www.elsevier.com/locate/jpedsurg

Intravascular administration of tumor tropic neuralprogenitor cells permits targeted delivery of interferon-bband restricts tumor growth in a murine model ofdisseminated neuroblastoma

Paxton V. Dicksona,b, John B. Hamnera,b, Rebecca A. Burgerc, Elizabeth Garciaf,Annastasia A. Oumae, Seung U. Kimd, Catherine Y.C. Nga, John T. Graye,Karen S. Aboodyf, Mary K. Danksc, Andrew M. Davidoffa,b,*

aDepartment of Surgery, St. Jude Children’s Research Hospital, Memphis, TN 38105, USAbDepartment of Surgery, The University of Tennessee–Memphis Health Science Center, Memphis, TN 38163, USAcDepartment of Molecular Pharmacology, St. Jude Children’s Research Hospital, Memphis, TN 38105, USAdDivision of Neurology, UBC Hospital, University of British Columbia, Vancouver, Canada V6T 2B5eDivision of Experimental Hematology, St. Jude Children’s Research Hospital, Memphis, TN 38105, USAfDivisions of Hematology and Hematopoietic Cell Transplantation and Neurosciences and Beckman Research Institute,

City of Hope National Medical Center, Duarte, CA 91010, USA

0022-3468/$ – see front matter D 2007

doi:10.1016/j.jpedsurg.2006.09.050

Presented at the 37th Annual Meetin

* Corresponding author. Departmen

495 2176.

E-mail address: andrew.davidoff@st

Index words:Interferon;

Neural progenitor cells;

Neuroblastoma

AbstractBackground: Interferon-b (IFN-b) has potent antitumor activity; however, systemic toxicity has limited

its clinical use. We investigated the potential of targeted delivery using tumor-tropic neural progenitor

cells (NPCs) transduced to express human IFN-b (hIFN-b).Methods: Disseminated neuroblastoma was established in SCID mice by tail vein injection of tumor

cells. Fourteen days after tumor cell inoculation, systemic disease was confirmed with bioluminescence

imaging (BLI). Mice were then treated by intravenous injection of human F3.C1 NPCs that had been

transduced with a replication deficient adenovirus to overexpress hIFN-b (F3-IFN-b). Two injections

were given: the first at 14 days and the second at 28 days following tumor cell injection. Control mice

received NPCs transduced with empty vector adenovirus at the same time points. Progression of disease

was monitored using BLI. At sacrifice, organ weights and histology further evaluated tumor burden.

Results: After initiation of therapy, BLI demonstrated a significant decrease in the rate of disease

progression in mice receiving F3-IFN-b. At necropsy, control mice had bulky tumor replacing the liver

and kidneys, as well as extensive retroperitoneal and mediastinal adenopathy. Impressively, these sites

within mice receiving F3-IFN-b therapy appeared grossly normal with the exception of small nodules

within the kidneys of some of the F3-IFN-b–treated mice. The accumulation of F3.C1 cells within sites

of tumor growth was confirmed by fluorescence imaging. Importantly, systemic levels of hIFN-b in the

treated mice remained below detectable levels.

Journal of Pediatric Surgery (2007) 42, 48–53

Elsevier Inc. All rights reserved.

g of the American Pediatric Surgical Association, May 20–24, 2006, Hilton Head, SC.

t of Surgery, St. Jude Children’s Research Hospital, Memphis, TN 38105, USA. Tel.: +1 901 495 4060; fax: +1 901

jude.org (A.M. Davidoff).

Intravascular administration of tumor tropic NPCs 49

Conclusions: These data indicate that in this model of disseminated neuroblastoma, the tumor-tropic

property of F3.C1 NPCs was exploited to target delivery of IFN-b to disseminated tissue foci, resulting

in significant tumor growth delay. The described novel approach for effective IFN-b therapy may

circumvent limitations associated with the systemic toxicity of IFN-b.D 2007 Elsevier Inc. All rights reserved.

as antiviral proteins, type I telencephalon of 15 weeks of gestation. Permission to use

Although originally described

interferons (a/b) also display pleiotropic antitumor effects

[1-5]. However, despite exciting preclinical results, these

agents have achieved only modest success in clinical trials.

Primary limitations include an extremely short half-life upon

administration of recombinant protein as well as significant

systemic toxicity at doses needed to achieve an antitumor

effect [6]. Alternative methods of interferon (IFN) adminis-

tration permitting sustained protein delivery at levels sufficient

to yield antitumor efficacywhile minimizing toxicity are likely

needed to help overcome these limiting factors.

Neural progenitor cells (NPCs) migrate toward sites of

pathology, including tumors, after introduction at local sites of

disease or after intravascular injection [7-9]. Therefore, the

ability to manipulate these cell lines to overexpress a

therapeutic protein of interest makes them potentially valuable

agents in anticancer gene therapy strategies. Early investiga-

tions using NPCs in preclinical models have primarily

examined tumors of and within the central nervous system

(CNS). Recently, work by Brown et al [10] demonstrated the

ability of intravenously administered NPCs to localize to and

infiltrate tumors of both neural and nonneural origin grown

subcutaneously in the flank of nude mice, suggesting the

possibility of targeting tumors outside of the CNS. Our interest

is in determining the potential of using NPCs to target

metastatic solid tumors, particularly neuroblastoma.

Despite development of increasingly aggressive multi-

modality therapies, the prognosis for children with high-risk

neuroblastoma remains poor. As a result, preclinical and

clinical trials evaluating immunotherapy, radionuclide ther-

apy, differentiating agents, and angiogenesis inhibitors are

underway. Prior work in our laboratory has demonstrated

that human IFN-b (hIFN-b) has significant antitumor

efficacy against neuroblastoma xenografts in both localized

and disseminated models of disease [11]. In previous

experiments, we used a liver-targeted adenoassociated viral

vector encoding the cDNA for hIFN-b to mediate contin-

uous, low-dose, systemic protein delivery. In the current

investigation, we have assessed the potential of the neural

progenitor cell line F3.C1 manipulated to overexpress hIFN-

b (F3-IFN-b) to localize to and produce antitumor effects

selectively at disseminated sites of neuroblastoma.

1. Methods

1.1. Cell lines

The F3.C1 neural progenitor cell line was derived from

the primary dissociated cell culture of a human fetal

autopsied human materials was granted (to SUK) by the

Clinical Screening Committee of the University of British

Columbia. Primary cultures were initially grown in UBC1

serum-free medium containing insulin, transferrin, seleni-

um, hydrocortisone, T3, and bFGF, as reported previously

[12]. The F3.C1 clonal cell line was immortalized by

retroviral transduction with v-myc, also as described

previously [12]. The cell line is maintained in Dulbecco’s

minimum essential medium supplemented with 10% fetal

bovine serum in a humidified atmosphere with 10% CO2.

The human neuroblastoma cell line NB-1691 was

engineered to constitutively express firefly luciferase, as

previously described [13]. These NB1691luc cells were

maintained in RPMI-1640 (Mediatech Cellgro, Herndon,

Va) supplemented with 10% heat-inactivated fetal

bovine serum (Mediatech Cellgro), 1% l-glutamine (Medi-

atech Cellgro), 100 U/mL penicillin (GIBCO BRL, Grand

Island, NY), and 100 lg/mL streptomycin (GIBCO BRL).

1.1.1. Adenoviral construction and transduction ofF3.C1 cells

To construct the adenoviral-hIFN-b vector (VAd-IFN-b),the human IFN-b gene was cloned from AAV-CAGG-

hIFN-b using EcoR I and Dpn I and ligated into the multiple

cloning site of pENTR1A (Invitrogen, Carlsbad, CA).

Subsequently, the gene was shuttled into the adenovirus

vector plasmid Ad-Gateway using Gateway recombination

methods (Invitrogen). The resulting E1-E3 deleted vector

expresses hIFN-b via a cytomegalovirus promoter/human

b-globin intron and polyadenylation sequence cassette [14].

Twenty-four hours before injection, F3.C1 cells in log

growth are transduced with VAd-IFN-b at a multiplicity of

infection of 20.

1.1.2. Animal modelAll murine experiments were performed in accordance

with a protocol approved by the Institutional Animal Care

and Use Committee of St Jude Children’s Research

Hospital. Disseminated neuroblastoma was established by

intravenous (IV) injection of 2 � 106 NB-1691luc into 4- to

6-week-old white C.B-17 SCID mice (Charles’ River

Laboratories, Wilmington, Mass). Fourteen days later,

animals were imaged using bioluminescence imaging

(BLI) and separated into 2 groups (n = 9 per group) with

equal signal intensity and disease distribution. Group 1

(treatment group) received tail vein injection of 2 � 106

F3-IFN-b at 14 and 28 days after tumor cell inoculation,

and group 2 (control group), 2 � 106 F3-cytotoxic

T-lymphocyte (CTL) cells on the same schedule. To

P.V. Dickson et al.50

determine circulating levels of hIFN-b in mice treated with

F3-IFN-b vs F3-CTL, mice were bled at 1, 4, 7, and 14

days after injection of these cells. Plasma was collected,

and a commercially available enzyme-linked immunosor-

bent assay (ELISA) (Biosource International, Camarillo,

Calif) was used to measure hIFN-b levels according to the

manufacturers protocol. The reported level of sensitivity for

hIFN-b is 0.1 ng/mL.

In cohorts of mice used to determine F3.C1 localization

to sites of tumor growth, the F3.C1 cells were labeled with

the lipophilic tracer CellTracker CM-DiI (C-7000; Invitro-

gen Molecular Probes, Eugene, Ore) immediately before

injection. Cells were incubated with CM-DiI working

solution, 1 lg/mL in 1� phosphate-buffered saline (PBS),

at 378C for 5 minutes and then 48C for 15 minutes. Cells

were then washed twice in PBS and resuspended for

injection. Animals were given 2 � 106 DiI-labeled F3.C1

cells, sacrificed 3 days later and organs harvested for

histology. DAPI and hematoxylin and eosin–stained sec-

tions were viewed and digitally photographed using an

Olympus U-SPT microscope (Olympus, Center Valley, PA)

equipped with both fluorescence and brightfield illumina-

tion with an attached CCD camera. Images were saved as

JPEG files for further processing in Adobe PhotoShop

(Adobe Systems Inc, San Jose, Calif).

For real-time in vivo assessment of tumor burden,

bioluminescence imaging was performed using an IVIS

Imaging System 100 Series (Xenogen Corporation, Alame-

da, Calif). Mice received an intraperitoneal injection of

d-luciferin (Xenogen) at 15 mg/mL in sterile PBS and were

imaged 5 minutes later. All animals were imaged in a supine

position at a range of 25 cm for 3 minutes. If an image was

saturated, the image time was reduced by 30-second

Fig. 1 Three days after administration to mice with disseminated ne

for the presence of CM-DiI–labeled F3.C1 cells. These cells (red) were d

(C), whereas lack of cells was detected in non–tumor-bearing intestine (D

shown (E-H).

intervals until saturation was eliminated. Acquired images

were analyzed using Living Image Software version 2.50

(Xenogen). Bioluminescence measurements were recorded

as photons per second.

1.1.3. Statistical analysisThe Sigmaplot program (SPSS Inc, Chicago, Ill) was used

to analyze and graphically present the data. Comparisons

between bioluminescence signal intensity as well as animal

and organ weights were made using the Student t test.

A P value of b .05 was considered statistically significant.

2. Results

2.1. Neural progenitor cell tropism for sites ofdisseminated tumor growth

To confirm localization of the F3.C1 cells to sites of

tumor growth, mice with disseminated neuroblastoma were

administered DiI-labeled F3.C1 cells by tail vein injection.

Three days later, animals were killed and organs harvested.

Fluorescence microscopy demonstrated the presence of

F3.C1 cells at sites of tumor growth including the liver,

kidneys, and bone marrow (Fig. 1A-C). These data indicate

that the F3.C1 cell line exhibits tropism to disseminated

neuroblastoma tumor sites within 3 days after IV admin-

istration. Importantly, when non–tumor-bearing organs such

as intestine (Fig. 1D) were assessed, NPCs were not

detected. Because the liver and kidney seemed like

potential sites of accumulation of NPCs following IV

administration, we also injected DiI-labeled F3.C1 cells

into naive (non–tumor-bearing) mice. Subsequent evalua-

tion of tissues revealed only a few NPCs within the liver

uroblastoma, tumor-bearing organs were harvested and evaluated

etectable in tumor bearing liver (A), bone marrow (B), and kidneys

). Corresponding hematoxylin and eosin–stained sections are also

Fig. 2 Bioluminescence imaging to assess tumor burden

demonstrated delay in tumor growth in F3-IFN-b–treated mice

over controls. *P = .028; **P = .001.

Fig. 4 At necropsy, livers of F3-CTL mice were replaced with

bulky disease (A), whereas those of F3-IFN-b–treated mice

appeared normal (B). Hematoxylin and eosin staining of these

organs further demonstrated almost complete replacement of

normal liver parenchyma in control mice (C), whereas only small

nests of tumor cells were detected in F3-IFN-b–treated animals

(D). The differences in liver weights are shown graphically (E).

**P b .001 vs both naive and control mice.

Intravascular administration of tumor tropic NPCs 51

sections compared with the number seen in tumor-bearing

mice and no F3.C1 cells within the kidneys sections of

naive mice.

2.2. Effect of IFN-bb–expressing NPCs ondisseminated disease

We next evaluated the therapeutic effects of F3-IFN-bcells in the model of disseminated neuroblastoma. Fourteen

days after NB-1691 tumor cell inoculation, animals were

divided into 2 groups with equal tumor burden, based on

BLI (8.74 F 1.57 � 105 vs 8.92 F 1.95 � 105 photons per

second, P = .94). Mice were then administered 2 � 106 F3-

IFN-b or F3-CTL cells by tail vein injection. Biolumines-

cence imaging 2 weeks after the initial administration of

NPCs revealed a significant restriction of tumor progression

in F3-IFN-b–treated mice (2.61 F 0.58 � 107photons per

second) vs controls (5.14 F 0.87 � 107photons per second,

P = .028) (Fig. 2). Animals were then administered a second

injection of F3-IFN-b or F3-CTL. Twelve days after this

Fig. 3 On day 40 after tumor cell injection, control mice were

ill-appearing and cachectic, whereas F3-IFN-b–treated mice

remained healthy and maintained normal body weights. *P = .015.

second treatment, 2 mice within the F3-CTL group

succumbed to disease, whereas while the remaining mice

within this group appeared ill and cachectic (animal weight

23.01 F 1.24 g). Impressively, all mice within the F3-IFN-bgroup appeared healthy at this time (animal weight 26.40 F0.51 g, P = .015 vs controls) (Fig. 3). Bioluminescence

imaging further demonstrated a significant delay in tumor

progression in F3-IFN-b–treated mice (1.69 F 0.69 � 109

photons per second) vs F3-CTL (3.13 F 0.50 � 1010pho-

tons per second, P b .001) (Fig. 2).

Due to the severity of disease progression in F3-CTL

mice, animals in both groups were killed for further

evaluation and comparison of disease burden on day 40

after tumor cell injection. Upon necropsy, mice within the

control group were found to have bulky tumor replacing the

liver and kidneys (Figs. 4A and 5A) (liver weight 3.77 F

Fig. 5 At necropsy, kidneys of F3-CTL mice also were replaced

with bulky disease (A), whereas those of F3-IFN-b–treated mice

appeared normal, with the exception of tumor nodules seen in

some mice (B). Hematoxylin and eosin staining of these organs

further demonstrated almost complete replacement of normal

kidney parenchyma in control mice (C), whereas less tumor

burden was detected in F3-IFN-b–treated animals (D). The

differences in kidney weights are shown graphically (E).

**P b .001 vs naive and control mice; #P = .02 vs naive mice.

P.V. Dickson et al.52

0.42 g; kidney weight 1.24 F 0.23 g, P b .001 vs both NPC-

IFN treated and naive mice) (Figs. 4E and 5E) as well as

extensive retroperitoneal and mediastinal adenopathy. In

contrast, these sites within mice receiving F3-IFN-b therapy

appeared grossly normal with the exception of small nodules

within the kidneys in some of these animals (Figs. 4B

and 5B) (liver weight 1.61 F 0.09 vs 1.55 F 0.04 g in naive

mice, P = .32; kidney weight 0.633F 0.03 vs 0.480F 0.03 g

in naive mice, P = .02) (Figs. 4E and 5E). Histologic analysis

of the livers and kidneys within these mice further

highlighted the tumor growth restriction within mice treated

with F3-IFN-b vs controls (Figs. 4C and D and 5C and D). In

mice bled at multiple time points after F3-IFN-b and F3-

CTL injection, plasma levels of hIFN-b were undetectable,

as assessed by ELISA. These data demonstrate that after

intravascular administration, F3-IFN-b cells localize to sites

of disseminated disease and express levels of hIFN-bsufficient to achieve a significant antitumor effect.

3. Discussion

The type I interferons are known to have multiple

antitumor effects including direct inhibition of tumor cell

proliferation through both cell cycle arrest and induction of

apoptosis [1,5] as well as indirect antitumor activity

through immunodmodulation [2,3] and inhibition of angio-

genesis [5,11]. Unfortunately, an extremely short half-life

and significant toxicity have hindered their use in the

clinical setting. An alternative dosing schedule directed at

achieving sustained low-level systemic or tumor-targeted

protein delivery is likely to help circumvent these

limitations. Gene therapy–based approaches for the deliv-

ery of IFNs have shown some promise in this regard. These

strategies have included both systemic [15] and intra-

tumoral [16] administration of adenoviral vectors, as well

as cationic liposomes [17] carrying the IFN-b gene.

Furthermore, we have recently shown that a liver-targeted

adenoassociated viral vector carrying the cDNA for IFN-bmediates continuous, low-dose, systemic protein delivery

and has significant antitumor activity against a variety of

tumor histologies [5,11]. In the current investigation, we

have demonstrated that intravascular administration of

tumor-tropic neural progenitor cells permits targeted

delivery of IFN-b and results in significant antitumor

efficacy without elevating systemic protein levels in a

murine model of disseminated neuroblastoma.

NPCs display remarkable tropism and migratory capacity

to sites of malignant growth. Importantly, these cells can be

modified to stably express a therapeutic transgene. Several

recent investigations have examined their use in cell-based

gene therapy strategies aimed at brain tumors [7,9,18].

Although initial approaches used local or direct intratumoral

administration of NPCs, further work has demonstrated

these cells retain their ability to bhome inQ on sites of tumor

growth after systemic administration [9,10]. Furthermore, it

has been shown that after IV injection, NPCs engineered to

express a lacZ reporter and the prodrug activating enzyme

cytosine deaminase migrated to established subcutaneous

SH-SY5Y human neuroblastoma xenografts [10]. In the

present study, the F3.C1 NPC line localized to sites of

disseminated tumor foci in a model of disseminated

neuroblastoma. When injected IV, NB-1691 neuroblastoma

cells consistently produce disease in the liver, kidneys, bone

marrow, and retroperitoneal lymph nodes. At a point when

these mice have miliary or microscopic disease at multiple

sites, 2 � 106 DiI-labeled F3.C1 cells were administered by

tail vein injection and animals killed 3 days later.

Subsequent histologic analysis demonstrated accumulation

of the F3.C1 cells in or near sites of tumor growth,

demonstrating the ability of these cells to migrate toward

Intravascular administration of tumor tropic NPCs 53

tumors outside of the CNS. Moreover, when injected into

naive mice, a few cells were detected in the liver and none

in the kidney. It is believed that in the absence pathology,

NPCs undergo apoptosis and are very short-lived. Although

only a 3-day time point after F3.C1 injection was evaluated

in the current investigation, NPCs have been found to

localize to sites of tumor growth as early as 30 minutes and

detected as far out as 7 days after injection [9,10]. Further

experiments are being designed to appropriately assess the

arrival and duration of survival of NPCs at sites of

disseminated neuroblastoma.

In this investigation, F3.C1 cells were engineered to

overexpress hIFN-b by adenoviral transduction. Application

of vector to these cells has been found to yield significant

transgene expression within 24 hours with hIFN-b levels

remaining stable in vitro for approximately 2 weeks (data

not shown). To assess the antitumor efficacy of F3-IFN-bcells in vivo, 2 � 106 cells were administered IV to SCID

mice with disseminated NB-1691 tumors. Significant tumor

growth restriction was demonstrated by serial BLI as well as

gross and histologic evaluation at necropsy. Importantly,

when mice were bled at 1, 4, 7, and 14 days after F3-IFN-binjection, plasma levels of hIFN-b were not detectable by

ELISA. These data are consistent with prior experiments

evaluating the antitumor effects of hIFN-b against NB-1691

xenografts [11] and highlight the use of a tumor tropic, cell-

based approach to achieve tumor targeted protein delivery.

The ability to effect sustained F3.C1 expression of hIFN-bat disseminated sites of microscopic disease represents a

novel therapeutic approach in a model designed to mimic a

minimal residual disease state.

In conclusion, the neural progenitor cell line F3.C1

displays tropism for sites of disseminated neuroblastoma

in a murine model and permits targeted delivery of

IFN-b,resulting in significant tumor growth restriction.

Future studies examining the mechanisms of NPC tumor

tropism, the timing and duration of localization, and

efficacy of F3-IFN-b against additional neuroblastoma cell

lines are ongoing.

References

[1] Stark GR, Kerr IM, Williams BR, et al. How cells respond to

interferons. Annu Rev Biochem 1998;67:227-64.

[2] Belardelli F, Ferrantini M, Proietti E, et al. Interferon-alpha in tumor

immunity and immunotherapy. Cytokine Growth Factor Rev 2002;13:

119-34.

[3] Dvorak HF, Gresser I. Microvascular injury in pathogenesis of

interferon-induced necrosis of subcutaneous tumors in mice. J Natl

Cancer Inst 1989;81:497-502.

[4] Slaton JW, Perrotte P, Inoue K, et al. Interferon-alpha–mediated down-

regulation of angiogenesis-related genes and therapy of bladder cancer

are dependent on optimization of biological dose and schedule. Clin

Cancer Res 1999;5:2726-34.

[5] Streck CJ, Dickson PV, Ng CY, et al. Antitumor efficacy of AAV-

mediated systemic delivery of interferon-beta. Cancer Gene Ther

2006;13:99-106.

[6] Einhorn S, Grander D. Why do so many cancer patients fail to respond

to interferon therapy? J Interferon Cytokine Res 1996;16:275-81.

[7] Benedetti S, Pirola B, Pollo B, et al. Gene therapy of experimental

brain tumors using neural progenitor cells. Nat Med 2000;6:447 -50.

[8] Aboody KS, Najbauer J, Schmidt NO, et al. Targeting of melanoma

brain metastases using engineered neural stem/progenitor cells.

Neurooncol 2006;8:119 -26.

[9] Aboody KS, Brown A, Rainov NG, et al. Neural stem cells

display extensive tropism for pathology in adult brain: evi-

dence from intracranial gliomas. Proc Natl Acad Sci U S A 2000;

97:12846-51.

[10] Brown AB, Yang W, Schmidt NO, et al. Intravascular delivery of

neural stem cell lines to target intracranial and extracranial tumors of

neural and non-neural origin. Hum Gene Ther 2003;14:1777-85.

[11] Streck CJ, Dickson PV, Ng CY, et al. Adeno-associated virus vector-

mediated systemic delivery of IFN-beta combined with low-dose

cyclophosphamide affects tumor regression in murine neuroblastoma

models. Clin Cancer Res 2005;11:6020-9.

[12] Kim SU, Nakagawa E, Hatori K, et al. Production of immortalized

human neural crest stem cells. Methods Mol Biol 2002;198:55 -65.

[13] Dickson PV, Ng CY, Zhou J, et al. In vivo bioluminescence imaging

for early detection and monitoring of disease progression in murine

model of neuroblastoma. J Am Coll Surg 1995;201:S53 -4.

[14] Ory DS, Neugeboren BA, Mulligan RC. A stable human-derived

packaging cell line for production of high titer retrovirus/vesicular

stomatitis virus G pseudotypes. Proc Natl Acad Sci U S A 1996;93:

11400-6.

[15] Choi EA, Lei H, Maron DJ, et al. Combined 5-fluorouracil/systemic

interferon-beta gene therapy results in long-term survival in mice with

established colorectal liver metastases. Clin Cancer Res 2004;10:

1535-44.

[16] Zhang F, Lu W, Dong Z. Tumor-infiltrating macrophages are involved

in suppressing growth and metastasis of human prostate cancer cells

by INF-beta gene therapy in nude mice. Clin Cancer Res 2002;8:

2942-51.

[17] Yoshida J, Mizuno M, Wakabayashi T. Interferon-beta gene therapy

for cancer: basic research to clinical application. Cancer Sci 2004;95:

858 -65.

[18] Yip S, Aboody KS, Burns M, et al. Neural stem cell biology may be

well suited for improving brain tumor therapies. Cancer J 2003;9:

189 -204.