Experimental and natural infections in MyD88- and IRAK-4-deficient mice and humans.

Most Toll-like-receptors (TLRs) and interleukin-1 receptors (IL-1Rs) signal via myeloid differentiation primary response 88 (MyD88) and interleukin-1 receptor-associated kinase 4 (IRAK-4). The combined roles of these two receptor families in the course of experimental infections have been assessed in MyD88- and IRAK-4-deficient mice for almost fifteen years. These animals have been shown to be susceptible to 46 pathogens: 27 bacteria, eight viruses, seven parasites, and four fungi. Humans with inborn MyD88 or IRAK-4 deficiency were first identified in 2003. They suffer from naturally occurring life-threatening infections caused by a small number of bacterial species, although the incidence and severity of these infections decrease with age. Mouse TLR- and IL-1R-dependent immunity mediated by MyD88 and IRAK-4 seems to be vital to combat a wide array of experimentally administered pathogens at most ages. By contrast, human TLR- and IL-1R-dependent immunity mediated by MyD88 and IRAK-4 seems to be effective in the natural setting against only a few bacteria and is most important in infancy and early childhood. The roles of TLRs and IL-1Rs in protective immunity deduced from studies in mutant mice subjected to experimental infections should therefore be reconsidered in the light of findings for natural infections in humans carrying mutations as discussed in this review.

Introduction

MyD88 was first described as a “macrophage differentiation marker” for which mRNA accumulated in murine M1 myeloleukemic cells upon activation with IL‐6 1, 2. Human MYD88 maps to chromosome 3p22‐p21.3 and contains five exons 1. The full‐length cDNA for human MYD88 encodes 296 amino acids forming a 33 kDa protein 2. Murine Myd88 maps to chromosome 9q119. It also has five exons and the full‐length cDNA encodes 296 amino acids forming a 33‐kDa protein. The MyD88 protein includes an N‐terminal “death domain” (DD) and a C‐terminal Toll‐interleukin receptor (TIR) domain, similar to the intracellular domains of TLRs and members of the IL‐1R superfamily, collectively referred to as TIR receptors 2, 3, 4, 5, 6. Human IRAK4 maps to chromosome 12q12 and contains 13 exons. The full‐length cDNA for human IRAK‐4 encodes 460 amino acids, forming a 52 kDa protein. The murine Irak4 gene maps to chromosome 15q94 and also contains 13 exons. The corresponding full‐length cDNA encodes 459 amino acids, forming a 52 kDa protein. The IRAK‐4 protein contains an N‐terminal DD and a central kinase domain 7.

MyD88 and IRAK‐4 are essential for signaling via all TLRs with the exception of TLR3 and, to some extent, TLR4, and for signaling via most IL‐1Rs, including IL‐1R1, IL‐18R, and IL‐33R (ST2) 8, 9, 10, 11, 12, 13, 14, 15, 16, 17. Following its activation, MyD88 binds to the IL‐1Rs and TLRs via its TIR domain, forming an oligomer; it then recruits IRAK‐4 to the receptor via its DD 5, 18, 19, 20, 21, mediating the activation of various transcription factors, including IRF5 and IRF7, AP‐1, and NF‐κB, depending in part on the cell type and the cell surface receptor stimulated 22 (Fig. 1). Thus, TIR‐MyD88‐IRAK‐4‐mediated signaling appears to be important for the innate recognition of pathogens and the ignition of inflammation, and, as such, is indispensable for the induction of protective immunity.

Figure 1

MyD88‐ and IRAK‐4‐signaling pathways. The TIR superfamily (TLRs/IL‐1Rs) is dependent on MyD88 and IRAK‐4 signaling for its regulation of gene transcription. MyD88 and IRAK‐4 control IRF (Interferon regulatory factors), MAPK (map kinases), and NEMO (NF‐κB essential modulator) that regulate AP‐1 and NF‐κB (p50/p65); the latter by stimulating the phosphorylation and degradation of IκBα so releasing NF‐κB from the inactive NF‐κB/IκBα complex. MyD88 also controls IRF5‐ and IRF7‐dependent signaling (not shown).

However, most demonstrations of the importance of TIR‐MyD88‐IRAK‐4‐dependent pathways for protective immunity have been based on studies of experimental infections in mice in which protective immunity against a broad range of infectious agents has been shown; however, the essential nature of the role of TIR signaling in such broad‐ranging immunity has been called into question by both clinical genetic and evolutionary genetic studies 23, 24, 25, 26, 27, 28. In particular, the identification of human IRAK‐4 and MyD88 deficiencies as immunological and clinical phenocopies has provided considerable insights 29, 30. Detailed immunological and clinical descriptions of a large cohort of patients with these deficiencies have led to a reassessment of the importance of the TIR‐MyD88‐IRAK‐4‐dependent pathway for general protective immunity in humans under natural conditions 31.

Impact of MyD88 and IRAK‐4 deficiencies on protective immunity in mice

Susceptibility to pathogens in MyD88‐ and IRAK‐4‐deficient mice

Myd88‐deficient mice are known to be susceptible to experimental infections with 45 pathogens: 27 bacteria 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, eight viruses 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, seven protozoa 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 105, 106, 107, and four fungi 108, 109, 110, 111, 112, 113. Enhanced pathogen growth in Myd88‐deficient mice has been observed for:

Six Gram‐positive bacteria: Bacillus anthracis (spores injected subcutaneously (s.c.)) 77, Listeria monocytogenes after i.v. or i.p. injection or infection via gavage 33, 34, 38, 59 , Staphylococcus aureus after i.v. or s.c. injection 32, 50, Streptococcus agalacticae after s.c. or i.p. injection 37, Streptococcus pneumoniae after i.v. or i.n. infection 45, 48, Streptococcus pyogenes after s.c. injection 72;

Eighteen Gram‐negative bacteria: Anaplasmataceae after i.p. injection 73, Borrelia burgdorferi after intradermal (i.d.) inoculation 42, 43, 51, Borrelia hermsii after i.p. injection 52, Brucella abortus after i.p. injection 46, 64, Burkholderia pseudomallei after i.n. inoculation 67, Campylobacter jejunii after stomach gavage 60, Chlamydia muridarum after i.n. inoculation 71, Chlamydia pneumoniae after i.n. application 47, Citrobacter koseri after the direct inoculation of live bacteria into the brain parenchyma by stereotactic injection 76, Citrobacter rodentium after ingestion of a suspension of the bacterium or gavage 61, 62, Franciscella tularensis after i.n. inoculation and i.d. deposition 53, 68, Haemophilus influenzae after i.n. inoculation or i.p. injection 49, 66, Klebsiella pneumoniae after intratracheal (i.t.) inoculation 69, Legionella pneumoniae after exposure to aerosols or after i.n. inoculation 54, 55, Mycoplasma pneumoniae after i.n. inoculation 74, Neisseria meningitidis after i.p. injection 56, Pseudomonas aeruginosa after exposure to aerosolized bacteria or i.n. inoculation 36, 44, 57, 63 , and Salmonella typhimurium after i.v. injection 70, 75;

Three mycobacteria after i.v., i.n., or aerogenic exposure (Mycobacterium avium, Mycobacterium bovis, Mycobacterium tuberculosis) 35, 39, 40, 41, 58, 65;

Eight viruses: Herpes simplex virus type 1 after aerogenic exposure 80, Herpes simplex virus type 2 after vaginal challenge 83, Influenza A virus after i.n. inoculation 84, Lymphocytic choriomeningitis virus after i.v. injection 81, 85, murine cytomegalovirus after i.p. injection 78, 79, 82, 86, rabies virus after intracranial injection 90, 91, SARS coronavirus after i.n. inoculation 89, and vesicular stomatitis virus after i.n. inoculation or i.v. injection 87, 88;

Seven parasites: Cryptosporidium parvum after gavage 101, Enterocytozoon bieneusi after oral inoculation 107, Leishmania braziliensis after the s.c. injection of stationary‐phase promastigotes 104, Leishmania major after the s.c. injection of stationary‐phase promastigotes 94, 95, 105, Toxoplasma gondii after i.p. injection 92, 93, 96, 98, 103, 106, Trypanosoma brucei after i.p. infection 99, Trypanosoma cruzii after i.p. injection 97, 100, 102;

Four fungi: Aspergillus after i.v. infection 108, Candida albicans after i.v or intragastric injection 108, 109, 112, Cryptococcus neoformans after i.n. inoculation or i.p. injection 110, 111, Paracoccidioides brasiliensis after i.t. inoculation 113 (see Table 1).

Table 1

Forty‐six pathogens displaying higher growth rates in vivo in MyD88‐deficient mice than in wild‐type controls in experimental conditions

Pathogen group	Strain	References
Gram‐positive bacteria	Bacillus anthracis	77
	Listeria monocytogenes	33, 34, 38, 59
	Staphylococcus aureus	32, 50
	Streptococcus agalacticae	37
	Streptococcus pneumoniae	45, 48
	Streptococcus pyogenes	72
Gram‐negative bacteria	Anaplasmataceae	73
	Borrelia burgdorferi	42, 43, 51
	Borrelia hermsii	52
	Brucella abortus	46, 64
	Burkholderia pseudomallei	67
	Campylobacter jejuni	60
	Chlamydia muridarum	71
	Chlamydia pneumoniae	47
	Citrobacter koseri	76
	Citrobacter rodentium	61, 62
	Francisella tularensis	53, 68
	Haemophilus influenzae	49, 66
	Klebsiella pneumoniae	69
	Legionella pneumoniae	54, 55
	Mycoplasma pneumoniae	74
	Neisseria meningitidis	56
	Pseudomonas aeruginosa	36, 44, 57, 63
	Salmonella typhimurium	70, 75
Mycobacteria	Mycobacterium avium	35
	Mycobacterium bovis	39
	Mycobacterium tuberculosis	40, 41, 58, 65
Viruses	Herpes simplex virus 1	80
	Herpes simplex virus 2	83
	Influenza A virus	84
	Lymphocytic choriomeningitis virus	81, 85
	Murine cytomegalovirus	78, 79, 82, 86
	Rabies virus	90, 91
	SARS coronavirus	89
	Vesicular stomatitis virus	87, 88
Parasites	Cryptosporidium parvum	101
	Enterocytozoon bieneusi	107
	Leishmania braziliensis	104
	Leishmania major	94, 95, 105
	Toxoplasma gondii	92, 93, 96, 98, 103, 106
	Trypanosoma brucei	99
	Trypanosoma cruzii	97, 100, 102
Fungi	Aspergillus spp.	108
	Candida albicans	108, 109, 112
	Cryptococcus neoformans	110, 111
	Paracoccidioides brasiliensis	113

IRAK‐4‐deficient mice showed enhanced pathogen growth when challenged with S. aureus i.p. 17.

Survival of MyD88‐ and IRAK‐4‐deficient mice

Survival of deficient mice

As greater pathogen growth in vivo is not always correlated with lower levels of survival, we consider here the published mortality data for experimental infections of MyD88‐deficient mice. Mortality due to experimental infections was greater in MyD88‐deficient mice than in wild‐type mice for 33 pathogens:

Six Gram‐positive bacteria: B. anthracis after i.p. injection of the toxin 77, L. monocytogenes after i.v injection 33, S. aureus after i.v injection 32, S. agalacticae after s.c. or i.p. injection 37, S. pneumoniae after i.v. or i.n. infection 45, 48, S. pyogenes after s.c. injection 72;

Eight Gram‐negative bacteria: Anaplasmataceae after i.p. injection 73, B. hermsii after i.p. injection 52, B. pseudomallei after i.n. inoculation 67, C. muridarum after i.n. inoculation 71, C. pneumoniae after i.n. application 47, F. tularensis after i.n. inoculation and i.d. deposition 53, 68, K. pneumoniae after i.t. inoculation 69, P. aeruginosa after exposure to aerosolized bacteria or i.n. inoculation 36, 44, 57, 63;

Three mycobacteria after i.v., i.n., or aerogenic exposure (M. avium, M. bovis, M. tuberculosis) 35, 39, 40, 41, 58;

Eight viruses: Herpes simplex virus type 1 after aerogenic exposure 80, herpes simplex virus type 2 after vaginal challenge 83, influenza A virus after i.n. inoculation 84, lymphocytic choriomeningitis virus after i.v. injection 81, 85, murine cytomegalovirus after i.p. injection 78, 82, Rabies virus after intracranial injection 90, 91, SARS coronavirus after i.n. inoculation 89, vesicular stomatitis virus after i.n. inoculation or i.v. injection 87, 88;

Four parasites: C. parvum after gavage 101, T. gondii after i.p. injection 92, 93, 96, 98, 103, 106, T. brucei after i.p. infection 99, T. cruzii after i.p. injection 97, 102;

Four fungi: Aspergillus after i.v. infection 108, C. albicans after i.v. or intragastric injection 108, 109, 112, C. neoformans after i.n. inoculation or i.p. injection 110, 111, P. brasiliensis after i.t. inoculation 113 (see Table 2).

Table 2

Thirty‐three pathogens for which the mortality of MyD88‐deficient mice in vivo was greater than that of wild‐type controls in experimental conditions

Pathogen group	Strain	References
Gram‐positive bacteria	Bacillus anthracis	77
	Listeria monocytogenes	33
	Staphylococcus aureus	32
	Streptococcus agalacticae	37
	Streptococcus pneumoniae	45, 48
	Streptococcus pyogenes	72
Gram‐negative bacteria	Anaplasmataceae	73
	Borrelia hermsii	52
	Burkholderia pseudomallei	67
	Chlamydia muridarum	71
	Chlamydia pneumoniae	47
	Francisella tularensis	53, 68
	Klebsiella pneumoniae	69
	Pseudomonas aeruginosa	36, 44, 57, 63
Mycobacteria	Mycobacterium avium	35
	Mycobacterium bovis	39
	Mycobacterium tuberculosis	40, 41, 58, 65
Viruses	Herpes simplex virus 1	80
	Herpes simplex virus 2	83
	Influenza A virus	84
	Lymphocytic choriomeningitis virus	81, 85
	Murine cytomegalovirus	78, 82
	Rabies virus	90, 91
	SARS coronavirus	89
	Vesicular stomatitis virus	87, 88
Parasites	Cryptosporidium parvum	101
	Toxoplasma gondii	92, 93, 96, 98, 103, 106
	Trypanosoma brucei	99
	Trypanosoma cruzii	97, 102
Fungi	Aspergillus spp.	108
	Candida albicans	108, 109, 112
	Cryptococcus neoformans	110, 111
	Paracoccidioides brasiliensis	113

IRAK‐4‐deficient mice displayed lower levels of survival than wild‐type mice following i.p. challenge with S. aureus 17.

Impact of MyD88‐ and IRAK‐4 deficiencies on protective immunity in humans

Susceptibility to pathogens in MyD88‐ and IRAK‐4‐deficient patients

The initial description of human IRAK‐4 deficiency was based on three patients 29 and that of human MyD88 deficiency was based on nine patients 30. These patients all carried either homozygous or compound heterozygous mutations of the IRAK4 or MYD88 gene that lead to nonfunctional proteins 29, 30. Given these small numbers of patients, only brief preliminary conclusions could be made concerning the infectious phenotype associated with the absence of MyD88‐IRAK‐4‐dependent signaling. The cumulative evidence from the large number of case reports since published and from the comprehensive description of a cohort of 76 patients (52 with IRAK‐4 deficiency and 24 with MyD88 deficiency) allow firmer conclusions about the infectious phenotype in the absence of MyD88‐IRAK‐4‐dependent signaling to be drawn 31, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 127, 128, 129. There are also an additional five patients with IRAK‐4 deficiency and two patients with MyD88 deficiency for whom no data have yet been published. The infectious pheno‐type of MyD88‐ and IRAK‐4‐deficient patients is dominated by invasive pyogenic infections. The most frequent of such infections are meningitis, sepsis, arthritis, and osteomyelitis, and the principal bacteria isolated in cases of invasive infection are S. pneumoniae, S. aureus, and Pseudmomonas aeruginosa (for a list of all the pathogens isolated from Myd88‐ and IRAK‐4‐deficient patients, see Table 3). These patients also typically suffer from deep tissue infections of the upper respiratory tract, such as severe tonsillitis due to P. aeruginosa in particular, and superficial skin infections, mostly caused by S. aureus.

Table 3

Invasive infections in patients with impaired MyD88‐IRAK‐4 signaling caused by six Gram‐positive and 13 Gram‐negative bacteria

Pathogen group	Strain	References
Gram‐positive bacteria	Staphylococcus aureus	29, 30, 31, 115, 117, 119, 121, 125, 126, 128
	Streptococcus agalacticae	30, 31, 127, 128
	Streptococcus milleri	125
	Streptococcus parasanguis	118
	Streptococcus pneumoniae	29, 30, 31, 114, 115, 116, 117, 118, 119, 120, 121, 122, 123, 124, 125, 126, 128
	Streptococcus pyogenes	31, 128
Gram‐negative bacteria	Acinetobacter baumanii	128
	Citrobacter freundii	31
	Clostridium septicum	31, 114, 116
	Escherichia coli	31
	Haemophilus influenzae	31
	Klebsiella pneumoniae	31
	Moraxella catarrhalis	31
	Neisseria meningitidis	31, 114, 116, 125
	Proteus spp.	30, 31
	Pseudomonas aeruginosa	30, 31, 120, 121, 122, 125, 128
	Salmonella enterica	30, 31
	Serratia marcescens	31
	Shigella sonnei	31, 119, 125, 127

We cannot rule out the possibility that the predominance of Gram‐positive bacteria in patients with MyD88 and IRAK‐4 deficiencies results at least in part from a patient recruitment bias. No patients with these deficiencies have yet been identified in the Indian subcontinent, in South America or in China. Patients with MyD88‐ or IRAK‐4‐deficiencies in these areas of the world might perhaps present a higher frequency of infections with Gram‐negative bacteria, as suggested by case reports of invasive infection with Shigella spp. during endemic diarrhea outbreaks 119, 127. However, by contrast to this uncertainty concerning positive associations about the infectious phenotype, we can highlight much more emphatically the negative associations drawn concerning the roles of MyD88 and IRAK‐4 in host anti‐pathogen defense. Most, if not all, of the 76 patients identified to date have been exposed to mycobacteria, viruses, Toxoplasma, Pneumocystis, and other fungi, but none of these pathogens caused invasive infection. This strongly suggests that MyD88 and IRAK‐4 are dispensable in humans for defense against these pathogens, contrary to expectations based on the results obtained in the mouse model 31.

Survival of MyD88‐ and IRAK‐4‐deficient patients

At the time of writing, 26 patients with MyD88 deficiency have been identified (22 published 31, two unpublished (von Bernuth, unpublished)). Nine of these patients have died since identification: five in infancy and four in early childhood. The youngest of these nine patients died at 1 month of age and the oldest died at 4 years of age. The 15 surviving patients are currently four, seven, 11, 14, and 20 years old. Fifty‐two patients with IRAK‐4 deficiency have been identified 31. Twenty‐one of these patients have died since identification: 10 in infancy and 11 in early childhood. The youngest of these 19 patients was 2 months old, and the oldest was 7 years old, at the time of death; the latter being patient P23 from the large cohort published in 2010 31 who recently died of S. pneumoniae meningitis (unpublished observation). The 31 surviving patients are currently two (two patients), three (one patient), four (one patient), five (two patients), six (two patients), seven (three patients), nine (one patient), 12 (one patient), 13 (two patients), 14 (two patients), 15 (two patients), 16 (two patients), 17 (two patients), 18 (one patient), 20 (one patient), 21 (one patient), 22 (one patient), 30 (one patient), 33 (two patients), and 38 (one patient) years old 31. Thus, it can clearly be seen that human MyD88‐ and IRAK‐4‐deficiencies are life threatening. MyD88 and IRAK‐4 are indispensable for survival in infancy and early childhood and, before the advent of vaccines and antibiotics, most if not all children with these defects would have died in the first few years of life. However, several individuals with MyD88 deficiency or its immunological phenocopy, IRAK‐4‐deficiency, who were given antibiotic prophylaxis and even sometimes IgG substitution following the genetic identification of the disease, have survived into adolescence and adulthood. Many of these patients have since stopped taking regular antibiotic prophylaxis, but have not yet developed invasive pyogenic infections. MyD88‐IRAK‐4‐dependent signaling, therefore, appears to be dispensable for survival after adolescence.

Closing remarks

The notion that TLR‐ and IL‐1R‐mediated innate immune recognition is indispensable for survival and protective defense against many pathogens — based largely on findings in mouse models of experimental infections — should be reconsidered in light of the naturally occurring infections in humans with MyD88‐ or IRAK‐4‐deficiency. By contrast to the broad susceptibility of MyD88‐deficient mice to 46 different bacteria, viruses, protozoa, and fungi (i.e. to almost nearly all the microbes tested), patients with MyD88‐ or IRAK‐4‐deficiencies are susceptible to invasive and noninvasive infections with only a few Gram‐positive and Gram‐negative bacteria. Moreover, MyD88‐IRAK‐4‐mediated TLR and IL‐1R immunity is undoubtedly vital in infancy and early childhood, but gradually becomes dispensable, from adolescence onwards. Overall, MyD88‐dependent TLR and IL‐1R immunity is vital in both mice and humans, but its role in the course of naturally occurring infections in humans seems to be much more restricted than initially inferred from experimental infections in mice, as humans lacking functional MyD88 or IRAK‐4 proteins are susceptible to a narrow range of pathogens, and only in infancy and early childhood. The different outcomes between experimental infections in mice and natural infections in humans may be due to species‐specific differences, or more likely to differences in the modes of infection. In that regard, the study of naturally occurring infections in MyD88‐ and IRAK‐4‐deficient mice would be insightful, as suggested by preliminary studies 130. In any case, the studies of MyD88‐ and IRAK4‐deficient humans neatly illustrate the value of dissecting inborn errors of immunity underlying pediatric infectious diseases for deciphering the redundant and nonredundant roles of host defense genes in natura 23, 24, 25, 26, 27, 28. Immunological redundancy is greater in the course of natural infections in outbred human populations than in the course of experimental infections in inbred mice. Genetic studies of this type will facilitate the burgeoning, long‐awaited investigation of the contribution of immunity to health, and disease in humans 131, 132, 133.

Conflict of interest

The authors have declared no conflict of interest.

death domain

Toll‐interleukin receptor