JCB

JCB/NYAS Meeting: Keynote Addresses

In addition to each of its sessions on Chromosomes, Cancer, and Stem Cells, the JCB Cell Biology of Disease Meeting had five powerful keynote addresses (two at the beginning and three at the end) that further highlighted exciting work underlying the cell biological basis of disease.

Opening Night

 

Our opening speaker on Wednesday evening was Ken Yamada (National Institutes of Health).  Yamada discussed work by his group recently published in the JCB which uses micropatterning to examine the effects of extracellular matrix dimensionality on mechanisms of cell motility (Doyle et al., 2009). Yamada and his team noticed that cells traveling in 3D often follow single fibrils of matrix for extended periods of time, and wondered what cell migration would look like if cells were given single tracks of matrix to follow- a matix they refer to as one-dimensional. In a nutshell, cell migration in 1D recapitulates migration in 3D, while migration in 2D remains distinct from both.  If a picture is worth a thousand words- our biosights videocast on this study speaks volumes, and will bring you up to speed in less than 8 minutes. 

 

Live cell imaging has revealed key differences in cell morphology, adhesion, receptor usage, and signal transduction pathways on flat or voluminous matrices.  For example, inhibition of myosin II doesn’t disrupt cells moving in 2D, but it does block migration in 3D and 1D.  One debate central to 3D migration is whether proteases are required for cell migration or tissue invasion in vivo.  Yamada’s group is now using fluorescence ratio imaging with a probe they’ve developed to detect metalloprotease activity in real time to look at adhesion dynamics and matrix degradation as cells migrate through 3D matrices.  Tissue organization during development is another fascinating way to study cell migration in a 3-dimensional context. 

 

Yamada described experiments his group is using to study branching morphogenesis in vivo.  It was fascinating to watch cells from a disrupted gland spontaneously aggregate and regenerate a new gland.  A second surprise was revealed by live cell imaging.  Although epithelial cells are commonly viewed as static in developed tissues, they are actually constantly moving, even in “fully developed” glands.  Yamada is trying to figure out how clefts and buds, characteristic structures of branching tissues, form:  What are the cell shape changes and the function of the matrix in forming these structures?  He also described some preliminary work characterizing a protein that promotes cleft formation.

 

   The next talk on opening night was given by Alan Hall , another member of the JCB Editorial Board.  How cells establish polarity and directionality during migration is one focus of Hall’s work.  Hall has a career’s worth of evidence to support a role for the Rho family GTPase Cdc42 in the regulation of cell polarity in tissue morphogenesis and in cell migration. 

Cdc42 regulates the actin and microtubule cytoskeletons during polarized cell migration.  Re-organization of actin is dependent on the Pak, an effector downstream of active Cdc42, while the Par6/aPKc pathway regulates microtubule polymerization.  The cdc42/Par6/aPKC complex is asymmetrically distributed in polarized cells (e.g. during migration or cell division), and localizes to the apical domain of polarized epithelial cells. 

 

Hall’s group used a 3D morphogenesis assay, similar to one described by Ken Yamada in the previous talk, to examine the function of Cdc42 in the regulation of epithelial cell polarity (See the biosights episode here).  When embedded in extracellular matrix, CaCo2 epithelial cells form fluid-filled cysts, and have an established polarity by a two-cell stage. As cells continue to divide, the apical markers remain centralized, suggesting that the orientation of the mitotic spindle and/or the site of cytokinesis specifies polarity.  To their surprise, depleting Cdc42 with siRNA didn’t prevent lumen formation altogether, rather Cdc42-depeleted cells form multiple lumens due to misorientation of the mitotic spindle (Jaffe et al., 2009). 

 

But Cdc42 is just one of 18 Rho family GTPases expressed in mammalian cells – Hall reasoned that other GTPases may also contribute to cell polarity establishment or maintenance.  In addition, Rho GTPases are regulated by more than 80 GTPase exchange factors (GEFs) and nearly 70 GTPase activating proteins (GAPs), which expands the candidate pool significantly.  Therefore, when Hall’s group conducted a shRNA screen for GTPase-related proteins involved in establishing cell polarity and tight junction formation, they expected more hits than they got – which was a disappointing “zero”.  However, a screen of more than 100 downstream Rho GTPase effector proteins turned up three candidates, Par6, Pak4, and aPKC – the same components important for establishing microtubule orientation and polarity during C. elegans embryogenesis.  In the CaCo2 system, Par6 or aPKC depletion phenocopies Cdc42 depletion, resulting in multiple-lumen formation.  However, depletion of aPKC also disrupts normal junction formation.  Finding out how all of these components work together to coordinate polarity and tissue morphogenesis is an ongoing project in Hall’s lab.

The Final Session

The final three talks of the meeting brought Chromosomes, Cancer, and Stem Cells together in one impressive session that beautifully illustrated the themes of the conference. 

 

Don Cleveland began the final session of the symposium by introducing what he calls “the centromere paradox”.   Centromeres, as we learned from Iain Cheeseman and others, are the DNA sequences that link chromosomes to spindle microtubules. Although the function of the centromere is conserved throughout evolution, the actual DNA sequence of centromeres is not.

 

Cleveland and others have argued that the conserved element is an epigenetic demarcation.  The nucleosomes that form at centromere DNA include a histone H3 variant, called CENP-A.  These centromeric nucleosomes are therefore structurally divergent and are conformationally more rigid than conventional nucleosomes that contain Histone H3. The exact architecture of CENP-A nucleosomes, however, is still up for debate (see Furuyama and Hennikof, 2009 and a recent JCB feature for details). 

 

To determine how the epigenetic mark for centromeres is replicated, Cleveland ’s group used SNAP-tagging to introduce a fluorescent substrate to detect when and where new CENP-A was synthesized.  They found that new CENP-A was synthesized and incorporated into centromeric chromatin during G1 (Jansen et al. 2007). Cleveland therefore argues that propagation of centromeric chromatin requires exit from mitosis.  

Now we know when CENP-A is synthesized, but how is it incorporated into nucleosomes?  Cleveland’s group used TAP affinity purification to identify components of the CENP-A prenucleosomal complex, which is significantly different from histone-H3 containing nucleosomes, and found a chaperone HJURP that loads CENP-A onto chromatin during G1 (Foltz et al., 2009). 

From CENP-A, Cleveland skipped through the CENP alphabet to CENP-E.  CENP-E is an unusual kinesin required for correct chromosome alignment.  Bound to kinetochores by its tail domain, its motor domain extends outwards from the chromosomes towards microtubules eminating from the spindle.  Cleveland describes CENP-E as a “tenacious, floppy tortoise”- because it is a very slow motor (50 times slower than other motors), but it is very processive – it remains bound to microtubules for minutes at a time.

Removal or reduced expression of CENP-E results in chromosme non-disjunction and high rates of aneuploidy.  In male mice heterozygous for CENP-E, the Y chromosome mis-segregates with high frequency.  To determine whether aneuploidy caused by loss of CENP-E can beget tumorigenesis, Cleveland’s team injected CENP-E +/- MEFs into nude mice, and indeed sees an increase in tumor development. Cleveland is excited about the promise of CENP-E inhibitors as anti-cancer therapeutic agents.  Unlike taxol, which inhibits mictrotubule depolymerization and therefore affects function of all cells, inhibitors of CENP-E would be extraordinarily specific for dividing cells.

After an energetic talk by Cleveland, no momentum was lost when Haifan Lin (Yale University) stepped up to the stage.  The discovery of piRNAs has turned junk DNA into a goldmine for Lin.

Lin found that most piRNAs are coded by stretches of the “junk” DNA that makes up ~98% of the human genome.  piRNAs are just one class of short RNAs, which differ from miRNAs and siRNAs because they bring their targets to Piwi proteins, not argonaute family endonucleases.  PiRNAs also are distinct due to their size (They’re a little bigger than miRNAs), and their shear number.  Lin’s lab has now identified more than 60,000 piRNAs in mouse testis, and 14,000 piRNAs in the fly genom

(Yin and Lin, 2007).

Identifiying the function of piRNAs is a daunting task since there are so many of them (Lin 2007).  However, where they are located in the genome is revealing – many piRNAs appear to regulate epigenomic templating. 

Piwi family members, known as Miwi in mice, are piRNA-specific proteins. In Drosophila, knockout of Piwi results in defects in ovarian stem cells, which tend to differentiate into two egg chambers, instead of one cell maintaining stem-cell capabilities for self-renewal.  In a yeast two-hybrid screen to identify other Piwi interacting proteins, Lin identified the heterochromatin protein 1A.  HP1A controls heterochromatin silencing and molds chromatin structure through interactions with Histone H3K9.  HP1A binding to chromatin requires Piwi complexed to piRNA, suggesting that piRNAs may target Piwi and chromatin remodeling proteins to specific areas of the genome.

 

 

Joan Massagué (Memorial Sloan Kettering Cancer Center) is something of a cellular archeologist.  Many metastatic cancers have preferential sites they colonize, for breast cancer cells, those sites include bone, lung, and brain.  In an ambitious comparative genomics project, Massagué isolated cells from each of these sites and compared them to cells isolated from the primary tumor, as well as healthy breast tissue.  The process of metastasis itself, he argues, may define the progression of acquired traits that make invasion, survival and colonization of tumor cells in alternate tissues possible.

When breast cancer cells reach bone, they influence the behavior of native cells (Lu et al., 2009).  Breast cancer cells induce osteoclasts to proliferate, differentiate, and resorb bone matrix. All of this activity releases additional growth factors that the cancer cells use to further proliferate (e.g. TGF-b, IGF-1).  Breast cancer cells derived from bone metastatic sites express a distinct set of genes, including MMP1, a metalloprotease associated with invasiveness, ostopointin, a protein that regulates osteoclast function, and IL-11, a cytokine that induces osteoclastogenesis.

 

The gene set expressed by breast cancer cells isolated from lungs have yet another distinct set of genes.  The lung metastasis signature identified by Massagué includes nearly 20 genes expressed in estrogen receptor negative tumors associated with high potential for relapse of lung metastasis. These genes include Cox2, a regulator of ROS production and inflammation, and Angiopoietin-like 4, important for breast cancer cell extravasation (Minn et al., 2005; Gupta et al., 2007; Padua et al. 2008). 

When compared to cells derived from the primary tumor, Massagué was surprised to find that Angptl4 was also expressed there.  Although Angptl4 doesn’t necessarily confer a growth advantage to cells in the primary tumor, when these cells reach the lung, it helps them to invade lung tissue by reducing cell-cell adhesions between pulmonary endothelial and epithelial cells –increasing their ability to invade into/through the epithelial layer. 

 

The prominence of breast-cancer derived metastasis to the brain is more than ten times the incidence of primary brain tumors.  Brain metastasis results in neurological disability, the tumors are resistant to treatments, are predominantly lethal, and poorly understood.  Massagué found that Breast cancer cells positive for BrMS are more likely to invade brain tissue.  A subset of genes are shared between lung and brain-metastatic cells, including Cox2 and Angplt4, which are both important for crossing the blood-brain-barrier.  In addition- brain-tropic cells express ST6GalNac5, a sialyl transferase normally only expressed by brain-derived cells.

JCB/NYAS Meeting: Cancer

Prevention, diagnosis, and treatment of cancer remains a challenging scientific and medical problem.  Our former Editor-in-Chief, Ira Mellman, opened the third session of the conference, focused on cancer on Friday morning with a discouraging figure. While the mortality rates from heart disease, stroke, influenza, and pneumonia have declined dramatically over the past 50 years, mortality due to cancer has changed little.  Mellman stressed that ‘translational’ cell biology is a two-way street, and there is much to be gained from the application of cell biological-based theories to therapeutic practices as well as understanding the mechanisms of clinical observations at the cell-biological level.

Ira Mellman (Genentech): Think twice about ligand binding and receptor dimerization

Mutations in the HER/EGF receptor gene family have been linked to human cancer. Monoclonal antibody therapeutics have been used clinically (e.g. Erbitux(Imclone), Vectibix(Amgen), and Herceptin(Genentech)), but the mechanism of action for all of these reagents – or why they sometimes fail- is not entirely clear. 

Signal transduction initiated by Her1/EGFR has been studied extensively, and the literature repeatedly iterates the dogma: EGF binding induces receptor dimerization and activation.  A high resolution structure of Her1/EGFR was recently published (Jura et al., 2009).  Indeed, ligand binding to the EGFR is predicted to change its confirmation from a closed to an open configuration, extending the intracellular dimerization loop to allow close apposition of receptor pairs for cross-phosphorylation and receptor activation. 

 

However, molecular imaging studies are providing new insights into how EGF sets its receptor going.  Using TIRFM, Mellman’s group at Genentech is following movement of the EGFR on the cell surface at the single molecule level.  Labeling the receptor with quantum dots allows the receptors to be tracked for extended periods of time in live cells.  The analysis revealed that receptors continuously and reversibly form dimers from monomers in the absence of ligand, with dimer formation being favored in the cell periphery -- despite an apparently even distribution of total receptors across the cell surface.  EGF binding requires dimerization competent but not activation competent EGFR, suggesting that dimerization of the receptor somehow favors ligand binding.  Mellman proposes that spontaneous dimerization primes the receptor for ligand-induced signaling.  The receptor predominantly exists as closed, low affinity monomers which “flicker” into inactive, higher affinity dimers primed for ligand binding and receptor activation.  Mellman hopes that similar studies in the future will be useful for understanding how anti-EGFR therapeutics affect function and regulation of this cancer-promoting receptor.

 

Lloyd Trotman (Cold Spring Harbor Laboratory): knocking the “two-hit hypothesis” with a dose of PTEN 

Trotman continued the trend of overturning the textbook version of cancer mechanisms.  The textbooks all say that it takes two hits – or multiple mutations - to make a tumor.  Trotman’s transgenic mice say it takes just one thing: active Akt.

 

Akt, a kinase that regulates all insidious cancer behaviors – proliferation, survival, polarization, and migration, is localized to the plasma membrane through association with the phospholipid PIP3.  In 2003, Trotman and colleagues generated mice which express varying levels of PTEN, a phosphatase that converts PIP3 to PIP2, and thus inactivates Akt.  They found that even a 30% reduction in PTEN levels resulted in a correlative increase in phosphorylated Akt – and formation of prostate tumors (Trotman et al., 2003).

 

In 2006, Trotman found that tumor incidence also increased in mice lacking the tumor suppressor PML (Trotman et al. 2006).  While Akt accumulates in the cytoplasm in PTEN-deficient cells, in PML-null cells it ends up in the nucleus where it inhibits pro-apoptotic and cell-cycle arrest programs.  PML, he showed, is required to recruit the PP2A phosphatase to nuclear bodies to turn off Akt and suppress uninhibited cell growth.

 

Using a search for proteins that favor the same PIP3-enriched neighborhoods as Akt, Trotman has now discovered a third phosphatase that regulates Akt. Relying on a tried-and-true approach, Trotman’s group has knocked the gene out in transgenic mice to see if they, too, will develop prostate tumors due to over-active Akt.

As an additional resource, Trotman has also joined forces with the MSKCC oncogenome project, which has obtained more than 150 primary tumor samples, including 34 metastatic biopsies.  The samples have been analyzed for alterations in Akt phosphatases by DNA copy number, mRNA expression levels, and immunohistochemistry. In these samples, there is frequent decrease in expression of PTEN.  However, in the metastatic samples, both PTEN and the newly identified Akt-phosphatase seem to be missing.  Understanding how loss of multiple phosphatases contributes to invasive phenotypes is an ongoing focus of research for Trotman’s group. For more information on Lloyd Trotman, see the P&I published in JCB last year.

 

Maria Blasco (CNIO, Madrid): The Long and the Short of iPS reprogramming
The next speaker was JCB Editorial Board Member Maria Blasco.  Blasco’s recent studies address the regulation of telomere length and its relationship to stem cell proliferative capacity and aging.  With every cell division, the telomeric repeats at the tips of chromosomes are shortened.  It’s the job of the shelterin complex, which binds to telomeres, to prevent chromosome instability and to compensate for loss of telomere length.  Mutations in shelterin complex components are associated with rare, human stem cell diseases which typically manifest as bone marrow failure, aplastic anemia, pulmonary fibrosis, and skin hyperpigmentation.

 

Normally, stem cells have long telomeres, however shortened telomeres are coincident with cells that leave the stem cell niche and differentiate.  In aged animals, the length of telomeres is generally decreased- even in stem cells.  Increasing telomerase activity in transgenic mice increased the animal’s life span, presumably through delay of telomere shortening.   

 

Blasco wondered what impact telomere shortening would have on stem cell function.  In mice missing the telomere component TERF1, mobilization of epithelial stem cells in response to TPA treatment is decreased, making the skin more resistant to carcinogenic treatment. However, these mice trade resistance to cancer for premature aging (Martinez et al. 2009). The mice also have a hyperpigmentation phenotype, and so mimic many of the characteristics typical of the human pathologies associated with loss of TERF function. 

 

What about reprogrammed, or induced pluripotent stem (iPS) cells?  Blasco’s team recently showed that expression of the iPS inducing transcription factors  Klf4, Oct3/4, Sox2, and myc,  “rejuvenated” telomeres.  Blasco also noted that the shorter telomeres are, the harder it was to get cells to re-program.  However, the decreased re-programming efficiency could be rescued by introduction of telomerase (Marion et al. 2009).  The DNA damage response factor p53 also was a limiting factor for reprogramming efficiency, which initiates programmed cell death pathways in response to telomeres that have become too short.  Blasco’s team showed that loss of p53 suppressed the accelerated aging phenotypes observed when there are defects in shelterin complex function and telomere protection or extension .

 

Senthil Muthuswamy (Ontario Cancer Institute): Too many cells does not a tumor make

Muthuswamy uses a 3D culture system to approximate acinar formation of human breast tissue in vitro.  He noticed that in addition to inducing rapid cell proliferation, expression of the breast-cancer inducing gene ErbB2 also disrupted the normal morphogenesis of these structures, which formed multiple, lumen-free acini mimicking the pre-malignant state observed in cancer patients.  However, when proliferation was induced by over-expression of cyclin D or inactivation of Rb, hyperplastic growth occurred – but the structure of the tissue was maintained. He was quick to point out that in 2D, under the same circumstances, no phenotype is observed.

 

Muthuswamy’s curiosity about how disruption of normal tissue morphogenesis affects tumor development has been a source of fixation throughout his career (Be on the look out for a P&I feature on Muthuswamy in the October 19 Issue of the JCB, Feigin and Muthuswamy 2009).  His suspicion that defects in regulation of cell polarity or a normal morphogenetic program may be an important component of metastatic tumor growth was confirmed by a recent set of experiments, which investigated the function of Scribble, a scaffolding protein that localizes to cell-cell junctions and is downregulated in certain cancer types.

 

Decreasing expression of scribble disrupts polarity and organization of epithelial cells grown in 3D culture.  Loss of scribble also blocks apoptosis during lumen formation, but does not have an effect on proliferation.  In vivo, loss of scribble disrupts mammary gland morphogenesis.  To find out what happens if you combine scribble deficiency with a proliferation-inducing oncogene, Muthuswamy expressed c-Myc in scribble knockout cells.  The result was strikingly similar to the neoplastic growths observed when ErbB2 is over-expressed (Zhan et al., 2008).

 

Alissa Weaver (Vanderbilt University): When it comes to invasion, tumors are inflexible.

Tissue composed of tumor cells is often more rigid than healthy tissue, perhaps due to increased cell packing, inflammatory or scar tissue infiltrates, or excess deposition of extracellular matrix.  In humans, this kind of dense tissue is associated with an increased risk of invasive breast cancer, which Weaver studies in her lab.

Weaver wondered if the stiffness of the extracellular matrix could influence tumor invasion by modifying the formation or function of invadopodia- specialized cellular protrusions that promote localized matrix degradation.  Formation of invadopodia correlates with invasion through basement membrane and tumor metastasis in nude mouse models. 

 

Weaver’s team in fact found that increased matrix density can increase both the number and degradative capacity of invadopodia in vitro (Alexander et al., 2008).  The ability to “sense” tension, through integrin-mediated activation of Myosin II and MLCK is important for invadopodia formation. Other candidates for mechanosensory transduction are p130 Cas and the focal adhesion kinase, FAK.  Phosphorylation of both molecules is increased in invadopodia, but not in the presence of the myosin inhibitor blebbistatin, which inhibits tension signaling. 

JCB/NYAS Meeting: Chromosomes

Rebecca Heald (UC Berkeley), one of JCB’s senior editors, kick-started Thursday afternoon’s session on Chromosomes.   Heald’s lab is broadly interested in understanding different aspects of cell division from chromosome organization to spindle assembly.  Most of her work is done in an in vitro system using crude extracts purified from eggs of the South African clawed frog Xenopus laevis.  Remarkably, these extracts can support most aspects of the cell cycle in a test tube.  This frog won’t win any beauty contests but watching spindle assembly in these egg extracts is a beautiful sight!   The system provides some unparalleled advantages, namely it is open to biochemical manipulation and different cell cycle stages can be studied independently.  In her talk, Heald described two ongoing projects in her lab on chromosome architecture and organelle size. 

Chromosome condensation is a hallmark of mitosis and crucial for accurate segregation.  DNA bound to octameric histones can be further compacted by linker histones such as histone H1.  In work that was published in the JCB, Heald and colleagues showed that H1 function is necessary for mitosis in Xenopus egg extracts.  In its absence, chromosomes become tangled up and do not segregate properly.  Heald presented new work comparing the function of embryonic and somatic isoforms of histone H1.  They examined whether the different isoforms can substitute for each other and the effects of this substitution on chromosome architecture.  Surprisingly, isoforms have different properties and regulation despite extensive homology between them.   By analyzing the proteins that interact specifically with one isoform vs. another, they are starting to get a handle on how these linker histones differentially control chromosome structure.

Heald then moved on to talk about her lab’s work on determinants of organelle size, namely the nucleus.  Previously in The JCB, Heald and colleagues observed that spindle size was affected by cytoplasmic factors.  They showed that spindles formed in egg extracts from the larger Xenopus laevis were bigger than those formed in extracts prepared from the smaller Xenopus tropicalis.  Curiously, mixing the extracts from these frogs resulted in spindles of intermediate length.  Now they are using this system to understand how the size of the nucleus is specified.  Like spindles, mixing extracts from the two frogs results in intermediate sized nuclei.   Heald’s lab is trying to uncover the molecular basis of this scaling.  Simultaneously, they are investigating another scaling phenomenon, i.e., as development proceeds why do nuclei become smaller?  Heald predicts that these two lines of inquiry will converge on the same molecular explanation.  Answers to these fundamental questions may prove useful therapeutically, as many cancer cells are characterized by alterations in cell size and nuclear morphology.

Next, Iain Cheeseman (Whitehead Institute) zoomed in on the kinetochore, the protein platform that connects chromosomes and microtubules during mitosis.  The kinetochore has a formidable task:  It has to ensure biorientation, i.e. only allow microtubules from opposite spindle poles to attach to opposing kinetochores.  While trying to achieve this orientation that is crucial for proper chromosome segregation, the kinetochore is bombarded by microtubules from all directions.  Many incorrect attachments (such as attachment to microtubules from the same pole) occur, which can have disastrous consequences (see summary on Pellman’s talk below).  In molecular terms, how does this structure translate these subtle differences in microtubule binding to stabilize the right connection and fix the wrong ones?  This question is the focus of Cheeseman’s current endeavors. 

Cheeseman presented his work on the KMN (KNL-1/Mis12 complex/Ndc80 complex) network, which consists of conserved protein complexes that are part of the outer kinetochore.  Cheeseman has previously shown that this network is essential for binding of microtubules to the kinetochore.  At the conference, he talked about how Aurora B regulates this network and presented a model for how this regulation could account for the ability of the kinetochore to molecularly sense correct and incorrect attachments.  To learn more about Cheeseman’s passion for the kinetochore, check out Ben’s People and Ideas on him in the September 21^st issue of JCB.

Tarun Kapoor (Rockefeller University) spoke next about his lab’s interdisciplinary efforts to study properties of the spindle.  Kapoor has done some important work relating to the cell biology of disease.  As a postdoc, he characterized a small molecule inhibitor of the kinesin Eg5 and evaluated its effects on spindle assembly (published in the JCB).  Kapoor started off his talk by mentioning that biotech companies are developing the use of kinesin inhibitors as chemotherapeutics.  He then moved on to discuss his more recent work on spindle architecture.

Kapoor and his colleagues devised a way to image single molecules of fluorescent tubulin incorporated into spindles formed in Xenopus egg extracts.  Using this single-fluorophore speckle imaging, they found that in contrast to the textbook view of long microtubules emanating from the spindle poles, the entire length of the spindle is composed of relatively short microtubules.  This certainly changes my view of what a spindle looks like.  They propose that the short microtubules may turn over quickly but through crosslinking enable force generation over long distances, which would be important for chromosome segregation.

Kapoor moved on to talk about the mechanical properties of the spindle.  In this work, he and his colleagues set up a dual cantilever system to exert force on the spindles in extracts (the same extracts that Heald uses, see above).  Essentially, this lets them grab spindles and squeeze them.  Through this approach, they found that they could deform spindles but upon release, the spindles recovered their shape and size.  Interestingly, they could break the spindle if they squeezed it pole-to-pole but not if they compressed the spindle along its width.  Once broken, the shape of the spindle recovers but not the size or the elasticity (it becomes stiffer).  These biophysical approaches will surely uncover more interesting properties of this structure that can be used to examine its function.

The fourth talk of the session was from David Pellman (Dana Farber Cancer Institute).  It is commonly believed that polyploidy, aneuploidy and other forms of chromosome instability (CIN) in cancer cells occurs via erroneous cell division.  However, Pellman pointed out that we still don’t know what causes CIN in human cancers.  There are correlations between CIN and both centrosome amplification and incorrect chromosome attachments to the spindle.  Pellman set out to test the popular hypothesis that CIN caused by the presence of multiple centrosomes is due to multipolar cell division.  They engineered cells to have multiple centrosomes and then observed their mitoses.  Interestingly, they found that most multipolar mitoses are transient; the extra centrosomes cluster together and allow ‘bipolar’ spindle assembly.  Moreover, multipolar anaphases were rare and if they did occur, the cells died.  So how do the extra centrosomes promote CIN?  They found that the presence of the extra centrosomes correlated with an increased frequency in merotelic attachments—when both kinetochores are attached to microtubules emanating from the same centrosome.  Pellman predicts these attachments only weakly activate the spindle assembly checkpoint and are therefore hard to correct.  These chromosomes lag behind the others during anaphase and seem to be the cause of the aneuploidy (for more details see Ganem et al. (2009)).  They also found that the lagging chromosomes get encapsulated in a micronucleus.  They are now trying to figure out what happens to these micronuclei, another hallmark of cancer cells, and how they contribute to genome instability.

The last talk of the session was from Wim Vermeulen (Erasmus Medical Center).  He focused on a different aspect of genome maintenance: DNA repair, particularly nucleotide excision repair (NER).  There are many human disorders, such as the UV hypersensitive disorder Xeroderma pigmentosum, that are associated with mutations in genes involved in NER.  NER requires transcription to fill in the excised base and a key factor of the machinery is the transcription factor TFIIH.  Although NER is well understood in vitro, how is it organized in a complex nuclear environment?

Vermeulen described his group’s work on the in vivo dynamics of NER proteins monitored by timelapse imaging and fluorescence recovery after photobleaching (FRAP).  They found that in cultured cells, these proteins are not preassembled into a complex but rather are distributed in the nucleus.  When damage occurs, they undergo rapid and sequential assembly at the damage site.  But what about the situation in vivo?  To address this, Vermeulen’s group engineered a mouse expressing YFP-tagged TFIIH.  These mice appear normal and Vermeulen presented some of their preliminary results examining NER in vivo in different tissues.  This mouse will be a valuable tool to monitor and visualize the molecular dynamics of transcription and repair in the presence of experimental and genetic manipulations that trigger disease.

Overall, this session was a really stimulating discussion of how compromising genome integrity can lead to disease.

JCB/NYAS Meeting: Stem Cells

Thursday morning's session of the JCB/NYAS meeting focused on stem cells, and began with a talk from long-time JCB editor and recent National Medal of Science awardee Elaine Fuchs, who presented a summary of her lab’s work from recent years. If you’re unfamiliar with the Fuchs lab’s pioneering work on epidermal stem cells (Disclaimer: I’m an ex-Fuchs lab member!), there are a couple of great reviews around to bring you up to speed. Some of the more recent work discussed in this session included the identification of the embryonic stem cell population that gives rise to all the lineages of the skin epithelium (hair follicles, sebaceous glands, and epidermal cells). These stem cells are specified early in embryonic development in a manner dependent on the transcription factor Sox9, and eventually give rise to the previously-identified adult skin stem cell population (Nowak et al, 2008).

Wnt signaling plays a key role in epidermal stem cell differentiation, promoting, for example, stem cell activation or hair follicle formation, depending upon the levels of Wnt activity. Wnt signaling leads to stabilization and nuclear translocation of the transcription factor β-catenin, which acts in conjunction with members of the Lef/Tcf family of transcriptional activators and repressors. The Fuchs lab recently studied Tcf3 and Tcf4, which are co-expressed in epidermal progenitor cells and surprisingly found that these two transcription factors also have Wnt-independent functions in skin homeostasis, helping to maintain stem cells in an undifferentiated state (Nguyen et al, 2009).

Fuchs is, of course, very excited at the prevalent use of adult skin cells as the starting material for induced pluripotent stem (iPS) cell generation, so it was fitting that her presentation was followed by a talk from Alex Meissner. He’s been at the forefront of this burgeoning field as both a student/postdoc with Rudolf Jaenisch and as a group leader at Harvard University and the Broad Institute – you can read more about his background and research interests in our recent profile article. Differentiated adult cells can be reprogrammed into multipotent iPS cells with just a handful of transcription factors. Meissner is interested in understanding the mechanism by which the reprogramming occurs, in order to improve the process’ efficiency. To do this, he’s studied the epigenetic profiles of both fully and partially reprogrammed cells, identifying the DNA and histone modifications corresponding to successful reprogramming.

Meissner is also using similar approaches to determine how similar iPS cells really are to embryonic stem (ES) cells (the answer seems to be “very, but with some small differences”). The most stringent test for pluripotency is a test called tetraploid complementation, which unambiguously demonstrates whether an iPS cell can develop into a live animal. This hurdle has just been cleared for mouse iPS cells (Zhao et al., 2009) but such tests are obviously not possible for human iPS lines. Studies on the epigenomes of human iPS cells will therefore be particularly important for monitoring both the efficiency and efficacy of reprogramming. To this end, Meissner is involved in efforts to produce a reference epigenomic map of human ES cells – this is revealing a lot of variation between different ES cell lines, meaning that there is a lot of work to do before we can draw conclusions about the quality of human iPS cells.

Next came a talk from Fiona Doetsch (Columbia University), who discussed her studies on adult brain stem cells. The presentation largely covered two recently published stories from the Doetsch lab: firstly, the identification of a vascular niche for stem cells in the neural subventricular zone (SVZ). SVZ stem cells are astrocytes that directly contact blood vessels and seem to receive activating signals from the blood due to specializations in the blood brain barrier in this region of the brain (Tavazoie et al., 2008). Secondly, Doetsch described the isolation of micro RNAs that regulate the differentiation of these neural stem cells. One of these, miR-124, is highly expressed in neuroblasts (developing neurons), and overexpressing the RNA promoted neuroblast formation. Knockdown of miR-124, on the other hand, kept SVZ cells dividing, but reduced their differentiation into neurons. miR-124 regulates the timing of SVZ cell differentiation by promoting cell cycle exit and inhibiting stem cell self-renewal. It does this by downregulating a number of targets identified by the Doetsch lab, including the transcription factor Sox9 (see Cheng et al, 2009 and cf. Elaine Fuchs’ epidermal progenitor cells above!).

Irina Conboy from The Berkeley Stem Cell Center was next up to the podium. Conboy is fascinated with aging – she began her talk by comparing the lifespans of rats (3 years) and squirrels (40 years). This huge difference between two ostensibly similar animals suggests to Conboy that aging is not simply controlled by the accumulation of damage over time. Conboy works on muscle, which repairs itself easily when young but not so well as time goes by, even though the number of stem cells (aka satellite cells) remains the same. Conboy has shown that the balance between Notch and TGFβ signaling is critical to determining whether the satellite cells can activate and expand to repair tissue damage (Carlson et al., 2008) and it now seems that a similar mechanism is at play in human cells too. Old muscle can be rejuvenated by restoring the Notch/TGFβ balance, either by knocking down SMADs (the downstream effectors of TGFβ) or by elevating Notch expression.

Parabiosis experiments (in which the circulatory systems of young and old mice are connected together) have shown that systemic factors in the old animal’s blood can prematurely age young muscle stem cells. TGFβ is elevated in the sera of older mice and humans, but Conboy showed that this signaling molecule isn’t acting systemically to inhibit regeneration. TGFβ only seems to act locally in determining aging – the hunt for the systemic factor that ages the stem cell niche goes on (Carlson et al., 2009). You can find out more about Irina Conboy’s work in an upcoming profile article, which will appear in the October 5th issue of the JCB.

The final speaker in the session was Michael Buszczak, from UT Southwestern. Buszczak is interested in histone modifications that silence genes in Drosophila germline stem cells (GSCs), and has used a high throughput “protein trapping” approach to identify proteins involved in GSC maintenance and differentiation. (GFP is randomly inserted throughout the genome, and fly embryos are screened for GFP-tagged proteins with interesting localization and expression patterns). One gene that Buszczak identified in this way is a histone deubiquitinating enzyme called Scrawny (Buszczak et al., 2009). Scrawny silences genes in GSCs (maintaining them in an undifferentiated state) by removing ubiquitin from histone H2B, which would otherwise serve to recruit RNA Pol II. Intriguingly, the human homolog of Scrawny is overexpressed in human ovarian cancers.

Buszczak is now complementing this random search for chromatin modifiers by systematically knocking out enzymes such as histone demethylases and deacetylases. Targeted knockouts in Drosophila are still fairly laborious, even using the new recombineering approach developed by Hugo Bellen’s lab so, in the meantime, Buszczak has begun analyzing a few existing mutants for GSC defects. Buszczak described one histone demethylase mutant that has a non-cell autonomous function in GSC differentiation.

The five talks covered a lot of ground but the connections to the rest of the meeting were clear. Stem cells, and the transcription factors and epigenetics that regulate them, are increasingly relevant to understanding disease from aging to cancer.

JCB Meeting: The Cell Biology of Disease

Our meeting at the New York Academy of Sciences is now underway. There's been some great talks so far - look out for blog updates with more details.
In the meantime, here's a snippet of JCB editor-in-chief Tom Misteli's opening remarks from yesterday evening: