Long-term Outcomes from a Phase 2 Study of Neoadjuvant Chemotherapy for Muscle-invasive Bladder Cancer (SWOG S1314; NCT02177695)

BACKGROUND

The COXEN gene expression model was evaluated for prediction of response to neoadjuvant chemotherapy for muscle-invasive bladder cancer (MIBC).

OBJECTIVE

To conduct a secondary analysis of the association of each COXEN score with event-free survival (EFS) and overall survival (OS) and by treatment arm.

DESIGN, SETTING, AND PARTICIPANTS

This was a randomized phase 2 trial of neoadjuvant gemcitabine-cisplatin (GC) or dose-dense methotrexate-vinblastine-adriamycin-cisplatin (ddMVAC) in MIBC.

INTERVENTION

Patients were randomized to ddMVAC (every 14 d) or GC (every 21 d), both for four cycles.

OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS

EFS events were defined as progression or death before scheduled surgery, a decision to not undergo surgery, recurrence, or death due to any cause after surgery. Cox regression was used to evaluate the COXEN score or treatment arm association with EFS and OS.

RESULTS AND LIMITATIONS

A total of 167 evaluable patients were included in the COXEN analysis. The COXEN scores were not significantly prognostic for OS or EFS in the respective arms, but the GC COXEN score had a hazard ratio (HR) of 0.45 (95% confidence interval [CI] 0.20-0.99; p = 0.047) when the arms were pooled. In the intent-to-treat analysis (n = 227), there was no significant difference between ddMVAC and GC for OS (HR 0.87, 95% CI 0.54-1.40; p = 0.57) or EFS (HR 0.86, 95% CI 0.59-1.26; p = 0.45). Among the 192 patients who underwent surgery, pathologic response (pT0 vs downstaging vs no response) was strongly correlated with superior postsurgical survival (5-yr OS 90%, 89% and 52%, respectively).

CONCLUSIONS

The COXEN GC score has prognostic value for patients receiving cisplatin-based neoadjuvant treatment. The randomized, prospective design provides estimates of OS and EFS for GC and ddMVAC in this population. Pathologic response (

PATIENT SUMMARY

In this study, we evaluated a biomarker to predict the response to chemotherapy. The results did not meet the preset study parameters, but our study provides information on clinical outcomes with the use of chemotherapy before surgery for bladder cancer.

Collapse

Key Words

• Biomarkers
• Cystectomy
• Neoadjuvant chemotherapy
• Urinary bladder cancer

Collapse

MESH Headings

• Humans
• Cisplatin
• Cystectomy/methods
• Deoxycytidine/therapeutic use
• Muscles/pathology
• Neoadjuvant Therapy/methods
• Neoplasm Invasiveness
• Retrospective Studies
• Treatment Outcome
• Urinary Bladder Neoplasms/drug therapy
• Urinary Bladder Neoplasms/surgery
• Urinary Bladder Neoplasms/pathology

Collapse

Grants

• U10 CA180821 NCI NIH HHS
• UG1 CA233324 NCI NIH HHS
• UG1 CA180830 NCI NIH HHS
• UG1 CA233160 NCI NIH HHS
• UG1 CA233373 NCI NIH HHS
• UG1 CA233196 NCI NIH HHS
• U10 CA180820 NCI NIH HHS
• U10 CA180888 NCI NIH HHS
• U10 CA180819 NCI NIH HHS
• UG1 CA233328 NCI NIH HHS

Collapse

Affiliation(s)

Thomas W Flaig School of Medicine, University of Colorado, Aurora, CO, USA.
Catherine M Tangen Fred Hutchinson Cancer Research Center, Seattle, WA, USA
Siamak Daneshmand Department of Urology, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA
Ajjai Shivaram Alva University of Michigan, Ann Arbor, MI, USA
M Scott Lucia School of Medicine, University of Colorado, Aurora, CO, USA
David James McConkey Johns Hopkins School of Medicine, Baltimore, MD, USA
Dan Theodorescu Cedars-Sinai Cancer, Los Angeles, CA, USA
Amir Goldkorn Division of Medical Oncology, Department of Medicine, Keck School of Medicine and Norris Comprehensive Cancer Center, University of Southern California, Los Angeles, CA, USA
Matthew I Milowsky University of North Carolina Lineberger Comprehensive Cancer Center, Chapel Hill, NC, USA
Rick Bangs Southwestern Oncology Group, San Antonio, TX, USA
Gary R MacVicar Illinois CancerCare PC, Peoria, IL, USA
Bruno R Bastos Cleveland Clinic Florida, Weston, FL, USA
Jared S Fowles Colorado State University, Fort Collins, CO, USA
Daniel L Gustafson Colorado State University, Fort Collins, CO, USA
Melissa Plets Fred Hutchinson Cancer Research Center, Seattle, WA, USA
Ian M Thompson CHRISTUS Medical Center Hospital, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA
Seth P Lerner Scott Department of Urology, Dan L Duncan Cancer Center, Baylor College of Medicine, Houston, TX, USA

Collapse

Identification of lineage-specific transcription factor defined subtypes in small cell/neuroendocrine bladder cancer

568

Background: Small cell/neuroendocrine bladder cancer (SCBC) is a rare and clinically aggressive disease. However, its molecular characteristics are poorly studied; hence, biology-based therapeutic options are not available. Strong pan-cancer similarity in gene expression signature of neuroendocrine (NE) tumors including SCBC and small cell lung cancer (SCLC) has been reported. Recent genomics studies discovered that SCLC can be grouped into lineage- specific transcription factor (TF)-defined subtypes. Therefore, we explored whether TF-defined molecular subtypes could also be observed in SCBC with the goal of developing subtype-specific therapeutic strategies. Methods: Whole transcriptome RNAseq on a cohort of n = 44 pure SCBCs collected at Johns Hopkins (n=20) and Tenon Hospital (Paris) (n=24) and DNA panel sequencing on subset of the tumors (n=24) were performed. Two expert genitourinary pathologists selected the cases for the analyses based on the histopathological characteristics of the tumor tissue. Results: SCBC tumors displayed overexpression of canonical NE markers and high mutation rates in TERT, TP53 and RB1 with 92%, 75% and 38%, respectively, confirming the concordance between genomic features and histopathologic characteristics.Lineage-specific TFs (ASCL1, NEUROD1, and POU2F3) defined three molecular subtypes of SCBC that resembled the previously reported SCLC subtypes. The three subtypes expressed variable levels of NE biomarkers and were heterogeneous with respect to distinct downstream therapeutic targets. The ASCL1 and NEUROD1 subtypes expressed the highest levels of NE lineage-specific gene expression and were associated with expression of known regulators of the neuroendocrine phenotype including FOXA2 and HES6, respectively. The ASCL1 subtype was also enriched with DLLs and other genes that control oncogenic Notch signaling, whereas tumors enriched with POU2F3, a master regulator of the NE-low SCBC subtype, expressed TRPM5, SOX9 and CHAT. Finally, an inverse association between NE lineage-specific gene expression and immune signatures associated with response to immunotherapy was observed. Conclusions: Three lineage-specific TFs (ASCL1, NEUROD1, and POU2F3) defined discrete molecular subtypes in SCBC. Each subtype was characterized by distinct therapeutic targets and gene expression signatures associated with response to immunotherapy. Future studies will be required to test the clinical relevance of these observations.

Atezolizumab with platinum and etoposide chemotherapy followed by cystectomy for patients with localized small cell neuroendocrine bladder cancer

TPS599

Background: Small cell bladder cancer (SCBC) is a rare aggressive bladder cancer (BC) variant comprising <1% of BC. Recommended front line etoposide and platinum chemotherapy for SCBC is extrapolated from small cell lung cancer (SCLCA) studies. Data describing small cell bladder cancer treatment are predominantly retrospective. Substitution of etoposide for gemcitabine with cis or carboplatin is widely accepted. This extrapolation is based on histologic similarity between the rare SCBC, and more common SCLCA. Recently, several teams have shown DNA mutation patterns are also more consistent between SCBC and SCLCA, compared to SCBC and urothelial bladder cancers, further supporting the divergence in treatment. These patterns which characterize this “neuronal” subgroup in the TCGA dataset, also reinforce this group has the poorest OS outcome vs other subtypes. Patients with SCBC are nearly uniformly excluded from BC trials, and thus the response to checkpoint inhibition and other novel therapies in SCBC is unknown. As SCBCs often respond to, then quickly recur post chemotherapy, novel active therapies are needed. Immune checkpoint blockade with anti-PD(L)1 antibodies is clinically active, leading to objective responses in approximately 15-25% of patients with urothelial cancer, and 33% in biomarker selected patients with SCLCA. In IMpower133, Atezolizumab + carboplatin + etoposide vs chemotherapy alone demonstrated significant OS advantage (12.3 vs 10.3 months) in frontline ES SCLCA, a practice changing milestone. Like urothelial bladder and SCLCA, we believe SCBC may have dramatic response to checkpoint inhibitors, as shown in small case series from our team and others. Methods: Patients with invasive (cT1-cT4) small cell/neuroendocrine carcinoma of the bladder (MIBC), with or without urothelial cancer component, who are eligible for radical cystectomy and platinum chemotherapy and immunotherapy will be enrolled. Patients with N1 disease within the true pelvis are eligible. Atezolizumab 1200 mg IV Day 1 of every 21 day cycle with etoposide and investigator choice cisplatin or carboplatin chemotherapy x 4 neoadjuvant cycles will be delivered. The primary objective is to assess pathologic complete response rate following protocol therapy. Following cystectomy, Atezolizumab maintenance Q 21 days will continue until unacceptable toxicity or loss of clinical benefit for up to 1 year (e.g., 16 cycles). With type I and type II error rates of 5 and 10% respectively, the first stage will consist of 15 patients. If 6 or fewer patients respond, the study will terminate after the first stage. Otherwise, accrual will continue to 34 patients. If 17 or fewer patients respond at the end of the second stage, the study will terminate after the second stage. If 18 or more patients respond, accrual will continue to 63 patients. Secondary endpoints include rates of DFS and OS. Clinical trial information: NCT05312671 .

Oncologic outcomes in patients with residual invasive upper tract urothelial carcinoma following neoadjuvant chemotherapy

475

Background: Emerging evidence supports use of neoadjuvant chemotherapy (NAC) prior to radical nephroureterectomy (RNU) for appropriately selected patients with upper tract urothelial carcinoma (UTUC). However, oncologic outcomes have not been well characterized for patients with residual muscle-invasive disease after NAC. Methods: We used a multi-institutional cohort from 24 centers in the U.S., Europe, and Japan to retrospectively identify patients who underwent RNU for UTUC from 1985-2022 and had high-grade muscle-invasive disease. We stratified the cohort based on receipt of NAC (>ypT2 vs. >pT2). Exclusion criteria included receipt of adjuvant chemotherapy, concurrent cystectomy with RNU, and distant metastatic disease. Baseline characteristics were compared between groups. Kaplan-Meier survival analysis with log-rank test was used to compare differences in recurrence-free survival (RFS), overall survival (OS), and cancer-specific survival (CSS). Multivariable Cox regression and Fine-Gray competing risk regression were used to determine predictors of these outcomes. Results: A total of 1,233 patients were included, 62 of whom received NAC prior to RNU. A platinum-based regimen was used in 90% of NAC recipients, and the median number of cycles administered was 4 (IQR: 3-5). Median follow-up time among all patients was 22 months (IQR: 8-47 mo.). NAC recipients were more likely to have pathologic node positivity (35% vs. 13%) and less likely to have positive tumor margins (8% vs. 28%). On Kaplan-Meier analysis, NAC recipients with residual >ypT2 disease had poorer outcomes than those with >pT2 disease (2-year RFS [NAC vs. no NAC]: 52% vs. 80%, p<0.001; 2-year OS: 60% vs. 78%, p=0.003; 2-year CSS: 61% vs. 86%, p<0.001). Multivariable analyses also showed a statistically significant association between residual muscle-invasive disease after NAC and poorer RFS, OS, and CSS (Table). Conclusions: NAC recipients with >ypT2 disease at RNU after exhibit poorer outcomes than stage-matched chemotherapy naïve counterparts. This may reflect effects of occult micrometastatic disease or chemoresistant primary tumors in non-responders. Our data highlight the need to improve prospective identification of candidates most likely to respond to NAC prior to RNU for UTUC in order to maximize its therapeutic benefit. [Table: see text]

Association of DNA damage repair (DDR) mutations (mts) and clinical outcomes in CALGB 90601 (Alliance)

4521

Background: Platinum-based chemotherapy is the standard 1st-line therapy for metastatic urothelial cancer (mUC). C90601 was a randomized phase III trial testing gemcitabine and cisplatin (GC) with bevacizumab (B) or placebo (P) in patients (pts) with untreated mUC. Median overall survival (OS) for GCB vs GCP was 14.5 months (mo) vs 14.3 mo (p=0.14) and median progression-free survival (PFS) was 8 vs 6.7 mo, respectively. DDR mts have been implicated in response and survival in mUC and were investigated in this negative trial. Methods: C90601 enrolled 506 pts randomized 1:1 to GCB or GCP from 7/15/09-12/2/14, with stratification for prior chemotherapy and visceral metastases. Consenting pts submitted archival FFPE tumor specimens and blood for matched germline (g)DNA. Tumor and gDNA were sequenced by MSK-IMPACT, a 468-gene exon capture assay, to detect mts in select DDR genes. The proportional hazards model was used to correlate mts in the DNA helicase ERCC2 (pre-specified hypothesis) and additional DDR gene panels being explored in prospective trials in muscle-invasive disease with OS and PFS, adjusting for tumor mt burden and stratification factors. Mts were categorized as deleterious (del) or non-del using pre-defined published criteria. Results: 208 pts underwent DNA sequencing. Clinical features and PFS/OS were comparable to the 506-pt cohort. Median sequencing coverage was 497X. Median mutation count was 13.2 and 8.8 for DDR mt and wild-type tumors, respectively. A non-significant improvement in OS and PFS was seen in pts with ERCC2 mts (HR 0.70), but the 5.3% frequency of ERCC2 mts was lower than in historical series. Neither del mts (table) nor any mts in DDR genes were associated with PFS/OS. Conclusions: DDR mts were not associated with improved outcomes in C90601. The reliance on archival specimens, lower-than-expected ERCC2 mt frequency, small sample sizes, and tumor genomic heterogeneity may have influenced the predictive capacity of DDR mts in this cohort. Similar analyses are underway in pts who received neoadjuvant chemotherapy prior to cystectomy from completed prospective trials. Support: U10CA180821, U10CA180882, Genentech.[Table: see text]

Association between molecular subtype membership or hypoxia-associated gene expression signatures and clinical outcomes in the CALGB 90601 (Alliance) phase 3 clinical trial of gemcitabine and cisplatin (GC) plus bevacizumab (B) or placebo (P)

4562

Background: Our previous work showed that basal tumors were associated with the best clinical outcomes in a Phase 2 clinical trial of neoadjuvant dose-dense MVAC plus B, and in other work we showed that basal tumors were enriched with hypoxia-associated gene expression signatures. Here we attempted to validate these findings in the C90601 Phase 3 clinical trial of GC plus B versus GC plus P. Methods: Whole transcriptome RNAseq was performed on all available tumors using Ion Torrent’s Ampliseq platform (n = 189). Tumors were assigned to molecular subtypes using 3 different classifiers - BASE47 (k=2), MDA oneNN (k=3), and the Consensus classifier (k=6). Tumor hypoxia signature enrichment was determined using 2 different gene expression signatures and gene set variation analysis (GSVA). The proportional hazards model was used to correlate molecular subtype calls and hypoxia signature enrichment with overall survival (OS) and progression-free survival (PFS) adjusting for stratification factors and treatment arm (for PFS). Results: The median OS & PFS by different signatures and the hazard ratios (HR) are presented in the Table. Conclusions: Predefined signatures associated with clinical benefit in the Phase-2 neoadjuvant clinical trial were not associated with benefit in C90601. Possible explanations include the lack of strong therapeutic effects of the treatments, potential heterogeneity (“subtype plasticity”) between the profiled tissue samples and the metastatic lesions under treatment pressure, and differences in biology associated with the disease states (muscle-invasive vs advanced/metastatic disease). Support: U10CA180821, U10CA180882, Department of Defense (CA160312), Genentech; ClinicalTrials.gov Identifier: NCT00942331. [Table: see text]

Cell-free DNA methylation as a predictive biomarker of response to neoadjuvant chemotherapy for patients with muscle-invasive bladder cancer in SWOG S1314

4506

Background: Neoadjuvant chemotherapy is the standard of care in muscle-invasive bladder cancer patients. However, treatment is intense, the overall benefit is small, and there is no established marker to identify patients who benefit most. The aim of the study is to characterize cell-free DNA (cfDNA) methylation from patients receiving neoadjuvant chemotherapy in SWOG S1314, a prospective cooperative group trial, and to correlate the methylation signatures with pathologic response. Methods: Blood samples were collected prospectively from 73 patients before and during standard neoadjuvant chemotherapy. At radical cystectomy, pathologic response was documented. Plasma cfDNA was profiled using Infinium MethylationEPIC BeadChip array. Differential methylation between pathologic responders (≤pT1N0M0) and non-responders was analyzed, and a Random Forest model was used to generate a classifier predictive of treatment response. Results: Using pre-chemotherapy plasma cfDNA, we developed a methylation-based response score (mR-score) predictive of pathologic response. The mR-score also could be calculated using plasma samples collected after the first cycle of neoadjuvant chemotherapy, resulting in a similar predictive ability. Furthermore, we used cfDNA methylation data to calculate the circulating bladder DNA fraction, which had a modest but independent predictive ability for treatment response. When we combined the mR-score and circulating bladder DNA fraction, we successfully predicted pathologic response outcomes in 79% of patients based on their plasma collected before chemotherapy and after 1 cycle of chemotherapy. Conclusions: Our study provides proof of concept that cfDNA methylation may be used to predict treatment response in bladder cancer patients receiving neoadjuvant chemotherapy. Clinical trial information: NCT02177695.

Association of DNA damage response (DDR) gene mutations (mts) and response to neoadjuvant cisplatin-based chemotherapy (chemo) in muscle-invasive bladder cancer (MIBC) patients (pts) enrolled onto SWOG S1314

4522

Background: Neoadjuvant cisplatin-based chemo followed by radical cystectomy (RC) is a standard of care treatment for pts with MIBC. DDR gene mts, including within ERCC2, a DNA helicase implicated in cisplatin sensitivity in MIBC, have been associated with higher pathologic (path) downstaging ( < pT2) and complete response (pT0) at RC and improved overall survival (OS) in retrospective series. S1314 randomized pts to one of 2 chemo regimens (dose dense MVAC or Gem/Cis) followed by RC. We sought to correlate ERCC2 and other DDR gene mts with response and survival in MIBC pts enrolled onto this prospective trial. Methods: Tumor and matched germline DNA from evaluable pts enrolled onto S1314 underwent exon capture sequencing of 505 cancer-associated genes (MSK-IMPACT). Both deleterious (del) mts and any mts in 9 DDR genes (ERCC2, ERCC5, BRCA1, BRCA2, RECQL4, ATM, ATR, RAD51C, FANCC) were correlated with clinical outcomes. The prespecified analyses included the association of mts with < pT2 and pT0 by logistic regression analysis and with progression-free survival (PFS) and OS by Cox proportional hazards regression. Results: 179 patients (median 61 years, 85% male, 87% white, and 87% clinical stage T2) who received >2 cycles of chemo and were evaluable for path response were included in the analysis. The pT0 rate was 28% and < pT2 was 41%. Del mts in ERCC2 were detected in 26 (14%) pts followed by ATM (n = 12, 7%), ATR (n = 3) and BRCA2 (n = 2). ERCC2 mts were associated with statistically significantly higher path responses with a 54% pT0 rate and 62% downstaging rate. Patients with any del mts had higher path response rates (51% pT0, 56% < pT2) and better PFS (Table) with a median follow-up of 53 months. There was a non-significant trend towards improved OS. Conclusions: In pts managed with neoadjuvant chemo and RC on S1314, both ERCC2 mts and del DDR gene mts correlated with pathologic response. Any del DDR gene mt was associated with improved PFS. These results are in line with retrospective analyses displaying a correlation between DDR gene mts and neoadjuvant chemosensitivity in MIBC and support ongoing genomically-informed organ sparing trials.[Table: see text]

S1314 correlative analysis of ATM, RB1, ERCC2, and FANCC mutations and pathologic complete response (pT0) at cystectomy after neoadjuvant chemotherapy (NAC) in patients with muscle invasive bladder cancer (MIBC): Implications for bladder preservation

4581

Background: SWOG S1314 (NCT02177695) was designed to validate the CoXEN classifier as a predictive biomarker in pts undergoing cystectomy after NAC. We repurposed banked DNA samples and prospective trial data from S1314 to further validate the predictive ability of the Philadelphia 4 gene signature (P4GS: any mutation in ATM, RB1, FANCC, ERCC2) to predict pT0 as previously reported (PMID: 26238431) and used in the RETAIN trial (NCT02710734). The RETAIN trial prospectively enrolled pts to receive NAC (DDMVAC) followed by allocation to bladder observation vs. intervention (cystectomy or RT) based on clinical evaluation and presence vs absence of P4GS. The primary objective of this correlative investigation was to determine whether presence of P4GS is predictive of pT0 at surgery. Methods: Eligibility for S1314 included cT2-T4a N0 M0 MIBC, cisplatin eligible, with plan for cystectomy; 237 pts were randomized between ddMVAC and gem/cis (GC) using standard dose/schedule. Of 167 pts who were evaluable for the original COXEN analysis (received 3+ cycles of chemo and evaluable for path response) adequate banked DNA was available for 105. Next-generation sequencing using the CARIS 592 Gene Panel (Caris Life Sciences, Phoenix, AZ) was performed. Pathogenic mutation or VUS of ATM, RB1, FANCC or ERCC2 was noted as present or absent for each pt and correlated with pT0 using logistic regression, adjusting for clinical stage. Results: Among the 105 pts, 51% ddMVAC, 49% GC. 15% female, 95% white, 15% clinical stage T3/T4a Prevalence of mutations: ATM (24%), ERCC2 (17%), FANCC (4%), RB1 (24%) and any variant 53%. Presence of any mutation correlated with pT0 (p = 0.0006), sensitivity 79%, specificity 59%. This association did not vary by treatment arm (MVAC vs. GC). The table below shows the contributions of each of the 4 genes with the greatest contribution from ATM and ERCC2. FANCC was non-contributory due to low prevalence. Conclusions: Patients with a mutation in ATM, RB1, FANCC or ERCC2 (P4GS) have a statistically significantly higher odds of a pT0 with GC or MVAC compared to those who do not have any variant. This signature was used to prospectively allocate patients to bladder observation as part of the RETAIN trial previously reported (ASCO GU 2021). RETAIN completed enrollment, final analysis of the primary endpoint – 2-year metastasis free survival – is expected later in 2022. Clinical trial information: NCT02177695.

Phase II trial of atezolizumab in BCG-unresponsive non-muscle invasive bladder cancer: SWOG S1605 (NCT #02844816)

4541

Background: Radical cystectomy (RC) is the standard of care for patients with BCG-unresponsive high risk non-muscle invasive bladder cancer (NMIBC), but many patients are unfit for surgery or elect bladder preservation. This trial was designed to evaluate the activity of atezolizumab in BCG-unresponsive high risk NMIBC. Methods: This single arm phase II registration trial testing systemic atezolizumab (1200 mg IV) every 3 weeks for one year aimed to enroll 135 (70 CIS and 65 non-CIS) eligible patients with histologically proven BCG-unresponsive high risk NMIBC who were unfit for or declined RC. Here we report the 18 month results for all eligible patients who received at least one protocol treatment. The co-primary endpoints were pathological complete response (CR) rate at 6 months in patients with CIS (reported at ASCO 2020), and event-free survival (EFS) in all patients at 18 months using Kaplan-Meier methods (KM), conditional on a positive CIS response rate. A sample size of 135 evaluable patients provided 93% statistical power for detecting a 30% 18-month EFS rate versus 20% using a one-sided alpha = 0.05. EFS in the subset with Ta/T1 disease and duration of response in CIS patients were secondary endpoints. Results: 172 patients were enrolled, 166 received at least one dose of atezolizumab and are included in the safety analysis, and, of those, 128 were eligible and included in the efficacy analysis. As previously reported, 20 (27%) out of 74 patients with CIS attained a pathologic CR at 6 months. The KM estimate of 12 month (actual 11.9 mo) duration of response after 6 month CR for CIS patients was 54% (95% CI 30%, 78%) and the median duration of response was 16.5 months. The KM EFS rate at 18 months in 74 patients with CIS was 17% (90% CI 9%, 25%). The 18 month KM EFS rate in the overall population of 128 patients with Ta, T1 and CIS was 29% (90% CI 22%, 36%). The 18 month actuarial EFS rate in 54 patients with Ta/T1 disease was 45% (90% CI 34, 57%). Any possibly or probably treatment-related adverse event (TRAE) was observed in 142 out of 166 (86%) patients who received any atezolizumab regardless of eligibilty. The most frequent TRAEs were fatigue 72 (43%), diarrhea 34 (20%), and anemia 38 (23%). Grade 3-5 TRAEs occurred in 28 (17%) patients, including rash in 4 (2%), hyponatremia in 4 (2%), hypertension in 3 (2%) and elevated liver function tests in 3 (2%). There were two treatment-related deaths (sepsis and respiratory failure due to myasthenia gravis). Conclusions: The observed response of atezolizumab at 6 and 18 months in patients with BCG-unresponsive CIS suggests that this could be a valuable treatment to address a critical unmet need in this patient population. The 18 month EFS in patients with Ta/T1 disease suggests activity in this patient subset. This trial provided no new safety concerns. Funding: NIH/NCI grants: CA180888, CA180819, CA180820, CA180821, CA180863 and in part by Genentech. Clinical trial information: NCT02844816.

Baseline circulating tumor cell (CTC) count as a prognostic marker of PSA response and progression in metastatic castrate sensitive prostate cancer (mCSPC): Results from SWOG S1216, a phase III randomized trial of androgen deprivation plus orteronel (cyp17 inhibitor) or bicalutamide

5506

Background: In mCSPC, androgen deprivation therapy (ADT) combined with chemotherapy or androgen receptor signaling inhibition (ARSI) is the new standard of care. Biomarkers that predict clinical outcomes with these therapies are needed. We hypothesized that CellSearch CTC count, an FDA-cleared biomarker in metastatic castrate resistant PC (mCRPC), may be a valuable biomarker in mCSPC. Methods: In S1216, peripheral blood was drawn with informed consent at registration (baseline), and CTCs were enumerated on the FDA-cleared CellSearch platform (Menarini) per standard manufacturer protocol. CTC counts were analyzed centrally for associations with 2 pre-specified trial intermediate endpoints: 7-month PSA (7mPSA) ≤ 0.2 ng/ml vs. 0.2–4.0 vs. > 4.0, (intermediate endpoint for overall survival, OS); and progression-free survival (PFS) < vs. > 2 years. Because OS data have not matured, analysis was pooled and equal numbers of samples were analyzed from each treatment arm and outcome measure (7mPSA and PFS) as stipulated by the Data Safety Monitoring Committee. Results: From 2014 to 2017, 523 baseline samples were collected. In the 7mPSA analysis (n = 264), CTCs were detected in 38% of men, with a median of 4 CTCs in those with detectable CTCs. In the PFS analysis (n = 336), CTCs were detected in 37% of men, with a median of 3 CTCs in those with detectable CTCs. Adjusting for disease burden (minimal vs. extensive) and ADT status (already initiated or not) at the time of CTC measurement, men with undetectable CTCs were 6.1-fold more likely to attain 7mPSA ≤ 0.2 (OR 6.1, 95% CI 2.1-17.2, p < 0.001) and 3.7-fold more likely to achieve > 2 years PFS (OR 3.7, 95% CI 1.7-8.1, p < 0.001) compared to men with baseline CTCs ≥ 5. Other cutpoints previously validated in mCRPC studies (CTC < 5 vs. ≥5 and CTCs 0 vs. ≥1) also strongly discriminated 7mPSA and PFS with statistical significance in this mCSPC cohort. Conclusions: CTC count at the start of treatment for mCSPC was highly prognostic of 7-month PSA response (intermediate endpoint for OS) and of PFS at 2 years. To our knowledge, this is the first such strong evidence from a prospective phase 3 trial of this magnitude. Additional analyses are planned when the trial is fully reported. Baseline CTC count may serve as a valuable prognostic marker to discriminate men likely to respond favorably to hormonal therapies from those who may benefit from early alternate interventions.

Phase II trial of atezolizumab in BCG-unresponsive non-muscle invasive bladder cancer: SWOG S1605 (NCT #02844816)

5022

Background: Radical cystectomy (RC) is the standard of care for patients with BCG-unresponsive high risk non-muscle invasive bladder cancer (NMIBC), but many patients are unfit for surgery or elect bladder preservation. Based on the reported efficacy of atezolizumab in metastatic urothelial carcinoma and the known expression of PD-L1 in NMIBC after BCG therapy, this trial was designed to evaluate the activity of atezolizumab in BCG-unresponsive high risk NMIBC. Methods: This single arm phase II registration trial testing systemic atezolizumab (1200 mg IV) every 3 weeks for one year aimed to enroll 135 (70 CIS and 65 non-CIS) eligible patients with histologically proven BCG-unresponsive high risk NMIBC who were unfit for or declined RC. Here we report on the subset with CIS (with or without concomitant Ta/T1) among patients who received at least one protocol treatment. The primary endpoint was pathological complete response (CR) rate at 6 months as defined by mandatory biopsy with a null hypothesis of 30% and alternative of 50% with a 1-sided alpha = 0.05 and 96% power. The 3 month CR rate, defined by cytology, cystoscopy and for-cause biopsy, is reported here as a secondary endpoint, in addition to safety. Results: Seventy-five eligible CIS patients were enrolled. Two received no treatment and are not evaluable. Of 73, median patient age was 73.4 years and median number of prior BCG doses was 12. Concomitant Ta/T1 tumor was found in 30 (41.1%) patients, including T1 disease in 16 (21.9%). A CR was observed in 30 (41.1%; 95% CI 29.7%, 53.2%) patients at 3 months and 19 (26.0%; 95% CI 16.5%, 37.6%) at 6 months. Any possibly or probably treatment-related adverse event (AE) was observed in 61 (83.6%) patients. The most frequent AEs were fatigue 36 (49.3%), pruritis 8 (11.0%), hypothyroidism 8 (11.0%), and nausea 8 (11.0%). Grade 3-5 AEs occurred in 9 (12.3%) patients and there was one treatment-related death (myasthenia gravis with respiratory failure and sepsis). Conclusion: The observed response to atezolizumab at 3 and 6 months in patients with BCG-unresponsive CIS was similar to that reported in recent similar trials and meets the benchmark for initial CR defined by the FDA guidance. This trial provided no new safety concerns. The duration of response will determine if this is a suitable treatment option for patients with BCG-unresponsive high risk CIS. Clinical trial information: 02844816 .

Association of molecular subtypes with pathologic response in a phase II study of co-expression extrapolation (COXEN) with neoadjuvant chemotherapy (NAC) for localized, muscle-invasive bladder cancer (SWOG S1314; NCT02177695)

5028

Background: Cisplatin-based NAC is recommended for patients with MIBC prior to radical cystectomy (RC) but the majority will not have a pathologic response. To identify responders the COXEN gene expression model with chemotherapy-specific scores (for DD-MVAC and GC) was developed and in a prospective rPII clinical trial (SWOG S1314) the GC score was associated with path downstaging in the pooled arms. We investigated RNA based molecular subtypes as additional predictive biomarkers for response to NAC in patients treated in S1314. Methods: Eligibility required cT2-T4a N0 M0, predominant urothelial, > 5 mm tumor, cisplatin eligible, and plan for RC and PLND. 237 patients were randomized between 4 cycles of ddMVAC and GC. Based on Affymetrix transcriptomic data used to assign COXEN scores, we determined subtypes using 3 classifiers: TCGA (k=5), Consensus (k=6), and MD Anderson (MDA; k=3). Primary objective was to assess subtype association with pathologic response to NAC in the pooled arms and to determine any association with COXEN. TCGA and Consensus classifiers were collapsed into 3 groups for ROC analyses. We tested whether each classifier contributed additional predictive power when added to a model based on pre-defined stratification factors (PS 0 vs. 1; T2 vs. T3, T4a). Results: 161 patients had adequate tissue and gene expression results, received at least 3 of 4 cycles of NAC and had pT-N response based on RC. Covariates were 78% PS=0, 89% T2, 84% male, median age 65, 51% randomized to ddMVAC, 49% GC with 33% pT0 and 52% downstaging. Although the TCGA 3 group classifier (Basal-Squamous (BS)/Neuronal, Luminal, Luminal infiltrated) and GC Coxen score yielded the largest AUCs (0.607, 0.610) for pT0 response, neither reached statistical significance (p=0.20, p=0.22). For downstaging (<pT2), the 3 category Consensus classifier (BS/NE-like, Luminal, Stroma-rich) significantly increased the AUC from 0.568 (strat factors alone) to 0.620 (p=0.044). The MDA classifier AUC was 0.640 and the GC Coxen score AUC was 0.626, but neither were significant (p=0.076, p=0.14. The MVAC Coxen score did not improve the AUC beyond the stratification factors. Conclusions: The Consensus classifier, which is based in part on the TCGA and MDA classifiers, modestly improved prediction for pathologic downstaging when added to clinical stage and PS. With additional followup, we will assess the association of COXEN scores and subtypes with overall survival. Clinical trial information: NCT02177695 .

Expression of nectin-4 in bladder cancer with variant histology

546

Background: The antibody-drug conjugate enfortumab vedotin is poised to change the bladder cancer (BC) treatment landscape by targeting Nectin-4, near ubiquitously expressed in urothelial cancer (UC). Less is known about this and other targets in BC with pure or mixed variant histology (VH). Methods: Immunohistochemistry (IHC) was performed on a Ventana Discovery Autostainer (Roche Diagnostics) using an ultraView DAB detection kit (Roche Diagnostics) and a Nectin-4 polyclonal antibody (1:100 dilution; Abcam, Cambridge, UK). The intensity and extent of Nectin-4 expression was determined by the histochemical scoring (H-score) used in preclinical testing, defined as the sum of the products of the staining intensity (score of 0–3) x % of cells (0–100) stained at a given intensity. Specimens were assessed by H score as: negative (0–14), weak (15–99), moderate (100–199), and strong (200–300). Results: Forty UC and VH BC were evaluated for Nectin-4 expression by IHC: 15 small cell (SCBC) (8 pure SCBC, 6 mixed SCBC/UC, 1 SCBC/CIS), 8 carcinosarcomas (CS) (7 pure CS, 1 HGUC/sarcomatoid features), and 17 pure HGUCs. Normal urothelium and stroma were negative. Eight of 8 (100%) pure SCBC were negative for Nectin-4. Six of 7 (85.3%) mixed SCBC+HGUC/CIS had weak staining and 1/7 (14.7%) had moderate staining in the urothelial components (comp) while 7/7 (100%) of the SCBC comp were negative. Seven of 7 (100%) pure CS were negative and 1/1 (100%) mixed CS+HGUC showed weak staining in the HGUC comp while the sarcomatoid comp was negative. Expression in UC was: 1/17 (5.9%) strong, 3/17 (17.6%) moderate, 10/17 (58.8%) weak, and 3/17 (17.6%) negative. Gene expression profiling confirmed Nectin-4 was downregulated in VH compared to UC samples, as was ERBB2 and Trop2. Conclusions: There is heterogeneity of expression of Nectin-4 and other targets in BC with VH compared to UC. This may have therapeutic implications, and highlights need for additional research in VH.[Table: see text]

Feasibility of digital pathology of circulating tumor cells with morphologic analysis in localized bladder cancer

525

Background: Circulating tumor cells (CTCs) are promising biomarkers in metastatic urothelial cancer (UC). Unfortunately, efforts in localized disease have been unsuccessful, in part due to limitations of existing technologies that rely on counting cells and epithelial-marker expression. Here, we applied a novel selection-free digital pathology platform in a localized UC cohort. To date, this platform has associated CTC morphology with differential therapeutic response in metastatic UC and castrate-resistant prostate cancer. If feasible in localized UC, we may potentially identify best candidates for adjuvant therapy or bladder sparing, as well as enable sensitive monitoring for recurrence. Methods: N=16 consecutive UC pts included 8 (50%) metastatic controls and 8 (50%) localized (3 (37%) at TURBT and 5 (63%) at cystectomy). Peripheral blood was processed with the Epic CTC platform (pan-CK/CD45/PD-L1/DAPI staining). Approximately 3 million cells per slide were imaged. Unsupervised clustering categorized CTCs into 5 subtypes based on 11 morphologic features (nuclear solidity, speckling, nucleoli and entropy; cytokeratin speckling and ratio; and cytoplasmic/nuclear circularity, area, and convex area ratio). Results: 119 CTCs were detected from 11/16 (69%) pts (5/8 (63%) localized (2 NMIBC, 6 MIBC) and 6/8 (75%) metastatic). All MIBC pts had cystectomy (4/6 (67%) received NAC). 2/8 (25%) metastatic pts had stable disease, 3/8 (38%) were progressing, and 3/8 (38%) had newly detected M1. Median (range) CTC count/mL was similar for localized and metastatic pts (0.4 (0-58.6), 0.75 (0-1.9)). CTCs were detected in a pt with CIS, but not in a pt with TaHG disease. 1/16 (6.3%) pts had a single PD-L1+ CTC. CTCs were successfully assigned into 5 subtypes with predominant features of large, small, or linear cells, high cytoplasmic circularity, and prominent nucleoli. Conclusions: Digital pathology and subtype assignment of CTCs is feasible in localized UC. Ongoing efforts at our center include application of this technology in localized patients receiving investigational checkpoint inhibitor therapy to potentially predict best responders or conversely those at the highest risk for recurrence.

Molecular characterization of neuroendocrine bladder cancer

561

Background: Neuroendocrine bladder cancer (NEBC) is a rare and aggressive variant that is associated with poor survival outcomes. Because NEBC is rare, the molecular features of NEBC remain poorly characterized. Therefore, we characterized NEBC at the molecular level to understand the underlying biology and identify novel therapeutic targets. Methods: Whole transcriptome RNAseq was performed on FFPE cores from 24 NEBCs and 51 conventional muscle-invasive bladder cancers (MIBCs) from Tenon Hospital in Paris. Results: Unsupervised cluster analysis of 75 tumors generated 2 distinct clusters that separated NEBCs from MIBCs. The NEBC tumors were strongly enriched with biomarkers for the characteristic of neuroendocrine or small cell malignancies, including DLL3, SOX2, and EZH2. In addition, E2F1 pathway is significantly enriched due to the impair of RB/p53 pathways. Further, the NEBCs were enriched with the TCGA’s neuronal differentiation genes that were associated with high response rates in patients treated in atezolizumab (anti-PDL1) within the context of the ImVigor 210 trial. Nevertheless, the NEBCs were characterized by suppressed immune pathway gene expression signatures, such as the Th1 pathway, effector T cell lymphocyte, and IFNg that are usually highly enriched in tumors that are sensitive to immune checkpoint blockade. Of candidate mechanisms, the suppressed TGFbpathway activity observed in the NEBCs was the most obvious explanation for sensitivity to checkpoint blockade. Conclusions: NEBCs are distinct from conventional MIBCs by gene expression signature. They are also characterized by overexpression of canonical neuroendocrine markers and inhibition of TGFb pathway activity.

Lentiviral interferon with immune checkpoint blockade: A novel method for gene therapy in bladder cancer

33

Background: Gene therapy for bladder cancer (BLCA) is rapidly evolving. We reported that intravesical adenoviral interferon-alpha (Ad-IFNα) produced a complete response in 35% of patients with BCG-unresponsive BLCA enrolled in a Phase II trial. Lentivirus (LV) is another potential vector for intravesical delivery of IFNα. LV can infect non-dividing cells and integrate into the host’s genome, making it more efficient gene delivery vectors. As treatment with IFN upregulates checkpoint inhibitors, we also wanted to investigate the role of checkpoint inhibitors with and without IFN gene therapy. Methods: Murine BLCA cell lines were transduced in-vitro with LV-IFNα (MOI 2:1). IFNα levels were measured by ELISA. Cell viability was assessed using Trypan blue dye exclusion. qPCR was used to identify expression of IFNα target genes. A LV-βGalactosidase reporter construct was delivered intravesically, and urinary IFNα levels were measured in mice treated with LV-IFNα or control virus to assess gene transfer. To assess survival benefit, the MB49 intravesical tumor model and p53+/- C57/B6 mouse model were employed. We also assessed the role of combination therapy with immune checkpoint blockade using PD1 antibody using our MB49 intravesical model. Results: Efficient LV-IFNα transduction of BLCA cells resulted in increased expression of IFNα and its target genes and reduced cell viability vs. controls (p<0.001). Mechanistically, TRAIL dependent cytotoxicity in the LV-IFNα cells was rescued by Caspase8 inhibition. Urinary IFNα levels were elevated in mice receiving LV-IFNα compared with control virus. Overall survival improved in the MB49 model and BBN model in treated mice. LV-IFN induced intratumoral CD8+ T cell infiltration, high expression of PD-L1, and inhibited angiogenesis in BBN model whereas in the MB49 tumor response was mediated by TRAIL. Combination therapy with PD1 resulted in further improved survival. Conclusions: LV-IFNα effectively upregulated IFNα target genes, was cytotoxic to murine BLCA cells, and improved the survival in mouse models. Combining it with PD1 therapy appears to further improve survival.

SWOG S1314: A randomized phase II study of co-expression extrapolation (COXEN) with neoadjuvant chemotherapy for localized, muscle-invasive bladder cancer

4506

Background: Both dose-dense Methotrexate-Vinblastine-Adriamycin/doxorubicin-Cisplatin (ddMVAC) and Gemcitabine-Cisplatin (GC) are accepted neoadjuvant regimens for muscle-invasive bladder cancer (BC). We investigated COXEN, a gene expression model, as a predictive biomarker. Methods: Eligibility included Stage cT2-T4a N0 M0, urothelial BC (mixed histology allowed), ≥ 5 mm of viable tumor, cisplatin eligible, with plan for cystectomy. 237 patients were randomized between ddMVAC, given every 14 days for 4 cycles, and GC, given every 21 days for 4 cycles. The primary objective was to assess whether the pre-specified dichotomous treatment-specific COXEN gene expression profile is prognostic of pT0 rate or ≤ pT1 at surgery, and to assess whether COXEN score is a predictive factor between regimens and response. Logistic regression was used to model response, adjusting for stratification factors. Results: 167 patients were included; the ddMVAC/GC arms had a median age of 65/64, PS = 0 in 80%/75%, Male proportion of 88%/79% and T2 stage of 87%/92%. All had at least 3 cycles of chemo and surgery/progression within 100 days of last chemo. There were favorable COXEN ddMVAC scores in 32% and GC score in 26%. The pT0 rates for ddMVAC and GC were 32% and 35%; the rates of ≤ pT1 were 55% and 49%, respectively. Conclusion: The COXEN scores were not significantly prognostic for response in their individual arms; The COXEN GC score was significant predictor for downstaging in pooled arms. There was no evidence of an interaction between COXEN score and regimen in predicting response. The prospective data and samples from this study will allow for further development of COXEN and other predictive biomarkers. Clinical trial information: NCT02177695. [Table: see text]

Lentiviral interferon: A novel method for gene therapy in bladder cancer

456

Background: Gene therapy for bladder cancer (BLCA) is rapidly evolving. We reported that intravesical adenoviral interferon-alpha (Ad-IFNα) produced a complete response in 35% of patients with BCG-unresponsive BLCA enrolled in a Phase II trial. Lentivirus (LV) is another potential vector for intravesical delivery of IFNα. Unlike the adenovirus, LV can infect non-dividing cells and integrate into the host’s genome, making it one of the most efficient gene delivery vectors. The objective of this study was to investigate lentiviral interferon-alpha (LV-IFNα) BLCA gene therapy in preclinical models. Methods: Murine BLCA cell lines were transduced in-vitro with LV-IFNα using a multiplicity of infection (MOI) of 2:1. IFNα levels were measured by ELISA. Cell viability was assessed using Trypan blue dye exclusion. qPCR was used to identify expression of IFNα target genes. A LV-βGalactosidase reporter construct was delivered intravesically, and urinary IFNα levels were measured in mice treated with LV-IFNα or control virus to assess gene transfer. To assess survival benefit, p53+/- C57/B6 mice were exposed to N-butyl-N-(4-hydroxybutyl)-nitrosamine (BBN) to induce CIS and then treated with LV-IFNα or control virus, and sacrificed when moribund. Results: Efficient LV-IFNα transduction of BLCA cells was observed at an MOI of 2:1, resulting in increased expression of IFNα and its target genes PDL-1, TRAIL, and IRF7 (p<0.001), and reduced cell viability vs. controls (p<0.001). Mechanistically, TRAIL dependent cytotoxicity in the LV-IFNα cells was rescued by Caspase 8 inhibition. βGal expression confirmed efficient transduction of murine urothelium. Urinary IFNα levels were elevated in mice receiving LV-IFNα compared with control virus. BBN mice treated with LV-IFNα had longer overall survival than mice treated with control virus (p=0.04). LV-IFNα induced intratumoral CD8+ T cell infiltration, high expression of PD-L1, and inhibited angiogenesis. Conclusions: LV-IFNα effectively upregulated IFNα target genes, was cytotoxic to murine BLCA cells, and improved the survival of BBN tumor-bearing mice. LV appears to be a promising vector for intravesical gene delivery.

S1605: Phase II trial of atezolizumab in BCG-unresponsive nonmuscle invasive bladder cancer

TPS527

Background: Radical cystectomy is the standard of care for patients with BCG-unresponsive high risk non-muscle invasive bladder cancer (NMIBC). Based on the reported efficacy of atezolizumab in metastatic urothelial carcinoma and the known expression of PD-L1 expression in NMIBC after BCG therapy, this trial will evaluate the activity of atezolizumab in BCG-unresponsive high risk NMIBC. Methods: This is a single arm phase II trial testing systemic atezolizumab (1200 mg IV) every 3 weeks for one year in 135 patients with BCG-unresponsive high risk NMIBC. The study will enroll 70 patients with CIS (with or without concomitant Ta/T1) and 65 with Ta/T1 only. Patients with CIS at baseline will undergo mandatory repeat biopsy at 6 months, and all other patients only for suspected recurrence. Patients with persistent CIS, high grade Ta/T1 recurrence or progression to muscle invasive or metastatic disease will be taken off treatment. The co-primary endpoints are: (1) complete response (CR) at 6 months in the CIS subgroup, and (2) event-free survival (EFS) at 18 months in the overall population. A hierarchical approach will be used to test the two co-primary endpoints. Secondary endpoints include duration of CR as well as progression-free, cystectomy-free, bladder cancer-specific, and overall survival in all patients. Response will be correlated to expression of PD-L1 and CD8 by IHC, and to molecular subtypes and immune signatures by RNA-sequencing. If ≥28 (40%) CIS patients respond, the agent will be considered promising. This design has a significance level of 4.6%, and a power of 96%. If the lower bound of the 90% confidence interval of the 18-month EFS excludes 20%, the investigators will conclude the regimen significantly improves EFS relative to historical data (type I error rate 0.05 and statistical power 0.93). Successful completion of this trial could lead to a new treatment paradigm for patients with BCG-unresponsive high risk NMIBC. Funding: NIH/NCI grants: CA180888, CA180819, CA180820, CA180821, and CA180863. Clinical trial registry: NCT02844816 Clinical trial information: NCT02844816.

A new 50-gene molecular subtype classifier: An evaluation of subtype stability and association with response to neoadjuvant chemotherapy in muscle-invasive bladder cancer

519

Background: The bladder cancer basal and luminal molecular subtypes have been associated with differential progression patterns and responses to neoadjuvant chemotherapy. However, they are typically identified by whole transcriptome expression profiling, which may be impractical for most academic centers. To make subtype classification more accessible, we developed a 50-gene basal and luminal subtype classifier and examined its performance in matched tumor specimens, including pretreatment biopsies from patients treated with neoadjuvant chemotherapy. Methods: We refined a 50-gene subtype classifier derived from a oneNN classifier containing >~2000 genes that we previously developed using unsupervised methods. We compared its accuracy against calls made using the original oneNN classifier, including 148 tumors from a neoadjuvant chemotherapy meta dataset. To test subtype stability, tumors in 2 different datasets (30 sets of matched primary and lymph node tumors and 43 sets of matched pre- and post-chemotherapy tumors) were assigned using 50-gene subtype classifier with Linear Discriminator Analysis (LDA) prediction algorithms. Results: In the NAC cohort, patients with tumors assigned to the basal subtype by the 50-gene classifier had better survival outcomes compared to the luminal tumors, consistent with the conclusion generated with the parent classifier (p<0.05). Basal subtype assignments were stable in 62.5 % of pairs, whereas luminal tumors displayed 100% stability. In matched pre- and post-chemotherapy tumors, basal tumors displayed 78% stability while luminal tumors showed 94% stability. Conclusions: Based on these preliminary data, it appears that basal tumors display higher plasticity than luminal tumors with these specific contexts. This plasticity may interfere with precise subtype predictions with tumors assigned to the basal subtype at biopsy.

A phase II trial of risk-adapted treatment for muscle invasive bladder cancer after neoadjuvant accelerated MVAC

TPS537

Background: Cisplatin-based neoadjuvant chemotherapy (NAC) followed by cystectomy (Cx) or chemoradiation (CRT) is the standard of care for urothelial carcinoma (UC) pts with muscle invasive bladder cancer (MIBC). Both Cx and CRT carry potential short and long-term toxicity and quality of life implications. Recent work has shown that mutations in DNA damage repair/response genes are predictive of pathologic downstaging after NAC at the time of Cx, with those pts achieving pT0 disease demonstrating excellent long-term survival (Van Allen et al. Cancer Discov. 2014; Plimack et al. Eur Urol. 2015; Liu et al. JAMA Oncol. 2016; Teo et al. CCR. 2017). Sparing pts Cx or CRT after NAC without compromising oncologic outcomes would improve quality of life and decrease morbidity. Methods: A phase II, parallel arm, multi-institutional clinical trial (NCT02710734) is being conducted to evaluate a risk-adapted approach to treatment of MIBC. Pts with cT2-T3N0M0 UC of the bladder, ECOG PS 0-1 and CrCl≥50 mL/min, undergo NAC with accelerated methotrexate, vinblastine, doxorubicin, and cisplatin. Simultaneously, the pre-NAC TURBT specimen is submitted for deep sequencing to identify variants in a panel of cancer-relevant genes (Caris Life Sciences, Phoenix, AZ). Those with an alteration in ATM, RB1, FANCC or ERCC2 and no clinical evidence of disease by restaging TUR and imaging post-NAC will begin a pre-defined active surveillance regimen that includes urinary cytological, cystoscopic, and radiographic evaluations. The remaining pts will undergo bladder-directed therapy at the discretion of the pt and clinician applying either intravesical therapy ( < cT2 post-NAC), CRT or Cx (≤cT2 post-NAC) or Cx (≥cT3 post-NAC). The primary objective is metastasis-free survival (MFS) at 2 years for all enrolled and evaluable pts. The trial has a non-inferiority design with a 14% margin between risk-adapted treatment (MFS = 78%) and standard-of-care (MFS = 64%) with a sample size of 70 pts, 82% power and a type I error of 0.045. Key secondary and translational objectives: assess the rate of UC recurrence in active surveillance pts; validate biomarkers of response to NAC; evaluate urinary biomarkers consistent with persistent UC. Clinical trial information: NCT02710734.

Muscle-invasive bladder cancer: Molecular subtypes and response to neoadjuvant chemotherapy

281

Background: Molecular subtypes of muscle-invasive bladder cancers (MIBC) have recently been discovered based on gene expression. We investigated the impact of different subtyping methods on response to neoadjuvant cisplatin-based chemotherapy (NAC) and developed a single sample model for subtyping. Methods: Transcriptome-wide microarray analysis was conducted on pre-NAC transurethral resection (TUR) specimens of 223 patients with MIBC who received NAC followed by cystectomy at 5 centers. The specimens were classified according to four published methods for molecular subtype (UNC, MDA, TCGA, Lund). Overall survival (OS) for each subtype was compared between NAC patients in this study and non-NAC patients from the provisional TCGA. A genomic classifier (GSC) was trained to predict subtype in a single sample model and validated in independent NAC (2 centers) and non-NAC datasets. Results: The models generated subtype calls similar to previously published ratios. Concordance of a given subtype between the different methods was high. Luminal tumors had the best OS independent of NAC. Patients with tumors classified as UNC basal, MDA basal and TCGA cluster III experienced the greatest improvement in OS after NAC compared to surgery alone. Tumors assigned as UNC claudin-low had the worst OS irrespective of treatment regimen (p=0.005). GSC accurately predicted four classes (luminal, luminal-infiltrated, basal, claudin-low) and the differential impact of a basal subtype on patient OS in NAC (3-yr survival of 75.2%; p=0.001) and non-NAC (3-yr survival of 42.4%; p=0.014) cohorts could be validated. Conclusions: The benefit of NAC varies between molecular subtypes. The good prognosis of luminal/cluster I tumors could not be improved with NAC, which suggests these patients may be managed best with surgery alone. The prognosis of patients with basal tumors improved the most when treated with NAC compared to surgery alone. Poor OS of claudin-low tumors even after NAC implies that these tumors are resistant to cisplatin-based chemotherapy, and these patients should be included in protocols investigating alternative treatment options like immunotherapy. Further validation prior to clinical implementation is needed.

The hotspot mutational landscape of upper tract and bladder urothelial cancers and correlation to survival

466

Background: The prevalence of potentially targetable molecular mutations in urothelial carcinoma of the upper tract and bladder carcinoma are beginning to emerge. Important differences of molecular alterations may exist between upper tract and bladder primary tumors necessitating different treatment approaches and drug development strategies for these patients. Methods: All patients with urothelial carcinoma at UT MD Anderson who had undergone the CLIA-certified 50 gene panel utilizing next generation sequencing for detection of hotspot mutations were retrospectively reviewed. The prevalence of mutations was tabulated and compared between upper tract and bladder primary tumor locations using the Fisher’s exact test. Survival comparisons were calculated using the Kaplan-Meier technique and the Log Rank Test. Results: 136 patients with urothelial carcinoma were included in the analysis of which 41 were upper tract and 95 had bladder primary tumors. Of the 136 patients, 132 had de novo or developed metastatic/recurrent disease. The most prevalent mutations in the upper tract primary cohort included FGFR3, TP53, EGFR, FBXW7, PIK3CA, and KRAS. The most prevalent mutations in the bladder primary tumors were TP53, PIK3CA, RB1, FGFR3, HRAS, KRAS, BRAF, and FBXW7. In both upper tract and bladder tumors, a mutation in FGFR3 was nearly mutually exclusive for having a TP53 mutation. Compared to bladder primary tumors, upper tract tumors were more likely to have an FGFR3 mutation (36.5% vs. 7.4%, p < 0.0001), while the prevalence of TP53 (34% vs. 49%) and or a mutation in PIK3CA, HRAS, or KRAS (12.2% vs. 20.8%) was not significantly different. In the 132 patients with metastatic disease, the presence of a mutation in PIK3CA, HRAS, or KRAS was associated with inferior survival (HR 2.342, 95% CI: 1.32-8.67, p = 0.012), while the presence of an FGFR3 or TP53 mutation was not associated with a statistically different survival when compared to patients with wild type mutational status. Conclusions: Patients with upper tract and bladder primary sites have different molecular mutation rates. Understanding how these mutations in differing sites affect tumor biology may have important clinical implications.

Differential expression of FOXF1 in bladder cancer metastases and association with established bladder cancer subtypes

464

Background: FOXF1 is a transcription factor that may act as a tumor suppressor by interaction with p53. We identified FOXF1 to be highly differentially expressed in a discovery set of primary bladder cancers and matched nodal metastases. Here we study FOXF1 in preclinical models and define the association of FOXF1 with clinical bladder cancer phenotypes in several independent cohorts. Methods: Whole genome mRNA gene expression profiling (GEP) using Illumina HT12v3-4 chips was performed on paired primary tumors and nodal metastases from a cystectomy cohort. Among the top differentially expressed genes, FOXF1 was identified as a candidate regulator of cancer cell invasion. Murine orthotopic xenografts were established from human bladder cancer cell lines (UC-3, UC-14), and mRNA expression was assessed by real-time PCR. Four independent clinical cohorts with GEP data were subtyped as luminal, p53-like, or basal according to a validated gene expression signature (Choi et al., Cancer Cell2014). Associations between FOXF1 expression with survival after cystectomy (adjusted Cox models) and bladder cancer subtype (ANOVA) were assessed. High gene expression was defined as expression > median. Results: In the initial discovery set, FOXF1 expression was 3.6-fold lower in nodal metastases than paired primary tumors (n = 33, p < 0.001). Among xenografts with nodal metastases, FOXF1 expression was 1.4-fold lower in the metastases (n = 9, p = 0.0886). In two clinical cohorts (MDA Discovery n = 73 and GSE13507 n = 55), high FOXF1 expression was associated with improved cancer specific survival (HR = 0.35, p = 0.046) and overall survival (HR = 0.45, p = 0.006), respectively. In two other cohorts (MDA Validation n = 57 and GSE32894 n = 93), FOXF1 was not associated with survival after treatment. In all four cohorts, FOXF1 was most highly expressed in tumors subtyped as p53-like (p ≤ 0.0024 for all cohorts). Conclusions: FOXF1 is differentially expressed between primary and metastatic bladder cancer lesions. Across multiple independent cohorts, FOXF1 is more highly expressed among p53-like tumors known to be resistant to neoadjuvant chemotherapy, suggesting that this may be a novel predictive biomarker.

Risk group stratification in patients with micropapillary bladder cancer treated with radical cystectomy and/or neoadjuvant chemotherapy

302

Background: Micropapillary bladder cancer (MPBC) is an aggressive variant of urothelial carcinoma. We have previously published clinical risk stratification groups for patients with conventional urothelial carcinoma and sought to identify if these were valid in patients with this variant histology. Methods: An IRB approved review of 1910 patients in our radical cystectomy database revealed 106 patients with preoperative diagnosis of ≤cT4aN0M0 MPBC between December 1992 and January 2012 who underwent upfront radical cystectomy (RC, n = 74) or neoadjuvant chemotherapy (NAC) followed by RC (n = 32). To determine whether patients with MPBC can be risk stratified using traditional risk factors, a recursive partitioning analysis (RPA) was performed. Results: In multivariate analyses, hydronephrosis (HR=3.1; p=0.01), and extent of MPBC at transurethral resection (TUR) (HR=1.9; p=0.04) were associated with shortened OS. In the reduced model, clinical stage also achieved significance (HR=2.8; p=0.03). Results were similar for DSS: hydronephrosis (HR=2.4, p=0.03), extent of MPBC (HR=2.1, p=0.03) and clinical stage (HR=4.7, p=0.02). Using the RPA analysis, following risk groups were identified according to OS or DSS: 1) cT1 disease with no hydronephrosis; 2) cT2 or higher with no hydronephrosis; or 3) hydronephrosis (with any cT stage). These groups corresponded to a low, intermediate and high-risk groups with 5-year OS and DSS rates of 85% and 91%, 50% and 57% and 16% and 17%, (p<0.001), respectively. We found these risk groups to hold true in those treated with NAC or upfront RC; those who received NAC trended towards better outcomes. Conclusions: In patients with MPBC, preoperative risk factors can help stratify patients into different risk groups similar to what is seen in patients with conventional UC. Presence of hydronephrosis is an especially ominous sign.

Evaluation of basal and luminal subtypes of urothelial carcinoma in African American and non-African American patients

305

Background: African American (AA) patients with urothelial carcinoma (UC) have been known to have a worse prognosis even when corrected for variables such as tumor stage and grade. Analysis of gene expression of several malignancies has resulted in the discovery of molecular subtypes with well-defined intrinsic biology. Recent studies in high grade (HG), muscle-invasive UC have led to the identification of two intrinsic, molecular subsets termed “luminal” and “basal” with characteristics of stages of urothelial differentiation, and that remarkably reflect the luminal and basal-like molecular subtypes of breast cancer. Patients with basal-like UC have a significantly worse overall survival. Methods: A total of 215 HG muscle-invasive UC tumors from the MDACC (n=75) and TCGA (n=140) were used to make intrinsic subtype calls using gene expression profiling (MDACC: DASL [cDNA-mediated Annealing, Selection, extension, and Ligation] and TCGA: RNA seq). Basal and luminal subtype calls were derived using previously published subtype classifiers (Damrauer et. al. PNAS, 2014 and Choi et. al. Cancer Cell, 2014). Patients were classified into AA and non-AA (white, Hispanic, or Asian) based upon self-reported race. Results: In total there were 16 and 199 tumors from AA and non-AA patients respectively. In non-AA patients, the proportion of tumors that were classified as basal and luminal were approximately equal (93 and 106 respectively), while in AA patients, there was enrichment of basal tumors (12 basal and 4 luminal) (p=0.03735, Fisher’s exact test). Conclusions: AA patients are enriched in the basal molecular subtype of UC. Similar findings have been previously documented in AA women with breast cancer. The enrichment of basal UC in AAs suggests that a biological explanation may in part underlie the poor outcomes seen in AA patients. Future studies will explore the prognostic and predictive implications of basal subtype in AA patients with UC.

The basal subtype to predict clinical benefit from neoadjuvant chemotherapy: Final results from a phase II clinical trial of DDMVAC plus bevacizumab

291

Background: Gene expression profiling (GEP) suggests 3 main subtypes of urothelial cancer: basal, which historically has the worst prognosis with high proliferation and HIF-1 expression; p53-like, with decreased proliferation and increased markers of extracellular matrix (ECM); and luminal which has increased proliferation compared to p53-like tumors. We hypothesized that GEP of transurethral resections (TUR) and cystectomy specimens from patients on a neoadjuvant trial would predict benefit from chemotherapy. Methods: Sixty patients enrolled on a neoadjuvant trial of DDMVAC+B. TUR and cystectomy specimens were available for gene expression profiling in 39 and 33 patients, respectively, with matched specimens in 23 patients. The validation set consisted of 49 patients treated with perioperative MVAC on a previously published clinical trial. Results: Chemotherapy was quite active with pT0N0 and ≤ pT1N0 down-staging rates of 38% and 53%, respectively. Basal tumors had improved survival compared to luminal and p53-like (5-year OS 91%, 73% and 36%, p=0.015). A validation cohort of patients treated with perioperative MVAC confirmed this survival benefit (5-year OS basal, luminal, and p53-like 77%, 57%, and 57%, respectively, p =0.027). The use of bevacizumab in basal tumors did not confirm evidence of significant benefit in these small numbers of patients (5-year OS bevacizumab: 91% vs MVAC: 77%, p=0.68) Bone metastases within 2 years associated exclusively with the p53-like subtype (p53-like: 100%, luminal: 0%, basal 0%, p≤0.001). The p53-like subtype was enriched at cystectomy (basal to p53-like in 3/5 (60%), luminal to p53-like in 5/7 (71%), suggesting chemo-resistance in p53-like tumors. Conclusions: In contrast to historical expectations, the basal subtype was predictive of clinical outcomes from neoadjuvant chemotherapy, reflecting the impact of chemotherapy on highly proliferative tumors. Bone metastases were associated with the p53-like subtype which is enriched for ECM. We can no longer think of urothelial cancer as one disease; subtyping should be considered for all tumors, and may have implications on selecting therapy. Clinical trial information: NCT00506155.

Extent of neutrophil (NTP) infiltration in benign lymph nodes (LNs) to predict survival in patients with muscle-invasive bladder cancer (MIBC)

e15522

Background: In preclinical models, NTPs appear to establish a pre-metastatic niche that fosters the invasion of metastases (Kowanetz et al. PNAS 2010). This observation still requires clinical validation in MIBC. Methods: Benign LN tissue was obtained from patients (pts) who had undergone cystectomy and LN dissection for documented MIBC. Immunohistochemical (IHC) staining for CD15 (a NTP marker) was performed. Interleukin-17 (IL-17) and phosphorylated signal transducer and activator of transcription 3 (pSTAT3), putative mediators of NTP recruitment, were assessed through the same method (Laan et al. J Immunol 1999; Fielding et al. J Immunol 2008). Positively staining cells were counted and averaged over 8 high power fields. Pts were stratified by the median cell count for each biomarker. Analyses of overall survival (OS) were performed using the Kaplan-Meier method and log-rank test. Results: Of 55 pts with MIBC, 19 pts received no neoadjuvant chemotherapy (NAC), while 36 pts had received NAC with either GC (n=17) or MVAC (n=19). CD15 and IL-17 expression was significantly lower in pts with prior NAC (P<0.001 for both), while expression of pSTAT3 was similar in both groups. Furthermore, across the whole cohort, a strong association was seen between expression of CD15 and IL-17 (r=0.73, P<0.001). Amongst patients with no prior NAC, median OS was higher in those pts with low CD15 v high CD15 (158.7 mos v 36.9 mos, P=0.02), and low pSTAT3 v high pSTAT3 (NR v 106.4 mos, P=0.04). Median OS was numerically higher in pts with low IL-17 v high IL-17 (114.5 v 36.9 mos; P=0.14). Patients with both low CD15/IL-17 had a particularly favorable outcome. Amongst pts with prior exposure to NAC, no difference in survival was noted based on CD15, pSTAT3 or IL-17. Conclusions: NTP recruitment to benign LNs (potentially mediated by IL-17) may be prognostic of OS in pts with MIBC who have not received NAC. Bolstered by these findings, the prognostic value of NTP recruitment will be examined prospectively in SWOG 1011, a trial comparing limited v extended LN dissection in pts with MIBC.

A novel phase I trial design featuring a two-dimensional dose-finding algorithm optimizing the dose of gemcitabine and doxorubicin with bortezomib in metastatic urothelial carcinoma (UC)

4548

Background: Preclinical studies suggested that bortezomib (B) enhanced the activity of gemcitabine and doxorubicin (GA) in UC; thus we sought to define possible combinations of bortezomib with this doublet. We employed a novel phase I trial design systematically exploring doses in 2 dimensions. The method estimates an isotoxic curve allowing not only a combination with approximately equal (with respect to single component MTD) contributions of the two components to be found, but also combinations emphasizing one component or the other. Methods: Since 11/06, 74 patients with previously treated metastatic cancer were enrolled (70 UC, 3 prostate, 1 renal). GA was treated as a single component and given in a fixed ratio to a maximum of 900 and 50 mg/m2, and B to a maximal dose of 1.6 mg/m2 IV, with dosing every 14 days. After determining the MTD along the diagonal, we then decreased the dose of B, increasing GA, and vice versa, exploring doses along an isotoxic curve aiming for ≤ 30% dose limiting toxicity (DLT) in cycle 1. The objective response rate (ORR) includes PR or CR, and excludes SD. Results: The MTD along the diagonal for GAB was 756, 42, and 1.4 mg/m2, respectively. Doses maximizing the GA (900, 50) required reduction of B to 1.2 mg/m2. Likewise, doses maximizing B (1.6) required reduction of GA to 559 and 33 mg/m2. The most common DLT were thrombocytopenia 14%, neutropenic fever 5%, and mucositis 1%. There was minimal activity at the on-diagonal MTD with an ORR 1/10. Of the tolerable doses along the isotoxic curve, the greatest activity was seen when maximizing B (1.5-1.6 mg/m2, ORR 7/12 (58%)). The ORR when maximizing GA was 4/10. The most frequent ≥ G3 toxicities include: thrombocytopenia (26%), neutropenia (26%), anemia (24%), fatigue (8%), and neutropenic fever or infection (12%). Treatment was tolerable in poor renal function; 36 patients (49%) had a GFR < 50 ml/min. Conclusions: The combination of GAB has promising activity at doses maximizing proteosome inhibition, despite relatively low doses of GA. Traditional phase 1 design dosing to the MTD "along the diagonal" would have lead to the incorrect conclusion that there was minimal activity. Clinical trial information: NCT00479128.

A STAT3- and p63-dependent transcriptional network to define a lethal basal subset of human bladder cancers

4538

Background: Muscle-invasive bladder cancers (MIBCs) are a heterogeneous group of tumors that display widely variable clinical outcomes and responses to conventional chemotherapy. Methods: We used whole genome mRNA expression profiling and unsupervised hierarchical cluster analyses on a cohort of 73 flash frozen primary tumors to identify 3 distinct subsets of muscle-invasive bladder cancer (MIBC). We confirmed the existence of these 3 subsets in a second cohort of 57 formalin-fixed, paraffin-embedded (FFPE) MIBCs and in 2 other public datasets. Analysis of primary tumors and mechanistic studies in human bladder cancer cell lines identified tumors that respond to FGFR inhibitors or chemotherapy. Results: The first subset was driven by an active "basal" EGFR-STAT3-p63 transcriptional network, and was associated with poor clinical outcomes. High miR-200c expression stratified the survival of these basal tumors. The second subset was characterized by active p53 pathway activation, and tumors and cell lines with these features were resistant to cis-platinum based chemotherapy. The third subset expressed "luminal" markers and active estrogen receptor (ER) and PPARγ signaling, and luminal cell lines were sensitive to fibroblast growth factor receptor (FGFR) inhibition. Conclusions: Molecular subtyping of MIBCs can be used to identify lethal cancers and enrich for tumors that will respond to FGFR inhibitors or conventional chemotherapy.

Augmentation of urothelial cancer cells' response to gemcitabine and cisplatin by a Smac mimetic

261

Background: Cisplatin-based chemotherapy is considered the gold standard for patients with advanced bladder cancer. Despite initial response, many patients will relapse; therefore, novel salvage treatment strategies are desperately needed. Herein, we studied a mechanism based treatment combination using a Smac mimetic with standard chemotherapy. Methods: A panel of 10 urothelial cancer cell lines was exposed to gemcitabine and cisplatin, a Smac mimetic, and the combination. Sensitivity was determined using a DNA fragmentation assay. Constitutive expression of select members of the IAP family (XIAP, cIAP-1, cIAP-2, Survivin), the BCL family (BCL-2, BCLXL, BAX) and Smac were evaluated by gene expression profiling and Western blotting. Changes in protein expression under treatment conditions were examined using Western blotting. TNF-a blocking antibody was used to explore the contribution of autocrine TNF-a in cellular death. Lastly, a mouse model of bladder cancer was used to determine the effectiveness of the drug combination. Results: UMUC-3, UMUC-13 and RT4v6 were considered sensitive to the combination of gemcitabine and cisplatin based on an apoptosis assay. Three additional cell lines were sensitized to gemcitabine and cisplatin with the addition of the Smac mimetic (UMUC-6, UMUC-12, UMUC-18). The constitutive RNA and protein expression of SMAC, select members of the IAP family, and members of the BCL family did not correlate to drug sensitivity. Autocrine TNF-a demonstrated a partial contribution to cell death in UMUC-12. In an in vivo mouse model, the Smac mimetic alone and in combination with gemcitabine and cisplatin resulted in decreased tumor volume and increased apoptosis compared to the chemotherapy treatment alone. The combination resulted in decreased cellular proliferation and decreased microvessel density. Conclusions: Smac mimetic is able to augment the activity of gemcitabine and cisplatin based chemotherapy in vitro and in vivo. This may be an effective treatment strategy in patients with urothelial carcinoma and warrants further investigation.

A novel phase I trial design featuring a two-dimensional dose-finding algorithm optimizing the dose of gemcitabine and doxorubicin with bortezomib in metastatic urothelial carcinoma (UC)

263

Background: Preclinical studies suggested that bortezomib (B) enhanced the activity of gemcitabine and doxorubicin (GA) in UC; thus we sought to define possible combinations of bortezomib with this doublet. We employed a novel phase I trial design systematically exploring doses in 2 dimensions. The method estimates an isotoxic curve allowing not only a combination with approximately equal (with respect to single component MTD) contributions of the two components to be found, but also combinations emphasizing one component or the other. Methods: Since 11/06, 74 patients with previously treated metastatic cancer were enrolled (70 UC, 3 prostate, 1 renal). GA was treated as a single component and given in a fixed ratio to a maximum of 900 and 50 mg/m2, and B to a maximal dose of 1.6 mg/m2 IV, with dosing every 14 days. After determining the MTD along the diagonal, we then decreased the dose of B, increasing GA, and vice versa, exploring doses along an isotoxic curve aiming for ≤ 30% dose limiting toxicity (DLT) in cycle 1. The objective response rate (ORR) includes PR or CR, and excludes SD. Results: The MTD along the diagonal for GAB was 756, 42, and 1.4 mg/m2, respectively. Doses maximizing the GA (900, 50) required reduction of B to 1.2 mg/m2. Likewise, doses maximizing B (1.6) required reduction of GA to 559 and 33 mg/m2. The most common DLT were thrombocytopenia 14%, neutropenic fever 5%, and mucositis 1%. There was minimal activity at the on-diagonal MTD with an ORR 1/10. Of the tolerable doses along the isotoxic curve, the greatest activity was seen when maximizing B (1.5-1.6 mg/m2, ORR 7/12 (58%)). The ORR when maximizing GA was 4/10. The most frequent ≥ G3 toxicities include: thrombocytopenia (26%), neutropenia (26%), anemia (24%), fatigue (8%), and neutropenic fever or infection (12%). Treatment was tolerable in poor renal function; 36 patients (49%) had a GFR < 50 ml/min. Conclusions: The combination of GAB has promising activity at doses maximizing proteosome inhibition, despite relatively low doses of GA. Traditional phase I design dosing to the MTD "along the diagonal" would have lead to the incorrect conclusion that there was minimal activity. Clinical trial information: NCT00479128.

Prediction of survival in patients with muscle-invasive bladder cancer (MIBC) by neutrophil (NTP) infiltration in benign lymph nodes (LNs)

273

Background: In preclinical models, NTPs appear to establish a pre−metastatic niche that fosters the invasion of metastases (Kowanetz et al PNAS 2010). This observation still requires clinical validation in MIBC. Methods: Benign LN tissue was obtained from patients (pts) who had undergone cystectomy and LN dissection for documented MIBC. Immunohistochemical (IHC) staining for CD15, a NTP marker, was performed for the entire cohort. Phosphorylated signal transducer and activator of transcription 3 (pSTAT3), vascular endothelial growth factor receptor−1 (VEGFR1), and CD68 (a macrophage marker) were further assessed in an initial cohort (detailed subsequently) via IHC. Positively staining cells were counted and averaged over 8 high power fields (hpfs). Pts were stratified by the median cell count for each biomarker. Analyses of overall survival (OS) were performed using the Kaplan−Meier method and log−rank test. Results: In an initial cohort of 19 pts who had received no neoadjuvant chemotherapy (NC), a median CD15 count of 284/hpf was noted. Median OS was higher in those pts with low CD15 as compared to high CD15 (158.7 mos v 36.9 mos, P=0.02). Median OS was also improved in those with high pSTAT3 v low pSTAT3 (not reached v 106.4 mos, P=0.04), but no difference was noted in OS in groups stratified by clinical stage, VEGFR1 staining, or CD68 staining. To determine if the prognostic value of CD15 staining was retained in pts with exposure to NC, the cohort was expanded to include an additional 36 pts who had received either preoperative GC (n=17) or MVAC (n=19) chemotherapy. In this group, no significant difference in OS was noted based using the previously applied CD15 cutoff. Conclusions: NTP recruitment to benign LNs may be prognostic of OS in pts with MIBC who have not received NC. pSTAT3, a putative mediator of NTP recruitment, may play a role in this phenomenon. VEGFR1 and CD68, which may mediate pre−metastatic niche through a different mechanism, do not predict OS in our dataset (Kaplan et al Nature 2005). Bolstered by these findings, the prognostic value of NTP recruitment will be examined prospectively in SWOG 1011, a trial comparing limited v extended LN dissection in pts with MIBC.

Relationship between high microRNA-200C expression and the risk of death from disease in muscle-invasive urothelial carcinoma of the bladder

279

Background: Bladder cancer represents both a common and highly morbid disease with limited tools for prognostication. Recently molecular studies have led to promising targets to inform disease behavior, including microRNAs (miRNA). miRNAs are non-coding RNAs, with widespread effects on cellular function. Our group has demonstrated that the miR200 family members play an integral role in epithelial to mesenchymal transition (EMT) via an inverse relationship with ZEB1 and a direct relationship with E-cadherin. EMT is seen as a necessary step for invasion and metastasis, however, studies have indicated a significant role for mesenchymal to epithelial transition (MET) to drive proliferation after cells have reached metastatic locations. Members of the miR200 family have been shown to induce MET and we hypothesize that miR200c, as a surrogate marker for MET, and will predict disease survival in muscle invasive urothelial carcinoma (MIUC).

Methods: A clinically diverse sample set was obtained consisting of 101 unique specimens upon which real-time PCR miRNA analysis was performed. Regression tree analysis and best-fit modeling was used to establish the most discriminating relative miR200c expression level to predict disease specific survival. Fisher exact test was carried out to compare clinical variables, Kaplan-Meier estimate of survival distribution based on miR200c expression and univariate log-rank test was used to compare survival distributions between groups. Multivariate analysis was done via the proportional hazards model. A p-value of <0.05 was considered significant for all statistical analyses.

Results: Patients with high miRNA200c had significantly more deaths (69 vs. 47%). In multivariable analysis of patients with MIUC miR200c expression was associated with the highest risk of death (RR 2.7). Lymph node involvement (RR 2.0) and age > 65yrs (RR 2.4) were also strong predictors of survival. High miR200c had lower median survival for all patients (59 vs 16 months; p = 0.039) and those with MIUC (41 vs 8 months; p = 0.0004).

Conclusions: High miR200c expression is associated with a higher risk of death from bladder cancer in patients with muscle invasive disease.

Marizomib, a proteasome inhibitor for all seasons: preclinical profile and a framework for clinical trials

The proteasome has emerged as an important clinically relevant target for the treatment of hematologic malignancies. Since the Food and Drug Administration approved the first-in-class proteasome inhibitor bortezomib (Velcade) for the treatment of relapsed/refractory multiple myeloma (MM) and mantle cell lymphoma, it has become clear that new inhibitors are needed that have a better therapeutic ratio, can overcome inherent and acquired bortezomib resistance and exhibit broader anti-cancer activities. Marizomib (NPI-0052; salinosporamide A) is a structurally and pharmacologically unique β-lactone-γ-lactam proteasome inhibitor that may fulfill these unmet needs. The potent and sustained inhibition of all three proteolytic activities of the proteasome by marizomib has inspired extensive preclinical evaluation in a variety of hematologic and solid tumor models, where it is efficacious as a single agent and in combination with biologics, chemotherapeutics and targeted therapeutic agents. Specifically, marizomib has been evaluated in models for multiple myeloma, mantle cell lymphoma, Waldenstrom's macroglobulinemia, chronic and acute lymphocytic leukemia, as well as glioma, colorectal and pancreatic cancer models, and has exhibited synergistic activities in tumor models in combination with bortezomib, the immunomodulatory agent lenalidomide (Revlimid), and various histone deacetylase inhibitors. These and other studies provided the framework for ongoing clinical trials in patients with MM, lymphomas, leukemias and solid tumors, including those who have failed bortezomib treatment, as well as in patients with diagnoses where other proteasome inhibitors have not demonstrated significant efficacy. This review captures the remarkable translational studies and contributions from many collaborators that have advanced marizomib from seabed to bench to bedside.

Proteasome inhibitors activate autophagy as a cytoprotective response in human prostate cancer cells

The ubiquitin-proteasome and lysosome-autophagy pathways are the two major intracellular protein degradation systems that work cooperatively to maintain homeostasis. Proteasome inhibitors (PIs) have clinical activity in hematological tumors, and inhibitors of autophagy are also being evaluated as potential antitumor therapies. In this study, we found that chemical PIs and small interfering RNA-mediated knockdown of the proteasome's enzymatic subunits promoted autophagosome formation, stimulated autophagic flux, and upregulated expression of the autophagy-specific genes (ATGs) (ATG5 and ATG7) in some human prostate cancer cells and immortalized mouse embryonic fibroblasts (MEFs). Upregulation of ATG5 and ATG7 only occurred in cells displaying PI-induced phosphorylation of the eukaryotic translation initiation factor 2 alpha (eIF2alpha), an important component of the unfolded protein responses. Furthermore, PIs did not induce autophagy or upregulate ATG5 in MEFs expressing a phosphorylation-deficient mutant form of eIF2alpha. Combined inhibition of autophagy and the proteasome induced an accumulation of intracellular protein aggregates reminiscent of neuronal inclusion bodies and caused more cancer cell death than blocking either degradation pathway alone. Overall, our data show that proteasome inhibition activates autophagy through a phospho-eIF2alpha-dependent mechanism to eliminate protein aggregates and alleviate proteotoxic stress.

Early RB94-produced cytotoxicity in cancer cells is independent of caspase activation or 50 kb DNA fragmentation

RB94, which lacks the N-terminal 112 amino-acid residues of the full-length retinoblastoma protein (RB110) is a more potent inhibitor of cancer cell growth than RB110, being cytotoxic to all cancer cell lines studied, independent of their genetic abnormalities. Although we initially thought RB94-induced cell death was caspase-dependent, such caspase activation now appears to be a late event. Cells that remained attached 48 h after transduction with Ad-RB94 showed, among other changes, nuclear enlargement, peripheral nuclear chromatin condensation and often micronucleation. In addition, the cells were TdT-mediated dUTP nick end labeling (TUNEL) positive but showed no cleavage of caspase 3 or 9. Only after the cells detached was cleavage of both caspase 3 and 9 observed. These TUNEL-positive cells showed neither cytochrome c mitochondrial translocation usually found in typical apoptotic cells nor DNA laddering indicative of oligonucleosomal DNA fragmentation. In addition, although 50 kb DNA fragmentation was produced in these TUNEL-positive cells, which was dependent on apoptosis-inducing factor (AIF), inhibiting this fragmentation by siAIF did not inhibit TUNEL formation or cytotoxicity. As RB94 will soon be used for gene therapy further understanding the molecular basis of these early changes in killing cancer cells is one of our particularly important present goals.

Adenoviral-mediated interferon alpha overcomes resistance to the interferon protein in various cancer types and has marked bystander effects

We have previously shown that intravesical administration of adenovirus encoding human interferon alpha-2b (Ad-IFN) induced a marked regression of superficial human bladder tumors derived from cells that are resistant to over 1 million units/ml of IFNalpha protein in vitro. In addition, Ad-IFN appeared to produce strong bystander effects. In this study, we show that Ad-IFN causes marked inhibition of cell growth and apoptosis in cells of various tumor types, all of which are resistant to IFNalpha protein. In addition, strong perinuclear IFN staining was seen in all cell lines following Ad-IFN transfection and was never observed after exposure to the IFN protein. Ad-IFN induced proteolytic processing of caspases 3, 8 and 9, indicative of enzymatic activation. However, the caspase-8-selective inhibitor, IETDfmk, blocked apoptosis only in the cell lines that were sensitive to the IFNalpha protein and had minimal effect on Ad-IFN-induced caspase-3 or -9 processing and cell death, indicating that death receptor-independent mechanism(s) were involved in the cytotoxic effects observed for cancer cell lines resistant to the IFNalpha protein. Moreover, we document that a yet to be identified soluble factor(s) is responsible for causing the bystander effect observed following Ad-IFN treatment in IFN protein-resistant cancer cells.

Increased mitochondrial biogenesis in primary leukemia cells: the role of endogenous nitric oxide and impact on sensitivity to fludarabine

B cell chronic lymphocytic leukemia (CLL) is the most prevalent adult leukemia in the Western hemisphere, yet many biological and molecular features of the disease remain undefined. CLL cells generate increased levels of radical species such as superoxide and nitric oxide (NO), which is associated with mitochondrial DNA mutations. Considering that NO levels can affect mitochondrial biogenesis, we hypothesized that the inherent nitrosative stress in CLL cells may lead to hyperactive mitochondrial biogenesis. Here we report that primary CLL cells contained significantly more mitochondria than normal lymphocytes and that their mitochondrial mass was significantly related to endogenous NO levels. Expression of the mitochondrial biogenesis factors nuclear respiratory factor-1 and mitochondrial transcription factor A was elevated in most CLL specimens examined and appeared to be related to cellular NO levels. Treatment of B cells with exogenous NO caused a substantial increase in mitochondrial mass. In vitro sensitivity of CLL cells to fludarabine was highly related to mitochondrial mass in that cells with greater mitochondrial mass were less sensitive to the drug. Taken together, our results suggest that NO is a key mediator of mitochondrial biogenesis in CLL and that modulation of mitochondrial biogenesis by NO may alter cellular sensitivity to fludarabine.

Surrogate markers in antiangiogenesis clinical trials

Novel antiangiogenic agents currently being developed may ultimately be more effective against solid tumours and less toxic than cytotoxic chemotherapy. As a result of the early clinical trials of angiogenesis inhibitors, investigators are beginning to appreciate the complexity of targeting angiogenesis and the realisation that developing clinically useful antiangiogenic therapy will be more challenging than originally thought. It is now apparent that new methods and surrogate markers to assess these agents' biological activity are crucial for their successful development. This review summarises the currently available clinical data on the development of surrogate markers of angiogenesis inhibitors.

p53 and Fas ligand are required for psoralen and UVA-induced apoptosis in mouse epidermal cells

A combination of 8-methoxypsoralen (8-MOP) and ultraviolet-A (UVA) radiation (320-400 nm) (PUVA) is widely used in the treatment of psoriasis and other skin diseases. PUVA is highly effective in eliminating hyperproliferative cells in the epidermis, but its mechanism of action has not been fully elucidated. In this study, we used immortalized JB6 mouse epidermal cells, p53(-/-), and Fas ligand deficient (gld) mice to investigate the molecular mechanism by which PUVA induces cell death. The results indicate that PUVA treatment induces apoptosis in JB6 cells. In addition, PUVA treatment of JB6 cells results in p53 stabilization, phosphorylation, and nuclear localization as well as induction of p21(Waf/Cip1) and caspase-3 activity. In vivo studies reveal that PUVA treatment induces significantly less apoptosis in the epidermis of p53(-/-) mice compared to p53(+/+) mice. Furthermore, FasL-deficient (gld) mice are completely resistant to PUVA-induced apoptosis compared to wild-type mice. These results indicate that PUVA treatment induces apoptosis in mouse epidermal cells in vitro and in vivo and that p53 and Fas/Fas ligand interactions are required for this process, at least in vivo. This implies that similar mechanisms may be involved in the elimination of psoriatic keratinocytes from human skin following PUVA therapy.

Calcium flux measurements in apoptosis

Early studies in apoptosis implicated an increase in cytosolic Ca2+ as a direct mediator of DNA fragmentation. However, efforts to delineate targets for this increase in Ca2+ have been slow in evolving. Several previous studies have implicated ER Ca2+ pool depletion in the initiation of apoptosis. Our own preliminary studies confirm that many (but not all) apoptotic stimuli empty the ER store via a mechanism that is blocked by BCL-2 expression. Furthermore, ER pool depletion is not affected by broad spectrum caspase inhibitors, indicating that it occurs via a caspase-independent mechanism. Finally, our data demonstrate that ER pool depletion occurs prior to release of cytochrome c from mitochondria. Given previous work demonstrating close coordination of ER and mitochondrial Ca2+ levels, we speculate that ER-dependent changes in mitochondrial Ca2+ serve as important signals for cytochrome c release. Alternative mechanisms include activation of caspase-12 and/or the JNK pathway, both of which can be directly stimulated by depletion of the ER Ca2+ pool. Although substantial improvements in intracellular Ca2+ imaging have emerged, compelling answers to many of the present questions related to the role of Ca2+ in apoptosis await future technical improvements. The development of organelle-specific, recombinant Ca2+ probes (targeted aequorins and cameleons) certainly should facilitate some of this work, although the target cell of interest must be amenable to molecular manipulation (transfection), which precludes straightforward analysis of primary cells. Pharmacological tools (i.e., thapsigargin and DBHQ) can provide conclusive data on ER pool status without requiring an overly sophisticated image analysis system. However, confocal microscopy allows for the effective analysis of Ca2+ pools as long as dye localization is homogeneous and properly controlled. However, current techniques should be considered semiquantitative at best and will remain so until specific organelle-targeted fluorescent dyes are developed and widely available.

Histone deacetylase inhibitor downregulation of bcl-xl gene expression leads to apoptotic cell death in mesothelioma

It has been shown that mesothelioma expresses the antiapoptotic protein BCL-XL, but not BCL-2, rendering bcl-xl gene expression a potential therapeutic target. Sodium butyrate (NaB) is a histone deacetylase inhibitor capable of alteration of bcl-2 family protein expression in other tumor types. Mesothelioma cell lines (REN, I-45) were exposed to NaB, and viability (colorimetric assay) and apoptosis (TUNEL, Hoescht staining, flow cytometry) were evaluated. Effects on bcl-2 family protein, fas-fas ligand, and caspases were examined by Western blot analysis and functional assay. An RNase assay evaluated bcl-2 family messenger RNA (mRNA) expression. Overexpressing BCL-XL mesothelioma clones were created by plasmid transfer. Cells were sensitive to NaB at low IC(50) (REN, 0.3 mM; I-45, 1 mM) and demonstrated apoptosis (percentage of cells below G1 phase by flow cytometry [sub-G1]: REN, 38.5%; I-45, 30.9%). A significant decrease in BCL-XL protein expression was noted with BAK, BAX, and BCL-2 unchanged, and this was corroborated at the transcriptional level with selectively decreased bcl-xl mRNA production after sodium butyrate exposure. Fas expression and fas-fas ligand sensitivity were unchanged. Caspases demonstrated low-level activation. Stable overexpressing BCL-XL clones were proportionally resistant to the NaB effect. This study suggests that mesothelioma cells are sensitive to the induction of apoptosis related to the attenuation of antiapoptotic bcl-xl gene and protein expression. Additional study of the therapeutic benefit of targeting bcl-xl gene expression in mesothelioma is warranted.

BCL2 expression correlates with metastatic potential in pancreatic cancer cell lines

BACKGROUND

Programmed cell death (termed apoptosis) regulates normal tissue homeostasis. Loss of local paracrine signals and intercellular adhesion molecules are potent inducers of apoptosis and thereby eliminate normal cells that may have escaped beyond the confines of the local organ environment. Dysregulation in the expression of the BCL2 gene family, the prototypic regulators of apoptosis, is a common occurrence in cancer and imparts resistance to standard triggers of apoptosis. Therefore, the authors sought to examine whether abnormal BCL2 gene family expression correlated with resistance to apoptosis and increased metastatic potential in pancreatic carcinoma.

METHODS

The authors examined BCL2 expression and apoptotic sensitivity in three panels of human pancreatic cancer cell lines that possess varying metastatic potential. Stable transfectants were generated that overexpress BCL2. These transfectants were then analyzed for differences in metastasis formation in athymic mice.

RESULTS

Among the isogenic panels of pancreatic cancer cell lines, BCL2 expression levels correlated with metastatic potential. Highly metastatic variants of each family of cell lines were more resistant to induction of apoptosis. Finally, using the BCL2 transfectant in a xenograft model, elevated BCL2 expression led to a higher incidence of metastases.

CONCLUSIONS

The authors conclude that increased BCL2 expression correlates with apoptotic resistance and metastatic potential; dysregulation of BCL2 expression may be involved in the metastatic progression of pancreatic carcinoma.

In vivo intracellular signaling as a marker of antiangiogenic activity

Alterations in endothelial cell (EC) signaling could serve as a marker of effective antiangiogenic therapy. We determined the effect of an antiangiogenic tyrosine kinase inhibitor, SU6668, on tumor EC signaling in liver metastases in mice. In vitro immunofluorescence verified that pretreatment of ECs with SU6668 before exposure to VEGF decreased in vitro phosphorylation of Erk and Akt. Using double-fluorescence immunohistochemistry, phosphorylated Erk and Akt were constitutively expressed in ECs in liver metastases in untreated mice, but SU6668 blocked activation of these signaling intermediates. Determining the activation status of the Erk and Akt signaling pathways in tumor ECs may serve as a surrogate marker for the effectiveness of antiangiogenic regimens.

Treatment with low-dose interferon-alpha restores the balance between matrix metalloproteinase-9 and E-cadherin expression in human transitional cell carcinoma of the bladder

Tumor invasion and metastasis are regulated by the expression of genes such as E-cadherin, which regulates cell adhesion, and matrix metalloproteinase-9 (MMP-9), which alters the integrity of the extracellular matrix. Both up-regulation of MMP-9 and down-regulation of E-cadherin correlate with bladder cancer metastasis. The purpose of this study was first to determine whether an imbalance between MMP-9 and E-cadherin expression correlates with metastasis from human transitional cell carcinoma (TCC) of the bladder after therapy with neoadjuvant chemotherapy and radical cystectomy and then to determine whether treatment of human TCC xenografts growing in nude mice with interferon (IFN)-alpha would restore this balance, thereby limiting tumor invasion and metastasis. We used in situ hybridization to evaluate the expression of several metastasis-related genes, including MMP-9 and E-cadherin, in paraffin-embedded biopsy specimens from 55 patients with muscle-invasive TCC treated with neoadjuvant methotrexate, vinblastine, doxorubicin, and cisplatin chemotherapy and radical cystectomy. By multivariate analysis, an MMP-9:E-cadherin ratio of >1.8 was an independent prognostic factor for disease progression. In vitro incubation of an IFN-resistant, highly metastatic human TCC cell line, 253J B-V(R) with noncytostatic concentrations of IFN-alpha down-regulated the activity of MMP-9, up-regulated E-cadherin, and inhibited in vitro invasion. 253J B-V(R) cells were implanted into the bladders of athymic nude mice. Systemic therapy with IFN-alpha (10,000 units s.c. daily) decreased the expression of MMP-9, increased expression of E-cadherin, reduced tumor volume, and inhibited metastasis. The MMP-9:E-cadherin ratio was 4.5 in untreated controls and 1.1 after IFN-alpha treatment. Moreover, systemic low-dose daily IFN-alpha potentiated the efficacy of paclitaxel. These studies indicate that in addition to its antiproliferative and antiangiogenic effects, IFN-alpha limits tumor invasion by restoring the normal balance between MMP-9 and E-cadherin and enhances the activity of systemic chemotherapy.

BCL2 expression correlates with metastatic potential in pancreatic cancer cell lines

BACKGROUND

Programmed cell death (termed apoptosis) regulates normal tissue homeostasis. Loss of local paracrine signals and intercellular adhesion molecules are potent inducers of apoptosis and thereby eliminate normal cells that may have escaped beyond the confines of the local organ environment. Dysregulation in the expression of the BCL2 gene family, the prototypic regulators of apoptosis, is a common occurrence in cancer and imparts resistance to standard triggers of apoptosis. Therefore, the authors sought to examine whether abnormal BCL2 gene family expression correlated with resistance to apoptosis and increased metastatic potential in pancreatic carcinoma.

METHODS

The authors examined BCL2 expression and apoptotic sensitivity in three panels of human pancreatic cancer cell lines that possess varying metastatic potential. Stable transfectants were generated that overexpress BCL2. These transfectants were then analyzed for differences in metastasis formation in athymic mice.

RESULTS

Among the isogenic panels of pancreatic cancer cell lines, BCL2 expression levels correlated with metastatic potential. Highly metastatic variants of each family of cell lines were more resistant to induction of apoptosis. Finally, using the BCL2 transfectant in a xenograft model, elevated BCL2 expression led to a higher incidence of metastases.

CONCLUSIONS

The authors conclude that increased BCL2 expression correlates with apoptotic resistance and metastatic potential; dysregulation of BCL2 expression may be involved in the metastatic progression of pancreatic carcinoma.