WO2020182995A1

WO2020182995A1 - Vlcfa-containing lipids as prognostic and predictive marker for cancer therapies

Info

Publication number: WO2020182995A1
Application number: PCT/EP2020/056895
Authority: WO
Inventors: Carsten GRÜLLICH; Andreas Mock; Stefanie ZSCHÄBITZ
Original assignee: Universität Heidelberg
Priority date: 2019-03-14
Filing date: 2020-03-13
Publication date: 2020-09-17

Abstract

The present invention relates to metabolic markers which are used as prognostic and/or predictive markers for a cancer therapy and/or for cancer therapy monitoring. The present invention further relates to methods for the prognosis and/or therapy monitoring of cancer, wherein said methods comprise determining the metabolome in a patient sample, in particular the amount of very long-chain fatty acid-containing lipids (VLCFA-containing lipids) in the metabolome.

Description

VLCFA-containing lipids as

prognostic and predictive marker for cancer therapies

BACKGROUND OF THE INVENTION

Metabolomics, besides genomics, epigenomics, transcriptomics, and proteomics, aims to characterize and quantify metabolites in organic liquids or tissues. The metabolome is influenced by both genetics and environmental factors like exercise, smoking, or nutrition.

There is an increasing interest in comparative metabolomic profiling for cancer research, as most published genomic and transcriptomic landscapes fell short of sufficiently explaining disease phenotypes or predicting therapy responses. The emerging hallmarks of cancer metabolism include dysregulation of glucose and amino acid uptake, opportunistic uptake of metabolites from the microenvironment, biosynthesis and NADPH production through glycolysis and citrate acid cycle, an increased need of nitrogen, and gene dysregulation caused by metabolites and other interactions with the microenvironment (Pavlova & Thompson, 2016). In fact, the phenomenon of altered glucose metabolism is routinely used to detect tumor tissues using the FDG-PET technique (Zhu et al ., 2011). Metabolomic signatures measured in serum or urine samples have been identified for various tumor entities as prognostic markers and as markers of early detection but are still not part of clinical routine Hakimi et al., 2016; Li et al, 2014; Sahu et al, 2017; Niziol et al, 2018).

The most commonly used technologies for measuring metabolites are mass spectroscopy (MS), enzyme-linked immunosorbent assay (ELISA), and nuclear magnetic resonance (NMR) spectroscopy. Depending on the method used, throughput, sensitivity, depth of coverage and cost vary significantly. While untargeted metabolomics describes the analysis of all measurable analytes in a given sample, targeted metabolomics is the approach of the target- oriented identification and absolute qualification of known metabolites. The metabolite portfolio encompasses amino acids, vitamins, bile acids, short-, medium-, and long-chain fatty acids (FA), and lipids such as (acyl-) carnitines, glycerophospholipids (GPL), sphingolipids, or lipid precursors.

As much as it is promising to study the tumor metabolism, in the age of cancer immunotherapies a special research focus has been placed on the interplay with the metabolism of immune cells, often referred to as immunometabolism. Deviant metabolites of tumor cells affect their microenvironment. For example, the activity and function of T cells is influenced by restriction of glucose availability or excessive lactate excretion by tumor cells (Haas et al, 2015; Lyssiotis & Kimmelman 2017).

Checkpoint inhibitors (CPI) have added to the treatment armamentarium in metastatic renal cell carcinoma (RCC) and urothelial cancer (UC) in the last years. While in UC the use of CPI in first line treatment has been restricted to programmed cell death ligand 1 (PD-L1) positive tumors in June 2018 following the results of the KEYNOTE-052 trial, the combination of nivolumab and ipilumumab in first line setting for intermediate and high-risk RCC has been FDA-approved and is expected to achieve EMA authorization in the near future. In second line, anti-PDl (programmed cell death protein 1) and anti-PD-Ll antibodies can be applied independent of risk profile in RCC and PD-L1 status in UC.

Positive PD-L1 expression (Teng et al, 2018), tumor mutational burden (TMB) (Yarchoan et al. , 2017), microsatellite status, and tumor infiltrating lymphocytes (TILs) (Galon et al. , 2006) are candidates for response prediction to CPI treatment. However, the predictive power of all these markers varies strongly between studies and cancer types and is also confused by differing methods and reagents used for biomarker testing.

There is emerging data that metabolomic profiles might also be predictive for treatment response to checkpoint inhibitors. In patients with melanoma and RCC treated with nivolumab, levels of kynurenine and adenosine, respectively, differed significantly between responders to CPI and non-responders (Giannakis et al. , 2017; Johnson et al. , 2016). There is a need in the art for providing improved means and methods which allow predicting the outcome or response to treatments methods of cancer, in particular immuno-oncologic treatments and therapies.

SUMMARY OF THE INVENTION

According to the present invention this object is solved by the use of very long-chain fatty acid-containing lipids (VLCFA-containing lipids) as prognostic marker or predictive marker for a cancer therapy and/or for cancer therapy monitoring.

According to the present invention this object is solved by an in vitro , in vivo or ex vivo method for the prognosis and/or therapy monitoring of cancer, comprising the following steps:

(a) providing a sample of a mammalian subject,

(b) determining the metabolome in said sample, and

(c) determining the amount of very long-chain fatty acid-containing lipids

(VLCFA-containing lipids) in the metabolome.

DESCRIPTION OF THE PREFERRED EMBODIMENTS OF THE INVENTION

Before the present invention is described in more detail below, it is to be understood that this invention is not limited to the particular methodology, protocols and reagents described herein as these may vary. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to limit the scope of the present invention which will be limited only by the appended claims. Unless defined otherwise, all technical and scientific terms used herein have the same meanings as commonly understood by one of ordinary skill in the art. For the purpose of the present invention, all references cited herein are incorporated by reference in their entireties.

Concentrations, amounts, and other numerical data may be expressed or presented herein in a range format. It is to be understood that such a range format is used merely for convenience and brevity and thus should be interpreted flexibly to include not only the numerical values explicitly recited as the limits of the range, but also to include all the individual numerical values or sub-ranges encompassed within that range as if each numerical value and sub-range is explicitly recited. As an illustration, a numerical range of "1 to 21" should be interpreted to include not only the explicitly recited values of 1 to 21, but also include individual values and sub-ranges within the indicated range. Thus, included in this numerical range are individual values such as 1, 2, 3, 4, 5 .... 17, 18, 19, 20, 21 and sub-ranges such as from 2 to 10, 8 to 15, etc. This same principle applies to ranges reciting only one numerical value, such as "at least 90%". Furthermore, such an interpretation should apply regardless of the breadth of the range or the characteristics being described.

VLCFA-containing lipids as prognostic and predictive markers

As outlined above, the present invention provides the use of very long-chain fatty acid- containing lipids (VLCFA-containing lipids) as prognostic marker.

As outlined above, the present invention provides the use of very long-chain fatty acid- containing lipids (VLCFA-containing lipids) as predictive marker.

Very long-chain fatty acids (VLCFA) refer to fatty acid with 22 or more carbons.

Long-chain fatty acids (LCFA) refer to fatty acid with an aliphatic chain of 14 to 21 carbons.

VLCFA and LCFA have been described to induce necroptosis and inflammatory signals through a variety of mechanisms (Fritsche 2015; Parisi et al., 2017). Unlike LCFA, VLCFA are not metabolized in the mitochondria but in peroxisomes, where they activate the peroxisome proliferation-activated receptor a (PPARa) signaling, controlling pivotal genes involved in lipid metabolism (Sonoda et al. , 2008).

In particular, the present invention provides the use of VLCFA-containing lipids as prognostic marker or predictive marker for a cancer therapy and/or for cancer therapy monitoring.

Said VLCFA-containing lipids are suitable as marker for the success of a cancer therapy and/or for cancer therapy monitoring.

Preferably, said cancer therapy is an immuno-oncological therapy. Immuno-oncological treatments or therapies refer to cancer treatments whose purpose is to strengthen the body’s own immune response against cancer. The main types of

immunotherapy being used to treat cancer include monoclonal antibodies, immune checkpoint inhibitors, cancer vaccines and other, non-specific immunotherapies.

Preferably, the use according to the present invention comprises response prediction to an immuno-oncologic therapy.

Preferably, the cancer comprises locally advanced and/or metastatic cancer, preferably renal cancer and cancer of the bladder, ureters, and renal pelvis.

In one embodiment, said immuno-oncologic therapy comprises treatment with checkpoint inhibitors (CPI).

For example, said CPI treatment comprises treatment(s) targeting PD1 (programmed cell death protein 1), PD-L1 (programmed cell death ligand 1), CTLA4 (cytotoxic T-lymphocyte- associated protein 4), or LAG3 (lymphocyte-activation gene 3).

For example, CPI treatment can comprise the treatment of programmed cell death ligand 1 (PD-L1) positive tumors.

CPI treatment can also comprise combinations of two or more CPIs (e.g. anti-CTLA4, anti- LAG3, etc.) or combinations with other targeted cancer therapeutics such as kinase inhibitors, anti-VEGF antibodies, PPAR agonists, and/or combinations with chemotherapy.

In one embodiment, said immuno-oncologic therapy comprises treatment with checkpoint inhibitors (CPI)

(a) as CPI monotherapy,

(b) as combinations of one CPI with a second CPI, and/or

(c) as combinations of one or more CPI with other targeted cancer therapeutics (e.g. kinase inhibitors, anti-VEGF antibodies, PPAR agonists etc.) and/or

(d) as combinations of one or more CPI with chemotherapy.

The use according to the present invention preferably comprises determining the metabolome in a sample of a mammalian subject. The sample preferably comprises blood, such as whole blood, plasma or serum.

The mammalian subject is preferably a human.

In a preferred embodiment, the metabolome is determined via mass spectrometry (MS), in particular via MS-based quantification.

Further methods are enzyme-linked immunosorbent assays (ELISA) or nuclear magnetic resonance (NMR) spectroscopy

In one embodiment, the use is in combination with further markers,

preferably high PD-L1 expression, tumor mutational burden (TMB), microsatellite status, HLA (human leukocyte antigen) type, tumor gene expression signatures, and tumor infiltrating lymphocytes and/or combinations thereof.

Methods for the prognosis and/or therapy monitoring of cancer

As outlined above, the present invention provides an in vitro , in vivo or ex vivo method for the prognosis and/or therapy monitoring of cancer. -

Said method comprises the following steps:

(a) providing a sample of a mammalian subject,

(b) determining the metabolome in said sample, and

(c) determining the amount of very long-chain fatty acid-containing lipids (VLCFA-containing lipids) in the metabolome.

Preferably, said cancer therapy is an immuno-oncologic therapy.

The method is preferably for the response prediction to an immuno-oncologic therapy.

Preferably, said cancer comprises locally advanced cancer and/or metastatic cancer, preferably renal cancer and cancer of the bladder, ureters, and renal pelvis.

In step (b), the metabolome is preferably determined via mass spectrometry (MS), enzyme- linked immunosorbent assays (ELISA) or nuclear magnetic resonance (NMR) spectroscopy. In one embodiment, in step (c), the amount of very long-chain fatty acids (VLCFA) in the metabolome is preferably determined via internal standard(s) (absolute quantification).

In one embodiment, in step (c), the amount of very long-chain fatty acids (VLCFA) in the metabolome is preferably determined relative to other measured metabolites in the same sample (relative quantification).

In one embodiment, the method further comprises step (d):

(d) grouping the VLCFA-containing lipids determined in step (b) and/or (c) into clusters,

optionally further comprises step (e):

(e) calculating the centroid of concentrations for the clusters.

Said centroid of concentrations usually corresponds to the median concentration.

For example, one cluster (such as named “cluster 2” below, see also e.g. Figure 3D) comprises phosphatidylcholines (PC), such as the following six: PC aa C38:0, PC aa C42:0, PC aa C42:2, PC ae C40:6, PC ae C42:3 and PC ae C44:6.

For example, one further cluster (such as named“cluster 1” below, see also e.g. Figure 3D) comprises sphingomyelins (SM), such as the following three: SM C26: l, SM C24: l and SM (OH) C22: l.

For example, the centroid concentration of such a cluster can be calculated in step (e), wherein said centroid concentration corresponds to the median concentration of e.g. the six phosphatidylcholines of cluster 2 (see also e.g. Figure 3D and 3E).

Preferably, the sample comprises blood, such as whole blood, plasma or serum; or saliva,

Preferably, the mammalian subject is human.

In one embodiment, said immuno-oncologic therapy comprises treatment with checkpoint inhibitors (CPI). For example, said CPI treatment comprises treatment(s) targeting PD1 (programmed cell death protein 1), PD-L1 (programmed cell death ligand 1), CTLA4 (cytotoxic T-lymphocyte- associated protein 4), or LAG3 (lymphocyte-activation gene 3).

(a) as CPI monotherapy,

(b) as combinations of one CPI with a second CPI, and/or

(d) as combinations of one or more CPI with chemotherapy.

In one embodiment, the method comprises determining or using further markers,

Preferred embodiments

- Abstract

The approval of checkpoint inhibitors (CPI) for a growing number of tumor entities has changed the therapeutic landscape of oncology significantly. Despite their success in a subset of patients, the majority receives an unsuccessful therapy while being exposed to the risks for an autoimmune adverse reaction. Tumor mutational burden, PD-L1 expression, and the extent of tumor-infiltrating lymphocytes have been found to be associated with successful CPI treatment, yet with inconclusive results.

We took a non-invasive metabolomic approach to understand immunotherapy response and failure in 28 urological cancer patients. In total, 134 metabolites were quantified in patient sera before the first, second, and third CPI dose. Modeling the association of metabolite concentrations with CPI response and patient characteristics revealed that all but one predictive metabolite (n=9/10) were very long-chain fatty acid-containing lipids (VLCFA- containing lipids). The best predictive performance was reached by a multivariate model of age and a centroid of VLCFA-containing lipids before the first immunotherapy (sensitivity: 0.850, specificity: 0.825, ROC: 0.935).

- Discussion

The field of immunooncology offers a fascinating new perspective for the treatment of cancer. However, the immune system is extremely complex and tightly regulated, and predictive biomarkers for CPI response are currently one of the most investigated areas of research in this field. The expression of PD-L1, the target of PD-L1/PD1 directed therapies, appears to be of limited predictive value especially in genitourinary tumors. While response rates are higher in PD-L1 positive tumors, PD-L1 negative tumors also respond to these therapies, albeit at lower rates. As another potential biomarker, TMB has been proposed because it may be a surrogate for neo-antigen expression-driven immunogenicity. However, data on TMB are still inconclusive. While a correlation with response was described in urothelial cancer for TMB- high tumors (Balar et al ., 2017), it could not be detected in renal cell cancer (Maia el al ., 2018).

The inventors took a longitudinal non-invasive metabolomic approach to reveal new insights about the mechanisms for immunotherapy response as well as to develop meaningful predictive biomarkers. They found that all but one predictive metabolite (n=9/10) were VLCFA-containing lipids.

In the present study, the best predictive performance was reached by combining patient age and a centroid of VLCFA-containing lipids before the first immunotherapy dose (sensitivity: 0.850, specificity: 0.825, ROC: 0.935). Though the numbers of patients are low, we made use of the necessary statistical means to avoid overfitting in our cohort by use of repeated 10-fold cross-validation.

The following examples and drawings illustrate the present invention without, however, limiting the same thereto. BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1.

(A) Graphical abstract of the study design. Serum was collected from tumor patients before the first, second, and third dose of an immune checkpoint inhibitor therapy.

(B) A total of 134 metabolites were quantified across measurements using the Absolut e//J(/ pi 80 kit (BIOCRATES Life Sciences AG, Innsbruck, Austria).

(C) Kaplan-Meier curve of the overall survival since CPI start stratified by the response groups defined according to the best response to immunotherapy.

(C) Kaplan-Meier curve of the time of ongoing CPI treatment stratified by the response groups defined according to the best response to immunotherapy.

Figure 2.

(A) tSNE plot comparing the metabolome of all patients at the three different time points.

(B) Distribution of measured concentrations per metabolite class stratified by the three different time points.

(C) Boxplot of patient age stratified for immunotherapy response (p=0.013).

(D) Summary heatmaps illustrating the number of significant metabolites associated with response and age stratified for metabolite class (p<0.05).

Figure 3.

(A) Scatter plot of fraction of potential isobars containing VLCFAs calculated for every lipid in the data set over the -log 10 p-value of response to immunotherapy. Nine VLCFA- containing lipids were identified to be predictive at time point 1 (p<0.05).

(B) Association of these nine VLCFA-containing lipids for immunotherapy response at Tl, T2, and T3.

(C) Fraction of potential isobars with VLCFA over the length of the two fatty acids of all measured uGPLs. Linear regression fit with confidence interval (r=0.9).

(D) Correlation heatmap for the nine predictive VLCFA-containing lipids. The two clusters comprise sphingomyelins (SMs; cluster 1) and phosphatidylcholines (PCs; cluster 2), respectively.

(E) ROC statistics of predictive modeling. Figure 4. Correlation heatmap for average correlation between unsaturated glycerophospholipids (GPL) of different fatty acid lengths.

Of note the length of the two fatty acids per GPL is displayed.

EXAMPLES

EXAMPLE 1

1. Materials and Methods

1.1 Study set

Patients with metastatic RCC or UC undergoing CPI treatment at Heidelberg University Hospital were consented for biobanking of blood samples starting from 24.02.2016. Paired serum samples were collected prior to first (28 patients), second (n=18 patients), and third dose (n=16 patients) of CPI treatment, aliquoted and stored at -80°C until used. Medical records of all patients were accessed through the electronic patient charts. From the date of consent, clinical data was prospectively collected. Clinical parameters assessed included histopathologic diagnosis, tumor stage and grading, location of metastases, types and duration of and responses to prior treatment, type and duration of CPI treatment, first and best response to CPI treatment, and side effects of CPI treatment.

Overall survival (OS) was calculated from the date of first CPI dose to the date of death or date of last follow-up, respectively (last assessed on 13.02.2019). As a secondary endpoint we further included time of ongoing CPI treatment. Patients with disease control during CPI treatment (responder) were defined as achieving a complete (CR), partial (PR) remission, or stable (SD) disease. In contrast patients with a progressive disease (PD) were defined as non responders to CPI. Response was evaluated using RECIST 1.1 criteria.

1.2 Ethics Committee Approval and Patient Consent

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. All patients provided written informed consent. Any information connected with the identity of individual subjects was excluded from further sample handling and data processing. The project was registered with the ethics committee of the University of Heidelberg (S023/2016).

1.3 Metabolomic analysis and data preprocessing Mass spectrometric metabolite quantification was performed at BIOCRATES Life Sciences AG, Innsbruck, Austria. All serum samples were analyzed with the Absolut e//J(/" pi 80 kit (BIOCRATES Life Sciences AG, Innsbruck, Austria). The kit enables the absolute quantification of 188 metabolite concentrations by using internal standards. Sample analysis was performed as previously described (Schmerler et al ., 2012) using MS technology (liquid chromatography MS for analysis of amino acids and biogenic amines and flow injection analysis MS for analysis of lipids, acylcamitines, and hexose. The limit of detection (LOD) was set to three times the median signal detected in the zero control samples as previously described (Tomas et al ., 2018). Metabolites were excluded when measurements were below 20% of LOD. The total number of metabolites available after data cleaning across the data set was 134 (9 acylcamitines, 21 amino acids, 10 biogenic amines, 1 hexose, 14 saturated GPLs, 14 sphingolipids, 65 unsaturated GPLs). To calculate the fraction of lipid isomers containing VLCFAs, we used the annotation of potential isobaric and isomeric lipid species measured with the AbsoluteZDO^® pi 80 kit provided by BIOCRATES Life Science AG. Before statistical analysis, measurements were log2 -transformed.

1.4 Statistical analysis

All statistical analyses were conducted within the R software environment (R version 3.4.2; www.r-project.org). The data was stored within a ' SummarizedExperimenf class object enabling coordinated representation of the metabolomic data and associated metadata. Heatmaps were illustrated using the ComplexHeatmap package and boxplots and scatter plots using the ggplot2 package. T-distributed stochastic neighbor embedding (tSNE) was performed using the Rtsne R package. To determine metabolites associated with therapy response, we employed multivariate linear mixed effects modelling. Linear mixed effects models (LMMs) extend the modelling of fixed effects (therapy response and patient age) by random effects (between patient variation). The nlme package offers the functionality of LMMs within R (Pinheiro 2009). In our predictive modelling, we compared the predictive performance of patient age and the median concentration (i.e. the centroids) of the VLCFA- containing lipids belonging to cluster 1 and cluster 2 (compare figure 2D) in univariate and multivariate models. To compensate for overfitting, we applied 10-fold cross-validation, which was repeated 10 times, for each model. In addition, we used a linear (logistic regression) and a non-linear (support vector machine; svm) approach. The predictive modelling was performed with the caret R package. All presented p-values are multiplicity unadjusted. A p-value <0.05 was considered significant. 2. Results

2.1 Study design and patient cohort

Serum samples were collected from tumor patient before the first, second, and third dose of an immune checkpoint inhibitor (time points Tl, T2, T3; Figure 1 A). A total of 134 metabolites were quantified across measurements (Figure IB). The study cohort encompassed 28 patients (Table 1). 25 patients (89.3%) had RCC with predominant clear cell characteristics, two patients (7.1%) had UC, and one patient was diagnosed with both UC and clear cell RCC (3.6%). With the exception of one RCC patient who was treated with atezolizumab and bevacizumab, all patients received treatment with nivolumab. Median age of our patient cohort was 64 years (range 41-79 years); 23 patients were male (82.1%) and 5 patients were female (17.9%). Best response to treatment was PR in 4 patients (14.3%), SD in 8 patients (28.6%), and PD in 15 patients (53.6%). No CR was observed. The overall survival since CPI initiation was significantly higher in responders to CPI (Figure 1C). The same observation hold true for the time of ongoing CPI treatment (Figure ID).

Table 1. Clinicopathological characteristics of patient cohort.

Variable Total [%]

Sex

Male 23 [82.1]

Female 5 [17.9]

Median age [Range] 64 [41-79]

Tumor type

Clear cell renal cell carcinoma 25 [89.3]

Transitional cell carcinoma 2 [7.1]

Transitional cell carcinoma and 1 [3.6]

renal cell carcinoma

Site of metastases

Lung 19 [67.9]

Bone 11 [39.3]

Liver 9 [32.1]

Brain 4 [14.3]

Pancreas 4 [14.3]

Peritoneum 5 [17.9]

Lymph nodes 13 [46.4]

Other 14 [42.9]

Type of treatment

Nivolumab 27 [96.4]

Atezolizumab + Bevacizumab 1 [3.6]

Response to first staging

Partial response 3 [10.7] Stable disease 8 [8,6]

Progressive di sease 16 [57.1 ]

Missing data 1 [3.6]

Best response to treatment

Partial response 4 [14.3]

Stable disease 8 [28.6]

Progressive disease 15 [53.6]

Missing data 1 [3.6]

2.2 Unsupervised analysis of serum metabolomes

As a first investigation, we visualized the metabolome-wide variance in the data set by means of t-distributed stochastic neighbor embedding (tSNE). The tSNE plot revealed greater variance in metabolomic measurements between patients than within the same patient over time (Figure 2 A). The 134 metabolites analyzed were assigned to one of seven metabolite classes (hexoses, amino acids, sphingolipids, unsaturated GPLs, saturated GPLs, biogenic amines, and acylcamitines). Calculating the average concentration for these classes showed that the pool of metabolites remained constant over the three measured time points (Figure 2B).

2.3 Metabolites associated with response to immunotherapy

Before comparing the metabolomes of immunotherapy responders and non-responders, we considered known prognostic factors in our patient cohort. In line with published studies (Betof et al ., 2017), we could identify responders to immunotherapy to be significantly younger than non-responders (Figure 2C). Hence, we applied a multivariate modelling approach revealing differentially abundant metabolites in the response groups independent of patient age. Figure 2D illustrates the number of metabolites significantly associated with response and patient age stratified for the three different time points as a heatmap. Of note, the highest number of metabolites associated with response could be identified before starting the first immunotherapy dose (time point Tl).

2.4 Very long-chain fatty acid-containing lipids are predictive for immunotherapy response

Intriguingly, all but one metabolite (n=9/10) associated with response to immunotherapy at Tl were very long-chain fatty acid-containing lipids (VLCFA-containing lipids; Table 2). This observation became especially clear when plotting the fraction of potential isomers containing VLCFAs for every lipid over the response p-value (Figure 3 A). In addition, we could observe that their predictive performance continuously decreased from Tl to T3 (Figure 3B). To corroborate the role of VLCFAs, we compared the fraction of potential VLCFA- containing lipids with the length of the two fatty acids of all unsaturated GPLs (uGPLs), yielding a strong correlation (Figure 3C, r=0.90). We used uGPLs for this comparison, as they were the largest group of lipids in the data set (65 metabolites). In addition, uGPLs with VLCFAs showed more correlation between each other than with uGPLs containing medium- or short-chain fatty acids (Figure 4). Interestingly, the nine predictive VLCFA-containing lipids revealed two distinct anticorrelated clusters (Figure 3D). The first cluster comprised three sphingomyelins (SM), whereas the second cluster was composed of six phosphatidylcholines (PC).

Table 2. Metabolites associated with immunotherapy response independent of patient age (p<0.05). å FA chain length = sum of fatty acid chain length; isomers > 22C=fraction of potential isomeric lipids containing at least one fatty acid with a chain length greater than or equal to 22; sGPL=saturated glycerophospholipid; uGPL=unsaturated glycerophospholipid; PC=phosphatidylcholine; SM=sphingomyelin. *PC aa C38:0 has two potential isomers with unsaturated fatty acids. log2 (fold-

A chain isomers >

Metabolite class change) p-value length 22C

PC aa C38:0* sGPL 0.488128877 0.027 38 7/12

PC aa C42:0 sGPL 0.566755571 0.031 42 3/4

PC aa C42:2 uGPL 0.437668994 0.013 42 6/6

PC ae C40:6 uGPL 0.420863679 0.047 40 4/6

PC ae C42:3 uGPL 0.615033724 0.005 42 6/8

PC ae C44:6 uGPL 0.512514563 0.024 44 2/2

SM (OH)

C22: l sphingolipid 0.444489484 0.047 22 4/4

SM C24: l sphingolipid 0.461333761 0.025 24 2/2

SM C26: l sphingolipid 0.517973234 0.029 26 1/1

SM C20:2 sphingolipid 0.4872320 0.039 20 0/2

2.5 Predictive modeling of therapy response

Due to the lack of an independent validation cohort, we performed rigorous predictive modelling in our patient set to estimate the predictive performance of the VLFCA-containing lipids. To compensate for overfitting, we applied 10-fold cross-validation, which was repeated 10 times, for each model. In addition, we used a linear (logistic regression) and a non-linear (support vector machine; svm) approach. The best performance was reached by a multivariate logistic regression model including age and the centroid 2, corresponding to the median concentration of the six phosphatidylcholines presented in cluster 2 of Figure 3D (sensitivity: 0.850, specificity: 0.825, ROC: 0.935; Figure 3E, Table 3).

Table 3. ROC statistics of predictive modeling svm: support vector machine.

Formula Statistic _ ROC Sensitivity Specificity age logistic regression 0.77125 0.805 0.645 centroid 1 logistic regression 0.6825 0.805 0.54 centroid2 logistic regression 0.79 0.805 0.64 age + centroid 1 logistic regression 0.825 0.815 0.64 age + centroid2 logistic regression 0.935 0.85 0.825 age + centroid 1 + centroid2 logistic regression 0.865 0.82 0.73 age svm 0.48375 0.89 0.115 centroid 1 svm 0.4225 0.845 0.09 centroid2 svm 0.6825 0.85 0.275 age + centroid 1 svm 0.725 0.79 0.46 age + centroid2 svm 0.78 0.79 0.65 age + centroid 1 + centroid2 svm 0.8925 0.81 0.75

REFERENCES

Balar AV, Castellano D, O'Donnell PH, Grivas P, Vuky J, Powles T, Plimack ER, Hahn NM, de Wit R, Pang L, Savage MJ, Perini RF, et al. First-line pembrolizumab in cisplatin- ineligible patients with locally advanced and unresectable or metastatic urothelial cancer (KEYNOTE-052): a multicentre, single-arm, phase 2 study. Lancet Oncol 2017; 18: 1483-92.

Betof AS, Nipp RD, Giobbie-Hurder A, Johnpulle RAN, Rubin K, Rubinstein SM, Flaherty KT, Lawrence DP, Johnson DB, Sullivan RJ. Impact of Age on Outcomes with

Immunotherapy for Patients with Melanoma. Oncologist 2017;22:963-71.

Fritsche KL. The science of fatty acids and inflammation. Adv Nutr 2015;6:293S-301 S.

Galon J, Costes A, Sanchez-Cabo F, Kirilovsky A, Mlecnik B, Lagorce-Pages C, Tosolini M, Camus M, Berger A, Wind P, Zinzindohoue F, Bruneval P, et al. Type, density, and location of immune cells within human colorectal tumors predict clinical outcome. Science

2006;313 : 1960-4. Giannakis M, Li H, Jin C, Gopal S, Desai K, Horak C, Wind-Rotolo M, Van Allen EM, Clish C, Hodi FS, Garraway LA, Choueiri TK. Metabolomic correlates of response in nivolumab- treated renal cell carcinoma and melanoma patients. J Clin Oncol 2017;35:3036-6.

Haas R, Smith J, Rocher-Ros V, Nadkarni S, Montero-Melendez T, D'Acquisto F, Bland EJ, Bombardieri M, Pitzalis C, Perretti M, Marelli-Berg FM, Mauro C. Lactate Regulates Metabolic and Pro-inflammatory Circuits in Control of T Cell Migration and Effector Functions. PLoS Biol 2015 ; 13 :e 1002202.

Hakimi AA, Reznik E, Lee C-H, Creighton CJ, Brannon AR, Luna A, Aksoy BA, Liu EM, Shen R, Lee W, Chen Y, Stirdivant SM, et al. An Integrated Metabolic Atlas of Clear Cell Renal Cell Carcinoma. Cancer Cell 2016;29: 104-16.

Johnson CH, Spilker ME, Goetz L, Peterson SN, Siuzdak G. Metabolite and Microbiome Interplay in Cancer Immunotherapy. Cancer Res 2016;76:6146-52.

Li B, Qiu B, Lee DSM, Walton ZE, Ochocki JD, Mathew LK, Mancuso A, Gade TPF, Keith B, Nissim I, Simon MC. Fructose- 1,6-bisphosphatase opposes renal carcinoma progression. Nature 2014;513:251-5.

Lyssiotis CA, Kimmelman AC. Metabolic Interactions in the Tumor Microenvironment. Trends Cell Biol 2017;27:863-75.

Maia MC, Almeida L, Bergerot PG, Dizman N, Pal SK. Relationship of tumor mutational burden (TMB) to immunotherapy response in metastatic renal cell carcinoma (mRCC). J Clin Oncol 2018;36:662-2.

Niziol J, Bonifay V, Ossolihski K, Ossolihski T, Ossolihska A, Sunner J, Beech I,

Arendowski A, Ruman T. Metabolomic study of human tissue and urine in clear cell renal carcinoma by LC-HRMS and PLS-DA. Anal Bioanal Chem 2018;410:3859-69.

Parisi LR, Li N, Atilla-Gokcumen GE. Very Long Chain Fatty Acids Are Functionally Involved in Necroptosis. Cell Chem Biol 2017;24: 1445-8. Pavlova NN, Thompson CB. The Emerging Hallmarks of Cancer Metabolism. CellMetab 2016;23:27-47.

Pinheiro J. nlme : Linear and Nonlinear Mixed Effects Models. http://cranr- projectorg/web/packages/nlme/indexhtml 2009.

Sahu D, Lotan Y, Wittmann B, Neri B, Hansel DE. Metabolomics analysis reveals distinct profiles of nonmuscle-invasive and muscle-invasive bladder cancer. Cancer Med

2017;6:2106-20.

Schmerler D, Neugebauer S, Ludewig K, Bremer- Streck S, Brunkhorst FM, Kiehntopf M. Targeted metabolomics for discrimination of systemic inflammatory disorders in critically ill patients. J Lipid Res 2012;53: 1369-75.

Sonoda J, Pei L, Evans RM. Nuclear receptors: decoding metabolic disease. FEBS Lett 2008;582:2-9.

Tomas L, Edsfeldt A, Mollet IG, Perisic Matic L, Prehn C, Adamski J, Paulsson-Beme G, Hedin U, Nilsson J, Bengtsson E, Gonqalves I, Bjorkbacka H. Altered metabolism

distinguishes high-risk from stable carotid atherosclerotic plaques. Eur Heart J

2018;39:2301-10.

Teng F, Meng X, Kong L, Yu J. Progress and challenges of predictive biomarkers of anti PD- 1/PD-Ll immunotherapy: A systematic review. Cancer Lett 2018;414: 166-73.

Yarchoan M, Hopkins A, Jaffee EM. Tumor Mutational Burden and Response Rate to PD-1 Inhibition. N Engl J Med 2017;377:2500-1.

Zhu A, Lee D, Shim H. Metabolic positron emission tomography imaging in cancer detection and therapy response. Semin Oncol 2011 ;38 : 55— 69.

Claims

U30910WO Universitat Heidelberg Claims

1. Use of very long-chain fatty acid-containing lipids (VLCFA-containing lipids) as prognostic marker or predictive marker for a cancer therapy and/or for cancer therapy monitoring.

2. The use of claim 1, wherein said cancer therapy is an immuno-oncologic therapy, and/or wherein said use comprises response prediction to an immuno-oncologic therapy.

3. The use of claim 1 or 2, wherein cancer comprises locally advanced and/or metastatic cancer,

preferably renal cancer and cancer of the bladder, ureters, and renal pelvis.

4. The use of any of claims 1 to 3, wherein said immuno-oncologic therapy comprises treatment with checkpoint inhibitors (CPI), such as treatments targeting PD1 (programmed cell death protein 1), PD-L1 (programmed cell death ligand 1), CTLA4 (cytotoxic T- lymphocyte-associated protein 4), or LAG3 (lymphocyte-activation gene 3),

(a) as CPI monotherapy,

(b) as combinations of one CPI with a second CPI, and/or

(c) as combinations of one or more CPI with other targeted cancer therapeutics (e.g. kinase inhibitors, anti-VEGF antibodies, PPAR agonists) and/or

(d) as combinations of one or more CPI with chemotherapy.

5. The use of any of claims 1 to 4, comprising determining the metabolome in a sample of a mammalian subject,

wherein the sample preferably comprises blood, such as whole blood, plasma or serum; and/or wherein the mammalian subject is human.

6. The use of claim 5, wherein the metabolome is determined via mass spectrometry (MS), enzyme-linked immunosorbent assays (ELISA) or nuclear magnetic resonance (NMR) spectroscopy.

7. The use of any of the preceding claims, wherein the use is in combination with further markers,

8. An in vitro , in vivo or ex vivo method for the prognosis and/or therapy monitoring of cancer, comprising the following steps:

(a) providing a sample of a mammalian subject,

(b) determining the metabolome in said sample, and

(c) determining the amount of very long-chain fatty acid-containing lipids

(VLCFA-containing lipids) in the metabolome.

9. The method of claim 8, wherein said cancer therapy is an immuno-oncologic therapy, and/or wherein the method is preferably for the response prediction to an immuno-oncologic therapy, and/or wherein cancer comprises locally advanced and/or metastatic cancer,

preferably renal cancer and cancer of the bladder, ureters, and renal pelvis.

10. The method of claim 8 or 9, wherein in step (b) the metabolome is determined via mass spectrometry (MS), enzyme-linked immunosorbent assays (ELISA) or nuclear magnetic resonance (NMR) spectroscopy.

11. The method of any of claims 8 to 10, wherein in step (c) the amount of VLCFA- containing lipids in the metabolome is determined

via internal standard(s) (absolute quantification), or

relative to other measured metabolites in the same sample (relative quantification).

12. The method of any of claims 8 to 11, further comprising step (d):

optionally further comprising step (e): (e) calculating the centroid of concentrations for the clusters.

13. The method of any of claims 8 to 12, wherein the sample comprises blood, such as whole blood, plasma or serum; or saliva,

and/or wherein the mammalian subject is human.

14. The method of any of claims 8 to 13, wherein said immuno-oncologic therapy comprises treatment with checkpoint inhibitors (CPI), such as treatments targeting PD1 (programmed cell death protein 1), PD-L1 (programmed cell death ligand 1), CTLA4 (cytotoxic T-lymphocyte-associated protein 4), or LAG3 (lymphocyte-activation gene 3),

(a) as CPI monotherapy,

(b) as combinations of one CPI with a second CPI, and/or

(d) as combinations of one or more CPI with chemotherapy

15. The method of any of claims 8 to 14, comprising determining further markers, preferably high PD-L1 expression, tumor mutational burden (TMB), microsatellite status, HLA (human leukocyte antigen) type, tumor gene expression signatures, and tumor infiltrating lymphocytes and/or combinations thereof.