Preventing acute asthmatic symptoms by targeting a neuronal mechanism involving carotid body lysophosphatidic acid receptors

Abstract

Asthma accounts for 380,000 deaths a year. Carotid body denervation has been shown to have a profound effect on airway hyper-responsiveness in animal models but a mechanistic explanation is lacking. Here we demonstrate, using a rat model of asthma (OVA-sensitized), that carotid body activation during airborne allergic provocation is caused by systemic release of lysophosphatidic acid (LPA). Carotid body activation by LPA involves TRPV1 and LPA-specific receptors, and induces parasympathetic (vagal) activity. We demonstrate that this activation is sufficient to cause acute bronchoconstriction. Moreover, we show that prophylactic administration of TRPV1 (AMG9810) and LPA (BrP-LPA) receptor antagonists prevents bradykinin-induced asthmatic bronchoconstriction and, if administered following allergen exposure, reduces the associated respiratory distress. Our discovery provides mechanistic insight into the critical roles of carotid body LPA receptors in allergen-induced respiratory distress and suggests alternate treatment options for asthma.

Fig. 1

LPA stimulates the carotid body. a RT-PCR: LPAr (L) 1, 3, 4, and 6 in carotid body (left), LPAr 3 and 6 in petrosal ganglia (center), and LPAr 1, 3, 4, and 6 in superior cervical ganglia (right) in reference to a housekeeping gene, hypoxyribosyltransferase (H; cDNA ladder, Ld, was used to identify tissues used for each gel, 100 bp marker indicated). b Calcium imaging with Fura-2 loaded carotid body glomus cells (12 cells on 5 cover slips) reveals an increase of intracellular calcium in response to 5 µM LPA (18:1 unless otherwise stated) and 20 mM K⁺. F_2,35 (one-way ANOVA) = 18.220, p < 0.001; Holm–Šidák post hoc: difference from baseline, ***p < 0.01. c En bloc perfused carotid body preparation used to record chemosensory afferents in the carotid sinus nerve (CSN). d LPA 18:1 causes an increase in CSN activity. e The dose-dependent relationship differed slightly between several species of LPA found in blood (16:0, 18:1, and 18:2); F_4,53 (two-way ANOVA: species × dose) = 3.380, p = 0.021. Holm–Šidák post hoc: all species 5 and 10 µM different from 2.5 µM, ***p < 0.001. f Effects of TRPV1 and LPAr antagonists on CSN response to LPA. No blockade (black trace), TRPV1 blockade (AMG9810 10 µM, teal trace, or LPAr blockade (BrP-LPA 1.5 µM, mauve trace; Ki16425 5 µM, fuscia trace). g Summary data. Responses 2.5 µM LPA: F_3,23 (one-way ANOVA) = 5.031, **p = 0.01. Holm–Šidák post hoc: LPA different from BrP-LPA (p = 0.026) and Ki16425 (p = 0.018). Responses 5 µM: F_3,23 (one-way ANOVA) = 24.547, ***p < 0.001. Holm–Šidák post hoc: LPA different from BrP-LPA, Ki16425, and AMG9810 (p < 0.001); AMG9810 different from Brp-LPA (p = 0.018) and Ki16425 (p = 0.044); Responses 10 µM: F_3,23 (one-way ANOVA) = 14.231, ***p < 0.001. Holm–Šidák post hoc: LPA different from BrP-LPA (p < 0.001); Ki16425 (p < 0.001); AMG9810 (p = 0.022); AMG9810 different from BrP-LPA (p = 0.022). h Summary data of 5 µM LPA (blue), with TRPV1 blockade (AMG9810, 10 µM; teal), LPAr blockade (BrP-LPA, 1.5 µM; mauve; Ki16425, 5 µM; fuscia), or LPAr + TRPV1 blockade (AMG9810 + BrPLPA red; AMG9810 + Ki16425 orange), F_5,35 (one-way ANOVA) = 26.164, p < 0.001. Holm–Šidák post hoc: AMG9810 + LPA significantly different from Ki16425 + AMG9810 + LPA, p = 0.005; LPA significantly different from all other groups, ***p < 0.001. All data are presented as mean ± sem

Fig. 2

LPA stimulation of the carotid body increases efferent vagal activity. a The dual perfused in situ preparation used to record phrenic and vagal activity in response to specific carotid body stimulation. b Phrenic (upper trace) and vagal (lower trace) activity with the carotid body intact under brainstem hypocapnia (black trace) causing cessation of phrenic firing (neural apnea) and vagal quiescence. Activity is restored following carotid body stimulation with 5 µM LPA (blue arrow/trace), with restoration of vagal activity. c In denervated preparations the stimulatory effect of LPA on vagal and phrenic activity is absent. d Expanded sections of traces from b indicated by the red boxes, illustrating phrenic and vagal activity prior to brainstem hypocapnia induced apnea, apnea and following LPA injections. e, f Summary data during hypocapnia in intact (solid) and carotid body (CB) denervated (crossed) preparations for phrenic frequency (Phrenic burst min⁻¹; Mann Whitney: U = 0, p = 0.002, ** median ± range) and vagal total activity (normalized to baseline normoxic activity, normalized units, independent t test: t₁₀ = 4.008, p = 0.0025, ** mean ± sem)

Fig. 3

In vivo carotid body-mediated bronchoconstriction. a–c The anesthetized in vivo preparation used to measure airway mechanics in carotid sinus nerve intact (solid, n = 6) or denervated (crossed, n = 6) preparations, in response to jugular vein injection of lysophosphatidic acid (LPA 5 µM, blue) or sodium cyanide (NaCN, orange; an independent test of carotid body function). Baseline lung resistance (R_L) prior to LPA or NaCN injection: intact = 0.31 ± 0.01 cmH₂O s ml⁻¹; denervated = 0.28 ± 0.02 cmH₂O s ml⁻¹ (independent t-test: t₁₀ = 1.668, p = 0.126). Baseline lung elastance (E_L) prior to LPA or NaCN injection: intact = 6.8 ± 0.3 cmH₂Oml⁻¹; denervated = 7.6 ± 0.4 cmH₂O ml⁻¹ (independent t-test: t₁₀ = 1.433, p = 0.182). b R_L in denervated compared to intact CB preparations—LPA-independent t-test: t₁₀ = 4.933, ***p < 0.001; NaCN-independent t-test: t₁₀ = 2.711, *p = 0.022. c E_L, in denervated compared to intact CB preparations—LPA-independent t-test: t₁₀ = 6.772, ***p < 0.00001; NaCN-independent t-test: t₁₀ = 4.202, **p < 0.01. In b and c, R_L and E_L are normalized to saline (baseline) measurements and presented as mean ± sem

Fig. 4

Ovalbumin-sensitized rats demonstrate increased airway remodeling, inflammatory markers, airway resistance and plasma LPA. a OVA-sensitization protocol (see Methods, OVA Cohort 1). b–e Typical asthmatic airway disease occurred in the OVA-sensitized rats (right c, e) compared to the naive group (left b, d) as indicated by the heightened presence of goblet cells (appearing pink with Periodic acid-Schiffs reagent; b, c) and thickening of airway smooth muscle (appearing brown when immuno-stained for smooth muscle actin; d, e). f qPCR of IL4, eotaxin (CCL11) and house-keeping gene β-actin for OVA (red) and naive (open) rats with calculated change in gene expression magnitude (2^−∆CTT). CT: cycles threshold. OVA vs naive CT for IL4-independent t-test, t₁₁ = 73.271, ***p < 0.00001; CCL11-independent t-test: t₁₂ = 6.369, ***p < 0.0001; and β-actin-independent t-test: t₁₂ = 0.403, p = 0.694. g Bradykinin increases R_L in OVA animals only (red; F_1,51 (two-way RM ANOVA) = 69.224, p < 0.001). Holm–Šidák post hoc: bradykinin increases R_L in OVA animals, ***p < 0.001; but not naive rats (p = 0.40). Prior to bradykinin R_L in naive (open) and OVA rats is not significantly different (p = 0.08), following bradykinin R_L is greater in OVA compared to naive rats (***p < 0.001). h Bradykinin increases LPA in OVA animals only (F_1,55 (two-way RM ANOVA) = 6.19, p = 0.02). Holm–Šidák post hoc: bradykinin increases LPA in OVA animals, ***p < 0.001; bradykinin has no effect on LPA in naive rats (p = 0.21), LPA in naive and OVA rats prior to bradykinin are not significantly different (p = 0.58); but LPA in OVA rats is greater than naive rats following bradykinin (**p = 0.01) (g, h See Methods, OVA Cohort 1). All data are presented as mean ± sem

Fig. 5

Plasma from ovalbumin-sensitized rats increases carotid body activity in LPA receptor-dependent manner. a Carotid sinus nerve activity from a naive en bloc carotid body preparation in response to plasma from naive (gray trace) and OVA-sensitized (red trace) rats. Application of dual LPAr and TRPV1 blockade (AMG9810, 10 µM + BrP-LPA, 1.5 µM) is indicated by the mauve arrow and subsequent trace. b Summary data of the effect of naive (open) and OVA (red) plasma as well as subsequent dual blockade (mauve). F_2,17 (one-way ANOVA) = 40.193, ***p < 0.001, mean ± sem. Holm–Šidák post hoc: OVA vs naive p < 0.001; OVA vs blockade p < 0.001

Fig. 6

Bradykinin-induced bronchoconstriction in ovalbumin-sensitized rats is dependent on the carotid body and LPA signaling. a OVA-sensitization protocol (see Methods, OVA Cohort 2) to test lung–carotid body–lung pathway. b OVA-sensitized and naive rats were exposed to nebulized saline (baseline) and three consecutive nebulizations of 0.4 mg bradykinin at 1 (solid), 10 (hatched), and 20 (crossed) min while measuring R_L. Bradykinin had group specific effects: See Methods, OVA Cohort 2; F_14,143 (two-way RM ANOVA: time × group) = 4.035, p < 0.001. Holm–Šidák post hoc: bradykinin caused a marked increase in R_L in OVA-sensitized (red; **p < 0.01) but not naive rats (white; p > 0.3) rats; carotid body (CB) denervated (blue), vagi (VaG) denervated (purple), TRPV1 blockade (AMG9810, orange), LPAr blockade (BrP-LPA, yellow and Ki16425, fuscia) and dual TRPV1 and LPAr blockade (AMG9810 + BrP-LPA, green), abolished the effects of bradykinin compared to OVA (Holm–Šidák post hoc: *p < 0.05; ***p < 0.001). c Chronic CB denervation (blue; but not sham-treatment, red) also abolished bradykinin-induced bronchoconstriction in OVA-sensitized animals (see Methods OVA Cohort 3; F_2,32 (two-way RM ANOVA time × group) = 5.418, p = 0.014; Holm–Šidák post hoc: Intact greater than denervated, 10 and 20 min greater than 1 min in Intact, **p < 0.01). d Bronchoconstriction in response to capsaicin does not involve the carotid body (see Methods OVA Cohort 4). R_L was measured 2 and 10 min after single capsaicin nebulization in OVA (red) and naive (gray) rats who had intact (solid) or resected vagi (hatched). All rats had denervated carotid bodies yet all, including naive animals, exhibited bronchoconstriction in response to capsaicin by 2 min (F_3,47 (two-way RM ANOVA treatment × group) = 8.435, p < 0.001). Holm–Šidák post hoc: capsaicin-induced bronchoconstriction was abolished by vagotomy: ***p < 0.001; different between 2 min and 10 min ***p < 0.001; **p = 0.002. In b–d, R_L is normalized to that during initial saline nebulization. All data are presented as mean ± sem

Fig. 7

LPAr + TRPV1 blockade abates acute asthmatic respiratory distress in conscious rats. a OVA-sensitization protocol (see Methods, OVA Cohort 6). b, c Inspiratory:expiratory time decreased (Ti:Te; F_35,431(two-way RM ANOVA time × group) = 8.577, p < 0.001, Holm–Šidák post hoc: p < 0.05, * different between groups at indicated time), expiratory time increased (Te; F_35,431 (two-way RM ANOVA time × group) = 3.948, p < 0.001, Holm–Šidák post hoc: p < 0.05, * different between groups at indicated time) in response to acute OVA provocation following OVA sensitization, confirming these parameters as indices of acute asthmatic respiratory distress in conscious animals. d 21-Day sensitization protocol to test dual LPAr + TRPV1 blockade on respiratory distress (see Methods OVA Cohort 7); LPAr + TRPV1 blockade (T), vehicle (V), or saline (S; randomized) were delivered 20 min after the onset of OVA (as indicated by the arrows). e Decrease in Ti:Te and f increase in Te caused by allergen provocation are rescued by dual blockade (blue). Ti:Te: F_70,1293 (two-way RM ANOVA group × time = 3.169, p < 0.001; Te: F_{35, 385} (two-way RM ANOVA time) = 10.590, p < 0.001, F_{2, 35} (two-way RM ANOVA group) = 7.393, p = 0.004). Holm–Šidák post hoc: dual block is significantly different from OVA-sensitized saline injected (red circles)* and vehicle injected (pale red triangles)^ groups, at indicated time points, p < 0.05. g The peak Ti:Te responses recorded 120 min after OVA exposure on day 21 in animals never having received dual blockade (red—from b), having dual blockade on day 21 and recorded on day 21 (acute treatment, light blue, from e), or having dual blockade on day 18 and recorded on day 21 (Long term treatment, dark blue, from e; F_2,15 (one-way ANOVA group) = 45.805, p < 0.001). h The peak Te (120 min) response recorded on day 21 following OVA exposure; groups as per g. Dual antagonist injection on days 18 or 21 reduced respiratory indices of acute bronchoconstriction; and remarkably, dual antagonist injection on day 18 also had beneficial effects three days later, on day 21, without a subsequent dual antagonist injection (F_2,15 one-way ANOVA group = 25.906, p < 0.001). Holm–Šidák post hoc: difference between indicated groups ***p < 0.001. Data are presented as mean ± sem