Effects of capillary red cell density on gas conductance of frog skin GARY M. MALVIN Oxygen Transport

AND

STEPHEN

MALVIN,GARY M., AND STEPHENC. W~~~.Effects lary red cell density

C. WOOD

Program, Lovelace Medical Foundation,

ofcupil-

of frog skin. J. Appl.

on gas conductance

Physiol. 73( 1): 224-233, 1992.-We tested experimentally the hypothesis that decreasing capillary red blood cell (RBC) density (d,,c) reduces the tissue diffusing capacity of frog skin to CO (Dtico) and 0, (Dti,,). The effects of dRBCon CO, transport were also assessed. C180, 0,, and CO, transport between the skin and a cutaneous sample chamber on the belly of anesthetized (halothane) frogs (Ram pipiens) was measured by mass spectrometry, and the cutaneous conductances to Cl80 (Gco), 0, (GO,), and CO, (Gco,) were calculated. The dRBCof the planar cutaneous capillary bed was measured by intravital fluorescent video microscopy. Dtico and Dtioz were calculated from a modification of the Roughton-Forster equation: l/G = l/Dti + l/(&c d,,c), where eRBCvalues were estimated from literature values. In one group of animals (n = 6), measurements were made before hemodilution (d,,c = 630 t 56 cells/ mm2), after one hemodilution (dRBC= 349 t 50 cells/mm2), and after a second hemodilution (d,,c = 150 t 31 cells/mm2). In controls, time had no effect on GCO, GO,, or Gco2 (P > 0.42). Before hemodilution, GCO, GO,, and Gco2 were 0.069 t 0.010, 0.088 t 0.0012, and 1.23 & 0.010 nmol. min-’ l Torr-‘.~rn-~, respectively, and lowering dRBCby hemodilution decreased all these parameters (P < 0.025). The mean slopes of GCO, GO,, and GCO, vs. dRBCwere 6.0 t 1.3 X 10S7,7.2 * 2.3 X 10S7,and 7.8 * 3.0 X 10B6nmol min-’ Torr-’ RBCY, respectively. Lowering dRBCalso decreased Dtico and Dtio, (P < 0.034). Dtico and Dtio were 0.080 t 0.012 and 0.096 t 0.013 nmol. min-’ $ or+ cmm2, respectively, before hemodilution. The mean slopes of Dtico and Dtio, vs. d,,, were 4.9 * 2.1 X low7 and 6.5 t 2.8 X 10S7nmol min-’ Torr-’ RBC?, respectively. Hemodilution had no effect on perfused capillary density (P = 0.38). These results indicate that tissue diffusive conductance is proportional to dRBc. Regulation of dRBCmay be an important mechanism modulating diffusive gas transport in tissue. l

l

l

l

l

l

l

l

l

hematocrit; cutaneous gas exchange; amphibians; diffusing capacity; Rana pipiens

THE HEMATOCRIT of blood within capillaries is much different from the hematocrit in large vessels. Capillary hematocrit is normally only lo-50% of the hematocrit in large blood vessels, and it can vary rapidly over a wide range, whereas large vessel hematocrit remains constant (4,17,29). For example, hyperemia due to tissue hypoxia or vasodilation causes capillary hematocrit to increase rapidly severalfold over prehyperemic levels (4, 17, 29). The precise significance of capillary hematocrit and changes in this parameter to tissue gas exchange is not known. It has been proposed that capillary hematocrit is an important determinant of diffusive gas transport, 224

0161-7567/92

$2.00 Copyright

Albuquerque,

New Mexico 87108

both in peripheral tissue (6,7,16) and across respiratory organs (1). Because 0, and CO, move across red cell membranes in gas uptake and delivery, the density of red cells within a capillary (dRBC) should determine the functional surface area available for gas transport into and out of blood. If this is so, then the diffusive conductance (diffusing capacity) of a capillary bed to respiratory gases should be a function of the d,,c. Modeling studies (6,7, 16) predict that tissue diffusing capacity (Dti) is proportional to dRBC, but to our knowledge, no study has actually measured both Dti and dRBCto test those predictions. The primary purpose of this study was to determine experimentally the relationship between dRBC and the Dti. The major hypothesis tested was that a change in dRBc alters Dti to gases carried in the blood primarily by red blood cells. Support for this hypothesis would indicate that Dti is regulated, in part, by d,,c. All experiments were performed on the skin of anesthetized frogs, a preparation well suited for measuring both Dti and dRBc of the cutaneous capillary bed. Frog skin structure is relatively simple and homogeneous. A rather uniform capillary network lies in a plane directly beneath the epidermis. The thickness of the epidermis is also constant. Most cutaneous gas exchange probably occurs between the environment and the capillaries. Little gas exchange is expected between the environment and the larger vessels, because the arterioles and venules run perpendicular to the skin surface and the larger arteries and veins lie approximately eight times farther from the skin surface than the capillaries (28). Frog skin is also readily accessible. Gas exchange across discrete regions of skin can be accurately measured (22), and the microvasculature can be observed without surgical intervention by use of a microscope. Thus frog skin is a simple relatively homogeneous respiratory organ in which gas exchange and microvascular parameters can be accurately measured. MATERIALS AND METHODS

We determined total conductances of frog skin to C180, 0,, and CO, by measuring the kinetics of gas transport into and out of a cutaneous sample chamber by mass spectrometry. Dti for 0, (Dti,,) and for CO (Dti,,) were calculated from these data. Cell density in the skin capillaries was assessed by intravital video microscopy immediately after the gas exchange measurements. Measurements were performed before and after dRBCwas reduced by hemodilution.

0 1992 the American

Physiological

Society

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

CAPILLARY

HEMATOCRIT

AND

Animals

Leopard frogs (Rana pipiens) weighing 71 t 11 (SD) g (n = 26) were kept in aquaria at -22OC 22 wk before experimental

use.-They were fed crickets twice per week.

Surgical Preparation

Frogs were anesthetized by submersion in a 0.3% MS(Sigma Chemical). A cannula (PE-50) was placed in the sciatic artery for blood withdrawal, injection, and arterial pressure measurement. Immediately after surgery, blood was withdrawn for labeling (see below). Frogs were returned to aquaria and allowed to recover for 24 h before experiments began.

TISSUE

GAS

CONDUCTANCE

225

sciences). After hydroethidine was added to the cells, the cells and solution were mixed and incubated at 23OC for 2 h in the dark. The cells were washed twice in diluted HB Basal medium, suspended in amphibian Ringer solution, and then injected into the frog via the sciatic arterial cannula. The Ringer solution contained (in mM) 76.1 NaCl, 30.7 NaHCO,, 1.4 NaH,PO,, 2.5 KCl, 3.1 MgSO,, and 5.6 glucose.

222 solution

Microcirculatory

Measurements

In an initial experiment, flux and velocity of both labeled and unlabeled cells through skin capillaries were measured to assess the in vivo rheological effects of cell labeling with ethidium. In experiments designed to test the effects of dRBCon cutaneous gas exchange, capillary Arterial Pressure, Heart Rate, and Body Temperature red cell velocity, density of cells in the cutaneous capilMeasurements lary bed, and density of perfused capillaries in the skin Arterial pressure was measured continuously by con- were determined. necting the sciatic arterial cannula to a pressure transAll microcirculatory measurements were performed ducer (Statham). Transducer output was recorded on a under anesthesia (halothane or methohexital sodium). chart recorder. Heart rate was determined from the arteThe animal was placed supine on a wire mesh beneath a rial pressure record. Skin and cloaca1 temperature were Nikon Optiphot microscope. The ventral pelvic skin was measured with thermocouples. observed under both bright-field and fluorescent epi-illumination (100-W mercury lamp, 540- to 560-nm excitaBlood Gases and Hematocrit tion filter, and 605-nm interference filter). Fluorescent was performed to record only the laArterial PO, (Pa,,) and pH (pH,) were measured at epi-illumination beled red blood cells. The magnified image was recorded 22OC with a Radiometer BMS 3 MK 2 blood micro sysby a videocamera (model CCD-72, Dage MTI), video tem on blood samples withdrawn from the sciatic artery. monitor (model VM-1220U, Hitachi), and video recorder Arterial PCO, (Pa,,,) was determined from the sample chamber PCO, measurements at the end of the l-h mea- (model AG-1950, Panasonic). Magnification from skin to surement period (see Experimental ProtocoZ). In all sam- monitor was x670 with an x20 objective. In experiments comparing the rheological properties of labeled and unple cha .mber gas mea surements, the chamber Pco, labeled cells, five to eight capillaries per animal were reached a con .stant level within 1 h. We assumed th at this steady state was caused by equilibration of chamber PCO, studied. In the other experiments, three different regions of the ventral skin were recorded for each experimental with arterial PacoO. In preliminary experiments, PacoM I 0 condition. values measured with this technique did not differ signifir Capillary red cell flux (cells/s) was determined from cantly from those obtained on arterial blood samples the videotapes by counting the number of red blood cells measured with a mass spectrometer blood gas probe. Arcrossing a transverse line placed in the middle of a capilterial hematocrit was measured on the blood withdrawn lary during a measured time period. The flux of all red from the sciatic artery after each gas exchange measureblood cells in a capillary was determined by counting all ment. Two 150-~1 capillary tubes were filled with blood cells under bright-field epi-illumination crossing the line and centrifuged at - 10,000 g for 5 min. during a 20-s period. Capillary red cell flux was estimated from labeled cells by counting under fluorescent epi-illuCell Labeling mination the number of labeled cells crossing the line Labeling a fraction of the total cells was necessary be- during a 2-min period. Labeled cell flux was then divided cause the high red cell density in many of the capillaries by the labeled cell fraction to obtain an estimate of total made it impossible to count and track individual cells cell flux through the capillary. when all cells were viewed under bright-field epi-illumiRed cell velocity (mm/s) was determined from the vidnation. After cannulation, -8% of a frog’s red blood cells eotapes from the time required for a red blood cell to were labeled with ethidium for fluorescent video microsmove a measured distance through a capillary. To comcopy of the cutaneous circulation. In this procedure, 0.55 pare labeled and unlabeled cells, the velocity of 20 cells ml of blood/100 g body wt were withdrawn and centriunder both bright-field and fluorescent epi-illumination fuged and the plasma was removed. For each milliliter of was determined in each capillary studied. In the other blood that was withdrawn, 2 ml of a hydroethidine (Polyexperiments, the velocities of 75 labeled cells were detersciences) solution were added to the cells. The hydroethmined in each skin region observed. Because three skin idine solution contained 0.15 mg hydroethidine/ml of di- regions were observed under each experimental condiluted (80% of original concentration) HB Basal medium tion, 225 cells were measured for each experimental con(Hana Biologicals). Hydroethidine was added to the HB dition. Basal medium from a stock solution containing 80 mg The density of red blood cells in the cutaneous capillarhydroethidine/ml of N’JV-dimethylacetamide (Polyies was determined by counting the number of labeled Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

226

CAPILLARY

HEMATOCRIT

AND

cells in a videoframe and then dividing that number by the fraction of total cells that were labeled and by the area of skin covered by the videoframe (0.11 mm2). Fifty different videoframes were analyzed for each region of skin observed (150 frames/experimental condition). Perfused capillary density (the length of perfused capillaries per unit area of skin) was determined by measuring the length of perfused capillaries in a video field and dividing that length by 0.11 mm2. Capillary length was measured by fitting line segments of 10 pm along the capillaries. This analysis was performed using images recorded under bright-field epi-illumination. The fraction of cells that were labeled was determined after each experimental period by dividing the number of labeled cells by the total number of cells counted in a blood sample taken from the sciatic artery. Cells were counted by diluting a blood sample ~20 times with amphibian Ringer solution and placing a drop of this cell suspension in a hemocytometer. Twenty different regions of the hemocytometer were recorded by video microscopy under both bright-field and fluorescent epi-illumination. Cell counts were made from the recorded images. Approximately 2,000 cells were counted in each blood sample. Red Cell Suspension Viscosity Measurements

To determine whether cell labeling had an effect on red cell aggregation, the viscosity of labeled and unlabeled cells suspended in amphibian Ringer solution was measured at 22.0 t 0.3*C with a Wells-Brookfield cone/plate viscometer (DV-II) at shear rates of 6,12,30, and 60 s-l. Viscosity was measured as a function of shear rate to assess possible changes in red cell aggregation. Gas Exchange Measurements

A frog was anesthetized with halothane and placed supine on an elevated wire mesh in a 5.1-liter Plexiglas box containing an open beaker of halothane (Fig. 1). Humidified air was drawn through the box at 700 ml/min. This arrangement produced a constant concentration of halothane in the box of -1.1%. This concentration of halothane maintained anesthesia at a constant level sufficient to eliminate lung and buccal ventilation. Consequently, all respiratory gas exchange occurred across the skin. A small stainless steel sample chamber was placed on the ventral abdomen. The chamber had a volume of 0.58 ml and covered 2.55 cm2 of skin (-2.5% of the total skin area). The weight of the chamber downward on the skin produced an airtight seal between the chamber and skin. Leaks of gas into the chamber were checked periodically by flowing helium around the chamber edges while monitoring chamber helium partial pressure by mass spectrometry. No leaks were ever detected. The chamber housed a small gas probe (AMIS A/S, Odense, Denmark) and a thermocouple for measuring chamber temperature. The gas probe consisted of a piece of sintered brass (1 mm diam, 4 mm long) covered by a 20-pm-thick polyethylene membrane. The membrane-covered brass was attached to a piece of stainless steel tubing that was connected to a quadrupole mass spectrometer. The 63% re-

TISSUE

GAS

CONDUCTANCE

sponse times of the probe for O,, CO, and CO, were 25, 23, and 33 s at 22*C. The O,, CO, and CO, conductances of the probe were 6, 6, and 36 pm010 min-l .Torr-l, respectively. These values are ~5% of the conductances of the area of skin measured in the experiments. Consequently, probe gas consumption had no significant effect on chamber partial pressures during the experiments. The sample chamber also had inflow and outflow ports fitted with gastight valves so that a gas mixture could be introduced into the chamber and then trapped beneath it. The mass spectrometer system was based on a Balzers QMA 120 mass analyzer. The inlet system, controller programs, and data acquisition system were produced by AMIS. The mass spectrometer was controlled by a computer with analog and digital interface cards. The mass spectrometer output was recorded on-line and stored on disk for later analysis. To measure cutaneous gas exchange, a gas mixture containing 99% air-l% Cl80 was introduced into the chamber and then trapped beneath it by closing valves on the inflow and outflow ports. Chamber PO,, Pco,, and Pc180 were measured for 1 h by the gas probe and mass spectrometer. Mass spectrometer outputs for each of the three gases were averaged over 5 s and then stored. The mass spectrometer was calibrated before and after each experimental measurement by flowing calibration gases through the chamber. Differences in calibration signals before and after the experimental measurement were always ~5%. These differences were assumed to have been due to a linear drift with respect to time, and experimental values were corrected accordingly. Calculation

of Cutaneous

The conductance lated as follows

Gas Conductance

of the skin to CO (Gco) was calcu-

Gco=-mln

V

(1)

where R is the universal gas constant, T is temperature (OK), V is chamber volume, Psccoi is initial sample chamber PCO, and Pscco, is sample chamber PCO at time t. This equation is not different from that used in the calculation of lung conductance to CO (10). In (Pscot/Psc,o.) lt was calculated from the slope of a leastsquares linear regression that was fit to In (Psc,,~) as a function of time (see Fig. 6). The cutaneous conductance to 0, (GO,) was calculated in a similar manner Go 2

-- V =

RTt

In (Psc,, - Pa,,),

(Psc,, - PaO2)i

(2)

Pa,, was determined immediately after the gas exchange measurements. The cutaneous conductance to CO, (Gco,) was calculated in a similar manner -- V

In (P%O~.f - paco2,, ) (3) Psc C02,f where Psccoz f is the sample chamber PCO, after it reached a steady level (see Fig. 5). Pscco2 data were Go 2

=

RTt

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

CAPILLARY

HEMATOCRIT

AND TISSUE

GAS CONDUCTANCE

227

AIR

Sample

FIG. 1. Experimental setup for measuring cutaneous gas exchange.

corrected for the probe response time as described in Grplnlund et al. (11). It was not necessary to correct PO, and PCO data for probe response times, because the probe rate constants were several hundred times larger than the rate constants for cutaneous 0, and CO uptake from the sample chamber. Calculation

of Dti,,

and Dti,,

. Dti co was calculated using a modification of the Roughton-Forster equation that was developed to assess the membrane diffusing capacity of the lung (27) -=-1

1

GCO

Dti,,

1 (4)

+

~~BC(CO)

l

&BC

where &&CO) is the CO uptake capacity (conductance) of a single red blood cell. Dti,, was calculated in the same manner. Use of Eq. 4 requires that gas uptake be diffusion limited. This assumption is probably valid for CO because of the extremely high affinity of hemoglobin for CO. This assumption has been used in numerous gas exchange studies on a variety of gas exchange organs. This assumption is probably valid also for 0,, because gas exchange studies by others indicate that 0, uptake across amphibian skin is almost entirely diffusion limited (24, 25). tiRBC(0,) and flRBC(CO) are not known for Rana pipiens, so these parameters were estimated. O&OQ) was estimated by calculating flRBC(0,) for a turtle at 22OC with the data of Yamaguchi et al. (32). ORBC(02) for the frog was then estimated with the relationship between mean cell volume (MCV) and 0&O,) (33): BRBC(02) (frog) = 8,& 0,) (turtle) [ MCV(turtle) /MCV( frog)]2’3. Values used for turtle (32) and frog (31) MCV were 330 and 680 Pm39 respectively. This procedure yielded a value for 8,,(0,) of 2.7 X lOa nmol min-’ Torr? B,,,(CO) was estimated by multiplying && 0,) by 0.5 (9). l

l

l

Experimental

Protocol

Effects of cell labeling on blood rheological properties. To assess the in vivo properties of labeled cells, frogs (n = 7) were anesthetized with methohexital sodium (15 mg/kg). Halothane was not used in these experiments because it would not maintain anesthesia long enough for these measurements to be completed. In each animal, between five and eight capillaries were examined under both bright-field and fluorescent epi-illumination. Capillaries chosen for observation were those that had the best red cell images. This capillary selection was necessary for accurate measurement of cells under bright-field epi-illumination. To assess in vitro the rheological properties of labeled cells, frogs (n = 8) were pithed and then blood (~1.5 ml) was collected by cardiac puncture into heparinized tubes. Half the blood from each frog was labeled with ethidium as described above. The other half was treated exactly the same, except ethidium was not added to the cells. Hematocrit was adjusted to 0.22. Viscosity as a function of shear rate was determined on both samples. Gas exchange and microcirculation experiments. Frogs were anesthetized with halothane, placed in the box, and instrumented as described above. The first gas exchange measurement was performed 30 min after arterial pressure reached a steady level (l-2 h after instrumentation). Immediately after the gas exchange measurements, the animal was placed under the microscope for measurements of dRBc, red cell velocity, and perfused capillary density. Recordings of the microcirculation were completed within 10 min. Then 1.8 ml of blood/100 g body wt (-50% of total blood volume) were withdrawn from the sciatic artery for blood gas, pH, and hematocrit measurements. The same volume of plasma from donor animals was infused into the animal. The animal was then returned to the box. One hour later (-30 min after arterial

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

228

CAPILLARY

HEMATOCRIT

AND TISSUE

GAS CONDUCTANCE

a.0 G

0

go.8 -I II n

v

V

I

0

2 Unlabeled4

0 . v v I

Cells,

VGiscosity

0

0

/

0

0.6

-

0

-1

60 s . -1 30 se, 12 s-, 6 s I “(cP)

I 10

FIG. 2. Comparison of viscosities of labeled and unlabeled red cell suspensions (n = 8). Solid line, line of identity. At each shear rate, there was no significant difference in viscosity between labeled and unlabeled cell suspensions (P > 0.67).

pressure reached a steady level) the above protocol was repeated. In four of the six animals, this protocol was performed again. Five frogs were used in time control experiments. These experiments were identical to the hemodilution experiments, except hemodilution was not performed. Measurements were performed twice in these experiments. The two measurement periods were separated by 3 h. The video-microscopy data were not analyzed in the time controls, because there was no effect of time on cutaneous gas exchange, blood gases, arterial pressure, heart rate, and hematocrit (see Table 2).

Q) 0.0 > 0.0

I

I

I

I

0.2

0.4

0.6

0.8

Velocity

(mm

l

s-l)

-

All

I 11.0.o

Cells

3. Velocity of labeled cells measured under epifluorescent illumination compared with velocity of red cells measured under brightfield conditions. Points represent mean red cell velocities determined in the same capillary. Forty-seven capillaries were examined from 6 different animals. There was no significant difference between the 2 velocity measurements (P = 0.80). FIG.

the total red blood cells were labeled. There was no significant difference between the velocity of labeled red blood cells measured under fluorescent epi-illumination and red cell velocities measured under bright-field epi-illumination (P = 0.80; Fig. 3). Capillary red cell fluxes determined from measurements on all cells and from labeled cells/labeled cell fraction were not significantly different (P = 0.15;Fig. 4).

Statistical Analysis

Gas Exchange and Microcirculation

The relationship between the cutaneous conductance to the three gases and dRBCwas evaluated by calculating the mean slope of conductance vs. dRBCfor each animal and then testing whether the slope was greater than zero with a paired t test. The relationships between Dti and dRBCwere assessed in the same way. The viscosity of labeled and unlabeled cell suspensions was compared at each shear rate with paired t tests. Red cell velocities of labeled and all cells were compared by calculating the mean labeled and combined cell velocity for each animal. Then the samples of mean velocities were compared with a paired t test. The same procedure was used to compare measured total cell flux and the total cell flux calculated using labeled cells. The effects of hemodilution on blood gases, pH, hematocrit, arterial pressure, heart rate, red cell velocity, and perfused capillary density were assessed by analysis of variance. Post hoc comparisons of multisample means were performed with the NewmanKeuls test. The effects of time on blood gases, pH, hematocrit, arterial pressure, and heart rate were determined with paired t tests. Statistical significance was set at

Sample chamber temperature was 22.4 t 0.2OC. There were no differences among chamber, cloacal, and skin temperatures. An experimental record from a gas exchange measurement is shown in Fig. 5. After 1 h, ~2040% of the initial sample chamber 0, and CO was taken up by the skin. Sample chamber PCO, usually reached a steady level after 20-30 min. The natural logarithm of the partial pressure differences for these gases between arterial blood and the sample chamber is plotted against time in Fig. 6. For all measurements, the r2 values for

P =

Experiments

0.05.

RESULTS lL

Comparison of Labeled and Unlabeled Red Blood Cells

The viscosity of labeled and unlabeled red cell suspensions from each frog at four different shear rates is shown in Fig. 2. At each shear rate, there was no significant difference in viscosity between the two cell suspensions (P > 0.67). In the in vivo experiments, 8.1t 0.5%of

0

5

Flux

10

(cells.

s-‘)

15

-

All

20

Cells

FIG. 4. Comparison of capillary red cell fluxes determined from measurements on all cells and from labeled cells/labeled cell fraction. Points represent red cell fluxes determined in the same capillary. Forty-seven capillaries were examined from 6 different animals. There was no significant difference between the 22 red cell flux measurements

(P = 0.15).

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

CAPILLARY

HEMATOCRIT

AND TISSUE

229

GAS CONDUCTANCE

(P = 0.35) were not affected by hemodilution (Table 2), but pH, was reduced (P = 0.009; Table 2). Reducing dRBC decreased the cutaneous GCO (P < 0.003), Go2 (P < 0.012), and GCO~ (P < 0.025; Fig. 8). The slopes of GCO, GO,, and GCO, vs. dRBC are 6.0 t 1.3 X 10 -7, 7.2 t 2.3 X 10m7, and 7.8 t 3.0 X 10e6 nmol min. Torr-l red blood cell-l, respectively. The calculated Dti,, (P = 0.034) and Dtioz (P = 0.032) were also decreased by reductions in dRBC(Fig. 9). Mean slopes for Dti,, and Dtio, vs. dRBCare 4.9 t 2.1 X loo7 and 6.5 t 2.8 X lo-’ nmol . min-l. Torr-l red blood cell-‘, respectively. l

01

I 0

10

I 20

I 30

I 40

l

I

I

50

60

DISCUSSION

This study provides the first experimental demonstration that reductions in dRBCdecrease both GCO, GO,, and GCO, of frog skin and Dti,, and Dti,,. The effect of d,c on Dtico and Dtio, supports the hypothesis that the sur-

n

u 0 ~60

-

n

40

-

1.9

20 I

0’ 0

10

I

I

20

I

30

40

I

50

I

1.7

60

15

aF -C-J 1.6 wa -c 1.5

F 10 0 22 04 En 5

1.4

. :.

I 1.3

I

' 0 r

4.8

10

I

I

20

I

30

40

I

50

9.. ,4 l

.

I

60

4.7

0

0

10

20

30 Time

FIG.

40

50

60

(min)

5. Mass spectrometer record of a gas exchange measurement.

2 4.6 no I 4.5 0”

=

least-squares linear regressions fit to these data were 0.88 t 0.05,0.99 t 0.01, and 0.99 t 0.01 (SD) for CO, 0,, and CO,, respectively. In the control experiments, time had no significant effect on cutaneous GCO (P = 0.83), Go2 (P = 0.43), and GCO (P = 0.42); Paoz (P = 0.18); Pace, (P = 0.85); pH, (P = 0.26); arterial hematocrit (P = 0.10); arterial pressure (P = 0.56); and heart rate (P = 0.33; Table 1, Fig. 7). Hemodilution reduced both sciatic arterial hematocrit (P < 0.0001) and dRBc in the skin (P < 0.0001; Table 2). The first hemodilution decreased arterial hematocrit to 52.0 t 4.5% and dRBc to 54.7 t 4.6% of the original values. After the second hemodilution, arterial hematocrit and dRBCwere 23.8 t 2.8 and 25.5 t 2.3% of original values, respectively. Red cell velocity was not affected by the first hemodilution (P > O.l), but after the second hemodilution, it fell to 78.5 t 4.8% of mean red cell velocity before hemodilution (P c 0.05; Table 2). The density of perfused capillaries was not affected by hemodilution (P = 0.38; Table 2). Hemodilution significantly reduced mean arterial pressure (P = 0.03) but had no effect on heart rate (P = 0.59; Table 2). Pao, (P = 0.12) and Pacoz

4.3 4.2 4.1 0

10

20

30

40

50

60

2

-2 0

10

20 Time

30

(min)

FIG. 6. Natural logarithm of partial pressure gradient between sample chamber and arterial blood as a function of time. Data are from the same record as in Fig. 5. Solid line, least-squares linear regression fit to data. Pscoz , PO, in sample chamber; Pa, , arterial PO,, Pscco, f, sample chamber PCO~ after it reached a steady fevel; Psccoz,, sample’chamber PCOz at time t.

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

230

CAPILLARY

HEMATOCRIT

AND TISSUE

0.10 cv ‘E ? 0.08 7L L f

0.06

7 .-z.

0.04

TABLE 1. Effect of time on blood gases, pH, Hct, arterial pressure, heart rate, and cutaneous gas conductances

iYE==-==:

0 V

V

5 0.02 0 0 urn 0.00

0.06

0.04

C

0.02 0” CY 0.00

1

2

2.0 E 0

. min-’

Torr-’

l

cmS2

l

l

l

8.3kO.6 13.1t1.5 7.64kO.05 22.5tl.O 17t2 42+1 0.055t0.003 0.073~0.004 1.1kO.l

13.3k1.8 7.62t0.04 25.0t1.8 18t2 4Ozkl 0.054t0.005 0.068kO.005 1.2kO.2

0.08

E

I

9 nmol

l

E

(v

@co

9.4t0.9

2nd Measurement Period

Values are means t SE of 5 frogs. Paq, and Pace,, arterial PO, and Pco,; pH,, arterial pH; Hct, hematocrit; Gsc,, Gso , and Gsc, , cutaneous conductance to CO, O,, and C02, respective I y, in samp Pe chamber. There are no significant differences in any of the parameters between the 1st and 2nd measurement periods (P > 0.10).

0.10

7i,

? Ton

l

';

‘; c .-

pace, PK

Gso,, nmol min-’ Torr-’ . crnm2 Torr-’ crnm2 GSCOz 9 nmol . min-’

0.12

=: 0 0

Paoz, Torr Hct, % Arterial pressure, Torr Heart rate, min-’

E

'E, .

1st Measurement Period

Parameter

z

n w

GAS CONDUCTANCE

1.8

2 Period FIG. 7. Cutaneous gas conductance (Gsco, Gso2, and GsCo2) determined in time control experiments. Lines connect values from the same animal. There were no significant differences in gas conductances between the 2 measurement periods (P > 0.42). 1

face area available for diffusion of gases between blood and tissue is determined, at least in part, by the density of red blood cells in the capillaries. The results support the modeling predictions of Federspiel and Pope1 (6), Federspiel and Sarelius (7), and Homer et al. (16). Thus the large and rapid changes in capillary hematocrit observed in other tissues may function to control Dti. Modeling of mammalian tissues (6,7,16) predicts that there is not a one-to-one relationship between Dti and d RBC.This is due to a number of microcirculatory factors that probably affect this relationship: the thickness of the plasma layer between red blood cells and the capillary wall, gas diffusion through the plasma gaps between red blood cells, convection in the plasma gaps, and the precise geometry of the capillary bed. Presumably, in this studv, these factors affected the relationship between Dti

and dRBC. However, because of a lack of information on these parameters in frog skin, further studies are needed to quantitate their roles in the gas conductance of tissue. The experimental preparation is well suited for studying the relationships between the microcirculation and gas exchange. The geometry of the cutaneous capillary bed and the overlying epidermis is simple, the capillary circulation is easily observed with video microscopy, and gas exchange can be accurately measured. The use of halothane anesthesia provided a stable experimental preparation. In the time controls, no significant changes in cardiovascular function, blood gases, or gas exchange occurred during the 3 h between measurements. The blood gas values and cardiovascular parameters measured in this study are consistent with previously reported values. In the time controls, pH,, Pac02, arterial blood pressure, and heart rate are all similar to those of conscious frogs (13,19). Although Pa,, was very low, it is similar to that of conscious diving frogs (5,30). This low Pa,, appeared adequate to meet the 0, requirements of the animal, because there were no changes in pH,, Pa,,, , arterial blood pressure, and heart rate during the time control experiments. Hemodilution produced decreases in arterial pressure and pH,. A decrease in arterial pressure with hemodilution also occurs in other animals (29) TABLE 2. Effect of hemodilution on blood gases, pH, Hct, arterial pressure, heart rate, VRBc, and perfused capillary density Parameter

Pao,, Torr pace, PH*

9 Torr

Hct, % Arterial pressure, Torr Heart rate, min-’ V RBC,mm/s Capillary density, d RBc, cells/mm

mm/mm2

No Hemodilution

1st Hemodilution

10.5k1.7 13.7kl.O 7.77kO.04

10.6-tl.2 13.7-to.7 7.63-tO.O5* 11.8t1.4*

22.5k1.3 21kl 41+2 1.03kO.08 19.6k0.8 63Ok56

18t2* 43kl 0.99kO.06 19.2kO.8 349*50*

2nd Hemodilution 1 l.OkO.7 15.5k1.4 7.58k0.08.t 5.3+0.8t$

15&2Q 43k2 0.77+0.07t$ 18.OkO.3 150k31Q

Values are means k SE of 6 frogs. VRBC,red cell velocity; dRBC,red cell density. * Significant difference between no hemodilution and 1st hemodilution (P < 0.05). t Significant difference between no hemodilution and 2nd hemodilution (P < 0.05). $ Significant difference between 1st hemodilution and 2nd hemodilution (P < 0.05).

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

CAPILLARY

HEMATOCRIT

AND TISSUE

0.12 hl k

0.10

.

‘; L + .

0.08

7t .F

0.06

-

0.02

8

-

In

”

0.00 0.16

n N

0 r

0.12

t -

0.10

-

0.08

-

0.06

-

0.04

-

0 N 0.02 “m 0.00

t

.

k 0 7 c .f . ‘i3 $ W

0 N ‘5

-i

1 600

I 800

I 1000

I 200

I 400

I 600

I 800

I 1000

0.14

0

-i

I 400

I

‘E ‘;

I 200

1.8 1.6 .

k0 * . c .-E . 5 E -c 0” 0 urn

1.4 1.2

231

GAS CONDUCTANCE

The method for determining the total conductance of the skin to the three gases also appears valid. This method is similar to the single-breath method for determination of lung diffusion capacities. In the lung, this method is adversely affected by a number of heterogeneities not considered by the gas exchange model [i.e., diffusing capacity-to-alveolar volume and alveolar ventilation-to-alveolar volume heterogeneities (14)]. Similar heterogeneities are not expected to be present in the cutaneous sample chamber used in this study. However, possible heterogeneities in capillary perfusion to Dti may exist that could affect the calculation of sample chamber GO, and Gco,. The high diffusion limitation of frog skin, however, should minimize the effects of this possible source of error. The excellent fit of the model to the data supports this hypothesis (Fig. 6). To calculate Dtico and Dti,, with Eq. 4, the following assumptions were made: I) CO and 0, uptake are diffusion limited, 2) correct values of flRBC(OP)and &&CO) are used, and 3) gas exchange with plasma is negligible. These assumptions are discussed below. CO uptake across respiratory organs is considered to be essentially diffusion limited because of the extremely high affinity of hemoglobin for CO (27). In mammalian lungs, this has been verified experimentally (26). We assume that CO behaves similarly in frog skin. In contrast to mammalian lungs, 0, uptake across amphibian skin appears to be at least 90% diffusion limited (24,25). Diffusion limitation for 0, in the skin is probably the combined result of a large epidermal thickness and high

1.0

0.12 04

0.8

‘E 0 .

0.6

L

0.4 0.2 0.0

0.10

I

0

200

I

I

400 Cells

600 l

I

800

0.08

h -v.

I

1000

mm-2

8. Cutaneous gas conductance as a function of capillary red cell density (dRBC)in hemodilution experiments. Lines connect values from the same animal. Reducing dRBc decreased cutaneous conductances to CO (P < 0.003), O2 (P < 0.012), and CO, (P < 0.025). FIG.

and may be due primarily to the reduction in blood viscosity. The reason for the fall in pH, is not known, but it may have been caused by a rise in anaerobic metabolism. In the hemodilution experiments, arterial hematocrit fell to very low values (5.3 t 0.8%) after the second hemodilution. Although this hematocrit would be fatal to mammalian species, amphibians are extremely tolerant of low or zero hematocrits (8), and all animals recovered from the experiments. Fluorescent video microscopy of ethidium-labeled cells appears to be a valid technique for determining dRBc and red cell velocity. This technique requires that labeled and unlabeled cells have identical rheological properties. Comparisons of labeled and unlabeled cells indicate that this requirement was met: labeling had no significant effect on capillary cell flux and velocity, and the viscosity of labeled and unlabeled cell suspensions was similar.

200

400

600

800

1000

0.16

n FJ

0* c;, ‘E

0.14 0.12

-

0.10

-

‘; c .E -6 r-

0.08

-

0.06

-

E

0.04

t

L 3 .

-cu .F 0

0.02 0.00

1 1 0

I

200

I

I

400 Cells.

600 mm

I

800

I

1000

-2

FIG. 9. Calculated tissue diffusive conductance (Dti,, and Dtico) as a function of dRBCin hemodilution experiments. Lines connect values from the same animal. Calculated Dtico (P = 0.034) and Dtio, (P = 0.032) were decreased by reductions in dRBC.

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

232

CAPILLARY

HEMATOCRIT

AND TISSUE

blood flow through the cutaneous capillaries. Epidermal thickness in Ranapipiens is -40 pm, which is -80 times greater than the alveolar-capillary membrane thickness of mammalian lungs. Although cutaneous capillary blood flow in our experimental preparation is not known, mean red cell velocities measured in this study were -2-10 times higher than red cell velocities in mammalian capillaries, including those in exercising skeletal muscle (17, 29). This suggests that cutaneous capillary blood flow is also high. Our estimates of 0&OQ) and &&CO) are difficult to assess. They were based on in vitro data on turtle red blood cells, on relationships between &&OS) and red cell size, and on relationships between 0(CO) and 0(0,) of mammalian blood. We also assumed that &&OS) and 0,c(CO) remained constant as the red blood cells moved through the cutaneous capillary bed. The turtle red cell data were collected using a rapid stopped-flow measurement technique. Although this technique yields valuable information, it probably underestimates 0 because of the presence of unstirred layers surrounding cells during the measurement (2, 12). The presence of unstirred layers probably causes 0(0,) to be underestimated by at least a factor of 2 (2, 12). In a recent study using thin films of human blood without significant plasma layers, Heidelberger and Reeves (12) found that 0,,,(0,) was highly dependent on hemoglobin saturation, hemoglobin-O, affinity, and PO, gradients. They concluded that a single value of &&OS) d oes not exist. However, almost all their 0i&OJ va 1ues under a variety of experimental conditions are above 0RBc(OZ) values determined from stopped-flow experiments. Consequently, if a single &&OQ) did not exist under our experimental conditions, we believe that our use of a single 0,,,(0,) value probably represents a lower limit for 0r&02). The relationship that we used between red cell size and &&OS) (33) is similar to that described by Holland and Forster (15). We assume that no significant errors occurred in this procedure. We estimated 0RBc(CO) by assuming that flRBc(Op) is approximately twice B,,c (CO) (9). However, this difference may be smaller (2), which would lead to an underestimation of 8&CO). The decrease in pH, with hemodilution is not expected to affect significantly 8,,J 0,) and &&CO). The pH change was small, 0.2 units, and pH has no effect on turtle O&Op) determined with the stopped-flow technique (32). Although the true values of 8,,(0,) and &&CO) are not known, we believe our estimates of these parameters may be too low. If &&OP) and &&CO) have been underestimated, then the relationship between Dti,, and Dti,, and dRBc will be stronger than that shown in Fig. 9. For example, if ORBc(OB)and eRBC(CO) are infinite, then Dt& and Dtico will equal GO, and GCO, respectively (Fig. 8). It is difficult to assess whether plasma gas exchange is negligible. However, modeling studies of mammalian capillaries (6,7,16) indicate that little 0, exchange between blood and tissue occurs within the plasma gaps between red blood cells. We are unaware of evidence indicating a significant contribution of plasma to 0, or CO transport. If all our assumptions are reasonable, then the ratio DtioJDti,, should equal the ratio of Krogh’s diffusion constants (K) for 0, and CO. In frog connective tissue,

GAS CONDUCTANCE

Ko&o = 1.33 (18). Dtio,/Dtico before hemodilution, after the first hemodilution, and after the second hemodilution was 1.23 t 0.08, 1.22 t 0.08, and 1.26 t 0.07, respectively. There were no significant differences between these values and 1.33 (P > 0.24). Consequently, we believe that our estimates of Dtio, and Dtico are reasonably accurate. The results also show that dRBCalso affects Gco,. We were unable, however, to determine Dtic,, because 1) unlike with 0, and CO, cutaneous CO, transport is probably significantly perfusion limited (21) and 2) there is not enough information available to estimate the CO, kinetics in frog blood. However, the results are consistent with dRBC having an effect on Dtic,,. Cutaneous GO, and GCO, have been determined by others in the bullfrog. Pinder (25) determined cutaneous GO, in paralyzed submerged bullfrogs at 5°C at different aquatic Po,‘s. At an aquatic PO, of 140 Torr, GO, was 0.036 nmol min-l Torr-’ g-l. By use of the relationship between cutaneous surface area and body weight for anurans (23), this value corresponds to 0.042 nmol. min-l Torr-1. cmS2. Lowering aquatic PO, to 30 Torr caused GO, to rise to 0.072 nmol min-l Torr-l cmm2. A concomitant rise in arterial hematocrit from 23.8 to 29.9% also occurred with aquatic hypoxia. Although the microcirculation was not observed in this study, it is likely that cutaneous dRBC was elevated under hypoxic conditions and may have contributed, in part, to the increase in GO,. These values of cutaneous GO, in the bullfrog are lower than GO, determined in this study on the leopard frog. This difference is probably due to 1) the thicker epidermis of the bullfrog (3), 2) the lower capillary density of the bullfrog (3), and 3) the lower temperature at which the bullfrog Go2 was determined. Cutaneous GCO, of the bullfrog at 20°C [0.77 nmol. mine1 . Tori? crnD2 (20)] is also lower than that measured in this study. Morphological differences also probably contribute to this difference. Our finding that dRBC affects Dtio, indicates that changes in capillary hematocrit will influence the diffusive conductance of tissue to 0,. Because capillary hematocrit can increase under hyperemic conditions, the diffusive conductance of tissue will increase with increasing perfusive conductance. This will help maintain the diffusive-to-perfusive ratio of tissue constant, which may help prevent, or at least ameliorate, tissue diffusion limitation to 0, during high tissue blood flows. l

l

l

l

l

l

l

l

This research was supported by National Heart, Lung, and Blood Institute Grants HL-38942 and HL-40537. Address for reprint requests: G. M. Malvin, Lovelace Medical Foundation, 2425 Ridgecrest SE, Albuquerque, NM 87108. Received 10 June 1991; accepted in final form 10 February 1992. REFERENCES BIJRROWS, B., AND A. H. NIDEN. Effects of anemia and hemorrhagic shock on pulmonary diffusion in the dog lung. J. Appl. Physiol. 18: 123-128, 1963. COIN, J. T., AND J. S. OLSON. The rate of oxygen uptake by human red blood cells. J. Biol. Chem. 254: 1178-1190, 1979. CZOPEK, J. Quantitative studies on the morphology of respiratory surfaces in amphibians. Acta Amt. 62: 296-323, 1965. DESJARDIN, C., AND B. R. DULING. Microvessel hematocrit: mea-

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.

CAPILLARY surement and implications

for capillary

HEMATOCRIT

oxygen transport.

AND TISSUE Am.

Physiol. 252 (Heart Circ. Physiol. 21): H494-H503, 1987. 5. EMILIO, M. G. Gas exchange and blood gas concentrations frog, Rana ridibunda. J. Exp. Biol. 60: 901-908, 1974. 6. FEDERSPIEL, W. J., AND A. S. POPEL. A theoretical analysis

diving bradycardia

J.

in the

GAS CONDUCTANCE

Wash. DC 159: lOl-103,196s. 9. FORSTER, R. E. Rate of gas uptake Physiology. Respiration. Washington,

Biochem. 24. PIIPER,

The Respiratory

System.

Gas Exchange.

J. Appl. Physiol. 12. HEIDELBERGER,

66: 1942-1955, E., AND R.

1989. REEVES.

B. blood O2 transfer kinetics: implications

Factors affecting whole for 0(0,). J. Appl. Physiol.

68: 1865-1874,199O. 13. HERMAN, C. E.,

D. 0. ROBLETO, P. L. MATA, AND M. D. LUJAN. Attenuated cardiovascular effects of prostaglandin I2 and prostaglandin F, in cold acclimated american bullfrogs, Rana catesbiana.

J. Exp. Zool. 14. HLASTALA, of Physiology.

238: 167-174,

1986.

In: Handbook Bethesda, MD: Am. Physiol. Sot., 1987, vol. IV, chapt. 12, p. 217-232. 15. HOLLAND, R. A. B., AND R. E. FORSTER. The effect of size of red cells on the kinetics of their oxygen uptake. J. Gen. Physiol. 49: 727-742, 16. HOMER,

M. P. Diffusing-capacity The

Respiratory

heterogeneity.

System.

Gas Exchange.

1966.

L. D., P. K. WEATHERSBY, AND L. A. KIESOW. gradients between red blood cells in the microcirculation.

vast. Res. 22: 308-323, 17. KLITZMAN, B., AND

Oxygen Micro-

1981. R. DULING.

B. Microvascular hematocrit and red cell flow in resting and contracting striated muscle. Am. J.

Physiol. 18. KROGH,

237 (Heart

Circ.

Physiol.

6): H481-H490,

1979.

A. The rate of diffusion of gases through animal tissues, with some remarks on the coefficient of invasion. J. Physiol. Lond.

52: 391-408, 19. LILLO. R. I

1919.

S. The effect of arterial-blood

PO,. PCO, and DH on

catesbiana.

Physiol.

2001.

1978.

M. Microvascular regulation of cutaneous gas exchange in amphibians. Am. Zool. 28: 999-1007, 1988. 22. MALVIN, G. M., AND M. P. HLASTALA. Effects of environmental O2 on blood flow and diffusing capacity in amphibian skin. Respir. Physiol. 76: 229-242, 23. MCCLANAHAN, L.

1989.

L., AND R. BALDWIN. Rate of water uptake through the integument of the desert toad, Bufo punctatus. Comp. Physiol.

28: 381-389,

1969.

J., R. GATZ, AND E. CRAWFORD, JR. Gas transport characteristics in an exclusively skin breathing salamander, Desmognathus fuscus (Plethodontidae). In: Respiration of Amphibious Vertebrates, edited by G. M. Hughes. London: Academic, 1976, p. 339-

vol. I, chapt. 32, p. 827-837. of gases across the alveolar membrane.

Bethesda, MD: Am. Physiol. Sot., 1987, vol. IV, chapt. 5, p. 71-88. 11. GRBNLUND, J., G. M. MALVIN, AND M. P. HLASTALA. Estimation of blood flow distribution in skeletal muscle from inert gas washout.

Rana

J., AND D. JACKSON. The effect of temperature on cutaneous COZ loss and conductance in the bullfrog. Respir. Phys-

iol. 32: 313-323, 21. MALVIN, G.

10. FORSTER, R. E. Diffusion In: Handbook of Physiology.

in the bullfrog

51: 340-346,197s. 20. MACKENZIE,

of the effect of the particulate nature of blood on oxygen release in capillaries. Microuasc. Res. 32: 164-189, 1986. 7. FEDERSPIEL, W. J., AND I. H. SARELIUS. An examination of the contribution of red cell spacing to the uniformity of oxygen flux at the capillary wall. Microvasc. Res. 27: 273-285, 1984. 8. FLORES, G., AND E. FRIEDEN. Induction and survival of hemoglobin-less and erythrocyte-less tadpoles and young bullfrogs. Science by red cells. In: Handbook of DC: Am. Physiol. Sot., 1964,

233

356. 25. PINDER,

A. W. Cutaneous diffusing capacity increases during hypoxia in cold submerged bullfrogs (Rana catesbiana). Respir. Physiol.

70: 85-95,1987. 26. ROSENBERG,

E., AND R. E. FORSTER. Changes in diffusing capacity of isolated cat lungs with blood pressure and flow. J. Appl. Physiol.

15: 883-892,1969. 27. ROUGHTON, F.

J. W., AND R. E. FORSTER. Relative importance of diffusion and chemical reaction rates in determining rate of exchange of gases in the human lung, with special reference to true diffusing capacity of pulmonary membrane and volume of blood in the lung capillaries. J. Appt. Physiol. 11: 290-302, 1957. 28. SAINT-AUBAIN, M. L. DE. The morphology of amphibian skin vascularization before and after metamorphosis. Zoomorphology 100: 55-63,1982. SARELIUS,

I. H. Microcirculation in striated muscle after acute reduction in systemic hematocrit. Respir. Physiol. 78: 7-17, 1989. 30. SHELTON, G., AND BOIJTILIER, R. G. Apnoea in amphibians and reptiles. J. Exp. Biol. 100: 245-273, 1982. 31. VANDEGRIFF, K. D., AND J. S. OLSON. Morphological andphysiological factors affecting oxygen uptake and release by red blood cells. 29.

J. Biol. Chem. 32. YAMAGUCHI,

259: 12619-12627,1984.

K., M. L. GLASS, P. SCHEID, AND J. PIIPER. transfer kinetics of red blood cells of the turtle Pseudemys

elegans. Respir. 33. YAMAGUCHI,

Physiol.

Oxygen scripta

75: 371-384.

K.,K.D. JURGENS, H. BARTELS,AND J. PIIPER.~XYgen transfer properties and dimensions of red blood cells in high altitude camelids, dromedary camel and goat. J. Comp. Physiol. B

Biochem.

Syst.

Environ.

Physiol.

157: l-9,

1987.

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (129.186.138.035) on January 19, 2019.