A COMPARATIVE STUDY OF THE MECHANISM OF WOUND HEALING.* LEO LOEB. (From the Department of Comparative Pathology of Washington University School of Medicine and the Department of Pathology of St. Louis University,') On the basis of our earlier studies on wound healing2 and of later comparative and quantitative studies, which have been carried out in our laboratory by Addison, Spain, Akaiwa and the writer' in the course of the last ten years, we intend to separate the various factors which are concerned in this process and to analyze each one separately. Wound healing is a composite process in which activities of various kinds enter and interact with each other, each showing definite quantitative variations. In order to obtain a full under- standing, it is not sufficient to follow the process as a whole. The mere observation with the naked eye does not allow a separation of the various factors. In studying, furthermore, the process in various species we will find quantitative differences in each species, and in com- paring the data thus obtained we may be able to separate the essential factors underlying this process from varying condi- tions of a more accidental nature. While at present this analysis is still far from being complete, a survey of our find- ings will enable us to gain a provisional orientation which may serve as the starting point for further investigations. I. Cell movements. - If we compare the cell movements in pigeon, rat and guinea-pig, we notice in each case certain irregularities which depend upon variable factors which enter into the individual experiments. The rapidity of the move- ment depends among other factors upon the character of the wound. At a corresponding period of time the epithelial tongues are much longer in smooth and shallow wounds than *Received for publication, August6, I9I9. 247 948 4LOEB. in uneven wounds, in which the unevenness of the wound or the firm adherence of the scab causes an obstacle to the move- ments of the cells. This has been brought out with particular clearness in the measurements of Akaiwa. It is furthermore probable that other factors, as, for instance, the situation of the wound, plays a part also. Perhaps the epidermis at the tip of the ear would move more quickly than in the center of the ear. Whether the age, general state of health and nour- ishment of the animal plays a part needs further investiga- tion. Notwithstanding these irregularities, which are as yet not fully understood, we may on the basis of our present data now draw certain conclusions. Thus it appears that in the pigeon the response of the epidermis to the wound stimulus is slowest during the greater part of wound healing, until between the ninth and eleventh day,justbefore the closure of the wound, a marked elongation of the epithelial tongues accompanied by a marked contraction of the wound takes place. Thus in the pigeon the closure of the wound takes place more slowly than in the guinea-pig and rat. The epi- thelial movements throughout are quickest in the guinea-pig; the results in this respect are similar in both series of experi- ments. In the rat, the results are somewhat different in both of our series. In the first, the movement is decidedly slower than in the guinea-pig, but faster, than in the pigeon; while in the second series the rapidity of movement approaches closely that of the guinea-pig. Accordingly we find in the first series the time of closure in the rat somewhat later than in the guinea-pig, while in the second series the time of closure is similar in both species. Altogether we may state that the motor response to the wound is most energetic in the guinea- pig and weakest in the pigeon. In the rat it stands between that of the pigeon and guinea-pig, but it may approach that of the guinea-pig. A similar order we find in the activities in the normal skin of the pigeon, guinea-pig and rat as ex- pressed in number of mitoses, thickness of stratum germina- tivum and cell size. The normal epidermis of the guinea-pig is most active; the noriral epidermis of the pigeon is least active; the epidermis of the normal rat stands between that MECHANISM OF WOUND HEALING. 249 of the pigeon and guinea-pig, but approaches more closely that of the pigeon. On the other hand, while in regard to the normal epidermis the rat is much more similar to the pigeon than to the guinea-pig, under the influence of the wound stimulus the rat epidermis may develop an energy of migra- tion which almost equals that of the guinea-pig. The re- sponse to the stimulus is therefore relatively much greater in the rat than in the pigeon. We may then conclude that a certain relation exists between the motor response to the wound stimulus and the prolifera- tive energy in the normal epidermis. The greater the latter, the greater the motor reaction; but in addition there seems to exist a reserve power which during wound healing makes it possible for the normally more sluggish epidermis to display an energy almost equal to that of the normally more active epidermis. 2. If we compare the rate of growth of the epithelial tongue at different periods, we notice again certain irregularities which make impossible, at the present time, a definite formu- lation of the law underlying the process. These irregularities are especially marked in the case of the pigeon. Itis probable that secondary variable factors complicate the conditions. But, making allowance for these complications, we may ten- tatively conclude that on the whole neither a marked retarda- tion nor a marked acceleration in the growth of the tongue takes place during the later as compared with the earlier periods of wound healing, but that an approximate propor- tionality exists between the increase in the length of the tongue and the duration of time of wound healing, and that this proportionality is independent of the phase of wound healing. Thus we find in the second guinea-pig series at five days the tongue 2.55 times as long as at two days, which satisfies the theoretical demands. In the same series after seven days the tongue is 3.3 times as long as after two days. We find, then, the following relation: Length of tongue at five days is to the length of tongue at seven days as 5.I:7. Theoretically we ought to expect, instead, 5:7 The agree- ment is satisfactory. 250 LOEB. In the first guinea-pig series the tongue is at five days 3.I times as long as at two days. The increase in this case is somewhat greater than should be expected, if direct propor- tionality existed. After seven days the tongue in the first guinea-pig is 4.4 times as long as after five days. The rela- tion is as follows: The length of the tongue at five days is to the length of the tongue at seven days as 4.97:7. Theo- retically it ought to be 5:7. The difference is not verygreat. In the first rat series we find after five days the tongue 2.6 times as long as after two days. In this case the agreement is good. After seven days the tongue is about 3.8 times as long as after two days. We find, therefore, the relation of 4.83:7 between the length of tongue at five and at seven days. Theoretically it ought to be 5:7. In the second rat series the tongue is about I.9 times as long after five days as after two days. In this case the growth is less than should be expected under the assumption of proportionality. At seven days the length in the second rat series is 2.5 times as great as at five days. The relation between the length at five and seven days is therefore 5.2:7, instead of 5:7 as theoretically required. In a provisional way we may then conclude that, while variable factors complicate the result, on the whole a direct proportionality seems to exist between the length of tongue and the duration of wound healing. Now, we must consider that several factors enter into the formation of the tongue, namely, (a) the ameeboid migration of the epithelial cells; (b) the increase in the size of the epi- thelial cells, and (c) the increase in the number of cells as a result of mitotic proliferation. While the mitotic multiplica- tion leads primarily to an increase in the number of cell rows, it contributes also to an extension of the tongue in a longitu- dinal direction. Both factors (b) and (c) contribute therefore to the growth of the tongue to some extent, yet the migration of the epithelial cells proper is in all probability the essential element in the enlargement of the tongue. If, as a result of the co6peration of all three factors, at best proportionality between length oftongueandduration of itsgrowth isattained, MECHANISM OF WOUND HEALING. 251 it is very probable that the energy of amoeboid activity shows on the whole a relative decrease in the later periods of wound healing just preceding the closure. That, however, this does not apply generally may be concluded from the rapid growth of the tongue just preceding closure in the case of the pigeon. 3. In the rat as well as in the guinea-pig, the rapidity with which the tongue enlarges depends upon the size of thewound. The motor response of the epidermis to the wound stimulus is in both species greater in the case of the larger wound. In the larger wound, the area of the defect as compared to the circumference of living epithelium is greater than in the smaller wound, where the circumference is relatively greater. It seems, therefore, that this relationship determines in some way the rapidity of movement of the epithelial cells. But, while in the case of the guinea-pig the larger wound closes earlier than the smaller, in the rat the greater rapidity with which the tongue enlarges in the large wound is not sufficient to lead to an earlier closure of the larger wound. In the rat, the shorter wound closes at an earlier date notwithstanding the more rapid elongation of the tongue in the larger wound. This difference between the effect of the size of the wound in the guinea-pig and rat depends upon the cooperation of a second factor, namely, the contraction of the wound. The intensity of contraction differs in the rat and in the guinea- pig, and being more marked in the guinea-pig, this factor is added to the greater rapidity of amceboid movement in the larger wound and leads to a more rapid closure of the larger wound in the case of the guinea-pig. 4. Contraction and retraction of the wound. -A contrac- tion of the wound occurs in all three species which we have examined so far; but it differs in strength in different species, and it differs also in the same species according to the length and probably also the depth and situation of the wound. There is one character in common to all the experiments in the various species: contraction does not set in to any con- siderable extent within the first few days after the defect has been made. It is not usually noticeable in the first four days 252 LOEB. of wound healing. In many cases it seems to become very marked at the period of wound closure and at the period directly following wound closure. This is especially notice- able in the second guinea-pig series. Here the closure occurs earlier in the four-millimeter series than in the two-millimeter series. The contraction sets in in accordance with the differ- ence in closure; it occurs earlier in the four-millimeter series, and the entire contraction throughout the whole period of wound healing is therefore much greater in the four-milli- meter series. In the pigeon also the principal contraction occurs between the ninth and eleventh day, which is the time of closure. In the rat, the main contraction takes place either at the time of closure or in the period following closure. In the first guinea-pig series contraction is also most marked in the period following closure. But in the pigeon and in the rat (second series) contraction during the essential processes of wound healing plays only a very insignificant part in our experiments. In the second rat series, contraction is not only very late but also quantitatively weak, and it merely suffices to overcome the initial retraction which took place, as is shown by Akaiwa. In the pigeon, as well as in the rat, not only does contrac- tion not occur in the early stages of wound healing, but on the contrary a retraction takes place which is due to the natural tension in the skin at certain places and to the relative looseness of the attachment of the skin proper to- the under- lying tissue. We see, tnen, that contraction is relatively most marked in the guinea-pig and least in the pigeon and rat. It occurs mainly in the later periods of wound healing, and is probably due to the activity of the connective tissue-which proliferates in the floor of the wound. In those cases in which an active contraction coincides with the time of closure, the active contraction which takes place at this time is probably the primary factor which makes conditions favorable for the occurrence of closure at this time. The second rat series is of special interest because here Dr. Akaiwa was enabled to study the factors in wound healing almost uncomplicated by MECHANISM OF WOUND HEALING. 253 contraction. In the guinea-pig, contraction showed the maximum effect; it was particularly strong in the four-milli- meter wounds and associated with the rapid outgrowth of the epithelium it enabled the larger wound to close more rapidly than the smaller wound. The effect of the size of the wound on the amount of con- traction of the granulations has been previously studied by A. Carrel.4 He finds that contraction is the more rapid the larger thewound. It is most rapid in theearlystages of contraction, and gradually becomes slower. It becomes about zero when the edges havereached a distance of about ten or fifteen milli- meters. In our experiments we find contraction even in wounds as small as four and two millimeters in diameter. Carrel believes that the wandering of the- epithelium sets in only after the period of granulous retraction has come to an end. We find, on the contrary, that the amceboid wander- ing of the epithelium is the first response of the tissue to the wound stimulus and that the contraction sets in only at a later period. Carrel states that the rate of epidermization is inversely proportional to the dimensions of the wound. We find that within the range of wounds examined by us the rapidity of outgrowth is greater in the larger than in the smaller wounds. A quantitative determination of the various part processes which constitute wound healing is only possible through microscopic determinations. 5. Closure of the wound. In regard to the time of closure of the-wounds we find marked variations in the different species, and within the same species under different condi- tions. In the pigeon, in which all activities show the least intensity, the wound closes later than in the other species, namely, between the ninth and eleventh day. It closes first -in the four-millimeter wound of the guinea-pig, but it may close almost as early in the two-millimeter wound of the rat, although in other cases it closes somewhat later in these. On -the whole, the time of closure is slightly later in the rat than in'the guinea-pig. Thus we find the fourth and eleventh days to be the extremes in the time of closure, and we also 254 LOEB. find a certain parallelism between the energy in the cell activi- ties in the normal and regenerating skin in the different species and the time of closure. In the shallow wounds the closure takes place earlier than in the deeper wounds; this explains the difference between the time of closure in the razor and in the trocar series, in the rat in Dr. Akaiwa's experiments. 6. Mitotic Proliferation. - In guinea-pig, rat and pigeon the mitotic proliferation of the epithelium shows definite similarities. Everywhere we find in the first period, which usually extends over approximately four days, only very rare mitoses in the new epithelium (the tongue); indeed, here the mitoses are more rare than in the normal skin. In the old epithelium adjoining the wound, on the other hand, we find a considerable increase as early as two days after operation. However, under certain conditions the new epithelium in the tongue may show a marked increase as early as three and a half days after operation, namely, in cases in which an early closure of the wound occurs. This is the case in the four- millimeter series of the guinea-pig. Shortly before closure a marked rise takes place in the new epithelium. Usually the closure of the wound occurs at a date later than three and a half days after operation. We invariably find a rise in the number of mitoses in the new epithelium, which reaches a maximum just preceding or at the time of closure and a de- cided fall directly following the closure. In one series (the two-millimeter series in the rat) the fall in the number of mitoses in the new epithelium was apparently somewhat de-. layed after closure. From the neighborhood of the wound the increase in mitotic activity extends into the old epithelium with gradually decreasing intensity. The greater the increase near the wound, the farther back the increase extends. The effect of the wound on the mitotic proliferation in the adjoin- ing epithelium appears to be cumulative; it increases more and more until the closure takes place. The curve of mitotic proliferation in the old epithelium varies somewhat in the different series. In the majority of cases the proliferation at MECHANISM OF WOUND HEALING. 255 this place is most marked after two days, and then gradually decreases, while in the new epithelium it still increases during this period; so that at the time ofclosure or directly following it the mitotic proliferation is usually more marked in the new- than in the old epithelium, while at an earlier period the con- dition was reversed. This holds good if we compare the number of mitoses in one half of the total area of the new epithelium with the number of mitoses in a unit area of the old epithelium. But, even if we compare the mitoses in a unit area of the new with the mitoses in a unit area of the old epithelium, the gradual increase in the mitotic proliferation in the new epithelium comes out quite clearly; and this in- crease continues until the maxima in both the old and new epithelium become at least equal. But in the second rat series a gradual increase took place in the mitotic proliferation, also in the old epithelium, and the proliferation reached a maximum just previous to the time of closure, at a somewhat earlier period than in the new epithe- lium; afterthemaximumhadbeenreached thenumbershowed a continuous fall. In a tentative way, we may explain these facts as follows: It is primarily the defect which causes the mitotic cell proliferation. Wherever the defect adjoins directly the epithelium or is separated from it only by a short piece of tissue, the cell proliferation is marked provided the conditions are otherwise favorable for cell proliferation. Such favorable conditions depend upon the character of the circulation and the proliferation of the underlying connective tissue. Both of these factors are deficient in the wound in the first few days of wound healing. The effect of the wound is cumulative. Therefore the mitotic activity increases in the new epithelium continuously until the time when the wound is closed. The cumulative effect of the defect would also extend to the old epithelium were it not that this effect is counteracted by the constantly growing distance between defect and the margin of the old epithelium. Usually, it- seems,thelatter factor overbalances the former, and then the maximum is reached in the old epithelium in the early periods- of wound healing. In the second rat series, however, the- 256 LOEB. second factor seems to have been relatively stronger and here a maximum was reached in the old epithelium.only a little earlier than in the new. However, even after the closure has taken place and thus the defect has been eliminated, the imitotic proliferation in the new and old epithelium does not abruptly return to its normal level, but, while a decided fall takes place, the cell multiplication continues with a more than ordinary-energy for some time. We must either assume that the stimulus previously exerted by the defect has an after-effect outlasting the defect, or that there exist other factors- causing cell proliferation in addition to the defect. The abnormal condition of the soil on which the epithelium, especially the new epithelium, is resting even after the closure of the wound may represent such an additional factor. If we consider the whole course of the wound healing, we find the greatest absolute cell proliferation in the new epi- thelium. But if we compare the cell proliferation in the unit areas in the old and new epithelium the maximum occurs in the first area of the old epithelium in the majority of experiments. 7. If we compare the mitotic proliferation in the two-milli- meter and four-millimeter series we find in the rat a greater absolute and relative number of mitoses in the four-millimeter series. The larger wound causes a.greater rise in cell multi- plication than the smaller wound.- It is different in the guinea-pig; here the total number of mitoses is considerably greater in the two-millimeter series. This difference between the guinea-pig and the rat is evidently due to the difference in the time of wound closure in guinea-pig and rat. In the rat, the four-millimeter wounds close later than the two-milli- meter wounds, while in the guinea-pig the four-millimeter wounds close considerably earlier. Now the time of closure determines the mitotic activity. With the closure there is everywhere a great reduction in the number of mitoses, and thus it comes about that in the four-millimeter series in the guinea-pig the increase in mitotic proliferation extends over a much shorter period than in the two-millimeter series. The