Perforin Kills Cancer Cells Without Its 'Ammunition': The Cell Perforin-Ferroptosis Study, Fact-Checked and Explained
For decades, immunology textbooks have carried a standard illustration: T cells and natural killer (NK) cells attack tumors, release perforin to punch holes in the cancer cell membrane, and granzymes pour through those holes to execute the cell. Perforin is the weapon; granzymes are the ammunition — both indispensable. A new *Cell* paper removes the ammunition from that picture: even when the main granzymes are missing, perforin by itself can kill tumor cells.
This post is a source-checked reworking of the original forum thread, with public corrections as well as credit where the original got things right.
Corrections to the Original Post
The date was wrong. The paper was published online in *Cell* on October 23, 2026 (Epub ahead of print, pii S0092-8674(26)01208-0; DOI: 10.1016/j.cell.2026.09.062). November 3 was the date Chinese media (China Science Daily / sciencenet.cn) republished the news, not the publication date — an eleven-day gap. Attribution of the volume/issue to 2027 in Chinese secondary sources reflects the same confusion.
The team has more than one lead. The study was led from U-M Rogel, with three co-senior/corresponding authors: Weiping Zou, Ilona Kryczek, and Judy Lieberman of Harvard Medical School. Crediting only the Zou lab is too narrow.
"Rewriting the textbook" is a media headline, not the authors' wording. The authors said the work "changes our understanding of one of the longest-standing mechanisms of T cell- and NK cell-mediated tumor clearance" — a more measured claim than the headlines.
"Major granzymes deleted" must be read precisely. What was removed were the two dominant scissors, GzmA/B (GzmB knockout in human NK cells, GzmA/B double knockout in mice, plus pharmacological pan-granzyme inhibition) — not the entire toolbox. The original post's phrasing was too broad.
The "sub-lytic concentration" qualifier — which the original post correctly flagged — deserves bolding. The paper specifies pores formed just enough to let granzymes pass and trigger calcium flux, without directly lysing the cell. At supraphysiological concentrations, perforin can kill cells outright; the "just right" condition is the precondition for the new finding. Most Chinese retellings omitted this.
The mechanism line checks out against the primary source: perforin pore formation drives transferrin receptor internalization, promotes endosome–mitochondria contacts, iron accumulates in mitochondria causing dysfunction and lipid peroxidation, ending in ferroptosis. Iron supplementation suppressed tumors in mice in a Prf1-dependent manner; on the patient side, the evidence is correlation in TCGA pan-cancer and ICB cohorts. Several details could not be verified in the primary source and are left blank here: the iron dosage in the high-iron diet, the exact mapping of the four mouse models to the diet experiments, cohort sample sizes, the verbatim basis for "membrane potential collapse," and the original wording of the claimed positive correlation between perforin and transferrin receptor expression.
The same-journal vitamin B2 study is not a companion paper. It was published online October 6, 2026 (DOI: 10.1016/j.cell.2026.09.020), by a different team on a different day — same journal, same month, nothing more. The directions also differ: in that paper, T cells die — tumors grab vitamin B2 via SLC52A2, T cells deprived of B2 undergo ferroptosis themselves, and immunity is weakened. In this paper, it is tumor cells that die. The two point in opposite directions and cannot be fused into a single "ferroptosis fights cancer" narrative.
Finally, hold the line on this: the press release states verbatim, "The takeaway is not to give all cancer patients an iron-rich diet," and Zou added, "We don't fully know which cancer types would be fed by iron." The iron supplementation in mice was a mechanism experiment proving the pathway depends on Prf1 — not a dietary prescription. A mechanism discovery is not a therapy; this caveat cannot be skipped here.
🗡️ Three Lines of Evidence
The study does not stop at in vitro observation; it connects three lines: in vitro experiments, mouse models, and patient data.
- In vitro: perforin without granzymes remains lethal to cancer cells. The killing mechanism is ferroptosis — an iron-dependent, lipid-peroxidation-driven programmed cell death formally named only in 2012.
- Mouse models (the surprise): mice on iron-rich diets showed slower tumor progression, and this suppression depended on perforin — in perforin-gene knockout mice, the protective effect vanished entirely. High-iron diets also synergized with anti-PD-L1 immunotherapy.
- Patient data: across colorectal cancer, melanoma, breast cancer, lung cancer and more, patients with high intratumoral perforin expression lived longer and responded better to immune checkpoint blockade. High perforin expression correlated positively with transferrin receptor expression, hinting that iron metabolism status may influence immunotherapy efficacy.
- A mechanism discovery is not a therapy. The paper itself labels clinical application "premature." Near-term directions are basic-research in nature: identifying what makes tumor cells susceptible to perforin, and testing whether multiple cell death pathways can cooperate. The intersection of immune checkpoint blockade and iron metabolism is a named future path.
- The ferroptosis picture is getting more complex. Around the same issue, *Cell* also published another ferroptosis study: tumors hijack vitamin B2, inducing ferroptosis in T cells themselves. Ferroptosis is both a weapon immune cells use against tumors and a tool tumors use to suppress immunity; the full map of this arms race is unfinished.
- What kind of news is this? It changes no current treatment guidelines and produces no immediately usable intervention. Its value is cognitive: a textbook mechanism running for decades gains a second pathway, and that pathway points to a variable — iron metabolism status — that is measurable and actionable.
- Bell, H.N., Nair, K., Tison, K. et al. Perforin triggers iron-induced mitochondrial damage to drive ferroptosis and anti-tumor immunity. Cell (2026). DOI: 10.1016/j.cell.2026.09.062
- Sciencenet.cn: "Immune system protein can kill cancer cells on its own" (Cell paper report), 2026-11-03
- WuXi AppTec / Academic Express: "Rewriting the textbook! Cell: The immune system has another trick for clearing tumors," 2026-11-02
- Sohu Health: "Cell: Challenging conventional wisdom, perforin itself is the cancer cell killer," 2026-11
Mechanistically, the study reveals how sub-lytic perforin works independently: membrane pores trigger transferrin receptor internalization, iron is delivered to mitochondria through endosome–mitochondria contact sites, causing iron accumulation, membrane potential collapse, and lipid peroxidation, culminating in ferroptosis. This pathway requires neither granzymes nor interferon.
⚠️ An Iron-Rich Diet Is Not a Prescription
"Mice on iron-rich diets grew tumors slower" is the easiest sentence to misread. The corresponding author's original words: "We need to be very careful. That's because there are some cancer types that love an iron-rich diet. For those tumors — which we don't yet know — an iron-rich diet may accelerate progression."
Iron is double-edged in cancer biology: on one hand, ferroptosis depends on iron and immune cells use it to execute tumors; on the other, cancer cell proliferation itself needs iron, and some tumors actively sequester it. Which cancers benefit and which are harmed — that boundary is not yet drawn. The conclusion is mechanistic; there is no dietary advice whatsoever, and a long road lies between mechanism and clinic.
🔗 Why the Textbook Picture Changes
The classical picture comes from decades of mature research: perforin punches the hole, granzymes enter and trigger apoptosis — an accepted default in immunology. Granzymes were the "ammunition," and a weapon without ammunition was assumed toothless.
The new finding says: the weapon itself is the ammunition. Beyond its old "pore-forming" role, perforin has an independent killing pathway operating through an entirely different mode of cell death (ferroptosis, not apoptosis). This explains an old puzzle — why some immune cells can still eliminate tumors when granzymes are insufficient — and reframes immunotherapy: immune checkpoint blockade has been shown to enhance T cell-mediated ferroptosis, and perforin sensitivity is an intervening variable tied to immunotherapy efficacy. In the authors' words, the work changes our understanding of one of the longest-standing mechanisms of T cell- and NK cell-mediated tumor clearance.
🧭 Boundaries and Next Steps
【Verdict】Perforin won't appear in any oncology prescription soon, but it may show up in future immunotherapy-response panels — if iron metabolism status truly predicts immunotherapy response, it is the next biomarker worth watching.
---
References: