Introduction

A recent and comprehensive report investigated the world trends in stem cell publications from 1996 to 2012 (Barfoot et al. 2013). Among the main results, the study shows a rapid growth both in the number of publications and in the number of active researchers—estimated by the number of authorship—in this subject. The increasing interest in stem cell (SC) research is related to its potential usage as a treatment in a variety of human diseases as well as in injuries caused by accidents. Such characteristic is associated to two main properties that differ SCs from other cell types: self-renewal, that is, they can generate copies of themselves, and differentiation, that is, they can become a specialized cell.

The use of SC in therapeutic practice is, however, not new. In leukaemia, for example, tissue (or adult) SCs have been used since the 1950’s to restore the patients’ blood and immune systems (Copelan 2006). But, it was after the isolation of embryonic stem cells from a mammalian in the 1980’s, that SC research expanded rapidly. This type of stem cell, derived from early-stage, is pluripotent, that is, it has the capacity to transform in most cell types. In 2007, a human SC was developed in laboratory: the induced pluripotent stem cells. This new type of SC is generated by using a technique called reprogramming, with which a human cell is genetically reprogrammed to become an embryonic SC. The ethical debate about the use of human embryonic for research proposes (Hyun 2010; McLaren 2001) as well as the potential usage of induced pluripotent SC and the ease with which these cells can be generated (Cyranoski 2008) are probably factors that contributed for the latter’s remarkable growth rate of publication (Barfoot et al. 2013).

The history of SCs reveals that in the beginning these cells have been used as a cell therapy in the treatment of diverse diseases or disorders like leukaemia, burnings or corneal damage. However, after the discovery of the induced pluripotent SCs, the potential use of these cells has turned to disease modelling, which includes drug discovery, and pre-clinical toxicological assessment of new drugs. However, such use requires many efforts on basic science research for a better understand the SCs’ mechanisms of division, differentiation and self-renewal. Hence, regarding SC research, the state of art of SC field suggests there are at least two dimensions: research with stem cell in which the focus is cell replacement or transplantation therapies developed by scientists from medical fields, such as cardiology, neurology, for instance; and research on stem cell that focuses on the basic science research.

Considering the growing importance of SC field, the present study seeks to map the scientific and cognitive structure of Brazilian stem cell publications. It is worth highlighting some landmarks towards the consolidation of Brazilian SC research. In 2001, the field started to be formally supported with a federal funding program through a national project called The Millennium Institute in Tissue Bioengineering. As early as 2005, a federal law was approved and determined norms for manipulating and developing research on embryonic SC. Some years later, in 2008, eight Cell Technology Centres and a Brazilian Network for Cell Therapy were founded in the country with the aim of developing cell therapy and SC studies. Hence, since 2005, hundreds of projects in this field have been approved and supported, mainly, with federal funds, which achieved approximately US$ 100 million. All these initiatives strengthened the novice stem cell research in the country in the beginning of 2000’s. Soon, Brazil became the leader of Latin American stem cell research with the highest number of personnel and institutions involved and number of papers published. Also, the country started to be part of a selective group of countries with the technology to produce induced pluripotent stem cells (iPSCs), a technique that allows the genetic reprogramming of adult cells to exhibit features of embryonic cells. (Pranke et al. 2014; Ruiz 2013; Rehen and Paulsen 2007)

Citation studies and the field of stem cell

Citation studies started in the 1920’s, with Gross & Gross considered the first to use “[…] citation counts to evaluate the importance of scientific work” (Bornmann and Daniel 2008, p. 45). During the following decades, the interest in studying citations to map and evaluate scientific performance increased notably, especially after Science Citation Index (SCI) was founded by Eugene Garfield in 1964. In the 1960’s, studies on citation shifted to a new stage: from simple analyses of the quantity of citations per document to more complex analyses focused on the relationship between citations. In 1963, Kessler proposed the technique of analysing word co-occurrence or bibliographic coupling, which takes into account the frequency with which a pair of documents (papers, for instance) shares one or more documents on their list of references. This technique assumes that the greater the number of shared references, the closer the pair is in terms of thematic and central concepts (Morris and Van der Veer Martens 2008).

In the 1970’s, additional co-occurrence techniques appeared. Small (1973) and Marshakova (1973) authored a bibliometric method, termed co-citation analysis, that focused on the co-occurrence of citations. In the following decades, other forms of co-citation analysis were proposed: author (White and Griffith 1981) and journal co-citation analysis (McCain 1991) and, more recently, issue co-citation analysis (de Moya-Anegón et al. 2004). Co-citation analysis can indicate the intellectual structure or, in other words, the main domains of a field. Such a technique considers the frequency with which pairs of documents, authors or journals are co-cited by other documents. The structural relationship demonstrated by co-citation analysis reveals theoretical and/or empirical similarities between the pairs analysed. Hence, a greater frequency of co-cited pairs suggests a closer relationship between the two.

The first study presenting a journal co-citation analysis dates to 1991, when Katherine W. McCain published Mapping economics through the journal literature: an experiment in journal cocitation analysis. This study illustrated that journal co-citation analysis, much like other types of co-citation analysis, may be useful to distinguish whether journals are more specialised or generalised and to identify, among other characteristics, their “methodological orientations, institutional affiliations, relative prestige” (McCain 1991, p. 291). For Morris and Van der Veer Martens (2008, p. 250), journal co-citation analysis helped to identify “the key journals and specialties from which base knowledge is drawn”. After McCain’s study, the technique of journal co-citation spread, being applied in a variety of subjects and fields.

Studies on SC scientific literature are relatively frequent in the fields of Bibliometrics and Scientometrics. Among SC publications that apply co-occurrence techniques, some are detailed. Zhao and Strotmann (2011) studied the trends in world scientific literature on SCs published from 2004 to 2009, by using author co-citation analysis The results revealed that world SC research is more concentrated in four areas: embryonic and pluripotent stem cells, neurogenesis, regenerative medicine, and cancer research. An and Wu (2011) applied co-word analysis on the world SC literature indexed in PubMed from 2001 to 2010. The authors found that regenerative medicine, gene therapy and neoplastic stem cells area emerging areas in SC research while cell biology, molecular biology, immunology and genetics display the faster growth rate. A more recent study authored by Cantos-Mateos et al. (2012) has also applied co-word analysis but on Spanish stem cell scientific literature from 1997 to 2007. Among the main results, authors found that areas of haematology, oncology and biophysics are the “outstanding areas” of SC study in the country.

In Brazil, studies on SC scientific literature are rare. A recent article addresses a comparison between the number world publications and patents in induced pluripotent SC (da Coelho et al. 2014). The study shows that US and China are the countries with the largest number of patents while US and Japan appear the largest number of papers, which are classified mostly in the fields of cell biology, biochemistry & molecular biology and biotechnology & and applied microbiology. In addition to this work, we call attention to the study entitled Trends of intellectual and cognitive structures of stem cell research: a study of Brazilian scientific (Machado and Leta 2013), which, to our knowledge, was the first to investigate Brazilian stem cell publications by applying co-occurrence analyses. In this paper, we found by applying author co-citation and co-word analysis that the profile of both authors and words during the study period changed from basic research, devoted mainly to health sciences, to applied research. In the present study, the analysis still focuses on Brazilian SC indexed literature, but the interest shifted towards a better understanding of how this promising field is structured in a country, where both research with and research on SC have high potential not only because of its international recognition but also because of its legal support.

Methodology

The journal co-citation analysis is the analytical strategy of this study and so journals are the main unit of analysis. In the following, the main methodological steps are presented.

The first step was the selection of a database and the retrieval strategy. For the present study, we selected the Web of Science (WoS) database, using “stem cell*” in the filter topic, that is, publications retrieved presented the term “stem cell*” in the title, abstract, author keywords or database keywords. It is important highlighting that stem cell has a unique meaning and there is no haziness or doubt when searching this term. Although there are different types of stem cells, they are all named as “stem cell”.

The filter topic was combined with publication year, from 2001 to 2010, address, Brazil or Brasil and publication type, article. All the information from the retrieved publications was downloaded simultaneously and collected in December 2013.

The next step was to categorize and cleaning the data manually. First, journal titles were assigned according to their topic category at Journal Citation Reports so it was possible to identify the most representative categories of each cluster (for instance: cell biology was the most representative category in cluster 1, corresponding to 18.3 % of the total). Then, some journals were excluded since they were assigned to the multidisciplinary sciences category, an indicative that they have a broad thematic interest. The excluded journals are: Anais da Academia Brasileira de Ciências (An Acad Bras Cienc), Nature, Proceedings of the National Academy of Sciences of the United States of America (P Natl Acad Sci USA), Plos One (Plos One) and Science. In this step, we have used Microsoft Excel only.

For the final step, processing the data, Microsoft Excel and VOSviewer (version 1.5.5), developed by Van Eck and Waltman (2010), were combined to carry out a descriptive and exploratory analysis. VOSviewer allows the visualisation of information in a four-mode bi-dimensional map: label view, density view, cluster density view and scatter view. Only the label map mode was selected for this study because it “[…] is particularly useful to get a quick overview of the important areas of a map” (Van Eck and Waltman 2013, p. 6). Hence, by applying VOS clustering technique, which unifies, according to Waltman et al. 2010, “a weighted and parameterized variation of modularity-based clustering”, clusters could be identified.

In this step, we have also used Pajek software, developed by Batagelj and Mrvar (2010), was used to obtain measures of centrality: degree, closeness and betweenness. Such measures allowed us to identify the main journal titles. The degree of centrality measures the number of direct links (co-citation) that a node (journal) establishes with another one (Hanneman and Riddle 2005). Thus, it identifies journal titles with the highest frequency of co-citation, in other words, journal titles with more strategic and therefore influential positions on the web. When analysing the map, these titles appear as bigger symbols. The other measures are complements. It is important to note that journal titles with high degrees of centrality do not necessarily have the highest closeness or betweenness. From 2001 to 2010, 731 Brazilian papers on stem cells were retrieved, citing 3415 journals in their reference lists. Only journals co-cited ≥10 times were considered, resulting in 454 journal titles, or 13.3 % of the total.

Results

The main results are presented in three sections: an overview of Brazilian and world stem cell publications, the most relevant journal categories and the main journal titles.

The growth of stem cell publications

From 2001 to 2010, the total number of world scientific publications on stem cell increased 269.8 % (from 5532 to 20,457), while Brazilian publications increased 1086.9 % (from 23 to 273). During this period (2001–2010), Brazilian stem cell publications registered in the WoS database summed up 1174, an amount that corresponds to 0.95 % of the world’s total number of stem cell publications (123,056).

Among these 1174 Brazilian publications, 731 (62.3 %) are original articles, meaning original contributions published in specialised and peer-reviewed journals, which account for 1.01 % of the world’s publications in this subject area (72,321). Table 1 presents the annual number of this selective set of publications. Although Brazilian performance is still low in terms of the quantity of original articles, it can be noted that not only did the quantity increase but also the country’s relative participation in the world from 0.42 % in 2001 to 1.48 % in 2010. In fact, in the later years, Brazilian original articles display a stronger growth rate (14-folds) than that of the world (4-folds). Such growth is also a result of the increase in the number of new Brazilian titles indexed in WoS database (Leta 2012).

Table 1 Number of Brazilian and world articles on stem cell (2001–2010)
Full size table

To better understanding Brazilian performance, a correlation analysis was developed using time and total number of publications as variables. The Pearson coefficient was applied together with coefficient of determination, (r 2) and t test (t). The results indicate r = 0.94 and r 2 = 88 %. In other words, 88 % of the observed increased can be explained by the variable time. The t test (p value < 0.05) indicates that total publications and year of publication are statistically correlated.

The most influential categories of co-cited journals

Figure 1 presents a relational map based on the co-cited journals extracted from the 731 Brazilian SC original articles published from 2001 to 2010. Taking into account the similarity index measured automatically by VOSViewer, the journals are grouped into eight clusters, which are coloured differently on the map. Each cluster has an internal structure formed by a core of co-cited journals, which share a high thematic homogeneity; outside of the clusters, co-cited journals are quite heterogeneous (Hair Jr et al. 2009). The most influential journal titles in each cluster can be seen easily: they are in bold and have bigger symbols.

Fig. 1
figure 1

Bibliometric map of co-cited journals of Brazilian stem cell articles (2001–2010). Colours indicate different thematic clusters while the size of the symbols represents the weight of the 454 co-cited journals for the entire analysis. VOSviewer software (version 1.5.5). (Color figure online)

Full size image

In the map centre, there is a set of journal titles that are well linked. Their proximity with other journals indicates that they are co-cited in a single paper (Van Eck and Waltman 2011). In the superior part of the map, there is another set of co-cited journals forming cluster 6 (blue), which is dedicated to fields that are more specific. Such a tendency may explain why journals in this cluster are so dispersed.

As a general observation, the map also presents isolated journal titles, Plant Cell (plant cell), Journal of the American Statistical Association (J Am Stat Assoc), and International Immunology (Int Immunol). These journals display a very low frequency of co-citation, suggesting that they are poorly or moderately related to stem cells.

There is a strong concentration of co-cited journals in the centre and lower right regions of the map, indicating that the most co-cited journals are located in these areas. These include: Blood, Bone Marrow Transpl (Bone Marrow Transplantation), Stem Cells, New Engl J Med (New England Journal of Medicine), J Biol Chem (Journal of Biological Chemistry), Cell and Circulation. These journals are grouped according to similarity index, which reflects a similarity in their content or intellectual approach (Verbeek et al. 2002). When grouped into categories, this set of journals represents the main specialties of a field or subject (Small and Griffith 1974). Another general feature is the presence of two isolated groups (in light blue and orange) at the higher region of the map, composed of dentistry and ophthalmology journals.

Table 2 quantifies co-cited journals and presents the two most prevalent associated categories. The dimensional structure of the map and the descriptive analysis of the clusters (Table 2) allow us to better understanding the clusters and how the stem cell thematic is distributed throughout the fields.

Table 2 Details of the clusters formed by co-cited journals in Brazilian stem cell articles (2001–2010)
Full size table

It can be noted that cluster 1 (red symbols in the centre of the map) is the largest and most diverse in terms of the number of co-cited journals (99), where Stem Cells, Cell and Development were the most co-cited. Cluster 1 is also the most diverse in terms of category (36) and displays a strong concentration of co-cited journals in the categories of cell biology and biochemistry & molecular biology. Also relevant (but not shown) is the presence of co-cited journals categorised as developmental biology (7.8 %), reproductive biology (5.9 %) and biology (2.6 %). This last category “[…] includes resources having a broad or interdisciplinary approach to biology” (Reuters 2012, online). Considering the scope of these categories and the most co-cited journals, we may infer that cluster 1 includes publications that are mostly focused on understanding different intracellular mechanisms related to cell division, differentiation and self-renewal. In order words, it reveals the most current and promising front: research on stem cells.

Cluster 2 (green symbols in the inferior part of the map) is the second largest cluster in terms of number of co-cited journals and categories. The most co-cited journals include Journal of Biological Chemistry (J Biol Chem), Journal of Neuroscience (J Neurosci), Neuron (Neuron) and Experimental Neurology (Exp Neurol), while the main categories are: neurosciences (31.6 %) and biochemistry & molecular biology (11.4 %). The category clinical neurology (10.5 %) is also important in this cluster (not shown). Considering the scope of these journals and categories as well as their proximity to cluster 1, we may infer that cluster 2 includes stem cell publications devoted to studying the nervous system using the two approaches: research on stem cells and research with stem cells.

Clusters 4 and 7 (yellow and grey symbols, respectively, both at the bottom of the map) show a high thematic affinity: stem cell research associated to diseases of the immune system and blood. The most co-cited journals are Blood, Bone Marrow Transplantation (Bone Marrow Transpl), New England Journal of Medicine (New Engl J Med), Journal of Clinical Investigation (J Clin Invest), Transplantation and Journal of Immunology (J Immunol) for cluster 4 and British Journal of Haematology (Brit J Haematol), Experimental Haematology (Exp Hematol) and Leukemia for cluster 7. Regarding the categories, we found that haematology weighs in both clusters 4 and 7, suggesting that stem cell publications included in both clusters are highly oriented towards the oldest stem cell research approach: the therapeutic usage of these cells in blood related diseases, such as leukaemia (Dupont 1997; Copelan 2006). Studies represented in these clusters, especially in cluster 7, are under the scope of the field onco-haematology (Bydlowski et al. 2009).

Cluster 3 is located at the interface of clusters 1, 4 and 7. The most co-cited journals are Cancer Research (Cancer Res), Journal of Clinical Oncology (J Clin Oncol) and American Journal of Pathology (Am J Pathol), which are classified under the categories oncology (25.0 %) and pathology (14.4 %). These characteristics suggest this cluster is related to the clinical usage of stem cell therapy, especially in cancer-related diseases, or, in other words, research with stem cells.

Clusters 5, 6 and 8 are more field-oriented. The most co-cited journals of cluster 5 (pink symbols in the centre of the map) include Circulation (Circulation), Nature Medicine (Nat Med), Lancet, Biochemical and Biophysical Research Communications (Biochem Bioph Res Co), Cell Transplantation (Cell Transplant) and Circulation Research (Circ Res), while the largest categories include cardiac & cardiovascular systems (19.7 %) and medicine, research & experimental (9.2 %). Interestingly, according to Reuters (2012) the last category is a field “with a particular emphasis on extremely novel techniques and clinical interventions in a broad range of medical specialisations […]”. Hence, considering the weight of both categories, we infer that the publications included in cluster 5 are more oriented towards the use of stem cells in the treatment of diseases or incidents associated with the cardiovascular system, such as heart attack or Chagas disease (Santos et al. 2004).

Cluster 6 (light blue in the superior part of the map) displays a strong association with the field of dentistry. The most co-cited journals in this cluster are: Biomaterials, Journal of Periodontology (J Periodontol) and Tissue Engineering (Tissue Eng), while the largest category is dentistry, oral surgery & medicine (28.1 %). The use of stem cell therapy to restore tissues or organs inside the oral cavity is relatively recent. Research with this approach marks a new scientific field Regenerative Dentistry, which started after clonogenic cells, cells with high regenerative power, were identified inside the dental pulp of adult patients (Gronthos et al. 2000).

Finally, cluster 8 is strongly associated with another specialty: ophthalmology. Among the most co-cited journals are: Investigative Ophthalmology & Visual Science (Invest Ophth Vis Sci) and Ophthalmology. Stem cell therapy has been used in certain corneal diseases, hereditary degeneration of the retina, macular degeneration and age-related ischemic retinopathy (Siqueira 2009).

A complementary analysis (not shown) was carried out to identify the top ten categories encompassing the largest share of co-cited journals (co-cited 19 or more times). We found that, in order of highest share, this selective group included the following categories: cell biology, neurosciences, biochemistry & molecular biology, immunology, haematology, oncology, medicine-research experimental, genetics & heredity, surgery, dentistry-oral surgery & medicine and pathology.

Having identified the top ten largest categories of co-cited journals, we checked their presence in each cluster (Table 3). There are co-cited journals classified in the categories Medicine-research experimental, cell biology and biochemistry & molecular biology in all clusters, except cluster 7. Also, co-cited journals classified in the category haematology are present in six clusters.

Table 3 Distribution (%) of the top 10 categories with the largest share of co-cited journals in the clusters formed by co-cited journals in Brazilian stem cell articles (2001–2010)
Full size table

The distribution of the top ten categories as well as the pattern found in the co-cited journals map suggest two main emphases for the structural basis for the Brazilian stem cell research: research on stem cells, which is oriented towards a better understanding of the processes of cell differentiation and regeneration and research with stem cells, which is focused on their clinical use for the treatment of primarily blood-related diseases.

The most influential co-cited journals

Using the relational map, the next and final step was to identify the most influential journals in this network. The primary measures of centrality that are commonly used in social network studies (Otte and Rousseau 2002) were applied here: degree, closeness and betweenness (Table 4).

Table 4 Centrality measures of the ten most co-cited journals in Brazilian stem cell articles (2001–2010)
Full size table

Considering all co-cited journals, we found that they display an average centrality degree of 153.76 (standard deviation = 79.84). Interestingly, 24.0 % of co-cited journals show a centrality degree lower than 100 and therefore appear in the periphery of the map (Fig. 1). Taking into account the centrality degree, we identified the ten most influential co-cited journals including Cell, Blood, Stem Cells and Journal of Biological Chemistry. These journals were also the ones with the largest centrality closeness, suggesting that they are highly central in the map structure.

While the centrality degree indicates the most central or influential co-cited journals, betweenness signifies the (inter) disciplinary nature of co-cited journals. According to Leydesdorff (2007, p. 1311), “journals that are more deeply embedded in a disciplinary cluster tend to show very low values on the betweenness”. This is supported by the data we report in Table 4. When compared, for instance, with the lowest co-cited journals, the top ten most influential co-cited journals have very low values for betweenness. In this case, we may assume that the most influential co-cited journals include more specialised journals, as Blood and Stem Cells, as well as some with a broader scope of interest, such as Lancet and New England Journal of Medicine. As a whole, the set of the most influential co-cited journals is a sum of disciplinary subjects, what gives Brazilian stem cell research the tendency to be an interdisciplinary field.

The three measures analysed highlight the relevance of Cell, Blood, Stem Cells, J Biol Chem and New Engl J Med. These are the core journals and are located at bottom of the map (Fig. 1) where the density of co-cited journals is higher. These journals occupy prominent positions in cluster 1, 2 and 7, the first two of which were characterised by research on stem cells while the last was characterised by research with stem cells.

Discussion and some remarks

Despite its limitations (such as the use of a single source for retrieving data and ≥10 co-citations as the threshold to select co-cited journals), the most relevant aspects of the present study are its originality and the effort to develop a type of research that is rare in Brazilian Scientometrics, due to its complexity in the analysis and interpretation of data as well as its interdisciplinary approach.

The results indicated that Brazilian stem cell research is structured in a diverse and interdisciplinary literature that is revealed by the eight clusters presented in Fig. 1. In this sense, we may infer that each cluster exposes a specialty, as suggested by Small (1973), or a discipline where stem cell research is being developed.

The high number of journal categories (69) and the profile of the top ten co-cited journals corroborate the interdisciplinary approach of stem cell field found also in previous studies (Barfoot et al. 2013; Cantos-Mateos et al. 2012, Rehen and Paulsen 2007). These findings also point to two primary research focuses: research with stem cell, which is oriented towards the clinical use of stem cells for the treatment of different diseases and research on stem cells, which is focused on elucidating the processes of cell differentiation and regeneration. These research orientations are clearly illustrated by the core co-cited journals, which include Cell, Blood, Stem Cells, Journal of Biological Chemistry and New England Journal of Medicine.

The set of results presented in this study illustrate a promising picture of Brazilian stem cell research because the country displays a broad spectrum of research interests, especially on stem cells. Such pattern, however, may be (or not) shared by other countries. Due to the legal and also the technological aspects required to carry research on stem cell, it is possible that most of the countries present an intellectual structure strongly based on co-cited journals from Hematology or other medical field, where the research with stem cell prevails, in other words, they are more oriented to the treatment and/or to use of these cells in a variety of human blood diseases and injuries.

The pattern towards the research on stem cells is not only its most recent focus, but it is also the one with the highest potential to affect world health. In Brazil, this type of research was driven by the approval of a federal law, which regulates embryonic stem cell use and research (Rehen and Paulsen 2007). Therefore, the analysis of stem cell research in other countries, especially those that have no formal regulations, could provide relevant data on the relationship between a country’s research focus and science policies.