3.3.1

DIY Fluorescence microscope

Fluorescence microscopy is the technique of choice in many science laboratories to image biological preparations. Labelled with fluorescent indicators, biological structures can be visualized over the homogeneous background with high contrast. Traditionally, fluorescent labelling required chemical stains, most of which are lethal to biological samples ⁹. In more recent years, the discovery of Green Fluorescent Protein (GFP)¹⁰ enabled scientists to perform fluorescent imaging in living samples, and even entire organisms ¹¹ through genetic engineering and other gene targeting approaches ¹². GFP was further modified to obtain fluorescent markers of different colors ¹³, as well as functional markers sensitive to transient cellular conditions, such as rising calcium waves ^{14, 15, 16} or voltage differences across the cell membrane ^{17, 18}. A newly introduced technique allowing tracking individual fluorescent molecules was recognized with the 2014 Nobel Prize in Chemistry. Unfortunately, despite the fact that fluorescent microscopes are widely used in research laboratories, their elevated cost prevents their use in most schools and educational institutions.

We modified the basic structure of the DIY compound microscope described above for trans-fluorescent imaging (Figure 4A). The white LED (Figure 1C) was replaced with a blue LED (450-460 nm) for sample illumination.

Figure 4:

DIY fluorescent microscopy A) diagram of the optical configuration used for fluorescence imaging and photo of the DIY microscope configured accordingly; B) diagram of larval drosophila nervous system, with the central nervous system (CNS) highlighted; C) fluorescence images of larval drosophila CNS: left captured using Raspberry Pi Camera, right captured using cell phone camera (Samsung Galaxy S6).

Additionally, a long-pass green filter (520 nm) was introduced in front of the camera. We used the DIY fluorescent microscope to image preparations of larval fruit fly (D. melanogaster) dissected along the medial line to expose the brain and mouth hooks (Figure 4B). This transgenic organism expressed GFP in its neurons, permitting detailed imaging of its nervous system. The images in figure 4C-D show the two lobes of the central brain. Figure 4C was captured using the low-cost Raspberry Pi camera while figure 4D was captured using a Samsung Galaxy S6 camera. The fine nerve fibers emerging from the central brain (5-10 μm) are visible in both images. Although the contrast is higher in the Raspberry Pi camera image (Figure 4C), the use of the phone camera has an advantage in terms of diffusion of the device (almost all students have a phone), as well as outreach and communication purposes. For example, phone images can be instantaneously shared through social media.

3.3.2

DIY infrared microscope for correction of chromatic aberration

One of the disadvantages of white light imaging is blurring and distortion of the image due to chromatic aberrations ¹⁹. This is caused by the unequal bending of the different wavelengths composing white light by lenses (Figure 5A). One method for avoiding this problem is to illuminate the sample with a narrower range of wavelengths.

Figure 5:

Infrared DIY microscope A) Diagram explaining chromatic aberration: parallel rays of different wavelengths (colors) are focused by a convex lens into slightly displaced points along the optical axis, deteriorating the resolution of the image and creating the characteristic “rainbow halo”; B) images of Golgi-stained mouse neurons captured using the DIY microscope: left using broadband white light illumination, right using IR illumination; C) mouse cortical neurons on a Golgi-stained brain slice imaged with the DIY IR microscope; D) compound eye of the freshwater crustacean Daphnia magna imaged with the DIY IR microscope; E) images of the aquatic plant Elodea: left using broadband white light illumination, right using IR illumination.

With the DIY optical kit assembled in a compound microscope configuration, we substituted the white LED with an IR one (850 nm). Because IR light is not visible to the human eye, samples can only be viewed through a camera when illuminated in this way. Most commercial cameras have an integrated IR-blocking filter. However, cameras intended for use at night or in low-light conditions use IR light and so do not have an IR filter. We used one of these cameras, the low-cost Raspberry Pi Noir camera, as our detector. The DIY microscope so modified was used to demonstrate the benefits of narrow spectrum illumination for high-resolution imaging.

The results are shown in Figure 5B and E, where the same objects are imaged with broadband white light illumination (left) and narrow-spectrum IR illumination (right). Despite the fact that the low-cost plano-convex lenses and the aspherical objective lens used for the DIY microscope had no correction for chromatic aberration, the narrower illumination spectrum made it possible to capture sharper images at high magnification. Images obtained with white light illumination showed colored halos around objects, hallmarks of chromatic aberration. Replacing the illuminator with a narrow spectrum IR source eliminated chromatic aberration, allowing for sharply resolved images. Most notably, individual cells in the Elodea leaf could be seen clearly, as well as granules and other structures within the cells (Figure 5E, right). Similarly, individual dendrites (1-10 μm) were clearly resolved in the infrared IR image of mouse neurons (Figure 5B, right and Figure 5C), and single ommatidia were visible in the eye of the freshwater crustacean Daphnia magna (Figure 4B, right).

3.3.3

Functional model of the human eye

We used the DIY optical kit to build a functional model of the human eye. The structure of the eye allows for a complex system of functions to coexist and create the images that we see ²⁰. Optically, however, it can be simplified and represented by a combination of two lenses, the cornea and the crystalline lens, forming an image in the retina ²¹, the neuronal tissue at the back of the eye that works as our “natural light sensor” (Figure 6A). The cornea allows for most of the focusing power, while the crystalline lens provides fine focus adjustments and forms the final image on the retina. Additionally, the crystalline lens confers the power of forming images of objects at different distances ²⁰. Here we show how the DIY optical kit can be used to demonstrate these fundamental properties of the human eye.

Figure 6:

DIY functional model of the human eye. A) The main optical components of the eye are illustrated: cornea, lens and retina. Note that the image of the object on the retina is inverted, according to the rules of image formation. B) A normal functioning eye focusing an object at a distance modeled with the DIY optical kit. From left to right: LED light source, object (a face drawn on a microscope slide), cornea lens (FL = 70 mm), crystalline lens (FL = 400 mm), and retina (white screen). C) Front view of the DIY optical kit aligned to model a normal eye focusing a faraway object on the retina. The crystalline lens is modeled with a FL = 400 mm lens. D) Front view of the DIY optical kit aligned to model a normal eye focusing a closeby object on the retina. The crystalline lens is modeled with a FL = 70 mm lens.

We modeled the basic structure of an eye using two lenses, corresponding to the cornea and crystalline lens, forming the image of an object on a paper screen representing the retina (Figure 6B). In our model, the cornea and crystalline lens remained stationary at a distance of 20 mm from each other, and 95 mm from the retina, mimicking the structure of a human eye where distances are determined by the anatomy and do not change. We began by modelling the cornea of a healthy eye with a convex lens of 70 mm focal length (FL). We then mimicked the crystalline lens using a second convex lens of FL = 400 mm. The object itself was a drawing painted on a microscope slide positioned 100 mm away from the cornea, and illuminated by a LED light source (Figure 6B). The image was focused by the two lenses onto the screen/retina (Figure 6C). Note that, as in a real eye, the image on the retina was upside down.

Our eye is able to form images of objects both closeby and far away. This is made possible by the ciliary muscles, that stretch and compress the crystalline lens thus modifying its effective focal length. This adjustment allows the images of objects at different distances to be focused in the retina. In our model, we mimicked the action of the ciliary muscles compressing the crystalline lens to focus on closeby objects by substituting the 400 mm FL crystalline lens (Figure 6C) with a thicker one, of 70 mm FL (Figure 6D). As expected, the object in focus on the retina was now closer to the “eye” (42 mm from the cornea vs. 100 mm in the faraway configuration). The size of the image on the retina also increased by a factor of 2, as it occurs in reality, where nearby objects appear larger than faraway ones.

3.3.4

Near and Farsightedness

Common eye defects such as myopia (nearsightedness) and presbyopia (age-related farsightedness), can also be easily modeled with the DIY microscope. Myopia often occurs when the eyeball is too long, causing the image of faraway objects to be formed in front of the retina. As a result, faraway objects appear blurry, contrary to closeby ones that appear sharp. Myopia affects around 30% of the U.S. population (American Optometric Association, 2018), and can be corrected by the use of concave glasses. To model this eye defect, we first aligned the DIY eye model in the “faraway” configuration, as described above (Figure 6B-C). We then moved the screen/retina backward by an additional 55 mm, from the original configuration, to mimic a longer eyeball (Figure 7A). As expected, the image on the screen/retina became blurry (Figure 7B). Because the crystalline lens is already in its most relaxed configuration (long FL) and cannot be extended further by the ciliary muscles, the eye is not able to naturally correct this defect. However, inserting a concave lens (FL = -45 mm) between the object and the cornea, mimicking a corrective glass or contact lens (Figure 7C), represtinated a sharp image in the screen/retina (Figure 7D).

Figure 7:

Eye defects. A) DIY optical kit arranged to model the eye of an individual affected by myopia, focusing on a faraway object. Note that the retina is farther away from the crystalline lens (150 mm vs. 95 mm in Figure 6B), to mimic the longer eyeball, typical of many forms of myopia. B) Front view of the screen/retina from A: as expected, there is no sharp image on the retina. C) DIY optical kit aligned as in A (myopia), but with the addition of a concave lens (FL = -45 mm) in front of the cornea to mimic correction glasses. D) Front view of the screen/retina from C: as expected, the image, an upside down smiley face, is back in focus on the retina. E) DIY optical kit arranged to model the eye of an individual affected by presbyopia, focusing on a closeby object. The crystalline lens is modeled by a lens with long FL = 400 mm: this reflects the inability of the ciliary muscle to compress the hardened crystalline lens, and reduce its effective FL, typical of presbyopia. F) Front view of the screen/retina from E: as expected, there is no sharp image on the retina. G) DIY optical kit aligned as in E (presbyopia), but with the addition of a convex lens (FL = 30 mm) in front of the cornea to mimic correction glasses. H) Front view of the screen/retina from G: as expected, the image, an upside down smiley face, is back in focus on the retina.

Presbyopia occurs when the image of nearby objects is formed at the back of the retina because of the hardening of the crystalline lens, which makes it impossible for the ciliary muscles to compress it and reduce its effective focal length. Presbyopia is a natural consequence of aging of the eye, with symptoms appearing progressively in people over 35 years old (National Eye Institute, 2010). To model this eye defect, we first aligned the DIY microscope in a “closeby” configuration (Figure 6D). We then substituted the short FL crystalline lens (FL = 70 mm) with a longer FL lens (FL = 400 mm, Figure 7E). This caused the image on the screen/retina to become blurry (Figure 7F). The effect could be reversed by introducing a convex lens (FL = 30 mm) between the object and the cornea, mimicking reading glasses (Figure 7G-H).

3.3.5

Age-Related Macular Degeneration

Subsequently, we used the DIY optical kit to model eye diseases. We chose to work on Age-Related Macular Degeneration (AMD), a debilitating disease affecting 1.75 million people in the U.S. ²². Since AMD primarily occurs in people over 60 years of age, many of the elderly find themselves losing independence as their eyesight deteriorates. Through the progression of AMD, the central and most sensitive area of the retina, called the macula, is impeded in performing its functions ²³. There are two main forms of AMD, non-neovascular or dry (D-AMD) and neovascular or wet (W-AMD). In D-AMD, tissue debris, called drusen, accumulate in the macula until scattered blind spots occur in the central vision. In W-AMD, the formation of drusen further triggers the development of new blood vessels in the macula, causing blood leakage and resulting in even more pronounced blind spots ²⁴.

We demonstrated the effects of AMD using our DIY model of the human eye. Drawing a black dot in the center of the screen/retina (macula), we mimicked the obstruction caused by drusen and growing blood vessels in a patient’s central visual field (Figure 8 A-B). This is a very simple, yet optically accurate depiction of AMD, that we used for educational purposes during public science outreach events.

Figure 8:

Age-related Macular Degeneration and the Implantable Miniature Telescope. A) Image of the object (smiley face painted on a microscopy slide) as formed by a “normal eye” aligned as in Figure 6B. B) Same object imaged by the eye of a patient affected by AMD: the drusen/outgrown vasculature are modeled by a black spot in the center of the screen/retina. Notice how the central visual field is completely obstructed. C) Model of an eye with Implantable Miniature Telescope (IMT): the crystalline lens was removed, as in real-life surgery. The IMT was modeled by a convex lens (FL = 70 mm) and a concave lens (FL = -45 mm), aligned in a Galilean telescope configuration (inter-lens distance 35 mm). D) Front view of the screen/retina from E: the image on the retina is enlarged by 1.6 times compared to B. As a result, the relative size of the drusen (black spot) is decreased, allowing for partial recovery of central vision. E) Model eye with IMT aligned as in C, but object in closeby configuration (42 mm from cornea). F) Front view of the screen/retina from E: as expected, there is no sharp image on the retina. G) Model eye with IMT aligned as in E (closeby configuration), but with the addition of a convex lens (FL = 30 mm) in front of the cornea to mimic correction glasses. H) Front view of the screen/retina from G: as expected the enlarged image is back in focus on the retina.

We then moved on to show how technology can be used to relieve the symptoms of AMD. In fact, although some drugs are currently available to reduce vasculature development in W-AMD ^{25, 26, 27, 28}, there is no pharmacological treatment for D-AMD ^{28, 29, 23}. As an alternative, an Implantable Miniature Telescope (IMT) can be inserted into the eye, after removal of the crystalline lens, to recover partial vision ^30–32. When implanted in the eye, the IMT creates a “zoomed in” image of objects in the central visual field, thus reducing the relative size of the blind spots created by drusens, which are not magnified since they reside physically on the retina. The IMT is designed for use in only one eye because the increased size of the central visual field comes at the expense of peripheral vision. After appropriate physical therapy, however, the brain is able to use the implanted eye for central vision, and the other eye for peripheral vision, represtinating quasi-normal sight ³⁰.

We mimicked the action of the IMT by using a variation of our DIY eye model, where the IMT itself was modeled by a convex lens (FL = 70 mm) and concave lens (FL = -45 mm) aligned to form a Galilean telescope (inter-lens distance 35 mm, equal to the sum of the focal lengths, Figure 8C). During surgery, the crystalline lens is removed and the IMT is implanted in its place. Accordingly, in our model, the crystalline lens was removed, and the telescope was added between the cornea and the screen/retina (Figure 8C). Since the original IMT was designed to allow perfect vision of faraway objects, we aligned the DIY microscope as in Figure 6B (faraway configuration). As expected, the image produced by the telescope on the retina was enlarged by 1.6 times (Figure 8D). This is in agreement with what happens in real life, where the insertion of the IMT would allow patients to see a magnified version of objects in their central visual field. As in real life, our model showed that the relative size of the blind spot was reduced in the final, enlarged image (Figure 8D).

Patients implanted with the IMT lose their natural ability to focus on closeby objects, since the crystalline lens is removed. They correct the problem by using reading glasses. As expected, in our model, if the object was placed in the closeby configuration (42 mm from the cornea), no sharp image could be formed in the retina (Figure 8 E-F). The use of a convex lens (FL = 30 mm) between the object and the cornea to mimic reading glasses repestinated normal vision (Figure 8 G-H). Note that once again, the image of closeby objects (Figure 8H) is larger than that of faraway objects (Figure 8D). In AMD patients, this further reduces the relative size of the drusen.