Command Palette

Search for a command to run...

The biomarker moved and the bones did not: INZ-701 in ENPP1 deficiency

OtherBiomarkerSeptember 25th, 2026·6 min read·10.5281/zenodo.20479005

An enzyme replacement built to restore inorganic pyrophosphate did exactly that through 52 weeks, then missed the radiographic rickets endpoint beside it. BioMarin discontinued the asset and the infant Phase 3 stopped at two participants.

Mechanism Risk Score

ComponentPoints
Phase-weighted failure burden11.8 / 30
Archetype severity8.5 / 25
Temporal recency4.3 / 15
Genetic evidence deficit2.3 / 15
Programmatic saturation2.7 / 15

For ENPP1 in ENPP1 deficiency presenting as generalized arterial calcification of infancy, the Mechanism Risk Score is 30/100 (yellow band). The score is a failure-burden index derived from Claidex post-mortems on this target–disease pair, not a probability of approval.

MRS 30/100 (YELLOW). One ENPP1 programme is on file, INZ-701 (BMN 401), a recombinant ENPP1-Fc enzyme replacement discontinued across all indications in August 2026 after the Phase 3 ENERGY 3 trial met its plasma pyrophosphate co-primary endpoint and missed the Radiographic Global Impression of Change co-primary. The Phase 3 ENERGY 2 infant study (NCT07473973) was terminated at 2 participants. The genetic deficit term contributes only 2.30 of a possible 15 because Open Targets scores ENPP1 against generalized arterial calcification of infancy type 1 at 0.8464 with genetic_association 0.967. Saturation contributes 2.72 because Open Targets records no other clinical programme against ENPP1. Almost the whole score comes from the failure occurring at Phase 3, and it reflects surrogate-endpoint risk rather than any doubt about the target biology.

This score does not predict whether the next trial will succeed. It flags how heavy the documented mechanistic failure record is before a new program is justified.

Primary figure supporting this claim (INZ-701 (BMN 401, rhENPP1-Fc) / ENPP1 / ENPP1 deficiency presenting as generalized arterial calcification of infancy): The biomarker moved and the bones did not: INZ-701 in ENPP1 deficiency

What was tried

ENERGY 2 (NCT07473973, protocol INZ701-105) was a multicentre, single-arm, open-label Phase 3 study of INZ-701 in infants aged one year or younger with genetically confirmed ENPP1 deficiency and clinical features of generalized arterial calcification of infancy. Inozyme Pharma was the sponsor and BioMarin Pharmaceutical the collaborator. INZ-701 is a recombinant fusion protein joining the extracellular domain of human ENPP1 to an IgG1 Fc fragment, given by weekly subcutaneous injection.

The study carried two primary outcomes at 52 weeks: change from baseline in plasma inorganic pyrophosphate, and overall survival measured from date of birth. Secondary outcomes covered left ventricular ejection fraction, incidence of heart failure, and coronary and aortic calcification on CT.

ENERGY 2 started on 26 March 2025. Primary completion and completion are both recorded as 31 August 2026, with 2 infants enrolled. The record was updated on 24 September 2026 with status TERMINATED and this reason: following review of the totality of program data, the sponsor decided to discontinue development of INZ701, not for safety.

The biological hypothesis

ENPP1 (ENSG00000197594) hydrolyses extracellular ATP to AMP and inorganic pyrophosphate. Pyrophosphate is the main endogenous inhibitor of hydroxyapatite crystal growth, so biallelic loss-of-function variants produce a pyrophosphate-deficient state in which mineral deposits where it should not. Infants present with generalized arterial calcification of infancy, and children who survive the vascular phase develop autosomal recessive hypophosphatemic rickets type 2 through elevated FGF23 and renal phosphate wasting. Ferreira and Carpenter set out this two-stage natural history in 2024, and Boyce and Gafni reviewed the management problem in 2020.

The genetic case is about as strong as target biology gets. Open Targets scores ENPP1 against generalized arterial calcification of infancy type 1 (MONDO_0008817) at 0.8464, with a genetic association datatype score of 0.967, genetic literature at 0.826 and animal model evidence at 0.489. Open Targets lists no approved or clinical-stage drug against ENPP1 and no safety liabilities for the target.

Preclinical work supported enzyme replacement directly. Nitschke and colleagues showed in 2018 that ENPP1-Fc prevented neointima formation in a model of the infantile disease, and Cheng and colleagues reported in 2021 that INZ-701 prevented ectopic tissue calcification and restored bone architecture and growth in ENPP1-deficient mice. The logic was clean. Restore circulating enzyme, restore pyrophosphate, prevent mineral deposition where it does not belong and correct the skeletal consequences.

What actually happened

The pyrophosphate part worked. ENERGY 3 (NCT06046820), a randomised 2:1 open-label Phase 3 in 27 children aged 1 to 12, met its first co-primary endpoint with statistically significant increases in plasma inorganic pyrophosphate from baseline through week 52. It missed the second co-primary, the Radiographic Global Impression of Change score at week 52, showing no improvement in rickets severity. BioMarin reported on 18 May 2026 that no positive trends appeared across the Rickets Severity Score or growth Z-scores for height, body length and weight, and that the drug was generally well tolerated with no new safety signals.

With its second quarter 2026 results in August 2026, BioMarin discontinued BMN 401, the company designation for INZ-701, across all indications, including ABCC6 deficiency and calcific uraemic arteriolopathy. The programme had arrived through the purchase of Inozyme Pharma for 270 million dollars, completed in May 2025.

Eight INZ-701 trials are registered, seven of which report actual enrolment totalling 81 participants. ENERGY 2 stopped at 2, and the ENABLE study (NCT06739980) was withdrawn without enrolling anyone. No openFDA FAERS record names INZ-701, expected for an agent never marketed.

Failure mechanism, best guess

The archetype is biomarker failure. A surrogate the drug was built to move did move, and the clinical endpoint attached to it did not follow.

Three explanations fit the public data and none can be excluded. The first is reversibility. The Radiographic Global Impression of Change asks whether existing radiographic disease improved, and it was measured in children who had accrued skeletal disease over years. Restoring pyrophosphate plausibly prevents new mineral deposition without remodelling what is already there, in which case the endpoint tested reversal while the drug does prevention.

The second is that pyrophosphate is necessary but not sufficient for the skeletal phenotype. Rickets in ENPP1 deficiency runs substantially through FGF23-mediated phosphate wasting, so an intervention aimed at the calcification axis can leave the mineralisation axis partly uncorrected.

The third is measurement. A global radiographic impression scored once in 27 children, split roughly 18 to 9 by the allocation, has limited resolution. This is the weakest of the three, because an underpowered endpoint usually still leaves a direction, and the Rickets Severity Score and growth measures showed none.

ENERGY 2 adds a design observation of its own. Overall survival as a co-primary endpoint in an ultra-rare infant population was not answerable at the enrolment this study reached.

How to prevent this next time

No participant-level data have been posted, so a quantitative model of the surrogate-to-outcome relationship cannot be built from public sources. The available levers are qualitative.

Qualify the surrogate before a confirmatory trial depends on it. The path from ENPP1 to pyrophosphate to ectopic calcification is mechanistic and well evidenced. The path from a 52-week rise in plasma pyrophosphate to a radiographic rickets score was an assumption that entered a pivotal design without a human exposure-response dataset linking the two.

Match the endpoint to the reversibility window, and separate the two diseases. A prevention mechanism tested with a reversal endpoint reads as a failure even when the biology is right, and generalized arterial calcification of infancy and autosomal recessive hypophosphatemic rickets type 2 share a gene but not an endpoint.

Pre-specify what a positive biomarker with a negative clinical endpoint would mean. That outcome was foreseeable and arrived with no agreed interpretation, which is how a programme ends on totality-of-data language rather than on a rule.

The single highest leverage change would have been running the infant prevention cohort to a vascular calcification endpoint before committing a pivotal trial to a radiographic rickets score in older children.

What this means for similar programs

This is the first ENPP1 entry in the Claidex graph. The recomputed mechanism risk score is 30, in the yellow band, from phase burden 11.8, archetype severity 8.52, recency 4.25, genetic deficit 2.30 and saturation 2.72. The genetic and saturation terms are small because the association is strong and Open Targets records no other clinical programme against ENPP1. Almost all of the score comes from the single recorded failure happening at Phase 3.

The read-across is not that ENPP1 is a bad target. It is that enzyme replacement in mineralisation disorders should treat the radiographic endpoint, not the biochemical one, as the thing to de-risk first. Enzyme replacement in this axis has worked before, since asfotase alfa is an approved treatment for hypophosphatasia, a disease of the opposing enzyme alkaline phosphatase. Programmes in ABCC6 deficiency and pseudoxanthoma elasticum inherit this result directly, since they were part of the same discontinued asset and rest on the same pyrophosphate restoration logic.

Open questions

How large was the pyrophosphate increase in ENERGY 3, and did participants reaching higher exposures show any movement on the Radiographic Global Impression of Change?

Did vascular calcification, the endpoint closest to the mechanism, change in any INZ-701 study? ENERGY 2 listed CT-based coronary and aortic calcification and posted no results.

Does the rickets phenotype respond to pyrophosphate restoration combined with FGF23 blockade, which no registered trial has tested?

Sources

  2. - Jacobs IJ, et al. INZ-701, a recombinant ENPP1 enzyme, prevents ectopic calcification in an Abcc6(-/-) mouse model of pseudoxanthoma elasticum. Exp Dermatol.

  3. - Nitschke Y, et al. ENPP1-Fc prevents neointima formation in generalized arterial calcification of infancy through the generation of AMP. Exp Mol Med.

  4. - Ferreira CR, Carpenter TO. ENPP1 in Blood and Bone: Skeletal and Soft Tissue Diseases Induced by ENPP1 Deficiency. Annu Rev Pathol.

  5. - Boyce AM, Gafni RI. Generalized Arterial Calcification of Infancy: New Insights, Controversies, and Approach to Management. Curr Osteoporos Rep.

  6. - Höppner J, Kornak U. Autosomal recessive hypophosphatemic rickets type 2 (ARHR2) due to ENPP1-deficiency. Bone.

  7. - Ralph D, van de Wetering K. Inorganic Pyrophosphate Deficiency Syndromes and Potential Treatments for Pathologic Tissue Calcification. Am J Pathol.

  8. - Open Targets Platform, ENPP1 (ENSG00000197594) association with generalized arterial calcification of infancy type 1 (MONDO_0008817), score 0.8464, retrieved 25 September 2026. https://platform.opentargets.org/target/ENSG00000197594 - ChEMBL, CHEMBL4650401 (INZ-701), maximum phase 3, retrieved 25 September 2026. https://www.ebi.ac.uk/chembl/compound_report_card/CHEMBL4650401/ - openFDA drug adverse event endpoint, queries for INZ-701 and INZ701 returned no matching records, 25 September 2026. https://open.fda.gov/apis/drug/event/.

Related failure claims

Linked claims sharing target, indication, or failure mechanism.