|Year : 2020 | Volume
| Issue : 7 | Page : 1656-1663
The prevalence of human papillomavirus in colorectal cancer and adenoma: A meta-analysis
Guanqun Chao1, Xiaojie Hong2, Xinli Chen2, Shuo Zhang2
1 Department of Family Medicine, Sir Run Run Shaw Hospital, Zhejiang University, Hangzhou, China
2 Department of Gastroenterology, The First Affiliated Hospital, Zhejiang Chinese Medical University, Hangzhou, China
|Date of Submission||15-May-2020|
|Date of Decision||13-Oct-2020|
|Date of Acceptance||30-Nov-2020|
|Date of Web Publication||9-Feb-2021|
Department of Gastroenterology, The First Affiliated Hospital, Zhejiang Chinese Medical University, Hangzhou
Source of Support: None, Conflict of Interest: None
Objective: The objective of this study is to further clarify the correlation between colorectal cancer (CRC) and human papillomavirus (HPV) through literature search and meta-analysis, which is conducive to the formulation of further prevention programs.
Methods: Searching Web of Science, PubMed, MEDLINE, Scopus, and CENTRAL for studies investigating the relationship between CRC and HPV. All analyses were performed through Revman (version 5.3, the Cochrane Collaboration, Oxford, UK). Data from selected studies were extracted into two by two tables. Moreover, all included studies were weighted and summarized.
Results: Eighteen studies were included. The expression of HPV in CRC tissues was obviously higher than that in nonmalignant tissues (odds ratio [OR] = 5.56, 95% confidence interval [CI] = 3.18–9.72, Z = 6.02, P < 0.00001). The expression of HPV in CRC tissues and adenoma tissues showed no significant abnormalities (OR = 1.74, 95% CI = 0.92–3.29, Z = 1.70, P = 0.09). The expression of HPV in CRC tissues was obviously higher than that in normal tissues (OR = 7.23, 95% CI = 3.89–13.42, Z = 6.27, P < 0.00001).
Conclusion: HPV infection in CRC tissues was obviously higher than that in nonmalignant tumor tissues and normal colon tissues, but there were no statistically significant abnormalities between CRC tissues and adenoma tissues.
Keywords: Adenoma, colorectal cancer, human papillomavirus, meta-analysis
|How to cite this article:|
Chao G, Hong X, Chen X, Zhang S. The prevalence of human papillomavirus in colorectal cancer and adenoma: A meta-analysis. J Can Res Ther 2020;16:1656-63
| > Introduction|| |
Colorectal cancer (CRC) is the third-most common malignancy in the world and the fourth leading cause of death in developing countries. In western countries, CRC is the second most common malignancy followed by skin cancer. Although the current level of medical care has significantly improved, the 5-year survival rate of CRC is 67%, compared with 64% for colon cancer. Even among survivors, CRC has a high recurrence rate and is prone to secondary tumors, especially in the digestive tract. The high mortality rate of CRC is ascribed to the late detection of CRC and the high recurrence rate to a great extent. The 5-year survival rate of CRC depends on the current stage of CRC. Although colonoscopy is widely used to detect early tumors, more studies were carried out to prevent CRC. Exploring the mechanism of CRC is a research hotspot recently. The researchers supposed that gene mutations in KRAS, NRAS, and BRAF are closely related to the occurrence of CRC. In addition to genetic factors, factors such as behavior and diet are also associated with the occurrence of CRC.
In recent years, due to increased concern about human papillomavirus (HPV), researchers have begun to pay attention to the correlation between HPV and CRC. HPV is a DNA tumor virus which is the main cause of sexually transmitted infections. HPV is considered a risk factor of many malignancies, such as cervical cancer, head and neck cancer. Through meta-analysis, the researchers identified a close relationship between HPV infection and esophageal squamous cell carcinoma. Studies carried out on HPV-related prostate cancer found that HPV colonization can also occur in benign tumors prior to the onset of prostate cancer. A new study suggests that HPV may be link to CRC, but appears to be influenced by location, diet, and lifestyle. Some researchers have also found that HPV16 was found in more than 50% of HPV positive colon cancer tissues, thus suggesting that the occurrence of local CRC is related to the high rate of HPV infection. However, a large case–control study shown that HPV was detected in both CRC tissues and control normal tissues, denying the relevance between CRC and HPV infection. Therefore, the correlation between CRC and HPV infection is still controversial. The target of this research is to further clarify the correlation between CRC and HPV through literature search and meta-analysis, which is conducive to the formulation of further prevention programs.
| > Methods|| |
For this meta-analysis, we searched Web of Science PubMed, MEDLINE, Scopus and CENTRAL with the goal of identifying the relevance of HPV and CRC. Data were collected from 1966 to 2020 (as of January). The terms we used in the search included two aspects. One included “colorectal cancer,” “colorectal tumor,” “rectal cancer,” “rectal tumor,” “colon cancer,” “colon tumor” and “colorectal carcinomas.” The other included “Human papillomavirus” and “HPV.” The search had a language restriction to English. We searched full-text articles through various websites, and collected dates as original as possible. All documents we reviewed were in duplicate, independently reviewed by the investigator.
Data were extracted by three reviewers from studies that met the inclusion criteria. All controlled studies investigating the relationship between HPV and CRC were included. Our inclusion criteria were: (1) random control; (2) tissues from patients who had a CRC diagnosis were regarded as the cases; (3) normal tissues and adenoma tissues were regarded as the controls; (4) HPV DNA was detected in the tissue. Our exclusion criteria were: (1) The researchers came from people with specific diseases, such as HIV; (2) the number of detected cases is not summarized in the literature; (3) animal experiments; and (4) overview or systematic analysis. Each article was a study reviewed by reviewers individually and ultimately selected. Research process, research method, research results, loss of follow-up, and adverse reactions must be included in these articles.
From four aspects (patient selection, index detection, reference criteria, and flow and time), quality appraisal was done using QUADAS-2 tools, for this study.
All analyses were performed through Revman (version 5.3, The Cochrane Collaboration, Oxford, UK). Data from selected studies were extracted into two by two tables. All included studies were weighted and summarized. We used a meta-analysis program (RevMan 5.3, The Cochrane Collaboration, Oxford, UK) to calculate the relative risk odds ratio (OR) and 95% confidence interval (CI) between cases and controls, and we performed a meta-analysis of the effect value (weighted mean difference). The l abbe plot was used to calculate the event rate of the experimental group compared to the control group to explore the heterogeneity of effect estimation. We used fixed-effects models to pool data for it had homogeneity; in addition, we used a random-effects model as well.
| > Results|| |
We identified 398 studies, of which 266 were aimed on other subjects, 35 were non-English articles, 17 were reviewers, 27 were experiments, 3 were case reports, 12 with special diseases, 1 was letter, 6 were repeated articles, 13 clinical studies did not meet the requirements of meta-analysis. Ultimately, 18 articles met the inclusion criteria and were included in the meta-analysis. [Figure 1] shows the flow chart of research selection.
All of 18 articles were included as qualified researches, which were published from 1993 to 2017. All of these articles were published in English. [Table 1] shows the characteristics of the included studies.
Publication bias assessment and quality assessment
From four aspects (patient selection, index detection, reference criteria, and flow and time), the quality appraisal of 18 studies was done using QUADAS-2 tools. [Figure 2]a shows the detailed appraisal results.
|Figure 2: Methodological quality diagram and publication bias assessment (a) Quality assessment based on the questions of QUADAS-2 quality assessment. Red, yellow, and green colors show the high, unclear, and low risk of bias and applicability concerns. (b) Red, yellow, and green colors show the high, unclear, and low risk of bias|
Click here to view
To investigate the underlying publication bias, we did the analysis by Revman 5.3. We found no publication bias [Figure 2]b. The reason why we analyzed publication bias is that in the process of publication, the authors tend to publish articles with positive results, while those with negative results tend to be omitted.
Colorectal cancer tissue versus nonmalignant tissue
The heterogeneity test was performed by the Chi-square test. χ2 = 40.23, P = 0.0002 [Figure 3]a. The results shown that heterogeneity was discovered in our meta-analysis. Thus, a random-effect model could be analyzed in the meta-analysis.
|Figure 3: The forest plot of meta-analysis. (a) The forest plot of meta-analysis of human papillomavirus between colorectal cancer tissue and nonmalignant tissue (b) The forest plot of meta-analysis of human papillomavirus between colorectal cancer tissue and adenoma tissue (c) The forest plot of meta-analysis of human papillomavirus between colorectal cancer tissue and normal tissue|
Click here to view
Colorectal cancer tissue versus adenoma tissue
The heterogeneity test was performed by the Chi-square test. χ2 = 2.59, P = 0.27 [Figure 3]b. The results shown that this study did not exist heterogeneity in our meta-analysis. Thus, a fixed-effect model could be analyzed in the meta-analysis.
Colorectal cancer tissue versus normal tissue
The heterogeneity test was performed by the Chi-square test. χ2 = 33.56, P = 0.0008 [Figure 3]c. The results shown that heterogeneity was discovered in our meta-analysis. Thus, a random-effect model could be analyzed in the meta-analysis.
Merging and meta-analysis
Colorectal cancer tissue versus nonmalignant tissue
Based on the meta-analysis and the forest map, we found that the rhombus was on the right side of the vertical line and did not intersect with it. Based on [Figure 3]a, we determined the following values: OR = 5.56, 95%CI = 3.18–9.72, Z = 6.02, P < 0.00001. Therefore, the OR value was statistically significant. It indicated that the expression of HPV in CRC tissues was significantly higher compared with nonmalignant tissues. However, the number of included articles was limited, more studies should be conducted.
Colorectal cancer tissue versus adenoma tissue
Based on the meta-analysis and the forest map, we found that the rhombus was crossed with the vertical line. Based on [Figure 3]b, we determined the following values: OR = 1.74, 95% CI = 0.92–3.29, Z = 1.70, P = 0.09. It indicated that the expression of HPV in CRC tissues and adenoma tissues showed no significant abnormalities. However, the number of included literature was limited, and more studies are needed.
Colorectal cancer tissue versus normal tissue
Based on the meta-analysis and the forest map, we found that the rhombus was on the right side of the vertical line and did not intersect with it. Based on [Figure 3]c, we determined the following values: OR = 7.23, 95% CI = 3.89–13.42, Z = 6.27, P < 0.00001. Therefore, the OR value was statistically significant. It indicated that the expression of HPV in CRC tissues was obviously higher than that in normal tissues. However, the number of included articles was limited, more studies should be conducted.
| > Discussion|| |
It is estimated that nearly 41% of CRC cases occur in the proximal colon, 22% in the distal colon, and 28% in the rectum. Currently, the incidence of CRC is generally in decline thanks to the improving medical care ability, but patients aged <50 years who were diagnosed with CRC has increased by 2%. By 2030, the rate of colonic and rectal cancer patients aged 20–34 years is projected to increase to 90.0% and 124.2%, respectively. Studies have shown that genes are related to the development of CRC. Compared with the general population, the incidence of CRC in people whose first-degree relatives have CRC is significantly doubled. Meta-analysis found that heavy drinking was positively correlated with CRC mortality. Obesity is related to the development of CRC, especially in women. Meanwhile, Vitamin D is thought to be associated with CRC, the higher Vitamin D content is, the lower incidence of CRC is. A recent study suggests that male smokers have higher probability to develop colon cancer on the left side, whereas female smokers have higher probability to develop colon cancer on the right side, and that female smokers have a higher risk of colon cancer than male smokers. Thus, genes, gender, and lifestyle may be relevant to the occurrence of CRC, but it is difficult to change.
Recent research on CRC has focused on early screening and prevention. Because CRC is associated with genes, multi-target stool DNA testing has also been studied for early risk detection. At the same time, a comprehensive risk model combining family history, personal characteristics, and environmental factors were also studied and developed for risk stratification and risk assessment of CRC. Previous studies have emphasized the importance of endoscopic examination in the early detection of CRC, but institutions with limited economic power need to bear a certain economic burden. Although, tens of studies have been done in the treatment aspect of CRC, and a number of methods have been found to predict risk early. However, CRC has a high mortality rate, more researchers are looking for related mechanisms to prevent CRC. In recent years, for CRC researchers, HPV has become the hotspot.
At present, studies on the mechanism of CRC have reached the molecular level, but its etiology remains unclear. A study has found that in Puerto Rican Hispanics, HPV has the characteristics of high popularity in CRC. Meanwhile, another study suggested that HPV infection is relevant to CRC and may be an important prognostic factor. HPV infection had been found a risk factor of CRC in studies of Chinese populations. Therefore, multiple researchers have found HPV infection is related to the development of CRC, which may be an important risk factor for CRC, but there are also controversial, and some studies hold the opposite results. Taherian et al. concluded that there exist no significant association between HPV infection and CRC occurrence, but the authors considered that it might be related to the lack of study samples. Ranjbar et al. also suggested that there was no causal relationship between HPV infection and CRC, but suggested that high-risk types of HPV may increase the aggressiveness of CRC.
Our meta-analysis showed that HPV infection does increase the risk of CRC: HPV infection in CRC tissues was obviously higher compared with nonmalignant tumor tissues and normal colon tissues, but no statistically significant difference exist between CRC tissues and adenoma tissues. In this study, 8 articles were included in the meta-analysis, thus the sample size was large, and the results were reliable. However, there are also some limitations: (1) Because the study believes that genes and lifestyle are related to the occurrence of CRC, this study did not conduct further stratification study on genes or lifestyle; (2) most of the research results are published with positive results, so some negative results may be missed; (3) in the included articles, HPV neither detected in the experimental group or the control group, so such literature was not selected in our meta-analysis. However, in accordance with our analysis, CRC and colorectal adenoma have a high probability of HPV infection, whereas the mechanism of CRC caused by HPV infection needs further research and follow-up with more cases. On the other hand, based on its correlation, the development of HPV vaccine may be effective in preventing CRC and also has important significance for social health care.
| > Conclusion|| |
Our findings support the hypothesis that HPV infection and CRC/adenoma share some disease links, which may indicate that further study of the relevant mechanism of HPV and the development of HPV-related vaccine may have great significance for the prevention of CRC.
Financial support and sponsorship
This research was supported by National National Natural Science Foundation of China (82074214, 81973598, 81573760); funding from Zhejiang Provincial Natural Science Foundation of China under Grant No.LY18H030001; the Medicine and Health Science and Technology Plan Projects in Zhejiang province (2017KY413), Traditional Chinese Medicine Science and Technology Plan of Zhejiang Province (2017ZA089, 2016ZB071, 2015ZZ012, 2014ZA030); ;Medical Health Platform Plan Projects of Zhejiang Province (2015RCA020); Zhejiang Provincial Natural Science Foundation of China (LY16H030010).
Conflicts of interest
There are no conflicts of interest.
| > References|| |
Arnold M, Sierra MS, Laversanne M, Soerjomataram I, Jemal A, Bray F. Global patterns and trends in colorectal cancer incidence and mortality. Gut 2017;66:683-91.
Nam S, Choi YJ, Kim DW, Park EC, Kang JG. Risk factors for colorectal cancer in Korea: A population-based retrospective cohort study. Ann Coloproctol 2019;35:347-56.
Siegel RL, Miller KD, Fedewa SA, Ahnen DJ, Meester RG, Barzi A, et al
. Colorectal cancer statistics, 2017. CA Cancer J Clin 2017;67:177-93.
Li C, Zuo D, Liu T, Yin L, Li C, Wang L. Prognostic and Clinicopathological Significance of MUC Family Members in Colorectal Cancer: A Systematic Review and Meta-Analysis. Gastroenterol Res Pract 2019;2019:2391670. doi: 10.1155/2019/2391670.eCollection 2019.
Yang Y, Qu A, Wu Q, Zhang X, Wang L, Li C, et al
. Prognostic value of a hypoxia-related microRNA signature in patients with colorectal cancer. Aging (Albany NY) 2020;12:35-52.
Luo Y, Cheng B, Liu S, Xu L, Wu X, Sun S. Relationship between CpG island methylation phenotype, microsatellite instability phenotype and mutation of KRAS, NRAS, and BRAF genes in colorectal cancer. Int J Clin Exp Pathol 2019;12:1101-7.
Mint Sidi Ould Deoula M, Huybrechts I, El Kinany K, Boudouaya H, Hatime Z, El Asri A, et al
. Behavioral, nutritional, and genetic risk factors of colorectal cancers in morocco: Protocol for a multicenter case-control study. JMIR Res Protoc 2020;9:e13998.
Krzowska-Firych J, Lucas G, Lucas C, Lucas N, Pietrzyk Ł. An overview of human papillomavirus (HPV) as an etiological factor of the anal cancer. J Infect Public Health 2019;12:1-6.
Bychkov VA, Nikitina EG, Ibragimova MK, Kaigorodova EV, Choinzonov EL, Litviakov NV. Comprehensive meta-analytical summary on human papillomavirus association with head and neck cancer. Exp Oncol 2016;38:68-72.
Hardefeldt HA, Cox MR, Eslick GD. Association between human papillomavirus (HPV) and oesophageal squamous cell carcinoma: A meta-analysis. Epidemiol Infect 2014;142:1119-37.
Glenn WK, Ngan CC, Amos TG, Edwards RJ, Swift J, Lutze-Mann L, et al
. High risk human papilloma viruses (HPVs) are present in benign prostate tissues before development of HPV associated prostate cancer. Infect Agent Cancer 2017;12:46.
Malekpour Afshar R, Deldar Z, Mollaei HR, Arabzadeh SA, Iranpour M. Evaluation of HPV DNA positivity in colorectal cancer patients in Kerman, Southeast Iran Asian Pac J Cancer Prev 2018;19:193-8.
Bernabe-Dones RD, Gonzalez-Pons M, Villar-Prados A, Lacourt-Ventura M, Rodríguez-Arroyo H, Fonseca-Williams S, et al
. High prevalence of human papillomavirus in colorectal cancer in hispanics: A case-control study. Gastroenterol Res Pract 2016;2016:7896716.
Vuitton L, Jaillet C, Jacquin E, Monnien F, Heberle M, Mihai MI, et al
. Human papillomaviruses in colorectal cancers: A case-control study in western patients. Dig Liver Dis 2017;49:446-50.
Venazzi A, Swardfager W, Lam B, Siqueira JO, Herrmann N, Cogo-Moreira H. Validity of the QUADAS-2 in assessing risk of bias in Alzheimer's disease diagnostic accuracy studies. Front Psychiatry 2018;9:221.
Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, et al
. QUADAS-2: A revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011;155:529-36.
McGregor B, Byrne P, Kirgan D, Albright J, Manalo P, Hall M. Confirmation of the association of human papillomavirus with human colon cancer. Am J Surg 1993;166:738-40.
Cheng JY, Sheu LF, Meng CL, Lee WH, Lin JC. Detection of human papillomavirus DNA in colorectal carcinomas by polymerase chain reaction. Gut 1995;37:87-90.
Yu HG, Shun LB, Luo HS, Huang H, Yu BP, Yu JP, et al
. Deletion of the FHIT gene in human colorectal cancer is independent of high-risk HPV infection. Int J Colorectal Dis 2002;17:396-401.
Bodaghi S, Yamanegi K, Xiao SY, Da Costa M, Palefsky JM, Zheng ZM. Colorectal papillomavirus infection in patients with colorectal cancer. Clin Cancer Res 2005;11:2862-7.
Pérez LO, Abba MC, Laguens RM, Golijow CD. Analysis of adenocarcinoma of the colon and rectum: Detection of human papillomavirus (HPV) DNA by polymerase chain reaction. Colorectal Dis 2005;7:492-5.
Buyru N, Tezol A, Dalay N. Coexistence of K-ras mutations and HPV infection in colon cancer. BMC Cancer 2006;6:115.
Pérez LO, Abba MC, Dulout FN, Golijow CD. Evaluation of p53 codon 72 polymorphism in adenocarcinomas of the colon and rectum in La Plata, Argentina. World J Gastroenterol 2006;12:1426-9.
Damin DC, Caetano MB, Rosito MA, Schwartsmann G, Damin AS, Frazzon AP, et al
. Evidence for an association of human papillomavirus infection and colorectal cancer. Eur J Surg Oncol 2007;33:569-74.
Militello V, Trevisan M, Squarzon L, Biasolo MA, Rugge M, Militello C, et al
. Investigation on the presence of polyomavirus, herpesvirus, and papillomavirus sequences in colorectal neoplasms and their association with cancer. Int J Cancer 2009;124:2501-3.
Salepci T, Yazici H, Dane F, Topuz E, Dalay N, Onat H, et al
. Detection of human papillomavirus DNA by polymerase chain reaction and southern blot hybridization in colorectal cancer patients. J BUON 2009;14:495-9.
Gornick MC, Castellsague X, Sanchez G, Giordano TJ, Vinco M, Greenson JK, et al
. Human papillomavirus is not associated with colorectal cancer in a large international study. Cancer Causes Control 2010;21:737-43.
Liu F, Mou X, Zhao N, Lin J, Teng L, Xiang C. Prevalence of human papillomavirus in Chinese patients with colorectal cancer. Colorectal Dis 2011;13:865-71.
Sun ZQ, Wang HJ, Zhao ZL, Wang QS, Fan CW, Fang F. Significance of HPV infection and genic mutation of APC and K-ras in patients with rectal cancer. Asian Pac J Cancer Prev 2013;14:121-6.
Ranjbar R, Saberfar E, Shamsaie A, Ghasemian E. The aetiological role of human papillomavirus in colorectal carcinoma: An Iranian population- based case control study. Asian Pac J Cancer Prev 2014;15:1521-5.
Taherian H, Tafvizi F, Fard ZT, Abdirad A. Lack of association between human papillomavirus infection and colorectal cancer. Prz Gastroenterol 2014;9:280-4.
Tanzi E, Bianchi S, Frati ER, Amicizia D, Martinelli M, Bragazzi NL, et al
. Human papillomavirus detection in paraffin-embedded colorectal cancer tissues. J Gen Virol 2015;96:206-9.
Cheng L, Eng C, Nieman LZ, Kapadia AS, Du XL. Trends in colorectal cancer incidence by anatomic site and disease stage in the United States from 1976 to 2005. Am J Clin Oncol 2011;34:573-80.
Bailey CE, Hu CY, You YN, Bednarski BK, Rodriguez-Bigas MA, Skibber JM, et al
. Increasing disparities in the age-related incidences of colon and rectal cancers in the Unite States, 1975-2010. JAMA Surg 2015;150:17-22.
Thanikachalam K, Khan G. Colorectal Cancer and Nutrition. Nutrients 2019;11:164. doi: 10.3390/nu11010164.
Cai S, Li Y, Ding Y, Chen K, Jin M. Alcohol drinking and the risk of colorectal cancer death: A meta-analysis. Eur J Cancer Prev 2014;23:532-9.
Thrift AP, Gong J, Peters U, Chang-Claude J, Rudolph A, Slattery ML, et al
. Mendelian randomization study of body mass index and colorectal cancer risk. Cancer Epidemiol Biomarkers Prev 2015;24:1024-31.
Ma Y, Zhang P, Wang F, Yang J, Liu Z, Qin H. Association between vitamin D and risk of colorectal cancer: A systematic review of prospective studies. J Clin Oncol 2011;29:3775-82.
Gram IT, Park SY, Wilkens LR, Haiman CA, Le Marchand L. Smoking and risk of colorectal cancer may differ by anatomical subsite and sex. Am J Epidemiol 2020. pii: Kwaa005.
Fang JH, Yu XM, Zhang SH, Yang Y. Effect of smoking on high-grade cervical cancer in women on the basis of human papillomavirus infection studies. J Cancer Res Ther 2018;14:S184-9.
Eckmann JD, Ebner DW, Kisiel JB. Multi-Target Stool DNA Testing for Colorectal Cancer Screening: Emerging Learning on Real-world Performance. Curr Treat Options Gastroenterol 2020. doi: 10.1007/s11938-020-00271-5. Online ahead of print.
Zheng Y, Hua X, Win AK, MacInnis RJ, Gallinger S, Le Marchand L, et al
. A new comprehensive colorectal cancer risk prediction model incorporating family history, personal characteristics, and environmental factors. Cancer Epidemiol Biomarkers Prev 2020. pii: Cebp. 0929.2019.
Panteris V, Vasilakis N, Demonakou M, Kornarou E, Ktenas E, Rapti E, et al
. Alarming endoscopic data in young and older asymptomatic people: Results of an open access, unlimited age colonoscopic screening for colorectal cancer. Mol Clin Oncol 2020;12:179-85.
Aghakhani A, Hamkar R, Ramezani A, Bidari-Zerehpoosh F, Sabeti S, Ghavami N, et al
. Lack of human papillomavirus DNA in colon adenocarcinama and adenoma. J Cancer Res Ther 2014;10:531-4.
[Figure 1], [Figure 2], [Figure 3]