Clinical Trials of Gene Therapy, Virotherapy, and Immunotherapy for Malignant Gliomas

REVIEW Clinical trials of gene therapy, virotherapy, and immunotherapy for malignant gliomas L Barzon1, M Zanusso2, F Colombo2 and G Palu` 1 1Department of Histology, Microbiology and Medical Biotechnologies, University of Padova, Padova, Italy and 2Division of Neurosurgery, San Bortolo Hospital, Vicenza, Italy

Despite advances in surgical and adjuvant therapy, the prognosis for malignant gliomas remains dismal. This gloomy scenario has been recently brightened by the increasing understanding of the genetic and biological mechanisms at the basis of brain tumor development. These findings are being translated into innovative therapeutic approaches, including gene therapy, virotherapy, and vaccination, some of which have already been experimented in clinical trials. The advantages and disadvantages of all these different therapeutic modalities for malignant gliomas will be critically discussed, providing perspective for future investigations. Cancer Gene Therapy (2006) 13, 539–554. doi:10.1038/sj.cgt.7700930; published online 6 January 2006 Keywords: cell therapy; immunotherapy; oncolytic virus; clinical trial; malignant glioma

Introduction article focuses on the results from these clinical trials, with particular emphasis on the most promising achievements, Malignant gliomas remain among the mostly lethal which should be kept as the basis for further investiga- cancers in humans, with an overall incidence of about 3 tions and for the design of controlled phase II and III per 100 000 people per year1 and a median survival of less studies. than 1 year.2 Standard treatment consists of surgery followed by radiotherapy and sometimes adjuvant chemotherapy. Eventually almost all malignant gliomas recur and, at recurrence, the median survival is 2–3 months.2 Gene therapy approaches The reason for the failure of conventional therapy is inherent to the intrinsic biological properties of malignant More than a decade has elapsed since the first clinical study of gene therapy for patients with malignant glioma cells, which are highly infiltrative into the brain, 3 thus preventing total tumor resection, and resistant to gliomas and more than 40 trials have been approved chemotherapy and radiotherapy. Thus, the goal of novel since then (Figure 1). Gene transfer techniques to the treatment strategies for malignant gliomas should be brain have been generally based on the direct injection of twofold: improvement of tumor debulking and enhance- the gene transfer vector or vector producer cells (VPCs) ment of tumor cell killing. These aims have been pursued into the tumor mass or into the surgical cavity margin by gene therapy, virotherapy, adoptive immunotherapy, after tumor debulking. The most frequently used ther- and vaccination studies, which have been translated into apeutic genes in clinical trials include prodrug activating several clinical trials. Although the results – in terms of genes (suicide genes) and cytokine genes, and, less tumor response – achieved so far from clinical trials are frequently, tumor suppressor genes (e.g., TP53). Results still rather poor, a positive remark derives from the are now available from several clinical trials. Unfortu- tolerability and safety of treatment that has been nately, they are rather disappointing in terms of clinical demonstrated in almost all studies. It should however be response, even though treatment procedures and gene transfer have been generally proven to be safe and well noted that the majority of clinical studies were pilot or 4,5 phase I protocols aiming at assessing the feasibility and tolerated (Table 1). Reasons for such failure have been safety of treatment, rather than efficacy. This review mainly ascribed to inefficient transduction of target tumor cells (as reported for other types of cancers), rather than to the gene therapy strategy per se.

Correspondence: Dr L Barzon, Department of Histology, Microbiology and Medical Biotechnologies, University of Padova, Gene transfer vectors Via A Gabelli 63, I-35121 Padova, Italy. In gene therapy clinical trials for malignant gliomas, genes E-mail: [email protected] are generally delivered by means of retroviral vectors or Received 21 July 2005; revised 21 October 2005; accepted 24 October adenoviral vectors, though nonviral vectors (liposomes) 2005; published online 6 January 2006 are also employed. Cell and gene therapy for malignant gliomas L Barzon et al 540 resection. However, at such high doses, their delivery may lead to severe side effects. In fact, phase I dose escalation trials showed that doses of 1012 vp may be accompanied by severe toxicity, with CNS symptoms, fever, leukocytosis, and hyponatremia,10 whereas lower doses are well tolerated.10,16,19,22 Toxicities associated with adenoviral vectors are generally consistent with a transient inflam- matory response to the vector, as observed in other clinical trials with this type of vectors.4 On the other hand, cellular and humoral immune response elicited by adenoviral vectors undoubtedly contributes to the antitumor activity of gene therapy, even though the antibody response against adenoviral vectors, which can pre-exist in the host or develop after treatment, even in the case of intracerebral injection, could interfere with vector deliv- Figure 1 Number of approved clinical trials of gene therapy, ery.10,19 Adenoviral vectors can transduce both dividing virotherapy, and vaccination for malignant gliomas (data from the and nondividing cells and are therefore more efficient literature and from www.wiley.co.uk/genmed/clinical/, last visit July 20, 2005). A gene transfer trial of the MDR1 multidrug resistance than retroviral vectors against tumors with low mitotic gene in patients undergoing high-dose chemotherapy and auto- index. In the site of vector injection, 95–100% of cells are logous stem-cell transplantation is also included.91 transduced and express the transgene, but transduced cells reside within only a short distance from the injection site, with a mean maximal distance of about 5 mm.19 Replica- Viral vectors. Retroviral vectors – As for retroviral tion-competent adenoviral vectors have never been vectors, Moloney-derived retroviral vectors are generally detected in any biological sample from glioma patients used and administered to the patient by direct injection as well as from patients with other types of malignancies of murine retroviral vector-producing cells (RVPCs) treated with adenoviral vector injection,10,16,19,22 even (Table 1). The rationale for using retroviral vectors is though vector sequences may be transiently detected in based on their property to transduce only dividing cells, the peripheral blood shortly after gene therapy.22 such as glioblastoma cells, and thus to spare neurons, which are typically resting cells. Transduction rate Nonviral vectors. Nonviral vectors, such as cationic achieved in vivo by intratumor RVPC injection is however liposomes, have the advantages of low immunogenicity rather low, and estimated to be below 0.00211 or 0.03%25 and toxicity, easy manufacturing, and long-term stability. in two different clinical studies. Transduction efficacy Cationic liposomes have been used to deliver the suicide could probably be improved, since much higher transduc- gene thymidine kinase of HSV-TK to eight patients with tion rates have been reported in animal models of GBM.20,21 The liposome–DNA complex was adminis- glioma26 and of other types of cancer.27 Inefficient tered by convection-enhanced delivery through catheters transduction in humans seems to be mainly due to rapid stereotactically implanted into the tumor and transcuta- inactivation of free retroviral vectors by complement and neously linked to an external pump.21 Vector infusion was due to lack of dissemination of RVPCs away from the accompanied by transient worsening of neurological injection sites. On the other hand, the positive aspect of symptoms, fever, and increase of transaminase levels. retroviral vectors is lack of significant side effects, with Expression of the HSV-TK gene could be monitored by doses up to 1 Â 109 RVPCs in a total volume of 10 ml. No positron emission tomography (PET) after intravenous replication-competent retroviral vectors have been ever administration of [124I]FIAU (I-124-labeled 20-fluoro-20- demonstrated, although transduction of circulating cells deoxy-1b-D-arabino-furanosyl-5-iodo-uracil), a specific by vectors may occur.6,8,9,14,24,28 The risk of cancer due to substrate for HSV-TK. Specific accumulation of FIAU insertional mutagenesis cannot be excluded in patients was detected at the infusion site in one of the patients, in treated with retroviral vectors. However, the presence of whom the greatest density of recurrent tumor cells had the herpes simplex virus type 1 (HSV-TK) gene in been documented histologically. In this patient, PET transduced cells allows to eliminate, by ganciclovir scans also showed the appearance of signs of local tumor (GCV) treatment, any cell that may have undergone necrosis after treatment with GCV.20 Thus, also the transformation. results of this clinical trial with liposome vectors showed Adenoviral vectors – First-generation E1-deleted adeno- that the therapeutic effect was restricted to a relatively viral vectors have been used more recently in gene therapy small volume around the infusion site;21 notwithstanding, clinical trials for malignant gliomas (Table 1). At variance a convection-enhanced delivery technique was used, with retroviral vectors, which are administered through which should allow delivery of molecules inside brain RVPCs because of their low titer, adenoviral vectors can tissue over long distances.29 be directly injected into the target site at high titer. In patients with gliomas, adenoviral vectors have been Methods for gene delivery injected at doses up to 3 Â 1012 viral particles (vp) into The surgical techniques employed for vector delivery to the tumor mass or into the surgical cavity after tumor brain tumors consist of direct injection of the vectors

Cancer Gene Therapy Table 1 Gene therapy clinical trials for malignant gliomas

Reference Pts Diagnosis (No. pts) Study design Gene(s) Vector Vector dose Protocol GT-related adverse Tumor response events

Izquierdo et al., 5 Recurrent GBM (5) Pilot study HSVTK/GCV Retrovirus 2 Â 108–7 Â 108 Intratumor injection of None 1 PR, 1 SD, 3 PD 19966 RVPC RVPC. i.v. GCV from day 7 to 21 Ram et al., 12 Small recurrent GBM Phase I/II HSVTK/GCV Retrovirus 2.5–10 Â 108 RVPC Intratumor injection of Intratumor 2 CR, 3 PR 19977 (9), AA (3) RVPC. i.v. GCV from haemorrhage (17%), day 7 to 21. Second seizure (17%) treatment in four patients Klatzmann 12 Recurrent GBM (12) Phase I/II HSVTK/GCV Retrovirus B9 Â 106/cm2 RVPC injection into the None 4 SD et al., 19988 surgical cavity margins after tumor debulking. i.v. GCV from day 7 to 21 Shand et al., 48 Recurrent GBM (48) Phase I/II HSVTK/GCV Retrovirus B1 Â 109 RVPC RVPC injection into the Seizures (1), 1-year survival rate 19999 surgical cavity margins hydrocephalus (1) 27%; seven after tumor debulking. patients free of i.v. GCV from day 14 recurrence for at to 27 least 6 months Trask et al., 13 Recurrent GBM (9), Phase I, dose HSVTK/GCV Adenovirus 2 Â 109–2 Â 1012 vp Intratumor vector CNS toxicity at doses 3SD 200010 GS (1), AA (3) escalation injection. i.v. GCV after of 2 Â 1012 vp 24 h for 14 days 6 Harsh et al., 5 Recurrent GBM (5) Phase I HSVTK/GCV Retrovirus 3 Â 10 RVPC Intratumor RVPC Brain abscess (1) 5 PD gliomas malignant Barzon for L therapy gene and Cell 200011 injection followed, after 5 days, by tumor

resection and RVPC al et injection into the surgical cavity Packer et al., 12 Recurrent pediatric Phase I, dose HSVTK/GCV Retrovirus RVPC injection into the Seizures and sign of NA 200012 malignant gliomas escalation surgical cavity margins elevated intracerebral (7), ependymomas after tumor debulking. pressure (4) (2), neuroectodermal i.v. GCV from day 14 tumors (3) to 27 Sandmair 14 Newly diagnosed or Phase I: comparison of HSVTK/GCV Retrovirus 1 Â 109 RVPCs; Tumor debulking Seizure and fever (two All PD in retrovirus et al., 200013 recurrent GBM (13), patients treated with (eight tumors); 3 Â 1010 pfu followed by injection in adenoviral vector group; 3 SD in AA (1) RVPCs (7) or adenovirus adenovirus into the surgical cavity group) adenovirus group; adenovirus (7) carrying (seven tumors) margins of RVPCs or significant longer HSV-TK. Seven control adenoviral vectors. survival for patients transduced i.v. GCV for 14 days adenovirus than acrGn Therapy Gene Cancer with LacZ marker gene retrovirus group with respect to the control group Rainov et al., 124 Newly diagnosed, Phase III, randomized: HSVTK/GCV Retrovirus B1 Â 109 RVPC RVPC injection into the Risk of cerebral No difference in 200014 previously untreated standard therapy surgical cavity margins hematoma (8 vs 1) and progression-free GBM (248) (surgery+radiotherapy) after tumor debulking. thromboembolic events survival and overall 541 542 acrGn Therapy Gene Cancer Table 1 (Continued)

Reference Pts Diagnosis (No. pts) Study design Gene(s) Vector Vector dose Protocol GT-related adverse Tumor response events

(N ¼ 124) vs standard i.v. GCV from day 14 (16 vs 13) higher in the survival with therapy+GT (N ¼ 124) to 27 GT group; control group complications due to central venous catheter for GCV infusion (3);

GCV-induced gliomas malignant for therapy gene and Cell neutropenia (1) Floeth et al., 27 Progressive single- Patients enrolled in HSVTK/GCV Retrovirus B1 Â 109 RVPC RVPC injection into the None NA 200115 lesion GBM (24), AA other phase II or III surgical cavity margins (1), GS (2) studies after tumor debulking. i.v. GCV from day 14 to 27 Smitt et al., 14 Recurrent GBM (11), Phase I, dose HSVTK/GCV Adenovirus 4.6 Â 108–4.6 Â 1011 Vector injection into the Cerebral hemorrhage No response in five 200316 AG (1), AO (2) escalation vp surgical cavity margins (1) evaluable patients after tumor debulking. Barzon L i.v. GCV after 48 h for 14 days Prados et al., 30 Recurrent GBM (30) Phase I/II HSVTK/GCV Retrovirus B1 Â 109 RVPC per RVPC injection into the Reservoir NA tal et 200317 cycle surgical cavity margins complications (16%); after tumor debulking acute ventricular/ and 7 days after meningeal reaction surgery through a (13%); pulmonary reservoir. i.v. GCV embolism (7%); from day 14 to 28. Up pancytopenia (3%) to four cycles of GT at 28-day intervals Germano et al., 11 Recurrent GBM (11) Phase I, dose HSVTK/GCV Adenovirus 2.5 Â 1011 –9 Â 1011 Vector injection into the None NA 200318 escalation vp surgical cavity margins after tumor debulking. i.v. GCV after 24 h for 7 days Lang et al., 15 Recurrent GBM (13), Phase I, dose TP53 Adenovirus 3 Â 1010 –3 Â 1012 Intratumor Ad-p53 Headache (53%), NA 200319 AA (1), AG (1) escalation vp injection via implanted fatigue (40%), fever catheter, followed, (27%), neurologic after 3 days, by en bloc events (27%) resection and Ad-p53 injection into the surgical cavity Jacobs et al., 8 Recurrent GBM (8) Phase I/II HSVTK/GCV Liposome NA Intratumor convection- Transient increase of 2PR 200120 and enhanced infusion liver enzymes (50%), Voges et al., of liposome–gene fever and CNS 200321 complex. i.v. GCV symptoms (37.5%) from day 4 to 18 Immonen et al., 17 Operable primary or Randomized study: HSVTK/GCV Adenovirus 3 Â 1010 pfu Vector injection into the Transient increase of Longer survival in 200422 recurrent GBM (13) surgery+GT (N ¼ 17) surgical cavity margins liver enzymes (18%) GT group than in Cell and gene therapy for malignant gliomas L Barzon et al 543 (either viral vectors, nonviral vectors, or VPCs) into the tumor mass either by stereotactic procedures or by open surgery. Each of these techniques has advantages and disadvantages: advantages of the stereotactic technique are reduced surgical trauma, accurate treatment planning, Tumor response control 2 PR, 4 MR, 42PD SD, and availability of precise spatial information that allows GCV: ganciclovir; an accurate measurement of the range of therapeutic effect.30 Its principal shortcomings are the lack of a debulking effect with no immediate improvement on mass e; pfu: plaque-forming unit; effect. Injection of vectors or VPCs after surgical removal of the tumor mass allows an accurate haemostasis control under microscopic vision and reduces the risk of injection into the ventricles and CSF spaces. The number of events and intracerebral edema during GCV (11%) Mild increase of transaminase levels (4), neurological symptoms (4), leukocytosis (2), leukopenia (1) injection sites and the volume of injected material can be increased, but, using visual control alone, planning of injections is forcibly less accurate, even when targeted to evident residual tumor.

Gene therapy strategies Tumor suppressor gene replacement. Based on the after tumor debulking. i.v. GCV from day 5 to 19 Injection of RVPC into the tumor (8) or into the surgical cavity margins after tumor debulking (4). i.v. GCV from day 7 to 21. Two cycles of treatment in four patients observation that inactivation of the TP53 tumor suppressor gene represents a critical event in the pathogenesis of malignant gliomas and that transfer of TP53 in animal 9 models inhibited growth of gliomas and restored their –10

8 sensitivity to radiotherapy and chemotherapy, a phase I 10 clinical trial of TP53 adenoviral vector-mediated gene Â therapy was undertaken in patients with recurrent RVPC malignant gliomas19 (Table 1). In this study, patients underwent intratumor stereotactic injection of the Ad-p53 adenoviral vector via an implanted catheter, followed by en bloc resection for analysis and treatment of the post- resection cavity. Of 15 patients enrolled, 12 underwent

Retrovirus 1.5 both treatment stages and could be evaluated for clinical toxicity and transduction efficiency. Toxicity of gene therapy was minimal and a maximum tolerated dose was not reached. Moreover, no systemic viral dissemination was observed. Analysis of tumor specimens showed expression of p53 protein within the nuclei of glioma HSVTK/ GCV+IL-2 cells, but transduced cells were only found within a short distance of the injection site. This is an important limitation, since the requirement for the success of tumor suppressor gene therapy is efficient transduction of the whole population of tumor cells. Due to the design of the study, tumor response could not be assessed. 19). RT for surgery alone ¼ N vs ( primary tumors Prodrug/suicide gene therapy. Most gene therapy clinical trials for malignant gliomas have been based on the delivery of the suicide gene HSV-TK, followed by GCV GBM

vs administration (Table 1) (for a detailed review on HSV- TK-based clinical trials on malignant gliomas, see Pulkkanen and Yla-Herttuala5). The HSV-TK enzyme phosphorylates the prodrug GCV into the active com- and AA (4) (18) and AA (1) pound GCV triphosphate, an inhibitor of DNA synthesis that leads to cell death. A useful feature of HSV-TK/GCV

12 Recurrent GBM (12) Phase I/II gene therapy is the ‘bystander effect’, that is, killing of nontransduced neighbor cells by GCV treatment. This

(Continued) effect is mediated by transfer of toxic metabolites and by et al.,

and stimulation of an immune response involving tumor et al., 23 24 31

` infiltrating T cells and natural killer (NK) cells. AA: anaplastic astrocytoma; AG: anaplastic mixed glioma; AO: anaplastic oligodendroglioma; CR: complete response; GBM: gliobastoma multiforme; GS: gliosarcoma; GT: gene therapy;PR: HSV-TK: thymidine partial kinase response; of herpes RVPC: simplex retroviral virus vector-producing type-1; i.v.: cells; intravenous; SD: MR: stable minor response; disease; NA: vp: not availabl viral particles. Colombo 2005 Table 1 Reference Pts Diagnosis (No. pts)Palu Study design1999 Gene(s) Vector Vector dose ProtocolNotwithstanding GT-related adverse the presence of the bystander effect,

Cancer Gene Therapy Cell and gene therapy for malignant gliomas L Barzon et al 544 this gene therapy approach gave poor results not only in superior to that for patients treated with RVPCs. Based patients with malignant gliomas but also in patients with on the results of these studies, gene therapy clinical other types of cancer.4 With a few anecdotal exceptions, protocols based on RVPC injection have been generally lack of efficacy was particularly observed in clinical trials replaced by the use of adenoviral vectors. based on the use of RVPCs as gene transfer tools.5 In a total of 79 evaluable patients with recurrent GBM Suicide/cytokine gene therapy. In order to amplify the enrolled in nine phase I/II studies of retroviral vector- antitumor immune response elicited by HSV-TK/GCV, mediated HSV-TK gene therapy, complete or partial combined delivery of the suicide gene together with the responses were observed in 13% of cases, and minor human IL-2 gene by means of a retroviral vector has been responses or stable disease in 18% (Table 1). In most attempted in patients with recurrent GBM.23,24 Treatment studies, treatment consisted in injection of RVPCs into was well tolerated and adverse events were generally mild the surgical cavity after tumor debulking and not in direct and mainly related to GCV administration, such as intratumor injection of the RVPCs. Therefore, response transient elevation of liver function tests. As observed in to treatment was difficult to assess. In the studies of HSV- HSV-TK/GCV clinical studies, also in this trial two cases TK gene therapy based on intratumor injection of 5,7 of fatal pulmonary embolism were reported, which could RVPCs, tumor regression was observed in a relatively conceivably be accounted to the increased thromboem- high percentage of patients, especially in those with small bolic risk of glioma patients, even though the contribution tumor recurrences. Unfortunately, tumor responses were of the systemic activation of cytokine cascades after gene limited to the site of RVPC injection, whereas untreated 7 therapy could not be excluded. Regression of tumor tumor masses progressed rapidly. masses was observed in 50% of cases, including the case Lack of efficacy of retroviral vector-mediated HSV- of a patient who showed complete regression of a distant TK/GCV gene therapy after tumor debulking was 4 ml untreated brain lesion, besides reduction of the confirmed by a phase III clinical trial, which showed no tumor mass injected with RVPCs. In this patient, tumor significant benefit of gene therapy over radiotherapy in 14 response was accompanied by a marked increase of newly diagnosed patients with GBM. In this study, circulating Th1 cytokine levels, especially interferon patients underwent gross surgical resection of the tumor (IFN)g and TNFa, immediately after RVPC injection, and radiotherapy. In the gene therapy arm, 124 patients and a persistent cytokine hypersecretion during follow-up were treated by RVPC injection into the wall of the observation.24 A similar profile of plasma cytokine levels resection cavity, followed by intravenous GCV infusion, was observed also in the other patients showing tumor whereas the 124 patients of the control arm were treated response to treatment. Moreover, analysis of tumor only by surgery and radiotherapy. Besides the lack of biopsies obtained after injection of RVPCs demonstrated efficacy of gene therapy, this study showed a higher the presence of tumor necrosis around the injection site rate of side effects, including cranial hematomas and and an intense cellular infiltration, represented mainly thromboembolic events, in the gene therapy arm than 14 by T-helper/inducer lymphocytes and activated cytotoxic in controls. Of note, this trial allowed the investigation T cells and macrophages.23 of the immunological component of HSV-TK/GCV gene therapy.32 Serum samples from treated patients showed in fact increased FasL and interleukin (IL)-12 levels,14 consistent with activation of a Th-1 response. A similar Oncolytic virus therapy response was observed in an adolescent patient with ependymoma after intracranial implantation of HSV-TK Limitations of replication-defective viral vectors, regard- RVPCs.33 Moreover, peripheral blood mononuclear cells ing inadequate delivery, insufficient expression of the cocultivated with autologous tumor cells or with RVPCs therapeutic gene, and short-term efficacy of treatment, demonstrated a transient (days 7–28 after RVPC injec- have been faced by the development of replication- tion) but specific-T1-type immune response.32 Develop- competent viruses, with the ability to selectively replicate ment of an immune response against retroviral vector and kill cancer cells (oncolytic viruses). Once the infected proteins and RVPCs has also been demonstrated.7,9 cell undergoes lysis, thousands of new vp’s are released Adenoviral vectors have been more recently investi- and new cells are infected and killed in successive rounds gated for gene therapy of malignant gliomas.13,16,18,19,22 of infection and cytolysis. Tumor selectivity has been A randomized, controlled study of adenovirus-mediated achieved by using attenuated viruses that preferentially HSV-TK gene therapy in patients with operable primary replicate in tumor cells, by inactivating viral genes that are or recurrent malignant gliomas demonstrated a significant necessary for replication in normal cells but dispensable in longer survival time in patients treated with standard tumor cells in which the missing function is supplied, or therapy plus HSV-TK/GCV gene therapy than to control by engineering viruses in which replication-essential genes patients receiving only standard treatment.22 A previous are placed under tumor-specific promoters.34 Most report by the same authors,13 which compared the safety oncolytic viruses have been engineered from adenovirus and efficacy of adenovirus- vs RVPC-mediated HSV-TK and HSV-1, although inherently tumor-selective viruses, gene therapy for patients with malignant glioma, showed such as Newcastle disease virus (NDV), reovirus, and that both treatments were well tolerated, but that the autonomous parvoviruses, are also being tested in clinical mean survival for the adenoviral vector-treated group was trials. Some of these vectors have been experimented in

Cancer Gene Therapy Cell and gene therapy for malignant gliomas L Barzon et al 545 pilot and phase I clinical studies in patients with and detected by PCR in eight other cases. Tumor tissue malignant gliomas (Table 2). from one of the patients inoculated with HSV1716 was cultured in vitro and tested for the presence of the virus, Oncolytic HSV-1 with negative results. When cells were reinfected in vitro, Replication-competent oncolytic HSV-1 used in clinical a small fraction was found to not undergo lysis and to protocols for gliomas include G207 and HSV1716. The continue to shed the virus.46 Thus, even though glioblas- double-mutant G207 herpes simplex virus harbors dele- toma cells are generally fully permissive for HSV1716 lytic tions of both copies of the neurovirulence gene g34.5 and replication, as demonstrated in primary GBM cell 38,47 contains an insertional inactivation in the UL39 locus, cultures, some could support persistent viral infection which encodes ICP6, a subunit of viral ribonucleotide and shed vp’s to the surrounding tumor cells.48 The reductase essential for viral replication in postmitotic HSV1716 virus was also used as adjuvant treatment in cells.35 The HSV1716 virus is similar to G207, since it patients with newly diagnosed or recurrent high-grade lacks both copies of g34.5, but it retains a wild-type ICP6 gliomas, by injection into the brain surrounding the gene.36 Besides its essential role in HSV-1 pathogenicity in surgical cavity, after tumor debulking38 (Table 2). As neurons, the g34.5 protein product ICP34.5 is involved in clinically indicated, patients proceeded to further radio- overcoming host cell defenses against infection mediated therapy or chemotherapy. Of note, also in this case no by protein kinase R (PKR). Upon infection, PKR toxicity related to HSV1716 administration was observed, activation shuts down translation in the infected cell by thus confirming the safety profile of oncolytic HSV-1. phosphorylating and inactivating eukaryotic initiation Moreover, three of the 12 treated patients had a long factor-2a (eIF-2a). ICP34.5 binds and recruits protein disease-free survival, including a patient who showed phosphatase 1-a to dephosphorylate eIF-2a, so that tumor size reduction at follow-up imaging evaluation.38 protein synthesis may proceed. g34.5-null viruses replicate in cells with defective PKR pathways, such as cancer cells Oncolytic adenoviruses with ras overexpression, whereas they cannot efficiently Adenoviruses carrying mutations of the early genes E1A replicate in cells with a functional PKR pathway.42 or E1B, which are responsible for binding and inactivat- Moreover, deletions of the g34.5 gene also remove ing several proteins involved in cell cycle control and sequences encoding latency-activated transcripts and thus apoptosis, such as pRB family members and p53, impair the ability of these viruses to establish latency preferentially replicate and lyse cancer cells. ONYX-015 within infected cells.43 Inactivation of ICP6 provides (dl1520), one of the first oncolytic adenoviruses to be further safety properties, that is, ability to replicate only designed, has been experimented against a variety of in dividing cells44 and increased sensitivity to acyclovir human cancers in phase I/II clinical trials, which and GCV.45 demonstrated its safety and anticancer activity, especially The G207 virus was experimented in a phase I dose- in combination with chemo- and radiotherapy.4,34 escalation study in patients with recurrent malignant ONYX-015 virus preferentially replicates in p53-deficient gliomas35 (Table 2). All patients were treated by stereo- cells because it carries a deletion of the sequence encoding tactic intratumor injection of viruses at doses up to the p53-inactivating protein E1B-55K. E1B-55K is 3 Â 109 pfu. No patient developed HSV encephalitis nor dispensable in tumor cells that are deficient for p53 any other significant toxicity, besides mild neurological function, where viral replication can proceed unim- deterioration, and a maximally tolerated dose could not paired.34 A phase I dose-escalation trial of ONYX-015- be established. Moreover, 8 of 20 patients had reduced based virotherapy was performed in 24 patients with enhancement volumes of their tumors at MRI performed recurrent malignant gliomas39 (Table 2). The virus was 1 month after surgery as compared to the immediate injected at 107–1010 pfu in a total of 10 sites within the postinoculation scans. Out of these patients, one showed surgical cavity after tumor resection. None of the patients almost complete tumor regression and three others had experienced serious adverse events related to ONYX-015, relatively long survival times.35 and the maximum tolerated dose was not reached. Safety and toxicity of the HSV1716 virus were first Unfortunately, safety of treatment was not accompanied evaluated in a phase I dose-escalation study in nine by similar positive results regarding efficacy, since all patients with recurrent high-grade glioma, who did not patients progressed, with the exception of a patient with show any sign of encephalitis, even at the highest dose anaplastic astrocytoma who had stable disease. The of 105 pfu.36 No virus shedding was demonstrated in median time to tumor progression was 46 days and the biological samples, nor evidence of viral genomes in median survival time 6.2 months. Interestingly, in two tumor biopsies and post-mortem material obtained some patients who underwent a second resection 3 months after months after treatment. As for tumor response to ONYX-015 injection, a lymphocytic and plasmacytoid virotherapy, five patients had stable disease, whereas the cell infiltrate was demonstrated at the site of injection.39 others progressed. Efficacy of HSV1716 was evaluated in a subsequent study in 12 patients with high-grade glioma Oncolytic NDV (Table 2), who, 4–9 days after intratumor injection of the NDV is an avian paramyxovirus, which is nonpathogenic virus, underwent surgical removal of the tumors to assess to humans, except for rare transient respiratory infection viral replication.37 Infectious HSV1716 was recovered among poultry farmers. Some attenuated NDV strains from the resected tumors at the injection site in two cases have marked oncolytic properties against human tumors,

Cancer Gene Therapy 546 acrGn Therapy Gene Cancer

Table 2 Clinical trials of oncolytic virus therapy for malignant gliomas eladgn hrp o ainn gliomas malignant for therapy gene and Cell Reference Pts Diagnosis (No. Study design Genetic Virus Dose Protocol Virotherapy-related Tumor response pts) modification adverse events

Markert et al., 21 Recurrent GBM Phase I, dose g34.5, RR-null G207 HSV 106–3 Â 109 pfu Stereotactic intratumor Mild neurological Eight cases of 200035 (15), AA (4) escalation injection of the virus symptoms following reduced virus injection enhancement volumes of tumors at MRI Rampling et al., 9 Recurrent GBM Phase I, dose g34.5-null HSV1716 103–105 pfu Stereotactic intratumor Mild neurological 5 SD, 4 PD

200036 (8), AA (1) escalation injection of the virus symptoms following Barzon L virus injection in one Papanastassiou 12 Recurrent GBM Phase I g34.5-null HSV1716 105 pfu Stereotactic intratumor None 1 CR, 3 PR

et al., 200237 (10), AA (1), injection of the virus. al et newly diagnosed Surgical removal of the GBM (1) tumors 4–9 days after inoculation Harrow et al., 12 Newly Phase I g34.5-null HSV1716 105 pfu Virus injection into the None 1 PR with long 200438 diagnosed GBM surgical cavity after progression free (6), recurrent tumor debulking survival, 1 SD, GBM (4), AA (1), 10 PD AO (1) Chiocca et al., 24 Recurrent GBM Phase I, dose E1B-del ONYX-015 107–1010 pfu Virus injection into the None 1 SD, 23 PD 200439 (17), AA (6), AO escalation surgical cavity after (1) tumor resection Csatary et al., 14 Recurrent GBM Pilot study Attenuated Newcastle Up to 107 pfu i.v. injection, daily or None 2 CR, 1 SD 199940 and (3), AA (1), MTH-68/H strain disease virus daily, for up to 7 weekly, continuously 200441 (including years pediatric pts) AA: anaplastic astrocytoma; AO: anaplastic oligodendroglioma; CR: complete response; GBM: gliobastoma multiforme; i.v.: intravenous; pfu: plaque-forming unit; PD: progressive disease; PR: partial response; SD: stable disease; vp: viral particles. Cell and gene therapy for malignant gliomas L Barzon et al 547 while sparing non-neoplastic cells.49 The mechanism of immunotherapy and active immunotherapy (vaccination) the oncolytic action of attenuated NDV is still unknown, (Table 3). Adoptive immunotherapy for gliomas has even though activation of Ras signaling pathway in generally involved transfer of nonspecific lymphokine- cancer cells has been hypothesized to play a role, as has activated killer (LAK) cells. Active immunotherapy been suggested for g34.5-defective HSV-1. NDV also approaches have been based on ex vivo or in vivo exhibits pleiotropic immune-modulatory properties50 and sensitization of patient’s APCs or naive T cells to induces production of double-stranded RNA, IFNs, and tumor-associated antigens, used as vaccines. If successful, chemokines in infected cells.51 The idea to use live NDV active vaccination results not only in tumor eradication, to treat cancer dates back to the 60s, when anecdotal cases but also in establishment of long-term immunity against of tumor regression after exposure to viral infection were tumor cells.73 observed.52,53 The oncolytic NDV PV701 was utilized in a phase I trial including 79 patients with various advanced Adoptive immunotherapy solid cancers.54 The results of this study established that Several approaches of adoptive immunotherapy have optimal treatment to avoid side effects should start with been attempted (Table 3). Initial studies, based on low doses, followed by escalation of doses. The most intratumor or intrathecal administration of autologous common adverse events were fever and flu-like symptoms, immune cells, either alone74,75 or combined with IFN but also serious adverse events were reported, including administration,76 gave poor response. More promising a treatment-related death. Another oncolytic NDV, results were achieved from the use of LAK cells and IL-2 MTH-68/H, has been investigated in phase I/II studies injected into the surgical cavity after tumor resection, but in patients with advanced cancer53,55,56 and in 14 patients significant toxicity was also observed, with symptoms of with high-grade gliomas41 (Table 2). Treatment consisted aseptic meningitis and increased intracranial pressure.77–79 of intravenous injection of daily doses of about A complete tumor response and two partial responses, out 2 Â 107 pfu of MTH-68/H. The doses were gradually of six treated patients, were observed after intratumor increased up to 2 Â 108 pfu daily and, in the presence of infusion of autologous tumor-infiltrating lymphocytes, clinical benefit, maintained during the patient’s lifetime expanded ex vivo in the presence of IL-2, with concurrent with less frequent administrations. No adverse events IL-2 infusion.80 Tumor-infiltrating lymphocytes, how- were reported in this study throughout the course of ever, demonstrated poor specificity and scarce antitumor treatment. Of 14 patients, four patients with GBM had activity, conceivably because of the immunosuppressed survival times between 5 and 9 years, including two environment typical of malignant gliomas. An alternative patients who demonstrated an almost complete tumor approach to the use of autologous T-cell populations for shrinkage.41 These results look promising, but controlled adoptive immunotherapy was the transfer of allogeneic studies are needed to definitely define the safety and T cells.81 This strategy was based on the rationale that oncolytic activity of NDV. MHC-mismatched CTLs would produce a more potent tumoricidal response against glioma cells that express MHC. This approach, however, did not confirm the Cell-mediated immunotherapy encouraging preclinical results when assessed in a pilot clinical study.82 The highly disseminated nature of malignant gliomas represents the major obstacle for effective eradication of Active immunotherapy all residual intracranial tumor reservoirs. Treatment Activation of an efficient antigen-specific T-cell response approaches aimed at enhancing the patient’s antitumor requires presentation of tumor antigens by professional immune response, which have already demonstrated APCs, which provide appropriate costimulatory signals. promising results in prelinical investigations and in pilot Activated T cells undergo clonal expansion with genera- studies in humans,57 could represent an effective adjuvant tion of CTLs, with cytotoxic activity against tumor cells to conventional therapy to eradicate brain tumors. On the and memory cells that confer long-lasting antitumor other hand, immunotherapy for malignant gliomas is a immunity. An important obstacle to the design of an hard task, since these neoplasms are characterized by efficient vaccination strategy against human malignant potent immunosuppressive properties, probably mediated gliomas is represented by the fact that immunologically by chemokines produced by the tumor, such as trans- relevant glioma-specific antigens have not yet been forming growth factor beta and IL-10, that depress host identified. antitumor immune response.57 In fact, patients with malignant gliomas typically show lymphopenia, depressed Autologous tumor cell vaccination strategies. Initial T-cell responsiveness and antibody production and vaccination strategies for glioma consisted in subcuta- impaired antigen-presenting cell (APC) function. More- neous inoculation of irradiated autologous tumor cells, over, glioma cells downregulate expression of major even engineered to produce immunostimulatory cyto- histocompatibility complex molecules, while overexpres- kines, such as IL-2 and GM-CSF. The first patient to be sing Fas ligand on their surface, thus inhibiting antitumor vaccinated for a glioblastoma received subcutaneous cytotoxic T lymphocyte (CTL) activity. injections of a mixture of irradiated autologous tumor Cell-mediated immunotherapy strategies that have been cells with fibroblasts genetically modified to secrete IL-2 designed for malignant brain tumors include adoptive by retroviral gene transfer.83 Treatment resulted in

Cancer Gene Therapy 548 acrGn Therapy Gene Cancer Table 3 Clinical trials of active immunotherapy for malignant gliomas Reference Pts Study design Diagnosis Protocol Adverse events Outcome (no. pts)

Holladay et al., 199658 15 Phase I Recurrent grade Vaccination with autologous irradiated tumor None 7 PR III/i.v. cells and BCG. Selective expansion of CD4+ astrocytoma (15) and CD8+ T lymphocytes from PBMCs taken 14 days after vaccination and cultured with irradiated autologous tumor cells and IL-2. i.v.

transfer of stimulated PBMCs7IL-2 gliomas malignant for therapy gene and Cell Plautz et al., 199859 10 Phase I Progressive Vaccination with irradiated autologous tumor None 2 PR, 1 SD primary or cells and GM-CSF. i.v. transfer of lymphocytes recurrent obtained from draining lymph nodes, surgically malignant resected 7 days after vaccination, and gliomas stimulated with staphylococcal enterotoxin A, anti-CD3, and IL-2 Plautz et al., 200060 12 Phase I Newly Vaccination with irradiated autologous tumor Mild transient fever, nausea, 4 PR, 2 SD diagnosed grade cells and GM-CSF. i.v. transfer of lymphocytes myalgias. Erythema at

II astrocytoma obtained from draining lymph nodes, surgically injection site Barzon L (2), AA (4), resected 8–10 days after vaccination, and GBM (6) stimulated with staphylococcal enterotoxin A

and IL-2 al et Wood et al., 200061 9 Pilot study Recurrent grade Vaccination by subcutaneous inoculation of Transient fever and nausea. 3 PR, 6 PD III/i.v. autologous irradiated tumor cells and BCG. Erythema or necrotic ulcers at astrocytoma (9) Activation and expansion of T lymphocytes from vaccine injection site PBMCs taken 14 days after vaccination by treatment with anti-CD3 and IL-2. i.v. transfer of stimulated PBMCs Andrews et al., 200162 12 Pilot study Recurrent GBM Vaccination with autologous glioma cells treated Deep venous thrombosis 2 CR, 4 PR (8), AA (4) ex vivo with IGF-IR antisense oligodeoxynucleotide, encapsulated in diffusion chambers, reimplanted in the rectus sheath within 24 h of craniotomy, and retrieved after 24-h in situ incubation Schneider et al., 200163 11 Phase I GBM (11) Vaccination by subcutaneous inoculation of Erythema at vaccine injection No improvement of autologous tumor cells infected by NDV and site survival in comparison inactivated by cisplatinum+IL-2. One application with controls receiving of inactivated noninfected tumor cells, followed radiotherapy and by four applications of NDV-infected tumor cells chemotherapy in 2-week intervals and the fifth after 3 months. Two weeks after the fourth vaccination, patients received four control injections containing carrier medium only, NDV only, IL-2 only and inactivated noninfected tumor cells Steiner et al., 200464 23 Pilot study Newly Vaccination by subcutaneous inoculation of Mild fatigues and palpable 1 CR; significantly diagnosed (15) autologous tumor cells infected by NDV and indurations at the vaccination longer progression-free and recurrent irradiated+IL-2. Up to eight applications: site. Mild fever in one; and overall survival GBM (8) vaccinations 1 to 4 were given in 3-week transient elevation of than nonrandomized intervals, followed by 4-week intervals for transaminases in one controls vaccinations 5 to 8 days Table 3 (Continued)

Reference Pts Study design Diagnosis Protocol Adverse events Outcome (no. pts)

Yu et al., 200165 9 Phase I Newly Intradermal vaccination with DCs pulsed with Fever, nausea, vomiting in Longer survival time diagnosed GBM peptides eluted from autologous glioma cells. one; lymphadenopathy in one than in historical (7), AA (2) Three vaccinations at intervals of 14 days controls Weeler et al., 200466 25 Phase I/II Newly Intradermal vaccination with DCs pulsed with NA 3 PR in diagnosed GBM tumor peptides or tumor lysate. Three vaccine+chemotherapy (25) vaccinations at intervals of 14 days. patients. Longer overall Vaccine+chemotherapy in 13, only vaccine survival in in 12 vaccine+chemotherapy patients than in patients receiving vaccine or chemotherapy alone Yamanaka et al., 10 Phase I/II GBM (7), AA (3) Vaccination with DCs pulsed with autologous Mild headache in one, mild 2 MR, 4 SD 200367 tumor lysate injected intradermally close to a erythema at the injection site in cervical lymph node and/or intratumorally via three a reservoir. Up to 10 vaccinations at 3-week intervals Yu et al., 200468 14 Phase I Recurrent GBM Vaccination with DCs pulsed with autologous Transient headache in three; Longer survival time (9), AA (3); tumor lysate injected s.c. three vaccinations at fatigue in two; seizures in two; than in historical newly diagnosed 2-week intervals local erythema in one controls GBM (1), AA (1) Rutkowski et al., 200469 12 Phase I Recurrent GBM Intradermal vaccination with DCs pulsed with Severe neurological symptoms 1 PR, 1 SD

(12) autologous tumor lysate. Vaccinations at due to peritumoral edema in gliomas malignant Barzon for L therapy gene and Cell intervals of 4 weeks one; mild hematotoxicity in 1, nocturnal sweating (1) and

meningeal irritation (1) after al et vaccination Yamanaka et al., 24 Phase I/II Recurrent GBM DCs, pulsed with autologous tumor lysate and Mild headache in one; mild 1 PR, 3 MR, 10 SD, 10 200570 (18) and AA (6) KLH protein, were injected intradermally, or both erythema and swelling at the PD. Longer survival of intratumorally and intradermally every 3 weeks. injection site in seven patients whose DCs Up to 10 vaccinations. In the phase II protocol, were matured with OK- DCs were cultured with tumor lysate followed by 432 than those without overnight incubation with penicillin-killed OK-432 maturation Streptococcus pyogenes (OK-432) Kikuchi et al., 200171 8 Phase I Recurrent GBM Intradermal injection of DCs and glioma cells Mild erythema at the injection 2PR (5), AA (2), fused by using polyethylene glycol. Up to nine site in one AO (1) administrations Kikuchi et al., 200472 15 Phase I Recurrent GBM DCs and glioma cells fused by using Transient mild fever in four; 4PR,1MR,2SD (6), AA (7), polyethylene glycol. Intradermal injection of general convulsion in one; acrGn Therapy Gene Cancer AO (2) vaccine on day 1, followed by subcutaneous erythema and induration at the injection at the same site of recombinant human injection site in 13; transient IL-12 on days 3 and 7. Three vaccinations at liver dysfunction in six, 2-week intervals leukocytopenia in seven BCG: Bacillus of Calmette and Guerin; IL-2: interleukin-2; AA: anaplastic astrocytoma; AG: anaplastic mixed glioma; AO: anaplastic oligodendroglioma; CR: complete response; DC: dedritic cell; GBM: gliobastoma multiforme; GS: gliosarcoma; i.v.: intravenous; MR: minor response; NA: not available; NDV: Newcastle disease virus; PBMC: peripheral blood mononuclear cell; PR: partial response; SD: stable disease. 549 Cell and gene therapy for malignant gliomas L Barzon et al 550 marked tumor necrosis and increased CTL-mediated Dendritic cell vaccinations. Taken together, the vacci- antitumor immunity. Several vaccination studies with nation protocols based on autologous tumor cells have irradiated autologous tumor cells were performed there- given promising results, notwithstanding the well-known after in patients with recurrent gliomas following surgical poor antigen-presenting capacity of glioma cells. To resection of the tumor mass58,61 (Table 3). In these induce a more efficient and specific immune response protocols, immunogenicity of the vaccine was enhanced against malignant gliomas, professional APCs, such as by mixing tumor cells with bacillus Calmette-Gue´ rin and dendritic cells (DCs), have been used.87 DCs, character- by combining vaccination with intravenous IL-2 admin- ized by high expression of costimulatory molecules, are istration and adoptive transfer of stimulated autologous present in peripheral tissues, where they efficiently T cells. Treatment was tolerated with limited toxicity and capture antigens. As DCs move towards the draining some patients demonstrated partial response. Moreover, secondary lymphoid organs, they process and present development of delayed-type hypersensitivity occurred in antigens to naive T cells, thereby inducing a cellular all patients.58,61 immune response that involves both CD4 þ T-helper cells A similar vaccination strategy was pursued by Plautz and CD8 þ CTLs. DCs can also activate NK cells and et al.,59,60 who added GM-CSF to the vaccine prepara- induce humoral immunity by activation of naive and tion, instead of BCG, and used as immune effector cells memory B cells.88 In vaccination studies against gliomas lymphocytes obtained from lymph nodes draining the (Table 3), DCs are obtained from peripheral blood stem vaccination site and cultured with staphylococcal enter- cells of patients and pulsed either with peptides eluted otoxin A, anti-CD3, and low doses of IL-2. Also, this from the surface of cultured autologous brain tumor vaccination protocol was safe and, in some cases, led to cells65,66 or with autologous tumor lysates.67,69,70,89 DCs partial tumor regression.59,60 Preliminary data of safety have been also fused with autologous glioma cells to and efficacy have been also reported with an autologous generate vaccines.71,72 glioma cell vaccine admixed with IL-4 gene-transfected Vaccination with DCs pulsed with autologous tumor fibroblasts.84 peptides or autologous tumor lysates has been performed An interesting vaccination strategy employed antisense in a total of 70 patients with malignant glioma (Table 3). oligodeoxynucleotides to inhibit expression of the type I Treatment has been shown to be safe and to enhance insulin growth factor receptor (IGF-IR) in cultured peripheral tumor-reactive CTL activity, and to signifi- glioma cells. The rationale for this approach was based cantly increase the number of tumor-infiltrating lympho- on the observation that inhibition of IGF-IR in experi- cytes.65 Although the number of objective tumor mental glioma models results in massive apoptosis of responses was low, the association of vaccination with tumor cells, marked inhibition of tumorigenesis and chemotherapy seemed to significantly prolong the pro- metastasis, and stimulation of an immune response gression-free and overall survival times as compared with against cancer cells.85,86 The clinical study was performed each treatment alone66 or with historical controls.68 With in 12 patients operated for recurrent GBM or anaplastic respect to the preparation of peptide-pulsed DCs, which is astrocytoma, who were vaccinated by implantation a long and laborious procedure, preparation of DCs into the rectus sheath of irradiated autologous stimulated by autologous tumor lysates is more easy glioma cells encapsulated in diffusion chambers, after and does not require previous identification of tumor- incubation with IGF-IR antisense oligodeoxynucleo- associated antigens. Tumor-lysate pulsed DCs have been tides.62 Treatment was associated with a rather high also safely employed for intradermal and intratumor vac- incidence of deep vein thrombosis, so that anticoagulant cination in patients with recurrent gliomas.67,70 The idea prophylaxis was included in the protocol, but also of injecting DCs intratumorally aims at promoting the with a relatively high rate of clinical and radiological processing of tumor antigens in situ and at initiating improvements. Interestingly, histological analysis of tumor-specific immune response in local lymphoid tumors resected from patients with disease progression organs.90 In one protocol, intratumor administration of revealed necrosis, lymphocytic infiltration, and micro- immature DCs was combined with intradermal vaccina- vessel thrombosis.62 tions with mature DCs, differentiated with the strepto- The immune-modulatory properties of NDV have been coccal preparation OK-43273 (Table 3). It was exploited as adjuvants to increase the immunogenicity of demonstrated the GBM patients receiving matured DCs cultured autologous glioblastoma cells in pilot and phase experienced longer survival than patients without OK-432 I studies63,64 (Table 3). After vaccination with autologous maturation, and that the GBM patients with both tumor cells cultured in vitro and infected with NDV, the intratumoral and intradermal administration had longer patients showed activation of specific CTL and memory overall survival times than the patients with intradermal antitumor immune responses and the presence of CD8 þ administration.70 cytotoxic tumor-infiltrating lymphocytes, but not after DCs have been also fused with cultured autologous administration of inactivated tumor cells alone.63,64 One glioma cells and injected intradermally for the treatment of the patients demonstrated objective tumor remission of patients with malignant glioma.71,72 Also, this vaccina- after vaccination and, altogether, patients had a signifi- tion protocol was not associated with any serious adverse cantly longer progression-free survival and overall survi- effect and induced antitumor immune responses and val than a nonrandomized matched group of GBM partial regression of the tumor mass in some cases.71 patients treated only by surgery and radiotherapy.64 Subcutaneous injection of recombinant human IL-12 at

Cancer Gene Therapy Cell and gene therapy for malignant gliomas L Barzon et al 551 the same site after vaccination was safe and seemed to be new molecular targets for vaccination, gene delivery, or associated with a higher tumor response rate.72 virotherapy. But, even in this horizon, in order to be effective and contribute to the improvement of patient’s outcome, gene therapy and vaccination will have to play an adjuvant role to conventional treatment modalities. Concluding remarks This is what the most recent gene therapy and vaccination clinical trials have taught us. Several pilot and phase I/II clinical studies of gene therapy, virotherapy, and active immunization for malignant gliomas have been undertaken and results are now Acknowledgements available. If the feasibility and safety of the different therapeutic modalities have been sufficiently documented, This work was supported by grants from FIRB no. efficacy has not yet been convincingly proven, despite RBNE0127YS-006 and from IOV (Istituto Oncologico reports of improved survival of patients and objective Veneto) to G Palu` . tumor responses. Indeed, the only phase III randomized multicenter clinical protocol of the evaulation of efficacy of retroviral vector-mediated HSV-TK suicide gene therapy for GBM failed to demonstrate the superiority References of gene therapy over conventional therapy. Thus, well- designed phase II and phase III trials are urgently needed 1 De Angelis LM. Brain tumors. N Engl J Med 2001; 344: to assess the efficacy of the other promising gene therapy 114–123. approaches, such as adenoviral vector-mediated delivery 2 Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, of suicide genes or the delivery of combinations of suicide Taphoorn MJ et al. Radiotherapy plus concomitant and and cytokine genes. adjuvant temozolomide for glioblastoma. N Engl J Med It is becoming more and more evident from clinical 2005; 352: 987–996. 3 Oldfield EH, Ram Z, Culver KW, Blaese RM, DeVroom studies on conventional chemotherapy and radiotherapy HL, Anderson WF. Gene therapy for the treatment of brain that multi-modality treatment strategies have a higher tumors using intra-tumoral transduction with the thymidine chance of success, conceivably because they simulta- kinase gene and intravenous ganciclovir. Hum Gene Ther neously hit different tumor targets. This paradigm is true 1993; 4: 39–69. also for gene therapy. Indeed, a poly-gene therapy 4 Barzon L, Boscaro M, Palu` G. Endocrine aspects of cancer approach, including delivery of suicide, cytokine, and gene therapy. Endocr Rev 2004; 25: 1–44. antiangiogenic genes, should have a higher chance to 5 Pulkkanen KJ, Yla-Herttuala S. Gene therapy for malignant knock down all the molecular alterations that accumulate glioma: current clinical status. Mol Ther 2005; 12: during tumor formation, whereas a combination of 585–598. different immunotherapeutic strategies should prevent 6 Izquierdo M, Martin V, de Felipe P, Izquierdo JM, Perez-Higueras A, Cortes ML et al. Human malignant the risk that tumor cells evade immune clearance. brain tumor response to herpes simplex thymidine kinase Perhaps, the most promising results have been achieved (HSVtk)/ganciclovir gene therapy. Gene Therapy 1996; 3: from pilot studies with oncolytic viruses, but, even in this 491–495. case, a long way still has to be gone before finding the best 7 Ram Z, Culver KW, Oshiro EM, Viola JJ, DeVroom HL, oncolytic virus with optimal safety and therapeutic Otto E et al. Therapy of malignant brain tumors by requirements. Again, reports from clinical trials of viral intratumoral implantation of retroviral vector-producing oncolysis for other types of malignancies teach us that cells. Nat Med 1997; 3: 1354–1361. virotherapy is more effective when combined with other 8 Klatzmann D, Valery CA, Bensimon G, Marro B, Boyer O, therapeutic modalities, such as chemotherapy and radio- Mokhtari K et al. A phase I/II study of herpes simplex virus therapy. type 1 thymidine kinase ‘suicide’ gene therapy for recurrent glioblastoma. Study Group on Gene Therapy for Glioblas- As for active immunotherapy, vaccination with primed toma. Hum Gene Ther 1998; 9: 2595–2604. DCs seems to be superior to conventional vaccination 9 Shand N, Weber F, Mariani L, Bernstein M, Gianella- with inactivated autologous tumor cells, and, besides Borradori A, Long Z et al. A phase 1–2 clinical trial of gene intradermal DC administration, intratumor injection of therapy for recurrent glioblastoma multiforme by tumor immature DCs seems to be also effective. DC vaccination transduction with the herpes simplex thymidine kinase gene protocols for gliomas, however, still need to be optimized, followed by ganciclovir. GLI328 European–Canadian Study regarding, for example, the type of antigen for priming, Group. Hum Gene Ther 1999; 10: 2325–2335. the method for antigen loading, the route of delivery, etc. 10 Trask TW, Trask RP, Aguilar-Cordova E, Shine HD, Wyde Thus, in a scenario that remains gloomy for patients PR, Goodman JC et al. Phase I study of adenoviral delivery with malignant gliomas, notwithstanding advances in of the HSV-tk gene and ganciclovir administration in patients with current malignant brain tumors. Mol Ther radiation therapy, chemotherapy, and surgical techni- 2000; 1: 195–203. ques, hope still relies on the promises of these new 11 Harsh GR, Deisboeck TS, Louis DN, Hilton J, Colvin M, molecular therapies. Success will probably come with the Silver JS et al. Thymidine kinase activation of ganciclovir in application of global investigation strategies to dissect recurrent malignant gliomas: a gene-marking and neuro- the mechanisms at the basis of cancer and to identify pathological study. J Neurosurg 2000; 92: 804–811.

Cancer Gene Therapy Cell and gene therapy for malignant gliomas L Barzon et al 552 12 Packer RJ, Raffel C, Villablanca JG, Tonn JC, Burdach SE, titer retroviral vector system and their subsequent migration Burger K et al. Treatment of progressive or recurrent in brain tumors. Gene Therapy 1998; 5: 1698–1704. pediatric malignant supratentorial brain tumors with herpes 27 Stefani A-L, Barzon L, Castagliuolo I, Guido M, Pacenti M, simplex virus thymidine kinase gene vector-producer Parolin C et al. Systemic efficacy of combined suicide/ cells followed by intravenous ganciclovir administration. cytokine gene therapy in a murine model of hepatocellular J Neurosurg 2000; 92: 249–254. carcinoma. J Hepatol 2005; 42: 728–735. 13 Sandmair AM, Loimas S, Puranen P, Immonen A, Kossila M, 28 Long Z, Li LP, Grooms T, Lockey C, Nader K, Mychkovsky Puranen M et al. Thymidine kinase gene therapy for human I et al. Biosafety monitoring of patients receiving intracer- malignant glioma, using replication-deficient retroviruses or ebral injections of murine retroviral vector producer cells. adenoviruses. Hum Gene Ther 2000; 11: 2197–2205. Hum Gene Ther 1998; 9: 1165–1172. 14 Rainov NG. A phase III clinical evaluation of herpes simplex 29 Bobo RH, Laske DW, Akbasak A, Morrison PF, Dedrick virus type 1 thymidine kinase and ganciclovir gene therapy as RL, Oldfield EH. Convection-enhanced delivery of macro- an adjuvant to surgical resection and radiation in adults with molecules in the brain. Proc Natl Acad Sci USA 1994; 91: previously untreated glioblastoma multiforme. Hum Gene 2076–2080. Ther 2000; 11: 2389–2401. 30 Colombo F, Zanusso M, Casentini L, Cavaggioni A, 15 Floeth FW, Shand N, Bojar H, Prisack HB, Felsberg J, Franchin E, Calvi P et al. Gene stereotactic neurosurgery Neuen-Jacob E et al. Local inflammation and devasculariza- for recurrent malignant gliomas. Stereotact Funct Neurosurg tion – in vivo mechanisms of the ‘bystander effect’ in VPC- 1997; 68: 245–251. mediated HSV-Tk/GCV gene therapy for human malignant 31 Barba D, Hardin J, Sadelain M, Gage FH. Development of glioma. Cancer Gene Ther 2001; 8: 843–851. anti-tumor immunity following thymidine kinase-mediated 16 Smitt PS, Driesse M, Wolbers J, Kros M, Avezaat C. killing of experimental brain tumors. Proc Natl Acad Sci Treatment of relapsed malignant glioma with an adenoviral USA 1994; 91: 4348–4352. vector containing the herpes simplex thymidine kinase gene 32 Rainov NG, Kramm CM, Banning U, Riemann D, followed by ganciclovir. Mol Ther 2003; 7: 851–858. Holzhausen HJ, Heidecke V et al. Immune response induced 17 Prados MD, McDermott M, Chang SM, Wilson CB, Fick J, by retrovirus-mediated HSV-tk/GCV pharmacogene therapy Culver KW et al. Treatment of progressive or recurrent in patients with glioblastoma multiforme. Gene Therapy glioblastoma multiforme in adults with herpes simplex virus 2000; 7: 1853–1858. thymidine kinase gene vector-producer cells followed by 33 Kramm CM, Korholz D, Rainov NG, Niehues T, Fischer U, intravenous ganciclovir administration: a phase I/II multi- Steffens S et al. Systemic activation of the immune system institutional trial. J Neurooncol 2003; 65: 269–278. during ganciclovir treatment following intratumoral herpes 18 Germano IM, Fable J, Gultekin SH, Silvers A. Adenovirus/ simplex virus type 1 thymidine kinase gene transfer in an herpes simplex-thymidine kinase/ganciclovir complex: pre- adolescent ependymoma patient. Neuropediatrics 2002; 33: liminary results of a phase I trial in patients with recurrent 6–9. malignant gliomas. J Neurooncol 2003; 65: 279–289. 34 Chiocca EA. Oncolytic viruses. Nat Rev Cancer 2002; 2: 19 Lang FF, Bruner JM, Fuller GN, Aldape K, Prados MD, 938–950. Chang S et al. Phase I trial of adenovirus-mediated p53 gene 35 Markert JM, Medlock MD, Rabkin SD, Gillespie GY, Todo therapy for recurrent glioma: biological and clinical results. T, Hunter WD et al. Conditionally replicating herpes simplex J Clin Oncol 2003; 21: 2508–2518. virus mutant, G207 for the treatment of malignant glioma: 20 Jacobs A, Voges J, Reszka R, Lercher M, Gossmann A, results of a phase I trial. Gene Therapy 2000; 7: 867–874. Kracht L et al. Positron-emission tomography of vector- 36 Rampling R, Cruickshank G, Papanastassiou V, Nicoll J, mediated gene expression in gene therapy for gliomas. Lancet Hadley D, Brennan D et al. Toxicity evaluation of 2001; 358: 727–729. replication-competent herpes simplex virus (ICP 34.5 null 21 Voges J, Reszka R, Gossmann A, Dittmar C, Richter R, mutant 1716) in patients with recurrent malignant glioma. Garlip G et al. Imaging-guided convection-enhanced delivery Gene Therapy 2000; 7: 859–866. and gene therapy of glioblastoma. Ann Neurol 2003; 54: 37 Papanastassiou V, Rampling R, Fraser M, Petty R, Hadley 479–487. D, Nicoll J et al. The potential for efficacy of the modified 22 Immonen A, Vapalahti M, Tyynela K, Hurskainen H, (ICP 34.5(À)) herpes simplex virus HSV1716 following Sandmair A, Vanninen R et al. AdvHSV-tk gene therapy intratumoural injection into human malignant glioma: a with intravenous ganciclovir improves survival in human proof of principle study. Gene Therapy 2002; 9: 398–406. malignant glioma: a randomised, controlled study. Mol Ther 38 Harrow S, Papanastassiou V, Harland J, Mabbs R, Petty R, 2004; 10: 967–972. Fraser M et al. HSV1716 injection into the brain adjacent to 23 Palu G, Cavaggioni A, Calvi P, Franchin E, Pizzato M, tumour following surgical resection of high-grade glioma: Boschetto R et al. Gene therapy of glioblastoma multiforme safety data and long-term survival. Gene Therapy 2004; 11: via combined expression of suicide and cytokine genes: a 1648–1658. pilot study in humans. Gene Therapy 1999; 6: 330–337. 39 Chiocca EA, Abbed KM, Tatter S, Louis DN, Hochberg 24 Colombo F, Barzon L, Franchin E, Pacenti M, Pinna V, FH, Barker F et al. A phase I open-label, dose-escalation, Danieli D et al. Combined HSV-TK/IL-2 gene therapy in multi-institutional trial of injection with an E1B-attenuated patients with recurrent glioblastoma multiforme: biological adenovirus, ONYX-015, into the peritumoral region of and clinical results. Cancer Gene Ther 2005; 12: 835–848. recurrent malignant gliomas, in the adjuvant setting. Mol 25 Long Z, Lu P, Grooms T, Mychkovsky I, Westley T, Ther 2004; 10: 958–966. Fitzgerald T et al. Molecular evaluation of biopsy and 40 Csatary LK, Bakacs T. Use of Newcastle disease virus autopsy specimens from patients receiving in vivo retroviral vaccine (MTH-68/H) in a patient with high-grade glioblas- gene therapy. Hum Gene Ther 1999; 10: 733–740. toma. JAMA 1999; 281: 1588–1589. 26 Tamura M, Ikenaka K, Tamura K, Yoshimatsu T, Miyao Y, 41 Csatary LK, Gosztonyi G, Szeberenyi J, Louis DN, Kishima H et al. Transduction of glioma cells using a high- Hochberg FH, Barker F et al. MTH-68/H oncolytic viral

Cancer Gene Therapy Cell and gene therapy for malignant gliomas L Barzon et al 553 treatment in human high-grade gliomas. J Neurooncol 2004; cellular immunotherapy in patients with grade III/IV 67: 83–93. astrocytoma. J Neurooncol 1996; 27: 179–189. 42 Farassati F, Yang A, Lee PWK. Oncogenes in Ras signaling 59 Plautz GE, Barnett GH, Miller DW, Cohen BH, Prayson pathway dictate host-cell permissiveness to herpes simplex RA, Krauss JC et al. Systemic T cell adoptive immunother- type 1. Nat Cell Biol 2001; 3: 745–750. apy of malignant gliomas. J Neurosurg 1998; 89: 42–51. 43 Whitley RJ, Kern ER, Chatterjee S, Chou J, Roizman B. 60 Plautz GE, Miller DW, Barnett GH, Stevens GH, Maffett S, Replication, establishment of latency, and induced reactiva- Kim J et al. T cell adoptive immunotherapy of newly tion of herpes simplex virus gamma 1 34.5 deletion mutants diagnosed gliomas. Clin Cancer Res 2000; 6: 2209–2218. in rodent models. J Clin Invest 1993; 91: 2837–2843. 61 Wood GW, Holladay FP, Turner T, Wang YY, Chiga M. A 44 Goldstein DJ, Weller SK. Factors present in herpes simplex pilot study of autologous cancer cell vaccination and cellular virus type 1-infected cells can compensate for the loss of the immunotherapy using anti-CD3 stimulated lymphocytes large subunit of the viral ribonucleotide reductase: char- in patients with recurrent grade III/IV astrocytoma. J Neuro- acterization of an ICP6 deletion mutant. Virology 1988; 166: oncol 2000; 48: 113–120. 41–51. 62 Andrews DW, Resnicoff M, Flanders AE, Kenyon L, Curtis 45 Mineta T, Rabkin SD, Martuza RL. Treatment of malignant M, Merli G et al. Results of a pilot study involving the use of gliomas using ganciclovir-hypersensitive, ribonucleotide an antisense oligodeoxynucleotide directed against the reductase-deficient herpes simplex viral mutant. Cancer Res insulin-like growth factor type I receptor in malignant 1991; 54: 3963–3966. astrocytomas. J Clin Oncol 2001; 19: 2189–2200. 46 Harland J, Papanastassiou V, Brown SM. HSV1716 63 Schneider T, Gerhards R, Kirches E, Raimund Firsching R. persistence in primary human glioma cells in vitro. Gene Therapy Preliminary results of active specific immunization with 2002; 9: 1194–1198. modified tumor cell vaccine in glioblastoma multiforme. 47 McKie EA, MacLean AR, Lewis AD, Cruickshank G, J Neurooncol 2001; 53: 39–46. Rampling R, Barnett SC et al. Selective in vitro replication of 64 Steiner HH, Bonsanto MM, Beckhove P, Brysch M, herpes simplex virus type 1 (HSV-1) ICP34.5 null mutants in Geletneky K, Ahmadi R et al. Antitumor vaccination of primary human CNS tumours – evaluation of a potentially patients with glioblastoma multiforme: a pilot study to assess effective clinical therapy. Br J Cancer 1996; 74: 745–752. feasibility, safety, and clinical benefit. J Clin Oncol 2004; 22: 48 Detta A, Harland J, Hanif I, Brown SM, Cruickshank G. 4272–4281. Proliferative activity and in vitro replication of HSV1716 in 65 Yu JS, Wheeler CJ, Zeltzer PM, Ying H, Finger DN, Lee PK human metastatic brain tumours. J Gene Med 2003; 5: et al. Vaccination of malignant glioma patients with peptide- 681–689. pulsed dendritic cells elicits systemic cytotoxicity and 49 Reichard KW, Lorence RM, Cascino CJ, Peeples ME, intracranial T-cell infiltration. Cancer Res 2001; 61: Walter RJ, Fernando MB et al. Newcastle disease virus 842–847. selectively kills human tumor cells. J Surg Res 1992; 52: 66 Wheeler CJ, Das A, Liu G, Yu JS, Black KL. Clinical 448–453. responsiveness of glioblastoma multiforme to chemotherapy 50 Schirrmacher V, Haas C, Bonifer R, Ahlert T, Gerhards R, after vaccination. Clin Cancer Res 2004; 10: 5316–5326. Ertel C. Human tumor cell modification by virus infection: 67 Yamanaka R, Abe T, Yajima N, Tsuchiya N, Homma J, an efficient and safe way to produce cancer vaccine with Kobayashi T et al. Vaccination of recurrent glioma patients pleiotropic immune stimulatory properties when using New- with tumour lysate-pulsed dendritic cells elicits immune castle disease virus. Gene Therapy 1999; 6: 63–73. responses: results of a clinical phase I/II trial. Br J Cancer 51 Washburn B, Schirrmacher V. Human tumor cell infection 2003; 89: 1172–1179. by Newcastle Disease Virus leads to upregulation of HLA 68 Yu JS, Liu G, Ying H, Yong WH, Black KL, Wheeler CJ. and cell adhesion molecules and to induction of interferons, Vaccination with tumor lysate-pulsed dendritic cells elicits chemokines and finally apoptosis. Int J Oncol 2002; 21: antigen-specific, cytotoxic T-cells in patients with malignant 85–93. glioma. Cancer Res 2004; 64: 4973–4979. 52 Cassell WA, Garrett RE. Newcastle disease virus as an 69 Rutkowski S, De Vleeschouwer S, Kaempgen E, Wolff JE, antineoplastic agent. Cancer 1965; 18: 863–868. Kuhl J, Demaerel P et al. Surgery and adjuvant dendritic 53 Csatary LK. Viruses in the treatment of cancer. Lancet 1971; cell-based tumour vaccination for patients with relapsed 2: 825. malignant glioma, a feasibility study. Br J Cancer 2004; 91: 54 Pecora AL, Rizvi N, Cohen GI, Meropol NJ, Sterman D, 1656–1662. Marshall JL et al. Phase I trial of intravenous administration 70 Yamanaka R, Homma J, Yajima N, Tsuchiya N, Sano M, of PV701, an oncolytic virus, in patients with advanced solid Kobayashi T et al. Clinical evaluation of dendritic cell cancers. J Clin Oncol 2002; 20: 2251–2266. vaccination for patients with recurrent glioma: results of a 55 Csatary LK, Eckhardt S, Bukosza I, Czegledi F, Fenyvesi C, clinical phase I/II trial. Clin Cancer Res 2005; 11: 4160–4167. Gergely P et al. Attenuated veterinary virus vaccine for the 71 Kikuchi T, Akasaki Y, Irie M, Homma S, Abe T, Ohno T. treatment of cancer. Cancer Detect Prev 1993; 17: 619–627. Results of a phase I clinical trial of vaccination of glioma 56 Csatary LK, Moss RW, Beuth J, Torocsik B, Szeberenyi J, patients with fusions of dendritic and glioma cells. Cancer Bakacs T. Beneficial treatment of patients with advanced Immunol Immunother 2001; 50: 337–344. cancer using a Newcastle disease virus vaccine (MTH-68/H). 72 Kikuchi T, Akasaki Y, Abe T, Fukuda T, Saotome H, Ryan Anticancer Res 1999; 19: 635–638. JL et al. Vaccination of glioma patients with fusions of 57 Ehtesham M, Black KL, Yu JS. Recent progress in dendritic and glioma cells and recombinant human inter- immunotherapy for malignant glioma: treatment strategies leukin 12. J Immunother 2004; 27: 452–459. and results from clinical trials. Cancer Control 2004; 11: 73 Ehtesham M, Kabos P, Gutierrez MA, Samoto K, Black 192–207. KL, Yu JS. Intratumoral dendritic cell vaccination elicits 58 Holladay FP, Heitz-Turner T, Bayer WL, Wood GW. potent tumoricidal immunity against malignant glioma in Autologous tumor cell vaccination combined with adoptive rats. J Immunother 2003; 26: 107–116.

Cancer Gene Therapy Cell and gene therapy for malignant gliomas L Barzon et al 554 74 Young H, Kaplan A, Regelson W. Immunotherapy with 83 Sobol RE, Fakhrai H, Shawler D, Gjerset R, Dorigo O, autologous white cell infusion (‘lymphocytes’) in the treat- Carson C et al. Interleukin-2 gene therapy in a patient with ment of recurrent glioblastoma multiforme: a preliminary glioblastoma. Gene Therapy 1995; 2: 164–167. report. Cancer 1977; 40: 1037–1044. 84 Okada H, Lieberman FS, Edington HD, Witham TF, Wargo 75 Neuwelt EA, Clark K, Kirkpatrick JB, Toben H. Clinical MJ, Cai Q et al. Autologous glioma cell vaccine admixed studies of intratechal autologous lymphocyte infusions in with interleukin-4 gene transfected fibroblasts in the treat- patients with malignant glioma: a toxicity study. Ann Neurol ment of recurrent glioblastoma: preliminary observations in 1978; 4: 307–312. a patient with a favorable response to therapy. J Neurooncol 76 Vaquero J, Martinez R, Ramiro J, Salazar FG, Barbolla L, 2003; 64: 13–20. Regidor C. Immunotherapy of glioblastoma with intra- 85 Resnicoff M, Sell C, Rubini M, Coppola D, Ambrose D, tumoural administration of autologous lymphocytes and Baserga R et al. Rat glioblastoma cells expressing an human lymphoblastoid interferon. A further clinical study. antisense RNA to the insulin-like growth factor-1 (IGF-1) Acta Neurochir (Wien) 1991; 109: 42–45. receptor are nontumorigenic and induce regression of wild- 77 Hayes RL, Koslow M, Hiesiger EM, Hymes KB, Hochster type tumors. Cancer Res 1994; 54: 1235–1242. HS, Moore EJ et al. Improved long term survival after 86 Resnicoff M, Tjuvajev J, Rotman HL, Abraham D, Curtis intracavitary interleukin-2 and lymphokine-activated killer M, Aiken R et al. Regression of C6 rat brain tumors by cells cells for adults with recurrent malignant glioma. Cancer expressing an antisense insulin-like growth factor I receptor 1995; 76: 840–852. RNA. J Exp Ther Oncol 1996; 1: 385–389. 78 Merchant RE, Grant AJ, Merchant LH, Young HF. 87 Akasaki Y, Black KL, Yu JS. Dendritic cell-based immu- Adoptive immunotherapy for recurrent glioblastoma multi- notherapy for malignant gliomas. Expert Rev Neurother forme using lymphokine activated killer cells and recombi- 2005; 5: 497–508. nant interleukin-2. Cancer 1988; 62: 665–671. 88 Banchereau J, Palucka K. Dendritic cells as therapeutic 79 Barba D, Saris SC, Holder C, Rosenberg SA, Oldfield EH. vaccines against cancer. Nat Rev Immunol 2005; 5: 296–306. Intratumoral LAK cell and interleukin-2 therapy of human 89 De Vleeschouwer S, Van Calenbergh F, Demaerel P, Flamen gliomas. J Neurosurg 1989; 70: 175–182. P, Rutkowski S, Kaempgen E et al. Transient local response 80 Quattrocchi KB, Miller CH, Cush S, Bernard SA, Dull ST, and persistent tumor control in a child with recurrent Smith M et al. Pilot study of local autologous tumor malignant glioma: treatment with combination therapy infiltrating lymphocytes for the treatment of recurrent including dendritic cell therapy. Case report. J Neurosurg malignant gliomas. J Neurooncol 1999; 45: 141–157. 2004; 100(Suppl Pediatrics): 492–497. 81 Kruse CA, Schiltz PM, Bellgrau D, Kong Q, Kleinschmidt- 90 Tsugawa T, Kuwashima N, Sato H, Fellows-Mayle WK, DeMasters BK. Intracranial administrations of single or Dusak JE, Okada K et al. Sequential delivery of interferon-a multiple source allogeneic cytotoxic T lymphocytes: chronic gene and DCs to intracranial gliomas promotes an effective therapy for primary brain tumors. J Neurooncol 1994; 19: antitumor response. Gene Therapy 2004; 11: 1551–1558. 161–168. 91 Hesdorffer C, Ayello J, Ward M, Kaubisch A, Vahdat L, 82 Kruse CA, Cepeda L, Owens B, Johnson SD, Stears J, Balmaceda C et al. Phase I trial of retroviral-mediated Lillehei KO. Treatment of recurrent glioma with transfer of the human MDR1 gene as marrow chemoprotec- intracavitary alloreactive cytotoxic T lymphocytes tion in patients undergoing high-dose chemotherapy and and interleukin-2. Cancer Immunol Immunother 1997; 45: autologous stem-cell transplantation. J Clin Oncol 1998; 16: 77–87. 165–172.

Cancer Gene Therapy