Gene therapy enhances chemotherapy tolerance and efficacy in glioblastoma patients

BACKGROUND. Temozolomide (TMZ) is one of the most potent chemotherapy agents for the treatment of glioblastoma. Unfortunately, almost half of glioblastoma tumors are TMZ resistant due to overexpression of methylguanine methyltransferase (MGMT^hi). Coadministration of O⁶-benzylguanine (O⁶BG) can restore TMZ sensitivity, but causes off-target myelosuppression. Here, we conducted a prospective clinical trial to test whether gene therapy to confer O⁶BG resistance in hematopoietic stem cells (HSCs) improves chemotherapy tolerance and outcome.

METHODS. We enrolled 7 newly diagnosed glioblastoma patients with MGMT^hi tumors. Patients received autologous gene-modified HSCs following single-agent carmustine administration. After hematopoietic recovery, patients underwent O⁶BG/TMZ chemotherapy in 28-day cycles. Serial blood samples and tumor images were collected throughout the study. Chemotherapy tolerance was determined by the observed myelosuppression and recovery following each cycle. Patient-specific biomathematical modeling of tumor growth was performed. Progression-free survival (PFS) and overall survival (OS) were also evaluated.

RESULTS. Gene therapy permitted a significant increase in the mean number of tolerated O⁶BG/TMZ cycles (4.4 cycles per patient, P < 0.05) compared with historical controls without gene therapy (n = 7 patients, 1.7 cycles per patient). One patient tolerated an unprecedented 9 cycles and demonstrated long-term PFS without additional therapy. Overall, we observed a median PFS of 9 (range 3.5–57⁺) months and OS of 20 (range 13–57⁺) months. Furthermore, biomathematical modeling revealed markedly delayed tumor growth at lower cumulative TMZ doses in study patients compared with patients that received standard TMZ regimens without O⁶BG.

CONCLUSION. These data support further development of chemoprotective gene therapy in combination with O⁶BG and TMZ for the treatment of glioblastoma and potentially other tumors with overexpression of MGMT.

TRIAL REGISTRATION. Clinicaltrials.gov NCT00669669.

FUNDING. R01CA114218, R01AI080326, R01HL098489, P30DK056465, K01DK076973, R01HL074162, R01CA164371, R01NS060752, U54CA143970.

Glioblastoma multiforme (GBM), the most common and aggressive primary brain tumor in adults, is almost uniformly fatal, carrying a 47% chance of survival at one year from diagnosis and a median overall survival (OS) of 12 to 14 months (1). Unfortunately, over the past decade, new therapies in GBM have accomplished little to improve OS. To date, the most effective chemotherapies to treat GBM are alkylating agents such as temozolomide (TMZ), which is the current standard chemotherapy regimen (2, 3). Despite TMZ potency, median survival in all GBM patients remains less than two years from diagnosis (median 21 months) (2, 3). Moreover, approximately 50% of GBM tumors overexpress methylguanine methyltransferase (MGMT^hi), indicated by hypomethylation of the promoter controlling MGMT expression. MGMT^hi tumor cells are resistant to TMZ due to the ability of the catalytic MGMT protein to repair TMZ-induced cytotoxic DNA damage, conferring poor median survival (12.6 months) (2, 3). Notably, MGMT overexpression in alkylating agent-resistant tumor cells is also found in many other tumor types (reviewed in ref. 4).

The small nucleoside inhibitor, O⁶-benzylguanine (O⁶BG), effectively depletes MGMT activity by mimicking the methylated guanine nucleotide base targeted by MGMT, thereby binding MGMT protein and causing structural changes that mark the protein-nucleotide complex for degradation. Concurrent administration of O⁶BG restores tumor cell sensitivity to alkylating agents such as TMZ, which was previously investigated as a strategy to improve TMZ efficacy in recurrent GBM patients (5, 6). However, early phase studies revealed that O⁶BG/TMZ administration caused severe off-target myelosuppression. Specifically, patients administered TMZ at the maximum tolerated dose (MTD) (472 mg/m²) in combination with O⁶BG exhibited 46% incidence of grade 4 neutropenia (5, 6). The hematopoietic-specific toxicity was attributed to low-to-nonexistent levels of MGMT detected in hematopoietic stem cells (HSCs) and progenitor cells (7).

Previously, we and others have shown that expression of the O⁶BG-resistant MGMT mutant P140K by hematopoietic cells provides significant chemoprotection against hematopoietic toxicity from O⁶BG/alkylator chemotherapy (8). Based on our preclinical studies, we initiated a prospective phase I/II clinical trial in which newly diagnosed MGMT^hi GBM patients were transplanted with autologous P140K gene–modified hematopoietic CD34⁺ cells to prevent hematopoietic toxicity during combination O⁶BG/TMZ chemotherapy (9). We recently reported the feasibility and safety of this approach based on observations in the first three patients enrolled on this trial (9). We have now treated a total of seven patients to date. Here, we report the efficacy of this approach, which includes increased chemotherapy tolerance and improved outcome in these patients.

Patient data and treatment. As detailed in Figure 1, a total of eight patients with newly diagnosed, MGMT promoter unmethylated malignant gliomas (WHO grade IV) were enrolled into the first phase of the study. One patient (patient 5) was classified as recursive partitioning analysis (RPA) class III (expected median survival 17.1 months, 70% OS rate at 1 year from diagnosis); five patients (patients 1, 2, 3, 4, 6, and 7) were classified as RPA class IV (expected median survival 11.2 months, 46% 1 year OS rate); and one patient (case patient 8) was classified as RPA class V (expected median survival 7.5 months, 28% 1 year OS rate) (Table 1 and ref. 10). All patients received a greater than 50% surgical resection of the primary tumor, followed by radiation therapy (RT) without concomitant TMZ, since this treatment is unlikely to significantly benefit patients with MGMT promoter unmethylated tumors and would predictably compromise the HSC pool available for subsequent gene modification. Following RT, patients were mobilized with granulocyte colony-stimulating factor (G-CSF) and underwent leukapheresis; autologous peripheral blood CD34⁺ cells from each patient were enriched. One patient (case patient 8) did not sufficiently mobilize to allow for collection of 2.5 × 10⁶ or more CD34⁺ cells per kilogram (kg) of body weight and was removed from the study. For all other patients, enriched CD34⁺ cells were genetically modified with a retroviral vector encoding the P140K transgene (Figure 2A). Following gene transfer, the observed mean level of colony-forming cells containing provirus was 65% (range 41%–90%). Cell products were infused at doses ranging from 3.8 × 10⁶ to 21.9 × 10⁶ CD34⁺ cells/kg within 24 to 48 hours after administration of single-agent carmustine (BCNU) at a dose of 600 mg/m². Following hematopoietic recovery from BCNU, study case patients received between 2 and 9 cycles (mean = 4.4 cycles) of adjuvant chemotherapy, including O⁶BG and TMZ at or above the previously established MTD of TMZ (472 mg/m²) (Table 1). To maximize potential therapeutic benefit, intrapatient dose escalation of TMZ was permitted. Criteria for dose escalation required at least two prior cycles at the MTD of TMZ with hematopoietic recovery (absolute neutrophil count [ANC] ≥ 1,500/μl and platelets [PLTs] ≥ 100,000/μl) within 42 days from the previous cycle and no grade 4 hematologic toxicity lasting longer than 7 days as well as no grade 4 nonhematologic toxicity associated with the therapy. Additionally, patients had to have had at least 1% detectable gene-modified granulocytes present in peripheral blood at the time of chemotherapy. Two enrolled patients (case patients 1 and 7) received a total of three TMZ dose-escalated cycles (590 mg/m²) based on these criteria. Patient response was measured every two months during treatment, and O⁶BG/TMZ therapy was discontinued if clinical evidence for disease progression while on therapy was observed. No dose-limiting toxicities, as defined in the clinical protocol, were observed in any enrolled patients (see clinical protocol included in Supplemental Material; supplemental material available online with this article; doi:10.1172/JCI76739DS1).

Figure 1

Flow diagram of study protocol.

Figure 2

Chemoresistance of wbc in vivo after infusion of P140K gene-modified CD34⁺ cells and O⁶BG/TMZ chemotherapy. (A) Schematic of the P140K-encoding gammaretrovirus vector used for transduction of autologous human CD34⁺ cells. (B–H) Blood cell counts for chemotherapy treatment per protocol (ANC and PLTs) for case patients 1 to 7 (B–H, respectively). Each graph shows ANC (white circles) and PLT counts (black squares) as measured from peripheral blood (y axis) over time (x axis) and chemotherapy with O⁶BG/TMZ (inverted triangles). Black inverted triangles in B (case patient 1) and H (case patient 7) represent dose-escalated chemotherapy cycles with TMZ dose of 590 mg/m². Upper and lower dashed lines represent the threshold counts for each hematopoietic cell type required for chemotherapy retreatment in the study.

Table 1

Patient demographics, treatment, and survival

Myelosuppression associated with single agent BCNU conditioning. With infusion of gene-modified CD34⁺ cells, hematopoietic suppression associated with BCNU conditioning was moderate: the lowest ANC observed was 30/microliter (μl) (case patient 6) and the lowest PLT count was 28,000/μl (case patient 1). An ANC of less than 100/μl was observed in four of seven patients lasting no more than 4 days (range 1–4 days) and without fever. Mean hematopoietic recovery following BCNU, measured as both an ANC greater than 1,500/μl and PLT greater than 100,000/μl, was 29 days after transplant (range 17–39 days) and no complications requiring intervention during the engraftment period were observed (Figure 2, B–H).

Myelosuppression associated with posttransplant O⁶BG/TMZ chemotherapy. Myelosuppression following O⁶BG/TMZ chemotherapy was observed in five of seven patients and occurred most frequently after the first, second, and third cycles, likely due to hematopoietic reconstitution by gene-modified short-term progenitors with limited repopulation capacity (Table 2). At 100 days after transplantation, the only incidence of significant chemotherapy-associated myelosuppression occurred in case patient 6, who displayed little to no gene-modified (i.e., chemoprotected) cells in circulation at the time of chemotherapy (Figure 3). The unanticipated loss in long-term engraftment of gene-modified cells in this patient resulted from a change in the lot formulation of commercial medium used during ex vivo culture of the cells for gene transfer, which is discussed in more detail below. Importantly, no significant myelosuppression was observed after the first 100 days following transplant in the patient with the longest follow-up (case patient 1), who displayed long-term engraftment of gene-modified cells and received an unprecedented 9 cycles of O⁶BG/TMZ chemotherapy at or above the previously established MTD for TMZ (Supplemental Table 1).

Figure 3

Persistence of P140K gene-modified cells in vivo. Percentage of gene-modified wbc within circulating peripheral blood of patients 1 to 7 are represented in A–G, respectively. Each graph represents the percentage of circulating bulk wbc (black diamonds), granulocytes (white circles), and lymphocytes (asterisks), with integrated transgene-containing provirus by PCR assay (y axis) over time (x axis). Chemotherapy (inverted triangles) is shown as well.

Table 2

Chemotherapy-associated myelosuppression

Gene-modified cell responses. Transient increases in gene-modified granulocytes after chemotherapy corresponded with transient increases observed in total wbc populations and coincided with nadir neutropenia and thrombocytopenia after each cycle, demonstrating protection of gene-modified peripheral blood cells (compare Figure 2 with Figure 3). In three patients (case patients 4–6), a loss of detectable gene-modified cells in the peripheral blood was observed between 40 and 100 days after transplant (Figure 3, D–F). Further review of manufacturing records revealed a correlation between the commercial lot of HSC culture medium used during ex vivo gene transfer into CD34⁺ hematopoietic cells from these patients compared with the first three patients, all of whom demonstrated persistent gene-modified cell engraftment. There were no observable significant differences in cell fitness at the time of transplant based on cell dose and viability, efficiency of gene transfer, or colony-forming capacity among these 6 patients (Supplemental Table 2); thus, the observed differences in long-term engraftment in vivo were unanticipated. Through communication with the manufacturer, the difference in media lot performance was attributed to bovine serum albumin used during formulation. We therefore switched to a xenogen-free version of HSC supportive medium for the subsequent gene modification of autologous CD34⁺ cells from case patient 7. Following gene transfer, we observed no significant difference in cell dose and viability, efficiency of gene transfer, or colony-forming capacity compared with the previous 6 patients (Supplemental Table 2), but we did observe robust, long-term engraftment of gene-modified cells in this patient throughout chemotherapy (Figure 3G). Interestingly, in the four patients with durable gene marking throughout chemotherapy (patients 1, 2, 3, and 7), we observed a gradual decline in circulating gene-modified peripheral blood cell levels immediately following discontinuation of chemotherapy, suggesting a potential disadvantage for gene-modified stem cells in the absence of chemoselective pressure.

Patient responses. All seven case patients were evaluable for response throughout treatment. We observed one partial response (PR) per Revised Assessment in Neuro-Oncology (RANO) (11) criteria while on study treatment (case patient 1). While this response was durable throughout therapy, declining levels of circulating gene-modified cells observed in this patient between cycles seven and nine warranted discontinuation of O⁶BG/TMZ treatment. This patient received no further treatment for GBM and more recently demonstrated complete response (CR) (Figure 4). Six out of seven patients demonstrated progressive disease while on O⁶BG/TMZ therapy with a median progression-free survival (PFS) from diagnosis of 9 months (Table 1). While on therapy, five of these six patients demonstrated a best response of stable disease by RANO criteria and one patient (case patient 6) demonstrated progressive disease. Median OS was 20 months for the seven patients treated to date, and we observed 100% survival of enrolled patients at one year from diagnosis. Six of the seven patients are evaluable for two-year survival from diagnosis at the time of this writing. We observed three of seven patients alive at two years, a significant improvement compared with historical MGMT^hi patients receiving RT alone as upfront therapy followed by adjuvant TMZ (1 of 54; P = 0.02) (2). Moreover, all seven patients surpassed the median survival for patients in the same RPA class: reported median OS is 17.1 months, 11.2 months, and 7.5 months for RPA classes III, IV, and V, respectively (10).

Figure 4

Patient-specific biomathematical modeling of tumor growth and treatment response. (A) UVC for predicted tumor growth if left untreated for case patient 1. Two initial pretreatment MRI scans (blue x’s) were used to determine the predicted tumor growth (solid line) if left untreated. The first MRI was obtained on the day of clinical presentation, and the second MRI was obtained as part of neuronavigational guidance protocol on the day of resection (day 10). Tumor radius (y axis) is determined from follow-up posttreatment MRIs (black x’s) and plotted against time since the first MRI in days (x axis). The treated tumor radius is assessed relative to the projected untreated growth to determine the response measured as DG or RTR. Note: Increase in enhancement observed by MRI at 182 days (D, below) was determined to be consistent with pseudoprogression (PSP) by second surgical resection. (B–G) Observed clinical response in case patient 1. Contrast-enhanced MRI scans obtained, in chronological order, at initial diagnosis (B, second gray x denoted in A), after RT and transplant (before O⁶BG/TMZ chemotherapy) (C), following cycle II of O⁶BG/TMZ chemotherapy (interval increase representative of pseudoprogression corresponding to second black x in A) (D), following second surgical resection, which indicated pseudoprogression (for pathology, see Supplemental Figure 3) (E), following cycle 9 of O⁶BG/TMZ chemotherapy (F), and at 2 years since diagnosis (11 months after therapy discontinuation) (G).

Biomathematical modeling to evaluate O⁶BG benefit. To determine potential therapeutic gain achieved with concomitant O⁶BG administration during the adjuvant phase of TMZ therapy, we applied patient-specific biomathematical modeling of tumor growth (12). We examined two pretreatment MRI scans (typically at presentation and immediately prior to surgical resection), preferably on the same MRI machine to accurately compute model kinetics. We then predicted the growth of each patient’s tumor, if left untreated, in terms of millimeters (mm) of radial tumor growth per day (Figure 4). The resulting patient-specific “untreated virtual control” (UVC) provided the baseline against which treatment effect was measured (Figure 4, plot of untreated radial tumor growth versus time) (12, 13). Treatment effect was described in two ways: first, as the x axis deflection from the UVC, days gained (DG, defined as number of days the treatment delayed tumor progression), and second, as the y axis deflection from the UVC, radial treatment response (RTR, defined as the radius of tumor growth prevented). Of the seven patients treated, four had the required pretreatment imaging necessary to reliably model untreated tumor growth (case patients 1–4).

We observed a robust deflection of tumor growth from the patient-specific UVC at each time point of response evaluation analyzed for all four patients, (Supplemental Figure 1 and Supplemental Figure 2). For the three patients for whom multiple response evaluations were available (case patients 1–3), we observed increased deflection from the UVC over time and with additional TMZ administration (Table 3).

Table 3

Treatment benefit values for each patient analyzed by biomathematical modeling

We then assessed whether this observed deflection in tumor growth with combined TMZ and O⁶BG provided a substantial treatment benefit over standard therapy (TMZ alone) by comparing treatment responses observed in study (case) patients with those of five control patients who received standard therapy and were treated at our institution (Table 3). We matched patients for comparison by similarity of model-determined disease kinetics (net invasion rate, D, and net proliferation rate, ρ) and similar prognostic clinical characteristics, including diagnosis of GBM, extent of surgical resection (EOR), RPA class, age, and Karnofsky Performance Status (KPS) (14) at time of diagnosis. MGMT promoter methylation status for control patients was also determined, since this variable has been prognostic for survival benefit achieved with standard care TMZ regimens (Supplemental Table 3). For patient 1, two suitable controls were identified (control 1A and 1B). Velocity of tumor growth was greater in case patients (mm/d: range 0.11–0.66, mean 0.34) compared with controls (mm/d: range 0.035–0.26, mean 0.13), but was not significantly different in distribution (P > 0.10). All case patients demonstrated greater average DG and prevention of radial tumor growth per mg of TMZ administered compared with matched controls, which was sustained for the duration of the treatment course (Figure 5 and Table 3). Case patient 4 displayed a single value for both parameters measured that was greater than that observed over all time points for the corresponding matched control patient 4, despite receiving fewer TMZ chemotherapy cycles.

Figure 5

Biomathematical modeling reveals greater response to treatment as a function of TMZ administered in study case patients versus matched controls. (A) Box and whisker plots illustrating the significant difference in mean DG per milligram of TMZ received observed between study patients (case) and matched patients receiving standard care regimens (control). (B) Box and whisker plots illustrating the significant difference in mean RTR per milligram of TMZ received observed between study patients and matched patients receiving standard care regimens. For each plot, box bounds represent the first quartile (lower bound) and third quartile (upper bound). Lines within the box represent the median. Whiskers represent the difference from the minimum value observed in the data set to the first quartile (lower whisker) and the difference from the third quartile to the maximum value observed (upper whisker).

Here, we show that gene therapy to protect HSCs and progenitor cells from the toxic effect of O⁶BG/TMZ chemotherapy is feasible and allows for the administration of significantly more chemotherapy cycles in patients with newly diagnosed, poor-prognosis GBM than previously described in patients without the benefit of chemoprotected blood cells (6). We observed no significant hematologic toxicity from chemotherapy following the first 100 days after transplant of chemoprotected, gene-modified cells, permitting up to nine cycles of O⁶BG/TMZ chemotherapy. Importantly, we demonstrate that the ability to give significantly more O⁶BG/TMZ cycles was associated with a significant improvement in overall outcome. Specifically, we observed a significant improvement in two-year survival in patients treated with gene therapy compared with historical patients receiving TMZ without the benefit of O⁶BG (2).

Given the high frequency of GBM patients with unmethylated MGMT promoters in tumor cells (~50% of both young and elderly patients), there is a substantial need to develop treatment strategies that improve the standard care approach (3, 15). In patients classified as Radiation Therapy Oncology Group (RTOG)/RPA class IV, median OS is 11.2 months, and 1-year survival rates are 46% (10). Six of the seven patients treated in this category were classified as RPA class IV (case patients 1, 2, 3, 4, 6, and 7), and all surpassed this median expected survival (range 14–57⁺ months). In the single RPA class III patient enrolled and treated (case patient 5), we still observed an improvement in survival over the expected: RPA class III median survival is 17.1 months, with a 70% 1-year survival rate, while patient 5 survived 28 months. For all categories, MGMT^hi status is considered a poor prognostic indicator for OS; thus, the 100% survival rate at one year and the 33% survival rate at two years from diagnosis observed in our study are highly significant. In addition, case patients 1, 3, and 5 have demonstrated significantly extended OS, surpassing even median survival for MGMT^lo tumor status patients receiving standard therapy (21.7 months) (2). Though these phase I study results are encouraging, we cautiously interpret clinical benefit compared with historical response, as true clinical benefit over standard therapy can only be detected following a phase III, randomized, double-blinded clinical trial.

Most notably, addition of O⁶BG to the adjuvant TMZ regimen permitted the observed improvement in OS to be achieved at far lower cumulative TMZ doses than typically received by GBM patients in standard care regimens. Case patient 1, who displayed the most robust response, received 9,480 total mg of TMZ during nine cycles of chemotherapy treatment, whereas his corresponding institutional matched controls received 4,900 and 9,650 mg TMZ, respectively, in only three cycles of treatment.

To characterize the treatment benefits achieved with O⁶BG added to the TMZ regimen, we applied a precision medicine, biomathematical model to compare the treatment strategy employed in study patients with standard care regimens in patients treated at our institution (12, 13, 16–20). We show that model-predicted, untreated tumor growth was deflected by a greater degree in study patients compared with matched control patients, at lower cumulative TMZ doses when administered with O⁶BG.

There are several assumptions implicit in using modeling approaches, such as the approach applied here. In this case, we assumed that patient-individualized predictions of tumor growth, if left untreated, were consistent with a fixed/intrinsic tumor kinetic. With treatment, those rates can change, but the UVC is informed by this baseline expectation of consistency. We have shown that this kinetic is consistent over as much as 15 years of imaging follow-up without treatment in low-grade gliomas (21) and for over 1 year in rare, untreated GBM tumors (19). However, we were only able to predict untreated tumor growth in four of seven treated patients, owing to inconsistencies in pretreatment imaging. Addressing this will likely require improvements in both pretreatment imaging standards and biomathematical modeling. Importantly, for the patients evaluated, tumor growth velocity was similar between treatment groups and, thus, does not explain the differences in treatment gains observed. For the study patients evaluated, increasing response values per mg of TMZ suggest a potentiating efficacy with successive rounds of TMZ. This supports our original hypothesis that the ability to deliver multiple cycles of TMZ in combination with O⁶BG results in improved tumor response and overall outcome in poor-prognosis GBM patients.

In a prior phase II clinical trial of O⁶BG and TMZ chemotherapy in recurrent GBM patients with TMZ-resistant tumors, a significant improvement in OS and PFS was not observed, primarily owing to a dose reduction of TMZ required (50%) to avoid hematologic toxicity (6). While our study treated newly diagnosed GBM patients with TMZ-resistant tumors based on MGMT promoter methylation status, our findings suggest that the ability to administer O⁶BG/TMZ at or above the previously described MTD for TMZ by chemoprotection of the hematopoietic system can improve response. Moreover, our data suggest that between six and nine cycles of this chemotherapy regimen may be all that is needed to elicit a prolonged therapeutic benefit.

The observed decline in gene-modified cells following therapy discontinuation in patients with durable gene marking throughout treatment could indicate that constitutive MGMT expression in blood cells may not be ideal. We envision that this limitation could be overcome by use of a less active or regulated (chemotherapy inducible) promoter to drive MGMT expression in gene-modified cells. Furthermore, the loss in long-term repopulating gene-modified cells resulting from lot-to-lot variation in commercial cell culture medium used during gene transfer highlights the need to reduce xenogen exposure during gene transfer, as well as to develop more potent tests for HSC fitness ex vivo. Importantly, given the improved therapeutic benefit and OS observed in all patients treated in the current study, these findings also suggest that permanent gene modification may not be required to elicit an improvement in outcome. Based on our results, six to nine cycles of O⁶BG/TMZ chemotherapy can substantially improve outcome; therefore, prolonged gene modification of peripheral blood cells is not necessary.

The results of this study indicate that gene modification with the O⁶BG-resistant P140K mutant MGMT can chemoprotect the blood and bone marrow from myelosuppression associated with combined O⁶BG and TMZ chemotherapy. Moreover, addition of O⁶BG to TMZ in the post-RT milieu can provide significant treatment gains per mg of TMZ dose compared with TMZ alone. This consistent pattern of superior treatment response in study patients provides strong evidence for increased therapeutic benefit of the study protocol over standard of care regimens for MGMT^hi GBM patients, supporting continued investigation of this approach.

In conclusion, our results show that P140K-modified HSCs and progenitor cells, in combination with dose-intense TMZ and O⁶BG, is a clinically feasible alternative to standard therapy for poor prognosis, MGMT^hi GBM patients. Importantly this also has implications for other MGMT-overexpressing brain tumors or even other types of tumors for which alkylating agent chemotherapy could be used for treatment.

View Supplemental data

View Clinical Protocols

View ICMJE disclosure forms

We thank Kenneth Cornetta and the Indiana University Virus Production Facility for production of the clinical grade virus used in this study. Devikha Chandrasekaran, Christina Ironside, Nathaniel Williams, and Garrick Horn are acknowledged for technical assistance in conduct of this work. We thank the Gene and Cell Therapy Laboratory of UW, the Cell Processing Facility of the Fred Hutchinson Cancer Research Center, and the Cellular Therapy Laboratory of the SCCA for support during postcollection processing and transduction. We also thank all clinical service teams involved in the conduct of this study including the Neuro-Oncology and Infusion Service teams of the UW Medical Center and SCCA, the Neurological Surgery and Radiation Oncology teams of the UW Medical Center and the Gold Autologous Transplant Team of the SCCA. We appreciate Grace Choi’s assistance in the collection of clinical data and preparation of the manuscript. The authors are grateful for research funding from the NIH (grants R01CA114218, R01AI080326, R01HL098489, P30DK056465, and R01HL074162 to H.-P. Kiem; K01DK076973 to B.C. Beard; R01CA164371, R01NS060752, and U54CA143970 to K.R. Swanson). H.-P. Kiem is a Markey Molecular Medicine Investigator and received support as the inaugural recipient of the José Carreras/E. Donnall Thomas Endowed Chair for Cancer Research. K.R. Swanson is a Robert H. Lurie Comprehensive Cancer Center Investigator at Northwestern University. K.R. Swanson was a recipient of the James D. Murray Endowed Chair and Academic Pathology Fund, both at UW, and Zell Scholar and Wirtz Scholar awards at Northwestern University. K.R. Swanson is a recipient of the James S. McDonnell Foundation award.

Address correspondence to: Hans-Peter Kiem, Fred Hutchinson Cancer Research Center, Mail Stop D1-100, P.O. Box 19024, Seattle, Washington 98109-1024, USA. Phone: 206.667.4425; E-mail: hkiem@fhcrc.org.

Conflict of interest: The authors have declared that no conflict of interest exists.

Role of funding source: The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH nor its subsidiary Institutes and Centers.

Reference information: J Clin Invest. 2014;124(9):4082–4092. doi:10.1172/JCI76739.

1. Tanaka S, Louis DN, Curry WT, Batchelor TT, Dietrich J. Diagnostic and therapeutic avenues for glioblastoma: no longer a dead end? Nat Rev Clin Oncol. 2013;10(1):14–26.
   View this article via: PubMed Google Scholar
2. Stupp R, et al. Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial. Lancet Oncol. 2009;10(5):459–466.
   View this article via: PubMed CrossRef Google Scholar
3. Stupp R, et al. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med. 2005;352(10):987–996.
   View this article via: PubMed CrossRef Google Scholar
4. Liu L, Gerson SL. Targeted modulation of MGMT: clinical implications (Review). Clin Cancer Res. 2006;12(2):328–331.
   View this article via: PubMed CrossRef Google Scholar
5. Quinn JA, et al. Phase I trial of temozolomide plus O6-benzylguanine 5-day regimen with recurrent malignant glioma. Neuro Oncol. 2009;11(5):556–561.
   View this article via: PubMed CrossRef Google Scholar
6. Quinn JA, et al. Phase II trial of temozolomide plus o6-benzylguanine in adults with recurrent, temozolomide-resistant malignant glioma. J Clin Oncol. 2009;27(8):1262–1267.
   View this article via: PubMed CrossRef Google Scholar
7. Allay JA, Dumenco LL, Koc ON, Liu L, Gerson SL. Retroviral transduction and expression of the human alkyltransferase cDNA provides nitrosourea resistance to hematopoietic cells. Blood. 1995;85(11):3342–3351.
   View this article via: PubMed Google Scholar
8. Beard BC, Trobridge GD, Ironside C, McCune JS, Adair JE, Kiem H-P. Efficient and stable MGMT-mediated selection of long-term repopulating stem cells in nonhuman primates. J Clin Invest. 2010;120(7):2345–2354.
   View this article via: JCI PubMed CrossRef Google Scholar
9. Adair JE, et al. Extended survival of glioblastoma patients after chemoprotective HSC gene therapy. Sci Transl Med. 2012;4(133):133ra57.
   View this article via: PubMed Google Scholar
10. Li J, et al. Validation and simplification of the Radiation Therapy Oncology Group recursive partitioning analysis classification for glioblastoma. Int J Radiat Oncol Biol Phys. 2011;81(3):623–630.
    View this article via: PubMed CrossRef Google Scholar
11. Wen PY, et al. Updated response assessment criteria for high-grade gliomas: response assessment in neuro-oncology working group. J Clin Oncol. 2010;28(11):1963–1072.
    View this article via: PubMed CrossRef Google Scholar
12. Neal ML, et al. Discriminating time to progression and survival using a response metric tuned to patient-specific glioblastoma kinetics. PLoS One. 2013;8(1):e51951.
    View this article via: PubMed CrossRef Google Scholar
13. Neal ML, et al. Resonse classification based on a minimal model of glioblastoma growth is prognostic for clinical outcomes and distinguishes progression from pseudoprogression. Cancer Res. 2013;73(10):2976–2986.
    View this article via: PubMed CrossRef Google Scholar
14. Karnofsky DA, Burchenal JH. The clinical evaluation of chemotherapeutic agents in cancer. In: Macleod CM, ed. Evaluation of Chemotherapeutic Agents. New York, NY: Columbia University Press; 1949:191–205.
    View this article via: PubMed Google Scholar
15. Yin AA, et al. The predictive but not prognostic value of MGMT promoter methylation status in elderly glioblastoma patients: a meta-analysis. PLoS One. 2014;9(1):e85102.
    View this article via: PubMed CrossRef Google Scholar
16. Swanson KR, et al. Velocity of radial expansion of contrast-enhancing gliomas and the effectiveness of radiotherapy in individual patients: a proof of principle. Clin Oncol (R Coll Radiol). 2008;20(4):301–308.
    View this article via: PubMed CrossRef Google Scholar
17. Swanson KR, Rostomily RC, Alvord EC Jr. A mathematical modelling tool for predicting survival of individual patients following resection of glioblastoma: a proof of principle. Br J Cancer. 2008;98(1):113–119.
    View this article via: PubMed CrossRef Google Scholar
18. Wang CH, et al. Prognostic significance of growth kinetics in newly diagnosed glioblastomas revealed by combining serial imaging with a novel biomathematical model. Cancer Res. 2009;69(23):9133–9140.
    View this article via: PubMed CrossRef Google Scholar
19. Swanson KR, Alvord EC Jr, Murray JD. Virtual brain tumours (gliomas) enhance the reality of medical imaging and ^hilight inadequacies of current therapy. Br J Cancer. 2002;86(1):14–18.
    View this article via: PubMed CrossRef Google Scholar
20. Rockne R, et al. Predicting the efficacy of radiotherapy in individual glioblastoma patients in vivo: a mathematical modeling approach. Phys Med Biol. 2010;55(12):3271–3285.
    View this article via: PubMed CrossRef Google Scholar
21. Mandonnet E, et al. Continuous growth of mean tumor diameter in a subset of grade II gliomas. Ann Neurol. 2003;53(4):524–528.
    View this article via: PubMed Google Scholar
22. Beard BC, Mezquita P, Morris JC, Kiem H-P. Efficient transduction and engraftment of G-CSF-mobilized peripheral blood CD34⁺ cells in nonhuman primates using GALV-pseudotyped γ-retroviral vectors. Mol Ther. 2006;14(2):212–217.
    View this article via: PubMed CrossRef Google Scholar
23. Swanson KR, Alvord EC Jr, Murray JD. A quantitative model for differential motility of gliomas in grey and white matter. Cell Prolif. 2000;33(5):317–329.
    View this article via: PubMed CrossRef Google Scholar
24. Harpold HL, Alvord EC Jr, Swanson KR. The evolution of mathematical modeling of glioma proliferation and invasion. J Neuropathol Exp Neurol. 2007;66(1):1–9.
    View this article via: PubMed Google Scholar
25. Szeto MD, et al. Quantitative metrics of net proliferation and invasion link biological aggressiveness assessed by MRI with hypoxia assessed by FMISO-PET in newly diagnosed glioblastomas. Cancer Res. 2009;69(10):4502–4509.
    View this article via: PubMed CrossRef Google Scholar
26. Shaw EG, et al. Reexamining the radiation therapy oncology group (RTOG) recursive partitioning analysis (RPA) for glioblastoma multiforme (GVM) patients [abstract]. Int J Radiat Oncol Biol Phys. 2003;57(:2, Supplement.)