Development of a Reproducible Model of Murine Pneumonia with Diverse Strains of Penicillin-Resistant Streptococcus pneumoniae (PRSP) Beatriz E. Salazar, Maria Agudelo, Carlos A. Rodriguez, Andrea Restrepo, Andres F. Zuluaga, Omar Vesga GRIPE 1 , Section of Infectious Diseases, University of Antioquia Medical School, Medellín, Colombia. Streptococcus pneumoniae leads the etiology of acute respiratory tract infection and invasive disease, killing each year more than a million children under 5 years of age. Although this organism was susceptible to penicillin for almost half a cen- tury, the past two decades have witnessed an alarming increase in the number of strains resistant to this and other antibiot- ics. In Colombia, penicillin resistance among pediatric invasive isolates increased from 10 to 56% between 1994 and 2001. Although excellent animal models of pneumonia and sepsis are available for penicillin -susceptible strains, models of pneu- monia with PRSP lack reproducibility, mainly because PRSP strains display low virulence in rodents. In the present study, we addressed this particular problem to obtain a model of PRSP pneumonia that is simple, practical, reliable and reproduci- ble for in vivo evaluation of antibiotics. Fundamental for success with this model was the development of a supplemented culture medium to optimize bacterial growth in the infecting inoculum, and its further modification to enhance virulence and induce an infection closely resembling pneumococcal pneumonia in humans. Background: experimental murine pneumonia can be induced with most susceptible strains, but results with PRSP are much harder to reproduce. By optimizing in vitro growth of PRSP, we developed a reproducible murine model of pneumonia, useful to evaluate in vivo efficacy of antibiotics. Meth- ods: we used 6 clinical strains of PRSP representing serotypes 19F, 9V, 14 and 6B as well as S. pneumoniae ATCC 49619 as standard strain. After opti- mization of culture conditions to obtain maximal growth of PRSP, neutropenic animals were infected by nasal instillation of 50 µL log-phase bacteria. MPF Udea:ICR(CD-1) female mice were 6 weeks-old and weighted 25±2g when inoculated with 10 7.7-8.4 log 10 CFU/mL of each strain. Groups of 3 mice were sacrificed and their lungs removed at 1, 2, 4, 6, 12, 14, 16, 24, 36, and 48 h after infection in order to determine the dynamics of bacterial infection by appropriate homogenization, dilution, serial plating, and culture on TSA 5% blood agar. A second model, in which 50% of the inoculum was re- placed by 10% porcine mucin, was similarly evaluated. Results: culture optimization led to growth of 9-10 log 10 CFU/mL, avoiding autolysis. One hour after infection, mice had 6.20-7.79 log 10 CFU/g of lung, a count that decreased to 3.6-7.6 log 10 CFU/g at 17±5 h (nadir) and then increased up to 6.8-9.6 log 10 CFU/g at 42±6 h (zenith), giving a net growth of 2.0-3.2 logs in 25 h. Histopathology confirmed pneumonia in all animals with all PRSP strains, but most mice recovered from the infection after 5 days. Mucin enhanced the virulence of all PRSP strains by transforming the model into a uniformly le- thal pneumonia by 29 to 79 hours, with bacterial dissemination to all vital organs. Conclusion: reproducible induction of PRSP pneumonia was attained with diverse clinical strains. Virulence enhancement with 10% porcine mucin allowed a lethal pneumonia model. ABSTRACT INTRODUCTION Presentation # B-741 Control Number # 3470 Mouse virulence varied widely among PRSP strains and was not correlated with the pattern of susceptibility to penicillin or the degree of human virulence, as judged by the serotype and the specimen from which the different strains were isolated (Table 1). Also, low virulence in immunocompetent mice did not predict failure to induce pneumonia with this model (see below). RESULTS Bacteria and culture media: eight invasive strains of S. pneumoniae tested the reproducibility of this pneumonia model; S. pneumoniae ATCC 49619 was included as a reference penicillin-resistant strain. These isolates belong to a nationwide col- lection administered by Instituto Nacional de Salud in Bogotá, Colombia; all came from CSF or blood of sick patients, and included serotypes 19F, 9V, 14, and 6B; 4 were penicillin-resistant, 2 penicillin-intermediate, and 2 penicillin-susceptible. Pneumococcal strains were kept frozen at -70ºC in skim milk and resuscitated as needed. Previous to animal experiments, we optimized culture conditions for PRSP under different combinations and found a three-phased (Phases 0, 1 and 2) cul- ture method that produced an inoculum large and young enough to cause pneumonia in the mice before activation of autol- ysis mechanisms [1]. Susceptibility tests: MIC and MBC were determined by broth microdilution following CLSI and ASM guidelines for penicillin, vancomycin and ceftriaxone. Animals: 6 week-old, female, murine pathogen free (MPF) Swiss mice from the strain Udea:ICR(CD-1) weighting 23-27 g were rendered neutropenic by two intraperitoneal injections of cyclo- phosphamide (Cytoxan®, BMS, Princeton, NJ) 4 days (150 mg/kg) and 1 day (100 mg/kg) before infection. The animals were maintained and cared in accordance with international requirements and were provided sterile food and vitaminized water ad libitum. Murine virulence: to see if successful induction of pneumonia was dependent on murine virulence, we de- termined such trait by 3 successive passes of each strain, inoculating ~10 7 CFU in the thighs of immunocompetent male mice of the same strain, age and weight of those employed in the pneumonia model. Bacteria were quantified 24 h later in the thighs, lungs, liver, kidneys and spleen to verify local growth and dissemination capacity of each strain; furthermore, we designed a virulence scale based in clinical signs of dissemination (local infection, edema, sepsis, or death within 24 h), grading dead mice with 4+, sick mice with signs of systemic infection with 3+, mice with local but not systemic signs of infec- tion with 2+, and healthy mice with 1+. Experimental pneumonia: mice were first anesthetized by 0.1 mL intramuscular in- jection of a mixture containing respectively 100+10 mg/kg ketamine (Ketalar®, Parke-Davis, Quito, Ecuador) and xylazine (Rompun®, Bayer, Sao Paulo, Brazil), then inoculated by intranasal instillation of 50 µL bacterial suspension from Phase 2, and finally suspended vertically for 10 minutes hanging from their incisors on a nylon string. Infected mice were sacrificed in groups of 3 animals at 1, 2, 3, 4, 5, 6, 12, 18, 24, 32, 40, and 48 h after infection to determine the dynamics of bacterial growth in the lung; another group of 3 mice was left to evaluate survival 120 h after infection. Once the nadir and the zenith of bacterial growth in the lung were established, additional time-points included 14, 16, 21, 30, 36, 42, 60, and 120 h after infection. Histopathology studies: lungs were removed at the time-point defined as the zenith of bacterial growth, fixed in 10% formalin buffer, sectioned, and stained with hematoxilin-eosin and Gram for microscopic confirmation of the pneumonia process. Figure 2. Impact of inoculum growth phase on virulence of S. pneumoniae INS-E611 in the lungs of neutropenic mice after intranasal instillation Table 1. Microbiological characteristics of PRSP strains employed in the animal model We also found that early log-phase bacteria must be inoculated to the neutropenic animals in order to induce pneumonia. Bacteria in early log-phase started at 7.02 log 10 CFU per mouse (50 µL inoculum), decreased to a nadir 24 h later (3.65 ± 0.12 log 10 CFU/g), and reached the zenith at 48 h (6.18 ± 0.43 log 10 CFU/g), with a net growth of 3.16 log 10 CFU/g be- tween hours 24 and 48. On the contrary, late log-phase bacteria started at 6.39 log 10 CFU per mouse, but grew only 0.49 log 10 CFU/g between hours 24 (nadir: 5.78 ± 0.62) and 48 (zenith: 6.27 ± 0.42 log 10 CFU/g). This incapacity of late log- phase bacteria to grow in lung tissue did not change by washing the cells and replacing their supernatant by sterile sa- line: each mouse was inoculated with 7.05 log 10 CFU, but bacterial growth in lung tissue did not take place (Figure 2). Figure 3. Growth dynamics of 5 PRSP strains in the lungs of neutropenic mice after intranasal instillation of a pure culture (blue squares and line) or a 1:1 mix of the same inoculum plus 10% porcine mucin (red circles and line) Pneumonia was effectively established with all PRSP strains tested (INS-E611, E674, E683, E684, and ATCC 49619) with an optimized early log-phase inoculum of 6.25 - 7.32 log 10 CFU per mouse. The nadir and the zenith were observed respectively 12-18 and 36-42 hours after infection, with a growth of 1.51 to 3.08 log 10 CFU/g. However, most mice cleared the lung infec- tion and recovered their health without treatment after 120 hours. Lethal pneumonia was induced in 80 -100% of the animals infected by adding 10% porcine mucin to the bacterial inoculum in a 1:1 proportion (final mucin concentration, 5%), and in most cases it also improved the net growth of PRSP in the lungs (Figure 3). RESULTS (CONT.) Table 2. Histopathology analysis of mice lungs after intra-nasal instillation with PRSP, mucin, or PRSP plus mucin CONCLUSION  The optimization of PRSP culture conditions allowed us to obtain an inoculum that was highly virulent for mice and less prone to autolysis. After direct nasal instillation to neutropenic mice, these bacteria consistently grew in the lungs more than 2 log 10 CFU/g in 24 hours, causing pneumonia with all strains tested. Mixing the inoculum in a 1:1 proportion with 10% porcine mucin caused a progressive, lung destructive, uniformly fatal infection.  The model with PRSP alone can be applied to antibiotic dose-effect studies which endpoint is the bacterial counts re- maining in the lungs after treatment. Experiments with survival endpoints cannot be designed with such model, be- cause animals recover spontaneously once the neutrophil count starts to normalize (day 5 after infection).  The model with PRSP plus 10% mucin (1:1) is uniformly lethal with strains belonging to different serotypes and dis- playing different murine virulence, a useful characteristic for survival or bacterial counts endpoints.  There is a relatively wide variation in the nadir points with this pneumonia model due to the different murine virulence of PRSP strains. Given that the nadir signals the time to start treatment in dose-effect experiments, it is always advisa- ble to run a pilot experiment to determine the nadir and the zenith for each particular strain. Addition of mucin to the bacterial inoculum also helps to reduce this source of variation. REFERENCES 1.Restrepo AV, Salazar B, Agudelo M, Rodriguez CA, Zuluaga AF, O Vesga. Optimization of culture conditions to obtain maximal growth of penicillin resistant Streptococcus pneumoniae. BMC Microbiol 2005; 5:34. Available from: URL: http:// www.biomedcentral.com/content/pdf/1471-2180-5-34.pdf. 2.Zuluaga AF, Salazar BE, Rodriguez CA, Zapata AX, Agudelo M, Vesga O. Characterization of neutropenia induced by a sim- plified low-dose cyclophosphamide regimen in outbred mice. BMC Infectious Diseases. Submitted, November 2005. ACKNOWLEDGEMENTS This study was funded by research grants from (1) Colciencias No. 1115-04-12981, (2) University of Antioquia, and (3) Rodrigo Vesga-Meneses Scientific Foundation No. UA-013. Time after infection (h) 01 12 24 48 Bacterial load in the lungs (log 10 CFU/g) 3 4 5 6 7 8 3 4 5 6 7 8 Nadir A Nadir B Nadir C Inoculum A: Early log-phase Inoculum B: Late log-phase Inoculum C: Late log phase (washed) The duration of agranulocytosis is critical for this model. Figure 1 shows that our protocol for mouse immunosuppression was successful at rendering the animals absolutely neutropenic by the day of intranasal instillation of S. pneumoniae, but this effect lasted only 4 days. It implies that pneumonia should be firmly established in that period, and that mortality should be assessed not later than day 4 after infection if selected as an endpoint [2]. Figure 1. Neutrophil count obtained from 6 week-old Swiss Udea:ICR(CD-1) MPF mice after immunosuppression with 2 intraperitoneal injections of cyclophosphamide (n=10 female mice) MATERIALS AND METHODS Table 2 shows the morphological study of lung tissues from agranulocytic mice with clinical signs of pneumonia sacrificed 38 hours after intranasal instillation of (A) 5% mucin, slight histological changes compatible with aspiration chemical pneumonit- is, these animals never showed signs of disease; (B) S. pneumoniae INS-E611, lymphocytic interstitial and hemorrhagic pneumonitis, these animals looked sick, but recovered spontaneously 5 days after infection; and (C) S. pneumoniae INS- E611 plus 5% mucin, the lungs showed much more severe damage than that described for mice instilled with PRSP alone, with extensive septum edema, necrosis, and destruction of the alveolar structure. Polymorphonuclear infiltrate and subse- quent lung consolidation is not seen in this model because the animals are rendered severely neutropenic with cyclophospha- mide, as described above. S. pneumoniae Strain Strain Serotype Diagnostic Specimen MIC/MBC (mg/L) Mouse Virulence Penicillin Ceftriaxone Vancomycin INS-E611 6B Blood 2.00/2.00 1.00/1.41 0.35/0.35 ++++ INS-E674 14 Blood 2.00/2.00 1.00/1.00 0.25/0.25 + INS-E676 14 CSF 1.00/1.41 1.00/1.41 0.35/0.70 + INS-E678 14 CSF 0.12/0.12 0.03/0.03 0.25/0.25 + INS-E682 6B CSF 0.01/0.01 0.01/0.01 0.25/0.25 ++++ INS-E683 9V CSF 1.00/1.00 0.35/0.35 0.35/0.35 ++ INS-E684 14 CSF 1.00/1.00 0.50/0.50 0.25/0.25 ++ INS-E685 1 Blood 0.01/0.01 0.01/0.01 0.25/0.25 + ATCC 49619 19F Sputum 0.46/0.46 0.06/0.12 0.33/0.44 +++ RESULTS (CONT.) Lung Pathology 5% Mucin PRSP PRSP: 5% Mucin Congestion Mild Severe Severe Atelectasia A few foci with lymphocytes Areas Extensive, with proteinaceous material with mononu- clear infiltrate Alveolar distension Absent Focal Plus focal alveolar edema Septum edema with infiltrates of mononuclears and megacariocytes A few lymphocytes Focal, perivascular and peribrochiolar Extensive, gross septum with lymphocyte and mono- nuclear infiltrate and perisvascular bacteria; dis- persed bacterial groups in alveolar septa Intra-alveolar hemorrhage or erythrocyte extravasa- tion Absent Focal intra-alveolar hem- orrhage Extensive, with perivascular erythrocyte extravasa- tion Terminal bronchial epithe- lia Unaffected Unaffected Ulcerated with lymphoid infiltrate and bacterial accu- mulation Gram-Positive Cocci (GPC) Absent 2 of 3 mice Abundant GPC accumulated within fibrin, forming proteinaceous material, heavier around terminal bronchioli, but present intra-alveolar. No epithelial cells mixed with bacteria or in the middle of fibrin Fibrin Absent 1 of 3 mice Abundant, forming focal hyaline membranes full of bacteria Mononuclear infiltrate Absent Only in the septum Focal, but multiple foci Necrosis Absent Absent Multiple foci of necrosis in alveolar septa, with se- vere changes in extension, number and concentra- tion of GPC Histopathology diagnosis Non-specific lym- phocytic interstitial pneumonitis Lymphocytic interstitial pneumonitis with alveolar hemorrhage and lung ate- lectasis Severe interstitial pneumonitis with abscess for- mation