Age-dependent differences in efferocytosis determine the outcome of opsonophagocytic protection from invasive pathogens

Summary

In early life, susceptibility to invasive infection skews towards a small subset of microbes, while other pathogens associated with disease later in life, including Streptococcus pneumoniae (Spn), are uncommon among neonates. To delineate mechanisms behind age-dependent susceptibility, we compared age-specific mouse models of invasive Spn infection. We show enhanced CD11b-dependent opsonophagocytosis by neonatal neutrophils improved protection against Spn during early life. Augmented function of neonatal neutrophils was mediated by higher CD11b surface expression at the population level due to dampened efferocytosis, which also resulted in more CD11b^hi ‘aged’ neutrophils in peripheral blood. Dampened efferocytosis during early life could be attributed to lack of CD169⁺ macrophages in neonates, and reduced systemic expression of multiple efferocytic mediators, including MerTK. Upon experimentally impairing efferocytosis later in life, CD11b^hi neutrophils increased and protection against Spn improved. Our findings reveal how age-dependent differences in efferocytosis determine infection outcome through modulation of CD11b-driven opsonophagocytosis and immunity.

Graphical Abstract

eTOC Blurb

Many pathogens rarely cause invasive disease during neonatal life. Bee et al. delineate an immunologic determinant of this phenomenon. During early life, developmental impairments in macrophage function (efferocytosis) alter neutrophil homeostasis to augment CD11b-dependent opsonophagocytosis. This results in increased protection against certain pathogens, and accounts for age-related patterns of susceptibility.

Introduction

While newborns are generally considered more susceptible to infection, this trend is predominantly seen with a small subset of microbes, clinically defined as neonatal pathogens. Many microbes known to cause systemic infection at later stages in life are infrequently problematic during the neonatal period¹. While it has been assumed that this protection is due to maternally-acquired passive immunity, active mechanisms in neonates may also protect against invasive infection by these ‘non-neonatal pathogens’. This epidemiological observation is at odds with the idea of hypertolerogenic responses thought to characterize neonatal immunity or a weak immune system, as seen in elderly or immunocompromised individuals. To delineate this dichotomy of susceptibility versus resistance in early life, a better understanding of immune mechanisms that characterize the neonatal period is needed. Much of our understanding of host defenses are founded on evidence relying solely on adult mouse models, which do not account for environmental and developmental challenges faced by neonates, or human cord blood leukocytes, which may not be representative of immune cells during immediate postnatal life².

This dichotomy between neonatal and ‘non-neonatal pathogens’ is demonstrated by the major pathogenic streptococci. Streptococcus agalactiae (GBS) is a leading cause of invasive neonatal infection, while Streptococcus pneumoniae (Spn) and Streptococcus pyogenes (GAS) invasive infections are uncommon during this period^3-7. In the case of Spn, rates of nasopharyngeal carriage peak during early childhood⁸, and therefore the rarity of invasive infection among neonates cannot be attributed to a lack of environmental exposure during early life^9,10. Spn carriage and invasive infection following colonization can be modeled in mice where maternal immunity from prior exposure is not a confounding factor. We have previously demonstrated that for a strain of Spn that causes invasive disease in adults¹¹, mice are protected during the first two weeks of life regardless of the postnatal age by which they were nasally challenged¹². Using Spn as a model ‘non-neonatal pathogen,’ we sought to investigate immune mechanisms that facilitate systemic resistance to pathogens during early life. We then extended our findings to other invasive streptococcal species. Herein, we show that dampened efferocytosis in early life alters neutrophil homeostasis to augment CD11b-mediated responses, whereby the extent of dependence on CD11b effector activity dictates invasive infection outcome during the neonatal versus non-neonatal period of life.

Results

Neonatal mice are resistant to systemic infection by Spn, a ‘non-neonatal pathogen’

To model age-dependent susceptibility to invasive infection by a ‘non-neonatal pathogen’, we adopted an intraperitoneal (IP) model of Spn infection to induce systemic challenge in neonatal (6 - 9 days of life) and adult (6 - 14 weeks of life) mice, with adjustment of dose based on relative weight of the animal. The survival rate of neonatal mice upon IP challenge was significantly higher than that of adult mice (Fig 1A), correlating with our previous observation of a protective window in neonates following intranasal challenge with an invasive Spn strain¹². At early time points of 24 and 32 hours post-Spn IP challenge, neonatal mice had significantly lower bacterial burden in the blood and spleen compared to adult mice (Fig 1B-C). To exclude the possibility that improved protection in neonates was solely attributed to adjusted doses based on body weight, we infected neonatal and juvenile (21 - 28 days of life) mice with an equivalent dose of Spn. Even when challenged with the same dose, neonates had higher survival rates than the larger juveniles (Fig 1A), and we continued to observe significantly lower bacterial burden in both blood (Fig 1D) and spleen (Fig 1E) of neonates compared to juveniles. Together, these results suggested that control of Spn dissemination is more efficient in neonatal (early in life) compared to juvenile and adult mice (later in life).

Figure 1: Neonates are resistant to systemic infection by Streptococcus pneumoniae, a ‘non-neonatal pathogen’.

[A] Survival of WT neonatal (NNT, solid circle black line, n = 26), juvenile (JUV, half solid circle grey spotted line, n = 25) and adult mice (AD, open circle grey line, n = 18) following intraperitoneal (IP) challenge with P2406, an Spn strain with invasive potential in humans adapted in mice, at 5 x 10¹ – 1 x 10² colony forming units (CFU) per gram of body weight.

[B] Recoverable live Spn in blood of NNT and AD to assess systemic dissemination at two early time points, 24 hours (n = 16 for NNT, n = 11 for AD) and 32 hours (n = 19 for NNT, n = 13 for AD) post-infection, following IP challenge with Spn at 5 x 10¹ – 1 x 10² CFU per gram of body weight.

[C] Similar as described in [B], but for spleen at 24 hours (n = 16 for NNT, n = 11 for AD) and 32 hours post-infection (n = 17 for NNT, n = 11 for AD); BW, body weight.

[D] Recoverable live Spn in blood of NNT (n = 10) and JUV (n = 13) to assess systemic dissemination at 32 hours post-infection, following IP challenge with Spn at 5 x 10¹ – 1 x 10² CFU per gram of neonatal body weight.

[E] Similar as described in [D] but for spleen (n = 10 for NNT, n = 13 for JUV).

[F] Blood neutrophil counts from mice treated with an IgG isotype control clone 2A3 (n = 6) or neutrophil-specific depletion antibody clone 1A8 (n = 6) following IP challenge with Spn at 5 x 10¹ – 1 x 10² CFU per gram body weight at 32 hours post-infection.

[G] Recoverable live bacteria in spleen from mice described in [F].

[H] Blood monocyte counts from mice that were either wild-type, Ccr2^+/+ (n = 15), or CCR2-deficient, Ccr2^−/− (n = 9), following IP challenge with Spn at 5 x 10¹ – 1 x 10² CFU per gram body weight at 32 hours post-infection.

[I] Recoverable live bacteria in spleen from mice described in [H]. Each data point represents an individual mouse, and data are representative of experiments repeated at least twice. Data with error bars are presented as mean ± SEM. N.S., not significant.

Neutrophils facilitate resistance against invasive Spn infection in neonatal mice

Next, we asked what cellular mediator is directly responsible for the improved protection in neonates. The quick onset of protection in neonates as early as 24 hours post-IP challenge, combined with absence of pre-existing Spn-specific antibodies (as dams were not exposed to Spn), suggested the importance of innate, rather than adaptive, immunity in this process. We turned our attention to phagocytic cells, specifically neutrophils, given their established importance as the initial line of defense against invading extracellular microbes, including Spn^13,14. While the role of neutrophils in controlling Spn dissemination is established in non-neonatal stages of life^15,16, we questioned if this holds true in the neonatal period. We tested this hypothesis by depleting neutrophils in neonates with the 1A8-clone monoclonal antibody (which specifically recognize Ly6G, a pan-neutrophil marker), followed by IP challenge with Spn. In the 1A8-treated neonates, an approximate 50% reduction in blood neutrophil numbers was associated with a significant increase in splenic bacterial load compared to isotype control, 2A3-treated neonates (Fig 1F-G), supporting a role for neutrophils in conferring immune protection against Spn in early life. To further ascertain the role of neutrophils as the primary cellular mediator of this protection, we infected neonatal mice that were deficient in CCR2, which have impaired monocyte trafficking¹⁷. As expected, monocyte counts were lower in Ccr2^−/− neonates compared to wild-type (Ccr2^+/+) neonates (Fig 1H). We did not observe differences in bacterial burden between Ccr2^−/− and Ccr2^+/+ neonates (Fig 1I), indicating that neutrophils were the primary circulating myeloid cellular mediator conferring protection against Spn during the neonatal period.

Increased CD11b expression on early life neutrophils drives systemic immunity against Spn

Considering the importance of neutrophils in controlling bacterial infection in both neonatal and non-neonatal stages of life, we sought to distinguish the features of neonatal neutrophils that accounted for improved protection in early life. We utilized flow cytometry to phenotype neutrophils in the bone marrow (BM) of mice at different ages. Gating strategy is listed in Fig S1A.

We found that surface expression of CD11b on BM neutrophils was higher in neonatal mice, and exhibited an age-dependent decay (Fig 2A). Higher CD11b surface expression was also observed on Ly6C-expressing neonatal monocytes (Fig S1B). In juvenile mice, surface expression of CD11b on myeloid cells were similar to that of adults (Fig 2A & Fig S1B). This correlated with the increased susceptibility to Spn systemic infection in juveniles compared to neonates, despite receiving similar infectious doses. We did not observe differences in CD11b mRNA transcript amounts in neutrophils from neonatal versus adult BM, indicating that increased CD11b surface expression on neonatal neutrophils is driven by post-transcriptional mechanisms (Fig 2B). The early life increase in CD11b on neutrophils did not reflect a transient phase in development as neutrophils from peripheral blood of neonatal mice also exhibited a similar phenotype (Fig 2C). Our observations in mice were recapitulated in humans. Mass cytometry analysis of peripheral blood neutrophils from human neonates during the first week of life (1 – 5 days of life) also exhibited higher CD11b surface expression than those from the immediate post-neonatal period (31 – 80 days of life), although this was not observed in monocytes (Fig 2D & Fig S1C). These results underscore in both mice and humans that surface expression of CD11b is increased on neutrophils during neonatal life.

Figure 2: Increased CD11b expression in early life enhances neutrophil function and drives systemic immunity against Spn.

[A] Flow cytometry analysis of CD11b surface expression via relative median fluorescence intensity (MFI) on bone marrow neutrophils (gated on Singlets⁺ CD45⁺ CD11b⁺ Ly6G⁺ Ly6C^mid) from mice at different ages [3 days old (3d), n = 4; 6 days old (6d), n = 5; 9 days old (9d), n = 5; JUV, n = 5; AD, n= 10].

[B] RT-PCR analysis of CD11b (Itgam) mRNA transcript in NNT (n = 15) versus AD (n = 11) bone marrow neutrophils isolated via density centrifugation.

[C] MFI of CD11b on peripheral blood neutrophils from NNT (n = 10) versus AD mice (n = 7).

[D] CD11b marker expression of neutrophils from mass cytometry experiment comparing newborns aged 0 – 5 days (n = 75) and 31 – 81 days after birth (n = 31), mean expression from binned days used for coloring.

[E] Opsonophagocytic uptake of mouse sera opsonized GFP-fluorescent Spn (P2531) by neutrophils from bone marrow suspensions of NNT versus AD mice at multiplicity of infection (MOI) of 25 (n = 8 for both groups).

[F] MFI of GFP on GFP⁺ neutrophils from [E] (n = 8 for both groups).

[G] Recoverable live bacteria in blood (left) and spleen (right) of WT (Itgam^+/+) and CD11b-deficient (Itgam^−/−) NNT mice challenged IP with Spn at 32 hours post-infection (n = 8 for Itgam^+/+ n = 10 for Itgam^−/−).

[H] MFI of CD11b on neutrophils from Itgam^+/+ NNT (n = 3), CD11b-heterozygous (Itgam^+/−) NNT (n = 6), and Itgam^+/+ AD (n = 6). Itgam^−/− AD is included as negative control (n = 2).

[I] Survival of Itgam^+/+ NNT (solid circle black line; n = 20), Itgam^+/− NNT (solid circle orange line; n = 17), Itgam^−/− NNT (solid circle green line; n = 17) and Itgam^−/− AD (open circle muted green line; n = 12) mice challenged IP with Spn at 5 x 10¹ – 1 x 10² CFU/g body weight.

[J] Neutrophil counts in peripheral blood from Itgam^+/+ (n = 10), Itgam^+/− (n = 10) and Itgam^−/− (n = 9) NNT mice.

[K] Survival of Itgam^+/+(solid circle dotted black line, n = 9) and Itgam^−/− (solid circle dotted green line, n = 9) NNT mice treated with 25μg cobra venom factor at 24 hours prior infection, followed by IP challenge with Spn at 5 x 10¹ – 1 x 10² CFU/g body weight.

[L] Survival of WT NNT (solid circle black line; n = 18) versus AD mice (open circle grey line; n = 20), and both groups of mice treated with 25μg cobra venom factor (WT NNT, solid circle dotted black line, n = 9; WT AD, open circle dotted grey line, n = 11) following IP challenge with complement sensitive capsule switch mutant, P2453, at 5 x 10¹ – 1 x 10² CFU/g body weight.

[M] Recoverable live bacteria in blood of Itgam^−/− neonate recipients receiving either Itgam^+/+ neonatal neutrophils (solid circle, n = 9), Itgam^+/− neonatal neutrophils (solid orange circles, n = 8) or Itgam^−/− neonatal neutrophils (solid green circle, n = 8) relative to that of recipients receiving Itgam^+/+ adult neutrophils (open circle; n = 9, n = 9, n = 7, respectively) at 20 hours post infection with Spn via IP. Each data point represents a biological replicate (refer to STAR Methods for number of mice pooled), and data are representative of experiments repeated at least twice. Data with error bars are presented as mean ± SEM. N.S., not significant. See also Figure S1.

Given the importance of the microbiota in regulating immune system development, we postulated that microbial cues in early life enhance CD11b surface expression on neonatal neutrophils. Contrary to our prediction, CD11b expression was similar between BM neutrophils from neonatal mice housed in germ-free and specific pathogen free conditions (Fig S1D). This suggests that the increase in CD11b expression on neonatal neutrophils was driven by developmentally-intrinsic mechanisms that are independent of microbial cues.

CD11b forms part of a heterodimeric β₂ integrin complex that recognizes complement protein C3 deposited on bacterial surfaces, leading to neutrophil phagocytosis¹⁸; in humans, recognition of C3-tagged Spn by CD11b on neutrophils is important for control of Spn^19-21. We hypothesized that increased CD11b surface expression on neonatal neutrophils enhances C3-dependent phagocytic capacity. To test this, we utilized an ex vivo phagocytosis assay using GFP-expressing Spn opsonized with mouse sera, and quantified bacterial uptake using flow cytometry. We observed that co-localization of neutrophils with a GFP signal was significantly higher in neonatal over adult neutrophils (Fig 2E). This difference was abrogated when Spn was incubated with heat-inactivated sera, in which complement proteins are degraded, supporting a critical role for C3-CD11b interactions in this process. For neutrophils with a GFP signal (co-localized with Spn), median fluorescence intensity (MFI) of GFP did not differ between adult and neonatal mouse neutrophils, suggesting that there was no difference in events following surface binding of Spn (Fig 2F). These trends were similarly observed with mouse monocytes (Fig S1E-F). In contrast to neutrophil CD11b expression, sera from neonates and adults deposited similar amounts of C3 on the surface of Spn (Fig S1G). Combined, these observations demonstrate that on the population level, more neonatal neutrophils are able to take up Spn, an observation that correlates with increased CD11b expression on early life myeloid cells.

Next, we tested whether the age-dependent increase in CD11b drives systemic immune resistance to Spn in early life in vivo. We first confirmed that CD11b-deficient (Itgam^−/−) neonates had significantly higher bacterial burden than wild-type (Itgam^+/+) neonates at 32 hours post-IP challenge with Spn (Fig 2G). To modulate expression of CD11b on neutrophils, we generated heterozygous CD11b (Itgam^+/−) mice. Surface expression of CD11b on neutrophils from Itgam^+/− neonates significantly decreased compared to Itgam^+/+ neonates and phenocopied that of Itgam^+/+ adults (Fig 2H). As predicted, upon IP Spn challenge both Itgam^−/− and Itgam^+/− neonates were more susceptible compared to WT neonates (Fig 2I). While Itgam^+/− mice appear to have a survival advantage over the Itgam^−/− mice during the first 72 hours post-infection, these mice ultimately succumb to the infection similarly. We also did not observe a difference in survival between Itgam^−/− neonatal and Itgam^−/− adult mice (Fig 2I). The effect of CD11b surface expression on bacterial density and neonate survival was not attributable to differences in neutrophil numbers (Fig 2J).

To confirm the protective effect of CD11b is dependent on interactions with complement C3, we first showed that Itgam^+/+ and Itgam^−/− neonates treated with cobra venom factor to deplete C3 were equally susceptible to Spn (Fig 2K). The polysaccharide capsule of Spn inhibits the deposition of C3 on the bacterial surface^22,23. To further resolve C3-dependent CD11b phagocytic interactions as the critical mechanism driving age-dependent differences in resistance to Spn, we performed infection experiments using a serotype switch Spn mutant strain, whereby the locus for capsule expression was switched while preserving the genomic background of the strain. To this end, we switched the native serotype 4 capsule for the serotype 6A capsule, the latter of which has been demonstrated to have higher deposition of C3 on its surface²⁴. We predicted that the increase in C3 deposition on surface of Spn overcomes the CD11b-C3 interactions bottleneck imposed during infection in adults, and thus, result in improved protection. In support of this hypothesis, we found that neonates and adults now become equally protected following infection with this capsule switch Spn strain (Fig 2L). This protection in both neonates and adults is lost following cobra venom factor pre-treatment to deplete C3 (Fig 2L).

Finally, we demonstrated that elevated CD11b expression on neutrophils specifically drove resistance to Spn in vivo during the neonatal period. To do so, we performed experiments whereby neutrophils isolated from either neonatal or adult donor bone marrow were adoptively transferred into Itgam^−/− neonate recipients infected with Spn. Transfer of neutrophils from Itgam^+/+ neonate donors resulted in improved bacterial control compared to that of Itgam^+/+ adult donors, but not recipients receiving neutrophils from either Itgam^+/− neonates or Itgam^−/− neonates (Fig 2M & Fig S1H). As predicted, recipients receiving Itgam^+/− neonatal neutrophils had similar bacterial burden to those receiving Itgam^+/+ adult neutrophils; this confirms that other age-dependent but CD11b-independent phenotypic differences in neutrophils are not responsible for improved neonatal protection (Fig 2M & Fig S1H). Collectively, these in vivo observations illustrate the protective benefit against Spn of enhanced neutrophil CD11b surface expression in neonates.

Circulating neutrophils in neonatal mice bias towards an ‘aged’ phenotype

Similar to BM neutrophils, we also observed increased CD11b surface expression on early life blood neutrophils compared to those from later in life (Fig 2C). This suggested that increased CD11b expression did not merely reflect a transient phase in development within the BM and that circulating neutrophils are capable of enhanced CD11b-dependent effector function. This observation prompted us to ask about other characteristics of circulating neutrophils in early life.

There is increasing appreciation that neutrophils can specialize and adapt to different environments, a feature that can lead to generation of functional diversity. For example, neutrophils which have circulated in blood extensively and nearing the end of their lifespan, known as ‘aged’ neutrophils, have increased pro-inflammatory potential^25,26. ‘Aged’ neutrophils are canonically marked by gradual loss of CD62L expression, which limits their migration into lymphatics, and a concomitant increase in CXCR4 expression, which promotes their migration into CXCL12-expressing tissues. These features help ‘aged’ neutrophils infiltrate tissues for degradation²⁷. Since increased expression of CD11b is a characteristic feature of ‘aged’ neutrophils, we determined whether neonatal blood neutrophils have an ‘aged’ phenotype.

Expression of CD62L was significantly decreased on neonatal compared to adult mouse peripheral neutrophils (Fig 3A). By gating for CD62L^lo (‘aged’) versus CD62L^hi (‘non-aged’) neutrophils, we found that neonates had a significantly higher proportion of CD62L^lo neutrophils compared to adults (Fig 3B). As expected, expression of CXCR4 and CD11b were higher in the CD62L^lo compared to the CD62L^hi population of neonatal neutrophils (Fig 3C-D). Other known phenotypic features of ‘aged’ neutrophils were observed in the CD62L^lo population in neonates, including higher CD45 expression and lower Ly6G expression (Fig 3E-F), as well as lower side and forward scatters, suggesting reduced granularity and cell size, respectively (Fig 3G-H). As additional validation of our gating scheme, we observed a similar expression profile in the CD62L^lo versus CD62L^hi populations of adult circulating neutrophils (Fig S2A-G). These results demonstrate that a higher proportion of ‘aged’ neutrophils exist in neonatal blood.

Figure 3: Circulating neutrophils in neonatal mice bias towards an ‘aged’ phenotype.

[A] MFI of CD62L on peripheral blood neutrophils from NNT (n = 10) versus AD mice (n = 7).

[B] Frequency of CD62L^lo ‘aged’ neutrophils from peripheral blood of NNT (n = 10) versus AD mice (n = 7) as proportion of all hematopoietic cells [Live, CD45⁺ cells] (left) and of all neutrophils [Ly6G⁺ cells] (right); FMO, fluorescence-minus-one control.

[C-H] MFI analysis of known markers between CD62L^lo ‘aged’ (red dots) and CD62L^hi ‘non-aged’ (blue dots) neutrophils in neonatal blood (n = 10). Each data point represents a biological replicate (refer to STAR Methods for number of mice pooled), and data are representative of experiments repeated three times. Data with error bars are presented as mean ± SEM. N.S., not significant. See also Figure S2.

Impaired efferocytosis in early life increases ‘aged’ neutrophil numbers in blood

Recently, a mechanism involving myeloperoxidase and T-cell death in influencing neutrophil lifespan has been described²⁸. However, these parameters do not exhibit an age-dependent component (Fig S3A-B), suggesting this mechanism does not explain the ‘aged’ phenotype bias during neonatal life. Further, mediators of known neutrophil-intrinsic pathways, CXCR2- and CXCR4-dependent signaling²⁹, involved in the ‘aging’ program did not appear to exhibit age-dependent differences (Fig S3C-D). Expression of CD11b itself also did not determine the proportion of ‘aged’ neutrophils present in blood (Fig S3E).

In addition to phenotypic differences of peripheral blood neutrophils between neonates and adults, we also observed that neonates have higher numbers of neutrophils in blood (Fig 4A). This raised the question of whether mechanisms to remove ‘aged’ neutrophils are impaired in neonates. Degradation of ‘aged’ cells, or efferocytosis, is mediated by macrophages including those of hematopoietic origin that express CD169³⁰. While a previous study noted the absence of CD169⁺ macrophages in neonatal spleens³¹, the functional consequences of this observation on neutrophil homeostasis in early life was not explored.

Figure 4: Impaired efferocytosis in early life increases ‘aged’ neutrophil numbers in blood.

[A] Neutrophil counts in peripheral blood from NNT (6d old, n = 7; 9d old, n = 10), JUV (n = 6) and AD (n = 11) mice.

[B-F] Transcriptional analysis of whole spleen from naïve NNT versus AD for [B] Cd169, [C] Lxra, [D] Mertk, [E] Abca1 and [F] Gas6. Expression was calculated as fold change in NNT relative to AD (n = 5 for both groups).

[G] Confocal images of naive NNT (n = 5) and AD (n = 4) lungs immunostained for CD169 (blue), MerTK (red), F4/80 (yellow), Ly6G (green) and CD3 (cyan). Scale bar is at 50μm. Quantification of densitometry values are adjacent to the images. Each data point represents an individual mouse, and data are representative of experiments repeated at least twice. Data with error bars are presented as mean ± SEM. N.S., not significant. See also Figure S1, S3-S4.

Besides the spleen³², there are other sites for efferocytosis of’aged’ neutrophils, such as the lungs³³ and bone marrow²⁷. Because expression of Cd169 is restricted to macrophages in mice³⁴, we assessed tissue expression of Cd169 in spleens, lungs and BM from neonate and adult mice. We found significant downregulation of Cd169 across all three neonatal tissues, which suggests decreased numbers of CD169⁺ macrophages at these sites in the neonatal period (Fig 4B & Fig S4A). Expression of Lxra, which regulates development of CD169⁺ macrophages in the spleen³¹ and facilitates efferocytosis^32,35, was similarly downregulated across neonatal tissues (Fig 4C & Fig S4B). This trend was also observed with genes both facilitating and induced upon efferocytosis, Mertk^32,35,36 and Abca1^37-38 (Fig 4D-E & Fig S4C-D). Expression of Gas6, a bridging molecule which facilitates recognition of phosphatidylserine on ‘aged’ cells by MerTK, induced during efferocytosis through LXRα-independent pathways^32,39 was decreased in lungs and spleens of neonates, but not BM (Fig 4F & Fig S4E). Collectively, our transcriptional analyses demonstrate that expression of mediators driving efferocytosis is dampened in the neonatal period. We also performed confocal imaging of immunostained spleen sections. Firstly, we observed significantly decreased CD169 staining in neonatal compared to adult spleens (Fig 4G). This validated our transcriptional analysis and supports a reduced density of CD169⁺ macrophages in neonates. We also found that Ly6C-expressing monocytes, which can seed tissues to develop into CD169⁺ macrophages³¹, were increased in blood circulation and within the spleen (Fig S1I-H). This suggests that absence of CD169⁺ macrophages was not due to defective seeding by progenitors; rather, the developmental signals (i.e. induction of LXRα-signaling pathways) to promote their functional maturation were lacking (Fig 4C). Moreover, LXRα expression in the spleen has been shown to increase over the first weeks of life³¹. We also observed that the defined localization of CD169 staining in the marginal zone of adult spleens was not observed in neonatal spleens. This disorganized architecture in conjunction with decreased numbers of CD169⁺ macrophages in neonatal spleens potentially contribute to impaired efferocytosis. Secondly, while there was no difference in staining of macrophages by F4/80 in neonatal versus adult spleens, neonatal spleens exhibited significantly reduced MerTK co-staining. Even though numbers of F4/80⁺ macrophages may be similar to that in adults, this suggests that other tissue resident non-CD169 expressing macrophages in early life, such as those in the red pulp, may also exhibit a defect in efferocytosis. Combined, these observations support a dampening of efferocytic processes in early life, which contributes to increased ‘aged’ neutrophil numbers in blood of neonates.

Impairing efferocytosis later in life confers systemic protection against Spn dependent on CD11b-effector ability

To test if we could improve protection against invasive Spn infection in adult mice, we utilized CD169-diphteria toxin receptor (DTR) mice to transiently impair efferocytosis³⁰, and predicted this would increase the generation of ‘aged’ neutrophils. First, we confirmed loss of CD169 staining cells in spleens of diphtheria toxin (DT)-treated CD169-DTR mice (Fig S5A). In line with previous work²⁵, we found that DT-treated CD169-DTR mice had higher proportions and absolute counts of ‘aged’ neutrophils than their WT counterparts (Fig 5A-B). Importantly, these ‘aged’ neutrophils (CD62L^lo) in the DT-treated CD169-DTR mice also exhibited increased CD11b expression over those of their ‘non-aged’ (CD62L^hi) counterparts (Fig 5C). We also observed that neutrophils from the BM of DT-treated CD169-DTR mice had higher CD11b surface expression than those from WT mice (Fig 5D). This was similarly observed with Ly6C-expressing monocytes (Fig S5B-C).

Figure 5: Impairing efferocytosis later in life protects against Spn in a CD11b-effector activity dependent manner.

[A] Frequency of CD62L^lo ‘aged’ neutrophils from peripheral blood of WT versus CD169-DTR adult mice per all Live, CD45⁺ cells (n = 9 for both groups).

[B] Absolute counts of CD62L^lo ‘aged’ neutrophils in peripheral blood of WT versus CD169-DTR adult mice (n = 9 for both groups).

[C] MFI of CD11b surface expression on CD62L^lo ‘aged’ and CD62L^hi ‘non-aged’ neutrophils in blood of WT versus CD169-DTR adults (n = 9 for both groups).

[D] MFI of CD11b surface expression on neutrophils in bone marrow of WT (n = 11) versus CD169-DTR adults (n = 8).

[E] Recoverable live Spn in blood (left) or spleen (right) of WT and CD169-DTR adults to assess systemic dissemination at 32 hpi (n = 9 for both groups) following IP Spn challenge.

[F] Survival of WT (open circle grey line, n = 12), CD169-DTR (open circle pink line, n = 13) and both groups of mice treated with 25μg of cobra venom factor (WT, open circle dotted grey line, n = 9; CD169-DTR, open circle dotted pink line, n = 9) following IP Spn challenge.

[G] Frequency of CD62L^lo ‘aged’ neutrophils from peripheral blood of Mertk^fl/fl (n = 6) versus Lyz2^Cre/+ Mertk^fl/fl (n = 9) adult mice per all Live, CD45⁺ cells.

^[H] Recoverable life Spn in blood (left) or spleen (right) of Mertk^fl/fl (n = 4) versus Lyz2^Cre/+ Mertk^fl/fl (n = 9) adult mice to assess systemic dissemination at 32hpi following IP Spn challenge. Each data point represents an individual mouse, and data are representative of experiments repeated at least twice. Data with error bars are presented as mean ± SEM. N.S., not significant. See also Figure S5-S6.

Following IP challenge of Spn in DT-treated CD169-DTR and WT adult mice, CD169-DTR adults were more protected than WT adults. CD169-DTR adults had significantly lower bacterial burden in both blood and spleen at 32 hours post infection (Fig 5E), and improved survival over WT adults (Fig 5F). Upon depletion of C3 using cobra venom factor, we found that the protective benefit conferred by depleting CD169⁺ macrophages was abrogated, and both WT and CD169-DTR mice were equally susceptible. This indicated that CD11b recognition of C3 by myeloid cells, such as neutrophils, was crucial for improved protection in CD169-DTR mice. To eliminate the possibility that the CD169 receptor itself was responsible for differences observed between WT and CD169-DTR adults, we infected Cd169^−/− adults (lacks the CD169 molecule but retains this subset of macrophages) and compared them to WT adults. We observed that Cd169^−/− mice were similarly susceptible to WT mice, and had similar circulating myeloid cell counts (Fig S5D-H).

MerTK is a receptor which facilitates macrophage efferocytosis of ‘aged’ neutrophils³², of which we observed reduced expression during neonatal life (Fig 4D, 4G & S4C). As additional validation on the role of efferocytosis in augmenting CD11b-driven immunity, we generated mice with floxed Mertk alleles expressing Cre recombinase driven by the Lyz2 promoter (Lyz2^Cre/+ Mertk^fl/fl) to genetically ablate MerTK from lysozyme-expressing macrophages. Despite (1) targeting only one of many receptors that facilitate efferocytosis and (2) an incomplete deficiency of MerTK from all macrophages in Lyz2^Cre/+ Mertk^fl/fl adult mice (Fig S6), we still observed an increase in CD11b^hi ‘aged’ neutrophils in these mice over Cre-negative (Mertk^fl/fl) adult mice (Fig 5G) and improved systemic control of Spn (Fig 5H). Thus, our genetic and pharmacologic approaches to impair macrophage-dependent efferocytosis demonstrate its role in modulating CD11b-driven immunity and consequently, generation of protective immunity against Spn.

Dependence on CD11b-driven immunity predicts streptococcal infection outcome in early life

Our data suggested that neonates adopt enhanced CD11b-dependent immunity as a strategy for systemic protection against potential pathogens. We infected neonatal and adult mice with Streptococcus pyogenes (GAS), another ‘non-neonatal pathogen’. Following IP challenge with GAS, neonatal mice were more protected compared to adult mice, and protection was CD11b dependent (Fig 6A). As predicted, the protection seen in Itgam^+/+ neonates over the Itgam^−/− neonates was abrogated if C3 is depleted using cobra venom factor prior infection with GAS (Fig 6A), supporting the importance of CD11b-C3 interactions as shown with Spn. This suggested our model where enhanced CD11b functionality drives early life microbial immunity was relevant to a broader array of ‘non-neonatal pathogens’. In contrast to the streptococcal species examined above, Streptococcus agalactiae (GBS) is the leading cause of invasive bacterial infection in neonates. We postulated that systemic defense against GBS is independent of CD11b-mediated responses and that might in part explain increased susceptibility to GBS in neonates, as mechanisms in place to enhance CD11b-mediated responses are no longer advantageous to the host. To test this, we infected neonatal and adult mice with GBS via an IP route. In contrast to our observations with Spn and GAS, we found neonates to be significantly more susceptible to GBS compared to adults. Importantly, in adults, the more protected group over neonates, Itgam^−/− adult mice were equally protected as Itgam^+/+ counterparts (Fig 6B).

Figure 6: Dependence on CD11b-driven immunity predicts streptococcal infection outcome in early life.

[A] Survival of WT NNT (solid circle black line, n = 17), Itgam^−/− NNT (solid circle green line, n = 10) and both groups of mice treated with 25μg of cobra venom factor (WT NNT, solid circle dotted black line, n = 12; Itgam^−/− NNT, solid circle dotted green line, n = 11) and WT AD mice (open circle grey line, n = 11) following IP challenge with GAS5, a Streptococccus pyogenes strain with necrotizing fasciitis potential in humans, at 5 – 50 CFU per gram of body weight.

[B] Survival of WT NNT (solid circle black line, n = 13), WT AD (open circle grey line, n = 7) and Itgam^−/− (open circle muted green line, n = 8) mice following IP challenge with GBS5, a Streptococcus agalactiae strain at 3 x 10⁵ CFU per gram of body weight. Each data point represents an individual mouse, and data are representative of experiments repeated at least twice. N.S., not significant.

Discussion

Microbial colonization of neonates begins immediately after exit from the sterile environment of the womb. To accommodate this transition, the neonatal immune system is typically thought of as hypertolerant. Many have proposed that both strict and opportunistic pathogens exploit this feature to cause disease. However, this hypothesis fails to explain the lack of association between numerous common or leading adult pathogens and neonatal sepsis, which implies that mechanisms exist in early life to protect against such ‘non-neonatal pathogens,’ such as Spn.

In the case of Spn, opsonophagocytosis driven by complement component C3–integrin CD11b interactions on myeloid cells, such as neutrophils, plays a pivotal role in host immunity and is often the limiting factor in generating a protective response^40,41. Since neonatal mice were more resistant than adult mice in our systemic Spn infection model, we hypothesized that this aspect of immune function might be augmented during early life. By profiling cells in the bone marrow and blood, we discovered that neonatal neutrophils have enhanced CD11b-dependent opsonophagocytic function. Our data shows that impaired efferocytosis in early life, as mediated by CD169⁺ macrophages, increased relative CD11b surface expression on neutrophils during development in BM and promoted retention of CD11b^hi expressing ‘aged’ neutrophils in blood. Consequently, these drove a CD11b-dependent systemic immune advantage against Spn in neonates. We found this advantage to be lost by weaning age (~21 days of life), as juvenile mice were more susceptible to Spn than neonatal mice. This correlated with the observation where CD11b surface expression on myeloid cells from juvenile mice phenocopied those of adult mice, and CD169⁺ macrophages have populated tissues such as the spleen³¹. The importance of this augmented CD11b-dependent immunity was demonstrated in our experiments where partially impairing efferocytosis in DT-treated CD169-DTR adult mice rescued protection against Spn in a C3-dependent manner.

Our results also provide a mechanistic explanation for age-dependent biases underscoring etiology of invasive streptococcal infections. We extrapolated on our observations with Spn and speculated that augmented CD11b immunity in neonates improves protection against microbes requiring CD11b for efficient control, but not against microbes that did not. We infected mice with GAS, which is rarely associated with human neonatal sepsis, and showed that neonatal mice are more protected over adult mice in a CD11b-dependent manner, similar to Spn. However, with GBS, a leading cause of human neonatal sepsis, we found neonatal mice to be more susceptible compared to adult mice. Contrary to Spn and GAS, we showed protection against GBS systemic infection was independent of CD11b, suggesting the augmented CD11b responses in neonates was inessential during systemic GBS infection. Protection in the case of GBS could instead require direct activity of the siglec CD169, which recognizes sialic acid⁴²; sialic acid is a surface feature of the GBS polysaccharide capsule⁴³. The age-dependent developmental delay in CD169 expression was thus disadvantageous for immunity against GBS, as previously proposed^44,45. We expanded on this model and propose that reduced CD169⁺ macrophages in early life is not wholly detrimental, but akin to a double-edged sword. Decreased CD169 expression would impair control of sialic acid coated bacteria, such as GBS. On the other hand, decreased efferocytosis enhances CD11b-effector ability, and improves bactericidal control against Spn and GAS, which both lack cell surface sialic acid⁴⁶. It is tempting to speculate that dependence on CD11b-versus-CD169 functionality can also predict susceptibility to other non-streptococcal microbes during early life. Another prominent invasive neonatal pathogen, K1-antigenic Escherichia coli (K1EC), is also covered by a sialic acid capsule⁴⁶. CD169⁺ macrophages have also been implicated in the pathogenesis of Listeria monocytogenes (LM)⁴⁷. Reduced numbers of macrophages expressing CD169 could therefore account for impaired control of the three major neonatal invasive pathogens (GBS, K1EC and LM).

Previous studies on neonatal immunity depend heavily on human cord blood as a source of immune cells. An advance in our study is the ability to phenotype myeloid cells in neonatal mice, which required obtaining bone marrow from mice as young as three, and blood at six, days of age. Our findings in mice starkly contrast those using human cord blood neutrophils. In the latter, CD11b expression and function are shown to be reduced compared to blood neutrophils from later stages of life, and proposed as the basis for increased susceptibility to infections in neonatal life^48-50. However, there are two caveats with cord blood studies. First, is their prenatal origin; human neonatal peripheral blood leukocytes are phenotypically distinct from cord blood leukocytes as early as one week into life, as demonstrated by a recent mass cytometry study². We overcame this first limitation by direct assessment of peripheral blood neutrophils from human infants during the neonatal versus immediate post-neonatal period. Second, cord blood studies have relied primarily on purified microbial ligands or heat killed microbes⁵¹, which oversimplify the complexity of pathogens in the setting of infection. Our use of (i) bone marrow and blood directly from neonatal mice and (ii) live bacteria for in vivo infections or ex vivo assays revealed aspects of neonatal immunity that may have been overlooked in human cord blood experiments. Further, because dams in our experiments were never infected or vaccinated, we were able to investigate immune mechanisms without the confounding effects of maternal specific antibodies, which is difficult to accomplish with human studies. This allowed us to directly examine neonate-intrinsic molecular pathways, such as kinetics of myeloid development and consequences for microbial susceptibility.

We primarily focused on neutrophils, given their well-established role in immune control of gram-positive bacteria¹³. Neutrophils were historically thought to comprise a homogenous population due to their short lifespan and non-proliferative capacity, but there is increasing appreciation for their ability to undergo phenotypic diversification⁵². Few studies have investigated such diversity in the context of early life, and those that do have focused on myeloid-derived suppressor cells, or MDSCs⁵³. We expanded on this by showing that neutrophils from another source of phenotypic diversification, ‘aged’ neutrophils, were prevalent in neonatal blood. Paralleling observations in adult mice, ‘aged’ neutrophils in neonates also possessed higher CD11b expression than ‘non-aged’ neutrophils, which further sustained the CD11b-driven protective advantage generated during neutrophil development in the bone marrow. Increased CD11b expression was also observed on Ly6C-expressing monocytes in neonatal mice, although (i) this increase was not observed on human neonatal monocytes in peripheral blood and (ii) Ccr2^−/− and Ccr2^+/+ neonatal mice had similar control of bacterial burden. These two observations suggested an unlikely role for Ly6C-expressing monocytes in mediating the protective effect against ‘non-neonatal’ pathogens, such as Spn. Because our model of invasive streptococcal infection spans only a few days, a direct contribution of neonatal adaptive immunity is unlikely. This is in line with existing evidence suggesting that lymphocytes are dispensable for systemic microbial immunity in early life⁵⁴.

Molecular mechanisms driving increased CD11b surface expression on neonatal neutrophils as they develop in the bone marrow are likely multifaceted. Our data suggests downstream effectors actively inhibited by efferocytosis are involved in this process, as relief of their inhibition in early life would promote CD11b expression. Of note, one biological pathway inhibited by efferocytosis is the IL23–IL17–GCSF axis, which promotes granulopoeisis⁵⁵. Efferocytosis of ‘aged’ neutrophils is a biological readout for sufficient neutrophil output in peripheries, and informs the hematopoietic compartment to reduce generation of neutrophils. A fundamental step in relaying this information is inhibition of GCSF production by bone marrow stromal cells, which is otherwise constitutive. Because neonates exhibit dampened efferocytosis, we speculate that differences in GCSF concentrations (and thus tonic sensing of GCSF by neutrophil progenitors) in the bone marrow microenvironment may drive differential CD11b expression. The potential effect of GCSF on priming CD11b-dependent neutrophil function could be decoupled from its effect on neutrophil numbers, as Itgam^+/− neonates remained more susceptible to invasive Spn infection than Itgam^+/+ neonates despite similar peripheral neutrophil counts.

Environmental factors can also influence host myeloid development and function. We and others have shown that the microbiota can functionally prime innate immunity under steady state conditions in both neonates and adults, promoting immunity against infectious insults at both mucosal and systemic tissues^11,56-58. We saw no difference in CD11b expression on neutrophils from neonatal mice in our germ free versus conventional specific pathogen free facility, which excluded a role of the microbiota in priming CD11b expression and function. Other factors inherently different in early compared to later in life, such as diet, may also be at play. Neonates solely consume milk, which is high in dietary fat. Consumption of high fat diets in adults have been shown to dysregulate hematopoietic programs by promoting a myeloid over lymphoid shift in the bone marrow^59,60, and such mice exhibit monocytosis and neutrophilia compared to adult mice fed regular chow⁶¹. It is possible that altered amounts of lipid metabolism byproducts in early life can prime CD11b expression and function. This correlates with the diminished CD11b surface expression in neutrophils from juvenile mice, which have begun consumption of solid foods as they wean off maternal milk.

Taken altogether, our results establish a mechanistic link between early life myeloid development, innate immune function and age-related patterns of infection susceptibility. We elucidate a previously unappreciated facet of neonatal immunity where an early life developmental defect (efferocytosis by macrophages) can prime enhanced bactericidal function (CD11b-dependent opsonophagocytosis by neutrophils) to resist invasive infection by two major pathogenic streptococci (Streptococcus pneumoniae & Streptococcus pyogenes). This is an immunological strategy to optimize the odds of survival as a newborn exits the womb into an environment filled with opportunistic microbes.

Limitations of Study

As there are no experimental models of invasive infection in humans, our study relied exclusively on murine models. Correlating mouse age with that of humans is inherently difficult without a more complete understanding of the relevant developmental and environmental factors affected by age. Many other questions remain. Multiple reports have documented that efferocytosis is impaired in the geriatric population, where defects in clearance of ‘aged’ neutrophils have also been observed^62-64. The elderly are highly susceptible to invasive infection by ‘non-neonatal pathogens’, suggesting that phenotypic alterations to antimicrobial immunity that stem from impaired efferocytosis differ in old age compared to newborns. The effects of efferocytosis on infection susceptibility during old age were not addressed in our study. Finally, the stage of neutrophil development in early life where priming of CD11b surface expression occurs remains unknown.

STAR Methods

Resource Availability

Lead Contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Jeffrey N. Weiser (Jeffrey.Weiser@nyulangone.org).

Materials Availability

This study did not generate new unique reagents.

Data and Code Availability

This study did not generate new datasets or code.

Experimental Model and Subject Details

Ethics statement

This study was conducted according to guidelines outlined by the National Science Foundation Animal Welfare Requirements and the Public Health Service Policy on the Humane Care and Use of Laboratory Animals. The New York University Langone Health IACUC oversees the welfare, well-being, and proper care and use of all vertebrate animals.

Mice

Wild-type C57BL/6J mice obtained from Jackson Laboratories (Bar Harbor, ME), Itgam^−/−, Itgam^+/−, CD169-DTR, Cd169^−/− mice, and Lyz2-cre Mertk^fl/fl mice were bred and maintained in conventional animal facilities. For littermate breeders, Lyz2^Cre/+ Mertk^fl/fl mice were crossed with Mertk^fl/fl mice⁶⁵ to generate Cre+ and Cre− littermates. For non-littermate breeders, Lyz2^Cre/Cre Mertk^fl/fl mice were crossed with Mertk^fl/fl mice to generate Lyz2^Cre/+ Mertk^fl/fl progeny, with sex- and age-matched Mertk^fl/fl mice as controls. Littermates and non-littermates were used in experiments. Germ-free C57BL/6J mice were bred and maintained in flexible film isolators, and absence of microbes was confirmed via (1) aerobic culture in brain heart infusion, sabaraud and nutrient broth, and (2) qPCR for bacterial 16S and eukaryotic 18S ribosomal RNA through stool sampling on a monthly basis. Unless otherwise stated, neonatal mice used were between 6 – 9 days of life and housed with a dam throughout the duration of infection experiments. Juvenile mice were used between 3 – 4 weeks of life. Adult mice used were between 6 – 14 weeks of life. Both sexes of mice were used in experiments, except for GAS infection experiments, where only males were used⁶⁶. Mice for control groups were sex- and age-matched to the experimental groups. For experiments involving CD169-DTR mice, both CD169-DTR and WT mice were administered 40ng/g body weight of diphtheria toxin (DT), with a pulse three days later at 4ng/g of diphtheria toxin followed by euthanasia or Spn infection another three days after (as described below).

Invasive streptococcal infection models

For Spn, P2406 (serotype 4, strain TIGR4; clinical isolate with streptomycin resistance) was used in infection experiments⁶⁷, or P2453 (serotype 6, strain TIGR4; capsule switch mutant from parental strain P2406)⁶⁸. Spn was grown statically in tryptic soy (TS) broth at 37°C to mid-log phase (optical density of 1.0 at 620nm). Quantification of Spn bacterial counts was done via serial plating on TS agar plates supplemented with 100μL catalase (30,000 U/mL, Worthington Biochemical) and 200μg/mL streptomycin, incubated overnight at 37°C and 5% CO₂. For GAS, GAS5 (M-type 3, strain 950771; clinical isolate from a child with necrotizing fasciitis and sepsis) was used in infection experiments⁶⁹. GAS was grown statically in Todd Hewitt broth supplemented with yeast extract (THY) at 37°C to an optical density of 1.0 at 620nm at a 1:10 subculture from an overnight 10mL culture inoculated with a single colony of GAS streaked on a TS agar plate supplemented with 5% sheep’s blood. Quantification of GAS bacterial counts was done via serial plating on TS agar plates supplemented with 5% sheep’s blood, incubated overnight at 37°C and 5% CO₂. For GBS, GBS5 (serotype 3, COH-1; clinical isolate from newborn with sepsis) was used in infection experiments⁷⁰. GBS strain was grown statically overnight in tryptic soy broth at 37°C. Quantification of GBS was done via serial plating on TS agar plates, incubated overnight at 37°C and 5% CO₂.

For the intraperitoneal model of infection, neonatal, juvenile or adult mice were inoculated with 20uL, 40uL and 100uL of streptococci in Dulbecco’s phosphate buffered saline (PBS), respectively, at the following doses: Spn, 50-100 CFU/g body weight; GBS, 3 x 10⁵ CFU/g body weight; GAS, 5-50 CFU/g body weight. Unless otherwise stated, dosing of Spn was normalized to body weight. For experiments involving depletions, either 150μg of neutrophil-depletion 1A8 antibody, 150μg of the isotype 2A3 control, or 25μg of cobra venom factor was administered IP 24 hours prior infection. At time points indicated following challenge, mice were euthanized via CO₂ asphyxiation followed by cardiac puncture, and tissues harvested for appropriate analyses.

Method Details

Myeloid cell phenotyping & flow cytometry

For bone marrow cell phenotyping, bone marrow was collected from mice at postnatal day 3 (pooled tibia and femur from 3 - 4 mice per biological replicate) and day 6 (pooled tibia and femur from 2 −3 mice per biological replicate). Bones from mice at 9 days of age, juveniles and adults were representative of 1 mouse per biological replicate. Bone marrow was flushed out into 1x Hank’s Balanced Salt Solution (Corning). Bone marrow cell suspension was filtered using a 40μm strainer (Falcon), prior pelleting for other applications. For blood cell phenotyping, blood was obtained via cardiac puncture from mice at postnatal day 6 (pooled from 4 −7 mice per biological replicate), day 9 (pooled from 2 – 4 mice per biological replicate) and adult (1 mouse per biological replicate) mice.

Following lysis of red blood cells with ACK lysis buffer, single cell suspensions were pelleted. All incubations were performed at 4°C. Cells were re-suspended in LIVE/DEAD Fixable Aqua (ThermoFisher) at 1:1000 in PBS for 30 minutes for assessment of cell viability, followed by anti-CD16/CD32 at 1:200 in PBS + 1% bovine serum albumin (BSA) for 15 minutes for blocking of Fc receptors. Surface marker staining was performed for 30 minutes, with antibodies against anti-anti-CD45-APC/Cy7 (clone 30-F11, BD Biosciences), anti-CD11b-V450 (clone M1/70, BD Biosciences), anti-Ly6G-PerCP/Cy5.5 (clone 1A8, BD Biosciences) and (i) for bone marrow myeloid cells: anti-Ly6C-APC (clone HK1.4, Biolegend) or (ii) for blood neutrophils: anti-CXCR4-PE (clone L276F12, Biolegend) and anti-CD62L-APC (clone MEL-14, Biolegend) at 1:150 dilution in PBS + 1% BSA. Cells were then fixed in 4% paraformaldehyde in PBS for 15 minutes. All cytometry was performed on an LSRII flow cytometer (BD Biosciences) and data analysis was performed using FlowJo software (Tree Star). All median fluorescence intensity (MFI) data is normalized to values from adult mice for comparison across different ages, unless noted otherwise.

Complete blood count differentials

To enumerate white blood cells in peripheral blood, 20μL of blood was harvested through cardiac puncture of either neonatal (6 or 9 days of life), juvenile (3 – 4 weeks of life) or adult (6 – 12 weeks of life) mice. The blood was diluted in 80μL of cold PBS in an EDTA-coated microtainer tube (BD Biosystems), and samples were run on the Element HT5 Veterinary Hematology Analyzer (Heska) for data acquisition.

Neutrophil adoptive transfer

Bone marrow from either neonatal or adult mice were harvested and processed as described above. Resulting suspensions were then overlayed onto a Histopaque-1119 and Histopaque-1077 gradient and cells were spun at 2000rpm for 30 minutes at room temperature, with deceleration breaks turned off. Neutrophils were isolated from the interface of Histopaque-1119 and Histopaque-1077, washed in cold dPBS and resuspended to 5 x 10⁷ cells/mL. Assessment of the purity in the granulocyte fraction from adult and neonatal mice, respectively, using flow cytometry yielded > 85%. At 4hpi, Spn-infected Itgam^−/− recipient neonatal mice (infection as described above) were inoculated with 20uL of the suspension (~10⁶ neutrophils) via the intraperitoneal route. Recipient mice were euthanized at 20 hpi to assess bacterial burden in the blood and spleen.

RNA isolation and quantitative RT-PCR

For spleen, lung and liver, tissues were harvested from neonatal and adult mice, then snap frozen in dry ice and stored at −80°C. Frozen tissues (~100mg to 200mg) were homogenized in 1mL of Trizol, and standard phenol-chloroform extraction was performed. For bone marrow and isolated neutrophils, cells were lysed in 600μL RLT lysis buffer (Qiagen) supplemented with 1% β-mercaptoethanol, and RNA was extracted using an isolation kit (RNeasy, Qiagen) as per manufacturer protocols. cDNA was generated using High Capacity cDNA Reverse Transcriptase Kit (Applied Biosystems, Thermo Fisher Scientific) as per manufacturer protocols. Reaction samples contained ~10ng cDNA and 0.5μM primers in Power SYBR Green PCR Master Mix (Applied Biosystems, Thermo Fisher) and tested in duplicates. qRT-PCR reactions were run in a 384 well plate (Bio-Rad) using CFX384^™ Real-Time System (Bio-Rad). All genes were normalized to expression of Gapdh and fold changes in expression of relevant genes were quantified using the ΔΔCt method. Primer sequences are detailed in Table S1.

Tissue preparation for confocal microscopy

Tissues were processed as previously described^47,71. Briefly, tissues were flash frozen in OCT media at −80°C. 20-um frozen tissue sections were sectioned using Leica CM3050S cryostat. Tissue were fixed with ice cold acetone for 15 minutes in −20°C followed by washing with 1x PBS. FcR blocked with anti-CD16/32 Fc block antibody (Clone 93, Biolegend) diluted in PBS containing 2% goat serum, and 2% fetal bovine serum (FBS) for 1 hour at room temperature. Sections were stained with the antibodies anti-CD169-ef660 (Clone Ser-4, Invitrogen), anti-MerTK-PE (clone DS5MMer, Biolegend), anti-F4/80-BV605 (clone BM8, Biolegend), anti-Ly6G-488 (clone 1A8, Biolegend), and anti-CD3-BV421 (clone 145-2C11, Biolegend) that were diluted in PBS containing 2% goat serum, 2% FBS, and 0.05% Fc block for 1 hour at room temperature. Images were acquired using a Zeiss LSM 880 confocal microscope (Carl Zeiss) with the Zen Black software. The imaging data were processed and analyzed using Imaris software version 9.0.1 (Bitplane; Oxford Instruments).

Ex vivo Spn opsonophagocytic and complement deposition assay

For the opsonophagocytosis assay, bone marrow cells were counted by trypan blue staining and adjusted to a density of 6 x 10⁶ cells/mL in PBS. GFP-expressing Spn used in experiments, P2531, is a serotype 23, clinical isolate transformed with chromosomal DNA from GFP-expressing serotype 2 strain D39 (P2521 in our collection)⁷². Mid-log phase GFP-Spn was adjusted to a density of 1.5 x 10⁸ cells/mL in PBS. 20μL of Spn was incubated with 20μL fresh mouse sera (as complement source) for 45 minutes at 180 rpm at 37°C. Following that, 20μL of bone marrow cells were added to the opsonized-bacteria, and incubated for 20 minutes at 180 rpm at 37°C, then immediately incubated at 4°C to stop the reaction. Cells were stained as described above, and neutrophil uptake is assessed as % of neutrophils with GFP-positive signal. Heat-inactivated sera (to degrade complement) was generated by incubating fresh adult mouse sera for 30 minutes at 56°C.

For the complement deposition assay, 2.5 x 10⁶ Spn strain P1121⁷³ or P2406 in 25μL was incubated with 25μL mouse sera for 30 minutes at 37°C, followed by addition of 50μL of PBS in 20mM EDTA and immediate incubation at 4°C to stop the reaction. Spn was pelleted and then stained with 50μL of diluted FITC anti-mouse C3 antibody (MP Biomedicals) at 1:200 dilution for 30 minutes at 4°C. Spn was then fixed in 100μL of 4% paraformaldehyde in PBS for 20 minutes at room temperature, pelleted and resuspended for flow cytometry analysis.

Mass cytometry analysis of human peripheral blood neutrophils

Blood samples were drawn from newborns at the Karolinska University Hospital in Stockholm, Sweden. Samples were mixed with a stabilizer⁷⁴ within 3 hours and cryopreserved. Mass cytometry experimental methods are detailed in Henrick et al. manuscript⁷⁵.

All FCS-files were unrandomized using CyTOF software (version 6.0.626) and transferred with no further processing. Grid, an automated cell classification supervised algorithm, was used to first define reference subpopulations manually and then train a learning algorithm (XGBoost) to recognize the same subsets of cells in novel data⁷⁶. Outputs from all mass cytometry experiments were merged, and batch differences removed using limma⁷⁷ Immune cell data was reduced to include infant samples collected between 0 to 5 days and 31 to 81 days after birth, resulting in n = 75 and n = 31 samples respectively. The mean of each bin window was calculated and colored accordingly, and the package ggpubr was used to perform a Welsh T-test to compare both groups.

Quantification and Statistical Analysis

Differences in bacterial counts from infected mice were assessed using a non-parametric, Mann-Whitney test. Differences between two groups in median fluorescence intensity, percentages and absolute cell counts were analyzed using either an unpaired or paired Student’s t-test. Differences between three or more groups was analyzed using a one-way ANOVA test with Tukey’s multiple comparisons. Survival time course was compared between groups using the Kaplan-Meier’s log-rank test. All statistical analyses were performed on GraphPad Prism 8. All experiments were repeated at least twice. A p-value of < 0.05 is deemed statistically significant.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
APC/Cy7 Rat anti-Mouse CD45 (clone 30-F11)	BD Biosciences	Cat #: 557659; RRID: AB_396774
V450 Rat anti-Mouse CD11b (clone M1/70)	BD Biosciences	Cat #: 560455; RRID: AB_1645266
APC Rat anti-Mouse Ly6C (clone HK1.4)	Biolegend	Cat #: 128016; RRID: AB_1732076
PerCP/Cy5.5 Rat anti-Mouse Ly6G (clone 1A8)	BD Biosciences	Cat #: 560602; RRID: AB_1727563
APC Rat anti-Mouse CD62L (clone MEL-14)	Biolegend	Cat #: 104411; RRID: AB_313098
PE Rat anti-Mouse CD184/CXCR4 (clone L276F12)	Biolegend	Cat #: 146505; RRID: AB_2562782
Rat anti-Mouse CD16/32 (clone 93)	Biolegend	Cat #: 101302; RRID: AB_312801
ef660 Rat anti-Mouse CD169 (clone SER-4)	Invitrogen	Cat #: 50-5755-82; AB_2574241
PE Rat anti-Mouse MERTK (clone DS5MMER)	ThermoFisher Scientific	Cat #: 12-5751-82; RRID: AB_2572623
BV605 Rat anti-Mouse F4/80 (clone BM8)	Biolegend	Cat #: 123133; RRID: AB_2562305
AF488 Rat anti-Mouse Ly6G (clone 1A8)	Biolegend	Cat #: 127625; RRID: AB_2561340
BV421 Hamster anti-Mouse CD3 (clone 145-2C11)	Biolegend	Cat #: 100335; RRID: AB_10898314
PE Rat anti-mouse F4/80 (clone BM8)	Biolegend	Cat #: 123110 RRID: AB_893486
PerCP-eFluor 710 Rat anti-Mouse MERTK (clone DS5MMER)	Invitrogen	Cat #:46-5751-82; RRID: AB_2688094
PE Rat anti-Mouse CD169 (clone SER-4)	ThermoFisher Scientific	Cat #: 12-5755-82; RRID: AB_2572625
FITC Goat anti-Mouse Complement C3	MP Biomedicals, Inc	Cat #: 0855500; RRID: AB_2334931
Anti-Ly6G (clone 1A8)	BioXCell	Cat #: BE0075-1; RRID: AB_1107721
IgG2a control (clone 2A3)	BioXCell	Cat #: BE0089; RRID: AB_1107769
Bacterial and virus strains
Streptococcus pneumoniae Serotype 4 (TIGR4) Strain P2406	(Zafar et al., 2016)	N/A
Streptococcus pneumoniae GFP+ Serotype 23F Strain P2531	This paper	N/A
Streptococcus pneumoniae Serotype 23F Strain P1121	(McCool et al., 2002)	N/A
Streptococcus pneumoniae TIGR4 expressing Serotype 6A Capsule P2453	(Zafar et al., 2017)	N/A
Streptococcus pyogenes GAS5 (M-type 3, strain 950771)	(Ashbaugh et al., 1998)	N/A
Streptococcus agalactiae GBS5 (serotype 3, WT COH1)	(Wilson and Weaver, 1985)	N/A
Chemicals, peptides, and recombinant proteins
Diphteria toxin	Sigma Aldrich	Cat #: D0564
Catalase	Worthington Biochemical	Cat #: LS001896
Cobra Venom Factor, Naja naja kaouthia	Millipore Sigma	Cat #: 233552
Power SYBR Green PCR Master Mix	Applied Biosystems, ThermoFisher Scientific	Cat #: 4309155
ACK Lysing Buffer	Gibco	Cat #: A10492-01
LIVE/DEAD Fixable Aqua	ThermoFisher Scientific	Cat #: L34966
BD Microtainer Tubes w/ K2E	BD Biosystems	Cat #: 365974
Histopaque 1077	Sigma	Cat #: 10771
Histopaque 1119	Sigma	Cat #: 11191
QIAshredder	Qiagen	Cat #: 79656
Critical commercial assays
RNeasy Mini Kit	Qiagen	Cat #: 74106
High Capacity cDNA Reverse Transcriptase Kit	Applied Biosystems, ThermoFisher Scientific	Cat #: 4368813
MiniElute PCR Purification Kit	Qiagen	Cat #: 28006
DNA-free Kit, DNAse Treatment and Removal Reagents	Ambion, Life Technologies	Cat #: AM1906
Experimental models: organisms/strains
Mouse: C57BL/6J	Jackson Laboratory	Stock #: 000664
Mouse: Itgam−/− : B6.129S4-Itgamtm1Myd/J	Jackson Laboratory	Stock #: 003991
Mouse: Itgam+/−	This paper	N/A
Mouse: CD169-DTR: B6;129-Siglec1<tm1(HBEGF)Mtka>	(Miyake et al., 2007)	RikenBRC: 04395
Mouse: Cd169−/−: CD169-DTR KO	(Miyake et al., 2007)	RikenBRC: 04395
Mouse: Mertkfl/fl	(Fourgeaud et al., 2016)	N/A
Mouse: Lyz2-Cre: B6.129P2-Lyz2tm1(cre)Ifo/J	Jackson Laboratory	Stock #: 004781
Oligonucleotides
See Table S1 for qRT-PCR primer sequences	See Table S1	N/A
Software and algorithms
GraphPad Prism	GraphPad	http://www.graphpad.com
FlowJo	FlowJo, LLC	http://www.flowjo.com
Imaris Software Version 9.0.1.	Bitplane; Oxford Instruments	http://www.bitplane.com

Highlights.

• Neonates resist Streptococcus pneumoniae invasive infection

• Neonatal neutrophils have enhanced opsonin C3–integrin CD11b dependent phagocytosis

• Dampened efferocytosis increases CD11b^hi ‘aged’ neutrophils during early life

• Dependence on CD11b opsonophagocytosis predicts age-related patterns of infection