Temozolomide analogs with improved brain/plasma ratios – Exploring the possibility of enhancing the therapeutic index of temozolomide

Article abstract of DOI:10.1016/j.bmcl.2016.08.064

Temozolomide is a chemotherapeutic agent that is used in the treatment of glioblastoma and other malignant gliomas. It acts through DNA alkylation, but treatment is limited by its systemic toxicity and neutralization of DNA alkylation by upregulation of the O⁶-methylguanine-DNA methyltransferase gene. Both of these limiting factors can be addressed by achieving higher concentrations of TMZ in the brain. Our research has led to the discovery of new analogs of temozolomide with improved brain:plasma ratios when dosed in vivo in rats. These compounds are imidazotetrazine analogs, expected to act through the same mechanism as temozolomide. With reduced systemic exposure, these new agents have the potential to improve efficacy and therapeutic index in the treatment of glioblastoma.

Full text of DOI:10.1016/j.bmcl.2016.08.064

Accepted Manuscript
Temozolomide analogs with improved brain/plasma ratios – exploring the pos-
sibility of enhancing the therapeutic index of temozolomide
Roopa Rai, Monali Banerjee, Darren H. Wong, Emma McCullagh, Ashu Gupta,
Shailendra Tripathi, Eduardo Riquelme, Ramnivas Jangir, Shyamraj Yadav,
Mohd. Raja, Pankaj Melkani, Vikas Dixit, Umesh Patil, Ritesh Shrivastava,
Sandip Middya, Felipe Olivares, Javier Guerrero, Arjun Surya, Son M. Pham,
Sebastián Bernales, Andrew A. Protter, David T. Hung, Sarvajit Chakravarty
PII:
S0960-894X(16)30897-6
DOI:
http://dx.doi.org/10.1016/j.bmcl.2016.08.064
BMCL 24190
Reference:
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
21 July 2016
17 August 2016
19 August 2016
Please cite this article as: Rai, R., Banerjee, M., Wong, D.H., McCullagh, E., Gupta, A., Tripathi, S., Riquelme, E.,
Jangir, R., Yadav, S., Raja, Mohd., Melkani, P., Dixit, V., Patil, U., Shrivastava, R., Middya, S., Olivares, F.,
Guerrero, J., Surya, A., Pham, S.M., Bernales, S., Protter, A.A., Hung, D.T., Chakravarty, S., Temozolomide analogs
with improved brain/plasma ratios – exploring the possibility of enhancing the therapeutic index of temozolomide,
Bioorganic & Medicinal Chemistry Letters (2016), doi: http://dx.doi.org/10.1016/j.bmcl.2016.08.064
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Bioorganic & Medicinal Chemistry Letters
journal homepage: www.els evier.com
Temozolomide analogs with improved brain/plasma ratios – exploring
the possibility of enhancing the therapeutic index of temozolomide
a*
Roopa Rai , Monali Banerjee , Darren H. Wong , Emma McCullagh , Ashu Gupta , Shailendra Tripathi ,
b
c
a*
d
d
e
d
d
d
d
d
Eduardo Riquelme , Ramnivas Jangir , Shyamraj Yadav , Mohd. Raja , Pankaj Melkani , Vikas Dixit ,
d
Umesh Patil , Ritesh Shrivastava , Sandip Middya , Felipe Olivares , Javier Guerrero , Arjun Surya , Son
a
M. Pham , Sebastián Bernales , Andrew A. Protter , David T. Hung and Sarvajit Chakravarty
b
b
e
e
b
a
a
a
a
a
b
c
th
Medivation, 525 Market Street, 36 Floor, San Francisco, CA 94105 USA
Curadev Pvt. Ltd., B-87, Sector 83, Noida, Uttar Pradesh, 201305 India
Concert Pharmaceuticals, 99 Hayden Ave, Suite 500, Lexington, MA 02421
d
Integral BioScience Pvt. Ltd., C-64, Hosiery Complex Phase II Extension, Noida Uttar Pradesh, 201306 India
Fundación Ciencia y Vida, Avenida Zañartu 1482, Ñuñoa, Santiago, Chile 7780272
e
ARTICLE INFO
ABSTRACT
Article history:
Received
Temozolomide is a chemotherapeutic agent that is used in the treatment of glioblastoma and
other malignant gliomas. It acts through DNA alkylation, but treatment is limited by its systemic
6
Revised
toxicity and neutralization of DNA alkylation by upregulation of the O -methylguanine-DNA
Accepted
methyltransferase gene. Both of these limiting factors can be addressed by achieving higher
concentrations of TMZ in the brain. Our research has led to the discovery of new analogs of
temozolomide with improved brain:plasma ratios when dosed in vivo in rats. These compounds
are imidazotetrazine analogs, expected to act through the same mechanism as temozolomide.
With reduced systemic exposure, these new agents have the potential to improve efficacy and
therapeutic index in the treatment of glioblastoma.
Available online
Keywords:
Temozolomide
Glioblastoma
2
009 Elsevier Ltd. All rights reserved.
Chemotheraphy
6
O -methyl-2’-deoxyguanosine (O -MedG)
6
2
The standard of care for glioblastoma multiforme (GBM)
is debulking surgery followed by radiation and alkylating
chemotherapy, frequently temozolomide (TMZ). Patients
are treated with a standard regimen of 150 or 200
TMZ treatment. Therefore, an ability to dose higher would
mitigate the issue of MGMT upregulation.
TMZ is a prodrug that harbors an imidazotetrazine which
is stable at the acidic pH of the stomach. In patients, it is
readily absorbed with 100% oral bioavailability within 1-2
hours of administration.
alkaline environment of the blood and tissues causes
spontaneous hydrolysis to form the active metabolite MTIC
2
mg/m /day for five consecutive days, to be repeated every
four weeks. As detailed in a recent review this treatment is
1
3, 4
Once in circulation, the slightly
palliative rather than curative, with TMZ providing an
overall survival increase of just 2.5 months over radiation
alone. The median overall survival in GBM is only 14.6
months and a large percentage of patients have negligible
[
3-methyl-(triazen-1-yl) imidazole-4-carboxamide]; MTIC
1
rapidly breaks down to form the reactive methyldiazonium
ion (diazomethane) that alkylates DNA. 5-Aminoimidazole-
4-carboxamide (AIC) is formed as a side product. This is
benefit from TMZ treatment.
Known toxicities from TMZ treatment include
thrombocytopenia, lymphopenia and neutropenia. These
1
6
illustrated in Figure 1 where O -Methyl-2’-deoxyguanosine
(O -MedG) is highlighted as the end-product since it is the
6
dose-limiting toxicities prevent dose escalation beyond a
certain point, limiting a potentially curative treatment. The
development of secondary or acquired resistance, whereby
a patient fails to respond following initial response, may be
related to this limitation. Both primary and secondary
resistance to TMZ treatment is most often due to
overexpression or upregulation of the DNA repair enzyme
6
most cytotoxic. Interestingly, O methylation of guanine
represents only 5% of methylated adducts following TMZ
administration. N methyl guanine represents 70% and N
4
7
1
3
and N methyl adenine and N methyl cytosine together
3
7
form 25% of adducts. N and N -methyl purines are rapidly
3
4
repaired by base excision repair.
6
O -methylguanine-DNA methyltransferase (MGMT), which
effectively mitigates the cytotoxicity that would result from

PK Parameters
Molecule
AUCbrain
AUCplasma
AUCbrainTMZ/
AUCbrainMol
1.0
B/P ratio
(ng·h/g)
(ng·h/ml)
16853.70
959.21
TMZ
5660.37
638.33
750.18
0.34
0.67
0.75
0.82
1.36
3.79
10.42
1
2
3
4
5
6
8.9
7.5
998.19
4005.62
525.05
4892.32
386.30
1.4
Figure 1. Mechanism of action of TMZ, illustrating the release of
10.8
2.3
diazomethane, which alkylates DNA.
2467.21
11183.14
650.33
1073.59
0.5
The discovery of TMZ represented an improvement over
agents such as dacarbazine. Both dacarbazine and TMZ
work through the active metabolite MTIC, but dacarbazine
5
requires metabolic dealkylation while TMZ generates
MTIC through spontaneous hydrolysis in basic media. The
5
small molecular weight of TMZ facilitates penetration into
the central nervous system (CNS). In humans,3, 6the
AUCCSF/AUCplasma has been reported to be 20-29%.
In
rats, the AUCbrain/AUCplasma (B/P ratio) has been measured
at 35-39% with TMZ being widely distributed in tissues as
well. We hypothesized that if TMZ levels could be focused
7
primarily to the brain, with limited systemic exposure, such
a profile would attenuate toxicities, leading to a possible
increase in the maximum tolerated dose (MTD). A higher
dose of TMZ with reduced side-effects can mitigate against
adaptive resistance due to MGMT upregulation.
Figure 2. Average drug levels in the plasma and brains of rats dosed with 10
mg/kg TMZ. Rats (n=3 per time point) were dosed IV with TMZ and brain
and plasma samples were collected at 30, 90 and 180 min post dose. Levels of
TMZ were analyzed by LC-MS/MS. Values shown in the graph are the
average TMZ levels in plasma (blue) and brain (red). Error bars represent the
standard deviation of the three replicates.
8
-10
Our experience with β- and γ-carbolines
had led to
the knowledge that compounds that embody these
scaffolds have inherently superior brain penetration.
Additionally, such scaffolds have been used to explore anti-
We next monitored the activity of TMZ in vitro in the U87
MG human glioblastoma cell line by quantifying the amount
11
12, 13
6
of O -MedG present after incubation with the compound.
Alzheimer’s agents as well as anti-psychotic agents
,
indications which necessitate good brain penetration. We
proposed the use of such privileged structures as “carrier
agents” to target TMZ to the brain. Our goal was to
synthesize and test compounds with increased brain
concentration of the imidazotetrazine analogs, with reduced
plasma exposure compared to TMZ.
Cells were incubated with 10 or 100 µM TMZ for 3 h at 37
°C and then washed and lysed with a commercial buffer
(ThermoScientific Cat # K0721). Genomic DNA was
extracted using an equal volume of isopropyl alcohol. DNA
was washed with isopropyl alcohol, dried and hydrolyzed.
To hydrolyze the DNA, 100 µg of genomic DNA was
incubated at 37 °C for 30 min in ddH
buffer (50 mM ammonium acetate, 0.2 nM ZnCl
0 µl nuclease P1 (0.4 units/µl) and 8 µl alkaline
2
O (2 mg/mL), 50 µl
Before embarking on the synthesis and screening of our
proposed molecules, we used TMZ to establish protocols
for in vivo and in vitro monitoring of compound levels and
activity. We measured levels of TMZ in brain and plasma
and demonstrated its efficacy. Healthy male Sprague
Dawley rats (n = 3) were dosed intravenously with 10
mg/kg of TMZ and drug levels in the brain and plasma were
analyzed by LC-MS/MS at three time points (30, 90 and
2
, pH 5.3),
1
phosphatase (1 unit/µl in 10 mM Tris pH 7.4). An internal
standard, 15 ppb [H]O -MedG, was added to hydrolyzed
DNA and the mixture was centrifuged for 15 min at 15,000
6
g at 4 °C and analyzed by LC-MS/MS. Incubation of cells
with TMZ showed a dose-dependent increase in O -MedG
6
(Figure 3) indicating that TMZ hydrolyzes in vitro and the
released diazomethane methylates guanine residues in cell
180 min post dose). While TMZ was present in the brain
and plasma, concentrations in the brain were much lower
than the plasma (Figure 2). The B/P ratio was 0.34 (Table
1
, Supplementary Table 1), matching previously reported
7
levels and suggesting that TMZ does not efficiently pass
through the blood brain barrier (BBB).
Table 1. Mean PK parameters from the 3-point screening rat
PK studies. Rats (n=3 per time point) were dosed with 10
mg/kg of compound and brain and plasma samples were
collected 30, 90 and 180 min post dose and analyzed by LC-
MS/MS for drug levels. The AUC was calculated by summing
the area under the three point curve created by connecting the
culture.
30, 90 and 180 time points with straight lines as shown in
Figure 2. The B/P ratio was calculated by dividing the
Figure 3. Treatment of U87 MG glioblastoma cells with TMZ results in
6
dose-dependent methylation of to form O -MedG. U87 MG cells were
AUCbrain by the AUCplasma
.
cultured and treated with 10 or 100 µM TMZ for 3 h. Columns represent the
6
mean (n=3) O -MedG levels and error bars display the standard deviation.

7
We also monitored for N methyl-2’-deoxyguanosine (N -
7
the 2-position. As with TMZ, these compounds were
administered IV to rats and compound levels were
monitored at the previously specified time points (30, 90
and 180 min). Compound 1 showed a B/P ratio of 0.67
(Table 1), which was an improvement over TMZ (B/P ratio
of 0.34). A comparison of drug concentration in the brain
revealed that at an identical dose, TMZ had 9-fold higher
amounts of the drug in the brain than 1 (Table 1,
Supplementary Table 1). Compound 2 had a B/P ratio of
MedG). We found that treatment of U87 MG cells with TMZ
resulted in methylation at N , and the extent of N -MedG
7
7
6
formation was more abundant than O -MedG (data not
4
shown) as reported in the literature. However, O -MedG
6
monitoring is more relevant in the context of TMZ
treatment, since the formation of N -MedG, while more
abundant, is rapidly repaired by base excision repair.
7
4
The primary premise of our hypothesis required our new
imidazotetrazine analogs to have improved brain
penetration relative to TMZ and to efficiently methylate
DNA. We, therefore, decided to use these two assays to
screen new compounds for improved B/P ratios relative to
TMZ and for efficacy in the U87 MG cellular assay. This
would allow us to select a compound for further testing in
vivo.
0
.75 but absolute concentration in the brain was low (7.5-
fold below TMZ). Our SAR progression led to 3 (the 5-
methyl analog of 2) that maintained a high B/P ratio of 0.8
while simultaneously improving the absolute levels of
compound in the brain (TMZ levels were only 1.4-fold
higher than levels of 3). This result correlates with the
calculated physicochemical properties of
methylating the free N-H of 2 decreases the topological
2
and 3;
The structure of TMZ (Figure 1) offers limited options for
making analogs. Derivatizing at the 6-position is
synthetically complex; moreover, depending on the
substitution, it could perturb the electronics of the ring and
may affect the reactivity of the molecule. We did not want
to modify the 3-methyl group as this is the site of action of
TMZ. The carboxamide group at the 8-position provided a
possible handle for attachment of carboline moieties, with
or without an intervening linker. The synthesis of 3-methyl-
polar surface area (TPSA) from 110.98 to 100.12 and
increases cLogP from 0.57 to 0.9. We expected this
change to enhance the brain penetration of 3. With
16
4
carboxylic acid was previously reported.
-oxo-3,4-dihydroimidazo[5,1-d][1,2,3,5]tetrazine-8-
14, 15
The
carboxamide is distant from the 3-methyl group, which
eventually forms diazomethane. We postulated that the
reactive 8-position carboxamide analogs would afford
compounds that are not chemically compromised from
generating diazomethane, and hence would be capable of
acting through the same mechanism as TMZ. As a quick
check, the 8-position N,N-dimethyl carboxamide analog of
TMZ was made and tested in U87 MG cells for its capacity
compound 3, we accomplished our original goal of utilizing
a carboline as a carrier molecule to favor brain partitioning
of an imidazotetrazine analog.
6
to hydrolyze and produce O -MedG. This compound
performed similarly to TMZ in this assay (data not shown).
Figure 4. SAR progression of γ-carbolines
A general procedure for the synthesis of carboline-
imidazotetrazine analogs and their corresponding AICs is
described in Scheme 1. The key step was an amide
coupling
We next turned our attention to β-carbolines, as
exemplified by compounds 4, 5 and 6 (Figure 5A). Going
from compound 4 to 5 and 6, we made some major
modifications to the physicochemical properties by: (a)
using a shorter linker between the imidazotetrazine and the
carboline (which reduced the molecular volume and
flexibility), (b) converting a basic amine to a neutral amide
and (c) introducing two lipophilic chlorines (which resulted
in an increased cLogP). Such changes have been
documented to increase brain penetration. Compound 5
and 6 have B/P ratios of 3.79 and 10.42, respectively,
compared to a B/P ratio of 0.34 for TMZ (Table 1). Brain
levels of these two compounds were comparable to that of
TMZ (Table 1 and Supplementary Table 1). Importantly,
compounds 5 and 6 had comparable or better brain levels
with concomitantly reduced plasma levels as compared to
TMZ (Figure 5B). These leads could be useful as tool
compounds to explore in vivo efficacy.
Scheme 1. General process for synthesis of carboline-imidazotetrazines and
1
7
the corresponding AIC
of a carboline with an appropriately placed amine “handle”
with the 8-acid analog of TMZ using PyBOP (benzotriazol-
1
-yl-oxytripyrrolidinophosphonium hexafluorophosphate).
The final product was isolated by extraction and drying,
followed by trituration with acetonitrile. This
imidazotetrazine amide analog was further used to
synthesize the corresponding AIC by warming a solution in
phosphate buffer at 60 °C for 4 h followed by purification by
reverse phase HPLC. The synthesis of the β- and γ-
carbolines (Compounds 1-6) are described in the
Supplemental Information.
In the γ-carboline series, we explored attachment of the
imidazotetrazine at the 2- and the 5-positions through an
amino ethyl linker, as shown in Figure 4. Compound 1 has
the imidazotetrazine attached to the γ-carboline at the 5-
position, while compounds 2 and 3 have the attachment at

improvement to the enhanced cell permeability of these
lead compounds as compared to TMZ, under the
assumption that the same physicochemical properties that
result in improved brain penetration would also improve cell
permeability. We chose to do further evaluations with
compound 5.
6
Figure 6. O -Methylation of guanine by TMZ and its carboline analogs in
U87 MG glioblastoma cells. U87 MG cells were incubated for 3 h with TMZ
or the carboline analogs (10 µM shown in black or 100 µM shown in grey)
6
and formation of O -methylguanine was monitored by LC-MS/MS. The
average peak area is shown in the graph. Error bars (where shown) represent
the standard deviation of at least 2 replicates. Where error bars are not shown,
only singlicate experiments were done. Due to the inherent complexity of
this assay, duplicates were done for compounds that were contenders for
further in vivo evaluation, e.g., compound 5.
Figure 5
A. SAR progression of β-carbolines
B. Graphical representation of the improvement in brain:plasma ratio
We compared the calculated and observed properties of
the γ- and β-carbolines with that of TMZ (Table 2). Optimal
brain penetration occurs in cLogP range of 1.5-2.7. While
TMZ has a negative cLogP due to the number of polar
heteroatoms, the small molecular weight and compact
molecular volume allow for some brain penetration,
resulting in a B/P ratio of 0.34. Attachment of the
imidazotetrazine to a carboline increased cLogP; this could
explain the improvement in B/P ratios. Major improvements
to B/P ratios were accomplished by reducing the molecular
volume as in compounds 5 and 6. In these molecules, we
attached the imidazotetrazine directly to the carboline
through an amide bond at the carboline-nitrogen. In doing
so, the basicity of this nitrogen is abrogated as well.
TMZ kills cells by hydrolyzing and releasing the active
diazomethane that methylates purine bases in DNA to such
an extent that cells are unable to repair and apoptosis is
induced. TMZ is known to have little effect on the
18
proliferation of U87 MG cells at 100 µM. Given that our
lead compounds showed enhanced O -MedG formation in
6
U87 MG cells as compared to TMZ, we thought that they
could be more cytotoxic than TMZ and inhibit cell
proliferation. To test this hypothesis, we measured the
proliferation of U87 MG cells in the presence of 100 µM
TMZ or our compounds for 72 h (MTS CellTiter 96 Cell
Proliferation Assay, Promega). We confirmed that 100 µM
TMZ did not inhibit the proliferation of U87 MG cells (Figure
Table 2. Physico-chemical properties like cLogP and pKa can
explain improved brain penetration of compounds 5 and 6.
7). Compounds 1, 3 and 4 did not significantly affect the
proliferation of U87 MG cells but compound 5 inhibited the
growth by 27.3% (Figure 7).
a
Carboline
N pKa
B/P
ratio
0.34
Compound
Structure
cLogP
a
TMZ
-0.99
1.16
1
γ-carboline
γ-carboline
γ-carboline
β-carboline
β-carboline
β-carboline
9.0
8.8
8.8
8.8
-4.2
-4.3
0.67
0.75
0.82
1.36
3.17
10.42
2
3
4
5
6
0.57
0.9
1.21
1.04
1.77
a
ACD Labs PhysChem profiler was used to calculate cLogP and pKa.
We next tested our lead compounds for their ability to
6
form O -MedG in vitro in U87 MG cells as described above
for TMZ. All compounds were tested at both 10 and 100
µM except for 6, which could not be dosed at 100 µM due
to solubility limitations. A positive result in this assay would
reveal the ability of compounds to hydrolyze and release
diazomethane and alkylate DNA. Compounds 1 and 2 had
Figure 7. Inhibition of growth of U87 MG glioblastoma cells by TMZ and
compounds 1-5. U87 MG cells were treated with 100 µM compound for 72 h.
Growth of drug-treated cells was compared with growth of vehicle-treated
cells using the CellTiter Cell Proliferation Kit (Promega). Growth inhibition
by compounds was calculated by dividing the average growth of drug treated
cells (n=3) by the growth of vehicle treated cells (n=3, normalized to 100%).
Error bars represent the standard deviation of the three replicates.
6
similar levels of O -MedG as TMZ, while our lead
compounds from our in vivo screen that had improved B/P
6
ratios over TMZ (3, 5 and 6) had increased levels of O -
MedG as compared to TMZ (Figure 6). We attributed this

We proceeded forward with our lead compound 5 to test
its efficacy in an in vivo orthotopic mouse xenograft model
of glioblastoma with U87 MG cells. One million luciferase
expressing U87 MG cells were surgically implanted into the
brains of female Harlan Nude mice and tumor volume was
monitored by MRI over time. Dosing began on Day 26 post
3
implantation when the mean tumor volume was 10.1 mm
for each group (Figure 8). TMZ was formulated in 50%
PEG400 and administered IV at 50 µmol/kg as a positive
control. The corresponding vehicle group was also dosed
IV. Compound 5 was formulated in 20% N,N-
dimethylacetamide/35% PEG400/30% propylene
glycol/15% 0.9% sterile saline and administered IP at 100
µmol/kg (IP administration was used to avoid tail necrosis
induced by IV dosing). The corresponding vehicle group
was also dosed IP. Compounds were dosed once daily for
five days (QDx5) and animals were monitored for tumor
volume, body weight and mortality (mice found in moribund
condition were euthanized).
1
9-21
As expected,
TMZ slowed tumor growth and
significantly improved survival (Figure 8A, B). Compound 5,
however, did not slow tumor growth (Figure 8A) and the
effects on survival were indistinguishable from those of the
vehicle control groups (Figure 8B). It was not possible to
increase the dose of compound 5 due to tolerability issues
including mortality, lack of grooming and low body carriage.
With this negative result in hand, we looked for an
Figure 9. TMZ efficiently methylated guanine in brain tissue in vivo while
compound 5 did not, despite similar exposure levels. Rats (n=1 per compound
per dose) were dosed with three different doses of TMZ or compound 5.
Brain tissue was collected 3 h post dose and analyzed for drug concentration
explanation that would guide us in future experiments.
6
6
and O -MedG. A. Peak area from the LC-MS/MS analysis of the O -MedG
quantification of rats dosed with TMZ (blue line) or compound 5 (green line).
B. Drug concentration present in brain of rats administered the indicated
doses of TMZ (blue line) or compound 5 (green line).
We sought an explanation for the lack of in vivo activity of
compound 5 (as measured by the formation of O -MedG)
6
since our in vitro data suggested that this compound was
capable of entering cells, hydrolyzing and releasing the
diazomethane to alkylate DNA and cause cytotoxicity. One
possible explanation is that, while compound 5 indeed
reaches the brain when administered to rats, it is
unavailable to enter cells because it is bound to lipid-rich
brain tissue. TMZ, on the other hand, is present in the brain
but is more available in the free fraction to enter cells and
alkylate DNA. To test the proclivity of our compounds to
bind to brain tissue, we incubated compounds at 10 µM
with 200 µL homogenized mouse brain tissue in HBSS
buffer for 1 h at 37 °C (1% final DMSO concentration).
Samples were centrifuged and the percentage of
compound in the insoluble tissue fraction and in the soluble
supernatant were determined by LC-MS/MS. 50.3% (peak
area 3322500) of TMZ was found in the soluble fraction
while 49.7% (peak area 3279500) was found in the
insoluble tissue fraction (Figure 10, Table 3,
Supplementary Table 2). Compound 5 tested under the
same conditions was found to bind much more strongly to
the tissue fraction with 97.8% of compound found in the
insoluble tissue fraction (peak area 2111500) and only
2.2% (peak area 47345) found in the soluble fraction
(Figure 10, Table 3, Supplementary Table 2). The other β-
carboline compounds were similarly highly tissue-bound
(Figure 10, Table 3, Supplementary Table 2). While the
importance of the unbound brain concentration has been
Figure 8. TMZ but not compound 5 inhibited tumor growth and increased
survival in an orthotopic U87 mouse xenograft model. Days of dosing are
indicated by the green line. A. Average tumor volume of animals that
survived to the day 39 is graphed. Error bars represent the standard deviation
of the volumes of the number of animals indicated. Compounds were dosed
3
QDx5 starting on day 26 when the average tumor volume was 10.1 mm . B.
Groups were monitored for survival out to day 63 and the survival curves are
plotted. Survival curves were compared with log-rank Mantel-Cox test (p =
0.004). The survival curve for TMZ treated group was compared to its vehicle
control group (p = 0.0059) and the curves for compound 5 treated group and
its vehicle control were not statistically different.
We demonstrated that compound 5 was present in rat
brain. However, was it able to methylate DNA in the brain?
To address this, TMZ and 5 were administered IV to rats at
10, 20 and 40 mg/kg (51.5, 103 and 206 µmol/kg for TMZ
and 26.5, 53 and 106 µmol/kg for 5) and brain tissue was
6
collected and analyzed for O -MedG and drug6
concentration at 3 h post dose. Analysis of O -MedG
levels revealed that while TMZ efficiently methylated DNA
in vivo, dosing compound 5 did not result in measurable
6
amounts of O -MedG at any dose tested (Figure 9A). The
in vivo concentration of each parent compound, however,
was comparable at equimolar doses (103 µmol/kg TMZ
and 106 µmol/kg 5) (Figure 9B).
22
emphasized and discussed in the literature, we had
assumed the inherent instability of imidazotetrazines would
result in viable concentrations of the reactive diazomethane
at the site of action. Given this, we also measured the

amount of the AIC in the tissue-bound and soluble
fractions. The AIC of TMZ is similarly distributed in the
tissue and soluble fractions as TMZ itself but the AIC
molecules of our β-carboline compounds including
compound 5 were highly tissue bound (Figure 10, Table 3,
Supplementary Table 2). This suggests that the hydrolysis
of the imidazotetrazine is occurring in the tissue and any
released diazomethane was likely quenched locally in the
lipids and did not reach the intended site of action. We
when we evaluated the proclivity of our compounds for
binding brain tissue. In an ex vivo brain fractionation
experiment, our lead compound 5 was almost completely
tissue bound and unlikely to be available and reach the
appropriate cells in vivo. Based on these results, replacing
the carboline with other, less lipophilic brain penetrating
groups should afford compounds with improved efficacy.
Alternatively, the use of cleavable linkers which would
release TMZ after entering the brain would be a novel way
of trafficking TMZ while obviating the lipophilic nature of the
carrier.
therefore concluded that
a
high unbound brain
concentration is an important feature in the design of new
compounds.
Funding: This research did not receive any specific grant
from funding agencies in the public, commercial or non-for-
profit sectors
Acknowledgements: The authors thank Dr. Anil Agarwal for
careful reading and editing of the manuscript.
1
2
2
3
. Chamberlain, M.C. Expert Rev Neurother. 2010, 10, 1537-1544.
. Zhang, J.; Stevens, M.F.; Bradshaw, T.D. Curr Mol Pharmacol.
012, 5, 102-114.
. Ostermann, S.; Csajka, C.; Buclin, T.; Leyvraz, S.; Lejeune, F.;
Decosterd, L.A.; Stupp, R. Clin Cancer Res. 2004, 10, 3728-3736.
. Newlands, E.S.; Stevens, M.F.; Wedge, S.R.; Wheelhouse, R.T.;
Brock, C. Cancer Treat Rev. 1997, 23, 35-61.
. Reid, J.M.; Kuffel, M.J.; Miller, J.K.; Rios, R.; Ames, M.M. Clin
4
Figure 10. Carboline analogs of TMZ bind strongly to insoluble brain tissue.
TMZ, compounds 4, 5, 6 and their corresponding AICs were incubated at 10
µM for 1 h with homogenized mouse brain tissue. Drugs levels were
measured in the soluble supernatant fraction and the insoluble brain tissue
fraction after centrifugation. The percentage of total TMZ, 4, 5 or 6 present
in the insoluble, or bound, fraction is shown in grey and the percentage of
total TMZ, 4, 5 or 6 present in the soluble, or free, fraction is shown in black.
The percentage of total AIC compounds detected in the bound fraction is
shown in turquoise and the percentage of total AIC compounds found in the
soluble, or free fraction, is shown in dark blue.
5
Cancer Res. 1999, 5, 2192-2197.
6. Stupp, R., Ostermann, S., Leyvraz, S., Csajka, C., Buclin, T.,
Decosterd, L.A. Cerebrospinal Fluid Levels of Temozolomide as a
Surrogate Marker for Brain Penetration, ASCOSan Francisco, 2001.
7
. Reyderman, L.; Statkevich, P.; Thonoor, C.M.; Patrick, J.; Batra,
V.K.; Wirth, M. Xenobiotica. 2004, 34, 487-500.
8
9
. Sweetlove, M. Pharmaceutical Medicine. 2012, 26, 113-115.
. Wollowitz, S.K., Erik W. Methods for preparing pyridylethyl-
Table 3. The percentages of TMZ or carboline analogs and
their AICs that was present in the soluble supernatant or
bound to mouse brain tissue. The percentages illustrated in
Figure 10 are listed.
substituted carbolines, in: W.I.P. Organization (Ed.)United States,
2009.
10. Hung, D.T.P., Andrew Asher; Chakravarty, Sarvajit; Jain,
Rajendra Parasmal; Duger Sundeep Pyrido [3, 4-B] Indoles and
Methods of Use, in: W.I.P. Organizaion (Ed.), 2009.
%
Compound
1
1. Otto, R.; Penzis, R.; Gaube, F.; Winckler, T.; Appenroth, D.;
Compound # Supernatant Tissue Bound
Fleck, C.; Trankle, C.; Lehmann, J.; Enzensperger, C. Eur J Med
Chem. 2014, 87, 63-70.
TMZ
AIC TMZ
4
50%
48%
11%
7%
50%
52%
89%
93%
98%
99%
97%
99%
1
2. Abou-Gharbia, M.; Patel, U.R.; Moyer, J.A.; Muth, E.A. J Med
Chem. 1987, 30, 1100-1105.
3. Abou-Gharbia, M.; Patel, U.R.; Webb, M.B.; Moyer, J.A.;
Andree, T.H.; Muth, E.A. J Med Chem. 1987, 30, 1818-1823.
4. Horspool, K.R.; Stevens, M.F.; Newton, C.G.; Lunt, E.; Walsh,
1
10 (AIC 4)
1
5
2%
R.J.; Pedgrift, B.L.; Baig, G.U.; Lavelle, F.; Fizames, C. J Med
Chem. 1990, 33, 1393-1399.
1
1 (AIC 5)
6
1%
1
5. Arrowsmith, J.; Jennings, S.A.; Clark, A.S.; Stevens, M.F. J Med
Chem. 2002, 45, 5458-5470.
6. Physicochemical properties werecalculated usingACD labs
software.
3%
1
12 (AIC 6)
1%
1
1
7. Pajouhesh, H.; Lenz, G.R. NeuroRx. 2005, 2, 541-553.
8. Hanif, F.; Perveen, K.; Jawed, H.; Ahmed, A.; Malhi, S.M.;
In conclusion, the compounds presented here are
imidazotetrazine analogs with B/P ratios that are up to 30-
fold improved over TMZ, with concomitantly reduced
plasma levels when dosed in rodents. Lead compounds6
were tested for their ability to alkylate DNA and form O -
MedG, and found to be more efficacious than TMZ in U87
MG cells. While these results were encouraging, they did
not bear out when tested in vivo in an orthotopic U87
mouse xenograft model and in a DNA methylation assay in
rat. We found a possible explanation for this lack of efficacy
Jamall, S.; Simjee, S.U. Cancer Cell Int. 2014, 14, 133.
9. Candolfi, M.; Curtin, J.F.; Nichols, W.S.; Muhammad, A.G.;
1
King, G.D.; Pluhar, G.E.; McNiel, E.A.; Ohlfest, J.R.; Freese, A.B.;
Moore, P.F.; Lerner, J.; Lowenstein, P.R.; Castro, M.G. J
Neurooncol. 2007, 85, 133-148.
2
0. Mathieu, V.; De Neve, N.; Le Mercier, M.; Dewelle, J.; Gaussin,
J.F.; Dehoux, M.; Kiss, R.; Lefranc, F. Neoplasia. 2008, 10, 1383-
392.
1

21. Kim, J.T.; Kim, J.S.; Ko, K.W.; Kong, D.S.; Kang, C.M.; Kim,
M.H.; Son, M.J.; Song, H.S.; Shin, H.J.; Lee, D.S.; Eoh, W.; Nam,
D.H. Oncol Rep. 2006, 16, 33-39.
2
2. Rankovic, Z. J Med Chem. 2015, 58, 2584-2608.

*Graphical Abstract (for review)
Graphical Abstract
Temozolomide analogs with improved
brain/plasma ratios – exploring the
possibility of enhancing the therapeutic
index of temozolomide
Leave this area blank for abstract info.
Roopa Rai, Monali Banerjee, Darren H. Wong, Emma McCullagh, Ashu Gupta, Shailendra Tripathi, Eduardo
Riquelme, Ram Nivas, Shyamraj Yadav, Mohd. Raja, Pankaj Melkani, Vikas Dixit, Umesh Patil, Ritesh
Shrivastav, Sandip Middya, Felipe Olivares, Javier Guerrero, Arjun Surya, Son M. Pham, Sebastián Bernales,
Andrew A. Protter, David T. Hung and Sarvajit Chakravarty