STUE‘IES ON INDUCEQ REGRESSMN . V Ni LAR‘JAL AMBYSTOMA' LEMBS Dissefiation for the fiegree of Ph. D. WEHEGAN STATE UNWERSITY LUCELLE G. ADAIR 1973 LIB? '-.RY( Michiga. . State Univerity This is to certify that the A thesis entitled STUDIES ON INDUCED REGRESSION IN LARVAL AMBYSTGMA LIMBS presented by Lucille Adair has been accepted towards fulfillment of the requirements for ”MB Aegree in 199.1% /\‘. 1/ L llaknrproflaunu' Date July 26, 1973 0-7639 Aauunue“uv " HOAB & SflNS' 800K p’HDERY INC. [JERA‘H’ BINDERS r ...,:.wcumgy We? I Aasrmt - '0. ‘ - h S! JHL‘b 0N lumen REGRESSMI t ”fun . U. 1 LAWN. WYSIW LWS .‘ | h} ‘0‘ '3 brim! .. mir ‘..I'M"‘, TL' V ‘ ‘r . g -' ' The regressive uhafié i.“ an undated Zarva‘. tat-pads. m: not if I“ '.‘ ."v' t ‘fm uunsive and the ent‘r-e "one resorted 1! -~. flaw f; ‘g: mutated or denervned ("t-tone“ reqnssuv m «3334 :g‘ amen!» .. a abrasion injury to We 53mm} muses cl’ an: at . 1am- ‘ V‘Ilflcbfluktn mesa: em. mxmngwtv g "In of the more: stag ‘2“ at? Mt 1h ,L,‘ V-V div» ‘_' ”‘3“ v-v‘ _>~»'T'u>‘ ABSTRACT STUDIES ON INDUCED REGRESSION LARVAL AMB$§TOMA LIMBS By Lucille G. Adair The regressive phase of an amputated larval urodele limb will become extensive and the entire limb resorbed if it is also x-irradiated or denervated. Extensive regression can also be induced by a compression injury to the mesodermal tissues of the limb, involving a break in the skeleton, and accompanied by denervation. Reinnervation of the regressing limb will halt the disintegrative processes and in some cases lead to regeneration of the limb. The experiments in this investigation were designed to study (1) regression induced by a subdermal homoplastic implant of a piece of humerus in a larval limb without injury to the musculo-skeletal system and with subsequent denervation, and (2) the effect of rein- nervation in these limbs. Both histological studies and a statistical analysis of humeral volume changes in the experimental left limbs and untreated right limbs were made. The limbs of the Ambystoma gpggum_larvae in Series I were untreated and provided, therefore, the normal base for the histology and humeral volumes for this species. Lucille G. Adair The left limbs of the Ambystoma gpgggm_larvae in Series 11 received homoplastic implants of a piece of humerus. The limbs were denervated four days later and kept denervated throughout the experiment. Histological disintegrative events were initiated and there was a significant loss in humeral volume. The untreated right limbs of the Ambystoma maculatum larvae in Series III constituted the controls for this species. The in- nervated left limbs of these larvae received homoplastic implants of a piece of humerus. The implant without denervation did not induce extensive tissue regression. and the localized muscle destruc- tion was soon repaired. Periosteal cartilage cuffs were induced on the host humerus. Though these were included in the humeral volume measurements, the resulting increase was not significant. Denervation alone of the left limbs of the Ambystoma maculatum larvae of Series IV led to muscle atrophy and destruction. The skeletal elements remained normal in appearance, but the humeral volumes did not show an increase with time. The left limbs of the Ambxstoma maculatum larvae of Series V were both implanted and denervated as in Series II, but the nerves were permitted to regrow after the final denervation on day 19. Reinnervation. which began about the twenty-eighth day, reversed the disintegrative and resorptive processes evident in the first part of the experiment. Periosteal cartilage formation resulted in an increase in humeral volumes up to day 22. From days 22 to 28. how- ever. humeral volumes decreased in spite of additional periosteal Lucille G. Adair cartilage deposition. Following reinnervation the humeral volumes showed a marked increase. This investigation has demonstrated that limb regression is induced by a subdermal homoplastic implant of a piece of humerus in a larval Ambystoma limb without injury to the musculo-skeletal system and with subsequent denervation and that reinnervation effects recovery in the regressing limbs. It has also shown that there are some distinct differences in the response in the two species used in the experiments. Nerve and muscle degeneration and humeral resorption occur more rapidly in Ambystoma gpgggm. The implant is phagocytized and resorbed more rapidly in Ambystoma maculatum and its complete removal coincides with the appearance of periosteal cartilage deposits on the host humerus. The implant is not completely removed in Ambystoma gpgggm_during the time course of the experiment and periosteal cartilage is not seen in this species. The humeri in the untreated right limbs do not show normal growth when the brachial nerves are transected on the contralateral side of the body, though they appear normal histologically. This effect occurs later in Ambystoma maculatum than in Ambystoma gpgggm, STUDIES ON INDUCED REGRESSION IN LARVAL AMBYSTOMA LIMBS By _X_£ V' Lucille GS‘Adair A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Zoology 1973 V'l “KITS A V m?! ’l J." ' "ill 1m:- n; .14..“ dr'fll' 3:36;“. 4;”) (18‘ 51.. . I ext’my LMnk' um ‘u": dish-as to the MS o.’ aw “in Canittee Jr. t't'av‘le‘. Ln’hornton, Dr. John Show. 3.~' ”chard Fennel». Md -.l a-‘flflfltfl, They may have m M! ‘f '3: record. ';; ' L .._. Cy amial thanks in {.r, my ’52'4‘.‘ for his .if‘ncrgson : ‘~ {w it a substitute ain‘t {1:2 umber un vrry fan‘- (flee :2 ‘ I sincerely appreciate tnv “new an‘ sum and meta ' ~ -- I 1”“ ml: at various Hm. .‘ v». is": in! 9' m V I . ; Qllm: Lonnie Ella-ed. are». .Igm...:v.5.;an, 6.0 w. _ , in to larch, Dr. John in :1. ‘ift v... flew u- teri: m. ‘ rm’r.m INI" ”“-’(§ To m in new personal «was! wane-Wt an .0 I . "If M . ’cmr Inn“ 8: sw‘fifi 131 W' "flan Mm. tutu m in w, ”as... w £3351. ACKNOWLEDGMENTS "all was done by grace alone“ Zech. 4:7b (LB) I extend my thanks and best wishes to the members of my Graduate Committee: Dr. Charles S. Thornton, Dr. John Shaver. Dr. Richard Fennell, and Dr. Charles Taban. They may have set an endurance record. My special thanks to Dr. Roy Tassava for his willingness to serve as a substitute committee member on very short notice. I sincerely appreciate the timely assistance and advice given by many people at various stages of this work. Some of these are as follows: Lonnie Eiland, Barbara Johnson-Muller, Judy Warner, Dr. Stephanie Barch, Dr. John Gill, Dr. Tom Connelly, Larry Weidner, Dr. James Asher. and Thelma Arnette. To those who gave personal interest, encouragement and prayer, I am deeply indebted. I cannot begin to express my grati- tude to Vivian Adams, Thelma Arnette. The Lyons, The Larnards, "Mac" Henderson, Velma Croff, Margaret Ferguson, and many others. iii TABLE OF CONTENTS LIST OF TABLES ......................... LIST OF FIGURES ......................... INTRODUCTION .......................... MATERIALS AND METHODS ...................... Basic Operations ...................... Experiments ......................... Histological Preparations .................. Measurements ........................ Statistics ......................... RESULTS ............................. Preliminary Experiments ................... Histology .......................... Analysis of Humeral Volume Changes ....... . ..... DISCUSSION ........................... SUMMARY ............................. LITERATURE CITED ........................ vi To 34 Table LIST OF TABLES Summary of the length changes in the denervated left limbs of the preliminary experiments ........ Means, standard deviations, and standard errors for the humeral volumes in Ambystoma opacum ....... Means, standard deviations, and standard errors for the humeral volumes in Ambystoma maculatum . . . . . Analysis of variance table for the three-factor factorial analysis of interaction in Series III. IV, and V ........................ Page 36 37 38 Figure 1. LIST OF FIGURES Page A longitudinal section of a normal humerus of Ambystoma opacum .......... . ......... 41 Higher power view of the diaphysis seen in Figure l. The two layers of cells composing the periosteum can be clearly seen ............. 41 A longitudinal section of a portion of the upper arm of an Ambystoma o acum larva eleven days after implantation ans seven days after denerva- tion (Series II) . . .................. 41 A longitudinal section of a portion of the upper arm of an Amb stoma opacum larva showing the implanted piece 0 humerus four days after implantation (Series II) ................ 41 A longitudinal section of an upper arm in Amb stoma o acum at day 23 (Series II) from the s e o t e humerus closest to the implant showing a large number of osteoclasts .......... 43 A longitudinal section of a lower arm in Ambystoma o acum at day 20 (Series II) in which the presence 0% osteoclasts and wrinkling of the periosteal bone evidence resorption in the radius ............ 43 A longitudunal section of an upper arm in Amb stoma opacum at day 14 (Series II) showing the multg- layered periosteum with osteoclasts along the humerus . . . . . . . . . . . , .......... . . . 43 A longitudinal section of an upper arm in Amb stoma o acum at day 17 (Series II) showing wrinkling o? tEe periosteal bone and telescoping at the junction of the diaphysis and the growth region . ........ 43 vi Figure Page 9. A longitudinal section of an upper arm of an Amb stoma maculatum larva nine days after implantation and five days after denervation (Series V) showing the implant and the multi- layered periosteum ................... 45 l0. A longitudinal section of an upper arm in Amb stoma maculatum at day 19 (Series V) showing a small periosteal cartilage formation and the multilayered periosteum along the humerus ........ 45 ll. A longitudinal section of an upper arm in Amb stoma maculatum at day 19 (Series V) showing a flbrocellular scar at the implant site ........ 45 12. A longitudinal section of an upper arm in AmbEstoma maculatum at day 25 (Series V) showing at exten51ve damage to the humerus and periosteal cartilage ........... . ...... 45 l3. A longitudinal section of an upper arm in Amb stoma maculatum at day 34 (Series V) showing a large periosteal cartilage callus along the humerus ..................... . . . . 47 14. A longitudinal section of an upper arm in Amb stoma maculatum at day 37 (Series V) showing periosteal Bone deposition on the surface of the extensive cartilage callus ............. 47 l5. A longitudinal section of a lower arm in Amb stoma maculatum at day 31 (Series V) showing osteoclastic damage to the radius and ulna ....... 47 16. A longitudinal section of an upper arm of an Ambystoma maculatum larva four days after , implantation in an innervated limb (Series III) showing the early increase in the number of layers in the periosteum ..... . .......... 47 17. A longitudinal section of an upper arm in Amb stoma maculatum at day l5 (Series III) showing the multilayered periosteum containing many large basophilic cells (chondroblasts) which apparently produce the periosteal cartilage calluses found in this series ....................... 49 vii Figure 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Page A longitudinal section of an upper arm in Ambystoma maculatum at day 21 (Series III) showing an extensive periosteal cartilage callus, some of which is covered with a deposit of bone . . . . . . . . . 49 A longitudinal section of an upper arm in Amb stoma maculatum at day l5 (Series III) showing '_—JLT§—-' a num er of multinucleate giant cells concentrated at the implant site ................... 49 A longitudinal section of an upper arm of an Amb stoma maculatum larva 20 days after a__JL_____ enervation (Series IV) showing macrophages among the thin, widely separated muscle fibers ..... 49 Series I. Amb stoma o acum. Left limb humeral volumes of t e untreate control animals ........ 5l Series I. Ambystoma opacum. Right limb humeral volumes of the untreate control animals ........ 53 Series II. Ambystoma opacum. Humeral volumes of the experimental left limbs ............. 55 Series II. Ambystoma opacum. Humeral volumes of the right limbs ................. . . 57 Series III. Ambystoma maculatum. Humeral volumes of the right untreated control limbs ....... . . . 59 Series III. Ambystoma maculatum. Humeral volumes of the implanted left limBs ............... 6l Series IV. Amb stoma maculatum. Humeral volumes of the denervated left limBs .............. 63 Series IV. Amb stoma maculatum. Humeral volumes of the right limbs ................... 65 Series V. Amb stoma maculatum. Humeral volumes of the experimental le t 1m 5 ............. 67 Series V. Ambystoma maculatum. Humeral volumes of the right limbs .................. . 69 viii INTRODUCTION Following amputation and wound closure, a urodele limb enters a regressive phase. During this time (approximately one week in larvae and two weeks in adults) the debris resulting from the injury is removed by phagocytes, histolysis of the formed structures sub- jacent to the amputation plane occurs, and morphologically dediffer- entiated cells begin to appear in increasing numbers. Hhat initiates and controls this period of histolysis, which is a vital prerequisite to regeneration, and what finally reverses the trend in order that a regenerate may be formed, is far from understood. Experiments in which full thickness (dermis and epidermis) skin flaps are sewn over the amputation surface indicate that the formation of a wound epidermis (apical epidermal cap) is essential to this regressive phase since this technique inhibits dedifferentia- tion (Tornier, 1906; Taube, 1921; Godlewski, 1928; Efimov, 1933). Thornton (1958) prevented the formation of an apical epidermal cap in amputated Ambystoma larvae by daily irradiations of the wound skin with ultraviolet light and found that dedifferentiation of the under- lying mesodermal tissues was delayed. Carlson (1969) found that the formation of an apical epidermal cap was delayed and dedifferentiation was reduced when axolotl skin was treated with actinomycin D prior to amputation. Hhen UV treatments were stopped in the case of Thornton's work or the effects of the treatment wore off in the case of Carlson's work an apical epidermal cap was formed and dediffer- entiation and subsequent regeneration proceeded. It has been demonstrated that the wound epidermis is involved in the removal of wound debris. This may be accomplished by phago- cytosis (Rose, 1948; Taban. 1955; Scheuing and Singer, 1957; Bodemer, 1958; Schmidt, 1958; Norman and Schmidt, 1967) or by the movement of cellular and particulate matter through the wound epidermis (Singer and Ray, 1957; Bodemer, 1958; Riddiford, 1960). There is also evidence for some lytic activity by the wound epidermis (Orechowitsch and Bromley, 1934; Adova and Feldt, 1939; Taban. 1955; Hroblewski and Grzybeck, 1961). Schmidt (1968) suggested that the high cathep- sin activity in the wound epidermis found by Hroblewski and Grzybeck (1961) "could well be responsible for a large measure of the cathep- sin activity reported in the preblastemic phases of regeneration" (p. 201). A collagenolytic enzyme, whose activity is maximal during the dedifferentiative phase, has been isolated from regenerating newt limb tissue, but the site of production has not been identified (Grillo, Lapiére. Dresden, and Gross, 1968; Dresden and Gross, 1970). Though the wound epidermis is necessary and is active during the regenerative phase it is not known whether it plays a primary initiating and/or controlling role or whether the activities demon- strated are responses to stimuli from the injured meSodermal tissues. Butler (1933) found that the regressive phase which follows amputation would continue unabated in larval limbs that were also x-irradiated. In subsequent studies Butler and Puckett (1940) showed that if the establishment of a blastema was suppressed by either x-irradiation or UV irradiation the formed structures of the limb dedifferentiated extensively. If x-irradiation was delayed until after a blastema was fully established, the limb tissues became static with neither excessive dedifferentiation nor regeneration occurring. 0n the other hand, ffthey suppressed its activity at the time the blastema was forming dedifferentiation continued and went to an extreme. These authors proposed the hypothesis that "dedifferentia- tion normally is checked through activity of the blastema" (p. 237). Similarly designed experiments using different treatments have given support to this hypothesis. For example, resection of the nerves of the brachial plexus in larval urodeles at the time of amputation inhibited blastema formation and all the formed structures of the limb degenerated and were resorbed. If a blastema was allowed to become established, then the limb denervated, the excessive regres- sion did not occur (Schotté, Butler and Hood, 1941). Thornton and Kraemer (1951) and Thornton (1953) induced extensive regression in larval urodele limbs by means of an injury to the mesodermal tissues involving a break in the skeleton along with denervation. Hhen reinnervation of the limb was permitted regression was halted and in some cases a blastema formed and the limb regenerated. These experiments along with those of Schotté and Butler (1944) demonstrate that nerves are necessary to check the regressive tenden- cies of a limb. In previous studies of the “regressive phase" of regeneration or extensive regression of the larval limb an injury to the larval musculo-skeletal system was involved. The experiments in this investigation were designed to study (1) regression induced by a subdermal homoplastic implant of a piece of humerus in a larval limb without injury to the musculo-skeletal system and with subsequent denervation, and (2) the effect of reinnervation in these limbs. Both histological studies and a statistical analysis of humeral volume changes in the experimental left limbs and the untreated right limbs were made. MATERIALS AND METHODS Eggs of the urodele species used in this study, Ambystoma gggggn and Ambystoma maculatum, were obtained from Mr. Glenn Gentry of Donelson, Tennessee, and raised in the laboratory. As the larvae reached the feeding stage they were transferred to individual plastic dishes (60 x 15 mm) containing aerated tap water. Young larvae were fed brine shrimp daily. They were transferred to clean dishes and fresh aerated tap water within 12 hours after feeding. For older larvae this routine was followed on alternate days. At the beginning of the experiments the larvae of both species were approximately 40 to 50 mm in length. There was an age difference between the species, however; the Ambystoma gpggum_larvae being approximately two months old and the Ambystoma maculatum larvae approximately four and one-half months old. The larvae were kept in an incubator at 20° C i 2° during the course of the experiments. Basic Operations The left limb was always used as the experimental limb which received either a homoplastic implant, was denervated, or both. The right limb was untreated. The operations are described in the fol- lowing paragraphs. 7' Two animals which were to receive homoplastic transplants of pieces of humerus were anesthetized along with a donor animal in MS-222 (1:3000). The animals were then placed in a petri dish lined with a gauze pad moistened with MS-222. Horking under a stereoscopic dissecting microscope, the skin was removed from the upper forelimb of the donor animal with finely sharpened watchmakers' forceps. Muscles and nerves were stripped away and the humerus removed. Any tissue clinging to the humerus was carefully pulled away, the epiphyses removed, and the remaining diaphysis divided into two approximately equal portions. Then through a small skin tear in the lateral aspect of the left upper forelimb of each of the recipi- ents a subcutaneous tunnel was made with the tip of the forceps directed toward the elbow into which the implant was inserted. The other forearm of the donor animal supplied humeral pieces for two additional recipients. The implanted humeral pieces ranged from 0.7 mm to 0.9 mm in length as measured by an ocular grid in serial sections of the limbs which were fixed 24 hours after the trans- plantation. After four days of healing, the implanted left limbs of the experimental groups were denervated. This procedure was accomplished by cutting spinal nerves three, four, and five with finely sharpened watchmakers' forceps with tips bent to function as microscissors. The animals were anesthetized in MS-222 (1:3000) and, working under a stereoscopic dissecting microscope, a small tear was made in the skin of the shoulder region. The muscles were separated to expose the nerves which were cut as close to their entrance into the limb as possible. The nerves were recut every five days, with each succeed- ing cut being made a little closer to the spinal cord. Experiments Preliminary Experiments Homoplastic implants of untreated pieces of humerus and scapula, pieces of humerus treated in various ways, and in one instance, pieces of polyethylene plastic, were used in some preliminary experiments using 25 Ambystoma pppppm_and 69 Ambystoma maculatum larvae. In Ambystoma ppggpm_eight larvae received implants of untreated humeral pieces and 17 received implants of whole skin. The implants were made on day zero, the first denervation on day four, with subsequent denervations at five day intervals, and the larvae were fixed on days 26 and 29 in the first group and day 25 in the second group . Pieces of humerus, which had been alternately frozen and thawed, were used as implants in eight Ambystoma maculatum larvae. This treatment was accomplished by dipping a shell vial containing the humeral pieces into liquid nitrogen for five minutes then thawing for five minutes. This procedure was repeated four times and was done just prior to implantation. The schedule of this experiment was the same as that given above for Ambxstoma pppppp, with fixation on day 25. The remainder of the experiments also used Ambystoma maculatum larvae and the time schedule differs only in that the larvae were fixed on day 32 instead of day 25. Pieces of scapular cartilage were used as the implant in 17 larvae. Humeral pieces that had been lyophylized served as the implant in 13 larvae. Pieces of humerus placed in a small pouch of gauze and held at the surface of boiling water for approximately five minutes were used to implant 14 larvae, and in 16 larvae, pieces of polyethylene plastic constituted the implant. The size of all implants was kept as nearly the same as possible. The length from the axilla to the tip of the longest digit of the left and right limbs was measured at the time of implantation of the left limb and just before fixation. For the measurements, which were made through a stereoscopic dissecting microscope, the orientation of the larvae and their limbs was kept constant. The fixed limbs were paraffin embedded and a small sample (two to four) of limbs from each experiment sectioned and stained for a brief look at the histology. The balance will be processed later for use in further studies. Series I The limbs of the Ambystoma opacum larvae in this series, which were untreated, provide the normal base for the histology and humeral volumes for Series II. The first sample was taken on the same day that the experimental animals in Series II received the homoplastic implants (day zero). This sampling date was recorded as day one, however, for purposes of the computer analysis of the humeral volumes. Additional samples were taken on days four, nine, and fifteen . (N = 19). Series II A piece of humerus was inserted subdermally in the lateral aspect of the upper left forelimb of the Ambystoma pppppm_larvae of this series on day zero. On day four the left limbs were denervated, with subsequent denervations at five day intervals; namely, days nine, fourteen and nineteen. Samples in this series were taken on days one, four, five, eight, eleven, fourteen, seventeen, twenty, and twenty- three. (N = 31). Series III The right limbs of the Ambystoma maculatum larvae of this series serve as the control limbs for the experiments involving this species. The upper left forearm of these larvae received an implant of a piece of humerus on day zero. There was no subsequent treatment and samples were taken on days four, nine, fifteen, and twenty-one. (N = 21). Series IV The larvae in this series (Ambystoma maculatum) were denervated on days zero, five. ten, and fifteen with no other treatment. Samples were taken on days five, ten, fifteen, and twenty. (N = 21). Series V A piece of humerus was inserted subdermally in the upper left forelimb of the Ambystoma maculatum larvae of this series on day zero. Denervations were made on days four, nine, fourteen, and the final one on day nineteen. Nerves were permitted to regrow after day 19. Samples were taken on days 9, l4, 19, 22, 25, 28, 31, 34, and 37. (N = 46). Histological Preparations Whole larvae were fixed in Bouin's fixative. Left and right limbs, including the shoulder region, were removed, dehydrated, embedded, and serially sectioned longitudinally at either six or eight microns. An adaptation of the Goldner-Foote modification of Masson's Trichrome Stain was used for the histological studies. As a check on the presence of nerves in the experimental or control groups which were denervated (Series II, IV, and V), the left and right limbs from one animal at each sampling time were stained for nerves by the Bodian method. Detailed instructions for the pro- cedure used were obtained from Dr. Marcus Singer, Case Western Reserve University, Cleveland, Ohio. Measurements As an indication of skeletal changes occurring in the regressing limb, the volume of the humerus in the left and right limb of each larva was measured. The volume was determined by first 11 projecting the serially sectioned limb on to unruled paper from a constant height and tracing the outline of every fifth section through the humerus. The areas of these drawings were measured with a K & E Compensating Polar Planimeter, Model 62 0022. Each area was measured twice and the sum was divided by two. The magnification of these areas was calculated by projecting a stage micrometer and measuring a one mm2 projection with the planimeter. 20.1 cm2 of projected area equaled one mm2 of actual area. After the projected humeral areas were converted to actual areas these were multiplied by the thickness of the sections to obtain actual humeral volumes. The mean difference of five limbs in which every section was measured as opposed to the measurement of every fifth section was t 5.5%. Statistics The means, standard deviations, and standard errors of the humeral volumes are given in Tables IIand III. Essentially all of the volume measurements fall within two standard errors of estimate (the standard deviation of the vertical deviations from the least squares line). The scatter diagrams (Figures 21 to 30) of the humeral volume measurements include the rectilinear and/or curvilinear line of the regression of humeral volume on time. The fact that the rectilinear line is the only significant one is indicated by the use of a solid line in Figures 26 and 28. Both the solid curvilinear line and the superimposed dashed rectilinear line are significant in Figures 23, 12 25, 29, and 30. In Figures 21, 22, and 24, though the solid curvilinear line is the only significant one, a dotted rectilinear line calculated by the least squares method is included for visual comparison. The dotted least squares rectilinear line on Figure 27 does not deviate significantly from zero. Various formulae using the t_distribution were used to determine significant differences in mean volumes, regression coef- ficients (slopes of rectilinear regression lines), and to test the significance of the polynomial terms. L,-l - RESULTS Preliminary Experiments A preliminary series of experiments was designed to establish the most consistent implant material for producing extensive regres- sion in limbs with uninjured musculo-skeletal systems. Table I summarizes the length changes in the denervated left limbs of these preliminary experiments. Only the skin implant failed to induce some regression. Histological examination of a small sample of limbs (two to four) from each experiment reveals that the skin is the only implant that does not create a chronic inflammatory reaction attracting macrophages and multinucleate giant C6115 (M90). Since an untreated piece of humerus was consistent in inducing regression it was used in all the experiments reported in this paper. All the right limbs from the preliminary series showed growth during the course of the experiments, increasing an average of one millimeter in length. It is planned to continue the analysis of the nature of the implant and the mechanism of induction of regression. 14 Histology The untreated left and right limbs of the Ambystoma pppgflp larvae of Series I and the right limbs of Ambystoma maculatum larvae of Series III provide the frame of reference for the normal histology of the larval limbs during the course of the experiments. Details of the normal histology will not be catalogued here, but since the humerus will be referred to a great deal, its structure and histology will be reviewed. The proximal and distal epiphyses of the humerus consist of spherical heads of cartilage in which the chondroblasts are arranged in concentric layers, and growth regions in which the chondroblasts are arranged in closely packed parallel rows (as viewed in longitudinal section, Figure l). The spherical heads, which are not covered with a perichondrium, increase in size by proliferation of the peripheral layers of chondroblasts and the secretion of intercellular matrix. The humerus increases in length by a proliferation of the cells of the growth region. The periosteum of the diaphysis is attached to the epiphyses just apical to the growth plate cartilage. The diaphysis consists of a hyaline cartilage core surrounded by periosteal bone. The bony collar is thick at the center of the diaphysis, but becomes very thin as it nears the growth regions of the epiphyses. The periosteum is composed of one to two layers of elongate, flattened cells which are indistinguishable from fibrocytes (Figure 2). At times areas of the inner layer are observed to consist of cells with 15 rounded nuclei and slightly basophilic cytoplasm, indicative of growth and/or normal remodelling processes. The following histological descriptions of the left limbs in the various control and experimental groups will consist of the general picture presented by the several limbs at each stage. Series II: Ambystoma opacum-- Implanted and Denervated The homoplastic implantation of a fragment of humerus, with subsequent denervation of the limb, leads to regressive changes. By the end of the first day an acute inflammatory reaction can be seen centered around the injured ends of the implant. Monocytes and/or macrophages are also found in the general vicinity of the implant. The inflammatory reaction continues through the fourth day with increased numbers of macrophages in the implant area. A fibrous capsule with fibroblasts is developing around the implant (Figure 4). During the following seven days the implant undergoes con- siderable destruction with many macrophages infiltrating the cartilaginous matrix through the broken ends of the implant. A number of multinucleate giant cells (MGC) are principally found con- tiguous with the calcified parts of the implant. There is also destruction of muscle in the region of the implant with macrophages seen among the disintegrating muscle fibers. Hallerian degeneration of the severed nerves is evident during this period, but a few axons still take the silver stain. Of particular interest is a gradual 16 increase in the thickness of the periosteum of the host humerus, beginning along the side closest to the implant (Figure 3), then extending around the diaphysis. As the number of layers of the periosteum increases, the cells nearest the diaphysis are no longer fusiform, but rounded with slightly basophilic cytoplasm. By the fourteenth day, the thickened and distended periosteum, along with layers of connective tissue, form a thick capsule around the diaphysis of the humerus. Osteoclasts can be seen in enlarged Howslip's lacunae and on the surface of the periosteal bone of the humerus. In two of four cases some slight damage to the humerus is apparent with the appearance of shallow eroded areas and osteoclastic penetration of the cartilage at the distal end of the diaphyseal bony collar (Figure 7). Most of the muscle of the upper arm has disap- peared with only small amounts remaining in the anterior and medial areas. Much loose connective tissue, with numerous large fibroblasts, fills the remainder of the upper arm. Most of the cartilaginous matrix of the implant has been destroyed, but at least one-half of the calci- fied portion of the implant remains attended by a large number of MGCs. Nerve bundles are almost completely degenerated. During succeeding days, the implant becomes reduced to a small nodule of cartilage and small pieces of bone with the fibrous capsule undergoing dispersion. Macrophages and MGCs remain in conjunction with the implant remnants. Upper arm muscle completely disappears and disintegration proceeds in the muscle of the lower arm. Continuing osteoclastic activity, which is always greater on the side of the 17 implant, is revealed by a reduced thickness and wrinkling of the periosteal bone. Changes in the character of the cartilage matrix in the diaphysis are indicated by a difference in its staining capacity. These degenerative changes have resulted in a slight telescoping of the diaphysis over the growth regions, both proximally and distally (Figure 8). Penetrations through the periosteal bone do not exceed normal and the humerus remains intact. An osteoclastic attack on the radius and ulna begins and destruction seems to proceed more rapidly than in the humerus (Figure 6). Finally, by the twenty-third day most of the muscle in both the upper and lower limb has disappeared and, though the host humerus is still intact, continued osteoclastic resorption has now resulted in destruction of some portions of the diaphyseal bone. The greater concentration of osteoclasts is still on the side of the implant (Figure 5). Series III: —i—’l‘.‘l§ 2:223 -——-'3§l“x‘““'"" It is clear from the description of the limbs of Series II that the homoplastic implantation of a piece of humerus followed by denerva- tion initiates a series of histological disintegrative events in the limb. The question arises as to whether an implant alone would be sufficient to induce similar destructive events in a limb; therefore, in this series of experiments, homoplastic implants of humeral frag- ments were made to limbs with intact innervation. 18 As in the previous series, the implant invokes an inflammatory response in the host limb which is still evident on the fourth day following implantation. At this same time an apparent reaction on the part of the periosteum can already be seen, with an increase in the number of cell layers and a change in the size and form of many of the cells (Figure 16). By the fifteenth day virtually all of the cartilaginous matrix of the implant has been destroyed. Just some small pieces of bone accompanied by many macrophages and MGCs remain (Figure 19). The periosteum has continued to thicken and, in two of four limbs, a faintly stained cartilage matrix can be detected between some of the enlarged cells of the inner layers adjacent to the humerus on the side closest the implant (Figure 17). Most of the muscle in the lateral and posterior aspects of the limb (region of the implant) has disappeared, but there is little indication of disintegration extending to other regions of the limb. By the end of three weeks the muscle tissue has been largely restored to the implant area and a fibrocellular scar marks the site of the implant in most limbs. Periosteal cartilage calluses have been produced along the humerus in all of the limbs (Figure 18). Though occasional osteoclasts are seen along the humerus in this series, there is no apparent osteoclastic activity above that normal to the growth and remodelling process, and, except for the added periosteal cartilage, the humerus is normal. The effects of the implant that can be observed histologically are confined to the 19 upper arm and the histology of the lower arm remains normal throughout the experiment. Series IV: Ambystoma maculatum-- Tl——_ Denervated n y From the foregoing experiment it is concluded that an implant of a piece of humerus without limb denervation does not induce exten— sive tissue regression, and the muscle destruction which occurs in the region of the implant is soon repaired. The question remains, however, as to whether denervation alone may lead to limb regression. To test this possibility is the purpose of Series IV. Figure 20, which is a portion of a longitudinal section through the upper arm twenty days after denervation. is characteristic of the histological effects which result from keeping the limbs in a denervated state. The periosteum increases in thickness slightly and the inner layers of cells show some enlargement, but no periosteal cartilage is formed and neither do osteoclasts appear in large num- bers. The structure of the skeletal elements appears normal through- out the experiment. The principal effect of denervation is upon the muscle tissue. The muscle becomes atrophic and is phagocytized by macrophages. Many macrophages can be seen among the very thin, widely separated muscle fibers in Figure 20. Hallerian degeneration of the severed nerves is evident at the first observation period (five days post-denervation), with the complete degeneration of axons and dispersal of Schwann sheath cells occurring between the fifteenth and twentieth 20 days. These results indicate that denervation alone does not lead to limb regression. Series V: Ambystoma-maculatum-- Implanted and Denervated This series was designed to analyze the effect of nerve ingrowth during the late stages of regression due to implantation of a piece of humerus and subsequent denervation. During the course of the experiment unexpected and striking differences were noted between the Ambystoma ppagpm limbs similarly treated (Series II) and the limbs of Ambystoma maculatum. These differences will be emphasized. Many macrophages have infiltrated the cartilaginous matrix of the implant by the time of the first observations at the ninth day, and there has been considerable destruction of the matrix. A few MGCs are found on the implant, which is surrounded by a fibrous capsule. As in the other humeral implant experiments, muscle destruc- tion begins in the area of the implant, and at this time it is limited to this area. The periosteum has also responded, as in the other experiments involving an implant, with an increase in the number of layers and a few enlarged basophilic cells adjacent to the humerus. Most of these features can be seen in the longitudinal section of an upper arm shown in Figure 9. Hallerian degeneration of the nerves has proceeded down the nerve bundles of the upper arm, but is not as exten- sive in the lower arm, where some axons still take the silver stain. 21 By the fourteenth day the muscle has disappeared in the lateral aspect of the limb (the area of the implant) and destruction is now evident in the anterior and posterior regions. Muscle disintegration in Ambystoma maculatum is occurring more slowly than in Ambystoma ppggpm, By this time, most of the muscle had disappeared from the upper arm in the latter species. The periosteum has increased only slightly in thickness and is about one-half the thickness of the periosteum in Ambystoma ppggpm_at this time. Enlarged basophilic cells and osteoclasts are quite numerous along the humerus in only one case in three. Evidence of demineralization along with some signs of erosion of the periosteal bone is seen in this one case. The matrix of the implant has been essentially destroyed, but some of the calcified part and many MGCs are still present. Nerve degenera- tion has proceeded into the lower arm, but is not as advanced as in Ambystoma m. A nodule of fibrocellular scar tissue has formed in two of three limbs by the nineteenth day (Figure 11). The undulated outline of the humerus indicates a minimal amount of resorption by the osteo- clasts. Though some resorption of the periosteal bone has occurred, many of the basophilic cells along the diaphysis have become active chondroblasts. Varying amounts of cartilaginous matrix have been deposited, particularly on the side of the humerus nearest the site of the implant (Figure 10). Periosteal cartilage calluses were formed in Ambystoma maculatum larvae of the "implant only“ series (Series III). but were not found in the Ambystoma pppgpm_larvae which received the 22 same treatment as this series. There is very little muscle left in the upper arm, and that which remains in the anterior or medial aspects of the limb shows a great deal of breakdown. Muscle disintegration and removal has also begun in the lower arm by this time. No stainable axons remain and the Schwann sheath cells are widely separated in the bundles of the upper arm. Muscle disintegration continues throughout the limb during the next nine days and by the twenty-eighth day the soft tissues of the limb consist almost entirely of loose connective tissue. Fibrocellular scar tissue at the original implant locus, and periosteal cartilage along the humerus are found in a majority of the limbs. In most cases the humerus has sustained only minor resorption shown by shallow eroded areas in the periosteal bone and distal wrinkling of the diaphysis. In a few cases, however, larger areas of the bone along with some of the cartilage core of the humerus have been resorbed (Figure 12). Periosteal cartilage is usually present even in such instances. Rein- nervation of the limbs is beginning by the end of this period. During the succeeding six days the nerve bundles of the upper arm become more compact with increasing numbers of axons. Reinnerva- tion and bundle organization is not as advanced in the lower arm, however. Some degenerative changes continue during this time concur- rent with evidence of some tissue repair. Some muscle disintegration, humeral resorption, as well as osteoclastic removal of the skeletal elements of the lower arm (Figure 15), can be seen in several limbs. 0n the other hand, some muscle repair and reorganization has occurred 23 and rather extensive periosteal cartilage calluses formed by the thirty-fourth day (Figure 13). Limbs on the final day of the experiment (day 37) are characterized by almost normal nerve bundles throughout the upper and lower arms, muscle repair, and periosteal bone deposition on the surface of the extensive cartilage calluses (Figure 14). The limbs appear to be moving toward full recovery from the regressive changes which occurred during the period of implantation and denervation. The right limbs from all series presented a normal histo- logical picture throughout the course of the experiments. Applysis of Humeral Volume Changes Changes in humeral volume through time is one parameter of the limb which can be measured and which would reflect the regressive effects indicated in the histological studies. Biochemical reactions leading to humeral volume changes would also be reflected in such measurements. The quantitative data from the volume measurements of both left and right humeri of all series provide insights into the regression phenomenon not otherwise noted. Regression analysis was used to show the form and quantitative relationship between humeral volume and time. The linear regression reveals the trend through time, i.e., whether there is a significant increase or decrease in volume. The rectilinear relationship is sig- nificant in most cases, but the irregular pattern exhibited by the volumes in some series suggested that the relationship may be 24 curvilinear. Curvilinear regressions were then fitted to determine if the variations represented such a relationship or whether they were due to chance. The significance of all the fitted curves indi- cates that there is a very low probability that the irregular response pattern is due to chance. In the normal limbs of Ambystoma pppgpm_($eries I) humeral volumes show a rather slow rate of increase during the 15-day period (Figures 21 and 22) which coincides with the first 15 days of Series II. The loss in humeral volume in the implanted and denervated left limbs of Ambystoma pppppm_larvae in Series II is statistically significant (Figure 23). Resorption by osteoclasts with the accompany- ing erosion, wrinkling, and telescoping is, no doubt, responsible for some of the volume loss. Osteoclasts were not seen in histologi- cal sections, however, until the fourteenth day of the experiment; yet volumes showed a decrease by the eighth day (three days post- denervation). Volume loss, therefore, may be due to resorption by unidentified mononucleate osteoclasts, and/or possible enzymic or other biochemical reactions in the humerus, in addition to resorption by the identifiable multinucleate osteoclasts. The right limbs in Series II did not receive any treatment, but the humeral volumes failed to show an overall increase during the 23-day experiment (Figure 24). The significance of the curve fitted on these data indicates that the irregularity of the response is not likely to be due to chance. This response is probably the result of the action of various intrinsic and systemic factors which control skeletal growth. 25 The humeral volumes of the right untreated control limbs of the Ambystoma maculatum larvae of Series III show a statistically sig- nificant increase through time (Figure 25). The rate of increase is considerably greater than in the control limbs of Ambystoma pppggp (p 2 0.005). Since the left limbs of these larvae were not denervated and since the histological effects of the cartilage implant in the left limbs were confined to the upper arm of these limbs, it is believed that the humeral volume changes in the right limbs represents the normal growth rate for Ambystoma maculatum. The graph (Figure 25) of the measurements from these right humeri is used, therefore, for com- parisons with the experimental groups of this series. Accretions of periosteal cartilage, which are characteristic in Ambystoma maculatum larvae which receive an implant, were included in the humeral measurements wherever found. The rate of humeral volume increase in the implanted left limbs of Series III (Figure 26) is not significantly different from that in the control right limbs (p < 0.40), but the deposition of periosteal cartilage in many of the limbs on the fifteenth and twenty-first days has produced some increase in the slope of the regression line. It was noted in the histological studies that the structure of the humerus in the left limbs of Series IV remained normal throughout the experiment. Figure 27 reveals, however, that if Ambystoma maculatum larval limbs are kept in a denervated state an increase in humeral volumes with time does not occur. An understanding of the reason for this failure to grow would be of great interest; but, though 26 possible causes could be speculated upon, further research is necessary to provide any concrete information on this question. The slope of the regression line of the right humeral volumes of this series (Figure 28) indicates some volume increase with time. From the ninth to the twenty—second day the humeral volumes of the implanted and denervated left limbs of the Ambystoma maculatum larvae of Series V show an increase (Figure 29). As mentioned pre- viously, periosteal cartilage cuffs or calluses are included in the humeral volume measurements. Such cartilage additions are found on both days 19 and 22 and are, no doubt, responsible for the humeral volume increase during a time when a decrease would normally be expected. In spite of an increase in frequency and size of the periosteal cartilage formations, the humeral volumes do show a decrease between days 22 and 28. The increase in humeral volumes after the twenty—eighth day cor- relates with the beginning of reinnervation at about that time. Histological sections show that humeral resorption continues for some time following the beginning of reinnervation, but a big increase in periosteal cartilage deposition yields a net increase in volume. A layer of periosteal bone on the surface of the cartilage calluses points to a return to a more normal skeletal growth toward the end of the experiment. The right humeral volumes of Series V (Figure 30) present a pattern similar to that of the left humeri, with the coincident loss of volume between the twenty-second and twenty-eighth day. 27 A three-factor factorial analysis of variance was used to test for a poSsible interaction between the implant and denervation in Series III, IV, and V. An interaction, however, is not indicated (Table IV). DISCUSSION Subdermal homoplastic implantation of a piece of humerus without injury to the musculo-skeletal system followed by denerva- tion initiates regression of the limb in the two species used in this study, Ambystoma ppgppm and Ambystoma maculatum. Nevertheless, some striking differences in the response of the two species, as shown in the results of Series II and Series V, were found. These differences will be explored in the following discussion. The rate at which the muscle degenerates and is removed con- stitutes one of the differences. Muscle degeneration always begins in the area of the implant and continues until virtually all of the muscle tissue has disappeared from the limb in both species. The breakdown and removal of muscle in Ambystoma ppagpm, however, is accomplished from five to seven days earlier than in Ambystoma maculatum. Electron microscope studies of the development of muscle and its subsequent breakdown in aneurogenic larvae of these two species are being conducted in the laboratory of C. S. Thornton. These studies have shown that, though muscle develops more slowly in Ambystoma gpgppm_than in Ambystoma maculatum, it breaks down more rapidly (Thornton, personal communication). In Ambystoma pppgpm_the nerve axons and nerve bundles of the severed nerves degenerate more rapidly than in Ambystoma maculatum. 28 n 29 By day 14 there are no stainable axons and the bundles are almost completely degenerated in the former. In Ambystoma maculatum, on the other hand, this stage is not reached until at least the nineteenth day. Gutmann (1968) states that "increased proteolytic activity after denervation occurs later in a muscle with a long peripheral nerve stump than in a muscle with a short nerve stump. This experiment shows that this change is not linked to loss of nerve impulse activity. It depends apparently on chemical systems which regulate degradation of proteins, are supplied by the nerve and are exhausted earlier in a muscle with a short nerve stump" (p. 237). A difference in the length of the nerve stump is not involved in these larvae, but a differential in the rate of Hallerian degeneration would likely have the same effect. The close correlation between the rates of nerve degeneration and muscle degeneration in each of the species would tend to support such an effect. The rate of degeneration is not uniform even within Species. Singer and Steinberg (1972) noted in their studies of Nallerian degeneration in transected nerves in adult Triturus viridescens limbs that the "speed of degeneration varied greatly among the fibers." The histological studies of Series II and Series V reveal that osteoclasts are found along the periosteal bone by the fourteenth day in the two species used. but resorption proceeds more rapidly in Ambystoma ppggpm, There is a significant loss in humeral volume in this species during the experiment with a decrease recorded soon after denervation. This loss in volume may be due to resorption by 30 unidentified mononucleate osteoclasts and/or enzymic or other biochemical regressive reactions in addition to the resorption by the many osteoclasts observed histologically. The activity of such lysosomal hydrolases as acid phosphatase (Schmidt and Weary, 1963; Weiss and Rosenbaum, 1967; Miller and Wolfe, 1968), cathepsins (Deuchar, Weber, and Lehmann, 1957; Wroblewski and Grzybeck, 1961), and B-glucuronidase (Dukiet and Niwelinski, 1960) has been shown to be high during the regressive phase of regenerating amphibian limbs or tails. Of these three enzymes, however, only acid phosphatase has been studied in the regressing limb. In addition to intense acid phosphatase activity in the epidermis, Weiss and Rosenbaum (1967) found some increase in activity within chondrocytes, some areas of the cartilage matrix, and the perichondrial envelope in regressing limbs of Ambystoma maculatum larvae. They suggested that "during limb resorption, increased synthesis of enzymic protein could occur in chondrocytes and could be reflected both in the increased acid phosphomonoesterase activity we visualized and in activity of other hydrolases as well" (p. 211). The humeral volumes in the Ambystoma maculatum larvae of Series V do not show a net loss during the part of the experiment in which the limbs were kept denervated. In fact, there is an increase in volume to day 22 which is followed by a decrease between days 22 and 28. The evidence of humeral resorption that can be detected in the histological sections indicates that it progresses more slowly than in Ambystoma opacum. This correlates with the slower rate of nerve and 31 muscle degeneration already noted for the species. The presence of periosteal cartilage on days 19 and 22 is very likely responsible for the increase in humeral volume during this time. In spite of an increase in frequency and size of the periosteal cartilage formations between days 22 and 28, the humeral volumes show a decrease. This increase in the rate of resorption of the humerus apparently begins shortly after complete nerve degeneration and coincides with the period of most extensive muscle disintegration. An inflammatory reaction accompanies the implantation of a piece of humerus in both Ambystoma pppppm_and Ambystoma maculatum. Bodemer (1959) noted that implants of various tissues into adult Triturus viridescens limbs were usually attended by inflammation, but that it was most pronounced with cartilage and bone implants. Fol- lowing the initial acute inflammatory period macrophages invade the matrix of the implant and multinucleate giant cells appear principally on the calcified portions of the implant. The implant is gradually removed by the phagocytic and resorptive action of these cells. In Ambystoma ppgppm_fragments of bone and matrix can still be seen in the limbs at the end of the experiment on day 23. In Ambystoma maculatum, however, by day 19 only a nodule of fibrocellular scar tissue marks the site of the implant in a majority of the limbs. The induction of periosteal cartilage on the host humerus was found only in Ambystoma maculatum. This occurred, not only in the implanted and denervated limbs of Series V, but also in the "implant only" limbs of Series III. It did not occur in the Ambystoma maculatum 32 larvae which were "denervated only." Periosteal cartilage calluses usually form on skeletal elements which have received a mechanical injury, such as a fracture (Pritchard and Ruzicka, 1950) or an ampu- tation (Schmidt, 1958). Such additions to the host skeleton have also been reported to occur frequently in kidney or liver implanted adult Triturus viridescens limbs (Ruben and Stevens, 1963; Stevens, Ruben, Lockwood, and Rose, 1965; Carlson, 1967; Carlson and Morgan, 1967). Particularly pertinent to this investigation is the report by Stevens et al. (1965) that the appearance of the bone and cartilage additions was coincident with the disappearance of the implant material. In the Ambystoma maculatum larvae of Series III and Series V the appear- ance of periosteal cartilage correlates with the complete disappear- ance of the implant. The fact that the implant was not completely destroyed by the end of the Series II experiment may be the reason that periosteal cartilage was not seen in Ambystoma opacum. This species difference, therefore, may lie in the speed of implant removal rather than in a capability or stimulus response difference. The histology of the untreated right limbs remained normal in all series during the course of these experiments. The humeral volumes of the right limbs of some of the series, however, did not show normal growth. In each of these series the contralateral left limbs had been denervated. These results are in accord with the report of Tweedle (1971) that the transaction of the brachial nerves of one forelimb in the adult newt will bring about a significantly slower rate of regeneration in the contralateral forelimb. This 33 effect is thought to be due to a decrease in the trophic ability of the brachial innervation. The response of the right limb humeral volumes in Ambystoma gpacum and Ambystoma maculatum seem to reflect the difference in the rate of nerve degeneration in the two species. In Ambystoma gpacum (Series II) the transneuronal effect of denerva- tion of the left limb is apparently rapid enough that the right humeral volumes fail to show a normal increase during the course of the experiment. The initial increase in the right humeral volumes in Ambystoma maculatum (Series V) followed by a decrease between days 22 and 28 would correlate with the slower rate of nerve degeneration and slower effect in the contralateral limb. The fact that the right humeral volumes of the Ampystoma maculatum larvae of the "denervation only" series (Series IV) also show an increase up to day 20 is consis- tent with this suggestion. Various species of Ambystoma larvae have been used by a number of investigators for many years, but very few have reported differences among them. Thornton and Tassava (1969) found that grafted limbs of Ambystoma maculatum undergo duplication more readily than do trans- planted Ambystoma opacum or Ambystoma mexicanum larvae. Weis and Weis (1970) reported that though the amputated limbs of both Ambystoma pppppp_and Ambystoma maculatum responded to treatment with nerve growth factor, the "overall ppagpm_rate was much slower." A clear difference in the rate of regeneration of these two species has also been shown, with that of Ambystoma opacum being considerably slower (Weis, Weis, and Graham, 1970). These reports do not appear to be at variance with the results of this investigation. SUMMARY This investigation has demonstrated that subdermal homoplastic implantation of a piece of humerus in a larval Ambystoma limb without injury to the musculo-skeletal system and with sub- sequent denervation of the limb will initiate regression. It has also shown that there are some distinct differences in the response in the two species (Ambystoma gpacum and Ampystoma maculatum) used in the experiments. Nerve and muscle degeneration and humeral resorption occur more rapidly in Ampystoma opacum. The implant is phagocytized and resorbed more rapidly in Ambystoma maculatum and its removal coincides with the appearance of periosteal cartilage deposits on the host humerus. The implant is not completely removed in Ambystoma ppgppm_during the time course of the experiment and periosteal cartilage is not seen in this species. The recovery of regressing limbs following reinnervation provides another demonstration of the importance of nerves for tissue repair and the maintenance of structural integrity. The implantation of a piece of humerus in an Ambystoma maculatum limb which is fully innervated does not lead to regression of the limb, and the localized muscle destruction is soon repaired. Denerva- tion of Ambystoma maculatum limbs without further treatment leads to the atrophy and removal of muscle. It also prevents a normal increase 34 35 in humeral volumes, but there is no evidence of humeral resorption during the period of the experiments. The humeri in the untreated right limbs do not show normal growth when the brachial nerves are transected on the contralateral side of the body, though they appear normal histologically. This effect occurs later in Ambystoma maculatum than in Ambystoma gpacum. 36 ovummpa o m m N o_ mcmpxcpmx_oa wmmpwugwu o m m e eF _mgmszs ump_om mumpwugmu Ass m.v _ o m m m_ enemas; umNLngaoss mmmFFuLmu o _ m m_ up ampsaaom emanates: mmmpwpgmo ngms:; 1. Ass m.v F e o m m umzmcpumNmmgm Ezpm_:ume .< Ase _V F m_ o 0 NF . seem mmmFFuemu .. o o o w m Pegasus nmpmmepc: Enummo .< spasm; cumcmb :_ as F l as m. ages so as _ m_me_c< pcmpasH we quh mmvumam cw mmmmeucm wmcmcu oz spasm; cw mmmmeumo mo consaz GEO mm; mpcmewgmaxm mg“ eo acmccmmmn ms“ pm zumcm— new, mmmem>m ugh .mpcmewemaxm xeocwevpmca mgu co mnewp peep umum>gmcmn mg» cw mmmcmzu cumcmF mg“ mo acmesam H m4m

60 .wom .Qom €602 awom acum EGG: XMQ ms:_o> Foamsoz pomwm mE:_o> FoLmE:I pmmo we:_o> poomEoI psmwm mso_o> ~ocmEoI pew; oooooaooo oooo ooooo .ompo>gmcwo new uopcoPoEH new; pemouuHH mmwgmm mpoepcoo venomgpcauna mowemm .saumam meoumxae< cm mmszpo> pong=g msu no» mgogem egoccopm coo .cowuo_>wu ugmucoum .mcomz HH u4m mE=Fo> mE=Fo> mas—o> mEo_o> asapo> FoLoEo: ooomo Poemsoz pmmo _oooEo: ooooo magmas: new; oooooo: “comm Poemsoz ammo oooootoeo oooo ooomo .Am_ one co compo> ugmcmo umoov umpo>emcmo ooo ooooooooo memo oooo--> moooom vmoomepca osmo ooooo .opoo oooo>ooooo mooo oooo-->o momoom umpowgpcz mo_o ooomo .oooo oooooposo memo oooo--ooo mooeom .Ezpomaume msoumxae< cw mmsopo> Pegasus one so» mgogem vgoucoum woo .mcomuow>mu cgoucoum .mcomz HHH m4mooooo x pooposo oom. oopoo. oopoo. _ oomoosooooo x ooo_oso omo. omooo. oomoo. o oemo x oomoo>aoooo o_o. ooooo. omooo. o memo x oomposo mooo. omo_o. moooo. o memo mooo. ooooo. ooooo. F oomoo>ooooo ooo. ooooo. ooooo. _ oooooso m mo mooooom mmeooom eooomgm comuomeo> mo mogzom mucmummmcmwm com: mo mssm mo mmmgmmo .> ooo .oo .Hoo momaom co oomoooaooom eo mmmz_ooo Fowgopuoo eopuomummgcu esp Low mpoop mucooxo> >H m4m 0 cl 0 o 0.05 Jr 24 21 15 12 DAY 10 18 FIGURE 23 56 Figure 24.--Series II. Ambystoma opacum. Humeral volumes of the right limbs (N = 31). These limbs were untreated but the contra- lateral left limbs were both implanted and denervated. The quintic regression coefficient of the curvilinear line is significant to the 0.001 level and the correlation coefficient (r) of the line is 0.719. The humeral volumes failed to show an overall increase during the 23-day experiment, as indicated by the dotted least squares recti- linear line. The irregular response may be due to the continuing, though hampered, effects of intrinsic and systemic factors affecting skeletal growth. 57 .11.} 0.35- 0.30" 0.25- 0 mozzv moomzox do ozooo > 0.10- 0.05” 21 18 15 12 DAY FIGURE 24 58 Figure 25.--Series III. Ampystoma maculatum. Humeral volumes of the right untreated control limbs (N = 16). Both the recti- linear and quadratic regression coefficients are signifi- cant (p < 0.0005 and p = 0.004). The correlation coefficient (r) of the curvilinear line is 0.921. The rate of humeral volume increase is considerably greater than in Ampystoma opacum. 59 1.4.64): 0.35 L 0.30 - 0.25 - b 0 2 o 0.15 r 0 Ammo: mass: no 232, 0.10 - 0.05 - 0‘ 24 21 18 15 12 DAY FIGURE 25 60 Figure 26.--Series III. Ambystoma maculatum. Humeral volumes of the implanted left limbs (N = 21). A homoplastic implant of a humeral fragment was inserted subdermally in the upper left forelimb of each larva on day zero. There was no further treatment. The rectilinear regression coefficient is significant to the 0.0005 level and the correlation coeffi- cient (r) is 0.841. There is no significant difference (p < 0.40) between the slope of this line and that of the rectilinear regression line of the contralateral (control) right limbs of Figure 25 (i.e., the rates of volume increase are comparable). 61 0.35L 0.30 ' 0.25 l b 0 2 00 0.15 * Amzzv moomzo: mo ozoooo 0.10 b 0.05 . 0d 24 21 18 15 12 DAY FIGURE 26 Mum. 62 Figure 27.--Series IV. Ambystoma maculatum. Humeral volumes of the denervated left limbs (N = 21)T' These limbs were denervated on days zero, five, ten, and fifteen. The dotted least squares line does not deviate significantly from zero (i.e., there is no significant increase or decrease in humeral volumes with time). 63 .r4. 2 l a I] . 2 av av. “xv . _ .. _ J8 - 1| — j - I av .u‘—v . m” _ . _ . _ _ .m _ a vuv aVv - . . o. _ .. _ 1 _ — 16 v vuvvv . l 13 1 J :0 h p b n u p p 5 a mo no m w w m 0 0 O 0. 0. O. 0. $25 852:: o £32 DAY FIGURE 27 64 Figure 28.--Series IV. Ambystoma maculatum. Humeral volumes of the right limbs (N = 17). These limbs were untreated, but the contralateral left limbs were kept denervated. The recti- linear re ression coefficient just approaches significance (p==0.0651 and the correlation coefficient (r) is 0.457. The slope of this line is not significantly different (p z 0.10) from the least squares line of the denervated contralateral left limbs. Paradoxically, neither is it significantly different (p < 0.25) from the slope of the rectilinear regression line of the humeral volumes of the control limbs (Figure 25). 65 VVVV VvV 21 15 12 DAY finu 0.35 - 0.30 l 0.25 - . . 0 5 2 1| 0 0.10 ' o 0 0 A15 86%: no ozo._o> 0.05 - 0.. 18 FIGURE 28 66 .x_:o compoNWFo:m_> some“ Foemcmm Low omoopoom mm w:m_ mmgoocm pmoop Lomcmpmpumg moo .mmm.o mm ALV pcmwuwwmmoo comuopmeeou mop woo Fm>mp Noo.o on» ou poooowmcmmm mm pcmmummwwou commmmemmg umoou mop .m_ moo memo zoemwc ou omuumELmo mom: mm>gmc mop .cmmumcoo ooo .cmmuezow .m:_c .eoow moon :0 umuo>emcmu can» mo: new. mo» .oemo moo co o>eop zoom mo os__woom pomp smog: mop cw oppoeemnoom ompommCm mo: magmas; mo momma a mo acoFQEw umpmorooeo; < .Aoe u zv moEm_ pomp Popcwspemoxm mg“ mo mmE:_o> Poemsoz .EopoFouoE weepmxoe< .> momemmln.mm mcommm mm $36: .33 mm mm om kw vm ~N m~ m~ N~ m m m o 67 kw om.o mm.c Qv.o v.o 68 .xpco :o_p loompoomm> moms“ Poemcmm so; umo:_u:w mm mcm— mmemoom “mom, Lomompmpume mob .~m~.o mo ALV pcmmumwwmou comumFmeLou mow moo Pm>m~ No.o mg» op ucmmmwmommm mm pcmmumm -wmom commmmgmmg umoou mg» .mp mom :0 meme momma comuo>emcmn chmo mgu mumpo>emcmu oco umpco—oem coon mgmz moEPF pomp _ogmpmpmgucom ma“ non .omuomemca mam: moemp mmmch .APN u zv mosmp “some mom mo mmsogo> poems:z .Eopopouoe oeoumxoe< .> mmmgmmuu.om meommm (59 Pm om mmauHm >