HGF (hepatocyte growth factor) is not a peptide. It is a large secreted glycoprotein of roughly 90 kDa that circulates as a heterodimer of an α- and a β-chain joined by a disulfide bridge. The confusion is common because HGF often appears in the same catalogues as short synthetic peptides such as BPC-157 or GHK-Cu, but in size, manufacture and behaviour it is a fundamentally different molecule.

A protease that forgot how to cut

The most interesting structural fact about HGF is its ancestry. The precursor pro-HGF (728 amino acids) is made as an inactive single chain and is activated by extracellular proteolytic cleavage carried out by serine proteases, HGF activator (HGFA), matriptase, hepsin. Unlike a peptide that is active the moment it is synthesised, HGF requires a separate activation step in the tissue.

Its sequence is about one third identical to plasminogen: four kringle domains in the α-chain and a serine-protease-like domain in the β-chain. But that protease domain is catalytically dead, it has lost the catalytic triad residues and cleaves nothing. HGF is essentially a protease scaffold that evolution repurposed into a signalling molecule: it kept the shape of an enzyme without the function.[8]

One protein, three independent discoveries

HGF was discovered three times, by different laboratories, under different names, for different reasons:

  • Hepatocyte growth factor, a mitogen for hepatocytes found in rat serum after partial hepatectomy.
  • Scatter factor (SF), a factor that made compact epithelial colonies "scatter" and disperse.
  • Tumor cytotoxic factor, an activity that killed certain tumour cell lines.

The three activities looked unrelated. Molecular cloning in 1989[1] and subsequent biochemistry showed they were the same protein. This is more than a historical anecdote: it explains immediately why HGF appears to "do a bit of everything", proliferation, cell motility, morphogenesis, survival. These are not side effects but different readings of one programme.[8]

Ligand of an oncogene: the central inconvenience

In 1991 the HGF receptor was identified as the product of the c-Met (MET) proto-oncogene[2], a receptor tyrosine kinase. HGF remains the only known ligand of c-Met, and c-Met the only known receptor for HGF, a strictly monogamous pair, which is rare in biology.

c-Met activation triggers what is called the invasive growth programme: cells loosen their junctions, become motile, invade matrix, build tubular structures. In development and repair this is useful. In a tumour it is metastasis.[5][7]

Hence the paradox worth understanding before any work with HGF: the pharmaceutical industry is moving in the opposite direction. The approved drugs acting on this axis are c-Met inhibitors for non-small-cell lung cancer with MET exon 14 skipping mutations, capmatinib[11] and tepotinib[12]. Billions have gone into switching this receptor off. No c-Met agonist is approved anywhere.

What genetics and liver regeneration show

HGF is not a subtle modulator. Mice lacking Hgf or Met die in utero with placental and liver defects and impaired migration of muscle progenitors.[3][4] This is an essential developmental gene.

In the adult the role is narrower but concrete: conditional deletion of Met in hepatocytes impairs full liver regeneration after partial hepatectomy.[6] That body of work, together with preclinical models in which HGF opposes TGF-β1-driven fibrogenesis and hepatocyte apoptosis, is what established HGF as an anti-fibrotic, pro-regenerative axis.

A useful contrast: anti-fibrotic effects are also pursued through entirely different receptors, B7-33, a relaxin analogue, acts via RXFP1 rather than c-Met. Different entry points into the same connective-tissue problem.

Why recombinant HGF barely works as a systemic drug

This is the bottleneck that shaped everything afterwards. As a systemic medicine recombinant HGF carries several unfavourable properties at once:

  • Very short plasma residence, the protein is cleared rapidly, largely by the liver.
  • Strong binding to heparan sulfates in the extracellular matrix, injected protein diffuses poorly and stays near the injection site.
  • Difficult manufacture, a glycosylated two-chain protein that additionally needs correct proteolytic activation; nothing like solid-phase peptide synthesis.
  • An oncology caveat, sustained systemic c-Met activation runs against everything known about invasive growth.[7]

Clinical development therefore avoided "inject the protein systemically" and took two detours.

Detour 1: deliver the gene, not the protein

VM202 (Engensis, donaperminogene seltoplasmid) is a non-viral plasmid DNA encoding two HGF isoforms (HGF728 and HGF723) via the hybrid HGF-X7 construct, injected intramuscularly so the tissue itself produces the protein locally.

What the studies actually showed, without embellishment:

  • Phase 2 in painful diabetic peripheral neuropathy[9]: in the lower-dose arm (8 mg per leg) 48.4% of participants reached ≥50% pain reduction at day 90 versus 17.6% on placebo. But the higher dose (16 mg) gave 27.8% and was not statistically significant, a non-monotonic dose response, which by itself weakens confidence in the finding.
  • Phase 3 in the same indication[13]: the primary endpoint was not met, superiority over placebo was not demonstrated.
  • The programme was also studied in diabetic foot ulcers[14], ALS and Charcot-Marie-Tooth disease 1A[15], largely early or interim data, not established efficacy.

Net position: tolerability in these studies was acceptable; efficacy at the strongest level of evidence was not confirmed.

Detour 2: deliver the protein locally

If systemic pharmacokinetics are hopeless, the alternative is to place the protein directly into the compartment of interest. Kringle Pharma is developing KP-100IT, intrathecal recombinant human HGF for acute spinal cord injury. A published double-blind randomised phase I/II trial reported an acceptable safety profile, it is a safety and exploratory-efficacy study, not proof of clinical benefit.[10] The same logic underlies neuroregenerative programmes such as ARA-290: act locally where systemic exposure is undesirable.

What remains unresolved

  • No approved HGF agonist exists, protein or gene-based, in any jurisdiction.
  • The therapeutic window is unknown. Regeneration and invasive growth are the same c-Met programme. How to obtain the first without the second is still unsolved.
  • Dose response is inconsistent, the non-monotonic phase 2 result has no convincing explanation.
  • Human data outside local delivery and gene therapy are essentially absent. Most "HGF repairs tissue" claims rest on rodent and cell-culture preclinical work.

A practical note on reagent identity

Research catalogues sell several different molecules under the name "HGF", and they are not interchangeable:

  • full-length two-chain HGF, the active form;
  • single-chain pro-HGF, requires protease activation and otherwise behaves differently in assay;
  • NK1 / NK2 fragments, truncated variants whose agonist or antagonist behaviour depends on context and heparin concentration.

The expression system (mammalian cells versus E. coli) determines glycosylation and folding. For reproducibility the form, expression system and activation method must be fixed in the protocol, otherwise results from different lots are not comparable.

Status of this material

HGF is presented here strictly as a research-use-only (RUO) reagent. It is not a medicine, not a dietary supplement, not a cosmetic ingredient, and is not intended for human consumption or veterinary use. The information above describes published scientific research and is not medical advice, a dosing regimen, or a claim of therapeutic benefit.

Related catalogue section: healing and regeneration.