Patent Application
20150126621
The instant application contains a Sequence Listing which has been submitted in ASCII format via EFS-Web and is hereby incorporated by reference in its entirety. Said ASCII copy, created on May 8, 2013, is named 243735.000020_SL.txt and is 4,431 bytes in size.
The invention relates to miRNA-based methods for prognosis of recurrence of melanoma and related methods and kits.
Melanoma originates from uncontrolled proliferation of specialized melanocytes normally responsible for producing pigments in the epithelial layer. Though typically associated with the skin, these cells can also be present in the eye, bone and heart1, 2, 3, and cancer lesions can develop in any of these locations.
Staging of the melanoma at the time of diagnosis incorporates thickness, mitotic index, ulceration, and sentinel lymph node status, and is generally indicative of clinical outcome4-8. Melanoma is curable for most patients whose primary tumors are adequately removed; however, many patients recur and progress to advanced disease and death. The vast majority of recurrent patients present with metastatic disease, from which they eventually succumb. Accordingly, 7.1% and 32.8% of stage I and II patients suffer disease recurrence, respectively9.
Clinical staging incorporates information about the primary tumor, regional lymphatics, and distant metastatic sites into the AJCC (American Joint Committee on Cancer) TNM staging system and is the primary means of assessing prognosis10. It is however insufficient to account for within-stage heterogeneity of disease outcome. Thickness remains the most robust predictor of survival in localized melanoma, but the morphologically-based staging system only partly explains the variability in the natural history of melanoma. Although the AJCC has incorporated the mitotic index into the staging criteria, the biomarkers are limited by inter-observer variability and lack of standardization and hence, have not been integrated into clinical practice11-13.
Advanced malignant melanoma remains a disease with poor prognosis, with median survival of 8.5 months and a 5-year survival rate of less than 5%15. This is likely a reflection of the absence of effective treatment for late stage melanoma and hence, the early identification of patients at highest risk for the development of aggressive disease is critical. The identification of biomarkers to aid in the diagnosis and prognosis of the cancer would impact mortality from melanoma16.
The classic tumorigenesis model posits that stepwise accumulation of genetic changes gradually results in the acquisition of metastatic potential by tumor cells. Recurrence and metastasis occur from primary melanomas that frequently are histologically equivalent to non-recurrent lesions at the time of diagnosis and excision, suggesting that genomic or epigenetic alterations that predetermine a tumor's potential to spread may be acquired early in tumor progression. Evidence has been accumulating in support of such a deterministic model of tumor evolution: mRNA expression profiling in breast and prostate carcinoma have shown that expression profiles of paired primary and metastatic tumors were more similar to each other than to other patient tumors17. Further, gene expression signatures identified in primary prostate or breast tumors have been predictive of disease recurrence or progression to metastasis18-20.
A recent study identified pro-metastatic genes in melanoma, which are recurrently amplified in primary tumor and also act as classic oncogenes, suggesting that molecular events involved in tumor initiation can dictate clinical outcome21. Further, a sequencing study of a primary acral melanoma and its metastasis found that the majority of genetic alterations present in the metastasis were detectable in the primary tumor22.
There is an unmet need in the art for treatment of recurrent melanoma. Identifying molecular alterations that can be measured at the time of melanoma diagnosis that are predictive of disease recurrence would be clinically useful for developing individualized treatment plans and/or to uncover novel therapeutic targets.
The alteration of miRNA expression correlates with cancer progression, and the perturbation of individual miRNAs can functionally impact cancer cell metastasis24-30. A study by inventors and co-workers identified several miRNAs that were prognostic in metastatic melanoma and were also found to be altered in primary tumors23 suggesting that melanoma metastasis may not strictly be a consequence of stepwise accumulation of molecular alterations resulting in rare cells that gain metastatic capacity. Rather, a larger population(s) of cells with metastasis-initiating events may be present at early stages of melanomagenesis.
As follows from the Background section, above, there is an unmet need in the art for compositions and methods for prognosis and treatment of recurrent melanoma.
The present invention addresses these and other needs by providing a method for predicting the likelihood of recurrence of melanoma (including distal metastasis and locoregional recurrence) in a subject diagnosed with melanoma, said method comprising:
In one specific embodiment, the above method comprises measuring the level of miR-10a, miR-1285, miR-374b*, miR-377*, miR-513b, miR-342-3p, miR-625*, SNORD3A, miR-1204, miR-574-3p, let-7a-2*, miR-615-3p, miR-564, miR-154*, miR-7, miR-215, miR-382, miR-663, miR-516b, miR-99b, and miR-1276. In another specific embodiment, the above method comprises measuring the level of miR-374b*, miR-377*, miR-1285, and miR-1276. In yet another specific embodiment, the above method comprises measuring the level of miR-374b*, miR-377*, miR-1285, and miR-1204. In a further specific embodiment, the above method comprises measuring the level of miR-382, miR-1276, and miR-615-3p. In another specific embodiment, the above method comprises measuring the level of miR-215, miR-374b*, miR-382, miR-516b, and miR-7. In yet another specific embodiment, the above method comprises measuring the level of miR-382, miR-516b, and miR-7.
The combined control levels used in the above method can be any suitable control (e.g., a predetermined standard or the combined levels of the same miRNAs in a non-recurrent melanoma sample [e.g., determined by the statistical measure, Youden's Index of the Receiving Operative Characteristic (ROC) curve, see, e.g., Zhou et al. (2011) Statistical Methods in Diagnostic Medicine, 2nd Edition, Wiley, N.J.]).
In one specific embodiment, the subject is human. In another specific embodiment, the subject is an experimental animal.
In one embodiment, the above method comprises a step of collecting the melanoma sample from the subject.
In the above method, the levels of the miRNAs can be determined using any method known in the art (e.g., hybridization [e.g., to miRNA arrays], RT-PCR, sequencing, etc.). In one embodiment, prior to measuring miRNA level, the miRNA is purified from the melanoma sample. In another embodiment, the method of the invention further comprises the step of reducing or eliminating degradation of the miRNA.
In one embodiment, the above method is followed by administering to the subject determined as being at high risk of melanoma recurrence a melanoma treatment. Any melanoma treatment can be used. Non-limiting examples of treatments include, e.g., Interleukin 2 (IL2), Aldesleukin (Proleukin), Dacarbazine (DTIC-Dome), Ipilimumab (Yervoy), temozolomide, Vemurafenib (Zelboraf), and any combinations thereof.
In another embodiment, the above method is followed by recruiting the subject in a clinical trial.
In conjunction with the above prognostic method, the invention provides a kit comprising primers or probes specific for four or more miRNAs selected from the group consisting of miR-10a, miR-1285, miR-374b*, miR-377*, miR-513b, miR-342-3p, miR-625*, SNORD3A, miR-1204, miR-574-3p, let-7a-2*, miR-615-3p, miR-564, miR-154*, miR-7, miR-215, miR-382, miR-663, miR-516b, miR-99b, and miR-1276.
In one specific embodiment, such kit comprises primers or probes specific for miR-10a, miR-1285, miR-374b*, miR-377*, miR-513b, miR-342-3p, miR-625*, SNORD3A, miR-1204, miR-574-3p, let-7a-2*, miR-615-3p, miR-564, miR-154*, miR-7, miR-215, miR-382, miR-663, miR-516b, miR-99b, and miR-1276. In another specific embodiment, such kit comprises primers or probes specific for miR-374b*, miR-377*, miR-1285, and miR-1276. In yet another specific embodiment, such kit comprises primers or probes specific for miR-374b*, miR-377*, miR-1285, and miR-1204. In a further specific embodiment, such kit comprises primers or probes specific for miR-382, miR-1276, and miR-615-3p. In another specific embodiment, such kit comprises primers or probes specific for miR-215, miR-374b*, miR-382, miR-516b, and miR-7. In yet another specific embodiment, such kit comprises primers or probes specific for miR-382, miR-516b, and miR-7. Any of the above kits can optionally further comprise miRNA isolation or purification means. Any of the above kits can optionally further comprise instructions for use.
In a separate aspect, the invention provides a method for treatment of a melanoma recurrence (including distal metastasis and locoregional recurrence) in a subject in need thereof (e.g., human of experimental animal) comprising increasing the level and/or activity of at least one miRNA selected from the group consisting of miR-215, miR-374b*, miR-382, miR-516b, and miR-7 in the melanoma cells of the subject.
The current invention is based on the observation that there are different levels of certain miRNAs in different patients at the time of diagnosis of melanoma. As discussed in more detail in the Examples, below, 204 primary melanoma tumors were analyzed, and top differentially-expressed miRNAs in recurrent versus non-recurrent patients were identified, as well as miRNAs whose expression correlated with tumor thickness (Tables 4a-d). Selected miRNAs were further tested in an in vitro screen in two melanoma cell lines to determine which miRNAs functionally impact melanoma metastasis. A subsequent screen of 13 of the miRNAs in both in vitro invasion and cell proliferation assays revealed miR-215, miR-374b*, miR-382, miR-516b, and miR-7 as being less expressed in recurrent as opposed to non-recurrent and/or thicker versus thinner tumors and were deemed potent suppressors of in vitro invasion. It has been shown that alteration of miRNA expression correlates with cancer progression, and the perturbation of individual miRNAs can functionally impact cancer cell metastasis.
Based on these observations, the present invention provides a novel highly sensitive method for predicting the likelihood of recurrence of melanoma (including distal metastasis and locoregional recurrence) at the time of diagnosis in a subject, said method comprising:
a. measuring the levels of four or more miRNAs selected from the group consisting of miR-10a, miR-1285, miR-377*, miR-513b, miR-342-3p, miR-625*, SNORD3A, miR-1204, miR-574-3p, let-7a-2*, miR-615-3p, miR-564, miR-154*, miR-663, miR-99b, miR-1276, miR-215, miR-374b*, miR-382, miR-516b, and miR-7, in a melanoma sample collected from the subject;
b. calculating combined levels of the miRNAs measured in step (a);
c. comparing the combined levels of the miRNAs measured in step (a) with the corresponding combined control levels of said miRNAs, and
d. (i) identifying the subject as being at high risk of melanoma recurrence if the combined levels of the miRNAs measured in step (a) are higher than the corresponding combined control levels or (ii) identifying the subject as being at low risk of melanoma recurrence if the combined levels of the miRNAs measured in step (a) are same or lower than the corresponding combined control levels.
The method of the invention makes possible early prognosis of recurrence of melanoma, e.g., at the time of the diagnosis prior to occurrence of major morphological changes and/or metastasis associated with the disease, allowing for early application of treatments to prevent such morphological changes and/or metastasis. Patients determined to be at low risk of melanoma recurrence will likely receive no additional treatment after primary tumor excision, and will simply be subject to annual or semi-annual check-ups (e.g., physical and skin screening). On the other hand, patients determined to be at high risk of melanoma recurrence, after primary tumor excision, will be likely subject to a more frequent and detailed surveillance (e.g., involving MRI or other imaging modality to identify local and distal recurrences), more extensive primary tumor staging (e.g., involving sentinel and/or regional lymph node mapping) and may be subject to a melanoma treatment (e.g., Interleukin 2 (IL2), Aldesleukin (Proleukin), Dacarbazine (DTIC-Dome), Ipilimumab (Yervoy), temozolomide, Vemurafenib (Zelboraf), and any combinations thereof).
The method of the invention also allows for more precise identification of various groups of patients who can be then recruited in clinical trials to develop and/or test new treatments to prevent melanoma recurrence.
In addition, cellular pathways regulated by the prognostic miRNAs identified herein are potential molecular therapeutic targets for control of melanoma recurrence. Thus, in conjunction with the prognostic method, the present invention also provides a method for treatment of a melanoma recurrence in a subject in need thereof comprising increasing the level and/or activity of at least one miRNA selected from the group consisting of miR-215, miR-374b*, miR-382, miR-516b, and miR-7 in the melanoma cells of the subject. Such increase in the level and/or activity of said miRNAs can be achieved using any method known in the art (e.g., over-expressing miRNA or mature miRNA mimic [an oligonucleotide, usually with some structural change(s), of the same sequence as the mature miRNA], e.g., using viral constructs; inhibiting negative or activating positive miRNA regulators [transcriptional or epigenetic], etc.).
The methods of the invention involve measuring miRNA levels. Examples of useful methods for measuring miRNA level in solid tumors include hybridization with selective probes (e.g., using Northern blotting, bead-based flow-cytometry, oligonucleotide microchip [microarray] (e.g., from Agilent, Exiqon, Affymetrix), or solution hybridization assays such as Ambion mirVana miRNA Detection Kit), polymerase chain reaction (PCR)-based detection (e.g., stem-loop reverse transcription-polymerase chain reaction [RT-PCR], quantitative RT-PCR based array method [qPCR-array]), or direct sequencing by one of the next generation sequencing technologies (e.g., Helicos small RNA sequencing, miRNA BeadArray (Illumina), Roche 454 (FLX-Titanium), and ABI SOLiD). For review of additional applicable techniques see, e.g., Chen et al., BMC Genomics, 2009, 10:407; Kong et al., J Cell Physiol. 2009; 218:22-25.
In some embodiments, miRNAs are purified prior to quantification. miRNAs can be isolated and purified from solid tumors by various methods, including, e.g., Qiazol or Trizol extraction or the use of commercial kits (e.g., miRNeasy kit [Qiagen], MirVana RNA isolation kit [Ambion/ABI], miRACLE [Agilent], High Pure miRNA isolation kit [Roche], and miRNA Purification kit [Norgen Biotek Corp.]), concentration and purification on anion-exchangers, magnetic beads covered by RNA-binding substances, or adsorption of certain miRNA on complementary oligonucleotides.
In some embodiments, miRNA degradation in solid tumor samples and/or during miRNA purification is reduced or eliminated. Useful methods for reducing or eliminating miRNA degradation include, without limitation, adding RNase inhibitors (e.g., RNasin Plus [Promega], SUPERase-In [ABI], etc.), use of guanidine chloride, guanidine isothiocyanate, N-lauroylsarcosine, sodium dodecylsulphate (SDS), or a combination thereof. Reducing miRNA degradation in samples is particularly important when sample storage and transportation is required prior to miRNA quantification.
In conjunction with the prognostic method, the present invention also provides various kits comprising primer and/or probe sets specific for the detection of biomarker miRNAs. Non-limiting examples of primer or probe combinations in kits are as follows:
1. Primers or probes specific for four or more miRNAs selected from the group consisting of miR-10a, miR-1285, miR-374b*, miR-377*, miR-513b, miR-342-3p, miR-625*, SNORD3A, miR-1204, miR-574-3p, let-7a-2*, miR-615-3p, miR-564, miR-154*, miR-7, miR-215, miR-382, miR-663, miR-516b, miR-99b, and miR-1276.
2. Primers or probes specific for miR-10a, miR-1285, miR-374b*, miR-377*, miR-513b, miR-342-3p, miR-625*, miR-1204, miR-574-3p, let-7a-2*, miR-615-3p, miR-564, miR-154*, miR-7, miR-215, miR-382, miR-663, miR-516b, miR-99b, and miR-1276.
3. Primers or probes specific for miR-374b*, miR-377*, miR-1285, and miR-1276.
4. Primers or probes specific for miR-374b*, miR-377*, miR-1285, and miR-1204.
5. Primers or probes specific for miR-382, miR-1276, and miR-615-3p.
6. Primers or probes specific for miR-215, miR-374b*, miR-382, miR-516b, and miR-7.
7. Primers or probes specific for miR-382, miR-516b, and miR-7.
Such kits can be useful for direct miRNA detection in primary melanoma tumor samples isolated from patients or can be used on purified miRNA samples.
A kit of the invention can also provide reagents for primer extension and amplification reactions. For example, in some embodiments, the kit may further include one or more of the following components: a reverse transcriptase enzyme, a DNA polymerase enzyme (such as, e.g., a thermostable DNA polymerase), a polymerase chain reaction buffer, a reverse transcription buffer, and deoxynucleoside triphosphates (dNTPs). Alternatively (or in addition), a kit can include reagents for performing a hybridization assay. The detecting agents can include nucleotide analogs and/or a labeling moiety, e.g., directly detectable moiety such as a fluorophore (fluorochrome) or a radioactive isotope, or indirectly detectable moiety, such as a member of a binding pair, such as biotin, or an enzyme capable of catalyzing a non-soluble colorimetric or luminometric reaction. In addition, the kit may further include at least one container containing reagents for detection of electrophoresed nucleic acids. Such reagents include those which directly detect nucleic acids, such as fluorescent intercalating agent or silver staining reagents, or those reagents directed at detecting labeled nucleic acids, such as, but not limited to, ECL reagents. A kit can further include miRNA isolation or purification means as well as positive and negative controls. A kit can also include a notice associated therewith in a form prescribed by a governmental agency regulating the manufacture, use or sale of diagnostic kits. Detailed instructions for use, storage and troubleshooting may also be provided with the kit. A kit can also be optionally provided in a suitable housing that is preferably useful for robotic handling in a high throughput setting.
The components of the kit may be provided as dried powder(s). When reagents and/or components are provided as a dry powder, the powder can be reconstituted by the addition of a suitable solvent. It is envisioned that the solvent may also be provided in another container. The container will generally include at least one vial, test tube, flask, bottle, syringe, and/or other container means, into which the solvent is placed, optionally aliquoted. The kits may also comprise a second container means for containing a sterile, pharmaceutically acceptable buffer and/or other solvent.
Where there is more than one component in the kit, the kit also will generally contain a second, third, or other additional container into which the additional components may be separately placed. However, various combinations of components may be comprised in a container.
Such kits may also include components that preserve or maintain DNA or RNA, such as reagents that protect against nucleic acid degradation. Such components may be nuclease or RNase-free or protect against RNases, for example. Any of the compositions or reagents described herein may be components in a kit.
As used herein in connection with melanoma, the term “recurrence” refers to a return of the disease, either locally (e.g., where it used to be before resection) or distally (e.g., metastasis).
The terms “microRNA” or “miRNA” as used herein refer to a class of small approximately 22 nt long non-coding RNA molecules. They play important roles in the regulation of target genes by binding to complementary regions of messenger transcripts (mRNA) to repress their translation or regulate degradation (Griffiths-Jones Nucleic Acids Research, 2006, 34, Database issue: D140-D144). Frequently, one miRNA can target multiple mRNAs and one mRNA can be regulated by multiple miRNAs targeting different regions of the 3′ UTR. Once bound to an mRNA, miRNA can modulate gene expression and protein production by affecting, e.g., mRNA translation and stability (Baek et al., Nature 455(7209):64 (2008); Selbach et al., Nature 455(7209):58 (2008); Ambros, 2004, Nature, 431, 350-355; Bartel, 2004, Cell, 116, 281-297; Cullen, 2004, Virus Research., 102, 3-9; He et al., 2004, Nat. Rev. Genet., 5, 522-531; and Ying et al., 2004, Gene, 342, 25-28). Information on most currently known miRNAs can be found in the miRNA database miRBase (available at the world wide web at mirbase.org). For the purposes of the present invention, the terms “microRNA” or “miRNA” include, in addition to the miRNAs described above, SNORD3A small RNA.
The term “miRNA array” refers to a multiplex technology used in molecular biology and in medicine. It consists of an arrayed series of multiple (e.g., up to 2000) microscopic spots of oligonucleotides, each containing a specific sequence (probe) complementary to a particular target miRNA. After probe-target hybridization under high-stringency conditions the resulting hybrids are usually detected and quantified by quantifying fluorophore-, silver-, or chemiluminescence-labeled targets to determine relative abundance of miRNA. In the methods of the present invention, both custom-made and commercially available miRNA arrays can be used. Non-limiting examples of useful commercially available miRNA arrays (based on various methods of target labeling, hybrid detection and analysis) include arrays produced by Exiqon, Affymetrix, Agilent, Illumina, Invitrogen, Febit, and LC Sciences.
The term “next generation sequencing technologies” broadly refers to sequencing methods which generate multiple sequencing reactions in parallel. This allows vastly increased throughput and yield of data. Non-limiting examples of commonly used next generation sequencing platforms include Helicos small RNA sequencing, miRNA BeadArray (Illumina), Roche 454 (FLX-Titanium), and ABI SOLiD.
The term “combined levels of the miRNAs” as used herein refers to the linear combinations of miRNAs levels. The linear combinations of miRNAs levels can be calculated using any method known in the art. For example, as specified in the Examples section, below, the coefficients required for such linear combinations can be calculated using the logistic regression method, i.e., β1*x1+β2*x2+ . . . +βk*xk, where β's are the coefficients from logistic regression model, and x's are the levels of miRNAs (see, e.g., Steyerberg (2009) Clinical Prediction Models, Springer, N.Y.). The coefficients can be positive or negative. The “combined levels”, or the score, is always positively associated with the risk of recurrence. Thus, patients with higher score have a higher probability of recurrence. The logistic regression method can also include more clinical predictors if necessary. The performance of logistic regression models is measured using the area under the Receiving Operative Characteristic (ROC) curves: the larger the area under the ROC curve, the better performance of the model. The best performed model would yield the coefficients required to calculate the linear combination of levels of candidate miRNAs.
In the context of the present invention insofar as it relates to melanoma and melanoma recurrence, the terms “treat”, “treatment”, and the like mean to relieve or alleviate at least one symptom associated with such condition, or to slow or reverse the progression of melanoma or melanoma recurrence, or to arrest, prevent or delay the onset (i.e., the period prior to clinical manifestation) and/or reduce the risk of developing or worsening of melanoma or melanoma recurrence.
An “individual” or “subject” or “animal”, as used herein, refers to humans, veterinary animals (e.g., cats, dogs, cows, horses, sheep, pigs, etc.) and experimental animal models of melanoma. In a preferred embodiment, the subject is a human.
The term “purified” as used herein refers to material that has been isolated under conditions that reduce or eliminate the presence of unrelated materials, i.e., contaminants, including native materials from which the material is obtained. For example, RNA purification includes elimination of proteins, lipids, salts and other unrelated compounds. Besides, for some methods of analysis a purified miRNA is preferably substantially free of other RNA oligonucleotides contained in tumor samples (e.g., rRNA and mRNA fragments, etc.). As used herein, the term “substantially free” is used operationally, in the context of analytical testing of the material. Preferably, purified material substantially free of contaminants is at least 50% pure; more preferably, at least 90% pure, and still more preferably at least 99% pure. Purity can be evaluated by chromatography, gel electrophoresis, composition analysis, biological assay, and other methods known in the art.
As used herein, the term “similarly processed” refers to samples (e.g., tumor samples or purified miRNAs) which have been obtained using the same protocol.
The term “associated with” is used to encompass any correlation, co-occurrence and any cause-and-effect relationship.
The term “about” or “approximately” means within a statistically meaningful range of a value. Such a range can be within an order of magnitude, preferably within 50%, more preferably within 20%, still more preferably within 10%, and even more preferably within 5% of a given value or range. The allowable variation encompassed by the term “about” or “approximately” depends on the particular system under study, and can be readily appreciated by one of ordinary skill in the art.
In accordance with the present invention there may be employed conventional molecular biology, microbiology, and recombinant DNA techniques within the skill of the art. Such techniques are explained fully in the literature (e.g., ref 32-59).
The present invention is also described and demonstrated by way of the following examples. However, the use of these and other examples anywhere in the specification is illustrative only and in no way limits the scope and meaning of the invention or of any exemplified term. Likewise, the invention is not limited to any particular preferred embodiments described here. Indeed, many modifications and variations of the invention may be apparent to those skilled in the art upon reading this specification, and such variations can be made without departing from the invention in spirit or in scope. The invention is therefore to be limited only by the terms of the appended claims along with the full scope of equivalents to which those claims are entitled.
501MEL cells (Halaban et al., PLoS One. 2009, 4(2):e4563) were obtained from Yale University and were cultured in Optimem+5% fetal bovine serum (FBS). 451Lu cells derived from metastatic melanoma were obtained from Dr. Meenhard Herlyn at Wistar Institute (Smalley et al., Mol. Cancer Ther., 2006, 5(5):1136-1144). The cells were cultured in Tu2%, which contains 80% MCDB153 (Sigma Aldrich) and 20% L15 (Cellgro), and were supplemented with ˜2% FBS, 1.68 mM CaCl2, and 5 μg/mL bovine insulin. SK-MEL-147, SK-MEL-173, and SK-MEL-28 cells were obtained from Dr. Alan Houghton at Memorial Sloan-Kettering Cancer Center (see Houghton et al., J Exp Med., 1982, 156(6):1755-1766 and Segura et al., Proc. Natl. Acad. Sci. USA, 2009, 106(6):1814-1819) and were cultured in DMEM+10% FBS. All cells were grown in a humidified incubator at 37° C. and 5% CO2.
Pelleted cells were stored at −20° C. until RNA extraction. RNA was extracted using miRNeasy mini kits (Qiagen) following manufacturer's recommendations. Briefly, pelleted cells were thawed on ice. 700 μL per tube of Qiazol (Qiagen) was added, tubes were vortexed for ˜60 sec, and incubated at room temperature for 5 minutes. 140 μL chloroform was added and tubes were shaken for 15 sec, followed by 2 minutes incubation at room temperature. Tubes were centrifuged at 12,000×g at 4° C. for 15 minutes. Aqueous phase (˜350 uL) was transferred to a fresh microcentrifuge tube. 1.5× volumes of 100% EtOH were added and mixed by vortexing briefly. 700 μL at a time were transferred to RNeasy mini spin columns and centrifuged at 13,000 rpm for 30 sec. Repeat with remainder of sample, discarding flow-through. 350 μL of buffer RWT were added per column and centrifuged at 13,000 rpm for 30 seconds. Flow-through was discarded. 80 μL of DNAse I (Qiagen) was added to each column and incubated at room temperature for 15 minutes. 350 μL of buffer RWT were added to column, followed by centrifugation at 13,000 rpm for 30 sec. Two washes with 500 μL buffer RPE were performed discarding flow-through each time, followed by centrifugation at 13,000 rpm for 2 minutes to remove all traces of ethanol from RPE buffer. Columns were transferred to 1.5 mL RNA collection tubes and 30 to 50 μL RNase-free H2O was added per column for RNA elution. After 1 minute incubation at room temperature, columns were centrifuged at 13,000 rpm for 1 minute. Eluted RNA was quantified by Nanodrop 2000 (Thermo Scientific) and stored at −80° C.
FFPE Melanomas or Nevi.
5 μm sections (4-12) were attached to PEN-Membrane 2.0 μm slides (Leica) designed for laser capture microdissection. Primary melanoma tissues were macroscopically dissected using disposable scalpels (Feather No. 11) under a dissecting microscope and guided by hematoxylin and eosin (H&E) staining of consecutive sections. Cut sections were stored in microcentrifuge tubes until RNA extraction. RNA extraction was performed with miRNeasy FFPE kit (Qiagen) following manufacturer's recommendations. RNA was quantified by Nanodrop 2000 (Thermo Scientific) and stored at −80° C.
miRNA Array Profiling
miRNA expression profiling of FFPE-extracted RNA from primary melanomas was performed by Exiqon. Briefly, the quality of the total RNA was verified by an Agilent 2100 Bioanalyzer profile (Agilent). For each cohort, a reference sample was generated by mixing an equal amount of all samples analyzed. 300 ng total RNA from sample and reference was labeled with Hy3™ and Hy5™ fluorescent label, respectively, using the miRCURY™ LNA Array power labelling kit (discovery cohort) or miRCURY LNA™ microRNA Hi-Power Labeling Kit (validation cohort) (Exiqon, Denmark) by following the procedure described by the manufacturer. The Hy3™-labeled samples and a Hy5™-labeled reference RNA sample were mixed pair-wise and hybridized to the miRCURY™ LNA array version 11.0 (discovery cohort) or miRCURY LNA™ microRNA Array 6th generation (validation cohort) (Exiqon, Denmark), which contain capture probes targeting all miRNAs for human, mouse or rat registered in the miRBASE version 14.0 or 16.0, respectively, at the Sanger Institute (http://www.mirbase.org/). The hybridization was performed according to the miRCURY™ LNA array manual using a Tecan HS4800 hybridization station (Tecan, Austria). After hybridization, the microarray slides were scanned and stored in an ozone-free environment (ozone level below 2.0 ppb) in order to prevent potential bleaching of the fluorescent dyes. The miRCURY™ LNA array microarray slides were scanned using the Agilent G2565BA Microarray Scanner System (Agilent Technologies, Inc., USA) and the image analysis was carried out using the ImaGene 8.0 or 9.0 software (BioDiscovery, Inc., USA). The quantified signals were background corrected (Normexp with offset value 10—Ritchie et al., Bioinformatics, 2007, 23(20):2700-2707) and normalized using the global Lowess (LOcally WEighted Scatterplot Smoothing) regression algorithm (Cleveland, W. S., 1979, J. Amer. Statist. Assoc., 74:829-836).
After Lowess normalization, scale normalization was performed such that each array has the same median expression level and inter-quartile range 1 (the third quartile minus the first quartile). The 339 miRNAs with highest expression levels (based on both Hy3 and Hy5 signals) among 867 miRNA probes in the discovery cohort were used for variable selection.
Predictive Signature for Recurrence from Discovery Cohort.
Seventy stage I/II patients with at least 3 years of follow-up in the discovery cohort were used to identify a predictive signature for 3-year recurrence. The miRNAs were ranked (adjusted for stage) according to three related endpoints: 3-year recurrence (logistic regression, Table 4a), tumor thickness (linear regression, Table 4b), and recurrence-free survival (RFS) (Cox PH regression, Table 4c). Starting from top ranked miRNAs in each of the rank list, multiple logistic regression models were developed using miRNAs as predictors with adjustment for stage. Since the cohort size is relatively small, each significant model has 7 to 11 miRNAs constructed by maximizing the area under the receiver operating characteristic (ROC) curve for 3-year recurrence with 4-fold cross validation. To minimize premature exclusion of promising miRNAs, 4 logistic regression models (
Evaluation of the Predictive Signature in the Validation Cohort.
For independent validation, the recurrence potential score formula obtained from the discovery cohort were directly applied to the validation cohort and used to obtain an ROC, the same cut-off point was used to classify the validation cohort patients into high and low risk groups. Kaplan-Meier curves were plotted for the RFS of the two groups and log-rank test p-value was obtained.
From the discovery cohort, differential expression of miRNAs between thick and thin or recurrent and non-recurrent tumors was determined using two-tailed t-testing using the Benjamini Hochberg method (Benjamini, Yoav and Hochberg, Yosef, Journal of the Royal Statistical Society, Series B (Methodological), 1995, 57(1): 289-300).
Validation of miRNA expression in tissues.
Mature miRNA expression validation in RNA extracted from FFPE primary melanomas was performed using Exiqon reagents following manufacturer's recommendations. 20 samples from discovery cohort were used for analysis. Reverse transcription was performed using miRCURY LNA Universal cDNA synthesis kit (Exiqon). Briefly, 25 ng RNA was diluted to 10 μL in nuclease-free H2O in 96-well PCR plates (Biorad). Reverse transcription master mixes were made with 5× reaction buffer, H2O, Spike RNA control, and +/− reverse transcriptase. 10 μL master mix was added to each 10 μL aliquot of diluted RNA and mixed gently. Tubes were incubated for 60 minutes at 42° C. followed by 5 minutes at 95° C. Duplicate room temperature reactions were performed for each sample tested. Reverse transcription products were used immediately or briefly stored at −20° C. until use. PCR was carried out using miRCURY LNA SYBR green mastermix and miRNA-specific LNA primers. Briefly, cDNA was diluted 80× in nuclease-free H2O and ROX passive reference dye (Invitrogen). PCR master mixes were made with 2× miRCURY LNA SYBR green master mix (5 μL) and LNA primers (1 μL). 4 μL diluted cDNA or H2O was added to wells of 384-well plates containing 6 μL of PCR master mix. Duplicate wells of each cDNA were run. PCR reactions were performed using a 7900 HT (Applied Biosystems) as follows: 10 min at 95° C., and 40 cycles of 10 sec at 95° C. followed by 60 sec at 60° C. with ramp rate of 1.6° C./sec. Data was analyzed in Excel (Microsoft) and relative expression determined by the method of Livak (Livak K J, Schmittgen T D, Methods. 2001, 25(4):402-408). Ct values from duplicate room temperature reactions were averaged. The data were normalized to the geometric mean of 3 internal reference miRNAs (miR-146b-3p, miR-let-7e, and miR-485-3p) selected due to their low deviation across samples in the original arrays and to the RNA spike-in control (as a measure of room temperature efficiency). Array log 2 expression ratios and qPCR expression was expressed as relative to an arbitrary sample (05-061) and plotted using Graphpad PRISM (Graphpad; www.graphpad.com). Correlation (R) values were calculated by Pearson correlation.
miRNA Overexpression in Cultured Cells
Mature miRNA expression was quantified using Taqman miRNA assays (Applied Biosystems) following manufacturer's recommendations. Briefly, RNA was diluted to 12.5 ng/μL. miRNA-specific reverse transcription (room temperature) master mixes were made with H2O, 10× RT buffer, miRNA RT primers, RNase inhibitor, dNTPs, and with or without reverse transcriptase. 14 μL RT master mix was added per well and 1 μL appropriate 12.5 ng/μL RNA stock was added to master mix, followed by incubation on ice for 5 minutes. Reverse transcription (RT) was carried out in a thermal cycler with 30 min at 16° C., 30 minutes at 42° C., and 15 minutes at 85° C. RT products were stored at −20° C. if not used immediately. Polymerase chain reaction (PCR) master mixes were made with miRNA-specific 20× Taqman primers, (2×) Taqman universal PCR master mix, fluorescein (Molecular Probes), and H2O. 18.66 μL PCR master mix was added per well in 96-well PCR plates (Biorad) followed by addition of 1.33 μL per well of appropriate RT or −RT reaction product or H2O. PCR reaction was performed on an iCycler equipped with a MyIQ Real-time PCR Imaging system (Biorad). Cycling was performed as follows: 10 sec and 95° C., and 40 cycles of 15 sec at 95° C. and 60 sec at 60° C. Ct threshold was selected with amplification curves in log scale. Relative expression were analyzed by Livak method using U6 snRNA or RNU44 as internal controls and plotted with Graphpad PRISM (Graphpad; www.graphpad.com).
4×106 293 T cells were seeded per 10 cM tissue culture dish and incubated overnight at 37° C. and 5% CO2. 16-20 hrs after seeding, 293T were co-transfected with lentiviral expression constructs (15 μg), viral packaging plasmid (psPAX2, 10 μg), and viral envelope plasmid (pMD2.G, 5 μg) using Lipofectamine2000 (Invitrogen) following manufacturer's recommendations. Viral supernatant was collected and 0.45 μm filtered at 36 hrs post-transfection and stored at 4° C. for short-term use (1-5 days) or −20° C. for long-term storage (5-30 days).
Target cells were seeded and incubated overnight prior to infection. Medium was replaced with 1:2 diluted viral supernatant with 4 μg/mL polybrene and incubated for 6-8 hrs, followed by replacement with growth medium. Cells were checked for GFP expression on subsequent days to ensure pure populations of GFP-bright transduced cells.
Fluorescent Cell Generation.
Lentiviral supernatant was generated as previously described (Segura et al., Proc. Natl. Acad. Sci. USA, 2009, 106(6):1814-1819) of green fluorescent protein (GFP) expression constructs (pGIPZ, Openbiosystems or pMIRH, Systems Biosciences). All cell lines were transduced at high efficiency to generate pure, GFP bright cell populations for use in invasion assays.
Reverse Transfection.
Transfection conditions were optimized for each cell line using dy547 or fluorescein labeled oligos (Dharmacon, dy547). Liposomal transfection complexes with miRNA mimics (Dharmacon, 50 nM final) or siRNA pools (Dharmacon, Smart Pools, 50 nM) were generated with Lipofectamine 2000 (Invitrogen, 0.2 μL per well) in at least triplicate in 96-well plates following manufacturer's recommendations. Replicate wells were scattered on the plate to limit technical bias. GFP expressing cells were seeded at specific densities (501MEL-25,000 cells/well, SK-MEL-147, SK-MEL-28, and 451Lu—30,000 cells/well, SK-MEL-173-40,000 cells/well) into wells containing liposomal complexes followed by overnight incubation in a humidified incubator at 37° C. and 5% CO2. Media was changed after incubation with liposomal complexes. 48-hours after initiation of transfection, cells were used for invasion assay seeding.
Invasion Assay Seeding.
Optimization was performed for each cell line to identify assay time length. Additionally, to identify the optimal seeding density, a 2-fold dilution series of each cell line was performed to test the linear range of the assay. 48 hour post-transfection cell counts were performed in initial experiments to ensure optimal cell quantities were transferred from transfection plate to invasion assay plate. Prior to invasion assay seeding, 96-well Fluoroblok inserts (Becton Dickinson) were coated with 10 μg/mL fibronectin in PBS for 60 min at room temperature, followed by PBS supplemented with 2.5% bovine serum albumin at RT until cell seeding (10-30 minutes). Cells were washed 1× with PBS, dissociated from 96-well plates using small volumes of 0.05% Trypsin-EDTA (Invitrogen) or Cell Dissociation Buffer (PBS-based, Invitrogen), and quenched with the described basal growth media, but supplemented with only 1/10 the volume of FBS and bovine insulin (451Lu) (top chamber media). Single cell suspensions were generated by gentle, repetitive (40×) pipetting using an 8-channel multipipette. 12.5 μL, 25 μL, or 50 μL of cell suspension (cell line dependent) were transferred using an 8-channel multipipette to the upper chamber of the 96-well Fluorblok inserts to yield resulting cell inputs in the previously defined optimal range. Top chamber media was supplemented to 50 μL for each insert well. Cells were allowed to settle then 200 μL growth media per well was added to the lower chamber of 96-well Fluoroblok inserts. An equivalent volume of cell suspension as used in the invasion assay was transferred to a standard 96-well tissue culture microplate as a cell input control. Invasion assay and cell input control plates were maintained at 37° C. and 5% CO2 until automated assay quantification. Cell input control plates were imaged and counted ˜30 min after seeding, except SK-MEL-173 which was imaged and counted at 40 hrs post-seeding. Invasion assay plates were imaged and counted 8-20 hrs post-seeding, except SK-MEL-173, which was imaged and counted at 40 hrs post-seeding.
Invasion Quantification.
Invasion assay and cell input control plates were scanned using a Cellomics ArrayScan VTI HCS Reader (Cellomics), a high-content inverted fluorescent microscope system with companion software. A 5× objective was used for imaging. Four fields per insert, which covered >95% of the insert membrane bottom were imaged for invasion assay plates, while seven fields per well were imaged for cell input control plates. GFP-labeled cells were counted by GFP fluorescence using a version of the TargetActivation_v3 protocol (Cellomics) modified to optimally capture individual cells. Modified parameters included: fixed threshold of 25-50, exposure length, size exclusion criteria, smoothing factor, and segmentation. Cell counts for each well were normalized to the average counts (of replicate wells) for the corresponding condition in the cell input plate to control cell proliferation effects that may have occurred between initiation of transfection and assay seeding.
Cell Proliferation.
Indicated cells were reverse transfected (n=6) following previously established conditions. 48-hours after transfection, cells were washed 1× with PBS, dissociated from well with 30 μL per well of 0.05% Trypsin-EDTA (Invitrogen). Trypsin was quenched with 270 μL growth media. 30-50 μL per well were transferred to replicate plates (n=6) to initiate growth curve. After cell attachment (4-6 hrs), 1 plate was fixed as a zero time point, and subsequent plates were fixed every 24 hours thereafter. Plates were fixed with 1% glutaradlehyde (Sigma) in PBS for 15 minutes at room temperature, followed by storage in PBS at 4° C. All plates were stained with 0.5% crystal violet in PBS for 2 hrs followed by extensive wash with diH2O. Crystal violet staining was dissolved in 15% acetic acid and measured at absorbance 595 nm in a standard plate reader. Data are plotted as relative proliferation normalized to time zero for each condition.
In Vivo Experiments.
451Lu cells transduced with lentiviral supernatants containing miRNA expression constructs (miRH backbone, Systems Biosciences) were resuspended in growth media at a concentration of 2×106 cells/150 μL, aliquoted into eppendorf tubes (150 μL) and maintained on ice until injection. Immediately prior to injection, cell aliquots were mixed with 150 μL Matrigel (Becton Dickinson). Cell/Matrigel suspensions were injected subcutaneously in the right flank of NOD/Shi-scid/IL-2Rγnull (NOG) mice (Jackson Laboratory) (n=9 per group). When tumors were palpable (14 days), length and width measurements were made with calipers 3 times weekly until the animals were sacrificed. Tumor volume was calculated by the following formula: a2*b/2, where a is the width and b is the length. Tumor did not develop in one animal of the miR-374b/b* group for technical reasons and was discarded from subsequent analyses. 6 weeks after cell injection all animals were sacrificed to assess tumor mass and quantify lung metastasis. Tumors were extracted, weighed and imaged. Lungs, liver, spleen, and kidney were removed for analysis of metastasis. Ventral and dorsal macroscopic images of metastasis-bearing lungs were taken with a fluorescent dissecting microscope equipped with a black and white camera. Images were processed in Photoshop (Adobe) by inversion followed by conversion to duotone. Duotone parameters for black were adjusted as follows: 5:10%, 10:30%, 20:70%, 30:100%. Macroscopic metastases were quantified by counting lesions in 4 boxes of equal size (210 by 210 pixels) per lung per side and averaged per mouse. Data were plotted using Graphpad PRISM and significance determined by one-tailed t-testing.
3′UTR Reporter Luciferase Assay.
Full length 3′UTR luciferase reporter clones of indicated genes were purchased (CTTN, PIK3CD, AKT3, MYO9B, RAC1) (Switchgear Genomics). 3′UTR of NCAPG2 was cloned downstream of Renilla luciferase in psiCHECK2 (Promega) cut with XhoI using the In Fusion HD cloning kit (Clontech) following manufacturer's recommendations, followed by sequence verification. Primers used to amplify the NCAPG2 3′UTR were:
293T cells were seeded in 96-well plates at 30,000 cells/well and incubated at 37° C. and 5% CO2 for 16-24 hours. 293T cells were co-transfected with 200 ng 3′UTR reporter plasmid and 50 nM indicated mimic or control miRNA (Dharmacon) using Lipofectamine2000 (Invitrogen) following manufacturer's recommendations. Liposomal complexes of 3′UTR construct and miRNA mimic were prepared separately in 50 μL volumes, then added consecutively to appropriate wells of the 96-well plate. Cells were incubated at 37° C. and 5% CO2 overnight. Media was aspirated from the wells and replaced with PBS. Luciferase assay was performed using Dual Glo Luciferase Assay kit (Promega—for NCAPG2) or Lightswitch Assay Reagent (Switchgear Genomics—for all others) following manufacturer's recommendations. Luminescence was measured in an Envision Multilabel plate reader (Perkin Elmer). Raw ratios of Renilla to Firefly luciferase (NCAPG2) or Renilla luciferase (Switchgear constructs) were normalized to empty vector and are relative to mock treatment (no transfection of miRNA mimic or control). Data represent average readings from replicate experiments (n>3). Data was plotted and significance determined in Graphpad Prism using 1-way ANOVA with Dunnett's multiple comparison post-testing using two different scrambled oligonucleotides as controls:
Western Blotting.
Protein lysates were generated using RIPA buffer (Thermo Fisher) supplemented with protease inhibitors (Complete EDTA-free, Roche) and phosphatase inhibitors (PhosStop, Roche) for 20 minutes on ice, followed by centrifugation for 15 minutes at 13,000 rpm at 4° C. Protein-containing supernatant was transferred to fresh microcentrifuge tubes and stored below −20° C. until further use. Protein was quantified using DC Protein Assay (Biorad) following manufacturer's recommendations, with standard curves generated with bovine serum albumin (Sigma Aldrich). 40 μg or 10 μg (CTTN) of total protein lysate was loaded per lane of 4-20% Novex tris-glycine polyacrylamide mini gels (Invitrogen). SDS-PAGE was run at 150V for 1.5 to 2 hrs. Proteins were transferred to nitrocellulose or PVDF membranes using an iBlot semidry transfer system (Invitrogen) for 7 min on program 3. Membranes were washed 1× in diH2O quickly followed by blocking with 5% non-fat dry milk (Carnation) in PBS for 60 minutes at room temperature. After blocking, membranes were washed briefly with PBS. Membranes were cut appropriately to examine multiple proteins per gel. Membranes were then incubated on a plate shaker overnight at 4° C. with primary antibodies diluted in Tris-buffered saline supplemented with 0.1% Tween-20 (TBS-T). Membranes were washed extensively with TBS-T (minimum 4× for 5 minutes), followed by incuation with appropriate horseradish peroxidase-conjugated secondary antibodies diluted in TBS-T+2% non-fat dry milk for 30-60 minutes at room temperature on a plate shaker. Membranes were washed extensively with TBS-T (minimum 4× for 5 minutes). Signal was detected using ECL Plus Chemiluminescent detection system (GE Healthcare) following manufacturer's recommendations. The following primary antibodies were used: NCAPG2 (Sigma Atlas), Tubulin (Sigma), CTTN (Millipore, clone 4F11), CDC42 (Cell Signaling, #2462), and PIK3CD (Santa Cruz, clone A-8). Secondary antibodies were HRP conjugated anti-mouse or rabbit IgG (GE Healthcare).
The miRNA expression was profiled by microarray of a well-annotated cohort of 92 primary melanomas with minimum patient follow-up of three years for surviving individuals to discover metastasis relevant miRNAs and develop predictive models of recurrence. miRNA expression profiling of FFPE (formalin fixed paraffin embedded)-extracted RNA from primary melanomas was performed. The quantified signals were background corrected (Normexp with offset value 10—Ritchie et al., Bioinformatics, 2007, 23(20):2700-2707) and normalized using the global Lowess (LOcally WEighted Scatterplot Smoothing) regression algorithm (Cleveland, W. S., 1979, J. Amer. Statist. Assoc., 74:829-836).
Forty candidates were selected from the array analyses and tested in a high-throughput in vitro invasion screen in two metastatic melanoma cell lines (501MEL, SK-MEL-147) to identify miRNAs that may functionally impact melanoma metastasis. Thirteen of these miRNAs were further screened across three additional melanoma cell lines (451Lu, SK-MEL-28, SK-MEL-173) in both in vitro invasion and cell proliferation assays. These analyses showed miR-215, miR-374b*, miR-382, miR-516b, and miR-7 as being less expressed in recurrent as compared to non-recurrent and/or thicker as compared to thinner tumors (
451Lu cells, which are highly metastatic to mouse lungs, were used to assess the impact of some of these miRNAs in vivo. miR-382, miR-516b, and miR-7 were ectopically expressed using lentiviral expression constructs containing the pre-miRNA and a GFP tracer to test their ability to suppress lung metastasis in mice. Primary tumor growth was unaffected by expression of miR-382 or miR-7, but was significantly decreased by miR-516b expression (
miR-516b, which also suppressed proliferation in several cell lines in vitro, was found to inhibit tumor growth in vivo. In addition, miR-516b potently suppressed lung metastasis in this model. The possibility that the effect of miR-516b on metastasis is a by-product of reduced tumor size cannot be excluded; however, there was minimal correlation of primary tumor size with metastatic burden (
To better understand the mechanisms by which miR-382, miR-516b, and miR-7 function to suppress invasion and metastasis, the present inventors sought to identify direct targets that could mediate antimetastatic phenotype. Potential downstream mediators of these miRNAs were selected by mRNA array analysis and tested in a secondary invasion screen. mRNA expression array analysis of two melanoma cell lines (SK-MEL-28 and 501MEL) over-expressing a control or individual invasion-suppressive miRNA was performed. Transcripts downregulated by each specific miRNA relative to scrambled control were identified in both cell lines and overlapped this list with that of the miRNA's predicted targets (Targetscan v5.2 [Targetscan] or miRANDA [http://www.microrna.org]) and Clip-seq reads mapped to predicted target binding sites (starbase.sysu.edu.cn) (refs 29-33). 40 candidate genes were selected from the resulting lists. These candidates were tested in this automated in vitro invasion assay by siRNA-mediated depletion in four melanoma cell lines to identify putative miRNA targets whose silencing could also suppress invasion (
A panel of cell lines was tested to ensure effects were applicable to most, if not all, melanomas. Analyses identified five miRNAs (miR-215, miR-374b*, miR-382, miR-516b, and miR-7) that consistently repressed invasion. Of the five miRNAs identified, evidence that three (miR-382, miR-516b, and miR-7) are suppressors of metastasis in vivo was shown. Further, analysis of the clinical data showed miR-374b*, miR-382, miR-516b expression independently correlates with overall survival of these patients, highlighting their importance in melanoma progression (
Further, the miRNAs identified have lower expression in aggressive primary tumors; thus in order to more closely recapitulate what occurs in the primary tumor, miR-382, miR-516b, and miR-7 were inhibited in a poorly invasive cell line to probe for effects on invasion. Inhibition of miR-382 and miR-516b alone or in combination enhanced the invasive capacity of these cells, further supporting the biological relevance of the present findings (
Finding a signature to robustly and accurately classify early stage patients by risk of disease progression is of great clinical importance. In order to address this question from the microRNA expression profile of 91 primary melanomas (discovery cohort), prognostic models of recurrence for stage I/II patients were developed. Risk models were developed using the 70 stage I/II patients (28 recurred, 42 not recurred) with at least three years of follow-up present in this cohort. Prognostic models using only clinical variables showed that the best, which included stage, thickness, and ulceration, had an AUC=64% under the receiver operating characteristic (ROC) curve with none of the predictors significant (
To validate the described model, miRNA expression of an independent cohort of primary melanomas (n=113) was profiled including 69 stage I and II tumors, of which 30 patients recurred while 39 patients have not recurred and 15 of the 39 have at least 3 years of follow-up (Table 3). Applying the classifier developed using the discovery cohort to predict risk for recurrence in this validation cohort yielded an AUC=95%, 95% CI: (0.88, 0.99) of the ROC curve (
In the discovery cohort, most surviving patients have at least 3 years follow-up; therefore almost all patients for model selection were able to be used to ensure good statistical power. However, it is arbitrary to dichotomize patients into recurrent vs non-recurrent at the 3-year mark. Five-year or 10-year recurrence could be also important endpoints for stage I/II patients. Generally, it is of interest to identify a predictive signature capable of robustly classifying patients into low vs. high risk groups corresponding to long vs. short recurrence-free survival (RFS). Towards this goal, miRNAs were ranked not only based on their univariate association with 3-year recurrence with adjustment of tumor stage (logistic model), but also with RFS (Cox PH model). In addition, miRNAs were also ranked by their association with thickness or ulceration, since it is well known that primary tumor thickness and ulceration are associated with melanoma patient RFS and overall survival. Therefore, the 339 highly expressed miRNAs were ranked according to four endpoints: 3-year recurrence (Table 4a. Univariate logistic regression of 3-year recurrence, with adjustment of stage), tumor thickness (Table 4b. Univariate linear regression of thickness, with adjustment of stage), recurrence-free survival (RFS) (Table 4c. Univariate Cox proportional hazard model of recurrence-free survival, with adjustment of stage) and ulceration (Table 4d. Univariate logistic regression of ulceration, with adjustment of stage).
Those miRNAs that are ranked high on such lists provided initial candidates for predictors in selecting multivariate models to predict RFS.
Starting from the top ranked miRNAs in each list in Tables 4a-4d, multivariate logistic regression models were constructed via upward and downward model selection to maximize area under the receiver operating characteristic (ROC) curve with 4-fold cross-validation. Therefore, starting from top ranked miRNAs in the first list (Table 4a), the following Model 1 was selected by maximizing area under the receiver operating characteristic (ROC) curve for 3-year recurrence with 4-fold cross-validation. Note that, within Model 1, hsa-miR-1204, hsa-miR-342-3p, hsa-miR-374b* and hsa-miR-625* are among top 30 of Table 4a. hsa-miR-516b is among top 10 of Table 4b. Model 2 was similarly selected starting from the top ranked miRNAs in Table 4a, by maximizing AUC with cross validation. Note that, within Model 2, hsa-miR-1204, hsa-miR-342-3p and hsa-miR-374b* are among top 10 of Table 4b. hsa-miR-663 and SNORD3A are among top 30 of Table 4c. Model 3 was selected starting from the top ranked miRNAs in Table 4b, by maximizing AUC with cross validation. Note that, within Model 3, hsa-miR-513b is top 1 and hsa-miR-215 is top 24 in Table 4b. hsa-miR-615-3p and hsa-miR-154* are among top 20 of Table 4a. Model 4 was selected starting from top ranked miRNAs in Table 4c, by maximizing AUC with cross validation. Note that, within Table 4(d) the miRNAs hsa-miR-1204, hsa-miR-374b*, hsa-miR-382 and hsa-miR-1276 are among top 15 of Table 4(c)). Each of the four models achieved an area under the ROC between 94% and 96% in the discovery cohort and between 84% and 96% in the validation cohort. Some top-ranked miRNAs remained significant in the selected predictive models while others were replaced by multivariately-significant predictive miRNAs not ranked so high among the four univariately-ranked lists. Given the limited sample size in the discovery stage, we would like to avoid eliminating potentially high-value miRNAs prematurely, thus 4 logistic regression models were selected with a combined total of 21 miRNAs, as the predictive signature set. Each of the four models achieved an area under the ROC between 94% and 96% in the discovery cohort. The risk scores from the four models were averaged to form the final classifier which was discussed earlier. The models (
The data herein support that a paradigm of combining the 1) identification of molecular alterations from large datasets generated from human tissue with 2) a functional screening platform is a more robust way to filter important events in tumorigenesis than either one alone. As such, from this set of differentially expressed miRNAs, 40 candidates were screened in an automated in vitro invasion assay, with careful control of cell proliferation effects, to identify potential metastasis modulators. A panel of cell lines was tested to ensure effects were applicable to most, if not all, melanomas. The analyses identified five miRNAs (hsa-miR-215, hsa-miR-374b*, hsa-miR-382, hsa-miR-516b, and hsa-miR-7) that consistently repressed invasion. Of the five miRNAs identified, it was observed that three (miR-382, miR-516b, and miR-7) are suppressors of metastasis in vivo. Further, analysis of the clinical data shows that miR-374b*, miR-382, miR-516b expression independently correlates with overall survival of these patients, highlighting their importance in melanoma progression (
The miRNAs identified have lower expression in aggressive primary tumors; thus in order to more closely recapitulate what occurs in the primary tumor, miR-382, miR-516b, and miR-7 were inhibited in a poorly invasive cell line to probe for effects on invasion Inhibition of miR-382 and miR-516b alone or in combination enhanced the invasive capacity of these cells, further supporting the biological relevance of the findings (
Source of miR and SNORD3A sequences: http://www.mirbase.org/
The present invention is not to be limited in scope by the specific embodiments described herein. Indeed, various modifications of the invention in addition to those described herein will become apparent to those skilled in the art from the foregoing description. Such modifications are intended to fall within the scope of the appended claims.
All patents, applications, publications, test methods, literature, and other materials cited herein are hereby incorporated by reference in their entirety as if physically present in this specification.
This application claims priority from U.S. Provisional Application Ser. No. 61/647,471, filed on May 15, 2012, which is incorporated herein by reference in its entirety.
This invention was funded, in part, by Department of Defense (DOD) Collaborative Award CA093471. Accordingly, the U.S. government has certain rights to this invention.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/US13/41216 | 5/15/2013 | WO | 00 |
| Number | Date | Country | |
|---|---|---|---|
| 61647471 | May 2012 | US |
