Patient-specific Immune States before Surgery Are Strong Correlates of Surgical Recovery

Abstract

Background: Recovery after surgery is highly variable. Risk-stratifying patients based on their predicted recovery profile will afford individualized perioperative management strategies. Recently, application of mass cytometry in patients undergoing hip arthroplasty revealed strong immune correlates of surgical recovery in blood samples collected shortly after surgery. However, the ability to interrogate a patient's immune state before surgery and predict recovery is highly desirable in perioperative medicine.

Methods: To evaluate a patient's presurgical immune state, cell-type-specific intracellular signaling responses to ex vivo ligands (lipopolysaccharide, interleukin [IL]-6, IL-10, and IL-2/granulocyte macrophage colony-stimulating factor) were quantified by mass cytometry in presurgical blood samples. Selected ligands modulate signaling processes perturbed by surgery. Twenty-three cell surface and 11 intracellular markers were used for the phenotypic and functional characterization of major immune cell subsets. Evoked immune responses were regressed against patient-centered outcomes, contributing to protracted recovery including functional impairment, postoperative pain, and fatigue.

Results: Evoked signaling responses varied significantly and defined patient-specific presurgical immune states. Eighteen signaling responses correlated significantly with surgical recovery parameters (|R| = 0.37 to 0.70; false discovery rate < 0.01). Signaling responses downstream of the toll-like receptor 4 in cluster of differentiation (CD) 14 monocytes were particularly strong correlates, accounting for 50% of observed variance. Immune correlates identified in presurgical blood samples mirrored correlates identified in postsurgical blood samples.

Conclusions: Convergent findings in pre- and postsurgical analyses provide validation of reported immune correlates and suggest a critical role of the toll-like receptor 4 signaling pathway in monocytes for the clinical recovery process. The comprehensive assessment of patients' preoperative immune state is promising for predicting important recovery parameters and may lead to clinical tests using standard flow cytometry.

Conflict of interest statement

Conflicts of Interest: G.P.N. has a personal financial interest in Fluidigm (South San Francisco, CA, USA), the manufacturer of the mass cytometer used in this manuscript. All other authors have no conflict of interest.

Figures

Figure 1. Flow-chart summarizing the experimental approach

Whole blood was obtained one hour before surgery from patients undergoing primary hip arthroplasty (PANEL 1) Separate whole blood aliquots were stimulated ex vivo with extracellular ligands (LPS, lipopolysaccharide; IL-6, interleukin 6; IL-10, interleukin 10; or a combination of IL-2 and GMCSF, interleukin 2 and granulocyte monocyte colony-stimulating factor) or left untreated (U) (PANEL 2a). Using mass cytometry, the expression of 23 cell surface markers and the phosphorylation states of 11 intracellular signaling proteins were measured in single cells from blood samples (PANEL 2b). Unsupervised hierarchical clustering and manual gating strategies were applied to visualize and quantify patient-specific signaling responses in immune cell subsets spanning the entire immune system. Shown is a visual representation of a cluster hierarchy plot (PANAL 3). Signaling responses that correlated significantly with clinical recovery parameters were identified by significance analysis of microarrays (PANAL 4). Arrows indicate results for the two hypothetical patients (PANAL 1).

Figure 2. Evoked signaling responses in pre-surgical whole blood samples

(A) The cluster hierarchy plot represents immune cell subsets across the entire peripheral immune system. The cluster hierarchy plot is annotated and contoured based on the expression of canonical lineage markers indicative of major immune cell subsets. As an example, the monocyte branch is highlighted in yellow based on the expression of the cell surface marker CD14. (B) STAT3 (phosphorylated signal transducer and activator of transcription 3) signaling response to IL-6 (interleukin 6): Individual clusters are colored based on the median signal induction of pSTAT3. A significant response was evoked in CD14+ monocytes, dendritic cells, CD4+ T cells, and a small fraction of CD8+ T cells. (C) NFκB (phosphorylated nuclear factor kappa-light-chain-enhancer of activated B cells) signaling response to LPS (lipopolysaccharide): Individual clusters are colored based on the median signal induction of pNFκB. A significant response was evoked in CD14+ monocytes and dendritic cells. (D) CREB (phosphorylated cAMP response element-binding protein) signaling response to LPS: Individual clusters are colored based on the median signal induction of pCREB. A significant response was evoked in CD14+ monocytes and dendritic cells.

Figure 3. Variability of evoked signaling responses in pre-surgical whole blood samples

Cell-type specific signaling responses were quantified in manually gated cell subsets. Values reflect the arcsinh ratio, i.e., the difference between arcsinh transformed values (inverse hyperbolic sine) obtained in unstimulated and stimulated conditions. Asterisks (*) denote statistical significance based on the Bonferroni-adjusted p value (p=8.92×10−4). (A) The pSTAT3 (phosphorylated signal transducer and activator of transcription 3) signal in CD14+ monocytes increased significantly in response to IL-6 (interleukin6; p=2.40×10−14) and IL-10 (interleukin 10; p=2.27×10−13), but not in response to LPS (lipopolysaccharide) or the combination of IL-2 (interleukin 2) and GMCSF (granulocyte monocyte colony-stimulating factor). The signaling response varied 3.5-fold among patients. (B) The pCREB (phosphorylated cAMP response element-binding protein) signal in CD14+ monocytes increased in response to LPS (p=1.22×10−15), the combination of IL-2 and GMCSF (p=4.80×10−14), and IL-6 (p=3.03×10−12), but not in response to IL-10. The signaling responses varied by 3.5- to 5.5-fold among patients. (C) The pNFκB (phosphorylated nuclear factor kappa-light-chain-enhancer of activated B cells) signal in CD14+ monocytes increased in response to LPS (p=4.95×10−13), the combination of IL-2 and GMCSF (p=1.65×10−9), and IL-6 (p=1.87×10−8), but not in response to IL-10. The signaling responses varied 5-fold among patients.

Figure 4. Patient-specific evoked signaling responses in myeloid cell subsets

Patient-specific immune states before surgery were characterized along three axes including eight cell types, eleven signaling proteins, and the activity of these signaling proteins in response to four different extracellular ligands. Signaling activity is indicated by the color-coded arcsinh ratio, i.e., the difference between arcsinh transformed values (inverse hyperbolic sine) obtained in unstimulated and stimulated conditions. A total of 352 parameters described each patient’s immune state. Shown are the three-dimensional heat maps characterizing and differentiating the immune states of two exemplary patients. Abbreviations: IL-10, interleukin 10; IL-2, interleukin 2; GMCSF, granulocyte monocyte colony-stimulating factor; IL-6, interleukin 6; LPS, lipopolysaccharide; pSTAT1/3/5, phosphorylated signal transducer and activator of transcription 1/3/5; pCREB, phosphorylated cAMP response element-binding protein; pNFkB, phosphorylated nuclear factor kappa-light-chain-enhancer of activated B cells; pPLCg2, phosphorylated phospholipase C gamma 2; pP38, phosphorylated P38 map kinase; pMAPKAPK2, phosphorylated mitogen activated protein kinase-activated protein kinase-2; pERK1/2, phosphorylated extracellular regulated kinase 1/2; prpS6, phosphorylated ribosomal protein S6; pp90RSK, phosphorylated ribosomal S6 kinase; pDCs, plasmacytoid dendritic cells; cDCs, classical dendritic cells.

Figure 5. Immune correlates of functional impairment of the hip

(A) P-values are plotted as the negative of the log base 10 for correlations between ligand-evoked intracellular signaling responses and clinical recovery parameters. The lowest 50 p-values are shown. The pMAPKAPK2 (phosphorylated mitogen activated protein kinase-activated protein kinase-2) response to LPS (lipopolysaccharide) in CD14+ monocytes (MCs) and classical dendritic cells (cDCs) passed the Bonferroni correction as depicted in red. (B) pMAPKAPK2 response to LPS in cDCs strongly correlated with functional recovery of the hip (R=0.70, p=9.17×10−5). (C) pMAPKAPK2 response to LPS in CD14+ MCs strongly correlated with functional recovery of the hip (R=0.69, p=1.39×10−4). Shown in red are the regression lines with adjacent 95% confidence intervals.

Figure 6. A model summarizing the TLR4-dependent signaling responses implicated in surgical recovery

In response to tissue damage, damage-associated molecular patterns (DAMPs) molecules including alarmins such as HMGB1 (high-mobility group protein B1) and HSPs (heat shock proteins) bind to and activate the TLR4 receptor (toll-like receptor 4) on the surface of innate immune cells. The TLR4 signaling pathway is also induced by binding of LPS (lipopolysaccharide) to the TLR4 receptor and to co-receptors CD14 (cluster of differentiation 14) and MD2 (myeloid differentiation factor 2). The TLR4 receptor engages a complex signaling cascade (only partially represented) leading to the activation of the p38 (phosphorylated P38 mitogen-activated protein kinase) and ERK/MAP (extracellular regulated kinase and mitogen activated protein kinase) and the NFκB (phosphorylated nuclear factor kappa-light-chain-enhancer of activated B cells) signaling pathways. Signaling proteins that changed significantly in response to surgery are indicated by blue circles. A subset of signaling proteins modulated after surgery (MAPKAPK2, mitogen-activated protein kinase-activated protein kinase 2; rpS6, ribosomal protein S6; CREB, cAMP response element-binding protein) overlapped with pre-surgical signaling responses that correlated with surgical recovery parameters and are indicated by red circles. Other abbreviations: MEKKs, mitogen activated protein kinase kinase kinase; IKK, IκB kinase; IκB, inhibitor of NFκB.