TWI531655B - Prognosis prediction for acute myeloid leukemia by a 3-microrna scoring system - Google Patents

Description

Using a micro-ribonucleotide scoring system to predict the prognosis of acute myeloid leukemia

The present invention relates to a scoring method for predicting the prognosis of patients with acute myeloid leukemia (AML) comprising three microribosomal nucleic acids (microRNAs).

MicroRNAs are small, non-coding RNA molecules derived from the role of their precursor RNAs by protein complexes including Dicer and Drosha. It can inhibit the occurrence of translation or degrade mRNA by binding to the 3' untranslated region of the target gene, and regulate the expression of the gene at the post-transcription level. The role of microRNAs in the mechanisms that contribute to the development of cancer is important and complex.

In AML, microRNAs are involved in the differentiation, proliferation, and survival of hematopoietic cells, and affect the effects of treatment response and patient prognosis. Different microRNAs have different gene expression profiles in different types of genetic variants of AML; in addition, AMLs with specific gene mutations also have many different microRNA signatures. The role of microRNA expression in prognosis is an important feature. More and more studies have shown that microRNAs are involved in both positive and negative regulation of genes and affect prognosis and AML leukemia production. A highly expressed single microRNA, such as miR-181a , is considered to be an independent good prognostic factor in cytogenetically normal AML. Conversely, high performance amounts of miR191 , miR-199a or miR-155 alone and low-performance miR-212 or miR-29 have been reported to be factors with poor prognosis in AML. We speculate that in the mechanism of specific physiological regulation, there are usually multiple microRNAs involved, and all of them may affect the chemotherapy response of AML patients. Therefore, the integration of these microRNAs may be more accurate than considering only a single microRNA. Effectively predict the prognosis of these patients. As a highly heterogeneous disease, AML needs to be better classified by risk in order to get the best treatment outcome for patients. Combining the performance of multiple relevant microRNAs during patient survival analysis and taking into account the specific gravity of microRNAs should provide more information on prognostic judgments.

In order to solve the above problems, it is an object of the present invention to provide a simple and easy to use prognostic method for AML patients. After repeated calculations of the data obtained by the patient, we obtained the following formula: Risk = 0.4908 [hsa-miR-9-5p performance] + 0.2243 [hsa-miR-155-5p performance]- 0.7187 [hsa-miR-203 performance] . The prognostic predictability of this formula can also be confirmed in other independent ethnic groups, from the Western American Cancer Genome Atlas (TCGA).

The method for scoring includes: (a) detecting the expression of microRNAs mir-9 , mir-155 and mir-203 in AML patients, wherein MAMMU6 gene is recommended as an endogenous control gene; (b) Three microRNAs scoring formulas are used to calculate the risk index for AML patients (detailed below); and (c) to determine the patient's prognostic risk category.

The scoring system is designed to pass the z-converted microRNA expression as input data, and the formula requires the population average of the three microRNAs in the population and the standard deviation of the population. In order to actually apply the scoring system to other clinical institutions or other hospitals that do not have group data, the present invention provides a calculation formula for the data set of the hospital: risk = 0.4908 (- Δ C _{t hsa-miR-9- 5p} + 15.71) / 3.60 + 0.2243 (- Δ C _{t hsa-miR-155-5p} + 6.94) / 1.45-0.7187 (- Δ C _{t hsa-miR-203} + 17.16) / 2.66 . The above-described system will microRNA △ C _t C _t of the endogenous control C _t by subtracting the obtained gene, the endogenous gene is preferably controlled to MAMMU6; 15.71 and 3.60 respectively, the average △ C _hsa-miR-9-5p of _t The value is the same as the standard deviation. The same calculation method can be used in mir-155 and mir-203 . For each newly diagnosed patient, the use of four sets of immediate polymerase chain reaction (detection of hsa-miR-9-5p , hsa-miR-155-5p , hsa-miR-203, and MAMMU6 ) is sufficient to assess the risk of prognosis. The index, the risk index obtained is compared with the median population risk index of 0.0031, and the prognostic risk is classified. If the risk index is less than or equal to 0.0031, it indicates that the patient has a good prognosis for survival; otherwise, the patient has a poor prognosis for survival.

In a preferred embodiment of the invention, the AML patient is a primary ( de novo ) AML patient.

Another object of the present invention is to provide a kit for detecting the amount of expression of microRNAs, wherein the kit comprises oligonucleic acids which can be used to detect microRNAs mir-9 , mir-155 and mir-203 .

In a preferred embodiment, the kit further comprises an oligonucleic acid useful for detecting endogenous control genes. In another preferred embodiment, the endogenous control gene is MAMMU6 .

Non-limiting and non-comprehensive embodiments will be described in conjunction with the drawings. The following drawings are only described in accordance with the embodiments of the present invention, and thus are not intended to limit the scope of the present invention. Wherein: FIG. 1 is a flow chart of analyzing by using microRNA in the embodiment of the present invention.

Figure 2(a) is a risk index distribution map of 138 patients in the NTUH detailed group; Figure 2(b) is a risk index distribution map of the TCGA group.

Figure 3(a) compares the overall survival of patients with a lower risk index and high risk index in the AML patient population of the NTUH group; Figure 3(b) shows a lower risk index for the AML patient population in the TCGA group. The overall survival rate of patients with high-risk index was compared; Figure 3(c) is the comparison of the overall survival rate of patients with lower risk index and high-risk index among AML patients with normal karyotype in NTUH; d) is an exponential map showing that the higher risk index will have a lower chance of complete remission (CR).

The disclosure of the present invention will be described in detail by the following examples and examples, and reference to the accompanying drawings.

However, it must be noted that the disclosure of the present invention is not limited to the embodiments in the specification, but may be implemented in other different ways. In the drawings, parts that are not relevant to the present invention have been omitted to improve the clarity of the drawings, and please refer to the disclosed drawings.

The use of the terms "comprises" or "includes" in the entire specification means that one or more other components are not excluded, except for the components, steps, operating instructions and/or elements described. Steps, operational instructions, and/or presence or additional elements. The word "about" or "approximately" is used to mean that there is a near or permissible range of error for avoiding the use of the precise or absolute value disclosed herein by an unknown third party. .

The inventors of the present invention provide a method based on the expression pattern of one or more microRNAs to predict clinical outcomes (eg, survival of prognosis) in AML patients.

The scoring method is as follows: recruiting a group of AML patients, measuring the amount of expression of multiple microRNAs in bone marrow cells according to conventional techniques, such as real-time PCR or microarray analysis. The amount of expression of each microRNA was corrected by comparison with the amount of expression of endogenous control genes such as MAMMU6, and the amount of corrected microRNAs was obtained in the same patient.

[Examples]

Hereinafter, the disclosure of the present invention will be described by way of embodiments and drawings. However, the disclosure of the present invention is not limited to the embodiments and the drawings.

[Materials and Methods]

(a) patient

A total of 195 consecutive adult patients (> or = 15 years old) who were newly diagnosed with de novo AML at the National Taiwan University Hospital (NTUH) between 1995 and 2007. There are a large number of frozen bone marrow cells that can be used for microRNA analysis. Among them, patients with antecedent hematologic disease and treatment-related acute myeloid leukemia (therapy-related AML) have been excluded. This experiment was carried out in accordance with the Helsinki Declaration and approved by the NTUH Ethics Committee. Of these primary patients, 138 (70.7%) received standard intensive chemotherapy, and the remaining 57 received tranquil care or low-dose chemotherapy due to poor performance status or patient wishes. All 195 patients included a correlation analysis between specific microRNAs relative to the performance of other parameters, but only for 138 patients who received standard intensive chemotherapy. AML population data obtained from TCGA (Cancer Genome Atlas) was used, which contained a publicly available microRNA database as a validation population (Figure 1).

(b) Quantification of microRNA expression

Mononuclear cells were isolated from the bone marrow samples obtained from the diagnosis and then stored frozen. RNA was extracted using the TriZol method. 1 μg of RNA was tested using the TaqMan MicroRNA Reverse Transcriptome (Applied Biosystems), and the microRNA assay was performed using a TaqMan Array Human microRNA A card (Applied Biosystems) with a 7900HT Instant Polymerase Chain Reaction Reactor. The amplification curve was converted by the Applied Biosystems SDS2.3 software to a digital representation. MicroRNAs that could not be detected within 40 cycles (C _T >40) are labeled as "undetermined".

(c) Establishment of a risk scoring method

In order to establish a risk scoring method based on microRNA expression level, the relationship between overall survival (OS) and individual microRNAs performance levels was first analyzed. Here, the prognostic significance of the performance of each microRNA for survival is measured by a univariate Cox proportional hazards regression model. MicroRNAs with significant performance relative to survival (univariate Cox P < 0.005) were applied to the multivariate Cox model to look for microRNAs whose performance was independently used to predict survival. A method for establishing a risk score by multivariate testing with a significant significance (multivariate Cox P <0.1) relative to the total survival rate was selected, in which the constituent microRNAs were subjected to round and round multivariate Cox regression tests to obtain The beta value of the specific gravity coefficient. MicroRNAs that are more significant for the importance of prognosis have a higher specific gravity coefficient. The microRNA-based risk scoring method is defined as risk ( p )= Beta _i . miRNA _i ( p ), where P is the patient's identification number, Beta _{i is} defined as the specific gravity of the microRNA probe i , and miRNAi is the value obtained by logarithmic conversion of the amount of expression obtained by the microRNA probe i . The sum of Beta _i Σ miRNA(p) for all microRNAs is the estimated risk index for patient P.

In order to confirm the performance of the risk scoring method of the present invention, 10,000 random permutation tests were performed, and in each random cycle, the same number of microRNAs were randomly selected from the microRNA data set to establish a "random scoring method", and according to the above The content specifies the appropriate weight. Each random score used to assess the importance of prognosis was tested against a univariate Cox model. After 10,000 cycles, the proposed P- value obtained by the risk assessment system can be calculated as a small fractional randomization method that achieves a better multivariate Cox P value than the estimated risk system. The smaller the empirical P value obtained, the better the proposed risk scoring method is than the random combination of microRNAs.

(d) Statistical analysis

The overall survival rate is calculated by calculating the date of the first diagnosis to the date of death or the final traceability for any reason. To rule out any confounding factors that may be caused by allogeneic hematopoietic stem cell transplantation (HSCT), patients who underwent this procedure were censored on the day of cell entry. Survival curves were plotted using the Kaplan-Meier estimate and log rank tests were used to examine differences between different study populations. The correlation between the risk-based scoring methods based on microRNA performance and the clinical features was included in the overall patient population; however, only patients receiving standard intensive chemotherapy were analyzed for survival. All computational analyses were performed using the XLSTAT Computational Analysis Software version 2010.5.02 (Addinsoft, Deutschland, Germany).

[result]

(a) Establishment of a 3-microRNA risk assessment method

To define a predictive microRNA scoring method for AML patient risk, a correlation screen for individual microRNA performance versus overall survival was performed in the NTUH population (as a discovery database, n=138). Eleven microRNAs with a significant correlation with total survival ( P < 0.005) were defined by a univariate Cox proportional hazard regression model, including: hsa-miR-9-5p , hsa-miR-146a , hsa-miR-222 , hsa-miR-128 , hsa-miR-181a , hsa-miR-125b , hsa-miR-196b , hsa-miR-155-5p , hsa-miR-224 , hsa-miR-203 and hsa-miR-339- 3p (sorted according to the size of the Cox P value). In order to further investigate the strong and robust survival prediction by microRNA expression, we introduced the 11 microRNA expressions into the multivariate Cox model, and found hsa-miR-9-5p (accession number: MIMAT0000441) and hsa- miR-155-5p (MIMAT0000646) was independently associated with poor overall survival, while hsa-miR-203 (MIMAT0000264) showed a good overall survival trend with multivariate Cox P values of 0.005, 0.033 and 0.080. Focusing on these three microRNAs to establish a risk scoring method is as follows: risk = 0.4908 [hsa-miR-9-5p] + 0.2243 [hsa-miR-155-5p]-0.7187[hsa-miR-203] , in which microRNAs The specific gravity is the multivariate Cox analysis and z-conversion (ie, after subtracting the average, divided by the unit standard deviation), so that each microRNA has a zero mean and unit standard deviation. According to data from 138 patients in the NTUH study population, the risk index distribution was close to the normal distribution, ranging from -2.01 to 1.97, with median, mean, and standard deviation of 0.01, 2.87e-15, and 0.78, respectively (Figure 2 (Figure 2 ( a)). The same applies to the TCGA group (Fig. 2(b)). This risk index has good predictive effect in our database (univariate Cox P =1.41 x 10 ^-7 , Log-rank P =1.03 x 10 ^-5 ). The scoring method of the present invention is almost superior to the random selection of 10,000 results (experience P = 6 x 10 ^-4 ), indicating that the proposed system is highly efficient and non-accidental.

(b) Correlation between clinical and molecular feature scoring methods

The higher index is positively correlated with higher age, high white blood cells, platelets and buds, but the chromosomal changes with better prognosis are mutually exclusive (as shown in Table 1).

Gene mutation profiles are also different in the high and low index groups: patients with high indices usually have NPM1 , FLT3- ITD, and MLL- PTD mutations, and rarely have CEBPA mutations (Table 2).

(c) Survival analysis

The overall survival rate of AML patients with high index was significantly lower than those with low index (median 13.5 months vs. not reached, P < 0.0001, Figure 3 (a)). The prognostic importance score was validated by the TCGA AML population (median 12.2 vs 26.4 months, P = 0.008, Figure 3(b)), which is the only group that publicly available AML survival and microRNA data. . When we restricted the patients to the normal karyotype, the higher the index, the overall survival rate remained poor (median 17.0 months vs. did not reach, P = 0.006, Figure 3(c)). Patients with a higher index of patients with a lower index were more likely to achieve complete remission (CR) after induction chemotherapy ( P = 0.0001, Figure 3(d)).

(d) Multivariate analysis

Since the high index appears to be associated with changes in other poor prognosis (Tables 1 and 2), we sought to investigate whether the index obtained by this scoring method can be used as an independent factor. We included some known prognostic impact factors as a common variable, and the post-high index appeared to be a highly independent risk factor. It is worth noting that the independence of the scoring method of the present invention also has demonstrated confidence in the TCGA cohort (Table 3). By integrating the covariates for further analysis, we found that patients with a poor risk index had a shorter overall survival rate in both our patients and patients with TCGA.

(e) Ratio of scoring methods to single microRNA expression for prognosis More

To understand whether the scoring method is more potent than a single microRNA, we increased the amount of microRNA composition, including hsa-miR-9-5p , hsa-miR-155-5p , hsa-miR-203, and hsa. -miR-181a , the increase in the expression of this microRNA is highly correlated with the good prognosis of AML, and the multivariate analysis was performed in addition to the covariate analysis of these microRNA expressions, as shown in Table 3. The results show that the 3-microRNA characterization is superior to the performance of all individual microRNAs (Tables 4-7).

(f) Clinical application scoring method using real-time polymerase interlocking reaction microRNA detection

Since the scoring method is designed to use the z-transferred microRNA expression hierarchy as input data, the mean and population standard deviation of the three microRNAs is required for this formula. To this scoring method actually used in clinical evaluation and non-hospital groups of the present invention, will be provided below using the NTUH data is calculated using the following formula: Risk _{= 0.4908 (- △ Ct hsa-} miR-9-5p +15.71) / _{3.60 + 0.2243 (- △ Ct hsa} -miR-155-5p +6.94) /1.45-0.7187 (- △ Ct hsa-miR-203 +17.16) /2.66.

Here, △ Ct value minus the Ct of the endogenous microRNA Ct of the control gene MAMMU6; 15.71 and 3.60 respectively, the average and standard _△ Ct _hsa-miR-9-5p difference. The values of hsa-miR-155-5p and hsa-miR-203 are also the same. For each newly diagnosed patient, a 4-well real-time polymerase-interlocking microRNA assay (detecting hsa-miR-9-5p , hsa-miR-155-5p , hsa-miR-203, and MAMMU6 ) is sufficient for prognosis The score can be compared to the median of our ethnic risk index of 0.0031 for risk group classification.

In the present invention, there are advantages in integrating the clinical, genetic, and array data of the population of the present invention to obtain a simple but effective 3-microRNA signature to predict postoperative outcomes, considering only the performance and specific gravity of the three microRNAs. It is screened through a series of statistical calculations to ensure its strong and independent influence on prognosis. The efficacy of this index was confirmed by the TCGA ethnic group, an independent patient validation combination that was studied in different platforms via different microRNA quantification platforms. Thereby, the scoring method of the present invention is independent in both the patient population and the quantitative method. Although both NTUH and TCGA are non-predictable patient populations, 3-microRNA signatures have a very high impact on predicting prognostic outcomes, and therefore have a high degree of susceptibility to risk classification. It is worth noting that the integrated 3-microRNA scoring method of the present invention is superior to the microRNA hsa-miR-9-5p , hsa-miR-155-5p and hsa-miR-203 alone, and the microRNAs in the scoring method are composed. , and hsa-miR-181a, its increased performance is highly correlated with good prognostic factors for AML. In the scoring method of the present invention, the low index system is associated with a good prognostic factor, such as good risk classification and gene mutation, and the multivariate analysis of the NTUH group and the TCGA group confirms that three independent microRNA features are independent of other important factors. Prognostic parameters.

From a practical point of view, the mean and standard deviation are added to the scoring method so that the results of each new diagnosis of AML patients can be predicted via a simple qPCR-based experimental procedure, even if the laboratory does not It is also possible to have a population mean or standard deviation. All are commercially available materials and are fast and suitable for high throughput applications.

miR-155 , miR-9 and miR-203 are the three core components of the scoring method of the present invention. In hematopoietic cancer, miR-155 is known to be an oncogene and has a poor prognosis, but another study has the opposite conclusion. The signal transmission pathway involved in miR-155 is very complex, but it may have a large-scale impact on cell reproduction, cell cycle progression, cell infiltration/metastasis. Among the populations of the present invention, miR-155 is an independent poor prognostic factor consistent with previous studies. miR-9 has been shown to inhibit the tumorigenicity of cancer, but this molecule can also promote the metastasis of solid cancer. More complicated, it is mentioned in one report that an increase in the amount of expression is shown to be a good prognostic factor in medulloblastoma, but in another report it is a poor outcome in glioma. Prognostic molecule. For AML, compared to other types of AML, miR-9, especially in MLL-gene rearrangement AML, is a specific amount of microRNA that is a target of MLL fusion protein, and its performance is directly related to disease. The progress is related. In the present invention, miR-9 is a factor of poor prognosis. There are few studies on miR-203 , which acts as an inhibitor of skin stem cells by regulating the metastasis between proliferating basal precursor cells and terminally differentiated basal cells. However, its significance for the prognosis of AML is not yet Clear, but it can target ABL1 and inhibit BCR-ABL1 in some chronic myelogenous leukemia or some acute lymphocytic leukemia. Among other cancers, it is often used as a tumor suppressor, but the increased expression of miR-203 is associated with tumor proliferation and poor prognosis in ovarian cancer; in the present invention, miR-203 is a prognosis. Favorable factor.

In summary, the present invention proposes a simple and easy-to-use 3-microRNA feature for AML as a factor that can effectively predict the prognostic effect, which is through a plurality of statistical analysis of the cycle of the population data of the present invention, and further by Confirmed by other independent patient populations. This scoring method is superior to the prediction method that only performs multivariate analysis of a single microRNA. Pairwise analysis of microRNA-mRNA is presumed to be associated with this feature and generally associated with cancer-associated molecular regulation Related to the trail.

Although the examples have been disclosed in the examples herein, it is still to be understood that there are still many possible variations. The changes in the embodiments are not to be construed as a departure from the spirit and scope of the invention, and the scope of the invention is intended to be included within the scope of the invention.

Claims (8)

A method for predicting the prognosis of patients with acute myeloid leukemia (AML), including: (a) detecting the expression of microRNAs mir-9 , mir-155, and mir-203 in patients with acute myelogenous leukemia; (b) calculating the risk index of patients with acute myeloid leukemia according to three microRNAs scoring formula; and (c) determining the prognostic risk category of the patient; wherein the risk index is calculated as follows (A) or (B) Show: Formula (A): Risk = 0.4908 [hsa-miR-9-5p expression amount] + 0.2243 [hsa-miR-155-5p expression amount] -0.7187 [hsa-miR-203 expression amount] ; Formula (B) : Risk = 0.4908 (- △ C _{t hsa-miR-9-5p} + 15.71) / 3.60 + 0.2243 (- △ C _{t hsa-miR-155-5p} + 6.94) / 1.45-0.7187 (- △ C _{thsa-miR- 203} + 17.16) / 2.66 , wherein each ΔC _t value is obtained by subtracting the C _t value of the endogenous control gene from the C _t value of the microRNA. The scoring method of claim 1, wherein the risk index of the formula (A) is less than or equal to 0.01, represents that the patient has a good prognosis survival prospect. The scoring method of claim 1, wherein the endogenous control gene of the formula (B) is MAMMU6 . The scoring method of claim 3, wherein the risk index of the formula (B) is less than or equal to 0.0031, represents that the patient has a good prognosis survival prospect. The scoring method of claim 1, wherein the acute myeloid leukemia patient is a primary ( de novo ) patient. A kit for detecting the amount of microRNAs present, wherein the kit comprises a polymerase chain reaction reagent and an oligonucleic acid that can be used to detect microRNAs mir-9 , mir-155 and mir-203 , wherein the detection The resulting microRNAs mir-9 , mir-155 and mir-203 are expressed in a risk index as in the formula (A) or formula (B) of claim 1. The kit of claim 6 further comprising an oligonucleic acid useful for detecting an endogenous control gene. The kit of claim 7, wherein the endogenous control gene is MAMMU6 .