Stability, Activity, and Application of Topical Doxycycline Formulations in a Diabetic Wound Case Study

Introduction. Tetracycline molecules comprise a group of broad-spectrum antibiotics whose primary mechanism of action is the inhibition of protein synthesis through the binding of the bacterial ribosome. In addition, tetracyclines inhibit matrix metalloproteases (MMPs), a family of zinc-dependent proteases that contribute to tissue remodeling, inflammation, and angiogenesis and are overexpressed in certain pathophysiologies such as diabetic foot ulcers (DFUs). Objective. This study aims to develop a liquid chromatography and mass spectrometry (LC-MS/MS) doxycycline quantification methodology to facilitate the development of a stable topical doxycycline hyclate (DOXY) formulation as well as evaluate the topical DOXY formulation for the efficacy in MMP-9 inhibition in vitro and in a clinical application of diabetic lower extremity wounds. Materials and Methods. A simple quantification method utilizing LC-MS/MS was used to develop a topical DOXY formulation, a sample of which was analyzed in stability testing. The formulation was evaluated in vitro for MMP-9 activity using a commercial assay and compared with internal kit controls as well as in a clinical setting for wound healing. Results. Two formulations of 2% (w/w) DOXY demonstrated acceptable stability (±10% target concentration) for 70 days when stored at 4°C. Using an in vitro assay of MMP-9 enzyme activity, the 2% DOXY formulation imparted a ~30% decrease in MMP-9 inhibitory potential as compared with the control drug alone (IC₅₀ values 62.92 µM and 48.27 µM, respectively). This topical product was evaluated for clinical utility in a patient with a DFU, and preliminary data suggest this intervention may promote wound healing. Conclusions. In summary, novel DOXY formulations may be stable and biologically active tools amenable to complex wound care. 

RELATED CONTENT

Doxycycline as a Modulator of Inflammation in Chronic Wounds

Evaluation of a Novel Multilayer Wound Dressing That Transports Chronic Wound Fluid Away From the Wound

Treatment of Recalcitrant Wounds of Diverse Etiology With a New Hydroactive Gel

Diabetes affects hundreds of millions of people worldwide. Disease prevalence has increased dramatically over the past 2 decades,¹ with the number of patients with diabetes mellitus (DM) increasing from 30 million cases in 1985 to 285 million in 2010 and estimated to reach 360 million people by 2030.^2,3 Individuals with DM exhibit a documented impairment in acute wound healing.⁴ Moreover, this population is prone to developing chronic, nonhealing diabetic foot ulcers (DFUs), which are estimated to occur in 15% of persons with DM and precede 84% of all diabetes-related lower leg amputations.⁵

Offloading is the mainstay of diabetic foot care as excessive pressure on the injured foot severely retards ulcer healing.^6,7 Irremovable total contact casts are considered to be the gold standard in treatment.⁶ Offloading is critical because patients with diabetic neuropathy and foot ulcers are unable to feel pain; the sensation of pain causes a non-neuropathic person to limp, instinctively offloading the injured foot to prevent further damage. This trigger is absent in a patient with diabetic neuropathy, hence why they will likely maintain pressure on the ulcered foot and hinder wound healing.⁸

The impaired healing of both DFUs and acute cutaneous wounds in persons with diabetes involves multiple complex physiological mechanisms. It has been hypothesized that persistently high concentrations of cytokines within the wound maintain the inflammatory response and induce high concentrations of proteases, which degrade growth factors, receptors, and matrix proteins that are essential for proper wound healing.^9–12 The most common proteases over expressed in DFUs are members of the matrix metalloproteinase (MMP) family.^9,13 These particular proteases are enzymes involved in various biological reorganization events, including wound healing.^14,15 Natural substrates for different MMPs vary substantially but include important extracellular matrix (ECM) proteins such as collagen, gelatin, and proteoglycans. The ECM constituents are essential components of the wound repair phenomenon; therefore, MMPs play an important role in normal wound healing.¹⁶ High levels of MMPs in wound fluids and biopsies have been reported to correlate with poor chronic wound healing, and decreased MMP levels are observed as chronic wounds begin to heal.^9,13 Although there are 23 MMP family members, Widgerow et al¹⁷ reported that the activities of MMP-2 and MMP-9 are several times higher in nonhealing wounds and that inhibition of these enzymes helps to enhance the healing process. This has given rise to numerous attempts to design MMP inhibitory drugs and the observation that tetracycline antibiotics harbor a latent level of inhibitory activity.^17,18 

Doxycycline hyclate (DOXY), a semisynthetic and chemically modified analog of tetracycline, is widely used to treat infections caused by both gram-negative and gram-positive microorganisms,¹⁹ but it exerts biological effects independent of its antimicrobial activity.²⁰ One such effect includes the inhibition of MMPs²¹ through the chelation of structural (calcium) or catalytic (zinc) ions necessary for enzymatic function.^22,23 Oral DOXY has been used in a pilot study²⁵ involving chronic leg ulcers with preliminary success; however, the drug levels within the wound were noted to be insufficient to inhibit MMP activity. Serra et al²⁶ found that topical DOXY reduced proteolytic activity in the wound fluid of chronic venous leg ulcers, and clinical application of topical 1% DOXY treatment showed a beneficial effect on healing.²⁴ Combined, preliminary evidence for positive outcomes in the treatment of chronic wounds with DOXY, both topical and systemic, has been established.^25,26 

The stability of DOXY and its degradation profile has been investigated as a function of pH, temperature, time, light exposure, and in combination with a multitude of stabilizing additives.^27–30 Of particular note, KuKanich et al³⁰ evaluated a variety of United States Food and Drug Administration-approved doxycycline formulations with acceptable stability for dried materials (eg, capsules), whereas semisolids and compounded liquids failed to meet United States Pharmacopeial Convention standards after 21 days of storage.³⁰ With these topics in mind, the authors set about developing a quantification method to support DOXY topical formulation development, evaluate drug stability and formula activity with respect to MMP-9, and gain preliminary insight into the product’s utility in a clinical setting.

Chemicals, reagents, and test articles
Used within the study are doxycycline hyclate, usp (0434-04; Medisca, Montreal, Quebec, Canada), doxycycline-d₃hyclate (major; D561503; Toronto Research Chemicals, North York, Ontario, Canada), methanol (MX0488-1; MilliporeSigma, Burlington, MA), acetonitrile (1405-7-40; Caledon Laboratories Ltd, Halton Hills, Ontario, Canada), high performance liquid chromatography (HPLC) water (8801-7-40; Caledon Laboratory Chemicals Ltd), Arlasolve dimethyl isosorbide (DMI; Croda International plc, Snaith, Yorkshire, United Kingdom), DelivraSR (topical vehicle control [TVC]; 13750-2; Delivra Corp, Hamilton, Ontario, Canada), sodium hydroxide (S8045; MilliporeSigma), hydrochloric acid (HCl; HCl333.500; BioShop Canada Inc, Burlington, Ontario, Canada), trifluroroacetic acid (299537; MilliporeSigma), magnesium chloride (MgCl₂; M8266; MilliporeSigma), and (2-hydroxypropyl)-β-cyclodextrin (β-HPCD; 332593; MilliporeSigma).

Formulations
A topical formulation of DOXY was prepared by first dissolving 100 mg of the drug in 1 mL of DMI then adjusting the pH of the solution to between 5.5 and 6.5 with a 2-M aqueous solution of sodium hydroxide (NaOH). Then, 4 g of TVC was added to the solution, and the mixture was homogenized using a vortex mixer, sonication, and manual stirring. Three formulations were stored in airtight, light-protected containers at -20°C, 4°C, and 25°C, respectively, for 70 days until analysis.

A second topical formulation of DOXY was prepared by first dissolving 100 mg of the drug and 268 mg of β-HPCD in 1 mL of distilled water. The pH of the solution was adjusted to between 5.5 and 6.5 with a 2-M aqueous solution of NaOH. Then, 4 g of TVC was added to the solution, and the mixture was homogenized using a vortex mixer, sonication, and manual stirring. Three formulations were stored in airtight, light-protected containers at -20°C, 4°C, and 25°C, respectively, for 70 days until analysis.

A third topical formulation of DOXY was prepared by first mixing 100 mg of the drug and 18.5 mg of MgCl₂ in 1 mL of DMI then adjusting the pH of the solution to between 5.5 and 6.5 with a 2-M aqueous solution of NaOH. The, 4 g of TVC was added to the solution, and the mixture was homogenized using a vortex mixer, sonication, and manual stirring. Three formulations were stored in airtight, light-protected containers at -20°C, 4°C, and 25°C, respectively, for 70 days. 

Instrumentation
High-performance liquid chromatography and tandem mass spectrometry (HPLC-MS/MS) was carried out on 5500 QTRAP Mass spectrometer with a TurboV source (SCIEX, Concord, Ontario, Canada), equipped with Agilent 1260 Infinity HPLC system (Agilent Technologies, Santa Clara, CA). 

Stability testing
A sample of each formulation to be analyzed was weighed into scintillation vials in triplicate, then enough extraction solvent (50:50 MeOH:H₂O adjusted to a pH of 2 with 1-M HCl) was added to make a 1-mg/mL solution of the test. The samples then were sonicated for 30 minutes at room temperature and centrifuged at 11 000 rpm for 10 minutes. After, 10 µL of the supernatant was added to 990 µL of a 126-ng/mL solution of doxycycline-d₃ hyclate in methanol/water (50:50). The resulting solution was mixed thoroughly and submitted for analysis within the authors’ laboratory. 

Stock solutions of DOXY and d₃-doxycycline were prepared by dissolving accurately weighted standards in 60:28:12 of water to acetonitrile to methanol, respectively, to generate 1-mg/mL solutions, and all other working solutions were prepared from these. A solution of 1-µg/mL DOXY + 125-ng/mL d₃-doxycycline in 50:50 methanol:water, respectively, was serial diluted with a 125-ng/mL d₃-doxycycline solution to give a standard series of 3.9, 7.8, 15.6, 31.2, 62.5, 125, 250, 500, and 1000 ng/mL of DOXY and a constant concentration of 125 ng/mL doxycycline-d₃hyclate. Ratio of peak area of doxycycline to peak area of d₃-doxycyline was plotted against doxycycline concentrations and used to produce a standard curve in the form of Y = a + bX using weighted least squares linear regression. Each batch of unknown samples was accompanied by a standard series and a set of quality control samples. Blank control samples were prepared from TVC cream, processed as previously described, and evaluated for specificity. The TVC was spiked with doxycycline at a concentration of 2% w/w and then extracted using the above method to evaluate the precision, accuracy, and recovery of the method. A blank extract of TVC was spiked with doxycycline at a concentration of 2% w/w to assess matrix effects.

Conditions for HPLC-MS/MS
Isocratic chromatographic separation was performed on a C18 column (4.6 mm x 150 mm, 5 µm; Eclipse XDB-C18, USKH095544; Agilent) with guard using a mobile phase of acetonitrile (0.1% trifluoroacetic acid) to water (0.1% trifluoroacetic acid) (30:70) at a flow rate of 0.75 mL/minutes for 10 minutes. The first 2 minutes were sent to the waste and doxycycline was eluted between 6.5 minutes to 7.5 minutes. There was no post time. The column temperature was 30°C with an injection volume of 5 µL.

A spectrometer system equipped with an electrospray ionization probe was used in the positive ion mode with multiple reaction monitoring for the quantitative analysis. Nitrogen was used as the collision gas and the curtain gas. The curtain gas was 10.00 psi, the collision gas was medium, the ion spray voltage was 5000 V, the temperature was 500°C, and gas sources 1 and 2 were 45 psi and 50 psi, respectively. The declustering potential was 40 V, the exit potential was 10.00 V, the focusing lens 1 was -10.50 V, and the cell exit potential was 4.00 V. Quantification was performed using the transitions m/z 444.8 J 428.2 (CE = 28 V, 100 msec) for doxycycline and 447.8 J 431.2 (CE = 28 V, 100 msec) for d₃-doxycycline with low resolution. Analytical data were acquired, and quantification processing was performed using Analyst (SCIEX) software. 

MMP-9 activity assay
The MMP-9 activity was measured using a commercially available MMP-9 inhibitor screening assay kit (ab139449; abcam, Cambridge, MA), and the procedure was completed as described within the product manual.³¹ Briefly, using the provided 96-well microplate, 70 µL of MMP-9 enzyme (0.013 U/µL; final concentration 0.009 U/µL) was added to the assay buffer along with 20-µL aliquot of assay buffer that included various concentrations of test articles (see below) or positive controls. The microplate reactions were mixed and prewarmed to 37°C, 10 µL of substrate (40 µM, final concentration 4 µM) was added, and real-time kinetics were collected over 10 minutes at 1-minute intervals using excitation and emission wavelengths of 328 nm and 420 nm, respectively. Relative fluorescence units per minute were fitted to a linear equation (y = mx + b) and graphical representations of reaction velocity (m) and drug concentration along with IC₅₀ analysis were completed using GraphPad Prism (GraphPad Software, La Jolla, CA).

Both DOXY in solution and within a topical formulation were evaluated for MMP-9 inhibition. The TVC was assessed separately for any independent MMP-9 activity. For experiments involving DOXY in solution, a 1.0-mM stock solution was prepared in deionized water. Using this stock, a 1:1 serial dilution was constructed with a range of 1000 µM to 15.6 µM in assay buffer. For each serial dilution concentration, 20 µL was added in duplicate to enzyme reaction wells (previously described) for a final drug range of 200 µM to 3.1 µM. The entire experiment was conducted on 3 independent days to yield an average IC₅₀ value.

The 2% DOXY formulation (27.7 mg) was weighed in microtube along with 1.0-mL deionized water and then vortexed to yield a homogenous solution (1-mM concentration). The solution was centrifuged at 11 000 rpm for 1 minute, and the supernatant underwent a 1:1 serial dilution with a range of 1000 µM to 15.6 µM in assay buffer. For each serial dilution concentration, 20 µL was added in duplicate to enzyme reaction wells (previously described) for a final drug range of 200 µM to 3.1 µM. The entire experiment was conducted on 2 independent repeats to yield a final IC₅₀ value.

Clinical chart review
The study was conducted in accordance with Good Clinical Practice guidelines and reviewed and approved by Institutional Review Board services (protocol number: Pro00022337) prior to study commencement. Data were collected between January 1, 2014, and July 5, 2017, for patients with lower limb ulcers at the Mayer Institute (Hamilton, Ontario, Canada) treated with the DOXY formulation. General patient information, wound size, wound classification, ulcer location, and photography were deidentified and provided to the Research and Development department of Delivra Corp (Charlottetown, Prince Edward Island, Canada) to maintain patient confidentiality. Wound size was calculated with Tissue Analytics software (Tissue Analytics, Baltimore, MD) and classified using the University of Texas Diabetic Wound Classification (UTDWC) staging system.

The development of a LC-MS/MS method for the quantification of DOXY in formulation preceded and supported the development of a stable formulation of DOXY that was evaluated for efficacy in vitro and in vivo.  

Doxycycline quantification method validation
The linear range for DOXY using this method was 3.9 ng/mL to 1000 ng/mL, with a linear regression equation of y = 0.00572x + 0.0698 and a correlation coefficient (r²)of 0.9925. Lower limit of detection (3x average blank peak area) and lower limit of quantification (10x average blank peak area) for DOXY were both < 3.9 ng/mL. The interday variation was < 5% relative standard deviation, which is well within the precision limits. Acceptable values for accuracy and recovery of control samples were 85% to 115%, and matrix effects were negligible (< 5%). No interfering peaks were observed in chromatogram of blank samples (Figure 1) at the retention time of DOXY (6.5 min).

Doxycycline stability and excipient usage
Real time stability of DOXY within prepared topical formulations was evaluated over a period of 70 days (Figure 2). A DOXY in DMI formulation, as well as a DOXY and β-HPCD in water formulation, had drug concentrations within 15% of the target concentration of 2% w/w after 70 days of storage at 4°C (Figure 2). The incorporation of magnesium chloride in the formulation had a deleterious effect on the stability of DOXY at room temperature (Figure 2). Storage temperature was the most important factor in terms of stability of the formulations when light was controlled. All formulations destined for clinical use had a measured DOXY concentration within 15% of the intended 2% w/w over the course of the analysis (1–3 months) when stored at 4°C in light-shielding containers. 

MMP-9 inhibition
The stable DOXY formulations were evaluated for their MMP-9 inhibitory bioactivity using an in vitro assay (Figure 3). The DOXY control solution produced an IC₅₀ value of 48.27 µM (Figure 3). This result is consistent with the doxycycline IC₅₀ value reported by Bannikov et al³² of 48 µM in neutrophil-purified MMP-9. The IC₅₀ of DOXY within formulation was slightly higher (62.92 µM) than the pure compound, and the TVC did not inhibit the enzyme at the concentrations tested (> 250 µM; Figure 3). 

Observational clinical utility
Evaluation of the clinical utility of the DOXY formulation naturally followed the stability and in vitro MMP-9 assessments. An observational chart review of a female patient with a refractory wound treated with the DOXY formulation was performed (Figure 4) at the Mayer Institute. The 62-year-old woman with type 2 diabetes had a medical history of hypertension, hyperlipidemia, and obesity. The second digit of her right foot had been amputated 2 years prior to presentation. At the beginning of the current case, she was receiving concurrent treatment for a then 180-day-old UTDWC A2 (infected) ulcer measuring 0.9 cm² on the left lateral ankle, which remained active for the duration of this case. The ulcer in this case study, a pressure-induced diabetic ulcer, was not present at initial presentation to the Mayer Institute. At 1-week follow-up, the ulcer in question was characterized with a surface area of 0.3 cm² and a UTDWC of A1 (no infection or ischemia).

At first, treatment consisted of standard of care (SOC; ie, debridement when required, cleaned, offloading, and an INADINE dressing [Systagenix, an Acelity Company, Gatwick, West Sussex, UK] application to prevent infection and maintain a moist wound bed) alone. The ulcer continued to increase in size to 1.3 cm² over the next 54 days; at clinical visit 6 (treatment day 55), topical 2% DOXY application was added to the treatment regimen. The wound began to respond to treatment, reducing in size by 23% within 2 weeks and closing within 98 days of the commencement of topical 2% DOXY (Figure 4). The ulcer was treated with SOC and 2% DOXY until closure was achieved 5 months post presentation. This result is consistent with previous reports^24,25 of topical doxycycline promoting healing in DFUs.

The mechanism of DOXY inhibition on MMP-9 is hypothesized to occur via chelation of the catalytic zinc.²³ The observed reduction of MMP-9 activity from the DOXY formulation compared with pure doxycycline is potentially due to this event being muted by the formulation matrix. Given the matrix itself has little effect on MMP-9 activity, TVC excipients compete for DOXY binding with MMP-9, thus decreasing the drug's effective inhibitory capacity. Despite the decrease in bioactivity in its finished format, application of 2% DOXY is hypothesized to yield sufficient drug levels to ensure MMP inhibition within the wound itself.

Clinical evaluation of 2% DOXY formulation was performed in a case study involving a wound that not only failed to progress towards closure in the first 55 days of SOC treatment but also expanded in size over that time period (Figure 5A, 5B). In this study, SOC was defined as cleaning, debridement, offloading, and the use of a nonadherent polyethylene glycol dressing containing 10% povidone iodine for broad-spectrum infection prevention.³³ The addition of topical 2% DOXY to the treatment regimen had a positive effect on reducing the size of the wound within the first 2 weeks of use, which continued until the wound was completely healed (Figure 5C, 5D). Note that the 2% DOXY formulation treatment was performed in conjunction with SOC, which included proper wound offloading prior to application, without which the studied therapy may not be successful. This case study corroborates healing outcomes previously observed by Chin et al,²⁴ justifying the further investigation of a topical wound treatment containing doxycycline.

A number of variables that may have an impact on wound healing were not reviewed, such as oral medication use (consistent or otherwise). Moreover, these limited observational data are not indicative of outright efficacy, but rather supportive of further investigations within a randomized placebo-controlled clinical setting. 

Formulations of doxycycline were prepared in TVC, and the formulation stability over a 70-day period was evaluated under specific storage conditions using an LC-MS/MS quantification method. The use of DMI or β-HPCD supported drug stability when maintained at 4°C, as temperature proved a key variable to ensure product stability. The DOXY formulation possessed slightly diminished MMP-9 inhibition compared with pure DOXY in vitro, and despite this observation, preliminary data gathered in the present observational case study support further clinical evaluation of the product's utility in DFUs. 

Authors: Simona Gabriele, MSc¹; Beth Buchanan, PhD²; Azoy Kundu, PhD²; Heather C. Dwyer, MSc²; Joseph P. Gabriele, PhD²; Perry Mayer, MB. BCh, BAO³; and David Charles Baranowski, PhD²

Affiliations: ¹Department of Medical Science, McMaster University, Hamilton, Ontario, Canada; ²Research and Development, Delivra Corp, Charlottetown, Prince Edward Island, Canada; and ³The Mayer Institute, Hamilton, Ontario, Canada

Correspondence: David Charles Baranowski, PhD, Delivra Corp, 550 University Avenue, NRC-UPEI Building Suite 407, Charlottetown, PE, Canada, C1A4P3; dbaranowski@delivrainc.com

Disclosure: The authors are/were employees of Delivra Corp, which has a pending patent (Transdermal formulations for delivery of doxycycline, and their use in the treatment of doxycycline-responsive diseases and conditions; Patent no. WO 2017/035665 A1). The authors disclose no financial or other conflicts of interest beyond this pending patent and employee status.