Discovery and Development of Cyclic Peptide Inhibitors of CIB1

Abstract

Calcium and integrin binding protein 1 (CIB1) is a small, intracellular protein recently implicated in survival and proliferation of triple-negative breast cancer (TNBC). Considering its interactions with PAK1 and downstream signaling, CIB1 has been suggested as a potential therapeutic target in TNBC. As such, CIB1 has been the focus of inhibitor discovery efforts. To overcome issues of potency and stability in previously reported CIB1 inhibitors, we deploy mRNA display to discover new cyclic peptide inhibitors with improved biophysical properties and cellular activity. We advance UNC10245131, a cyclic peptide with low nanomolar affinity and good selectivity for CIB1 over other EF-hand domain proteins and improved permeability and stability over previously identified linear peptide inhibitor UNC10245092. Unlike UNC10245092, UNC10245131 lacks cytotoxicity and does not affect downstream signaling. Despite this, UNC10245131 is a potent ligand that could aid in clarifying roles of CIB1 in TNBC survival and proliferation and other CIB1-associated biological phenotypes.

Calcium and integrin binding protein 1 (CIB1) is a small 22 kDa cytoplasmic protein that was originally discovered as a binder of the cytoplasmic tail of integrin αIIb.¹ Since then, CIB1 has been validated to bind multiple protein partners, including additional α-integrin tails and regulatory kinases, and has been shown to regulate a number of cellular processes.² In particular, CIB1 has been shown to play an important role in signaling pathways in cancer, especially triple negative breast cancer (TNBC). In TNBC, CIB1 is hypothesized to regulate cell proliferation and survival via interaction with PAK1 and associated downstream signaling to PI3K/AKT and Ras/Raf/MEK/ERK pathways.^3,4 Specifically, CIB1 depletion via RNA interference has been shown to decrease activity of AKT and ERK. This decrease in AKT and ERK activation leads to induction of cell death in multiple cancer cell lines, including 8 of 11 TNBC cell lines.^4,5 A common feature among CIB1-depletion sensitive TNBC cell lines is the presence of elevated levels of pAKT as well as distinctive reductions in phosphatase and tensin homologue, PTEN. CIB1 depletion effectively decreased tumor volume in mice with TNBC xenograft tumors, indicating CIB1 may function as a potential therapeutic target in TNBC containing activated AKT and/or PTEN inactivation or loss.^4,5 Based on these and other data, identification of a chemical probe or inhibitor for CIB1 would increase understanding of the mechanism behind cancer-specific cell death in CIB1-depletion sensitive and insensitive cells as well as the role of CIB1 in normal cells and its interactions with different binding partners.

Much of the structural information about how CIB1 binds its protein partners and how these interactions might be inhibited has been garnered by study of its integrin interactions. Structurally, CIB1 is composed of 10 alpha helices, 8 of which form 4 EF-hand domains with the two most C-terminal domains capable of binding Ca²⁺.^6,7 CIB1 interacts with the αIIb cytoplasmic tail via a hydrophobic pocket normally hidden by a C-terminal autoinhibitory-helix (helix 10).^6,8,9 A 15-mer stretch on the N-terminus of the cytoplasmic tail of αIIb was found to be necessary for CIB1 binding, mainly through hydrophobic interactions with this CIB1 pocket.^6,8,9 A more extensive analysis, including docking studies and in vitro binding assays, revealed that CIB1 can bind multiple α-integrin tails through this same hydrophobic pocket, but that the mode and exact positioning of the integrin tails within the pocket varies based on sequence and hydrophobicity of the binding partner.¹⁰ Cumulatively, this work established a broad binding surface within CIB1 that allows it to interact with various partners.

Because CIB1 does not possess any known intrinsic activity, the primary mode of regulation likely depends on CIB1 making protein–protein interactions (PPIs) through its large binding surface. Initial efforts to find small molecule probes of CIB1 have thus far proved unsuccessful. However, additional efforts in small molecule discovery have continued as more is uncovered about the CIB1 binding pocket, although many challenges still remain. Due to advantages that peptides hold over small molecules when inhibiting PPIs,¹¹⁻¹³ an effort to discover an inhibitor or probe for CIB1 was performed using a fully randomized 10-mer phage display library, resulting in the identification of a high affinity 16-mer peptide, UNC10245092, that binds CIB1 with a K_D of 29.4 nM and inhibits CIB1 with an IC50 of 46 nM.¹⁴ To better understand the binding mode of this interaction, UNC10245092 was cocrystallized with CIB1. The 2.1 Å crystal structure revealed that the peptide displaces α-helix 10 and binds in the same hydrophobic cleft as the integrin αIIb tail.^{2,6,8−10,15,16} To determine if UNC10245092 could mimic the effects of CIB1 depletion in cells, UNC10245092 was synthesized with an N-terminal TAT sequence, giving UNC10245350. Alanine mutations at W2 and Y10 in UNC10245350 resulted in negative control peptide UNC10245351. UNC10245350 showed up to 60% cell death in CIB1-sensitive cells but not CIB1-insensitive cells. UNC10245351 showed minimal to no cell death in either cell line.¹⁴ This observation provides strong evidence for CIB1 as a potential therapeutic target for TNBC.

Although UNC10245092 validates CIB1 inhibition as a potentially viable means of therapeutic intervention against TNBC, the peptide has some limitations as a cellular probe and therapeutic starting point. UNC10245092 required an N-terminal TAT sequence and induced significant cell death at concentrations of only 25 μM or higher.¹⁴ These features suggest that permeability and/or stability may limit the effectiveness of UNC10245092, as has been seen with other linear peptide inhibitors. Thus, we set out to identify alternative cyclic peptide inhibitors of CIB1 as potentially more stable and penetrant cellular probes. Peptide cyclization is a well-explored strategy to improve peptide cell penetrance and stability.¹⁷⁻¹⁹ We chose to use mRNA display of cyclic peptides to discover a novel binder of CIB1 that could act as an inhibitor and/or probe to better study the biology of CIB1 and its many roles in various cell types. Here, we report the discovery and characterization of UNC10245131, a novel cyclic peptide inhibitor of CIB1, with improved cell permeability and stability over UNC10245092. UNC10245131 does not, however, induce the CIB1-specific cytotoxicity of UNC10245350, suggesting that depletion of CIB1 or other means of CIB1 partner inhibition may be necessary to impede this specific biological function.

A macrocyclic peptide discovery campaign was initiated with the RaPID (random nonstandard peptide integrated discovery) system (Figure 1A, B),²⁰⁻²² which takes advantage of flexizyme technology to incorporate the unnatural amino acid, chloro-acetyl-l-tyrosine (ClAcY), in place of methionine at the initiation position during translation. Incorporation of a downstream cysteine residue allows a cotranslational S_N2 reaction between the thiol and chloroacetyl groups, leading to a thioether linked macrocycle (Figure 1B).²³ Using the RaPID system, we panned two different libraries, one presenting sequences of 6, 7, or 8 randomized residues (NNK6, 7, 8) between the thiol and chloroacetyl groups and another presenting sequences of 9, 10, and 11 randomized residues (NNK9, 10, 11), against immobilized CIB1. After four rounds of selection, sequences of hits started to resolve into families of related peptides (Figure 1C, Figure S1). Four peptides from the most highly enriched families and sequences were chosen for validation. To validate mRNA display hits, peptides were synthesized and tested in a time-resolved fluorescence resonance energy transfer (TR-FRET) assay as previously described.¹⁴ Of the cyclic peptides selected for validation, all showed inhibitory activity below 100 nM and three of the four peptides exhibited IC₅₀s below 20 nM (Figure 1D, Figure S2). This TR-FRET assay is based on displacement of a labeled version of phage peptide UNC10245092 as the FRET partner. Thus, these data suggest that the cyclic peptides bind the same hydrophobic cleft as UNC10245092 and other known CIB1 interacting proteins.^2,9,14,15 The top three peptides had similar inhibitory activity and we chose to focus on UNC10245131, which exhibited a K_d of 6 nM by isothermal calorimetry (ITC) (Figure S3), for further development.

Figure 1.

mRNA display results. (A) Outline of mRNA display selection process. (B) S_N2 reaction between C-terminal cysteine residue and N-terminal ClAcY residue to form thioether cyclization linkage during translation. (C) Families of CIB1 binding peptides are highlighted in blue. Bold peptides are representative hits from each family analyzed in this study. An asterisk indicates a detected stop codon. Red indicates an amino acid mutation from the bold peptide. (D) IC50 values for the bold peptides from panel C as determined by TR-FRET against CIB1.

To better understand structure–activity relationships between UNC10245131 and CIB1, we performed a synthetic alanine scan and molecular modeling to dock the peptide into the CIB1 crystal structure. The alanine scan showed that the Y1A, W10A, and C11A, a linear version of UNC10245131, mutations had the greatest impact on CIB1 binding and inhibition (Figure 2A). Each of these replacements resulted in more than 100-fold decrease in IC₅₀ in the TR-FRET assay. We used this alanine scanning data in combination with the knowledge that UNC10245131 could displace a UNC10245092-based probe in the TR-FRET assay to inform computational docking to the crystal structure of CIB1; specifically, into the hydrophobic cleft that binds phage peptide UNC10245092. The most energetically favorable docking conformation suggested similarities in binding poses between UNC10245131 and UNC10245092. Perhaps most importantly, in the model, UNC10245131’s residues Y1 and W10 make similar contacts as UNC10245092’s residues Y10 and W2, respectively, the same two positions that led to inactive UNC10245092 variants when mutated to alanines (Figure 2D, F).¹⁴ Additionally, cyclization of UNC10245131 appears to enable Y1 and W10 π-stacking and hydrogen bonding in the two pockets on either end of the hydrophobic cleft, in agreement with the severe decrease in affinity in the C11A replacement (Figure 2A, B). The model also suggests differences between the two peptides, in particular engagement of peptide side chains with the sides of the hydrophobic cleft. In UNC10245092, residues A5 and M6 are tucked against the interior of the pocket, whereas in UNC10245131, the analogous residues Y5 and W6 extend past the upper lip of the pocket to hydrogen bond with N67 and L135 of CIB1 (Figure 2B, C, E).¹⁴ Lastly, it should be noted that in the cellular probe based on UNC10245092, an N-terminal TAT tag likely projects along a portion of CIB1 that is untouched by UNC10245131 (Figure 2C, E).

Figure 2.

Structure and activity of UNC10245131. (A) Table of IC50 values for alanine mutants of UNC10245131; data represent average and standard deviation of a minimum of six replicates. (B) Interaction map of UNC10245131 when docked into the CIB1 binding pocket. (C) Structure of UNC10245092 (phage peptide) bound to CIB1 (PDB 6OD0) and (D) UNC10245092 bound to CIB1; dotted lines represent the interactions between Y10 and W2 with CIB1 residues K107, F115, and S160.¹⁴ (E) UNC10245131 docked into the CIB1 binding pocket and (F) UNC10245131 docked into CIB1; dotted lines represent the interactions of Y1 and W10 with CIB1 residues K107, S160, and F173.

Based on the alanine scan and docking model, we prepared a number of analogues of UNC10245131. Analogs were designed to interrogate the docked conformation and to increase lipophilicity and decrease hydrogen-bond donors (HBD) and hydrogen-bond acceptors (HBA), as these changes could contribute to better cell permeability, a known problem for large peptide therapeutics.^17,24,25 Based on the alanine scan results, K2, Q3, and N9 appeared readily amenable to replacement. We replaced these residues with lipophilic isosteres that contained fewer HBDs and HBAs (Table 1, Table S1). Since all three residues tolerated alanine mutations individually, we tested two combinations where all three residues were mutated to alanine or a similar lipophilic natural amino acid. Neither of these peptides maintained comparable activity to the parental peptide in TR-FRET (Table 1, entries 2 and 3). Mutation of N9 to a lipophilic isostere, β-(2-furyl)-l-alanine, showed very little activity compared to peptides containing N9L or N9A (Table 1, entries 5, 12, 16, and 17 compared to entries 3, 10, and 14). Loss of hydrogen bonding capabilities of both K2 and Q3 side chains was generally unfavorable with one exception, entry 11 (Table 1, entries 2, 3, 5, 13, 15, and 17). To decrease HBD of the peptide backbone, we used N-methylated versions of different residues in the peptide, as docking indicated that few backbone nitrogens were potentially involved in hydrogen bonding with CIB1. N-Methyl-N9 is not well tolerated, whereas N-methyl-W6 and N-methyl-K2A are better tolerated with only moderate losses in activity (Table 1, entries 6–8). N-Methyl-K2A and N-methyl-W6 in combination with other mutations show varying effects on the IC50 values (Table 1, entries 11, 15, and 18 compared to entries 16 and 17). Finally, N-methylation of one of the tyrosine residues is generally tolerated, with one exception (entry 13), whereas N-methylation of both tyrosines is not (Table 1, entries 10 and 11 compared to entries 9 and 12). Of the 17 analogues synthesized, UNC10245496 showed the best retention in activity in TR-FRET with the greatest increase in cLogP (Table 1, Table S1) and was chosen for further characterization alongside UNC10245131 and the K2A mutant UNC10245234.

Table 1. TR-FRET Determination of Inhibitory Activity of UNC10245131 Analogues.

entry	name	sequencea	IC50 (μM)	rep
1	UNC10245131	-S-Ac-YKQPYWLINWC-NH2	0.012 ± 0.007	16
2	UNC10245479	-S–Ac-YAAPYWLIAWC-NH2	0.82 ± 0.71	8
3	UNC10245480	-S–Ac-YAAPFWLILWC-NH2	0.30 ± 0.30	8
4	UNC10245481b	-S-Ac-YKQPY(Bzt)LIN(Bzt)C-NH2	0.23 ± 0.21	8
5	UNC10245482c	-S–Ac-Y(Nle)(Hle)PYWLI(Fua)WC-NH2	1.5 ± 0.85	8
6	UNC10245485	-S-Ac-YKQPYWLI(MeN)WC-NH2	1.8 ± 0.61	8
7	UNC10245486	-S-Ac-YKQPY(MeW)LINWC-NH2	0.059 ± 0.055	8
8	UNC10245487	-S–Ac-Y(MeA)QPYWLINWC-NH2	0.051 ± 0.063	6
9	UNC10245491c	-S-Ac-(MeY)A(Hle)P(MeY)WLINWC-NH2	1.8 ± 0.20	8
10	UNC10245492	-S–Ac-YAQP(MeY)WLILWC-NH2	0.067 ± 0.017	11
11	UNC10245493c	-S-Ac-(MeY)(Nle)(Hle)PY(MeW)LINWC-NH2	0.072 ± 0.031	14
12	UNC10245494c	-S-Ac-(MeY)(Nle)QP(MeY)WLI(Fua)WC-NH2	3.9 ± 1.4	8
13	UNC10245495c	-S-Ac-(MeY)(Nle)(Hle)PYWLINWC-NH2	0.84 ± 0.081	8
14	UNC10245496c	-S–Ac-Y(Nle)QPYWLILWC-NH2	0.059 ± 0.016	11
15	UNC10245497c	-S–Ac-Y(MeA)(Hle)PY(MeW)LINWC-NH2	0.19 ± 0.024	8
16	UNC10245498	-S–Ac-Y(MeA)QPY(MeW)LI(Fua)WC-NH2	0.19 ± 0.042	8
17	UNC10245499c	-S–Ac-Y(MeA)(Hle)PYWLI(Fua)WC-NH2	4.2 ± 1.4	7
18	UNC10245500	-S–Ac-Y(MeA)QPY(MeW)LINWC-NH2	0.054 ± 0.0060	8

^aPeptides were cyclized via the thioether moiety described in Figure 1B between the N-term acyl group and underlined C-term cysteine.

^bBzt: 3-(3-benzothienyl)-l-alanine.

^cFua: β-(2-furyl)-l-alanine, Nle: norleucine, Hle: homoleucine.

Given the similarities between cyclic peptide UNC10245131 and linear peptide UNC10245092 in binding CIB1, we sought to probe three additional characteristics: (1) cell permeability, (2) cell lysate stability, and (3) selectivity. We first measured permeability of two cyclic peptides, UNC10245234 and UNC10245496, and three linear peptides, UNC10245350, UNC10245351, and UNC10245092, using the chloroalkane penetration assay (CAPA) developed by Kritzer et al.²⁶ This assay relies on cytosolic delivery of compound analogues that are tagged with a chloroalkane that acts as a covalent ligand for strategically positioned HaloTag “receptors” within reporter cells. More penetrant compounds exhibit higher occupancy and thereby prevent fluorophore labeling of the same receptors in a subsequent chase experiment. To this end, we prepared chloroalkane-tagged analogues of the five peptides by making use of free amines at the N-terminus in the linear peptides and an introduced free lysine at the C-terminus of the cyclic peptides; because of this, UNC10245234 had to be used as a surrogate compound for UNC10245131 (see Supporting Information). Notably, both newly synthesized cyclic peptides showed better permeability as measured by CP₅₀, compared to the chloroalkane-tagged linear peptides: 0.438–0.440 μM for cyclic peptides relative to 2.086–3.291 μM for linear peptides (Table 2, Figure S4B). To put this in context, the cyclic peptides perform better in CAPA than two different cell-penetrating peptides (CPPs), HIV-TAT and DP1,²⁷ whereas our linear TAT conjugates show permeability similar to what has already been reported for TAT by itself.²⁶ Next, we measured cell lysate stability of several of the same peptides plus UNC10245131, by incubating the peptides in MDA-MB-468 lysates at 37 °C. After 24 h, cyclic peptides exhibited higher percentages of intact peptide as compared to linear peptides: >20% peptide remaining vs 0% peptide remaining (Table 2, Figure S4C), suggesting that cyclization improves both permeability and stability.

Table 2. Extended in Vitro Activity of Select Compounds.

entry	compound	description	IC50 (nM)	CP50 (nM)a	stability (%)b	% cell deathc	% cell deathd
1	UNC10245092	phage display peptide	46 ± 13	3291 ± 397	n.d.	1.7 ± 2.0	9.5 ± 3.2
2	UNC10245350	TAT-UNC10245092	150 ± 40	2086 ± 423	0	57.1 ± 13.4	9.6 ± 8.8
3	UNC10245351	TAT-UNC10245092 neg	>10 000	2225 ± 124	0	1.1 ± 0.7	1.9 ± 1.3
4	UNC10245131	mRNA display hit 2	12 ± 7	n.d.	48.6 ± 21.4	2.7 ± 2.5	2.0 ± 2.2
5	UNC10245234	UNC10245131 K2A	9 ± 3	440 ± 69	21.9 ± 19.0	2.7 ± 1.5	1.5 ± 1.3
6	UNC10245496	UNC10245131 analogue	59 ± 16	438 ± 89	62.2 ± 8.3	3.4 ± 2.4	1.1 ± 0.4

^aCP₅₀ measured using chloroalkane-tagged compounds

^bStability as percent peptide remaining after 24 h in cell lysate.

^cMDA-MB-468 (CIB1-sensitive cell line) cell death.

^dMDA-MB-231 (CIB1-insensitive cell line) cell death.

To compare selectivity of UNC10245131 and UNC10245092, we generated biotinylated versions of both peptides and performed pulldowns from cell lysates of wild-type CIB1 MDA-MB-436 cells. Both peptides potently pulled down CIB1 and the interaction could be readily competed off by soluble peptide (Figure 3B). The proteins that were pulled down were run through proteomic mass spectrometry to identify other potential protein binders of the two peptides. In four replicate assays with UNC10245131, CIB1 proved the dominant and only consistent protein observed (Figure 3C, Table S2). UNC10245092 also potently pulled down CIB1 (Table S2). Importantly, neither peptide pulled down additional, known EF-hand domain-containing proteins or close homologues of CIB1 (Table S2), suggesting high target class specificity for both peptides. Overall, UNC10245131 showed higher permeability, better stability, and similar selectivity for EF-hand domain proteins compared to UNC10245092.

Figure 3.

Cellular activity, target engagement, and proteomic pulldown. (A) Western blots of 6 h- (left) and 24 h-treated (right) MDA-MB-231 and MDA-MB-468 cells. (B) UNC10245092 and UNC10245131, immobilized via streptavidin–biotin interaction, pulled down CIB1 from MDA436 cell lysates. (C) Venn diagrams indicating the number of proteins found in 1, 2, 3, or 4 replicates of the proteomic pulldown. Colors of the Venn diagram each indicate a unique replicate.

Lastly, we tested the peptides for cytotoxicity against two different TNBC cell lines: MDA-MB-468, a CIB1-depletion sensitive cell line, and MDA-MB-231, a CIB1-depletion insensitive cell line.⁵ Because MDA-MB-231 cells do not exhibit increased cell death upon CIB1 depletion,⁵ this cell line serves as an important control for nonspecific toxicity. Cytotoxicity was assessed by trypan blue exclusion test for cell viability. Cyclic peptides showed less than 10% cell death over 24 h (Table 2, Figure S4A), compared to UNC10245350, which showed an average cell death of 57% over 24 h (Table 2, Figure S4A). Decreases in phosphorylated AKT and phosphorylated ERK are two of the key characteristics of CIB1 depletion in the CIB1-depletion sensitive TNBC cell lines,⁵ indicating a decrease in AKT and ERK signaling. Thus, we asked if CIB1 inhibition also affected the activity of AKT and ERK in peptide treated cells. At 6 and 24 h, UNC10245131 showed no effect on the activity of AKT or ERK relative to DMSO control (Figure 3A). In contrast, at both time points, UNC10245350 showed a decrease in both pAKT and pERK compared to DMSO control, consistent with our previous report (Figure 3A).¹⁴ Notably, in addition to decreases in pAKT and pERK at the 6 h time point, UNC10245350 treated cells also showed a substantial decrease in CIB1 levels, not seen in cells treated with UNC10245131 (Figure 3A).

Calcium and integrin binding protein 1, CIB1, has been suggested as a potential therapeutic target for the treatment of triple-negative breast cancer (TNBC). Here, we report the discovery of a new peptide inhibitor of CIB1, UNC10245131. UNC10245131 is a cyclic peptide with high affinity and good selectivity, and increased permeability and stability over earlier, linear peptide inhibitors, UNC10245092/UNC10245350. The low stability of UNC10245350 may account for the high concentrations of this peptide needed to induce a response in cells. Together, the TR-FRET data and docking studies on UNC10245131 suggest that both UNC10245131 and UNC10245092 similarly take advantage of two deeper pockets at either end of the known binding cleft of CIB1. Additionally, UNC10245131 potently and selectively pulls-down CIB1 from cellular lysates (Figure 3B). These findings led us to hypothesize that UNC10245131 should be able to disrupt CIB1 interactions and effect AKT/ERK signaling in TNBC, similar to UNC10245350. However, UNC10245131 does not induce cell death in CIB1-sensitive TNBC cells, whereas UNC10245350 induces approximately 57% cell death in the same CIB1-sensitive TNBC cell line, while leaving CIB1 insensitive cells healthy. UNC10245131 does not appear to alter AKT and ERK phosphorylation levels while treatment with UNC10245350 results in decreases in pAKT and pERK.

Cumulatively, these results suggest that UNC10245350 and UNC10245131 engage CIB1 differently in cells. It is possible that the difference in activity between the two peptides is due to UNC10245131’s inability to effectively engage CIB1 in a cellular context. However, our MS data demonstrates that UNC10245131 engages CIB1 potently and selectively in the relatively complex environment of cell lysates so there are potentially other causes for this difference in cellular activity. A closer look at CIB1 protein levels after treatment with UNC10245350 showed a consistent decrease in CIB1 protein levels not seen in cells treated with UNC10245131. This may highlight major differences in how the peptides are interacting with or directing CIB1 in cells. This decrease in CIB1 protein levels could be the result of at least two different mechanisms. First, the N-terminal TAT sequence on UNC10245350, which helps improve permeability, may also affect localization of CIB1 within the cell, leading to its degradation, whether that is proteasomal, endosomal, or other. UNC10245092, the initial peptide discovered via phage display, exhibits some permeability in CAPA, and yet does not induce cell death (Table 2), but the addition of the TAT sequence significantly improves activity of the peptide. Future efforts to elaborate UNC10245131 into a proteolysis targeting chimera (PROTAC) or other form of degrader ligand could cement this hypothesis. Alternatively, the TAT tag on UNC10245350 may alter its binding pose and/or help to block different interactions relative to UNC10245131. We previously showed that addition of a different CPP, penetratin, to the N-terminus of UNC10245092 completely abrogated binding,¹⁴ so it is possible the TAT sequence affects the binding pose of UNC10245350. We were able to solve the structure of UNC10245092 bound to CIB1, but not TAT-tagged UNC10245350. Significantly, prior work has shown that different integrin tails can be positioned very differently within the extended binding pocket of CIB1.¹⁰ Addition of the N-term TAT tag might further prevent access to CIB1 partners or else push the C-terminus of UNC10245092 further along the extended binding interface, thereby disrupting additional CIB1 interactions. These differences in binding modes between peptides could account for the differences in cellular activity reported here.

In conclusion, UNC10245131 is a high affinity, permeable, and cell stable ligand for CIB1 and could serve to study CIB1 interactions and biology in greater detail. CIB1 has been connected to a number of roles in different cell lines.² In addition to AKT/ERK signaling in TNBC, CIB1 has been implicated in integrin biology, including control of adhesion, spreading,^28,29 and cell migration,^30,31 and in other related cellular processes such as intracellular Ca²⁺ signaling.^2,32−34 A potent CIB1 inhibitor or cellular probe such as UNC10245131 could be useful in dissecting these additional cellular roles of CIB1. Within cancer, future studies to discern the mechanistic basis of differences in activity between UNC10245350 and UNC10245131 could involve the synthesis of PROTAC or other differentially modified analogues of UNC10245131 as well as structural studies on both UNC10245350 and UNC10245131.

Glossary

Abbreviations

CIB1

calcium and integrin binding protein 1

TNBC

triple negative breast cancer

PTEN

phosphatase and tensin homologue

PPI

protein–protein interaction

ClAcY

chloroacetyl-l-tyrosine

RaPID

random nonstandard peptide integrated discovery

TR-FRET

time-resolved fluorescence resonance energy transfer

ITC

isothermal calorimetry

HBD

hydrogen bond donors

HBA

hydrogen bond acceptors

CAPA

chloroalkane penetration assay

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.1c00438.

• Experimental details, synthetic schemes, purity traces, and figures (PDF)

The authors declare no competing financial interest.