A Molecular Biophysical Approach to Diclofenac Topical Gastrointestinal Damage

Abstract

Diclofenac (DCF), the most widely consumed non-steroidal anti-inflammatory drug (NSAID) worldwide, is associated with adverse typical effects, including gastrointestinal (GI) complications. The present study aims to better understand the topical toxicity induced by DCF using membrane models that mimic the physiological, biophysical, and chemical environments of GI mucosa segments. For this purpose, phospholipidic model systems that mimic the GI protective lining and lipid models of the inner mitochondrial membrane were used together with a wide set of techniques: derivative spectrophotometry to evaluate drug distribution at the membrane; steady-state and time-resolved fluorescence to predict drug location at the membrane; fluorescence anisotropy, differential scanning calorimetry (DSC), dynamic light scattering (DLS), and calcein leakage studies to evaluate the drug-induced disturbance on membrane microviscosity and permeability; and small- and wide-angle X-ray scattering studies (SAXS and WAXS, respectively), to evaluate the effects of DCF at the membrane structure. Results demonstrated that DCF interacts chemically with the phospholipids of the GI protective barrier in a pH-dependent manner and confirmed the DCF location at the lipid headgroup region, as well as DCF's higher distribution at mitochondrial membrane contact points where the impairment of biophysical properties is consistent with the uncoupling effects reported for this drug.

Keywords: DLS; DSC; NSAIDs; SAXS; WAXS; derivative spectrophotometry; diclofenac; gastrointestinal topical toxicity; steady-state and time-resolved fluorescence; steady-state anisotropy.

Figure 1

(A) Schematic representation of the gastrointestinal (GI) epithelium containing extracellular and cellular phospholipid linings rich in phosphatidylcholine (PC) in the gel phase (L_β′), which constitute a protective barrier of the GI mucosa against the attack of luminal acid (H⁺). The mitochondria outer and inner membranes (OMM and IMM) containing typical crista and contact points rich in bilayer prone phospholipids (e.g., PC) that produce lamellar phases (L_α) and non-lamellar prone lipids (e.g., Cardiolipin (CL) and phosphatidylethanolamine (PE)) that produce inverted hexagonal phases (H_II) are also represented. (B) Schematic representation of GI topical damage depicting the disruptive effect of non-steroidal anti-inflammatory drugs (NSAIDs) on the GI protective phospholipid lining. On the right are represented the possible mechanisms by which NSAIDs cause mitochondrial toxicity: inhibition of respiratory chain; inhibition of ATP synthase and translocase; blockage of cation transporters; and membrane permeabilization with the formation of membrane permeability transition pores (MPT).

Figure 2

(A) DCF (40 μM) absorption spectra in aqueous buffered phase pH = 5.0 (red spectrum) and after incubation with increasing concentrations of LUVs of DMPC (0–1000 μM) prepared in the same buffer pH = 5.0 (black spectra); (B) third derivative of the absorbance spectra where it is possible to observe isosbestic points (arrows) and a shift of λ values at the peaks of the derivative spectra (Δλ); and (C) nonlinear fitting of derivative absorbance values at λ = 321 nm as a function of lipid concentration (DMPC).

Figure 3

(A) Schematic representation of the possible location/orientation of the fluorescent probes TMA-DPH and DPH (yellow structures) and the drug DCF within PC molecules of the lipid bilayer. (B) Excitation and emission spectra of the fluorescence probe TMA-DPH incorporated in LUVs of DMPC. Red arrow shows the fluorescence quenching effect of increasing concentrations of DCF added to the labeled system at a temperature of 37 °C and pH 5.0. (C) Stern-Volmer plots of the probes TMA-DPH and DPH incorporated in LUVs of DMPC at a temperature of 37 °C and pH 5.0 as a function of DCF membrane concentrations. Grey symbols (■) were obtained from fluorescence lifetime measurements of TMA-DPH; black closed symbols (●) were obtained from fluorescence steady-state measurements of TMA-DPH; and red open symbols (◯) were obtained from fluorescence steady-state/lifetime measurements of DPH. All are average values ± standard deviation of at least three independent measurements. Lines are the best fit according to Equation (2). (D) DCF molecule with the quenching moieties (O, N, and Cl) highlighted.

Figure 4

Normalized count rate of DMPC lipid vesicles unloaded and DCF-loaded (40 µM) as a function of temperature. Continuous lines are the best-fitted curves according to Equation (8). Values reported are the mean ± standard deviation of three replicate samples. T_p and T_m stand for pre-transition temperature and main phase transition temperature, respectively.

Figure 5

(A) Rotational correlation time (θ) of TMA-DPH (closed black symbols) and DPH (open red symbols) in DOPC:DOPE:CL (1:1:1) liposomes at L_α phase (circles) and H_II phase (triangles) as a function of DCF concentration. (B) Calcein release from DOPC:DOPE:CL (1:1:1) liposomes at L_α phase as a function of DCF concentration. Calcein release (%) was calculated according to Equation (7). Results are representative of at least three independent determinations ± standard deviation. Dashed lines are guidelines only.

Figure 6

Long and short spacing obtained, respectively, by SAXS (A,C) and WAXS (B,D) measurements of DPPC multilayers made at several temperatures and at pH 5.0 in the absence (open symbols) and in the presence (closed symbols) of DCF. Values are presented as average spacing ± standard deviation calculated from all diffraction peaks. Dashed lines are guidelines only, and star symbols are temperatures at which no Bragg peaks appear in the WAXS region. L_β, L_β’, P_β’, and L_α stand for lamellar gel phase (β stands for untilted and β’ stands for tilted), ripple phase, and lamellar fluid phase, respectively.

Figure 7

(A) WAXS patterns recorded in static exposures at 20 °C and at pH 5.0 for multilayers of DPPC in the absence (lower figure) and presence of DCF. Solid red lines give the best fit of the Lorentzian’s analysis model to the scattered intensities. (B) Schematic representation of the effect of DCF penetrating the headgroup region of DPPC causing the loss of tilt angle with consequent increase of the long spacing (d) and decrease of cross-sectional area (A₀) of the aliphatic lipid chain.