This invention is related to the area of cancer management. In particular, it relates to methods for testing, stratifying, and treating cancers.
In the immune system, the critical balance between rejection and self-tolerance is maintained by a finely tuned series of co-regulatory receptor-ligand interactions. Recent attention has focused on the programmed death (PD)-1:PD-1 ligand (PD-L1, B7-H1) pathway as a key mediator of tumor immune tolerance. Under physiologic conditions, the inhibitory PD-1 receptor is expressed on activated immune effector cells, including T, B and NK cells. Through interactions with its ligands PD-L1 and PD-L2, normally expressed on antigen presenting cells (APCs), immune effector activity in peripheral tissues during inflammatory processes is self-limited (Keir et al., 2008). This inhibitory system is fundamental to protecting healthy tissues and non-infected cells during clearance of viral and bacterial intracellular infections. However, many human cancers have been shown to express PD-1 ligands, thus inducing immune tolerance locally in the tumor microenvironment (TME) and facilitating tumor cell escape from immune attack (Dong et al., 2002; Topalian et al., 2015). Two general mechanisms promoting expression of PD-L1 on tumor cells have been postulated (Pardoll, 2012). In some tumors, aberrant signaling pathways can constitutively up-regulate PD-L1 expression, a phenomenon termed “innate immune resistance”. In others, the expression of PD-L1 is an adaptive mechanism that occurs in response to inflammatory cytokines produced in the TME during an antitumor immune response (“adaptive immune resistance”, Taube et al., 2012). These mechanisms of PD-L1 expression are not mutually exclusive, i.e., constitutive PD-L1 expression on tumor cells may be further up-regulated by cytokines such as interferon-gamma (IFN-g) (Lyford-Pike et al., 2013).
In renal cell carcinoma (RCC) and some other tumor types, monoclonal antibodies (mAbs) blocking the interaction of PD-1 and its ligands, either by targeting PD-1 (e.g., nivolumab, pembrolizumab) or PD-L1 (e.g., MPDL3280A/atezolizumab, MEDI4736/durvalumab), can restore the efficacy of tumor-specific T cells within the TME leading to substantial and sustained tumor regressions (Brahmer et al., 2010; Brahmer et al., 2012; Topalian et al., 2012; Hamid et al., 2013; Herbst et al., 2014). Approximately 20-30% of patients with advanced RCC experience durable objective tumor regressions following PD-1 pathway blockade (Motzer et al., 2014; McDermott et al., 2015). This has revolutionized treatment algorithms and has focused attention on identifying biomarkers to predict response or resistance to this form of therapy. We previously identified PD-L1 expression on the tumor cell surface as one factor associated with the clinical activity of anti-PD-1 in RCC and other tumors (Topalian et al., 2012). This observation was supported by a recent study of anti-PD-1 (nivolumab) in RCC, showing an objective response rate (ORR) of 31% in patients whose pre-treatment tumor specimens were PD-L1+, and 18% in those that were PD-L1(−) (Motzer et al., 2014).
Notably, a significant number of patients with PD-L1+ RCC still do not respond to PD-1 pathway blockade, suggesting that additional intratumoral factors may influence treatment outcomes. There is a need in the art to develop ways of determining which patients will respond so that they can be treated and which patients will not respond so that they will not be unnecessarily treated. Moreover, there is a need in the art to provide effective methods to treat patients that are identified as non-responders.
One aspect of the invention is a method to predict non-responsiveness to an anti-PD-1 or anti-PD-L1 immunotherapy agent in PD-L1+ renal cell carcinoma (RCC). A sample from a PD-L1+ RCC tumor is tested for expression level of one or more genes selected from the group consisting of aldo-keto reductase family 1, member C3 (AKR1C3); CD24 molecule (CD24); cytochrome c oxidase subunit Va (COX5A); cytochrome P450, family 4, subfamily F, polypeptide 11 (CYP4F11); ectonucleotide pyrophosphatase/phosphodiesterase 5 (ENPP5); coagulation factor II (thrombin) receptor-like 1 (F2RL1); UDP-N-acetyl-alpha-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase 14 (GALNT14); potassium inwardly-rectifying channel, subfamily J, member 16 (KCNJ16); mal, T-cell differentiation protein (MAL); solute carrier family 23 (nucleobase transporters), member 1 (SLC23A1); solute carrier family 37 (glucose-6-phosphate transporter), member 4 (SLC37A4); solute carrier organic anion transporter family, member 3A1 (SLCO3A1); UDP glucuronosyltransferase 1 family, polypeptide A1 (UGT1A1); UDP glucuronosyltransferase 1 family, polypeptide A3 (UGT1A3); and UDP glucuronosyltransferase 1 family, polypeptide A6 (UGT1A6). Expression of protein, mRNA, or both is tested. An increased expression level relative to a control gene whose expression does not substantially vary in response to anti-PD-1 immunotherapy is detected. The increased expression predicts non-responsiveness to anti-PD-1 or anti-PD-L1 immunotherapy.
Another aspect of the invention is a method to predict responsiveness to an anti-PD-1 or anti-PD-L1 immunotherapy agent in PD-L1+ renal cell carcinoma (RCC). A sample from a PD-L1+ RCC tumor is tested for expression level of one or more genes selected from the group consisting of BTB and CNC homology 1, basic leucine zipper transcription factor 2 (BACH2); bone morphogenetic protein 1 (BMP1); calcium channel, voltage-dependent, beta 1 subunit (CACNB1); chemokine (C—C motif) ligand 3 (CCL3); E2F transcription factor 8 (E2F8); interleukin 11 receptor, alpha (IL11RA); latent transforming growth factor beta binding protein 1 (LTBP1); myosin light chain kinase 2, skeletal muscle (MYLK2); nuclear factor of activated T-cells, cytoplasmic, calcineurin-dependent 1 (NFATC1); paired-like homeodomain 2 (PITX2); plectin 1, intermediate filament binding protein 500 kDa (PLEC); protein phosphatase 2 (formerly 2A), regulatory subunit B (PPP2R3B); tumor necrosis factor receptor superfamily, member 19 (TNFRSF19); uncoupling protein 3 (mitochondrial, proton carrier) (UCP3), nuclear gene encoding mitochondrial protein (UCP3); and Wolf-Hirschhorn syndrome candidate 1 (WHSC1). Expression of protein, mRNA, or both is tested. Increased expression relative to a control gene whose expression does not substantially vary in response to anti-PD-1 immunotherapy is detected. The increased expression predicts responsiveness to anti-PD-1 or anti-PD-L1 immunotherapy.
Yet another aspect of the invention is a method to treat a PD-L1+ RCC tumor that is non-responsive to anti-PD-1 or anti-PD-L1 immunotherapy. An inhibitor of one or more proteins is administered to the RCC patient. The one or more proteins are selected from the group consisting of aldo-keto reductase family 1, member C3 (AKR1C3); CD24 molecule (CD24); cytochrome c oxidase subunit Va (COX5A); cytochrome P450, family 4, subfamily F, polypeptide 11 (CYP4F11); ectonucleotide pyrophosphatase/phosphodiesterase 5 (ENPP5); coagulation factor II (thrombin) receptor-like 1 (F2RL1); UDP-N-acetyl-alpha-D-galactosamine:polypeptide N-acetylgalactosaminyltransferase 14 (GALNT14); potassium inwardly-rectifying channel, subfamily J, member 16 (KCNJ16); mal, T-cell differentiation protein (MAL); solute carrier family 23 (nucleobase transporters), member 1 (SLC23A1); solute carrier family 37 (glucose-6-phosphate transporter), member 4 (SLC37A4); solute carrier organic anion transporter family, member 3A1 (SLCO3A1); UDP glucuronosyltransferase 1 family, polypeptide A1 (UGT1A1); UDP glucuronosyltransferase 1 family, polypeptide A3 (UGT1A3); and UDP glucuronosyltransferase 1 family, polypeptide A6 (UGT1A6). An anti-PD-1 or anti-PD-L1 immunotherapy agent is also administered to the RCC patient.
An additional aspect of the invention is a method to treat a patient with a PD-L1+ RCC tumor that is non-responsive to anti-PD-1 or anti-PD-L1 immunotherapy. An enhancer of a protein is administered to the RCC patient. The protein is selected from the group consisting of BTB and CNC homology 1, basic leucine zipper transcription factor 2 (BACH2); bone morphogenetic protein 1 (BMP1); calcium channel, voltage-dependent, beta 1 subunit (CACNB1); chemokine (C—C motif) ligand 3 (CCL3); E2F transcription factor 8 (E2F8); interleukin 11 receptor, alpha (IL11RA); latent transforming growth factor beta binding protein 1 (LTBP1); myosin light chain kinase 2, skeletal muscle (MYLK2); nuclear factor of activated T-cells, cytoplasmic, calcineurin-dependent 1 (NFATC1); paired-like homeodomain 2 (PITX2); plectin 1, intermediate filament binding protein 500 kDa (PLEC); protein phosphatase 2 (formerly 2A), regulatory subunit B (PPP2R3B); tumor necrosis factor receptor superfamily, member 19 (TNFRSF19); uncoupling protein 3 (mitochondrial, proton carrier) (UCP3), nuclear gene encoding mitochondrial protein (UCP3); and Wolf-Hirschhorn syndrome candidate 1 (WHSC1). An anti-PD-1 or anti-PD-L1 immunotherapy agent is also administered to the RCC patient.
According to one aspect of the invention a combination regimen is provided that comprises:
As yet another aspect of the invention a second combination regimen is provided. This combination regimen comprises:
According to another aspect of the invention a method is provided. The method comprises the steps of:
According to yet another aspect of the invention a method is provided that comprises:
According to still another aspect of the invention a method is provided that comprises:
According to another aspect of the invention a method is provided that comprises:
According to an additional aspect of the invention a kit is provided for predicting clinical response or non-response to anti-PD-1 or anti-PD-L1 antibody therapy in kidney cancer. The kit comprises:
These and other aspects which will be apparent to those of skill in the art upon reading the specification provide the art with methods and tools for testing, stratifying, and treating cancers.
PD-L1 expression by tumor cells prior to treatment correlates highly with response to anti-PD-1 monotherapy (for example, nivolumab (Bristol-Myers Squibb), pembrolizumab (Merck)) and anti-PD-L1 therapy (for example, MPDL3280A (Genentech/Roche)). Nonetheless, the majority of patients with PD-LI(+) tumors do not respond to PD-1 pathway blockade. The inventors have identified distinct gene profiles associated with differential response to nivolumab in patients with PD-L1+ kidney cancer. In particular, a strong up-regulation of genes involved in metabolic functions and pathways was found in patients not responding to the therapy. Additionally, a down-regulation of genes involved in cellular migration functions was found in the same group of patients (non-responders). Specific biomarkers can be used to stratify responders from non-responders for PD-1 pathway blocking drugs. Additionally, the biomarkers are therapeutic targets for anti-PD-1 combination therapy, and companion diagnostic products for such combination therapies.
Any means of determining expression of the mRNA or protein may be used. One can use the any of the markers identified and reported here. There are a host of assays available to those of skill in the art for determining expression, and these can be used as is convenient to the skilled worker. Such tests include using expression arrays for RNA, cDNA, or protein analysis, qRT-PCR, ELISA assays, in situ hybridization assays, tagless assays, such as using mass spectrometry and MRI, Northern or Western blots, serial analysis of gene expression, bead emulsion amplification, immunohistochemistry, and immunofluorescence. The particular choice of assay technology is not critical. The test samples may be tissue samples, whole cells, isolated RNA, cDNA, isolated protein, for example. The test samples may be in suspension or solution or they may be affixed to a solid support. Similarly any specific reagents for detecting expression products may be in solution or affixed to a solid support. For examples, tissue samples may be on slides. Tissue samples may be prepared in any manner, including but not limited to formalin-fixed, paraffin embedded tissues, fresh frozen tissues, dissociated specimens, such as fine needle aspirates or enzymatically digested fresh solid tumors. Nucleic acid probes may be on beads or chips or nanoparticles. The amino acid sequences and RNA sequences for these markers are known and can be obtained from GenBank.
Reporter systems can be any that are known in the art, as is convenient to the skilled worker. Reporter systems may involve chromagens, radioactive isotopes, or fluorochromes, for example. Dyes may be used for staining proteins or nucleic acids. Specific primers and probes may be used to detect nucleic acid expression products. Primary antibodies used in assays may be directly labeled, or may be detected by a secondary antibody that is directly labeled. Secondary antibodies can be directed against the constant portion of the antibody; they may be anti-isotype antibodies. Other secondary detection systems such as a cascade system may also be used. Such systems may amplify a signal, for example by nucleic acid amplification.
Kits may contain specific instructions for performing any of the assays that are described here or that can be used to detect the markers for kidney cancer responsiveness. The instructions may be in any format, included printed or recorded to an electronic medium or referencing to information on the internet. Kits are typically a single container that comprises one or more elements. The elements may be mixed or separate. The kits may comprise a solid support to which specific reagents are linked or can be linked. The kit may comprise one or more reagents of a certain category or a mixture of categories, such as both an antibody and a nucleic acid probe. The kit may contain specific reagents for each of 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, or 27 markers. The kit may contain more than one specific reagent for any of the markers. Some of the markers are associated with increased expression in responders and some are associated with increased expression in non-responders. Combinations of such types of markers can be used or just one or the other type can be used. Reagents may be in any physical state, such as dried, frozen, in solution, or aerosolized. Useful ancillary reagents may also be included in the kits, including tubes, plates, enzymes, such as reverse transcriptase or DNA polymerase. Antibodies specific for PD-1 or PD-L1 may also be included for analytical or preparatory uses. Cascade systems may be used to detect primary reagents and these can be included in the kits as well.
Test samples may be from any type of cancer or body fluid. Cancer cells may be obtained from plasma, urine, or stool, for example. Alternatively they can be obtained from biopsy samples. Any type of kidney cancer may be tested, including renal cell carcinoma. Other tumor types may be tested as well, including without limitation, bone cancer, bowel cancer, colon cancer, melanoma, basal cell carcinoma, lymphoma, glioblastoma, oligodendroglioma, astrocytoma, lung cancer, esophageal cancer, breast cancer, testicular cancer, prostate cancer, pancreatic cancer, ovarian cancer, uterine cancer, cervical cancer, gastric cancer.
Endogenous genes or proteins that are used as references or controls will generally be selected for their constancy of expression. A range of expression can be pre-defined within which the control genes might vary. It is preferred that the control gene have a small variation in expression, if any, and that this variation not correlate with response to anti-PD-1 immunotherapy. These genes are sometimes referred to as house-keeping genes. Examples of suitable genes or proteins are the 18S rRNA, beta-actin, PTPRC/CD45, and GUSB. Any can be used as is congenial for the purpose.
Antibodies as employed in the invention may be modified. For example, they may be humanized to reduce immunological rejection. They may have modified glycosylation due to the cell type in which they have been produced. They may be truncated or fused to other antibodies or proteins. They may be bifunctional antibodies or single chain antibodies. They may be engineered to be better discriminators, such as by affinity maturation. Any such modifications from the natural product may be used.
For some assays it may be useful to preselect or simultaneously analyze samples for their expression of PD-L1 or PD-1. Any assay may be used for this purpose as is convenient. However, one need not prescreen. The level which is determined for such expression may vary with the assay used. Additionally, such expression may be used to dissect portions of a tissue sample for those that express or do not express these markers or for those that express more or less of these markers. This may enhance the discrimination of the marker expression determination of the invention.
A combination regimen is a course of therapy in which two or more agents are administered, whether in combination in a single composition, separately in a serial fashion, or simultaneously by different routes. The two or more agents are administered to the same individual. Inhibition of a target that is overexpressed in non-responders would expand the population of responders. Similarly, inhibition of such targets in responders or weak responders can be used to increase the response intensity or duration. Conversely, enhancement of expression or activity of targets that are under-expressed in non-responders or over-expressed in responders will expand the population of responders. Similarly, enhancement of expression or activity of such targets can be used to increase the response intensity or duration in responders or weak responders. Inhibitory agents of the markers can be antagonist antibodies or chemical entities. Inhibitory agents known in the art for these protein markers can be used in the combination regimen. Antibodies may comprise all or part of an antibody molecule so long as it retains specific binding of its cognate antigen. Other moieties may be attached by translational or post-translational means to antibodies molecules. For example, a toxin or a reporter moiety may be attached to an antibody. Enhancers may include, for example, expression vectors for the marker or chemical entities. When using antibodies in a therapeutic manner, whether to inhibit or enhance a treatment, antibodies will be selected for their ability to access their targets. Thus antibodies that bind to surface proteins are preferred. Such antibodies will preferably bind to epitopes of a surface protein that are accessible to the antibody from the extracellular milieu.
If the target marker is a receptor, for example, the ligand or a synthetic ligand molecule can be used as an agonist (stimulator). For example, the natural ligand for TLR3 is double stranded DNA, and a chemical mimic (poly I:C) can be used to stimulate this ligand. Additionally, agonist monoclonal antibodies can provide stimulation when they bind to their target. Those of skill in the art can routinely make synthetic ligands and antibodies with agonistic properties.
When an expression signature is detected that indicates that the patient will be a responder (or does not indicate that the patient will be a non-responder) to therapy that involves blockade of PD-1 and/or PD-L1, then such therapy may be administered. If the expression signature indicates that the patient will be a non-responder, then such therapy may not be administered and alternative therapies that act by other mechanisms may be considered and prescribed. Examples of therapies that involve blockade of PD-1 and/or PD-L1 are monoclonal antibodies to either the receptor or the ligand, recombinant proteins such as AMP-224, a PD-L2/Fc fusion protein, peptides, anti-sense RNA or anti-sense expression constructs, or small molecule inhibitors. See, e.g., US 20130309250, US 20140205609, the disclosures of which are expressly incorporated herein. Exemplary therapeutics include pembrolizumab (formerly known as lambrolizumab) (MK-3475), nivolumab (BMS-936558), pidilizumab (CT-011), AMP-224MEDI4736, MPDL3280A, and BMS-936559 (also known as MDX-1105). Ipilimumab or tremelimumab, inhibitors of CTLA4, may be administered in combination with an anti-PD-1 or anti-PD-L1 agent.
The expression signature of the cancer cells may be used to stratify patients. Patients may be put into groups or cohorts of similarly signatured patients. Cohorts may be used, for example, for testing new therapies, for studying long term outcomes of therapies or disease progression, for testing new ways of administering therapies, for testing new ways to monitor or manage disease.
Expression of the immunosuppressive ligand PD-L1 in pre-treatment tumor biopsies has been shown to correlate with favorable clinical outcomes to PD-1 and PD-L1 blocking therapies (Topalian et al., 2012; Herbst et al., 2014; Garon et al., 2015). This can be understood by viewing PD-L1 as a surrogate marker for an immune-reactive tumor milieu, since inflammatory cytokines such as IFN-g are major drivers of PD-L1 expression on tumor and stromal cells. In this model, blocking the PD-1/PD-L1 interaction unleashes an immune response that was already properly trained and poised to attack cancer cells, but was being held in check by this immunosuppressive pathway. Despite the therapeutic impact of this approach in many patients with certain cancer types, the majority of patients with PD-L1+ tumors still do not respond to anti-PD-1/-PD-L1 drugs. This implicates the involvement of additional factors in the tumor immune microenvironment, and/or factors intrinsic to tumor cells themselves, conspiring to maintain local tumor immunosuppression. The current study attempts to identify such factors by exploring the gene expression landscape of PD-L1+ kidney cancers derived from patients with divergent clinical outcomes after anti-PD-1 therapy, and identifies groups of metabolic and immunologic factors associated with adverse or favorable clinical outcomes, respectively. Our findings suggest that an intricate balance between metabolic and immune factors may determine the eventual outcome of anti-PD-1 therapy in patients with RCC.
RCC has been characterized as a metabolic disease, with the signature up-regulation of factors adapting to hypoxia and functioning to meet the bioenergetic demands of cell growth and proliferation (Linehan et al., 2010). We here describe a metabolic shift in RCCs resistant to anti-PD-1 therapy, with overexpression of molecules associated with glucuronidation and the transport of solutes and nutrients. This shift mirrors the Warburg metabolic phenotype which has been associated with poor prognosis in primary RCC (Cancer Genome Atlas Research, 2013). We found that UGT1A6, whose principal role is to promote cellular clearance of toxins and exogenous lipophilic chemicals (Wells et al., 2004), was the single most highly overexpressed molecule associated with anti-PD-1 treatment resistance, and that other UGT1A family members were also up-regulated. Although this may simply reflect an activated cell phenotype and further investigation is needed, one might hypothesize that the heightened clearance of toxins from tumor cells may specifically allow them to evade immune attack mediated by secreted molecules such as lytic factors (e.g., perforin, granzyme B) and cytokines. Indeed, it has been shown that the coordinate regulation of UGT1A family members and drug/solute transporters represents an essential component of the chemical “defensome” providing cells with protection against various external stressors (Wells et al., 2004). Interestingly, UGT1A6 mRNA expression does not appear to correlate with overall survival in the general population of patients with RCC, based on an analysis of published TCGA data derived from a large patient cohort. This suggests a specific intersection between UGT1A6 and other metabolic factors with immunologic phenomena mediated by anti-PD-1.
The general approach to identifying markers predicting clinical response to PD-1-targeted therapies has focused on immunologic factors in the TME, such as modulatory receptors and ligands (e.g., PD-L1, PD-L2, LAG-3, TIM-3), T cell infiltrates (intensity and subsets), and soluble molecules (lymphokines, chemokines). However, our data suggest that a deeper level of investigation is warranted for individual tumor types against which these new therapies are being applied with some success. For instance, in melanoma, a recent report associated over-expression of beta-catenin with decreased infiltration of tumor-specific T cells, postulated to be due to a barrier effect (Spranger et al., 2015). Our study suggests that certain metabolic factors in RCC, a cancer type in which metabolic aberrations are a hallmark, may support anti-PD-1 resistance mechanisms uniquely characteristic of this tumor. Future studies will address the potential processes underlying these mechanisms. A greater knowledge of such mechanistic markers may reveal new therapeutic targets for combination regimens based on PD-1 pathway blockade, and useful biomarkers for selecting patients most likely to respond to these therapies.
The above disclosure generally describes the present invention. All references disclosed herein are expressly incorporated by reference. A more complete understanding can be obtained by reference to the following specific examples which are provided herein for purposes of illustration only, and are not intended to limit the scope of the invention.
Case Selection.
Pre-treatment tumor expression of PD-L1 has been shown to correlate with favorable clinical outcomes following PD-1 or PD-L1 blocking therapies, yet the majority of patients with PD-L1+ tumors do not respond to treatment. In order to understand mechanisms underlying the failure of anti-PD-1 targeted therapies in patients with positive tumor expression of PD-L1, patients with advanced metastatic renal cell cancer (RCC; kidney cancer) who had received nivolumab (anti-PD-1) monotherapy at Johns Hopkins and whose treatment outcomes were known were selected for analysis. Pre-treatment tumor biopsies were assessed for PD-L1 expression, using an immunohistochemistry assay developed in our laboratories. Formalin-fixed, paraffin-embedded (FFPE) biopsy material was retrieved from the Johns Hopkins Medical Center archives or from referring hospitals. Incisional or excisional tumor specimens or core needle biopsies, but not fine needle aspirates, were allowable for analysis. A PD-L1 positive (PD-L1+) specimen was defined as having ≧5% of tumor cells with cell surface PD-L1 expression, consistent with our previous publications (Brahmer et al., J Clin Oncol 2010; Taube et al., Science Transl Med 2012; Topalian et al., NEJM 2012; Taube et al., Clin Cancer Res 2014). A total of 13 PD-L1+ RCC specimens from 13 patients, including patients who did or did not respond to nivolumab therapy, were selected for further analysis. Macroscopic areas of PD-L1+ tumor were excised from FFPE tissue sections on glass slides by scraping with a sterile scalpel, while microscopic focal areas of PD-L1 expression were removed by laser capture microdissection (Taube et al., Science Transl Med 2012). RNA was isolated and reverse transcription reactions were conducted with the High Pure RNA Paraffin Kit (Roche, Indianapolis, Ind.) and qScript cDNA SuperMix (Quanta Biosciences, Gaithersburg, Md.), respectively.
Following RNA isolation, several molecular analyses were performed, as described below.
1) Analysis of Immune-Inhibitory Networks in RCC Specimens Using a Custom Quantitative RT-PCR (qRT-PCR) Multiplex Array.
Expression of immune-related molecules often found in the tumor microenvironment was analyzed by multiplex qRT-PCR using Custom Taqman Low-Density Array (TLDA) microfluidic cards (Applied Biosystems) containing 60 unique gene targets, including those that were previously found to be associated with PD-L1 expression in melanoma (Young et al., AACR 2013, abstr.). In addition to the 60 genes of interest, 4 endogenous controls genes (PTPRC/CD45, GUSB, 18S rRNA, and B-actin) were included in the array for a total of 64 genes. The pre-amplification and Taqman PCR reactions followed the Applied Biosystems Custom Taqman PreAmp Pool protocol for microfluidic cards (http://tools.lifetechnologies.com/content/sfs/manuals/cms_088987.pdf). TLDA card reactions were acquired by a QuantStudio 12k Flex real-time PCR system in the Johns Hopkins University Genetics Core Research Facility, and data were analyzed with Expression Suite Software (v. 1.0.4, Applied Biosystems). Samples were grouped according to clinical response [responder (R) vs. non-responder (NR), according to RECIST criteria], where responders had partial (PR) or complete (CR) tumor regressions, and non-responders (NR) had stable or progressive disease (Topalian et al., NEJM 2012).
To normalize the amount of source RNA, PTPRC/CD45 transcript was used as internal reference reflecting immune cell content in each specimen. Each targeted transcript was evaluated using the comparative Ct method for relative quantification (ΔCt) to the amount of the common reference gene. The results showed that none of the immune genes previously associated with positive expression of PD-L1 in melanoma (comparing PD-L1 positive vs. negative tumors) was significantly associated with clinical outcomes in RCC specimens that were pre-selected for positive expression of PD-L1. Similar results were obtained by using B-actin, 18S rRNA, or GUSB as the reference gene (not shown).
2) Analysis of Molecular Pathways Associated with Clinical Response to Anti-PD-1 in PD-L1+ RCC.
In order to assess differential gene expression in PD-L1+ RCC according to response or non-response to anti-PD-1 therapy, global gene expression profiling of tumor specimens from 11 RCC patients (two of the 13 initial specimens were not included due to insufficient RNA) was performed by using a whole genome DASL (cDNA-mediated Annealing, Selection, extension, and Ligation; Illumina) microarray including >29,000 gene targets. This BeadChip features content covering more than 29,000 annotated genes derived from RefSeq (Build 36.2, Release 38). Global gene expression was analyzed using BRBArrayTools developed by the Biometric Research Branch, NCI (http://linus.nci.nih.gov/BRB-ArrayTools.html) and Partek Genomics Suite (St. Louis, Mo.). The transcriptional profiles derived from R (n=4) and NR patients (n=7) were compared using class comparison analysis. This analysis identified 234 transcripts differentially expressed between the two groups, using an expression fold-change of ≧1.5 and p value ≦0.01 by Student's T test (
3) qRT-PCR Validation of 60 Genes Differentially Expressed in PD-L1+ Tumor Specimens from RCC Patients with Divergent Clinical Outcomes.
Following global gene expression profiling, validation of differential gene expression was performed by multiplex qRT-PCR using Custom Taqman Low-Density Array (TLDA) microfluidic cards (Applied Biosystems). Among the 234 genes previously found to be differentially expressed patients with divergent clinical outcomes by DASL global microarray, 60 unique gene targets were selected for screening. Several criteria were adopted for gene selection including:
In addition to the 60 genes of interest, 4 endogenous control genes (PTPRC/CD45, GUSB, 18S and B-actin) were included in the array, for a total of 64 genes. To normalize the amount of source RNA, GUSB transcript was used as an internal reference. Each targeted transcript was validated using the comparative Ct method for relative quantification (ΔCt) to the amount of the common reference gene. Results confirmed the differential expression of many of the 60 genes selected based on whole genome expression profiling. Genes over-expressed in RCC non-responders included those involved in metabolic pathways and carbohydrate transport, as well as molecules involved in mitochondrial functions and certain immunological pathways (
A complete list of the 64 genes included in the custom multiplex qRT-PCR assay that was constructed based on RCC DASL array data is provided below:
EXPERIMENTAL PROCEDURES—The following procedures were used in the experiments that are described below.
Tumor Specimens
Consenting patients with unresectable metastatic RCC received nivolumab anti-PD-1 monotherapy at the Johns Hopkins Kimmel Cancer Center, on one of four clinical trials (NCT00441337, NCT00730639, NCT01354431, NCT01358721) under approval by the Johns Hopkins Institutional Review Board. Patients were classified as responders (R) or non-responders (NR) to anti-PD-1 therapy based on radiographic staging according to Response Evaluation Criteria in Solid Tumors (RECIST) (Therasse et al., 2000). Non-responders included patients whose disease progressed as well as those with stable disease (SD). Responding (R) patients included patients with complete or partial responses (CR, PR). From among 35 potential pretreatment tumor specimens derived from 26 patients, 21 formalin-fixed paraffin-embedded (FFPE) tumor specimens were available for study. They were characterized for PD-L1 expression by immunohistochemistry (IHC) as previously described (Taube at al., STM, 2012, Topalian et al NEJM 2012). In brief, a tumor specimen was defined as PD-L1+ if ≧5% of tumor cells showed cell surface staining with the murine anti-human PD-L1 mAb 5H1 (from Lieping Chen, Yale University). Among the 21 tumors examined, 14 specimens (67%) from 13 unique patients demonstrated PD-L1 expression. One specimen from each of 13 patients was selected for further analysis. Only 12 specimens yielded sufficient RNA for gene expression analyses
Immunohistochemical Analysis
Serial 5 um-thick sections from PD-L1+ FFPE tumor specimens were stained for expression of selected markers with specific mAbs. The molecules CD3, CD4, CD8, CD68 and FoxP3 were detected with standard automated MC methods. MC for PD-1, PD-L2, and LAG-3 was performed as previously described (Taube et al., 2014 and 2015). TIM-3 was detected with a primary murine anti-human TIM-3 mAb (clone F38-2E2; Biolegend, San Diego, Calif.) at 1.5 ug/ml, after antigen retrieval for 10 min in citrate buffer, pH 6.0 at 120° C.; a secondary anti-mouse IgG1 was used at 1.0 ug/ml, amplification was performed with the CSA kit (DAKO #1500, Carpinteria, Calif.), and visualization was accomplished with DAB (Sigma, St. Louis, Mo.). UGT1A6 expression was detected using the same antigen retrieval conditions, with application of a primary rabbit anti-human UGT1A6 mAb (clone EPR11068, Abcam, Cambridge, Mass.) at 1.25 ug/ml (1:250), followed by application of the Novolink anti-rabbit polymer detection system (RE7112, Leica, Buffalo Grove, Ill.) and visualization with DAB.
The intensity of immune cell infiltrates was scored as mild, moderate or severe, as previously described (Taube et al., 2014). CD3 and CD68 immunostains were performed on each specimen and were used to guide assignment of an intensity score for immune infiltrates and to determine which cell types were expressing PD-1 ligands. Intratumoral CD4:CD8 ratios were estimated at 1:1, 1:2, 1:4, or 2:1. The proportion of TILs expressing PD-1, LAG-3, TIM-3 or FoxP3 was scored as “none”, “focal” (isolated, <5% of lymphocytes), “moderate” (5-50% of TILs), or “severe” (>50% of TILs). PD-L2 expression on infiltrating immune cells (TILs or histiocytes) was scored on the same semi-quantitative scale of “none”, “focal”, “moderate” or “severe”. Positive UGT1A6 staining in tumor cells was scored at 5% intervals.
Multiplex qRT-PCR Assays and Statistical Analyses
PD-L1+ tumor areas, identified with IHC on neighboring tissue sections, were either manually dissected by scraping with a scalpel, or were laser-capture microdissected from 5-um FFPE tissue sections as previously described (Taube et al., 2012 and 2015). (
Whole Genome Expression Profiling and Analysis 1601 Global gene expression in tumor specimens from anti-PD-1 responders (R, n=4) and non-responders (NR, n=7) was measured by DASL (cDNA-mediated Annealing, Selection, extension, and Ligation) assays arrayed on the Illumina HumanHT-12 WG-DASL V4.0 R2 expression beadchip, per the manufacturer's specifications (Illumina, San Diego, Calif.). This platform detects 29,670 annotated transcripts and is designed to detect partially degraded mRNAs such as typically found in FFPE tissue specimens. Briefly, total RNA was reverse transcribed into 1st-strand cDNA and then annealed with an assay-specific oligo pool for 2nd-strand cDNA synthesis. The cDNA was further amplified by PCR using universal primers. PCR products were then purified and denatured to obtain labeled single-strand DNA for DASL array hybridization, after which the BeadChip was washed and scanned to acquire the intensity data. A single intensity (expression) value for each Illumina probe on the DASL array was obtained using Illumina GenomeStudio software with standard settings and no background correction. For each sample, the expression values for all the probes were scaled to have median 256 (28) and were then log (base 2) transformed before performing statistical analysis. Gene expression data were deposited in NCBI's Gene Expression Omnibus and are accessible through GEO Series accession number GSE67501. Gene expression was further analyzed using BRBArrayTools developed by the Biometric Research Branch, NCI and Partek Genomics Suite (St. Louis, Mo.). The transcriptional profile derived from R vs. NR patients was compared using class comparison analysis and Student's t-test (p value ≦0.01, fold change ≧1.5). Lists of genes passing specified distinguishing criteria were examined for significant enrichment in gene annotation categories, and in functionally related categories including KEGG pathways, using the DAVID web tool (Huang et al., 2007). Principal component analysis (PCA) was also conducted to compare gene expression in tumors from R vs. NR, using Partek Software (St. Louis, Mo.). PCA is defined as a statistical procedure that uses an orthogonal transformation to convert a set of observations of possibly correlated samples into a set of values of linearly uncorrelated variables called principal components (PCs) (Jolliffe, 2002).
RCC Cell Lines
The twelve cultured RCC lines used in this study included four that were established from operative kidney cancer specimens (RCC-MO, RCC-WH, RCC-WA and RCC-BR; obtained from Dr. James Yang, National Cancer Institute, Bethesda, Md.) and 8 commercially available lines [ACHN, UO-31, TK-10, A498, RXF-393, SN12C, 786-0 and Caki-1; American Type Culture Collection, Manassas, Va. (http://www.atcc.org/)]. The former were cultured in DMEM+10% heat-inactivated FBS with 10% tryptose phosphate broth, 1% HEPES buffer, 1% L-glutamine, 1% penicillin/streptomycin, 1% insulin/transferrin/selenium, and 1% sodium pyruvate. The latter were cultured in RPMI 1640+10% heat-inactivated FBS supplemented with 10 mM HEPES buffer and 1% antibiotic/antimycotic solution (Life Technologies, Grand Island, N.Y.). All cell cultures were maintained at 37° C., 5% CO2 and confirmed to be mycoplasma-free with the Venor®GeM Mycoplasma Detection kit (Sigma Aldrich). In some experiments, cells were cultured in the presence of IFN-g 250 IU/ml (Biogen, Cambridge, Mass.) for 48 hrs prior to assessing gene expression.
In Silico Correlation of Gene Expression with Overall Survival in RCC
To investigate potential associations between differentially expressed genes identified in this study with RCC clinical stage and the overall survival of patients with RCC, RNA sequencing data from The Cancer Genome Atlas project (TCGA), including 444 clear cell RCC samples and 72 matched normal kidney samples, were used for in silico analysis. Level 3 RSEM normalized data were downloaded from the TCGA Data Portal (https://tcga-data.nci.nih.gov/tcga/). Analysis was performed using R/Bioconductor software with the survival package and custom routines for data analysis (Gentleman et al., 2004). Association of gene expression level with tumor stage was tested by fitting a linear model using continuous expression level of a gene and the tumor stage as a numeric value (Wilkinson, 1973). The linear model coefficient and p-value were adjusted by Benjamini-Hochberg procedure (false discovery rate, FDR) (Hochberg and Benjamini, 1990). For survival analysis, the estimated logarithm of the hazard ratio and p-value were computed from the Cox regression model (Andersen et al., 1985) using continuous expression values and the tumor stage for all tumor samples, or just using continuous expression values for stage IV samples. All p-values were adjusted by Benjamini-Hochberg procedure. Kaplan-Meier curves were made with the help of the survfit function from the survival package using the median expression level to split samples into two groups: high or low expression of the gene of interest.
Immune-Related Genes Over-Expressed in PD-L1+ Melanomas are Uniformly Expressed in PD-L1+ RCCs Regardless of Clinical Outcome
In a prior study of archival melanoma specimens, we identified immune-related genes that were coordinately overexpressed in PD-L1+ compared to PD-L1(−) tumors (Taube et al., 2015). They included genes associated with CD8+ T cell activation (CD8A, IFNG, PRF1, CCL5), antigen presentation (CD163, TLR3, CXCL1, LYZ), and immunosuppression [PD1, CD274 (PD-L1), LAG3, IL10]. In the current study of PD-L1+ RCC, we first sought to examine whether these or other candidate immune-related genes were differentially expressed in tumors from patients responding or not to nivolumab therapy. The same 60-gene multiplex qRT-PCR array employed in our prior melanoma study was used to analyze PD-L1+ RCC specimens obtained from 12 patients before treatment with anti-PD-1, including 4 patients who responded to therapy (responders, R) and 8 who did not (non-responders, NR) (Table 1). Gene expression was normalized to the pan-immune cell marker PTPRC (CD45). We found that genes which were over-expressed in PD-L1+ vs. PD-L1(−) melanomas were also expressed in PD-L1+ RCC, and none of the screened molecules was significantly differentially expressed according to clinical outcomes after nivolumab therapy (data not shown). Additionally, we used immunohistochemistry (IHC) to examine protein expression of a more focused group of these immune-related molecules, including PD-1, PD-L2, LAG-3, TIM-3, and to identify infiltrating immune cell subsets (FoxP3, CD4:CD8 ratios) (
aAll samples were obtained by surgical resection.
bPrimary tumor refers to nephrectomy specimen.
cEvaluated according to Response Evaluation Criteria in Solid Tumors (RECIST) (Therasse et al., 2000).
dSample used only for immunohistochemical (IHC) analyses.
eSample used for qRT-PCR but not microarray analysis due to lack of sufficient RNA.
Increased Intratumoral Expression of Genes with Metabolic Functions is Associated with Resistance of PD-L1+ RCC to Anti-PD-1 Therapy
Because analysis of a selected panel of 60 immune-related genes did not reveal significant differences between PD-L1+ RCCs that were responsive or resistant to anti-PD-1 therapy, we next turned to unbiased analysis with whole genome expression profiling. For this analysis, we employed the DASL microarray platform (cDNA-mediated Annealing, Selection, extension, and Ligation; Illumina) designed for use with partially degraded mRNAs such as those isolated from formalin fixed paraffin-embedded (FFPE) tissues. Eleven available RCC specimens from among the original cohort were analyzed for expression of 29,670 gene targets (Table 1). By comparing tumors from 4 responding and 7 non-responding patients, we identified 234 probe sets corresponding to 226 genes that were differentially expressed between the two groups, based on a p-value of ≦0.01 and expression fold change ≧1.5. Among them, 116 probe sets corresponding to 113 genes were up-regulated in tumors from responding patients, and 118 probe sets corresponding to 113 genes were up-regulated in tumors from non-responding patients (
aPer DAVID web tool. See Huang et al., 2007, 2009a, and 2009b. In particular, Additional Data File 7 in Huang et al., 2007, contains a list of the 14 annotation categories used by the DAVID Functional Classification Tool, with associated web links.
bDAVID adjustment of the Fisher exact test (hypergeometric distribution) p-value.
aRefers to the official gene name from NCBI.
bObtained from HUGO Gene Nomenclature Committee website
cData were analyzed using the comparative Ct method (ΔΔCt), normalized to either GUSB (beta-glucuronidase) or PTPRC (CD45, pan immune cell marker). A 2-tailed, unpaired Student's t-test was used to determine the statistical significance of FC values.
aGene ID refers to official gene symbol
bParametric p-value derived from Student's t-test, analyzed with BRBArrayTools). Cutoff criteria were p-value ≦0.01 and expression fold change (FC) ≧1.5, comparing anti-PD-1 responders (R) versus non-responders (NR).
cFold change (FC), the ratio R/NR signal intensity detected by DASL. Genes are ordered based on ascending FC. Negative values indicate genes up-regulated in NR. NR, non-responder; R, responder.
dOfficial gene description
eTranscripts evaluated by different Illumina probe sets for the same gene.
Validation of Differentially Expressed Genes with Multiplex qRT-PCR
Following global gene expression profiling, a Custom Taqman Low-Density Array (TLDA; Applied BioSystems, Waltham Mass.) was designed to validate differential expression of 60 selected unique gene targets (Table 5). Criteria employed for gene selection included the following: expression fold-change ≧2, comparing tumors from NR vs. R; p-value ≦0.01; little or no overlap in the relative expression values of individual samples in the 2 groups; and biological associations. By considering results obtained with each of the four endogenous gene controls, 25 among the 60 queried genes were confirmed to be differentially expressed in the two groups of patients with divergent clinical outcomes (Table 3). Similar results were obtained when using 18S, ACTB, GUSB, or PTPRC to normalize gene expression. In particular, up-regulation of molecules associated with metabolic and solute transport functions was found in non-responders (
aAs provided at NCBI. Listed alphabetically.
bAs provided at NCBI..
AKR1C3
BACH2
BMP1
CACNB1
CCL3
CD24
COX5A
CYP4F11
E2F8
ENPP5
F2RL1
GALNT14
IL11RA
KCNJ16
LTBP1
MAL
MYLK2
NFATC1
PITX2
PLEC
PPP2R3B
SLC23A1
SLC37A4
SLCO3A1
TNFRSF19
UCP3
UGT1A1
UGT1A3
UGT1A6
WHSC1
aRefers to the official gene name (NCBI) Genes indicated in bold or italics are up-regulated or down-regulated, respectively, in RCCs from non-responding patients.
bRefers to the official gene description (NCBI)
cData were analyzed using the comparative Ct method (ΔΔCt). Results were normalized to 4 different internal control genes: GUSB, beta-glucuronidase; 18S, 18S ribosomal RNA; ACTB, beta-actin; and PTPRC, Protein Tyrosine Phosphatase, Receptor type, C (CD45, pan immune cell marker). Genes significantly and differentially expressed according to the cutoff criteria of fold change (FC) ≧2 and p-value ≦0.1 (2-tailed, unpaired Student's t-test) and are labelled as YES. Genes which do not meet these criteria are labelled as NO.
Genes Up-Regulated in PD-L1+ RCCs from Patients Resistant to Anti-PD-1 Therapy are Also Expressed by Kidney Cancer Cell Lines
The RCC TME is a complex milieu containing many different cell types. To understand whether metabolic genes that were over-expressed in tumor specimens from non-responding patients were specifically associated with renal carcinoma cells, we evaluated their expression in 12 established kidney cancer cell lines using qRT-PCR. Results confirmed that cultured renal carcinoma cells expressed the metabolic genes of interest (data not shown). RCC cell lines were also briefly exposed to IFN-g in vitro, to mimic an inflammatory in situ tumor milieu. Following exposure, expression of the metabolic factors UGT1A6 and KCNJ16 decreased by 2.4-fold and 2.5-fold, respectively (p=0.002 and 0.004 respectively, paired Student's t-test), suggesting the potential for cross-talk to occur between immunologic and metabolic factors found in the same TME.
UGT1A6 Protein is Over-Expressed in PD-L1+ RCCs Associated with Non-Response to Anti-PD-1 Therapy
UGT1A6 is involved in the chemical “defensome” and detoxifies exogenous and stress-related lipids. In whole genome expression profiling and qRT-PCR validation, it was the most highly over-expressed gene associated with non-response to anti-PD-1 (˜8-fold and ˜300-fold, p≦0.005 with multiple probes and p=0.007, respectively). Therefore, the expression of UGT1A6 in RCC was further explored at the protein level with IHC, in the same 12 specimens as those used for gene expression profiling. UGT1A6 protein expression was observed in renal epithelial cells and not stromal cells. Expression levels varied widely among the specimens (
Expression of UGT1A6 is not Generally Associated with Survival in Patients with RCC
To assess whether the over-expression of UGT1A6 is generally associated with poor prognosis in patients with kidney cancer, in silico analysis was performed on RNA expression data obtained from 444 clear cell RCC specimens in The Cancer Genome Atlas project (TCGA) (Cancer Genome Atlas Research, 2013). Kaplan-Meier curves were generated using the median expression value for UGT1A6 to segregate samples into high or low expressers. As shown in
The disclosure of each reference cited is expressly incorporated herein.
Clauses
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/US2015/050084 | 9/15/2015 | WO | 00 |
Number | Date | Country | |
---|---|---|---|
62050379 | Sep 2014 | US |