Effectiveness of Vitamin D Analogues in Treating Large Tumors and During Prolonged Use in Murine Retinoblastoma Models

Home , Vitamin D analogues

LABORATORY SCIENCES Effectiveness of Vitamin D Analogues in Treating Large Tumors and During Prolonged Use in Murine Retinoblastoma Models

Daniel M. Albert, MD, MS; Amit Kumar, MD; Stephen A. Strugnell, PhD; Soesiawati R. Darjatmoko, MS; Janice M. Lokken; Mary J. Lindstrom, PhD; Sarit Patel, MD

Objective: To investigate the effectiveness of the vita- smaller tumor size compared with its control (PϽ.001). min D analogues 1,25-(OH)2-16-ene-23-yne vitamin D3 No significant difference was seen between the 16,23-D3 ␣ ␣ (16,23-D3) and 1 -hydroxyvitamin D2 (1 -OH-D2)inin- group and its control. Some toxic effects related to hy- hibiting retinoblastoma growth in large tumors in a xe- percalcemia were seen in both studies. nograft model and with prolonged use in a transgenic model. Conclusions: In athymic mice in the large-tumor study, ␣ both 1 -OH-D2 and 16,23-D3 were effective in inhibit- Methods: For the large-tumor study, the xenograft athy- ing tumor growth compared with controls. In the long- ␣ mic mouse/human retinoblastoma cell (Y-79) model was term study, 1 -OH-D2 inhibited tumor growth but used. Subcutaneous tumors were allowed to grow to an 16,23-D3 did not. Effective doses of both compounds average volume of 1600 mm3. Systemic treatment with caused hypercalcemia and a significant increase in mor- 1 of the vitamin D analogues or with vehicle (control tality. groups) was carried out for 5 weeks. For the long-term ␤ ␣ study, transgenic –luteinizing hormone–large T anti- Clinical Relevance: Use of 1 -OH-D2 inhibited tu- gen (LH␤-Tag) mice were systemically treated with 1 of mor growth in large tumors and with long-term treat- the 2 compounds or vehicle (control groups) for up to ment compared with controls. Because of hypercalcemia- 15 weeks. Tumor size and signs of toxicity were as- related toxic effects seen in the present experiments, in sessed. clinical trials, serum calcium levels should be carefully monitored. This analogue may require use with drugs that Results: In the large-tumor study, tumor volume ratios lower serum calcium levels or use of relatively lower doses ␣ for the 1 -OH-D2 and 16,23-D3 groups were signifi- or skipped doses. The ideal alternative solution would cantly lower than those for controls (PϽ.002). No sig- be to identify vitamin D analogues that retain the anti- nificant differences in tumor volume were seen between neoplastic action without the calcemic activity. ␣ the 1 -OH-D2 and 16,23-D3 groups (P=.15). In the long- ␣ term study, the 1 -OH-D2 group showed significantly Arch Ophthalmol. 2004;122:1357-1362

ITAMIN D COMPOUNDS (1,25-dihydroxyvitamin-D3) and ergocal- have been recognized for ciferol (vitamin D2), inhibited tumor many years as having growth but with marked hypercalcemia- potential usefulness in the related toxic effects in mouse xenograft treatment of a variety of and transgenic models. Also demon- cancers,V including retinoblastoma.1 A strated was a similar inhibitory effect with major obstacle to their use in human stud- the newer synthetic vitamin D analogues ␣ ␣ 6-9 ies has been drug-induced hypercalce- 1 -hydroxyvitamin D2 (1 -OH-D2) and 1 From the Departments of mia. Vitamin D analogues have previ- 1,25-(OH)2-16-ene-23-yne vitamin D3 10,11 Ophthalmology and Visual ously been shown in preclinical studies (16,23-D3), with reduced calcemic ef- Sciences (Drs Albert, Kumar, fects. All of these studies were carried out and Patel and Mss Darjatmoko See also page 1365 using small tumors treated for a relatively and Lokken) and Biostatistics short time (5 weeks). The present studies and Medical Informatics of retinoblastoma using xenograft and were carried out to determine whether vi- (Dr Lindstrom), University of 2 Wisconsin Medical School, transgenic models to cause apoptosis and tamin D analogues are effective in the 3 Madison; and Bone Care to inhibit angiogenesis. Previous preclini- treatment of retinoblastoma in larger tu- International, Middleton, Wis cal studies4,5 demonstrated that 2 com- mors and remain effective with more pro- (Dr Strugnell). monly used forms of vitamin D, calcitriol longed administration.

(REPRINTED) ARCH OPHTHALMOL / VOL 122, SEP 2004 WWW.ARCHOPHTHALMOL.COM 1357

All research using mouse models of retinoblastoma con- The LH␤-Tag mice are a well-characterized transgenic model of formed to the guidelines set by the Research Animal retinoblastoma.16 A total of 170 LH␤-Tag transgene-positive mice Resources Center of the University of Wisconsin and the were randomized by sex and litter into 2 treatment groups and Association for Research in Vision and Ophthalmology State- 2 control groups corresponding to the 4 treatment groups de- ment for the Use of Animals in Ophthalmic and Vision scribed for the large-tumor athymic study in the previous sub- Research. These guidelines permit large-tumor studies to be ␤ section. The presence of the transgene was confirmed by poly- carried out in athymic mice but not in –luteinizing 16 ␤ merase chain reaction. These mice were maintained on the same hormone–large T antigen (LH -Tag) transgenic mice because vitamin D– and calcium-restricted diet as noted in the previous of the relatively small size of the tumors occurring in the eyes subsection. Each mouse in the 1␣-OH-D treatment group re- and the morbidity associated with orbital extension. Likewise, 2 ceived 0.2 µg/d. Each mouse in the 16,23-D3 treatment group long-term studies were suitable for the transgenic but not received 0.5 µg/d. Treatment was given 5 times per week. One athymic mice. third of the mice in each group were treated for 5 weeks, one third for 10 weeks, and one third for 15 weeks. Baseline body COMPOUND PREPARATION weights were recorded before the first treatment, and animals were reweighed twice per week during treatment and before eu- ␣ Pure crystalline 1 -OH-D2 was provided by Bone Care Inter- thanization on the last treatment day. Details of the oral GV and national and was prepared for administration with drug con- IP methods of treatment with vitamin D analogues have been centrations confirmed in the manner previously described previously described elsewhere.12,14 Doses were skipped for up 12 ␣ elsewhere. A solution of 1 -OH-D2 was diluted in coconut to 2 consecutive days in mice that either developed weight loss oil to a concentration of 0.2 µg/0.1 mL for use in this study. (Ͼ20% of baseline weight) or severe lethargy, and treatment was ␣ Each treated mouse received 0.2 µg of 1 -OH-D2 (approxi- resumed when the affected animals regained the lost weight or mately 10 µg/kg) per treatment. This dose has previously been resumed normal activity. Investigators were masked to histo- demonstrated in toxicity and dose-response studies1,12,13 to be pathologic examination findings and tumor measurement but the effective dose with the least toxicity. Pure crystalline not to drug delivery method. 16,23-D3 (provided by ILEX Oncology Inc, San Antonio, Tex) was prepared for injection as previously described else- TUMOR MEASUREMENT AND 1,14,15 where. This drug was diluted in mineral oil to a concen- HISTOPATHOLOGIC STUDY tration of 0.5 µg/0.1 mL. Each mouse in the treatment group received 0.5 µg of 16,23-D3 (approximately 25 µg/kg) per In athymic mice, tumor size was measured daily 5 times a week treatment. This dose was found in previous toxicity and dose- in 3 dimensions (length, width, and height) by means of calipers 1,14,15 response studies to be the effective dose with the least measuring to the nearest millimeter, and the volume was ap- toxicity. proximated by multiplying the 3 measurements. After euthanization, each tumor was excised, measured using calipers in 3 di- LARGE-TUMOR XENOGRAFT mensions, and weighed. The tumors were fixed in 10% neutral MODEL TREATMENT buffered formalin for standard histopathologic processing. Five- micrometer sections were cut and stained with hematoxylin- A total of 119 athymic “nude” mice (4-6 weeks old) were given eosin. In addition, von Kossa–stained preparations were made. dorsal subcutaneous injections of 2ϫ107 Y-79 human retino- These were examined microscopically, and the histologic fea- blastoma cells suspended in 0.5 mL of Iscove culture medium tures were recorded as previously described elsewhere.1,13,14 supplemented with 20% fetal bovine serum and basement mem- In transgenic mice, after euthanization the eyes were enucle- brane matrix suspensions. Details of culture methods have been ated and placed in 10% neutral buffered formalin for standard previously described elsewhere.1,14 After tumor cells were in- histopathologic sectioning. Four serially sectioned 5-µm- jected, all mice were maintained on a vitamin D– and calcium- thick sections were cut from each of the superior, middle, and restricted diet (PD; Purina Mills Inc, St Louis, Mo). This diet inferior areas of the globe and were stained with hematoxylin- was used to minimize hypercalcemia and to control for any effect eosin. The 4 sections from each globe area were examined un- of vitamin D in the diet. Mice were randomized to 1 of 4 treat- der a microscope, and the section with the largest tumor from ␣ ment groups: (1) 1 -OH-D2 (0.2 µg per mouse; 40 animals), each of the 3 areas was used for measurement. The outline of (2) gavage (GV) control (0.1 mL of coconut oil; 20 animals), the tumor in each section was traced from a microscopically (3) 16,23-D3 (0.5 µg per mouse; 40 animals), or (4) intraperi- digitized image, and the area was measured using image analy- toneal (IP) control (0.1 mL of mineral oil; 19 animals). The sis software (Optimas version 6.5; Media Cybernetics, Silver method of administration conformed with instructions pro- Spring, Md). The 3 tumor areas from each representative por- vided by the respective pharmaceutical companies supplying tion of the globe were averaged together to obtain the mean the drugs to obtain maximum blood levels. Tumors were al- tumor measurement of each eye (expressed in square microm- lowed to grow for 19 days to an average volume of 1600 mm3 eters). The measurements from both eyes were then averaged ␣ before initiation of treatment with either oral GV (1 -OH-D2 to provide the mean tumor area per mouse. Other histopatho- 1,12,15 and GV control groups) or IP (16,23-D3 and IP control groups). logic features were also evaluated. Treatment was given 5 times per week for 5 weeks. Baseline tumor volume and animal body weight measurements were de- TOXIC EFFECT ASSESSMENT termined before the first treatment. Each animal was weighed, and its tumor was measured using calipers, twice per week dur- Serum samples from representative mice in each group were ing treatment and just before euthanization on the last treat- obtained just before euthanization from the axillary vessels and ment day as previously described elsewhere.1,13,14 Investiga- were analyzed for calcium levels by Marshfield Laboratories, tors were masked to final tumor measurements and Marshfield, Wis. histopathologic examination findings but not to drug delivery Kidneys, lungs, and liver were harvested from represen- method. tative mice in each group in both arms of the study and were

(REPRINTED) ARCH OPHTHALMOL / VOL 122, SEP 2004 WWW.ARCHOPHTHALMOL.COM 1358

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 Table 1. Dose-Response Study of 1␣-OH-D2 and 16,23-D3 in Inhibiting Growth of Large Tumors

Tumor Size Change in Serum Survival, at End of Study, Tumor Volume, Calcium, Weight Dose Group Mice, No. No. (%) Mean ± SD, mm3* Mean ± SD, mm3 mg/dL Change, %

1␣-OH-D2 (0.2 µg), 10 µg/kg 40 26 (65) 3407 ± 421 2.56 ± 0.39 10.2 +2.90 GV control 20 18 (90) 6450 ± 780 5.24 ± 0.98 9.3 +3.30

16,23-D3 (0.5 µg), 25 µg/kg 40 39 (98) 2506 ± 278 1.96 ± 0.24 10.2 +14.9 IP control 19 17 (89) 6308 ± 767 5.22 ± 0.98 8.7 +14.8

Abbreviations: 1␣-OH-D2,1␣-hydroxyvitamin D2; 16,23-D3, 1,25-(OH)2-16-ene-23-yne vitamin D3; GV, gavage; IP, intraperitoneal. SI conversion factor: To convert serum calcium to millimoles per liter, multiply by 0.25. *Values are rounded to the nearest whole number.

Table 2. Survival of Animals in the Large-Tumor and Long-term Treatment Studies

Tumor Size ؋ 103,m3 Survival at 15 Serum Calcium, Dose Group Mice, No. wk, No. (%) 5wk 10wk 15wk mg/dL, at 15 wk

1␣-OH-D2 (0.2 µg), 10 µg/kg 40 33 (83) 19.25 79.42 463.62 9.4 GV control 20 20 (100) 30.24 936.15 1526.22 8.6

16,23-D3 (0.5 µg), 25 µg/kg 40 36 (90) 28.80 161.99 397.85 9.5 IP control 21 16 (76) 20.59 310.36 224.18 8.9

Abbreviations: See Table 1. SI conversion factor: To convert serum calcium to millimoles per liter, multiply by 0.25.

processed histologically. The number of organs harvested var- RESULTS ied among the groups. Kidneys were stained with von Kossa stain in addition to hematoxylin-eosin stain to ascertain the severity of renal calcification. Two sections of each kidney were TUMOR SIZE IN THE LARGE-TUMOR STUDY examined by masked reviewers (D.M.A. and A.K.), and the number of calcium deposits was determined and averaged for each In the athymic mice with large tumors, mean change in kidney. Each kidney was graded according to the following scale: tumor volume was analyzed as a proportional change dur- grade 0, no calcifications; grade I, 1 to 7 foci of calcification; ing treatment (volume at treatment end divided by vol- grade II, 8 to 15 foci of calcification; and grade III, greater than ume at treatment start) (Table 1). The tumor volume ␣ 15 foci of calcification. ratio for the 1 -OH-D2 group was significantly lower than Toxic effects were assessed by using the following vari- that for the GV control group, and the ratio for the ables: survival, changes in body weight, serum calcium levels, 16,23-D3 group was significantly lower than that for the and degree of kidney calcification. Animals that died before IP control group (PϽ.002 for each). No statistically sig- completion of the treatment protocol were not examined with nificant differences in tumor volume were seen between regard to the latter 3 categories. Further details regarding meth- ␣ ods involved in toxic effect assessment are described in detail the 1 -OH-D2 and 16,23-D3 groups (P=.15). elsewhere.12-15 TUMOR SIZE IN THE LONG-TERM STATISTICAL ANALYSIS TREATMENT STUDY

Tumor weight, tumor volume or area, animal weight, serum In the long-term study with LH␤-Tag mice, the tumor calcium level, and kidney calcification were analyzed using 1-way sizes after 5, 10, and 15 weeks of treatment are given in analysis of variance to assess statistical differences among the ␣ groups. Pairwise comparisons were then performed to detect Table 2. The 1 -OH-D2 group showed significantly statistical differences between particular dose groups. The data smaller tumor areas compared with the GV control group from various measurements were transformed to the log scale at each time point (PϽ.001). However, no significant dif- before analysis to stabilize the variance. The change in animal ference was seen between the 16,23-D3 group and the IP weight (from first to last measurement) was restricted to ani- control group at any of the 3 times. mals that survived until the last measurement. The effect of dose of vitamin D analogues on mortality was assessed using a gen- TOXIC EFFECTS IN THE LARGE-TUMOR eralized linear model assuming binomial variability. AND LONG-TERM TREATMENT STUDIES The effect of “layer” (ie, animal batch) on response was accounted for in all analyses by including a blocking term in The survival data for the study are presented in Tables 1 the model. All significant global tests for effect of dose were and 2. There were significantly fewer survivors among followed by pairwise analyses to assess differences between spe- ␣ cific dose groups. athymic mice in the 1 -OH-D2 group in the large- In the long-term arm of study, comparisons of each drug’s tumor study than in the GV control group (65% vs 90%; efficacy over each time point (5, 10, or 15 weeks) were also P=.03) (Table 1). There were also fewer survivors among ␤ ␣ performed. the LH -Tag mice in the 1 -OH-D2 group in the long-

(REPRINTED) ARCH OPHTHALMOL / VOL 122, SEP 2004 WWW.ARCHOPHTHALMOL.COM 1359

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 ␣ term study than in the GV control group (83% vs 100%; with controls, with the 1 -OH-D2 groups exhibiting P=.048) (Table 2). greater severity of calcification compared with the Serum calcium levels for both study arms are given 16,23-D3 groups. Gross and histopathologic examina- in Tables 1 and 2 and Table 3. All treatment groups had tion of lung and liver tissues from randomly selected mice significantly higher serum calcium levels compared with in each treatment group showed no evidence of meta- their respective control groups (PՅ.02), except both of static lesions or other abnormalities. the 15-week treatment groups, which did not show any Ͼ difference (P .25). No difference in serum calcium lev- COMMENT els was found between the 2 treatment modalities. The LH␤-Tag mice developed well-differentiated retinoblas- Retinoblastoma is the most common intraocular malig- toma, with rosettes, some calcification, and areas of ne- nancy of childhood, occurring once in every 20000 live crosis. Treated animals had smaller tumors, but the de- births worldwide.17 Current methods of treatment in- gree of differentiation and the proportional amounts of clude enucleation, external beam radiotherapy, scleral calcification and necrosis were similar in treated and con- plaque brachytherapy, cryotherapy, photocoagulation, and trol animals. The athymic/Y-79 mice had poorly differ- chemotherapy.17 Although current treatment methods entiated tumors, with an extensive and diffuse compo- have achieved survival of 90%, there remains a need for nent of dead cells. The diffuse nature of the necrosis is improved treatment alternatives to provide better visual similar to that seen in large intraocular retinoblastoma results and to decrease the risk of secondary nonocular in enucleated human eyes. Because of the diffuse distri- cancers in hereditary retinoblastoma.18 bution of dead cells, the precise percentage of the tumor In recent years, there has been strong interest in sys- they represent cannot be accurately determined using the temic chemotherapy and, particularly, “chemoreduc- present techniques. It is our impression, however, that tion” in the treatment of retinoblastoma.19-21 There are 4 dead cells generally represented one third to two thirds general circumstances in the treatment of retinoblas- of the cells present and were more numerous in treated toma in which chemotherapy is used20,21: (1) to shrink eyes. The gradations of the kidney calcifications are given tumors in eyes with visual potential that are too large to in Table 4. The treatment groups in both arms of the treat with focal methods to a size at which photocoagu- study showed a greater degree of calcification compared lation, cryotherapy, thermotherapy, or radioactive plaques can be administered; (2) in patients younger than 1 year Table 3. Serum Calcium Levels in the Large-Tumor who have advanced bilateral tumors that require exter- and Long-term Treatment Studies nal beam radiotherapy for cure; (3) as a single modality (which rarely obtains a permanent response)20; and (4) 21 Serum Calcium, mg/dL for the treatment of extraocular spread. Numerous se- rious short-term sequelae and less severe complications Long-term Treatment Study 20 Large-Tumor can occur. These drugs are mutagenic, and the devel- Dose Group Study 5wk 10wk 15wk opment of secondary nonocular cancers is a well- documented risk.20,22-28 New retinoblastomas have been 1␣-OH-D2 (0.2 µg), 10.2 11.7 10.96 9.39 10 µg/kg reported to develop in the eyes of patients undergoing GV control 9.3 9.1 8.1 8.63 treatment with systemic chemotherapy.20,29 30 16,23-D3 (0.2 µg), 10.2 10.6 9.95 9.5 Verhoeff, in 1966, hypothesized that calcification 25 µg/kg induced spontaneous regression in retinoblastoma and IP control 8.7 8.43 8.53 8.92 suggested that treatment with vitamin D might prove effective. Vitamin D has since been shown to inhibit the Abbreviations: See Table 1. 31 SI conversion factor: To convert serum calcium to millimoles per liter, growth of retinoblastoma in vitro and in vivo in the athy- multiply by 0.25. mic mouse model4 and in the transgenic mouse model.5

Table 4. Severity of Kidney Calcifications in Animals at the End of the Study*

Long-term Treatment Study

Large-Tumor Study 5wk 10 wk 15 wk

Mice, Mice, Mice, Mice, Dose Group No.† 0 I II III No.† 0 I II III No.† 0 I II III No.† 0 I II III

1␣-OH-D2 (0.2 µg), 180 639121704 115303 275958 10 µg/kg GV control 10 8 2 0 0 7 5200 5 5000 2725200

16,23-D3 (0.5 µg), 2041402 127401 128400 1212000 25 µg/kg IP control 8 6 2 0 0 5 3200 5 5000 5 5000

Abbreviations: See Table 1. *Kidney calcifications: 0, grade 0 (none); I, grade I (mild), 1 to 7; II, grade II (moderate), 8 to 15; III, grade III (severe), greater than 15. †Number of animals whose kidneys, liver, and lungs were examined.

(REPRINTED) ARCH OPHTHALMOL / VOL 122, SEP 2004 WWW.ARCHOPHTHALMOL.COM 1360

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 The effect, however, is unrelated to either high serum cal- effects of hypercalcemia were still observed. These ear- cium levels or calcium deposition in the tumor, and, in lier experiments with small tumors in short-term therapy fact, the clinical usefulness of vitamin D is limited by the simulated the conditions in which chemoreduction is most toxic effects associated with hypercalcemia.1,32 The an- commonly used. tineoplastic mechanism of action of vitamin D com- The experiments described in the present article were pounds used in models of human retinoblastoma has been intended to determine the effectiveness and the toxic ef- ␣ demonstrated to be apoptosis due to increased expres- fects of treatment with 16,23-D3 and 1 -OH-D2 in large sion of p53 and p212 and also inhibition of angiogen- tumors and in tumors receiving long-term therapy. The esis.3 Vitamin D receptor messenger RNAs, which are nec- size of tumors that could be treated and the duration of essary for this antineoplastic effect, have been treatment were limited by the guidelines of the Univer- demonstrated in Y-79 retinoblastoma cells, WERI-1 cells, sity of Wisconsin Medical School Animal Care and Use LH␤-Tag tumors, and 23 consecutive freshly removed hu- Committee regarding suffering and stress caused to ani- man retinoblastoma specimens using reverse transcrip- mals. Specifically, we wanted to determine whether drug- tase polymerase chain reactions.1 resistant cell lines develop with prolonged use of these Our initial studies were with calcitriol (the active drugs or with large tumors, as well as the drugs’ com- 4,5,33 form of vitamin D3) and ergocalciferol (vitamin D2). parative effectiveness in these situations. These experi- These compounds consistently inhibited tumor growth ments simulate the circumstances of treatment for ex- by more than 50% compared with controls. In the case traocular spread. In the large-tumor studies, 16,23-D3 of transgenic mice, animals were observed in the treat- therapy achieved 62% growth inhibition compared with ␣ ment groups with total regression of tumors. This was controls, and 1 -OH-D2 therapy achieved 51% inhibi- proven by serially sectioning the globes. All the nega- tion of tumor growth compared with controls. These de- tive animals were rechecked to make sure that they car- terminations are based on tumor size alone. However, as ried the transgene. In this study,3 6 of 10 animals in the determined by histopathologic analysis, there was a dif- high-dose calcitriol group and 4 of 12 animals in the low- fuse distribution of dead cell in the tumors that consti- dose calcitriol group showed no tumors, whereas 16 con- tuted one third to two thirds of the cells present, and we trol animals had tumors. The difference was statistically observed that the dead cells were more numerous in significant as assessed by a ␹2 test for independence. How- treated animals. We are adapting techniques to accu- ever, only 25% to 50% of animals survived 5 weeks of rately quantify numbers of viable, common, necrotic, and treatment in the calcitriol and ergocalciferol treatment apoptotic cells. groups given doses adequate to cause tumor inhibition In the long-term treatment studies using trans- ␣ or regression. genic mice only, 1 -OH-D2 therapy seemed to be effec- In the athymic/Y-79 xenograft treated with cal- tive. In these experiments, there was 92% regression at citriol, 81% inhibition of tumor growth was seen com- 10 weeks and 70% regression at 15 weeks compared with pared with controls, and vitamin D2 inhibition ranged from controls. Some treated eyes seemed to have totally re- 58% to 66%, but no regression was seen in these tu- gressed tumors on the basis of the slides examined, but mors. The real-time polymerase chain reaction expres- because serial sections were not examined, this cannot sion of vitamin D receptors is barely perceptible in the be stated with certainty. All of the groups that were ef- Y-79 retinoblastoma line. It is far lower than the 23 reti- fectively treated showed statistically significant hyper- noblastoma samples taken from enucleated eyes from dif- calcemia and associated increased mortality compared ferent patients.1 In contrast, the LH␤-Tag tumors have a with controls (Tables 1, 2, and 3). Although these toxic greater degree of vitamin D receptor expression. This dif- effects and mortality rates are considerably less than those ference seems to be related to the degree of drug re- seen with calcitriol and vitamin D2 treatment, it remains sponse. We concluded that calcitriol and ergocalciferol a potential deterrent to proceeding to clinical trials. showed impressive antineoplastic activity but caused hy- An Investigational New Drug application for 1␣- percalcemic activity that was too excessive for them to OH-D2 as a cancer treatment was submitted to the Food be used in human clinical trials. and Drug Administration in 1996, and an application for We then turned our attention to 2 synthetic vita- 16,23-D3 was submitted in 1999. In phase 1 trials for treat- ␣ ␣ min D analogues, 16,23-D3 and 1 -OH-D2, which have ment of prostate cancer, 1 -OH-D2 exhibited tumor antineoplastic effects similar to calcitriol and ergocalcif- growth suppression for stabilization with low but revers- erol but with reduced hypercalcemic activity. In short- ible toxic effects.35 This drug is currently being used in term experiments (5 weeks) in small tumors, 16,23-D3 phase 2 human trials of prostate cancer (George Wild- therapy caused 55% inhibition in LH␤-Tag transgenic ing, MD, oral communication, October 7, 2003). mice34 and 58% inhibition in athymic mice.14 The ad- Because of their mechanism of action, which dif- ministration of effective doses of 16,23-D3 resulted in 11% fers from that of existing retinoblastoma chemothera- mortality in transgenic mice34 and 25% mortality in athy- peutic agents, and the fact that they are nonmutagenic 14 ␣ mic mice compared with controls. Use of 1 -OH-D2 re- agents, we believe that these compounds have potential sulted in 81% tumor growth inhibition in transgenic mice12 use in eye-preserving and, in cases of extraocular spread, and 62% tumor growth inhibition in athymic mice13 com- life-preserving treatment as a component of multidrug ␣ pared with controls. Use of 1 -OH-D2 resulted in 12% therapy or possibly as a potential single-modality treat- 12 13 ␣ mortality in transgenic mice and 38% mortality in athy- ment. In clinical trials of 1 -OH-D2 and 16,23-D3, hy- mic mice. Thus, although the mortality rate was lower percalcemia-related toxic effects would need to be care- than with use of calcitriol and ergocalciferol, the toxic fully monitored, controlling the dose given and possibly

(REPRINTED) ARCH OPHTHALMOL / VOL 122, SEP 2004 WWW.ARCHOPHTHALMOL.COM 1361

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/27/2021 skipping doses in response to elevated serum calcium lev- D3 prolongs survival time of leukemic mice. Proc Natl Acad Sci U S A. 1990;87: els. Drugs that lower serum calcium levels may need to 3929-3932. 12. Dawson DG, Gleiser J, Zimbric ML, et al. Toxicity and dose-response studies of be used. The ideal alternative solution would be to iden- 1-␣ hydroxyvitamin D2 in LH␤-Tag transgenic mice. Ophthalmology. 2003;110: tify vitamin D analogues that retain the antineoplastic ac- 835-839. tion but have no calcemic activity. 13. Grostern RJ, Bryar PJ, Zimbric ML, et al. Toxicity and dose-response studies of 1␣-hydroxyvitamin D2 in a retinoblastoma xenograft model. Arch Ophthalmol. Submitted for publication November 13, 2003; final revi- 2002;120:607-612. 14. Sabet SJ, Darjatmoko SR, Lindstrom MJ, Albert DM. Antineoplastic effect and sion received January 23, 2004; accepted January 26, 2004. toxicity of 1,25-dihydroxy-16-ene-23-yne-vitamin D3 in athymic mice with Y-79 This study was supported by grant EY01917 from the human retinoblastoma tumors. Arch Ophthalmol. 1999;117:365-370. National Eye Institute, Bethesda, Md, and an unrestricted 15. Shternfeld IS, Lasudry JGH, Chappell RJ, Darjatmoko SR, Albert DM. Antineo- grant from Research to Prevent Blindness Inc, New York, plastic effect of 1,25-dihydroxy-16-ene-23-yne-vitamin D3 analogue in trans- NY. genic mice with retinoblastoma. Arch Ophthalmol. 1996;114:1396-1401. 16. Windle JJ, Albert DM, Marcus, et al. Retinoblastoma in transgenic mice. Nature. We thank Daniel Dawson, MD, and Joel Gleiser, MD, 1990;343:665-669. for treating animals and measuring tumor size and Kirsten 17. Ferris FL III, Chew EY. A new era for the treatment of retinoblastoma [editorial]. Hope for contributing to preparation of the article. Arch Ophthalmol. 1996;114:1412. Correspondence: Daniel M. Albert, MD, MS, Depart- 18. O’Brien JM. Alternative treatment in retinoblastoma [editorial]. Ophthalmology. ment of Ophthalmology and Visual Sciences, University of 1998;105:571-572. 19. Gallie B, Budnig A, De Boer G, et al. Chemotherapy with focal therapy can cure Wisconsin Medical School, F4/344 Clinical Science Cen- intraocular retinoblastoma without radiotherapy. Arch Ophthalmol. 1996;114: ter, 600 Highland Ave, Madison, WI 53792-3284 1321-1328. ([email protected]). 20. Abramson DH, Schefler AC. The treatment of retinoblastoma. In: Albert DM, Po- lans A, eds. Ocular Oncology. New York, NY: Marcel Dekker Inc; 2003:353-376. 21. Pucetti D. Treatment of extraocular retinoblastoma. In: Albert DM, Polans A, eds. REFERENCES Ocular Oncology. New York, NY: Marcel Dekker Inc; 2003:489-498. 22. Travis LB, Curtis RE, Stovall M, et al. Risk of leukemia following treatment for 1. Albert DM, Nickells RW, Gamm DM, et al. Vitamin D analogs, a new treatment non-Hodgkin’s lymphoma. J Natl Cancer Inst. 1994;86:1450-1457. for retinoblastoma: the first Ellsworth Lecture. Ophthalmic Genet. 2002;23:137- 23. Hawkins MM, Wilson LM, Stovall MA, et al. Epipodophyllotoxins, alkylating agents, 156. and radiation and risk of secondary leukaemia after childhood cancer. BMJ. 1992; 2. Audo I, Darjatmoko SR, Schlamp CL, et al. Vitamin D analogues increase p53, 304:951-958. p21, and apoptosis in a xenograft model of human retinoblastoma. Invest Oph- 24. Turner AR, Melynk A, Clark G. MDS and acute monocytic leukaemia after reti- thalmol Vis Sci. 2003;44:4192-4199. noblastoma [case report]. Br J Haematol. 1996;92:249. 3. Shokravi MH, Marcus DM, Alroy J, Egan K, Saornil MA, Albert DM. Vitamin D 25. Felice MS, Zubizarreta PA, Chantada GL, et al. Acute myeloid leukemia as a sec- inhibits angiogenesis in transgenic murine retinoblastoma. Invest Ophthalmol ond malignancy: report of 9 pediatric patients in a single institution in Argen- Vis Sci. 1995;36:83-87. tina. Med Pediatr Oncol. 1998;30:160-164. 4. Cohen S, Saulenas A, Sullivan C, Albert DM. Further studies of the effect of vi- 26. White L, Ortega JA, Ying KL. Acute non-lymphocytic leukemia following multi- tamin D on retinoblastoma inhibition with 1,25-dihydroxycholecalciferol. Arch modality therapy for retinoblastoma. Cancer. 1985;55:496-498. Ophthalmol. 1988;106:541-543. 27. Rivera GK, Pui CH, Santana VM, et al. Epipodophyllotoxins in the treatment of 5. Albert DM, Marcus DM, Gallo JP, O’Brien JM. The antineoplastic effect of vita- childhood cancer. Cancer Chemother Pharmacol. 1994;34:S89-S95. min D in transgenic mice with retinoblastoma. Invest Ophthalmol Vis Sci. 1992; 28. Draper GJ, Sanders BM, Kingston JE. Second primary neoplasms in patients with 33:2354-2364. retinoblastoma. Br J Cancer. 1986;53:661-671. 6. Jones G, Byford V, Makin HLJ, et al. Anti-proliferative activity and target cell ca- 29. Scott IU, Murray TG, Toledano S, et al. New retinoblastoma tumors in children tabolism of the vitamin D analog 1␣,24(S)-(OH)2D2 in normal and immortal- undergoing systemic chemotherapy. Arch Ophthalmol. 1998;116:1685-1686. ized human epidermal cells. Biochem Pharmacol. 1996;52:133-140. 30. Verhoeff FH. Retinoblastoma undergoing spontaneous regression: calcifying agent 7. Knutson JC, Hollis BW, LeVan LW, et al. Metabolism of 1␣-hydroxyvitamin D2 suggested in treatment of retinoblastoma. Am J Ophthalmol. 1966;62:573-574. to activated dihydroxyvitamin D2 metabolites decreases endogenous 1␣- 31. Saulenas A, Cohen S, Key L, Winter C, Albert DM. Vitamin D and retinoblas- dihydroxyvitamin D3 in rats and monkeys. Endocrinology. 1995;136:4749- toma: the presence of receptors and inhibition of growth in vitro. Arch Ophthal- 4753. mol. 1988;106:533-535. 8. Strugnell S, Byford V, Makin HLJ, et al. 1␣,24S-dihydroxyvitamin D2: a biologi- 32. Reichel H, Koeffler H, Norman A. The role of the vitamin D endocrine system in cally active product of 1␣-hydroxyvitamin D2 made in the human hepatoma, HEP3B. health and disease. N Engl J Med. 1989;320:980-991. Biochem J. 1995;310:233-241. 33. Albert DM, Saulenas AM, Cohen SM. Verhoeff’s query: is vitamin D effective against 9. Coburn JW, Tan AU Jr, Levine BS, et al. 1␣-hydroxy-vitamin D2: a new look at retinoblastoma? Arch Ophthalmol. 1988;106:536-540. an “old” compound. Nephrol Dial Transplant. 1996;11:153-157. 34. Wilkerson CL, Darjatmoko SR, Lindstrom MJ, Albert DM. Toxicity and dose- 10. Zhou JY, Norman AW, Lubbert M, et al. Novel vitamin D analogs that modulate response studies of 1,25-(OH)2-16-ene-23-yne vitamin D3 in transgenic mice. leukemic cell growth and differentiation with little effect on either intestinal cal- Clin Cancer Res. 1998;4:2253-2256. cium absorption or bone calcium mobilization. Blood. 1989;74:82-93. 35. Liu G, Oettel K, Ripple G, et al. Phase I trial of 1␣-hydroxyvitamin D2 in patients 11. Zhou JY, Norman AW, Chen DL, et al. 1,25-dihydroxy-16-ene-23-yne-vitamin with hormone refractory prostate cancer. Clin Cancer Res. 2002;8:2820-2827.

(REPRINTED) ARCH OPHTHALMOL / VOL 122, SEP 2004 WWW.ARCHOPHTHALMOL.COM 1362