Direct Injection of Liposome-Encapsulated Doxorubicin Optimizes Chemomyectomy in Rabbit Eyelid

Abstract

purpose. Doxorubicin chemomyectomy presently represents the only permanent, nonsurgical treatment for blepharospasm and hemifacial spasm. The major deterrent to an otherwise extremely effective treatment protocol is the development in patients of localized inflammation, discomfort, and skin injury over the injection site. As a potential alternative therapy, Doxil (Sequus, Menlo Park, CA), a liposome-encapsulated form of doxorubicin that displays tissue-selective therapeutic effects compared with free doxorubicin, was examined. These effects have been related to its increased retention in tissues and its sustained release over time. For the skin, Doxil is classified as an irritant rather than a vesicant.

methods. Rabbits received direct injections of 1, 2, or 3 mg Doxil alone or in sequence with other agents directly into the lower eyelids. The treated eyelids were examined daily for signs of skin injury. One month after the last injection, the rabbits were euthanatized, and their eyelids were examined histologically for the effect of Doxil on the orbicularis oculi muscle and the skin.

results. At equivalent milligram doses of free doxorubicin, Doxil spared the skin from injury. Doxil was only approximately 60% as effective in killing muscles as the same milligram dose of free doxorubicin. However, either two injections of Doxil spaced 2 months apart or preinjury of the lid with bupivacaine before a single dose of Doxil treatment resulted in increased muscle loss compared with a single dose of Doxil alone and was as effective as free doxorubicin. Higher doses of Doxil did not increase the desired myotoxic effect; apparently, the dose effect levels off at a maximum. Signs of skin injury were minimal; there were small or no adverse skin changes at the maximum effective myotoxic doses.

conclusions. Injection of Doxil resulted in significant reduction of skin injury compared with doxorubicin alone. Although single injections of Doxil were myotoxic, multiple exposure of the eyelid to the liposome-encapsulated form substantially improved myotoxicity while sparing the skin. Repeated doses of the liposome-encapsulated form of doxorubicin may be as clinically effective as free doxorubicin injections and may produce fewer unwanted side effects.

Blepharospasm and hemifacial spasm are muscle spasm diseases that are characterized by forceful, involuntary contractions of the muscles of the eyelids and surrounding face. They result in functional blindness in affected patients, and although not life threatening are seriously debilitating. Blepharospasm has a prevalence of 1:25,000, and hemifacial spasm has a prevalence of between 7.4 (male) and 14.5 (female) per 100,000. ¹ A number of treatment options are available for patients with blepharospasm, hemifacial spasm, and other focal dystonic disorders. Repeated injections of botulinum A toxin and surgical removal of the spastic muscles are the most commonly used treatments. The former is a temporary treatment, but the latter may yield permanent relief or partial amelioration of the muscle spasms. Long-term studies have shown that direct injection of doxorubicin into the eyelid of patients with blepharospasm and hemifacial spasm was an effective treatment for such patients, providing permanent or long-term amelioration of symptoms, best evidenced by the lack of patient requests for short-term rescue therapy with botulinum toxin. ^{2 3} When biopsy samples of eyelid tissue from a patient 2 years after the last doxorubicin injection were examined, there were either small foci of muscle tissue or no myofibers at all in the histologic sections from the treated eyelids. ⁴ Doxorubicin chemomyectomy remains the only permanent, nonsurgical treatment for these patients. As with any medical treatment, there are side effects that are caused by these injections. ^{2 3 5} Although a few patients were treated successfully with a single dose of doxorubicin, most patients required up to three separate injection series in each eyelid for complete localized abatement of the muscle spasms. Each successive exposure to doxorubicin increases the risk of skin injury. The major patient deterrent to the use of local doxorubicin injections into the eyelid for treatment of these muscle spasm diseases is the possible development of local skin inflammation, ulcers, contracted scars, and hyperpigmentation. Of course, localized soreness and skin changes also occur after surgical removal of muscle from the eyelid (orbicularis myectomy), and these changes require secondary surgical touch-ups in some patients. Moreover, some areas of symptomatic periocular facial muscles are not as easily removed with myectomy surgery—particularly, the orbital portions of the orbicularis oculi muscles underlying the eyebrows, the eyelid depressors, corrigators, and procerus muscles. 

We have investigated the ability of a variety of anti-inflammatory mediators, such as cyclosporin ⁶ and corticotropin releasing factor, ^{7 8} to decrease the localized inflammation after doxorubicin treatment. These are effective in reducing either edema ⁶ or localized inflammatory cell infiltration ^{7 8} and in reducing the incidence of skin injury. Although these treatments represent improvements to the free doxorubicin chemomyectomy protocol, we wanted to examine other forms of doxorubicin as they became available. One potential approach to improving doxorubicin chemomyectomy would be to administer doxorubicin in a liposome-encapsulated form. This method of administration has improved treatment effectiveness in certain cohorts of cancer patients by increasing tissue retention while substantially reducing systemic side effects. ⁹ Systemic administration of liposome-encapsulated doxorubicin in specific groups of cancer patients reduced short-term side effects, such as nausea, vomiting, and alopecia, while maintaining therapeutic effectiveness. ¹⁰ Liposome encapsulation of doxorubicin significantly reduced cardiomyopathy and other long-term side effects of systemic application as well ¹¹ and thus may represent a safer alternative than free doxorubicin to patients. Liposome-encapsulated doxorubicin was reported to have few side effects at sites of infiltration of intravenous solutions compared with free doxorubicin. ^{12 13} In fact, regarding its effect on human skin, Doxil (Sequus, Menlo Park, CA), a liposome-encapsulated form of doxorubicin, is classified as an irritant, whereas free doxorubicin is a vesicant. One hypothesis for the increased therapeutic effectiveness of liposome-encapsulated doxorubicin compared with free doxorubicin is prolonged exposure of the tumor to the liposomal doxorubicin because of its local accumulation and slow release at the tumor site. ^{14 15 16}  

Doxil was injected directly into the eyelids of rabbits to determine its chemomyectomy effect after local application. Rabbits were monitored daily for changes in the skin over the injection site. One month after the last treatment, the eyelids were assessed for muscle loss induced by the Doxil treatment and for any changes in the overlaying skin. 

Materials and Methods

The New Zealand White rabbits used for this study were obtained from Birchwood Valley Farms (Redwing, MN) and housed in the Research Animal Resources facility at the University of Minnesota. All research conformed to the guidelines published by the National Institutes of Health and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

The rabbits were anesthetized with an intramuscular injection of ketamine-xylazine, 1:1, (10 mg/kg and 2 mg/kg, respectively). Proparicaine drops were placed in the conjunctival cul-de-sac before eyelid injection. Doxil, liposome-encapsulated doxorubicin, was injected into the lower eyelids. Care was taken to ensure that the injection extended from the medial to the lateral extent of the eyelid. The red coloration of the Doxil solution allows visual confirmation of the injection site. The Doxil was injected in the following doses: 0.5, 1, or 2 mg in 1 ml sterile isotonic saline or 3 mg in 1.5 ml. A second set of animals received two injections of 1, 2, or 3 mg Doxil, administered 1 month apart. A third set of rabbits received two sets of injections into both lower eyelids of 150 U hyaluronidase (Wydase; Wyeth Ayerst, Philadelphia, PA) in 1 ml 0.75% bupivacaine HCl in isotonic saline with 1:200,000 epinephrine (Sensorcaine; Astra Pharmaceutical Products, Westborough, MA) spaced 18 hours apart. ¹⁷ Previous studies have shown that the peak of satellite cell activation in the orbicularis oculi muscle occurs 2 days after a local anesthetic injury to the eyelid, ¹⁸ and that injection of doxorubicin 2 days after injury significantly increases muscle loss. ¹⁹ The two doses of local anesthetic treatment were followed 2 days later with an injection of either 1 or 2 mg Doxil. Animals were examined daily for skin injury. At least four eyelids were examined for each dose of Doxil and for each distinct set of treatment parameters. Doxil data were compared with doxorubicin-only data prepared with the identical injection and analysis procedures. ^{6 7} Additional control eyelids were prepared by injection of saline only. 

One month after the final injections, the animals were euthanatized with an overdose of barbiturate anesthesia. Samples were removed from the lateral, middle, and medial portions of the treated and control lids and frozen in 2-methylbutane chilled to a slurry on liquid nitrogen. Cross-sections were cut at 12 μm, and the tissue sections were stained by myosin adenosine triphosphatase (ATPase) histochemistry to detect type 1 and type 2 muscle fibers. Additionally, eyelid cross-sections were stained with hematoxylin and eosin so that the skin epithelium nuclei could be examined for precancerous cellular atypia. 

Muscle fiber number and cross-sectional areas were determined by computer-assisted morphometric analysis using commercial software (Bioquant; R & M Biometrics, Nashville, TN). ²⁰ Results were compared with control data from doxorubicin-only injections administered in the same manner. ^{6 7} All injection and analysis methods were unchanged from the original free doxorubicin studies, and all injections were administered by the principal investigator. Statistical significance was determined using an unpaired, two-tailed t-test using statistical analysis software (Prism and Statmate; Graphpad, San Diego, CA), with significance at P < 0.005. An F-test indicated that the variances of the control and experimental groups were not significantly different. 

Results

Safety

Injection of free doxorubicin alone resulted in localized inflammation characterized by edema, erythema, and the development of a skin ulcer followed by scab formation in 100% of treated rabbits. ^{6 7} None of the eyelids that received injections of 0.5 or 1 mg Doxil into their lower eyelids showed any sign of skin surface breakdown (Fig. 1) . At a dose of 2 mg Doxil, only two of eight eyelids injected showed a small scab. One appeared on day 8 and was present for 3 days, and one appeared on day 14 and was no longer visible the next day. Neither of these scabs was preceded by a recognizable blister or ulcer. The additional 6 eyelids showed no signs of skin injury. These results were in marked contrast to the extent and duration of eyelid ulcers and skin crusting after free doxorubicin injections of 2 mg, a higher dose than the recommended therapeutic dose in patients. After free doxorubicin injections of 2 mg, skin ulcers and scabs were large and routinely lasted up to 1 month. ^{6 7} At a dose of 3 mg Doxil, a dose too toxic to the skin to administer as free doxorubicin, one of the five treated eyelids showed, for only 1 day, a small scab between 2 and 3 mm in diameter. When two sets of injections of Doxil were administered 1 month apart, there was no increase in the likelihood of development of injury of the eyelid skin compared with the single doses of Doxil. Moreover, the eyelids on which scabs developed the first time did not have a scab after the second injection. There was some edema in the treated lids, and as has been seen with local free doxorubicin injections in the eyelid, eyelid hair loss but not eyelash loss was seen after the higher doses of Doxil. The remainder of the eyelid anatomy appeared normal, including the skin, nerves, and tarsal glands (Figs. 2 and 3) . Histologic examination of the eyelid skin epithelium showed no signs of atypia, often a prelude to carcinogenesis. 

None of the treated eyelids showed scabs after bupivacaine preinjury of the orbicularis oculi muscle 2 days before Doxil injection, a method previously demonstrated to improve substantially the myotoxicity of free doxorubicin. ¹⁹  

Effectiveness

Single injections of Doxil were approximately 60% as myotoxic to the myofibers of the orbicularis oculi compared with single injections of the same dose of doxorubicin (Figs. 3 and 4) . Whereas single free doxorubicin injections demonstrated a dose-dependent myotoxicity, single Doxil injections resulted in the same myotoxic effect at 0.5-, 1-, and 2-mg injection doses. Only at the 3-mg Doxil injection dose was there a slight increase in myotoxicity compared with the other doses, but the increase was not significant. 

Increased myotoxicity was seen when 1 or 2 mg Doxil was administered at two separate times spaced 1 month apart compared with single injections (Fig. 5) . At both 1 and 2 mg Doxil these differences were statistically significant. Preinjury of the orbicularis oculi with bupivacaine 2 days before Doxil treatment significantly improved muscle loss over single injections of Doxil alone. Injection of 1 mg free doxorubicin alone resulted in loss of approximately 74% of the myofibers and 2 mg free doxorubicin alone resulted in loss of approximately 87% of the myofibers. Single-injection comparable doses of Doxil resulted in loss of approximately 60% of the treated myofibers. Treatment with either preinjury with bupivacaine before a single Doxil injection or two injections of Doxil spaced 1 month apart both resulted in loss of approximately 74% of the treated myofibers, essentially the same result as was obtained with 1 mg doxorubicin alone, except that the skin was spared from injury. 

Discussion

Overall, Doxil, a liposome-encapsulated form of doxorubicin, when injected directly into the rabbit eyelid, substantially increased the dermatologic safety of chemomyectomy without much sacrifice in the efficacy of muscle loss from each injection session compared with a 1-mg dose of doxorubicin alone. Moreover, Doxil treatment at the maximum effective dose resulted in little or no skin injury. Although the myotoxicity of Doxil was improved by either serial injections of Doxil or by pretreatment of the eyelid with bupivacaine 2 days before Doxil injection, chemomyectomy was never as complete as that seen with doxorubicin alone. 

The liposome-encapsulated doxorubicin formulation substantially alters the toxic and therapeutic profile of doxorubicin. Liposome encapsulation lessens the toxicity of doxorubicin, reducing cardiomyotoxicity, which is the major dose-limiting consideration in the use of free doxorubicin as a first-line cancer chemotherapeutic agent. ²¹ Liposome encapsulation also reduces the acute adverse effects seen with free doxorubicin that lead to ulceration and scarring at sites of inadvertent extravasation at percutaneous intravenous sites. In fact, this has led to the administration of doxorubicin by central venous catheter, whereas Doxil is routinely administered by percutaneous injection. Doxil does not possess the vesicant properties of doxorubicin and has been demonstrated to be safer for the skin. ²²  

Skin Toxicity

Local injection of free doxorubicin produces acute skin toxicity ^{3 4} that is substantially reduced when an equal dose is injected in a liposome-encapsulated form. This study examined both the acute inflammatory injury and the possibility of cellular atypia after direct local injections of Doxil into rabbit eyelid. Although milligram-equivalent doses of liposome-encapsulated doxorubicin were less myotoxic than free doxorubicin alone, the sparing of the skin over the injection site represents a major advantage to the use of liposome-encapsulated doxorubicin for localized injection. After free doxorubicin injections at doses of 1 or 2 mg in rabbits, all eyelids showed skin injury, and the scabs that developed remained for up to 1 month. ^{6 7} Because of the development of injury to the skin, in the clinical protocol the threshold toxic dose of doxorubicin is 1.5 mg per eyelid per injection session. ³ At milligram-equivalent doses of Doxil, no skin injury developed in the treated rabbits. Even at twice the clinical toxic dose of free doxorubicin of 3 mg or after serial injections, if scabs developed, they were small and disappeared within 2 to 3 days. It is noteworthy that the second of two injections of Doxil were not associated with skin injury, even if skin injury occurred after the first injection. This protective effect is in direct contrast to the effects of multiple exposure of the eyelid skin to free doxorubicin, in which each subsequent exposure increased the risk of skin injury. ^{2 3 5}  

The relative absence of injury to rabbit eyelid skin in the present application is interesting, because the dose-limiting toxicity for high systemic doses of liposome-encapsulated doxorubicin is cutaneous erythema and hyperemia. ^{21 23} When Doxil is administered intravenously in humans, the drug is selectively extravasated into the tumor tissues and into the palms of the hands and the soles of the feet where it can cause an erythematous reaction and scaling. We did not observe these changes in the rabbits at any time after Doxil injection. 

Although a long-term assessment of Doxil is beyond the scope of this study, it is important to note that there have been no published reports of Doxil-induced skin cancer, even after high-dose systemic use for cancer treatment. In the patients who received doxorubicin chemomyectomy, many of whom have now had more than 10 years elapse since treatment, no precancerous skin changes have been seen in the regions over the injection sites. Furthermore, a number of studies have been published indicating the efficacy of systemic injections of either doxorubicin or Doxil specifically as an anticancer drug to shrink skin cancer ^{24 25} and other solid tumors. ²⁶ Doxil, in its use as a systemic high-dose chemotherapeutic agent, has fewer side effects than does free doxorubicin. ²¹ Free doxorubicin, although used primarily as a chemotherapeutic agent, may have carcinogenic potential, and this potential is the basis for the recommendations for the safe handling of doxorubicin and other anthracyclines by health care personnel. In spite of perhaps thousands of episodes of extravasation, we found only one report in the literature of human skin cancer that occurred 10 years after an untreated ulcer caused by extravasation of doxorubicin during chemotherapy. ²⁷ There is evidence that encapsulation of potential mutagens markedly diminishes and even abolishes their genotoxic effects. ²⁸  

We have now locally injected two patients affected by blepharospasm with Doxil, three times each in the same eyelids. Both patients have had no skin ulceration or blisters over the injection sites at any time after the Doxil injections. 

Chemomyectomy

Single local injections of Doxil into the eyelid were not as myotoxic as the milligram-equivalent dose of free doxorubicin. The maximal dose for a single injection of free doxorubicin for patients is 1.5 mg per eyelid per injection, which is insufficient for total relief of muscle spasms in most patients. ³ The patients have usually required one or two additional injection sessions, for an average cumulative dose of approximately 3 mg per eyelid for complete abatement of muscle spasms. To minimize skin injury, botulinum toxin injections are incorporated into the 2- to 3-month cycle, and the free doxorubicin treatment is administered over the course of 6 to 9 months. Serial injections of Doxil and preinjury followed sequentially by Doxil treatment increased myotoxicity over single injections of Doxil alone, and the skin was largely spared from injury. Thus, Doxil should be as effective a chemomyectomy agent as free doxorubicin but with a reduced incidence of the concomitant skin changes. It should be pointed out that in this study, Doxil was administered locally within the eyelid, not systemically into the blood stream, as it is when used for chemotherapeutic purposes. Although we did not study empty liposomes as a control, we note that a number of studies have demonstrated that empty liposomes injected intramuscularly do not cause muscle tissue damage. ^{29 30}  

Mechanism of Action

The mechanism of cellular uptake of liposome-encapsulated drugs is not known. ¹⁶ Liposome-encapsulated doxorubicin is composed of sterically stabilized liposomes, which improves stability and drug retention. ³¹ Liposomes slowly disintegrate, and this increases the length of drug exposure at the site of liposome accumulation. Subcutaneous administration of liposomes, in particular, results in more sustained drug release, particularly for drugs such as doxorubicin, which have a very short half-life as free drug. ³² Of note, by themselves empty liposomes of the same composition as used in the present study do not induce myotoxicity when directly injected into muscles. ^{29 30} Liposomes themselves are not taken up intracellularly by fusion with plasma membranes. ³³ Recent studies also suggest that sterically stabilized liposomes are not taken up by cells through endocytosis. ^{14 34} Their enhanced cytotoxic effects appear to be caused by increased concentrations in the interstitial fluid at the site of their accumulation and their ability to display sustained release over time. ^{16 29} Yet, at the same time, the liposomal encapsulation protects the surrounding tissue from the type of damage seen after injection of free doxorubicin. ³⁰ The local injection protocol of Doxil reduces the potential for adverse systemic side effects compared with systemic administration. The potential movement of Doxil from the site of local administration into the eyelid is currently under investigation. 

The direct injection of Doxil could represent a major improvement to the doxorubicin chemomyectomy protocol for the treatment of blepharospasm and hemifacial spasm. Doxil treatment spared the skin from the severe type of inflammation and development of scabs that may occur after the injection of free doxorubicin directly into the eyelid. Subsequent exposure of a previously treated eyelid did not increase the risk of the eyelid to the development of skin injury. Although Doxil was less myotoxic to the orbicularis oculi muscle than was free doxorubicin, it still caused a significant loss of muscle compared with that in the control eyelids. By preinjuring the eyelid muscle with bupivacaine or treating the eyelid with more than one injection session of Doxil, myotoxicity was the same as the results after lower doses of free doxorubicin. The ability to maintain a high degree of muscle loss while minimizing or preventing skin injury represents a major advantage to the patients with these conditions. Sparing the eyelid skin from inflammation and skin injury should improve patient acceptance of this new treatment protocol. 

 Supported by Grant EY07935 from the National Institutes of Health, the Minnesota Lions and Lionesses, the Frank Burch Chair in Ophthalmology (JDW), and an unrestricted grant to the Department of Ophthalmology from Research to Prevent Blindness.

 Submitted for publication December 29, 1998; revised April 29, 1999; accepted June 21, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

Onset and duration of detectable skin injury in all treated eyelids after Doxil (Sequus, Menlo Park, CA) injections; 2×, two injection sessions spaced 1 month apart tabulated after the second injection; preinj, injection of bupivacaine 2 days before injection with Doxil.* At this dose, no skin injury developed on eyelids; # the results after the second injection only; + all first injections at this dose.

Figure 2.

 

The skin (A) and nerves (B) within the treated eyelids showed normal morphology 1 month after 2 mg Doxil (Sequus, Menlo Park, CA) treatment. Arrowheads indicate basal lamina of the skin. Arrows indicate nerves in longitudinal section. Bar, 100 μm.

Figure 3.

 

Orbicularis oculi muscle in the untreated (A) and Doxil (Sequus, Menlo Park, CA)-injected (B) eyelid cross-sections. Doxil injections caused a significant loss of myofibers (arrows) in the treated orbicularis oculi muscle. No pathologic changes were seen in the other eyelid structures, including the skin and tarsal glands. Bar, 50 μm.

Figure 4.

 

Muscle loss in orbicularis oculi after a single injection of doxorubicin or Doxil (Sequus, Menlo Park, CA). Chemomyectomy effects of single injections of Doxil are compared with those of free doxorubicin ^{6 7} injections based on muscle fiber counts. Single injections of 1 or 2 mg Doxil were significantly less myotoxic than equivalent doses of free doxorubicin (**). *This injection dose was not used.

Figure 5.

 

Chemomyectomy effects at 1- and 2-mg injections of free doxorubicin (dxn) ^{6 7} ; 1, 2, or 3 mg Doxil (Sequus, Menlo Park, CA); eyelids preinjured with bupivacaine (Doxil/preinj); and two sequential injections of Doxil (Doxil/2×) compared with control. *Statistically significant difference when compared with single injection of liposome-encapsulated doxorubicin; **statistically significant difference from all other treatment parameters; dxn, doxorubicin; 2×, two injection sessions spaced 1 month apart; preinj, injection of bupivacaine 2 days before injection with Doxil.