Neutrophil-derived catecholamines mediate negative stress effects on bone

Abstract

Mental traumatization is associated with long-bone growth retardation, osteoporosis and increased fracture risk. We revealed earlier that mental trauma disturbs cartilage-to-bone transition during bone growth and repair in mice. Trauma increased tyrosine hydroxylase-expressing neutrophils in bone marrow and fracture callus. Here we show that tyrosine hydroxylase expression in the fracture hematoma of patients correlates positively with acknowledged stress, depression, and pain scores as well as individual ratings of healing-impairment and pain-perception post-fracture. Moreover, mice lacking tyrosine hydroxylase in myeloid cells are protected from chronic psychosocial stress-induced disturbance of bone growth and healing. Chondrocyte-specific β2-adrenoceptor-deficient mice are also protected from stress-induced bone growth retardation. In summary, our preclinical data identify locally secreted catecholamines in concert with β2-adrenoceptor signalling in chondrocytes as mediators of negative stress effects on bone growth and repair. Given our clinical data, these mechanistic insights seem to be of strong translational relevance.

Fig. 1. Association of different aspects of psychosomatic health with tyrosine hydroxylase (TH) expression in the human fracture hematoma and outcome after upper ankle fracture.

a Correlational analyses of TH immunoreactivity in the fracture hematoma of human patients suffering upper ankle fracture with different aspects of psychosomatic health assessed by established questionnaires (somatic symptoms: Som, PHQ15; anxiety: Anx, GAD7; depression: Depri, PHQ9; stress load: Stress, PHQS; social functioning: socfunct, SF36; pain: SF36; childhood adversity: CTQ sum) on the days after the surgery. b Representative images (scale bars: 50 µm) visualizing TH immunofluorescent stainings in the fracture hematoma of a patient with low (0; left panel) and high (5; right panel) stress scores. n_{human patients} = 20–21. c Representative images (scale bars: 50 µm) visualizing TH and CD16 immunofluorescent double-stainings in the fracture hematoma of a patient with low (0; left panel) and high (5; right panel) stress scores. Fluorescent channels were adjusted equally for both groups. d Correlational analyses of TH immunoreactivity in the fracture hematoma at the day of surgery with the degree of mobility limitation, as well as the healing process impairment documented both at 3, 6, 9 and 12 months post surgery using visual analog scales (VASs). e Correlational analyses of different aspects of psychosomatic health assessed by established questionnaires (somatic symptoms: Som, PHQ15; anxiety: Anx, GAD7; depression: Depri, PHQ9; stress load: Stress, PHQS; social functioning: socfunct, SF36; pain: SF36; childhood adversity: CTQ sum) on the days following surgery with pain assessed by SF36 questionnaire at 3, 6, 9 and 12 months post surgery. Spearman correlation analyses, *P ≤ 0.05; **P ≤ 0.01, ***P ≤ 0.001 as significant correlation. n_{human patients} = 16 for the follow-up time points. Source data, exact n-numbers, exact p-values and used statistical tests per panel are provided in the Source Data file.

Fig. 2. Effects of different durations of chronic subordinate colony housing (CSC) on bone homeostasis in wild-type (WT) mice.

a Experimental timeline for WT mice exposed to 7 d of single-housed control (SHC) conditions or CSC (partly created with BioRender.com). Mice were single-housed for one week before the start of the CSC paradigm on Day 1 and euthanized on Day 8 of CSC (=7 d CSC exposure). b Trabecular thickness, c trabecular number, d trabecular tissue mineral density (Tb. TMD), e trabecular separation, f cortical TMD, g growth plate thickness, h cortical thickness, i bone volume/tissue volume (BV/TV), j femur length, k tibia length and l representative images of tyrosine hydroxylase (TH) immunofluorescent staining (scale bar: 100 µm) in the bone marrow (BM) of unfractured femora in WT mice euthanized on Day 8 of SHC/CSC. n = 7–8. m Experimental timeline for WT mice exposed to 19 d of SHC/CSC conditions (partly created with BioRender.com). Mice were single-housed for one week before the start of the CSC paradigm on Day 1. The aggressor mouse was changed on Days 8 and 15 and the experimental mice were euthanized on Day 20. Number of BM n TH⁺Ly6G⁺ cells, o TH⁺CD8⁺ cells, p TH⁺CD4⁺ cells, and q TH⁺F4/80⁺ cells assessed using flow cytometry in WT mice euthanized on Day 20 of SHC/CSC exposure. n = 7–8. r Experimental timeline for WT mice exposed to 19 d of SHC/CSCconditions followed by 21 d of single housing (SH) (partly created with BioRender.com). Mice were single-housed for one week before the start of the CSC paradigm on Day 1. The aggressor mouse was changed on Days 8 and 15, and on Day 20 all experimental mice were single-housed for 21 consecutive days before being euthanized on Day 22 of SH. s Trabecular thickness, t trabecular number, u Tb. TMD, v trabecular separation, w BV/TV, x growth plate thickness, y femur length in WT mice exposed to 19 d of SHC/CSC + 21 d SH. n = 6–8. Data are presented as mean + SEM including individual values. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 versus SHC condition. n.s. not significant. Source data, exact n-numbers, exact p values and used statistical tests per panel are provided in the Source Data file.

Fig. 3. Effects of 19 d of chronic subordinate colony housing (CSC) on bone homeostasis in TH^flox/flox/CD11b-Cre (TH^flox/Cre) mice.

a Experimental timeline for TH^flox/Cre⁻ and Cre⁺ mice exposed to 19 d of single-housed control (SHC)/CSC conditions (partly created with BioRender.com licensed to SOR). Mice were single-housed for one week before the start of the CSC paradigm on Day 1. The aggressor mouse was changed on Days 8 and 15 and the experimental mice were tested for anxiety-like behavior in the open field/novel object (OF/NO) test on Day 19 and euthanized on Day 20. b Femur length, c tibia length, d representative images of growth plates (scale bars: 100 µm), e growth plate thickness, f representative 3D images of analyzed volume of interests (VOIs), g trabecular tissue mineral density (Tb. TMD), h trabecular thickness and i bone volume to tissue volume ratio (BV/TV) of unfractured femora of TH^flox/Cre^- and Cre⁺ SHC/ CSC mice. n = 6–8. Data are presented as mean + SEM including individual values. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 versus respective SHC condition; ^##P ≤ 0.01, ^###P ≤ 0.001 versus respective TH^flox/Cre⁻ group. n.s. not significant. Source data, exact n-numbers, exact p values and used statistical tests per panel are provided in the Source Data file.

Fig. 4. Effects of 19 d of chronic subordinate colony housing (CSC) on fracture healing in TH^flox/flox/CD11b-Cre (TH^flox/Cre) mice.

a Experimental timeline for TH^flox/Cre^- and Cre⁺ exposed to 19 d of single-housed control (SHC)/CSC conditions and femur osteotomy on Day 20 (partly created with BioRender.com licensed to SOR). Mice were single housed for one week before the start of the CSC paradigm on Day 1. The aggressor mouse was changed on Days 8 and 15 and the experimental mice were tested for anxiety-like behavior in the open field/novel object (OF/NO) test on Day 19. The experimental mice underwent femur osteotomy on Day 20 and were euthanized 1 d, 10 d or 21 d post surgery. Number of CD11b⁺Ly6G⁺ neutrophils in b the fracture hematoma and c the bone marrow and d plasma CXCL1 concentration of TH^flox/Cre^- and Cre⁺ mice exposed to CSC/SHC conditions and femur osteotomy 1 d post fracture. e percentage of Runx2⁺ hypertrophic chondrocytes (HTCs) in the fracture callus, f representative immunohistochemical images of Runx2 immunostainings in fracture callus sections (scale bars: 50 µm), g number of osteoclasts (N.Oc) per bone perimeter (B.Pm) and h osteoblast surface per bone surface (Oc.S/BS) in the fracture callus, i representative images of tartrate-resistant acid phosphatase (TRAP)-stained fracture callus sections (scale bars: 50 µm), j Col10a1⁺ cells and k vascularization in the fracture callus of TH^flox/Cre^- and Cre⁺ SHC/CSC mice 10 d post-fracture. l Tissue mineral density (TMD), m BV/TV, n relative cartilage area in the fracture callus and o representative images of Safranin-O-stained (scale bars: 1000 µm) fracture callus sections of TH^flox/Cre⁻ and Cre⁺ SHC/CSC mice 21 d post fracture. n = 4–8. Data are presented as mean + SEM including individual values. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 versus respective SHC condition; ^#P ≤ 0.05 versus respective TH^flox/Cre⁻ group. n.s. not significant. Source data, exact n-numbers, exact p-values and used statistical tests per panel are provided in the Source Data file.

Fig. 5. Effect of catecholamines (CAs) or conditioned medium from single-housed control (SHC) and chronic subordinate colony housing (CSC) mice and different adrenoceptor (AR) antagonists on in vitro chondrocyte-to-osteoblast transdifferentiation.

a Experimental setup of testing different synthetic CAs on in vitro chondrocyte-to-osteoblast transdifferentiation. b Heatmap visualization of the effects of norepinephrine (NE), epinephrine (EPI) and dopamine (DOP) on the expression of pluripotency and osteogenic marker genes in transdifferentiating ATDC5 cells versus. respective control conditions. n = 4. c Experimental setup of testing conditioned medium of CD11b⁺ myeloid bone marrow cells isolated from TH^flox/flox/CD11b-Cre^- and Cre⁺ SHC/CSC mice in combination with antagonists for β-Adrenoceptor (β-AR), α-Adrenoceptor (α-AR), α₁-AR, α₂-AR, class 1 dopaminergic receptor (D_1/5) and class 2 dopaminergic receptor (D_2/3/4) signaling. Expression of d, e pluripotency (d: Sex determining region Y box (Sox)2; e: Nanog) and f–h osteogenic marker genes (f: Core-binding factor alpha (Cbfa)1; g: Sp7; h: Alkaline phosphatase (Alpl)) in transdifferentiating ATDC5 cells after 24 h of osteogenic differentiation in conditioned medium with receptor antagonists. n = 6–9. Data are presented as mean + SEM. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 versus respective SHC condition; ^#P ≤ 0.05, ^##P ≤ 0.01 versus respective control group. n.s. not significant. Source data, exact n-numbers, exact p values and used statistical tests per panel are provided in the Source Data file.

Fig. 6. Effects of 19 d of chronic subordinate colony housing (CSC) on bone homeostasis in Adrb2^flox/flox/Col2a1-Cre (Adrb2^flox/Cre) mice.

a Experimental timeline for Adrb2^flox/Cre⁻ and Cre⁺ exposed to 19 d of single-housed control (SHC)/CSC conditions (partly created with BioRender.com licensed to SOR). Mice were single-housed for one week before the start of the CSC paradigm on Day 1. The aggressor mouse was changed on Days 8 and 15 and the experimental mice were tested for anxiety-like behavior in the open field/novel object (OF/NO) test on Day 19 and euthanized on Day 20. b Relative adrenal weight, c femur length, d tibia length, e trabecular thickness, f trabecular tissue mineral density (Tb. TMD), g bone volume to tissue volume ratio (BV/TV) of Adrb2^flox/Cre^- and Cre⁺ SHC/ CSC mice. n = 7–8. Data are presented as mean + SEM including individual values. *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001 versus respective SHC condition; ^#P ≤ 0.05, ^###P ≤ 0.001 versus respective Adrb2^flox/Cre⁻ group. n.s. not significant. Source data, exact n-numbers, exact p values and used statistical tests per panel are provided in the Source Data file.

Fig. 7. Graphical abstract.

While mental trauma in general promotes bone marrow (BM) myelopoiesis, tyrosine hydroxylase (TH) expression and, consequently, the capacity to produce/ secrete catecholamines (CAs) is specifically facilitated in neutrophil granulocytes (NGRs). Neutrophil-derived CAs locally in the BM activate α/β-adrenoceptors (ARs) and dopaminergic receptors (DRs) on chondrocytes (CCs) and compromise their transdifferentiation into osteoblasts and, thus, bone metabolism. Neutrophil-derived CAs in an autocrine manner further promote their own BM emigration and, in case of a fracture, facilitate their own immigration into the fracture hematoma, likely in a paracrine manner by increasing CXCL1 release from hematoma mast cells and macrophages (MAs) which are two main CXCL1 producing cell types. In the fracture hematoma (FH), neutrophil-derived CAs again activate α/β-ARs and DA receptors on CCs and, consequently, compromise their transdifferentiation into osteoblasts (OBs) and, thus, adequate bone repair. (partly created with BioRender.com licensed to SOR). GP growth plate, TMD tissue mineral density, BV/TV bone volume/ tissue volume ratio, TB trabecular, GP growth plate, MPC myeloid progenitor cell, MO monocyte, RBA relative bone area, RCA relative cartilage area, RSTA relative soft tissue area, Oc.S/BS osteoclast surface/bone surface, N.Oc/ B.Pm number of osteoclasts/ bone perimeter.