Russo, Paul

Abstract Background Tumor cell phenotypes and anti-tumor immune responses are shaped by local metabolite availability, but intratumoral metabolite heterogeneity (IMH) and its phenotypic consequences remain poorly understood. In vitro mechanistic studies have demonstrated that the anti-tumor activity of lymphoid and myeloid cell populations is mediated by metabolite availability and signaling in the TME, raising the possibility that the immune response and metabolism of ccRCC tumors coevolve and jointly influence the likelihood that a patient responds to therapy.. However, both the broad patterns of coordination between metabolite abundance and TME cellular composition, as well as the precise cell populations producing metabolic phenotypes of interest, remain unknown. Methods To study IMH, we multiregionally profiled the metabolome, transcriptome, and genome of 187 tumor/normal regions from 31 clear cell renal cell carcinoma (ccRCC) patients. Using these measurements and additional multimodal metabolomic/transcriptomic profiling of ccRCC and other diseases, we developed computational models that can be used to understand RNA-metabolite covariation and ultimately impute metabolite levels from RNA sequencing data. Results Analysis of intratumoral metabolite-RNA covariation revealed that the immune composition of the microenvironment, and especially the abundance of myeloid cells, drove intratumoral metabolite variation. Motivated by the strength of RNA-metabolite covariation and the clinical significance of RNA biomarkers in ccRCC, we deployed and benchmarked a machine learning method (MIRTH) to impute metabolite levels directly from RNA sequencing data of primary and metastatic ccRCC tumors. We inferred metabolomic profiles from RNA sequencing data of ccRCC patients enrolled in 6 clinical trials, ultimately identifying specific metabolite biomarkers associated with response to anti-angiogenic agents. Conclusions Local metabolic phenotypes therefore emerge in tandem with the immune microenvironment and associate with therapeutic sensitivity. CDMRP DOD Funding: yes

Chromophobe renal cell carcinoma (chRCC) typically shows ~7 chromosome losses (1, 2, 6, 10, 13, 17, and 21) and ~31 exonic somatic mutations, yet carries ~5%–10% metastatic incidence. Since extensive chromosomal losses can generate proteotoxic stress and compromise cellular proliferation, it is intriguing how chRCC, a tumor with extensive chromosome losses and a low number of somatic mutations, can develop lethal metastases. Genomic features distinguishing metastatic from nonmetastatic chRCC are unknown. An integrated approach, including whole-genome sequencing (WGS), targeted ultradeep cancer gene sequencing, and chromosome analyses (FACETS, OncoScan, and FISH), was performed on 79 chRCC patients including 38 metastatic (M-chRCC) cases. We demonstrate that TP53 mutations (58%), PTEN mutations (24%), and imbalanced chromosome duplication (ICD, duplication of ≥ 3 chromosomes) (25%) were enriched in M-chRCC. Reconstruction of the subclonal composition of paired primary-metastatic chRCC tumors supports the role of TP53 , PTEN , and ICD in metastatic evolution. Finally, the presence of these 3 genomic features in primary tumors of both The Cancer Genome Atlas kidney chromophobe (KICH) ( n = 64) and M-chRCC ( n = 35) cohorts was associated with worse survival. In summary, our study provides genomic insights into the metastatic progression of chRCC and identifies TP53 mutations, PTEN mutations, and ICD as high-risk features.

513 Background: Chromophobe renal cell carcinoma (chRCC) is the third most common histologic subtype of kidney cancer. While most of these tumors have an indolent behavior, 7% of patients with chRCC develop metastases, with no currently available standard of care. The Cancer Genome Atlas characterized chRCC, highlighting pathognomonic single copy chromosomal losses of 1, 2, 6, 10, 13 and 17, as well as a minimal mutation burden distinguishing it from all other cancer types. However, only 15% of the analyzed patients had advanced disease. We analyzed metastatic chRCC to further characterize these tumors and elucidate mechanisms leading to aggressive disease using a variety of next generation and whole genome sequencing. Methods: Our cohort of metastatic chRCC consisted of 40 patients with available clinical and pathologic data. Whole genome sequencing (WGS) was performed on 6 patients (4 primary tumors and 2 metastases), 42 additional samples from 33 patients were analyzed using targeted next-generation sequencing (MSK-IMPACT). Notably, we were able to collect and analyze matched primary and metastatic tumors from 7 patients. As control cohort 27 non-metastatic chRCC tumors were sequenced with MSK-IMPACT. Copy number patterns were computed with OncoSNP seq and FACETS. Results: The most commonly mutated genes in the aggressive chRCC tumors were TP53 and PTEN (WGS: TP53 67 %, PTEN 33%; MSK-IMPACT: TP53 61%, PTEN 27%). No other genes were mutated frequently. Primary tumor samples of chRCC did show the typical pattern of chromosomal losses in 1, 2, 6, 10, 13 and 17. Interestingly, these canonical losses could not be detected in the metastases even when accounting for tumor purity. Conclusions: TP53 and PTEN mutations are highly enriched in both primary and metastatic tumors of aggressive chRCC compared to the non-aggressive tumors and likely play a critical role in disease progression. More intriguingly, the observation of differential copy numbers in matched primary and metastatic tumors suggest whole genome or whole chromosome events in these samples. We are currently employing different bioinformatic and cytogenetic platforms to validate our novel hypothesis of chromosomal events as driver for metastatic development in chRCC.

321 Background: Factors that determine renal function after partial nephrectomy (PN) are not well defined, including the impact of cold vs. warm ischemia and the relative importance of modifiable and non-modifiable factors. We studied these determinants in a large cohort of patients with a solitary functioning kidney undergoing PN. Methods: In 1980–2009, 660 PN were performed at 4 centers for tumor in a solitary-functioning kidney under cold (n=300) or warm (n=360) ischemia. Data were collected in IRB-approved registries; follow- up averaged 4.5 years. Pre- and postoperative glomerular filtration rates (GFR) were estimated via CKD-EPI equation. Results: At 3 months after PN, median GFR decreased by equivalent amounts with cold or warm ischemia (21% vs. 22%, respectively, p=0.7) although median cold ischemic times were much longer (45 vs. 22 min. respectively, p<0.001). In multivariable analyses, increasing age, larger tumor size, lower preoperative GFR, and longer ischemia time were associated with decreased postoperative GFR (p<0.05). When percentage of parenchyma spared was incorporated into the analysis, this factor and preoperative GFR proved to be the primary determinants of ultimate renal function, and duration of ischemia lost statistical significance. Conclusions: This non-randomized comparative study suggests that long-term renal function after PN is determined primarily by the quantity and quality of renal parenchyma that can be preserved. Within the relatively strict parameters of conventional practice, i.e. predominantly short ischemic intervals and liberal use of hypothermia, ischemia time was not an independent predictor of ultimate renal function after PN. Nevertheless, type and duration of ischemia remain the most important modifiable factors during PN, and mandate further study. [Table: see text]

Abstract B7x is the newest member of the B7-CD28 family and is thought to dampen immune responses via negative costimulation. Tumor expression of B7x was recently described in renal cell carcinoma (RCC) and was associated with poor outcome. We developed an assay to detect serum-soluble B7x (sB7x) and investigated 101 patients with clear cell RCC who underwent nephrectomy between 2003 and 2007. For controls, we obtained serum from 101 sex-matched blood donors within the same age range. Following an ELISA for sB7x, detectable levels (&gt;0.1 ng/mL) of sB7x were observed in 53 RCC patients compared with 18 controls (P &lt; 0.001). Median (range) concentrations of sB7x for RCC patients and controls were 14.4 ng/mL (0.1–56.9) and 2.7 ng/mL (0.2–37.1), respectively. For RCC patients with detectable sB7x, median levels were significantly higher for patients with a tumor thrombus (19.2 versus 6.6 ng/mL; P = 0.007), positive lymph nodes (41.3 versus 10.3 ng/mL; P = 0.018), and distant metastases at nephrectomy (43.3 versus 8.5 ng/mL; P = 0.002) and tended to be higher in patients with high-grade tumors (18.8 versus 8.5; P = 0.090). Additionally, median sB7x levels for tumor-node-metastasis stage I to IV RCC were 6.6, 10.3, 14.5, and 43.3 ng/mL, respectively (P = 0.012). In this first evaluation of sB7x in RCC, we show that RCC patients are more likely to have detectable sB7x compared with controls and higher sB7x levels correlate with advanced tumor stage. These early results merit further investigation of this serum marker for potential diagnostic and prognostic purposes. [Cancer Res 2008;68(15):6054–8]