The numerous natural products and their bioactivity potentially afford an extraordinary resource for new drug discovery and have been employed in cancer treatment. However, the underlying pharmacological mechanisms of most natural anticancer compounds remain elusive, which has become one of the major obstacles in developing novel effective anticancer agents. In this case, we developed CancerHSP which could be used to address these unmet needs. Here, a case study of taxol is provided to illustrate how to use CancerHSP.

---

 

Part 1. Data collection

 

1)      Select “Chemical name” option in the search box and use “Taxol” as a keyword for searching;

Figure 1

 

2)      The browser will display the page of chemical entries which contains the highlighted “taxol” substring in the chemical names;

 

Figure 2

 

 

3)      CancerHSP provides a very useful gadget which is a feasible way to filter and sort data, thus user could easily find the full matched entry in this page;

Figure 3

 

4)      After clicking the “taxol” in “Name & synonyms” column, molecule information page of taxol will be visible, where molecule structure, pharmacological and molecular properties, anticancer activities, targets, related herbs, references, etc., were well organized and displayed. In this page, user could collect the ADME properties, anticancer activities based on cancer cell lines, targets, etc.

 

Figure 4

---

 

Part 2. Data analysis

 

1)      ADME analysis

 

By analyzing the ADME parameters for all compounds in CancerHSP, the basic distribution profiles show a very interesting phenomenon (Figure 5). For example, for the molecule weight (MW) profile of the compounds in CancerHSP, the median is 426.80, and 90% of them fall into 230.33 to 971.02 g/mol which is different from Lipinski's rule of five. This phenomenon indicates that anticancer drugs may have a specific MW profile.

Figure 5. The distribution profiles of ADME parameters for compounds in CancerHSP

 

 

Table 1. The minimum, 5% percentile, median, average, 95% percentile and maximum values of the parameters for compounds in CancerHSP.

 

By mapping the parameters of taxol to whole dataset distribution, we can roughly know which parameters need to be modified when developing a new anticancer drug.

 

Figure 6. The box plots of parameters for compounds in CancerHSP. The black triangles represent taxol. The red plus signs are outliers.

 

 

2)      Sensitivity analysis

 

The sensitivity data (anticancer activity based on cell lines) in CancerHSP can be analyzed after transforming by the following formula:

Where x_i is the i_th value of sensitivity data, ED₅₀ for example, which are unified by the units; X denotes one kind of sensitive data set, ED₅₀ in this case, for one compound; and median is the function of calculating the median value in X.

 

Figure 7. Sensitivity of cancer cell lines for taxol. In figure 7ABC, generally, the cell lines with negative S values are more sensitive than the positive ones. In contrast, the cell lines with positive S values are more sensitive than the negative ones in the figure 7D, due to they are Inhibition rates of cell lines under the treatment of 30μg/ml taxol.

 

3)      Target analysis

 

There are 37 targets in CancerHSP which were predicted by our previous methods ^1,2, 16 of them were proved by the literatures, BindingDB ³ or DrugBank database ⁴. However, there are also 11 targets need to be proved by the “wet” experiments. All these targets provide clues to look inside why the taxol shows different sensitivities to a range of cancer cell lines. In addition, the drug action mode provide more precise drug-target interaction detail which is a new opportunity to uncover the mechanisms when using network methods ⁵. By analyzing the ontology ^6,7 and pathway ^8,7 which these protein targets involved, the relationships between these protein to biological process, molecular function and cellular component will be emerging.

---

 

Part 3. Other usages

 

1)      The anticancer chemicals in CancerHSP can be used as the lead compounds for developing anticancer drugs.

2)      The compound structures can be download for further study by building statistical models.

3)      It is also valuable for searching anticancer compounds in traditional Chinese medicines.

 

---

 

References

 

1 Yu, H. et al. A systematic prediction of multiple drug-target interactions from chemical, genomic, and pharmacological data. PLoS One 7, e37608 (2012).

2 Zheng, C. et al. Large-scale direct targeting for drug repositioning and discovery. Submitted (2014).

3 http://www.bindingdb.org/bind/index.jsp, access time: January 16, 2014

4 http://www.drugbank.ca/, accessed on October 1st 2013

5 Zheng, C. et al. Systimg/s-pharmacology dissecting holistic medicine for treatment of complex diseases: an example using cardio-cerebrovascular diseases treated by TCM. Submitted (2014).

6 http://david.abcc.ncifcrf.gov/

7 http://amp.pharm.mssm.edu/Enrichr/

8 http://www.genome.jp/kegg/pathway.html