Inhibition of mTOR signaling using the rapalog everolimus is an FDA-approved targeted therapy for patients with lung and gastroenteropancreatic neuroendocrine tumors (NET). However, patients eventually progress on treatment, highlighting the need for additional therapies. We focused on pancreatic NETs (pNET) and reasoned that treatment of these tumors upon progression on rapalog therapy, with an mTOR kinase inhibitor (mTORKi), such as CC-223, could overcome a number of resistance mechanisms in tumors and delay cardiac carcinoid disease. We performed preclinical studies using human pNET cells in vitro and injected them subcutaneously or orthotopically to determine tumor progression and cardiac function in mice treated with either rapamycin alone or switched to CC-223 upon progression. Detailed signaling and RNA sequencing analyses were performed on tumors that were sensitive or progressed on mTOR treatment. Approximately 57% of mice bearing pNET tumors that progressed on rapalog therapy showed a significant decrease in tumor volume upon a switch to CC-223. Moreover, mice treated with an mTORKi exhibited decreased cardiac dilation and thickening of heart valves than those treated with placebo or rapamycin alone. In conclusion, in the majority of pNETs that progress on rapalogs, it is possible to reduce disease progression using an mTORKi, such as CC-223. Moreover, CC-223 had an additional transient cardiac benefit on valvular fibrosis compared with placebo- or rapalog-treated mice. These results provide the preclinical rationale to further develop mTORKi clinically upon progression on rapalog therapy and to further test their long-term cardioprotective benefit in those NET patients prone to carcinoid syndrome. Mol Cancer Ther; 16(11); 2432–41. ©2017 AACR.

The incidence of pancreatic neuroendocrine tumors (pNETs) is increasing (1). These tumors vary in their clinical presentation, depending on whether they secrete hormones and have metastasized to the liver (2). At the genetic level, mutations in key players of the mTOR pathway have been identified in pNET patient tumors by whole-exome sequencing (3, 4). Downregulation of tumor suppressors of this pathway, including PTEN and tuberous sclerosis complex 2 (TSC2), leads to mTOR activation in a number of human pNETs (5). In addition, nutrients, amino acids and glucose, as well as hormones, insulin and IGF1, mediate mTOR activation. The clinical relevance of targeting mTOR signaling in these tumors is highlighted by the FDA approval of the allosteric mTOR inhibitor everolimus (RAD001), for the treatment of pNETs. The rapalog, RAD001, by targeting mTOR complex 1 (mTORC1) has both antiproliferative and antiangiogenic properties in tumors (6). In the phase III RAD001 in Advanced Neuroendocrine Tumors (RADIANT-3) clinical trial, treatment of pNET patients with RAD001 increased median progression-free survival by approximately 6.4 months compared with placebo, leading to its FDA approval for the treatment of pNET patients with progressive, unresectable, locally advanced, or metastatic disease (7). More recently, the follow-up phase III RADIANT-4 trial data reported that RAD001 significantly improved the median progression-free survival in patients with progressive gastrointestinal or lung NETs by approximately 7.1 months (8). Nonetheless, the response to treatment is not durable, most likely due to incomplete and unsustained inhibition of mTORC1 signaling by rapalogs, and/or activation of mTOR complex 2 (mTORC2), underscoring the need for alternative therapies and drug combinations (9).

Consistent with our studies in cell culture (10), analysis of patient biopsies treated with RAD001 suggested that activation of protein kinase B (PKB referred to as Akt), due to loss of the mTOR/40S ribosomal protein S6 kinase 1 (S6K1) negative feedback loop, could contribute to tumor progression (11). Irrespective of treatment response, S6K1-mediated phosphorylation (p-) of ribosomal protein S6, an indicator of mTORC1 activation, was significantly decreased in matched neuroendocrine tumor patient biopsies after 2 weeks of RAD001/octreotide treatment (12), suggestive of effective target inhibition. However, RAD001 affects substrate specificity not kinase activity. Indeed, RAD001 can abolish S6K1 signaling, while having little impact on other mTORC1 substrates such as translation initiation factor 4E (eIF4E) binding protein 1 (4E-BP1) and unc-51 like kinase 1 (ULK1); (refs. 10, 13). Thus, incomplete inhibition of mTORC1 substrates and activation of survival effector Akt have the potential to lead to drug resistance. This resistance could potentially be overcome by PI3K and/or ATP-site–competitive mTOR inhibitors, which have an added therapeutic advantage in solid tumors as they inhibit the catalytic activity of both mTOR complexes (14–16). However, there is a lack of systematic evidence demonstrating in vivo whether a switch to an mTOR kinase inhibitor (mTORKi) will be beneficial upon progression on rapalog therapy in pNETs.

Preclinical mouse models for pNETs have been valuable as predictors of clinical responses to drug combinations and targeted therapies (17–20). More recently, nude mice injected with BON cells, which were derived from a lymph node metastasis of a human pancreatic tumor (21), not only form primary tumors and metastases to the liver and abdominal lymph nodes (20), but also exhibit carcinoid syndrome. This includes cardiac carcinoid disease that phenotypically recapitulates the human disease (22). The cardiac impairment and fibrosis that develops in some pNET patients ultimately leads to congestive heart failure (23). Although increased serotonin levels secreted by the tumor have been postulated to play a role in cardiac carcinoid disease, the detailed molecular mechanism remains unclear (24).

On the basis of these premises and using in vivo mouse models, we investigated whether mTOR kinase inhibition provided benefit over rapalog treatment in pNETs that progress on rapalog therapy. Moreover, we determined whether the benefit from inhibiting mTOR signaling would extend beyond inhibiting tumor progression to the heart, as it has been reported to be adversely impacted by the growth of the primary tumor and its metastases.

Cell culture

BON cells were a kind gift from Dr. Courtney Townsend (UTMB, Galveston, Texas) received in July, 2013. QGP1 cells were purchased in February, 2015 from the Japanese Collection of Research Bioresources Cell Bank. Both cell lines were expanded and frozen into multiple low passage aliquots. BON and QGP1 cells were passaged using DMEM: F12K media or RPMI respectively, routinely tested for mycoplasma contamination and not maintained in culture for more than 1.5 months after thawing of a frozen vial. Expression of NET markers, including chromogranin A and synaptophysin, were confirmed by Western blotting for both cell lines. All media contained 10% FBS and penicillin/streptomycin antibiotics. Firefly luciferase virus obtained from GenTarget (LVP-326) was used to infect BON cells followed by selection of stable pool of cells using blasticidin (Life Technologies). TAK-228, rapamycin, and RAD001 were obtained from LC Laboratories and AZD2014 from Selleckchem.

Mouse experiments

All in vivo studies were performed following procedures approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Cincinnati (Cincinnati, OH). Five-week-old nude male mice (J:NU stock 007850) were obtained from Jackson Laboratories and housed at a barrier facility in our institution with ad libitum access to food and water. Mice were injected with 2 × 106 cells subcutaneously or 1 × 106 cells in 20 μL orthotopically into the pancreas (25). Subcutaneous tumor measurements were done using digital calipers and the volume was calculated using the formula (π/6 × length × width2). Rapamycin was administered in drinking water at [10 mg/kg body weight (BW) daily]. The water was refreshed every other day. Progression on rapamycin was defined according to a modified Response Evaluation Criteria in Solid Tumor (mRECIST) by an increase of at least 20% in diameter in the majority of the mice. Mice were taken off rapamycin for 24 hours prior to the switch to CC-223 to allow time for the rapamycin to be cleared to avoid dosage with the combination of both drugs. CC-223, a generous gift from Celgene was gavaged twice daily at 15 mg/kg BW. The percent change in tumor volume from baseline was calculated on the basis of the formula: [100 × (tumor volume at time of sacrifice – tumor volume at time of switch)/tumor volume at time of switch]. Mice were starved for 3 hours prior to the time of euthanasia at which point all tumors and tissues collected were snap-frozen using liquid nitrogen and/or fixed in formalin. Serum serotonin levels were measured using an ELISA assay (IBL-America) following the manufacturer's protocol.

Imaging

Mice were injected with potassium luciferin (Gold Biotechnology) and 10 minutes later were anesthetized using isoflurane. Bioluminescence imaging was performed using a 2D Carestream (Kodak) MultiSpectral FX (2D).

Echocardiographic studies

Mice were anesthetized with isoflurane and cardiac function and structure was measured via high frequency echocardiography with M-mode imaging (26), and valvular function with B-mode and color Doppler imaging (27) as previously described using a Vevo2100 ultrasound system. Postprocessing of the images was done using the VevoStrain software (Visualsonic, Vevo 2100, v1.1.1 B1455) by an investigator blinded to the treatment groups.

IHC/Modified Masson trichrome staining and scoring of valve thickness

The heart was trimmed longitudinally through both ventricles after formalin fixation. The tissue was then sectioned (at a thickness of 5 μm) until the valves were visible. At which point, 20 μm was skipped and another section was saved until the tissue was sectioned throughout to get an even representation throughout the whole valve. These sections were then stained with a modified Masson trichrome staining kit (ScytTek Laboratories) following the manufacturer's protocol. The stained slides were scored for cardiac valvular and endocardial histopathologic lesions by a veterinary pathologist blinded to the treatment groups.

Western blot analyses

Tumors were extracted as previously described except for the lysis buffer contained 1% NP-40 (10). Quantitative Western blot analyses were performed using a LI-COR Odyssey imaging system. All total and phosphorylated antibodies were obtained from Cell Signaling Technology except for chromogranin A which was from Thermo Fisher Scientific.

RNA sequencing

RNA-seq of tumors was performed at the UC Genomics, Epigenomics and Sequencing Core following a standardized method. Briefly, RNA was isolated from each approximately 50 mg of each tumor sample using the mirVana miRNA Isolation Kit (Thermo Fisher Scientific) according to the manufacturer's protocol. Target RNA enrichment was performed from approximately 750 ng of RNA by isolation of polyA RNA using NEBNext Poly(A) mRNA Magnetic Isolation Module (New England BioLabs). PrepX PolyA script was run for automated polyA RNA isolation. RNA-seq library preparation was done using NEBNext Ultra Directional RNA Library Prep Kit (New England BioLabs) with dUTP for cDNA synthesis to maintain strand specificity, indexing via 11 cycles of PCR enrichment and finally cleaning up of the amplified library by AMPure XP beads for quality control analysis. The quality and yield of the purified library was analyzed with a Bioanalyzer (Agilent) and a DNA high sensitivity chip; then the concentration was quantified by qPCR using the Kapa Library Quantification kit (Kapa Biosystem). Individually indexed and compatible libraries were proportionally pooled (∼25 × 106 million reads per sample) for clustering with a cBot system (Illumina). Libraries at the final concentration of 15 pmol/L were clustered onto a single read flow cell using Illumina TruSeq SR Cluster kit v3, and sequenced to 50 bp using TruSeq SBS kit on an Illumina HiSeq system. The laboratory for Statistical Genomics and Systems Biology at University of Cincinnati (Cincinnati, OH) performed the bioinformatics analyses by aligning the sequence reads to the genome using a standard Illumina sequence analysis pipeline. A list of significantly different genes between the mTOR inhibitors treatment groups was generated by a multivariate regression analysis with a significant P ≤0.006. Comparative networks and functional analyses were performed using the NetWalker software (28). The GEO accession number for the RNA-seq data is GSE102246.

Statistical analyses

Statistical analyses were performed using the GraphPad Prism 5 Software. Unless otherwise specified in the figure legends, data are presented as means ± SEM. Statistical significance was determined by a one-way ANOVA followed with a Bonferroni multiple group comparison test when applicable unless otherwise stated in the figure legend.

Rapalog treatment using rapamycin or its analogue RAD001 has been reported to delay tumor growth of BON cells subcutaneously injected into mice (12). To follow tumor progression in the mouse, we generated BON cells (29) that stably express luciferase (BON-luci) for in vivo bioluminescence imaging (BLI). In addition, to recapitulate the tumor microenvironment for a pNET, the BON-luci cells were injected into the pancreas. To our knowledge, we are the first to test the preclinical efficacy of mTOR inhibitors on BON cells or human pNET cell lines orthotopically rather than subcutaneously (12, 30) or intrasplenically (22). Mice were randomized into 2 groups of relatively equal tumor volume based on BLI measurements at approximately 5 weeks postimplantation of the cells, and treated either with placebo or rapamycin. Rapamycin treatment decreased tumor burden significantly compared with placebo (Fig. 1A and B) after 25 days, albeit the tumors were still progressing. As previously reported for subcutaneously injected BON tumors, orthotopic tumors of rapamycin-treated mice had a significant decrease in p-S6, no inhibition of 4E-BP1 and activation of p-Akt compared with placebo-treated mouse tumors (Fig. 1C).

Figure 1.

Mice-bearing orthotopically injected BON tumors had a significant decrease in tumor burden when treated with rapamycin compared with placebo. A, Representative BLI images of mice treated with placebo or rapamycin at the indicated times. All images were obtained with 1-minute exposure. B, Mean tumor weight (±SEM) at the time of sacrifice of mice treated with placebo (n = 9) and rapamycin (n = 11) for 25 days. *, P < 0.05 using a t test. C, Western blot analyses of tumor lysates with each lane representing a mouse tumor. Actin was used as a loading control.

Figure 1.

Mice-bearing orthotopically injected BON tumors had a significant decrease in tumor burden when treated with rapamycin compared with placebo. A, Representative BLI images of mice treated with placebo or rapamycin at the indicated times. All images were obtained with 1-minute exposure. B, Mean tumor weight (±SEM) at the time of sacrifice of mice treated with placebo (n = 9) and rapamycin (n = 11) for 25 days. *, P < 0.05 using a t test. C, Western blot analyses of tumor lysates with each lane representing a mouse tumor. Actin was used as a loading control.

Close modal

Given this incomplete suppression of mTOR signaling in rapamycin-treated BON tumors, we reasoned that treatment of BON tumors upon progression on rapalog therapy with an mTORKi would overcome a number of resistant mechanisms such as those mediated by activation of Akt or 4E-BP1; owing to their more potent inhibition of these and other mTOR substrates. Indeed, treatment of BON cells in vitro with clinically relevant mTORKi(s) including CC-223, MLN0128 (also known as TAK-228, sapanisertib or INK128) and AZD2014 (vistusertib) decreased phosphorylation of a number or mTOR substrates, including p-4E-BP1, p-Akt and p-ULK1 in a dose-dependent manner compared with vehicle or rapalogues (Fig. 2A; Supplementary Fig. S1; refs. 31–33). These mTORKis were equally effective in inhibiting mTORC1/2 signaling in QGP1 cells, another human pNET cell line (Fig. 2B; Supplementary Fig. S1). To test our hypothesis in vivo, we injected mice with BON-luci cells, subcutaneously, to allow for quantitative volumetric monitoring of tumor growth. Once tumor volumes averaged approximately 180 mm3, mice were randomized into 2 groups of equal tumor volume for treatment with either placebo or rapamycin (Fig. 3A). Rapamycin treatment resulted in a significant decrease in tumor volume compared with placebo treatment, although tumor progression continued similar to our findings with the orthotopically injected tumors. Once the tumors progressed on rapalogue therapy according to a modified RECIST criteria, these mice were divided into 2 cohorts of similar tumor distribution and either maintained on rapamycin or switched to CC-223. Placebo-treated mice were euthanized on day 80 as not to exceed the tumor volume limited by IACUC. Tumor progression in mice switched to CC-223 was decreased compared with rapamycin (Fig. 3B). More importantly, a waterfall plot of the percent change in tumor volume from the time of switch to CC-223 of each mouse revealed that while some mice responded to the switch to CC-223 others continued to progress (Fig. 3C). Mice in the CC-223-treatment group were further separated into two subgroups termed “CC-223-responders” based on the clinical criteria of either stable disease or a better response rate to treatment; while mice with a change from baseline greater than 20% were grouped as “CC-223 nonresponders” (Fig S2).

Figure 2.

mTORKis result in more potent inhibition of mTOR signaling compared to rapalogs in a dose-dependent manner in the human pNET cell lines (A) BON and (B) QGP1 cells. Cells were treated with vehicle, RAD001 (20 nmol/L), rapamycin (50 nmol/L), or mTORKis (100, 200, 400, and 800nmol/L) for 2 hours prior to harvesting them for signaling analysis by Western blotting.

Figure 2.

mTORKis result in more potent inhibition of mTOR signaling compared to rapalogs in a dose-dependent manner in the human pNET cell lines (A) BON and (B) QGP1 cells. Cells were treated with vehicle, RAD001 (20 nmol/L), rapamycin (50 nmol/L), or mTORKis (100, 200, 400, and 800nmol/L) for 2 hours prior to harvesting them for signaling analysis by Western blotting.

Close modal
Figure 3.

A switch to CC-223 resulted in significant decrease in tumor burden in a subset of BON tumors that progressed on rapalog therapy. A, Schematic diagram of timeline [day (D)] for subcutaneous injection of cells and drug treatments. B, Changes in tumor volume (mean ± SEM) over time as a result of treatment with placebo (n = 8), rapamycin (n = 15), and CC-223 (n = 15). There were two deaths during the course of the experiment: 1 rapamycin-treated mouse died due to a technical gavage error and 1 CC-223–treated mouse that initially developed diarrhea on rapamycin treatment and had to be euthanized for persistent symptoms. C, Percent change in tumor volume from baseline (day 78) until the time of euthanasia (day 94) of mice treated with rapamycin or CC-223.

Figure 3.

A switch to CC-223 resulted in significant decrease in tumor burden in a subset of BON tumors that progressed on rapalog therapy. A, Schematic diagram of timeline [day (D)] for subcutaneous injection of cells and drug treatments. B, Changes in tumor volume (mean ± SEM) over time as a result of treatment with placebo (n = 8), rapamycin (n = 15), and CC-223 (n = 15). There were two deaths during the course of the experiment: 1 rapamycin-treated mouse died due to a technical gavage error and 1 CC-223–treated mouse that initially developed diarrhea on rapamycin treatment and had to be euthanized for persistent symptoms. C, Percent change in tumor volume from baseline (day 78) until the time of euthanasia (day 94) of mice treated with rapamycin or CC-223.

Close modal

To gain insight into the mechanism by which CC-223 treatment led to inhibition of tumor progression in some mice and not others, we performed signaling analyses on tumors from all treatment groups to determine whether the extent of inhibition of mTOR signaling would correlate with the level of tumor response. Overall, mTOR substrate inhibition of rapamycin-treated tumors compared with placebo-treated tumors was equivalent between the subcutaneously injected tumors and orthotopic ones (Figs. 1C and 4). CC-223 treatment induced a more pronounced inhibition of p-4E-BP1, p-Akt S473, and p-ULK1 S758 than rapamycin-treated tumors (Fig. 4). Moreover, although there was a modest increase in phosphorylation of extracellular signal-regulated kinase (ERK) at T202/Y204 in CC-223–treated tumors when compared with placebo, there was no apparent correlation between MAPK/ERK signaling and tumor response (Fig. 4).

Figure 4.

Compared with placebo- or rapamycin-treated tumors, CC-223–treated mice had a more complete inhibition of mTOR targets irrespective of tumor response. Western blot analyses of tumor lysates. Each lane was loaded with lysates from individual mouse tumors. The green and red bars indicate, respectively, tumors that either responded to treatment or did not respond based on mRECIST criteria. Total actin was used as a loading control.

Figure 4.

Compared with placebo- or rapamycin-treated tumors, CC-223–treated mice had a more complete inhibition of mTOR targets irrespective of tumor response. Western blot analyses of tumor lysates. Each lane was loaded with lysates from individual mouse tumors. The green and red bars indicate, respectively, tumors that either responded to treatment or did not respond based on mRECIST criteria. Total actin was used as a loading control.

Close modal

As inhibition of these selected targets did not explain the differences in response to CC-223, we performed RNA sequencing (RNA-seq) analyses as an unbiased means to assess changes in gene expression. Our major goal was to identify the differences in tumor mRNAs between the CC-223- and non-CC-223 responders compared with the rapamycin alone treatment arm (Fig. 5A). A total of 477 genes were significantly differentially expressed between CC-223 and rapamycin-treated tumors and clustered into 3 groups highlighted in the heat map (Fig. 5A). Interestingly, RNA-seq analyses revealed that in comparison with rapamycin and the CC-223 responders, the expression of ribosomal proteins and other genes involved in translation, were increased in a number of CC-223 nonresponders (cluster 1) (Fig. 5A and B), consistent with the accelerated growth of these tumors. Indeed, Western blot analyses revealed a modest increase in levels of ribosomal proteins (RP) L5 and S19 and eIF4E in a number of tumors of CC-223 nonresponders compared with CC-223 responders (Fig. 5C). Tumors are dependent on protein synthesis to meet high demands of growth and proliferation. It is not surprising that inhibition of mTOR signaling, a pathway implicated in regulating translation at multiple signaling and transcriptional nodes would alter the translational landscape within tumors (32). Nonetheless, it is interesting that transcription of certain proteins involved in translation was upregulated in a number of CC-223 nonresponders tumors under mTOR inhibition. As translation is a major energy consuming process in cells (34), we determined the activation state of AMP-activated protein kinase (AMPK) a major energy sensing kinase (35). Overall, the level of p-AMPK in mTORKi-treated tumors appeared to inversely correlate with the tumor response rate and upregulation of some of the translational machinery.

Figure 5.

Distinct significant changes occurred at the genetic and signaling levels of CC-223 responders and nonresponders compared with rapamycin-treated tumors. A, Hierarchical clustering of genes in rapamycin-treated tumors and tumors of CC-223 responders (green line), or nonresponders (red line). Each row represents an individual tumor. B, A network diagram of translation proteins which were significantly upregulated in tumors from CC-223 nonresponders mice compared with tumors from rapamycin-treated mice obtained using the NetWalker software. Blue and black lines indicate reactome and protein–protein interactions respectively. C, Western blot analyses of tumor lysates (same as Fig. 4).

Figure 5.

Distinct significant changes occurred at the genetic and signaling levels of CC-223 responders and nonresponders compared with rapamycin-treated tumors. A, Hierarchical clustering of genes in rapamycin-treated tumors and tumors of CC-223 responders (green line), or nonresponders (red line). Each row represents an individual tumor. B, A network diagram of translation proteins which were significantly upregulated in tumors from CC-223 nonresponders mice compared with tumors from rapamycin-treated mice obtained using the NetWalker software. Blue and black lines indicate reactome and protein–protein interactions respectively. C, Western blot analyses of tumor lysates (same as Fig. 4).

Close modal

Next, we set out to determine whether mTOR inhibition had additional benefits beyond the direct effects on tumor progression; focusing on cardiac carcinoid disease which has been linked to decreased survival in pNET patients (36). Interestingly, the RNA-seq data revealed that transcript levels of L-DOPA decarboxylase (DDC), an enzyme involved in the synthesis of serotonin, was significantly decreased in tumors switched to CC-223 compared with rapamycin-treated tumors (Supplementary Fig. 3A). Consistent with the transcriptional downregulation of DDC, protein levels of DDC were also decreased in CC-223–treated tumors irrespective of the response rate (Fig. 5C). To address the relative impact of mTOR inhibitors on cardiac impairment, we implanted BON-luci cells into the pancreas; as this better recapitulates human tumor growth at the primary site with occurrence of metastases to the liver and other organs upon disease progression. Two weeks after orthotopic injection, the mice were randomized into 2 groups of comparable tumor burden, based on BLI measurements, to be treated with either placebo or rapamycin for 22 days (Fig. 6A). On day 24, rapamycin-treated mice were divided into 2 groups of relatively equal tumor burden, again based on BLI, such that one group remained on rapamycin while the other was switched to CC-223 for approximately 26 days (Fig. 6A; Supplementary Fig. S3B). Cardiac function and structure were measured by high frequency echocardiography with M-mode imaging (26), and valvular function with B-mode and color Doppler imaging (27) on day 34 (D10 post-switch to CC-223) (Fig. 6B) and ∼2 days prior to sacrifice on D48 (D24 post-switch to CC-223) (Fig. 6C). Both CC-223 and rapamycin-treated mice had a significant improvement in left ventricular dilation (diastolic volume) compared to placebo within ∼10 days of the switch to CC-223 (Fig. 6B) but this conferred cardiac protection was not statistically significant within ∼24 days of the switch to CC-223 (Fig. 6C). Systolic cardiac function (as measured via ejection fraction) was not significantly different among the groups at either time point. However, diastolic function (as measured by E/E') was slightly better in mice switched to CC-223 treatment than those remaining on rapamycin at D34 (D10 postswitch; Fig. 6B), but was less apparent by day 48 (day 24 postswitch to CC-223; Fig. 6C).

Figure 6.

Mice switched to CC-223 treatment developed less histopathologic lesions of their heart valves and fibrosis than in rapamycin or placebo treatment. A, Schematic diagram of timeline for orthotopic injection of cells and drug treatments. B and C, Mean diastolic volume and E/E' (±SEM) on approximately day 10 and day 24, respectively, from the switch to CC-223. *, P < 0.001. D, Representative images of the modified Masson trichrome staining of the heart. The arrowheads point to the tricuspid valve leaflets with the blue color staining collagen. E, Percent thickened valves based on scoring of cardiac valvular histopathologic lesions by a pathologist blinded to the treatment groups (detailed in Supplementary Table S1). F, Model for potential development of resistance that could occur in pNETs leading to increased tumor growth, valvular fibrosis, and cardiac impairment.

Figure 6.

Mice switched to CC-223 treatment developed less histopathologic lesions of their heart valves and fibrosis than in rapamycin or placebo treatment. A, Schematic diagram of timeline for orthotopic injection of cells and drug treatments. B and C, Mean diastolic volume and E/E' (±SEM) on approximately day 10 and day 24, respectively, from the switch to CC-223. *, P < 0.001. D, Representative images of the modified Masson trichrome staining of the heart. The arrowheads point to the tricuspid valve leaflets with the blue color staining collagen. E, Percent thickened valves based on scoring of cardiac valvular histopathologic lesions by a pathologist blinded to the treatment groups (detailed in Supplementary Table S1). F, Model for potential development of resistance that could occur in pNETs leading to increased tumor growth, valvular fibrosis, and cardiac impairment.

Close modal

Importantly, histopathologic blind scoring of trichrome-stained consecutive heart sections showed that only mice that were switched to CC-223 treatment had less histopathologic lesions of their heart valves (especially the mitral and tricuspid ones) as compared with either placebo- or rapamycin-treated mice (Fig. 6D and E; Supplementary Table S1). Although rapamycin-treated mice or those switched to CC-223 had a significant decrease in tumor burden compared with placebo (Supplementary Fig. S3C), there was no difference in tumor sizes between the 2 mTOR inhibitors that could account for the difference in valvular pathology. A smaller number of histopathologic valve lesions in the hearts of mice switched to CC-223 was associated with decreased expression of DDC in the tumor (Supplementary Fig S3D). Although it was possible that DDC levels might influence serotonin synthesis, the expression of the rate-limiting serotonin synthesis enzyme tryptophan hydroxylase 1 (TPH1) was unchanged (Supplementary Fig S3D), and serum serotonin levels were not significantly different between any of the treatment groups (Supplementary Fig S3E). Together, these results indicate that mTOR inhibitors can potentially delay progression of cardiac impairment and that the sustained efficacy of mTORKi(s) may be more prolonged than that of rapalogs.

As RAD001 is FDA-approved for the treatment of pNETs, and a number of mTORKi(s) are currently in phase 1–2 clinical trials, it is translationally relevant to determine whether an mTORKi would serve as next line therapy to RAD001 clinically. Our data suggest that switching to an mTORKi, here CC-223, in a subset of mice that progress on rapalog therapy, provides potential clinical benefits. Recently published data from a clinical trial using the dual mTOR/PI3K inhibitor BEZ235 to treat pNET patients who had progressed on an mTOR inhibitor reported at least stabilization of disease in greater than 50% of the patients albeit with numerous challenges (37). The latter consisted of multiple adverse effects, treatment interruptions, and poor pharmacokinetics of BEZ235. A number of issues have dampened the initial excitement in the clinical development of PI3K and/or mTOR kinase inhibitors; ranging from higher toxicity profiles compared with rapalogs, to clinical testing of inhibitors with suboptimal formulations, pharmacokinetic, and pharmacodynamic properties. Nevertheless, a number of mTORKi(s) with improved pharmacokinetic/pharmacodynamic are being actively pursued in the clinic. One of these, TAK-228 is currently in a phase II clinical trial for pNETs. Moreover, a third generation of mTOR inhibitors has been developed which could be best described as a hybrid of the first two generations of inhibitors (38) with recently reported improved efficacy compared to rapalogs (39). In summary, based on our preclinical data, treatment of pNETs with a rapalog until progression then switching to an mTORKi was associated with tumor responses in the majority of the cases.

Activation of the MAPK/ERK pathway may signal the beginning of acquired resistance to mTOR inhibitors and has been reported in biopsied patient tumors after RAD001 treatment (40). We found that p-ERK modestly increased in tumors of mice switched to CC-223 compared with control tumors, irrespective of tumor response. This is consistent with previous reports showing activation of MAPK following treatment of BON cells with BEZ235 (30). Future studies will be needed to determine whether MAPK/ERK activation could be clinically exploited upon development of resistance to mTORKi.

A number of additional resistance mechanisms to rapalogs have been reported to occur both in tumors extrinsically as well as in a tumor cell–intrinsic manner (41). Incomplete inhibition of p-4E-BP1 is one potential mechanism of resistance to rapalogs that can be overcome by mTORKi(s). In the context of mTORKi treatment, the hypophosphorylated 4E-BP1 sequesters eIF4E, thus preventing the assembly of a productive eIF4F translation initiation complex (42). Availability of eIF4E is rate limiting for translation, and its overexpression is sufficient to induce both transformation and tumor formation (43, 44). Interestingly, a number of CC-223 nonresponders had increased levels of total eIF4E and ribosomal proteins compared with CC-223-responders in addition to the decreased AMPK activity, consistent with upregulation of the translational machinery and anabolic processes, which actively promote tumor growth (Fig. 6F).

As a secondary endpoint, our study also tested the potential cardiac benefits provided by mTOR inhibitors, as the heart is one of the organs damaged by growth of NET tumors. Earlier literature reported more than half (66%–∼77%) of NET patients developed cardiac carcinoid disease (24, 36, 45). However more recent data suggest that this percentage is lower, potentially due to earlier diagnosis and treatment with somatostatin analogues (46). Once valvular carcinoid disease develops, the affected valve can be replaced with either a mechanical or tissue valve of bovine or porcine origin. The disadvantage of a tissue valve is that it also may become damaged due to the underlying carcinoid disease. Our results suggest that mTOR inhibitors decrease progression of cardiac impairment due to valve thickening. Although serum serotonin mean levels were lower in both rapamycin- and CC-223–treated mice than in placebo-treated mice at the time of sacrifice; the differences were not significant. This suggests that either (i) the levels of serotonin had increased by that time point above a minimum threshold required to induce valvular thickening and fibrosis, (ii) mTOR inhibition decreased the expression of serotonin receptors on valve cells, and/or (iii) other players are implicated including bradykinin or chromogranin A (45). Further investigations beyond the scope of these studies are needed to decipher the mechanism of mTOR inhibitors on serotonin synthesis and secretion by tumor cells and uptake by valve cells. Altogether, it might be more suitable to treat pNET patients who are at a greater risk of developing cardiac carcinoid disease with mTOR inhibitors in lieu of other FDA-approved targeted therapies such as sunitinib which has been associated with cardiotoxicity (47). Teasing out the impact of everolimus on carcinoid syndrome is complicated as many patients treated with everolimus usually receive octreotide or other somatostatin analogues in conjunction to manage their carcinoid syndrome.

Altogether, our results reveal potential applicability to other gastrointestinal and lung NETs given the efficacy of everolimus based on the RADIANT4 trial. Our findings warrant deeper future studies to test the efficacy of mTORKi(s) and/or to determine selection biomarkers for pNET patients who would more likely benefit from a switch to an mTORKi upon progression on rapalog therapy.

No potential conflicts of interest were disclosed.

Conception and design: H.M. Dahche, X. Qi, L.B. Anthony, H.E. Thomas

Development of methodology: M.A. Orr-Asman, X. Qi, H.E. Thomas

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.A. Orr-Asman, M. Jiang, K. LaSance, S.E. Koch, C.A. Mercer, J. Rubinstein, H.E. Thomas

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.A. Orr-Asman, M. Jiang, M. Worley, S.E. Koch, V.S. Carreira, K. Komurov, L.B. Anthony, J. Rubinstein, H.E. Thomas

Writing, review, and/or revision of the manuscript: M.A. Orr-Asman, S.E. Koch, V.S. Carreira, D.R. Plas, X. Qi, C.A. Mercer, L.B. Anthony, J. Rubinstein, H.E. Thomas

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Z. Chu

Study supervision: H.E. Thomas

Other (designed small-animal imaging protocol and acquired animal imaging data): K. LaSance

The authors thank Dr. Courtney Townsend for kindly providing BON cells, Celgene for providing CC-223 and Dr. Ellen Filvaroff for discussions, Drs. Sara Kozma and George Thomas for valuable discussions and critical reading of the article, members of the laboratories of Drs. Carol Mercer, Atsuo Sasaki, Sara Kozma and David Plas for project discussions, Dr. Olowokure Olugbenga for clinical insights and valuable comments, Joanne Tetens-Woodring, Carissa Lester, and all LAMS staff for animal care and advice. The authors also thank The UC Genomics and Epigenomics Core, especially Drs. Xiang Zhang and Jenny Chen and Tamara Hagen for laboratory management assistance.

This work was supported by grants to H.E.Thomas from the NANETS (YIA award), Just-in-Time funding from the Cincinnati Cancer Center, the Department of Internal Medicine Junior Faculty Pilot Project Award and Division of Hematology and Oncology Translational Science Awards. The UC Genomics, Epigenomics and Sequencing Core is partially funded by an NIEHS CEG administrative core grant P30ES006096.

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.

1.
Yao
JC
,
Hassan
M
,
Phan
A
,
Dagohoy
C
,
Leary
C
,
Mares
JE
, et al
One hundred years after "carcinoid": epidemiology of and prognostic factors for neuroendocrine tumors in 35,825 cases in the United States
.
J Clin Oncol
2008
;
26
:
3063
72
.
2.
Kulke
MH
,
Anthony
LB
,
Bushnell
DL
,
de Herder
WW
,
Goldsmith
SJ
,
Klimstra
DS
, et al
NANETS treatment guidelines: well-differentiated neuroendocrine tumors of the stomach and pancreas
.
Pancreas
2010
;
39
:
735
52
.
3.
Jiao
Y
,
Shi
C
,
Edil
BH
,
de Wilde
RF
,
Klimstra
DS
,
Maitra
A
, et al
DAXX/ATRX, MEN1, and mTOR pathway genes are frequently altered in pancreatic neuroendocrine tumors
.
Science
2011
;
331
:
1199
203
.
4.
Scarpa
A
,
Chang
DK
,
Nones
K
,
Corbo
V
,
Patch
AM
,
Bailey
P
, et al
Whole-genome landscape of pancreatic neuroendocrine tumours
.
Nature
2017
;
543
:
65
71
.
5.
Missiaglia
E
,
Dalai
I
,
Barbi
S
,
Beghelli
S
,
Falconi
M
,
della Peruta
M
, et al
Pancreatic endocrine tumors: expression profiling evidences a role for AKT-mTOR pathway
.
J Clin Oncol
2010
;
28
:
245
55
.
6.
Zaytseva
YY
,
Valentino
JD
,
Gulhati
P
,
Evers
BM
. 
mTOR inhibitors in cancer therapy
.
Cancer Lett
2012
;
319
:
1
7
.
7.
Yao
JC
,
Shah
MH
,
Ito
T
,
Bohas
CL
,
Wolin
EM
,
Van Cutsem
E
, et al
Everolimus for advanced pancreatic neuroendocrine tumors
.
N Engl J Med
2011
;
364
:
514
23
.
8.
Yao
JC
,
Fazio
N
,
Singh
S
,
Buzzoni
R
,
Carnaghi
C
,
Wolin
E
, et al
Everolimus for the treatment of advanced, non-functional neuroendocrine tumours of the lung or gastrointestinal tract (RADIANT-4): a randomised, placebo-controlled, phase 3 study
.
Lancet
2016
;
387
:
968
77
.
9.
Chan
J
,
Kulke
M
. 
Targeting the mTOR signaling pathway in neuroendocrine tumors
.
Curr Treat Options Oncol
2014
;
15
:
365
79
.
10.
Thomas
HE
,
Mercer
CA
,
Carnevalli
LS
,
Park
J
,
Andersen
JB
,
Conner
EA
, et al
mTOR inhibitors synergize on regression, reversal of gene expression, and autophagy in hepatocellular carcinoma
.
Sci Translat Med
2012
;
4
:
139ra84
.
11.
O'Reilly
KE
,
Rojo
F
,
She
QB
,
Solit
D
,
Mills
GB
,
Smith
D
, et al
mTOR inhibition induces upstream receptor tyrosine kinase signaling and activates Akt
.
Cancer Res
2006
;
66
:
1500
8
.
12.
Meric-Bernstam
F
,
Akcakanat
A
,
Chen
H
,
Do
KA
,
Sangai
T
,
Adkins
F
, et al
PIK3CA/PTEN mutations and Akt activation as markers of sensitivity to allosteric mTOR inhibitors
.
Clin Cancer Res
2012
;
18
:
1777
89
.
13.
Choo
AY
,
Yoon
SO
,
Kim
SG
,
Roux
PP
,
Blenis
J
. 
Rapamycin differentially inhibits S6Ks and 4E-BP1 to mediate cell-type-specific repression of mRNA translation
.
Proc Natl Acad Sci U S A
2008
;
105
:
17414
9
.
14.
Feldman
ME
,
Apsel
B
,
Uotila
A
,
Loewith
R
,
Knight
ZA
,
Ruggero
D
, et al
Active-site inhibitors of mTOR target rapamycin-resistant outputs of mTORC1 and mTORC2
.
PLoS Biol
2009
;
7
:
e38
.
15.
Thoreen
CC
,
Kang
SA
,
Chang
JW
,
Liu
Q
,
Zhang
J
,
Gao
Y
, et al
An ATP-competitive mammalian target of rapamycin inhibitor reveals rapamycin-resistant functions of mTORC1
.
J Biol Chem
2009
;
284
:
8023
32
.
16.
Maira
SM
,
Stauffer
F
,
Brueggen
J
,
Furet
P
,
Schnell
C
,
Fritsch
C
, et al
Identification and characterization of NVP-BEZ235, a new orally available dual phosphatidylinositol 3-kinase/mammalian target of rapamycin inhibitor with potent in vivo antitumor activity
.
Mol Cancer Ther
2008
;
7
:
1851
63
.
17.
Hanahan
D
. 
Heritable formation of pancreatic beta-cell tumours in transgenic mice expressing recombinant insulin/simian virus 40 oncogenes
.
Nature
1985
;
315
:
115
22
.
18.
Pietras
K
,
Hanahan
D
. 
A multitargeted, metronomic, and maximum-tolerated dose "chemo-switch" regimen is antiangiogenic, producing objective responses and survival benefit in a mouse model of cancer
.
J Clin Oncol
2005
;
23
:
939
52
.
19.
Chiu
CW
,
Nozawa
H
,
Hanahan
D
. 
Survival benefit with proapoptotic molecular and pathologic responses from dual targeting of mammalian target of rapamycin and epidermal growth factor receptor in a preclinical model of pancreatic neuroendocrine carcinogenesis
.
J Clin Oncol
2010
;
28
:
4425
33
.
20.
Scholz
A
,
Wagner
K
,
Welzel
M
,
Remlinger
F
,
Wiedenmann
B
,
Siemeister
G
, et al
The oral multitarget tumour growth inhibitor, ZK 304709, inhibits growth of pancreatic neuroendocrine tumours in an orthotopic mouse model
.
Gut
2009
;
58
:
261
70
.
21.
Evers
BM
,
Townsend
CM
 Jr
,
Upp
JR
,
Allen
E
,
Hurlbut
SC
,
Kim
SW
, et al
Establishment and characterization of a human carcinoid in nude mice and effect of various agents on tumor growth
.
Gastroenterology
1991
;
101
:
303
11
.
22.
Jackson
LN
,
Chen
LA
,
Larson
SD
,
Silva
SR
,
Rychahou
PG
,
Boor
PJ
, et al
Development and characterization of a novel in vivo model of carcinoid syndrome
.
Clin Cancer Res
2009
;
15
:
2747
55
.
23.
Strosberg
J
. 
Evolving treatment strategies for management of carcinoid tumors
.
Curr Treat Options Oncol
2013
;
14
:
374
88
.
24.
Dobson
R
,
Burgess
MI
,
Pritchard
DM
,
Cuthbertson
DJ
. 
The clinical presentation and management of carcinoid heart disease
.
Int J Cardiol
2014
;
173
:
29
32
.
25.
Chu
Z
,
Abu-Baker
S
,
Palascak
MB
,
Ahmad
SA
,
Franco
RS
,
Qi
X
. 
Targeting and cytotoxicity of SapC-DOPS nanovesicles in pancreatic cancer
.
PloS One
2013
;
8
:
e75507
.
26.
Basford
JE
,
Koch
S
,
Anjak
A
,
Singh
VP
,
Krause
EG
,
Robbins
N
, et al
Smooth muscle LDL receptor-related protein-1 deletion induces aortic insufficiency and promotes vascular cardiomyopathy in mice
.
PLoS One
2013
;
8
:
e82026
.
27.
Rubinstein
J
,
Lasko
VM
,
Koch
SE
,
Singh
VP
,
Carreira
V
,
Robbins
N
, et al
Novel role of transient receptor potential vanilloid 2 in the regulation of cardiac performance
.
Am J Physiol Heart Circul Physiol
2014
;
306
:
H574
84
.
28.
Komurov
K
,
Dursun
S
,
Erdin
S
,
Ram
PT
. 
NetWalker: a contextual network analysis tool for functional genomics
.
BMC Genomics
2012
;
13
:
282
.
29.
Evers
BM
,
Ishizuka
J
,
Townsend
CM
 Jr
,
Thompson
JC
. 
The human carcinoid cell line, BON. A model system for the study of carcinoid tumors
.
Ann New York Acad Sci
1994
;
733
:
393
406
.
30.
Valentino
JD
,
Li
J
,
Zaytseva
YY
,
Mustain
WC
,
Elliott
VA
,
Kim
JT
, et al
Cotargeting the PI3K and RAS pathways for the treatment of neuroendocrine tumors
.
Clin Cancer Res
2014
;
20
:
1212
22
.
31.
Guichard
SM
,
Curwen
J
,
Bihani
T
,
D'Cruz
CM
,
Yates
JW
,
Grondine
M
, et al
AZD2014, an Inhibitor of mTORC1 and mTORC2, is highly effective in ER+ breast cancer when administered using intermittent or continuous schedules
.
Mol Cancer Thera
2015
;
14
:
2508
18
.
32.
Hsieh
AC
,
Liu
Y
,
Edlind
MP
,
Ingolia
NT
,
Janes
MR
,
Sher
A
, et al
The translational landscape of mTOR signalling steers cancer initiation and metastasis
.
Nature
2012
;
485
:
55
61
.
33.
Mortensen
DS
,
Fultz
KE
,
Xu
S
,
Xu
W
,
Packard
G
,
Khambatta
G
, et al
CC-223, a potent and selective inhibitor of mTOR kinase: in vitro and in vivo characterization
.
Mol Cancer Ther
2015
;
14
:
1295
305
.
34.
Rolfe
DF
,
Brown
GC
. 
Cellular energy utilization and molecular origin of standard metabolic rate in mammals
.
Physiol Rev
1997
;
77
:
731
58
.
35.
Hardie
DG
,
Ross
FA
,
Hawley
SA
. 
AMPK: a nutrient and energy sensor that maintains energy homeostasis
.
Nat Rev Mol Cell Biol
2012
;
13
:
251
62
.
36.
Pellikka
PA
,
Tajik
AJ
,
Khandheria
BK
,
Seward
JB
,
Callahan
JA
,
Pitot
HC
, et al
Carcinoid heart disease. Clinical and echocardiographic spectrum in 74 patients
.
Circulation
1993
;
87
:
1188
96
.
37.
Fazio
N
,
Buzzoni
R
,
Baudin
E
,
Antonuzzo
L
,
Hubner
RA
,
Lahner
H
, et al
A Phase II Study of BEZ235 in Patients with Everolimus-resistant, Advanced Pancreatic Neuroendocrine Tumours
.
Anticancer Res
2016
;
36
:
713
9
.
38.
Rodrik-Outmezguine
VS
,
Okaniwa
M
,
Yao
Z
,
Novotny
CJ
,
McWhirter
C
,
Banaji
A
, et al
Overcoming mTOR resistance mutations with a new-generation mTOR inhibitor
.
Nature
2016
;
534
:
272
6
.
39.
Fan
Q
,
Aksoy
O
,
Wong
RA
,
Ilkhanizadeh
S
,
Novotny
CJ
,
Gustafson
WC
, et al
A kinase inhibitor targeted to mTORC1 drives regression in glioblastoma
.
Cancer Cell
2017
;
31
:
424
35
.
40.
Carracedo
A
,
Ma
L
,
Teruya-Feldstein
J
,
Rojo
F
,
Salmena
L
,
Alimonti
A
, et al
Inhibition of mTORC1 leads to MAPK pathway activation through a PI3K-dependent feedback loop in human cancer
.
J Clin Invest
2008
;
118
:
3065
74
.
41.
Kurmasheva
RT
,
Huang
S
,
Houghton
PJ
. 
Predicted mechanisms of resistance to mTOR inhibitors
.
Br J Cancer
2006
;
95
:
955
60
.
42.
Sonenberg
N
,
Hinnebusch
AG
. 
Regulation of translation initiation in eukaryotes: mechanisms and biological targets
.
Cell
2009
;
136
:
731
45
.
43.
Lazaris-Karatzas
A
,
Montine
KS
,
Sonenberg
N
. 
Malignant transformation by a eukaryotic initiation factor subunit that binds to mRNA 5′ cap
.
Nature
1990
;
345
:
544
7
.
44.
Ruggero
D
,
Montanaro
L
,
Ma
L
,
Xu
W
,
Londei
P
,
Cordon-Cardo
C
, et al
The translation factor eIF-4E promotes tumor formation and cooperates with c-Myc in lymphomagenesis
.
Nat Med
2004
;
10
:
484
6
.
45.
Lundin
L
,
Norheim
I
,
Landelius
J
,
Oberg
K
,
Theodorsson-Norheim
E
. 
Carcinoid heart disease: relationship of circulating vasoactive substances to ultrasound-detectable cardiac abnormalities
.
Circulation
1988
;
77
:
264
9
.
46.
Anthony
L
,
Vinik
AI
. 
Evaluating the characteristics and the management of patients with neuroendocrine tumors receiving octreotide LAR during a 6-year period
.
Pancreas
2011
;
40
:
987
94
.
47.
Chu
TF
,
Rupnick
MA
,
Kerkela
R
,
Dallabrida
SM
,
Zurakowski
D
,
Nguyen
L
, et al
Cardiotoxicity associated with tyrosine kinase inhibitor sunitinib
.
Lancet
2007
;
370
:
2011
9
.