In vivo screening of subcutaneous tolerability for the development of novel excipients

Abstract

To develop safe subcutaneous formulations and minimize the risk of local irritation, it is essential to optimize the composition of active pharmaceutical ingredients and excipients. Depending on the physicochemical properties of the active pharmaceutical ingredient, additional excipients may be required to improve the stability and solubility of the active pharmaceutical ingredient. However, some of these excipients may not have been previously used in injectable drugs. Owing to the lack of safety data for such excipients, especially those used in subcutaneous dosing, it is important to evaluate their potential for local irritation during the early stages of formulation development. We evaluated the tolerability of 44 formulations with 24 candidate novel excipients, such as surfactants, polymers, and lipids, in a single subcutaneous dose in rats. Excipient formulations were administered as single bolus subcutaneous injections with an injection volume of 1 mL. The injection sites were observed for 2 days, and macroscopic and microscopic examinations were conducted. Local tolerability was evaluated on the basis of severity, incidence, and pathophysiology of each finding. Formulations that caused tissue degeneration or necrosis, which is indicative of tissue injury, were determined to be irritative and poorly tolerated. A single-dose subcutaneous screening study in rats was considered effective in ranking the safety of candidate excipients during the formulation optimization phase.

In pharmaceutical development, the subcutaneous (SC) route is often selected for its slow absorption rate¹ and easy injectability for patients. Depending on the physicochemical properties of the active pharmaceutical ingredient (API), additional excipients may be required to improve the stability and solubility of the API². It is essential to optimize the API and excipient composition to minimize the risk of local irritation, especially for formulations requiring daily dosing, because medication that causes irritation should not be subcutaneously administered¹; adverse reactions, such as tissue necrosis or abscess, may occur at injection sites, and this could be a critical issue for subjects and patients. There are a limited number of excipients registered in the FDA Inactive Ingredient Database that can be used in formulations for SC administration³. In addition, there is relatively little information about the non-clinical safety of excipients administered subcutaneously compared with the information available for excipients administered via other routes, such as the oral and intravenous routes⁴. Thus, early safety evaluation of candidate excipients is important for the development of safer formulations. In this study, we conducted in vivo screening of novel excipients (in single doses formulations), such as surfactants, polymers, and lipids, in rats to rank candidate excipients based on local tolerability. In addition to the candidate excipient formulations, we prepared two vehicles containing marketed SC products, insulin glulisine (Apidra^®; Sanofi-Aventis U.S. LLC, Bridgewater, NJ, USA)⁵ and liraglutide (Victoza^®; Novo Nordisk A/S, Bagsvaerd, Denmark)⁶, according to the label information, to use as references for measuring the significance of histopathological findings. Additionally, using vehicles of marketed SC products as references helped determine the developability of the novel excipients.

The care and use of the animals and experimental protocols used in this study were approved by the Institutional Animal Care and Use Committee (IACUC) of BoZo Research Center Inc. (approval No. G190297). In this study, we used 90 6-week-old male Sprague-Dawley rats (Crl:CD(SD); Charles River Laboratories Japan, Inc., Kanagawa, Japan). Each dose group was assigned two rats for local tolerability screening. The rats were housed in a solid-floored plastic cage (W 440 × D 275 × H 180 mm; Hannyu Co., Saitama, Japan) with bedding (ALPHA-dri, Shepherd Specialty Papers, Inc., Watertown, TN, USA) under the following environmental conditions: the temperature was maintained between 20 °C and 26 °C, relative humidity ranged from 30% to 70%, the rate of air exchange was between 10 and 15 times/h, and a 12-h light/dark cycle was implemented (lights on from 7:00 a.m. to 7:00 p.m., 300 lux or less). The rats were allowed free access to pelleted diets (γ-irradiated CR-LPF; Oriental Yeast Co. Ltd., Tokyo, Japan) and tap water. Appropriate environmental enrichment was provided to the rats in accordance with the guidelines of the IACUC. The data in this report was derived from three separate toxicity studies that had identical study procedures.

For this screening study, we produced 45 formulations (including two reference vehicles and a negative control mentioned below) using 24 excipients, including novel surfactants, polymers, and lipids, for SC formulation development, as shown in Table 1. Each excipient was prepared at a concentration assumed to confer clinical benefits, such as improvement of stability or solubility of the drug load. To rule out the possibility of local irritation due to physicochemical properties, most of the formulations were adjusted to a pH between 5.5 and 8.5¹ using acid or base, and physiological osmolality was adjusted using phosphate-buffered saline or mannitol. Please note that pH of polyvinyl alcohol at 4.6% w/v was 5.3, which can be considered to be within the injectable pH range of saline⁷. The pH values of the following formulations were outside the pH range of 5.5–8.5: gentisic acid at 0.2% w/v (pH 3.2), benzoic acid at 0.2% w/v (pH 4.0), levulinic acid at 0.2% w/v (pH 4.4), and oleyl sarcosine at 1% / Transcutol at 10% (pH 4.7). These pH conditions were selected to maintain the native charged state of the main formulation component (API), which would improve the physicochemical properties of potential APIs (solubilization through co-crystal/co-amorphous formation or molecular assemblies). Several excipient formulations had a measured osmolality that was over 600 mOsm/kg, which is reported to be the threshold for pain in humans⁸. These included: trimethyl glycine at 8% (740 mOsm/kg), glycine at 5% (609 mOsm/kg), and sucrose at 20% (657 mOsm/kg). We were unable to measure the osmotic pressure of Transcutol and oleyl sarcosine/Transcutol formulations owing to technical limitations. The reference SC vehicles were prepared using the description on the information labels of Apidra^®5 and Victoza^®6, and the osmotic pressure and pH were measured. Physiological saline (Otsuka Pharmaceutical Factory, Inc., Tokushima, Japan), the specification of which was according to the Japanese Pharmacopoeia⁷, was used as a negative control. The EMA guidelines⁹ recommend a maximum applicable volume for SC tolerability evaluation, and the dose volume in this study was set at 1 mL/site/animal, approximately 5 mL/kg for 6-week-old male rats (the body weight range at dosing: 154–225 g), which is the administration volume commonly used in SC toxicity studies¹⁰. Each test article was administered subcutaneously via the clipped abdominal skin (outside the median) using a needle (25G or 27G). We chose the abdominal skin as the injection site, rather than the dorsal skin, to minimize diffusion of the injected liquid into the mobile SC connective tissue of the dorsal skin¹¹. To evaluate skin irritation reactions, erythema, and edema (dermal swelling indicating fluid retention) at the injection site (abdominal skin), the Draize test was used to score the reactions¹². Observations were performed five times on the day of dosing: at predose, 1, 2, 4, and 8 h post dose; and once a day on days 1 and 2 after dosing. Following clinical observations 2 days after dosing, the animals were euthanized by exsanguination from the abdominal aorta under isoflurane anesthesia. Thereafter, macroscopic examination of the injection site was conducted. The abdominal skin (including injection site) was dissected, fixed, and preserved in phosphate buffered 10% formalin. The abdominal skins from all the rats, which were trimmed to include both the injection site and intact area, was embedded in paraffin, sectioned, stained with hematoxylin and eosin, and examined microscopically. Each microscopic finding was graded as follows: a sparse or focal finding in one region of the subcutis, dermis, or epidermis was graded as “minimal”; a multifocal or diffuse finding in one region of the subcutis, dermis, or epidermis was graded as “mild”; and a multifocal or diffuse finding extending to adjacent tissues, such as findings extending from the subcutis to the dermis and/or epidermis, was graded as “moderate”. In this study, we chose only one timepoint, day 2 post dose, for pathological examination to observe the acute phase effect after a single dose. However, the dose regimen or timing of the terminal necropsy could be arranged with flexibility depending on the intended clinical dose regimen or study purpose, such as the evaluation of repeat dose or recovery.

Table 1. Test Article Information and Local Tolerability Data.

The results of skin observations, necropsy, and histopathology of all the rats are summarized in Table 1; additionally, we listed findings observed in at least one rat from each dose group. If both rats in the group had similar findings, but one of the rats was more severely affected than the other, the more severe grade was recorded in the table.

During the clinical observations, edema was observed after the administration of various formulations. The formulations included both irritative and non-irritative formulations (described in the next paragraph), which indicated that external edema was not evidence of SC irritation, especially when observed at earlier time points, such as 1 to 8 h after dosing. On the other hand, external black or dark red discoloration and crust formation, most of which were observed at later time points, such as one to two days after dosing, were observed only in the irritative formulations. In addition, macroscopic findings of crust formation and black focus in the epidermis and/or subcutis were observed during necropsy of the rats that had received the irritative formulations. Dark red foci in the subcutis, which were noted with several formulations, seemed to be related to microscopic hemorrhage.

Twenty-eight out of forty-four excipient formulations that showed necrosis/degeneration, indicating tissue injury, were considered irritative and poorly tolerated as single SC injections (examples represented in Fig. 1). In these 28 formulations, other histopathological findings, such as mononuclear/mixed cell infiltration, granulocytic cell debris, regeneration of muscle fibers, thrombus, hemorrhage, and/or crust formation were observed. Two formulations (HPC at 14% and vitamin E TPGS at 0.18%) showed only mild mononuclear cell infiltration, which was considered a less irritative test-article-related finding owing to a lack of apparent tissue injury. Although minimal mononuclear cell infiltration, hemorrhage, and/or crust formation were observed after administration of trimethyl glycine at 3% and 8%, glycine at 2%, 1, 2-Distearoyl-sn-glycero-3-phosphoethanolamine-Polyethylene glycol 5000 (PEG5k-DSPE) at 0.4%, Polyxamer 188 at 0.54%, Kolliphor KL at 0.04% and 0.12%, Transcutol at 10%, benzoic acid at 0.2%, and levulinic acid at 0.2%, they were considered to be spontaneous or SC injection procedure-related findings. In addition to these formulations, sucrose and glycine at 5% had no microscopic findings, and these formulations were well tolerated.

Fig. 1.

Histopathological features of injection site (subcutaneous tissue), 2 days after single dose. Saline: No remarkable changes were noted. Bar=250 µm. Vehicle of Victoza: Mild degeneration/necrosis (*) and minimal mixed cell infiltration were observed. Bar=250 µm. Kollidon VA (4.6%): Minimal degeneration/necrosis (*) and mild mononuclear cell infiltration were observed. The picture on the right is a highly magnified image of these findings. Bar=250 µm (left) and 50 µm (right). Polyvinyl alcohol (4.6%): Minimal degeneration/necrosis (*) and moderate mixed cell infiltration were observed. The picture on the right is a highly magnified image of these findings. Bar=250 µm (left) and 50 µm (right).

To avoid the risk of tissue damage, a pH range of 5.5–8.5 is recommended¹. Low pH values (<5.5) could be related to the SC irritation observed after the administration of some formulations, such as polyvinyl alcohol at 4.6%, gentisic at 0.2%, and oleyl sarcosine at 1% / Transcutol at 10%; however, no irritative findings were noted with benzoic acid at 0.2% and levulinic acid at 0.2%, which had low pH values. Furthermore, a correlation between the degree of hypertonicity and the sensation of pain in humans was reported, and an upper limit of 600 mOsm/kg was proposed to minimize the risk of pain⁸. In this evaluation, 3 formulations, trimethyl glycine at 8%, glycine at 5%, and sucrose at 20% did not appear to cause any irritation despite having high osmolality values (>600 mOsm/kg).

In this screening assessment, vehicles of marketed products (Apidra^® and Victoza^®), which are subcutaneously administered daily in clinical settings, were used as reference vehicles. Clinical local tolerability data of the reference vehicles is not available; however, they showed minimal to mild tissue degeneration or necrosis in rats (Fig. 1). Although irritative candidate novel excipients should be deprioritized for safer clinical formulation development, comparison of tolerability profiles of these excipients with those of marketed reference vehicles, in a risk-benefit analysis, could support their use. For example, the developability of novel excipients could be considered if their tolerability is similar to that of the reference vehicles, and the benefits of the excipients are significant for patients; such benefits include improvement of API stability and solubility for better pharmacodynamic and pharmacokinetic profiles.

In conclusion, evaluation of the SC tolerability of novel excipients, using a single-dose, in rat models identified a useful way to rank the safety of novel candidate formulations at the early stage of development.

Disclosure of Potential Conflicts of Interest

The authors declare that there are no conflicts of interest associated with this manuscript.