PfSnf2L liver test models

sporozoite bottleneck !!

Cell lines

Phenotyping the parasite in HC-04.J7 where signal is highest, and then confirming in non-dividing PHH where host toxicity is interpretable. Second, none of these models relieves the sporozoite bottleneck — every one of them consumes salivary-gland sporozoites at broadly similar rates, so your insectary access, not the hepatocyte model, sets the scale of what you can do.

Immortalised human lines.

HepG2 is essentially non-permissive for P. falciparum — it is used for P. berghei. The exception is HC-04, a spontaneously immortalised normal human hepatocyte line and the only human line shown to support complete exo-erythrocytic development of both P. falciparum and P. vivax, with merozoites infectious for human erythrocytes; the original infection rates were 0.066% (Pf) and 0.041% (Pv) (DOI — Sattabongkot et al. 2006, Am J Trop Med Hyg 74:708). The subclone HC-04.J7 was selected for invasion susceptibility and reaches 3–5% invasion, three- to five-fold better than other hepatoma models, with glypican-3 implicated as an entry receptor (DOI). Practically, this is the cheapest possible entry: HC-04 comes from BEI/MR4, grows like any cell line, and the Rao organoid paper itself uses it for sporozoite batch QC in 384-well format. For dose–response curves and for phenotyping a conditional Snf2L line, it is where I would start.

Primary human hepatocytes, 2D.

Plain PHH monolayers work but give infection rates of roughly 0.01–0.75% with strong donor variability. Two engineered versions fix most of that. The MIT micropatterned co-culture (MPCC — PHH islands surrounded by 3T3-J2 fibroblasts) supports full Pf and Pv liver stages including merosome release and subsequent reticulocyte infection (DOI), and was later used to culture P. vivax hypnozoites, test anti-relapse drugs and generate the hypnozoite transcriptome (DOI). Worth knowing: infection efficiency in that platform improves markedly under hypoxia, via host HIF-1α, because standard incubator oxygen is far above sinusoidal pO₂ (DOI) — a cheap variable to optimise in any of these systems.

The other is the 384-well cryopreserved-PHH platform from the Adams/Kyle groups, built deliberately from commercially available materials so that other labs can adopt it. Hepatocyte physiology holds for at least 30 days, and it supports complete maturation of both Pf and Pv schizonts plus Pv hypnozoites, with immunofluorescence and drug readouts (DOI). For medium-throughput Pf liver-stage pharmacology on human hepatocytes, this is currently the most transferable option and the closest thing to a standard.

iPSC-derived hepatocyte-like cells.

These support Pb, Py, Pf and Pv, with permissiveness appearing already at the hepatoblast stage, and they metabolise antimalarials — though they remain functionally immature unless chemically matured (DOI). Their real advantage for you is genetic tractability on the host side: if you need to show that your inhibitor's effect is on the parasite and not on human SMARCA1/SMARCA5, an editable human hepatocyte background is exactly the tool for that.

Bioengineered formats.

Spheroids, bioprinted constructs, perfused microfluidic chips and zonated devices all exist and have been applied to malaria among other hepatotropic infections; the Khetani reviews are the practical entry points (DOI, DOI). None of them is more accessible than the two PHH platforms above, and most add complexity you do not need to answer an essentiality question.

In vivo human hepatocytes.

FRG-huHep or uPA-SCID liver-chimeric mice reconstituted with human hepatocytes remain the only system that gives you Pf liver stage in human hepatocytes plus the liver-to-blood transition — the endpoint that neither the organoids nor MPCC can deliver.

Sources (retrieved via PubMed): Sattabongkot 2006, Tweedell 2019, March 2013, Gural 2018, Ng 2013, Roth 2018, Ng 2015, Kukla & Khetani 2021, Yuan 2023

If it's useful, I can put all of this — the six platforms side by side with infection efficiency, throughput, what each supports, sporozoite cost per condition, and who to contact — into a one-page comparison you could drop into a grant or a collaboration email.

Organoid Model - difficult...

38770342 /24 to difficult

I read the full paper (Rao et al., Heliyon 2024; DOI, via PubMed). Short version: the platform is technically undemanding for an organoid lab, but it is a low-signal, high-variance assay, and it cannot see the endpoint you probably care most about.

How hard the assay itself is. The organoids are adult intrahepatic ductal (cholangiocyte) organoids, expanded in the standard Huch-type cocktail and differentiated to hepatocyte-like cells over six days. They can be biobanked, so you screen donors once and then work with the permissive lines — that is the model's real advantage. Infection is a three-hour co-incubation at MOI 4 (75,000 sporozoites per well of ~150 organoids ≈ 18,750 HLCs), a 600 g spin, then five to six washes and reseeding in BME drops on chamber slides. Readout is fixed-cell IFA with anti-PfHSP70 (all intrahepatic forms) and anti-PfMSP-1 (merozoite maturation), confocal tile scans, and their published ImageJ/StarDist macro. Nothing exotic.

The catch is the numbers. Infection rate was 26 ± 26 EEFs per 10,000 HLCs — 0.26%, with a standard deviation equal to the mean across eight donors. They needed ≥2,500 HLC nuclei imaged per condition just to pin the infection rate to a 95% CI with 0.1 margin, and actually imaged >3,900. EEFs are ~9 µm at day 6, maturation is scored as relative MSP-1 intensity, and the day 9 drop is probably egress. So every condition costs you ~75,000 sporozoites and a confocal tile-scan session, and the effect you need to detect must survive donor-to-donor noise of that magnitude.

Sporozoites are the actual bottleneck, not the organoids. They bought PfNF54 sporozoites from TropIQ and used them within five hours of salivary-gland dissection at room temperature. For a wild-type inhibitor experiment that is a solved problem — buy them, or use cryopreserved PfSPZ. For a transgenic Snf2L line it is not: you need an insectary willing to feed your parasites, and the dissected sporozoites must reach the organoids the same morning. Zurich (Marti/Venugopal, already on your Nature paper) or Nijmegen are the realistic options; a full experiment at four conditions × three timepoints × three donors × duplicates is ~70 wells ≈ 5 × 10⁶ sporozoites, i.e. hundreds of infected mosquitoes per repeat.

What testing PfSnf2L essentiality would technically require. First, confirm it is expressed in liver forms — the Kappe/Vaughan LS transcriptome from liver-chimeric mice will answer that in an afternoon and decides whether the experiment is worth doing at all.

Then a conditional line, because you cannot get a null through blood and mosquito stages. Given that this assay is a three-hour infection followed by a fixed endpoint, DiCre is the best fit: a snf2l-loxP line in a DiCre background, rapamycin pulsed onto salivary-gland sporozoites (or during the three-hour co-incubation), with the same line plus DMSO as the paired control on the same organoid batch — which neutralises the donor variability, since both arms come from one preparation. A ligand-controlled degron (FKBP-DD/Shield-1, knocksideways, AID) is the alternative and has the advantage that you can deplete at a chosen hour post-invasion and separate invasion from schizogony from maturation; the risk is ligand penetration into BME drops. TetR-DOZI is the wrong tool here, because the parasite spends the whole mosquito phase without aTc.

Around that you need: an HA/GFP tag for excision and protein-loss verification on sporozoites, excision PCR, and sporozoite fitness controls — salivary-gland yield, gliding, and the HC-04 traversal assay they use for batch QC — so that an EEF deficit is not just a sick-sporozoite artefact. Then EEF counts at day 3 (invasion and early development), EEF size at day 3/6, and MSP-1 maturation at day 6/9.

Two warnings from their own data. The paper's putative hit, K98, is an ARID1A/BAF inhibitor, and it reduced EEFs mainly by de-differentiating the hepatocytes — CYP3A4, albumin and NTCP all dropped. For a chromatin-remodeller inhibitor in a hepatocyte model that is exactly the trap you are walking into, and it applies doubly to a Snf2L inhibitor, since the human ISWI ATPases SMARCA5/SNF2H and SMARCA1/SNF2L are precisely the enzymes your lab works on. Any liver-stage result with your compound will need the host-marker qPCR panel, CellTiter-Glo 3D, and — which you can do better than anyone — in vitro remodelling assays against human SNF2H/ACF to show selectivity.

Second, the model cannot do liver-to-blood transition; the authors say so explicitly, blaming the ~5-minute invasive window of merozoites. So "essential in the liver" from organoids means intrahepatic development only. For causal prophylaxis you still need FRG-huHep mice with human RBCs, or P. berghei FLP/FRT in mice, which remains the cheap, high-power route and would let you do dose-ranging and merosome counts for a fraction of the cost.

My honest read: run the inhibitor in the organoid first — it needs no genetics, uses purchasable NF54 sporozoites, and is a few weeks of work once you have organoid lines — and reserve the conditional-genetics version for either P. berghei or a well-resourced collaboration with Zurich.

Sources (retrieved via PubMed): Rao et al. 2024, Zanghí et al. 2025, Lacroix et al. 2011

Want me to turn this into a one-page feasibility sheet with the reagent list, sporozoite budget and a decision tree? It would drop straight into a grant or a collaboration email.

How to close the gap

NH125 (or a more selective analog) on P. falciparum/P. berghei liver stages in primary human hepatocytes or HC-04, with a separate sporozoite-invasion arm, needs no new parasite line. Genetically, the inducible nature of the DiCre line is the opening: transmit unexcised parasites, then induce in sporozoites or infected hepatocytes and score EEF development — with excision efficiency verified in sporozoites/EEFs rather than assumed, and gametocyte/sporozoite yields from the parental line checked given the conversion defect.

Wording that current data support: essential for asexual blood-stage development and early gametocytogenesis (PMID 39972139), transcribed at high level in sporozoites and throughout liver-stage development (PMID 39891010), pre-erythrocytic requirement not yet tested.