Live Attenuated Whole-Cell Vaccines

23.2.1 Overview

The current trend towards testing novel live at­tenuated MAP strains (LAV) produced by mo­lecular genetic techniques as vaccine candidates follows the growing evidence that specific at­tenuation of M.

Initial attempts to identify candidate tar­gets determining MAP virulence have relied in the most part, on in vitro assays to provide a preliminary screen before entering animal stud­ies. Techniques employed have included the on­set of auxotrophy (Cavaignac et al., 2000) and, more widely, differential survival of candidate LAV vs wild-type strain controls in short-lived monocyte-derived macrophage (MDM) as­says involving either uptake and entry, short­time survival or ability to induce apoptosis (Scandurra et al., 2009; Park et al., 2011).

Genomic manipulation approaches used to generate novel LAV candidates have includ­ed random transposon mutagenesis (Harris et al., 1999; Shin et al., 2006) and directed al­lelic exchange through specialized shuttle plas­midphage transduction, first developed in M. tuberculosis (Bardarov et al., 2002; Park et al., 2008). Strain libraries generated by random in­sertion mutagenesis using a kanamycin-marked transposon (TN5367) were pre-screened for potential attenuation using auxotrophic media (Cavaignac et al., 2000) or ip. challenge in mice 12 weeks/goats 6-12 months (Shin et al., 2006; Scandurra et al., 2010). This identified a short­list of LAV candidates (Table 23.1) with defined MAP gene knockouts (MAP1566, MAP0053c, MAP0011, MAP3963, MAP2408c) that were chosen for downstream challenge studies in animals (Scandurra et al., 2010; Settles et al.,

2014).

Despite these high likelihood ‘indicator’ pa­rameters, each of eight LAV chosen when tested in mice (CL57BL/6) using a low-dose immuniza­tion given ip. followed by low-dose ip. challenge after 6 weeks, failed to show any significant protection 12 weeks post-challenge relative to a phosphate buffered saline (PBS) sham vacci­nated group (Bannantine et al., 2014a). Parallel study of two further strains LAV MAP3963 and MAP2408c using a higher dose immunization but administered sc., boosted after 2 weeks, fol­lowed by low-dose ip. challenge given 5 weeks later showed only minor improvement in pro­tection, with only one log decrease in mucosal tissue loads relative to PBS sham vaccination 18 weeks post-challenge (Settles et al., 2014).

These observations are somewhat in line with studies on other LAV candidates. LAV strains MAP3025c, MAP3291c and MAP1534 screened for attenuated persistence in C57BL/6 mice using a high ip. dose, showed significant attenuation (but no clearance) after 12 weeks (Chen et al., 2012). High ip. dose immunization of C57BL/6 mice with these candidates followed by high oral dose challenge, 3 weeks post­immunization, provided a decrease in visible acid-fast organisms by microscopy 12 weeks post-challenge relative to PBS sham vaccina­tion. Further study of LAV MAP3025c in goats using high sc. dose immunization, boosted at 3 weeks, followed by high oral dose challenge, 3 weeks post-immunization, provided a decrease in tissue loads 6 months post-challenge and some evidence of decreases in faecal shedding relative to PBS sham but no reduction superior to commercial (Mycopar) vaccination (Faisal et al., 2013a). Of note here is the choice to de­liver LAV immunizations ip., a route previously shown to be ineffective for live 316F-based vac­cines (Griffin et al., 2009).

A separate study using LAV MAP1566, MAP0053c and MAP0011 showed differential induction of apoptosis in MDM and reduced persistence after 12-day infections relative to wild-type strains (Kabara and Coussens, 2012).

Study of single-gene allelic exchange mu­tants MAP4201 or MAP3323c, both global signalling factors, showed LAV attenuation in vitro using MDM at 7 days and some attenua­tion when given ip. at high dose into immunode­ficient BALB/c mice (Ghosh et al., 2013, 2014,

2015). Testing in C57BL/6 mice using a low sc.

Notably, despite some attempts at stand­ardizing methodology and encouraging the use of comparable controls, none of the studies de­scribed above followed the same protocol for any species of animal and mouse models using a va­riety of strains. In some studies power calcula­tions on group sizes prior to testing were absent or not reported and test groups in larger animals were often very low. Other parameters seemed to have been arbitrarily chosen. Immunization dos­ing ranged from 1 ? 10⁴ - 2 ? 10⁹ CFU/animal and multiple variations of ip., iv., sc., id., im. and oral were employed to immunize and/or chal­lenge. Some LAV were boosted and some, par­ticularly those given orally, were only single dose. Several different challenge strains were used among the published studies. Times between the last immunizing dose and challenge ranged from 3- 8 weeks, longitudinal monitoring of LAV and challenge strain excretion was often omitted or deemed impossible to perform and the important final time point chosen to evaluate the degree of protection post-challenge were even greater in range. The use of Silirum heat-killed vaccine in mice was reported to be ineffective (Bannantine et al., 2014a) thus in many cases there was no valid control group allowing comparison with one of the current commercial killed whole-cell vaccine ‘gold standards’ and protections rela­tive to sham vaccination with PBS were used. Surprisingly, LAV were delivered in most cases without adjuvant, a major factor in subunit­based vaccines. In the only case where one was used, comparison was made on one LAV only (Ghosh et al., 2015).

The reasons for not considering adjuvanted LAV are unclear. Early MAP vaccines using clas­sically attenuated strains were all delivered sc., live in aggressive adjuvants such as mineral oil or pumice and, as a result, stimulated a wide range of immunological responses, in a similar way to the dead whole vaccines that have re­placed them. One attempt to induce protection using a non-adjuvanted live strain of MAP316F in sheep however was unsuccessful (Begg and Griffin, 2005). This is contrary to the situa­tion pertaining to the use of BCG in protecting against tuberculosis, where live BCG without an adjuvant gives very much better protection than dead BCG administered in a mineral oil adjuvant (Griffin et al., 1999). Using an appropriate ad­juvant however could be critical. A study com­paring adjuvanted vaccination in sheep (Griffin et al., 2009) suggested standard mineral oil­based vaccine sc. delivery of a killed 316F strain invoked vigorous cell-mediated and humoral reactivity while live 316F delivered in lipid- based adjuvant formulations invoked lower cell- mediated responses and no humoral response.

The development of novel live MAP strains to use as vaccines has several important regula­tory, bio-safety and diagnostic issues that will need to be addressed, similar to some of those pertaining to live tuberculosis vaccines (Walker et al., 2010), before commercialization is feasi­ble. Nevertheless, the use of well-characterized stable attenuated MAP mutants as vaccines re­mains attractive as they would be cheaper than subunit-based vaccines to manufacture and at this moment appear to offer the opportunity of delivering superior protection possibly without the need for a logistically difficult requirement for administering a boost dose.

23.3