Pennisi et al. BMC Bioinformatics 2019, 20(Suppl 6):504 https://doi.org/10.1186/s12859-019-3045-5 RESEARCH Open Access Predicting the artificial immunity induced by RUTI® vaccine against tuberculosis using universal immune system simulator (UISS) Marzio Pennisi1†, Giulia Russo2†, Giuseppe Sgroi1, Angela Bonaccorso2, Giuseppe Alessandro Parasiliti Palumbo1, Epifanio Fichera3, Dipendra Kumar Mitra4, Kenneth B. Walker5, Pere-Joan Cardona6,7,8, Merce Amat6, Marco Viceconti9 and Francesco Pappalardo2* From 2nd International Workshop on Computational Methods for the Immune System Function Madrid, Spain. 3-6 December 2018 Abstract Background: Tuberculosis (TB) represents a worldwide cause of mortality (it infects one third of the world’s population) affecting mostly developing countries, including India, and recently also developed ones due to the increased mobility of the world population and the evolution of different new bacterial strains capable to provoke multi- drug resistance phenomena. Currently, antitubercular drugs are unable to eradicate subpopulations of Mycobacterium tuberculosis (MTB) bacilli and therapeutic vaccinations have been postulated to overcome some of the critical issues related to the increase of drug-resistant forms and the difficult clinical and public health management of tuberculosis patients. The Horizon 2020 EC funded project “In Silico Trial for Tuberculosis Vaccine Development” (STriTuVaD) to support the identification of new therapeutic interventions against tuberculosis through novel in silico modelling of human immune responses to disease and vaccines, thereby drastically reduce the cost of clinical trials in this critical sector of public healthcare. Results: We present the application of the Universal Immune System Simulator (UISS) computational modeling infrastructure as a disease model for TB. The model is capable to simulate the main features and dynamics of the immune system activities i.e., the artificial immunity induced by RUTI® vaccine, a polyantigenic liposomal therapeutic vaccine made of fragments of Mycobacterium tuberculosis cells (FCMtb). Based on the available data coming from phase II Clinical Trial in subjects with latent tuberculosis infection treated with RUTI® and isoniazid, we generated simulation scenarios through validated data in order to tune UISS accordingly to STriTuVaD objectives. The first case simulates the establishment of MTB latent chronic infection with some typical granuloma formation; the second scenario deals with a reactivation phase during latent chronic infection; the third represents the latent chronic disease infection scenario during RUTI® vaccine administration. (Continued on next page) * Correspondence: francesco.pappalardo@unict.it Marzio Pennisi and Giulia Russo. These authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors. 2 Department of Drug Sciences, University of Catania, Italy, 95125 Catania, Italy Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Pennisi et al. BMC Bioinformatics 2019, 20(Suppl 6):504 Page 2 of 10 (Continued from previous page) Conclusions: The application of this computational modeling strategy helpfully contributes to simulate those mechanisms involved in the early stages and in the progression of tuberculosis infection and to predict how specific therapeutical strategies will act in this scenario. In view of these results, UISS owns the capacity to open the door for a prompt integration of in silico methods within the pipeline of clinical trials, supporting and guiding the testing of treatments in patients affected by tuberculosis. Keywords: Computational modeling framework, In silico clinical trials, Tuberculosis, Vaccine, Immunity, Therapeutic strategies Background reduce the probability of transmission, along with other Tuberculosis (TB) is a transmissible airborne disease trig- public health measures [15]. The anti-tuberculosis gered by the Mycobacterium tuberculosis (MTB) acid-fast drugs can be classified as drugs with bactericidal and bacillus that most often affect the lungs [1, 2]. Even sterilizing effect. The first one is fundamental in the though TB has existed for millennia, it still represents a early phase giving a significant reduction of the bac- worldwide health problem and one of the major causes of terial load; the second one is more efficient in the morbidity and mortality in developing and developed continuation phase, because it is directed to kill countries. In 2016, 10.4 million of new cases and 1.6 mil- latent mycobacteria. The secondary consequence of this lion of TB-causing deaths were globally estimated; of activity is the decrease of the probability of selecting drug- these, 2.79 million were new cases in India and 435,000 resistant strains [16]. cases dead for TB [3]. For these reasons, India is the coun- The mainstay of TB antibiotics consists on a first-line try with the highest burden of TB and leads the world in therapy for active tuberculosis. The standard treatment deaths from tuberculosis in terms of absolute number of for people at low risk for drug-resistance is isoniazid cases [4]. TB is spread through the air from person to per- (INH), rifampicin (RMP), pyrazinamide (PZA) and eth- son, in particular when people with lung TB cough, sneeze ambutol (EMB) or streptomycin (SM) for the initial two- or spit. The lifetime risk of falling ill with TB for people month phase followed by isoniazid and rifampin for 4 previously infected with TB bacilli is about 5–15%. How- months to 7. The main properties of these anti-tubercu- ever, compromised immune systems subjects, such as losis drugs are the bactericidal and sterilizing activity people living with diabetes, HIV or malnutrition condi- along with a reduction in the development of drug re- tions have a much higher risk of falling ill with TB [5–9]. sistance [17, 18]. For an organism encountering M. tuberculosis bacilli Second line drugs are used for the treatment of drug there are different possible outcomes: firstly, the bacillus resistant TB patients (i.e., for single or multidrug-re- can be instantly killed by the host’s innate immune re- sistant tuberculosis (MDR-TB) as well as extensively sponse [10, 11]. Secondly, of every 10 people infected drug-resistant tuberculosis (XDR-TB)) or and the with M. tuberculosis, one may develop an active infec- World Health Organization (WHO) has divided them tion in their lifetime within a finite time frame, from 1 in 2, 3 and 4 group [19]. Aminoglycosides and poly- to 3 years. This category probably lacks the capability peptides belong to WHO group 2 (i.e., amikacin to both control the early infection and develop a pro- (AMK) and capreomycin); fluoroquinolones are attrib- tective response in time in order to prevent the dis- uted to WHO group 3 such as ciprofloxacin (CIP), ease. Cough with sputum and blood at times, fever, levofloxacin (LVX), moxifloxacin (MXF) while thioa- weakness, chest pains, weight loss and night sweats mides and cycloserine are representative drugs of are the major common symptoms of active pulmonary WHO group 4 for example with ethionamide [20, 21]. TB [12]. Finally, we observe latent tuberculosis infec- Currently treatment regimens vary in duration be- tion (LTBI) when a persistent immune response to tween 6 months in drug susceptible TB to 24 months stimulation by Mycobacterium tuberculosis antigens in drug resistant TB [22]. occurs without evidence of clinically manifested active Rifamycin-based regimens for LTBI have been success- tuberculosis. One-quarter of the global population is ful in preventing progression to TB disease [23]. However, infected with LTBI and individuals with LTBI repre- treatment of LTBI needs a long period of chemotherapy, sent a reservoir for active TB cases [13, 14]. approximately 9 months, leading to an extremely difficulty In order to manage the disease, antibiotic treatment re- in terms of management and treatment-compliance duces the bacterial load in the lungs and can be helpful to therapy [24]. Pennisi et al. BMC Bioinformatics 2019, 20(Suppl 6):504 Page 3 of 10 The lengthy TB chemotherapy regimen likely reflects eliciting the complete clearance of the infection or, even- the inability of current chemotherapy drugs to eradicate tually, allowing the chronic establishment of MTB reser- subpopulations of Mycobacterium bacilli, allowing these voir inside the host due to both MTB characteristics and population bacteria to persist in infected individuals, genetic features of the host). surviving in granulomas with a central necrotic core and Based on the available data coming from phase II Clin- an external layer of foamy macrophages (FM) as the ical Trial in subjects with latent tuberculosis infection main immunosuppressive barrier [25]. treated with RUTI® [29] and then with 1 month of iso- Moreover, the dramatic changes from an epidemio- niazid treatment, we analyzed three simulation scenarios logical point of view during the last two decades and the in order to tune UISS through previous validated data increase of drug-resistant forms of tuberculosis consider- and accordingly to STriTuVaD objectives. ably complicate the clinical and public health manage- The first scenario simulates the establishment of MTB ment of the patients and of their relative contacts [26]. latent chronic infection with some typical granuloma Therapeutic vaccinations have been suggested to overcome formation as a reservoir of MTB infection; the second these critical issues and several candidates are already in clin- scenario deals with a reactivation phase during latent ical phases of development [27]. In this regard, efforts to chronic infection; the third represents the latent chronic shorten treatment duration have been successful, as shown disease infection scenario during RUTI® vaccine adminis- in recent clinical trials, evaluating the role of several immu- tration in order to enhance antibiotic therapy. notherapeutic vaccines, such as RUTI® [28, 29]. The safety and immunogenicity of RUTI® vaccine have been already Implementation tested in Phase I and II studies in healthy volunteers and The simulation framework LTBI infected patients and it was observed its capacity to UISS has a long history of development. It branches from C- stimulate host immune effectors against bacilli improving IMMSIM [44] to specialize the simulation framework to deal outcomes under chemotherapy treatment [30]. with tumor immunology. Then it held different stages of im- Inside the “In Silico Trial for Tuberculosis Vaccine mune system features development and was applied to a Development” (STriTuVaD) project, the anti-TB vaccine wide disease modeling scenarios. Nowadays, a lot of compu- RUTI® is under evaluation in under evaluation in studies tational efforts are leading to an incredible boost in biomedi- reducing the time to negativity of Sputum Culture Con- cine research [45, 46]. Technically speaking, UISS is an agent version (SCC) [31] in TB drug sensitive patients and TB based model (ABM) [47]. It can be thought as a two-compo- multidrug resistance patients, favorably responding to nent software. The first one is represented by a core that standard TB treatment, with the aim to shorten treat- implements all the immune system features that is ment and reduce the frequency of recurrence. In this continuously updated to take into account all the state-of- context, the application of in silico computational mod- the-art advancements in immunology. It owns anatomical eling strategies in the pipeline of standard clinical trials compartments, cells and molecules, immune system reper- can helpfully increase the high likelihood chance of suc- toire, molecular affinity, haematopoiesis with generation of cess of promising vaccine through the simulation of each cells, cell maturation and thymus selection, Hayflick limit in individual of the reference population and the in silico the number of duplications [48], aging and memory of past predictions in terms of vaccine efficacy, recurrence and infections, hyper-mutation of antibodies [49], bystander ef- reactivation of the disease [32–34]. fect [50], cell activation and anergy, cell interaction and co- Within the STriTuVaD consortium, the Universal Im- operation, antigen digestion and presentation. From the mune System Simulator (UISS) platform (previously suc- point of the view of physical representation, UISS maps bio- cessfully applied to a large number of disease modeling logical tissues, lymphatic system and peripheral blood into a scenarios [35–42]) simulates the MTB infection dynam- mathematical structure called cartesian lattice (two or three ics and its interactions with the host immune system. dimensional). Thymus and bone marrow are taken into ac- Hence, the in silico trial platform will be delivered to the count in an implicit way. This means that positive and nega- STriTuVaD multidisciplinary consortium in order to tive selection (that are the main features of thymus) are simulate the relevant individual human physiology and implemented to educate T cells. A procedure that generates physiopathology in patients affected by Mycobacterium mature B cells, instead, models the bone morrow cells pro- tuberculosis and to predict how specific vaccinations duction. UISS playground is then only populated by im- strategies will be effective or not. munocompetent entities. An important working assumption Here, we present the current advances of UISS devel- that allows UISS to deal with the incredible and fascinating opment and its extension in the context of tuberculosis variability of the immune system repertoire (including its [43], focusing on specific scenarios that show the cap- capacity to adapt and consequently mount immune re- ability of UISS in simulating the intrinsic immune sys- sponse) is the stochastic bitstring polyclonality. This means tem behavior against MTB infection (eliciting or not that epitopes and paratopes are modeled using binary Pennisi et al. BMC Bioinformatics 2019, 20(Suppl 6):504 Page 4 of 10 calculus. Each of them is represented by a set of binary a virtual injection of MTB bacilli into a lattice point in the strings. So, having N bits allow a potential immune system lung compartment. Then the bacilli are free to move and dis- repertoire diversity of 2N. Stochasticity is implemented as seminate randomly. When an AM becomes infected, it se- probabilistic rules that produces a logical causal/effect se- cretes IL-1, TNF-α, IL-12, IL-6 and chemokines. Depending quence of events culminating in the immune response and on MTB strains and their virulence, infected AM can play a development of immunological memory. different role in determining the downstream pathways lead- The second part consists on the disease model. It en- ing to the induction of either apoptosis or necrosis and to capsulates all the characteristics that are needed to re- the final outcome of the infection. In this context, lipoxin A4 produce the biological scenario one would like to (LXA4) promotes necrosis, while prostaglandin E2 (PGE2) simulate. In this specific case, the disease model takes is a proapoptotic factor. When necrosis process is favored, into account the entities, the interactions and the dy- the AM becomes necrotic and contributes to the MTB namics of MTB against its natural target and with the spread. Otherwise, when the AM becomes apoptotic, sim- immune system of the host. ultaneously three specific scenarios can occur. Firstly, AM To simulate the effects of RUTI® vaccination strategy, apoptotic can interact with a lung resting macrophage (M) we put into the model the liposome dynamics and how and lead to efferocytosis of macrophages, in other words it interacts with the immune system. In particular we ex- an engulfment of AM apoptotic by M, essential for tissue tended UISS framework to include: activation of PRRs homeostasis and immunity. This means switches from that triggers the initiation of the innate immune re- “resting” to “active” status. sponse; activated CTLs that recognize peptides bound to Secondly, AM apoptotic cells can encounter a lung den- the major histocompatibility complex class I and II mol- dritic cell (DC). AM apoptotic can be taken up by DC that ecules (MHC-I, MHC-II), which express antigenic pep- capture antigens (Ag) through a process called nibbling; tides on APCs and bind to T cells via the T-cell then, DC will process and present the resulting fragments to receptor. TH cells provide help to antigen-specific B antigen-specific T lymphocytes in the context of molecules cells, resulting in antibody production. of the major histocompatibility complex of class I (MHC-I) or related proteins. From this point forward, MTB–antigen The disease conceptual model processing DCs, migrate to the local lung-draining lymph To model the entire dynamics of MTB and interactions with nodes (by 8–12 days post infection) driving naïve T cell the immune system machinery, we selected all the players polarization. This migration is influenced by IL-2 release and that have a role in TB in order to include the disease model other chemokines, except when IL-10 is present and is able into the simulation framework. The starting point was the to block this moving. A third interaction dealt with the secre- implementation of all the biological entities involved into the tion of MTB debris from AM apoptotic: MTB debris will dynamics of MTB, both at cellular and molecular scale. To interact with DCs, in status resting, that will process and analyze the behavior of a possible therapeutic intervention, present the resulting fragments to antigen-specific T lym- we selected one of the vaccines we are going to test inside phocytes in the context of molecules of the major histocom- the STriTuVaD project i.e., RUTI® vaccine. To this end, we patibility complex class II (MHC-II) or related proteins. implemented inside UISS the main mechanism of action When MTB infects a lung N, N produces and secretes (MoA) of this immunotherapeutic vaccine. IL-1 and other chemokines. Just like AM, also for N the Our model takes into account both innate and adaptive MTB strain can lead to a different role in the induction immunity (both cellular and humoral) and immune memory. of either apoptosis or necrosis and to the final outcome Figure 1 depicts all the entities implemented within the of the infection. simulation framework, especially immune cells, cytokines Both AM and N effector functions, can be negatively mod- and specific biological processes along with their peculiar ulated by IL-10 induction during MTB infection. Respect- properties in TB dynamics. We used Unified Modeling Lan- ively, IL-10 can lead to the inhibition of macrophage and guage (UML) to design the conceptual model. UML is a gen- neutrophil effector functions, with reduction of bacterial kill- eral-purpose modeling language that is extensively used in ing and impairing of secretion of cytokines and chemokines. the field of software engineering to provide a standard way As previously said, IL-10 can also block chemotactic to visualize the design of a system. All the entities of the TB factors that control DC moving to the lung-draining disease model interact each other, and are appropriately lo- lymph nodes. cated in two specific compartments: the lung (pulmonary al- The scenario inside the lymph node depicts the DC veoli) and the peripheral lymph nodes. cells in antigen presenting cell status secreting IL-12, The starting point of our TB disease model consists on the Type 1 IFN, IL-6 and IL-23 and driving naïve T cell dif- aerosol droplets of MTB that reach lung alveolar macro- ferentiation toward a Th1, Th2 or Th17 phenotype. Th phages (AMs) on one side and neutrophils (N) on the other cell population differentiation can be negatively modu- one. The initial challenge of MTB is implemented simulating lated by IL-10 and regulatory T cells (TReg). Protective Pennisi et al. BMC Bioinformatics 2019, 20(Suppl 6):504 Page 5 of 10 Fig. 1 The pulmonary tuberculosis – immune system interaction disease model. Conceptual description of the main entities and interactions of MTB – immune system. The main two compartments are represented: the lung and the peripheral lymph nodes. The representation depicts both cellular and humoral response when MTB droplets infect alveolar macrophages resident in the lung. The cascade of cytokines and chemokines is also represented with possible different behaviors depending on the virulence of the MTB strain antigen-specific Th1 cells migrate back to the lungs node, B cells interact also with MTB; after that B cells be- about 14–17 days after the initial exposure and infection come active and secrete immunoglobulins type M (IgM). to MTB and under a chemokine gradient (except when IL-10 blocks this process). In the lung, activated Th1 cell Results and discussion population, produce and secrete IFN-γ, causing macro- With the aim to obtain accurate in silico predictions about phage activation, relative cytokine production (IL-12 and the efficacy of therapeutical interventions directed against TNF-α) and bacterial control. It is worth to mention that MTB, we run different simulations to evaluate the degree in this context, IL-10 can block macrophage activation of precision of UISS when applied to a specific interven- and consequent cytokine secretion. Also TReg negatively tional vaccine scenario based on a promising vaccine i.e., modulate Th1 population effector functions. RUTI® vaccine. We used publicly available data coming from For the sake of completeness, the Th1 cell population in- phase II Clinical Trial in subjects with latent tuberculosis in- teracts also with B cells leading to three specific processes at fection treated with RUTI® [29], to evaluate the immunogen- the same time: after a successful interaction, B cells duplicate, icity. We analyzed three simulation scenarios. The first one differentiate in memory B and secrete immunoglobulins type is represented by the establishment of MTB latent chronic G (IgG). infection with some typical granuloma formation as a reser- Similarly, Th2 cell migration and interaction with B cells voir of MTB infection. The second one deals with a reactiva- leads to B cells duplication, differentiation in memory B cells tion phase during latent chronic infection where a possible and secretion of immunoglobulins type A (IgA). For what breakdown of the granuloma may lead to the spread of the concern Th17 cell population, their migration and inter- bacilli and to the reactivation of the disease, with an in- action with B cells will finally lead to cell duplication and se- creased necrotic burden. The third consists on the latent cretion of immunoglobulins type E (IgE). Inside the lymph chronic disease infection scenario during RUTI® vaccine Pennisi et al. BMC Bioinformatics 2019, 20(Suppl 6):504 Page 6 of 10 administration: here we show the evolution in time of M and (red lines). After the formation of the granuloma that repre- CD4 T cell populations, in conjunction with the temporal sents a reservoir of MTB infection, the infection remains evolution of IFN-γ that owns an important role as a bio- mostly latent, and also the CD4 T response that is recruited marker of protection. at the initial stages of infection (panel B) tends towards base- In Fig. 2 panel A, an initial infection of a virulent strain of line levels. This is also present in the Interferon-gamma plots MTB is supposed to happen at day 40. MTB rapidly infects (panel C) where after an initial peak, its level drastically both AM and M that become infected (blue lines). A similar drops. behavior (not shown here) also occurs for N. Then, MTB In Fig. 3 a breakdown of about the 50% of the total grows inside them eventually leading to either apoptosis or granulomatous burden is supposed to happen approxi- necrosis of the infected cell, accordingly to the levels of mately after 400 days after primary infection. This trans- LXA4 and PGE2 in the site of infection. In particular, the lates in a rapid drop in the number of entities that are virulent strain will drive towards the production of the pro part of the granuloma (yellow lines) and in a reactivation necrotic LXA4 (panel D) and thus to higher percentages of of the disease with a final worsening of the total granu- necrotic cells that will become part of the granuloma mass loma burden, even if an immune response is present A B C D Fig. 2 In silico latent tuberculosis infection scenario. Panel A depicts AM population dynamics. In particular, it is simulated the time course of the MTB-alveolar macrophages infected levels, the apoptotic ones and those forming granuloma. One can appreciate the typical granuloma behavior in LTBI as a potential protective immune system instrument from one side and the MTB reservoir of infection from the other one. Panel B shows Helper T cells dynamics. Subtypes (TH-1, TH-2 and TH- 17) are reported. LTBI induces a TH-2 switch instead of TH-1. This leads to a lesser immune system activation that could overcome the MTB spread. Panel C describes IFN-γ dynamics. After the first spike (representing the initial inflammatory response soon after the MTB infection), IFN-γ levels are kept almost at low level, indicating a latent typical scenario of immune system tolerance. In panel D LXA4 and PGE2 detailed dynamics are reported. The simulator correctly predicts a predominant LXA4 level indicating a pro-necrotic induction commonly observed in virulent strain of MTB causing a LTBI scenario. For all the simulated scenarios, time has been set to 720 days (2 years) and the virtual patient has been challenged with MTB at day 40 Pennisi et al. BMC Bioinformatics 2019, 20(Suppl 6):504 Page 7 of 10 A B C D Fig. 3 In silico latent tuberculosis infection with reactivation event scenario. Panel A depicts AM population dynamics. In particular, it is simulated the time course of the MTB-alveolar macrophages infected levels, the apoptotic ones and those forming granuloma. One can appreciate the breakdown of the granuloma leading to a successive reactivation at day 450. Moreover, granuloma increases in dimension indicating a worsening of the disease. Panel B shows Helper T cells dynamics. Subtypes (TH-1, TH-2 and TH-17) are reported. At the reactivation time, it is reported an evident TH-2 switch. However also a TH-1 driven response is present. Panel C describes IFN-γ dynamics. The peak of IFN- γ indicates an attempt of immune system to contain the MTB spreading. In panel D LXA4 and PGE2 detailed dynamics are reported. The simulator correctly predicts a predominant second peak of LXA4 indicating a pro-necrotic induction commonly observed in virulent strain of MTB. For all the simulated scenarios, time has been set to 720 days (2 years) and the virtual patient has been challenged with MTB at day 40 (see panels B and C), that is however unable to deal with response mirroring the IFN-γ, CD4 T cells, LXA4 and the spread of the bacteria. PGE2 dynamics. Moreover, form the spatial point of Finally, in Fig. 4 we show the effects of the vaccination view, the granuloma dynamics with MTB spread were with the RUTI® vaccine in an individual with latent TB appropriately reproduced during the two most wide- infection. Even if the vaccine seems to be mostly inef- spread scenarios characterizing pulmonary tuberculosis fective against the existing granuloma burden (panel A), i.e., latent and latent with reactivation. a strong Th1 induced response is elicited with the induc- tion of immunological memory (panel B). High levels of Conclusions IFN-γ are present, in good agreement with the results Computational models are important for the under- presented in [29] for a 25μg dosage of the RUTI® vaccine standing of biological systems. Such models can be ap- in latent patients. plied to enhance or predict therapeutic effects at the Overall, the simulations results are in very good agree- organism level. The pharmaceutical companies suggest ment with available data coming from clinical trials. In that computational biology can play an excellent role in particular, the in silico framework shows reliability in this field. In silico models can afford answers to the gen- capturing the main landmarks of both induced immune eral behaviour of the immune system, the analysis of Pennisi et al. BMC Bioinformatics 2019, 20(Suppl 6):504 Page 8 of 10 A B C D Fig. 4 In silico latent tuberculosis infection with RUTI vaccine administration. Panel A depicts AM population dynamics. In particular, it is simulated the time course of the MTB-alveolar macrophages infected levels, the apoptotic ones and those forming granuloma. Panel B shows Helper T cells dynamics. Subtypes (TH-1, TH-2 and TH-17) are reported. The RUTI vaccine has been administered at day 450 and at day 478 (two inoculations, 28 days interval). CD4 T cells (TH-1 subtype) are the predominant one, indicating a strong immune system response induced by the therapeutical intervention. Panel C describes IFN-γ dynamics. Here one can appreciate the presence of two peaks of IFN-γ indicating the vaccine activity in stimulating pro-inflammatory and TH-1 mediated immune system response. Moreover, the level of IFN-γ dynamics prediction mirrors the one observed in the RUTI phase II clinical trial. In panel D LXA4 and PGE2 detailed dynamics are reported. The simulator correctly predicts a predominant peak of LXA4 indicating a pro-necrotic induction commonly observed in virulent strain of MTB. For all the simulated scenarios, time has been set to 720 days (2 years) and the virtual patient has been challenged with MTB at day 40 cellular and molecular interactions, the effects of treat- a personalization of the virtual patient to reproduce the ments, and the course of diseases. The main goal of the physiology and the pathophysiology of the subject. STriTuVaD project is to create a computational infra- This UISS simulation platform will be able, when com- structure able to predict the general outcome of two dif- pared to the experimental results of the clinical trial ferent vaccination strategies against M. tuberculosis. In phase II, to predict with the necessary accuracy the bac- this work, we presented UISS framework implementa- terial load in the sputum smear. tion able to simulate the artificial immune response elic- ited by the vaccines. In particular we implemented the Availability and requirements components of RUTI® vaccine. Project name: UISS for TB. Further steps deal with the creation of a set of virtual Project home page: https://www.combine-group.org/software subjects with the aim to reproduce biological diversity of Operating system(s): Platform independent. the subjects we will simulate. Programming language: C and Python. Through a “vector of features” that combines both bio- Other requirements: none. logical and pathophysiological parameters we will obtain Any restrictions to use by non-academics: not applicable. Pennisi et al. BMC Bioinformatics 2019, 20(Suppl 6):504 Page 9 of 10 Abbreviations Italy, 95125 Catania, Italy. 3Etna Biotech S.r.l., 95121 Catania, Italy. 4 ABM: Agent Based Modeling; Ag: Antigens; AM: Alveolar Macrophage; Department of Transplant Immunology and Immunogenetics, All India AMK: Amikacin; APC: Antigen Processing Cell; CIP: Ciprofloxacin; Institute of Medical Sciences, New Delhi 110029, India. 5TuBerculosis Vaccine DC: Dendritic Cell; EMB: Ethambutol; FCMtb: Fragments of Mycobacterium Initiative (TBVI), Lelystad 8219, The Netherlands. 6Archivel Farma, S.L, 08916 tuberculosis cells; FM: Foamy macrophages; HLA: Human Leukocyte Antigen; Badalona, Spain. 7Experimental Tuberculosis Unit (UTE), Fundació Institut IgA: Immunoglobulins type A; IgE: Immunoglobulins type E; Germans Trias i Pujol (IGTP), Universitat Autònoma de Barcelona (UAB), IgG: Immunoglobulin class G; IgG: Immunoglobulins type G; Badalona, Spain. 8Centro de Investigación Biomédica en Red (CIBER) de IgM: Immunoglobulins type M; INH: Isoniazid; LTBI: Latent Tuberculosis Enfermedades Respiratorias, Madrid, Spain. 9Department of Industrial Infection; LVX: Levofloxacin; LXA4: Lipoxin A4; M: Macrophage; MDR- Engineering, Alma Mater Studiorum, University of Bologna, 40136 Bologna, TB: Multidrug-resistant Tuberculosis; MHC-I: Major Histocompatibility Italy. Complex class I; MHC-II: Major Histocompatibility Complex class II; MoA: Mechanism of Action; MTB: Mycobacterium tuberculosis; Received: 12 August 2019 Accepted: 21 August 2019 MXF: Moxifloxacin; N: Neutrophils; PGE2: prostaglandin E2; PZA: Pyrazinamide; Published: 12 December 2019 RMP: Rifampicin; SCC: Sputum Culture Conversion; SM: Streptomycin; STriTuVaD: In Silico Trial for Tuberculosis Vaccine Development; TB: Tuberculosis; TReg: Regulatory T cells; UISS: Universal Immune System References Simulator; WHO: World Health Organization; XDR-TB: Extensively Drug- 1. Furin J, Cox H, Pai M. Tuberculosis. Lancet. 2019;393(10181):1642–56. resistant Tuberculosis 2. Koch A, Mizrahi V. Mycobacterium tuberculosis. Trends Microbiol. 2018;26(6): 555–56. Acknowledgements 3. WHO. WHO Global Tuberculosis Report 2018. 2018. Not applicable. 4. Central TB Division. Annual Status report of Tb India 2016, Revised National TB Control Programme. J Chem Inf Model. 2013;53:1689–99. About this supplement 5. Fogel N. Tuberculosis: a disease without boundaries. Tuberculosis. 2015; This article has been published as part of BMC Bioinformatics Volume 20 95(5):527–31. Supplement 6, 2019: Towards computational modeling on immune system 6. Donald PR, Marais BJ, Barry CE. Age and the epidemiology and function. The full contents of the supplement are available online at https:// pathogenesis of tuberculosis. Lancet. 2010;375(9729):1852–54. bmcbioinformatics.biomedcentral.com/articles/supplements/volume-20- 7. Cadena AM, Fortune SM, Flynn JL. Heterogeneity in tuberculosis. Nat Rev supplement-6. Immunol. 2017;17(11):691–702. 8. Narasimhan P, Wood J, Macintyre CR, Mathai D. Risk factors for Authors’ contributions tuberculosis. Pulmonary Medicine. 2013;2013:828939. https://doi.org/1 MP: implemented the model, performed numerical simulations, analyzed 0.1155/2013/828939. data, wrote the manuscript. GR: developed the conceptual model, checked 9. Lönnroth K, Jaramillo E, Williams BG, Dye C, Raviglione M. Drivers of tuberculosis the biological adherence and meaning, analyzed data, wrote the manuscript. epidemics: the role of risk factors and social determinants. Soc Sci Med. 2009; GS: performed numerical simulations, developed the python scripts for 68(12):2240–46. results generations. AB: gave biomedical knowledge, wrote the manuscript. 10. Dheda K, Schwander SK, Zhu B, van RN Z-S, Zhang Y. The immunology of GAPP: performed numerical simulations, developed the python scripts for tuberculosis: from bench to bedside. Respirology. 2010;15:433–50. https:// results generations. EP: checked the biological adherence and meaning. doi.org/10.1111/j.1440-1843.2010.01739.x. DKM: gave immunological knowledge and useful data about pulmonary 11. O’Garra A, Redford PS, McNab FW, Bloom CI, Wilkinson RJ, Berry MPR. The tuberculosis. KBW: gave useful insights about pulmonary tuberculosis and immune response in tuberculosis. Annu Rev Immunol. 2013;31:475–527. granuloma formation. MA: gave biomedical knowledge and useful data https://doi.org/10.1146/annurev-immunol-032712-095939. about RUTI® vaccine. PJC: gave biomedical knowledge and useful data about 12. Cailleaux-Cezar M, De D, Xavier GM, CLG DS, FCQ DM, Ruffino-Netto A, et al. RUTI® vaccine. MV: gave useful insights about in silico clinical trial Tuberculosis incidence among contacts of active pulmonary tuberculosis. development. FP: supervised, conceived and developed the model, gave Int J Tuberc Lung Dis. 2009;13(2):190–95. computational immunological knowledge, drafted the manuscript. All 13. Lin PL, Flynn JL. Understanding Latent Tuberculosis: A Moving Target. J authors have read and approved the final manuscript. Immunol. 2010;185(1):15–22. 14. Tufariello JAM, Chan J, Flynn JAL. Latent tuberculosis: mechanisms of host Funding and bacillus that contribute to persistent infection. Lancet Infect Dis. 2003; Authors of this paper acknowledge support from the STriTuVaD project. The 3(9):578–90. STriTuVaD project has been funded by the European Commission, under the 15. Rockwood N, du Bruyn E, Morris T, Wilkinson RJ. Assessment of treatment contract H2020-SC1–2017- CNECT-2, No. 777123. The information and views response in tuberculosis: Expert Review of Respiratory Medicine; 2016;10(6): set out in this article are those of the authors and do not necessarily reflect 643–54. the official opinion of the European Commission. Neither the European Com- 16. Sotgiu G, Centis R, D’Ambrosio L, Battista Migliori G. Tuberculosis treatment mission institutions and bodies nor any person acting on their behalf may and drug regimens. Cold Spring Harb Perspect Med. 2015;5(5):a017822. be held responsible for the use which may be made of the information con- 17. Diel R, Hitte Nl, Schaberg T. Cost effectiveness of treating multi-drug tained therein. Publication costs are funded by the STriTuVaD project. resistant tuberculosis by adding Deltyba™ to background regimens in Germany. Respiratory Medicine. 2015;109(5):632–41. Availability of data and materials 18. Kerantzas CA, Jacobs WR. Origins of combination therapy for tuberculosis: The main computational framework is fully described in the paper. Lessons for future antimicrobial development and application. mBio. 2017; 8(2):e01586–16. Ethics approval and consent to participate 19. Tiberi S, Scardigli A, Centis R, D’Ambrosio L, Muñoz-Torrico M, Salazar- Not applicable. Lezama MÁ, et al. Classifying new anti-tuberculosis drugs: rationale and future perspectives. Int J Infect Dis. 2017;56:181–84. Consent for publication 20. Tiberi S, Muñoz-Torrico M, Duarte R, Dalcolmo M, D’Ambrosio L, Migliori GB. Not applicable. New drugs and perspectives for new anti-tuberculosis regimens. Rev Port Pneumol (English Edition). 2018. Competing interests 21. Nath H, Ryoo S. First– and second–line drugs and drug resistance. In: The authors declare that they have no competing interests. Tuberculosis - Current Issues in Diagnosis and Management; 2013. 22. Zumla A, Chakaya J, Centis R, D’Ambrosio L, Mwaba P, Bates M, et al. Author details Tuberculosis treatment and management-an update on treatment 1 Department of Mathematics and Computer Science, University of Catania, regimens, trials, new drugs, and adjunct therapies. Lancet Respir Med. 2015; 95125 Catania, Italy. 2Department of Drug Sciences, University of Catania, 3(3):220–34. Pennisi et al. BMC Bioinformatics 2019, 20(Suppl 6):504 Page 10 of 10 23. Haas MK, Belknap RW. Updates in the treatment of active and latent 44. Castiglione F, Duca K, Jarrah A, Laubenbacher R, Hochberg D, Thorley- tuberculosis. Semin Respir Crit Care Med. 2018;39:297–309. Lawson D. Simulating Epstein-Barr virus infection with C-ImmSim. 24. Getahun H, Matteelli A, Abubakar I, Abdel Aziz M, Baddeley A, Barreira D, et al. Bioinformatics. 2007;23(11):1371–77 Management of latent Mycobacterium tuberculosis infection: WHO guidelines 45. Hafsi S, Candido S, Maestro R, L F, Soua Z, Bonavida B, et al. correlation for low tuberculosis burden countries. Eur Respir J. 2015;46(6):1563–76. between the overexpression of yin Yang 1 and the expression levels of 25. Silva Miranda M, Breiman A, Allain S, Deknuydt F, Altare F. The tuberculous miRNAs in Burkitt’s lymphoma: a computational study. Oncol Lett. 2016;11: 607granuloma: An unsuccessful host defence mechanism providing a safety 1021–5. https://doi.org/10.3892/ol.2015.4031. 608shelter for the bacteria? Clin Dev Immunol. 2012;2012(2012):139127. 46. Polo A, Crispo A, Cerino P, Falzone L, Candido S, Giudice A, et al. 26. Oxlade O, Schwartzman K, Behr MA, Benedetti A, Pai M, Heymann J, et al. Environment and bladder cancer: molecular analysis by interaction Global tuberculosis trends: a reflection of changes in tuberculosis control or networks. Oncotarget. 2017;8:65240–52. https://doi.org/10.18632/ in population health? Int J Tuberc Lung Dis. 2009;13(10):1238–46. oncotarget.18222. 27. Schrager LK, Harris RC, Vekemans J. Research and development of new 47. An G, Fitzpatrick BG, Christley S, Federico P, Kanarek A, Neilan RM, et al. tuberculosis vaccines: a review. F1000Research. 2019;7:1732. Optimization and control of agent-based models in biology: a perspective. 28. Prabowo SA, Painter H, Zelmer A, Smith SG, Seifert K, Amat M, et al. Bull Math Biol. 2017;79(1):63–87. RUTI Vaccination Enhances Inhibition of Mycobacterial Growth ex vivo 48. Shay JW, Wright WE. Hayflick, his limit, and cellular ageing. Nat Rev Mol and Induces a Shift of Monocyte Phenotype in Mice. Front Immunol. Cell Biol. 2000;1(1):72–76. 2019;10:1–10. 49. Related Mechanisms of Antibody Somatic Hypermutation and Class Switch 29. Nell AS, D’lom E, Bouic P, Sabaté M, Bosser R, Picas J, et al. Safety, Recombination. In: Mobile DNA III. 2015. tolerability, and immunogenicity of the novel Antituberculous vaccine RUTI: 50. Hall EJ. The bystander effect. In: Health Physics. 2003. randomized. Placebo-Controlled Phase II Clinical Trial in Patients with Latent Tuberculosis Infection PLoS One. 2014;9:e89612. https://doi.org/10.1371/ journal.pone.0089612. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in 30. Vilaplana C, Montané E, Pinto S, Barriocanal AM, Domenech G, Torres F, et published maps and institutional affiliations. al. Double-blind, randomized, placebo-controlled phase I clinical trial of the therapeutical antituberculous vaccine RUTI®. Vaccine. 2010;28:1106–16. https://doi.org/10.1016/j.vaccine.2009.09.134. 31. Javaid A, Ahmad N, Afridi AK, Basit A, Khan AH, Ahmad I, et al. Validity of time to sputum culture conversion to predict cure in patients with multidrug-resistant tuberculosis: a retrospective single-center study. Am J Trop Med Hyg. 2018;98(6):1629–36. 32. Pappalardo F, Russo G, Tshinanu FM, Viceconti M. In silico clinical trials: concepts and early adoptions. Brief Bioinform. 2018:1–10. https://doi.org/1 0.1093/bib/bby043. 33. Carlier A, Vasilevich A, Marechal M, De Boer J, Geris L. In silico clinical trials for pediatric orphan diseases. Sci Rep. 2018;6;8(1):2465. 34. Viceconti M, Henney A, Morley-Fletcher E. In silico clinical trials: how computer simulation will transform the biomedical industry. Int J Clin Trials. 2016;3:37. https://doi.org/10.18203/2349-3259.ijct20161408. 35. Pappalardo F, Russo G, Pennisi M, Sgroi G, Alessandro G, Palumbo P, et al. Agent based modeling of relapsing multiple sclerosis: A possible approach to predict treatment outcome. In: Proceedings - 2018 IEEE international conference on bioinformatics and biomedicine, BIBM 2018; 2019. 36. Pappalardo F, Forero IM, Pennisi M, Palazon A, Melero I, Motta S. SimB16: modeling induced immune system response against B16-melanoma. PLoS One. 2011;6:e26523. https://doi.org/10.1371/journal.pone.0026523. 37. Pennisi M, Russo G, Ravalli S, Pappalardo F. Combining agent based-models and virtual screening techniques to predict the best citrus-derived vaccine adjuvants against human papilloma virus. BMC Bioinformatics. 2017;18:544. https://doi.org/10.1186/s12859-017-1961-9. 38. Palladini A, Nicoletti G, Pappalardo F, Murgo A, Grosso V, Stivani V, et al. In silico modeling and in vivo efficacy of cancer-preventive vaccinations. Cancer Res. 2010;70:7755–63. https://doi.org/10.1158/ 0008-5472.CAN-10-0701. 39. Gullo F, van der Garde M, Russo G, Pennisi M, Motta S, Pappalardo F, et al. Computational modeling of the expansion of human cord blood CD133 + hematopoietic stem/progenitor cells with different cytokine combinations. Bioinformatics. 2015;31:2514–22. https://doi.org/10.1093/bioinformatics/btv172. 40. Pennisi M, Russo G, Motta S, Pappalardo F. Agent based modeling of the effects of potential treatments over the blood–brain barrier in multiple sclerosis. J Immunol Methods. 2015;427:6–12. https://doi.org/10.1016/j.jim.2 015.08.014. 41. Pennisi M, Rajput A-M, Toldo L, Pappalardo F. Agent based modeling of Treg-Teff cross regulation in relapsing-remitting multiple sclerosis. BMC Bioinformatics. 2013;14(Suppl 16):S9. https://doi.org/10.1186/1471-21 05-14-S16-S9. 42. Pappalardo F, Motta S, Lollini PL, Mastriani E. Analysis of vaccine’s schedules using models. Cell Immunol. 2006;244(2):137–40. 43. Pappalardo F, Russo G, Pennisi M, Sgroi G, Alessandro G, Palumbo P, et al. An agent based modeling approach for the analysis of tuberculosis - Immune system dynamics. In: Proceedings - 2018 IEEE international conference on bioinformatics and biomedicine, BIBM 2018; 2019.