INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IN ADULT MAMMALS, normal hematopoietic activity is mostly confined to the bone marrow. It has traditionally been difficult to study the bone marrow in vivo because it is enclosed within the rigid bone cavity. With the use of the atraumatic, radioactive microsphere method, we and others (7, 14, 15, 22) showed that augmented hematopoietic activity, induced by blood loss or by growth factors, increased blood flow to the intact bone marrow in rabbits and rats. These findings might be translated into a clinical context because patients treated with growth factors due to severe bone marrow failure frequently experience pain at the site of hematopoietic activity (2, 11). It is possible that the hyperemia after addition of growth factors increases blood volume and/or enhances transcapillary passage of fluid and thus raises the intramedullary pressure. It is well known (5, 8, 30) that the bone marrow is innervated with nerve fibers terminating in close proximity to blood vessels, progenitor cells, and stromal cells. Although no convincing role has been ascribed to the efferent part of this innervation in murine animals (3), a raised intramedullary pressure will possibly induce excitation of sensory neurons in the patients and hence cause pain. However, there is to the best of our knowledge no data on the pressure-volume dynamics in the intact bone marrow. Most estimates of intravascular volume (IVV) within soft tissues stem from experiments where a plasma tracer (e.g., albumin) has been allowed a brief (usually 5 min) circulation period, balanced between adequate intravascular mixing and minimal outward transport into the interstitium (25). This outward transport is predominantly determined by size as the smaller compounds leak out, whereas larger compounds will be confined to the microcirculation. The vascular anatomy within the bone marrow is complicated, with several arterial branches that, via sinusoids, converge on the emissary veins (13). The structure of the sinusoidal wall, with its discontinuous basement membrane, allows virtually unrestricted exchange also of macromolecules, but not of cells (17, 20, 26). This is in contrast to the continuous capillaries found in, e.g., skeletal muscle and skin. After injection, inert microspheres with diameters slightly below the diameters of capillaries and sinusoids will recirculate without being entrapped in the microcirculation. Hence we wanted to determine bone marrow intravascular volume with microspheres. To validate this method, we included a comparison with low-density lipoproteins (LDL) because the plasma concentration of these particles remains nearly constant and with virtually no outward transport during the first 10-15 min in the rat (6, 24).

Both the brain and the dental pulp are encased in rigid tissues, and their presumably low interstitial compliances might resemble that in the bone marrow. Measurements of interstitial fluid pressure (IFP) by using a servo-controlled counter-pressure system connected to micropipettes inserted into either the brain or the dental pulp have allowed detailed studies (12, 29) of baseline pressure characteristics of these tissues and how the pressures respond to various forms of neuronal, respiratory, and circulatory changes in vivo. An important aim of the present study was thus to examine whether this micropuncture technique would be suitable for the bone marrow.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Rats. The experimental protocol was approved by the local ethics committee. We used male Wistar rats (350-475 g body wt). Anesthesia was induced with barbital sodium (50 mg/kg ip) and maintained with regular supply (2-3 mg/kg ia). The rats were placed supine on a heating pad to maintain normal body temperature, as measured with a rectal probe. A polyethylene-50 catheter was inserted into one common carotid artery and advanced into the aortic root for blood sampling, injection of tracers, and measurements of blood pressure and heart rate. Any unwanted blood loss (<1 ml) was immediately replaced with equal amounts of 0.9% NaCl (37°C). Arterial blood pressure was recorded continuously with an oscillograph (model 7402AQ, Hewlett-Packard). The rats breathed spontaneously, and, if necessary, oxygen was delivered to the inspired air to ensure adequate arterial blood gases (range: PO₂, 11.5-13.7; PCO₂, 4.8-5.7 kPa). To prevent any movement, both hind legs were fixated without compromising the circulation. We made a small incision in the groin and the femoral arteries were exposed. To induce changes in the bone marrow blood flow, one artery was clamped with a thin snare placed around it. The contralateral hind leg served as a control.

Measurements of tissue extracellular volumes. To measure IVV, we gave an intra-arterial injection of ¹³¹I-labeled human serum albumin (diameter ~7 nm, 1 µCi, specific activity 10 mCi/mg; Institute for Energy Technique; Kjeller, Norway), ¹²⁵I-labeled rat LDL (diameter 20-25 nm, 2 µCi, specific activity 15 mCi/mg), or microspheres (~500,000 per injection) with diameters of either (means ± SE) 5.2 ± 0.1 or 8.3 ± 0.2 µm and labeled with either ⁴⁶Sc or ¹⁰³Ru (NEN-DuPont; Boston, MA). Rat LDL was prepared and iodinated as detailed elsewhere (4). Iodide-labeled compounds were dialyzed before each experiment, and trichloloroacetic acid precipitation showed that the amount of either free ¹²⁵I or free ¹³¹I never exceeded 0.2%. To prevent possible aggregation, albumin, LDL, or microspheres were injected into separate rats. Erythrocyte volume was determined with injection of ⁵⁹Fe-tagged erythrocytes (1 µCi, specific activity 13 mCi/mg) that had been prepared as previously described (21). To measure interstitial fluid volume (IFV), a constant plasma concentration of ⁵¹Cr-labeled EDTA (specific activity 30 mCi/mg, Institute for Energy Technique) was achieved by a continuous infusion (0.01 µCi/min ia) over 90 min. Separate experiments showed that the plasma equivalent volumes of EDTA in the studied tissues had reached a stable plateau at this time point (data not shown).

Calculations. All of the volumes were determined as milliliters per gram of tissue wet weight. Plasma equivalent volumes (V_pl) in the tissues were calculated as

(1)

where counts per minute (cpm) denote the radioactivity of the injected tracer. Tissue erythrocyte volume (V_e) was determined as

(2)

where Hct denotes the systemic hematocrit expressed as a fraction. Tissue IVV was then obtained as

(3)

We estimated tissue IFV as

(4)

Measurements of bone marrow IFP. The IFP was measured with a servo-controlled counter-pressure system originally described by Wiederhielm et al. (28) with these minor modifications. After induction of anesthesia, the proximal part of the tibial bone close to the tuberositas was exposed. Under stereomicroscopic guidance, a small cavity (diameter 200-250 µm) was carefully drilled until the endosteal surface became visible. A sharpened glass micropipette (tip diameter 3-5 µm) filled with 0.5 M NaCl colored with Evans blue was thereafter gently advanced ~0.5-1 mm into the marrow. Care was taken to minimize tissue damage and bleeding. A 30-min stabilization period was allowed before any measurements. The IFP was recorded continuously with a pressure transducer connected to an amplifier and recorder (Hewlett-Packard). The transducer was calibrated immediately before each experiment, and zero pressure at the bone surface was checked regularly by placing the pipette in the saline covering the marrow exposure. A recording of IFP was accepted provided that the following criteria were fulfilled: 1) the feedback gain of the servo-controlled counter-pressure system could be altered without interfering with the measured pressure; 2) by applying pump suction, the resistance across the glass pipette increased indicating fluid communication between the pipette and the interstitium; 3) the zero pressure level did not change during a recording; and 4) injection of a small amount of Evans blue remained in the tissue and did not disappear, indicating that the tip was not placed intravascularly. This experimental setup did not allow simultaneous bilateral measurements of bone marrow IFP.

Measurement of local blood flow. A laser-Doppler flowmeter (model 4001, Periflux Master, Perimed KB; Jarfalla, Sweden) with a wavelength of 780 nm was used to measure the laser-Doppler flux (LDF) as an indicator of local blood flow to the tibial skin. The flowmeter was equipped with a needle probe (fiber diameter 125 µm with separation 500 µm; Probe Periflux PF 415:10). The probe was placed in gentle contact with a shaved region ~1-2 cm from the marrow exposure, where the IFP measurements were performed. A second needle probe was placed at a similar position on the contralateral hind leg. Both probes were positioned to give the largest resting flux signals. Motility standard calibration of the instrument and fiberoptic probes was done according to the manufacturer. Zero flux was determined as the values recorded from a stationary white card with the same intensity of reflected light as was present when recorded with the probe positioned at the leg after cardiac arrest. The flowmeter set constant was 0.03 s with both an upper and a lower bandwidth of 20 kHz. The values were obtained in arbitrary perfusion units and are presented as percentages of control measurements.

Post mortem handling of tissue samples. After the volume measurements were completed, some rats were deeply anesthetized intraperitoneally with barbital sodium. We then cannulated the thoracic aorta. This catheter was connected to a peristaltic pump, and the rat was perfused with heparinized 0.9% NaCl until the fluid content from the cut distal end of the inferior caval vein was devoid of blood. After we concluded the experiments, all of the rats were killed with an intraperitoneal overdose of barbital sodium. To determine the extracellular volumes, we carefully removed the hind legs before we isolated the soleus muscle with overlying skin. The tibial bone marrow was then separated from the tibial bone as previously detailed (14). Radioactivity was determined with a gamma counter (Auto-Gamma 5220, Packard Instruments). Different decay rates for the various nuclides and spillover between different nuclide channels were corrected for. At least 5,000 counts above an insignificant background were obtained for any nuclide.

Statistics. The values were means ± SE. Differences were evaluated with the two-tailed Wilcoxon rank sum test for paired samples or the Kruskall-Wallis nonparametric ANOVA test with the Bonferroni correction. P < 0.05 indicated statistical significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Determination of intravascular and interstitial volumes. To assess whether we could use the smallest microspheres (5.2-µm diameter) as a plasma tracer, we injected either albumin (6 rats), LDL (6 rats), or microspheres with diameters of either 8.3 µm (6 rats) or 5.2 µm (6 rats). Figure 1 shows the plasma concentrations of the radioactively labeled tracers obtained after the injections. After an initial brief (~2 min) decrease, most likely caused by intravascular mixing, the plasma concentrations of both albumin and LDL decreased only insignificantly over the 150-min observation period (Fig. 1A). The plasma concentration of 5.2-µm microspheres followed a similar concentration-time pattern, whereas the plasma concentrations of the larger microsphere batch (8.3 µm) as expected rapidly approached zero, due to entrapment in the microcirculation (Fig. 1B). Hence, the larger-sized microspheres were not used for further experiments. On the basis of the values for plasma concentrations obtained at 5 min for albumin, LDL, or 5.2-µm microspheres and the systemic hematocrit, total blood volumes were (n = 6, P > 0.05) 56.4 ± 2.8, 57.1 ± 3.0, and 56.9 ± 2.9 ml/kg body wt for the albumin, LDL, and microspheres, respectively. This is in accordance with previous estimates that used an injection of tagged erythrocytes into rats (1).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Time course of the concentrations of the various plasma tracers. A: plasma concentrations of both albumin () and low-density lipoproteins (LDL) () remained virtually unchanged during the 150-min observation period. B: plasma concentrations of the 8.3-µm () microspheres decreased rapidly after the injections, whereas the 5.2-µm () microspheres remained intravascularly for 150 min. cpm, Counts per minute. Values are means ± SE; n = 6 rats. *P < 0.05 and **P < 0.01, compared with the 5-min measurement.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Intravascular and interstitial fluid volumes within various tissues during reactive hyperemia. A: intravascular volume (IVV) increased during peak hyperemia (solid bars) compared with control (open bars) in skeletal muscle, skin, and bone marrow, whereas it remained unaltered in the bone. B: intersitial fluid volume (IFV) increased during hyperemia in skeletal muscle and skin. Whereas a reduction was noted in bone marrow IFV, the bone IFV remained unchanged. Values are means ± SE; n = 6 rats. *P < 0.05 for hyperemia vs. control.

Bone marrow IFP. During baseline conditions, a pulsative pattern of IFP synchronous to the pulsations in arterial blood pressure and the LDF was noted (Fig. 3A and 4A). Figure 3A shows recordings of bone marrow IFP, skin LDF, and arterial blood pressure before, during, and after clamping of a femoral artery. The MAP and IFP obtained during baseline conditions were 116 ± 6 and 9.7 ± 0.4 mmHg (n = 11), respectively.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Reactive hyperemia increases interstitial fluid pressure (IFP) within the bone marrow. A: simultaneous recordings of arterial blood pressure (AP), bone marrow IFP, and skin LDF in ipsi- and contralateral hind leg before, during, and after clamping of the ipsilateral femoral artery. Start and end of clamping are indicated by arrows. B: both the skin LDF and the bone marrow IFP dropped significantly (*) during arterial clamping, whereas mean AP (MAP) remained unchanged. Values are means ± SE (n = 6 rats) given as percentage of baseline (preclamping) measurements.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Hypotensive hypovolemia decreases IFP within the bone marrow. A: simultaneous recordings of AP, bone marrow IFP, and skin LDF before, during, and after a blood loss of ~25% of total blood volume. Arrowheads, periods of MAP recordings. Solid arrows, loss of 5 ml of blood. Open arrow, reinfusion of 5 ml of blood. Pulsatile IFP recordings are in phase with the AP and spontaneous variations in the latter are paralleled in the IFP. B: marked blood loss decreased both skin LDF, MAP, and bone marrow IFP. Values are means ± SE; n = 6 rats. *P < 0.05 for blood loss vs. baseline (prebleeding) measurements.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This is the first report detailing measurements of extracellular volumes and IFP within the intact bone marrow. Because the marrow is encased in the rigid bone compartment with a complex vascular morphology and lack of lymphatic vessels, in vivo measurements of plasma volume within the bone marrow requires that a steady-state concentration of free tracer compound is rapidly achieved and that the tracer does not exit the vasculature during passage through this organ. Both the 5.2-µm microspheres and LDL fulfilled these criteria. In contrast, the larger-sized microspheres embolized in the microcirculation. Albumin extravasated within the bone marrow to rapidly equilibrate between plasma and the interstitium, as recently noted (19). This is in contrast to skeletal muscle and skin which both have low permeabilities to albumin in their continuous capillaries (25). In the present study, we showed that the injected LDL remained intravascular within the studied organs in the rat for at least 10 min, thus supporting other observations (6, 24). Previous studies (20) are also in line with our findings, showing that the uptake of LDL through the bone marrow sinusoids in murine animals is quite small and insignificant compared with the hepatic clearance.

IFV was measured after EDTA had equilibrated in the entire bone marrow extracellular volume. EDTA is an inert molecule subjected to renal elimination (10) and not bound to plasma proteins (9). After density centrifugation of crude bone marrow cell suspensions, no EDTA-related radioactivity was detected in the low-, intermediate-, or high-density fractions (data not shown), indicating no cellular uptake. Consequently, EDTA most likely is a suitable tracer for IFV measurements in the bone marrow.

The mean baseline value for bone marrow IVV (1.1 ml/100 g) was clearly lower than for both skeletal muscle (5.4 ml/100 g) and skin (2.9 ml/100 g) but higher than for bone (0.4 ml/100 g). The mean baseline value for bone marrow IFV (6.3 ml/100 g) was also lower than for skeletal muscle (12.8 ml/100 g) and skin (37.2 ml/100 g). Our IFV estimate for bone (1.9 ml/100 g) compares favorably with McCarthy (19), who obtained an extravascular volume after albumin equilibration of 1.2 ml/100 g.

Use of the micropuncture technique to continuously measure both baseline values and acutely induced changes in IFP has successfully been applied for low-compliant tissues, such as the brain and dental pulp (12, 29). By adopting a similar procedure, we now present direct measurements of IFP in the bone marrow. The procedure for inserting the pipette is virtually atraumatic, and indeed we did not observe any alteration in IFP or MAP on administration of indomethacin, indicating no inflammatory reaction. Because of the smaller outer diameter of the pipette tip relative to the diameter of the bone marrow exposure, compliance might have been slightly increased and thus led to a minor underestimation of IFP. Furthermore, to adequately measure IFP, the tip of the pipette needs to be placed extracellularly. It is not possible to precisely localize the tip during a measurement. Nevertheless, if some of our recordings should reflect intracellular rather that IFP, the measured pressure will still most likely be equal to IFP because significant hydrostatic pressure gradients are unlikely across the soft plasma membrane of cells (18). IFP was measured at one site only and any regional variation in IFP throughout the bone marrow is unknown. Our estimate was higher than those of the rat brain (3.4 mmHg; see Ref. 29) and cat dental pulp (6.3 mmHg; see Ref. 12).

The clamping of one femoral artery and the subsequent reactive hyperemia increased bone marrow IFP and induced marked fluid shifts within the extracellular space of the bone marrow without affecting the total extracellular volume. In contrast, both IVV and IFV remained unaltered in the surrounding bone. The numbers and volumes of bone marrow-derived cells sampled from the ipsilateral leg were not different from those of the contralateral leg (data not shown), indicating no major intracellular volume shifts in the experimental hind leg after either hyperemia or blood loss. The increase in IVV resulted mainly from an increase in the plasma volume. The concomitant decrease in IFV indicates a net transport of fluid from the bone marrow interstitium into the intravascular compartment. Furthermore, as albumin rapidly equilibrated across the extravascular space, any change in oncotic pressures is a less likely explanation for the observed changes in bone marrow IFP and extravascular volumes. Because no change in cellular volume took place and the bone marrow is devoid of lymphatic drainage, we postulate that the hyperemia led to a transient distension of smaller arterial vessels and proximal sinusoids, thereby causing an increase in IFP and hence a hydrostatic gradient favoring an initial transport of fluid from the interstitium to the intravascular compartment. Whether the reactive hyperemia led to an increase in shear stress and thus augmented hydraulic conductivity of the bone marrow microvasculature (23) is not known. The estimates of IVV and IFV rely on tissue deposition of tracer molecules; hence, no continuous measurements can be performed. We therefore cannot detail possible time-dependent extracellular fluid shifts during the reactive hyperemia.

To ensure an adequate function, the bone marrow must be able to rapidly adapt its metabolic activity to changing demands such as, e.g., blood loss or inflammatory processes. Iversen et al. (14, 15) showed that both normovolemic and hypovolemic anemia as well as hematopoietic growth factor supplements rapidly enhance blood flow to the bone marrow. Our present finding that enhanced blood flow increased bone marrow IFP may explain the clinical observations that patients treated with such growth factors regularly experience bone pain.

In conclusion, the present study shows that perfusion changes have marked impact on the extracellular volume-pressure relation within the bone marrow microenvironment. In the absence of lymphatic drainage and with a stiff bone encapsulement, volume-pressure relations will be adjusted by altered perfusion or transport of bone marrow-derived cells. Any malfunctioning of these two mechanisms may impair bone marrow function. This can be seen, i.e., during the progression of acute myeloid leukemia in rats, where the blood flow to the bone marrow gradually decreases with a concomitant hypoxic and acidic microenvironment that is ultimately fatal (16, 27).