The Short-Lived Guardian

Programmed cell death, or apoptosis, is an evolutionarily conserved physiological process essential for embryonic morphogenesis, tissue homeostasis, and the elimination of damaged or premalignant cells. The intrinsic (mitochondrial) pathway of apoptosis is governed by the BCL-2 protein family and centers on the integrity of the mitochondrial outer membrane. At steady state, the pro-apoptotic executioner proteins BAX and BAK exist as inactive, monomeric conformers. Upon exposure to cytotoxic stimuli or severe DNA damage, conformational activation causes BAX and BAK to homo-oligomerize, creating toroidal lipid-protein pores that drive mitochondrial outer membrane permeabilization (MOMP). MOMP allows cytochrome c release from the intermembrane space into the cytosol, where it binds APAF-1 in the presence of dATP to assemble the heptameric apoptosome. This platform recruits and activates initiator procaspase-9, subsequently cleaving and activating downstream executioner caspases (caspase-3 and caspase-7) to orchestrate proteolytic cellular disassembly. That non-transformed cells evade this lethal cascade is due to continuous sequestration of pro-apoptotic signals by anti-apoptotic BCL-2 family members.

A Brief Primer on the Apoptosis Machinery

BAX and BAK are the executioners. Once activated, they punch the hole in the mitochondrial membrane that no cell survives.

MOMP (mitochondrial outer membrane permeabilization) is the point of no return, the moment BAX/BAK rupture the membrane and commit the cell to death.

BH3-only proteins — including BIM, PUMA, NOXA, and BID — are the messengers. They're produced in response to specific stress (DNA damage, growth factor withdrawal, ER stress) and their job is to activate BAX/BAK, directly or indirectly.

The BCL2 family of guardians — chiefly BCL2, MCL1, and BCL-XL — are the brakes. Each one works by physically gripping BH3-only messengers in a shallow surface groove, muffling the death signal before it reaches BAX and BAK.

The Guardian on a Deadline: MCL1 vs. BCL2

BCL2 itself carries historical weight: it was the first oncogene ever cloned from a human lymphoma, identified through the chromosomal translocation t(14;18). But BCL2 isn't the only guardian, and in many fast-dividing tissues it isn't even the most important one. That role often belongs to a relative called MCL1 (myeloid cell leukemia 1) — a protein built on a fundamentally different timescale.

BCL2 is comparatively stable, often persisting in a resting cell for close to a full day. MCL1 is not built to last. Its functional half-life has been measured at roughly thirty minutes. A cell expressing MCL1 isn't storing a shield, it's continuously rebuilding one. Stop the construction, even briefly, and the dam it maintains starts to fail almost immediately.

When the Dam Is Reinforced Too Much

Tumors exploit this system directly. The gene for MCL1 sits at 1q21 on chromosome 1, a region amplified across a striking range of malignancies: acute myeloid leukemia, multiple myeloma, lung adenocarcinoma, hepatocellular carcinoma. The amplification is often not subtle, with some tumors carrying dozens of extra copies. The result floods the cell with guardian protein, turning a dynamic, actively-rebuilt dam into something closer to a permanent wall. Chemotherapy and radiation still generate their usual BH3-only distress signals. The signals just get absorbed without consequence.

Normal Apoptotic Restraint:

Guardian Exhausted (appropriate death):

Cancer-Reinforced State (1q21 amplification):

The Geometric Problem, Again

For years, pharmaceutical chemists considered MCL1 essentially undruggable, for reasons that will sound familiar from other famously "impossible" targets. Its BH3-binding groove is shallow and hydrophobic, with none of the deep catalytic cleft that makes a kinase such an easy target. Worse, the protein doesn't sit still. It shifts constantly between conformations in solution, so a molecule engineered to fit one structural snapshot can miss the target entirely a moment later. High-throughput screening campaigns repeatedly failed. The field largely gave up on direct inhibition and chased indirect strategies instead, trying to choke off MCL1's transcription or accelerate its natural degradation, without ever confronting the guardian head-on.

The Turn: Wedging Fragments into a Moving Target

High-resolution crystallography eventually revealed that despite its shallow appearance, MCL1's groove has distinct sub-pockets. A rigid naphthol group could wedge into one corner. An indole structure fit an adjacent space. Linking these fragments together produced the first genuinely selective MCL1 inhibitors — compounds like AMG 176 and AZD5991 — engineered specifically for MCL1's architecture rather than repurposed from BCL2 chemistry. These molecules displaced BH3-only messengers from the groove with sub-nanomolar affinity and, in preclinical models, killed MCL1-dependent cancer cell lines within hours. A target the field had written off was, for the first time, directly and selectively drugged.

Translational Limiting Factor: On-Target Myocardial Toxicity

Translational validation revealed severe on-target, off-tumor safety liabilities. MCL-1 is not merely an anti-apoptotic oncogene; it is a pleiotropic homeostatic factor required for the maintenance of hematopoietic stem cells, mature neutrophils, and adult cardiac tissue. Cardiomyocytes are post-mitotic, non-dividing cells characterized by extraordinarily high mitochondrial volume density (~40% of cardiomyocyte cellular volume) to satisfy the continuous energetic demands of excitation-contraction coupling. In the myocardium, MCL-1 localizes not only to the outer mitochondrial membrane to neutralize BH3-only pro-apoptotic proteins, but also resides within the inner mitochondrial membrane (IMM) matrix space, where an amino-terminally truncated isoform stabilizes OPA1 and maintains mitochondrial cristae ultrastructure and ATP synthesis. Targeted genetic disruption of Mcl1 in adult murine cardiomyocytes triggers rapid disruption of mitochondrial respiration, acute sarcomere disorganization, troponin release, and fatal dilated cardiomyopathy within days.

That finding translated directly into the clinic. Amgen's oral MCL1 inhibitor, later named murizatoclax (AMG 397), was placed on FDA clinical hold in 2019 after a cardiac toxicity signal emerged. Its intravenous counterpart, AMG 176, showed troponin elevations at higher doses, though not at the lowest tested doses. AstraZeneca's AZD5991 fared worse: a retrospective analysis of troponin elevations and cardiovascular risk factors led to the trial being terminated early, and its combination cohort with venetoclax recorded a death from cardiac arrest among its adverse events. AbbVie's ABBV-467 was efficacious in tumor models but likewise triggered cardiac troponin increases in patients. Across four independently developed molecules from four different companies, the same organ kept objecting.

Racing the Rebuild: Intermittent Dosing

MCL1's defining biological feature, its rapid turnover, is also what makes it so hard to inhibit safely, and the field's current strategy tries to use that same feature as an escape route. Rather than maintaining constant blockade, some programs have shifted toward a sharp intravenous pulse of drug followed by a longer recovery window, intended to knock out MCL1-dependent tumor cells during the pulse while giving slower-cycling, less MCL1-hungry cardiac tissue time to recover between doses. Preclinical work on ABBV-467 specifically supported this logic, suggesting that a short-half-life molecule given intermittently could preserve antitumor activity while giving investigators more room to control for cardiac safety. Whether that margin holds up at scale, across a full patient population and a full treatment course, remains the open clinical question.

Combination Warfare: Venetoclax and the MCL1 Escape Hatch

The BCL2 inhibitor venetoclax transformed treatment for chronic lymphocytic leukemia and acute myeloid leukemia, but resistance predictably follows the same logic as everywhere else in this story: cancer cells under BCL2 blockade frequently respond by upregulating MCL1, shifting their dependence from one guardian to the other. The obvious next move is to hit both simultaneously, and several trials have paired venetoclax with an MCL1 inhibitor to try to do exactly that. The approach runs directly into the same wall as monotherapy, only doubled: bone marrow, gut epithelium, and cardiac tissue all depend on some baseline level of BCL2-family protection, and stripping away two guardians at once narrows the safety margin further rather than resolving it.

Therapeutic Window: Pharmacokinetic Pulsing vs. Cardiomyocyte Mitochondrial Reserves

The short intracellular half-life of MCL-1 presents a unique pharmacokinetic and pharmacodynamic challenge. Rapid turnover is not a structural vulnerability, but an evolved signaling mechanism allowing instantaneous cellular responses to trophic factor deprivation. When survival signaling halts, ongoing MCL-1 translation ceases, proteasomal degradation mediated by the E3 ligase MULE clears remaining pools within minutes, and unrestrained BH3-only proteins initiate MOMP. Small-molecule BH3-mimetics chemically replicate this physiological collapse.

However, developing clinically viable MCL-1 inhibitors requires establishing a workable therapeutic index between oncogenic dependencies and baseline organ physiology. While high-affinity chemical scaffolds like AMG 176, AZD5991, and ABBV-467 demonstrated potent tumor regression, their clinical translation was constrained by on-target myocardial injury. Current translational approaches focus on pulsed, intermittent dosing regimens designed to achieve transient cytotoxic target coverage in malignant cells while preserving the continuous mitochondrial energetics required for cardiomyocyte survival.