CN107299129B

CN107299129B - Application of circulating nucleic acid as breast cancer biomarker

Info

Publication number: CN107299129B
Application number: CN201610234813.6A
Authority: CN
Inventors: 不公告发明人
Original assignee: Kunshan Industrial Technology Research Institute Of Small Nucleic Acid Biotechnology Research Institute Co ltd
Current assignee: Kunshan Industrial Technology Research Institute Of Small Nucleic Acid Biotechnology Research Institute Co ltd
Priority date: 2016-04-15
Filing date: 2016-04-15
Publication date: 2021-07-27
Anticipated expiration: 2036-04-15
Also published as: CN107299129A

Abstract

The invention relates to application of circulating nucleic acid as a breast cancer biomarker. Through a large amount of literature research, 96 miRNAs which are differentially expressed in breast cancer tissues and body fluids or have function verification at the breast cancer cell line level are screened out to be used as candidate genes for research; particularly, the invention researches the prognosis relation of circulating miRNA and breast cancer long-term follow-up population, and binary regression analysis based on machine training of 182 patients reveals that 2-10miRNA combinations have distinguishing effect on the prognosis of breast cancer patients, and multifactor Cox regression analysis shows that the combination of 2-7 miRNAs is used as an independent risk factor for disease-free survival and distant disease survival of the breast cancer patients, and especially in lymph node negative patients, the combination of 4-5 miRNAs has better prediction effect on the survival of high-risk and low-risk breast cancer patients. This study provides the possibility for the application of circulating mirnas as tumor markers.

Description

Application of circulating nucleic acid as breast cancer biomarker

Technical Field

The present invention relates to the use of circulating nucleic acids as biomarkers for breast cancer. Specifically, the invention relates to the application of circulating microRNA (miRNA) in breast cancer prognosis; more specifically, the invention relates to an application of a combination of 2-10 different miRNAs in breast cancer prognosis, which comprises preparing a kit for breast cancer prognosis based on the combination of 2-10 different miRNAs, a microarray and a method for predicting breast cancer prognosis by using the kit and the microarray.

Background

Breast cancer is the cancer with the highest incidence in women worldwide, accounting for 23% of all cancers, and is also the leading cause of cancer death in women. According to the statistics of the world health organization 2012 (GLOBOCAN 2012), the global incidence of breast cancer increased by 20% and the mortality increased by 14% during the last four years compared to 2008. In European and American countries, the annual incidence rate of female breast cancer is 100-110/10 ten thousands. Although the total annual incidence rate of breast cancer of Chinese women is 30/10 ten thousand, which is about 1/3 of European and American countries, the annual death rate is close to that of European and American countries. Moreover, the incidence of breast cancer of Chinese women is on a remarkable rising trend in recent 10 years, and is increasing at a rate of 3% per year, and the incidence of breast cancer of cities is increasing at a rate of 7.5% per year. Meanwhile, the onset age of breast cancer of Chinese women shows a youthful trend, the onset peak age is 40-50 years old, 10 years earlier than that of western countries, and the health of women in China is seriously threatened.

In addition, breast cancer is a highly heterogeneous disease, mainly manifested by its different genotypes, clinical features, pathological processes and different responses to drugs and survival. Different types of breast cancer patients have different survival and different molecular patients have different drug reactions. In addition to the heterogeneity manifestation among different breast cancer patients, recent studies have found that heterogeneity exists within the tumor, i.e., the gene level changes vary greatly among different cells derived from the same breast cancer patient tissue. The phenomenon that the breast cancer patients with the same molecular pathological type receive the same treatment but have different effects can be explained to a certain extent; on the other hand, due to the high heterogeneity of breast cancer individuals, clinically patients of the same stage and pathological type may differ significantly in prognosis and responsiveness to adjuvant therapy, suggesting that they are still of great importance and far in the prognosis of breast cancer.

In particular, in early stage breast Cancer in which axillary lymph nodes are negative, 70% of patients can survive for a long time without any adjuvant treatment, but according to the national Comprehensive Cancer network (nccn) guidelines, more than 80% -90% of patients with early stage breast Cancer receive adjuvant treatment and are severely over-treated. Therefore, the high-risk breast cancer patients are searched for targeted auxiliary treatment, so that the life cycle of the patients can be prolonged, and the over-treatment of low-risk people can be avoided.

To date, the classical risk factors for breast cancer prognosis have been mainly based on histological level, including Tumor size, grade and lymph Node status, and the independent prognostic manifestations of these factors have been integrated into TNM (Tumor, Node, Metastasis) staging and npi (nottingham cognitive index) systems. However, these conventional pathological indicators have their own limitations and are not suitable as prognostic indicators for the clinical outcome of patients. The molecular level marker is more effective for the differentiation of breast cancer patient subtypes, and can be used as a marker for the prognosis of breast cancer. For example, immunohistochemical factors IHC4 (including ER, PR, HER2, and Ki67) have guiding significance for the treatment of breast cancer, but these risk factors have limitations in differentiating between high-risk and low-risk breast cancer patients and, subject to intratumoral heterogeneity in breast cancer patients, the molecules described above reflect poorly patient-specific characteristics.

Therefore, there is an urgent need in the art to find new prognostic markers for breast cancer to accurately classify its heterogeneity. In this regard, the detection of molecular or cellular levels in body fluids is of increasing interest with its unique advantages. Firstly, the sampling does not need tissues of primary tumor or metastasis, so that traumatic material taking and possible diffusion caused in the material taking process are avoided; secondly, sampling can be carried out at different time for a long time along with the treatment process, so that the detection of drug effect, tumor recurrence and the like is facilitated; again, based on the detection of genomic and transcriptome levels in body fluids, it is possible to eliminate the problem of intra-tumor heterogeneity.

Wherein the miRNA is an endogenous non-coding RNA with the length of about 22 nt. Besides playing a role in the normal physiological process, miRNA also plays an important role in the development and proliferation processes of tumor cells by regulating the expression of cancer-promoting genes or cancer-inhibiting genes. Moreover, studies have shown that differential expression of mirnas can distinguish breast cancer tissues from paracancerous tissues, suggesting that miRNA may become a new marker for breast cancer diagnosis, prognosis and prediction, but traumatic detection limits the application of tissue mirnas as markers to some extent. Therefore, the discovery of circulating miRNA in body fluid opens a new space for the research of novel noninvasive markers of breast cancer.

Compared with traditional marker molecules, circulating miRNAs in body fluids have the following advantages: 1) and (3) more stable: the circulating miRNA has good stability in blood plasma and blood serum, and the content of the circulating miRNA still keeps relatively stable after being treated under various extreme conditions such as boiling, repeated freeze thawing, strong acid, strong base, DNase, RNase and the like; 2) more sensitive: researches find that some miRNA changes in early stage of diseases and can indicate the occurrence of the diseases, and the development of nucleic acid in-vitro amplification technology enables the detection of low-abundance molecules to be possible; 3) more specifically: the miRNA has the specificity of tissue expression and pathological process, different diseases have respective specific circulating miRNA expression profiles, and sequence specificity amplification based on base pairing avoids technical false positive; 4) more convenient: compared with the detection of protein markers which need to screen and prepare specific antibodies, miRNA can be directly detected.

However, although the research on circulating miRNA related to breast cancer has been based on a certain foundation, the current results mostly focus on the role of differentiating breast cancer patients from healthy people, i.e., the research on circulating miRNA as an early diagnosis marker of breast cancer, and the research on circulating miRNA as a prognosis marker of breast cancer, although there is a little research on the correlation between metastasis, molecular subtype and the like of breast cancer patients, the existing research has a small sample amount, and the repeatability of results among different subject groups is poor. The research on clinical results and survival conditions of circulating miRNA and breast cancer patients is still few, and especially the research on patients who follow-up for a long time based on larger samples is very deficient.

Disclosure of Invention

In order to solve the problems, the invention researches the relation of circulating miRNA used as a prognostic marker of breast cancer, particularly the relation of circulating miRNA and recurrence survival, and researches 300 cases of serum of breast cancer patients (with 7-year follow-up data) in a breast cancer sample bank and 200 cases of healthy control serum from physical examination population by cooperating with the breast cancer prevention and treatment center of Beijing university tumor hospital. Through comprehensive literature research, 96 candidate miRNA molecules related to the occurrence of breast cancer diseases or differentially expressed in tissues are determined. The expression profiles of 96 candidate mirnas in the sera of 182 breast cancer patients were analyzed by using a machine-trained binary regression model (Binreg analysis) developed by the university of ducker. 40 of them are test sets, and 142 are verification sets. The combination of 2-10 miRNAs with good distinguishing effect is obtained by screening through Principal Component Analysis (PCA), Probit regression curve fitting and cross validation Analysis. Then, in 142 cases of the validation samples, the Disease-free survival (DFS) and Distant Disease-free survival (DDFS) analysis was performed on the high-risk and low-risk breast cancer patients predicted by the combination of 2-10 mirnas, and as a result, the combination of 4-5 mirnas was found to have the best survival prediction effect. Multifactor Cox regression analysis showed that combinations of 2-7 mirnas were independent risk factors for breast cancer prognosis, with the highest combined risk ratio of 4-5 mirnas. Then, in the survival analysis of lymph node negative and positive patients, the combination of 4-5 miRNAs has a better prediction effect on the survival of lymph node negative breast cancer patients, thereby completing the invention.

According to a first aspect of the invention, the invention provides a kit for predicting breast cancer prognosis, which is characterized in that the kit comprises a circulating nucleic acid detection reagent, wherein the circulating nucleic acid comprises circulating miR-203 and circulating miR-205.

Preferably, the circulating nucleic acid further comprises one or more of the following circulating mirnas: circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452 and circulating miR-193 b.

According to a second aspect of the invention, the invention provides a use of a circulating nucleic acid detection reagent in the preparation of a kit for predicting breast cancer prognosis, wherein the circulating nucleic acid comprises circulating miR-203 and circulating miR-205.

According to a third aspect of the invention, the invention provides a microarray for predicting breast cancer prognosis, which is characterized in that the microarray comprises a circulating nucleic acid detection probe, wherein the circulating nucleic acid comprises circulating miR-203 and circulating miR-205.

According to a fourth aspect of the invention, the invention provides the use of a circulating nucleic acid detection probe in the preparation of a microarray for the prognosis of breast cancer, wherein the circulating nucleic acid comprises circulating miR-203 and circulating miR-205.

According to a fifth aspect of the invention, there is provided use of a circulating nucleic acid in predicting breast cancer prognosis, characterised in that the circulating nucleic acid comprises circulating miR-203 and circulating miR-205.

According to a sixth aspect of the invention, the invention provides a method for predicting breast cancer prognosis, which is characterized in that the method comprises a step of detecting circulating nucleic acids in body fluid, wherein the circulating nucleic acids comprise circulating miR-203 and circulating miR-205.

Advantageous effects

According to the invention, the prognosis relation between circulating miRNA and breast cancer population with long-term follow-up is firstly researched by utilizing large sample follow-up population, and the specific combination of 2-10miRNA is found to have good differentiation on breast cancer patients with different prognoses. Wherein, the combination of 2-8 miRNAs predicts that the disease-free survival (DFS) and the disease-free survival (DDFS) of the breast cancer patients with high and low risks are significantly different. In particular, the combination of 4-5 mirnas is the best predictor of survival prediction, and can be used as an independent predictor of prognosis of high-risk and low-risk breast cancer patients. Compared with the traditional prognostic risk factors (such as lymph node state, tumor size and the like) in the histological level, the specific miRNA combination can more conveniently and remarkably predict the survival effect of the breast cancer patient, provides theoretical guidance for prognosis evaluation of the breast cancer and selection of a treatment scheme, and has practical and precious application value.

Drawings

FIG. 1 shows the relative expression of circulating miRNAs normalized by an internal reference (miR-103a/miR-132) in the serum of a breast cancer patient. BC-1 to BC-4 represent different breast cancer patients.

Figure 2 shows principal component analysis of good and poor prognosis breast cancer patients in the training set. Breast cancer patients are scattered based on 2-10miRNA combination signatures. Factor 1 is the distinguishing element and factor 2 is the display. The clinical prognosis effect of each patient is represented by a number, wherein the numbers 1 to 20 are clinically good patients, and the numbers 21 to 40 are clinically poor patients.

FIG. 3 shows a leave-one-out cross-validation analysis in the training set based on the results of the factorial regression analysis. The X-axis is the total score, the Y-axis is the classification probability corresponding thereto, and the dashed line is the probability corresponding to the 90% confidence interval. Wherein a metagene score greater than 0 is a patient with poor predicted prognosis and less than 0 is a patient with good predicted prognosis. The clinical prognosis effect of each patient is represented by a number, wherein the numbers 1 to 20 are clinically good patients, and the numbers 21 to 40 are clinically poor patients.

FIG. 4 shows the results of Kaplan-Meier Disease Free Survival (DFS) analysis of two groups of patients with high risk and low risk predicted by 2-10miRNA combinations in the test samples. Significance analysis was performed on survival differences by Log-rank.

FIG. 5 shows the results of Kaplan-Meier distance free disease survival (DDFS) analysis of two groups of patients with high risk and low risk predicted by 2-10miRNA combinations in the validation sample. Significance analysis was performed on survival differences by Log-rank.

FIG. 6 shows the results of Kaplan-Meier DFS and DDFS analysis of lymph node positive (A, B) and negative (C, D) breast cancer patients in the validation set, with significance analysis of survival differences by Log-rank, for high-risk and low-risk patients predicted based on 5 miRNA combinations.

Detailed Description

1. Extraction of circulating miRNAs

Compared with miRNA in tissues or cells, the content of circulating miRNA in body fluid is much lower, meanwhile, the miRNA is only 22-25nt, small segments are not easy to enrich in the extraction process, and the circulating miRNA exists in a form of being combined with protein or vesicles, so that great difficulty is brought to extraction and purification of the circulating miRNA.

The extraction of circulating miRNA mainly adopts a reagent for liquid extraction and a kit based on an adsorption column. The reagents are commonly used

And

LS Reagent (Invitrogen), after combining serum or plasma with Reagent, removing protein and other factors which may affect subsequent reaction by phenol/chloroform extraction, and then carrying out ethanol precipitation, but different methods have differences in the process, and miRNA is lost to different degrees.

Another extraction method is mainly based on column binding kit, and is commonly used

miRNA isolation kit (Ambion), mirVana PARIS kit (Ambion) and

miRNA kit (QIAGEN). These kits mainly distinguish the sizes of RNAs by their affinities to silica or glass fiber columns after mixing the RNAs with alcohols of different concentrations, and usually enrich RNAs of 200nt or less. Compared with

The reagent means, the kit has certain advantages, mainly the relatively small amount of the sample required at the beginning.

2. Detection of circulating mirnas

The strategy for the study of circulating mirnas as disease markers is generally divided into two steps: (1) discovery processes based on high-throughput or multigene screening; (2) the results were verified by the qRT-PCR method.

The high-throughput screening method mainly comprises miRNA microarray, deep sequencing and qRT-PCR array. Each of these three approaches has its advantages and disadvantages. miRNA microarray is a microarray of a large number of DNA sequences immobilized on a solid substrate, and is analyzed by fluorescent signals of probes hybridized to target sequences. However, because the sequence of the miRNA is short and the similarity of the miRNA sequences of the same family is high, the false positive rate of the miRNA microarray for expression profiling analysis is high. deep sequencing can solve the problems of short miRNA sequence and high homology, and can research and discover the polymorphism of miRNA and unknown miRNA, but deep sequencing requires a large amount of RNA, a large amount of body fluid is required for RNA extraction, and meanwhile deep sequencing costs a lot, which can limit the incorporation of multiple samples in the discovery period. In many studies, sequencing is performed by mixing a fluid sample, but this method may make some miRNA differences difficult to detect, and the method requires complicated data analysis and is time-consuming. miRNA qRT-PCR array is the shortest time consuming of the three methods, and it is analyzed by qRT-PCR on 754 known mirnas in two 384-well plates. However, because of the low abundance of circulating mirnas, pre-amplification is usually performed before qRT-PCR array, which may cause some mirnas to be supersaturated in the detection and introduce bias. Meanwhile, the method of qRT-PCR array needs a proper internal reference for correction, but no proper unified internal reference is found in the current research of circulating miRNA.

The validation phase of circulating miRNA studies is typically performed by fluorescent quantitative-reverse transcription PCR (qRT-PCR). Because the miRNA sequence is short, which is a difficulty of detection, two strategies are mainly adopted for the detection, wherein one strategy is to adopt a Stem-loop primer (Stem-loop primer) in reverse transcription to obtain a cDNA with an extended sequence; another strategy is to first extend the poly (A) tail of miRNA and then reverse transcribe the extended RNA using oligo dT. There are different ways to detect miRNA after obtaining cDNA. Mainly comprises a Taqman probe method based on sequence specificity, an LNA probe and a non-specific probe

Green dye method.The sensitivity of the Taqman probe and LNA probe methods is high, but the cost is high;

the Green dye method has relatively low sensitivity, but low cost, and is convenient for large-scale verification and analysis.

3. Internal reference selection for circulating mirnas

In the research of circulating miRNA, the relative quantification is almost carried out by adopting qRT-PCR in the verification stage, and the proper internal reference selection is crucial to the analysis of data, but at present, no housekeeping gene in serum or plasma can be used as a stable internal reference, and the discovery and selection can refer to the research method of tissue and cell level, and the following methods are mainly adopted:

(1) genes reported in literature are selected as internal references, mainly miR-16 and small nuclear RNA RNU6(U6), RNU44 and RNU 48. Although the above genes are widely used as internal references, there are some studies that find that their levels vary under different physiological or pathological conditions. And the other is that exogenous miRNA is added as internal reference in the extraction process, such as nematode miRNA cel-miR-39, cel-miR-54 and cel-miR-328, so that the extraction efficiency can be detected.

(2) Screening the gene with stable expression as internal reference. The main analytical methods are geNorm and NormFinder. GeNorm is the ratio of one gene to the other, assuming that the mean of the standard deviations of all ratios is defined as the M value, and the two genes with the smallest M value are combined to the final reference, see Vandersampele, J., et al, Accurate normalization of real-time quantitative RT-PCR data by geometrical averaging of multiple internal control genes, Genome Biol,3(7), RESEARCH0034 (2002). NormFinder is the assessment of the stability of genes (Standard definition, SD), see the literature Andersen, C.L., J.L.Jensen, and T.F.Orntoft, Normalization of real-time quantitative transcription-PCR data a model-based variance estimation from identification genes for Normalization, applied to adaptor and color calculator data, Cancer Res,64, 5245-5250 (2004).

(3) The mean value of all gene expressions was used as an internal control. This approach presupposes that the RNA addition is the same for different samples.

4. Data processing and statistical method

(1) Screening of internal reference

miRNA stability analysis was performed on the expression profiles of 96 mirnas from 25 breast cancer patients and 20 healthy controls by the miniserum immune extraction multiplex qRT-PCR method (see detailed description below). Evaluation was performed using geNorm and NormFinder in the geneex software, and the reference miRNA was selected and normalized by the expression of other mirnas, i.e. Δ Ct ═ Ct^{target miRNA}-Ct^{reference miRNA}Relative expression amount of 2^-ΔCt。

(2) Machine learning binary regression model analysis

In studies of circulating mirnas as prognostic markers for breast cancer, selection of miRNA combinations was analyzed by the BinReg 2.0 software package based on the Bayesian binding regression algorithm. This analytical method was developed by the university of duck in 2001 and was carried out by MATALAB. The analysis of BinReg mainly comprises the steps of dividing samples into a training set and a verification set, establishing a classification model through known information of the training set, and then verifying the classification combination in the verification set. Due to the limitation of the number of experimental samples, overfitting may occur when a model is trained directly using Maximum Likelihood Estimation (MLE), resulting in a reduction in the generalization capability of the model. Therefore, this study was analyzed using the Bayesian framework (Bayesian Regularization) to select γ_iThe prior distribution of (i ═ 1, 2, …, r) is a standard normal distribution N (0, 1) independent of each other, and γ is obtained from a posterior distribution ═ likelihood function × prior distribution_iPosterior distribution of (2). The posterior distribution is then sampled using standard MCMC (Markov Chain Monte Carlo) methods, and an estimate of the posterior distribution is obtained using the sampled data. Finally, model parameters are obtained according to the principle of maximum a posteriori distribution (MAP).

The model selection process is as follows: the parameters to be selected are the space dimensions r after dimension reduction, namely the number of singular values, and the model is verified by adopting a standard leave-one-out cross validation (LOOCV) method.Specifically, for each model, an example sample x is selected_iAs a verification set, the residual samples are used as a training set, and the samples x to be verified are predicted through a trained model_iThe result is denoted as I (z)_i＝z_predict) The prediction result is correctly 1 and the error is 0. Traversing all training set samples by the verification set to obtain the accuracy of model prediction as follows:

and (5) utilizing the lambda as an evaluation standard of different models, thereby completing the process of model selection.

(3) Statistical analysis

The prognosis of breast cancer patients is predicted by circulating miRNA combination, the patients are divided into high-risk groups and low-risk groups, the classification effect of the two groups of people is evaluated by using Sensitivity and Specificity, and the survival condition of the two groups of people is another important index. And (3) analyzing the predicted high-risk group and low-risk group for disease-free survival (DFS) and disease-free survival (DDFS) by a Kaplan-Meier method respectively, and performing significance analysis on the survival difference of the two groups of people by a Log-rank test. Here, DFS means that no in situ recurrence, distant metastasis and death events occurred by the end of the follow-up; whereas DDFS means that no distant metastasis and death events occurred by the end of the follow-up visit. Patients who remained alive at the end of the follow-up were considered as Censored data, as were breast cancer patients who died for other reasons. Survival analysis of circulating miRNA combinations and other pathological features was performed by multivariate Cox risk regression of SPSS 18.0 to assess whether risk factors are independent risk factors for survival prediction.

All statistical tests were Two-tailed (Two-side tailed) tests, with differences of P <0.05 being statistically significant.

5. Establishment of trace serum extraction-free multiple qRT-PCR method

To accomplish the detection of trace amounts of serum mirnas, in addition to the methods described in

items

1 and 2 of this section, the present study further attempted to eliminate the extraction step, reverse transcription directly after serum heating, followed by fluorescent quantitative PCR, and surprisingly found that this method can detect the expression of circulating mirnas more efficiently. Furthermore, the method can avoid the loss of RNA in the process of extracting serum or plasma, and meanwhile, the method for trace detection can also be convenient to use a biological sample library for research, so that the experimental result is more reliable, and the information is more comprehensive.

In the discovery phase of research markers, high-throughput miRNA detection methods are very important. In studies of tissue sample mirnas, Multiplex cDNA synthesis (Multiplex cDNA synthesis) simultaneous detection of 48 different mirnas has proven feasible. Based on the above, the research explores a method for detecting circulating miRNA by trace serum extraction-free multiplex qRT-PCR. The 10miRNA multiplex cDNA synthesis was evaluated by a serum free-draw method. In the serum of a healthy control, 9 miRNAs and U6 can be detected by the method, the Ct value is 25-35, the abundance of miR-150 is highest, the abundance of miR-191 is lowest in the miRNAs to be detected, and the qRT-PCR product is subjected to sequencing verification. Subsequent studies found that 20 miRNA multiplex cDNA synthesis was also feasible and the stability of the method was tested by gradient dilution of the cDNA. After 2, 4 and 8 times of gradient dilution are carried out on cDNA, Ct values detected by 4 miRNAs are in a good linear relation, and a correlation coefficient is between 0.896 and 0.999.

On this basis, the synthesis of 20 miRNA multiplex cDNAs is traversed to the synthesis of 96 miRNA multiplex cDNAs. Specifically, a minimal amount of serum-immune extraction multiplex qRT-PCR primers were prepared by mixing specific stem-loop RT primers for 96 different mirnas (10 μ l each, 50 μ M each). The mixed primers were concentrated and nuclease-free ddH was used₂O was adjusted to a final concentration of 5. mu.M. Mu.l of serum sample was mixed with 2.5. mu.l of nuclease-free ddH₂O mixed and heated to 60-80 deg.C (e.g., 70 deg.C) followed by reverse transcription using 2. mu.l of the aforementioned minimal serum extraction multiplex qRT-PCR primers via reverse transcription kit (Transgene, Beijing, China) (final volume 20. mu.l).

After obtaining cDNA, ddH was applied without nuclease₂O the cDNA product was diluted 10-fold and centrifuged at 13200rpm for 5min to collect the supernatant. Followed byAnd performing fluorescent quantitative PCR reaction by using the supernatant. Mu.l of cDNA was used in a 25. mu.l qPCR reaction, which also contained SYBR Green I dye and miRNA-specific detection primers. The PCR product was cloned into pGM-T vector (Tiangen, Beijing, China) and verified by sequencing. Serum samples from each patient were subjected to three replicates.

Among them, the miRNA-specific stem-loop primers used in the reverse transcription process (such as, but not limited to, the miRNA-specific stem-loop primers shown in table 16) can be appropriately designed and synthesized by those skilled in the art according to the known miRNA sequences; likewise, miRNA-specific detection primers used in PCR processes (such as, but not limited to, the miRNA-specific detection primers shown in table 17) can also be appropriately designed and synthesized by one skilled in the art based on known miRNA sequences.

6. Kits and microarrays of the invention

The kit and the microarray are used for detecting circulating nucleic acid, wherein the circulating nucleic acid comprises circulating miR-203 and circulating miR-205; preferably, the circulating nucleic acid further comprises one or more (e.g., 1, 2, 3, 4, 5, 6, 7, and all 8) of the following circulating mirnas: circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452 and circulating miR-193 b; more preferably, the circulating nucleic acids comprise circulating miR-203, circulating miR-205, circulating miR-450a and circulating miR-16; most preferably, the circulating nucleic acids comprise circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16 and circulating miR-34 a.

Without limitation, a circulating nucleic acid detecting primer or a circulating nucleic acid detecting probe may be used to prepare the kit or microarray of the present invention, for example, one skilled in the art may prepare the kit or microarray of the present invention based on the sequences shown in table 16 and table 17, but the present invention is not limited thereto.

In addition, the kit of the present invention may further comprise a DNA purification reagent (e.g., Nucleon)^TMKit, lysis buffer, protease solution, etc.); PCR reagents (e.g., reaction buffer, thermostable polymerase, dNTPs, etc.), but the present invention is not limited theretoHere, the process is repeated. Probes for detecting circulating nucleic acids can be labeled with different detectable means. Such detectable means refer to compounds, biomolecules or biomimetic materials that can be conjugated, linked or attached to the probes to provide quantifiable indices (e.g., density, concentration, quantity, etc.). Examples of detectable means include fluorescent labels, luminescent materials, bioluminescent materials and radioisotopes, but the invention is not limited thereto. Details and choices of detectable tools will be apparent to those skilled in the art.

Various aspects of the invention may be defined by any one of the following numbered paragraphs:

1. a kit for predicting breast cancer prognosis, comprising a circulating nucleic acid detection reagent, wherein the circulating nucleic acid comprises circulating miR-203 and circulating miR-205.

2. The kit of paragraph 1, wherein the circulating nucleic acid further comprises one or more of the following circulating mirnas: circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452 and circulating miR-193 b.

3. The kit of paragraph 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205 and circulating miR-450 a.

4. The kit of paragraph 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a and circulating miR-16.

5. The kit of paragraph 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16 and circulating miR-34 a.

6. The kit of paragraph 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a and circulating miR-20 a.

7. The kit of paragraph 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a and circulating miR-373.

8. The kit of paragraph 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, and circulating miR-519 c.

9. The kit of paragraph 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, and circulating miR-452.

10. The kit of paragraph 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452, and circulating miR-193 b.

11. The kit according to any one of paragraphs 1 to 10, wherein the detection object of the kit is a body fluid.

12. The kit of paragraph 11 wherein the body fluid is serum or plasma.

13. The kit of paragraph 12, wherein, in the detection of circulating nucleic acid in the serum using the kit, the serum is not extracted, but the serum is directly heated and then reverse-transcribed to obtain cDNA for detection.

14. The kit of any of paragraphs 1-13, wherein the kit further comprises reagents for detecting at least one of miR-107, miR-103a and miR-132.

15. The kit of paragraph 14 wherein the kit contains reagents for detecting miR-103a and miR-132.

16. The kit of any of paragraphs 1-15, wherein said kit comprises a circulating nucleic acid detection primer selected from at least one of the primers shown as SEQ ID No. 14-SEQ ID No. 40.

17. Use of a circulating nucleic acid detection reagent in preparation of a kit for predicting breast cancer prognosis, wherein the circulating nucleic acid comprises circulating miR-203 and circulating miR-205.

18. The use of paragraph 17, wherein the circulating nucleic acid further comprises one or more of the following circulating mirnas: circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452 and circulating miR-193 b.

19. The use of paragraph 18, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205 and circulating miR-450 a.

20. The use of paragraph 18, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a and circulating miR-16.

21. The use of paragraph 18, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16 and circulating miR-34 a.

22. The use of paragraph 18, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a and circulating miR-20 a.

23. The use of paragraph 18, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a and circulating miR-373.

24. The use of paragraph 18, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, and circulating miR-519 c.

25. The use of paragraph 18, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, and circulating miR-452.

26. The use of paragraph 18, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452, and circulating miR-193 b.

27. The use according to any of paragraphs 17-26, wherein the subject of detection of the kit is a body fluid.

28. The use of paragraph 27 wherein the body fluid is serum or plasma.

29. The use of paragraph 28 wherein, in the detection of circulating nucleic acids in said serum using said kit, said serum is not extracted but directly heated and then reverse transcribed to obtain cDNA for detection.

30. The use of any of paragraphs 17-29, wherein the kit further comprises reagents for detecting at least one of miR-107, miR-103a and miR-132.

31. The use of paragraph 30 wherein the kit comprises reagents for detecting miR-103a and miR 132.

32. The use of any of paragraphs 17-31, wherein the kit comprises a circulating nucleic acid detection primer selected from at least one of the primers set forth in SEQ ID No.14 to SEQ ID No. 40.

33. A microarray for predicting prognosis of breast cancer, comprising a circulating nucleic acid detection probe, wherein the circulating nucleic acid comprises circulating miR-203 and circulating miR-205.

34. The microarray of paragraph 33, wherein said circulating nucleic acids further comprise one or more of the following circulating mirnas: circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452 and circulating miR-193 b.

35. The microarray of paragraph 34 wherein said cycling nucleic acids comprise cycling miR-203, cycling miR-205 and cycling miR-450 a.

36. The microarray of paragraph 34 wherein said cycling nucleic acids comprise cycling miR-203, cycling miR-205, cycling miR-450a and cycling miR-16.

37. The microarray of paragraph 34 wherein said cycling nucleic acids comprise cycling miR-203, cycling miR-205, cycling miR-450a, cycling miR-16 and cycling miR-34 a.

38. The microarray of paragraph 34 wherein said cycling nucleic acids comprise cycling miR-203, cycling miR-205, cycling miR-450a, cycling miR-16, cycling miR-34a and cycling miR-20 a.

39. The microarray of paragraph 34 wherein said cycling nucleic acids comprise cycling miR-203, cycling miR-205, cycling miR-450a, cycling miR-16, cycling miR-34a, cycling miR-20a, and cycling miR-373.

40. The microarray of paragraph 34 wherein said cycling nucleic acids comprise cycling miR-203, cycling miR-205, cycling miR-450a, cycling miR-16, cycling miR-34a, cycling miR-20a, cycling miR-373, and cycling miR-519 c.

41. The microarray of paragraph 34 wherein said cycling nucleic acids comprise cycling miR-203, cycling miR-205, cycling miR-450a, cycling miR-16, cycling miR-34a, cycling miR-20a, cycling miR-373, cycling miR-519c and cycling miR-452.

42. The microarray of paragraph 34 wherein said cycling nucleic acids comprise cycling miR-203, cycling miR-205, cycling miR-450a, cycling miR-16, cycling miR-34a, cycling miR-20a, cycling miR-373, cycling miR-519c, cycling miR-452, and cycling miR-193 b.

43. The microarray of any of paragraphs 33-42, wherein the subject of said microarray is a bodily fluid.

44. The microarray of paragraph 43 wherein the bodily fluid is serum or plasma.

45. The microarray of paragraph 44 wherein, in the detection of circulating nucleic acids in the serum by the microarray, the serum is not extracted but directly heated and then reverse transcribed to obtain cDNA for detection.

46. The microarray of any of paragraphs 33-45, wherein said microarray further comprises probes for detecting at least one of miR-107, miR-103a and miR-132.

47. The microarray of paragraph 46 wherein the microarray comprises probes for detecting miR-103a and miR-132.

48. Use of a circulating nucleic acid detection probe in the preparation of a microarray for predicting breast cancer prognosis, wherein the circulating nucleic acid comprises circulating miR-203 and circulating miR-205.

49. The use of paragraph 48, wherein said circulating nucleic acid further comprises one or more of the following circulating miRNAs: circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452 and circulating miR-193 b.

50. The use of paragraph 49, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205 and circulating miR-450 a.

51. The use of paragraph 49, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a and circulating miR-16.

52. The use of paragraph 49, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16 and circulating miR-34 a.

53. The use of paragraph 49, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a and circulating miR-20 a.

54. The use of paragraph 49, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a and circulating miR-373.

55. The use of paragraph 49, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, and circulating miR-519 c.

56. The use of paragraph 49, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, and circulating miR-452.

57. The use of paragraph 49, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452, and circulating miR-193 b.

58. The use of any of paragraphs 48-57, wherein the subject of detection of the microarray is a bodily fluid.

59. The use of paragraph 58 wherein the body fluid is serum or plasma.

60. The use of paragraph 59 wherein, in the detection of circulating nucleic acids from the serum using the microarray, the serum is not extracted but is directly heated and then reverse transcribed to obtain cDNA for detection.

61. The use of any of paragraphs 48-60, wherein the microarray further comprises probes for detecting at least one of miR-107, miR-103a and miR-132.

62. The use of paragraph 61 wherein the microarray comprises probes for detecting miR-103a and miR-132.

Examples

The present invention will be described in further detail with reference to examples, but the present invention is not limited to these examples.

1. Clinical sample information

In the study of the prognosis relationship between circulating miRNA and breast cancer patients, 189 breast cancer patients are breast cancer patients treated by the tumor hospital of Beijing university from 2005 to 2009, and serum is sampled before the patients receive treatment. 96 miRNA expression profile detections are carried out on 189 breast cancer patients, and preliminary analysis is carried out on data, wherein the experimental results of 182 breast cancer patients meet the data requirements and are included in the subsequent analysis. The patients were in the age range of 28-85 years with a median age of 52 years. Tumor staging was performed according to the TNM staging protocol of UICC (Union International Conter Le cancer). The lymph node metastasis state is determined by histological examination after lymph nodes are cleared. Of the 182 breast cancer patients, 174 patients had follow-up data with a follow-up time of 5-83 months and a median follow-up time of 58.5 months. This study was supported by the ethical research institute of the tumor hospital, Beijing university and was informed by the participants in written consent.

2. Serum RNA extraction

LS extraction: 500 μ l serum was added to 3 volumes

LS (Invitrogen), denaturation at room temperature, phenol/chloroform extraction, ethanol precipitation, addition of 20. mu.l DEPC H to air-dried RNA₂And O. By using

And (5) carrying out subsequent qRT-PCR after measuring the concentration of ND-1000.

And (3) miRNA extraction kit: 500. mu.l of serum was taken and extracted according to the kit instructions. The main process is as follows: the serum was mixed with an equal volume of denaturing solution and extracted with phenol/chloroform. Mixing with 80% ethanol, passing through glass fiber column for enrichment, washing, and dissolving in 20 μ l DEPC H₂And (4) in O. By using

The subsequent qRT-PCR experiment is carried out after the ND-1000 concentration measurement.

PARIS extraction kit: the process is the same as

miRNA extraction kit.

Detection of serum miRNA expression profile by Human MicroRNA Array

1ml of patient plasma was passed through

miRNA extraction kitRNA extraction was performed as described above. Total RNA (200ng) was mixed with 2 Megaplex reverse transcription primers (Human Pool A)&B, Applied Biosystems), by

MicroRNA reverse transcription kit (Applied Biosystems) was synthesized covering 377 miRNAs and 6 internal controls per Megaplex reverse transcription primer. 7.5. mu.l reaction System containing RNA, 10 × reverse transcription primer, 10 × reverse transcription solution, MgCl₂(25mM) and RNase inhibitor (20U/. mu.l). The reaction process is as follows: 2 minutes at 16 ℃, 1 minute at 42 ℃ and 1 second at 50 ℃, and the three steps are carried out for 40 cycles; followed by heating at 85 ℃ for 5 minutes. The cDNA obtained is directly passed through without pre-amplification

Human MicroRNA Array set A&B (applied biosystems) A384-well fluorescent quantitative PCR reaction was performed. The cDNA was diluted to 50. mu.l and mixed with 50. mu.l

Universal PCR Master Mix (No UNG, Applied Biosystems). The thermal cycling process is as follows: heating at 94.5 deg.C for 10 min; followed by 40 cycles comprising 97 ℃ for 30 seconds and 59.7 ℃ for 1 minute. The above reactions were performed on a 7900HT Fast Real-Time PCR instrument (Applied Biosystems). The cycle number Ct of the resulting miRNA was normalized by the internal control.

4. Micro serum extraction-free multiple qRT-PCR (quantitative reverse transcription-polymerase chain reaction) detection miRNA

Mixing reverse transcription stem-loop primers of 96 miRNAs, mixing 10 μ l reverse transcription primer solution (50 μ M) of each miRNA, drying in a freeze dryer for 1 hr, adding 100 μ l DEPC H₂And O, forming a miRNA multiple reverse transcription primer solution with the final concentration of 5 mu M.

Mu.l of serum was denatured by heating (70 ℃ C.), and 20. mu.l of a reverse transcription reaction containing 2. mu.l of the above-mentioned minute amount of serum without extraction of multiple reverse transcription primers (5. mu.M) was carried out by adding a reverse transcription system (all-type gold, Transgene). After obtaining cDNA, the cDNA was diluted 10 times, and 2. mu.l of the cDNA was used for the subsequent fluorescent quantitative PCR reaction. Fluorescent probeQuantitative PCR reaction in

The experimental procedure was performed on a StepOne Plus instrument as follows: 25 ul of fluorescent quantitative PCR system contained 10 XHotMaster Taq enzyme buffer (Tiangen), 20 uM forward and reverse primers, 2.5U/. mu.l HotMaster Taq enzyme (Tiangen),

Green I (Invitrogen) and DEPC H₂And O. The thermal cycling process is as follows: preheating for 10 minutes at 95 ℃; melting curve collection was then performed after denaturation at 95 ℃ for 15 seconds, annealing at 60 ℃ for 60 seconds (40 cycles). The non-templated qRT-PCR served as a negative control for the experiment.

Example determination of 196 candidate miRNAs

In order to find circulating mirnas that are effective as prognostic markers for breast cancer, the present study screened multiple mirnas associated with breast cancer in already reported databases or studies, and their circulating forms have not been studied. Using this as a candidate gene, expression profiling was performed in a multi-sample training set. Source and selection criteria for candidate genes: 1) in breast cancer and paracarcinoma tissues, chip or functional experiments suggest that the expression of miRNA has difference (>2 times); 2) miRNA related to prognostic indexes such as breast cancer typing and lymph node state; 3) demonstrating in a cell or animal model mirnas that participate in the development of breast cancer; 4) a potentially effective miRNA has been detected in a small sample of serum from breast cancer populations; 5) mirnas are widely present in serum plasma. Through the above analysis, 96 mirnas were screened as candidate molecules for subsequent detection.

Example 2 screening of internal controls and normalization of miRNA expression levels

Randomly selecting 25 breast cancer patients and 20 healthy controls matched with the breast cancer patients in age, detecting 96 miRNA micro serum immune extraction multiplex qRT-PCR expression profiles, and evaluating the stability of the obtained miRNA Ct values through NormFinder and geNorm software.

NormFinder analysis showed that miR-132 is the most stably expressed miRNA in patients and healthy controls, while geNorm analysis suggested that miR-103a/miR-107 is suitable as an internal reference. Evaluation of the stability of the combination of the three miRNAs (miR-132, miR-103a and miR-107) shows that the miR-103a/miR-132 has the strongest stability and is suitable for serving as an internal reference.

After internal controls were selected, remaining mirnas were normalized for breast cancer patients and healthy controls. As shown in figure 1, different mirnas are expressed in serum in varying abundance.

Example 3 acquisition of circulating miRNA expression profiles in breast cancer patients

In order to obtain an effective miRNA combination for predicting breast cancer outcome, this study screened 96 miRNA raw data from 189 breast cancer patients according to the following criteria: 1) the Ct value of circulating miRNA is between 15-35; the Ct value STD of miRNA in the serum repeated experiment of the same patient is less than 2; 2) the Ct values of mirnas were between the mean Ct ± 3 of 189 breast cancer patients. Finally obtaining the expression profiles of 82 serum miRNAs of 182 breast cancer patients meeting the requirements. Clinical information for 182 breast cancer patients is shown in table 1, with a median age of 52 years, of which 55 breast cancer patients had poor clinical outcomes including in situ recurrence, distant metastasis and death.

Ct values of other 80 miRNAs were normalized for subsequent analysis with miR-103a/miR-132 as an internal reference.

TABLE 1182 clinical characteristics of breast cancer patients

Example 4 different miRNA combinations obtained by training set

In order to more accurately find the mirnas with good prediction effect on clinical outcome of breast cancer patients, the study analyzes the normalized 80 miRNA expression profiles of 182 breast cancer patients selected in example 3 through a machine-trained binary regression model. Among 55 cases of breast cancer patients with poor prognosis, 20 cases of patients were randomly selected; and of 127 breast cancer patients with good prognosis, 20 patients with matched clinical characteristics are selected as a control to jointly form a training set, the specific clinical characteristics are shown in table 2, and the rest 142 breast cancer patients are used as a verification set.

TABLE 2 clinical information from 40 breast cancer patients in a training set

In a binary regression model of a training set, a Principal Component Analysis (PCA) based on Singular Value Decomposition (SVD) is performed on the relative expression amounts of 80 mirnas of 40 patients with different prognoses to obtain a combination of 2-10 mirnas, as shown in fig. 2, a factor 1(factor 1) is a distinguishing element, and a factor 2(factor 2) is displayed. The clinical prognosis effect of each patient is represented by a number, wherein the numbers 1 to 20 are clinically good patients, and the numbers 21 to 40 are clinically poor patients. It can be seen that 2-10miRNA combinations are well differentiated for breast cancer patients with different prognoses.

In order to further screen stable and reliable combinations, the study performed Leave-one-out cross validation (Leave-one-out cross validation) on 2-10miRNA combinations obtained by PCA analysis in the training set, the probability of prediction of two types of breast cancer patients by different miRNA combinations is shown in fig. 3, and when the gene score is greater than 0, the prediction is poor prognosis patients, and when the gene score is less than 0, the prediction is good prognosis patients. The clinical prognosis effect of each patient is represented by a number, wherein the numbers 1 to 20 are clinically good patients, and the numbers 21 to 40 are clinically poor patients. Statistical analysis was performed on the accuracy of differentiation of 2-10 mirnas in the training set for breast cancer patients, and as shown in table 3, the combination of 6 mirnas was found to have the highest sensitivity and specificity.

TABLE 32-10 differentiation of miRNA combinations between good and poor prognosis in training set for breast cancer patients

The composition of the different miRNA combinations is shown in table 4.

TABLE 42-10 composition of miRNA combinations and contribution of each miRNA to classification

Ce (coefficience) refers to the coefficient of miRNA in the binary regression equation, i.e. its contribution to patient classification.

In this machine-trained model analysis of binary regression, the combinations of 2-10 mirnas presented in table 4 are the best combinations among the similar combinations for differentiating the prognosis effects of breast cancer patients. Taking 2 miRNA combinations as an example, the machine-trained model combines every two mirnas of 80 mirnas (3160 total), and evaluates the classification ability of the combinations on the patient prognosis effect respectively to obtain the combination with the best differentiation effect as miR-203/miR-205, and other combinations are not presented here. Example 5 differentiation and prediction of survival of different miRNA combinations for validated focused breast cancer patients

Subsequently, the present study evaluated the discrimination of different combinations of 2-10 mirnas in 142 validation sets. Of the 142 validation samples, 35 were poor-prognosis breast cancer patients and 107 were good-prognosis breast cancer patients. The discrimination of the 2-10miRNA combinations on the two types of patients is shown in Table 5, the predicted discrimination of 4 miRNAs and 5 miRNAs is the highest, and the sensitivity and the specificity are 89% and 62% respectively.

TABLE 52-10 differentiation of miRNA combinations on validation of the Subjects with good and poor prognosis in the set

To more directly and objectively assess the association of 2-10miRNA combinations as predictors with clinical outcome, this study performed a correlation analysis of differential miRNA combinations with patient survival. The validation set was divided into two classes of high risk (poor predicted prognosis, greater than 0) and low risk (good predicted prognosis, less than 0) patients according to the predicted metagene score, and these patients were analyzed for Disease Free Survival (DFS) and disease free survival (DDFS) using Kaplan-Meier survival analysis, and the survival difference between the two classes of patients was significantly analyzed by log-rank test, with p <0.05 indicating that the survival difference was statistically significant. As a result, the DFS of the high-risk breast cancer patients and the low-risk breast cancer patients predicted by 2-10miRNA combinations are remarkably different, and the risk ratio (HR) is 2.550-5.796 (figure 4), wherein the DDFS of the high-risk breast cancer patients and the low-risk breast cancer patients predicted by 2-8 miRNA combinations is also remarkably different, and the risk ratio is 2.329-4.756 (figure 5). As can be seen, 2-10miRNA combinations have a distinguishing effect on the prognosis of a breast cancer patient, wherein the combination of 4 miRNAs and 5 miRNAs has the best prediction effect on the poor breast cancer patient with good prognosis, and the distinguishing degree of the 5 miRNA combinations on the two breast cancer patients is better than that of the 4 miRNA combinations by combining the performance in the training set, so that the 5 miRNA combinations (miRNA-16, miR-205, miR-450a, miR-203 and miR-34a) are the best prediction factors (Table 4).

Examples 62-7 miRNA combination as a prognostic independent risk factor for breast cancer

To test whether a combination of 5 mirnas could be an independent predictor of high-risk and low-risk breast cancer patients, this study examined them by Cox risk ratio regression. After balancing age, tumor size, ER, PR, HER2 and lymph node status of the test samples, 5 miRNA combinations were found to be independent risk factors for breast cancer prognosis, as shown in table 6, the predicted risk ratio for DFS was 6.891 with a confidence interval of 2.291-20.724, and a P-value of 0.001; the predicted risk ratio for DDFS was 5.960, confidence interval was 1.899-18.369, and P-value was 0.002. TABLE 6 multifactor COX regression analysis validation DFS and DDFS (5-miRNA) of breast cancer-focused persons

In addition to the combination of 5 mirnas, the multifactor Cox regression analysis described above found that lymph node status was also an independent risk factor for DFS and DDFS in breast cancer patients in this validation sample. To this end, the study performed a one-way risk analysis of DFS and DDFS, respectively, on patients with positive and negative lymph nodes in the test samples, based on the high-risk and low-risk patient cohort predicted by the 5 miRNA combinations by the Kaplan-Meier method (fig. 6). The 5 miRNA combinations were more significant for risk prediction of DFS and DDFS in lymph node negative patients compared to lymph node positive patients (fig. 6A and 6B) (fig. 6C and 6D).

In addition, the research also carries out multifactor Cox regression analysis on other miRNA combinations (Table 7-Table 14), and the result shows that 2-7 miRNA combinations are also independent risk factors of DFS and DDFS of breast cancer patients, but the combined risk ratio of 5 miRNAs and 4 miRNAs is the highest. For the patients with positive and negative lymph nodes in the test sample, based on the high-risk and low-risk patient groups predicted by the 4 miRNA combinations, the Kaplan-Meier method is used for carrying out single-factor risk analysis on the DFS and the DDFS of the patients respectively, the result is similar to that in the figure 6, and the 4 miRNA combinations are more remarkable in risk prediction on the DFS and the DDFS of the patients with negative lymph nodes.

TABLE 7 multifactor COX regression analysis validation DFS and DDFS (2-miRNA) of breast cancer-focused persons

12 breast cancer patients with partial loss of clinical data were removed and 130 breast cancer patients were enrolled in the COX regression analysis.

TABLE 8 multifactor COX regression analysis validation DFS and DDFS (3-miRNA) of breast cancer-focused persons

TABLE 9 multifactor COX regression analysis validation DFS and DDFS (4-miRNA) of breast cancer-focused humans

TABLE 10 multifactor COX regression analysis validation DFS and DDFS (6-miRNA) of breast cancer-focused humans

TABLE 11 multifactor COX regression analysis validation DFS and DDFS (7-miRNA) of breast cancer-focused persons

TABLE 12 multifactor COX regression analysis validation DFS and DDFS (8-miRNA) of breast cancer-focused humans

TABLE 13 multifactor COX regression analysis validation DFS and DDFS (9-miRNA) in breast cancer patients in focus

TABLE 14 multifactor COX regression analysis validation DFS and DDFS (10-miRNA) in breast cancer patients in focus

As shown in tables 6 and 9, the predicted risk ratio of lymph node status to DDFS was 2.898, and p value was 0.023; the predicted risk ratio of 4 mirnas or 5 miRNA combinations for DDFS was 5.960 with a p-value of 0.002. Therefore, compared with the traditional histological classical prognostic risk factor, namely the lymph node state, the noninvasive tumor marker, namely the circulating nucleic acid (4 miRNA or the combination of 5 miRNA) in blood, the prognosis effect of the breast cancer patient is more remarkable. Therefore, the combination of 2-7 miRNAs can be used as a prognosis independent risk factor of a high-risk breast cancer patient, and based on the prognosis independent risk factor, a kit or a microarray for breast cancer prognosis can be developed.

Table 15 sequences of mirnas used in the present invention

miRNA	Sequence (5' ->3’)	SEQ ID NO.
			hsa-miR-203	GUGAAAUGUUUAGGACCACUAG	1
hsa-miR-205-5p	UCCUUCAUUCCACCGGAGUCUG					2
			hsa-miR-450a-5p	UUUUGCGAUGUGUUCCUAAUAU	3
hsa-miR-16-5p	UAGCAGCACGUAAAUAUUGGCG					4
			hsa-miR-34a-5p	UGGCAGUGUCUUAGCUGGUUGU	5
hsa-miR-20a-5p	UAAAGUGCUUAUAGUGCAGGUAG					6
			hsa-miR-373-3p	GAAGUGCUUCGAUUUUGGGGUGU	7
hsa-miR-519c-3p	AAAGUGCAUCUUUUUAGAGGAU					8
			hsa-miR-452-5p	AACUGUUUGCAGAGGAAACUGA	9
hsa-miR-193b-3p	AACUGGCCCUCAAAGUCCCGCU					10
			hsa-miR-107	AGCAGCAUUGUACAGGGCUAUCA	11
hsa-miR-103a-3p	AGCAGCAUUGUACAGGGCUAUGA					12
			hsa-miR-132-3p	UAACAGUCUACAGCCAUGGUCG	13

TABLE 16 sequences of miRNA-specific stem-loop primers used in the invention

TABLE 17 sequences of miRNA-specific detection primers used in the present invention

Claims

1. The application of the circulating nucleic acid detection reagent in preparing a kit for predicting breast cancer prognosis is characterized in that the circulating nucleic acid comprises circulating miR-203, circulating miR-205 and circulating miR-450a, wherein the detection object of the kit is serum or plasma.

2. The use of claim 1, wherein the circulating nucleic acid further comprises one or more of the following circulating mirnas: circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452 and circulating miR-193 b.

3. The use of claim 2, wherein said circulating nucleic acids comprise circulating miR-203, circulating miR-205, circulating miR-450a, and circulating miR-16.

4. The use of claim 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, and circulating miR-34 a.

5. The use of claim 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, and circulating miR-20 a.

6. The use of claim 2, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, and circulating miR-373.

7. The use of claim 2, wherein said circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, and circulating miR-519 c.

8. The use of claim 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, and circulating miR-452.

9. The use of claim 2, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452, and circulating miR-193 b.

10. The use according to any one of claims 1 to 9, wherein, in the circulating nucleic acid detection of the serum using the kit, the serum is not extracted, but the serum is directly heated and then reverse-transcribed to obtain cDNA for detection.

11. The use of any one of claims 1-9, wherein the kit further comprises reagents for detecting at least one of miR-107, miR-103a, and miR-132.

12. The use of claim 10, wherein the kit further comprises reagents for detecting at least one of miR-107, miR-103a and miR-132.

13. The use of claim 11, wherein the kit comprises reagents for detecting miR-103a and miR 132.

14. The use of claim 12, wherein the kit comprises reagents for detecting miR-103a and miR 132.

15. The use according to any one of claims 1 to 9 and 12 to 14, wherein the kit comprises a circulating nucleic acid detection primer selected from at least one of the primers shown as SEQ ID No.14 to SEQ ID No. 40.

16. The use of claim 10, wherein the kit comprises a circulating nucleic acid detection primer selected from at least one of the primers shown as SEQ ID No.14 to SEQ ID No. 40.

17. The use of claim 11, wherein the kit comprises a circulating nucleic acid detection primer selected from at least one of the primers shown as SEQ ID No.14 to SEQ ID No. 40.

18. The application of a circulating nucleic acid detection probe in preparing a microarray for predicting breast cancer prognosis is characterized in that the circulating nucleic acid comprises circulating miR-203, circulating miR-205 and circulating miR-450a, wherein a detection object of the microarray is serum or plasma.

19. The use of claim 18, wherein the circulating nucleic acid further comprises one or more of the following circulating mirnas: circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452 and circulating miR-193 b.

20. The use of claim 19, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, and circulating miR-16.

21. The use of claim 19, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, and circulating miR-34 a.

22. The use of claim 19, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, and circulating miR-20 a.

23. The use of claim 19, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, and circulating miR-373.

24. The use of claim 19, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, and circulating miR-519 c.

25. The use of claim 19, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, and circulating miR-452.

26. The use of claim 19, wherein the circulating nucleic acid comprises circulating miR-203, circulating miR-205, circulating miR-450a, circulating miR-16, circulating miR-34a, circulating miR-20a, circulating miR-373, circulating miR-519c, circulating miR-452, and circulating miR-193 b.

27. The use according to any one of claims 18 to 26, wherein, in the detection of circulating nucleic acids in the serum by the microarray, the serum is not extracted, but is directly heated and then reverse-transcribed to obtain cDNA for detection.

28. The use of any one of claims 18-26, wherein the microarray further comprises probes for detecting at least one of miR-107, miR-103a, and miR-132.

29. The use of claim 27, wherein the microarray further comprises probes for detecting at least one of miR-107, miR-103a and miR-132.

30. The use of claim 28, wherein the microarray comprises probes for detecting miR-103a and miR-132.

31. The use of claim 29, wherein the microarray comprises probes for detecting miR-103a and miR-132.