YIV-906: From Ancient Formula to Cancer Adjuvant
In 2010, a paper published in Science Translational Medicine reported something unusual. Researchers at Yale had taken a gastrointestinal formula first written down in China roughly 1,800 years ago, standardized it under pharmaceutical-grade manufacturing conditions, and demonstrated in animal models that it could substantially reduce the intestinal damage caused by irinotecan — one of the most widely used chemotherapy agents in colorectal cancer — while simultaneously enhancing its antitumor activity. The formula had four ingredients. None of them were irinotecan. None of them, administered alone, produced the same result. The effect required the complete polychemical system.
That formula is Huang Qin Tang, a classical TCM preparation historically used to treat diarrhea, nausea, vomiting, and abdominal spasm. Its pharmaceutical-grade derivative, developed by Yung-Chi Cheng at the Yale School of Medicine, is now known as YIV-906 (also referred to as PHY906 or KD018). It is composed of four herbs: Scutellaria baicalensis Georgi (skullcap), Paeonia lactiflora Pall. (peony), Glycyrrhiza uralensis Fisch. (licorice), and Ziziphus jujuba Mill. (jujube), combined at a fixed weight ratio and prepared to consistent chemical and biological specifications.
Over the past two decades, YIV-906 has been evaluated in more than nine Phase I/II to Phase II clinical trials across institutions including Yale University, Stanford University, the UPMC Hillman Cancer Center, City of Hope, and Memorial Sloan Kettering. More than 250 patients with liver, pancreatic, colorectal, and rectal cancers have been treated. It has received FDA orphan drug designation for hepatocellular carcinoma. A randomized, double-blind, placebo-controlled Phase II trial in HBV-positive liver cancer patients was actively enrolling at 22 study sites across the United States, mainland China, Hong Kong, and Taiwan.
YIV-906 is not a cure. It is not a standalone cancer treatment. What makes it scientifically significant — and what makes it a useful case study for the principles introduced in this series — is what it does as an adjuvant: it appears to make chemotherapy work better, hurt less, and do both simultaneously. Unpacking how requires engaging with several layers of biology that the single-target model was not designed to address.
Two Problems, One Formula
Cancer therapy faces a structural tension that has never been satisfactorily resolved. The agents most effective at killing tumor cells tend also to damage healthy tissue. For the gastrointestinal tract, this problem is acute: the same rapid cellular turnover that makes intestinal epithelium sensitive to chemotherapy is the reason the gut maintains its integrity at all. Drugs that hit dividing cells hard will hit gut lining hard, regardless of where the tumor is. Irinotecan, which works by inhibiting topoisomerase I in cancer cells, produces severe GI toxicity — diarrhea, nausea, mucositis — in a significant proportion of patients. So does capecitabine. So does radiation.
The conventional approach to this problem is dose reduction: accept less efficacy to avoid intolerable toxicity. What YIV-906 appears to offer is a different trade — one in which GI protection and antitumor enhancement are not alternatives but simultaneously achievable ends. This is not intuitive. An agent that protects normal tissue from chemotherapy might be expected to also protect tumor tissue. YIV-906 seems to do the opposite: it selectively modulates the tissue responses in ways that favor normal recovery in the gut and unfavorable outcomes for tumors.
The mechanistic explanation for why this selectivity exists goes to the heart of what a multi-target, polychemical system can do that a single compound cannot.
The Gut: Anti-Inflammatory and Regenerative
In the intestine, irinotecan's damage proceeds through a cascade of inflammatory signaling. YIV-906 was shown to interrupt this cascade at multiple nodes simultaneously: it inhibits NF-κB, COX-2, iNOS, and IL-6, all of which are activated during chemotherapy-induced gut inflammation. But suppressing inflammation is only part of the story. The gut also needs to repopulate its damaged epithelium, and YIV-906 was found to potentiate the Wnt signaling pathway — promoting the repopulation of intestinal stem and progenitor cells in the injured mucosa. Protection and repair, operating through different mechanisms, delivered by the same formulation.
Clinical data have broadly supported this picture. Across trials in colorectal, pancreatic, and hepatocellular cancers — using irinotecan, capecitabine, sorafenib, and chemoradiation — patients receiving YIV-906 consistently showed reduced grade 3 and 4 GI toxicities. In the Taiwan Phase II trial of PHY906 plus capecitabine in advanced hepatocellular carcinoma, the incidence of severe nausea and emesis was lower than with capecitabine alone, and quality of life scores did not deteriorate significantly from baseline during the combination therapy. Only two patients discontinued treatment due to adverse effects. This GI protection profile has been observed across cancer types and chemotherapy classes, suggesting the mechanism operates independently of the specific cytotoxic agent involved.
The Tumor: Inflammation Inverted
In normal tissue, YIV-906 dampens inflammation. In tumor tissue, it appears to do something almost opposite: it converts tumors from immunologically "cold" — largely ignored by the immune system — to "hot," characterized by immune infiltration and active cytotoxic pressure. This is not a contradiction. It reflects a tissue-specific and context-specific biology that a single-target agent is poorly equipped to navigate.
The sorafenib work, published in Scientific Reports in 2015, illustrated this clearly. Sorafenib alone reduced HepG2 liver tumor growth by about 35% in mouse models. The combination with YIV-906 completely inhibited tumor growth and drove tumor shrinkage — an effect that sorafenib at double or quadruple the dose could not replicate. The mechanism centered on the tumor microenvironment. YIV-906 induced upregulation of MCP-1, a macrophage-recruiting chemokine, drawing immune cells into the tumor. Once there, those macrophages polarized toward the M1 phenotype — the tumor-rejecting, pro-inflammatory phenotype — rather than the M2 phenotype, which supports tumor growth. When macrophages were depleted using clodronate liposomes, YIV-906's enhancement of sorafenib's antitumor activity was substantially reduced, confirming that the macrophage infiltration was causally central to the effect.
Component analysis confirmed that the full formula was required. Removing the S herb (Scutellaria baicalensis) or the P herb (Paeonia lactiflora) each eliminated the synergistic activity with sorafenib. No single herb delivered the combination's effect. This is a recurring finding across the YIV-906 literature: the formula as a whole does what no subset of it can.
The Immune System: A New Frontier
The immunology of YIV-906 has become increasingly well-characterized, and it opens up a dimension of the formula's action that was not anticipated from its original GI indication.
Work published in 2023 by the Cheng laboratory demonstrated that YIV-906 enhances T cell activation through a mechanism centered on SHP1 and SHP2 — two phosphatase enzymes that normally suppress T cell receptor signaling. When PD-1 on a T cell binds PD-L1 on a tumor cell, it recruits SHP2 to dephosphorylate the downstream signaling cascade, effectively switching off the T cell's anti-tumor activity. This is the molecular basis of one of cancer's most clinically significant immune evasion strategies — and the target of the approved anti-PD-1 checkpoint inhibitor antibodies.
YIV-906 was found to inhibit both SHP1 and SHP2 enzymatic activity, with the S herb again playing the primary role. The result is that the TCR downstream cascade — Lck, Zap70, LAT, Fyn, Pyk2 — remains phosphorylated and active even in the presence of PD-1/PD-L1 interaction. NFAT, the transcription factor that drives T cell activation, stays elevated. CD69, a T cell activation marker, is upregulated. The depressed T cell activity caused by PD-1/PD-L1 signaling is rescued.
Critically, this mechanism does not depend on the T cell receptor itself. When TCRαβ was knocked out using CRISPR/Cas9, YIV-906 and the S herb continued to promote NFAT activity with similar magnitude, confirming that the formula acts directly on signaling downstream of the receptor. This means that the same SHP1/2 inhibitory mechanism that rescues PD-1-suppressed T cells should also enhance CAR T cell function — and in cell culture models using CAR T cells targeting CD19, YIV-906 was shown to increase NFAT-driven killing activity on CD19-expressing tumor cells, including under conditions of PD-L1 overexpression.
The clinical significance of SHP1/2 dual inhibition is worth pausing on. Most SHP2 inhibitors currently in clinical trials are selective for SHP2, but SHP1 can compensate for SHP2 when the latter is knocked out or inhibited. An agent that inhibits both — as YIV-906 appears to do — may be better positioned to overcome the immune resistance that arises from this redundancy. Combined with YIV-906's documented ability to modulate IDO activity, reduce monocytic MDSCs in the tumor, and promote M1 macrophage polarization, the formula operates across multiple immune suppression axes simultaneously.
The Formula Across Cancer Types
One of the striking features of the YIV-906 literature is the breadth of its activity across different cancer types and chemotherapy regimens. The same formula — manufactured to the same standard — has shown adjuvant activity in combination with irinotecan-based regimens in colorectal cancer, capecitabine in pancreatic and hepatocellular cancers, sorafenib in liver cancer, and chemoradiation in rectal cancer. This is not what one would expect from a targeted agent with a specific molecular mechanism. It is what one would expect from a system that modulates the tumor microenvironment and host immune biology in ways that amplify a broad range of cytotoxic insults.
The basic science corroborates this breadth. The mechanisms identified — NF-κB and COX-2 suppression in gut tissue, MCP-1 upregulation and M1 macrophage recruitment in tumors, SHP1/2 inhibition in T cells, Wnt pathway potentiation in intestinal progenitors, ERK1/2 phosphatase inhibition in liver cancer cells — are not cancer-type-specific. They operate on biology that is shared across the oncological landscape. A formula that modulates this biology distributed across multiple compounds and multiple pathways does not need a different mechanism for each indication. Its breadth follows from its architecture.
More recent work confirms the formula is still yielding new mechanistic insights. Studies using the colitis-associated colorectal cancer model have shown that Huang Qin Tang intervenes at each stage of the inflammation-to-cancer progression: reducing IL-6, IL-1β, and TNF-α during the inflammatory phase; restoring amino acid homeostasis — particularly glutamine, glutamic acid, and arginine — during proliferation; and suppressing the PI3K/AKT/mTOR pathway that drives tumor cell growth and survival. In NSCLC models with EGFR-TKI resistance, the formula attacks cancer stem cell markers by suppressing the STAT3/GPX4 axis and raising reactive oxygen species levels, sensitizing resistant cells to targeted therapy. In colorectal cancer liver metastasis, network pharmacology and single-cell sequencing have identified GOT1, CYP1A2, and CA2 as hub targets through which the formula may disrupt glutamine metabolism and reshape the immunosuppressive tumor microenvironment. The same four herbs. Different cancers. Different mechanisms — and yet a coherent underlying logic of network-level modulation.
What the Clinical Data Actually Show
It is worth being precise about what the clinical evidence does and does not establish. The Phase II trials completed to date are relatively small, mostly non-randomized or single-arm, and conducted in late-stage patients with limited treatment options. They have consistently demonstrated a favorable safety profile and reduction in GI toxicities. They have shown encouraging survival signals — particularly in the HBV-positive, systemic-therapy-naive subgroup across the U.S. and Taiwan HCC studies, where the combined median overall survival reached 16.5 months in evaluable Asian patients, compared to the 6.5 to 10.7 months typical of sorafenib alone in comparable populations. But they have not yet delivered the definitive randomized efficacy data that would establish YIV-906 as a standard adjuvant.
The pancreatic cancer results are also notable in context. Advanced pancreatic cancer is arguably the hardest oncological setting in which to show clinical benefit: patients who have already failed gemcitabine face a median survival of roughly two months on best supportive care, and there is no established standard second-line therapy. In the first Phase II trial to evaluate any herbal formulation in gemcitabine-refractory pancreatic cancer — 25 patients at Yale Cancer Center treated with YIV-906 plus capecitabine — median overall survival for the full intent-to-treat cohort was 21.6 weeks, rising to 28 weeks for the 20 patients who completed at least two cycles. Six-month survival was 44 percent. Two patients achieved confirmed partial responses with survival lengths of 69 and 84 weeks. Quality of life scores, tracked using the Edmonton Symptom Assessment System, improved meaningfully from baseline across fatigue, appetite, nausea, and sense of well-being, with patients who remained on study longer continuing to maintain those gains.
The pancreatic cancer trial also yielded an unexpected finding from correlative cytokine profiling. Of the sixteen cytokines, chemokines, and growth factors measured in patient plasma, only IL-6 showed a statistically significant difference between short- and long-term survivors (p < 0.001): higher baseline IL-6 correlated strongly with shorter overall survival. This is not surprising given IL-6's established roles in tumor progression, cachexia, and metastasis, but it raises an intriguing possibility — that baseline inflammatory cytokine status may be a useful predictor of who benefits most from the YIV-906 combination. Identifying patient subgroups likely to respond is one of the central challenges facing all adjuvant trials, and having a candidate biomarker to investigate is clinically valuable.
That data, if it comes, will come from the ongoing randomized, placebo-controlled Phase II trial of YIV-906 plus sorafenib in HBV-positive HCC patients — the trial designed to resolve the most important outstanding questions. Its design reflects the lessons accumulated across a generation of clinical investigation: double-blind, placebo-controlled, with clear inclusion criteria and a defined patient population where the biological rationale for YIV-906's benefit is strongest.
A Formula as a Systems Drug
The conceptual framework Posts 1 and 2 established — the inadequacy of single-target pharmacology for complex diseases, the emergence of network pharmacology as a way of asking better questions about botanical medicines — finds its most concrete expression in YIV-906. Here is a drug that is not a drug in the conventional sense. It has no single active compound. It has no single target. It has no single mechanism. It is a polychemical system that works on biological networks — in the gut, in the tumor, in the immune microenvironment — simultaneously and through mechanisms that require the full formulation to operate.
What it demonstrates, more than any preclinical model could, is that this approach is clinically tractable. It can be standardized. It can be manufactured consistently. It can be taken through regulatory processes. It can be combined with conventional drugs without additive toxicity. It can be studied in controlled trials. The challenges are real — the pharmacokinetics of a polychemical formula are more complex to characterize than a single molecule, the mechanism studies more difficult to complete, the regulatory pathway less well-defined. But none of these challenges are insurmountable, and YIV-906's clinical history demonstrates they have been navigated before.
The deeper significance of YIV-906 is not just what it does, but what it represents: proof that a 1,800-year-old clinical observation about a gastrointestinal formula can be translated, with appropriate scientific rigor, into a sophisticated cancer adjuvant whose mechanisms of action illuminate some of the most active frontiers of contemporary oncology — checkpoint immunotherapy, cancer stem cell biology, metabolic reprogramming, tumor microenvironment remodeling. The ancients did not know about SHP2. They did not know about NFAT or PD-L1 or M1 macrophage polarization. But they knew that this formula, given to sick people, produced better outcomes than its components alone. Two thousand years of clinical observation compressed into a four-herb prescription. What WE Medicine is trying to do is explain why.
"PHY906 may potentiate various anti-tumor agents that work synergistically in the tumor microenvironment."
— Lam et al., Scientific Reports, 2015