FIGURE 1.

Computational model system for predicting cell-mediated immune responses in Memory Design Space for multiple Ag specificities. (A) Known cell-mediated immune responses generated by vaccine or natural infection to various infections. Reported relative numbers of central memory (CM) and effector memory (EM) T cells following vaccination are shown for smallpox in humans (6 months post-vaccination) (Miller et al., 2008) and listeria in mice (day 35) (Busch and Pamer, 1999; Pamer, 2004; Wong et al., 2004). Similarly, memory T cell populations generated following human vaccination with BCG (10 weeks) (Fletcher, 2007; Soares et al., 2008), MVA85A boosting BCG (24 weeks (Beveridge et al., 2007) or 56 days (Scriba et al., 2010)), and H56 (100 days after the 2nd of two boosts (Luabeya et al., 2015) are plotted, as well as T cells generated as a result of natural infection (marked by asterisks) with M. tuberculosis in patients with active TB disease (Goletti et al., 2006; Wang et al., 2010), latent TB infection (LTBI) (Wang et al., 2010), or successfully treated TB (1 month post-treatment) (Wang et al., 2010). We refer to a plot EM and CM T cells remaining in blood and peripheral tissues after infection has cleared and memory is established (time point t = 30 days) as Memory Design Space. Both the size and skew of the memory population can be easily visualized in Memory Design Space. (B) Schematic of computational model. Our 3-compartment hybrid model comprises an agent-based model of the lymph node and systems of ordinary differential equations representing blood and peripheral tissues. Ag-specific naïve T cells (N) in the LN may be primed by dendritic cells (DCs) and differentiate (graded arrows) to memory subtypes. In blood, cells may die, transit to other compartments, and some EM cells convert to CMs. Effector cells (E) and EMs that enter peripheral tissues do not reenter circulation. (C) Model quantities capturing antigen presentation. DCs display p_iMHCs from several antigens (shown by different shapes). The number of pMHC complexes displayed on each Ag-bearing DC that enters the LN is P_tot, the number of complexes for each antigen (p_iMHC) is x_i and the fraction of total complexes for each antigen is m_i. (D) Binding probability, P_i(bind) describes the probability that a DC with an available pMHC receptor will bind a T cell in its neighborhood. This depends on x_i, the number of p_iMHC displayed that match the T cell’s specificity, and a_i, the number of p_iMHC necessary to achieve a 50% binding probability according to the equation: P_i(bind) , as was developed for previous versions of the model (Riggs et al., 2008; Linderman et al., 2010).