Advertisement

Commentary Free access | 10.1172/JCI63939

How many neutrophils are enough (redux, redux)?

Samuel C. Silverstein1,2 and Raul Rabadan3,4

1Department of Physiology and Cellular Biophysics, 2Department of Medicine, 3Department of Biomedical Informatics, and 4Center for Computational Biology and Bioinformatics, College of Physicians and Surgeons, Columbia University, New York, New York, USA.

Address correspondence to: Samuel C. Silverstein, Department of Physiology and Cellular Biophysics, Columbia University College of Physicians and Surgeons, 630 W. 168th Street, Rm. 11-411, New York, New York 10032, USA. Phone: 212.305.3546; Fax: 212.305.5775; E-mail: scs3@columbia.edu.

Find articles by Silverstein, S. in: PubMed | Google Scholar

1Department of Physiology and Cellular Biophysics, 2Department of Medicine, 3Department of Biomedical Informatics, and 4Center for Computational Biology and Bioinformatics, College of Physicians and Surgeons, Columbia University, New York, New York, USA.

Address correspondence to: Samuel C. Silverstein, Department of Physiology and Cellular Biophysics, Columbia University College of Physicians and Surgeons, 630 W. 168th Street, Rm. 11-411, New York, New York 10032, USA. Phone: 212.305.3546; Fax: 212.305.5775; E-mail: scs3@columbia.edu.

Find articles by Rabadan, R. in: PubMed | Google Scholar

Published July 23, 2012 - More info

Published in Volume 122, Issue 8 on August 1, 2012
J Clin Invest. 2012;122(8):2776–2779. https://doi.org/10.1172/JCI63939.
© 2012 The American Society for Clinical Investigation
Published July 23, 2012 - Version history
View PDF

Related article:

Evidence for bistable bacteria-neutrophil interaction and its clinical implications
Roy Malka, … , Eliezer Shochat, Vered Rom-Kedar
Roy Malka, … , Eliezer Shochat, Vered Rom-Kedar
Research Article

Evidence for bistable bacteria-neutrophil interaction and its clinical implications

Abstract

Neutropenia, which may develop as a consequence of chemotherapy, increases the risk of bacterial infection. Similarly, increased risk of bacterial infection appears in disorders of phagocytic functions, such as the genetic disorder chronic granulomatous disease. To elucidate the organizing principles behind these distinct immunodeficiency conditions, we investigated the interaction between in vitro bacteria and human neutrophils by experiments and mathematical modeling. The model and the experiments showed that the in vitro bacterial dynamics exhibit bistability for a certain range of neutrophil concentration and function. Thus, there is a critical bacterial concentration above which infection develops, and below which neutrophils defeat the bacteria. Whereas with normal neutrophil concentration and function, an infection may develop when the initial bacterial concentration is very high, under neutropenic conditions or when there is neutrophil dysfunction, the critical bacterial concentration can be lower, within the clinically relevant range. We conclude that critical bacterial concentration has clinically relevant implications. The individual maximum bearable bacterial concentration depended on neutrophil concentration, phagocytic activity, and patient barrier integrity; thus, the resulting maximal bearable bacterial concentration may vary by orders of magnitude between patients. Understanding the interplay between neutrophils and bacteria may enhance the development of new therapeutic approaches to bacterial infections.

Authors

Roy Malka, Baruch Wolach, Ronit Gavrieli, Eliezer Shochat, Vered Rom-Kedar

×
Abstract

Many chemotherapeutic regimens produce neutropenia, which predisposes to microbial infection. However, not all neutropenic individuals develop infections, so the ability to predict this outcome would be a powerful clinical tool. In this issue of the JCI, Malka et al. describe a dynamic system model of neutrophil bactericidal activity that confirms and extends the concept of critical neutrophil concentration. The authors demonstrate that when the neutrophil concentration approaches the critical concentration, bacterial populations in contact with them exhibit bistability. Their experimental findings raise the intriguing possibility of greater variability in bactericidal activity of neutrophils from healthy adults than heretofore recognized; their model predicts that this could have life-and-death consequences.

Although the link between neutropenia and infection risk is clear, the precise number of neutrophils required to maintain health remains a topic of intense research interest. Neutrophils continuously patrol the luminal surfaces of endothelial cells, searching for signs of infection or inflammation. Such signs stimulate them to emigrate from blood into extravascular compartments (1). There, their armamentarium of chemoattractant and phagocytosis-promoting plasma membrane opsonin and pattern-recognition receptors enable them to phagocytose approximately 40–50 bacteria/neutrophil (2), and their preformed granule proteins and capacity to produce high intravacuolar concentrations of reactive oxygen species (3) enable them to kill their bacterial prey. The extracellular release of DNA-histone antimicrobial protein–containing nets (4) that entrap bacteria and kill yeast/fungi (5) extends their bactericidal and fungicidal activities into their afterlives and further enhances their efficacy as guardians of tissue sterility.

Chemotherapeutic agents that inhibit neutrophil biogenesis (6–8) and/or reduce their bactericidal activity (7) produce neutropenia (i.e., <500,000 neutrophils/ml blood). Although neutropenia predisposes to infection, it has little direct effect on the sterility of blood because under most (9) — but not all (10) — conditions, hepatic and splenic macrophages are the cells primarily responsible for clearing bacteria from the circulation. Indeed, as suggested by Crosby (11) and supported by the studies of Wright et al. (12) and Koene et al. (13), it is the tissue neutrophil concentration (N) that determines whether the small number of bacteria that breech mucosal surfaces each day will find fertile soil for growth, or be engulfed and killed. Thus, while a blood N value of 5 × 105 neutrophils/ml is a call for vigilance, it is an imprecise measure of the likelihood of infection. This is so because blood is primarily the conduit by which neutrophils travel from bone marrow to tissues, and the blood N reflects the sum of the rates at which neutrophils are produced and released from bone marrow into the circulation, and the rates at which they are consumed in tissues (14) and/or recycled to spleen and bone marrow for destruction (15).

A critical concentration of neutrophils is required to produce a net reduction in bacterial concentration

Since Robertson and Sia’s 1924 report (16), most in vitro studies of neutrophil-bactericidal activity have used neutrophils stirred continuously for 15–120 minutes at 37°C in plasma- or serum-containing medium with known concentrations of bacteria. The neutrophils then are lysed, and the concentration of viable bacteria remaining is measured by bacteriological colony assay. This system mimics neutrophil-bactericidal activity in blood. Comparatively few investigators have studied neutrophil-bactericidal activity in tissues of living animals or in three-dimensional gels composed of extracellular matrix proteins. The findings that extracellular matrix proteins exert potent physiological effects on phagocytes (17, 18), and that the chemoattractant N-formyl-methionine-leucine-phenylalanine (fMLP) (19) inhibits neutrophil migration in fibrin-containing matrices, but not in collagen matrices, prompted Li et al. to examine neutrophil-bactericidal activity in tissues and tissue-like environments (20). They discovered that neutrophil-bactericidal activity in dermis of living rabbits and in fibrin gels is controlled by N and is independent of bacterial concentration (B), up to approximately 106 CFU/ml (20). They derived an equation that describes bacterial killing as a function of N and bacterial growth rate (Figure 1A) and showed that it accurately described neutrophil-bactericidal activity in suspension (21), fibrin gels, and dermis of living rabbits (20). Additionally, they found the ratio of the experimentally determined bacterial growth rate (g) and neutrophil-bacterial killing rate constant (k) defined a new and extremely useful parameter: the critical neutrophil concentration (CNC), namely, the N required to hold B constant (Figure 1A and refs. 20, 21). At N values below the CNC, B increases; at N above the CNC, B decreases. The CNC in stirred suspensions is approximately 4 × 105 neutrophils/ml (21), very close to the clinically determined value of 5 × 105 neutrophils/ml that defines neutropenia and characterizes neutropenic patients in danger of infection (6–8). The CNC in fibrin gels and rabbit dermis is approximately 1 × 106 and 8 × 106 neutrophils/ml, respectively (20), 2- to 20-fold larger than in stirred suspensions. Apparently, evolution has taken note of the CNC, as the blood neutrophil concentration is no less than approximately 5 × 105 neutrophils/ml in all adult vertebrates (22–24).

Comparative behaviors of bacterial populations at or near the CNC, as prediFigure 1

Comparative behaviors of bacterial populations at or near the CNC, as predicted by the equations of Li et al. (20, 21) and Malka et al. (25). (A) Li et al. derived an equation that includes an experimentally determined bacterial killing constant and the bacterial growth rate (20, 21). It predicts the CNC is constant at all B values. Their experiments show, however, that the bacterial growth rate and killing constant decrease, and the CNC increases (curved dashed line), at B values greater than 107 CFU/ml in stirred suspension and greater than 106 CFU/ml in fibrin gel (not shown). Most importantly, this equation predicts B will decrease or increase at all N values below or above the CNC, respectively, and remain constant at the CNC. (B) In contrast, as reported here by Malka et al., the bacterial population is bistable at N at or very near the CNC, creating favorable conditions for uncontrolled bacterial growth (25). N1 and N2 and the dashed line indicate the N and B values that define the zone of bistability in the bacterial population. For N below the CNC (solid yellow line), the bacterial population grows exponentially. Note also that in fibrin gels and rabbit dermis, the CNC at all N values is at least 2-fold larger than in stirred suspensions, a reflection of the effect of the former environment on the bacterial killing constant.

What is the physical meaning of the CNC? The answer is evident from the value and units of k (ml/neutrophil/min), which for these purposes can be viewed as the volume of medium or tissue a neutrophil can explore per minute. For neutrophils in stirred suspension and in dermis of living rabbits, k is equivalent to 2 × 10–8 and 4 × 10–9 ml/neutrophil/min, respectively. In stirred suspension, 1 neutrophil can explore approximately 20 pl medium/min, whereas in rabbit dermis, it can only explore approximately 4 pl dermis/min. Accordingly, 5 × 106 neutrophils/ml — the average concentration in human blood — is required for these cells to explore 1 ml of a stirred bacterial suspension in 10 minutes (i.e., 2 × 10–8 ml/neutrophil/min × 5 × 106 neutrophils × 10 min = 1 ml). Likewise, N of 2.5 × 107 neutrophils/ml or neutrophils/g rabbit dermis is required for neutrophils to explore 1 milliliter or gram of dermis in 10 minutes. Higher concentrations of neutrophils or longer exploration times are required when B exceeds 107 per milliliter stirred suspension (21) or 106 per milliliter or gram of tissue (20).

Dynamic system model of neutrophil bactericidal activity: effects of bistability

The equation Malka et al. report here (25) and previously (26), namely dB(t)/dt = ρB(t)/[1 + βB(t)] + s – δB(t) — (αNB(t)/[1 + γB(t) + ηN]), more accurately describes the rate of bacterial clearance than does the one previously reported by Li et al. (20, 21). This is because Malka et al. included the saturation rate for bacterial growth (β) and the spontaneous rate of bacterial death [δB(t)]. The bacterial growth rate (ρ) behaves like a carrying capacity and saturates (βB) at high B values. The authors also included in their equation the rates of bacterial influx into the circulation or site of infection(s) and of spontaneous bacterial death. If these rates could be measured, their inclusion would add significantly to the predictive power of the equation; however, there are no established methods at present for measuring them. Hence, Malka et al. assign them a value of 0, thereby eliminating this difference between the two groups’ equations.

The equation’s quadratic term represents the killing rate (–αNB(t)/[1 + γB(t) + ηN]). It involves the law of mass action: the neutrophil-bactericidal rate is proportional (α) to the product of N and B multiplied by the amount of time (t; in minutes) the neutrophils interact with bacteria. It saturates at high concentrations of both neutrophils (ηN) and bacteria (γB). The system yields two very different outcomes depending on the magnitude of the initial parameters: it is monostable when N is well above or below the critical concentration, but becomes bistable when N fluctuates close to the CNC. Bistability is a characteristic of dynamic systems in which two stable coexisting fixed points exist for a range of parameters (Figure 1B), and the system’s outcome depends on its initial conditions.

Unresolved questions

Although the equation derived by Malka et al. (25, 26) provides a more complete description of neutrophil-bactericidal activity than that of Li et al. (21), additional experiments are needed to assess its clinical utility. First, the conclusion that there is substantial patient-to-patient variability in neutrophil-bactericidal activity relies on studies of the staphylococcicidal activity of neutrophils from only four healthy adults (25). Li et al. observed an approximately 2-fold variation in values of k and CNC for neutrophils from more than 20 healthy adult donors killing S. epidermidis in stirred solutions (21), and variations in these parameters of similar magnitude for neutrophils killing S. aureus, E. coli, and P. aeruginosa in stirred suspensions in studies are reported by others. Variations in neutrophil-bactericidal activity of this magnitude near or at the CNC are certainly sufficient to produce bistability in the bacterial population. Nonetheless, additional measurements of the efficiency of killing of both Gram-positive and -negative bacteria by neutrophils, from a larger number of uninfected and infected eupenic and neutropenic donors than was tested by Malka et al., are needed to define precisely the frequency and magnitude of variation in bactericidal activity of neutrophils. Second, the authors measured neutrophil-bactericidal activity in stirred suspensions (25), a condition that mimics neutrophil-bactericidal activity in blood. However, neutrophil-bactericidal activity in tissues is of greatest relevance for neutropenic patients (8, 11–13). Thus, the method Malka et al. use might better predict the likelihood of infection in neutropenic patients if it employed fibrin gels to determine their neutrophils’ bactericidal activity. Li et al. showed that the kinetics of neutrophil-bactericidal activity in these gels closely mimics that for neutrophils in tissues in vivo (20). Third, Malka et al. ascribed the variability they observed in killing of S. aureus to differences in bactericidal activity of neutrophils from different healthy adult donors (25). However, their studies used 10% autologous serum, close to the concentration at which opsonin concentration becomes limiting (21, 27); 20%–40% serum would have been a better choice (21). Moreover, Malka et al. did not control for variations in opsonin concentration in each leukocyte donor’s serum. Thus, the variations in killing efficiencies observed could reflect variations in serum opsonin concentrations (27, 28), not in neutrophil bactericidal activity. From both mechanistic and practical standpoints, this is an important matter to resolve.

Clinical experience now suggests that the use of G-CSF to prevent bone marrow–depressive effects of chemotherapy in patients at high risk for neutropenia is far more efficacious than using it to treat neutropenia once it has occurred (8, 29). Accordingly, it may be that the most useful clinical application of this new equation will be to use it prospectively in conjunction with other predictive tools (29), including low-tech methods such as the one devised by Wright et al. (12), to assess the likelihood that chemotherapy will cause the tissue bactericidal activity of a patient’s neutrophils to fall below the CNC and/or into regions of bacterial bistability.

These concerns notwithstanding, we believe the findings that neutrophil-bactericidal activity can be described by equations very similar to those describing enzyme-substrate interactions provide an entirely new and quantitative way of viewing cellular defense against microbial pathogens. Budhu et al. reported that the equation described by Li et al. (21) also describes antigen-specific CD8+ T cell killing of cognate antigen–expressing melanoma cells in vitro and in living mice (30), which suggests that the same physical principles that govern neutrophil-bacterial interactions also govern the interactions of cytolytic lymphocytes with eukaryotic target cells. Indeed, it seems likely that these principles govern the effector activities of all immune cells. Consistent with this idea, Babbs (31) and Lejune et al. (32) report bistability of tumor cell growth under conditions in which the intratumoral concentration of tumor antigen-specific CD8+ T cells fluctuates around the critical CD8+ T cell concentration.

Conclusions

Malka et al. have derived a dynamic system model (25, 26) that describes neutrophil-bactericidal activity in stirred solutions more accurately than did its linear predecessor (20, 21). Although further work is needed to determine whether the variability in bactericidal activity described for neutrophils from normal adults applies to neutrophils from neutropenic and/or cytotoxic drug–treated patients, this model may enable clinicians to better distinguish patients who are at risk of infection from those who are not, thereby reducing the morbidity, mortality, and expense associated with unnecessary hospitalizations and/or administration of antibiotics and/or G-CSF.

Acknowledgments

This work was supported by NIH grant AI-20516-026 to S.C. Silverstein and by grants U54 CA121852-05 from the NIH and 1R01LM010140-01 from the National Library of Medicine to R. Rabadan. The authors note that two previous reviews of neutropenia have used the approximate title, “How many neutrophils are enough?” (11, 33); hence their use of “redux, redux” in this title.

Address correspondence to: Samuel C. Silverstein, Department of Physiology and Cellular Biophysics, Columbia University College of Physicians and Surgeons, 630 W. 168th Street, Rm. 11-411, New York, New York 10032, USA. Phone: 212.305.3546; Fax: 212.305.5775; E-mail: scs3@columbia.edu.

Footnotes

Conflict of interest: Raul Rabadan is a consultant for Novartis and the Geneva Foundation. Samuel C. Silverstein owns stock in Roche Pharmaceuticals and Teva Pharmaceuticals.

Reference information: J Clin Invest. 2012;122(8):2776–2779. doi:10.1172/JCI63939.

See the related article at Evidence for bistable bacteria-neutrophil interaction and its clinical implications.

References
  1. Muller WA. Mechanisms of transendothelial migration of leukocytes. Circ Res. 2009;105(3):223–230.
    View this article via: PubMed CrossRef Google Scholar
  2. Leijh PC, van den Barselaar MT, van Zwet TL, Dubbeldeman-Rempt I, van Furth R. Kinetics of phagocytosis of Staphyloccocus aureus and Escherichia coli by human granulocytes. Immunology. 1979;37(2):453–465.
    View this article via: PubMed Google Scholar
  3. Hampton MB, Kettle AJ, Winterboum CC. Inside the neutrophil phagosome: Oxidants, myeloperoxidase, and bacterial killing. Blood. 1998;92(9):3007–3017.
    View this article via: PubMed Google Scholar
  4. Ma AC, Kubes P. Platelets, neutrophils, and neutrophil extracellular traps (NETS) in sepsis. J Thromb Haemost. 2007;6(3):415–420.
    View this article via: PubMed CrossRef Google Scholar
  5. Urban CF, et al. Neutrophil extracellular traps contain calprotectin, a cytosolic protein complex involved in host defense against Candida albicans. PLoS Pathog. 2009;5(10):e1000639.
    View this article via: PubMed CrossRef Google Scholar
  6. Bodey GP, Buckley M, Sathe YS, Freireich EJ. Quantitative relationship between circulating leukocytes and infection in patients with acute leukemia. Ann Intern Med. 1966;64(2):328–340.
    View this article via: PubMed Google Scholar
  7. Lejeune M, Sariban E, Cantineaux B, Ferster A, Devalck C, Fondu P. Granulocyte functions in children with cancer are differentially sensitive to the toxic effect of chemotherapy. Pediatr Res. 1996;39(5):835–842.
    View this article via: PubMed CrossRef Google Scholar
  8. Crawford J, Dale DC, Lyman GH. Chemotherapy induced neutropenia. Cancer. 2003;100(2):228–243.
    View this article via: PubMed CrossRef Google Scholar
  9. Benacerraf B, Sebestyen MM, Schlossman S. A quantitative study of the kinetics of blood clearance of P32-labeled Escherichia coli and Staphylococci by the reticuloendothelial system. J Exp Med. 1959;110(1):27–48.
    View this article via: PubMed CrossRef Google Scholar
  10. Kurahashi K, et al. Depletion of phagocytes in the reticuloendothelial system causes increased inflammation and mortality in rabbits with Pseudomonas aeruginosa pneumonia. Am J Physiol Lung Cell Mol Physiol. 2009;296(2):L198–L209.
    View this article via: PubMed CrossRef Google Scholar
  11. Crosby WH. How many polys are enough? Arch Intern Med. 1969;123(4):722–723.
    View this article via: PubMed CrossRef Google Scholar
  12. Wright DG, Meierovics AI, Foxley JM. Assessing the delivery of neutrophils to tissues in neutropenia. Blood. 1986;67(4):1023–1030.
    View this article via: PubMed Google Scholar
  13. Koene HR, de Haas M, Kleijer M, Huizinga TW, Roos D, von dem Borne AE. Clinical value of soluble IgG Fc receptor type III in plasma from patients with chronic idiopathic neutropenia. Blood. 1998;91(1):1–7.
    View this article via: PubMed Google Scholar
  14. Watson RWG, Redmond HP, Wang JH, Condron C, Bouchier-Hayes D. Neutrophils undergo apoptosis following ingestion of Escherichia coli. J Immunol. 1996;156(10):3986–3992.
    View this article via: PubMed Google Scholar
  15. Furze RC, Rankin SM. The role of the bone marrow in neutrophil clearance under homeostatic conditions in the mouse. FASEB J. 2008;22(9):3111–3119.
    View this article via: PubMed CrossRef Google Scholar
  16. Robertson OH, Sia RHP. Studies on pneumococcus growth inhibition: ii. A method for demonstrating the growth-inhibitory and bactericidal action of normal serum-leucocyte mixtures. J Exp Med. 1924;39(2):219–244.
    View this article via: PubMed CrossRef Google Scholar
  17. Wright SD, Licht MR, Craigmyle LP, Silverstein SC. Communication between receptors for different ligands on a single cell: Ligation of fibronectin receptors induces a reversible alteration in the function of complement receptors on cultured human monocytes. J Cell Biol. 1984;99(1 pt 1):336–339.
    View this article via: PubMed CrossRef Google Scholar
  18. Nathan CF. Neutrophil activation on biological surfaces. J Clin Invest. 1987;80(6):1550–1560.
    View this article via: JCI PubMed CrossRef Google Scholar
  19. Li Y, et al. The bacterial peptide N-Formyl-Met-Leu-Phe (fMLP) inhibits killing of Staphylococcus epidermidis by human neutrophils in fibrin gels. J Immunol. 2002;168(2):816–824.
    View this article via: PubMed Google Scholar
  20. Li Y, Karlin A, Loike J, Silverstein SC. A critical concentration of neutrophils is required to block growth of Staphylococcus epidermidis in fibrin gels. J Exp Med. 2004;200(5):613–622.
    View this article via: PubMed CrossRef Google Scholar
  21. Li Y, Karlin A, Loike JD, Silverstein SC. A critical concentration of neutrophils is required for efficient bacterial killing in suspension. Proc Natl Acad Sci U S A. 2002;99(12):8289–8294.
    View this article via: PubMed CrossRef Google Scholar
  22. von Vietinghoff S, Ley K. Homeostatic regulation of blood neutrophil counts. J Immunol. 2008;181(8):5183–5188.
    View this article via: PubMed Google Scholar
  23. Orun I, Dorusu M, Yaziak H. Haematological parameters of three Cyprinid fish species from Karakaya Dam Lake, Turkey. J Biol Sci. 2003;3(3):320–328.
    View this article via: CrossRef Google Scholar
  24. Cutts LH, Krause WJ. Leukocytes in the blood of the developing opossum. J Anat. 1980;130(pt 1):113–120.
    View this article via: PubMed Google Scholar
  25. Malka R, Wolach B, Gavrieli R, Shochat E, Rom-Kedar V. Evidence for bistable bacteria-neutrophil interaction and its clinical implications. J Clin Invest. 2012;122(8):3002–3011.
    View this article via: JCI Google Scholar
  26. Malka R, Schochat E, Rom-Kedar V. Bistability and bacterial infections. PLoS One. 2010;5(5):1–10.
    View this article via: PubMed Google Scholar
  27. Johnston RB Jr, Klemperer MR, Alper CA, Rosen FS. The enhancement of bacterial phagocytosis by serum. The role of complement components and two cofactors. J Exp Med. 1969;129(6):1275–1290.
    View this article via: PubMed CrossRef Google Scholar
  28. Brouwer N, et al. Mannose-binding lectin (MBL) substitution: Recovery of opsonic function in vivo lags behind MBL serum levels. J Immunol. 2009;183(5):3496–3504.
    View this article via: PubMed CrossRef Google Scholar
  29. Aapro M, Crawford J, Kamioner D. Prophylaxis of chemotherapy-induced febrile neutropenia with granulocyte colony-stimulating factors: where are we now? Support Care Cancer. 2010;18(5):529–541.
  30. Budhu S, et al. CD8+ T-cell concentration determines the efficiency of killing of cognate antigen-expressing syngeneic mammalian cells in vitro and in tissues. J Exp Med. 2010;207(1):223–235.
    View this article via: PubMed CrossRef Google Scholar
  31. Charles FB. Predicting success or failure of immunotherapy for cancer: insights from a clinically applicable mathematical model. Am J Cancer Res. 2012;2(2):204–213.
    View this article via: PubMed Google Scholar
  32. Lejeune O, Chaplain MAJ, El Akili I. Oscillations and bistability in the dynamics of cytotoxic reactions mediated by the response of immune to solid tumors. Math Comput Model. 2008;47(5–6):649–662.
    View this article via: CrossRef Google Scholar
  33. Dale DC, Liles C. How many neutrophils are enough? Lancet. 1998;351(9118):1752–1753.
    View this article via: PubMed CrossRef Google Scholar
Version history
  • Version 1 (July 23, 2012): No description
  • Version 2 (August 1, 2012): No description
Advertisement
Advertisement