Derek Ingram recently asked about using nuclear volume measurements to
determine ploidy levels. Setting aside the volume determination for the moment, the real question is, can one accurately determine ploidy level based on nuclear area as projected in an image? This subject was extensively researched in the 1960's, and one should look up papers by Swartz related to such studies in rodent liver. These older studies demonstrated that only in the case of mammalian liver could nuclear diameter (in thick sections) be demonstrated to have a consistent relationship to nuclear ploidy, as determined by Feulgen absorption photometry. Remember that all of this absorption photometry (called microspectrophotometry at the time) was done using end window photomultipliers, with small spot sizes on the optical axis of the device, or with scanning stages or Nipkow disc type scanners. Anyone contemplating doing Feulgen ploidy determinations should become familiar with the extensive literature of the 1950's and 1960's. With careful work, CV's of approximately 1% (as good as flow cytometry) can be achieved with careful work. Remember that the flow cytometer was invented specifically to speed up ploidy determinations. However, by the time the instrument was sufficiently developed, diagnosticians had lost interest in ploidy (even though the literature was clear that this was a powerful diagnostic technique). The flow cytometer simply languished until the advent of monoclonal antibodies. With respect to fluorescence, this technique was introduced into microspectrophotometry as a way to avoid the necessity for scanning systems to integrate signal. One of the authors who contributed greatly to this effort (1960's) was Frank. The many problems inherent in fluorescence made this a problematic technique. Issues are stochiometry of dye binding, fading of dues, evenness of illumination, energy transfer, and stability of light sources. Many of these issues are still very real, yet are generally ignored by modern workers. The issue of even illumination is a major problem in excitation of fluorescence. In absorption photometry (brightfield), field correction is necessary. Assuming a Feulgen stained specimen, only nuclei should be stained. Any area of the slide that does not contain stained or unstained specimen should have the exact same value, it integrated nuclear measurements are to be meaningful. Microscope optics simply cannot do this, without a correction for field illumination. The routines often coded into cameras or capture software are not very good at this. Seldom do they provide an constant value for "background". In a transmitted light system, every "white" area of the specimen must have an identical value for integrated stained specimen components to be useful. In the case of stained nuclei, one way to check validity of data is to measure the same object (nucleus) at many different locations in the field of view. The obtained value should be the same, assuming identical illumination intensity, exposure, etc. The issue of immunohistochemical staining raises a number of problems that are ignored in current literature, and in fact, there are many papers in the literature in the recent past that are simply misleading because the technique used to generate the data are based on false assumptions. First, there are no current methods in use that can control the multiple steps of amplification that are inherent in immunostaining. Without this type of control, any derived data is meaningless. Even worse, the current practice in the U. S. is to employ peroxidase as a readout enzyme system, with DAB as the chromogen. DAB cannot be used for absorption photometry, as it does not meet the Beer-Lambert criteria for a phometric material. It is not a true absorbing material, as it is actually a particulate (the reason it is an effective stain for electron microscopy). As the concentration of DAB in a stained specimen increases, more and more light is scattered outside the capture cone of the objective, and therefore any measured signal is very nonlinear. It might be assumed that one could create a calibration chart, much as was done with photographic grains for counting of autoradiographs. This does not work for DAB, as the particle size changes from manufacturer to manufacturer and with age of the DAB solution. Whether this is a real change of particle size, or is due to aggregation of particles has yet to be determined. I hope this clarifies some of the recent issues raised regarding ploidy and stochiometry. Al Floyd 23126 South Shore Drive Edwardsburg, MI 49112 Phone: 269.699.7182 Mobile: 574.215.0703 |
It sounds like fluorescence of nuclei would be the way to go if the
mentioned factors could be adequately controlled. I wasn't involved in the work but the professor I worked for in the early 70's used the Feulgen staining and fluorescence of the Schiff's reagent and a photometer to measure ploidy levels in root tissues; it was a Leitz Ploem system with pinholes to isolate the nucleus to be measured. I recall that there were preps of chicken RBCs (?) as a calibration/standard. In root tips with rapid cell division I recall a nice bimodal distribution was typical; blocking with colchicine gave a switch to the higher peak. Older root cells often had multiples of the 2C level. It all seemed very reliable even back then - I recall them letting everthing warm up and stabilize thouroughly. Certainly instruments are more stable today, and LED light sources should fix the excitation stability issue. Capturing a widefield image can occur fairly quickly and so little bleaching should occur while an image is captured. The pararosanilin-based Schiff's reagent is very fluorescent (green excitation, Rhodamine set works nicely) and seems very stable. Fluorescence should allow a much higher SNR relative to abasorbance measurement. Antifade could be used to make it better. A digital widefield image (with no saturation and a flat prep) should allow multiple nuclei to be measured in one field using the ImageJ features to integrate the pixel values in each delineated nucleus. Dale Alton D Floyd wrote: > ---------------------- Information from the mail header ----------------------- > Sender: ImageJ Interest Group <[hidden email]> > Poster: Alton D Floyd <[hidden email]> > Subject: Re; Nuclear volume, ploidy and Stain stochiometry > ------------------------------------------------------------------------------- > > Derek Ingram recently asked about using nuclear volume measurements to > determine ploidy levels. Setting aside the volume determination for the > moment, the real question is, can one accurately determine ploidy level > based on nuclear area as projected in an image? This subject was > extensively researched in the 1960's, and one should look up papers by > Swartz related to such studies in rodent liver. These older studies > demonstrated that only in the case of mammalian liver could nuclear > diameter (in thick sections) be demonstrated to have a consistent > relationship to nuclear ploidy, as determined by Feulgen absorption > photometry. Remember that all of this absorption photometry (called > microspectrophotometry at the time) was done using end window > photomultipliers, with small spot sizes on the optical axis of the > device, or with scanning stages or Nipkow disc type scanners. Anyone > contemplating doing Feulgen ploidy determinations should become familiar > with the extensive literature of the 1950's and 1960's. With careful > work, CV's of approximately 1% (as good as flow cytometry) can be > achieved with careful work. Remember that the flow cytometer was > invented specifically to speed up ploidy determinations. However, by the > time the instrument was sufficiently developed, diagnosticians had lost > interest in ploidy (even though the literature was clear that this was a > powerful diagnostic technique). The flow cytometer simply languished > until the advent of monoclonal antibodies. > > With respect to fluorescence, this technique was introduced into > microspectrophotometry as a way to avoid the necessity for scanning > systems to integrate signal. One of the authors who contributed greatly > to this effort (1960's) was Frank. The many problems inherent in > fluorescence made this a problematic technique. Issues are stochiometry > of dye binding, fading of dues, evenness of illumination, energy > transfer, and stability of light sources. Many of these issues are still > very real, yet are generally ignored by modern workers. The issue of > even illumination is a major problem in excitation of fluorescence. > > In absorption photometry (brightfield), field correction is necessary. > Assuming a Feulgen stained specimen, only nuclei should be stained. Any > area of the slide that does not contain stained or unstained specimen > should have the exact same value, it integrated nuclear measurements are > to be meaningful. Microscope optics simply cannot do this, without a > correction for field illumination. The routines often coded into cameras > or capture software are not very good at this. Seldom do they provide an > constant value for "background". In a transmitted light system, every > "white" area of the specimen must have an identical value for integrated > stained specimen components to be useful. In the case of stained nuclei, > one way to check validity of data is to measure the same object (nucleus) > at many different locations in the field of view. The obtained value > should be the same, assuming identical illumination intensity, exposure, > etc. > > The issue of immunohistochemical staining raises a number of problems > that are ignored in current literature, and in fact, there are many > papers in the literature in the recent past that are simply misleading > because the technique used to generate the data are based on false > assumptions. First, there are no current methods in use that can control > the multiple steps of amplification that are inherent in immunostaining. > Without this type of control, any derived data is meaningless. Even > worse, the current practice in the U. S. is to employ peroxidase as a > readout enzyme system, with DAB as the chromogen. DAB cannot be used for > absorption photometry, as it does not meet the Beer-Lambert criteria for > a phometric material. It is not a true absorbing material, as it is > actually a particulate (the reason it is an effective stain for electron > microscopy). As the concentration of DAB in a stained specimen > increases, more and more light is scattered outside the capture cone of > the objective, and therefore any measured signal is very nonlinear. It > might be assumed that one could create a calibration chart, much as was > done with photographic grains for counting of autoradiographs. This does > not work for DAB, as the particle size changes from manufacturer to > manufacturer and with age of the DAB solution. Whether this is a real > change of particle size, or is due to aggregation of particles has yet to > be determined. > > I hope this clarifies some of the recent issues raised regarding ploidy > and stochiometry. > > Al Floyd > 23126 South Shore Drive > Edwardsburg, MI 49112 > Phone: 269.699.7182 > Mobile: 574.215.0703 |
In reply to this post by Alton D Floyd
Dr. Floyd,
I appreciate your explanation on the stoichiometric problem. Regarding this part: > it might be assumed that one could create a calibration chart, much as was > done with photographic grains for counting of autoradiographs. This does > not work for DAB, as the particle size changes from manufacturer to > manufacturer and with age of the DAB solution. Should I take it then that in autoradiographic emulsions the silver particles are uniform in size in and do not change from manufaturer to manufaturer, age, etc? If so can you recommend some reading on this? I searched but found nothing. > Al Floyd > 23126 South Shore Drive > Edwardsburg, MI 49112 > Phone: 269.699.7182 > Mobile: 574.215.0703 -- Hugo Arruda de Moura Torres ================================== Departamento de Biofísica Universidade Federal de São Paulo Rua Botucatu 862 7o. andar CEP 04023-062 Vila Clementino São Paulo - SP - Brasil Tel:+55 (11) 5576 4530 r.220 Fax: 55 11 5571 5780 |
In reply to this post by Alton D Floyd
For autoradiographic grain counting, the type of emulsion, developer
used, and method of development all influences the grain size and morphology. When this technique was widely used, in the early 1960's, most investigators used developers that produced tiny, round grains. As in any other type of quantitative measurement, the specimen preparation and grain development must be highly controlled. When I was actively involved in this type of work, we found that we had to create new calibrations each time we used a new batch of emulsion. The calibration was created by making a series of slides, coating with emulsion, and exposing for increasing lengths of time. Then we took reflectance measurements, and compared these to actual grain counts in the measured areas. With such a calibration, we found excellent correlation between actual counts and reflectance measurements. Obviously, the measurements were much faster, and permitted studies such as localization of radioactive tagged materials in serial sections of organs, such as rodent brains. I don't remember all of the citations to these type of studies. I will check my old citation card file and see if I can come up with specific citations that will give you a starting point for a citation search. Al Floyd 23126 South Shore Drive Edwardsburg, MI 49112 Phone: 269.699.7182 Mobile: 574.215.0703 |
First of all I would like to thank everyone for their input and interest
regarding my initial posting. I should elaborate on my initial question. Annually the company I work for evaluates ploidy via Propidium Iodide and flow cytometry. Alternatively we would like to develop a technique to determine ploidy in various strains of salmonids using IMAGEJ and preferably unstained blood smears. I would be more then happy to forward some journal articles to whomever is interested regarding ploidy determination through means of nuclear measurement. It has been done before. The reason I am trying to utilize IMAGEJ software is to speed up the processing of samples. I have no problem measuring total cell measurements with other software. The issue I have with IMAGEJ is isolating the nucleus, I'm sure this should be a simple task, but due to my inexperience with the software I am having some issues. Again I would like to thank everyone who is responding. Derek Ingram -----Original Message----- From: ImageJ Interest Group [mailto:[hidden email]] On Behalf Of Alton D Floyd Sent: Wednesday, February 18, 2009 10:49 AM To: [hidden email] Subject: Re: Re; Nuclear volume, ploidy and Stain stochiometry For autoradiographic grain counting, the type of emulsion, developer used, and method of development all influences the grain size and morphology. When this technique was widely used, in the early 1960's, most investigators used developers that produced tiny, round grains. As in any other type of quantitative measurement, the specimen preparation and grain development must be highly controlled. When I was actively involved in this type of work, we found that we had to create new calibrations each time we used a new batch of emulsion. The calibration was created by making a series of slides, coating with emulsion, and exposing for increasing lengths of time. Then we took reflectance measurements, and compared these to actual grain counts in the measured areas. With such a calibration, we found excellent correlation between actual counts and reflectance measurements. Obviously, the measurements were much faster, and permitted studies such as localization of radioactive tagged materials in serial sections of organs, such as rodent brains. I don't remember all of the citations to these type of studies. I will check my old citation card file and see if I can come up with specific citations that will give you a starting point for a citation search. Al Floyd 23126 South Shore Drive Edwardsburg, MI 49112 Phone: 269.699.7182 Mobile: 574.215.0703 |
In reply to this post by Alton D Floyd
On Wed, 2009-02-18 at 13:48 -0500, Alton D Floyd wrote:
> For autoradiographic grain counting, the type of emulsion, developer > used, and method of development all influences the grain size and > morphology. When this technique was widely used, in the early 1960's, > most investigators used developers that produced tiny, round grains. As > in any other type of quantitative measurement, the specimen preparation > and grain development must be highly controlled. > > When I was actively involved in this type of work, we found that we had > to create new calibrations each time we used a new batch of emulsion. > The calibration was created by making a series of slides, coating with > emulsion, and exposing for increasing lengths of time. Then we took > reflectance measurements, and compared these to actual grain counts in > the measured areas. With such a calibration, we found excellent > correlation between actual counts and reflectance measurements. > Obviously, the measurements were much faster, and permitted studies such > as localization of radioactive tagged materials in serial sections of > organs, such as rodent brains. > I don't remember all of the citations to these type of studies. I will > check my old citation card file and see if I can come up with specific > citations that will give you a starting point for a citation search. Excellent! I am interest in how those control experiments should be conducted in more detail, I could start by reading one of your papers. Would you recommend one in particular? Setting aside the problem of silver particle size in emulsion radiography for a moment and back to the problem of DAB semi-quantification, Dr. Al Floyd said previously: > ... DAB cannot be used for absorption photometry, as it does not meet the Beer-Lambert criteria for > a phometric material. It is not a true absorbing material, as it is > actually a particulate (the reason it is an effective stain for electron > microscopy). As the concentration of DAB in a stained specimen > increases, more and more light is scattered outside the capture cone of > the objective, and therefore any measured signal is very nonlinear. Dr Floyd, sir, after reading this statement I am really intrigued: an article from the American Journal of Pathology (Open Access PMID: 11159179) claims that at least three times the linearity between DAB absorbance or concentration staining and antigen concentration was established by incorporating antigens and antibodies at different concentration combinations in mounting media and therefore suitable for densitometry (references number 32,34,43). Unfortunately I don't have access to these works from my institution and I can't ponder on their quality as of now. I did find another work though, where a similar approach was taken and it seemed pretty robust for me. Now, I am not particularly experienced, so I would like if you could take a critical look (or anyone else in this list, I gladly want to hear): PMID 8773563 Also mentioned: > It might be assumed that one could create a calibration chart, much as > was done with photographic grains for counting of autoradiographs. > This does not work for DAB, as the particle size changes from > manufacturer to manufacturer and with age of the DAB solution. True, but by creating standards one would still be able to compare, semi-quantitatively, relative amounts of antigen between experimental and control groups, say, expression levels of neuropeptide P between transgenic and naive mice groups. -- Hugo Arruda de Moura Torres ================================== Departamento de Biofísica Universidade Federal de São Paulo Rua Botucatu 862 7o. andar CEP 04023-062 Vila Clementino São Paulo - SP - Brasil Tel:+55 (11) 5576 4530 r.220 Fax: 55 11 5571 5780 |
Free forum by Nabble | Edit this page |