The B. pertussis strains Tohama 1 (WT B. pertussis) and BPH101, an isogenic pertussis toxin-deficient derivative (B. pertussisΔptx) have been previously described (22, 30–32). Bacteria were maintained on Bordet-Gengou agar (Difco) supplemented with 10% defibrinated sheep blood (Hema Laboratories). Liquid cultures were grown overnight in Stainer-Scholte broth at 37°C to mid-log phase then maintained in 20% glycerol stocks at -80°C for use as inoculum. Purified pertussis toxin (PTx) was obtained from Sigma-Aldrich as a lyophilized powder and resuspended with 500 μl PBS (P7208-5OUG). The purified PTx was added to the described inoculum at a concentration of 10 ng/μl.

Six- to eight-week old female and male C57BL/6J (00664) mice were procured from the Jackson Laboratory (Bar Harbour, ME) and bred in the Harvill mouse colony (University of Georgia, GA). All mice were maintained in specific pathogen-free facilities, and all experiments were conducted following institutional guidelines. Pups were utilized at the indicated ages (five-, seven-, eleven-, fourteen-, twenty-five-, and twenty-eight-days-old) and six- to eight-week old mice were utilized for adult experiments. Mice were lightly sedated with 5% isoflurane (Pivetal) and inoculated (10⁴ CFU suspended in either 15 μl PBS for neonates and 50 μl PBS for adults) by pipetting the inoculum as droplets into their external nares to be inhaled. At the indicated timepoints mice were euthanized via CO₂ inhalation and/or decapitation. Organs were excised and homogenized in 1 ml PBS, serially diluted, and plated on BG agar to quantify bacterial numbers. Colonies were counted following incubation for five days at 37°C.

Lungs were processed and stained as previously described (33). Viable cells were identified with Zombie Aqua (Biolegend). 0.35 μl of each extracellular antibody was added to each sample. Antibodies to identify T and B cell populations included anti-CD45 AF700 (clone: 30-F11, Biolegend), anti-CD3 APC (clone:17A2, Biolegend), anti-CD4 VF450 (clone:RM4-5, Tondo Biosciences), anti-CD8 APC Fire 750 (clone:53-6.7, Biolegend), and anti-CD19 PerCP/Cy5.5 (clone:1D3/CD19, Biolegend). Antibodies to identify virtual memory T cells (T_VM) and naïve adult T cells (CD44^loCD49d^lo) (34) included anti-CD8 BV650 (clone: 53-6.7, BD Biosciences), anti-CD4 AF700 (clone:RM4-55, Biolegend), anti-CD44 PE/Cy7 (clone: IM7, Biolegend), anti-CD49d APC Fire 750 (clone:R1-2, Biolegend), anti-CD3 FITC (clone: 17A2, Biolegend), anti-TCRg/d PE/Cy5 (clone: GL-3, Invitrogen), anti-NK1.1 PE (clone:PK136, Biolegend), anti-Eomes APC (clone: Dan11mag, Invitrogen), and anti-CXCR3 BV750 (clone: CXCR3-173, BD Biosciences). Intracellular staining for Eomes was performed as previously described (35). To identify neutrophils (CD45⁺Ly6G⁺CD11b⁺) and macrophages (CD45⁺Ly6G^- Siglec-F^-MHCII⁺), the gating strategy from Misharin et al. and the following antibodies were utilized; anti-CD45 AF 700 (clone: 30-F11, Biolegend), anti-MCHCII VF450 (clone:M5/114.15.2, Tonbo Biosciences), anti-CD11b PE (clone: M1/70, Tonbo Biosciences), anti-CD11c BV605 (clone: N418, Biolegend), anti-CD69 BV711 (clone: H1.2F3, Biolegend), anti-Ly6G FITC (clone: RB6-8C5, Tonbo Biosciences), and anti-Siglec F BV650 (clone: E50-2440, BD Biosciences). The acquisition of the data was performed using the Quanteon (Agilent) and analysis was performed with NovoExpress (Agilent) following the gating strategies in Supplementary Figures 2 and 3 and Misharin et al. (36)

Twelve, 5-day-old, female, C57BL/6J mice were randomly divided into 3 groups of 4 mice each. Mice received PBS (control) or 10⁴ CFU WT B. pertussis or B. pertussisΔptx in 15 μl PBS intranasally one time. Three days after inoculation, the animals were sacrificed with CO₂ euthanasia. The lungs were infused with neutral-buffered, 10% formalin fixative solution and immersed in the same fixative. One week following necropsy, tissues were removed from formalin solution, immersed in 70% ethanol, and remained in alcohol until ready for processing. Tissues were subsequently processed, embedded in paraffin, sectioned at approximately 5 μm, and stained with hematoxylin and eosin (HE). A board-certified pathologist (UBM) performed all microscopic evaluations of the HE stained sections.

Microscopic exam consisted of evaluation of the lung for the presence or absence of inflammation. Microscopically, lesion (tissue change or alteration) incidence, severity, and distribution were recorded. If absent (i.e., histologically normal), a score of 0 was assigned. If present, the severity of the lesions was recorded as minimal, mild, moderate, or severe, with severity scores of 1 through 4, respectively, based on an increasing extent and/or complexity of change, unless otherwise specified. Lesion distribution was recorded as focal, multifocal, or diffuse, with distribution scores of 1, 2, or 3, respectively. The Group Histopathological Score was calculated by adding individual animal severity + distribution scores.

Supernatant was taken from lung samples utilized for flow cytometry experiments. Samples were collected and stored at -20°C for cytokine analysis. 100 μl of supernatant from each lung sample was assessed for concentrations of IL-4 and IL-17 utilizing R&D Systems DuoSet ELISA kits following manufacturer’s instructions.

For experiments determining differences in bacterial loads in the organs of mice, the following statistical analyses were performed using GraphPad PRISM (GraphPad Software, Inc): Two-tailed unpaired Student t-tests, One-way ANOVA, and Two-way ANOVA. For experiments determining differences in immune cell populations, the following statistical analyses were performed using GraphPad PRISM (GraphPad Software, Inc): Two-tailed unpaired Student t-tests, One-way ANOVA, and Two-way ANOVA.

This study was carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Institutional Animal Care and Use Committees at The University of Georgia at Athens, GA (A2022 04-001-Y1-A0 Bordetella-Host Interactions, A2022 04-025-Y1-A0 Breeding Protocol, and A2022 04-022-Y1-A0 Neonatal Models of Bordetella infection, transmission, and immunity). Mice were consistently monitored for signs of distress over the course of the experiments to be removed from the experiment and euthanized using carbon dioxide inhalation to prevent unnecessary suffering.

To identify the appropriate aged mouse to assess neonatal susceptibility to B. pertussis (Bp), C57BL/6J mouse pups were inoculated with 10⁴ CFU Bp at five-, seven-, eleven-, and twenty-five-days-post birth (P5, P7, P11, and P25). Those inoculated at P7, P11, and P25 had roughly 10,000 CFU in their lungs 3 days post inoculation (dpi) ( Figure 1 ), demonstrating growth similar to that observed in adult mice ( Supplementary Figure 1 ). In contrast, mice inoculated at 5-days-old (P5) had approximately 5,000,000 CFU of Bp at 3 dpi ( Figure 1 ), over 500-fold higher numbers than any older mice, indicating that P5 pups are significantly more susceptible to this neonatal pathogen than mice even 2 days older (P7). These results suggest there is a significant shift in immunological development distinguishing mice inoculated on P5 and P7.

To assess shifts in the pulmonary immune response that distinguishes P5-P8 mice from more mature mice, populations of B and T cells were assessed in the lungs of C57BL/6J mice at various ages. We observed significantly higher numbers of CD3⁺ T cells and CD19⁺ B cells in the lungs of P8 than P10, P14, and P28 mice, suggesting that despite high numbers of lymphocytes in the periphery, P5 mice are still highly susceptible to Bp ( Supplemental Figure 2 ). Thus, it is likely that some feature of these lymphocytes in P5-P8 pups differ from those in P7, P10, P14, and P28 mice which results in significant susceptibility to Bp.

A key feature of immunological development is the maturation of the thymus and the T cells which develop therein. Therefore, subsets of T cells were characterized in the lung of P8, P10, P14, and P28 mice. A comprehensive flow cytometry panel was designed to assess shifts in populations of virtual memory T cells (T_VM). Using the gating strategy in Supplemental Figure 3 , we first identified populations of single positive CD3 α/β T cells via expression of CD3⁺NK1.1^-TCRγδ^- and CD4⁺ or CD8⁺, then T_VM were identified via CD44⁺CD49d^-CXCR3⁺Eomes⁺ expression. The panel and gating strategy were validated for specific binding via positive and negative controls ( Supplemental Figure 4 ). Shifts in populations of CD44^hiCD49d^- to CD44^hiCD49d⁺ antigen experienced T cells were increasingly observed from P8 and P10 to P14 and P28 mice ( Figure 2A ). The decreased expression of Eomes and CXCR3 in these CD44⁺CD49d^- populations followed a similar trend as mouse age progressed, indicating that the highest proportions of T_VM were observed in P8 mice ( Figure 2B ). T_VM comprised a significantly larger proportion of single-positive T cells (15%) in the lungs of P8 mice, while only comprising 1-4% of T cells in P10, P14, and P28 mice ( Figure 2C ). This trend was also reflected in the total T_VM numbers in the lungs of P8 mice, which had ~75x greater numbers than P10, P14, or P28 mice ( Figure 2D ). Conversely, naïve adult T cells (CD44^loCD49d^lo) made up only ~8% of single-positive T cells in the lungs of P8 mice, whereas they composed ~38% of single-positive T cells in the lungs of P10 mice ( Supplemental Figure 5 ).

Since there is no objective cutoff distinguishing “neonatal” mice from later stages of development, we here use the dramatically different sensitivity to Bp, predominance of T_VM, and relative lack of naive T_adult, as the basis for distinguishing P5-P8 mice, herein referred to as “neonatal”, from older mice referred to as “juvenile” (P10 to P21) or “adolescent” (P22 to P30) ( Figure 2 ). Utilizing this framework, these data demonstrate that neonatal mice are much more susceptible to Bp and that increased populations of T_adult observed in P10 mice are associated with control of Bp.

To better understand the extraordinary susceptibility of five-day-old (P5) mice to Bp, we inoculated them as above and followed the progression over time in the respiratory tract as well as the spleen, which is colonized in severe infections. At 2 hours post inoculation, most of the initial inoculum was deposited in the lungs, confirming successful inoculation ( Figure 3A ). Within 3 days, Bp grew in the lungs over 100-fold to levels exceeding 1,000,000 CFU, indicating a failure of these neonates to control infection. Bp grew similarly unrestrained in the nasal cavities, from ~100 CFU at 2 hours post inoculation to over 100,000 CFU by 3 dpi ( Figure 3B ). The extraordinary susceptibility of P5 mice was further underscored by the sporadic splenic colonization on 1 dpi that grew to nearly 1,000 CFU in all mice by 3 dpi, indicating consistent systemic dissemination ( Figure 3C ), reflecting a serious pneumonic infection and failure of systemic immune control. These results demonstrate that when delivered to neonatal mice (P5), a relatively low dose of Bp can efficiently and consistently grow rapidly in the nose and lungs and disseminate to the spleen within 3 days post inoculation, characteristic of severe neonatal disease.

Multiple studies have identified pertussis toxin (PTx) as contributing to severe disease in juvenile (P14) and adult mice (22, 29, 32), but its effects have not been examined in neonatal (P5-P8) mice, as defined here. To assess the role of PTx in the exceptional virulence of Bp in our neonatal model, we compared the wild-type parental strain to an isogenic mutant with an in-frame deletion of the coding region of ptx (BpΔptx) (32). Inoculation with WT Bp and BpΔptx demonstrated similar recovery from the lungs at 2 hours post-inoculation, confirming equivalent delivery of both strains to the lungs ( Figure 3A ). Within 3 days, BpΔptx was nearly cleared from the lungs of neonatal mice, with only approximately 100 CFU remaining, while WT Bp grew nearly 1000x the original inoculation dose in this same period ( Figure 3A ). Despite consistent colonization of the spleen by WT Bp, there was no detected systemic dissemination to the spleen by BpΔptx ( Figure 3C ). Complementation of the BpΔptx mutant via co-inoculation with purified PTx resulted in partial recovery of the phenotype in the lungs ( Supplemental Figure 6 ). This indicates that a single bolus delivery of PTx does not precisely mimic the continual secretion from the site of Bp microcolonies required for its effects. Together, these results indicate that the neonatal immune system can efficiently and rapidly control lung infection with the BpΔptx strain, but WT Bp substantially disrupts this ability via secretion of PTx.

Conventional infection models inoculate mice with supernaturally high doses of Bp (5x10⁵ CFU), potentially overcoming the most relevant host immune responses. Though C57BL/6 adult mice are approximately 10x the size of P5 pups, our much lower inoculation dose of 10⁴ CFU is 50x less than that delivered to adults. To examine the potential effects of high inocula in our neonatal mouse model, we used doses equivalent to those used in the conventional adult assays. Increasing the dose of BpΔptx delivered to P5 pups 10- and 100-fold, to 10⁵ and 10⁶ CFU, respectively, resulted in the death of most animals ( Supplemental Figure 7 ). These results demonstrate that the neonatal immune system can be effective against moderate numbers resembling natural infection but can be overwhelmed by unnaturally large inocula. They also demonstrate that extremely high doses can disrupt the ability to study and understand the natural function of the host immune system.

PTx has been demonstrated to delay immune cell recruitment to the site of infection by approximately one week in adolescent and adult mice (22, 29). To quantify the effects of PTx on populations of immune cells recruited to neonatal lungs, P5 pups inoculated with WT Bp or BpΔptx were assessed via flow cytometry at 2 hours, 1 day, and 3 days post inoculation. At 2 hours post inoculation, pups inoculated with BpΔptx had significantly higher neutrophil counts in the lungs than pups inoculated with WT Bp ( Figure 4A ), indicating that PTx interferes with rapid neutrophil recruitment. By 1 dpi, BpΔptx was present in much lower numbers than WT Bp but had recruited significantly higher (3x more) numbers of CD3⁺ T cells into the lungs ( Figure 4B ), though numbers of T_VM in the lungs remained consistent across infection groups ( Supplemental Figure 8 ). By 3 dpi, the mutant was nearly cleared from the lungs and the neutrophil and T cell numbers had already decreased, having substantially resolved the infection. In contrast, WT Bp had grown over 100-fold in number, but did not result in substantial increase in the populations of T cells in the lungs ( Figure 4B ). Instead, WT Bp resulted in significantly higher numbers of macrophages and neutrophils in lungs at 3 dpi, but these levels were not sufficient to clear or control the infection, as the bacterial load was 100-fold higher than the original inoculum and had already disseminated to the spleen ( Figures 4A, C ).

Histopathology of the lungs of infected mice was assessed as a measure of disease severity. To assess resulting inflammatory lesions in the lungs, P5 mice inoculated as above with WT Bp, BpΔptx, or PBS were histologically assessed at 3 dpi. H&E staining of lung sections of infected or naïve pups revealed that infection with BpΔptx did not result in observable inflammatory lesions and was indistinguishable from the PBS control pups ( Figures 5A, C, D , and Supplemental Table 1 ). In contrast, WT Bp induced substantial inflammation, as evidenced by the presence of inflammatory lesions in 100% of mice assessed ( Figures 5B, D and Supplemental Table 1 ).

Neonatal immune responses are largely anti-inflammatory, which minimizes collateral tissue damage that might result from uncontrolled inflammatory responses to the onslaught of mostly harmless bacteria encountered at birth. We therefore assessed concentrations of representative anti-inflammatory (IL-4) and pro-inflammatory (IL-17) cytokines in neonates infected with either WT Bp or BpΔptx. At 1 dpi, pups inoculated at P5 with BpΔptx had significantly higher concentrations of both IL-4 and IL-17 isolated from lung supernatant than naïve and WT Bp-inoculated pups ( Figures 6A, B ). By 3 dpi, when the infection was nearly controlled, the concentrations of these cytokines in pups inoculated with BpΔptx returned to levels similar to the naïve control ( Figures 6A, B ). Surprisingly, pups inoculated with WT Bp did not have increased IL-4 or IL-17 in lung supernatants at 1 or 3 dpi, despite very high numbers of bacteria, significant immune cell recruitment, and formation of inflammatory lesions. These results suggest that PTx plays a pivotal role in suppressing the production of IL-4 and IL-17 in neonatal mice, thereby allowing the pathogen to grow to high numbers.