Development of a murine orthotopic model of leukemia - Nature

© 2000 Nature America, Inc. 0929-1903/00/$15.00/⫹0 www.nature.com/cgt

Development of a murine orthotopic model of leukemia: Evaluation of TP53 gene therapy efficacy Gianluca Bossi,1 Raffaella Scardigli,1 Piero Musiani,2 Roberta Martinelli,1 Maria Pia Gentileschi,1 Silvia Soddu,1 and Ada Sacchi1 1

Molecular Oncogenesis Laboratory, Regina Elena Cancer Institute - Centro Richerche Sperimentali, Rome, Italy; and 2Department of Oncology and Neurosciences, “G. D’Annunzio” University, Chieti, Italy. The onco-suppressor gene TP53 has potential use in the gene therapy of many human cancers including leukemias. The latter indication derived from numerous experimental reports of p53-mediated suppressing effects on human and murine leukemia cells in vitro. However, few in vivo experiments have been performed, and those that have used a subcutaneous injection of p53-transduced leukemia cells. Thus, we developed an orthotopic leukemia model in adult, syngenic mice to evaluate the feasibility of TP53-mediated therapeutic approaches. We found that among other cells, v-src-transformed 32D myeloid progenitors induce leukemia when injected intravenously in syngenic mice. The resulting malignancy resembles the clinical manifestations of human acute myeloid leukemia because it is characterized by a massive invasion of bone marrow compartments, splenomegaly, generalized lymphadenopathy, and a macroscopic or microscopic infiltration of the kidneys, liver, and lungs. When these 32Dv-src cells were infected with a TP53-recombinant retrovirus before intravenous injection, we found a decreased mortality and, in those animals that develop leukemia, a drastic reduction of the generalized organ infiltration, suggesting that exogenous TP53 expression might be used for ex vivo bone marrow purging from leukemia cells. Cancer Gene Therapy (2000) 7, 135–143

Key words: TP53; leukemia, gene therapy; bone marrow.

C

onsiderable experimental and epidemiological evidence has accumulated in the past 10 years showing that the product of the TP53 gene is a tumor suppressor.1 The expression of an exogenous wild-type (wt)-p53 protein in a great variety of tumor cells has been shown to suppress their transformed phenotypes.2 Moreover, some clinical trials have been undertaken that attempt to transduce the TP53 gene in some solid tumors including hepatocellular carcinomas, metastatic liver tumors, nonsmall cell lung cancers, and head and neck squamous carcinomas.3 Alterations of the p53 pathway, including direct TP53 gene mutations4 or p53 protein inactivation by posttranslational alterations,5–7 are present in a considerable number of hematological malignancies and are associated with a poor prognosis and a poor response to treatment.8 Moreover, strong tumor-suppressing effects were found in in vitro studies with a number of leukemia cell lines expressing an exogenous wt-p53 protein.9 These effects include growth arrest, apoptosis, and/or differentiation.10 On the basis of these findings, leukemia cells are also thought to be a target for TP53Received September 8, 1998; accepted March 14, 1999. Address correspondence and reprint requests to Dr. Silvia Soddu or Dr. Ada Sacchi, Molecular Oncogenesis Laboratory, Regina Elena Cancer Institute - Centro Richerche Sperimentali, Via delle Messi d’Oro 156, 00158 Rome, Italy. E-mail address: soddu64dotto.ifo.it or sacchi64dotto.ifo.it

Cancer Gene Therapy, Vol 7, No 1, 2000: pp 135–143

mediated gene therapy.9 However, thus far, only a few in vivo experiments have been performed to evaluate the feasibility of this type of treatment, and these experiments only used subcutaneous (s.c.) injections of leukemia cells. In some leukemia cell lines, wt-p53 overexpression has been shown to induce apoptotic death that can be reverted in vitro by the presence of specific cytokines, indicating that the apoptosis in these tumor cells is caused by a TP53-mediated restoration of survival factor dependence.11–13 Because s.c. and muscle tissues, in which the in vivo tumorigenicity of tumor cell lines is usually measured, have only low levels of these cytokines, it is possible that the suppressing effects observed in vitro or in vivo on wt-p53 overexpression are, at least in part, due to the experimental conditions (e.g., low levels of cytokines). In contrast, the bone marrow (BM) environment, from which leukemia develops, is characterized by the presence of numerous cytokines that support the survival, proliferation, and differentiation of hemopoietic cells.14 Because of these qualities, the BM environment might weaken or even repress the suppressing effects that can be observed on exogenous wt-p53 overexpression in leukemia cells maintained in other environments. Thus, to address this issue and assess the feasibility of TP53-mediated gene therapy of leukemias, an orthotopic leukemia model in adult animals is required. To our knowledge, no such model has been used

135

136

BOSSI, SCARDIGLI, MUSIANI, ET AL: TP53 GENE THERAPY EFFICIENCY IN A MURINE MODEL

thus far to evaluate the effects of TP53-mediated gene therapy of hematological malignancies. Here, we report on the development of an orthotopic model of murine leukemia and on the data obtained by an ex vivo TP53-mediated gene therapy of these leukemic cells. Because of the need for TP53 gene transduction of recombinant retroviruses, to avoid interference between viruses we used hemopoietic cells transformed by activated oncogenes rather than by leukemogenic viruses such as Friend leukemia virus-transformed cells. We first evaluated the capacity of different oncogenetransformed myeloid progenitors to induce leukemic syndrome in syngenic animals. Although all of the cell types assessed were tumorigenic in immune-incompetent nude mice after s.c. injection, we found that only the v-src-transformed 32D myeloid progenitors were able to induce a clearly leukemic syndrome in immune-competent, syngenic adult mice after intravenous (i.v.) injection. We subsequently used this murine orthotopic leukemia model to evaluate the efficacy of an ex vivo TP53-recombinant retrovirus-mediated gene therapy. MATERIALS AND METHODS

Cell cultures 32Dv-src,15 32Dv-abl,16 and 32Dc-fms (S301, F969)17 cells were cultured in RPMI 1640 medium (Life Technologies, Grand Island, NY) supplemented with 10% heat-inactivated fetal bovine sera (FBS) (Life Technologies). The GP ⫹ E-LXSN and GP ⫹ E-Lp53SN ecotropic packaging cell lines18 were cultured in Dulbecco’s modified Eagle’s medium (Life Technologies) supplemented with 10% FBS. The GP ⫹ Ebased packaging cell lines are helper-independent lines with low virus recombination efficiency.19

Retroviral vectors and packaging cells The construction of the pLXSN retroviral vector has been described previously.20 Briefly, pLXSN is a defective Moloney leukemia virus-based vector that contains the bacterial neo gene downstream of the simian virus 40 early promoter. The pLp53SN vector was obtained by inserting human wt-p53 cDNA, derived from pCMVp53,21 into the unique BamHI site of the pLXSN vector. The generation of the virus-producing cells GP ⫹ E-LXSN and GP ⫹ E-Lp53SN has been described previously.18

Table 1. Evaluation of Tumorigenicity of 32D Cells Transformed by Different Oncogenes

Mice

Site of injection

nu/nu C3H/HeJ C3H/HeJ

s.c. s.c. i.v.

Cells 32Dv-src

32Dv-abl

24* † Yes‡ (5.0⫻105)§ Yes‡ (2.0⫻107)§ Yes‡ (⬍2.0⫻104)§ Yes‡ (2.0⫻107)§

32Dc-fms 20* No¶ NT㛳

* Reference number. † M. Valtieri and G. Rovera, unpublished observations. ‡ Tumorigenic. § TPD50, cell dose inducing tumors in 50% of the injected animals. ¶ Not tumorigenic. 㛳 NT, not tested.

Retroviral infections 32Dv-src cells were incubated for 1 hour with 8 ␮g/mL polybrene and subsequently infected by in vitro cocultivation with the same number of mitomycin-treated retrovirus-producing packaging cells (GP ⫹ E-Lp53SN or GP ⫹ E-LXSN). After 16 hours of cocultivation, 32Dv-src-Lp53SN and 32Dvsrc-LXSN cells were removed from the packaging monolayer and replated in fresh medium. After an additional 24 hours, cells were selected with G418 (1.5 mg/mL) for 5 days to eliminate the uninfected cells. The absence of replicating viruses from our lines was assessed by infection and G418 selection of NIH-3T3 cells with the supernatant of 32Dv-src infected cells.

Indirect immunofluorescence Approximately 4 ⫻ 104 cells were spun onto a glass slide with a cytocentrifuge. Cytospin preparations were air-dried, fixed with 2% formaldehyde for 10 minutes at room temperature, permeabilized with 0.25% Triton X-100 for 5 minutes, incubated for 4 hours at 37°C with sheep anti-p53 sera (Ab-7, Oncogene Science, Uniondale, NY). After washing, cytospins were incubated for 1 hour at 37°C with rhodamine-conjugated rabbit anti-sheep immunoglobulin G (Cappel, Durham, NC).

Tumorigenesis analysis Serial dilutions of logarithmically proliferating cells were injected s.c. between the scapulae or injected i.v. into the tail vein in syngenic C3H/HeJ mice22 (6- to 7-week-old males). Mice were examined twice a week for the appearance of tumor

Table 2. Pathology of Mice Injected I.v. with 32Dv-src Cells

Cells injected 2.0 ⫻ 10 2.0 ⫻ 105 2.0 ⫻ 106 4

No. of mice with leukemia and % of tumor take 11/14 (78.5%) 18/24 (75 %) 14/14 (100 %)

Tumor latency (days)

Spleen weight (g ⫾ SD)

80 ⫾ 13 74 ⫾ 31 63 ⫾ 6

0.71 ⫾ 0.20 0.64 ⫾ 0.33 0.63 ⫾ 0.30

Lymphadenopathy*

Leukemic liver and/ or kidneys†

BM invasion‡

Vertebral column paralysis

63% 71% 71%

63% 23% 29%

100% 100% 100%

18% 23% 29%

* Mice were considered positive for lymphadenopathy when at least three lymphoid compartments were found to be enlarged by macroscopic examination. † By macroscopic examination. ‡ By cell explant and G418 resistence in culture.

Cancer Gene Therapy, Vol 7, No 1, 2000


137

Figure 1. Histological examination of the lung (a), liver (b), and BM (c) from a mouse injected i.v. with 32Dv-src cells. A massive leukemic infiltrate widens the lung intra-alveolar septa (a) and is present in foci near the central vein of the hepatic lobule (b). The BM of the vertebral body (c) is completely replaced by neoplastic cells, which erode the cancellous and the cortical bone and infiltrate the surrounding tissue. Magnification is at ⫻630.

masses in the case of s.c. injection, or for general illness in the case of i.v. injection. Autopsy and macroscopic examination of organs were performed on dead mice and on ill mice sacrificed by cervical dislocation. After autopsy of sacrificed mice, cells were explanted from the lymph nodes, BM, spleen, and vertebral column and cultured in RPMI 1640 with 10% FBS supplemented with 2 ␮g/mL puromycin and subsequently with 1.5 mg/mL G418.

Morphological analysis Animals were sacrificed by cervical dislocation. For histological evaluation, tissues were fixed in 10% neutral-buffered formalin, embedded in paraffin, sectioned at 4 ␮m, and stained with hematoxylin-eosin and Giemsa. For in situ staining of apoptotic cells, a terminal deoxynucleotidyltransferase-catalyzed DNA nick-end labeling assay was performed using a kit from Boehringer Mannheim (Mannheim, Germany) according to the manufacturer’s recommendations. Sections were briefly counterstained with hematoxylin. Quantitative studies of mitoses and apoptotic cells were performed independently by three pathologists in a blind fashion on three or more samples from distinct mice by evaluating 10 randomly chosen fields in each sample. Mitoses and apoptotic cells were counted under a microscope at ⫻400 high-power (HP) fields (a ⫻40 objective and a ⫻10 ocular lens; 0.180 mm2 per field). A total of 10 randomly chosen fields were counted for each sample. For additional morphological evaluation of the differentia-

tion state, 32Dv-src cells infected with LXSN or Lp53SN retroviruses were explanted from tumors, selected with puromycin and G418, cytocentrifuged, and stained by May-Grunwald Giemsa.

Western blotting Protein extracts, sodium dodecyl sulfate-polyacrylamide gel electrophoresis, blotting, and immunoreactivity were performed as described previously.15 The anti-p53 monoclonal antibody PAb1801 (Oncogene Science), which preferentially recognizes p53 of human origin, was employed. Immunoreactivity was detected using an enhanced chemiluminescence kit (Amersham, Arlington Heights, Ill) according to the manufacturer’s instructions.

Statistical analysis Student’s t test and Fisher’s double-tailed test for the comparison of two proportions (with approximation of binomial distribution by means of normal distribution and Yates’ correction for continuity) were used for statistical analysis.

RESULTS

Establishment of a murine orthotopic leukemia model The interleukin-3 (IL-3)-dependent 32D myeloid progenitor cells,22 which have an endogenous wt TP53 gene

Figure 2. Indirect immunofluorescence of exogenous, human p53 protein in infected 32Dv-src cells. The cells were infected with Lp53SN (A,B) or LXSN (C,D) and selected for 5 days in the presence of G418; the surviving cells were purified by Ficoll gradient centrifugation and subsequently cytocentrifuged on slides. A,C: Nuclei counter-stained by Hoechst. B,D: The same fields as in (A,C), respectively, in which p53 protein was detected with specific Ab7 sera.


138


and expressed a very low level of p53 protein,15 can be transformed by some activated oncogenes, such as v-src, v-abl, and c-fms (S301, F969); after transformation, these cells can induce s.c. tumors in nude mice (M. Valtieri and G. Rovera, unpublished observations).17,23 Overexpression of an exogenous wt-p53 protein was shown to inhibit the transformed phenotype of each of these cell lines in vitro.15 We first evaluated whether these cells were tumorigenic in syngenic adult mice after s.c. injection. Only 32Dv-src and 32Dv-abl cells induced s.c. tumors in immune-competent syngenic mice (Table 1), and these two cell lines were then used for i.v. injections. As shown in Table 1, after i.v. injection, 32Dv-src cells induced a neoplastic disease with a greater efficiency than 32Dv-abl cells; the dose needed to produce tumors in 50% of the animals was ⬍2 ⫻ 104 and 2 ⫻ 107, respectively. Upon morphological examination, the mice injected with 32Dv-src cells presented with severe splenomegaly, generalized lymphadenopathy, leukemic lungs, kidneys, and liver, and massive infiltration of BM compartments (Table 2 and Fig 1). Approximately 25% of the injected animals died due to posterior paralysis resulting from infiltration of the vertebral bodies by the 32Dv-src cells, as assessed by histology (Fig 1), and dissection of the tumor mass and in vitro cell culture in the presence of puromycin (data not shown), to which our 32Dv-src cells are resistant.15 Because this neoplastic disease resembles the human acute myeloid leukemia, the 32Dv-src cells provide a suitable orthotopic model to evaluate the efficacy of new therapeutic approaches in adult syngenic mice.

Effect of TP53-recombinant retrovirus infection on tumorigenicity of 32Dv-src leukemia cells We found previously that despite the wt status of the endogenous TP53 gene, overexpression of exogenous wt-p53 protein in 32Dv-src cells maintained in vitro suppresses transformation by differentiation through the monocytic pathway at 3 weeks postinfection.15,18 This differentiation is not terminal because it is not associated with a proliferation arrest.15 To evaluate the effects of TP53 on in vivo tumorigenicity, 32Dv-src cells were infected with LXSN or Lp53SN retroviruses carrying the neomycin resistance gene alone or together with human TP53 cDNA. Because exogenous wt-p53 protein expression does not obviously modify the proliferation rate of 32Dv-src cells in vitro,15 the infected cells were maintained for 5 days in the presence of G418 selection before injection. This period of selection was sufficient to kill 100% of the mock-infected cells. Western blot analysis (data not shown) and indirect immunofluorescence (Fig 2) were performed to evaluate the expression of the exogenous p53 protein. As expected for the polyclonal population of transduced cells, heterogeneous levels of p53 expression were found in Lp53SNinfected 32Dv-src cells. Aliquots of 200 ␮L containing from 2 ⫻ 104 to 2 ⫻ 106 total number of G418-resistant infected cells were injected i.v.; next, mice were examined twice a week for 1

Figure 3. Tumorigenicity of 32Dv-src cells after retroviral infection. The indicated concentrations of 32Dv-src cells infected with Lp53SN (f) or LXSN (䡺) recombinant retroviruses were injected i.v. in C3H/Hej syngenic adult mice. The survival curves of the injected animals are reported. The numbers near the symbols indicate the ratio of tumor-bearing mice to the total number of injected mice.

year. A total of 104 animals were injected and analyzed. The mice often started to be cachectic ⬃1 week before death. The ill animals (72 of 104 (50%) of total number of injected mice) were subjected to autopsy and, in 21 of 72 (30%) of the cases, cells from spleens, lymph nodes, BM, and vertebral masses were explanted and cultured in the presence of G418 and puromycin to assess their origin. Double resistant cells were selected from all of the organ explants analyzed, indicating that 32Dv-src cells resistant to puromycin15 that had been infected with the recombinant retroviruses, which induce G418 resistance, were responsible for organ invasion. Autopsies were also performed on animals that survived and were apparently healthy for 1 year. In this group, all animals were macroscopically free of disease. As shown by the survival curves reported in Figure 3, a 50.8% reduction in mortality was present in Lp53SN-treated mice versus LXSN-treated mice at the lowest number of cells injected (2 ⫻ 104, P ⬍ .05 by Fisher’s test). At the


139


Table 3. Effects of p53 Gene Transduction on the Morbidity of the 32Dv-src Leukemia Model* Cell number and cell type injected

Lymphadenopathy†

Leukemic liver and/or kidneys‡

BM invasion§

Vertebral column paralysis

0.42 ⫾ 0.35¶ 0.64 ⫾ 0.32

38%㛳 50%

38%¶ 75%

100%¶ 100%

38%¶ 25%

69 ⫾ 12# 88 ⫾ 32

0.60 ⫾ 0.38** 0.38 ⫾ 0.36

45%㛳 18%

30%¶ 18%

100%¶ 100%

55%¶ 41%

57 ⫾ 7** 64 ⫾ 8

0.74 ⫾ 0.37# 0.36 ⫾ 0.24

75%㛳 33%

16%¶ —

100%¶ 100%

25%¶ 44%

Tumor latency (days)

Spleen weight (g ⫾ SD)

2.0 ⫻ 104/mouse 32Dsrc-LXSN 32Dsrc-Lp53SN

106 ⫾ 33¶ 97 ⫾ 25

2.0 ⫻ 105/mouse 32Dsrc-LXSN 32Dsrc-Lp53SN 2.0 ⫻ 106/mouse 32Dsrc-LXSN 32Dsrc-Lp53SN

* The data are from ill animals; surviving, healthy mice are not included in this analysis. † Mice were considered positive for lymphadenopathy when at least three lymphoid stations were found to be enlarged by macroscopic examination. ‡ By macroscopic examination. § By cell explant and G418 resistence in culture. ¶ Not statistically significant. 㛳 P ⬍ .05 by Fisher’s test; the data have been analyzed together to make the test applicable. # P ⬍ .01 by Student’s t test. ** P ⬍ .05 by Student’s t test.

higher number of cells injected (2 ⫻ 105 and 2 ⫻ 106), a weaker suppression of mortality was observed (19% and 31%, respectively), which was still statistically significant when the results of the two different injections were analyzed together (P ⬍ .05 by Fisher’s test). Moreover, a significant inhibition of morbidity with reduced splenomegaly, lymphadenopathy, and organ infiltration was found (Table 3 and Fig 4). Only the BM compartments were constantly found to be invaded by either type (LXSN or Lp53SN) of infected cells in the ill mice (Table 3). The reduction of mortality at 2 ⫻ 104 cells injected with no reduction of morbidity in the animals that developed leukemia, as well as the opposite results obtained at the higher number of cells injected, might be determined by different causes. We excluded the appearance of recombinant, replication-competent retroviruses, whose concentration would have been higher in the injections with the largest number of cells, by cultivating NIH-3T3 cells in the presence of the supernatant of 32Dv-src-infected cells, before and after in vivo growth, and by subsequent G418 selection. No G418resistant NIH-3T3 cells were found, indicating that no detectable replication-competent viruses developed in our packaging and 32Dv-src-infected cells. Another explanation might rely on the different tumor latencies and the constant (100%) invasion of BM compartments. Although some mice die of posterior paralysis, it is conceivable that the others die because of the consequences of BM infiltration. Because at 2 ⫻ 104 cells injected, the tumor latency is longer even for the control mice, it is possible that those Lp53SN-treated animals that are not cured by wt-p53 protein expression have enough time to develop organ infiltration. In contrast, at the higher doses of cells injected, although wt-p53 protein is less active in suppressing mortality, the shorter


tumor latency allows the evaluation of the TP53 efficacy on morbidity. Taken together, these results indicate that an ex vivo transduction of the TP53 gene can reduce the tumorigenicity (mortality plus morbidity) of leukemia cells in the hemopoietic compartment.

Figure 4. Comparison of spleens and lymph nodes from normal mice and mice injected with 32Dv-src, 32Dv-src-LXSN, and 32Dvsrc-Lp53SN cells. The injected mice were sacrificed when cachectic. Representative examples of spleens and lymph nodes are reported in comparison with organs from normal, noninjected mice of the same age.

140


Table 4. Analysis of Relationship Between Exogenous p53 Detection and Pathological Status Lymph nodes† Cell injected

Days*

Spleen g

Inguinal

Axillary

Mesentary

Peyer’s patch

Liver

Kidneys

Paralys. vertebr.

p53 detection‡

2.0 ⫻ 10 1

89

0.38

⫹

⫹

—

—

—

Leuk.¶

Yes

No

2.0 ⫻ 105 1 2 3 4 5 6 7

68 74 75 82 86 95 106

0.8 0.38 0.14 0.14 0.23 0.46 1.42

⫹⫹ ⫹ — — ⫹ ⫹ ⫹

⫹⫹ ⫹ — ⫹⫹⫹ — — ⫹

⫹⫹ ⫹ — — — — ⫹⫹

⫹ — ⫹ — — — ⫹

— — — — — — —

— — — — Leuk. Leuk. —

No Yes Yes No Yes No No

Yes Yes Yes Yes Yes No No

2.0 ⫻ 106 1

67

0.38

⫹⫹

⫹⫹

⫹

—

—

—

No

Yes

4

* Days after tumor cell injection in which the clinical features of leukemia start to appear. † By macroscopic examination: ⫹, lightly; ⫹⫹, moderately; ⫹⫹⫹, notably increased. ‡ p53 detection was performed with the anti-p53 monoclonal antibody PAb1801, which preferentially recognizes p53 of human origin. ¶ Leuk. ⫽ Leukemic kidneys.

Analysis of exogenous p53 protein in leukemia-derived cells Among the mice injected with 32Dv-src-Lp53SN cells, there was a heterogeneity in the suppression of the leukemic syndrome. Excluding the animals that did not develop the leukemia, the ill mice presented a larger variability in the degree of organ infiltration (Table 4) compared with the control animals (data not shown). To evaluate whether these differences could be due to loss of the exogenous TP53 gene, we assessed the expression of the exogenous p53 protein by Western blot analysis on cells explanted from the spleen, lymph nodes, and BM of nine different mice. No correlation was found between the maintenance of exogenous p53 expression and milder tumorigenic phenotypes (Table 4), but we cannot exclude the presence of other mechanisms of exogenous p53 protein inactivation.

Evaluation of the biological effects induced in vivo by exogenous TP53 expression Overexpression of wt-p53 protein in several leukemia cells in vitro was shown to suppress the transformed phenotype by a variety of mechanisms, including inhibition of cell proliferation, induction of differentiation, and/or restoration of cytokine-dependence.10 The latter effect can promote apoptosis when the required cytokines are lacking.11–13 We have shown previously that exogenous wt-p53 overexpression in in vitro cultured 32Dv-src cells induces their differentiation through the monocytic pathway with no modification of survival and proliferation capacities.15 Because the environment is known to play a critical role in the definition of the TP53-mediated outcomes,15,24,25 we evaluated the effect(s) induced by wt-p53 overexpression in 32Dv-src cells in the tumor masses. Histological analysis of different leukemia-invaded organs was performed on samples from mice injected with 32Dv-src-LXSN or 32Dv-src-

Lp53SN cells after hematoxylin-eosin and Giemsa staining (for details, see Materials and Methods). In agreement with the in vitro observation, a more differentiated phenotype of the infiltrating cells was found in the mice injected with Lp53SN-infected cells compared with the controls (Fig 5A). To further analyze the differentiation state, the cells explanted and selected with puromycin and G418, as described above, were cytocentrifuged and stained by May-Grunwald Giemsa for morphological examination. As shown in Figure 5B, differentiation through the monocytic pathway was present only in 32Dv-src cells infected with Lp53SN retrovirus and maintained in vivo. Interestingly, the in vivo analyses of Lp53SN-treated mice versus LXSN-treated mice also showed a reduced rate of mitotic figures (⬍3.2 ⫾ 1.3/HP field in Lp53SNtreated mice and ⬎10.6 ⫾ 2.4/HP field in LXSN-treated mice; P ⬍ .01 by Student’s t test) and an increased number of apoptotic cells (⬎3.9 ⫾ 1.3/HP field in 6 of 10 Lp53SN-treated mice and ⬍0.7 ⫾ 0.4/HP field in LXSNtreated mice; P ⬍ .01 by Student’s t test) (Fig 5A). The latter results were confirmed by the terminal deoxynucleotidyltransferase-catalyzed DNA nick-end labeling assay (data not shown). These results together with those obtained previously in vitro on the same 32v-src cells15 strongly support an environmental contribution in the determination of p53 final outcomes. Moreover, the data show that the TP53-transduced cells have a reduced rate of proliferation and can undergo differentiation and apoptosis in the presence of the physiological amount of cytokines present in the hemopoietic compartment. DISCUSSION We evaluated the efficacy of ex vivo TP53 gene therapy by recombinant retrovirus in an orthotopic leukemia



141

Figure 5. A: Representative samples of a histological examination of tumors from mice injected with 32Dv-src-LXSN or 32Dv-src-Lp53SN cells. 32Dv-src-LXSN-derived tumors (a) consist of large cells with vesicular nuclei, prominent nucleoli, basophilic cytoplasm, and a highproliferation fraction; one mitosis is shown (arrowhead). In contrast, 32Dv-src-Lp53SN-derived tumors (b) are composed of medium-sized cells with round nuclei, fine chromatin, and occasionally, an evident nucleolus. Mitoses are rare, whereas cells with apoptotic features are frequent. In the latter cells, chromatin is pyknotic and packed into smooth masses applied against the nuclear membrane. Nuclei with the characteristic half moon of condensed chromatin are shown (arrowheads). Magnification is at ⫻1000. B: Cytospin preparation of explants from the same tumors as in (A) (in vivo) compared with the relative cells before in vivo injection (in vitro). Cells were stained with May-Grunwald Giemsa for morphological analysis. The cells always have the morphology of undifferentiated myeloblasts, with the exception of in vivo 32Dv-src-Lp53SN, in which cells show characteristics of differentiated monocytes (eccentric and reniform nuclei as well as less basophilic cytoplasms). Magnification is at ⫻630.


142


model. We found that TP53-recombinant retrovirus infection of 32Dv-src leukemia cells before they are injected ir i.v. into syngenic adult mice can significantly reduce morbidity and mortality compared with controlinfected cells. To our knowledge, this is the first study in which the suppressing effects of the TP53 gene were assessed in vivo on leukemia cells in their own environment. It has been shown previously that overexpression of wt-p53 protein can suppress the transformed phenotypes of a large number of leukemia cells in vitro.11,12,26 –30 Besides the numerous in vitro experiments, only few in vivo studies were performed, and those exclusively used s.c. injections of transduced leukemic cells.27,28 We believe that an orthotopic model is required to evaluate the TP53-mediated effects on leukemias for two main reasons. The first concerns the environmental influence on TP53-induced biological outcomes. IL-3-dependent hemopoietic progenitors undergo apoptosis or proliferation arrest on the same stressing stimulus (e.g., ultraviolet irradiation), depending upon the presence or absence of IL-3 or on the activation of pathways mimicking the presence of cytokines.24 Different outcomes (differentiation or proliferation arrest) are induced by exogenous wt-p53 overexpression in cells of the same line transformed by different activated oncogenes.15 Because hemopoietic and s.c. environments are completely different, and because wt-p53 overexpression induces survival factor dependence, leukemia cells may undergo apoptosis in vitro or after in vivo injection into tissues deficient in hemopoietic cytokines, whereas the same leukemia cells might continue to survive and proliferate in the cytokine-rich BM compartment. We found that 32Dv-src cells transduced with the TP53-recombinant retrovirus were suppressed in their tumorigenicity when injected i.v. in syngenic mice, indicating the possibility of developing p53-mediated gene therapy of leukemias. Interestingly, the histological analysis showed that exogenous wt-p53 expression can induce more types of suppressing effects in vivo than in vitro, including differentiation, a reduction of mitotic figures, and an increased number of apoptotic cells; only differentiation was observed in vitro, without modification of proliferation or apoptosis. These results confirm the importance of the environment in the determination of the suppressing effects of TP53. At present, the selective transduction of leukemia cells in vivo cannot be achieved with the transfection or infection systems available. However, we have recently found that the introduction of exogenous TP53 in murine primary BM cells does not affect their proliferation and differentiation capacities.18 Thus, normal BM cells are apparently not affected by TP53 transduction, whereas several leukemia cells are suppressed in their transformed phenotypes; consequently, we have proposed that BM purging from p53-responsive leukemia cells might be achieved ex vivo by delivering TP53 to all (normal and leukemic) marrow cells.18,31 This procedure would allow a functional targeting only of the neoplastic cells and might be used before autologous BM trans-

plantation.32 Thus, the second main reason to develop an orthotopic leukemia model to evaluate TP53-mediated gene therapy is represented by the requirement of assessing the efficacy of this BM purging strategy. The results reported here indicate that TP53-recombinant retrovirus infection can reduce the morbidity and mortality of mice injected with 32Dv-src cells. Although these are promising in vivo results, the partiality of the suppressing effect indicates that, for an efficient BM purging, an ex vivo apoptotic outcome would be the most desired TP53-mediated effect. Because the mechanism(s) by which TP53 can induce survival factor dependence and eventually apoptosis is still unknown, as are those that cause proliferation arrest or differentiation, for therapeutic purposes, at present, it is only possible to empirically evaluate which leukemia responds to exogenous wt-p53 protein expression by apoptosis. We have recently found that in leukemia cells corresponding to different stages of myeloid differentiation, forced expression of the TP53 gene causes apoptosis only in the cells in the earlier stages of differentiation and independently of the status of the endogenous TP53 gene.33 The development of the orthotopic model of leukemia can also allow the evaluation of alternative or combined treatments. These might include (a) the de novo expression of the TP53 gene together with radiotherapy or chemotherapy;34 (b) the use of modified p53 protein, such as the chimeric tumor suppressor 1, which is more resistant to inactivation and possesses enhanced proliferation-suppressing and apoptotic-promoting activities compared with wt-p53;35 and (c) the contemporary transduction of two different onco-suppressor genes.36 In all of these cases, for an efficient BM purging, it will be necessary to evaluate whether the treatment is not detrimental for normal BM cells. We are currently evaluating some of these strategies in our leukemia model, which we believe to be a useful tool for studying the in vivo efficacy of new therapeutic approaches for leukemias.

ACKNOWLEDGMENTS We thank Frank L. Graham and Silvia Bacchetti for critical revision of the manuscript and for helpful advice and discussions. We also thank Alessandro Porrello for statistical analysis. G.B. is the recipient of a fellowship from Fondazione Italiana per la Ricerca sul Cancro. The support provided by Associazione Italiana per la Ricerca sul Cancro, Italy-United States Finalized Project and by the National Research Council Biotechnology Project is gratefully acknowledged.

REFERENCES 1. Gottlieb TM, Oren M. p53 in growth control and neoplasia. Biochim Biophys Acta. 1996;1287:77–102. 2. Ozturk M, Ponchel F, Pulsleux A. p53 as a potential target in cancer therapy. Bone Marrow Transplant. 1992;9(suppl 1):164 –170.



3. Roth JA, Cristiano RJ. Gene therapy for cancer: what have we done and where are we going? J Natl Cancer Inst. 1997;89:21–39. 4. Imamura J, Miyoshi I, Koeffler HP. p53 in hematologic malignancies. Blood. 1994;84:2412–2421. 5. Slingerland JM, Minden MD, Benchimol S. Mutation of the p53 gene in human acute myelogenous leukemia. Blood. 1991;77:1500 –1507. 6. Zhang W, Hu G, Estey E, Hester J, Deisseroth A. Altered conformation of the p53 protein in myeloid leukemia cells and mitogen-stimulated normal blood cells. Oncogene. 1992;7:1645–1647. 7. Seliger B, Papadileris S, Vogel D, et al. Analysis of the p53 and MDM-2 gene in acute myeloid leukemia. Eur J Haematol. 1996;57:230 –240. 8. Wattel E, Preudhomme C, Hecquet B, et al. p53 mutations are associated with resistance to chemotherapy and short survival in hematologic malignancies. Blood. 1994;84:3148–3157. 9. Prokocimer M, Rotter V. Structure and function of p53 in normal cells and their aberrations in cancer cells: projection on the hematologic cell lineages. Blood. 1994;84:2391–2411. 10. Hansen R, Oren M. p53: from inductive signal to cellular effect. Curr Opin Genet Dev. 1997;7:46 –51. 11. Yonish-Rouach E, Resnitzky K, Lotem J, Sachs L, Kimchi A, Oren M. Wild-type p53 induces apoptosis of myeloid leukemic cells that is inhibited by interleukin 6. Nature. 1991;352:345–347. 12. Johnson P, Chung S, Benchimol S. Growth suppression of Friend virus-transformed erythroleukemia cells by p53 protein is accompanied by hemoglobin production and is sensitive to erythropoietin. Mol Cell Biol. 1993;13:1456 – 1463. 13. Gottlieb E, Haffner R, Yonish-Rouach E, von Ruden T, Wagner EF, Oren M. Wild type p53 activity contributes to dependence on hematopoietic survival factors. In: Mihic E, Schimke RT, eds. Apoptosis. NewYork: Plenum Press; 1994:31– 45. 14. Sachs L. The control of hematopoiesis and leukemia: from basic biology to the clinic. Proc Natl Acad Sci USA. 1996;93:4742– 4749. 15. Soddu S, Blandino G, Scardigli R, et al. Wild-type p53 induces diverse effects in 32D cells expressing different oncogenes. Mol Cell Biol. 1996;16:487– 495. 16. Rovera G, Valtieri M, Mavilio F, Reddy EP. Effect of Abelson murine leukemia virus on granulocytic differentiation and interkeukin-3 dependence of a murine progenitor cell line. Oncogene. 1987;1:29 –35. 17. Pierce JH, Di Marco E, Cox GW, et al. Macrophagecolony-stimulating factor (CSF-1) induces proliferation, chemotaxis, and reversible monocytic differentiation in myeloid progenitor cells transfected with the human c-fms/ CSF-1 receptor cDNA. Proc Natl Acad Sci USA. 1990;87: 5613–5617. 18. Scardigli R, Bossi G, Blandino G, Crescenzi M, Soddu S, Sacchi A. Expression of exogenous wt-p53 does not affect normal hematopoiesis: implication for bone marrow purging. Gene Ther. 1997;4:1371–1378.


143

19. Markowitz D, Goff S, Bank A. A safe packaging line for gene transfer: separating viral genes on two different plasmids. J Virol. 1988;62:1120 –1124. 20. Miller AD, Rosman GJ. Improved retroviral vectors for gene transfer and expression. Biotechniques. 1989;7:980 – 982. 21. Baker SJ, Fearon ER, Nigro JM, et al. Chromosome 17 deletions and p53 gene mutations in colorectal carcinomas. Science. 1989;244:217–221. 22. Greenberger JS, Eckner RJ, Sakakeeny A, et al. Interleukin 3-dependent hematopoietic progenitor cell lines. Fed Proc. 1983;42:2762–2771. 23. Kruger A, Anderson SM. The v-src oncogene blocks the differentiation of a murine myeloid progenitor cell line and induces a tumorigenic phenotype. Oncogene. 1991;6:245– 256. 24. Canman CE, Gilmer TM, Coutts SB, Kastan MB. Growth factor modulation of p53-mediated growth arrest versus apoptosis. Genes Dev. 1995;9:600 – 611. 25. Polyak K, Waldman T, He TC, Kinzler KW, Vogelstein B. Genetic determinants of p53-induced apoptosis and growth arrest. Genes Dev. 1996;10:1945–1952. 26. Soddu S, Blandino G, Citro G, et al. Wild-type p53 gene expression induces granulocytic differentiation of HL-60 cells. Blood. 1994;83:2230 –2237. 27. Shaulsky G, Goldfinger N, Pedel A, Rotter V. Involvement of wild-type p53 in pre-B cell differentiation in vitro. Proc Natl Acad Sci USA. 1991;88:8982– 8986. 28. Cheng J, Yee JK, Yeargin J, Friedmann T, Haas M. Suppression of acute lymphoblastic leukemia by the human wild-type p53 gene. Cancer Res. 1992;52:222–226. 29. Feinstein E, Gale RP, Reed J, Canaani E. Expression of the normal p53 gene induces differentiation of K562 cells. Oncogene. 1992;7:1853–1857. 30. Ryan JJ, Danish R, Gottlieb CA, Clarke MF. Cell cycle analysis of p53-induced cell death in murine erythroleukemia cells. Mol Cell Biol. 1993;13:711–719. 31. Soddu S, Sacchi A. p53: prospects for cancer gene therapy. Cytokines Cell Mol Ther. 1998;4:77–185. 32. Champlin R. Purging: elimination of malignant cells from autologous blood or marrow transplants. Curr Opin Oncol. 1996;8:79 – 83. 33. Rizzo MG, Zepparoni A, Cristofanelli B, et al. Wt-p53 action in human leukaemia cell lines corresponding to different stages of differentiation. Br J Cancer. 1998;77: 1429 –1438. 34. Lowe SW. Cancer therapy and p53. Curr Opin Oncol. 1995;7:547–553. 35. Conseiller E, Debussche L, Landais D, et al. CTS1: a p53-derived chimeric tumor suppressor gene with enhanced in vitro apoptotic properties. J Clin Invest. 1998; 101:120 –127. 36. Sandig V, Brand K, Herwig S, Lukas J, Bartek J, Strauss M. Adenovirally transferred p16INK4/CDKN2 and p53 genes cooperate to induce apoptotic tumor cell death. Nat Med. 1997;3:313–319.