Granulin (Ov-GRN-1)

In brief

What granulin is, and which granulin we mean

Granulin is not one molecule. It is a small cysteine-rich module, generally held in shape by six disulfide bonds ²³, and the proteins that carry it are chains of those modules. The full chain is called progranulin, and enzymes can cut it into separate 6-kilodalton granulin peptides; both the intact chain and some of the single modules are biologically active, which is why the literature uses the two names almost interchangeably and why that is confusing ². The module is old. It turns up in certain plant proteases, in sponges, and in single-celled relatives of animals, so it is among the oldest signaling proteins multicellular animals still use. Mammals carry one progranulin gene; bony fish carry three to six ².

Human progranulin is a 593-amino-acid glycoprotein, written by the gene GRN on the short arm of chromosome 17 at 17q21.31, and secreted as a 68.5-kilodalton protein ^4,5. It has three jobs that the rest of this page keeps colliding into each other. It makes epithelial cells divide, and overexpressing it turns them tumorigenic ⁴. It is a wound protein: after injury its message rises in the inflammatory cells, the fibroblasts and the blood vessels of the wound, and applying it brings in more of all of those ¹⁶. And in the cell's own interior it is part of the endolysosomal system, the compartments that break down and recycle material, which is the job that explains why losing it damages the brain ^17,24.

Ov-GRN-1 is the liver fluke's version. A proteomic screen of what adult Opisthorchis viverrini releases into its surroundings — its excretory-secretory products, the whole mixture of proteins and vesicles a worm sheds — found a homolog of human granulin, meaning a protein descended from the same ancestral gene in a different species ^1,2. It is a single granulin domain, running at about 9 kilodaltons on a western blot ³, expressed most strongly in the worm's gut and tegument, which is the living skin covering a fluke, and it was detected in place on the surface of the bile duct lining cells of experimentally infected hamsters ¹. Those lining cells are cholangiocytes, and cholangiocarcinoma is the cancer they give rise to ²⁵.

There are therefore at least four granulins in this story, and keeping them apart is most of the work. There is human progranulin, the host's own. There is Ov-GRN-1, the small secreted one the worm sends out. And there are two further fluke granulin genes: a second single-domain granulin, Ov-GRN-2, and a full progranulin — a gene of thirteen coding stretches whose protein is 846 amino acids long and carries seven complete granulin modules plus a partial eighth. The first pair was reported with the fluke's genome in 2014, and the 2015 mechanism paper noted both while adding that neither product was detectable in the secreted proteome ^22,26; the large one was characterized in detail only in 2023, and only in a preprint that has not been peer reviewed, where it is most strongly expressed in the infective larval stage ²⁷. That preprint counts seven and a half granulin modules where the genome report describes eight subunits. Deleting Ov-GRN-1 does not delete either of them, which is one plain reason the knockout experiments attenuated disease rather than abolishing it.

The whole history, including every reversal

This molecule has been discovered three times by people who were not looking for each other's results. Cancer biologists found it in the 1990s as a growth factor that makes epithelial cells tumorigenic ⁴. Wound researchers found it in 2003 as a mediator of repair ¹⁶. Neurologists found it in 2006, from the opposite direction, when they learned that having only one working copy of the gene destroys the frontal and temporal lobes ^5,6. Parasitologists found the fluke's version in 2009 ¹. Each field spent years describing a different animal, and the reason the show's sentence came out muddled is that the whole literature is built out of partial views.

The reversals are instructive and they are not minor. The protein that looked like a cancer driver turned out to be necessary for neurons to survive. The receptor that made the mechanism tidy — progranulin as a natural blocker of tumor necrosis factor — was reported in Science in 2011 ²⁹ and then, in a deliberate replication that reproduced the paper's other interactions by the same methods, could not be found at all ³⁰. The simple reading of gene dose, that one broken copy gives a mild version of what two broken copies give, is wrong here: one copy gives dementia under the age of 65, two give neuronal ceroid lipofuscinosis, a disease whose age of onset and pathology are markedly different ^5,7. And in the parasite literature, the 2014 claim that the fluke cannot live without its granulin did not survive contact with knockout worms that lived ^20,31.

The parasite thread runs separately until 2019, when it becomes the most interesting thing in the field. Two papers appeared in eLife on the same day, from overlapping teams: one editing a gene in the blood fluke Schistosoma mansoni, one editing granulin in the liver fluke ^3,32. They were the first reports of programmed gene editing in any flatworm, which an independent commentary in the same journal said could usher in a new era for research on these parasites ³³. Three years later the liver fluke experiment was rerun with a chemical carcinogen added and taken out to twenty-four weeks, and this time cancers were counted ¹⁰. That second paper is the one the show should cite when it says the word cancer.

Where the field stands in 2026 is worth stating precisely, because it is less finished than the headlines suggest. No heritably transgenic liver fluke line exists; every result so far comes from mosaic editing of individual worms, and the laboratory that built the method says so in print ¹¹. On the human side, no treatment yet changes the course of frontotemporal dementia, and the strategies being tried to raise progranulin all carry the same stated worry in the other direction: too much of a growth factor is an oncogenic risk ^34,35.

The granulin CRISPR experiment, said plainly

Here is the sentence to keep. The gene that was cut out belonged to the worm, not to the patient. In 2019 a team used CRISPR/Cas9 to mutate Ov-grn-1 — the liver fluke's own granulin gene — inside living Opisthorchis viverrini, and then put those edited parasites into hamsters. Nothing in the hamster's genome was touched, and no human gene was involved ³. CRISPR/Cas9 is a pair of parts: a short guide made of RNA that matches a chosen stretch of DNA, and an enzyme, Cas9, that cuts the DNA where the guide lands; the cell's own error-prone repair then usually leaves small insertions or deletions that wreck the gene. A knockout is a gene disabled in that way. The plasmid carrying both parts was pushed into the worms by electroporation, a brief electrical pulse that makes membranes temporarily leaky.

The 2019 experiment measured precancerous change only, and it is important that this is not oversold. There was no chemical carcinogen and there were no tumors. Thirty hamsters were infected with wild-type or edited flukes, or left uninfected, and the readouts were the thickness of the bile duct lining, the amount of collagen, and a marker of activated scar-forming cells. Edited flukes established and matured normally, though the fluke counts behind that claim come from two livers per group at day 60. At fourteen days, wild-type infection thickened the duct lining to roughly 500% of the thickness measured in uninfected animals while the knockout infection reached only 145%; by sixty days the gap had narrowed to 216% against 162%, which the authors call significant but less marked than during acute infection at day 14. At sixty days collagen was about a quarter lower in the knockout group, a figure the paper's own text gives as 28% and its figure legend as 23% ³. Hyperplasia is tissue thickening by cell multiplication where the cells still look normal; fibrosis is the laying down of scar tissue. Both are risk states, not cancer.

The 2022 experiment is the one that measured cancer. The same parasite-gene knockout was combined with dimethylnitrosamine in the drinking water — a chemical carcinogen — and run to twenty-four weeks, which is how this model is known to produce bile duct cancer. Cholangiocarcinoma appeared in 10 of 12 control hamsters against 7 of 13 in the knockout group, with high-grade tumors in 8 against 6. Scarring immediately around the worms was lower, the fraction of duct cells caught dividing fell roughly fivefold, and the proportion carrying mutant p53 fell from 61.1% (95% confidence interval 43.6–68.3) to about 7.5% (95% confidence interval 2.3–27.5), roughly a seventh of the control level ¹⁰. The abstract states the same findings in words: markedly less periductal fibrosis, minimal biliary cell proliferation, significantly fewer TP53 mutations and fewer high-grade cancers. What did not change is as informative as what did. Cancer still appeared in more than half the knockout-infected hamsters. The tumors that did appear were the same kind in the same places: no substantial difference among the groups in location or histological subtype, 21 of 26 of the tubular type, the right lobe commonest at 20 of 26. Egg output tracked worm burden only loosely, and no significant correlation between egg counts and the severity of disease was apparent in any group ¹⁰.

Three honest limits travel with those numbers, and the authors raise all three themselves. First, the editing was mosaic: the median worm sequenced carried insertions or deletions in about 3% of read-pairs at the target site, with individual worms ranging from none to 91%, so no worm was a clean null and every result is an average over a mixed population ^10,11. Second, worm burden differed: 38.9% fewer knockout worms were recovered, so some of the reduced damage could in principle reflect fewer parasites rather than absent granulin. That count comes from the study's separate fourteen-week arm; in the cancer arm itself the worms were never counted, because the livers had to be fixed whole for sectioning, so the worm burden in the animals that developed tumors is simply unknown ¹⁰. Two things narrow the worry. The authors measured scarring and cell division immediately adjacent to individual flukes rather than across the whole liver; and the second knockout group lost even more worms — 64.5% fewer, 21.3 against 60.0 per hamster — yet showed no attenuation whatever, with cancer in 9 of 13 hamsters, four of them high grade, and mutant p53 in 67.7% of duct cells, slightly above the controls' 61.1% ¹⁰. Fewer worms, on this study's own evidence, does not by itself produce less cancer. Third, the intended comparator gene, the surface tetraspanin Ov-tsp-2, turned out to be essential: those worms survived in far smaller numbers and the adults that did survive carried near-normal transcript levels, implying that only the unedited ones made it, so the control arm the design depended on could not do its job — as a later paper showing that this knockout leaves the worm's tegument eroded and vacuolated explains ^10,20.

And the separate experiment that causes the confusion: in 2020 the same group did knock out a human gene, progranulin, using CRISPR/Cas9 — but in a cultured human bile duct cell line, achieving more than 80% reductions in the host protein and in the vesicles those cells released. The purpose was to silence the host's own progranulin so the fluke protein's effect could be read without interference. No animal was edited, no parasite was edited, and nothing about disease was measured ³⁷. Parasite gene edited in a living worm: 2019 and 2022. Host gene edited in a dish: 2020. Any sentence that merges them will be wrong.

What it does to host cells, tier by tier

The biology divides cleanly into three tiers of evidence, and the discipline this page tries to keep is never to borrow a strong word from a weak tier — never to say the molecule “causes” anything in people on the strength of what it does to cells in a dish. In human cells in culture, the molecule is a straightforward mitogen: nanomolar concentrations of correctly folded monomeric protein make both cancerous and non-cancerous bile duct cell lines divide ^1,8. In animals, recombinant protein accelerates wound healing in mouse skin and grows new vessels in a quail egg membrane ²², and in hamsters the removal of the gene from the worm reduces the damage the infection causes ^3,10. In people, there is no measurement of this protein at all.

The mechanistic picture is unusual. Labeled Ov-GRN-1 was internalized by human cholangiocytes and located, by high-resolution microscopy, between the cell's upper and lower actin filaments, without colocalizing with lysosomes, early or late endosomes, Golgi or endoplasmic reticulum — which argues for direct entry into the cytosol rather than the familiar pattern of a growth factor signaling at a surface receptor ²². The human protein's best-characterized receptor does the opposite job: sortilin binds progranulin and takes it in for disposal in the lysosome, which is why blocking sortilin raises progranulin levels and is being pursued as a treatment ^28,34.

Downstream, the signal looks like a wound program. Proliferation driven by the protein was blocked by an inhibitor of MAP kinase kinase, placing it on the mitogen-activated protein kinase pathway ¹. Exposure changed the cholangiocyte's output of chemokines and altered proteins associated with wound healing and cancer pathways ²². When host progranulin was removed from the cells so that the parasite protein could be read alone, fluke granulin induced cancer-associated RNAs in the vesicles those cells released, including members of the MAPK and AKT pathways, two of the chains of signals a cell uses to decide whether to grow and survive, and those vesicles then changed naive recipient cells ³⁷.

Granulin is not the only secreted molecule doing this, and the page would be misleading if it implied otherwise. The same worm releases extracellular vesicles that cholangiocytes take up, driving proliferation and secretion of interleukin-6 and interleukin-8, two signaling proteins that call inflammatory cells in ¹²; antibodies against the vesicles' surface tetraspanins, and drugs that block the cell's uptake machinery, reduce both that uptake and the proliferation that follows, while silencing those tetraspanin genes cuts how many vesicles the worm releases at all ³⁸; recombinant loops of both surface proteins make bile duct cells migrate, and the Ov-TSP-3 loop also raises interleukin-6 and interleukin-8 and drives proliferation ³⁹. Not all of those vesicles do the same thing: in a preprint not yet peer reviewed, the small-vesicle fraction drove cholangiocyte proliferation while the larger microvesicle fraction produced no effect at all, and the two had different protein compositions ⁴⁰. That is a reason to read “fluke vesicles” in any one paper as shorthand for a particular preparation. And in the related fluke Opisthorchis felineus, an independent Russian group finds exosome-like vesicles enriched in particular tRNA-derived small RNAs whose predicted human targets are over-represented among the genes that change in cholangiocytes treated with that worm's secretions — a prediction plus a correlation rather than a demonstrated mechanism, and one the authors say does not carry across species, since O. viverrini's dominant fragments are different ones ⁴¹. The field's own synthesis treats secreted mitogens, inflammation and repeated mechanical wounding as three strands twisting together ^18,19.

Two granulins in one bile duct, and a second fluke gene

A tidy story would have the worm supplying granulin and the host supplying none. That is not what happens. In the hamster model, the animal's own granulin rises as the lesions progress, measured by immunohistochemistry (staining a tissue slice with antibodies so that a particular protein shows up under the microscope), western blot (separating a tissue's proteins by size and probing for one of them) and RT-PCR — reverse transcription polymerase chain reaction, which copies a gene's message and amplifies it so that small amounts can be counted — across 120 animals in four groups, and most prominently in the hamsters given both fluke and nitrosamine, tracking a marker of cell division ⁴². Both granulins are therefore present and both increase, which is the biological reason the 2020 experiment had to delete the host's progranulin from its cells before it could read the parasite protein's effect at all ³⁷.

The worm also has more than one granulin gene. A preprint, not peer reviewed, characterizes a full fluke progranulin whose gene spans 36,463 base pairs of DNA across thirteen coding stretches, and whose protein is 846 amino acids long with seven granulin modules and part of an eighth, most strongly expressed in the metacercaria, the infective larval stage encysted in fish flesh. A peptide fragment of it made human cholangiocytes divide and raised interleukin-6 and interleukin-8 ²⁷. It was not a new discovery: the 2014 genome had already reported both this progranulin and a second single-domain granulin, Ov-GRN-2, and the 2015 mechanism paper noted both while adding that neither product was detectable in the secreted proteome ^22,26. If that holds up in peer review, then knocking out Ov-grn-1 removes one granulin from a worm that still has another, which is an obvious candidate explanation for why the knockout attenuated disease instead of abolishing it.

How much granulin a given bile duct receives is not fixed by the worm's genome either. In diabetic hamsters, worm establishment correlated with fasting plasma glucose, the worms themselves were stunted and laid fewer eggs, and yet virulence genes including granulin, tetraspanins and thioredoxins were significantly upregulated — 4,314 differentially expressed genes at twenty-one days and 567 at thirty-five ⁴⁴. That is association inside an animal model, from a group largely separate from the gene-editing authors, and it introduces a modifier the knockout experiments could not see: the host's metabolic state appears to turn the parasite's growth factor up.

One boundary is worth drawing clearly, because it separates two mechanisms that are often blurred. Ov-GRN-1 is a soluble secreted protein and is not detected in the worm's extracellular vesicles; the surface tetraspanin Ov-TSP-2 is a structural protein of the worm's skin and of those vesicles ²⁰. Knocking out the tetraspanin wrecks the tegument, cuts vesicle release and cuts uptake by human cells in culture ²⁰, and in hamsters those worms did not survive to the adult stage at all ¹⁰; knocking out granulin is not lethal ^10,20. Two different mechanisms, two different vaccine logics, and one reason the 2022 experiment's comparator arm failed.

How it is found and measured

Nothing about this molecule was found by looking for it. It was found by collecting what adult flukes release into culture medium, breaking that mixture into its component proteins and identifying them by mass spectrometry, and recognizing one of them as a granulin by its sequence ¹. That is the standard route for parasite virulence factors, and it has a selection bias worth naming: a protein that is abundant in the mixture and easy to detect gets found first, which is one reason granulin dominates this literature while the worm's other granulin-family genes have stayed at the margins of it. They were not hidden: the 2014 genome ²⁶ already reported a second single-domain granulin, Ov-GRN-2, and a full progranulin, and the 2015 mechanism paper noted both while adding that neither product was detectable in the secreted proteome ²². Only in 2023, and only in a preprint, was the large one characterized in detail ²⁷.

Making the protein to test it was itself an obstacle. Granulins are knots of disulfide bonds and refuse to fold properly in bacteria, so the first preparation had to be recovered from inclusion bodies and refolded, and a dedicated paper was needed to produce monomeric protein at useful scale ⁸. That difficulty is why the therapeutic program moved to short synthetic fragments instead of the whole molecule ^13,36, and it is also why claims about the full-length protein's behavior rest on a small number of preparations.

In the editing experiments, the measurement that matters is how much of the gene was actually cut. Both papers amplified the target region from the worms' DNA and sequenced millions of reads, counting mutations against unedited controls ^3,10. The numbers are sobering and the authors report them plainly: in the 2022 work the median adult fluke carried insertions or deletions in about 3% of read-pairs at the target site, with individual worms ranging from none to 91% ¹⁰. The message and the protein fell far further than that 3% would predict — adult knockout worms held about a tenth of normal transcript — which the group explains by the editing plasmid reaching the worm's multinucleated surface layer, where these genes are expressed ²⁰. Neither paper reports a genome-wide search for off-target cuts. The guides were chosen because they were predicted to have little or no off-target match in the fluke's genome, and each knockout had a second, untargeted gene measured to show the effect was gene-specific; but no one sequenced the whole genome to look for stray cuts, and the authors say in print that they cannot conclude that off-target mutations did not occur ¹⁰.

In the host tissue, everything is read by eye and by stain: hematoxylin and eosin for the thickness of the bile duct lining, a collagen stain for fibrosis, an antibody for activated scar-forming cells, a labeled nucleotide for cells caught dividing, and an antibody for mutant p53 ^3,10. Two veterinary pathologists graded the 2022 lesions blinded, which is the right way to do it. These are image-based counts on sampled sections, so they inherit the ordinary limits of sampling: a section is a thin slice of one lobe of one liver.

And in people, the measurements available are about the infection, not the protein. Stool microscopy finds eggs, antibody tests are used for screening, and molecular methods in reference laboratories distinguish the species ²⁵. The only granulin assay in clinical use measures the human protein in blood or cerebrospinal fluid, for frontotemporal dementia, where carriers run up to about 50% below normal and the result guides classification and monitoring ^34,35. Nobody has a test that tells a patient in Khon Kaen how much fluke granulin their bile duct is receiving.

• Plasma or cerebrospinal fluid progranulinThe one granulin measurement in routine clinical use, and it is about the brain, not the bile duct: carriers of a loss-of-function GRN mutation run up to about 50% below normal ³⁴.To support or refute a genetic diagnosis of frontotemporal dementia, and to follow treatments meant to raise the level ^34,35.Low levels track the haploinsufficiency well enough that progranulin is used for classification, prognosis and monitoring in this disease ³⁵.It says nothing about the parasite protein. No clinical assay measures Ov-GRN-1 in a patient.
• Deep sequencing of the edited geneIn both knockout experiments the proof that editing happened came from sequencing millions of reads across the target site and counting the mutations there — overwhelmingly single-letter substitutions in 2019, insertions and deletions in 2022 ^3,10.To report what fraction of the sequenced reads from the parasite's DNA carried a mutation at the cut site.It counts DNA molecules precisely; in the 2022 experiment the median adult worm carried insertions or deletions in about 3% of read-pairs at the target site, with individual worms ranging from none to 91% ¹⁰.A percentage of reads is not a worm. Editing was mosaic throughout, no heritable knockout line exists, and that is the field's own stated next milestone ¹¹.
• Transcript and protein measurement in the worm and its secretionsQuantitative PCR — polymerase chain reaction, which copies and amplifies a chosen stretch of genetic material so that small amounts can be counted — for the message and western blot for the roughly 9-kilodalton protein, in whole worms and in what they secrete ^3,10,20.To show that cutting the gene actually removed the product; adult knockout worms held around a tenth of normal transcript ¹⁰.The protein fell much further than the measured DNA editing rate would predict, which the authors explain by the plasmid reaching the worm's multinucleated surface, where these genes are expressed ²⁰.It is an expression control, not a genome-wide check. The guides were chosen because they were predicted to have little or no off-target match in the fluke's genome, and each knockout had a second, untargeted gene measured to show the effect was gene-specific. But no one sequenced the whole genome to look for stray cuts, and the authors say so themselves: they cannot conclude that off-target mutations did not occur.
• Tests for the infection itselfFinding the worm, not its granulin: stool microscopy for eggs in surveys, antibody tests for screening, and molecular methods in reference laboratories for telling the species apart ²⁵.Case finding and treatment decisions in endemic areas.Adequate for population surveys; species identification needs the molecular methods ²⁵.A positive stool says a worm is present. It says nothing about how much granulin that worm is secreting, and nothing about the fibrosis already laid down ¹⁸.

What is done about it, and what is being built

Nothing in the clinic targets granulin. What exists for the infection is praziquantel, which kills the worm: a network meta-analysis of 26 trials in 3,340 participants, only seven of them in Opisthorchis viverrini, put the best predicted cure rate for that species at 93.8% (95% confidence interval 85.7–97.5) for 50 mg/kg plus 25 mg/kg given in one day and 92.1% for a single 50 mg/kg dose, in a trial literature the authors themselves grade as low quality ⁴⁵. Control programs combine that drug with education and environmental measures ²⁵. Killing the worm removes the source of the granulin, but it does not undo scarring already laid down, and reinfection is routine in endemic areas ¹⁸.

A vaccine is being built, and it is important to say that it is not a granulin vaccine. Every tested antigen in this program is a tetraspanin from the worm's surface and its vesicles. Hamsters immunized with vesicles or with recombinant tetraspanin loops yielded significantly fewer adult flukes, with egg counts falling significantly only for recombinant Ov-TSP-2 ¹⁴. An oral version made of Bacillus subtilis spores gave 56% reductions in both worm counts and fecal eggs ¹⁵. Transferred monoclonal antibody alone cut adult flukes by 74.5% and eggs per gram by 67.4%, showing the protection is antibody-mediated ⁴⁶. The case for a granulin vaccine rests, so far, only on the 2009 finding that antibodies against it block the proliferation its secretions cause ¹.

In the other direction, the molecule is being developed as a drug. Because the full protein is hard to make, shortened peptides from its front end were synthesized; both two- and three-disulfide versions drove human cholangiocyte proliferation and healed mouse wounds potently, and the chemistry was then refined around the proline residues that control folding ^13,36. The most recent paper in that line found that a fragment from the tail end works without folding into the granulin shape at all ²³. The target is chronic wounds in people with diabetes, tobacco users and the elderly ⁹. A growth factor from the most carcinogenic parasite in medicine is a strange drug lead, and the authors say so themselves.

On the human protein, the therapeutic traffic runs the opposite way: upward. Because frontotemporal dementia from GRN mutation is a deficiency state, the strategies are adeno-associated virus gene therapy, compounds that read through a premature stop codon, inhibitors of the sortilin receptor that clears progranulin, and antisense oligonucleotides, all aimed at restoring the missing roughly 50% ³⁴. Two cautions sit in that same review and both belong on this page: translational success in trials so far has been limited, and overexpression carries an oncogenic risk ³⁴. Two of those strategies have now been taken into people or to the edge of it. Latozinemab, a monoclonal antibody that blocks the sortilin receptor so progranulin is not cleared, was given to eight symptomatic people carrying a loss-of-function GRN mutation in a first-in-human phase 1 trial: it was well tolerated and raised progranulin in both blood and cerebrospinal fluid to roughly the levels seen in healthy volunteers ⁴⁷. AVB-101, an adeno-associated viral vector carrying the human gene delivered directly into the thalamus, reversed pathology in progranulin-deficient mice and spread the protein through the cortex of sheep and monkeys with no adverse events over six months, which its developers present as the basis for going into patients ⁴⁸. Raising the level is therefore no longer hypothetical; changing the course of the disease still is. No disease-modifying therapy for frontotemporal dementia exists; treatment remains symptomatic ³⁵.

What was believed, and is not

Several claims in this area were stated confidently, taught for years, and then corrected, and one of them is the claim the show's own script carried into the episode. Listing them is not a swipe at the researchers who made them; in most of these cases the correction came from the same group that made the original claim, published in print under their own names, which is exactly how this is supposed to work.

There is also a methodological belief worth retiring quietly: that the vesicle measurements in this field are directly comparable between papers. The authors who made most of them say otherwise, noting that purification and characterization standards for helminth extracellular vesicles are generally ill defined, which is a reason to treat fold-changes in vesicle counts across different laboratories with care rather than to average them ³⁸. The same caution applies to the editing percentages, though in the opposite direction from pooling: the 2019 figures come from DNA pooled across 7 to 10 worms per group ³, while the 2022 figures were read from 13 individual adult flukes ¹⁰. The sequencing is precise about DNA molecules either way, and silent about which cells in which worm carried them — which is why no result here is a clean gene-deletion phenotype ^10,11.

What is still unknown

The central gap is human. Every causal result on this molecule comes from a hamster, a mouse, a quail embryo or a cell line. Nobody has measured Ov-GRN-1 in the bile of infected people and followed them to see who develops cholangiocarcinoma, and until that is done the human link is inference from animals plus a strong epidemiological association with the worm itself ^12,18. The study that would settle it is straightforward to describe and hard to run: a prospective cohort in northeastern Thailand, biliary granulin measured at baseline, cancer counted fifteen to twenty years later.

The second gap is the cleanliness of the tool. There is no heritable knockout line in any liver fluke; editing has been mosaic in every experiment, with a median of about 3% of read-pairs at the target site carrying insertions or deletions in the cancer study, and the laboratory that built the method names heritable lines as the next milestone ^10,11. Until such a line exists, no result here is a clean gene-deletion phenotype, and the gap between a 3% edit rate and a near-total loss of protein rests on a plausible but inferential explanation about where the plasmid lands in the worm's body ²⁰.

The third gap is independence, and it is narrower in one direction than it looks and wider in another. It applies to the parasite half of this page rather than the human progranulin half, where the primary papers ⁴–² come from unrelated laboratories. Essentially every primary paper on liver fluke granulin shares a core group of authors across Khon Kaen University, George Washington University and James Cook University, and no outside laboratory has replicated the Ov-grn-1 knockout itself. But the underlying claim — that a liver fluke granulin drives malignant change in human bile duct cells — has been reached independently. A group at Sun Yat-Sen University in Guangzhou showed first that the granulin of the other Group 1 carcinogenic fluke, Clonorchis sinensis, makes normal human liver cells behave like tumor cells through the epidermal growth factor receptor and the RAS/MAPK/ERK and PI3K/Akt pathways ⁴⁹, and then that the same protein drives malignant transformation of normal human bile duct lining cells, in culture and in mice, by pushing macrophages into a tumor-supporting state and raising interleukin-6, which switches on the STAT3 and MEK/ERK pathways ⁵⁰. That is convergence on the mechanism from a different fluke, a different country and a different laboratory. It is not a replication of the gene-editing result, which still has none. The other genuinely external items are a commentary ³³, a transcriptomic study from a largely separate Khon Kaen group whose first author has nonetheless co-authored two papers with the editing team ⁴⁴, and a vesicle analysis from a Russian group working on a different fluke species ⁴¹. There is one sound reassurance to set against all this: a systematic check of retraction flags across this literature turns up exactly one retracted paper, an unrelated risk-mapping study, and none of the papers behind the granulin story carries a retraction or an expression of concern ⁵¹.

Smaller unknowns, each of which would change how the page reads. Whether the worm's other granulin genes really reach host tissue in quantity. A preprint reports that a peptide from the large one makes cholangiocytes divide ²⁷, but neither that protein nor Ov-GRN-2 was detectable in the secreted proteome when the question was last looked at ²², and peer review has tested neither. Whether the reduction in disease in knockout-infected hamsters is partly explained by the 38.9% drop in worm numbers, which per-worm measurements narrow, and which the comparator arm's own result argues against, since worms there fell by 64.5% without any attenuation of cancer ¹⁰. Whether the molecule signals at a receptor at all or enters the cytosol directly, since the microscopy argues against the usual route and no fluke granulin receptor has been identified ²². And whether raising human progranulin therapeutically is safe over years, given that the same protein's overexpression is tumorigenic in mice and the risk is stated openly by the people developing the treatments ^4,34.

One method also went quiet, and the silence is itself an answer. RNA interference — soaking worms in double-stranded RNA to silence a gene without cutting it — was how granulin was first switched off, in 2014 ³¹, and it is still used on the worm's tetraspanins ³⁸. But no RNA-interference study of Ov-grn-1 has been published since gene editing arrived in 2019. The field moved to CRISPR and did not go back, which means the one method that could knock the gene down reversibly, and in adults rather than larvae, has not been used to check the editing results.

How this connects to parasites and cancer, at its true strength

Start with what is not in doubt. Opisthorchis viverrini was classified a Group 1 human carcinogen — the category for agents where the evidence of human carcinogenicity is sufficient rather than suspected — in 1994, and Clonorchis sinensis in 2009 ¹⁸. Roughly 25 million people worldwide carry one of the three Opisthorchiidae liver flukes, with about 10 million of those carrying O. viverrini ^18,45. The current global accounting counts both flukes among the twelve Group 1 infectious agents and estimates that about 2.3 million new cancers worldwide in 2024, some 12% of all cancer cases, were attributable to infection — with the flukes a small share of a very large total dominated by Helicobacter pylori and human papillomavirus ⁵². Over 12 million people are estimated to be at risk of opisthorchiasis, and the associated bile duct cancer is credited with roughly 20,000 deaths a year in Thailand alone ⁵³. The adults live in the bile ducts for 10 to 30 years, and Khon Kaen province in northeastern Thailand has the highest recorded incidence of this cancer in the world ¹⁸. That part of the story needs no help from molecular biology.

What granulin adds is a step from association toward mechanism, and the step is real but smaller than the headlines. The argument runs: the worm secretes a growth factor; that factor makes bile duct cells divide and wounds close; delete the gene from the worm and the infected animal gets less thickening, less scarring, less cell division, fewer cells with a damaged p53 brake, and fewer high-grade cancers ^1,3,10. That is a genuine intervention experiment, not a correlation, and it is the strongest causal evidence in the whole parasite-cancer field. It is also a hamster experiment with a mosaic knockout, a comparator arm that failed, a difference in worm burden and no independent replication of the editing experiment ^10,11. One point cuts the other way and belongs in the same breath: the carcinogen was dimethylnitrosamine at 12.5 parts per million in the drinking water from the second to the tenth week, a dose the authors describe as otherwise sub-carcinogenic. On its own it does not produce these tumors. The worm is doing work the chemical cannot do alone ¹⁰.

The right summary sentence is therefore narrow. One parasite protein makes a measurable contribution to the precancerous damage, and to cancer incidence in an animal model that pairs it with a dose of nitrosamine too low to produce these tumors by itself. It is a contribution, not the cause: disease was attenuated, not abolished, the worm carries at least one other granulin-family gene ²⁷, and the field's own synthesis puts secreted mitogens alongside inflammation and repeated mechanical wounding rather than above them ^18,19. Nothing here says granulin causes cancer in people, and nothing here lets anyone dismiss it either.

Two wider implications are worth carrying into the season's question about counting. First, the knockout experiments are a template that can be pointed at other parasites and other proteins; the same week's companion paper did it in a blood fluke, and the method is now being used across flatworms ^11,32. If the only parasites credited with causing cancer are the ones somebody has built this kind of experiment for, then the list of parasite-attributable cancers reflects where the tools were aimed as much as where the biology is. Second, the fluke's genome is the reason this was possible at all, and it was published only in 2014 ²⁶. The tools are young, which is a reason to expect the list to grow, and a reason not to read today's list as the final number.

And one last thing, said kindly because the owner was right to find it odd. The strangest fact on this page is that the same molecule is a drug candidate for chronic wounds in people with diabetes, and the people developing it are the people who showed it helps turn a bile duct cancerous ^9,13. That is not a contradiction. A signal that tells epithelium to repair itself is useful for a few days in a skin wound and dangerous for thirty years in a bile duct. Dose, site and duration are the whole difference, and saying so is more honest than either half of the story told alone.

Where it connects

Sources

53 sources, numbered as they are cited. Every one was checked against PubMed or its publisher before it was cited here; the note under each says what it shows and what it does not.

1. 1
   Smout MJ, Laha T, Mulvenna J, Sripa B, Suttiprapa S, Jones A, Brindley PJ, Loukas A. A granulin-like growth factor secreted by the carcinogenic liver fluke, Opisthorchis viverrini, promotes proliferation of host cells.PLoS Pathog · 2009 · 5(10):e1000611doi:10.1371/journal.ppat.1000611 · PMID 19816559

   The discovery paper: Ov-GRN-1 is a secreted granulin homolog in the fluke's excretory-secretory products, concentrated in gut and tegument, detected on the biliary lining of infected hamsters, and, on the paper's own conclusion, the major growth factor in those secretions, since antibodies to it inhibited the proliferation they caused — a ranking the paper makes, not a measured fraction. It shows proliferation of mouse fibroblasts and a human cancer cell line; it does not show that the protein causes cancer.

2. 2
   Palfree RG, Bennett HP, Bateman A. The Evolution of the Secreted Regulatory Protein Progranulin.PLoS One · 2015 · 10(8):e0133749doi:10.1371/journal.pone.0133749 · PMID 26248158

   Explains what a granulin is: progranulins are chains of small cysteine-rich modules that proteolysis can release as separate 6-kDa granulin peptides, both the whole protein and some single modules being active. The module is ancient, present in sponges, in some plant proteases and in single-celled relatives of animals, and mammals carry one gene where fish carry three to six.

3. 3
   Arunsan P, Ittiprasert W, Smout MJ, Cochran CJ, Mann VH, Chaiyadet S, Karinshak SE, Sripa B, Young ND, Sotillo J, Loukas A, Brindley PJ, Laha T. Programmed knockout mutation of liver fluke granulin attenuates virulence of infection-induced hepatobiliary morbidity.eLife · 2019 · 8:e41463doi:10.7554/eLife.41463 · PMID 30644359

   The experiment the show's sentence is about: CRISPR/Cas9 was used on the parasite's own Ov-grn-1 gene, the edited flukes still colonized hamster bile ducts and matured, and the infection produced less biliary thickening and less fibrosis. No carcinogen was given and no tumors were counted, so this paper measures precancerous change only. Its Results text and its Figure 4 legend give the fall in collagen as 28% and 23% respectively, an internal discrepancy only the authors can settle.

4. 4
   He Z, Bateman A. Progranulin gene expression regulates epithelial cell growth and promotes tumor growth in vivo.Cancer Res · 1999 · 59(13):3222-9PMID 10397269

   The human protein as a tumor driver: progranulin is a 593-amino-acid glycoprotein, and forcing cells to make more of it produced anchorage-independent growth and markedly increased the tumorigenicity of those cells as xenografts in nude mice, while reducing it impaired growth. Xenografts of transfected cell lines into immune-deficient mice, not human disease, and an increase in tumorigenicity is not the same as a count of tumors.

5. 5
   Baker M, Mackenzie IR, Pickering-Brown SM, Gass J, Rademakers R, Lindholm C, Snowden J, Adamson J, Sadovnick AD, Rollinson S, Cannon A, Dwosh E, Neary D, Melquist S, Richardson A, Dickson D, Berger Z, Eriksen J, Robinson T, Zehr C, Dickey CA, Crook R, McGowan E, Mann D, Boeve B, Feldman H, Hutton M. Mutations in progranulin cause tau-negative frontotemporal dementia linked to chromosome 17.Nature · 2006 · 442(7105):916-9doi:10.1038/nature05016 · PMID 16862116

   One of the two papers that made progranulin a brain protein: mutations likely to create null alleles cause frontotemporal dementia, the second commonest dementia under 65. It gives the gene's position 1.7 megabases from MAPT on 17q21.31 and the protein's description as a 68.5-kDa secreted growth factor involved in development, wound repair, inflammation and tumorigenesis.

6. 6
   Cruts M, Gijselinck I, van der Zee J, Engelborghs S, Wils H, Pirici D, Rademakers R, Vandenberghe R, Dermaut B, Martin JJ, van Duijn C, Peeters K, Sciot R, Santens P, De Pooter T, Mattheijssens M, Van den Broeck M, Cuijt I, Vennekens K, De Deyn PP, Kumar-Singh S, Van Broeckhoven C. Null mutations in progranulin cause ubiquitin-positive frontotemporal dementia linked to chromosome 17q21.Nature · 2006 · 442(7105):920-4doi:10.1038/nature05017 · PMID 16862115

   The companion paper, which names the mechanism as haploinsufficiency: one working copy is not enough. In a Belgian series of familial frontotemporal dementia, progranulin mutations were 3.5 times more frequent than mutations in the tau gene.

7. 7
   Smith KR, Damiano J, Franceschetti S, Carpenter S, Canafoglia L, Morbin M, Rossi G, Pareyson D, Mole SE, Staropoli JF, Sims KB, Lewis J, Lin WL, Dickson DW, Dahl HH, Bahlo M, Berkovic SF. Strikingly different clinicopathological phenotypes determined by progranulin-mutation dosage.Am J Hum Genet · 2012 · 90(6):1102-7doi:10.1016/j.ajhg.2012.04.021 · PMID 22608501

   Two siblings homozygous for a progranulin frameshift had neuronal ceroid lipofuscinosis, a different disease with a different age of onset from the dementia seen in people carrying one copy of the same kind of mutation. Dose, not presence or absence, decides which illness appears. A single family.

8. 8
   Smout MJ, Mulvenna JP, Jones MK, Loukas A. Expression, refolding and purification of Ov-GRN-1, a granulin-like growth factor from the carcinogenic liver fluke, that causes proliferation of mammalian host cells.Protein Expr Purif · 2011 · 79(2):263-70doi:10.1016/j.pep.2011.06.018 · PMID 21757010

   Establishes that the proliferative effect belongs to correctly folded monomeric protein, not to misfolded aggregate, and that it works at nanomolar doses on both cancerous and non-cancerous human bile duct cell lines. It is a methods paper; it contains no animal or human outcome.

9. 9
   Haugen B, Karinshak SE, Mann VH, Popratiloff A, Loukas A, Brindley PJ, Smout MJ. Granulin Secreted by the Food-Borne Liver Fluke Promotes Angiogenesis in Human Endothelial Cells.Front Med (Lausanne) · 2018 · 5:30doi:10.3389/fmed.2018.00030 · PMID 29503819

   Moves the blood-vessel claim from a bird embryo into human umbilical vein endothelial cells, where 10 and 40 nM recombinant protein significantly stimulated proliferation and tubule formation. PubMed's copy of this record drops the italicized species name from the title and the abstract; the fluke is Opisthorchis viverrini, whose granulin Ov-GRN-1 the paper tests. No animal or human endpoint.

10. 10
    Chaiyadet S, Tangkawattana S, Smout MJ, Ittiprasert W, Mann VH, Deenonpoe R, Arunsan P, Loukas A, Brindley PJ, Laha T. Knockout of liver fluke granulin, Ov-grn-1, impedes malignant transformation during chronic infection with Opisthorchis viverrini.PLoS Pathog · 2022 · 18(9):e1010839doi:10.1371/journal.ppat.1010839 · PMID 36137145

    This is the paper that measured cancer: the same parasite-gene knockout plus dietary nitrosamine, run to 24 weeks. Fewer hamsters developed high-grade bile duct cancer, fibrosis and proliferation fell, and mutant p53-positive cells fell sharply. It also walks back two earlier claims: knockout worms were recovered 38.9% less often, and the intended comparator gene turned out to be essential, so that control arm collapsed.

11. 11
    Ittiprasert W, Brindley PJ. CRISPR-based functional genomics for schistosomes and related flatworms.Trends Parasitol · 2024 · 40(11):1016-1028doi:10.1016/j.pt.2024.09.010 · PMID 39426911

    The state of the method, from the laboratory that built it. It confirms that gene knockout showed a liver fluke growth mediator contributes to disease progression, and it states plainly that the next milestone, heritably transgenic lines, has not been reached. That is the sentence that keeps every result here provisional.

12. 12
    Chaiyadet S, Sotillo J, Smout M, Cantacessi C, Jones MK, Johnson MS, Turnbull L, Whitchurch CB, Potriquet J, Laohaviroj M, Mulvenna J, Brindley PJ, Bethony JM, Laha T, Sripa B, Loukas A. Carcinogenic Liver Fluke Secretes Extracellular Vesicles That Promote Cholangiocytes to Adopt a Tumorigenic Phenotype.J Infect Dis · 2015 · 212(10):1636-45doi:10.1093/infdis/jiv291 · PMID 25985904

    The second, separate arm of the secretion story, and the only place in this chain where fluke molecules were found in human tissue: vesicle proteins were detected in bile from infected hamsters and from infected people. Internalized vesicles drove proliferation and interleukin-6 release, and anti-tetraspanin antibodies blocked uptake.

13. 13
    Bansal PS, Smout MJ, Wilson D, Cobos Caceres C, Dastpeyman M, Sotillo J, Seifert J, Brindley PJ, Loukas A, Daly NL. Development of a Potent Wound Healing Agent Based on the Liver Fluke Granulin Structural Fold.J Med Chem · 2017 · 60(10):4258-4266doi:10.1021/acs.jmedchem.7b00047 · PMID 28425707

    Four shortened versions of the molecule's front end were made and solved by nuclear magnetic resonance; both the two- and three-disulfide peptides drove human cholangiocyte proliferation and healed mouse wounds. This is the paper that turns a parasite carcinogen into a drug lead, and it is also the clearest statement of the translational interest the authors hold.

14. 14
    Chaiyadet S, Sotillo J, Krueajampa W, Thongsen S, Brindley PJ, Sripa B, Loukas A, Laha T. Vaccination of hamsters with Opisthorchis viverrini extracellular vesicles and vesicle-derived recombinant tetraspanins induces antibodies that block vesicle uptake by cholangiocytes and reduce parasite burden after challenge infection.PLoS Negl Trop Dis · 2019 · 13(5):e0007450doi:10.1371/journal.pntd.0007450 · PMID 31136572

    The vaccine idea tested: every vaccinated group yielded significantly fewer adult flukes than adjuvant-only controls, egg counts fell significantly only with recombinant Ov-TSP-2, and worms were shorter. Partial efficacy against infection intensity, not protection, and the antigen is a tetraspanin, not granulin.

15. 15
    Phumrattanaprapin W, Chaiyadet S, Brindley PJ, Pearson M, Smout MJ, Loukas A, Laha T. Orally Administered Bacillus Spores Expressing an Extracellular Vesicle-Derived Tetraspanin Protect Hamsters Against Challenge Infection With Carcinogenic Human Liver Fluke.J Infect Dis · 2021 · 223(8):1445-1455doi:10.1093/infdis/jiaa516 · PMID 32813017

    An oral vaccine of recombinant Bacillus subtilis spores gave 56% reductions in both adult flukes and fecal eggs in hamsters. It is the strongest translational result in this literature, and it comes from the tetraspanin arm.

16. 16
    He Z, Ong CH, Halper J, Bateman A. Progranulin is a mediator of the wound response.Nat Med · 2003 · 9(2):225-9doi:10.1038/nm816 · PMID 12524533

    The human protein as a wound protein: its message rose in the inflammatory cells, fibroblasts and vessels of mouse puncture wounds, and applying it drew in neutrophils, macrophages, fibroblasts and new vessels while making isolated cells divide, migrate and form tubules. Mouse skin, not human wounds.

17. 17
    Gillett DA, Wallings RL, Uriarte Huarte O, Tansey MG. Progranulin and GPNMB: interactions in endo-lysosome function and inflammation in neurodegenerative disease.J Neuroinflammation · 2023 · 20(1):286doi:10.1186/s12974-023-02965-w · PMID 38037070

    Sets out the modern, non-growth-factor job of progranulin: losing it disturbs the cell's waste-handling compartments and raises another disease-linked protein with age. A review of mechanism, with no clinical outcome.

18. 18
    Buisson Y. [Control of Opisthorchis viverrini infection for cholangiocarcinoma prevention].Bull Soc Pathol Exot · 2017 · 110(1):61-67doi:10.1007/s13149-017-0544-8 · PMID 28105582

    A French-language review with an English abstract that carries the dates: the International Agency for Research on Cancer classified Opisthorchis viverrini as a Group 1 human carcinogen in 1994 and Clonorchis sinensis in 2009. It also gives the worm's 10- to 30-year residence in the bile ducts, the resulting hyperplasia and periductal fibrosis, and the three-strand mechanism in which secreted mitogens are one strand.

19. 19
    Smout MJ, Laha T, Chaiyadet S, Brindley PJ, Loukas A. Mechanistic insights into liver-fluke-induced bile-duct cancer.Trends Parasitol · 2024 · 40(12):1183-1196doi:10.1016/j.pt.2024.10.012 · PMID 39521672

    The current synthesis of the causal chain, and the cleanest statement of it: secreted growth factors, enzymes and vesicles, plus inflammation and repeated wounding where the worm grazes, give hyperplasia, fibrosis and finally malignant transformation. Four of its five authors are senior authors on the primary papers, so treat it as the field's own framing rather than independent adjudication.

20. 20
    Chaiyadet S, Ittiprasert W, Smout MJ, Khowawisetsut L, Ruangsuwast A, Brindley PJ, Loukas A, Laha T. Gene Editing of a Carcinogenic Liver Fluke Tetraspanin Impairs Parasite Surface Biogenesis and Extracellular Vesicle Uptake by Human Host Cells.J Infect Dis · 2026 · 233(2):e510-e520doi:10.1093/infdis/jiaf466 · PMID 40905941

    The newest primary paper in this line. Knocking out the surface protein Ov-TSP-2 left the tegument eroded and vacuolated, with tegumental area down 40% to 47%, and cut vesicle release and uptake by human cells — all of it in culture, with no hamster infected here, which explains why worms of that genotype failed to reach adulthood in the 2022 cancer study [12]. It also draws the dividing line this page needs: Ov-GRN-1 is a soluble secreted growth factor, is not carried in the vesicles, and its knockout is not lethal to the worm in vivo.

21. 21
    Insinga G, Tomassetti C, Satasiya V, Gimigliano F, Giordano A, Giurisato E, Morrione A. Progranulin and Inflammation-Induced Cancer: An Important Player in the Tumor Microenvironment?J Cell Physiol · 2026 · 241(5):e70188doi:10.1002/jcp.70188 · PMID 42112954

    The current review of human progranulin in tumors: proliferation, migration and epithelial-to-mesenchymal transition, plus reprogramming of the surrounding tissue toward immune suppression, activation of fibroblasts, remodeling of the matrix and new vessels. Note that it still describes signaling through a tumor necrosis factor receptor, a route a dedicated replication study failed to confirm, so the receptor question is not closed.

22. 22
    Smout MJ, Sotillo J, Laha T, Papatpremsiri A, Rinaldi G, Pimenta RN, Chan LY, Johnson MS, Turnbull L, Whitchurch CB, Giacomin PR, Moran CS, Golledge J, Daly N, Sripa B, Mulvenna JP, Brindley PJ, Loukas A. Carcinogenic Parasite Secretes Growth Factor That Accelerates Wound Healing and Potentially Promotes Neoplasia.PLoS Pathog · 2015 · 11(10):e1005209doi:10.1371/journal.ppat.1005209 · PMID 26485648

    The mechanism paper: silencing the gene measurably weakened the flukes' secretions — wounds in cholangiocyte and cancer-cell monolayers closed significantly more slowly and proliferation fell by about 48%, rather than the activity being abolished — recombinant protein accelerated healing of mouse skin wounds and was angiogenic in the quail egg membrane, and labeled protein was taken up inside human cholangiocytes without passing through lysosomes or endosomes. The wound and vessel endpoints were measured in mouse skin and a bird embryo, not in a bile duct.

23. 23
    Dastpeyman M, Wilson DT, Loukas A, Smout MJ, Daly NL. The C-terminal region of granulin-1 promotes cell proliferation in the absence of regular secondary structure.Aust J Chem · 2025 · 78(11)doi:10.1071/CH25079 · PMID 42428249

    The most recent chemistry on the molecule: a peptide from its tail end does not fold into the granulin shape at all and still makes cells divide, so sequence may matter more than three-dimensional structure. PubMed's title drops an italicized element before “granulin-1”; the peptide is derived from Ov-GRN-1. In vitro only.

24. 24
    Takahashi H, Strittmatter SM. The role of endolysosomal progranulin and TMEM106B in neurodegenerative diseases.Mol Neurodegener · 2025 · 20(1):86doi:10.1186/s13024-025-00873-6 · PMID 40713630

    A 2025 review placing progranulin in the cell's digestive compartments alongside TMEM106B, including the regulation of complex lipids, and surveying the preclinical models and treatment attempts. Review, not primary data.

25. 25
    Qian MB, Keiser J, Utzinger J, Zhou XN. Clonorchiasis and opisthorchiasis: epidemiology, transmission, clinical features, morbidity, diagnosis, treatment, and control.Clin Microbiol Rev · 2024 · 37(1):e0000923doi:10.1128/cmr.00009-23 · PMID 38169283

    The most recent authoritative synthesis of the three liver flukes: complex life cycles, chronic infection complicated most importantly by cholangiocarcinoma, stool microscopy for surveys, immunologic and molecular methods in reference laboratories, and praziquantel-based preventive chemotherapy. It names the research gaps honestly, including burden and diagnostics.

26. 26
    Young ND, Nagarajan N, Lin SJ, Korhonen PK, Jex AR, Hall RS, Safavi-Hemami H, Kaewkong W, Bertrand D, Gao S, Seet Q, Wongkham S, Teh BT, Wongkham C, Intapan PM, Maleewong W, Yang X, Hu M, Wang Z, Hofmann A, Sternberg PW, Tan P, Wang J, Gasser RB. The Opisthorchis viverrini genome provides insights into life in the bile duct.Nat Commun · 2014 · 5:4378doi:10.1038/ncomms5378 · PMID 25007141

    The 634.5-megabase draft genome and transcriptomes, showing a metabolism tuned to the fat-rich diet of bile, and providing further evidence that these flukes secrete proteins that directly drive host cell proliferation. It is the reference that made targeted gene editing in this species possible.

27. 27
    Hembasat T, Chaiyadet S, Ittiprasert W, Smout MJ, Young ND, Loukas A, Brindley PJ, Laha T. Peptide derived from progranulin of the carcinogenic liver fluke, Opisthorchis viverrini stimulates cell hyperproliferation and proinflammatory cytokine production.Res Sq · 2023 · preprint, not peer revieweddoi:10.21203/rs.3.rs-2586058/v1 · PMID 36993607

    A preprint, flagged as such. It characterizes in detail the fluke's full progranulin: a gene of thirteen coding stretches whose protein is 846 amino acids long, counted here as seven and a half granulin modules, most strongly expressed in the infective larva, and a peptide from it made cholangiocytes divide and raised interleukin-6 and interleukin-8. It did not discover the gene: the 2014 genome [28] had already reported this progranulin, as eight granulin subunits, together with a second single-domain granulin, Ov-GRN-2, and the 2015 mechanism paper [4] noted both while adding that neither product was detectable in the secreted proteome. This preprint has not passed peer review, and whether either protein reaches host tissue in quantity is unsettled.

28. 28
    Hu F, Padukkavidana T, Vægter CB, Brady OA, Zheng Y, Mackenzie IR, Feldman HH, Nykjaer A, Strittmatter SM. Sortilin-mediated endocytosis determines levels of the frontotemporal dementia protein, progranulin.Neuron · 2010 · 68(4):654-67doi:10.1016/j.neuron.2010.09.034 · PMID 21092856

    Identifies sortilin as the receptor that pulls progranulin into cells and delivers it to lysosomes: mice lacking sortilin carry 2.5- to 5-fold more progranulin, and the 50% shortfall seen in patients was fully corrected in mice by removing sortilin. Mouse and cell work; it is the basis of the drug strategy of blocking the receptor rather than replacing the protein.

29. 29
    Tang W, Lu Y, Tian QY, Zhang Y, Guo FJ, Liu GY, Syed NM, Lai Y, Lin EA, Kong L, Su J, Yin F, Ding AH, Zanin-Zhorov A, Dustin ML, Tao J, Craft J, Yin Z, Feng JQ, Abramson SB, Yu XP, Liu CJ. The growth factor progranulin binds to TNF receptors and is therapeutic against inflammatory arthritis in mice.Science · 2011 · 332(6028):478-84doi:10.1126/science.1199214 · PMID 21393509

    The much-cited claim that progranulin binds tumor necrosis factor receptors directly and antagonizes inflammation, with an engineered fragment protecting mice from arthritis. Mouse models; the binding claim was contradicted two years later, and the dispute is still visible in the current cancer literature.

30. 30
    Chen X, Chang J, Deng Q, Xu J, Nguyen TA, Martens LH, Cenik B, Taylor G, Hudson KF, Chung J, Yu K, Yu P, Herz J, Farese RV, Kukar T, Tansey MG. Progranulin does not bind tumor necrosis factor (TNF) receptors and is not a direct regulator of TNF-dependent signaling or bioactivity in immune or neuronal cells.J Neurosci · 2013 · 33(21):9202-13doi:10.1523/JNEUROSCI.5336-12.2013 · PMID 23699531

    The failed replication, done deliberately and reported in detail: using the same methods the authors reproduced progranulin binding to sortilin and TNF binding to its receptors, but found no binding of progranulin to those receptors and no antagonism of their signaling. A negative result in cells and mouse-derived cells; it does not rule out indirect effects.

31. 31
    Papatpremsiri A, Smout MJ, Loukas A, Brindley PJ, Sripa B, Laha T. Suppression of Ov-grn-1 encoding granulin of Opisthorchis viverrini inhibits proliferation of biliary epithelial cells.Exp Parasitol · 2014 · 148:17-23doi:10.1016/j.exppara.2014.11.004 · PMID 25450776

    The pre-CRISPR silencing experiment: double-stranded RNA knocked the gene down 95% by day 3 and about 100% by day 7, and co-culture with silenced flukes cut host cell hyperproliferation by 25% in normal cholangiocytes and 92% in a cancer line. It also reported that silenced flukes were less viable in culture, a claim the later in vivo knockouts did not support.

32. 32
    Ittiprasert W, Mann VH, Karinshak SE, Coghlan A, Rinaldi G, Sankaranarayanan G, Chaidee A, Tanno T, Kumkhaek C, Prangtaworn P, Mentink-Kane MM, Cochran CJ, Driguez P, Holroyd N, Tracey A, Rodpai R, Everts B, Hokke CH, Hoffmann KF, Berriman M, Brindley PJ. Programmed genome editing of the omega-1 ribonuclease of the blood fluke.eLife · 2019 · 8:e41337doi:10.7554/eLife.41337 · PMID 30644357

    The companion paper published the same day, in Schistosoma mansoni: editing the omega-1 gene abolished Th2 polarization and shrank lung granulomas, and it achieved homology-directed repair, which the liver fluke work did not. It is why “first gene editing in a flatworm” belongs to a pair of papers. PubMed's title drops the italicized species name.

33. 33
    McVeigh P, Maule AG. Can CRISPR help in the fight against parasitic worms?eLife · 2019 · 8:e44382doi:10.7554/eLife.44382 · PMID 30702425

    An independent commentary from Queen's University Belfast on the two 2019 papers, by authors involved in neither. It is the only outside voice on the gene-editing work itself, and it adds no data. A small external commentary and bioinformatics literature on Ov-GRN-1 does exist beyond it (for example Botelho 2016, Parasitol Res 115:2913-4, PMID 27130317; Machicado 2016, Springerplus 5:773, PMID 27386259), but none of it is a replication.

34. 34
    Bagyinszky E, An SSA. Targeting Granulin Haploinsufficiency in Frontotemporal Dementia: From Genetic Mechanisms to Therapeutics.Int J Mol Sci · 2025 · 26(20):9960doi:10.3390/ijms26209960 · PMID 41155255

    The current review of what is being done about too little progranulin: loss-of-function mutations cut levels by up to about 50%, and gene therapy delivered by adeno-associated virus, stop-codon readthrough compounds, sortilin inhibitors and antisense oligonucleotides are all being tried. It states the two problems squarely: limited success in trials so far, and the oncogenic risk of pushing levels too high.

35. 35
    Raj K, Kumari N, Bharti P, Kumar R, Kumar D. Decoding Frontotemporal Dementia: Genes, Proteins and Prospects for Precision Medicine.Curr Alzheimer Res · 2026doi:10.2174/0115672050478591260821105132 · PMID 42708260

    Cited for one narrow point: progranulin is among the fluid markers now used in frontotemporal dementia for classification, prognosis and monitoring, in a field where no disease-modifying therapy yet exists and treatment remains symptomatic. A narrative review.

36. 36
    Dastpeyman M, Bansal PS, Wilson D, Sotillo J, Brindley PJ, Loukas A, Smout MJ, Daly NL. Structural Variants of a Liver Fluke Derived Granulin Peptide Potently Stimulate Wound Healing.J Med Chem · 2018 · 61(19):8746-8753doi:10.1021/acs.jmedchem.8b00898 · PMID 30183294

    Shows that proline residues govern how the shortened peptide folds and how well it works, which is chemistry in service of yield rather than biology of infection. It says nothing about cancer.

37. 37
    Arunsan P, Chaidee A, Cochran CJ, Mann VH, Tanno T, Kumkhaek C, Smout MJ, Karinshak SE, Rodpai R, Sotillo J, Loukas A, Laha T, Brindley PJ, Ittiprasert W. Liver fluke granulin promotes extracellular vesicle-mediated crosstalk and cellular microenvironment conducive to cholangiocarcinoma.Neoplasia · 2020 · 22(5):203-216doi:10.1016/j.neo.2020.02.004 · PMID 32244128

    The other knockout, and the likely source of the confusion: here CRISPR/Cas9 inactivated the human progranulin gene in a cultured human bile duct cell line, so that the fluke protein's effect could be read without the host's own progranulin in the way. Nothing living was edited, no animal was used, and no disease was measured.

38. 38
    Chaiyadet S, Sotillo J, Krueajampa W, Thongsen S, Smout M, Brindley PJ, Laha T, Loukas A. Silencing of Tetraspanin Gene Expression Results in Reduced Secretion of Extracellular Vesicles.Front Cell Infect Microbiol · 2022 · 12:827521doi:10.3389/fcimb.2022.827521 · PMID 35223551

    Silencing the two tetraspanins cut vesicle secretion, and blocking antibodies or endocytosis inhibitors cut vesicle uptake and the proliferation that follows. The authors also note that purification standards for helminth vesicles are ill defined, which is a reason to treat vesicle counts across papers with care. PubMed's title drops the italicized species name.

39. 39
    Ruangsuwast A, Smout MJ, Brindley PJ, Loukas A, Laha T, Chaiyadet S. Tetraspanins from the liver fluke Opisthorchis viverrini stimulate cholangiocyte migration and inflammatory cytokine production.Folia Parasitol · 2023 · 70:2023.017doi:10.14411/fp.2023.017 · PMID 37752807

    Recombinant loops of both tetraspanins significantly enhanced bile duct cell migration; the Ov-TSP-3 loop alone also raised interleukin-6 and interleukin-8 gene expression and drove proliferation at 24 and 48 hours. It shows that granulin is not the only fluke protein with mitogenic and inflammatory effects on host cells. Cell culture only.

40. 40
    Chaiyadet S, Sotillo J, Smout M, Cooper M, Doolan DL, Waardenberg A, Eichenberger RM, Field M, Brindley PJ, Laha T, Loukas A. Small extracellular vesicles but not microvesicles from Opisthorchis viverrini promote cell proliferation in human cholangiocytes.bioRxiv · 2023 · preprint, not peer revieweddoi:10.1101/2023.05.22.540805 · PMID 37292777

    Flagged as a preprint: the 120k small-vesicle fraction promoted H69 cholangiocyte proliferation at different concentrations while the 15k microvesicle fraction did not, with proteomic differences between them and microRNAs in the active fraction predicted to target inflammation, immune and apoptosis pathways. Target prediction is computational. Not peer reviewed as of October 2026.

41. 41
    Lishai E, Pakharukova M. The Liver Fluke Opisthorchis felineus Exosomal tRNA-Derived Small RNAs as Potential Mediators of Host Manipulation.Biomolecules · 2026 · 16(2):244doi:10.3390/biom16020244 · PMID 41750314

    A current paper from an independent Russian group, and the one source here that is about a different fluke: Opisthorchis felineus, not O. viverrini. Its exosome-like vesicles are enriched in particular tRNA-derived small RNAs whose predicted human targets are over-represented among the genes that change in cholangiocytes treated with O. felineus secretions. The targeting is computational prediction plus a correlation, not a demonstrated mechanism, and the authors state that O. viverrini's dominant fragments are different ones, so the result does not transfer. PubMed's title drops the italicized species name.

42. 42
    Upontain S, Sereerak P, Laha T, Sripa B, Tangkawatana P, Brindley PJ, Tangkawatana S. Granulin Expression in Hamsters during Opisthorchis viverrini Infection-Induced Cholangiocarcinogenesis.Asian Pac J Cancer Prev · 2018 · 19(9):2437-2445doi:10.22034/APJCP.2018.19.9.2437 · PMID 30255697

    The host's own granulin also rises as the lesions progress, most of all in hamsters given both fluke and nitrosamine, tracking a marker of cell division. It is immunohistochemistry and western blot in 120 hamsters, reported without p-values in the abstract, and it is association, not cause.

43. 43
    Chaiyadet S, Krueajampa W, Hipkaeo W, Plosan Y, Piratae S, Sotillo J, Smout M, Sripa B, Brindley PJ, Loukas A, Laha T. Suppression of mRNAs encoding CD63 family tetraspanins from the carcinogenic liver fluke Opisthorchis viverrini results in distinct tegument phenotypes.Sci Rep · 2017 · 7(1):14342doi:10.1038/s41598-017-13527-5 · PMID 29084967

    Identifies Ov-TSP-2 and Ov-TSP-3 as surface tetraspanins present at all life stages and recognized by sera from infected hamsters and infected people, and shows that silencing them deforms the worm's skin. Read in hindsight it is the warning that this gene would be a poor comparator in a knockout experiment.

44. 44
    Chaidee A, Charoenram N, Sengthong C, Dangtakot R, Pinlaor P, Pongking T, Pinlaor S. Transcriptome changes of liver fluke Opisthorchis viverrini in diabetic hamsters.Parasite · 2024 · 31:54doi:10.1051/parasite/2024056 · PMID 39269256

    A largely separate Khon Kaen group, arriving at granulin from the opposite direction: in diabetic hamsters the worms were stunted and laid fewer eggs, yet virulence genes including granulin, tetraspanins and thioredoxins were turned up. Its first author has co-authored two papers with the gene-editing team [10][13], so it is not fully independent of them. This is association within an animal model, with no causal test, and the abstract gives differentially expressed gene counts (4,314 at 21 days, 567 at 35) without the fold-changes for granulin itself.

45. 45
    Qian MB, Patel C, Palmeirim MS, Wang X, Schindler C, Utzinger J, Zhou XN, Keiser J. Efficacy of drugs against clonorchiasis and opisthorchiasis: a systematic review and network meta-analysis.Lancet Microbe · 2022 · 3(8):e616-e624doi:10.1016/S2666-5247(22)00026-X · PMID 35697047

    The honest version of “praziquantel works”: roughly 25 million people infected, predicted cure rates for O. viverrini of 93.8% for a split 50 plus 25 mg/kg day and 92.1% for a single 50 mg/kg dose, from a trial literature the authors themselves grade as low quality. Curing the infection is not the same as undoing the damage, which this analysis does not address.

46. 46
    Phumrattanaprapin W, Pearson M, Pickering D, Tedla B, Smout M, Chaiyadet S, Brindley PJ, Loukas A, Laha T. Monoclonal Antibodies Targeting an Extracellular Vesicle Tetraspanin Protect Hamsters against Challenge Infection.Vaccines (Basel) · 2021 · 9(7):740doi:10.3390/vaccines9070740 · PMID 34358156

    The first passive immunization against this infection: of two anti-TSP-2 IgM monoclonals, 1D6 and 3F5, transferred 3F5 alone, without vaccination, cut adult flukes by 74.5% and eggs per gram by 67.4% against an isotype control. Both blocked vesicle uptake by human bile duct cells in vitro. It shows the protection is antibody-mediated. PubMed's title drops the italicized species name.

47. 47
    Ward M, Carter LP, Huang JY, Maslyar D, Budda B, Paul R, Rosenthal A. Phase 1 study of latozinemab in progranulin-associated frontotemporal dementia.Alzheimers Dement (N Y) · 2024 · 10(1):e12452doi:10.1002/trc2.12452 · PMID 38356474

    The proof that the sortilin strategy works pharmacologically in humans: eight symptomatic FTD-GRN participants, favorable safety and pharmacokinetics, progranulin in plasma and cerebrospinal fluid raised to approximately healthy-volunteer levels. Eight people, no clinical endpoint, and sponsored by the company developing the drug.

48. 48
    Lee YB, Lee DY, Walker Z, Allison J, Kaliszewska A, Fernandes A, Zhang KY, Bekele Y, Hosseini P, Joubert R, Furtado J, Gumus M, Brock O, Bienemann A, Gill S, Murray S, Mitchell N, Bouchard P, Bloom A, Isaac J, Miranda CJ, Shaw CE. Intrathalamic delivery of adeno-associated viral vector expressing progranulin as gene therapy for GRN-related frontotemporal dementia.Mol Ther Methods Clin Dev · 2025 · 33(4):101622doi:10.1016/j.omtm.2025.101622 · PMID 41311476

    The gene-therapy arm of the same effort, still preclinical: AVB-101 reversed pathology in Grn-deficient mice, reached most prefrontal cortical neurons in sheep when given into the thalamus but not into the cisterna magna, and was tolerated in macaques for six months. Animals only; written by the developing company.

49. 49
    Wang C, He Q, Yin Y, Wu Y, Li X. Granulin Promotes Malignant Transformation of Hepatocyte Through EGFR-Mediated RAS/MAPK/ERK and PI3K/Akt Signaling Pathways.Front Cell Infect Microbiol · 2021 · 11:734750doi:10.3389/fcimb.2021.734750 · PMID 34858869

    An independent group at Sun Yat-Sen University, no overlap with the Ov-GRN-1 authors: overexpressing the Clonorchis sinensis granulin in a normal human liver cell line drove proliferation, migration, invasion and epithelial-to-mesenchymal transition markers through EGFR-mediated RAS/MAPK/ERK and PI3K/Akt signaling. Transfected cell lines only, and hepatocytes rather than cholangiocytes. PubMed's title drops the italicized species name.

50. 50
    He Q, Pan X, Yin Y, Xu A, Yi X, Wu Y, Li X. Clonorchis sinensis granulin promotes malignant transformation of human intrahepatic biliary epithelial cells through interaction with M2 macrophages via regulation of STAT3 phosphorylation and the MEK/ERK pathway.Parasit Vectors · 2023 · 16(1):139doi:10.1186/s13071-023-05765-6 · PMID 37095578

    The nearest thing to independent corroboration in this whole story: a separate Chinese group shows the sister fluke's granulin produces hyperplasia and abnormal proliferation of normal human bile duct lining cells and biliary damage in mice, working through M2 macrophage polarization, interleukin-6, JAK2/STAT3 and MEK/ERK, with the MEK/ERK inhibitor PD98059 reversing it. A different species, so not a replication of the Ov-grn-1 knockout, and no cancer endpoint in an animal.

51. 51
    Rujirakul R, Ueng-arporn N, Kaewpitoon SJ, Loyd RA, Kaewthani S, Kaewpitoon N. Risk Areas of Liver Flukes in Surin Province of Thailand using Geographic Information System.J Med Assoc Thai · 2015 · 98 Suppl 4:S22-6PMID 26201130

    Recorded for transparency, not used for any claim: a systematic check of the retraction flag across this literature returns exactly one retracted paper, this risk-mapping study by a group unrelated to the granulin work. None of the papers behind the granulin story carries a retraction or an expression of concern in PubMed.

52. 52
    Rumgay H, Georges D, Huang Y, Hirabayashi M, Shah R, de Martel C, Park JY, Soerjomataram I, Clifford GM. Global burden of cancer attributable to infections in 2024: a worldwide incidence analysis.Lancet Oncol · 2026 · 27(10):1237-1248doi:10.1016/S1470-2045(26)00307-4 · PMID 42805198

    The current IARC count of infection-attributable cancer: 2.3 million new cases in 2024, 12% of all cancers, across twelve Group 1 agents including Opisthorchis viverrini and Clonorchis sinensis. It is a modeled population-attributable-fraction estimate built on GLOBOCAN incidence, not a direct measurement, and the fluke share is small beside H. pylori and human papillomavirus.

53. 53
    Taweesin W, Kulwong S, Jamnongkan W, Taron W, Siriprayong A, Sitthirak S, Techasen A, Pitaksakulrat O, Namwat N, Klanrit P, Loilome W, Wangwiwatsin A. Transcriptomics analysis of in vivo Opisthorchis viverrini: Stage-specific gene expression and novel genes with stable expression in mammalian stages for opisthorchiasis intervention.PLoS Negl Trop Dis · 2025 · 19(12):e0013714doi:10.1371/journal.pntd.0013714 · PMID 41411373

    RNA sequencing of the fluke inside hamsters at 14, 42 and 180 days, from a largely separate Khon Kaen group, identifying stage-specific programs and 2,011 stably expressed genes as drug, diagnostic and vaccine candidates. Cited here for the burden figures and for showing how much of the worm's expression shifts between juvenile and chronic adult stages — the context in which any one measurement of granulin output sits.