BI-3406, a potent and selective SOS1::KRAS interaction
inhibitor, is effective in KRAS-driven cancers through
combined MEK inhibition
Marco H. Hofmann 1* , Michael Gmachl1*, Juergen Ramharter1*, Fabio Savarese1, Daniel
Gerlach1, Joseph R. Marszalek3, Michael P. Sanderson1, Dirk Kessler1, Francesca Trapani1,
Heribert Arnhof1, Klaus Rumpel1, Dana-Adriana Botesteanu1, Peter Ettmayer1, Thomas
Gerstberger1, Christiane Kofink1, Tobias Wunberg1, Andreas Zoephel1, Szu-Chin Fu4, Jessica
L. Teh3, Jark Boettcher1, Nikolai Pototschnig1, Franziska Schachinger1, Katharina Schipany1,
Simone Lieb1, Christopher P. Vellano3, Jonathan C. O’Connell2, Rachel L. Mendes2, Jurgen
Moll1, Mark Petronczki1, Timothy P. Heffernan3, Mark Pearson1, Darryl B. McConnell1,
Norbert Kraut1.
1 Boehringer Ingelheim RCV GmbH & Co KG, Vienna, Austria
2 Forma Therapeutics, Watertown, MA, USA
3 TRACTION Platform, Therapeutics Discovery Division, The University of Texas MD
Anderson Cancer Center, Houston, TX 77030, USA
4 Department of Genomic Medicine, The University of Texas MD Anderson Cancer Center,
Houston, TX 77030, USA
*These authors contributed equally to the work
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Supplementary Figures and Legends S1-S519
20
21
Figure S1: Discovery of BI-3406, a potent and selective SOS1::KRAS interaction
inhibitor
a, AlphaScreen and TR-FRET dose response data for the confirmed actives from the
AlphaScreen HTS. Out of the 714 hits with IC50 values of ≤ 10 µM in one of the assay
formats, only 52 (7%) inhibited both read outs. Among these 52 hits were the 18 quinazoline
hits. BI-68BS is highlighted in a dark blue circle. b, Co-crystal x-ray structure of BI-68BS (in
yellow) bound to the catalytic site of SOS1 (in grey as surface representation). Key
interactions of BI-68BS with SOS1: Pi-Staking of the quinazoline core with His905,
interaction of the phenethyl substituent with a lipophilic area defined by Phe890, Leu901 and
Met878, polar interaction of the NH with Asn879 and interaction of the 6-methoxy substituent
with Tyr884. Structure and potency are shown to the lower right. c, Overlay of the co-crystal
22
23
24
25
26
27
28
29
30
31
32
33
x-ray structures of BI-68BS and SOS1 with the previously described SOS1 activator (PDB:
4NYI) (1) and SOS1 demonstrates binding of both compounds to the same pocket. d, The
sphere representation of BI-68BS (in yellow) bound to SOS1 (key amino acid residues shown
in yellow) superimposed with the published SOS1::RAS complex (Tyr884SOS1 and Arg73RAS
shown in orange, PDB: 4NYI) clearly illustrates, that the methoxy substituent of BI-68BS
competes with Arg73RAS for interaction with Tyr884SOS1. Thus, when bound to SOS1, BI-
68BS would clash with KRAS. This prevents KRAS from binding and causes inhibition of
SOS1. e, Synthesis of key intermediate 6: i.) 1.3 eq tributyl(1-ethoxyvinyl)tin, 0.1 eq
Pd(PPh3)2Cl2, 2.0 eq NEt3, dioxane, 80°C, 12 hours, then work-up with 1N HCl, 64% yield;
ii.) 1.5 eq (R)-(+)-2-methyl-2-propane-sulfinamide, 2.5 eq Ti(OEt)4, THF, 80°C, 5 hours, 72%
yield; iii) 1.8 eq NaBH4, THF/H2O = 50/1, -50°C, 5 hours, 63% yield; iv.) 2 eq HCl in
dioxane, rt, 5 hours, 87% yield; v.) 0.05 eq Pd/C, H2, MeOH, rt, 12 hours, 92% yield. f,
Synthesis of BI-3406: vi.) 10 eq CH3CN, 8 eq HCl in dioxane, 60°C, 5 hours, 75% yield; vii.)
1.1 eq (R)-tetrahydrofuran-3-yl 4-methylbenzenesulfonate, 1.2 eq Cs2CO3, DMF, 100°C, 12
hours, 69% yield; viii.) 1.2 eq 2,4,6-triisopropylbenzenesulfonyl chloride; 0.1 eq DMAP, 3 eq
NEt3, DCM, room temperature, 65% yield; ix.) 1.5 eq intermediate 6, 5 eq NEt3, DMSO,
90°C, 6 hours, 79% yield. g, Overlay of SOS1 with BI-3406 and SOS2 (PDB ID 6EIE) in
green revealing clash of BI-3406 (with dotted surface) and Val 903 in SOS2 (spheres shown
around side chain). h, Biochemical protein-protein interaction assays (AlphaScreen). Effect of
BI 3406 on SOS1 wildtype and SOS1 binding site mutants. Recombinant wild type SOS1 and
various point mutations (Y884A, H905V and Y884A / H905V) were tested for the ability of
BI 3406 to disrupt the interaction with KRAS (KRASG12D). i, FLAG-SOS1 wild-
typeFLAG-SOS1 H905V and H905I mutant variants and empty vector control transgenes
were transiently transfected into HEK293 cells. 40 hours after transfection, cells were treated
with DMSO solvent, BI-3406 (1 µM) or trametinib (50 nM) for 4 hours. Protein lysates were
analyzed using a capillary immunodetection assay to quantify phosphoERK and loading
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
control alpha-actinin signals (n=3, mean and single data points of 3 biological repeats are
displayed). Inlay: Lane views of FLAG-SOS1 transgene expression in HEK293 cells and
loading control alphaactinin. j, Cells were starved for one day in medium with 0.5% FCS.
One the next day cells were either treated with 1 ng/mL EGF for exactly 2 minutes or they
were treated with BI-3406 for 1 hour before addition of 1 ng/mL of EGF. Lysate of the cells
was analyzed by Western blotting (left) for reduction of pERK levels and by GLISA (right,
n=2-3, mean±s.d.) for reduction of KRAS-GTP levels. k, Western blotting of cell lysate from
NCI-H358 cells (parental) or NCI-H358 cells carrying a genetically engineered SOS1 (NCI-
H358 SOS1 KO) or SOS2 (NCI-H358 SOS2 KO) knockout to proof absence of SOS1 or
SOS2 in the respective cell lines. The antibody for SOS2 shows two bands on the immunoblot
with one disappearing in the SOS2 KO cell. l, Effect of BI-3406 on RAS-GTP levels in NCI-
H358 cells with SOS2 knockout is more pronounced than in parental cells. NCI-H358 cells
(parental) or NCI-H358 cells carrying a genetically engineered SOS2 knockout (NCI-H358
SOS2 KO) were treated with 1 µM BI-3406 or DMSO for 2 h and RAS-GTP levels relative to
DMSO were quantified (n=2, mean±s.d.). m, Effect of BI-3406 on pERK levels is reduced in
NCI-H358 SOS1 knockout cells and enhanced in NCI-H358 SOS2 knockout cells compared
to NCI-H358 parental cells. NCI-H358 parental, SOS1 or SOS2 CRISPR engineered
knockout cells were incubated with increasing concentrations of BI-3406 for 1 h. pERK and
total ERK levels were analyzed. Dose response curves of pERK levels relative to total ERK
levels are shown (n=2, mean±s.d.). n, in vitro sensitivity of parental NCI-H358, NCI-H358
SOS1 KO or NCI-H358 SOS2 KO cells with BI-3406 in a 3D proliferation assay (n=2,
mean±s.d.). o, Time course of GTP RAS and pERK after BI 3406 treatment. NCI-H358 cells
were treated with 1µM BI 3406 and samples were harvested after 1, 3, 7 and 24h. RAS GTP
and pERK / total ERK levels were quantified.
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
Figure S2: SOS1/SOS2 expression in different models/patients, phosphoprotein response
for BI-3406 inhibition, and compound effects in non-tumorous human cells
a-b, Gene expression of SOS1 and SOS2 across 33 TCGA cancer type cohorts. The dashed
line shows an expression cutoff of 3 TPM (transcripts per million). a, SOS1
(ENSG00000115904) and b, SOS2 (ENSG00000100485). c, Expression of SOS1 plotted
against expression of SOS2 for cell lines (CCLE) and tumors (TCGA). All data was processed
84
85
86
87
88
89
90
with identical processing pipelines (see Material and Methods). BI-3406 sensitive and
resistant cell lines are color-coded in the figure. d, Immunoassays performed with capillary
simple western system using lysate of NCI-H358 (NSCLC) and A549 (NSCLC) cell lines
treated for 2 hours with different concentrations of BI-3406 (medium with 10% FCS) e,
Phosphoprotein levels for ERK and AKT in five cell lines and two time points (2 h and 24 h)
at four dose levels of BI-3406. The DMSO control groups are shown in grey, while BI-3406
groups are shown in shades of blue following increasing compound concentrations.
Measurements are normalized according to (p-ERK/total-ERK) / mean of (p-ERK-c1/total-
ERK-c1, p-ERK-c2/total-ERK-c2) such that the mean value in the DMSO groups is 100%.
Other measurements are then computed in relation to this standard. Each group comprises two
technical replicates (n=2). Mean values of each group are plotted as bars, while error bars
show the mean ± SD. Statistical significance was computed for the 100 nM treatment groups
using a two-sided t-test using a confidence level of 0.95 and setting the true value of the mean
to 100. Asterisks indicate p-values ≤ 0.05. f)-h) To estimate a difference in sensitivity of
tumor versus normal cells for BI-3406, we assessed compound effects on a series of normal
cells (n=3, mean±s.e.m). In vitro sensitivity was tested with BI-3406 on f) human
immortalized epithelial cells (hTERT-RPE-1) g) primary smooth muscle cells (HSMC) and h)
normal foreskin cells (BJ) in a 3D, 7 days proliferation assay in softagar.
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
Figure S3: SOS1 inhibition suppresses tumor growth and KRAS/MAPK signaling in
xenograft models of KRAS-driven cancers
a, Mouse pharmacokinetic (PK) analysis in comparison to IC50 unbound in A549 cells.
Female NMRI-Foxn1 nude mice (n=3 per group, mean±s.e.m) were dosed with BI-3406
orally (50 mg/kg, bid) or intravenously (1 mg/kg) and the plasma concentrations (total and
unbound) were measured at the indicated time points using LC-MS. Unbound IC50 value for
an anti-proliferative effect in A549 cells are shown in comparison. b-c, Modulation of ERK
phosphorylation b, or RAS-GTP c, in subcutaneous A549 tumors following treatment with 50
109
110
111
112
113
114
115
116
117
mg/kg BI-3406 prior treatment and 2h, 7h and 24h post-treatment. Treatment groups comprise
5 animals per group (mean±s.e.m). d, Following b.i.d. (t= 0 and 6h) treatment with 50mg/kg
of BI-3406 mice were sacrificed to explant the tumor to analyze biomarker modulation (see
also Fig. 3a, b, c) and blood was collected to measure plasma levels at the respective time
points (n=5 per group, mean±s.e.m).
e, Representative IHC pictures from d, of pERK in mouse skin: pERK was labeled with DAB
(Brown) Vehicle (A), BI-3406 at 4h (B), BI-3406 at 10h (C) and BI-3406 at 24h.
Magnification: 20x f, Gene expression profiling of pharmacodynamic biomarkers in a MIA
PaCa-2 CDX biomarker experiment. Subset of nine genes showing time-dependent
modulation after BI-3406 (50 mg/kg) monotherapy treatment and sampling at 4 h, 10 h, and
24 h post-first-dose. Dosing was performed b.i.d. (t= 0 and 6h). Treatment groups comprised
3-5 animals. Gene expression fold-changes are color-coded in the heat map and changes
between -1.25-fold down- and +1.25-fold up-regulation are overlaid with a white box. g-h,
Gene expression of SOS1 and SOS2 in the MIA PaCa-2 biomarker experiment described
below. Dosing was performed as described in Fig. S3f g, SOS1 (ENSG00000115904) with a
median gene expression of TPM 50 across treatment groups and h, SOS2
(ENSG00000100485) with a median gene expression of TPM 15 across treatment groups. i,
Anti-tumor effect of BI-3406 (50 mg/kg, twice daily, orally) in the A549 (KRAS
G12S/BRAF wt) non-small cell lung cancer xenograft model (n=7 mice/group, mean±s.e.m,
Student's t-test; one-tailed). j, Median body weight of mice treated as described in k). i, No
significant anti-tumor effect of BI-3406 (50 mg/kg, twice daily, orally) in the A375 (KRAS
wt/BRAF V600E) melanoma cell cancer xenograft model (n=7 mice/group).
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
Figure S4: 3D Proliferation assay with BI-3406 and trametinib and suppression of
tumor growth and body weight curves in xenograft models of KRAS-driven cancers and
a, 3D Proliferation assay with BI-3406 and trametinib in MIA PaCa-2 and DLD1 cells. Cells
were treated in vitro with different concentrations of the SOS1::KRAS inhibitor BI-3406 and
with different concentrations of the MEKi trametinib. Cell growth inhibition (CGI) was
calculated. A CGI of >100% is indicative of net cell death. Bliss excess was calculated. Bliss
excess CGI values of > 0 are indicative of more than additive effects on cell growth
inhibition. (n=2, values calculated from the mean). b, Median body weight of mice bearing
subcutaneous MIA PaCa-2 tumors (n=7 per group). Treatment as described in Figure 4a). c,
Median body weight of mice bearing subcutaneous LoVo tumors (n=7 per group). Treatment
as described in Figure 4c. d-e, Box and whiskers graph of the relative tumor volume from d,
C1047 (KRASG12C) PDX tumors and e, B8032 (KRASG12C) PDX tumors over the course
of treatment, with measurements from individual tumors plotted. The mean tumor volume is
represented by a horizontal line within each box (*p≤0.05, **p<0.005 and ***p<0.0005,
statistical significance was determined using an unpaired t-test per row and the Holm-Sidak
method to correct for multiple comparisons). f-g) Pancreatic cancer PDX tumor growth in
mice treated with Vehicle, BI-3406 (50 mg/kg, bid), trametinib (0.1 mg/kg, bid), or the
combination for the models PATX53 (f) and PAT216 (g) tumors (n=5-7 animals per group,
means±s.e.m., statistical significance was determined using an unpaired t-test per row and the
Holm-Sidak method to correct for multiple comparisons. h-k, The average relative body
weight change for each treatment group in the PDX study with (h) C1047 (colorectal
KRASG12C), (i) with B8032 (colorectal, KRASG12C), j) with (j) PATX53 (pancreatic,
KRASG12V) and (k) with PAT216 (pancreatic, KRASQ61K) PDX tumors. l-m, Gene
expression of SOS1 and SOS2 after 21 days of treatment in the B8032 PDX efficacy study.
Dosing was performed as described in Figure 4c. (l) SOS1 (ENSG00000115904) with a
median gene expression of TPM 51.5 for the control and a median gene expression of TPM
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
54.5 for the combination. (m) SOS2 (ENSG00000100485) with a median gene expression of
TPM 43 for the control and a median gene expression of TPM 49 for the combination.
168
169
170
Figure S5: Biomarker modulation upon combined blockade of MEK and SOS1
a, PD marker modulation upon MEKi/SOS1 combination treatment in vitro. MIA PaCa-2
cells (2D growth) were treated with BI-3406 (1 µM), trametinib (25 nM) or the combination
for 2, 4, 8 hours. Whole cell extracts were subjected to Western blot analysis and compared
with untreated cells using pMEK1/2 (Ser217/221) antibodies. b, Lysate of MIA PaCa-2 cells
(2D growth) treated with BI-3406 (1 µM), trametinib (9 nM and 25 nM) or the combination
for 24, 48, 72 hours was analyzed for modulation of total MEK protein. c, PD marker
modulation upon MEKi/SOS1i combination treatment in vitro. MIA PaCa-2 cells (2D
growth) were treated with BI-3406 (1 µM), trametinib (1, 3 or 9 nM) or the combination for 2,
4, 8 hours. Whole cell extracts were subjected to Western blot analysis and compared with
untreated cells using antibodies against pERK1/2 (Thr202/Tyr204) and the total protein of
ERK1/2. d, PD marker modulation upon MEKi/SOS1i combination treatment in vitro. NCI-
23 cells (2D growth) were treated with BI-3406 (1 µM), trametinib (1, 3 or 9 nM) or the
combination for 2, 4, 8 hours. Whole cell extracts were subjected to Western blot analysis and
compared with untreated cells using antibodies against pERK1/2 (Thr202/Tyr204) and the
total protein of ERK1/2 (a-d) n=3, means±s.e.m., Student's t-test; one-tailed). e-h, PD marker
modulation of ERK and pMEK analyzed with capillary immunodetection assay. e,
Immunoblot of cell lysate of MIA PaCa-2 cells grown in 3D on coated plates (medium with
1%FCS) treated with BI-3406. f, Immunoblot of untreated MIA PaCa-2 cells grown in 3D on
coated plates, and cells treated with 10nM of trametinib alone or in combination with 100 or
300 nM of BI-3406. g, Immunoblot of cell lysate of NCI-H23 cells grown in 3D on coated
plates (medium with 1%FCS) treated with BI-3406. h, Immunoblot of cell lysates of untreated
NCI-H23 cells grown in 3D on coated plates, and cells treated with 10nM of trametinib alone
or in combination with 100 or 300 nM of BI-3406. i, PDX tumor bearing (C1047 or B8032)
mice were treated with Vehicle, trametinib (0.1 mg/kg), BI-3406 (50 mg/kg), or the
combination on a bid schedule, with each dose given 6 hours apart (Supplementary Table
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
S12, n=4 per group). Samples were collected 4 hours after the second dose and tumors were
harvested for RPPA analysis (Supp. A heatmap of log2 normalized protein expression for all
four markers averaged from both samples within treatment groups is shown, with the numbers
denoting fold-change expression over the corresponding Vehicle control. Samples were
analyzed for phospho-ERK (Thr202/Tyr204), phospho-S6 (Ser235/236) and Phospho-AKT
(Ser473). j, MIA PaCa-2 tumor bearing mice (n = 4 mice per group) were treated with either a
vehicle (0.5% Natrosol), 50 mg/kg of BI-3406, 0.1 mg/kg of trametinib or the combination.
Tumors were explanted four hours post treatment and DUSP6 and GAPDH mRNA levels
were measured by QuantiGene single plex assays (one-tailed, t-test). The DUSP6 levels of the
individual tumors were normalized to their respective GAPDH levels (means±s.d.). k, DLD1
cells were treated in vitro for 24, 48 or 72 hours with trametinib, BI-3406 or the combination.
Cells lysate was analyzed for cleaved PARP using the MSD Assay System (mean±s.d.)
Apoptosis Panel Whole Cell Lysate Kit). l, NCI-H358 cells (KRAS G12C) were treated with
solvent control DMSO, the KRAS G12 inhibitor AMG 510 (20 nM), BI-3406 (1 µM), or the
SHP2 inhibitor SHP099 (5 µM) alone or in combinations of AMG 510 with BI-3406 or
SHP099. pERK and alpha-actinin loading control levels were quantified by capillary
immunodection assays in protein lysates were prepared at the indicated time points and
analyzed. The pERK to alpha-actinin signal ratio is displayed in the graph over time (n=2
mean ± SD).
References for supplementary figures and legends
1. Burns MC, Sun Q, Daniels RN, Camper D, Kennedy JP, Phan J, et al. Approach for targeting Ras with small molecules that activate SOS-mediated nucleotide exchange. Proceedings of the National Academy of Sciences of the United States of America 2014;111(9):3401-6 doi 10.1073/pnas.1315798111.
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218219220221
222