A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and elicits inflammatory responses in tumor explants

Article

A potent and selective PARP14 inhibitor decreases

protumor macrophage gene expression and elicits

inﬂammatory responses in tumor explants

Graphical abstract

Authors

Laurie B. Schenkel, Jennifer R. Molina,

Kerren K. Swinger, ...,

Victoria M. Richon, Heike Keilhack,

Kevin W. Kuntz

Correspondence

kkuntz@ribontx.com

In brief

Schenkel et al. disclose the ﬁrst highly

potent and selective inhibitor of PARP14,

RBN012759, an in vitro and in vivo

chemical probe. RBN012759 reverses IL-

4-driven protumor gene expression in

macrophages and induces an

inﬂammatory mRNA signature in primary

human tumor explants, suggesting utility

of PARP14 inhibitors for treating cancer.

Highlights

d

The ﬁrst highly potent and selective inhibitor of PARP14,

RBN012759, is described

d

The proﬁle of RBN012759 enables its use as an in vitro and

in vivo chemical probe

d

RBN012759 reversed IL-4-driven protumor gene expression

An inﬂammatory gene signature was activated with

RBN012759 treatment

Schenkel et al., 2021, Cell Chemical Biology 28, 1–11

August 19, 2021 ª 2021 Elsevier Ltd.

https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

A potent and selective PARP14 inhibitor decreases

protumor macrophage gene expression and elicits

inﬂammatory responses in tumor explants

Laurie B. Schenkel,^1,3Jennifer R. Molina,²Kerren K. Swinger,^1,4Ryan Abo,^2,5Danielle J. Blackwell,¹Alvin Z. Lu,²

Anne E. Cheung,^2,6W. David Church,¹Kaiko Kunii,²Kristy G. Kuplast-Barr,²Christina R. Majer,¹Elena Minissale,²

Jan-Rung Mo,²Mario Niepel,²Christopher Reik,^1,7Yue Ren,¹Melissa M. Vasbinder,¹Tim J. Wigle,¹Victoria M. Richon,^1,2

Heike Keilhack,²and Kevin W. Kuntz^1,8,

*

¹Department of Molecular Discovery, Ribon Therapeutics, Inc., Cambridge, MA 02140, USA

²Department of Biological Sciences, Ribon Therapeutics, Inc., Cambridge, MA 02140, USA

³MOMA Therapeutics, Cambridge, MA 02142, USA

⁴Xilio Therapeutics, Waltham, MA 02451, USA

⁵Obsidian Therapeutics, Cambridge, MA 02138, USA

⁶A2Empowerment, Arlington, MA 02474, USA

⁷Bain & Company, Boston, MA 02116, USA

⁸Lead contact

*Correspondence: kkuntz@ribontx.com

https://doi.org/10.1016/j.chembiol.2021.02.010

SUMMARY

PARP14 has been implicated by genetic knockout studies to promote protumor macrophage polarization

and suppress the antitumor inﬂammatory response due to its role in modulating interleukin-4 (IL-4) and inter-

feron-g signaling pathways. Here, we describe structure-based design efforts leading to the discovery of a

potent and highly selective PARP14 chemical probe. RBN012759 inhibits PARP14 with a biochemical half-

maximal inhibitory concentration of 0.003 mM, exhibits >300-fold selectivity over all PARP family members,

and its proﬁle enables further study of PARP14 biology and disease association both in vitro and in vivo. In-

hibition of PARP14 with RBN012759 reverses IL-4-driven protumor gene expression in macrophages and in-

duces an inﬂammatory mRNA signature similar to that induced by immune checkpoint inhibitor therapy in

primary human tumor explants. These data support an immune suppressive role of PARP14 in tumors and

suggest potential utility of PARP14 inhibitors in the treatment of cancer.

INTRODUCTION

Based on the proposed opposing effects on IFN-g- and IL-4-

mediated gene expression, PARP14 has been implicated in

PARPs (poly(ADP-ribose) polymerases) are a family of 17 en- macrophage polarization. Genetic PARP14 inactivation in mac-

zymes known to regulate fundamental cellular processes, rophages skews them toward a pro-inﬂammatory IFN-g-driven

including gene expression, protein degradation, and multiple M1 phenotype while reducing the IL-4-driven M2 phenotype

cellular stress responses (Vyas and Chang, 2014). PARPs utilize (Iwata et al., 2016). In line with these results, PARP14 knockout

nicotinamide adenine dinucleotide (NAD⁺) to post-translationally (KO) showed therapeutic effects in Th2 cytokine-driven models

modify substrates, including DNA, RNA, and protein, via transfer of allergic airway disease (Cho et al., 2013; Mehrotra et al., 2013).

of a single ADP-ribose unit (mono-ADP-ribosylation or MARyla-

Several PARP14 inhibitors have been previously disclosed,

tion) in the case of the PARP-monoenzyme subfamily, or chains including compounds that bind the NAD⁺pocket in the catalytic

of ADP-ribose units (poly-ADP-ribosylation or PARylation) in the domain (Ekblad et al., 2015; Andersson et al., 2012; Peng et al.,

case of the PARP-polyenzymes (Cohen and Chang, 2018). 2017; Upton et al., 2017; Yoneyama-Hirozane et al., 2017; Wang

PARP14 is

a macrodomain-containing PARP-monoenzyme et al., 2014; Iansante et al., 2015; Holechek et al., 2018), as well

that was originally identiﬁed as BAL2 (B aggressive lymphoma as those that bind the PARP14 macrodomains (Ekblad et al.,

2), a gene associated with poor outcome of diffuse large B cell 2018; Moustakim et al., 2018; Schuller et al., 2017). While in

lymphoma (Juszczynski et al., 2006; Aguiar et al., 2000, 2005). some cases the catalytic inhibitors inhibit PARP14 with

PARP14 is an interferon-stimulated gene (ISG), whose mRNA biochemical half-maximal inhibitory concentration (IC₅₀) values

expression is increased by all three types of interferons (Rusi- in the sub-micromolar range, achieving signiﬁcant PARP14

nova et al., 2013). It has been identiﬁed as a negative down- selectivity has remained challenging due to high sequence sim-

stream regulator of interferon-g (IFN-g) and STAT1 transcription, ilarity among the PARP family NAD⁺binding pockets and lack of

and as a positive regulator of the interleukin-4 (IL-4) and STAT6 access to a complete panel of sensitive, high-throughput PARP

signaling cascade (Goenka et al., 2007; Mehrotra et al., 2011). assays to fully evaluate selectivity (Thorsell et al., 2017;

Cell Chemical Biology 28, 1–11, August 19, 2021 ª 2021 Elsevier Ltd.

1

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Figure 1. Structure-based optimization and PARP family selectivity of PARP14 inhibitors

(A) Structures of compounds 1–4.

(B) PARPs structure and sequence alignments for key residues in the active site. The numbering of amino acid residues is in reference to PARP14 residue

numbering. PARPs that have been characterized by X-ray crystallography are highlighted in bold letters and were aligned based on their three-dimensional

structures. PARPs that have not been structurally characterized were aligned based on sequence. Residues that could not be predicted by sequence alignment

are represented as an X. Canonical HYE motifs for PARP-polyenzymes are colored orange. For PARP-monoenzymes, the residue that overlaps with the canonical

E residue is colored in green (Leu1782 for PARP14) and comprises part of the MAR/PAR pair of residues (1782/1721). Conserved residues that contribute to

binding of most PARP ligands are colored in magenta. The residue in the PARP14 D loop that contributes to selectivity (Y1701) is in yellow and the Asp/Ser pair

(1685/1688) targeted for PARP14 selectivity is in purple.

(C) PARP14 TR-FRET IC₅₀values, represented as geometric means SD, and PARP family selectivity of compounds 2, 3, and RBN012759, represented as fold

selectivity for PARP14.

(D and E) Cartoon diagrams of PARP14 crystal structures with ligands drawn as cyan sticks, compound 2 (D) and RBN012759 (E). Hydrogen bonds are drawn as

red dashes. Residues are colored as described in (C)

Schweiker et al., 2018). Limited PARP family selectivity data for the quinazolinone NH and Gly1683 backbone carbonyl, the qui-

previous PARP14 inhibitors has been disclosed and, when nazolinone carbonyl and Ser1722 side-chain OH, and a pi-stack-

measured, the selectivity for PARP14 over the PARP-monoen- ing interaction with Tyr1727 (Figure 1D). The thiopyridine moiety,

zymes in particular is typically less than 10-fold. We sought to which is positioned in an area of the pocket lined on one side by

identify a highly potent and selective inhibitor of PARP14 to the ﬂexible D loop (Wahlberg et al., 2012), does not make inter-

enable exploration of PARP14-mediated biology. Here, we actions with the protein and the D loop is disordered.

describe the discovery and characterization of an in vitro and

Despite having relatively weak PARP14 potency and poor

in vivo chemical probe, RBN012759 (1), which inhibits PARP14 PARP family selectivity (Figure 1C; Table S2), 2 provided a hit

catalytic activity with an IC₅₀of 0.003 mM and with >300-fold that enabled modular synthesis of analogs to probe PARP14 po-

selectivity over all PARP family members. Catalytic inhibition of tency and selectivity structure-activity relationships (SAR). Com-

PARP14 with RBN012759 leads to repolarization of M2-like pound 2 also presented vectors that were well suited to pursue

macrophages to a less immunosuppressive phenotype and in- structure-based hypotheses for improving potency and selec-

duction of inﬂammation markers in kidney cancer explants.

tivity in two key regions of the binding pocket. The thiopyridine

moiety of 2 is near an area of the pocket formed on one side

by Ser1688 and Asp1685, residues that are unique among the

PARP family to only PARP14 and PARP15 (Figure 1B). Ser1688

and Asp1685 engage each other via a hydrogen bond and

RESULTS

Structure-based optimization to identify RBN012759

Screening a PARP-preferring compound collection with a previ- interact with a water molecule, an interaction that we hypothe-

ously described PARP14 self-modiﬁcation DELFIA assay (Wigle sized could be displaced by an inhibitor. Optimization of the thi-

et al., 2019) identiﬁed compound 2 (Figure 1A), a close structural opyridine portion of 2 included derivatives designed to engage

analog of a known pan-PARP inhibitor (Wahlberg et al., 2012), the Asp1685/Ser1688 motif and add a favorable interaction to

with an IC₅₀of 1 mM. An X-ray co-crystal structure of 2 bound improve potency and selectivity. Diverse thioether substituents

to PARP14 conﬁrmed that it binds in the NAD⁺pocket and were synthesized while holding the methylquinazolinone con-

makes several interactions, including hydrogen bonds between stant. PARP14 potency for all analogs was determined using a

2

Cell Chemical Biology 28, 1–11, August 19, 2021

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

PARP14 probe displacement assay with time-resolved ﬂuores- RBN012759 X-ray co-crystal structures (Figure S1B) illustrates

cence energy transfer (TR-FRET) readout (Wigle et al., 2020), that optimization in the MAR/PAR difference region had little ef-

and to build selectivity SAR, analogs were screened in a com- fect on the thioether portion of the molecule, as the trans-cyclo-

plete panel of PARP family biochemical assays (Wigle et al., hexanol groups are well overlaid. Based on its superior PARP14

2019). The trans-cyclohexanol-containing compound 3 was pre- potency and PARP family selectivity, we characterized

pared (Figure 1A), and it provided a 3-fold improvement in RBN012759 further to conﬁrm its suitability as a chemi-

PARP14 potency (IC₅₀= 0.3 mM) versus 2 with a signiﬁcant cal probe.

improvement in selectivity over all PARP-monoenzymes (Fig-

ure 1C; Table S2).

Biophysical and pharmacokinetic characterization of

A second region of the PARP14 binding pocket targeted to RBN012759

gain selectivity over the PARP-polyenzymes, referred to as the The PARP14 binding afﬁnity of RBN012759 (K_d= 0.002 mM) was

MAR/PAR difference region, due to key residue differences be- determined by surface plasmon resonance (SPR) using the im-

tween the PARP-polyenzymes and PARP-monoenzymes, is mobilized catalytic domain of PARP14 (Figure 2A), which corre-

located at a site deﬁned by Tyr1721 and Leu1782. While the lates well with the IC₅₀(0.003 mM) determined from the PARP14

PARP-polyenzymes contain a salt bridge formed by Lys and TR-FRET probe displacement assay. RBN012759 retained at

Glu residues (Vyas et al., 2014), where the Glu is part of the least this level of potency when measured in a full-length

conserved HYE motif, most PARP-monoenzymes contain hy- PARP14 enzymatic auto-modiﬁcation assay in the DELFIA

drophobic residues at these positions (Figure 1B). Seeking to format, in which it returned an IC₅₀value that was below the limit

exploit this polarity difference between the PARP-monoenzymes of quantiﬁcation of the assay (<0.005 mM). In a thermal shift

and -polyenzymes, we incorporated hydrophobic moieties into assay, RBN012759 stabilized the human PARP14 catalytic

the quinazolinone core that would project into the MAR/PAR dif- domain by 11^ꢀC (Figure 2B). RBN012759 inhibited a mouse

ference region. Keeping the trans-cyclohexanol moiety of 3 in PARP14 TR-FRET probe displacement assay with an IC₅₀value

place, it was determined that substitution off the 7-position of (0.005 mM) that is comparable with its human PARP14 potency,

the quinazolinone core improved potency. Optimization led to suggesting that RBN012759 is also a suitable chemical probe for

incorporation of a 7-cyclopropylmethanol substituent and a evaluating the role of PARP14 in mouse systems. RBN012759

ﬂuorine atom at the 5-position of the quinazolinone, with no sub- has solubility at pH 7.4 of 198 mM, with good permeability and

stitution at the 6- and 8-positions, to give compound 1 minimal efﬂux in MDR1 overexpressing MDCK cells (P_app,A-B

(RBN012759, Figure 1A). RBN012759 has an IC₅₀of 0.003 mM 15 3 10^ꢁ6cm/s, efﬂux ratio = 2) and Caco-2 cells (P_app,A-B

=

in the PARP14 TR-FRET probe displacement assay and has 19 3 10^ꢁ6cm/s, efﬂux ratio = 1). After intravenous administration

>300-fold selectivity over the other PARP-monoenzymes and in CD-1 mice, RBN012759 had a clearance of 54 mL/min/kg with

>1,000-fold selectivity over the PARP-polyenzymes (Figure 1C; a half-life of 0.4 h and steady state volume of distribution of 1.4 L/

Table S1). Several previously reported PARP14 inhibitors were kg. RBN012759 demonstrated oral bioavailability (30% F) at

proﬁled in our PARP assay panel and were substantially less 100 mg/kg that decreased to 10% F at 500 mg/kg. Despite being

potent and selective PARP14 inhibitors than RBN012759 (Table modestly orally bioavailable, RBN012759 maintained exposure

S4).

that provided 90-fold coverage of the mouse PARP14-free IC₅₀

An X-ray co-crystal structure of RBN012759 bound to PARP14 value (15 ng/mL) at 8 h following a 500 mg/kg oral dose.

illustrates its binding mode (Figure 1E). RBN012759 binds to the

PARP14 NAD⁺pocket and retains the hydrogen bonds and pi- Cellular PARP14 target engagement with RBN012759

stacking interaction observed with 2. Relative to the thiopyridine The ability of RBN012759 to inhibit intracellular PARP14 was

moiety of 2, the conformation of the thio-trans-cyclohexanol conﬁrmed via orthogonal methods. A PARP14 NanoBRET probe

group of RBN012759 is shifted toward the Asp1685 and displacement assay was developed in 293T cells that measures

Ser1688 residues, and the alcohol of the cyclohexyl group inter- inhibition of bioluminescence resonance energy transfer (BRET)

acts with Asp1685, displacing the water molecule observed in between NanoLuc-tagged PARP14 and an active site probe

the structure of PARP14 bound to 2. This conformational change containing a ﬂuorescent group. RBN012759 exhibited concen-

of the thioether enables the D loop in the RBN012759 structure to tration-dependent inhibition of the BRET signal with an IC₅₀value

adopt an ordered, ‘‘closed’’ conformation in which the trans-cy- of 0.003 mM (Figure 2C). When tested across a panel of PARP-

clohexanol also engages the D loop residues Tyr1701 and monoenzyme NanoBRET assays (PARP7, PARP10, PARP11,

Lys1704 in van der Waals interactions. Overlay of the compound PARP12, and PARP16), RBN012759 showed similar levels of po-

2/PARP14 and RBN012759/PARP14 X-ray co-crystal structures tency as that observed in enzyme assays (Table S1), suggesting

(Figure S1A) illustrates that a steric clash would occur between that it is also highly selective in cells.

the thiopyridine of 2 in its binding conformation and the D loop

in the closed conformation.

Since PARP14 is a known ISG, assays were developed in IFN-

g-stimulated mouse RAW264.7 cells and primary human macro-

The co-crystal structure of RBN012759 with PARP14 also phages to demonstrate target engagement on endogenous

highlights determinants for the observed improvement in PARP14. Treatment with RBN012759 led to a concentration-

PARP-polyenzyme selectivity (Figure 1E). The 7-cyclopropyl- dependent increase in PARP14 protein, a phenomenon that

methanol group makes several van der Waals interactions in was described previously in other models (Yoneyama-Hirozane

the hydrophobic region lined by the MAR/PAR difference resi- et al., 2017; Scaltriti et al., 2009), with half-maximal effective con-

dues and part of the D loop, including residues Leu1782, centration values of 0.01 mM in RAW264.7 cells and 0.08 mM in

Ala1706, and Tyr1714. Overlay of the compound

3

and human macrophages (Figures 2D and 2E).

Cell Chemical Biology 28, 1–11, August 19, 2021

3

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Figure 2. Biophysical characterization and intracellular engagement of PARP14 with RBN012759

(A) PARP14 SPR assay data. PARP14 catalytic domain having an N-terminal biotin-Avi tag was immobilized on a streptavidin-coated sensor chip and

RBN012759 binding was analyzed using single-cycle kinetics.

(B) PARP14 thermal shift assay data. The red curve represents PARP14 melting in the presence of 3% DMSO, and the blue curve represents PARP14 melting in

the presence of 167 mM RBN012759. Data are representative of two independent experiments.

(C) Ten-point IC₅₀curve for RBN012759 in the PARP14 NanoBRET biophysical cellular assay. The NanoBRET half-maximal effective concentration (EC₅₀) is

0.003 mM SD 0.0009 mM from six independent experiments.

(D) Western blot and IC₅₀curve with treatment of RAW264.7 cells with varying concentrations of RBN012759 following stimulation with 20 ng/mL IFN-g.

(E) Western blot and EC₅₀curve with treatment of primary human macrophages with varying concentrations of RBN012759 following stimulation with 20 ng/mL

IFN-g.

(F) PARP14 and MAR/PAR protein expression by western blot analysis of human primary macrophages stimulated with 20 ng/mL IFN-g and treated with

increasing concentrations of RBN012759 for 48 h.

We next evaluated the ability of RBN012759 to inhibit Figure S2D). The same effect was not observed with niraparib

PARP14-speciﬁc self-MARylation. MARylation was detected treatment, a PARP1 inhibitor, in IFN-g-stimulated CFPAC-1 cells

by immunoblot using an antibody that binds to both MAR and (Figure S2E). The major effect of RBN012759 was observed un-

PAR (Lu et al., 2019) and a PARP14 antibody in IFN-g-stimulated der IFN-g-stimulated conditions rather than unstimulated base-

primary human macrophages (Figure 2F). RBN012759 line conditions (Figures S2E and S2D). Altogether these data

decreased the MAR/PAR signal corresponding to PARP14- show that the reduced MARylation signal after RBN012759

self-MARylation in a concentration-dependent manner, further treatment corresponds to PARP14 self-MARylation.

supporting its activity on endogenous PARP14 (Figure 2F). A mi-

To further demonstrate target engagement, compound 4 (Fig-

nor MAR/PAR signal remained after treatment, suggesting that ure 1A), a PARP14 catalytic inhibitor that is a structural analog of

PARP14 may be a substrate of additional PARP enzymes (Fig- RBN012759 with a similar potency and selectivity proﬁle (Table

ure 2F). To reinforce the speciﬁcity of PARP14 autoMARylation, S2), was used in a substrate identiﬁcation experiment with

we used the CFPAC-1 cell line due to its high endogenous IFN-g-stimulated primary human macrophages. Proteins identi-

PARP14 baseline level and response to IFN-g stimulation by ﬁed in the vehicle-treated samples and reduced in the PARP14

increasing PARP14 and ADP-ribosylation. First, endogenous inhibitor-treated samples were considered as potential

PARP14 was immunoprecipitated from IFN-g-stimulated PARP14 substrates (Figure 3A). A subset of interferon-related

CFPAC-1 cells and probed for MAR/PAR (Figure S2C). These proteins, including OAS2, STAT2, STAT1, and MX1, showed

data demonstrated that PARP14 auto-MARylation is stimulated reduced spectral counts in the PARP14 inhibitor-treated sam-

by IFN-g and robustly inhibited by RBN012759 (Figure S2C). In ples, suggesting that these proteins may be potential substrates

addition, CFPAC-1 cells were stimulated with IFN-g to increase of PARP14 (Figure 3B; Table S5). Our data identiﬁed previously

PARP14 expression and MARylation and treated with increasing reported PARP14 putative substrates, such as STAT1, pointing

concentrations of RBN012759 (Figure S2D). RBN012759 to its role in macrophage polarization and immune responses

decreased the MARylation signal in a concentration-dependent (Iwata et al., 2016; Caprara et al., 2018; Carter-O’connell

manner, similar to that observed in human macrophages (see et al., 2018).

4

Cell Chemical Biology 28, 1–11, August 19, 2021

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Figure 3. Compound 4 is active in cells and inhibits MARylation of inﬂammation-related proteins

(A) MAR/PAR and PARP14 protein expression signals were determined by immunoblot analysis in human primary macrophages prepared as described in

Figure 2E and treated with 10 mM of compound 4. Protein lysate pre- and post-immunoprecipitation with Af1521-conjugated magnetic resin was analyzed by

immunoblot and putative PARP14 substrates are marked by the white asterisks (*).

(B) Scatterplot of log-transformed spectral counts from PARP14 substrate identiﬁcation screen in primary human macrophages. Interferon GeneSet genes

enriched in both substrate and transcriptional experiments upon PARP14 inhibitor treatment are highlighted in red.

In vivo target engagement with RBN012759 M1-like polarization may involve domains other than the catalytic

The ability of RBN012759 to engage PARP14 in vivo was also domain of PARP14 (Iwata et al., 2016).

evaluated. C57BL/6 mice were treated with RBN012759 at 300

IL-4-stimulated human primary macrophages were used to

and 500 mg/kg BID PO for 7 days, during which it was well toler- investigate if catalytic inhibition of PARP14 decreased protumor

ated with no signiﬁcant body weight loss observed. The mice M2-like macrophage polarization, as suggested by the pub-

were sacriﬁced at 12 h following the ﬁnal dose, and spleen, a tis- lished PARP14 KO data in mouse macrophages (Iwata et al.,

sue with endogenous PARP14 expression (Cho et al., 2009), was 2016) (Figure 4A). Gene expression changes of IL-4-polarized

collected for determination of PARP14 protein levels. As M2-like macrophages treated with RBN012759 clustered with

observed with treatment of cells in vitro (see Figure 2D), treat- naive-M0 macrophage signatures (Figure 4B), suggesting that

ment with RBN012759 led to an increase in PARP14 protein PARP14 inhibition reversed the IL-4-driven skewing closer to

in vivo. The 500 mg/kg treatment group showed increased the naive M0 phenotype. Unsupervised clustering analysis un-

PARP14 protein, while the 300 mg/kg group did not, correlating covered ﬁve distinct clusters of differentially expressed genes

with RBN012759 plasma exposures at the 12 h time point in that distinguished IL-4 vehicle (M2-like) versus RBN012759-

which the mouse PARP14-free IC₅₀value was increased by 6- treated macrophages (Figure 4C; Table S6). Cluster 4 revealed

and <1-fold, respectively (Figures S2F–S2H). These data support signiﬁcantly increased genes, including CD209 in RBN012759-

the use of RBN012759 as an in vivo chemical probe that selec- treated M2-like macrophages. CD209 is a pathogen-recognition

tively engages PARP14 in tissue.

receptor that mediates the endocytosis function of macro-

phages (Schulz et al., 2019). Several markers typically expressed

by inﬂammatory actively phagocytic macrophages, MMP-1

(cluster 4) and CLTC (cluster 1) (Huang et al., 2012; Steenport

Catalytic inhibition of PARP14 reverses IL-4-driven

gene expression in M2-like macrophages

With a suitable chemical probe in hand, we next investigated et al., 2009; Gong et al., 2008; Liu et al., 2014), were also upregu-

whether PARP14 catalytic inhibition phenocopies the published lated with RBN012759. Cluster 3 revealed a group of downregu-

effects of PARP14 knockout on IFN-g (M1-like)- and IL-4 (M2- lated mitochondria metabolism-related genes, such as ATP5D,

like)-driven macrophage polarization. Treatment of primary hu- ETFB, and VDAC2. These data align with studies suggesting

man macrophages with IFN-g elicited typical interferon response that inﬂammatory macrophages increase glycolytic activity while

gene expression changes (Figure S3A). RBN012579 co-treat- decreasing oxidative phosphorylation of mitochondria (Jung

ment for 6 and 24 h did not change IFN-g-driven gene expression et al., 2018; Viola et al., 2019). Cluster 5 revealed upregulation

dramatically (Figure S3B). This contrasts with data published in of a key inﬂammatory chemokine, CCL13, which attracts mono-

mouse macrophages, where genetic PARP14 KO exacerbates cytes, lymphocytes, basophils, and eosinophils to inﬂamma-

the IFN-g response and suggests that the effects on the tory sites.

Cell Chemical Biology 28, 1–11, August 19, 2021

5

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Figure 4. Inhibition of PARP14 with RBN012759 dampens the M2-like macrophage protumor phenotype

(A) M2-like polarization of primary human macrophages with IL-4 (15 ng/mL; M2-like) and treatment with RBN012759 to determine impact on gene expression.

(B) Heatmap and hierarchical clusters representing the transcriptional distances between vehicle and RBN012759-treated samples based on a Poisson distance

metric.

(C) Heatmap and hierarchical clusters using normalized gene expression values from differentially expressed genes with IL-4 + vehicle or RBN012759 + IL-4

treatments.

(D) Levels of IL-10, CCL24, and IL-1ra in naive M0 or M2-like (IL-4 + IL-13) macrophage culture supernatants (three biological replicates) were determined 48 h

after treatment with RBN012759. Statistical signiﬁcance we determined by Welch’s t-test.

RBN012759 treatment also decreased the protein level of regulator of the M2-like protumor phenotype. These data sug-

three M2-like proteins, IL-10, CCL24, and IL1-ra, in the superna- gest that catalytic inhibition of PARP14 with RBN012759

tants of IL-4-stimulated macrophages to a level closer to changes M2-like macrophages toward a less immunosuppres-

baseline (i.e., not stimulated with IL-4, Figure 4D). IL-10 is an sive phenotype.

IL-4-elicited anti-inﬂammatory protumor cytokine that is

secreted by various immune cell types in the tumor microenvi- RBN012759 promotes inﬂammatory gene expression in

ronment (TME), including tumor-associated macrophages human kidney cancer tumor explants

(Cheng et al., 2019). CCL24 and IL1-ra are also associated The observation that treatment with RBN012759 led to a

with the M2-like phenotype (Shapouri-Moghaddam et al., decreased M2-like immunosuppressive macrophage phenotype

2018; Pechkovsky et al., 2010; Makita et al., 2015). As expected, prompted us to explore a model system that includes a complete

IL-4 alone or in combination with IL-13 increased STAT6 phos- tumor immune inﬁltrate and may more closely resemble the hu-

phorylation in M2-like macrophages; however, no changes in man tumor setting. To do so, PARP14 gene expression was eval-

phospho-STAT6 were observed with RBN012759 treatment at uated in a variety of tumors in The Cancer Genome Atlas (TCGA)

the evaluated time point (Figure S3F). This is in contrast with a and found to be signiﬁcantly overexpressed in kidney renal clear

previously reported negative effect on STAT6 phosphorylation cell carcinoma (KIRC) as compared with normal kidney tissue

with PARP14 genetic depletion and points to a potential differ- (Figure S4A). Moreover, analysis of TCGA KIRC patient data re-

ence between catalytic inhibition and complete protein depletion vealed an association between high PARP14 gene expression

(Iwata et al., 2016).

and a signiﬁcantly lower probability of survival (Figure S4B). Tu-

Apart from a minor subset of genes, RBN012759 treatment mor PARP14 expression positively associated with signatures of

mostly led to negative changes in gene expression in the M2- various immune cells, including macrophages, and negatively

like macrophage model system (Figures 4C, S4C, and S4D). associated with tumor purity, as assessed by the Tumor Immune

RBN012759 effects were observed across genes with both Estimation Resource (Li et al., 2016) (Figure S4C). This suggests

high and low expression levels (Figure S3E). The ﬁnding that that the increased PARP14 expression in kidney cancer is at

PARP14 catalytic inhibition reverses IL-4-driven gene expres- least in part due to increased expression in the tumor immune

sion in macrophages suggests a role of PARP14 as a positive inﬁltrate. For these reasons, we investigated the effects of

6

Cell Chemical Biology 28, 1–11, August 19, 2021

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Figure 5. RBN012759 induces antitumor inﬂammation in kidney cancer tumor explant models

(A) Cartoon depiction of renal clear cell carcinoma tumor explant culture assay.

(B) Biplots of the treatment-based gene expression changes comparing PARP14 inhibitor treatment (1 mM RBN012759) and anti-PD-1 (3 mg/mL) + anti-CTLA-4

(3 mg/mL) treatment for the ten bulk explant tumor samples.

(C) Heatmap and hierarchical clusters showing expression changes in ten bulk explant tumor samples from PARP14 inhibitor treatment (1 mM RBN012759) or

anti-PD-1 (3 mg/mL) + anti-CTLA-4 (3 mg/mL) treatment.

(D) Barplots indicating the treatment-based expression fold changes for select genes from PARP14 inhibitor treatment with RBN012759 (red) and anti-PD-1 +

anti-CTLA-4 treatment (blue) for the 10 bulk explant tumor samples.

RBN012759 on KIRC human tumor explants (Page et al., 2018; 60% (Santoni et al., 2018; Motzer et al., 2018). Altogether, these

Mediavilla-Varela et al., 2018). Ten KIRC tumors were treated data revealed that PARP14 inhibition with RBN012579 elicited

with RBN012759, vehicle, or the combination of anti-PD-1 and similar gene expression changes as ICI combination therapy,

anti-CTLA-4 (immune checkpoint inhibitor [ICI]), and RNA was an approved treatment for kidney cancer, in subsets of kidney

isolated 24 h after treatment initiation for gene expression anal- cancer tumor explants (Motzer et al., 2018).

ysis (Figures 5A and S4D). A positive correlation of gene expres-

sion changes between ICI and RBN012759 treatments was DISCUSSION

observed across all tumors tested (Figure 5B), but neither

RBN012759- nor ICI-induced expression changes correlated Here, we describe the development of RBN012759, a chemical

with those elicited by co-treatment with phorbol-12-myristate- probe that inhibits PARP14 with an IC₅₀value of 0.003 mM and

13-acetate and ionomycin, a general inﬂammatory stimulus Fig- has high selectivity over all PARP family enzymes. Medicinal

ure S4E). Unsupervised clustering analysis of gene expression chemistry efforts that began with an unselective, micromolar

data revealed that a subset of RBN012759-treated tumors clus- screening hit were enabled by a detailed understanding of

tered with ICI-sensitive tumors (Figure 5C). Among the genes PARP family NAD⁺binding pockets. Hit 2 provided a template

signiﬁcantly modulated with RBN012759 were several chemo- for targeting interactions in two key regions of the PARP14

kines associated with antitumor inﬂammation, including NAD⁺binding pocket. The Ser1688/Asp1685 motif, which is

CXCL11, CCL19, and CCL13 (Figure 5D). Toll-like receptor 7 unique to PARP14 and PARP15, was targeted to improve

(TLR7), a key component of innate and adaptive immunity, was PARP14 potency and selectivity over the other PARP-monoen-

also upregulated with RBN012759 or ICI treatment in tumors 5 zymes. The MAR/PAR difference region, which contains a

and 9 (Figure 5D). The immune checkpoint ligand PD-L2, a conserved HYE motif in which the Glu forms a salt bridge in

known marker of T cell activation, was upregulated in tumors 1 the PARP-polyenzymes but contains HY(I/L/Y) in the PARP-

and 9 upon treatment with either RBN012759 or ICI (Figure 5D). monoenzymes, was targeted with hydrophobic substituents. It

The variability in RBN012759-driven gene expression changes was anticipated that this strategy would disfavor interactions

that correlate to response to ICI treatment is not entirely unex- with the PARP-polyenzymes, and indeed a series of PARP11 in-

pected, since KIRC tumors are characterized by highly variable hibitors for which selectivity over the PARP-polyenzymes was

tumor inﬁltrating lymphocytes (TILs) and macrophage inﬁltration obtained in a similar manner has been described (Kirby

and are known to have response rates ranging from 20% up to et al., 2018).

Cell Chemical Biology 28, 1–11, August 19, 2021

7

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Optimization of the thiopyridine moiety in compound 2 to inhibition results in a less immunosuppressive phenotype. Of

target displacement of the water atom bound to Ser1688/ note, catalytic inhibition of PARP14 in human macrophages did

Asp1685 led to 3, which incorporates a trans-cyclohexanol moi- not entirely phenocopy the positive stimulatory effect of

ety that was maintained in RBN012759. X-ray co-crystal struc- PARP14 KO on IFN-g-driven gene expression observed in mouse

tures of 3 and RBN012759 with PARP14 show an interaction macrophages (Iwata et al., 2016), at least under the conditions

between the alcohol and Asp1685. The change of conformation and time points tested in our studies. This points to a possible

of the thioether in going from the pyridine of 2 to the trans-cyclo- non-catalytic function of PARP14 in the IFN-g pathway in normal

hexanol also enabled the PARP14 D loop to adopt an ordered, macrophages. However, RBN012579 increased markers associ-

closed conformation, which may be preferred, as both 3 and ated with active phagocytosis, such as CD209, MMP-1, and

RBN012759 engage the D loop in favorable van der Waals inter- CLTC, in addition to reversing IL-4-driven gene expression.

actions. The interaction with Tyr1701 may contribute to the This suggests thatPARP14 inhibition has thepotential to increase

observed selectivity over PARP15, which has a Cys residue at the antitumor activity of macrophages in the tumor microenviron-

this position (Figure 1B). Optimization of the portion of the inhib- ment. Although cancer immunotherapy currently beneﬁts a sig-

itors that binds in the Asp/Ser region of the pocket enabled inter- niﬁcant fraction of cancer patients, many have primary or develop

actions with the protein that both increased PARP14 potency secondary (immunotherapy induced) resistance, which has

and decreased potency against the other PARP-monoenzymes. been, in part, attributed to the presence of macrophages in the

Efforts to project hydrophobic groups into the MAR/PAR dif- tumor microenvironment (Pathria et al., 2019). Targeting tumor-

ference region led to incorporation of a cyclopropylmethyl ether associated macrophages has been proposed as a viable thera-

at the 7-position of the quinazolinone, which gave signiﬁcant peutic strategy to reverse immunosuppression. Repolarization

improvements in selectivity favoring PARP14 over the PARP- of tumor-associated macrophages is one of the most promising

polyenzymes coupled with additional improvement in PARP14 strategies currently being explored and targeting PARP14 may

potency. Since the thioether moieties of 2 and RBN012759 be a viable strategy for this therapeutic hypothesis.

adopt the same conformation in the binding pocket (Figure S1B),

The immunosuppressive role of PARP14 was evaluated in hu-

improvements in potency and selectivity are attributed to inter- man tumor explants to recapitulate the human tumor setting,

actions in the MAR/PAR difference region. RBN012759 is where PARP14 is overexpressed, more closely than in cellular

>300-fold selective over all PARP-monoenzymes and demon- or mouse systems. Catalytic inhibition of PARP14 with

strates similarly high selectivity over the PARP-polyenzymes RBN012759 in KIRC human tumor explants revealed signiﬁcant

(>1,000-fold), a substantial improvement over previously re- increases in several pro-inﬂammatory chemokines, including

ported inhibitors (Table S4).

CCL13, which was also upregulated in cultured normal macro-

Additional characterization of RBN012759 demonstrated its phages after PARP14 inhibition. TLR7, also upregulated with

utility as both an in vitro and in vivo chemical probe. RBN012759 treatment, is being explored as a target to achieve

RBN012759 has sufﬁcient solubility at pH 7.4, is highly perme- macrophage repolarization via agonist molecules that stimulate

able, and is not likely a substrate of Pgp or BCRP. Multiple the innate immune response (Pathria et al., 2019). Intriguingly,

orthogonal methods were used to demonstrate that PARP14 inhibition elicited similar gene expression changes in tu-

RBN012759 inhibits PARP14 intracellularly. Its potency in the mor explants as ICI combination therapy, an active regimen in

PARP14 NanoBRET probe displacement assay correlated well kidney cancer (Motzer et al., 2018). This suggests potential utility

with both enzymatic assay data and binding afﬁnity generated of PARP14 inhibitors in this cancer type as single agent and/or in

by SPR, and the selectivity for PARP14 observed in enzyme combination with ICI therapy.

assays was reproduced in a panel of PARP-monoenzyme bio-

The discovery of RBN012759, a potent and highly selective

physical cellular assays. Robust, concentration-dependent sta- in vitro and in vivo chemical probe of PARP14, enabled the ef-

bilization of PARP14 and inhibition of MARylation were observed forts described here to link PARP14 with suppression of the anti-

in IFN-g-stimulated primary human macrophages, conﬁrming tumor immune response. It is our hope that RBN012759 will

that RBN012759 inhibits endogenous PARP14. Dose-depen- serve as a useful tool for further exploration of PARP14 functions

dent stabilization of PARP14 was also observed in mouse spleen in cell biology and as a drug target.

tissue. With a pharmacokinetic proﬁle enabling continuous

target coverage with oral, twice-daily dosing, RBN012759 can SIGNIFICANCE

be utilized to evaluate the effects of PARP14 in the in vivo setting.

RBN012759 provides the ability to evaluate the effects of PARP14 is an interferon-stimulated gene (ISG) that is overex-

PARP14 catalytic inhibition with a potent, selective chemical pressed in tumors compared with normal tissues and has

probe in both in vitro and in vivo settings.

been implicated by genetic knockoutstudies to promote pro-

Clinical and experimental studies have demonstrated that tumor macrophage polarization and suppress the antitumor

macrophages are amply present in the TME of most tumor types inﬂammatory response due to its role in modulating IL-4 and

and have a key regulatory role in tumor progression (Noy and IFN-g signaling pathways. Efforts to further explore and vali-

Pollard, 2014). Given the general abundance of macrophages in date the role of PARP14 have been hampered by a lack of

most cancers, repolarization of macrophages toward an inﬂam- potent, highly selective PARP14 inhibitors. Here, we describe

matory function may prove efﬁcacious for multiple oncology indi- the discovery of a chemical probe, RBN012759, whose proﬁle

cations. Our studies demonstrated that catalytic inhibition of represents a substantial improvement over previously re-

PARP14 with RBN012759 led to decreased IL-4-driven M2-like ported tool compounds. RBN012759 inhibits PARP14 with

gene expression in macrophages, suggesting that PARP14 an IC₅₀of 0.003 mM, exhibits >300-fold selectivity over all

8

Cell Chemical Biology 28, 1–11, August 19, 2021

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

ACKNOWLEDGMENTS

mono- and PARP-polyenzyme family members, and is a cell-

permeable, soluble probe that achieves continuous target

We thank Jennifer Downing for collecting analytical data for all compounds,

coverage in mice with oral dosing. RBN012759 inhibits MAR-

Andrew Santospago for compound handling and plating, Jeff Song for man-

ylation in primary human macrophages and demonstrates

agement of the tumor explant study, and Mark Boothby and James Audia

robust stabilization of endogenous PARP14 in cells and

for strategic input.

in vivo. Inhibition of PARP14 with RBN012759 reverses IL-4-

driven protumor gene expression in macrophages and in-

AUTHOR CONTRIBUTIONS

duces an inﬂammatory mRNA signature similar to immune

L.B.S., J.R.M., and K.K.S. prepared the manuscript with input from all authors.

checkpoint inhibitor therapy in primary human tumor ex-

L.B.S. designed the compounds, supervised the chemistry team, and guided

plants. These data support an immune suppressive role of

medicinal chemistry strategy. J.R.M. conceived, conducted, and oversaw

PARP14 in tumors and suggest potential utility of PARP14 in-

biology experiments and guided biology strategy. K.K.S. oversaw X-ray crys-

hibitors in the treatment of cancer.

tallography and guided strategy. R.A. conducted computational biology ana-

lyses. D.J.B. and Y.R. developed and performed the NanoBRET assays.

T.J.W. and W.D.C. developed and performed the PARP14 enzyme and ther-

STAR+METHODS

mal shift assays. T.J.W., W.D.C., and C.R.M. developed, performed, and over-

saw the PARP family selectivity assays. T.J.W. oversaw the SPR assay. K.K.

Detailed methods are provided in the online version of this paper

and include the following:

oversaw the in vivo target engagement study. K.G.K.-B. oversaw the KIRC

explant study. A.Z.L. conducted substrate identiﬁcation experiments with

guidance from M.N. A.E.C. conducted the macrophage gene expression

experiment. E.M. conducted the self-MARylation expression analysis. J.-

R.M. and C.R. conducted cell assays. M.M.V., V.M.R., H.K., and K.W.K.

guided strategy and supervised the work.

d KEY RESOURCES TABLE

d RESOURCE AVAILABILITY

B Lead contact

B Materials availability

DECLARATION OF INTERESTS

B Data code and availability

d EXPERIMENTAL MODEL

L.B.S., K.K.S., M.M.V., and K.W.K. are co-inventors on pending patent appli-

cation WO 2019/126443 and issued U.S. Pat. No. 10,562,891, both of which

relate to quinazolinone inhibitors of PARP14 for the treatment of cancer and

other diseases and are assigned to Ribon Therapeutics, Inc., of Cambridge,

MA, USA.

B Cell lines

B In vivo animal studies

d METHOD DETAILS

B Assay and chemistry reagents

B Chemistry

Received: June 26, 2020

B Protein puriﬁcation and crystallography

B Human PARP14 catalytic domain in vitro active site

probe displacement (TR-FRET) assay

B Human full-length PARP14 self-modiﬁcation enzy-

matic activity (DELFIA) assay

Revised: December 18, 2020

Accepted: February 11, 2021

Published: March 10, 2021

REFERENCES

B Mouse catalytic domain PARP14 in vitro active site

probe displacement (TR-FRET) assay

B PARP14 SPR assay

Aguiar, R.C., Takeyama, K., He, C., Kreinbrink, K., and Shipp, M.A. (2005). B-

Aggressive lymphoma family proteins have unique domains that modulate

transcription and exhibit poly(ADP-ribose) polymerase activity. J. Biol.

Chem. 280 , 33756–33765.

B Human PARP14 catalytic domain thermal shift

assay (TSA)

Aguiar, R.C.T., Yakushijin, Y., Kharbanda, S., Salgia, R., Fletcher, J.A., and

Shipp, M.A. (2000). BAL is a novel risk-related gene in diffuse large B-cell lym-

phomas that enhances cellular migration. Blood 96 , 4328–4334.

B PARP14 NanoBRET cellular biophysical assay

B Human primary macrophage culture

B Cytokine and chemokine measurements

B Immunoblotting

Andersson, C.D., Karlberg, T., Ekblad, T., Lindgren, A.E., Thorsell, A.G., Spjut,

S., Uciechowska, U., Niemiec, M.S., Wittung-Stafshede, P., Weigelt, J., et al.

(2012). Discovery of ligands for ADP-ribosyltransferases via docking-based

virtual screening. J. Med. Chem. 55 , 7706–7718.

B NanoString Assay and analysis

B RNA sequencing and analysis

B Substrate identiﬁcation

Caprara, G., Prosperini, E., Piccolo, V., Sigismondo, G., Melacarne, A.,

Cuomo, A., Boothby, M., Rescigno, M., Bonaldi, T., and Natoli, G. (2018).

PARP14 controls the nuclear accumulation of a subset of type I IFN-inducible

proteins. J. Immunol. 200 , 2439–2454.

B Human tumor explant culture and RNA extraction

B TCGA analysis

B In vivo pharmacokinetics and bioanalysis

B In vivo target engagement

Carter-O’connell, I., Vermehren-Schmaedick, A., Jin, H., Morgan, R.K., David,

L.L., and Cohen, M.S. (2018). Combining chemical genetics with proximity-

dependent labeling reveals cellular targets of poly(ADP-ribose) polymerase

14 (PARP14). ACS Chem. Biol. 13 , 2841–2848.

d QUANTIFICATION AND STATISICAL ANALYSIS

B % Inhibition

B IC₅₀curve ﬁtting

Cheng, H., Wang, Z., Fu, L., and Xu, T. (2019). Macrophage polarization in the

development and progression of ovarian cancers: an overview. Front. Oncol.

9 , 421.

B Pass/fail criteria for screening plates

SUPPLEMENTAL INFORMATION

Cho, S.H., Goenka, S., Henttinen, T., Gudapati, P., Reinikainen, A., Eischen,

C.M., Lahesmaa, R., and Boothby, M. (2009). PARP-14, a member of the B

aggressive lymphoma family, transduces survival signals in primary B cells.

Blood 113 , 2416–2425.

Supplemental information can be found online at https://doi.org/10.1016/j.

chembiol.2021.02.010.

Cell Chemical Biology 28, 1–11, August 19, 2021

9

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Cho, S.H., Raybuck, A., Wei, M., Erickson, J., Nam, K.T., Cox, R.G.,

Trochtenberg, A., Thomas, J.W., Williams, J., and Boothby, M. (2013). B

cell-intrinsic and -extrinsic regulation of antibody responses by PARP14, an

intracellular (ADP-ribosyl)transferase. J. Immunol. 191 , 3169–3178.

regulates T(H)2 differentiation and allergic airway disease. J. Allergy Clin.

Immunol. 131 , 521–531.e1-12.

Mehrotra, P., Riley, J.P., Patel, R., Li, F., Voss, L., and Goenka, S. (2011).

PARP-14 functions as a transcriptional switch for Stat6-dependent gene acti-

vation. J. Biol. Chem. 286 , 1767–1776.

Cohen, M.S., and Chang, P. (2018). Insights into the biogenesis, function, and

regulation of ADP-ribosylation. Nat. Chem. Biol. 14 , 236–243.

Motzer, R.J., Tannir, N.M., Mcdermott, D.F., Aren Frontera, O., Melichar, B.,

Choueiri, T.K., Plimack, E.R., Barthelemy, P., Porta, C., George, S., et al.

(2018). Nivolumab plus ipilimumab versus sunitinib in advanced renal-cell car-

cinoma. N. Engl. J. Med. 378 , 1277–1290.

Ekblad, T., Lindgren, A.E., Andersson, C.D., Caraballo, R., Thorsell, A.G.,

Karlberg, T., Spjut, S., Linusson, A., Schuler, H., and Elofsson, M. (2015).

Towards small molecule inhibitors of mono-ADP-ribosyltransferases. Eur. J.

Med. Chem. 95 , 546–551.

Moustakim, M., Riedel, K., Schuller, M., Gehring, A.P., Monteiro, O.P., Martin,

S.P., Fedorov, O., Heer, J., Dixon, D.J., Elkins, J.M., et al. (2018). Discovery of

a novel allosteric inhibitor scaffold for polyadenosine-diphosphate-ribose po-

lymerase 14 (PARP14) macrodomain 2. Bioorg. Med. Chem. 26 , 2965–2972.

Ekblad, T., Verheugd, P., Lindgren, A.E., Nyman, T., Elofsson, M., and Schuler,

H. (2018). Identiﬁcation of poly(ADP-ribose) polymerase macrodomain inhibi-

tors using an AlphaScreen protocol. SLAS Discov. 23 , 353–362.

Goenka, S., Cho, S.H., and Boothby, M. (2007). Collaborator of Stat6 (CoaSt6)-

associated poly(ADP-ribose) polymerase activity modulates Stat6-dependent

gene transcription. J. Biol. Chem. 282 , 18732–18739.

Noy, R., and Pollard, J.W. (2014). Tumor-associated macrophages: from

mechanisms to therapy. Immunity 41 , 49–61.

Page, M., Kreahling, J., and Altiok, S. (2018). Innovative combinatorial

approach to characterize the immune landscape and analyze the tumor

response after anti-PD-1 blockade in a 3D ex-vivo tumoroid system of non-

small cell lung cancer. J. Immunother. Cancer 6 , 114.

Gong, Y., Hart, E., Shchurin, A., and Hoover-Plow, J. (2008). Inﬂammatory

macrophage migration requires MMP-9 activation by plasminogen in mice.

J. Clin. Invest. 118 , 3012–3024.

Holechek, J., Lease, R., Thorsell, A.G., Karlberg, T., Mccadden, C., Grant, R.,

Keen, A., Callahan, E., Schuler, H., and Ferraris, D. (2018). Design, synthesis

and evaluation of potent and selective inhibitors of mono-(ADP-ribosyl)trans-

ferases PARP10 and PARP14. Bioorg. Med. Chem. Lett. 28 , 2050–2054.

Pathria, P., Louis, T.L., and Varner, J.A. (2019). Targeting tumor-associated

macrophages in cancer. Trends Immunol. 40 , 310–327.

Pechkovsky, D.V., Prasse, A., Kollert, F., Engel, K.M., Dentler, J., Luttmann,

W., Friedrich, K., Muller-Quernheim, J., and Zissel, G. (2010). Alternatively acti-

vated alveolar macrophages in pulmonary ﬁbrosis-mediator production and

intracellular signal transduction. Clin. Immunol. 137 , 89–101.

Huang, W.C., Sala-Newby, G.B., Susana, A., Johnson, J.L., and Newby, A.C.

(2012). Classical macrophage activation up-regulates several matrix metallo-

proteinases through mitogen activated protein kinases and nuclear factor-

kappaB. PLoS One 7 , e42507.

Peng, B., Thorsell, A.G., Karlberg, T., Schuler, H., and Yao, S.Q. (2017). Small

molecule microarray based discovery of PARP14 inhibitors. Angew. Chem. Int.

Ed. 56 , 248–253.

Iansante, V., Choy, P.M., Fung, S.W., Liu, Y., Chai, J.G., Dyson, J., Del Rio, A.,

D’santos, C., Williams, R., Chokshi, S., et al. (2015). PARP14 promotes the

Warburg effect in hepatocellular carcinoma by inhibiting JNK1-dependent

PKM2 phosphorylation and activation. Nat. Commun. 6 , 7882.

Rusinova, I., Forster, S., Yu, S., Kannan, A., Masse, M., Cumming, H.,

Chapman, R., and Hertzog, P.J. (2013). Interferome v2.0: an updated database

of annotated interferon-regulated genes. Nucleic Acids Res. 41 ,

D1040–D1046.

Iwata, H., Goettsch, C., Sharma, A., Ricchiuto, P., Goh, W.W., Halu, A.,

Yamada, I., Yoshida, H., Hara, T., Wei, M., et al. (2016). PARP9 and PARP14

cross-regulate macrophage activation via STAT1 ADP-ribosylation. Nat.

Commun. 7 , 12849.

Santoni, M., Massari, F., Di Nunno, V., Conti, A., Cimadamore, A., Scarpelli, M.,

Montironi, R., Cheng, L., Battelli, N., and Lopez-Beltran, A. (2018).

Immunotherapy in renal cell carcinoma: latest evidence and clinical implica-

tions. Drugs Context 7 , 212528.

Jung, S.B., Choi, M.J., Ryu, D., Yi, H.S., Lee, S.E., Chang, J.Y., Chung, H.K.,

Kim, Y.K., Kang, S.G., Lee, J.H., et al. (2018). Reduced oxidative capacity in

macrophages results in systemic insulin resistance. Nat. Commun. 9 , 1551.

Scaltriti, M., Verma, C., Guzman, M., Jimenez, J., Parra, J.L., Pedersen, K.,

Smith, D.J., Landolﬁ, S., Ramon Y Cajal, S., Arribas, J., and Baselga, J.

(2009). Lapatinib, a HER2 tyrosine kinase inhibitor, induces stabilization and

accumulation of HER2 and potentiates trastuzumab-dependent cell cytotox-

icity. Oncogene 28 , 803–814.

Juszczynski, P., Kutok, J.L., Li, C., Mitra, J., Aguiar, R.C., and Shipp, M.A.

(2006). BAL1 and BBAP are regulated by a gamma interferon-responsive bidi-

rectional promoter and are overexpressed in diffuse large B-cell lymphomas

with a prominent inﬂammatory inﬁltrate. Mol. Cell Biol. 26 , 5348–5359.

Schuller, M., Riedel, K., Gibbs-Seymour, I., Uth, K., Sieg, C., Gehring, A.P.,

Ahel, I., Bracher, F., Kessler, B.M., Elkins, J.M., and Knapp, S. (2017).

Discovery of a selective allosteric inhibitor targeting macrodomain 2 of polya-

denosine-diphosphate-ribose polymerase 14. ACS Chem. Biol. 12 ,

2866–2874.

Kirby, I.T., Kojic, A., Arnold, M.R., Thorsell, A.G., Karlberg, T., Vermehren-

Schmaedick, A., Sreenivasan, R., Schultz, C., Schuler, H., and Cohen, M.S.

(2018). A potent and selective PARP11 inhibitor suggests coupling between

cellular localization and catalytic activity. Cell Chem. Biol. 25 , 1547–1553 e12.

Li, B., Severson, E., Pignon, J.C., Zhao, H., Li, T., Novak, J., Jiang, P., Shen, H.,

Aster, J.C., Rodig, S., et al. (2016). Comprehensive analyses of tumor immu-

nity: implications for cancer immunotherapy. Genome Biol. 17 , 174.

Schulz, D., Severin, Y., Zanotelli, V.R.T., and Bodenmiller, B. (2019). In-depth

characterization of monocyte-derived macrophages using a mass cytometry-

based phagocytosis assay. Sci. Rep. 9 , 1925.

Liu, C., Wang, J., and Zhang, X. (2014). The involvement of MiR-1-clathrin

Schweiker, S.S., Tauber, A.L., Sherry, M.E., and Levonis, S.M. (2018).

Structure, function and inhibition of poly(ADP-ribose)polymerase, member

14 (PARP14). Mini Rev. Med. Chem. 18 , 1659–1669.

pathway in the regulation of phagocytosis. PLoS One 9 , e98747.

Lu, A.Z., Abo, R., Ren, Y., Gui, B., Mo, J.R., Blackwell, D., Wigle, T., Keilhack,

H., and Niepel, M. (2019). Enabling drug discovery for the PARP protein family

through the detection of mono-ADP-ribosylation. Biochem. Pharmacol.

167 , 97–106.

Shapouri-Moghaddam, A., Mohammadian, S., Vazini, H., Taghadosi, M.,

Esmaeili, S.A., Mardani, F., Seiﬁ, B., Mohammadi, A., Afshari, J.T., and

Sahebkar, A. (2018). Macrophage plasticity, polarization, and function in health

and disease. J. Cell Physiol. 233 , 6425–6440.

Makita, N., Hizukuri, Y., Yamashiro, K., Murakawa, M., and Hayashi, Y. (2015).

IL-10 enhances the phenotype of M2 macrophages induced by IL-4 and con-

fers the ability to increase eosinophil migration. Int. Immunol. 27 , 131–141.

Steenport, M., Khan, K.M., Du, B., Barnhard, S.E., Dannenberg, A.J., and

Falcone, D.J. (2009). Matrix metalloproteinase (MMP)-1 and MMP-3 induce

macrophage MMP-9: evidence for the role of TNF-alpha and cyclooxyge-

nase-2. J. Immunol. 183 , 8119–8127.

Mediavilla-Varela, M., Page, M.M., Kreahling, J., and Altiok, S. (2018).

Characterization of the immune landscape and analysis of tumor response af-

ter anti-PD-1 blockade in a 3D ex vivo system of non-small lung cancer.

Cancer Res. 78.

Thorsell, A.G., Ekblad, T., Karlberg, T., Low, M., Pinto, A.F., Tresaugues, L.,

Moche, M., Cohen, M.S., and Schuler, H. (2017). Structural basis for potency

and promiscuity in poly(ADP-ribose) polymerase (PARP) and tankyrase inhib-

itors. J. Med. Chem. 60 , 1262–1271.

Mehrotra, P., Hollenbeck, A., Riley, J.P., Li, F., Patel, R.J., Akhtar, N., and

Goenka, S. (2013). Poly(ADP-ribose) polymerase 14 and its enzyme activity

10 Cell Chemical Biology 28, 1–11, August 19, 2021

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Upton, K., Meyers, M., Thorsell, A.G., Karlberg, T., Holechek, J., Lease, R.,

Schey, G., Wolf, E., Lucente, A., Schuler, H., and Ferraris, D. (2017). Design

and synthesis of potent inhibitors of the mono(ADP-ribosyl)transferase,

PARP14. Bioorg. Med. Chem. Lett. 27 , 2907–2911.

Wigle, T., Blackwell, D.J., Schenkel, L.B., Ren, Y., Church, W.D., Desai, H.J.,

Swinger, K.K., Santospago, A.G., Majer, C.R., Lu, A.Z., et al. (2020). In vitro

and cellular probes to study PARP enzyme target engagement. Cell Chem.

Biol. 27 , 877–887.e14.

Viola, A., Munari, F., Sanchez-Rodriguez, R., Scolaro, T., and Castegna, A.

(2019). The metabolic signature of macrophage responses. Front. Immunol.

10 , 1462.

Wigle, T.J., Church, W.D., Majer, C.R., Swinger, K.K., Aybar, D., Schenkel,

L.B., Vasbinder, M.M., Brendes, A., Beck, C., Prahm, M., et al. (2019).

Forced self-modiﬁcation assays as a strategy to screen MonoPARP enzymes.

SLAS Discov. 25 , 241–252.

Vyas, S., and Chang, P. (2014). New PARP targets for cancer therapy. Nat.

Rev. Cancer 14 , 502–509.

Witten, D.M. (2011). Classiﬁcation and clustering of sequencing data using a

Vyas, S., Matic, I., Uchima, L., Rood, J., Zaja, R., Hay, R.T., Ahel, I., and Chang,

P. (2014). Family-wide analysis of poly(ADP-ribose) polymerase activity. Nat.

Commun. 5 , 4426.

Poisson model. Ann. Appl. Stat. 5 , 2493–2518.

Yoneyama-Hirozane, M., Matsumoto, S.I., Toyoda, Y., Saikatendu, K.S.,

Zama, Y., Yonemori, K., Oonishi, M., Ishii, T., and Kawamoto, T. (2017).

Identiﬁcation of PARP14 inhibitors using novel methods for detecting auto-ri-

bosylation. Biochem. Biophys. Res. Commun. 486 , 626–631.

Wahlberg, E., Karlberg, T., Kouznetsova, E., Markova, N., Macchiarulo, A.,

Thorsell, A.G., Pol, E., Frostell, A., Ekblad, T., Oncu, D., et al. (2012). Family-

wide chemical proﬁling and structural analysis of PARP and tankyrase inhibi-

tors. Nat. Biotechnol. 30 , 283–288.

Yu, G., Wang, L.G., Han, Y., and He, Q.Y. (2012). clusterProﬁler: an R package

for comparing biological themes among gene clusters. OMICS 16 , 284–287.

Wang, P., Li, J., Jiang, X., Liu, Z., Ye, N., Xu, Y., Yang, G., Xu, Y., and Zhang, A.

(2014). Palladium-catalyzed N-arylation of 2-aminobenzothiazole-4-carboxyl-

ates/carboxamides: facile synthesis of PARP14 inhibitors. Tetrahedron 70 ,

5666–5673.

Zhang, J.H., Chung, T.D., and Oldenburg, K.R. (1999). A simple statistical

parameter for use in evaluation and validation of high throughput screening as-

says. J. Biomol. Screen 4 , 67–73.

Cell Chemical Biology 28, 1–11, August 19, 2021 11

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

STAR+METHODS

Detailed methods are provided in the online version of this paper and include the following:

KEY RESOURCES TABLE

REAGENT or RESOURCE

Antibodies

SOURCE

IDENTIFIER

PARP14 antibody clone 15A6-1 (for Western

blotting)

Genscript

U9897EA250-9

U0106DJ020-2

PARP14 antibody clone 15B10-1 (for

immunoprecipitation)

Beta-Actin

Phospho-STAT6

STAT6

Cell Signaling Technologies

Cat#8457; RRID: AB_10950489

Cat#9361; RRID: AB_331595

Cat#9362; RRID: AB_2271211

Biological Samples

Leukopak

STEMCELL Technologies

Peprotech

Cat#70500.2

Cat#78057

Cat#31505

Cat#78020

Cat#78045

Cat#78029

Human recombinant M-CSF

Mouse recombinant IFN-gamma

Human recombinant IFN-gamma

Human recombinant IL-4

Human recombinant IL-13

STEMCELL Technologies

Chemicals, Peptides, and Recombinant Proteins

PARP1 (human, full-length)

PARP2 (human, full-length)

PARP3 (human, full-length)

PARP4 (human, truncate)

PARP5a (human, truncate)

PARP6 (human, truncate)

PARP7 (human, truncate)

PARP8 (human, truncate)

PARP9 (human, full-length)

PARP10 (human, truncate)

PARP11 (human, full-length)

PARP12 (human, truncate)

PARP13 (human, full-length)

PARP14 (human, truncate)

PARP14 (human, full-length)

PARP14 (mouse, truncate)

PARP14 (mouse, full-length)

PARP15 (human, truncate)

PARP16 (human, full-length)

RBN010860

Wigle et al., SLAS Discovery (2019)

This paper

N/A

This paper

N/A

This paper

N/A

Wigle et al., SLAS Discovery (2019)

Wigle et al., Cell Chem Biol (2020)

This paper

N/A

RBN011198

N/A

RBN012759 (Compound 1)

Compound 2

N/A

This paper

N/A

Compound 3

This paper

N/A

Compound 4

Biotin-NAD⁺

This paper

N/A

Biolog

Cat#N012

Cat#D3072

Cat#J60712

(Continued on next page)

Doxycycline

Millipore Sigma

HEPES pH = 7.5

Alfa Aesar

e1 Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Continued

REAGENT or RESOURCE

NaCl

SOURCE

IDENTIFIER

Quality Biological

Thermo

Cat#351-036-721

Cat#28320

Tween 20

DPTA-puriﬁed BSA

Dithiothreitol

Perkin Elmer

Fisher Scientiﬁc

Hoefer

Cat#CR84-100

Cat#BP172-25

Cat#GR154-1

Cat#IBB-645

Cat#A412-1

TBS-T

PBS-T

Boston Bioproducts

Fisher Scientiﬁc

Millipore Sigma

Thermo

Methanol

DMSO

Cat#D8418

MEM + Glutamax media

OptiMEM media

DMEM media

Fetal bovine serum

FuGENE(R) HD Transfection Reagent

Methylcellulose

Tween 80

Cat#41090036

Cat#11058021

Cat#10569010

Cat#97068-085

Cat#E2312

Thermo

VWR

Promega

Sigma-Aldrich

Shanghai Titan Scientiﬁc Co. Ltd

Cat#M7027-250G

Cat#G8919113

Critical Commercial Assays

Europium-labeled streptavidin

ULight-labeled streptavidin

Europium-labeled anti-His

Ulight-labeled anti-His

384-well nickel-NTA coated microplates

DELFIA Eu-N1 Streptavidin

DELFIA Assay Buffer

Perkin Elmer

Thermo

Cat#AD0063

Cat#AD0062

Cat#AD0111

Cat#TRF0105

Custom

Perkin Elmer

Promega

Cat#1244-360

Cat#1244-111

Cat#1244-105

Cat#N2161

DELFIA Enhancement Solution

IntracellularTE Nano-Glo(R) Substrate/Inhibitor

IL-10 quantikine ELISA kit

Luminex Cytokine Human 25-Plex Panel

RIPA lysis buffer

R&D systems

Invitrogen

Millipore

Cat#D1000B

Cat#LHC0009

Cat#20-188

SDS

Amresco

Cat#E719

HALT protease/phosphatase inhibitor

Sample buffer

ThermoFisher

LICOR

Cat#23225

Cat#92840004

Cat#IB24001

Cat#NP0001

Cat#3450124

Cat#92632213

Cat#AM9738

Cat#PanCancer IO 360ꢀ Panel

Cat#79565

PVDF membranes

ThermoFisher

BIORAD

MOPS running buffer

4-12% Bis-Tris gels

Blocking buffer

LICOR

TRI reagent

ThermoFisher

nanoString

QIAGEN

PanCancer IO 360ꢀ Panel

QIAshredder columns

Mag MAX kit

Ambio

Cat#1830M

TRIxol reagent

Invitrogen

Cat#15596018

Deposited Data

Structure of PARP14 bound to Compound 2

Structure of PARP14 bound to Compound 3

Structure of PARP14 bound to RBN012759

This paper

PDB: 6WE4

PDB: 6WE3

PDB: 6WE2

Experimental Models: Cell Lines

293T cells

ATCC

Cat#CRL-3216

Cat#TIB-71ꢀ

Cat# CRL-1918

RAW264.7 cells

CFPAC-1 cells

(Continued on next page)

Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021 e2

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Continued

REAGENT or RESOURCE

SOURCE

IDENTIFIER

N/A

Experimental Models: Organisms/Strains

Mus Musculus / C57BL/6

Oligonucleotides

Shanghai Lingchang BioTech

Activating dsDNA for PARP1 (5’ ACCCTGCTGTGG

GC/ideoxyU/GGAGAACAAGGTGAT and 3’

ATCACCTTGTTCTCCAHGCCCACAGCAGGGT)

IDT DNA

N/A

Activating dsDNA for PARP2 (5’/phosphate/

IDT DNA

N/A

GCCTATAGGC and 3’/phosphate/GCCTATACCG)

Activating ssDNA for PARP3 (/phosphate/GCTG

GCTTCGTAAGAAGCCAGCTCGCGGTCAGC

TTGCTGACCGCG)

Recombinant DNA

pcDNA3.1 plasmid

pcDNA3.1 with human PARP1 amino acids 1-1014

with C-terminal or N-terminal NanoLuc tag

Wigle et al., Cell Chem Biol (2020)

N/A

pcDNA3.1 with human PARP3 amino acids 1 – 533

with C-terminal or N-terminal NanoLuc tag

pcDNA3.1 with human PARP7 amino acids 1-657

insert

with C-terminal or N-terminal NanoLuc tag

pcDNA3.1 with human PARP7 amino acids

456-657 insert with C-terminal or N-terminal

NanoLuc tag

Wigle et al., Cell Chem Biol (2020)

N/A

pcDNA3.1 with human PARP10 amino acids

808-1025

with C-terminal or N-terminal NanoLuc tag

pcDNA3.1 with human PARP10 amino acids 1-

1025

with C-terminal or N-terminal NanoLuc tag

pcDNA3.1 with human PARP11 amino acids 1-338

with C-terminal or N-terminal NanoLuc tag

Wigle et al., Cell Chem Biol (2020)

N/A

pcDNA3.1 with human PARP12 amino acids 1-701

with C-terminal or N-terminal NanoLuc tag

pcDNA3.1 with human PARP14 amino acids

1611-1801

with C-terminal or N-terminal NanoLuc tag

pcDNA3.1 with human PARP14 amino acids

1-1801

Wigle et al., Cell Chem Biol (2020)

N/A

with C-terminal or N-terminal NanoLuc tag

pcDNA3.1 with human PARP16 amino acids 1-322

with C-terminal or N-terminal NanoLuc tag

Software and Algorithms

Odyssey Image Studio

R statistical software

DESeq2

Licor https://www.licor.com/bio/image-studio/

r-project.org

N/A

Version 3.6.1

Version 1.18.17

Love et al., Genome Biol (2014) https://

bioconductor.org/packages/release/

bioc/html/DESeq2.html

Salmon

Patro et al., Nat Meth (2017) https://

Version 0.11.26

Version 2.2.0

combine-lab.github.io/salmon/

ComplexHeatmap

Gu et al., Bioinformatics (2016) http://

bioconductor.org/packages/release/

bioc/html/ComplexHeatmap.html

(Continued on next page)

e3 Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Continued

REAGENT or RESOURCE

Other

SOURCE

IDENTIFIER

Envision platereader

Perkin Elmer

GE Healthcare

Licor

Cat#2101-0010

Cat#28975001

Cat#29104992

Cat#28994951

Cat. # 9140

Biacore T200 SPR instrument

Streptavidin SPR chips

NTA SPR chips

Odyssey CLX infrared imager

BioRad BioPlex MAGPIX multilex reader system

BioRad TransBLot Turbo Transfer system

BioRad

Cat#171000201

Cat#1704150

BioRad

RESOURCE AVAILABILITY

Lead contact

Further information and requests for reagents and resources should be directed to the lead contact, Kevin W. Kuntz (kkuntz@

ribontx.com).

Materials availability

Reagents generated in this study are available from the Lead Contact with a completed Materials Transfer Agreement.

Data code and availability

The coordinates for the protein structures reported in this manuscript have been deposited in Protein Data Bank (http://rcsb.org/pdb)

under ID codes 6WE4, 6WE3, and 6WE2.

EXPERIMENTAL MODEL

Cell lines

293T cells (ATCC; sex: female) expressing PARP-NanoLuc fusion proteins, RAW 267.4 cells (ATCC; sex: male), and CFPAC-1 cells

(ATCC; sex: male) were grown in Dulbecco’s modiﬁed Eagle’s medium (DMEM) supplemented with 10% heat inactivated fetal bovine

serum (VWR) in a 5% CO₂environment at 37^ꢀC.

In vivo animal studies

All procedures related to animal handling, care and treatment were performed according to the guidelines approved by the Institu-

tional Animal Care and Use Committee (IACUC) following the guidance of the Association for Assessment and Accreditation of Lab-

oratory Animal Care (AAALAC). For the pharmacokinetics studies, male CD1 mice were obtained from the Animal Facility of Beijing

Vital River Laboratory Animal Technology Co., Ltd. For the in vivo target engagement studies, C57Bl/6 J female mice were obtained

from Shanghai Lingchang BioTech Co., Ltd. All mice were kept in individual ventilation cages under a 12 hour light/dark cycle with

free access to food and water at constant temperature (20-25^ꢀC) and humidity (40-70%), and with 3 or 5 animals in each cage. Mice

were between 6 to 8 weeks of age, weighed 20-25 g (male CD1 mice) and 18-20 g (female C57Bl/6 mice), and were randomly as-

signed to experimental groups.

METHOD DETAILS

Assay and chemistry reagents

TR-FRET reagents used in the in vitro active site probe displacement TR-FRET assays, DELFIA reagents used in the self-modiﬁcation

enzyme assays and 384-well proxiplates were purchased from Perkin Elmer. NaCl, HEPES, 384-well PCR microplates and thermal

shift dye were purchased from VWR. Tween-20, dithiothreitol (DTT) and 384-well nickel-NTA coated microplates were purchased

from Thermo Fisher Scientiﬁc. Tris-buffered saline was purchased from Hoefer. NAD⁺was purchased from Millipore-Sigma.

Biotin-NAD⁺was purchased from BIOLOG Life Sciences Institute. All reactions were carried out in commercially available glassware

using standard synthetic chemistry methods, and reagents were purchased commercially and used without further puriﬁcation.

RBN010860 and RBN011198 were prepared as previously described(Wigle et al., 2020).

Chemistry

List of abbreviations:

ACN (acetonitrile); DCM (dichloromethane); DMSO (dimethylsulfoxide); EtOAc (ethyl acetate); g (gram); h (hours); HRMS (high reso-

lution mass spectrometry); LCMS (liquid chromatography-mass spectrometry); MeOH (methanol); MHz (megahertz); mL (milliliter);

mmol (millimole); NMP (N-methyl-2-pyrrolidinone); NMR (nuclear magnetic resonance); RT (room temperature); THF (tetrahydro-

furan); TFA (triﬂuoroacetic acid); TLC (thin layer chromatography)

Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021 e4

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Synthesis of 8-methyl-2-((pyridin-4-ylthio)methyl)quinazolin-4(3H)-one (2):

A mixture of 2-(chloromethyl)-8-methylquinazolin-4(3H)-one (300 mg, 1.44 mmol, 1.0 equiv), pyridine-4-thiol (190 mg, 1.73 mmol,

1.2 equiv), NaI (216 mg, 1.44 mmol, 1.0 equiv), and Cs₂CO₃(1.41 g, 4.32 mmol, 3.0 equiv) in 25 mL acetone was heated at reﬂux for 2 h

under an atmosphere of nitrogen. The mixture was diluted with 100 mL water and extracted with EtOAc (2 x 100 mL). The combined

organic portion was washed with 100 mL brine, dried over Na₂SO₄, and concentrated under reduced pressure. The crude material

was puriﬁed by preparative TLC using 20:1 DCM:MeOH to give the title compound as a yellow solid (200 mg, 49%). LCMS: [M+H]⁺

284.1; ¹H NMR (400 MHz, DMSO-d₆) d 12.43 (s, 1H), 8.40 (d, J = 6.0, 2H), 7.92 (d, J = 7.2, 1 H), 7.64 (d, J = 5.2, 1 H), 7.55-7.54 (m, 2H),

7.39-7.35 (m, 1H), 4.30 (s, 2H), 2.48 (s, 3H); ¹³C NMR (100.5 MHz, DMSO-d₆) d 162.26, 152.44, 149.51, 147.77, 147.10, 135.58,

135.32, 126.64, 123.86, 121.48, 121.36, 33.94, 17.45; HRMS: [M+H]⁺Calcd for C₁₅H₁₄N₃OS 284.0858; found 284.0858.

C

LCMS: [M+H]⁺305.1;

¹H NMR (400 MHz, DMSO-d₆) d 12.27 (s, 1H), 7.95 – 7.93 (m, 1H), 7.67 (d, J = 6.8 Hz, 1H), 7.38 (t, J = 7.6 Hz, 1H), 4.55 (d, J = 4.4 Hz,

1H), 3.64 (s, 2H), 3.39 - 3.34 (m, 1H), 2.82 - 2.75 (m, 1H), 2.51 (s, 3H), 2.02 (d, J = 12.3 Hz, 2H), 1.83 (d, J = 12.0 Hz, 2H), 1.28 – 1.19

(m, 4H).

¹³C NMR (400 MHz, DMSO-d₆) d 162.543 (1C), 154.250 (1C), 147.471 (1C), 135.433 (1C), 135.194 (1C), 126.242 (1C), 123.872 (1C),

121.252 (1C), 68.455 (1C), 42.294 (1C), 35.463 (1C), 33.137 (1C), 31.216 (2C), 17.549 (2C). HRMS: [M+H]⁺Calcd for C₁₆H₂₀N₂O₂S

305.1324; found 305.1312.

Synthesis of 7-(cyclopropylmethoxy)-5-ﬂuoro-2-((((1r,4r)-4-hydroxycyclohexyl)thio)methyl)quinazolin-4(3H)-one

(RBN012759):

trans-4-Mercaptocyclohexanol.

To a solution 7-oxabicyclo[2.2.1]heptane (1 g, 10.2 mmol, 1.0 eq) in ethanol (10 mL) was added p-TSA (2.91g, 15.3mmol) and thio-

urea (1.2 g, 15.8 mmol, 1.5 eq) and the mixture was heated at reﬂux for 21 h. After cooling to RT, NaOH (1.3 g) and water (3 mL) were

added and the solution was heated at reﬂux for a further 2 h. The mixture was cooled to RT, NaOH (1.3 g) and water (3 mL) were added

and the solution was heated at reﬂux for a further 2 h, then allowed to cool to RT and concentrated under reduced pressure. The

residue was diluted with water (15 mL) and adjusted to pH 3-4 with 1 M HCl and extracted with EtOAc (50 mL x 3). The combined

organic layers were washed with brine (50 mL), dried over Na₂SO₄and concentrated under reduced pressure. The residue was pu-

riﬁed by column chromatography (Petroleum ether:EtOAc, 10:1, v/v) to afford the title compound (500 mg, 37%) as a yellow oil.

¹HNMR (400 MHz, DMSO-d₆) d 6.18 (br s, 1H), 4.51 (br s, 1H), 3.41 - 3.36 (m, 1H), 2.73 - 2.64 (m, 1H), 1.96 - 1.86 (m, 2H), 1.72 -

1.81 (m, 2H), 1.36 - 1.26 (m, 2H), 1.23 - 1.17 (m, 2H).

e5 Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Methyl 4-(cyclopropylmethoxy)-2,6-diﬂuorobenzoate.

A mixture of methyl 2,6-diﬂuoro-4-hydroxybenzoate, (bromomethyl)cyclopropane (102 mL, 1.05 mol) and K₂CO₃(330 g, 2.39 mol)

in DMSO (1 L) was heated at 80^ꢀC overnight. The mixture was diluted with water (5 L) and extracted with EtOAc (1 L x 3). The com-

bined organic extracts were washed with water (800 mL), brine (800 mL), dried over Na₂SO₄and concentrated under reduced pres-

sure to afford the title compound (214 g, 92%) as a brown oil. LCMS: [M+H]⁺243.1.

Methyl 4-(cyclopropylmethoxy)-2-((2,4-dimethoxybenzyl)amino)-6-ﬂuorobenzoate.

A mixture of methyl 4-(cyclopropylmethoxy)-2,6-diﬂuorobenzoate (214 g, 881 mmol), (2,4-dimethoxyphenyl)methanamine

(139 mL, 926 mmol) and K₂CO₃(243 g, 1.76 mol) in NMP (1 L) was heated at 80^ꢀC overnight. The mixture was poured into water

(5 L), and the resulting precipitate was collected by ﬁltration and washed with water (800 mL). The ﬁlter cake was dissolved in

DCM (2.5 L) and washed with brine (800 mL). The organic layer was dried over Na₂SO₄and concentrated under reduced pressure

to give the title compound (343 g, 99%) as an off-white solid. LCMS: [M+Na]⁺412.1.

Methyl 2-amino-4-(cyclopropylmethoxy)-6-ﬂuorobenzoate.

To a solution of methyl 4-(cyclopropylmethoxy)-2-((2,4-dimethoxybenzyl)amino)-6-ﬂuorobenzoate (1.6 g, 4.11 mmol, 1.0 eq) in

DCM (8.0 mL) was added TFA (4.0 mL) and the mixture was stirred at RT for 2 h. The mixture was concentrated under reduced pres-

sure and the residue was puriﬁed by column chromatography (DCM:MeOH, 20/1, v/v) to afford the title compound (0.9 g, 91%) as a

brown solid. LCMS: [M+H]⁺240.1.

2-(Chloromethyl)-7-(cyclopropylmethoxy)-5-ﬂuoroquinazolin-4(3H)-one

To a solution of methyl 2-amino-4-(cyclopropylmethoxy)-6-ﬂuorobenzoate (900 mg, 3.76 mmol, 1 equiv) in 4 M HCl/dioxane (8 mL,

32 mmol, 8.51 equiv) was added 2-chloroacetonitrile (1.19 mL, 18.8 mmol, 5 equiv). The mixture was heated at 80^ꢀC for 3 h. The

mixture was concentrated and puriﬁed by silica gel column (DCM/MeOH = 40/1,v/v) to afford the title compound (600 mg,

2.12 mmol, 56% yield) as a white solid. LCMS: [M+H]⁺283.1.

7-(Cyclopropylmethoxy)-5-ﬂuoro-2-(((trans-4-hydroxycyclohexyl)thio)methyl)quinazolin-4(3H)-one.

Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021 e6

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

To a solution of 2-(chloromethyl)-7-(cyclopropylmethoxy)-5-ﬂuoroquinazolin-4(3H)-one (300 mg, 1.06 mmol, 1.0 eq) in THF (5 mL)

under a N₂atmosphere was added trans-4-mercaptocyclohexanol (168 mg, 1.27 mmol, 1.2 eq) and 2 M NaOH (2 mL) and the mixture

was stirred at RT overnight. The mixture was poured into water (30 mL) and extracted with EtOAc (20 mL x 3). The combined organic

layers were washed with brine (30 mL), dried over Na₂SO₄and concentrated under reduced pressure. The residue was puriﬁed by

C18 reverse phase column (Biotage, 40% ACN in water) to afford the title compound (130 mg, 32%) as a white solid. LCMS: [M+H]⁺

379.1; ¹H NMR (400 MHz, DMSO-d₆) d 12.1 (s, 1H), 6.89-6.86 (m, 2H), 4.52 (d, J = 4.4 Hz, 1H), 3.96 (d, J = 7.2 Hz, 2H), 3.57 (s, 2H),

3.40-3.38 (m, 1H), 2.74-2.67 (m, 1H), 1.97-1.94 (m, 2H), 1.82-1.80 (m, 2H), 1.28-1.11 (m, 5H), 0.60-0.58 (m, 2H), 0.37-0.33 (m, 2H);

¹³C NMR (400 MHz, DMSO-d₆) d 163.841 (d, J = 14.1 Hz, 1C), 162.076 (d, J = 259.5 Hz, 1C), 159.080 (d, J = 2.9 Hz, 1C), 157.199

(1C), 152.549 (1C), 106.176 (d, J = 2.4 Hz, 1C), 104.392 (d, J =7.1 Hz, 1C), 102.798 (d, J = 23.4 Hz, 1C), 73.580 (1C), 68.320 (1C),

42.737 (1C), 35.276 (1C), 33.211 (2C), 31.401 (2C), 10.276 (1C), 3.622 (2C).

HRMS: [M+H]⁺Calcd for C₁₉H₂₃FN₂O₃S 379.1492; found 379.1493.

Synthesis of 7-(Cyclopentylamino)-2-(((trans-4-hydroxycyclohexyl)thio)methyl) quinazolin-4(3H)-one (4) was carried

out following the procedures detailed in PCT Int. Appl. (2019) WO 2019126443 A1 20190627

LCMS: [M+H]⁺374.2; ¹H NMR (400 MHz, DMSO-d₆) d 11.6 (br s, 1H), 7.70 (d, J = 8.0 Hz, 1H), 6.69 (dd, J = 8.0, 2.0 Hz, 1H), 6.50 (m,

1H), 6.45 (d, J = 2.0 Hz, 1H), 4.52 (br s, 1H), 3.74 - 3.82 (m, 1H), 3.52 (s, 2H), 3.39 - 3.35 (m, 1H), 2.76 - 2.64 (m, 1H), 2.03 – 1.90 (m, 4H),

1.82 - 1.76 (m, 2H), 1.73 - 1.63 (m, 2H), 1.62 - 1.53 (m, 2H), 1.52 – 1.42 (m, 2H), 1.24 - 1.13 (m, 4H); 13C NMR (300 MHz, DMSO-d₆)

d 162.214 (d, J = 128 Hz, 1C), 159.076 (d, J = 1.65 Hz, 1C), 156.248 (1C), 153.866 (d, J = 6.90 Hz , 1C), 152.443 (1C), 102.306 (1C),

99.510 (1C), 99.253 (d, J =3.98 Hz, 1C), 68.344 (1C), 53.801 (1C), 42.735 (1C), 35.342 (2C), 33.215 (1C), 32.716 (2C), 31.432 (2C),

24.105 (2C). HRMS: [M+H]⁺Calcd for C₂₀H₂₆FN₃O₂S 392.1808; found 392.1802.

Protein puriﬁcation and crystallography

Recombinant enzymes were expressed and puriﬁed as described previously. PARP14 (NM_017554, human and NM_001039530.3,

mouse) recombinant proteins were puriﬁed by the same method using an N-terminal hexahistidine (His6) tag(Wigle et al., 2019).

PARP1 (NP_001609.2), amino acids 1 – 1014, was cloned into a pFastBac1 insect expression vector with N-terminal FLAG and

hexahistidine tags, with a tobacco etch virus (TEV) cleavage site between the FLAG tag and the His tag. PARP1 was expressed in

Sf9 insect cells and harvested at 48 h. The cells were resuspended in lysis buffer (20 mmol/L 2-amino-2-(hydroxymethyl)propane-

1,3-diol (Tris)-HCl (pH = 7.5), 200 mmol/L sodium chloride (NaCl), 5% glycerol, 5 mmol/L beta-mercaptoethanol) and lysed by

sonication. The protein was puriﬁed by nickel-nitrilotriacetic acid (Ni-NTA) afﬁnity gel chromatography and anti-FLAG afﬁnity gel

chromatography and stored in 20 mmol/L Tris-HCl (pH = 7.5), 200 mmol/L NaCl, 5 mmol/L beta-mercaptoethanol and 5% glycerol.

PARP2 (NP_005475.2), amino acids 1 – 583, was cloned into a pFastBac1 insect expression vector with N-terminal FLAG and hex-

ahistidine tags, and a TEV cleavage site between the FLAG tag and the His tag. PARP2 was expressed in Sf9 insect cells and har-

vested at 48 h. The cells were resuspended in lysis buffer (25 mmol/L 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES)

(pH = 7.5), 500 mmol/L NaCl, 10% glycerol, 1 mmol/L Tris(2-carboxyethyl)phosphine (TCEP), 10 mmol/L magnesium chloride

(MgCl2) and 1 unit/mL DNase) and lysed by sonication. The protein was puriﬁed by anti-FLAG afﬁnity gel chromatography and stored

in 25 mmol/L HEPES (pH = 7.5), 300 mmol/L NaCl, 1 mmol/L TCEP, 200 mg/mL FLAG peptide and 10% glycerol.

PARP3 (NP_001003931.3) amino acids 1 – 533 was cloned into pTriIJ-HV expression vector with an N-terminal hexahistidine tag

followed by a TEV cleavage site, streptavidin binding protein (SBP) sequence and PreScission protease (PRX) cleavage site. The pro-

tein was expressed in Sf21 insect cells and harvested at 72 h. Cells were resuspended in lysis buffer (25 mmol/L Tris-HCl (pH = 8),

300 mmol/L NaCl, 5% glycerol, 1 mmol/L TCEP, 1X Roche protease inhibitor cocktail, 1 mmol/L phenylmethylsulfonyl ﬂuoride

(PMSF), 10 units/mL benzonase, and 0.25% 3-[(3-Cholamidopropyl) dimethylammonio]-1-propanesulfonate (CHAPS)) and lysed

by sonication. The protein was puriﬁed by Ni-NTA afﬁnity chromatography, followed by S-200 size exclusion chromatography. Pu-

riﬁed protein was stored in 25 mmol/L Tris-HCl (pH = 7.5), 300 mmol/L NaCl, 1 mmol/L TCEP and 5% glycerol.

PARP4 (NP_006428.2) amino acids 226 – 566 was cloned into pTriIJ-HV expression vector with an N-terminal hexahistidine tag,

and a TEV cleavage site and glycine residue between the hexahistidine tag and the gene. The protein was expressed in Sf21 insect

cells and harvested at 72 h. Cells were resuspended in lysis buffer (25 mmol/L Tris-HCl (pH = 7.5), 300 mmol/L NaCl, 5% glycerol,

1 mmol/L TCEP, 1X Roche protease inhibitor cocktail, 1 mmol/L PMSF, 10 units/mL and 0.25% CHAPS) and lysed by sonication. The

e7 Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

protein was puriﬁed by Ni-NTA afﬁnity chromatography, followed by S-200 size exclusion chromatography. Puriﬁed protein was

stored in 25 mmol/L Tris-HCl (pH = 7.5), 300 mmol/L NaCl, 1 mmol/L TCEP and 5% glycerol.

PARP5b (NM_025235) amino acid 667 – 1166 was purchased from BPS Biosciences (catalog #80505). The protein construct con-

tained an N-terminal glutathione S-transferase (GST) tag and was expressed and puriﬁed from Sf9 insect cells, then was stored in

40 mmol/L Tris-HCl (pH 8), 210 mmol/L NaCl, 2.2 mmol/L potassium chloride (KCl), 0.04% Tween20, 3 mmol/L dithiothreitol

(DTT) and 20% glycerol.

PARP6 (NP_001310451.1) amino acids 321 – 630 was cloned into pTriJ-HV expression vector with an N-terminal hexahistidine tag.

The protein was expressed in E. coli BL21-Gold(DE3) cells by addition of 0.5 mmol/L isopropyl b-D-1-thiogalactopyranoside (IPTG)

for 16 h at 16^ꢀC. Cells were resuspended in lysis buffer (20 mmol/L Tris-HCl (pH = 7.5), 300 mmol/L NaCl, 5% glycerol, 1 mmol/L

TCEP and 1X Roche protease inhibitor cocktail) and lysed by sonication. The protein was puriﬁed by Ni-NTA afﬁnity chromatography

followed by cation exchange and S-200 size exclusion chromatography. Puriﬁed protein was stored in 25 mmol/L Tris-HCl (pH = 7.5),

300 mmol/L NaCl, 1 mmol/L TCEP and 5% glycerol.

PARP7 (NP_001171646.1) amino acids 456 – 657 was cloned into pET21b expression vector with N-terminal hexahistidine and Avi

tags, and a TEV cleavage site between the hexahistidine and Avi tags. The protein was expressed in inclusion bodies in E. coli BL21-

CodonPlus (DE3)-RIL cells by addition of 0.3 mmol/L IPTG for 4 h at 37^ꢀC. Cells were resuspended in lysis buffer (40 mmol/L Tris-HCl

(pH = 8), 110 mmol/L NaCl, 2.2 mmol/L KCl, 5 mmol/L beta-mercaptoethanol, 1X Roche protease inhibitor cocktail and 20% glycerol)

and lysed by sonication. Inclusion bodies were washed with 50 mmol/L Tris-Cl (pH 8.0), 1 mol/L urea, 500 mmol/L NaCl, 1 mmol/L

Ethylenediaminetetraacetic acid (EDTA) and 0.2% (w/v) Triton X-100. The denatured protein was puriﬁed by Ni-NTA chromatog-

raphy, then the protein was then refolded by dilution into 20 mmol/L Tris-HCl (pH = 7.4), 50 mmol/L NaCl, 1 mmol/L glutathione di-

sulﬁde, 10 mmol/L glutathione, 200 mmol/L L-arginine, 400 mmol/L sucrose and 15% glycerol, followed by passage over a Ni-NTA

column for a second time. Puriﬁed protein was stored in 20 mmol/L Tris-HCl (pH = 7.5), 200 mmol/L NaCl, 5 mmol/L beta-mercap-

toethanol, 300 mmol/L imidazole and 10% glycerol.

PARP8 (NP_001171526.1) amino acids 630 – 854 was cloned into pTriJ-HV expression vector with an N-terminal hexahistidine tag.

The protein was expressed in E. coli BL21-Gold(DE3) cells by addition of 0.5 mmol/L IPTG for 16 h at 16^ꢀC. Cells were resuspended in

lysis buffer (25 mmol/L Tris-HCl (pH = 7.5), 300 mmol/L NaCl, 1 mmol/L TCEP, 0.5 % CHAPS, 10 mmol/L MgCl₂, 1 unit/mL DNase and

5% glycerol) and lysed by sonication. The protein was puriﬁed by Ni-NTA chromatography followed by cation exchange chromatog-

raphy and S-200 size exclusion chromatography. Puriﬁed protein was stored in 25 mmol/L Tris-HCl (pH = 7.5), 300 mmol/L NaCl,

1 mmol/L TCEP and 5% glycerol.

PARP9 (NP_001139574.1) amino acids 1 – 854 was cloned into a pFastBac1 insect expression vector with an N-terminal maltose

binding protein (MBP) tag and a TEV cleavage site between the MBP tag and the gene. DTX3L (NP_612144.1) amino acids 1 – 740

was cloned into a pFastBac1 insect expression vector with an N-terminal hexahistidine tag and a thrombin cleavage site between the

hexahistidine tag and the gene. PARP9 and DTX3L were co-expressed in High Five insect cells and harvested at 48 h. Cells were

resuspended in lysis buffer (20 mmol/L HEPES (pH = 7.5), 200 mmol/L NaCl, 1 mmol/L TCEP, 1X Roche protease inhibitor cocktail,

1 unit/mL DNase, 0.5% CHAPS and 5% glycerol) and lysed by passing through a high-pressure homogenizer at 200 bar then sub-

sequently at 300 bar. The protein was puriﬁed by Ni-NTA chromatography followed by MBP chromatography. Puriﬁed protein was

stored in 20 mmol/L HEPES (pH = 7.5), 200 mmol/L NaCl, 1 mmol/L TCEP and 5% glycerol.

PARP10 (NP_0011617821) amino acids 808 – 1025 was cloned into pET21b expression vector with an N-terminal hexahistidine tag

and a TEV cleavage site between the hexahistidine tag and the gene. The protein was expressed in E. coli BL21-CodonPlus (DE3)-RIL

cells by addition of 0.5 mmol/L IPTG for 18 h at 18^ꢀC. Cells were resuspended in lysis buffer (20 mmol/L HEPES (pH = 7.5), 500 mmol/

L NaCl and 5 mmol/L imidazole) and lysed by sonication. The protein was puriﬁed by Ni-NTA chromatography followed by anion ex-

change chromatography and S-200 size exclusion chromatography. Puriﬁed protein was stored in 20 mmol/L HEPES (pH = 7.5),

200 mmol/L NaCl and 2 mmol/L TCEP.

PARP11 (NP_0001273450.1) amino acids 1 – 337 was cloned into pTriIJ-HV expression vector with an N-terminal hexahistidine tag

and a C-terminal MBP tag, including a TEV cleavage site between the gene and the MBP tag. The protein was expressed in Sf21

insect cells and harvested at 72 h. Cells were resuspended in lysis buffer (25 mmol/L Tris-HCl (pH = 8), 300 mmol/L NaCl,

1 mmol/L TCEP, 1X Roche protease inhibitor cocktail, 1 mmol/L PMSF, 10 unit/mL benzonase, 0.25% CHAPS and 5% glycerol)

and lysed by sonication. The protein was puriﬁed by Ni-NTA afﬁnity chromatography, then MBP chromatography, followed by S-

200 size exclusion chromatography. Puriﬁed protein was stored in 25 mmol/L Tris-HCl (pH = 7.5), 300 mmol/L NaCl, 1 mmol/L

TCEP and 5% glycerol.

PARP12 (NP_073587.1) amino acids 489 – 684 was cloned into pTriIJ-HV expression vector with an N-terminal hexahistidine tag,

plus a TEV cleavage site and residues serine and methionine between the hexahistidine tag and the gene. The protein was expressed

in Sf21 insect cells and harvested at 72 h. Cells were resuspended in lysis buffer (25 mmol/L Tris-HCl (pH = 8), 300 mmol/L NaCl,

1 mmol/L TCEP, 1X Roche protease inhibitor cocktail, 1 mmol/L PMSF, 10 units/mL benzonase, 0.25% CHAPS and 5% glycerol)

and lysed by sonication. The protein was puriﬁed by Ni-NTA afﬁnity chromatography, followed by S-200 size exclusion chromatog-

raphy. Puriﬁed protein was stored in 25 mmol/L Tris-HCl (pH = 7.5), 300 mmol/L NaCl, 1 mmol/L TCEP and 5% glycerol.

PARP14 (NP_060024.2) amino acids 1611 – 1801 was cloned into pET21b expression vector with an N-terminal hexahistidine tag

and a TEV cleavage site between the His tag and the gene or with N-terminal hexahistidine tag, Avi tag and a TEV cleavage site be-

tween the His tag and the gene. The protein was expressed in E. coli BL21-CodonPlus (DE3)-RIL cells by addition of 0.5 mmol/L IPTG

for 16 h at 18^ꢀC. Cells were resuspended in lysis buffer (20 mmol/L HEPES (pH = 7), 500 mmol/L NaCl, 5 mmol/L imidazole and 1X

Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021 e8

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

Roche protease inhibitor cocktail) and lysed by sonication. The protein was puriﬁed by Ni-NTA chromatography, followed by S-200

size exclusion chromatography. Avi-tagged constructs were biotinylated by BirA either in vitro using recombinant enzyme and

conﬁrmed to have close to 100% modiﬁcation via mass spectrometry. Puriﬁed protein was stored in 20 mmol/L HEPES (pH =

7.5), 200 mmol/L NaCl and 2 mmol/L TCEP.

PARP15 (NP_001106995.1) amino acids 481 – 678 was cloned into pET21b expression vector with an N-terminal hexahistidine tag,

plus an additional sequence SSGVDLGT, and a TEV cleavage site between the additional sequence and the gene. The protein was

expressed in E. coli BL21-CodonPlus (DE3)-RIL cells by addition of 0.5 mmol/L IPTG for 16 h at 18^ꢀC. Cells were resuspended in lysis

buffer (25 mmol/L Tris-HCl (pH = 7.5), 300 mmol/L NaCl, 1 mmol/L TCEP, 1X Roche protease inhibitor cocktail, 1 mmol/L PMSF, 1

unit/mL DNase, 0.25% CHAPS and 5% glycerol) and lysed by sonication. The protein was puriﬁed by Ni-NTA chromatography fol-

lowed by S-200 size exclusion chromatography. Puriﬁed protein was stored in 20 mmol/L HEPES (pH = 7.5), 200 mmol/L NaCl and

2 mmol/L TCEP.

PARP16 (NP_001303872.1) amino acids 5 – 279 was cloned into pET21b expression vector with an N-terminal hexahistidine tag

and a TEV cleavage site between the hexahistidine tag and the gene. The protein was expressed in E. coli BL21-CodonPlus (DE3)-RIL

cells by addition of 0.5 mmol/L IPTG for 16 h at 16^ꢀC. Cells were resuspended in lysis buffer (20 mmol/L Tris-HCl (pH = 7.5), 300 mmol/

L NaCl, 1 mmol/L TCEP, 1X Roche protease inhibitor cocktail and 5% glycerol) and lysed by sonication. The protein was puriﬁed by

Ni-NTA chromatography, followed by cation exchange and S-200 size exclusion chromatography. Puriﬁed protein was stored in

20 mmol/L HEPES (pH = 7.5), 300 mmol/L NaCl, 2 mmol/L TCEP and 10% glycerol.

UBA1 (NP_003325.2) amino acids 1 – 1058 was cloned into a pFastBac1 insect expression vector with an N-terminal FLAG tag.

UBA1 was expressed in Sf9 insect cell lines and harvested at 48 h. The cells were resuspended in lysis buffer (20 mmol/L HEPES

(pH = 7.5), 200 mmol/L NaCl, 1 mmol/L TCEP, 1X Roche protease inhibitor cocktail, 10 mmol/L MgCl₂, 1 unit/mL DNAse and 5%

glycerol) and lysed by sonication. Protein was puriﬁed by FLAG chromatography, followed by S-200 size exclusion chromatography.

Puriﬁed protein was stored in 20 mmol/L HEPES (pH = 7.5), 200 mmol/L NaCl, 1 mmol/L TCEP and 5% glycerol.

UBE2D1 (NP_003329.1) amino acids 1 – 147 was cloned into pET21W7 expression vector with an N-terminal hexahistidine tag and

a TEV cleavage site between the hexahistidine tag and the gene. The protein was expressed in E. coli BL21-CodonPlus (DE3)-RIL

cells by addition of 0.5 mmol/L IPTG for 16 h at 16^ꢀC. Cells were resuspended in lysis buffer (25 mmol/L HEPES (pH = 7.8),

250 mmol/L NaCl, 1 mmol/L TCEP, 1X Roche protease inhibitor cocktail and 5% glycerol) and lysed by sonication. The protein

was puriﬁed by Ni-NTA chromatography, followed by TEV protease cleavage, then a second Ni-NTA chromatography puriﬁcation,

followed by S-200 size exclusion chromatography. Puriﬁed protein was stored in 25 mmol/L HEPES (pH = 7.8), 250 mmol/L NaCl,

1 mmol/L TCEP and 5% glycerol.

Ubiquitin (NP_066289.3) amino acids 1 – 76 was cloned in pET21b expression vector with an N-terminal hexahistidine tag. The

protein was expressed in E. coli BL21-CodonPlus (DE3)-RIL cells by addition of 0.5 mmol/L IPTG for 16 h at 16^ꢀC. Cells were resus-

pended in lysis buffer (25 mmol/L Tris-HCl (pH = 8), 200 mmol/L NaCl, 1 mmol/L TCEP and 5% glycerol) and lysed by passing through

a high-pressure homogenizer at 400 bar once and at 900 bar twice. The protein was puriﬁed by Ni-NTA chromatography, followed by

S-200 size exclusion chromatography. Puriﬁed protein was stored in 20 mmol/L Tris-HCl (pH = 8), 300 mmol/L NaCl, 1 mmol/L TCEP

and 5% glycerol.

Fusions tags for the four proteins used were as follows: human PARP14 catalytic domain for enzyme assay was

MHHHHHHSSGVDLGTENLYFQS followed by PARP14 fragment 1611-1801, human catalytic domain for SPR assay was

MHHHHHHSSGVDLGTENLYFQSNAGLNDIFEAQKIEWHE followed by PARP14 fragment 1611-1801, mouse catalytic domain

was MHHHHHHSSGVDLGTENLYFQS followed by PARP14 fragment 1627-1817, and human full length protein was

MHHHHHHSSGVDLGTENLYFQS followed by PARP14 residues 2-1801. Catalytic domains were expressed in E. coli while the full

length protein was puriﬁed from insect cells. Puriﬁcation steps were nickel column chromatography followed by size exclusion chro-

matography. For crystallization and SPR usage of the human catalytic domain proteins, the His6 tag was cleaved using TEV protease

prior to size exclusion chromatography, and for SPR, the protein was also biotinylated prior to SEC chromatography. For crystalli-

zation, puriﬁed, tag-cleaved PARP14 (30 mg/mL, 20 mM HEPES, 200 mM NaCl, 2 mM TCEP, pH 7.5) was incubated with RBN012579

or compound 3 as a mixture with its cis isomer at a ﬁnal concentration of 2 mM (ﬁnal DMSO concentration of 2%) for 1 h at 4^ꢀC. For

compound 3 (the cis isomer was not observed in the crystal), the protein ligand complex was crystallized via vapor diffusion using the

sitting drop method at 18^ꢀC from the following well conditions: 0.6 M ammonium sulfate, 0.1 M MES pH 6.5, 10% v/v Jeffamine M-

600 using a drop size of 0.5 mL reservoir solution and 0.5 mL well solution. Crystals appeared after 5 days. For RBN012759, the protein

ligand complex was crystallized via vapor diffusion using the sitting drop method at 18^ꢀC from the following well conditions: 0.2 M

ammonium sulfate; 0.1 M tri-sodium citrate pH 5.6; 25% w/v PEG 4000 using a drop size of 0.3 mL reservoir solution and 0.3 mL well

solution. Crystals appeared after 7 days. The structure of compound 2 was solved by crystal soaking. Crystals were grown by the

sitting drop method above and 2-methyl-1,5,6,7-tetrahydro-4H-cyclopena[4,5]thieno[2,3-d]pyrimidin-4-one, which can be pur-

chased commercially from Aldrich. The well conditions were 0.2 M lithium nitrate, 14% w/v PEG 3350. After formation, crystals

were soaked in well solution containing 5 mM compound 2 for 24 h. Crystals were cryoprotected in a solution containing 80% mother

liquor and 20% glycerol prior to vitriﬁcation in liquid nitrogen. Data reduction and scaling was performed using XIA (compound 2),

HKL3000 (compound 3), and XDS (RBN012759). Structure determination was performed by molecular replacement utilizing Phaser

program from the CCP4 software package. After manual ligand placement, iterative cycles of reﬁnement and model building were

performed using REFMAC5 and COOT, respectively. There were no Ramachandran outliers. The crystal structures have been depos-

e9 Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

ited into the Protein Data Bank with PDB codes 6WE4, 6WE3, and 6WE2 for compound 2, compound 3, and RBN012759, respec-

tively. Data collection and reﬁnement statistics are shown in Table S3.

Human PARP14 catalytic domain in vitro active site probe displacement (TR-FRET) assay

Displacement of a biotinylated small molecule probe from the NAD⁺-binding site of human PARP14 catalytic domain was measured

in vitro using a TR-FRET assay. A Mosquito (STP Labtech) was used to add 20 nL of a dose response curve of each test compound in

DMSO into black 384-well polystyrene Proxiplates (Perkin Elmer), and positive control wells contained RBN010860 and negative

control wells contained DMSO (ﬁnal concentration, f.c. = 0.2%). A Multidrop Combi (Thermo Fisher) was used to add the rest of

the reagents. Reactions were performed in an 8 mL volume by adding 6 mL of PARP14 catalytic domain (f.c. = 6 nM) and probe

(f.c. = 2 nM) in 1X assay buffer (20 mM HEPES pH = 8, 100 mM NaCl, 0.1% bovine serum albumin, 2 mM DTT and 0.002% Tween20),

incubating with test compound at 25^ꢀC for 30 min, then adding 2 mL of ULight-anti 6xHis (f.c. = 10 nM) and LANCE Eu-W1024 labeled

streptavidin (f.c. = 0.25 nM). Binding reactions were equilibrated at 25^ꢀC for an additional 60 min, then read on an Envision plate

reader equipped with a LANCE/DELFIA top mirror using excitation = 320 nm and emission = 615 nm and 665 nM with a 90 ms delay

(Perkin Elmer). The ratio of the 665/615 nm emission was calculated for each well to determine the relative amounts of complex of

PARP14 and the probe formed in each well.

Human full-length PARP14 self-modiﬁcation enzymatic activity (DELFIA) assay

Inhibition of full-length human PARP14 enzymatic activity was performed using an assay format that has been previously described

for PARP-monoenzymes (Wigle et al., 2019). Reactions were performed in a 25 mL volume in 384-well white polystyrene Ni-NTA

coated microplates at 25^ꢀC. 1X enzyme assay buffer was 20 mM HEPES (pH = 7.5), 100 mM NaCl, 2 mM DTT, 0.1% DTPA-puriﬁed

BSA and 0.002% Tween 20. Compounds were stored in 100% DMSO and 0.5 mL were dry-spotted into the microplates. Uninhibited

control wells contained DMSO (f.c. = 2%) and fully inhibited control wells contained RBN010860 (f.c. = 200 mM). His-tagged PARP14

was added in a 20 mL volume (f.c. = 10 nM) to the microplates and incubated for 30 min before the addition of 5 mL of biotinylated-

NAD⁺(f.c. = 3 mM) to initiate the enzymatic reaction. The reaction was stopped after 3 h by addition of 5 mL of NAD⁺(f.c. = 2 mM) to

outcompete the incorporation of biotinylated-NAD⁺. Quenched reactions were washed ﬁve times using 100 mL of Tris-buffered

saline + Tween 20 (TBS-T), followed by addition of 1:1000 DELFIA Eu-N1 streptavidin diluted in DELFIA assay buffer, then incubated

for 30 min at 25^ꢀC to allow the streptavidin to bind to the incorporated biotin. Next, the reactions were washed ﬁve times with 100 mL

TBS-T, followed by addition of 25 mL of DELFIA enhancement solution. Microplates were incubated 15 minutes, then the DELFIA

signal was read on an Envision plate reader (excitation = 340 nm, emission = 615 nm).

Mouse catalytic domain PARP14 in vitro active site probe displacement (TR-FRET) assay

Displacement of a biotinylated small molecule probe from the NAD⁺-binding site of mouse PARP14 catalytic domain was measured

in vitro using a TR-FRET assay similar to the one used for human catalytic domain PARP14, with the following ﬁnal concentrations:

PARP14 = 10 nM, probe = 50 nM, ULight-streptavidin = 4 nM and LANCE Eu-W1024 Anti-His6 = 1 nM.

PARP14 SPR assay

Biotin-Avi-His-TEV-PARP14^1611-1801was diluted into running buffer (50 mM HEPES (pH = 7.5), 50 mM NaCl, 1 mM TCEP, 0.05%

Tween20 and 2% DMSO) at a concentration of 1 mM and immobilized to a level of ꢂ2000 RU on a streptavidin sensor chip in a Biacore

T200 (GE Life Sciences; Marlborough, MA) using a ﬂow rate of 10 mL/min. Residual streptavidin binding sites were blocked with an

injection of PEG-biotin (Thermo Scientiﬁc; Waltham, MA) for 2 min at a ﬂow rate of 10 mL/min and a reference channel blocked with

PEG-biotin was also produced in parallel. RBN012759 binding was analyzed using single-cycle kinetics at a ﬂow rate of 80 mL/min

using an association times of 150 s, dissociation times of 300 s between injections, and a ﬁnal dissociation time of 1800 s. Following

subtraction of blank and solvent correction, data was analyzed using single-cycle kinetics ﬁtting via Biacore software.

Human PARP14 catalytic domain thermal shift assay (TSA)

Stabilization of the PARP14 catalytic domain by compound binding was measured using a thermal shift assay in a LightCycler 480 II

(Roche). Reactions proceeded in a 15 mL volume in thermal shift assay buffer (20 mM HEPES pH = 7.5 and 100 mM NaCl). PARP14

(f.c. = 3 mM) and dye (1X ﬁnal concentration) were incubated with 0.5 mL of compound (f.c. = 167 mM) stored in DMSO (f.c. = 3%) for

30 minutes at 25^ꢀC. The assay plate was sealed with an optically clear seal and heated on the LightCycler 480 II with a temperature

ramp of 0.11^ꢀC/s. The melting temperature (T_m) was determined using the ﬁrst derivative function and DT_mwas calculated by sub-

tracting the T_mof PARP14 + RBN012759 from PARP14 + DMSO.

PARP14 NanoBRET cellular biophysical assay

Displacement of RBN011198 binding to NanoLuc-tagged PARP-monoenzyme was measured in live cells using a bioluminescence

resonance energy transfer (NanoBRET) assay. A PARP14 expression plasmid was created by inserting a truncated form of PARP14

(1611-1801) into pc DNA3.1(-) vectors coding for N-terminal NanoLuc fusion. The plasmid was prepared for transfection by diluting in

OptiMEM +6% FuGENE HD and adding empty vector. A volume of 2.4 mL of diluted plasmid was added to 2 x 10⁷293T cells, incu-

bated for 24 h under standard growth conditions and then used in the NanoBRET assay. Transfected cells were resuspended in

phenol-red free OptiMEM to a concentration of 5 x 10⁵cells/mL and the NanoBRET probe was added. Next, 40 mL of cells were

Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021 e10

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

added to white polystyrene 384-well non-binding surface microplate (Corning) using a Multidrop Combi and 40 nL of a dose response

curve of each test compound in DMSO was added to the cell plate using a Mosquito. The plate was incubated in a 5% CO₂environ-

ment at 37^ꢀC for 2 h, then 20 mL per well of a solution consisting of a 1:166 dilution of NanoBRET substrate (Promega) and a 1:500

dilution of NanoLuc extracellular inhibitor (Promega) in phenol-red free OptiMEM was added to each well. Filtered luminescence was

measured on an Envision plate reader equipped with a dual 585 nm mirror, 460 40 nm bandpass ﬁlter (donor) and 610 50 nm

longpass ﬁlter (acceptor) (Perkin Elmer). Experiments were repeated at least three times.

Human primary macrophage culture

Human PBMCs (sex of sample not available) were isolated from a Leukopakꢁ (STEMCELL Technologies) by diluting with an equal

volume of EasySep^TMbuffer (STEMCELL Technologies), gently mixing, and spinning at 120 g for 10 minutes at room temperature.

The cell pellet was dislodged, washed in EasySep^TMbuffer twice, and centrifuged as described above. Red blood cells (RBC) were

lysed with ammonium chloride solution (STEMCELL Technologies) according to the manufacturer’s instructions. After RBC lysis,

PBMCs were resuspended into fresh EasySep^TMbuffer for counting and again resuspended at 5.0 x 10⁷cells/mL. Monocytes

were isolated from the PBMC cell suspension with the EasySep^TMhuman monocyte isolation kit (STEMCELL technologies) accord-

ing to the manufacturer’s instructions. Monocytes were resuspended in ImmunoCult macrophage media (STEMCELL technologies)

containing 50 ng/mL human recombinant M-CSF (STEMCELL technologies) and plated in 12-well plates for Western immunoblotting

assays (1.0 x 10⁶cells/well in 1 mL of media) or 6 well plates for RNAseq experiments (2.5 3 10⁶cells/well in 2.5 mL of media); and

allowed to differentiate for 7 days with fresh media and M-CSF (half volume) added on day 4. Macrophages were ready for down-

stream assays on day 7.

Cytokine and chemokine measurements

IL-10 concentrations in the supernatant of macrophage cultures were determined with the human IL-10 Quantikine ELISA kit (R&D

Systems) using the manufacturer’s instructions and with 3 biological replicates. For experiments with the KIRC tumor explants, cul-

ture supernatants were collected at 24- and 48 h time points for the detection of cytokine concentrations using the Invitrogenꢀ

Luminexꢀ Cytokine Human 25-Plex Panel. All samples and standards were prepared and assayed following the manufacturer’s pro-

tocol. The plate was read, and concentrations of each cytokine determined, using the BioRadꢁ BioPlexꢀ MAGPIX multiplex reader

system.

Immunoblotting

RAW264.7 or CFPAC-1 cells were plated at a density of 4.0⁵x 10 cells/well in 12-well dishes one day prior to treatment and allowed to

adhere overnight. The next day, cells were serum starved for 24 h in 0.1% FBS and subsequently treated with the indicated concen-

trations of compounds (RBN012759, compound 4, or Niraparib (AdooQ Bioscience)using DMSO as the vehicle (0.1% ﬁnal concen-

tration in each well) 6 h prior to stimulation with 20 ng/mL mouse IFN-g (Peprotech) for an additional 24 h. Cells were rinsed with PBS

on ice prior to being scraped in 150 mL of 13 radioimmunoprecipitation assay (RIPA) lysis buffer supplemented with 0.1% sodium

dodecyl sulfate and 13 Halt^TMProtease and Phosphatase Inhibitor Cocktail (Thermo Fisher Scientiﬁc). Clariﬁed lysates were as-

sayed for protein concentration by Pierce BCA assay according to the manufacturer’s instructions. To run Western blots, 30 mg of

protein in sample buffer (LI-COR) were loaded per lane on 18-well 4-12% Bis-Tris gels (Bio-Rad) and run in 3-(N-morpholino)pro-

panesulfonic acid (MOPS) buffer. Gels were transferred to polyvinylidene diﬂuoride (PVDF) membranes using a dry transfer device

(Bio-Rad). At completion of transfer, blots were rinsed with water before blocking in Odyssey blocking buffer (LI-COR) and rocking for

1 h at room temperature. Blots were then incubated by rocking with primary antibody diluted in blocking buffer overnight at 4^ꢀC. The

next day, blots were rinsed three times with 2-amino-2-hydroxymethylpropane-1,3-diol (TRIS) buffered saline with Tween-20 (TBST)

prior to incubation with secondary antibody diluted in blocking buffer and rocking at room temperature for 45 minutes. Finally, blots

were rinsed three times with TBST before imaging on an Odyssey CLx (LI-COR). Analysis was performed with Odyssey Image Studio

(LI-COR). Polarized M2-like human macrophages were lysed and processed in the same manner as described above. For the spleen

tissue samples, frozen spleen tissue was sectioned on a cold surface with a clean scalpel and homogenized in T-PER (Thermo Fisher

Scientiﬁc) lysis buffer containing 1x HALT inhibitor and 1 ml/mL benzonase (Sigma-Aldrich). The homogenized lysate was incubated

on ice for 15 minutes and centrifuged at 30,000 at 4^ꢀC for 15 minutes. Cleared supernatants were further processed as described

above. Experiments were repeated at least two times.

NanoString Assay and analysis

KIRC tumor explants were cultured as described below in the explant tumor section. Total RNA was isolated from tumor explants

using TRIzolꢀ Reagent (Thermo Fisher Scientiﬁc) following the manufacturer’s protocol. After ﬁnal isolation steps, the RNA pellet

was resuspended in warm (37^ꢀC) diethyl pyrocarbonate (DEPC)-treated water. The amount of isolated RNA was quantiﬁed using

the Invitrogen^TMQubit^TM3 Fluorometer (Thermo Fisher Scientiﬁc) and stored at -80^ꢀC for additional analyses. Total RNA was loaded

on a custom 100 gene NanoString array (Canopy Biosciences) (see list below). The NanoString raw data was normalized using

nSolver software. Expression changes for each gene and tumor treatment condition were calculated by the log-transformed ratio

between compound treated and DMSO-treated sample values. The expression change proﬁles were characterized using a heatmap

with a k-means clustering of the gene changes (k = 3) and hierarchical clustering of samples specifying the order of rows and col-

umns, respectively (R package ComplexHeatmap (v2.2.0)). The following additional genes were added to the panel and include:

e11 Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

APP, CTSA, CTSB, CTSC, CTSD, EPHA2, EPHB1, EREG, HLA-DQB2, IL3RA, KYNU, LY75, LYVE1, MMP14, MSR1, NRP2, PARP1,

PARP10, PARP11, PARP14, PARP15, PARP16, PARP2, PARP3, PARP6, PARP8, PLAUR, TIPARP, TNKS2, and ZC3HAV1.

RNA sequencing and analysis

Human primary macrophages were cultured as referenced above and plated at a density of 1.0 3 10⁶cells/well in 1 mL of macro-

phage media in 6-well dishes. Macrophages were treated with 1 mM or 0.1 mM of compound 6 h prior to stimulation with 15 ng/mL of

IL-4 (M2) or IFN- g (M1) or left in the naı¨ve condition for an additional 18 h. Sample set up included 3 biological replicates per condition.

Samples were collected 6 h and 18 h post stimulation by lysing and scraping in 350 mL/well of TRIzol^TMReagent. RNA was isolated

with QIAshredder columns (QIAGEN) and the Mag MAX kit (Ambio) according to the manufacturer’s instructions. RNA sequencing

was performed at the BioMicroCenter at MIT with a single lane of an Illumina NextSeq. Illumina software produced the raw, de-multi-

plexed sequencing intensity ﬁles.

Transcriptome sequences were mapped and quantiﬁed using Salmon software (v0.11.26). The mapping-based mode with a Gen-

code (release 29) human reference transcriptome index was used for alignment. To characterize the sequenced samples based on

their measured whole-transcriptome proﬁles, hierarchical clustering was performed using a Poisson distance metric (R package Poi-

ClaClu (v1.0.2.1)(Witten, 2011). Transcriptome read counts for each sequenced sample were then analyzed to identify differentially

expressed genes using R package DESeq2 (v1.18.17). Comparisons were made between PARP14 inhibitor-treated samples with

and without IL-4 stimulation and untreated samples with and without stimulation. The genes with signiﬁcant expression changes

(multiple testing adjusted p-values < 0.05 and an absolute log2 fold-change > 2) in either the vehicle with IL-4 stimulation compared

to vehicle and no stimulation comparison, or the PARP14 inhibitor treatment with IL-4 stimulation compared to vehicle with IL-4 stim-

ulation were extracted and their expression values normalized against the mean of the vehicle with no stimulation expression values.

A k-means clustering (k = 5) of the normalized expression values and heatmap visualization were performed with R package Com-

plexHeatmap (v2.2.0).

Substrate identiﬁcation

Cell lysis, immunoprecipitation, and mass spectrometry analysis methods have been described(Lu et al., 2019). In short, IFN-g-

treated human primary macrophages were treated with DMSO or 10 mM compound 4 for 48h. ADP-ribosylated proteins were immu-

noprecipitated using Af1521-conjugated magnetic resin (Tubip BioLabs). After incubation for 1 h at 4^ꢀC, the resin was washed 3 times

in PBS followed by elution in 100 mL of 1X sample loading buffer at 65^ꢀC for 5 min. Af1521 immunoprecipitation eluates were sub-

mitted to MS Bioworks (Ann Arbor, MI) for mass spectrometric sample preparation and analysis(Lu et al., 2019).

The detected proteins and their corresponding spectral counts were compared to the genes with opposing expression changes

with and without PARP14 inhibitor treatment in the IL-4 stimulation setting. The comparison between these sets was performed at the

gene set level, whereby overlapping enriched gene sets from the two datasets were identiﬁed. For the substrate identiﬁcation data-

set, all the proteins with a reduction in spectral count in the treated sample compared to the untreated sample were input into an

enrichment analysis with Gene Ontology gene sets using R package clusterProﬁler (v3.14.0) (Yu et al., 2012). The enrichment analysis

is based on a hypergeometric test to determine over- or under-represented gene sets among the input list of proteins. The same

enrichment analyses were performed with signiﬁcant opposing expression effects genes (adjust p-value < 0.05). For PARP14 immu-

noprecipitation, 10 mg of PARP14 antibody was conjugated to 100 mL of Protein G Dynabeads (Thermo Fisher Scientiﬁc) according to

manufacturer recommended protocol. Two mg of CFPAC-1 lysate from cells treated with or without 1 mM RBN012759 followed stim-

ulation with or without 20 ng/mL human IFN-g were incubated with resin for 1 h at 4^ꢀC. After washing, proteins were eluted by incu-

bating resin with sample buffer (LI-COR) for 5 min at 65^ꢀC then separated on 4–20% TGX polyacrylamide gels (Bio-Rad). Proteins in

the gel were transferred to a piece of Immobilon-FL PVDF membrane (EMD Millipore) using the TransBlot Turbo (BioRad) semi-dry

transfer setup. The membrane was blocked in TBS Odyssey Blocking Buffer (LI-COR) for overnight, followed by incubation with pri-

mary antibody and the QuickWestern detection reagent (LICOR) for 2 h at room temperature. After washing with TBS-T, the blot was

scanned on the Odyssey CLx infrared imaging system (Lu et al., 2019).

Human tumor explant culture and RNA extraction

The human tumor explants were grown as previously described (Page et al., 2018; Mediavilla-Varela et al., 2018). Brieﬂy, procedures

were carried out at Nilogen Oncosystems. Tumor samples (sex of patients from which tumoroids were made from is not available)

were obtained with informed consent and relevant IRB approval from patients with renal cell carcinoma. Fresh tumors were pro-

cessed into tumoroids of approximately 150 microns in size using a proprietary method, and approximately 400 tumoroids from

each individual tumor were pooled in culture plates. Tumoroids were treated as referenced in Figure 5A for 24h. Total RNA was iso-

lated from tumor organoids by using the TRIzolꢀ Reagent (Invitrogen) following the manufacturer’s protocol. After ﬁnal isolation

steps, the RNA pellet was resuspended in warm (37^ꢀC) Diethyl pyrocarbonate (DEPC)-treated water (Quality Biological). The amount

of isolated RNA was quantiﬁed using the Invitrogen^TMQubit 3^TMFluorometer (Thermo Fisher Scientiﬁc) and stored at -80^ꢀC for addi-

tional analyses.

TCGA analysis

TCGA gene expression data were extracted from Genomics Data Commons Portal using the HTSeq – Counts workﬂow analysis

(https://portal.gdc.cancer.gov). Tumor immune cell inﬁltrate and purity estimates were obtained from the TIMER resource(Li et al.,

Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021 e12

Please cite this article in press as: Schenkel et al., A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and

elicits inﬂammatory responses in tumor explants, Cell Chemical Biology (2021), https://doi.org/10.1016/j.chembiol.2021.02.010

ll

Article

2016). R packages survminer (v0.4.3) and RTCGA (v1.10.0, clinical and rnaseq 20160128) were used to perform the survival-expres-

sion analyses. Speciﬁcally, survival curves were compared for ‘‘high’’ and ‘‘low’’ PARP14 expression level groups, deﬁned by the

maxstat R package (v0.7.25).

In vivo pharmacokinetics and bioanalysis

51 mg RBN012759 was dissolved into 1.0 mL 0.5% methylcellulose and 0.2% Tween 80 in water with vortexing, sonicating and stir-

ring at room temperature for 20 min to obtain a homogenous suspension with a concentration of 50 mg/mL. The formulated com-

pound (500 mg/10mL/kg) was dosed slowly via oral gavage to n = 3 male CD-1 mice. 0.03 mL blood was collected at each time point

via submandibular bleed. Each sample was transferred into plastic microcentrifuge tubes containing Heparin-Na and mixed well,

then placed on wet ice prior to centrifugation. Samples were centrifuged at 4000 g for 5 min at 4^ꢀC to obtain plasma and stored

at -75 15^ꢀC prior to analysis.

RBN012759 compound levels were determined by HPLC-MS/MS analysis. A stock solution of RBN012759 was prepared at 1 mg/

mL in DMSO. For undiluted plasma samples, an aliquot of 30 mL sample was added to 200 mL internal standard (dexamethasone,

50 ng/mL) in acetonitrile. The mixture was vortexed for 5 min at 750 rpm and centrifuged 14,000 rpm for 5 min. An aliquot of

5.0 mL supernatant was injected for LC-MS/MS analysis. For diluted plasma samples, an aliquot of 10 mL and 3 mL sample was added

with 20 mL and 27 mL of blank plasma for 3-fold and 10-fold dilutions, respectively. For 50-fold diluted samples, an aliquot of 2 mL

sample was ﬁrst added with 8 mL of blank plasma. After vortexing for 5 min at 750 rpm an aliquot of 3 mL of diluted sample was added

with 27 mL of blank plasma for second step dilution. The diluted samples were added to 200 mL internal standard (dexamethasone,

50 ng/mL) in acetonitrile. The mixture was vortexed for 5 min at 750 rpm and centrifuged at 14,000 rpm for 5 min. An aliquot of 5.0 mL

supernatant was injected for LC-MS/MS analysis.

In vivo target engagement

C57Bl/6 J female mice were obtained from the Shanghai Lingchang BioTech Co., Ltd and kept in individual ventilation cages with free

access to food and water at constant temperature and humidity with 3 or 5 animals in each cage. Mice were between 6 to 8 weeks of

age. Mice were administered vehicle by oral gavage twice a day (BID) for 5 days or either 300 or 500 mg/kg of RBN012759 by oral

gavage twice a day (BID) for 7 days. RBN012759 was dissolved in 0.5% methylcellulose (Sigma-Aldrich) + 0.2% Tween 80 (Sigma-

Aldrich). Each dose was delivered in a volume of 0.2 mL/20 g mouse (10 mL/kg) and adjusted for the last recorded weight of individual

animals. Mice were monitored and body weight was measured daily. Plasma was collected at 2 and 12 h after the ﬁrst or the last dose.

Spleen tissues were collected at 12 h after the last dose and stored at -80^ꢀC until further analysis, described in the Immunoblotting

method above.

QUANTIFICATION AND STATISICAL ANALYSIS

In general, data were analyzed in Excel or Prism. Details of replicates and data analysis for each experiment can be found in the ﬁgure

legends or methods section.

% Inhibition

The % inhibition for all assays was calculated as shown below:

signal_cmpdꢁ signal_min

% inhibition = 100 3

signal_maxꢁ signal_min

where signal_cmpdis the assay signal from the compound treated well, signal_minis the assay signal from the positive control well and

signal_maxis the assay signal from the DMSO-treated negative control well.

IC₅₀curve ﬁtting

The % inhibition values were plotted as a function of compound concentration and the following 4-parameter ﬁt was applied to derive

the IC₅₀values:

ðTop ꢁ BottomÞ

ꢀ

ꢁ

Y = Bottom +

Hill Coefficient

1 + ðX=IC₅₀

Þ

Typically, the 4-parameters were allowed to ﬂoat, however in some cases the bottom or top of the curves were ﬁxed at 0% or 100%

respectively. IC₅₀data reported as geometric means throughout the manuscript, and geometric means standard deviations and

number of experiments is listed in the supplemental information.

Pass/fail criteria for screening plates

Plates were failed if the Z’ (Zhang et al., 1999) was below 0.5, as well as if the IC₅₀value for a reference inhibitor included on each plate

did not register to within 3-fold of its historical averaged IC₅₀

.

e13 Cell Chemical Biology 28, 1–11.e1–e13, August 19, 2021

Article Doi

A potent and selective PARP14 inhibitor decreases protumor macrophage gene expression and elicits inflammatory responses in tumor explants

DOI: 10.1016/j.chembiol.2021.02.010

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.chembiol.2021.02.010

Full text of DOI:10.1016/j.chembiol.2021.02.010

Products guided by the article

R&D Labs maybe for 6522-35-6

Relevant to this article

Hot Product