- Email: [email protected]
Evaluation of the Time Course of Plasma Extravasation in the Skin by Digital Image Analysis
Evaluation of the Time Course of Plasma Extravasation in the Skin by Digital Image Analysis Hilda Leonor Gonzalez, Nicole Carmichael, Jonathan O. Dostrovsky, and Milton P. Charlton Abstract: Plasma extravasation (PE) can be triggered by neurotransmitters as part of a neuroinflammatory response. We present a technique based on video digital image processing that provides a simple, noninvasive, reliable, and quantitative method for measuring the time course and extent of PE in the skin. After intravenous infusion of Evans Blue dye, stimulation of the saphenous nerve caused the skin on the dorsomedial region of the hind paw to become dark blue. The change in reflectance of the skin was recorded with a monochrome video camera. Images were digitized and analyzed with inexpensive or public domain software. The change in pixel intensity was determined in a selected region. Stimulation at 4 Hz caused greater darkening of the skin than at 1 Hz, and this was confirmed with spectrophotometric measurements of Evans Blue content. The NK1 receptor antagonist CP-99, 994 blocked saphenous nerve and substance P–induced darkening of the skin. The results indicate that our measurement gives results similar to those obtained with classic methods that are widely accepted as an indication of PE. This simple and quick method reveals the extent, time course, and location of PE, is cheap to implement and easy to learn, and thus represents a useful and alternative tool for studies of PE and its modulation. Perspective: This article presents a simple technique with which to evaluate the time course and extent of plasma extravasation in the skin of animal models of neuroinflammation. The technique is well suited to answer questions about basic physiologic mechanisms of neuroinflammation and should also be useful in drug testing studies. © 2005 by the American Pain Society Key words: Substance P, capsaicin, neurogenic inflammation, endothelium, nociceptor, pain cycle.
P
eripheral release of neuropeptides from C fibers produces neurogenic inflammation (NI), which is characterized by vasodilation and plasma extravasation (PE).1,11,12,14,24 The underlying physiologic mechanisms of neurogenic inflammation include the peripheral release of substance P (SP) and calcitonin gene-related peptide (CGRP) after C-fiber activation. SP binds to the NK1 receptor located on endothelial cells and provokes an increase in vascular permeability that allows plasma proteins to leave the blood vessels. SP also acts on mast cells and leukocytes, inducing the release of inflammatory substances such as histamine, serotonin, prostaglandins, and cytokines, which amplify the inflammatory cascade. On the other hand, CGRP binds CGRP-1 receptors located in arterioles and inReceived March 23, 2005; Revised June 7, 2005; Accepted June 24, 2005. From the Department of Physiology, University of Toronto, Toronto, Ontario, Canada. Supported by CIHR grant (M.P.C.) and a CIHR training grant (M2C) to J.O.D. for support of N.C. H.L.G. was supported by Universidad Autónoma de Bucaramanga (Colombia). Dr. Gonzalez’s current address is Universidad Autónoma de Bucaramanga, Colombia. Address reprint requests to Milton P. Charlton, Physiology Department, University of Toronto, Medical Science Building, Rm 3308, 1 King’s College Circle, Toronto, Ontario M5S1A8, Canada. E-mail: [email protected] 1526-5900/$30.00 © 2005 by the American Pain Society doi:10.1016/j.jpain.2005.06.004
duces vasodilation.2,3,6,9,13 NI can lead to the sensitization of nociceptors, and the resulting positive feedback thereby prolongs inflammation and pain. In the skin, experimental NI can be evoked by intense thermal or mechanical stimulation, by application of irritant chemicals such as capsaicin or mustard oil, or by antidromic nerve stimulation.6,9,10,24 Of the 2 components of NI, vasodilation can be studied easily by using noninvasive methods such as laser Doppler flowmetry or infrared thermography. However, the classic method of spectrophotometry used to assess PE in animal models is invasive, time-consuming, and often restricted to evaluation of one point in time, which is at the end of the experiment. The evaluation of PE often relies on the use of dyes such as Evans Blue (EB) that bind to albumin, the main plasma protein. After induction of NI, albumin-bound EB dye leaks out of blood vessels, turning the region involved in this inflammatory process dark blue. This allows for direct visualization of PE, which is later quantified by spectrophotometric methods after extracting the dye from the tissue sample.25,26 Although microdialysis22 and colored image analysis8,18 can be used to quantify PE in the skin, spectrophotometry is still in frequent use. The aim of this study was to develop a simple digitized image analysis system that can be used to measure the time course of saphenous nerve–induced PE in the skin. This method allowed the amount, location, and time course of PE to be mea-
The Journal of Pain, Vol 6, No 10 (October), 2005: pp 681-688
681
682
Video Measurement of Plasma Extravasation
sured in response to different stimulating parameters and the NK-1 antagonist, CP-99,994. This study provides researchers with an alternative method to assess PE that is noninvasive, quantitative, easy, and inexpensive.
solution (in 93% saline and 7% Tween80) was injected subcutaneously in the dorsal region of the paw 3 minutes after EB infusion.
Materials and Methods
To image the skin, we used a monochrome CCD video camera (Cohu 4915, resolution: 768 ⫻ 494 pixels; Cohu Inc, San Diego, Calif) with a 28- to 70-mm zoom lens (Tamron, Commack, NY). The gain and black level of the camera were set by a manual control and were constant during each experimental series. The camera gamma was set to 1. The camera was connected to a frame grabber (ATI All-in-Wonder, Markham, Ontario, Canada; model Radeon 7500) to acquire and digitize images in a PC. The images were acquired at regular intervals controlled by the program WinTLV (C3 Systems, UK). The AVI files stored on hard disk were uncompressed by VirtualDub (http://www.virtualdub.org) and analyzed with ImageJ (Wayne Rasband, Research Services Branch, National Institute of Mental Health, Bethesda, Md [http://rsb.info.nih.gov/ij/]). WinTLV is now unavailable, but the program Studiosurveillance (www.studio86designs.co.uk) works well with this hardware to acquire images and to uncompress AVI files. Illumination with white light was provided via a circular fiberoptic ring held 15 cm above the leg. The intensity of the light source and its distance from the leg were constant in an experimental series, but the source was sometimes tilted slightly to reduce glare on the skin. The room lighting was extinguished during experiments. The 8-bit resolution of the frame grabber provided a grey scale of 0 to 255, with black represented by 0 and white represented by 255. The lighting conditions and camera controls were set so that pixel intensity (PI) was always well within the linear range of the system; intensities higher or lower than those reported could have been measured. The region of interest (ROI) for analysis was chosen to exclude as much glare as possible. Although glare sometimes reduces the dynamic range of a measurement, it is relatively constant. When EB is extravasated, the skin color changes to a dark blue, and the reflectance of the skin is thus greatly diminished. We measured extravasation as the change in intensity of light reflected from the skin. The light intensity was measured on a pixel by pixel basis (PI) in the digitized images. PI is reported as the average PI in a ROI. To facilitate presentation of data, the grey scale of the images was inverted (Fig 1) by using the program ImageJ (inverted PI ⫽ 255 – measured PI). Two to 4 minutes before electrical stimulation or SP/ capsaicin injection, 50 mg/kg of EB dye (10 mg/mL) was injected intravenously. In saphenous nerve stimulation experiments, measurements of reflectance were made from a rectangular ROI of innervated skin approximately 15 ⫻ 35 mm on the dorsomedial part of the paw. For subcutaneous injections of SP or capsaicin, a circular area of skin (about 1-mm diameter centered on the injection site) was used as the ROI to measure reflectance. When the skin was pretreated with CP-99994 or saline, the area of injection was included as the center of the ROI. Images were recorded at 10-second intervals beginning 1
Animals and Surgical Procedures Experiments were performed under protocols approved by the University of Toronto animal care committee in accordance with the regulations of the Ontario animal research act (Canada). Male Sprague-Dawley rats weighing 250 to 400 g were housed in pairs in constant humidity, temperature of 20°C, and light and dark cycles of 12-hour duration. Anesthesia was induced with an intraperitoneal (IP) injection of sodium pentobarbitone (50 mg/kg). An adequate anesthetic level was evaluated at regular intervals throughout the experimental procedure by absence of nociceptive reflexes; when necessary, an additional injection (5 mg/kg) of pentobarbitone was applied IP. Two hours before antidromic nerve stimulation the rats were injected with 10 mg/kg IP of guanethidine to inhibit vasoconstriction due to activation of the autonomic nervous system and simplify interpretation of the findings. Both hind limbs were shaved with an electrical shaver, and the hair of the dorsal part of the paws was completely removed by using a commercial depilatory cream (Andrea; Faulding Healthcare P/L, Melbourne, Victoria, Australia). An endotracheal tube was used to facilitate spontaneous breathing, and the jugular vein was exposed and cannulated for infusion of EB. Rectal temperature was held constant at 37°C to 38°C by use of a heating pad that was feedback controlled; heart rate was monitored during the experiments. At the end of each experiment animals were killed by intravenous injection of 0.3 mL/kg of T-61 (Hoechst-Roussel Canada Ltd, Montreal, Quebec, Canada), a fast-acting euthanasia agent that contains 200 mg/mL embutraminde, 50 mg/mL mebozonium iodide, and 5 mg/mL tetracaine hydrochloride.
Induction of Plasma-Protein Extravasation Both saphenous nerves were dissected free in the thigh, cut proximally, and suspended over bipolar platinum electrodes connected to a constant current stimulus isolation and stimulator (World Precision Instruments, Sarasota, Fla; models A360 and 310). The skin flaps of the wound were raised, and paraffin oil was applied to cover the nerve. Constant current stimulation (3 mA, 0.5 millisecond) was delivered at a frequency of 4 or 1Hz for a period of 10 minutes. In a subset of experiments, SP (EMD Biosciences, Inc, San Diego, Calif) and/or capsaicin (Calbiochem, San Diego, Calif) was used to induce PE. A 28nmol/L solution of SP was prepared in 5% acetic acid and 0.9% sterile saline. Twenty-five microliters of SP was subcutaneously injected in the dorsal region of the paw after EB infusion. Twenty-five microliters of 1% capsaicin
Evaluation of Plasma Extravasation
ORIGINAL REPORT/Gonzalez et al
683
Figure 1. Appearance of the skin before (left panel) and after 10 minutes of stimulation of the saphenous nerve at 4 Hz (middle panel). Because nerve stimulation caused the skin gradually to become dark blue, the reflectance of white light decreased. To facilitate the analysis, the images were inverted (right panel); thus the areas with the darkest blue color appear brightest (highest PI) in the inverted images. Note that there was no change on the right leg, which was not stimulated. The rectangle indicates the area selected to analyze the change in PI (ROI).
minute before intravenous injection of EB. After EB injection, images were recorded for 4 minutes to obtain a baseline and then for another 10 minutes during electrical stimulation of the saphenous nerve (Fig 2). The same protocol was used for imaging effects of SP and capsaicin, but recording was prolonged to 15 minutes and 20 minutes, respectively.
Measurement of EB Concentration To compare our video-based measurements to those obtained with the classic technique for the measurement of PE, we measured the concentration of EB in excised skin by using spectrophotometry. At the end of an experiment, 0.2 mL of heparin (heparin sodium injection; BP Leo Laboratories, Canada Ltd, Guelph, Ontario, Canada) was injected intravenously followed by death by injection of T-61. The circulatory system was then flushed by transcardial perfusion with 300 mL of warmed (37°C) saline (0.9% NaCl). Excised skin tissue (30 to 50 mg) was immersed in 1 mL of formamide (BDH, Inc, Toronto, Ontario, Canada) at 60°C for 24 hours to extract the EB. The absorbance of the extracted EB solution ( ⫽ 620nm), which is proportional to EB concentration, was measured by a spectrophotometer (Shimadzu, UV 1601, Columbia, Md). The absorbencies of solutions containing known concentrations of EB were determined to construct a concentration calibration curve from which unknown concentrations could be determined. The concentration of dye was then calculated per gram weight of tissue. For each experiment EB concentration in the skin excised from the nonstimulated side was subtracted from the EB concentration in the skin excised from the stimulated side to compensate for EB that remained in blood vessels despite perfusion of the circulatory system.
Modulation of NI A 33-mmol/L or 2-mmol/L solution of CP-99,994 (donation from Pfizer) was prepared by dissolving the antagonist in 0.9% sterile saline. Twenty-five microliters of the solution was locally injected subcutaneously into the dorsomedial region of one hind paw by using a 30-gauge
Figure 2. Frequency-dependent changes in skin reflectance are similar to frequency-dependent EB accumulation. (A) Stimulation began at t ⫽ 0 minutes, and the skin gradually darkened. The baseline obtained after EB infusion was subtracted so that only stimulus-dependent changes are shown. A significant difference (P ⬍ .05, ANOVA, post hoc Tukey) in PI was observed after 2 minutes of saphenous nerve stimulation at 4 Hz. Each point represents the mean change in the average PI ⫾ SEM (n ⫽ 6). The maximum change in PI with 4-Hz stimulation (closed symbols) was significantly larger (75.6 ⫾ 3.0 PI) than stimulation at 1 Hz (54 ⫾ 2.8 PI) (open symbols) (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6). (B) The absorbance ( ⫽ 620 nm) of EB extracted from tissue samples (same animals and legs as in A) showed significant differences (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6) when the saphenous nerve was stimulated for 10 minutes at 4 Hz compared with 1 Hz (403 ⫾ 60 g/g and 194 ⫾ 35.9 g/g, respectively).
684 needle. The same volume of saline (vehicle for CP-99,994) was injected into the same region on the contralateral paw, and this side served as a control for any possible vehicle injection effects. Nerve stimulation or SP injection began 35 minutes after pretreatment with CP-99,994 and saline. The contribution of change in vessel diameter to the overall change in PI was evaluated in a subset of experiments by injecting 25 L of (0.1 mg/mL) noradrenalin subcutaneously into the dorsomedial region of the paw after PE had reached a maximum after electrical stimulation. Vasoconstriction resulting from the injection of noradrenalin was confirmed with laser Doppler blood flow monitor (Moor Instruments Inc, Wilmington, Del).
Statistical Analysis Infusion of EB dye darkened the skin. Therefore, the mean PI in a ROI recorded during 2 to 4 minutes after injection of EB determined the baseline intensity. PE induced by electrical stimulation of saphenous nerve and SP and capsaicin was calculated by subtracting the mean baseline PI resulting from EB infusion from each subsequent measurement. The image analysis figures show the change in average PI (⫾standard error of the mean [SEM]) in the ROI for each frame. Values for maximum change in PI were determined by averaging the peak PI values during the last 2 to 4 minutes of recording or when values reached a plateau. Significant differences for PI at each recorded time point between the experiments were determined with repeated-measures analysis of variance (ANOVA); if significant differences were found, post hoc Tukey tests were performed. A P value of ⬍.05 was used to determine statistical significance. All statistical comparisons were done with SigmaStat 2.03 software (SPSS Inc, Chicago, Ill).
Results Imaging Neurogenic Inflammation Intravenous injection of EB caused all areas of the skin to darken slightly. On electrical stimulation of the saphenous nerve the dorsomedial region of the paw turned dark blue. There was no change in the appearance of the contralateral (nonstimulated) side after saphenous nerve stimulation. The change in the appearance of the paw was well represented in the images obtained by the video camera (Fig 1). Stimulation of the saphenous nerve at 4 Hz caused a rapid darkening of the skin within the first 3 minutes, after which point the rate of change in PI decreased. By 10 minutes the dye was homogeneously distributed in the dorsomedial region of the paw. When the nerve was stimulated at 1 Hz, the time course of the change in average PI was similar to the change observed stimulating at 4 Hz, but the response was patchier, and the peak PI value was significantly less (54 ⫾ 2.8 vs 75.6 ⫾ 3.0 PI) (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6) (Fig 2A). To verify that the video measurements reflect PE, spectrophotometry was also used to measure PE in the same
Video Measurement of Plasma Extravasation skin observed in Fig 2A. Consistent with the video measurements, the concentration of EB dye was greater in the tissue samples from 4-Hz experiments than those from 1-Hz experiments (403 ⫾ 60 and 194 ⫾ 35.9 g/g, respectively) (Fig 2B). Although this difference is qualitatively similar to that obtained with the video measurement of PE, the measurement of EB concentration showed a 108% increase between 1 and 4 Hz, whereas the video technique showed a 41% increase in PI.
NK1 Receptor Blocker CP-99994 Modulates Saphenous Nerve–Induced Plasma Extravasation It is widely believed that nerve-evoked PE is caused by neuropeptides such as SP released from nociceptor endings. SP acts via NK1 receptors to cause an increase in vascular permeability.10 If our video recordings represent PE, then the effects of nerve stimulation should be blocked by a NK1 receptor blocker such as CP-99,994. CP-99,994 (2 or 33 mmol/L) or saline was injected subcutaneously 35 minutes before saphenous nerve stimulation. Injection of 33 mmol/L of CP-99,994 almost completely blocked saphenous nerve–induced PE around the injection site (radius, approximately 0.5 cm). A patchy response appeared surrounding this area, and after 10 minutes of stimulation a more intense darkness was observed in the area distal to the injection site (Fig 3). The video images from the group pretreated with 2 mmol/L CP-99,994 show an initial rapid change in average PI, similar to the initial increase observed in the saline group, but the rate of change slowed after 1 minute of stimulation, and the peak level was obtained by 9 minutes. Statistical analysis revealed a significant difference in the change of average PI between the saline group and 33-mmol/L group after 40 seconds of stimulation and onwards (72.9 ⫾ 4.7 and 11.5 ⫾ 3.2 PI), between the saline and 2-mmol/L groups after the second minute of stimulation (72.9 ⫾ 4.7 and 41.5 ⫾ 3 PI), and between 33-mmol/L and 2-mmol/L groups after 40 seconds of stimulation (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6). Because the blocker of NK1 receptors partially blocked the stimulus-dependent changes in skin reflectance, similar to blockade of classically measured PE,19 we concluded that these changes represent PE.
SP-Induced PE If SP released from nociceptor endings causes PE, then intradermal injection of the peptide should give a similar result. Intradermal injection of SP (25 L, 28 nmol/L) into the skin produced a circular bleb (diameter, 0.5 mm). SP injection after infusion of EB caused the area of injection to rapidly turn blue (Fig 4A). Analysis of the time course of reflectance changes in the injected area showed an initial rapid change in PI during a period of 4 minutes, followed by a slower increase. The peak change in PI induced by SP was 106.9 ⫾ 8.2 PI, and the maximum change was obtained by approximately 9 minutes. At this time, a uniform and intense blue color was observed around the injected site in a circular area of approxi-
ORIGINAL REPORT/Gonzalez et al
685
Figure 3. Stimulus-evoked PE requires NK1 receptor activation. The time course and extent of saphenous nerve–induced (4 Hz for 10 minutes) PE were determined with and without pretreatment with the NK1 receptor blocker CP-99,994. CP-99,994 reduced saphenous nerve–induced PE in a dose-dependent manner. There were significant differences between all groups (P ⬍ .05, ANOVA, post hoc Tukey). Data are expressed as mean ⫾ SEM (n ⫽ 6). Inset figures show typical pictures of legs taken at the end of experiments with saline or CP 99,994 (2 or 33 mmol/L) pretreatment. The rectangle represents the ROI used for measurements.
mately 8 mm in diameter. After this time the PI remained constant (Fig 4C). Local subcutaneous injection of SP after CP-99,994 (25 L, 2 mmol/L) caused only a slight darkening of the skin (Fig 4B). There was a significant difference in the maximum change in PI on the CP-99,994 pretreated side compared with the contralateral side that was pretreated with saline (43.6 ⫾ 5 PI and 106.9 ⫾ 8.2 PI, respectively) (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6) (Fig 4C). Image analysis showed a slower time course of PE in the skin pretreated with CP-99,994. Statistical analysis revealed significant differences between the saline and CP-99,994 pretreated groups after 3.25 minutes (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6) (Fig 4B and C). Therefore, both nerve-evoked and SP-evoked skin reflectance changes depend on activation of NK1 receptors, as does classically measured PE.19
Capsaicin-Induced PE Capsaicin triggers NI and PE7 and should therefore cause skin darkening in our assay. Subcutaneous injec-
tion of 1% capsaicin (25 L) into the skin produced a circular bleb as described for SP injection, and after EB infusion the skin rapidly darkened. Capsaicin and electrical stimulation at 4 Hz produced the same maximum change in PI (73.8 ⫾ 9.4 and 74.8 ⫾ 4 PI) but significantly less than SP-induced PE (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6). The maximum change occurred within 3 minutes of capsaicin injection (Fig 5), but the effect began to wane slightly 10 minutes after injection.
Effect of Blood Vessel Diameter on PI We wondered whether part of the optical change our technique detects is due to changes in blood vessel diameter. We tried, by injection of guanethidine, to avoid nerve-evoked vasoconstriction, but if there had been large vasodilation, a larger fraction of the skin would have appeared dark blue independent of loss of plasma protein. To evaluate this possibility we caused severe vasoconstriction by injection of noradrenalin when the skin had reached its darkest state after nerve stimulation. If much of the darkness was due to vasodilation, then the
Figure 4. PE caused by SP requires NK1 receptors. (A) After infusion of EB, a local subcutaneous injection of saline was given, followed 35 minutes later by a local injection of SP. The injection of SP caused intense darkening of the skin. (B) Local subcutaneous injection of SP after CP-99,994 caused only a slight darkening of the skin. The circles indicate the ROI chosen to analyze the change in PI. (C) Time course of the SP-induced PE in areas pretreated with saline or CP-99,994. The change in PI in the ROI was determined by subtracting the average PI during the baseline period from the average PI at each recorded time after SP injection. Data represent the mean ⫾ SEM. CP-99,994 (25 L, 2 mmol/L) significantly decreased the maximum change in PI compared with the contralateral side that was only pretreated with saline (43.6 ⫾ 5 and 106.9 ⫾ 8.2 PI, respectively) (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6). Responses to SP injection diverged between the saline group and the CP-99,994 group approximately 3 minutes after injection.
686
Figure 5. Capsaicin causes PE. EB was infused 3 minutes before subcutaneous injection of 1% capsaicin (25 L) into one hind paw and 0.9% saline (25 L) into the contralateral side. Capsaicin was injected at time 0, and a significant difference (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6) in PE occurred in the capsaicintreated paw compared with the contralateral (saline) paw (73.8 ⫾ 9.4 and 18.2 ⫾ 8 PI). Data are expressed as mean ⫾ SEM.
skin should have become much lighter when noradrenalin was injected. Intradermal injection of noradrenalin (25 L, 100 g/mL) into the dorsomedial region after electrical stimulation produced vasoconstriction of cutaneous blood vessels and 50% lower blood flow as confirmed by a laser Doppler blood flow monitor applied to the area. Noradrenalin did not cause a significant change in PI (n ⫽ 3; Fig 6). Therefore, few of the stimulusdependent changes in skin appearance are due to changes in blood vessel diameter.
Discussion Does Digitized Monochrome Image Analysis Detect PE? Stimulation of the saphenous nerve10 in rats has become a widely used animal model to study NI. Different stimulation parameters have been used in studies of nerve-induced NI in the skin of rats. It has been generally accepted that low frequencies (1 to 10 Hz) and high intensities are the best parameters to induce PE.23 Few studies have evaluated variations in evoked NI within the commonly suggested range of stimulating parameters. One study compared the effects of 2 sets of stimulation parameters, high voltage and frequency (25 V, 2 milliseconds, 10 Hz for 15 minutes) and lower voltage and frequency (10 V, 1 millisecond, 2 Hz) on saphenous nerve– induced PE. High stimulation parameters were found to produce less extravasation compared with low stimulation parameters. The decreased extravasation produced by high stimulation parameters was explained as the result of sympathetic activation, because this effect was reversed by guanethidine pretreatment.25 In the present study, possible effects of activation of sympathetic fibers by electrical stimulation of the saphenous nerve were prevented by pretreating the rats with guanethidine.
Video Measurement of Plasma Extravasation The different magnitude of PE we observed after electrical stimulation at 4 or 1 Hz indicates that the release of neuropeptides from the saphenous nerve terminal is dependent on the nerve stimulation parameters used. The use of monochrome digitized image analysis is therefore sensitive enough to detect these differences. Although it is clear that high stimulation intensities are needed to activate C fibers, the frequency of stimulation is another important parameter to be considered. This is of interest for future studies involving this model. Although a correlation analysis between tissue spectrophotometry and image analysis could not be performed because of the small number (n ⫽ 6) of animals used, we found qualitatively similar changes in skin reflectance and in EB accumulation in skin with 1-Hz and 4-Hz stimulation. However, for both frequencies tested the change in EB accumulation measured photometrically was greater than the peak change in PI. One factor that might contribute to this difference is that in the classic technique, stained blood is removed before measurement, whereas in our technique the stained blood remains in place. Exsanguination could be done at the end of an experiment to evaluate the contribution of residual blood to the measurements made with our technique. It is also possible that, unlike the tissue spectrophotometry technique, our video technique does not detect PE in deeper layers of the skin. Because the skin depth from which the changes in PI are measured is not known and these measurements might be biased toward surface effects, our technique might not show an exact linear correlation with tissue spectrophotometry, which measures EB in the entire tissue specimen. However, rat skin is sufficiently translucent that subcutaneous blood vessels are visible, suggesting that the changes monitored with this technique are likely not just from the surface layers of the skin. Future studies could use transillumination of skin to ensure that the absorbance of all layers is measured. Our finding of a qualitative relation-
Figure 6. Vasal tone has little effect on measurement of PE. Intradermal injection of noradrenalin at the end of electrical stimulation (4 Hz, minute 10) should have produced intense vasoconstriction, but no significant change (P ⬎ .05, ANOVA) in PI occured. Intradermal injection of saline at minute 10 had no effect on PI. Data are expressed as mean ⫾ SEM (n ⫽ 3).
ORIGINAL REPORT/Gonzalez et al ship between grey scale digitized imaging and spectrophotometry is evidence that the 2 methods can provide different measurements of the same phenomenon. The ability of this technique to detect changes in PE was verified by evaluating the modulatory effect of the NK1 antagonist CP-99, 994. Analysis of the change in PI confirmed that video imaging detects changes caused by saphenous nerve activity and that CP-99, 994 blocked this response. These results are in agreement with earlier findings that demonstrated the efficacy of other NK1 antagonists to inhibit saphenous nerve–induced PE.25,26 This is the first study to demonstrate the effect of CP99,994 on saphenous nerve–induced PE, because other studies have focused on its inhibitory effect in the conjunctiva and airways19 of guinea pigs, as well as in rat dura mater21). The concentrations of CP-99,994 (2 and 33 mmol/L) tested here were based on an early study that administered the antagonist subcutaneously in rats before formalin challenge.5 Although the antagonist was not used to block PE, the dose response curve determined in their study indicates that the maximum effective dose on nociceptive scores was 10 mmol/L, and 33 mmol/L CP-99,994 produced no additional effect. In our study 33 mmol/L produced significantly more reduction in PE than 2 mmol/L. High concentrations of CP-99,994 can act on calcium channels in vitro,17 and we cannot rule out the possibility that 33 mmol/L CP-99,994 might produce nonspecific effects. Although the exact specificity of CP-99,994 is beyond the scope of this study, the concentrations injected would have been higher than those actually reaching blood vessels because of dilution and diffusion in the skin after injection. In addition to sensory nerve activity, PE can also be induced by application of capsaicin or SP directly into the skin, and our technique detects appropriate changes with these challenges. Our finding that SP produces a greater change in PI compared with capsaicin and electrical stimulation indicates that the peak values observed with capsaicin and electrical stimulation between 6 to 10 minutes are not due to an experimental limitation. Furthermore, our finding also suggests that capsaicin and electrical stimulation do not release enough SP to saturate all of the NK1 receptors. In summary, our technique detects optical changes consistent with classically measured PE resulting from 3 different kinds of stimuli. We wondered whether the change in PI after stimulation of the nerve reflects not only PE but also the change in diameter of the skin blood vessels. CGRP-induced vasodilation occurs when the saphenous nerve is stimulated, and this response could contribute to the change in PI. However, when the NK1 antagonist was injected to block PE, no changes in PI were recorded in the area closest to the injected site (data not shown). These findings indicate that any vasodilation induced by CGRP did not change the light reflectance. Furthermore, under conditions of maximum vasoconstriction induced by injection of noradrenalin, no significant decrease in PI was observed. Although the subcutaneous blood vessels are no longer visible after extravasation of EB, we have ruled
687 out the possibility that any change in PI produced by changes in blood vessel diameter interfered with the ability of this technique to accurately quantify PE.
Comparison With Other Techniques The most common method to evaluate saphenous nerve–induced PE has been quantification of extravasated albumin-bound EB dye in the skin by using spectrophotometry.25,26 However, this method requires removal of the experimental tissue, thus precluding repetitive measurements in the same animal during the time course of stimulation. Recently NI has been measured with less invasive techniques. For instance, a method of computer-assisted analysis of color photographs was developed to evaluate thermal-induced erythema16 and the effects of anesthetics on burn injury in human beings.15 In this method 35-mm slides were digitized by using an image scanner, and the color (normalized red, green, and blue) and attributes (hue, saturation, and intensity) of the digitized images were analyzed. The same method was also used to follow thermal-induced extravasation of EB in skin of rats.8 Unlike these studies the present method uses grey scale computerized image analysis to study PE. The change in light reflectance produced by extravasation of the dye into the skin was evaluated by analyzing the change in PI in digitized monochrome images. Compared with slide scanning, online capture of the images is quick, and results are available immediately. The capture of images at specified intervals enabled the time course of PE to be determined. Color space analysis of video images18 has been used in several studies on human skin.4,18,20 We did not make a comparison between colored versus grey scale imaging, but analysis of color space can potentially provide more information than our method. However, the analysis of monochrome digitized video images is easier and more intuitive than the analysis of color space. Our technique cannot quantify the amount of SP (or other substances) released in the skin or its local concentration, but it simply reports the effect of released SP. Techniques available to provide a more direct assessment of SP release include microdialysis22 and radioimmunoassays. Our technique provides a very different measurement of related phenomenon and therefore cannot be compared directly. Researchers must decide which attributes of these techniques are most important.
Conclusion Analysis of monochrome digitized images provided numeric expression of the degree, time course, and location of PE occurring in the skin with 3 challenges that are accepted to produce PE as evaluated spectrophotometrically. This technique represents an alternative method to evaluate neurogenic inflammation by using a noninvasive, simple, objective, and highly reproducible method. This technique would be beneficial to those wishing to obtain quick and accurate in-
688
Video Measurement of Plasma Extravasation
formation about the effect of various drugs on PE before investing time and money in more complicated techniques such as microdialysis. Furthermore, we pro-
pose that this technique can be used to image PE in other areas, and we are currently using it to image PE in the plantar surface of the hind paw and dura mater.
References
14. Lynn B: Neurogenic inflammation caused by cutaneous polymodal receptors. Prog Brain Res 113: 361-368, 1996
1. Bayliss WM: On the origin from the spinal cord of the vasodilator fibers of the hind limb, and on the nature of these fibers. J Physiol (Lond) 26:173-209, 1901
15. Mattsson U, Cassuto J, Tarnow P, Jonsson A, Jontell M: Intravenous lidocaine infusion in the treatment of experimental human skin burns: Digital colour image analysis of erythema development. Burns 26:710-715, 2000
2. Brain SD, Newbold P, Kajekar R: Modulation of the release and activity of neuropeptides in the microcirculation. Can J Physiol Pharmacol 73:995-998, 1995 3. Brain SD, Williams TJ, Tippins JR, Morris HR, MacIntyre L: Calcitonin gene-related peptide is a potent vasodilator. Nature 313:54-56, 1985 4. Forster C, Greiner T, Nischik M, Schmelz M, Handwerker HO: Neurogenic flare responses are heterogeneous in superficial and deep layers of human skin. Neurosci Lett 185:3336, 1995 5. Henry JL, Yashpal K, Pitcher GM, Chabot JG, Coderre TJ: Evidence for tonic activation of NK-1 receptors during the second phase of formalin test in the rat. J Neurosci 19:65886598, 1999 6. Holzer P: Local effector functions of capsaicin-sensitive sensory nerve endings: Involvement of tachykinins, calcitonin gene-related peptide and other neuropeptides. Neuroscience 24:739-768, 1988 7. Jancso N, Jancso-Gabor A, Szolcsanyi J: Direct evidence for neurogenic inflammation and its prevention by denervation and by pretreatment with capsaicin. Br J Pharmacol 31:138-151, 1967 8. Jonsson A, Mattsson U, Cassuto J, Heyden G: Quantification of burn induced extravasation of Evans blue albumin based on digitized image analysis. Comput Biol Med 28:153167, 1998 9. Kress M, Guthmann C, Averbeck B, Reeh PW: Calcitonin gene-related peptide and prostaglandin E2 but not substance P release induced by antidromic nerve stimulation from rat skin in vitro. Neuroscience 89:303-310, 1999 10. Lembeck F, Holzer P: Substance P as neurogenic mediator of antidromic vasodilation and neurogenic plasma extravasation. Naunyn Schmiedebergs Arch Pharmacol 310:175-183, 1979 11. Lewis T: The blood vessels of the human skin and their responses. London, UK, Shaw and Sons, 1927 12. Lin Q, Wu J, Willis W: Dorsal root reflexes and cutaneous neurogenic inflammation after intradermal injection of capsaicin in rats. J.Neurophysiol 82:2602-2611, 1999 13. Louis SM, Jamieson A, Russell NJ, Dockray GJ: The role of substance P and calcitonin gene-related peptide in neurogenic plasma extravasation and vasodilatation in the rat. Neuroscience 32: 581-586, 1989
16. Mattsson U, Jonsson A, Jontell M, Cassuto J: Digital image analysis (DIA) of colour changes in human skin exposed to standardized thermal injury and comparison with laser Doppler measurements. Comput Methods Programs Biomed 50:31-42, 1996 17. McLean S, Ganong A, Seymour PA, Snider RM, Desai MC, Rosen T, Bryce DK, Longo KP, Reynolds LS, Robinson G, Schmidt AW, Heym J: Pharmacology of CP-99,994; a nonpeptide antagonist of the tachykinin neurokinin-1 receptor. J Pharmacol Exp Ther 267:472-479, 1993 18. Nischik M, Forster C: Analysis of skin erythema using true-color images. IEEE Trans Med Imaging 16:711-716, 1997 19. Rodger IW, Tousignant C, Young D, Savoie C, Chan CC: Neurokinin receptors subserving plasma extravasation in guinea pig airways. Can J Physiol Pharmacol 73:927-931, 1995 20. Rukwied R, Nischik M, Forster C, Heyer G, Handwerker HO: Flare responses of atopic eczema patients analysed with true colour images. Inflamm Res 46:336-341, 1997 21. Shepheard SL, Williamson DJ, Williams J, Hill RG, Hargreaves RJ: Comparison of the effects of sumatriptan and the NK1 antagonist CP-99,994 on plasma extravasation in dura mater and c-fos mRNA expression in trigeminal nucleus caudalis of rats. Neuropharmacology 34:255-261, 1995 22. Schmelz M, Luz O, Averbeck B, Bickel A: Plasma extravasation and neuropeptide release in human skin as measured by intradermal microdialysis. Neurosci Lett 230:117120, 1997 23. Szolcsanyi J: Capsaicin-sensitive sensory nerve terminals with local and systemic efferent functions: Facts and scopes of an unorthodox neuroregulatory mechanism. Prog Brain Res 113:343-359, 1996 24. Szolcsanyi J: Antidromic vasodilatation and neurogenic inflammation. Agents Actions 23:4-11, 1988 25. Towler PK, Brain SD: Activity of tachykinin NK1 and bradykinin B2 receptor antagonists, and an opiod ligand at different stimulation parameters in neurogenic inflammation in the rat. Neurosci Lett 257:5-8, 1998 26. Xu XJ, Salsgaard CJ, Maggi CA, Wiesenfeld-Hallin Z: NK-1, but not NK-2 tachykinin receptors mediate plasma extravasation induced by antidromic C-fiber stimulation in rat hindpaw: Demonstrated with the NK-1 antagonist CP96,345 and the NK-2 antagonist Men 10207. Neurosci Lett 139:249-252, 1992
P
eripheral release of neuropeptides from C fibers produces neurogenic inflammation (NI), which is characterized by vasodilation and plasma extravasation (PE).1,11,12,14,24 The underlying physiologic mechanisms of neurogenic inflammation include the peripheral release of substance P (SP) and calcitonin gene-related peptide (CGRP) after C-fiber activation. SP binds to the NK1 receptor located on endothelial cells and provokes an increase in vascular permeability that allows plasma proteins to leave the blood vessels. SP also acts on mast cells and leukocytes, inducing the release of inflammatory substances such as histamine, serotonin, prostaglandins, and cytokines, which amplify the inflammatory cascade. On the other hand, CGRP binds CGRP-1 receptors located in arterioles and inReceived March 23, 2005; Revised June 7, 2005; Accepted June 24, 2005. From the Department of Physiology, University of Toronto, Toronto, Ontario, Canada. Supported by CIHR grant (M.P.C.) and a CIHR training grant (M2C) to J.O.D. for support of N.C. H.L.G. was supported by Universidad Autónoma de Bucaramanga (Colombia). Dr. Gonzalez’s current address is Universidad Autónoma de Bucaramanga, Colombia. Address reprint requests to Milton P. Charlton, Physiology Department, University of Toronto, Medical Science Building, Rm 3308, 1 King’s College Circle, Toronto, Ontario M5S1A8, Canada. E-mail: [email protected] 1526-5900/$30.00 © 2005 by the American Pain Society doi:10.1016/j.jpain.2005.06.004
duces vasodilation.2,3,6,9,13 NI can lead to the sensitization of nociceptors, and the resulting positive feedback thereby prolongs inflammation and pain. In the skin, experimental NI can be evoked by intense thermal or mechanical stimulation, by application of irritant chemicals such as capsaicin or mustard oil, or by antidromic nerve stimulation.6,9,10,24 Of the 2 components of NI, vasodilation can be studied easily by using noninvasive methods such as laser Doppler flowmetry or infrared thermography. However, the classic method of spectrophotometry used to assess PE in animal models is invasive, time-consuming, and often restricted to evaluation of one point in time, which is at the end of the experiment. The evaluation of PE often relies on the use of dyes such as Evans Blue (EB) that bind to albumin, the main plasma protein. After induction of NI, albumin-bound EB dye leaks out of blood vessels, turning the region involved in this inflammatory process dark blue. This allows for direct visualization of PE, which is later quantified by spectrophotometric methods after extracting the dye from the tissue sample.25,26 Although microdialysis22 and colored image analysis8,18 can be used to quantify PE in the skin, spectrophotometry is still in frequent use. The aim of this study was to develop a simple digitized image analysis system that can be used to measure the time course of saphenous nerve–induced PE in the skin. This method allowed the amount, location, and time course of PE to be mea-
The Journal of Pain, Vol 6, No 10 (October), 2005: pp 681-688
681
682
Video Measurement of Plasma Extravasation
sured in response to different stimulating parameters and the NK-1 antagonist, CP-99,994. This study provides researchers with an alternative method to assess PE that is noninvasive, quantitative, easy, and inexpensive.
solution (in 93% saline and 7% Tween80) was injected subcutaneously in the dorsal region of the paw 3 minutes after EB infusion.
Materials and Methods
To image the skin, we used a monochrome CCD video camera (Cohu 4915, resolution: 768 ⫻ 494 pixels; Cohu Inc, San Diego, Calif) with a 28- to 70-mm zoom lens (Tamron, Commack, NY). The gain and black level of the camera were set by a manual control and were constant during each experimental series. The camera gamma was set to 1. The camera was connected to a frame grabber (ATI All-in-Wonder, Markham, Ontario, Canada; model Radeon 7500) to acquire and digitize images in a PC. The images were acquired at regular intervals controlled by the program WinTLV (C3 Systems, UK). The AVI files stored on hard disk were uncompressed by VirtualDub (http://www.virtualdub.org) and analyzed with ImageJ (Wayne Rasband, Research Services Branch, National Institute of Mental Health, Bethesda, Md [http://rsb.info.nih.gov/ij/]). WinTLV is now unavailable, but the program Studiosurveillance (www.studio86designs.co.uk) works well with this hardware to acquire images and to uncompress AVI files. Illumination with white light was provided via a circular fiberoptic ring held 15 cm above the leg. The intensity of the light source and its distance from the leg were constant in an experimental series, but the source was sometimes tilted slightly to reduce glare on the skin. The room lighting was extinguished during experiments. The 8-bit resolution of the frame grabber provided a grey scale of 0 to 255, with black represented by 0 and white represented by 255. The lighting conditions and camera controls were set so that pixel intensity (PI) was always well within the linear range of the system; intensities higher or lower than those reported could have been measured. The region of interest (ROI) for analysis was chosen to exclude as much glare as possible. Although glare sometimes reduces the dynamic range of a measurement, it is relatively constant. When EB is extravasated, the skin color changes to a dark blue, and the reflectance of the skin is thus greatly diminished. We measured extravasation as the change in intensity of light reflected from the skin. The light intensity was measured on a pixel by pixel basis (PI) in the digitized images. PI is reported as the average PI in a ROI. To facilitate presentation of data, the grey scale of the images was inverted (Fig 1) by using the program ImageJ (inverted PI ⫽ 255 – measured PI). Two to 4 minutes before electrical stimulation or SP/ capsaicin injection, 50 mg/kg of EB dye (10 mg/mL) was injected intravenously. In saphenous nerve stimulation experiments, measurements of reflectance were made from a rectangular ROI of innervated skin approximately 15 ⫻ 35 mm on the dorsomedial part of the paw. For subcutaneous injections of SP or capsaicin, a circular area of skin (about 1-mm diameter centered on the injection site) was used as the ROI to measure reflectance. When the skin was pretreated with CP-99994 or saline, the area of injection was included as the center of the ROI. Images were recorded at 10-second intervals beginning 1
Animals and Surgical Procedures Experiments were performed under protocols approved by the University of Toronto animal care committee in accordance with the regulations of the Ontario animal research act (Canada). Male Sprague-Dawley rats weighing 250 to 400 g were housed in pairs in constant humidity, temperature of 20°C, and light and dark cycles of 12-hour duration. Anesthesia was induced with an intraperitoneal (IP) injection of sodium pentobarbitone (50 mg/kg). An adequate anesthetic level was evaluated at regular intervals throughout the experimental procedure by absence of nociceptive reflexes; when necessary, an additional injection (5 mg/kg) of pentobarbitone was applied IP. Two hours before antidromic nerve stimulation the rats were injected with 10 mg/kg IP of guanethidine to inhibit vasoconstriction due to activation of the autonomic nervous system and simplify interpretation of the findings. Both hind limbs were shaved with an electrical shaver, and the hair of the dorsal part of the paws was completely removed by using a commercial depilatory cream (Andrea; Faulding Healthcare P/L, Melbourne, Victoria, Australia). An endotracheal tube was used to facilitate spontaneous breathing, and the jugular vein was exposed and cannulated for infusion of EB. Rectal temperature was held constant at 37°C to 38°C by use of a heating pad that was feedback controlled; heart rate was monitored during the experiments. At the end of each experiment animals were killed by intravenous injection of 0.3 mL/kg of T-61 (Hoechst-Roussel Canada Ltd, Montreal, Quebec, Canada), a fast-acting euthanasia agent that contains 200 mg/mL embutraminde, 50 mg/mL mebozonium iodide, and 5 mg/mL tetracaine hydrochloride.
Induction of Plasma-Protein Extravasation Both saphenous nerves were dissected free in the thigh, cut proximally, and suspended over bipolar platinum electrodes connected to a constant current stimulus isolation and stimulator (World Precision Instruments, Sarasota, Fla; models A360 and 310). The skin flaps of the wound were raised, and paraffin oil was applied to cover the nerve. Constant current stimulation (3 mA, 0.5 millisecond) was delivered at a frequency of 4 or 1Hz for a period of 10 minutes. In a subset of experiments, SP (EMD Biosciences, Inc, San Diego, Calif) and/or capsaicin (Calbiochem, San Diego, Calif) was used to induce PE. A 28nmol/L solution of SP was prepared in 5% acetic acid and 0.9% sterile saline. Twenty-five microliters of SP was subcutaneously injected in the dorsal region of the paw after EB infusion. Twenty-five microliters of 1% capsaicin
Evaluation of Plasma Extravasation
ORIGINAL REPORT/Gonzalez et al
683
Figure 1. Appearance of the skin before (left panel) and after 10 minutes of stimulation of the saphenous nerve at 4 Hz (middle panel). Because nerve stimulation caused the skin gradually to become dark blue, the reflectance of white light decreased. To facilitate the analysis, the images were inverted (right panel); thus the areas with the darkest blue color appear brightest (highest PI) in the inverted images. Note that there was no change on the right leg, which was not stimulated. The rectangle indicates the area selected to analyze the change in PI (ROI).
minute before intravenous injection of EB. After EB injection, images were recorded for 4 minutes to obtain a baseline and then for another 10 minutes during electrical stimulation of the saphenous nerve (Fig 2). The same protocol was used for imaging effects of SP and capsaicin, but recording was prolonged to 15 minutes and 20 minutes, respectively.
Measurement of EB Concentration To compare our video-based measurements to those obtained with the classic technique for the measurement of PE, we measured the concentration of EB in excised skin by using spectrophotometry. At the end of an experiment, 0.2 mL of heparin (heparin sodium injection; BP Leo Laboratories, Canada Ltd, Guelph, Ontario, Canada) was injected intravenously followed by death by injection of T-61. The circulatory system was then flushed by transcardial perfusion with 300 mL of warmed (37°C) saline (0.9% NaCl). Excised skin tissue (30 to 50 mg) was immersed in 1 mL of formamide (BDH, Inc, Toronto, Ontario, Canada) at 60°C for 24 hours to extract the EB. The absorbance of the extracted EB solution ( ⫽ 620nm), which is proportional to EB concentration, was measured by a spectrophotometer (Shimadzu, UV 1601, Columbia, Md). The absorbencies of solutions containing known concentrations of EB were determined to construct a concentration calibration curve from which unknown concentrations could be determined. The concentration of dye was then calculated per gram weight of tissue. For each experiment EB concentration in the skin excised from the nonstimulated side was subtracted from the EB concentration in the skin excised from the stimulated side to compensate for EB that remained in blood vessels despite perfusion of the circulatory system.
Modulation of NI A 33-mmol/L or 2-mmol/L solution of CP-99,994 (donation from Pfizer) was prepared by dissolving the antagonist in 0.9% sterile saline. Twenty-five microliters of the solution was locally injected subcutaneously into the dorsomedial region of one hind paw by using a 30-gauge
Figure 2. Frequency-dependent changes in skin reflectance are similar to frequency-dependent EB accumulation. (A) Stimulation began at t ⫽ 0 minutes, and the skin gradually darkened. The baseline obtained after EB infusion was subtracted so that only stimulus-dependent changes are shown. A significant difference (P ⬍ .05, ANOVA, post hoc Tukey) in PI was observed after 2 minutes of saphenous nerve stimulation at 4 Hz. Each point represents the mean change in the average PI ⫾ SEM (n ⫽ 6). The maximum change in PI with 4-Hz stimulation (closed symbols) was significantly larger (75.6 ⫾ 3.0 PI) than stimulation at 1 Hz (54 ⫾ 2.8 PI) (open symbols) (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6). (B) The absorbance ( ⫽ 620 nm) of EB extracted from tissue samples (same animals and legs as in A) showed significant differences (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6) when the saphenous nerve was stimulated for 10 minutes at 4 Hz compared with 1 Hz (403 ⫾ 60 g/g and 194 ⫾ 35.9 g/g, respectively).
684 needle. The same volume of saline (vehicle for CP-99,994) was injected into the same region on the contralateral paw, and this side served as a control for any possible vehicle injection effects. Nerve stimulation or SP injection began 35 minutes after pretreatment with CP-99,994 and saline. The contribution of change in vessel diameter to the overall change in PI was evaluated in a subset of experiments by injecting 25 L of (0.1 mg/mL) noradrenalin subcutaneously into the dorsomedial region of the paw after PE had reached a maximum after electrical stimulation. Vasoconstriction resulting from the injection of noradrenalin was confirmed with laser Doppler blood flow monitor (Moor Instruments Inc, Wilmington, Del).
Statistical Analysis Infusion of EB dye darkened the skin. Therefore, the mean PI in a ROI recorded during 2 to 4 minutes after injection of EB determined the baseline intensity. PE induced by electrical stimulation of saphenous nerve and SP and capsaicin was calculated by subtracting the mean baseline PI resulting from EB infusion from each subsequent measurement. The image analysis figures show the change in average PI (⫾standard error of the mean [SEM]) in the ROI for each frame. Values for maximum change in PI were determined by averaging the peak PI values during the last 2 to 4 minutes of recording or when values reached a plateau. Significant differences for PI at each recorded time point between the experiments were determined with repeated-measures analysis of variance (ANOVA); if significant differences were found, post hoc Tukey tests were performed. A P value of ⬍.05 was used to determine statistical significance. All statistical comparisons were done with SigmaStat 2.03 software (SPSS Inc, Chicago, Ill).
Results Imaging Neurogenic Inflammation Intravenous injection of EB caused all areas of the skin to darken slightly. On electrical stimulation of the saphenous nerve the dorsomedial region of the paw turned dark blue. There was no change in the appearance of the contralateral (nonstimulated) side after saphenous nerve stimulation. The change in the appearance of the paw was well represented in the images obtained by the video camera (Fig 1). Stimulation of the saphenous nerve at 4 Hz caused a rapid darkening of the skin within the first 3 minutes, after which point the rate of change in PI decreased. By 10 minutes the dye was homogeneously distributed in the dorsomedial region of the paw. When the nerve was stimulated at 1 Hz, the time course of the change in average PI was similar to the change observed stimulating at 4 Hz, but the response was patchier, and the peak PI value was significantly less (54 ⫾ 2.8 vs 75.6 ⫾ 3.0 PI) (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6) (Fig 2A). To verify that the video measurements reflect PE, spectrophotometry was also used to measure PE in the same
Video Measurement of Plasma Extravasation skin observed in Fig 2A. Consistent with the video measurements, the concentration of EB dye was greater in the tissue samples from 4-Hz experiments than those from 1-Hz experiments (403 ⫾ 60 and 194 ⫾ 35.9 g/g, respectively) (Fig 2B). Although this difference is qualitatively similar to that obtained with the video measurement of PE, the measurement of EB concentration showed a 108% increase between 1 and 4 Hz, whereas the video technique showed a 41% increase in PI.
NK1 Receptor Blocker CP-99994 Modulates Saphenous Nerve–Induced Plasma Extravasation It is widely believed that nerve-evoked PE is caused by neuropeptides such as SP released from nociceptor endings. SP acts via NK1 receptors to cause an increase in vascular permeability.10 If our video recordings represent PE, then the effects of nerve stimulation should be blocked by a NK1 receptor blocker such as CP-99,994. CP-99,994 (2 or 33 mmol/L) or saline was injected subcutaneously 35 minutes before saphenous nerve stimulation. Injection of 33 mmol/L of CP-99,994 almost completely blocked saphenous nerve–induced PE around the injection site (radius, approximately 0.5 cm). A patchy response appeared surrounding this area, and after 10 minutes of stimulation a more intense darkness was observed in the area distal to the injection site (Fig 3). The video images from the group pretreated with 2 mmol/L CP-99,994 show an initial rapid change in average PI, similar to the initial increase observed in the saline group, but the rate of change slowed after 1 minute of stimulation, and the peak level was obtained by 9 minutes. Statistical analysis revealed a significant difference in the change of average PI between the saline group and 33-mmol/L group after 40 seconds of stimulation and onwards (72.9 ⫾ 4.7 and 11.5 ⫾ 3.2 PI), between the saline and 2-mmol/L groups after the second minute of stimulation (72.9 ⫾ 4.7 and 41.5 ⫾ 3 PI), and between 33-mmol/L and 2-mmol/L groups after 40 seconds of stimulation (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6). Because the blocker of NK1 receptors partially blocked the stimulus-dependent changes in skin reflectance, similar to blockade of classically measured PE,19 we concluded that these changes represent PE.
SP-Induced PE If SP released from nociceptor endings causes PE, then intradermal injection of the peptide should give a similar result. Intradermal injection of SP (25 L, 28 nmol/L) into the skin produced a circular bleb (diameter, 0.5 mm). SP injection after infusion of EB caused the area of injection to rapidly turn blue (Fig 4A). Analysis of the time course of reflectance changes in the injected area showed an initial rapid change in PI during a period of 4 minutes, followed by a slower increase. The peak change in PI induced by SP was 106.9 ⫾ 8.2 PI, and the maximum change was obtained by approximately 9 minutes. At this time, a uniform and intense blue color was observed around the injected site in a circular area of approxi-
ORIGINAL REPORT/Gonzalez et al
685
Figure 3. Stimulus-evoked PE requires NK1 receptor activation. The time course and extent of saphenous nerve–induced (4 Hz for 10 minutes) PE were determined with and without pretreatment with the NK1 receptor blocker CP-99,994. CP-99,994 reduced saphenous nerve–induced PE in a dose-dependent manner. There were significant differences between all groups (P ⬍ .05, ANOVA, post hoc Tukey). Data are expressed as mean ⫾ SEM (n ⫽ 6). Inset figures show typical pictures of legs taken at the end of experiments with saline or CP 99,994 (2 or 33 mmol/L) pretreatment. The rectangle represents the ROI used for measurements.
mately 8 mm in diameter. After this time the PI remained constant (Fig 4C). Local subcutaneous injection of SP after CP-99,994 (25 L, 2 mmol/L) caused only a slight darkening of the skin (Fig 4B). There was a significant difference in the maximum change in PI on the CP-99,994 pretreated side compared with the contralateral side that was pretreated with saline (43.6 ⫾ 5 PI and 106.9 ⫾ 8.2 PI, respectively) (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6) (Fig 4C). Image analysis showed a slower time course of PE in the skin pretreated with CP-99,994. Statistical analysis revealed significant differences between the saline and CP-99,994 pretreated groups after 3.25 minutes (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6) (Fig 4B and C). Therefore, both nerve-evoked and SP-evoked skin reflectance changes depend on activation of NK1 receptors, as does classically measured PE.19
Capsaicin-Induced PE Capsaicin triggers NI and PE7 and should therefore cause skin darkening in our assay. Subcutaneous injec-
tion of 1% capsaicin (25 L) into the skin produced a circular bleb as described for SP injection, and after EB infusion the skin rapidly darkened. Capsaicin and electrical stimulation at 4 Hz produced the same maximum change in PI (73.8 ⫾ 9.4 and 74.8 ⫾ 4 PI) but significantly less than SP-induced PE (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6). The maximum change occurred within 3 minutes of capsaicin injection (Fig 5), but the effect began to wane slightly 10 minutes after injection.
Effect of Blood Vessel Diameter on PI We wondered whether part of the optical change our technique detects is due to changes in blood vessel diameter. We tried, by injection of guanethidine, to avoid nerve-evoked vasoconstriction, but if there had been large vasodilation, a larger fraction of the skin would have appeared dark blue independent of loss of plasma protein. To evaluate this possibility we caused severe vasoconstriction by injection of noradrenalin when the skin had reached its darkest state after nerve stimulation. If much of the darkness was due to vasodilation, then the
Figure 4. PE caused by SP requires NK1 receptors. (A) After infusion of EB, a local subcutaneous injection of saline was given, followed 35 minutes later by a local injection of SP. The injection of SP caused intense darkening of the skin. (B) Local subcutaneous injection of SP after CP-99,994 caused only a slight darkening of the skin. The circles indicate the ROI chosen to analyze the change in PI. (C) Time course of the SP-induced PE in areas pretreated with saline or CP-99,994. The change in PI in the ROI was determined by subtracting the average PI during the baseline period from the average PI at each recorded time after SP injection. Data represent the mean ⫾ SEM. CP-99,994 (25 L, 2 mmol/L) significantly decreased the maximum change in PI compared with the contralateral side that was only pretreated with saline (43.6 ⫾ 5 and 106.9 ⫾ 8.2 PI, respectively) (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6). Responses to SP injection diverged between the saline group and the CP-99,994 group approximately 3 minutes after injection.
686
Figure 5. Capsaicin causes PE. EB was infused 3 minutes before subcutaneous injection of 1% capsaicin (25 L) into one hind paw and 0.9% saline (25 L) into the contralateral side. Capsaicin was injected at time 0, and a significant difference (P ⬍ .05, ANOVA, post hoc Tukey) (n ⫽ 6) in PE occurred in the capsaicintreated paw compared with the contralateral (saline) paw (73.8 ⫾ 9.4 and 18.2 ⫾ 8 PI). Data are expressed as mean ⫾ SEM.
skin should have become much lighter when noradrenalin was injected. Intradermal injection of noradrenalin (25 L, 100 g/mL) into the dorsomedial region after electrical stimulation produced vasoconstriction of cutaneous blood vessels and 50% lower blood flow as confirmed by a laser Doppler blood flow monitor applied to the area. Noradrenalin did not cause a significant change in PI (n ⫽ 3; Fig 6). Therefore, few of the stimulusdependent changes in skin appearance are due to changes in blood vessel diameter.
Discussion Does Digitized Monochrome Image Analysis Detect PE? Stimulation of the saphenous nerve10 in rats has become a widely used animal model to study NI. Different stimulation parameters have been used in studies of nerve-induced NI in the skin of rats. It has been generally accepted that low frequencies (1 to 10 Hz) and high intensities are the best parameters to induce PE.23 Few studies have evaluated variations in evoked NI within the commonly suggested range of stimulating parameters. One study compared the effects of 2 sets of stimulation parameters, high voltage and frequency (25 V, 2 milliseconds, 10 Hz for 15 minutes) and lower voltage and frequency (10 V, 1 millisecond, 2 Hz) on saphenous nerve– induced PE. High stimulation parameters were found to produce less extravasation compared with low stimulation parameters. The decreased extravasation produced by high stimulation parameters was explained as the result of sympathetic activation, because this effect was reversed by guanethidine pretreatment.25 In the present study, possible effects of activation of sympathetic fibers by electrical stimulation of the saphenous nerve were prevented by pretreating the rats with guanethidine.
Video Measurement of Plasma Extravasation The different magnitude of PE we observed after electrical stimulation at 4 or 1 Hz indicates that the release of neuropeptides from the saphenous nerve terminal is dependent on the nerve stimulation parameters used. The use of monochrome digitized image analysis is therefore sensitive enough to detect these differences. Although it is clear that high stimulation intensities are needed to activate C fibers, the frequency of stimulation is another important parameter to be considered. This is of interest for future studies involving this model. Although a correlation analysis between tissue spectrophotometry and image analysis could not be performed because of the small number (n ⫽ 6) of animals used, we found qualitatively similar changes in skin reflectance and in EB accumulation in skin with 1-Hz and 4-Hz stimulation. However, for both frequencies tested the change in EB accumulation measured photometrically was greater than the peak change in PI. One factor that might contribute to this difference is that in the classic technique, stained blood is removed before measurement, whereas in our technique the stained blood remains in place. Exsanguination could be done at the end of an experiment to evaluate the contribution of residual blood to the measurements made with our technique. It is also possible that, unlike the tissue spectrophotometry technique, our video technique does not detect PE in deeper layers of the skin. Because the skin depth from which the changes in PI are measured is not known and these measurements might be biased toward surface effects, our technique might not show an exact linear correlation with tissue spectrophotometry, which measures EB in the entire tissue specimen. However, rat skin is sufficiently translucent that subcutaneous blood vessels are visible, suggesting that the changes monitored with this technique are likely not just from the surface layers of the skin. Future studies could use transillumination of skin to ensure that the absorbance of all layers is measured. Our finding of a qualitative relation-
Figure 6. Vasal tone has little effect on measurement of PE. Intradermal injection of noradrenalin at the end of electrical stimulation (4 Hz, minute 10) should have produced intense vasoconstriction, but no significant change (P ⬎ .05, ANOVA) in PI occured. Intradermal injection of saline at minute 10 had no effect on PI. Data are expressed as mean ⫾ SEM (n ⫽ 3).
ORIGINAL REPORT/Gonzalez et al ship between grey scale digitized imaging and spectrophotometry is evidence that the 2 methods can provide different measurements of the same phenomenon. The ability of this technique to detect changes in PE was verified by evaluating the modulatory effect of the NK1 antagonist CP-99, 994. Analysis of the change in PI confirmed that video imaging detects changes caused by saphenous nerve activity and that CP-99, 994 blocked this response. These results are in agreement with earlier findings that demonstrated the efficacy of other NK1 antagonists to inhibit saphenous nerve–induced PE.25,26 This is the first study to demonstrate the effect of CP99,994 on saphenous nerve–induced PE, because other studies have focused on its inhibitory effect in the conjunctiva and airways19 of guinea pigs, as well as in rat dura mater21). The concentrations of CP-99,994 (2 and 33 mmol/L) tested here were based on an early study that administered the antagonist subcutaneously in rats before formalin challenge.5 Although the antagonist was not used to block PE, the dose response curve determined in their study indicates that the maximum effective dose on nociceptive scores was 10 mmol/L, and 33 mmol/L CP-99,994 produced no additional effect. In our study 33 mmol/L produced significantly more reduction in PE than 2 mmol/L. High concentrations of CP-99,994 can act on calcium channels in vitro,17 and we cannot rule out the possibility that 33 mmol/L CP-99,994 might produce nonspecific effects. Although the exact specificity of CP-99,994 is beyond the scope of this study, the concentrations injected would have been higher than those actually reaching blood vessels because of dilution and diffusion in the skin after injection. In addition to sensory nerve activity, PE can also be induced by application of capsaicin or SP directly into the skin, and our technique detects appropriate changes with these challenges. Our finding that SP produces a greater change in PI compared with capsaicin and electrical stimulation indicates that the peak values observed with capsaicin and electrical stimulation between 6 to 10 minutes are not due to an experimental limitation. Furthermore, our finding also suggests that capsaicin and electrical stimulation do not release enough SP to saturate all of the NK1 receptors. In summary, our technique detects optical changes consistent with classically measured PE resulting from 3 different kinds of stimuli. We wondered whether the change in PI after stimulation of the nerve reflects not only PE but also the change in diameter of the skin blood vessels. CGRP-induced vasodilation occurs when the saphenous nerve is stimulated, and this response could contribute to the change in PI. However, when the NK1 antagonist was injected to block PE, no changes in PI were recorded in the area closest to the injected site (data not shown). These findings indicate that any vasodilation induced by CGRP did not change the light reflectance. Furthermore, under conditions of maximum vasoconstriction induced by injection of noradrenalin, no significant decrease in PI was observed. Although the subcutaneous blood vessels are no longer visible after extravasation of EB, we have ruled
687 out the possibility that any change in PI produced by changes in blood vessel diameter interfered with the ability of this technique to accurately quantify PE.
Comparison With Other Techniques The most common method to evaluate saphenous nerve–induced PE has been quantification of extravasated albumin-bound EB dye in the skin by using spectrophotometry.25,26 However, this method requires removal of the experimental tissue, thus precluding repetitive measurements in the same animal during the time course of stimulation. Recently NI has been measured with less invasive techniques. For instance, a method of computer-assisted analysis of color photographs was developed to evaluate thermal-induced erythema16 and the effects of anesthetics on burn injury in human beings.15 In this method 35-mm slides were digitized by using an image scanner, and the color (normalized red, green, and blue) and attributes (hue, saturation, and intensity) of the digitized images were analyzed. The same method was also used to follow thermal-induced extravasation of EB in skin of rats.8 Unlike these studies the present method uses grey scale computerized image analysis to study PE. The change in light reflectance produced by extravasation of the dye into the skin was evaluated by analyzing the change in PI in digitized monochrome images. Compared with slide scanning, online capture of the images is quick, and results are available immediately. The capture of images at specified intervals enabled the time course of PE to be determined. Color space analysis of video images18 has been used in several studies on human skin.4,18,20 We did not make a comparison between colored versus grey scale imaging, but analysis of color space can potentially provide more information than our method. However, the analysis of monochrome digitized video images is easier and more intuitive than the analysis of color space. Our technique cannot quantify the amount of SP (or other substances) released in the skin or its local concentration, but it simply reports the effect of released SP. Techniques available to provide a more direct assessment of SP release include microdialysis22 and radioimmunoassays. Our technique provides a very different measurement of related phenomenon and therefore cannot be compared directly. Researchers must decide which attributes of these techniques are most important.
Conclusion Analysis of monochrome digitized images provided numeric expression of the degree, time course, and location of PE occurring in the skin with 3 challenges that are accepted to produce PE as evaluated spectrophotometrically. This technique represents an alternative method to evaluate neurogenic inflammation by using a noninvasive, simple, objective, and highly reproducible method. This technique would be beneficial to those wishing to obtain quick and accurate in-
688
Video Measurement of Plasma Extravasation
formation about the effect of various drugs on PE before investing time and money in more complicated techniques such as microdialysis. Furthermore, we pro-
pose that this technique can be used to image PE in other areas, and we are currently using it to image PE in the plantar surface of the hind paw and dura mater.
References
14. Lynn B: Neurogenic inflammation caused by cutaneous polymodal receptors. Prog Brain Res 113: 361-368, 1996
1. Bayliss WM: On the origin from the spinal cord of the vasodilator fibers of the hind limb, and on the nature of these fibers. J Physiol (Lond) 26:173-209, 1901
15. Mattsson U, Cassuto J, Tarnow P, Jonsson A, Jontell M: Intravenous lidocaine infusion in the treatment of experimental human skin burns: Digital colour image analysis of erythema development. Burns 26:710-715, 2000
2. Brain SD, Newbold P, Kajekar R: Modulation of the release and activity of neuropeptides in the microcirculation. Can J Physiol Pharmacol 73:995-998, 1995 3. Brain SD, Williams TJ, Tippins JR, Morris HR, MacIntyre L: Calcitonin gene-related peptide is a potent vasodilator. Nature 313:54-56, 1985 4. Forster C, Greiner T, Nischik M, Schmelz M, Handwerker HO: Neurogenic flare responses are heterogeneous in superficial and deep layers of human skin. Neurosci Lett 185:3336, 1995 5. Henry JL, Yashpal K, Pitcher GM, Chabot JG, Coderre TJ: Evidence for tonic activation of NK-1 receptors during the second phase of formalin test in the rat. J Neurosci 19:65886598, 1999 6. Holzer P: Local effector functions of capsaicin-sensitive sensory nerve endings: Involvement of tachykinins, calcitonin gene-related peptide and other neuropeptides. Neuroscience 24:739-768, 1988 7. Jancso N, Jancso-Gabor A, Szolcsanyi J: Direct evidence for neurogenic inflammation and its prevention by denervation and by pretreatment with capsaicin. Br J Pharmacol 31:138-151, 1967 8. Jonsson A, Mattsson U, Cassuto J, Heyden G: Quantification of burn induced extravasation of Evans blue albumin based on digitized image analysis. Comput Biol Med 28:153167, 1998 9. Kress M, Guthmann C, Averbeck B, Reeh PW: Calcitonin gene-related peptide and prostaglandin E2 but not substance P release induced by antidromic nerve stimulation from rat skin in vitro. Neuroscience 89:303-310, 1999 10. Lembeck F, Holzer P: Substance P as neurogenic mediator of antidromic vasodilation and neurogenic plasma extravasation. Naunyn Schmiedebergs Arch Pharmacol 310:175-183, 1979 11. Lewis T: The blood vessels of the human skin and their responses. London, UK, Shaw and Sons, 1927 12. Lin Q, Wu J, Willis W: Dorsal root reflexes and cutaneous neurogenic inflammation after intradermal injection of capsaicin in rats. J.Neurophysiol 82:2602-2611, 1999 13. Louis SM, Jamieson A, Russell NJ, Dockray GJ: The role of substance P and calcitonin gene-related peptide in neurogenic plasma extravasation and vasodilatation in the rat. Neuroscience 32: 581-586, 1989
16. Mattsson U, Jonsson A, Jontell M, Cassuto J: Digital image analysis (DIA) of colour changes in human skin exposed to standardized thermal injury and comparison with laser Doppler measurements. Comput Methods Programs Biomed 50:31-42, 1996 17. McLean S, Ganong A, Seymour PA, Snider RM, Desai MC, Rosen T, Bryce DK, Longo KP, Reynolds LS, Robinson G, Schmidt AW, Heym J: Pharmacology of CP-99,994; a nonpeptide antagonist of the tachykinin neurokinin-1 receptor. J Pharmacol Exp Ther 267:472-479, 1993 18. Nischik M, Forster C: Analysis of skin erythema using true-color images. IEEE Trans Med Imaging 16:711-716, 1997 19. Rodger IW, Tousignant C, Young D, Savoie C, Chan CC: Neurokinin receptors subserving plasma extravasation in guinea pig airways. Can J Physiol Pharmacol 73:927-931, 1995 20. Rukwied R, Nischik M, Forster C, Heyer G, Handwerker HO: Flare responses of atopic eczema patients analysed with true colour images. Inflamm Res 46:336-341, 1997 21. Shepheard SL, Williamson DJ, Williams J, Hill RG, Hargreaves RJ: Comparison of the effects of sumatriptan and the NK1 antagonist CP-99,994 on plasma extravasation in dura mater and c-fos mRNA expression in trigeminal nucleus caudalis of rats. Neuropharmacology 34:255-261, 1995 22. Schmelz M, Luz O, Averbeck B, Bickel A: Plasma extravasation and neuropeptide release in human skin as measured by intradermal microdialysis. Neurosci Lett 230:117120, 1997 23. Szolcsanyi J: Capsaicin-sensitive sensory nerve terminals with local and systemic efferent functions: Facts and scopes of an unorthodox neuroregulatory mechanism. Prog Brain Res 113:343-359, 1996 24. Szolcsanyi J: Antidromic vasodilatation and neurogenic inflammation. Agents Actions 23:4-11, 1988 25. Towler PK, Brain SD: Activity of tachykinin NK1 and bradykinin B2 receptor antagonists, and an opiod ligand at different stimulation parameters in neurogenic inflammation in the rat. Neurosci Lett 257:5-8, 1998 26. Xu XJ, Salsgaard CJ, Maggi CA, Wiesenfeld-Hallin Z: NK-1, but not NK-2 tachykinin receptors mediate plasma extravasation induced by antidromic C-fiber stimulation in rat hindpaw: Demonstrated with the NK-1 antagonist CP96,345 and the NK-2 antagonist Men 10207. Neurosci Lett 139:249-252, 1992