Detailed evidence-based analysis (claims & support)

1) Core claim: conserved executioners, variable upstream order

Park et al. correctly emphasize that caspases (ced-3 / ICE family) and Bcl‑2 family members are conserved effectors across taxa and paradigms; they marshal genetic (C. elegans ced genes), biochemical and inhibitor studies to support that caspase activation is the execution step and Bcl‑2 family acts upstream in many but not all contexts. This syntheses remains accurate: subsequent structural and biochemical work (apoptosome/Apaf-1 and cytochrome c) validated the mitochondrial–Apaf-1–caspase axis anticipated in the note added in proof Park et al., core conserved elements

2) Paradigm-level ordering — what they argued and evidence strength

• Fas receptor signaling: short receptor-to-caspase pathway via FADD and FLICE (caspase‑8 homologs), rapid and protein‑synthesis independent; evidence then included biochemical recruitment of FLICE to Fas/FADD and sequential ICE-like → CPP32-like activity in some cells (e.g., Enari et al.). The review highlighted uncertainties about how FLICE is activated after recruitment — a fair caveat given then-available data Fas pathway ordering (Park 1997)
• CTL/granzyme B: proposed as a 'rapid bypass' to effectors — granzyme B can directly cleave and activate multiple caspases; Bcl‑2 often fails to protect here, suggesting mitochondrial regulation is bypassed in many CTL killings. Evidence at the time included granzyme B activation of caspases in vitro and granzyme B-deficient CTL studies CTL/granzyme pathway (Park 1997)
• DNA damage / p53: p53 placed upstream as a checkpoint; mix of transcriptional and non-transcriptional roles was acknowledged; links to ATM, PARP, DNA‑PK and ceramide/JNK were proposed but described as incomplete. Strength: moderate — later work solidified p53 role yet also emphasized context dependence p53/genotoxic pathway (Park 1997)
• Trophic-factor (neuronal) deprivation: evidence for Bcl‑2/Bax regulation upstream of caspases, involvement of JNK and inappropriate cell-cycle re-entry; the review accurately lifted neuronal-specific ordering and suggested caspases act downstream of cell-cycle elements in some neuronal models NGF deprivation ordering (Park 1997)

3) Strengths of the review

• Timely, integrative framing across model organisms (C. elegans → mammals) and multiple death triggers, useful for hypothesis generation and experimental design (clear identification of open mechanistic gaps).
• Balanced: acknowledges conserved effectors while not over-generalizing upstream ordering — repeatedly flags context-dependence and unresolved causal links.
• Supports later mechanistic discoveries: the review's note cited Apaf-1/cytochrome c (Zou et al., 1997), which later matured into a robust apoptosome model (see later reviews) Apaf-1 note (Park 1997)

4) Limitations, blindspots and modern context

1. As a narrative review (1997) it depends on the published studies available then — selection bias and publication bias could skew weighting of pathways; authors acknowledge this Narrative review limitations (Park 1997).
2. Mechanistic resolution: some key molecular connections (how FLICE is activated at the DISC, precise mitochondrial pore composition, protease hierarchy) were unsettled in 1997; subsequent structural/biochemical work (apoptosome literature) resolved some but not all uncertainties. For a modern mechanistic treatment see later apoptosome syntheses (reviewed below) Apoptosome review (2007).
3. Quantitative ordering absent: the review is qualitative — it does not supply kinetic, stoichiometric, or systems-level models (ordinary differential or stochastic) that would enable predictive falsification. The field later moved in that direction (kinetic modeling of caspase cascades) but that is not present here.
4. Cell-type heterogeneity: experiments across cell lines, primary cells and organisms were pooled conceptually; differences in expression, isoforms, adaptor composition and compartmentation make 'ordering' non-transferable without targeted cross-system tests (an acknowledged caveat in the text).

5) How well have predictions held up? — short verdict

Major predictions held: caspases and Bcl‑2 family are central conserved players; mitochondria–cytochrome c–Apaf-1 involvement was correct and is now well-established; granzyme B can directly activate caspases in CTL-mediated killing — all supported by subsequent work. The key unknown (precise causal links and cell-type specific thresholds that determine which upstream signals engage caspases) remains an active research area, consistent with the authors' stated priorities Review foresight (Park 1997).

6) Recommendations for experiments that would have satisfied the paper’s open priorities

• Systematic genetic perturbation + timed caspase activity readouts across matched cell types (Fas-expressing vs Fas-resistant; primary neurons vs PC12) to define order and threshold relationships between adaptors (FADD/FLICE), Bcl‑2 family shifting, ceramide/JNK activation and caspase activation.
• Reconstitution biochemistry of the DISC and of mitochondrial release — purified components to test whether recruitment is sufficient for activation (particularly for FLICE/caspase-8) and whether Bcl‑2 family members directly block Apaf-1 oligomerization.
• Quantitative systems-biology models (kinetic ODEs) that include protease activation, feedback amplification and inhibitor strengths (e.g., IAPs) to produce falsifiable predictions about amplitude and timing of caspase activation in different paradigms.

7) Relationship to later literature (examples)

Apoptosome concept: later reviews consolidated the cytochrome c–Apaf-1–caspase‑9 platform as central to intrinsic apoptosis and gave structural/kinetic detail that the 1997 review could only foreshadow Apoptosome structural/functional review.

Caspase ordering reviews (1999): subsequent syntheses refined initiator vs effector caspase families and hierarchies, supporting the cascade and parallel execution views described in Park et al. Caspase hierarchy review (1999)

8) Critical blindspots, biases, and how they could mislead

• Narrative selection bias: the review’s examples reflect prominent studies (Fas, granzyme B, p53, NGF withdrawal) but rarer or negative results may be underrepresented.
• Cross-species extrapolation risk: heavy reliance on C. elegans genetics to infer mammalian ordering can obscure vertebrate-specific regulators (IAPs, mitochondrial complexity) — authors mention conservation but also context dependence.
• Therapeutic extrapolation risk: while the review mentions therapeutic aims, direct translation is nontrivial because of pathway redundancy and tissue specificity.

9) Practical takeaways for researchers today

1. Use the four paradigms as a conceptual checklist when designing apoptosis experiments (ask: which upstream modules are present/absent?).
2. Combine genetic (CRISPR), biochemical reconstitution and high-time-resolution caspase activity assays to determine ordering, not just end-point apoptosis.
3. Capture heterogeneity: include primary cells and matched cell lines and quantify expression of Bcl‑2 family, adaptors and IAPs to contextualize ordering claims.

10) What would disprove the core thesis?

Strong disproof would require (a) robust demonstrations that caspase-dependent execution is not conserved across multiple paradigms, or (b) evidence that a single upstream mechanism (e.g., one regulator) uniformly controls all four paradigms regardless of cell type. Neither has been shown; instead, literature since 1997 supports conserved caspase execution with diverse upstream triggers (so the review's central thesis remains well supported) Core thesis persistence.