Meeting News

Tools to better assess tumor heterogeneity will ‘shed new light’ on cancer research, treatment

John D. Carpten PhD
John D. Carpten

ATLANTA — Emerging technology is providing researchers with the tools to better understand the relationships among cells in the tumor microenvironment that can inform the individualized treatment of patients with cancer, according to a presentation during the opening plenary session of American Association for Cancer Research Annual Meeting.

It has been established that tumor heterogeneity exists among individuals, with data demonstrating “spatial transcriptional clonal heterogeneity,” John D. Carpten, PhD, professor and chair of translational genomics and director of Institute of Translational Genomics at Keck School of Medicine at USC, and AACR program committee chair, said during his presentation.

“We know heterogeneity exists, so hopefully now we can move toward new and improved technologies that allow us to expand our ability to assess tumor heterogeneity,” he said.

Importance of tumor heterogeneity

There are two important components of assessing tumor heterogeneity among patients with cancer, Carpten said in an interview with HemOnc Today.

The first is the tumor cells.

“The tumor cells in a cancer can either be relatively clonal — meaning there is one cell that divides, and all of the progeny resemble the original progenitor cell — or there can be high degree of clonality,” he said. “This means that as the cancer is growing and dividing, new mutations are occurring in daughter cells and those daughter cells become new clones, giving birth to a group of similar cells.”

There can be multiple clones growing in a given tumor, Carpten added.

“Those different clones might actually acquire changes that make them grow faster, grow in a region without a lot of blood flow or oxygen, or make them more resistant to different types of chemotherapy, for instance,” he said. “If we aren’t able to really understand the full breadth of the cancer cells growing in the tumor, we could miss some really important molecular features that will help to better manage that cancer.”

The second component is the noncancer cells.

“As a tumor is growing, it has blood vessels going through it to feed it oxygen and other nutrients,” Carpten said. “Another important aspect are the immune cells that are there as the body tries to mount a defense against the tumor. There are different types of immune cells present that send signals to go into the tumor to fight it, whereas the tumor is sending the opposite signal, to tell the blood vessels to not let anything in. There is a tug and pull between the various types of nontumor cells and the tumor cells.”

Therefore, when assessing nontumor cells, researchers may want to understand if there are activated T cells in the tumor that can mount an effective immune response, which is central to efficacy of immunotherapy.

“There is a complex interplay between the different cell types that make up the tumor bed, and the more we can understand that, the better we can manage or treat the disease,” Carpten said.

New tools

Previous tools limited researchers’ ability to understand the full complexity of the tumor microenvironment. However, new tools allow researchers to make these measurements “in an unprecedented way,” Carpten said.

Looking at a sample from ovarian cancer, for instance, it is known that various types of cells, including tumor cells and microenvironment cells, build the tumor mass.

“We also know that we can take these sections, disassociate them mechanically and chemically into single cells, and then utilize various technologies to conduct single-cell profiling,” Carpten said during his presentation. “We can use aqueous-phase approaches and microfluidics that allow us to perform whole-transcriptome sequencing within a single droplet, and utilize that information based on molecular probes and molecular barcodes to actually create information on tumor heterogeneity and create subpopulations of cells.”

However, simply disassociating the tumor loses important spatial context, Carpten said.

“We can now imagine the opportunity to take serial sections of a specimen, lay them out on special slides to map, and create whole-transcriptome sequencing libraries on the slides for small groups of cells,” he said.

Taking these libraries, cleaving them off the slide, and performing transcriptome sequencing can capture the heterogeneity of cells across the section, maintaining spatial context, Carpten said, adding that this is crucial for understanding the complexity of cancer and tumor heterogeneity across individuals and patients.

Carpten and colleagues have been using this approach to conduct spatial whole-transcriptome sequencing of serious ovarian carcinoma with serial primary and metastatic pairs, and serial primary and recurrent pairs.

Researchers can take these various tumor clusters based on their transcriptome profiles and perform molecular annotation using tools to look at immune cluster profiling — such as CIBERSORT and xCELL — and measuring inflammation and immune scores, or study the tumor clusters with tools such as gene set enrichment analysis (GSEA) and IPA.

“We can take these data and look at immune effector cell profiling and gene expression of various genes that define these tumor inflammation scores at the single-gene level and interpolate across the specimens to look for hot and cold regions of a tumor tumor,” he said. “Tools like CIBERSORT can be used to look closely at the actual immune effector cell profiles within various microenvironment regions.”

When looking at one specific ovarian sample, researchers were surprised to find the region defined as tumor histopathologically actually could be broken out into three distinct transcriptional profiles that could not be deduced by histopathological methodology, Carpten said.

“We actually saw different gene expression profiles, processes and pathways driving these three distinct regions of tumor that make up a really intriguing transcriptional substructure across the sample,” he said. “Importantly, we’re also measuring this across the depth of the tumor in these cases using serial sections.”

‘This is the future’

Data gleaned from this deep sequencing can help clinicians to better tailor and personalize treatments.

For instance, an analysis of a tumor might reveal one group of cells expressing genes that suggest growth due to cell cycles, but a different population of cells is growing due to hormone receptor activation.

“By knowing that information, a doctor can say, ‘OK, there is a population of cells that might respond to tamoxifen, but another population of cells that might not. I might need to use chemotherapy plus tamoxifen to battle this cancer,’” Carpten told HemOnc Today. “Whereas, another cancer might be more clonal, that might be driven by a different process, like hypoxia, and we can go in and attack that cancer very specifically.”

Clinicians can understand the disease better by understanding the different tumor clones.

“Doctors can better understand which patients are likely to fail, why they are likely to fail and, if they do fail, what might be an effective way to treat that cancer when it comes back,” Carpten said. “This is the future of cancer genome science.”

Overall, these technologies that are under development can help researchers to better understand the relationships of cells that define a cancer and its microenvironment, Carpten said during his presentation.

“Integrating these cutting-edge approaches going forward will undoubtedly shed new light on cancer and will open up new and innovative ideas on how to best approach disease prevention, diagnosis and treatment for all patients,” he said. – by Alexandra Todak

Reference:

Carpten JD. Wrap-up and opportunities for the future. Presented at: AACR Annual Meeting; March 29-April 3, 2019; Atlanta.

Disclosure: Carpten reports a collaboration agreement with 10X Genomics.

John D. Carpten PhD
John D. Carpten

ATLANTA — Emerging technology is providing researchers with the tools to better understand the relationships among cells in the tumor microenvironment that can inform the individualized treatment of patients with cancer, according to a presentation during the opening plenary session of American Association for Cancer Research Annual Meeting.

It has been established that tumor heterogeneity exists among individuals, with data demonstrating “spatial transcriptional clonal heterogeneity,” John D. Carpten, PhD, professor and chair of translational genomics and director of Institute of Translational Genomics at Keck School of Medicine at USC, and AACR program committee chair, said during his presentation.

“We know heterogeneity exists, so hopefully now we can move toward new and improved technologies that allow us to expand our ability to assess tumor heterogeneity,” he said.

Importance of tumor heterogeneity

There are two important components of assessing tumor heterogeneity among patients with cancer, Carpten said in an interview with HemOnc Today.

The first is the tumor cells.

“The tumor cells in a cancer can either be relatively clonal — meaning there is one cell that divides, and all of the progeny resemble the original progenitor cell — or there can be high degree of clonality,” he said. “This means that as the cancer is growing and dividing, new mutations are occurring in daughter cells and those daughter cells become new clones, giving birth to a group of similar cells.”

There can be multiple clones growing in a given tumor, Carpten added.

“Those different clones might actually acquire changes that make them grow faster, grow in a region without a lot of blood flow or oxygen, or make them more resistant to different types of chemotherapy, for instance,” he said. “If we aren’t able to really understand the full breadth of the cancer cells growing in the tumor, we could miss some really important molecular features that will help to better manage that cancer.”

The second component is the noncancer cells.

“As a tumor is growing, it has blood vessels going through it to feed it oxygen and other nutrients,” Carpten said. “Another important aspect are the immune cells that are there as the body tries to mount a defense against the tumor. There are different types of immune cells present that send signals to go into the tumor to fight it, whereas the tumor is sending the opposite signal, to tell the blood vessels to not let anything in. There is a tug and pull between the various types of nontumor cells and the tumor cells.”

PAGE BREAK

Therefore, when assessing nontumor cells, researchers may want to understand if there are activated T cells in the tumor that can mount an effective immune response, which is central to efficacy of immunotherapy.

“There is a complex interplay between the different cell types that make up the tumor bed, and the more we can understand that, the better we can manage or treat the disease,” Carpten said.

New tools

Previous tools limited researchers’ ability to understand the full complexity of the tumor microenvironment. However, new tools allow researchers to make these measurements “in an unprecedented way,” Carpten said.

Looking at a sample from ovarian cancer, for instance, it is known that various types of cells, including tumor cells and microenvironment cells, build the tumor mass.

“We also know that we can take these sections, disassociate them mechanically and chemically into single cells, and then utilize various technologies to conduct single-cell profiling,” Carpten said during his presentation. “We can use aqueous-phase approaches and microfluidics that allow us to perform whole-transcriptome sequencing within a single droplet, and utilize that information based on molecular probes and molecular barcodes to actually create information on tumor heterogeneity and create subpopulations of cells.”

However, simply disassociating the tumor loses important spatial context, Carpten said.

“We can now imagine the opportunity to take serial sections of a specimen, lay them out on special slides to map, and create whole-transcriptome sequencing libraries on the slides for small groups of cells,” he said.

Taking these libraries, cleaving them off the slide, and performing transcriptome sequencing can capture the heterogeneity of cells across the section, maintaining spatial context, Carpten said, adding that this is crucial for understanding the complexity of cancer and tumor heterogeneity across individuals and patients.

Carpten and colleagues have been using this approach to conduct spatial whole-transcriptome sequencing of serious ovarian carcinoma with serial primary and metastatic pairs, and serial primary and recurrent pairs.

Researchers can take these various tumor clusters based on their transcriptome profiles and perform molecular annotation using tools to look at immune cluster profiling — such as CIBERSORT and xCELL — and measuring inflammation and immune scores, or study the tumor clusters with tools such as gene set enrichment analysis (GSEA) and IPA.

“We can take these data and look at immune effector cell profiling and gene expression of various genes that define these tumor inflammation scores at the single-gene level and interpolate across the specimens to look for hot and cold regions of a tumor tumor,” he said. “Tools like CIBERSORT can be used to look closely at the actual immune effector cell profiles within various microenvironment regions.”

PAGE BREAK

When looking at one specific ovarian sample, researchers were surprised to find the region defined as tumor histopathologically actually could be broken out into three distinct transcriptional profiles that could not be deduced by histopathological methodology, Carpten said.

“We actually saw different gene expression profiles, processes and pathways driving these three distinct regions of tumor that make up a really intriguing transcriptional substructure across the sample,” he said. “Importantly, we’re also measuring this across the depth of the tumor in these cases using serial sections.”

‘This is the future’

Data gleaned from this deep sequencing can help clinicians to better tailor and personalize treatments.

For instance, an analysis of a tumor might reveal one group of cells expressing genes that suggest growth due to cell cycles, but a different population of cells is growing due to hormone receptor activation.

“By knowing that information, a doctor can say, ‘OK, there is a population of cells that might respond to tamoxifen, but another population of cells that might not. I might need to use chemotherapy plus tamoxifen to battle this cancer,’” Carpten told HemOnc Today. “Whereas, another cancer might be more clonal, that might be driven by a different process, like hypoxia, and we can go in and attack that cancer very specifically.”

Clinicians can understand the disease better by understanding the different tumor clones.

“Doctors can better understand which patients are likely to fail, why they are likely to fail and, if they do fail, what might be an effective way to treat that cancer when it comes back,” Carpten said. “This is the future of cancer genome science.”

Overall, these technologies that are under development can help researchers to better understand the relationships of cells that define a cancer and its microenvironment, Carpten said during his presentation.

“Integrating these cutting-edge approaches going forward will undoubtedly shed new light on cancer and will open up new and innovative ideas on how to best approach disease prevention, diagnosis and treatment for all patients,” he said. – by Alexandra Todak

Reference:

Carpten JD. Wrap-up and opportunities for the future. Presented at: AACR Annual Meeting; March 29-April 3, 2019; Atlanta.

Disclosure: Carpten reports a collaboration agreement with 10X Genomics.

    See more from American Association for Cancer Research Annual Meeting