This chapter describes some of the fundamental pathophysiology of skin cancer. I explain why ultraviolet radiation (UVR) causes erythema, and why we can use erythema as a biomarker for DNA damage and skin cancer. You will never think about sunburn or freckles the same way again. As ever, you might want to check out the one minute video first, watch the longer video, and then consolidate these by reading the full explanations in the text. 


This chapter in one minute

Sunburn, DNA and skin cancer (9 minutes)

Ultraviolet radiation causes erythema and sunburn

The picture below shows the author’s back 24 hours after exposure to ultraviolet radiation (UVR). The striking physical sign is the redness, reflecting the large increase in dermal blood flow (erythema). Note that the reaction is a direct one, in that it is precisely limited to those areas exposed to the ultraviolet radiation — the lettering template ‘The Importance of Being Red’ shielded the areas of skin that remained pale. This reaction is therefore not a systemic one, nor it is one that spreads more than a few millimetres across the skin’s surface.

The erythema from UVR develops within hours of UVR exposure, and although it may not yet be visible to our eyes, it can be measured with a Doppler flow meter which is more sensitive than our eyes (to changes in blood flow). The erythema peaks at 8-24 hours, before subsiding over up to several days. The more intense the erythema, the sooner it is detectable, and the longer it takes to subside. After a large UVR exposure erythema can still be detected over a week later. If the erythema is substantial, there may be accompanying pain and an increased sensitivity to stimuli that do not normally cause pain such as light touch, but which are now ‘felt’ as pain (allodynia). If the UVR dose is very large, there may be oedema and even blister formation, reflecting widespread keratinocyte cell death. This is the sunburn reaction, familiar to many pale-skinned people.

Why and how does UVR cause sunburn?

The obvious questions to ask are why, and how does UVR  cause erythema, and what is its medical significance? And if I ask you to accept on trust, a fact I will later justify, why is it that those who develop the most erythema in response to ultraviolet radiation are most at risk of skin cancer?

The figure below shows the ability of ultraviolet radiation of various wavelengths to induce erythema  in human skin. The ‘X’ axis refers to the wavelength of ultraviolet radiation in nanometres, and the ‘Y’ axis the amount of erythema or DNA damage. Note that the ‘Y’ axis is on a logarithmic scale, spanning at least three orders of magnitude.

This sort of representation, linking some output (e.g. erythema) with wavelength is known as an action spectrum. What is shown above is the action spectrum for erythema.

The graph below adds in the amount of DNA damage caused by UVR of different wavelengths (shown in black). We are therefore showing two action spectra: that of erythema, and that of DNA damage.

There are two observations we can make about the graph.

  • First, both the amount of erythema and DNA damage caused by UVR varies strongly with wavelength. If this was not the case, we would expect a horizontal line.
  • Second, the two graph lines — one for DNA damage and the other for erythema — are remarkably similar. There is an obvious speculation to make: is sunburn due to DNA damage?

Establishing causality: the example of xeroderma pigmentosum (XP)

The action spectrum of DNA damage and UVR induced erythema is suggestive that the former causes the latter, but it is not proof.

Note, I am not suggesting that erythema is always due to DNA damage, or that causality is the other way round — otherwise blushing at a risqué joke or vigorous exercise might be considered a cause of skin cancer ( they are not, in case you were worried…..).

So how might we prove the relation between DNA damage and erythema is causal? As so often in medicine, insight comes from the close study of patients with a particular disease. In this instance, the disease is xeroderma pigmentosa (OMIM #278700). The image below shows a child with xeroderma pigmentosum.

By James Halpern, Bryan Hopping and Joshua M Brostoff – James Halpern, Bryan Hopping and Joshua M Brostoff: Photosensitivity, corneal scarring and developmental delay: Xeroderma Pigmentosum in a tropical country. In: Cases Journal 2008, 1:254 doi:10.1186/1757-1626-1-254 published under CC-by-2.0, CC BY-SA 3.0, Link

Although XP is very rare (~1:500,000) it is unforgettable once you have seen a patient with it. The disorder is inherited as an autosomal recessive and patients often present in early childhood with severe sunburn after trivial exposure to sunshine. These individuals are said to be photosensitive, that is they are abnormally sensitive to sunshine. Some patients with xeroderma pigmentosum are well over  one hundred times more sensitive to UVR induced erythema than normal individuals. In some individuals with xeroderma pigmentosum the sunburning episodes are less prominent than the large number of freckles that develop within the first few years of life on sun exposed sites. Some individuals with xeroderma pigmentosum are over 1000 times more likely to develop skin cancers, cancers that may that curtail their life, even in childhood.

The DNA repair defect in XP is specific to UVR induced damage

If XP patients go red after small doses of  sunshine, does that prove that it is a sensitivity to UVR? No, it doesn’t, and indeed there are other disorders in which patients develop erythema in response to visible light.

But in XP, we can show either in vivo or in cell culture, that the cells are sensitive to UVR, particularly the shorter wavelengths in the UVB range. This defect is specific to UVR and is not seen with other types of DNA damage such as X-rays (for example). Nor do these cells react abnormally to visible light.

the normal process of DNA repair after UVR exposure is grossly impaired in patients with XP.

Finally, if we fuse xeroderma pigmentosum cells in culture with cells that are known to have normal DNA repair systems, the sensitivity to UVR is corrected in the ‘XP’ cells — the normal cells are said to complement the abnormal xeroderma pigmentosum cells.

Whether in patients with XP or normal individuals, freckles and UVR induced erythema are a proxy for DNA damage.

Sunburn, DNA and cancer: the argument laid bare

The data I have presented on the action spectra for erythema and DNA damage, and the example of xeroderma pigmentosum allow us to argue as follows:

  • UVR causes DNA damage
  • DNA damage varies with the wavelength of UVR
  • DNA damage causes erythema (but there are other causes of erythema, too)
  • failure to repair UVR induced DNA damage produces a phenotype characterised by increased UVR induced erythema, and freckling in the short term, and skin cancer in the long term.
  • repair of DNA damage accounts for the reduction in UVR induced erythema over time (this is why sunburn in not permanent).
  • failure to repair DNA damage after UVR is causally related to increased skin cancer rates

This all implies that we can use the acute response to UVR including erythema (and the development of freckles) as a proxy or intermediate measure for the chronic consequence, skin cancer. But lets put it another way:

erythema or freckles, are a biomarker for skin cancer.

What part of the UVR in sunshine causes erythema and skin cancer?

Before leaving the subject of the relation between UVR and erythema there is one further point that needs consideration.

Examination of the action spectrum for both erythema and DNA damage showed that shorter wavelengths were more potent at inducing erythema. What happens in natural sunshine? Is the erythema due to shorter wavelengths in the UVB range or to other longer (UVA) wavelengths? And what is the distribution of UVR in natural sunlight? A short(ish) aside on the physics of UVR is required — no equations, needed 🙂 .  It will consolidate some of the above, too.

Ultraviolet radiation: physics 101 (for medics) 20 mins

Ultraviolet radiation (UVR) is defined as electromagnetic radiation between 400nm and 100nm. As dermatologists we  ignore radiation below ~ 290nm, because it is blocked by the atmosphere (we are not spacemen).

Conversely, radiation of wavelengths greater than 400nm ( and up to 800nm) is visible radiation (‘light’). For example radiation at 420nm will be visible and appear violet, whereas radiation at 360 is invisible to the human eye.

One simple way to classify ultraviolet radiation is to split the UVR component of electromagnetic radiation by wavelength into three further wavebands :

  • UVA (400-320nm),
  • UVB (320-290nm),
  • UVC (290-100nm).

(the exact figures used sometimes vary a little depending on scientific context e.g 315nm instead of 320nm, but we can ignore these details)

This classification is widely quoted — and is useful — but remember it is slightly artificial, as activity is usually a continuous function of wavelength. For example,  UVR of 325nm and 365nm may  vary greatly in biological activity, despite both wavelengths being described as UVA. The figure below makes this clear.


The figure above shows both the action spectrum for erythema (as we have seen before, in red) and the amount of ultraviolet radiation by wavelength falling on the earths surface again measured on a logarithmic scale (‘Y’ axis) in blue. Also shown are the wavebands (UVC, UVB, and UVA); and the green dotted line marks the boundary between UVB and UVA.

What is immediately apparent is that most of the UVR in sunlight is in the UVA range with only a small component falling within the UVB range (note, once again, the ‘Y’ axis is logarithmic).

Can we therefore assume that a sunburn due to natural sunshine is mainly due to UVA? No, we cannot.

Indeed, given how much more potent the shorter wavelengths are at inducing erythema we can say that most (~80%) of the sunburn is actually due to the UVB component of natural sunlight even though on a joule to joule basis UVA comprises most (>90%) of the incident radiation. (Check you follow the logic, but not the details of the two numbers I have quoted.)

What wavelengths are most important in causing skin cancer? Here there are differences of opinion and some uncertainties. As we learned earlier, DNA damage mirrors erythema, and shorter wavelengths (UVB) are more potent than longer wavelengths (‘UVA’) at causing DNA damage. For the most common types of skin cancer (‘non-melanoma skin cancer’), it seems likely that UVB is the most important causative UVR waveband. For melanoma, there is less consensus, and for reasons I will not go into here, many believe that UVA is important in addition to UVB (note the phrase, ‘in addition to UVB’).

Some annoying little details

  1. Xeroderma pigmentosum. Xeroderma pigmentosum is conceptually important, but too rare for medical students to know all the details. There are a couple of ‘gaps in my logic’ in the above account that may bug some of you, so here goes….Xeroderma pigmentosum can be caused by mutations in a variety of different genes all important for DNA repair. The severity of the clinical phenotype varies depending on which gene is affected, and patient UVR exposure. Some types of DNA repair are more important for genes that are being actively transcribed and others important for genes that are not being actively transcribed. If you affect the former, an abnormal erythemal response to UVR predominates; if you affect the latter pathways, the (grossly) abnormal freckling phenotype may predominate. In many patients both phenotypes are very obvious.
  2. Why is UVR (<400nm) ‘ultra’ when the wavelengths are ‘smaller’ than visible violet light (>400nm)? Frequency (Hz) and wavelength (nm) are inversely related and physicists (who really understand the world…) often prefer using frequency rather than wavelength.
  3. It may seem strange to imagine that people may be sensitive to visible light, but clinical dermatology is wonderfully full of strange phenomena, and some patients with photodermatoses are sensitive to visible light. For instance, in a condition called solar urticaria, some patients may come out in a nettle sting like weals upon exposure to visible  light (or UVR). 
  4. What are freckles? Freckles are areas of focal accumulation of melanin. We will revisit them in the next chapter. They are also of great interest to leprechauns.

Questions: Sunburn, DNA and cancer

  1. In sunshine which waveband predominates, UVA or UVB?
  2. In sunshine which waveband contributes most to sunburn, UVA or UVB?
  3. Are patients with xeroderma pigmentosum abnormally sensitive to X-irradiation as well as UVR?
  4. Does UVR of less than 310nm penetrate the atmosphere? 
  5. When does UVR induced erythema peak?
  6. Why does UVR induced erythema subside over time?
  7. Name two biomarkers of skin cancer that follow acute UVR exposure
  8. The term action spectrum refers only to DNA damage and erythema. True or false?
  9. Why is sunburn so well demarcated?
  10. UVA is the main causative waveband for skin cancer. Discuss.

A PDF  containing the above questions and the answers is here . The video talkover below goes through the questions and answers in greater depth.


Skincancer909 by Jonathan Rees is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Where different rights apply for any figures, this is indicated  in the text.