Sepsis continues to be one of the most common and devastating morbidities among preterm infants, with a mortality rate of ∼40% and poor neurodevelopmental outcomes (1–3). It is characterized by a systemic bacterial invasion followed by inflammatory responses. The pathogenesis of sepsis is believed to be mediated by multiple factors including oxidative stress and apoptosis (4). Distinct differences in immune responses between neonates and adults have been reported, possibly contributing to the higher mortality observed in neonatal sepsis (5, 6). In particular, in human preterm infants, host defense maintained by the innate immunity is immature and not fully intact as in term infants (7). In fact, the mortality rate from sepsis in preterm infants has not significantly decreased (3), and may be partly due to the unavailability of appropriate animal models to study its pathophysiology.
Several animal models have been developed to study sepsis. The cecal ligation and puncture (CLP) model is the widely used “gold standard” in adult rodents (8). However, it has some limitations, such as large investigator variability (2) and difficulty in comparing animals with variable cecum shapes and sizes (9). For neonatal mice, the use of a surgical procedure is not feasible because of their fragility and small size. Recently, an alternative “cecal slurry” (CS) model was established by Wynn et al. (2, 10), which involves the intraperitoneal (IP) administration of adult cecal contents suspended in dextrose. This model is regarded to appropriately mimic the septic condition in a human neonate (9), but has been used only in moderate-sized newborn pups (5–7-days old) (2, 10), whose immune system is comparable to that of a term human newborn (5). An attenuated anti-inflammatory response and antioxidant activity has been reported in this model when compared with adult counterparts (10). Recently, a new CS preparation protocol has been created by Starr et al. (9), which allows the storage of a stock CS preparation at –80°C without loss of bacterial viability. Here, we combined these two models in order to create a model of preterm sepsis using 4-day-old mouse pups, whose age is equivalent to that of a human preterm infant (5, 11).
Heme oxygenase (HO), the rate-limiting enzyme in the heme catabolic pathway, produces equimolar amounts of free iron, biliverdin, and carbon monoxide (CO) (12, 13). Biliverdin is rapidly reduced to bilirubin, a potent endogenous antioxidant (14). In addition, CO has protective anti-apoptotic and anti-inflammatory properties (15, 16). Among the 3 HO isoforms, HO-1 (a stress-response protein) is known to play key roles in physiologic and pathologic conditions, including adult sepsis (17, 18). The expression pattern of HO-1 is developmentally regulated and variable between individual organs (19, 20). In addition, a recent report from our laboratory has demonstrated a protective role of HO-1 against neonatal intestinal inflammation (21). Moreover, HO-1 promoter polymorphisms (i.e., those that lead to increased HO-1 expression) have been implicated as protective in pediatric sepsis (22), which suggests that an induction of HO-1 may be involved in defense mechanisms of this immunologically immature population.
In this study, we first validated our modified CS model by characterizing the progression and severity of preterm sepsis. Then, we investigated if the induction of HO-1 can attenuate the severity of preterm sepsis.
MATERIALS AND METHODS
FVB wild-type (WT) breeders (6–8-weeks old) were obtained from Charles River Laboratories (Wilmington, Mass). Mice were maintained on standard rodent food and water ad libitum. Pups were kept with their mothers throughout the course of the study. For each study, pups were randomized on an individual basis within each litter in order to eliminate any litter bias effects. All studies were approved by Stanford University's Institutional Animal Care and Use Committee (Protocol # 30910). The experiments were performed in adherence to the National Institutes of Health Guidelines on the Use of Laboratory Animals.
Stock CS preparation
Adult WT FVB mice were sacrificed and whole cecums were harvested. Cecal contents were collected, pooled, weighed, and then mixed with 0.5 mL of sterile water per 100 mg of cecal content. The mixture was filtered through a 100-μm mesh strainer and then mixed with an equal volume of 30% glycerol in PBS. One milliliter of CS stock was transferred to cryovials and stored at −80°C, which has been shown to keep bacterial viability for ∼6 months (9). For each study, a CS aliquot was thawed and vortexed prior to injection.
To induce sepsis, 4-day-old WT FVB pups were given various doses [1.0–4.0 mg/g body weight (BW)] of CS IP and then closely monitored for health and survival for 7 days. When pups show the following signs of impending death: pale coloring, scattering (positioning away from remainder of the litter), or absence of milk in the stomach (failure to feed) (2), they were euthanized.
At 24 h (i.e., 5 days of life) post-sepsis induction, livers, lungs, and spleens were removed and rinsed with ice-cold 0.1 M KPO4. One hundred milligrams of each tissue was diluted ×10 with 0.1 M KPO4. For spleen samples <20 mg, 200 μL of KPO4 was added. Tissues were sonicated at 50% power with a Microson Ultrasonic Cell Disruptor (Misonix, Farmingdale, NY). The sonicates were diluted serially in PBS as follows: liver: undiluted, 1:10, and 1:100; lung: undiluted and 1:10; and spleen: undiluted, 1:10, 1:100, and 1:1000 (23). Fifty microliters of samples was then streaked onto agar plates containing BHI broth as described previously (9). Bacteria were grown at 37°C for 24 h and quantified as CFU/mL as follows: dilution factor × CFUs × 1000 μL/mL divided by 50 μL.
Hematology and serum biochemistry
At 24 h (i.e., 5 days of life) post-sepsis induction, pups were sacrificed via CO2 inhalation and 30 to 80 μL of blood pooled from two or three pups was either immediately collected after decapitation by directly placing the pup body (without squeezing) over a Microtainer tube containing EDTA (BD Sciences, San Jose, Calif) for complete blood counts (CBCs) with automated differentials and blood smears or in lithium-heparin gel tubes (Capiject, Terumo Medical Corp, Elkton, Md) to separate red blood cells for serum biochemistry (serum aspartate aminotransaminase [AST] and blood urea nitrogen [BUN]) measurements. All measurements were performed by the Diagnostic Laboratory in the Department of Comparative Medicine (Stanford, Calif).
At 24 h (i.e., 5 days of life) post-sepsis induction, pups were first sacrificed by CO2 inhalation and then the pulmonary artery was perfused with PBS. Livers and lungs were harvested, sectioned, and placed in 10% (vol/vol) neutral buffered formalin (Fisher Scientific, Waltham, Mass) for 24 h. Fixed tissues were then embedded in paraffin according to standard protocols. Tissues 6-μm thick were cut from paraffin-embedded blocks by use of a microtome. Following deparaffinization, sections were stained with H&E (American Master Tech Scientific, Lodi, Calif) and histological changes were visualized by light microscopy (Nikon Eclipse E800, Tokyo, Japan). Liver injury was assessed at ×400 magnification by the presence of cytoplasmic vacuolization in hepatocytes according to the method by Leelahavanichkul et al. (24). In brief, in three randomly selected, nonoverlapping fields, number of hepatocytes with vacuoles was counted and scored blindly using a 4-point scale as follows: 0: no cells with vacuolization; 1: <10 cells with vacuolization; 2; 10 to 20 cells with vacuolization; and 3: >20 cells with vacuolization. The three scores were then averaged for each slide. Lung injury was assessed at ×100 magnification by the presence of increased lung air space, which can be due to impaired postnatal alveolar development (25). In brief, percentage of air space area was measured by using ImageJ (National Institutes of Health, Bethesda, Md) in three randomly selected, nonoverlapping fields as previously described (26), and was expressed as a mean±SD percentage of the total area.
At 6 h (i.e., 4 days of life) post-sepsis induction or vehicle (Veh) treatment, pups were sacrificed and 5 × 5 × 1-mm pieces of liver and spleen were immediately placed in liquid nitrogen and stored at –80°C until use. Total RNA was extracted according to standard laboratory procedures using the RNAeasy Mini Kit (Qiagen, Valencia, Calif). cDNA was synthesized by use of a RT2 First Strand Kit (Qiagen). PCR array kits (Qiagen) for mouse Innate & Adaptive Immune Responses (catalog no. PAMM-052Z), which screens for 84 genes, were used. Real-time PCR was performed with RT2 Real-Time SYBR Green/ROX PCR Master Mix (Qiagen) on a Mx-3005 Quantitative PCR System (Stratagene, Cedar Creek, Tex). Fold changes in gene expression levels in CS or Heme + CS pups over Veh levels, and Heme + CS pups over CS levels were then calculated using the ΔΔCt method as previously described (27).
Twenty-four hours (i.e., 4 days of life) after heme administration, liver (100 mg), lung (100 mg), and spleen (20 mg, pooled from 2 to 4 pups) were harvested, and then rinsed and sonicated in phosphate buffer. HO activity was then determined by gas chromatography and calculated as pmol CO/h/mg fresh weight (FW) as previously described (28).
Statistical analyses were performed using log-rank test for survival curves, unpaired Student two-tailed t test for PCR array analyses, Mann–Whitney test or chi-square test for comparisons of two groups, and Kruskal–Wallis test with Dunns multiple comparison test for comparisons of three groups or more. Differences were deemed statistically significant when P <0.05.
Non-surgical preterm sepsis model
To induce sepsis, 4-day-old WT FVB pups were given various doses (1.0–4.0 mg/g BW) of the stock CS solution IP and then closely monitored for health and survival for 7 days. In pilot studies, consistent mortality rates in pups given 1.5 mg/g CS prepared from fresh cecums harvested from adult mice could not be achieved (data not shown). Therefore, to ensure that each pup received a consistent CS preparation for each experiment, we used the storage protocol described by Starr et al. (9). In brief, cecal contents from 5 to 10 freshly harvested cecums from adult FVB mice (6–8-week-old) were pooled and suspended in 15% glycerol/PBS. 1.0-mL aliquots were then transferred to 2.0-mL vials, which were then placed in a Cryo 1°C Freezing Container (Nalgene, Rochester, NY) for 2 h at –80°C, removed, and then further stored at –80°C until use.
To establish the optimal dose for sepsis induction, 1.0 to 4.0 mg/g BW of CS was administered to 4-day-old newborn WT FVB pups. At least three different litters were used for each dose. Health of pups (such as pale coloring, scattering, and absence of milk in the stomach), BW, and survival were monitored daily for 7 days. We found a dose-dependent increase in mortality with increasing CS dose. The LD40 was interpolated to be 2.0 mg/g (red dotted line, Fig. 1A). There was also inverse relationship between average survival days and CS dose, with survival decreasing with increasing dose. Because the mortality rate for human preterm infants with early-onset sepsis is reported to be ∼40% (3), a CS of 2.0 mg/g (LD40 established above) was used for all subsequent studies.
BW changes post-sepsis induction
In order to monitor the progression of sepsis development, BWs were recorded daily for 7 days beginning 24 h after CS administration. We found that BWs of CS-treated (CS) pups significantly decreased (–2.1±6.2%, n = 21) compared with age-matched vehicle-treated (Veh) controls (21.5 ± 8.2%, n = 29, P <0.001, Fig. 2). When we categorized pups based upon BW changes, BW loss correlated with increased mortality (43.8%, n = 16) compared with BW gain (20.0%, n = 5), but was not statistically significant (P = 0.34).
Organ bacterial colonization post-sepsis induction
Subsets of pups were sacrificed 24 h post-sepsis induction to assess the extent of bacterial colonization in selected organs. Livers, lungs, and spleens were harvested, placed in phosphate buffer, and then sonicated. Tissue sonicates were then serially diluted with sterile PBS, streaked onto BHI-agar plates, and incubated at 37°C for 24 h. CFUs on each plate were then counted. Moderate variations in colony sizes and CFU counts were observed, which is consistent with a polymicrobial infection. Bacteria were equally detected in all organs (n = 9, for all organs, Fig. 3).
Hematological changes post-sepsis induction
Twenty-four hours post-sepsis induction, white blood cells (WBC) and platelet counts were found to be significantly lower in CS pups (2,821 ± 931 and 204 ± 82 ×103 per μL, n = 8) compared with Veh pups (6,298 ± 1,806 and 613 ± 118 ×103 per μL, n = 6, P <0.01, respectively, Figs. 4, A and C). Hemoglobin levels were not statistically different between the groups (8.2 ± 0.6 vs. 8.1 ± 0.5 g/dL, P = 0.85, Fig. 4B). When we compared the levels of neutrophils and lymphocytes, which play key roles in maintaining immune responses, both cell types were significantly lower in CS pups (751 ± 511 and 1,854 ± 849 per μL, respectively, n = 8 for each group) versus that of Veh pups (2,891 ± 910 and 2,937 ± 1,232 per μL, n = 6, P < 0.01, and = 0.04, respectively, Figs. 4, D and E).
Serum biochemistry post-sepsis induction
To determine the effects of sepsis induction on the liver and kidney, serum AST and BUN levels, respectively, were measured at 24 h post-sepsis induction. AST levels in CS pups (229 ± 95 U/L, n = 13) were 34% higher compared with Veh pups (171 ± 52 U/L, n = 9, P = 0.12), though statistical significance was not reached. However, when we categorized the pups based on BW changes as above, a significant 1.8-fold difference in AST levels between pups with BW loss (265 ± 91 U/L, n = 9) compared with those without BW loss (146 ± 30 U/L, n = 4, P = 0.02, Fig. 5A) was found.
In addition, BUN levels of CS pups (65.5 ± 26.8 mg/dL, n = 13) were higher compared with Veh pups (52.0 ± 5.4 mg/dL, n = 6, P = 0.56), although not statistically significant. However, when we categorized the pups based on BW changes, there was a significant difference between pups with BW loss (76.8 ± 29.2 mg/dL, n = 8) compared with those without BW loss (47.6 ± 3.2 mg/dL, n = 5, P <0.01, Fig. 5B).
Histological changes post-sepsis induction
To assess histological changes 24 h post-sepsis induction, lungs and livers were harvested, fixed with formalin, and stained using hematoxylin and eosin (H&E). In livers of newborn mice, there was an abundance of hepatoblasts and hepatocytes, and parenchymal cells demonstrated various degrees of vacuolization, which occurs naturally in developing mouse livers (29). No massive neutrophil infiltration was detected in either CS or Veh groups. When we compared the degree of vacuolization defined as mean vacuolization score as a marker of liver injury (24), no significant difference between the CS (5.3 ± 1.6, n = 7) and Veh (6.0 ± 2.5, n = 6, P = 0.70, Fig. 5C) pups was found. In lungs of newborn mice, there were differences in saccular structure and alveolar formation, and no massive neutrophil infiltration was detected in either group. We then compared the percentage of air space area to gauge alveolar space enlargement and alveolar wall thinning as a means to assess impaired postnatal alveolar development (25). We found no significant difference between CS (55.7 ± 7.4%, n = 7) and Veh (54.3 ± 13.6%, n = 6, P = 0.90, Fig. 5D) pups.
Expression profiles of genes involved in innate and adaptive immunity post-sepsis induction
To determine expression levels of genes involved in innate and adaptive immune responses, we isolated RNA from livers and spleens harvested 6 h post-sepsis induction. Using RT-PCR arrays, we found that of the 84 genes screened by the kit, 15 gene transcripts significantly increased and 1 gene transcript significantly decreased in livers of CS compared with Veh mice (fold change >4.0, P <0.05, Table 1). The increased genes were cytokines (Ccl5, Cxcl10, IL-1a, IL-1b, Tnf, and Ifng), pattern-recognition receptors (Cd14, Nod2, Tlr2, and Nlrp3), and other several related to immune response (Cd40, Icam1, Mx1, Myd88, and Nfkbia). The only gene with decreased expression was the pattern-recognition receptor of Tlr5. On the other hand, eight gene transcripts significantly increased and no gene transcript significantly decreased in spleens, (fold change >4.0, P <0.05, Table 1). Interestingly, the genes that increased in spleens (except IL-10) matched those that increased in livers, including cytokines (Cxcl10, IL-1a, IL-1b, IL-10, and Ifng), pattern-recognition receptors (Cd14 and Nlrp3), and other related to immune response (Nfkbia). The anti-inflammatory cytokines, IL-4 and IL-13, did not significantly change (data not shown).
Induction of HO-1 is protective against the development of preterm sepsis
HO-1 induction decreased mortality rate in preterm sepsis
To induce HO-1 expression, we gave 3-day-old mice 30 μmol of heme/kg BW subcutaneously (SC), a dose we have previously shown to increase liver HO activity (28), 24 h prior to sepsis induction. Twenty-four hours after heme administration, we found that HO activity (expressed as pmol CO/h/mg FW) in the liver (496 ± 85, n = 10) and spleen (510 ± 91, n = 5) significantly increased 64% and 50% over age-matched controls (302 ± 41, n = 23, P <0.001 and 341 ± 13, n = 4, P = 0.02, respectively). However, slight increases in lung HO activity were observed (134 ± 15 vs. 123 ± 10, n = 10 in both group, P = 0.05, Fig. 6).
To determine if the induction of HO-1 has a protective role against preterm sepsis, we then administered heme (30 μmol/kg BW) to 3-day-old pups 24 h prior to sepsis induction. We then monitored pup health, BW, and survival daily for 7 days as described above. When we compared the mortality rates of heme-treated (Heme + CS) to CS pups, we found that the induction of HO-1 significantly reduced mortality from 40.9% (CS, n = 22) to 6.3% (Heme + CS, n = 32, P < 0.005, Fig. 1B).
HO-1 induction improved BW gain in preterm sepsis
To assess the effect of HO-1 induction on BW change, we compared BW changes between Heme + CS and CS pups. We found a significant increase in BW in Heme + CS pups (7.0 ± 10.0%, n = 32) compared with CS pups (–2.1 ± 6.2%, n = 21, P <0.05, Fig. 2). However, BW gain in Heme + CS pups (7.0 ± 10.0%, n = 32) was significantly lower than that of Veh pups (21.5 ± 8.2%, n = 29, P <0.001).
HO-1 induction decreased bacterial colonization in preterm sepsis
To investigate the effect of HO-1 induction on bacterial colonization, we compared CFU counts in each organ 24 h post-sepsis induction between Heme + CS and CS pups. We found no significant differences in CFU counts (expressed as CFU/mL) in the liver (Heme + CS: 5.82 ± 6.27 × 105 (n = 11) vs. CS: 6.20 ± 7.98 × 105 (n = 9), P = 0.98) and lung (1.40 ± 2.76 × 105 (n = 11) vs. 2.29 ± 3.26 × 105 (n = 9), P = 0.23) between groups. However, CFU counts in spleens of Heme + CS pups were significantly less than that of CS pups (1.81 ± 4.12 × 105 (n = 11) vs. 4.19 ± 9.15 × 106 (n = 9), P = 0.03, Fig. 3).
HO-1 induction attenuated pro-inflammatory gene expressions in preterm sepsis
We then compared gene expression profiles of Heme + CS and CS pups 6 h post-sepsis induction. We found that most of the genes upregulated in livers and spleens post-sepsis induction were down-regulated after heme pretreatment. Particularly, four cytokines (Ccl5, Cxcl10, IL-1b, Ifng) and two other genes related to immune response (Mx1, Nod2) were significantly down-regulated (>2-fold) in livers of Heme + CS compared with CS pups. In spleens of Heme + CS pups, only CxCl10 was significantly down-regulated (>2-fold) compared to CS pups (P <0.05, Table 2).
Because preterm sepsis is still one of the most intractable morbidities for neonates, understanding its pathophysiology is important. In this study, we adapted the non-surgical CS mouse model described by Wynn et al. (2) to 4-day-old mice, an age comparable to human preterm infants (5, 11, 28). During the first week of life, the proportions of leukocyte subpopulations involved in innate immune responses dramatically change, being initially immature and then maturing to adult levels by 7 days of life (6). Thus, neonatal mice aged 5- to 7-day-old used in previous studies (2, 10) have an immune system comparable to that of human term neonates, and may not be ideal for the study of sepsis in preterm infants (5, 11). Until now, few neonatal or preterm sepsis mouse models have been developed, using intravenous (IV) or intramuscular (IM) injections of a single strain of bacteria or LPS to suckling mice (23, 30). However, the disadvantages of these models are that bolus bacterial infusions do not accurately simulate the human circumstance because a septic focus that persistently challenges the body with bacteria or the diversity of infectious agents in human sepsis is not present (8). In addition, the hypodynamic cardiovascular state immediately caused by a bolus injection of LPS does not mimic hemodynamic changes observed in human sepsis (31).
The CS model is a non-surgical and less invasive sepsis model, which was originally proposed in adult pigs and adapted to mice by Wynn et al. (2). In addition to its less invasive nature, the advantage of this model is that the IP route of administration causes abdominal abscesses, a focus of infection (the usual cause of human sepsis) and is polymicrobial in nature, which is similar to that in the CLP model (32). Major disadvantages of the CS model are that adult cecums need to be freshly harvested for each experiment and there is a lack of a necrotic tissue component as well as the clinical abdominal sepsis due to ulcer or diverticula rupture compared with the CLP model. For example, in our preliminary experiments using this model, we observed inconsistencies in bacterial viability, which resulted in variable mortality rates. In 2014, Starr et al. (9) published a protocol that describes the long-term storage of a stock CS preparation, such that the same CS can be used over the course of a study to increase reproducibility and decrease inter-CS variability. We therefore combined these procedures, which allowed us to have consistent bacterial viabilities and mortality rates for a given CS dose to newborn pups.
To validate our preterm sepsis model, we first demonstrated a dose-dependent mortality as a function of CS dose. In addition, we found consistent bacterial colonization in each organ, which is similar to the bacterial load observed in livers and spleens of neonatal mice using a model of infection (23). In addition, septic pups showed hematological changes such as leukocytopenia, thrombocytopenia, and lymphocytopenia (all consistent with the presence of an ongoing systemic infection) as well as partial damage in livers and kidneys. Associated with these findings, we also observed an increase in gene expression profiles of cytokines, pattern-recognition receptors, and other genes related to immune responses. Interestingly, most of the cytokines with significant increases were those involved in pro-inflammatory responses (Ccl5, CxCl10, Ifng, IL-1a, IL-1b, and Tnf), with the exception of IL-10, an anti-inflammatory cytokine, which significantly increased only in the spleen. Other anti-inflammatory cytokines (IL-4 and 13) did not increase post-sepsis induction. Taken together, we postulated that the development of severe sepsis in our preterm sepsis model was caused by systemic inflammatory responses rather than specific organ failure. Further studies are warranted to fully substantiate this speculation as disparities in previously undescribed gene expression profiles of isolated single organs and circulating leukocytes, especially in mouse models of inflammation using inbred strains, and their translation to human disease (33, 34).
Because HO-1 has been known to possess anti-inflammatory as well as antioxidant and anti-apoptotic properties (16), we then investigated if the induction of HO-1 was protective against preterm sepsis. HO-1 is believed to play roles in macrophage survival and function, and also known to mediate anti-inflammatory IL-10 (16). In addition, we have previously reported that induction of HO-1 significantly increases the ratio of regulatory T-cells to effector T-cells in a necrotizing enterocolitis (NEC) model, and mediates protection against intestinal inflammation (21).
Several investigators have investigated the role of HO-1 in adult sepsis. Chung et al. (15) have shown an enhanced mortality in adult HO-1 knockout mice undergoing CLP in comparison with WT mice. In addition, HO-1 deficiency has been reported to be associated with increased disease severity in newborn mouse models of bronchopulmonary dysplasia (20) and NEC (27). Recently, our laboratory reported not only an increased susceptibility to intestinal injury in HO-1-deficient (HO-1+/−) mice (21); but also a decreased incidence of intestinal injury following the induction of HO-1 in WT pups prior to injury (27).
In this study, we induced HO-1 via heme administration 24 h prior to sepsis induction, and found significant HO induction in the liver and spleen, the two main organs affected in sepsis. The spleen is the site of primarily localization of bacteria and endotoxin, and the primary source of circulating pro-inflammatory cytokines (35). It has been shown that the bacterial colonization does indeed occur in the spleen following infection and is time dependent (23). We found a significantly improved survival rate, BW changes, decreased bacterial colonization in the spleen, and attenuated pro-inflammatory gene expression changes in the liver and spleen. The decrease in CFUs only in the spleen may be due to the fact that the spleen contains primarily HO-1-expressing cells and can be more induced by heme administration compared with other organs, which have an equal distribution of HO-1 and the constitutive HO-2, such as the liver and lung. Specifically, we observed significant down-regulation of four pro-inflammatory cytokines (Ccl5, Cxcl10, IL-1b, and Ifng) in the liver and one pro-inflammatory cytokine (Cxcl10) in the spleen after HO-1 induction. There were no significant changes in anti-inflammatory cytokines. We therefore speculate that HO-1 induction with heme prior to sepsis induction can attenuate pro-inflammatory cytokine production, resulting in a reduced systemic inflammatory response.
Several studies have reported using heme to induce HO-1 in the adult CLP model, but have shown conflicting results (17, 18, 36). Freitas et al. (36) reported an increase in mortality following repeat injections of hemin (15 μmol/kg) 30 min before and 6 h after sepsis induction. In contrast, Luo et al. (17) reported an attenuation of the severity of sepsis-induced acute lung injury using hemin (43 μmol/kg) 12 h prior to sepsis induction. Similarly, Fei et al. (18) reported that 12 h prior HO-1 induction via hemin pretreatment (50 μmol/kg) inhibited sepsis-induced thrombosis and suppressed the production of TNF-α and IL-6. We also observed significant improvement in survival with 24 h prior heme treatment in our preterm sepsis model. In our studies, methemalbumin (MHA), which is albumin bound to heme, was used because free heme is known for its toxicity and pro-oxidant activity (37). In addition, we administered MHA 24 h prior to sepsis induction, a time when peak HO-1 expression and HO activity is achieved as shown previously (28). The dose used by Freitas et al. may not have been sufficient to induce HO-1, and therefore did not confer any protective effects in their model. Furthermore, their time of heme administration may have been too close to the induction of sepsis and maximum HO-1 induction was not reached. This could in turn lead to an accumulation of toxic free heme, resulting in a worsening septic course and mortality. To further investigate the causal relationship about HO-1 expression and protection of sepsis, we are currently extending our studies by applying this preterm sepsis model to HO-1 heterozygous knockout mice, which have a partial deficiency in HO-1, and hence mimic the human HO-1-deficient circumstance.
In this study, we did not monitor the severity of sepsis for each pup because we could not categorize the severity based on physical examination alone as previously reported for adult rodents. Since newborn pups lack fur, have a small body size, and have subtle movements, the adult criteria of scoring sepsis severity based on posture and behavior was not possible. In addition, the physical changes in the newborn mice at this age are transient and not distinct. However, intriguingly, we found that BW loss 24 h post-sepsis induction was associated with higher mortality rates as well as increased organ damage and AST levels (especially in those with weight loss). This simple physical parameter appeared to accurately reflect the severity of sepsis, and may be a useful adjunct to using only physical signs as an assessment of morbidity in newborn pups. However, not unexpectedly, we did not see obvious histopathological changes in the livers and lungs 24 h post-sepsis induction, when pathological changes are generally observed in adult sepsis (24). It may have been due to the fact that we could not differentiate between normal physiologic cell death or apoptosis that occurs naturally during normal development in newborn mouse livers from pathologic cell death (necrosis) (29). Similarly, 5-day-old newborn mouse lungs are undergoing normal transition from a saccular to alveolar stage (38) making it difficult to see clear differences in air space areas due to damage or just normal variations of a developing mouse lung. Because liver HO activity is high during the neonatal period (19), and lung HO-1 protein is maximally expressed at birth (20), we may have observed less damage in our newborn pups compared with that seen in adults. Although a later histological examination may have been more appropriate, we could not harvest tissues later than 24 h post-sepsis induction because the majority of pups die between 24 and 48 h post-induction, especially in the severest of cases.
In summary, we conclude that our non-surgical CS model can be used for the study of the pathogenesis of preterm sepsis. Furthermore, because the induction of HO-1 significantly reduced mortality in association with a reduction of systemic inflammatory responses, we speculate that HO-1 may confer protection against sepsis in preterm infants.
The authors thank Dr Koshi Kunimoto, Dr Hidekazu Nishikii, Dr David N. Cornfield (Stanford University), and Dr Hiroshi Saito (University of Kentucky) for their invaluable advice on our experimental design. The authors thank Pauline Chu (Stanford University) for her technical expertise in preparing and staining slides for histology.
1. Wynn JL, Wong HR. Pathophysiology and treatment of septic shock in neonates. Clin Perinatol
2. Wynn JL, Scumpia PO, Delano MJ, O’Malley KA, Ungaro R, Abouhamze A, Moldawer LL. Increased mortality and altered immunity in neonatal sepsis produced by generalized peritonitis. Shock
3. Stoll BJ, Hansen NI, Sanchez PJ, Faix RG, Poindexter BB, Van Meurs KP, Bizzarro MJ, Goldberg RN, Frantz ID 3rd, Hale EC, et al. Eunice Kennedy Shriver National Institute of Child H, Human Development Neonatal Research N. Early onset neonatal sepsis: the burden of group B Streptococcal and E. coli disease continues. Pediatrics
4. Hotchkiss RS, Karl IE. The pathophysiology and treatment of sepsis. N Engl J Med
5. Adkins B, Leclerc C, Marshall-Clarke S. Neonatal adaptive immunity comes of age. Nat Rev Immunol
6. Sun CM, Fiette L, Tanguy M, Leclerc C, Lo-Man R. Ontogeny and innate properties of neonatal dendritic cells. Blood
7. McElroy SJ, Weitkamp JH. Innate immunity in the small intestine of the preterm infant. Neoreviews
8. Dejager L, Pinheiro I, Dejonckheere E, Libert C. Cecal ligation and puncture: the gold standard model for polymicrobial sepsis? Trends Microbiol
9. Starr ME, Steele AM, Saito M, Hacker BJ, Evers BM, Saito H. A new cecal slurry
preparation protocol with improved long-term reproducibility for animal models of sepsis. PLoS One
10. Gentile LF, Nacionales DC, Lopez MC, Vanzant E, Cuenca A, Cuenca AG, Ungaro R, Szpila BE, Larson S, Joseph A, et al. Protective immunity and defects in the neonatal and elderly immune response to sepsis. J Immunol
11. Hagberg H, Peebles D, Mallard C. Models of white matter injury: comparison of infectious, hypoxic-ischemic, and excitotoxic insults. Ment Retard Dev Disabil Res Rev
12. Maines MD. Heme
oxygenase: function, multiplicity, regulatory mechanisms, and clinical applications. FASEB J
13. Tenhunen R, Marver HS, Schmid R. The enzymatic conversion of heme
to bilirubin by microsomal heme
oxygenase. Proc Natl Acad Sci USA
14. Stocker R, Yamamoto Y, Mcdonagh AF, Glazer AN, Ames BN. Bilirubin is an antioxidant of possible physiological importance. Science
15. Chung SW, Liu X, Macias AA, Baron RM, Perrella MA. Heme
oxygenase-1-derived carbon monoxide enhances the host defense response to microbial sepsis in mice. J Clin Invest
16. Zhao H, Ozen M, Wong RJ, Stevenson DK. Heme
oxygenase-1 in pregnancy and cancer: similarities in cellular invasion, cytoprotection, angiogenesis, and immunomodulation. Front Pharmacol
17. Luo YP, Jiang L, Kang K, Fei DS, Meng XL, Nan CC, Pan SH, Zhao MR, Zhao MY. Hemin inhibits NLRP3 inflammasome activation in sepsis-induced acute lung injury, involving heme
oxygenase-1. Int Immunopharmacol
18. Fei D, Meng X, Zhao M, Kang K, Tan G, Pan S, Luo Y, Liu W, Nan C, Jiang H, et al. Enhanced induction of heme
oxygenase-1 suppresses thrombus formation and affects the protein C system in sepsis. Transl Res
19. Johnson JD, Wetmore DL, Martinez CW, Ostrander CR. Developmental changes in bilirubin production in the rat. J Pediatr Gastroenterol Nutr
20. Zhuang T, Zhang M, Zhang H, Dennery PA, Lin QS. Disrupted postnatal lung development in heme
oxygenase-1 deficient mice. Respir Res
21. Schulz S, Chisholm KM, Zhao H, Kalish F, Yang Y, Wong RJ, Stevenson DK. Heme
oxygenase-1 confers protection and alters T-cell populations in a mouse model
of neonatal intestinal inflammation. Pediatr Res
22. Vazquez-Armenta G, Gonzalez-Leal N, JV-dlT M, Munoz-Valle JF, Ramos-Marquez ME, Hernandez-Canaveral I, Plascencia-Hernandez A, Siller-Lopez F. Short (GT)n microsatellite repeats in the heme
oxygenase-1 gene promoter are associated with antioxidant and anti-inflammatory status in Mexican pediatric patients with sepsis. Tohoku J Exp Med
23. Kronforst KD, Mancuso CJ, Pettengill M, Ninkovic J, Power Coombs MR, Stevens C, Otto M, Mallard C, Wang X, Goldmann D, et al. A neonatal model of intravenous Staphylococcus epidermidis infection in mice <24 h old enables characterization of early innate immune responses. PLoS One
24. Leelahavanichkul A, Somparn P, Panich T, Chancharoenthana W, Wongphom J, Pisitkun T, Hirankarn N, Eiam-Ong S. Serum miRNA-122 in acute liver injury induced by kidney injury and sepsis in CD-1 mouse models. Hepatol Res
25. Massaro GD, Massaro D. Postnatal treatment with retinoic acid increases the number of pulmonary alveoli in rats. Am J Physiol
26. Zhao J, Chen H, Peschon JJ, Shi W, Zhang Y, Frank SJ, Warburton D. Pulmonary hypoplasia in mice lacking tumor necrosis factor-alpha converting enzyme indicates an indispensable role for cell surface protein shedding during embryonic lung branching morphogenesis. Dev Biol
27. Schulz S, Wong RJ, Jang KY, Kalish F, Chisholm KM, Zhao H, Vreman HJ, Sylvester KG, Stevenson DK. Heme oxygenase-1 deficiency
promotes the development of necrotizing enterocolitis-like intestinal injury in a newborn mouse model
. Am J Physiol Gastrointest Liver Physiol
28. Fujioka K, Kalish F, Wong RJ, Stevenson DK. Inhibition of heme
oxygenase activity using a microparticle formulation of zinc protoporphyrin in an acute hemolytic newborn mouse model
. Pediatr Res
29. Yu H, Bauer B, Lipke GK, Phillips RL, Van Zant G. Apoptosis and hematopoiesis in murine fetal liver. Blood
30. Weisman LE, Leeming AH, Kong L. Appropriate antibiotic therapy improves Ureaplasma sepsis outcome in the neonatal mouse. Pediatr Res
31. Buras JA, Holzmann B, Sitkovsky M. Animal models of sepsis: setting the stage. Nat Rev Drug Discov
32. Sam AD 2nd, Sharma AC, Law WR, Ferguson JL. Splanchnic vascular control during sepsis and endotoxemia. Front Biosci
33. Osuchowski MF, Remick DG, Lederer JA, Lang CH, Aasen AO, Aibiki M, Azevedo LC, Bahrami S, Boros M, Cooney R, et al. Abandon the mouse research ship? Not just yet!. Shock
34. Seok J, Warren HS, Cuenca AG, Mindrinos MN, Baker HV, Xu WH, Richards DR, McDonald-Smith GP, Gao H, Hennessy L, et al. Genomic responses in mouse models poorly mimic human inflammatory diseases. Proc Natl Acad Sci USA
35. Ge Y, Ezzell RM, Clark BD, Loiselle PM, Amato SF, Warren HS. Relationship of tissue and cellular interleukin-1 and lipopolysaccharide after endotoxemia and bacteremia. J Infect Dis
36. Freitas A, Alves-Filho JC, Trevelin SC, Spiller F, Suavinha MM, Nascimento DC, Pestana CR, Dal-Secco D, Sonego F, Czaikoski PG, et al. Divergent role of heme
oxygenase inhibition in the pathogenesis of sepsis. Shock
37. Jeney V, Balla J, Yachie A, Varga Z, Vercellotti GM, Eaton JW, Balla G. Pro-oxidant and cytotoxic effects of circulating heme
38. Amy RW, Bowes D, Burri PH, Haines J, Thurlbeck WM. Postnatal growth of the mouse lung. J Anat