Predictive biomarkers

Kidney cancer. Principles and practice. Second edition. Primo N. Lara, Jr. Eric Jonasch (Editors). Springer International Publishing (2015)

With the abundance of approved therapies for RCC, oncologists now have the luxury to choose individualized therapy for each patient. Traditional immunotherapy should be retailored to fit selected patients better. Targeted therapies not only have invigorated RCC oncologic practice but also have changed the approaches used to predict response to therapy and to measure clinical outcome. In the next section, we differentiate and discuss biomarkers according to different therapies (Table 4.2).

Predictive biomarkers for immunotherapy

Despite the advances of targeted therapy, traditional immunotherapy is not obsolete. Immunotherapy offers the possibility of a complete and durable response for a small number of patients with favorable disease factors. However, the toxicities from immunotherapy are significant and the disease factors, which favor immunotherapy, are uncertain. Immunotherapy is therefore often not considered a reasonable option. A reliable biomarker would be ideal to select patients who are likely to have a good response or less toxicity to immunotherapy, as well as to monitor their progress. In addition, the introduction of anti-PD1 (programmed death-1) therapy could also uncover predictive biomarkers for this therapy in the near future.

  • RCC subtyping

It is clear that RCC subtyping for clear cell histology is an important predictive biomarker for immunotherapy [88–90]. The Cytokine Working Group performed a retrospective analysis of tumor tissue from 231 RCC patients treated with interleukin (IL)-2 immunotherapy. The response rate to IL-2 was 21% in patients with ccRCC, compared with 6% with non-ccRCC [90]. Similar results were found in the SELECT trial, with zero out of five patients responding [91]. Among the patients with ccRCC, those with >50% alveolar and no granular or papillary feature had the best response to IL-2 [90].

Table 4.2. Potential predictive biomarkers of response to targeted therapies for renal cell carcinoma (RCC)

IL indicates interleukin, CAIX carbonic anhydrase IX, VEGFR vascular endothelium growth factor receptor, NGAL neutrophil gelatinase-associated lipocalin bFGF basic fibroblast growth factor, HIF hypoxia-inducible factor, TNF tumor necrosis factor, MMP matrix metallopeptidase 9, VHL von Hippel-Lindau gene, CXCR chemokine receptor, TGF transforming growth factor, WT wild type, HGF hepatic growth factor, LDH lactate dehydrogenase

  • CAIX

CAIX expression had initially been reported as a predictive biomarker of response to IL-2 [32, 92].

High CAIX expression (>85% of tumor cells) was observed in 78% of patients responding to IL-2, compared with only 51% in nonresponders after examination of 66 RCC patients (27 responders). However, the role of CAIX as a predictive biomarker was further studied in the prospective SELECT trial in combination with histologic features but failed to predict responsiveness to IL-2 [91].

  • Genetic studies

Genetic studies as predictive biomarkers have also been explored for immunotherapies. Pantuck and colleagues reported an expression panel of 73 genes potentially useful to identify complete responders from nonresponders after IL-2 therapy [32]. Interestingly, complete responders to IL-2 possessed unique expression patterns of genes including CAIX, PTEN, and CXCR4. An analysis of a-CGH in ccRCC showed that tumors from complete responders to IL-2 had fewer whole chromosome losses than nonresponders. The loss of chromosome 9p was present in 65% of nonresponders vs. 0% of complete responders [93]. Pioneering work using SNP genotyping to predict the response to IFN-α has also been reported [94]. A stepwise logistic regression analysis revealed that the SNPs in signal transducer and activator 3 (STAT3) were significantly associated with better response to IFN-α. All of these findings from exploratory retrospective analyses remain to be validated in prospective studies.

  • PD1

PD1 (programmed death) is a T-cell immune checkpoint receptor thought to be involved in tumor-mediated immunosuppression. Preliminary data from phase II studies suggest that patients with RCC that expresses PDL1 (the ligand that binds T-cell PD1) on their tumors may benefit from anti-PD1 therapy more than those without, although these results need to be validated in future studies [95].

Predictive biomarkers for VEGF-targeted therapy

  • Clinical biomarkers

It is intriguing to note that hypertension (HTN), a frequent side effect of VEGF-targeted therapy, has been strongly associated with clinical outcome in the setting of VEGF-directed agents. Rini et al. reported that HTN could be used as a predictive biomarker of efficacy in patients treated with sunitinib [96]. Patients with a maximum systolic blood pressure (SBP) of 140 mmHg or more had a greater improvement in both PFS (12.5 vs. 2.5 months; p < 0.0001) and OS (30.5 vs. 7.8 months; p < 0.0001), when compared with patients with lower SBP. Similar results were found in studies of interferon and bevacizumab treatment when patients who developed grade 2 or more HTN had both improved PFS and OS [25, 96–98].

  • VHL mutation

VHL gene mutation is a key event of tumorigenesis of ccRCC, a highly vascular neoplasm. Although the incidence of this lesion is >90%, it has been postulated that VHL gene status may serve as a predictive biomarker for ccRCC patients in monitoring of response to VEGFtargeted agents. Recently, Choueiri et al. examined 123 ccRCC patients treated with VEGF-targeted monotherapy with sunitinib, sorafenib, axitinib, or bevacizumab [99]. In multivariate analysis, patients with VHL mutational events obtained a significant response rate of 52% (when missense mutations were excluded) compared to those with wild-type VHL who had a response rate of 31% (p = 0.04). Interestingly, no responses were noted in patients with wildtype VHL receiving sorafenib or bevacizumab. However, VHL mutation status did not seem to affect the responses seen in patients treated with potent VEGFR inhibitors sunitinib or axitinib. Other small studies did not provide strong evidence to support the predictive value of VHL mutation as a biomarker. In 13 RCC patients treated with axitinib, no correlation was seen between somatic VHL mutational status and response [100]. In another study, VHL gene status of 78 RCC patients treated with pazopanib was examined, but no association was found between VHL gene status and response [101]. VHL mutational status did not predict treatment benefit in a large phase III study of sorafenib in advanced RCC, although only a minority of patients had known VHL mutation status [102]. Taken together, it remains uncertain whether any correlation exists between VHL status and VEGF therapy response, and definitive studies are awaited.

  • HIF levels

Patel and colleagues used Western blot to measure HIF expression level in 43 ccRCC specimens prior to sunitinib treatment. Twelve (92%) of 13 patients with high HIF-2α expression (>50% compared to cell line control) responded to sunitinib, whereas only 4 (27%) of 15 patients with low expression of HIF-2α showed response to sunitinib [103]. A recent abstract reported that both HIF-1α and HIF-2α (H1H2)-positive expressions were correlated with improvement in PFS and OS, as well as response rate to first-line VEGF TKI therapy [84]. This is somewhat contradictory to a previous study by Klatte et al. that showed patients with higher expression levels of HIF-1α had significantly worse overall survival than those with low expression [83]. Studies further establishing the role of classifying tumors according to HIF expression profile have been hindered by technical limitations of antibody nonspecificity, rapid oxidation, and degradation of HIF proteins in improperly handled specimens. In addition, microdeletions in HIF-1α in some cases can lead to nonfunctional protein that retains the domains and features for antigen detection by traditional immunostaining [104].

  • VEGF/soluble VEGF receptor levels

The value of plasma VEGF levels as a predictive biomarker for antiangiogenesis therapies was addressed in the TARGET trial [105, 106]. High baseline VEGF level was an independent prognostic factor (p = 0.014) as patients with high baseline VEGF had poorer prognosis. This has been validated in several other trials [107–109]. In another trial, both patients with high VEGF levels and low VEGF levels at baseline benefitted from sorafenib therapy, although those with high VEGF levels had a trend toward more pronounced benefit [102].

A phase 2 trial investigating circulating biomarker changes after sunitinib treatment in cytokine-refractory disease demonstrated significant changes in VEGF, sVEGFR-2, and sVEGFR-3 levels in patients with objective tumor response compared with those with stable disease or disease progression [110, 111]. This finding was similar to findings that lower baseline levels of sVEGFR-3 and VEGF-C were associated with longer PFS and better tumor response in patients receiving sunitinib following disease progression on bevacizumab [109]. Similarly, biomarker studies in a phase 2 trial with pazopanib showed that sVEGFR-2 decrease at day 14 of therapy predicted a better outcome in terms of response and PFS [101].

There has also been some evidence of cross talk between the VEGF pathways and CXCR4 pathway, and one small study has suggested that low CXCR4 expression correlates with improved responsiveness to sunitinib therapy [112].

  • Cytokines and angiogenic factors

Thus far, no single cytokine or angiogenic factor has emerged as reliably predictive of response to VEGF-targeted therapy. However, several studies have explored using clusters of cytokines and angiogenic factors (CAFs) to predict response to therapy. One study found a six-marker baseline signature of factors correlated with improved PFS on sorafenib [113]. However, another study showed no difference in PFS or OS with pazopanib treatment based on CAF signature with similar included factors [107].

Predictive biomarkers for mTOR-targeted therapy

  • RCC subtyping

RCC subtyping could be an important predictive biomarker for mTOR inhibitors as well. In contrast to immunotherapies, mTOR inhibitors seem more effective in non-ccRCC.

In a subset analysis of a randomized phase 3 trial, median overall survival of patients with nonccRCC (75% of whom had the papillary subtype) was 11.6 months in the temsirolimus group vs. 4.3 months in the IFN group [114]. The favorable activity of temsirolimus in non-ccRCC is also different from what was observed with the VEGFR antagonists sorafenib and sunitinib, both of which have demonstrated only limited activity against non-ccRCCs [115]. In the RECORD-3 trial, patients with non-ccRCC had worse PFS than ccRCC when treated with either sunitinib or everolimus as first-line therapy [116]. Patients with nonccRCC had a longer PFS on first-line sunitinib than everolimus (7.2 vs. 5.1 months), suggesting that perhaps mTOR inhibitors are not more effective than VEGF inhibitors in the non-ccRCC subtypes. A study randomizing patients with non-clear cell histologies between sunitinib and everolimus showed superiority of sunitinib over everolimus [Ref]. The ongoing phase II ASPEN trial comparing sunitinib and everolimus in non-ccRCC will potentially confirm these findings.

  • PTEN loss

The tumor suppressor gene PTEN (phosphatase and tensin homologue) encodes a dual specific protein and phospholipid phosphatase that is involved in tumorigenesis and is one of the most commonly lost tumor suppressors in human cancer. It has been reported that PTEN loss could be associated with poor prognosis in RCC [117], although interest has focused on PTEN deletion as a potential indicator of response to mTOR inhibitor therapy. However, clinical studies have not substantiated either the prognostic role of PTEN loss in RCC or any correlation between tumor PTEN expression to either tumor response, OS, or PFS in patients treated with temsirolimus [116–118].

  • Phospho AKT/phospho S6K

AKT regulates cell growth and survival mechanisms by phosphorylating a wide spectrum of cellular substrates, including mTOR [119]. Previously, phospho AKT (p-AKT) expression was shown to be correlated with pathologic variables and survival, with higher levels of cytoplasmic p-AKT expression compared with nuclear p-AKT in primary RCC [120]. A recent study found cytoplasmic p-AKT to be significantly correlated to other pathway markers and to nuclear p-AKT in RCC metastases. Unlike primary RCC, p-AKT staining was not prognostic in that cohort of RCC patients [121]. Recent clinical trial data showed that a higher level of p-AKT is associated with both decreased PFS and OS in general in patients with RCC [35].

When mTOR is activated, it phosphorylates two proteins, 4E-BP1 and S6 kinase, to start the cell cycle protein translation process. In primary RCC, phospho S6 kinase (pS6K) expression has been associated with T stage, nuclear grade, incidence of metastasis, and cancer-specific survival [120]. Cho and colleagues investigated VHL mutation, p-AKT, and pS6K expression in archival tumor specimens from 20 RCC patients treated with temsirolimus [122]. Although there was no correlation seen between VHL mutation and treatment response, protein expression of p-AKT and pS6K, two important proteins indicating activity of the mTOR pathway, was positively associated with response to mTOR-directed treatment. This has been further validated in recent studies with correlation of p-4E-BP1 expression with PFS on mTOR therapy [37]. Another study found that phosphorylation of mTOR and S6RP (the 40S ribosomal protein S6 which increases mRNA transcription in response to mTOR activation) was related to response to mTOR therapy (PFS). However, that study did not show a correlation between expression levels of p-4E-BP1 and efficacy of mTOR therapy [38].

  • Genetic biomarkers

A recent case series explored the genetic signatures of several patients who were long-term responders to mTOR inhibitor therapy. Genomic alterations with an activating effect on mTOR signaling were detected in 11 of 14 specimens through alterations in two genes (TSC1 and MTOR) [123].

Predictive biomarkers for other targeted therapies

  • MET

MET germline mutations have been suggested to play a predictive role in response to new MET inhibitor or multikinase inhibitor therapy in papillary RCC. A recent phase II trial showed up to 50% partial response rate in papillary RCC with MET germline mutations compared to 9% in those without mutations [167]. These results need to be validated with further studies, but provide one of the most promising rational biomarker/therapy combinations on the horizon.

[contact-form-7 id=»5168″ title=»Контактная форма 1″]


Добавить комментарий

Войти с помощью: 

Ваш e-mail не будет опубликован. Обязательные поля помечены *