Characterising the gut microbiome in veterans with Gulf War Illness: a protocol for a longitudinal, prospective cohort study.

INTRODUCTION: Approximately 25%-35% of the 1991 Gulf War Veteran population report symptoms consistent with Gulf War Illness (GWI), a chronic, multi-symptom illness characterised by fatigue, pain, irritable bowel syndrome and problems with cognitive function. GWI is a disabling problem for Gulf War Veterans, and there remains a critical need to identify innovative, novel therapies.Gut microbiota perturbation plays a key role in the symptomatology of other chronic multi-symptom illnesses, including myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS). Given similarities between ME/CFS and GWI and the presence of gastrointestinal disorders in GWI patients, Veterans with GWI may also have gut abnormalities like those seen with ME/CFS. In this longitudinal cohort study, we are comparing the diversity (structure) and the metagenomes (function) of the gut microbiome between Gulf War Veterans with and without GWI. If we find differences in Veterans with GWI, the microbiome could be a target for therapeutic intervention to alleviate GWI symptoms. METHODS AND ANALYSIS: Participants answer questions about diet, exercise and lifestyle factors. Participants also complete a questionnaire (based on the Kansas case definition of GWI) regarding their medical history and symptoms; we use this questionnaire to group participants into GWI versus healthy control cohorts. We plan to enrol 52 deployed Gulf War Veterans: 26 with GWI and 26 healthy controls. Participants provide stool and saliva samples weekly for an 8-week period for microbiome analyses. Participants also provide blood samples at the beginning and end of this period, which we will use to compare measures of inflammation markers between the groups. ETHICS AND DISSEMINATION: The protocol was approved by the University of Wisconsin-Madison Health Sciences Institutional Review Board and the William S. Middleton Memorial Veterans Hospital Research and Development Committee. Results of this study will be submitted for publication in a peer-reviewed journal.

Apnea causes microvascular perfusion maldistribution in isolated rat lungs.

Obstructive sleep apnea is associated with significant cardiovascular disease, yet little is known about the effects of OSA on pulmonary microvascular perfusion. In a recent report, we showed that pulmonary microvascular perfusion was significantly mal-distributed in anesthetized, spontaneously breathing rats exposed to five episodes of obstructive apnea. We quantified microvascular perfusion by analyzing trapping patterns of 4 µm diameter fluorescent latex particles infused into the pulmonary circulation after the last episode. We could not determine if the perfusion maldistribution was due to the effects of large subatmospheric intrapleural pressures during apnea, or to precapillary OSA hypoxic vasoconstriction. To address this, we repeated these studies using isolated, buffer-perfused rat lungs (Ppulm art , 10 cm H2 O) ventilated in a chamber (-5 to -15 cm H2 O, 25 breaths/min; Ptrachea  = 0). We simulated apnea by clamping the trachea and cycling the chamber pressures between -25 and -35 cm H2 O for five breaths. After five apnea episodes, we infused 4 µm diam. fluorescent latex particles into the pulmonary artery. The number of particles recovered from the venous effluent was 74% greater in nonapneic isolated lungs compared to apneic lungs (P ≤ 0.05). Apneic lungs also had perfusion maldistributions that were 73% greater than those without apnea (P ≤ 0.05). We conclude that simulated apnea in isolated, perfused rat lungs produces significantly greater particle trapping and microvascular perfusion maldistribution than in nonapneic isolated lungs. We believe these effects are due to the large, negative intrapleural pressures produced during apnea, and are not due to hypoxia.

Obstructive apnea causes microvascular perfusion maldistribution in the lungs of rats.

Obstructive sleep apnea (OSA) is associated with significant cardiovascular consequences, including pulmonary hypertension, yet little is known about its effects on pulmonary microvascular perfusion. To investigate effects of OSA on pulmonary microvascular perfusion, we clamped the tracheal cannulas of anesthetized, spontaneously breathing rats to simulate obstructive apnea. The clamp remained in place for 10 breaths before it was released to allow the animals to again breathe spontaneously. We repeated this protocol every 20 s until the rat experienced a total of five apneic episodes of 10 breaths each. We then infused into a femoral vein 108 fluorescent latex particles (4 µm diameter), which became trapped within the pulmonary microcirculation. We removed the lungs, allowed them to air-dry, and quantified the particle distributions in sections of the lungs using dispersion index (DI) analysis, a method we developed previously. The log of the DI (logDI) is a measure of perfusion maldistribution. Greater log(DI) values correspond to greater maldistribution. Apneic lungs had average logDI values of 1.28 (SD 0.24). Rats not subjected to apnea had average logDI values of 0.85 (SD 0.08) ( P ≤ 0.05). Rats that received latex particles 10 min or 24 h after apnea had average logDI values of 0.97 (SD 0.31) and 0.84 (SD 0.38), respectively (not significant). Our results demonstrate, for the first time, that a few apneic events produced significant, but temporary, perfusion maldistribution within the pulmonary microcirculation. Repeated nightly episodes of apnea over months and years may explain why human patients with OSA suffer from significantly greater cardiovascular disease than those without OSA.

Positive pressure ventilation compresses pulmonary acinar microvessels but not their supply vessels.

Pulmonary alveolar septal capillaries receive their perfusion from a web of larger surrounding acinar vessels. Using 4 µm diam. Latex particles, we showed that positive pressure ventilation narrowed the acinar vessels as evidenced by venous 4 µm particle concentrations and perfusate flows <50% of particle concentrations in negative pressure ventilated lungs. We aimed to understand the effects of positive and negative pressure ventilation on flows of larger particles through the lung. Isolated, ventilated rat lungs (air, transpulmonary pressures of 15/5 cm H2O, 25 breaths/min) were perfused with a hetastarch solution at Ppulm art/PLA pressures of 10/0 cm H2O. Red latex 7 µm (2.5 mg, 4.8 × 106) or 15 µm (2.5 mg, 5.5 × 105) particles were infused into each lung during positive or negative pressure ventilation. An equal number of green particles was infused 20 min later. Flows through lungs infused with 7 µm and 15 µm particles were not different from flows through lungs infused with 4 µm particles (p = 0.811). Venous particle concentrations of 7 µm particles relative to infused particles were lower in positive pressure lungs (0.03 ±â€¯0.03%) compared to negative pressure lungs (0.17 ±â€¯0.12%) (p = 0.041). Venous particle concentrations of 15 µm particles were not different between positive (2.3 ±â€¯1.9%) and negative (3.3 ±â€¯1.8%) pressure ventilation (p = 0.406). Together with our previous study, we conclude that 4 µm and 7 µm particles both enter acinar vessels but that the 7 µm particles are too large to flow through those vessels. In contrast, we conclude that 15 µm particles bypass the acinar vessels, flowing instead through larger intrapulmonary vessels to enter the venous outflow. These findings suggest that intrapulmonary vessels are organized as a web that allows bypass of the acinar vessels by large particles and, that these bypass vessels are not compressed by positive pressure ventilation.

Negative pressure ventilation enhances acinar perfusion in isolated rat lungs.

We compared acinar perfusion in isolated rat lungs ventilated using positive or negative pressures. The lungs were ventilated with air at transpulmomary pressures of 15/5 cm H2O, at 25 breaths/min, and perfused with a hetastarch solution at Ppulm art/PLA pressures of 10/0 cm H2O. We evaluated overall perfusability from perfusate flows, and from the venous concentrations of 4-µm diameter fluorescent latex particles infused into the pulmonary circulation during perfusion. We measured perfusion distribution from the trapping patterns of those particles within the lung. We infused approximately 9 million red fluorescent particles into each lung, followed 20 min later by an infusion of an equal number of green particles. In positive pressure lungs, 94.7 ± 2.4% of the infused particles remained trapped within the lungs, compared to 86.8 ± 5.6% in negative pressure lungs ( P ≤ 0.05). Perfusate flows averaged 2.5 ± 0.1 mL/min in lungs ventilated with positive pressures, compared to 5.6 ± 01 mL/min in lungs ventilated with negative pressures ( P ≤ 0.05). Particle infusions had little effect on perfusate flows. In confocal images of dried sections of each lung, red and green particles were co-localized in clusters in positive pressure lungs, suggesting that acinar vessels that lacked particles were collapsed by these pressures thereby preventing perfusion through them. Particles were more broadly and uniformly distributed in negative pressure lungs, suggesting that perfusion in these lungs was also more uniformly distributed. Our results suggest that the acinar circulation is organized as a web, and further suggest that portions of this web are collapsed by positive pressure ventilation.

Arterio-venous anastomoses in isolated, perfused rat lungs.

Several studies have suggested that large-diameter (>25 µm) arterio-venous shunt pathways exist in the lungs of rats, dogs, and humans. We investigated the nature of these pathways by infusing specific-diameter fluorescent latex particles (4, 7, 15, 30, or 50 µm) into isolated, ventilated rat lungs perfused at constant pressure. All lungs received the same mass of latex (5 mg), which resulted in infused particle numbers that ranged from 1.7 × 107 4 µm particles to 7.5 × 104 50 µm particles. Particles were infused over 2 min. We used a flow cytometer to count particle appearances in venous effluent samples collected every 0.5 min for 12 min from the start of particle infusion. Cumulative percentages of infused particles that appeared in the samples averaged 3.17 ± 2.46% for 4 µm diameter particles, but ranged from 0.01% to 0.17% for larger particles. Appearances of 4 µm particles followed a rapid upslope beginning at 30 sec followed by a more gradual downslope that lasted for up to 12 min. All other particle diameters also began to appear at 30 sec, but followed highly irregular time courses. Infusion of 7 and 15 µm particles caused transient but significant perfusate flow reductions, while infusion of all other diameters caused insignificant reductions in flow. We conclude that small numbers of bypass vessels exist that can accommodate particle diameters of 7-to-50 µm. We further conclude that our 4 µm particle data are consistent with a well-developed network of serial and parallel perfusion pathways at the acinar level.

Inhaled thrombolytics reduce lung microclot and leukocyte infiltration after acute blood loss.

We showed previously that a 30% blood loss in rats, without resuscitation, caused significant accumulation of microthrombi and leukocytes within the pulmonary circulation by 24 h. We hypothesized that the microthrombi formed spontaneously as a consequence of hemorrhage-induced stasis within the low-pressure pulmonary circuit and that the leukocytes were attracted to them. This suggested that elimination of the microthrombi, using an inhaled thrombolytic agent, could prevent the neutrophil sequestration after blood loss. To test this hypothesis, we removed 30% of the calculated blood volume from isoflurane-anesthetized, male Sprague-Dawley rats (350-500 g) over 5 min and allowed them to recover. Six hours later, we re-anesthetized the rats and nebulized tissue plasminogen activator (80 or 320 µg/kg), lactated Ringer's solution (LRS), or ipratropium bromide (i-bromide) into their lungs. We used i-bromide as a control after we discovered that nebulized LRS had thrombolytic properties. At 24 h, we removed and fixed the lungs and prepared sections for immunohistochemistry using antibodies against fibrinogen (microthrombi) and CD16 (leukocytes). Digital images of each section were obtained using a confocal microscope. Pixel counts of the images showed significantly less accumulation of microthrombi and leukocytes in lungs nebulized with tissue plasminogen activator or LRS than in non-nebulized lungs or in lungs nebulized with i-bromide (P ≤ 0.05). Lactated Ringer's solution becomes positively charged when nebulized (unlike i-bromide), suggesting that it eliminated microthrombi by fibrin depolymerization. We confirmed this using an in vitro assay. Our results demonstrate that lyses of microthrombi that accumulate in the lung after acute blood loss prevent subsequent leukocyte sequestration.

Evidence for active control of perfusion within lung microvessels.

Vasoconstrictors cause contraction of pulmonary microvascular endothelial cells in culture. We wondered if this meant that contraction of these cells in situ caused active control of microvascular perfusion. If true, it would mean that pulmonary microvessels were not simply passive tubes and that control of pulmonary microvascular perfusion was not mainly due to the contraction and dilation of arterioles. To test this idea, we vasoconstricted isolated perfused rat lungs with angiotensin II, bradykinin, serotonin, or U46619 (a thromboxane analog) at concentrations that produced equal flows. We also perfused matched-flow controls. We then infused a bolus of 3 µm diameter particles into each lung and measured the rate of appearance of the particles in the venous effluent. We also measured microscopic trapping patterns of particles retained within each lung. Thirty seconds after particle infusion, venous particle concentrations were significantly lower (P ≤ 0.05) for lungs perfused with angiotensin II or bradykinin than for those perfused with U46619, but not significantly different from serotonin perfused lungs or matched flow controls. Microscopic clustering of particles retained within the lungs was significantly greater (P ≤ 0.05) for lungs perfused with angiotensin II, bradykinin, or serotonin, than for lungs perfused with U46619 or for matched flow controls. Our results suggest that these agents did not produce vasoconstriction by a common mechanism and support the idea that pulmonary microvessels possess a level of active control and are not simply passive exchange vessels.

Bacteremia does not affect cellular uptake of ultrafine particles in the lungs of rats.

To assess the effects of intra-abdominal bacteremia on lung cellular function in vivo, we used electron microscopy to quantify the uptake of 6 nm diameter, albumin-coated colloidal gold particles (overall diam. 20.8 nm) by cells in the lungs of rats made septic by the introduction of live bacteria (E.coli and B. fragilis) into their abdomens. Gold particles were instilled into the trachea 24 hr after bacteremia induction, and lungs were harvested and prepared for electron microscopy 24 hr later. Because bacteremia produces an increase in metabolism, we hypothesized that this might be associated with increased cellular uptake of these particles and also with increased permeability of the alveolar epithelial barrier to them, as bacteremia is also associated with lung injury. We quantified particle uptake by counting particle densities (particles/µm²) within type I and type II epithelial cells, capillary endothelial cells, erythrocytes and neutrophils in the lungs of five septic rats and five sham-sepsis controls. We also counted particle densities within organelles of these cells (nuclei, mitochondria, type II cell lamellar bodies) and within the alveolar interstitium. We found particles to be present within all of these compartments, although we found no differences in particle densities between bacteremic rats and sham-sepsis controls. Our results suggest that these 6 nm particles were able to freely cross cell and organelle membranes, and further suggest that this ability was not altered by bacteremia.

Microthrombus formation may trigger lung injury after acute blood loss.

We showed previously that acute blood loss, without resuscitation, caused marked maldistribution of interalveolar perfusion. Because hemorrhage is a known risk factor for the development of lung injury, the goal of our present studies was to determine if there was a correlation between perfusion maldistribution and the subsequent development of lung injury after blood loss. Specifically, we wanted to know if the perfusion maldistribution might be due to microthrombus formation and/or leukocyte sequestration within the pulmonary microcirculation. We bled rats (30% blood loss) and harvested their lungs 45 min or 24 h later. Lungs were prepared for perfusion distribution analysis, Western blot analysis to measure whole-lung fibrinogen concentrations, and for immunohistochemistry to measure fibrin deposition and leukocyte deposition (CD16 fluorescence). Perfusion was significantly maldistributed at 45 min and 24 h (P < 0.05). At 45 min, whole-lung fibrinogen concentrations were less than half that in controls (P < 0.05), whereas numbers of fibrin microthrombi were 2.5-fold greater than control by 45 min (not statistically significant) and were 4.5-fold greater by 24 h (P = 0.01). Leukocyte deposition was two-fold greater than control by 45 min (not statistically significant) and was 4-fold greater by 24 h (P = 0.02). Fibrin-to-leukocyte nearest-neighbor distances remained unchanged (18.1 [SD, 1.1] µm) even as the numbers of both increased with time after blood loss. Our results suggest that soluble fibrinogen polymerized to insoluble fibrin within minutes after acute blood loss, which caused perfusion maldistribution and attracted leukocytes. The development of lung injury after blood loss may be a consequence of leukocyte chemoattraction to fibrin microthrombi that seem to form within minutes after blood loss.

Bacteremic sepsis disturbs alveolar perfusion distribution in the lungs of rats.

OBJECTIVE: Sepsis often leads to lung injury, although the mechanisms that initiate this are unclear. One preinjury phenomenon that has not been explored previously is the effect of bacterial (nonlipopolysaccharide) sepsis on the distribution of alveolar perfusion. The goals of our studies were to measure this. DESIGN: Randomized, controlled, prospective animal study. SETTING: University animal laboratory. SUBJECTS: Male Sprague-Dawley rats (450-550 g). INTERVENTIONS: We induced sepsis by placing gelatin capsules containing Escherichia coli and Bacteroides fragilis into the abdomens of rats (n = 9). Empty capsules (n = 6) were placed into the abdomens of controls. After 24 hrs, 4-microm-diameter fluorescent latex particles (2 x 10(8)) were infused into the pulmonary circulation. Sepsis was induced in additional rats and controls to assess lung injury, as follows: Lung histology was performed on eight septic rats and on seven controls; lung lavage was performed on three septic rats and three controls after their plasma albumin had been labeled with Evans blue dye. MEASUREMENTS AND MAIN RESULTS: Confocal microscopy was used to prepare digital maps of latex particle trapping patterns (eight per lung). Analysis of these patterns revealed statistically more clustering (perfusion inhomogeneity) down to tissue volumes less than that of ten alveoli in septic lungs compared with controls (p < or = .05). Bacterial counts and neutrophil counts were significantly higher in the circulation of septic rats (p < or = .05). Blood pressures and arterial PO2s were unchanged. Cell counts in histological images were three-fold higher in septic lungs than in controls (p < or = .05). Lung lavage revealed 0.41 +/- 0.03 mL of plasma in the lungs of septic rats, and 0.06 +/- 0.05 mL in the lungs of controls (p < or = .05). CONCLUSIONS: Bacterial sepsis caused significant maldistribution of interalveolar perfusion in the lungs of rats in the absence of significant lung injury.

Acute hypoxia does not alter inter-alveolar perfusion distribution in unanesthetized rats.

Effects of hypoxic vasoconstriction on inter-alveolar perfusion distribution (< or =1000 alveoli) have not been studied. To address this, we measured inter-alveolar perfusion distribution in the lungs of unanesthetized rats breathing 10% O(2). Perfusion distributions were measured by analyzing the trapping patterns of 4 microm diameter fluorescent latex particles infused into the pulmonary circulation. The trapping patterns were statistically quantified in confocal images of the dried lungs. Trapping patterns were measured in lung volumes that ranged between less than 1 and 1300 alveoli, and were expressed as the log of the dispersion index (logDI). A uniform (statistically random) perfusion distribution corresponds to a logDI value of zero. The more this value exceeds zero, the more the distribution is clustered (non-random). At the largest tissue volume (1300 alveoli) logDI reached a maximum value of 0.68+/-0.42 (mean+/-s.d.) in hypoxic rats (n = 6), 0.50+/-0.38 in hypercapnic rats (n.s.) and 0.48+/-0.25 in air-breathing controls (n.s.). Our results suggest that acute hypoxia did not cause significant changes in inter-alveolar perfusion distribution in unanesthetized, spontaneously breathing rats.

Hemorrhage progressively disturbs interalveolar perfusion in the lungs of rats.

Acute hemorrhage is often followed by devastating lung injury. However, why blood loss should lead to lung injury is not known. One possibility is that hemorrhage rapidly disturbs the distribution of microvascular perfusion at the alveolar level, which may be a triggering event for subsequent injury. We showed previously that a 30% blood loss in rats caused significant maldistribution of interalveolar perfusion within 45 min (J Trauma 60:158, 2006). In this report, we describe results of further exploration of this phenomenon. We wanted to know if perfusion distribution was disturbed at 15 min, when vascular pressures were significantly reduced by the blood loss, compared with those at 45 min, when the pressures had returned substantially toward normal. We hemorrhaged rats by removing 30% of their blood volume. We quantified interalveolar perfusion distribution by statistically analyzing the trapping patterns of 4-microm-diameter fluorescent latex particles infused into the pulmonary circulation 15 (red particles) and 45 min (green particles) after blood removal. We used confocal fluorescence microscopy to digitally image the trapping patterns in sections of the air-dried lungs and used pattern analysis to quantify the patterns in tissue image volumes that ranged from 1,300 alveoli to less than 1 alveolus. LogDI, a measure of perfusion maldistribution, increased from 1.00 +/- 0.15 at 15 min after blood loss to 1.62 +/- 0.24 at 45 min (P < 0.001). These values were 0.86 +/- 0.22 (15 min) and 1.12 +/- 0.24 (45 min) in control rats (P = 0.03). Hemorrhage caused the green (45 min)-to-red (15 min) particle distance to decrease from 35.9 +/- 6.5 to 28.0 +/- 5.1 microm (P = 0.024) and the red-to-green particle distance to remain unchanged (30.2 +/- 5.7 microm [red]; 31.5 +/- 10.0 microm [green] [n.s.]). We conclude that hemorrhage caused a progressive increase in interalveolar perfusion maldistribution over 45 min that did not correspond to reduced arterial pressures or altered blood gases. Our particle distance measurements led us to further conclude that this maldistribution occurred in areas that were perfused at 15 min rather than in previously unperfused areas .

Hemorrhage causes interalveolar perfusion maldistribution in the lungs of anesthetized rats.

BACKGROUND: Lung injury often occurs following hemorrhage and we hypothesized that this might be due to the effects of hemorrhage on perfusion distribution among alveoli. To test this, we measured interalveolar perfusion distribution in anesthetized, spontaneously breathing rats subjected to blood losses of 0%, 10%, 20%, or 30% of calculated blood volume. METHODS: We measured interalveolar perfusion distribution by analyzing trapping patterns of 4-mum diameter fluorescent latex particles infused into the pulmonary circulation. The particles (2 x 10) were infused 1 hour after each animal had been bled, and the lungs were then removed and air-dried. Using a confocal fluorescence microscope, we collected images of the particles in eight sections of each lung. Each image encompassed 3,360 x 3,360 x 100 microm (approximately 5,000 alveoli), and included 3-4,000 particles. Particle distributions in the images were measured using the method of dispersion index (DI) analysis. A DI value of zero corresponds to a statistically random distribution; the more DI exceeds zero, the more the distribution is clustered or inhomogenous. RESULTS: The largest DI values for the four groups were: 0%, 0.69 +/- 0.41; 10%, 0.57 +/- 0.58; 20%, 0.72 +/- 0.34; 30%, 1.38 +/- 0.41. The 30% blood loss group had a max DI value approximately twofold greater than those of the other three (p < 0.0001). CONCLUSIONS: Our results suggest that interalveolar perfusion distribution becomes markedly maldistributed at blood losses of 30%. This contributes to ventilation-perfusion mismatching, and may be a precipitating event for lung injury following hemorrhage.

Thromboxane receptor analog, U-46619, redistributes pulmonary microvascular perfusion in isolated rat lungs.

Effects of vasoconstriction on the distribution of perfusion among alveoli are not well understood. To address this, we used a new method we developed to determine how microvascular perfusion distribution was affected by a potent vasoconstrictor, the thromboxane receptor analog U-46619. Our method was to infuse 4-microm-diameter fluorescent latex microspheres into the circulation of isolated rat lungs vasoconstricted with U-46619. We used a confocal microscope to image trapping patterns of the particles in dried sections of the lungs and then used dispersion index analysis to quantify the particle patterns in the images, which encompassed approximately 2,000 alveoli. Dispersion indexes revealed significantly more particle clustering (inhomogeneous distribution) in vasoconstricted lungs than in normal flow controls or in controls in which flow was reduced by either lowering pulmonary arterial pressure or raising left atrial pressure. These results suggest that vasoconstriction occurred in the microvessels themselves, which are much smaller vessels than those previously thought to be capable of vasoconstriction.

Perfusion heterogeneity in rat lungs assessed from the distribution of 4-microm-diameter latex particles.

Pulmonary vascular perfusion has been shown to follow a fractal distribution down to a resolution of 0.5 cm(3) (5E11 microm(3)). We wanted to know whether this distribution continued down to tissue volumes equivalent to that of an alveolus (2E5 microm(3)). To investigate this, we used confocal microscopy to analyze the spatial distribution of 4-microm-diameter fluorescent latex particles trapped within rat lung microvessels. Particle distributions were analyzed in tissue volumes that ranged from 1.7E2 to 2.8E8 microm(3). The analysis resulted in fractal plots that consisted of two slopes. The left slope, encompassing tissue volumes less than 7E5 microm(3), had a fractal dimension of 1.50 +/- 0.03 (random distribution). The right slope, encompassing tissue volumes greater than 7E5 microm(3), had a fractal dimension of 1.29 +/- 0.04 (nonrandom distribution). The break point at 7E5 microm(3) corresponds closely to a tissue volume equivalent to that of one alveolus. We conclude that perfusion distribution is random at tissue volumes less than that of an alveolus and nonrandom at tissue volumes greater than that of an alveolus.