see whether patients who have tumors that express DLL3 will have a better outcome when a DLL3-targeted therapy was used.

Anja C. Roden, MD, is Professor of Laboratory Medicine and Pathology and is a board-certified anatomic and clinical pathologist with special interest in thoracic pathology.  

Dr Roden received her medical training at the Humboldt University in Berlin and Technical University in Dresden, Germany. After training in general surgery and working in basic immunology science, she completed her anatomic and clinical pathology residency, surgical pathology fellowship, and pulmonary Mayo Clinic scholarship at Mayo Clinic in Rochester, US. Subsequently, she joined the staff of the Department of Laboratory Medicine and Pathology at Mayo Clinic in Rochester. Dr Roden also serves as Pulmonary Pathology Fellowship director and head of the thoracic transplant section.