These are notes from lecture 12 of Harvard Extension’s Cell Biology course.

what are stem cells

A stem cell is a cell with two key properties:

  • self-renewal
  • differentiation

In other words, they can divide to form more stem cells or to form more committed lineages of cells.  Stem cell populations are maintained either by asymmetric mitosis, where one daughter cell becomes differentiated and the other remains as a stem cell, or by stochastic differentiation, such that some mitosis events in a population will result in two stem cells while others will lead to two differentiated cells.

The use of the word “stem” here is believed to have arisen from analogy to the stem of a plant, which can branch out into terminal leaves.

Here is a lengthy introductory video:

how pluripotent is pluripotent?

We can think of a hierarchy of stem cells, in terms of how diverse of lineages they can give rise to:

  • Totipotent – can make all possible cell types of an organism. (zygote, morula)
  • Pluripotent – can make cells of any germ layer (endoderm, mesoderm, ectoderm), but not extraembryonic tissue, ex. the placenta.  ESCs and iPSCs in the lab cannot become placental tissue (which requires CDX2 expression).
  • Multipotent – can make different cell types within a tissue. For instance hematopoetic stem cells (HSCs) can make a variety of blood cell types.
  • Unipotent – can only make one cell type.  These do not have the property of differentiation, but do have the property of self-renewal.

As cells specialize, they lose their pluripotency.  This is hypothesized to have evolved as a defense against diseases.

Tissue-specific stem cells can be slow cycling or quiescent at steady state, but in response to stimuli can rapidly ramp up production of themselves.  An example is hematopoietic stem cells.  Differentiated blood cells have high turnover – neutrophils last ~12h, red blood cells last ~120d.  These are the best-characterized adult stem cells, well-studied since the 1960s.  They are used clinically in bone marrow transplantation.   By the way, “cord blood” is largely hematopoietic stem cells.

Stem cells live in a stem cell ‘niche’: a supportive environment made of cells and extracellular material that regulate the stem cells’ fate decisions.  Cellular signals may be secreted or transmitted through direct cell-cell contact.

stem cells in the lab

Embryonic stem cells (ESCs) used in laboratory settings are derived from culturing the inner cell mass of a blastocyst.  Because the inner mass does not include the Cdx2-positive outer layer destined for placental cell fate, ESCs are pluripotent, not totipotent.

Ways to reprogram cells:

All of these properties result in epigenetic changes to DNA that give rise to pluripotency.

experimental methods

To determine whether an isolated cell population is multipotent, label them (e.g. with GFP), then transplant them into a non-labeled irradiated recipient animal.  Irradiation prevents the host animal from rejecting the transplanted tissue and clears the stem cell niche.  Then watch to see if the GFP-expressing cells populate the host animal, and what cell types they end up being.  For example, if you see erythroid, myeloid and T, B, and NK cells, then that means you got HSCs to start with.  If you only get erythroid and myeloid cells, you had probably isolated the CMPs.

Cell fate mapping is well-studied in C. elegans because the origin and fate of every cell are known.

random grab-bag of facts about cancers

Many cancers display stem-like characteristics, but without the ‘checks and balances’: apoptosis is inhibited, proliferation is largely unchecked, and differentiation may be blocked.  (By the way, warts, moles and skin tags are all benign tumors).

Here are six changes that often (but not necessarily) occur in cancers:

  • Proliferation – tendency to divide
  • “Enabling replicative immortality” – (how is this any different from the previous bullet?)
  • Evasion of growth suppressors – ways of ignoring anti-growth signals from nearby cells
  • Angiogenesis – ability to induce vascularization so that the tumor continues to have access to nutrients.
  • Resisting death – ways of ignoring pro-apoptotic signals from within or without
  • Invasion & metastasis – ways of escaping from own tissue and invading other tissues

Two major classifications of cancer:

  • Carcinoma – arise from epithelia (endoderm) or ectoderm.  Gut, skin, nervous system.
  • Sacroma – arise from mesoderm. Muscle, blood, and their precursors.

Glioblastoma is not quite either of these, arising from glial cells in the brain.

Tumors can be characterized in vitro as having undergone “transformation” if they can grow indefinitely in the absence of growth factors on a variety of media.  Human tumors are often studied by xenograft into model organisms, to study angiogenesis, metastasis and other processes in real-time.

Sometimes inherited germline mutations contribute to cancer, even if not oncogenic on their own.  Other genetic lesions contributing to cancer are somatic mutations, arising spontaneously or in response to mutagens such as cigarette smoke, asbestos, etc.

How do proto-oncogenes convert to full-on oncogenes?  Four types of mutations often contribute:

  1. SNPs that created a constitutively active protein product.  Examples include mutations in RTK that mimic ligand binding by introducing an amino acid substitution that mimics the phosphorylated state of the protein, and mutations that make Ras less able to hydrolyze GTP, thus making it ‘activated’ by GTP binding all the time.
  2. Gene fusion by chromosomal translocations, creating a novel fused protein product. For instance, a chr9-chr22 translocation creates a BCR-ABL fusion protein under the BCR promoter, resulting in overexpression of a constitutively active ABL mutant which promotes cell proliferation, causing chronic myelogenous leukemia.  Imatinib targets the BCR-ABL fusion protein very specifically (it can’t bind to wild-type ABL) and is a very effective treatment.
  3. Chromosomal translocation moving a gene under a new regulatory region, resulting in overexpression or ectopic expression.
  4. Duplication of a gene leading to overexpression.

Tumor suppressor genes are usually haplosufficient, so both copies must be lost in order to give rise to a cancer.  Most tumor suppressors fall into one of the following categories:

  1. Cell cycle regulators (ex. p16, Rb)
  2. Receptors or signal transducers (ex. TGF-beta).  Examples include a V>Q HER2 mutation which allows kinase activity in the absence of a ligand, and a deletion of the Erb2 extracellular ligand-binding domain, allowing the intracellular domains to dimerize and activate without ligand.  Receptors may also be activated by viral proteins that mimic signaling molecules.
  3. Checkpoint controllers (ex. p53)
  4. Pro-apoptotic.  For instance, loss of Fas and Fas ligand. Mutations in p53 also lead to a loss of Bax expression.  (Gain of function of anti-apoptotic genes will also do: overexpression of Bcl-2 can happen in cancers).
  5. Caretaker (ex. DNA repair enzymes)

Here’s a detailed example involving TGF-beta.  Mutations in TGF-beta receptors can cause a loss of incoming pro-apoptotic signal, shutting off the Smad signaling pathway and causing a loss of p15 and Pai-1 exepression.  Loss of p15 deregulates the cell cycle, and loss of Pai-1 allows invasion of the extracellular matrix and escape from the tissue for metastasis.   In order to metastasize, cancer cells must get through the extracellular matrix.  They harness cofilin and WASP in order to form an “invadopodium”.  You can see hair-like structures on some tumor cells in micrographs, which are precursors of invadopodium.

Most cancers are at least ‘two-hit’, involving both activation of oncogenes and inactivation of tumor suppressors.  One interesting line of evidence for a ‘multi-hit’ model of tumor formation is the age-related incidence of cancers.  For a variety of cancers, incidence rises exponentially with age from < 1/100,000 individuals in their 20s – 40s up to ~100-500/100,000 by people’s 70s and 80s. This suggests it is the accumulation of multiple somatic mutations in the same cell which finally allow tumor formation.  For instance, in colon cancer, loss of APC alone will cause small benign polyps. Only when combined with loss of function of K-ras will it cause larger (but still benign) polyps.  Finally when p53 is lost, then the tumor is capable of metastasizing.  Current estimates for the most lethal cancers suggest it takes 5-6 ‘hits’ i.e. 5-6 different genes being mutated, before a malignant tumor can form.  Mouse models also support the multi-hit model.  Mice overexpressing myc or expressing mutant rasD will both get tumors, but double mutant mice with both of these genes get tumors much earlier in life.

One consequence of the multi-hit nature of cancers is that highly specific cancer drugs such as imatinib, which targets BCR-ABL fusion protein, may not be potent enough to stop the whole tumor.  This is why specific drugs are still often combined with global therapies (“controlled poisons”).  Radiation and chemotherapy just target all rapidly proliferating cells, which is why people lose their hair.