Hepatocellular carcinoma: towards personalized medicine.

Cancer Sci. 2012 Feb 16;

Authors: Miki D, Ochi H, Nelson Hayes C, Aikata H, Chayama K

Abstract
Over the past several years, the success of genome-wide association studies (GWAS) and pharmacogenomics have gradually begun to enable personalized medicine in some fields. In the field of liver diseases, host genetic factors are now very useful in clinical practice for predicting treatment outcome and adverse reactions for pegylated interferon plus ribavirin combination therapy against chronic hepatitis C virus (HCV) infection. Recently, three virus-related hepatocellular carcinoma (HCC) GWAS studies were reported from Asia. One study examined hepatitis B virus (HBV)-related HCC in China, where HBV is very prevalent, while the other two examined HCV-related HCC in Japan. We identified a common variant in the DEPDC5 locus associated with HCV-related HCC, and another group identified an association involving the MICA locus. In this review, we compare the results of these GWAS studies and earlier candidate gene studies. Further research is needed to determine the role of these single nucleotide polymorphisms on HCC risk, but identification of these markers may make it possible to assess the magnitude of the risk of cancer based on each patient's genetic background. Consideration of the genetic background of the patients will likely play a role in personalized medicine for HCC, and understanding the mechanism underlying the association may suggest novel promising therapeutic targets in the future. © 2012 Japanese Cancer Association.

PMID: 22339805 [PubMed - as supplied by publisher]

Personalized medicine: a personal view.

Clin Pharmacol Ther. 2012 Mar;91(3):373-5

Authors: Meyer UA

Abstract
Personalized medicine is a strategy to prevent, diagnose, and treat disease so as to achieve an optimal result for the individual. The sequencing of the human genome and other technological advances have revealed the extent of genetic diversity and the relative contribution of genetic and nongenetic factors to human health, disease, and drug response. The challenge is to translate this knowledge into tangible benefits for the patient.

PMID: 22343810 [PubMed - in process]

Industry panel discusses status of personalized medicine.

Am J Health Syst Pharm. 2012 Mar 1;69(5):366

Authors: Thompson CA

PMID: 22345411 [PubMed - in process]

Recent successes of cancer immunotherapy: a new dimension in personalized medicine?

Target Oncol. 2012 Feb 15;

Authors: Caux C, Zitvogel L

PMID: 22350488 [PubMed - as supplied by publisher]

Egr-ly awaiting a "personalized medicine" approach to treat scleroderma.

J Cell Commun Signal. 2012 Feb 15;

Authors: Leask A

Abstract
Scleroderma, an autoimmune disorder characterized by skin and organ fibrosis, has no treatment. Although over the past decade valuable insights into the molecular mechanisms underlying scleroderma have been generated, results in clinical trials have been disappointing. This issue is likely to arise due to the heterogeneity of scleroderma. Molecular insights into the heterogeneity of this disease have been provided by genome-wide expression profiling. In a recent paper, Bhattacharyya and colleagues (PLOS One 6:e23082, 2011b) to show that the overexpression of a range of "fibroproliferative" genes in diffuse cutaneous scleroderma patients are likely to be caused by the overexpression of transcription factor Early growth response (Egr)-1. Only a minority of Egr-1-regulated genes were also found to be regulated by TGF-ß. Moreover, Greenblatt and colleagues (Am J Pathol., 2012) have shown that the overexpression of "inflammatory" genes overexpressed in "localized" scleroderma and a small subset of limited and diffuse scleroderma patients is likely to be due to the activity of interleukin-13 (IL-13). Intriguingly, at a gene expression level, murine sclerodermatous graft-versus-host disease (sclGVHD) approximates this inflammatory subset of scleroderma. These data suggest that targeting Egr-1 expression/activity might be a novel therapeutic strategy to control fibrosis in a subset of diffuse scleroderma patients, and further emphasize that notion that elevated canonical TGFβ signaling is insufficient to explain the fibrosis observed in scleroderma. Moreover, targeting IL-13 expression/activity might be a novel therapeutic strategy to target the inflammation leading to "localized" scleroderma.

PMID: 22350706 [PubMed - as supplied by publisher]

Data from the 2010 census showed that about 13% of people living in the United States self-identify as African American, but from a genetic point of view, ethnicity isn’t so black and white. Most African Americans have genetic ancestry tracing back to both Africa and Europe but the exact proportion can vary widely. And conversely, many Americans who consider themselves of completely European descent may actually have some African ancestry as well.

A study published a few years ago found that the average amount of DNA in an African American’s genome that could be traced back to West Africa was about 77%, but ranged from as little as one percent to as much as 99%.

At 23andMe, researchers have looked at the genetic ancestry of about 78,000 customers likely to consider themselves as entirely of European ancestry and found that somewhere between three to four percent of those people have “hidden” African ancestry ranging from 0.5 to 0.75 percent. These findings suggest that this group of customers have an African ancestor who lived about six generations, or about 200 years, ago.

The United States is often referred to as a “melting pot” and although African Americans comprise the largest racial minority, this country is also home to many other genetically mixed groups. Indeed, Americans with ties to Latin America typically have DNA deriving from Native American, European and sometimes African populations. By studying genetic ancestry, researchers can break down the proportion of a person’s DNA that traces to different parts of the world and learn more about our unique and interconnected histories.

February is Black History Month — stay tuned for more posts!

Did you know? provides tidbits of information about genetics (in humans and in other animals) and explains how DNA relates to both ancestry and health.

By Emily Drabant

As the Research Manager for the 23andMe Parkinson’s Community, I spend my days thinking about ways to advance this research faster, cheaper and more effectively than ever before. We launched this project in 2009 and since that time have been going full steam to make this the largest and most scientifically productive Parkinson’s research cohort in the world. We’ve succeed on the first goal and already have the world’s largest group of people with Parkinson’s participating in genetics research. And we are making strides towards the second goal with significant scientific discoveries last year.

While our progress is incredible, we have not yet reach our primary milestone of enrolling 10,000 people with Parkinson’s. Our team of scientists believes that having a group of that size will put us in a new era of Parkinson’s research. This will dramatically increase the odds of really understanding the genetics behind the disease. Scientists call this the “statistical power” of research. We call it the power of our research revolution.

As enrollment slowed around 6,000 people, I began to brainstorm on ways to get more people with Parkinson’s involved. We’ve worked directly with many patient advocacy groups to spread the word — especially with the Michael J. Fox Foundation, the National Parkinson Foundation, and The Parkinson’s Institute. But I wondered: how could we reach people who aren’t linked up with an advocacy group or a support group, and perhaps aren’t even very public about the fact that they have Parkinson’s? I began to think that if we did something really big — if we made a splash in a really public way — that it would truly raise awareness about this research and give people who want to participate the chance to do so. I hear from people with Parkinson’s all the time saying ‘I had no idea this was available. I didn’t know I could advance research from home”.

In considering ways to make this research public in a really big way, we naturally started to think about working with a celebrity with Parkinson’s. At the top of that list is Muhammad Ali, one of the world’s most recognized athletes, who has been fighting Parkinson’s for 25 years. We spoke to one of our collaborators, Dr. Bill Langston, and he suggested we contact Muhammad’s wife, Lonnie. “When Lonnie says something is going to happen, she will make it happen,” he said.

Those words were completely true. In December, my colleague, Melissa Del Sol, and I flew to Arizona to film Muhammad and interview Lonnie at their home. It was a magical experience and you can read more in Melissa’s blog.

I didn’t plan to conduct the interview — the director of the video was going to handle that — but at the last minute he decided it made more sense for me to do it. I did it off the cuff and asked questions that I wanted to know as a scientist, as someone with personal connections to Parkinson’s, and as an admirer of Muhammad Ali.

We talked about Muhammad’s diagnosis of Parkinson’s. When he first started showing symptoms they were quite minor — a tremor in his finger. The doctor said it was Parkinson’s syndrome, which meant that it would never progress into full blown Parkinson’s, but rather was just a Parkinson’s-like symptom. But over time his symptoms did progress. When they went to a Parkinson’s specialist he confirmed that it was in fact Parkinson’s disease. I asked Lonnie how they both reacted to this.

“Muhammad deals with adversity a lot different than most people. To him, I don’t think it really registered, because he was still doing a lot of what he wanted to do,” she said. “He’s always had that thing about mind over matter … where he thought he could overcome this on his own. It took a real long time for him to realize that it wasn’t going to work that way.”

“However for me it was a little bit different because when you say disease it is a lot different than hearing syndrome. Any “disease” sounds very fatalistic. So I had to do a bit of investigating.”

As Lonnie learned more, she and Muhammad started to get involved in research. When he realized that he could advocate for others with Parkinson’s, he decided to go public with his diagnosis. Muhammad and Lonnie founded the Muhammad Ali Parkinson’s Center in Arizona in 1997 and have made tremendous contributions to research.

“Muhammad and I have sort of grown up with the research,” Lonnie said. “A lot of research has been done to advance our understanding of what this disease is and where it comes from and some of the factors that may contribute to it. So we are very hopeful about the research. But when we first started out it was funny because we were told that in ten years we would have a cure and it’s just a function of money. Well now we know that’s not true. Ten years have come and gone and there has been a lot of money, but it’s a lot more complex than that. … We are really hopeful that this is one of the neurological diseases that will be given a cure.”

This point is not lost on us at 23andMe. Indeed, the slow pace of research into Parkinson’s and many other diseases was what motivated us to launch our online research platform for all in the first place. We wanted to speed up research and do it bigger than ever before. We knew we needed to change the research model if we were going to make progress in the same lifetime as those who are currently battling Parkinson’s.

“I think 23andMe is cutting edge technology that has been introduced to address this illness and to look at it in a different lens; getting to the genetic basis of what causes this illness, why do some people get it and why is it that some people don’t. The sooner we understand that,” said Lonnie, “the sooner we will be able to address treatment — better treatment — options for people.”

She continued, “Looking at the genetics of the illness and the way 23andMe is doing it – getting a broad base of people to participate in the research – is extremely important to finding a cure. Then all the added things 23andMe does with regards to the person – giving them information about their personal health history and how to keep them healthy aside from Parkinson’s, because what people forget is that people with Parkinson’s don’t just have Parkinson’s, they have a lot of other illnesses. They may have hypertension, they may have diabetes. They could have a lot of other things which agitate the Parkinson’s that they have to live with. So getting all that information back about works best for you on an individual basis … is very important. It’s a wonderful thing that 23andMe has done, trying to partner with other researchers around the country to try to answer some of these complex questions.”

I left Arizona feeling inspired, full of gratitude and hope for the future of Parkinson’s research. When I was a neuroscientist at Stanford doing brain imaging research I never could have imagined that one day I would be working with Muhammad Ali on a video project. But there is no doubt in my mind that this video, and Muhammad and Lonnie’s involvement with 23andMe, will push 23andMe’s Parkinson’s research to new heights. Parkinson’s disease can be devastating for those with the illness, but sometimes we forget that such a daunting diagnosis can also inspire great courage.

In this job working with the Parkinson’s community, I am continually awed by the reservoir of spirit found among those facing the disease. Just as Muhammad Ali inspired and continues to inspire generations of people with his courage and spirit, those with Parkinson’s who are bound together to fight this disease can inspire us all.

Learn more and see the video at www.23andme.com/pd

Personalized medicine: Bring clinical standards to human-genetics research.

Nature. 2012 Feb 16;482(7385):300-1

Authors: Lyon GJ

PMID: 22337032 [PubMed - in process]

Targeted Therapy in the Management of Advanced Gastric Cancer: Are We Making Progress in the Era of Personalized Medicine?

Oncologist. 2012 Feb 14;

Authors: Wong H, Yau T

Abstract
AbstractBackground. Gastric cancer is one of the leading causes of cancer death. With greater understanding of the molecular basis of carcinogenesis, targeted agents have led to a modest improvement in the outcome of advanced gastric cancer (AGC) patients.Methods and Results. We conducted an overview of the published evidence regarding the use of targeted therapy in AGC patients. Thus far, the human epidermal growth factor receptor (HER) pathway, angiogenic pathway, and phosphatidylinositol-3-kinase (PI3K)-Akt-mammalian target of rapamycin pathway have emerged as potential avenues for targeted therapy in AGC patients. The promising efficacy results of the Trastuzumab for Gastric Cancer trial led to the approved use of trastuzumab-based therapy as first-line treatment for patients with HER-2(+) AGC. On the other hand, the Avastin® in Gastric Cancer trial evaluating bevacizumab in combination with chemotherapy did not meet its primary endpoint of a longer overall survival duration despite a significantly higher response rate and longer progression-free survival time in patients in the bevacizumab arm. Phase III data are awaited for other targeted agents, including cetuximab, panitumumab, lapatinib, and everolimus.Conclusion. Recent progress in targeted therapy development for AGC has been modest. Further improvement in the outcome of AGC patients will depend on the identification of biomarkers in different patient populations to facilitate the understanding of gastric carcinogenesis, combining different targeted agents with chemotherapy, and unraveling new molecular targets for therapeutic intervention.

PMID: 22334453 [PubMed - as supplied by publisher]

By David Hinds, Chuong Do, and Shirley Wu

In January we wrote about the challenging search for genetic influences on human longevity, touching on two of the most recent studies as examples of how elusive solid findings have been. Because one of these studies was a new version of a paper that was previously retracted, we took a particular interest to see if the concerns raised by the previous analysis were addressed. What follows is a technical review of “Genetic Signatures of Extreme Longevity in Humans”, published last month in PLoS ONE.

Although Paola Sebastiani and Thomas Perls (the main authors on both papers) do not directly discuss their previous study, their new version does appear to resolve many of the issues from the prior manuscript, and their GWAS results are substantially improved. They now report just one SNP clearly associated with longevity, rs2075650 near APOE. They also present a revised genetic model for predicting longevity based on 281 SNPs representing the common genetic variants most significantly associated with longevity in an analysis of 801 centenarians and 914 healthy control individuals.

As described in the paper, the performance of the model remains quite impressive:

• 89% sensitivity and 89% specificity (“area under the curve”=AUC=0.95) on the discovery set,
• 60% sensitivity and 58% specificity (AUC=0.58) on a replication set of 253 independent cases and 341 genetically-matched controls, and
• 78% sensitivity and 61% specificity (AUC=0.74) on a replication set of 60 independent cases and 2,863 unmatched controls.

We wanted to understand the meaning of these results and to determine whether this model could be used to predict longevity for 23andMe customers. To this end, we examined a few specific aspects of the study in closer detail.

1. Performance on discovery data.

Our first area of interest was in understanding the authors’ proposed methodology for constructing the predictive models. In the paper, the authors describe a procedure for constructing an ensemble of nested predictive models, using SNPs that were identified as the top associations in a GWAS. To choose between different models, the authors described a procedure that involved repeatedly splitting the discovery set into training and testing folds; however, in each case, the set of SNPs that were candidates for inclusion were taken from the top hits in the GWAS of the entire discovery set. We were concerned that constructing models in this way could lead to extremely biased estimates of accuracy since the candidate SNP selection process uses information from the testing folds.

To evaluate this, we simulated genome-wide association studies of 801 cases and 914 controls across 200,000 independent SNPs, one representing the true association with rs2075650, and the other 199,999 not associated with longevity (reduced from the 243,980 SNPs in the paper to account for potential non-independence in the original set). In each simulation, we selected the 281 most strongly associated SNPs and built nested prediction models as described in the longevity study. Across 1,000 simulated datasets, when we evaluated the resulting models by 10-fold cross validation, we observed an average sensitivity of 88%, an average specificity of 89%, and an average AUC of 0.95, very close to the reported values. Thus, the reported performance in the discovery data seems consistent with the null hypothesis that rs2075650 is the only informative SNP in the model, and the performance difference between a 281-SNP model and a 1-SNP model could be entirely explained by overfitting.

Performance of risk model ensembles from Sebastiani et al. 2012, trained on simulated random data containing only one true association. The left plot shows 10-fold cross-validation performance when SNPs are pre-selected using the entire discovery sample. The right plot shows 10-fold cross-validation performance when no SNPs are pre-selected. Shaded regions denote 95% confidence bands. Note the similarity of the sensitivity and specificity curves in the left plot with Figures 4B and S4A from the paper, which the authors use to justify including 281 SNPs in their model. In contrast, in the right plot, model accuracy (in terms of AUC) decreases with increasing model size.

Although the authors do acknowledge the possibility of overfitting, we suspect that some readers may be surprised at the extent to which “pre-selecting” candidate SNPs for a model (based on the results of a genome-wide scan where the validation set has not been excluded) can lead to highly inflated estimates of accuracy, even when the underlying data is essentially random. In particular, these results cast doubt on the validity of the resampling experiments used by the authors to justify their choice of model size since the predictive accuracies estimated by the authors’ bootstrapping procedure are just as vulnerable to overfitting the discovery data as the cross-validation protocol in our simulation above.

Had SNP selection been performed in the “inner loop” of the validation (i.e., only using the training portion of each train/test split), then the estimates of performance for different models would not have suffered from the bias shown here and substantial overfitting could have been avoided. In the simulations above, if one uses a correct cross-validation procedure, model accuracy decreases with model size, consistent with the fact that there is just one true association in the data.

2. Population stratification in 23andMe data.

Of course, even a model that is overfit may have some amount of predictive power as long as the “noisy” predictions from the model show some meaningful correlation with true phenotypes. Our second step was to implement the 281-SNP model described in Table S1 of the paper and test it in a collection of over 80,000 23andMe research participants with predominantly European ancestry.

In preparation for assessing the performance of the model, we examined whether there was any evidence that the risks predicted by the model varied with ancestry in the 23andMe data. We used principal components analysis to extract the two most important dimensions of ancestry in these individuals, and plotted the proportion of individuals with predicted probability of exceptional longevity (according to the model) > 0.5 for 23andMe participants on these dimensions. Because living to age 100 is so rare, this plot effectively shows the false positive rate of the classifier, or “1-specificity”, in 23andMe participants, assuming a prior probability of exceptional longevity of 0.5 to be consistent with results in the paper.

Proportion of individuals in 23andMe database with predicted longevity probability > 0.5 based on Sebastiani’s model. (LB=Lebanon, IR=Iran, IT=Italy, GR=Greece, ES=Spain, FR=France, BE=Belgium, UK=United Kingdom, RO=Romania, HU=Hungary, AT=Austria, DE=Denmark, NO=Norway, SE=Sweden, CZ=Czech Republic, UA=Ukraine, PL=Poland, FI=Finland, LT=Lithuania).

In the figure to the right, we show results for areas of the plot with at least 50 participants. The labels indicate country of origin for participants with four grandparents from the same country, or ‘AJ’ for self-identified Ashkenazi Jews. We see a trend in risk scores, with lower probabilities of longevity assigned to Ashkenazi Jews, Southern, and Eastern Europeans compared to Western and Northern Europeans. Specificity of the model varies from >70% for Ashkenazi Jews to <50% for Scandinavians.

The plot provides evidence that the overall risk scores predicted by the model correlate with ancestry in the 23andMe population. Although this observation might seem surprising, given that the risk model was trained on cases and controls that were matched for genetic ancestry, there is no contradiction here — it is plausible that the genetic component of exceptional longevity might vary systematically with ancestry. Furthermore, it is certainly true that individual alleles in the risk model are in some cases correlated with ancestry; for instance, rs2075650 near APOE shows a gradient in minor allele frequency from south to north across Europe.

The fact that the risk score from the model stratifies by ancestry suggests that the latter should be taken into account as a potential confounding factor when measuring performance, especially when comparing performance in datasets with differing composition by ancestry. This result might complicate interpretation of the authors’ second replication experiment, because that experiment used cases and controls that were not matched for ancestry. (In the paper, the authors did look for such an effect, but Fig. S10 does not provide strong evidence for the absence of residual population stratification.)

3. Performance on 23andMe data.

From the 23andMe research database we selected a cohort of unrelated individuals of primarily European ancestry, including 31,547 participants with current age < 50, and 2,506 participants with age >= 80. Using logistic regression, we tested whether there was an association between longevity (age >= 80) and the predicted longevity score (converted to log-odds), adjusting for gender. Interestingly, despite the model overfitting described earlier, we found a weak but significant association between the two (P=0.021, odds ratio (OR) = 1.03).

However, we saw a stronger association between longevity and the single APOE SNP rs2075650 (P=2.8e-5, OR=0.83); this association was the only statistically genome-wide significant association in the authors’ analysis as well. We also saw a strong association between longevity and the first five principal components of ancestry (P= 100), the association with rs2075650 was no longer significant but had a larger effect size in the expected direction (P=0.075, OR=0.44). The longevity risk score was still not associated with age (P=0.97) after controlling for ancestry and the APOE SNP, a result which is not particularly sensitive to the age cutoff used (P=0.30 using 479 individuals with age >= 90, or P=0.28 using 141 individuals with age >= 95).

Median age of 23andMe research participants by self-identified (AJ) or 4-grandparent-determined ancestry.

There are multiple reasons why the risk score might not replicate in the 23andMe cohort. For example, our younger longevity phenotypes are less extreme than the phenotypes in the PLoS ONE study, and our small group of centenarians has lower power for detecting small effects on risk. In addition, our self-reported age data may be less reliable for extreme ages, though to reduce the risk of errors from incorrect age reporting, the above analyses were conservatively restricted to individuals who provided their date-of-birth in at least two separate forms or surveys on the 23andMe website and excluded those who reported conflicting birth years in any of these locations.

Interestingly, in our cohort, Ashkenazi Jewish ancestry is positively associated with longevity (for age >= 80: P=3.4e-4, OR=2.2), while in the PLoS ONE study, Ashkenazi Jewish ancestry seems to be negatively associated with longevity. This almost certainly reflects a difference in how individuals were selected for the two studies, rather than a difference in genetic predispositions. If we exclude Ashkenazi Jews from our regression analysis, the longevity score is slightly more predictive on its own (P=0.019, OR=1.03), but the association again goes away once we add the top five principal components and APOE to the model (P=0.22, OR=1.02).

Concluding thoughts.

The genetic model presented by Sebastiani and Perls appears to be overfit to their training data, and we further see very little correlation between the predictive scores generated from their model and longevity in the 23andMe research cohort, once the effects of ancestry and the significant APOE association are taken into account. Even without these additional corrections, the ability of the authors’ risk score to distinguish long-lived individuals in our database is poor.

It is worth noting that a model based on only the single APOE association in the paper achieves an AUC of 0.58 (0.52-0.63) among individuals with age >= 100 in the 23andMe cohort, consistent with the authors’ estimate of 0.62; the effect weakens with decreasing age cutoff (e.g., AUC=0.53 (0.49-0.56) for age >= 95). When combined with the 280 other SNPs in the authors’ model, however, the performance in our cohort drops to AUC=0.49 (0.39-0.60) among centenarians (or AUC=0.52 (0.47-0.56) for age >= 95), which is statistically indistinguishable from random guessing. While the ability of our cohort to detect meaningful association may be limited due to sample size, our confidence intervals are at least sufficiently narrow to exclude the point estimate of AUC=0.74 reported by the authors for their second replication study.

We remain concerned that despite clear improvements in the authors’ revised study, the analysis may nonetheless be susceptible to subtle biases, due to the way in which cases and controls were selected. Because of the numerous potential confounding factors related to the separate ascertainment of cases and controls and the fact that most of the controls were derived from a single source, it may be impossible to formally show that all sources of bias have been adequately controlled in the authors’ analyses.

An alternative analysis that at least partially addresses some of the concerns above (without genotyping additional samples) would be to repeat the discovery analysis using a random subset of NECS cases and Illumina controls (using a properly cross-validated model selection procedure), and reserving the remainder of the NECS cases and controls for replication. This would not resolve the potential for confounding in the model building step, but would at least reduce the potential for bias in the replication set. A more convincing demonstration would require constructing and evaluating the model using a dataset where such confounding factors are simply absent by design (i.e., where cases and controls are drawn from a single underlying population and genotyped together); this would certainly involve genotyping of more controls than in the study, but unlike centenarians, controls are relatively easy to come by (we have lots!).

In our view a predictive genetic model for longevity remains elusive but we are optimistic that continuing efforts like the New England Centenarian and Supercentenarian Studies will keep adding to our knowledge of this most human of traits.