CLEAR RAR - Reviewer Attention Report

Alternative paper-level hypotheses 2

Paper-level alternatives surfaced by HRAN and checked against the broader manuscript context. A reviewer would inspect whether the authors address these hypotheses across the paper.

Alternative paper-level hypothesis

β2m-deficient target-cell vulnerability in GVHD and tumor models

severity S partial

The phenotype attributed to enhanced direct susceptibility of β2m-deficient targets to CD4+ T cells may instead reflect a broader target-cell stress-hypersensitivity program with heightened IFNγ sensitivity and oxidative injury vulnerability, of which ferroptosis is one downstream manifestation rather than the singular defining mechanism.

Mechanism: broader inflammatory stress hypersensitivity beyond ferroptosis

Reason to inspect: This is the most consequential paper-level alternative. The paper's mechanistic narrative centers on ferroptosis as the key death pathway in β2m-deficient target cells exposed to CD4+ T cell-derived IFNγ. However, β2m deficiency has known consequences beyond MHC I loss: β2m associates with hemochromatosis protein HFE and is involved in iron homeostasis, and β2m-KO mice have documented iron overload in certain tissues. This creates a pre-existing vulnerability to oxidative stress that could sensitize cells to multiple death pathways, not just ferroptosis. The paper shows ferroptosis markers (lipid ROS, GPX4 changes) and DFX rescue, which are consistent with ferroptosis but also consistent with broader oxidative stress vulnerability where ferroptosis is one of several activated death programs. The paper partially addresses this by showing ferroptosis-specific markers, but does not appear to test whether blocking other death pathways (apoptosis, necroptosis, pyroptosis) also partially rescues the phenotype, which would indicate ferroptosis is one arm of a broader stress-death program rather than the singular mechanism. The partial status reflects that the paper has relevant ferroptosis data but has not excluded the broader stress-hypersensitivity interpretation.

Suggested experiment: Test inhibitors of apoptosis (zVAD-fmk), necroptosis (necrostatin-1), and pyroptosis (VX-765) alongside ferrostatin-1 in β2m-deficient IECs or organoids exposed to IFNγ or CD4+ T cells, to determine whether ferroptosis blockade alone fully rescues or whether multiple death pathways contribute. Additionally, test whether restoring HFE function in β2m-deficient cells normalizes iron handling and reduces ferroptosis sensitivity.

Relevant panels: Figure 2B Figure 3C Figure 3E Figure 3F Figure 3I Figure 4A Figure 4B Figure 4C Figure 4I

Alternative paper-level hypothesis

MHC I-low tumors

severity S unaddressed

The paper's MHC I-low tumor state may substantially overlap a pre-existing IFNγ-inflamed tumor class, rather than representing a distinct state in which low MHC I itself newly drives CD4-linked antitumor biology.

Mechanism: known IFNγ-inflamed tumor state overlap

Reason to inspect: This is a high-quality paper-level identity-overlap concern. MHC I expression is itself regulated by IFNγ signaling, and tumors with low MHC I expression could represent either (a) tumors with intrinsically low antigen presentation that are genuinely more susceptible to CD4-mediated immunity, or (b) tumors that have lost MHC I despite being in an IFNγ-inflamed microenvironment (e.g., through epigenetic silencing or β2m loss), which would mean the 'MHC I-low' label is capturing a subset of IFNγ-inflamed tumors already known to have favorable immunotherapy responses. The paper's observational human analyses group tumors by MHC I expression level, but the packet provides no evidence that the paper separates tumor-cell-intrinsic MHC I loss from the broader IFNγ-inflamed state. If MHC I-low tumors are simply a relabeling of a known inflamed/responsive tumor class, the translational novelty of the human findings is substantially diminished.

Suggested experiment: Within MHC I-low tumors, stratify by IFNγ signaling score and test whether CD4-associated survival benefit persists in MHC I-low tumors with low IFNγ signatures; alternatively, compare MHC I-low tumors arising from genetic β2m loss versus transcriptional downregulation to separate intrinsic MHC I loss from microenvironment-driven regulation.

Relevant panels: Figure 5A Figure 5B Figure 5C Figure 5D Figure 5F Figure 5G Figure 6G

Needs synthesis 3

Claims linked to several panels where neither panel in itself provides sufficient evidence. The useful review task is to inspect the panels together.

Figure 2

Figure 2C Figure 2D Figure 2E Figure 2F

Synthesis claim C47, C48

Author claim

changes in immune cell subsets or their activation were similar between the groups and, thus, not the driving factor in causing the pathway alternation within β2mΔIEC IECs

Specific points under review

C47changes in immune cell subsets were not the driving factor in causing the pathway alternation within β2mΔIEC IECs

C48immune cell activation was not the driving factor in causing the pathway alternation within β2mΔIEC IECs

Questions for Reviewer

Is there evidence for immune-variable to IEC-pathway relationship (not the driving factor in causing the pathway alternation within β2mΔIEC IECs), considering the panels together or the surrounding figure context?
Is there evidence for recipient bm12 → β2mΔIEC (comparison of immune cell subset/activation features between groups), considering the panels together or the surrounding figure context?
Could this alternative explanation be addressed: Whether total lamina propria CD4+ T cell numbers differ between WT and β2m-KO recipients. If they do, the paper's central claim that the phenotype is target-cell-intrinsic rather than effector-driven would be weakened?
Is there evidence for IEC pathway state:pathway alternation within β2mΔIEC IECs (the claim says immune activation was not the driving factor causing the pathway alternation), considering the panels together or the surrounding figure context?
Is there evidence for recipient bm12 → β2mΔIEC (to compare immune activation measurements between groups), considering the panels together or the surrounding figure context?

Figure 5

Figure 5C Figure 5D

Synthesis claim C77

Author claim

While melanoma showed a negative correlation

Specific point under review

melanoma showed a negative correlation

Questions for Reviewer

Is there evidence for cancer type:melanoma (melanoma), considering the panels together or the surrounding figure context?

Figure 5

Figure 5F Figure 5G

Synthesis claim C83

Author claim

participants with MHC I-low tumors with higher CD4+ T cell infiltration had significantly better survival than those with lower CD4+ T cell infiltration (Fig. 5f,g)

Specific point under review

participants with MHC I-low tumors with higher CD4+ T cell infiltration had significantly better survival than those with lower CD4+ T cell infiltration

Questions for Reviewer

Is there evidence for tumor MHC I status:MHC I-low (defines the subgroup in which the survival comparison is claimed), considering the panels together or the surrounding figure context?
Could this alternative explanation be addressed: Whether CD4 abundance is independently prognostic within MHC I-low tumors after adjusting for tumor type, stage, mutational burden, overall immune infiltration, treatment modality, and other known confounders?
Can the figure, legend, or methods clarify whether the plot does not visibly expose the exact model coefficient coding beyond the 0/1 legend labels?
Can the figure, legend, or methods clarify whether the y-axis term list is not fully shown beyond the generic label 'Variable'; likely only one predictor term is present, but the displayed row label is not provided verbatim?

Needs experiment 6

Alternative panel explanations or hypotheses, with a suggestion for a discriminating experiment. A reviewer would check if the authors address these hypotheses elsewhere in the paper.

Figure 1

Figure 1B

Experiment claim C3

Author claim

By contrast, all of the syngeneic WT and β2m-KO recipients survived with no GVHD

Alternative to consider: The excess mortality in allogeneic β2m-KO recipients could reflect host-wide vulnerability to conditioning-associated tissue injury plus allogeneic inflammatory stress, rather than stronger direct CD4+ T-cell-mediated target-cell killing per se.

Suggested experiment: Compare tissue injury markers (e.g., serum LDH, intestinal permeability, hepatic enzymes) early after HSCT in β2m-KO vs WT recipients receiving identical allogeneic transplants, before mortality divergence, to assess whether host tissues show differential injury sensitivity.

Questions for Reviewer

Could this alternative explanation be addressed: Whether the whole-body β2m-KO host has intrinsically reduced tissue tolerance to combined conditioning + allogeneic inflammation, independent of direct CD4-mediated killing. The IEC-restricted model partially addresses this but does not retroactively resolve Figure 1B's whole-body KO design?

Figure 2

Figure 2B

Experiment claim C27

Author claim

allogeneic β2m∆IEC mice demonstrated significantly greater mortality, body weight loss and GVHD score compared to β2mfl/fl (WT) mice (Fig. 2b and Extended Data Fig. 1h,i)

Alternative to consider: The higher mortality in bm12 → β2mΔIEC recipients could be driven by altered intestinal barrier resilience to irradiation and inflammatory stress, rather than uniquely greater direct CD4+ T-cell-mediated killing of IECs.

Suggested experiment: Measure intestinal barrier function (e.g., FITC-dextran permeability, tight junction protein expression) and IEC apoptosis/death markers in β2mΔIEC vs WT recipients at early time points after allogeneic HSCT, before mortality divergence, and compare with matched syngeneic controls to isolate alloreactive-dependent versus stress-dependent barrier failure.

Questions for Reviewer

Could this alternative explanation be addressed: Whether β2m-deficient IECs have intrinsically altered barrier resilience or stress tolerance that is unmasked by inflammatory stress, independent of enhanced direct CD4-mediated killing. The ferroptosis data show a downstream death pathway but do not distinguish whether it is triggered by direct CD4 killing versus indirect inflammatory stress vulnerability?

Figure 2

Figure 2F

Experiment claim C32

Author claim

the activated and IFNγ+CD4+ T cells in the intestinal lamina propria were comparable between WT and β2m-KO recipients (Fig. 2f)

Alternative to consider: Comparable percentages of activated and IFNγ+ lamina propria CD4+ T cells may mask differences in total infiltrate size or absolute pathogenic CD4+ T-cell burden between groups.

Suggested experiment: Quantify absolute numbers of total CD4+ and IFNγ+CD4+ T cells per unit length of intestine or per lamina propria preparation in WT versus β2m-KO allogeneic recipients.

Questions for Reviewer

Could this alternative explanation be addressed: Whether total lamina propria CD4+ T cell numbers differ between WT and β2m-KO recipients. If they do, the paper's central claim that the phenotype is target-cell-intrinsic rather than effector-driven would be weakened?

Figure 4

Figure 4C

Experiment claim C58

Author claim

improved survival and body weight and reduced clinical severity of GVHD in allogeneic β2m∆IEC recipients when compared to diluent-treated β2m∆IEC mice (Fig. 4c and Extended Data Fig. 4g)

Alternative to consider: The survival benefit from DFX could arise from broader systemic protection against transplant-associated inflammatory or oxidative injury, not specifically from blocking IEC ferroptosis downstream of β2m loss.

Suggested experiment: Use genetic ferroptosis rescue (e.g., GPX4 overexpression or ferrostatin-1 targeted to IECs) or IEC-specific iron chelation to test whether IEC-local ferroptosis blockade is sufficient for survival rescue.

Questions for Reviewer

Could this alternative explanation be addressed: Whether the survival benefit is specifically attributable to blocking IEC ferroptosis versus broader systemic anti-inflammatory/anti-oxidant protection. The genotype interaction helps but does not fully resolve this because β2mΔIEC hosts may also have systemically altered iron handling or inflammatory responses?

Figure 4

Figure 4I

Experiment claim C71

Author claim

in vivo administration of anti-IFNγ antibody abolished the difference of lipid peroxidation in IECs between WT and β2m-KO allogeneic recipients (Fig. 4i)

Alternative to consider: The anti-IFNγ effect on relative IEC lipid ROS could reflect a global reduction in intestinal inflammatory stress rather than proving IFNγ is the unique direct proximal trigger of ferroptosis in β2m-deficient IECs.

Suggested experiment: Present absolute lipid ROS values for WT and β2m-KO IECs under IgG and anti-IFNγ conditions separately; complement with in vitro IFNγ treatment of β2m-deficient versus WT organoids to test direct IFNγ-to-ferroptosis signaling in the absence of inflammatory milieu.

Questions for Reviewer

Could this alternative explanation be addressed: Whether IFNγ acts directly on β2m-deficient IECs to trigger ferroptosis or whether its blockade reduces overall intestinal inflammatory burden, secondarily reducing oxidative stress in all IECs including β2m-deficient ones?

Figure 5

Figure 5F

Experiment claim C83

Author claim

participants with MHC I-low tumors with higher CD4+ T cell infiltration had significantly better survival than those with lower CD4+ T cell infiltration (Fig. 5f,g)

Alternative to consider: The survival advantage of the CD4-high subgroup within MHC I-low tumors may reflect broader favorable tumor or treatment context correlated with CD4 abundance, rather than a direct beneficial role of CD4+ T cells.

Suggested experiment: Multivariable Cox regression within MHC I-low tumors adjusting for tumor type, stage, total immune infiltration, CD8 abundance, mutational burden, and treatment to test independent prognostic value of CD4 abundance.

Questions for Reviewer

Could this alternative explanation be addressed: Whether CD4 abundance is independently prognostic within MHC I-low tumors after adjusting for tumor type, stage, mutational burden, overall immune infiltration, treatment modality, and other known confounders?

Evidence question 5

CLEAR could not find sufficient evidence for the claim in the referenced panel. A reviewer should investigate the strength of the claim vs the experimental setup.

Figure 1

Figure 1B

Evidence claim C3

Author claim

By contrast, all of the syngeneic WT and β2m-KO recipients survived with no GVHD

Specific point under review

all of the syngeneic WT recipients survived with no GVHD

Questions for Reviewer

Is there evidence for GVHD status/readout in syngeneic WT recipients (all of the syngeneic WT recipients survived with no GVHD), considering the panels together or the surrounding figure context?
Could this alternative explanation be addressed: Whether the whole-body β2m-KO host has intrinsically reduced tissue tolerance to combined conditioning + allogeneic inflammation, independent of direct CD4-mediated killing. The IEC-restricted model partially addresses this but does not retroactively resolve Figure 1B's whole-body KO design?

Figure 2

Figure 2I

Evidence claim C36

Author claim

While growth kinetics were similar in the absence of Trp1 T cells

Specific point under review

growth kinetics were similar in the absence of Trp1 T cells

Questions for Reviewer

Is there evidence for kinetics (growth kinetics), considering the panels together or the surrounding figure context?

Figure 4

Figure 4A

Evidence claim C53, C54

Author claim

IECs from allogeneic recipients demonstrated increased lipid peroxidation in multiple irradiated or nonirradiated GI-GVHD models when compared to syngeneic controls

Specific points under review

C53IECs from allogeneic recipients demonstrated increased lipid peroxidation in multiple irradiated GI-GVHD models when compared to syngeneic controls

C54IECs from allogeneic recipients demonstrated increased lipid peroxidation in nonirradiated GI-GVHD models when compared to syngeneic controls

Questions for Reviewer

What comparison or control would help interpret this point: Syngeneic control required by the claim is absent?
Is there evidence for the allogeneic-recipient comparison (allogeneic recipients demonstrated increased lipid peroxidation when compared to syngeneic controls), considering the panels together or the surrounding figure context?
Is there evidence for the syngeneic-recipient comparison (when compared to syngeneic controls), considering the panels together or the surrounding figure context?
Is there evidence for more than one irradiated GI-GVHD model (in multiple irradiated GI-GVHD models), considering the panels together or the surrounding figure context?
What comparison or control would help interpret this point: Missing syngeneic observational reference for the claimed allogeneic-versus-syngeneic comparison?
Is there evidence for nonirradiated GI-GVHD (in nonirradiated GI-GVHD models), considering the panels together or the surrounding figure context?

Figure 4

Figure 4D

Evidence claim C61

Author claim

β2m-KO B16 tumors exhibited significantly higher lipid peroxidation compared to both scramble B16 control tumors and those that did not receive Trp1 T cells (Fig. 4d)

Specific point under review

β2m-KO B16 tumors exhibited significantly higher lipid peroxidation compared to scramble B16 control tumors

Questions for Reviewer

What comparison or control would help interpret this point: The panel lacks explicit discrimination of Trp1 transfer status for the scramble and β2m-KO groups used in the claimed comparison?
Is there evidence for explicitly resolved for the direct scramble vs β2m-KO comparison (the claim compares β2m-KO B16 tumors to scramble B16 control tumors as a like-for-like tumor comparison), considering the panels together or the surrounding figure context?

Figure 6

Figure 6G

Evidence claim C90

Author claim

scRNA-seq confirmed a significantly higher IFNγ response signature in MHC I-low tumor cells (Fig. 6g)

Specific point under review

scRNA-seq confirmed a significantly higher IFNγ response signature in MHC I-low tumor cells

Questions for Reviewer

Is there evidence for tumor cells (the claim is specifically about tumor cells), considering the panels together or the surrounding figure context?

Figure synthesis refinement 1

The claim appears to require figure-level synthesis across multiple panels rather than a missing-evidence question for one panel.

Figure 2

Figure 2F

Synthesis note claim C45, C46

Author claim

collectively suggest that changes in immune cell subsets or their activation were similar between the groups

Specific points under review

C45changes in immune cell subsets were similar between the groups

C46their activation were similar between the groups

Figure text check 1

The panel structure is usable, but the figure text, labels, or legend leave a local ambiguity that should be checked before the claim is interpreted.

Figure 3

Figure 3G

Figure text claim C43

Author claim

the β2mΔIEC group exhibited significant enrichment in lipid peroxidation (Fig. 3g)

Specific point under review

the β2mΔIEC group exhibited significant enrichment in lipid peroxidation

Questions for Reviewer

Can the figure, legend, or methods clarify whether individual GO term labels are not provided in the panel input?
Can the figure, legend, or methods clarify whether no explicit graph-visible baseline/reference category is shown?

Unclear baseline 1

The referenced panel needs a clear baseline or reference condition before the claim can be reviewed against it.

Figure 5

Figure 5B

Baseline claim C76

Author claim

This was further supported by a significant negative correlation between mean MHC I expression and CD4+ T cell scores across these datasets (Fig. 5b)

Specific point under review

there was a significant negative correlation between mean MHC I expression and CD4+ T cell scores across these datasets

Questions for Reviewer

Which baseline or reference condition should be used to interpret this claim in Figure 5B?

Supplementary lookup 9

The author claim points to supplementary or Extended Data evidence that should be inspected before judging the main-figure panel.

Figure 1

Figure 1D Figure 1E Figure 1F Figure 1G Figure 1H

Supplementary claim C7, C8

Author claim

Compared to WT control, allogeneic β2m-KO recipients showed comparable percentages of donor CD4+IFNγ+ or CD8+IFNγ+ cells

Specific point under review

Compared to WT control, allogeneic β2m-KO recipients showed comparable percentages of donor CD4+IFNγ+ or CD8+IFNγ+ cells

Questions for Reviewer

The source points to Extended Data Figure 1 panel(s) e; donor CD4+IFNγ+ cell percentage. Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?
The source points to Extended Data Figure 1 panel(s) e; donor CD8+IFNγ+ cell percentage. Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?

Figure 2

Figure 2B

Supplementary claim C29

Author claim

allogeneic β2m∆IEC mice demonstrated significantly greater mortality, body weight loss and GVHD score compared to β2mfl/fl (WT) mice (Fig. 2b and Extended Data Fig. 1h,i)

Specific point under review

allogeneic β2m∆IEC mice demonstrated significantly greater GVHD score compared to β2mfl/fl (WT) mice

Questions for Reviewer

The source points to Extended Data Figure 1 panel(s) h, i; GVHD clinical score. Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?

Figure 3

Figure 3I

Supplementary claim C50

Author claim

While total intestinal iron content (Fe2+ and Fe3+) remained similar

Specific point under review

total intestinal iron content (Fe2+ and Fe3+) remained similar

Questions for Reviewer

The source points to Extended Data Figure 3 panel(s) g; total intestinal iron content (Fe2+ and Fe3+). Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?

Figure 3

Figure 3I

Supplementary claim C51

Author claim

Fe2+ levels were elevated in allogeneic β2mΔIEC recipients compared to WT

Questions for Reviewer

The source points to Extended Data Figure 3 panel(s) h; Fe2+ levels. Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?

Figure 4

Figure 4A

Supplementary claim C55, C56

Author claim

IEC-specific KO of major ferroptosis inhibitor glutathione peroxidase 4 (GPX4∆IEC) exacerbated GVHD mortality and severity when compared to allogeneic corn-oil-treated controls (GPX4fl/fl)

Specific point under review

IEC-specific KO of major ferroptosis inhibitor glutathione peroxidase 4 (GPX4∆IEC) exacerbated GVHD mortality and severity when compared to allogeneic corn-oil-treated controls (GPX4fl/fl)

Questions for Reviewer

The source points to Extended Data Figure 4 panel(s) e, f; GVHD mortality (GPX4ΔIEC vs GPX4fl/fl + corn oil). Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?
The source points to Extended Data Figure 4 panel(s) e, f; GVHD clinical severity (GPX4ΔIEC vs GPX4fl/fl + corn oil). Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?

Figure 4

Figure 4C

Supplementary claim C59

Author claim

improved survival and body weight and reduced clinical severity of GVHD in allogeneic β2m∆IEC recipients when compared to diluent-treated β2m∆IEC mice (Fig. 4c and Extended Data Fig. 4g)

Specific point under review

the administration of DFX improved body weight in allogeneic β2m∆IEC recipients when compared to diluent-treated β2m∆IEC mice

Questions for Reviewer

The source points to Extended Data Figure 4 panel(s) g; body weight. Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?

Figure 4

Figure 4G Figure 4H

Supplementary claim C66, C67

Author claim

Upon IFNγ stimulation, β2m-KO B16 cells exhibited higher IRF1 mRNA and protein levels

Specific point under review

Upon IFNγ stimulation, β2m-KO B16 cells exhibited higher IRF1 mRNA and protein levels

Questions for Reviewer

The source points to Extended Data Figure 4 panel(s) j, k; IRF1 mRNA (IFNγ-stimulated vs unstimulated). Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?
The source points to Extended Data Figure 4 panel(s) j, k; IRF1 protein (IFNγ-stimulated vs unstimulated). Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?

Figure 4

Figure 4I

Supplementary claim C72

Author claim

the in vivo scRNA-seq data revealed a trend toward higher Acsl4 mRNA levels in allogeneic β2mΔIEC mice

Questions for Reviewer

The source points to Extended Data Figure 4 panel(s) m; Acsl4 mRNA (scRNA-seq). Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?

Figure 4

Figure 4I

Supplementary claim C73, C74

Author claim

we observed an increase in ACSL4 expression at both transcript and protein levels in β2m-KO cells upon IFNγ stimulation

Specific point under review

we observed an increase in ACSL4 expression at both transcript and protein levels in β2m-KO cells upon IFNγ stimulation

Questions for Reviewer

The source points to Extended Data Figure 4 panel(s) n, o; ACSL4 transcript (IFNγ-stimulated vs unstimulated). Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?
The source points to Extended Data Figure 4 panel(s) n, o; ACSL4 protein (IFNγ-stimulated vs unstimulated). Should that supplementary or Extended Data evidence be inspected before judging this claim from the main figure panels?

Claim wording check 1

The claim needs a more explicit tested point before the figure evidence can be reviewed cleanly.

Figure 5

Figure 5C Figure 5D

Wording claim C78

Author claim

other cancer types exhibited variable trends

Questions for Reviewer

Can this claim be rewritten or split so the reviewer can identify the exact cancer type, trend direction, and readout being tested?