On 2014 Mar 10, Madhusudana Girija Sanal commented:

You observed that significantly more CNVs are present in early-passage human iPS cells than intermediate passage human iPS cells or human ES cells. Can the reason for the observed improvements in genetic quality be a conscious selection by a trained eye for ideal hES cell morphology resulting in improvement of the quality of iPSC over passages-(although the investigators were not knowing that they are selecting for genetically/epigenetically normal cells)!

It has been observed by several groups that the genetic/epigenetic quality of iPSC improves over passages (1,2). The reasons could be DNA repair mechanisms acting over time, natural selection or human (artificial) selection over passages. Primary cells become adapted over time in culture conditions, short term adaptations due to changes in regulation and expression of various genes at transcription and translational levels and later by acquiring genetic changes including mutations, copy number variations, deletions and probably epigenetic changes (3-5). It is observed that cells will accumulate mutations under the ‘natural’ selection pressure which favor in-vitro growth and proliferation and most of the robust cell lines have mutations which make them successful for in-vitro growth. This is a contrast to iPSC which improves the genetic quality over passage. Morphology identified by human eye is very powerful in identifying subtle differences in phenotype which in turn reflects the changes in the genotype. This is why a trained person can identify types of cancer from a simple H&E stained section without going for genetic markers or one can distinguish between the identical twins. In case of iPSC, every time the investigator observes the culture dishes under microscope, he discards ‘bad looking’ colonies and selects ‘good-healthy’ colonies. In this process he unknowingly selects for genetically/epigenetically normal cells. It may be noted that all of the established human embryonic stem (hES) cell lines have undergone several passages and hence careful human selection and hence less genetic defects compared to relatively ‘fresh’ (low passage number) iPSC. Conscious selection by a trained eye for ideal hES cell morphology could be the prime reason for the observed improvement of the quality of iPSC over passages.

"majority of the cells, whether during the early or late passages, had the ideal hES morphology" The first author notes. But majority is a relative term. Anyway, the majority of the IPSC has normal genetics also.

Please see an extract from first and second page of your article : "A control probe was selected from a chromosome 1 location that showed normal copy number across all human iPS cell lines that were tested. During early passages, the test probe demonstrated a significantly higher fraction of cells with aberrant copy number state than the control probe. The fraction of aberrant cells was also significantly higher in early-passage human iPS cells (18%) than in fibroblasts (3%) or in later-passage human iPS cells (9%)".

"The median number of CNVs in human iPS cell lines was about twofold higher than in human ES cell lines and fibroblasts" This is expected because most of the established human ES lines have undergone several passages compared to relatively "freshly brewed iPSC" and hence more rounds of selection by an expert's eye!

It may be noted that morphology is one or the most important markers of ES cells, all others- alkaline PO4ase staining, expression of Nanog, TRAs are non-specific. They are present in several cancer cell lines also. In contrast, typical ES cell morphology is quite rare among non-ES cells.

A note on the methods: "consistently passaged the cells enzymatically (Collagenase or dispase), and as such this alleviates bias that may occur by manual passaging of ideal hESCs"

Suppose you have a dish of IPSC cells, say 40-60 colonies and you find one or two colonies looking "bad" or atypical or 'differentiated', then you will most likely discard the entire dish. You will not attempt an enzymatic splitting and passaging. This is what we do in our lab. Suppose the ES/IPSC line is precious, then we would do? We would manually pick good colonies by cutting them out using a pipette tip. In both cases we are doing the same.

You believe that "most of the selection pressure may actually occur through the fact that several of the mutations observed in early passage hiPSCs (as highlighted in your paper, supplementary table 8) carry the possibility of hindering hESC maintenance and self-renewal."

I will say several of the mutations which hinders with maintenance/self-renewal will not be selected because as soon as they show these phenotypic changes the experimenter will discard them!

"hiPSCs are genetically mosaic, more so in early passage than in later passages". Moreover, SNP arrays (as highlighted in your paper) would detect CNVs that are found in the majority of the cells.

I would explain this fact, this way: This can be because of inefficiency of the method -the positive predictive value versus negative predictive value of the test! Again, our key question why more defects in early passages? My answer is that with every passage genetically/epigenetically bad colonies were discarded by the experimenter. It is possible that mosaics are generated during enzyme treatment (collagenase/trypsin/dispase) when two or more monoclonal colonies gets mixed up. So these facts cannot be followed up without studying and following up iPSC clonal selection and proliferation.

It is possible that genetically defective cells, parasite on (or may be complementing!) the growth and maintenance of IPSC colonies which is facilitated by mosaic-ism.

References:

1. Mayshar Y et al. Identification and classification of chromosomal aberrations in human induced pluripotent stem cells. Cell Stem Cell. 7:521-31(2010).

2. Hussein, S. M. et al. Nature 471, 58–62 (2011).

3. http://www.sanger.ac.uk/genetics/CGP/CellLines/

4. Freshney, R.I. (2005) Culture of Animal Cells, a Manual of Basic Technique, 5th Ed. Hoboken NJ, John Wiley & Sons.

5. Chen J et al. Evaluation of x-inactivation status and cytogenetic stability of human dermal fibroblasts after long-term culture. Int J Cell Biol.2010:289653 (2010).

On 2014 Mar 10, Madhusudana Girija Sanal commented:

You observed that significantly more CNVs are present in early-passage human iPS cells than intermediate passage human iPS cells or human ES cells. Can the reason for the observed improvements in genetic quality be a conscious selection by a trained eye for ideal hES cell morphology resulting in improvement of the quality of iPSC over passages-(although the investigators were not knowing that they are selecting for genetically/epigenetically normal cells)!

It has been observed by several groups that the genetic/epigenetic quality of iPSC improves over passages (1,2). The reasons could be DNA repair mechanisms acting over time, natural selection or human (artificial) selection over passages. Primary cells become adapted over time in culture conditions, short term adaptations due to changes in regulation and expression of various genes at transcription and translational levels and later by acquiring genetic changes including mutations, copy number variations, deletions and probably epigenetic changes (3-5). It is observed that cells will accumulate mutations under the ‘natural’ selection pressure which favor in-vitro growth and proliferation and most of the robust cell lines have mutations which make them successful for in-vitro growth. This is a contrast to iPSC which improves the genetic quality over passage. Morphology identified by human eye is very powerful in identifying subtle differences in phenotype which in turn reflects the changes in the genotype. This is why a trained person can identify types of cancer from a simple H&E stained section without going for genetic markers or one can distinguish between the identical twins. In case of iPSC, every time the investigator observes the culture dishes under microscope, he discards ‘bad looking’ colonies and selects ‘good-healthy’ colonies. In this process he unknowingly selects for genetically/epigenetically normal cells. It may be noted that all of the established human embryonic stem (hES) cell lines have undergone several passages and hence careful human selection and hence less genetic defects compared to relatively ‘fresh’ (low passage number) iPSC. Conscious selection by a trained eye for ideal hES cell morphology could be the prime reason for the observed improvement of the quality of iPSC over passages.

"majority of the cells, whether during the early or late passages, had the ideal hES morphology" The first author notes. But majority is a relative term. Anyway, the majority of the IPSC has normal genetics also.

Please see an extract from first and second page of your article : "A control probe was selected from a chromosome 1 location that showed normal copy number across all human iPS cell lines that were tested. During early passages, the test probe demonstrated a significantly higher fraction of cells with aberrant copy number state than the control probe. The fraction of aberrant cells was also significantly higher in early-passage human iPS cells (18%) than in fibroblasts (3%) or in later-passage human iPS cells (9%)".

"The median number of CNVs in human iPS cell lines was about twofold higher than in human ES cell lines and fibroblasts" This is expected because most of the established human ES lines have undergone several passages compared to relatively "freshly brewed iPSC" and hence more rounds of selection by an expert's eye!

It may be noted that morphology is one or the most important markers of ES cells, all others- alkaline PO4ase staining, expression of Nanog, TRAs are non-specific. They are present in several cancer cell lines also. In contrast, typical ES cell morphology is quite rare among non-ES cells.

A note on the methods: "consistently passaged the cells enzymatically (Collagenase or dispase), and as such this alleviates bias that may occur by manual passaging of ideal hESCs"

Suppose you have a dish of IPSC cells, say 40-60 colonies and you find one or two colonies looking "bad" or atypical or 'differentiated', then you will most likely discard the entire dish. You will not attempt an enzymatic splitting and passaging. This is what we do in our lab. Suppose the ES/IPSC line is precious, then we would do? We would manually pick good colonies by cutting them out using a pipette tip. In both cases we are doing the same.

You believe that "most of the selection pressure may actually occur through the fact that several of the mutations observed in early passage hiPSCs (as highlighted in your paper, supplementary table 8) carry the possibility of hindering hESC maintenance and self-renewal."

I will say several of the mutations which hinders with maintenance/self-renewal will not be selected because as soon as they show these phenotypic changes the experimenter will discard them!

"hiPSCs are genetically mosaic, more so in early passage than in later passages". Moreover, SNP arrays (as highlighted in your paper) would detect CNVs that are found in the majority of the cells.

I would explain this fact, this way: This can be because of inefficiency of the method -the positive predictive value versus negative predictive value of the test! Again, our key question why more defects in early passages? My answer is that with every passage genetically/epigenetically bad colonies were discarded by the experimenter. It is possible that mosaics are generated during enzyme treatment (collagenase/trypsin/dispase) when two or more monoclonal colonies gets mixed up. So these facts cannot be followed up without studying and following up iPSC clonal selection and proliferation.

It is possible that genetically defective cells, parasite on (or may be complementing!) the growth and maintenance of IPSC colonies which is facilitated by mosaic-ism.

References:

1. Mayshar Y et al. Identification and classification of chromosomal aberrations in human induced pluripotent stem cells. Cell Stem Cell. 7:521-31(2010).

2. Hussein, S. M. et al. Nature 471, 58–62 (2011).

3. http://www.sanger.ac.uk/genetics/CGP/CellLines/

4. Freshney, R.I. (2005) Culture of Animal Cells, a Manual of Basic Technique, 5th Ed. Hoboken NJ, John Wiley & Sons.

5. Chen J et al. Evaluation of x-inactivation status and cytogenetic stability of human dermal fibroblasts after long-term culture. Int J Cell Biol.2010:289653 (2010).