NF-YB Regulates Spermatogonial Stem Cell Self-Renewal and Proliferation in the Planarian Schmidtea mediterranea

Gametes are the source and carrier of genetic information, essential for the propagation of all sexually reproducing organisms. Male gametes are derived from a progenitor stem cell population called spermatogonial stem cells (SSCs). SSCs give rise to male gametes through the coordination of two essential processes: self-renewal to produce more SSCs, and differentiation to produce mature sperm. Disruption of this equilibrium can lead to excessive proliferation of SSCs, causing tumorigenesis, or can result in aberrant differentiation, leading to infertility. Little is known about how SSCs achieve the fine balance between self-renewal and differentiation, which is necessary for their remarkable output and developmental potential. To understand the mechanisms of SSC maintenance, we examine the planarian homolog of Nuclear Factor Y-B (NF-YB), which is required for the maintenance of early planarian male germ cells. Here, we demonstrate that NF-YB plays a role in the self-renewal and proliferation of planarian SSCs, but not in their specification or differentiation. Furthermore, we characterize members of the NF-Y complex in Schistosoma mansoni, a parasitic flatworm related to the free-living planarian. We find that the function of NF-YB in regulating male germ cell proliferation is conserved in schistosomes. This finding is especially significant because fecundity is the cause of pathogenesis of S. mansoni. Our findings can help elucidate the complex relationship between self-renewal and differentiation of SSCs, and may also have implications for understanding and controlling schistosomiasis.


Introduction
Spermatogenesis is highly prolific, relying on SSCs for continual production of progeny. This prodigious output must employ multiple mechanisms to maintain the fine balance between SSC self-renewal and differentiation. Understanding the mechanisms of SSC maintenance is crucial for the treatment of several physiological and disease conditions. Self-renewal of SSCs without differentiation can result in tumor formation. For instance, seminoma-like growth of undifferentiated spermatogonia is seen upon expression of activated RAS, or overexpression of GDNF, or Cyclins D2 and E1, or BCL6B [1][2][3]. In contrast, aberrant development and differentiation of spermatozoa, due to insufficient sperm production, inadequate sperm motility, or abnormal sperm morphology, are the principal causes underlying male infertility [4].
The maintenance of germline stem cells is also a key feature behind the fecundity of trematodes such as Schistosoma mansoni, a causative agent of schistosomiasis, a disease affecting over 200 million people worldwide. The pathogenicity of schistosomiasis is due to the body's immune response to eggs laid by adult worms in their human hosts. The spermatogenic output of these parasites is clear from the observation that individuals with schistosomiasis can pass eggs over 30 years after initial infection [5,6]. Thus, in addition to illuminating the causes behind infertility and tumorigenesis, a better understanding of molecules that play a role in the maintenance of SSCs may provide new approaches for preventing and treating schistosomiasis.
One such molecule is a planarian homolog of Nuclear Factor-Y B (NF-YB), which belongs to the NF-Y family of transcription factors [7][8][9]. The NF-Y complex has been studied in several developmental contexts in D. melanogaster [10][11][12], C. elegans [13], and D. rerio [14], and a function in germ cells for this gene family has been described in the freshwater planarian Schmidtea mediterranea [15]. More recent work has shown that members of this complex also play roles in somatic stem cell maintenance in the asexual strain of S. mediterranea [16]. In the sexual strain, upon NF-YB knockdown, animals initially lost their SSC pool followed by more differentiated male germ cells. After over a month of NF-YB(RNAi), mature sperm were seen in sperm ducts of sexual planarians, and some animals had small testes filled with mostly spermatids and some sperm. Thus, NF-YB(RNAi) animals appeared to complete the initial rounds of spermatogenesis, but failed to maintain sperm production over time, possibly due to the loss of SSCs. This phenotype is strikingly similar to that seen in Plzf and TAF4b mutant mice [17,18]. How NF-YB coordinates the balance between self-renewal and differentiation decisions of SSCs at both cellular and molecular levels needs further exploration.
In this study, we provide a phenotypic characterization of planarian NF-YB(RNAi) using new markers to track individual stages of spermatogenesis [19]. Our experiments indicate that in S. mediterranea, NF-YB does not control germ cell specification or differentiation, but instead promotes self-renewal and proliferation of early germ cells. Interestingly, the NF-YB (RNAi) phenotype in the male germline is strikingly similar in both S. mediterranea and the trematode S. mansoni. Our findings provide mechanistic insight into the role of NF-YB, show the conserved function of this molecule in the testes of both free-living and parasitic flatworms, and may have implications for combating schistosomiasis.

Results
NF-YB(RNAi) results in progressive loss of male germ cells in Schmidtea mediterranea starting from the stem cell population To observe the different stages of NF-YB(RNAi) phenotype progression in the male germ cells of S. mediterranea (schematic Fig 1A), we tracked the following male germ cell populations and their respective signature transcripts: SSCs (= nanos), spermatogonia (= germinal histone H4/gH4), spermatocytes (= tektin1/tkn-1), and spermatids (= protein kinase A/pka) [20][21][22]. Loss of SSCs and spermatogonia was observed at the earliest stages of NF-YB(RNAi), indicated by loss of nanos and gH4 labeling (Figs 1B, 1C and S1). Although the spermatocyte layer was initially unaffected, upon continued knockdown, a reduction in tkn-1-labeled spermatocytes was seen (Figs 1D and S1). NF-YB(RNAi) animals also show varying degrees of mature spermatozoa loss during the RNAi timecourse. At later time points the testes only contained clusters of spermatids, labeled with pka (Figs 1E and S1) and some sperm. Eventually, there was a complete loss of all male germ cells after NF-YB(RNAi). Although all animals showed a progressive loss of male germ cells starting with the least differentiated cells (SSCs and spermatogonia), there was some variability both between samples and within samples in NF-YB(RNAi) animals. We hypothesize that this variability could be a reflection of the NF-YB mRNA/protein half-life in the system, or possibly reflect the variability of germ cell turnover among animals and between different testis lobes (S1 Fig, S1 Table).
Since NF-YB is part of a hetero-trimeric complex, requiring its partners NF-YA and NF-YC for transcriptional activation or repression [23][24][25], we also examined whether other components of the planarian NF-Y complex function in the gonad. We identified and cloned two planarian paralogs of NF-YA (A1 and A2), two of NF-YB (B and B2) and one of NF-YC. ClustalW analyses showed a high degree of conservation between the histone-fold motifs of these proteins with their human counterparts (S2A Our observation is consistent with experiments performed in the asexual strain of S. mediterranea [16]. Due to the early lethality of these RNAi treatments, we could not ascertain whether these genes also play roles in testes maintenance, and if they phenocopy NF-YB(RNAi). NF-YA2(RNAi) had no somatic or germ cell phenotype (S2C and S2D Fig), and its function may be redundant with NF-YA1. NF-YB appears to be the only subunit of the planarian NF-Y complex with a germline-specific function and this gene belongs to the relatively small group of planarian genes required for early germ cell maintenance. Thus, we directed our focus on its functional characterization.
NF-YB is not required for the specification of nanos-expressing cells in S. mediterranea The early germ cell loss seen in NF-YB(RNAi) animals is reminiscent of the knockdown of planarian nanos [21], a gene with conserved germ cell functions across metazoans (S3 Fig). The similar phenotypes of NF-YB(RNAi) and nanos(RNAi) led us to speculate that the two genes might act in concert to control germ cell development. The presence of a CCAAT box, the NF-Y DNA binding motif, -118bp upstream of the nanos transcription start site made NF-YB an attractive candidate regulator of nanos expression.
In S. mediterranea sexual hatchlings, expression of nanos is detected within 3 days post hatching [21]. If NF-YB plays a role in the regulation of nanos expression, we reasoned that NF-YB expression would precede that of nanos. In situ analyses showed that the NF-YB transcript was seen in the soma in early hatchlings. However, germline NF-YB transcript expression is observed only at later time points relative to nanos expression (Fig 2A). This observation suggests that NF-YB does not activate the expression of nanos.
Planarians specify germ cells from tissue fragments completely devoid of reproductive structures [21,26,27]. We modified a previously established experimental paradigm [27] to further test whether NF-YB is required for the specification of nanos-expressing cells. Briefly, sexually mature planarians were fed NF-YB double-stranded (dsRNA) 2-3 times and amputated anterior to the ovaries. The resulting head fragments (lacking reproductive structures at the time of amputation) were monitored for the re-appearance and maintenance of nanos-expressing cells at various regeneration time points (Fig 2B).
To ensure that NF-YB protein levels were depleted below the threshold required for the maintenance of nanos + cells, we performed NF-YB(RNAi) in sexually mature planarians (6 feedings over a month) until nanos + cells were lost (S4A Fig). At fifteen days of regeneration, both control (RNAi) (n = 11/11) and NF-YB(RNAi) (n = 11/11) ( Fig 2C) head fragments showed de novo nanos expression, indicating that NF-YB is not required for the respecification of nanos + SSCs. There was no significant difference in the number of respecified nanos + cells between control and NF-YB(RNAi) animals at this early time point (n = 11/11 for both, S5A Fig). The respecified nanos + cells in NF-YB(RNAi) animals persisted through regeneration for over a month (Fig 2D-2E'). In later stages of regeneration (45 days post amputation), control animals had numerous nanos + clusters and many nanos + cells per cluster, indicating proliferation of SSCs (n = 10/10, Figs 2C-2E', S5B and S5C). By contrast, NF-YB(RNAi) animals had fewer SSC clusters and the nanos + cells remained mostly as single cells in these clusters (n = 10/10, Figs 2C-2E', S5B and S5C). nanos transcripts in male germ cells were not detected in dmd1(RNAi) animals (S4B Fig), consistent with the previously reported role for this gene in SSC specification [24]. We validated the effectiveness and specificity of NF-YB knockdown with quantitative real time PCR and in situ hybridization experiments to ensure that nanos expression in NF-YB knockdown animals was not due to residual NF-YB, defective regeneration, or off-target effects (S4C and S6 Figs). Together, these data suggest that NF-YB is not required for SSC specification, but may function later in SSC self-renewal or proliferation.

NF-YB is required for the proliferation of planarian SSCs and spermatogonia in S. mediterranea
The NF-Y complex is associated with cell cycle regulation in other systems [28][29][30][31][32], and is enriched in many stem cell populations [33][34][35][36][37]. After ruling out a role for NF-YB in SSC specification, we tested if NF-YB is required for cell cycle progression of SSCs and spermatogonia and whether early germ cell loss in NF-YB knockdown was through differentiation or apoptosis. We knocked down NF-YB in juvenile sexual planarians; in these animals, the testes contain clusters of SSCs and spermatogonia, but lack the more differentiated germ cells. Thus, aberrations in spermatogonial differentiation are more easily assayed in these animals compared to mature sexual animals that already possess differentiated male germ cells. The dsRNA-fed animals were processed at early and late knockdown timepoints, cryosectioned, and sections of the same animal were used for phospho-histone H3 (PH3S10) and TUNEL labeling (Fig 3A).
In planarian testes, the SSCs give rise to spermatogonia, undergoing three rounds of mitosis with incomplete cytokinesis, which can be easily visualized by PH3S10 staining [38,39]. NF-YB (RNAi) animals showed a dramatic reduction in mitotic cell number following two feedings of dsRNA (Fig 3B and S1 Table). We confirmed that NF-YB(RNAi) animals do not show a reduction in nanos + SSCs or the nanos transcript at this RNAi timepoint (S7 Fig). Not surprisingly, this difference was more pronounced after four feedings, following loss of the mitotic spermatogonial layer (Fig 3C and S1 Table). Our analysis clearly indicates a reduction in proliferation of SSCs and spermatogonia upon NF-YB knockdown.
We next tested whether germ cells were undergoing apoptosis in NF-YB(RNAi) animals. We found low levels of TUNEL labeling in NF-YB(RNAi) animals after two feedings of dsRNA ( Fig 3D, S1 Table). However, following four feedings of NF-YB dsRNA there was an increase in apoptosis of the differentiated male germ cells (Fig 3E, S1 Table). NF-YB(RNAi) animals in later stages of RNAi (after four feedings) still have differentiated male germ cells (spermatocytes and spermatids), indicating that the early germ cells are unlikely to be undergoing apoptosis themselves and are capable of differentiation. Conditional deletion of both NF-YB alleles in primary mouse embryonic fibroblasts causes a block in progression of the cell cycle and induction of apoptosis [40]. We speculate that a similar mechanism of NF-YB-mediated testismaintenance could be acting here.

Sm-NF-YB(RNAi) results in fewer proliferating male germ cells in the parasite Schistosoma mansoni
We have previously shown that molecular similarities exist between planarian and schistosome somatic stem cells [41][42][43]; however, similarities between the germ cells of these two flatworms remain unexplored. To test if the NF-Y complex plays a similar role in the gonads of free-living and parasitic flatworms, we examined the role of NF-Y components in the parasite S. mansoni. We were especially interested in this comparison because the morbidity associated with schistosomiasis is a result of the tremendous reproductive output of the parasite. Inhibiting fertilization or propagation by blocking germ cell production may open novel avenues for treating this disease. Although schistosomes are dioecious, we restricted our analysis to male schistosomes (Schematic in Fig 4A) due to the testis-specific function of NF-YB in planarians [15].
In situ hybridization revealed that Sm-NF-YB, Sm-NF-YA, and Sm-NF-YC (Figs 4A and S8A) were enriched in the parasite testes, with possible low levels of somatic expression. Further, we found that Sm-nanos-1 is expressed in the testes of S. mansoni ( Fig 4A). Next, we asked if NF-YB plays a similar role in schistosome testis maintenance. We knocked down the NF-Y complex components in schistosomes by culturing the worms in vitro in the presence of dsRNA [41]. Similar to the planarian NF-YB knockdown phenotype, we found that following seven days of dsRNA treatment, Sm-NF-YB(RNAi) animals, and surprisingly Sm-NF-YA(RNAi) and Sm-NF-YC(RNAi) animals, showed a loss of Sm-nanos-1 labeling (Figs 4B and S8B).
To determine whether loss of nanos expression is due to reduced proliferation of the male germ cells, we performed a 24-hour EdU pulse in vitro at early (7 days) and late (14 days) RNAi time points. After 7 days of knockdown, control (RNAi) animals showed a large number of EdU + cells in the testes (Fig 4C). In contrast, Sm-NF-YB(RNAi) animals had fewer EdU + male germ cells (Fig 4C), while maintaining an intact testis structure (assessed using DAPI labeling). Similarly, Sm-nanos-1(RNAi) animals had fewer proliferating male germ cells but greater male germ cell loss compared to Sm-NF-YB(RNAi) animals (Fig 4C). Following 14 days of RNAi, most male germ cells were lost in both Sm-NF-YB(RNAi) and Sm-nanos-1(RNAi) animals (Fig 4D and 4E). Sm-NF-YA and Sm-NF-YC knockdown animals showed a similar RNAi phenotype (S8C Fig). Together, we conclude that the role of NF-YB in male germ cell proliferation is conserved in both free-living and parasitic flatworms.

Discussion
Significant progress has been made in understanding post-transcriptional regulation and the role of RNA-binding proteins in the germline [44][45][46]. However, mechanisms of transcriptional regulation in germ cells have received relatively less attention. A variant of TFIIA, ALF or TFIIAτ, is expressed in male and female gonads in mice and Xenopus, and can substitute for TFIIA in core promoters [47,48]. A TATA-Binding Protein (TBP) variant, TLF/TRF2 is not  required for the female reproductive system, but loss of TLF/TRF2 results in an inability to complete spermatogenesis in males [49][50][51]. Several TBP-associated factors (TAFs) have germ cell-specific roles. Deletion of TAF4b in both female [52] and male [19] gonads results in sterility. A TAF-II-80 homolog, cannonball, is expressed in Drosophila spermatocytes, and mutations in this gene block spermatid differentiation [53]. Here, we studied a planarian male germ cell-specific transcription factor, NF-YB, exploring the cellular mechanisms of NF-YB-mediated maintenance of early planarian male germ cells. We also found gonadal enrichment of NF-Y components in male schistosomes, and a requirement for NF-Y genes in the proliferation and maintenance of male germ cells in these parasites.
Members of the NF-Y complex are enriched in stem cell populations in many other systems. In proliferating skeletal muscle cells the NF-Y complex, and its target cyclin B1, are expressed at high levels; however, in terminally differentiated cells there is a loss or reduction in NF-Y components [33]. In hematopoietic stem cells the NF-Y complex activates HOXB4, a homeobox gene that is expressed abundantly in primitive HSCs. During HSC differentiation there is a decline in NF-Y binding to the HOXB4 promoter and a concomitant reduction in HOXB4 transcript levels [54]. Subsequent work showed that NF-Ya overexpression in HSCs shifts the balance towards HSC-self-renewal rather than differentiation [34]. By contrast, deletion of NF-Ya caused an accumulation of HSCs in G2/M phase of the cell cycle, followed by apoptosis, possibly as a result of dysregulation of key genes involved in cell cycle control, apoptosis, and self-renewal [36]. In human embryonic stem cells (ESCs), the NF-Y complex is required for proliferation and isoforms of NF-Ya are differentially expressed during differentiation [35]. A recent study showed that, in addition to its housekeeping functions, the NF-Y complex regulates ESC identity by coordinating the binding of ESC master transcription factors to core self-renewal and pluripotency genes [37]. Our finding that planarian NF-YB is necessary for self-renewal and proliferation of SSCs and spermatogonia is consistent with the known functions of the NF-Y complex in these other stem cell systems, and provides insight into the role of this transcription factor family in germ cells. We also performed functional characterization of the NF-Y complex in the parasite Schistosoma mansoni. Previous work reported that the S. mansoni NF-YA protein is expressed in both male and female gonads and levels of Sm-NF-YA decreased as maturation of the male germ cells progressed [55]. We found that NF-Y components are necessary for the proliferation of male germ cells in S. mansoni.
The NF-YB(RNAi) phenotype in planarians is strikingly similar to those observed in Plzf [17,18] and TAF4b [19] mutant mice, both of which undergo progressive loss of spermatogonia with age. This progressive loss, from the least differentiated to the most differentiated germ cells, is not expected in the case of meiotic or maturation defects, strongly indicating that all three genes function in early male germ cell maintenance. Both Plzf and TAF4b mutant mice are born with normal number of gonocytes/primordial germ cells, indicating proper specification of the germ cells; planarian tissue fragments lacking germ cells and NF-YB activity regenerate normal number of SSCs, consistent with proper specification of these cells. Plzf and TAF4b mutant mice complete the initial rounds of spermatogenesis but show decreasing fertility with age; in planarians, NF-YB(RNAi) SSCs and spermatogonia are able to differentiate initially but fail to do so over time. Plzf and TAF4b mutant mice show a decrease in proliferative spermatogonia over time; NF-YB(RNAi) animals also show reduced proliferation of male germ cells in both S. mediterranea and S. mansoni. Given the striking similarities between Plzf, TAF4b, and NF-YB functions in the male gonad, it is not unreasonable to speculate that these three genes might be controlling similar targets, or genes that perform similar functions.
We also observed an increase in apoptotic germ cells in later stages of NF-YB(RNAi). NF-YB (RNAi) results in loss of nanos + SSCs, but these cells do not show obvious TUNEL labeling. Several possibilities exist to explain this observation. SSCs may be undergoing apoptosis but the signal may be too weak or transient to be detected, or they may use a non-apoptotic mechanism of cell death. It is also possible that the early germ cells could be entering the differentiation pathway aberrantly, resulting in apoptosis of the differentiating cells. NF-YB(RNAi) results in loss of elongated spermatids and sperm in addition to early germ cells (Figs 1 and S1). Although the primary phenotype of NF-YB(RNAi) is the loss of early germ cells, the expression of NF-YB transcript in the more differentiated male germ cells leaves open the possibility that NF-YB regulates additional target(s) vital for the survival of these differentiated cells.
The identification of a testis-specific component of the planarian NF-Y complex, and our finding that NF-YB is required for the maintenance of planarian SSCs, provides a valuable tool for understanding the dynamics of early, undifferentiated germ cells. Putative SSC-specific targets of NF-YB will help reveal the function of the NF-Y complex in the planarian, in which tissue-specific knockdown is not possible. Several genes required for the maintenance of various stages of planarian germ cell development have homologs in other species [15]. Many of these genes, such as rap55, ELAV, and MSY4 [56][57][58][59][60], are known to play roles in germ cell development in vertebrates. Given the observation that the NF-YB transcript is upregulated in mouse spermatogonia relative to other male germ cells [59], we predict that the role of NF-YB in SSC self-renewal and proliferation might be conserved in vertebrates. Thus, a better understanding of SSC-specific NF-YB targets in planarians is expected to yield insight into the workings of early germ cells across different systems.

Ethics statement
In adherence to the Animal Welfare Act and the Public Health Service Policy on Humane Care and Use of Laboratory Animals, all experiments with and care of vertebrate animals were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Illinois at Urbana-Champaign (protocol approval number 13017).

Cloning of NF-Y complex
Coding DNA sequences of the NF-Y complex were obtained from SmedGD [63]. Mixtures of sexual and asexual planarian cDNA were used as templates to clone NF-Y components, using primers in S2 Table. qRT-PCR Total RNA was extracted using TRIzol (Invitrogen) according to manufacturer's instructions, DNase treated (Fisher Scientific) and purified using an RNA clean-up kit (Zymo) before reverse transcription (iScript, Bio-Rad). Prior to RNA extraction, animals were starved for 7 days after the last RNAi feeding to ensure that any remnant dsRNA was cleared from the system. qPCR was performed using GoTaq qPCR master mix (Promega) using Applied Biosystems StepOne Plus RT-PCR system. All experiments were done in biological and technical triplicates. Transcript levels were normalized to β-tubulin (primers in S2 Table). Relative mRNA levels were calculated using ΔΔCT [64].

In situ hybridization
Whole-mount in situ hybridization of planarians was performed as previously described [66] with modifications for the sexual strain [21,22,67].

Imaging
Samples developed through the NBT/BCIP colorimetric method were mounted in 80% glycerol and imaged using a Leica M205A stereomicroscope (Leica, Wetzlar, Germany), equipped with Leica DFC420 camera. Whole-mount FISH animals were mounted in Vectashield (Vector Laboratories, Burlingame, CA) and imaged on a Zeiss Stereo Lumar V12 (Carl Zeiss, Germany). For confocal FISH images, samples were mounted in Vectashield and imaged using a Zeiss LSM710 confocal microscope running ZEN 2011. Images were processed using Adobe Photoshop CS5.

Schistosoma culture and RNAi
Schistosoma mansoni, Strain NMRI-exposed Swiss Webster mice (NR-21963) were provided by the NIAID Schistosomiasis Resource Center at the Biomedical Research Institute (Rockville, MD) through NIH-NIAID Contract HHSN272201000005I for distribution through BEI Resources. Mice were perfused with DMEM containing 10% heat-inactivated serum and schistosomes were cultured in vitro [41]. In situ hybridization was performed as previously described [41]. For RNAi, animals (in quadriplicates) were soaked in dsRNA generated by in vitro transcription (30 μg RNA per 10-12 pairs in 3 ml of Basch 169 medium [70,71]). Animals were incubated for 7 and 14 days at 37°C. EdU pulse chase and detection were performed as described previously [41]. nanos + SSCs were the first cells to be lost in NF-YB(RNAi) animals, after 14 days of RNAi (4 feedings), closely followed by gH4 + SSCs and spermatogonia. At these early stages, very few testis lobes showed reduced tkn-1 + spermatocytes and pka + spermatids. By 23 days of RNAi (6 feedings), more testis lobes showed reduced nanos and gH4 expression, and the number of testis lobes with reduced tkn-1 labeling increased slightly. 32 days after starting RNAi (8 feedings), all testis lobes examined lacked nanos and gH4 labeling, many testis lobes showed reduced tkn-1 expression and about half the lobes showed reduced pka expression. By 42 days (10 feedings) almost all germ cells were lost. Elongated spermatids and sperm were lost early (between 14-23 days, 4-6 feedings) and this loss was visualized using DAPI. Ten testis lobes per animal (n = 4-6) were counted for each testis marker per RNAi time point. We quantified SSCs in control and NF-YB(RNAi) animals to ensure that the reduced PH3S10 labeling was not due to fewer nanos + cells in NF-YB(RNAi) animals. Following 2 feedings of dsRNA (6 animals each, 4-8 testis lobes per animal), percentage of nanos + cells per testis lobe in NF-YB(RNAi) animals (28.2 ± 1.4, n = 44) was not significantly different (P<0.05) from control (RNAi) animals (24.8 ± 1.2, n = 42). Unpaired parametric T-test with Welch's correction was performed. Scatter plot shows mean with SD. (C) qRT-PCR assay showing that nanos mRNA levels were unaffected following 2 feedings of NF-YB dsRNA. Unpaired parametric two-tailed T-test with Welch's correction was performed to determine significance ( Ã = P value 0.01-0.1; n.s. = not significant). (TIF)