Trimodulin supports antibacterial defence and restricts inflammation in preclinical pneumonia models

Graphical abstract

Overview of the study. CAP: community-acquired pneumonia; ctr: control; hpi: hours post infection; HVT: high tidal volume ventilation; i.n.: intranasal; i.p.: intraperitoneal; i.v.: intravenous; NV: non-ventilated; OD: optical density; OPKA: opsonophagocytosis killing assay; sCAP: severe community-acquired pneumonia; S. pneumoniae: Streptococcus pneumoniae; VILI: ventilator-induced lung injury.

Abstract

Background

Severe community-acquired pneumonia remains a global health challenge with high mortality despite advances in antibiotic therapy and supportive care. Immunoglobulin therapies, especially IgM-containing ones, have shown promise in enhancing host defence and reducing inflammation. The CIGMA trial highlighted the potential of trimodulin to lower mortality in patients with severe community-acquired pneumonia with high C-reactive protein and low IgM levels.

Methods

We investigated the protective effects of trimodulin on clinical status, bacterial burden, lung integrity and inflammatory responses in murine models of lung injury, including both ventilator-induced lung injury and infection-induced models with nonsterile inflammation.

Results

In mice, trimodulin significantly protected against lethal pneumococcal pneumonia by reducing bacterial burden and disease severity while preserving alveolar barrier integrity and limiting lung oedema. The antibacterial action of trimodulin was mediated through opsonophagocytosis, and its anti-inflammatory effects operated independently of the latter. When combined with ampicillin, trimodulin exhibited enhanced suppression of inflammation.

Conclusion

Our findings in preclinical pneumonia models suggest that trimodulin could be a promising therapy for severe community-acquired pneumonia. We provide evidence that trimodulin enhances host defence, reduces detrimental pulmonary inflammation and barrier dysfunction, and limits pulmonary oedema, which may explain the beneficial effects observed in patients with severe community-acquired pneumonia.

Shareable abstract

Trimodulin, a human plasma-derived polyvalent immunoglobulin preparation containing IgM and IgA for intravenous administration, effectively treats severe pneumococcal pneumonia by enhancing bacterial clearance and reducing inflammation https://bit.ly/40Zeoy2

Introduction

Pneumonia causes >2.49 million deaths annually worldwide [1, 2]. Among patients with community-acquired pneumonia (CAP), up to 10% are hospitalised; of those, up to 20% require intensive care [3]. 30-day mortality after hospitalisation for CAP in Germany varies from 2.8% for adults younger than 60 years to 26.8% for those aged 60 years and older and with comorbid conditions [4]. CAP is a leading cause of infectious disease-related deaths and a significant driver of healthcare costs worldwide [3].

Pneumonia is an inflammatory lung condition affecting the alveoli and distal airways. Various viral, bacterial and fungal pathogens may cause pneumonia [2]. Streptococcus pneumoniae is the most frequent causal bacterial pathogen in pneumonia. When infectious microorganisms invade the lungs, they provoke tissue damage and trigger host immune defence mechanisms that induce inflammation. Notably, lung inflammation and injury may develop despite effective antimicrobial therapy [1, 2]. More precisely, infection and inflammation may cause a breakdown of the alveolar–capillary barrier, eventually leading to protein-rich alveolar oedema, impaired gas exchange and reduced lung compliance. Thus, pneumonia may progress to acute respiratory distress syndrome, requiring life-saving respiratory support like mechanical ventilation [5]. The stress then imposed on lung tissue by the mechanical ventilation may itself aggravate lung tissue damage, causing ventilator-induced lung injury (VILI), subsequently enhancing the severity of the underlying lung pathology [6]. Adjunctive therapies, given in addition to antibiotics, aim to counteract the development of severe inflammation and lung failure, and ultimately to reduce morbidity and mortality.

Trimodulin, a polyvalent immunoglobulin preparation containing IgM (∼23%), IgA (∼21%) and IgG (∼56%), has emerged as a potential adjunctive immunomodulatory therapy in the treatment of severe CAP (sCAP) [7, 8]. Trimodulin has been investigated for the treatment of patients with certain indications, such as sCAP (CIGMA trial [7, 8]) and SARS-CoV-2-induced pneumonia (ESsCOVID trial [9]). Post hoc analyses of these two clinical phase 2 studies have demonstrated the efficacy of trimodulin in subgroups of patients showing elevated signs of inflammation [7–9]. Currently, trimodulin is being investigated for its efficacy and safety as an adjunctive treatment to standard of care in two phase 3 clinical trials (ESsCAPE (NCT05722938) [10, 11] and TRICOVID (NCT05531149) [12]. In these trials, trimodulin is compared to placebo in patients with sCAP requiring invasive mechanical ventilation (ESsCAPE trial) and in hospitalised patients with CAP, including COVID-19 pneumonia (TRICOVID trial), respectively.

Preclinical in vitro and in vivo studies point towards three possible modes of action by trimodulin: opsonisation for enhanced clearance of pathogens, neutralisation of bacterial exotoxins, and modulation of host inflammatory responses [13–15]. However, empirical data quantifying the specific contribution of each mode of action remain limited.

In our current study, we explored the therapeutic efficacy and mechanisms of action of trimodulin in severe pneumococcal pneumonia using murine models of S. pneumoniae infection [16, 17] and VILI [18]. We focused on pneumococcal pneumonia given its high prevalence in sCAP [19] and because bacterial pneumonia provides a clinically relevant model in which the proposed mechanisms of actions of trimodulin can be effectively investigated. Trimodulin demonstrated multifaceted protective effects in our murine models, improving survival, reducing bacterial growth and inflammation, and ameliorating lung barrier function. When used adjunctively with ampicillin, trimodulin significantly mitigated inflammation. Furthermore, trimodulin helped to maintain normal physiological parameters during mechanical ventilation and moderately limited VILI in the mouse model. These findings underscore the potential of trimodulin as an adjunctive therapy to antibiotics for improving outcomes in severe pneumonia.

Methods

Animals, housing and ethics statements

Specific pathogen-free female C57BL/6J mice (Charles River, Sulzfeld, Germany) aged 9–12 weeks were used for the experiments. All animal experiments were approved by institutional (animal welfare officer) and governmental (Landesamt für Gesundheit und Soziales Berlin) authorities and were in accordance with the Federation of European Laboratory Animal Science Associations (FELASA) guidelines and recommendations for the care and use of laboratory animals, which are equivalent to the US Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines. Handling of mice occurred under sterile conditions. Mice were randomly assigned to the appropriate experimental groups and were kept in closed, individually ventilated cages with filter hoods, under specific pathogen-free conditions, at room temperature with a 12 h light/dark cycle and free access to food and water. Mice undergoing experimental procedures were monitored for body temperature and weight changes and scored for humane end-point criteria at ∼12 h intervals. Mice that reached predefined end-point criteria scores were euthanised in accordance with official animal health guidelines to prevent onset of severe symptoms (supplementary tables S2 and S3).

Murine pneumococcal pneumonia model

Mice were anaesthetised with ketamine and xylazine. In the murine pneumococcal model, mice were transnasally infected with 5×10⁶ colony-forming units (CFU) of S. pneumoniae serotype 3 in the presence of hyaluronidase; control mice received PBS with hyaluronidase as previously described [20]. Bacteria were incubated at 37°C in liquid culture until the exponential growth phase. The suspension was washed with PBS. Bacterial concentration was adjusted and hyaluronidase was added. Each mouse received 5×10⁶ CFU in 20 µL. An optical density at 600 nm (OD₆₀₀) of 1 corresponds to 1×10⁹ CFU·mL⁻¹. The infection dose was controlled by plating serial dilutions. The disease score was assessed as depicted in supplementary table S2. Briefly, 0 disease points represented healthy symptom-free mice; mice that reached a B criteria four times or a C criteria once were classified as having met humane end-point criteria, equivalent to 8 disease points (category A: 1 point; category B: 2 points; category C: 8 points).

Murine VILI

Mice were ventilated as previously described [21]. Before ventilation, mice were anaesthetised with fentanyl, midazolam and medetomidine. Anaesthesia was maintained during ventilation with additional doses via an intraperitoneal (i.p.) catheter, and body temperature was maintained at 37°C using a heating pad. VILI was induced by high tidal volume (HVT) ventilation at 35 mL·kg⁻¹ for 4 h, while non-ventilated control mice (NV ctr) were only ventilated during experimental procedures at a low tidal volume of 9 mL·kg⁻¹. All mice underwent tracheotomy and intubation at 9 mL·kg⁻¹, at a rate of 160 breaths·min⁻¹, with an inspiratory oxygen fraction (F_IO2) of 0.5%, an I:E ratio of 0.5 and 2 cmH₂O positive end-expiratory pressure (PEEP). A carotid artery catheter facilitated blood pressure monitoring and infusion of a balanced electrolyte solution containing 150 M trometamol hydrochloride. Ventilation was conducted using a rodent ventilator (flexiVent; SCIREQ, Montreal, QC, Canada), continuously recording airway pressure, respiratory rate and tidal volume. Dynamic elastance, resistance and compliance were measured every 5 min using the forced oscillation technique, with static compliance documented after the initial recruitment manoeuvre and post-exsanguination. All groups underwent a recruitment manoeuvre; NV ctr mice were ventilated at 9 mL·kg⁻¹ for 10 min to collect baseline data, while HVT mice were ventilated at 35 mL·kg⁻¹ for 4 h at 70 breaths·min⁻¹. Dead space in the respiratory tubing was adjusted for the HVT group to maintain consistent pH and arterial CO₂ tension levels across all groups.

Mean arterial pressure, heart rate and peripheral oxygen saturation were monitored every 10 min. 90 min before termination, mice received 1 mg of human serum albumin via the arterial catheter. After 4 h of mechanical ventilation, F_IO2 was increased to 1.0% for 10 min. Heparin was injected through the arterial catheter 5 min before exsanguination, and mice were euthanised via the carotid artery catheter. No mice reached the humane end-point criteria or succumbed to the ventilation protocol [22].

Murine severe pneumococcal CAP with ventilation model

Mice were infected with 5×10⁶ S. pneumoniae serotype 3 or mock-infected with PBS (PBS ctr), both with hyaluronidase present. At 24 h post infection (hpi), mice underwent 6 h of mechanical ventilation at 12 mL·kg⁻¹, with the experimental analysis conducted at 30 hpi. Initial anaesthesia was administered via i.p. injection of fentanyl, midazolam and medetomidine, while body temperature was monitored using a rectal probe on a warming mat. Anaesthesia was maintained through an i.p. catheter, with isoflurane available for emergencies. To maintain fluid balance, 200 µL of Sterofundin Iso (Tham Köhler) was injected subcutaneously every 2 h as needed, and a bladder catheter ensured constant urine flow. A tracheostomy was performed by inserting and subsequently fixing a tracheal cannula, and ventilation was started, followed by a total lung capacity manoeuvre that briefly increased PEEP to 20 cmH₂O.

Ventilation continued for 360 min at a tidal volume of 12 mL·kg⁻¹, a PEEP of 2 cmH₂O, F_IO2 at 75% and an I:E ratio of 0.5, with body temperature and peak inspiratory pressure (PIP) recorded every 10 min. After 6 h of ventilation, anaesthesia was deepened and heparin injected into the right ventricle and cardiac arrest was induced via exsanguination through the caudal vena cava. Ventilation was stopped, followed by bronchoalveolar lavage (BAL) via the tracheal cannula, and organ sampling for analysis [23].

Trimodulin, placebo and ampicillin treatments in murine in vivo models

To ensure bioavailability comparable to human infusion, trimodulin was administered intravenously (i.v.) at 384 mg·kg⁻¹ of body mass (BM), with a buffer control given in equivalent volume. Ampicillin (0.4 mg) was delivered by i.p. injection, with a NaCl solution control group. For i.v. injections via the lateral tail vein, mice were restrained and the tail was warmed with infrared light. In the pneumococcal pneumonia model, mice received two i.v. injections of trimodulin or buffer control, one at −2 h and one at 22 hpi. For the VILI model, a single preventive was given 30 min before ventilation. In the severe pneumococcal CAP model, a therapeutic protocol involved injecting trimodulin or buffer control with ampicillin or NaCl control solution at 22 hpi after pneumonia had fully developed.

Statistical analysis

PRISM GraphPad software (versions 9 and 10) was used for statistical analysis. p-values <0.05 were considered statistically significant. For parametric data, two groups were compared using an unpaired t-test, and three or more groups by one-way ANOVA. For non-parametric data, two groups were compared using a Mann–Whitney test. Grouped analyses of parametric and non-parametric data were performed using two-way ANOVA.

Role of funders

Funding sources had no role or influence in the design, execution, data analysis or interpretation of the results of the study.

Results

Prophylactic trimodulin protects against lethal pneumococcal pneumonia in a murine model

We assessed the efficacy of trimodulin by evaluating its impact on pathogen burden and host immunity after preventive intravenous treatment (figure 1a). To align with the CIGMA trial dosage for patients with sCAP undergoing invasive mechanical ventilation [7, 8], we performed pharmacokinetic and efficacy analyses in mice (supplementary figure S1). Pharmacokinetic analysis showed a faster decline of IgM in mice compared to humans, confirming previous findings of shorter half-lives for human immunoglobulins in murine models [24]. Administering the highest dose twice (supplementary figure S1c) resulted in significant improvements in disease parameters, including body temperature and weight, as well as reduced CFU in lung and blood samples (supplementary figure S1d–g). Consequently, a dose of 384 mg·kg⁻¹ given 2 h prior (−2 hpi) and 22 h post infection (22 hpi) was selected for subsequent experiments in mice.

FIGURE 1.

Trimodulin alleviated pneumococcal pneumonia burden and reduced lung oedema. a) Preventive trimodulin treatment scheme in murine pneumococcal pneumonia. B6J mice were intravenously (i.v.) treated with trimodulin (7.7 µL·g⁻¹=384 mg·kg⁻¹ body mass) or buffer control 2 h prior to (−2 h post infection (hpi)) and 22 hpi S. pneumoniae serotype 3 or PBS mock infection. b) Body weight (% of start), c) rectal body temperature and d) clinical score (supplementary table S2) over time. Data are presented as curves with mean±sem, n=12. Statistical significance determined with mixed-effects models (restricted maximum likelihood) and Šídák's multiple comparisons test between infected groups. Time point 72 hpi not represented by buffer control (ctr) + S. pneumoniae-infected group and therefore excluded from the statistical analysis. e) Per cent survival using Kaplan–Maier curves. A log-rank (Mantel–Cox) test was done between infected groups. f) Aspartate aminotransferase (AST) and g) blood urea levels in plasma of mice, measured by Cobas 8000 C701 (Roche Diagnostics). Statistical significance determined using one-way ANOVA with Šídák's multiple comparisons test between PBS and S. pneumoniae-infected groups, n=0–6. h) Bronchoalveolar lavage fluid (BALF) samples from different time points were analysed for protein content as a representation of lung barrier function. Statistical significance determined using two-way ANOVA with Tukey's multiple comparisons test. i, j) Haematoxylin and eosin-stained lungs sections were analysed and scored by a pathologist for i) alveolar oedema and j) perivascular oedema. ctr: control; i.n.: intranasal. *: statistical difference between buffer ctr + S. pneumoniae versus trimodulin + S. pneumoniae groups; ^#: statistical difference between buffer ctr + mock PBS versus buffer ctr + S. pneumoniae groups; ^¶: statistical difference between trimodulin + mock PBS versus S. pneumoniae-infected groups. *: p<0.5; **^/##/¶¶: p<0.01; ***^/###/¶¶¶: p<0.001; ****^/####: p<0.0001.

To evaluate the effects of trimodulin on disease progression following S. pneumoniae infection, we monitored murine vital signs over time. Assessments were conducted at the start (0 hpi) and at 12 to 24-h intervals. Uninfected mice maintained stable health, while infected mice showed continuous weight loss, decreased body temperature and increased disease scores, ultimately reaching humane end-point criteria within 60 hpi (figure 1e). Infected mice treated with trimodulin exhibited significantly improved outcomes across all parameters compared to controls (figure 1b–e). Kaplan–Meier curves indicated a survival rate of approximately 75% in trimodulin-treated mice (figure 1e). Additionally, mild increases in aspartate aminotransferase, a marker of liver damage, were observed at 48 hpi but were prevented by trimodulin (figure 1f). No significant changes in blood urea, an indicator of kidney dysfunction, were noted at this time (figure 1g). Liver damage may potentially limit hepatic urea synthesis from amino acid catabolism [25].

We measured protein extravasation into the alveolar compartment over time as an indicator of lung-vascular permeability. At 6 hpi, protein concentrations in BAL fluid (BALF) of infected and uninfected mice were similar (figure 1h). By 24 hpi, protein levels had significantly increased in infected versus uninfected controls yet remained unaffected by treatment. At 48 hpi, buffer-treated mice showed a steady rise in BALF protein concentrations, reaching approximately 1 mg·mL⁻¹, while trimodulin treatment significantly prevented this increase (figure 1h). Histopathological assessment of alveolar and perivascular oedema scores revealed similar trends in trimodulin-treated mice (figure 1i, j, supplementary figure S2).

Prophylactic trimodulin reduces bacterial burden in a pneumococcal pneumonia mouse model

To assess the antimicrobial effects of trimodulin treatment during pneumococcal pneumonia (figure 2a), we measured bacterial burden in primary and secondary target organs (figure 2b–e). Consistent with prior efficacy pilot results (supplementary figure S1f, g), trimodulin at 384 mg·kg⁻¹ BM significantly reduced bacterial burden in BALF, blood and spleen within 24 hpi, whereas a reduction in the lungs was observed at 48 hpi. Thus, trimodulin effectively prevented the development of bacteraemia in S. pneumoniae-infected mice from 24 hpi onwards (figure 2b–e).

FIGURE 2.

Trimodulin reduces bacterial burden. a) Prophylactic treatment scheme in murine pneumococcal pneumonia. B6J mice were treated intravenously (i.v.) with trimodulin (7.7 µL·g⁻¹=384 mg·kg⁻¹ body mass) or buffer control 2 h prior to (−2 h post infection (hpi)) and 22 hpi with S. pneumoniae serotype 3 or PBS mock infection. b–e) Bacterial burden in b) bronchoalveolar lavage (BAL), c) lungs, d) blood and e) spleen. CFU data were log transformed and data are presented as mean±sem, n=12. Statistical significance determined using two-way ANOVA with Šídák's multiple comparisons test. Dotted lines indicate limit of detection. f–i) Lung sections were immunohistochemically stained for pneumococci and the presence of bacteria was scored in four lung compartments: f) alveolar parenchyma, g) perivascular space, h) subpleural space and i) mediastinal adipose tissue. Scores presented as mean±sem, n=0–6. j–m) Corresponding representative images of infected, buffer control-treated (top row) and infected trimodulin-treated mice (bottom row) at 48 hpi of the j) alveolar parenchyma, k) perivascular space, l) subpleural and m) mediastinal adipose tissue. Free streptococci (red) marked with solid ovals. Phagocytosed streptococci (red) within neutrophils marked with small, solid circles or within macrophages marked with large, dotted circles. Original magnification ×600, scale bar: 100 µm. ctr: control; i.n.: intranasal; S.pn.: S. pneumoniae. ****: p<0.0001.

We subsequently conducted immunohistochemical staining of pneumococci in lung tissue to cross-validate our previous findings and to explore bacterial distribution patterns during and after treatment (figure 2f–m, supplementary figure S3). Murine lungs were analysed at 24, 48 and 72 hpi. At 24 hpi, bacteria were dispersed in the lung parenchyma of all animals and found in macrophages and neutrophils in buffer-treated infected mice (figure 2f, supplementary figure S3). By 48 hpi, bacteria were clustered in all compartments examined (lung parenchyma, perivascular space, subpleural space, mediastinal adipose tissue) with an increase in macrophage-associated bacteria. By contrast, trimodulin-treated animals showed free bacteria only in the alveolar parenchyma at all time points, along with a decrease in bacterial load in this compartment during the course of infection (48 hpi, 72 hpi) (figure 2f, supplementary figure S3). No bacteria were detected in the lungs of uninfected controls (supplementary figure S3).

The antibacterial effect of trimodulin depends on opsonophagocytosis

Before investigating the mechanism behind the antibacterial effects of trimodulin, we first assessed its ability to reach the site of infection for local action. BALF samples were collected at 6, 24 and 48 hpi to measure human IgM, IgA and IgG antibodies (figure 3a–d). By 24 hpi, significant levels of trimodulin were present in the alveolar compartment. Notably, the concentration of human IgM and IgA in plasma remained higher in S. pneumoniae-infected animals, but did not significantly affect human IgG levels (figure 3e–g).

FIGURE 3.

Trimodulin translocates into alveolar spaces in vivo and eliminates S. pneumoniae via opsonophagocytosis in vitro. a) Preventive trimodulin treatment scheme in murine pneumococcal pneumonia. B6J mice were intravenously (i.v.) treated with trimodulin (7.7 µL·g⁻¹=384 mg·kg⁻¹ body mass) or buffer control 2 h prior to (−2 h post infection (hpi)) and 22 hpi with S. pneumoniae serotype 3 or PBS mock infection. b–d) Levels of human b) IgM, c) IgA and d) IgG levels in bronchoalveolar lavage fluid (BALF) of experimental mice. e–g) Levels of human e) IgM, f) IgA and g) IgG levels in plasma of experimental mice. Ig levels were determined by semi-quantitative ELISA measurements. For b–g, data presented as mean±sem. Statistical significance determined by two-way ANOVA with Šídák's multiple comparisons test between trimodulin-treated groups, n=10–13. h) Growth curves of different S. pneumoniae serotypes (STs) in liquid media in the presence of buffer (negative ctr), 5 mg·mL⁻¹ trimodulin or 3 µg·mL⁻¹ ampicillin (positive ctr). Optical density at 600 nm (OD₆₀₀) was measured every 30 min. Three independent experiments were performed in triplicate. Representative graph shown with data presented as mean±sem, n=3. i) Opsonophagocytic killing assay of buffer ctr, trimodulin or World Health Organization (WHO) anti-streptococcus sera (pneumococcal vaccinated donors), using neutrophil-like HL60 cells as phagocytes, and live S. pneumoniae as target cells. Four independent experiments were performed in triplicate. Data are presented as mean±sem, n=4. Statistical significance determined by two-way ANOVA with Tukey's multiple comparisons test. There was no significant difference between trimodulin and WHO (positive ctr). ctr: control; i.n.: intranasal; S.pn.: S. pneumoniae. ^#: trimodulin versus buffer ctr; ^¶: WHO versus buffer ctr; *: p<0.5; **^/##: p<0.01; ***^/¶¶¶: p<0.001; ****: p<0.0001.

Next, we investigated whether trimodulin had a direct effect on S. pneumoniae growth. S. pneumoniae wild-type strains and mutants lacking a capsule were cultured with either 5 mg·mL⁻¹ trimodulin, formulation buffer or 3 µg·mL⁻¹ ampicillin. Growth curve analyses showed that neither trimodulin nor the formulation buffer affected S. pneumoniae growth, while ampicillin effectively inhibited growth of all strains, regardless of serotype or mutant status (figure 3h). Given the lack of direct antibacterial effect in vitro, we proceeded with an opsonophagocytic killing assay, because trimodulin is expected to exert its antibacterial effects through cell-mediated opsonophagocytosis.

The opsonophagocytic killing assay demonstrated that trimodulin effectively aids in the elimination of S. pneumoniae with the help of neutrophil-like granulocytes and complement, similar to the positive control (serum from vaccinated individuals). In contrast, the formulation buffer used as a negative control had no effect on the CFU count of S. pneumoniae. The strongest phagocytic effect was observed at dilutions of trimodulin and positive control (World Health Organization) serum ranging from 1:30 to 1:100 (figure 3i).

Trimodulin and ampicillin therapy each reduce bacterial burden and together limit systemic inflammation

We found that prophylactic trimodulin treatment is life-saving in a model of pneumococcal pneumonia. To evaluate its therapeutic potential, we tested trimodulin in a complex mouse model simulating critically ill CAP patients on mechanical ventilation after S. pneumoniae infection. We combined trimodulin with ampicillin to determine if it provides additional benefits beyond reducing bacterial burden. Mice were infected with S. pneumoniae or sham-infected with PBS, then treated 22 h later with placebo (buffer and NaCl), trimodulin, ampicillin or both, followed by either ventilation or no ventilation for 6 h until a 30 hpi analysis (figure 4a).

FIGURE 4.

Trimodulin and ampicillin therapy alone and in combination reduced bacterial burden. a) Trimodulin therapeutic treatment scheme in murine severe community-acquired pneumonia models. Mice were intranasally (i.n.) infected with a dose of 5×10⁶ S. pneumoniae serotype 3 or PBS mock infection. At 22 h post infection (hpi), mice received treatment according to the group assignments (384 mg·kg⁻¹ body mass trimodulin or buffer control (ctr) intravenously (i.v.) and 0.4 mg per mouse of ampicillin or NaCl ctr intraperitoneally (i.p.)). At 24 hpi, corresponding groups underwent ventilation for 6 h with a tidal volume of 12 mL·kg⁻¹, a positive end-expiration pressure of 2 cmH₂O and an inspiratory oxygen fraction of 75%, or remained unventilated (NV ctr). All groups were analysed at 30 hpi. b) Peak inspiratory pressure (PIP) was measured every 10 min in mice receiving mechanical ventilation. Data are presented as the mean. c) Body temperature of NV ctr mice was assessed as a disease parameter at 30 hpi. Data presented as mean±sem. Statistical significance was determined using one-way ANOVA with Tukey's multiple comparisons test. Significant differences are reported only for the placebo (buffer ctr and NaCl ctr) groups when comparing PBS-treated mice with S. pneumoniae-infected mice. If applicable, significant differences among various S. pneumoniae experimental groups are provided for all group comparisons. d–g) Bacterial burden was determined at 30 hpi in all experimental groups in d) bronchoalveolar lavage (BAL), e) lungs, f) blood and g) spleen. Data presented as mean±sem. Statistical significance was determined using two-way ANOVA with Šídák's multiple comparisons test, tested between NV ctr and ventilated groups, n=6–12. ctr: control; i.n.: intranasal; S.pn.: S. pneumoniae. *: p<0.05; **: p<0.01; ***: p<0.001; ****: p<0.0001.

In the ventilated groups (tidal volume of 12 mL·kg⁻¹ for 360 min), the PIP was recorded every 10 min, showing no differences among infected treatment groups (figure 4b). All animals were monitored for body weight and temperature at 0 hpi, and 22 hpi (treatment time). During the 6 h ventilation period, anaesthetised animals were kept at steady body temperature using a warming mat, so temperature changes from 22 hpi to 30 hpi could only be evaluated in non-ventilated groups. In this group, trimodulin treatment alone or in combination with ampicillin prevented significant drops in body temperature. Ampicillin alone showed a similar trend but did not reach statistical significance (figure 4c).

As previously demonstrated, prophylactic trimodulin treatment reduced the bacterial burden of mice infected with S. pneumoniae (figure 2). Here, we confirmed that trimodulin significantly decreased bacterial burden in lungs, blood and spleen even when administered during a fully established infection (at 22 hpi), within 8 h of treatment (30 hpi analysis). As anticipated, the antibacterial effect was enhanced when trimodulin was combined with ampicillin (figure 4d–g).

To assess lung organ integrity at 30 hpi, an independent group of mice underwent the same experimental and therapeutic treatment for histopathological analysis (figure 5a). Lungs were stained with haematoxylin and eosin and evaluated for signs of pneumonia severity by pneumococcal pneumonia-adapted scores by Li Bassi et al. [26] and Matute-Bello et al. [27] to assess acute lung injury. In both non-ventilated and ventilated groups, S. pneumoniae infection increased pneumonia severity and affected lung areas. Combined treatment with trimodulin and ampicillin resulted in the lowest pneumonia scores, acute lung injury scores and affected lung areas in non-ventilated animals (figure 5b–e).

FIGURE 5.

Combined trimodulin and ampicillin therapy reduced pulmonary and systemic inflammation. a) Trimodulin therapeutic treatment scheme in murine severe community-acquired pneumonia models. Mice were intranasally (i.n.) infected with a dose of 5×10⁶ S. pneumoniae serotype 3 or PBS mock infection. At 22 h post infection (hpi), mice received treatment according to the group assignments (384 mg·kg⁻¹ body mass trimodulin or buffer control (ctr) intravenously (i.v.) and 0.4 mg ampicillin or NaCl ctr). At 24 hpi, corresponding groups underwent ventilation for 6 h with a tidal volume of 12 mL·kg⁻¹, a positive end-expiration pressure of 2 cmH₂O and an inspiratory oxygen fraction 75%, or remained unventilated (NV ctr). All groups were analysed at 30 hpi. b) Degree of pneumonia modified from Li Bassi et al. [26]. Grading ranged from 0 (no pneumonia) to 1 (minimal), 2 (mild), 3 (moderate), 4 (severe) or 5 (massive). c) Acute lung injury (ALI) score by Matute-Bello et al. [27]. d) Lung area affected (%). Data presented in b–d as mean±sem, n=1–4. e) Scanned images of cross sections of lungs stained with haematoxylin and eosin. Scale bar: 4 mm. f–i) Cytokine measurements by multiplex ELISA in plasma. Data presented as mean±sem, n=5–8. Statistical significance determined by two-way ANOVA with Dunnett's multiple comparison test, tested against placebo (buffer ctr and NaCl ctr) versus S. pneumoniae-infected groups. ctr: control; S.pn.: S. pneumoniae. *: p<0.05; **: p<0.01; ***: p<0.001; ****: p<0.0001.

Inflammatory mediators such as interferon-γ (IFNγ), tumour necrosis factor (TNF), interleukin-6 (IL6) and interleukin-1β (IL1β) and various CXCL and CCL chemokines in BALF were not reduced by the different treatments (supplementary figure S4a–i). However, systemic inflammatory protein levels were significantly reduced in the ventilated groups (figure 5e–h, supplementary figure S4j–m).

To explore the inflammation-modulating potential of the treatments at higher resolution, we performed RNA sequencing (RNA-seq) on blood cells from non-ventilated groups, in which protein levels of inflammatory mediators were low (figure 6a, supplementary figure S5a).

FIGURE 6.

Trimodulin and ampicillin combination therapy reduced inflammatory gene expression upon S. pneumoniae infection. a) Trimodulin therapeutic treatment scheme in a murine severe community-acquired pneumonia model. Mice were infected with a dose of 5×10⁶ S. pneumoniae serotype 3. At 22 h post infection (hpi), mice received treatment according to group assignments (384 mg·kg⁻¹ body mass trimodulin or buffer control (ctr) intravenously (i.v.) and 0.4 mg ampicillin or NaCl ctr). All groups were analysed at 30 hpi. b) Volcano plot displaying differentially expressed genes (DEGs) between trimodulin and ampicillin combination therapy versus placebo (buffer ctr and NaCl ctr) treatment. DEGs with increased (log₂ fold changes >0.5 and adjusted p<0.05) or decreased expression (log₂ fold changes <−0.5 and adjusted p<0.05) are denoted in red and blue, respectively. c) Hallmark gene sets of DEGs in the indicated experimental comparisons. In the case of trimodulin alone, no enriched gene sets were detected. When using the combination therapy of ampicillin and trimodulin, lower adjusted p-values and a higher number of regulated genes were enriched compared to ampicillin therapy alone. d) Dot plot of log₂ fold changes for different comparisons for the selected key genes, which belong to the gene sets “complement activation” and “inflammatory response”. Differential expression analyses were performed with DESeq2 and the Wald test was used to identify DEGs. p-values were adjusted for multiple testing using the Benjamini–Hochberg method. Pathway enrichment analyses were performed using the R package tmod (www.r-project.org). S.pn.: S. pneumoniae; Tnfɑ: tumour necrosis factor ɑ; NfκB: nuclear factor κB; Ifn: interferon.

Principal component (PC) analysis revealed that infection status primarily drove differences between experimental groups (PC1: 56% variance, supplementary figure S5b). Differential gene expression analysis showed that trimodulin treatment alone regulated only a few genes (supplementary figure S5c). However, when combined with ampicillin, this therapy had a greater impact on gene expression than ampicillin alone, downregulating several inflammatory mediators such as Cxcl2, Cxcl10, the neutrophil activation marker Cd177 and interferon-related genes (figure 6b, supplementary figure S5c).

Gene set enrichment analysis [28] using “hallmark” gene sets [29] identified enriched sets for the combined trimodulin and ampicillin treatment, including reactive oxygen species pathway (M5938), inflammatory response (M5932), complement (M5921), IFNγ response (M5913), IFNα response (M5911), IL6 JAK STAT3 signalling (M5897), and TNFα signalling via NF-κB (M5890) (figure 6c). These sets were not enriched with trimodulin treatment alone. Given that trimodulin regulates complement pathways [14], we further investigated key regulated genes contributing to the inflammatory response (M5932) and complement (M5921) gene sets (supplementary figures S6 and S7).

The expression of the C3, C5a receptor 1 (C5ar1) and Toll-like receptor 2 (Tlr2) genes was significantly decreased compared to the infected placebo group. Fibronectin (Fn1) expression was reduced with both ampicillin and combined treatments. Cxcl10 and Cxcl2 expression consistently decreased across multiple comparisons, with the strongest reduction observed in the combined trimodulin and ampicillin versus placebo (figure 6d).

In summary, despite unchanged protein levels in the treated, non-ventilated groups (NV ctr), transcriptome analysis revealed significant shifts in gene expression, indicating that the combination of trimodulin and ampicillin effectively attenuates the inflammatory response.

Trimodulin prophylaxis mildly alleviates ventilation-induced lung damage

In the next experiments, we evaluated the effect of trimodulin treatment on VILI during mechanical ventilation. Naïve mice were treated with trimodulin or buffer control and placed on a ventilator, with some subjected to 240 min of HVT ventilation while others were connected for only 10 min to establish baseline parameters (NV ctr). Clinical parameters were monitored every 10 min to assess the animal's clinical status (figure 7a).

FIGURE 7.

Trimodulin diminished ventilator-induced lung injury. a) Preventive treatment scheme in murine ventilator-induced lung injury model. Mice were treated intravenously (i.v.) with trimodulin (384 mg·kg⁻¹ body mass) or buffer control (ctr) 30 min prior to ventilation start. b) Mean arterial pressure (MAP), c) heart rate (HR) and d) oxygen saturation (S_pO2) measured at 10-min intervals during 240 min of high tidal volume (HVT) ventilation. Statistical significance was determined using two-way ANOVA. e–i) Blood gas parameters e) pH, f) bicarbonate (HCO₃⁻), g) standardised arterial oxygen tension (stP_aO2), h) arterial carbon dioxide tension (P_aCO2) and i) standard base excess (SBE) in arterial blood measured after no ventilation (NV) or 240 min of HVT ventilation. Statistical significance determined by two-way ANOVA with Šídák's multiple comparisons test for selected comparisons (between treatments and between ventilation regimes). j) Airway opening pressure (AOP), k) compliance (C) and l) elastance (E) normalised to 1 and measured at 5-min intervals during 240 min of HVT ventilation, analysed using mixed-effect modelling (restricted maximum likelihood). Data presented as mean±sem, n<15. m) Mouse lung permeability measured as the ratio of mouse albumin concentration in plasma and bronchoalveolar lavage fluid (BALF) samples. n) Proportions and numbers of BALF neutrophils (Ly6G⁺CD11b^hi) measured by flow cytometry. Statistical significance determined in m and n by two-way ANOVA with Šídák's multiple comparisons test for selected comparisons (between treatments and between ventilation regimes). *: p<0.05; **: p<0.01; ****: p<0.0001.

During ventilation, the mice exhibited a general decrease in mean arterial pressure and an increase in heart rate (figure 7b, c). Oxygen saturation remained stable for the first 180 min, but declined in the final 60 min (figure 7d). Compared to the NV ctr group, HVT ventilation resulted in increased pH, similar arterial blood bicarbonate (HCO₃⁻) levels, decreased standardised partial pressure of oxygen, decreased partial pressure of carbon dioxide and slightly higher base excess values (figure 7e–i). Blood sodium (Na⁺) levels increased upon HVT, while potassium (K⁺) and calcium (Ca²⁺) levels remained stable, and lactate levels rose (supplementary figure S8). Overall, differences in arterial blood gas and biochemical parameters between HVT and NV groups were primarily related to ventilation status, with trimodulin treatment not affecting these parameters (figure 7b–i, supplementary figure S8).

We used SCIREQ flexiVent software to monitor lung function by recording ventilator parameters every 10 min during HVT ventilation. Parameters measured included tracheal pressures, resistance, compliance and elastance (figure 7j–l, supplementary figure S8f). Airway opening pressure (AOP) was controlled in the trachea to prevent barotrauma. Trimodulin-treated mice showed a statistically significant reduction in AOP levels compared to controls during HVT ventilation (figure 7j), indicating reduced transpulmonary pressures and stress on the lung parenchyma.

Airway resistance measured at the chest wall and endotracheal tube showed no significant difference between groups during HVT ventilation (supplementary figure S8f). Compliance, reflecting the lung's ability to expand, decreased initially due to surface tension but was less affected in trimodulin-treated mice (figure 7k). Both groups exhibited an increase in elastance, but trimodulin-treated mice had significantly lower elastance compared to controls (figure 7l).

HVT ventilation modestly increased alveolo-capillary permeability in both groups, with trimodulin maintaining lower permeability indices, though differences were not statistically significant (figure 7m). Histological analysis indicated trends towards reduced perivascular oedema and alveolar wall injury in trimodulin-treated mice, although alveolar oedema was not significantly different (supplementary figure S9).

To assess anti-inflammatory cell-modulating properties, we monitored alveolar neutrophil recruitment. In our VILI model of sterile lung inflammation, trimodulin treatment significantly reduced neutrophils proportions, consistent with the overall neutrophil counts (figure 7n). In murine models of pneumococcal pneumonia and ventilated sCAP, neutrophil recruitment showed more variability due to the presence of S. pneumoniae (supplementary figure S10).

In conclusion, our study demonstrated robust protective effects of trimodulin against pneumococcal pneumonia in mice. Trimodulin enhanced opsonophagocytosis and limited inflammation independently of its antimicrobial properties and in conjunction with antibiotics. Additionally, in a sterile lung injury model of VILI, trimodulin protected lung function and reduced alveolar neutrophil recruitment.

Discussion

Our results demonstrate that trimodulin offers significant therapeutic benefits for treating severe pneumococcal pneumonia and VILI in mice, highlighting its potential as an adjunctive treatment to reduce mortality and inflammation.

In the earlier phase 2 CIGMA trial, adjunctive trimodulin therapy was studied in pneumonia patients in need of invasive mechanical ventilation. Trimodulin improved survival in patient subgroups with immune dysregulation, particularly in patients with high C-reactive protein levels, low lymphocyte counts or high neutrophil-to-lymphocyte ratios at baseline. The effect was even more pronounced in patients with low IgM levels [7, 8]. Based on these results, we analysed the mechanisms that support the beneficial effects of trimodulin in severe pneumonia and the interaction between trimodulin and antibiotic treatment across various mouse models.

In vivo, we found that trimodulin significantly reduced bacterial burden, inflammation, pulmonary vascular permeability, lung oedema and disease severity, including mortality, in pneumococcal pneumonia. When used alongside ampicillin, trimodulin further decreased inflammation independent of bacterial load. Trimodulin also mitigated the recruitment of inflammatory neutrophils and improved lung compliance and oxygenation in a murine model of sterile pulmonary immune stimulation due to mechanical ventilation, correlating with clinical observations that neutrophil counts in patients with sCAP rapidly decreased with trimodulin treatment [8]. Notably, while trimodulin did not exhibit direct antibacterial effects in vitro, it enhanced phagocytosis of viable S. pneumoniae in an opsonophagocytosis assay, indicating effective support of the host's antimicrobial killing mechanisms. This is consistent with recent findings that trimodulin enhances opsonophagocytosis of fluorescent Staphylococcus aureus bioparticles [30]. In addition to reducing bacterial burden in murine pneumococcal pneumonia, trimodulin preserved alveolar barrier integrity, reduced lung oedema, prevented liver injury, helped to maintain body weight and body temperature, decreased disease score and improved survival rates. Previous research demonstrated that BT086, the predecessor of trimodulin [31], reduced systemic Escherichia coli burden in a rabbit model, enhancing overall disease outcome [13]. Furthermore, we and others have shown that trimodulin protects against platelet destruction caused by the S. pneumoniae exotoxin pneumolysin, which may contribute to the preservation of lung barrier integrity [16, 32, 33].

To assess whether the benefits of trimodulin in S. pneumoniae-infected mice stemmed solely from reduced bacterial burden or from additional effects, we combined antibiotic treatment with adjunctive trimodulin therapy. Ampicillin was selected as a first-line β-lactam antibiotic because it effectively targets S. pneumoniae, significantly reducing bacterial load and associated morbidity in pneumonia models. While its use in patients with sCAP varies depending on patient characteristics and local microbiological resistance patterns, ampicillin is well tolerated and serves as a reliable component of a broader therapeutic strategy, thus making it a critical choice for empirical treatment [34].

To model the clinical situation of respective patients, we mechanically ventilated mice with pneumococcal pneumonia and initiated trimodulin treatment once severe pneumonia developed (22 hpi) [12, 22, 35]. While effectively contributing to host bacterial defence in vivo, trimodulin was less effective in bacterial killing compared to ampicillin. However, when combined with ampicillin, trimodulin significantly reduced inflammation in both ventilated and non-ventilated animals without further enhancing bacterial killing. Additionally, levels of specific cytokines such as IL12, IFNγ, TNF, IL6 and IFNα were lowered in the plasma. This finding aligns with previous in vitro studies such as from Bohländer et al. [15], who demonstrated that trimodulin, compared to i.v. Ig (comprising >95% IgG), exerted anti-inflammatory effects, such as the suppression of IL8, using neutrophil-like HL-60 cells exposed to SARS-CoV-2 spike-protein-coated beads. Similarly, when neutrophil-like HL-60 cells were challenged with S. aureus bioparticles, trimodulin treatment led to a reduction in IL8 secretion compared to i.v. Ig [36]. In vivo, Pentaglobin, another IgM-enriched polyvalent antibody preparation (containing 12% IgM, 12% IgA and 76% IgG), reduced alveolar TNF levels in rats in an in vivo pneumonia model developing shock, which is based on a combination of injurious mechanical ventilation and intratracheal lipopolysaccharide treatment [37]. Consistent with the above findings, trimodulin was shown to attenuate ex vivo endotoxin-induced immune reactions during early hyperinflammation, further supporting its potential anti-inflammatory properties [38]. Although still speculative, these observations suggest that the enhanced IgM and/or IgA content in Ig preparations such as Pentaglobin and trimodulin may contribute to their immunomodulatory effects. Notably, Pentaglobin contains multimeric, J-chain-positive forms of IgA and IgM, which are essential for transcytosis across mucosal epithelium via the polymeric immunoglobulin receptor, thereby enabling the formation of secretory IgA [39]. Given that trimodulin is derived from human plasma, it is likely to contain similar multimeric forms capable of enhancing mucosal barrier protection. In fact, in vitro evidence suggests that, unlike IgG, secretory IgA is generally capable of preventing excessive neutrophil activation and protecting epithelial barrier integrity in a co-culture model involving Calu-3 epithelial cells and neutrophils exposed to E. coli [40], likely by acting on the Fc receptor for IgA on neutrophils.

IgA is the predominant immunoglobulin at mucosal surfaces, and is most effective in neutralisation [41]; it has been implicated in both pro- and anti-inflammatory pathways [42]. IgM plays a critical role in early immune defence and is most potent in activating the complement pathway. Importantly, both natural and immune IgM also facilitate the clearance of immune complexes, apoptotic cells and cellular debris, and exhibit recognised anti-inflammatory properties [43]. Notably, each immunoglobulin class has distinct and overlapping affinities and avidities for pathogens and interacts with different Fc receptors. As a result, antibody preparations containing IgM, IgA and IgG offer enhanced and more versatile immunoprotective effects than IgG-only formulations.

Taken together, the diverse and potentially synergistic actions of IgM, IgA and IgG likely contribute to the improved outcomes observed in both ventilated and NV mice with severe pneumonia in our study. The clinical relevance of these findings is supported by previous studies showing that immunoglobulin preparations enriched in IgM and IgA are effective in treating infection-associated sepsis across various patient populations [44–47]. These data also align with post hoc analyses from the phase 2 CIGMA trial, which indicated that trimodulin treatment led to a more rapid and sustained normalisation of inflammation markers, neutrophil counts, and levels of C-reactive protein and procalcitonin [8].

We further explored the anti-inflammatory and barrier-stabilising effects of trimodulin in our VILI mouse model [48, 49]. Here, it reduced neutrophil recruitment and improved lung compliance, highlighting its anti-inflammatory capability in noninfectious lung inflammation and injury.

Despite effective antibiotic therapy, pneumonia can lead to complications like sepsis, acute respiratory distress syndrome and multiorgan dysfunction [3]. Various pathophysiological mechanisms contribute to this deterioration, including bacterial components damaging alveolar epithelial and endothelial cells, bacterial spreading, excessive inflammation and organ injury. Current adjunctive therapies focus on either enhancing antimicrobial coverage or reducing inflammation; however, the former may be insufficient, the latter risks secondary infections. An ideal adjunctive therapy would enhance immune-mediated pathogen killing, limit hyperinflammation and improve lung barrier function. Our experimental data suggest that trimodulin possesses these three key properties.

The main limitations of the murine studies presented here are the anatomical, physiological and immune system differences between mice and humans, which may affect the applicability of our findings to clinical outcomes. While both species share basic immunoglobulin isotypes (IgG, IgM, IgA, IgD and IgE), variations in subclass function and Fc receptor specificity can lead to divergent immune responses and therapeutic efficacy [50, 51]. Additionally, the mice analysed in our studies were young, female and healthy, which is not representative of typical human pneumonia cohorts, and thus warrants follow-up studies [52]. Moreover, the progression of pathological processes differs significantly; for example, acute lung injury develops more rapidly in mice [53]. In our study, mice developed VILI after just 4 h of ventilation, whereas in humans, VILI typically arises from cumulative lung damage over a longer time period. To extend the observation window for mechanistic analysis, we used a prophylactic treatment strategy, despite its limited clinical relevance. Pneumonia was evaluated in time windows at a maximum of 72 hpi, which may not capture late-stage inflammatory alterations in cell populations or in fibrosis, both being key features of advanced human pneumonia that are likely to be underrepresented in our murine models. Nonetheless, our murine protocols provide maximum experimental control over confounding variables in vivo while effectively modelling lung injury from various causes.

Another limitation of our study is the exclusive use of S. pneumoniae as the infectious agent. Although S. pneumoniae is the most common cause of CAP, its pathophysiological effects and the associated immune response differ from those induced by other bacterial, viral or fungal pathogens. While bacterial and viral infections share common immune mechanisms, such as antibody-mediated neutralisation, opsonisation and complement activation, their reliance on these mechanisms differs. Additionally, pneumococcal elimination depends more heavily on phagocytosis, whereas viral elimination relies more on cytotoxic responses [54]. Therefore, our findings should not be generalised to pneumonia caused by other microorganisms [55].

Our findings highlight the potential of trimodulin as a multifaceted therapeutic agent against lethal pneumococcal pneumonia in murine models. The ability of trimodulin to reduce bacterial burden through enhanced opsonophagocytosis, mitigating harmful inflammation, and to maintain lung barrier integrity underscore its promising role in critical care. Further research is necessary to fully understand its mechanisms of action and to optimise its use, especially in combination with other antimicrobial agents, to maximise therapeutic benefits.

Supplementary material

Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.

Supplementary material

ERJ-00392-2025.Supplement

Data availability

Publicly available datasets used in this manuscript can be found at GEO under accession number GSE284045.