Respiration Physiology (1981). 44. 61-86Eiscvicr/Norih-Hoiiand Biomedical Presi

DESIGN OF THE MAMMALIAN RESPIRATORY SYSTEM.V. SCALING MORPHOMETRIC PULMONARY DIFFUSING CAPACITY

TO BODY MASS: WILD AND DOMESTIC MAMMALS*

PETER GEHR1, DETER K. MWANGI2, ALEX AMMANN,GEOFFREY M.O. MAI.OIY1, C. RICHARD TAYLOR4

and EWALD R. WEIBEL

Deportment of Anatomy. University of Heme. Berne. Switzerland

Abstract.Thii papa utilizes a comparative approach 10 establish the rcldiffusing capacity for oxygen (Dio,) and maxii

i on the lame 21 individuals in African mammals spanning a range in body mass from_ Ve confirmed earlier findings that Dto, was proportional lo Mb°" while V0llluI was

proportional lo Mj*. Thus, the ratio of DlOi/V0i„„, is approximately proportional to M(,"M. Weconclude that large animals require a larger pulmonary diffusing capacity to transfer oxygen at thesame rate from air to blood.

Lung morphologyOxygen consumptionPulmonary diffusing capacity

The pulmonary diffusing capacity Tor 02, Di.0, is the conductance for O, flowfrom air to blood in the lung, driven by the partial pressure difference betweenalveolar air and capillary blood (fig. I). Il is determined, in part, by some

;il Science Foundation (3.394.74 and

rd School of Public Health, 665 I luntington Avenue.Boston, MA 02115. U.S.A.'Present address: ILRAD, P.O. Box 30709, Nairobi. Kenya." dicine. University of Nairobi, P.O. Box 29053, Kabelc,Kenya.' Present address: Museum of Comparative Zoology. Harvard University, Cambridge, MA 02138. U.S.A.

0034-5687/81,00O0-O0O0/S02.50 © Elscvier/North-Holland Biomedical Press

DESIGN OF THE MAMMALIAN RESPIRATORY SYSTEM

0, FLOW _ PRESSLNE , nMUClMICSRA1E _ GRADIENT

;;fe (A) = (Pl-P^Oi^fl?,

\L (Si* <Fa-P»W.;VB-/JB •

S C A L I N G O F D l 0 i T O B O D Y M A S S 6 3

we believe that marine mammals should be considered separately for reasons givenin the discussion.

Consequently, the two approaches proposed to test whether Dl0, was matched10 vo, 6>ve paradoxical results: whereas Dl0i is closely proportional to V„, withina size class, V0l and Dl,0j scale differently to body mass over a large size range.

It seemed possible that this paradox resulted from the fact that in the comparativestudies the measurement of V0j and Dl„, have not been done on the same animals,or not even on a homogeneous and comparable population of animals. In orderto remove this the present study combines Dt^,. studied by morphometry in thelungs of the same specimens of African bovids and vivcrrids on which Taylor cl al.(1981) had obtained measurements of V0imlI. The inclusion of wild and domesticatedbovids in that study also allowed us lo compare Dl.0) with V„JlnjI for animals ofsimilar size but different metabolic needs.

,ided into three compartments with 02 flow_ra.es "function

r r° T"p„ i i I o. b.« to «m»' —" " ",e r"!'

size ran? of nammab (2 s <*fc,"<T.,8'u0v°';'simitar «b,i„„ is ro„„<l f«-

mrfta on >h= m»n,m„l,.„ lung m found 1tot t.,, », 1)r„p„„io„al

t:l«^U—.s „,.« J— -U-* I*. SW »nd bod, m

MATERIALS AND METHODS

This study was performed on 27 African mammals (15 species) ranging in bodymass (M,,) from 0.42 lo 251 kg as specified in table 1. The animals were recruitedfrom vivcrrids (dwarf mongoose, banded mongoose, genet cat), wild bovids (stint,dik-dik, Grant's gazelle, Thomson's gazelle, wildebeest or gnu, waterbuck, eland,giraffe) and domestic ruminants (Masai goat, Masai sheep, zebu cattle, camel).Combined estimates of maximal oxygen consumption, V0imai, and morphomctricdiffusing capacity for oxygen, DL,,,, were obtained on 22 of these animals. V0m„measurements were obtained while the animals were running on a treadmill by themethod described by Sccherman el al. (1981). The detailed data of this part of thestudy arc presented in the companion paper by Taylor et al. (1981).

After completion of the physiological studies the animals were sacrificed andtheir lungs fixed by tracheal instillation of a 2.5% potassium-phosphate bufferedglutaraldehydc solution at a head pressure of 25 cm above the chest with the animalsin supine position. The lungs were removed from the chest in tow, their volumeestimated by a water displacement method, following which we proceeded lo tissuesampling and morphometric analysis as described in detail by Weibcl el al. (1981a).The sampling procedure adopted for this study deviated somehow from that usedon the older material introduced here for comparison, but it has been shown thatthe new method only affects the errors and not the estimates, so that the results ofthe new and of the older studies remain comparable (Weibcl cl at., 1981a).

The following physical coefficients were used together with the estimates ofmorphometric parameters to calculate Dl„, after the model of Wcibel (1970/71):

K, = K„ =4.1 .i0",0cm2-sec-' -mbar"1,

0Ol = 1.87 I0-' mlO, ml-' sec-' mbar-'

<a>

co-

I p £ ■= 5|0 =

-I'

a is .ap-g sO

T3

- re

—

I,."") XJISN30

KIVibfB BVX3ATY

«f ■« 8 ■= -o S, g - .=

3 -

~ n

c a

_g

B >

o =

•>

o s g ■- = = §sl

ls

i

-816E0i.f9

811010906850

S'LLZVSLt

("(« rioeVIU688?

O'OOSfl000£ilV6U

101cMR frnjopjmojp mpuio0)

jaono

8"! 11filleu0 89

esfoi8Z9'88W990£-9

IOZ'01610IftOtcro

USO99500£K091 t'O

8'9869'£Z8S'l8Vrtos

rit*l'S9£9TS££80E

9£0S8t-9EI'KCfEIC

0S£HIOi86ri018090'9£08

Oflt01 StOKIO'ISI

BUZ6113zitz(IKt output tog)

sjlira nqsz

▶-8I8111KL'I

oiro£££0

0850Itr'd

99011CS081

£5056£'S£

56 65trw

079910'6fr£l

OTZ9'K

80IZ60Kfsqjs tyi0)

daaqi uraujv

r6i6'9I

886_0£660

66£"0ato

two2150

80601era

ro'SfLira

0F9tLilt

0SS£l0f8£l

6'K661

OIKUKfmw;V t"<l"3)

leoS aaxijvon - JDp»"ff fyutoj

££H 0V01

9101 Z9989"0i Si0'9

£06 6150

9650 SW6

9610 OZf'O £T89CtrO c6f0 098?

- 6lc0 - tiro

££50 OO'tf

9ISS 19E9 O'OOOIZ

TE£8 809t- 0£SS 0S9901

0611 9£'S£

I'Uf 0T8C6£•£81 0'88Z9

8'06£ 08£9£

0056 0'0£9I

(si\vpiodo\3wm offniir))

puqg

(vmpp tnqo)l)ijonqjaiBAA |

Isnuijnm swoumuuoj)UMqapilM j

(iuonuot/1 ttflltBQ)3\\31<S s.uosnjouj,

ftlimiS oji»xdo).-i||J*cS S.1UE1Q :

(IPHI1 mbPBfV)wva

(miupttm sn2"Jot*N)mns

tptdt PH* - "Vt.og ■<1!"">J(aiySii vhm'O)

1K> 13U'D(oStmUl soSunjl)MOo8aoai papucg

(DjruJtd 3]v2o\>n)ssooSuoui Jicwrj

zQ ,uij) V (;0 !"■) I

^v^^^j^sX'St. '''!'UOJj d|da:« psynciu 91-8 »»» 4UU

. nana, ■(iki) ;o U JOJMJ. *M P*""1

SCALING OP Dl0, TO BODY MASS

for (C). 95% confiden

• for restricted population of African, ill 2.7 animals, (B) and (C) only thosea • Mk* (Mb in kg) for (A) and (II) and y •

ell as correlation coefficient r, are reported

»L m l 66.13 58.27 . 7504 0 9 8 6 0.951. 1.020 i' tM«*) m> 3.577 soar. 4.255 0.918 0171 . 0.966 0.99]5(0 m ' ■ 2.988 2.488 . J.S8S 0.924 0 8 7 4 . 0.974 0.991V(c) m l 4 190 3.231. 5.432 0.971 0 9 0 0 . 1042 0.9S5«K0 urn 0 44J 0.400. 0.494 0028 -0001 . 0.057 0.364D l O , ml 0,/sec • mbar 0.055 0.044. 0 0 6 9 0.950 0 8 8 9 . 1.012 0.911V(c)/(SA) m l / m i 1 171 0936 . 1.465 0.053 - 0 . 0 0 9 . 0.114 0.333IB)SlM-Mb m ' / k g 3.474 2MS0. 4.189 -0 .066 - 0 . 1 2 0 . - 0 0 1 3 - 0 5 1 0Dl-Oi/Mb ml Qjltec -kR-mbar 0.051 0.043. 0.063 -0 .003 - 0 0 5 7 . 0.052 -0.022* O i « « / M b n i l O , / s « k g 1.71 1.43 , 1.976 -0 .221 - 0 . 2 6 2 . -0 .179 - 0 9 3 2

toS(A) 0)1 1.879 1.455. 2.428 1.188 1.107. 1.268 0.990D<-0l ml 0,/sec mbar 0.027 0.020. 0.037 1.265 1.168 . 1.363 0.987

R H i ^ ^ " W ^ |

. ^ : m, nowcr electron micrograph or wildebeest lung showing dense packing of capillaries <Q

r ; o t ^with fibroblast processes. Magmficalion: 9000 x.

B00Y MASS (kg)Pig. 6. Allometric plot of lung volume. African mammals. Symbols sec fig. 4.

D E S I G N O P T H E M A M M A L I A N R E S P I R A T O R Y S Y S T E MS C A L I N G O P D t n , T O B O D Y M A S S

BOOY MASS (kg). ....r,,.-, .-..ca. African mammals. Symbols sec fig. 4.

tlCHSOM G4ZEUE C*

'9LAj S S I C t P t ! ) '

( flAn rVoMcuE

• ' S U N K ? ) ' V 0 j > n a >( / s l c « « : 0 7 SCAHJI A , 'p • qanoed Hoeoosr.

BOOY MASS (kg)

Pig. 9. Al lometric plot of pulmonary diffusing capacity. Afr ican mammals. Symbols see fig. 4.

SHEEJM2) * /IHOMSON GAZEUC Jr^ , , , ,

y gemEI CAT<2)/ • BANDED MONGOOSE

/* DWARf MONGOOSCU)

a S o o i o o i o i « * ™ 1 0 0 °BODY MASS (kg)

Fig. 8. Allometric plot of capillary volume. African mammals. Symbols see fig. 4.

was that found for the harmonic mean thickness of the tissue barrier (r,„) whosemean was of the order of 0.5 pm with a range from 0.37 lo 0.65 urn (table 2).

Prom these data the morphometric pulmonary diffusing capacity for oxygen,Dl.0l, was calculated and found to vary between 0.017 and 11.1 mlO.-s"1 -mbar"';it scaled linearly with M„ (b = 0.95), as shown in fig. 9. It is noteworthy that theslopes for diffusing capacity, alveolar and capillary surface areas, and capillaryvolume were close to I, and that they were all significantly different from the slopeof 0.78 obtained on the same animals by Taylor et al. (1981) for V0ima, (table 2b).Figure 10 finally plots diffusing capacity against maximal oxygen consumptionfor the 21 animals on which both measurements had been obtained; Dl0, increaseswith the power 1.27 of maximal oxygen consumption (tabic 2c). Thus, as shownin fig. II, the flow of oxygen across the unit alveolar surface area or the unitdiffusing capacity is smaller in large than in small animals.

Finally, comparing the wild and domestic bovids comprised in this study forwhich both V0)-M and Dl„, estimates were available (fig. 12) it is clear that thelarge domestic animals had consistently a lower mass-specific V0 „,, and also alower mass-specific DLq,. It was surprising that goats and sheep did not show areduced V0lin„/Mb and Dl̂ ,,/Mb compared to gazelles.

DESIGN OP THE MAMMALIAN RESPIRATORY SYSTEM

2EBUU)

WAltRSJCK /■

SCALING OF Di0, TO BODY MASS

/ • BANDED MO«00SE

,* DWARF MONGOOSE (2): VfhjH, and specific Dl

3. COMPARATIVE MORPHOMETRY OF MAMMALIAN LUNGS

pu.monary diffusing capacity against %,„, for African mammal, on which V„..j :-j:..:a....ii.. rrav|or el al.. 1981).

1' 10

*| 02i? 01

- sloe*: - 015L

00DY MASS (kg)

Rg. II. Plot of maximal O, flow per unit alveolar su.facass for African mammals on which V0jnul was m.

unit diffusing cap"vidually (Taylor ■■

An important question to answer is whether the group of Kenyan mammals considered in this study is representative of a broader population of mammalian lungs.For this purpose we have compared the present data with those obtained on othermammalian species using the same basic method. Rather than presenting individualanimal data table 3 and figs. 13-16 report average values for species estimates.The species considered range in body weight from 2 g in the Etruscan shrew to500 kg in tlie horse and 700 kg in a Swiss cow. Of the 32 species 15 representthose investigated in the present study. It is clearly seen from figs. 13-16 thatthe data obtained on the African ruminants and vivcrrids fit well into the generalpopulation, and that the allometric regressions Tor all parameters are identical forthe Kenyan mammals and the overall population (tables 2 and 4). In additionthe broader population reveals that the slight increase in thickness of the air-bloodbarrier with body mass is significant; it should be noted that the allometricregression coefficient for t,,, found in the Kenyan animals is the same as that forthe overall population. Also the capillary loading of the alveolar surface area,estimated by the ratio V(c)/S(A), is significantly increasing with body mass.

LUNG FINE STRUCTURE AND BODY SIZE

Allhough the animals investigated span a body size range of almost three ordersnf Magnitude, there were no striking size-dependent differences in the fine structure

e.crs of pulmonary gas exchange apparatus for mammalian specie,ed are marked by asterisk (cf. Taylor el a,.. 1981). Because of the-lively, on the basis of M*. The two groups for the rats were.

. for *c dogs were: mean 5.4 kg. range 2.6-8.2; men U.6 kg.in 46.1. range 36.0-57.0)

(Ciocubaa fffarii)

9.11 0.0240.61 0.1040.10 0.017

(0.001810.22 0.0310.39 0.0620.49 0.067

1.13 0.1562.7 0.25

t.45 0.12510.006)

0.571 0.063

(O.oos)

6.34 0.388

(0.054)

0.01700770.0130.004)0.023

0.0110.0600.0116

(0.OO4)0.018

0.260.290.27(0.04)0.37

0.0420.050

0.0340.043

0.36"0.42

0.123 0.135 0.380.202 0.245 0.35

0.113f"0.<WU

0.147(0.007)

0.29(0.012)

0.054(ami

0.065(0.017)

0.26(0.005)

0.407(0056)0.816

(0.019)

o.4So(0.063)1.48

(0.0431

0.37(0.039)

0.40(0.021)

Family LrporiJaeR*bbit

0.91 0.74 1.460.11) (0.09) (0.32)

5.86 4.70 7.151114) (OutS) (1*8)

D o g(Canisfamilujru) 18.2 14.1 26.0 0 43

(13.5) (U . l ) 124.9) (0.02)43.2 35.5 55.8 0.47tt-9) (8.8) (21.1) (003)82.7 65.5 110.0 0.50(20J) 115.2) (33.0) I0O1)76.9 131.9 233.7 0.53(45.6) (37J) (68.7) (0.03)

illy GlraffulceGiraffe

: l o p a r d a l i s ) I

(2) awm

OA

9Nm(,uj)

aawans

avioBATv

(in)

|1W

O

'Mf

0-9

SK

W"

DESIGN OP THE MAMMALIAN RESPIRATORY SYSTEMSCALING OF Din, TO BODY MASS

rMOMSOti GAZIUE7/000C O A I /D O G O /

GRAMS GAZEUE <

OWARf MONGOOSE •/• BANDED MONGOOSEGUrCA PO CO RAI

BOOY MASS (kg)Fig. 15. Allometric plot of capillary volume for all species. Sec fig. 13.

*is' y i'&lu

Allometric regression for morphoincti ic rcspualoiy variables l<(table 3). Reported as in

It extended |table 2

]H»pul;itimi based on species data

P a r a m e t e r U n i t s Coefficicnl a

lean (95% eort /UaKtinltrval)

Eipone 1 1 b (

(95% confUcnce

1.059 1.031 . 1087 0.9973.342 2.97/ , 3.759 0.949 0 917.0 980 0.9967.727 2.456. 3.027 0.952 0 924.0.980 0.9963.198 2 JOS. 4437 1.000 0.912.1.037 0.9700.416 0J96. 0.437 0.050 0.037.0064 0.7960.049 0055

ofthe lung parenchyma. The main difference lies perhaps in the number of alveolarpores which arc sparse in bovids but very dense and conspicuous in the viverridlungs; the latter resemble shrew or dog lungs with their notoriously large numberof pores, occupying nearly every capillary loop (Gchr et at., 1980a,b). Therewere also considerable differences in the dimensions of terminal airways and ofalveoli, but these did not show any obvious rclalion with body size.

This was confirmed by the morphometric analysis which revealed a great variationin the volume density of tissue and capillaries, as well as in the alveolar surfacedensity, the largest and smallest values differing by 2-3 times. But with the soleexception ofthe capillary volume density which was slightly larger in bigger animals,these morphomclric variables were not correlated with body size. It is interestingto note, however, that some of these variables arc significantly different betweendomesticated and wild animals, so that Ihcy may be related to differences in thelevel of 02 consumption.

noo a/GRANIS OAZEIlE •

ik-d* .V**-MONKI.Y °sl__

/ o riAUOil•KMir CAT

/• IWNOID MONGOOSI

»«tw »° H005£

BODY MASS (kg) |

Fig. 16. Allometric plot of pulmonary diffusing cap..

SCALING OF MORPHOMETRIC VARIABLES

The present study was undertaken to compare the scaling of maximal metabolicrate with that of the major morphometric determinants of pulmonary diffusingcapacity. The study was done on a group of mammals ranging in body mass from400 g to 250 kg. The first uuestion to answer is whether the morphometric datathus obtained are representative of the data obtained on a more extended populationranging from 2 g to 700 kg and derived from 34 different species (Weibcl, 1972, 1973,1979).

The morphometric method used in this present study was slightly modified ascompared to that used in previous studies, particularly with respect to the samplingprotocol. In a companion paper (Weibcl ci at., 1981a) we have shown that the

DESIGN OF THE MAMMALIAN RESPIRATORY SYSTEM

new and the old method yield comparable results. The two data sets can thereforebe considered jointly. ^ ^^.^ dcrJvcd for ,hc rcslrictcd and

extended populations respectively; none of the correlations ***»*"*between restricted and extended population. We can there ore «***"»range of African mammals considered in the present study ,s rcprcscn at.v ofhe extended population. On the other hand, due to the limited weight angc

coifdenc/interva, for the regression coefficients for the Ker.ynn^annnawider than those for the extended population; this had however no effect on thetat stica significance of the correlations, with the exception of the scaling o

ZSSonSc mean barrier thickness, xh„ which is significant only for the extended

Telung volume, VL, corresponding to submaximal inflation was found tosea e i X with body mass with a narrow confidence limit This IS . few*

', "videtL which also shows that lung we*ht and ^.^«g»£linearly related to M„ (Brody, 1945; Tcnncy and Remmers 1963, Stahl, 1965, 1967,Weibcl 1972; Barllctt and Areson, 1977; Gehr « at., 1980b).

h alveolar and capillary surface areas both scaled with a factor of 0.92 wdiffered significantly from 1.0. In contrast the capillary blood volume scaled w„h

t or which was not significantly different from 1.0, and as a consequencethe capillary loading or the alveolar surface, estimated by the quotient V(c)/S(A)how ". tendency .o increase slightly with body weight. The morphometry

puhnonary diffusing capacity, Dl0:, sca.ed linearly with Mb, the exponent 0.95not beim; statistically different from 1.

The present study has thus, in essence, confirmed the previously estabhshedfindmg .1 a the main morphometric parameters of the pulmonary gas exchangeapparatus, namely alveolar surface area, capillary volume, and the compoundparameter DL,„. all scale about linearly with Mb.

RELATION OF MORPHOMETRIC VARIABLES TO V„!ln,j

The main purpose ofthe present study was to see how the morphometr. - >dfile the 1 una as a gas exchanger related to maximal oxygen consumption. Asset o, it introduction two approaches are possible: (l)onc can comparenim so comparable size but different oxygen requirements; (2) one car,.comp.™

Tiling of V, With that of morphometric variables over an extended rangeof size classes. In our previous studies these two approaches had y.elded confi.cttng

- S X Z S K Z * — « " s i z e s u c h a s « ^ h o r « J O e h r a n dErni 1980) or normal and waltzing mice (Geelhaar and Weibcl. 1971, Hugonn 111977) we had found that the morphometric variables *****»£diffusing Capacity were closely related to the animals' oxygen requ.retnents. In the

SCALING OF Di.0,TO IIODY MASS

present study this finding was confirmed by comparing wild and domesticatedAfrican bovids (fig. 12): the large domesticated bovids had a tendency to loweroxygen needs and smaller pulmonary diffusing rapacity.

This study has however also confirmed the finding, that V0l„„ and the morphometric variables determining Dl.0l scale differently to body mass. Restrictingthis analysis to the 12 species (21 animals) for which both estimates of V0inJ1 andlung morphometry were available we find that V0;11I<„ scales with the powerb = 0.78 of body mass, with a 95% confidence interval for this slope of 0.72-0.83.In contrast, alveolar surface area scales with b = 0.92 (confidence interval 0.87-0.97)and DL0l with b = 0.95 (confidence interval 0.89-1.01). It is evident from this thatthe slopes for S(A) and Dl.0l arc significantly different from the slope of V0im,,.

If we relate Dl0i directly lo V0imil wc find DLq, = 0.027 • V0jnm'".It should be noted that the same general relationship also holds if wc compare

the scaling of morphometric variables in the extended population with the scalingof vo.n»« f°r an extended population as presented by Taylor et at. (1981). Thesetwo extended studies cover roughly the same size range, but were for the majorpart done on different animals.

From these results wc conclude, that bigger animals require a larger pulmonarydiffusing capacity than smaller animals in order to admit the flow of oxygenrequired by the organism during heavy work. For example, a wildebeest weighingabout 100 kg requires a diffusing capacity five times larger than a mongoose of1 kg body weight to transfer the same amount of oxygen from air to blood(see fig. 11).

COMPARISON WITH OTHER DATA

When comparing the results of the present study with those of other authorswc first observe that physiological estimates of Dlco and Dl„, have repeatedly beenfound to scale about linearly with Mh; Stahl (1967) obtained slopes of about1.1 whereas O'Ncil and Lcith (1980) recently found a linear relation between Dlcoand Mb.

Our findings are in apparent conflict with those of Tcnney and Remmers (1963)who had found the alveolar surface area lo be linearly related to V0l in a rangeof mammals spanning six orders of magnitude from the bat to the whale, whereaswe find S(A) lo be related to V0j''. We have said before (Weibcl, 1973) that thisapparent conflict cannot result from differences in the morphometric estimates:comparing those species that arc contained in both studies one finds good agreementbetween the data; the fact that Tenncy and Remmers (1963) obtained theirmeasurements on air-dried lungs by light microscopy whereas we use electronmicroscopy should have no effect on a comparative study of this kind. Thedifference seems lo lie essentially in the selection of animals. Tcnncy and Remmers(1963) included in their study terrestrial and marine mammals whereas our own

z i111 «

_ 2

; o

_s

-3 M

Oi

il i

"i i

ill-

S 5

G

IS.

8 J.

E -

o o

o =

= *

* o