In vitro method for the diagnosis of a mycoplasma pneumoniae infection and / or for the differential diagnosis between a mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias

Abstract

The present invention refers to an in vitro method for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between a Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias and / or for assessing whether to treat a patient suffering from pneumonia with a macrolide antibiotic.

Description

[0001] IN VITRO METHOD FOR THE DIAGNOSIS OF A MYCOPLASMA PNEUMONIAE INFECTION AND / OR FOR THE DIFFERENTIAL DIAGNOSIS BETWEEN A MYCOPLASMA PNEUMONIAE PNEUMONIA AND OTHER BACTERIAL OR VIRAL PNEUMONIAS

[0002] FIELD OF THE INVENTION

[0003] The present invention refers to the medical field. Particularly, it refers to an in vitro method for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between a Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias. The present invention also refers to the possibility of selecting the antibiotic treatment for a patient suffering from pneumonia. This comprises determining whether the patient is suffering from Mycoplasma pneumoniae pneumonia by following the method of the invention, wherein, if the patient is suffering from Mycoplasma pneumoniae pneumonia a treatment with a macrolide antibiotic may be recommended and a treatment with beta-lactam antibiotic might be initially discarded, and / or wherein if the patient is not suffering from Mycoplasma pneumoniae pneumonia other clinical decisions could be considered.

[0004] STATE OF THE ART

[0005] Mycoplasma pneumoniae commonly causes upper respiratory infections in infants and young adults, presenting with clinical manifestations ranging from asymptomatic to pneumonia. M. pneumoniae is one of the most common causes of atypical pneumonia. The incidence of M. pneumoniae pneumonia is uncertain and is typically described as endemic, marked by cyclic epidemics occurring every 3 to 5 years. The COVID-19 pandemic, along with associated nonpharmaceutical interventions, had a significant impact on the circulation of respiratory pathogens, including M. pneumoniae. A global surveillance network noted a decrease in M. pneumoniae detection from April 2020 to March 2021. Despite a resurgence in other pathogens, M. pneumoniae infections remained low until March 2022-23. Concerns about a potential upsurge in mycoplasma infections have arisen due to waning herd immunity. Recent outbreaks of M. pneumoniae infections have been reported during the last months in different European countries, including Spain, Denmark, France and the Netherlands. In November 2023, the World Health Organization (WHO) reported an increase in outpatient consultations and hospitalizations for pneumonia in China since May, along with rises in RSV, adenovirus, and influenza virus M. pneumoniae cases since October. Establishing the microbiological ethology of pneumonia remains a challenge for clinicians within healthcare facilities, primarily due to the complexity of obtaining direct microbiological samples that require invasive methods. Additionally, there is a delay in obtaining results from other indirect tests such as pneumococcal antigen in urine, PCR in respiratory swabs, or serology if available. In the case oiM. pneumoniae, cultures are usually the gold-standard for bacterial diagnosis; however, due to the difficulty of obtaining a representative sample, the slow growth, and the low sensitivity, they are not useful for clinical and treatment decisions. Serology for M. pneumoniae is another common approach, but tests show low specificity, and IgM can remain positive for months after the infection, with frequent cross-reactivity with other pathogens. This has led to its gradual replacement by PCR. Specific diagnosis of M. pneumoniae is important since, due to its lack of cellular wall, it does not respond to betalactams such as amoxicillin, which is the first-line treatment for typical community acquired pneumonia CAP. Azithromycin is the first line of treatment instead. Thus, clinicians usually start with empirical antibiotic treatment and adjust it later based on microbiological test results.

[0006] So, there is an unmet medical need of finding reliable tools aimed at diagnosing Mycoplasma pneumoniae infection, giving the possibility of differentiating it from other bacterial or viral pneumonias, and of assessing whether a patient suffering from pneumonia should be treated with a beta-lactam antibiotic, discarding a treatment with beta-lactam antibiotic when it is confirmed that the patient has a Mycoplasma pneumoniae pneumonia.

[0007] The present invention is focused on solving this problem by exploring host gene expression mechanisms specifically triggered by M. pneumoniae and evaluating them in the context of pneumonia molecular patterns generated by other pathogens, thus giving rise to a host- transcriptome signature specific for M. pneumoniae infection that may be used to overcome the limitations of current clinical and non-clinical procedures.

[0008] DESCRIPTION OF THE INVENTION

[0009] Brief description of the invention

[0010] The present invention refers to an in vitro method for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between ^Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias and / or for selecting the antibiotic treatment for a patient suffering from pneumonia. Particularly, the inventors of the present invention generated predictive signatures specific for M. pneumoniae using a lasso regression simulation approach and blood microarray data from 107 pneumonia children (including 30 M. pneumoniae). 8 different signatures, ranging from 3 to 10 transcripts, were identified (see Table 1) that differentiate mycoplasma pneumonia from other bacterial and viral pneumonias with high accuracy (AUC: 0.84-0.95).

[0011] Table 1. Model coefficients and weights (%) for the genes included in each of the signatures; with the genes (DIO3, GALM, and TXNDC11) appearing in at least five transcript signatures bolded, n: number of transcripts in the optimal signature. Additionally, the present invention demonstrates that existing signatures for broadly distinguishing viral / bacterial infections and viral / bacterial pneumonias were ineffective in distinguishing pneumonia caused by mycoplasma. The new mycoplasma signatures were successfully validated in an independent cohort of children with pneumonia, demonstrating their robustness. The present invention indicates that AT. pneumoniae infection induces specific transcriptomic alterations, enabling the development of robust diagnostic signatures. These signatures offer promising potential for enhancing the diagnosis and management of M. pneumoniae pneumonia, particularly with the possibility of integration into point-of-care diagnostic tools, thereby improving treatment outcomes.

[0012] On the other hand, kindly note that the eight signatures of Table 1 were considered as proof- of-concept and that the present invention offers reliable data showing that the individual use of any of the following genes PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, GALM, DNAI7, CCL23, ELL2, UAP1, CDCA2, PTGDR2, CADM1, GPR107 and / or MAPRE3, or any combination thereof comprising between 2 and 18 genes, can be used in the context of the present invention for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between a Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias and / or for selecting the antibiotic treatment for a patient suffering from pneumonia. In this regard, please refer to Table 2 below wherein the AUC value of each of the genes PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, DNAI7, CCL23, ELL2, UAP1, CDCA2, GALM, PTGDR2, CADM1, GPR107 and / or MAPRE3 ; and several combinations between 2 and 10 genes are provided as proof-of-concept.

[0013] The “special technical feature” conferring unity of invention is precisely the use of host gene expression mechanisms specifically triggered by pneumoniae, particularly for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between a Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias and / or for selecting the antibiotic treatment for a patient suffering from pneumonia. Since these genes have not been used before for this specific purpose, the use of alternative genes fulfils the requirements of unity of invention.

[0014]

[0015] So, the first embodiment of present invention refers to an in vitro method for selecting or identifying biomarkers signatures for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between a Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias which comprises: a) Assessing the level of expression of at least one gene, or any combination thereof comprising between two and eighteen genes, selected from the list consisting of PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, GALM, DNAI7, CCL23, ELL2, UAP1, CDCA2, CADM1, PTGDR2, GPR107 and / or MAPRE3 in a biological sample obtained from the subject, wherein the identification of a statistically significant variation of the level of expression with respect a pre-established threshold level determined in subjects with viral pneumonia, is an indication that the biomarkers signatures may be used for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between a Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias. The second embodiment refers to an in vitro method for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between ^Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias, the method comprising determining the level of expression of at least one gene, or any combination thereof comprising between two and eighteen genes, selected from the list consisting of PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, GALM, DNAI7, CCL23, ELL2, UAP1, CDCA2, CADM1, PTGDR2, GPR107 and / or MAPRE3, in a biological sample obtained from the subject.

[0016] The third embodiment of the present invention refers to an in vitro method for selecting the antibiotic treatment for a patient suffering from pneumonia which comprises determining whether the patient is suffering from Mycoplasma pneumoniae pneumonia by following the method of the invention, wherein, if the patient is suffering from Mycoplasma pneumoniae pneumonia a treatment with a macrolide antibiotic may be recommended and a treatment with beta-lactam antibiotic can be initially discarded, and / or wherein if the patient is not suffering from Mycoplasma pneumoniae pneumonia other clinical decisions could be considered.

[0017] In a preferred embodiment, the beta-lactam antibiotic is selected, for instance, from: Amoxicillin, ampicillin, or cefotaxime and / or the macrolide antibiotic is selected, for instance, from: Azithromycin, clarithromycin or erythromycin.

[0018] The fourth embodiment of the present invention refers to the in vitro use of at least one gene, or any combination thereof comprising between two and eighteen genes, selected from the list consisting of: PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, GALM, DNAI7, CCL23, ELL2, UAP1, CDCA2, PTGDR2, CADM1, GPR107 and / or MAPRE3; or of a kit comprising reagents for the determining the level of expression of the genes; for the diagnosis of ^Mycoplasma pneumoniae infection and / or for the differential diagnosis between a Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias and / or for selecting the antibiotic treatment for a patient suffering from pneumonia.

[0019] In a preferred embodiment, the method comprises determining the level of expression of at least a combination of genes selected from the list shown in Table 1 or Table 2.

[0020] In a preferred embodiment, the method comprises determining the level of expression of at least one combination of genes selected from: DIO3, GALM and TXNDC11; FKBP11, TXNDC11 and DIO3; GAS6, HY0U1, CAV1 and DIO3; or PRRll, DNAI7, GALM, GINS4, TXNDC11 and CAV1.

[0021] In a preferred embodiment, the method comprises determining the level of expression of at least one combination of genes that comprises the genes DIO3, GALM, and TXNDC11 selected from: DNAI7, GALM, GINS4, TXNDC11 and DIO3; CCL23, GALM, TXNDC11, ELL2, DIO3, UAP1 and CDCA2; CCL23, PRR11, GAS6, GALM, TXNDC11, CAV1, DIO3 and CADM1; CCL23, PRR11, DNAI7, GALM, GINS4, TXNDC11, CAV1, DIO3 and PTGDR2; or CCL23, PRR11, DNAI7, GALM, GINS4, TXNDC11, MAPRE3, DI03, GPR107 and CADM1.

[0022] In a preferred embodiment, the identification of a higher level of expression with respect a pre- established threshold level determined in subjects with viral pneumonia, is an indication that the subject has ^.Mycoplasma pneumoniae infection and / or that the subject has ^Mycoplasma pneumoniae pneumonia rather than another bacterial or viral pneumonias and / or that a treatment with a macrolide antibiotic may be recommended and a treatment with beta-lactam antibiotic can be initially discarded.

[0023] In a preferred embodiment, the biological sample is selected from blood, plasma or serum.

[0024] In a preferred embodiment, the present invention is a computer-implemented invention, wherein a processing unit (hardware) and a software are configured to: a) Receive the expression level values of any of the above cited biomarkers or signatures, b) process the concentration level values received for finding substantial variations or deviations, and c) provide an output through a terminal display of the variation or deviation of the expression level, wherein the variation or deviation of the expression level indicates that the subject may be suffering from Mycoplasma pneumoniae pneumonia.

[0025] In addition to the embodiments described above, further embodiments of the invention are provided herein, which may be practiced independently or in combination with any other embodiment disclosed.

[0026] In one embodiment, the invention provides a method for treating a subject suffering from pneumonia, comprising: i) Determining the expression level of at least one gene, or any combination thereof comprising between two and eighteen genes selected from the group consisting of PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, GALM, DNAI7, CCL23, ELL2, UAP1, CDCA2, PTGDR2, CADM1, GPR107 and MAPRE3, in a biological sample obtained from the subject; and b) administering a macrolide antibiotic to the subject if the expression level is indicative of Mycoplasma pneumoniae infection; or administering a beta-lactam antibiotic if Mycoplasma pneumoniae infection is not indicated.

[0027] In certain embodiments, the macrolide antibiotic is selected from azithromycin, clarithromycin, or erythromycin, and the beta-lactam antibiotic is selected from amoxicillin, ampicillin, or cefotaxime. In another embodiment, the invention provides a method for dynamically adjusting antibiotic therapy in a subject suffering from pneumonia, comprising: i) Periodically determining the expression level of at least one gene, or any combination thereof comprising between two and eighteen genes selected from the group consisting of PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, GALM, DNAI7, CCL23, ELL2, UAP1, CDCA2, PTGDR2, CADM1, GPR107 and MAPRE3, in a biological sample obtained from the subject; and ii) modifying or maintaining the administered antibiotic regimen based on changes in gene expression over time, wherein persistence or increase in expression levels associated with Mycoplasma pneumoniae infection indicates continuation or escalation of macrolide antibiotic treatment.

[0028] Such dynamic monitoring allows for real-time optimization of the therapeutic regimen based on host response, thereby improving clinical outcomes.

[0029] In another embodiment, the invention provides a computer-implemented system for assisting healthcare professionals in selecting appropriate antibiotic therapy for a subject with pneumonia, the system comprising: i) A hardware processing unit configured to: a) receive gene expression data from a diagnostic assay, for at least one gene or gene combination thereof comprising between two and eighteen genes selected from the group consisting of PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, GALM, DNAI7, CCL23, ELL2, UAP1, CDCA2, PTGDR2, CADM1, GPR107 and MAPRE3; b) compare said data with reference threshold values indicative of Mycoplasma pneumoniae infection; and c) generate a treatment recommendation comprising administration of a macrolide antibiotic if Mycoplasma pneumoniae infection is indicated, or administration of a beta-lactam antibiotic if not indicated; and ii) a user interface configured to present the treatment recommendation to a healthcare provider.

[0030] In some embodiments, the processing unit may further comprise a trained machine learning algorithm configured to improve diagnostic accuracy based on accumulated data.

[0031] In yet another embodiment, the invention provides a diagnostic and therapeutic kit for use in pneumonia management, the kit comprising: i) reagents for determining expression levels of at least one gene or gene combination thereof comprising between two and eighteen genes selected from the group consisting of PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, GALM, DNAI7, CCL23, ELL2, UAP1, CDCA2, PTGDR2, CADM1, GPR107 and MAPRE3; and ii) written instructions for administering a macrolide antibiotic upon detection of gene expression levels indicative of Mycoplasma pneumoniae infection; optionally, pre-packaged macrolide and / or beta-lactam antibiotic formulations for administration based on test results.

[0032] The availability of the kit enables rapid diagnosis and immediate initiation of targeted therapy at the point of care.

[0033] In a further embodiment, the invention provides a method for improving clinical outcomes in patients suffering from pneumonia, comprising: i) Performing the diagnostic gene expression assay as described herein; ii) selecting the appropriate antibiotic treatment based on the obtained gene expression profile; ii) thereby reducing unnecessary use of beta-lactam antibiotics in cases of Mycoplasma pneumoniae infection, minimizing the risk of antibiotic resistance development, and improving therapeutic efficacy.

[0034] For the purposes of the present invention the following terms defined:

[0035] • The term "comprising" means including, but it is not limited to, whatever follows the word "comprising". Thus, use of the term "comprising" indicates that the listed elements are required or mandatory, but that other elements are optional and may or may not be present.

[0036] • The term "consisting of’ means including, and it is limited to, whatever follows the phrase “consisting of’. Thus, the phrase "consisting of’ indicates that the listed elements are required or mandatory, and that no other elements may be present.

[0037] • According to the present invention, a reference value can be a “pre-established threshold value” or a “cut-off value”. Typically, a "threshold value" or "cut-off value" can be determined experimentally, empirically, or theoretically. According to the present invention, the “pre-established threshold” value refers to a value previously determined in subjects with viral pneumonia. A “threshold value” can also be arbitrarily selected based upon the existing experimental and / or clinical conditions, as would be recognized by a person of ordinary skilled in the art. The “threshold value” has to be determined in order to obtain the optimal sensitivity and specificity according to the function of the test and the benefit / risk balance (clinical consequences of false positive and false negative). Typically, the optimal sensitivity and specificity (and so the “threshold value”) can be determined using a Receiver Operating Characteristic (ROC) curve based on experimental data. Description of the figures

[0038] Figure 1. Overall study design. Differential expression was investigated between M. pneumoniae and virus datasets from Wallihan et al. (2018). We employed a re-sampling approach to generate 1,000 transcript signatures of size n. These signatures were utilized to compute 1,000 AUC values on both the training (TA) and test (TE) sets, as well as on the nonmycoplasma bacteria and virus datasets from the EUCLIDS cohort. Next, Signature Scores (SS) were obtained for the different n-transcript signatures (from n=3 to 10) based on AUCTA, AUCTE, and AUCEU values, allowing to derive the best signature of size n; n-SSMAX. AUC values were then computed with the best n-transcript signatures for the comparison between M. pneumoniae vs. virus and vs. virus + other pathogens from Wallihan et al. For comparative purposes, we also assessed pneumoniae the performance of other available signatures in the literature. Finally, the best signatures were subsequently validated in the DIAMOND / PERFORM cohorts.

[0039] Figure 2. PCA of transcriptomic profiles of different pneumonia cohorts and differential expression analysis. A) PCA of transcriptome profiles from blood samples of viral, M. pneumoniae, co-infections or other bacteria pneumonia infections. The two first principal components (PCI and PC2) are shown. B) Volcano plot showing the DEGs between conditions: M. pneumoniae pneumonia vs. viral pneumonia. Genes with an adjusted P-value < 0.05 and a log2FC > |1| were colored in orange, genes with an adjusted P-value < 0.05 but a log2FC < |1| were colored in dark blue, genes with an adjusted P-value >0.05 but a log2FC > |1| were colored in light blue and non-significant genes were colored in grey. Only genes included in the final signatures were labeled in the graphic.

[0040] Figure 3. Transcriptomic signatures obtained from the discovery cohort. AUCs and ROC curves from the density plots (left panels) of the AUC values computed on the 999 training resamples and using the optimal n-transcript signature; red dashed vertical lines in the left panels indicate the median values. ROC curves and AUC values (central panels) for the total cohort of M. pneumoniae vs. viral pneumonia dataset (black line; AUCTO|M-V) and for the total cohort of M. pneumoniae vs. all non-mycoplasma pneumonias (violet line; AUCTO|M- A). Boxplots of the predicted values using each optimal model in the total cohort with Wilcoxon P-values (right panels). Red dashed line represents the optimal cutpoint. Abbreviations: A: all non-mycoplasma pneumonias; B: bacterial; M: AL pneumoniae,' TO: total sample; V: virus. Figure 4. Validation of the transcriptomic signatures. ROC curves with AUC values for the three tested validation subsets: M. pneumoniae vs. viral pneumonias (black line; AUCM-V), M. pneumoniae vs. all non-mycoplasma pneumonias (violet line; AUCM-A) and M. pneumoniae vs. non-mycoplasma bacterial pneumonias (red line; AUCM-B) (left panels). Boxplots of the predicted values using each model in the three validation subsets (right panels). Wilcoxon P-values are also displayed. The red dashed line represents the optimal cutpoint. Abbreviations as in legend of Figure 3.

[0041] Figure 5. Differential pathways analysis. A) Volcano plot showing the genes in DEPs between conditions: M. pneumoniae pneumonia vs. viral pneumonia. Pathways meeting the criteria of an adjusted P-value < 0.05 and a log2FC > |0.1| were colored in yellow, pathways with an adjusted P-value < 0.05 but a log2FC < |0. 11 were colored in dark blue, genes with an adjusted P-value >0.05 but a log2FC > |0.1| were colored in light blue and non-significant genes were colored in grey. B) Boxplots representing the pathway activity of the DEPs for the mycoplasma and viral groups in which genes from the transcriptomic signatures are involved (left panel). Bubble plot of the DEPs in which genes included in the transcriptomic signatures are involved (right panel). Numeric identifies indicate DEPs in which genes from the transcriptomic signatures are involved.

[0042] Detailed description of the invention

[0043] The present invention is illustrated by means of the Examples shown below without the intention of limiting the scope of protection.

[0044] Example 1. Material and methods

[0045] Example 1.1. Samples and study design

[0046] We investigated the microarray blood transcriptomic profiles of 122 pneumonia children and 20 healthy controls. These cases involved pneumonia caused by M. pneumoniae (n = 30), various viral infections (n = 77), pyogenic bacteria (n = 5) and different co-infections (n = 10). Samples from coinfected pneumonias caused by M. pneumoniae with other pathogens and samples from non-etiology pneumonias were disregarded for downstream analysis. Data were downloaded from the Gene Expression Onmibus (GEO) database with accession number GSE103119.

[0047] Additionally, we have used RNAseq data from blood of other pediatric patients with pneumonia not caused by M. pneumoniae recruited in the European Union Childhood Life- threatening Infectious Diseases Study (EUCLIDS). The EUCLIDS cohort contains RNAseq data for 39 definitive non-my coplasma bacterial and 9 definitive viral pneumonia infections. For validation purposes, we have generated new RNAseq data from an additional pediatric cohort comprising blood samples obtained from children infected with pneumoniae (n = 9), along with samples from viral (n = 10) and bacterial (n = 22) pneumonias. These samples were recruited under the umbrella of the PErsonalised Risk assessment in Febrile illness to Optimise Real-life Management across the European Union (PERFORM https: / / www.perfomi2020.org / ) and the Diagnosis and Management of Febrile Illness using RNA Personalised Molecular Signature Diagnosis (DIAMONDS https: / / www.diamonds2020.eu) consortiums.

[0048] Example 1.2. RNAseq analysis

[0049] Whole blood was collected into PAXgene blood RNA tubes (PreAnalytiX), and stored at - 80 °C. Total RNA was isolated using PAXgene blood miRNA isolation kit according to the manufacturer’s instructions (Qiagen). An additional DNAse treatment was carried out with the RNA clean & concentrator kit (Zymo Research) prior to sequencing. RNA was quantified using RiboGreen (Invitrogen) on the FLUOstar OPTIMA plate reader (BMG Lab tech) and the integrity analyzed on the TapeStation 2200(Agilent, RNA ScreenTape). After a normalization step, a strand specific library preparation was completed using NEBNext® Ultra™ II mRNA kit (NEB) and NEB rRNA / globin depletion probes following manufacturer’s recommendations. Individual libraries were normalized using Qubit, pooled together and diluted. The sequencing was performed using a 150 paired-end configuration in aNovaseq6000 platform. Quality control of raw data was carried out using FastQC, alignment and read counting were performed using STAR, alignment filtering was done with SAMtools and read counting was carried out using FeatureCounts.

[0050] Example 1.3. Statistical analysis

[0051] Microarray data pre-process and normalization was performed using the HluminaHumanv4.db and Umma packages. RNAseq data was processed for batch correction using control samples and COMBAT-Seq package. Data was subsequently normalized with DESeq2 package. Principal component analysis (PCA) was conducted to explore the different groups in the data and check for potential outliers. A differential expression (DE) analyses was carried out using Umma package and accounting for differences in age and sex, in order to compare the blood transcriptome of pneumonia patients with M. pneumoniae infection (n = 30) v.s. viral pneumonia infections (n = 77).

[0052] In order to identify subsets of genes that could serve as predictive transcriptomic signatures differentiating M. pneumonia from other pneumonias of viral etiology, we randomly split the dataset (n = 107) into 1,000 independent subsets comprising 70% of the samples (training datasets; TA) and other 1,000 subsets containing the remaining 30% of the samples (test datasets; TE); step 2 in Figure 1. A predictive transcriptomic signature was computed using the R package glmnet for each of the 1,000 TA. To do that, a logistic LASSO regression model was fitted with the alpha parameter set to 1 and a 10-fold cross validation (step 3 in Figure 1); 276 differentially expressed genes (DEGs) were included as input for the logistic regression based on |Log2FC| > 1, adjusted / J- value < 0.01, and a Log2 average expression > 2.

[0053] The accuracy of the predictive transcriptomic signatures was measured by calculating the area under the receiver operating characteristic curve (AUC) with 95% confidence intervals (CI) using the pROC package. The optimal cut-point value (cut-off) that maximize sensitivity and specificity, was calculated using the OptimalCutPoints R package.

[0054] AUC values were computed for the 1,000 TA and the corresponding TE datasets (step 4 in Figure 1), and also for the non-A7. pneumoniae pneumonia infections (EUCLIDS cohort); step 5 in Figure 1. Among the 1,000 signatures of size n transcripts, we selected those (z) of size n = 3 to 10 transcripts. Next, on these selected z signatures, we computed the score zz-SS;= (0.4 x AUCTA) + (0.4 x AUCTE) - (0.2 x AUCEU) (step 6 in Figure 1); this score considers the AUC value on training and test samples but also simultaneously penalize the performance of the signature in other non-A7. pneumoniae pneumonia infections (EUCLIDS cohort). We selected the best transcriptomic signature of size n = 3 to 10 ( Z-SSMAX) among the different signatures with the same number of transcripts. The AUC values for the signatures with the Z-SSMAX values, were then obtained in the 999 complete training resamples (1,000 minus the one used to generate the Z-SSMAX) (step 7 in Figure 1). In addition, AUC values were also calculated for the full pneumonia datasets of z) mycoplasmas vs. virus, and zz) mycoplasmas vs. virus+others (step 8 in Figure 1).

[0055] The performance of other signatures to differentiate M. pneumoniae from viral pneumonias were also investigated. In particular, we tested the 5-transcript signature recently developed to differentiate non-Mycoplasma pneumonia from viral pneumonia, and the 2-transcript signature developed to differentiate viral from bacterial infections; step 9 in Figure 1. Example 1.4. Validation of the best signatures

[0056] The performance and accuracy of the predictive transcriptomic signatures were validated using new blood RNAseq data generated from an additional pediatric cohort (PERFORMDIAMONDS cohort); step 10 in Figure 1. Coefficients and intercepts from each LASSO model were applied to the new data to perform ROC analysis and calculate the AUC, sensitivity, and specificity of the signatures. Three data comparisons were tested: mycoplasma vs. viral pneumonia, mycoplasma vs. bacterial pneumonia, and mycoplasma vs. all other pneumonias including both viral and non-mycoplasma bacterial.

[0057] Example 1.5. GSVA pathway analysis

[0058] Biological pathways differentially involved in viral and M. pneumoniae pneumonia were inferred from gene expression data using the GSVA algorithm included in the Gene Set Variation Analysis (GSVA) R package. Gene Ontology (GO) biological pathways gene set collection from the Molecular Signatures Database (MSigDB) was used as reference database. Significantly differentially expressed pathways (DEPs) were determined using the Umma package with the viral groups as a reference and a threshold of adjusted / J- value < 0.05. All graphics were created using R software v.4.3.2 (www-r-proj ect.org).

[0059] Example 2. Results

[0060] Example 2.1. Transcriptomic analysis

[0061] We conducted a PCA on a subset of the 500 most variable genes with the entire pneumonia cohort; which comprised four different groups: bacterial infection, viral infection, M. pneumoniae infection and co-infections. The first principal component (PCI), explaining 16.85 % of the variation, clearly shows the segregation of the M. pneumoniae samples from pneumonias caused by other bacteria. As expected, co-infected samples are distributed evenly between the two groups. Viral profiles intermingle with other profiles in the plot, including AL pneumoniae and bacterial transcriptomes; Figure 2A. PC2 (accounting for 8.16% of the variation) makes a subtle distinction between bacterial infections on one pole of the component, with most co-infections included in this cluster. M. pneumoniae samples are slightly displayed towards the center of the component, while viral samples are evenly distributed along the whole component, reproducing the same behavior as in PCI (Figure 2A). In a comparative analysis of transcriptomes in children with pneumonia caused by M. pneumoniae vs. viral pneumonia, we identified 3,783 DEGs, using a significance threshold of False Discovery Rate (FDR) 5%. Among these DEGs, 2,288 were found to be upregulated, while 1,495 were downregulated (Figure 2B).

[0062] Example 2.2. Diagnostic signature discovery

[0063] To identifying the best minimal signatures for differentiating M. pneumoniae infection from viral infections, we constructed a LASSO model using the top 276 DEGs, applying the criteria of / value < 0.01, Log2FC > 111, and Log2 Average Expression > 2. Subsequently, we generated 1,000 different transcriptomic signatures using the 1,000 randomized TA. We then calculated the AUC for each signature in the TA, TE and the EUCLIDS cohort (non-mycoplasma bacterial and viral pneumonias). Among the 1,000 signatures, we computed the signature scores (SS) to obtain the best signatures of size 3 to 10. Eight transcriptomic signatures emerged as the best candidates for distinguishing pneumonia caused by M. pneumoniae from viral pneumonia (Table 1). These signatures include genes in common, totaling 18 different transcripts, with 15 over-expressed and 3 under-expressed in mycoplasma pneumonias compared to pneumonias from viral etiology. Notably, all these signatures demonstrated reliable performance when employed to compare AL pneumoniae samples against all non-mycoplasma samples, including cases of co-infections and other bacterial pneumonias.

[0064] We assessed the performance for each of the eight selected signatures across all randomized datasets (TA), excluding those in which each of the signatures were generated (therefore, 999 subsets). The median AUC values ranged from 0.84 (for the 3-transcript signature) to 0.95 (for the 10-transcript signature) (Figure 3), demonstrating the overall high-accuracy of the signatures to discriminate between both phenotypes. Additionally, we evaluated each transcriptomic signature in the complete dataset to test their performance in differentiating mycoplasma pneumonias from all other pneumonias, represented with a ROC curve and their respective AUC values. We set the optimal cut-off for each signature to optimize the discrimination of the two categories (Figure 3). The AUC values from this analysis aligned with the AUC median values observed across all TA datasets (Figure 3).

[0065] Example 2.3. Validation of M. pneumoniae signatures in an independent cohort

[0066] The diagnostic accuracy of the proposed RNA signatures was evaluated using additional gene expression data generated from a new paediatric cohort of pneumonia samples and a different technology (RNAseq). The results confirmed that all RNA signatures can discriminate between viral and mycoplasma pneumonias with AUCs higher than 0.68 (7-transcript), being the most predictive the 9-transcript signature (AUC = 1 [CI: 1-1]; sensitivity: 1; specificity: 1); Figure 4. Unexpectedly, the third most predictive model was the 3-transcript signature (AUC = 0.86 [CI: 0.65-1]; sensitivity: 0.89; specificity: 0.9) (Figure 4). Remarkably, all signatures can also accurately differentiate mycoplasma pneumonia from other bacterial pneumonias, with AUCs ranging from 0.74 (CI: 0.55-0.92) for the 4-transcript signature to 0.85 (CI: 0.73-0.99) for the 10-transcript signature. Considering all viral and non-my coplasma bacterial pneumonias against mycoplasma pneumonias, the AUC values were similar to those obtained from the comparison mycoplasma vs. bacterial comparison. In this case, the 9-transcript signature yielded the best performance of all the signatures tested (AUC = 0.89 [CI:0.79-0.99]); Figure 4

[0067] As in the case of the discovery dataset, we determined the optimal cut-off for each signature in the validation that establishes the threshold to discriminate between pneumonias from different aetiologies in the three comparisons groups (Figure 4).

[0068] Example 2.4. Performance of other available signatures in M. pneumoniae samples

[0069] As M. pneumoniae pediatric pneumonias appear to induce a different alteration in the transcriptome compared to other bacterial infections causing pneumonia in children, we opted to examine the performance of two available signatures designed to differentiate viral and bacterial pediatric infections in samples from children with pneumonia caused by M. pneumoniae', the two-transcript signature (IFI44L!FAM89A capable of differentiating between pediatric viral from bacterial infections; and the novel 5-transcript signature, designed to specifically distinguish between viral and bacterial CAP in children. We observed that neither of these two tested signatures could effectively differentiate M. pneumoniae as a bacterial infection or a bacterial pneumonia, yielding AUC values of 0.56 and 0.52, respectively.

[0070] Example 2.5. Differentially expressed pathway analysis between M. pneumoniae and viral pneumonia

[0071] To detect biological pathways responsible for the different response to M. pneumoniae and viral pneumonia in our study cohort, we performed a GSVA analysis directly from gene expression data. We identify 525 significantly (adjusted / J- value < 0.05) DEPs between both categories, with over half upregulated in AT. pneumoniae pneumonia (358 / 525; 68%); Figure 5A. Among the top 20 pathways, the most notable changes in pathways activity were mainly represented by up-regulated processes in atypical pneumonia (18 / 20; 90%).

[0072] The most differentially activated pathway (DEP) was “SRP dependent co-translational protein targeting to membrane signal sequence recognition” (adjusted E- value = 1 x 1 O’06) followed by ’’interleukin- 12 mediated signaling” pathways (adjusted -value= 2xlO'08), and a group of biological routes related to polysaccharides and glycolipids synthesis, including “nucleotide sugar biosynthetic process” (adjusted E- value = 9x1 O'05), “UDP N-Acetylglucosamine metabolic process” (adjusted E- value = 9x1 O'05), “amino sugar biosynthetic process” (adjusted E- value = 2x1 O'05), “UDP N-Acetylglucosamine biosynthetic process” (adjusted E- value = 4xlO'05) and “GDP -mannose metabolic process” (adjusted E- value = 9xlO'05).

[0073] Afterwards, we investigated the functional involvement of the genes included in the transcriptomic signatures by examining the DEPs related to these genes. We found that 10 out of the 18 genes (UAP1, PTGDR2, CAV1, HYOU1, GAIAP GAS6, GINS4, CDCA2, MAPRE3, DIO3, Figure 5B) participate in 9 significantly DEPs. Notably, among these pathways, “the nucleotide sugar biosynthetic process” emerged as one of the most significant pathways in the overall analysis (adjusted E- value = 9x1 O'05), Figure 5B. Additionally, the involvement of two of the predictive genes in the same DEP (CAV1 and HYOUP) was only detected for the “response to endoplasmic reticulum stress process” (adjusted E- value = 0.002).

Claims

CLAIMS1. In vitro method for selecting or identifying biomarkers signatures for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between a Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias which comprises: a) Assessing the level of expression of at least one gene, or any combination thereof comprising between two and eighteen genes, selected from the list consisting of: PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, GALM, DNAI7, CCL23, ELL2, UAP1, CDCA2, CADM1, PTGDR2, GPR107 and / or MAPRE3 in a biological sample obtained from the subject, wherein the identification of a statistically significant variation of the level of expression with respect a pre- established threshold level determined in subjects with viral pneumonia, is an indication that the biomarkers signatures may be used for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between a Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias.

2. In vitro method for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between a Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias, the method comprising determining the level of expression of at least one gene, or any combination thereof comprising between two and eighteen genes, selected from the list consisting of: PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, GALM, DNAI7, CCL23, ELL2, UAP1, CDCA2, CADM1, PTGDR2, GPR107 and / or MAPRE3, in a biological sample obtained from the subject.

3. In vitro method for selecting the antibiotic treatment for a patient suffering from pneumonia which comprises determining whether the patient is suffering from Mycoplasma pneumoniae pneumonia by following the method of claim 2, wherein, if the patient is suffering from Mycoplasma pneumoniae pneumonia a treatment with a macrolide antibiotic may be recommended and a treatment with beta-lactam antibiotic can be initially discarded, and / or wherein if the patient is not suffering from Mycoplasma pneumoniae pneumonia other clinical decisions could be considered.

4. In vitro method, according to claim 3, wherein the beta-lactam antibiotic is amoxicillin, ampicillin or cefotaxime and / or the macrolide antibiotic is azithromycin, clarithromycin or erythromycin.

5. In vitro use of at least one gene, or any combination thereof comprising between two and eighteen genes, selected from the list consisting of: PRR11, FKBP11, TXNDC11, DIO3, GAS6, HY0U1, CAV1, GINS4, GALM, DNAI7, CCL23, ELL2, UAP1, CDCA2, CADM1, PTGDR2, GPR107 and / or MAPRE3; or of a kit comprising reagents for the determining the level of expression of the genes; for the diagnosis of a Mycoplasma pneumoniae infection and / or for the differential diagnosis between a Mycoplasma pneumoniae pneumonia and other bacterial or viral pneumonias and / or for selecting the antibiotic treatment for a patient suffering from pneumonia.

6. In vitro method, or in vitro use, according to any of the claims 2 to 5, the method comprising determining the level of expression of at least a combination of genes selected from the list shown in Table 1 or Table 2.

7. In vitro method, or in vitro use, according to any of the claims 2 to 6, which comprises determining the level of expression of at least one combination of genes selected from: DIO3, GALM and TXNDCll; FKBP11, TXNDC11 and DIO3; GAS6, HY0U1, CAV1 and DIO3; or PRR11, DNAI7, GALM, GINS4, TXNDC11 and CAVL8. In vitro method, or in vitro use, according to any of the claims 2 to 7, which comprises determining the level of expression of at least one combination of genes that comprises the genes DIO3, GALM, and TXNDC11 selected from: DNAI7, GALM, GINS4, TXNDC11 and DIO3; CCL23, GALM, TXNDC11, ELL2, DIO3, UAP1 and CDCA2; CCL23, PRR11, GAS6, GALM, TXNDC11, CAV1, DIO3 and CADM1; CCL23, PRR11, DNAI7, GALM, GINS4, TXNDC11, CAV1, DIO3 and PTGDR2; or CCL23, PRR11, DNAI7, GALM, GINS4, TXNDC11, MAPRE3, DIO3, GPR107 and CADM1.

9. In vitro method, or in vitro use, according to any of the claims 2 to 8, wherein the identification of a higher level of expression with respect a pre-established threshold level determined in subjects with viral pneumonia, is an indication that the subject has a Mycoplasma pneumoniae infection and / or that the subject has a Mycoplasma pneumoniae pneumonia rather than another bacterial or viral pneumonias and / or that a treatment with a macrolide antibiotic may be recommended and a treatment with betalactam antibiotic can be initially discarded.

10. In vitro method, or in vitro use, according to any of the claims 2 to 9, wherein the biological sample is selected from blood, plasma or serum.

Images