The HIV field is NOT drugging LEDGF: LEDGINs and ALLINIs bind integrase, and the proof is that they still work in a LEDGF knockout cell
also called LEDGIN misconception · ALLINI target · NCINI
People often say the HIV field has spent twenty years making drugs against LEDGF, so a leukaemia drug should be easy. It has not. Those drugs bind the virus's own enzyme, on the face where LEDGF would dock, and they still work in cells that have no LEDGF at all.
The same thing, for a scientist
This is a correction to a premise that recurs whenever the LEDGF axis is proposed as a leukaemia target, including in this project's own dispatch. **LEDGINs (also ALLINIs, NCINIs, INLAIs) bind HIV-1 integrase, not LEDGF.** Their pocket is on the integrase catalytic core domain dimer interface - the face integrase presents to the LEDGF IBD - and their dominant antiviral mechanism is now understood to be drug-induced integrase hypermultimerisation during maturation, with the compounds acting as molecular glue that nucleates a CCD-CTD interface using integrase residues Tyr226, Trp235 and Lys266. The decisive experiment was published in 2012 and is unusually clean: in a human somatic **LEDGF/p75 knockout** cell line in which exons 11-14 encoding the IBD were deleted, LEDGINs *'remained active even in the absence of LEDGF/p75 by blocking the interaction with the IBD of HRP-2'*. A compound that retains activity when its supposed target has been deleted is not an inhibitor of that target. So the leukaemia programme inherits **no tool compound** from HIV. What it does inherit is real but smaller: the AlphaScreen/AlphaLISA assay format for the IN-IBD interaction, the structural map of the IBD surface, and one published proof that the IBD can be competed off the menin-MLL complex by a ligand.
Which molecule binds which protein
| Agent class | Binds | Consequence | Transferable to leukaemia? |
|---|---|---|---|
| LEDGINs / ALLINIs / NCINIs (BI-D, GS-9822, pirmitegravir, BI-224436, GSK3739936) | HIV-1 integrase CCD dimer pocket | blocks IN-IBD contact; induces IN hypermultimerisation | No. There is no integrase in a leukaemia cell |
| Cyclic peptide CP65 | LEDGF IBD (the MLL-binding face) | impairs clonogenic growth of primary murine MLL-AF9 blasts | Yes in principle; a peptide, not a drug |
| eGFP-LEDGF(325-530), LEDGF(375-386), LEDGF(424-435) | acts as a competing IBD, dominant-negative | impairs clonogenic growth; extends latency | Yes in principle; a genetic construct |
| Ebselen | LEDGF/p75, via selenium-sulphur bonds to reduced thiols | disrupts the IN-LEDGF interaction, DTT-reversible | technically the only direct LEDGF binder; see drug-ebselen-is-the-only-reported-direct-ledgf-binder |
| HRP-2/HDGFL2 PWWP fragments; LEDGF PWWP fragment series | LEDGF and HRP-2 PWWP | occupies the H3K36me2/3 cage at double-digit micromolar | Yes; see target-ledgf-pwwp-fragment-chemistry-and-its-ceiling |
The row that matters is the first. Every compound the phrase 'the HIV field has been drugging this protein for years' refers to is in that row.
The knockout experiment, and why it is decisive
Schrijvers et al. built a human somatic cell line in which PSIP1 exons 11-14, encoding the IBD, were removed by homologous recombination 1. HIV-1 replication was severely delayed for laboratory strains and defective for clinical isolates, and the residual replication ran through HRP-2 (HDGFL2), the only other cellular protein carrying an IBD. LEDGINs remained active in those cells.
Two conclusions follow, and the second is the one this graph keeps needing:
- LEDGINs act on integrase. Target deletion does not abolish their activity.
- HRP-2 is a functional substitute for the LEDGF IBD in at least one biological system. That is the paralogue-buffering problem this beat records separately in
mechanism-hdgfl2-hrp2-buffers-the-ledgf-ibd, and it was demonstrated in virology a decade before anyone asked it in leukaemia.
What actually is transferable
- Assay format. The IN-LEDGF interaction was screened at scale in AlphaScreen luminescent proximity format as early as 2008 3. The same format, run against the IBD and a KMT2A 111-160 phosphopeptide, is a ready-made primary screen for the leukaemia interface, and this project's structural beat already recorded an AlphaLISA competition on exactly this pair.
- The structural map. PDB 2B4J resolves the integrase-IBD complex at 2.02 A, and the integrase footprint overlaps the KMT2A IBM2 site, with integrase Trp131 sitting where KMT2A Phe148 goes 4.
- One proof of principle for displacement. Integrase, as a protein, disrupts the IBD-menin-MLL interaction with IC50 301 nM in AlphaLISA 4. That establishes the site is competable. It does not establish that a small molecule can do it, and nobody has reported one that does.
What is not transferable is the twenty years of medicinal chemistry, because all of it was optimised against a viral protein that does not exist in a leukaemia cell.
Why the confusion is systematic rather than careless
The naming is the trap. 'LEDGIN' is a contraction of LEDGF-IN, meaning an inhibitor of the LEDGF-integrase interaction, and the literature that introduced the term titles papers as 'Inhibitors of the Interaction Between HIV-1 Integrase and LEDGF/p75' 2 - accurate, and easy to compress into 'LEDGF inhibitors' by a reader who has not looked at the co-crystal. This graph should treat any claim of the form 'chemical matter already exists for LEDGF' as requiring a structure or a binding measurement against LEDGF itself, not against the interaction.
| correction | LEDGINs/ALLINIs/NCINIs bind HIV-1 integrase, not LEDGF |
|---|---|
| decisive evidence | LEDGINs retain activity in a PSIP1 IBD-exon knockout cell line |
transferable from hiv
[ "AlphaScreen/AlphaLISA assay format for the IBD interaction", "PDB 2B4J structural map of the IBD surface", "proof that the IBD can be competed off menin-MLL (IC50 301 nM, by integrase protein)", "the demonstration that HRP-2 substitutes for the LEDGF IBD" ]
not transferable
[ "any LEDGIN/ALLINI compound as a LEDGF-directed tool", "the medicinal chemistry optimised against integrase" ]
- Repurposing a LEDGIN into leukaemia would waste a programme on a compound whose target is absent from the cell, and the naming of the class actively invites that mistake.
- The cross-field precedent that does transfer is the HRP-2 result, and it is unfavourable: the paralogue substitutes for the LEDGF IBD, so an LEDGF-selective agent may be buffered.
- The only reported direct LEDGF binder is a promiscuous thiol-reactive selenium compound, which is chemical matter in the narrowest sense and would fail any selectivity counter-screen.
- Has any LEDGIN ever been counter-screened against a leukaemia line for an integrase-independent effect? It should be null, and it is the kind of null nobody publishes, so the absence here is unbounded.
- The AlphaLISA competition shows the IBD-menin-MLL complex can be broken by a protein. What is the smallest entity that does it? CP65 is a cyclic peptide and the gap between a cyclic peptide and a small molecule at this interface has never been bridged for any IBD partner.
- Ebselen binds LEDGF through reactive cysteines. Which cysteines, and are any of them in or near the IBD? If the answer is yes, a covalent-fragment screen against the IBD is an obvious and cheap next step.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In a human somatic LEDGF/p75 knockout cell line generated by deleting the exons encoding the IBD, HIV-1 replication was severely delayed or defective, the residual replication was mediated by HRP-2, and LEDGINs remained active in the absence of LEDGF/p75 by blocking the interaction with the IBD of HRP-2.
LEDGF/p75-independent HIV-1 replication demonstrates a role for HRP-2 and remains sensitive to inhibition by LEDGINs. · Schrijvers R, De Rijck J, et al. · PLoS Pathog 2012What this evidence cannot show
A virology experiment. It establishes what LEDGINs bind and that HRP-2 can substitute for LEDGF in HIV integration; it says nothing directly about leukaemia. Note that a correction notice for this paper exists in PubMed (PMID 32870932), which was not retrieved, so the primary record was used and the correction's content is unknown here.
Evidence that argues against this (1)
- Residual activity in a knockout could in principle reflect incomplete knockout rather than a different target; the authors address this by showing the activity runs through HRP-2's IBD specifically, but a reader should note the inference is indirect. Schrijvers R et al. 2012
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
LEDGINs are described by their originating laboratory as small molecule inhibitors of the interaction between the integrase binding domain of LEDGF/p75 and HIV integrase, proposed as block-and-lock agents that retarget HIV integration.
LEDGINs, Inhibitors of the Interaction Between HIV-1 Integrase and LEDGF/p75, Are Potent Antivirals with a Potential to Cure HIV Infection. · Debyser Z, Bruggemans A, et al. · Adv Exp Med Biol 2021What this evidence cannot show
A review by the group that named the class. Cited for the definition and self-description of the class, not for efficacy claims.
Evidence that argues against this (1)
- The block-and-lock mechanism attributed to this class is itself contested in scope: the maturation defect from integrase multimerisation may dominate over integration-site retargeting in determining antiviral potency. Singer MR et al. 2023
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The integrase-LEDGF/p75 interaction has been screened for inhibitors in AlphaScreen luminescent proximity format since 2008, giving the leukaemia interface a ready-made high-throughput assay template.
Screening for antiviral inhibitors of the HIV integrase-LEDGF/p75 interaction using the AlphaScreen luminescent proximity assay. · Hou Y, McGuinness DE, et al. · J Biomol Screen 2008What this evidence cannot show
An assay-development paper. Its transferability to a KMT2A peptide-IBD pair is an inference, though the same format has since been used for exactly that pair in AlphaLISA.
Evidence that argues against this (1)
- AlphaScreen is notoriously prone to singlet-oxygen quenchers and coloured-compound artefacts, and a hit rate against a micromolar protein-protein interaction is dominated by false positives, so the assay's availability is not the bottleneck it is sometimes presented as. Hou Y et al. 2008
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
HIV-1 integrase and the KMT2A IBM2 motif use overlapping surfaces on the LEDGF IBD, with integrase Trp131 occupying the position of KMT2A Phe148, and integrase disrupts the IBD-menin-MLL interaction with IC50 301 nM in AlphaLISA.
The same site on the integrase-binding domain of lens epithelium-derived growth factor is a therapeutic target for MLL leukemia and HIV. · Murai MJ, Pollock J, et al. · Blood 2014What this evidence cannot show
The competitor is a protein at high local affinity, not a drug-like small molecule, so the IC50 is a proof that the site is competable rather than a chemical starting point.
Evidence that argues against this (1)
- Because integrase binds the IBD roughly two orders of magnitude more tightly than the menin-MLL complex does, the displacement says more about integrase's affinity than about the vulnerability of the leukaemic interface. Murai MJ et al. 2014
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Co-crystal structures of HIV-1 integrase with the ALLINIs BI-D and pirmitegravir show the compounds bound in a pocket on the integrase catalytic core domain dimer, acting as molecular glue that nucleates an interface with the integrase C-terminal domain through residues Tyr226, Trp235 and Lys266 - an integrase-intrinsic mechanism with no LEDGF component.
The Drug-Induced Interface That Drives HIV-1 Integrase Hypermultimerization and Loss of Function. · Singer MR, Dinh T, et al. · mBio 2023What this evidence cannot show
Structural virology. Cited to establish where the compounds bind and that their principal mechanism is integrase multimerisation, which is the specific point that rules out repurposing.
Evidence that argues against this (1)
- The same pocket is the LEDGF-binding site, so these compounds do block the IN-LEDGF interaction as well; the claim here is about which protein they bind, not about whether they affect the interaction. Debyser Z et al. 2021
corrects what the HIV chemistry on this interface actually targets
no approved or preclinical agent binds this protein
the HIV precedent does not supply a tool compound for this target
the single exception that proves the class distinction
the knockout experiment that proves both points at once