The structure of capillaries and the unmyogenic character of rouget cells (pericytes) in the omentum of rabbits and in the web of living frogs.код для вставкиСкачать
THE STRUCTURE O F CAPILLARIES AND T H E UNMYOGENIC CHARACTER O F ROTTGET CELLS (PEBICYTES) I N THE ORXENTUM O F RABBITS AND I N THE WEB O F LIVING FROGS NICIIOLAS A. MICITELB Daniol Buugk Imtitute of Anatomy, Jefferson Medical College ELEVEN FIGURES The acceptance of Rouget cells as the elements responsible for capillary contraction is unwarranted for the simple reason that the muscular character and sympathetic innervation of these cells have to date not been definitely established on an anatomical basis. So-called contractile cells about capillaries were first described by Rouget in 1873 in the hyaloid membrane of the frog’s eye. It was his conviction that they constituted a typical muscular coat of which he says: “Cette tunique n’etant que la continuit6 des tuniques musculaires des art6res et des veines, il en r6sulte que tout le syst6me vasculaire sanguin, du coeur aux capillaires inclusivement est env6lopp6 dans une tunique contractile. ’’ The theory of Rouget was forgotten until 1902 when Mayer resuscitated it after observing the intestine of a mouse stained with methylene blue. Later Vimtrnp ( ’22)’ Bensley and Vimtrup (’28) gave new impetus to the theory in their assertions that in vital preparations of frog tissue (methylene blue, Janus green B) the Rouget cells show distinct myofibrils and exhibit contractions with electric stimulation. Icrogh gave the fullest expression to the concept that the Rouget cells effect capillary contraction and believed that the cells have a sympathetic nerve supply . . . . a point subsequently given 99 100 NICHOLAS A. NLICHELS a purported anatomical support by his pupil Busch (’29). Zimmermann ( ’23) labeled the pericapillary cells, pericytes. He advocated the muscular character of many of these structures on the basis of the existence of transition stages from the arteriolar muscle cells. Other investigators who have upheld the theory are Heubner ( ’23), Gurwitsch (’23), Tannenberg (’26), Plenk ( ’27), and Schaly ( ’26), the latter claiming the cells to be visible in fixed preparations. The main objections to Rouget cells as contractile elements may be summarized as follows : 1. They do not show myofibrils (Benninghoff), nor has the transition from muscle cell to Rouget cell ever satisfactorily been explained (Zweifach, ’34). 2. They are connective tissue cells, for they store vital dyes as do histiocytes and pericytes (Dominici, Aschoff and Kyono, Ferrata, Marschand, Ohno, Volterra, Woollard, Hassin, the Clarks, Stilwell). 3. They are modified sloughed-off endothelial cells (Marschand, Ohno). Favoring this view is the fact that the cytoplasm of Rouget cells can only be seen when capillaries are contracted ; their processes are faint or indistinguishable when capillaries are dilated (Krogh and Vimtrup). 4. They are not visible in fixed preparations. I n vital preparations, amording to Busch ( ’29) “it is impossible from the nucleus alone to recognize a particular cell as a Rouget cell.” Rouget stained only the nuclei, the protoplasm was observed in an unstained condition. 5. They are not sufficiently numerous to cause capillary contraction (Florey and Carleton), while during contraction of capillaries the endothelium moves away from the Rouget cells (the Clarks). 6. They are not present in higher animals, especially in man, but occur only in lower forms, particularly in the frog. Busch did not find them in man, although Vimtrup claims to have observed them. 7. They do not cause capillary contraction when directly stimulated in the living state (the Clarks, Zweifach), CAPILLARIES A N D PEBICAPILLARY CELLS 101 8. Their innervation has never been definitely proved, the evidence submitted by Busch being very meager while Bensley and Vimtrup ( '28) could not determine the same. 9. They have often been mistaken f o r pericapillary nerves (nucleated segments of Remak fibers) ; the extreme length of some of the Rouget cells (217 p, Zimmermann; 200 p, Vimtrup) favors this view (Sfichels, '33 ; Boeke, '33 ; Jones, '36). 10. They are not needed for capillary contraction since capillary cndothelium itself has the inherent power to contract f Thomson, 1835) ; Stricker, 1865 ; Vulpian, 1875 ; Severini, 1878; Golubew, 1869; Roy and Brown, 1879; Marschand, '23 ; Ebbeckc, '23 ;Huebner, '23 ;Ohno, '24 ;the Clarks, '25 ; Barksdale, '25 ; Tannenberg, '26 ; Florey and Carleton, '26 ; Heimberger, '26 ; Zweif ach, '34). Regarding lower vertebrates (frogs) in which the thesis of capillary contraction through Rouget cells was largely established by Rouget, Bensley and Vimtrup, the Clarks ('25) emphatically denied the contractile role of Rouget cells in the capillaries of living amphibian larvae and showed that capillary endothelium alone was capable of active spontaneous contractions in the absence of any kind of adventitial cells. Identical conclusions were arrived at by Zweifach ('34) in an extensive micromanipulative study of the capillaries of the tongue, nictitating membrane, mesentery and intestinal wall of frog. The most prodigious effort to substantiate Krogh's concept of the contraction of capillaries through Rouget cells is the work of Busch, a pupil of Krogh. It is based on investigations in frog, guinea pig, rabbit and man. Of the results Busch writes : It was, however, necessary almost exclusively to study the innervation of these cells in u n k e d preparations, the methylene blue staining of nerves and of Rouget cells not coinciding as to time, the nerve staining being nearly always faded when the protoplasm of the Rouget-cells colours. True, the nuclei of these cells as a rule stain simultaneously with the nerves, but, as mentioned above, it is impossible from the nucleus alone to recognize a particular cell as a Rouget cell. 102 NICHOLAS A. MICHELS In human preparations, however, I have not been able to demonstrate the Rouget cells, even if I have frequently seen nuclei, bearing a remarkable resemblance to the Rouget-nuclei ; neither I nor Dr. Med. B. Vimtrup (who has most kindly examined some of these preparations) dare make any definite assertion on this point. I n human preparations many different nuclei are found around and on the capillary wall, and many of these nuclei, stained with methylene-blue, present the same spotted appearance, the dye being precipitated in the nuclear substance. Even if many of these nuclei strikingly resemble the Rouget-nuclei, no certain conclusion can be drawn from this. Two entirely new concepts regarding the structure of capillaries, significance of Rouget cells, and the mechanism of capillary caliber changes have recently been propounded, viz. that of the Italian school (Volterra, '25; Fieschi and Storti, '29) and its followers (Hueella, Plenk, Goldner) and that of Tudor Jones of Liverpool. Volterra claims that every capillary consists of two tubes, an inner composed of true endothelium, an outer composed of a reticular membrane reinforced by argentophil fibers and known as the perithelial reticulum or adventitia of the capillary. The latter contains histiocytes and adventitial cells but no Rouget cells, the existence of which as muscular elements is denied. The pericapillary reticulum functions as a digusion membrane between the capillary wall and the surrounding connective tissue. It gives rigidity to the capillary wall, maintains its tonicity and plays a trophic role. Through its permeability properties it effects capillary caliber changes these being associated with hydrogen ion concentration changes in the surrounding tissue. Contraction of capillaries is primarily a function of the elasticity of the reticular capillary fibers, characterizing features of which are their net-like arrangement and selective affinity for silver impregnation methods (Bielschowsky, Hortega, etc.) Using silver methods Fieschi and Storti have shown that in the omentum of the rabbit every capillary has its own reticular adventitia composed of extremely h e interlacing fibrillae which histochemically considered are fully comparable to the reticular fibers . of the blood forming organs. Tlie conccyt of Volte1-1~w a s further developed by Goldner (’34). On the hasis of ohservatiom in graiiuIatioii tissue he claims that tlie argent opliil fibers of tlie capillar>-reticnlum are associated with the neoformation of c*apillarics in the sense tliilb tht>J-precede the nev-1:- formed capillaries, orientate and monld their future (:ourso by attracting rasoformative cells (fihroblasts, monocytes) toward them ; tlicse cells then glide along the fihers arid gradually transform into enclotliclium. Qoldner ascribes not only a ‘role morphorrgnlatcn~’to the argent opliil capill a r y fibers but also ii tropliic oiie, i.e., by imbidiiig wtler, retairiiiig colloids and transporting humorti1 substances they a d as ‘ultracapillaries. ’ ‘I’he viewpoint of Jones is its follows : Capillaries arc composctl orily of two elements, viz., non-medullated nerves and fusiform muscle cells. \l;hat anthors have described as Itoiigct cells art) iicurilemma iiuclci, x-hicah stand out from the ~ ( ~ s s e 1 wall, the p ~ t a i r i i i i gcapillary ncrvos arising from a common nrlt. lVliat histologists have described a s capillary c~ndotlielial cclls wit11 silver salt cctmeiit lines aro elongitted sl)irallF twined mnsc4e cells in vvliich the nerves ramify. ‘I’he caldlary tlieii is a iieiiromusciiltir rnechanism, i.c., nothing hut innervated sniootli muscle. Accmdiiigly thcro is “110 need to iiivoko aiiy esrcptionnl apparaiiis in order to cq)lniii capillary consti*icdion.” At transition stages from arterioles t o capillaries (usually iiot visible) the loiigitudiiial muscle cells disappear, ilie circiilar types arc continued ; they become fusiform, assume a spiral for.matiou, the riids of the cells overlapping oiie aiiother. The muscle cells are held togcther hy a rcdicdum o r stroma, itlent ical with that found in any otllcr smooth muscle organ. Jones suhsiantiatcts his speculation with the contention that in the yolk sac of salmon l a r ~ a rthe earliest rcssel s are non-cc~lhilar,i.c., spaces in the rcticulnm ahout wliich there is much cviclencc! of a developing neryolls mechanism. This led him to conclude thiit “iiinermtion is previous not only to the appearance of a contractile mcchAuism, hut to tlie acquisition of a celliilar vascdar wtll. ” ,Jones T H E A N A T O M I C A L H%LOI<I), TOL. 6 5 , S O . 1 used vital methylene blue as his predominant stain. His matcrial comprised tissnos from fish to laboratory animals, best results being obtained from the iris of the rat and the yolk sac of salmon larvac. METHODS AKl) ORSERVATTCISS The material comprises whole mount prcparat ioris tnlren from, a ) normal young and adult rabbits, b) rabbits which ht1d becn snhjcctcd to a single erj-thema dose of x-rays (400 r ) over the omciital area, rabbits which had received several iiitrawnous injections of India ink, d ) rabbits stained intravitally by injection of Berlin blue (0.25 per cent soliitioii) throngh the aorta. 'Che wholc, mount preparations were taken from omerita x-hich had been fixed in ioto with Zmker's solution. Thcy wcre stained with the hc3matoxvlin-aznr-eosiii combination and with various iieurological stains (methods of (ywjc?I, Kissl, Rortegu, Globus, Penfield). 111 the living stale thc pericapillary cells were stntlicd in the web of frog both under riormal and abnormal conditions, the latter comprising tho locd application of chloroform, mechanical stimulatiori, lic!ating antl stretrhing. Tlir omoiitnm of rabbit aflords ziri esccllent o1)portuiiity for tlic study of capillilrics antl their adjoining cells. 1) The organ is sufficiently iliiii and transparwit to be studied i n tvhole mount preparations, 1)rovideci the larger perivascnlar fat bocliw be remored. 2 ) T n the omeiitiim the capillary bcd is extmsively tlevclopod, exhibiting tuft format ion> in various phases of their progression a i d retrogression. 3) No otlicr tissue in the body contains such a multiplicity of (.onnoctive tissue cells and displays suc:li varicd cytogtiiietic interrelationships l)et,wwn them as the omcntum (hlichels, '33 ; llatta and Rutlcdgc, '35). The discordant iiot ioris regarding the iiatare of the Rougct cells are due primarily t o the fact that authors havo not rcstricted their observations solely to capillaries but h a w i n cluded in their tlcscript ions arterioles a i d prccapi 1lary arterioles, structures wlikh contain numerow intermittently (2) placed c:ircular muscle ctAlls. The notion that all minutc vesscds with longitudinal aligrirncnl of endothelial cells are capillaries is orrorieous. In many omental capillary areas fnlly one-half of the minute vessels pertain t o arterioles and p e capillary iir*terioles,ilnd to x-clliilw and prwapillary Ycnules. 111ii capilltirp b d with both arterial a i d vcrtoiis blood supply plainly visible (fig. 1) it can rpadily he asccrtaiucd that Fig. 1 4 rapillary Iwd i n the onientumi of rabbit. x 60. I t illustrates the mode in which branc.hes from the paicbiita I vessels beconic. tlivitlcd iiitn numerou4 prc.c.apillerg arterioles aiid precapillary venules to attain the size and character of capillaries. Four low fields upward the two vessels end in a capillary tuft. 4, arteriole with tht, circuisr muscle cell nurlei placed at rcgular intervals ; V, veniile in which the imdothelial riuelri a r e oval, more nurnwnus and irregularly distrihntcd, the muscle cclls either absent or very sparse. To the right of the lettcr A is a prwapillary veiiul(~ a branch of which unitc.s w i t h a ( apillary cwning directly from the arteriole. Vrnous eapillarirs at I’ and 7). Pig. 2 A ( apillnry and its :td,joiiiiiig connwti\e tisme c.c.11~ in the oinenturn of norrnal railhit, illustrating its single cndothelial character. x 1000. A, plasmodia1 capillary nerve (Renisk fiber) which ma:‘ be followed in figurt, 3. K, fibroc.ytrs, onr under the iicwe; C, nuclei of mexothelial cells; I), pcricytes or so called Rouget cells : 14 2iistior.ptt.x or resting wandering cells; F, retirular adwntitia of Volterra; G, 13, profile and faw viev of cmlothrlial 11ucIc4,the onc in profile could rradilp inistalten f o r a liouget (ell nucleus. the heads being outside of the endothelium (fig. 11). With few csceptioris the mdtii a i - set ~ a t right angles to the vcsscl; their chromatin lias a net-like arrangement whereas in eiidotlielial nuclei it is clotted. The protoplasm of the miiscltl cells is usnally not visiblth; whcii stained it is restricted t o R narrow rim around the niwlcus id on 110 occasioii 1)resrnts i,he liool)-like pcixc1ol)otl foi-rnations of Ronget or tlie s1,iral Fig. 3 Thc~s i m c capillary in :i contiguous field. X 1000. I t is not a spiral complex (Joucs) for endothrlial nuclei :ire lacking on tlict right sidr. A, ncrve fibcr crossing a capillary; B, pericytes generated from outlying 1nol)ilizrrl m e w thelial cells ; C, niesothrlial crlls, one diflcrentiating into :t fibrocytc ; n, pericytrs or so-called Rouget c.clls ; E, histiocj trs ; F, pericapillarp retivular adventitia ; G , If, profile and fact. vic.w of ent1othc~li:rlmiclri; X, process of undrrlying crll; Y, c*urwd rndothrlial nnrlcns simulating a Kongct cell nuelcns of Krogh a i i d Vimtrup. 108 SICHOTAS A. JIICHET,S twitling of Jones. The prewpillary arterioles give rise to nummnis capillaries ~vl-hicharc c?ndothclial tubes devoid of the circnlar m i w l ( ~c ~ l iiuclei l (fig. 11). Tlic transition l o tlie caapillary is abrupt. I%rt:ntal capilla rips l w o m e rcsolvecl into iiumerous capillary loops which ultimatcly tie iip with Fig. 4 Omr~ntalrapillaries a n d adjoining connective tissue cells in tlie normal rabbit. X 600. The oblique one shows no cvidente of being composed of spirally twined miisclt wlls i i b claimed bg .Jones. A , I%, sehwannian nucleated plasmodia1 nerve strand (Reinak fiber) running parallel to a rapillary, the iiiterral being 1 7 p. If placed elosrr to the capillary its nue1e:tted portion eould readily be riiistakeii f o r a perjpyte two of ahieh are shown at (’ ( i 0 k ) . There is no indication that the pericytes ( Rougrt ctlls) are innerv:ttcyl l y the n c r w strand (Jiusch). Capillary t o the right shows a sloughcd off cndothcliiil (211. Thcx upper forking present? tlie perirapillarp reticula1 adventitja. veiioiis capillaries, these in t~u’n with pmcapillary venules and veiiulcs. T n arterial capillaries the cndothelial nnclci a r e sparse, oblong, iiatprow, far apart and mwirily 1)eripheral; in venons capillarks they a r e more numerous, m o w closc~ly packed, round or oval and irregularly (list ribntecl (fig. 11). With iiicrease in size of the venulc the mtlothrliwl nnolci Fig. A portion of the omchntal capillary bed in t h e normal rabbit showing how nuclcatcd wgments of capillary nerves may simiilate ‘ lioiiget cells.’ x 478. Strand A, IS not a pwicyte or Kouget re11 but :t schwannian nucleated segment of a Reniak fiber (rnpillary nrrvc) it being continuous with thc nucleatcd nerve strand J3, which again shows pericapillar!. s(hw:mni:in nuclei a t C and 11. At E, thr strand is lost i n c~iidothelium. F, perirytes; Cr, fihrocgtes. The lower capillary is somtwh:it constriatcil. I n this physioI[tgieal statr its constituent c r l b lime a sllperficial rcwinhlance t o a spirally arranged mechanism. become still more nnmcrous ; circular muscle (:ell nuc1t.i arc absent or very sparse, they being ahsent at times for a disl a i i c ~of 5 mm. of the venulc return. Ti1 a single tuft tlie main arteriole map give rise to 20 to 50 prccapillary arterioles wch of which may hranch into 2 t o 6 caldlaries. ‘I’liesc may run indepciicicntly for 100 t o 500 p to join a veiiule; t h c ? may ~ ~ form a plexixs of 5 to 10 loops or pass directly into a veuule. A capillaq- may be eiitirely arterial and vice versa cuti rely venous. .Distally a 1)recapillary Fig. 6 A nervlp formcd capillary enrlotliclial sprout devoid of pcricytcls and Rouget cells. X 1.500. The black dots ( A ) represents carbon partrcles which wvrc ~~hdgocytosed by endotlwlium and pericytes after intravenous injection of IIiggin’s India ink. Ti, short filamentous growth tip ; C, erythrocytes t*rtravasatcd after x-raying; D, pericptw, one with ingested carlmn ; E, cmdot hclial nucleus. arteriole may give rise t o 2 to 25 c.iipillarv loops. L o n g hefore it has spent itself the pcrcapillary arteriole may give off numerous capillaries to tlie center of the tuft. The most distal capillaries are nearly d\\Tilys naked cndothelial tubes. These point P properly evaluated lead to the irifereiicc that \ v h t Rrogli aiid Vimtrup have described as capillaries Jvit 11 Kouget cells a r e arlcriolrs arid prccapillary arteriolcs with oircnla-r. Fig. 7 Collapsing capi1larit.s from an x-rayed omrntum showing the cssriitial single c*~llulztrconipositim of the eiidotl~elialwall. x 1200. A , grouping of u11dothelial nuclei, one in prophaw mitosis ; K, non-cellular strand with partial rt,tcntion of luinen; C, entrapped erythrocytes; I), pyknotic chromatin granules; 15, sloughed off enrlothrliel cell and rounded u p pericyte containing yellow pigmrnt granules. X 1000. Comparable fields may be seen in the retrogression (Ruckbildung) of newly formed capillaries. 112 NIC HOLM A. MTCHICT,S iriuscle cell iiuclri (their fig. 7 is lalwlcd ill1 arteriole, fig. 8, large ( ~ i ~ ~ ) i l l a r y ) . The second great source of cwnfnsion regarding Roilget cells is the fact thilt ilutllorfi have paid too little attention to the tissiic in which c~ipillai-ieseour.se. Oiice all cellular. constituents of an organ have been earrnarked it is a matter of simple elimiiiat ion a s to which I)(lr.tiiin to the capillai*iefi. For Fig. 8 A capillary unit of the omentuin of rabbit undergoing collapsc~ i d rc4rogression after x-ray treatnient. The lumen cwntains five deep staining r r y t h r o r ~ t e s ,the endothelial cells (the lower in prnpliase mitosis) a r e not spiriilly twined rriusclc cells, the stroma between the eiidothelial cells is not of a rcxtirular nature (strurturcs shown are four nail-staining nyrhroq-tvq). three of the forking strands a r e nun-cellular, the field contains debris, somc in granule form-all data against the liypotlic~sisof Jones. tlie cell types present in the omentum a i d the c1iaractt.r of its aoiincctivct tissiie fibers 1 rclfcr i h c reader t o a previous cornmuiiic~~tion ( ’33). In the aacompaiiying ~ ~ h o t o r n i c r o g r ~all ~ ~ lphascls is of llie Rouget cell problems are presmt ed, viz., 1) the fibrocyie-like cltaractt~rof Ihe mesothelium, which may contribute a share of the pericapillary c d s , 2) the diversity of cell tyl)es obtaiiiing in cu1dIury arcas aiiti their rela1 ionships to thc capillary Fig. 9 (‘ollapwd orriciital c:tpillnrirs from the SIIIIU x-rayed animal, illustrating their single endothclial cllaracter. X (500. A, c~ndothelialnuclei without any sigri of being spirally twined or having tlir cliaractcr of smooth muscle; I<, entrapped eryt1iroc)tc.s ; (2, non-cellular region of endotlirlial strands. Tlirse are comnlon piic,noniena in rrtrogression of capillaries a s a r c likewise t h c ‘granular extensions’ set’ri at I). (Here rellular debris am1 not aspects of nc,iirogt,ncGs.) 114 XICIlOLBS A . MICllELS wall, 3) the structure of the capillaries in their fully differentiated state, in stages of hutlcling and collapso, 4) the nucleated non-medullated capillary nerves, 5 ) tho pericapillary atlventitial 1-eticnlar fibers of Voltcrra, ti) absence of arteriolar and veiiule mnscle cells, the regions representing only capillaries. Fig. 10 Two oniental collapsd eapillaric~sfrom the same animal cxltibitiilg the extreme Imgtli into which endothc~ltalcells may he drawn during rctrogression. Evidence of the single character of the endothe1i:jl wa1L X 750. A, naktvi en dothclial strand with a single endotht4ial xtueltws ahout which a r c sevrrnl ( i t h i s grannles ; H , non-cellular region of an cmrlothelial strand with intcrm~ittently orclucletl lumen; C, group of endothelial nuclei ; I), entrapped erythroeytcs; E, clasmot ocyte with granular debris, products of rctrogrcssion ; X, lumina a t forking. A and J3 tie up with a field comparable t o t h a t in figure 7. The half canalized, e r y t h r o q t e containing, collapsed cndotlwlial strands ( 2 t o ti p ) can readily be distingushed from the solid, evenly eontourcd, 1 L./ wide, srhwannian nucleated (Kemak) fibers seen in figures 2 t o 4. T ~ Ithe dililt cd stat(’ fully differentiated capillaries havc a remarkably uniform architccture (figs, 2 to 5 ) . The enclothclial wall itself is very thin but due t o a juxlaposition of the r.pti<Lnlar fibers and cytoplasmic proct.ssc~sof pericytcis it often appears t o bc considcrahly thicker thaii is actually the. case. The eiidotlielial nuclei arc provailiiiglp oval aiid oblong, iheir longitutlinal axis in most instances being orieiitcd in the tlirectioii of the capillaries. 111 figures 2 anti 3 oval-shaped nuclei in profile view are shown at G, in face view at 1%. Curved endothelial iii.iclei ( Y , fig. 3) and those iii profile view ( G ) oftcii hare a striking similarity t o Roiiget cell nixlei figured in the Iitctrat,ure. ‘rho existence of c:ontracted, inclented, (wrvcd, swollen, rnfllcd and ronnded-up mdoihelial niiclei is explainable on the basis o f consonance of form with physiological activity and the state in which they were at thcl time of fixation. From the freqnc?ncy of cell types partially attaclied to the capillary wall (figs. 4 and 5 ) it may r.easoriahlp he Ca0nCh7d~?d that cndothelial cctlls often become cicsquamated and t hereaficr porsist either as pcricj-tes or undergo transfornintiol1 into cell types iririistiiiguishshlt. from fibi*oblasts ( M aximow) . Ko better opportnnity for a detcrmination of the c~cllliilar constituents of the capillaries is afforded than in their stages of bizcltling and collapse (figs. 6 and 7). Budding ciipillarics grow out from preexisting capillaries a s single o r cloiible st rands of endotlielium and only subsequently acquire pericyt es, while collapsed capillaries rctain the morphological cliarticteristics of endothelium long after the circnlatioii lias c w ~ e t l . The ultimate fate of the vast majority of endotlielial ccills of irreversibly collapsed capillaric~sis tliei r dediff ercntiation into mc~sonchpmal(>ells(resting wandering wlls, polyblit~ts). 1’ericyfc.s. Muc.li 2ias bccii writtcw r(>gardiiigthe morphology, gcncsis and function of pcricptcs but to date 110 definite conwpt of these cells prevails. The notion that there is a gradual transitioii from typical smooth miiscle cc?lls f o u n d 011 arteriolcs and veiiules to tliose obtaining on capillaries :IS Bonget cells caniiot he ciitert aiiied histologically. h’o one has convincingly d e mo n s t r a t d the existence of partly myofibrilirial (TI coniiective tissnc cells in fixed material, while in vital preparations it is highly probahle that intercellular cement lines have erroneoiisly been interpreted a s mpofibrils. Secondly, under t he term Rougct cell (pericyte) a variety of cellular clement s ~ R V Cbeen included. I have already shown how smooth miisclc cells 011 arterioles and precayillary arterioles have becii counted a s Rougct cells. Other elements that have been iiicluded are crescent-shaped endotlielial nuclei and those in profile view, sloughed off endothelial CCI~S, emigrating lymphoid cells, pericapillary fibroblasts, clasmatocytes, polgblasts, advcntitial cells, undifferentiated mesenchyme cells, reticidar cells, dotigated Idasma cells, mast cells and even nuclpated segmented of perionpillary nerves (Remak fibers). A study of the photomicrographs illustrates amply the participation of connective tissue cells in the formation of peric~apillaryoblong cells. TVhidi of these should be termcd pcricytes and which not lias Ixen a matter of individnal opinion f or to date no spccaific morphological o r functional criterion has been found to distinguish a pcricapillarp fihrocyic from a pericytc, nor a pericyte from a sloughed off endothrilial cell or any other type of inidifferentiated oblong Fig. 11 Onientunr of noriiial rabbit. Iiorte~l-Globus-P(,nfieldstain. Carnwn lucida drawings. Objective 2 nim., 1.4, oculars, X 70, 15. 1 ) Arteriole illustrating the marginal scriation, transverse position and brnt character of t h r circular muscle cells. Their nuclei h e w n chromatic pattern which is nct like, thrir protoplasm is e\ciily contoured. I n endothelial nurlei the chromatin ia dust-like, the protoplasm fusiformlp arranged. 2 ) Prccapillary arteriolc with characteristic far spread orientation of circular muscle erll nuclei. Krogh and Vimtrup interpret thest. as brlonging t o Kouget cells, but their morphology is identical with that obtaining in arteriolar miiscle cclls. Thc reticulum of Volterra is seen. 3 ) An arterial capillary. I t consists only of longitudinally disposed eiidothrlial cells, one in mitosis. Four ergthroca?tc’s. 4) A venous capillary. l’hr endothelial nuclei ~ T Cbulkier, predominantly oral, mow numerous, anti inore closely packed. 3 ) Venous capillary showing mitosis in an endothelial c ~ l l . 6 ) 1’ree:ipillary arteriole with intermittent circular muscle cell nuelci and adjoining conneeti\ c, tissnc cclls. A, pcric~apillarpundiff wentinted niesenc~hyniecrll ; H, pcricptr with phagoc ytosed material; C, rryting wandering ccll ; D, fibrocyte ; E, plasma c ~ l ;l F, niesothelial riuc*leus; G , large lymphocyte ; H, rlasniatocpte. 117 118 SIClLOLAS A. iVIlCRETA5 niesciwhymal cell. Abseiice of large dark niw~eoliin endotheliid nuclci choiistitute practically the only morpliological difference between tlic nuclei of fibroblasts aiid entlothelial cells, f o r in both types the chromatin granules have a dustlilw distribii t i on. Hevxog ('lfi),AIaximow ( '27) i111tl others htivc i~dVo(~~Itcd a rlistinctioii bctween two kinds of adventitial o r pericapillary c*clls: 1)Ail inner group usually contiguous with the surface of t h e enclothelial tube consisting of flat syinclle-shaped cells wif h oval endothelial imclei, aiid sharp cytoplasmic processes u-hich do not store dye granules, 2) an outer group comprising c d l s comparable to restiiig wandering cells 01- histiocytes in tliut they liave a varied morphology, vacnolated cytoplasm, round cytoplasmic proccwes and liavca dye storing proporties. According to ltaximow, the first group u c undifferentiated (TIIS Jvhich nor.m;illy sliotv no traiisitioa to the second groul) hut iindcr inflamm;-\tory conditions and in tissiie c i d t w e s bewrne uctivatcd gci~cmtingnew resting 11-aidering cells, histiocytes ant1 fibrocytes. 7'his cutcgorization of pericapillary cells is more spaciid 1han morpliogeiietic f o r the site relations and functional roles of tho ~ ~ 1 are 1 s often rtvci-setl. Typical histiocytcs and restiiig wandering cells ma? have the position of pericytcs (figs. 2 and 3) j mcsenchymisl fibrocytw often coristitute the outer rather than the inner group of ptlricapillary cells (fig. 2). Pericytes may he generated from outlying mobilizcd mesothclial cells (R, fig. 3), dcscpamatetl endothelial cells ant1 cmigratcd lymphoid cells. Functionally considered the jniicr perkytcs are not incapable of phagoc>-tosis as claimed by ,\I ilximow for after intravenous in,jeotions of d l o i d i ~ carbon l (fig. 6) pcricytcs often contain carbon particles. In silvcl-ed aiid intravitally stained preparations interccllular cement lines in endothelium may have a superficial resemblance to myofibrils (precipitation artefacts). Pcricytes never contain iiitracellular inyofibrils, iior do they possess tapering processes which eiicircle the ca1)iIlary tube like hooks, as Ronget cells a r e supposed to do. CAPILLARIES AND PERICAPILLARY CELLS 119 Although firmly enrooted in the literature the term Rouget cell should be dropped for histologically considered the concept of a pericapillary contractile cell is extremely vague and contradictory. The distinction which Krogh and Vimtrup make between Rouget cells and pericytes does not hold for a goodly number of their Rouget cells are circular muscle cells on precapillary arterioles. Rouget originally stated that the distinguishing feature of the cell is its nucleus. The latest worker (Busch) takes an opposite stand in the statement “it is impossible from the nucleus alone to recognize a particular cell as a Rouget cell.” Furthermore, it is highly probable that many of the so-called Rouget cells have been mistaken for nucleated segments of capillary Remak fibers. The extreme length of some of the Rouget cells reported in the literature (217 p, Zimmermann ; 200 p, Vimtrup, 60 to 200 p, Krogh) favors this view for pericytes and fibrocytes rarely exceed 90 p in length, while smooth muscle cells are considerably shorter. Early in my studies on capillary innervation I learned to pay particular attention to numerous short. and long undulations of capillary nerves, for they enabled me on many occasions to distinguish the nerve strands from capillary endothelium, connective tissue fibers and cell processes. When the capillary plasmodial nerve strands run very near (1 to 2 p) and parallel to the capillary, their undulations repeatedly touch the capillary wall, eventually being lost on numerous occasions in the endothelium (figs. 2 to 5). I n viewing such pericapillary nerve strands at sites of their Remak nuclei in a single microscopic oil immersion field, they appeared as Rouget cells. However, after a projected drawing of the regional capillary bed was made and the pericapillary nerve strand was followed in its relation to the capillary bed, it was often noted that it eventually tied up with the general capillary nerve net. Figure 5 is a typical illustration of these points. The long filamentous structure ( A ) running along the upper capillary for 105 p is not a Rouget cel1, or pericyte, but a continuation of the nucleated plasmodia1 nerve strand THE ANATOMICAL RECORD, VOI.. 65, NO. 1 120 IiICHOMS A. MICHELS seen in between the capillaries, the strand having crossed the upper capillary in the wall of which it eventually becomes lost. Following the intercapillary strand downward, one notes that it exhibits numerous short undulations which touch the endothelium and that it contains two oval nuclei which if studied alone might be mistaken for the nuclei of Rouget cells. Instances in which a nerve fiber of the type shown in figure 4 is more closely applied to the capillary wall (fig. 2) could readily be mistaken for an extremely long pericyte. In view of what has been said there is reason to believe that Busch in his contention that Rouget cells have a sympathetic nerve supply, mistook Remak nuclei for Rouget-cell nuclei. Since in my projected drawings of capillary nerves no evidence was found that the nucleated plasmodia1 nerve strands upon reaching the capillary wall, cross one cell type more frequently than another, it may reasonably be concluded that the selective orientation of capillary nerves toward the nuclei of pericytes (Rouget cells) as propounded by Busch is erroneous. I n short, the Rouget cells are not muscular elements nor have they a sympathetic nerve supply. With regard to the recent work of Jones the following comments may be made. Entia non sunt reducenda (as opposed to his ‘multiplicanda’) praeter necessitatem. The staining of capillaries by methylene blue is ‘fickle’ as he admits. I n his own preparations he observed capillaries and nerves, nerves without capillaries, capillaries without nerves, independent nerve nets, smooth muscle with and without nerves. This marked variation of histological data is anything but ‘ideal’ for the isolation and determination of capillary components. In his methylene blue preparations “arterioles do not appear in their continuity neither do the capillaries.” Only in one or two cases were transitions from arterioles to capillaries seen. I n my material the transition stages are numerous and clear cut, especially in the HortegaGlobus-Penfield preparations. These in particular present no evidence that arteriolar circular muscle cells become continuous as spirally twined fusiform endothelial cells. The CAPILLARIES AND PERICAPILLARY CELLS 121 statement that “never more than two smooth muscle fibers can be seen a t any level in the capillary system” is erroneous. Capillaries frequently have double rows of endothelial nuclei a t upper and lower peripheries. Phenomena of grouping (two to five) and near spacing of endothelial nuclei are common especially in the venous capillaries ; while in intimate contact with the capillary wall may be seen a pericyte, a sloughed off endothelial cell, an undifferentiated mesenchymal cell, a segment of a schwannian nucleated Remak fiber and an emigrating lymphoid cell. Not all capillaries have a nerve supply. I n my report on the plexus omentalis I have shown by means of numerous micrometric measurements that the meshes of the capillary nerve plexus are considerably larger than those of the capillary bed itself, the result being that many capillaries consist only of naked endothelial tubes. The alleged noncellular character of primitive blood vessels in the yolk sac can be explained on the basis of retrogression (Ruckbildung) of capillaries. It has long been known that not all newly formed capillaries attain functional permanency. This is especially apparent in the embryo. Maximow covers this topic fully in his large German work and gives convincing figures. Jones’s statement that in vital preparations the developing nerves appear as “exceedingly minute rods and granules forming an extended and delicate tracery around the lumina of the vascular channels” and above all his figure 16 indicates degenerative phenomena and extravasations rather than developmental aspects of specific neural entities. Finally the injection method is not the best to study capillary structure. It has led us into many errors regarding lymphatics and is apt to do the same in regard to capillaries. I n my omental preparations, i n j d e d intravitally through the aorta with Berlin blue, a very high percentage of the capillaries are distorted, showing partial collapse, spiral twining and intermittent bulging ;these phenomena extending even to arterioles. I n support of the single character of the capillary wall I offer figures 7 to 10 representing capillary units undergoing retrogression after x-ray treatment. The non-cellular charac- 122 NICHOLAS A. MICHELS ter of some of the endothelial strands, the sparsity of endothelial nuclei in many others and the extreme length and thinness into which single endothelial cells may be drawn during retrogression is convincing evidence that capillary endothelial cells are not spirally twined smooth muscle cells. On the other hand the erythrocyte content of endothelial strands, the grouping of endothelial nuclei, the retention of partial lumina, the presence of cellular debris (some in the form of rods and granules) constitute never failing differential criteria whereby one can distinguish collapsed capillaries from schwannian nucleated capillary nerves. LIVING MATERIAL My studies of the behavior of the capillaries in the web of living frogs under normal and experimental conditions revealed the following blood flow alterations : 1) Pulsating, 2) steady, 3) intermittent, 4) very slow, 5) very fast, 6 ) partial and complete stasis. As a rule, the cited current alterations were accompanied by caliber changes in the capillaries. I n the normal frog it was relatively seldom that all capillaries were functionally active at the same time. Often capillaries and capillary units, although open, showed no blood flow at all. Open capillaries on many occasions were seen to contract, leaving distal short filamentous strands which subsequently became reopened. Collapsed capillaries were invariably short and, as a rule, were restricted to anastomosing units. Reversal of the direction of the blood flow was accompanied by marked diameter changes in the regional capillaries, the result in instances being partial or complete collapse of the capillaries. Frequently capillaries and capillary loops functioned as reservoirs for white corpuscles which temporarily were being removed one by one from the blood stream. As the number of white corpuscles increased the lumen between the endothelial tubes became extended and remained so, sometimes for hours, until the white corpuscles reentered the blood stream. Whenever in any given capillary the number of white corpuscles became too numerous to pass CAPILLARIES AND PERICAPILLARY CELLS 123 through the lumen, the adjoining endothelial wall would become buldged out, thereby forming an open pocket until such time as the supernumerary white corpuscles could be swept back into the blood stream. After local application of chloroform by the drop method, the regional capillaries dilated and the blood flow became progressively slower, until it finally ceased. Maximal dilatation of capilIaries was obtained with heat, continued application of which led to complete stasis. Mechanical stimulation and stretching resulted in a variety of capillary caliber changes, frequent phenomena being partial and complete collapse of the regional capillaries. I n all the capillary caliber changes observed there was no evidence of a contractile role being played by pericapillary cells or so-called Rouget cells. Since with every dilatation and contraction of the capillary wall, the movement of the oblong pericapillary connective tissue cells (pericytes) was passive the inference is fully warranted that naked endothelium in virtue of its inherent ameboid activity is primarily responsible for caliber changes in capillaries. Whether the movement of the endothelium is initiated by differences of capillary blood pressure (Landis), changes in the hydrogen ion concentration in the surrounding tissues (Volterra) or through nerve stimulation either directly or through the liberation of a neurohumoral substance remains an unsolved problem. For further data on capillary innervation and my own investigations on its underlying anatomy I refer the reader to my recently published paper “The plexus omentalis and its relation to capillary innervation. ? ’ 124 NICHOLAS A. MICHELS CONCLUSIONS A capillary is redefined as a naked endothelial tube. It must be distinguished from a precapillary arteriole and precapillary venule; the former contains many circular muscle cells, the latter few or none. Neither in fbed nor in vital preparations of the omentum of the rabbit was any evidence found for the muscular character or sympathetic innervation of pericapillary cells. Since no one knows what a ‘Rouget cell’ is, the term should be dropped. Pericapillary connective tissue cells should henceforth be termed pericytes as originally suggested by Zimmermann. Morphogenetically considered pericytes comprise the following : primarily, fibrocytes and undifferentiated mesenchymal cells, secondarily, histiocytes, resting wandering cells, emigrating lymphoid cells and sloughed off endothelial cells. When closely applied to the capillary wall segments of schwannian nucleated capillary nerves (Remak fibers) may readily be mistaken for pericytes or Rouget cells (Jones). No evidence was found for the view of Jones that capillary endothelial cells represent elongated spirally twined circular arteriolar muscle cells carried over into the capillary bed. In the living capillaries of the web of frog under normal and experimental conditions capillary caliber changes may be effected by the regional endothelium without the aid of pericytes or Rouget cells. Acknowledgment is made of my indebtedness to Dr. J. Parsons Schaeffer for his interest and clarifying suggestions extended during the work. CAPILLARIES AKD PERICAPILLARY CELLS 125 LITERATURE CITED References not cited here will be found in the bibliography published by Krogh and Vimtrup (’34) in their article, The Capillaries, Special Cytology vol. 1, p. 475. Hoeber, New York. EBBECKE, U. 1923 (fefiissreacktionen. Ergeb. d. Physiol., Bd. 22, S. 401. 1923 Endothelzellen, ‘bugetzellen’ und Adventitialzellen in ihrer Beziehung zur Contractilitat der Capillaren. Kl. Woch., Bd. 2, 5. 1341. FIEscm, A. AND .& STORTI 1929 Richerehe sui capllari sanguigni e sul teSSUt0 reticolare dell-omento. Bol. SOC. Med. Cir. Pavia, T. 43, p. 523. GOLDKER, J. 1934 Sur la neoformation des capillaires dans lee tissus inflammatoiles. Anal. d’ant. path. et d’ant norm., T. 11, p. 461. HASSIN,G. 1929 The nerve supply of the cerebral blood vessels. A histologic study. Arch. of Neur. and Psych., vol. 22, p. 375. HEUBNEE, W. 1923 Physiologic und Pharmakologie der Blutcapillaren. Klin. Wochens., Bd. 2, S. 1965. JOKES,T. 1936 The structure and mode of innervation of capillary blood vessels. Am. J . Anat., vol. 58, p. 227. LAWA,J. AKD D. RGTLEDGE1935 The reaction of omental tissues to trypan blue injected intraperitoneally, with special reference to interrelationships between cell types. Am. J. Anat., vol. 56, p. 481. MICHELS, N. A. 1933 Susceptibility of the omenturn of rabbits t o a single erythema dose (400r) of Roentgen rays. Am. J. Anat., vol. 52, pp. 333-395. 1935 The plexus omentalis and its relation to capillary innervation in the omentum of rabbit. Am. J. Anat., vol. 57, pp. 205-258 (additional references). NESTWW, A. 1925 fJber Contractiltkit der Blutcapillaren beim Menschen. Arch. f. d. ges. Physiol., Bd. 209, 5. 465. OHNO, Y. 1924 Beitrage zur R a g e der neuropathologischen Entziindungslehre. Ziegler’s Beitrage, Bd. 72, 8. 722. PARRISIUS, W. 1921 Zur Frage der Contractilitiit des mensehliehen Hautcapillaren. Arch. f. ges. Physiol., Bd. 191, 8. 217. R O G m , J. 1932 Observations on the pericapillary cells in the mesenteries of rabbits. Anat. Rec., vol. 54, p. 1. ZWEVACII, B. 1934 A micro-manipulative study of blood capillaries. Anat. Rec., vol. 59, p. 83.