Preclinical Data Suggest Novel Molecule Enhances Immunotherapy Response in CRC

Jessica Hergert
Jessica Hergert

Assistant Editor, OncLive®
Jessica joined the company in August 2019 and is one of the point contacts for the OncLive On Air™ podcast. She is a Rider University alumna and holds a degree in journalism and biology. Prior to joining MJH Life Sciences, she interned with the Ireland-based social media monitoring agency Olytico and served as a copy editor and writer for The Rider News. Email:

Partner | Cancer Centers | <b>Columbia University Medical Center</b>

Timothy Cragin Wang, MD, discusses the advent of immunotherapy and how it has revolutionized multiple areas of cancer treatment; however, the transformative potential of the modality has been mild in the field of colorectal cancer.

The advent of immunotherapy has revolutionized multiple areas of cancer treatment, said Timothy Cragin Wang, MD, who added that the transformative potential of the modality has been mild at best in the field of colorectal cancer (CRC).

“In particular, apart from the microsatellite instability–high (MSI-H) and mismatch repair deficient (dMMR) tumors, most sporadic CRC has shown no response to checkpoint inhibitor therapy,” said Wang. “It is thought that many solid tumors could or should respond [to immunotherapy] under the right conditions.”

During the 2020 AACR Virtual Annual Meeting II, preclinical findings showed that fusing the trefoil factor 2 (TFF2) anti-inflammatory peptide with 2 carboxyl-terminal peptides (CTP) and 3 Flag motifs alone or in combination with a PD-1 inhibitor led to a marked reduction in tumor growth in CRC mouse models compared with PD-1 inhibition alone.

Taken collectively, the findings from the study suggest that combining TFF2-CTP with checkpoint inhibitors could enhance response rates to immunotherapy in CRC, potentially through suppression of myeloid-derived suppressor cells (MDSCs).

In an interview with OncLive, Wang, the Dorothy L. and Daniel H. Silberberg Professor of Medicine, chief of the Division of Digestive and Liver Diseases, and co-leader of the Tumor Biology and Microenvironment Program at the Herbert Irving Comprehensive Cancer Center at Columbia University Medical Center, discussed the current state of immunotherapy in CRC, the rationale to evaluate a novel TFF2-CTP molecule, and how early findings have the potential to impact the future of CRC treatment.

OncLive: Could you provide insight on the challenges of utilizing checkpoint inhibitors in CRC?

Wang: Checkpoint inhibitor therapy has been a huge breakthrough, leading to great responses in melanoma, non–small cell lung cancer, renal cell carcinoma, and even some cases of gastric cancers. However, several tumors, including pancreatic cancer and CRC, have shown themselves to be quite resistant to immunotherapy. 

We believe the reason these tumors do not respond is because of immune resistance, which is the accumulation of immunosuppressive cells. These could be regulatory T cells,

MDSCs, macrophages, or fibroblasts. [In all cases], the tumor microenvironment is thought to promote this resistance to some extent.

Our lab has been particularly focused on MDSCs, and the idea that if we can decrease that particular immunosuppressive cell compartment, tumors will respond better. We’ve seen promising results in a number of different models.

Could you shed light on the TFF2-CTP molecule that was evaluated in this study?

One of the therapies we have been looking at is a naturally occurring molecule called TFF2. Originally, this was called spasmolytic polypeptide for a variety of reasons, but it is known as TFF2 today. It is a small, cloverleaf-shaped protein that is very stable. It is secreted mainly by mucous cells in the gastrointestinal tract and elsewhere. We found that it is expressed by memory T cells in the spleen and elsewhere, and it appears to regulate the production of immature myeloid cells/MDSCs.

Interestingly, we also found that when we knocked out this gene, we saw more rapid tumor development and an accumulation of MDSCs within and around the tumor. When we suppress these cells—perhaps by overexpressing TFF2 transgenically or delivering it through an adenovirus—we [see] smaller tumors and less robust tumor growth.

Therefore, we have been trying to develop this into a protein molecule that could be delivered therapeutically. We came up with TFF2-CTP, which is TFF2 with the addition of a CTP tag. The CTP tag seems to stabilize the protein because, otherwise, it has a fairly short half-life. I’ve been trying to develop this first through adenoviral delivery and then through recombinant proteins, given exogenously to mice mainly by intraperitoneal injection. That is what we tried to do in this particular study.

What was the rationale for evaluating the role of PD-L1 inhibition in this setting?

In this particular study, we also looked at the effect of PD-L1 expression. Something we noticed in our TFF2 knockout mice is that in the absence of TFF2, there was an accumulation of MDSCs that highly expressed PD-L1. Therefore, PD-L1 served as sort of a marker for the most suppressive form of MDSCs. That was interesting because, of course, that is one of the major markers or targets for immune checkpoint therapy.

To mimic this in normal mice, we generated a knockin mouse model where PD-L1 was overexpressed in a targeted way. In this model, there is a lox-stop-lox—not to get too technical—signal in front of the PD-L1 reading frame. Through Cre targeting, we could express PD-L1 either in the epithelial compartment or in the myeloid compartment. Interestingly, PD-L1 was originally recognized by all the large pharmaceutical companies as a biomarker for tumors that would, in theory, be more responsive to anti–PD-1/PD-L1 therapy. It hasn’t been as much of a biomarker in mouse models. We’ve tried to mimic the human condition more by overexpressing it. In this particular study, we decided to use mice that overexpressed PD-L1 in the myeloid compartment because we felt that that was a good model.

We were also testing whether that would accelerate tumorigenesis. One idea is that PD-L1 is one of the checkpoints for the development of tumors initially and that many tumors may be initially suppressed by the normal T-cell response in the body. The immune surveillance in the body would lead to the suppression of most cancers that develop until they are able to express PD-L1. In fact, we found in this and other studies that PD-L1 accelerates the development of tumors in response to various carcinogens, such as azoxymethane (AOM).

We generated these PD-L1–positive mice and found that they develop tumors more rapidly. Therefore, we thought it was a good model to also test the effect of TFF2-CTP in [combination with PD-1 inhibition].

What was found from this preclinical work?

In this particular study, we used this mouse model of colon carcinogenesis called AOM/dextran sodium sulfate (DSS). AOM was the initiator or the mutagen, and DSS caused a form of colitis or injury to the mice that was a promoting agent for tumorigenesis.

Using this combination, we examined immune changes that occurred. It hadn’t been well studied in terms of the changes that occur in myeloid cells and T cells during carcinogenesis in a colon mouse model. We confirmed that during this process of carcinogenesis of the colon, that MDSCs markedly increased during tumor progression, which got bigger and more advanced. There was also an increase, to some extent, in regulatory T cells and, more importantly, a decrease in CD8-positive T cells relative to regulatory T cells during tumor progression.

This confirmed that one reason tumors were developing is that the CD8-positive T-cell response was not robust enough and was being inhibited by both MDSCs and regulatory T cells.

Early on, we also found that overexpressing PD-L1 in myeloid cells promoted the development of early colorectal tumorigenesis with larger and earlier developing tumors. This again indicated that immune suppression was part of the normal process of colorectal carcinogenesis.

We began to try to reverse this by giving TFF2 in various ways. In the first part of the study, we looked at the effect of overexpressing TFF2 transgenically. To do this, we targeted TFF2 to the T cells in the spleen using the CD2 promoter. We also showed that by knocking out the TFF2 gene, we saw much larger tumors. This showed that TFF2 is something that suppresses colorectal carcinogenesis.

Finally, we began with our therapeutic molecule, TFF2 hooked up to a stabilizing peptide, CTP. We gave either TFF2-CTP alone, anti–PD-1 therapy alone, or the 2 agents in combination. Interestingly, TFF2-CTP alone was somewhat effective, even more so than anti–PD-1 alone. The combination in particular was very effective in PD-L1–overexpressing tumors in suppressing the growth and development of colon tumors in this AOS/DSS model.

When we began to look at the correlates of response to this combination, TFF2-CTP plus anti–PD-L1, we found it was an increased CD8-positive T-cell response and a decrease in the number of MDSCs. This indicates that by knocking down MDSCs, [we could derive] a stronger response by CD8-positive T cells.

Although this research is in early stages, what could the future look like with this agent?

This study was sponsored by Tonix Pharmaceuticals, and the intellectual property has been licensed to them. I’ve been working with them to see if we can develop TFF2-CTP as a biologic therapy for patients with cancer, perhaps initially for patients with CRC. The company has been working to produce this [agent] in a larger amount, so we are trying to get all the pieces together. I would love to see a phase 1 study with TFF2-CTP. Eventually, I would love to see the agent brought to the clinic and used as a possible combination therapy with checkpoint inhibitors.

Could TFF2-CTP have efficacy in other tumor types beyond CRC?

The idea is that it would be most effective in tumors with high levels of PD-L1 expression and tumors where MDSCs or an inflammatory component play a major role. I think that describes many digestive cancers. So, in theory, [TFF2-CTP could have efficacy in] liver cancer, gastric cancer, esophageal cancer, some subsets of pancreatic cancer, and other inflammatory-associated tumors.

Certainly, if there is already a signal within a cell response, checkpoint inhibitor therapy would be the most promising [treatment modality]. Again, those tumors that have shown little neural signaling may require different targets in order to make immune therapy effective. I’m optimistic that it could be effective in more than just CRC, but rather many different solid tumor types.


Woosook K, Fu N, Duddempudi P, Dubeykovskaya Z, et al. Abstract 6640: stabilized recombinant trefoil factor 2 (TFF2-CTP) enhances anti-tumor activity of PD-1 blockade in mouse models of colorectal cancer. Cancer Res. 2020;80(16). doi:10.1158/1538-7445.AM2020-6640