Dr Bergman is presently affiliated with Bright Heart Veterinary Centers, Armonk, NY. Presented in part at the 28th Annual Conference of the Veterinary Cancer Society, Seattle, Washington, October 2008.
Xenogeneic Murine Tyrosinase DNA Vaccine for Malignant Melanoma of the Digit of Dogs
Article first published online: 8 DEC 2010
Copyright © 2010 by the American College of Veterinary Internal Medicine
Journal of Veterinary Internal Medicine
Volume 25, Issue 1, pages 94–99, January/February 2011
How to Cite
Manley, C.A., Leibman, N.F., Wolchok, J.D., Rivière, I.C., Bartido, S., Craft, D.M. and Bergman, P.J. (2011), Xenogeneic Murine Tyrosinase DNA Vaccine for Malignant Melanoma of the Digit of Dogs. Journal of Veterinary Internal Medicine, 25: 94–99. doi: 10.1111/j.1939-1676.2010.0627.x
- Issue published online: 11 JAN 2011
- Article first published online: 8 DEC 2010
- Submitted January 13, 2010; Revised May 23, 2010; Accepted September 13, 2010.
- Melanoma staging;
- Tumor immunotherapy
Background: Malignant melanoma of dogs is a highly aggressive neoplasm and is the 2nd most common digit tumor. Metastatic disease is a common sequela for which few effective treatment options exist. Studies show that xenogeneic tyrosinase DNA vaccination yields immune responses and prolongation of survival in dogs with oral malignant melanoma.
Objectives/Hypothesis: Describe clinical findings and tumor characteristics of a cohort of dogs with digit malignant melanoma, and evaluate the prognostic utility of a proposed staging system. Determine if a novel xenogeneic DNA vaccine is safe and potentially effective for treatment of dogs with digit melanoma.
Animals: Fifty-eight dogs with digit malignant melanoma treated at the Animal Medical Center between 2004 and 2007.
Methods: Retrospective, medical records review of dogs with digit melanoma treated with xenogeneic DNA vaccine.
Results: Overall median survival time (MST) for dogs treated with loco-regional control and xenogeneic DNA vaccine was 476 days with a 1-year survival rate of 63%. MST for dogs presenting with metastasis was 105 days versus 533 days for dogs presenting without metastasis (P < .0001). Forty-eight percent of the dogs in the latter group were alive at 2 and 3 years. A proposed staging system proved prognostic with stages I–IV dogs surviving >952, >1,093, 321, and 76 days, respectively.
Conclusions and Clinical Importance: The xenogeneic murine tyrosinase DNA vaccine was safe and appears effective when used in conjunction with local and regional disease control. The proposed staging system was prognostic in this study and future studies might benefit from utilizing this staging system.
Malignant melanoma in dogs is a highly aggressive and frequently metastatic neoplasm that represents approximately 4% of all tumors in dogs.1 Commonly affected sites include the oral cavity, mucocutaneous junction, nail bed, and footpad. Cutaneous melanomas that are not in proximity to mucosal margins often behave in a benign manner. Melanoma is the 2nd most common tumor affecting the digit.1–3 Melanomas of the digit, specifically those of nail bed origin, often have multiple histopathologic criteria of malignancy.4,5 While histologic features have not been strongly correlated with outcome in dogs with digit malignant melanoma, these tumors commonly display a high degree of local invasiveness and a high metastatic propensity.4,5 Upon diagnosis, the metastatic rate is 32–40% and subsequent local and distant metastasis after amputation of the digit is a common occurrence.2,3
Dogs treated with surgery (digit amputation or lumpectomy) alone have survival times of approximately 12 months.2,6 The 1- and 2-year survival rates are 42–57 and 11–36%, respectively.2,3,6,7 While a variety of chemotherapeutic agents have been attempted to delay the spread of disease, including dimethyl triazeno imidazole carboxamide, melphalan, doxorubicin, and cyclophosphamide, little data exist to demonstrate the effectiveness of adjuvant therapy beyond surgery for melanoma of the digit.3 Furthermore, other forms of malignant melanoma in both dogs and humans have responded poorly to chemotherapy, holding little promise for its future efficacy.8–10
A murine xenogeneic DNA vaccine for canine oral malignant melanoma was recently evaluated.11,12 The vaccine targets the melanosomal glycoprotein tyrosinase, which is normally expressed on melanocytes and is essential in melanin synthesis. Immunization with xenogeneic DNA encoding tyrosinase family proteins has been shown to induce antibody, T cell, and antitumor responses in mice and a small cohort of dogs.13–15 With the addition of vaccine to local-regional control, survival times appear to have increased with no major toxicoses.11,12 The human variant of the xenogeneic vaccine (Oncept) has recently received full licensure for oral stages II and III melanoma.
The purpose of this retrospective study was to determine if this novel xenogeneic DNA vaccine is safe and potentially effective for dogs with digit malignant melanoma. Additionally, the clinical findings and tumor characteristics of this group of dogs were summarized, and a proposed staging system for malignant melanoma of the digit was evaluated for prognostic utility.
Materials and Methods
The medical records of 58 dogs with histologically confirmed malignant melanoma of the digit treated with xenogeneic murine tyrosinase DNA vaccine at The Animal Medical Center between March 2004 and April 2007 were reviewed. Dogs included for evaluation had a histologically confirmed diagnosis of malignant melanoma affecting the dermis of the digit or the nail bed and received at least 1 vaccination. Dogs were excluded if their tumors originated from the footpad or interdigital region.
Pretreatment evaluation included complete physical examination, a CBC, serum chemistry profile, lactate dehydrogenase (LDH), and antinuclear antibody (ANA), and these diagnostics were repeated monthly during vaccination. LDH was evaluated because of its prognostic nature in the human form of the disease and ANA was monitored to ensure the vaccine did not incite an autoimmune reaction in patients. Staging diagnostics including 3 view radiographs of the thorax and regional lymph node aspirates or biopsy specimens or both were performed at the time of 1st vaccination and were repeated at monthly intervals for the first 2 months and every 3 months thereafter. Patient characteristics including age, breed, weight, and sex, tumor characteristics including site, size, and tumor type (melanotic versus amelanotic), stage based on the World Health Organization (WHO) Tumor, Node, Metastasis guidelines for digit tumors (Table 1), and vaccine adverse effects were also recorded.16 The WHO clinical staging system for tumors of the oral cavity was applied to this population to evaluate for prognostic significance (Table 2) given that there is no accepted clinical staging system for digit tumors or digit melanomas. Written consent was obtained from each dog's owner before subjects receiving the vaccine.
|T: Primary tumor|
|T1||Tumor <2 cm in diameter superficial/exophytic|
|T2||Tumor 2–5 cm in diameter minimum invasion|
|T3||Tumor >5 cm or invading subcutis|
|T4||Tumor invading fascia or bone|
|N: Regional lymph nodes|
|N0||No evidence of tumor found|
|N2||Moveable contralateral or bilateral|
|(1) No tumor cells found|
|(2) Tumor cells found|
|M: Distant metastasis|
|M1||Distant metastasis found specify site(s)|
|IV||Any T||Any N||M1|
The murine tyrosinase cDNA was previously cloned and sequenced at Memorial Sloan-Kettering Cancer Center.17 The DNA has been inserted in the pING plasmid vector, which contains a cytomegalovirus promoter and kanamycin resistance selection marker. The vaccine was produced and released from the Gene Transfer and Somatic Cell Engineering Facility at the Memorial Sloan-Kettering Cancer Center with permission from the U.S. Department of Agriculture.
Each dog received either 50 mcg (n = 57) or 100 mcg (n = 1) of xenogeneic plasmid DNA encoding murine tyrosinase IM every 14 days for a total of 4 vaccinations in the left caudal thigh with the Vitajeta spring-loaded needle-free injection device. A booster vaccine was administered every 6 months thereafter.
Survival time was defined as the time from receiving the 1st xenogeneic murine tyrosinase DNA vaccination until death. Overall survival time was defined as the survival time from histopathologic diagnosis until death. Vaccine delay time was defined as the time period from diagnosis until the vaccine was started. Median survival times (MST) were calculated via the Kaplan-Meier life table analysis with Statview statistical analysis software.b Recorded variables were evaluated for their effect on survival time by log-rank analysis and both univariate and multivariate Cox's proportional hazards analysis. Dogs were censored if alive at data collection, lost to follow-up, or died due to nonrelated disease. A P-value of ≤.05 was considered significant.
Fifty-eight dogs diagnosed with malignant melanoma of the digit were evaluated. The dogs ranged in age from 3.5 to 16.5 years, with a median age of 9.6 years and mean age of 9.7 years, respectively. The cases were evenly distributed between males and females with 30 spayed females, 26 castrated males, and 2 intact males. Upon diagnosis, the median and mean weight was 30.6 and 30.5 kg, respectively, with a range from 2.8 to 54.5 kg. Fifty-three purebreds and 5 mixed breed dogs were recorded. The most frequently represented breeds seen in this study included the Labrador Retriever (n = 13), Golden Retriever (n = 10), Rottweiler (n = 6), and Scottish Terrier (n = 5).
Three of the 58 tumors (5%) were histologically classified as amelanotic. Thirty-four tumors were located on the forelimbs (left—22, right—12) and 24 tumors were located on the hindlimbs (left—14, right—10). No tumors were located on the 1st digit, and the remaining digits were equally represented with 15 tumors affecting digits 2–4, respectively, and 13 tumors affecting digit 5. Thirty-six tumors were located in the subungual region and 22 tumors were located on the associated dermis of the digit.
Local tumor stage T, defined as tumor size, depth, and bony involvement, was available for 50 of 58 tumors.16 There were 13 T1 tumors, 9 T2 tumors, and 28 T4 tumors. Bony invasion diagnosed radiographically or histologically or both, was present in 28 of 48 tumors (58%) available for analysis. Metastasis was reported in 16 (27%) dogs upon presentation representing nodal alone (n = 5), distant sites alone (n = 5), and both nodal and distant sites (n = 6). Margin analysis was available for 55 tumors after amputation. Incomplete margins were noted in 23 dogs initially and 10 dogs went on to have a surgical revision, which ultimately led to clean margins in those dogs. Of the remaining 13 dogs, 7 dogs presented with local metastases (n = 2), distant metastases (n = 4), or both (n = 1) and local reexcision was not recommended. Only 1 dog had macroscopic disease at the primary tumor site at the time of 1st vaccination.
Fifty-seven of 58 dogs were treated with digit amputation for local-regional control. Radiation therapy was administered to 2 dogs, 1 with an incompletely excised digit tumor and 1 with an incompletely excised metastatic prescapular lymph node. Hypofractionated radiation was delivered on a weekly basis with a Cobalt-60 unit for a total dose of 24 or 32 Gy (3–4 fractions, respectively). Before presentation, 3 dogs received between 2 and 5 doses of Carboplatin. The 1st dog, an 11-year-old male castrated Rottweiler, had previously received 2 doses of Carboplatin. He presented without metastatic disease and died 7 days after start of vaccine therapy because of a gastric dilatation volvulus. The other 2 dogs, a 10-year-old female spayed Scottish Terrier and a 10-year-old female spayed Cocker Spaniel, both previously received 5 doses of Carboplatin. They both presented without metastatic disease; the former was alive 1,093 days after start of vaccine therapy and the latter died 418 days after the start of vaccine therapy because of progressive disease. There were no systemic toxicoses observed from the vaccine, and only mild local pain responses at the vaccination site were noted in a minority of cases (n = 9).
The MST, independent of stage, metastasis, and other factors not controllable at presentation, from the date of 1st vaccination was 351 days as determined by Kaplan-Meier life table analysis with 50% alive at 1 year and 36% alive at 2 and 3 years (Fig 1). The overall MST from the date of diagnosis was 476 days with 63% alive at 1 year and 32% alive at 2 and 3 years (Fig 2). The median and mean follow-up time from date of 1st vaccination was 241 days and 334 days, respectively (range = 7–1,052 days).
Variables identified by univariate Kaplan-Meier life table analysis and Cox's proportional hazards model that were prognostic for survival are outlined in Table 3. Dogs that presented without lymph node or distant metastasis had a statistically significant survival advantage over dogs presenting with metastasis with a MST of 533 days and 105 days, respectively (log rank P < .0001—Cox's PH P < .001). Forty-eight percent of the dogs without metastasis were alive at 2 and 3 years after 1st vaccination. Specifically, dogs that presented with nodal metastasis had a MST of 110 days versus 459 days for dogs presenting without nodal metastasis (log rank P= .0019—Cox's PH P= .0034). Similarly, dogs that presented with distant metastasis had a MST of 76 days and dogs without distant metastasis had a median survival of 459 days (log rank P < .001—Cox's PH P < .001).
|Variable||N||MST (days)||P-Value Cox's PH versus Log Rank||HR||95% CI|
|Amelanotic (all stage III or IV)||3||158||.0349||3.7||1.1−12.7|
|Metastasis upon presentation|
|Regional lymph node metastasis|
|Presence of macroscopic disease (local or distant) at first vaccination|
|Development of distant metastasis after completion of vaccine|
|Completion of vaccine protocol|
|Vaccine delay time||.0043||1.0||1.001−1.007|
The development of metastasis during or after the vaccine protocol was also significantly associated with survival. Dogs that developed metastasis after the initial vaccine protocol (n = 29) lived a median of 238 days versus those dogs that did not develop metastasis (n = 29) with a MST of ≥1,152 days (median not yet reached, log rank P= .0002—Cox's PH P= .0006). The median time to develop metastases for this cohort was 62 days (range = 13–423 days). Additionally, if they did not complete the protocol (n = 9) their survival time was significantly shorter than those that did, with an MST of 63 days and 418 days, respectively (log rank P= .0003—Cox's PH P= .0009). Moreover, when a significant delay occurred between diagnosis and the start of vaccine treatment (vaccine delay time), dogs had an increased risk of death versus those that started vaccine immediately (Cox's PH P= .0043, HR = 1.004, 95% CI = 1.001–1.007). The mean and median time to start vaccine was 70 days and 32 days, respectively, with a range of 7–434 days.
Stage, as extrapolated from the WHO oral tumor staging system, was statistically significant with survival time (Fig 3). Dogs presenting at time of 1st vaccination with stage I disease (n = 12) had an MST of >952 days (median not yet reached), dogs presenting with stage II disease (n = 7) had an MST of >1,093 days (median not yet reached), dogs with stage III disease (n = 29) had an MST of 321 days, and those with stage IV disease (n = 10) had an MST of 76 days.
Dogs that presented with amelanotic tumors (n = 3) had survival time of 158 days versus 418 days for those with melanotic tumors (Cox's PH P= .0349—log rank P= .0233, HR = 3.7, 95% CI = 1.098–12.719). One of these dogs presented with stage III disease and the other 2 dogs presented with stage IV disease. Results should be interpreted in light of this fact.
Stepwise multivariate Cox's proportional hazards analysis revealed the presence of metastasis at presentation (P= .0275, HR = 8.9, 95% CI = 1.27–62.5), development of metastasis after completion of vaccine (P= .0007, HR = 8.6, 95% CI = 2.46–30.3), and lack of completion of the vaccine protocol (P < .0001, HR = 28.5, 95% CI = 5.92–137.14) to all be statistically significant independent prognostic variables for survival.
This retrospective study suggests prolongation of survival for dogs with melanoma of the digit treated with xenogeneic DNA melanoma vaccine compared with historical controls treated with surgery alone (MST 365 days).2 While the survival time from 1st vaccination to death (351 days) was comparable to historical controls (365 days), the overall survival time from histopathologic diagnosis to death (476 days) was extended compared with previous studies with a 1-year survival rate of 63% and a 2 and 3-year survival rate of 32%.2,3 A recent study which described a population with a similar stage distribution as this study (40% metastasis upon presentation and 40% T4 tumors) reported a 1-year survival rate of 44% and a 2-year survival rate of 11%.3
Given that the vaccine was not commercially available at the time these dogs were treated, one third of the cases had a delay of 2 months or longer (up to 14 months) before receiving the vaccine. Furthermore, as the number of days before vaccine treatment increased, there was an increased risk of death for those dogs. As a result, the delay in vaccination might account for the modest improvement in survival time over historical controls treated with surgery alone.
Additionally, the proposed staging system adapted from the oral tumor staging system proved prognostic. Dogs with advanced stage disease or macroscopic disease or both on presentation had a significantly worse outcome compared with those dogs with locally controlled stage I or II disease, with MSTs > 2.5 years. Likewise, those dogs that did not complete the vaccine protocol, most often because of progressive disease, fared significantly worse than those that completed the protocol. This not only implies that early detection and treatment are critical, but also that the vaccine may be more efficacious against micrometastatic disease. Moreover, the immune response created by the vaccine may be overwhelmed by a large tumor burden.
Similar to previous reports, dogs with malignant melanoma of the digit in this study were primarily large breeds and the majority were >20 kg.3 Also, comparable to other findings, tumors equally involved digits other than the 1st and the forelimbs were more frequently affected than the hindlimbs.3,6 Both increased carcinogen exposure and increased weight-bearing function of the forelimbs have been proposed mechanisms for the increased predilection at this site.3
In this study, approximately 58% of tumors invaded into the bone, in contrast to previous studies, which had bone invasion in only 5–40% of cases.2,6 The difference might be because of the small sample size in these studies, but is also likely a result of our inclusion criteria. Only melanomas affecting the subungual region or the dermis of the digit were included, excluding the pad and the interdigital region. Previous reports have included these latter regions in which tumors located here would be less likely to invade into the bone.2,6 Dogs with melanoma of the footpad, nail bed, and cutaneous digit regions were recently evaluated. Only 1 dog of 7 (14%) with footpad melanoma died of their disease while 4 of 9 (44%) of dogs with subungual melanoma died of their disease. One dog was identified with a malignant cutaneous digit melanoma, which behaved aggressively and led to that dog's death.5 Therefore, it is possible that we accrued a more malignant tumor population. In this report there was no statistical difference in outcome between melanomas affecting the subungual region or those affecting the dermis of the digit.
While the dogs with amelanotic tumors had a statistically significant disadvantage in outcome over the dogs with melanotic tumors, results were confounded by those cases presenting with higher stage disease. Whether this result is due to type I error or whether amelanotic tumors are in fact more aggressive and therefore are more likely to present with higher stage disease would be best evaluated in a prospective manner. Amelanotic tumors in other locations (ie, oral) have not been found to behave more aggressively and therefore the latter is less likely.18
Although there was not a statistically significant difference in outcome between dogs with incomplete surgical margins and those with complete margins, achieving good local control appears beneficial for long-term control. Of the 16 dogs that presented with local or distant metastases, approximately half (43%) of those dogs had incomplete margins on presentation. The MST for those 13 dogs that never achieved a complete excision was only 105 days. While presentation with metastatic disease likely precluded a surgical revision from taking place in many of these cases, it is also possible that the inability to obtain good local control early on in the disease process correlates with the onset of early metastatic disease.
One limitation of this study owing to its retrospective nature was that a single pathologist did not evaluate all tumor samples and therefore variation regarding histologic features of malignancy, invasiveness, and bony involvement may exist. However, conflicting studies regarding whether malignant features are predictive of clinical behavior with digit malignant melanoma suggest that testing beyond routine light microscopy is necessary. Negative histologic determinants for survival in dogs with malignant melanomas affecting the feet and lips were recently evaluated and findings suggest that mitotic index, nuclear atypia, tumor score, inflammation, necrosis, and junctional activity are prognostic for survival.5 However, a separate study reports that histologic subtype and histologic features are not associated with survival in dogs specifically with distal extremity malignant melanomas.7 While a further study found that all nail bed melanomas had histologic features of malignancy but these features did not correlate with outcome.4 Given the predictive value of Ki67 for canine cutaneous melanomas, evaluation of proliferative indices for digital melanomas may prove prognostic.19 Furthermore, these indices may isolate lesions that more closely resemble melanocytic nevi. Closer evaluation of depth of invasion and ulceration, both features highly prognostic in human melanoma, may be helpful in identifying patients requiring more aggressive treatment and monitoring.
The xenogeneic vaccine used in this study was of murine DNA origin unlike the current commercially available product, which is of human DNA origin. However, the authors believe that the immune response and therefore response to vaccine treatment should be the same. It has been shown that vaccination of mice with DNA encoding cancer differentiation antigens is ineffective when using syngeneic DNA as a result of self-tolerance; however, tumor immunity is induced when xenogeneic DNA is used.13 A prospective study with the commercially available product would be beneficial to further investigate this. Additional studies to evaluate this novel therapeutic with malignant melanoma affecting other sites and species are also warranted.
In conclusion, these pilot retrospective data demonstrate that the xenogeneic murine tyrosinase vaccine is safe and suggests that the vaccine has efficacy when used in conjunction with local-regional control for dogs with digit melanoma. Additional controlled prospective studies are recommended to further evaluate efficacy. Based on our results, a survival advantage was noted for those dogs that were vaccinated near the time of diagnosis over those dogs that had a significant delay between diagnosis and vaccination. Therefore we recommend commencing vaccination as early as possible. Extrapolation of the oral staging system to digit melanomas was prognostic in this study and future studies may benefit from utilizing this staging system.
aVitajet, Bioject Inc, Tualatin, OR
bStatview statistical package, SAS Institute, Cary, NC
The authors thank Drs Brenn, Gill, Finora, Taylor, Camps, McKnight, and Farrelly for their contributions to these cases.
Conflict of interest: The senior investigator (Philip J. Bergman) reports that he consults and speaks for a number of animal health companies including the manufacturer of Oncept (Human Tyrosinase DNA vaccine; Merial Inc). Furthermore, he receives royalties associated with a xenogeneic patent, which has been licensed by Merial with distribution through the Animal Medical Center.
- 3Canine digital tumors: A veterinary cooperative oncology group retrospective study of 64 dogs. J Vet Intern Med 2005;19:720–724., , , et al.
- 7Distal extremity melanocytic nevi and malignant melanomas in dogs. J Am Anim Hosp Assoc 1990;26:605–612.,
- 14Vaccination with human tyrosinase DNA induces antibody responses in dogs with advanced melanoma. Cancer Immun 2007;6:1–10., , , et al.
- 16TNM Classification of Tumors in Domestic Animals. Geneva, Switzerland: World Health Organization; 1980.
- 18Of mice and men (and dogs): Development of a xenogeneic DNA vaccine for canine oral malignant melanoma. Cancer Therapy 2008;6:817–826.,