If humans began systematically burning after they arrived, this w

If humans began systematically burning after they arrived, this would diminish the effects of fire as lighting

more fires increases their frequency but lowers their intensity, since fuel loads are not increased. Flannery (1994:230) suggested that the extinction of large herbivores preceded large scale burning in Australia and the subsequent increase in fuel loads from unconsumed vegetation set the stage for the “fire-loving plant” communities that dominate the continent today. A similar process may have played out much later in Madagascar. Burney et al. (2003) used methods similar to Gill et al. (2009) to demonstrate that CX-5461 manufacturer increases in fire frequency postdate megafaunal decline selleck kinase inhibitor and vegetation change, and are the direct result of human impacts on megafauna communities. Human-assisted extinctions of large herbivores in Madagascar, North America, and Australia, may all have resulted in dramatic shifts in plant communities and fire regimes, setting off a cascade of ecological changes that contributed to higher extinction rates. With the advent of agriculture, especially intensive agricultural

production, anthropogenic effects increasingly took precedence over natural climate change as the driving forces behind plant and animal extinctions (Smith and Zeder, 2013). Around much of the world, humans experienced a cultural and economic transformation from small-scale hunter–gatherers to larger and more complex agricultural communities. By the Early Holocene, domestication of plants and animals was underway in several regions including Southwest Asia, Southeast Asia, New Guinea, and parts of the Americas. Domesticates quickly spread from these centers or were invented independently with local wild plants and

Y-27632 order animals in other parts of the world (see Smith and Zeder, 2013). With domestication and agriculture, there was a fundamental shift in the relationship between humans and their environments (Redman, 1999, Smith and Zeder, 2013 and Zeder et al., 2006). Sedentary communities, human population growth, the translocation of plants and animals, the appearance and spread of new diseases, and habitat alterations all triggered an accelerating wave of extinctions around the world. Ecosystems were transformed as human subsistence economies shifted from smaller scale to more intensified generalized hunting and foraging and to the specialized and intensive agricultural production of one or a small number of commercial products. In many cases, native flora and fauna were seen as weeds or pests that inhibited the production of agricultural products. In tropical and temperate zones worldwide, humans began clearing large expanses of natural vegetation to make room for agricultural fields and grazing pastures.

In all animals exposed to alumina dust the presence of alumina cr

In all animals exposed to alumina dust the presence of alumina crystals in the lung (alveolar spaces and airways) was qualitatively evaluated under polarized light (Axioplan, Zeiss, Oberkochen, Germany) at

1000× magnification. The right lungs were homogenized in 1 mL of PBS with protease inhibitors (1 μg/mL leupeptin and 1 μg/mL pepstatin). Homogenates were centrifuged (Centrifuge 5415R, Hamburg, Germany, 4 °C, 6700 × g, 15 min) and then, the supernatant was collected for transforming growth factor beta (TGF-β) and interleukin-1beta (IL-1β) assays by ELISA (R&D Systems Inc., Minneapolis, MN, USA), according to the manufacturer’s protocol. Total protein concentration in lung homogenates was determined by Bradford’s method ( Bradford, 1976). Concentration of cytokines in lung homogenates selleck chemicals was further normalized to protein concentration in the samples and expressed PARP inhibitor as picograms per milligram of protein. Optical density was measured at 450 nm by a microplate reader (SpectraMax 190, Molecular Devices, Sunnyvale, CA, USA). The normality of the data and the homogeneity of variances were tested by Kolmogorov–Smirnov test with Lilliefors’ correction and Levene median test, respectively. In all instances both conditions were satisfied and parametric

tests were run. One-way ANOVA was used to compare the values of body weight measured every 7 days, throughout 4 weeks in each group. Weight differences between control and exercise groups at every 7 days were

evaluated by Student’s t-test. Two-way ANOVA was applied to the remaining parameters (factors: exercise and alumina). For all ANOVAs, the Student–Newman–Keuls Temsirolimus was used as a post hoc test. The morphometric data, originally expressed as percent, underwent an arcsine transformation, in order to generate a normal distribution. The statistical analyses were carried out by the SigmaStat 9.0 software (SYSTAT, Point Richmond, CA, USA). In all instances p < 0.05 was considered a statistically significant difference. Metal composition of alumina dust is presented in Table 1. A high concentration of the element Al, followed by Fe and Hg was found. Scanning electron micrographs of particles are shown in Fig. 1, demonstrating the frequency distribution of diameters of particle sample. 90% of particles diameter are under 150 μm, being 50% below 100 μm and 10% smaller than 57 μm. A progressive increase in body weight was observed along time in animals not submitted to physical exercise. In exercising group, a decrease in body weight occurred during the first week of aquatic training, but thereafter the values did not differ from those in control mice (Fig. 2). All mechanical parameters (ΔP2, ΔE and Est) but ΔP1 were higher after alumina dust exposure in animals not submitted to physical exercise. Additionally, exercise training before particle exposure caused no changes in resistive and viscoelastic components, but Est increased in this group ( Fig. 3).

Moira Elizabeth Schöttler and Scientific Linguagem for their assi

Moira Elizabeth Schöttler and Scientific Linguagem for their assistance in editing the manuscript. Carfilzomib mouse
“Asthma is a chronic inflammatory disease affecting the airways and lung parenchyma (Bateman et al., 2008), associated with remodeling characterized by the following ultrastructural changes: subepithelial fibrosis, mucous metaplasia, airway wall thickening, smooth muscle

cell hypertrophy and hyperplasia, myofibroblast hyperplasia, vascular proliferation, and extracellular matrix abnormalities (Al-Muhsen et al., 2011). These changes accelerate decline in lung function (Holgate, 2008) despite treatment with corticosteroids. Since lung remodeling is usually related to established inflammation, it may be hypothesized that early treatment with immunoregulatory agents could prevent damage. Recent studies have demonstrated the Bacillus Calmette–Guérin (BCG) vaccine to be effective at reducing inflammation and hyperresponsiveness in animal models (Lagranderie et al., 2008) and in humans with asthma (Choi and Koh, 2002, Choi and Koh, 2003 and Cohon et al., 2007). However, the effectiveness of this treatment seems to be affected

by aspects of vaccine delivery: DAPT experimental studies report better control of the inflammatory process of asthma with intranasal administration compared to the intradermal route (Choi et al., 2007 and Erb et al., 1998), even though the latter is more commonly used in humans (Sarinho et al., 2010 and Shirtcliffe et al., 2004). Furthermore, there is controversy regarding the best time of BCG administration before induction of allergy (Erb et al., 1998, Nahori et al., 2001 and Ozeki et al., 2011). Additionally, a strain-dependent effect of BCG cannot be ruled out. Mirabegron In this line, the Moreau strain, which is widely used for tuberculosis control in Brazil (Benevolo-de-Andrade et al., 2005), has been observed to induce an adaptive immunity while increasing cytokines from T helper 1 (Th1) and regulatory T cells (Treg) (Wu et al., 2007), suggesting that this vaccine could be a potential tool for prevention of allergic asthma. Based on the aforementioned, we used

a murine model of allergic asthma to analyze the effects of different routes of administration and application times of the BCG-Moreau strain on pulmonary inflammation, remodeling process, and lung function. Moreover, possible mechanisms of action were investigated. This study was approved by the Ethics Committee of the Carlos Chagas Filho Institute of Biophysics, Health Sciences Center, Federal University of Rio de Janeiro, Brazil (CEUA-CCS, IBCCF 019). A total of 168 newly weaned male BALB/c mice (10–15 g) were randomly divided into two groups. The first group (n = 84) received 25 μL of a solution of 106 UFC lyophilized BCG Moreau strain resuspended in saline while the second group (n = 84) received saline. BCG or saline were intradermally (n = 42) or intranasally (n = 42) injected one or two months before the induction of allergic asthma.

The increase in channel slope, a metric of channel adjustment, le

The increase in channel slope, a metric of channel adjustment, leads to an increase in the shear stress available to transport sediment between an initial time (t1) when Robinson Creek was near the elevation of the current terrace surface and the present time (t2) with Robinson Creek characterized by incision. Assuming that PD0325901 supplier grain size distributions are similar at t1 and t2, using Eqs. (1) and (2) shows that the transport

capacity increased by about 22% and using equation 3 shows that the excess shear stress increased by 24% between t1 and t2. During the three-year period between 2005 and 2008, two segments of this reach showed significant changes in bed elevation (Fig. 11) in two locations. Downstream of Lambert Lane bridge, the thalweg lowered up to 0.7 m; in contrast, downstream of the Mountain View Road bridge, near the confluence with Anderson Creek, the thalweg aggraded up to 0.7 m. The sediment eroded from the channel in the zone Fluorouracil clinical trial that incised during the 2006 flood was likely transported downstream and deposited at the mouth of Robinson Creek—indicating spatial variability in geomorphic response to the same environmental

forcing factor. Changes in other portions of the study reach were less pronounced during this short period. The Robinson Creek case study illustrates the challenge of attribution of incision to a single extrinsic cause such as tectonic, climatic, or landuse changes. Tectonics is not considered a factor in the active incision of Robinson Creek; however,

climate variability and anthropogenic landuse changes are linked over similar temporal and spatial scales and it is difficult to separate their effects. Historical rain gage and paleo-records document that climate variability is a factor characterizing California’s north coastal region that operated before the “Anthropocene,” and it contributed to the landscape template the Euro-Americans encountered before agriculture, grazing, and logging activities began in Anderson Valley. However, oral histories indicate that incision and bank erosion in Robinson Creek occur during decadal floods, suggesting that California’s characteristic climate variability selleck chemicals llc facilitates incision processes. Nonetheless, because climate variability governed the region before the landuse-transformation of Anderson Valley, we hypothesize that anthropogenic disturbances were likely significant in initiating incision processes in Robinson Creek. Determining the validity of this assertion depends on the extent to which the timing for the initiation of incision can be accurately established. This task is a challenge in an ungagged watershed with limited consistent quantitative historical bed elevation measurements. Repetitive bridge cross section data from Anderson Creek (which represents the baselevel for Robinson Creek) suggest that incision of almost a meter has occurred since 1960.

anthropogenic conditions on both delta plain and delta front and

anthropogenic conditions on both delta plain and delta front and the examine how similar changes may affect maintenance of deltas

in general and wave-dominated SCH727965 chemical structure deltas in particular. The Danube delta, built in the northwestern Black Sea over the last ∼9000 years (Giosan et al., 2009), comprises of two distinct morphological regions (Antipa, 1915). The internal “fluvial delta” was constructed inside the former Danube Bay, whereas the external “marine delta” developed into the Black Sea proper once this paleo-bay was filled (Fig. 1). The modern delta plain preserves surface morphological elements as old as ∼5500 years indicating that sea level did not vary much since then and that subsidence has been minimal when considered at the scale of the whole delta (Giosan et al., 2006a and Giosan et al., 2006b). The fluvial delta is an amalgamation of river-dominated bayhead and lacustrine lobes characterized by networks of successively branching channels and numerous lakes (Fig. 1). Wave-dominated lobes, characterized by beach ridge and barrier plains composed of alongshore-oriented sand ridges, are typical for the marine delta (Fig. 1). Although the youngest region of the marine delta, Chilia III, started as a

river-dominated lobe, it has come under wave-dominance in the first half of 20th century when sediment delivered by Proteasome function Chilia branch became insufficient relative to its size (Giosan et al., 2005). Much of

the late development of the delta may be due to expansion of deforestation in the drainage basin in the last 1000 years (Giosan et al., 2012) leading to an overextended Danube delta. The high density of the fossil and active channel network (Fig. 1) suggests that after construction, the natural delta plain was fed by fluvial sediments through overbank flooding and avulsion in the fluvial sector, but primarily via minor overbank flooding in the marine sector. In the latter waves have tended to suppress avulsion and, thus, channel development (Bhattacharya and Giosan, 2003 and Swenson, 2005). The fluvial sediment delivery to the internal delta was probably relatively small compared to the sediment delivered to the coast Sclareol even with secondary channels present there. For example, Antipa (1915) described the internal delta after his comprehensive campaign of mapping it at the beginning of the last century as a “vast shallow lake” covered by floating reed islands and with marshes along its edges. Even today hundreds of lakes dot the fluvial delta (Giosan et al., 2005). Antipa’s “vast lake” was bounded by the high banks of the three large Danube distributaries (i.e., the Chilia, Sulina, and St. George from north to south) and the sand ridges of the marine delta, and internally segmented by the minor levees of some more prominent secondary channels.

, 2008) and the UK (Brown, 1997) However, many studies of alluvi

, 2008) and the UK (Brown, 1997). However, many studies of alluvial fills in both the Old World and New Worlds have revealed a mid or late Holocene (sensu Walker et al., 2012) hiatus in sedimentation that is both traceable within valleys and regionally. Although interpreted by the authors as evidence for climatic control on floodplain sedimentation, time-series of cumulative density functions of dates reveals not only peaks related to events or series of events but also an overall trend when these

dates are converted into rates ( Macklin et al., 2010; Fig. 2). All Holocene catchments have a Lateglacial check details inheritance which although dominated by climatic forcing (Gibbard and Lewin, 2002) may have been influenced to a minor extent by human activity (Notebaert and Verstraeten, 2010). Since catchment

size can be assumed to have remained constant during the Holocene it follows that changes in floodplain deposition must reflect the sum of the input of sediment to and export from the reach – the basis of the sediment budget approach to fluvial geomorphology. Allowing for geometric considerations, changes in the rate of sediment deposition within valley must then reflect changing inputs (Hoffmann et al., 2010). An important result of the occurrence of relatively small basins and relatively uniform erosion rates is buy NSC 683864 high levels of retention of anthropogenic sediments on the lower parts of hillslopes as colluvium or 0 order valleys (Brown, 2009 and Dotterweich et al.,

2013) and in 1st order valley floors (Brown and Barber, 1985 and Houben, 2003). In a recent study of a small catchment in Germany 62% of the sediment produced by 5000 years Tyrosine-protein kinase BLK of cultivation still resides in the catchment as colluvium amounting to 9425 t ha−1 (Houben, 2012). This represents an approximate average of 2.6 t ha−1 yr−1 (equivalent to 0.2 mm yr−1) which is close to the median for measured agricultural soil erosion rates (Montgomery, 2007b). Two small catchments are used here to show the existence of a major sedimentary discontinuity associated with human activity within two contrasting valley chronostratigraphies. The catchments of the Culm and Frome are both located in England but are 100 km apart. They are similar in size, altitude, relative relief and even solid geology (Table 1; Fig. 3). The methods used in both studies are standard sedimentary and palaeoecological analytical procedures and can be found in Brown et al. (2011) and will not be detailed here, except for the geophysical and GIS methodology which are outlined below. In both catchments sediment logging from bank exposures and coring was augmented by ground penetrating radar transects.

Hillslope failure, river channel widening, and/or construction ac

Hillslope failure, river channel widening, and/or construction activity may mobilize sediment from deeper (i.e., meters) sources. Aeolian deposition may be a third source, although

no evidence supports aeolian deposition as a significant source to the rivers studied here. The relative contributions from these sources may change both temporally and spatially in a river. These changes allow only limited AZD8055 price conclusions to be drawn from a single data point, limiting the success of a mitigation effort that is applied uniformly across a watershed. Contemporary sediment sources are frequently augmented and supplemented by legacy sediment. Legacy sediment comes from anthropogenic sources and activities, such as disturbances in land use/cover and/or surficial processes (James, 2013). For rivers, legacy sediments can originate from incised floodplains (Walter and Merritts, 2008), impoundments behind dams (Merritts et al., 2011), increased hillslope erosion due to historic deforestation (DeRose et al., 1993 and Jennings et al., 2003), and anthropogenic activities

such as construction learn more and land use changes (Wolman and Schick, 1967 and Croke et al., 2001). Legacy sediment can also deliver high loads of contaminants to river systems (Cave et al., 2005 and Lecce et al., 2008). The current supply of sediment is high (Hooke, 2000), as humans are one of the greatest current geomorphic agents. Consequently, combining legacy sediment with increased anthropogenic geomorphic activity makes it even more important to identify the source of sediments in rivers. Sediment sources can be distinguished until using the radionuclides lead-210 (210Pb) and cesium-137 (137Cs). 210Pb is a naturally-occurring isotope resulting from the decay of 238Uranium in rock to eventually 222Radon. This gas diffuses into the atmosphere and decays into excess 210Pb, which eventually settles to the ground. This diffusion process creates a fairly consistent level of excess 210Pb in

the atmosphere and minimizes local differences that exist in the production of radon. Rain and settling can subsequently result in the deposition of excess 210Pb, with a half-life of 22.3 years. This atmospheric deposition of excess 210Pb, is added to the background levels that originate from the decay of radon in the soil. “Excess” atmospheric 210Pb occurs because, if the material (in this case the sediment) is isolated from the source (i.e., the atmosphere), this level will decay and decrease in activity. As this excess 210Pb is then correlated with the time of surficial exposure, it is commonly used as a sediment tracer (e.g., D’Haen et al., 2012, Foster et al., 2007, Whiting et al., 2005 and Matisoff et al., 2002). 137Cs is also used as a sediment tracer, although its source is different. It is the byproduct of nuclear fission through reactors and weapon activities, and is not naturally found in the world.

In Japan, the main island of Honshu also has several sites that c

In Japan, the main island of Honshu also has several sites that contain obsidian obtained from Kozu Island (Izu Islands) by 32,000 years ago ( Habu, 2010). Overall, the evidence from Sunda and Sahul demonstrates

significant maritime voyaging, ocean navigation, and island colonization by the Late Pleistocene. Somewhat later in time, colonization of California’s Channel Islands at least 11,000 B.C. (all B.C./A.D./B.P. dates are calibrated calendar ages unless otherwise GW3965 manufacturer noted) required boats and was achieved by some of the earliest people to live in the Americas (Erlandson et al., 2011a and Erlandson et al., 2011b). Early coastal sites in California, elsewhere on the Pacific Rim, and in Chile have helped support the coastal migration theory for the initial peopling of the Americas (Erlandson et al., 2007). Colonization of several Mediterranean islands

occurs about this same time, with hunter-gatherers or early agriculturalists expanding to several islands and traveling to Melos to obtain obsidian during the Terminal Pleistocene and Early Holocene (Cherry, 1990, Patton, 1996 and Broodbank, 2006). During the Middle and Late Microbiology inhibitor Holocene, there is an explosion of maritime exploration and island colonization, facilitated by major advances in sailing and boat technology (Anderson, 2010). The Austronesian expansion of horticulturalists out of island Southeast Asia, through Near Oceania and into Remote Oceania (ca. 1350 B.C.) begins several millennia of island colonization in the vast Pacific, culminating in the Polynesian colonization of Hawaii, Easter Island, and New Zealand during the last millennium

(Kirch, 2000 and Anderson, 2010). Human settlement of Caribbean islands began at least 7000 years ago, initially by mafosfamide hunter-gatherers and later by horticulturalists expanding primarily, if not exclusively, out of South America (Keegan, 2000, Fitzpatrick and Keegan, 2007 and Wilson, 2007). In the North Atlantic, Mesolithic peoples began an expansion into the Faroes and elsewhere that increased during the Viking Age, with voyages to Iceland, Greenland, and northeast North America (see Dugmore et al., 2010 and Erlandson, 2010a). Other islands in southern Chile and Argentina, northeast Asia, the Indian Ocean, and beyond were all colonized by humans during the Holocene, each starting a new anthropogenic era where humans often became the top predator and driver of ecological change. A final wave of island colonization occurred during the era of European exploration, when even the smallest and most remote island groups were visited by commercial sealers, whalers, and others (Lightfoot et al., 2013). Early records of human colonization of islands are often complicated by a small number of archeological sites and fragmentary archeological record, which is hindered by interglacial sea level rise that left sites submerged offshore. Consequently, the early environmental history of colonization can be difficult to interpret.

Global deposits of relatively high 137Cs activity also correspond

Global deposits of relatively high 137Cs activity also correspond to the nuclear accidents in Chernobyl, Ukraine in 1986 and Fukushima, Japan in 2011. As its half-life of 30.2 years is similar to 210Pb, 137Cs is often used in parallel with excess 210Pb to identify the sources of sediment. Sediment derived from shallow, surficial erosion, such as through overland flow, would typically have higher amounts of excess 210Pb than sediment from deeper sources that have been isolated from the atmosphere for a longer time. Samples with higher activity readings of excess 210Pb indicate sources from upland/surface Icotinib supplier erosion, while samples with lower readings suggest sources from depths that have not recently

been exposed to the atmosphere (Feng et al., 2012). Surficial sources eroded in the uplands and/or floodplains contribute to higher activity levels. Deeper sources, with lower or nonexistent see more excess 210Pb levels, might come from sources that expose and transport sediment, such as hillslope failure or river bank erosion.

Many previous studies have used radionuclides to determine sediment sources (e.g., reviewed in Brown et al., 2009, D’Haen et al., 2012 and Mukundan et al., 2012) for more than 20 years (e.g., Joshi et al., 1991). These studies have used tracers in mountain streams to determine particle transit times (Bonniwell et al., 1999), watershed sediment budgets (Walling et al., 2006), sources of suspended sediments (Collins et al., 1998 and Mukundan et al., 2010), floodplain deposition and erosion (Humphries et al., 2010), and land use changes (Foster et al., 2007). Information for sediment sources derived from 210Pb and 137Cs has also been combined with numerical models to produce sediment budgets for watersheds. Generally,

these studies have used radionuclides and/or other sediment tracers with some combination of transport, mixing, storage, and depositional models with a randomization component (e.g., Monte Carlo simulation) to determine potential contributing sources to the sampled sediment. This approach identifies the often diffuse nature of sediment sources from the sediment sample. For example, numerical modeling elucidated the percent contributions of sediment (and associated Isotretinoin possible statistical deviations) from various catchment land uses (Collins et al., 2012b and Collins et al., 2012c). However, model limitations include the amount and timing of storage in system (Parsons, 2012), assumptions about unmeasured terms (Parsons, 2012), and the need for validated input data (Collins and Walling, 2004). Like any scientific model, the limitations and assumptions should be recognized to prevent over-reaching. In a previous study, the authors validated the regional correlation between excess 210Pb with urban watersheds and little to none excess 210Pb with channel/bank areas. Feng et al.

1-, 8 7-, 5 4- and 4 3-fold higher, respectively, than that in th

1-, 8.7-, 5.4- and 4.3-fold higher, respectively, than that in the head kidney Dabrafenib mouse (Fig. 4). mRNA of the Fas is expressed in a variety of tissues, including thymus, liver, lung, heart and ovary of the adult mouse [41], and is observed in some human cell lines [42]. In birds, erythrocytes

can perish rapidly via lysis or pyknosis [43], the latter being the most characteristic expression of programmed cell death or apoptosis [43], [44] and [45]. Most recent experiments suggest that injured (anucleated) erythrocytes display phosphatidylserine on their surface [46], [47] and [48], a key feature of apoptosis in nucleated cells. Under different pathological conditions, oxidative stress results in the activation of Fas and initiates the extracellular pathway of apoptosis in nucleated cells [49]. The possibility exists, given that teleost erythrocytes are nucleated, that these cells can undergo apoptosis or programmed cell death.

During mitogen-stimulated PBL expression analyses, we observed the induction of RbFas expression between 1 and 36 h after the PBLs were stimulated with lipopolysaccharide (LPS) and from 1 and 3 h after the PBLs were stimulated with poly I:C Fulvestrant molecular weight (Fig. 5). Additionally, the mRNA expressions of the RbFas in the kidney and the spleen were examined under bacterial and viral challenge via real-time RT-PCR analysis. Symptoms of the disease were first apparent on approximately day 4 postinjection, and each pathogen was reconfirmed via PCR (red sea bream iridovirus, RSIV) and cell culture (bacteria). The experimental challenge of the rock bream with S. iniae or RSIV resulted in significant increases in the RbFas mRNA in the kidney and the spleen. In the case of bacterial injection, the RbFas transcript

peaked 6 h after injection in both the kidney and the spleen ( Fig. 6A). Otherwise, the RbFas transcript peaked after 1 h in spleen and 6 h in kidney following the injection with RSIV ( Fig. 6B). LPS stimulates the polyclonal proliferation of salmonid lymphocytes, the respiratory burst and phagocytic activity of macrophages [50] P-type ATPase and, in addition, it has adjuvant properties. It has been shown to elicit the expression of cytokines like IL-1 in channel catfish [51] and TNF α in rainbow trout [52]. Cell death pathway can be initiated by a variety of cytotoxic agents, such as LPS, which induce activation of pro-inflammatory cytokines, caspases and other signalling pathways that ultimately lead to apoptosis and cell death [53]. Poly I:C is a synthetic analogue of dsRNAs that are generally produced during virus reproduction and triggers antiviral responses in host cells [54]. Robert et al. reported that a single poly I:C challenge is sufficient to induce an acute increase in apoptosis [55]. The LPS activation systems have previously been used in the analysis of molecular determinants of fish leucocyte proliferation.