Electrochemical detection of histamine and 5-hydroxytryptamine at

Nov 1, 1995 - buffer, histamine is oxidized at carbon fiber microelec- trodes at potentials close to the background in a chemi- cally irreversible pro...
1 downloads 0 Views 2MB Size
Anal. Ghem. 1995, 67, 4514-4521

Electrochemical Detection of Histamine and 5Hydroxytryptamine at Isolated Mast Cells Karin Pihel, Showchien Hsieh, James W. Jorgenson, and R. Mark Wightman*

Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-3290

The electrochemical oxidation of histamine has been investigated as an analytical tool. In a physiological buffer, histamine is oxidized at carbon fiber microelectrodes at potentials close to the background in a chemically irreversible process. Cylindrical carbon fiber electrodes were used as ampemmetric detectorsfor histamine separated with a reversed-phase capillary column, and detection limits of 240 am01 were achieved. Electrodes with beveled tips were used as real-time Sensors by monitoring with repetitive cyclic voltammograms at a scan rate of 800 V/s with a 16.7-ms repetition rate, and detection limits of 1.4 pM were achieved. Both techniques were used to probe histamine and 5-hydroxytqptamine (5-HT)stored in rat peritoneal mast cells. The content in single cells was measured by capillary HPLC, and both substances were found in single cells. Although the analysis revealed a large cell-to-cellvariation in the amount of histamine and 5-€IT,the average amount was 150 and 4 fmol of histamine and 5-€IT, respectively. Release of histamine and 5-€ITwas measured with the electrode placed 1 pm from the cell surface. Release was observed as a series of sharp concentration spikes, consistent with corelease of the two substances from individual vesicles following exocytosis. Carbon fiber microelectrodes enable measurements of very small amounts of biologically important compounds. The small size of these sensors allows them to be coupled with microcolumns to serve as detectors of separated Alternatively, they can be used as chemical sensors implanted in tissue8 or adjacent to single biological cells?JO Microelectrodetechniques have been used extensively for the detection of easily oxidizable compounds such as the catecholamine^^-^^ and 5hydroxytryptamine (5-HT),13-15which serve as chemical messengers in (1) Kennedy, R. T.; Oates, M. D.; Cooper, B. R.; Nickerson. B.; Jorgenson, J. W. Science 1 9 8 9 , 246, 57-63. (2) Cooper, B. R.; Jankowski, J. A; Leszczyszyn, D. J.; Wightman, R. M.; Jorgenson, J. W. Anal. Chem. 1992, 64, 691-694. (3) Cooper, B. R; Wightman, R M.; Jorgenson, J. W. J. Chromatogr. B 1 9 9 4 , 653, 25-34. (4) Ewing, A G.; Mesaros, J. M.; Gavin, P. F. Anal. Chem. 1 9 9 4 , 66, 527A537A (5) Olefirowicz, T. M.; Ewing, A G. Anal. Chem. 1 9 9 0 , 62, 1872-1876. (6) Chien, J. B.; Wallingford, R. A; Ewing, A G.J. Neurochem. 1990,54,633638. (7) Wallingford, R. A; Ewing, A G. Anal. Chem. 1 9 8 8 , 60, 1972-1975. (8) Stamford, J. A.; Palij, P.; Davidson, C.; Jorm, C. M.; Phillips, P. E. M. In Voltammetric Methods in Brain Systems; Boulton, A A, Baker, G. B., Adams, R N., Eds.; Humana Press: Totowa, NJ, 1995; pp 81-116. (9) Wightman, R. M.; Jankowski, J. A; Kennedy, R. T.; Kawagoe, K. T.; Schroeder, T. J.; Leszczyszyn, D. J.; Near, J. A,; Diliberto, E. J.; Viveros, 0. H. Proc. Natl. Acad. Sci. U.S.A. 1 9 9 1 , 88, 10754-10758. (10) Chow, R. H.; von Ruden, L.; Neher. E. Nature 1992, 356, 60-63.

4514 Analytical Chemistry, Vol. 67, No. 24, December 15, 1995

biological systems. More recently, these techniques have been extended to compounds that are more difficult to oxidize, such as insulin.16 In this work, we examine the utility of carbon fiber-based electrochemical techniques for the detection of histamine, a chemical messenger important in the nervous and immune systems. The work focuses on the measurement of the secretion and storage of histamine in mast cells, which trigger the immune response in mammalian systems.17 The importance of histamine in this role makes it desirable to develop chemical probes to monitor it. Traditionally, histamine has been measured by derivatization with fluorescent reagents followed by chromatographic separation18J9or by a radioenzymatic transfer assay.20 However, these techniques have not been developed for use at the level of individual cells, a condition for which carbon fiber microelectrode techniques are particularly well suited. Mast cells originate in the bone marrow and &e distributed throughout the body in such places as the skin, ear, nose, throat, lungs, peritoneal cavity, intestines, heart, tonsils, and adenoids.z1 Mast cells synthesize,2zstore, and release histamine, a neurohormone that triggers a variety of allergic reactions such as congestion of the airways, wheezing, breathing mculty, itching and pain in the skin, swelling, shock, and even death.I7 Fluorometric studies of rat mast cells have reported an average of 140-240 fmol of hi~tamine/cell'8J3-~~ as well as 2-24 fmol of 5 H T / ~ e l l . ~ * - ~ ~ The average cell is 8 pm in diameterz3and contains -1020 (11) Ciolkowski, E. L.; Cooper, B. R; Jankowski, J. A; Jorgenson, J. W.; Wightman, R. M. J. Am. Chem. Soc. 1 9 9 2 , 114, 2815-2821. (12) Pihel, K.; Schroeder, T. J.; Wightman, R. M. Anal. Chem. 1994, 66,45324537. (13) Jackson, B. P.; Dietz, S. M.; Wightman, R M.Anal. Chem. 1995,67,11151120. (14) Tatham, P. E.; Duchen, M. R; Millar, J. pfugers Arch. 1 9 9 1 , 419, 409414. (15) Alvarez de Toledo, G.; Femandez-Chacon, R; Femandez, J. M. Nature 1993, 363, 554-558. (16) Kennedy, R T.; Huang, L.; Atkinson, M. A; Dush, P. Anal. Chem. 1 9 9 3 , 65, 1882-1887. (17) Lichtenstein, L. M. Sci. Am. 1 9 9 3 , 269, 117-124. (18) van Haaster, C. M. C. J.; Engels, W.; Lemmens, P. J. M. R; Homstra, G.; van der Vusse, G. J. J Chromatog?, 1993, 617, 233-240. (19) Egger, D.; Reisbach, G.; Hultner, L. J. Chromatogr. B 1994,662,103-107. (20) Holgate, S. T.; Robinson, C.; Church, M. K. In Allergy. principles andpractice, 4th ed.; Middleton, E., Reed, C. E., Ellis, E. F., Adkinson, N. F., Yunginger, J. W., Busse, W. W., Eds.; Mosby: St. Louis, MO, 1993; Vol. 1, pp 267301. (21) Pearce, F. L. Pharmacology 1986,32, 61-71. (22) Bauza, M. T.;Lagunoff, D. Biochem. Pharmacol. 1981,30, 1271-1276. (23) Barrett, K. E.; Pearce, F. L. Znt. Arch. Allergy Appl. Immunol. 1983, 72, 234-238. (24) Purcell, W. M.; Cohen, D. L.; Hanahoe, T. H. P. Znt. Arch. Allergy Appl. Immunol. 1 9 8 9 , 90,382-386. (25) Wingren, U.; Wasteson, A; Enerback, L. Znt. Arch. Allergy Appl. Immunol. 1 9 8 3 , 70, 193-199. (26) Bergendorff, A; Uvnas, B. Acta Physiol. Scand. 1 9 7 2 , 8 4 , 320-331.

0003-2700/95/0367-4514$9.00/0 0 1995 American Chemical Society

vesiclesz7in which histamine and 5HT are stored. The vesicles have an average diameter of 780 nmz7 and contain a matrix composed of heparin, a proteoglycan with a high concentration of sulfate residues.28 In vivo, release from mast cells usually occurs when an allergen binds and cross-links two or more immunoglobulin E antibodies attached to receptors on the mast cell.17 This triggers a series of intracellular events that result in the fusion of vesicles with the cell membrane and extrusion of the vesicular contents. This release process, termed exocytosis, is used by many types of cells to release chemical messengers, and the mast cell has been an important model cell for investigation of exocytotic mechanisms. In this work, we first establish the voltammetric properties of histamine at carbon fiber microelectrodes. Electrochemical detection is used for analysis of the content of histamine and 5HT in individual mast cells following separation with capillary chromatography. In addition,we demonstrate with fast-scan voltammetric techniques that histamine and 5HT are coreleased from individual vesicles in mast cells. Previous indirect evidence has suggested histamine and 5HT are stored in the same vesicles in mast ce11s.z5~z9~30~31 For example, the storage of both histamine and 5HT in mast cell vesicles, breaks down at pH 4, and 5HT is capable of displacing tritiated histamine from mast cell vesicles which suggests that the two species might share the same binding sites in the vesicles.z5~z9~30 The measurements here demonstrate corelease from a common storage site at the level of individual vesicular events. EXPERIMENTAL SECTION

Cyclic Voltammetry. Carbon fiber disk electrodes were constructed as previously described32from 10-pm-diameter carbon fibers m o m e l P-55, Amoco Corp., Greenville, SC). The tips were polished at a 45" angle on a micropipet beveling wheel (IC T. Brown Type, Sutter Instrument Co., Novato, CA) to give a 1-pm thickness of glass surrounding the carbon fiber at the tip. Electrodes were electrochemically pretreated before use by application of a triangle potential waveform from +l00 to +1400 to +lo0 mV at 800 V/s every 30 ms in 0.1 M NaOH for 15 s based on the method of A n j ~ . ~ ~ Electrodes were characterized and calibrated in a flow injection apparatus.34The electrodes were used with an E1400 potentiostat (Ensman Instrumentation, Bloomington, IN) in the twoelectrode mode, and measurements were made with respect to a sodiumsaturated calomel reference electrode (SSCE). Fast-scan cyclic voltammetry employed a scan rate of 800 V/s repeated at 16.7ms intervals. Data were acquired at 30 ps/point. The current response was filtered at 5 kHz and digitized with a 12-bit analog to digital converter (Labmaster DMA, Scientific Solutions, Solon, (27) Helander, H. F.; Bloom, G. D. J. Micrusc. 1974,100, 315-321. (28) Schwartz, L.; Huff, T. In Allergy: principles and practice, 4th ed.; Middleton, E., Reed, C. E., Ellis, E. F., Adkinson, N. F., Yunginger, J. W., Busse, W. W., Eds.; Mosby: St. Louis, MO, 1993; Val. 1, pp 135-168. (29) Uvnas, B.; Aborg, C.; Bergendorff, A Acta Physiul. Scand. 1970,78 (Suppl. 336), 1-26. (30) Bergendorff, A; Uvnas, B. Acta Physiol. Scand. 1972,84, 320-331. (31) Siraganian, R P. In Allergy: Principles and practice, 4th ed.; Middleton, E., Reed, C. E., Ellis, E. F., Adkinson, N. F., Yunginger, J. W., Busse, W. W., Eds.; Mosby: St. Louis, MO, 1993 Vol. 1, pp 105-134. (32) Kawagoe, K T.; Zimmerman, J. B.; Wightman, R M. J Neurusci. Methods 1993,48, 225-240. (33) Anjo, D. M.; Kahr, M.; Khodabakhsh, M. M.; Nowinski, S.; Wanger, M. Anal. Chem. 1989,61, 2603-2608. (34) Baur, J. E.; Kristensen, E. W.; May, L. J.; Wiedemann, D. J.; Wightman, R M. Anal. Chem. 1988,60, 1268-1272.

OH) interfaced to an IBMcompatible personal computer with commercially available software (CV6, Ensman). Cyclic voltammograms were background-subtractedwith voltammograms collected immediately before introduction of histamine and 5HT to the electrode surface. Current-time traces of 5HT oxidation were constructed from the average current collected in the 1WmV range around the peak oxidation potential for this wave. Similarly, current-time traces for histamine oxidation were constructed from the data around its peak potential. Cell Culture Preparation. Mast cells were isolated by the method of Atkinson and co-w0rkers.3~Male Sprague-Dawleyrats (350-700 g) were anesthetized with ether and decapitated, and 20-30 mL of tris buffer at 37 "C was injected into the peritoneal cavity. M e r the abdomen was massaged for 1-2 min, the peritoneal cavity was cut open to allow removal of -15 mL of the buffer with a pipet. The buffer was centrifuged at 200g for 5 min and the supematant discarded. The cells were resuspended in fresh buffer and transferred into plastic cell culture dishes (Falcon 3001, Becton Dickinson, Lincoln Park, NJ) that were placed in an incubator for at least 1h before use to allow the cells to adhere to the culture dishes. Cyclic voltammetry experiments were done at room temperature on the same day that the cells were isolated. Chromatographic System. The chromatography columns used were 27- or 40-pm inner diameter fused-silica capillaries (Polymicro Technologies, Inc., Phoenix, AZ), 40 cm long, and slurry packed with YMGODSAQ 5-pm spherical C-18 silica particles (YMC, Morris Plains, NJ). The packing procedure is similar to that described before36with minor modifications. 'The frit for the column was formed by tapping the end of the capillary into a pile of 5pm spherical silica particles. To make room for carbon fiber electrode placement, the particles were forced 1mm into the column with a 25pm tungsten wire. These were formed into a frit by sintering in the flame of a match. The open end of the capillary was placed in a high-pressure stainless steel reservoir containing a slurry of 1:30 (w/v) packing to hexane. An Altex Model llOA pump forced isopropyl alcohol into the reservoir at 207 bar. The packing procedure was typically complete in two hours. The microcolumn LC setup was similar to that described A commercial pump (Model LC-600,Shimadm Corp., Kyoto, Japan) was used to deliver a flow rate of 0.08 mL/ min. Matched inner diameter splitter capillaries were used so that the flow rate through the packed capillary column was -60 nWmin. A cylindrical (1-mm length, gpm diameter) carbon fiber electrode was used as a detector. The current was measured with a Model 427 current amplifier (Keithley Instrument Inc., Cleveland, OH) connected directly to a l6bit analog to digital converter and IBMcompatible computer. The time constant of the current amplifier was 300 ms, and the data collection rate to the computer was 1point/s. Histamine and 3,4dihydroxybenzylamine(DHBA) were detected at an oxidation potential of +1300 mV vs Ag/AgCl. Shortly after elution of DHBA, the voltage was switched to +800 mV vs Ag/AgCl to enable the gain on the current amplifier to be increased by 1order of magnitude for 5HT detection. The voltage change also improved selectivityfor 5HT because fewer species are oxidized at +800 mV than at +1300 mV. (35) Atkinson, G.; Ennis, M.; Pearce, F. L. BY.J. Phamacol. 1979,65, 395402. (36) Kennedy, R T.; Jorgenson, J. W. Anal. Chem. 1989,61, 1128-1135.

Analytical Chemistty, Vol. 67, No. 24, December 15, 1995

4515

Chromatographicpeaks were quantified with calibration runs. In each calibration run, the amount of DHBA remained constant while the amounts of histamine and 5HT were varied. Histamine and 5HT were quantified using DHBA as an internal standard. Measurements at S i e Cells. A culture plate containing the mast cells was placed on the stage of an inverting microscope (Axiovert 35, Zeiss, Eastern Microscope, Raleigh, NC). An electrode was positioned next to a single cell with a piezoelectric micropositioner (Kopf Model 640, Tujunga, CA) . Cell-electrode distance was determined by placing the electrode in contact with the cell surface as visually detected by the cell's deformation and then retracting the electrode 1 pm away. A pressure ejection device (Picospribr, General Valve Corp., Fairfield, NJ connected to a micropipet with a l@pm inner diameter tip was used to introduce secretagogues to the cell. Pressure ejection (0.7 bar for 6 s) solutions consisted of either 40 pM digitonin or 5 pM calcium ionophoreA23187 dissolved in buffer. A freshly polished and electrochemicallytreated electrode was used to take measurements at each cell. Measured currents were converted to concentrations based on the average of pre- and postcalibrations. Typically, the postcalibration was not less than half of the calibration obtained before measurements were taken at the cell. Measured values are reported as means i the standard error of the mean. Contour plots of the results from secretion experiments were constructed in Spyglass (Spyglass, Inc., Champaign, IL). Current was collected every 24 mV during cyclic voltammograms, and cyclic voltammograms were collected every 16.7 ms. For chromatography experiments, single cells were removed and transferred to 3WnL microvials with the aid of a micropipet with a l@pm inner diameter as previously de~cribed.3~Each transfer was accompanied by several nanoliters of balanced salt solution. The microvials were capped with several layers of paraffm-coated wrapping film and placed in dry ice. Five to 10 cells were removed from each plate. Exactly 20.0 nL of a solution of 10pM DHBA in 0.3 N HC104was added as an internal standard to each microvial with a specially constructed microdispenser pipet3$ The capped microvials were centrifuged at 12 O0Og for 8 min. The supernatant from individual vials was injected directly into the chromatography column with a microinjector pipet that has been described previou~ly.~~ Reagents and Solutions. Histamine dihydrochloride, Shydroxytryptamine hydrochloride, and all other compounds were obtained from S i a (St Louis, MO). Solutions of histamine were prepared by dilution of 50 mM stock solutions in 0.1 N HClOh and those of 5HT were similarly prepared from a stock solution of 2.5 mM. The buffer used in the flow injection system consisted of 200 mM Tris-HC1 (PH 7.4) and 150 mM NaCl. Cyclic voltammetric experiments at single cells were conducted in pH 7.4 physiological buffer with 12.5 mM Tris-HC1, 150 mM NaC1, 4.2 mM KCl, 5.6 mM glucose, 1.5 mM CaC12, and 1.4 mM MgC12. Solutions were prepared in distilled, deionized water (Mega Pure System MP3A Corning Glass Works, Coming, NY). For chromatography experiments, the mobile phase was prepared by dilution of 85%phosphoric acid to 75 mM and addition of 0.3 mM hexanesulfonicacid. The pH was then adjusted to 3.0 with NaOH, and the solution was 6ltered through a 0.4Spm Nylon (37) Kennedy, R T.;Jorgenson, J. W.Anal. Chem. 1989,61, 436-441. (38) Kennedy, R T.; Jorgenson, J. W.Anal. Chem. 1988,60,1521-1524.

4516 Analytical Chemistty, Vol. 67, No. 24, December 15, 1995

B

A

, 1000 500

, 0

1000 500

0

mV vs. SSCE

D

C

1000 500

0

5

mV vs. SSCE

10

15

seconds

Flgure 1. Electrochemistry of histamine at a beveled carbon fiber electrode. (A) Cyclic voltammetry (100 mV/s) in tris buffer at an electrochemically treated electrode. (B) Cyclic voltammetry in the presence of 100 p M histamine. (C) Cyclic voltammogram in (A) subtracted from the cyclic voltammogram in (B). Potential limits: +lo0 and +1200 mV. (D) Amperometry of 10pM histamine at 1200 mV at an untreated electrode.

filter (Alltech, Deeriield, IL). These solutions were made in deionized, 0.2-pm filtered water (Barnstead, Dubuque, IA), RESULTS AND DISCUSSION

Cyclic Voltammetry of Histamine. The oxidation of histamine occurs near the anodic potential limit in pH 7.4 buffer (Figure 1). At a scan rate of 100 mV/s, the cyclic voltammogram recorded at an electrochemically treated, carbon fiber electrode in buffer shows the background oxidation at a potential of -1100 mV ( F i i r e l&. Addition of 100 pM histamine (Figure 1B) results in additional oxidation current near the background process. Subtraction of these two voltammograms yields the signal due to histamine (Figure IC). Similar results are obtained at untreated electrodes although the amplitude is smaller. This poorly defined, irreversible cyclic voltammogram does not provide sufficient information to allow investigation of the oxidation mechanism. It is sufficiently robust, however, that it can be used for analytical purposes. Amperometric detection of 10 pM histamine (Eapp = 1200 mv) is shown in Figure 1D. The decrease in current during the oxidation suggests passivation of the electrode surface by electrogenerated products. However, the passivation is short-lived since subsequent exposure of the electrode to histamine results in an identical current amplitude. Because of this temporal distortion, amperometric detection of histamine secreted from cells was not attempted. However, this form of detection is suitable for the detection of histamine separated on microlumns (vide infra). More rapid scan rates are required to monitor secretion from single cells with cyclic voltammetry. At a scan rate of 800 V/s, repeated at 16.7-ms intervals, voltammogramsfor histamine can be obtained using the background subtraction procedure.34The positive potential limit in these scans (1400 mv) was optimum for histamine detection. It is considerably more positive than normally used for catecholamine detection, and it may activate the electrode on each scan. Nevertheless, the sensitivity to

,- B

A

/

b

10nA

7 2

A 5 nA/

- untreated electrode

II

1 I "

1

1300 to 1200 m~

\

treated electrode 1200

,

,

I

,

1

700

200

1200

700

200 100 to 200 mV

mV vs. SSCE

D

I

I

1200

700

___---

I

Figure 2. Background-subtractedcyclic voltammograms at 800 V/s at an untreated (solid line) and electrochemically treated (dashed line) electrode. (A) 20 pM histamine. (B) 0.25 pM 5-HT. (C) A mixture of 20 pM histamine with 0.25 pM 5-HT. The potential scan limits are +lo0 and +1400 mV. (D) Background current at the electrode with no electroactive species present.

Figure 3. Current-time traces of 20 pM histamine obtained in a flow injection apparatus. (A) The upper trace shows the current averaged between +1300 and +1200 mV on the reverse scan, the potential at which the oxidative histamine peak occurs. The lower trace shows the current averaged between +lo0 and $200 mV on the forward scan. (B) The larger current trace is at the peak oxidative potential for histamine after the noise has been subtracted out as explained in the text. The smaller and slower current trace represents the current measured between 900 and 1000 mV.

histamine was l@foldgreater when the measurements were made with electrodes that had been previously electrochemicallypretreated (Figure 2A). At this scan rate the background current is large at all potentials (Figure 2D). However, it is suf6ciently stable to allow the background subtraction procedure. Note that HEPES and other buffers often used with biological systems contain tertiary amines that are electroactive at the potential at which histamine is oxidized and, thus, cannot be used. In Figure 2 4 the cyclic voltammogram of histamine also exhibit a second oxidative peak at 900 mV. This second wave increases in response to a histamine bolus with a slower rate than the main oxidative wave for histamine (vide infra). Thus, it must result from a product formed during the initial oxidation. This second oxidative peak, therefore, cannot be used to quantify histamine since it is time dependent. However, it does provide additional information that histamine is the substance detected. Since 5HT secretion also occurs from mast cells, the background-subtracted, cyclic voltammogramsof this substance were investigated under identical conditions (Figure 2B). Again, the electrochemical pretreatment enhanced sensitivity for the peak at 600 mV. Identiiication of 5HT is also aided by the occurrence of reductive current near the end of the reverse scan. More negative sweeps allow characterization of this wave; however, they degrade the response to histamine. Background-subtracted,cyclic voltammograms of a mixture of the two compounds at a concentration ratio similar to that anticipated from cells30 give signals for each that are additive, thus allowing simultaneous detection and quantification (Figure 2C). Evaluation of Temporal Concentration Changes. To evaluate temporal changes in concentration with fast-scan cyclic voltammetry, the current at a single potential in successive voltammograms can be plotted as a function of time. This is illustrated in Figure 3A for a bolus of histamine introduced into the flow injection system. Since the peak for histamine exhibits hysteresis, the largest amplitude is actually found for the current recorded in the initial portion of the reverse scan (over the range

1300-1200 mv). Thus, this current was used to evaluate histamine concentrations. The likely origin of the hysteresis is a transient activation of the carbon surface at the very positive potentials employed. Thus, although the electrode is less positive than the switching potential, histamine can be oxidized more easily at the activated surface. The direction of the scan must be reversed before histamine oxidation is complete because if the electrode is scanned beyond 1400 mV, the background current increases dramatically at the carbon fiber surface. A large increase in background current is undesirable because it causes a loss in signal-to-noise ratio. The current noise was greater at the switching potential than at potentials intermediate in the scan. This decreases the signalto-noise ratio for histamine, which is detected in this region. Examination of the current-time curves over the whole range of potentials revealed that the increased noise is also present in the initial portion of the scan (100-200 mV on the forward scan, Figure 3A) but not in the more intermediate ranges where 5HT has its oxidative peak. The noise in each region was identical with only a difference in amplitude. Since the changes in background current are largest in the regions of largest noise, a slight timing error in the initiation of the potential scan may be the origin. The current measured in the 100-2o(lmV interval can be used to remove the noise at the potential where histamine is detected. The amplitude of the current is adjusted by multiplication by the ratio of the standard deviation of the noise at the potential where the oxidation of histamine occurs to that at 100200 mV, where no faradaic signals occur. This normalized current was then subtracted from the current averaged between 1300 and 1200mV on the reverse wave to remove the noise. This procedure is quite effective: in Figure 3B, the noise was reduced by a factor of 4. It is not required for the 5HT determination,which is simply obtained from the current over the potential interval of 600-700 mV. With these procedures, the detection limit for histamine was 1.4,uM (S/N = 5) and 12 nM for 5HT. The remarkable detection limits for 5HT are due to its adsorptive nature.I6 Note that this

200

1200

700

200

mV vs. SSCE

Analytical Chemistty, Vol. 67,No. 24, December 15, 1995 4517

Table 1. Amounts of Histamine and S=HTDetected in Individual Mast Cells by Capillary Chromatography

1. Histamine 2. DHBA 3. 5-HT

10

20

30

50 60 minutes

40

concn (fmol)

70

80

90

8.

T 6.0

7.0

8.0

800 DA

9.0

10.0

11.0

12.0

13.0

14.0

15.0

minutes

C.

cell no.

histamine

5-HT

histamine/&HT

1 2 3 4 5 6 7 8 9 10 11 12 mean f SEM

170 201 133 169 150 116 76.2 128 98.7 278 58.5 218 150 i 18

2.5 0.0 2.1 0.0 2.1 1.3 4.9 0.0 8.9 15 4.6 4.2 3.8 i 1.3

68

I " " I " " 1 " " I ' " ' " ' ' 1 " ' ' I ' " I " ' ' I ' " ' 1

87.0

86.0

88.0

89.0

90.0

91.0

92.0

93.0

94.0

minutes

Figure 4. Chromatogram of 50 fmol of histamine, 200 fmol of DHBA, and 50 fmol of 5-HT. (A) The full chromatogram. At 17 min, the potential at the electrode was switched from f1300 to +800 mV vs Ag/AgCI. (6) This trace shows the histamine peak at 6.3 min and the DHBA peak at 13.4 min in greater detail. (C) This trace shows the 5-HT peak at 89 min in detail.

1

A.

/I

Senaitivitv

1. Histamine 2. DHBA 3. 5-HT

100 to 200 mV 200 to 300 mV 300 to 400 mV 400 to 500 mV 500 to 600 mV ... .. ...

900 to I000 mV 1000 to 1100 mV

1100to1200mV 1200 to 1300 mV 1300to 1400 mV

1400 to 1300 mV

-

11 18 13 52 33

2nA

85.0

-

63 71 89 16

1

2 seconds

L--

3 , ' ~ " 1 " ' ' 1 ' " ' 1 ' " ' 1 " ' ~ I " ~ ' l ' ~ ' ' l ~ ~ ' ' I '

20

10

30

50 60 minutes

40

80

70

1300 to I200 mV

100

90

1200 to 1100 mV 1100 to 1000 mV 1000to 900 mV 900 to 800 mV 800 to 700 mV 700 to 600 mV 70

6.0

8.0

9.0

10.0

11.0

12.0

13.0

14.0

15.0

minutes

C.

T*o DA 1 l

'

85.0

~

"

I

'

"

88.0

~

1

'

87.0

600 to 500 mV 500 to 400 mV 400 to 300 mV 300 to 200 mV 200 to I00 mV

Figure 6. Current-time traces of data measured at a mast cell at all potentials along the waveform. All data are averaged over a 100mV range. "

l

'

"

~

88.0

l

"

~

'

89.0

1

'

'

"

90.0

I

'

~

~

91.0

'

I

'

"

'

92.0

I

'

'

'

93.0

~

i

94.0

minutes

Figure 5. Chromatogram of cell No. 5 in Table 1. (A) The full chromatogram. (6) This trace shows the histidine, histamine, and DHBA peaks in detail. (C) This trace shows the 5-HT peak.

waveform is not optimum to minimize filming of 5HT, which can affect its time response. Figure 3B shows the electrode response to a histamine bolus. The oxidative wave that occurs at 800 mV increases at a slower rate than the main oxidative wave. Chromatographic Measurement of 5-€ITand Histamine. Amperometric detection of 5HT and histamine separated on microcolumns was investigated with the potential program described in the Experimental Section. Figure 4 shows a chromatogram obtained following injection of 50 fmol of both 5 H T and 4518 Analytical Chemistry, Vol. 67, No. 24, December 15, 1995

histamine and 200 fmol of DHBA The results with the cylindrical carbon fiber gave a detection limit of 300 amol for 5HT and 270 amol for histamine (S/N = 5). The chromatogram of the supernatant from a single cell (cell 5 in Table 1) with DHBA added as an internal standard is shown in Figure 5. Since only part of the sample in the microvial was removed and only part of the sample in the pipet was injected onto the column, DHBA allowed for correction of transfer efficiency and the electrochemical response of histamine and 5HT. Buffer that was removed from near the cells in the cell plate as a control was also chromatographically analyzed. These injections revealed that there were 4 fmol of histamine, and no detectable 5HT, present in a solution volume that is approximately twice as large as normally removed during cell retrieval.

SdO

Id0

mV vs. SCCE

750

Id0

Id00 ,14W

I

Sd0

\ I000

7

I 100

5b0

I

\

1000

I

1400

\

t

100

I 1400

I

I

k

-1-

I I

I

I

I

I

I

O.OOnA

-1.25nA

,

I

,

8

9

10

, 11

, 12

13

14

15

16

17

tlme(SeC) Figure 7. Contour plot of the data in Figure 5. Horizontal line 1 indicates the peak histamine oxidative current that murs at +1250 mV on the reverse scan. The direction of the scan is from the bottom to the top of the figure. Horizontal line 3 marks the peak oxidative current for 5-HT that occurs at +650 mV. Horizontal line 2 indicates a secund oxidation peak that is Observed for histamine. The reduction of 5-HT can be observed at 100 mV on the reverse Scan (top of graph). Cyclic voltammograms are shown for three time wink. indicated bv vertical dashed lines, in the contour plot.

Table 1 summarizes the chromatographic results from 12 individual cells. The average amount of histamine and 5HT per cell was 150 and 4 fmol. respectively. This falls into the range reported p r e v i o u s l ~ . ~The ~ ~ average - ~ ~ ~ ratio of histamine to 5HT in a cell was calculated by addition of the ratio of 5HT to histamine in each cell divided by the number of cells. The inverse of this number reveals that the average ratio of histamine to 5HT was 33 for 12 cells. The amino acid histidine is a precursor to histamine formation and is expected to be present in the cell cytoplasm outside the vesicles. Histidine is oxidized at the Same potential as histamine and was detected in half of the cells studied. The average amount of histidine per cell was 22 f 9 fmol. Histidine is not expected to be observed during cyclic voltammetric measurements at individual cells because the amount of histidine inside the cell is small compared to the amount of histamine present and histidine is not w e d to be inside the vesicles. Furthermore, sensitivitytoward histidine is less than that toward histamine because unlike histamine. which is positively charged at physiological pH, histidine is neutd. Release Measured from Individual Mast Cells by cyclic Voltammehy. The chromatographic analysis conrirms that histamine and 5HT are costored in individual mast cells. Cyclic

voltammetric measurements recorded at the cell surface allow examination of the characteristics of their secretion. Figure 6 shows a series of currents recorded with an electrode adjacent to a rat peritoneal mast cell that was exposed to digitonin for 3 s to allow entry of extracellular Caz' into the cell. These m c e s were constructed from successive cyclic voltammograms. and each trace is obtained from a different potential window, each of lo& mV width. In the potential region from 100 to 500 mV on the anodic scan, changes in current are not observed or they consist of small. high-frequency changes. However, at more positive potentials, the oxidation current exhibits a series of spikes at later times. Similar, coincident features can be seen in the cathodic portion of the scans. Such concenhation changes in the form of spikes are observed at adrenal medullary cells during exocytotic secretion of catecholamines.L10~'2Thus, the pattern of concenhation changes observed at mast cells is that expected for a series of exocytotic release events. An alternate view of these data is given by a contour plot ( F i r e 7). This plot c o ~ that s the spikes occur only at potentials where histamine and 5HT have cyclic voltammetric features. The horizontal asymmetry of the spots during the Current spikes shows that the current rises quickly but decays more slowly during a spike. The 5HT spikes are more d ~ c u l t Anawical Chemistry, Vol. 67, No. 24, December 15, 1995 4519

to see than the histamine spikes because the 5HT current is smaller than that of histamine. The reduction of 5HT can also be seen at 100 mV on the reverse scan. The simultaneous release of 5HT and histamine is readily apparent from these primary data. The current observed in the 600-700-mV range on the forward scan (horizontal line 3) is associated with an increase in reductive current in the 200-100 mV range, consistent with the voltammetry of 5HT described above. The oxidation of histamine can be observed in the 13001200mV range on the reverse scan (horizontal line 1). In addition, the secondary wave for histamine oxidation emerges in the 9001000-mV-range during the forward scan (horizontal line 2). Because the current traces are averaged in 1WmV ranges, the peaks for 5HT oxidation and the peak at 900-1OOO mV are not as well resolved as when the individual, background-subtracted cyclic voltammograms are examined. Cyclic voltammograms are shown for three time points in the contour plot. The close temporal relation of histamine and 5HT spikes seen in Figure 6 indicate that both are secreted from the same subcellular organelles. Usually, at the time of a spike, both substances are detected. Thus, these results provide direct evidence of costorage and secretion of these substances in mast cells. Chemical messenger cosecretion is thought to be a common mode of extracellular communication of biological cells, but these measurements represent the first time it has been directly observed at the level of single vesicular events. Characteristics of Release Events. The characteristics of release were examined at several individual cells. Release was induced by exposure of the cell to digitonin, as above, or A23187, a calcium ionophore that allows its selective entry into the cell. The frequency of release events was similar with each agent (an example of release induced by A23187 is shown in Figure 8), but there was a longer delay between stimulation with A23187 and release. The delay before release was typically 30-60 s, which is similar to previous reports.39 Release normally continued for 2-3 min with both secretagogues. After release, the cell was visually altered and had a more poorly defined cell membrane. The characteristics of 667 histamine and 5HT spikes from 11 cells were examined in more detail. The width at half-maximum of the spikes was 118 f 2 ms for histamine and 153 5 3 ms for 5HT. A difference in widths of the 5HT and histamine spikes is observed because the electrode response to 5HT is slower because of adsorption. However, these spikes are significantly wider than those found for catecholamine secretion where the widths were 70 i 2 ms.12 This indicates that the kinetics of secretion are slower in mast cells with their larger vesicles. The ratio of the histamine concentration to the 5HT concentration is shown for individual vesicles in Figure 8C. This ratio varies randomly with time. The average concentration observed at the spike maxima of 11 cells stimulated with A23187 was 26 f 2 nM 5HT (all data reported as mean f SEM) and 3.68 f 0.11 pM histamine. If the ratio of 5HT to histamine concentration is averaged for 667 spikes (including those spikes where the 5HT measurement was below a signal-to-noise ratio of 5 and therefore counted as undetectable),the inverse of this ratio is 106 f 9. This ratio is -3 times larger than the ratio measured for cells by capillary chromatography. One source of error is that the ratio is an overestimate because many of the 5HT concentrations were (39) Garland, L.G.; Mongar, J. L. Int. Arch. Allergy Appl. Immunol. 1976,50, 27-42.

4520 Analytical Chemistty, Vol. 67, No. 24,December 15, 1995

100

1

do

I

70

ao

90

100 110 120 130 140 160

seconds Figure 8. Current spikes obtained at a mast cell stimulated with calcium ionophore A231 87. The currents have been converted to concentration values from calibrations. (A) Current averaged between +1300 and +1200 mV on the reverse scan. This current corresponds to histamine release. (6) Current averaged between +600 and +700 mV on the forward scan. This current corresponds to 5-HT release. (C) Ratio of the histamine current to the 5-HT current at the peak of the spikes. The bars with negative values indicate histamine spikes whose corresponding 5-HT spikes had a signal-to-noise ratio of less than 5.

taken as zero. Furthermore, because the time response of the electrode to 5HT was slower than the time response to histamine, the calculated ratio would be expected to be higher than the actual ratio because the true 5HT peak concentration was never measured by the electrode. 5HT may be present in the cell cytoplasm at a higher concentration than histamine. Previous reports have shown that 5HT may be stored in the cell cytoplasm and can be released nonexocytotically from the cell without concurrent histamine relea~e.~O