ISSN 1866-8836
Клеточная терапия и трансплантация

The lymphocyte as a stem cell, common to different blood elements in embryonic development and during the post-fetal life of mammals

By Alexander A. Maximow

doi 10.3205/ctt-2009-en-000032.01


Lecture with a demonstration, held at a special meeting of the Berlin Hematological Society on 1 June 1909.

Translated from: Maximow A, Der Lymphozyt als gemeinsame Stammzelle der verschiedenen Blutelemente in der embryonalen Entwicklung und im postfetalen Leben der Säugetiere. (Demonstrationsvortrag, gehalten in der ausserordentlichen Sitzung der Berliner Hämatologischen Gesellschaft am 1. Juni 1909), Folia Haematologica, 8, 1909, 125-134.
Translated by Claudia Koltzenburg, Alexey Chukhlovin, Athanasius Anagnostou, and Carol Stocking.

Although every attempt is made to ensure precision in the translation into English of the material in these articles, we do not guarantee nor imply their absolute accuracy.

The first blood cells are known to emerge from the so-called blood islands, i.e., irregularly shaped aggregates of peripheral mesenchymal mesoblasts, forming a network of cells in the area opaca. The peripheral cells of the blood islands flatten out, becoming endothelial cells, whereas the inner cells round up, becoming the primary blood cells that are released into the liquid blood plasma. I have now found that these primitive blood cells (which is what I call them), contrary to what would be commonly expected, are not erythroblasts but completely undifferentiated elements with a round bright nucleus and narrow basophilic cytoplasm. These are neither red nor white blood corpuscles. They could be considered white blood corpuscles since they occasionally, especially in the chicken, appear ameboid, and seem very similar to large lymphocytes. They proliferate extensively, although the increasing numbers during the early stages may in part due to the detachment of endothelial cells in the primitive vessels.

After some time, one observes how these primitive blood cells differentiate into two kinds of cells. One type – which makes up the majority – produces hemoglobin in their cytoplasm, and thus become the so-called primitive erythroblasts. These are large and rapidly proliferating cells that, ultimately, grow to rather hemoglobin-rich cells with relatively small nuclei. They serve the organism for a long period of time, but die out gradually and are replaced by the definitive erythroblasts and erythrocytes.

Another fraction of the primitive blood cells remains hemoglobin-free. These cells now possess a large bright nucleus with nucleoli, and a thin, ameboid, strongly basophilic rim of cytoplasm. In histological terms, they resemble large lymphocytes. These are the first embryonic leukocytes that first appear as lymphocytes.

In what follows, we will see how these intravascular lymphocytes become the starting point of erythropoiesis in the area vasculosa. Through "heteroplastic proliferation" (differentiation) they produce secondary erythroblasts; first to appear are megaloblasts that are variable in size and have bright nuclei; later generations of these cells increasingly resemble the normoblasts, and, finally within the vessels of area vasculosa, a mixed population arises composed of primitive erythroblasts that are rich in hemoglobin, basophilic lymphocytes, and large numbers of intensely proliferating megaloblasts and normoblasts growing in clumps.

However, despite the fact that these lymphocytes produce erythroblasts, they should not be considered erythroblasts themselves; as, in addition to the production of hemoglobin-containing cells, they also give rise to megakaryocytes and diverse other elements in the yolk sack, which have nothing to do with the red blood corpuscles.

Such secondary erythroblasts are quite distinct to the primitive erythroblasts, differing from them by their smaller size and, in contrast to normoblasts, by their smaller and darker nucleus. Ultimately, this nucleus becomes pyknotic and leaves the cell in a degenerative state.

Here, I avoid consciously the issue of erythroblast enucleation, since the current discussions do not correspond to the factual material at hand. In my opinion, all known facts speak in favor and not against nucleus expulsion, whereas no direct proof can be presented for its intra-cellular disappearance – at least in normal hematopoiesis. Although pale shadows of nuclei are often visible, as for example in hemoglobin-rich primitive erythroblasts, this is but due to the fact that the basic dye cannot penetrate a thick erythrocyte envelope. However, as soon as the nucleus leaves the cell, it immediately acquires dark color. 

The vessel network of area vasculosa is, therefore, the first blood-forming organ in the mammalian embryo. It is from here that lymphocytes, erythrocytes, and megakaryocytes emerge; however, granulocytes are never produced here.

In the course of events in the extra-embryonic areas outlined above, the first freely migrating cells appear in the mesenchyme of an organism, which is initially entirely free of wandering cells. This occurs at a very early stage, e.g. in rabbit or guinea pig embryos with a length of 4–5 mm. They emerge by rounding off and separating from the common undifferentiated, branched mesenchymal cells.

Generally, the first migrating cells seem to be similar to lymphocytes, i.e., they look like lymphocytes found in the area vasculosa vessels. Immediately upon their first emergence and even more so at somewhat later stages, migrating cells of other types can be found all over the mesenchyme, e.g. cells with pale, ameboid and often vacuolized plasma, and small, irregularly folded light or dark nuclei. Hence, the migrating cells in mesenchyme are manifold and highly polymorphic, with many different transitional forms between them. Such histological differences do not have any particular significance, since the basic feature of these cells, i.e., their progressive developmental ability, always remains unchanged and all migrating cells of mesenchyme are of equal value.

It is of the highest importance to note that the wandering cells in the mesenchyme are also identical—morphologically and physiologically—to the lymphocytes circulating in vessels and in blood in the area vasculosa. Both are free, ameboid, undifferentiated mesenchymal cells, even though their appearance can change significantly, depending on their environmental conditions.

Similar to those observed in the vessels of the area vasculosa, lymphocytes in the mesenchyme can be observed producing erythroblasts and megakaryocytes. However, in the mesenchyme, the wandering cells or lymphocytes can differentiate even further—with some of them differentiating into granulated myelocytes and leukocytes. These often develop into small, abortive leukocytes with polymorphous nuclei that are scattered in the tissue and very quickly subject to degeneration or phagocytosis.

There is one more fact that proves that mesenchymal migrating cells are identical with the lymphocytes of the area vasculosa. That is the fact that the endothelium of certain vessels (in particular the aortic endothelium) proliferates intensely at certain stages and in particular areas, whereby large clumps of cells emerge that project into the lumen, are then washed away by blood, and, finally are incorporated into the circulating blood. Here, they cannot be distinguished from the lymphocytes that originate from the area vasculosa.

At this point, I also wish to make a short comment on the circulating blood. Contrary to common opinion, it is a fact that white blood corpuscles – i.e., large lymphocytes – already exist in the blood from its earliest stages of development, and in significant quantities. In the blood-producing vessel network of the area vasculosa, most of the lymphocytes are held back as producers of erythroblasts, yet some of them enter the blood circulation.

The liver is the second blood-producing organ in a mammalian embryo. As is well known, erythrocytes, megakaryocytes, and granulocytes are produced extravascularly, between the liver cells. The question arises: where is the starting point of this hematopoiesis to be found? An investigation of appropriate stages shows that migrating cells initially appear between the liver cells and the vessel endothelium. These look exactly like the migrating cells in the rest of mesenchyme of the body. Some of these are similar to lymphocytes; some are pale and have small nuclei. If we go back further, to examine those stages in which cords of liver cells grow into the mesenchyme of the septum transversum, we will be convinced that the migrating cells are derived from this mesenchyme. The mesenchymal cells emerge either as such, or already as migrating cells, from between the liver cells and endothelial walls of growing vessels. Here, they remain unchanged for a short while, but soon unveil an amazing capability of development. At first, the migrating cells transform primarily into proliferating large "lymphocytes," which produce large amounts of erythroblasts and erythrocytes. A smaller number of the cells transform into granulocytes and megakaryocytes. Hence, in the liver, too, we note the same non-differentiated migrating mesenchymal cell, the lymphocyte, the starting point of hematopoiesis. The liver cell environment provides rather favorable conditions for the lymphocyte, in which it proliferates and produces a large variety of blood elements.

Finally, the third, final blood-producing organ, which takes over for the liver, is the bone marrow. I have also followed its origin from the very beginning. Here again, we observe that in the young undifferentiated mesenchyme, which invades the cartilage and resorbs it, some of the fixed cells become migrating cells that, at first, appear highly polymorphic. In this case, too, almost all of them ultimately achieve the appearance of typical lymphocytes, and, again, these cells become the starting point of blood formation, which proceeds much as it does in the liver as an extravascular process. However, in contrast to the liver, it continues here lifelong. Here, the lymphocytes, by means of differentiating proliferation, produce erythroblasts, megakaryocytes, and granulocytes of three various types, too. However, some of them produce their own sort, typical agranular lymphocytes, i.e., they function not only as myeloblasts, but also as lymphoblasts at the same time.

Until now, in terms of blood formation, we have actually only observed the emergence of the so-called myeloid tissue: erythrocytes, megakaryocytes, and granulocytes. One might propose, and Schridde actually says so, that the cells I have thus far called lymphocytes, are not lymphocytes at all, but myeloblasts. Indeed, the cellular elements I observed are histologically identical with the lymphocytes, but one might argue that only those cells should be described as lymphocytes or lymphoblasts that can be shown to produce typical small lymphocytes. According to Schridde, these cells, i.e., the true lymphoblasts, should appear only much later and look completely different.

While the single migrating cells described above might be more or less similar to the typical small lymphocytes from very early stages onwards, it is also true to say that the latter arise only relatively late in the organism in large quantities. In the bone marrow, we observe quite frequently, and later, more commonly, numerous progeny of the proliferating large lymphocytes that acquire an appropriate appearance. However, especially large quantities of small lymphocytes emerge in the thymus. So, here, I must also say a few things about this organ. Improved knowledge of thymic histogenesis is quite important in order to have a comprehensive understanding of the significance of lymphocytes in an organism.

At first, the thymus is epithelial only. Then, at a quite early stage, large lymphocytes occur in its mesenchymal environment, as in other areas of the organism, and, sometimes, pale migrating cells with small nuclei are detectable. All these ameboid cells then pass into the epithelial anlage, where they transform into typical large lymphocytes. Hence, the initial stages are exactly the same as in the liver, i.e., the first lymphocytes of the thymus are undoubtedly morphologically identical to the first granulocyte-producing lymphocytes in the liver. It is just that the conditions that exist for these cells are, apparently, quite different; since the lymphocytes in the thymus, irrespective of their exceptionally strong proliferation, never produce erythroblasts, and only very few granulocytes, and always only their own kind. Soon they infiltrate the entire organ. Upon proliferation, they become smaller and smaller, and, finally, evolve into massive numbers of typical small lymphocytes, which are washed out into the blood.

In the case of the developing lymph nodes, one again observes the differentiation of small, densely populated, undifferentiated mesenchymal cells into small ameboid migrating cells. Here, too, from the beginning, strong polymorphic features are evident among these migrating cells. Rather quickly, single large lymphocytes emerge; but for the most part, very small, ameboid, elements with bright nuclei and scarce cytoplasm initially appear. They proliferate, turning partially into typical small lymphocytes with dark nuclei and migrate to lymphatic crevices. On the other hand, one might occasionally observe them transform into large, even giant, lymphocytes that may then produce small lymphocytes, similarly to the thymus. Therefore, it must be pointed out with certainty that the large lymphocytes are not essential to produce the typical small lymphocytes in the embryo.


An investigation of the fetal blood formation thus teaches us that one cannot distinguish between myeloblasts and lymphoblasts. A single cell type exists: a ubiquitous, non-differentiated, polymorphic, migrating mesenchymal cell, which, when influenced by specific existence conditions, has a variable appearance and may produce a variety of differentiation products. Likewise, the lymphoblasts and myeloblasts in embryo cannot be distinguished from each other through merely histological means.


When considering blood formation of adult organisms from our present point of interest, two questions related to non-granulated cells need to be resolved.

The first question addresses changing relationships between the large and the small lymphocytes. Those two terms were created on the basis of studies in adult organisms. Currently, the general opinion is that, in an adult organism, the small lymphocytes actually arise via proliferation of larger lymphocytes in the germinal centers; however, they are themselves incapable of further reproduction, and, in particular, they cannot re-transform themselves into large lymphocytes.

However, based on my studies, I must take a different viewpoint. Indeed, in an adult organism, the small lymphocytes mostly develop through proliferation of larger cells. For some time immediately after their emergence, they are, in fact, unable to proliferate. Most likely, this state depends on the special relation between nucleus and the cytoplasm, caused by the previous intensive proliferation. I am absolutely certain that these small mature lymphocytes are able to proliferate further. They enter the blood and circulate, and where they find appropriate conditions they can function again as fully non-differentiated mesenchymal cells and present a starting point for various developmental events; they can certainly transform themselves, in a hypertrophic way, into large lymphocytes capable of division. In my opinion, the reason for such a strange phenomenon, that the majority of lymphocytes in adult organisms must pass through the stage of a small cell incapable of proliferation for a certain time period, is that small lymphocytes can be easily transported from one place to another in blood and lymph flow, thus reaching all organs and tissues everywhere. Weidenreich has also recently expressed this opinion.

Hence, the small and the large lymphocytes are merely transitory stages in the life of one and the same type of cell, i.e., of a lymphocyte in the broadest sense of the word.

The second question addresses the distinction between special types of lymphoblasts and myeloblasts in an adult organism. If this difference, as we have seen, is not justified in the embryo, one does not have to conclude, that, a priori, it is not possible in the adult organism. A number of authors, starting with Schridde, also postulate that non-granular cells in lymphoid tissue are not the same large lymphocytes found in the myeloid tissue, but represent two different types of cells, namely lymphoblasts and myeloblasts.

When defining criteria for identification of the two cell types, histological features should certainly be defined first, and, secondly, physiological characteristics, especially their prospective developmental potential.

In regard to histological characteristics of the two cell types, I asked Sir Dr. S. Tschaschin in my laboratory to check in detail the differences reported by Schridde.

As far as we can judge from the results obtained to date, in most cases one is able to note certain differences in newborn animals. However, these differences are small. In general, the lymphoblasts have a thinner and more homogenous cytotoplasmic rim; in the nucleus, larger nucleoli are found which are, as a rule, densely colored. The so-called myeloblasts have mostly, though not always, a broader cytotoplasmic rim of a lighter, reticular structure, with a widely variable degree of basophilia. The nucleus always contains nucleoli, but these are smaller and their color is less distinct. Generally, the myeloblasts appear more polymorphic than the lymphoblasts, and the differences among the myeloblasts are often more prominent than those between myeloblasts and lymphoblasts.

In particular, special attention was given to the Altmann-Schridde staining technique, which has been described by Schridde as the most important method for discrimination. It turns out that the Eosin-Azure-staining of the large cells in the adenoid tissue and in the bone marrow, i.e., Schridde’s lymphoblasts and myeloblasts, detected both cells with and without granules; the majority of which contain only a few granules. This is in contrast to Schridde, who states that lymphoblasts should always contain granules, while myeloblasts never do. The small and middle-sized lymphocytes, however, always contain very clear and numerous granules. The specific granules and the eosinophilic granules are stainable as well. This method generally gives exactly the same images as the well-known original technique by Altmann, and it seems to me that it is especially inappropriate for the studies of blood cells. The different details of staining that Schridde refers to cannot be seriously considered for distinguishing between different types of the cells. It goes without saying that all these granulation images, in general, cannot have any particular significance because, undoubtedly, the granules can emerge de novo or disappear in one and the same cell, depending on its functional state, even if they exist in vivo.

So even if there are certain unstable histological differences that are difficult to define, one must consider that the cells in lymph nodes and in marrow exist in very different environments, and, therefore, this factor alone might already present a sufficient explanation for the histological differences. Moreover, we saw clearly that the lymphocytes from the very first embryonic stages onwards are characterized by an extreme polymorphism, even though all of them look completely similar. Hence, a clear separation between lymphoblasts and myeloblasts is not justified on histological basis alone. Such a separation would only be possible if we could succeed in proving that cells of one type can never transform (differentiate) into cells of the other type and, vice versa, the (cell) products of differentiation are completely different for both cell types under any possible conditions.

It is these physiological, or rather reproductive, cytogenetic characteristics of our cells that we would now like to scrutinize. If the lymphocytes in adenoid tissue and marrow lymphocytes are similar and produce normally divergent differentiation products solely due to different conditions of existence, one could try to recreate such conditions artificially so that the lymphocytes of adenoid tissue, the supposed lymphoblasts, could differentiate into granulocytes and erythroblasts. It is known, however, that myeloid transformation can take place in adenoid tissue on various occasions. Additionally, it is generally accepted that the latter originate from autochthonous elements. But the question remains, from which cells? It is known that it is not the cells of the germinal centers that differentiate into myelocytes and erythroblasts, but cells located in the trabeculae of the lymph nodes and the red pulp of the spleen. For the dualists, it is exactly this that proves that their view is correct. According to them, myeloid elements are derived from either pre-existing myeloblasts that are very different from the lymphoblasts or directly from the vessel-lining cells. Others assume that, in this case, some special adventitial undifferentiated mesenchymal cells may become a starting point of the transformation.

In my laboratory, I asked Mme Dr. H. Babkin to perform special experiments in animals, in order to get closer to the answer to this issue. In the spleen it was easy to induce, in part, the myeloid transformation, i.e., myelo- and megakaryocytopoiesis—it sufficed to introduce an aseptic foreign body into the spleen tissue; numerous myelocytes and megakaryocytes soon emerge in its environment. In the lymph nodes, however, this and other methods have so far been unsuccessful in triggering a myeloid transformation. Similarly, in the spleen, the Malpighi’s bodies remained unchanged: the myelocytes appeared only in red pulp and venous sinuses.

In the first instance, these experiments seem to confirm the difference between lymphocytes and myeloblasts. However, I do not believe that these preliminary results are to be interpreted in this way after all. We must take into account that considering the very special conditions existing in the adenoid tissue, these areas would be more suitable than other parts of the body for homoplastic proliferation of undifferentiated mesenchymal cells and/or lymphocytes. The pre-requisites for a myeloid transformation of lymphocytes are normally completely missing in such "breeding" environments. Two different conditions are necessary, on the one hand for a "homoplastic" proliferation in an unchanged, undifferentiated state and, on the other hand, for "heteroplastic" development (differentiation) towards myeloid elements, that, apparently, cannot co-exist in the adult organism. This is why neither the cells of the germinal centers, nor the young small lymphocytes can be artificially induced to directly transform into granulocytes and erythroblasts at the place of their emergence. As it is known, the homoplastic proliferation ceases, and the germinal centers disappear wherever the myeloid transformation begins.

The young age of the bulk lymphocyte population in adenoid tissue may be a probable obstacle for the myeloid transformation. It may be that for such cells a certain period of time has to elapse in order for them to become capable of myeloid differentiation, and, moreover, for this purpose, additional special favorable conditions must exist. For example, it can be speculated that their blood circulation makes the lymphocytes originating in the adenoid tissue more susceptible to myeloid transformation.

While all of this evidence may represent only indirect and probably doubtful proof that lymphocytes of the lymphoid tissue and those of the myeloid tissue are equivalent with regard their prospective potential to develop in an adult organism, in my opinion, there is another direct, albeit preliminary argument, which is often ignored by the various authors who have described heterotopic formation (differentiation) of myeloid tissue.

Some time ago I studied the histogenesis of myeloid tissue that develops in the rabbit kidney after the ligature of its main vessels. This approach is favorable in the sense that lymphoid elements are apparently absent in the scarce stroma of the kidney. It turned out that all of the bone marrow elements—granulocytes, megakaryocytes, and erythroblasts—emerge from the lymphocytes circulating in blood, i.e., from cells that are proven to originate from adenoid tissue and its germinal centers. In the course of this process, the small blood lymphocytes transform themselves into large lymphocytes again and migrate into the tissue, either as small cells, or as preformed large cells. While still inside the vessels, or only after they have left them, they either produce myelocytes by accumulation of granules in their cytoplasm, or form erythroblasts by producing hemoglobin. Actual myeloblasts do not seem to exist in normal peripheral blood, even though K. Ziegler believes the large mononuclear leukocytes to be such continuously undifferentiated cells with myeloid differentiation capacity. But these, too, according to more recent findings, emerge from banal small lymphocytes.

In my opinion, it can be assumed that the heterotopic emergence of myeloid elements in humans may occur at the expense of ubiquitous lymphocytes in the circulating blood or to lymphocytes of the connective and adenoid tissue, which are completely equivalent, but not at the cost of latent myeloblasts, or problematic proliferating adventitial cells, or the cells of vessel walls. 

All in all, my final conclusion is that for adult organisms, as with the developing embryo, there is no reason to assume the existence of two clearly distinct cell types, i.e., the myeloblasts and the lymphoblasts. In mammalian organisms a single type of cell exists, i.e., the lymphocyte, in the broadest sense of the term, which both looks different and may produce a variety of differentiation progeny depending on their current location and survival factors. The lymphocytes are ubiquitous, equivalent for all places, and are indistinguishable from one another by means of histological or hematological approaches. In adenoid tissue, during homoplastic proliferation, they always produce lymphocytes. An easily transportable cellular form emerges: a small lymphocyte, which circulates in the blood and lymph flow, moving throughout the body, until, after a certain period of inactivation, it unfolds its full ability for development.

Originally published in: Folia Haematologica 8.1909, 125-134. (English translation prepared by Claudia Koltzenburg, Alexey Chukhlovin, Athanasius Anagnostou, and Carol Stocking for Cellular Therapy and Transplantation, Vol. 1, No. 3, 2009. Although every attempt is made to ensure precision in the translation into English of the material in these articles, we do not guarantee nor imply their absolute accuracy.)

Volume 1, Number 3

Download PDF version

doi 10.3205/ctt-2009-en-000032.01

Back to the list