"Star sign Capricor: when astrology does drug development"

"Executive Summary

Capricor Therapeutics (NASDAQ:CAPR) is expected to announce the results of a phase III clinical trial of its only asset, Deramiocel (CAP-1002), in Duchenne Muscular Dystrophy in Q42025.

In this work, we present an air-tight case challenging the safety and efficacy of Deramiocel as a treatment for Duchenne Muscular Dystrophy.

Capricor has changed the stated mechanism of action multiple times throughout the development of deramiocel, and they have failed to characterize the cell product leading up to this pivotal readout. We do not find the assembled data compelling or rooted in the scientific process.

The repeated failure of deramiocel in cardiac indications clearly show that it has little effect on human tissue, and the purported HOPE-2 efficacy signal was spurious and likely confounded by concomitant medication. Beyond these cardiac readouts we have found clear baseline imbalances in the functional readouts for DMD which confound the stated pre-specific primary endpoints in HOPE-2 and HOPE-3.

We also believe the safety signals of this product are not discussed enough publicly. We have found an alarmingly high rate of anaphylaxis and severe immune mediated reactions as a result of taking this allogeneic cell product. Worse still, are Capricor’s attempts to hide this data from their published works.

They have yet to demonstrate clear efficacy signals and the FDA issued a CRL in July of 2025, requiring Capricor to finally unblind themselves to the HOPE-3 topline data which they have had available to them for over a year, and which they have continuously punted since they first said they would publish it in Q42024.

Capricor has been (unsuccessfully) developing Deramiocel for the past 15 years across 8 indications, and Duchenne Muscular Dystrophy may be the last. Upon the announcement of Phase III failure, the stock price should convert to the company’s cash value or below, representing a >60% decline from its current price of ~$5.80 per share (landing near $2)."



AI summary on the report:

• Setup & timing

  • Capricor Therapeutics (CAPR) is awaiting Phase III HOPE-3 data for deramiocel (CAP-1002) in Duchenne muscular dystrophy (DMD).

  • Author expects HOPE-3 to fail, and CAPR equity to re-rate toward cash (~$1.8–2/share vs. ~$5.8 today).

Biology & mechanism

• Drug: Deramiocel = allogeneic cardiosphere-derived cells (CDCs), originally pursued for cardiac regeneration.

• Original “cardiac stem cell” story rests partly on now-retracted / discredited work from the Anversa group; field now views the “resident cardiac stem cell” concept very skeptically.

• Capricor/Marbán have shifted the narrative to a paracrine / exosome mechanism:

  • Inject CDCs → CDCs secrete exosomes → exosomes carry mixed RNA/protein payload → supposedly anti-fibrotic, anti-inflammatory, immunomodulatory.

• The report argues this is biologically implausible:

  • Only ~1% of injected CDCs reach the heart in animal models; most are trapped in lungs or cleared.

  • Exosomes have minute-scale half-lives in blood; many carry no payload; cargo is heterogeneous.

  • Independent groups have failed to reproduce meaningful benefit of CDCs or their exosomes in cardiac models.

PK / PD “Russian-doll” problem

• Authors build a back-of-envelope PK model for IV CDC dosing used in HOPE-2/3:

  • Assumptions heavily favour CAPR (1% of cells to heart, high exosome secretion, 10% uptake, etc.).

• Even under generous assumptions:

  • Only ≈ 0.24% of the estimated “effective” payload is delivered to target cardiac cells.

  • More realistic assumptions (lower cargo per EV, 1% uptake) make effective dose orders of magnitude too low.

• Implication: at current dosing, very unlikely that deramiocel delivers enough active cargo to meaningfully affect heart or systemic skeletal muscle.

Preclinical DMD data

• Key DMD work used female mdx mice, a mild disease model that overestimates efficacy vs. severe human DMD.

• More stringent models (mdx/utrn-/-; aged animals) show that cell therapies can fail or even worsen disease in advanced settings.

• Positive CDC data are largely confined to Marbán-authored studies; independent labs in Spain/Brazil found no benefit of CDCs or their exosomes in cardiac failure / MI models.

Clinical track record (non-DMD)

• ALLSTAR (post-MI) and ALPHA (pulmonary hypertension):

  • Randomized, double-blind, placebo-controlled.

  • No significant benefit on scar size, LVEF, hemodynamics, or functional endpoints.

• These failures undermine the idea that CDCs have robust cardiac efficacy, even with intracoronary delivery.

DMD clinical program

HOPE (Phase I/II, intracoronary CDCs)

• One-time 75M CDCs directly into coronary arteries; various cardiac and functional endpoints.

• No clinical benefit demonstrated.

• If direct coronary delivery fails, IV delivery is even less likely to work.

HOPE-2 (Phase II, IV CDCs, n=20)

Design & topline:

• Randomized, double-blind, placebo-controlled; 150M CDCs IV every 3 months x4.

• Originally powered for 80 patients, but enrollment stopped at 20 after an interim futility analysis + cash constraints.

• Pre-specified primary endpoint: change in mid-level PUL 1.2 at 12 months.

• Initial readout: primary endpoint p=0.08 (failed); PUL2.0 secondary p=0.05.

Key criticisms:

1. Post-hoc stats rescue

   • Later, Capricor switches to a percentile-rank (non-parametric) MMRM citing non-normal residuals.

   • With this new method, mid-PUL1.2 becomes p=0.014 and is re-labeled the “primary” endpoint; PUL2.0 ~p=0.04.

   • Author views this as an after-the-fact salvage, not a pre-specified strategy.

2. Baseline imbalance & floor effect

   • CAP-1002 arm has more advanced patients (lower PUL entry scores) → less room to decline on mid-PUL1.2.

   • Placebo arm includes more patients at stages where elbow loss drives big point drops (elbow = 34 of 74 PUL1.2 points).

   • PUL2.0, which handles later-stage function better, shows weaker drug effects, supporting the idea that baseline imbalance, not drug, drives apparent signal.

3. Internal inconsistency in reported numbers

   • Using published spaghetti plots and integer constraints, author reconstructs 12-month scores.

   • They can match either the means or the SDs, but not both simultaneously, implying at least one set of published values is wrong or inconsistent with the plots.

4. Cardiac endpoints: over-tested & confounded

   • SAP was amended to add ~50 secondary/exploratory endpoints (incl. 26 post-hoc MRI metrics) without multiplicity control; FDA later flags this as inflated Type I error.

   • Crucially, baseline concomitant cardioprotective meds (ACEi, ARBs, MRAs, β-blockers) are not reported for HOPE-2, though collected.

   • These drugs are well known to slow LVEF decline in DMD; any imbalance could explain small LVEF differences.

5. Underpowered + noisy imaging

   • n=20 vs planned 80; MRI LVEF variation ~2–3% between readers.

   • With typical annual treatment effects 1–3%, results are fragile and easily distorted.

Net view: HOPE-2 data are not robust, heavily confounded, and not regulatory-grade.

Safety & immunogenicity

• CAP-1002 is an allogeneic product → immune reactions are expected.

• In HOPE-2 and HOPE-OLE:

  • Around 30–37% of CAP-1002 patients suffered Grade 3–4 immune / allergic reactions, including life-threatening anaphylaxis with hospitalizations.

  • Some reactions appeared on later infusions despite premedication, indicating de novo anti-product antibodies.

• Protocol amendments:

  • Original language explicitly used “anaphylactic” and “severe”; later amendments soften or remove these terms and delete detail on first-infusion symptoms later followed by a severe reaction.

  • Prophylactic meds added to “minimize risk of severe allergic reactions.”

• Author sees this pattern as downplaying serious risk and ethically troubling, given:

  • Modest/unproven efficacy.

  • Availability of guideline-backed cardioprotective drugs with much better risk–benefit.

Regulatory & CMC

• CRL (July 2025):

  • FDA: HOPE-2 failed its pre-specified primary endpoint and secondaries; post-hoc functional and cardiac analyses + non-randomized OLE are insufficient to establish efficacy.

  • Notes extensive exploratory endpoint fishing without multiplicity adjustment.

  • Raises (partly redacted) manufacturing / CMC issues.

• Form 483 findings (May 2025) summarized:

  • Inadequate process & facility qualification under realistic conditions.

  • Repeated deviations (~27) involving missing instrument/data entries over 2 years.

  • Non-conforming product reports and CAPAs left unresolved beyond SOP timelines.

  • Documentation issues vs ALCOA+; cleanroom not maintained in full “state of good repair”.

• Author believes significant CQA, batch-release and potency-assay work is still needed; current CMC package is not commercial-ready.

Data integrity risk (investigator WL)

• HOPE-3 investigator Han Phan, MD previously received an FDA Warning Letter (a serious step beyond a Form 483) for another trial, involving:

  • Incorrect administration and scoring of primary efficacy assessments.

  • Documentation so poor it was “impossible to retrospectively account for all instances of noncompliance”.

  • FDA questioning validity and integrity of data from that site.

• Author does not hinge the thesis on fraud, but expects heightened FDA scrutiny of HOPE-3 data coming from this site.

HOPE-3 & investment conclusion

• HOPE-3 design:

  • ~102 patients, 1:1 randomization, IV 150M CDCs q3m as in HOPE-2.

  • Primary endpoint: Total PUL2.0 at 12 months.

  • Key secondary: LVEF by cMRI at 12 months, plus multiple additional secondaries.

• Why the author expects failure:

  • PUL2.0 is less biased toward elbow, more robust to floor effects; larger n should balance baselines.

  • With appropriate multiplicity control, scattered nominal cardiac/pulmonary endpoints are unlikely to survive.

  • Same PK/biology problems and immunogenicity risks persist.

  • They expect similar anaphylaxis/Grade 3–4 AE rates (~30%), further harming benefit–risk.

• Valuation frame:

  • If HOPE-3 fails and deramiocel is effectively worthless, company should trade at (or below) net cash.

  • Q3’25: $98.5m cash; 45.7m shares → $2.18/share; adjusting for partial Q4 burn → ~$1.78/share.

  • From ~$5.8 today, implied downside >60%.
    
    
    Source:
    https://anthonystaj.substack.com/p/star-sign-capricor-when-astrology