# STAT6 degrader plus TYK2 inhibitor for the IL-4/IL-13 and IL-22 signals in atopic dermatitis

> Two oral drugs, each assigned to a different inflammatory route, could give people whose disease single drugs leave active a next option, and give developers a clear pairing to test.

- Authors: AI3 Discovery
- Category: Patent filed
- Published: 2026-11-17
- Updated: 2026-11-17
- Language: en
- Canonical: https://ai3discovery.com/research/stat6-degrader-tyk2-inhibitor-atopic-dermatitis

> **Patent filed · AI3's own research.** The technology described here was researched by AI3 on its bio research platform BioQ, and AI3 has filed a patent application on it. We are looking for companies to develop or license it.

Atopic dermatitis is a chronic inflammatory skin disease involving itch and damage to the skin barrier. When inflammation persists, the skin can thicken and the barrier weakens.

The cause is not one signal. Several interleukins (signalling molecules that immune cells pass to one another), including IL-4, IL-13 and IL-22, act on skin cells together. Some patients do not respond well enough to existing systemic therapy, and they keep living with residual inflammation and itch.

The question we set out to answer: if the disease runs on at least two distinct signalling routes, can two oral drugs, each aimed at a different route, be paired by design rather than by guesswork?

## Why it is hard today

Current options each address one route. Injectable antibodies block IL-4 and IL-13 signalling. Oral JAK1 inhibitors (upadacitinib and abrocitinib) are FDA-approved and in use.

Three gaps stand out in the reports AI3 analysed:

- **Signal left behind.** After three months of dupilumab, its effect on IL-22-related pathways was reported to be limited.
- **Combination evidence gap.** Public reports centre on single-drug use, one drug at a time.
- **Hard to choose a partner.** Of five JAK inhibitors tested, only one reduced the IL-22 side, so the choice of partner drug matters.

So patients keep the inflammation and itch that a single drug does not reach. Clinicians and developers have no obvious next option for those whom a single drug cannot control.

## What AI3 found

AI3 designed an oral combination for atopic dermatitis that pairs the STAT6 degrader KT-621 with the TYK2 inhibitor deucravacitinib. The STAT6 degrader is meant to lower the protein at the end of the IL-4/IL-13 signal. The TYK2 inhibitor is meant to act on the IL-22 side. The invention also includes an evaluation method and a way to supply information on whether a patient is a candidate, based on an IL-22 axis marker.

AI3 did this analysis on BioQ, its bio research platform. BioQ holds 97 public biological databases, each checked against its original, indexed for direct queries and cross-linked by identifier. For this study AI3 drew on ChEMBL (measured drug-target activity), Open Targets (target-disease evidence and drug development stages), GTEx (gene expression in normal tissues), the Human Protein Atlas (where proteins and RNA are found in tissues and cells) and ClinicalTrials.gov (registered clinical trials).

The data showed three things that shaped the design:

- **The two routes differ.** The IL-4/IL-13 pathway and the IL-20-family pathway share 7 genes, and each keeps its own: 104 genes unique to the first, 19 to the second. Of the 26 genes in the IL-22-family pathway, 7 overlap with the IL-4/IL-13 pathway. This limited overlap supports AI3's rationale for giving the two drugs different roles; it is not proof of how either drug affects the other pathway.
- **On the IL-22 side, TYK2 is the key.** Of five JAK inhibitors tested, only deucravacitinib reduced the viral susceptibility raised by IL-22, by 76 ± 19% as reported. That is why AI3 chose it over other JAK inhibitors.
- **Both drugs already exist at human stage.** The STAT6 degrader has reached Phase 2, and the TYK2 inhibitor is approved for psoriasis. This describes each drug on its own, not evidence for the combination.

One caution shaped the design. The STAT6 degrader's developer describes it as blocking STAT6 function completely. AI3 therefore limited TYK2's role to the IL-22 side and built an evaluation method to examine that division of labour.

The parts of the invention are:

1. the specific pair, KT-621 with deucravacitinib, for atopic dermatitis;
2. an evaluation method that measures the added effect of a TYK2 inhibitor in keratinocytes (the main cells of the outer skin) in which STAT6 has been lowered;
3. a method of supplying information on whether to treat, based on an IL-22 axis marker.

What is new: the closest earlier work either tests single drugs or lists the two drugs as separate entries among more than a thousand co-administered drugs. AI3 found no earlier report that identifies this pair for atopic dermatitis.

![Four-step flow from two signals to an oral drug pair](/uploads/agent/p3a96dd0079-fig1-en-b7c5957f.png)

**Figure 1. From two signals to one pairing.** How the data led from two signalling routes to a specific oral pairing.

> **What AI3 found.** AI3 designed a pairing of the STAT6 degrader KT-621 and the TYK2 inhibitor deucravacitinib, assigning one drug to the IL-4/IL-13 signal and the other to the IL-22 signal in atopic dermatitis. **What is new:** The closest earlier work examined concerned single drugs or listed these two as separate items among more than a thousand co-administered drugs; AI3 found no earlier report identifying this pair for atopic dermatitis.

## How it works

In keratinocytes, STAT6 is the transcription factor (a protein that regulates gene activity) at the end of IL-4 and IL-13 signalling. The degrader removes the STAT6 protein itself instead of blocking its activity. TYK2 is a kinase (an enzyme that relays signals by adding phosphate groups to other proteins) on the IL-22 route, and the inhibitor reduces its activity.

The expected chain: STAT6 falls, so IL-4/IL-13 signalling weakens. TYK2 is blocked, so IL-22 signalling weakens. Inflammation arriving by two separate paths would then drop together.

Think of two doors into the same room. Antibodies and degraders address one. The second needs a different key, and the data points to TYK2 as that key.

## What could change if it is realised

For patients whom a single drug does not control, an all-oral pair would be a new option, taken as pills rather than as an injection plus a second therapy.

For clinicians, the IL-22 axis marker could become information on who the pair suits, so that a second drug is added where the IL-22 route is active.

For developers, a company holding a STAT6 degrader or a TYK2 inhibitor would get a specific, reasoned pairing instead of an open field of possible partners. Companies with IL-22 receptor antibodies are further potential development partners. If the added effect holds, it gives a defined starting point for co-development talks. The market size is hard to estimate at this stage.

![Comparison of single-route targeting today and the proposed two-route oral pairing](/uploads/agent/p3a96dd0079-fig2-en-0e152772.png)

**Figure 2. Targeting one route vs two.** The proposed pairing assigns each signalling route to its own drug.

## The road ahead

AI3's pairing is a computational design based on public data. The next step is cell-based validation: a cell test of whether TYK2 inhibition adds an effect beyond STAT6 degradation. Alongside it, the IL-22 marker used to select patients needs refining.

The partner this step needs holds a STAT6 degrader or a TYK2 inhibitor, or has dermatology development and keratinocyte assay capability.

> **The data behind this research.** AI3 carried out this research on [BioQ](https://bioq.bio), its bio research platform. BioQ holds [97 public biological databases](https://bioq.bio/catalog), each checked against its original, indexed for direct questions and cross-linked by identifier. For this study AI3 analysed ClinicalTrials.gov (registered clinical trials), ChEMBL (measured drug-target activity), GTEx (gene expression and its genetic control in normal tissues), IMPC and MGI (mouse model phenotypes), Open Targets (target-disease evidence and drug development stages) and GWAS Catalog (genetic associations with traits and diseases) together.

## Let's take it further together

AI3 carried out this research on its bio research platform BioQ and has filed a patent application for it. We are looking for companies to take the next step with us. If your company could turn this into a product, a service or a development program, we would like to hear from you, whether you are thinking of **joint research and development** or **licensing the technology**.

[Send an inquiry to AI3 Discovery](/partnerships?type=joint&type=license#contact) and tell us what you have in mind. We will share a detailed briefing on the research.

*This post introduces AI3's computational research. Nothing here is medical advice.*
