NEWS & VIEWS RESEARCH regulate body temperature. Although Puente and colleagues single out changes in atmospheric oxygen as the cause of cardiomyocyte cell-cycle withdrawal, other factors — such as metabolic stress resulting from periods of starvation between birth and the initiation of nursing or intermittent feeding — could also contribute to cardiomyocyte maturation. However, the finding that oxygen-level manipulation is sufficient to affect the timing of cellcycle withdrawal demonstrates that oxygen is a powerful stimulus in the developing heart. Much of what is known of neonatal development is based on research in mice, which have a gestation period of 19–21 days. Humans are born with more-mature organs than mice, after a nine-month gestation period. The observation that neonatal mice can regenerate part of the heart after injury3 has led to much speculation over potential treatment of human congenital heart defects. However, little is known about how human cardiomyocytes proliferate and mature during late fetal or neonatal development, and it is not known to what extent the development of the newborn mouse heart corresponds to that of a human heart. It is possible that exposure to atmospheric oxygen at birth has a comparable effect on human cardiomyocytes and their potential to regenerate, but studies of large animals with longer gestation periods than mice, or of nonhuman primates, will be necessary to determine whether there is likely to be comparable flexibility in heart growth or regeneration in the first weeks of human life. Oxidative stress at birth leads to the production of chemically reactive molecules called reactive oxygen species (ROS) and induction of a DNA-damage response. Puente et al. identify these processes as the stimuli for cardiomyo cyte cell-cycle withdrawal in a high-oxygen environment in newborn mice, and show that inhibition of oxidative stress through treatment with a ROS scavenger (N-acetylcysteine) prolongs the ability of the mouse heart to regenerate to 21 days after birth4, compared with the normal 7-day window3. But even this window of proliferation closed after this point, indicating that inhibition of oxidative stress can delay, but not prevent, cardiomyocyte cell-cycle withdrawal. Delayed withdrawal for 1–2 weeks after birth has been observed in several mouse models, but in most cases the cells eventually stop dividing2. The overriding mechanism that ultimately prevents cardiomyocytes from proliferating is not fully understood and remains an area of intense research interest. Although Puente and colleagues show that ROS scavenging can prolong the heart’s regenerative capacity in newborn mice, they do not test whether this treatment promotes regeneration of the adult mouse heart after injury. Antioxidant therapy using the same ROS scavenger has been tested in humans as a means of reducing muscle damage after a heart attack5, but limited benefit was observed, and
this therapeutic approach is not in wide use. Because adult heart muscle is distinct from that of neonates in terms of metabolism and cell-cycle regulation, it seems unlikely that, under conditions of oxidative stress, adult cardiomyocytes will show an increased regenerative response to ROS scavenging. However, this is certainly an area worthy of future investigation. It is interesting to speculate that the evolutionary transition to terrestrial life in atmospheric oxygen led to the inability of mammals to regenerate heart muscle. Zebrafish inhabit a relatively hypoxic environment and have a greater capacity to regenerate their hearts and limbs than mammals. Although it may not be possible to test the effects of atmospheric levels of oxygen or hyperoxia in fish, manipulation of oxygen levels in amphibians, which can also regenerate their hearts, could be used to test this hypothesis3. The definitive test might be to examine heart regeneration in terrestrial lizards, which can regenerate their tails (although the new tail is not exactly the same as the original)6.
In mammals, it seems that exposure to atmospheric oxygen at birth leads to development of the more powerful muscle cells needed in a closed circulatory system, with the tradeoff being a relative inability to regenerate heart muscle after injury or disease in adulthood. But further understanding of the transitions that occur in newborn mammals, such as that provided by Puente et al., could reveal approaches by which to rejuvenate or repair diseased hearts. ■ Katherine E. Yutzey is in the Division of Molecular Cardiovascular Biology, Cincinnati Children’s Medical Center, Cincinnati, Ohio 45229, USA. e-mail: [email protected] 1. Xin, M., Olson, E. N. & Bassel-Duby, R. Nature Rev. Mol. Cell Biol. 14, 529–541 (2013). 2. Kikuchi, K. & Poss, K. D. Annu. Rev. Cell Dev. Biol. 28, 719–741 (2012). 3. Porrello, E. R. et al. Science 331, 1078–1080 (2011). 4. Puente, B. N. et al. Cell 157, 565–579 (2014). 5. Sochman, J. J. Am. Coll. Cardiol. 39, 1422–1428 (2002). 6. Ritzman, T. B. et al. Anat. Rec. 295, 1596–1608 (2012).
I M M U N O LO GY
To affinity and beyond Tracking B cells in germinal centres — hotspots of B-cell proliferation and mutation during an immune response — reveals that those cells presenting the most antigen on their surface are programmed to dominate. See Letter p.637 D AV I D M . TA R L I N T O N
O
n page 637 of this issue, Gitlin et al.1 provide compelling insight into the processes driving the extraordinary increases that can occur in the affinity of antibodies for their target antigen molecules during vertebrate immune responses. The study resolves the question of how the rare B cells that produce high-affinity antibodies, which arise at random in a sea of average-affinity B cells, can dominate the response numerically within days. The answer is proliferation, but not through increased frequency or speed in the cells’ basic response cycle of ‘proliferate, mutate and select’. Instead, they are programmed by T cells, on the basis of the amount of antigen presented on the B-cell surface, to divide more times than lesser-affinity B cells in between rounds of affinity-based selection. During an immune response, B cells undergo random mutations in their immunoglobulin variable-region (IgV) genes, which encode the antigen-binding region of antibody proteins. B cells in which this somatic (non-germline) mutation results in enhanced antibody affinity are selected for further rounds of proliferation2. This process of affinity maturation occurs in
germinal centres (GCs) — regions of high cell density in secondary lymphoid organs, such as the lymph nodes and spleen, that form in the early stages of an immune response. The characterization of an enzyme known as activation-induced cytidine deaminase3 provided a biochemical mechanism for the somatic mutation of IgV genes. Subsequently, experiments in many systems4,5 suggested that affinity-based selection is probably driven by B-cell competition in the GC for interactions with follicular helper T cells (TFH cells), which stimulate multiple signalling pathways6, including those involved in survival, division, differentiation and migration. But how changes in the affinity of a B cell for antigen are translated into changes in relative representation of that cell in the immune response remained unclear. Germinal centres form spatially distinct light and dark zones. B cells in the light zone are considered to be neither dividing nor undergoing IgV-gene mutation, but to be receiving signals from TFH cells and thus being subjected to affinity-based selection. Darkzone B cells, by contrast, are thought of as the survivors of previous rounds of selection, and to be actively proliferating and mutating. Affinity maturation occurs by B cells cycling 2 9 M AY 2 0 1 4 | VO L 5 0 9 | NAT U R E | 5 7 3
© 2014 Macmillan Publishers Limited. All rights reserved
RESEARCH NEWS & VIEWS Increasing affinity, increasing antigen presentation
Dying B cell
TFH cell B cell IgV gene Light zone Dark zone
Mutated IgV gene
Figure 1 | Selection in the germinal centre. During an immune response to a specific antigen, B cells undergo antibody-affinity maturation in germinal centre (GC) regions of the lymphoid organs. This process involves the introduction of mutations in the B cells’ IgV genes as the cells divide in the dark zone of the GC. The B cells then migrate to the GC light zone, where interactions with follicular helper T cells (TFH cells) determine which cells re-enter the dark zone — B cells in which IgV-gene mutation leads to the generation of higher-affinity antibodies survive and migrate, whereas those with diminished affinity die. Gitlin et al.1 show that the amount of antigen captured by a light-zone B cell and presented on its surface to TFH cells determines the number of rounds of division and mutation that the cell subsequently undergoes in the dark zone, and the time taken for this, before re-migration. Because higher-affinity cells, which correlate with more mutations, capture more antigen, this results in a feed-forward mechanism, ensuring that these cells dominate the response.
from zone to zone, undergoing proliferation and mutation, and then selection, as they move between zones. But what actually happens to a positively selected B cell in the light zone to promote its dominance? Is it subject to enhanced survival, increased movement between zones or enhanced proliferation that bypasses the cyclical routine? More pragmatically, how can affinity maturation be analysed, when it may occur at a low frequency with unpredictable timing, and when the high-affinity cells thus generated may be identifiable only retrospectively and as a population, making their behaviour around the time of selection difficult to follow? Gitlin et al. address these very questions. First, they studied the effect, in mice, of delivering extra antigen to a small minority of B cells in established GCs. They observed increased proliferation of the boosted antigen-specific B cells, initially in the dark zone and to an extent proportional to the amount of antigen delivered. Although this clearly demonstrated that extra antigen enhances the proliferation of B cells in the dark zone, it did not distinguish between an acceleration of the normal cyclic migration of B cells between zones and an increased number of divisions per dark-zone B cell before migration to the light zone. To address this issue, the authors isolated B cells from GCs and quantified the relative proportion in the ‘actively dividing’ or ‘noncycling’ phases of the cell cycle. They found unboosted B cells in the actively dividing phases in both GC zones and observed that, once in the
dark zone, the cells underwent, on average, two cell divisions. Boosting with antigen increased the proportion of actively dividing B cells in the dark zone from 60% to almost 90%, while halving the migration of cells from the dark zone to the light zone. Together, these findings show that B cells undergoing selection in the light zone can be programmed to divide in the dark zone a variable number of times on the basis of their previous interactions with TFH cells; such interactions are, in turn, determined by the amount of antigen the B cell presents to the TFH cells (Fig. 1). These in vivo observations are strikingly similar to the finding7 that the number of divisions that B cells undergo in vitro is dictated by the strength of the signal passing through the CD40 pathway — one of the drivers of B-cell behaviour in GCs that is provided by TFH cells. To relate these observations to ‘real-world’ immune responses, Gitlin and colleagues applied their GC B-cell division tracking to cases in which antigen availability was not manipulated after the initial immunization. They found that the B cells undergoing the most proliferation had six times more affinityenhancing mutations in their IgV genes than the least-proliferating cells. Cells undergoing the most proliferation also contained the highest number of somatic mutations. These findings confirmed the relationship between proliferation, mutation and affinity maturation. The ultimate conclusion of this work is that affinity maturation works through a feed-forward mechanism, in which improved affinity
5 7 4 | NAT U R E | VO L 5 0 9 | 2 9 M AY 2 0 1 4
© 2014 Macmillan Publishers Limited. All rights reserved
begets a ‘stronger’ TFH-cell signal that begets increased dark-zone B-cell proliferation and mutation, begetting further improvements in affinity, antigen acquisition and thus even more proliferation and mutation. Such a mechanism allows an ever more rapid expansion of highaffinity cells in the population, overwhelming both low-affinity and nonspecific B cells. Is this the end of the road for GC biology? Thankfully, no. Although it provides extra ordinary insight into the mechanics of B-cell dominance, Gitlin and colleagues’ study leaves unexamined the means by which such favoured GC B cells extricate themselves from the GC and become circulating memory B cells and long-lived antibody-secreting cells in the bone marrow. There is evidence for stringent, affinity-based selection of GC B cells into the long-lived bone-marrow population8, but how such selective interactions differ from those measured by the authors remains unclear. Moreover, it has been shown9 that inter actions with TFH cells mediate the differentiation of GC B cells into antibody-secreting cells, which cease affinity maturation and leave the GC to migrate to the bone marrow. So it is puzzling that when GC B cells are provided with almost unlimited amounts of TFH-cell-derived signals, as in Gitlin and colleagues’ study, they prefer to divide rather than differentiate. Perhaps this is an issue of kinetics, an incorrect hypothesis of what induces differentiation or an indication of nuanced types of signalling through TFH cells or other, regulatory, T cells. Resolving such issues will further our understanding of how the orchestration of B-cell behaviour in GCs helps to optimize immediate and long-term immune protection. ■ David M. Tarlinton is at the Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria 3052, Australia, and in the Department of Medical Biology, University of Melbourne, Parkville. e-mail: [email protected] 1. Gitlin, A. D., Shulman, Z. & Nussenzweig, M. C. Nature 509, 637–640 (2014). 2. MacLennan, I. C. Annu. Rev. Immunol. 12, 117–139 (1994). 3. Muramatsu, M. et al. Cell 102, 553–563 (2000). 4. Victora, G. D. & Nussenzweig, M. C. Annu. Rev. Immunol. 30, 429–457 (2012). 5. Allen, C. D. C. et al. Nature Immunol. 5, 943–952 (2004). 6. Crotty, S. Annu. Rev. Immunol. 29, 621–663 (2011). 7. Turner, M. L., Hawkins, E. D. & Hodgkin, P. D. J. Immunol. 181, 374–382 (2008). 8. Smith, K. G. C., Light, A., Nossal, G. J. V. & Tarlinton, D. M. EMBO J. 16, 2996–3006 (1997). 9. Tarlinton, D. & Good-Jacobson, K. Science 341, 1205–1211 (2013).
CORRECTION In the News & Views article ‘Sensory biology: Radio waves zap the biomagnetic compass’ by Joseph L. Kirschvink (Nature 509, 296–297; 2014), Figure 1a was wrongly credited. It should have been credited to Marianne Hanzlik.
this therapeutic approach is not in wide use. Because adult heart muscle is distinct from that of neonates in terms of metabolism and cell-cycle regulation, it seems unlikely that, under conditions of oxidative stress, adult cardiomyocytes will show an increased regenerative response to ROS scavenging. However, this is certainly an area worthy of future investigation. It is interesting to speculate that the evolutionary transition to terrestrial life in atmospheric oxygen led to the inability of mammals to regenerate heart muscle. Zebrafish inhabit a relatively hypoxic environment and have a greater capacity to regenerate their hearts and limbs than mammals. Although it may not be possible to test the effects of atmospheric levels of oxygen or hyperoxia in fish, manipulation of oxygen levels in amphibians, which can also regenerate their hearts, could be used to test this hypothesis3. The definitive test might be to examine heart regeneration in terrestrial lizards, which can regenerate their tails (although the new tail is not exactly the same as the original)6.
In mammals, it seems that exposure to atmospheric oxygen at birth leads to development of the more powerful muscle cells needed in a closed circulatory system, with the tradeoff being a relative inability to regenerate heart muscle after injury or disease in adulthood. But further understanding of the transitions that occur in newborn mammals, such as that provided by Puente et al., could reveal approaches by which to rejuvenate or repair diseased hearts. ■ Katherine E. Yutzey is in the Division of Molecular Cardiovascular Biology, Cincinnati Children’s Medical Center, Cincinnati, Ohio 45229, USA. e-mail: [email protected] 1. Xin, M., Olson, E. N. & Bassel-Duby, R. Nature Rev. Mol. Cell Biol. 14, 529–541 (2013). 2. Kikuchi, K. & Poss, K. D. Annu. Rev. Cell Dev. Biol. 28, 719–741 (2012). 3. Porrello, E. R. et al. Science 331, 1078–1080 (2011). 4. Puente, B. N. et al. Cell 157, 565–579 (2014). 5. Sochman, J. J. Am. Coll. Cardiol. 39, 1422–1428 (2002). 6. Ritzman, T. B. et al. Anat. Rec. 295, 1596–1608 (2012).
I M M U N O LO GY
To affinity and beyond Tracking B cells in germinal centres — hotspots of B-cell proliferation and mutation during an immune response — reveals that those cells presenting the most antigen on their surface are programmed to dominate. See Letter p.637 D AV I D M . TA R L I N T O N
O
n page 637 of this issue, Gitlin et al.1 provide compelling insight into the processes driving the extraordinary increases that can occur in the affinity of antibodies for their target antigen molecules during vertebrate immune responses. The study resolves the question of how the rare B cells that produce high-affinity antibodies, which arise at random in a sea of average-affinity B cells, can dominate the response numerically within days. The answer is proliferation, but not through increased frequency or speed in the cells’ basic response cycle of ‘proliferate, mutate and select’. Instead, they are programmed by T cells, on the basis of the amount of antigen presented on the B-cell surface, to divide more times than lesser-affinity B cells in between rounds of affinity-based selection. During an immune response, B cells undergo random mutations in their immunoglobulin variable-region (IgV) genes, which encode the antigen-binding region of antibody proteins. B cells in which this somatic (non-germline) mutation results in enhanced antibody affinity are selected for further rounds of proliferation2. This process of affinity maturation occurs in
germinal centres (GCs) — regions of high cell density in secondary lymphoid organs, such as the lymph nodes and spleen, that form in the early stages of an immune response. The characterization of an enzyme known as activation-induced cytidine deaminase3 provided a biochemical mechanism for the somatic mutation of IgV genes. Subsequently, experiments in many systems4,5 suggested that affinity-based selection is probably driven by B-cell competition in the GC for interactions with follicular helper T cells (TFH cells), which stimulate multiple signalling pathways6, including those involved in survival, division, differentiation and migration. But how changes in the affinity of a B cell for antigen are translated into changes in relative representation of that cell in the immune response remained unclear. Germinal centres form spatially distinct light and dark zones. B cells in the light zone are considered to be neither dividing nor undergoing IgV-gene mutation, but to be receiving signals from TFH cells and thus being subjected to affinity-based selection. Darkzone B cells, by contrast, are thought of as the survivors of previous rounds of selection, and to be actively proliferating and mutating. Affinity maturation occurs by B cells cycling 2 9 M AY 2 0 1 4 | VO L 5 0 9 | NAT U R E | 5 7 3
© 2014 Macmillan Publishers Limited. All rights reserved
RESEARCH NEWS & VIEWS Increasing affinity, increasing antigen presentation
Dying B cell
TFH cell B cell IgV gene Light zone Dark zone
Mutated IgV gene
Figure 1 | Selection in the germinal centre. During an immune response to a specific antigen, B cells undergo antibody-affinity maturation in germinal centre (GC) regions of the lymphoid organs. This process involves the introduction of mutations in the B cells’ IgV genes as the cells divide in the dark zone of the GC. The B cells then migrate to the GC light zone, where interactions with follicular helper T cells (TFH cells) determine which cells re-enter the dark zone — B cells in which IgV-gene mutation leads to the generation of higher-affinity antibodies survive and migrate, whereas those with diminished affinity die. Gitlin et al.1 show that the amount of antigen captured by a light-zone B cell and presented on its surface to TFH cells determines the number of rounds of division and mutation that the cell subsequently undergoes in the dark zone, and the time taken for this, before re-migration. Because higher-affinity cells, which correlate with more mutations, capture more antigen, this results in a feed-forward mechanism, ensuring that these cells dominate the response.
from zone to zone, undergoing proliferation and mutation, and then selection, as they move between zones. But what actually happens to a positively selected B cell in the light zone to promote its dominance? Is it subject to enhanced survival, increased movement between zones or enhanced proliferation that bypasses the cyclical routine? More pragmatically, how can affinity maturation be analysed, when it may occur at a low frequency with unpredictable timing, and when the high-affinity cells thus generated may be identifiable only retrospectively and as a population, making their behaviour around the time of selection difficult to follow? Gitlin et al. address these very questions. First, they studied the effect, in mice, of delivering extra antigen to a small minority of B cells in established GCs. They observed increased proliferation of the boosted antigen-specific B cells, initially in the dark zone and to an extent proportional to the amount of antigen delivered. Although this clearly demonstrated that extra antigen enhances the proliferation of B cells in the dark zone, it did not distinguish between an acceleration of the normal cyclic migration of B cells between zones and an increased number of divisions per dark-zone B cell before migration to the light zone. To address this issue, the authors isolated B cells from GCs and quantified the relative proportion in the ‘actively dividing’ or ‘noncycling’ phases of the cell cycle. They found unboosted B cells in the actively dividing phases in both GC zones and observed that, once in the
dark zone, the cells underwent, on average, two cell divisions. Boosting with antigen increased the proportion of actively dividing B cells in the dark zone from 60% to almost 90%, while halving the migration of cells from the dark zone to the light zone. Together, these findings show that B cells undergoing selection in the light zone can be programmed to divide in the dark zone a variable number of times on the basis of their previous interactions with TFH cells; such interactions are, in turn, determined by the amount of antigen the B cell presents to the TFH cells (Fig. 1). These in vivo observations are strikingly similar to the finding7 that the number of divisions that B cells undergo in vitro is dictated by the strength of the signal passing through the CD40 pathway — one of the drivers of B-cell behaviour in GCs that is provided by TFH cells. To relate these observations to ‘real-world’ immune responses, Gitlin and colleagues applied their GC B-cell division tracking to cases in which antigen availability was not manipulated after the initial immunization. They found that the B cells undergoing the most proliferation had six times more affinityenhancing mutations in their IgV genes than the least-proliferating cells. Cells undergoing the most proliferation also contained the highest number of somatic mutations. These findings confirmed the relationship between proliferation, mutation and affinity maturation. The ultimate conclusion of this work is that affinity maturation works through a feed-forward mechanism, in which improved affinity
5 7 4 | NAT U R E | VO L 5 0 9 | 2 9 M AY 2 0 1 4
© 2014 Macmillan Publishers Limited. All rights reserved
begets a ‘stronger’ TFH-cell signal that begets increased dark-zone B-cell proliferation and mutation, begetting further improvements in affinity, antigen acquisition and thus even more proliferation and mutation. Such a mechanism allows an ever more rapid expansion of highaffinity cells in the population, overwhelming both low-affinity and nonspecific B cells. Is this the end of the road for GC biology? Thankfully, no. Although it provides extra ordinary insight into the mechanics of B-cell dominance, Gitlin and colleagues’ study leaves unexamined the means by which such favoured GC B cells extricate themselves from the GC and become circulating memory B cells and long-lived antibody-secreting cells in the bone marrow. There is evidence for stringent, affinity-based selection of GC B cells into the long-lived bone-marrow population8, but how such selective interactions differ from those measured by the authors remains unclear. Moreover, it has been shown9 that inter actions with TFH cells mediate the differentiation of GC B cells into antibody-secreting cells, which cease affinity maturation and leave the GC to migrate to the bone marrow. So it is puzzling that when GC B cells are provided with almost unlimited amounts of TFH-cell-derived signals, as in Gitlin and colleagues’ study, they prefer to divide rather than differentiate. Perhaps this is an issue of kinetics, an incorrect hypothesis of what induces differentiation or an indication of nuanced types of signalling through TFH cells or other, regulatory, T cells. Resolving such issues will further our understanding of how the orchestration of B-cell behaviour in GCs helps to optimize immediate and long-term immune protection. ■ David M. Tarlinton is at the Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria 3052, Australia, and in the Department of Medical Biology, University of Melbourne, Parkville. e-mail: [email protected] 1. Gitlin, A. D., Shulman, Z. & Nussenzweig, M. C. Nature 509, 637–640 (2014). 2. MacLennan, I. C. Annu. Rev. Immunol. 12, 117–139 (1994). 3. Muramatsu, M. et al. Cell 102, 553–563 (2000). 4. Victora, G. D. & Nussenzweig, M. C. Annu. Rev. Immunol. 30, 429–457 (2012). 5. Allen, C. D. C. et al. Nature Immunol. 5, 943–952 (2004). 6. Crotty, S. Annu. Rev. Immunol. 29, 621–663 (2011). 7. Turner, M. L., Hawkins, E. D. & Hodgkin, P. D. J. Immunol. 181, 374–382 (2008). 8. Smith, K. G. C., Light, A., Nossal, G. J. V. & Tarlinton, D. M. EMBO J. 16, 2996–3006 (1997). 9. Tarlinton, D. & Good-Jacobson, K. Science 341, 1205–1211 (2013).
CORRECTION In the News & Views article ‘Sensory biology: Radio waves zap the biomagnetic compass’ by Joseph L. Kirschvink (Nature 509, 296–297; 2014), Figure 1a was wrongly credited. It should have been credited to Marianne Hanzlik.
