pith. sign in

arxiv: 2604.23917 · v1 · submitted 2026-04-27 · 📊 stat.ME

MR-CCC: Bayesian Mendelian Randomization for Causal Cell--Cell Communication

Bitan Sarkar , Yang Ni This is my paper

Pith reviewed 2026-05-08 02:33 UTC · model grok-4.3

classification 📊 stat.ME
keywords Bayesian Mendelian randomizationcausal inferencecell-cell communicationligand-receptor interactionscis-eQTLsingle-cell RNA sequencingspike-and-slab prior
0
0 comments X

The pith

A Bayesian Mendelian randomization method uses cis-eQTLs as instruments to infer causal ligand-receptor signaling and receptor-modulated interaction effects while controlling false discoveries under confounding.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Standard approaches to cell-cell communication infer signaling from ligand-receptor co-expression, but this cannot separate true causal effects from shared regulation or other confounders. MR-CCC applies Mendelian randomization by treating cis-eQTLs as instruments for ligand and receptor expression levels. It adds an interaction term so that the ligand's causal effect can change depending on receptor abundance. A spike-and-slab prior produces posterior inclusion probabilities for each potential signaling pair, and the method is shown in benchmarks to limit false positives better than regression or other MR approaches while still detecting true signals. When applied to NK-cell to monocyte data, it recovers eight specific causal pathways across several signaling families.

Core claim

MR-CCC establishes causal cell-cell communication by instrumenting ligand and receptor expression with cis-eQTLs inside a Bayesian model that explicitly includes a receptor-modulated interaction term, yielding posterior inclusion probabilities that quantify evidence for causal signaling and allowing separate estimation of main ligand effects and interaction effects.

What carries the argument

Bayesian Mendelian randomization model that instruments ligand and receptor expression via cis-eQTLs, incorporates a ligand-by-receptor interaction term, and uses a spike-and-slab prior to select causal effects.

If this is right

  • MR-CCC controls false discoveries under confounding while retaining high power, outperforming naive regression, MVMR, and MR-BMA in benchmarks.
  • The model uniquely recovers both the ligand main effect and the receptor-modulated interaction effect for each pair.
  • Applied to OneK1K NK-to-monocyte data, it identifies eight causal discoveries spanning GABA, interferon, interleukin, and prostaglandin signaling.
  • It detects stoichiometry-dependent effects such as dissociation of the two IL-18 receptor chains and joint discovery of both IFN-γ receptor subunits.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The interaction term could be used to predict how receptor blockade would alter ligand-driven outcomes in therapeutic settings.
  • Extending the framework to additional cell-type pairs or disease states would likely surface causal networks missed by co-expression methods alone.
  • The posterior inclusion probabilities offer a natural ranking for prioritizing experimental validation of the discovered signaling pairs.

Load-bearing premise

The selected cis-eQTLs function as valid instruments for ligand and receptor expression levels with no unmeasured pleiotropy or other Mendelian randomization assumption violations.

What would settle it

A simulation or real dataset in which ligand and receptor expression are correlated solely through an unmeasured confounder but no direct causal signaling occurs, and MR-CCC nevertheless assigns high posterior probability to a causal link.

Figures

Figures reproduced from arXiv: 2604.23917 by Bitan Sarkar, Yang Ni.

Figure 1
Figure 1. Figure 1: Overview of the MR-CCC framework, simulation benchmark, and lead biological discovery. (1) Ligand–receptor co-expression between a sender and a receiver cell can be driven by an unmeasured confounder U, so correlation alone cannot establish causal communication. (2) MR-CCC uses cis-eQTLs G and H as instruments for ligand X and receptor Z and adjusts for covariates V , blocking U. The working outcome model … view at source ↗
Figure 2
Figure 2. Figure 2: Communication scores across scenarios, sample sizes, and methods. Boxplots of communication scores across 20 replicates, stratified by scenario (rows: S1–S3) and sample size (columns: n ∈ {500, 1000, 10,000, 30,000}). For MR-CCC the score is the posterior inclusion probability Pr(γ = 1 | data). For MR-BMA the score is MIPX. For OLS the score is 1 − pF , where pF is the joint F-test p-value for H0 : βX = βX… view at source ↗
Figure 3
Figure 3. Figure 3: Estimates of the ligand main effect βX across scenarios, sample sizes, and methods. Boxplots of the estimated ligand main effect βˆX across 20 replicates, stratified by scenario and sample size. Rows correspond to scenarios S1–S3; columns correspond to sample size. For MR-BMA the plotted value is the model-averaged causal effect (MACE) for the ligand exposure. The red dashed line indicates the true value o… view at source ↗
Figure 4
Figure 4. Figure 4: Estimates of the receptor-modulated interaction effect βXZ across sce￾narios, sample sizes, and methods. Boxplots of the estimated ligand–receptor interac￾tion effect βˆXZ across 20 replicates, stratified by scenario and sample size. MR-BMA and MVMR are omitted as neither models the interaction term βXZ. The red dashed line indi￾cates the true value (0 for S1 and S3; 0.3 for S2). Under S1 and S3, where the… view at source ↗
Figure 5
Figure 5. Figure 5: Posterior inclusion probabilities for the NK cells → monocytes analysis. Posterior inclusion probabilities for all 41 ligand–receptor–pathway triplets in the NK cells → Monocytes analysis (n = 651 donors), ranked from highest to lowest PIP. Points are coloured by pathway class. Black rings identify the eight discoveries with PIP > 0.5: the two IL18 receptor-chain pairs lead at PIP = 0.98 and 0.93, followed… view at source ↗
Figure 6
Figure 6. Figure 6: Pathway-specific posterior mean effects for the NK cells → Monocytes analysis. The left panel shows |βˆ (s) X |, the absolute standardized ligand main effect, and the right panel shows |βˆ (s) XZ|, the absolute standardized receptor-modulation effect. Point size encodes effect magnitude; color encodes PIP on a 0–1 scale, with black borders marking the eight discoveries (PIP > 0.5). Discoveries concentrate … view at source ↗
Figure 7
Figure 7. Figure 7: Receptor-modulated ligand effect curves for the eight NK cells → monocytes discoveries. Receptor-modulated ligand effects for the NK cells → Monocytes analysis (n = 651 donors). Each curve represents βˆ (s) X + βˆ (s) XZ ·(Z/sd(Z)), plotted over the observed donor range of standardized receptor expression. Only the eight discovery pairs (PIP > 0.5) are displayed, grouped into the four pathway panels with c… view at source ↗
Figure 8
Figure 8. Figure 8: Causal diagram of the MR-CCC structural model. The causal effect of ligand expression on pathway activity is βX +βXZZi , allowing receptor-modulated communi￾cation. Genetic instruments Gi and Hi identify ligand and receptor expression respectively; observed covariates Vi and unmeasured confounders Ui may affect all three observed biolog￾ical variables (Xi , Zi , Yi). Solid arrows indicate causal effects; d… view at source ↗
read the original abstract

Cell--cell communication (CCC) is commonly inferred from ligand--receptor co-expression, an associational paradigm that cannot distinguish causal signaling from shared regulation or confounding. We propose MR-CCC, a Bayesian Mendelian randomization framework that uses cis-eQTLs as instruments for ligand and receptor expression and explicitly models receptor-modulated ligand effects through an interaction term, so the causal effect of a ligand can vary with receptor abundance. A spike--and--slab prior yields posterior inclusion probabilities quantifying evidence for causal signaling, and an efficient Gibbs sampler provides scalable inference. Benchmarked against naive regression, MVMR, and MR-BMA, MR-CCC controls false discoveries under confounding while retaining high power, and uniquely estimates both the ligand main and receptor-modulated interaction effects. Applied to the OneK1K NK cells $\to$ monocytes axis, MR-CCC identifies eight discoveries across GABA, interferon, interleukin, and prostaglandin signaling, including a stoichiometry-dependent dissociation of the two IL-18 receptor chains and co-discovery of both obligate IFN-$\gamma$ receptor subunits.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper proposes MR-CCC, a Bayesian Mendelian randomization framework for inferring causal cell-cell communication (CCC). It treats cis-eQTLs as instruments for ligand and receptor expression, introduces an interaction term to allow the ligand effect to be modulated by receptor abundance, uses a spike-and-slab prior to obtain posterior inclusion probabilities (PIPs) for causal signaling, and employs an efficient Gibbs sampler. The method is benchmarked in simulations against naive regression, MVMR, and MR-BMA, claiming improved false-discovery control under confounding while preserving power and uniquely recovering both main and interaction effects. In an application to the OneK1K NK-cell-to-monocyte axis, it reports eight discoveries across GABA, interferon, interleukin, and prostaglandin pathways, including stoichiometry-dependent dissociation of IL-18 receptor chains and joint detection of both IFN-γ receptor subunits.

Significance. If the central claims hold, the work would be a useful methodological advance for moving CCC inference from purely associational co-expression analyses toward causal statements, while providing interpretable estimates of receptor-modulated effects. The spike-and-slab formulation and Gibbs sampler are practical strengths that enable scalable posterior inference and direct quantification of evidence via PIPs. The real-data application yields biologically plausible and specific findings (e.g., the IL-18 and IFN-γ receptor results) that could be followed up experimentally. Significance is tempered by the fact that the method inherits all standard MR assumptions; any violation would propagate directly to the reported discoveries and false-discovery claims.

major comments (3)
  1. [§4] §4 (Simulation studies): The reported false-discovery control is demonstrated only under simulated confounding that preserves instrument validity; no simulations or sensitivity analyses incorporate realistic pleiotropy or other violations of the exclusion restriction that are common for cis-eQTLs. Because the central claim of superior FDR control rests on the instruments remaining valid, this omission is load-bearing for the benchmarking conclusions.
  2. [§5.2–5.3] §5.2–5.3 (Real-data application): The eight reported discoveries (Table 2 or equivalent) are presented with PIP thresholds but without any empirical assessment of horizontal pleiotropy (e.g., MR-Egger intercept, heterogeneity tests, or leave-one-out analyses) for the selected cis-eQTL instruments. Given that the method’s causal interpretation and the specific biological claims (IL-18 chain dissociation, IFN-γ subunit co-discovery) depend on the absence of unmeasured pleiotropy, this gap directly affects the credibility of the applied results.
  3. [Eq. (3)–(5)] Eq. (3)–(5) (Model specification): The interaction term is introduced to capture receptor-modulated ligand effects, yet the paper provides no robustness checks against alternative functional forms (e.g., nonlinear or threshold modulation) or against misspecification of the main-effect versus interaction decomposition. Because the unique selling point is the ability to estimate both effects separately, the lack of such checks weakens the claim that the interaction term is reliably identified.
minor comments (2)
  1. [§3.1] Notation for the ligand–receptor interaction coefficient is introduced without an explicit statement of its identifiability conditions under the chosen instrument set; a short paragraph clarifying the rank conditions would improve readability.
  2. [§5.3] The abstract states that MR-CCC “uniquely estimates both the ligand main and receptor-modulated interaction effects,” but the main text does not compare the magnitude or uncertainty of these two coefficients across the eight discoveries; adding this comparison would strengthen the applied section.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive and detailed review. The comments identify important gaps in simulation coverage, empirical validation of assumptions in the application, and robustness of the interaction modeling. We address each point below and will incorporate the suggested additions in the revised manuscript.

read point-by-point responses

  1. Referee: [§4] §4 (Simulation studies): The reported false-discovery control is demonstrated only under simulated confounding that preserves instrument validity; no simulations or sensitivity analyses incorporate realistic pleiotropy or other violations of the exclusion restriction that are common for cis-eQTLs. Because the central claim of superior FDR control rests on the instruments remaining valid, this omission is load-bearing for the benchmarking conclusions.

    Authors: We agree that the current simulation design focuses on valid instruments under confounding and does not yet include direct pleiotropy or exclusion-restriction violations. This limits the strength of the FDR-control claims under realistic cis-eQTL scenarios. In the revised manuscript we will add simulation scenarios in which instruments have direct pleiotropic effects on the outcome (both constant and heterogeneous pleiotropy). We will report FDR, power, and PIP calibration for MR-CCC versus the benchmark methods under these conditions, and will discuss the implications for the method's robustness. revision: yes

  2. Referee: [§5.2–5.3] §5.2–5.3 (Real-data application): The eight reported discoveries (Table 2 or equivalent) are presented with PIP thresholds but without any empirical assessment of horizontal pleiotropy (e.g., MR-Egger intercept, heterogeneity tests, or leave-one-out analyses) for the selected cis-eQTL instruments. Given that the method’s causal interpretation and the specific biological claims (IL-18 chain dissociation, IFN-γ subunit co-discovery) depend on the absence of unmeasured pleiotropy, this gap directly affects the credibility of the applied results.

    Authors: We acknowledge that the applied results currently lack direct empirical checks for horizontal pleiotropy on the selected instruments. While the method inherits the standard MR assumptions, the specific biological interpretations (e.g., stoichiometry-dependent IL-18 receptor dissociation and joint IFN-γ receptor detection) would be strengthened by such diagnostics. In the revision we will add MR-Egger intercept tests, Cochran’s Q heterogeneity statistics, and leave-one-out analyses for the cis-eQTLs underlying the eight reported discoveries. Any evidence of pleiotropy will be discussed in relation to the causal claims. revision: yes

  3. Referee: [Eq. (3)–(5)] Eq. (3)–(5) (Model specification): The interaction term is introduced to capture receptor-modulated ligand effects, yet the paper provides no robustness checks against alternative functional forms (e.g., nonlinear or threshold modulation) or against misspecification of the main-effect versus interaction decomposition. Because the unique selling point is the ability to estimate both effects separately, the lack of such checks weakens the claim that the interaction term is reliably identified.

    Authors: The interaction term is biologically motivated by receptor stoichiometry, yet we did not examine sensitivity to alternative functional forms or to misspecification of the main-effect/interaction split. We agree this is a gap. In the revised manuscript we will include sensitivity analyses that (i) replace the linear interaction with a quadratic term, (ii) use a threshold-based modulation, and (iii) alter the prior decomposition between main and interaction effects. We will report how the posterior inclusion probabilities and effect estimates for the eight discoveries change under these alternatives. revision: yes

Circularity Check

0 steps flagged

MR-CCC derivation uses standard Bayesian MR with interaction term; no reduction of results to inputs by construction

full rationale

The framework defines causal effects via Mendelian randomization assumptions on cis-eQTL instruments, an explicit interaction term for receptor modulation, and a spike-and-slab prior for inclusion probabilities. These are standard modeling choices whose outputs (posterior probabilities and effect estimates) are obtained by applying the model to observed data rather than being tautologically equivalent to the inputs. Benchmarking and real-data discoveries follow from the fitted posterior, not from any self-definitional loop or load-bearing self-citation that collapses the claimed causal inferences back to the model specification itself. The derivation chain remains self-contained against external data and assumptions.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on the validity of cis-eQTLs as instruments and standard MR assumptions; no free parameters or invented entities are quantified in the abstract.

axioms (2)
  • domain assumption cis-eQTLs are valid instruments for ligand and receptor expression
    Used explicitly as instruments to identify causal effects in the MR framework.
  • domain assumption No unmeasured pleiotropy or confounding violates the MR assumptions
    Required for the causal interpretation of the posterior inclusion probabilities.

pith-pipeline@v0.9.0 · 5484 in / 1456 out tokens · 47869 ms · 2026-05-08T02:33:17.872087+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

41 extracted references · 34 canonical work pages

  1. [1]

    Grace Gordon, Stacey Andersen, Qinyi Lu, Antonia Rowson, Thomas R

    Seyhan Yazar, Jose Alquicira-Hernandez, Kristof Wing, Anne Senabouth, M. Grace Gordon, Stacey Andersen, Qinyi Lu, Antonia Rowson, Thomas R. P. Taylor, Linda Clarke, Katia Maccora, Christine Chen, Anthony L. Cook, Chun Jimmie Ye, Kirsten A. Fairfax, Alex W. Hewitt, and Joseph E. Powell. Single-cell eQTL mapping identifies cell type-specific genetic control...

  2. [2]

    doi: 10.1126/science.abf3041

    work page doi:10.1126/science.abf3041

  3. [3]

    Waldmann, and Yutaka Tagaya

    Sigrid Dubois, Jill Mariner, Thomas A. Waldmann, and Yutaka Tagaya. IL-15Ralpha recycles and presents IL-15 in trans to neighboring cells.Immunity, 17(5):537–547, 2002. doi: 10.1016/S1074-7613(02)00429-6

    work page doi:10.1016/s1074-7613(02)00429-6 2002

  4. [4]

    Fehniger and Michael A

    Todd A. Fehniger and Michael A. Caligiuri. Interleukin 15: biology and relevance to human disease.Blood, 97(1):14–32, 2001. doi: 10.1182/blood.V97.1.14

    work page doi:10.1182/blood.v97.1.14 2001

  5. [5]

    Waldmann

    Thomas A. Waldmann. The biology of interleukin-2 and interleukin-15: implications for cancer therapy and vaccine design.Nature Reviews Immunology, 6(8):595–601, 2006. doi: 10.1038/nri1901. 106

    work page doi:10.1038/nri1901 2006

  6. [6]

    Tall and Laurent Yvan-Charvet

    Alan R. Tall and Laurent Yvan-Charvet. Cholesterol, inflammation and innate immu- nity.Nature Reviews Immunology, 15(2):104–116, 2015. doi: 10.1038/nri3793

    work page doi:10.1038/nri3793 2015

  7. [7]

    Anton M. Jetten. Retinoid-related orphan receptors (RORs): critical roles in develop- ment, immunity, circadian rhythm, and cellular metabolism.Nuclear Receptor Signaling, 7:e003, 2009. doi: 10.1621/nrs.07003

    work page doi:10.1621/nrs.07003 2009

  8. [8]

    Yang, Bhanu P

    Xuexian O. Yang, Bhanu P. Pappu, Roza Nurieva, Askar Akimzhanov, Hong Soon Kang, Yeonseok Chung, Liang Ma, Bhavana Shah, Aris D. Panopoulos, Kimberly S. Schluns, Stephanie S. Watowich, Qiang Tian, Anton M. Jetten, and Chen Dong. T helper 17 lineage differentiation is programmed by orphan nuclear receptors ROR alpha and ROR gamma.Immunity, 28(1):29–39, 200...

    work page doi:10.1016/j.immuni.2007.11.016 2008

  9. [9]

    Hales, and Daniel L

    Jian Tian, Christine Chau, Thomas G. Hales, and Daniel L. Kaufman. GABA(A) receptors mediate inhibition of T cell responses.Journal of Neuroimmunology, 96(1): 21–28, 1999. doi: 10.1016/s0165-5728(98)00264-1

    work page doi:10.1016/s0165-5728(98)00264-1 1999

  10. [10]

    GABA, a natural im- munomodulator of T lymphocytes.Journal of Neuroimmunology, 205(1–2):44–50, 2008

    Helena Bj¨ urstr¨ om, Jin Wang, Idun Ericsson, Mikael Bengtsson, Yang Liu, Suresh Kumar-Mendu, Shohreh Issazadeh-Navikas, and Bryndis Birnir. GABA, a natural im- munomodulator of T lymphocytes.Journal of Neuroimmunology, 205(1–2):44–50, 2008. doi: 10.1016/j.jneuroim.2008.09.015

    work page doi:10.1016/j.jneuroim.2008.09.015 2008

  11. [11]

    Hertzog, Timothy Ravasi, and David A

    Kate Schroder, Paul J. Hertzog, Timothy Ravasi, and David A. Hume. Interferon- gamma: an overview of signals, mechanisms and functions.Journal of Leukocyte Biol- ogy, 75(2):163–189, 2004. doi: 10.1189/jlb.0603252

    work page doi:10.1189/jlb.0603252 2004

  12. [12]

    Kunkel, Ulrike Gosslar, et al

    Jianlin Pan, Eric J. Kunkel, Ulrike Gosslar, et al. A novel chemokine ligand for CCR10 and CCR3 expressed by epithelial cells in mucosal tissues.Journal of Immunology, 165 (6):2943–2949, 2000. doi: 10.4049/jimmunol.165.6.2943

    work page doi:10.4049/jimmunol.165.6.2943 2000

  13. [13]

    CCL27–CCR10 interactions regulate T cell-mediated skin inflammation.Nature Medicine, 8(2):157–165, 2002

    Bernhard Homey, Harri Alenius, Andreas M¨ uller, et al. CCL27–CCR10 interactions regulate T cell-mediated skin inflammation.Nature Medicine, 8(2):157–165, 2002. doi: 10.1038/nm0202-157. 107

    work page doi:10.1038/nm0202-157 2002

  14. [14]

    Wnt/β-catenin signaling in development and disease.Cell, 127(3):469– 480, 2006

    Hans Clevers. Wnt/β-catenin signaling in development and disease.Cell, 127(3):469– 480, 2006. doi: 10.1016/j.cell.2006.10.018

    work page doi:10.1016/j.cell.2006.10.018 2006

  15. [15]

    Kennedy, Mary Glaccum, Shelley N

    Melanie K. Kennedy, Mary Glaccum, Shelley N. Brown, Edward A. Butz, Joanne L. Viney, Monica Embers, Noriko Matsuki, Kimberly Charrier, Louise Sedger, Charles R. Willis, Kenneth Brasel, Patrick J. Morrissey, Kent Stocking, Johanna C. L. Schuh, Se- bastian Joyce, and Jacques J. Peschon. Reversible defects in natural killer and memory CD8 T cell lineages in ...

    work page doi:10.1084/jem.191.5.771 2000

  16. [16]

    Schluns and Leo Lefrancois

    Kimberly S. Schluns and Leo Lefrancois. Cytokine control of memory T-cell development and survival.Nature Reviews Immunology, 3(4):269–279, 2003. doi: 10.1038/nri1052

    work page doi:10.1038/nri1052 2003

  17. [17]

    Heinrich, Iris Behrmann, Serge Haan, Heike M

    Peter C. Heinrich, Iris Behrmann, Serge Haan, Heike M. Hermanns, Gerhard M¨ uller- Newen, and Fred Schaper. Principles of interleukin (IL)-6-type cytokine signalling and its regulation.Biochemical Journal, 374(1):1–20, 2003. doi: 10.1042/bj20030407

    work page doi:10.1042/bj20030407 2003

  18. [18]

    Inhibition of natural killer cell cytotoxicity by interleukin-6: implications for the pathogenesis of macrophage activation syndrome

    Loredana Cifaldi, Giusi Prencipe, Ivan Caiello, Claudia Bracaglia, Franco Franceschini, Fabrizio De Benedetti, and Franco Locatelli. Inhibition of natural killer cell cytotoxicity by interleukin-6: implications for the pathogenesis of macrophage activation syndrome. Arthritis & Rheumatology, 67(11):3037–3046, 2015. doi: 10.1002/art.39313

    work page doi:10.1002/art.39313 2015

  19. [19]

    Regulation of immune responses by prostaglandin E2.Journal of Immunology, 188(1):21–28, 2012

    Pawel Kalinski. Regulation of immune responses by prostaglandin E2.Journal of Immunology, 188(1):21–28, 2012. doi: 10.4049/jimmunol.1101029

    work page doi:10.4049/jimmunol.1101029 2012

  20. [20]

    Davies, Maaike J

    Martin Stacey, Gin-Wen Chang, Janet Q. Davies, Maaike J. Kwakkenbos, Ralph D. Sanderson, J¨ org Hamann, Siamon Gordon, and Hsi-Hsien Lin. The epidermal growth factor-like domains of the human EMR2 receptor mediate cell attachment through chondroitin sulfate glycosaminoglycans.Blood, 102(8):2916–2924, 2003. doi: 10.1182/ blood-2003-04-1320

    2003

  21. [21]

    Dustin, Reiner Rothlein, Atul K

    Michael L. Dustin, Reiner Rothlein, Atul K. Bhan, Charles A. Dinarello, and Timo- thy A. Springer. Induction by IL 1 and interferon-gamma: tissue distribution, biochem- 108 istry, and function of a natural adherence molecule (ICAM-1).Journal of Immunology, 137(1):245–254, 1986

    1986

  22. [22]

    Myones, James G

    Barry L. Myones, James G. Dalzell, Nancy Hogg, and Gordon D. Ross. Neutrophil and monocyte cell surface p150,95 has iC3b-receptor (CR4) activity resembling CR3. Journal of Clinical Investigation, 82(2):640–651, 1988. doi: 10.1172/JCI113643

    work page doi:10.1172/jci113643 1988

  23. [23]

    IL-6 in inflammation, immunity, and disease.Cold Spring Harbor Perspectives in Biology, 6(10):a016295,

    Toshio Tanaka, Masashi Narazaki, and Tadamitsu Kishimoto. IL-6 in inflammation, immunity, and disease.Cold Spring Harbor Perspectives in Biology, 6(10):a016295,

  24. [24]

    doi: 10.1101/cshperspect.a016295

    work page doi:10.1101/cshperspect.a016295

  25. [25]

    Springer

    Timothy A. Springer. Adhesion receptors of the immune system.Nature, 346(6283): 425–434, 1990. doi: 10.1038/346425a0

    work page doi:10.1038/346425a0 1990

  26. [26]

    J¨ org Hamann, Barbara Vogel, Gijs M. W. van Schijndel, and Ren´ e A. W. van Lier. The seven-span transmembrane receptor CD97 has a cellular ligand (CD55, DAF).Journal of Experimental Medicine, 184(3):1185–1189, 1996. doi: 10.1084/jem.184.3.1185

    work page doi:10.1084/jem.184.3.1185 1996

  27. [27]

    Brent Oppmann, Rosanne Lesley, Bj¨ orn Blom, Joseph C. Timans, Yuqin Xu, Brent Hunte, Fernando Vega, Nancy Yu, Jian Wang, Kathleen Singh, Fran¸ coise Zonin, Ed- ward Vaisberg, Tatyana Churakova, Mingde Liu, Daniel Gorman, Janet Wagner, San- dra Zurawski, Yong-Jun Liu, Jeffrey S. Abrams, Kevin W. Moore, Donna Rennick, Ren´ e de Waal-Malefyt, Charles Hannum...

    work page doi:10.1016/s1074-7613(00)00070-4 2000

  28. [28]

    Thromboxane A2: physiology/pathophysiology, cellular signal transduction and pharmacology.Pharmacology & Therapeutics, 118(1):18–35, 2008

    Norimichi Nakahata. Thromboxane A2: physiology/pathophysiology, cellular signal transduction and pharmacology.Pharmacology & Therapeutics, 118(1):18–35, 2008. doi: 10.1016/j.pharmthera.2008.01.001

    work page doi:10.1016/j.pharmthera.2008.01.001 2008

  29. [29]

    Semaphorins and their 109 receptors in immune cell interactions.Nature Immunology, 9(1):17–23, 2008

    Kotaro Suzuki, Atsushi Kumanogoh, and Hitoshi Kikutani. Semaphorins and their 109 receptors in immune cell interactions.Nature Immunology, 9(1):17–23, 2008. doi: 10. 1038/ni1553

    2008

  30. [30]

    Bach, Michel Aguet, and Robert D

    Elizabeth A. Bach, Michel Aguet, and Robert D. Schreiber. The IFN gamma receptor: a paradigm for cytokine receptor signaling.Annual Review of Immunology, 15:563–591,

  31. [31]

    doi: 10.1146/annurev.immunol.15.1.563

    work page doi:10.1146/annurev.immunol.15.1.563

  32. [32]

    1996, ARA&A, 34, 645, doi:10.1146/annurev

    Charles A. Dinarello. Immunological and inflammatory functions of the interleukin- 1 family.Annual Review of Immunology, 27:519–550, 2009. doi: 10.1146/annurev. immunol.021908.132612

    work page doi:10.1146/annurev 2009

  33. [33]

    Mutual activa- tion of natural killer cells and monocytes mediated by NKp80-AICL interaction.Nature Immunology, 7(12):1334–1342, 2006

    Stefanie Welte, Sabine Kuttruff, Inja Waldhauer, and Alexander Steinle. Mutual activa- tion of natural killer cells and monocytes mediated by NKp80-AICL interaction.Nature Immunology, 7(12):1334–1342, 2006. doi: 10.1038/ni1402

    work page doi:10.1038/ni1402 2006

  34. [34]

    Ahuja and Philip M

    Sunil K. Ahuja and Philip M. Murphy. The CXC chemokines growth-regulated oncogene (GRO) alpha, GRObeta, GROgamma, neutrophil-activating peptide-2, and epithelial cell-derived neutrophil-activating peptide-78 are potent agonists for the type B, but not the type A, human interleukin-8 receptor.Journal of Biological Chemistry, 271(34): 20545–20550, 1996. doi...

    work page doi:10.1074/jbc.271.34.20545 1996

  35. [35]

    Okamura, H

    H. Okamura, H. Tsutsui, T. Komatsu, M. Yutsudo, A. Hakura, T. Tanimoto, K. Torigoe, T. Okura, Y. Nukada, K. Hattori, et al. Cloning of a new cytokine that induces IFN- gamma production by T cells.Nature, 378(6552):88–91, 1995. doi: 10.1038/378088a0

    work page doi:10.1038/378088a0 1995

  36. [36]

    Born, Erik Thomassen, Timothy A

    Tara L. Born, Erik Thomassen, Timothy A. Bird, and John E. Sims. Cloning of a novel receptor subunit, AcPL, required for interleukin-18 signalling.Journal of Biological Chemistry, 273:29445–29450, 1998. doi: 10.1074/jbc.273.45.29445

    work page doi:10.1074/jbc.273.45.29445 1998

  37. [37]

    Immunological functions of the neuropilins and plexins as receptors for semaphorins.Nature Reviews Immunology, 13(11):802–814,

    Atsushi Kumanogoh and Hitoshi Kikutani. Immunological functions of the neuropilins and plexins as receptors for semaphorins.Nature Reviews Immunology, 13(11):802–814,

  38. [38]

    doi: 10.1038/nri3545. 110

    work page doi:10.1038/nri3545

  39. [39]

    Penning, Stephan Steckelbroeck, David R

    Trevor M. Penning, Stephan Steckelbroeck, David R. Bauman, Matthew W. Miller, Yi Jin, Donna M. Peehl, Adriane E. Thigpen, and Hsueh-Kung Lin. Aldo-keto reduc- tase (AKR) 1C3: role in prostate disease and the development of specific inhibitors. Molecular and Cellular Endocrinology, 248(1–2):182–191, 2006. doi: 10.1016/j.mce. 2005.12.009

    work page doi:10.1016/j.mce 2006

  40. [40]

    Hata and Richard M

    Akiyoshi N. Hata and Richard M. Breyer. Pharmacology and signaling of prostaglandin receptors: multiple roles in inflammation and immune modulation.Pharmacology & Therapeutics, 103(2):147–166, 2004. doi: 10.1016/j.pharmthera.2004.06.003

    work page doi:10.1016/j.pharmthera.2004.06.003 2004

  41. [41]

    Identification of CD72 as a lymphocyte receptor for the class IV semaphorin CD100: a novel mechanism for reg- ulating B cell signaling.Immunity, 13(5):621–631, 2000

    Atsushi Kumanogoh, Chiharu Watanabe, Inhwa Lee, et al. Identification of CD72 as a lymphocyte receptor for the class IV semaphorin CD100: a novel mechanism for reg- ulating B cell signaling.Immunity, 13(5):621–631, 2000. doi: 10.1016/S1074-7613(00) 00062-5. 111

    work page doi:10.1016/s1074-7613(00 2000