A selective BCL-XL PROTAC degrader achieves safe and potent antitumor activity

Article abstract of DOI:10.1038/s41591-019-0668-z

B-cell lymphoma extra large (BCL-X_L) is a well-validated cancer target. However, the on-target and dose-limiting thrombocytopenia limits the use of BCL-X_L inhibitors, such as ABT263, as safe and effective anticancer agents. To reduce

Full text of DOI:10.1038/s41591-019-0668-z

Articles
https://doi.org/10.1038/s41591-019-0668-z
A selective BCL-XL PROTAC degrader achieves
safe and potent antitumor activity
Sajid Khan1,10, Xuan Zhangꢀ ꢀ2,10, Dongwen Lvꢀ ꢀ1,10, Qi Zhang3, Yonghan He1, Peiyi Zhang2, Xingui Liu1,
Dinesh Thummuri1, Yaxia Yuan1, Janet S. Wiegand1, Jing Pei1, Weizhou Zhang4, Abhisheak Sharmaꢀ ꢀ5,
Christopher R. McCurdy2, Vinitha M. Kuruvilla3, Natalia Baranꢀ ꢀ3, Adolfo A. Ferrando6, Yong-mi Kim7,
Anna Rogojina8, Peter J. Houghton8, Guangcun Huangꢀ ꢀ9, Robert Hromas9, Marina Konopleva3,
Guangrong Zhengꢀ ꢀ2,11* and Daohong Zhouꢀ ꢀ1,11*
B-cell lymphoma extra large (BCL-XL) is a well-validated cancer target. However, the on-target and dose-limiting thrombo-
cytopenia limits the use of BCL-XL inhibitors, such as ABT263, as safe and effective anticancer agents. To reduce the toxicity
of ABT263, we converted it into DT2216, a BCL-XL proteolysis-targeting chimera (PROTAC), that targets BCL-XL to the Von
Hippel-Lindau (VHL) E3 ligase for degradation. We found that DT2216 was more potent against various BCL-XL-dependent
leukemia and cancer cells but considerably less toxic to platelets than ABT263 invitro because VHL is poorly expressed in plate-
lets. In vivo, DT2216 effectively inhibits the growth of several xenograft tumors as a single agent or in combination with other
chemotherapeutic agents, without causing appreciable thrombocytopenia. These findings demonstrate the potential to use
PROTAC technology to reduce on-target drug toxicities and rescue the therapeutic potential of previously undruggable targets.
Furthermore, DT2216 may be developed as a safe first-in-class anticancer agent targeting BCL-XL.
he evasion of apoptosis is a key hallmark of cancer1, which is in
PROTACs are bivalent small molecules containing a ligand that
part attributable to the overexpression of anti-apoptotic proteins recognizes the target protein linked to an E3 ligase ligand. Such
in the BCL-2 family, including BCL-2, BCL-XL and myeloid cell molecules can recruit the target protein to the E3 ligase, promote
T
leukemia 1 (MCL-1)2,3. Inhibition of these BCL-2 family proteins with proximity-induced ubiquitination of the target protein and lead to
small molecules has been extensively investigated as a therapeutic its degradation through the ubiquitin proteasome system (UPS).
strategy for cancers4–9, resulting in the discovery of ABT263 (navi- PROTACs act catalytically to induce protein degradation in a sub-
toclax, a BCL-2 and BCL-XL dual inhibitor), ABT199 (venetoclax, a stoichiometric manner. Their effect is not limited by equilibrium
BCL-2 selective inhibitor) and several BCL-XL and MCL-1 mono- occupancy, and therefore requires less total drug exposure27–31. They
selective inhibitors as promising anticancer drug candidates10–16
.
generally have longer-acting activity but reduced toxicity compared
Currently, ABT199 is the only Food and Drug Administration (FDA)- with typical occupancy-driven protein inhibitors and are increas-
approved antitumor agent targeting the BCL-2 family proteins17,18 ingly being used to develop more effective antitumor agents32–38
ABT263 is not approved because inhibition of BCL-XL induces on- More importantly, because PROTACs rely on E3 ligases to induce
target and dose-limiting thrombocytopenia19-21
protein degradation, they can achieve cell and/or tissue selectiv-
.
.
.
Although ABT199 is useful for the treatment of certain hema- ity even when their target proteins are ubiquitously expressed.
tological malignancies, such as chronic lymphocytic leukemia and They achieve this by targeting the proteins to an E3 ligase that is
acute myeloid leukemia, it has limited utility for the treatment of differentially expressed in tumor cells compared with normal tis-
solid tumors22–24, because most solid tumor cells are not depen- sues. Therefore, we hypothesized that PROTAC technology can be
dent on BCL-2 for survival23,24. In contrast, BCL-XL is predomi- exploited to reduce the thrombocytopenia induced by BCL-XL inhi-
nantly overexpressed in many solid tumor cells and also in a subset bition, by converting a BCL-XL inhibitor into a BCL-XL PROTAC
of leukemia cells23,24, and its expression is highly correlated with that targets BCL-XL to an E3 ligase that is minimally expressed in
resistance to cancer therapy, independent of a person’s tumor pro- platelets. Here we report the first proof-of-concept evidence to our
tein P53 (TP53, also known as P53) mutational status25,26 To date, knowledge using PROTAC technology to generate a cell-selective
BCL-XL stands as one of the most important validated canc.er targets BCL-XL PROTAC, termed DT2216, which has improved antitumor
without a safe and effective therapeutic. In this context, develop- activity but reduced platelet toxicity compared with ABT263. This
ment of a platelet-sparing BCL-XL-targeting agent has the potential is achieved by targeting BCL-XL to the Von Hippel-Lindau (VHL)
to transform the treatment of BCL-XL-dependent malignancies.
E3 ligase, which is minimally expressed in platelets. In addition,
1Department of Pharmacodynamics, University of Florida, Gainesville, FL, USA. 2Department of Medicinal Chemistry, College of Pharmacy, University of
Florida, Gainesville, FL, USA. 3Department of Leukemia, University of Texas M.D. Anderson Cancer Center, Houston, TX, USA. 4Department of Pathology,
Immunology and Laboratory Medicine, College of Medicine, University of Florida, Gainesville, FL, USA. 5Department of Pharmaceutics, College of
Pharmacy, University of Florida, Gainesville, FL, USA. 6Department of Pediatrics, Pathology, Cell Biology and Systems of Biology and Institute for Cancer
Genetics, Columbia University, New York, NY, USA. 7Department of Pediatrics, Children’s Hospital Los Angeles, Los Angeles, CA, USA. 8Greehey Children’s
Cancer Research Institute, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA. 9Department of Medicine, the Long School of
Medicine, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA. 10These authors contributed equally: Sajid Khan, Xuan Zhang,
Dongwen Lv. 11These authors jointly supervised this work: Guangrong Zheng, Daohong Zhou. *e-mail: zhengg@cop.ufl.edu; zhoudaohong@cop.ufl.edu
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
DT2216 does not degrade BCL-2, yet can synergistically kill a vari- Data Fig. 3a) and DT2216-induced BCL-XL degradation (Fig. 2b,c).
ety of cancer cells that are not solely dependent on BCL-XL for sur- However, DT2216 had no effect on the levels of BCL-XL in VHL-
vival when combined with a BCL-2 inhibitor, an MCL-1 inhibitor null 786-O renal cell carcinoma (RCC) cells (Fig. 2d). Furthermore,
or a chemotherapeutic agent. These findings support the potential DT2216NC, which lacks VHL binding, failed to form the BCL-XL–
of DT2216 to be developed as a safe first-in-class BCL-XL-targeting DT2216–VHL ternary complex (Extended Data Fig. 3b) and degrade
antitumor agent. More broadly, our results define a new general BCL-XL (Fig. 2e), and inhibition of proteasome activity with MG132
strategy to convert antitumor agents with on-target tissue-specific abolished the degradation of BCL-XL induced by DT2216 (Fig. 2f).
and dose-limiting toxicities to tumor-selective, less toxic PROTACs Moreover, the effect of DT2216 on MOLT-4 cell viability was also
by engaging a tumor- or tissue/cell-specific E3 ligase.
dependent on its PROTAC activity because VHL-L alone was not
cytotoxic to the cells, and it did not have any additive or synergis-
tic effect on MOLT-4 cell viability when combined with ABT263
Results
DT2216 has increased antitumor activity but reduced platelet (Fig. 2g), but it did reduce the cytotoxicity of DT2216 (Fig. 2h).
toxicity. By analyzing human platelet RNA-sequencing data, we Finally, DT2216NC showed no cytotoxicity against MOLT-4 cells
identified VHL as an E3 ligase minimally expressed in human (Fig. 2i). Collectively, these data confirm that DT2216 acts as a
platelets39,40. We confirmed this finding by immunoblotting, PROTAC that depends on the VHL E3 ligase and proteasomes to
which showed that VHL expression was barely detectable in plate- degrade BCL-XL and induce apoptosis in MOLT-4 cells.
lets. However, high levels of expression of VHL and BCL-XL were
detected in multiple human tumor cells, with the exception of VHL- DT2216 is a specific BCL-XL PROTAC. ABT263 has a high bind-
null 786-O renal cell carcinoma (RCC) cell line (Extended Data ing affinity for BCL-XL and BCL-2 but binds more weakly to other
Fig. 1a,b). This was consistent with high levels of VHL and BCL2 members in the BCL-2 family, such as BCL-W or MCL-1 (Extended
like 1 (BCL2L1) mRNA expression in a variety of human malignan- Data Fig. 4a)13,24. The binding affinities of DT2216 for BCL-XL,
cies, as we analyzed using The Cancer Genome Atlas (TCGA) data- BCL-2 and BCL-W were all reduced by about seven- to ninefold
base (Extended Data Fig. 1c)41,42. On the basis of this finding, we compared with that of ABT-263, but those for BCL-XL and BCL-2
rationally designed and synthesized a series of BCL-XL PROTACs remained high. To investigate the ability of DT2216 to degrade not
that target BCL-XL to VHL for ubiquitination and degradation by only BCL-XL, but also BCL-2 and other anti-apoptotic BCL-2 fam-
linking the BCL-2/BCL-XL binding moiety (BCL-2/XL-L) derived ily members, we selected tumor cell lines that express high levels of
from ABT263 to a VHL ligand (VHL-L) (Fig. 1a and Extended BCL-XL, BCL-2, BCL-W and/or MCL-1 (Fig. 3a and Extended Data
Data Fig. 1d). In addition, a BCL-XL PROTAC negative control Fig. 4b,c). Our analyses revealed that DT2216 selectively degraded
(DT2216NC) compound that cannot bind to VHL was synthesized BCL-XL, but did not change the levels of BCL-2 in all cells examined
as a control. Among these BCL-XL PROTACs, DT2216 was selected even though it has a higher binding affinity for BCL-2 than BCL-
as a lead because of its high potency in inducing BCL-XL degrada- XL. In addition, DT2216 had no effect on the levels of BCL-W and
tion in MOLT-4 T-cell acute lymphoblastic leukemia (T-ALL) cells MCL-1 in these cells (Fig. 3b and Extended Data Fig. 4b,c). These
with the half-maximal degradation concentration (DC50) of 63nM findings are in agreement with the observations that DT2216 exhib-
and maximum degradation (Dmax) of 90.8% (Fig. 1b). Notably, we ited an eightfold reduction in its potency against RS4 B-cell ALL
observed minimal reduction (Dmax, 26%) in BCL-XL levels in plate- cells, which primarily depend on BCL-2 for survival, compared
lets after incubation with up to 3µM of DT2216 (Fig. 1c). The with ABT263 (ref. 24) (Extended Data Fig. 4d and Supplementary
induction of BCL-XL degradation by DT2216 in MOLT-4 cells was Table 1). DT2216 was also not cytotoxic to MCL-1-dependent H929
rapid and long lasting (Extended Data Fig. 2a,b). Because both myeloma cells16 (Extended Data Fig. 4e and Supplementary Table
MOLT-4 cells and platelets are solely dependent on BCL-XL for sur- 1). To further validate the specificity of DT2216, we used the stable-
vival19,24,43, we next evaluated the effects of DT2216 on the viability isotope labeling with amino acids in cell culture (SILAC) and liquid
of MOLT-4 cells and platelets in comparison with ABT263. As pre- chromatography–tandem mass spectrometry (LC–MS/MS)-based
viously reported, ABT263 was highly toxic to both MOLT-4 cells proteomics to analyze the changes in proteins in cells after DT2216
and platelets (Fig. 1d)24,43. In contrast, DT2216 (half-maximal effec- and DT2216NC treatment (Extended Data Fig. 4f). The SILAC and
tive concentration (EC50), 0.052μM) was about fourfold more cyto- LC–MS/MS results show that DT2216, but not DT2216NC, reduced
toxic to MOLT-4 cells than ABT263 (EC50, 0.191μM), and exerted the levels of BCL-XL, but both agents did not affect expression of
almost no effect on the viability of platelets up to a concentration any other proteins (Fig. 3c), demonstrating that DT2216 is a spe-
of 3µM (Fig. 1d). Both DT2216 and ABT263 killed MOLT-4 cells cific BCL-XL PROTAC. The lack of BCL-2 degradation by BCL-
by caspase-3-mediated induction of apoptosis in a BCL-2 homol- XL PROTACs is not unique to DT2216, because the other BCL-XL
ogous antagonist killer (BAK)- and BCL-2-associated X protein PROTAC we synthesized that is similar to DT2216, but targets a dif-
(BAX)-dependent manner (Fig. 1e–h and Extended Data Fig. 2c,d). ferent E3 ligase (such as cereblon), degraded BCL-XL but not BCL-2
However, ABT263 functions as a BCL-XL inhibitor that inhibits the (Extended Data Fig. 4g).
interaction of BCL-XL with BAK, BAX and BIM indiscriminately in
The selectivity of a PROTAC is largely determined by its ability
both MOLT-4 cells and platelets, whereas DT2216 acts as a BCL-XL to form a stable and cooperative ternary complex with its targeted
PROTAC that degrades BCL-XL selectively in MOLT-4 cells but protein(s) and E3 ligase44,45. To elucidate the mechanism of selectivity
not in platelets (Fig. 1i,j). These findings confirm that DT2216 is a of DT2216, we first measured the formation of the BCL-2– or BCL-
BCL-XL PROTAC that has improved antitumor potency and XL–DT2216–VHL ternary complexes in vitro using the AlphaLISA
reduced toxicity to platelets compared with ABT263.
assay44. These analyses revealed that both BCL-2 and BCL-XL can
form stable ternary complexes with DT2216 and VHL in cell-free
DT2216 induces proteasomal degradation of BCL-XL via VHL. conditions (Fig. 3d,e). However, only BCL-XL, but not BCL-2, seems
To further confirm that DT2216 degrades BCL-XL via VHL and to be able to form stable ternary complexes with DT2216 and VHL
proteasome, we first examined the effects of ABT263 and VHL-L in live cells, as determined by nanoBRET assay, a bioluminescence
alone and in combination on BCL-XL levels in MOLT-4 cells and resonance energy transfer (BRET)-based assay to measure the inter-
found that none of these treatments affected the levels of BCL-XL action of two binding proteins in live cells46 (Fig. 3f), even though
(Fig. 2a). In addition, we found that preincubation of MOLT-4 cells both proteins could bind to DT2216 in a cellular thermal-shift
with an excess amount of ABT263 or VHL-L inhibited the forma- assay (Extended Data Fig. 4h). These results are consistent with
tion of the BCL-XL–DT2216–VHL ternary complex (Extended the ability of DT2216 to selectively induce BCL-XL, but not BCL-2,
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
a
c
b
d
MOLT-4 + DT2216 (µM)
Platelets + DT2216 (µM)
Active VHL E3 ligase
ligand
MOLT-4
125
100
75
50
25
0
Linker
DT2216
DT2216 (EC50 = 0.052 µM)
ABT263 (EC50 = 0.191 µM)
KDa
30
KDa
30
BCL-XL
BCL-XL
β-actin
45
β-actin
45
0.001 0.01 0.1
1
10
BCL-2/BCL-XL ligand
DT2216NC
Inactive VHL E3 ligase
ligand
Concentration (µM)
Linker
125
100
75
50
25
0
125
100
75
50
25
0
Platelets
125
100
75
50
25
0
DC50 = 0.063 µM
max = 90.8%
D
DT2216 (EC50 > 3 µM)
ABT263 (EC50 = 0.237 µM)
DC50 > 3 µM
D
max = 26%
BCL-2/BCL-XL ligand
0.0
0.5
1.0
1.5
0
1
2
3
4
0.01
0.1
1
10
DT2216 (µM)
DT2166 (µM)
Concentration (µM)
g
e
j
i
DT2216
(µM)
Veh 0.1 0.3 0.1 0.3
ABT263
(µM)
MOLT-4
IP: BCL-XL
Platelets
IP: BCL-XL
125
100
IP: IgG
Input
Input
KDa
75
50
25
0
Caspase-3
35
sgCTRL (EC50 = 0.086 µM)
sgBAX + BAK #1 (EC50 > 2 µM)
sgBAX + BAK #2 (EC50 > 2 µM)
KDa
KDa
20
Cleaved caspase-3
20
25
23
BAX
BAX
17
0.001 0.01 0.1
1
10
25
23
BAK
BAK
116
89
PARP1
DT2216 (
µ
M)
Cleaved PARP1
BIMEL
BIMEL
h
125
100
75
50
25
0
23
30
PUMA
45
β-actin
BCL-XL
30
45
BCL-XL
β-actin
β-actin
sgCTRL (EC50 = 0.03 µM)
sgBAX + BAK #1 (EC50 > 2 µM)
sgBAX + BAK #2 (EC50 > 2 µM)
45
Veh
QVD
f
150
100
50
0
0.001 0.01 0.1
1
10
ABT263 (
µ
M)
0
62.5
125
DT2216 (nM)
Fig. 1 | DT2216, a BCL-XL PROTAC, selectively induces BCL-XL degradation and apoptosis in BCL-XL-dependent MOLT-4 T-ALL cells but not in platelets.
a, Chemical structures of DT2216 and negative-control DT2216NC showing a dual BCL-2 and BCL-XL ligand (BCL-2/BCL-XL ligand) linked to a VHL ligand
via an optimized linker. DT2216NC has the inactive VHL ligand that does not bind to VHL. b,c, DT2216 selectively degrades BCL-XL in MOLT-4 cells (b) but
not in platelets (c) after treatment with increasing concentrations of DT2216 as indicated for 16ꢀh. Top, Representative immunoblot. Bottom, Densitometric
analyses of BCL-XL expression. Data points represent the mean (nꢀ=ꢀ2 and 3 independent experiments for MOLT-4 and platelets, respectively). DC50, the
drug concentration causing 50% protein degradation; Dmax, the maximum level of degradation. d, Viability of MOLT-4 cells and human platelets were
determined after they were incubated with increasing concentrations of DT2216 and ABT263 for 72ꢀh. The data are presented as meanꢀ ꢀs.d. from six and
three replicate cell cultures in a representative experiment for MOLT-4 and platelets, respectively. Similar results were also observed in two additional
independent experiments. For the platelet-viability assay, each experiment used platelets from one individual donor. EC50 values are the average of three
independent experiments. e, A representative of two independent immunoblot analyses of cleaved and full-length caspase-3 and poly(ADP-ribose)
polymerase 1 (PARP1) in MOLT-4 cells 24ꢀh after they were treated with vehicle (Veh), DT2216 or ABT263. f, Cell viability of MOLT-4 cells was determined
after the cells were pretreated with the pan-caspase inhibitor Q-VD-OPh (QVD, 10ꢀμM) and were then treated with DT2216 for 72ꢀh at the indicated
concentrations. Data are presented as meanꢀ ꢀs.d. from six replicate cell cultures in a representative experiment. Each symbol represents data from an
individual replicate. Similar results were also observed in one additional independent experiment. g,h, Viability of non-targeting small guide RNA (sgRNA)-
transfected (sgCTRL) and BAX and BAK (BAK1) double-knockout (KO; represented as sgBAXꢀ+ꢀBAK) H146 cells was determined after cells were incubated
with increasing concentrations of DT2216 or ABT263 for 72ꢀh. Data are presented as meanꢀ ꢀs.d. from three replicate cell cultures in a representative
experiment. Similar results were also observed in one additional independent experiment. i,j, MOLT-4 cells and human platelets were treated with
either DT2216 (DT, 1ꢀµM) or ABT263 (1ꢀµM) for 6ꢀh. MOLT-4 cells were pretreated with MG132 (MG, 1ꢀµM) for 1ꢀh in order to block BCL-XL degradation.
Immunoblots of BAX, BAK and BH3-only pro-apoptotic proteins, BIM-extra large (BIMEL) and PUMA, after immunoprecipitation with BCL-XL and in whole-
cell lysates (Input) are shown from a single experiment. β-actin was used as an equal loading control in all immunoblot analyses shown in b, c, e, i and j.
The uncropped immunoblot images related to this figure are provided in separate source data file.
ubiquitination and subsequent degradation by proteasomes
Inspection of BCL-XL–ABT263 X-ray crystal structures (protein
(Fig. 3g) and are in agreement with previous reports demonstrating data bank (PDB) entry: 4QNQ) (Fig. 3h)14 revealed Lys 16, Lys 20, Lys
that the specificity of a PROTAC can be determined in part by its 87 and Lys 157 in BCL-XL as solvent-exposed residues and potential
ability to form ternary complexes with its target proteins44,45. The ubiquitination sites. The conformations of other two lysine residues
differential formation of ternary complexes in vitro and in cell may (Lys 205 and Lys 233) are not resolved, probably because they are
be attributable to numerous factors, including the inherent differ- buried in the transmembrane region of mitochondria, which pre-
ences between these two assays. For example, the full-length BCL-XL cludes them from being potential ubiquitination sites47. Analyses of
and BCL-2 were expressed in cells for the nanoBRET assay, whereas BCL-XL mutants with a lysine-to-arginine substitution revealed that
recombinant-transmembrane-domain-deleted proteins were used DT2216 can degrade Flag-tagged wild-type BCL-XL and the K157R
for the AlphaLISA assay. In addition, some other proteins may inter- mutant, but not the BCL-XL mutants with all lysines mutated to
fere with the interaction between BCL-2, DT2216 and VHL in live arginines or with a K87R or K87H single mutant (Fig. 3i,j). These
cells. However, the exact reason for the limited capacity of DT2216 results support the notion that K87 is required for the degradation
to form ternary complexes with BCL-2 in cells is not known at this of BCL-XL by DT2216. Consistently, only K87-ubiquinated BCL-XL
point and would be an important topic for future research.
could be detected in DT2216-treated cells by proteomics. Moreover,
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
a
b
c
DT2216 (µM)
DT2216 (µM) – 0.1
1
–
–
+
0.1
+
1
+
ABT263 (1 µM)
VHL-L (10 µM)
–
–
+
–
–
+
+
+
–
–
0.1
–
1
–
–
+
0.1
+
1
+
ABT263 (1 µM)
VHL-L (10 µM)
– –
KDa
30
KDa
30
KDa
30
BCL-XL
BCL-XL
BCL-XL
β-actin
45
β-actin
45
β-actin
45
e
d
g
f
VHL-null cells + DT2216 (µM)
–
–
0.1
–
1
–
0.1
+
1
DT2216 (µM)
KDa
30
KDa
30
KDa
30
MG132 (1 µM)
–
+
+
BCL-XL
BCL-XL
BCL-XL
45
45
β-actin
45
β-actin
β-actin
h
i
DT2216
DT2216NC
Veh
Veh
VHL-L (1 µM)
150
100
50
150
100
50
150
100
50
VHL-L (1 µM)
0
0
0
0
0
0.25
ABT263 (µM)
0.125
DT2216 (µM)
0
0.125 0.25 0.5
Concentration (µM)
1
Fig. 2 | DT2216 degrades BCL-XL in a VHL- and proteasome-dependent manner. a, ABT263 and/or VHL-L cannot induce BCL-XL degradation in MOLT-4
cells. An immunoblot analysis of BCL-XL in MOLT-4 cells is shown after the cells were treated with ABT263, VHL-L or the combination of both for 16ꢀh.
b, Pretreatment with ABT263 blocks BCL-XL degradation by DT2216. An immunoblot analysis of BCL-XL in MOLT-4 cells after the cells were either left
untreated or pretreated with ABT263 for 1ꢀh and then were treated with or without DT2216 as indicated for 16 h. c, Pretreatment with VHL-L blocks the
BCL-XL degradation by DT2216. An immunoblot analysis of BCL-XL in MOLT-4 cells is shown. The cells were either left untreated or pretreated with VHL-L
for 1ꢀh and then were treated with or without DT2216 as indicated for 16ꢀh before being assayed. d, An immunoblot analysis of BCL-XL in VHL-null 786-O
cells treated with vehicle or increasing concentrations of DT2216 for 16ꢀh. e, An immunoblot analysis of BCL-XL in MOLT-4 cells is shown. The cells were
treated with 1ꢀµM of DT2216 and its negative control DT2216NC for 16ꢀh before being assayed. f, Proteasome inhibition blocks the BCL-XL degradation by
DT2216. A representative of two immunoblot analyses of BCL-XL in MOLT-4 cells after they were either left untreated or pretreated with the proteasome
inhibitor MG132 for 1ꢀh, and then were treated with or without DT2216 for 16ꢀh. g, Cell viability of MOLT-4 cells treated with or without ABT263 in the
presence or absence of VHL-L for 72ꢀh. h, Cell viability of MOLT-4 cells after they were either left untreated or pretreated with VHL-L for 1ꢀh, and then
were treated with DT2216 for 72ꢀh. The data presented in g and h are meanꢀ ꢀs.d. from three replicate cell cultures in one representative experiment. Each
symbol represents data from an individual replicate. Similar results were obtained in an additional independent experiment. i, Cell viability of MOLT-4 cells
after treatment with increasing concentrations of DT2216 or its negative control DT2216NC for 72ꢀh. The data represent meanꢀ ꢀs.d. from six replicate
cell cultures in one representative experiment. Each symbol represents data from an individual replicate. Similar results were obtained in one additional
independent experiment. β-actin was used as an equal loading control in immunoblot analyses shown in a–f. The uncropped immunoblot images related to
this figure are provided in separate source data file.
BCL-XL that only retains Lys 87 (K87-only, that is all the lysines are ranging from 25 to 100 mg per kg body weight caused severe throm-
mutated to arginines except Lys 87) could be degraded by DT2216 bocytopenia in mice after 6h of administration (Fig. 4c), which
(Fig. 3j,k). These findings suggest that DT2216 degrades BCL-XL raises the risk of spontaneous hemorrhage as reported previously49.
in a Lys 87-ubiquitination-dependent manner, which is in agree- Platelet count gradually recovered 3 d after ABT263 administra-
ment with a recent report indicating that the formation of a ternary tion and then exceeded the normal value thereafter, before com-
complex is necessary but not sufficient for a PROTAC to induce its ing back to the normal level 10 d after the treatment. This rebound
target ubiquitination and degradation48.
can potentially increase the risk of thrombosis50. In contrast, platelet
counts after treatment with DT2216 were mildly reduced and were
DT2216 is a more potent and safer antitumor agent than ABT263 not followed by reactive thrombocytosis (Fig. 4c). These results
in vivo. DT2216 is metabolically stable and has a favorable pharma- demonstrate that, at therapeutically equivalent doses, DT2216 is
cokinetic property for in vivo study via either intraperitoneal (i.p.) less toxic to platelets than ABT263 in mice.
or intravenous (i.v.) injection, but is not bioavailable by oral (per os,
Dose-escalation studies showed that DT2216 after weekly
p.o.) administration (Extended Data Fig. 5a–c and Supplementary administration at 15 mg per kg body weight was more effective at
Tables 2 and 3). A single dose of i.p. injection of DT2216 at 15mg suppressing the growth of MOLT-4 T-ALL xenografts in mice than
per kg body weight produced an intratumoral concentration of 7.5 mg per kg body weight DT2216 was (Extended Data Fig. 6a–c).
DT2216 in MOLT-4 T-ALL xenografts significantly greater than We next compared the effects of repeated treatments with DT2216
the cellular EC50 values of DT2216 for the cells for more than one (15mg per kg body weight, every 7 d (q7d), i.p.) and ABT263 (50 mg
week, which led to a sustained reduction in BCL-XL expression per kg body weight, qd, p.o.) on blood platelet counts and MOLT-4
(Fig. 4a,b, Extended Data Fig. 5d,e and Supplementary Table 4). T-ALL xenograft growth (Fig. 4d–g and Extended Data Fig. 7). In
Next, we examined the effects of different doses of DT2216 on these analyses, daily ABT263 treatment induced severe and per-
platelet levels in mice in comparison with the therapeutically equiv- sistent thrombocytopenia with only moderate effects on MOLT-4
alent doses of ABT263; DT2216 is about fourfold more potent than T-ALL xenograft growth as shown previously13. In contrast, once
ABT263 against MOLT-4 cells in vitro (Fig. 1d)13. ABT263 doses weekly dosing of DT2216 almost completely inhibited MOLT-4
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
a
c
b
14
12
10
8
H146 EJM
RS4 H929
14
12
10
8
DT2216
DT2216NC
KDa
30
BCL-XL
KDa
30
BCL-XL
BCL-XL
BCL-2
MCL-1
6
6
28
BCL-2
4
4
28
40
40
18
MCL-1
VHL
2
2
0
0
0
0.5
1
1.5
0
0.5
1
1.5
β-actin
45
45
β-actin
Fold change
Fold change
e
d
f
BCL-XL
BCL-2
18
18
16
14
12
10
8
BCL-XL
250,000
200,000
150,000
100,000
50,000
0
150
100
50
BCL-XL
BCL-2
BCL-2
800,000
DT2216
DMSO
16
14
12
10
8
DT2216
DMSO
600,000
400,000
200,000
0
6
6
0
0.001 0.01 0.1
1
10 100
0.001 0.01 0.1
1
10 100
10
10
0.039
0.625
0.039
0.625
Concentration (µM)
Concentration (µM)
0.0024
0.0024
pH
0.00015
0.00015
DT2216 (µM)
DT2216 (µM)
g
h
i
Flag-BCL-XL
Flag-BCL-2
Flag-BCL-2
HA-Ub
DT2216 –
Flag-BCL-XL
HA-Ub
+
+
–
+
+
+
+
+
–
+
+
+
+
+
+
+
+
+
+
+
+
+
–
–
+
+
–
+
–
+
+
+
+
Flag-BCL-XL WT
HA-Ub
DT2216
HA-Ub +
DT2216
DT2216
MG132
Lys 157
Lys 16
–
DT2216
KDa
~30
Lys 20
Lys 87
Flag
HA
HA
HA
HA
45
β-actin
BCL-XL
with BCL-2/BCL-XL ligand
Flag
Flag
β-actin
Flag
Flag
IP: Flag
WT
IP: Flag
WT
β-actin
Input
WT
Input
Flag-BCL-XL
(
79EVIPMAAVKGGQALR91
y10
)
k
100%
j
y10+2H
770.44 m/z, 2+, 1,538.86 Da, (Parent error: 18 ppm)
K+114
Q
A
L
R
E
R
V
P
M
A
A
V
V
K+114
A
V
L
A
Q
A
M
P
I
E
K157R
Flag BCL-XL
K87R
K87H
K-ko
WT
K87-only
b2
y10+1
–
+
–
+
–
+
–
+
–
+
–
+
–
+
–
+
–
+
DT2216
Flag
KDa
~30
y8
a2
y7
y5
b7
y8+1
y6
b8
y9
y4
y1
b1
by33
y2
β-actin
b10
b4
b5
y11
b6
45
0%
0
250
500
750
m/z
1,000
1,250
1,500
Fig. 3 | DT2216 is a BCL-XL-specific PROTAC and induces BCL-XL degradation through Lys 87 ubiquitination. a, A representative of two immunoblot
analyses of BCL-XL, BCL-2, MCL-1 and VHL in distinct tumor cell lines. b, A representative immunoblot analysis of BCL-XL, BCL-2 and MCL-1 in H146 SCLC
cells after they were treated with 0.1ꢀµM DT2216 for 48ꢀh, and in RS4 B-ALL cells and EJM and H929 multiple myeloma cells after they were treated with
1ꢀµM DT2216 for 16ꢀh. Similar results were obtained in one additional independent experiment. c, Proteomic analysis showing specificity of DT2216 for
BCL-XL degradation in comparison with its negative-control DT2216NC in WI-38 cells. d, Ternary complex formation of BCL-XL or BCL-2 with DT2216 and
VHL, determined by AlphaLISA assay. Data are expressed as the mean of a single experiment (nꢀ=ꢀ2 technical replicates). Similar results were obtained in
two more independent assays performed with BCL-XL. e, pH stability of the ternary complex formed by DT2216 and the VHL complex and BCL-XL or BCL-
2, as measured by AlphaLISA assay. Data are expressed as mean (nꢀ=ꢀ2 technical replicates). Similar results were obtained in one additional independent
experiment. f, NanoBRET ternary complex formation of BCL-XL and BCL-2. Ternary complex formation was determined in 293T cells after they transiently
expressed HiBit-BCL-XL, LgBit and HaloTag-VHL or HiBit-BCL-2, LgBit and HaloTag-VHL and were then treated with a serial dilution of DT2216. Data
are expressed as meanꢀ ꢀs.e.m. of three independent experiments. g, Representative immunoblot of hemagluttinin (HA), Flag and β-actin following Flag
immunoprecipitation of protein extracts from 293T cells that were cotransfected as indicated with Flag-BCL-XL and HA-tagged ubiquitin (HA-Ub) or Flag-
BCL-2 and HA-ubiquitin plasmids, and then were treated with or without DT2216 (1ꢀµM) and MG132 (10ꢀµM) as indicated for 4ꢀh. Data are representative
of three independent experiments. h, Crystal structure of BCL-XL with ABT263 (BCL-2/BCL-XL ligand). The lysines are colored blue. i, A representative
immunoblot analysis of BCL-XL showing that DT2216 induces wild-type (WT) BCL-XL degradation in a proteasome-dependent manner. Flag-BCL-XL-WT
plasmid was transfected into 293T cells for 40ꢀh, and then the cells were treated with or without DT2216 (1ꢀµM) and MG132 (10ꢀµM) as indicated for
6ꢀh. Similar results were obtained in two more independent experiments. j, Representative immunoblot analysis of BCL-XL showing that DT2216 induces
BCL-XL degradation dependent on Lys 87 ubiquitination. For the analysis, Flag-BCL-XL-WT and the K87R and K157R (lysine to arginine), K87H (lysine to
histidine), K-ko (all the lysines in BCL-XL were mutated to arginines) and K87-only (all the lysines in BCL-XL were mutated to arginines except Lys 87) Flag-
BCL-XL mutant plasmids were transfected into 293ꢀT cells for 40ꢀh, and then the cells were treated with or without 1ꢀµM DT2216 for 6ꢀh. Similar results were
obtained in two more independent experiments. k, Lys 87 is the only ubiquitination site triggered by DT2216. 293T cells were co-transfected with Flag-
BCL-XL and HA-ubiquitin vectors. Extracts were immunoprecipitated with anti-Flag affinity resin, followed by trypsin and AspN digestion and tandem mass
spectrometry, as described in the Methods. A fragmentation spectrum of ubiquitinated EVIPMAAVKQALR peptide (ubiquitinated Lys 87 residue) of BCL-
XL. The parent ion corresponding to this peptide has been subjected to higher-energy collisional dissociation in mass spectrometer. The detected b- and
y-fragment ion series have been labeled. The results were obtained from a single experiment. β-actin was used as equal loading control in immunoblotting
experiments shown in a, b, g, i and j. The uncropped immunoblot images related to this figure are provided in separate source data file.
T-ALL xenograft growth with only moderate and transient reduc- at a lower dose level (15mg per kg body weight, q7d, i.p.) that did
tion in platelet counts and no significant change in body weight. not cause severe thrombocytopenia, effectively responded to treat-
MOLT-4 T-ALL xenografts progressing under ABT263 treatment ment with DT2216 (15mg per kg body weight, q4d, i.p.) without
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
a
c
b
DT 7.5 mpk
DT 15 mpk
DT 22.5 mpk
Vehicle
1,200
1,000
800
600
400
150
100
50
Veh
DT2216
1,400
1,200
1,000
800
ABT 25 mpk
ABT 50 mpk
ABT 100 mpk
600
400
200
0
200
EC50 for MOLT-4 cells
0
0
1
2
4
7
14
0
5
10
15
0.25
1
2
3
4
5
7
10
Days after treatment
Days after treatment
Post-treatment time (d)
d
e
f
DT2216 (15 mpk, q7d, i.p.)
ABT263 (50 mpk, qd, p.o.)
Veh
1,000
800
600
400
200
0
25
2,000
1,500
1,000
500
Veh
DT2216
ABT263
20
15
10
5
P < 0.0001
Veh
DT2216
ABT263
P < 0.0001
P = 0.0033
0
0
0
10
20
30
0
10
20
30
Post-treatment time (d)
Post-treatment time (d)
Post-treatment time (d)
g
h
Veh
DT2216
ABT263
KDa
30
P < 0.0001
2.5
2.0
1.5
1.0
0.5
0.0
P < 0.0001P = 0.0031
BCL-XL
BCL-2
MCL-1
β-actin
28
40
45
Veh DT2216 ABT263
Fig. 4 | DT2216 is more potent against MOLT-4 T-ALL xenografts and less toxic to platelets than ABT263 in mice. a, Concentration of DT2216 in MOLT-4
tumors after a single DT2216 administration (15 mg per kg body weight/i.p.). Data represent the average of two mice in a group at each time point.
b, Densitometric analysis of BCL-XL protein levels in tumors (mean, nꢀ=ꢀ2 mice in a group at each time point) at different durations after a single vehicle
or DT2216 administration (15 mg per kg body weight, i.p.). Each symbol represents data (% of vehicle) from an individual animal. The representative
immunoblots are shown in Extended Data Fig. 5e. c, Numeration of platelets (PLT) 0.25, 1, 2, 3, 4, 5, 7 and 10 d after a single i.p. injection with DT2216 or
p.o. dosing with ABT263 at the indicated doses. Data are represented as meanꢀ ꢀs.e.m. (nꢀ=ꢀ3 mice in each group). Each symbol represents data from an
individual animal. mpk, mg per kg body weight. d, Numeration of platelets after the mice were continuously treated with DT2216 or ABT263 as indicated.
Data are represented as meanꢀ ꢀs.e.m. (nꢀ=ꢀ3 mice in each group until day 12.25; nꢀ=ꢀ7 mice in vehicle, 7 mice in DT2216 and 8 mice in ABT263 at day 17.25;
nꢀ=ꢀ7 mice in vehicle, 6 mice in DT2216 and 8 mice in ABT263 at day 24.25). Each symbol represents data from an individual animal. In c and d, the green
and red dashed lines indicate average platelet count in vehicle-treated mice and platelet count at or below which, respectively, there is an increased risk of
hemorrhage. e,f, Body-weight and tumor-volume changes in mice after the start of treatment with vehicle, DT2216 (15ꢀmg per kg body weight, q7d, i.p.) or
ABT263 (50ꢀmg per kg body weight, qd, p.o.). Data presented are meanꢀ ꢀs.e.m. (nꢀ=ꢀ7 mice in vehicle, 7 mice in DT2216 and 8 mice in ABT263 at the start
of treatment). Statistical significance was determined by two-sided unpaired Student’s t-test. g, MOLT-4 tumor-bearing mice were euthanized 25 d after
treatment initiation (4 d and 1 d after last dose of DT2216 and ABT263, respectively). Tumor weights at the end of study are shown. Data are presented
as meanꢀ ꢀs.e.m. (nꢀ=ꢀ7 mice in vehicle, 6 mice in DT2216 and 8 mice in ABT263). Each symbol represents data from an individual animal, and the middle
horizontal line represents the mean. Statistical significance was determined by two-sided unpaired Student’s t-test. h, Immunoblot analysis of BCL-XL, BCL-2
and MCL-1 in MOLT-4 tumors (nꢀ=ꢀ3 mice in each group). The uncropped immunoblot images related to this figure are provided in a separate source data file.
significant changes in body weight and blood platelet counts
Next, we examined whether DT2216 can be combined with
(Extended Data Fig. 8a–f). The potent antitumor activity of DT2216 ABT199 to more effectively eradicate tumors dependent on both
was correlated with its ability to induce BCL-XL degradation BCL-XL and BCL-2, such as H146 SCLC xenografts, than ABT263
(Fig. 4h). These findings support that DT2216 is a safer and more can (Fig. 5c). Treatment of H146 SCLC xenografts with DT2216,
potent antitumor agent than ABT263. It also demonstrated that ABT199, ABT263 and DT2216 plus ABT199 induced no significant
DT2216 has the potential to be used as a single agent to treat tumors changes in body weight (Fig. 5d). Administration of DT2216 at 15
solely dependent on BCL-XL.
mg per kg body weight per week by i.p. injection alone or in com-
bination with ABT199 at 50mg per kg body weight per day by p.o.
Synergy of DT2216 in combination with BCL-2 or MCL-1 inhibi- administration did not cause any significant reduction in platelet
tors. Hematological and solid tumors frequently engage more levels, whereas i.p. administration of 15 mg per kg body weight per
than one member of the BCL-2 family proteins for survival24,51,52
.
week of ABT263 did (Fig. 5e). Treatment with DT2216, ABT199
For example, NCI-H146 (H146) human small-cell lung cancer or ABT263 alone significantly inhibited tumor growth, resulting
(SCLC) cells are dependent on both BCL-XL and BCL-2 for sur- in 81.6%, 57.9% and 59.5% mean tumor growth inhibition (TGI)
vival, whereas EJM myeloma cells depend on MCL-1 and BCL-XL at the end of the experiment (Fig. 5f,g and Extended Data Fig. 9),
for survival16 (Fig. 5a,b and Supplementary Table 1). Inhibition of respectively. Moreover, the combination of DT2216 and ABT199
BCL-2 and MCL-1 with their respective inhibitors ABT199 and induced markedly increased antitumor effects resulting in almost
S63845 had moderate effects on the viability of H146 and EJM cells, complete suppression of tumor growth (mean 98.2% TGI). The TGI
respectively. However, their effect was significantly augmented by induced by DT2216 was also associated with a significant reduction
the addition of DT2216, resulting in synergistic killing of both cell in BCL-XL expression in the tumors harvested at the end of DT2216
lines (Fig. 5a,b).
treatment (Fig. 5h). These findings suggest that DT2216 can be
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
a
b
EJM
H146
125
Compound
DT2216
EC50 (µM)
0.16
125
100
75
50
25
0
EC50 (µM)
>2
CI values
0.15
Compound
DT2216
S63845
DT+S
CI values
100
DT2216 (DT)
ABT199 (199)
DT+199
EC75
0.039
0.036
0.037
EC75
DT2216 (DT)
S63845 (S)
DT+S
75
50
25
0
ABT199
DT+199
0.57
EC90
0.183
0.503
0.057
EC90
0.043
Average
0.166
Average
0.001 0.01
0.1
1
10
0.001 0.01
0.1
1
10
Concentration (µM)
Concentration (µM)
c
d
e
P = 0.7511
P = 0.6306
P = 0.0147
P = 0.1139
P = 0.0262
P = 0.9965
Treatments and observations
H146 SCLC engraftment
25
20
15
10
5
P = 0.1652
P = 0.3719
Time
1,200
1,000
800
600
400
200
0
–30
–20
–10
0
10
20
30
1,200
(d)
Veh
1,000
800
600
400
200
0
DT2216
ABT263
ABT199
DT+199
Treatment started
1. Vehicle
2. DT2216 (15 mpk, q7d, i.p.)
3. ABT263 (15 mpk, q7d, i.p.)
4. ABT199 (50 mpk, qd, p.o.)
5. DT+199
0
CB-17 SCID mice
5 × 106 H146 cells,
s.c.
Blood
Blood
Termination
0
10
20
30
collection
for CBC
collection
for CBC
Post-treatment time (d)
Veh
Veh
DT2216ABT263ABT199DT+199
DT2216ABT263ABT199DT+199
g
h
f
Veh
DT2216
KDa
30
2.0
1,000
800
600
400
200
0
Veh
BCL-XL
BCL-2
MCL-1
DT2216
ABT263
ABT199
DT+199
1.5
1.0
28
*
a
b
a
b
c
40
45
0.5
0.0
c
β-actin
0
10
20
30
Veh
Post-treatment time (d)
DT2216
ABT263ABT199
DT+199
Fig. 5 | Synergy of DT2216 with other BCL-2 family protein inhibitors. a, Percentage of viable H146 SCLC cells after 72ꢀh treatment with increasing
concentrations of DT2216 (DT) and ABT199 (199, a BCL-2 specific inhibitor) alone or the combination of these two at equimolar concentrations (1:1)
as indicated is presented in the graph on the left. EC50 values for each of these treatments, the combination index (CI) at EC75 and EC90 values, and the
average CI are presented in the table on the right. Data are presented as meanꢀ ꢀs.d. from six replicate cell cultures in one representative experiment.
Similar results were obtained in one additional independent experiment. b, Percentage of viable EJM multiple myeloma cells after 72ꢀh treatment with
increasing concentrations of DT2216 (DT) and S63845 (S, a MCL-1 specific inhibitor) alone or the combination of these two at equimolar concentrations
(1:1) is presented in the graph on the left. EC50 values for each of these treatments, CI at EC75 and EC90 values, and the average CI are presented in the
table. Data are presented as meanꢀ ꢀs.d. from six replicate cell cultures in one representative experiment. Similar results were obtained in one additional
independent experiment. c, Illustration of the experimental design of H146 SCLC xenograft model. d, Body weight changes in H146 SCLC tumor-bearing
mice after the start of treatment with vehicle, DT2216, ABT263 or ABT199 alone or the combination of DT2216 (DT) and ABT199 (199) as shown in c.
Data are presented as meanꢀ ꢀs.e.m. (nꢀ=ꢀ6 mice per group at the start of treatment). e, Blood platelets were numerated 1 d after first treatment with all
the agents (left), and 1 d after last treatment with DT2216 or ABT263 and 22nd dose of ABT199 (right) as shown in c. Data are presented as meanꢀ ꢀs.e.m.
(nꢀ=ꢀ5 mice each for DT+199 in the left graph and vehicle in the right graph, nꢀ=ꢀ6 mice each in other groups). Each symbol represents data from an
individual animal. Statistical significance was determined by two-sided unpaired Student’s t-test. f, Changes in tumor volume over time after the start
of treatment as shown in c. Data are presented as meanꢀ ꢀs.e.m. (nꢀ=ꢀ6 mice per group at the start of treatment). a (Pꢀ=ꢀ0.0381 ABT263 versus vehicle,
Pꢀ=ꢀ0.0400 ABT199 versus vehicle); b (Pꢀ=ꢀ0.0069 versus vehicle, Pꢀ=ꢀ0.0152 versus ABT263, Pꢀ=ꢀ0.0044 versus ABT199); and c (Pꢀ=ꢀ0.0022 versus
vehicle, Pꢀ=ꢀ0.0001 versus DT2216, Pꢀ=ꢀ0.0003 versus ABT263, Pꢀ<ꢀ0.0001 versus ABT199) determined by two-sided unpaired Student’s t-test at post-
treatment day 27. g, The average wet weight of excised tumors from each group. Data are presented as meanꢀ ꢀs.e.m. (nꢀ=ꢀ5 mice in vehicle, and 6 mice
each in other groups). Each symbol represents data from an individual animal. a (Pꢀ=ꢀ0.0524 versus vehicle); b (Pꢀ=ꢀ0.0086 versus vehicle, Pꢀ=ꢀ0.0116
versus ABT263, Pꢀ=ꢀ0.0072 versus ABT199); and c (Pꢀ=ꢀ0.002 verus vehicle, Pꢀ<ꢀ0.0001 versus DT2216, Pꢀ=ꢀ0.0006 versus ABT263, Pꢀ<ꢀ0.0001 versus
ABT199) determined by two-sided unpaired Student’s t-test. h, A representative of two immunoblot analyses of BCL-XL, BCL-2 and MCL-1 in the H146
SCLC tumors excised at the end of experiment (nꢀ=ꢀ6 mice in each group). β-actin was used as a loading control. The uncropped immunoblot images
related to this figure are provided in separate source data file.
combined with ABT199 to more effectively treat tumors dependent cells to docetaxel, doxorubicin and vincristine in vitro (Fig. 6a). A
on both BCL-XL and BCL-2 than either agent alone or ABT263, similar chemosensitizing effect was also observed in PC-3 prostate,
without causing significant platelet toxicity. Of note, the combina- HepG2 liver and SW620 colon cancer cells (Supplementary Table 5),
tion of DT2216 and S63845, a selective MCL-1 inhibitor, was lethal but not in 786-O RCC cells (Fig. 6b), which lack VHL expression
to mice because the liver cells are also dependent on both BCL-XL (Fig. 2d and Extended Data Fig. 1b).
and MCL-1 for survival53, which limits the use of this combination
as a systemic therapy to treat cancer.
To determine whether DT2216 can sensitize MDA-MB-231 BC
cells to docetaxel in vivo, we generated MDA-MB-231 BC xeno-
grafts in mice and then treated the mice with vehicle, docetaxel
Synergy between DT2216 and conventional chemotherapy. High and/or DT2216 as shown in Fig. 6c. We found that DT2216 alone
levels of BCL-XL expression are associated with chemotherapy resis- had minimal effect on tumor growth, while docetaxel was able to
tance across multiple tumor types25,26, and inhibition of BCL-XL with substantially inhibit it. However, the combination of docetaxel and
ABT263 can improve the therapeutic efficacy of various chemo- DT2216 was more effective at suppressing the tumor growth than
therapeutic agents24,54,55. However, the on-target thrombocytopenia docetaxel alone without causing significant changes in body weight
resulting from BCL-XL inhibition prevents the use of combination (Fig. 6d,e).
therapy of ABT263 with other cytotoxic chemotherapeutic agents
Patient-derived xenograft (PDX) tumor models can better reca-
in clinic24,54,55. To evaluate the potential of DT2216 to overcome che- pitulate tumor biology of human diseases than conventional tumor
moresistance, we evaluated the ability of this agent to enhance the xenograft models employing cancer cell lines can. They are also
effects of chemotherapy. In these experiments, DT2216 sensitized more predictive of clinical outcomes of experimental therapeutic
drug-resistant triple-negative MDA-MB-231 breast cancer (BC) agents than the latter56. Therefore, we used CUL76 T-ALL PDX to
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
a
b
MDA-MB-231
MDA-MB-231
786-O
MDA-MB-231
120
120
100
80
60
40
20
0
120
100
80
60
40
20
0
120
100
80
60
40
20
0
Doxorubicin (DOX)
DT2216 (DT)
ABT263 (ABT)
DOX+DT
Vincristine (VIN)
DT2216 (DT)
ABT263 (ABT)
VIN+DT
Docetaxel (DTX)
DT2216 (DT)
ABT263 (ABT)
DTX+DT
Docetaxel (DTX)
100
80
60
40
20
0
DT2216 (DT)
ABT263 (ABT)
DTX+DT
DOX+ABT
VIN+ABT
DTX+ABT
DTX+ABT
0.001 0.01 0.1
1
10 100
0.001 0.01 0.1
1
10 100
0.001 0.01 0.1
1
10 100
0.001 0.01 0.1
1
10 100
Concentration (µM)
Concentration (µM)
Concentration (µM)
Concentration (µM)
e
c
d
Treatments and observations
MDA-MB-231 BC engraftment
–10
Time
(d)
30
25
20
15
10
5
–20
0
10
20
30
35
Veh
1,000
Docetaxel (DTX)
DT2216 (DT)
800
600
400
200
0
Vehicle
DT+DTX
Docetaxel (DTX)
DT2216 (DT)
DT+DTX
Treatment started (DTX was started 2 d after DT)
1. Vehicle
2. Docetaxel (DTX, 7.5 mpk, q14d, i.v.)
3. DT2216 (DT, 15 mpk, q4d, i.p.)
4. DT+DTX
NOD-SCID mice
5 × 106 MDA-MB-231
cells, s.c.
Termination
0
0
10
20
30
40
0
10
20
30
40
Post-treatment time (d)
Post-treatment time (d)
g
f
h
Treatments and observations
PDX T-ALL engraftment
Median survival
(d)
Veh
100
Time
(d)
0
10
20
40
50
80
DT+199
DT+Chemo
199+Chemo
ABT199
DT2216
Chemo
100
50
Veh
47
ABT199 40
DT2216 55
Chemo 47
DT+199 54
DT+Chemo 72
199+Chemo 47
80
60
40
20
0
=
0
.
2
=
0.4516
P
P
Treatment
D14–D46
Treatment started
Termination
250 cGY
irradiation
Treatment
stopped
1. Vehicle
0
1
2
3
4
2. ABT199 (100 mpk, qd, p.o.)
3. DT2216 (15 mpk, q4d, i.p.)
4. Chemotherapy (Chemo)*
5. DT+199
6. DT+Chemo
7. 199+Chemo
0
20
40
60
80
NSG mice
1 × 106 CUL76
cells, i.v.
Time after initial treatments (weeks)
Time after engraftment (d)
*Vincistrine 0.15 mpk, dexamethasone 5 mpk,
and L-asparaginase 1,000 Upk, i.p., q7d
Fig. 6 | Synergy of DT2216 with chemotherapy. a, Percentage of viable MDA-MB-231 triple-negative breast cancer cells after 72ꢀh of treatment with
increasing concentrations of DT2216, ABT263, docetaxel (DTX), doxorubicin (DOX) or vincristine (VIN) alone or the combination of DT2216 (DT) or
ABT263 (ABT) with one of the chemotherapeutic agents. Data are presented as meanꢀ ꢀs.d. from three replicate cell cultures in one representative
experiment. Similar results were obtained in two additional independent experiments for DTX. b, Percentage of viable VHL-null 786-O renal cell carcinoma
cells after 72ꢀh treatment with docetaxel (DTX), DT2216 or ABT263 alone, or the combination of DT2216 (DT) or ABT263 (ABT) with DTX as indicated.
Data are presented as meanꢀ ꢀs.d. from three replicate cell cultures of a single experiment. For the combination treatment, cells were treated with an
equimolar ratio of two drugs (1:1), except the combination of DTX with DT2216 or ABT263 in MDA-MB-231 cells, where the ratios of DTX and DT2216 or
ABT263 were 1:10. c, Illustration of the experimental design of MDA-MB-231 breast cancer xenograft model. d,e, Body weight and tumor volume changes in
MDA-MB-231 tumor-bearing mice after the start of treatment with vehicle, DTX, DT2216 or the combination of these two as shown in c. Data are presented
as meanꢀ ꢀs.e.m. (nꢀ=ꢀ10 mice in each group at the start of treatment). Statistical significance was determined by two-sided unpaired Student’s t-test at
post-treatment day 26. f, Illustration of the experimental design of CUL76 T-ALL PDX model. Upk, units per kilogram body weight. g, Percentage of CUL76
T-ALL cells in mouse blood collected at various times after the initiation of the treatments as shown in f. Data are presented as meanꢀ ꢀs.e.m. (nꢀ=ꢀ5 mice in
each group at the start of treatment). Each symbol represents data from an individual animal, and the middle horizontal line represents the mean. h, Kaplan–
Meier survival curve showing the survival of mice after CUL76 T-ALL engraftment. Median survival time within each treatment group is presented along
with statistical analysis results (nꢀ=ꢀ5 mice in each group at the start of treatment). Statistical significance was determined by log-rank (Mantel–Cox) test.
validate whether DT2216 can be combined with ABT199 or chemo- PROTAC technology. Specifically, we show that we can convert a
therapy to more effectively inhibit the growth of tumor cells from non-selective BCL-2 and BCL-XL inhibitor that is highly toxic to
people with relapsed or refractory T-ALL. CUL76 T-ALL PDX was platelets into a BCL-XL-specific PROTAC with significantly reduced
highly resistant to conventional chemotherapy and expressed high platelet toxicity. This was achieved by linking the BCL-2/BCL-XL
levels of BCL-XL, BCL-2 and MCL-1 (Fig. 6f and Supplementary binding moiety derived from ABT263 to a VHL-L with an empiri-
Fig. 1a,c). This PDX grew rapidly after being transplanted into cally optimized linker. The resulting BCL-XL-specific PROTAC lead
NOD-SCID IL2Rgnull (NSG) mice (Fig. 6g and Supplementary compound, DT2216, can selectively induce BCL-XL degradation in
Fig. 1b). Mice with CUL76 T-ALL PDX had a median survival time various tumor cells, but not in platelets, because it targets BCL-XL to
of 47, 40 and 47 d after the initiation of treatments with vehicle, the VHL E3 ligase that is minimally expressed in platelets.
ABT199 and a standard T-ALL chemotherapy regimen consisting of
These findings may have important clinical implications. First,
vincristine, dexamethasone and l-asparaginase (VDL), respectively BCL-XL is a well-validated cancer target. However, the on-target and
(Fig. 6h). Treatment with DT2216 alone or DT2216 plus ABT199 dose-limiting thrombocytopenia induced by BCL-XL inhibition has
moderately prolonged the survival of the mice, whereas the combi- limited the clinical use of BCL-XL inhibitors2,4,20,21. By converting a
nation of DT2216 with VDL chemotherapy substantially increased BCL-2/BCL-XL dual inhibitor into a BCL-XL-specific PROTAC, we
survival (Fig. 6h). The extension of survival correlated with reduced developed a safer and more effective BCL-XL-targeting antitumor
circulating tumor burden of CUL76 T-ALL cells serially measured agent. Our lead BCL-XL degrader, DT2216, significantly reduces the
by flow cytometry in the blood (Fig. 6g and Supplementary Fig. 1c). on-target and dose-limiting platelet toxicity resulting from conven-
Similar results were also observed in two additional T-ALL PDX mod- tional BCL-XL inhibition. In addition, BCL-XL-specific PROTACs
els (Extended Data Fig. 10a–d). These findings confirm that DT2216 should have another advantage compared to BCL-2/BCL-XL dual
can sensitize drug-resistant tumor cells to chemotherapy in vivo.
inhibitors or BCL-XL monoinhibitors, because PROTACs act cata-
lytically to induce protein degradation in a substoichiometric man-
Discussion
ner, and their effect is not limited by equilibrium occupancy27–31
.
Here we report a novel strategy to reduce the on-target and Therefore, it is not surprising that DT2216 is more potent than
dose-limiting normal tissue toxicity of an antitumor agent using ABT263 in inducing tumor cell apoptosis in vitro and inhibiting
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
10. Delbridge, A. R., Grabow, S., Strasser, A. & Vaux, D. L. Tirty years of BCL-2:
translating cell death discoveries into novel cancer therapies. Nat. Rev. Cancer
16, 99–109 (2016).
tumor growth in vivo. These findings suggest that DT2216 has
greater clinical potential than ABT263 or other BCL-XL inhibitors.
However, it is unlikely that DT2216 will be totally devoid of platelet
toxicity, because it retains a relatively reasonable binding affinity to
BCL-XL, and thus can still function as a moderate BCL-XL inhibitor
that could potentially induce mild thrombocytopenia if concentra-
tions are high enough. Therefore, identifying effective and tolerable
DT2216 doses and a schedule for administration will be an impor-
tant task in future clinical-oncology trials.
11. Delbridge, A. R. & Strasser, A. Te BCL-2 protein family, BH3-mimetics and
cancer therapy. Cell Death Difer. 22, 1071–1080 (2015).
12. Oltersdorf, T. et al. An inhibitor of Bcl-2 family proteins induces regression
of solid tumours. Nature 435, 677–681 (2005).
13. Tse, C. et al. ABT-263: a potent and orally bioavailable Bcl-2 family inhibitor.
Cancer Res. 68, 3421–3428 (2008).
14. Souers, A. J. et al. ABT-199, a potent and selective BCL-2 inhibitor, achieves
antitumor activity while sparing platelets. Nat. Med. 19, 202–208 (2013).
15. Tao, Z. F. et al. Discovery of a potent and selective BCL-XL inhibitor with
in vivo activity. ACS Med. Chem. Lett. 5, 1088–1093 (2014).
16. Kotschy, A. et al. Te MCL1 inhibitor S63845 is tolerable and efective in
diverse cancer models. Nature 538, 477–482 (2016).
Interestingly, while ABT263 inhibits both BCL-XL and BCL-2,
DT2216 is BCL-XL-specific and does not induce BCL-2 degradation.
In fact, proteomic analysis demonstrated that DT2216 is highly spe-
cific for BCL-XL degradation. Similarly, the pan-bromodomain and
extra-terminal (BET) inhibitor JQ1, which binds and inhibits the
BET proteins bromodomain-containing protein 2 (BRD2), BRD3
and BRD4, has been converted into a selective BRD4 PROTAC57.
Others have shown that a promiscuous ligand for multi-kinases
could induce only a subset of its targets for degradation when con-
verted into PROTACs38,58. As seen with other selective PROTACs
reported previously, the lack of BCL-2 degradation by DT2216 is at
least partially attributed to its inability to form a stable target pro-
tein–PROTAC–E3 ligase ternary complex in cells44,45. We anticipate
that because DT2216 is a specific BCL-XL PROTAC, it will have lim-
ited effect on cancers that depend on both BCL-XL and BCL-2 for
survival24. However, we showed that this hurdle can be overcome by
combining DT2216 with other selective inhibitors of BCL-2 family
proteins (such as ABT199) or with standard chemotherapy.
17. Deeks, E. D. Venetoclax: frst global approval. Drugs 76, 979–987 (2016).
18. Roberts, A. W. et al. Targeting BCL2 with venetoclax in relapsed chronic
lymphocytic leukemia. N. Engl. J. Med. 374, 311–322 (2016).
19. Mason, K. D. et al. Programmed anuclear cell death delimits platelet life span.
Cell. 128, 1173–1186 (2007).
20. Schoenwaelder, S. M. et al. Bcl-xL-inhibitory BH3 mimetics can induce a
transient thrombocytopathy that undermines the hemostatic function of
platelets. Blood 118, 1663–1674 (2011).
21. Kaefer, A. et al. Mechanism-based pharmacokinetic/pharmacodynamic
meta-analysis of navitoclax (ABT-263) induced thrombocytopenia.
Cancer Chemother. Pharmacol. 74, 593–602 (2014).
22. Itchaki, G. & Brown, J. R. Te potential of venetoclax (ABT-199) in chronic
lymphocytic leukemia. Adv. Hematol. 7, 270–287 (2016).
23. Perini, G. F., Ribeiro, G. N., Neto, J. V. P., Campos, L. T. & Hamerschlak, N.
BCL-2 as therapeutic target for hematological malignancies. J. Hematol.
Oncol. 11, 65 (2018).
24. Leverson, J. D. et al. Exploiting selective BCL-2 family inhibitors to dissect
cell survival dependencies and defne improved strategies for cancer therapy.
Sci. Transl. Med. 7, 279ra40 (2015).
Our results present the proof of principle for the use of tissue-
specific E3 ligases to direct tissue- or disease-specific degradation 25. Amundson, S. A. et al. An informatics approach identifying markers
of chemosensitivity in human cancer cell lines. Cancer Res. 60,
of a target protein using PROTAC technology. We demonstrate here
6101–6110 (2000).
that this strategy could rescue BCL-XL as an anticancer target from
26. Vogler, M. Targeting BCL2-proteins for the treatment of solid tumours.
its on-target and dose-limiting toxicity. This approach could also be
J. Adv Med. 2014, 943648 (2014).
applied to convert other toxic antitumor agents into tumor-selective
PROTACs by engaging E3 ligases more abundantly expressed in
tumor cells than in normal tissues.
27. Lai, A. C. & Crews, C. M. Induced protein degradation: an emerging drug
discovery paradigm. Nat. Rev. Drug Discov. 16, 101–114 (2017).
28. Runcie, A. C., Chan, K. H., Zengerle, M. & Ciulli, A. Chemical genetics
approaches for selective intervention in epigenetics. Curr. Opin. Chem. Biol.
33, 186–194 (2016).
Online content
29. Deshaies, R. J. Protein degradation: prime time for PROTACs. Nat. Chem.
Biol. 11, 634–635 (2015).
Any methods, additional references, Nature Research reporting
summaries, source data, extended data, supplementary informa-
tion, acknowledgements, peer review information, details of author
contributions and competing interests, and statements of code and
data availability are available at https://doi.org/10.1038/s41591-019-
0668-z.
30. Churcher, I. Protac-induced protein degradation in drug discovery: breaking
the rules or just making new ones? J. Med. Chem. 61, 444–452 (2018).
31. Ohoka, N., Shibata, N., Hattori, T. & Naito, M. Protein knockdown
technology: application of ubiquitin ligase to cancer therapy. Curr. Cancer
Drug Targets 16, 136–146 (2016).
32. Lu, J. et al. Hijacking the E3 ubiquitin ligase cereblon to efciently target
BRD4. Chem. Biol. 22, 755–763 (2015).
33. Bondeson, D. P. et al. Catalytic in vivo protein knockdown by small-molecule
PROTACs. Nat. Chem. Biol. 11, 611–617 (2015).
Received: 11 February 2019; Accepted: 28 October 2019;
Published: xx xx xxxx
34. Lai, A. C. et al. Modular PROTAC design for the degradation of oncogenic
BCR-ABL. Angew. Chem. Int. Ed. 55, 807–810 (2016).
References
35. Raina, K. et al. PROTAC-induced BET protein degradation as a therapy
for castration-resistant prostate cancer. Proc. Natl Acad. Sci. USA 113,
7124–7129 (2016).
1. Hanahan, D. & Weinberg, R. A. Hallmarks of cancer: the next generation.
Cell. 144, 646–674 (2011).
2. Singh, R., Letai, A. & Sarosiek, K. Regulation of apoptosis in health and
disease: the balancing act of BCL-2 family proteins. Nat. Rev. Mol. Cell Biol.
20, 175–193 (2019).
36. Saenz, D. T. et al. Novel BET protein proteolysis-targeting chimera exerts
superior lethal activity than bromodomain inhibitor (BETi) against
post-myeloproliferative neoplasm secondary (s) AML cells. Leukemia 31,
1951–1961 (2017).
3. Igney, F. H. & Krammer, P. H. Death and anti-death: tumour resistance to
apoptosis. Nat. Rev. Cancer 2, 277–288 (2002).
37. Winter, G. E. et al. Drug development. Phthalimide conjugation
as a strategy for in vivo target protein degradation. Science 348,
1376–1381 (2015).
4. Ashkenazi, A., Fairbrother, W. J., Leverson, J. D. & Souers, A. J. From basic
apoptosis discoveries to advanced selective BCL-2 family inhibitors.
Nat. Rev. Drug Discov. 16, 273–284 (2017).
38. Huang, H. T. et al. A chemoproteomic approach to query the degradable
kinome using a multi-kinase degrader. Cell Chem. Biol. 25, 88–99.e6 (2018).
39. Bray, P. F. et al. Te complex transcriptional landscape of the anucleate
human platelet. BMC Genomics 14, 1 (2013).
5. Adams, J. M. & Cory, S. Te Bcl-2 apoptotic switch in cancer development
and therapy. Oncogene 26, 1324–1337 (2007).
6. Reed, J. C. Bcl-2-family proteins and hematologic malignancies: history and
future prospects. Blood 111, 3322–3330 (2008).
40. Kissopoulou, A., Jonasson, J., Lindahl, T. L. & Osman, A. Next generation
sequencing analysis of human platelet PolyA+ mRNAs and rRNA-depleted
total RNA. PLoS One 8, e81809 (2013).
7. Tomas, S. et al. Targeting the Bcl-2 family for cancer therapy. Expert Opin.
Ter. Targets 17, 61–75 (2013).
8. Opfermann, J. T. Attacking cancer’s Achilles heel: antagonism of antiapoptotic
BCL-2 family members. FEBS J. 283, 2661–2675 (2016).
9. Garner, T. P., Lopez, A., Reyna, D. E., Spitz, A. Z. & Gavathiotis, E. Progress
in targeting the BCL-2 family of proteins. Curr. Opin. Chem. Biol. 39,
133–142 (2017).
41. Cerami, E. et al. Te cBio cancer genomics portal: an open platform
for exploring multidimensional cancer genomics data. Cancer Discov. 2,
401–404 (2012).
42. Gao, J. et al. Integrative analysis of complex cancer genomics and clinical
profles using the cBioPortal. Sci. Signal. 6, pl1 (2013).
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
52. Berger, S. et al. Computationally designed high specifcity inhibitors delineate
the roles of BCL2 family proteins in cancer. eLife. 5, e20352 (2016).
53. Hikita, H. et al. Mcl-1 and Bcl-xL cooperatively maintain integrity of
hepatocytes in developing and adult murine liver. Hepatology 50, 1217–1226.
54. Chen, J. et al. Te Bcl-2/Bcl-X(L)/Bcl-w inhibitor, navitoclax, enhances the
activity of chemotherapeutic agents in vitro and in vivo. Mol. Cancer Ter. 10,
2340–2349 (2011).
43. Vogler, M. et al. BCL2/BCL-X(L) inhibition induces apoptosis, disrupts
cellular calcium homeostasis, and prevents platelet activation. Blood. 117,
7145–7154 (2011).
44. Gadd, M. S. et al. Structural basis of PROTAC cooperative recognition for
selective protein degradation. Nat. Chem. Biol. 13, 514–521 (2017).
45. Nowak, R. P. et al. Plasticity in binding confers selectivity in ligand-induced
protein degradation. Nat. Chem. Biol. 14, 706–714 (2018).
46. Riching, K. M. et al. Quantitative live-cell kinetic degradation and
mechanistic profling of PROTAC mode of action. ACS Chem. Biol. 13,
2758–2770 (2018).
47. Farmer, T., O’Neill, K. L., Naslavsky, N., Luo, X. & Caplan, S. Retromer
facilitates the localization of Bcl-xL to the mitochondrial outer membrane.
Mol. Biol. Cell 30, 1138–1146 (2019).
48. Smith, B. E. et al. Diferential PROTAC substrate specifcity dictated by
orientation of recruited E3 ligase. Nat. Commun. 10, 131 (2019).
49. Morowski, M. et al. Only severe thrombocytopenia results in
bleeding and defective thrombus formation in mice. Blood 121,
4938–4947 (2013).
55. Ackler, S. et al. Te Bcl-2 inhibitor ABT-263 enhances the response of
multiple chemotherapeutic regimens in hematologic tumors in vivo. Cancer
Chemother. Pharmacol. 66, 869–880 (2010).
56. Pompili, L., Porru, M., Caruso, C., Biroccio, A. & Leonetti, C. Patient-derived
xenografs: a relevant preclinical model for drug development. J. Exp. Clin.
Cancer Res. 35, 189 (2016).
57. Zengerle, M., Chan, K. H. & Ciulli, A. Selective small molecule induced
degradation of the BET bromodomain protein BRD4. ACS Chem. Biol. 10,
1770–1777 (2015).
58. Bondeson, D. P. et al. Lessons in PROTAC design from selective degradation
with a promiscuous warhead. Cell Chem Biol. 25, 78–87.e5 (2018).
50. Rinder, H. M. et al. Correlation of thrombosis with increased platelet
turnover in thrombocytosis. Blood 91, 1288–1294 (1998).
51. Koch, R. et al. Biomarker-driven strategy for MCL1 inhibition in T-cell
lymphomas. Blood 133, 566–575 (2018).
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
© The Author(s), under exclusive license to Springer Nature America, Inc. 2019
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
(2.0ml, 26.1mmol). The mixture was stirred at room temperature for 1h, and
solvents were removed under reduced pressure. The solid residue was washed with
diethyl ether to afford the title compound 7 (752mg, yield 100%) as a white solid.
1H NMR (400MHz, CDCl3): δ 7.41–7.33 (m, 2H), 7.31–7.26 (m, 2H), 7.23–7.15
(m, 1H), 4.74 (s, 2H), 3.73–3.41 (m, 4H), 3.20–2.66 (m, 5H), 2.58–2.28 (m, 6H),
1.84–1.57 (m, 2H). LC-MS (ESI): m/z 440.1 [M+H]+.
Methods
Chemical synthesis. Te chemical structures and synthetic schemes for DT2216
and DT2216NC are presented in Extended Data Fig. 1d. Detailed synthetic
procedures are provided below.
General methods. Tetrahydrofuran (THF), dichloromethane (DCM), toluene
and acetonitrile were obtained via a solvent-purification system by filtration
through two columns packed with activated alumina and a 4-Å molecular sieve,
respectively. All other chemicals obtained from commercial sources were used
without further purification. Flash chromatography was performed using silica
gel (230–400 mesh) as the stationary phase. Reaction progress was monitored by
thin-layer chromatography (silica-coated glass plates) and visualized by ultraviolet
(UV) light, and/or by LC–MS. Nuclear magnetic resonance (NMR) spectra were
recorded in CDCl3 at 400MHz for 1H NMR. Chemical shifts (δ) are given in ppm
using tetramethylsilane as an internal standard. Multiplicities of NMR signals
are designated as singlet (s), broad singlet (br s), doublet (d), doublet of doublets
(dd), triplet (t), quartet (q) and multiplet (m). All final compounds for biological
testing were of ≥98.0% purity as analyzed by LC–MS, performed on an Advion
AVANT LC system with the expression compact mass spectrometer using a
Thermo Accucore Vanquish C18+UHPLC Column (1.5µm, 50×2.1mm) at 40°C.
Gradient elution was used for UHPLC with a mobile phase of acetonitrile and
water containing 0.1% formic acid.
Preparation of 2,2,2-trichloroethyl (R)-4-(4-(phenylthio)-3-((4-sulfamoyl-2-
((trifluoromethyl)sulfonyl)phenyl)amino)butyl)piperazine-1-carboxylate (compound
9). A mixture of compounds 7 (752mg, 1.36mmol) and 8 (417mg, 1.36mmol)
and TEA (945μl, 6.80mmol) in acetonitrile (20ml) was stirred under reflux for
4h. Solvents were evaporated under reduced pressure and the crude product
was purified by silica gel flash column chromatography using ethyl acetate and
hexanes as eluents to afford the title compound (780mg, yield 79%) as a white
solid. 1H NMR (400MHz, CDCl3) δ 8.24 (d, J=2.2Hz, 1H), 7.84 (d, J=9.1Hz,
1H), 7.42–7.37 (m, 2H), 7.36–7.27 (m, 3H), 7.05 (d, J=8.6Hz, 1H), 6.65 (br s,
1H), 5.13 (br s, 2H), 4.76 (s, 2H), 4.02–3.88 (m, 1H), 3.75–3.40 (m, 4H), 3.16–2.97
(m, 2H), 2.82–2.26 (m, 6H), 2.19–2.05 (m, 1H), 1.85–1.77 (m, 1H). LC–MS (ESI):
m/z 727.0 [M+H]+.
Preparation of 2,2,2-trichloroethyl (R)-4-(3-((4-(N-(4-(4-((4′-chloro-4,4-dimethyl-
3,4,5,6-tetrahydro-[1,1′-biphenyl]-2-yl)methyl)piperazin-1-yl)benzoyl)sulfamoyl)-
2-((trifluoromethyl)sulfonyl)phenyl)amino)-4-(phenylthio)butyl)piperazine-1-
carboxylate (compound 11). A mixture of compounds 9 (780mg, 1.07mmol) and
10 (470mg, 1.07mmol), 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI)
(411mg, 2.14mmol) and 4-dimethylaminopyridine (DMAP) (262mg, 2.14mmol)
in DCM (40ml) was stirred at room temperature overnight. Solvent was
evaporated under reduced pressure, and the crude product was purified by silica
gel flash column chromatography using DCM and methanol (MeOH) as eluents to
afford the title compound (859mg, yield 70%) as a white solid. 1H NMR (400MHz,
CDCl3) δ 8.37 (d, J=2.0Hz, 1H), 8.13 (d, J=9.2Hz, 1H), 7.62 (d, J=8.9Hz,
2H), 7.41–7.27 (m, 6H), 7.12 (d, J=8.7Hz, 1H), 6.98 (d, J=8.3Hz, 2H), 6.79 (d,
J=9.0Hz, 2H), 6.58 (d, J=9.4Hz, 1H), 4.75 (s, 2H), 3.95–3.81 (m, 1H), 3.63–3.38
(m, 4H), 3.33–3.23 (m, 4H), 3.11 (dd, J=13.8, 4.9Hz, 1H), 3.00 (dd, J=13.9,
7.5Hz, 1H), 2.81 (s, 2H), 2.46–2.01 (m, 15H), 1.75–1.65 (m, 1H), 1.46 (t, J=6.3Hz,
2H), 0.98 (s, 6H) ppm. LC–MS (ESI): m/z 1,147.1 [M+H]+.
Preparation of (R)-3-((tert-butoxycarbonyl)amino)-4-hydroxybutyric acid benzyl
ester (compound 2). N-Methylmorpholine (4.41ml, 40.1mmol) and isobutyl
chloroformate (4.43ml, 34.2mmol) were added into a stirred solution of N-Boc-
d-aspartic acid 4-benzyl ester (compound 1) (10.0g, 30.9mmol) in THF (250ml)
at −25°C. The resulting mixture was stirred at −25°C for 30min and allowed to
warm to −15°C. A solution of NaBH4 (2.94g, 77.7mmol) in water (100ml) was
then added to the mixture in one portion resulting in evolution of gas. The mixture
was stirred for 30min at −15°C and quenched with 1N HCl (aq.). The solution
was extracted with ethyl acetate three times, and the combined organic layers were
washed with brine, dried over Na2SO4, filtered and evaporated to dryness under
reduced pressure. The crude product was used directly in the next step. 1H NMR
(400MHz, CDCl3) δ 7.48–7.29 (m, 5H), 5.17 (br s, 1H), 5.12 (s, 2H), 4.05–3.93 (m,
1H), 3.69 (t, coupling constant (J)=5.2Hz, 2H), 2.67 (d, J=6.1Hz, 2H), 2.38 (br s,
1H), 1.42 (s, 9H). LC–MS (with electrospray ionization, ESI): m/z 310.3 [M+H]+.
Preparation of (R)-4-(4-((4′-chloro-4,4-dimethyl-3,4,5,6-tetrahydro-[1,1′-biphenyl]-
2-yl)methyl)piperazin-1-yl)-N-((4-((1-(phenylthio)-4-(piperazin-1-yl)butan-2-yl)
amino)-3-((trifluoromethyl)sulfonyl)phenyl)sulfonyl)benzamide (compound 12).
Zinc powder (960mg, 14.8mmol) was added to a mixture of compound 11
(316mg, 0.28mmol) and acetic acid (600μl, 10.5mmol) in THF (20ml). The
reaction mixture was stirred at room temperature for 5h. The solid was removed
by filtration, and the filtrate was poured into water and extracted with ethyl acetate.
The combined organic phases were washed with brine, dried over Na2SO4, filtered
and evaporated to dryness under reduced pressure. The crude product was purified
by silica gel flash column chromatography using DCM, methanol and TEA as
eluents to afford the title compound (210mg, yield 78%). 1H NMR (400MHz,
CDCl3) δ 8.21 (s, 1H), 7.93 (d, J=9.2Hz, 1H), 7.85 (d, J=8.6Hz, 2H), 7.33–7.24
(m, 2H), 7.22–7.08 (m, 5H), 6.92 (d, J=8.3Hz, 2H), 6.77 (d, J=8.4Hz, 1H), 6.66
(d, J=8.7Hz, 2H), 6.46 (d, J=9.3Hz, 1H), 3.83–3.67 (m, 1H), 3.17–3.08 (m, 4H),
3.02–2.92 (m, 5H), 2.89–2.78 (m, 1H), 2.72 (s, 2H), 2.64–2.13 (m, 12H), 2.04–1.91
(m, 3H), 1.62–1.49 (m, 1H), 1.39 (t, J=6.3Hz, 2H), 0.91 (s, 6H) ppm. LC–MS
(ESI): m/z 973.2 [M+H]+.
Preparation of benzyl (R)-3-((tert-butoxycarbonyl)amino)-4-(phenylthio)butanoate
(compound 3). A mixture of compound 2 (30.9mmol), diphenyl disulfide (8.8g,
40.2mmol) and Bu3P (9.9ml, 40.2mmol) in toluene (150ml) was heated at
80°C under N2 overnight. The mixture was cooled to room temperature and
concentrated under reduced pressure. The crude product was purified by silica gel
flash column chromatography using ethyl acetate and hexanes as eluents to afford
the title compound (7.1g, yield 57% in two steps). 1H NMR (400MHz, CDCl3)
δ 7.44–7.09 (m, 10H), 5.15 (br s, 1H), 5.08 (s, 2H), 4.24–3.97 (m, 1H), 3.23 (dd,
J=13.7, 5.5Hz, 1H), 3.08 (dd, J=13.6, 7.3Hz, 1H), 2.80 (dd, J=16.2, 5.1Hz, 1H),
2.67 (dd, J=16.4, 5.7Hz, 1H), 1.40 (s, 9H). LC-MS (ESI): m/z 402.2 [M+H]+.
Preparation of tert-butyl N-[(2R)-4-oxo-1-(phenylsulfanyl)butan-2-yl]carbamate
(compound 4). Diisobutylaluminium hydride (DIBAL-H; 1.2M in toluene, 34.0ml,
40.8mmol) was added dropwise to a solution of compound 3 (7.1g, 17.7mmol)
in toluene (80ml) at −78°C and stirred for 3h. The reaction mixture was then
quenched with NH4Cl (aq.), warmed to room temperature, and diluted with
ethyl acetate. The resulting mixture was filtered, and the filtrate was poured
into water and extracted with ethyl acetate. The combined organic phases were
washed with brine, dried over Na2SO4, filtered and evaporated to dryness under
reduced pressure. The crude product was purified by silica gel flash column
chromatography using ethyl acetate and hexanes as eluents to afford the title
compound (4.16g, yield 80%). LC–MS (ESI): m/z 296.2 [M+H]+.
Preparation of ethyl 7-(((S)-1-((2S,4R)-4-hydroxy-2-(((S)-1-(4-(4-methylthiazol-5-yl)
phenyl)ethyl)carbamoyl)pyrrolidin-1-yl)-3,3-dimethyl-1-oxobutan-2-yl)amino)-7-
oxoheptanoate (compound 15a). Compounds 13a and 13b were prepared according
to reported procedures in the literature35. A mixture of compound 13a (100mg,
0.207mmol), 7-ethoxy-7-oxoheptanoic acid (compound 14) (44mg, 0.234mmol),
HATU (82mg, 0.216mmol) and TEA (160μl, 1.15mmol) in DCM was stirred
at room temperature for 1h. The reaction mixture was poured into water and
extracted with DCM. The combined organic layers were washed with aq. NH4Cl
solution and saline, dried over Na2SO4, and concentrated under vacuum. The
crude product was purified by silica gel column chromatography to afford the title
compound (105mg, yield 83%). 1H NMR (400MHz, CDCl3) δ 8.68 (s, 1H), 7.50–
7.32 (m, 5H), 6.26 (d, J=8.7Hz, 1H), 5.14–5.03 (m, 1H), 4.72 (t, J=7.9Hz, 1H),
4.59–4.46 (m, 2H), 4.14–4.08 (m, 3H), 3.61 (dd, J=11.3, 3.7Hz, 1H), 2.57–2.45 (m,
4H), 2.32–2.18 (m, 4H), 2.11–2.05 (m, 1H), 1.66–1.58 (m, 4H), 1.48 (d, J=6.9Hz,
3H), 1.36–1.22 (m, 5H), 1.04 (s, 9H) ppm. LC–MS (ESI): m/z 615.4 [M+H]+.
Preparation of 2,2,2-trichloroethyl (R)-4-(3-((tert-butoxycarbonyl)amino)-4-
(phenylthio)butyl)piperazine-1-carboxylate (compound 6). To a mixture of
compound 4 (592mg, 2.00mmol), compound 5 (753mg, 2.88mmol) and
triethylamine (TEA) (1.12ml, 8.05mmol) in DCM (15ml) was added to
NaBH(OAc)3 (638mg, 3.00mmol). The resulting solution was stirred at room
temperature overnight before being poured into water and extracted with DCM.
The combined organic phases were washed with brine, dried over Na2SO4, filtered
and evaporated to dryness under reduced pressure. The crude product was purified
by silica gel flash column chromatography using ethyl acetate and hexanes as
eluents to afford the title compound (733mg, yield 68%). 1H NMR (400MHz,
CDCl3) δ 7.43–7.36 (m, 2H), 7.32–7.27 (m, 2H), 7.19 (t, J=7.3Hz, 1H), 5.44 (br s,
1H), 4.76 (s, 2H), 3.99–3.84 (m, 1H), 3.72–3.49 (m, 4H), 3.23 (dd, J=13.3, 4.6Hz,
1H), 3.10–2.95 (m, 1H), 2.61–2.31 (m, 6H), 1.96–1.61 (m, 2H), 1.43 (s, 9H). LC–
MS (ESI): m/z 540.1 [M+H]+.
Preparation of ethyl 7-(((S)-1-((2R,4S)-4-hydroxy-2-(((S)-1-(4-(4-methylthiazol-5-yl)
phenyl)ethyl)carbamoyl)pyrrolidin-1-yl)-3,3-dimethyl-1-oxobutan-2-yl)amino)-7-
oxoheptanoate (compound 15b). Compound 15b was prepared using the procedure
described for the synthesis of compound 15a by using compound 13b instead of
compound 13a. Yield 80%. 1H NMR (400MHz, CDCl3) δ 8.65 (s, 1H), 7.35 (d,
J=9.2Hz, 5H), 6.13–6.03 (m, 1H), 5.11–5.00 (m, 1H), 4.68 (dd, J=8.5, 3.8Hz, 1H),
4.55–4.44 (m, 1H), 4.33 (d, J=7.1Hz, 1H), 4.13–4.04 (m, 3H), 3.59 (dd, J=10.4,
5.3Hz, 1H), 2.51 (s, 3H), 2.47–2.38 (m, 1H), 2.27–2.15 (m, 4H), 2.11–2.01 (m, 1H),
Preparation of 2,2,2-trichloroethyl (R)-4-(3-amino-4-(phenylthio)butyl)piperazine-
1-carboxylate trifluoroacetic acid salt (compound 7). To a mixture of compound
6 (733mg, 1.36mmol) in DCM (5ml) was added to trifluoroacetic acid (TFA)
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
1.62–1.53 (m, 4H), 1.40 (d, J=7.0Hz, 3H), 1.32–1.19 (m, 5H), 1.06 (s, 9H) ppm.
cat. no. 12430054, Thermo Fisher Scientific) with 10% FBS, 100U ml–1 penicillin
and 100µg ml–1 streptomycin. All the cell lines were maintained in a humidified
incubator at 37°C and 5% CO2.
LC–MS (ESI): m/z 615.5 [M+H]+.
Preparation of (2S,4R)-1-((S)-2-(7-(4-((R)-3-((4-(N-(4-(4-((4′-chloro-4,4-
dimethyl-3,4,5,6-tetrahydro-[1,1′-biphenyl]-2-yl)methyl)piperazin-1-yl)benzoyl)
sulfamoyl)-2-((trifluoromethyl)sulfonyl)phenyl)amino)-4-(phenylthio)butyl)
piperazin-1-yl)-7-oxoheptanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-((S)-
1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)pyrrolidine-2-carboxamide (DT2216).
Compound 15a (105mg, 0.171mmol) was dissolved in methanol (5ml) and treated
with LiOH monohydrate (43mg, 1.02mmol) in water (0.5ml) for 2h. The pH
of the reaction mixture was slowly adjusted to 5–6 with 1.0N HCl. The mixture
was concentrated directly under reduced pressure to afford the corresponding
acid, which was dissolved in DCM (5ml) and mixed with compound 12 (158mg,
0.162mmol), HATU (68mg, 0.179mmol) and TEA (113μL, 0.810mmol). The
resulting solution was stirred at room temperature for 1h before being poured
into water and extracted with DCM. The combined organic layers were washed
with NH4Cl solution (aq.) and saline, dried over Na2SO4, and concentrated under
reduced pressure. The crude product was purified by silica gel flash column
chromatography to afford DT2216 (115mg, yield 46%). 1H NMR (400MHz,
CDCl3) δ 8.67 (s, 1H), 8.34 (d, J=2.3Hz, 1H), 8.08 (d, J=9.1Hz, 1H), 7.69 (d,
J=8.6Hz, 2H), 7.46–7.27 (m, 12H), 7.14–6.93 (m, 3H), 6.76 (d, J=8.6Hz, 2H),
6.61 (d, J=9.4Hz, 1H), 6.34 (d, J=8.7Hz, 1H), 5.13–5.00 (m, 1H), 4.79–4.69 (m,
1H), 4.60 (d, J=8.7Hz, 1H), 4.49 (s, 1H), 4.10 (d, J=11.4Hz, 1H), 3.96–3.83 (m,
1H), 3.72–2.81 (m, 13H), 2.55–2.00 (m, 24H), 1.77–1.30 (m, 12H), 1.03 (s, 9H),
0.97 (s, 6H) ppm. LC–MS (ESI): m/z 1,542.0 [M+H]+.
Immunoblotting. Cells were treated with DT2216 and other compounds at
indicated concentrations and durations. Untreated and treated cells were harvested
and washed once with ice-cold phosphate-buffered saline, pH 7.2 (PBS, cat. no.
20012027; Thermo Fisher Scientific). Adherent cells were harvested using 0.25%
Trypsin–EDTA solution (cat. no. 25200056, Thermo Fisher Scientific) and washed
with ice-cold PBS. The cell pellets were lysed in radioimmunoprecipitation
assay (RIPA) lysis buffer (50mM Tris-HCl pH 7.4, 150mM NaCl, 1% NP-40,
0.5% sodium deoxycholate, 0.1%SDS, 5mM EDTA, 1mM EGTA; cat. no. BP-
115DG, Boston Bio Products, Ashland, MA, USA) supplemented with 1%
protease inhibitor cocktail (cat. no. P8340, Sigma-Aldrich, St. Louis, MO, USA)
and 1% phosphatase inhibitor cocktail (cat. no. P0044, Sigma-Aldrich). Protein
lysates were incubated on ice for 15min and then kept at −80°C overnight to
allow for complete lysis. The samples were thawed on ice and then centrifuged
at maximum speed (14,000r.p.m.) for 15min in a refrigerated centrifuge. The
protein concentration in the supernatants was determined using the Pierce
BCA protein Assay kit (cat. no. 23225, Thermo Fisher Scientific). The protein
concentration was normalized and the samples were reduced in 4× Laemmli’s
SDS-sample buffer (cat. no. BP-110R, Boston Bio Products) and denatured at 95°C
in a heated block. An equal amount of protein samples (20–40 μg per lane) were
resolved using precast 4–20% Tris-glycine gels (Mini-PROTEAN TGX, cat. no.
456–1094, Bio-Rad), and resolved proteins were transferred onto 0.2-µm pore size
polyvinylidene difluoride (PVDF) blotting membranes (cat. no. LC2002, Thermo
Fisher Scientific) using mini trans-blot electrophoretic transfer cell (Bio-Rad). The
membranes were blocked with non-fat dry milk (5% wt/vol) in 1× Tris-buffered
saline-Tween-20 (TBST, cat. no. J77500, Affymetrix) for 1h at room temperature,
and were subsequently probed with primary antibodies at a predetermined optimal
concentration in non-fat dry milk (5% wt/vol in TBST) overnight at 4°C. The
membranes were washed three times (5–10min each) in TBST and then incubated
with horse radish peroxidase (HRP)-conjugated secondary antibodies for 1–1.5h
at room temperature. Following sufficient washing with TBST, the membranes
were exposed with chemiluminescent HRP substrate (cat. no. WBKLS0500,
MilliporeSigma), and the signal was detected using autoradiography (SRX-101,
Konica) or the ChemiDoc MP Imaging System (Bio-Rad). The immunoblots
were quantified by densitometry using ImageJ software (NIH) and the data were
expressed as relative band intensities normalized to equal loading control.
Preparation of (2R,4S)-1-((S)-2-(7-(4-((R)-3-((4-(N-(4-(4-((4′-chloro-4,4-
dimethyl-3,4,5,6-tetrahydro-[1,1′-biphenyl]-2-yl)methyl)piperazin-1-yl)benzoyl)
sulfamoyl)-2-((trifluoromethyl)sulfonyl)phenyl)amino)-4-(phenylthio)butyl)
piperazin-1-yl)-7-oxoheptanamido)-3,3-dimethylbutanoyl)-4-hydroxy-N-((S)-1-
(4-(4-methylthiazol-5-yl)phenyl)ethyl)pyrrolidine-2-carboxamide (DT2216NC).
DT2216NC was prepared using the procedure described for the synthesis of
DT2216 by using compound 15b instead of compound 15a. Yield 40%. 1H NMR
(400MHz, CDCl3) δ 8.65 (s, 1H), 8.36 (d, J=1.9Hz, 1H), 8.11–8.04 (m, 1H), 7.69
(d, J=8.8Hz, 2H), 7.44–7.26 (m, 12H), 7.08 (d, J=8.5Hz, 1H), 6.98 (d, J=8.3Hz,
2H), 6.76 (d, J=9.0Hz, 2H), 6.62 (d, J=9.4Hz, 1H), 6.17 (d, J=6.9Hz, 1H), 5.13–
5.02 (m, 1H), 4.68 (dd, J=8.5, 3.9Hz, 1H), 4.54–4.44 (m, 1H), 4.34 (d, J=6.9Hz,
1H), 4.13–4.02 (m, 1H), 3.97–3.84 (m, 1H), 3.72–3.56 (m, 2H), 3.43–3.22 (m, 7H),
3.09–2.79 (m, 4H), 2.51–2.00 (m, 24H), 1.73–1.24 (m, 12H), 1.07 (d, J=4.8Hz,
9H), 0.98 (s, 6H) ppm. LC–MS (ESI): m/z 1,541.8 [M+H]+.
Viability assays in cancer cells. Cancer cells in complete cell culture medium
were seeded in 96-well plates (100µl per well) at the optimized densities
(50,000–100,000 suspension cells, 3,000–5,000 adherent cells). Suspension
cells were treated 30min after seeding, whereas adherent cells were allowed to
adhere overnight and were then treated. Compound treatments were prepared in
complete cell culture medium and 100µl of 2× treatment-containing medium were
added to each well. Complete cell culture medium without treatment was added
in control wells, and wells containing medium without cells served as background.
The outer wells of the 96-well plate were not used for treatment and were filled
with 200µl of PBS to reduce evaporation of medium from inner wells. Each
compound and/or combination was tested at nine different concentrations with
three to six replicates, unless otherwise specified. For combination treatments with
DT2216 and ABT199 or DT2216 and S63845, cells were treated with equimolar
concentrations in twofold serial dilutions. For combination treatment with
DT2216 and chemotherapeutics, cells were treated at equimolar concentrations
and threefold serial dilutions, unless otherwise specified. The cell viability was
measured after 72h by tetrazolium-based MTS assay. MTS reagent (2mg ml–1
stock; cat. no. G1111, Promega) was freshly supplemented with phenazine
methosulfate (PMS, 0.92mg ml–1 stock, cat. no. P9625, Sigma-Aldrich) in 20:1
ratio, and 20µl of this mixture was added to each control and treatment well.
The cells were incubated for 4 h at 37°C and 5%CO2, and then the absorbance
was recorded at 490nm using Biotek’s Synergy Neo2 multimode plate reader
(Biotek). The average absorbance value of background wells was subtracted from
absorbance value of each control and treatment well, and percent cell viability
((At / A0)×100)) was determined in each treatment well, where At is the
absorbance value of the treatment well and A0 is the average absorbance value of
control wells after background subtraction. The data were expressed as average
percentage cell viability and fitted in non-linear regression curves using GraphPad
Prism 7 (GraphPad Software, La Jolla, CA, USA).
Preparation of (2S,4R)-1-((S)-2-acetamido-3,3-dimethylbutanoyl)-4-hydroxy-N-
((S)-1-(4-(4-methylthiazol-5-yl)phenyl)ethyl)pyrrolidine-2-carboxamide (VHL-L).
To a solution of compound 13a (52mg, 0.108mmol) and TEA (42μl, 0.302mmol)
in DCM (5ml) was added acetic anhydride (10.4μl, 0.110mmol). The resulting
solution was stirred at room temperature for 1h. The reaction was quenched
with water (5ml) and extracted with DCM. The organic layer was washed with
NH4Cl solution (aq.) and saline, dried over Na2SO4 and concentrated under
reduced pressure. The crude product was purified by silica gel flash column
chromatography to afford VHL-L (41mg, yield 78%). 1H NMR (400MHz, CDCl3)
δ 8.69 (s, 1H), 7.47–7.34 (m, 5H), 6.34 (d, J=8.8Hz, 1H), 5.17–5.03 (m, 1H), 4.70
(t, J=7.9Hz, 1H), 4.61–4.45 (m, 2H), 4.06 (d, J=11.3Hz, 1H), 3.70–3.52 (m, 1H),
2.57–2.42 (m, 4H), 2.10–2.03 (m, 1H), 1.98 (s, 3H), 1.48 (d, J=6.8Hz, 3H), 1.05 (s,
9H) ppm. LC–MS (ESI): m/z 487.2 [M+H]+.
Chemical compounds and antibodies. The information for commercially
available chemical compounds, anticancer drugs and antibodies are provided in
Supplementary Tables 6 and 7.
Cell lines and cell culture. Human T-ALL MOLT-4 (cat. no. CRL-1582), B-ALL
RS4;11 (RS4, cat. no. CRL-1873), SCLC NCI-H146 (H146, cat. no. HTB-173),
breast cancer MDA-MB-231 (cat. no. HTB-26), prostate cancer PC-3 (cat. no. CRL-
1435), hepatocellular carcinoma HepG2 (cat. no. HB-8065), colorectal carcinoma
SW620 (cat. no. CCL-227), renal cell carcinoma 786-O (cat. no. CRL-1932), lung
fibroblasts WI-38 (cat. no. CCL-75) and epithelial kidney HEK 293T (293T, cat.
no. ACS-4500) cell lines were purchased from American Type Culture Collection
(ATCC). Human multiple myeloma cell lines (EJM and NCI-H929 (H929)) were a
kind gift from E. Tian at the Winthrop P. Rockefeller Cancer Institute at University
of Arkansas for Medical Sciences. MOLT-4, RS4 and H146 cell lines were
cultured in RPMI 1640 medium (cat. no. 22400–089, Thermo Fisher Scientific)
supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum (FBS, cat. no.
S11150H, Atlanta Biologicals, Flowery Branch, GA, USA), 100U ml–1 penicillin
and 100µg ml–1 streptomycin (penicillin–streptomycin, cat. no. 15140122, Thermo
Fisher Scientific). EJM cells were cultured in RPMI medium with 10%FBS, 100U
ml–1 penicillin and 100µg ml–1 streptomycin supplemented with 1% MarrowMAX
Bone Marrow Medium (cat. no. 12260014, Thermo Fisher Scientific). H929 cells
were cultured in RPMI medium with 10% FBS and 100U ml–1 penicillin and 100µg
ml–1 streptomycin supplemented with 0.05mM 2-mercaptoethanol. All other cell
lines were cultured in complete Dulbecco’s modified Eagle medium (DMEM,
Platelet viability assays. Human platelet-rich plasma (PRP) was purchased from
Zenbio (cat. no. SER-PRP-SDS). PRP was used for experiments immediately after
delivery. PRP was transferred into 50-ml polypropylene tubes, each containing
5ml acid citrate buffer (cat. no. sc-214744, Santa Cruz Biotechnology). To prevent
clotting, prostaglandin E1 (PGE1, cat. no. sc-201223A, Santa Cruz Biotechnology)
and apyrase (cat. no. A6237, Sigma-Aldrich) were added to final concentrations
of 1µM and 0.2 units per ml, respectively. After gently mixing the solution,
platelets were pelleted by centrifugation at 1,200g for 10min. Pelleted platelets
were gently washed without disruption in 2ml HEPES-buffered Tyrode’s solution
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
(10mM HEPES, 135mM NaCl, 2.8mM KCl, 1mM MgCl2, 2mM CaCl2, 12mM
NaHCO3, 0.4mM NaH2PO4, 0.25%BSA and 5.5mM glucose, pH 7.4; cat. no. PY-
921WB, Boston BioProducts) containing 1µM PGE1 and 0.2 units ml–1 apyrase.
After washing, pellets were slowly resuspended in 10ml HEPES-buffered Tyrode’s
solution containing 1µM PGE1 and 0.2 units ml–1 apyrase. The number of platelets
was counted using a HEMAVET 950FS hematology analyzer (Drew Scientific,
Inc.). For viability assays, platelet number was adjusted to 2×108 per ml in HEPES-
buffered Tyrode’s solution containing 1µM PGE1, 0.2 units ml–1 apyrase and 10%
FBS. Each treatment was given in 2ml platelet suspension in 15-ml polypropylene
tubes for 72h. The tubes were placed on a rotating platform at room temperature
for the duration of treatment. After 72h treatment, 200µl of untreated or treated
platelets were plated in each well of 96-well plates, and the viability was measured
by MTS assay as described in the previous method.
were performed. The assays were performed at room temperature and reagents
were diluted in a buffer containing 250mM HEPES pH 7.5, 1M NaCl, 1%BSA and
0.05% Tween-20. Purified recombinant histidine (His)-tagged BCL-XL (0.1nM,
cat. no. SRP0187, Sigma-Aldrich), BCL-2 (0.2nM, cat. no. SRP0186, Sigma-
Aldrich) or BCL-W (0.2nM, cat. no. B1059, Sigma-Aldrich) were incubated with
increasing concentrations of DT2216 or ABT263 and 15nM biotin-tagged BAD
(biotin-LWAAQRYGRELRRMSDEFEGSFKGL amino-terminal, AnaSpec) or
30nM BIM peptides (biotin-MRPEIWIAQELRRIGDEFNA N-terminal, AnaSpec)
to a final volume of 40µl in 96-well PCR plate. BAD peptide was used to assess the
binding affinity of compounds towards BCL-XL, whereas BIM peptide was used
for BCL-2 and BCL-W. After 24h incubation, 5µl 6× His-acceptor beads (cat. no.
AL128M, PerkinElmer) were added to each well at 20µg ml–1 final concentration
and incubated for 1h. Thereafter, 5µl streptavidin-donor beads were added (cat.
no. 6760002, PerkinElmer) to each well at 20µg ml−1 final concentration and
incubated for 30min. At the end of the incubation period, 17µl of each sample
was transferred in adjacent wells of 384-well proxy plate (cat. no. 6008280,
PerkinElmer). The plate was scanned using the Alpha program on Biotek’s Synergy
Neo2 multimode plate reader. The inhibition constant (Ki) was calculated using the
non-linear regression, one site, competitive binding, Fit Ki function on GraphPad
Prism 7 software, based on experimentally determined dissociation constant (KD)
for each protein–peptide pair.
Determination of EC50 and combination index. Dose–response curves (Variable
slope, four parameters) were generated for each test compound, and their EC50
values were determined using GraphPad Prism 7. The combination indices (CIs)
were determined using Compusyn version 1.0 software (http://www.combosyn.
com). CI<0.3 was considered to be a strong synergistic effect induced by two
compounds when they were used in combination. The average of CI at EC75
(concentration with 75% loss of cell viability) and EC90 (concentration with 90%
loss of cell viability) are presented in Fig. 5 and Supplementary Table 5 as a more
accurate measure of synergy by different combinations.
Proteomics analysis. Sample preparation and LC–MS/MS analysis. To label cells
with stable isotopic amino acids (SILAC), WI-38 cells were propagated in DMEM
SILAC medium defcient in both l-lysine and l-arginine (cat. no. 88364, Termo
Fisher Scientifc) and supplemented with light-labeled lysine ([12C614N2]Lys) and
arginine ([12C614 N4]Arg) for light state (cat. no. 89987 and no. 89989; Termo
Fisher Scientifc), and [13C615N2]Lys and [13C615N4]Arg for heavy-state labeling (cat.
no. 88209 and no. 89990, Termo Fisher Scientifc). Cells were cultured for at least
six doubling times for complete incorporation. Te light-labeled WI-38 cells were
untreated (DMSO), and the heavy-labeled WI-38 cells were treated with 1 μM
DT2216 or DT2216NC for 6h, respectively. Reverse labeling was used in the
second biological replicate. Te untreated and DT2216- or DT2216NC-treated
cells were collected by centrifugation at 500g for 5min. Pellets were washed twice
by resuspension in 1ml of ice-cold PBS. Te cell pellets were resuspended in 20ml
freshly prepared lysis bufer (2%SDS, 100mM Tris/HCl pH 7.6) containing MS-
SAFE Protease and Phosphatase Inhibitor (cat. no. MSSAFE-5VL, Sigma-Aldrich)
for sonication. Te lysate was centrifuged at 15,000g for 10min at 18°C. Te
supernatant was stored at −80°C for proteomics analysis. Protein concentration
was measured by BCA assay (cat. no. 23227, Termo Fisher Scientifc), and
SILAC pairs were mixed in equimolar amounts. Purifed proteins were reduced,
alkylated and digested using flter-aided sample preparation60. Tryptic peptides
were separated into 36 fractions on a 100×1.0mm Acquity BEH C18 column (cat.
no. 186002350, Waters) using an UltiMate 3000 UHPLC system (Termo Fisher
Scientifc) with a 40-min gradient from 99:1 to 60:40 bufer A (0.1% formic acid,
0.5% acetonitrile):B (0.1% formic acid, 99.9% acetonitrile) ratio under basic pH
conditions, and then consolidated into 12 super-fractions. Each super-fraction was
then further separated by reverse phase XSelect CSH C18 2.5-μm resin (cat. no.
186006109, Waters) on an in-line 150×0.075mm column using an UltiMate 3000
RSLCnano system (Termo Fisher Scientifc). Peptides were eluted using a 60-min
gradient from 97:3 to 60:40 bufer A:B ratio. Eluted peptides were ionized by
electrospray (2.15kV) followed by MS/MS analysis using higher-energy collisional
dissociation (HCD) on an Orbitrap Fusion Lumos mass spectrometer (Termo
Fisher Scientifc) in top-speed data-dependent mode. MS data were acquired using
the FTMS analyzer in profle mode at a resolution of 240,000 over a range of 375
to 1,500m/z. Following HCD activation, MS/MS data were acquired using the ion
trap analyzer in centroid mode and normal mass range with precursor mass-
dependent normalized collision energy between 28.0 and 31.0.
Caspase-3 activity assay. The caspase-3 activity in untreated and treated cells
was measured using EnzChek caspase-3 assay kit no. 2, Z-DEVD-R110 substrate
(cat. no. E13184, Thermo Fisher Scientific) following manufactures’ instructions.
Briefly, MOLT-4 cells were seeded in 60-mm dishes (2.5×106 cells in 5ml complete
cell culture medium per dish), and then treated with indicated concentrations of
DT2216 or ABT263 for 24h. Cells were lysed in 1× cell lysis buffer by subjecting
them to a freeze–thaw cycle in ice-ethanol bath or by 30-min incubation on ice.
After centrifugation at 5,000r.p.m. for 5min, 50µl of supernatant from each
sample was transferred into duplicate microplate wells, and 50µl of 2× substrate
solution was added to each well. 50µl of lysis bufferand50µl of 2× substrate
solution was added to background wells. After incubation for 30min at room
temperature, fluorescence was measured (excitation/emission ~ 496/520nm)
using a Biotek’s Synergy Neo2 multi-mode plate reader. The data were expressed
as percent caspase-3 activity ((Ft / F0)×100)), where Ft is the fluorescence value
of the treatment well and F0 is the average fluorescence value of control wells after
background subtraction.
BAK (BAK1) and BAX double-knockout by CRISPR–Cas9 genomic editing.
To deplete BAK and BAX, the sgRNAs targeting human BAK and BAX
were designed and cloned into lentiCRISPR v2 vector (a gift from F. Zhang;
Addgene plasmid no. 52961). Packaging 293T cells were transfected with
BAK and BAX sgRNAs or negative control (non-targeting sgRNA-sgCTRL)59
and helper vectors (pMD2.G and psPAX2; Addgene nos. 12259 and 12260)
using lipofectamine 2000 reagent (cat. no. 11668019, Life Technologies).
Medium containing lentiviral particles and 8 μg ml–1 polybrene (Sigma-
Aldrich) was used to infect H146 cells. Infected cells were selected in
medium containing 1 μg ml–1 puromycin. The target guide sequences are as
follows: BAK-sg1, forward (5′-CACCGTCATCGGGGACGACATCAAC-3′)
and reverse (5′-AAACGTTGATGTCGTCCCCGATGAC-3′); BAK-
sg2, forward (5′-CACCGCTGCAACCTAGCAGGTGAGC-3′) and
reverse (5′- AAACGCTCACCTGCTAGGTTGCAGC-3′); BAX-sg1,
forward (5′-CACCGGGATCGAGCAGGGCGAATGG-3′) and reverse
(5′-AAACCCATTCGCCCTGCTCGATCCC-3′); BAX-sg2: forward
(5′-CACCGAGCGAGTGTCTCAAGCGCAT-3′) and reverse (5′-
AAACATGCGCTTGAGACACTCGCTC-3′).
Raw data processing. Proteins were identified and SILAC ratios determined using
MaxQuant with a parent ion tolerance of 3 ppm and a fragment ion tolerance of
0.5Da. The derived peak list was searched with the built-in Andromeda search
engine against the reference human proteome downloaded from Uniprot (http://
www.uniprot.org/) on 13 March 2018. The search parameters for both algorithms
included carbamidomethylation of cysteine residues as a fixed modification and
N-terminal acetylation, oxidation at methionine, and SILAC labeling [13C615N2]Lys
and [13C615N4]Arg as variable modifications. Trypsin was specified as the protease,
and a maximum of two missed cleavages were allowed. The data were screened
against a target decoy database and the false discovery rate (FDR) was set to 1% at
the peptide level and contained at least 2 identified peptides. Protein probabilities
were assigned by the Protein Prophet algorithm61. Statistical analysis was performed
using Perseus software62. A threshold with SILAC ratio greater than 1.5-fold and
q value of 0.05 for the Benjamini–Hochberg false-discovery rate were used to
identify the proteins with significant changes. The MS data have been deposited to
the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org)
through the PRIDE partner repository63 with the dataset identifier PXD010878.
Immunoprecipitation. Human cancer cells or platelets were lysed in Pierce IP lysis
buffer (25mM Tris-HCl pH 7.4, 150mM NaCl, 1mM EDTA, 1% NP-40 and 5%
glycerol; cat. no. 87787; Thermo Fisher Scientific) supplemented with 1% (vol/vol)
protease and phosphatase inhibitor cocktail (cat. no. PPC1010; Sigma-Aldrich).
Lysates were precleared by incubating with 1µg of mouse IgG (cat. no. sc-2025;
Santa Cruz Biotechnology) and 20µl of protein A/G-PLUS agarose beads (25%
vol/vol; cat. no. sc-2003; Santa Cruz Biotechnology) for 30min at 4°C. 1,000µg of
cell lysates were incubated with 2µg of anti-BCL-XL (cat. no. sc-56021; Santa Cruz
Biotechnology) or anti-IgG for 1h followed by addition of 25µl of protein A/G-
PLUS agarose beads and incubated overnight at 4°C. Immunoprecipitates were
collected by centrifugation and washed 2–3 times with IP lysis buffer. Samples were
mixed with Laemmli’s SDS-buffer, denatured and analyzed by SDS–polyacrylamide
gel electrophoresis (SDS–PAGE) and immunoblotting. An anti-rabbit HRP-
conjugated Fc-fragment-specific secondary antibody directed against primary
antibodies was used to detect immune complexes in immunoblotting.
AlphaScreen for the determination of DT2216-BCL-XL, BCL-2 and BCL-W
binding affinity. To evaluate the binding affinities of DT2216 and ABT263
towards BCL-XL, BCL-2 and BCL-W, AlphaScreen competitive binding assays
Cellular thermal shift assay. The cellular thermal shift assay (CETSA) was
adapted from Smith et al.48. Briefly, 2.5×107 MOLT-4 or RS4 cells were treated
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
with 1µM DT2216 or DMSO for 6h, and then were collected, washed with PBS
and resuspended in PBS containing protease and phosphatase inhibitors. For
MOLT-4 cells, 10µM MG132 was added in to prevent the degradation of BCL-XL.
The resuspended cells were freeze–thawed four times with liquid nitrogen. After
each freeze–thaw cycle, lysate was vortexed briefly to ensure homogenous thawing.
The soluble fraction was separated from cell debris by centrifugation at 20,000g
for 30min at 4°C and then heated at gradient temperature from 42°C to 69.5°C
for 3min and cooled down to 25°C for another 3min. The treated samples were
centrifuged at 20,000g for 30min at 4°C to remove the denatured proteins and then
were analyzed by immunoblotting.
CGGAAATGAAGATCCAGATCTTATTAAAGCAGAAC-3′) and reverse
(5′-CACCAGCTCCCGGTTGCTCTGAGACATAGATCCCTTGTCA-
TCGTCGTCCTTGTAGTC-3′). The DNA sequence of the plasmid was
authenticated by automatic sequencing. 293T cells were cotransfected with Flag-
BCL-XL and HA-tagged ubiquitin (a gift from T. Dawson, Addgene, Plasmid no.
17608) or Flag-BCL-2 (a gift from C. Distelhorst, Addgene plasmid no. 18003)
and HA-ubiquitin for 40h, and then were treated with DMSO or 1 μM DT2216
for 4h, and 10 μM MG132 was added to prevent protein degradation. Proteins
were extracted by using immunoprecipitation lysis buffer (cat. no. 87788, Thermal
Fisher Scientific) and then subjected to immunoprecipitation using Anti-FLAG M2
Magnetic Beads (cat. no. M8823, Sigma-Aldrich) according to the manufacturer’s
protocol. Anti-FLAG M2 Magnetic Beads was washed with 1× TBS three times and
then added to protein samples, and the mixture was incubated at 4°C with rotation
overnight. The magnetic beads were collected and then washed three times with
1× TBS. Immunoprecipitated samples were eluted with 2× SDS sample buffer and
boiled 5min at 95°C.
AlphaLISA assay for ternary complex formation. We used AlphaLISA assay
to monitor ternary complex formation between target protein, PROTAC
and E3 ligase, as described previously44,58. To validate the formation of the
ternary complex of BCL-XL or BCL-2, DT2216 and the VHL E3 ligase, a fixed
concentration of His-tagged recombinant proteins (100nM BCL-XL and 10 nM
BCL-2) and recombinant active glutathione-S-transferase (GST)-tagged VHL–
elongin B–elongin C–Cul2 complex (50nM for BCL-XL, 5nM for BCL-2; cat.
no. 029641, US Biological) was incubated with varying concentrations of test
compounds in fourfold serial dilutions, to a final volume of 40µl in a 96-well PCR
plate. After 30min of incubation at room temperature, 5µl alpha glutathione-
donor beads (cat. no. 6765300, PerkinElmer) were added to each well to 20µg ml–1
final concentration and incubated for 15min. Thereafter, 5µL 6× His-acceptor
beads were added to each well at 20µg ml–1 final concentration and incubated for
an additional 45min at room temperature. Thereafter, 17µl of each sample was
transferred in adjacent wells of 384-well proxy plate, and the plate was scanned
using the Alpha program on Biotek’s Synergy Neo2 multimode plate reader. The
data were expressed as average AlphaLISA signal and plotted against different
concentrations of compounds.
Mutational analysis. Flag-BCL-XL-K87R, Flag-BCL-XL-K87H, Flag-
BCL-XL-K157R, Flag-BCL-XL-K-ko, and Flag-BCL-XL-K87-only were
constructed based on pSG5-Flag-BCL-XL (pDL2009) (mentioned above)
using QuikChange II site-directed mutagenesis kit (Agilent Technologies)
and the following primer sets: BCL-XL-K87R forward (5′-GTGATCCCCAT-
GGCAGCAGTAAGGCAAGCGCTGAGGGAGGCAGGC-3′) and reverse
(5′-GCCTGCCTCCCTCAGCGCTTGCCTTACTGCTGCCATGGGGATCAC-3′);
BCL-XL-K87H forward (5′-CCCATGGCAGCAGTACATCAAGCGCTGAGG
GAG-3′) and reverse (5′-CTCCCTCAGCGCTTGATGTACTGCTGCCATGG
G-3′); BCL-XL-K157R forward (5′-TGTGCGTGGAAAGCGTAGACAGGGA
GATGCAGG-3′) and reverse (5′-CCTGCATCTCCCTGTCTACGCTTTCCA
CGCACA-3′). BCL-XL-K-ko was constructed by mutating each of the six Lys
residues (Lys 16, Lys 20, Lys 87, Lys 157, Lys 205, and Lys 233) to Arg by the
primers mentioned above and the following primer sets: BCL-XL-K16R and
-K20R forward (5′-GTGGTTGACTTTCTCTCCTACAGGCTTTCCCAGA
GAGGATACAGCTGGAGTCAGT-3′) and reverse (5′-ACTGACTCCAG
CTGTATCCTCTCTGGGAAAGCCTGTAGGAGAGAAAGTCAACCAC-3′);
BCL-XL-K205R forward (5′-CAATGCAGCAGCCGAGAGCCGAAGGGGCC-
AGGAACGCTTCAACCGC-3′) and reverse (5′-GCGGTTGAAGCGTTCCT
GGCCCCTTCGGCTCTCGGCTGCTGCATTG-3′); BCL-XL-K233R forward
(5′-CTGGGCTCACTCTTCAGTCGGAGATGAAGATCCAGATCTTATTAA-3′)
and reverse (5′-TTAATAAGATCTGGATCTTCATcTCCGACTGAAGAGT
GAGCCCAG-3′). BCL-XL-K87-only was constructed by mutating each of the five
Lys residues (Lys 16, Lys 20, Lys 157, Lys 205 and Lys 233), excluding Lys 87, to
Arg by the primers mentioned above. DNA sequences in all these plasmids were
authenticated by automatic sequencing.
nanoBRET ternary complex formation assay. Plasmids were purified on
miniprep columns according to the manufacturer’s protocol (cat. no. 27106
Qiagen). Plasmids HaloTag-VHL (cat. no. CS1679C155) and CMV LgBit (cat.
no. CS1956B03) were purchased from Promega. HiBit-BCL-XL (pDL2288) and
HiBit-BCL-2 (pDL2306) were constructed based on pBiT3.1 N-HiBiT (cat. no.
N2361, Promega), pLX304-BCL-XL (DNASU plasmid cat. no. HsCD00437924)
and Flag-BCL-2 (a gift from C. Distelhorst, Addgene plasmid no.18003) plasmids
through Gibson assembly method (Gibson Assembly Master Mix, cat. no. E2611S,
NEB) using the primers as follows. pBiT3.1-N-BCL-XL-FP1: 5′-TGGCTCGAGC-
GGTGGGAATTCTGGTATGTCTCAGAGCAACCGGGAGCTGGTG-3′;
pBiT3.1-N-BCL-XL-RP1: 5′-TCTTCCGCTAGCTCCACCGGATCCTCCTCATT
TCCGACTGAAGAGTGAGCC-3′; pBiT3.1-N-BCL-XL-Vector-FP1: 5′-GGCTC-
ACTCTTCAGTCGGAAATGAGGAGGATCCGGTGGAGCTAGCGGAAGA-3′;
pBiT3.1-N-BCL-XL-Vector-RP1: 5′-CACCAGCTCCCGGTTGCTCTGAGACATA-
CCAGAATTCCCACCGCTCGAGCCA-3′; pBiT3.1-N-BCL-2-FP1: 5′-TGGCTC-
GAGCGGTGGGAATTCTGGTATGGCGCACGCTGGGAGAACAGGGTAC-3′;
pBiT3.1-N-BCL-2-RP1: 5′-TCTTCCGCTAGCTCCACCGGATCCTCCTCACT
TGTGGCCCAGATAGGCACCCAG-3′; pBiT3.1-N-BCL-2-Vector-FP1: 5′-CTG
GGTGCCTATCTGGGCCACAAGTGAGGAGGATCCGGTGGAGCTAGCGG
AAGA-3′; pBiT3.1-N-BCL-2-Vector-RP1: 5′-GTACCCTGTTCTCCCAGCGTG
CGCCATACCAGAATTCCCACCGCTCGAGCCA-3′. DNA sequences in these
plasmids were authenticated by automatic sequencing. 293T cells (8×105) were
transfected with Lipofectamine 2000 reagent (cat. no. 11668019, Thermo Fisher
Scientific) and 1μg HaloTag-VHL, 10ng HiBit-BCL-XL and 10ng LgBit or 1μg
HaloTag-VHL, 10ng HiBit-BCL-2 and 10ng LgBit. After 24h, 2×104 transfected
cells were seeded into white 96-well tissue culture plates in FluoroBrite DMEM
(cat. no. A18967-02, Thermo Fisher Scientific) containing 4% FBS with or without
HaloTag NanoBRET 618 Ligand (cat. no. PRN1662, Promega) and incubated
overnight at 37°C, 5% CO2. The following day, different doses of DT2216 were
added into the medium, and plates were incubated at 37°C, 5%CO2, for 6h.
After treatment, NanoBRET Nano-Glo Substrate (cat. no. N1662, Promega) was
added into the medium, and the contents were mixed by shaking the plate for 30s
before measuring donor and acceptor signals on Biotek plate reader. Dual-filtered
luminescence was collected with a 450/50nm bandpass filter (donor, NanoBiT-
BCL-XL protein or NanoBiT-BCL-2 protein) and a 610-nm longpass filter
(acceptor, HaloTag NanoBRET ligand) using an integration time of 0.5s. mBRET
ratios were calculated following the NanoBRET Nano-Glo Detection System (cat.
no. N1662, Promega).
Identifcation of ubiquitination site in BCL-XL. Enrichment of Flag-tagged
protein for mass spectrometry. 293T cells were cotransfected with Flag-BCL-XL
(pDL2009) and HA-ubiquitin vectors, or Flag-BCL-2 and HA-ubiquitin vectors.
Proteins were extracted by immunoprecipitation lysis bufer and then subjected
to immunoprecipitation using Pierce Anti-DYKDDDDK Afnity Resin (cat. no.
A36801, Termal Fisher Scientifc) according to the manufacturer’s protocol. Anti-
DYKDDDDK Afnity Resin was washed with 1× TBS three times and then added
to protein samples, and the mixture was incubated at 4°C with rotation for 4h. Te
immunoprecipitated samples were centrifuged at 1,000g for 1min at 4°C, and then
washed three times with 1× TBS. Immunoprecipitated samples were eluted with
0.1M glycine HCl (pH 2.8).
Sample preparation and LC–MS/MS. Purified proteins were reduced, alkylated
and digested using filter-aided sample preparation60. Resulting peptides were then
separated by reverse phase XSelect CSH C18 2.5-µm resin (Waters) on an in-line
150×0.075mm column using an UltiMate 3000 RSLCnano system (Thermo
Fisher Scientific). Peptides were eluted using a 90-min gradient from 97:3 to
60:40 buffer A:B ratio A (buffer A: 0.1% formic acid and 0.5% acetonitrile in
water, buffer B: 0.1% formic acid in acetonitrile). Eluted peptides were ionized by
electrospray (2.15kV), followed by MS/MS analysis using higher-energy collisional
dissociation (HCD) on an Orbitrap Fusion Lumos mass spectrometer (Thermo
Fisher Scientific) in top-speed data-dependent mode. MS data were acquired using
the FTMS analyzer in profile mode at a resolution of 240,000 over a range of 375
to 1,500m/z. Following HCD activation, MS/MS data were acquired using the ion
trap analyzer in centroid mode and the normal mass range with precursor-mass-
dependent normalized collision energy between 28 and 31.
Coimmunoprecipitation assay for the determination of BCL-XL and
BCL-2 ubiquitination. Plasmid pSG5-Flag-BCL-XL (pDL2009) was
constructed based on pSG5-Flag vector (a gift from D. Hayward at Johns
Hopkins University, which was originally generated by Stratagene) and
pLX304-BCL-XL (DNASU plasmid cat. no. HsCD00437924) through
the Gibson assembly method. The following two primers sets were used:
primer set-1, forward (5′-GACTACAAGGACGACGATGACAAGGGATCT
ATGTCTCAGAGCAACCGGGAGCTGGTG-3′) and reverse (5′-GTTCTGCTTT
AATAAGATCTGGATCTTCATTTCCGACTGAAGAGTGAGCCCAGCAGA
AC-3′); primer set-2, forward (5′-GTTCTGCTGGGCTCACTCTTCAGT-
Proteomics data analysis. Proteins were identified by a database search against the
human Uniprot database (73,911 proteins, 16 March 2019) using Mascot with a
parent ion tolerance of 3 ppm and a fragment ion tolerance of 0.5Da. Methionine
oxidation (+15.99492Da), protein N-terminal acetylation (+42.03670) and
lysine ubiquitination (+114.04293Da) were variable modifications; cysteine was
assigned a fixed carbamidomethyl modification (+57.021465Da). Percolator
was used to filter the peptide spectrum matches to a false-discovery rate of 1%.
Scaffold (Proteome Software) was used to verify MS/MS-based peptide and protein
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
identifications. Protein identifications were accepted if they could be established
with a false-discovery rate that was <1% and contained at least 2 identified
peptides. Protein probabilities were assigned by the Protein Prophet algorithm61.
The MS data have been deposited to the ProteomeXchange Consortium (http://
proteomecentral.proteomexchange.org) through the PRIDE partner repository61
with the dataset identifier PXD015454.
100mg per kg body weight). Approximately 50µl of blood was collected at diferent
time points from each mouse as described in the Fig. 4c legend via submandibular
plexus route in EDTA tubes (RAM Scientifc), and platelets were enumerated using
an automated hematology analyzer HEMAVET 950FS (Drew Scientifc).
Daily dosing with ABT263 or once a week with DT2216. Female CB-17 SCID-
beige mice were treated with ABT263 (50 mg per kg body weight per day, p.o.) or
DT2216 (15 mg per kg body weight per week, i.p.). Approximately 50µl of blood
was collected from each mouse 6h (0.25 d) after each dose of ABT263 or collected
from each mouse after 0.25 d, 1.25 d, 2.25 d, 3.25d, 4.25 d, 5.25 d, 6.25 d, 7.25
d, 8.25 d, 9.25 d, 10.25 d, 11.25 d, 12.25 d, 17.25 d and 24.25 d of once per week
dosing with DT2216. Platelets were enumerated using HEMAVET 950FS.
Mice. Female CB17/Icr-Prkdcscid/IcrIcoCrl (CB-17 SCID), CB17.Cg-PrkdcscidLystbg-J
Crl (CB-17 SCID-beige) or NOD.CB17-Prkdcscid/NCrCrl (NOD-SCID) mice aged
5–6 weeks were purchased from Charles River Laboratories, and were housed
in the Assessment and Accreditation of Laboratory Animal Care (AAALAC)-
accredited animal facilities at University of Arkansas for Medical Sciences
(UAMS) or University of Florida (UF) under pathogen-free conditions. Female
NOD-SCID IL2Rgnull (NSG) mice aged 6–7 weeks were purchased from Jackson
Laboratory (Bar Harbor, Main, USA) and were housed in the AAALAC-accredited
animal facilities at MD Anderson Cancer Center (MDACC) under pathogen-free
conditions. Mice received food and water ad libitum and were allowed to acclimate
for 1–2 weeks before being used for experiments at the age of 6–8 weeks. All
animal work was approved and done in accordance with the approvals from the
Institutional Animal Care and Use Committees of UAMS, UF and MDACC with
the exception of the pharmacokinetic (PK) studies that were done by BioDuro,
a global contract research organization, through a contract. All animal studies
complied with the ethical regulations and humane endpoint criteria according to
the NIH Guidelines for the Care and Use of Laboratory Animals.
/
Complete blood-cell counts. Approximately 50µl of blood was collected from
each mouse in EDTA tubes via submandibular plexus or retro-orbital bleeding.
The mice were anaesthetized with isoflurane inhalation for retro-orbital bleeding.
The blood was immediately used for complete blood counts (CBCs) using an
automated hematology analyzer HEMAVET 950FS. The data were expressed as
number of different blood cells or platelets per µl of blood.
MOLT-4 T-ALL xenograft mouse model. To test the effect of DT2216 on tumor
growth in MOLT-4 T-ALL xenografts, MOLT-4 T-ALL cells were collected
and suspended in regular RPMI medium and mixed with Matrigel (1:1)
(cat. no. 356231, Corning). The cells (5×106 cells) suspended in 100µL of RPMI
medium–matrigel mixture were subcutaneously (s.c.) implanted in the right
flank of CB-17 SCID mice. Tumor growth was monitored daily, and tumors
were measured twice per week using Vernier caliper or digital calipers. Tumor
volume was determined using the formula ((L×W2)×0.5), where L is length in
mm and W is the width in mm. The treatment started once the average tumor
volume reached 150–200mm3. The animals were randomly assigned into separate
groups (n=6–8 animals per group) such that each group had nearly equal starting
average tumor volume. Mice were weighed twice a week, and the treatments were
given according to average mouse weight within each group before initiation of
treatment. DT2216 and ABT263 for i.p. administration were formulated in 50%
phosal 50 PG, 45% miglyol 810N and 5% polysorbate 80. DT2216 and ABT263
were administered via i.p. injection at 15 mg per kg body weight per week in 100µl
vehicle (Extended Data Fig. 8). ABT263 for oral administration was formulated in
10% ethanol, 30% polyethylene glycol (PEG) 400 and 60% phosal 50 PG (Fig. 4).
Control mice received 100µl vehicle via i.p. injection. A mouse was euthanized
when its maximum tumor size reached the humane endpoint, defined according
to institutional policy concerning tumor endpoints in rodents. In addition, to
prevent excessive pain or distress, the mice were euthanized if the tumors became
ulcerated or if the mice showed any signs of ill health. Mice were euthanized by
CO2 suffocation followed by cervical dislocation, and then various tissues including
tumors were collected for further analyses.
Analysis of DT2216 in tumor samples. A bioanalytical method was developed
for the quantification of DT2216 in tumor homogenates using a Waters Xevo
TQ-S Micro triple quadrupole mass spectrometer and Acquity Class I UPLC.
Chromatographic separation was achieved using a mobile phase consisting of
aqueous ammonium acetate buffer (10mM, pH 3.5) (pump A) and acetonitrile
(pump B) on a Waters Acquity BEH C18 column (1.7μm, 2.1×50mm). A gradient
method started with pump B supplying 15% of acetonitrile up to 0.5min. The
composition of acetonitrile was linearly increased to 80% until 1.5min and kept
constant until 3.0min. The composition of acetonitrile was immediately decreased
to 15% at 3.0min and kept constant until 3.5min. The flow rate of the mobile
phase was 0.35ml min–1 and the injection volume was 2µl. Analysis of DT2216 was
performed using ESI in the positive mode. Capillary voltage, cone gas, desolvation
gas, desolvation temperature and source temperature were optimized for maximum
analyte response and held at 3.0kV, 50L per h (flow rate), 900L per h (flow rate),
450°C and 150°C, respectively. The mass spectral analysis of DT2216 and the
internal standard (amiodarone) was achieved by multiple-reaction monitoring
(MRM). Compound parameters are detailed in Supplementary Table 8. Data
acquisition and quantitation were performed using MassLynx and TargetLynx
XS software version 4.2. For the bioanalysis of DT2216, pieces of tumor samples
were weighted and homogenized in water (1:4, %wt/vol) using a tissue homogenizer
(Dremel, BioSpec Products). Sample cleanup (20µl) was performed by a protein-
precipitation method using acetonitrile containing the internal standard (40ng ml–1)
as a quenching solution (80µl). Calibration (5, 10, 50, 100, 200, 400 and 500ng ml–1)
and quality-control standards (5, 7.5, 250 and 450ng ml–1) were prepared in
drug-free tumor homogenate and analyzed using the ultra-performance LC–MS/
MS method. The method was linear for a calibration range of 5–500ng ml–1. The
accuracy and precision calculated for the method were within the specified limits64
Test samples were analyzed along with freshly prepared calibration and quality-
control standards (Extended Data Fig. 5d). Tumor concentration-time data were
subjected to non-compartmental analysis, and pharmacokinetic parameters were
calculated using Phoenix WinNonlin, version 6.1.
H146 small-cell lung cancer xenograft model. To test the effect of DT2216 alone
or in combination with ABT199 on tumor growth in H146 SCLC xenografts,
5×106 H146 SCLC cells were suspended in regular RPMI medium, mixed with
Matrigel, and s.c. implanted in the right flank of female CB-17 SCID mice as
described in the previous method. Tumor growth was monitored, tumors were
measured and the tumor volume was determined as mentioned in the previous
method. The treatment started after 4 weeks of cell implantation as shown in Fig. 5c.
The animals were randomly assigned into groups treated with vehicle, DT2216
15 mg per kg body weight per week (i.p. injection), ABT263 15 mg per kg body
weight per week (i.p. injection), ABT199 50 mg per kg body weight per day (p.o.),
and DT2216 + ABT199. DT2216 and ABT263 were formulated as described above.
ABT199 was formulated for oral dosing in 60% phosal 50 PG, 30% PEG 400 and
10% ethanol. All the animals were euthanized in accordance with institutional
policy, and various tissues were collected. Tumors were weighed and used for BCL-
Pharmacodynamics of BCL-XL degradation by DT2216. MOLT-4 xenografts
were established in female CB-17 SCID-beige mice as described in following
methods. Tumor-bearing mice were treated with single injection of vehicle or
DT2216 (15mg per kg body weight, i.p.) when the tumors were ~600mm3 in size.
Two mice each from vehicle- and DT2216-treated groups were euthanized at each
time point as indicated in the figure legend of Fig. 4b, and tumors were collected.
The proteins were extracted from tumors and used for BCL-XL degradation by
immunoblot analysis. Some portion of these tumors was used for DT2216 analysis
as described in the previous method.
X
L, BCL-2 and MCL-1 expression by immunoblotting.
MDA-MB-231 BC xenograft model. To test the effect of DT2216 in combination
with docetaxel on tumor growth in MDA-MB-231 breast cancer xenografts,
5×106 MDA-MB-231 cells were suspended in regular RPMI medium, mixed with
matrigel, and s.c. implanted in the right flank of female NOD-SCID mice. The
animals were randomly assigned to vehicle, DT2216 (15 mg per kg body weight,
q4d, i.p.), docetaxel (7.5 mg per kg body weight, q14d, i.v.) and DT2216+docetaxel
groups when the average tumor volume reached ~130mm3. Docetaxel was
dissolved in 5% DMSO+30% PEG 300+5% Tween 80+60% distilled H2O. The
solution was filter sterilized to obtain clear solution for intravenous administration.
DT2216 was administered 2 d before starting dosing with docetaxel, as also
mentioned in Fig. 6c.
Protein extraction from tumors for immunoblotting. Following resection, the
tumors were snap frozen and stored at −80°C until later use. Tumor tissue (~50mg
wet tissue) was lysed in 1ml of RIPA lysis buffer supplemented with 1% protease
inhibitor cocktail and 1% phosphatase inhibitor cocktail. The tissues were kept on
ice and were homogenized with the help of a hand homogenizer. The samples were
incubated on ice for 2–3h and then kept at −80°C overnight to allow for complete
lysis. The subsequent sample preparation and immunoblotting were performed as
described in the immunoblotting method.
T-ALL patient-derived xenograft models. CUL76 and 332×-Luci T-ALL PDX
models were kind gifts from A. Ferrando (Columbia University, New York, NY, USA)
and K. Yong-Mi (Children’s Hospital Los Angeles, Los Angeles, CA), respectively.
The D115 T-ALL PDX model was from M. Konopleva’s Lab (UTMDACC, Houston,
TX). The use of these models was in accord with the Declaration of Helsinki using
In vivo platelet toxicity assays. Single dosing with DT2216 or ABT263. Female 5–6
weeks old CB-17 SCID-beige mice were treated with single i.p. doses of DT2216
(7.5, 15 and 25mg per kg body weight) or single p.o. doses of ABT263 (25, 50 and
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
an Institutional Review Board (IRB)-approved protocol from MDACC. To establish
T-ALL PDX mouse models, 8-week-old NSG mice were sublethally irradiated
(0.25Gy) 24h prior to cell inoculation. PDX cells (1×106) were suspended in PBS
and injected into mice through the tail vein. Tumor engraftment was determined
by costaining with human and murine anti-CD45 (BioLegend, San Diego, CA) in
bone-marrow aspiration samples (CUL76 and D115 T-ALL) or in peripheral blood
(332X-Luci) at 10 d post injection. Mice were randomized into different groups.
CUL76 PDX were treated with DT2216 (15mg per kg body weight, q4d, i.p.),
ABT199 (100mg per kg body weight, qd, p.o.), chemotherapy (vincristine 0.15 mg
per kg body weight+dexamethasone 5 mg per kg body weight+l-asparaginase
1,000 Units per kg body weight, i.p., weekly), a combination of DT2216 and ABT199
or chemotherapy or ABT199 plus chemotherapy as shown in Fig. 6f. D115 and
332X-Luci PDXs were treated with DT2216, chemotherapy or a combination
of DT2216 and chemotherapy as shown in Extended Data Fig. 10. Mice were
monitored for disease progression by weekly assessment of leukemic (hCD45+) cells
in the peripheral-blood or bone-marrow aspirates and followed for survival.
62. Tyanova, S. et al. Te Perseus computational platform for comprehensive
analysis of (prote) omics data. Nat. Methods 13, 731–740 (2016).
63. Vizcaíno, J. A. et al. Te PRoteomics IDEntifcations (PRIDE) database and
associated tools: status in 2013. Nucleic Acids Res. 41, D1063–D1069 (2013).
64. US Food and Drug Administration. Bioanalytical Method Validation,
Guidance for Industry (US Department of Health and Human Services Food
and Drug Administration, 2018).
Acknowledgements
This study was supported by US National Institutes of Health (NIH) grants R01
CA211963 (D.Z.), R01 CA219836 (D.Z.), R01 GM109645 (R.H.), R01 CA205224 (R.H.),
R01 CA200673 (W.Z.), R01 CA203834 (W.Z.), R21 CA223371 (G.Z.), R35 CA210065
(A.F.), R01 CA172809 (Y.-M.K.), Department of Defense grant BC180227 (W.Z.), the
Schwab Charitable fund (M.K.) and CPRIT grant RP160716 (P.J.H.). Mass spectrometric
support was provided by Alan J. Tackett and Samuel G. Mackintosh in the UAMS
Proteomics Core Facility.
Statistical analyses. All the graphs presented in this manuscript were made and
statistical analyses were performed using GraphPad Prism 7 software. For analysis
of means of three or more groups, analysis of variance (ANOVA) tests were
performed. In the event that ANOVA justified post hoc comparisons between
group means, the comparisons were conducted using Tukey’s multiple-comparisons
test. Two-sided unpaired Student’s t-test was used for comparisons between the
means of two groups. A Kaplan–Meier test was used for the survival rate in Fig. 6h
and Extended Data Fig. 10b,d, and the data were statistically analyzed using log-
rank (Mantel–Cox) test. P<0.05 was considered to be statistically significant. The
precise P values have been provided wherever possible and appropriate.
Author contributions
S.K. designed, performed and analyzed most of the experiments and wrote the
manuscript; X.Z. designed, synthesized and analyzed BCL-XL PROTACs and wrote the
manuscript; D.L. designed, performed and analyzed the mass-spectrometry experiments,
nanoBRET assay and wrote the manuscript; Y.H., D.T., J.S.W., J.P., W.Z., A.S., C.R.M,
V.M.K., N.B., A. R. and G.H. performed and analyzed some experiments; P.Z and X.L.
designed, synthesized and analyzed BCL-XL PROTACs; A.A.F. and Y.-M.K. provided the
PDX T-ALL model and revised the manuscript; Q.Z. and M.K. designed and performed
the PDX T-ALL study and analyzed and interpreted data and revised the manuscript;
Y.Y., P.J.H, W.Z. and R.H. interpreted data and revised the manuscript; G.Z. designed
and supervised the study, analyzed and interpreted data and revised the manuscript. D.Z.
conceived, designed and supervised the study, analyzed and interpreted data, and wrote
the manuscript. All authors discussed the results and commented on the manuscript.
Reporting Summary. Further information on research design is available in the
Nature Research Reporting Summary linked to this article.
Data availability
The MS data have been deposited to the ProteomeXchange Consortium (http://
proteomecentral.proteomexchange.org) through the PRIDE partner repository
with the dataset identifiers PXD010878 and PXD015454. The raw immunoblot
images and files from statistical analyses are supplied as source data. Other raw and
analyzed data files are available from the corresponding author upon reasonable
request. A data availability statement is also included in the reporting summary
linked to this article.
Competing interests
S.K., X.Z, Y.H., P.Z., G.Z. and D.Z. are inventors of two pending patent applications
for use of BCL-XL PROTACs as senolytic and antitumor agents. R.H., G.Z. and D.Z.
are co-founders of and have equity in Dialectic Therapeutics, which develops BCL-XL
PROTACs to treat cancer.
Additional information
Extended data is available for this paper at https://doi.org/10.1038/s41591-019-0668-z.
References
59. Lv, D.-W., Zhang, K. & Li, R. Interferon regulatory factor 8 regulates aspase-1
expression to facilitate Epstein–Barr virus reactivation in response to B cell
receptor stimulation and chemical induction. PLoS Pathog. 14, e1006868 (2018).
60. Wiśniewski, J. R. et al. Universal sample preparation method for proteome
analysis. Nat. Methods 6, 359–362 (2009).
Supplementary information is available for this paper at https://doi.org/10.1038/
s41591-019-0668-z.
Correspondence and requests for materials should be addressed to G.Z. or D.Z.
Peer review information Javier Carmona was the primary editor on this article and
managed its editorial process and peer review in collaboration with the rest of the
editorial team
61. Nesvizhskii, A. I., Keller, A., Kolker, E. & Aebersold, R. A statistical model for
identifying proteins by tandem mass spectrometry. Anal. Chem. 75,
4646–4658 (2003).
Reprints and permissions information is available at www.nature.com/reprints.
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
Extended Data Fig. 1 | expression of anti-apoptotic BCL-2 members and VHL in cancer, and the synthetic schemes of DT2216 and DT2216NC. a, A
representative of three immunoblot analyses of VHL in three different human tumor cell lines and platelets from three different individuals (indicated
as units 1–3). b, Immunoblot analyses of the basal protein levels of BCL-XL, BCL-2, MCL-1 and VHL in different solid tumor cells. Data are representative
of two independent experiments. c, BCL2L1 and VHL mRNA expression (log2 transformed) and mutational status were analyzed using TCGA PanCancer
Atlas studies via cBioPortal. d, Synthetic schemes of DT2216, DT2216NC and VHL-L. Reagents and conditions: (i) (1) N-methylmorpholine, isobutyl
chloroformate, THF, –25 °C then −15 °C; (2) NaBH4, H2O, –15 °C; (ii) Bu3P, diphenyl disulfide, toluene, 80 °C; (iii) DIBAL-H, toluene, –78 °C (iv) compound
5, NaBH(OAc)3, TEA, DCM; (v) TFA, DCM; (vi) compound 8, TEA, acetonitrile, reflux; (vii) compound 10, EDCI, DMAP, DCM; (viii) Zn, HOAc, THF; (ix)
HATU, TEA, DCM; (x) (1) LiOH monohydrate, methanol, H2O; (2) compound 12, HATU, TEA, DCM; (xi) Ac2O, TEA, DCM.
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
Extended Data Fig. 2 | The BCL-XL degradation by DT2216 is rapid and long lasting. a, Immunoblot analysis of BCL-XL expression in MOLT-4 cells after
they were treated with DT2216 for various durations as indicated. A representative immunoblot is presented on the top panel. Densitometric analysis of
BCL-XL expression is presented on the bottom panel as mean of two independent experiments. Each symbol represents data (% of 0 h) from an individual
experiment. b, Analysis of BCL-XL expression by immunoblot in MOLT-4 cells treated with DT2216 for 16 h followed by drug withdrawal and then cultured
without DT2216 for 0 to 48 h as indicated. A representative immunoblot is on the top panel. Densitometric analysis of BCL-XL expression is presented
on the bottom panel as the mean of two independent experiments. Each symbol represents data (% of Veh) from an individual experiment. c, Caspase-3
activity in MOLT-4 cells was measured 24 h after they were treated with 1 µM of DT2216 or ABT263. Data are presented as the mean (n = 2 technical
replicates) of a representative experiment. Each symbol represents data (% of Veh) from an individual replicate. Similar results were obtained in an
additional independent experiment. d, Representative immunoblot to confirm CRISPR–Cas9-mediated double knockout of BAX and BAK in H146 cells. The
experiment was repeated independently one more time with similar results.
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
Extended Data Fig. 3 | ABT263 or VHL-L blocks the formation of ternary complex of BCL-XL, DT2216 and VHL. a, Recombinant His-tagged BCL-XL protein
(100 nM) was incubated with the VHL-Elongin B/C complex (50 nM) with 0.039 µM of DT2216. The ternary complex formation was abrogated in the
presence of ABT263 (1 µM) or VHL-L (10 µM). Data are expressed as mean (n = 2 technical replicates). Each symbol represents data from an individual
replicate. b, The negative-control of DT2216 (DT2216NC) cannot form a ternary complex with BCL-XL and the VHL complex. Recombinant His-tagged
BCL-XL protein (100 nM) was incubated with the VHL-Elongin B/C complex (50 nM) with increasing concentrations of DT2216 or DT2216NC. Data are
expressed as mean (n = 2 technical replicates).
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
Extended Data Fig. 4 | DT2216 binds to both BCL-XL and BCL-2 but cannot degrade BCL-2. a, The binding affinities of DT2216 and ABT263 towards
BCL-XL, BCL-2 and BCL-W were measured by AlphaScreen and are represented in terms of inhibition constant (Ki). The data are average of two
independent experiments each performed in duplicates. b,c, Immunoblot analyses of BCL-2 and BCL-W are shown after the RS4 cells were treated with
indicated concentrations of DT2216 for 16 h (top), or with 1 µM of DT2216 for indicated durations (bottom). d, Cell viability of BCL-2-dependent RS4 cells
after treatment with increasing concentrations of DT2216, ABT199 or ABT263 for 72 h. Data are presented as mean s.d. from six replicate cell cultures
in one experiment. Similar results were obtained in two additional independent experiments for DT2216. The EC50 of DT2216 is the average of three
independent experiments. e, Cell viability of MCL-1-dependent H929 cells after treatment with increasing concentrations of DT2216 or S63845 for 72 h.
Data are presented as mean s.d. from six replicate cell cultures in one experiment. f, Schematic representation of the proteomic assay shown in Fig. 3c.
g, Representative immunoblot analyses are shown after the cells were treated with indicated concentrations of PZ-15227 for 16 h (top) or with 0.1 µM of
PZ-15227 for indicated time points (bottom). Similar results were obtained in one more independent experiment. h, CETSA assay for BCL-XL and BCL-2.
MOLT-4 (for BCL-XL) or RS4 (for BCL-2) cells were treated with DMSO, or 1 µM of DT2216 for 6 h. Raw band intensities are mean of two independent
experiments.
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
Extended Data Fig. 5 | Drug metabolism and pharmacokinetic/pharmacodynamic profile of DT2216. a, DT2216 stability in mouse blood and plasma.
b, DT2216 liver microsomal stability. c, Plasma concentrations of DT2216 after a single administration of 2 mg per kg body weight (i.v. injection),
20 mg per kg body weight (i.p. injection) or 20 mg per kg body weight (p.o. administration) are presented as mean s.d. (n = 3 mice per group).
These studies were done by BioDuro, a global contract research organization, through a contract. d, MRM chromatograms of (A) DT2216 in drug-free
brain homogenate, (B) internal standard in spiked drug-free tumor homogenate (40 ng ml–1), (C) DT2216 in spiked tumor homogenate (5 ng ml–1; LLOQ),
(D) DT2216 in tumor sample taken from vehicle dosed mouse at 24 h and (E) DT2216 in tumor sample taken at 24 h after 15 mg per kg body weight,
i.p. administration. e, Representative immunoblot analysis of BCL-XL in tumors at different durations after DT2216 (DT, 15 mg per kg body weight, i.p.)
administration (n = 2 mice in vehicle and DT2216 groups at each time points). Similar results were obtained in two more immunoblot experiments.
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
Extended Data Fig. 6 | DT2216 induces BCL-XL degradation and apoptosis in MOLT-4 T-ALL xenografts and suppresses their growth in a dose-
dependent manner. a,b, T-ALL xenograft tumors were collected 2 d after female CB-17 SCID mice received 1 i.p. injection of DT2216 at 7.5 mg per kg body
weight and 15 mg per kg body weight. A single immunoblot analysis of BCL-XL, BCL-2, MCL-1, cleaved and full length caspase-3 and PARP1 in the tumors
is shown. c, Changes in tumor volume over time after the start of treatment with vehicle (Veh), or DT2216 (7.5 or 15 mg per kg body weight, q7d, i.p.).
All the data presented are mean s.e.m. (n = 8 mice in vehicle, 8 mice in 7.5 mg DT2216 per kg body weight and 7 mice in 15 mg DT2216 per kg body
weight). The data from the 15 mg DT2216 per kg body weight group are also presented in Extended Data Fig. 8f, in which the tumor size in these mice were
continuously monitored till day 55 post-treatment.
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
Extended Data Fig. 7 | images of MOLT-4-tumor-bearing mice and excised tumors. The images shown (quantification data are shown in Fig. 4f,g) were
captured at the end of the experiment, when the mice were treated with vehicle, DT2216 (15 mg per kg body weight, q7d, i.p.) or ABT263 (50 mg per kg
body weight, qd, p.o.). The tumor-bearing mice (shown in the top) and harvested tumors (shown in the bottom) are not placed in an identical order.
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
Extended Data Fig. 8 | DT2216 induces regression of MOLT-4 xenografts without causing thrombocytopenia. a, Illustration of the experimental design
of the MOLT-4 T-ALL xenograft mouse model. b, Blots that are representative of two independent immunoblot analyses of BCL-XL, cleaved and full-length
caspase-3 and PARP1 in MOLT-4 T-ALL collected 2 d after tumor-bearing mice were treated with a single injection of vehicle or DT2216 (15 mg per kg
body weight, i.p.). c, Blood platelets (PLT), white blood cells (WBC), lymphocytes (LYM) and granulocytes (GRA) were numerated 1 d after first treatment
with vehicle, DT2216 (15 mg per kg body weight, i.p.) or ABT263 (15 mg per kg body weight, i.p.) as shown in a. a and b represent statistical significance
versus Veh and DT2216, respectively, determined by one-way ANOVA and Tukey’s multiple-comparison test. d, Body-weight changes in mice after the
start of treatment with vehicle, DT2216 or ABT263 as shown in a. e, Numeration of PLT 1 d after the sixth dose of DT2216 (15 mg per kg body weight, q7d,
i.p. or 15 mg per kg body weight, q4d, i.p.). Data are presented as mean s.e.m. (n = 2 mice in vehicle, 7 mice in DT2216 q7d and 6 mice in DT2216 q4d).
f, Changes in tumor volume over time after the start of treatment with vehicle, DT2216 or ABT263 as shown in a. Data presented in c, d and f are mean
s.e.m. (n = 6 mice for vehicle group and 7 mice each for DT2216 and ABT263 groups). Each symbol in c and e represents data from an individual animal.
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
Extended Data Fig. 9 | images of excised H146 SCLC tumors. The tumor images shown for the quantification data presented in Fig. 5g. Mice were treated
as shown in Fig. 5c.
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Nature MediciNe
Articles
Extended Data Fig. 10 | Anti-leukemic activity of DT2216 alone and in combination with chemotherapy in T-ALL PDX models. a, Eight-week-old
female NSG mice (n = 20 mice) were injected with 1 × 106 D115 cells 24 h after 0.25-Gy irradiation. Upon engraftment, mice were randomized to receive
vehicle, DT2216 (15 mg per kg body weight, i.p., q4d for 3 weeks), chemotherapy (Chemo (vincristine 0.15 mg per kg body weight + dexamethasone 5
mg per kg body weight + l-asparaginase 1,000 units per kilogram body weight, i.p., q7d for 3 weeks)), or the combination of DT2216 with chemotherapy.
Disease burden was followed by engraftment in bone marrow by checking hCD45% in bone marrow aspiration samples through flow cytometry. Data are
presented as mean s.e.m. (n = 5 mice in each group). Each symbol represents data from an individual animal, and the middle horizontal line represents
the mean. b, Mice survival was followed and statistical significance was determined by log-rank test (n = 5 mice in each group). c, Eight-week-old female
NSG mice (n = 20 mice) were injected with 1 × 106 332X-luci cells 24 h after 0.25-Gy irradiation. Upon engraftment, mice were randomized to receive
vehicle, DT2216, chemotherapy or the combination of DT2216 with chemotherapy as mentioned in a. Disease burden was followed by checking hCD45%
in peripheral blood samples (retro orbital) through flow cytometry. Data are presented as mean s.e.m. (n = 5 mice in each group). Each symbol
represents data from an individual animal, and the middle horizontal line represents the mean. d, Mice survival was followed, and statistical significance
was determined by log-rank test (n = 5 mice in each group).
NATuRe MeDiCiNe | www.nature.com/naturemedicine

Corresponding author(s):
Last updated by author(s):
Daohong Zhou
10-25-2019
Reporting Summary
Nature Research wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency
in reporting. For further information on Nature Research policies, see Authors & Referees and the Editorial Policy Checklist.
Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.
n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement
A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly
The statistical test(s) used AND whether they are one- or two-sided
Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested
A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons
A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient)
AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals)
For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted
Give P values as exact values whenever suitable.
For Bayesian analysis, information on the choice of priors and Markov chain Monte Carlo settings
For hierarchical and complex designs, identification of the appropriate level for tests and full reporting of outcomes
Estimates of effect sizes (e.g. Cohen's d, Pearson's r), indicating how they were calculated
Our web collection on statistics for biologists contains articles on many of the points above.
Software and code
Policy information about availability of computer code
Image Lab Touch version 2.2.0.08 Software (Bio-rad, Hercules, CA, USA) was used for scanning some of the immunoblots on the
ChemiDoc MP imaging system; Gen5 version 3.04 software (BioTek, Winooski, VT, USA) was used for absorbance, fluorescence and
luminescence measurements on Synergy Neo2 multi-mode plate reader; FlowJo V10 software was used for the acquisition of flow
cytometry data; MassLynx and TargetLynx XS version 4.2 software (Waters, Milford, MA, USA) was used for data acquisition and
quantification to determine DT2216 concentration in tumor samples.
Data collection
ImageJ version 1.52a software (NIH) was used for the quantification of immunoblots; GraphPad Prism version 7 (GraphPad Software, La
Jolla, CA, USA) was used for the preparation of all the graphs, determination of half maximal effective concentration (EC50) values,
Inhibition constant (Ki), and the statistical analyses; Compusyn version 1.0 software (http://www.combosyn.com) was used for the
determination of combination index (CI); MaxQuant version 1.6.2.10 software (https://maxquant.org/maxquant/) was used for the
analysis of mass spectrometry data; FlowJo V10 software was used to analyze the flow cytometry data; Perseus version 1.6.8.0 software
(https://maxquant.org/perseus/) was used for the statistical analysis of proteomics data; Scaffold version 4.10.0 (Proteome Software,
Portland, OR, USA) was used to verify MS/MS based peptide and protein identifications; Phoenix WinNonlin version 6.1 (Certara,
Princeton, NJ, USA) was used for the calculation of pharmacokinetic parameters presented in Table-4.
Data analysis
For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors/reviewers.
We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Research guidelines for submitting code & software for further information.
1

Data
Policy information about availability of data
All manuscripts must include a data availability statement. This statement should provide the following information, where applicable:
- Accession codes, unique identifiers, or web links for publicly available datasets
- A list of figures that have associated raw data
- A description of any restrictions on data availability
The Mass Spectrometry (MS) data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) through the PRIDE
partner repository with the dataset identifiers PXD010878 and PXD015454. These data have been released for public access. The raw immunoblot images and
statistics data are supplied as source data linked with this article. Raw densitometric analyses files and plate-reader data files are available from the corresponding
author upon reasonable request.
Field-specific reporting
Please select the one below that is the best fit for your research. If you are not sure, read the appropriate sections before making your selection.
Life sciences
Behavioural & social sciences
Ecological, evolutionary & environmental sciences
For a reference copy of the document with all sections, see nature.com/documents/nr-reporting-summary-flat.pdf
Life sciences study design
All studies must disclose on these points even when the disclosure is negative.
In our own experience 5-10 animals are sufficient to observe significant differences among different groups. For in vivo platelet toxicity
studies, we used three animals in each group because the effect of treatments on platelet counts was consistent among different animals in a
group. For PDX studies, the number of animals were based on sample availability and experimental feasibility. In other experiments, sample
sizes were determined on the basis of previous studies that had used similar sample sizes to successfully determine a difference between
groups. No statistical methods were used to predetermine sample sizes.
Sample size
In general, no data were excluded from the analyses unless there were obvious technical reasons. For example, in two samples the blood got
clotted, hence could not be used for CBCs. One animal was excluded because of unsuccessful tumor engraftment.
Data exclusions
Replication
Adequate measures were taken to verify the reproducibility of findings. An effect was confirmed across multiple cell lines and tumor samples
with independent experiments and using several biological and technical replicates. The compounds were tested at multiple concentrations in
vitro and the in vivo results were confirmed in different model systems such as cell line- and patient-derived xenografts. Some results were
confirmed by different investigators and in different laboratory conditions. All attempts at replication were successful.
Animals were randomized in different groups on the basis of tumor volumes in a way that each group had nearly equal average tumor volume
at the start of treatment.
Randomization
Blinding
Investigators were not blinded during randomizing the mice in different groups. Blinding was not possible because the animals were
segregated into separate labeled cages with nearly equal average tumor volumes in each treatment group. When randomizing the animals in
different treatment groups, the animals with lowest tumor sizes were first assigned to each group and so on to ensure that each group had
animals with similar tumor sizes.
Reporting for specific materials, systems and methods
We require information from authors about some types of materials, experimental systems and methods used in many studies. Here, indicate whether each material,
system or method listed is relevant to your study. If you are not sure if a list item applies to your research, read the appropriate section before selecting a response.
Materials & experimental systems
n/a Involved in the study
Antibodies
Methods
n/a Involved in the study
ChIP-seq
Eukaryotic cell lines
Flow cytometry
Palaeontology
MRI-based neuroimaging
Animals and other organisms
Human research participants
Clinical data
2

Antibodies
Antibodies used
For immunoblotting: the rabbit polyclonal VHL (Cat. No. 68547S, dilution 1:1000), BCL-XL (Cat. No. 2762S, dilution 1:2000),
caspase-3 (Cat. No. 9662S, dilution 1:1000), cleaved Caspase-3 (clone Asp175, Cat. No. 9661S, dilution 1:1000), cleaved PARP1
(clone Asp214, Cat. No. 9541S, dilution 1:1000), Flag tag (Cat. No. 2044S, dilution 1:1000), BAX (Cat. No. 2772S, dilution 1:1000),
and β-tubulin (Cat. No. 2146S, dilution 1:5000); the rabbit monoclonal MCL-1 (clone D35A5, Cat. No. 5453S, dilution 1:1000),
BCL-2 (clone 50E3, Cat. No. 2870S, dilution 1:500), BCL-W (clone 31H4, Cat. No. 2724S, dilution 1:500), PARP1 (clone 46D11, Cat.
No. 9532S, dilution 1:1000), BAK (clone D4E4, Cat. No. 12105S, dilution 1:1000), BIM (clone C34C5, Cat. No. 2933S, dilution
1:1000), PUMA (clone D30C10, Cat. No. 12450S, dilution 1:1000), HA tag (clone C29F4, Cat. No. 14031S, dilution 1:1000), and β-
actin (clone 13E5, Cat. No. 4970L, dilution 1:5000) were purchased from Cell Signaling Technology. The mouse monoclonal β-
actin (clone C4, Cat. No. 8691001, dilution 1:10000) was purchased from MP Biomedicals. The goat anti-rabbit (Cat. No. 7074S,
dilution 1:3000) and horse anti-mouse (Cat. No. 7076S, dilution 1:3000) HRP-conjugated secondary antibodies were purchased
from Cell Signaling Technology. The goat anti-rabbit IgG, Fc fragment specific (Cat. No. 111-035-046, dilution 1:10000) was
purchased from Jackson ImmunoResearch.
For immunoprecipitation: the mouse monoclonal BCL-XL (clone 7B2.5, Cat. No. sc-56021, dilution 1:50) and normal mouse IgG
(Cat. No. sc-2025, dilution 1:100) were purchased from Santa Cruz Biotechnology.
For flow cytometry: the rat IgG2b, κ, APC mCD45-APC (clone 30-F11, Cat. No. 103111, dilution 1:250) and Mouse IgG1, κ, FITC
hCD45-FITC (clone H130, Cat. No. 304038, dilution 1:250) were purchased from Biolegend.
All the used antibodies are commercially available. The antibodies used in a specific species or application have been validated
by manufacturers to be used in that species/application and this information is provided in their website and/or antibody
datasheets.
Validation
Eukaryotic cell lines
Policy information about cell lines
All the cell lines were purchased from American Type Culture Collection (ATCC), except EJM and NCI-H929 cells which were
kind gift from Dr. Erming Tian at the Winthrop P. Rockefeller Cancer Institute at University of Arkansas for Medical Sciences.
MOLT-4 (ATCC, Cat. No. CRL-1582); RS4;11 (ATCC, Cat. No. CRL-1873); NCI-H146 (ATCC, Cat. No. HTB-173); MDA-MB-231
(ATCC, Cat. No. HTB-26); PC-3 (ATCC, Cat. No. CRL-1435); HepG2 (ATCC, Cat. No. HB-8065); SW620 (ATCC, Cat. No. CCL-227);
786-O (ATCC, Cat. No. CRL-1932); WI-38 (ATCC, Cat. No. CCL-75); HEK 293T (ATCC, Cat. No. ACS-4500).
Cell line source(s)
The cell lines have been validated by the suppliers. The morphology and growth of the used cell lines were verified with the
supplier's data sheets, and their pharmacological responses were matched with the available literature.
Authentication
Cell lines were recently purchased from ATCC and were not further tested for mycoplasma contamination in our laboratory.
Mycoplasma contamination
Commonly misidentified lines
(See ICLAC register)
NCI-H929 was used to generate some supporting data (Fig 3a, b and supplemetary Fig. 4e). This cell line is known to solely
depend on MCL-1 for their survival and has also been previously used for this purpose (Kotschy, A. et al. Nature. 2016, 538,
477–482). The data pertaining to this cell line does not have any significant impact on the outcomes of our study. None of the
other cell lines used are listed in ICLAC database.
Animals and other organisms
Policy information about studies involving animals; ARRIVE guidelines recommended for reporting animal research
Female CB17/Icr-Prkdcscid/IcrIcoCrl (CB-17 SCID), CB17.Cg-PrkdcscidLystbg-J/Crl (CB-17 SCID-beige), or NOD.CB17-Prkdcscid/
NCrCrl (NOD-SCID) mice aged 5-6 weeks were purchased from Charles River Laboratories. Female NOD-scid IL2Rgnull (NSG)
mice aged 6-7 weeks were purchased from Jackson Laboratory.
Laboratory animals
The study did not involve wild animals.
Wild animals
The study did not involve samples collected from the field.
Field-collected samples
Ethics oversight
All animal work was approved and done in accordance with the Institutional Animal Care and Use Committees of University of
Arkansas for Medical Sciences, University of Florida, and MD Anderson Cancer Center with the exception of the pharmacokinetic
(PK) studies that were done by BioDuro (San Diego, CA, USA), a global contract research organization, through a contact. All
animal studies were complied with the ethical regulations and humane endpoint according to the NIH Guidelines for the Care
and Use of Laboratory Animals.
Note that full information on the approval of the study protocol must also be provided in the manuscript.
3

Flow Cytometry
Plots
Confirm that:
The axis labels state the marker and fluorochrome used (e.g. CD4-FITC).
The axis scales are clearly visible. Include numbers along axes only for bottom left plot of group (a 'group' is an analysis of identical markers).
All plots are contour plots with outliers or pseudocolor plots.
A numerical value for number of cells or percentage (with statistics) is provided.
Methodology
Peripheral blood samples were collected through retro orbital bleeding, and red blood cells were lysed at room tempreture for
30 min. The samples were washed once with PBS, and stained in antibody cocktail for 15 min at room temperature. After
washing with PBS, the samples were re-suspended in PBS plus DAPI and immediately analyzed by flow cytometry.
Sample preparation
Gallios, Beckman Coulter
Instrument
FlowJo, V10.
Software
All samples were collected at least 50,000 events of live cells.
Cell population abundance
Gating strategy
FSC/SSC was used to distinguish the cell size. DAPI-negative was used to gate the live cells followed by mCD45 and hCD45 to gate
mouse and human white blood cells, respectively.
Tick this box to confirm that a figure exemplifying the gating strategy is provided in the Supplementary Information.
4