Page 100 - MI-2-2
P. 100

Microbes & Immunity                                               Big data and DNN-based DTI model in CHP




            Table 1. The regulation ability, toxicity, and sensitivity information of candidate drugs
            Biomarker                               Candidate drug       Regulation    Sensitivity   Toxicity
                                                                        ability (L1000)  (PRISM)   (LC50, mL/kg)
            AKT Serine/Threonine Kinase 1 (up-regulation)  Azathioprine   0.108309      −0.86865      4.558
                                                    Clobenpropit          0.264513      0.349028      4.223
                                                    Primaquine            0.64545       −0.34658      3.236
            Chemokine (C-C motif) ligand 20 (down-regulation)  Zosuquidar  −0.55168     −0.02332      7.732
                                                    Masitinib             −0.81919      0.034718      5.465
                                                    Mycophenolate-mofetil  −0.61481     −0.70094       4.33
            Cell division cycle 23 (down-regulation)  Azathioprine        −0.92617      −0.86865      4.558
                                                    Masitinib             −0.20676      0.034718      5.465
                                                    Elesclomol            −0.5743       −2.86051      4.321
            Chemokine (C-X-C motif) ligand 1 (down-regulation)  Azathioprine  −0.9438   −0.86865      4.558
                                                    Masitinib             −0.21064      0.034718      5.465
                                                    Primaquine            −0.14379      −0.34658      3.236
            Nuclear factor kappa B subunit 1 (up-regulation)  Imatinib    0.224935      −0.80213      5.419
                                                    Masitinib             0.21467       0.034718      5.465
                                                    Primaquine            0.310101      −0.34658      3.236
            Tumor necrosis factor (down-regulation)  Azathioprine         −0.54562      −0.86865      4.558
                                                    Masitinib             −0.19718      0.034718      5.465
                                                    Mycophenolate-mofetil  −0.59299     −0.70094       4.33
            Abbreviation: LC : Lethal concentration 50.
                       50
            Table 2. A potential multi‑molecular drug for chronic
            hypersensitivity pneumonitis

            Target    AKT1  CCL20  CDC23  CXCL1  NFKB1  TNF
            Drug       ()   ()   ()    ()    ()  ()
            Azathioprine  ✓         ✓      ✓           ✓
            Masitinib         ✓     ✓      ✓     ✓     ✓
            Primaquine  ✓                  ✓     ✓
            Drug           Sensitivity        Toxicity
            Azathioprine    −0.86865           4.558
            Masitinib       0.034718           5.465
            Primaquine      −0.34658           3.236
            Note: ✓ denotes the existence of interaction between the molecular
            drugs and biomarkers, ↑ denotes up-regulation, and ↓ denotes
            down-regulation.

            analysis revealed aberrant activation of the three major
            MAPK pathways – extracellular signal-regulated kinase   Figure  8. The receiver operating characteristic curve could measure
                                                               the probability of the prediction accuracy of the deep neural network-
            (ERK), p38, and Jun N-terminal kinase (JNK) – in lung tissue   based drug-target interaction model, with an area under the curve of the
            specimens from CHP patients.  The p38 MAPK pathway   receiver operating characteristic score of 0.981.
                                    28
            plays a pivotal role in driving the inflammatory milieu. 29,30    Abbreviation: AUC: Area under the curve; DNN: Deep neural network;
            Phosphorylation of p38α and p38β isoforms leads to the   DTI: Drug-target interaction.
            activation of key transcription factors, such as NF-κB, which
            subsequently induces the transcription of pro-inflammatory   beta (TGF-β)-induced myofibroblast differentiation and
            cytokines (e.g., TNF-α) and chemokines.  In addition, p38   extracellular matrix production, contributing to the fibrotic
                                            30
            MAPK signaling mediates transforming growth factor-  process. 31

            Volume 2 Issue 2 (2025)                         92                               doi: 10.36922/mi.4620
   95   96   97   98   99   100   101   102   103   104   105