Rapid onset of severe septic shock in the pregnant mouse

Julia Zöllner1,2,3, Simon Lambden3,4, Noor Mohd Nasri1,2,3, James Leiper3,5 and Mark R. Johnson1,2

1 Chelsea and Westminster Hospital, 369 Fulham Road, London, SW10 9NH, UK

2 Institute of Reproductive and Developmental Biology, DuCane Road, London W12 0NN, UK

3 Nitric Oxide Signalling Group, MRC Clinical Sciences Centre, Imperial College London, Hammersmith

Hospital Campus, DuCane Road, London W12 0NN, UK

4 Department of Medicine, University of Cambridge, 5th Floor, Addenbrooke’s Hospital, Cambridge,

CB20QQ

5 Institute of Cardiovascular and Medical Sciences, University of Glasgow, University Avenue, Glasgow

G12 8QQ

Correspondence to: Mark Johnson, Imperial College Parturition Research Group, Department of

Obstetrics and Gynaecology, Imperial College School of Medicine, Chelsea and Westminster Hospital,

369 Fulham Road, London, SW10 9NH, UK.

Telephone: 44 20 88 46 78 78 Fax: 44 20 88 46 77 96

Email: mark .johnson@imperial.ac.uk

Word count: 6278.

1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

Abstract:

Aims:

Globally, sepsis is a major cause of mortality through the combination of cardiovascular collapse and

multi-organ dysfunction. Pregnancy appears to increase the risk of death in sepsis, but the exact

reason for the greater severity is unclear. In this study, we used polymicrobial sepsis induced by

caecal ligation and puncture (CLP) and high-dose intraperitoneal lipopolysaccharide (LPS; 10 or 40mg,

serotype 0111: B4) to test the hypotheses that pregnant mice are more susceptible to sepsis and that

this susceptibility was mediated through an excessive innate response causing a more severe

cardiovascular collapse rather than a reduction in microbe killing.

Methods and Results:

Initial studies found that mortality rates were greater and that death occurred sooner in pregnant

mice exposed to CLP and LPS. In pregnant and non-pregnant CD1 mice monitored with

radiotelemetry probes, cardiovascular collapse occurred sooner in pregnant mice, but once initiated,

occurred over a similar timescale. In a separate study, tissue, serum and peritoneal fluid (for protein,

flow cytometry, nitric oxide and bacterial load studies) were collected. At baseline, there was no

apparent Th1/Th2 bias in pregnant mice. Post CLP, the circulating cytokine response was the same,

but leukocyte marginationinfiltration in the lung was greater in pregnant mice, but only TNFα levels

were greater in lung tissue. The bacterial load in blood and peritoneal fluid was similar in both

groups.

Conclusion:

Sepsis-related mortality was markedly greater in pregnant mice. Cardiovascular collapse and organ

dysfunction occurred sooner in pregnancy, but bacterial killing was similar. Circulating and tissue

cytokine levels were similar, but off-target immune cell extravasation margination into other organs

was greater in pregnant mice. These data suggest that an excessive innate immune system response,

2

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

as shown by the exaggerated lung margination infiltration of leukocytes may be responsible for the

greater mortality. Approaches that reduce off-site trafficking may improve the prognosis of sepsis in

pregnancy.

3

46

47

48

49

50

Introduction

Sepsis in pregnancy is the leading direct cause of maternal mortality in the UK and, together with

haemorrhage and pre-eclampsia, it is a major cause of maternal mortality worldwide [1]. However,

the underlying mechanisms that result in the increased maternal mortality are not well understood.

Despite improvements in the overall management of sepsis, the rate of sepsis has increased over the

last few decades [2]. In the United States, between 1991 and 2003, the rates of maternal sepsis and

of sepsis-related death doubled [3-5]. Furthermore, with the emergence of increasingly virulent

strains and greater anti-microbial resistance, coupled with the trend for women to delay pregnancy

until later life, when obesity and type 2 diabetes are more common, it seems likely that maternal

deaths from sepsis will increase [6].

The immune system during pregnancy is uniquely modified to allow the development and growth of

a healthy pregnancy, while retaining the ability to fight infection. Traditionally, pregnancy was seen

as an immunosuppressed state [7]. However, the evidence indicates that the susceptibility to

infection during pregnancy is not increased, but the ability to fight infection once it has been

acquired is compromised. This suggests that we should actually consider pregnancy as a “modulated

immunologic condition” [8, 9]. However, the exact mechanism(s) responsible for the increase in

sepsis-related mortality during pregnancy are not understood. Indeed, little is known about the

impact of pregnancy on the response to infection. For example, cardiovascular function changes

dramatically during pregnancy with a marked decrease in peripheral resistance [10], which may

exacerbate any sepsis-induced hypotension. Cardiovascular dysfunction is a major driver of mortality

in sepsis and overproduction of nitric oxide (NO) is a mechanism for this [11]. Normally, a septic

insult triggers an inflammatory response involving the activation of inflammatory cells,

predominantly macrophages and neutrophils, which release cytokines and activate complement and

other mediators that include reactive oxygen species [12]. In tandem with the inflammatory

response, a reciprocal anti-inflammatory response occurs, which limits the extent of the pro-

inflammatory response. Dysregulation in this process leads to adverse outcomes in patients with 4

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

sepsis [13]. One previous study suggested that the innate response to LPS was enhanced during

pregnancy, with a greater pro-inflammatory TNFα and anti-inflammatory IL-10 response [14].

However, we recently studied the response in greater detail and found that the increase in

circulating cytokines after LPS treatment was similar in both pregnant and non-pregnant (NP) mice,

but that inflammatory cell tissue infiltration was more marked and that this was associated with a

greater hypotensive response [15].

There are no published basic science studies that have investigated the cardiovascular and immune

response in sepsis during pregnancy. In the current study, we have established for the first time a

reproducible rodent model of polymicrobial sepsis during pregnancy using caecal ligation and

puncture (CLP) in conjunction with radio telemetry to investigate the haemodynamic and innate

immune response to sepsis. In this study we explored the hypothesis that pregnancy is associated

with altered immune and cardiovascular responses to life threatening infection.

5

77

78

79

80

81

82

83

84

85

86

87

88

Methods

Animals: All experiments were performed under UK Home Office License 70/7372 or 70/8237, and in

accordance with the regulations of the UK Animals (Scientific Procedures) Act of 1986. NP female and

male (for breeding purposes only) CD1 outbred mice were obtained from Charles River (Margate, UK)

at 6-8 weeks of age and acclimatised for at least 48 hours before mating and 7 days prior to

operating. All mice were maintained in open cages at 21 ± 1oC, with ad libitum access to food and

water, and a 12:12 light/dark cycle regimen.

Time mating: Female mice were placed overnight into cages with male studs. The following morning

the female mice were inspected for a copulatory plug. The day of the plug would indicate the

gestational day 0 (E0). From previous work in our group E16 compares to about 34 weeks gestation in

human pregnancy.

Cardiovascular telemetry: NP CD1 female mice had a Pa-C10 implantable murine radiotelemetry

pressure catheter (Data Science International, USA) implanted in the left carotid artery. The

operation was conducted under isoflurane anaesthetic (5% for induction and 2% maintenance of

isoflurane, 400mL/min). Pre-operative medication was given prior to induction (Buprenorphine

0.02mg/kg s.c.; Enrofloxacin 150uL of 5% m/v). The left common carotid artery was isolated, a small

incision made, and the catheter tip inserted and advanced distally towards the aortic arch.

Postoperatively the mice were monitored very closely and 7 days recovery was allowed. Following

this, a continuous 24 hour recording was commenced to collect baseline data (mean arterial

pressure, heart rate and body temperature). This allowed the animal to be their own control.

Subsequently (about 12 days post- operatively), female mice were time-mated if appropriate. Data

was recorded at pre-determined time points in each experiment. Implantable temperature

transponders IPTT-300 (BMDS, USA) were placed subcutaneously (s.c) to allow non-invasive

temperature measurements.

6

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

CLP: Mice were subjected to the CLP or sham operation in both pregnant and NP groups. Animals

were prepared and induced as above (see cardiovascular telemetry). A longitudinal midline incision

was made to expose the caecum carefully and 70% ligated. Double caecal puncture was performed

using an 18G needle midway between the ligation and the tip of the caecum. The caecum was then

replaced into the abdominal cavity and the muscle layers and skin closed. In the sham operation the

caecum is identified, presented onto the abdomen and replaced again without intervention.

Following the operation, animals were monitored using telemetric measurements of blood pressure,

heart rate and temperature to determine the onset of sepsis and to enable euthanasia at a pre-

defined and humane endpoint (see Suppl. 1). Euthanasia was performed by cervical dislocation and

cease of circulation was used as confirmation method as per the Animals (Scientific Procedures) Act

1986 (ASPA) and the European Directive 2010/63/EU.

Tissue Collection: Blood and tissue was collected in a separate experiment at 3 and 6 hours following

CLP or sham operation. Blood was obtained by cardiac puncture and spun at 13,000rpm for 20

minutes for serum extraction. Tissue was collected, snap frozen and stored at -80⁰C until further use.

LPS dose response: Lipopolysaccharide (LPS, 0111:B4, Sigma-Aldrich, UK) endotoxin was injected i.p.

to cohort of NP and pregnant animals. The doses used were 10mg, and 40mg per kg.

Flow cytometry: Following tissue collection as described above, the lung tissue was homogenised in

1mL of IC fixation buffer (eBioscience, UK) using gentleMACS M tubes (Miltenyi Biotex, UK) for 1

minute. The cell suspension was filtered using 40μm nylon cell strainer (BD Falcon, UK). This filtered

suspension was centrifuged twice at 2000 rpm for 5 minutes at 4⁰C to extract leukocytes and re-

suspended in facs wash buffer (FWB) before finally re-suspended in 400uL of perm wash buffer

(PWB). The cell suspension was incubated with antibody mix for 30 minutes at 4ᵒC. Fluorophore-

conjugated rat anti-mouse monoclonal antibodies were used (see suppl. Table 1). Cell counts were

determined using AccuCheck counting beads (Invitrogen.UK), 20uL were added per sample. The

7

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

gating strategy used has been described previously [16] and [15]. Samples were run using the BD LSR

II flow cytometer (BD, UK). Data were analysed using FlowJo dongle version V10.

Plasma and peritoneal bacterial load: All samples were collected 8 hours following CLP in NP and

pregnant animals. Peritoneal samples were obtained by filling the peritoneal cavity with 3mls of

sterile PBS. Aliquots of whole blood and peritoneal washouts were serially diluted with PBS and

plated on tryptic soy agar plates (BD, UK). All plates were incubated overnight at 37˚C. The following

day, all plates were inspected and counted for Colony Forming Units (CFU) per ml.

Multiplex assay: To assess cytokine and chemokines levels during sepsis a 23-plex cytokine and

chemokine assay was performed (Bio-rad, UK) using serum samples. The manufacture’s protocol was

strictly followed. Analytes were quantified using a Luminex MAGPIX instrument with xPonent 4.2

software (Luminex Corp., USA). The concentrations of analytes were calculated by comparison to

standard curves.

Nitric Oxide Analyser: To quantify nitric oxide production a Sievers nitric oxide analyser (NOA, GE

Analytical Equipment, UK) was used. To define nitric oxide production, concentrations of both nitrite

and nitrate (NOx) levels were determined. These results were used to measure total NOx

concentration in mouse serum and tissue homogenates. In brief, samples were prepared by protein

precipitation with four volumes of methanol and centrifugation at 16,000 rpm for 30 minutes. To

obtain the total concentrations of NO, nitrites and nitrates in each sample is reduced back to NO.

This occurred in a reaction with VCl2 (Sigma, UK) in 1M HCl (Sigma, UK) at 90⁰C before passing into

the NOA system to react with ozone (O3). The production of a chemiluminescent signal was

quantified by a photo-multiplier system. The concentration of NOx was calculated by comparison to a

calibration curve of known nitrate samples.

Statistical analysis

The Kolmogoroff–Smirnoff test was used to assess the distribution of the data. Unpaired t-test or

Mann Whitney test were used where appropriate, to examine the differences between the groups.

8

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

For repeated measures or >2 groups an analysis of variance (ANOVA) was used followed by

Bonferroni correction. A p-value of <0.05 was considered to be statistically significant. Data was

analysed using GraphPad PRISM software (V.5).

9

163

164

165

166

Results

Mortality following CLP and high dose LPS: All mice were randomly allocated to naïve, sham or CLP

groups. Post-operatively, all animals were monitored closely for a 5 day period or until a humane

end-point was reached. The humane endpoint was defined as a temperature below 25 0C or the

animal demonstrating signs of severe sepsis in their coat/appearance, behaviour, posture/mobility,

breathing and movements (Suppl. Table 1). TAs expected, the mortality rate, assessed over 5 days,

following a septic insultCLP compared to sham operation during pregnancy was 92% (p<0.0001) over

a 5 day period in comparison to 50% (p=0.048) in non-pregnant (NP) mice, with a median survival of

23 hours versus 100 hours, respectively (Figure 1A). These findings established that a polymicrobial

septic insult in the form of CLP of the same magnitude resulted in worse outcomes during pregnancy

as compared to NP mice (p=0.0042). Similarly, mortality rates were observed in pregnant and non-

pregnant mice after 40mg/kg of LPS (serotype 0111:B4; Figure 1B).

Haemodynamic response to CLP: A comparison of all non-pregnant and pregnant mice following CLP

regardless of outcome, demonstrated that there was no statistical difference between mean MAP in

NP CLP 99.56 ± 1.587 vs Pregnant CLP 97.43 ± 1.988 (Suppl. Figure 1A). However, over 24 hours a

significant change difference in heart rate (HR; beats per minute, bpm) was noted in NP CLP 546 ±

14.11 vs Pregnant CLP 421 ± 12.45, p<0.0001 (Suppl. Figure 1B). The CLP NP group was then divided

into non-survivor and survivor groups, whereas the CLP pregnant group was not separated as there

was only one survivor. In the group monitored by telemetry Oforver a 100 hours telemetry period,

the median survival was 30 hours during pregnancy vs. 100 hours in NP mice. A delta analysis of the

mean arterial blood pressure (Figure 2A) and heart rate (Figure 2B) over the 24 hours preceding the

experimental endpoint was conducted. As compared to their sham groups, a statistically significant

decrease was noted in MAP and HR over the 24 hours in the NP non-survivor group and pregnant

CLP, p<0.0001 and p<0.0001, respectively. The maximum increase in ∆MAP was higher in pregnancy,

∆2.02 ± 1.26 mmHg in NP non-survivor mice vs. ∆7.62 ± 1.43 mmHg in pregnant mice as compared to

their respective sham control (NP non-survivor p<0.001, pregnant CLP p<0.05; Figure 2C). The 10

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

maximum decrease in ∆MAP was ∆-34.84 ± 8.97 mmHg in NP non-survivor mice and ∆-25.50 ± 5.16

mmHg in pregnant mice (NP non-survivor p<0.01, pregnant CLP p<0.05; Figure 2D). Comparing the

CLP NP non-survivor and CLP pregnant group there was no statistical difference over 24 hours or at

the experimental endpoint in either MAP or HR.

Organ function: Worsening organ function during septic pregnancy resulted in a The higher base

excess deficit was greater in pregnant mice post CLP (-7.4 ± 1.5) at 6 hours following CLP as

compared to NP mice (-2.3 ± 0.3, p=0.049; Suppl. Figure 2F5A). The bicarbonate levels tended to be

lower in the CLP pregnant group (Pregnant CLP 21.7 ± 1.7 vs. non-pregnant CLP 25.7 ± 0.6; Suppl.

Figure 25EB).

Inflammatory cell trafficking: In separate groups of mice treated in an identical manner, flow

cytometry was performed as previously described [15]. Comparisons were made between pregnant

and NP mice in each group (naïve, sham and CLP). Measurements were made at 3 and 6 hrs after

surgery.

Circulating: At 3 and 6 hours after CLP or sham operations, circulating Ly6CHigh monocytes tended to

increaseddecline, while declined and neutrophils remained unchanged , but both were comparable

in both NP and pregnant mice (Figure 3A-D); similarly, for circulating Ly6CLow monocytes and natural

killer cells (Suppl. Figure 2A-D).

Lung: Lung leukocyte margination infiltrationnumbers increased to a greater extent was more

marked in pregnant CLP mice at both time points (Ly6CHigh monocytes: pregnant CLP vs. NP CLP,

p=0.004 and p=0.028; neutrophils: pregnant CLP vs. NP CLP, p=0.017 and p=0.025; Figure 3E-H). For

Ly6cLow monocytes and natural killer cells, there were no significant differences at 3 hours (Suppl.

11

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

Figure 3A-D), but at 6 hours, lung Ly6cLow monocytes were increased in the pregnant CLP group

(p=0.028; Suppl. Figure 3C).

Peritoneum: NLeukocyte trafficking into peritoneal fluid was assessed 6 hours post CLP (Figure 3I-L).

In pregnant mice, nNeutrophil numbers in the peritoneal fluid were increased higher at 6 hours in

pregnant mice following CLP (p=0.033; Figure 35J).

Circulating and lung cytokine levels: At baseline in the circulation, TNFα and CCL-2 levels were

significantly higher in the naïve pregnant mice (p=0.0004 and 0.045 respectively, Figure 4C&F). At 6

hours post sham surgery, IL-6 levels were marginally higher in the NP mice (p=0.0337, Figure 4B). At

6 hours following CLP, NP mice had higher levels of TNFα, IL-12 p40 and VEGF (p=0.002, p=0.037 and

p=0.025 respectively; Figure 5C and Suppl. Figure 4A&B).

In the lung at 6 hours, cytokine levels were very similar in all groups, with the exception of TNF α,

where the levels were higher in pregnant mice, but this appeared to be due to a decline in levels in

the NP group (p<0.0005, Figure 5C).

Bacterial numbers: The number of colony forming units in whole blood and peritoneal fluid were

similar 8 hours after CLP. In peritoneal fluid, pregnant mice tended to have higher bacterial loads (4.1

x 104 [interquartile range [IQR], 3.8 x103 – 1.4 x 105]) compared to NP mice (1.3 x 103 colony-forming

unit (CFU)/ml [IQR, 283 – 5.9 x 104; Figure 6A]). In whole blood, no significant difference was seen

between pregnant and NP mice (484 [IQR, 225 – 792] vs. 1250 [IQR, 208 – 5075, Figure 6B]).

Organ function: Worsening organ function during septic pregnancy resulted in a higher base excess

deficit (-7.4 ± 1.5) 6 hours following CLP as compared to NP mice (-2.3 ± 0.3, p=0.049; Suppl. Figure

5A). The bicarbonate levels tended to be lower in the CLP pregnant group (Pregnant CLP 21.7 ± 1.7

vs. non-pregnant CLP 25.7 ± 0.6; Suppl. Figure 5B).

12

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

Total nitrates/nitrites: In a separate group of mice treated in an identical manner, no differences in

total nitrates and nitrites (NOx) concentration were found in NP and pregnant mice following CLP as

compared to their respective sham groups 3 or 6 hours following CLP (Suppl. Figure 56).

13

239

240

241

242

Discussion

To date, few studies have compared the response to sepsis in pregnant and NP rodents and those

that have used LPS as a stimulus [14, 15]. Here, we have compared the response to polymicrobial

sepsis induced by CLP in pregnant and NP mice, showing that the rate of onset of sepsis is quicker

and the mortality greater in pregnant mice. The CLP model has significantly contributed to the

understanding of the pathophysiology of sepsis and its associated molecular mechanisms 17. We used

both high-dose LPS and CLP to model sepsis in pregnancy. High-dose LPS resulted in the rapid onset

of septic shock onset underwithin 24 hours and consequently maternal deaths. Here we used

bothThe CLP model and high-dose LPS, finding demonstrated that mortality rates were comparable

however, resulted in maternal deaths over 30 hours. The increased maternal mortality in both

models perhaps suggests that the maternal response perhaps suggesting that it is the maternal

response rather than deficient bacterial killing, which ismay be responsible for the adverse outcome.

Significantly, post CLP, although cardiovascular collapse occurred earlier in pregnant mice, we found

that the degree of hypotension was similar in pregnant and NP non-surviving mice, implying that

although the blood pressure is lower during pregnancy, sepsis does not result in a greater

hypotensive response. Neutrophil migration and bactericidal activity both appeared to be intact, with

similar bacterial numbers in blood and peritoneal fluid, suggesting that the high rates of mortality

with pregnancy were not related to overwhelming infection, in agreement with as suggested by the

LPS data. The key difference we found was that pregnant mice show greater off-target inflammatory

cell margination extravasationnumbers in into off-target (tissues unaffected by the initial insult)other

organs areas, including like the the lungs, which has been implicated in adverse outcomes in other

studies of sepsis 18-21.

In the current study, we observed a more rapid cardiovascular collapse in pregnant mice, but that in

the period prior to reaching the humane end-point, the decline in blood pressure was comparable in

non-surviving pregnant and NP mice, meaning that once septic shock had set in, the rate of

deterioration was similar in pregnant and NP mice. There are 2 key differences: the proportion of 14

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

survivors, pregnant 8% vs. NP 50%, and the speed of deterioration after CLP, pregnant median

survival time 24 hours vs. NP 100 hours. These differences suggest that the response to the same

septic insult is more rapid and severe in pregnancy. Consequently, we opted to focus on the acute

response to sepsis, reasoning that we would distinguish differences in the first 6 hours which would

indicate why the onset of sepsis in pregnancy is more acute and ultimately, more severe. The earliest

deaths in the pregnancy cohort occurred at 6 hours post CLP, consequently, we chose to study

animals at 3 and 6 hours post CLP rather than later time points, in order not to introduce bias into

our data by the loss of some more severely affected mice.

Pregnancy and models of sepsis: There are no comparative studies using CLP in pregnancy. We

previously studied the impact of low-dose LPS on the immune and cardiovascular response in

pregnant and NP mice and found that in pregnancy the hypotensive effect was more marked in

association with greater monocyte margination in the pulmonary vasculature; while markers of

cardiac dysfunction and circulating cytokine levels were similar in pregnant and NP mice suggesting

that the greater monocyte activation/margination could impair vascular function in pregnant mice 15.

The only other study to investigate this used high dose LPS during pregnancy, demonstrated a higher

mortality (76% pregnant vs. 25% NP mice) with increased TNFα and decreased IL-10 immune

response 14. The higher mortality is consistent with our observations, although we found that TNF α

was higher in NP pregnant mice and that IL-10 was similar in both groups. Overall, this paper

supports the hypothesis that sepsis results in increased mortality during pregnancy.

Septic shock: The CLP model shows that the haemodynamic dysfunction in NP and pregnant mice is

closely related to poor outcome and mortality just as in human sepsis 22. It appears that maintaining

a critical threshold of perfusion is important and deteriorating below that provokes terminal decline.

Lower basal blood pressure in pregnancy may contribute to greater susceptibility to septic shock.

Indeed, in our first study, using low-dose LPS, the blood pressure declined markedly in pregnant but

not in the NP mice 15. These data supported the hypothesis that the cardiovascular response to

polymicrobial sepsis would be more marked in pregnancy. Indeed, the hypotensive response in 15

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

pregnant mice appears to occur earlier and result in worse outcomes. However, if only the non-

survivors are considered, then the cardiovascular response is comparable. This suggests that the

ability to maintain BP above a threshold level is the key determinant of survival. Consistent with this,

circulating NO levels were comparable between NP and pregnant mice and VEGF levels were actually

greater in NP mice, perhaps contributing to the lower blood pressure (in the non-survivors) at the

humane endpoint. With regards to the similarity in NO levels, the relatively early time points may

have missed any later increase in NO production or reflect the difficulty to detect changes in plasma

nitrite/nitrates (NOx) associated with specific NOS isoforms 23.

There is evidence in the literature that suggests that immune cell function is impaired during sepsis

24. Further evidence suggests that pregnancy also alters the immune response to infection and it is

likely that pregnancy hormones such as oestrogen or progesterone play a role in these alterations 8.

TheTherefore, the role of the innate immune response following CLP was investigated to evaluate

potential mechanisms to explain the higher mortality noted in this study.

Circulating changes in inflammatory cells were similar; with a decline in Ly6C High monocytes being

apparent in both pregnant and NP mice post CLP. The circulating leukocyte response was also similar

between pregnant and NP mice, but in terms of monocytes at 3 hours, was associated with a decline

in circulating numbers, which tended to be greater for the NP in terms of Ly6C LO and pregnant mice

for the Ly6CHI monocytes. The key differences were found in lung extravasationmarginationnumbers

of Ly6CHigh monocytes and neutrophils, which were greater in pregnant mice at 3 and 6 hours post

procedure.

Indeed, tThe literature describes how non-specific excessive infiltration of leukocytes may be

damaging 18-21. Other studies, using CLP in NP rodents, have been informative and shown changes in

neutrophil function in severe sepsis, which have beenare suggested to contribute to the adverse

outcome. Neutrophils have been shown to migrate less well to infected areas and this dysfunction to

be associated with increased numbers of bacteria in both blood and the peritoneal fluid 25. This

16

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

failure of neutrophil migration is suggested to be NO-mediated 25, 26 and to be associated with greater

off targetinfiltrationnumbers of neutrophils marginalisation in the lung and liver 27. The greater

marginalisation infiltrationinflammatory cell numbers may 27 or may not 28 cause direct damage to

lung or liver, but exposure to a second insult, either LPS or IgG immune complexes, caused greater

lung damage in the context of sepsis 28. In terms of our work, we found that pregnancy was not

associated with impaired peritoneal neutrophil migration, the numbers were in fact greater, or

reduced bacterial killing, where the numbers of bacteria in the blood and peritoneal fluid were the

same.

Where the pregnancy data and these findings are similar is in the greater off-target trafficking of

neutrophils and monocytes into other organs like lung. The “off-target” migration of inflammatory

cells is off target effect has been shown to be CCR-2 mediated, to be prevented in CCR-2 knockout

mice or mice treated with a CCR2 antagonist with an associated improvement in outcome 29. We

have seen a similar greater off-target response in our comparison of the effects of low dose LPS in

pregnant and NP mice, which was associated with a marked hypotensive response 15. Interestingly,

when we studied the effects of RU486, the progesterone and glucocorticoid antagonist, we found

that CCR2 mediated both neutrophil and monocyte myometrial recruitment, suggesting that during

pregnancy neutrophils are CCL2 responsive 16. This may explain why we see an increase in both

monocyte and neutrophil numbers in the lung. Recruiting neutrophils to the site of infection is

important to improve survival outcomes and to initiate a bactericidal immune response 26. Here we

found that the peritoneal neutrophil infiltration was greater and of Ly6C High monocytes similar in

pregnant mice. It is theoretically possible, that the capacity of monocytes to phagocytose and

neutrophil burst activity function in pregnant mice may be reduced, but the similarity in bacterial

colony forming units in the peritoneal fluid and blood suggests that this is not the case.

Using LPS, we observed greater numbers of inflammatory cells in the lung {Edey, 2016 #393}. These

cells were found to be marginalised, to remain in the circulation, but to be adherent to the lung

vasculature prior to potentially transmigrating into the lung tissue {Edey, 2016 #393;Patel, 2015 17

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

#408}. It is possible that CLP induces a greater effect, such that the inflammatory cells actually

migrate into the lung tissue, but this remains to be studied.

In terms of circulating cytokines, the response to CLP was similar in pregnant and NP mice with the

exception of TNFα, which was higher in the NP mice (as mentioned above). Lung cytokine levels

were broadly similar with the exception of TNFα levels, which were much greater in pregnant

animals.

Bacterial Killing: Interestingly, our findings suggest that the increased mortality noted during

pregnancy was unrelated to an increased bacterial load. The similarity in the bacterial numbers

suggests that the pressure exerted by the gravid uterus did not appear to increase or reduce the

release of faecal material into the peritoneal cavity. Studies have shown that sex steroids may

modulate the immune system A study exploring sex differences to Mycobacterium marinum

demonstrated that macrophages from female mice appeared to be more activated and

demonstrated a greater migration to the site of infection 30; this suggests that sex steroid hormones

modulate the immune response. Indeed, they have been suggested to modulate the release of pro-

and anti-inflammatory cytokines, TLR expression and antibody production 31. These data suggest that

the higher sex steroid levels in pregnant mice may be responsible for the increased migration of

neutrophils to the site of injury. Whether they also have a role in the greater off-target effects

remains to be clarified. The fall in base deficit was greater in pregnant mice post CLP, consistent with

the clinical observation that increased lactate and a greater base deficit are associated with increased

mortality and further septic complications 32-34.

This study showed that sepsis is more aggressive in pregnant mice. The markedly greater off-target

inflammatory cell extravasationmargination in the lungs suggests that this off-target effects may

mediate the high rates of mortality and that treatments targeted at reducing off target inflammatory

cell migration may be of benefit. Given the role played by NO in this process 25, 26, and the importance

18

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

of NO in blood pressure regulation, a correctly targeted NO antagonist in combination with the usual

supportive therapy, may be effective.

Funding: This work was supported by a grant from the Chelsea and Westminster Health Charity

Borne.

Acknowledgement: None.

Conflict of Interest: none declared.

19

370

371

372

373

374

375

376

References

1. Knight, M., D. Tuffnell, S. Kenyon, J. Shakespeare, R. Gray, and K.J.E.o.b.o. MBRRACE-UK, Saving Lives, Improving Mothers’ Care - Surveillance of maternal deaths in the UK 2011-13 and lessons learned to inform maternity care from the UK and Ireland Confidential Enquiries into Maternal Deaths and Morbidity 2009-13. 2015, National Perinatal Epidemiology Unit: University of Oxford.

2. Bamfo, J.E.A.K., Managing the risks of sepsis in pregnancy. Best Pract. Res. Clin. Obstet. Gynaecol., 2013. 27(4): p. 583-595.

3. Bauer, M.E., B.T. Bateman, S.T. Bauer, A.M. Shanks, and J.M. Mhyre, Maternal Sepsis Mortality and Morbidity During Hospitalization for Delivery: Temporal Trends and Independent Associations for Severe Sepsis. Anesth. Analg., 2013. 117(4): p. 944-950.

4. Cordioli, R.L., E. Cordioli, R. Negrini, and E. Silva, Sepsis and pregnancy: do we know how to treat this situation? Rev Bras Ter Intensiva, 2013. 25(4): p. 334-344.

5. Acosta, C.D., M. Knight, H.C. Lee, J.J. Kurinczuk, J.B. Gould, and A. Lyndon, The Continuum of Maternal Sepsis Severity: Incidence and Risk Factors in a Population-Based Cohort Study. PLoS One, 2013. 8(7): p. e67175.

6. Holmes, A.H., L.S.P. Moore, A. Sundsfjord, M. Steinbakk, S. Regmi, A. Karkey, P.J. Guerin, and L.J.V. Piddock, Understanding the mechanisms and drivers of antimicrobial resistance. The Lancet. 387(10014): p. 176-187.

7. Medawar, P., Some immunological and endocrinological problems raised by the evolution of viviparity in vertebrates. Symp. Soc. Exp. Biol., 1953. 7: p. 320-338.

8. Kourtis, A.P., J.S. Read, and D.J. Jamieson, Pregnancy and Infection. N. Engl. J. Med., 2014. 370(23): p. 2211-2218.

9. Pazos, M., R. Sperling, T. Moran, and T. Kraus, The influence of pregnancy on systemic immunity. Immunol. Res., 2012. 54(1-3): p. 254-261.

10. Chapman, A.B., W.T. Abraham, S. Zamudio, C. Coffin, A. Merouani, D. Young, A. Johnson, F. Osorio, C. Goldberg, L.G. Moore, T. Dahms, and R.W. Schrier, Temporal relationships between hormonal and hemodynamic changes in early human pregnancy. Kidney Int., 1998. 54(6): p. 2056-2063.

11. Davis, J.S., C.J. Darcy, T.W. Yeo, C. Jones, Y.R. McNeil, D.P. Stephens, D.S. Celermajer, and N.M. Anstey, Asymmetric Dimethylarginine, Endothelial Nitric Oxide Bioavailability and Mortality in Sepsis. PLoS One, 2011. 6(2): p. e17260.

12. Abraham, E. and M. Singer, Mechanisms of sepsis-induced organ dysfunction. Crit. Care Med., 2007. 35(10): p. 2408-2416.

13. Daniel, G.R., Cytokine Therapeutics for the Treatment of Sepsis: Why has Nothing Worked? Curr. Pharm. Des., 2003. 9(1): p. 75-82.

14. Vizi, E., J. Szelenyi, Z. Selmeczy, Z. Papp, Z. Nemeth, and G. Hasko, Enhanced tumor necrosis factor-alpha-specific and decreased interleukin-10-specific immune responses to LPS during the third trimester of pregnancy in mice. J. Endocrinol., 2001. 171(2): p. 355-361.

20

377

378379380381382

383384

385386387

388389

390391392

393394395

396397

398399

400401

402403404405

406407408

409410

411412

413414415

15. Zöllner, J., L.G. Howe, L.F. Edey, K.P. O’Dea, M. Takata, F. Gordon, J. Leiper, and M.R. Johnson, The response of the innate immune and cardiovascular systems to LPS in pregnant and nonpregnant mice†. Biol. Reprod., 2017. 97(2): p. 258-272.

16. Edey, L.F., K.P. O'Dea, B.R. Herbert, R. Hua, S.N. Waddington, D.A. MacIntyre, P.R. Bennett, M. Takata, and M.R. Johnson, The Local and Systemic Immune Response to Intrauterine LPS in the Prepartum Mouse1. Biol. Reprod., 2016. 95(6): p. 125, 1-10-125, 1-10.

17. Buras, J., B. Holzmann, and M. Sitkovsky, Animal Models of Spesis: Setting the Stage. Nat Rev Drug Discov, 2005. 4: p. 854-865.

18. Babayigit, H., C. Kucuk, E. Sozuer, C. Yazici, K. Kose, and H. Akgun, Protective effect of β-glucan on lung injury after cecal ligation and puncture in rats. Intensive Care Med., 2005. 31(6): p. 865-870.

19. Grailer, J.J., M. Kalbitz, F.S. Zetoune, and P.A. Ward, Persistent neutrophil dysfunction and suppression of acute lung injury in mice following cecal ligation and puncture sepsis. J. Innate Immun., 2014. 6(5): p. 695-705.

20. Reutershan, J., A. Basit, E.V. Galkina, and K. Ley, Sequential recruitment of neutrophils into lung and bronchoalveolar lavage fluid in LPS-induced acute lung injury. Am. J. Physiol. Lung Cell Mol. Physiol., 2005. 289(5): p. L807-L815.

21. Asaduzzaman, M., S. Zhang, S. Lavasani, Y. Wang, and H. Thorlacius, LFA-1 and MAC-1 mediate pulmonary recruitment of neutrophils and tissue damage in abdominal sepsis. Shock, 2008. 30(3): p. 254-259.

22. Brun-Buisson, C., F. Doyon, J. Carlet, and et al., Incidence, risk factors, and outcome of severe sepsis and septic shock in adults: A multicenter prospective study in intensive care units. JAMA, 1995. 274(12): p. 968-974.

23. Lauer, T., M. Preik, T. Rassaf, B.E. Strauer, A. Deussen, M. Feelisch, and M. Kelm, Plasma nitrite rather than nitrate reflects regional endothelial nitric oxide synthase activity but lacks intrinsic vasodilator action. Proc. Natl. Acad. Sci. U. S. A., 2001. 98(22): p. 12814-12819.

24. Hotchkiss, R.S., G. Monneret, and D. Payen, Sepsis-induced immunosuppression: from cellular dysfunctions to immunotherapy. Nat. Rev. Immunol., 2013. 13(12): p. 862-874.

25. Alves-Filho, J.C., A. de Freitas, F. Spiller, F.O. Souto, and F.Q. Cunha, The role of neutrophils in severe sepsis. Shock, 2008. 30(7): p. 3-9.

26. Benjamim, C.F., S.H. Ferreira, and F.d.Q. Cunha, Role of Nitric Oxide in the Failure of Neutrophil Migration in Sepsis. J. Infect. Dis., 2000. 182(1): p. 214-223.

27. Mercer-Jones, M.A., M. Heinzelmann, J.C. Peyton, D.J. Wickel, M. Cook, and W.G. Cheadle, The Pulmonary Inflammatory Response to Experimental Fecal Peritonitis: Relative Roles of Tumor Necrosis Factor-α and Endotoxin. Inflammation, 1997. 21(4): p. 401-417.

28. Czermak, B.J., V. Sarma, C.L. Pierson, R.L. Warner, M. Huber-Lang, N.M. Bless, H. Schmal, H.P. Friedl, and P.A. Ward, Protective effects of C5a blockade in sepsis. Nat. Med., 1999. 5(7): p. 788-792.

21

416417418

419420421

422423

424425426

427428429

430431432

433434435

436437438

439440441

442443

444445

446447

448449450

451452453

29. Souto, F.O., J.C. Alves-Filho, W.M. Turato, M. Auxiliadora-Martins, A. Basile-Filho, and F.Q. Cunha, Essential Role of CCR2 in Neutrophil Tissue Infiltration and Multiple Organ Dysfunction in Sepsis. Am. J. Respir. Crit. Care Med., 2011. 183(2): p. 234-242.

30. Yamamoto, Y., H. Saito, T. Setogawa, and H. Tomioka, Sex differences in host resistance to Mycobacterium marinum infection in mice. Infect. Immun., 1991. 59(11): p. 4089-4096.

31. Garcia-Gomez, E., B. Gonzalez-Pedrajo, and I. Camacho-Arroyo, Role of Sex Steroid Hormones in Bacterial-Host Interactions. BioMed Research International, 2013. 2013: p. 10.

32. Duebendorfer, C., A. Billeter, B. Seifert, M. Keel, and M. Turina, Serial lactate and admission SOFA scores in trauma: an analysis of predictive value in 724 patients with and without traumatic brain injury. Eur. J. Trauma Emerg. Surg., 2013. 39(1): p. 25-34.

33. Kellum, J.A., Fluid resuscitation and hyperchloremic acidosis in experimental sepsis: Improved short-term survival and acid-base balance with Hextend compared with saline. Crit. Care Med., 2002. 30(2): p. 300-305.

34. Scheiermann, P., D. Ahluwalia, S. Hoegl, A. Dolfen, M. Revermann, B. Zwissler, H. Muhl, K.A. Boost, and C. Hofstetter, Effects of intravenous and inhaled levosimendan in severe rodent sepsis. Intensive Care Med., 2009. 35(8): p. 1412-1419.

22

454455456

457458

459460

461462463

464465466

467468469

470

471

Figure Legends

Figure 1: Survival in non-pregnant (NP) and Pregnant CD1 mice after CLP or sham operation (CLP, A)

and following administration of i.p. 10 or 40mg/kg of LPS (LPS, B). Data are expressed as mean ± SEM.

A log-rank (Mantel-Cox) test was conducted to analyse survival data. *p<0.05, **p<0.01, ***p<0.001

(Sham Non-pregnant n=6, Sham Pregnant n=9, CLP non-pregnant n=10, CLP Pregnant n=12; NP

10mg/kg n=6, Pregnant 10mg/kg n=6, NP 40mg/kg n=9, Pregnant 40mg/kg n=6).

Figure 2: The effect of CLP on pregnant and non-pregnant mice on delta mean arterial pressure

(∆MAP, A), heart rate (∆HR, B), max increase MAP (∆MAP, C), max decrease MAP (∆MAP, CD), max

increase HR (∆HR, E) and, max decrease HR (∆HR, DF), and organ function following CLP operation (E)

base excess (BEecf) (F) bicarbonate (HCO3-). Serum samples were taken 8 hours following CLP in non-

pregnant and pregnant mice. Data are expressed as mean ± SEM. Recordings were analysed as hourly

averages. One-way ANOVA analysis with Bonferroni’s post-test, and two group analysis using an

unpaired student’s t-test or Mann Whitney U tests depending on the data distribution. ***p<0.001

(Sham NP n=6, CLP NP survivor n=6, CLP NP non-survivor n=6, Sham Pregnant n=6, CLP Pregnant

n=5).

Figure 3: Circulating cell count at 3 hours (A) Ly6Chigh monocytes; (B) Neutrophils and 6 hours (C)

Ly6Chigh monocytes; (D) Neutrophils ; lung leukocyte cell count at 3 hours (E) Ly6Chigh monocytes; (F)

Neutrophils; and 6 hours (G) Ly6Chigh monocytes; (H) Neutrophils; peritoneal leukocyte cell count

densities at 6 hours (I) Ly6Clow monocytes; (J) Ly6Chigh monocytes; (K) natural killer cells; (L) neutrophil

cells following CLP in pregnant (Preg) and non-pregnant (NP) mice. Data are expressed as mean ±

SEM. Two group analysis using an unpaired student’s t-test or Mann Whitney U tests depending on

the data distribution. *p<0.05, (NP Naïve n=6, Preg Naïve n=7, NP Sham n= 7, Preg Sham n=7, NP CLP

n=13, Preg CLP n=8).

Figure 4: Circulating concentrations of pro-inflammatory cytokines (IL-1b, IL-6, TNFa), anti-

inflammatory (IL-4, IL-10) and Chemokines, (CCL2, CCL5, CXCL1, CXCL2) in serum of non-pregnant

23

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

(NP) and pregnant (E16) mice following CLP or sham operation. (A) IL-1b (B) IL-6 (C) TNFa (D) IL-4 (E)

IL-10 (F) CCL2 (G) CCL5 (H) CXCL1 (I) CXCL2. Serum was collected 3 and 6 hours after CLP or sham

operation and in naïve mice. Data shown as mean ± SEM. Two group analysis using an unpaired

student’s t-test or Mann Whitney U tests depending on the data distribution. **p<0.01, (Naïve NP

n=6, Naïve E16 n=6, Sham NP 3HR n=6, Sham E16 6HR n=8, Sham NP 6HR n=7, Sham E16 6HR n=6,

CLP NP 3HR n=9, CLP E16 3HR n=6, CLP NP 6HR n=10, CLP E16 6HR n=6).

Figure 5: Lung concentrations of pro-inflammatory cytokines (IL-1b, IL-6, TNFa), anti-inflammatory

(IL-4, IL-10) and Chemokines, (CCL2, CCL5, CXCL1, CXCL2) in serum of non-pregnant (NP) and

pregnant (E16) mice following CLP or sham operation. (A) IL-1b (B) IL-6 (C) TNFa (D) IL-4 (E) IL-10 (F)

CCL2 (G) CCL5 (H) CXCL1 (I) CXCL2. Lung tissue was collected 6 hours after CLP or sham operation and

in naïve mice. Data shown as mean ± SEM. Two group analysis using an unpaired student’s t-test or

Mann Whitney U tests depending on the data distribution. ***p<0.001, (Naïve NP n=6, Naïve E16

n=6, Sham NP 6HR n=6, Sham E16 6HR n=6, CLP NP 6HR n=6, CLP E16 6HR n=6).

Figure 6: Bacterial count in blood and peritoneal fluid of non-pregnant (NP) and pregnant (E16) mice

following CLP operation. (A) Peritoneal fluid (B) Blood. Bacterial count was assessed in samples taken

8 hours following CLP in non-pregnant and pregnant mice. Data are expressed as median ±

interquartile range, Mann Whitney U test was used to compare NP and E16 samples (NP n=8, E16

n=8). CFU: colony forming unit.

24

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514