Scholarly article on topic 'Clinical and local biological effects of an intratumoral injection of (IL24; INGN 241) in patients with advanced carcinoma: a phase I study'

Clinical and local biological effects of an intratumoral injection of (IL24; INGN 241) in patients with advanced carcinoma: a phase I study Academic research paper on "Clinical medicine"

CC BY-NC-ND
0
0
Share paper
Academic journal
Mol Ther
OECD Field of science
Keywords
{""}

Academic research paper on topic "Clinical and local biological effects of an intratumoral injection of (IL24; INGN 241) in patients with advanced carcinoma: a phase I study"

Clinical and Local Biological Effects of an Intratumoral injection of mda-7 (IL24;INGN 241) in Patients with Advanced Carcinoma: a Phase I Study

C. Casey Cunningham1'* Sunil Chada2,y James A. Merritt2 Alex Tong1 Neil Senzer1 Yuan Zhang1 Abner Mhashilkar2 Karen Parker2 Sasha Vukelja3 Don Richards3 Jill Hood2 Keith Coffee2 John Nemunaitis1

1Mary Crowley Medical Research Center, Dallas, TX 75246, USA 2Introgen Therapeutics, Inc., Houston, TX 77030, USA, and U.S. Oncology, Houston, TX, USA 3Tyler Cancer Center, Tyler, TX, USA

*To whom correspondence and reprint requests should be addressed at 5th Floor, Collins Building, 3535 Worth Street, Dallas, TX 75246, USA. Fax: +1 214 370 1886. E-mail: casey.cunningham@usoncology.com.

yTo whom correspondence and reprint requests should be addressed at Introgen Therapeutics, Inc., Houston. TX 77030, USA. E-mail: s. chada@introgen.com.

The melanoma differentiation-associated gene-7 (mda-7; approved gene symbol IL24) is a tumor suppressor gene whose expression induces selective apoptosis in tumor cells. To characterize the safety and biologic activity of mda-7 gene transfer, we conducted a phase I trial using intratumoral injections of an adenovirus containing the mda-7

1012 vp) in 28 patients with resectable solid tumors. One hundred percent of injected lesions demonstrated INGN 241 vector transduction, transgenic mRNA, elevated MDA-7 protein, and apoptosis induction, with the highest levels near the injection site. Apoptosis of cells in injected tumors was consistently observed even in heavily pretreated patients. INGN 241 vector DNA and mRNA were detected more than 1 cm from the injection site, whereas MDA-7 protein and bioactivity were more widely distributed. Toxicity attributable to the injections was self-limiting and generally mild; however, one patient experienced a grade 3 SAE possibly related to the study drug. Evidence of clinical activity was found in 44% of lesions with the repeat injection schedule, including complete and partial responses in two melanoma patients. Thus intratumoral administration of INGN 241 is well tolerated, induces apoptosis in a large percentage of tumor cells, and demonstrates evidence of clinically significant activity.

Key Words: mda-7, IL-24, apoptosis, bystander, ER, stress, cytokine, secretion, adenovirus, cancer gene therapy, IL-10, IL-19, IL-20, IL-22, receptor

Introduction

The melanoma differentiation-associated gene-7 (mda-7; approved gene symbol IL24) is a tumor suppressor gene identified by subtraction hybridization from human melanoma cells induced to terminally differentiate by IFN-h and mezerein [1]. The mda-7 cDNA encodes a novel 24-kDa protein with low amino acid homology to interleukin 10 (IL-10) and greater homology to the IL-10 family members: IL-19, IL-20, IL-22, and IL-26 [2,3]. Based upon its localization within the IL-10 family cluster at 1q32.2 and cytokine-like properties, mda-7 was designated by HUGO as interleukin 24 [4]. The MDA-7 protein contains a consensus signal sequence and pro-

teolytic cleavage site and mda-7-transfected cells secrete a soluble form of the protein [3,5].

Expression of MDA-7 by adenoviral gene transfer (Ad-mda7;INGN 241) induces apoptotic cell death in cells from a variety of solid tumor types, including melanoma, lung, breast, colorectal, and prostate [3,5-11]. However, the mechanism by which apoptosis occurs varies, depending upon the cell type studied. Further investigations of MDA-7-induced apoptosis are ongoing, but the process is clearly independent of p53, Rb, Ras, Bax, and caspase 3 [5,8]. Significantly, levels of expression of MDA-7 similar to those that induce apoptosis in tumor cells have little or no effect in normal human mammary epithelial cells,

human skin fibroblasts, human endothelial cells, or rat embryo fibroblasts [3,11], implying selective activity limited to malignant cells. mda-7 gene transfer thus offers promise as a new and potentially widely applicable antitumor therapeutic.

Preclinical animal models support this promise. Inhibition of tumorigenicity was seen in a model using ex vivo treatment of the breast cancer line MCF-7 with mda-7 gene transfer [9]. Intratumoral injections of INGN 241 into subcutaneous xenografts of H1299 or A549 lung cancer cells in nude mice demonstrated significant tumor growth inhibition and apoptotic tumor cell death compared to control injections [10,12]. The injected tumors were removed after 48 h and exhibited colocalization of MDA-7 protein and apoptotic marker expression. Intratumoral injection of INGN 241 into Her-2/neu over-expressing breast cancer xenografts resulted in significant inhibition of tumor growth [13]. No toxic effects of INGN 241 administration in these pharmacology studies were identified. GLP toxicology studies of INGN 241 in mice found effects only at the highest dose (5 x 1012 viral particles (vp)/kg;equivalent to a human dose of 3 x 1014 vp), which consisted of decreases in body weight, mild liver toxicity, and transient decreases in platelet counts (unpublished data). Therefore, to begin to characterize the safety and biologic activity of mda-7 gene transfer in a clinical setting, we conducted a phase I trial using intratumoral injections of INGN 241 into resectable solid tumor lesions. In this report, we detail our findings from that study.

Results

Patient Characteristics and Safety Analyses After INGN 241 Treatment

After study entry, all patients underwent a baseline biopsy procedure to provide control uninjected tumor. We treated patients in a prespecified dose-escalation

schema and they received 2 x 1010 to 2 x 1012 vp delivered into the central region of the target tumor; injected tumors were resected to evaluate efficiency of gene transfer and biologic endpoints (Fig. 1). Twenty-eight patients were enrolled in the trial; 22 patients completed at least one cycle of treatment (Table 1). Patients included 13 females and 9 males and ranged in age from 38 to 92, with a median age of 66. Date of initial cancer diagnosis ranged from 1982 to 2002;all patients had been heavily pretreated as shown in Table 1. The majority had received surgery in addition to chemotherapy and/or radiotherapy. The inclusion criteria for this study allowed enrollment of patients with a variety of solid tumors. A total of 15 different tumor types were enrolled in the study: melanoma, breast, SCCHN, colorectal, lymphoma, hepatoma, NSCLC, adenocarcinoma, sarcoma, and carcinomas of the adrenal, bladder, parotid, lip, kidney, and penis. Malignant melanoma (21.4%), SCCHN (18%), breast carcinoma (14%), and colorectal carcinoma (7%) were the most frequent tumor types. Adverse events were generally mild (Table 1); however, there was one grade 3 serious adverse event (SAE) involving fatigue in a cohort 8 patient that was possibly related to the study drug. We removed this patient from the study. We observed no other SAEs considered possibly related to INGN 241. Of the adverse events specifically related to injection of the study drug, the most common were injection site pain and fever, V 2, occurring within 24 h of tumor injection. We saw these effects more consistently at higher doses of INGN 241 and they generally resolved by 48 h postinjection. In two patients in cohort 8 (twice-weekly injections), we noted marked skin erythema surrounding the injected lesion within 24 h of injection. This then resolved over the following 96 h. A maximum tolerated dose was not attained in this study. Overall, the accumulated data from 22 patients completing treatment indicate that INGN 241 is well

fig. 1. Study schema for clinical protocol. The dose levels of INGN 241 and biopsy schedules are indicated. bx, biopsy; rxn, resection.

fig. 2. Excisional biopsy procedure. (1) INGN 241 vector admixed with Isosulfan blue allows identification of the site of injection. (2) Lesion resected 24 h after injection. (3) Postresection processing. The bisected lesion is serially sectioned and the left portion is fixed and evaluated by immunohistochemistry. The right half is sectioned and immediately flash-frozen for quantitative PCR analyses.

tolerated when administered in single or multidosing regimens.

Vector Distribution and Kinetic Profiles After Injection

We analyzed gene transfer and biologic effects elicited by intratumoral injection of INGN 241 in cohorts 1-5 (patients 1-12), in which we resected tumors after 1-4

days. Treated tumors ranged in size from 1.8 x 1.2 to 11.0 x 8.1 cm (average area was 22.7 cm2). We performed quantitative analyses of vector-specific DNA and RNA at the point of injection (center of lesion) and in serial sections to the periphery of each lesion. To facilitate identification of the injection site, we admixed Isosulfan blue dye with vector just prior to administration. Upon resection of tumors (1-4 days later), injec-

fig. 3. Pharmacodynamics of INGN 241 vector and expression. (A) Dose response of INGN 241 vector DNA and RNA. Tumor sections were obtained from preinjected (Pre), proximal, and distal sections from cohort 1-3 patients and vector-specific signals evaluated by quantitative DNA- and RT-PCR. RNA was not available from cohort 2 tumor. (B) Dose response of INGN 241 transgenic MDA-7 expression and TUNEL reactivity. Tumor sections were obtained from preinjected (Pre), proximal, and distal sections from cohort 1-3 patients and vector-specific signals evaluated. (C) Spread of INGN 241 DNA and (D) RNA from injection site. Genomic DNA and RNA were isolated from tumor sections and analyzed using vector-specific primers for INGN 241. Signals were quantitated using real-time PCR. Data are plotted to indicate signals compared to distance from injection site. Correlation coefficient for DNA decay = 0.9 (P < 0.02) and for RNA decay = 0.82 (P < 0.05). (n > 40 samples were used in the analysis.)

tion sites could be identified by the intense blue staining (Fig. 2). Pretreatment (uninjected) lesions exhibited low vector PCR signals: median pretreatment samples contained 100 vector DNA copies/Ag tumor genomic DNA, whereas these lesions had 213 RNA copies/Ag tumor RNA. All injected lesions showed high levels of vector-specific DNA signals in tumors (Fig. 3). We compared vector DNA, vector RNA, transgenic MDA-7, and TUNEL signals across the dose range of 2 x 10102 x 1012 vp (Figs. 3A and 3B). Although the sample size was small, there was indication of a dose-dependent increase in the number of INGN 241 vector DNA copies/Ag DNA in the tumor (7 x 107 copies/Ag in the proximal (central) sections of cohort 1 compared to an average of 2.2 x 108 copies/Ag in cohort 3). We observed a similar indication of dose response for vector RNA, with 2.0 x 104 vector RNA copies found in the central section of tumors injected with low dose INGN 241, whereas 2.3 x 107 copies/Ag were observed in high-dose tumors (Fig. 3A). Note that the distal regions of tumors uniformly showed lower vector signals than the proximal sections. We evaluated parallel sections for trans-genic MDA-7 protein and apoptosis induction via TUNEL assay. All patients demonstrated undetectable MDA-7 staining in the preinjected baseline samples, whereas we found MDA-7 immunostaining in 20% of tumor cells from low dose, 30% of cells at intermediate dose, and 53% of cells from tumors injected with high dose of INGN 241. We did not observe MDA-7 staining in distal regions of tumors injected with low dose vector, whereas we found 5 and 30% positive cells in mid- and high-dose tumors, respectively. Apoptosis induction also trended higher with increased dose, except for the cohort 2 patient (colorectal carcinoma) who exhibited unusually high TUNEL reactivity of 75% after injection of INGN 241 (Fig. 3B).

We evaluated the dynamics of vector spread and subsequent gene expression. We found the highest number of DNA vector copies at the area of injection (e.g., the center of injected lesions averaged greater than 1 x 108 vector DNA copies/Ag) and signal decreased significantly (P < 0.02) with distance from the injection site; 1 cm from the injection site, vector DNA levels had fallen by almost 90% (Fig. 3C). We could detect low levels of vector DNA up to 3 cm from the injection site; however, only 3 tumors showed DNA signals greater than 1 x 106 copies/Ag at 1 cm from the injection site. We also quantitated vector RNA levels and they showed a pattern of distribution similar to that of the vector DNA (Fig. 3D). Vector RNA also was distributed distally from the injection site and, similar to vector DNA, vector RNA signals were significantly reduced (P < 0.05) with distance from the injection site. Both vector DNA and RNA signals showed exponential decay with distance from the injection site. Regression analyses indicated very strong correlations between vector DNA and RNA

signals and distance from injection site (correlation coefficient = 0.82-0.9).

We evaluated the kinetics of vector DNA and trans-genic mRNA for patients who received a single dose of INGN 241. Vector DNA reached maximum levels at the point of injection 24-48 h after injection. If one assumes these signals are cell-associated, then median signals approximated 1000 vector DNA copies per cell, a value 6 logs above preinjection controls (Fig. 4A). Vector DNA signals decreased by almost 3 logs by day 4 and by greater than 4 logs by day 30, although DNA was still detectable above background levels at 30 days after injection (Fig. 4A). Vector-specific mRNA exhibited a distribution and kinetic profile similar to that of vector DNA (Fig. 4B). Median vector-specific RNA signals were more than 4 logs greater in the center of injected lesions compared to uninjected control samples (Figs. 3A and 4B). Vector-specific RNA signals decreased by more than 2 logs by day 4 (Fig. 4B). Samples were not available for day 30 analysis.

fig. 4. Intratumoral pharmacokinetics of vector DNA and mRNA. (A) Decay of INGN 241 vector at injection site. The median number of DNA copies at each time point postinjection is shown; number of patients per sample is indicated below. The number of DNA copies/Ag genomic DNA was converted to illustrate the average number of vector DNA copies per cell—shown above graph. (B) Decay of INGN 241 vector RNA at injection site. The median number of RNA copies/Ag at each time point is shown; number of patients per sample is indicated below.

Transgenic Gene Expression After INGN 241 Injection

We serially sectioned excised tumors and evaluated them for transgenic MDA-7 protein expression and apoptosis. Pretreatment samples were uniformly negative for MDA-7 protein expression and most were negative for TUNEL reactivity (average TUNEL signal was <6% in cohorts 1-5). After INGN 241 injection, all tumors demonstrated substantial MDA-7 immunostaining that ranged from 20 to 90% positive tumor cells in the center of the lesion (Fig. 5A). We detected transgenic MDA-7 beyond the injection site: five of eight (62%) tumors had detectable MDA-7 at 1 cm from the single injection site, and MDA-7 expression was detected more than 3 cm from the injection site. We detected up to 25% MDA-7-stained cells 16 mm from the point of injection. Reproducibly, within each tumor MDA-7 immunostaining was reduced with distance from the injection site. With the exception of patient 2, >90% of MDA-7-staining cells in all biopsies exhibited malignant histological features. The remaining MDA-7-positive cells comprised infiltrating lymphocytes and/or histocytic/reticuloendothelial cells.

Apoptosis staining varied, with up to 80% of tumor cells at the center of the lesion demonstrating TUNEL reactivity. Apoptosis declined with distance from the injection site;five of seven (71%) tumors exhibited TUNEL reactivity beyond 1 cm from the injection site (Fig. 5B). Both MDA-7 and TUNEL staining demonstrated

60-,—

■E oJ-----■------

£ 0 5 10 15 20 25 30 35 n

m distance (mm)

0 5 10 15 20 25 30 35

fig. 5. Spread of MDA-7 protein and biological effect. MDA-7 protein expression correlates with apoptosis. Serial sections from each tumor were evaluated for (A) MDA-7 expression and (B) TUNEL reactivity using immunohistochemistry. Data are plotted to indicate signals compared to distance from injection site. Both MDA-7 and TUNEL staining show strong correlation with distance; correlation coefficient for MDA-7 = 0.69 and for TUNEL = 0.77 (P < 0.02). (n > 30 samples were used in the analysis.)

significant (P < 0.02) linear decay with distance from the injection site. Regions of tumor exhibiting TUNEL staining strongly corresponded to those regions having MDA-7 protein expression;those samples with distal MDA-7 staining also showed distal TUNEL reactivity. Both protein expression and apoptosis reactivity reached maximum levels by 4 days postinjection but had returned to baseline, preinjection levels by day 30 (31, data not shown).

A representative example is shown in Fig. 6, which illustrates the high level of MDA-7 immunostaining after INGN 241 injection and the decay in protein, DNA, and RNA signals with distance from injection site. This patient with a 5 x 5 cm SCCHN lesion was injected with 2 x 1012 vp INGN 241 and resected at 24 h. We observed no MDA-7 signal in the pretreatment lesion, whereas the injection site (center section) exhibited up to 75% of tumor cells staining for MDA-7. Staining intensity decreased with distance, but the distal lesion (18 mm from injection site) still showed regions of strong MDA-7 staining (Fig. 6). Parallel sections were evaluated for apoptosis. The central section showed 50% TUNEL reactivity, whereas the distal section showed 30% staining. The levels of vector-specific DNA and RNA showed strong signals at the injection site, which decreased markedly (more than 1000-fold) at the periphery (Fig. 6). Overall, vector-specific DNA and RNA, MDA-7 protein levels, and TUNEL reactivity demonstrated similar dose response and kinetic and radial concentration gradients, although MDA-7 and TUNEL signals persisted longer and showed enhanced distribution.

Clinical Responses to INGN 241 Injection

In the first six patient cohorts, we excised the injected lesions 24 to 96 h after injection;therefore no conclusions about clinical activity can be drawn. Although we saw minor changes in morphology in the injected lesions of several cohort 7 patients, none qualified as an objective PR using RECIST guidelines. Tumor measurement data were available for only three cohort 7 patients at the end of study: the tumors exhibited SD (stable disease), with 0, 0, and 23% reduction in tumor size (melanoma, colorectal carcinoma, and SCCHN, respectively). Off-study tumor measurements were taken 5-7 weeks from screening. Time from first injection to death for these patients was 347, 51, and 401 days, respectively.

Cohort 8 patients received injections twice weekly for 3 weeks (in a 28-day cycle); five patients completed at least one cycle of treatment. All patients in this cohort had failed multiple prior therapies (Table 2). Two of five patients demonstrated a clinically significant response to INGN 241 injections consisting of at least partial regression of the injected lesion. The most dramatic of these responses was in a 64-year-old female with widely metastatic melanoma (at study entry, she

fig. 6. MDA-7 transgene expression correlates with distribution of vector throughout the tumor. One-half of the tumor was analyzed for MDA-7 protein expression and the other half for vector-specific DNA and RNA levels. The number of DNA copies/Ag genomic DNA and number of RNA copies/^g total RNA are shown for each tumor section. TUNEL reactivity was 50% in the central section and 30% at the periphery (Section 3).

had >10 distinct lesions). Her initial site of treatment was a supraclavicular node measuring 2 x 2 cm at baseline. No appreciable change was noted for the first five injections but by the sixth (and final) injection, a clear decrease in the size of the lesion was apparent and was associated with erythema over the anterior chest (Fig. 7). The erythema resolved and regression continued over the next 2 weeks until there was no clinical evidence of disease at that site (Fig. 7). Subsequently, we began a second course of injections on a lesion on the dorsum of the right hand. The baseline measurement was 1.8 x 2.3 cm and regression was evident by the fifth injection (84% reduction in lesion area). After completing six injections, we excised the residual lesion and on microscopic examination found it to have a marked inflammatory lymphoplas-

macytic infiltrate throughout the residual nodule and surrounding tissue with extensive coagulative necrosis in the tumor. We then treated a third lesion on the anterior right thigh with two cycles of injections. We saw regression of this lesion also after the first course of injections (baseline measurements 3.5 x 3.25 cm decreasing to 2.4 x 3.1 cm after the injection course; 35% reduction in lesion area) but a second set of injections produced no further response. Interestingly, several distant uninjected melanoma lesions also became erythematous during the course of injection of the target lesion, although clinical regression was not seen at these distant sites. This patient is still alive >600 days after initiating INGN 241 treatment. An additional melanoma patient exhibited a partial response (33% decrease by RECIST).

fig. 7. Objective clinical response to INGN 241 in a patient with metastatic melanoma. Cohort 8 patient with metastatic melanoma. Injected lesion was on right clavicle (dashed circle in (A)). (B) By day 4, region is inflamed. (C) At the end of cycle 1 (day 30), lesion has completely regressed. This patient is still alive >600 days after initiating treatment.

TABLE 1: Patient characteristics and dosing

Cohort No. of patientsa Age (range) Previous treatmentsb Dose (vp) Adverse eventsc

1 1 (1) 49 S, C, Ca, N, H, G 2 > < 1010 2

2 1 (1) 44 S, F, I 2 > < 1011 1

3 3(3) 74 (66-76) S, RT, T, P, C, V, D, Fl, Cl 2> < 1012 0/1/1

4 3 (3) 45 (38-60) S, A, C, T, Ta, N, P, RT, E 2> < 1012 0/1/0

5 4 (3) 75 (65-75) S, Ta, RT, I, D 2> < 1012 0/2/2

6 1 (1) 57 S, A, T, Ta 2> < 1012/divided doses 0

7 7 (5) 86 (76-92) S, M, V, A, P, T, G, RT, F, I 2> < 1012 2/0/0/2/2

8 8 (5) 64 (62-91) S, RT, IF, IT, P, T, F 2> < 1012 repeated 2 x/week 2/2/2/1/1/d

for 3 weeks

a Number of patients enrolled per cohort is indicated. The number of patients completing at least one cycle of treatment is shown in parentheses.

b Prior treatments: S, surgery; A, adriamycin; C, cyclophosphamide; Ca, capecitabine; Cl, chlorambucil; D, dacarbazine; E, etoposide; F, 5-fluorouracil; Fl, fludarabine; G, gemcitabine; H, herceptin; I, irinotecan; IF, IFN-a; IT, immunotherapy; M, methotrexate; N, navelbine; P, platinum; RT, radiotherapy; T, taxane; Ta, tamoxifen; V, vincistine. c Adverse events possibly or probably related to INGN 241 administration are indicated. d One cohort 8 patient experienced a grade 3 SAE and withdrew from the study.

We saw a less dramatic response in another patient with squamous cell carcinoma of the penis with multiple skin nodules in the groin and right hip area. Injection of INGN 241 into one of the upper right hip lesions (2.5 x 3.0 cm at baseline) produced significant central necrosis with surrounding erythema by the fifth injection. However, the lesion continued to expand peripherally so that by completion of the first set of six injections, there was an indurated rim of erythematous tissue surrounding the central area and the total measurement of the lesion was now 3.0 x 4.0 cm. A large central portion of the lesion (approximately two-thirds of the total area) remained blackened and necrotic at the end of cycle 1. Other, new lesions were rapidly appearing in the region, so we removed the patient from the study as PD and he went on to other therapy. Three additional patients (with adenocarcinoma, NSCLC, and lip carcinoma) exhibited stable disease after INGN 241 injection and one patient with SCCHN exhibited disease progression (20% increase by RECIST). Of the nine lesions treated in cohort 8 patients, four demonstrated objective response (CR or PR by RECIST criteria). Five patients are still alive 600-1160 days after INGN 241 injection.

Discussion

mda-7 is a tumor suppressor gene encoding a proinflammatory cytokine. mda-7 gene transfer demonstrates potent growth arrest and cell death in a variety of preclinical tumor models [3,11]. Perhaps most significantly, expression of MDA-7 protein at levels giving rise to these anti-tumor effects does not produce similar cytotoxicity in a variety of normal cell lines. In this report, we detail the initial clinical experience with MDA-7 in patients with refractory cancer. Direct injection of an adenovirus containing the mda-7 cDNA (INGN 241) into a variety of solid tumors was generally well tolerated, with injection site pain and erythema being noted locally in some patients. One patient

experienced a grade 3 SAE of fatigue and was discontinued from study.

Analysis of serial sections of injected lesions demonstrated that MDA-7 DNA and RNA were detectable in 100% of the injected lesions, with the highest concentrations found at the site of injection, as expected (Fig. 3). However, MDA-7 protein was detectable in the periphery of injected lesions (greater than 3 cm from injection site), beyond the area of DNA spread, suggesting that the MDA-7 protein can diffuse from transduced cells (compare Figs. 3C, 3D, 5, and 6). Immunohistochemical staining of tumor sections showed regions of intense punctate MDA-7 staining surrounded by regions of more diffuse staining, which may be reflective of active secretion of MDA-7 from transduced cells (Fig. 6). These data are consistent with our preclinical studies in which we have demonstrated active secretion of glycosylated MDA-7 after INGN 241 transduction [5,17]. Glycosylated human MDA-7 mediates induction of immune, antiangiogenic, and cytotoxic bystander effects [3-5,18,30].

Apoptosis, as measured by TUNEL assay, was significant in the injected lesions and correlated geographically with MDA-7 protein expression (Figs. 5 and 6). Both MDA-7 transgenic protein expression and apoptosis decayed with similar intratumoral dynamics (compare Figs. 5 and 6). In lesions injected and then not biopsied until day 30, protein expression and apoptosis had returned to baseline levels (data not shown). Such a decrease is to be expected since the INGN 241 virus is nonreplicating, but also illustrates the need for repeat dosing to maintain levels of protein expression. In this regard, it is significant that the objective clinical responses were seen in the cohort of patients receiving repeat injections. If protein expression, and by extension apoptosis, is kinetically limited then repeat dosing will provide greater therapeutic benefit.

A unique feature of this study was the inclusion of Isosulfan blue as marker dye to localize injection site. The ability to identify precisely the injection site

facilitated comprehensive pharmacodynamic and kinetic analyses of vector distribution within human tumors. One hundred percent of injected tumors evaluated in our study demonstrated transgenic MDA-7 expression and elevated apoptosis induction compared to untreated control tumors. In contrast, adeno-viral-mediated gene transfer of p53 by injection into non-small-cell lung tumors resulted in vector detection by DNA PCR in 86% of patients' tumors and increased apoptosis in 46% of biopsy specimens [19]. Indeed, as most previous gene transfer studies have not been structured to assess the geographic extent of vector distribution and biologic effect, it is difficult to compare INGN 241's effects directly with other studies. A recent report has indicated that administration of Ad-p53 to glioma is hampered by the limited spread (<5 mm) of transduced cells [20]. Both INGN 241 DNA and transgenic MDA-7 protein distribute to a greater extent (see Figs. 3, 5, and 6) and demonstrate strong correlations with distance from injection site (P < 0.02). Nucleic acid signals decay by 50% at 4-6 mm from the injection site whereas MDA-7 protein and apoptosis signals decay by 50% at 18-20 mm. Therefore, either INGN 241 is more potent than other therapeutic constructs at inducing apoptosis in human tumors or the study design allowed us to capture signals that were lost in other studies. The levels of apoptosis induction observed in this study are substantially higher (average = 45%) than reported with other anti-cancer drugs. Apoptosis induction correlates with loss of Ki-67 staining [31]. It is noteworthy that seven different tumor types were treated in cohorts 1-6 and all showed high levels of MDA-7 expression and subsequent apoptosis induction. As discussed above, all these patients had been heavily pretreated and failed multiple treatment regimens (Table 1). It would be predicted that this group of chemo- and radio-resistant tumors would have acquired resistance to apoptosis. The fact that MDA-7 induces apoptosis in such a spectrum of advanced tumor types mirrors our

preclinical studies and suggests that INGN 241 may have broad utility.

MDA-7 (IL-24) also acts as an interleukin, stimulating IL-6, TNF-a, GM-CSF, and IFN-g production [4]. An interesting avenue for further investigation is then the potential systemic immunomodulatory effects of intra-tumoral MDA-7 [3,29,32]. The significant apoptosis in concert with downregulation of TGF-h [10] suggests the potential for enhanced dendritic cell priming and maturation at the local injection site. We have previously demonstrated MDA-7 immunostaining in dendritic cells in human tumors and tonsil [3,21], and recent data indicate an adjuvant effect of mda-7 gene transfer in syngeneic murine tumor models [3]. Therefore, it is intriguing that one patient developed erythema of distant lesions following injection of one primary site. If this represented an immune response, however, it was not robust enough to lead to clinical response of the distal lesions.

Preclinical studies evaluating activity of Ad-mda7 in various tumor models have revealed a complex interplay between direct tumor cell killing and other complementary (bystander) anti-tumor mechanisms. For example, Ad-mda7 is able to kill tumor cells by activating a variety of proapoptotic signaling pathways (p53, PKR, PTEN, TRAIL, JNK) and concomitantly inhibiting survival/onco-genic pathways (PI3K, bcl-2, h-catenin) [13,17,22-28]. MDA-7 can elicit death in tumor cells via either intra-cellular or extracellular pathways [17,30]. Ad-mda7 induces a stress response signal from the endoplasmic reticulum via caspases 7 and 12, leading to mitochondrial disruption and intrinsic apoptosis [17]. This pathway appears to predominate in cells that lack receptors for MDA-7. In cells expressing cognate MDA-7 receptors (IL-20R1/IL-20R2 and IL-22R1/IL-20R2), ligand engagement activates alternate signaling pathways that result in tumor cell death ([30], unpublished data). Thus human MDA-7 protein can kill tumor cells in either a receptor-dependent or a receptor-independent fashion. Recombinant MDA-7 produced from nonmammalian sources does not share

TABLE 2: Responses in cohort 8 patients

Patient Gender/age Diagnosis Date Previous treatments No. of Response Time to

diagnosed injections deatha

81 M/71 Adeno-carcinoma 1998 Carboplatin/taxane/ gemzar 2b SD n.d.

83 F/64 Melanoma 1994 Surgery/RT/Immuno-tx/IFN-a 24c CR, PR, SD >600

84 F/62 Melanoma 2000 Surgery/RT/Immuno-tx/IFN-a 12 PR 309

85 M/64 Penile carcinoma 2002 Surgery, RT/CDDP/ taxane 6 PD 75

86 F/66 NSCLC 2001 RT/carboplatin/taxane 3b SD 180

87 M/62 SCCHN 1992 Surgery/RT/CDDP/ taxane/ 6 PD 185

5-FU/carboplatin

88 M/91 Lip carcinoma 1982 Surgery/RT 6 SD 181

SD, stable disease; CR, complete response; PR, partial response; PD, progressive disease. a Time in days from first injection of INGN 241 until death. b Patients 81 and 86 did not complete one full course of treatment (6 injections). c Patient received 12 injections on compassionate use protocol.

these properties [17,30]. The active secretion and receptor utilization of MDA-7 may help to explain both the wide distribution of ectopic MDA-7 in human tumors and the widespread radius of effect. We hypothesize that intra-tumoral injection of INGN 241 results in high local concentration of INGN 241 within tumors and diffusion of bioactive MDA-7 protein throughout the tumor. Furthermore, vector DNA and RNA signals decrease faster than MDA-7 protein and apoptosis (Figs. 3C, 3D, and 6), supporting a bystander activity for MDA-7 [30].

In addition to apoptosis induction of tumor cells and immunostimulation, additional bystander activities have been reported for MDA-7. Ad-mda7 was reported to inhibit endothelial differentiation in vitro, implying antiangiogenic activity [10]. Further studies demonstrated antiangiogenic activity of Ad-mda7 in vivo [12], in that MDA-7 expression significantly repressed angiogenic mediators, including VEGF, basic FGF, and IL-8. Studies using purified MDA-7 demonstrated that it functions via the IL-22R1 on endothelial cells [18] and MDA-7 was 50-fold more active than endostatin or angiostatin. When Ad-mda7 was combined with XRT, synergistic inhibition of tumor growth was observed, with significant reduction in microvessel density and pronounced apoptotic response in tumors [12]. Two recent preclinical studies have provided support for intriguing observations about the role of MDA-7 in melanoma disease progression and metastasis. Initial studies found that MDA-7 protein was expressed in normal human melanocytes and benign nevi, but expression was lost in metastatic melanoma [6]. Subsequent studies evaluated larger patient groups and concluded that MDA-7 protein expression is progressively lost as melanoma tumors invade and become more metastatic and aggressive [21]. These authors speculated that MDA-7 must play a role in maintenance of normal physiology of melanocytes and that inhibition of MDA-7 results in transformation and progression from a local, nonmetastatic primary tumor to a highly metastatic phenotype. Additional studies have shown that Ad-mda7 can inhibit cell migration and invasion of lung and breast tumor cells by downregulating the PI3K pathway and inhibiting production of FAK and matrix metalloproteases [22,27]. Tumor cells expressing MDA-7 demonstrated significantly reduced lung metastases compared to control cells [27,32]. It is clear that tumor cells can develop resistance to cytotoxic therapies, and reports are now emerging of acquisition of resistance to pathway-specific molecularly targeted therapeutics. The anti-tumor effects mediated by MDA-7 encompass a variety of signaling pathways, and it is anticipated that redundant proapoptotic signals are activated.

In summary, INGN 241 can induce apoptosis in a large percentage of tumor volume following a single intra-tumoral injection;however, clinically significant responses are primarily seen with repeat injection. Future

studies, therefore, will concentrate on repeat dosing of INGN 241, particularly in malignant melanoma, in which the greatest clinical activity was seen in this study. Other future directions include developing systemic administration of INGN 241, given its widespread tumor selectivity, exploration of its immunopotentiating effects, and its use in combinatorial strategies.

Patients and Methods

The clinical protocol used in this study was reviewed by the Biosafety Committee of each participating institution and the U.S. Food and Drug Administration. Written informed consent was obtained from all patients stating that they were aware of the investiga-tional nature of this study, in keeping with institutional and federal guidelines. For inclusion in this study, patients had to be at least 18 years of age with histologically confirmed carcinoma and at least one lesion accessible for needle injection. A Karnofsky performance status of >70% was required as was acceptable hematologic, renal, and hepatic function. No patient with active CNS metastases, chronic immuno-suppressive use, or prior participation in a therapy requiring the administration of adenovirus was allowed.

Description of agent

INGN 241 is constructed from an adenoviral vector deleted in the E1 and a small portion of the E3 regions to render it replication defective. The vector contains a transgene region encoding a wild-type human mda-7 cDNA driven by the cytomegalovirus immediate early promoter and ending in an SV40 polyadenylation signal and has been described previously [5]. INGN 241 was provided as a frozen vial suspension (3.0-ml vial) at a concentration of 1 x 1012 vp/ml (5 x 1010 PFU/ml) in a neutral buffer containing saline and 10% glycerol. The vials were stored at <60°C. Prior to injection, vials were thawed to ambient temperature and then mixed with 5% glucose solution to obtain the appropriate number of viral particles in a total volume of 2.0 ml. To aid in localization of the injection site, 0.4 ml of Isosulfan blue dye was added just prior to injection. Previous studies demonstrated that Isosulfan blue did not negatively impact vector trans-duction (unpublished data).

Study design

Patients were treated in the outpatient research centers of the Mary Crowley Medical Research Center at Baylor University Medical Center (Dallas, TX, USA) or the Tyler Cancer Center (Tyler, TX, USA). All patients signed the U.S. Oncology Institutional Review Board approved consent form prior to entry on the trial.

Patients were initially enrolled in sequential cohorts receiving doses of the INGN 241 construct escalating

from 2 x 1010 to 2 x 1012 vp (see Table 1 and Fig. 1). The injections were placed in the center of the accessible target tumor lesion. In the first three cohorts (encompassing the dose range 2 x 1010 to 2 x 1012 viral particles), the injected lesion was resected 24 h after injection. The next two patient cohorts continued to receive 2 x 1012 vp without escalation, but with tumor resection at either 48 (cohort 4) or 96 h (cohort 5) postinjection. The resected lesions were bisected (see schematic in Fig. 2): one hemilesion was serially sectioned, fixed in paraformaldehyde, and paraffin embedded (for immunohistochemical evaluation) and the other hemilesion was serially sectioned and immediately frozen in liquid nitrogen (for quantitative nucleic acid evaluation). Samples were analyzed to determine the radius of diffusion of injection solution, the distribution and concentration of the viral agent, mRNA and protein expression, and the resultant biologic effects on the tumor cells (Fig. 2). Blood was also sampled for vector DNA. A single patient cohort (cohort 6) received a total dose of 2 x 1012 vp divided into 10 deposit sites with injections of 0.2 ml at each site, with subsequent resection of the lesion 48 h postinjection.

To assess longer term effects of MDA-7 expression, as well as the effects of repeat injections, the final two cohorts did not undergo excision but instead had incisional biopsies performed pretreatment and 30 days posttreat-ment. In one cohort (cohort 7), a single injection was carried out on day 1 and then an incisional biopsy was performed 30 days later. In the other cohort (cohort 8), 2 x 1012 vp were injected twice weekly for 3 weeks of a 28-day cycle, with a maximum of two cycles permitted. Tumor responses were measured at the end of each cycle and an incisional biopsy was carried out 30 days after the last injection. Tumor responses were evaluated using the RECIST methodology for each indicator lesion. All patients were monitored throughout for development of toxicity related to either the agent or the injection. The full treatment schema is outlined in Fig. 1.

Laboratory analyses

From the biopsy specimens, frozen tissue was obtained for quantitative DNA PCR for INGN 241 DNA and for RT-PCR for mda-7 mRNA. Paraffin sections of biopsy specimens were analyzed for morphology, immunohis-tochemistry staining for the MDA-7 protein, and TUNEL analysis of the percentage of apoptotic cells. A previously described automated immunoperoxidase staining technique [14] was used to characterize protein expression following antigen retrieval, with the use of the Ventana 320 ES System (Ventana Medical Systems, Tucson, AZ, USA) and the avidin-biotin-complexed immunoperoxidase reaction (DAB detection kit;Ven-tana Medical Systems) following initial incubation with affinity-purified rabbit anti-human MDA-7 antibody (Introgen Therapeutics). A TUNEL method (DeadEnd

Colorimetric Apoptosis Detection System; Promega) was used to detect DNA fragmentation in situ. The frequency of MDA-positive or TUNEL-positive cells was determined as an averaged value of the proportion of positive-cells in three 100 x power microscopic fields of representative staining pattern. Tumor cell types were distinguished from normal cells and/or infiltrating lymphoid cells by histological criteria by a trained histopathologist.

To characterize systemic immune activation, ELISAs (R&D Quantikine kits;Minneapolis, MN, USA) were used to quantify patient serum cytokine levels at defined time points before and after treatment [15]. In addition, flow cytometric immunophenotype analysis was carried with a two-color immunofluorescence reaction to determine the effect of INGN 241 on peripheral blood frequency distribution of T, B, and NK subsets [16]. More extensive descriptions of these studies are provided in another article [31].

RECEIVED FOR PUBLICATION MAY 18, 2004; ACCEPTED SEPTEMBER 28, 2004.

References

1. Jiang, H., et al. (1995). Subtraction hybridization identifies a novel melanoma differentiation associated gene, mda-7, modulated during human melanoma differentiation, growth and progression. Oncogene 11: 2477-2486.

2. Kotenko, S. V. (2002). The family of IL-10-related cytokinesand their receptors: related, but to what extent? Cytokine Growth Factor Rev. 13: 223-240.

3. Chada, S., et al. (2004). MDA-7/IL-24 is a unique cytokine in the IL-10 family. Int. Immunopharmacol. 4: 649-667.

4. Caudell, E. G., et al. (2002). The protein product of the tumor suppressor gene, melanoma differentiation-associated gene 7, exhibits immunostimulatory activity and is designated IL-24. J. Immunol. 168: 6041 -6046.

5. Mhashilkar, A. M., et al. (2001). Melanoma differentiation associated gene-7 (mda-7): a novel anti-tumor gene for cancer gene therapy. Mol. Med. 7: 271 -282.

6. Ekmekcioglu, S., et al. (2001). Down-regulated melanoma differentiation associated gene (mda-7) expression in human melanomas. Int. J. Cancer 94: 54-59.

7. Saeki, T., et al. (2001). Tumor-suppressive effects by adenovirus-mediated mda-7 gene transfer in non-small cell lung cancer cell in vitro. Gene Ther. 7: 2051 -2057.

8. Jiang, H., et al. (1996). The melanoma differentiation associated gene mda-7 suppresses cancer cell growth. Proc. Natl. Acad. Sci. USA 93: 9160-9165.

9. Su, Z. Z., et al. (1998). The cancer growth suppressor gene mda-7 selectively induces apoptosis in human breast cancer cells and inhibits tumor growth in nude mice. Proc. Natl. Acad. Sci. USA 95: 14400-14405.

10. Saeki, T., et al. (2002). Inhibition of human lung cancer growth following adenovirus-mediated mda-7 gene expression in vivo. Oncogene 21: 4558-4566.

11. Fisher, P. B., et al. (2003). mda-7/IL-24, a novel cancer selective apoptosis inducing cytokine gene: from the laboratory into the clinic. Cancer Biol. Ther. 2(4 Suppl. 1): S23 -S37.

12. Nishikawa, T., Ramesh, R., Chada, S., and Meyn, R. E. (2004). Suppression of tumor growth and angiogenesis by adenovirus-mediated mda-7/IL-24 gene transfer in combination with ionizing radiation. Mol. Ther. 9: 818-828.

13. McKenzie, T., Liu, Y., Fanale, M., Swisher, S. G., Chada, S., and Hunt, K. K. (2004). Combination therapy of Ad-mda7 and trastuzumab increases cell death in Her-2/Neu-overexpressing breast cancer cells. Surgery 136: 437-442.

14. Nemunaitis, J., et al. (1998). Prognostic value of K-ras mutations, ras oncoprotein, and c-erb B-2 oncoprotein expression in adenocarcinoma of the lung. Am. J. Clin. Oncol. 21: 155-160.

15. Nemunaitis, J., et al. (2001). Intravenous infusion of a replication-selective adenovirus (0NYX-015) in cancer patients: safety, feasibility and biological activity. Gene Ther. 8: 746-759.

16. Zhang, X., Qian, J., and Li, J. (2000). [Clinical and experimental analysis of myeloid antigen positive childhood acute lymphoblastic leukemia]. Zhongguo Shi Yan Xue Ye XueZa Zhi 8: 280-282.

17. Sieger, K. A., et al. (2004). The tumor suppressor activity of MDA-7/IL-24 is mediated by intracellular protein expression in NSCLC cells. Mol. Ther. 9: 355-367.

18. Ramesh, R., et al. (2003). Melanoma differentiation-associated gene 7/interleukin (IL)-

24 is a novel ligand that regulates angiogenesis via the IL-22 receptor. Cancer Res. 63: 5105-5113.

19. Swisher, S. G., et al. (1999). Adenovirus-mediated p53 gene transfer in advanced non-small-cell lung cancer. J. Natl. Cancer Inst. 91: 763-771.

20. Lang, F. F., et al. (2003). Phase I trial of adenovirus-mediated p53 gene therapy for recurrent glioma: biological and clinical results. J. Clin. Oncol. 21: 2508-2518.

21. Ellerhorst, J. A., et al. (2002). Loss of MDA-7 expression with progression of melanoma. J. Clin. Oncol. 20: 1069-1074.

22. Mhashilkar, A. M., et al. (2003). MDA-7 negatively regulates the h-catenin and PI3K signaling pathways in breast and lung tumor cells. Mol. Ther. 8: 207-219.

23. Pataer, A., et al. (2002). mda-7 induces apoptosis via upregulation of the double-stranded RNA dependent kinase PKR. Cancer Res. 62: 2239-2243.

24. Pataer, A., Chada, S., Hunt, K. K., Roth, J. A., and Swisher, S. G. (2003). Adenoviral MDA-7 induces apoptosis in lung cancer cells through mitochondrial permeability transition independent cytochrome c release. J. Thoracic Cardiovasc. Surg. 126: 1325 -1328.

25. Cao, X. X., et al. (2002). Adenoviral transfer of mda-7 leads to BAX up-regulation and apoptosis in mesothelioma cells, and is abrogated by over-expression of BCL-XL. Mol. Med. 8: 869 -876.

26. Ekmekcioglu, S., et al. (2003). Negative association of melanoma differentiation

associated gene (mda-7) and inducible nitric oxide synthase (iNOS) in human melanoma: MDA-7 regulates iNOS expression in melanoma cells. Mol. Cancer Ther. 2: 9-17.

27. Ramesh, R., Ito, I., Gopalan, B., Saito, Y., Mhashilkar, A. M., and Chada, S. (2004). Ectopic production of MDA-7/IL-24 inhibits invasion and migration of human lung cancer cells. Mol. Ther. 9: 510-518.

28. Kawabe, S., Nishikawa, T., Munshi, A., Roth, J. A., Chada, S., and Meyn, R. (2002). Adenovirus mediated mda-7 gene expression radiosensitizes non-small lung cancer cells via p53-independent mechanisms. Mol. Ther. 6: 637-644.

29. Chada, S., Ramesh, R., and Mhashilkar, A. M. (2003). Cytokine- and chemokine-based gene therapy for cancer. Curr. Opin. Mol. Ther. 5: 463-474.

30. Chada, S., et al. (2004). Bystander activity of Ad-mda7: human MDA-7 protein kills melanoma cells via an IL-20 receptor-dependent but STAT3-independent mechanism. Mol. Ther. (in press).

31. Tong, A., et al. (2004). Intratumoral injection of INGN 241, a non-replicating adenovector expressing the melanoma-differentiation associated antigen-7 (MDA-7/IL-24): biologic outcome in advanced cancer patients. Mol. Ther., (in press).

32. Ramesh, et al. (2004). Liposome-mediated mda7/IL-24 gene delivery inhibits human lung cancer growth in in vivo. DNA Cell Biol., (in press).