Supplementary Materials for

Phase 1 trial of vocimagene amiretrorepvec and 5-fluorocytosine for

recurrent high-grade glioma

Timothy F. Cloughesy, Joseph Landolfi, Daniel J. Hogan, Stephen Bloomfield,

Bob Carter, Clark C. Chen, J. Bradley Elder, Steven N. Kalkanis, Santosh Kesari,

Albert Lai, Ian Y. Lee, Linda M. Liau, Tom Mikkelsen, Phioanh Leia Nghiemphu,

David Piccioni, Tobias Walbert, Alice Chu, Asha Das, Oscar R. Diago, Dawn Gammon,

Harry E. Gruber, Michelle Hanna, Douglas J. Jolly, Noriyuki Kasahara, David McCarthy,

Leah Mitchell, Derek Ostertag, Joan M. Robbins, Maria Rodriguez-Aguirre,

Michael A. Vogelbaum*

*Corresponding author. Email: vogelbm@ccf.org

Published 1 June 2016, Sci. Transl. Med. 8, 341ra75 (2016)

DOI: 10.1126/scitranslmed.aad9784

This PDF file includes:

Materials and Methods

Fig. S1. Simplified schematic for the phase 1 trial (NCT01470794).

Fig. S2. Tumor-specific Toca 511 staining in re-resected tumors after multiple

cycles of Toca FC.

Fig. S3. Evidence of potential pseudoprogression on resected tumor after Toca

511 and Toca FC treatment.

Fig. S4. Comparison of subjects with glioblastoma at first or second recurrence

treated with Toca 511 and Toca FC to lomustine.

Fig. S5. Peripheral blood CD4+ T cell modulation after Toca 511 and Toca FC

dosing.

Fig. S6. Tracking of Toca 511 DNA and RNA signal in whole blood and plasma

over time.

Fig. S7. Molecular classification of tumor samples from study subjects based on

mRNA expression.

Fig. S8. Intratumor versus intertumor heterogeneity in RNA expression profiles.

Fig. S9. No relationship between variation in mRNA expression and confounding

technical factors.

Fig. S10. No correlation between neural subtype and prognosis in newly

diagnosed glioblastoma.

www.sciencetranslationalmedicine.org/cgi/content/full/8/341/341ra75/DC1

Fig. S11. Identification of an mRNA profile (SRNS) associated with longer

survival after Toca 511.

Fig. S12. No correlation between SRNS and prognosis in newly diagnosed

glioblastoma.

Fig. S13. Study survival–related neural subtype samples likely derive from

nonenhancing regions of tumors.

Fig. S14. Negative expression between SPOC1 expression and subject survival

time with Toca 511 and Toca FC therapy.

Fig. S15. MGMT promoter methylation.

Table S1. Toca 511 and Toca FC: Baseline demographic and clinical

characteristics.

Table S2. Additional baseline demographic and clinical characteristics.

Table S3. Dosing cohorts.

Table S4. Detection of Toca 511 in re-resected tumors.

Table S5. Best overall response in the efficacy evaluable population.

Table S6. Adverse events and serious adverse events related to Toca 511 and

Toca FC.

Table S7. Related adverse events: Toca 511 and Toca FC compared to lomustine.

Table S8. Adverse events regardless of attribution: Toca 511 and Toca FC

compared to lomustine.

Table S9. Summary of multivariate analysis for survival (TCGA neural

signature).

Table S10. Summary of multivariate analysis for survival (SRNS signature).

References (45, 46)

Other Supplementary Material for this manuscript includes the following:

(available at www.sciencetranslationalmedicine.org/cgi/content/full/8/341/341ra75/DC1)

Table S11 (Microsoft Excel format). RNA sequencing normalized results (reads

per kilobase of transcript per million mapped reads values).

Table S12 (Microsoft Excel format). SRNS gene list.

Table S13 (Microsoft Excel format). Subject summary.

Supplemental Materials and Methods

qPCR and qRT-PCR for Toca 511 DNA and RNA in human blood and plasma

DNA was extracted from previously frozen whole blood using DNA extraction kits and

following the manufacturer’s suggested protocol (Qiagen). RNA was extracted from

previously frozen plasma using RNA extraction kits and following the manufacturer’s

suggested protocol (Qiagen). qPCR and qRT-PCR protocols were previously described

(7, 8).

RNA isolation and quantification

Tumor pieces were separated from frozen bulk tumor with a scalpel on dry ice. Total

RNA was isolated either using Maxwell 16 Total RNA Purification Kit (Promega) or via

sequential purification with SurePrep RNA/DNA/Protein Purification Kit (Fisher),

followed by DNase treatment and further purification via miRneasy kit (Qiagen). RNA

was quantified on the Nanodrop 8000. RNA integrity was determined using Tapestation

(Agilent).

Library prep and sequencing

100-200 ng total RNA was depleted of ribosomal RNA (NEB) and converted into

barcoded sequencing libraries using TruSeq Stranded mRNA LT (Illumina). Libraries

were sequenced on an Illumina Hiseq machine generating ~3 x 107; 100-130 base

single-end reads per sample on average.

Alignment and post-alignment processing

These steps were performed on an Amazon Web Services instance running an Ubuntu

14 Linux environment. Raw sequencing results in the form of FASTQ files were used as

input for alignment of the sequencing data to the human ENSEMBL reference RNA

database GRCh37. ENSEMBL and human genome (hg19) annotation files were

retrieved from the Tophat website, which links to the Illumina iGenome collection.

Alignments were performed with Tophat2, which produced BAM alignment files to the

human genome (hg19) (45). Tophat BAM files were used directly as input into Cuffdiff2

to generate normalized reads per kilobase of transcript per million mapped reads

(RPKM) values (table S11).

Sequencing data analyses

Sequencing analyses focused on transcripts annotated as “protein_coding.” Gene-

centric normalized RPKM values were filtered as follows: mRNAs with average RPKM <

1 were removed, remaining mRNAs were log2 (RPKM +1) transformed, and each

mRNA was then normalized by subtracting its mean log2 (RPKM+1) value. Clustering

was performed with Cluster 3.0 using average-linkage uncentered Pearson correlation,

and results were visualized with Java Treeview. Other analyses and graphs were

produced in R.

TCGA data

Normalized Affymetrix microarray data and metadata were retrieved from TCGA data

portal. Only flash frozen recurrent tumor samples were available for RNA sequencing

analysis and therefore TCGA –related subtype assignments were not able to follow

TCGA histologic guidelines regarding the degree of necrosis and tumor cell

contamination by normal cells. This was partially controlled for by analyzing multiple

samples from each subject and comparison of RNA expression to normal brain.

Molecular classification

Updated classification gene lists were provided by Roeland Verhaak (MD Anderson),

who developed the classification system and is a member of the TCGA GBM

consortium. Samples were classified by hierarchical clustering using uncentered

Pearson correlation as the distance metric. SRNS signature gene list is provided in

table S12 and classification of samples described in this study is documented in table

S13.

Retrieval and alignment of additional GBM RNA sequencing data

SRA compressed raw sequencing data were retrieved from the GEO website

(GSE59612) (23). SRA files were converted to FASTQ files and processed following the

same protocol as performed for the study samples. Normalized RPKM values were

generated from these 94 samples as well as study samples using Cuffnorm, and the

normalized RPKM values were processed as described above. mRNAs that

systematically tracked with the experimental dataset were identified from an

unsupervised hierarchical cluster (standard deviation > 0.8) and removed. The

“proliferation signature” used in Figure S12 was obtained from Sandberg et al. (46).

Methylated DNA detection and DNA fragmentation

DNA samples were digested with AluI for 3 hours at 37°C. DNA digests were processed

with the MethylMagnet mCpG DNA Fractionation kit (MM101K RiboMed

Biotechnologies) without purification.

Methylated DNA fractionation

Methylated DNA fragments were separated from unmethylated DNA using the

MethylMagnet kit (Ribomed Biotechnologies, Inc) following the manufacturer’s protocol

with the exception that a 10x binding buffer was used to minimize the dilution of the

samples. DNA samples were incubated with magnetic beads bearing a GST-MBD2-

DNA-binding-domain protein for 1 hour at room temperature with mixing at 1000 rpm in

an Eppendorf thermomixer. The supernatant fraction containing unmethylated DNA

fragments was recovered and stored at -20°C. The beads were washed as specified in

the kit manual followed by incubation at 65°C for 10 min with mixing at 1000 rpm to

elute the methylated DNA. The eluted DNA fraction (methylated fragments) was

separated from the beads with a magnet and stored at -20°C.

Coupled abscription PCR (CAP)

End-point, hot-start PCR was performed with 0.5 units Maxima Taq, Maxima hot-start

buffer (ThermoFisher), 800 µM dNTPs, 2 mM MgCl2, and 6% DMSO. Both the

methylated and unmethylated DNA fractions were amplified with PCR primers (RiboMed

Biotechnologies) targeting the AluI fragment that maps between -110 and +149 relative

to the MGMT transcription start site (Forward: APC-DN-

GCGCACCGTTTGCGACTTGG, Reverse: UR-AGCGAGGCGACCCAGACACT). One

primer in the set contained a single-stranded inactive abortive promoter cassette at its 5’

end, which was activated for abortive transcription (Abscription) when made duplex in

the course of copying the target. PCR cycling conditions were: Step 1 94°C, 2 min, Step

2, 57°C, 3 min, Step 3, 72°C, 30 sec, Step 4, repeat Steps 1-3, 2x, Step 5 88.9°C, 5

sec, Step 6, 57.7°C, 15 sec, Step 7, 72°C, 15 sec, Step 8, repeat Steps 5-7, 25x or 27x

for 29 or 31 total cycles, respectively. PCR reaction mixes with 1 µl DNA inputs below

100 genomic copies were amplified for 31 cycles. PCR reaction mixes with 1 µl DNA

inputs between 100 and 500 genomic copies were amplified for 29 cycles. The PCR

reaction volume was 10 µl.

The amplified samples were mixed with equal volumes of Abscription master mix

(RiboMed Biotechnologies) and were incubated at 40-41°C for 15 one minute cycles in

an ABI 7000 Prism qPCR cycler. Abortive transcription of the synthetic promoters

produced short RNAs (abscripts) that annealed to a quenched fluorescein labeled

molecular beacon. Extension of an annealed abscript by the CAP-DNA polymerase

opened the beacon allowing for light emission. Fluorescein fluorescence readings were

taken every minute. The rate of fluorescence increase was proportional to the number

of genomes sampled by PCR. Fluorescence signals were expressed as the slopes of

linear curves relating fluorescence increase to abscription time. Experimental samples

were calibrated in terms of genomic copies using a parallel titration of HeLa DNA.

Percent methylation was calculated as (methylated DNA copies)/(methylated DNA

copies + unmethylated DNA copies)*100. A percentage greater than 11% was

considered positive for methylation.

Supplemental Figures

Figure S1. Simplified schematic for the phase 1 trial (NCT01470794).

This trial involved dose escalation of Toca 511, an investigational non-lytic, retroviral

replicating vector, and Toca FC (an investigational extended-release version of 5-

fluorocytosine, 5-FC). Eligible subjects enrolled in the trial underwent surgical resection

of the recurrent high grade glioma. Subsequently, Toca 511 was injected into the

resection cavity wall via multiple injections to maximize coverage area using a blunt tip

needle with a side port. Virus was allowed to spread for approximately 6 weeks, and

then subjects started repeated cycles of Toca FC.

Fig. S2. Tumor-specific Toca 511 staining in re-resected tumors after multiple

cycles of Toca FC.

IHC staining with anti-CD antibody on resected tumor sections after Toca 511 delivery

and multiple cycles of Toca FC. (A) Section of normal brain where no positive staining is

observed after Toca 511 treatment. (B) Positive staining of CD protein in a tumor

section after Toca 511 treatment. CD protein is stained brown; nuclei are in blue.

Fig. S3. Evidence of potential pseudoprogression on resected tumor after Toca

511 and Toca FC treatment.

(A) H&E stained tumor section with evidence of necrosis and a hypocellular region after

treatment with Toca 511 and Toca FC. (B) Lymphocytic infiltration observed in a

cellular section from the same tumor.

Fig. S4. Comparison of subjects with glioblastoma at first or second recurrence

treated with Toca 511 and Toca FC to lomustine.

Fig. S5. Peripheral blood CD4+ T cell modulation after Toca 511 and Toca FC

dosing.

Changes in absolute CD4+ T cell counts per mL of blood were followed over time using

the CD4+ T cell counts before surgery as baseline for each subject. A significant

increase in CD4+ T cell counts was observed after week 21 (p=0.019 using Wilcoxon

Signed Rank Test).

Fig. S6. Tracking of Toca 511 DNA and RNA signal in whole blood and plasma

over time.

DNA extracted from whole blood and RNA extracted from plasma were analyzed by

qPCR and qRT-PCR, respectively, for Toca 511 polymerase gene signal. The

percentage of subjects with quantitative (%Quant.) DNA (left panel) or RNA (right panel)

Toca 511 signal was plotted as a function of time.

Fig. S7. Molecular classification of tumor samples from study subjects based on

mRNA expression.

mRNA expression patterns of sixty five study tumor samples from twenty seven

biopsies organized by hierarchical clustering of 257 mRNAs used to classify newly

diagnosed glioblastoma samples by molecular subtype. mRNAs associated with each

subtype and samples classified in each subtype are color coded: Classical (C) is black,

Mesenchymal (M) is red, Neural (N) is green, Proneural (P) is blue. Variation in mRNA

expression extends from 0.125-fold (green) to 8-fold (red) from the mean across all

samples (−2 to +2 in log2 scale).

Fig. S8. Intratumor versus intertumor heterogeneity in RNA expression profiles.

(A) Boxplot comparing intra-tumor Pearson correlations in mRNA expression vs inter-

tumor Pearson correlations in mRNA expression. The dark line is the mean, the boxes

confine the interquartile range and the whiskers are 1.5X the interquartile range.

Samples from the same biopsy tended to have a higher correlation in mRNA expression

than samples from other subjects (p = 1e-11, one-sided t-test). (B) Strip chart showing

intra-tumor Pearson correlation(s) in red and inter-tumor correlations in black for each

subject. For instance, two samples from subject 02 were profiled and thus there is one

red dot illustrating that the Pearson correlation across all mRNAs between these two

samples was 0.5, whereas the black dots indicate pairwise Pearson correlations

between each of these two samples compared to the samples from all other subjects.

Fig. S9. No relationship between variation in mRNA expression and confounding

technical factors.

Bar plot showing the coefficient of determination (R2) comparing RNA integrity number

(RIN, black), number of reads (red), and percentage of mapped reads (gray) versus the

first five principal components of the RNA sequencing dataset. None of the correlations

were significant at p < 0.05.

Fig. S10. No correlation between neural subtype and prognosis in newly

diagnosed glioblastoma.

Kaplan-Meier plot of newly diagnosed glioblastoma subjects whose tumor RNA

expression was profiled by TCGA consortium. Patients are color-coded based on TCGA

classifications, including Classical = 130 (black), Mesenchymal = 143 (red), Proneural =

91 (dark blue), Neural = 75 (green), G-CIMP = 31 (cyan) (19). G-CIMP refers to

glioblastoma tumors that are Proneural and also have glioma-CpG island methylator

phenotype, which confers a favorable prognosis (38).

Figure S11. Identification of an mRNA profile (SRNS) associated with longer

survival after Toca 511.

(A) Supervised hierarchical cluster of mRNA expression across sixty four tumor

samples from twenty six efficacy evaluable biopsies based on the relative expression of

890 mRNAs in Cluster 5, Figure 3. Samples segregated into two main groups, which

are color-coded in the dendrogram (black, non-SRNS; magenta, SRNS). (B) Bar plot

representation of the number of samples from glioblastoma tumors with SRNS (purple)

and non-SRNS (black) in subjects that survived less than twelve months (left), greater

than twelve months (middle), or not yet determined (right). (C) Bar plot representation

of the number of samples from glioblastoma tumors with SRNS (purple) and non-SRNS

(black) grouped according to tumor molecular subtype (Figure 4).

Figure S12. No correlation between SRNS and prognosis in newly diagnosed

glioblastoma.

(A) Kaplan-Meier plot of newly diagnosed glioblastoma subjects whose tumor RNA

expression was profiled by TCGA consortium grouped by SRNS (purple - 108) and not

SRNS (black - 164); p = 0.66 log-rank test. This analysis focused on patients treated

with temozolomide who were G-CIMP negative. (B) Bar plot representation of the

number of samples from TCGA newly diagnosed glioblastoma subjects with SRNS

(purple) and non-SRNS (black) grouped according to tumor molecular subtype.

Fig. S13. Study survival–related neural subtype samples likely derive from

nonenhancing regions of tumors.

(A) Unsupervised hierarchical clustering of RNA expression profiles from study subjects,

grouped by molecular subtype, and tumor samples from a recent publication, grouped

by tumor region (23) using 3544 mRNAs with the largest variation in expression. Color

coding of samples is noted to the right of the heatmap. (B) Box plot of “proliferation

signature” mRNA measurements across study tumors, grouped by molecular subtype,

and tumor samples from a recent publication, grouped by tumor region (upper left) or

box plots showing the relative expression of SOX2, MYC, MKI67, ANXA2, and

CDKN2D, respectively.

Fig. S14. Negative correlation between SPOC1 expression and subject survival

time with Toca 511 and Toca FC therapy.

Scatterplot showing relative tumor SPOC1 mRNA expression (x-axis) and subject

survival (y-axis). The error bars are standard error of mean across different tumor

samples from the same subject. Only subjects with glioblastoma are included. The red

line is a least squares fit of mRNA expression vs log10[survival]. Pearson correlation is -

0.71.

Fig. S15. MGMT promoter methylation.

(A) MGMT methylation was determined by measuring the % MGMT promoter

methylation across two regions of the promoter. (B) Any region with >11% is

considered methylation positive. No survival benefit was seen for MGMT methylation

status when subjects underwent Toca 511 and Toca FC therapy.

.

Supplemental Tables

Table S1. Toca 511 and Toca FC: Baseline demographic and clinical characteristics.

Total (N=45)

N (%) Characteristics Age (years) Median 56 Min, Max 24, 72

Sex Male 35 (77.8) Female 10 (22.2)

Race Asian 1 (2.2) Black or African American 1 (2.2) White 43 (95.6)

Baseline Karnofsky performance score 70 4 (8.9) 80 7 (15.6) 90 26 (57.8) 100 8 (17.8)

Initial tumor histology Glioblastoma* 37 (82.2) Anaplastic astrocytoma 5 (11.1) Other glioma** 3 (6.7) Type of primary surgical procedure Gross total resection 33 (73.3) Subtotal resection 12 (26.7) Number of recurrences including current recurrence 1 23 (51.1) 2 10 (22.3) 3 or greater 12 (26.6)

*includes gliosarcoma

**anaplastic ganglioglioma (1), anaplastic oligoastrocytoma (1), anaplastic

oligodendroglioma (1)

Table S2. Additional baseline demographic and clinical characteristics.

Total (N=45)

N (%)

Months since initial diagnosis of HGG

Median 15.5

Min, Max 5.1, 118.6

Total radiation delivered (Gy)

Median 60.0

Min, Max 42.5, 95.0

Number of courses of maintenance

temozolomide

Median 6.0

Min, Max 1.0, 48.0

Prior therapy*

Investigational agent 21 (46.7)

Temozolomide 16 (35.6)

Bevacizumab 8 (17.8)

Lomustine 7 (15.6)

Novocure TTF 5 (11.1)

Gliadel wafer 4 (8.9)

Carboplatin 3 (6.7)

Dendritic cell vaccine 2 (4.4)

Etoposide 2 (4.4)

Radiation 2 (4.4)

* In 2 or more subjects

Table S3. Dosing cohorts.

φ For subjects enrolled in Cohorts 6a and 6b, an additional 6.4 mL Toca 511 was mixed with Surgifoam powder and packed in the resection cavity.

† Toca 511 dosage varies with the titer of each manufactured lot.

Table S4. Detection of Toca 511 in re-resected tumors.

Cohort Dose TU/g Tumor samples positive/total (%)

for vector DNA

Highest DNA value (copies/µg)

Tumor samples positive/ total (%)

for RNA

Highest RNA value (copies/µg)

1 9.5 x 103 10/11 (91) 3728 6/11 (55) 16,273

1 9.5 x 103 4/7 (57) <100 (BLOQ) 0 ND

3 1 x 105 3/5 (60) 479 NT NA

3 1 x 105 47/47 <100 (BLOQ) NT NA

3 1 x 105 0/1 (0) ND 0 ND

5 1 x 106 1/4 (25) <100 (BLOQ) 2/4 (50) <250 (BLOQ)

5 1 x 106 0/7 (0) ND 0 ND

5 1 x 106 5/7 (71) 259 1/7 (14) 20,500

6 3.2 x 106 2/2 (100) <100 (BLOQ) 0 ND

Toca 511 transducing units per gram of brain (TU/g), Below Limit Of Quantification for the assay (BLOQ),

Not available (NA), Not detected (ND), Not tested (NT)

Cohort # N Dosing cohort in TU Toca 511 (mL) Toca FC dose (mg/kg/day)

1 3 1.4 x 107 0.8

130 x 8 days

2 4 3.8 x 107 2.0

3 6 1.5 x 108 0.9

4 6 4.8 x 108 1.1

5a 7

1.5 x 109

3.1

5b 4 3.1 135 x 14 days

5c 3 3.1 170 x 7 days

6a 4 4.8 x 109 3.1 + 6.4 φ

6b 5 220 x 7 days

7a 3 1.2 x109† 3.6

Table S5. Best overall response in the efficacy evaluable population.

Total (N=43)

Cohorts 4-7a (N=30)

n (%) n (%)

Best overall tumor response

Complete response (CR) 2 ( 4.7) 2 ( 6.7)

Partial response (PR) 2 ( 4.7) 2 ( 6.7)

Stable disease (SD) 8 (18.6) 7 (23.3)

Progressive disease 31 (72.0) 19 (63.3)

Overall response rate (CR+PR) 4 (9.3) 4 (13.3)

Clinical benefit rate (CR+PR+SD) 12 (27.9) 11 (36.7)

Table S6. Adverse events and serious adverse events related to Toca 511 and Toca FC.

Toca 511 N=45

N (%)

Toca FC N=43

N (%)

All Grades Grade ≥ 3 All Grades Grade ≥ 3

Subjects reporting at least one adverse event 45 (100.0) 28 (62.2) Subjects reporting at least one AE 42 (97.7) 19 (44.2)

Related treatment emergent adverse events 8 (17.8) 1 (2.2) Related treatment emergent adverse events 16 (37.2) 0 (0.0)

General disorders and administration site conditions 6 (13.3) 1 (2.2) Gastrointestinal disorders 9 (20.9) 0 (0.0)

Nervous system disorders 2 (4.4) 0 (0.0) General disorders and administration site conditions 9 (20.9) 0 (0.0)

Psychiatric disorders 2 (4.4) 0 (0.0) Nervous system disorders 3 (7.0) 0 (0.0)

Eye disorders 1 (2.2) 0 (0.0) Metabolism and nutrition disorders 1 (2.3) 0 (0.0)

Gastrointestinal disorders 1 (2.2) 0 (0.0)

Related serious adverse event 2 (4.4) 0 (0.0) Related serious adverse event 0 (0.0) 0 (0.0)

General disorders and administration site conditions 1 (2.2) 0 (0.0) Adverse event resulting in Toca FC discontinuation 1 (2.2)* 0 (0.0)

Asthenia 1 (2.2) 0 (0.0)

Pyrexia 1 (2.2) 0 (0.0)

Nervous system disorders 1 (2.2) 0 (0.0)

Convulsion 1 (2.2) 0 (0.0) *Grade 2 erythematous pruritus and Grade 2 facial

swelling probably related to Toca FC

Table S7. Related adverse events: Toca 511 and Toca FC compared to lomustine.

Toca 511 & Toca FC

N=27

n(%)

Lomustine

N=84

n(%)

Preferred term Grade 2 Grade 3 to 4 Grade 2 Grade 3 to 4

Nonhematologic toxicity

Fatigue 2 (7.4%) 0 4 (4.8%) 0

Asthenia 0 1 (3.7%) 0 0

Decreased appetite 1 (3.7%) 0 0 0

Disorientation 1 (3.7%) 0 0 0

Dysgeusia 1 (3.7%) 0 0 0

Nausea 1 (3.7%) 0 2 (2.4%) 0

Pyrexia 1 (3.7%) 0 1 (1.2%) 0

Wound infection 1 (3.7%) 0 0 0

Alanine aminotransferase increased 0 0 1 (1.2%) 1 (1.2%)

Anal ulcer 0 0 0 1 (1.2%)

Candidiasis 0 0 1 (1.2%) 0

Cellulitis 0 0 0 1 (1.2%)

Clostridium difficile colitis 0 0 0 1 (1.2%)

Cushing's syndrome 0 0 1 (1.2%) 0

Diarrhea 0 0 1 (1.2%) 0

Dyspnea 0 0 0 1 (1.2%)

Epistaxis 0 0 1 (1.2%) 0

Febrile neutropenia 0 0 0 1 (1.2%)

Gamma-glutamyltransferase increased 0 0 0 1 (1.2%)

Herpes ophthalmic 0 0 1 (1.2%) 0

Hypoxia 0 0 0 1 (1.2%)

Localized infection 0 0 0 1 (1.2%)

Toca 511 & Toca FC

N=27

n(%)

Lomustine

N=84

n(%)

Preferred term Grade 2 Grade 3 to 4 Grade 2 Grade 3 to 4

Malaise 0 0 1 (1.2%) 0

Edema peripheral 0 0 2 (2.4%) 0

Pneumonia 0 0 0 1 (1.2%)

Pulmonary function test decreased 0 0 1 (1.2%) 0

Purpura 0 0 1 (1.2%) 0

Stomatitis 0 0 1 (1.2%) 0

Transaminases increased 0 0 0 1 (1.2%)

Vomiting 0 0 1 (1.2%) 0

Hematologic toxicity

Anemia 0 0 1 (1.2%) 2 (2.4%)

Hemoglobin decreased 0 0 2 (2.4%) 0

Leukopenia 0 0 3 (3.6%) 4 (4.8%)

Lymphopenia 0 0 2 (2.4%) 0

Neutropenia 0 0 4 (4.8%) 11 (13.1%)

Neutrophil count decreased 0 0 0 6 (7.1%)

Platelet count decreased 0 0 0 1 (1.2%)

Thrombocytopenia 0 0 9 (10.7%) 20 (23.8%)

White blood cell count decreased 0 0 1 (1.2%) 2 (2.4%)

Table S8. Adverse events regardless of attribution: Toca 511 and Toca FC compared to lomustine.

Toca 511 &Toca FC

N=27

n (%)

Lomustine

N=84

n (%)

Preferred term Grade 2 Grade 3 to 4 Grade 2 Grade 3 to 4

Nonhematologic toxicity

Fatigue 10 (37.0%) 3 (11.1%) 7 (8.3%) 1 (1.2%)

Depression 7 (25.9%) 0 2 (2.4%) 2 (2.4%)

Hemiparesis 7 (25.9%) 4 (14.8%) 2 (2.4%) 2 (2.4%)

Hyperglycemia 7 (25.9%) 3 (11.1%) 4 (4.8%) 1 (1.2%)

Incision site pain 7 (25.9%) 0 0 0

Procedural pain 5 (18.5%) 0 1 (1.2%) 0

Anxiety 4 (14.8%) 0 1 (1.2%) 0

Cognitive disorder 4 (14.8%) 0 2 (2.4%) 1 (1.2%)

Headache 4 (14.8%) 3 (11.1%) 8 (9.5%) 3 (3.6%)

Blood glucose increased 3 (11.1%) 0 0 0

Gait disturbance 3 (11.1%) 2 (7.4%) 0 0

Memory impairment 3 (11.1%) 0 2 (2.4%) 0

Nausea 3 (11.1%) 0 4 (4.8%) 0

Aphasia 2 (7.4%) 1 (3.7%) 1 (1.2%) 1 (1.2%)

Arthralgia 2 (7.4%) 0 1 (1.2%) 1 (1.2%)

Asthenia 2 (7.4%) 2 (7.4%) 0 0

Constipation 2 (7.4%) 0 0 0

Convulsion 2 (7.4%) 0 9 (10.7%) 2 (2.4%)

Deep vein thrombosis 2 (7.4%) 1 (3.7%) 0 3 (3.6%)

Fall 2 (7.4%) 0 1 (1.2%) 1 (1.2%)

Fine motor delay 2 (7.4%) 0 0 0

Toca 511 &Toca FC

N=27

n (%)

Lomustine

N=84

n (%)

Preferred term Grade 2 Grade 3 to 4 Grade 2 Grade 3 to 4

Hemianopia homonymous 2 (7.4%) 0 0 0

Hydrocephalus 1 (3.7%) 2 (7.4%) 0 1 (1.2%)

Hypotension 2 (7.4%) 0 0 0

Insomnia 2 (7.4%) 0 2 (2.4%) 0

Mental status changes 0 2 (7.4%) 0 1 (1.2%)

Migraine 2 (7.4%) 0 0 0

Muscular weakness 2 (7.4%) 0 3 (3.6%) 2 (2.4%)

Pulmonary embolism 0 2 (7.4%) 0 2 (2.4%)

Somnolence 2 (7.4%) 0 1 (1.2%) 0

Tremor 2 (7.4%) 0 1 (1.2%) 0

Upper respiratory tract infection 2 (7.4%) 0 2 (2.4%) 0

Urinary incontinence 2 (7.4%) 0 5 (6.0%) 1 (1.2%)

Urinary tract infection 2 (7.4%) 0 0 1 (1.2%)

VIIth nerve paralysis 2 (7.4%) 0 0 0

Vasogenic cerebral edema 2 (7.4%) 0 0 0

Vomiting 2 (7.4%) 0 4 (4.8%) 0

Abdominal pain upper 1 (3.7%) 0 1 (1.2%) 0

Abscess soft tissue 1 (3.7%) 0 0 0

Abulia 1 (3.7%) 0 0 0

Agitation 0 1 (3.7%) 0 1 (1.2%)

Alanine aminotransferase increased 1 (3.7%) 0 2 (2.4%) 1 (1.2%)

Amnesia 1 (3.7%) 0 0 1 (1.2%)

Anesthesia 1 (3.7%) 0 0 0

Ataxia 1 (3.7%) 0 0 0

Balance disorder 1 (3.7%) 0 0 0

Toca 511 &Toca FC

N=27

n (%)

Lomustine

N=84

n (%)

Preferred term Grade 2 Grade 3 to 4 Grade 2 Grade 3 to 4

Benign prostatic hyperplasia 1 (3.7%) 0 0 0

Blindness unilateral 1 (3.7%) 0 0 0

Blood phosphorus decreased 0 1 (3.7%) 0 0

Brain edema 1 (3.7%) 1 (3.7%) 1 (1.2%) 2 (2.4%)

Candidiasis 1 (3.7%) 0 2 (2.4%) 0

Cerebral cyst 0 1 (3.7%) 0 0

Chest pain 1 (3.7%) 0 0 0

Clonus 1 (3.7%) 0 0 0

Confusional state 1 (3.7%) 0 4 (4.8%) 2 (2.4%)

Deafness bilateral 1 (3.7%) 0 0 0

Decreased appetite 1 (3.7%) 0 0 0

Dehydration 1 (3.7%) 0 0 0

Diabetes mellitus 1 (3.7%) 0 0 0

Diarrhea 1 (3.7%) 0 1 (1.2%) 0

Disorientation 1 (3.7%) 0 1 (1.2%) 0

Dysarthria 1 (3.7%) 0 0 1 (1.2%)

Dysgeusia 1 (3.7%) 0 0 0

Dyspepsia 1 (3.7%) 0 0 0

Dyspnea 0 1 (3.7%) 0 3 (3.6%)

Ecchymosis 1 (3.7%) 0 0 0

Executive dysfunction 1 (3.7%) 0 0 0

Facial paresis 1 (3.7%) 0 0 0

Flushing 1 (3.7%) 0 0 0

Gamma-glutamyltransferase increased 1 (3.7%) 1 (3.7%) 0 2 (2.4%)

Grand mal convulsion 0 1 (3.7%) 0 0

Toca 511 &Toca FC

N=27

n (%)

Lomustine

N=84

n (%)

Preferred term Grade 2 Grade 3 to 4 Grade 2 Grade 3 to 4

Hemorrhage intracranial 1 (3.7%) 0 0 0

Hemorrhoids 1 (3.7%) 0 0 0

Hemiplegia 0 1 (3.7%) 0 0

Hemisensory neglect 1 (3.7%) 0 0 0

Hepatic enzyme increased 0 1 (3.7%) 0 0

Herpes zoster 1 (3.7%) 0 0 0

Hypertension 1 (3.7%) 0 0 0

Hypoesthesia 1 (3.7%) 0 1 (1.2%) 0

Hypocalcaemia 1 (3.7%) 0 0 0

Hypokalemia 1 (3.7%) 0 1 (1.2%) 1 (1.2%)

Hyponatremia 0 1 (3.7%) 0 0

Impulsive behavior 1 (3.7%) 0 0 0

Intracranial hypotension 0 1 (3.7%) 0 0

Lacunar infarction 1 (3.7%) 0 0 0

Lethargy 1 (3.7%) 0 0 0

Lobar pneumonia 1 (3.7%) 0 0 0

Loss of proprioception 1 (3.7%) 0 0 0

Malaise 1 (3.7%) 0 1 (1.2%) 0

Medication error 1 (3.7%) 0 0 0

Meningitis 0 1 (3.7%) 0 0

Monoplegia 0 1 (3.7%) 0 0

Musculoskeletal discomfort 1 (3.7%) 0 0 0

Myalgia 1 (3.7%) 0 0 0

Myopathy 1 (3.7%) 0 0 0

Nephrolithiasis 0 1 (3.7%) 0 0

Toca 511 &Toca FC

N=27

n (%)

Lomustine

N=84

n (%)

Preferred term Grade 2 Grade 3 to 4 Grade 2 Grade 3 to 4

Neurologic neglect syndrome 1 (3.7%) 1 (3.7%) 0 0

Neuropathy peripheral 1 (3.7%) 0 0 0

Edema peripheral 1 (3.7%) 0 3 (3.6%) 0

Oral candidiasis 1 (3.7%) 0 0 0

Oropharyngeal candidiasis 1 (3.7%) 0 0 0

Pain 1 (3.7%) 0 0 0

Pain in extremity 1 (3.7%) 0 0 0

Pain in jaw 1 (3.7%) 0 0 0

Pain of skin 1 (3.7%) 0 1 (1.2%) 0

Parkinsonism 0 1 (3.7%) 0 0

Peroneal nerve palsy 1 (3.7%) 0 0 1 (1.2%)

Pneumocephalus 1 (3.7%) 0 0 0

Pneumonia 1 (3.7%) 1 (3.7%) 0 2 (2.4%)

Procedural headache 1 (3.7%) 0 0 0

Purulent discharge 1 (3.7%) 0 0 0

Pyrexia 1 (3.7%) 0 1 (1.2%) 0

Rash 1 (3.7%) 0 0 0

Sensation of heaviness 1 (3.7%) 0 0 0

Shunt infection 0 1 (3.7%) 0 0

Skin laceration 1 (3.7%) 0 0 0

Stomatitis 1 (3.7%) 0 1 (1.2%) 0

Subdural hematoma 1 (3.7%) 0 0 0

Subdural hygroma 1 (3.7%) 0 0 0

Tachycardia 1 (3.7%) 0 0 0

Thrombophlebitis 1 (3.7%) 0 0 0

Toca 511 &Toca FC

N=27

n (%)

Lomustine

N=84

n (%)

Preferred term Grade 2 Grade 3 to 4 Grade 2 Grade 3 to 4

Tooth fracture 1 (3.7%) 0 0 0

Upper motor neuron lesion 1 (3.7%) 0 0 0

Urinary retention 1 (3.7%) 0 0 0

Vision blurred 1 (3.7%) 0 0 0

Visual field defect 1 (3.7%) 0 0 0

Weight decreased 1 (3.7%) 0 0 0

Wound infection 1 (3.7%) 0 1 (1.2%) 0

Wound infection staphylococcal 0 1 (3.7%) 0 0

Abdominal pain 0 0 1 (1.2%) 0

Altered visual depth perception 0 0 1 (1.2%) 0

Anal ulcer 0 0 0 1 (1.2%)

Arrhythmia supraventricular 0 0 1 (1.2%) 0

Back pain 0 0 1 (1.2%) 0

Blindness 0 0 0 1 (1.2%)

Blindness transient 0 0 1 (1.2%) 0

Bronchitis 0 0 1 (1.2%) 0

Cellulitis 0 0 0 3 (3.6%)

Clostridium difficile colitis 0 0 0 1 (1.2%)

Condition aggravated 0 0 1 (1.2%) 0

Coordination abnormal 0 0 1 (1.2%) 0

Cranial neuropathy 0 0 1 (1.2%) 0

Cushing's syndrome 0 0 1 (1.2%) 0

Diplopia 0 0 1 (1.2%) 0

Dizziness 0 0 1 (1.2%) 1 (1.2%)

Dysphagia 0 0 1 (1.2%) 0

Toca 511 &Toca FC

N=27

n (%)

Lomustine

N=84

n (%)

Preferred term Grade 2 Grade 3 to 4 Grade 2 Grade 3 to 4

Dysphemia 0 0 1 (1.2%) 0

Ear pain 0 0 1 (1.2%) 0

Epilepsy 0 0 0 1 (1.2%)

Epistaxis 0 0 1 (1.2%) 0

Erysipelas 0 0 0 1 (1.2%)

Euphoric mood 0 0 1 (1.2%) 0

Eye pain 0 0 1 (1.2%) 0

Fecal incontinence 0 0 1 (1.2%) 1 (1.2%)

Febrile neutropenia 0 0 0 1 (1.2%)

Femoral neck fracture 0 0 0 1 (1.2%)

General physical health deterioration 0 0 1 (1.2%) 1 (1.2%)

Glucose tolerance impaired 0 0 0 1 (1.2%)

Hematoma 0 0 0 1 (1.2%)

Herpes ophthalmic 0 0 1 (1.2%) 0

Hiccups 0 0 1 (1.2%) 0

Hypoxia 0 0 0 1 (1.2%)

Immobile 0 0 0 1 (1.2%)

Incontinence 0 0 1 (1.2%) 0

Increased appetite 0 0 1 (1.2%) 0

Infection 0 0 1 (1.2%) 0

Intracranial venous sinus thrombosis 0 0 1 (1.2%) 0

Keratoconjunctivitis sicca 0 0 1 (1.2%) 0

Localized infection 0 0 0 1 (1.2%)

Loss of consciousness 0 0 0 1 (1.2%)

Lower respiratory tract infection 0 0 0 1 (1.2%)

Toca 511 &Toca FC

N=27

n (%)

Lomustine

N=84

n (%)

Preferred term Grade 2 Grade 3 to 4 Grade 2 Grade 3 to 4

Motor dysfunction 0 0 1 (1.2%) 0

Neurological symptom 0 0 0 1 (1.2%)

Optic neuropathy 0 0 1 (1.2%) 0

Paresthesia 0 0 1 (1.2%) 0

Partial seizures 0 0 1 (1.2%) 0

Peripheral motor neuropathy 0 0 1 (1.2%) 0

Pharyngolaryngeal pain 0 0 1 (1.2%) 0

Pollakiuria 0 0 0 1 (1.2%)

Post-traumatic pain 0 0 0 1 (1.2%)

Pulmonary function test decreased 0 0 1 (1.2%) 0

Purpura 0 0 1 (1.2%) 0

Pyramidal tract syndrome 0 0 0 1 (1.2%)

Radiation injury 0 0 0 1 (1.2%)

Restrictive pulmonary disease 0 0 1 (1.2%) 0

Speech disorder 0 0 3 (3.6%) 1 (1.2%)

Tinnitus 0 0 1 (1.2%) 0

Transaminases increased 0 0 0 1 (1.2%)

Wheezing 0 0 1 (1.2%) 0

Hematologic toxicity

Lymphopenia 0 1 (3.7%) 3 (3.6%) 0

Anemia 0 0 1 (1.2%) 2 (2.4%)

Hemoglobin decreased 0 0 2 (2.4%) 0

Leukopenia 0 0 4 (4.8%) 5 (6.0%)

Neutropenia 0 0 4 (4.8%) 12 (14.3%)

Toca 511 &Toca FC

N=27

n (%)

Lomustine

N=84

n (%)

Preferred term Grade 2 Grade 3 to 4 Grade 2 Grade 3 to 4

Neutrophil count decreased 0 0 0 6 (7.1%)

Platelet count decreased 0 0 0 1 (1.2%)

Thrombocytopenia 0 0 9 (10.7%) 20 (23.8%)

White blood cell count decreased 0 0 1 (1.2%) 2 (2.4%)

Table S9. Summary of multivariate analysis for survival (TCGA neural signature).

Variable Reference

group

P-value Hazard

ratio

95% hazard ratio

confidence

limits

Subtype Neural 0.0097 0.108 0.020 0.583

Recurrences Recurrences<=2 0.0226 0.262 0.083 0.828

HGG grade Grade 3 0.0021 0.044 0.006 0.321

Table S10. Summary of multivariate analysis for survival (SRNS signature).

Variable Reference group P-value Hazard

ratio

95% hazard

ratio

confidence

limits

Subtype SRNS 0.0032 0.109 0.025 0.476

Recurrences Recurrences<=2 0.0023 0.152 0.045 0.509

HGG grade Grade 3 0.0011 0.047 0.008 0.294