Ocular and systemic toxicity of high-dose intravitreal topotecan in rabbits: implications for retinoblastoma treatment

Abstract

Although many more eyes of children with retinoblastoma are salvaged now compared to just 10 years ago, the control of vitreous seeding remains a challenge. The introduction of intravitreal injection of melphalan has enabled more eyes to be salvaged safely but with definite retinal toxicity. Intensive treatment with high-dose intravitreal topotecan may be a strategy to control tumor burden because of its cell cycle-dependent cytotoxicity and the proven safety in humans. Therefore, we evaluated the ocular and systemic safety of repeated high-dose intravitreal injections of topotecan in rabbits.

Systemic and ocular toxicity was assessed in non-tumor-bearing rabbits after four weekly injections of three doses of topotecan (10μg, 25μg, and 50μg) or vehicle alone. Animals were evaluated weekly for general and ophthalmic clinical status. One week after the last injection, vitreous and plasma samples were collected for drug quantification and the enucleated eyes were subjected to histological assessment.

Weight, hair loss, or changes in hematologic values were absent during the study period across all animal groups. Eyes injected with all topotecan doses or vehicle showed no signs of anterior segment inflammation, clinical or histologic evidence of damage to the retina, and ERG parameters remained unaltered throughout the study. Vitreous and plasma topotecan lactone concentrations were undetectable.

Four weekly intravitreal injections of topotecan up to 50μg in the animal model or a 100μg human equivalent dose were not toxic for the rabbit eye. High doses of topotecan may show promising translation to the clinic for the management of difficult-to-treat retinoblastoma vitreous seeds.

Introduction

The landmark development of a safety-enhanced technique for intravitreal injection (IVi) of chemotherapy has revolutionized the conservative management of advanced intraocular retinoblastoma leading to a higher number of eyes in which enucleation due to the presence of vitreous seeds may be avoided (Munier et al., 2012b, 2012a).

Tumor burden can be controlled by IVi as this route of drug administration bypasses the blood-retinal barrier directly exposing the avascular vitreous and, probably, subretinal seeds to high concentrations of chemotherapy as shown in animal models (Buitrago et al., 2016, 2010). Also, it results in low plasma exposure and no systemic drug-related toxicity in humans (Abramson et al., 2019; Buitrago et al., 2016, 2010; Francis et al., 2017, 2014). More than 30 years ago, Inomata et al. reported in vitro sensitivity to conventional chemotherapy agents and showed that nitrogen mustard melphalan at a concentration of 4μg/ml provided the highest cytotoxicity and most homogeneous response against a panel of primary and commercial cell lines of retinoblastoma (Inomata and Kaneko, 1987). Based on these results and the absence of ocular toxicity after a single IVi dose of 10μg in albino rabbits or a human equivalent dose of 23μg, Japanese researchers implemented the use of up to 16μg of IVi melphalan in children with the longest follow-up to date (Kaneko, 1999; Suzuki et al., 2015). Subsequent reports by ophthalmologists of the main clinical centers around the world showed that the use of 20 to 30μg per dose of melphalan resulted in globe salvage in eyes that one decade ago would never have been treated conservatively (Francis et al., 2017, 2015; Mirzayev et al., 2020; Munier et al., 2012; Shields et al., 2016; Yousef et al., 2021). Nonetheless, this great success is offset by permanent retinal toxicity as well as other lower-incidence toxicities, including pupillary synechiae, iris atrophy, and optic atrophy (Aziz et al., 2017; Francis et al., 2017, 2015, 2014; Narala et al., 2019; Yousef et al., 2021). Interestingly, pigmented eyes (brown) have shown more retinal dysfunction than lighter eyes (blue) in patients which is consistent with irreversible retinal damage found in pigmented rabbits (Francis et al., 2017). A probable explanation is that melphalan binds to melanin of the retinal pigment epithelium resulting in a reservoir-effect with slow drug release and therefore prolonged drug exposure to the retina (Francis et al., 2014; Rimpelä et al., 2018).

Thus, in order to intensify drug treatment in eyes with advanced or refractory retinoblastoma, in cases of poor patient compliance, or to preserve vision in a single-remaining eye in bilateral patients, alternative non-toxic drugs should be explored. In this sense, the S-phase inhibitor topotecan has proven to be safe after intravitreal injection of up to 30μg/dose in humans (Bogan et al., 2021; Francis et al., 2017; Kiratli et al., 2020; Nadelmann et al., 2021; Rao et al., 2018). A key element for optimizing its efficacy is the highly schedule-dependent cytotoxicity. Protracted schemes of intravenous administration showed better tumor control than a single high dose, as the former increased the probability of targeting actively-replicating tumor cells due to its cell-cycle-phase dependence (Houghton et al., 1995). Thus, high and prolonged vitreous exposure would be desirable to leverage antitumor efficacy without the need of further IVi injections. We previously reported that after only 5μg IVi topotecan in rabbits or a human equivalent dose of 10μg, pharmacologically active concentrations in the vitreous represented by topotecan lactone moiety exceeded the 50% inhibitory concentration (IC50) of a commercial cell line for more than 12h (Buitrago et al., 2010). Despite the prolonged exposure, no retinal toxicity was observed in the rabbits (Buitrago et al., 2013). However, concerns about IVi topotecan efficacy at the current clinical doses and the S-phase cytotoxicity warrant the evaluation of higher doses to maintain the vitreous concentrations above the cytotoxic effect for prolonged intervals.

Therefore, we aimed to evaluate the ocular and systemic safety and toxicity of repeated high-dose intravitreal injections of topotecan in rabbits for the potential translation to difficult-to-treat retinoblastoma vitreous seeds.

Methods

Fifteen New Zealand rabbits weighing between 1.5 and 2.1 kg were included. The animals were housed under a 12-hour light/dark cycle and given free access to food and water. All eyes underwent pupillary mydriasis that was induced with 5% phenylephrine hydrochloride and 0.5% tropicamide (Fotorretin, Poen Lab., Argentina) and topical anesthesia with 0.5% sterile proparacaine hydrochloride (Anestalcom, Poen Lab., Argentina). Before IVi injection and ophthalmic examination animals were anesthetized as previously described (Buitrago et al., 2013; Francis et al., 2014). Animals were assigned to one of three groups (n=4 in each group) that received four weekly injections of three different doses of topotecan (Group A, 10μg/dose; Group B, 25μg/dose; Group C, 50μg/dose, yielding human equivalent doses of approximately 20, 50, and 100μg/dose considering the vitreous volume of rabbits and humans of 1.7–2ml and 4ml, respectively) into the right eye and vehicle into the left eye (Group D). The vehicle was prepared as a saline solution with mannitol and tartaric acid at the same concentration and pH described for the commercial product and diluted accordingly to match the dilutions of the corresponding topotecan solution injected in the contralateral eye. A fourth group served as controls (n=3, Group E) and received 50μl of vehicle in the right eye and normal saline was administered to the left eye. In all cases, IVi injections were performed without anterior chamber paracentesis as previously described using a 31-gauge needle directed to the optic nerve and the same volume of 50μl of vehicle, saline, or topotecan solution were injected.

The lowest IVi dose of 10μg corresponds to a human equivalent dose of about 20 to 23μg that is currently used in the clinics. In addition, to determine the maximum study dose we estimated the dose of topotecan that would result in an at least 2-fold increase in the interval of time during which vitreous topotecan lactone concentrations would be above the IC50 in a primary cell culture (HPG-RBT-12L) of retinoblastoma and the topotecan-resistant isogenic model. The resistant cell model was derived from the parental cell line after exposure to three weekly doses of topotecan and thereafter, the resistant phenotype was confirmed by an increment of the IC50 from 13.2nM in the parental cell line to 79.4nM in the resistant variant. For these estimations, we used our previous pharmacokinetic data (Buitrago et al., 2010) in the vitreous of rabbits after a single IVi dose of 5μg topotecan and simulations were performed using ADAPT II (D’Argenio David Z., Schumitzky Alan, 2009). Our results indicated that a dose of 50μg would lead to a 2- to 5- fold increase in the interval of time during which topotecan concentrations are above the IC50 in both parental and topotecan-resistant cells and thus we selected this as the maximum dose.

Before the start of the study and on a weekly basis, all animals underwent complete clinical inspection regarding weight loss, hair loss, general condition, and complete blood count. A peripheral blood sample was collected for lactone topotecan quantification by high-performance liquid chromatography as previously described (Buitrago et al., 2010). In the anesthetized animals (37.5 mg/kg ketamine hydrochloride and 5mg/kg xylazine), ophthalmic clinical evaluations for signs of ocular inflammation, retinal damage or cataract formation included direct and indirect ophthalmoscopy, biomicroscopy, and mixed rod-cone scotopic electroretinography recordings (Akonic BIO-PC, Akonic, Argentina) were performed at baseline and one week after the last administration as detailed elsewhere (Buitrago et al., 2013).

At termination of the follow-up period, vitreous and blood samples were collected for topotecan quantification; thereafter, the rabbits were euthanized by intravenous injection of an overdose of 40% sodium pentobarbital and 5% diphenylhydantoin (Euthanyle, Brouwer, Argentina) and both eyes were enucleated and processed for routine histological evaluation.

Animal weight, hematologic values, and ERG parameters (a- and b-wave amplitude and implicit times) were compared between animal groups or treated eyes by means of two-way repeated measures ANOVA to account for treatment and time with the Bonferroni test a posteriori (significance set at p<0.05). Statistical analysis and graphs were performed using GraphPad Prism v.8., R software and RStudio Version 1.3.959, 2020, Inc.

All experimental procedures adhered to the Association for Research in Vision and Ophthalmology Statement on Animal Use and were approved by the Animal Care and Use Committee of the Faculty of Veterinary Sciences, UNCPBA, Argentina.

Results

Changes in weight or hematologic values were absent during the study period across all animal groups as shown in Figure 2A. Eyes injected with all topotecan doses or vehicle showed no signs of anterior segment inflammation and no retinal or optic nerve changes were evidenced by fundus examination (Figure 1A). All study eyes remained normotensive throughout the study period.

Figure 2.

A. ERG parameters for vehicle- and topotecan-treated eyes. All values are shown as expressed as mean (SD). B. Weight and hematologic parameters of animals treated with topotecan 10μg in the right eye and vehicle in the left eye (light grey circles), topotecan 25μg in the right eye and vehicle in the left eye (grey circles), topotecan 50μg in the right eye and vehicle in the left eye (black circles), vehicle in the right eye and saline in the left eye (group D).

Figure 1.

A. Representative fundoscopic images one week after the fourth IVi of vehicle, 10μg, 25μg, and 50μg of topotecan. All eyes showed normal morphology of the retina and optic disc and unaltered quantity and quality of retinal and choroidal vessels. Both topotecan- and vehicle-treated eyes were free of signs of inflammation. B. Representative micrographs of retinal sections of eyes treated with vehicle, 10μg, 25μg, and 50μg of topotecan. Objective lens x40, hematoxylin and eosin.

ERG parameters remained unaltered in all groups of eyes injected with topotecan or vehicle throughout the study follow-up and there was no difference among topotecan-, vehicle-, and saline-injected eyes (Figure 2B, p>0.05). In addition, selective impairment of the inner retinal function was not evident as no reduction in b-to-a wave amplitude was observed at any of the topotecan doses.

Histological examination showed that both vehicle- and topotecan-treated eyes appeared normal. All eyes showed conserved retinal thickness with no histologic evidence of damage to the optic nerve and the photoreceptor and ganglion cell layers (Figure 1B).

Discussion

Our study shows that four weekly intravitreal injections of topotecan up to 50μg in the animal model or 100μg human equivalent cause no systemic or ocular toxicity in a rabbit model.

After four weekly doses of 10-to-50μg of topotecan per dose, retinal dysfunction was absent as the ERG recordings remained stable during the study period in each group of drug-treated eyes and no significant changes were observed among topotecan-, vehicle-, and saline-treated eyes. Moreover, we observed no signs of inflammation or other histological changes of the eye structures related to the acidic vehicle of the commercial product that contained topotecan or after topotecan at any of the studied doses. In line with our previous study, in which we reported that topotecan plasma concentrations were low after a single IVi of 5μg and undetectable after multiple weekly IVi in rabbits, we here show that plasma topotecan concentrations were below the limit of quantitation one week after the last dose and animals presented no signs of hematological or other systemic toxicity.

The outcome of eye globe salvage in the conservative management of advanced intraocular retinoblastoma has greatly improved with the introduction of IVi probably related to the selective exposure to chemotherapy of seeds in the avascular vitreous environment (Francis et al., 2017; Munier et al., 2012a; Shields et al., 2016). The introduction of an anti-reflux technique to avoid extraocular tumor spread after IVi injection and the original reports of the Japanese experience with IVi melphalan as the most cytotoxic agent against retinoblastoma cells, resulted in a fundamental change in the treatment modality worldwide showing excellent tumor response and ocular survival (Munier et al., 2012b). Nonetheless, the therapeutic window with melphalan is narrow and at doses presently being used toxicity is common and occasionally severe. That is the reason for investigating other drugs for intravitreal use. In this regard, topotecan has shown promising advantages due to the well-documented activity in pediatric malignancies including retinoblastoma, the lack of ocular and systemic toxicity after IVi in animals and children used as a single-agent of combined with melphalan, and favorable pharmacokinetics with prolonged exposure in the vitreous of preclinical models (Bogan et al., 2021; Buitrago et al., 2013, 2010; Francis et al., 2017; Nadelmann et al., 2021; Rao et al., 2018; Shields et al., 2016). Nevertheless, in order to treat recurrent or refractory vitreous seeds or cases of poor patient compliance, higher doses are required. For IVi topotecan, current clinical doses range from 20 to 30μg. Although these doses are devoid of ocular toxicity after single-drug or concomitant treatment with melphalan, clinical efficacy of topotecan at these doses has not been confirmed. Thus, clinical outcome may not be so impressive because of the use of doses of IVi topotecan that are lower than those required in cases that need treatment intensification. Direct drug delivery into the vitreous of high doses of chemotherapy may expose the retina to toxic concentrations of nonselective DNA-damaging agents with associated retinal dysfunction and hence, ocular as well as systemic safety has to be assessed before clinical translation. Therefore, we studied the ocular toxicity of high-dose IVi topotecan and we observed that even after multiple injections of 3-fold the highest clinical dose (corresponding to 50μg in the rabbit and a human equivalent dose of 100μg) there were no signs of ocular toxicity in the animal model.

Our results are in line with those from other recent report showing the lack of ocular toxicity after three weekly IVi topotecan at a maximum of 30μg per dose (Bogan et al., 2021). Nonetheless, the design of our study was intended to intensify drug treatment with a rational design based on the simulated concentrations of the active form of topotecan (lactone topotecan) that may be attained after IVi and in relation to the concentrations needed to exert cytotoxicity in the worst clinical scenario represented by topotecan-resistant tumors. Although useful, defining a dose based on the cytotoxicity of chemotherapy in commercial cell cultures may be misleading as they do not completely resemble the heterogeneity of pharmacological responses as seen in vivo. Thus, primary cell cultures and models of the drug resistant variant provide a more realistic approximation and were therefore used in the present study for defining the dose to be evaluated for potential translation into the clinic. We calculated that to double the time above the IC50 in these primary and chemotherapy-resistant cells, a dose of 50μg would be required and based on this finding and considering the worst scenario of refractory or relapsed tumors that may need an intensive scheme of chemotherapy, we decided to evaluate the toxicity after multiple high doses of IVi topotecan.

In conclusion, multiple high doses of up to 50μg of IVi topotecan are safe in the rabbit eye and may show promising translation to the clinic for the treatment of resistant retinoblastoma vitreous seeds and special cases of poor patient compliance.