Considerations for development of therapies for cutaneous neurofibroma

Current treatments for cNF

Current clinical management of cNF involves surveillance or some form of procedure-based treatment (table 1). Conventional surgical resection (often termed blade-based) allows for complete removal of the lesions and allows the application of techniques that reduce scarring. This is particularly beneficial for large lesions. There are, however, drawbacks, which include variable scarring (which is influenced by skin type and closure technique), the time required for the procedure and wound healing, and the limited number of lesions that can be treated at a single surgical session.

Alternative approaches have been developed to address these drawbacks, including electrodessication, laser-based treatments (e.g., laser photocoagulation, CO₂ laser), radiofrequency ablation, and photodynamic therapy (table 1). While each of these approaches provides possible treatment options for patients, there are several limitations to the trials that have evaluated these methods. These include a lack of detailed enrollment criteria, distinguishing among the different phenotypes of cNF evaluated, comparisons of treated vs untreated areas, duration of follow-up, and standardized endpoints including the use of validated questionnaires for pain and patient satisfaction. Additional challenges not fully addressed in the literature are the cost and availability of interventions.

Electrodessication is a method in current clinical use that entails tissue desiccation through dehydration and denaturation of the tumor.^8,9 The primary advantage of this approach is the ability to remove large numbers (hundreds to thousands) of lesions in a single procedure over a couple of hours. The limitations are that electrodessication requires the use of general anesthesia, and results in variable scarring, with the rare formation of hypertrophic scarring.¹⁰ Skin pigmentation changes appear to be a more common effect of electrodessication. Further, the number of centers and expert surgeons or dermatologists offering electrodessication under general anesthesia is limited. Laser photocoagulation uses a nonablative laser to finely cauterize blood vessels. It has been reported to be an effective tool for treatment of small cNF.¹¹ CO₂ laser ablation has been used for a number of years, in many indications, though only a few reports have evaluated the efficacy of this method in NF1. A retrospective study of 106 patients with multiple small, superficial (<1 cm) cNF treated with CO₂ laser was recently reported.¹² Patient satisfaction as assessed by questionnaires was reported as >90%, with patients reporting improvements in pain and where 94% of treated patients stated they would use CO₂ laser treatment in the future. While encouraging, a larger prospective study is needed to confirm these preliminary results. Radiofrequency ablation, where high energy flux around the electrode tip imparts a localized thermal effect without damage to adjacent local tissue, has also been utilized to treat cNF.¹³ Although the number of patients reported from these studies was smaller, reported outcomes indicate high to very high patient satisfaction as assessed by patient- and observer-based questionnaires, and the ability to remove several lesions (80 per session) from different parts of the body.¹³ Photodynamic therapy (PDT) has been utilized in the treatment of certain cancers, and is currently being investigated for superficial cNF.¹⁴ PDT consists of a binary process entailing selective accumulation of a photosensitizer followed by its light-mediated activation resulting in the generation of reactive oxygen species and subsequent localized tumor cell destruction. A phase I study for cNF using PDT involving topical application of the known photosensitizer aminolevulinic acid and red light established the safety and maximal tolerated dose for this method (NCT01682811), and provided the basis for a single-site phase 2 study targeting adolescents and young adults (ages 14–30) with cNF (NCT02728388). This age range was selected for the phase 2 study as it is hypothesized that this is a life stage associated with accelerated cNF growth, providing an opportunity to prevent or diminish tumor development. Limitations associated with this approach include pain during treatment, achieving a sufficient depth of penetrance, and limited number of lesions that can be treated per session. In order to address depth of penetrance to reach the deeper portion of the cNF lesions, intralesional injection by microneedles accompanied by delivery-driving technologies like iontophoresis has been considered.¹⁵ In this regard, local drug/device combinations represent an exciting potential approach for cNF. Approaches such as fractional lasers or devices that vaporize tissue to create fenestrations deep into skin to enhance topical delivery are possibilities for therapy delivery to be explored.¹⁶

While procedural methods look promising, they are not risk-free. In the CO₂ laser study, for example, a local infection rate of 15% was reported.¹² As noted by the authors of the study, this might be prevented by prophylactic treatment with a short course of periprocedural oral antibiotic treatment. The approaches described above are all in use or being evaluated at various centers around the world, but limited by a lack of prospective data about the short- and long-term adverse events or efficacy (table 1) of treatment.

Going forward, important goals in the field include generating standardized outcome assessments of these approaches regarding their ability to control cNF growth, understanding patterns of regrowth, effects of tumor-associated symptoms, and effects on neighboring skin, all with awareness of the skin type and body region treated and while assessing short- and long-term patient perceived benefit with patient-reported outcomes.

Experimental drug therapies for cNF

Beyond procedural-based techniques, there have been a few interventional drug trials specifically focused on cNF (table 1). Early work with the histamine 1 receptor inhibitor ketotifen has been explored, exploiting the mast cell stabilization properties of this drug as a means of controlling neurofibroma growth and the associated pain and itch.^17,–,19 This treatment showed modest effectiveness in case series and observational studies, but the generation of rigorous modern clinical trial data will be necessary in order to formally evaluate this therapy for cNF.

To target new vessel growth within the cNF tumor, the vascular endothelial growth factor (VEGF) targeting monoclonal antibody ranibizumab was evaluated in a pilot study of patients with cNF whose lesions were discrete, amenable to measurement with calipers (5–20 mm) and photographs (NCT00657202). The primary objective was to determine the biologic effect of inhibiting local VEGF on tumor volume. The secondary objectives were to identify angiogenic factors upregulated in cNFs treated with the drug and to conduct correlative studies using immunohistochemistry for angiogenic proteins. Data indicated that there was tremendous variability with respect to tumor volume as assessed with calipers over time (both within and across tumors) limiting data interpretation, and the overall effect of treatment was minimal. It was concluded that intralesional injection is an effective approach for studying tumor biology and potentially for therapeutics for specific lesions, but that the priority is development of validated measurement techniques for cNF given the observed variability in size assessment with calipers.

The TLR7 agonist imiquimod (5%) was evaluated in patients with cNF, based on its demonstrated efficacy for actinic keratosis, basal cell carcinoma, and external genital warts (NCT00865644). The primary objective was to assess the effect of topical treatment on tumor volume of cNFs in adults, and the secondary objectives were to correlate inflammatory infiltrate adjacent to treated lesions during treatment with tumor response, and determine the effect on circulating Tregs. By scaling tumors to the mean, a 15% reduction in volume (as measured by caliper) was observed over 4 months, while the control group had an approximately 10% reduction in volume, thus there appeared to be little effect on tumor volume with TLR7 agonism. Importantly, physical examination revealed lower than expected rates of skin inflammation after treatment (5%–10%), which suggests it may be difficult to generate an immune response against NF1 lesions.

Continuously activated Ras driving unrestrained growth of cells is a central feature of NF1, and so components of the Ras-MEK pathway are logical therapeutic targets for NF1-driven tumors.²⁰ Repurposing established drugs or development of novel molecules against new targets in these pathways are both logical strategies. A study exploring the use of topical rapamycin conducted for the treatment of tuberous sclerosis complex on facial angiofibromas was expanded to include patients with NF1 with cNF.²¹ This study showed that topical rapamycin applied daily for 6 months resulted in no detectable systemic absorption of rapamycin, and no significant changes in white blood cell, red blood cell, or platelet counts, indicating that low systemic side effects are to be expected with similar topical therapies.²¹ Subsequently, patients with cNF were enrolled for a study of both systemic and topical rapamycin with cNF lesional improvement as assessed by digital photography as the primary endpoint (NCT01031901). Enrollment is completed and data are expected.

The investigational MEK1/2 inhibitor AZD6244 (Selumetinib) has shown activity in a phase 1 study of children with NF1-associated pNF.²² Inspired by these results, a study evaluating AZD6244 in adults with NF1 and cNF has recently started, with the primary objective of assessing if the drug can shrink cNF tumors as assessed by calipers (NCT02839720).

Areas yet to be explored for cNF are biologically based drugs, although such approaches are under active investigation for pNF.²³ One could also consider cNF-specific antigens as potential targets, and the use of cNF-selective antibodies or CAR-modified T cells configured to recruit the immune system to destroy cNF Schwann cells.²⁴ Another approach may be to reduce cNF-directed immunosuppression by identification of important checkpoint blockers in cNF lesions and application of appropriate immune checkpoint blockers.²⁵ The identification of cNF selective antigens and cNF checkpoint inhibitors would be a key first step in the process. However, short- and long-term toxicities of such approaches must be considered.

Considerations for therapies development for cNF

In considering the relatively limited numbers of patients with cNF, the variability of cNF presentation and symptoms, and the likely need for long-term therapy, there are specific considerations pertaining to cNF drug development, including the following (table 2):

1. Safety: cNF are a major source of morbidity for patients with NF1 but are not associated with mortality. While disfigurement, reduced QoL, and restricted social function are devastating for patients with cNF,^26,27 patients, caregivers, physicians, and regulators may view serious adverse events to be unacceptable for a nonmalignant condition. Thus, new drugs will need to be well-tolerated over long treatment periods with a low frequency of local and systemic side effects.

2. Route of administration: This is an important consideration for cNF, where treatment of hundreds (or thousands) of lesions may be required. While systemic therapy is typically preferred by patients for convenience, this benefit must be weighed against the occurrence of toxicities. Drugs delivered via IV and subcutaneous routes have been approved for other dermatologic conditions; however, these drugs are less convenient for patients and caregivers to administer and are associated with infusion site reactions. Furthermore, such products are very expensive to develop, manufacture, and typically require cold chain storage. Thus there is an overall higher cost of goods. Topical therapies allow localized delivery of a drug to the target lesion, thereby reducing the systemic exposure. However, topical delivery is limited by the cutaneous penetration of the active substance as well as the solubility of the drug in topical formulations. Variable depth of the apparent forms of cNF tumors is another consideration for topical delivery. For example, topical drugs are likely better suited for application to smaller visible areas (i.e., hands, neck, face). For patients with a great many lesions, and in nonvisible areas, a topical delivered medicine may not be ideal.

3. Phenotype: To develop treatments for cNF, the classification, distribution, and size of cNF will be important considerations guiding development of therapeutic approaches.²⁸ Additional factors to consider include anatomic location, cNF-associated symptoms, age, sex, skin type, and hormone effects, as described in table 3.

4. Goal of treatment: As cNF develop over a lifetime, and have variable functional effects and associated symptoms, treatment approaches may have a range of goals, including prevention of cNF development, halting the progression of established cNF, relieving cNF associated symptoms, or regression of the cNF. The goal of the drug being developed should be matched to its mechanism of action and considered throughout preclinical to clinical therapeutic development.

In order to accelerate the development of a drug with the highest likelihood of biologic effect, one promising approach is to test several candidate drugs against several well-defined drug targets via small clinical discovery studies assessing intralesion pharmacokinetic and pharmacodynamic effects across multiple cNF lesion types, skin types, in people of various ages, and in various regions of the body. The goal of this approach is to rapidly rule in or out individual drugs, drug targets, and conditions of use. The best agents from these evaluations could then be advanced to efficacy testing. Such an approach may provide an early proof-of-concept regarding drug mechanism and required concentration, and dosing frequency with low toxicity. Another approach that mitigates the concern about a potential disproportionate risk of systemic therapy (e.g., MEK inhibitor) for benign tumors is to include cNF endpoints in clinical trials designed for other NF1 manifestations such as pNF or malignant peripheral nerve sheath tumors. This is being done to some extent with the ongoing studies assessing MEK inhibitors in adults and children with pNF in which digital photography is incorporated as an exploratory endpoint (NCT03231306, NCT01362803). As ongoing efforts to define the most accurate and reproducible endpoints for cNF through groups like the Response Endpoints in Neurofibromatosis and Schwannomatosis (REiNS) International Consortium yield data, it will be increasingly feasible to incorporate assessment of cNF response into other clinical trials.