Formulation and drug-content assay of microencapsulated antisense oligonucleotide to NF-κB using ATR-FTIR

Antisense oligonucleotides to NF-κB have been shown to mitigate the pathophysiology of experimental septic shock [1–5]. The therapeutic potential of antisense oligonucleotides to NF-κB and thus, tumor necrosis factor has been limited by inadequate intracellular permeability [6]. Microencapsulation has been performed within a non-immunogenic, biodegradable bovine serum albumin (BSA) matrix in a 1.0–8.0 μm size range [7–10]. This formulation was optimal for phagocytosis, [11]. The antisense oligonucleotides to NF-kB were added to a cross-linked BSA solution to achieve 5, 10 and 15% microsphere drug loading [6]. Spray drying was performed using a Büchi 191 mini spray drying with parameter settings for optimum drug microencapsulation and yield^TM [3, 12, 13].

The microencapsulation of biopharmaceuticals such as proteins, peptides and oligonucleotides in a biodegradable polymer matrix provides for a method of controlled release delivery system [14, 15].

Microencapsulation enhances endocytotic cell uptake and also protects the biopharmaceutical drugs from enzymatic degradation [11, 14, 16, 17]. Microencapsulation of oligonucleotides such as the antisense to NF-κB presents a challenge with respect to the determination of drug content and, consequently, encapsulation efficiency and formulation consistency. The consistency of a pharmaceutical formulation preparation is critical to active pharmaceutical ingredients (API) performance. The bioavailability and release profiles of the pharmaceutical formulation are dependent on the formulation's consistency. Although, HPLC analysis has been used for determining the content analysis, we are trying to explore the use of attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) to monitor and quantify the microencapsulated oligonucleotide drug.

A solid-state analytical technique utilizing ATR-FTIR was developed [18–20]. Mid-IR spectroscopy is an example of a vibrational spectroscopic technique. Mid-IR spectroscopy is sensitive to functional groups in molecules, with each functional group producing a specific vibrational pattern. The resultant IR spectrum is unique for each molecular structure [19]. The sensitivity of mid-IR is such that the IR spectrum produced is representative of the functional groups and structural interaction of molecules within a compound as well as the concentration of the molecule [19].

For quantitative analysis, antisense to NF-κB oligonucleotide in albumin solid microsphere formulations of known oligonucleotide concentrations was prepared. Infrared spectra of each standard were collected from the samples. Principal component analysis (PCA) of the sample spectra data and multivariate analysis were used to determine the influence of oligonucleotide concentration in the formulations on spectra [20–22]. A specific absorption band unique to the neat drug (non-encapsulated) oligonucleotide and microencapsulated oligonucleotide, but absent in the blank (non-oligonucleotide containing) albumin microspheres was identified. The peak areas were calculated for the various standards. A standard curve of corrected peak area versus microencapsulated oligonucleotide concentration was generated [22].

To the best of our knowledge, use of ATR-FTIR as a method of quantification of microencapsulated oligonucleotide has not been reported elsewhere. This paper will demonstrate the methodology involved developing ATR-FTIR as a process analytical technology (PAT) for content analysis of microencapsulated oligonucleotide formulation. The formulation process will be described along with the sample handling, data acquisition and processing.

2.1. Chemicals

2.2. Preparation of antisense oligonucleotide to NF-κB microspheres

2.3. Data acquisition

ATR-FTIR full spectra of the pure antisense, antisense microspheres and blank microspheres were collected. The equation describing the depth of beam penetration for ATR is as follows:

$$\begin{eqnarray*} &&{\rm Dp}\,({\rm nm}) = \frac{{\lambda \,({\rm nm})}}{{2\pi n_{{\rm atr}} }}\sqrt {\sin \,\theta ^2 - ({{n_{{\rm sample}} }/{n_{{\rm atr}} }})^2,} \end{eqnarray*}$$

depth of penetration (Dp) depends on the wavelength (directly) and the refractive indices of the crystal (n_atr) and sample (n_sample).

Full spectra ATR and baseline correction was performed. n_sample can vary depending on the sample composition. ATR correction transforms spectra collected by ATR into those resembling standard transmission spectra. ATR correction accounted for sample-to-sample biases caused by different beam depths of penetration related to refractive index variations with concentration. Baseline correction simply smoothed the baseline for quantization.

The spectra for the pure antisense-NF-κB and blank BSA microspheres are displayed in figure 1. A clear dependence on the antisense-NF-κB was shown in the spectra in the fingerprint region range of 974–915 cm⁻¹, with a peak at approximately 948 cm⁻¹. The pure antisense-NF-κB and blank BSA microsphere spectra were then deconvoluted with respect to the carboxylic acid group, out-of-plane O–H bend, at 974–915 cm⁻¹ as shown in figures 2 and 3. Deconvolution was used to resolve the peaks within this fingerprint region.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

An antisense-NF-κB microsphere formulations sample set (0–15% antisense-NF-κB loading) was analyzed as shown in figure 4. The calibration sample spectra correlated to antisense-NF-κB concentration of antisense-NF-κB microsphere formulations sample. Full spectra were then analyzed by a standard deviation of mean absorbance as shown in figure 5. The standard deviation of mean absorbance at each wave number of the spectra allowed the identification of peaks showing the most significant differences between the formulations. The narrower was the standard deviation of mean absorbances, the more significant were the differences in peak areas between the formulations of different concentrations. The peak at approximately 948 cm⁻¹ increased in intensity with increased loading of antisense-NF-κB in the albumin microspheres.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The antisense-NF-κB microsphere formulations sample set peak areas for the carboxylic acid group, out-of-plane O–H bend, at 974–915 cm⁻¹ were calculated as shown in table 1. These peak areas were then used to develop a quantitative method for predicting antisense-NF-κB content in albumin microsphere formulation samples. A standard curve of the peak area against antisense-NF-κB percentage loading in albumin microspheres is shown in figure 6. The calibration plot of the peak areas against per cent antisense-NF-κB loading was linear with a correlation coefficient (R²) value of 0.9674 and a sample mean standard error of less than 1 confirming the high precision of this analytical technique.

Zoom In Zoom Out Reset image size

Antisense-NF-κB microsphere formulations of 0, 2.5, 5.0, 7.5, 8.5, 10 and 12.5% loading were prepared as verification samples. The verification sample spectra were collected and peak areas collected at 974–915 cm⁻¹ as shown in table 2 and plotted as in figure 7. The verification plot of actual versus predicted antisense to NF-κB percentage loading in albumin microspheres gave a correlation coefficient (R²) of 0.984, a mean sample standard error and co-efficient of variance less than 1, figure 7. The antisense-NF-κB per cent loading in albumin microspheres in the verification samples was predicted to be within 2% (w/w) of the actual value.

Zoom In Zoom Out Reset image size

This study has demonstrated that mid-IR ATR-FTIR can be successfully used to quantify the antisense oligonucleotide to NF-κB in albumin microencapsulated formulations. In developing this ATR-FTIR method, several factors had to be considered.

Problems associated with the size and handling of the calibration samples were overcome by full spectra ATR and baseline correction. The calibration samples were critical in establishing a valid quantitative methodology. This was achieved by preparing microsphere samples that were both homogenous and contained drug loadings, adequate enough to produce a valid calibration curve.

The quantification method developed utilized the peak area for the carboxylic acid group, out-of-plane O–H bend, at 974–915 cm⁻¹. A calibration plot of the peak areas against percent antisense-NF-κB loading was linear with a correlation coefficient (R²) value of 0.9674 and a sample mean standard error of less than 1%. The linearity indicated the accuracy of the ATR-FTIR analytical technique for albumin microencapsulated antisense-NF-κB quantization. Unbiased samples were prepared to verify the model. The verification plot of actual versus predicted values produced a correlation coefficient (R²) value of 0.984.

ATR-FTIR is an easy-to-use vibrational spectroscopic method which is a simple, rapid, non-destructive and reliable method for sample quantization. As a viable alternative to other solid state techniques such as NMR, equal to or better results than Raman ATR-FTIR can be expected. This procedure also allows for rapid analytical sample turnaround. With respect to oligonucleotide microencapsulation ATR-FTIR can provide a real-time in-process analytical tool to enable process validation.

This study was supported by Dr Carl W Oettinger from the Dialysis Clinics of Atlanta and Dr Martin J D'Souza at the College of Pharmacy and Health Sciences, Mercer University, Atlanta, GA.