Efficacy of Rac and Cdc42 Inhibitor MBQ-167 in Triple-negative Breast Cancer

Triple-negative breast cancer (TNBC) comprises approximately 15% of all breast cancers in Caucasians and approximately 35% in African Americans. It is often diagnosed in younger women and prognosis for long-term survival is very poor for patients with metastatic TNBC (1). Only few FDA-approved therapeutics are available due to its heterogenous nature and the absence of the three main current membrane receptors for drug intervention: estrogen receptor (ER)/progesterone receptor (PR) and HER2 (2). In recent years, the immune checkpoint therapies atezolizumab (tecentriq) and pembrolizumab (keytruda) were approved for treatment of TNBC in combination with chemotherapy, but their efficacy is dependent on expression of programmed death ligand (PDL-1) in tumor cells or immune cells in the microenvironment (3). The April 2020 FDA approval of sacituzumab govitecan-hziy (Trodelvy) for previously treated metastatic patients with TNBC that do not respond to standard chemotherapy is highly promising. However, this antibody–drug conjugate of a Trop-2 antibody with a topoisomerase inhibitor, may not target dormant stem-like cells with the capacity for metastasis. Moreover, patients on this drug exhibited myelotoxicity as a side effect, which can result in anemia and fatigue (4). Alternate therapies that target signaling molecules in the PI3-K/Akt and MAPK pathways are in various stages of development, but none have reached FDA approval (5). Therefore, the current prognosis for metastatic TNBC still remains poor, highlighting the urgent need for efficacious metastasis targeted therapies (6).

Herein, we describe our studies targeting the key signaling GTPases Rac and Cdc42, which are regulated by several oncogenic cell surface receptors, such as growth factor receptors (not only HER2) that activate guanine nucleotide exchange factors (GEF) for Rac and Cdc42. Activated Rac and Cdc42 (GTP bound) stimulate p21-activated kinase (PAK), PI3-K/Akt, and MAPK pathways leading to cancer metastasis (7). Rac and Cdc42 are ideal metastatic cancer drug targets because they direct the formation of motile structures required for intravasation into the circulatory system to initiate metastasis (8). In addition to orchestrating the extension of invasive actin structures, Rac and Cdc42 have also been implicated in tumor initiation, progression, and therapy resistance due to their central roles in cell-cycle progression, survival, membrane trafficking, and gene transcription (9). Rac and Cdc42 proteins are generally not mutated in cancer but rather overexpressed or hyperactivated. So far, over 80 potential Rho family GEFs are known and have been identified as oncogenes. Therefore, targeting Rac and Cdc42, which share about 70% structural homology and have similar, as well as distinct GEFs and downstream effectors, is a viable approach. Nevertheless, to date no Rac/Cdc42 inhibitor has received regulatory approval for the treatment of cancer (10, 11).

To fill this need, we developed MBQ-167 (9-Ethyl-3-(5-phenyl-[1,2,3]triazol-1-yl)-9H-carbazole), which inhibits Rac 1/2/3 activity with an IC₅₀ of 103 nmol/L, and Cdc42 activity with an IC₅₀ of 78 nmol/L (12). MBQ-167 reduces high-grade cancer cell polarity and migration, resulting in subsequent detachment from the substratum. At longer incubation time (96 hours) or higher concentration (500 nmol/L), MBQ-167–mediated loss of polarity, and detachment from the substratum culminates in anoikis (apoptosis due to dissolution of integrin signaling) in metastatic breast cancer cell lines. In contrast, epithelial MCF10A or non-metastatic MCF7 cells are not affected. Moreover, MBQ-167 reduces mammosphere formation in TNBC cells; thus, indicating a suppression of stem cell–like activity. Accordingly, MBQ-167 inhibits mammary tumor growth, metastasis, and angiogenesis in a HER2-type mouse model (12). Moreover, MBQ-167 has a bioavailability of approximately 35% in plasma and tissue, including tumor tissue, following intraperitoneal or per oral administration (13).

The current study demonstrates the efficacy, safety, and availability of MBQ-167 in in vitro and in vivo TNBC models. Our previous report demonstrated that MBQ-167 is a specific and potent inhibitor of the activation of Rac and Cdc42 and their downstream effector PAK. Herein, we have further explored the specificity and mechanism of action of MBQ-167 using phosphokinome arrays and show that initial increases in certain oncogenic pathways in response to MBQ-167 are not sustained over prolonged treatment. Taken together, this study validates future development of MBQ-167 as a TNBC therapeutic agent.

MBQ-167 was synthesized to >99% purity at Millipore Sigma, using an adaptation of our published protocol (12).

GFP-tagged MDA-MB-231, MDA-MB-468 (ATCC), and 4T-1 (Anticancer, Inc.) were cultured and maintained, as described previously (12). All cells were authenticated by short tandem repeat DNA profiling, and were routinely tested for the presence of Mycoplasma by PCR.

MDA-MB-231 cells were treated for 24 hours with vehicle (0.04% DMSO), or varying concentrations of MBQ-167 (250–500 nmol/L). The attached and detached (D) cell populations were recovered and equal number of cells plated for 7 days. Colonies (defined as a cluster of 50 or more cells) were fixed and stained with 50% MeOH, 50% crystal violet and quantified at 4× magnification.

The CellTiter 96 Non-Radioactive Assay (Promega) was used according to manufacturer's instructions. Briefly, cells were seeded in a 24-well plate and treated for 48 hours with vehicle or MBQ-167 at the indicated concentrations. After incubation, the MTT (3-(4,5-dymethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide) reagent was added, and plates were incubated for 4 hours at 37°C. Formazan absorbance was measured at 570 nm.

Rac activity was analyzed from MDA-MB-231, MDA-MB-468, and 4T-1 cell lysates by pulldown assays. The P21-binding domain of PAK 1 was used to isolate active GTP-bound Rac, as described previously (14). Active and total Rac GTPases were separated in a 12% SDS-PAGE gel and identified by Western blotting using Rac specific antibodies (Cell Signaling Technology, Inc).

As described in ref. 12, equal total protein amounts from cell lysates were run on SDS-PAGE gels and Western blotted using specific antibodies to Rac, Cdc42, phospho-Jun (Ser 63), c-Jun, phospho-Jun kinase (JNK) (Thr 183/Tyr 185), JNK, phospho-CREB (Ser 133), CREB, Zeb, survivin, and cyclin D. Anti-β-actin was used for normalization. The integrated density of positive bands of total and phospho bands were quantified using Image J software, as per routine laboratory protocols.

The Human Phospho-Kinase Array (ARY003B) from R&D Systems with antibodies to 45 specific phosphoproteins was used. MDA-MB-231 cells were treated with MBQ-167 for 24 hours at 200 nmol/L and the lysates were processed and quantified as per manufacturer's directions.

All animal studies were conducted under approved animal protocols by the University of Puerto Rico Medical Sciences Campus Institutional Animal Care and Use Committee, under protocols A8180118 and A8180218, in accordance with the principles and procedures outlined in the NIH Guideline for the Care and Use of Laboratory Animals. Four to 5 weeks- old female SCID or BALB/c mice (Charles River Laboratories, Inc.) were housed under pathogen-free conditions in high efficiency particulate air-filtered cages and kept on a 12-hour light/dark cycle, and controlled temperature (22°C–24°C), and humidity (25%).

GFP-tagged MDA-MB-231 (∼5 × 10⁵) or 4T-1 cells (2.5 × 10⁵) in Matrigel (BD Biosciences) were injected at the fourth right mammary fat pad of SCID mice under isofluorane inhalation. After tumor establishment (∼100 mm³, 1 week post-inoculation), animals were randomly divided into treatment groups (n = 10).

A total of 3 × 10⁵ GFP-expressing 4T1 cells were injected via the tail vein of Balb/c mice. Treatments began after 24 hours and were administered by intraperitoneal injections 5× a week.

For the experiments with the MDA-MB-231 cell line, SCID mice were treated with vehicle (12.5% ethanol, 12.5% Cremophor (Sigma-Aldrich), 75% 1× PBS pH 7.4; formulation A), or 1 or 10 mg/kg body weight (BW) MBQ-167 (dissolved in vehicle) by intraperitoneal injection in a 100 μL volume 3× a week.

In the experiment with the 4T-1 cell line in BALB/c mice, for oral gavage, MBQ-167 was dissolved in 0.5% methyl cellulose, 0.1% Tween 80 in PBS and mixed by homogenization and administered in a 100 μL volume 5× a week (formulation B). Treatments began when the tumors were approximately 100 mm³ and continued for 4 weeks until sacrifice.

Mammary tumor growth was quantified as changes in the integrated density of GFP fluorescence. Mice were imaged weekly following breast cancer cell inoculation (on day 1 of treatment administration) and once a week thereafter. The FluorVivo small animal in vivo imaging system (INDEC Systems, Inc.) was used for whole-body imaging of GFP fluorescence. Tumor fluorescence intensities were analyzed using Image J software (NIH, Bethesda, MD). Relative tumor growth was calculated as the integrated density of fluorescence of each tumor on each day of imaging relative to the integrated density of fluorescence of the same tumor on day 1 of treatment administration, relative to vehicle controls.

Following sacrifice, lungs were excised and immediately stored in liquid N₂. Stored organs were thawed and analyzed by fluorescence microscopy, as described previously (12).

Serum from BALB/c mice following administration of MBQ-167 3× a week by intraperitoneal at 10 mg/kg BW was subjected to a blood chemistry panel for indicators of liver toxicity, using a commercial source (CorePlus Labs).

A previously validated method was used to quantify MBQ-167 in plasma and tumor tissues using an Acquity UPC^2TM system (Waters Corp.), coupled to a triple quadrupole MS-MS (13). Thirty-five BALB/c mice were injected with 2.5 × 10⁵ 4T1 murine metastatic breast cancer cells at the mammary fat pad under isoflurane inhalation (1%–3% in oxygen using an inhalation chamber at 2 L/minute) to produce primary tumors (2 weeks following cell inoculation, ∼200 mm³ tumors). Mice were treated with a single dose of MBQ-167 (10 mg/kg BW) or oral gavage (100 mg/kg BW) and 5 mice/group were sacrificed by cervical dislocation at 0.5, 1, 3, 6, 9, 12, and 24 hours. Following sacrifice, blood was collected by cardiac puncture and tumors were snap-frozen in liquid nitrogen, and stored at −80°C. Data analysis was performed by noncompartmental analysis, according to a uniform weighing scheme, using Phoenix WinNonlin professional software, Version 8.1 (13).

Rac1 construct was expressed in transformed BL21 Escherichia Coli using pGEX-2T plasmid encoding Rac1 with a GST fusion tag. For isotopically labeled Rac1, we used M9 minimal media with ¹⁵N-ammonium chloride (¹⁵NH₄Cl from Cambridge Isotope) as the sole nitrogen source. For unlabeled Rac1, we grew the bacteria using Luria broth media. Labeled amino acids were added 30 minutes prior to induction, where the cell culture was induced at an optical density of 0.7 with 0.1 mmol/L of Isopropyl β-D-1-thiogalactopyranoside at 20°C for 16 hours. After harvesting and sonication, all buffer used for the step and beyond contained an excess of GDP to preserve the bound conformation of Rac1. The protein was purified by chromatography on Glutathione Sepharose 4B resin (GE). To cut the GST tag, we used Thrombin Clean Kit (Sigma-Aldrich) and separated Rac1 from GST using again Glutathione Sepharose 4B GST. Rac1 was further purified by size exclusion chromatography using HiLoad 26/60 Superdex 75 pg (GE). Samples were concentrated using Amicon Ultra Centrifugal Filter Ultracel 10 kDa (Merck Millipore). Purity of the sample was validated by SDS Page.

To prepare samples for nuclear magnetic resonance (NMR) spectroscopy, we exchanged the buffer to 20 mmol/L Hepes-NaOH (pH 7.0) 50 mmol/L NaCl, 5 mmol/L MgCl₂, and 0.5 mmol/L GDP using a Corning Spin-X Concentrator. The NMR titration experiments were performed on a Bruker Avance 700 MHz spectrometer equipped with a QCI cryoprobe. All spectra were processed and analyzed with Bruker TopSpin and NMRFAM-Sparky software (15). MBQ-167 was diluted in DMSO for 10 mmol/L stock samples. For titration experiments, we used a fixed concentration of 100 μmol/L of Rac1-GDP and added aliquots for final concentrations of MBQ-167 of 100, 200, 400, and 800 μmol/L. For each added concentration, we sonicated the sample and let it incubate at room temperature for 1 hour. For each concentration, we recorded at two-dimensional ¹H-¹⁵N heteronuclear single quantum coherence (HSQC) spectrum at 25°C to correlate the chemical shifts of amide proton and amide nitrogen nuclei. Partial resonance assignments were obtained by comparison with a published NMR structure (PDB ID: 6agp, BMRB accession number: 27577). Each ¹H-¹⁵N HSQC spectrum was compared with spectra of Rac1 with corresponding aliquots of DMSO added without any drug. Data were analyzed by overlaying each spectrum and quantifying the chemical shift and signal intensity changes as a function of added drug.

Statistical comparisons between vehicle-treated and MBQ-167–treated samples were done by Student t test using GraphPad Prism 6 or Excel. Differentially expressed proteins were selected at >1.5-fold expression, statistical significance of P < 0.05.

The small molecule MBQ-167 inhibits Rac 1/2/3 activity with an IC₅₀ of 103 nmol/L, and Cdc42 activity with an IC₅₀ of 78 nmol/L in the MDA-MB-231 TNBC cell line. Following Rac/Cdc42 inhibition, prolonged MBQ-167 treatment results in cell-cycle arrest, reduced cell growth, and increased apoptosis. The specific inhibition of Rac and Cdc42 activation by MBQ-167 blocks signaling via their downstream effector PAK, which leads to the reduction of motile actin structures, attachment with the extracellular matrix, and metastatic cancer cell viability and migration. Consequently, MBQ-167 significantly reduced tumor growth, metastasis, and angiogenesis in a HER2 (+) metastatic breast cancer mouse model (12). Herein, we report on the effects of MBQ-167 in TNBC cell lines in vitro and in vivo.

MBQ-167 treatment of TNBC, HER2 (+) breast, gastric, pancreatic, and ovarian cancer and neuroblastoma, as well as yeast cells that are dependent on Cdc42 for cell polarity, result in a loss of cell viability and polarity and detachment from the substrate (12, 16). This phenotype is demonstrated in Fig. 1A, where MDA-MB-231 cells respond to MBQ-167 treatment for 24 hours by cell rounding and detachment from the substratum. In addition, MBQ-167 causes 20%–42% of the total cells to detach from the substratum. From this detached cell population, only 30%–50% are viable. Previously, using a MTT cell viability assay, we reported a GI₅₀ of 130 nmol/L for MDA-MB-231 cells following 120 hours of MBQ-167 treatment, where all of the detached cells were nonviable after 120 hours (12). Clonogenic assays of separated attached and detached populations following treatment of 250 or 500 nmol/L MBQ-167 for 24 hours showed that compared with vehicle treatments, MBQ-167, at 250 or 500 nmol/L, reduced colony formation from the attached population by approximately 50% and by approximately 75% from the detached cell population (Fig. 1B and C). In addition, both MDA-MB-468 and 4T-1 TNBC cells demonstrated >80% inhibition of cell viability, as measured by a MTT assay, following MBQ-167 (250 nmol/L) for 48 hours (Fig. 1D and E) To confirm that MBQ-167 exerts its effect on aggressive TNBC cells via inhibition of Rac and Cdc42 activation, similar to MDA-MB-231, Rac activity was determined in the human TNBC MDA-MB-468 cells and the 4T-1 mouse TNBC cell line following 250 or 500 nmol/L MBQ-167 for 24 hours. All TNBC cells tested demonstrated a 70%–80% decrease in GTP bound Rac with no corresponding change in total Rac expression, thus validating the role of MBQ-167 as a Rac/Cdc42 inhibitor (Fig. 1F and G).

Decreased Rac and Cdc42 activities in response to MBQ-167 resulted in reduced signaling of their downstream effector PAK, as was ascertained by specific total and phospho antibodies to PAK isoforms and specific phospho residues, as well as the PAK downstream effectors LIM kinase and cofilin (12). In this study, to demonstrate specificity of MBQ-167 to the Rac/Cdc42/PAK pathway, we tested the effects of 200 nmol/L MBQ-167 (double the IC₅₀) for 24 hours in MDA-MB-231 cells using a commercially available phosphokinome array. Combined attached and detached live cells were lysed and subjected to a kinome array with mono-specific antibodies to 45 phosphoproteins associated with cancer cell malignancy. As shown in Fig. 2A, the phospho-(active) status of all kinases tested were not altered in response to MBQ-167 and no phosphoproteins were reduced below the cutoff of ±0.5-fold change or P < 0.05 from vehicle controls. However, MBQ-167 increased the phospho (active) levels of transcription factors CREB and c-Jun by 1.75-fold and 0.75-fold, and P values of 0.01 and 0.025, respectively. Because Jun Kinase (JNK), which phosphorylates and activates c-Jun is a potential PAK target (17), the array results were confirmed by Western blot analysis for phospho-c-Jun and c-Jun. c-Jun expression did not change in response MBQ-167, but phospho-c-Jun levels were upregulated by approximately 1.7-fold at 24 hours following 200 nmol/L MBQ-167. To determine whether c-Jun was phosphorylated because of an upregulation of JNK activity, we determined phospho and total JNK levels by Western blot analysis and found no changes in either JNK expression or activation in response to 200 nmol/L MBQ-167 for 24 hours (Fig. 2B and C).

The cellular response to MBQ-167 is characterized by a loss of cell polarity and substrate attachment followed by apoptosis (anoikis; Fig. 1A; ref. 12). Therefore, after MBQ-167 treatment, detached and attached MDA-MB-231 cell populations were separated and analyzed by phosphokinome array and Western blot analysis. Phosphokinome array of separated attached and detached cell lysates following 200 nmol/L MBQ-167 still demonstrated a >2-fold increase in phospho-c-Jun levels at 24 hours. However, CREB and P-CREB levels in attached and detached populations were only increased by 0.5-fold in phosphokinome arrays (Supplementary Fig. S1), and Western blot analysis showed that P-CREB levels dropped to vehicle control levels (0.9-fold) by 48 hours in 200 nmol/L MBQ-167 (Fig. 3A). A time-course analysis via Western blot analysis revealed that the initial increase in c-Jun phsophorylation is transient, peaking at 48 hours to rapidly decline at 96 hours. In addition, total c-Jun protein expression did not change in the attached cell population, but decreased in the detached population at 24 hours following MBQ-167 treatment to drop to undetectable levels by 48 hours. Therefore, 96 hours of MBQ-167 treatment resulted in statistically significant decreases in p-c-Jun and c-Jun, as well as the Jun and CREB transcriptional targets: Zeb1 and cyclin D, especially in the detached cell population (Fig. 3A,–D). MBQ-167 treatment also did not significantly change the CREB target survivin expression, which remained at slightly lower levels even after 96 hours treatment in both attached and detached cell populations (Fig. 3).

To determine whether the initial increases in p-c-Jun and p-CREB and their transcriptional pro-cancer targets were a mechanism of therapy resistance, we attempted to create MBQ-167–resistant MDA-MB-231 cell lines. MDA-MB-231 cells were treated with MBQ-167 three times per week with treatment concentrations ranging from 100–500 nmol/L, but all of the cells detached and ultimately lost viability, and no resistant cells were recovered after 3 weeks. Therefore, we posit that the initial increase in the activity of mitogenic transcription factors c-Jun and CREB is a transient resistant mechanism to MBQ-167 treatment that is not sustained over time.

We employed NMR chemical shift mapping to determine the interaction site of MBQ-167 by using ¹H-¹⁵N HSQC to detect the backbone of Rac1 in a GDP-bound conformation, as described in refs. 18, 19. Site-specific chemical shifts were assigned by comparison with chemical shift resonances already reported in the literature under similar conditions (BMRB Entry Access: 27577 and 6970). Supplementary Figure S2 shows the spectra of Rac1-GDP where chemical shift changes were observed with increasing concentrations of MBQ-167. Signals were selected for analysis when they were at least three times more intense than the signal-to-noise ratio. Chemical shift changes were evident for Gly-12, Val-85, and Asn-92 at MBQ-167 concentrations of 200 and 400 μmol/L. The small chemical shift deviation is consistent with subtle conformation changes to the protein backbone. The chemical shift at Gly-12 is highly relevant because it is in the binding site of GDP and GTP, which is critical for Rac1 function, Additional chemical shift changes outside the GDP/GTP binding region may be due to protein flexibility or secondary binding at high MBQ-167 concentrations.

We determined whether the decrease in Rac/Cdc42 activation, pro-cancer proteins, and cell viability also resulted in reduced tumor progression in TNBC. The HER2⁺ cell line we used previously expresses high Rac1 and Rac3 proteins, with enhanced activity, and is a high metastatic variant of the MDA-MB-435 cell line (14). Herein, we determined whether MBQ-167 can be used to treat TNBC, by using the MDA-MB-231 TNBC cell line, which expresses Rac1 at lower levels without expressing the constitutively active Rac1B isoform (20). Immunocompromised mice bearing GFP-MDA MB-231 TNBC mammary fat pad tumors were treated with 0 or 10 mg/kg BW MBQ-167 3× a week by intraperitneal injection for 65 days. We observed a 96% decrease in tumor size following MBQ-167 treatment where the relative average tumor growth for vehicle-treated mice was 37.5 ± 10 and 1.85 ± 0.4 for MBQ-167–treated mice (Supplementary Fig. S3).

Because 10 mg/kg BW MBQ-167 treatment resulted in such a drastic reduction in tumor growth in this pilot study, we investigated the effective concentration of MBQ-167 in the MDA-MB-231/SCID mouse model by determining the effect of 0, 1, 5, or 10 mg/kg BW on GFP-MDA-MB-231 mammary tumor growth and metastasis following 3× a week treatment via intraperitoneal. Average tumor growth, as quantified by integrated density of GFP fluorescence on each day of imaging compared with the individual fluorescent tumor images on the first day of MBQ-167 treatment, was decreased by approximately 80%–90%, which saturated at an MBQ-167 dose of 1 mg/kg BW (Fig. 4A and B). The increase in tumor fluorescence intensity measurements was validated by quantifying average tumor weight following necropsy, where 1 mg/kg BW treatment resulted in approximately 100% decrease in tumor weight (Supplementary Fig. S2B). This drastic reduction in tumor growth was accompanied by an approximately 90% decrease in lung metastasis, in this spontaneous metastasis model (Fig. 4C). To determine whether the reduced metastasis was solely due to the reduction in tumor growth, we performed an experimental metastasis assay by tail vein injections of GFP-tagged MDA-MB-231 mouse TNBC cells in SCID mice. One week following tail vein injection, MBQ-167 was administered by intraperitoneal at a dose of 5 mg/kg BW, three times a week for 4 weeks. Figure 4D shows that this experimental model of metastasis also resulted in a significant 60% reduction of lung metastases in response to MBQ-167 treatment. We report here that 1–10 mg/kg BW MBQ-167 administered by intraperitoneal in the formulation A (homogenous solution Cremophor/ethanol) is sufficient to reduce tumor growth and metastasis in a human TNBC mouse model.

Formulation studies conducted at Charles River Labs indicated that higher concentrations of MBQ-167 can be achieved by an oral formulation in a suspension in 0.5% methyl cellulose, 0.1% Tween 80 (formulation B). Using this formulation, in dose-escalation and repeat dose studies, MBQ-167 was not toxic to rodents at a dosage as high as 4,000 mg/kg BW (Personnel communications, Charles River Labs). Therefore, to determine the efficacy and utility of oral administration of MBQ-167 at a range of concentrations in immunocompetent mice, MBQ-167 was administered orally in formulation B to BALB/c mice bearing GFP-tagged 4T-1 mammary fat pad tumors (∼300 mm³) 5× a week for 28 days. Tumor growth, as assessed by fluorescence image analysis, was reduced for all MBQ-167 treatments compared with vehicle-treated tumors. Mice treated with 25, 50, or 100 mg/kg BW MBQ-167 demonstrated reduced tumor growth kinetics compared with vehicle controls (Fig. 5A). At the day of sacrifice (day 28), the tumors of vehicle mice increased by 7-fold while the tumors from 5 or 25 mg/kg BW MBQ-167 treated mice demonstrated a nonstatistically significant 25% and 40% decrease compared with vehicle, while the tumors from the 50 and 100 mg/kg BW MBQ-167 treated tumor growth was inhibited by 60% in a statistically significant manner (Fig. 5B).

Fluorescent image analysis of excised lungs following 4 weeks of MBQ-167 treatment showed no changes in lung metastases at 0, 5, or 25 mg/kg BW MBQ-167 treatment, while the mice treated with 50 or 100 mg/kg BW MBQ-167 demonstrated approximately 80%–90% reduction in metastasis (Fig. 5C). Figure 5D demonstrates representative lungs from this experiment, where fluorescent metastatic foci were observed in lungs from mice treated with vehicle, but not MBQ-167. Therefore, higher levels of MBQ-167 are required to reduce lung metastasis from this immunocompetent spontaneous metastasis model.

We previously reported that immunocompromised mice bearing mammary fat pad tumors, did not demonstrate significant changes in weight, following 10 mg/kg BW administration of MBQ-167 for 65 days (12). Similarly, immunocompetent BALB/c mice bearing 4T-1 mouse tumors did not show a significant change in weight following 100 mg/kg BW MBQ-167 treatment for 4 weeks (Fig. 6A). Moreover, blood cell counts and clinical chemistry, as well as indicators of liver toxicity: total bilirubin, aspartate aminotransferase, albumin, alanine aminotransferase, alkaline phosphatase were not significantly altered in the serum of BALB/c immunocompetent mice, after MBQ-167 treatment for 4 weeks (Fig. 6B).

We recently published the pharmacokinetics and plasma and tissue distribution of MBQ-167 in BALB/c mice following 10 mg/kg BW MBQ-167 via both intraperitoneal and oral administration (13). MBQ-167 demonstrated an acceptable pharmacokinetic/pharmacodynamic profile with approximately 35% bioavailability in the formulation A used for the in vivo study in Fig. 4. However, we could not use concentrations higher than 10 mg/kg BW with this ethanol/cremophor formulation A. The study in Fig. 5 was conducted at higher concentrations of MBQ-167 by oral using the Methyl cellulose/Tween 80 formulation B. Administration of formulation B at 100 mg/kg BW MBQ-167 by oral resulted in a 1 hour time to peak (T_max) with a half-life (T_1/2) of 4.6 hours (Fig. 6C).

We reported that the plasma drug concentration of MBQ-167 (in formulation A) after oral administration of 10 mg/kg BW MBQ-167 is approximately 4.5 times lower compared with the same dosage given by intraperitoneal administration (13). Therefore, the lower percentage of TNBC tumor growth inhibition after oral gavage (Fig. 5) may be due to the lower bioavailability of MBQ-167 via this route of administration with the methyl cellulose/Tween 80 formulation B. Figure 6C shows the plasma concentration of MBQ-167 following oral treatment at 100 mg/kg BW in BALB/c mice. Compared with the 10 mg/kg BW ethanol/Cremophor formulation A, which gave a C_max of MBQ-167 of 619 ng/mL, the 100 mg/kg BW methyl cellulose/Tween formulation B resulted in a lower C_max of 377 ng/mL MBQ-167 after oral administration. The T_max was 0.5 hours for oral administration of 10 mg/kg BW of formulation A, whereas for 100 mg/kg BW formulation B, the time to peak was 1.04 hours. For 10 mg/kg BW formulation A, the T_1/2 is 2.6 hours, while it is 4.6 hours for 100 mg/kg BW formulation B. Systemic drug exposure as measured by the area under the curve (AUC_0-t) was 585 ng·hour/mL for 10 mg/kg BW formulation A, and 2,938 ng·hour/mL for 100 mg/kg BW formulation B after oral administration of MBQ-167, respectively (Fig. 6C; Supplementary Table S1). Therefore, the higher dose of MBQ-167 in formulation B showed an increase in systematic drug exposure compared with the lower formulation A dose, which may explain the observed significant inhibition of lung metastasis following 50 and 100 mg/kg BW MBQ-167 but not at lower doses.

TNBC is a highly fatal form of breast cancer characterized by local invasion and distant metastases, which drastically reduce survival rates. Very few targeted therapies exist for TNBC, due to the absence of ER, PR, and HER2 expression (21). A number of clinical trials are ongoing for several growth factor receptor and PI3-K/Akt/mTOR targeted therapeutics that are upregulated in TNBC (22), as well as immune checkpoint therapies (23); however, their success has been confounded by the heterogenous nature of TNBC, and the acquisition of therapy resistance (2–6, 17–19). Herein, we describe the potential use of MBQ-167, a Rac/Cdc42-targeted therapeutic, which is in preparation for the submission of an investigative new drug application to the FDA, as a putative TNBC therapeutic agent.

Aberrations in Rac and Cdc42 activity are not often due to mutations but rather due to overexpression or overactivation by oncogenic GEFs in most aggressive cancers, with the exception of the P29S mutation in melanoma and the constitutively active Rac1B splice variant (10). From TCGA data, the breast METABARIC study reported 1.5% Rac1 amplifications in primary breast tumors (24), while Rac1 is amplified in approximately 4% of metastatic breast cancers from the Metastatic Breast Cancer Project (CBioPortal); therefore, Rac1 upregulation in breast cancer may be associated with metastatic potential. Moreover, a study that analyzed Rac1 expression in breast cancer reported that Rac1 is particularly upregulated in ER-negative breast tumors and may represent up to 20% in basal breast cancers indicating that targeting Rac1 is a viable therapeutic strategy for TNBC (25). Because the upstream effectors of Rac/Cdc42 such as growth factor receptors (EGFR, VEGFR) and PI3-K activate Rac/Cdc42 via their GEFs, which can also be oncogenic, inhibition of Rac/Cdc42 can impede signaling from multiple upstream effectors that are upregulated in TNBC.

Previously, we reported that the small molecule compound MBQ-167 targets Rac/Cdc42 activation to inhibit actin cytoskeletal extensions, migration, viability, proliferation, cell-cycle progression, and induce apoptosis in HER2 (+) and MDA-MB-231 TNBC cells (12). In this study, we focused on TNBC cell lines and show that MBQ-167 also reduces clonogenecity in MDA-MB-231 and viability in MDA-MB-468 and 4T-1 breast TNBC cell lines. In addition, we show that MBQ-167 effectively inhibits Rac and Cdc42 activation by GTP loading in all TNBC cell lines tested, demonstrating that MBQ-167 targets Rac and Cdc42.

Our previous study demonstrated that at 24 hours, MBQ-167 at 250 nmol/L reduced the activation of the Rac/Cdc42 direct downstream effectors PAK1 and PAK2 by decreased phosphorylation and showed that reduced PAK activity resulted in decreased phosphorylation of its downstream effectors LIM kinase and cofilin, as well as STAT3. However, we did not see an effect of MBQ-167 on p38 or p42/44 MAP kinases or Akt activation (12). Herein, phosphokinome arrays were performed with the MDA-MB-231 TNBC cell line to elucidate potential off-target effects of MBQ-167. Data show that MBQ-167 did not affect other kinases at a cutoff of −1.5-fold or +1.5-fold but increased the phosphorylation of c-Jun and CREB transcription factors by >1.5-fold. This observation was confirmed via Western blotting for phospho and total forms of c-Jun and CREB using mono-specific antibodies. However, we also show that sustained MBQ-167 administration (>48 hours) or higher dosage decreases this transient activation of c-Jun and CREB and expression of their transcriptional targets. It should also be noted that MBQ-167 induces a loss in polarity in MDA-MB-231 cells followed by apoptosis to undergo anoikis by 96–120 hours. Moreover, we reported that 250 nmol/L MBQ-167 resulted in decreased expression of pro-survival proteins Bcl-XL, Bcl-2, and MCL-1 that are also transcriptional targets of of c-Jun and CREB, especially in the detached breast cancer cell population (12).

Consequently, data presented herein show that the initial increases in phospho c-Jun and phospho-CREB at 24 hours following 200 nmol/L MBQ-167 were not sustained with prolonged treatment. The major kinase that phosphorylates c-Jun on Ser63 and Ser73, Jun kinase [JNK, stress activated protein kinase (SAPK)] are activated by cytokines and stress to result in increased AP-1 (c-Jun and c-Fos) transactivation of pro-cancer genes (26). However, expression or phospho status of JNK was not changed by MBQ-167 treatment. c-Jun can be activated by kinases other than JNK, such as P42/44 MAP kinases and cyclin dependent kinases (27, 28). Nevertheless, our studies clearly show that p42/44 MAP kinases are not activated by MBQ-167. Similarly, while CREB is phosphorylated by a number of kinases including protein kinase A (PKA), extracellular regulated kinase (ERK), p90 ribosomal S6 kinase (p90RSK), MAPK and SAPK 1 (MSK1), and glycogen synthase kinase 3β (GSK3β) (29); ERK1/2, p90RSK1/2/3, MSK1/2, and GSK3α/β levels remained unchanged by MBQ-167. Even though we did not see changes in phoshpholipase Cγ (PLCγ) phosphorylation in the phosphokinome array conducted with combined attached and detached cells (N = 4), an increase in phospho PLCγ was observed in the separated attached cell population, which may result in increased cAMP to activate the observed CREB phosphorylation via PKA activity.

To determine whether this activation of c-Jun and CREB were present in both the attached cells and the subpopulation of TNBC cells that initially detach from the substrate to ultimately undergo anoikis, we separated the cell populations and performed Western blotting for c-Jun and CREB transcriptional targets that contribute to cancer progression (29, 30). Notably, ZEB1, Cyclin D1, P-c-Jun, and c-Jun expression was reduced by 200 nmol/L MBQ-167 at >48 hours treatment, especially in the detached cell population. Survivin, a CREB target, was not affected by MBQ-167 at all times tested, indicating additional mechanisms of regulation for this survival factor, which is under complex regulation (31). Because all cells eventually detach and die in response to prolonged MBQ-167, we conclude that the initial increases in c-Jun and CREB-mediated transcriptional activation of pro-survival factors are not sustained over prolonged MBQ-167 treatment. The most likely explanation for this early elevation of phospho-c-Jun and phospho-CREB levels in response to low levels of MBQ-167 is that it is a response to the stress due to inhibition of Rac and Cdc42 signaling, which leads to temporary activation of these growth-related transcription factors as a final effort to survive before ultimately undergoing anoikis. Accordingly, we were unable to isolate MBQ-167–resistant variants, due to total cell death during the first week of selection. Therefore, we posit that the increase in active forms of these critical pro-cancer transcription factors is due to initial resistance mechanisms, which are overcome by prolonged MBQ-167 treatments.

The mechanism by which MBQ-167 ultimately induces cell death was attributed to inhibition of the Cdc42-regulated cell polarity and cell-cycle arrest at G₂–M stage, and the loss of Rac/Cdc42 mediated integrin signaling to result in a dissolution of focal adhesions with the substratum. This dissociation ultimately leads to anoikis, with consequent decreases in caspase 3 activity, and Bcl2 family anti apoptotic proteins Bcl2, Bcl-XL, and MCL-1 expression following 48 hours in 250 nmol/L MBQ-167. We also previously demonstrated that MBQ-167 treatment inhibits Rac/Cdc42/PAK signaling to the actin cytoskeleton by reducing activated (phospho) PAK, and thus, the phosphorylation of LIM kinase and the actin binding protein cofilin (12). Studies are underway to identify specific Rac/Cdc42 interactors that are inhibited by MBQ-167, similar to the studies conducted in (32, 33).

To demonstrate direct interaction of MBQ-167 with Rac1 in vitro, we used NMR. Our data support the physical interaction of MBQ -167 with Rac1 via residues Gly-12, Val-85, and Asn-92, which are located closely in a pocket region formed by an alpha-helix and a loop in the protein. This region of the protein shows a deep, hollow morphology where non-polar molecules, such as MBQ-167, can interact. Moreover, several amino acids from this region have been reported as altered in cancer. The highly conserved Gly-12 acts as a hotspot for oncogenic mutations in the Ras family by maintaining Rac1 in an active GTP-bound state. Also, the N92I point mutation at Asn-92 is a spontaneous oncogenic mutation in a wide array of cancer (34). Therefore, MBQ-167 may inhibit Rac and Cdc42 activation by binding to the nucleotide binding site and displacing GDP/GTP binding, similar to the mechanism described for EHT-1864, an established Rac inhibitor. Studies are underway to biochemically demonstrate the role of MBQ-167 as a nucleotide binding inhibitor, as described in ref. 35.

Our in vivo data further validates the efficacy of MBQ-167, where treatment of SCID mice bearing MDA-MB-231 mammary fat pad tumors, or BALB/c mice bearing 4T-1 mammary fat pad tumors, resulted in a statistically significant decrease in tumor growth and metastasis to the lung. In the slower growing MDA-MB-231 model in immunocompromised mice, we observed a drastic approximately 90% reduction in tumor growth and metastasis to the lung at all concentrations tested. In this experiment, the mice were treated by intraperitoneal with lower concentrations (1–10 mg/kg BW) of MBQ-167 3× a week for 3.5 months. Our data show that at these concentrations, a formulation of MBQ-167 in Cremaphor/Ethanol (A) was detected in plasma, tumor tissue, and other organs in sufficient levels to act as an effective inhibitor of Rac and Cdc42, and thus, tumor cell viability and migration/invasion (13). Accordingly, pharmacokinetic/pharmacodynamic modeling of these data predict that the observed inhibition of tumor growth in the MDA-MB-231 mammary fatpad tumors in mice following 1 or 10 mg/kg BW of MBQ-167 3× a week, result in an IC₅₀ of 0.1 nmol/L, and demonstrates a higher efficacy for inhibiting TNBC growth compared with HER2⁺ breast tumors (36).

We also found that at 10 mg/kg BW, MBQ-167 not only inhibits spontaneous metastasis from mammary tumors to the lung but also reduces experimental metastasis to the lung when directly introduced into the plasma. These data indicate that MBQ-167 inhibits intravasation from the primary tumor into the vasculature, as well as extravasation into the lung microenvironment and establishment of metastases. As expected, the efficiency of lung metastasis inhibition was less (60%) by experimental metastasis since the cancer cells introduced directly into the blood stream travel faster to the lung to establish metastases. The reduced metastasis inhibition by this method may also be explained by the highly aggressive nature of the 4T-1/BALB-c syngeneic, immunocompetent mouse breast cancer model.

The effect of MBQ-167 was also tested in a spontaneous metastasis assay in the 4T-1/BALB/c model, using oral administration with the Methyl cellulose/Tween-80 formulation B that enabled dosing at higher concentrations of MBQ-167. The formulation B was tested in toxicity/safety studies contracted to Charles River Labs and showed no toxic effects of MBQ-167 up to 4,000 mg/kg BW in rats. Likewise, 100 mg/kg BW MBQ-167 was administered orally 5× a week in our mouse experiments with no apparent toxicity. This novel formulation and dosage schedule of MBQ-167 resulted in a 4.6 hours half-life (∼2× longer than 10 mg/kg BW MBQ-167 by oral gavage with formulation A) and a 5× higher AUC in plasma, indicating prolonged exposure and a higher concentration of MBQ-167. Therefore, we expect effective quantities of MBQ-167 to reach the tumors via oral gavage of this formulation. At lower concentrations, tumor growth was similar to vehicle controls until approximately 15 days of treatment, after which MBQ-167 treatment resulted in a sharp decline in tumor growth. This may reflect the early resistant mechanisms that were observed in vitro with enhanced c-Jun and CREB transcriptional activities, which are overcome with prolonged treatment of MBQ-167. Moreover, in this immunocompetent mouse model, which is more complex than the previously used SCID mouse model, MBQ-167 may also affect immune cells which could possibly reduce the inhibitory action on tumor growth.

A concentration-dependent inhibitory effect was observed on lung metastasis inhibition from the spontaneous metastasis model in immunocompetent mice, where concentrations up to 25 mg/kg BW MBQ-167 had no effect on metastasis but higher concentrations of 50 or 100 mg/kg BW inhibited lung metastases by 80%. In contrast, 1–10 mg/kg BW MBQ-167 administered via intraperitoneal resulted in >80% inhibition of lung metastasis at all concentrations. Therefore, this observation may be due to a slower rate of absorption of MBQ-167 into lung tissue by oral gavage compared with intraperitoneal.

In summary, the data presented show that MBQ-167 is a specific Rac/Cdc42 inhibitor, with minimal off-target effects and resistance mechanisms that are overcome by prolonged treatment. In mouse models of both experimental and spontaneous metastasis, MBQ-167 is an efficient inhibitor of TNBC tumor growth and lung metastasis, with effective quantities reaching the target tissues. Therefore, further development of MBQ-167 as a TNBC therapeutic is warranted.

H. Picón reports personal fees from MBQ Pharma during the conduct of the study. L.D. Borrero-Garcia reports other support from MBQ Pharma, Inc. during the conduct of the study; other support from MBQ Pharma Inc. outside the submitted work. E Hernandez-O-Farrill reports other support from MBQ Pharma, Inc. during the conduct of the study; in addition, E. Hernandez-O-Farrill has a patent for US 8,884,006 B2 issued, a patent for US 9,278,956 B1 issued, a patent for US 9,981,980 B2 issued and licensed to UPR MSC and MBQ Pharma, Inc., a patent for PCT/US2017/029921 issued, a patent for US 10,392,396 B2 issued, and a patent for US 10,729,689 B2 issued. C.P. Vlaar reports non-financial support from MBQ Pharma during the conduct of the study; other support from MBQ Pharma outside the submitted work; in addition, C.P. Vlaar has a patent for US9,981,980 issued and licensed to MBQ Pharma. S. Dharmawardhane reports non-financial support from MBQ Pharma, Inc. during the conduct of the study; in addition, S. Dharmawardhane has a patent for US9,981,980 licensed to UPR MSC, MBQ Pharma, Inc. and a patent for US10,392,396 licensed to UPR MSC, MBQ Pharma, Inc. No disclosures were reported by the other authors.

A. Cruz-Collazo: Conceptualization, resources, formal analysis, supervision, funding acquisition, investigation, methodology, writing–original draft, project administration, writing–review and editing. J.F. Ruiz-Calderon: Conceptualization, formal analysis, supervision, investigation, methodology, writing–review and editing. H. Picon: Conceptualization, software, formal analysis, supervision, investigation, methodology, writing–review and editing. L.D. Borrero-Garcia: Software, formal analysis, investigation, methodology, writing–review and editing. I. Lopez: Formal analysis, investigation, writing–review and editing. L. Castillo-Pichardo: Conceptualization, formal analysis, supervision, investigation, writing–review and editing. M. del Mar Maldonado: Conceptualization, software, formal analysis, supervision, investigation, writing–review and editing. J. Duconge: Software, formal analysis, investigation. J.I. Medina: Conceptualization, resources, software, formal analysis, supervision, funding acquisition, validation, investigation, methodology, writing–original draft. M.J. Bayro: Conceptualization, resources, formal analysis, supervision, funding acquisition, validation, investigation, methodology, writing–original draft, writing–review and editing. E. Hernández-O'Farrill: Conceptualization, resources, software, formal analysis, supervision, funding acquisition, investigation, writing–original draft, project administration, writing–review and editing. C.P. Vlaar: Conceptualization, resources, formal analysis, supervision, funding acquisition, validation, investigation, methodology, writing–original draft, writing–review and editing. S. Dharmawardhane: Conceptualization, resources, formal analysis, supervision, funding acquisition, validation, investigation, methodology, writing–original draft, project administration, writing–review and editing.

This study was partially funded by MBQ Pharma, Inc. Additional support was provided by NIH/NIGMS P20GM103475 to J.F. Ruiz-Calderon, UPR RCM NIH/NIMHHD R25GM061838 (to A. Cruz-Collazo, M. del Mar Maldonado, J. I. Medina), NIH/NIGMS SC3GM116713 (to C.P. Vlaar), NIH/NIGMS SC3GM084824, US Army Breast Cancer Research Program W81XWH2010041, NIH/NCI U54 CA096297, Susan Komen for the Cure OGI70023, and Puerto Rico Science, Technology, and Research Trust (PRSTRT) grants (to S. Dharmawardhane), and PRSTRT award 2020-00128 (to M.J. Bayro), and the UPR Institutional Funds for Research (FIPI) Program (to M.J. Bayro). We wish to thank Mariano de Socarraz, MD (Core Plus Labs) for the blood chemistry services.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.